Experimental Innovations in Surface Science: John T. Yates JR

John T. Yates Jr.
Experimental
Innovations in
Surface Science
A Guide to Practical Laboratory
Methods and Instruments
Second Edition
Experimental Innovations in Surface Science
John T. Yates Jr.
Experimental Innovations
in Surface Science
A Guide to Practical Laboratory Methods
and Instruments
Second Edition
123
John T. Yates Jr.
Department of Chemistry
University of Virginia
Charlottesville, VA
USA
ISBN 978-3-319-17667-3 ISBN 978-3-319-17668-0 (eBook)

DOI 10.1007/978-3-319-17668-0
Library of Congress Control Number: 2015941351
Springer Cham Heidelberg New York Dordrecht London

© Springer International Publishing Switzerland 1998, 2015
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made.
Printed on acid-free paper
Springer International Publishing AG Switzerland is part of Springer Science+Business Media

(www.springer.com)
Preface to the Second Edition
A number of new experimental innovations have been added in over 50 new

chapters. I thank each of the innovators for their contributions which are
acknowledged in the text. I very much appreciated the graphical work of Dr. Ana
Stevanovic who skillfully converted hand drawings to computer graphics. I also
deeply appreciate the assistance of Dr. Zhen Zhang in assembling the new and
former material to produce a well-organized second edition. And finally, I wish to
thank Dr. Winfried Heichler and his son Jan Heichler for important high-resolution
optical scanning of the first edition, allowing its fusion with new material in the
second edition. I also thank the Department of Chemistry at the University of
Virginia for financial support of the graphics effort. And finally, as in the first
edition, I thank my loving and caring wife, Kerin Yates, for her continued support
and constructive criticism of the new book.
Virginia John T. Yates Jr.
v
Preface to the First Edition
The modern era of surface science, developed in the last 30 years, features hundreds
of new physical and chemical measurement methods such as low-energy electron
diffraction and Auger electron spectroscopy, which form a vital foundation for the
field. These measurement methods constitute the basis for a number of books used
by beginning students as their introduction to experimental research.
This book is not about this particular experimental core of the field. Instead, it is
about those practical laboratory methods used throughout the field, but seldom
communicated coherently in writing and drawings to beginning students. These
practical methods of design, construction, and measurement also form their own
foundation for making possible significant research. They constitute a gold mine of
clever ideas and “tricks of the trade” about how to use basic materials and com-
ponents effectively for research and measurement in surface science.
Approximately 250 practical methods are described in the book in enough detail
that one could design a new experimental device from the book alone.
This book will be useful to beginning research workers in surface science as well
as to those who work with similar methods in fields such as semiconductor tech-
nology, materials science, and physical electronics.
The contents of the book have been selected from unpublished information
supplied by more than 100 workers in the field as well as from the literature, dating
back to the mid-1960s. I have attempted to communicate the core ideas from
published and unpublished accounts of research in easy-to-comprehend pictorial
form, believing that pictures are the most effective mode for communication of
information of this type. Thus, the reader can learn much just by looking at the
pictures. The text associated with each drawing gives more details. In many cases,
the text and its citations indicate the originators and subsequent contributors to the
innovation, but in other cases the origin of the key ideas is unknown, making
accurate historical citations impossible.
There are some topics that have not been addressed in the book, primarily
because they are currently highly specialized and fast-moving activities that are best
learned by advanced research students who have previously mastered some of the
topics covered in this book. Untreated topics include many details of the scanning
vii
viii Preface to the First Edition
tunneling microscope (STM), complex molecular beam techniques, and complex

laser-based measurements.
Updated information on United States and international suppliers of materials
and components used in surface science research is featured at the end of the book.
Here the reader will find a guide to product sources as well as addresses and other
communications information for all companies mentioned in the text footnotes as
well as many other company listings. I thank Dr. Jun Yoshinobu and Dr. Joachim
Ahner for help with these compilations.
I very much appreciate the involvement of many friends and colleagues who
have contributed to this book by sending me accounts of their innovative methods.
This includes many of my own students and postdoctoral who have developed
some of the experimental methods described here. I also thank the publishers of the
Journal of Vacuum Science and Technology for permission to use many published
contributions dating back to 1964. These were organized for me by Professor
Michael Bozack and Mr. Kenneth W. Bryant at Auburn University, to whom I am
greatly indebted, and were invaluable in writing a portion of this book. I thank Dr.
Fred Dylla and Mrs. Maria Taylor for their encouragement during work with AIP
and with Springer-Verlag as the book progressed. Mrs. Marge Augenstein and Mrs.
Mary Beth Merenick supplied accurate assistance in organizing the collection and
verification of information for the book at the University of Pittsburgh, and in
checking the accuracy of the citations. Ms. Hana Novak took my hand-drawn
mechanical drawings and skillfully converted them to computer graphics. The book
would not have been produced without the financial assistance of the Alexander
von Humboldt Stiftung, which made possible my stay in the Department of Physics,
Philipps University, Marburg, Germany. I am indebted also to the University of
Pittsburgh for a coincident sabbatical leave to write the book. I wish to thank my
many German friends for their hospitality, particularly Dr. Heinz J. Jaensch and
Professor Dieter Fick for their friendship during my seven months in Marburg.
I also thank my parents, Mr. and Mrs. J. Thomas Yates, Sr., for their hospitality in
Hagerstown, Maryland, during the completion of the book. And finally, and most
importantly, I thank my wife, Kerin, for all of her support. Without her under-
standing and constructive criticism, writing and illustrating the book would have
been impossible.
March 1997 John T. Yates, Jr.

R.K. Mellon Professor
Department of Chemistry
and Department of Physics and Astronomy
Director, Surface Science Center
University of Pittsburgh
Contents
Part I Vacuum System Technology
1 Motion in Vacuum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 3

1.1 Crystal Manipulator with Six Degrees
of Motional Freedom. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Rotation in Ultrahigh Vacuum . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Crystal Positioning Using Capacitance . . . . . . . . . . . . . . . . . 11
1.4 UHV Rotary Manipulator with Arcsec Resolution . . . . . . . . . 13
1.5 Solenoid Driven Linear Motion Device . . . . . . . . . . . . . . . . 15
1.6 An Inexpensive Linear Motion Device. . . . . . . . . . . . . . . . . 15
1.7 Translating Auger Spectrometer . . . . . . . . . . . . . . . . . . . . . 16
1.8 Simple Multiple Motion Manipulator . . . . . . . . . . . . . . . . . . 18
1.9 Simple Device for Small Rotations in Ultrahigh
Vacuum—Grating Application . . . . . . . . . . . . . . . . . . . ... 18
1.10 Turntable Rotation in Ultrahigh Vacuum . . . . . . . . . . . . ... 20
1.11 Small Motions in UHV Systems Using Shape Memory
Effect Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.12 Rotary Shutter Device Driven by Magnetic Eddy Current . . . 22
1.13 High-Speed Ultrahigh Vacuum Motor . . . . . . . . . . . . . . . . . 24
1.14 High-Speed Rotary Feedthrough for UHV Operation . . . . . . . 25
1.15 Self-lubricating Bearings in UHV . . . . . . . . . . . . . . . . . . . . 26
1.16 Lubrication for Heavy Sliding Loads
in Ultrahigh Vacuum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
1.17 Simple UHV Bearing for High-Speed Shafts . . . . . . . . . . . . 29
1.18 Flange-Mounted UHV Variable Aperture . . . . . . . . . . . . . . . 30
1.19 Linear Motion Platform (LMP) . . . . . . . . . . . . . . . . . . . . . . 31
1.20 Vacuum Trolley Transport System . . . . . . . . . . . . . . . . . . . 33
1.21 Thermally Compensated STM with Repeatable
Sample Positioning . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 34
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 36
ix
x Contents
2 Sample Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 39

2.1 Sample Transfer from High Pressure to Ultrahigh
Vacuum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 39
2.2 Novel Sample Transfer from UHV Chamber
to External Cell . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 45
2.3 UHV Sample Transfer Device with Low-Temperature
Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
2.4 Internal Cup High-Pressure Cell . . . . . . . . . . . . . . . . . . . . . 49
2.5 High-Pressure Transfer Cell . . . . . . . . . . . . . . . . . . . . . . . . 50
2.6 Sample Transfer with Disconnect-I . . . . . . . . . . . . . . . . . . . 52
2.6.1 Transfer in Standard UHV Systems. . . . . . . . . . . . . . 52
2.6.2 Transfer in STM Systems . . . . . . . . . . . . . . . . . . . . 54
2.7 Sample Transfer with Disconnect-II. . . . . . . . . . . . . . . . . . . 54
2.8 Vibration-Free Translatable Device . . . . . . . . . . . . . . . . . . . 55
2.9 Mirror Port for Enhanced Sample Transfer . . . . . . . . . . . . . . 56
2.10 UHV Access Port . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3 Electrical Connections. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.1 Tungsten Ribbon—Attachment to Power Leads . . . . . . . . . . 61
3.2 Wrap Connection—Tungsten to Tungsten . . . . . . . . . . . . . . 62
3.3 Cold-Formed Wire Connector . . . . . . . . . . . . . . . . . . . . . . . 64
3.4 Ultrahigh Resistance Vacuum Feedthroughs . . . . . . . . . . . . . 64
3.5 Sliding Metal Electrical Contact . . . . . . . . . . . . . . . . . . . . . 66
3.6 Low-Profile Electrical Lead Clamp . . . . . . . . . . . . . . . . . . . 67
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4 Pumping and Trapping. . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 69

4.1 Water Aspirator/Sorption Pump Combination for Efficient
UHV System Evacuation . . . . . . . . . . . . . . . . . . . . . . . ... 69
4.2 Ballast Pumping—Vibration Free . . . . . . . . . . . . . . . . . ... 71
4.3 The Use of Appendage Titanium Sublimation Pumps
for Pumping Low Gas Loads in Ultrahigh Vacuum . . . . . . . . 72
4.4 Pressure-Equalizing Device in Complex Systems . . . . . . . . . 72
4.5 Diode Ion Pump Performance in He Pumping. . . . . . . . . . . . 75
4.6 Vacuum Applications of Metal Foams . . . . . . . . . . . . . . . . . 75
4.7 Cleaning of Ion Pumps by Chemical Etching . . . . . . . . . . . . 78
4.8 An Efficient Liquid Nitrogen Trap . . . . . . . . . . . . . . . . . . . 79
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
5 Bakeout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
5.1 Bakeout of Metal Ultrahigh Vacuum Systems. . . . . . . . . . . . 83
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Contents xi
6 Behavior of UHV Systems . . . . . . . . . . . . . . . . . . . . . . ........ 87

6.1 Wall Passivation in Stainless Steel Ultrahigh
Vacuum Systems . . . . . . . . . . . . . . . . . . . . . . . ........ 87
6.2 Minimizing Wall Reactions in Ultrahigh Vacuum
Systems—Gas Dosers . . . . . . . . . . . . . . . . . . . . ........ 89
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........ 91
7 Mechanical Action on Samples . . . . . . . . . . . . . . . . . . . . . . . . . . 93

7.1 Simple UHV Crystal Cleaver . . . . . . . . . . . . . . . . . . . . . . . 93
7.2 Piezoelectric Fatigue Apparatus for UHV Operation . . . . . . . 95
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
8 Gasket Seals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
8.1 Gasket Seals for Ultrahigh Vacuum Systems . . . . . . . . . . . . 99
8.2 Cryogenic Gasket Seals . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
8.3 Copper Gasket Removal Devices . . . . . . . . . . . . . . . . . . . . 102
8.4 Unconventional Compression Seals for Ultrahigh Vacuum . . . 103
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
9 Leak Repairs and Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

9.1 Coaxial Pumped He Leak Detection Probe . . . . . . . . . . . . . . 107
9.2 Temporary Leak Sealing of Welded Bellows . . . . . . . . . . . . 108
9.3 Atmospheric Permeation—Viton O-Ring . . . . . . . . . . . . . . . 108
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
10 Specialized UHV Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

10.1 Aluminum Ultrahigh Vacuum System . . . . . . . . . . . . . . . . . 113
10.2 Surface Electrochemistry Apparatus. . . . . . . . . . . . . . . . . . . 114
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Part II Mechanical Fabrication Techniques
11 Grids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
11.1 Hemispherical Grids—Formation and Piercing . . . . . . . . . . . 119
11.2 Making Flat Mesh Grids of Large Size . . . . . . . . . . . . . . . . 120
11.3 Grid Fabrication Techniques—Conical Grid . . . . . . . . . . . . . 121
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
12 Conductive Coatings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

12.1 Deposition of Electrically Conductive SnO2 Films. . . . . . . . . 123
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
xii Contents
13 Phosphor Screens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

13.1 Sedimentation Method for Depositing Phosphor Screens . . . . 127
13.2 Dusting Method for Coating Phosphor Screens . . . . . . . . . . . 129
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
14 Thermionic Emitters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 133

14.1 Thoriated Thermionic Emitters . . . . . . . . . . . . . . . . . . . ... 133
14.2 Lanthanum Hexaboride Thermionic Emitters—Deposition
on Metals. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 135
14.3 Directly Heated Lanthanum Hexaboride
Thermionic Emitters . . . . . . . . . . . . . . . . . . . . . . . . . . ... 137
14.4 Thermionic Emitter Mounting. . . . . . . . . . . . . . . . . . . . ... 139
14.5 Indirectly Heated Cathodes for High-Temperature
Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 141
14.6 Filament Power Supply Tester . . . . . . . . . . . . . . . . . . . ... 141
14.7 Bayard-Alpert Ionization Gauge with Corrections
for External Radiation and External Electron Flux Effects ... 142
14.8 Replacing Filaments in Glass Bayard-Alpert Gauges . . . . ... 145
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 146
15 Shielding . . . . . ....................... . . . . . . . . . . . . . . 149

15.1 Magnetic Shielding in Ultrahigh Vacuum. . . . . . . . . . . . . . . 149
15.2 Electrical Isolation of UHV Components . . . . . . . . . . . . . . . 151
References. . . . . ....................... . . . . . . . . . . . . . . 152
16 Single Crystal Fabrication/Orientation . . . . . . . . . . . . . . . . . . . .. 155

16.1 Making Small Au Single Crystals from Au Wire for STM
and Other Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 155
16.2 Single Crystal Orientation, Grinding, and Polishing . . . . . . .. 156
16.3 A Simple Goniometer for Cutting Single Crystals . . . . . . . .. 159
16.4 Ion Beam Polishing of Crystals
to Subnanometer Roughness . . . . . . . . . . . . . . . . . . . . . . .. 160
16.5 Crystal Optical Alignment . . . . . . . . . . . . . . . . . . . . . . . .. 161
16.6 Measurement of the Angles of Incidence
in a LEED Experiment . . . . . . . . . . . . . . . . . . . . . . . . . .. 162
16.7 Measurement of Angle of Incidence of an Electron Beam
on a Single Crystal . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 163
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 165
17 Tip Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167

17.1 Fabrication of Metal Tips Using Zone Electropolishing . . . . . 167
17.2 Etching STM Tips Reproducibly . . . . . . . . . . . . . . . . . . . . . 168
17.3 STM Tip Flasher . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
Contents xiii
17.4 Reverse Bias Method for Sharp Tip Etching . . . . . . . . . . . . . 170

17.5 Sharpening Single Crystal Metal Tips
by Ion Bombardment. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174
18 Spot-Welding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
18.1 Spot-Welding Difficult Junctions. . . . . . . . . . . . . . . . . . . . . 175
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
19 Mechanical Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179

19.1 Stretching Thin Metal Foils to a Wrinkle-Free Condition . . . . 179
19.2 Making Micron-Size Holes for Supersonic Nozzles . . . . . . . . 180
19.2.1 Method I—Fig. 19.2a . . . . . . . . . . . . . . . . . . . . . . . 180
19.2.2 Method II—Fig. 19.2b. . . . . . . . . . . . . . . . . . . . . . . 182
19.3 Straightening and Braiding Wires . . . . . . . . . . . . . . . . . . . . 183
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Part III Measurement Methods
20 Electrons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 187

20.1 Electron Gun Design and Behavior . . . . . . . . . . . . . . . . ... 187
20.2 Low-Energy Electron Gun for Broad-Beam Irradiation. . . ... 189
20.3 Low-Energy Electron Gun for Broad-Beam
Irradiation—Cylindrical Symmetry . . . . . . . . . . . . . . . . . . . 190
20.4 Electron Energy Analyzers . . . . . . . . . . . . . . . . . . . . . . . . . 191
20.5 Electron Spectrometer Calibration Using Graphite . . . . . . . . . 196
20.6 Measurement of Electron Beam Angular Divergence . . . . . . . 198
20.7 Measuring Anisotropy of Electrical Conductivity
of a Single Crystal—Four Point Probe . . . . . . . . . . . . . . ... 200
20.8 Spot Photometer for LEED Intensity Measurements. . . . . ... 201
20.9 Modified Faraday Cup for Electron Current Measurement ... 202
20.10 HREELS-Discrimination Between Single
and Multiple Electron Scattering Events. . . . . . . . . . . . . . . . 205
20.11 Rejuvenation of Cu/Be Electron Multipliers . . . . . . . . . . . . . 205
20.12 Geiger-Müller Counter for Inverse Photoemission . . . . . . . . . 206
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
21 Ions . ....................................... . . . . . . 211

21.1 Ion Sputter Gun with Low Gas Emission . . . . . . . . . . . . . . . 211
21.2 Alternate Ion Bombardment Sources . . . . . . . . . . . . . . . . . . 213
21.3 Ion Gun Based on Bayard-Alpert Gauge . . . . . . . . . . . . . . . 214
21.4 Sputter Ion Gun—Cold Cathode . . . . . . . . . . . . . . . . . . . . . 216
21.5 Broad Ion Sources—Improved Stability with Oxygen . . . . . . 218
xiv Contents
21.6 Focusing of Low-Energy Ions by Magnetic Field . . . . . . . . . 220

21.7 Low Energy Ion Gun—Construction and Performance . . . . . . 220
21.8 Measuring Ion Beam Dimensions . . . . . . . . . . . . . . . . . . . . 223
21.9 Beam Position Monitor—Charged Particles . . . . . . . . . . . . . 224
21.10 Pulse-Counting LEED/ESDIAD Analyzer—Using
MicroChannel Plate Detection. . . . . . . . . . . . . . . . . . . .... 225
21.11 Time-of-Flight Measurements for Ions Generated
in ESDIAD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 227
21.12 Ion Detection with Kinetic Energy and Mass Resolution
Using a Hemispherical Energy Analyzer . . . . . . . . . . . .... 229
21.13 Time-of-Flight Detection for Laser-Ionized Neutrals . . . .... 231
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 232
22 Photons (UV) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 235

22.1 Hollow Cathode Resonance Lamp for He(II)
(40.8 eV) Photoemission . . . . . . . . . . . . . . . . . . . . . . . . . . 235
22.2 He(I) and He(II) Ultraviolet Resonance Source . . . . . . . . . . . 236
22.3 Photochemistry on Surfaces Using an Ultraviolet Lamp . . . . . 238
22.4 Use of Polarized Ultraviolet Light for Photochemistry . . . . . . 240
22.5 Photoreduction of Ag on Doped TiO2—Making
Enhanced Photoactivity Visible. . . . . . . . . . . . . . . . . . .... 242
22.6 Reducing Radiation Damage in Raman Spectroscopy
of Solids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 243
22.7 Lyman-α Lamp Calibration Using a Chemical
Actinometer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 245
22.8 Pulsed Vacuum UV Light Source . . . . . . . . . . . . . . . . .... 247
22.9 Grazing Incidence X-Ray Photoemission Spectroscopy
(GIXPS): Enhanced Surface Sensitivity . . . . . . . . . . . . .... 247
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 250
23 Kinetics of Adsorption/Adsorbate Coverages . . . . . ........... 253

23.1 Absolute Adsorption Uptake Measurements
from a Calibrated Effusive Molecular Beam . . ........... 253
23.2 Measurement of Absolute Sticking Coefficient
at Desorption Temperature . . . . . . . . . . . . . . ........... 255
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........... 257
24 Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 259

24.1 Shielded Quadrupole Mass Spectrometers—Temperature-
Programmed Desorption. . . . . . . . . . . . . . . . . . . . . . . . ... 259
24.2 Enhanced Signal-to-Noise in Quadrupole
Mass Spectrometers. . . . . . . . . . . . . . . . . . . . . . . . . . . ... 262
24.3 Elimination of Spurious Electron Emission
from a Quadrupole Mass Spectrometer . . . . . . . . . . . . . ... 265
Contents xv
24.4 Ta Mass Spectrometer Filament . . . . . . . . . . . . . . . . . . ... 265

24.5 Modulated QMS Ion Source for Enhanced Sensitivity . . . ... 267
24.6 Line-of-Sight Mass Spectrometry for Thermal Desorption ... 269
24.7 Mass Spectrometer Calibration
by Molecular Decomposition . . . . . . . . . . . . . . . . . . . . ... 270
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 272
25 Temperature-Programmed Desorption/Reaction . . . . . . . . . . . . . . 273

25.1 Thermal Desorption—Angular Resolved . . . . . . . . . . . . . . . 273
25.2 Magic-Angle Thermal Desorption Mass Spectroscopy . . . . . . 275
25.3 High-Temperature Thermal Desorption Spectroscopy . . . . . . . 277
25.4 Automated Temperature Programmed Desorption
(TPD) Apparatus. . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 278
25.5 Scanning Kinetic Spectroscopy—A Survey Method
for Investigation of Surface Reaction Processes. . . . . . . . ... 280
25.6 Temperature-Programmed Reaction Spectroscopy (TPRS) ... 282
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 284
26 Gas Chromatography—Enhanced Sensitivity . . . . . . . . . . . . . . . . 287

26.1 Gas Chromatography—Sensitivity Enhancement . . . . . . . . . . 287
Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
27 Work Function, Tunneling Spectroscopy and Ellipsometry . . . ... 289

27.1 Work Function Changes Using Retarding Diode Method . ... 289
27.2 Work Function Measurements Using an Electron Gun . . . ... 291
27.3 Work Function and Electron Reflection Coefficient
Measurements with the Shelton Diode . . . . . . . . . . . . . . . . . 293
27.4 Improved Piezoelectric Drive Kelvin Probe . . . . . . . . . . . . . 295
27.5 Improved Electromagnetic Drive Kelvin Probe . . . . . . . . . . . 297
27.6 Kelvin Probe—High Temperature Flow Reactor . . . . . . . . . . 299
27.7 Scanning Tunneling Spectroscopy (STS)—Compensation
for Density of States of the Tip. . . . . . . . . . . . . . . . . . . ... 300
27.8 Ellipsometry as a Tool for Studying Adsorption . . . . . . . ... 302
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 303
28 Radioactive Adsorbates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305

28.1 Radiotracer Techniques for Surface Studies . . . . . . . . . . . . . 305
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
29 Thin Film Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309

29.1 Thin Film Deposition to Absolute Surface Coverages . . . . . . 309
29.2 Gold Evaporation Source . . . . . . . . . . . . . . . . . . . . . . . . . . 312
29.3 Ionization Gauge Measurement of Film Evaporation Rate . . . 314
xvi Contents
29.4 Reusable Quartz Crystals for Film

Thickness Measurement . . . . . . . . . . . . . . . . . . ......... 314
29.5 Hybrid Thin Film Deposition Processes—Stable
and Metastable Films. . . . . . . . . . . . . . . . . . . . ......... 317
29.6 Sputter-Coating Inner Walls of Tubes . . . . . . . . ......... 317
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ......... 319
30 Infrared Spectroscopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321

30.1 Infrared Reflection Absorption Spectrometer. . . . . . . . . . . . . 321
30.2 IRAS Measurements at High Pressure . . . . . . . . . . . . . . . . . 323
30.3 Internal Reflection IR Spectroscopy—Silicon . . . . . . . . . . . . 324
30.4 Internal Reflection IR Spectroscopy—Gallium Arsenide. . . . . 327
30.5 Infrared Spectroscopy—Continuous Pumping
of Detector Dewar. . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 328
30.6 Rapid IR Spectrometer Purge After Cell Transfer—Using
Rubber Sleeves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 329
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 331
31 Calorimetric Heats of Adsorption—Single Crystals. . . . . . ...... 333

31.1 Measurement of the Calorimetric Heat of Adsorption
on Ultrathin Metal Single Crystals. . . . . . . . . . . . . . ...... 333
31.2 Improved Adsorption Calorimeter for Single Crystals
at 100–350 K . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...... 334
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...... 336
32 Surface Debye Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337

32.1 Measuring the Surface Debye Temperature
of a Single Crystal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
33 Friction and Fracture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 341

33.1 UHV Tribometer for Measuring the Coefficient
of Friction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
33.2 Pin-on-Disk Measurement—UHV . . . . . . . . . . . . . . . . . . . . 343
33.3 UHV Fracture Stage for Surface Analysis After Fracture . . . . 343
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
Part IV Thermal Control
34 Heating Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349

34.1 Electronic Temperature Programmer—Metal Crystals . . . . . . 349
34.2 Temperature Programming of Metal
and Semiconductor Crystals . . . . . . . . . . . . . . . . . . . . . . . . 351
Contents xvii
34.3 Mounting Nonweldable Crystals for Resistive Heating

and Cooling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 354
34.4 Strain-Free Single Crystal Mounting Azimuthal Motion . . ... 355
34.5 Heating Silicon Crystals and Temperature Measurements . ... 356
34.6 Avoiding Extraneous Electrical Effects Owing
to Heating Crystals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
34.7 Electron Bombardment Crystal Heating . . . . . . . . . . . . . . . . 361
34.8 Indirect Sample Heating. . . . . . . . . . . . . . . . . . . . . . . . . . . 363
34.9 Indirect Heating of Compound Semiconductors. . . . . . . . . . . 365
34.10 Heating Design—Insulator Crystals . . . . . . . . . . . . . . . . . . . 366
34.11 Cement Mounting—Semiconductor Crystals . . . . . . . . . . . . . 368
34.12 Low-Stress Mount for Fragile Semiconductor Crystals . . . . . . 369
34.13 Radiation Heating of Crystals Through Glass Windows . . . . . 371
34.14 Pyrolytic Graphite Heating Element for UHV . . . . . . . . . . . . 372
34.15 Heater—Paint-on Type . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
34.16 Mounting High-Temperature Tungsten Filaments . . . . . . . . . 375
34.17 Rotatable Single Crystal. . . . . . . . . . . . . . . . . . . . . . . . . . . 375
34.18 Radiation-Heated Evaporator . . . . . . . . . . . . . . . . . . . . . . . 378
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378
35 Cooling Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 381

35.1 6 K—Cryogenic Crystal Holder . . . . . . . . . . . . . . . . ..... 381
35.2 Cooling Crystals by Contact with a Thermal Reservoir ..... 383
35.3 Cooling Samples on Manipulators—Thermal
Conductivity Issues . . . . . . . . . . . . . . . . . . . . . . . . . ..... 383
35.4 Vacuum-Jacketed Cryogenic Manipulator Rod . . . . . . ..... 386
35.5 Manipulator Component with Cryogenic
Azimuthal Rotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 388
35.6 Cooling Using Direct Contact . . . . . . . . . . . . . . . . . . . . . . . 389
35.7 Cooling Reservoir—Horizontal Manipulator . . . . . . . . . . . . . 390
35.8 Cryogenic Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . 392
35.9 Sample Cooling Using Thermal Braid . . . . . . . . . . . . . . . . . 392
35.10 Flexible Cooling Reservoir . . . . . . . . . . . . . . . . . . . . . . . . . 394
35.11 Enhanced Cooling Using Liquid Nitrogen . . . . . . . . . . . . . . 394
35.12 Level Alarm-Liquid Nitrogen . . . . . . . . . . . . . . . . . . . . . . . 396
35.13 Liquid Nitrogen Reservoir—Non-leaking . . . . . . . . . . . . . . . 398
35.14 Cryogenic Cooling to Below 4 K—Photon Stimulated
Desorption of Adsorbed He . . . . . . . . . . . . . . . . . . . . . . . . 399
35.15 Static Temperature Control by Mixing Gas Streams. . . . . . . . 401
35.16 Gas Thermal Switch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404
xviii Contents
36 Temperature Measurements . . . . . . . . . . . . . . . . . . . . . . . . . ... 405

36.1 Tungsten-Rhenium Thermocouples—Calibration
Over a Wide Temperature Range . . . . . . . . . . . . . . . . . ... 405
36.2 Measuring Temperatures of Hot Filaments . . . . . . . . . . . ... 407
36.3 Pyrometric Measurement of Temperatures . . . . . . . . . . . ... 409
36.4 Pyrometer Measurements Through Windows Coated
Inadvertently with Variable-Thickness Thin Films . . . . . . . . . 411
36.5 Calibration Point for Optical Pyrometer Used for Silicon . . . . 413
36.6 Control of Silicon Temperature Using Resistivity . . . . . . . . . 415
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 417
Part V Delivery of Adsorbates to Surfaces
37 Gases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 421

37.1 Design and Performance of Microcapillary Array
Beam Doser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421
37.2 Gas Handling System for Array Beam Doser . . . . . . . . . . . . 423
37.3 Calibration and Use of Array Beam Doser . . . . . . . . . . . . . . 425
37.4 Sticking Coefficient and Surface Reactivity . . . . . . . . . . . . . 426
37.5 Beam Doser Shutoff—Condensable Gases . . . . . . . . . . . . . . 428
37.6 Neutral Beam Intensity Measurement Using a Stagnation
Detector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 429
37.7 Low Deadspace Aperture Valve for Repetitive
Gas Dosing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 430
37.8 Gas Flow Regulation—Squeeze Valve . . . . . . . . . . . . . . . . . 433
37.9 Retractable Beam Doser with Convenient On-Off Control . . . 434
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 435
38 Evaporation Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437

38.1 Degassing Evaporation Sources. . . . . . . . . . . . . . . . . . . . . . 437
38.2 Metal Evaporation Sources for Downward Evaporation . . . . . 439
38.3 Ultralow Coverage Metal Evaporation . . . . . . . . . . . . . . . . . 440
38.4 Congruent Evaporation of Multicomponent Materials. . . . . . . 442
38.5 Evaporator for Eutectic-Forming Metals . . . . . . . . . . . . . . . . 444
38.6 Simple Lithium Metal Evaporation Source . . . . . . . . . . . . . . 444
38.7 Aluminum Evaporation Sources—Long Lived . . . . . . . . . . . 446
38.8 Cr Evaporator Sources of Compact Design. . . . . . . . . . . . . . 448
38.9 Electron Beam Evaporator for Refractory Materials . . . . . . . . 449
38.10 Small Electron Beam Evaporator for Refractory Metals . . . . . 451
38.11 Electron Beam Evaporator . . . . . . . . . . . . . . . . . . . . . . . . . 453
38.12 Break-Seal Ampoule Doser . . . . . . . . . . . . . . . . . . . . . . . . 455
38.13 Cadmium Sulfide Evaporation Source . . . . . . . . . . . . . . . . . 456
38.14 Arsenic Atom Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . 456
Contents xix
38.15 Portable Microevaporator for Depositing Low Coverages

of Single Atoms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 459
38.16 Nanoscale Templating of Close-Packed
C60 Nanostructures . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 460
38.17 Quantitative Molecular Beam Dosing of Slightly
Volatile Organics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 460
38.18 Organic Doser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 464
38.19 Vapor Deposition of Organic Thin Films . . . . . . . . . . . . . . . 464
38.20 Evaporator for Organic Film Deposition. . . . . . . . . . . . . . . . 465
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 468
39 Ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 471
39.1 Solid State Cesium Ion Gun Source. . . . . . . . . . . . . . . . . . . 471
39.2 Alkaline Earth Metal Ion Sources . . . . . . . . . . . . . . . . . . . . 473
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 475
40 Active Gases and Species . . . . . . . . . . . . . . . . . . . . . . . . . ..... 477

40.1 High Flux Atomic Hydrogen Sources—Thermal . . . . . ..... 477
40.2 Atomic Hydrogen Beam Source—Radio
Frequency Driven . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 480
40.3 Atomic Fluorine Source . . . . . . . . . . . . . . . . . . . . . . . . . . . 481
40.4 Alkali Metal Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 483
40.5 Thermal Source of Methyl Free Radicals . . . . . . . . . . . . . . . 485
40.6 Production, Storage, and Use of Ozone . . . . . . . . . . . . . . . . 486
40.7 Ozone Generator and Molecular Beam Doser Source. . . . . . . 488
40.8 Detection of Ozone or Atomic Oxygen by Oxidation
of Silver Films . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 490
40.9 Purity of Ir and Pt Hot Filament Sources
for Atomic Oxygen Production . . . . . . . . . . . . . . . . . ..... 491
40.10 Pure Hydrogen Peroxide Doser for Ultrahigh Vacuum . ..... 493
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 494
41 Electrochemical Sources and Devices . . . . . . . . . . . . . . . . . . . . . . 497

41.1 Electrochemical Sulfur Source . . . . . . . . . . . . . . . . . . . . . . 497
41.2 Electrochemical Halogen Sources . . . . . . . . . . . . . . . . . . . . 498
41.3 Electrochemical I2 Doser . . . . . . . . . . . . . . . . . . . . . . . . . . 501
41.4 Electrochemical Device Working in UHV . . . . . . . . . . . . . . 502
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504
42 Gas Purification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 505

42.1 Gas Purification for Experiments at High Pressures . . . . . ... 505
42.2 Production and Purity Measurement
for Extremely Pure He. . . . . . . . . . . . . . . . . . . . . . . . . ... 507
42.3 Purification of Gases by Cryogenic or Gettering Methods ... 509
xx Contents
42.4 Gas Purification by Permeation Through Metals .......... 511

42.5 O2 Source—Solid State . . . . . . . . . . . . . . . . . .......... 513
42.6 Reusable Sample Holder for Hydrogen
Permeation Studies . . . . . . . . . . . . . . . . . . . . .......... 513
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......... 514
43 Liquid Handling in UHV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 517

43.1 Adding Liquid to Atomically Clean Surfaces . . . . . . . . . . . . 517
Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 518
Part VI UHV Windows
44 Spectroscopic Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 521

44.1 Mounting Infrared Windows on UHV Systems . . . . . . . . ... 521
44.2 Windows for Vacuum UV Transmission into Ultrahigh
Vacuum Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 524
44.3 Construction of Be Windows for Ultrahigh Vacuum Use . . . . 526
44.4 Bakeable Aluminum Ultrahigh Vacuum Window . . . . . . . . . 528
44.5 Metal Thin Foil Windows—Welding . . . . . . . . . . . . . . . . . . 529
44.6 Preparation of Large-Area Si Single Crystal Windows . . . . . . 529
44.7 Strain-Free Optical Window Mounting. . . . . . . . . . . . . . . . . 532
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 533
45 Observation Windows . . . . . . . . . . . . . . . . . . ............... 535

45.1 Shielded Observation Window . . . . . . . ............... 535
45.2 Heated Internal Window for Prevention
of Film Deposition . . . . . . . . . . . . . . . ........ . . . . . . . 536
45.3 UHV Glass Window Seal—Low Optical Distortion. . . . . . . . 537
45.4 Zero-Length Window Assembly . . . . . . ........ . . . . . . . 539
References. . . . . . . . . . . . . . . . . . . . . . . . . . . ........ . . . . . . . 540
Part VII Surface Preparation Methods
46 Cleaning Metal and Semiconductor Crystals—Examples . . . . . .. 543

46.1 Surface Segregation and the Cleaning
of Metal Single Crystals. . . . . . . . . . . . . . . . . . . . . . . . . .. 543
46.2 Chemical Method to Measure and Remove Carbon
from Pd . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 546
46.3 Surface Cleaning Procedures—Silicon . . . . . . . . . . . . . . . . . 548
46.4 Preparation of Atomically Clean Tungsten Single Crystals . . . 551
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 551
Contents xxi
Part VIII High-Area Solids
47 Infrared Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 555

47.1 Infrared Cell for Adsorption Studies
on Supported Catalysts . . . . . . . . . . . . . . . . . . . . . . . . ... 555
47.2 Wide-Temperature-Range IR Cell for High-Area Solids . . ... 557
47.3 Measurements of Diffusion Through Powders
Using Transmission Infrared Spectroscopy . . . . . . . . . . . ... 560
47.4 Using Transmission Infrared Spectroscopy to Witness
Diffusion into Core-Shell Structures . . . . . . . . . . . . . . . ... 562
47.5 Diffusion in Powders as Measured by Transmission
IR Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 562
47.6 Transmission IR Spectroscopy Through Opaque Materials ... 564
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 567
48 Adsorption/Desorption—Thermal . . . . . . . . . . . . . . . . . . . . . . . . 569
48.1 Adsorption and Desorption from High-Area Solids . . . . . . . . 569
48.2 Thermal Desorption from High Area Materials . . . . . . . . . . . 571
48.3 Microcalorimetric Studies of Adsorption Heat on Powders . . . 574
48.4 Chemical Vapor Synthesis of Oxide Nanoparticles . . . . . . . . 576
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 577
Part IX Safety
49 Protecting the Vacuum System . . . . . . . . . . . . . . . . . . . . . . .... 581

49.1 Delay Circuit for Turbopumping Protection
Against Vacuum Loss from Power Interruptions . . . . . . . . . . 581
49.2 Fast-Closing Beam Valve for UHV Chamber Protection . . . . 583
49.3 Safety System for Oil Diffusion-Pumped UHV Systems. . . . . 584
49.4 Folding Linear Magnetic Translator . . . . . . . . . . . . . . . . . . . 586
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 587
50 Protecting Personnel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 589

50.1 Electrical Shocks in the Laboratory . . . . . . . . . . . . . . . . . . . 589
50.2 Accidental Electrical Charging from Ionization Gauge . . . . . . 590
Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 592
Appendix A. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 593
Appendix B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 597
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 643
Part I
Vacuum System Technology
Chapter 1
Motion in Vacuum
1.1 Crystal Manipulator with Six Degrees of Motional

Freedom
This design [1] is shown in two parts to separate the azimuthal motion (Fig. 1.1)
device from the tilt motion device (Fig. 1.2). A central rod extends from a typical
x, y, z-plus rotary motion manipulator. The azimuthal and tilt motion (pushing and
pulling) devices are connected to the same central rod and are free to slide up and
down on this rod. The central rod is slotted to keep the crystal mount linking fittings
from rotating with respect to the rod. The two push-and-pull actuators permit the
central rod that passes through them to rotate in two ball-bearing (BB) mounts.
Azimuthal crystal motion is transmitted by the push-and-pull actuator, which,
through two coupled universal joints, turns the crystal mounting platform on its
own ball bearing mount behind. Tilt motion is transmitted by the push-and-pull
actuator, which tilts the crystal mounting platform about the axis of the hexagonal
nut shown in the Tilt Motion diagram. The degrees of freedom for motion at various
points are shown by double-headed arrows. This design does not result in a
decoupling of tilt motion from other degrees of freedom, which is a disadvantage in
some cases, but the design has been used for more than ten years with success.
A more recent design uses a circular slot to guide a pin in the crystal tilt mechanism
that eliminates the crystal translational motion associated with the tilt motion. It is
of interest to note that in this design, ordinary unlubricated ball bearings are used in
ultrahigh vacuum with success and without seizure or binding.
The crystal is mounted on electrical and thermal isolation standoffs on the crystal
mounting platform and may be heated by electron bombardment or by ohmic
heating. Various thermocouples may be connected to the crystal, and cooling to
about 40 K is achieved using a copper braid connected between the crystal
mounting assembly and a liquid He cryostat. Appropriate provisions are made to
shield the mounting assembly from emitted radiation generated during crystal
heating so that the desorption of impurities from the mounting is minimized.
© Springer International Publishing Switzerland 2015 3

J.T. Yates Jr., Experimental Innovations in Surface Science,
DOI 10.1007/978-3-319-17668-0_1
4 1 Motion in Vacuum
Fig. 1.1 Azimuthal motion

1.1 Crystal Manipulator with Six Degrees of Motional Freedom 5
Fig. 1.2 Tilt motion
Because of the complexity of the design of a device of this sort, it is recom-

mended that a working model first be constructed before final design parameters
are set.
Many other designs allowing cooled crystals to be azimuthally rotated in
ultrahigh vacuum have been published [2–12] and some are treated in this book.
1.2 Rotation in Ultrahigh Vacuum
The provision for rotation of samples in ultrahigh vacuum (UHV) is centrally

import ant in most investigations, since the analytical facilities in most research
chambers are located on different azimuths from the center line of the vacuum
chamber. It is often desired to provide other motions, as well as heating and cooling
of the crystal, but this section is concerned only with axial rotation. Historically, the
problem was addressed in the first sixty inch cyclotron by Wilson [13], long before
ultrahigh vacuum was achieved in metal systems. It is of interest to note that the
differential pumping method employed by Wilson is still fundamentally important
to UHV technology.
Figure 1.3 shows the Wilson seal as currently used in ultrahigh vacuum to rotate
a long cylindrical cryostat on which the crystal is mounted using ceramic feed-
throughs [14]. Two Viton gaskets, with a hole cut smaller than the rotary seal tube,
are held in place in a brass housing that bolts onto a specially shaped lower Conflat
flange having an upward-facing conical section to push the Viton gasket upward as
shown. An assembly plate, below the upper Viton seal, also has an upward-directed
conical support section. These two Viton seals make a strained contact with the
polished-mirror-finished (600 grit) outer portion of the rotary seal, and provide the
vacuum tight joint. To prevent significant leakage through the lower Viton seal
gasket, the region between the two gaskets is differentially pumped with a rotary
pump or a small ion pump. The cryostat tube, welded to a mini-Conflat flange is
bolted to a mini-Conflat knife-edge sealing surface on top of the rotary seal. Two
stabilizer bearings prevent wobbling of the central components; these may be made
of phosphor bronze [14], Teflon, or they may even be ball bearings. Wires for
heating the sample electrically, and thermocouple wires conveniently thread down
the cryostat tube and exit from the bottom. As is obvious from the drawing, the
bottom Con-flat flange must be bolted on from below, meaning that a flanged
extension tube from the chamber flange must be employed under the device.
A handle added to the flange of the rotary seal (not shown) is helpful in applying
torque, and the upper section of the brass housing is divided into degrees of arc for
accurate angular positioning, using the pointer. Bakeout temperatures of 150 °C are
nondamaging to the Viton seals. For higher temperature bakeout, KALREZ per-
fluoro elastomer might be used [15]. The gaskets can most conveniently be made by
use of a homemade punching jig to accurately locate and cut the center hole to size
as well as to place the six radial holes through which pass the bolts clamping the
assembly plates down on the Viton. A pressure rise of about 1 × 10−10 Torr was
observed on slow rotation.
The seal in Fig. 1.3 is designed for liquid nitrogen cooling, and in order to
achieve long cryogenic storage time, a 1-L storage Dewar with a hole in its bottom
was fabricated [16]. To make a liquid nitrogen seal each day with the cryostat tube,
a platform support for the Dewar is covered with some cotton wool that is saturated
with water; sufficient ice forms to make an excellent seal. Ice formation under the
support platform by condensation from the atmosphere is a problem that can be
1.2 Rotation in Ultrahigh Vacuum 7
Fig. 1.3 Differentially pumped rotary seal—flat Viton gaskets
reduced by using a heater tape around the brass housing and possibly a fan to
evaporate liquid water runoff. The apparatus can be inefficiently used with the
gaseous discharge from a liquid He Dewar to achieve temperatures near 30 K, and
the momentary transfer of liquid He brings the temperature to 15 K [14, 16].
Another rotary seal based on this principle has been reported [17].
Figure 1.4 shows a second design for a rotary seal in which O-rings are
employed as the sealing elements, with one stage of differential pumping between
the O-rings [18]. A polished rotary-seal tube rotates against the slightly compressed
Fig. 1.4 Differentially pumped rotary seal—Viton O-ring
Viton O-rings, which are held in place by tightening the flange on the compression
pieces, which are made of copper and which slide in between the outer wall and the
rotary seal. These two compression pieces contain a number of holes to provide a
pumping opportunity between the inside and outside of the cylindrical pieces. The
O-rings are lubricated with a trace of graphite powder. A different lubricant reported
to be satisfactory for Viton O-rings consists of a mixture of graphite powder and
Santovac 5 diffusion pump oil (a polyfluo-roether) [19]. Differential pumping with a
rotary pump or a small ion pump is used for the inner section holding the lower
compression piece. A Teflon spacer at the top supplies a bearing surface for the
rotary seal. Inside the rotary seal, a cryostat tube similar to that in Fig. 1.3 may be
inserted. A pressure rise of about 3 × 10−11 Torr was observed on rotation [18].
Fig. 1.5 Differentially pumped rotary seal—spring-loaded Teflon seals spring-loaded Teflon
Rotation using large differentially pumped seals is often employed to move

heavy instruments inside the vacuum system, with a large-diameter rotating plate
being moved through 360°. Auerbach [20] and his colleagues first reported the use
of spring-loaded Teflon seals in this role in ultrahigh vacuum systems. Ordinary
Teflon has a high thermal expansion coefficient, poor creep characteristics, and high
permeability in static systems, so that differential pumping is required, along with a
mechanical trick in which a spring is embedded inside the Teflon ring [21] to
compensate for creep. These seals can also be obtained with Teflon filled with
MoS2 lubricant particles for ultimate lubrication of an already lubricious material
[20]. However, the filled polymer has a higher gas permeability than pure Teflon. In
Fig. 1.5, a diagram of a double-stage differentially pumped Teflon seal is shown
[20]. The large-diameter rotating plate is centered with two ball bearings, which
have been split at the inner and outer races to prevent distortion and binding owing
to differences in expansion coefficient between the bearings and the stainless steel.
Beneath the ball bearings are three spring-loaded Teflon seals. The space between
the upper two seals is pumped with a rotary pump, while a small ion pump [11 L/s]
pumps the space above the inner seal and the center seal. The system is baked to
200 °C, and the pressure rise upon rotation is less than 10−10 Torr and usually
unobservable. The pressure at the small ion pump is in the 10−8 Torr range. The
estimated gas leakage under static conditions is 10−13 Torr L/s [20]. Seals based on
this principle are now commercially available. Others have reported the use of this
type of seal for rotary manipulators [22–24].
Sometimes rapid rotation in ultrahigh vacuum is desired, where rather massive
components must turn in bearings. Figure 1.6 shows a magnetically driven rotation
Fig. 1.6 Internal ball bearings

device that drives an inner cylinder suspended in vacuum. Two vacuum ball
bearings are used at either end of the inner cylinder. These are ordinary ball
bearings that are unlubricated, and they contain large cross section ball races. These
bearings are routinely baked at 220 °C, and proved satisfactory for 18 months,
having undergone 5 × 106 rotations without any degradation in performance. The
evolution of gases is proportional to the rotation speed as shown in the lower part of
Fig. 1.6, and the rise in total pressure is about 2 × 10−11 mbar at 100 rpm. Most of
the gas evolution was due to H2O and H2 liberation [25].
Other designs for achieving independent rotational motions around two axes are
described in the literature [26–28]. Reference [26] describes an electrical readout
for the angular motions accurate to 0.5° and based on motion of a wire-wound
potentiometer.
1.3 Crystal Positioning Using Capacitance
The ability to position a single crystal precisely can be useful in many measurement
procedures in ultrahigh vacuum. This is a problem when the positioning must be
done at accuracies beyond the capability of micrometer drives on manipulators.
Factors such as thermal expansion effects, backlash in precision screw drives, and
vibration interfere with accurate positioning methods.
The topic of capacitive positioning of a single crystal is treated in the context of
locating the crystal accurately in front of a mass spectrometer that is housed in an
apertured shield for accurate and reproducible temperature programmed desorption
measurements [29]. This connects to the topic on p. 259.
Figure 1.7a shows a glass shield [29, 30] mounted over a quadrupole mass
spectrometer ionization source. A Ta ring is rigidly supported near the aperture of
the shield and acts as the capacitive pickup terminal. The capacitance is of the order
of 0.1 pF for a 1-mm spacing between the Ta ring and the crystal. The capacitance
measuring circuit, to be described below, can measure capacitance changes in the
range of 0.1 %, meaning that the crystal can be positioned to 1 μm reproducibility,
corresponding to an error in the measurement of the desorption rate from the crystal
of only about 0.05 %. Small tilting of the crystal will reduce this reproducibility,
and the demonstrated reproducibility of actual measurements was better than 1 %
[30].
Figure 1.7b shows a slight variation on the design. Here, an Ag ring is employed
as the capacitance pickup terminal, and it is fixed on the end of the glass shield as
shown. The Ag plate, (12 mm OD; 6 mm ID, 1 mm thick) is held in place by use of
3 Ta wires that are spotwelded to the plate and wound around three glass eyelets on
the shield. The Ag plate is connected through a Kapton-insulated and shielded lead
to a BNC connector that exits the vacuum system. To avoid static charging, the
glass shield is surrounded by a very thin grounded metal mesh [31]. An Aquadag
[32] (colloidal graphite) coating can also be used for this purpose, or a metal shield
like that shown on p. 260 may be used. The spacing employed between the crystal
Fig. 1.7 Crystal positioning using capacitance. a Ring capacitance. b Affixed capacitance.
c Capacitance measurement—block diagram
1.3 Crystal Positioning Using Capacitance 13
and the aperture is reported to be only 10 μm in this design, and this maybe
reproduced to 1 μm [31].
The capacitance measuring circuit [30, 31] is shown in schematic block diagram
form in Fig. 1.7c. It is a phase-shifted bridge circuit, working at 20 kHz [30].
A sinusoidal signal (20 Vpp) is applied to the crystal and the capacitive pickup is
made to the ring aperture, since the crystal is often connected to unshielded leads
and electronics outside of the vacuum system. The detection system is used as a
simple lock-in amplifier.
1.4 UHV Rotary Manipulator with Arcsec Resolution
In some cases, the use of a commercial crystal manipulator is inappropriate because

of the poor resolution and reproducibility inherent in these designs. For certain
cases, such as in X-ray standing wave spectroscopy (XSW), a high-precision rotary
stage is needed [33].
Figure 1.8a shows a side view of an ultrahigh vacuum system containing both a
coarse and a fine adjustment for rotating a single crystal about the horizontal axis.
The crystal holder is mounted on a platform such that its surface is in a plane that
contains the axis of the shaft that holds the crystal holder. This shaft, made of
precision-ground 304 stainless steel, rotates and translates inside of two
linear/rotary ball bearings that are mounted on the top flange. With a conventional
rotary feedthrough and rotary manipulator contacted to the shaft by a U-joint,
coarse sample manipulation in 7 arcsec increments can be achieved with a stepping
motor/speed reducer.
Figure 1.8b shows the principle of the high-precision rotary device, where the
shaft may be engaged by a lever, and where precise variations in the position of the
top of the shaft may be used to make azimuthal rotations to better than arcsec
accuracy and precision.
Figure 1.8c shows the detailed structure of the high-precision rotary device
coupled to the upper flange by a bellows. The rotary feedthrough drives an actu-
ating lever that pushes a pushrod and the lower clutch body against two springs.
When the force on the pushrod is released, the upper and lower clutch bodies grab
the shaft. The rotation of the shaft about its axis is determined by the setting of the
upper micrometer which is driven by a stepping motor. A weight is used to maintain
tension and contact between the upper micrometer and the movable flange, and
incremental motion of 0.88 arcsec is achieved with a lever arm of 292 mm length
and a micrometer displacement of 1.25 × 10−3 mm [33].
Fig. 1.8 Rotary motion—

arcsec resolution. a Side view
—coarse and fine
adjustments. b Schematic—
fine angular adjustment.
c Detail—fine angular
adjustment
1.5 Solenoid Driven Linear Motion Device 15
Fig. 1.9 Linear motion actuator
1.5 Solenoid Driven Linear Motion Device
A rapid linear motion inside an ultrahigh vacuum system may be achieved by using
a solenoid-power-driven actuator like that shown in Fig. 1.9. All moving parts are
inside the vacuum system, and the external solenoids do not require vacuum
encapsulation. The actuator is a barrel-shaped plunger made of high-purity iron
sized to make a loose fit inside the stainless steel tube. In this device a stroke of
3 cm was achieved. Each solenoid consists of 1400 turns of #24 magnet wire
wound on an aluminum spool and giving a dc resistance of 14 Ω. Short electrical
pulses are used on either solenoid to move the actuator in or out; the current is then
turned off to prevent overheating or stray magnetic fields. In operation, the actuator
will tilt as shown in the drawing when it moves inward. Measurements made on the
actuator show that a torque of 500 g cm can be achieved at the internal pivot point.
The closing time was measured to be 50 ms. The original paper [34] also provides a
design for either an automatic or manual actuator circuit.
The solenoids are removed for system bakeout by releasing the lock ring on the
back end. A similar magnetically driven shutter assembly using a Nd Fe-B mag-
netic encapsulated slug and an external ceramic-ring magnet has also been
described [35].
1.6 An Inexpensive Linear Motion Device
A homemade linear motion device may be constructed quite easily from an inex-
pensive machinist’s device as shown in Fig. 1.10. The vise is made of high-grade
gray iron and the jaws as received from the manufacturer are precision ground for
squareness and parallelism to ±0.0001 cm per cm. The hardened jaw pieces are
Fig. 1.10 Linear translation device
removed from the vise and replaced with aluminum clamps that have an inner lip
machined in them to prevent the Conflat flanges from slipping through. These
clamps are mounted on the vise and then tightened around the flanges holding the
welded bellows. Machinist’s vises giving up to 7.5 cm linear travel are available
[36] at a cost of about $200–$300, and this device, while heavy, gives smooth and
repeatable translation. The precise position can be determined using a vernier
caliper between the clamps, or repeatable positions can be established by using
simple spacer bars. The overall price is about a quarter that of a comparable
commercial bellows manipulator [37].
1.7 Translating Auger Spectrometer
It is sometimes appropriate to move an Auger spectrometer or other large instru-

ment over a large distance. This may be done in a conventional fashion by using a
large-bellows translation device, which is both heavy and expensive. An alternate
method for achieving the same result is to employ internal linear ball bearings and
to translate with a small bellows-type linear translator. This requires that the
electrical connections to the instrument be flexible over rather long extension
distances.
1.7 Translating Auger Spectrometer 17
Fig. 1.11 Translating CMA auger spectrometer
Figure 1.11 shows a design [38] for translating a cylindrical mirror analyzer
(CMA) Auger spectrometer over a 14-cm distance. This is done using a l/2-in.-
diameter linear ball bearing [39] that slides on a polished stainless support rod at the
top, and is guided along a 1/4-in.-diameter rod-linear bearing assembly on the
bottom. The use of the smaller rod on the bottom allows the system to have a little
flexibility to prevent seizure due to slight misalignment. The first upper support
bushing is clamped with a stainless steel ring to the μ-metal shield of the CMA
spectrometer using three set screws at 120° to each other. The second support
bearing is attached to a baseplate that rigidly supports the CMA by rods. The entire
assembly is mounted on an 8-in.-diameter Conflat flange. Translation is achieved
using a small linear translator, mounted on a small flange, which pushes and pulls
on the baseplate at a point close to the upper support rod to avoid torque on the
upper bushing during motion.
The flexible electrical connecting wires are looped to give the needed extension
and are made of Teflon-coated coaxial cable [40]. The coaxial cable is made UHV
compatible by stripping off the outer cover, leaving the cable shield, and contributes
no significant outgassing problem.
Another design for a translating Auger spectrometer may be found in [41]. In
this case, the miniature instrument (electron optics 30 mm diameter; 45 mm length)
is held on a single linear translator mounted on a 2 3/4-in.-diameter Conflat flange.
The resolution of this instrument is R = AE/Eo = 1.5 % [41].
Fig. 1.12 Simple multiple motion
1.8 Simple Multiple Motion Manipulator
The provision of various degrees of motional freedom for samples, apertures, etc.,
in ultrahigh vacuum systems is often accomplished with complex manipulators
such as that shown on p. 3. When complex motion is required and the measurement
of the position of the sample is not required, more simple devices are possible. In
Fig. 1.12, a device involving the combination of a port aligner and a linear trans-
lator is used to give multiple degrees of freedom for a rod. The port aligner allows
±10° motion within a solid cone, which is schematically shaded. The linear
translator permits motion back and forth within this solid cone [42].
1.9 Simple Device for Small Rotations in Ultrahigh

Vacuum—Grating Application
The rotation of gratings and other optical components through small angles in ultra
high vacuum for the purpose of alignment is not straightforward. This device
permits the adjustment of a grating through an angular range of less than 1° with
ease [43]. Adjustments in one direction or in two are possible, depending on the
structure used.
Figure 1.13a shows a side view of a device permitting one-dimensional rotation
about the x-axis. Two planar broad-leaf springs are connected to a base support, and
to a movable platform. A plane extrapolation of the two springs intercepts at the
origin, o, which is the center of the grating. Here the x-, y-, and z-axes are defined.
An adjustment rod is connected to the base support, and to an external adjustment
screw. Bellows, as shown, isolate the mounting system from the atmosphere.
Adjustment of the screw in the z-direction rotates the grating about the x-axis.
1.9 Simple Device for Small Rotations in Ultrahigh Vacuum—Grating Application 19
Fig. 1.13 Small rotations in UHV. a One-axis device. b Two-axis device. c Result—Z rotation
The principle can be extended to rotation about two axes, as shown in the
schematic (Fig. 1.13b). Here, two sets of orthogonal leaf springs are used, and
screws for rotation of the adjustment rod operate in the z- and x-directions, leading
to grating rotation about the x- and z-axis.
Figure 1.13c shows the use of the two-axis device for optimization of the res-
olution of a UHV spectrometer. In this case, an angular resolution of less than 0.1°
was achieved for optimum spectrometer resolution adjustment [43].
1.10 Turntable Rotation in Ultrahigh Vacuum
Figure 1.14 shows a turntable for the rotation of heavy equipment in UHV with
high angular precision and repeatability, and without the need for coated or
lubricated bearings. Rotation is achieved using a rotary feedthrough to move the
axle of the turntable. The flat bottom of the turntable rests on three spheres made of
a metal carbide mounted in a V-groove. The spheres are of 10-mm diameter and the
three define the plane on which the turnable rotates. The mechanical stability is
excellent, and several tens of kilograms may be supported. In addition, friction
is low in ultrahigh vacuum. The danger of the spheres coming close to each other is
small, especially when forward and backward motion is used alternatively. The
system is uninfluenced by bakeouts at 200 °C [44].
1.11 Small Motions in UHV Systems Using Shape

Memory Effect Alloys
An unusual phenomenon present in certain alloys may be used to make small

motions in ultrahigh vacuum systems by the application of a small heating current
to a helix of the alloy material. The shape memory effect (SME) is so named
Fig. 1.14 Rotary platform

1.11 Small Motions in UHV Systems Using Shape Memory Effect Alloys 21
because shape memory alloy (SMA) wire helixes can be made to change from the
“deformed” cold shape to the “preferred” hot shape reversibly. The restoring force.
is quite large and may be used to achieve small motions as illustrated in the two
examples shown here. There are numerous alloys that exhibit the SME, and a
favorite SMA for UHV applications is NiTi. This material is easily formed into the
hot shape and has a convenient transformation temperature. The shape change is
caused by a transformation from a martensitic to an austenitic phase, with the high
temperature austenitic material being produced at about 60 °C, while the reverse
transformation occurs at about 30 °C. Deformations producing less that 2 % local
strain in the martensitic state can be reversed by heating through the transformation
temperature. By setting the hot shape in a helical wire it is possible to get extensions
of several hundred percent without exceeding this strain limit.
Figure 1.15a shows a shutter arrangement made with a 0.45-mm-diameter NiTi
wire [45]. The hot shape of the left-hand helix was set by winding it on a cylindrical
mandrel and then heating to 500 °C for 1 h in a vacuum annealing oven to prevent
oxidation. The helix was then stretched between two stainless steel posts and pinch
clamped to each. The helix was therefore stretched from its preferred hot shape.
Upon heating with a current of 2.5 A, the left segment of the helix contracts to its
preferred hot shape, moving the attached shutter to the left in a period of less than
0.5 s. This operation stretches the right-hand cold helix beyond its plastic limit,
which means that it has little restoring force. Thus, the heating current can be turned
off after the shutter has moved to the left. To restore the shutter to its original
position, one heats the right-hand portion of the helix.
The effect of bakeouts at 215 °C on a severely stretched (300 %) helix was tested
over a 24-h period, and it was found that this treatment did not affect the hot shape
memory [45].
A second example in which the SME is used to open and close a clamp on a
manipulator is shown in Fig. 1.15b [46]. Here, the jaws of the clamp are opened
when the NiTi SMA wire is heated. The particular wire [47] used here is of small
diameter (0.15 mm) and has been treated in manufacture to exhibit a shorter length
at high temperature and a longer length at low temperature. For strong clamping,
four such wires were braided together and connected to a copper wire by a simple
overhand knot above the ceramic insulator tube, which passes through the back end
of the stationary upper jaw. The other end of the wire bundle is grounded through
the jaw. A restoring force to close the jaws when the heating current is off is
supplied by a spring mounted on a screw that is fixed in the upper jaw but that does
not touch the lower jaw. When the lower jaw opens, this spring is compressed.
Electrical heating to a wire temperature of 115 °C causes the jaws of the clamp to
open up by about 2 mm when 0.21 A is passed through the four wires in vacuum. If
the current is taken too high (0.23 A in vacuum), the annealing temperature will be
exceeded, and the SMA wires will lose their pulling ability. The reader is referred to
texts on the SME for more details [48, 49]. Also engineering studies of SME test
structures are described in [50].
Fig. 1.15 Small motions using shape memory effect. a Shape Memory effect—Shutter. b shape
memory effect—Clamp
1.12 Rotary Shutter Device Driven by Magnetic Eddy

Current
The use of magnetic eddy currents in an ultrahigh vacuum system to drive a low
friction shutter is illustrated in Fig. 1.16. The eddy current magnet, shown without
its housing in the upper part of the Figure, drives the stepped copper rotor at
1.12 Rotary Shutter Device Driven by Magnetic Eddy Current 23
Fig. 1.16 Eddy current rotor
300 rpm when ball bearings are used, and 60-Hz power is applied to the coil. The
copper disk is 0.8 mm in thickness. The stepped openings on the perimeter permit
the operator to vary the open duty cycle for evaporations through the openings from
an evaporation source mounted above the disk; operation near the perimeter gives
larger fractional transmission than with the substrate placed more toward the axis.
The lower end of the iron cylinder (5 cm dia. × 7.5 cm long) contains a semi-
circular piece of high-purity copper, 0.6 cm thick. It is wrapped with 150 turns of
#14 (1.6 mm dia.) magnet wire with Formvar insulation. This is housed in a
stainless steel can to prevent outgassing, and the can is welded shut using the
ceramic feedthrough for electrical contact to the coil inside.
The device works on the basis of the lag of the magnetic field at position b from
that at position a. This means that when the magnetic field is maximized during the
cycle at a, the eddy currents centered at b will also be maximized. A force on the
volume element of the copper disk near a will be in the direction from a to b, and
the disk will move. One quarter cycle later, the magnetic field at b will interact with
the eddy currents centered at a and the force is again in the direction from a to
b. During the next half cycle, both the magnetic field and eddy current directions
are reversed, again giving a force in the same direction. A mechanical brake is
employed to stop the rotation.
The shutter device is used to reduce the flux of evaporating material arriving at
the substrate by a precise geometrical method involving only the dimensions of the
transmission slots relative to the circumference of the sector disk. Thus, thick films
deposited near the outer edge of the disk are proportionately reduced to thinner
films deposited at a known radius under the disk. Measurements of the film
thickness deposited above the shutter can be converted to sector thicknesses below
the shutter [51]. A similar device, driven conventionally, for extending the range of
a quartz microbalance has been described previously [52].
1.13 High-Speed Ultrahigh Vacuum Motor
A modified hysteresis-synchronous motor was designed to eliminate most of the

outgassing that occurs during motor operation [53]. There are two sources of gas
evolution in motors working in vacuum. The first is outgassing of the bearings. The
second is outgassing of the stator winding insulation owing to heating during
operation.
The design shown in Fig. 1.17 eliminates both of these effects to a high degree.
The bearings [54] in the motor [55] were lubricated with a dry lubricant. The stator
was isolated from the vacuum system and the rotor by a container sealed with two
O-ring seals at either end. This container is filled with nitrogen to 1 atm to facilitate
heat conduction away from the motor. Cooling coils that are brazed to the outside of
the isolation container maintain low temperatures at the motor.
The main modification in the motor involved increasing the gap between rotor
and stator to allow the isolation container wall (0.010 in. thick) to fit. This may be
done in two ways. Either the inside diameter of the stator is increased, in which case
the isolation container wall must be of magnetic stainless steel (this was done), or if
the outer diameter of the rotor is decreased, the isolation container wall between
rotor and stator must by made of nonmagnetic stainless steel.
Motors modified in this way have been working for more than one year at
200 Hz and exhibit a pressure rise of less than 1 × 10−9 Torr [53].
To reduce the loss of gas still further, the motor housing could be filled with
nitrogen gas at about 1 Torr pressure, which is still adequate for heat transfer.
Alternatively, UHV sealants might be used to seal the motor [56].
1.14 High-Speed Rotary Feedthrough for UHV Operation 25
Fig. 1.17 High-speed UHV Motor
1.14 High-Speed Rotary Feedthrough for UHV Operation
A simple rotary feedthrough, suitable for high-speed rotation in an ultrahigh vac-

uum system, is shown in Fig. 1.18. The device uses a hysteresis motor, in which the
stator is externally mounted, for driving a shaft at speeds from 1200 to 15,000 rpm
[57]. The pressure rise during operation is less than 1 × 10−10 Torr, since the stator,
which generates gas on heating, is isolated from the vacuum system.
In the Figure, a shaft, held between two unlubricated bearings [58], is attached to
a rotor and both pieces are ground together to assure good dynamic balance. Two
spring washers eliminate axial play. The stator is held in place by plastic supports.
The torque delivered is limited by the motor size chosen [59] and by the gap
between stator and rotor, which is 1.5 mm here. After 1200 h of intermittent
operation, no need for bearing replacement was encountered.
Fig. 1.18 Very high speed UHV motor
1.15 Self-lubricating Bearings in UHV
The use of boron nitride, BN, as a bearing lubricant has been demonstrated for light
loads in a combination rotary-translator bearing [60]. This is especially convenient
for magnetically driven shafts where only small forces may be applied.
Figure 1.19 shows an exploded view of the bearing assembly. The rotary and
translation shaft, made of stainless steel, is supported at some distance from the
1.15 Self-lubricating Bearings in UHV 27
Fig. 1.19 Exploded view—self-lubricating translation-rotation bearing
bearing by V-blocks that can be moved into contact with the shaft, and these are not
shown here. The shaft is held by WC bearings (0.48 cm diameter) which them-
selves are held in a BN ball separator cylinder containing holes that accommodate
the WC balls. This BN ball separator provides continuous lubrication to the WC
balls and the rod, since BN is a dry lubricant. The balls are held in their place by a
type 6061 Al alloy retainer which slips over the top of the balls keeping them from
touching each other. An outer-bearing sleeve then slips over all of the components
and is stabilized by contact with the inside of the vacuum wall tubulation. The
bearing assembly is loosely retained between the magnetic drive mechanism and a
stop on its other end, not shown.
The system can be repeatedly baked at 275 °C. Since BN is hygroscopic, when
the chamber is open to atmosphere, the bearing is kept at 80–100 °C using heating
tape on the outside of the tubulation. No contamination of samples by BN has been
detected using various surface analysis methods.
1.16 Lubrication for Heavy Sliding Loads in Ultrahigh

Vacuum
The lubrication of metal/metal interfaces for sliding motion in ultrahigh vacuum is a

special problem related to the fact that practical surfaces that often slide together
with low friction in air will seize in vacuum because of the absence of adsorbed
molecular layers of water lubricant or other materials.
Figure 1.20a shows a test device used to evaluate the lifetime of a special coating
on stainless steel surfaces that was able to lubricate bearing surfaces successfully
for over 200,000 cycles. The comparable unlubricated surfaces failed after 1740
cycles [61].
A ballscrew that translates over spherical bearings in a race was installed, along
with a parallel identical partner ballscrew behind the device plane (not shown), in a
device which drove a 2-kg mass up and down in ultrahigh vacuum of the order of
Fig. 1.20 Heavy load lubrication in UHV. a Ballscrew test device. b Wear track patterns
10−10 Torr. The vacuum condition was achieved after a bakeout at 140 °C. The
motion drove the mass upward and downward 45 mm, at a speed of 2.25 mm/s. The
screws and nuts of the ballscrews are made of 440C stainless steel, finished using
electrochemical buffing to a roughness of about 0.1 μm, and then coated as
described below.
The ballscrew was coated with a TiC film, 1.5 μm in thickness, and with a
micro-Vickers hardness of about Hv 1100. An Ag coating of 0.3-μm thickness was
then applied. The ball bearings mating to the ballscrew were also made of 440C
stainless steel, coated with an Ag film of 0.3-μm thickness. The unlubricated sur-
face, used as a control, was not treated with either TiC or Ag.
1.16 Lubrication for Heavy Sliding Loads in Ultrahigh Vacuum 29
Figure 1.20b shows the wear track observed visually after 200,000 cycles. The
thin-film Ag coating is still retained on sections of the screw that did not undergo
contact with the balls. The TiC underlayer is visible in the contact regions and was
not worn through. Surprisingly, when the device was operated in air, the untreated
ballscrew performed well but the coated ballscrew did not move smoothly.
Plasma-coated Ag films on stainless steel screws for use in UHV have been
found to resist cold welding [62]. Also, ordinary Ag plating of either the bolt
threads or the nut threads of stainless steel bolts used internally or externally in
baked UHV systems is reported to be an excellent lubricant using coatings of
10-μm thickness [63]. It has been found that coating metals such as Ta, Cu, Mo, and
stainless steel with an oxide film, by heating to redness in air, will provide pro-
tection against galling during motion in UHV [64].
A good review on tribology and lubrication in the vacuum environment in space
is found in [65].
1.17 Simple UHV Bearing for High-Speed Shafts
Steel ball bearings are often used unlubricated in ultrahigh vacuum. At high rota-
tional speeds, over time, these bearings tend to become sticky before final seizure.
The development shown in Fig. 1.21 uses polytetrafluoroethylene (PTFE or Teflon)
as a bearing material for continuous rotation at 10–1000 rpm. The bearing is
machined using sharp oil-free tools, and care is taken not to introduce metallic
particles into the PTFE. Cleaning is carried out in organic solvents, and traces of
metal and oil may be removed by use of hot nitric acid solution, followed by
soaking in distilled water.
The design has three essential components. The first is the use of relief slots
made by a small, round file at 90° intervals around the inside of the bearing.
Removal of about one-third of the contact area was found to be essential. The
second feature is the placing of pumping holes in the PTFE bearing, and more of
these can be supplied on the upper housing. The third feature of the design is the
use of an 0.25–0.50-mm-diameter wire squeezed between the bearing and the
housing as a friction point to prevent the bearing from rotating when torque is
supplied from the shaft.
Bakeout below 200 °C is appropriate, and temperatures above about 250 °C are
to be avoided. The bearing should be operated near room temperature and the
mechanical admission to the bearing surface of flakes of metal from thin film
deposition in the chamber should be avoided. Small partial pressures of fluorine and
chlorine were seen initially, and these decreased with time. Initial use also produced
small partial pressures of methane and carbon monoxide which soon disappeared.
The rotary bearing was used with a 400 rpm shaft for a million rotations without
any apparent degradation [66].
Fig. 1.21 Internal rotary bearing
1.18 Flange-Mounted UHV Variable Aperture
The need for an internal aperture that can be adjusted in ultrahigh vacuum exists
both in optical experiments and also in vacuum systems where variable pumping
speed control is desired. An inexpensive aperture device is shown in Fig. 1.22. It
consists of a variable iris aperture [67] that may be purchased as a component,
mounted inside a 2 3/4-in.-diameter Conflat flange. The aperture is mounted on an
1.18 Flange-Mounted UHV Variable Aperture 31
Fig. 1.22 Adjustable UHV Aperture
adapter ring that is tack welded into the flange to avoid distortion. The aperture is
fixed from rotation by mounting pins that fit into the adapter ring and through the
wall of the aperture housing. The aperture is tightened into place by a threaded ring
containing holes for a spanner wrench. The original Conflat flange has had an inside
slot milled around its inner circumference for the motion of the aperture control.
This is accomplished by a linear actuator coupled to the aperture control lever by a
Cu-Be strip (0.3 mm thick) that extends around the aperture housing inside the
milled slot. The aperture is degreased before use and relubricated with a tiny
amount of MoS2 applied lightly with a cloth.
The iris opening ranges from 1 to 25 mm and can be reproduced to a precision of
2 % if one is careful to avoid backlash by moving always in the same direction for
final adjustment [68].
1.19 Linear Motion Platform (LMP)
A clever mechanical design to convert a linear screw motion into a linear platform
motion with reduced amplitude has been described by Dr. H.P. Rust and
Mr. L. Cramer [69]. The design is used as an approach mechanism for an STM tip
in a low-temperature instrument [70].
Fig. 1.23 Linear movable platform. a 3D-sectional view of portion of LMP. b Top and sectional
view (x–x)
The device shown in Fig. 1.23a, b is made entirely from a single piece of bronze,
which, after fabrication, is rigidly attached to a base plate also made of bronze. The
active LMP part is milled with three orthogonal slots in the form of a C, and then
6 “milled hinge” cavities are also produced as shown in Fig. 1.23b by dashed lines.
These milled hinge cavities have only a thin section of bronze remaining at the top
1.19 Linear Motion Platform (LMP) 33
of each cavity, giving excellent bending response through a few degrees. The LMP
is then rigidly attached to the two projecting flats from the baseplate, and a screw
motion, schematically shown as a thumbscrew, is added on the left.
The device works as follows. As the screw is moved up or down, the plate
section above it moves up or down, and we will assign a length of motion a as
shown in Fig. 1.23a. This motion is reduced at point b by virtue of the relative
length of the arms involved as the milled hinge is bent near the point b. The section
of the LMP device labeled c is rigidly connected to the outer plate section con-
taining the a and b motions, and there is therefore an equivalent c motion as shown,
i.e., b = c. As section c moves up, a milled hinge, midway between c and c’ flexes
as a result of the forces it experiences. This causes a symmetrical distortion in the
bridged section c–c’ and leads to equal angular displacement on both ends of the
platform, which occurs as a result of distortions in the region of the pair of milled
hinges on each end of the platform. The small motion of the platform is sche-
matically indicated by d. Because of the symmetry, the motion of the platform up
and down will be in the normal direction without any tilting, which is the desired
result.
The function of the device can best be understood by remembering that the light
gray region, shown in Fig. 1.23b, and containing a, b, and c and the symmetrically
similar regions, is rigid and undergoes identical angular distortion away from the
baseplate when the screw is advanced. The mechanical advantage is given by the
ratio of the lever arms, l1 and l2.
1.20 Vacuum Trolley Transport System
The transport of samples from one vacuum system to another, where different
preparation and measurement methods are used, is a common necessity in UHV
measurements, as only so many techniques may be used in a single vacuum
chamber. A trolley system can be used to make this transport, and the system shown
has the capacity to transport 4 samples at one time, dropping them off amongst
5 vacuum systems. Magnetic transfer is used to insert a sample holder into the
trolley system from the loading chamber at the third position from the left. A screw
advance mechanism moves the samples laterally in 4 special sample holders which
are designed as locking devices. The locking sample holders, shown schematically
on the trolley in the insert to Fig. 1.24, are used to hold the samples. Engagement of
each sample holder by the magnetic transfer device allows the sample holder to be
inserted and locked by rotation on centrally-located sample stages in each system.
While this device has maximum versatility for simple analysis at various locations,
heating of the samples inside the vacuum chambers will require reversible electrical
connections for both electrical heating power and thermocouples as described
elsewhere in this book [71].
Fig. 1.24 Vacuum trolley transport system—4 samples
1.21 Thermally Compensated STM with Repeatable

Sample Positioning
Two problems common to most STM designs have to do with the thermal drift of
the relative position of the STM tip and the sample and the inability to replace a
sample in the same position so that the STM tip can return to the same image point.
Both of these problems have been solved by a design that addresses these issues
[72].
Figure 1.25a shows a side view of the STM design. Two piezotubes are mounted
on the same base. The inner tube (scan piezotube) is used for scanning the tip, while
the outer tube (translation piezotube) connects to the sample support assembly and
is used to support and translate this assembly. Both of these piezotubes are of
exactly the same length, so that thermal drifts are compensated in the z-direction;
their circular symmetry eliminates thermal drifts in the x, y-directions. The outer
translation piezotube is surrounded by a thermal shield, made of oxygen-free high
conductivity (OFHC) Cu to eliminate temperature gradients, which would defeat
the principle of having two identical-length piezotubes for compensation of thermal
drift effects. The outer tube is connected to the sample holder through a quartz
mounting assembly that consists of an outer quartz tube and a quartz or
metal/ceramic sample holder. The sample holder slides (by means of five spherical
bearing contacts) on two metallized quartz rails that are mounted inside the outer
quartz tube as shown in Fig. 1.25b. Operation of the translator piezodrive causes the
sample to jump toward the tip as a result of the friction and inertia effects in the
sample holder, and adjustment of the piezo activation waveform can be used to
1.21 Thermally Compensated STM with Repeatable Sample Positioning 35
Fig. 1.25 Thermally compensated STM—repeatable sample positioning. a Side view. b Sample
holder. c Detail, probe translator
produce jumps from about 104 Å to less than 5 Å. By working in the kHz frequency
domain, translation rates up to 1 mm/s are possible.
The sample holder may be removed from the outer quartz tube containing the
rods and the sample may be transferred and heated in a furnace or in a load lock.
Upon return of the sample holder to the quartz rods, the design permits repositioning
to an accuracy of several thousand Å, depending upon how much the sample is
bumped during transfer. This is accomplished by using gravity to align the sample
on the rails, which are tilted with respect to the horizontal plane. The repositioning
scheme, using gravity to set the sample on the rails in the same position, is based on
kinematic design principles that are well described elsewhere [73].
The details of the construction of the probe translator are shown in Fig. 1.25c.
The thermal drift experienced by this design is about 1 Å/h.
A piezo-driven translator for the tip, which moves in the x, y-directions over
several millimeters by a stepping and gripping action, thereby eliminating
mechanical positioning devices, has been described as a modification of this device
[74].
References
1. P. Toennies, Max Planck Institut Stroemungsf, Bunsenstr. 20, D-37073 Göttingen, Germany
(private communication)
2. D. Venus, Rev. Sci. Instrum. 62, 1361 (1991)
3. Y. Dai, H. Li, F. Jona, Rev. Sci. Instrum. 61, 1724 (1990)
4. T. Engel, D. Braid, E.H. Conrad, Rev. Sci. Instrum. 57, 487 (1986)
5. J.J. Zinck, W.H. Weinberg, Rev. Sci. Instrum. 56, 1285 (1985)
6. K.D. Jamison, F.B. Dunning, Rev. Sci. Instrum. 55, 1509 (1984)
7. N.J. Wu, A. Ignatiev, Rev. Sci. Instrum. 56, 752 (1985)
8. J.A. Stroscio, W. Ho, Rev. Sci. Instrum. 55, 1672 (1984)
9. J. Larscheid, J. Kirschner, Rev. Sci. Instrum. 49, 1486 (1978)
11. H. Durr, Th Fauster, R. Schneider, J. Vac. Sci. Technol. A8, 145 (1990)
12. M. Michaud, P. Cloutier, L. Sanche, Rev. Sci. Instrum. 66, 2661 (1995)
13. R.R. Wilson, Rev. Sci. Inst. 12, 91 (1941)
14. V.R. Unwin, K. Horn, P. Geng, Vakuum Tek 29, 149 (1980)
15. M.F. Zabielski, R.R. Blaszuk, J. Vac. Sci. Tech. 13, 644 (1976)
16. J.T. Yates, Jr. (unpublished)
17. J.J. Zinck, W.H. Weinberg, Rev. Sci. Inst. 56, 1285 (1985)
18. A. Pararas, S.T. Ceyer, J.T. Yates Jr, J. Vac. Sci. Tech. 21, 1031 (1982)
19. E. Puckrin, J.K. Fowler, A.J. Slavin, J. Vac. Sci. Tech. A7, 2818 (1989)
20. D.J. Auerbach, C.A. Becker, J.P. Cowin, L. Wharton, Rev. Sci. Inst. 49, 1518 (1978)
21. The spring-loaded Teflon seals are available in many sizes from the Fluorocarbon Corporation,
10871 Kyle Street, P.O. Box 520, Los Alamitos, CA 90720. Their descriptive literature gives
many details about the seals and the preparation of the metal surfaces on which they ride
22. S.M. George, J. Vac. Sci. Tech. A4, 2394 (1986)
23. J.A. Stroscio, W. Ho, Rev. Sci. Inst. 55, 1672 (1984)
24. P.A. Thiel, J.W. Anderegg, Rev. Sci. Instrum. 55, 1669 (1984) (A particularly nice design of a
manipulator with an easily demountable cold finger end for crystal replacement can be found)
References 37
25. R.A. Kubiak, P. Stonestreet, S.M. Newstead, E.H.C. Parker, T. Whall, N. Naylor, J. Vac. Sci.
Technol. A9, 2797 (1991)
26. J. Larscheid, J. Kirschner, Rev. Sci. Instrum. 49, 1486 (1978)
28. H. Durr, Th Fauster, R. Schneider, J. Vac. Sci. Technol. A8, 145 (1990)
29. P. Feulner, D. Menzel, J. Vac. Sci. Technol. 17, 662 (1980)
30. H. Schlichting, D. Menzel, Surface Sci. 285, 209 (1993)
31. H. Schlichting (private communication)
32. Obtainable from Acheson Colloids, 1607 Washington Avenue, P.O. Drawer 611747, Port
Huron, MI 48061-1747
33. T. Gog, S.M. Durbin, Rev. Sci. Instrum. 60, 3030 (1989)
34. R.W. Bertram, J. Vac. Sci. Tech. 7, 612 (1970)
35. C.M. Rouleau, R.M. Park, J. Vac. Sci. Tech. A11, 464 (1993)
36. MSC, 151 Sunnyside Boulevard, Plainview, Long Island, New York, 11803
37. H. Gutleben, J.T. Yates Jr, J. Vac. Sci. Tech. A9, 170 (1991)
38. M.C. Reuter, R.M. Tromp, J. Vac. Sci. Technol. A6, 2575 (1988)
39. 1/2 in. -Thompson A-81420-SS; 1/4 in.—Thompson A-4812-SS
40. Custom Cable Corp., 242 Butler Street, P.O. 1050, Westbury, NY 11590-3193. Cable type:
RG 316/V
41. D.N. Mcllroy, P.A. Dowben, A. Knop, E. Rühl, J. Vac. Sci. Technol. B13, 2142 (1995)
42. Dr. Oliver Schaff, Fritz Haber Institut der Max Planck Gessellschaft, Faraday Weg 4-6,
D-14195 Berlin (private communication)
43. W.A.M. van Bers, A.J.J. Franken, P.K. Larsen, Rev. Sci. Instrum. 54, 637 (1983)
44. W. Stocker, K.H. Rieder, J. Vac. Sci. Tech. A2, 1606 (1984)
45. A.P. Jardine, M. Ahmad, R.J. McClelland, J.M. Blakely, J. Vac. Sci. Tech. A6, 3017 (1988);
This idea is based on an earlier paper which also characterizes the behavior of a similar device
as a function of repeated cycling; see B.J. Mulder, Vacuum 26, 31 (1976)
46. W.F. Smith, E.E. Ehrichs, A.L. de Lozanne, J. Vac. Sci. Tech. A10, 576 (1992) (This article
contains additional references to other devices constructed with SME alloys)
47. The thicker SME wire used in Figure 11A was obtained from RayChem, 300 Constitution
Drive, Menlo Park, CA 94025
48. J. Perkins (ed.) Shape Memory Effects in Alloys (Plenum, New York, 1975)
49. T.W. Duerig, K.N. Melton, Proc. Mater. Res. Soc, Int. Meet. Adv. Mater. 9, 581 (1989)
50. M. Nishikawa, M. Kawai, T. Yokoyama, T. Hoshiya, M. Naganuma, M. Kondo,
K. Yoshikawa, K. Watanabe, J. Nucl. Matls. 179–181, 1115 (1991)
51. J.S. Brooks, R. Meservey, R.K. MacNabb, J. Vac. Sci. Tech. 20, 243 (1982)
52. F.C. Wilson, D.B. Sandstrom, H.B. Serreze, Rev. Sci. Instrum. 46, 1593 (1975)
53. T. Engel, D. Braid, Rev. Sci. Instrum. 56, 1668 (1985)
54. Barden Bartemp bearings, SR 188 SSTB
55. Globe motor, model 75 A 1003-2
56. Professor Thomas Engel, Department of Chemistry, BG-10 University of Washington, Seattle,
WA 98195 (private communication)
57. T. Engel, Rev. Sci. Instrum. 52, 301 (1981)
58. The bearings were from Barden Bartemp, Type SR 188 SSTB
59. The rotor and stator were obtained from KaVo Electrotechnisches Werk GmbH D-7970
Leutkirch, Germany, Type EV 48-15
60. G.J. Collet, E.L. Garwin, R.E. Kirby, Rev. Sci. Instrum. 58, 479 (1987)
61. H. Saeki, J. Ikeda, I. Kohzu, H. Ishimaru, J. Vac. Sci. Technol. A8, 3360 (1990)
62. R.A. MacGill, R.A. Castro, X.Y. Yao, I.G. Brown, J. Vac. Sci. Technol. A12, 601 (1994)
63. A. Contaldo, Rev. Sci. Instrum. 36, 1510 (1965)
64. W.F. Egelhoff, National Institute of Standards and Technology, Gaithersburg, MD
20899-0001 (private communication)
65. E.W. Roberts, M.J. Todd, Wear 136, 157 (1990)
66. R.A.A. Kubiak, P. Driscoll, V. Manning, R. Houghton, J. Vac. Sci. Tech. A4, 1951 (1986)
67. Edmund Scientific Company, 101 East Gloucester Pike, Barrington, NJ 08007-1380
68. F. Homan, J. Vac. Sci. Tech. 17, 664 (1980)
69. Registered design: H.-P. Rust, L. Cramer, Deutsches Patentamt Gebrauchsmuster No. G94 12
112.5, (27 July 1994)
70. H.-P. Rust, J. Buisset, E.K. Schweizer, L. Cramer, Rev. Sci. Instrum. 68(1), Jan 1997
71. J. Szuber, private communication
72. J.W. Lyding, S. Skala, J.S. Hubacek, R. Brockenbrough, G. Gammie, Rev. Sci. Instrum. 59,
1897 (1988); also, U.S. Patent 4,841,148
73. E.B. Wilson, An Introduction to Scientific Research, 1st edn. (McGraw Hill, NY, 1952)
74. R.T. Brockenbrough, J.W. Lyding, Rev. Sci. Instrum. 64, 2225 (1993)
Chapter 2
Sample Transfer
2.1 Sample Transfer from High Pressure to Ultrahigh

Vacuum
The ability to transfer samples from high-pressure reaction conditions into ultrahigh
vacuum for surface analysis or for crystal preparation has its origin in the work of
the Somorjai group at Berkeley [1], who devised apparatus to do this for studies of
the kinetics of catalytic reactions on metal single crystal surfaces. Figure 2.1 shows
an early version of this scheme. A single crystal mounted on a flange is surrounded
by a movable reactor cup that seals the crystal into a small cavity capable of being
pressurized. The reactor cup is sealed under strong force by a drive rod that operates
inside a stainless steel welded bellows. Because of the strong sealing forces
involved, the body of the ultrahigh vacuum chamber, shown cross-hatched in
Fig. 2.1, is made of extra thick stainless steel to avoid significant distortion when
the sealing force is applied. The crystal can be heated electrically in the reactor cup
and its temperature measured with a thermocouple welded to the crystal edge. In
some cases the inside of the high-pressure chamber is coated with a gold film to
reduce wall reactivity. Gas circulation pumps cause mixing of the high-pressure gas
in order to mimic stirred flow conditions. The design shown in Fig. 2.1 would not
permit Auger analysis of the crystal surface because of its fixed geometry, but later
work permitted the crystal itself to be moved inside the ultrahigh vacuum system
after the reactor cup was retracted [2–5].
Figure 2.2 shows an alternate design that has been widely used [6]. The crystal is
mounted on a ceramic feedthrough at the end of a transfer tube housed inside a
welded stainless steel bellows. The bellows conveniently moves on an external
translation carriage containing two parallel stainless rods that accommodate linear
bearings connected to the outer flange (not shown) of the retraction device. When
retracted, the crystal sits in a reaction chamber which can be pressurized to about
two atmospheres. The reaction chamber is closed off by two all metal sealed valves,
one leading to the surface analysis/preparation chamber, and the second leading to

DOI 10.1007/978-3-319-17668-0_2
40 2 Sample Transfer
Fig. 2.1 Movable reactor cup—stationary crystal
the high-pressure chamber pumps. As in Fig. 2.1, the entire apparatus can be
brought into ultrahigh vacuum condition by means of bakeout. The crystal can be
translated into the analysis/surface preparation chamber where surface analysis and
gas dosing can be carried out under well-controlled conditions. In the high-pressure
reactor chamber, gas analysis is made by extraction of a small quantity of the gas
either into the sampling loop of a flame ionization gas chromatograph, or into the
mass spectrometer of the main ultrahigh vacuum chamber. For many analyses, the
gas chromatograph is superior in sensitivity and selectivity to the mass spectrom-
eter. Compared to the reactor cup design, the internal volume of the high-pressure
cell shown in Fig. 2.2 is large and wall effects may in principle be larger. For certain
reactions, the wall effect has been measured and found to be small [7]. Viton O-ring
sealing valves are likely to give poorer results because of absorption and liberation
of gases from the Viton. Often, wall cooling is necessary when the single crystal
operates at high pressure and at elevated temperature because of efficient heat
transport to the s through the dense gas. It may be necessary to cool the outer wall
of the reactor chamber in these cases. Such cooling is not easy to achieve in the
reactor cup design of Fig. 2.1. High-pressure chambers are often useful in chemical
2.1 Sample Transfer from High Pressure to Ultrahigh Vacuum 41
Fig. 2.2 High pressure cell—movable crystal
cleaning of crystals because of the high gas density achievable, let us say for
oxygen or hydrogen treatment of a metal single crystal. There is a danger with all
stainless steel reactor chambers when operated at high pressure with carbon mon-
oxide gas. The carbon monoxide will produce nickel and iron carbonyls from the
stainless walls that transfer their metal atoms to the hot metal surface (see pp. 89
and 505).
Another type of sample transfer device was marketed by Leybold-Heraeus and is
schematically shown in Fig. 2.3. A 2-cm-diameter polished stainless steel rod is
arranged to move laterally by sliding through a number of differentially pumped
O-ring seals that permit rod translation rapidly from a high-pressure reactor into the
ultrahigh vacuum system. Differential pumping occurs by means of a mechanical
pump in the first stage between the reactor cell and the ultrahigh vacuum system,
and high-pressure gases used in the reactor cell would be carried away here. The
second stage is pumped with a turbomolecular pump, and achieves a pressure of the
order of 10−8 mbar [8, 9]. This design was used by others [10, 11] with minor
modifications, and in the lower section of Fig. 2.3 can be seen the manner in which
a metal single crystal is mounted [10]. The high-pressure reactor is only a few cubic
centimeters in volume and was entirely made of stainless steel and gold-plated pan
bronze, except for Viton/Teflon seals. The small volume allows an enhanced sen-
sitivity for product gases evolved in catalytic reactions at high pressures [10, 11].
Fig. 2.3 Sliding rod seal
Leaks due to poor O-ring lubrication or to scratching of the stainless rod is avoided
to some degree by providing Teflon liners, in the form of tubes which slide on the
rod between the O-rings [12].
A more recent type of arrangement for sample transfer between an ultrahigh
vacuum system and a high-pressure chamber is called the “rod-through-chamber”
geometry. As shown in Fig. 2.4, a single crystal is mounted on a long feedthrough
that contains a special conical shoulder that mates to a Teflon gasket when the
crystal has been advanced into the high-pressure chamber. The conical contour
shoulder is polished to a mirror finish, and acts like a plunger when closing the
2.1 Sample Transfer from High Pressure to Ultrahigh Vacuum 43
Fig. 2.4 Rod-through chamber
apparatus to isolate the two sections. This Teflon closure can seal against 1 atm
pressure with a leak rate less than 10−9 Torr L/s. When the high-pressure chamber is
evacuated and the probe is withdrawn, a gate valve is closed to completely isolate
the ultrahigh vacuum system from the high-pressure chamber. While bakeout
temperature limits are not specified in the original paper [13], temperatures in the
150–200 °C range would not be expected to disturb the Teflon. This type of seal
requires that a rather large force be applied to the rod in order to make a gas tight
junction with the Teflon gasket.
In some cases, it is not appropriate to employ high forces along the rod to make a
metal-to-Teflon seal. Fig. 2.5 shows a method for making a high-force seal using
another approach, on a crystal holder that moves vertically and that is not so robust as
that shown in Fig. 2.4. Here the crystal is mounted on electrical leads that are cooled
in a reservoir containing liquid nitrogen supplied through a tube entering from the
left-hand side. The gas tight seal is again made by contact between a polished
stainless steel surface on the bottom of the plunger, which is of spherical contour
mating up with a square-edge Teflon gasket. The upper portion of the plunger is also
of spherical contour, and when the plunger is seated properly on the Teflon gasket,
tapered steel wedges are able to be brought in from the side to mate with the upper
spherical surface. This gives a strong downward force, making the lower
Teflon-stainless steel junction gas tight. The steel wedges are driven laterally by two
separate linear motion feedthroughs. They enter through heavy guide blocks to
provide strength needed to make the compression seal. In order to provide the
nonrigid character needed to accommodate the downward motion produced by the
Fig. 2.5 Rod-through chamber seal—local sealing Force
tapered wedges, the plunger assembly is coupled to the liquid nitrogen reservoir by
means of an internal bellows. This seal also has a leak rate less that 10−9 Torr L/s [13].
A number of other designs for transferring samples between vacuum systems
have been devised and are discussed in [14–19].
A novel design for a continuous stirred flow reactor for use in studying surface
reaction kinetics at high pressure, but which cannot be used for placing the catalyst
surface in a position for surface analysis, is discussed in [20]. The criteria for
deter-mining that the reactor is behaving as a stirred flow reactor are given in [20].
Another catalytic reactor, which permits the measurement of the spatial distribution
of reactant and product gases, has been described in [21]. This reactor also does not
permit the analysis of the catalyst surface composition by electron spectroscopies.
In this design, by measuring the leak rate for the reactor through a capillary,
the temperature of the gas in the vicinity of the catalyst surface can be measured
[21, 22].
2.2 Novel Sample Transfer from UHV Chamber to External Cell 45
2.2 Novel Sample Transfer from UHV Chamber

to External Cell
A novel design of an ultrahigh vacuum system permitting the rotation of a

manipulator containing a single crystal sample to several analytical ports, while also
conveniently allowing sample access to an external cell, has been described [23]. In
Fig. 2.6a, an octagonal UHV chamber is shown in side view. The different ports on
the octagonal faces contain analytical capabilities of various sorts. The manipulator
is mounted on a doubly differentially pumped rotary seal on a 14-in.-diameter
flange (see p. 9 for one design), and its center line is 5 in. off center, allowing the
sample manipulator to be rotated through 360° on a 5-in. radius circle to face the
various analytical probes mounted on the octagonal faces of the chamber. The
bellows manipulator has 18 in. of motion in the vertical direction, and is attached to
a standard x, y, z-translating stage. The hollow tube holding the crystal may be
rotated through 360° also, allowing axial rotation of the sample, and this rotational
motion is achieved with a differentially pumped rotary seal at the top of the
translating stage (see pp. 7 and 8).
The unique access to an external chamber occurs by pushing the sample through
three spring-loaded Teflon seals [24] that are doubly differentially pumped and that
act as a sliding seal, as in the design shown on Fig. 2.3c. Accurate adjustment of the
position of the vertical manipulator with the x-y-stage allows this to be done without
strain. The Teflon seal section is separated from the large chamber above it with a
small gate valve. The external cell is located below the sliding seal section and has a
small volume. In various cells, high-pressure surface reactions, infrared reflection
absorption spectroscopy, and electrochemistry [25] maybe carried out, or if the cell
is removed, repairs to the exposed crystal and its mounting maybe made. This
permits insertion of new crystal samples into the UHV chamber without requiring
bakeout of the chamber, and the design is therefore a tremendous time saver.
The bottom portion of the feedthrough is polished to allow easy motion through
the spring-loaded Teflon seals and the achievement of a vacuum seal as shown in
Fig. 2.6b. When the sample has been translated back into the UHV chamber, the
gate valve may be closed to prevent outgassing from the external cell. A pressure of
2 atm in the external cell causes no measurable degradation of vacuum in the UHV
chamber (typical pressure = 2 × 10−10 Torr).
2.3 UHV Sample Transfer Device with Low-Temperature

Capability
The transfer of a sample from a manipulator in a preparation chamber to a

manipulator in an ultrahigh vacuum is a classic problem, and this has been dealt
with on pp. 52 and 54. However, when excellent cooling is required to temperatures
below about 100 K, special designs must be used to permit rapid heat transfer
Fig. 2.6 Novel sample transfer to external cell. a UHV chamber. b Sample transport device
2.3 UHV Sample Transfer Device with Low-Temperature Capability 47
through a demountable joint. Various designs to do this are found starting on

p. 383. A particularly appealing design, because of its simplicity, is shown in
Fig. 2.7 [26].
Figure 2.7a shows the basic vacuum system with a sample preparation chamber
separated from the UHV chamber by a straight-through valve. The sample, initially
held at the end of a horizontal manipulator in the sample preparation chamber, can
be moved through the valve into the ultrahigh vacuum chamber, where it is
transferred to the vertically oriented manipulator. This is done by transfer of a
sample slide wedge, shown in Fig. 2.7b. This wedge is made of Mo and is highly
polished on its bottom contact surface to avoid cold welding when assembled. The
sample is attached to a flat pad (0.3 × 0.4 in.2) on the top, using spring clips, screws,
or welding. A long Mo screw, fixed in the sample slide wedge, and rounded on the
threaded end, is tightened by rotation of the transfer rod drawing the wedge shaped
surfaces on the slide and the base together. The base is also made of Mo to avoid
differential thermal expansion, and its upper surface facing out of the drawing is
also polished to receive the slide wedge contact surface. Two screws at the top and
bottom of the heating/cooling base act as stops for the sample slide wedge. With the
correct adjustment of dimensions, there is actually no sliding of the wedge against
the base contact surface; i.e., when the surfaces come together there is no room left
for any sliding.
The heating/cooling base is connected by three screws to a copper cooling
finger, which may be refrigerated with liquid He. A pancake radiation heater (see
p. 364) is attached by a Mo screw to the back of the base [27]. The pancake heater
is surrounded by a radiation shield for maximum efficiency in heating. The cutout
section of the base between the sample region and the copper block provides
sufficiently poor thermal contact when heating to allow a high temperature to be
reached, yet also sufficient thermal conduction for cooling the sample.
Transfer of the sample slide wedge is carried out using the inverted screwdriver,
which rotates the Mo screw, drawing the slide wedge to the base. Two retaining
pins help to align the sample slide wedge and then disengage when the screw is
tightened and the transfer rod is retracted. A clutch spring permits rotation of the
inverted screwdriver while the pins are engaged with the slide wedge.
The temperature range available was shown to be from 32 K to about 1270 K
using 100 W of indirect heating. It was found that Aquadag [28] was needed to
prevent galling of the contact surfaces other than the polished Mo wedge contact
surfaces. A disadvantage of this design is that attachment of a thermocouple to the
sample and transfer of the thermocouple is not provided.
Other transfer systems incorporating low-temperature capabilities for work at
cryogenic temperatures are given in [29, 30].
Fig. 2.7 UHV transfer—low-temperature capability. a System configuration. b Transfer devices

2.4 Internal Cup High-Pressure Cell 49
2.4 Internal Cup High-Pressure Cell
Here is described the provision of an internal cup high-pressure cell that accom-
modates a sample mounted to move in a normal fashion to various analytical ports
in an ultrahigh vacuum chamber [31]. A schematic diagram of the apparatus is
shown in Fig. 2.8a, where a sample, mounted in a special cup, can rotate to various
analytical ports. One of its positions is opposite a reactor cup assembly that can be
Fig. 2.8 Internal cup high-pressure cell. a Schematic of rotating sample + cup assembly.
b Schematic of high-pressure reaction cell
moved laterally by hydraulic pressure. The sample cup is rigidly held by a retainer
assembly, connected to the reactor cup assembly, which prevents the application of
large forces to the manipulator shaft when the seal is made using hydraulic pressure.
Figure 2.8b shows an enlarged schematic view of the reactor cup assembly and its
hydraulic cylinder. When the sample has been rotated into its retainer assembly, the
reactor cup assembly is pressed against the sample cup, and an annealed Cu gasket
produces the seal at a mini-Conflat sealing surface. This operation is assisted by an
internally tapered guiding surface on the outer section of the reactor cup assembly.
This method of sealing allows repetitive use of the same annealed gasket. The
background pressure in the UHV system was found to rise into the low 10−9 Torr
range when the cell was pressurized to 100–2500 Torr. The reactor cell (volume
45 cm3), could be evacuated through gas lines passing through the heavy stainless
steel reactor cup assembly, or more rapid evacuation could be achieved by opening
the pressurized cell inside the UHV system, causing a pressure surge that decreased
into the 10−9 Torr range in 1–2 min. The original paper [31] contains a brief review
of other methods for building high-pressure cells.
A different design, permitting a single crystal catalyst sample to be moved into a
small reactor cup, sealed with a Viton O-ring, has been described previously [32–
34]. In this design also, the single crystal may be moved in an ordinary fashion to
various positions to face different analytical probes mounted radially in an ultrahigh
vacuum system. The design in [33] cleverly permits the reactor cup to be sealed off
from the UHV system when the sample is removed, preventing outgassing effects
from the inner walls from contaminating the UHV system.
2.5 High-Pressure Transfer Cell
A high-pressure transfer cell that is designed to have a small (30 cm3) volume and
provision for sample cooling is shown in Figs. 2.9a and b. Pressures as high as
15 bar are able to be sustained in the high-pressure cell. Sealing is accomplished by
a stainless steel/Teflon seat. Impurities on the crystal that might possibly come from
the Teflon under compression are eliminated in this design [35].
Figure 2.9a shows a schematic design for the transfer rod. The sample is elec-
trically mounted on Mo connector rods with W, Ta, or Mo wires. A thermocouple is
also welded to the sample. The transfer rod enters the chamber through differen-
tially pumped spring loaded Teflon seals (not shown) such as those mentioned on
pp. 6–11 [36]. These seals permit both translation and rotation of the sample when
it is in the UHV chamber. The transfer rod is self aligning as it enters the Teflon
sealing position. The compression of a spring permits the force on the Teflon seal to
be measured and reproduced at about 400 lb. Heating of the sample in vacuum to
1300 K is possible by resistive heating. Heating of the sample to about 700 K is
possible in 10 bar of H2 if cooling is achieved by using flowing air rather than liquid
nitrogen. Cooling to about 130 K can be achieved with liquid nitrogen coolant. In
order to prevent extreme cooling of the transfer rod when liquid nitrogen is used, a
2.5 High-Pressure Transfer Cell 51
Fig. 2.9 High-pressure transfer-cell. a Transfer rod for UHV. b High-pressure cell
U-shaped Cu tube is inserted inside the transfer rod wall so that it is held com-
pressively against the inner wall, and warm, dry air is continuously circulated
through this tube when liquid nitrogen is used. This prevents failure of the sliding
seal mechanism owing to cooling. Teflon spacers are inserted equidistantly inside
the transfer rod using a pusher rod with calibrated markings. The outside of the
transfer rod is polished to the correct rms roughness following the directions of the
spring-loaded Teflon seal manufacturer [36]. The inside of the transfer rod is
evacuated with a dual-stage mechanical pump when liquid nitrogen cooling is
employed. The outside of the rod is protected with a hypalon/nylon bellows [37]
and this bellows is purged continuously with nitrogen gas.
Figure 2.9b shows the transfer rod sealed into the high-pressure chamber. An
alignment insert, H, is brazed between the vacuum housing and the high-pressure
cell. It is made of a bakeable and UHV-compatible soft aluminized bronze [38] so
that it does not scratch or damage the transfer rod. The annular space between the
Teflon seat and the high-pressure chamber is of the order of 0.05 cm, and its length
is of the order of 2.5 cm. Thus, at high pressure, possible diffusion of impurities
from the Teflon to the sample is minimized by the long diffusion time. Calculations
of the diffusion time indicate that at 10 bar of gas pressure, diffusion times of
greater than 30 min are involved over the 6.5 cm distance from the Teflon to the
sample.
The high-pressure cell is directly attached to a gas handling manifold that serves
to introduce gases as well as to extract them for gas chromatographic analysis. In
addition, direct pumping of the high-pressure cell by a small turbopump may be
carried out.
Proof of the performance was made in two ways. First it was shown that a
pressure of 10 bar of H2 led to an increase of pressure of only 1 × 10−9 Torr in the
UHV chamber. Secondly, storage of an atomically clean Cu crystal for up to 30 min
in the evacuated high-pressure cell does not lead to measurable contamination as
measured by XPS or TPD methods. Also, long reaction periods at high pressure
occur with constant reaction kinetics on a Cu single crystal surface [35].
2.6 Sample Transfer with Disconnect-I
2.6.1 Transfer in Standard UHV Systems
Often it may be desirable to transfer a sample from one chamber to another and to
disconnect the sample from the transfer device after transfer. In Fig. 2.10 one may
see that transfer from a reaction chamber to an ultrahigh vacuum chamber is
accomplished by use of a magnetic transfer rod, which is commercially available.
The sample, mounted on a ceramic sample holder ring at the end of the transfer rod,
sits between three tines on the fork at the end of the transfer arm. The hooks of the
tines lock into the sample holder ring by a slight twist, which may be accomplished
2.6 Sample Transfer with Disconnect-I 53
Fig. 2.10 Longitudinal sample transfer—disconnect
with the transfer magnet. The electrical connecting pins are machined from 6–32
screws as shown in the central Figure, and these pins insert into homemade sockets
made if necessary from alloys designed for thermocouple connections, etc. It is
stated that the electrical connections do not have to be particularly tight and that
ordinary misalignments in the pins supply the friction level needed [39]. The
ceramic in this case was cast from machinable ceramic [40] and fired in a hobbyist’s
ceramic kiln.
The sample is transferred to a manipulator that has a central translating rod used
to drive the tilt mechanism. This tilting action is carried out by moving the central
manipulator rod up and down, causing the base holding the sample to tilt backward
and forward as desired. A locking nut on this rod prevents overdriving the tilting
mechanism. In the bottom portion of Fig. 2.10, two types of samples are shown. On
the left, a metallic foil substrate is welded to the screwhead ends of the mounting
pins in the base. On the right, a single crystal of a metal is welded to a circular piece
of foil which is welded to the mounting pins. A circular hole in the backing foil will
permit electron bombardment of the back of the crystal. One might provide a
radiation shield between the emitter filament and the ceramic base to reduce heating
of the ceramic and outgassing.
The proper location of viewports for monitoring the sample transfer is recom-
mended, or else two persons may be needed to transfer the sample.
2.6.2 Transfer in STM Systems
The special considerations needed for the transfer of samples being used for scan-
ning tunneling microscopy (STM) studies as well as other more conventional surface
science studies have been reported by Sullivan and Kummel [41]. Here, a design
coupling the STM sample to a cold bath, electron beam heating, two degrees of
rotation, and electrical isolation, along with sample transfer are described, and a
review of other manipulator designs is given [41]. Noteworthy is a simple design for
transferring sample holders between forklike manipulators [42]. (See also p. 45.)
2.7 Sample Transfer with Disconnect-II
A simple and inexpensive method to transfer a sample with disconnect [43] is

shown in Fig. 2.11. The sample is mounted on stainless steel or other metal posts
using Ta clamps that bolt into the base of the holder as shown. This assembly is
placed on a transfer screw rod in the the loadlock chamber, and then, after loadlock
chamber evacuation, the valve between chambers is opened and the sample
assembly is passed into the UHV chamber and plugged into the female banana plug
receptacles. These receptacles are mounted on a rod that is connected to the rotary
manipulator in the UHV chamber. Electrical power is supplied through Cu braid
(see p. 392) as shown. The sample transfer screw rod is then unscrewed and
withdrawn. The UHV banana plugs [44] tend to allow the sample to sag a little, and
2.7 Sample Transfer with Disconnect-II 55
Fig. 2.11 Sample transfer with disconnect
if this is unacceptable, then a mechanical modification to the design near the banana
plugs can be added.
In cases where ohmic heating of the sample is inappropriate, it may be backed
with a section of a Si wafer (approximately 0.1 Ω cm) for indirect resistive heating.
2.8 Vibration-Free Translatable Device
It is often necessary to perform a delicate translation under conditions where, after

adjustment, the translated device is vibrationally decoupled from the UHV system.
The example shown in Fig. 2.12 is used to make mechanical/thermal contact of a
Fig. 2.12 Vibration-free translatable device

detector head (consisting of a polymer pyroelectric foil material) to the back side of
a sample experiencing adsorption from a molecular beam, as for example in a
microcalorimeter as described in Chap. 163. The sample is held in an
electrically-isolated Mo sample holder and the movable detector head may be
adjusted on a Cu platform to make contact with the back of the sample, using an
Allen wrench to engage a translation screw. The Allen wrench can be rotated using
a rotatable wobble stick. The entire assembly is mounted on a vibration damping
stack which rests on a plate on the bottom of the chamber. Thus, when the trans-
lation has been achieved and the Allen wrench is disengaged, the detector and the
attached sample is free of vibration [45].
2.9 Mirror Port for Enhanced Sample Transfer
Often sample transfer from one manipulator to another must be achieved in a

loadlock chamber in order to pass the sample through a gate valve and on to another
chamber for study. The transfer process can be improved by having a mirror
mounted on an external viewport which gives a wide angle view throughout the
chamber when one looks through small viewports in the load lock chamber [46]
(Fig. 2.13).
2.10 UHV Access Port
The entry into an ultrahigh vacuum chamber can often be achieved through a
flanged entry port that contains a standard copper gasket. For large ports that permit
access to the entire chamber, one wall of the system may be designed to be
reversibly removable without a need for gasket replacement [47].
The design for a removable wall for a rectangular access port is shown in
Fig. 2.14. The port contains spring-loaded Teflon [48] seals in separate wells. The
regions between the Teflon seals are differentially pumped to prevent small leakage
from influencing the vacuum conditions in the chamber. Thus, an ion pump is used
on the pumping region closest to the chamber, and a roughing pump is used on the
pumping region closest to the atmosphere. The access port is arranged, by means of
a recessed section, to be pulled down tightly against the mating surface, providing
some further sealing at the very inner section of the seal.
2.10 UHV Access Port 57
Fig. 2.13 Mirror port for enhanced sample transfer
Fig. 2.14 UHV access port

References
1. D.R. Kahn, E.E. Petersen, G.A. Somorjai, J. Catal. 34, 294 (1974)
2. D.W. Blakely, E.I. Kozak, B.A. Sexton, G.A. Somorjai, J. Vac. Sci. Tech. 13, 1091 (1976)
3. G.A. Somorjai, Surf. Sci., 89, 496 (1979)
4. A.L. Cabrera, N.D. Spencer, E. Kozak, P.W Davies, G.A. Somorjai, Rev. Sci. Instrum. 53,
1888 (1982)
5. T.G. Rucker, K. Franck, D. Colomb, M.A. Logan, G.A. Somorjai, Rev. Sci. Instrum. 58, 2292
(1987)
6. D.W. Goodman, R.D. Kelley, T.E. Madey, J.T. Yates, Jr., J. Catal. 63,226 (1980); see also, D.
W. Goodman, J. Vac. Sci. Technol. 20,522 (1982)
7. D.W. Goodman, J.T. Yates Jr, J. Catal. 82, 255 (1983)
8. H.J. Krebs, H.P. Bonzel, G. Gamer, Surf. Sci. 88, 269 (1979)
9. H.P. Bonzel, G. Broden, H.J. Krebs, Appl. Surf. Sci. 16, 373 (1983)
10. C.T. Campbell, M.T. Paffett, Surf. Sci. 139, 396 (1984)
11. C.T. Campbell, J. Vac. Sci. Tech. A2, 1024 (1984)
12. Dr. Renato Zenobi, Analytical Chemistry, Universitätstr. 16, ETH Zentrum CHN C59,
CH-8092 Zurich, Switzerland (private communication)
13. T. Yamada, T. Misono, K.I. Tanaka, Y. Murata, J. Vac. Sci. Tech. A7, 2808 (1989)
14. C.A. Crider, G. Cisneros, P. Mark, J.D. Levine, J. Vac. Sci. Technol. 13, 1202 (1976)
15. J.P. Hobson, E.V. Kornelsen, J. Vac. Sci. Technol. 16, 701 (1979)
16. R.E. Clausing, L. Heatherly, L.C. Emerson, J. Vac. Sci. Technol. 16, 708 (1979)
17. J. Klebanoff, V.H. Ritz, R.E. Thomas, J. Vac. Sci. Technol. A2, 1396 (1984)
18. G.S. Chottiner, W.D. Jennings, K.I. Pandya, J. Vac. Sci. Technol. A5, 2970 (1987)
19. R. Raval, M.A. Harrison, D.A. King, G. Caine, J. Vac. Sci. Technol. A9, 345 (1991)
20. J.J. Vajo, W. Tsai, W.H. Weinberg, Rev. Sci. Instrum. 56, 1439 (1985)
21. S. Lundgren, K.E. Keck, B. Kasemo, Rev. Sci. Instrum. 65, 2696 (1994)
22. B. Kasemo, Rev. Sci. Instrum. 50, 1602 (1979)
23. (a) R.A. Campbell and D.W. Goodman, Rev. Sci. Instrum. 63, 172 (1992). (b) J. Szanyi and
D.W. Goodman, Rev. Sci. Instrum. 64, 2350 (1993). These articles also contain useful
references to other external high-pressure cell designs
24. Furon Co., 3336 E. Lapaima Ave., Anaheim, CA 92806
25. L.-W.H. Leung, T.W. Gregg, D.W. Goodman, Rev. Sci. Instrum. 62, 1857 (1991)
26. N.J. DiNardo, J.E. Demuth, W.A. Thompson, P.G. Ledermann, Rev. Sci. Instrum. 55, 1492
(1984)
27. Custom as well as standard heaters are available from Spectra-Mat, Inc., 100 Westgate Drive,
Watsonville, CA 95076. The version used here has W/Re alloy wires embedded in an alumina
matrix and housed in a Mo can
28. Aquadag is a colloidal graphite paint, which can be brushed onto a surface and then outgassed
by bakeout. It provides lubrication and is electrically conductive. Obtain Dag 154 from
Ache-son Colloids, 1607 Washington Avenue, P.O. Drawer 611747, Port Huron, MI
48061-1747
29. G.S. Chottiner, W.D. Jennings, K.I. Pandya, J. Vac. Sci. Technol. A5, 2970 (1987)
30. R. Raval, M.A. Harrison, D.A. King, G. Caine, J. Vac. Sci. Technol. A9, 345 (1991)
31. Y.-N. Wang, R. McAllister, R.G. Herman, G.W. Simmons, K. Klier, Rev. Sci. Instrum. 63,
5767 (1992)
32. J.J. Weimer, F.A. Putnam, Rev. Sci. Instrum. 55, 238 (1984)
33. K.-E. Keck, B. Kasemo, A. Hoglund, Rev. Sci. Instrum. 54, 574 (1983)
34. M.J. Mummey, L.D. Schmidt, Surf. Sci. 91, 301 (1980)
35. A. Ludviksson, J. Yoshihara, C.T. Campbell, Rev. Sci. Instrum. 66, 4370 (1995). This paper
also contains references to many other transfer devices in the literature
36. Obtained from: Fluorocarbon Co., 10871 Kyle Street, P.O. Box 520, Los Alamitos, CA 90720
37. Gortiflex Protector Bellows
References 59
38. No. UNS C95400, Copper and Brass Sales, 17401 Ten Mile, East Detroit, MI 48021
39. A. Garcia, M.E. Kordesch, J. Vac. Sci. Tech. A11, 461 (1993)
40. Aremco, P.O. Box 429,23 Snowden Avenue, Ossining, NY 10562-0429
41. D.J.D. Sullivan, A.C. Kummel, Rev. Sci. Instrum. 63, 4285 (1992)
42. B.J. Mulder, J. Phys. E12, 908 (1979)
43. Dr. Ray Kaplan, Mail Code 6863, Naval Research Laboratory, Washington, DC 20375
44. UHV banana plugs (Model Nos. BPBPM and BPBPF) can be obtained from: Surface/Interface
Inc., 110 Pioneer Way, Suite H, Mountain View, CA 94041
45. S. Schauermann, private communication
46. Kejing Li, private communication
47. D.J. Auerbach, C.A. Becker, J.P. Cowin, L. Wharton, Rev. Sci. Instrum. 49, 1518 (1978)
48. The spring-loaded Teflon seals are available from the Fluorocarbon Co., 10871 Kyle Street,
RO. Box 520, Los Alamitos, CA 90720. Many sizes of the seals are available, and instructions
for their use are given by the company
Chapter 3
Electrical Connections
3.1 Tungsten Ribbon—Attachment to Power Leads
Heated tungsten ribbons are used in many experiments as radiation sources, metal
deposition sources, electron bombardment sources, and molecular cracking sources.
Most commonly, the ribbons are attached to heating leads by spot welding. This
method requires that the mounted power leads be removed from the vacuum system
for ribbon replacement, a time-consuming procedure. An alternative method for
ribbon attachment shown in Fig. 3.1 may be carried out inside the vacuum system
without welding [1].
A Ta feedthrough lead has a slot milled into it as shown in Fig. 3.1a. The length
of the slot is about 1.5 times the ribbon width, and the width of the slot is about 1.5
times the ribbon thickness. A small cap is machined from Ta, for example, as
shown in Fig. 3.1b. The inside diameter of the cap is larger than that of the
feedthrough lead by one thickness of the ribbon. The cap length is three to four
times that of the ribbon width, and the depth of the bore is 2.5 times the ribbon
width, leaving plenty of material in the slotted upper region. The cutout on the cap
is centered on the cap so as to allow the ribbon to pass as shown in Fig. 3.1c. The
end of the lead extends internally above the cutout by about 0.5 mm, as shown by
the dotted lines in Fig. 3.1c, preventing the cap from tilting when rotated.
To assemble the device, the ribbon is passed through the slot in the lead with the
cap in place, and extends about one-half the lead diameter beyond the cap. To
secure the ribbon, the cap is rotated one-half turn as shown, wrapping the ribbon
around the lead under the cap, and locking it in place.
This method will work on very thin W ribbon and has been extended to
0.010-mm-thick ribbon material, 1.2 mm wide; this particular mounting was done
on 1.5-mm-diameter Ta leads. To keep the ribbon straight after it has been used, it
is common to strain the two leads toward each other before making the second cap
attachment. The strain in the two supporting leads will place tension on the ribbon
and keep it straight.

DOI 10.1007/978-3-319-17668-0_3
62 3 Electrical Connections
Fig. 3.1 Tungsten ribbon—attachment to power leads
3.2 Wrap Connection—Tungsten to Tungsten
Tungsten is a brittle metal that is difficult to bend or wrap. In addition, tungsten is

also difficult to weld. The method shown in Fig. 3.2 involves a wrapping technique
that produces tungsten coils that fit onto tungsten wires tightly without welding.
Such coils can be used as a friction support or locator in high-temperature appli-
cations [2].
Splintering or breakage during fabrication of tungsten wire above about
0.25 mm diameter is a common occurrence owing to its inherent brittle character.
While this may be avoided by bending tungsten wire while it is being heated to
redness in a flame, heavy oxidation will occur and also intricate bends are difficult
to achieve before the tungsten cools. The method shown here prevents splintering
of 0.5-mm-diameter W wire when it is bent into a tight coil. Such coils maybe used
to make strong friction contacts to 0.5-mm W wire as will be shown.
A hand drill is mounted in a vise, and a core wire of W, 0.5 mm in diameter, is
held in the jaws of the chuck. The second piece of similar W wire to be wrapped is
also inserted into the chuck between the edges of the jaw pieces as shown so that it
will be prevented from slipping when the wrapping begins. About two revolutions
of the chuck are carried out, giving preliminary contact between the core wire and
the wrapping wire; the wrapping wire does not start out directed normally from the
3.2 Wrap Connection—Tungsten to Tungsten 63
Fig. 3.2 Wrap connection—Tungsten to Tungsten. a Starting coil. b During winding. c Final coil
—slightly bent in middle
axis of rotation. A wood-forming clamp is next placed on the wrapping wire coil,
and while under pressure, the coil is formed to the desired length. The friction
between the wood and the W will prevent fracture. The ends of the wrap wire coil
can now be cut to the desired length using tungsten carbide end nippers and
dressing on a grinding wheel if necessary.
The coil is removed from the core wire, and a slight bend is placed in the center
of the coil, using either two pairs of pliers or a knife blade inserted partially into the
coil. This slight bend allows the coil to be forced onto a 0.5 mm W wire and to
remain fixed owing to the strain induced in the coil as it is straightened by the inner
wire. The coil will not move at high temperature either during ohmic heating or
bakeout.
Fig. 3.3 Cold-formed wire connector
3.3 Cold-Formed Wire Connector
The fabrication of electrical connections that are conductive for both electricity and
heat may be carried out using silver soldering, welding, or screw-type connectors.
Another method, involving a press fit of a hard wire into a soft material such as Cu
or Al, is shown in Fig. 3.3. Here a vise is used to press the junction, and the various
stages of compression are illustrated. The last stage (d) involves deformation of the
soft metal and enclosure of the hard wire. The resistance of these junctions between
tungsten and copper is typically about 0.05 Ω [3].
3.4 Ultrahigh Resistance Vacuum Feedthroughs
The measurement of the electrical conductivity of materials with a very high

resistance can be plagued by lower-resistance pathways to ground that interfere
with the electrical measurements. In particular, Al2O3 ceramic insulators exhibit
resistances of about 1012 Ω, and these may deteriorate in their resistance with time
due to the adsorption of water on the alumina or to the deposition of thin metal films
on the vacuum side of the insulator surface. Such problems are termed electrical
leakage problems, and prevent measurements into the 1016 Ω range, which are often
needed [4]. In some cases, using an enclosure containing drying agent on the
outside of the feedthrough will help, and the more complicated use of forepump
evacuation of this region will prevent water adsorption there. For the highest
resistivity feedthroughs, other methods must be employed, as shown in Fig. 3.4.
The Figure shows the construction principles of a doubly insulated and guarded
design that permits measurements down to the 1016 Ω range. A special crystalline
sapphire-metal feedthrough [5] is employed, and this is brazed into an ordinary Cu
3.4 Ultrahigh Resistance Vacuum Feedthroughs 65
Fig. 3.4 Ultrahigh resistance vacuum feedthrough
tubing feedthrough mounted on a flange by means of an alumina ceramic insulator.

The sapphire feedthrough contains a hollow tube. Into this tube is placed either a Ni
or Cu conductor wire, which is silver brazed under hydrogen into the hollow tube.
At the same time the brazing of the sapphire thimble and the flared Cu tubing is
carried out. Great care must be taken to avoid extreme thermal shocks at the
sapphire. The assembly on the vacuum side is provided with a second shield as
shown, and also shielding of the outer air side of the feedthrough is carried out as
shown.
The impedance to ground of the sapphire insulator approached 1016 Ω when the
sapphire was thoroughly cleaned [4]. Spectroscopic grade alcohol, followed by
distilled water spraying through a fine glass nozzle, is recommended, followed by
blotting the excess water and drying in an oven.
3.5 Sliding Metal Electrical Contact
A need for a sliding electrical contact can occur when flexible coils of ceramic
bead-insulated wire are unsatisfactory, perhaps because of the formation of ceramic
dust as motion occurs. Figure 3.5 shows a sliding stainless steel rod that moves
through a Teflon bearing that is mounted inside a re-entrant stainless tube connected
to a vacuum linear drive of the bellows or magnetic type. Electrical contact to the
rod is made by a phosphor bronze spring contact rigidly mounted on a stainless
support assembly, and electrically isolated by ceramic insulator cylinders that
contain small bolts for holding the spring contact strip. The phosphor bronze spring
strip is electrically connected to an electrical feedthrough as shown. The holes
drilled in the Teflon-bearing should be about 5 % larger than the sliding rod to
allow for deformation of the Teflon as well as for its large thermal expansion
coefficient. The contact can pass a 2-A current. The stainless steel rod can be moved
Fig. 3.5 Sliding metal electrical contact

3.5 Sliding Metal Electrical Contact 67
in and out at 5 cm/s without difficulty, and cold welding in ultrahigh vacuum does
not occur [1]. The authors also report that a Ta wire spring may be used success-
fully against a Mo slider in a rotary device [6].
3.6 Low-Profile Electrical Lead Clamp
There is often a need for a band clamp arrangement to anchor electrical leads to a
manipulator rod or other structure. Often, narrow clearance spacing is available
around the rod so that the binding on the electrical leads to the rod must be achieved
in a limited annular space. Figure 3.6 illustrates one solution to this problem, using
half-hard Cu-Be spring stock. This material is formed as desired by using
steel-forming wires of the correct diameter to produce the rounded grooves in the
Cu-Be stock by hammering the forming wire into the stock with the forming jig as a
guide to fabrication. An extra channel is provided as an additional spring expansion
channel to give the band extra elasticity. The tabs at the end of the band are each
about 3–4 mm in length. To install, the ceramic insulated wires are placed into
Fig. 3.6 Band clamp for electrical leads

position on the rod and temporarily anchored. Overlapping the Cu-Be tabs and
folding them into each other will result in a tension holding the band and the
ceramic insulated wires in place. The band will not soften or lose its spring quality
from −200 °C to bakeout temperatures [7].
References
1. Mr. Gerald L. Fowler, 6816 Kelly Ave., N.E., Albuquerque, NM 87109 (private
communication)
2. Mr. G.L. Fowler, 6816 Kelly Ave., N.E., Albuquerque, NM 87109 (private communication)
3. D. Edwards Jr, J. Vac. Sci. Tech. 16, 970 (1979)
4. A.H. Madjid, J.M. Martinez, A.A. Vaala, Rev. Sci. Instrum. 43, 1049 (1972)
5. Obtained from: Cory Lab, Inc., 823 5th Street, Menominee, MI 49858
6. S.M. Newstead, R.A.A. Kubiak, E.H.C. Parker, J. Vac. Sci. Tech. A2, 1603 (1984)
7. G.L. Fowler, J. Vac. Sci. Tech. A5, 2978 (1987)
Chapter 4
Pumping and Trapping
4.1 Water Aspirator/Sorption Pump Combination

for Efficient UHV System Evacuation
On ion-pumped UHV systems, it is of course necessary to bring the system pressure

down from atmospheric to about 10−2 Torr before starting the ion pumps. This may
be done in several ways, including the use of zeolite sorption pumps that are cooled
to liquid nitrogen temperatures in order to physically adsorb all of the atmospheric
gases. Because of the large quantity of air that must be removed by the zeolite, the
process of reaching the limiting pressure sometimes exceeds 1 h. A method to
remove the majority of the atmospheric gas has been reported that is free from the
possibility of causing oil contamination [1]. The use of a rotary pump does not
exclude this possibility.
Figure 4.1 shows a schematic diagram of the use of a water aspirator for
removing the majority of the atmospheric gas. The aspirator is prevented from
admitting water to the system (equilibrium vapor pressure of water = 24 Torr at
25 °C) by interposing a liquid nitrogen trap. After this aspirator evacuation, which
takes about 1–2 min, the aspirator is isolated from the pumping system by closing
valves V1 and V2. The sorption pump is then cooled with liquid nitrogen, and valve
V2 is opened. It is observed that the pressure in the vacuum system is reduced to
about 5 × 10−3 Torr in only about 10 min, since the capacity of the zeolite is much
greater than the quantity of air that remains to be pumped after the aspirator
evacuation. Several hours of pumping with the zeolite trap used alone would be
required to go to this pressure range.
Following the initial pumpout of the system, valves V4 and V3 are closed, and
the bake-out processing of the system is carried out to reach UHV conditions, as
discussed on p. 83. This procedure allows the sorption pump to be cycled several
times without the need for warmup and vaporization of the sorbed gases.

DOI 10.1007/978-3-319-17668-0_4
70 4 Pumping and Trapping
Fig. 4.1 Water aspirator/sorption pump combination

4.2 Ballast Pumping—Vibration Free 71
4.2 Ballast Pumping—Vibration Free
There are often situations in which the elimination of mechanical vibrations in an

ultrahigh vacuum system would require turning off the mechanical forepump. This
is feasible using the method to be described [2], since the gas load being pumped
from the system is very low if ultrahigh vacuum conditions exist there. The system
is fore-pumped by a large ballast volume of the order of 50 L, which is isolated
from the mechanical pump by a valve as shown in Fig. 4.2. The ballast will pump
for several days without significant degradation of the turbopump’s ability to
pump. It can be evacuated in minutes after several days’ operation where the
forepump has been turned off.
This arrangement is also useful for maintenance of the forepump, since it may be
disconnected from the ballast after the isolation valve is closed.
The same idea may be applied to diffusion pumps where large gas loads are not
being handled.
A variation on this idea is to back a turbopump with an ion pump over evenings
and the weekend. If the power should fail, problems with oil sucking back into the
system from the forepump will be avoided [3].
Fig. 4.2 Ballast pumping—vibration free

4.3 The Use of Appendage Titanium Sublimation Pumps

for Pumping Low Gas Loads in Ultrahigh Vacuum
Getter films of active metals have been used for many years in ultrahigh vacuum
technology, especially in the days when work was carried out in glass vacuum
systems that were sealed off from dynamic pumping [4]. This method of pumping
chemically active gases is superior to all other methods insofar as secondary effects
are concerned. The principle of getter film pumping is the strong irreversible
chemisorption of chemically active gases by clean active metal surfaces. Simple
considerations of monolayer capacity indicate that 1 cm2 of a getter surface will
pump of the order of 1–3 × 10−5 Torr L of an active gas. Film roughness factors of
2 or 3 will increase the gettering capacity accordingly.
One example of the use of a Ti sublimation pump (TSP) for pumping low flux
gases from a differentially pumped mass spectrometer is shown in Fig. 4.3. Here, a
thin film of Ti is deposited on the stainless steel wall of a cylindrical appendage
each morning before beginning work. The Ti film is kept at 77 K to maintain its
pumping efficiency to the highest level. A single Ti film will remain active all day
during use of the mass spectrometer at low gas loads. This particular apparatus also
contains a 60-L/s diode ion pump for maintenance of vacuum quality in the mass
spectrometer overnight when the titanium pump is not being used [5].
Careful studies, in which the Ti getter film pump has been compared with ion
pumping, show that regurgitation effects [6] involving methane liberation from the
ion pump are completely eliminated when the Ti pump is employed [5]. This may
be seen by looking at the thermal desorption spectra of methyl chloride from an
inert surface. The use of the TSP eliminates the CH4 evolution originating in the ion
pump. However, wall displacement effects caused by molecules entering the mass
spectrometer region from the single crystal when it is heated in thermal desorption
experiments can still be detected. These displaced molecules were H2, CO, and H2O
in the particular experiments carried out [5].
The detailed behavior of the residual gases in a liquid nitrogen-cooled titanium
sublimation pumping system has been reported [7].
4.4 Pressure-Equalizing Device in Complex Systems
Often in systems such as those used in molecular beam research, it is undesirable

for large pressure differences to exist during either pumpdown or during vacuum
failure in one section of the apparatus. Large pressure differences, and the associ-
ated flow of gas, can damage fragile parts such as thin walls, skimmers, and orifices.
The device shown in Fig. 4.4 is designed to relieve large pressure differences
between sections of a vacuum system that are pumped in parallel [8]. Thin metal
pressure-release sheets are designed to open in either direction if a large pressure
difference occurs. The thin metal sheets are mounted inside flanges that accept
4.4 Pressure-Equalizing Device in Complex Systems 73
Fig. 4.3 Appendage titanium sublimation pump
Viton O-ring seal rings. The O-ring seal ring is brazed to a disk containing two
holes, and the thin metal sheets are mounted on opposite sides of the disk in order to
relieve pressure differences in either direction.
Fig. 4.4 Pressure-equalizing device. a Chambers/pumps in parallel. b pressure-release device

4.5 Diode Ion Pump Performance in He Pumping 75
4.5 Diode Ion Pump Performance in He Pumping
There is sometimes a need to fill a vacuum system with a high-pressure inert purge
gas, and then to recover ultrahigh vacuum. Diode ion pumps have only a limited
pumping speed for inert gases such as He, and their performance is of interest in
this application; triode ion pumps are designed to have good pumping speed for
inert gases at a small sacrifice in total pumping speed.
In Fig. 4.5, a series of pumping tests were performed on an ultrahigh vacuum
system that could be filled with high-purity He at 1 atm pressure. The He was
pumped out with a trapped mechanical pump to about 1 × 10−5 Torr, and then the
diode ion pump was employed. Figure 4.5 shows selected pumping performance
curves after the first He filling and the fifth. The performance degrades in successive
tests because of traces of water vapor admitted with the high-purity He and also
because of a drop in efficiency of the diode ion pump owing to the regurgitation
effect in which previously pumped gas is partially evolved in successive pumping
periods. Baking out (BO) the ion pump (IP) only for 4 h at 250 °C restores some of
the original pumping speed; baking out the ion pump plus the vacuum system
restores all of the original performance seen in the first test.
The behavior shown in Fig. 4.5 indicates that He can be pumped well by a diode
ion pump from ultrahigh vacuum systems made of stainless steel using highly
purified He as a purge gas, and that short bakeouts at 250 °C restore system
pumping performance. It was suspected that residual water in the He (1.9 ppm) was
partially responsible for degradation of the pumping behavior in successive tests
[9].
4.6 Vacuum Applications of Metal Foams
Copper [10], nickel [10] and aluminum [11] foams are commercially available and
may be used in a number of vacuum applications [12]. The foams are highly
porous, they are easily shaped and machined, and they have some inherent strength.
They also can be easily brazed or soldered.
Four vacuum applications of foamed metals are illustrated in Fig. 4.6. In panel
A, a plug of foamed Cu is used as a hydrocarbon trap between a forepump and a
vacuum system. The high surface area and the torturous pathways in the foam
impede hydrocarbon passage through the trap. In panel B, a zeolite trap is retained
between porous Cu plugs, and serves also as a hydrocarbon trap. The zeolite must
be baked out prior to becoming active and acts as a molecular sieve for hydrocarbon
molecules. Panel C shows a plug of Cu foam used as a conductance limiter under
molecular flow conditions. To adjust the conductance accurately, additional Cu
foam segments can be added as needed. Finally, in panel D, a cylindrical piece of
metal foam is used as a diffuser for injection of gases in random directions into a
vacuum system.
Fig. 4.5 Test chamber and pumping behavior after He purge (1 atm)
4.6 Vacuum Applications of Metal Foams 77
Fig. 4.6 Vacuum applications of metal foams. a Foreline hydrocarbon trap. b Foreline zeolite
trap. c Molecular flow conductance limiter. d Gas diffuser
Metal foam has a cellular internal structure. Low-density foam has cell diameters
in the range 0.5–1.5 μm, whereas high density foam has cell diameters from 9 to
40 μm. Typical surface areas range from 0.025 to 0.038 m2/g. For low-density foam
the density is typically from 2 to 5.5 % of that of the parent metal. For high-density
foam, the density ranges from 20 to 60 % of the bulk metal. Low-density foam can
be compressed by up to a factor of 7.
Measurements of the conductance of 2.4 mm thick disks of 5 % density Cu
foams made by compression of 1.9 % density foam gave a conductance (for air?) of
3.2 L/s cm2.
4.7 Cleaning of Ion Pumps by Chemical Etching
There are a number of companies that will take in used ion pumps for recondi-
tioning at a fraction of the cost of purchasing a new ion pump [13]. In cases where
the pump is a large one, this may be done by disassembly and cleaning or
replacement of the individual components. In the case of small ion pumps, it may
be necessary to open the pump casing to get at the inner components, and then to
reweld the reconditioned components back into the cleaned casing.
The reconditioning of a dirty ion pump can also be done by controlled chemical
etching procedures as described for two cases below.
Case A: Chemical Etching After Disassembly
The pump casing is separated from the Ti cathode assembly as shown in Fig. 4.7.
The surface region of the Ti containing implanted impurities is removed by etching
the assembly, containing electrode connectors and supports, in a water solution
containing 9 % HNO3, 12 % HF(aq), and 0.4 % H2O2 by weight. A 3-s dip,
followed by repeated rinsing in deionized water, should produce shiny Ti surfaces.
The pump casing is etched in a standard solution (Varian) containing 30 %
Fig. 4.7 Chemical cleaning—Ion pump

4.7 Cleaning of Ion Pumps by Chemical Etching 79
HNO3(aq), and 3 % HF(aq) by volume. The carbon deposit must be removed

mechanically by fine abrasive paper; or, alternatively, it may float off with the Ti
film during etching. The pump elements are dried in warm air, assembled, and
pumped into a second ion pump while baking at 200 °C [14].
Case B: Chemical Etching Without Disassembly
In this situation, the fact that the acid attacks the Ti more readily than the
ceramic-metal seal is employed. First the pump is rinsed in deionized water. This is
reported [14] to seal the ceramic partially against the acid. A water solution of 30 %
HNO3(aq) and 3 % HF(aq) by volume is added and quickly rinsed away with
deionized water. The pump is dried by warm air and baked and pumped as in
Case A.
The original report indicates that the hydrogen background pressure from an
acid-cleaned pump will be somewhat higher than from one reassembled using new
elements [14].
4.8 An Efficient Liquid Nitrogen Trap
Liquid nitrogen traps are highly inefficient because of the discharge of liquid
nitrogen out of the trap when the trap is being refilled periodically. A design, shown
made of glass in Fig. 4.8, eliminates this problem [15], and reduces liquid nitrogen
consumption by severalfold, resulting in great economies over long periods of time.
The basic concept embodied in this small glass trap is transferable to large metal
traps.
The trap consists of an inner trap of conventional design, labeled the inner
chamber, surrounded by a second outer chamber, which is in turn surrounded by a
silvered Dewar flask. All of these are sealed together by glassblowing in this design.
A resistance temperature sensor is immersed in the second outer chamber.
When the trap is at a condition where refilling is to occur, all liquid nitrogen in
the Dewar has been lost, and the level of nitrogen in the outer chamber has dropped
to a level causing the sensor to turn on the liquid nitrogen supply, which is driven
by compressed nitrogen gas. Liquid nitrogen flows into the Dewar and spills into
the outer chamber through the lower hole; it then fills the outer chamber until the
sensor delivers a shutoff signal.
The avoidance of liquid nitrogen loss is attributed to the fact that the Dewar does
not contain liquid nitrogen at the time of filling, so that the burst of nitrogen gas at
the beginning of the fill sequence does not eject liquid nitrogen from the trap. In
addition, ice buildup does not occur at the inlet and outlet to the trap [15].
Fig. 4.8 Efficient liquid N2 trap
References
1. A.M. Russell, Rev. Sci. Instrum. 36, 854 (1965)

2. Dr. Oliver Schaff, Fritz Haber Institut der Max Planck Gessellschaft, Faradayweg 4–6,
D-14195, Berlin (private communication)
3. Dr. W. F. Egelhoff, National Institute of Standards and Technology, Gaithersburg, MD
4. P.A. Redhead, E.V. Kornelsen, J.P. Hobson, Can. J. Phys. 40, 1814 (1962)
5. G. Lu, A.L. Linsebigler, J.T. Yates Jr, J. Vac. Sci. Technol. A12, 384 (1994)
References 81
6. P.A. Redhead, J.P. Hobson, E.V. Kornelsen, The Physical Basis of Ultrahigh Vacuum (AIP,
New York, 1993), pp. 403–406
7. L.Y.L. Shen, Rev. Sci. Instrum. 43, 1301 (1972)
8. Dr. R. David, Institut fur Grenzflächenforschung und Vakuum physik, Forschungszentrum
Jülich GmbH, Postfach 1913, D-52425 Jülich, Germany (private communication)
9. P.M. Danielson, J. Vac. Sci. Technol. 5, 89 (1968)
10. Foametal, Inc., 37645 Vine Street, Willoughby, OH 44904
11. Energy Research and Generation, Inc., 900 Stanford Avenue, Oakland, CA 94608
12. B.R.F. Kendall, J. Vac. Sci. Technol. 17, 1385 (1980)
13. One company offering this service: Duniway Stockroom, Inc., 1600 N. Shoreline Boulevard,
Mountain View, CA 94043
14. P.H. Schmidt, L.Y.L. Shen, Rev. Sci. Instrum. 44, 1788 (1973)
15. R.S. Timsit, A.K.C. Kiang, Rev. Sci. Instrum. 44, 770 (1973)
Chapter 5
Bakeout
5.1 Bakeout of Metal Ultrahigh Vacuum Systems
The removal of adsorbed impurities from the walls of an ultrahigh, vacuum system
is carried out by heating the walls uniformly while pumping the system. Generally,
for stainless steel systems this is done in the temperature range 150–200 °C. For
preservation of effective ion pumping, bakeout into a turbopump is recommended
to avoid introducing a high gas load into the ion pump. Ion pumps have only a finite
capacity, and also suffer from regurgitation effects in which previously pumped gas
is evolved when another gas is being pumped-as would occur in an experiment in
which gas is admitted to the chamber.
Many methods are used to bake out metal systems. Heating tapes, IR radiation
heaters on the outside, clamp-on band heaters, glue-on rubberized heaters, and
internal light-bulb heaters have all been used. In order to lose as little heat as
possible, it is appropriate to surround the system with some sort of reflective
bakeout shield [1]. Figure 5.1 shows a schematic system in which aluminized
fiberglass [2] is used to shield the main system. The fiberglass tent is supported
internally by temporary metal supports. Radiative heaters are uniformly distributed
over the system inside the reflective shield, and for appendages such as the bellows
on the top, lightweight reflective metal cylinders may be used for radiation capture.
The system should be monitored over many positions with thermocouples to assure
uniform heating, and the wattage output of the various radiative heaters may be
varied to achieve uniformity. Experience in monitoring the pressure during bakeout
will supply information about the bakeout time needed to reach the ultimate low
pressure [3]. It is reported that external IR lamps bring the vacuum system to
bakeout temperatures much faster than other methods [1].
Quite nice tents made, of the aluminized fiberglass may be constructed, and a
convenient method for joining the sections is to use a poprivet device to make
excellent lap seams. The aluminum surface is placed on the inside of the tent. Even
nonaluminized fiber glass curtain material may be used by sandwiching aluminum

DOI 10.1007/978-3-319-17668-0_5
84 5 Bakeout
Fig. 5.1 Schematic bakeout assembly
foil between the layers. A small internal circulating fan inside the oven will lead to
improved temperature uniformity.
Experience indicates that clamp-on band heaters applied to metal flanges can
thermally shock the flanges to a point where a leak develops. In the event that these
heaters are used, a gradual increase in the applied voltage over time is recom-
mended. Internal lamp heaters, working inside the vacuum system, would seem to
be ideal for bakeout because of efficient radiation capture by the system walls, but if
breakage of the glass envelope occurs, the system will have to be completely
cleaned of glass slivers before continued use. Such slivers are very detrimental to
welded bellows and to the sealing surfaces of valves.
Turbopumps should be baked at the end of the bakeout period (for about 4 h).
Titanium sublimation pumps should be outgassed until the pressure reading reaches
a maximum; long outgassing at high temperature will oxidize the components near
the hot Ti. It has been found that cooling the room during measurements will cause
a substantial improvement in the limiting pressure under ultrahigh vacuum condi-
tions owing to the reduction in the outgassing rate from the system walls [4].
5.1 Bakeout of Metal Ultrahigh Vacuum Systems 85
Often it is appropriate to bake, out a new system, with all heat sensitive com-
ponents removed, to temperatures between 300 and 450 °C to remove impurities
left from chemical cleaning procedures.
References
1. B.V. Hess, T.E. Felter, J. Vac. Sci. Tech. A3, 2218 (1985)
2. Fyrepel, P.O. Box 518, Newark, OH 43055
3. R. David, Institut fur Grenzflächenforschung und Vakuum-physik, Forschungszentrum Jülich
GmbH, Postfach 1913, D-5170 Jülich, Germany (private communication)
4. Pittsburgh Surface Science Center
Chapter 6
Behavior of UHV Systems
6.1 Wall Passivation in Stainless Steel Ultrahigh

Vacuum Systems
To achieve the limiting pressure in an ultrahigh vacuum system, it is necessary to

reduce the rate of desorption of molecules from the stainless steel walls to the
lowest possible limit. These surface impurities originate from exposure of the walls
to impurities during opening and processing of the system. They also result from
the diffusion of impurities in the bulk of the stainless steel to the wall surface,
especially for the case of dissolved hydrogen [1].
Bakeout in the presence of traces of oxygen or water may result in further
oxidation of the walls, leading to some passivation. However, a more effective
technique involves the oxidative passivation of the stainless steel walls with a
strong gaseous oxidizing agent. The oxidation of the walls and the oxidizable
surface impurities is kinetically accelerated using a strong oxidizing molecule. In
Fig. 6.1, an energy level diagram (Gibbs free energy) is shown. It may be seen that
the nitric oxide molecule is more energetic than a mixture of O2 and N2 to which its
energy is referenced. The excess energy is 0.9 eV/NO at 300 K. This excess
chemical energy may be sufficient to allow the NO molecule to produce a rapid
surface oxidation process that proceeds only slowly over an activation barrier when
O2 is the oxidizing agent. The oxidation process on the surface is schematically
shown as the formation of a chemisorbed O atom (O(a)), but the formation of CO2
or H2O from surface carbon or from surface hydrogen could also be the end result
of the oxidation process.
These ideas have been used to clean up the walls of an ultrahigh vacuum system
during bakeout, achieving lower base pressure [2]. The bakeout at 200 °C was
carried out in a background pressure of NO of 5 × 10−7 mbar achieved by flowing
NO through the hot system. Initially, during the first phase of the treatment, high
levels of N2, H2O, and CO2 are detected as the NO acts to oxidize the surface and
the oxidizable impurities. After 30–180 min, a continuous increase of the partial

DOI 10.1007/978-3-319-17668-0_6
88 6 Behavior of UHV Systems
Fig. 6.1 Nitric oxide surface reactivity for passivation of vacuum system walls
pressure of NO occurs, and the oxidation of the surface approaches completion.

This treatment is particularly good for reducing the displacement of surface
impurities from the walls during subsequent adsorption experiments. A simple
equilibrium thermodynamics calculation shows that the use of the energetic NO
molecule at 5 × 10−7 mbar is equivalent to the use of O2 gas for the same purpose at
a pressure of about 106 atm!
6.2 Minimizing Wall Reactions in Ultrahigh Vacuum Systems … 89
6.2 Minimizing Wall Reactions in Ultrahigh Vacuum

Systems—Gas Dosers
The purity of gases used in surface science adsorption and surface reaction studies
is of paramount concern just as the production of ultrahigh vacuum is directed
toward the same end. Often gas dosers are employed to deliver pure gases to
surfaces, and for some gases, the displacement of adsorbed impurities, or the
reaction of the dosing gas with the internal surfaces of the gas dosing system itself
can be a problem. In the case of a surface reaction, it is sometimes possible to
passivate the internal surface of a gas doser to reduce the rate of the surface
reaction. Figure 6.2a shows a typical gas doser made mainly of stainless steel. The
construction of this doser and its calibration are discussed elsewhere (p. 421);
however, it may be seen that the conductance is controlled by a laser-drilled
aperture, and collimation of the exiting gas is achieved using a microcapillary array
made of glass. Both the internal metal surface and the internal surface of the
microcapillary array (which is many tens of cm2) provide reactive surface area that
can evolve impurities by displacement or reaction [3], Displacement processes,
such as the exchange of surface OH groups with OD groups in deuterated alcohols,
can often best be handled by overnight flushing of the doser with the isotopic gas.
In extreme cases of reaction with acid sites on the internal surfaces of the capillaries
[3], one must remove the glass microcapillary array from the doser [3, 4]. Reaction
processes with the stainless steel walls of the doser system can sometimes be
reduced in magnitude by passivation of the internal surfaces. This is vividly seen in
the behavior of the metal carbonyls, where decomposition of Fe(CO)5 on the inner
stainless steel walls has been shown to produce CO(g), leaving iron on the wall [4].
The surface reaction was studied, and it was found that predosing the doser and its
stainless steel delivery system with high fluxes of molecular oxygen resulted in
significant reduction of the decomposition reaction, as judged by mass spectro-
metric studies of the gas exiting the doser, line-of-sight, into the mass spectrometer
[4]. The internal walls of the ultrahigh vacuum system were found to have little
effect on the metal carbonyl, probably because it was exposed to oxygen each day
[4]. For the experiments, it was found that the glass microcapillary usually
employed with this design of gas doser had to be removed because of deleterious
effects on Fe(CO)5.
Another solution to this problem is to make the doser system from materials
other than stainless steel [5]. Figure 6.2b shows such a system, in which the glass
tube doser contains a capillary for the control of the flow rate from the connecting
tube into the ultrahigh vacuum system through the Pyrex doser tube. The
high-pressure (several Torr) section beyond the Teflon-glass valve is made of PFA
(polyfluoroalkoxy) plastic and fittings are readily available commercially [6]. The
paper describing this system indicates that the capillary tube was produced by hand
and was drawn down to a fine diameter. This particular design was used for the
delivery of hydrazine (N2H4). Thermometer tubing, made generally of lead glass,
has a *100–150-μm diameter capillary for sensitive thermometers (1 K/3.8 cm
90 6 Behavior of UHV Systems
Fig. 6.2 Gas doser designs. a Stainless doser+microcapillary array collimator. b Pyrex—plastic
doser. c Cleaning clogged glass capillaries
capillary length) and may be used if a lead glass to Pyrex glass graded seal is
employed. A Teflon doser for ultrahigh vacuum use also has been described [7], as
well as a heatable glass and Teflon doser for admission of organometallics of low
vapor pressure [8].
It is unfortunate that glass capillaries often clog up with traces of dust or grease
that are present within the vacuum system. A clever method to restore the capillary
to a clean, unblocked condition is shown in Fig. 6.2c [9]. Here, the capillary section
is surrounded on both ends of the capillary with pointed tungsten electrodes, easily
installed by the glassblower. When the capillary becomes blocked, micron pres-
sures are established on the high-pressure side of the capillary, and the vacuum-side
6.2 Minimizing Wall Reactions in Ultrahigh Vacuum Systems … 91
electrode is electrically grounded. The application of a Tesla coil to the

high-pressure tungsten electrode produces an internal plasma discharge, and
the plasma usually cleans out the obstruction in the capillary within seconds. We
have found that this harsh internal treatment changes the conductance of the glass
capillary slightly, and recalibration after cleaning is suggested.
A third solution that has been used to minimize reactions inside a gas doser is to
dilute the reactant gas with an inert diluent such as He or Ar [10]. This may be done
in a gas handling line that is dynamically pumped, causing the gas mixture to be
replenished rapidly, and removing products of reaction [10]. This method also
permits small doses of the reactive gas to be quantitatively monitored since pressure
measurements of the mixture may be converted to partial pressure measurements of
the active gas. One must remember that the composition of the gas mixture will
change under dynamic pumping conditions owing to different rates of pumping for
the gas components [10].
Reports of the use of a Teflon coating for passivation of the stainless steel walls
of an ultrahigh vacuum system are available [11, 12]. Briefly, a weak acid priming
treatment was used to remove oxide films to promote adhesion of the Teflon.
A Teflon suspension [13] was painted on the walls and cured in the air for several
days. The coating was then baked in vacuum at 200 °C. Small parts were cured
using a heat gun before assembly. This coating was effective in preventing TiCl4
decomposition on the walls. However, the Teflon adsorbs water from the air when
the system is opened, and longer bakeout is required to get to a base pressure in the
mid 10−9 Torr range than would be required with an uncoated vacuum chamber
[12].
References
1. P.A. Redhead, J.R. Hobson, E.V. Kornelsen, The Physical Basis of Ultrahigh Vacuum (AIP,
New York, 1993)
2. M. Grunze, G. Strasser, O. Eishazly, J. Vac. Sci. Technol. A4, 2396 (1986)
3. P.K. Leavitt, P.A. Thiel, J. Vac. Sci. Technol. A8, 148 (1990)
4. M.A. Henderson, R.D. Ramsier, J.T. Yates Jr, J. Vac. Sci. Technol. A9, 2785 (1991)
5. E. Apen, R. Wentz, F. Pompei, J.L. Gland, J. Vac. Sci. Technol. A12, 2946 (1994)
6. Swagelok fittings are available from Swagelok, 31400 Aurora Road, Solon, OH 44139
7. M.A. Mendicino, E.G. Seebauer, J. Vac. Sci. Technol. A10, 3590 (1992)
8. L.A. Jones, F.-X. Wei, D.F. Thomas, Rev. Sci. Instrum. 66, 4981 (1995)
9. D.A. King, (1970) (private communication)
10. E.C. Henn, M.E. Bussell, C.T. Campbell, J. Vac. Sci. Technol. A9, 10 (1991)
11. W. Muller-Markgraf, M.J. Rossi, Rev. Sci. Instrum. 61, 1217 (1990)
12. Professor E.G. Seebauer, Department of Chemical Engineering, University of Illinois, Urbana,
IL 61801. From thesis by M.A. Mendicino, Ph.D (1994) (private communication)
13. Teflon or PTFE suspension is obtainable from: Intek Services, Inc., 7 McMillan Way, Suite
101, Newark, DE 19713
Chapter 7
Mechanical Action on Samples
7.1 Simple UHV Crystal Cleaver
Often one of the best ways to produce an atomically clean surface is to expose an
inner cleavage plane by cleaving the crystal in ultrahigh vacuum. A simple
mechanical cleaver design is shown in Fig. 7.1a [1]. Here, a MgO single crystal is
mounted on a coolable manipulator inside a small rectangular compartment, held by
a screw that presses on an internal plate that distributes the screw pressure over the
side of the crystal. A portion of the crystal projects upward, ready for the cleaver
blade.
The cleaver is a spring-loaded tungsten carbide blade (25° cutter angle) that
slides in a stainless steel channel. Two springs are compressed by turning a pair of
screws prior to attaching the cleaver to the manipulator. The manipulator/cleaver
assembly is loaded into the vacuum system, and after system processing, the trigger
holding the blade back is released by moving the manipulator against the side of the
manipulator port. The blade cleaves the crystal, and caution should be exercised not
to allow the cleaved-off portion of the crystal to enter fragile apparatus as it bounces
around the chamber.
Using a closed cycle He cryocooler, this assembly can reach 44 K. Studies on
the freshly cleaved MgO(100) surface, involving Kr adsorption, are shown in
Fig. 7.1b, and the results are compared to MgO that has been cleaved outside the
vacuum system, transferred, and thermally processed. The sharp vertical step
indicating a 2D gas → 2D solid transformation, as studied by LEED, is charac-
teristic of the UHV-cleaved crystal; a less distinct transition is seen on the exter-
nally cleaved crystal because of defects produced by water adsorption that cannot
be removed by thermal processing. Indeed, after days of storage in UHV, the
UHV-cleaved crystal begins to yield an isotherm resembling that of the defective
crystal because of slow impurity adsorption.

DOI 10.1007/978-3-319-17668-0_7
94 7 Mechanical Action on Samples
Fig. 7.1 Simple UHV cleaver. a Cleaver design. b Comparison of Kr isotherms on MgO(100)
using LEED
7.1 Simple UHV Crystal Cleaver 95
The authors report that the only disadvantage of this design is that only one
cleave can be achieved per loading into the vacuum system because of the necessity
to compress the springs outside the vacuum system.
A cleavage device for mica has been described that can be used in ultrahigh
vacuum [2].
7.2 Piezoelectric Fatigue Apparatus for UHV Operation
There is much interest in the role of adsorption on the rate of fatigue of metals
undergoing periodic stress conditions. Usually these measurements are performed
under cycle rates of about 1 Hz. Hydraulic cycling methods often impart noise to
the stress-strain measurement system. A piezoelectric drive mechanism, capable of
working at 100 Hz frequency in UHV, is described below [3].
Figure 7.2a shows a schematic drawing of the apparatus. The specimen, coupled
to an extensomer, is clamped between two grips that have serrated clamping sur-
faces to prevent misalignment during testing. A load cell is placed above the
moving grip and below the piezoelectric actuator. The translator used has a capacity
of 10 kN in compression, and 3.5 kN in tension [4]. It is capable of translation
through about 40 μm over an applied voltage range of 0–1500 V. The load is
measured with a load cell having a range of 0–2000 N [5]. A strain of about 0.3 %
may be measured using an extensometer [6] in a gauge section of 5 mm, and the
specimen tested in this case was polycrystalline Ag. The control electronics are
described in [7]. Figure 7.2b shows a hysteresis curve measured for Ag in which the
stress and strain are measured during compression and expansion. The hysteresis
loop changes as the number of cycles increases, and is a measure of the specimen
fatigue. The data obtained with the piezodriver exhibits a better signal-to-noise ratio
than data taken with hydraulic devices.
The main disadvantage of the device is the small translation range of the driver.
The device must account for all of the translations taken up by the components
along the axis of the load. This limits the machine to small specimens, and for
high-strength materials, smaller gauge lengths and more narrow cross sections
would be needed to obtain sufficient strains for accurate measurement [3].
Another design, providing fatigue to a thin metal specimen in ultrahigh vacuum,
has been described [8]. It was found that clean Al(111) single crystals exhibited
changes in their ultraviolet photoelectron spectrum after 1000 fatigue cycles at
0.2 % strain. The involvement of chemisorbed species containing oxygen was
eliminated on the basis of the absence of O(2p) emission in the photoelectron
spectrum, and no explanation was given for this effect [8].
96 7 Mechanical Action on Samples
Fig. 7.2 Piezoelectric fatigue apparatus—UHV. a Schematic design. b Typical stress-strain curve
—polycrystalline Ag
References 97
References
1. T. Angot, J. Suzanne, J.Y. Hoarau, Rev. Sci. Instrum. 62, 1865 (1991)
2. K. Reichelt, Rev. Sci. Instrum. 44, 243 (1973)
3. T.S. Sriram, M.E. Fine, Y.-W. Chung, Rev. Sci. Instrum. 62, 2008 (1991)
4. Physik Instrumente, Model 243.37, Option 703.20, sold by PolyTec Optronics, 3001 Redhill
Ave., Costa Mesa, CA 92626
5. Model 31, SENSOTE Inc., 1202 Chesapeake Ave., Columbus, OH 43212
6. Model 632.29C, Option 011, MTS Systems Corporation, Box 24012, Minneapolis, MN 55424
7. W.J. Lee, Y.-W. Chung, M.E. Fine, J.P. Baker, Rev. Sci. Instrum. 57, 2854 (1986)
8. T.K.G. Swami, Y.W. Chung, Rev. Sci. Instrum. 51, 842 (1980)
Chapter 8
Gasket Seals
8.1 Gasket Seals for Ultrahigh Vacuum Systems
The primary ultrahigh vacuum flange design used universally is the stainless steel
Conflat sexless flange, which operates with two opposed knife edges that seal a
metallic gasket by deformation when the flange is tightened. A cross-sectional view
of the flange and a confined gasket is shown in Fig. 8.1.
In many cases where it may be difficult to seal the gasket by applying the
necessary torque to the flange bolts, or where tightening down to the maximum
force level is not desired, it is useful to use annealed copper gaskets. These maybe
purchased from the suppliers, but they can be made quite simply in the laboratory
beginning with a nonannealed gasket. The gasket is heated while rotating it around
its circumference in an oxygen-gas torch (cool flame) until dull red color is emitted
by the hot copper. The gasket is then quenched in water quickly, resulting in
softening to almost the softness of lead or tin. This gasket will deform in the seal
region with less torque on the bolts and also with less deformation of the flange
during tightening. Often, gaskets of this type are needed for spectroscopic windows
mounted in stainless steel flanges in order to prevent window cracking during
tightening owing to flange flexure.
The annealing process results in considerable oxidation of the copper gasket, and
this oxide layer must be removed before use. Two methods are available to do this.
In the first, the hot gasket, immediately after annealing, is dropped into methanol
instead of water. The methanol, a reducing agent, immediately strips off the oxide
from the hot metal, leaving a clean copper surface. Care should be exercised to
avoid damage if the methanol should ignite. The second method involves dropping
the cool gasket into an aqueous solution of citric acid. Here the citric acid forms a
soluble complex with the copper ions in the oxide film, and again clean copper
surfaces are produced.
Certain coatings are employed to prevent the oxidation of copper gaskets fol-
lowing relatively high temperature bakeouts. Heavy amounts of copper oxide can

DOI 10.1007/978-3-319-17668-0_8
100 8 Gasket Seals
Fig. 8.1 Gasket designs. a Conflat flange + Cu gasket. b Gold wire—shear seal
fall into the vacuum system when the flange is opened, and also seizure of the
gasket in its recess in the stainless steel flanges can occur. Silver plating has been
used to prevent this effect [1]. A minimum 5-μm-thick silver film, made by plating,
is required. The silver will stand repeated bakeouts at 300–350 °C. Contamination
of the silver film by sulfur from either the gasket or from sulfur-containing lubri-
cants is to be avoided as this causes discoloration and peeling of the silver film.
Thus, bolt lubricants such as molybdenum disulfide cannot be used and graphite or
other lubricants must be employed.
Another useful copper gasket coating is titanium nitride, TiN [2]. Using reactive
magnetron sputtering from a Ti target in an N2 + Ar mixture (20 % N2), films from
50 to 150 Å were deposited on copper gaskets. In tests, ten flanges were baked at
550 °C in a vacuum of 10−5 Torr for up to 200 h, and were then tested for leak
8.1 Gasket Seals for Ultrahigh Vacuum Systems 101
tightness with a He leak detector, finding no leaks. The flanges could then be
disassembled without seizure. Without the TiN film, separation must be done with
jacking screws and copper is left on the sealing surfaces, requiring abrasives for
removal. A similar coating has also been prepared using diamondlike carbon
coatings prepared by RF discharge techniques in methane gas [3].
One of the early methods of sealing stainless steel flanges employed gold wire
gaskets [4]. For flat flanges without a knife edge or for a stepped seal used for
crushing a metal gasket, gold wire will make an excellent seal. The wire, of about
0.25–0.5 mm diameter, is made into a circular loop on a forming mandrel; then a
microtorch is used to butt weld the ends together when the loop is removed from the
mandrel. This process is best accomplished by gently clamping together the squared
off ends of the gold wire and slowly melting the gold so that surface tension causes
liquid gold flow to even out the diameter of the heated region. Too much heating
will result in separation of the gold wires with the formation of balls of gold on the
ends of each. Practice is, therefore, necessary. The wire once formed is placed
between two smooth flange faces, and the bolts are tightened, making an ultrahigh
vacuum seal. A stepped seal may also be used with either gold or copper, and here
the seal produces a shear seal within the gasket as shown in Fig. 8.1. In the case of
gold wire shear flanges, the wire must be butt-welded to fairly exact dimensions
using a mandrel of the same size as the male flange for accurately cutting the gold
wire to appropriate length. A nice description of the method to make custom copper
wire gaskets is given in [5].
The reuse of large Cu gaskets for many cycles has been described [6]. Here, a
“feeler gauge” such as that used to measure spark plug gaps is employed, and
successive seals with the original gasket are made by diminishing the
flange-to-flange clearance in a systematic manner.
8.2 Cryogenic Gasket Seals
Stainless steel Conflat flanges are guaranteed to be vacuum tight from 77 K up to

their normal bakeout temperature limits. However, there are special situations
where vacuum components made of different metals are to be joined by flanges, and
in these cases cooling to 77 K may disrupt the vacuum seal because of differences
in thermal expansion coefficients.
This problem has been solved for 18-mm-diameter junctions between stainless
steel flanges and aluminum Conflat 34 flanges using a gold wire seal (Fig. 8.2). In
this design, the mismatch in expansion/contraction occurs in a radial direction along
the sealing plane, and this stress can be absorbed by the plastic deformation of the
gold wire. The seal is maintained through many cycles as a result of the gold
ductility and its good adherence against stainless steel and aluminum surfaces [7].
A cross sectional drawing of the seal is shown in Fig. 8.2. The gold wire ring
was 1 mm in diameter, prepared by butt-welding. The screws were 4 mm diameter,
made of stainless steel which prevents the need for retorquing after a thermal cycle.
102 8 Gasket Seals
Fig. 8.2 Gold wire seal—cryogenic design
The specially designed stainless steel flange has a precision entrance guide to
prevent damage to the aluminum knife edge and it is also machined to be very
smooth on its sealing surface. In the particular case shown here, the stainless steel,
exhibiting poor heat conduction, was needed to reduce the thermal conduction of
heat to the aluminum section therefore preserving liquid nitrogen which was housed
in the aluminum section.
8.3 Copper Gasket Removal Devices
There are three published methods that are directed toward easy removal of copper
gaskets from Conflat flanges after disassembly [8–10]. These methods are all
designed to prevent damage to the sharp sealing knife edges by tools being used to
pry up the gaskets. The first of these [8], shown in Fig. 8.3a, employs Conflat
flanges that have two opposed slots machined on the inside of the flange.
A screwdriver blade can be inserted into the slot and rotated to pop out the gasket
and the boundary of the slot prevents the blade from coming into contact with the
knife edge. The disadvantage of this design is that one must machine every flange
with the two slots [9]. A second design [9] (Fig. 8.3b), uses the leak-check groove
as a slot to control the motion of a specially wedge-shaped and hardened steel
blade, shown. The third design (Fig. 8.3c), shows a modified “parrot-nose” set of
8.3 Copper Gasket Removal Devices 103
Fig. 8.3 Tools for copper gasket removal. a Inside slot design. b Leak-check groove tool.
c Custom pliers
ignition pliers that can safely grip the gasket [10]. This design allows only upward
force on the gasket and also grips the gasket during removal to prevent it from
falling. In certain instances where there is insufficient clearance for the pliers on the
outside of the copper gasket, one must resort to a tool of the second design [9, 10].
8.4 Unconventional Compression Seals for Ultrahigh

Vacuum
Several unconventional compression seals have been designed for ultrahigh vac-
uum systems where flange mounting was inappropriate. These seals all have in
common the use of a material softer than stainless steel that is deformed under
pressure supplied by screwing down the re-entrant seal from the outside of the
vacuum system. This forms the vacuum seal, which is always on the inside of the
104 8 Gasket Seals
vacuum system. Figure 8.4a is a seal involving oxygen-free high conductivity

(OFHC) copper that is deformed, at the edge of a hole drilled into the vacuum
system wall. A stainless steel bolt, with a hole drilled axially through its length, is
machined to present a hemispherical surface under the head. Copper tubing is spun
down on a lathe to closely fit over the bolt. The bolt is inserted from the vacuum
side of the hole and tightened with the nut from the outside. Such seals can be
cycled between 873 and 77 K many times, remaining He-leak tight. The seals may
be loosened and used repeatedly for five to ten times. Ultimately, the copper
becomes thin and hardened as a result of the mechanical deformation. New copper
fittings may be used to replace the original used fitting [11].
A second re-entrant bakeable all-metal vacuum seal is shown in Fig. 8.4b. Here a
conventional Conflat flange design is used for the sealing surface and the force to
Fig. 8.4 Unconventional compression seals. a Hollow bolt seal. b Mini-Conflat seal. c Gauge
tube seal
8.4 Unconventional Compression Seals for Ultrahigh Vacuum 105
produce the seal at the copper gasket is applied externally by means of jacking
screws. The assembled seal was designed to apply 2000 lb/in. along the copper
gasket seal path. The seal is assembled from inside and initially tightened using the
threaded jack ring, which pushes down on a chrome plated thrust washer needed to
prevent distortion of the main sealing flange that contains the hole and also to
prevent galling of the jackbolts under torque. First torquing of ten bolts to 20 in. lb
resulted in a leak tight seal. Further torquing to 40 in. lb reached the plastic-flow
limited torque for the No. 8 stainless bolts. Disassembly and reassembly at 40 in. lb
torque produced a He-leak-tight seal, which withstood baking to 250 °C [12].
A third design, solving a major vacuum problem for sealing thermocouple
gauges into ultrahigh vacuum systems, is shown in Fig. 8.4c. For most applications
under non-stringent vacuum conditions, thermocouple gauges are sealed at the pipe
thread joint by use of Teflon tape, vacuum cements, etc. This is unacceptable for
ultrahigh vacuum operation owing to the possibility of gas trapping in the pipe
thread or cement. The design shown in Fig. 8.1c uses the pipe thread to compress
either a Teflon or aluminum or OFHC copper gasket at the bottom of the tubulation.
The end of the stainless steel gauge stem must be smoothed with 600-grit paper,
using a removable plug inside the tubulation to prevent entry of debris. Similarly,
the sealing surface in the receptor hole must also be polished. This method permits
thermocouple gauges to be used on ultrahigh vacuum systems and to withstand
normal bakeout temperatures [13].
References
1. T. Wikberg, J. Vac. Sci. Tech. 5, 205 (1968)

2. E.L. Garwin, E.W. Hoyt, A.R. Nyaiesh, J. Vac. Sci. Tech. A4, 2537 (1986)
3. A.R. Nyaiesh, E.L. Garwin, E.W. Hoyt, J. Vac. Sci. Tech. A5, 2977 (1987)
4. A. Roth, Vacuum Sealing Techniques (Pergamon, Oxford, 1966), p. 380
5. E.M. Beaver, Rev. Sci. Instrum. 51, 254 (1980)
6. L.Y.L. Shen, J.E. Rowe, E.E. Chaban, Rev. Sci. Instrum. 44, 1747 (1973)
7. F. Shimoshikiryo, Y. Takakuwa, N. Oyama, N. Miyamoto, J. Vac. Sci. Tech. A11, 2874
(1993)
8. J.J. Donelon, J.M. Baker, J. Vac. Sci. Tech. 10, 894 (1973)
9. E.E. Chaban, J. Vac. Sci. Tech. 12, 654 (1975)
10. B.J. Waclawski, J. Vac. Sci. Tech. Al, 99 (1983)
11. N. Raeuber, K. Stadler, M.J.O. Strutt, J. Vac. Sci. Tech. 8, 662 (1971)
12. P.H. Edmonds, W.D. Shipley, J. Vac. Sci. Tech. A6, 165 (1988)
13. F.N. Rabarchik, J. Vac. Sci. Tech. 16, 2116 (1979)
Chapter 9
Leak Repairs and Detection
9.1 Coaxial Pumped He Leak Detection Probe
He-leak checking is commonly used, either in the shop with a portable leak
detector, or in the laboratory where a mass spectrometer inside the vacuum system
is used for detection. A common problem involves the detection of a small leak at
some point, where subsequent investigation will show that there is really no leak at
that point. This confusing result is caused by the diffusion through the atmosphere
of He from the suspect point to the real leak, which is some distance away. The
device shown in Fig. 9.1 eliminates this effect by sweeping He away from the test
point using a coaxial vacuum sleeve.
Figure 9.1 shows the device, and the reader is referred to the original publication
for construction and assembly instructions [1]. The He flow rate is adjusted to one
or two bubbles of He per second as determined by placing the probe under water
with the valve-controlled vacuum pump turned off. The rate of pumping should be
adjusted so that the sensitivity for He is not diminished by too rapid pumping.
A pump with a displacement of 5–10 L/min is satisfactory, with a throttle valve for
pumping speed adjustment.
Once a leak is noted, a second test should be performed to verify its general
location. This consists of placing a plastic bag around the suspected part, sealing the
bag with tape, and then filling the bag with He, and observing the leak again. Using
this coaxial probe and then using the bag verification test will eliminate carrying
non-leaking components to the shop for repair.
In the event that leak sealing compounds are used for temporary repairs, only
very small quantities of the sealer should be applied at the leak location, and a tag
should be placed on the affected component telling the location of the leak and the
date of temporary repair. The leak should be repaired correctly as soon as possible
after the temporary repair. The tag is a guide for the memory when the vacuum
system is ready to be disassembled.

DOI 10.1007/978-3-319-17668-0_9
108 9 Leak Repairs and Detection
Fig. 9.1 Coaxial pumped he leak detection probe
9.2 Temporary Leak Sealing of Welded Bellows
Welded bellows often develop small leaks at times when it is inconvenient to

interrupt the work going on in the vacuum system. These leaks are most likely
located on the inner welded seal of the bellows and are difficult to locate precisely.
For small leaks a temporary seal may often be established by using a leak sealer
diluted with a solvent to reduce its viscosity and then spraying tiny amounts of the
liquid onto the leak region. However, continued flexure of the seal after drying can
cause the leak to reappear. Therefore, a clever trick [2] was devised in order to
extend the lifetime of the bellows while a new one was being ordered. This involves
immobilizing the small section affected by the leak. This is done by squeezing
gently several 0.5-mm-thick stainless steel clips around the bellows region of
interest immediately after spraying the diluted sealant on the bellows, as shown in
Fig. 9.2. Since one leak in a bellows is generally a signal of more to come, this
procedure should not be employed as a long-term solution to the leaking bellows
problem.
A variety of liquid leak sealers are commercially available, and the reader is
referred to the catalogs or to the original reference for a description of a silicone
sealer [2].
9.3 Atmospheric Permeation—Viton O-Ring
Viton [3] seals are used in ultrahigh vacuum applications where deformable met-
alseals cannot be used. Viton seals may be baked out at 200 °C and after a 24 h
bake-out an outgassing rate of the order of 2 × 10−9 Torr L/s cm2 (nitrogen
equivalent) has been reported [4]. After bakeout, the main gas evolved was water.
9.3 Atmospheric Permeation—Viton O-Ring 109
Fig. 9.2 Temporary welded bellows repair
Bakeout eliminates hydrocarbons initially detected upon installation of a fresh

Viton seal [4].
A significant property of Viton seals is their ability to allow water from the
atmosphere to permeate slowly through the outgassed Viton into the ultrahigh
vacuum system, causing the partial pressure of water to rise for many days after
bakeout. Figure 9.3a shows an apparatus where atmospheric gas permeation
through a Viton O-ring was investigated [5]. A stainless steel vacuum system,
equipped with a mass spectrometer and trapped diffusion pumping facilities, con-
tained a test Viton O-ring [4]. The top of the vacuum chamber was baked at about
100 °C, while it was estimated that the section containing the O-ring was baked at
50 °C. A base pressure of about 5 × 10−9 Torr was achieved after 3 days’ baking.
Following 2 days’ cooling, the mass spectrum of the residual gases in the chamber
was measured for an additional 30 days, as shown in the test schedule (Fig. 9.3b).
It was found that the major effect was a continuous rise in the partial pressure of
water during the 30-day period, with a decreasing rate after about 20 days.
Figure 9.3c, d schematically show these tests for conditions of low atmospheric
humidity and high atmospheric humidity. The rate of water permeation increases
for the humid conditions. There is no detectable amount of permeation of other
atmospheric gases such as N2, Ar, or CO2 during the 30-day test, although some O2
permeation was reported [5].
110 9 Leak Repairs and Detection
Fig. 9.3 Atmospheric permeation—Viton O-Ring. a Apparatus for O-Ring studies. b O-Ring test
schedule. c Low humidity—permeation. d High humidity—permeation
These results suggest that the best application for Viton seals in ultrahigh vac-
uum systems is in cases where exposure of the seal to the atmosphere does not
occur, i.e., where the seal is used internally for isolation of one part of the vacuum
system from another. This is the case with Viton sealed gate valves mounted
between sections of the vacuum system, where the seal is retained in vacuum during
the time when the valve is opened.
A number of studies of outgassing and permeation of vacuum materials have
been carried out [6–8].
References
1. G.L. Fowler, J. Vac. Sci. Technol. A5, 390 (1987)

2. W.F. Egelhoff Jr, J. Vac. Sci. Tech. A6, 2584 (1988)
References 111
3. Viton A is a copolymer between vinylidine fluoride and hexafluoropropylene with the formula
[-FC(CF3)-CF2-CH2-CF2-]n. Initial outgassing evolves water, oxygen, hydrogen, nitrogen, and
some hydrocarbons (most prominent mass peak at 43 amu, probably from C2F+). After bakeout,
the main gas observed was water [2]
4. R.S. Barton, R.R. Govier, J. Vac. Sci. Technol. 2, 113 (1965)
5. N. Yoshimura, J. Vac. Sci. Technol. A7, 110 (1989)
6. W.G. Perkins, J. Vac. Sci. Technol. 10, 543 (1973)
7. B.B. Dayton, Transaction of the American Vacuum Society 6th Vacuum Symposium, 101 (1959)
8. R.A. Redhead, J.R Hobson, E.V. Kornelson, The Physical Basis of Ultrahigh Vacuum
(American Institute of Physics, New York, NY, 1993), chapter 10
Chapter 10
Specialized UHV Systems
10.1 Aluminum Ultrahigh Vacuum System
There are advantages in using aluminum UHV systems. For proton synchrotrons,
the residual radioactivity produced in stainless systems may be reduced by one or
two orders of magnitude, 100 h after machine shutdown, by employing Al instead
of stainless steel. For electron synchrotron machines and electron storage rings,
aluminum is preferred because of good thermal conductivity, ease of manufacture
by extrusion, low residual radioactivity, nonmagnetic character, and low cost. The
design here is an all aluminum system made of two types of aluminum alloy;
aluminum flanges and gaskets are specially designed [1, 2]1.
A special aluminum alloy, 2219-T87, is found to have the necessary mechanical
properties for making the sealing flanges and nut/bolt sets. As shown in Fig. 10.1,
the sealing surface consists of a circular slot on each flange that houses a Helicoflex
[3] O-ring containing an elastic spring core inside a pure aluminum sheath. The
fixed depth of the O-ring groove was adjusted to correspond to the specified
compression radius of the Helicoflex element—a 4.5-mm diameter Helicoflex
required two 1.5-mm deep grooves.
The chamber body was made of a 6063-T alloy, which could be welded by an
AC TIG welding process. Distortion after welding was negligible and remachining
was not necessary. The entire chamber, except the O-ring sealing surfaces, was
anodized to a 30-μm thickness. Bellows made of 6063-T6 alloy could be made by a
rolling process yielding 0.5 mm thick walls. The bellows are employed for align-
ment purposes, not for dynamic motion. They also compensate for thermal
expansion effects.
The O-rings were sealed by torquing the nut/bolt combinations, bringing the
aluminum flange surfaces together. The torque (per unit area of the O-ring) used for
tightening was between 140 and 215 kg/cm, with 200 kg/cm being normal.
1
Which contains references to various papers on this subject.

DOI 10.1007/978-3-319-17668-0_10
114 10 Specialized UHV Systems
Fig. 10.1 Aluminum UHV chamber, flange, and bellows
Following bakeout at 200 °C, the leak rate at the Helicoflex seals was measured
with a He leak detector. It is reported that the leak rate is less than 10−11 atm cm/s,
where the units on this number suggest that a leak rate per cm2 of sealing surface
was calculated. The ultimate pressure in the system was 3 × 10−10 Torr after
bakeout and the residual gases consisted mainly of H2, CO, and CO2. The 38-L
system was pumped by a 220-L/s turbopump and a 160-L/s sputter ion pump. It is
indicated that the practical upper bakeout temperature is limited to 170 °C by the
6063-T6 aluminum alloy.
A study of the outgassing rate of an aluminum vacuum system during bakeout is
given in [4].
10.2 Surface Electrochemistry Apparatus
The development of the field of electrochemistry from the viewpoint of surface

science has depended upon the experimental ability to prepare atomically clean
single crystal surfaces and then to transfer these surfaces to an electrochemical
10.2 Surface Electrochemistry Apparatus 115
Fig. 10.2 Schematic of surface electrochemistry apparatus
environment for measurements of current versus voltage [5]. While a number of

designs have been provided for this transfer operation, the design schematically
shown in Fig. 10.2 is historically important and remains useful today.
A single crystal is mounted on a sample manipulator in the left-hand UHV
system, which contains various facilities for surface cleaning and surface charac-
terization. The sample, once cleaned and characterized, is unplugged in vacuum
from the sample manipulator (see pp. 89 and 93) and is captured on a bellows type
translator coming in from the right through an open gate valve. From here the
sample may be immersed into the electrolyte of the electrochemical cell by means
of moving the translator vertically and lowering the sample through the open gate
valve. Four types of pumps are utilized on the right-hand side of the apparatus: two
sorption pumps using zeolite pellets at 77 K; a cryogenic pump containing a
charcoal-covered surface cooled to 14 K by a Joule-Thomson refrigerator, which
lowers the pressure to about 10−7 Torr; ionization pumps to bring the pressure down
to about 10−9 Torr; and a titanium sublimation pump for final pumping and for
purification of Ar gas which is admitted to high pressure for the sample transfer to
the electrochemical cell.
The solution-handling subsystem must exclude dissolved contaminants. Liquids
are contained in deaerated (N2 purged) Pyrex bottles. Air is the most difficult
impurity to exclude, and Teflon tubing sheathed in Teflon bellows is used under
inert gas purging conditions inside the annular space to exclude air diffusion
through the Teflon to the electrolyte. Glass vacuum valves rather than stopcocks are
used to avoid contact with the air. Teflon compression fittings are used to make
tubing interconnections. All parts are soaked in chromic acid cleaning solution until
needed, then rinsed with highly purified distilled water.
116 10 Specialized UHV Systems
Potential-dependent interfacial properties of the electrode-sample are measured

using a Pt electrode introduced through a 3-mm-diameter section of Teflon tubing.
A silver wire, coated with silver chloride, and immersed in a chloride solution is a
convenient reference electrode. By applying a potential between the Pt electrode
and the sample, and monitoring this potential against the silver chloride/silver
electrode, it is possible to plot out the current-voltage curves and to correlate with
the nature of the electrode introduced.
A comprehensive listing of other types of electrochemical cells which are
coupled to UHV systems is given in [6, 7], which also describe X-ray photoelectron
spectroscopy (XPS) studies of the corrosive oxidation of a Ni surface.
References
1. H. Ishimaru, J. Vac. Sci. Technol. 15,1853 (1978)

2. H. Ishimaru, J. Vac. Sci. Technol. A7, 2439 (1989)
3. Helicoflex is a product of Helicoflex Co., P.O. Box 9889, Columbia, SC 29290
4. M. Suemitsu, T. Kaneko, N. Miyamoto, J. Vac. Sci. Technol. A5, 37 (1987)
5. A.T. Hubbard, in Comprehensive Chemical Kinetics, edited by C.H. Bamford, D.H.R Tipper,
and R.G. Compton (Elsevier, Amsterdam, 1988), pp. 1–67
6. Y. Liang, D.K. Paul, Y. Xie, P.M.A. Sherwood, Anal. Chem. 65, 2276 (1993)
7. P.M.A. Sherwood, Chem. Soc. Rev. 14, 1 (1985)
Part II
Mechanical Fabrication Techniques
Chapter 11
Grids
11.1 Hemispherical Grids—Formation and Piercing
Formed grids are employed in ion and electron optical devices. Often these grids must
be made of highly transparent woven stock, either of stainless steel or tungsten [1].
Woven grid material may be formed into a hemispherical shape using a mold or
mandrel, as shown in Fig. 11.1a. Here the grid material is stretched over the mandrel
and an annealed aluminum ring is slipped over the stretched grid, capturing it near
the base of the mandrel. A steel pressure ring is then seated on the aluminum ring and
pressure is applied causing the soft aluminum ring to distort into an L-shaped cross
section as shown. This also encases the grid material, fixing it in the aluminum [2].
Another method for attaching a stretched grid to a support ring is shown in
Fig. 11.1b. Here a mesh grid is stretched over a brass mandrel, and a stainless steel
plate is clamped in place to hold the grid on the mandrel. The plate is tack-welded
to the grid in about ten places and the clamps are released and the plate turned over.
The grid wires are trimmed and individually spot-welded [3]. A more complex
variation on this method is shown in Fig. 11.1c where the grid is captured between
two shaped rings, R, which are bolted inside a jig. Spot-welding takes place through
the holes, H, after the grid is stretched over the mandrel [4].
Grids must often be pierced in specific locations for the passage of hardware or
source beams through the grid. Figure 11.1d shows a nice way to do this job
accurately. The brass mandrel is punched with the location of the center of the hole.
A scribe is then used to scratch a circular shape on the mandrel locating the cutting
line on the grid. The grid is then stretched and spot welded, and the assembly is
immersed into aqueous NaCl solution (50 g/L). An electrode to a spot welder (S1) is
attached to the assembly, and a second electrode, S2, made of tungsten wire
(0.5 mm) is sharpened to a radius about twice the diameter of the mesh wires to be
cut. An electrical arc from the spot welder will cut the wires individually, usually
between adjacent wires perpendicularly crossing the wire to be cut. The salt water
prevents pitting on the brass mandrel by the electric arc [3].

DOI 10.1007/978-3-319-17668-0_11
120 11 Grids
Fig. 11.1 Hemispherical grids—formation and Piercing. a Grid capture—Al Ring. b Grid capture
—spot welding. c Grid capture—spot welding. d Piercing grid
This stretching method will not make perfect hemispherical grids, since some
relaxation will occur upon removal from the mandrel. If more rigid grid material is
needed, the stretched grid may be electroplated on the mandrel [5] or it may be fired
in an oven. The electroplating stiffens the contact points between the meshes while
firing will stiffen the wire itself. In the case of firing, a brass mandrel cannot be
used, and release agents on the mandrel will be needed. Either Aquadag (colloidal
graphite) or MoS2 will work as release agents.
11.2 Making Flat Mesh Grids of Large Size
Electroformed mesh [6], having up to 85 % transparency, may be conveniently used

as grid material in electron and ion optical applications. However, it is difficult to
construct flat grids of many centimeter dimensions, and irregularities in flatness will
influence the electrical field uniformity desired in front or between parallel grids.
11.2 Making Flat Mesh Grids of Large Size 121
Fig. 11.2 Making large grids planar
Figure 11.2 shows a convenient method [7] for using a Viton O-ring to produce
uniform stresses at the boundaries of electroformed grids, making them flat. The
lower plate has a special groove cut in it with the inner wall beveled 30° with
respect to the normal. When the sandwich involving the two plates, the grid, and the
O-ring is compressed by adjusting bolts around the perimeter, the motion of the
O-ring gradually pulls the mesh tight as the O-ring expands outward in the beveled
groove under pressure. This produces the desired flat grid.
11.3 Grid Fabrication Techniques—Conical Grid
The fabrication of a grid having a precise geometry is accomplished by using a

mandrel of the desired shape, and on p. 119, methods for making hemispherical
grids are shown. Here we describe the fabrication of a conical grid that is used as an
ion cage in a high-efficiency molecular beam ionization detector with a short ion-
ization region [8].
The conical cage, which was designed with specific ion optical properties, is
shown mounted on its lens plate in Fig. 11.3a. Four angled support wires are
welded to the plate and provide rigidity for the grid.
Figure 11.3b shows the metal mandrel, which has been fabricated to the desired
shape so that it can fit the lower electrode mount of a tabletop spot-welder. Ta wire
(0.1 mm diameter) is wound on the mandrel on a lathe, operated at slow speed or by
hand, and then the ends of the coil are fixed with tape or by a soluble glue dot. After
the four support wires are spot-welded to the grid on the mandrel, the grid assembly
may be removed by a short immersion in an ultrasonic bath. Following removal of
the grid from the mandrel, the angled support wires are welded to the lens plate,
using the mandrel for centering the grid assembly on the lens aperture [9].
122 11 Grids
Fig. 11.3 Grid fabrication techniques—conical grid. a Conical grid. b Mandrel + spot-welding
References
1. Unique Wire Weaving Co., Inc., 762 Ramsey Avenue, Hillside, NI07205-1094; Electronic
Space Products International, 1050 Bensen Way, Ashland, OR 97520
2. D.A. Shaw, D.A. Bennett, Rev. Sci. Instrum. 43, 1706 (1972)
3. R.A. Taylor, J. Vac. Sci. Technol. A6, 2583 (1988)
4. J. Koch, Rev. Sci. Instrum. 45, 1212 (1974)
5. P.W. Palmberg, D.C. Johnson, H.J. Boll, Rev. Sci. Instrum. 35, 244 (1964)
6. Electroformed mesh is a product of Buckbee Mears, 245 E. Sixth Street, St. Paul, MN 55101,
and it may be obtained in various materials, with various thicknesses and optical transmission
7. R.J. Noll, L.L. Ochalla, J.C. Weisshaar, Rev. Sci. Instrum. 62, 246 (1991)
8. K. Kuhnke, K. Kern, R. David, G. Comsa, Rev. Sci. Instrum. 65, 3458 (1994)
9. Dr. R. David, Institut fur Grenzflächenforschung und Vakuumphysik, Forschungszentrum
Chapter 12
Conductive Coatings
12.1 Deposition of Electrically Conductive SnO2 Films
Many devices used in surface science research require the deposition of electrically
conductive films. These films are often used as the backing for phosphor screens in
LEED apparatus and other electron beam display techniques.
Figure 12.1 shows the method used in many laboratories for the deposition of
semiconducting SnO2 films on Pyrex glass surfaces. A glass surface, appropriately
shielded on the inside and outside with Aquadag (colloidal graphite) [1] to prevent
unwanted deposition, is placed into a small furnace or glassblower’s annealing
oven. The oven is slowly brought to the annealing temperature of Pyrex, 550 °C.
A saturated solution of SnCl2 in isopropanol is made up and loaded into the
reservoir of an atomizer, as shown. Since we wish to deposit SnO2, the atomizer is
connected to a regulated oxygen tank, and oxygen gas is used to oxidize the tin
from Sn(II) to Sn(IV) upon deposition of the SnCl2 uniformly to the hot surface of
the glass. The spraying is carried out in short bursts, and the oven door is then
closed between bursts in order to rewarm the glass to the deposition temperature.
An experienced operator can judge the conductivity of the film from its color. Blue
and barely visible coatings possess approximately kΩ resistances between two
points a few centimeters apart, while yellow or brown coatings give less than 1 kΩ
resistance.
Caution should be taken not to breath in during the film deposition with the
SnCl2, since pungent HCl gas is a major product of the pyrolysis chemistry leading
to SnO2 deposition.
Following film deposition, the Aquadag can be removed using an abrasive
kitchen cleanser and a cotton swab in water, or by heating in a glass annealing oven
for long times, followed by washing.
The electrical contact to the conductive film can easily be made using a
glass-metal feedthrough fabricated by the glassblower. To the inner end of the metal
feedthrough, a small piece of platinum ribbon or wire is welded in such a manner

DOI 10.1007/978-3-319-17668-0_12
124 12 Conductive Coatings
Fig. 12.1 Deposition of conductive SnO2 films
that it makes good contact with the inner surface of the glass. The contact is made
permanent by applying either platinum bright or gold bright paint [2] with a small
camel’s hair brush, followed by firing in the oven to produce a shiny film of metal
which is both adherent and conductive. Alternate methods for making electrical
contacts through glass to a conductive SnO2 film are described in [3].
The work function of SnO2 films prepared in this way has been measured using a
retarding potential method for an incident electron beam. The coatings were found
to have a work function of 4.79 ± 0.07 eV. Spatial variations in the work function
averaged ±10 meV, and work function differences between samples were less than
100 meV [4].
The original scientific report of the use of SnO2 films for producing conductive
glass may be found in [5]. Here also is a study of the resistivity of a number of these
films from 1.5 to 500 K.
Tin oxide films can be removed by “nascent hydrogen” produced near the film
surface by dipping a cotton swab first in HCl(aq) and then in powdered zinc dust.
This reactive mixture, when rubbed on the SnO2 film, will efficiently remove it by a
reduction process. Thorough washing in H2O should then be carried out.
References 125
References
1. Aquadag colloidal suspensions may be obtained from Acheson Colloids Co., 1607 Washington
Avenue, P.O. Drawer 611747, Port Huron, MI 48061-1747
2. Organometallic paints containing either Pt or Au may be obtained from the Hanovia Division of
Engelhard Industries, 1 West Central Avenue, East Newark, NJ 07029. These paints contain the
metal chemically bound with organic ligands which are thermally unstable
3. R. Gomer, Field Emission and Field Ionization (AIP, American Institute of Physics, New York,
1993), pp. 169–170
4. J. Burns, E. Yelke, Rev. Sci. Instrum. 40, 1236 (1969)
5. R. Gomer, Rev. Sci. Instrum. 24, 993 (1953)
Chapter 13
Phosphor Screens
13.1 Sedimentation Method for Depositing Phosphor

Screens
A number of methods involving electron beam measurements require the use of

phosphor screens that are uniform and adherent. The method described below is a
variant of techniques originally described for coating cathode ray tube screens and
field emission microscope screens [1–4]. The method may be used for either
glass-based phosphor screens, which are viewed by transmission from the back
side, or metal-based screens, which are viewed from the front side [5].
The essential requirements are to deposit a phosphor powder uniformly along
with a binder substrate that will be adherent during bakeout and use. All such
screens are sensitive to mechanical disturbance, including touching with the fingers.
For deposition of metal-based screens the underlying metal substrate must be
pickled after chemical cleaning to prepare its surface. Two pickle agents are
described for stainless steel substrates: (a) 30 volume % HNO3 + 3 volume % HF in
water—warming above room temperature will increase its rate of action and the
substrate should be immersed for 30–60 s; (b) 10 volume % HNO3 + 2 volume %
HCl + 2 volume % HF in water-immerse screen for 5 min in an ultrasonic bath. One
pickle agent is described for aluminum substrates: 12 volume % HF—then rinse
with water and dip in 86 volume % HC1 for no longer than 60 s. If the aluminum
turns black during the HC1 dip, repeat the HF step followed by the HC1 step [5].
The clean and pickled substrate is then transferred to the 3-L sedimentation
chamber shown in Fig. 13.1, and is ready for application of the phosphor screen.
Two sedimentation solutions are prepared as described below:
Solution A This is an electrolyte solution that facilitates the gelation of the
phosphor particles with the silicate. 300–500 ml water + 1.8 g Ba
(NO3)2.

DOI 10.1007/978-3-319-17668-0_13
128 13 Phosphor Screens
Fig. 13.1 Sedimentation deposition of phosphor screen
Suspension B 150–175 ml water + 0.18 g silica gel [6] + 3.2 g phosphor.

This suspension should be agitated in an ultrasonic bath or an
ultrasonicator (not the same thing).
After mixing thoroughly, transfer suspension B to a large Erlenmeyer flask and

dilute to 500 ml with water, using a wash bottle to make a complete transfer. Add
solution A and make a final volume of about 1 L. Shake vigorously and transfer
to the ultrasonic bath for continued agitation for 10 min. Pour the suspension into a
3–4-L beaker and rinse all material into the beaker. Dilute to about 3 L, and stir
rapidly for 5–10 min using a magnetic stirbar of 3-in. size. Turn off the stirrer and
allow the suspension to partially settle, during which time the large particles will
preferentially settle to the bottom. Decant the liquid into another 3-L beaker, being
careful to leave the large particles on the bottom undisturbed. Allow this second
suspension to settle for about 7 min and then carefully decant the solution into the
sedimentation chamber containing the substrate. This should be done carefully,
directing the suspension down the side of the sedimentation chamber using a large
funnel and avoiding bubbles. Cover the sedimentation chamber and allow it to sit for
5–12 h without agitation or vibration. Following sedimentation, the most difficult
task is to drain off the liquid without mechanically disturbing the screen. The drain
stopcock should be opened very slightly so that a very slow drip is achieved, making
the flow rate so small that about 6 or more hours are required to drain the liquid.
13.1 Sedimentation Method for Depositing Phosphor Screens 129
The drained screen should be air dried in the sedimentation chamber for several
hours, keeping it covered to avoid dust. After drying, inspect for homogeneity.
A “ring” on the screen indicates a discontinuity owing to a change in the drip rate,
possibly due to a particle being trapped in the stopcock. Such discontinuities may
not be serious unless photometry is to be employed, where phosphor thickness must
be constant. Even without visible evidence of inhomogeneities, the phosphor
thickness will vary over the screen surface owing to the variable angle of inter-
ception of the settling particles by the curved screen. Following drying, the screen
should be baked in an oven at 80 °C for 30 min to remove additional water. Screens
treated in this manner are satisfactory for use in ultrahigh vacuum and show little or
no influence on the base pressure after a bakeout at 150 °C for 24 h [5].
For deposition of phosphor screens in glass vacuum envelopes, minor variations
of the procedure described above may be used. The sedimentation chamber may be
replaced by the envelope itself if the depth of the liquid suspension is sufficient.
Alternatively, the glass vacuum envelope can be immersed in a large sedimentation
vessel, and in either case the draining can be done by very, very slow tipping of the
envelope using a slowly driven motor that turns once in 24 h.
Screens prepared in this manner are approximately 10 μm thick and well suited
for electron beam imaging. Thicker screens may be subject to electrostatic charging.
An unusual electrostatic deposition and adhesion method has been described by
Cetronio and Cross [7].
13.2 Dusting Method for Coating Phosphor Screens
Phosphor coating of screens for imaging electron beams may be achieved using a
technique developed years ago at AT & T Bell Laboratories [8] and described
briefly elsewhere [9]. This method has the advantage over the sedimentation
method of producing more uniform phosphor coatings, since it does not sense a
curved surface during the deposition procedure.
As shown in Fig. 13.2, a glass or metal substrate, cleaned with a strong hot
alkaline detergent, has a stopper made of cork or plastic (but not rubber) that seals
the hole for the electron gun or other source. A sticky uniform layer of phosphoric
acid, H3PO4, is deposited by the following procedure:
Clean about 100 glass balls (3-mm diameter) and the screen with ethanol and
acetone to degrease. Then from a solution consisting of 3 drops of H3PO4 in 25 ml
of acetone, place 25 drops onto the screen substrate with the stopper in place. Place
the glass balls onto the screen substrate and roll them around as the acetone
evaporates. After 2–3 min, the acetone will have evaporated, and the phosphoric
acid will uniformly wet the substrate. If streaks are observed, repeat the process.
Then shake 1–2 cm3 of phosphor powder onto the wet substrate and tap the screen
until the phosphor coats the entire surface. Invert the screen and tap to remove the
loose powder. Remove the stopper and using a gentle jet of nitrogen, blow off any
130 13 Phosphor Screens
Fig. 13.2 Dusting method for deposition of phosphor screen
remaining loose phosphor. The screen should be baked in air at 200 °C for 2 h and
it is then ready for use in ultrahigh vacuum.
Caution should be exercised to prevent breathing of phosphor or contact with the
skin.
A unique method for electrostatically depositing and adhering a phosphor screen
is described in [10].
Another unique method for depositing a phosphor powder on a flat, tin
oxide-coated plate is described in [11]. Here, a glass plate which has been chilled in
air below the dew point is coated with phosphor powder from an atomizer. The
phosphor uniformly adheres to the moist plate. Drying the plate in a dessicator
removes the water, and the phosphor adheres to the plate without any adhesive. The
adhesion is sufficiently strong to permit the use of the phosphor screen in any
position in a vacuum system, yet the coating can easily be removed by wiping
under water.
References 131
References
1. F. Rosebury, Handbook of Electron Tube and Vacuum Techniques (American Institute of

Physics, New York, 1993), pp. 251–254
2. J. Millman, Vacuum-tube and Semiconductor Electronics (McGraw-Hill, New York, 1958)
3. K.R. Spangenberg, Vacuum Tubes (McGraw-Hill, New York, 1948), pp. 429–432, 821
4. R. Gomer, Field Emission and Field Ionization (American Institute of Physics, New York,
1993), p. 167
5. D.K. Flynn-Sanders, S.-L. Chang, P.A. Thiel, and R. Imbihl, J. Vac. Sci. Technol. A10, 413
(1992). In addition to the phosphor deposition procedure described here, this paper has a
description of several phosphors and their light emission properties. One should be especially
careful with beryllium-containing phosphors because of their poisonous nature and their
possible entry to the human body through cuts in the skin
6. Silica gel which works in this application a product of Davidson Chemical which is a division
of W.R. Grace, 7221 W. Parkland Court, Milwaukee, WI 53223. The surface of the silica is
hydroxylated and the total surface area is about 300 m2/g. In general silica gels have smaller
particle sizes than silica powders
7. A. Cetronio, J.A. Cross, J. Vac. Sci. Technol. 15, 1191 (1978)
8. R. Carr, E. Chaban, I. Vac, Sci. Technol. A10, 3595 (1992)
9. E.W. Muller, T.T. Tsong, Field Ion Microscopy (Elsevier, New York, 1969), p. 139
10. A. Cetronio, J.A. Cross, J. Vac. Sci. Technol. 15, 1191 (1978)
11. R.C. Abbott, P.M. Brown, Rev. Sci. Instrum. 38, 832 (1967)
Chapter 14
Thermionic Emitters
14.1 Thoriated Thermionic Emitters
Thermionic emitters are very widely used in surface science experiments as sources
of electrons. The emission of electrons in the absence of applied electric fields
follows the Richardson equation,
I=S ¼ AT 2 ee/=kT
where I/S is the electron current density, ϕ = the work function of the emitter, e is
the electronic charge, T is the temperature in K, and k is the Boltzmann constant.
For metals, A = 120 A/cm2 K2.
Various coatings on thermionic emitters are used to reduce the work function,
permitting the thermionic emitter to operate at lower temperatures while delivering
the same current density as an unmodified emitter at higher temperatures. Some
aspects of the various coatings and their advantages and disadvantages in ultrahigh
vacuum systems are described by Redhead et al. [1]. For various reasons, ThO2
coatings on W or Ir surfaces exhibit the most favorable properties for use in
ultrahigh vacuum. While long-term operation in an oxygen ambient is not possible
with either metal, Ir will withstand short accidental exposure to oxygen without
burnout.
Figure 14.1a shows the simple experimental procedure that may be used for
coating a thermionic emitter with a layer of ThO2 by cataphoretic deposition [2].
A suspension of 200 mesh ThO2 powder (5 g) in a solution containing 0.075 g of
thorium nitrate in 100 ml of 95 % ethanol is used. This suspension is allowed to
stand for a few minutes so that the large ThO2 particles settle to the bottom and are
removed from further considerations. The suspension is electrically deposited on
the filament at a negative potential of a few volts relative to the Pt anode. The
required current density depends on the shape of the surface being coated, and is
found by trial and error. It is best to electropolish tungsten substrates in NaOH

DOI 10.1007/978-3-319-17668-0_14
134 14 Thermionic Emitters
solution (tungsten positive) and to rinse in water before beginning the ThO2
deposition. The white ThO2 powder will form a very uniform deposit on the emitter
surface, and when fired in vacuum, this coating will be rather strongly adherent.
Coatings between 0.025 and 0.040 mm in thickness are satisfactory [2]. Firing in
vacuum to about 1800 K brightness temperature activates the thoria fully [2].
Reference [2] contains many interesting measured properties of ThO2.
The maximum current density that may be drawn from both clean W and ThO2
surfaces is shown in Fig. 14.1b. Since the average work function of W (about
4.5 eV) is reduced to about 3.0 eV by thoriation, a tremendous enhancement of
electron emission is achieved, as shown in Fig. 14.1b [1].
Fig. 14.1 Thoriated tungsten thermionic emitters. a Cataphoretic ThO2 deposition on a thermionic
emitter. b Maximum current density for thermionic emitters. c W Evaporation rate for thermionic
emitters
14.1 Thoriated Thermionic Emitters 135
One of the reasons for using thoriated thermionic emitters is to reduce the
operating temperature for electron emission to levels that involve negligible
evaporation of the metal. Figure 14.1c [1] shows the evaporation rate of W atoms
versus electron emission current density for both W and ThO2/W emitters. At the
higher emission current densities for pure W, significant evaporation of W will
occur over long experimental times, whereas ThO2/W emitters involve negligible
W evaporation rates.
It is important not to bombard the ThO2 coating with electrons, which will
cause damage. Even using AC accelerating voltages is detrimental in degassing
ionization gauge grids, when the hot grid can be a thermionic emitter on the reverse
half-cycle [1].
It has been reported that ThO2-coated W filaments often emit unknown impu-
rities that can be contaminants in surface science experiments [3]. This was studied
for ThO2/W filaments contaminated by F [4], and significant transfer of F or
F-containing species to a clean single crystal of W was observed. Other halogen
impurities such as Cl are also likely to be present in ThO2 coatings exposed to new
vacuum systems, since both HCl and HF are common pickling agents in cleaning
stainless steel [5]. These effects can be eliminated eventually by outgassing new
ThO2-coated emitters for long periods of time [4]. Such halogen contamination of
the grid surfaces of a Bayard-Alpert gauge renders the gauge useless for low
pressure measurements, since electron impact desorption from the halogen-covered
grid surface will supply spurious ion currents [6].
In addition to W wire thoriated by surface coatings of ThO2, it is also possible to
purchase W wire that contains alloyed Th at the l–2 % level [7, 8]. This small
amount of thoria was originally added to decrease crystallite slippage in the
polycrystalline wire as it was formed by drawing through a die. These wires also
exhibit greatly enhanced thermionic emission, compared to pure W. Despite high
vacuum conditions, these emitters are readily poisoned [9]. This poisoning process
may be diminished by carburizing the wire in a hydrocarbon gas, such as acetylene
at temperatures near 1700 °C [8]. An outer layer of W2C, about one-tenth the wire
diameter in thickness is produced [8].
A thoriated Ir thermionic emitter has been reported to have high resistance
to exposure to air and to exhibit low Th evaporation rates at operating
temperatures [10].
Thorium carbide thermionic emitters exhibiting an effective work function of
about 3.3 eV have been produced and will emit about 10 A/cm2 at 2150 K [11].
14.2 Lanthanum Hexaboride Thermionic

Emitters—Deposition on Metals
The use of cataphoretic deposition of thorium oxide, ThO2, on thermionic emitter

metal surfaces for the purpose of reducing the work function has been described on
p. 133. A similar technique can be used to produce LaB6-coated metal emitters.
Solid LaB6 thermionic emitters have been described on p. 137, but these are only
useful for well-defined geometries and are costly. Thus for small, disposable LaB6
emitters, cataphoretic deposition on metal ribbons or filaments provides an excel-
lent route to useful thermionic emitters.
Five factors have been identified as being of importance in LaB6 deposition.
They are:
1. The particle size and the relationship between particle weight and surface area.
2. The concentration of cations in the suspension holding the LaB6 particles.
3. The ability of cations to adsorb onto the LaB6 surface, providing surface charge
to permit particle migration under electrolytic conditions.
4. The electrostatic potential gradient causing the particles to migrate.
5. The adherence of the particles to the metal cathode surface.
The most efficient deposition occurs with particles having a high surface charge and
migrating in a strong electrostatic field.
The present study [12] has been built on earlier methods [13, 14]. Re ribbons,
0.0025 × 0.76 × 12.7 mm3 were used. These filaments were heated white hot in
vacuum to clean them before external cataphoretic deposition. A Ta anode was
used, and the current (0.2–1.0 mA) was adjusted to achieve optimum deposition of
LaB6 in 30–300 s. For best results, 10 g of very finely divided LaB6 was suspended
in 100 ml of methanol and about 0.05 g of NH4NO3 was added as the electrolyte,
dissolving this first in a small volume of methanol. These suspensions have a short
lifetime of the order of several weeks owing to hydrolysis of the LaB6, which
produces a white coating on the powder that interferes with adhesion to the cathode.
A Ta ribbon was used as an anode, and it was noted that the deposition occurs at the
Re ribbon edges owing to higher electric field strength there. In order to avoid this,
the geometry shown in Fig. 14.2 was used, with the Ta anode parallel to the
meniscus of the suspension, and the ribbon pulled almost out of the suspension to
reduce the tendency for deposition on the edges. A narrow gap between cathode
and anode is preferred to increase the electrostatic field. In addition, deposition
against the gravitational field leads to the selection of smaller LaB6 particles which
adhere better than large particles. The LaB6 thickness could be adjusted up to
0.5 mm, but little difference in performance was noted for various thicknesses.
Deposition on filaments can be done with a large anode and with the electrodes in a
vertical position, since the electric field at the cylindrical filament surface is uni-
form [12, 14].
Failure in the adhesion of LaB6 to Re supports has been observed following
cataphoretic deposition. It was found that either boron or boric oxide impurity in the
LaB6 was responsible for the adhesion loss, and a purification procedure was
described which produced LaB6 with excellent adhesion properties on Re surfaces
[13].
LaB6 may also be deposited on Ta foil or filament surfaces if the surface is
carburized first to produce a barrier layer of TaC that protects the underlying Ta
from attack by the LaB6 at high temperatures [15].
14.3 Directly Heated Lanthanum Hexaboride Thermionic Emitters 137
Fig. 14.2 LaB6(s) Cataphoretic deposition
14.3 Directly Heated Lanthanum Hexaboride

Thermionic Emitters
The work function of a tungsten thermionic emitter can be reduced significantly by

using a thoria coating (see p. 133). In addition, it is well known that LaB6 coatings
on refractory metals such as W or Ta serve to produce a low work function
thermionic emitter of limited lifetime [16] (see p. 135), and LaB6-coated refractory
metal emitters are widely used when high electron current densities are desired at
low emitter temperatures [16]. The LaB6 structure is also remarkably stable toward
damage by positive ion bombardment that may occur in some high-voltage devices
[16]. The degradation of LaB6 occurs by means of B diffusion into the W or Ta
substrate, accompanied by La vaporization. This effect can be minimized by car-
burizing the substrate before coating with LaB6. With rhenium as the base metal,
this degradation effect is minimized [17]. For bulk LaB6, the surface La atoms
vaporize slowly at high temperatures, and are replaced by La atoms from the bulk
[16]. Other authors have also investigated the properties of LaB6 as a thermionic
emitter [18, 19].
Bulk LaB6 may be used also as a thermionic emitter material. Questions about
the best mounting procedure to avoid stresses (owing to thermal expansion) and to
avoid degradation at the electrical contact points then come into play [17].
Figure 14.3a shows an unsuccessful mounting procedure for mounting a LaB6 strip
between two connecting electrodes made of graphite. While the emitter gave
Fig. 14.3 LaB6(s)

Thermionic emitters. a Rigid
mount. b Flexible. c Flexible.
d LaB6(s)—T versus I.
e LaB6(s)—Ie/S versus T.
f Wt. Loss versus Ie/S
14.3 Directly Heated Lanthanum Hexaboride Thermionic Emitters 139
copious emission at 1400 °C, it cracked upon cooling as a result of its high coef-
ficient of thermal expansion. To avoid this problem, rectangular and circular fila-
ments were constructed as shown in Fig. 14.3b. These structures survived
temperature cycling. The structure in Fig. 14.3b is mounted on graphite holders,
while that in Fig. 14.3c is mounted on Mo holders.
Figure 14.3d gives the temperature versus current for a LaB6 strip, 1.6 × 3 mm2
in cross-sectional dimension. The points are experimental points, and the line is a
theoretical line calculated on the basis of the ohmic heating and the radiative losses
[17]. Figure 14.3e shows the thermionic electron emission of LaB6 compared to Ta
and W emitters. Approximately four to five orders of magnitude difference in
current density is observed compared to the refractory metals.
As mentioned before, one is concerned with La evaporation from LaB6 when it
is heated to sufficient temperature. However, one is also concerned about metal
evaporation effects when refractory metals are used as thermionic emitters, since
they must work at very high temperatures (see p. 133). Figure 14.3f shows a
comparison of the evaporation rate from LaB6 compared to W and Ta, obtained
from published data, where the ordinate is in g cm−2s−1. On this basis, a LaB6
cathode can last for several thousands of hours at current densities of 10–
20 A cm−2. Above 50 A cm−2, the useful life may still be greater than 200 h [18].
A method for welding small LaB6 shapes to Re wire for making compact
high-efficiency thermionic emitters has been described [20]. In addition, a method
for protecting a W filament from LaB6 attack by interposing a layer of ThO2 has
been described [21].
A method for mounting conical tipped LaB6 cathodes for electron beam
instruments has been commercially developed. The cathode is mounted between
pyrolytic graphite blocks that are clamped between Mo/Re alloy support posts that
have a high modulus of elastically at high temperatures. The method is also used for
CeB6 cathodes [22].
14.4 Thermionic Emitter Mounting
The provision of long-lived thermionic emitters for electron bombardment requires

special mounting and connection procedures. Figure 14.4a shows the essentials of
mounting a thermionic emitter for stability and also for ease of replacement [23].
A fixed alumina tube with two holes fits a spiral emitter, made from W/3%Re,
about 0.2 mm in diameter. The coil is wound uniformly around a mandrel in such a
fashion that its ends can be slipped into the holes in the alumina tube without
producing any strain on the coil. If this is not done, the coil will distort severely
upon first heating. The W/3%Re alloy is preferred over pure W because of its
tendency to not become so brittle during heating.
Friction connections are made with heavier Mo conductor wires by slipping in
0.1–0.2-mm-diameter filler pieces from the back side until a tight fit is achieved
Fig. 14.4 Thermionic emitter mounting. a Open emitter. b Shielded emitter
between the emitter and the conductor wires. This method avoids welding and
allows the emitter coil to be disassembled for replacement with ease.
Connection to the Mo conductor wires is conveniently made with Cu braid, and
a soft Cu connector tube (annealed in a flame) that is crimped tight several times
around the braid and the Mo wire using pliers. The assembly should be arranged
such that the braid does not strain the Mo conductor wires by its weight or motion.
Figure 14.4b shows a similar device that utilizes a Mo electron repeller shield.
The alumina tube is flattened on two sides and is placed into a fitted channel at the
back of the Mo shield. The Mo shield will charge negatively during operation, but
for heavy electron bombardment heating, the crystal must be biased positively.
Macor ceramic supports are inappropriate for these purposes for two reasons.
The material degrades at high temperatures, and fluorine compounds are evolved at
high working temperatures [24].
14.5 Indirectly Heated Cathodes for High-Temperature Operation 141
14.5 Indirectly Heated Cathodes for High-Temperature

Operation
Cathodes for electron guns and other devices are often coated with low-work-
function materials that permit higher current densities to be drawn than would be
possible with pure refractory metal cathodes such as W or Mo (see p. 133). Coated
cathodes are often heated to temperatures in excess of 2000 K for activation and
operation. Directly heated cathodes may be employed, but the requirement for an
emitter to be at equipotential over its surface rules out direct ohmic heating.
Thus, indirectly heated cathodes are often used. The design shown in Fig. 14.5a
uses an ohmically heated W filament to heat a Mo disk containing a work-function-
lowering material on its surface [25].
A self-supporting W filament (0.38-mm diameter), wound in a double helix on a
mandrel into a coil 9.5 mm in length, is used to heat a Mo disk cathode mounted on
the helix axis. The Mo disk is at the end of a cathode cylinder, 3.8 mm in diameter
and 10 mm long. Mounted on the outside of the Mo cathode cylinder are 3 (0.1-mm
thick) polished Ta radiation shields. These shields reduce the radiation transmitted
to the outer stainless steel support cylinder such that a measurable color temperature
for the stainless steel sleeve is not reached.
Figure 14.5b shows the W heater life behavior versus the heater current for
operation in ultrahigh vacuum. The temperatures associated with each point were
measured pyrometrically on the Mo cathode surface.
In the example shown here, only radiation heating of the Mo cathode was used.
However, this may be assisted by applying a positive potential to the cathode
relative to the filament, and drawing electron bombardment current to the cathode.
This method of heating is well suited for fast flashing of the cathode which is often
necessary during activation of the cathode coating (about 2500 K for ThO2).
Electron bombardment also involves lower W heater temperatures, and will not
shorten the heater lifetime.
14.6 Filament Power Supply Tester
Often it is necessary to be able to test a filament power supply without having a

filament operating in vacuum use in the test. This circuit diagram shows a dummy
filament which may be used in the test, without actually having to connect to a
filament for the power supply test. The power supply filament and anode connectors
are connected as shown. For DC operation, one has a choice of simulating
a *2A-filament or a *3–4 A filament. Switch S2 controls the current range desired.
For AC operation, switch S1a and S1b are open and the circuit utilizes the bridge
rectifier, BR66. The part of the circuit to the right of the dashed line simulates the
dummy filament thermionic emission using an NPN transistor, MJE 340. Filaments
requiring about 2A involve tungsten wire of the order of 0.125 mm diameter; 3–4 A
filaments involve tungsten wire of about 0.2 mm diameter [26] (Fig. 14.6).
Fig. 14.5 Indirectly heated cathode for high temperature operation. a Cathode design, b Cathode
lifetime versus heater current
14.7 Bayard-Alpert Ionization Gauge with Corrections

for External Radiation and External Electron
Flux Effects
Many workers have noted that external electron sources in UHV systems cause
erratic readings in ionization gauges. In the case of ionization gauges in the vicinity
of synchrotron or other radiation sources, there are also effects due to external
radiation causing inaccurate gauge readings. In early work by Redhead [27], the
effects of soft-X-ray production leading to photoelectron generation inside the
14.7 Bayard-Alpert Ionization Gauge with Corrections … 143
Fig. 14.6 Filament power supply tester
gauge itself were eliminated in various ways, resulting in new designs for gauges
with for instance two collector electrodes (the modulated ionization gauge), or to
gauges in which the ion collector was electrically and spatially isolated from the
ionization region (suppressor and extractor gauges). These modifications also led to
methods to compensate for positive errors caused by electron stimulated desorption
from adsorbed gases on the ionization gauge grid.
A normal ion gauge operating in the vicinity of an external thermionic emitter or
in a region containing high energy radiation will suffer major errors, caused by
things like external electron collection by the grid or the collector. Also
externally-produced photons interacting with the electrodes can cause X-ray pro-
duction, leading to photo-emission within the gauge, and giving spurious ion
current measurements. In the new gauge design shown in Fig. 14.7, measurements
of the current due to externally-produced electrons are made at the correcting
electrode in an auxilliary gauge which does not contain a thermionic emitter fila-
ment, and these measurements, applied to the current measurement from the normal
gauge, can be used to make corrections to the measured ion current and calculated
pressure. The details of this are dealt with in a recent paper [28], and background
information will be found in earlier papers [29–35].
Figure 14.7 shows the side view of the gauge design and its logic. ICE is the
collector current from the ion collector in the correcting electrode (CE) attachment,
and IC is the current from the normal gauge. The measurement of ICEe, the external
electron current received by the grid (and hence available within the gauge structure
for ionization of gases), is also employed to make corrections to the measured ion
current in the normal gauge. Reference [28] gives a formula to measure a com-
pensated pressure in this gauge. Figure 14.7 shows the measurements of the
uncorrected gauge, BA-S4, compared to measurements of the corrected gauge,
BA-S4. Both gauges are subjected to a flux of externally-produced electrons which
enter the two gauges from the outside. The pressure measurements made with
NIG-2F are made for a gauge which does not see externally produced electrons.
The abscissa is a plot of the filament current causing thermionic emission in an
external electron source and large emitted electron currents are produced as one
moves from 2 to 3 A. The indicated emission current in the ion gauge being tested
remains essentially constant at 3.97 mA. It is seen that without compensation, the
ion gauge suffers considerable negative errors when externally-generated electrons
are present. The compensation circuit also includes a method for measuring the
effect of external radiation on currents generated in coaxial cable, ICErr and ICrr.
Fig. 14.7 Bayard alpert ionization gauge with corrections for external radiation and electron flux
effects
14.8 Replacing Filaments in Glass Bayard-Alpert Gauges 145
14.8 Replacing Filaments in Glass Bayard-Alpert Gauges
Filament replacement in a glass Bayard-Alpert ionization gauge is quite simple if a

glassblower will move the tubulation from the normal position to a position on the
bottom of the gauge where the filament attachment posts are located [36].
Figure 14.8 shows the location of the new port, and the position of the original
tubulation. The new tubulation is sealed to the outer envelope using Corning
No. 3320 glass as a grade to Pyrex, and the seal is made with an oxidizing flame,
Fig. 14.8 Replacing filaments—Bayard-Alpert Gauges

after thorough cleaning of the gauge with Alconox detergent. The gauge should be
annealed in the upright position before use.
Replacement filaments can be made of many materials, such as iridium, thoriated
iridium, thoriated tungsten, etc. (see p. 133). The use of 0.010-in. diameter W is
easiest. The bent filament is formed into a tight V-shape, and a hook is fabricated
from the filament wire by first twisting the apex region of the wire and then making
the hook as shown in Fig. 14.8. The hook is then inserted into a wire loop at the top
of the filament support rod. When softer filament materials such as iridium are used,
a tungsten hook made of 0.010-in. wire may be used for the upper support.
Hemostats are reported to be useful in the filament insertion procedure.
The lower attachment of the two ends of the filament to the filament attachment
posts can be made in three different ways:
1. Welding, using a tweezer spot welding assembly, and making small 0.002-in.
thick Ta tabs that will weld nicely to W and Ir filaments.
2. The filament post is filed round, and a l/8-in.-diameter stainless steel barrel is
fabricated as shown in Fig. 14.8. The small tungsten locking coil fits onto the
filament post, and when the barrel is twisted in one direction, binding occurs.
Twisting oppositely results in breaking the junction. A 0.020-in. clearance is
needed when the tungsten locking coil is made of 0.010-in. W. The emitter
filament is held under an 0–80 screw head placed in a tapped hole on the
opposite end of the barrel, using a jam nut as shown.
3. The W filament may just be wrapped several times around the two support
posts.
The gauge calibration seems to be well maintained after refabrication [36].
References
1. P.A. Redhead, J.P. Hobson, E.V. Kornelsen, The Physical Basis of Ultrahigh Vacuum
(American Institute of Physics, New York, 1993), Chapter 7, section 7.6
2. T.E. Hanley, J. Appl. Phys. 19, 583 (1948)
3. P.A. Redhead, J.P. Hobson, Br. J. Appl. Phys. 16, 1555 (1965)
4. J.T. Yates Jr, D.A. King, J. Vac. Sci. Technol. 9, 1256 (1972)
5. Stainless Steel Fabrication (Allegheny Ludlum Steel Corp., 1959), p. 309
6. H.E. Winters, J.W. Coburn, Surf. Sci. Rep. 14, 161 (1992)
7. I. Langmuir, Phys. Rev. 22, 357 (1923)
8. R.G. Murray, R.J. Collier, Rev. Sci. Instrum. 48, 870 (1977)
9. R.O. Jenkins, W.G. Trodden, J. Electron. Control 12, 1 (1963)
10. H.F. Gray, E.H. Harris, T. Pankey Jr, G.A. Haas, Rev. Sci. Instrum. 43, 1113 (1972)
11. H.H. Glascock Jr, Rev. Sci. Instrum. 43, 698 (1972)
12. J.E. Delmore, Rev. Sci. Instrum. 54, 158 (1983)
13. L.J. Favreau, D.F. Koenig, Rev. Sci. Instrum. 38, 841 (1967)
14. L.J. Favreau, Rev. Sci. Instrum. 36, 856 (1965)
15. M. Nasini and G. Redaelli, Rev. Sci. Instrum. 42
16. J.M. Lafferty, J. Appl. Phys. 22, 299 (1951)
17. K.N. Leung, R.A. Pincosy, K.W. Ehlers, Rev. Sci. Instrum. 55, 1064 (1984)
References 147
18. H. Ahmed, A.N. Broers, J. Appl. Phys. 43, 2185 (1972)

19. S.P. Gordienko, E.A. Guseva, V.V. Fesenko, Teplofiz. Vys. Temp. 6, 821 (1968)
20. K.N. Ramachandran, Rev. Sci. Instrum. 46, 1662 (1975)
21. L. Bakker, ThJ van Velzen, Rev. Sci. Instrum. 37, 1404 (1966)
22. Manufactured by FEI Company, P.O. Box 654, McMinnville, OR 97128
24. Personal experience
25. J.T. Jensen Jr, J. Klebanoff, G.A. Haas, Rev. Sci. Instrum. 38, 1178 (1967)
26. Michael Schmidt and Ulli Diebold, private communication
27. P.A. Redhead, J.P. Hobson, E.V. Kornelsen, “The Physical Basis of Ultrahigh Vacuum”
(Chapman and Hall Ltd, Great Britain, 1968), pp. 304–349; reprinted as an American Vacuum
Society Classic, AIP (1993)
28. H. Saeki, T. Magome, J. Vac. Sci. Technol. A29, 061601 (2011)
29. H. Saeki, T. Momose, J. Vac. Sci. Technol. A18, 244 (2000)
30. H. Saeki, T. Magome, T. Aoki, N. Gotoh, T. Momose, Rev. Sci. Instrum. 75, 5152 (2004)
31. T. Magome, H. Saeki, J. Vac. Sci. Technol. A23, 725 (2005)
32. H. Saeki, T. Aoki, T. Momose, J. Vac. Sci. Technol. A19, 1022 (2001)
33. H. Saeki, T. Magome, Radiat. Meas. 41, S271 (2007)
34. H. Saeki, T. Magome, Y. Shoji, J. Vac. Sci. Technol. A24, 1148 (2006)
35. H. Saeki, T. Magome, Rev. Sci. Instrum. 79, 055102-1 (2008)
36. D. Bedding, T. Moran, A. Brown, Rev. Sci. Instrum. 56, 2170 (1985)
Chapter 15
Shielding
15.1 Magnetic Shielding in Ultrahigh Vacuum
It is common to employ magnetic shields inside ultrahigh vacuum systems for the
elimination in electron spectrometers of stray magnetic fields caused by the Earth or
by magnets associated with ion pumps or other devices near the vacuum system.
A number of formulas have been presented for the calculation of the efficiency of a
magnetic shield [1–3]. These formulas do not consider deviations from pure
cylindrical geometry caused by lips, caps, and pumpout ports. The calculation
method studied here is able to approximate the nonideal conditions caused by
deviations from perfect cylindrical geometry [4].
A computer code, JASON [5], has been used. This code solves the linear Poisson
equation. The Earth’s magnetic field is of the order of 500 mG. It is desired in this
particular example to reduce the magnetic field to less than 1.0 mG inside the shield
shown in Fig. 15.1. The following constraints exist in this design as well as in
designs to be made in other ultrahigh vacuum systems: (1) A small number of
shield layers should be employed to reduce the additional surface area and out-
gassing which occurs from these surfaces; (2) Holes must be cut into the shield for
pumping by the ultrahigh vacuum system. The pumping speed through a hole for
atmospheric gases is about 11 L/cm2 s; (3) The sample must be moved into the
magnetically shielded spectrometer along the Z axis of the cylindrical shield.
The optimum shield material thickness, t, is
t 3a=2l ð15:1Þ
where a is the inner radius of the cylinder and μ is the permeability of the
magnetic shielding material. If the value of t needed, based on (15.1), is larger than
the thickness of available material, two concentric shells may be used in shielding,
at the expense of introducing more surface area.

DOI 10.1007/978-3-319-17668-0_15
150 15 Shielding
Fig. 15.1 Magnetic Shielding in UHV
Because the JASON program rotates a two-dimensional array about a cylinder

axis, Z, it is geometrically impossible to specify three-dimensional holes along the
shield perimeter. Instead, in a first-approximation method, a gap along the perimeter
is used to model the desired hole size where the open area in the gap is equivalent to
the hole area. This gap has the permeability of vacuum in the simulations. A second
approximation method uses a gap with three times the pumping hole area, con-
taining a magnetic shielding alloy that is two-thirds made of the shielding material,
and one-third made of vacuum, i.e., specifying a “hole alloy.” The actual hole
configuration consists of radial holes of 3-cm-diameter, placed 6 cm apart around
the ends of the two outer shields, as shown in Fig. 15.1.
The material chosen is a high permeability material, ad-mu 80 [6]. Its thickness,
t, is 0.25 mm. The Figure shows a cross section through the optimized shield made
of three concentric cylinders that contain pumpout ports as shown, as well as the
axial sample insertion region. Approximation (1) gives approximately a 1000-fold
attenuation of the Earth’s field; approximation (2) gives approximately a 10,000-
fold attenuation, both measured along the Z-direction at the center of the spec-
trometer. To make a rough comparison with actual measurements, the attenuation
of an AC magnetic field, naturally present in the spectrometer room, was measured.
Attenuation factors in the range 1900(X), 640(Y), and > 3600(Z) were measured in
general agreement with the calculations.
It should be remembered that mechanically straining magnetic shielding made of
mu metal will reduce its effectiveness. The effectiveness of annealing magnetic
shielding is dealt with in [7].
15.2 Electrical Isolation of UHV Components 151
15.2 Electrical Isolation of UHV Components
It is sometimes necessary to isolate electrically one ultrahigh vacuum component

from another. This may be done using a conventional ceramic electrical isolator,
connecting the two components with Conflat flanges on either side of the ceramic
isolator. This arrangement requires linear space, and also, small lateral forces on the
ceramic isolator may cause it to crack. A novel solution to this problem is presented
in Fig. 15.2 [8].
Polyimide film (Type H), known as Kapton [9], is used. The film thickness is
0.127 mm. Two polyimide washers are cut with punches that have a clearance of
0.0127 mm. Sleeves, made by rolling up the polyimide film, with 2-mm overlap,
are prepared and cemented with a fast-setting epoxy cement. This is best done using
bolt-clearance holes drilled in a polyethylene block to hold the tubes during glue
setting.
Two copper gaskets are sandwiched together around a polyimide spacer ring that
is glued between the gaskets using a Dow Corning No. 997 silicone varnish [10].
The varnish is applied, the gaskets are placed on either side of the polyimide spacer,
and the assembly is heated at 95 °C for 2 h in an oven for curing while being
clamped under pressure. The polyimide spacer protrudes slightly beyond the inner
and outer edges of the copper gaskets.
As shown in Fig. 15.2, the polyimide washers fit under stainless steel washers.
The polyimide washers mechanically and electrically separate the stainless washers
and bolt heads from the flange surface. The polyimide sleeves electrically separate
the bolts from the bolt holes. The polyimide spacer ring between the copper gaskets
separates the flanges from each other electrically.
Thermal shock tests between 260 and 10 °C resulted in no vacuum leak
development. The devices do not add gas load to a vacuum system, and the UHV
background mass spectrum did not change as a result of 10 of these devices being
installed on the same system. The electrical resistance is greater than 1011 Ω, and
electrical breakdown of the assembly did not occur below 17 kV. Breakdown does
not result in failure of the vacuum integrity of the seals.
152 15 Shielding
Fig. 15.2 Electrical isolation of UHV components
References
1. A.J. Mager, IEEE Trans. Mag. MAG-6, 67 (1970)

2. A. Mager, J. Appl. Phys. 39, 1914 (1968)
3. D.U. Gubser, S.A. Wolf, J.E. Cox, Rev. Sci. Instrum. 50, 751 (1979)
4. S. Dobscha, W. Siekhaus, Rev. Sci. Instrum. 57, 3123 (1986)
5. S. Sackett, JASON, UCRL-18721, Lawrence livermore national laboratory, Livermore, CA
94550
References 153
6. Ad-Vance Magnetics, Inc., 625 Monroe Street, P.O. Box 69, Rochester, IN 46975
7. D.L. Martin, R.L. Snowdon, Rev. Sci. Instrum. 46, 523 (1975)
8. J.C. Jones, Rev. Sci. Instrum. 46, 489 (1975)
9. Obtained from: Dupont, 1007 Market Street, N-2426-A, Wilmington, DE 19898; see Dupont
high performance films, Bulletin H-38492 (1983)
10. Obtained from: Dow Corning, P.O. Box 0994, Midland, MI 48640–0994
Chapter 16
Single Crystal Fabrication/Orientation
16.1 Making Small Au Single Crystals from Au Wire

for STM and Other Studies
The production of small spherical single crystals from Au wire provides a conve-
nient way of doing STM experiments on the various single crystal plane facets
which are exposed on the surface [1]. A procedure outlined below has been
employed in a number of STM studies on Au(111) and other crystal planes [2–10].
Low purity Au wire of 0.3 mm diameter was employed. The wire is first cleaned
in piranha solution (3:1 mixture of concentrated H2SO4 with 30 % H2O2) to remove
organic impurities and then in boiling concentrated HNO3 to remove transition
metals for about 20 min. The impurities from the bulk of the wire are removed by a
“zone refining” procedure like that used in the semiconductor industry. The wire is
clamped in a ceramic tweezer with the tweezer face horizontal and the wire hanging
down. Use a 3 cm portion of the wire. Then bring a tiny hydrogen flame slowly
from the lowest end of the wire and start melting it. Due to surface tension effects,
the Au melts into a bead. Do this slowly, and as soon as the bead size is about the
size of the wire diameter, remove the flame and allow the Au to crystallize. Then
reheat, remelt and move the melting front (readily visible) upwards away from the
bead. The impurities get stuck at the melting front and you then swipe them
upwards from the bead to the junction between the bead and the wire. The impu-
rities also segregate to the Au surface in this process.
After two or three cycles of this melting and sweeping procedure, repeat the
cleaning procedure (boiling in HNO3 and aqua regia—a mixture of concentrated
HNO3 and HCl which will dissolve Au) to dissolve a portion of the bead. The
surface of the Au bead and the interface region between the bead and wire, con-
taining zone-migrated impurities, will dissolve removing the impurities segregated
there and the etched bead will appear non-specular. Then repeat the whole proce-
dure several times and finally grow the bead to the desired size less than 1 mm in
diameter. Facets will appear on the bead. If you inspect a facet and see a single

DOI 10.1007/978-3-319-17668-0_16
156 16 Single Crystal Fabrication/Orientation
Fig. 16.1 Making small Au single crystal
visible particle, then the surface is not going to be STM clean. The single crystal
bead is conveniently mounted by its Au wire tail, spot welding to a Ta foil support.
The <111> facets will be several hundred µm in size. These facets feature zero
miscut angles since they are natural facets. The (111) terrace width can be as large
as several µm in the center of a facet. The edges of the facet approach the curved
surface of the bead providing a variety of vicinal Au(111) surfaces, where
single-atom steps form a superlattice.
This preparation should be done in a dust-free environment. Normally after
installation in the STM the Au surface will need to be ion bombarded with Ar+ and
then annealed in vacuum to present a clean surface (Fig. 16.1).
16.2 Single Crystal Orientation, Grinding, and Polishing
The preparation of oriented single crystal surfaces by X-ray diffraction followed by

grinding and polishing to a high degree of smoothness and flatness is of central
importance in surface science experiments, and there is a long interconnected
history of developments in this field [11–19]. At the present time, using the best
methods, an orientational accuracy of ±0.05° [5, 8] and a flatness of one-twentieth
the wavelength of sodium light [15] have been reported.
A brief history of the developments over almost 50 years follows. In 1949,
Walker et al. [11] described a cylindrical crystal holder that contained a crystal
holder platform that could be tilted so that the crystal could be crystallographically
oriented in a mounting attached to an X-ray diffraction apparatus. The holder then
could be transferred to cutting, grinding, and polishing apparatus without disturbing
the crystal alignment. The cylindrical crystal holder is remarkably similar to a holder
now marketed commercially as the Bond Barrell Holder [20]. In 1961, Bond [12]
described a jig that held a cylindrical crystal holder for grinding an oriented crystal
16.2 Single Crystal Orientation, Grinding, and Polishing 157
on a flat grinding plate, called a lapping plate, containing an abrasive powder.

Bond’s holder was attached to a hardened steel ring that made contact with the
lapping plate and provided a reference plane for the mechanical grinding and pol-
ishing processes. Bennett and Wilson [13], who used a grinding/polishing jig of
different design, were not interested in accurate crystal orientation, but showed that
lapping of hard crystals on soft metal plates containing embedded abrasive particles
could result in surface flatness of the order of one-eighth the wavelength of sodium
light. Butz et al. [15] described a goniometer-jig combination that was used for both
orientation and lapping of crystals to an accuracy of 0.25° and to a flatness of
one-twentieth the wavelength of sodium light. Berger et al. [17] described a
goniometer-jig combination that possessed 2° of crystal tilt freedom, plus transla-
tional and azimuthal orientation freedom. They ground and polished a W crystal on a
glass lap, with 15 μm A12O3, 5 μm Al2O3, and 6-μm and 3-μm diamond abrasive
paste. An interesting feature of this work was the production of a W crystal disk with
five desired facet-planes, which were convenient for plane-to-plane comparisons.
The accuracy of orienting and polishing single crystals has been brought to a very
high level by Linke and Poelsema [18], who showed by He diffraction methods that
Pt(lll) crystals oriented and polished as described previously [17] exhibited a mean
terrace width of 3000 Å and an orientational error of ±0.05°. The reader is referred to
the original papers for detailed designs of the goniometer-jig combinations.
Figure 16.2a shows a Bond Barrel Holder [20] mounted in a simple grinding-
polishing jig made of hardened steel [21]. The Pt crystal on the holder platform has
been crystallographically oriented using Laue X-ray diffraction, where the Bond
Barrel Holder was held in a mount that fitted onto an optical rail that was part of the
Laue apparatus. Typically, moving the Bond Barrel Holder from one mount to
another can be done with an error of less than 0.1°. The single crystal was ground
using a 600 grit abrasive paper mounted on a rotary wheel grinding and polishing
machine. The weight of the Bond Barrel Holder forces the crystal onto the abrasive
paper. The hardened steel jig is not appreciably affected by this process as should be
determined periodically using a micrometer to measure its thickness around its
circumference. When the crystal is flat, a series of polishing steps with smaller and
smaller abrasive particles is performed as shown in Fig. 16.2b. It is extremely
important to remove, by scrupulous washing and rinsing away from all parts of the
apparatus, the abrasive from each step, or scratches owing to inadvertent abrasive
particles will ruin the next steps in polishing with finer abrasive. The polishing cloth
should of course be changed each time as the abrasive particle size is reduced.
Figure 16.2c shows a measurement made by the observation of interference
fringes with an optical flat touching the finished crystal [21]. It was found, in
duplicate experiments, that the surface was not flat, but instead had a curved profile
with a typical radius from 4 to 2 m near the center, and a slope of from 0.05° to
0.08° at the edges. The limiting slope corresponds to one atomic step for about
every 1000 Pt atoms measured horizontally, and for many experiments this is
unimportant. Figure 16.2a shows a hypothesis about the origin of the slightly
curved figure of the crystal, where slight buckling of the flexible abrasive paper is
postulated to produce a wave that tends to round the crystal slightly. A small misfit
between the Bond Barrel Holder and the hardened steel jig can also result in a
curved crystal [22].
Grinding and polishing to make highly flat surfaces can be carried out most
effectively using rigid soft metal laps that are very flat and contain embedded
abrasive particles [13, 15, 17]. Such an arrangement is shown in Fig. 16.2d. A set of
laps, containing different particle size abrasive grains, is required, and
cross-contamination with other abrasive grain sizes is to be avoided. Brass and tin
are favorite lap materials. Grinding and polishing on laps has produced crystal
surfaces flat to one-twentieth the wavelength of sodium light or about 250 Å/cm,
which corresponds to only one atomic step for about 4 × 105 Pt atoms, measured
horizontally [15]. At this level of step density resulting from departures from pla-
narity, the step density owing to misalignment will be the dominant factor, being
about 2 orders of magnitude greater for a misalignment of 0.05°. A polishing jig
giving an accuracy of 0.02° in the orientation of a silicon single crystal has been
reported [23]. A method, involving the observation of a single Bragg reflection with
a diffractometer, and employing a crystal holder that can be rotated about its axis,
has been described, and the reported accuracy is better than 0.003° [24].
Combined mechanical and electrochemical polishing techniques are often used
and a description of one method for polishing copper single crystals and alloys is
given in [25] and [26].
Fig. 16.2 Single crystal orientation, grinding and polishing. a Typical crystal grinding. b Typical
crystal polishing. c Measured crystal curvature. d Crystal lapping for optimal flatness
16.2 Single Crystal Orientation, Grinding, and Polishing 159
A special holder for the grinding polishing jig, shown here, has been developed
which allows the device to be used in an automatic grinding-polishing wheel device
as is often employed to prepare metallurgical samples [27]. The holder uses ball
bearings inside a ring which make contact with the outside of the hardened steel shell
which makes contact with the polishing cloth and wheel. This ring-ball bearing
apparatus allows the jig to rotate, while applying downward and lateral force which
causes the polishing jig to translate and oscillate across the rotating wheel. Using this
jig, a spherical surface with 3.4 m radius was generated near the center of the crystal,
with surfaces fitting larger radii being produced near the crystal edge.
16.3 A Simple Goniometer for Cutting Single Crystals
A clever method for cutting single crystal boules along any plane involves a
goniometer of very simple design, as shown in Fig. 16.3 [28]. Here, a single crystal
boule is fastened into a holder with a crystal bond compound, using a plastic screw
that will not deform the crystal when tightened against its outer surface. The holder
is placed into a special wedge-shaped device with shoulders cut at angle α with
respect to the upper surface. This wedge-shaped device can rotate inside a tilted
cylinder with an upper surface inclined at angle a with respect to the base of the
device. When the two wedges are rotated relative to each other, the tilt angle of the
crystal, ω, will change.
The crystal is aligned by Laue back-reflection X-ray diffraction by manipulating
the wedged device. A rotation of 180° will result in swinging the crystal axis by a,
as shown in Fig. 16.3. Thus, if the saw cuts in a plane parallel to the baseplate of the
goniometer, the crystal may be cut at any angle over the range α. An engraved
angular scale is provided in two places as shown. The rotations are not simply
Fig. 16.3 Simple crystal cutting goniometer

related to the angles measured on the Greninger chart, used for X-ray diffraction
measurements, but the original paper gives a method to achieve the conversion
using a small programmable calculator [28].
16.4 Ion Beam Polishing of Crystals to Subnanometer

Roughness
Crystalline substrates can be polished to near atomic perfection by sputtering under

glancing incidence conditions [29–31] as shown in Fig. 16.4. If the glancing
incidence angle is chosen such that the surface normal component of the incident
beam kinetic energy is sufficiently low to ensure that the beam is reflected with unit
probability from atomically flat, defect-free terraces, sputtering only occurs at ad-
atom and step edge sites [30, 31]. Even at those defect sites, little damage to the
subsurface regions appears to occur. The smoothing technique can be combined
with annealing of the substrate, either simultaneously or after bombardment. By
monitoring the width and peak position of the angular distribution of the scattered
ion beam during sputtering, an indication of the surface quality and the change with
time is conveniently obtained. A narrow, near specularly scattered beam indicates a
low density of defects, on an atomic scale, over the length of the glancing incidence
trajectory at least.
Fig. 16.4 Ion-beam polishing of crystals to subnanometer roughness

16.5 Crystal Optical Alignment 161
16.5 Crystal Optical Alignment
The alignment of highly reflective single crystals by using simple optical methods is
commonly employed for the setting of vernier adjustments. In addition to the
optical methods, electron beams can also be used for crystal alignment, as shown on
pp. 253 and 255.
Figure 16.5a shows a very simple crystal alignment method [32], using only a
flashlight in a darkened room. An image of the pupil of the eye is reflected off of
the crystal and back into the eye. In addition, the image is also reflected from the
window to the eye. When the crystal surface is parallel to the window surface, the
two reflected images of the pupil will coincide.
Figure 16.5b shows the same method used outside the vacuum system to align a
crystal in its tilt angle so that the normal of the crystal is parallel to the manipulator
mounting flange face [32]. The manipulator is held in a carefully made holder that
is square. The image of the eye, illuminated by a flashlight filament, is reflected
from the crystal and passes back by a pointer. If the center of the image of the pupil
is coincident to the pointer when the pointer is translated along the table, then the
crystal normal is parallel to the tabletop and therefore to the manipulator reference
flange face.
Fig. 16.5 Crystal optical alignment. a Crystal alignment by eye. b Crystal alignment by eye on
table. c Crystal alignment by laser
Figure 16.5c shows the use of a He/Ne laser for crystal alignment. The laser is
moved until its reflection from the window is aligned so that the laser beam returns
to the laser as an aligned beam coincident with the incident beam. The crystal is
then manipulated in the vacuum system until the reflection of the laser beam from
the crystal is also coincident. In this case, then, the normal to the crystal and the
normal to the window will be parallel. The accuracy of the laser alignment pro-
cedure will increase as the optical path is increased, and the accuracy of coalign-
ment can easily be made to be 0.1°, or better. A partially reflecting glass slide may
be used to deflect a fraction of the two reflected laser beams to a card located off the
laser axis. The slight divergence of the laser ultimately limits the accuracy of the
alignment procedures.
It should be pointed out that the establishment of laboratory coordinates inside
an ultrahigh vacuum system, where reference is made to flange alignments, maybe
in error because of machining errors and welding errors made in the fabrication of
the chamber, as well as to distortion of the chamber which occurs both on welding
and due to the atmospheric pressure.
16.6 Measurement of the Angles of Incidence in a LEED

Experiment
It is often necessary to determine the angle of incidence of an electron beam in a

low energy electron diffraction (LEED) apparatus relative to the crystal being
studied. This can be done qualitatively using settings on the crystal manipulator
goniometers or by simple measurements of the LEED beam pattern on the screen.
However, for accurate work, it should be possible to use all or almost all of the
LEED beam positions to do this with high accuracy, exceeding the accuracy of the
average goniometer, which has an uncertain relationship to the crystal connected to
it. The method to be described [33] outperforms earlier methods involving the use
of an internal graduated circle and vernier [34], laser calibration of the manipulator
outside of the vacuum system, or measurement of the 0, 0 beam position using a
Faraday cup on a vernier circle.
Figure 16.6 shows a LEED screen with its axial electron gun, and the definition
of the laboratory directions, xl, yl, and zl. Similar angles relative to a crystallo-
graphic direction on the crystal are given by xc, yc, and zc. The azimuthal angle is Ψ.
The polar angle is given by Θ. The angle ε is the measured angle between the 0, 0
beam and an h, k beam.
By examination of a LEED photograph containing multiple LEED beams and
making a series of angular measurements using a transparent angular scale made
from circular graph paper, it is possible to map out many angles which permit the
accurate determination of Θ and Ψ. The computer program for doing this is
available from the authors [33], and accuracies of the order of 0.1° are achieved
owing to the use of multiple angular measurements.
16.7 Measurement of Angle of Incidence of an Electron Beam on a Single Crystal 163
Fig. 16.6 Measuring angle of incidence in LEED
16.7 Measurement of Angle of Incidence of an Electron

Beam on a Single Crystal
A simple method for measuring the angle of incidence of an electron beam on a

single crystal that involves only the measurement of the collected current is
described in Fig. 16.6.
When an electron beam of fixed kinetic energy is received by a single crystal, the
reflectivity of the electrons will fall off as the angle of incidence is changed. By
differentiating the collected current with angle, plots that are symmetrical about the
normal will be obtained, provided the plane of the angular scan is perpendicular to a
mirror plane of the crystal. This behavior is a consequence of the band structure of
the crystal, and the more free-electron-like a band is, the higher the electron cou-
pling probability is to this band. Inside the band gap region the reflectivity of
electrons with low energies can reach values close to unity.
The apparatus shown in Fig. 16.7a shows the electron gun and the electrical
circuit used for differentiating the collected current at the crystal. The measurements
Fig. 16.7 Measuring angle of incidence of electrons. a Electron gun + circuit. b First derivative—
collected current
16.7 Measurement of Angle of Incidence of an Electron Beam on a Single Crystal 165
were taken in the ΓXWK mirror plane of the Cu(100) crystal, perpendicular to
another ΓXWK mirror plane. The angle of incidence was varied by rotating the
electron gun around the crystal. The data shown in Fig. 16.7b clearly show
the symmetry with respect to the normal incidence angle for two electron energies.
The symmetry indicates the absence of external magnetic fields and the absence of
stray electrical fields. It is estimated that the position of the normal can be estimated to
±0.5°, although the features in the wings differ in their maxima in the two directions
by several degrees. The procedure can be used to determine the azimuthal orientation
of the crystal by scanning the azimuthal angle. For off-normal angles of incidence,
extrema should be observed when crossing a mirror plane of the crystal [35].
References
1. P. Maksymovych, private communication

2. P. Maksymovych, D.C. Sorescu, D. Dougherty, J.T. Yates Jr, J. Phys. Chem. B 109, 22463
(2005)
3. P. Maksymovych, J.T. Yates Jr, Chem. Phys. Lett. 421, 473 (2006)
4. P. Maksymovych, J.T. Yates Jr, J. Am. Chem. Soc. 128, 10642 (2006)
5. P. Maksymovych, D.C. Sorescu, J.T. Yates Jr, Phys. Rev. Lett. 97, 146103 (2006)
6. P. Maksymovych, D.C. Sorescu, J.T. Yates Jr, J. Phys. Chem. B 110, 21161 (2006)
7. P. Maksymovych, D. Dougherty, X.-Y. Zhu, J.T. Yates Jr, Phys. Rev. Lett. 99, 016101 (2007)
8. P. Maksymovych, J.T. Yates Jr, J. Am. Chem. Soc. 130, 7518 (2008)
9. P. Maksymovych, D. Sorescu, K. Jordan, J.T. Yates Jr, Science 322, 1664 (2008)
10. P. Masksymovych, O. Voznyy, D. Dougherty, D. Sorescu, J.T. Yates Jr, Prog. Surf. Sci. 85,
2906 (2010)
11. J.G. Walker II, J. Williams, R.M. Bozorth, Rev. Sci. Instrum. 20, 947 (1949)
12. W.L. Bond, J. Sci. Instrum. 38, 63 (1961)
13. G.A. Bennett, R.B. Wilson, J. Sci. Instrum. 43, 669 (1966)
14. D.E. Johnson, Rev. Sci. Instrum. 39, 238 (1968)
15. R. Butz, B. Krahl-Urban, K. Mench, Rev. Sci. Instrum. 44, 485 (1973)
16. J.F. Wendelken, S.P. Withrow, C.A. Foster, Rev. Sci. Instrum. 48, 1215 (1977)
17. S. Berger, U. Linke, J. Phys. E: Sci. Instrum. 11, 1097 (1978)
18. U. Linke, B. Poelsema, J. Phys. E: Sci. Instrum. 18, 26 (1985)
19. Y.K. Chang, Rev. Sci. Instrum. 53, 515 (1982)
20. Bond Barrel Holders may be purchased from Blake Industries, Inc., 660 Jerusalem Road,
Scotch Plains, NJ 07076
21. C.E. Tripa, J.T. Yates Jr, J. Vac. Sci. Technol. A14, 2544 (1996)
22. Dr. W.F. Egelhoff, Jr., NIST, Gaithersburg, MD 20899 (private communication)
23. S.I. Mathiesen, L. Gerward, O. Pedersen, Rev. Sci. Instrum. 45, 278 (1974)
24. M. Deutsch, I. Freund, Rev. Sci. Instrum. 46, 765 (1975)
25. J.S. Aherarn Jr, J.P. Monaghan Jr, J.W. Mitchell, Rev. Sci. Instrum. 41, 1853 (1970)
26. T.W. Snouse, Rev. Sci. Instrum. 36, 866 (1965)
27. P. Maksymovych, T. Gasmire, S. Ohno, J.T. Yates Jr, J. Vac. Sci. Technol. A23, 362 (2005)
28. G.L. Price, Rev. Sci. Instrum. 51, 1000 (1980)
29. Dr. K. J. Snowdon, Department of Physics, University of Newcastle, New Castle upon Tyne,
NE 17RU, UK (private communication). The first report of this effect is in U. von Gemmingen
and R. Sizmann, Surf. Sci. 114, 445 (1982)
30. M. Holzwarth, M. Wissing, D.S. Simeonova, S. Tzanev, K.J. Snowdon, and O.I. Yordanov,
Surf. Sci. 331-333, 1093 (1995). Here polycrystalline Cr films were smoothed with a reduction
in the rms roughness by a factor of 5 to 8 from 9.2 nm, as measured by the STM
31. M. Wissing, M. Batzill and K. Snowdon, Nanotechnology 8, 40 (1997). Here, CaF2 single
crystal surfaces were smoothed from an rms roughness of 0.6 nm to better than 0.1 nm (the
AFM noise level)
32. Dr. Oliver Schaff, Fritz Haber Institut der Max-Planck Gesellschaft, Faradayweg 4-6, D-14195
Berlin (private communication)
33. S.L. Cunningham, W.H. Weinberg, Rev. Sci. Instrum. 49, 752 (1978)
34. J.M. Burkstrand, Rev. Sci. Instrum. 44, 774 (1973)
35. R. Schneider, H. Durr, Th Fauster, V. Dose, J. Vac. Sci. Technol. A8, 3363 (1990)
Chapter 17
Tip Fabrication
17.1 Fabrication of Metal Tips Using Zone

Electropolishing
Metal tips are required in field electron and field ion microscopy, as well as in
scanning tunneling microscopy. One method for preparing such tips is to use zone
electropolishing while viewing the process under an optical microscope [1–3].
Figure 17.1 shows the setup employed for both wire samples and for irregular
specimens such as thin films stripped from their substrates or slivers of metal cut
from a sheet. The electrolyte is held on a loop by surface tension, and its volume
can be adjusted upward by transfer from a wet glass rod or downward by absorption
by a piece of tissue paper. This loop can be moved laterally using an x, y,
z micromanipulator as shown, and the specimen to be electropolished is attached
mechanically and conductively to a specimen holder. For a wire specimen, selected
regions of the tip may be electropolished by lateral motion of the loop to the desired
region as shown in Fig. 17.1, inset A. Very sharp tips may be produced by first
electropolishing a neck behind the tip apex and then reducing the necked-down
region to a sharp point in an electrolyte film of about 0.2-mm thickness. Additional
sharpening can be achieved by quick thrusts of the electrolyte film onto and away
from the tip.
Irregular specimens can be shaped as shown in Fig. 17.1, inset B. Here, working
on the sides of the sample, etching is achieved to bring the sample to a tip shape
prior to sharpening the tip as shown in Fig. 17.1, inset A.
Either AC or DC electropolishing can be used. In addition, instead of using a
micromanipulator, the lateral stage motion of the specimen holder or the stage of
the microscope can be used with the sample being held rigidly in place.
The microscope magnification is either ×l50 or ×600 with a long working dis-
tance objective. Note the protective coverglass for the objective to prevent etching
of the glass lens by droplets of electrolyte produced during electropolishing.

DOI 10.1007/978-3-319-17668-0_17
168 17 Tip Fabrication
Fig. 17.1 Etching Field Ion Microscopy Tips
Special considerations in tip construction apply for tips to be used under elec-
trolytes, where Faradaic currents may damage the tip. This damage problem may be
eliminated by the use of Apiezon wax to shield the tip shank [4].
An excellent review on methods for sharpening STM tips is given in [5].
The use of AC waveform control in shaping tips for field ion microscopy has
been described [6, 7]. Electrochemical etching of Ag tips has recently been
described, using electronic control of the tip movement out of the electrolyte [8, 9].
A review comparing the instrumental characteristics of various STM [10].
17.2 Etching STM Tips Reproducibly
A simple apparatus [11] for making a number of STM tips while automatically
replenishing the electrolyte is shown in Fig. 17.2. Section (a) shows a three-hole
electrolyte support in which thin films of the electrolyte are held by surface tension
effects. The central hole of 3-mm diameter is connected to smaller 2-mm diameter
supply and withdrawal reservoirs in a 1-mm-thick Cu plate. Section (b) shows the
beveled end of a (1-mm outside diameter, 0.45-mm inside diameter) supply tube for
fresh electrolyte being fed into the supply reservoir. A similar tube is used to
withdraw electrolyte from the other reservoir. The wire to be etched enters the
17.2 Etching STM Tips Reproducibly 169
Fig. 17.2 Etching STM Tips Reproducibly
central hole, and projects below the bottom of the electrolyte film. This permits the
etched tip to fall away under gravity when the etching is complete, and it is believed
that as the lower tip falls away the local strain is likely to produce an atomic
asperity [12] for STM imaging. Section (c) shows the application of AC current to
the etching apparatus, and the profile of a tip just being started. The thin neck below
the desired tip location is useful for making a detachable handle with which to hold
the tip, and which can be broken off later.
The flow rate of electrolyte and the configuration of the meniscus can be
adjusted by changing the level of electrolyte storage beakers on labjack platforms
that connect to the supply and withdrawal reservoirs. Tubing 30 cm in length is

approximately correct. For an electrolyte solution of KOH(aq), 4 wt%, a flow rate
of about 1 ml/h is correct for an etching current of 20 mA. A slightly concave upper
and lower surface in the etching cell is preferable, achieved by fine height
adjustments of the two beakers. Tips have had a wire diameter as small as 0.08 mm.
Many tips can be produced in succession by this method. It is helpful to use a wax
pencil to outline the three-hole device on both sides to prevent wetting and elec-
trolyte creep across the horizontal surfaces. Another design for tip etching is shown
on page 258 as well as others in [12, 13]. The use of Apiezon wax for protection of
tip shanks used in electrolytes from Faradaic current effects is discussed in [13].
A general review of the performance of various STM instruments is given in [14],
and a review of tip-making for field ion microscopy tips is given in [15]. An Auger
spectroscopy study of the surface composition of etched tungsten tips is given in
[16]. It was found that a 10-nm layer of tungsten oxide is mostly removed by
heating to about 1800 K. The tungsten oxide film from an electrochemically etched
W tip can also be removed by treatment in concentrated HF(aq) [17].
17.3 STM Tip Flasher
An Omicron STM tip carrier has been converted to a thermal tip flasher. In the
design shown in Fig. 17.3, a Ta sheet, 0.125 mm thick, is fabricated with a key hole
slot. This sheet, mounted by three holes which correspond in position to the holes in
the tip carrier body, is electrically isolated from the supporting bolts by use of
concentric ceramic tubes. The sheet is slightly bent as shown to be able to make
contact with the tip at some point when the tip is magnetically attracted to a
permanent magnet. Current is passed between the Ta sheet and the bottom of the
carrier. Nominal flashing conditions are: 7 A × 3.5 V for 10 s, and outgassing
causing <5 × 10−10 mbar pressure rise is reported [18]. A similar device was
recently described [19].
17.4 Reverse Bias Method for Sharp Tip Etching
Many methods have been devised for making sharp tungsten tips for STM and other
applications, and some of these are described in Sects. 17.1–17.3. It has been
discovered that tip blunting can occur during the last stages of tip etching [20]. The
method described here eliminates this blunting effect by control of the electro-
chemical potential during the moments just after tip separation in the etching
solution.
Figure 17.4a shows a typical etching apparatus for making tungsten tips. Using
2 M KOH (aq) as electrolyte, suspended in a lens shaped droplet in an Ir loop, the
17.4 Reverse Bias Method for Sharp Tip Etching 171
Fig. 17.3 STM Tip Flasher outgassing, tip flasher
tip is roughly etched to a point where the thinning of the neck is readily visible. The
etching is then continued more slowly in a fine etching process with 0.1 M KOH
(aq) solution until gravity separates the parts. At this point the etching potential is
changed from V1 to V2 as shown in Fig. 17.4b. If the power supply is a normal dc
supply, the applied potential can be dropped to zero at the moment of separation,
but it has been observed that there is still a natural electrical potential which exists
between the surface and the tip and the Ir ring, of order 0.5 V, and that charge will
still flow at currents in the 10−7–10−8 A range. This natural current causes blunting
of the tip. Using a power supply designed to apply the reverse potential from that
used for V1, the blunting effect can be eliminated. The reverse bias is applied within
500 ns of the separation time by electronic detection of the moment of separation
which triggers the application of the reverse bias. Figure 17.4c shows the measured
tip radii versus post etching time at the natural potential compared to experiments
where V2 is applied as a reverse bias, eliminating etching. Figure 17.4d shows an
electron micrograph of a tip etched with the reverse bias method, where a 4 nm
radius is achieved. The blunting is thought to be due to the recession of the tip,
Fig. 17.4 Reverse bias method for sharp tip etching. a STM tip machine. b Current versus time
during etching c. Tip radius versus post-etching time. d STM tip etching to r ˂ 4 nm by Reverse
Bias Technique. e Tip configuration if reverse bias is not used
during the post-separation etching, and the radius will be dependent on the cone
angle achieved. A schematic of the blunting is shown for times increasing from 1 to
5 in Fig. 17.4e after the separation.
17.5 Sharpening Single Crystal Metal Tips by Ion Bombardment 173
17.5 Sharpening Single Crystal Metal Tips by Ion

Bombardment
Both tungsten and molybdenum single crystal wires can be sharpened into tips by
using axial sputtering. The data shown here are for Mo(110) tips [21], but similar
results are reported for W(111)-oriented wires [22], except in that work the ion
beam direction was from behind the tip, producing very sharp pyramidal shapes.
The Mo(l10) crystal wire is mounted such that the axis of the wire is parallel to
the ion beam, as shown in Fig. 17.5a. In addition, because of variation in ion current
density across the ion beam, the tip should be centered in the beam to achieve
symmetrical sputter etching. The ion beam in this case is 350 μm in diameter. It is
found that Xe+ ions are more effective in making less roughened tips; Ar+ at low
kinetic energies can produce very sharp asperities also. High-energy Ar+ sputtering
produces increased surface roughness compared to Xe+.
Fig. 17.5 Sharpening tips by ion bombardment. a Apparatus. b Tip profiles by TEM
The profiles of tips produced both by 3 kV Xe+ and 0.5 kV Ar+ sputtering are
shown in Fig. 17.5b. It is estimated that the central asperity on the Ar+ sputtered tip
has a base diameter of about 10 Å.
Ion sputtering for sharpening field electron and field ion tips has been described
[22, 23].
Polycrystalline wires cannot be shaped so well by this technique because of the
presence of grain boundaries from which preferential sputter etching occurs [23].
However, single crystal tungsten tips have been ion milled to a point that is sharp
on the atomic scale [24, 25].
A recent example of the use of a focused ion beam of Ga+ ions for ion milling W
and Pt/Ir tips is given in [26], which also summarizes much other work in this area.
Copper wires of 30–50 μm diameter have been ion etched into sharp tips, and the
dynamics of the process have been studied [27].
References
1. A.J. Melmed and J.J. Carroll, J. Vac. Sci. Technol. A2, 1388 (1984)
2. A.J. Melmed, J. Vac. Sci. Technol. B9, 601 (1991)
3. H. Lemke, T. Goddenhenrich, H.P. Bochem, U. Hartmann, C. Heiden, Rev. Sci. Instrum. 61,
2538 (1990)
4. L.A. Nagahara, T. Thundat, S.M. Lindsay, Rev. Sci. Instrum. 60, 3128 (1989)
5. M. Fotino, Rev. Sci. Instrum. 65, 159 (1993)
6. H. Morikawa, K. Goto, Rev. Sci. Instrum. 59, 2195 (1988)
7. H. Bourque, R.M. Leblanc, Rev. Sci. Instrum. 66, 2695 (1995)
8. K. Dickmann, F. Demming, J. Jersch, Rev. Sci. Instrum. 67, 845 (1996)
9. A.A. Gorbunov, B. Wolf, Rev. Sci. Instrum. 64, 2393 (1993)
10. Y. Kuk, P.J. Silverman, Rev. Sci. Instrum. 60, 165 (1989)
11. J.O. Fiering, F.M. Ellis, Rev. Sci. Instrum. 61, 3911 (1990)
12. P.J. Bryant, H.S. Kim, Y.C. Zheng, R. Yang, Rev. Sci. Instrum. 58, 1115 (1987)
13. L.A. Nagahara, T. Thundat, S.M. Lindsay, Rev. Sci. Instrum. 60, 3128 (1989)
14. Y. Kuk, P.J. Silverman, Rev. Sci. Instrum. 60, 165 (1989)
15. E.W. Muller, T.T. Tsong, Field Ion Microscopy (Elsevier, New York, 1969)
16. A. Cricenti, E. Paparazzo, M.A. Scarselli, L. Moretto, G. Selci, Rev. Sci. Instrum. 65, 1558
(1994)
17. L.A. Hockett, S.E. Creager, Rev. Sci. Instrum. 64, 263 (1993)
18. B. Unal, A. Belianinov and P.A. Thiel, Private Communication
19. F. Bastiman, A.G. Cullis, M. Hopkinson, K.J. Briston, J. Vac. Sci. Technol. B28, 371 (2010)
20. O. Guise, J.W. Ahner, M.-C. Jung, P.C. Goughnour, J.T. Yates Jr, Nano Lett. 2, 191 (2002)
21. S. Morishita, F. Okuyama, J. Vac. Sci. Technol. A9, 167 (1991)
22. J.M. Walls, H.N. Southworth, G.J. Rushton, Vacuum 24, 475 (1974)
23. J.A. Kubby, B.M. Siegel, J. Vac. Sci. Technol. B4, 120 (1986)
24. D.K. Biegelsen, F.A. Ponce, and J.C. Tramontane Appl. Phys. Lett. 54, 1223 (1989)
25. D.K. Biegelsen, F.A. Ponce, J.C. Tramontna, S.M. Koch, Appl. Phys. Lett. 50, 696 (1987)
26. M.J. Vasile, D.A. Grigg, J.E. Griffith, EA. Fitzgerald, and RE. Russell, Rev. Sci. Instrum. 62,
2167 (1992)
27. H. Hasuyama, Y. Kanda, T. Soeda, K. Niiya, M. Kimura, Matls. Sci. and Engr. 90, 33 (1987)
Chapter 18
Spot-Welding
18.1 Spot-Welding Difficult Junctions
Many metal/metal junctions to be spot-welded are difficult to fabricate as a result of

failure to understand the nature of the spot-welding process. A spot-weld is made
when an electrical spark causes two surfaces in contact to fuse together as a result of
the high local heating. Ideally, the surfaces to be welded should make only a point
contact in order to focus the local heating at a single point. Thus, as shown in
Fig. 18.1a, two cylindrical wires, placed orthogonally to each other, will theoreti-
cally contact each other only at one point, leading to efficient welding. In contrast,
two wires oriented parallel to each other will contact along a line, making a broad
contact and leading to a weak weld.
Figure 18.1b shows a way to weld a thermocouple or other wire to a metal single
crystal. The crystal is placed on a rigid, clean, smooth surface that is a noncon-
ductor. A single spot welding electrode presses the wire against the back of the
crystal, and a second electrode also presses on the back of the crystal to make
electrical continuity.
Often it is found that the welding leads stick to the weld when it is completed.
This may be due to welding electrodes that have not been smoothed with emery
paper before welding, or the application of too much power to make the weld.
Sometimes sticking may be prevented by coating the welding electrode with carbon
black from a match or candle flame, as shown in Fig. 18.1c.
Oxidation during the spot weld is often a problem owing to local heating. This
may be prevented by welding under a droplet of methanol (Fig. 18.1d), which acts
as a shield from the air and is also a reducing agent in the presence of the hot metal
at the weld point.
In cases where a good weld is difficult to make, a sandwich weld is sometimes
used as an intermediate material between two refractory metals. Tantalum and
platinum foil are often good for this, as shown in Fig. 18.1e [1]. It is also reported

DOI 10.1007/978-3-319-17668-0_18
176 18 Spot-Welding
Fig. 18.1 Spot-welding difficult junctions. a Point contact weld. b Welding to single crystal.
c Preventing electrode sticking. d Welding under methanol. e Sandwich weld between refractory
metals. f In situ electropolishing+welding
that changing the welding tip polarity sometimes helps in making a difficult weld,
although the reasons for this are not understood [2].
It is very important to only try to weld metals when the surfaces have been
freshly cleaned. Thus, W and Mo are cleaned by electropolishing in NaOH(aq)
solution, about 1 M in concentration. The surface to be welded is positive during
electropolishing, or in some cases AC electropolishing may be used.
Figure 18.1f shows a clever way to electropolish the metal and then to use the
elec-tropolish solution as a protective boundary from air oxidation. Here, a Mo
single crystal is to be welded to a Mo wire at the crystal edge. The NaOH(aq)
solution in a droplet is placed on the crystal, and electropolishing of the edge is
carried out using a Mo wire electrode in series with a light bulb and a transformer.
The light bulb supplies a protective load if a short circuit to the crystal occurs. Once
the electropolishing is complete, the droplet of solution is left in place, and the
spot-weld of a We/Re thermocouple in this case is made under the electrolyte [3].
After spot-welding, the droplet of electrolyte is pipetted away, and the assembly is
washed well with distilled water. It was noted that lower power was needed to make
effective spot-welds in this manner [3].
18.1 Spot-Welding Difficult Junctions 177
It is important to make many practice spot-welds before assembling the desired

object. With practice, a careful welder can make ten successful practice welds in
series, therefore almost guaranteeing that the eleventh will be good. Adjustments of
the power per weld, the pressure used at the weld, and the composition of the
welding tips must be made to achieve a good weld. Once the proper conditions are
found, a single weld is all that is needed, and in fact multiple welds at the same
point can result in metal work hardening and weakening and should be avoided.
Metals with very high thermal conductivity (Al, Ag, Cu) are difficult to weld, and
slotted crystals of those metals are often used where the slot accommodates a
refractory heating/support wire [4]. It is reported that a Pt sandwich will facilitate
refractory metal spot-welds to Cu [2].
References
1. D.C. Carlyle, Rev. Sci. Instrum. 41, 599 (1970)

2. W.R Egelhoff, Jr., National Institute of Standards and Technology, Gaithersburg, MD
3. M.E. Jones, B.E. Koel, R.T. Weppner, Rev. Sci. Instrum. 60, 3067 (1989)
4. J.E. Crowell, J.G. Chen, J.T. Yates Jr, Surf. Sci. 165, 37 (1986)
Chapter 19
Mechanical Procedures
19.1 Stretching Thin Metal Foils to a Wrinkle-Free

Condition
In some cases, the provision of wrinkle-free metal foil for windows may be needed.
A very simple method, related to the differential thermal expansion of metal and
Pyrex glass, may be employed to achieve this objective [1].
Figure 19.1a shows a short cylinder of fused quartz or Pyrex glass ground flat on
its end, mounted on top of the wavy foil. The thin foil is supported on its bottom by
a horizontal substrate such as a lapped flat plate of aluminum, coated with a
fluorocarbon dry lubricant [2]. The foil is attached to the plate at the edges with an
adhesive tape to prevent curling. An epoxy adhesive (Torr seal [3]) is applied to the
lower edge of the glass tube and the tube is placed on the wrinkled foil. The
assembly is heated in an oven for several hours at about 60 °C, and then allowed to
slowly cool in the oven overnight. When the assembly is heated, the areal expan-
sion of the metal foil exceeds that of the glass cylinder due to the thermal expansion
coefficient differences. The light weight glass cylinder does not impede the foil
expansion, and wrinkling does not occur in the center of the glass form. Upon
cooling, the foil contracts more than the glass, resulting in a uniformly stretched foil
as shown in Fig. 19.1b. This technique can be used for stretching foil over an
aperture that cannot be heated; the glass form is larger than the aperture, and the
aperture is attached by cement to the foil, previously stretched as shown. A simple
consideration of the strain effects in the circumferential and the radial directions
indicates that for the circular geometry of the glass cylinder, the strain in both
directions is equal, meaning that no shear stress takes place on cooling. This leads
to a uniform tension and a flattening of the foil [1]. HAVAR foil (0.005 mm thick)
[4], aluminum foil (0.025 mm thick) and Mylar foil (0.025 mm thick) have all been
stretched flat in this manner.

DOI 10.1007/978-3-319-17668-0_19
180 19 Mechanical Procedures
Fig. 19.1 Stretching metal foils. a Before assembly. b After cure and cooling
19.2 Making Micron-Size Holes for Supersonic Nozzles
Two methods are outlined for producing micron-sized holes in the ends of tubes to
be used as supersonic jet sources in vacuum [5].
19.2.1 Method I—Fig. 19.2a
A tube of the appropriate length and inner and outer diameter is made with only one
end open. An inner rod, which fits into the tube, is made with its length just slightly
longer than the inside dimension of the tube. This rod is sharpened on one end, and
19.2 Making Micron-Size Holes for Supersonic Nozzles 181
Fig. 19.2 Making micron-size holes for supersonic nozzles. a Method I. b Method II
a cap is machined on its upper end. By varying the cone angle of the sharpened tip,
the hole size can be controlled. The sharpened rod is then punched with a hammer
until the cap is tightly in contact with the upper tube edge. This should produce a
cone-shaped indentation in the end of the tube as shown. The outer surface of the
tube is then sanded down carefully until a hole just begins to appear. A microscope
is used to judge the hole size using both reflected and transmitted light [6].
182 19 Mechanical Procedures
19.2.2 Method II—Fig. 19.2b
A tube of the appropriate length and inner and outer diameter is made with only one
end open. The closed end is cut down with a conical surface on a lathe until a tiny
hole appears on the end. A punch die with the same cone angle as the tube end is
made. The punch die is then hammered on top of the tube end to close down the
hole until the desired size is obtained. A microscope is used to monitor the changes
in the hole size using both reflected and transmitted light [7].
Fig. 19.3 Straightening and braiding wires. a Stretching a wire to straighten. b Making uniform
braided wire
19.3 Straightening and Braiding Wires 183
19.3 Straightening and Braiding Wires
It is often advisable to work with very straight segments of wire, for example, in
placing wires along the outside of a manipulator shaft (see p. 67), where bends
might interfere with the linear construction or where a tight fit through tubulations
is needed. Figure 19.3a shows a method which works for nonrefractory metal wires
such as-Cu, Ni, Ag, and stainless steel [8]. The end of the wire is rigidly clamped in
a vise, and with a pair of pliers, the wire is stretched only a fraction of a percent.
The elastic limit should not be exceeded, and one can feel the give in the wire as
stretching occurs below the elastic limit. It is wise to have prepared a place to store
the straightened wire and a storage tube is recommended for this. After straight-
ening, many wires become work hardened so that further straightening by this
method will not work unless the wire is annealed.
Thin refractory wire will not straighten by this method, and another method must
be used [8]. Lay the wire on a paper tablet and stroke the wire with a round tool at
right angles to the wire axis, keeping the arch upward by rotating the wire during
straightening. A pair of smooth jaw pliers will hold the wire in the desired position.
Work on only a few inches at a time.
Braided wire may be conveniently produced in a uniform bundle by placing the
wires together and holding one end in a vise. A hand-cranked drill is chucked to the
other end of the wire bundle and turned slowly to make a uniform braid, as shown
in Fig. 19.3b.
References
1. W.H. Holt, C.M. Blackmon, Rev. Sci. Instrum. 51, 671 (1980)
2. Obtained from Miller-Stephenson Chemical Co., Inc., George Washington Highway, Danbury,
CT 06810
3. Obtained from Varian Vacuum Products, 121 Hartwell Avenue, Lexington, MA 02173
4. Obtained from Hamilton Technology, Inc., 101 N. Queen Street, P.O. Box 4787, Lancaster, PA
17604
5. Professor Thomas Engel, Department of Chemistry, Box 351700, University of Washington,
Seattle, WA 98195-1700 (private communication). The ideas are not original from the Engel
group and are believed to originate from Dr. Adi Scheidemann, from Peter Toennies’ group,
Göttingen University, Göttingen, Germany
6. From Steve Moskovits, T. Engel group, University of Washington
7. From B. Holm, machinist, University of Washington
8. G.L. Fowler, 6816 Kelly Ave., N.E., Albuquerque, NM 87109 (private communication)
Part III
Measurement Methods
Chapter 20
Electrons
20.1 Electron Gun Design and Behavior
Many descriptions of the design and properties of electron guns are presented in the
literature [1–5] and it is not the purpose of this treatment to summarize this prac-
tically important field. Instead, an ordinary electron gun [6] is described to give the
reader a beginning picture of this technology [7].
Figure 20.1a shows a drawing of the arrangement of elements in an electron gun
like those used in LEED and other surface science probes. It consists of a therm-
ionic emitter filament or ribbon, usually coated with a material that lowers the work
function, yielding higher operating currents at lower filament temperatures. For this
directly heated tungsten cathode, either thoriated tungsten, or ThO2, or LaB6
coatings can be used, as discussed on pp. 206 and 210. Some electron guns, such as
the widely used Varian Model 981-2125, have the thermionic emitter mounted off
axis to avoid line-of-sight exposure to the crystal reducing the possibility of con-
tamination as well as reducing light output. The thermionic cathode is housed in a
Wehnelt cylinder that surrounds it except for the beam extraction aperture. The
Wehnelt cylinder can be biased either positively or negatively, and is usually kept
near the potential of the cathode. This may be done by placing a resistor between
the cathode and the Wehnelt, which results in self-biasing of the Wehnelt; a typical
value of 570 kΩ will produce a small negative bias of the cylinder relative to the
cathode and will maintain beam focus over a several hundred volt range. The
electron beam is extracted by the first lens, G1, which then focuses the initially
divergent beam. This lens forms an image of either the emitter or the aperture of the
Wehnelt cylinder, depending on geometrical factors and the potentials applied. The
focal length is of the order of 10–30 cm. The drift tube, the lens G3, and the exit
aperture form an Einzel lens that focuses the beam. The beam diameter is typically
about 1 mm, and this can be reduced by decreasing the first two aperture sizes
(0.4-mm apertures in guns working at 100 eV energy typically deliver 1 μA cur-
rent). For 0.1-μA beam currents, a minimum beam size of 0.2 mm and a divergence

DOI 10.1007/978-3-319-17668-0_20
188 20 Electrons
of 0.2° are possible. Changing the bias on the Wehnelt cylinder with respect to the
cathode drastically affects the beam current and its stability.
Figure 20.1b shows an experimental plot of the electron beam width for the
electron gun in Fig. 20.1a, c shows the beam divergence versus electron energy.
These properties were measured with the multiple Faraday cup assembly and
procedure shown on p. 198.
Modeling of the behavior of an electron gun with cylindrical symmetry such as
this one can be carried out using an electrostatic computer program that is readily
available [8].
High-quality electron guns for high-resolution LEED measurements have been
described and often involve field emission electron sources [9, 10].
A high-resolution low-energy electron source involving a pair of 127° cylin-
drical selectors and delivering 3 × 10−8 A with 30-meV resolution has been
described [11].
Fig. 20.1 Electron gun properties. a Ordinary electron gun. b Beam width. c Beam divergence
20.1 Electron Gun Design and Behavior 189
A simple technique for precise regulation of the emission current from an

electron gun has been published. It uses the returning ground current which is
matched to a reference current. The difference is optically coupled to control the
filament to Wehnelt cylinder potential allowing beam currents of 0.1 μA to be
regulated to 1 % [12].
A pulsed electron gun with picosecond pulses has recently been described. The
pulse is generated by a pulsed laser that produces photoemitted electrons [13].

Irradiation
Electron sources for irradiation of cm2 area surfaces often employ stray electrons
that originate from thermionic emitters in mass spectrometers and ionization gau-
ges, or from bare filaments some distance from the target. While this is often
satisfactory, without adequate modeling of the distribution of electrons originating
from these sources, one may be dealing with non-uniform current densities on the
surface being irradiated. Extended electron beams are employed in cases where
surface analysis over a broad target area is to be done following damage to the
surface by electron bombardment.
A simple electron gun [14] designed to give a uniform current density across an
extended surface is shown in Fig. 20.2a. The gun contains a thoriated tungsten
filament (0.019-cm diameter) (see p. 206) which is hidden from line-of-sight to the
target by an L-shaped shield. Electrons are accelerated through a mesh grid (80 %
transparency) and then decelerated on the way to the target. The SIMION elec-
trostatic lens and analysis program [15] was used to model the behavior of the
electron gun as a function of various potentials, and Fig. 20.2b shows some electron
trajectories calculated for various electron energies. A fairly uniform flux in the
vertical plane is observed for electron energies above 2 eV.
Because of end effects on the thoriated emitter [16], to achieve a uniform current
density laterally, it would be wise to employ a gun width that is perhaps two times
larger than the target width, as was done here.
The current collected by the mesh grid and the target is shown in Fig. 20.2c.
Currents at the crystal in the 2–20 μA range are measured for electron energies in
the 2–31 eV range. Even at the lowest electron energies, about 1 in 5 electrons
emitted are collected by the target for mesh potentials above 10 V.
190 20 Electrons
Fig. 20.2 Low-energy/broad-beam electron gun. a Electron gun design. b Electron trajectory
analysis. c Current/voltage performance

Irradiation—Cylindrical Symmetry
The electron gun described here provides a broad, cylindrically symmetric electron
beam, as opposed to a simpler gun described on p. 290, which has only planar
symmetry. Low-energy electron beams produced by guns of these designs are used
to cause surface chemical reactions by controlled electron beam damage, where the
entire surface of a macroscopic sample is to be irradiated with uniform current
density. Since electron-induced surface reaction is expected to happen at low
electron energies, special design provisions must be made for delivery of high
current densities at low energies.
20.3 Low-Energy Electron Gun for Broad-Beam Irradiation—Cylindrical Symmetry 191
Figure 20.3a shows a cylindrical gun design [17]. A BaO-coated disk cathode
[18] is mounted inside a cylindrical repeller electrode. An anode plate with a
2-mm-diameter, 4-mm-long cylindrical hole is located close to the emitter disk.
Four cylindrical lenses are used to focus the electron beam, and these are stacked
between ceramic spacers [19]. A single positioning ring at the end of the stack holds
the assembly together, and the assembly is housed inside a metal tube housing. To
minimize work function differences along the gun, the inside of the lens is gold
plated.
Figure 20.3b shows calculated electron trajectories for different electrode
potentials. At the top, an anode at high positive potential withdraws electrons from
the cathode, and the trajectories are shaped by the four lenses set for potentials from
+30 V to 0 V as shown. This focuses the beam on the crystal at ground potential.
Figure 20.3c shows the electrostatic potential in the cathode region in a
three-dimensional plot for electrons. The geometry and potential of the repeller and
anode create a steep potential drop for electrons. The shape of the potential is
designed to increase the lateral momentum of the electrons in order to achieve a
broad electron beam. The broad beam exhibits very uniform current density as may
been seen from the trajectory spacings in Fig. 20.3b.
In the simulations, it was found that the emission is very sensitive to the exact
position of the cathode disk relative to the repeller cap and the anode opening.
Figure 20.3d shows the performance of the gun for various electron beam
energies. At an electron energy of 10 eV, a current of 7 μA is obtained at the crystal,
comparable to the performance of the gun shown on p. 293. The optimized anode
and repeller potential are shown for various electron energies in Fig. 20.3e.
20.4 Electron Energy Analyzers
It is not the purpose of this book to discuss the details of the various forms of
electron energy analyzers that are used for the analysis of surface phenomena.
Instead, a practical guide to the selection, construction, and use of electron energy
analyzers is available [20]. In addition, an extensive bibliography dealing with the
various design analyzers is also presented in [20].
Six types of electron energy analyzers are described in [20], and these
constitute the most widely used instruments. The analyzers are briefly categorized
below.
Figure 20.4a—Time-of-Flight Analyzers
These analyzers are used primarily for ion detection owing to the extremely high
velocity of electrons. However, time-of-flight methods have also been used for
rapid pulsed sources of slow electrons. Electrons are allowed to drift in a field-free
region, and their arrival at a microchannel plate detector is recorded by a suitable
waveform digitizer. MicroChannel plate systems can provide gains of 106–107 and
192 20 Electrons
Fig. 20.3 Low-energy electron gun—broad beam a Design of the low-energy electron gun.
b Beam profile for different electron energies. c Shape of the potential near the cathode.
d Emission current from low-energy electron gun. e Potential settings for broad beam operation
and constant electrode potentials
20.4 Electron Energy Analyzers 193
Fig. 20.3 (continued)

194 20 Electrons
Fig. 20.4 Electron energy analyzers. a Time-of-flight analyzer. b Retarding grid analyzer.
c Tandem reflection—transmission analyzer. d Plane mirror deflection analyzer. e Cylindrical
mirror deflection analyzer. f Cylindrical deflection analyzer. g Spherical deflection analyzer
20.4 Electron Energy Analyzers 195
pulse widths of 1 ns. Drift-tube lengths of a few meters are employed, and in such
long tubes magnetic field effects may be significant.
Figure 20.4b—Retarding Grid Analyzers
Here, the source is located at the center of two hemispherical grids and a spherical
collector. The source and the inner grid are grounded to make a field-free inter-
vening region. The outer grid, biased at—VR, retards the electrons, and the collector
is biased positive relative to ground to collect all electrons that pass the retarding
grid. The energy resolution is estimated from the assumption that the source is
spherical, of radius r = a, and that the radius of the collector is r = b. A schematic of
the transmission of the device versus kinetic energy is shown when the retarding
grid is set at VR. The energy resolution, ΔE, at VR is given by the equation in
Fig. 20.4b.
Figure 20.4c—Tandem Reflection-Transmission Analyzer
This type of analyzer allows bandpass filtering of the energy distribution of elec-
trons emitted from the source. The first grid system is a low-pass filter. Electrons
having energies greater than eVREF can overcome the retarding field and are col-
lected on the ellipsoidal collector. Lower-energy electrons are reflected and focused
through a grounded aperture, and thence to a retarding grid analyzer something like
that shown in Fig. 20.4b. The two-dimensional image detector makes possible the
simultaneous detection of the energy and the emission angle of the electrons
received in the band pass of the two analyzers, as shown in the transmission
196 20 Electrons
diagram for the tandem analyzers. Analyzers of 100-meV resolution and large
angular acceptance have been constructed.
Figure 20.4d—Plane Mirror Deflection Analyzer
This represents the most elementary type of the deflection analyzers. Electrons,
constrained to a certain angular range by the apertures, enter from the left into a
region of width = h in which field-free conditions exist. From there, they enter a
second region of width = d, where a constant electric field exists, which field is
produced by making the upper plate negative. Here, the electrons travel in parabolic
paths in the uniform field. Electrons having a specific energy arrive at the exit slits
on the right and are measured below the lower grounded plate.
Figure 20.4e—Cylindrical Mirror Deflection Analyzer
Here principles similar to those used in the plane mirror analyzer apply. However,
because of the cylindrical symmetry, a wide solid angle of acceptance exists,
leading to high throughput and sensitivity. The cylindrical geometry is also con-
venient for the placement of the electron gun on axis, and has made this geometry
popular for Auger electron spectroscopy (AES).
Figure 20.4f—Cylindrical Deflector Analyzer
This type of analyzer is widely used in high-resolution electron energy loss spec-
troscopy (HREELS). The optimum angle between the entrance and exit slits is
θ = 127°. Analyzers of this type are particularly useful for measuring the angular
distribution of electrons emitted from a surface.
Figure 20.4g—Spherical Deflection Analyzer
This type of analyzer is widely used in photoelectron spectroscopy. The optimum
angle between the entrance and exit slits of the analyzer is θ = 180°.
Improvements in the resolution and throughput of these analyzers can be
achieved by preretarding the incident electrons prior to entry into the analyzer [20].
20.5 Electron Spectrometer Calibration Using Graphite
The accurate location of the Fermi level (EF, sample) is often an important
requirement for the calibration of electron spectrometers. Often in ultraviolet
photoelec-tron spectroscopy (UPS) the Fermi level can be determined by measuring
the Fermi edge feature of a clean metal surface. However, if UPS is not available,
the precise calibration of the Fermi edge position often cannot be carried out.
A method to do this, using the secondary electron emission from electron impact
excitation, X-ray photons, or UV photons on a graphite reference sample is
described below [21].
Figure 20.5 shows the secondary electron emission region of a sample of
graphite (highly oriented pyrolytic graphite, or HOPG) and a sample of Aquadag
20.5 Electron Spectrometer Calibration Using Graphite 197
Fig. 20.5 Electron spectrometer calibration
(colloidal graphite) [22] that have been excited by He I (21.22 eV) radiation. In
addition, the upper energy photoelectrons in the vicinity of the Fermi edge are also
shown from a cleaned Ta sample in electrical contact with the graphite samples.
The fact that the upper energy measured is at 21.22 eV indicates that the spec-
trometer energy scale is properly calibrated in this measurement. It can be seen in
the secondary electron emission region that there is a distinct emission feature
centered above the Fermi level at 7.57 ± 0.03 eV and superimposed on the broad
secondary emission tail. This feature is related to electronic excitation of the a state
present in the conduction band of the graphite. This state arises from the a electrons
198 20 Electrons
in the C–C bonds in the graphite lattice and has been well studied previously
[23, 24]. The energy of the σ peak from graphite is independent of the energy of the
exciting radiation or electrons and hence can be used to locate the Fermi level
accurately when it is otherwise not accessible.
The behavior of the a state is independent of the annealing temperature of the
Aquadag over the range 230–750 °C and is independent of air exposure. Also, the
energy is invariant with the angle of emission. From the low-energy cutoff in the
secondary electron spectrum, the sample or electron analyzer work function can be
determined [21].
20.6 Measurement of Electron Beam Angular Divergence
The response of an instrument measuring an electron diffraction beam is determined

by many factors. Included are the source extension (or the angular divergence of the
electron beam), the beam width, and the electron energy distribution. The most
difficult factor to measure is the angular divergence of the beam, and this factor has
generally been measured by the elimination of other parameters.
The method to be described can measure both the beam width and the angular
divergence of the beam. Figure 20.6 shows a schematic view of the apparatus [25].
A stationary electron gun is directed toward a set of three Faraday cups that are
translated by an electrically driven bellows. The Faraday cup detectors are located
at different distances along the electron beam, and each 2-mm-diameter hole is
half-covered with a razor blade acting as a knife edge. The current to each Faraday
cup is measured as the assembly is translated across the beam. Thus, the shape of
the electron beam is measured at three z-positions, and this permits the divergence
to be obtained by comparing the full width at half maximum (FWHM) at each of
the three cup positions. Placing one of the Faraday cups at the distance of the
sample in normal use allows the beam character to be measured at the sample
position. The loss of current through secondary electron emission is minimized by
using Faraday cups that are at least three times as deep as their 2 mm apertures, and
losses of a few percent are found in the 50–600-eV region [26]. The cups are not
biased, and coaxial cables are used inside and outside the vacuum system to
minimize noise.
An AC method may be used to make measurements of the first derivative of the
collected current versus translation position. A small sinusoidal modulation is
applied to one pair of deflection plates to provide a slight motion of the beam
perpendicular to the knife edge on the collector. A lock-in amplifier measures the
derivative signal as a function of translation position (measured electrically) and an
accurate beam profile is plotted out with improved stability and noise compared to
using the DC method. No mechanical instability was observed in measuring
20.6 Measurement of Electron Beam Angular Divergence 199
Fig. 20.6 Measurement of electron beam divergence
50–600-eV beams having a typical FWHM of 0.1 mm. A beam divergence of

less than 0.1° may be measured, and typical electron guns exhibit divergences of
about 0.2°.
The application of information obtained by the above method to actual LEED
instruments is described in an excellent review article [27].
For the approximate measurement of electron and ion beam (>400 eV) dimen-
sions, a phosphor screen may be used. A nice method to measure the beam
dimensions involves the use of a small stainless steel plate with a precise grid
overlay attached by welding. A grid that has 0.2 × 0.2 mm2 openings, photoli-
thographically etched, is ideal. A grid thickness of 0.025 mm is appropriate [28].
The phosphor, mixed with ethanol, is applied to the screen, and excess phosphor is
removed from the outer surface of the grid overlay. When dry, the phosphor will
adhere only inside the grid spacings, and the known dimensions of the grid provide
a direct calibration scale for the electron beam [29].
200 20 Electrons
20.7 Measuring Anisotropy of Electrical Conductivity

of a Single Crystal—Four Point Probe
The bulk electrical conductivity of single crystal semiconductor materials is of

importance in understanding their behavior in devices such as photocells. A four-
point probe can be employed for the measurement at 300 K in air as schematically
indicated in Fig. 20.7a [30]. Here four spring-loaded SiC tips are pressed into the
crystal to make good contact with the bulk. The contact breaks the surface of the
Fig. 20.7 Measuring anisotropy of electrical conductivity of a single crystal—four point probe.
a Four point probe. b Anisotropy in conductivity of TiO2(110). The radius observed is
proportional to the electrical conductivity of the single crystal in the measured direction
20.7 Measuring Anisotropy of Electrical Conductivity … 201
TiO2. The force on each tip was equivalent to 100 g. The SiC contact points are
located at the corners of a 1 mm square and are 0.1 mm in diameter. Current is
passed through two non-diagonal tips and the potential drop is measured through
the other two tips. From this information, the electrical conductivity is measured at
the angle of the probe with respect to the <001> crystal axis. The calculation
method of Yamashita [31] was employed. This method takes into account the finite
width and thickness of the crystal. The four-point probe is rotated in a systematic
manner and a new conductivity is measured at each azimuth. Figure 20.7b shows a
directional conductivity measurement for TiO2 measured on the well studied <110>
face [30].
20.8 Spot Photometer for LEED Intensity Measurements
The measurement of LEED beam intensities using a spot photometer may be

carried out in the absence of a Faraday cup or electronic image-enhancing display
method. The use of a uniform phosphor screen makes these measurements satis-
factory for many purposes.
The construction of a spot photometer is shown in Fig. 20.8. The detector is
equipped for synchronous detection, eliminating interference from background
Fig. 20.8 Spot photometer for LEED intensity measurements

202 20 Electrons
light. A lens focuses the light from a LEED beam on a photodiode that is mounted
behind an aperture placed in a thin reflective metal screen. The LEED pattern can be
observed visually through a port using the reflective screen and the spot photometer
is moved laterally for measurement of individual beams. A photodiode mounted
behind the aperture integrates the light received from a selected beam plus the
surrounding background. The diffraction pattern is modulated at 100 Hz by
applying a chopping voltage to the suppressor grid of the order of Ve + δ, which is
sufficient to suppress the LEED pattern [32].
At a spacing of 30 cm from the LEED screen, the photodiode sees a circular area
of about 5.4-mm diameter, about five times the average diameter of a diffraction
spot in the apparatus. The photodiode is a low-noise planar-diffused silicon sensor
photodiode [33]. The reader may see examples of the use of this apparatus in
[34, 35].
Newer methods involving vidicon detectors are widely employed for the study
of LEED patterns and intensities on phosphor screens. It is possible to employ a
bandpass optical filter to remove stray light from high temperature filaments and
emission from a hot sample. For a P11 phosphor, with a narrow band of emission at
459 nm, an interference filter with a central wavelength of 460 nm and a bandwidth
of 50 nm is excellent for removal of the stray light, permitting more accurate
measurement of LEED intensities on a background that is closely similar to the dark
current of the vidicon [36].
20.9 Modified Faraday Cup for Electron Current

Measurement
Faraday cups are employed for the collection of electrons in collimated beams. One
problem is that specularly reflected electrons from the bottom of the Faraday cup
may be reemitted, causing the measured collected current to be diminished. Various
methods for the suppression of specularly reflected and secondary electron emission
are employed, and one principle employed is to make the Faraday cup depth a
number of times the aperture width. But even this design criterion does not prevent
the reflection of specular electrons from the bottom of the cup. Surfaces are known
to have high reflectivities for low-energy electrons of a few eV energy.
To avoid this problem, a stainless steel cone was inserted in the cup [37], as
shown in Fig. 20.9a. The specularly reflected electrons from this conical surface
maintain their forward component of velocity along the optical axis of the
cup. These electrons have no velocity component along the optical axis toward
the aperture. Thus this design is the electron optical equivalent of a light trap. The
design invalidates the principle that the depth of the cup should be large compared
to the aperture.
20.9 Modified Faraday Cup for Electron Current Measurement 203
Fig. 20.9 Modified faraday

cup. a Faraday cup with
internal cone. b Electron
energy distribution—no cone.
c Electron energy distribution
—with cone
The energy distribution of the collected electrons (10−11 A) versus retarding

potential is shown in Figs. 20.9b, c for the unmodified Faraday cup and the Faraday
cup containing the inner conical surface. It is seen that the ratio I/V can be increased
by a factor of 2 from 2.5 to 5 when the cup contains the conical insert. At the
same time, AE increased from 1.2 to 1.6 eV. No explanation for the existence of
the plateau extending into the loss region by about 8 eV was presented in the
paper [37].
204 20 Electrons
Fig. 20.10 HREELS-discrimination between single and multiple electron scattering events.
a EELS and AES measurements of Al2O3/Al(111) at three different levels of oxidation.
b Correlation between the EELS intensities and the Al2O3 thickness
20.10 HREELS-Discrimination Between Single … 205
20.10 HREELS-Discrimination Between Single

and Multiple Electron Scattering Events
High resolution electron energy loss spectroscopy (HREELS) is widely used to

study the vibrational spectrum of adsorbed molecules and the phonon spectrum of
thin films. In addition to the excitation of fundamental vibrational modes
(ν = 0 → ν = 1), it is possible to excite more complex modes such as overtones and
combination losses. These processes differ in the number of vibrational quanta
excited or the number of scattering events the electron undergoes. In fundamental
and overtone loss processes, a single electron excites a single mode; in the case of
multiple electron energy loss processes, an electron undergoes two or more scat-
tering events. Combination loss processes often fall into the category where the
electron undergoes multiple scattering.
The multiple scattering events (here double scattering) should scale as the square
of the thickness of the film being investigated, whereas fundamental modes should
be linearly dependent on the thickness. This is demonstrated in the case of an Al2O3
film on an Al(111) substrate, as shown in Fig. 20.10a. Here the three low frequency
modes, ν1, ν2, and ν3 are fundamental phonon modes of the oxide which involve
single electron scattering. The higher frequency modes involve multiple scattering.
As the thickness of the Al2O3 film grows, as evidenced by the Auger spectra, the
relative intensity of the multiple loss modes increases non-linearly with respect to
the single scattering modes, as may be seen by the changing magnification scale
(from 6666x to 200x) needed to maintain the appearance of similar spectra for the
three film thicknesses. In Fig. 20.10b, plots of intensity versus the thickness, d, or
the thickness squared, d2, are shown. In Fig. 20.10b(b), the multiple scattering
modes scale in their later stage of development with d2, whereas in Fig. 20.10b(a),
the single scattering modes scale with d. This may also be seen in Fig. 20.10b(c)
where the ratio of intensities of the multiple to single scattering modes increases as
film thickness increases. The drop off of intensity seen in all three figures for thin
films is due to imperfections of the film thickness measurement using Auger
spectroscopy [38].
20.11 Rejuvenation of Cu/Be Electron Multipliers
The deposition of thin films onto electron multipler surfaces occurs after long use,
particularly in vacuum systems having a background of oil vapors or in systems that
have been used for reactive gases, such as fluorine [39]. A simple rejuvention
method has been reported for Cu/Be dynodes exposed to fluorine to the extent that
the multiplier gain had dropped to 1/300 of its original gain. Complete rejuvenation
back to the original gain was observed [39].
For the particular case here, it was decided to expose the dynodes to a plasma
discharge in hydrogen to remove the fluorine, and then to expose them to a plasma
206 20 Electrons
Fig. 20.11 Rejuvenation of Cu/Be electron multipliers
discharge in oxygen to reoxidize the dynode surfaces. This was done, of course, by
removing the multiplier and treating it in a separate metal vacuum system. In
Fig. 20.11, a sequence of rejuvenation steps is carried out, in which neighboring
dynodes are used as anode/cathode combinations, working up the multiplier chain.
For hydrogen plasma generation, a pressure of about 1 Torr of H2 was used, under
flow conditions, with a 350 V DC power supply connected through 5000 Ω
between adjacent dynodes. A Tesla coil was used to initiate the plasma, and the
pressure was increased in the range 1–5 Torr until about 30 mA of current passed.
Hydrogen plasma treatment was carried out for 2 min. After this, for O2 plasma
generation, about 10 mA of current was used for 1 min. The multiplier anode was
not treated.
A cleaning and oxidation treatment, designed to avoid copper oxide formation
on the dynode structures and to place beryllium oxide films on the dynodes, has
been described [40]. Here, oxidation in water vapor at pressures up to 5 × 10−2 Torr
was carried out in cycles in an RF-heated furnace at 625 °C. Beryllium oxide
exhibits a higher secondary electron yield than copper oxide [40].
20.12 Geiger-Müller Counter for Inverse Photoemission
Inverse photoemission involves the emission of ultraviolet photons from a surface

when the surface is excited by electron impact. This method samples the empty
density of states of the surface and near surface regions of a solid. The intensity of
photon emission is measured as a function of incident electron energy.
A Geiger-Müller counter, with windows passing ultraviolet light, is used for the
measurements [41, 42].
20.12 Geiger-Müller Counter for Inverse Photoemission 207
Fig. 20.12 Geiger-müller counter for inverse Photoemission
Figure 20.12 shows the counter, which consists of two parts, the housing unit
and the insertion unit. The housing unit is mounted on a 2 3/4-in. Conflat flange.
A conical ring presses either CaF2 or SrF2 windows [43] onto the housing con-
taining a Viton O-ring. The inner surface of the ring is beveled at 45° to minimize
photon emission from this surface by stray electrons. The window has a diameter of
20 mm and a thickness of 2 mm.
The insertion unit is mounted inside the housing unit, and can be removed for
bakeout. It too is sealed with a Viton O-ring at the 2 3/4-in. flange. Both the inner
wall of the insertion unit and the counting wire (1.5-mm diameter) are polished. The
counting wire is electrically isolated from the apparatus by a Macor insulator, which
is glued to the outer flange and to the counting wire with vacuum cement [44]. The
active length of the counting wire is 40 mm and its end is located 10 mm from the
window. The glue joint is made with the counter wire accurately centered in the
insertion unit. The port serves to evacuate the counter as well as being an inlet for
Ar background gas.
The counter is assembled with the help of a He leak detector to assure tight
sealing between the housing unit and the entrance window with minimum force on
the front-end bolts. After installation, the assembled unit is pumped overnight, then
opened and a few crystals of iodine are dropped in. After pumping again for 30 min,
Ar is introduced to a pressure of 100–200 mbar, and the counter is valved off.
Typical operating voltages are 400–500 volts. All pulses should be of identical
height (about 0.2 V). Occasionally the counter is pumped again and refilled with Ar.
Repeated fillings with Ar may also help if the counter does not work initially.
Deterioration of the counter performance (requiring a steady increase in voltage)
usually indicates insufficient vacuum conditions.
208 20 Electrons
References
1. K.R. Spangenberg, Vacuum Tubes (McGraw-Hill, New York, 1948)

2. J.A. Simpson, C.E. Kuyatt, J. Res. Nat. Bur. Stand. 67C, 279 (1963)
3. J.R. Pierce, Theory and Design of Electron Beams (D. Van Nostrand Co., Inc, New York,
1949)
4. F. Rosebury, Handbook of Electron Tubes and Vacuum Techniques (American Institute of
Physics, New York, 1993)
5. A particularly excellent account of practical aspects of electron optics is given in J.H. Moore,
C.C. Davis, M.A. Coplan, Building Scientific Apparatus (Addison-Wesley, Reading, MA,
1983), chapter 5
6. Cliftronics, Inc., 15 South Street, Clifton Springs, NY 14432. Model 406-S; Another sup-plier
of electron guns is Kimball Physics, Inc., 311 Kimball Hill Road, Wilton, NH 03086-9742
7. M.G. Lagally, J.A. Martin, Rev. Sci. Instrum. 54, 1273 (1983)
8. This program, which is periodically updated, is available from Idaho National Engineering
Laboratory, EG8cG Idaho, P.O. Box 1625, Idaho Falls, ID 83415
9. R.Q. Hwang, E.D. Williams, R.L. Park, Rev. Sci. Instrum. 60, 2945 (1989)
10. R.N. Lee, Rev. Sci. Instrum. 39, 1306 (1968)
11. M. Proulx, P. Marmet, R. Dutil, Rev. Sci. Instrum. 53, 778 (1982)
12. M.K. Debe, D.C. Johnson, Rev. Sci. Instrum. 47, 37 (1976)
13. M. Aeschlimann, E. Hull, J. Cao, C.A. Schmuttenmaer, L.G. Jahn, Y. Gao, H.E. Elsayed-Ali,
D.A. Mantell, M.R. Scheinfein, Rev. Sci. Instrum. 66, 1000 (1995)
14. J.-L. Liang, J.T. Yates Jr, J. Vac. Sci. Technol. A12, 2795 (1994)
15. This program, which is periodically updated, is available from Idaho National Engineering
Laboratory, EG&G Idaho, P.O. Box 1625, Idaho Falls, ID 83415
16. J.C. Lin, R. Gomer, Surf. Sci. 172, 183 (1986)
17. T.R. Barefoot, H.D. Ebinger, J.T. Yates Jr, J. Vac. Sci. Technol. A15, 2740 (1997)
18. BaO disk cathode, ES-015 on AEI base, Kimball Physics, Inc., 311 Kimball Hill Road,
Wilton, NH 03086-9642
19. A12O3 ceramic insulator—Finnigan # 40005-20070, Scientific Instrumentation Services, Inc.,
1027 Old York Road, Ringoes, NJ
20. J.L. Erskine, Experimental Methods in the Physical Sciences, vol. 29A (Academic Press, Inc.,
1995), p. 209
21. P. Oelhafen, I.L. Freeouf, J. Vac. Sci. Technol. A1, 96 (1983)
22. Aquadag may be obtained from Acheson Colloids, 1607 Washington Avenue, P.O. Drawer
611747, Port Huron, MI 48061-1747. In applications of this type, it may be painted onto the
sample to yield the desired reference energy above the Fermi level
23. J. Krieg, P. Oelhafen, H.-J. Guntherodt, Solid State Comm. 42, 831 (1982)
24. R.F. Willis, B. Feuerbacher, B. Fitton, Phys. Rev. B. 4, 2441 (1971)
25. J.A. Martin, M.G. Lagally, J. Vac. Sci. Technol. 18, 58 (1981)
26. M.G. Lagally, Ph.D. Dissertation, University of Wisconsin, 1968
28. Photolithograph W grid is available from Buckbee Mears Corp., 245 East 6th Street, St. Paul,
MN 55101-1959
29. Dr. R. David, Institut fur Grenzflächenforschung und Vakuum-physik, Forschungszentrum
30. O. Byl, J.T. Yates Jr, J. Phys. Chem. B. 110, 22966 (2006)
31. M. Yamashita, M. Agu, Jap. J. Appl. Phys. 23, 1499 (1984)
32. A.G. Schrott, M.D. Chinn, C.G. Shaw, S.C. Fain Jr, J. Vac. Sci. Technol. 21, 101 (1982)
33. PIN-5DP, United Detector, 12525 Chadron Avenue, Hawthorne, CA 90250
34. C.G. Shaw, S.C. Fain Jr, M.D. Chinn, M.F. Toney, Surf. Sci. 97, 128 (1980)
35. R.D. Diehl, M.F. Toney, S.C. Fain Jr, Phys. Rev. Lett. 48, 177 (1982)
36. R.J. Phaneuf, H.-C. Kan, Rev. Sci. Instrum. 65, 3871 (1994)
References 209
37. W.R. Miller Jr, N.N. Axelrod, Rev. Sci. Instrum. 37, 1096 (1966)
38. J.G. Chen, J.E. Crowell, J.T. Yates Jr, Rapid communications. Phys. Rev. B. 35, 5299 (1987)
39. G.E. Salser, Rev. Sci. Instrum. 37, 674 (1966)
40. M.B. Pongratz, Rev. Sci. Instrum. 43, 1714 (1972)
41. V. Dose, Rev. Sci. Instrum. 39, 1055 (1968)
42. V. Dose, Th Fauster, R. Schneider, Appl. Phys. A. 40, 203 (1986)
43. CaF2 and SrF2 windows are available from Harshaw Optical/Bicron, 6801 Cochran Road,
Solon, OH 44139
44. The vacuum cement used is Araldit ATI, Ciba-Geigy GmbH, Oeflinger Str. D-79664 Wehr
(Baden), Germany
Chapter 21
Ions
21.1 Ion Sputter Gun with Low Gas Emission
Normal ion guns operate in an ambient pressure of the gas to be used for sputtering.
This pressure needed to generate a satisfactory ion bombardment current is nor-
mally in the 10−6–10−5 Torr range [1]. Improved ion gun designs involve a direct
feed of the inert gas into the ionization region, and these designs normally involve
an ambient pressure in the 10−8 Torr range. However, any rise in the pressure in the
main chamber during the cleaning process is detrimental, since it is known that gas
release from the chamber walls, from the ionization pump (if left on) and from the
sublimation pump is caused by an increase in the pressure of even an inert gas. This
undesirable gas evolution depends upon the previous history of the vacuum system
and influences the time required to pump down to the base pressure after ion
sputtering.
The design shown in Fig. 21.1a involves direct injection of the sputtering gas
into the ion gun and direct pumping from the ion gun into a turbopump. The gas
beam inside the ion gun is created by a multicapillary array collimator (see p. 421
for gas doser designs employing this collimation principle). The angular distribu-
tion of sputter gas molecules exiting from the array may be calculated [2, 3]. The
sputtering gas is injected into the ion gun through transparent Ta grids that are
welded over windows cut into the walls of the outer shield of the ionization gun.
A multicapillary array of 13-mm diameter and 3-mm thickness is used, with 10-μm
diameter capillaries. Control of the gas pressure is achieved with a leak valve, as
shown in Fig. 21.1b. Two gate valves are used to isolate the main chamber from the
turbopump after sputtering when an ion pump is used for maintenance of UHV.
Normally the 360-L/s turbopump is used to pump the main chamber in parallel to
an ion pump and a Ti sublimation pump (not shown) [4].
While an IQ 10/35 ion gun is shown in the example here, many ion gun designs
will work with this modification.

DOI 10.1007/978-3-319-17668-0_21
212 21 Ions
Fig. 21.1 Ion sputter gun with low gas emission. a Modified ion gun. b Schematic of vacuum
system
21.2 Alternate Ion Bombardment Sources 213
21.2 Alternate Ion Bombardment Sources
Ion bombardment of elemental crystal surfaces is widely used for removing both
surface and bulk impurities. In the case of bulk impurities, cleaning by sputtering is
often accomplished at elevated temperatures, permitting bulk impurities to diffuse
to the surface for removal. Normally an ion gun is employed for this and guns
giving both focused beams and more diffuse beams are available. Admission of Ar
gas directly to the ionization region of the ion gun will reduce the gas load in the
vacuum system and hence reduce contamination effects during and after sputtering
(see pp. 87–91).
Alternately, very simple ion bombardment sources have been reported, and two
examples of these are illustrated here. In Fig. 21.2a, an open ion gauge is used as a
source of Ar+ ions [5]. The vacuum system is filled with Ar gas to about 10−3 Torr,
permitting ample gas density within the ionization region of the gauge for
Fig. 21.2 Two alternate ion bombardment sources. a Ion gauge sputter source. b Reflector sputter
source
214 21 Ions
producing an ion current which is collected at the sample biased at −600 V. It is

reported that for an electron emission current of 10 mA, ion currents of 10–15 μA
are collected at the crystal, depending on the negative bias. The gauge becomes
inaccurate for pressure readings under these conditions. In addition, sputtered metal
from the elements of the gauge will soon coat the insulator, which isolates the
central collector electrode, causing leakage currents that disable the gauge as a
pressure-measuring device. A shield below the grid structure preventing metal
deposition on the collector insulator has been used to prevent these effects. The
gauge is normally cleaned extensively by electron bombardment of the grid before
use as an ion source, keeping the sample positioned elsewhere without its electrical
bias.
Ion bombardment sources made without the use of carefully designed ion
extraction optics have the capability of delivering impurity metal to the sample from
sputtering processes inside and around the gauge components. Therefore, one
should be very certain in controlled, and high-sensitivity Auger studies, that these
contamination effects do not occur. For the design shown here, with the dimensions
and mounting geometry indicated in Fig. 21.2a, Pt and Cu crystals were success-
fully cleaned and comparisons with a conventional ion gun showed that no dif-
ferences in the surface preparation could be detected by Auger spectroscopy and
other surface characterization methods [5].
Figure 21.2b shows an even simpler ion bombardment source. A tungsten fila-
ment is mounted on two tungsten feedthrough wires between a curved anode and the
sample. Electron emission current of the order of 18 μA is collected by the curved
anode at a bias potential of +500 V, and at an Ar pressure of about 5 × 10−5 Torr, an
Ar+ current density of about 20 μA/cm2 is collected by the sample at ground
potential. At 3 × 10−4 Torr Ar pressure, ion current densities of 50–60 μA/cm2 are
reported. The ion source should be scrupulously cleaned by heating the emitter for a
long time to remove desorbing gases, and also by electron bombarding the curved
anode extensively to remove adsorbed gases that might undergo electron stimulated
desorption during use of the source for sputtering. Again, with this device, special
care should be taken to eliminate the possibility of metal deposition from extraneous
sources, although the report [6] indicates that this effect is not observed on a GaAs
wafer sample.
A more complex ion gun source made from a Bayard-Alpert gauge is described
on p. 214.
21.3 Ion Gun Based on Bayard-Alpert Gauge
A more complex ion gun than that described on p. 211 and 213 is presented here
[7]. A Bayard-Alpert gauge is modified in two ways: first, by removal of the top
part of the cylindrical grid cage for addition of the ion extraction electrodes; and
second, by the addition of a metal shield around the gauge body and the thermionic
emitter.
21.3 Ion Gun Based on Bayard-Alpert Gauge 215
Figure 21.3a shows a schematic of the gun design and the electrical components.
A tantalum cylindrical shield to repel electrons is mounted around the gauge body
and its electrical potential is maintained at the same potential as the positive end of
the thermionic emitter filament. An extractor lens shaped with a waist withdraws
ions from the region enclosed by the grid and these pass into an accelerator region,
and thence into a cylindrical lens and a quadrupole lens, where the ions are ejected
into vacuum and to the target. The flange holding the ion gun is equipped with
radial feedthroughs about its outer edge for supplying potentials to the ion gun
elements.
The extractor lens accelerates mainly off-axis ions to an on-axis focal point, and
its potential can be as high as −200 V. The divergence of the ion beam is controlled
by the bore of the accelerator electrode, and the final focusing is done by the lens.
Since all potentials in the gun, except for the lenses, are referenced to the primary
energy, the extraction of ions is independent of the final beam energy. At low
energies, the ions are of high energy when they arrive at the lens, and are decel-
erated there, while being focused. This leads to high currents at low ion energies.
The quadrupole lens consists of four plates at 90° to each other symmetrically
placed around the gun axis, but its use is not described in the original article [7].
Fig. 21.3 Ion gun based on Bayard-Alpert gauge. a Ion gun and circuit. b Ion spatial distribution.
c Current and current density. d Ion beam energy distribution
216 21 Ions
The spatial distribution of the ions may be adjusted as shown in Fig. 21.3b.
Either a sharply peaked ion beam with high current density at the center may be
produced, or alternatively, a flat-topped beam profile may be produced by adjusting
the acceleration and lens voltages.
In Fig. 21.3c, the current and current density are plotted for a pressure of
1 × 10−4 Torr Ar pressure. Current densities between 0.1 and 120 μA/cm2 are
obtained at various ion energies. The current collected near 1 keV energy is limited
by space charge effects in the extractor region. The collector in these cases was
located 4 cm from the end of the gun.
Figure 21.3d shows the energy distribution measured with a retarding field
analyzer and showing that at 1.2 keV, an ion beam halfwidth of 1.6 eV may be
obtained using an extractor voltage of about −30 V. The ion beam was also mass
analyzed, and at an electron energy of 70 eV, a 1:10 ratio of Ar++ to Ar+ was
measured.
An earlier version of this design for an ion gun has been published [8].
21.4 Sputter Ion Gun—Cold Cathode
A cold cathode ion gun, useful for sputter deposition, can be made, as shown in
Fig. 21.4 [9]. The gun is useful for sputter deposition from any conducting target
material. The ion gun was first described in [10], and this is a modification of that
design. The gun is capable of producing an ion beam of 1-mA current at a kinetic
energy of 2–6 keV, impinging on an area 5–10 mm in diameter. There is no
filament, and the gun can be operated for hundreds of hours without maintenance.
Currents in excess of 1 mA cause tip heating which shortens the tip life. Sputter
deposition rates of 1–4 Å/s can be obtained.
The beam gas is admitted from a pumped gas handling system through two
ceramic insulator nipples and through a metal frit particle filter into a Conflat
flange-mounted electrical feedthrough passing a 1/4-in. tube that is connected via
Swagelok connectors to a hollow tube made of l/16-in.-diameter tungsten or tung-
sten carbide alloy. This needle is 0.625-in. long and has an orifice of 0.05–0.075-mm
diameter. It is called a wire bonding tool [11]. The high electrical field at the tip
ionizes the gas as it passes through the tip, resulting in a plasma cone that extends to
the target. The diameter of the cone depends on the orifice diameter, supply gas
pressure, and target distance. A 2.5 Torr beam of gas in the gun, combined with a
tip-target distance of 6 mm and a 0.050-mm-diameter orifice produces a beam crater
diameter of 7–10 mm.
A principal difficulty is keeping particles out of the tip. The tip may be
unplugged by passing a 0.003-in.-diameter W wire from the back end. This wire
may be made from larger stock by electrolytic etching in NaOH (aq) solution. Final
cleaning of the tip is done in an ultrasonic bath with acetone.
The beam current and stability depend on the pressure of gas entering the gun.
Stable operation is obtained when the chamber pressure is in the 5 × 10−4 to
21.4 Sputter Ion Gun—Cold Cathode 217
Fig. 21.4 Sputter ion gun—cold cathode

218 21 Ions
3 × 10−3 Torr range. If the supply pressure is too low, arcing will occur in the
high-voltage nipples on the gas supply line. If the pressure is too high, or the orifice
diameter is too large, a glow discharge will occur in the chamber with little apparent
localization. The beam current increases with supply gas pressure up to the point of
glow discharge initiation. The high-voltage power supply should be current limited
to prevent more than 5–10 mA of current from flowing during such an event.
Another design of a cold cathode sputtering source for metal deposition maybe
found in [12].
21.5 Broad Ion Sources—Improved Stability with Oxygen
Broad-beam ion sources are employed for many purposes including sputter depo-
sition, etching, etc. Figure 21.5 schematically shows a typical conventional source,
which illustrates the basic principles of more complex sources [13, 14]. The
cylindrical chamber contains a cylindrical anode and an axially positioned cathode.
Gas for ionization is admitted from the rear of the source. A magnetic field, sup-
plied by either an electromagnetic winding or a permanent magnet, is aligned
axially, and the magnetic field diverges slightly as one approaches the front of the
source. The purpose of the magnetic field is to cause the electrons emitted from the
cathode to have a spiral path length sufficiently long such that each will cause a gas
phase ionization event before reaching the anode. The source has two grids on its
Fig. 21.5 Ion source—axial magnetic field

21.5 Broad Ion Sources— Improved Stability with Oxygen 219
front side—a screen grid and an accelerator grid. The discharge chamber is nor-
mally operated under low-pressure conditions where the mean free path of an
electron for ionization is much longer than the apparatus dimensions. The plasma
potential is usually within a few volts of the anode potential, and the primary
electrons emitted from the cathode therefore have energies close to the potential
difference between cathode and anode [13, 14].
When Kaufman sources [13, 14] are used with inert gases or reactive gases at
low power levels, they operate for long times without failure. However, at high
pressures with reactive gases like O2, the ion source lifetime is shortened. This is
due to failure of the refractory cathode (such as W or Ta), or to buildup of an
insulating oxide layer on the anode. Such an insulating layer on the anode reduces
the discharge current until it is no longer possible to support a discharge [15]. This
lifetime problem can be reduced significantly by using thorium oxide coated Ir
cathodes (see p. 133) instead of the more standard bare refractory filaments. For
electron emission current densities of 0.2 A/cm2, the thoriated filament lifetime is
1000 h or longer [16]. As the emission current density is increased, the lifetime
decreases as the square of the current density. Multiple filaments, used in series, can
extend the lifetime by reducing the current density at each filament [15].
To avoid the buildup of an insulating layer on the anode, it may be
sputter-coated with a 1-μm-thick film of IrO2, which is a conductive metallic oxide.
Because this oxide cannot be oxidized further, operation in an intense plasma will
not form an insulating layer, as with many anode metal materials. The source,
containing both a thoria-coated Ir filament and an IrO2-coated anode has been
operated at 100-W discharge power for about 50 h without degradation of
source performance. With an uncoated stainless steel anode, failure occurred in
30 min [15].
Other possibilities exist for using this idea. The anode may be coated with other
con-ductive metal oxides such as those from Ru or Re [17, 18]. Alternatively, the
anode could be made from the oxide by chemical vapor deposition. Finally, the
anode could be made from solid Ir, since an insulating oxide layer will result from
its oxidation by the atomic O and other active oxygen species produced in the
source [15].
Steinbruchel [19, 20] has studied the emission of both neutral and ionized
species produced from a Kaufman-type ion source operating with fluoro- and
chlorocarbon species, as well as Ar, Ne, O2, N2, and BCl3. This work involved a
line-of-sight quadrupole mass spectrometer as the detector.
Ensinger [21] has written an excellent review of a broad range of ion source
designs for many purposes.
220 21 Ions
21.6 Focusing of Low-Energy Ions by Magnetic Field
There is some interest in producing surface chemical reactions with low-energy ions
impinging on the surface. In order to do this effectively, ion current densities at very
low energies must be maintained at sufficient levels to carry out surface reactions in
a reasonable time. A method is described in which the efficiency of collection of a
beam of ions with a diameter of about 7 mm is markedly enhanced with an axial
magnetic field [22].
The ions are generated by a Colutron GD-2 ion beam source, and ions are
produced with an energy spread of as low as ±0.1 eV. The ions are extracted,
focused by an Einzel lens, and mass-filtered with a Wien filter. These ions are
directed at an aperture plate of 12-mm thickness containing a 7-mm-diameter hole,
as shown in Fig. 21.6a. A tungsten grid (90 % transparency) is placed across the
aperture plate. The gap between the grid and the sample to be bombarded is kept
between 2 and 5 cm and the ions are decelerated here. The difference between the
beam potential and the positively biased sample defines the ion energy at the sample
surface. A 2000 G Co-Sm permanent magnet (4-cm diameter) is placed behind the
sample at a distance of about 0.5 cm. This magnet can be electrically biased, and
sometimes a magnet potential of +200 V above that of the sample has been found to
increase the measured ion current collected by the sample. This effect may be due to
the collection of secondary electrons from the grid. The 0.5-cm gap is varied during
ion collection to maximize the collected ion current at the sample. The magnet can
also be removed when surface analytical methods influenced by external magnetic
fields are employed. Obviously the sample must be nonmagnetic.
Figure 21.6b shows the effect of magnetic focusing on an Ar+ ion beam incident
on a Pt sample. A dramatic increase in the collected current below an ion energy of
10 eV is noted, where the initial ion current at 1000 eV is 400 nA. The ion beam
energy spread at the sample is about ±0.5 eV. The paper [22] describing this work
contains an analysis of the effect of the cyclotron motion of the ions in the magnetic
field on their divergence. This method is simple and cost effective compared to
electrostatic focusing methods when a small ion beam spot diameter is not a critical
concern.
21.7 Low Energy Ion Gun—Construction

and Performance
The conversion of a conventional high energy ion gun into a low energy gun can be
carried out, preserving the ion current capability as well as the spatial and the
energy resolution capability [23]. Several low energy ion guns have been described
in Sects. 21.1–21.6 of this book, but all requizre the construction of new ion optics.
This design uses existing ion optics and electronic control capabilities of a com-
mercial ion gun. It can be used where it is desired to use low energy ions as reactive
21.7 Low Energy Ion Gun—Construction and Performance 221
Fig. 21.6 Focusing of low-energy ions by magnetic field. a Ion deceleration in magnetic field.
b Ion collection efficiency
species without significant sputter damage at the target. The chemistry measured
will be that of the neutral, ion because of efficient ion neutralization at surfaces.
Figure 21.7a shows a schematic of the original and modified ion guns.
Essentially, the conversion to the low ion energy regime is achieved by biasing the
entire ion gun to about −5 kV. This is achieved by the use of ceramic standoffs to
222 21 Ions
Fig. 21.7 Low energy ion gun. a Modification to the ion gun for low energy application. b Ion
potential energy. c Energy spread. d Ion beam profile
isolate the outer case of the ion gun at its 2¾ in. Conflat flange, and then to float the
gun at high negative potential. Great care should be used to prevent laboratory
workers from making accidental contact with the lethal retarding voltages on the
electrically isolated ion gun. This can be done by covering everything at high
potential with a plexiglass shield.
Figure 21.7b shows the potential energy surface experienced by a positive ion
under conditions when it is desired to deliver the ion to a surface with 10 eV of
kinetic energy. Most of the loss of ion energy occurs near the exit of the ion gun,
but small additional retardation occurs near the front of the sample as shown. In this
particular simulation [24], the sample and the walls of the vacuum are at ground.
The measured beam energy spread as a function of ion energy is shown in
Fig. 21.7c. About 2 eV change in halfwidth of the ion beam is seen over the energy
range 10–50 eV.
21.7 Low Energy Ion Gun—Construction and Performance 223
By moving the sample across the ion beam, an approximate estimate of the beam
profile could be made. Using the % FI potential shown, the F2 % potential is varied.
For F2 from 77.8 to 91.1 %, total beam widths from 3–4 mm are observed. As F2 is
raised to 96.3 % the beam width degrades to 8–9 mm. In addition the maximum ion
current received by the sample decreases at high values of F2. Ion beam profiles are
shown in Fig. 21.7d.
21.8 Measuring Ion Beam Dimensions
The size of ion beams used in research can be measured easily [25] by using the fact
that ion-irradiated Gd2(MoO4)3, called GMO, becomes passivated toward attack by
dilute hydrofluoric acid, HF(aq). This passivation occurs for the c-surfaces (001
plane) of GMO crystals, when irradiated with 1–2 keV He+, Ne+, and Ar+.
The GMO crystal is polished down to a surface finish achieved with 1-μm
alumina abrasive and then may be irradiated with the ion beam as shown in
Fig. 21.8. Following irradiation, the crystal is removed and etched with 1 % HF(aq).
The area exposed to the ion beam is almost unaffected by the HF, whereas the
surrounding areas are etched away. The imaging of the ion beam is reported to have
l-μm lateral resolution, and to be much more sensitive than the competitive tech-
nique in which a Ta2O5 anodic film on Ta is bombarded until metallic Ta is
exposed. The authors warn that the technique is an integrating technique, and the
pattern will therefore reflect any wandering of the ion beam during the irradiation
period. The technique is quite useful for grazing incidence ion beams.
Three more accurate methods for measuring ion beam sizes and the divergence of
an ion beam are given in [26]. Here, a HgCdTe substrate covered with a photoresist
is used in one method, or movable collectors are used in the latter two methods.
Fig. 21.8 Measuring ion beam dimensions

224 21 Ions
21.9 Beam Position Monitor—Charged Particles
The alignment of a beam of ions or electrons may be facilitated by the use of a

segmented aperture device [27] shown in Fig. 21.9. A simple bridge circuit is used,
with a current measuring device to monitor the imbalance in collected current
between opposing segments of the aperture. Only one of the two bridges is shown
in the Figure. Several of these monitors may be placed along the beam axis, and any
ions or electrons scattered by the edges of the apertures are removed by other
apertures further down the beam axis. The variable shunt around the monitor allows
the fraction of the collected current going to the current monitor to be adjusted.
Fig. 21.9 Beam position monitor—charged particles

21.9 Beam Position Monitor—Charged Particles 225
Varying the value of the resistance, R, over a wide range has little effect on the
meter deflection, but R should be kept as small as possible to avoid charge buildup
on the segments that could cause beam aberrations.
21.10 Pulse-Counting LEED/ESDIAD Analyzer—Using

MicroChannel Plate Detection
There are many advantages in pulse counting during the measurement of either
electron diffraction or ion angular distribution patterns produced by electron beam
excitation of surfaces (ESDIAD). Pulse counting affords the maximum sensitivity to
particle measurement, assuming)each pulse is counted. The digitized information
arising from pulse counting allows easy handling in computer analysis and error
analysis. But most importantly, the high sensitivity of pulse counting techniques
allows one to work with electron beams having low intensities, thereby minimizing
damage to the surface layer during measurement [28, 29].
Figure 21.10a shows a LEED/ESDIAD apparatus [30] that uses four grids in
front of a pair of microchannel plate (MCP) amplifiers [31]. Charged particles (and,
in some cases, excited neutral particles) are transmitted through the grid system to
the microchannel plate amplifiers, where each particle exciting the microchannel
plate produces a cascade of electrons that exit the back of the second microchannel
plate and are accelerated into a two-dimensional resistive anode detector [32]. This
pulse of about 105 or 106 electrons expands radially from the point of impact and is
collected at the four corners of the resistive anode. Since the local intensity of
expanding charge decreases as 1/r, the pulse height arriving at the four corners can
be used to calculate the coordinates of the point of impact (the centroid). The pulses
at the four corners are capacitively coupled into four preamplifiers, A-D. These
amplified pulses then travel into the position computer [33]. The position computer
then generates a pair of pulses which are proportional to the x- and y-coordinates of
the impact point on the resistive anode. The x- and y-pulses are then sent to a pair of
analog-to-digital converters that form the input to either a multichannel analyzer [34]
or to a computer. A three-dimensional display of the collected angular distribution of
the electron or ion or excited neutral pattern is then displayed. It is important to point
out that the excitation of the crystal surface by the electron beam will produce a
background due to soft X-rays, and this must be subtracted from the measured
patterns [35, 36]. The ultimate spatial resolution of the apparatus is about 0.3°.
Typical electron beam currents of the order of 1 × 10−9 A are employed over time
periods of the order of 100 s. Damage to even the most electron-beam-sensitive
surface species is generally less than 1 %. The positive ion patterns viewed in this
apparatus may be compressed by applying a positive potential, Vc, to the crystal.
Examples of both typical LEED patterns and ESDIAD patterns are shown in
Fig. 21.10b, c for the case Cu(110) and CO chemisorbed on Cu(110) [37]. An
earlier example of this technology applied to LEED measurements alone can be
found in [38].
226 21 Ions
Fig. 21.10 Pulse counting LEED/ESDIAD. a Pulse counting LEED/ESDIAD analyzer. b Typical
LEED pattern—Cu(110). c Typical ESDIAD pattern—CO/Cu(110)
Earlier examples of the use of microchannel plate amplification for ESDIAD or

LEED measurements involved the production of images on phosphor screens and
either photographic [39–42] or video camera [43–47] data acquisition. The video
camera method is efficient in the sense that a single video camera can be used on
multiple systems.
An alternate high-resolution position detector is the wedge-and-strip anode
[40, 41] that captures the electron pulses on a conductive sheet containing accu-
rately shaped wedge and strip conductive regions. The spatial resolution of these
devices exceeds the inherent resolution limitations of the microchannel plate
amplifiers [40, 41].
A LEED analyzer built along the principle of the MCP/resistive anode combi-
nation has been reported by others [46, 48, 49]. In addition, a gated movable
channeltron detector, coupled with a chopped electron beam, has been used to
reduce background effects [50].
21.10 Pulse-Counting LEED/ESDIAD Analyzer—Using MicroChannel Plate Detection 227
The counting efficiency for ions on a MCP detector has been carefully measured
using a field evaporation apparatus [51]. It was found that the application of a
positive bias on the front surface of the channel plate did not increase the
capture/counting efficiency much above that expected from the fractional area of
open channels present on the plate. It was, however, argued by others that a small
positive potential (20 V) facilitated capture of secondary electrons and hence
enhanced the counting efficiency [52], and this effect was verified in a careful paper
in which the absolute detection efficiencies for He+, H+, and O+ ions was directly
measured using a Faraday cup to compare with the count rate. Ion detection effi-
ciencies ranged from 18 to 63 % in the energy range 250–5000 eV [53]. These
workers also measured the detection efficiency as a function of incident angle,
finding that an incident angle of about 10° yielded maximum efficiency, with the
efficiency falling 24 % in the range 20–40° off the channel axis direction [53].
The basic physical properties of straight through MCP arrays, chevron MCP arrays,
and curved channel MCP arrays are discussed in some detail in [54].
Field penetration effects from one region of the LEED apparatus to another can
be prevented in most cases by using neighboring grids at the same bias potential to
screen elements at high potential (such as the front of the MCP) from the interior of
the grid system [55].
The hemispherical grid analyzer may be used as a retarding field energy analyzer
for either electrons or ions. For these measurements, the retarding potential ramp is
applied to a central grid along with a small sinusoidal modulating voltage, and
lock-in detection is applied to the transmitted current. This has been discussed in
detail in [56].
A planar grid ESDIAD apparatus has been used to monitor the angular distribution
of ions emitted from an insulator surface as a result of excimer laser pulses [57].
21.11 Time-of-Flight Measurements for Ions Generated

in ESDIAD
Section 21.10 describes the design and use of an ESDIAD apparatus in which
positive ions, produced in electron stimulated desorption (ESD), are imaged spa-
tially as they are ejected from surface species. These measurements are used to
determine the chemical bond directions in adsorbed species on single crystal sur-
faces [29] and the dynamical behavior of the molecules as they undergo low fre-
quency motions on their binding sites [58].
Detailed information about desorbing ions can be obtained if their momentum is
also measured, using a pulsed electron beam, and measuring the flight time between
the crystal and the detector. The flight time will directly yield the ion velocity,
except that it is safest to measure changes in flight times with various accelerating
potentials to derive the ion energy and mass (i.e. the momentum). An apparatus [59]
to measure the time-of-flight (TOF) of desorbing ions is shown in Fig. 21.11a. The
apparatus is capable of measuring the desorbing direction and the flight time for
228 21 Ions
Fig. 21.11 Time of flight measurements in ESDIAD. a Schematic circuit for TOF measurement of
positive ions in ESDIAD. b Time sequence of pulses and signals. c ESDIAD analyzer and pulsed
electron beam. d Time-of-flight separation of H+ and D+. e Benzoate on Cu(110). f ESDIAD
patterns of H+ and D+ from isotopic benzoate molecules
21.11 Time-of-Flight Measurements for Ions Generated in ESDIAD 229
each ion comprising the ESDIAD pattern. Time resolution is in the picosecond
range and data acquisition up to 100 kHz is achievable. This enables the decon-
volution of the spatial distribution and the momentum of each ion. The reader is
referred to [59] for details of the electronics shown in Fig. 21.11a.
Figure 21.11b shows the time sequences for various pulses. Initiated by a trigger
signal, an electron gun ejects an electron pulse of variable time width down to 5 ns.
Positive ions A and B arrive at the MCPs at different times. The flight time of an ion
is recorded along with the x and y coordinates of its arrival at the MCP. A drawing
of the hemispherical analyzer and the pulsed electron and ion beams is shown in
Fig. 21.11c. The ions, after being amplified and converted to electrons in the MCPs,
are counted by the resistive anode (RA) which yields the coordinates of ion arrival
there. For the adsorption of the benzoate ion, oriented normally on Cu(110) by
adsorption through the COO group, three hydrogenic ion beams are shown in
Fig. 21.11e corresponding to the 3-, 4-, and 5- positions on the phenyl ring. By
using the TOF method, the two H+ ions corresponding to the 3- and 5- C–H bonds
are measured at 60° to the normal. Using a deuterium label at the 4-position, a D+
beam is observed at the 4-position which produces a normally-oriented beam
[60, 61]. The D+ ion has a longer TOF due to its heavier mass as shown in
Fig. 21.11d. The separate ion angles are plotted in Fig. 21.11f as contour diagrams.
The difference in yield of H+ and D+ is due to several factors related to the departing
speed of the ion which strongly controls its neutralization and to smaller differences
in the rate of the neutralization of the departing ions originating from different
positions on the phenyl ring [61].
21.12 Ion Detection with Kinetic Energy and Mass

Resolution Using a Hemispherical Energy Analyzer
In photon or electron stimulated desorption one may require both mass analysis of
the ions as well as energy resolution. This problem was first addressed by Feulner
[62, 63], using a simple retarding field analyzer coupled to a microchannel plate
(MCP) assembly and readout by a multichannel analyzer (MCA), as shown in
Fig. 21.12a. The macrostructure of the ion peaks yields the masses of the ions in
each peak; the microstructure (peak shape) yields the energy distribution as sche-
matically indicated in the spectrum shown.
A more elaborate analyzer can be made from a hemispherical electron energy
analyzer as shown in Fig. 21.12b. If “single punch (pulsed) light” is not available,
an electrostatic gate is placed in front of the entrance aperture to the hemispherical
analyzer. It is a pair of deflection plates placed inside a screening box. A symmetric
square wave is applied to these deflection plates in order to modulate the trans-
mitted ion beam. The signals from the 5 channeltron multipliers—used for
230 21 Ions
Fig. 21.12 Ion detector—kinetic energy and mass resolution. a Retarding field TOF [58].
b Hemispherical-TOF analyzer
21.12 Ion Detection with Kinetic Energy and Mass Resolution … 231
increased sensitivity—are fed into multi-page, multi-channel scalars that are syn-
chronized with the modulation of the ions. In this manner, 2D fields are obtained
with the axes “kinetic energy” and “flight time”. From this information the kinetic
energy distributions of the individual ions can be extracted [62].
21.13 Time-of-Flight Detection for Laser-Ionized Neutrals
The detection of neutral atoms and molecules that have been generated from sur-
faces using scattering, electron- and photon-stimulated desorption, and ion-induced
desorption is facilitated by tunable nanosecond pulsed lasers. Laser excitation
permits particle detection with quantum state specificity using laser-induced fluo-
rescence (LIF) or resonance-enhanced multiphoton ionization (REMPI) methods.
The REMPI technique has proven to be more popular because of the high sensi-
tivity of the ion detection technique and the ability to image ion angular distribu-
tions analogous to the ESDIAD technique (see p. 225).
The apparatus shown in Fig. 21.13 was designed to detect laser-ionized neutral
atoms and molecules that have undergone electron stimulated desorption from a
single crystal surface [64, 65]. It is based on the design developed by N. Winograd
et al. [66] for the detection of neutral atoms desorbed by ion bombardment. The
Fig. 21.13 Time-of-flight detection for laser-ionized neutrals

232 21 Ions
pulsed electron gun shown in the Figure therefore can be replaced by a pulsed ion
gun or a pulsed laser, and the detection scheme is the same. The principle is as
follows. Neutral products that leave the surface as a result of excitation by the
pulsed source must travel a specified distance before they enter the ionizing laser
beam, which is directed parallel to the crystal surface. The laser generated ions are
then accelerated into the chevron electron multichannel plate array (CEMA)
detector [67]. The output of the detector is coupled to a phosphor screen. Provided
that the laser is focused in a sheet-like fashion, the visual display on the phosphor
screen provides angular resolved patterns of the desorbed neutrals, as in the
ESDIAD method, but now with full internal quantum state resolution [65, 66]. The
integrated signal can be detected with a gated integrator or the two-dimensional
image can be digitized using a charge-coupled device (CCD) camera.
By pulsing the electron beam rapidly (1 μs), the velocities of the neutrals can be
measured by the time delay between the pulses of the electron gun and the laser.
Desorbed ions can be a problem, but they can be swept away prior to the laser pulse
by an applied potential to the first grid and/or the crystal. It is also possible to
discriminate against unwanted signals by pulsing the applied voltage on one of the
multichannel plates. Recently several groups have oriented the detection assembly
90° to the surface normal, so that the spread in velocities is obtained simultaneously
with the angular distribution [68].
References
1. Data sheet for Ion source IQ 10/35, SPECS, GmbH (1996)

2. C.T. Campbell, S.M. Valone, J. Vac. Sci. Technol. A3, 408 (1985)
3. A. Winkler, J.T. Yates Jr, J. Vac. Sci. Technol. A6, 2929 (1989)
4. Design provided by Dr. J. Ahner and D. Mocuta, Surface Science Center, University of
Pittsburgh
5. T. McCabe, J. Somers, M.E. Bridge, D.R. Lloyd, J. Vac. Sci. Technol. A5, 119 (1987)
6. V.M. Bermudez, R.E. Thomas, J. Vac. Sci. Technol. A1, 1557 (1983)
7. J. Kirschner, Rev. Sci. Instrum. 57, 2640 (1986)
8. M. Khan, J.M. Schroeer, Rev. Sci. Instrum. 42, 1348 (1971)
9. A.R. Krauss, Argonne National Laboratory, 9700 South Cass Avenue, Argonne, IL 60439
10. J.F. Mahoney, J. Perel, A.T. Forrester, Appl. Phys. Lett. 38, 320 (1981)
11. The wire bonding tool is available from Technical Devices, Inc., 5721 Prescott Road, Prescott,
AZ 86301
12. C. Kim, D.F. Ogeltree, M.B. Salmeron, X. Godechot, G.A. Somorjai, I.G. Brown, Appl. Surf.
Sci. 59, 261 (1992)
13. H.R. Kaufman, J.J. Cuomo, J.M.E. Harper, J. Vac. Sci. Technol. 21, 725 (1982)
14. J.M.E. Harper, J.J. Cuomo, H.R. Kaufman, J. Vac. Sci. Technol. 21, 737 (1982)
15. C.R. Guarnieri, K.V. Ramanathan, D.S. Yee, J.J. Cuomo, J. Vac. Sci. Technol. A6, 2582
(1988)
16. G.A. Haas, Thermionic emission sources (AD266039 Electron Tubes Branch, Electronic
Division, U.S. Naval Research Laboratory, Washington D.C., 1961), p. 13. (For sale by
Superintendent of Documents, U.S. Government)
17. D.B. Rogers, R.D. Shannon, A.W. Sleight, J.L. Gillson, Inorg. Chem. 8, 841 (1969)
References 233
18. L.F. Mattheiss, Phys. Rev. 181, 987 (1969)

19. Ch. Steinbruchel, J. Vac. Sci. Technol. B2, 38 (1984)
20. Ch. Steinbruchel, J. Appl. Phys. 59, 4151 (1986)
21. W. Ensinger, Rev. Sci. Instrum. 63, 5217 (1992)
22. D.R. Strongin, J.K. Mowlem, K.G. Lynn, Y. Kong, Rev. Sci. Instrum. 63, 175 (1992)
23. I. Popova, R. Muha, Z. Chen, J.T. Yates Jr, J. Vac. Sci. Technol. A21, 401 (2003)
24. SimIon program 7.0
25. A.S. Bhalla, L.L. Tongson, J. Vac. Sci. Technol. 21, 1039 (1982)
26. S. Maniv, Rev. Sci. Instrum. 62, 1179 (1991)
27. K.J. Snowdon, C.H. Barber, J. Phys. E12, 923 (1979)
28. C.G. Pantano, T.E. Madey, Appl. Surf. Sci. 7, 115 (1981)
29. R.D. Ramsier, J.T. Yates Jr, Surf. Sci. Rep. 12, 243 (1991)
30. J.T. Yates, Jr., M.D. Alvey, K.W. Kolasinski, and M.J. Dresser, Nucl. Inst, and Methods in
Phys. Res. B27, 147 (1987). The first apparatus of this sort for low incident current LEED
studies was reported in P.C. Stair, Rev. Sci. Inst. 51, 132 (1980)
31. Galileo Electro-Optics Corp., Galileo Park, P.O. Box 550, Sturbridge, MA 01566
32. No. 3394D Quantar Technology, Inc., 3004 Mission Street, Santa Cruz, CA 95060–95700
33. No. 2401 Quantar Technology, Inc., 3004 Mission Street, Santa Cruz, CA 95060–95700
34. No 76B Canberra Industries, Inc., Nuclear Products Group, 800 Research Parkway, Meriden,
CT 06450
35. M.J. Dresser, M.D. Alvey, J.T. Yates Jr, Surf. Sci. 169, 91 (1986)
36. M.J. Dresser, M.D. Alvey, J.T. Yates Jr, J. Vac. Sci. Technol. A4, 1446 (1986)
37. J. Ahner, D. Mocuta, R.D. Ramsier, J.T. Yates Jr, J. Vac. Sci. Technol. A14, 1583 (1996)
38. P.C. Stair, Rev. Sci. Instrum. 51, 132 (1980)
39. J.J. Czyzewski, T.E. Madey, J.T. Yates Jr, Phys. Rev. Lett. 32, 777 (1974)
41. D.F. Ogletree, G.S. Black-man, R.Q. Hwang, U. Starke, G.A. Somorjai, J.E. Katz, Rev. Sci.
Instrum. 63, 104 (1992)
42. V.M. Bermudez, J. Vac. Sci. Technol. A5, 2975 (1987)
43. T.E. Madey, J. Vac. Sci. Technol. A4, 257 (1986)
44. T.E. Madey, Science 234, 316 (1986)
45. D.F. Ogletree, G.A. Somorjai, J.E. Katz, Rev. Sci. Instrum. 57, 3012 (1986)
46. R.A. Malic, Rev. Sci. Instrum. 59, 1951 (1988)
47. G. Held, S. Uremovic, C. Stellwag, D. Menzel, Rev. Sci. Instrum. 67, 378 (1996)
48. E.G. McRae, R.A. Malic, D.A. Kapilow, Rev. Sci. Instrum. 56, 2077 (1985)
49. F.-K. Men, B.L. Clothier, J.L. Erskine, Rev. Sci. Instrum. 64, 1883 (1993)
50. H. Niehus, B. Krahl-Urban, Rev. Sci. Instrum. 52, 56 (1981)
51. T. Sakurai, T. Hashizume, Rev. Sci. Instrum. 57, 236 (1986)
52. J.A. Panitz, J.A. Foesch, Rev. Sci. Instrum. 47, 44 (1976)
53. R.S. Gao, P.S. Gibner, J.H. Newman, K.A. Smith, R.F. Stebbings, Rev. Sci. Instrum. 55, 1756
(1984)
54. J.G. Timothy, Rev. Sci. Instrum. 52, 1131 (1981)
55. C.W. Caldwell Jr, Rev. Sci. Instrum. 36, 1500 (1965)
56. N.J. Taylor, Rev. Sci. Instrum. 40, 792 (1969)
57. J.T. Dickinson, S.C. Langford, J.J. Shin, D.L. Doering, Phys. Rev. Lett. 73, 2630 (1994)
58. J.T. Yates Jr, J. Ahner, D. Mocuta, Proc. Nat. Acad. Sci. 95, 443 (1998)
59. J. Ahner, D. Mocuta, J.T. Yates Jr, J. Vac. Sci. Technol., A 17, 2333 (1999)
60. J. Lee, O. Kuzmych, J.T. Yates Jr, Surf. Sci. 582, 117 (2005)
61. J. Lee, I.A. Balabin, D.N. Baratan, J.-G. Lee, J.T. Yates Jr, Chem. Phys. Lett. 412, 171 (2005)
62. R. Weimar, R. Romberg, S.P. Frigo, B. Kassühlke, P. Feulner, Surf. Sci. 451, 124 (2000)
63. P. Feulner, private communication
64. A.R. Burns, E.B. Stechel, D.R. Jennison, in Desorption Induced by Electronic Transitions,
DIET III, ed. by R.H. Stuhlen, M.L. Knotek (Springer, Heidelberg, 1988), p. 67
65. A.R. Burns, Surf. Sci. 280, 349 (1993)
234 21 Ions
66. R.H. Kobrin, G.A. Schick, J.P. Baxter, N. Winograd, Rev. Sci. Instrum. 57, 1354 (1986)
67. CEMA assemblies are available from Galileo Electro-Optics Corp., Galileo Park, P.O. Box
550, Sturbridge, MA 01566
68. M. Menges, B. Baumeister, K. Al-Shamery, B. Adam, T. Mull, H.-J. Freund et al.,
in Desorption Induced by Electronic Transitions, DIET V, ed. by A.R. Burns, E.B. Stechel,
D.R. Jennison (Springer, New York, 1993), p. 275
Chapter 22
Photons (UV)
22.1 Hollow Cathode Resonance Lamp for He(II)

(40.8 eV) Photoemission
The use of vacuum ultraviolet sources for valence band photoemission studies of the
surfaces of solids is widespread, and many such studies are made on synchrotron
sources. Small laboratory vacuum UV sources predated synchrotron sources, and are
still widely used. A high-intensity source delivering both He(I) and He(II) radiation,
depending upon the discharge conditions, is shown schematically in Fig. 22.1a [1].
He gas is admitted to a differentially pumped region containing a hollow cathode,
and is pumped away continuously by a trapped forepump. A second stage of dif-
ferential pumping occurs closer to the ultrahigh vacuum chamber using a diffusion
pump that is backed by the same forepump. The forepump (rated at 500 L/min)
speed was measured to be 100–200 L/min at the lamp and 30–70 L/min at the back
of the diffusion pump for typical operation [2]. The forepump was trapped with a
water-cooled Cu trap and a zeolite trap. An 8-cm-diameter port at the lamp leading to
the liquid nitrogen trap before the diffusion pump was used [1], as an aid to improve
pumping speed over that originally used [2]. This large port was made from a
straight-through valve. This permits the lamp to be isolated from the UHV system.
Figure 22.1b shows the details of the He lamp. A high-voltage UHV feed through
holds a hollow cathode cylinder made of 0.5 mm Ta sheet. This cathode, 2 cm in
diameter and 4–6 cm in length, is held inside a glass tube in a location in which a
stainless steel capillary (2.3-mm internal diameter) picks up radiation at a distance of
1–1.5 cm from the end of the cathode. The glass discharge region can be removed for
cleaning. A needle valve admits He into the side of a Conflat flange that mates to the
flange holding the first stainless capillary tube. Below the first capillary is the region
pumped by the forepump, region A-B. A second capillary, mounted between B and
C, directs the radiation to a third capillary which enters the UHV chamber. The He
pressure in the system with the lamp operating is approximately 1 × 10−8 Torr.
Contaminant gases are measured to be below 10−10 Torr, although such gases will be

DOI 10.1007/978-3-319-17668-0_22
236 22 Photons (UV)
Fig. 22.1 Hollow Cathode resonance lamp—He(II) source. a Pumping system. b VUV lamp.
c Au UPS spectrum
highly directed as a beam to the sample by the capillaries, and one should be alert to
the possible build up of surface contamination.
The performance of the lamp for both He(I) (21.2 eV) and He(II) (40.8 eV), and
for 48.4 eV radiation, is shown in Fig. 22.1c, where an evaporated Au film was
studied, using a double-pass CMA electron spectrometer for measurement of the
photoelectron spectrum at a constant resolution of 0.2 eV. The pressure of He in the
discharge was about 1 Torr for optimum He(I) intensity and 0.1–0.2 Torr for
40.8 eV and 48.4 eV photons. This lamp has improved He(II) intensity compared to
earlier lamps utilizing a microwave discharge [3].
22.2 He(I) and He(II) Ultraviolet Resonance Source
A laboratory source of vacuum ultraviolet radiation useful for photoelectron

spectroscopy in an ultrahigh vacuum chamber is described [4]. The source is based
on the principle of the charged particle oscillator [5–8]. The production of maxi-
mum He(I) (21.22 eV) and He(II) (40.81 eV) radiation occurs at a discharge
22.2 He(I) and He(II) Ultraviolet Resonance Source 237
Fig. 22.2 He(I) and He(II) UV source. a Cold cathode discharge UV source. b Differential
pumping system
pressure of He of about 8 × 10−2 Torr. By means of two stages of differential

pumping in the source, the pressure in the main UHV chamber can be maintained at
2 × 10−10 Torr during operation.
Figure 22.2a shows the UV source. It consists of a grounded outer cathode
cylinder made of 304 stainless steel. Two 3/16-in.-diameter tungsten rods serve as
the anode, and they are supported and separated from the outer cathode by
machinable ceramic fittings. The W rods can expand and slide in loosely fitting
holes in the ceramic as the lamp heats during operation. The ceramic also contains
larger holes for conductance of He. The anode potential of 0 to +5 kV is supplied
through a high-voltage feedthrough at the rear of the source. Grounded stainless
steel end plates cover the inner side of the ceramic fittings to reflect electrons
traveling in the axial direction back into the source volume. Photons emerge
238 22 Photons (UV)
through a 0.1-in.-diameter exit aperture between the anodes at the front of the
lamp. An aperture with a diameter less than one-third the inner diameter of the
cathode allows sampling of the radiation without allowing electrons to escape [9].
The gas used here was 99.9995 % He, which flows through an activated charcoal
trap at 77 K and a variable leak valve. A 0 to +5 kV DC power supply (0.500 mA)
is connected through a variable resistance ballast set at 16 kΩ. This prevents current
surges during initiation of the discharge whose resistance varies from infinity
(before activation) to several kΩ during operation.
The pumping system in the lamp consists of two stages of differential pumping,
followed by a third stage in the UHV chamber. The first stage, across a 1.5-mm-i.d.
Pyrex capillary, 3 cm long, is pumped by a chilled water-trapped 325-L/s oil
diffusion pump. This stage can be eliminated by opening a bypass valve between
entrance region and the first stage of diffusion pumping. The second stage is also
across a similar capillary and is pumped by a liquid-nitrogen-trapped 325-L/s oil
diffusion pump. The third stage of pumping across the exit capillary (2.5 mm i.d,
22.86 cm long) is pumped by the UHV system pumps, which consist of a 450-L/s
turbopump and a 500-L/s ion pump. A gate valve is used to isolate the source from
the UHV system. The original paper shows plots of photoelectron intensity from a
Au surface versus pressure [4].
Under optimal operating conditions, the He(I)/He(II) intensity ratio produces a
10/1 ratio of photoelectrons from the two spectral lines from the Au surface.
22.3 Photochemistry on Surfaces Using an Ultraviolet

Lamp
Although ultraviolet induced photochemistry research has been carried out for more
than 60 years in homogeneous phase, detailed studies of photochemical processes
on well-defined single crystal surfaces are much more recent. Because ultraviolet
lamps emit more photons over an extended exposure time than most laboratory
lasers, these lamps are often used for inducing photochemical processes on sur-
faces. Both intensity measurements and photon energy range selection using filters
are conveniently carried out, permitting good measurements of cross sections
versus photon energy for surface process induced photolytically.
Figure 22.3a shows an ultrahigh vacuum chamber [10] set up so that a single
crystal may be irradiated with ultraviolet light from a high-pressure Hg arc source
[11]. The single crystal is located in front of an aperture that leads to a quadrupole
mass spectrometer in a differentially pumped section, permitting the direct mea-
surement of photodesorbed species. In addition, the mass spectrometer maybe used
also for thermal desorption studies after ultraviolet irradiation to access the changes
that have occurred due to irradiation of an adsorbed layer. The ultraviolet light
passes through a 10-cm distilled water filter that filters out infrared radiation.
A 200-mm f/1 quartz condensing lens focuses the light onto the crystal through a
UV-grade sapphire window. Interference filters (Oriel Corporation) centered on the
22.3 Photochemistry on Surfaces Using an Ultraviolet Lamp 239
Fig. 22.3 Photochemistry using ultraviolet lamp. a Ultraviolet lamp for photochemistry using
filters and polarizer. b Typical cross section measurements from TPD analysis
240 22 Photons (UV)
intense lines of the mercury arc are often used to produce narrowband radiation and
have a transmission of from 15 to 50 % at the center of the band with a full width at
half maximum (FWHM) of 10 nm, except for the 237-nm filter, which has a
FWHM of 20 nm. One must be careful in using these filters for long irradiation
times, since the transmission band center can shift up to 2 nm when extensive
irradiation occurs.
The apparatus contains a quartz flat partial reflector used to measure continu-
ously the light intensity. The quartz flat reflects roughly 10 % of the light, and this is
directed to a neutral density filter (1 % transmission) and then into a blackened
housing which holds a calibrated photodiode [12]. The calibration of the photo-
diode relative to an absolute thermopile detector can be done externally on an
optical bench, with the thermopile being placed in the same position as would be
occupied by the crystal. Such measurements have found that the intensity of UV
radiation from the arc source is not uniform, and it is interesting to map out the
intensity pattern which gives a view of the arc location in the source in the most
intense region of emission.
A commercially available UV quartz prism polarizer may be placed into the
beam if either s- or p-polarized radiation is desired (see p. 240).
By combining the photoirradiation experiments at various photon energies with
analysis of the surface before and after irradiation (using thermal desorption in this
case) it was possible to observe two major separate photochemical processes for
molecular oxygen chemisorbed on Pd(111)-desorption of O2 and dissociation to
produce chemisorbed O. The cross sections for these two processes are plotted
versus photon energy in Fig. 22.3b, giving an idea of the random errors inherent in
this type of measurement [10]. These cross sections were also measured in another
laboratory using laser-induced photochemistry and the results were in agreement to
within a factor of 2–3 [13].
A photoreactor for studies of powders and single crystals has been described
in [14]. This device can operate at high or low gas pressures.
22.4 Use of Polarized Ultraviolet Light for Photochemistry
The use of polarized UV radiation to induce photochemical processes on surface is

of importance because of the ability to use this information to obtain information
about the mechanism of the electronic processes responsible for photoactivation of
the surface [15, 16]. The photochemical activation of chemisorbed molecular O2 is
a model system, discussed briefly on p. 238, and here we show how polarized
radiation may be used for studies of the same system.
Figure 22.4a shows the configuration of the photochemical system [17].
Ultraviolet light, originating from a lamp source, may be polarized by using a
UV-grade quartz prism polarizer, directing either p- or s-polarized radiation to the
single crystal surface. The single crystal may be used as a “calorimeter” to measure
the thermal energy deposited from the radiation incident on its surface. This
22.4 Use of Polarized Ultraviolet Light for Photochemistry 241
Fig. 22.4 Measurements
with polarized UV light.
a Configuration for irradiation
with polarized UV light.
b “Calorimetric”
measurement for p- and
s-polarization. c Effect of
polarization on oxygen
desorption state distribution
242 22 Photons (UV)
absorbed energy will vary in accordance with Fresnel’s laws and the optical con-
stants of the crystal absorber. Figure 22.4b shows a calorimetric experiment in
which polarized broad band radiation having 0.23 W/cm2 power density is allowed
to warm the crystal from 79 K. For an incidence angle of 60°, one can see that the
p-polarized radiation is more efficiently absorbed than the s-polarized radiation, as
indicated by the relative increases in temperature. The measured difference in
absorptivity is in fair agreement with the optical constants considering the errors
inherent in this type of experiment [17].
Figure 22.4c shows the comparative effect of s- and p-polarized radiation on a
molecular O2 layer on Pd(111) as studied by thermal desorption following various
periods of irradiation. The α1- and α2-molecular oxygen states desorb by phot-
ostimulation, and also photochemically dissociate to produce a thermal desorption
feature labeled β2. It may be seen that p-polarized radiation is more efficient in
causing these desorption and dissociation processes than is s-polarized radiation
[17]. This is consistent with a process in which electronic excitation of the metal by
ultraviolet light is responsible for the surface photochemical processes, but more
detailed studies would be needed to prove this.
22.5 Photoreduction of Ag on Doped TiO2—Making

Enhanced Photoactivity Visible
While many chemical methods may be used to witness photoactivity and its
enhancement on semiconductors, the production of clusters of Ag by photoreduction
of Ag+ ions in solution to produce Ag0 clusters is novel in its ability to statistically
witness the reaction, using AFM to measure the cluster density and size [18–21].
As shown in Fig. 22.5a, slices of a TiO2(110) crystal, glued to a microscope
slide, are immersed in a 0.1 M AgNO3(aq) solution. Exposure to about 1017
photons/cm2 causes photo-produced electrons in the TiO2 to reduce Ag+ ions to
Ag0 atoms which cluster together to produce AFM-visible deposits. Figure 22.5b
shows an AFM image of the undoped TiO2 compared to the doped TiO2 and the
enhanced density of silver clusters clearly indicates that doping was effective in
enhancing the frequency of Ag+ reduction events. The volume of the Ag clusters
produced allows an estimate of the quantum yield for the process, and Fig. 22.5c
shows a plot, versus photon energy, of the quantum yield which falls in the 1 %
range. In the case shown, doping was done by treating the TiO2(110) crystal in
NH3(g) at 1 atm and 870 K for 5 min. Initial heating and cooling of the crystal to
870 K was achieved in Ar(g) at 1 atm.
Figure 22.5d shows the N(1s) XPS spectrum achieved after doping. Spectrum a
corresponds to the removal of about 5 Å of the surface by sputtering in order to
achieve a clean surface after handling in air and NH3. After the removal of 60 Å of
material, two N(1s) features are observed. It was shown that the 399.6 eV feature,
due probably to N-H type species, is active for doping, giving a photon threshold
energy at about 2.4 eV which is 0.6 eV below the bandgap of rutile TiO2 [21].
22.6 Reducing Radiation Damage in Raman Spectroscopy of Solids 243
Fig. 22.5 Photoreduction of Ag on Doped TiO2—making enhanced photoactivity visible.

a Photoreduction of Ag on doped TiO2. b Ag photoreduction on TiO2; hν = 2.47 eV. c Action
curve for photoreduced Ag-deposition on TiO2(110). d N(1s) XPS spectra for NH3-treated
TiO2(110)
22.6 Reducing Radiation Damage in Raman Spectroscopy

of Solids
A serious problem often encountered in measuring the Raman spectrum of a solid is

the effect of the high power laser on damaging the sample [22]. A standard
approach to eliminating or minimizing the effect is to rotate the sample while it is
under study, thereby distributing the excitation over the sample and reducing the
local heating effect [23]. The method reported here is able to do this with a static
sample. Such sample holders are often more convenient when studying high area
solids which cannot be rotated because of other types of measurements which must
be made under static conditions [24].
The unique feature of the mounting procedure is the use of a tungsten mesh grid
to hold the powdered sample. This mesh is photoetched in 0.025 mm thick tungsten
foil and contains regular square openings of 0.22 × 0.22 mm2 area. The squares are
bonded by tungsten borders of 0.038 mm width. The powder is mechanically
pressed as described by Ballinger [25]. This method of holding powdered samples
244 22 Photons (UV)
has been widely used in IR spectroscopy as discussed elsewhere in this book. The
grid is held in a square frame with a 1.4 × 1.4 cm2 opening and consists (for
strength) of a front frame made of stainless steel (1.5 mm thick) and a rear Cu
(OFHC) frame for maximum thermal conductivity, which is brazed to Ni leads. Six
bolts and nuts clamp the frame to the grid. The sample is mounted in a Pyrex
Raman cell with provisions for cooling with liquid nitrogen.
The arrangement was compared to a conventional rotating Raman sample holder,
using an OMARS 89 triple-monochromator spectrometer (Dilor, Lille, France)
equipped with a thermoelectrically cooled CCD camera (Princeton Instruments).
A test was carried out with a WOx/ZrO2 catalyst. The Raman spectrum was recorded
with a rotating holder as shown in Fig. 22.6b, using three laser power levels from a
488-nm Ar+ ion laser (Spectra Physics, Type 2020). Since the spectrum does not
change during irradiation at any level, the rotating sample holder is able to prevent
radiation damage. In Fig. 22.6c, radiation damage is not seen for the mesh-supported
sample until the power is increased above 90 mW, where a new band at 1015 nm
appears. An Os3(CO)12/SiO2 was studied also, and the cryogenically-cooled sample
was stable to 24 mW power. These studies show that non-rotating cryogenically
cooled powdered samples held in a grid may be stabilized against laser damage, but
that the rotating sample method is superior for reducing damage [24].
Fig. 22.6 Reducing radiation damage in Raman spectroscopy of solids. a Cryogenically cooled
non-rotating Raman cell. b, c Raman spectrum—WOx/ZrO2
22.7 Lyman-α Lamp Calibration Using a Chemical Actinometer 245
22.7 Lyman-α Lamp Calibration Using a Chemical

Actinometer
The measurement of the light flux from vacuum UV lamps is important in a number
of areas of photochemistry such as astrochemistry and atmospheric chemistry.
Different methods are used to calibrate such lamps including photoelectron gen-
eration, photoionization of gases, and photodecomposition of gases [26–31]. There
are only a few reports of the use of chemical actinometry for the calibration of
Lyman-α lamps [29–36] (which operate at 10.2 eV). The results of a variety of
calibrations of these lamps are tabulated in [37] and show typical light fluxes
measured over a 47 year history. It can be seen that the various Lyman-α lamps
used in the past have been shown by various methods to deliver approximately
1015 photons s−1 cm−2 over a range of approximately 10-fold.
A chemical actinometer successfully used recently for the calibration involves
the photodecomposition of N2O and the use of transmission IR spectroscopy to
measure the N2O destruction as well as the production NO and NO2 [37]. N2O
decomposes when illuminated by Lyman-α radiation according to the reactions
(22.1)–(22.4):
N2 O þ hm ! N2 þ O ð22:1Þ
N2 O þ hm ! NO þ N ð22:2Þ
1=2O þ 1=2N2 O ! 1=2N2 þ 1=2O2 ð22:3Þ
1=2O þ 1=2N2 O ! NO ð22:4Þ
The quartz Lyman-α lamp was mounted with a Viton O-ring onto a stainless
steel Conflat flange containing a MgF2 window. The lamp was supplied with a
continuous flow of a 10 % H2/90 % Ar mixture at 1 Torr pressure and was excited
by a 60 W microwave supply. The MgF2 effectively cuts off radiation at wave-
lengths above 1100 Å, thereby removing Ar emission [38]. The vacuum UV
radiation passes into a bakeable high vacuum stainless steel transmission IR cell
(Volume = 1.9 L) equipped with turbopumping, Baratron capacitance manometer
pressure gauges, and a mass spectrometer, and having a base pressure of about
1 × 10−8 Torr as shown in Fig. 22.7a. For the lamp calibration, between 1 and
10 Torr of N2O was admitted to the cell and the lamp was turned on and stabilized.
The IR spectrum for N2O and the products during photolysis is shown in Fig. 22.7b.
The rate of consumption of N2O was measured by following the integrated
absorbance (at 2 cm−1 resolution) at two strong fundamental frequencies, ν1 and ν3,
where the IR beam travelled transversely through the cell. Separate pressure versus
integrated absorbance curves were measured for N2O at these frequencies, and the
consumption of N2O, calculated in Torr s−1, was determined. Figure 22.7c shows a
sample of the N2O depletion curves where about 10 % consumption of N2O was
measured, yielding linear plots. It may be seen that the values of dP/dt, measured
246 22 Photons (UV)
Fig. 22.7 Calibration apparatus for N2O photodecomposition by Lyman-α radiation. a FTIR cell
and Lyman-α source. b Spectral evolution of N2O, NO and NO2 features during the
photodecomposition of N2O. c Determination of photon flux
22.7 Lyman-α Lamp Calibration Using a Chemical Actinometer 247
from ν1 and ν3, are in excellent agreement, with less than 1 % difference. The flux
of the Lyman-α light was estimated using a measured efficiency from the literature
[39] of 0.69 ± 0.02 photons/N2O. One should not employ pressure changes in
estimating the light flux as heating effects were found to confuse such measure-
ments. It is interesting to recognize that in the pressure region from 1 to 10 Torr, all
Lyman-α radiation is absorbed in the first few cm from the MgF2 window and the
assumption of fast mixing of the gas was found to hold. In addition, a slight
pressure dependence of the photodecomposition kinetics was found and attributed
to wall effects which become more significant at low pressure [37].
22.8 Pulsed Vacuum UV Light Source
Photochemistry studies are being extended to higher and higher photon energies
because of scientific interest in processes involving deep ionization in molecules, in
solids and on surfaces. A source of radiation of this energy range can be UV lamps,
such as the Lyman-α lamp (10.2 eV), and various other gas lamps excited by
radiofrequency radiation (see Sect. 22.7). Another source of such radiation is syn-
chrotron radiation which extends far into the X-ray region, and which is commonly
monochromatized for use in photoelectron spectroscopy. A laser plasma light source
is another source of Vacuum UV (VUV) or Extreme UV (EUV) light. In the example
shown here, a Ta plasma is generated by a 1064 nm pulsed laser and light is
monochromatized by toroidal gratings in the vacuum spectrometer for the range 4–
108 nm (310–11.5 eV). The apparatus is more than 3 m in length and is pumped by
turbomolecular pumps (TMP) as well as Ti sublimation pumps (TSP) as shown.
As shown in the data below the spectra in Fig. 22.8c, 106 to 108 photons per
pulse are generated which over the duty cycle yields 5 × 107 to 5 × 109 photons/s
averaged over a long time period [40]. This is a far smaller number of photons than
can be generated by a gas discharge lamp such as a Lyman-α lamp, where over the
whole spectrum in the region of 10.2 eV, about 1015 photons/s are generated [37].
22.9 Grazing Incidence X-Ray Photoemission

Spectroscopy (GIXPS): Enhanced Surface Sensitivity
GIXPS was first described by Henke [41] and has seen limited use. The technique is
based upon the angle-dependent X-ray penetration depth at small incidence angles
(measured from grazing angle). X-rays that impinge on a flat surface at *1 keV at
sufficiently small angles undergo total external reflections and their penetration
depth is strongly attenuated. This means that the photoelectron yield contains an
increased contribution from the surface of the solid. As the angle of the X-rays
increases, increased contributions from deeper layers are observed. Therefore, depth
dependent XPS analysis is possible. The theory behind GIXPS is given in [42].
248 22 Photons (UV)
Fig. 22.8 Pulsed vacuum UV light source. a Top view. b Side view. c Intensity (photon/pulse)
Figure 22.9a shows a laboratory design for making GIXPS measurements, using
a small X-ray source. The X-ray source may be rotated in angle of incidence and the
center of rotation is the analysis position on the sample. The microchannel plate
amplifier is used to measure the reflected X-ray radiation. A double pass cylindrical
mirror electron analyzer is used. Normalized sample spectra for oxidized and
unoxidized GaAs(100) are shown in Fig. 22.9b, and it may be seen that at higher
angles (66 mR) the relative contribution from the bulk—derived As0(3d) feature
(binding energy = 41 eV) is enhanced compared to As5+ and As3+ which come from
the surface oxide film and overlap near a binding energy = 45 eV [42]. A patent has
been issued [43].
Figure 22.9c shows a schematic for the performance of GIXPS using synchro-
tron excitation. Here, the sample is rotated in the X-ray beam, maintaining the
22.9 Grazing Incidence X-Ray Photoemission … 249
Fig. 22.9 Grazing X-ray photoemission spectroscopy—enhanced surface sensitivity.

a Laboratory-based GIXPS. b As (3d) GIXPS spectra for oxidized GaAs(100).
c Synchrotron-based GIXPS
250 22 Photons (UV)
origin of photoelectrons at a constant position with respect to the cylindrical mirror

electron analyzer. This work has been focused on measuring the thickness and
chemistry of ultrathin gate oxide films on Si [44].
References
1. J.E. Rowe, S.B. Christman, E.E. Chaban, Rev. Sci. Instrum. 56, 759 (1985)
2. J.E. Rowe, S.B. Christman, E.E. Chaban, Rev. Sci. Instrum. 44, 1675 (1973)
3. T.V. Vorburger, B.J. Waclawski, D.R. Sandstrom, Rev. Sci. Instrum. 47, 501 (1976)
4. G.M. Lancaster, J.A. Taylor, A. Ignatiev, J.W. Rabalais, J. Electron Spectrosc. Relat. Phenom.
14, 143 (1978)
5. A.H. Mcllraith, Nature 212, 1422 (1966)
6. A.H. Mcllraith, J. Vac. Sci. Technol. 9, 209 (1972)
7. R.K. Fitch, T. Mulvey, W.J. Thatcher, A.H. Mcllraith, J. Phys. D3, 1399 (1970)
8. R.K. Fitch, T. Mulvey, W.J. Thatcher, A.H. Mcllraith, J. Phys. E4, 533 (1971)
9. A.H. Mcllraith, private communication to J.W. Rabalais
10. L. Hanley, X.-C. Guo, J.T. Yates Jr., J. Chem. Phys. 91, 7220 (1989)
11. Oriel, 250 Long Beach Boulevard, P.O. Box 872, Stratford, CT 06497-0872 (Model 66014,
350 watt)
12. The l-cm2-area circular photodiode-preamplifier package is sensitive from 185 to 1150 nm,
and is calibrated by the manufacturer with standards traceable back to the National Institute of
Science and Technology (EG and G HUV-4000B)
13. F. Weik, A. de Meijere, E. Hasselbrink, J. Chem. Phys. 99, 682 (1993)
14. A.Z. Moshfegh, A. Ignatiev, Rev. Sci. Instrum. 59, 2202 (1988)
15. S.A. Buntin, L.J. Richter, D.S. King, R.R. Cavanagh, J. Chem. Phys. 91, 6429 (1989)
16. E. Hasselbrink, S. Nettesheim, M. Wolf, A. Cassuto, G. Ertl, in Springer Series in Surface
Science, vol. 19, DIET IV (1990), p. 75
17. J. Yoshinobu, X.-C. Guo, J.T. Yates Jr., Chem. Phys. Lett. 169, 209 (1990)
18. P.D. Fleischauer, H.K.A. Kahn, J.R. Shepherd, J. Am. Chem. Soc. 94, 283 (1972)
19. P.A.M. Hotsenpiller, J.D. Bolt, W.E. Farneth, J.B. Lowekamp, G.S. Rohrer, J. Phys. Chem.
B 102, 3216 (1998)
20. J.B. Lowekamp, G.S. Rohrer, P.A.M. Hotsenpiller, J.D. Bolt, W.E. Farneth, J. Phys. Chem.
B 102, 7323 (1998)
21. O. Diwald, T.L. Thompson, T. Zubkov, E.G. Goralski, S.D. Walck, J.T. Yates, Jr., J. Phys.
Chem. B 108, 6004 (2004)
22. J.M. Stencel, Raman Spectroscopy for Catalysis (Van Nostrand Reinhold, New York, 1990),
p. 32
23. S. Chan, A.T. Bell, J. Catal. 89, 433 (1984)
24. E.A. Wovchko, J.T. Yates Jr., M. Scheithauer, H. Knozinger, Langmuir 14, 552 (1998)
25. T.H. Ballinger, J.C.S. Wong, J.T. Yates Jr., Langmuir 8, 1676 (1992)
26. T.A. Chubb, H. Friedman, Rev. Sci. Instrum. 26, 493 (1955)
27. J.J. Bollinger, F.M. Pipkin, Rev. Sci. Instrum. 52, 938 (1981)
28. K. Watanabe, F.F. Marmo, E.C.Y. Inn, Phys. Rev. 91, 1155 (1953)
29. H. Cottin, M.H. Moore, Y.B. Nilan, J. Astrophys. 590, 874 (2003)
30. M.S. Westley, R.A. Baragiola, R.E. Johnson, G.A. Baratta, Nature 373, 405 (1995)
31. P. Warneck, Appl. Opt. 1, 721 (1962)
32. H. Okabe, J. Opt. Soc. Am. 54, 478 (1964)
33. D. Davis, W. Braun, Appl. Opt. 7, 2071 (1968)
34. I. Yamashita, Jpn. J. Appl. Phys. 14, 70 (1975)
35. P. Weber, J.M. Greenburg, Nature 316, 403 (1985)
36. G.A. Baratta, G. Leto, M.E. Palumbo, Astron. Astrophys. 384, 343 (2002)
References 251
37. M. Rajappan, M. Buttner, C. Cox, J.T. Yates Jr., J. Phys. Chem. A 114, 3443 (2010)
38. A. Duncanson, R.W.H. Stevenson, Proc. Phys. Soc., London 72, 1001 (1958)
39. W.E. Groth, W.E. Schierholz, Planet. Space Sci. 1, 333 (1959)
40. I. Arakawa, private communication
41. B.L. Henke, Phys. Rev. A 6, 94 (1972)
42. M.A. Chester, T. Jach, J. Vac. Sci. Technol., B 11, 1609 (1993)
43. Jach et al., United States Patent 5,280,176 (1994)
44. T. Jach, E. Landree, J. Surf. Anal. 9, 339 (2002)
Chapter 23
Kinetics of Adsorption/Adsorbate
Coverages
23.1 Absolute Adsorption Uptake Measurements

from a Calibrated Effusive Molecular Beam
The measurement of the kinetics of adsorption and the absolute saturation coverage
of an adsorbate on a single crystal is basic to our understanding of surface chem-
istry. While relative surface coverages, compared to the saturation coverage, are
fairly easy to measure using various surface spectroscopies, the measurement of the
absolute coverage and the absolute sticking coefficient are more difficult. One of the
absolute methods employed in some cases involves the use of nuclear activation of
the adsorbed layer, where absolute counting of nuclear disintegration events after
activation may be used with high accuracy [1]. Unfortunately, not all adsorbates are
amenable to this method.
Figure 23.1a shows an apparatus that is capable of measuring both the absolute
sticking coefficient and the absolute surface coverage during an adsorption experi-
ment. The method is a version of the King and Wells method, developed in 1972 [2].
A multicapillary doser, with an absolutely known flux of effusing molecules in a
known angular distribution (see p. 421) is incident on a single crystal of known size,
located at a known distance from the doser. A quadrupole mass spectrometer,
shielded from direct line of sight from the crystal and doser, measures the random
flux of gas in the vacuum system as a function of the exposure of the crystal to the
incident beam. A shutter can be rotated to intercept the beam, shielding the crystal
from the direct flux, and converting all of the incident beam into a random flux by
reflection from its surface. In order to make an uptake measurement, the crystal is
cleaned, brought to adsorption temperature, and positioned behind the shutter. Gas is
admitted, as shown in Fig. 23.1b, and the system pressure rises quickly to a sharp
point where the rate of rise decreases significantly. At time t1, the shutter is opened,
and gas begins to be adsorbed on the crystal, causing the random flux of gas to
decrease. Region A of the uptake curve represents the first phase of adsorption, and
the horizontal trace indicates that the sticking coefficient is constant (and probably

DOI 10.1007/978-3-319-17668-0_23
254 23 Kinetics of Adsorption/Adsorbate Coverages
Fig. 23.1 Absolute adsorption uptake measurement from calibrated beam. a Apparatus for
absolute adsorption uptake measurement. b Typical uptake curve C2H2/Si(100). c Increase in C2H2
coverage beyond tb owing to continued slow adsorption
23.1 Absolute Adsorption Uptake Measurements … 255
unity) in this region. Following adsorption to point ta, it is observed that the random
flux begins to rise in region B as the sticking coefficient decreases below its value in
region A. At the end of the experiment, the random flux has asymptotically risen to
the dashed line, which indicates that, at the sensitivity level of this experiment, the
uptake is complete. Placing the shutter into the beam causes little effect at point tb,
since the complete reflection of incident molecules by the inert shutter is identical in
magnitude to the reflection of incident molecules by the saturated crystal. The area of
region A and region B may be used to compute the final absolute coverage at tb [3].
There are three issues that contribute to the uncertainty of this measurement:
1. The nonhorizontal character of the dashed line against which the integrals are
measured. This error is probably due to slow pumping by the vacuum system
walls, which decreases as the exposure is increased. The empirical solution to
this problem is to treat the dashed limiting line as a baseline; it is designated Peq.
2. The adsorption of gas from the random flux. A small correction can be made for
this effect, which typically amounts to 1 % in the region from t0 to t1. This initial
coverage is designated N(t)initial.
3. The low accuracy of determining the saturation point, tb. Here, Auger spec-
troscopy was employed to show that for long exposures beyond tb, the rise in
adsorbate coverage was 6 % as shown in Fig. 23.1c for the example chosen here.
The equation that governs the adsorption is
Z
NðtÞtotal ¼ NðtÞinitial þ ðF þ Fr Þ SðtÞdt; ð23:1Þ
where N(t)total = the total coverage at saturation; F = the beam flux; Fr = the random
flux; and S(t) = the sticking coefficient. S(t) = [Peq − P(t)]/fPeq and f = the fraction of
the beam intercepted by the crystal. The value N(t)total is then increased to account
for the small correction mentioned above in issue 3.
A variation on this technique uses a diffuse source of adsorbate from an
uncalibrated leak valve and a collimated detector used to measure the rate of
adsorbate reflection from the surface during adsorption [4].
23.2 Measurement of Absolute Sticking Coefficient

at Desorption Temperature
Conventional measurement techniques measure the net sticking probability, which

is related to the difference between the absolute rate of adsorption, ra, and the
absolute rate of desorption, rd, according to (23.2).
snet ¼ ðra rd Þ=z ð23:2Þ
Here z = the impingement rate for the incident gas.

256 23 Kinetics of Adsorption/Adsorbate Coverages
It is often desired to measure the absolute sticking coefficient, sabs, which is

given by (23.3).
sabs ¼ ra =z ð23:3Þ
The absolute sticking coefficient is equal to the net sticking coefficient only
when the rate of desorption is zero, so in order to achieve this condition, con-
ventional sticking coefficient measurements are made at temperatures below the
desorption temperature.
Substituting (23.2) into (23.3) gives (23.4).
sabs ¼ snet þ rd =z ð23:4Þ
Equation (23.4) yields
sabs ¼ lim snet ðzÞ ð23:5Þ

1=z!0
Normally, a plot of snet versus 1/z at constant temperature will give an intercept of
sabs and slope rd. Figure 23.2 shows a three-dimensional plot with the linear
extrapolation of the family of curves (black points) to the limit of 1/z = 0, yielding
the white points which form a trace of the absolute sticking probability in the limit
of very high impingement rate, z. Here the symbol ML = monolayer.
The data shown in the Figure are taken at 400 K, whereas the desorption tem-
perature for CO in this case is in the range 350–550 K [5]. The method works well
when rd < z, or correspondingly, when 1/rd > 1/z.
An alternate method, particularly useful for irreversible dissociative chemi-
sorption, has been described by Xia et al. [6].
Fig. 23.2 Absolute sticking

coefficient method at
desorption temperature at
desorption temperature
References 257
References
1. Z.H. Lu, K. Griffiths, P.R. Norton, T.K. Sham, Phys. Rev. Lett. 68, 1343 (1992)
2. D.A. King, M.G. Wells, Surf. Sci. 29, 454 (1972)
3. C.C. Cheng, R.M. Wallace, P.A. Taylor, W.J. Choyke, J.T. Yates Jr, J. Appl. Phys. 67, 3693
(1990)
4. A.M. Glines, M.A. DeAngelis, A.B. Anton, J. Vac. Sci. Technol. A10, 3179 (1992)
5. X.-C. Guo, D.A. King, Surf. Sci. 302, L251 (1994)
6. L.-Q. Xia, M.E. Jones, N. Maity, J.R. Engstrom, J. Vac. Sci. Technol. A13, 2651 (1995)
Chapter 24
Mass Spectrometry
24.1 Shielded Quadrupole Mass Spectrometers—

Temperature-Programmed Desorption
One of the most widely used and versatile techniques for the study of adsorbed
species on single crystal surfaces is temperature-programmed desorption (TPD),
often called thermal desorption spectroscopy (TDS). The method extends back to
the early work of Taylor and Langmuir, 60 years ago, who studied the desorption of
Cs from polycrystalline tungsten filaments [1]. The method was extended to other
adsorbates on polycrystalline surfaces by Becker and Hartman [2], Ehrlich [3],
and by Redhead [4]. A review of the early work can be found in [5].
Temperature-programmed desorption is now used extensively for studies on single
crystal surfaces of metals, semiconductors, and insulators. A comprehensive review
of ten methods for analyzing TPD data, and an evaluation of each was recently
written by de Jong and Niemantsverdriet [6].
The collection of TPD data has progressed from the early use of ionization
gauges to the almost universal employment of quadrupole mass spectrometers
(QMS) for the detection of desorbing gases. Using the mass spectrometer in the
multiplex mode, one can simultaneously measure the evolution of various gas
products as the temperature of the crystal is raised. This can also be extended to the
study of isotopic mixing between labeled adsorbates, as was first demonstrated for
adsorbed CO desorbing from polycrystalline W [7]. Large differences in the
desorption kinetics from different single crystal faces was first demonstrated for H2
desorbing from W(100), W(111), and W(110) by the combined works of various
authors [8–12].
The early work with single crystals employed unshielded mass spectrometers
that sampled the random flux of desorbing gases originating from the front of the
crystal, from the electrical support leads to the crystal, and from the back of the
crystal disk. These effects could be eliminated to various degrees by using a dif-
ferentially pumped and shielded mass spectrometer that preferentially collected

DOI 10.1007/978-3-319-17668-0_24
260 24 Mass Spectrometry
desorbing gas only from the front of the crystal. The enhanced sensitivity to
desorbing gas as a result of the shield is related to the fact that the shield retains gas
molecules for an average time period of τ = V/S, where V = volume and
S = pumping speed, both relating to the shield region itself. Values of τ = 50
−100 ms are common, permitting each trapped molecule to pass through the ion
source many times before being pumped away. The apertured shield is selective for
desorption from the front of the crystal, but it is not measuring significant
desorption for molecules making a single pass, compared to the contribution from
trapped molecules within the shield which make multiple passes through the ion-
ization source. The value of τ must be kept below the characteristic time for the
desorption process measured, which is usually of the order of seconds for tem-
perature scan rates of a few K/s.
Figure 24.1a shows a glass shield that was just slipped over the ionization source
of a mass spectrometer. Pumping occurs from the annular space at the rear of the
shield [13]. Variations on this design (Fig. 24.1b) involve a metal shield with
pumping holes present on the side of the shield behind the ionization source. For
N2(g) at 300 K, a 1-cm2 aperture will provide a pumping speed, S = 11 L/s, where
the total pumping aperture area considered should also include the area of the
entrance aperture. Figure 24.1c shows a quadrupole mass spectrometer mounted in
an apertured shield and differentially pumped by a dedicated pump [14], which may
be either an ion pump, a titanium sublimation pump (TSP) [15], or a turbopump.
The reader is referred to p. 72 for a discussion of the relative merits of an ion pump
and a TSP in this application.
The ionization source of the quadrupole mass spectrometer contains a thermionic
emitter producing electron emission at mA currents. A fraction of these electrons
escape from the axial aperture and may be collected by the crystal, causing electron
beam damage of the adsorbate. Stray currents of the order of 10 μA are often
observed. These effects may be eliminated by either biasing the crystal at a negative
potential to retard the collection of electrons, or by using a separately negatively
biased aperture, as shown in Fig. 24.1d [16]. The effect of electron beam damage
from stray electrons in the absence of these precautions can be profound in alter-
ating the chemistry on the crystal surface.
Figure 24.1d also shows a helpful device useful for reproducibly locating the
crystal at a fixed distance from the aperture [16]. A Ta contact with a gently
rounded end is connected to the electrically isolated aperture, and the connection to
the bias potential is opened. An ohmmeter, working on the MΩ scale, is connected
between the Ta contact and ground. The crystal is slowly advanced to the Ta
contact until MQ resistance is measured as a result of very light contact. This allows
one to calibrate the horizontal micrometer drive for the crystal at a fixed point in the
apparatus. The micrometer drive is then retracted by a measured amount, and
knowing the position of the Ta contact allows one to calculate the position of the
crystal relative to the aperture.
The fractional interception of desorbing gas from the crystal will depend on the
distance between the crystal and the aperture and also on the angular distribution of
the desorbing gas, expressed as cosnθ. For a crystal of 3.5 mm radius, and an
24.1 Shielded Quadrupole Mass Spectrometers—Temperature-Programmed … 261
Fig. 24.1 Shielded quadrupole mass spectrometers—temperature-programmed desorption.

a Glass shield. b Metal shield. c Differentially pumped shield. d Stray electron Suppression +
Crystal positioning contact. e Relative signal for desorption into aperture
aperture of 0.8 mm radius, the relative collected signal for various values of n is
shown in Fig. 24.1e. It is noted that only small losses are observed in the first 1-mm
crystal-aperture distance, and also that an error in distance will make a smaller
effect in the first 1-mm range compared to errors at larger ranges [17].
A critical review of the advantages and disadvantages of using a shielded mass
spectrometer for temperature programmed desorption studies has been given by
Schlichting and Menzel [18]. A shielded mass spectrometer such as that shown in
Fig. 24.1a will have a constant pumping speed determined by the geometry of the
aperture and the opening at the rear, whereas the pumping speed of a vacuum
system used with an unshielded mass spectrometer will be variable. Also effects due
to non-Boltzmann energy distributions and noncosine angular distributions in the

desorbing flux will be reduced by the shield, which is primarily sensing molecules
scattered from its inner walls. These effects can be totally eliminated by internally
blocking the ionization source from the sample by means of a shutter [18]. The
disadvantages are the dependence of the detected signal on the distance from the
crystal to the aperture (Fig. 24.1e), and the contribution of readsorption on the
crystal during the TPD measurement because of gas readmission from the shielded
mass spectrometer through the sampling aperture. By considering these factors, the
dynamic range of the TPD coverage and kinetic measurement technique has been
extended from monolayer detection down four or five orders of magnitude [18].
For the desorption of complex molecules, or mixtures of species, it is important
to employ multiplex methods in which various mass peaks are rapidly monitored in
sequence many times during the desorption process. A recent development involves
monitoring the complete mass spectrum consecutively for many cycles during the
evolution of thermal desorption products [19, 20].
A comparison between the sensitivity of a mass spectrometer in measuring gas
pulses and also in measuring steady-state pressures, compared with a nude ioni-
zation gauge, has been published [21]. This paper shows that for eight different
gases, the ionization gauge signal and the mass spectrometer signal are linear with
the gas pulse size over six orders of magnitude for pulses ranging from 10−11 to
10−5 atm cm3. The authors were surprised at the high sensitivity of the
Bayard-Alpert gauge in these measurements [21].
24.2 Enhanced Signal-to-Noise in Quadrupole Mass

Spectrometers
The quadrupole mass spectrometer is widely used in surface science for residual gas
analysis as well as for the study of desorbing species from surfaces. Thermal
desorption, and electron- and photon-induced desorption (including laser-induced
desorption) have been widely studied. For most quadrupole mass spectrometers, a
channeltron multiplier is employed for detection of the mass-separated ions.
A method is described here for improvement of the signal-to-noise ratio of a
channeltron detector; and in addition, the behavior of a channeltron detector is also
described [22].
An electron multiplier has a low inherent noise, leading to a dark current. In the
case of a channeltron multiplier mounted behind quadrupole rods with line of sight
to the ionization source, there are additional background effects that increase as the
electron emission current increases and also as the pressure in the mass spec-
trometer is increased. While these effects are not fully understood [22], they are
thought to be due to soft X-rays and UV photons produced in the ion source, and
also to excited neutral species, both of which pass down through the quadrupole
rods without deflection. The use of an off-axis ion multiplier enables one to
24.2 Enhanced Signal-to-Noise in Quadrupole Mass Spectrometers 263
eliminate these effects to a large degree, and the modification described here
accomplishes this while not causing unacceptable reduction of the ion collection
efficiency.
Figure 24.2a shows the off-axis channeltron detector located beneath the
quadrupole rods. Between the exit region of the quadrupole rods and the chan-
neltron entrance are two ion deflection plates, designed to steer ions into the
channeltron. The channeltron is located 10 mm off axis. A multiplier shield plate,
8 mm in diameter, is mounted 2 mm in front of the channeltron entrance cone.
A deflector plate, with a 45° angle to the shield plate, is mounted on the axis of the
quadrupole, and a second deflector plate is mounted in a direction perpendicular to
the shield plate, and 15 mm from the quadrupole axis. The distance between these
deflector plates and the exit aperture of the quadrupole lens is 2 mm. The 45°
deflector plate may be biased by an isolated lead, which is separately shielded. The
entire structure is housed in a shield can to prevent both photons and charged
particles from entering the channeltron.
Figure 24.2b shows the performance of the channeltron for various applied
multiplier voltages, and also for various positive deflector plate voltages. The large
enhancement of output current for a slight increase of multiplier voltage is evident
from the three curves; in addition, a large effect of the positive deflector plate is also
seen, with maximum transmission of ions being observed at +10 V plate bias.
Figure 24.2c shows the behavior of the dark current versus electron emission
current in the ionization source of the mass spectrometer. A slope of 1.3 in this log–
log plot indicates that slight nonlinearities exist as the filament temperature is raised
in the ionization source to produce a larger electron emission current. A 100-fold
improvement in the dark current is observed for the off-axis detector.
Figure 24.2d shows the behavior of the dark current as a function of Ar pressure
in the mass spectrometer, with data taken at 1 mA emission current. The dark
current rises as the 1.1 power of the pressure in both of these measurements. The off
axis detector reduced the dark current here by a factor of 140.
The performance of the mass spectrometer with the two configurations of the
chan-neltron detector can be summarized in (24.1) and (24.2) below:
For on axis detector : Ii ¼ 3:5 P1:1 Ie1:3 ð24:1Þ
For off axis detector : Ii ¼ 0:0025 P1:1 Ie1:3 ð24:2Þ
where Ii = measured ion current; P = Ar pressure; and Ie = electron emission

current.
An ion current of the order of 10−18 A can be detected with the off-axis
channeltron.
Earlier improvements in the detection efficiency of quadrupole mass spectrom-
eters are described in [23]. The regeneration of Ag-Mg and Cu-Be dynode multi-
pliers by heating to 300 °C in oxygen gas at 1 atm has been described [24].
Fig. 24.2 Enhanced signal-to-noise—QMS. a Off axis channeltron detector, b detector perfor-
mance, c Dark current behavior, d Dark current behavior
24.3 Elimination of Spurious Electron Emission from a Quadrupole … 265
24.3 Elimination of Spurious Electron Emission

from a Quadrupole Mass Spectrometer
Electrons emitted from the ion source of a mass spectrometer can be a source of error
in measurements when the electrons strike and damage a sample in front of the mass
spectrometer. For an enclosed mass spectrometer, with an aperture, placing a neg-
ative bias on the aperture will repel electrons and prevent this effect (see p. 259). The
effect of spurious electron emission may also be eliminated by biasing the sample
negative to repel electrons. A third method involves the use of a floating grid
surrounding the mass spectrometer ion source, which is automatically brought to the
potential of the thermionic emitter (usually −70 V in a mass spectrometer). This
occurs as a result of the charging of the isolated grid by the emitted electrons [25].
Figures 24.3a–c show the steps in rebuilding a mass spectrometer ionizer to
provide the floating grid. In Fig. 24.3a, an inner grid cap is placed over the existing
inner grid cage. In Fig. 24.3b, the existing electron repeller grid cage is placed on
the ion source. Inside the repeller grid is the thermionic emitter assembly, not
shown in the drawing. Finally, in Fig. 24.3c, the floating grid, consisting of a
cylindrical grid cage capped with a planar grid, is installed. The floating grid is
electrically isolated by means of a ceramic ring that can be bolted to the mass
spectrometer housing by means of three standoff bolts as shown.
The floating grid at negative potential is capable of extracting positive ions from
the ionizing region inside the inner grid. This is the reason for the addition of the
inner grid cap on the ionizing region, which acts as a shield to prevent positive ion
extraction.
The floating grid collects electron current so that the electron emission current to
the inner grid cage is somewhat reduced. Thus, in the UTI mass spectrometer, the
filament protect circuit must be deactivated, since it will trip when it senses too little
current at the inner grid. With the floating grid design, spurious emitted electron
currents are reduced from about 10 μA to less than 1 nA [25].
A large O+ signal and a large F+ signal may be observed in the mass spec-
trometer operating with the floating grid (the latter especially if fluorinated mole-
cules are introduced to the chamber deliberately). These signals probably originate
from electron stimulated desorption effects from the ceramic ring and from the grid
material itself [25]. Fluorine, a well-known surface contaminant in UHV systems,
with a very high cross section for electron-stimulated desorption as a positive ion, is
often seen when contaminated surfaces are electron bombarded [26].
24.4 Ta Mass Spectrometer Filament
Many mass spectrometers use either W, thoriated W, or thoriated Ir thermionic

emitters as described on p. 133. The use of Ta as an emitter offers some advantages
[27] as described below.
Fig. 24.3 Elimination of spurious electron emission from a Quadrupole Mass Spectrometer.
a Placing inner grid cap, b Placing repeller grid + emitter (not shown), c Placing floating grid
Ta is much more ductile than W at room temperature and can be wound and
welded easily compared to W. Ta wire (0.006-in. diameter) is wound around a rod
(typically 0.040-in. diameter) to form a coil before being installed into the mass
spectrometer as shown in Fig. 24.4.
A Ta emitter has a built-in property that is sometimes useful. Ta has a suffi-
ciently high vapor pressure that at filament operating temperature there is a
detectable signal from 181Ta, the most abundant isotope. This signal for the 181Ta+2
24.4 Ta Mass Spectrometer Filament 267
Fig. 24.4 Ta mass spectrometer filament
ion, observed at M/q = 90.5, may be used as a mass spectrometer calibration point,
and interference from other ions in this region will be absent. The lifetime of the Ta
filament will be years despite its slow evaporation during use.
24.5 Modulated QMS Ion Source for Enhanced Sensitivity
The enhancement of the sensitivity of a QMS for TPD measurements is important if

the TPD signal is low. For example, when a differentially-pumped shield is used
where incident molecules which strike the shield walls are adsorbed, the mass
spectrometer signal can be dramatically reduced. The circuit to be described here
modulates the ion signal in the ion source region and then uses lock-in detection of
the multiplier signal for enhancement of signal/noise. The mass spectrometer is
operated at a single mass. Normally the signal from the multiplier is converted
through a current-to-voltage converter and is then amplified and digitized. The
software periodically corrects for baseline offsets in the electronics by switching the
ion source off and then measuring the signal in the absence of ions. The circuit
shown in Fig. 24.5a is able to take advantage of this capability. In one application
the ions are extracted from the source by running it at about +3 V, and is then
switched off by decreasing the source voltage to about −4 V. The frequency is
selectable from 7.5 Hz to 1 kHz. The ICM 7555 timing chip generates a
TTL-compatible square wave. The square wave from the timing chip does not have
a 50 % duty cycle, and is therefore sent to a D-type flipflop circuit, with a square
Fig. 24.5 Modulated QMS ion source for enhanced sensitivity. a Modulation for circuit,
b Enhanced TPD spectrum
wave output at 50 % of the timing chip frequency. The 100 kΩ variable resistor is
used to vary the frequency. When the disable switch is manually closed the ion
source is not modulated and the QMS operates normally.
The advantage of measuring the modulated QMS signal is illustrated in
Fig. 24.5b. Here a fraction of a ML of iodine from Pt(111) is being monitored by
thermal desorption at a temperature sweep rate of 0.1 K/s. In (a), the unmodulated
signal is shown and is very noisy. A smoothing routine is applied to this signal as
shown in (b), and two desorption peaks are barely resolved. In (c), the modulated
signal is shown with good resolution of two desorption processes. By enhancing the
speed of the temperature ramp, further improvements in signal intensity will be
found, but the resolution of overlapping peaks will begin to degrade [28]. Of
course, the method may also be used to enhance the mass spectrometer sensitivity
in other types of measurements.
24.6 Line-of-Sight Mass Spectrometry for Thermal Desorption 269
24.6 Line-of-Sight Mass Spectrometry for Thermal

Desorption
A mass spectrometer may be used for line-of-sight detection if all arriving mole-
cules from the background are removed. By placing the mass spectrometer behind
apertures which connect to baffle surfaces at 77 K, many gases may be sampled in
this manner. Any molecule with an approximate desorption enthalpy of
≥ 18 kJ mol−1 will be pumped at 77 K, and this includes molecules such as water,
ethene and almost all higher molecular weight organic and inorganic compounds.
Molecules such as H2, O2, N2, CO, Ar and CO2 will not be pumped at 77 K.
Figure 24.6 shows a schematic of the cryo-shield [29–31]. The mass spec-
trometer should have its shield removed to prevent internal reflection from warm
walls. Liquid N2 fills the tank from a reentrant tube which leads to the outside. Two
cylindrical shields, heat stationed on the cryo-reservoir are cooled by conduction.
An entrance aperture, N, admits a beam of molecules originating from the crystal
surface during a desorption experiment. All molecules originating from outside the
focal area will hit a cold surface and be retained. For a gas being admitted from the
focal area at a pressure of 10−8 mbar, the pressure in the mass spectrometer ioni-
zation volume will be about 10−12 mbar at a sample-to-ionization volume distance
of *30 cm. A mass spectrometer with a multiplier operating at 106 amplification
factor typically has a sensitivity of better than 10−14 mbar. The exit aperture, M,
does not have to be maintained at 77 K, since molecules reflected from its surface
will reach the cooled inner cylindrical shield surrounding the mass spectrometer and
be captured. To eliminate the influence of background gas produced in the mass
spectrometer, chopping of the desorbing gas may be carried out.
This method of beam measurement eliminates desorption from the crystal sup-
port assembly and may be used to measure accurate thermal desorption spectra
[31–34]. It may also be used to measure the sticking probability [29–32, 34];
Fig. 24.6 Line-of-sight mass spectrometry for thermal desorption

surface reactions can be measured either isothermally or during heating [30, 31, 34,
35]; and radical desorption may be directly measured [30].
The upper pressure limit for the apparatus is about 10−4 mbar, where mean free path
considerations will disrupt the trajectories of molecules coming from the crystal.
24.7 Mass Spectrometer Calibration by Molecular

Decomposition
Many factors make it difficult to control the partial pressures of two or more gases
in a pumped vacuum system and to know the absolute ratios of partial pressures,
much less the absolute partial pressures themselves. The factors that prevent such
calibrations have to do with fundamental issues such as the pumping speed, for
individual gases, the relative differences in sticking coefficients on the chamber
walls for different molecules, and the absence of absolute knowledge of gas fluxes
from calibrated dosers, etc. One way to overcome some of these difficulties is to
utilize the thermal decomposition of molecules, containing known stoichiometric
ratios of the elements, and to measure the relative signals from the gaseous
decomposition products under dynamic pumping conditions [36].
Figure 24.7 shows a general arrangement that maybe utilized for these mea-
surements. A molecular beam consisting of the pure compound is directed at a
heated surface that is known to decompose the compound, producing only gaseous
products. A mass spectrometer samples the desorbing gas products and yields
signals of the mass spectrometer cracking products and parent ions of the product
molecules from the decomposition.
The absolute mass spectrometer sensitivity, βx, is defined as
bx ¼ kTg =Sx gx ð24:3Þ
Here we define ηx as a proportionality constant such that Px = ηxIx and is often

called the mass spectrometer sensitivity, converting an ion signal into a partial
pressure. The quantity ηx contains factors such as the ionization probability of the
molecule, the transmission probability of the ions through the mass spectrometer,
and the electron multiplier gain, which is generally considered to have a M−1/2
dependence, where M = molecular mass.
On this basis, the ratio βx/βy will be the relative mass spectrometer sensitivity for
the two species, x and y.
If we can generate the two species in a known stoichiometric ratio by decom-
position of a molecule on a hot surface, and if we know the pumping speeds for the
two molecules from their molecular weights (scaling as M−1/2), then the relative
sensitivities βx/βy can be determined. It is of course important to measure changes in
Ix above the background and to consider that a particular mass spectrometric ion
24.7 Mass Spectrometer Calibration by Molecular Decomposition 271
Fig. 24.7 Mass spectrometer calibration by molecular decomposition
may contain contributions from the cracking of more than one type of molecule. It
is also important to remember that this method assumes that the angular distribution
of all of the product molecules from the molecular decomposition is the same, and
probably of cosine form if surface equi-librium is achieved before release of the
product molecules.
The method has been used for two molecules to determine the relative sensitivies
for four gases as diagrammed below:
H2 CO ! H2 þ CO ð24:4Þ
4DCOOH ! D2 þ 2HD þ H2 þ other products ð24:5Þ
In the case of the isotopicallylabeled formic acid, an equilibrium mixture of the

three hydrogen isotopes will be produced, since possible kinetic isotope effects
changing this ratio are likely to be small at the high surface temperature of the
decomposition.
For the vacuum system used in [36], in which a UTI-100C mass spectrometer
was employed, the following relative sensitivities were determined:
bH2 : bHD : bD2 : bCO ¼ 1:00 : 0:80 : 0:63 : 0:98 ð24:6Þ
References
1. J.B. Taylor, I. Langmuir, Phys. Rev. 44, 423 (1933)

2. J.A. Becker, C.D. Hartman, J. Phys. Chem. 57, 153 (1953)
3. An excellent review may be found in: G. Ehrlich, Adv. Catal. 14, 255 (1963)
4. P.A. Redhead, Trans. Faraday Soc. 57, 641 (1961)
5. J.T. Yates Jr, Methods of Experimental Physics, vol. 22 (Academic Press, Orlando, FL, 1985),
p. 425
6. A.M. de Jong, J.W. Niemantsverdriet, Surf. Sci. 233, 355 (1990)
7. T.E. Madey, J.T. Yates Jr, R.C. Stern, J. Chem. Phys. 42, 1372 (1965)
8. P.W. Tamm, L.D. Schmidt, J. Chem. Phys. 51, 5352 (1969)
9. T.E. Madey and J.T. Yates, Jr., in Structure et Properties des Surfaces des Solides
No. 187 (CNRS, Paris, 1970), p. 155
10. P.W. Tamm, L.D. Schmidt, J. Chem. Phys. 54, 4775 (1971)
11. T.E. Madey, Surf. Sci. 29, 571 (1972)
12. T.E. Madey, J.T. Yates Jr, Surf. Sci. 63, 203 (1977)
14. S.M. Gates, J.N. Russell Jr, J.T. Yates Jr, Surf. Sci. 146, 199 (1984)
15. G. Lu, A.L. Linsebigler, J.T. Yates Jr, J. Vac. Sci. Technol. A12, 384 (1994)
16. V.S. Smentkowski, J.T. Yates Jr, J. Vac. Sci. Technol. A7, 3325 (1989)
18. H. Schlichting, D. Menzel, Surf. Sci. 285, 209 (1993)
19. L.H. Dubois, Rev. Sci. Instrum. 60, 410 (1989)
20. A.C. Liu, C.M. Friend, Rev. Sci. Instrum. 57, 1519 (1986)
21. W.M. Graven, D.E. Gilmartin, Rev. Sci. Instrum. 55, 1827 (1984)
22. F. Nakao, Rev. Sci. Instrum. 46, 1489 (1975)
23. P. Marchand, C. Paquet, P. Marmet, Rev. Sci. Instrum. 37, 1702 (1966)
24. J.R. Young, Rev. Sci. Instrum. 37, 1414 (1966)
25. Professor Patricia Thiel, Department of Chemistry, Iowa State University, Ames, Iowa 50011
(private communication); taken from the Ph. D. Thesis of M.M. Walczak, Iowa State
University (1989); the original idea comes from A.L. Johnson, Ph.D. Thesis, University of
California, Berkeley (1986)
26. J.T. Yates Jr, D.A. King, J. Vac. Sci. Technol. 9, 1256 (1972)
27. Professor J.B. Hudson, Materials Engineering Department, Rensselaer Polytechnic Institute,
Troy, NY 12180 (private communication)
28. S.A. Furman, M. Labayen, R.G. Dean, D.A. Harrington, J. Vac. Sci. Technol., A 19, 1032 (2001)
29. R.G. Jones, C.J. Fisher, Surf. Sci. 424, 127 (1999)
30. A.S.Y. Chan, M.P. Skegg, R.G. Jones, J. Vac. Sci. Technol., A 19, 2007 (2001)
31. R.G. Jones, C.A. Clifford, Phys. Chem. Chem. Phys. 1, 5223 (1999)
32. M.G. Roper, R.G. Jones, Phys. Chem. Chem. Phys. 10, 1336 (2008)
33. A. Deyko et al., Phys. Chem. Chem. Phys. 11, 8544 (2009)
34. M.G. Roper, R.G. Jones, Langmuir 21, 11684 (2005)
35. A.S.Y. Chan, R.G. Jones, Surf. Sci. 451, 232 (2000)
36. Y.-K. Sun, W.H. Weinberg, J. Vac. Sci. Technol. A7, 3121 (1989)
Chapter 25
Temperature-Programmed
Desorption/Reaction
25.1 Thermal Desorption—Angular Resolved
The measurement of the angular distribution of molecules desorbing from single

crystal surfaces provides information on the dynamics of the desorption process.
For desorption from completely equilibrated surface species, a cosine distribution is
expected. However, often a noncosine distribution is observed, and this may be
related to processes such as atom recombination at a point on a repulsive
adsorbate-surface potential energy curve in the region of the transition state.
A simple apparatus to measure the angular distribution is shown in Fig. 25.1a [1].
A collimator is used to select particles desorbing at a certain angle, Θ, with
respect to the surface normal. The crystal is of rectangular geometry
(10 mm × 10 mm × 1 mm), and the apertures have dimensions a1 = 1.5 mm × 6 mm;
a2 = a3 = 30-mm diameter; a4 = 6 mm × 10 mm). The length of the collimator is
100 mm and the distance to the crystal is 23 mm in these measurements. The
acceptance angle outside and inside the collimator is 2.1°, and a region on the crystal
measuring 3 × 9 mm2 will be sampled in the normal direction; at 70°, a 9 × 9 mm2
region will be sampled. An angular distribution of form cosnΘ will give a signal
proportional to cos(n−l)Θ.
The collimator is internally pumped by means of a Ti getter film which is
deposited by evaporation from a Ti wire source, and this will act as a very fast
internal pump for the collimator (see p. 72).
The ionization gauge is located off-axis so that only a random flux of gas,
thermally equilibrated with the walls of the chamber containing the gauge and ion
pump, will be measured. The ionization gauge was thermostatted by means of a
water bath, and the ion pump was operated with a stabilized DC power supply to
reduce fluctuations in pumping speed. It was found that a pulse of H2 into the
ionization gauge chamber resulted in a short period decrease in the background
signal, so for measurements of H2, a continuous minute constant flux of H2 was
added. No difficulties of this kind were observed with CO.

DOI 10.1007/978-3-319-17668-0_25
274 25 Temperature-Programmed Desorption/Reaction
Fig. 25.1 Thermal desorption—angular resolved . a Desorption apparatus. b Desorption spectra

—H2/Ni(111). c Measured angular distribution for H2 and CO
25.1 Thermal Desorption—Angular Resolved 275
Figure 25.1b shows typical desorption spectra for H2 from Ni(111) that has been
saturated with H2. The angular distributions of both H2 and CO from Ni(111) are
shown in Fig. 25.1c. It can be seen that CO desorbs in the two azimuths measured
with a cosΘ distribution, whereas H2 has a sharper distribution, described as cosnΘ,
where n = 4.6 or 4.8.
25.2 Magic-Angle Thermal Desorption Mass Spectroscopy
The thermal desorption of species from surfaces is widely used throughout the field
of surface science for studies of adsorbates on metal foils and single crystal surfaces
(see p. 273). Often an unshielded mass spectrometer is used for detection of the
desorbing particles and it will generally measure the density of gas produced in a
random flux within the vacuum chamber. More modern methods involve the use of
apertured mass spectrometers that are differentially pumped within the surrounding
shield (see p. 72). The aperture is often close-coupled to the crystal surface along
the surface normal in order to collect the majority of the desorbing gas evolved
from the region of the crystal viewed by the aperture. Under these conditions of
efficient collection, considerations of the angular distribution of the desorbing gas
are of second-order importance.
However, when the apertured mass spectrometer is located far from the crystal
surface, considerations of the angular distribution of the desorbing gases become
very important, especially if measurements are made of the relative signal from
different gases desorbing simultaneously (and sampled by multiplex methods).
Thus, gases are known to desorb with angular distributions approximated by cosnΦ
where Φ is the angle from the normal. Here n can have a value of unity, or values
running to about 20 in extreme cases. If the value of n differs significantly for two
species desorbing simultaneously, then errors in estimating the relative signal
intensities will occur, and these errors can approach 400 % for collection along the
normal direction when the mass spectrometer is distant from the crystal. Pauls and
Campbell [2] have considered this problem and have found that there is a magic
angle for the mass spectrometer location that will minimize the error in comparing
desorption signals with n varying over the range 1–9. Figure 25.2a shows the
geometry of the mass spectrometer collection optics, where r is the crystal radius,
and d is the distance to the aperture. The polar angle Φ is shown. Figure 25.2b
shows the magic angle, which minimizes the error in the relative magnitude of the
desorption signals having varying angular distribution, where an integration over
the emitting surface of the crystal has been carried out. The function shown varies
as the value of the ratio d/r varies. For typical values of d/r from 2 to 5, the magic
angle varies from 42° to 34°, and at large values of d/r, the optimal angle of
Φm = 34° is found. The variations in the signal for the range of n considered is only
about 26 % at Φm = 42°.
It should be mentioned that the mass spectrometer detection behavior may vary
between two limits. For very fast pumping within the shield, only a single pass of
Fig. 25.2 Magic-angle

thermal desorption
spectroscopy. a Mass
spectrometer aperture—
collection angle.
b Magic-angle, Φm, for
aperture collection
the desorbing species occurs through the ionizing optics, and considerations of the
molecular velocities of the gas molecules would be necessary in measuring relative
fluxes [3]. For very slow pumping in the shield, which is the normal method of
operating the shielded mass spectrometer at high sensitivity (see p. 259), the
desorbing particles will probably acquire the temperature of the ionizing source and
then these considerations of molecular velocity will be unneeded. The detector
signal is then truly proportional to the flux entering the aperture [2].
These considerations also apply to the measurement of any particle desorption
technique, for example sputtering of neutral species from surfaces [4].
25.3 High-Temperature Thermal Desorption Spectroscopy 277
25.3 High-Temperature Thermal Desorption

Spectroscopy
The extension of thermal desorption spectroscopy into temperature regions well

over 2000 K involves the use of electron bombardment heating rather than ohmic
heating. Electron bombardment heating introduces experimental complications into
the method, which have been overcome as shown below [5].
The experimental difficulties introduced by electron bombardment heating are
listed below:
1. The accelerating potential for electrons must be in the kV region. If the crystal is
grounded, and the emitter is biased negatively, then spurious desorption may
occur from regions other than the crystal that are also at ground potential.
Therefore, biasing the crystal positive is the preferred mode. This requires that
the thermocouple be measured using a high-voltage isolation amplifier, and the
amplifier must be able to handle short high-voltage common mode signals if
shorting occurs between the sample/thermocouple assembly and grounded
regions of the support assembly. A suitable high-voltage precision isolation
amplifier is commercially available [6].
2. Because of radiative heating of the emitter filament by the incandescent crystal,
causing enhanced thermionic emission, the heating current to the emitter must
be controlled by a computer working in feedback style. The LabVIEW computer
(see p. 351) and program are ideal for this, but another program was used in the
work described here.
3. The mechanical mounting of the spiral emitter filament is critical. It must be
rather close to the back of the crystal, and it should be mounted in a strain-free
condition prior to first heating. Even then, distortion of the filament may occur,
and because of its brittle character after heating, readjustment of its position by
bending the W is impossible. However, if the filament is mounted on ductile Ta
posts, they may be slightly adjusted by bending after the filament is embrittled,
moving the filament to the correct position. Spiral W emitters, wound to
specified dimensions, are commercially available [7].
4. The thermocouple wires, because of their small radius of curvature, will con-
centrate the electric field lines to themselves, and will undergo electron bom-
bardment with a higher current density than the flat crystal, causing excess local
heating. This may be avoided by mounting the thermocouple close to the top of
the crystal, and placing the emitter closer to the bottom as shown in Fig. 25.3a.
The desorption of Pt from a W(211) crystal is shown in Fig. 25.3b, where a
quadrupole mass spectrometer, working in the line-of-sight mode was employed for
detection [5].
Fig. 25.3 High-temperature

thermal desorption
spectroscopy. a Electron
bombardment design. b Pt
thermal desorption from W
(211)
25.4 Automated Temperature Programmed Desorption

(TPD) Apparatus
A computer-controlled system allowing acquisition of large data sets automatically

is shown [8]. The system has a single crystal, mounted off-center, a molecular beam
doser at position 1 and a shielded quadrupole mass spectrometer at position 2. The
crystal is moved, after automated cleaning by heating, to the doser, where for a
specified time it is dosed with gas. It is then moved to the mass spectrometer
(QMS) position for the measurement of the TPD spectrum as the sample temper-
ature is automatically programmed upwards. In between each separate adsorption
experiment, the mass spectrometer is recalibrated against the steady small back-
ground pressure of the gas being continuously dosed to compensate for drifts in the
multiplier gain.
Figure 25.4a shows a schematic of the apparatus. Two stepper motors (type
Nema 23, 1 A current rating) are used to rotate the sample and to translate the
sample as shown by the arrows on the drawing. The motors operate on an open loop
25.4 Automated Temperature Programmed Desorption (TPD) Apparatus 279
Fig. 25.4 Automated temperature programmed desorption. a Two position-automated TPD

apparatus. b Schematic of control system. c Automated TPD program. d Typical TPD. e Integrated
desorption yield
control system without position feedback to the computer. Figure 25.4b shows the
main function blocks for control. Through the DAQ board, the computer controls
the power supply that feeds the heating current to the crystal. It also reads the
amplified output of the thermocouple attached to the crystal. The stepper motors are
driven by two motor control boards (Motion Group, model MS2.0). The motors
provide an angular resolution of 1.8º translating into an angular resolution of 0.045º
inside the vacuum system. The resolution of the micrometer giving linear motion to
the crystal is in the µm range.
The timing of separate steps is shown in Fig. 25.4c. Dosing of the sample to
different doses of gas is accomplished in the constant-gas flow by adjusting the time
of exposure. QMS calibration is then performed with the crystal in a null position.
TPD is next carried out with the crystal at position 2 in front of the shielded QMS.
After the upward temperature ramp, the crystal is cooled before continuing the
cycle. Figure 25.4d shows a stack of typical TPD measurements made, where the
exposure is automatically increased by a factor of 21/2 in each step. Figure 25.4e
shows that the coverage increases linearly with exposure, to great accuracy. This
means that the adsorption process is governed by the production of a mobile
precursor species.
The automatic method is highly accurate since it does not involve human
intervention.
25.5 Scanning Kinetic Spectroscopy—A Survey Method

for Investigation of Surface Reaction Processes
Multiple reaction pathways are available to a polyatomic molecule interacting with

a solid surface. It is often useful, in a survey experiment, to delineate the temper-
ature regions where various reaction steps are either active or inactive. The scan-
ning kinetic spectroscopy (SKS) method [9] permits this to be done efficiently in
many cases [10, 11].
Figure 25.5a shows an SKS apparatus for the study of a molecule’s decompo-
sition routes on a single crystal surface. The cleaned single crystal is first covered
with a monolayer of the adsorbate to be studied at a temperature below the dis-
sociation temperature of any of the bonds in the molecule. At this point, the crystal
is rotated into a beam of the molecules originating from a collimated beam doser, as
described on p. 421. Initially, the incident beam is reflected off of the saturated
monolayer on the crystal, providing a reference signal in the mass spectrometer
which is located behind an apertured shield (see p. 259) and is differentially
pumped (see p. 72). The crystal is then programmed upward in temperature, and the
mass spectrometer records, in the multiplex mode, all of the relevant mass peaks
related to consumption of the parent reactant molecule and to liberation of product
molecules, as shown schematically in the bottom section of Fig. 25.5a.
Figure 25.5b shows a typical set of SKS measurements, made in multiplex
mode, which characterizes the temperatures where the various elementary steps
occur. The steps themselves are shown in Fig. 25.5c. Here, the reaction of the
methanol molecule on Ni(111) is investigated.
Let us follow the processes observed in Fig. 25.5b and outlined in Fig. 25.5c. In
this example, the isotopically labeled molecule, CD3OH, is being employed as the
reactant at a known flux onto the crystal. The covered crystal is rotated into the
CD3OH beam at the double-headed arrow on the left. In the top panel of Fig. 25.5b,
at temperatures below 187 K, CD3OH merely reflects off of the monolayer of
CD3OH previously adsorbed, and a flat line for mass 33 is observed in this region.
Upon beginning the temperature program, α-CD3OH is first observed to desorb,
originating from undissociated methanol molecules. At the dashed line, 290 K, a
25.5 Scanning Kinetic Spectroscopy—A Survey Method for Investigation … 281
Fig. 25.5 Scanning kinetic spectroscopy (SKS). a SKS apparatus. b Typical SKS data.
c Elementary reaction steps as inferred from B
β-CD3OH feature is sharply evolved. The β-process is due to an elementary step

involving the recombination of adsorbed H (from CD3O–H bond scission) with
adsorbed methoxy, CD3O. It is informative to note that when the isotopic molecule,
CD3OD, is monitored at mass 36, the β-process is also observed, and the D atom
involved here can only come from the breaking of a C–D bond in adsorbed CD3O.
This probably occurs by the production of adsorbed formaldehyde, CH2O (a),
shown (Fig. 25.5c) in process A, where the isotopic composition is not written for
any of the molecules in Fig. 25.5c. When mass 4 is monitored in Fig. 25.5b, D2 is
observed to desorb at 330 K, opening up sites on the surface. When mass 28 is
monitored in Fig. 25.5b, CO is observed to desorb at 420 K, also opening up sites.
Three processes labeled γ are also indicated on the top SKS spectrum. The
process γ1 for adsorbed CH3O formation by O–H bond scission is not defined, since
a volatile product is not liberated from this process, but O–H bond scission must
begin below 290 K since the β-recombination process begins at 290 K. The process
γ2 and γ3 involve the consumption of the reactant molecule since the reflected flux
of reactant falls below the reference line in this temperature region. These con-
sumption processes are closely correlated with the D2 and CO desorption processes,
shown near the bottom of Fig. 25.5b, which open Ni sites for adsorption and
reaction processes.
The SKS method is a fast survey method, giving one a partial view of the
characteristic temperatures for various adsorption, desorption, and surface disso-
ciation processes. It has been effectively used with intentional poisoning of the
surface by added species, and with intentional addition of known reactants to the
surface [9]. It has been successfully used to observe deuterium kinetic isotope
effects that give insight into the rate-controlling steps in surface reactions involving
the hydrogen moiety [9, 11]. The original concept of temperature scanning while
making reaction kinetic measurements on single crystal surfaces was first presented
in 1981 by Goschnick and Grunze [12]. The SKS method also resembles the
familiar temperature-programmed reaction spectroscopy (TPRS) [13, 14] (see
p. 282), but SKS has added advantages of working like a flow reactor where fresh
reactant is continuously supplied during the measurements.
The ability to rapidly scan many mass peaks and to record the time behavior of
selected mass peaks has been a familiar subject for a number of years. Recently, a
fast multimass measurement method has been introduced that permits up to 15 mass
spectra of 50 amu width to be recorded per second, using a quadrupole mass
spectrometer [15]. Another variation on this method scans up to 100 complete mass
spectra per thermal desorption experiment. Enhanced signal-to-noise is achieved by
integrating each of the mass peaks rather than by measurement of the peak height
only [16].
There are many other mass-spectrometer based methods used for the study of
surface reactions in which transient surface species may be detected. Among these
is laser induced desorption, in which a Fourier transform mass spectrometer is often
used [17]. Here, the laser may be used to desorb surface species without extensive
thermally induced fragmentation.
25.6 Temperature-Programmed Reaction Spectroscopy

(TPRS)
This method extends the basic concepts of temperature-programmed desorption

(TPD) (see p. 259) to examine complex surface reactions. Multiple mass channels
are monitored simultaneously during programmed heating. The method has been
widely used to unravel the kinetic steps in both decomposition and synthesis
reactions on single crystal surfaces. The method also resembles scanning kinetic
spectroscopy (SKS) described on p. 280.
25.6 Temperature-Programmed Reaction Spectroscopy (TPRS) 283
Fig. 25.6 Temperature-programmed reaction spectroscopy (TPRS) a TPRS data—

(CH16
2 OH)2 +
18
O/Ag(110). b Dissociative adsorption followed by H18
2 O evolution. c H2 O and
18
CH2 O evolution. d HCO O decomposition

16 18
An example of the use of TPRS is given here [18]. The interaction of ethylene
glycol, (CH16 18
2 OH)2, with fractional monolayers of chemisorbed O on Ag(110) has
been studied. First, up to 200 mass peaks are scanned to survey the products. Then,
using methods which scan 10 mass peaks in a multiplex mode in 20 ms, the
temperature dependence of evolution of a variety of products was measured as
shown in Fig. 25.6a. The two cross-hatched desorption processes occur in surface
reaction limited processes and contain no 18O, and are the end result of the oxi-
dative dehydrogenation of the ethylene glycol. The five remaining product peaks
are indicative of reactions involving the chemisorbed 18O.
The desorption of H182 O at 255 K occurs as a result of the direct reaction of O–H
bonds of the two O–H groups of the ethylene glycol with adsorbed 18O, as illus-
trated in Fig. 25.6b. The desorption together of H18 16
2 O and CH2 O with a peak
temperature of 300 K is indicative of a surface reaction controlled by C–H bond
scission in the CH2 groups of the 1,2 ethanedioxy species as shown in Fig. 25.6c.
The simultaneous evolution of H2 and C16O18O at 415 K is indicative of the
decomposition of adsorbed formate, HC18OO, on the surface. Similar reactions
were observed in separate experiments with adsorbed formate, and are illustrated in
Fig. 25.6d.
There are other possible mechanisms that might participate in the surface
reactions, and these were eliminated by using D substitution for H in the CH2
groups of the ethylene diol and looking to see whether a primary kinetic isotope
effect was seen, in which the scission of the C–D bond occurs at a significantly
higher temperature than the C–H bond in the unlabeled molecule. This isotope
effect was observed, pointing to C–H bond scission as being a key elementary step
in the surface chemistry.
Two reviews on TPRS have been written [14, 19].
References
1. H.R Steinruck, A. Winkler, K.D. Rendulic, Surf. Sci. 152/153, 323 (1985)
2. S.W. Pauls, C.T. Campbell, Surf. Sci. 226, 250 (1990)
3. G. Comsa, R. David, Surf. Sci. Rep. 5, 145 (1985)
4. A. Wucher, F. Novak, W. Reuter, J. Vac. Sci. Technol. A6, 2265 (1988)
5. K. Pelhos, T.E. Madey, Department of Physics and Astronomy, Laboratory for Surface
Modification, Rutgers, The State University, Piscataway, NJ 08855 (private communication)
6. Model AD210AN. Analog Devices, One Technology Way, Norwood, MA 02062-9106
7. C.M. Furnaces, Inc., 103 Dewey Street, Bloomfield, NJ 07003
8. P. Kondratyuk, J.T. Yates Jr, J. Vac. Sci. Technol. A23, 215 (2005)
10. I. Chorkendorff, J.N. Russell Jr, J.T. Yates Jr, J. Chem. Phys. 86, 4692 (1987)
11. P.L. Hagans, I. Chorkendorff, J.T. Yates Jr, J. Phys. Chem. 92, 471 (1988)
12. J. Goschnick, M. Grunze, J. Vac. Sci. Technol. 18, 561 (1981)
13. R.J. Madix, J. Benziger, Ann. Rev. Phys. Chem. 29, 285 (1978)
14. R.J. Madix, in Chemistry and Physics of Solid Surfaces, vol. II, ed by R. Vanselow (CRC
Press, Boca Raton, FL, 1979)
References 285
15. G. Zagatta, H. Muller, N. Bowering, U. Heinzmann, Rev. Sci. Instrum. 65, 359 (1994)
16. L.H. Dubois, Rev. Sci. Instrum. 60, 410 (1989)
17. D.P. Land, C.L. Pettiette-Hall, D. Sander, R.T. Mclver Jr, J.C. Hemminger, Rev. Sci. Instrum.
61, 1674 (1990)
18. A.J. Capote, R.J. Madix, Surf. Sci. 214, 276 (1989)
19. R.J. Madix, J. Benziger, Annu. Rev. Phys. Chem. 29, 285 (1978)
Chapter 26
Gas Chromatography—Enhanced
Sensitivity
26.1 Gas Chromatography—Sensitivity Enhancement
The sampling of gases for gas chromatography analysis suffers from the fact that
gas expansion in the sampling lines invariably causes loss of analytical sensitivity.
This behavior may be reversed by using a high-pressure pulse of the carrier gas to
compress the low-pressure gas sample into the sample loop, where it is then swept
into the gas chromatograph. A factor of 20 or more sensitivity enhancement may be
achieved. The application shown here involves sampling a high-pressure reactor
being used to study the kinetics of a catalytic reaction over a single crystal.1
As shown in Fig. 26.1a, the reactor volume to be sampled is connected by valve
V1 to a length of capillary tubing containing an aliquot volume. The tube is then
connected via a tee to a gas chromatography sampling valve. Gas from the reactor is
expanded into the aliquot volume and the thin tubing (V5 open) and its pressure is
read with a low-volume pressure gauge. The valve V2 to the pump is closed and
valve V3 is opened to the high pressure of the carrier gas regulator (about 800 Torr).
Valve V4 is then opened, and the carrier gas acts as a piston, compressing the
sample gas into the sample loop located on the sampling valve. The contents of the
sample loop are then injected into the gas chromatograph, as shown in Fig. 26.1b.
In the case shown here, the sampling volume is the aliquot volume plus the volume
of the thin tubing, valve, and sampling loop. By keeping valve V5 closed during
sampling, the sampling volume is just the aliquot volume. It was shown that two
consecutive analyses using this method differed by less than 4 %.
1
Personally contributed by Professor C.T. Campbell, Department of Chemistry, University of
Washington, Seattle, WA 98195. See also [1].

DOI 10.1007/978-3-319-17668-0_26
288 26 Gas Chromatography—Enhanced Sensitivity
Fig. 26.1 Gas chromatography—sensitivity enhancement. a “Fill loop”. b “Inject”
Reference
1. C.T. Campbell, K.A. Daube, J. Catalysis 104, 109 (1987)

Chapter 27
Work Function, Tunneling Spectroscopy
and Ellipsometry
27.1 Work Function Changes Using Retarding Diode

Method
The use of a retarding diode method for measuring work function changes upon
adsorption is one of the oldest and simplest methods for measurement in surface
science. The literature related to this method is mostly from the 1960s or 1970s
[1–6]. Other methods for measuring work function changes are described elsewhere
in Sects. 27.2–27.6.
In Fig. 27.1a is shown a thermionic emitter housed in an ultrahigh vacuum
system and facing a sample whose work function change, Δϕ, is to be measured as a
result of adsorption or desorption. The emission current collected by the sample as a
function of the applied voltage, U, is given by
IðUÞ ¼ AT 2 exp½eðU þ /Þ=kT ð27:1Þ
where A is a constant (proportional to the cathode area), T is the absolute emitter

temperature, k is the Boltzmann constant, and ϕ is the sample work function, and
e is the electronic charge. Equation (27.1) applies only for the initial section of the
retarding potential curve where space charge effects are absent [4].
In Fig. 27.1b, a plot of the collected current versus the retarding potential, U, is
shown for a sample which has undergone a work function change, Δϕ. The dif-
ference in the retarding potential measured at constant current on the linear sections
of the curves is equivalent to the work function change, assuming that the emitter
work function has not changed and that the emitter temperature remains constant.
This difference measurement is made where the two “linear” portions of the curve
are exactly parallel.
In practice, the resistance, Re, of the emitter is kept constant using a Wheatstone
bridge, and if the dimensions of the emitter filament have not changed due to
oxidation of the material of which it is made or to sublimation of the emitter, the

DOI 10.1007/978-3-319-17668-0_27
290 27 Work Function, Tunneling Spectroscopy and Ellipsometry
Fig. 27.1 Retarding diode for measuring work-function changes. a Schematic of circuit,
b Collected current versus retarding potential
temperature will also be constant. The temperature of the cathode emitter will be
sensitive to the heat loss through the leads to the emitter, and thermostating these at
constant temperature with a water cooling jacket will remove this variable.
There are several factors that can affect the measurements using this simple
device. One possible problem is the emission of impurity gases from the hot emitter
27.1 Work Function Changes Using Retarding Diode Method 291
filament, which will influence the work function of the sample. Second is the effect
of the hot filament on the residual gases in the vacuum system (for example, H2
dissociation) making them reactive with the sample. This effect can be eliminated
by turning off the emitter during gas admission. The method will not work well at
pressures above about 10−4 Torr, since otherwise the cathode might be cooled by
heat conduction through the gas. Finally, the effect of electron bombardment and
radiation on the surface layer on the sample must be understood when using this
method.
The retarding diode method has been employed in a conventional LEED
apparatus with continuous derivative measurement of the work function change
during adsorption. A sensitivity of about 5 meV was obtained, using the lock-in
amplifier method for making continuous measurements of the changing work
function [7]. Reference [7] also summarizes background papers in this area.
27.2 Work Function Measurements Using an Electron

Gun
One of the simplest methods to measure the work function change during
adsorption on a metal single crystal is to employ a retarding potential method, using
an ordinary electron gun and an interposed grid or aperture. Figure 27.2a shows a
simple device first used for the study of hydrogen, nitrogen and xenon adsorption
on tungsten surfaces [8–10]. An electron gun of ordinary design supplies current to
the crystal through an aperture made of coiled tungsten in this case. The coiled
tungsten is also used as an electron emitter for thermally cleaning the tungsten
crystal at 2500 K. The crystal is welded to a thin wire support which may be cooled
with liquid nitrogen. By using 25 eV electrons, and retarding between the aperture
and the crystal surface, I–V plots are obtained, and from the shift in electrical
potential of the knee region, the work function change may be deduced as gases are
adsorbed on the crystal. The reproducibility of duplicate measurements is
about ±20 meV.
Figure 27.2b shows the linear work function change versus coverage for H2
adsorption on W(100). The linearity versus measured coverage (using thermal
desorption spectroscopy) suggests that depolarization effects between neighbor H
atoms are small in this case [8]. Figure 27.2c shows measurements of the adsorption
of N2 on W(110) at 123 K. Here, the work function change is nonlinear with
coverage, and this indicates that depolarization effects exist between the adsorbed
molecular N2 molecules [9]. The plot shown in Fig. 27.2c is a fit to the immobile
Topping model, and the N2 polarizability may be obtained from this fit [9, 11].
The use of an electron gun in a display LEED apparatus for retarding potential
measurements of work function changes has been described [12].
Fig. 27.2 Work-function

changes using electron gun.
a Retarding potential
measurement system using an
electron gun, b Work function
change—H2/W(100),
c Topping model plot—N2/W
(100) work function
measurement
27.3 Work Function and Electron Reflection Coefficient Measurements … 293
27.3 Work Function and Electron Reflection Coefficient

Measurements with the Shelton Diode
The use of the Shelton diode [13, 14] to measure the work-function difference
between two surfaces, an emitter and a collector, has been employed to determine
the difference in work function between two crystal planes of tungsten. In addition,
the work-function change during adsorption of a gas on the collector, and the
energy-dependent reflection coefficient for electrons incident normally on the col-
lector surface, can be measured with this device [15].
Figure 27.3a shows the schematic arrangement used [15]. Two single crystals of
different crystallographic orientation are mounted by wires and a tungsten spring in
support rings. These crystals face a movable septum containing a small aperture.
Either crystal may be heated by electron bombardment to a temperature where
thermionic emission occurs, and a feedback stabilized controller is used to maintain
emitter and crystal temperature constant, using the current collected from the crystal
emitter by the septum for regulation of the temperature of the spiral tungsten emitter
located behind the emitter crystal. The entire assembly is placed in an axial mag-
netic field of 3600 G, causing the emitted electrons to travel in tight spiral paths
from the emitter to the collector crystal. The movable septum permits sampling of
different regions of the emitter and collector. The collector crystal current is plotted
against the retarding potential applied between emitter and collector, and a typical
retarding potential plot, as shown in Fig. 27.3b, may be measured. The collected
current follows the equation
I ¼ I0 exp½eðVR þ D/Þ=kT ð27:2Þ
where I0 is the emitted current from the emitter, VR is the retarding potential
applied, and Δϕ is the difference in work function between the collector and the
emitter, and e is the electronic charge. Thus as the work function of the collector
changes, the retarding potential plot (Fig. 27.3b) will shift. The shift of the “knee”
in the plot is a measure of the work-function change of the collector, and may be
used as a measure of the adsorption of a gas on the collector, the emitter remaining
clean at its high operating temperature. The device shown in Fig. 27.3a measures
collected currents of the order of 0.01 μA and to avoid problems with leakage
currents, the terminals to each crystal are shielded from evaporating metal films by
the hat-shaped shields shown on each feedthrough. The device is symmetric, so that
it may be operated with either crystal acting as the emitter, but such a symmetrical
arrangement is of course unnecessary if one is interested in only one crystal. In
contrast to similar devices [13, 14] that used ohmic heating of the emitter, this
device does not have a potential drop across the emitter region sampled by
the aperture and should therefore perform better by having less uncertainty of the
electrical potential distribution in the sampled emitter region.
By varying the electron energy and measuring the collected current, curves like
those shown in Fig. 27.3c may be generated. These curves are related to the
Fig. 27.3 Work function measurements—shelton thermionic diode. a Shelton diode, b Retarding
potential plot, c Reflection coefficient plot
effective electron reflection coefficient in the plateau region and have to do with
band structure and other possible effects at the collector.
The accuracy of work-function measurements made with this device is of the
order of ±20 meV over many months of operation with tungsten crystals.
27.4 Improved Piezoelectric Drive Kelvin Probe 295
27.4 Improved Piezoelectric Drive Kelvin Probe
The Kelvin-Zisman vibrating capacitor method for measuring changes in the work
function of a surface is widely used in surface science studies. The principles of this
method are well described in the literature [16, 17] and will not be discussed in
detail here. Basically, a reference electrode is caused to oscillate at audio frequency
and the electrical potential of the reference electrode is adjusted to yield a null
signal in the capacitance circuit. As the work function of the sample changes due to
adsorption or to changes in temperature, the potential applied to the reference
electrode is adjusted to maintain a null signal in the AC circuit. The change in the
applied potential is equal to the change in the work function or surface potential of
the sample. It is of course important to know that the reference electrode is not
changing its work function during the measurements, and inert electrodes such as
Au or SnO2 films are often used. The vibrational motion of the reference electrode
is commonly produced by mechanical excitation of the resonant frequency of the
device, by an electromagnetic driving mechanism, or by the use of a piezocrystal
oscillator.
The design shown in Fig. 27.4a involves the use of a piezoceramic foil actuator
[18]. This avoids the interaction of stray magnetic fields from an electromagnet
driver with the sample, which could cause problems if the sample is magnetic. In
addition, the measured signal is taken from the reference electrode, not the sample,
avoiding complications owing to the connection of the sample to both the heating
circuit and the thermocouple circuit. The device described in Fig. 27.4a is a
modification of the piezoelectric Kelvin probe built by Besocke and Berger
[19, 20]. The device was used to study magnetic materials and yielded an accuracy
of ±1 meV [21]. This compares to a reported sensitivity of better than 0.1 meV in
the original device built by Besocke and Berger [19]. The reader is referred to the
original paper [18] for dimensional details of the Kelvin probe shown in Fig. 27.4a.
The reference electrode is an L-shaped piece of gold, connected by a tungsten wire
to a quartz rod and thence to a shield plate. This assembly is actuated by a piez-
oceramic foil [22], containing vapor deposited Ag contacts on both surfaces, and
glued using UHV leak sealer to a tungsten reed, 20 mm × 4 mm × 0.1 mm thick.
The signal from the Au reference electrode is measured through a very thin wire
connected to a support ceramic before exiting the apparatus. The stable resonance
frequency of the reed is about 130 Hz, the amplitude of vibration is about 0.1 mm,
and the AC driving voltage is about 1.0 Vrms. The spacing between the reference
electrode and the sample is about 0.5 mm.
Before operating, the piezoceramic foil is polarized by a DC voltage pulse above
300 V. The polarization remains as long as the piezoceramic is not heated
significantly.
Figure 27.4b shows a block diagram of the self-compensating feedback circuit
which allows a continuous measurement of the work function change. It is esti-
mated that the drift is about 1 meV/h and the noise is about ±0.5 meV. The response
time constant of the system is less than 10 s. The variable DC power supply
Fig. 27.4 Improved kelvin probe—piezodrive. a Piezodriven kelvin probe, b Circuit for Δφ
measurements, c Work function change on cooling Ni (poly.)
supplies a variable compensating voltage, Vb. The authors note that electron
emission from thermionic sources can disturb the electrical isolation of the sample
and the reference electrode [18].
27.4 Improved Piezoelectric Drive Kelvin Probe 297
Figure 27.4c shows a record of the increase in the work function of a poly-
crystalline Ni sample as it cools through the Curie temperature, TC.
Excellent reviews and analysis of the effects of stray capacitance [23, 24] and
noise [25] on the performance of the Kelvin probe have been published. Kelvin
probes, driven by an electromagnetic voice coil and possessing superior noise
levels, have been described [26, 27].
The Kelvin probe has been used in dynamic measurements of the desorption of
one desorbing surface species from a single crystal surface [28]. The method was
shown to give superior resolution of overlapping desorption processes compared to
temperature programmed desorption into a large chamber with gas monitoring by a
quadrupole mass spectrometer [28]. Also, the dynamic work function change
method can be used over a wider range of dT/dt than the mass spectrometer method
[28].
For more complicated surface phenomena in which surface decomposition of a
molecule may occur without product desorption, the use of the Kelvin probe for
observing work-function changes associated with decomposition has proven helpful
[29]. The Kelvin probe may also be used to measure work-function changes on
semiconductor surfaces when adsorption occurs [30].
27.5 Improved Electromagnetic Drive Kelvin Probe
In comparison to the piezodriven Kelvin probe shown on p. 295, an

electromagnet-driven device is described here with somewhat different features.
In the design shown in Fig. 27.5a, great attention has been paid to the elimi-
nation of stray capacitances [31]. This is accomplished by electrical adjustments of
a potential applied to the electrical shielding as described in the original paper [31].
The reference electrode is a grid of 80 % transparent stainless steel mesh, stretched
across a 3–4-mm-diameter Mo ring, and coated with SnO2 (see p. 123) to produce
an inert surface of constant work function during gas adsorption. The spacing
between the reference electrode and the sample is 0.3–0.5 mm. The transparent grid
electrode is used in order to allow the sample to have almost full access to gas
during adsorption rather than having partial blocking by a solid electrode. The
frequency of the device is about 160 Hz, but it is advantageous to operate at higher
frequencies, so that an overtone frequency of 1150 Hz was used. Working at such
high frequencies has the effect of removing disturbances owing to low frequency
noise, caused by effects such as bumping the apparatus. Because the resonant
overtone frequency is sharp (<0.5 Hz) a quartz oscillator was used for stability. The
amplitude of the vibration is about 0.1 mm. The response time constant was
measured by applying a sudden potential to the sample, and was 20 ms.
The Kelvin probe is constructed using a light stainless steel tube to connect the
reference electrode to the holder. A lightweight mu metal platelet was welded onto
the tube in the position of the electromagnet, and the assembly was clamped in a
stainless steel holder, electrically isolated by ceramic insulators from the
Fig. 27.5 Improved kelvin probe—electromagnet drive. a Electromagnet-driven kelvin probe,

b Stability and noise, c Thermal desorption of CO/Ru(001) measured two ways
electromagnet mounting. The electromagnet is made of soft iron and is wrapped

with a few turns of aluminum winding wire, which has an anodized oxide insulating
surface film for isolation. All electrical measurements are made from the isolated
reference electrode assembly, using rigid partially screened conductor wires that are
free of microphonics, and that are connected to vacuum feed throughs. Figure 27.5b
shows a measure of the noise and long term stability of the device.
The quality of the measurements made with this apparatus permit one to study
the time derivative of the work-function change on the sample as adsorbate surface
coverage changes. Assuming that the work function is linearly proportional to
coverage, the thermal desorption spectrum obtained with the mass spectrometer
27.5 Improved Electromagnetic Drive Kelvin Probe 299
should agree with that obtained from the derivative of the work function. It was
found that an improvement in the resolution of the thermal desorption spectroscopy
could be achieved by using the Kelvin probe, compared to the mass spectrometer
detector of desorbing gas. However, the two kinetic curves do not agree because of
the nonlinear behavior of work function with coverage. Figure 27.5c shows data for
the thermal desorption of CO from Ru(001) measured both ways [31].
Reference [31] gives an excellent review of the development of the vibrating
capacitor method for work-function measurements.
An excellent analysis of the effect of stray capacitance on Kelvin probe mea-
surement is given in [32], and the effect of noise on Kelvin probe measurements is
given in [33]. Reference [34] gives a new design for a Kelvin probe involving a
voice coil driver. Reference [35] gives design and performance features for a Kelvin
probe with a reproducibility of 1 meV over long periods of time. In addition,
methods for measuring the work-function change during ohmic heating of a metal
crystal are described [35].
27.6 Kelvin Probe—High Temperature Flow Reactor
The vibrating Kelvin probe has been described on pp. 295–299. Both mechanical
and electromechanical activation of the motion of the reference electrode have been
employed. For electromechanical activation, a piezoceramic actuator is often used,
but because the upper temperature limitation is near 415 K for these actuators, they
are not generally useful for high-temperature studies.
Figure 27.6 shows a method for using a piezoceramic actuator [36] to drive a
Kelvin probe at some distance from a high-temperature sample [37]. The device,
with a balanced quartz tube lever of 180-mm length, has the actuator on one end
and the Au Kelvin probe electrode on the other. The quartz tube (2-mm diameter)
lever arm is balanced on sapphire rods, as shown in the drawing. The actuator is
coupled to the quartz tube lever with a light weight stainless steel spring. This
refinement avoids the addition of a large mass to the actuator, and the resonant
frequency of the device is close to that of the free actuator (in this case, 325 Hz).
The apparatus is used to study chemical reactions on a single crystal sample
placed under the Kelvin probe. This sample may be heated in flowing gas while the
work function is being measured. The large distance between the hot sample and
the actuator avoids heating of the actuator. In addition, the large separation distance
prevents electromagnetic pickup in the work function measurement from the pi-
ezodrive unit. A sampling tube permits the gas composition to be measured in the
vicinity of the sample, and this provides a measurement of the local rather than the
incident gas composition.
The resolution of the probe was estimated to be better than 5 meV, using known
applied potentials on the sample for calibration. It is believed that improvements in
the electrical measurement methods will lead to substantially lower noise and
higher resolution [37].
Fig. 27.6 Kelvin probe—high-temperature flow reactorhigh-temperature flow reactor
27.7 Scanning Tunneling Spectroscopy (STS)—

Compensation for Density of States of the Tip
Many examples if the use of I/V measurements for probing the local density of
electronic states (LDOS) of adsorbed atoms and molecules are now present in the
literature. Similarly, for alloy metal surfaces, STM-STS measurements are often
used for studying the individual metal atoms, either in the surface or the subsurface
layers of the metal. A common difficulty with these measurements is that the I–V
and dI/dV measurements versus V contain a contribution from the LDOS of the tip
atoms which is mixed in. In the method to be described here, a subtraction of STS
spectra is made in order to cancel out the LDOS contribution from the tip.
Figure 27.7a shows a schematic alloy of composition AxBy. The STM, through a
number of averaged scans across an identical region, provides first a dI/dV scan of
atom A, which contains, also the tip contribution. A similar number of scans across
atom B provides a second dI/dV scan, also containing the tip contribution. By
subtracting A−B, the tip contribution is cancelled out, and the difference spectrum
represents the change in dI/dV when one moves from atom A to atom B. These
differences are able to be handled more easily by theory than the convoluted spectra
from which the difference spectrum is derived [38].
27.7 Scanning Tunneling Spectroscopy (STS)—Compensation … 301
Fig. 27.7 Difference I/V spectra—STM. a Schematic differences in A and B spectra—STM,

b Pd-Cu alloy differences—STM
One example shown in Fig. 27.7b concerns the Cu-Pd alloy, where STM can
distinguish surface Cu atoms, surface Pd atoms, and subsurface Pd atoms. The data,
when subtracted give a difference spectrum for the two types of Pd. Here the Cu
spectrum has been used as a subtraction spectrum, removing the tip contribution.
The difference spectrum therefore represents the difference between Pd and Cu. It is
seen that at about −0.4 V, both types of Pd atoms exhibit a slightly lower LDOS
than Cu atoms. Consequently, charge transfer takes place from Pd to Cu atoms, just
the opposite from a prediction based on the higher work function of Pd (5.12 eV)
versus Cu (4.65 eV) [39].
27.8 Ellipsometry as a Tool for Studying Adsorption
Ellipsometry has been used for years to study adsorption on surfaces, especially in
the study of oxide films grown on metals. There are few comparisons of adsorption
measurements with other surface measurements, especially under ultrahigh vacuum
conditions for model systems with well-understood behavior.
The ellipsometer is an automatic null system with a polarizer-compensator-
sample-analyzer (PCSA) configuration as shown in Fig. 27.8a. Here, two ellips-
ometry parameters are measured: “phase shift” Δ, and “amplitude reflectance ratio”
ѱ, which are components of the optical constant of the surface. The relative change
in Δ, δΔ, measured versus the clean substrate is used as a measure of the amount of
adsorption. This is due to the fact that δΔ is much more sensitive than δѱ for a
physisorbed film thinner than a few 10s of nm [40]. The apparatus shown also
contains XLEED, which is a sensitive low energy electron diffraction method using
a low current electron beam with microchannel plate amplification of the diffraction
pattern. XLEED eliminates the problem of electron stimulated desorption (ESD).
Measurement of the coverage behavior of Xe is achieved by measuring the intensity
of the diffracted Ag(10) beam [41, 43].
Figure 27.8b shows a comparison of Ellipsometry and XLEED for the mea-
surement to the adsorption of Xe on both Ag(111) and Ag(100). At a constant
pressure of Xe of 1.1 × 10−5 Pa, the temperature of the crystal was slowly lowered
from 82 to 56 K. The broken line shows the behavior of the ellipsometry mea-
surement of δΔ, versus T, where equilibrium is maintained at all temperatures. The
initial increase is due to the adsorption of 1 ML of Xe, followed by a second
increase due to the formation of a second layer. The solid line shows the XLEED
behavior of the [10] beam. The two experimental methods compare two crystals,
but even in this case, it may be seen that the condensation temperature for the
second layer is identical to 0.1 K. The slight difference in condensation temperature
for the monolayer, measured by the two techniques, is not due to differences
between the techniques, but is due to slight differences in the polarization of Xe on
the two Ag surfaces, leading to a difference of 0.2−0.4 K in the position of the step
corresponding to monolayer completion.
References 303
Fig. 27.8 Ellipsometry for monolayer measurement. a Ellipsometry-XLEED apparatus,

b Adsorption of Xe on Ag(100) and Ag(111)
References
1. Z. Knor, V. Ponec, Coll. Czech. Chem. Commun. 31, 1172 (1966)

2. T. Jirsak, V. Nikolajenko, Z. Knor, Czech. Chem. Commun. 59, 1709 (1994)
3. See literature quoted in V. Ponec, Z. Knor, and S. Cerny, Adsorption on Solids (Butterworth,
London, 1974)
4. A. Eberhagen, Fortschritte d. Physik 8, 245 (1960)
5. E.C. Tompkins, Chemisorption of Gases on Metals (Academic Press, London, 1978)
6. G. Ertl, J. Küppers, Low Energy Electrons and Surface Chemistry (Verlag Chemie, Weinheim,
1974), Chap. 8
7. J.L. Taylor, W.H. Weinberg, J. Vac. Sci. Technol. 15, 1811 (1978)
8. T.E. Madey, J.T. Yates, Jr., Colloques Internationaux du Centre National de la Recherche
Scientifique, No. 187, 155 (1970)
9. J.T. Yates Jr, R. Klein, T.E. Madey, Surf. Sci. 58, 469 (1976)
10. M.J. Dresser, T.E. Madey, J.T. Yates Jr, Surf. Sci. 42, 533 (1974)
11. J.R. MacDonald, C.A. Barlow Jr, J. Chem. Phys. 39, 412 (1963)
12. J.H. Fritz, C.A. Haque, Rev. Sci. Instrum. 44, 394 (1973)
13. H. Shelton, Phys. Rev. 107, 1553 (1957)
14. P. Kisliuk, Phys. Rev. 122, 405 (1961)
15. T.E. Madey and J.T. Yates, Jr., Supplemento al Nuovo Cimento, No. 2, Series I, 5, 483, (1967)
16. G. Ertl and J. Küppers, Low Energy Electrons and Surface Chemistry (Verlag Chemie, GmbH,
D-694 Weinheim, 1974), p. 120
17. J. Holzl, F.K. Schulte, in Solid Surface Physics, ed. by G. Hohler, vol 85 (Springer, Berlin,
1979), p. 1
18. S. Saito, T. Soumura, T. Maeda, J. Vac. Sci. Technol. A2, 1389 (1984)
19. K. Besocke, S. Berger, Rev. Sci. Instrum. 47, 840 (1976)
20. G. Wedler, R. Ruhmann, Appl. Surf. Sci. 14, 137 (1982–1983)
21. S. Saito, T. Maeda, J. Vac. Soc. Jpn. 25, 190 (1981)
22. Piezofoil can be obtained from Kyoto Ceramic Ltd., Kyoto, Japan
23. I.D. Baikie, E. Venderbosch, J.A. Meyer, P.J.Z. Estrup, Rev. Sci. Instrum. 62, 725 (1991)
24. see also J.S.W. de Boer, H.J. Krusemeyer, N.C. Burhoven Jaspers, Rev. Sci. Instrum. 44, 1003
(1973)
25. I.D. Baikie, S. Mackenzie, P.J.Z. Estrup, J.A. Meyer, Rev. Sci. Instrum. 62, 1326 (1991)
26. I.D. Baikie, K.O. van der Werf, H. Oerbekke, J. Broeze, A. van Silfhout, Rev. Sci. Instrum.
60, 930 (1989)
27. C.S. Kumar, A. Subrahmanyam, J. Majhi, Rev. Sci. Instrum. 67, 805 (1996)
29. L. Hanley, X. Guo, J.T. Yates Jr, J. Phys. Chem. 93, 6754 (1989)
30. J. Abelson, G. de Rosny, J. Physique 44, 993 (1983)
31. H.A. Engelhardt, P. Feulner, H. Pfnur, D. Menzel, J. Phys. E10, 1133 (1977)
32. I.D. Baikie, E. Venderbosch, J.A. Meyer, P.J.Z. Estrup, Rev. Sci. Instrum. 62, 725 (1991)
33. I.D. Baikie, S. Mackenzie, P.J.Z. Estrup, J.A. Meyer, Rev. Sci. Instrum. 62, 1326 (1991)
34. I.D. Baikie, K.O. van der Werf, H. Oerbekke, J. Broeze, A. van Silfhout, Rev. Sci. Instrum.
60, 930 (1989)
35. M. Chelvayohan, R. Gomer, Surf. Sci. 172, 337 (1986)
36. A Phillips PXE 5 piezoactuator with length of 27 mm was used
37. S. Lundgren, B. Kasemo, Rev. Sci. Instrum. 66, 3976 (1995)
38. H.L. Tierney, A.E. Baber, E.C.H. Sykes, J. Phys. Chem. C 113, 7246 (2009)
39. A.E. Baber, H.L. Tierney, E.C.H. Sykes, ACS Nano 4, 1637 (2010)
40. A. Itakura, I. Arakawa, J. Vac. Sci. Technol., A 9, 1779 (1991)
41. S. Igarsshi, Y. Abe, Y. Irie, T. Hirayama, I. Arakawa, J. Vac. Sci. Technol., A 16, 974 (1998)
42. S. Igarashi, A. Tosaka, T. Hirayama, I. Arakawa, Langmuir 19, 4627 (2003)
43. A. Tosaka, I. Arakawa, Surf. Sci. 600, 1071 (2006)
Chapter 28
Radioactive Adsorbates
28.1 Radiotracer Techniques for Surface Studies
The high sensitivity of radiochemical methods suggests that these methods would
have utility in studies of chemisorption and reaction on single crystal surfaces. This
description of the use of a silicon surface barrier detector for the measurement of
14
C in chemisorbed layers on Pt(111) represents one of the few attempts to make
measurements of this type [1]. Earlier studies using a Geiger-Müller counter were
also successful [2, 3].
The surface barrier detector [4] is a large area diode consisting of a depleted slice
of n-type Si mounted in an insulator ring as shown in Fig. 28.1a. The front and back
surfaces are metallized. The thickness of the Au front surface coating is optimized
for low energy β counting (158 keV) from 14C. Also, the depth of the depletion
region (300 μm) is chosen to optimize the detection of the β-electrons from 14C,
which have a range of about 150 μm in Si. Free charge carriers produced by the β-
emission are separated by an applied electric bias (50 V). Integration of the current
induced between the front and back electrodes yields an output pulse that is pro-
portional to the ionizing energy. Either β-spectra may be collected or a ratemeter
may be used to count the level of radioactive material, using the appropriate dis-
criminator settings to eliminate noise.
The detector, *20-mm in diameter, is mounted on a rotary feedthrough that
allows it to be positioned in front of the Pt crystal at a distance of 5–10 mm, as
shown in Fig. 28.1b. In the opposite orientation, background counting may be done.
When a bias is applied to the detector, the temperature must be maintained between
−30 °C and +30 °C. This is done by using a flexible Cu braid (see p. 392) that is
anchored on a copper block maintained at the desired temperature. To avoid con-
tributions from detector noise, the lower discriminator was set at 35 keV, and this
resulted in rejection of about 40 % of the 14C β-signal as determined in separate
calibrations. The counting efficiency was determined as 2.9 ± 0.3 % for the
geometry and discriminator settings used here.

DOI 10.1007/978-3-319-17668-0_28
306 28 Radioactive Adsorbates
Fig. 28.1 Radiotracer techniques for surface science. a Si surface barrier detector. b. Detector
geometry. c C Auger signal versus radioactive 14C coverage absolute surface coverages
28.1 Radiotracer Techniques for Surface Studies 307
Table 28.1 Partial listing of Isotope Emax (keV) Half-life (years)

radioisotopes useful for 14
surface studies using a barrier C 158 5720
11
detector [1] Na 544 2.6
26
Al 1160 740,000
32
Si 210 710
35
S 167 0.24
36
Cl 714 300,000
137
Cs 510 30
Figure 28.1c shows the correlation between the measured count rate and the
Auger intensity from carbon deposited as 14C–C6H6 on the Pt (111) crystal. The
linear correlation between the two measurements indicates that both are measuring
the carbon coverage as expected.
The sensitivity of the 14C counting method depends upon many factors such as
geometry, detector parameters, and the specific activity of the carbon-containing
compounds used. While many hydrocarbons and other molecules may be obtained
with activities up to 64 mCi/mmol, a nominal value of 10 mCi/mmol is more
common. Assuming a detection efficiency of 3 %, and a background signal of 3
counts/min, the detection sensitivities for 20-min counting, at nominal labeling
levels, are 1 × 1013 (23 % standard deviation), 5 × 1013 (10 % standard deviation),
and 1 × 1014 (4 % standard deviation) molecules/cm2. Thus, coverages of 1 % of a
monolayer may be determined accurately. It is found that repeated adsorption
experiments lead to increasing background radioactivity during measurements over
a day; pumping overnight restores the background to a low level.
A second example of the use of this method for the study of hydrocarbon
chemistry on Pt(111) can be found in [5].
The use of radioactive adsorbates is not limited to those labeled with 14C as may
be seen in Table 28.1.
References
1. S.M. Davis, B.E. Gordon, M. Press, G.A. Somorjai, J. Vac. Sci. Technol. A19, 231 (1981)
2. K. Klier, Rev. Sci. Instrum. 40, 372 (1969)
3. K. Klier, A.C. Zettlemoyer, H. Leidheiser Jr, J. Chem. Phys. 52, 589 (1970)
4. Ortec TA-23-25-300. Obtained from Ortec, P.O. Box 1397, Easley, SC 29640
5. S.M. Davis, F. Zaera, B.E. Gordon, G.A. Somorjai, J. Catal. 92, 240 (1985)
Chapter 29
Thin Film Deposition
29.1 Thin Film Deposition to Absolute Surface Coverages
The ability to evaporate clean thin films and to achieve desired surface coverages
that are absolutely known is of fundamental importance in many areas of surface
science research. Fortunately, quartz thickness monitors (QTM) are commercially
available that have sensitivities of the order of a nominal thickness of 1 Å of
deposited material. With appropriate use, these monitors can be used to monitor,
absolutely, the deposition of thin films on substrates of interest.
Figure 29.1a shows a successful arrangement for producing absolute coverages of
metal on a single crystal [1]. In the case of this work, the metal deposited was Pt and
the substrate was atomically clean Mo(110). The metal deposition source consists of
a tungsten hairpin filament, mounted on degassing loops (see pp. 407 and 437),
which is shielded by a metal shield containing a 15-mm-diameter hole near the
source. A thermostated QTM is able to be moved quickly between three positions—
1, 2, and 3. When the QTM is in position 1, the Pt source is completely shielded, so
that metal does not escape into the vacuum chamber. Position 1 is used for through
outgassing of the source. When the QTM is moved to position 2, the QTM collects
evaporated metal and radiation from the hot source. Thermostatting with circulating
water from a temperature-regulated source maintains the QTM temperature constant
to ±0.5 K, and this condition is necessary for absolute accuracy of the QTM. When
the QTM is rotated to position 3, the crystal receives the metal vapor and the QTM is
out of line-of-sight with the metal source. Stable operation of the metal source can be
achieved by monitoring the current and voltage drop across the source and making
the product of these constant. In practice, the source is stable over hours of operation
without sophisticated wattage regulation. The relative flux at the crystal and the
QTM is determined from the distances by the ratio, d2QTM/d2crystal.
Figure 29.1b shows the duty cycle described above as monitored by the QTM. In
position 1, no Pt is received either by the QTM or the crystal. In position 2, the
mass of the deposited platinum on the QTM rises linearly with time. In position 3,

DOI 10.1007/978-3-319-17668-0_29
310 29 Thin Film Deposition
Fig. 29.1 Thin Film Deposition to Absolute Surface Coverage. a Metal evaporation/quartz
thickness monitor. b Duty cycle for evaporation/measurement. c Layer-by-layer growth kinetics Pt
deposition measurement
29.1 Thin Film Deposition to Absolute Surface Coverages 311
Pt is deposited on the Mo(110) substrate and the QTM again does not show any
change in mass. Finally, in position 2 for the second time, the QTM again receives
Pt, to check that the rate of evaporation has not changed significantly during
deposition. The two slopes in position 2 agree to within 0.4 %.
Following each deposition cycle, the Mo(110) crystal was rotated to the CMA
Auger spectrometer position for measurement of the Pt(NW) intensity, which is a
secondary measure of the extent of Pt deposition. It can be seen in Fig. 29.1c that
the Pt Auger signal grows linearly in at least three branches of different slope. The
break points, located analytically, correspond to the completion of a monolayer of
Pt and the start of the formation of a second layer of Pt as growth proceeds by a
layer-by-layer mechanism. It is interesting to note that the three labeled break points
correspond well, in absolute Pt coverage as measured with the QTM, with the
number of Pt atoms in a close-packed Pt(111) layer which has a coverage of

1.504 × 1015 Pt/cm2. The location of the break points were used to calculate
the average attenuation length for 64 eV Pt(NVV) electrons, using the bulk
(111) interplane spacing for Pt of 2.47 Å. These experiments gave an average
attenuation length of 4.1 ± 0.3 Å, in close agreement with calculated values for this
quantity [1, 2].
Even with the care taken in this experiment, a propagation of systematic and
random errors results in a ±16 % error estimate in the absolute coverages measured.
The use of quartz microbalance measurements stretch back to the pioneering
work of Sauerbrey [3], Slutsky and Wade [4], and to Kasemo and Tornqvist [5],
who showed the usefulness and sensitivity of these devices for adsorption and thin
film studies. A good review of the history of the development of the quartz
microbalance may be found in [6]. A paper concerned with improvement in the
accuracy of the quartz microbalance to allow the measurement of the acoustic
impedence ratio of the film to the crystal, using two non-harmonically related
resonant modes is given in [7].
A method to compensate for the effects of temperature changes owing to radiation
heating is described in [8, 9]. Here, a thermocouple pair measures the difference
between the temperature of the quartz crystal detector and a matched reference
detector in the vacuum system. A heater, driven by a feedback circuit working off the
thermocouple pair adjusts the reference crystal temperature to within 0.1 K. An
accuracy of thickness measurement of at least 0.01 nm is estimated [8, 9].
A description of the calibration of a quartz microbalance against an electrostatic
balance is given in [10]. The probable error of the microbalance was found to be
0.6 %, compared to the theoretical expectation.
A description of the use of a quartz microbalance at cryogenic temperatures is
given in [11–13].
A description of the use of a quartz microbalance for the calibration of an Auger
spectrometer used to measure oxygen chemisorption on Ni is found in [14, 15].
Recent advances in improving the sensitivity and accuracy of the QTM tech-
nology depend upon improved temperature stability (±0.03 K) over several hours.
In addition temperature stabilization of the electronics were investigated, as well as
grounding effects. A frequency stability of 1:108 (0.06 Hz) over 0.5 h. was
achieved, equivalent to a mass resolution of 2 % of an atomic monolayer of O over
0.5 h. It was reported that irreversible uptake of oxygen into the bulk of the quartz
crystal on its rear face may be a factor causing error [16].
29.2 Gold Evaporation Source
The evaporation of Au is sometime difficult because Au droplets on a W wire have

a tendency to detach. A neat trick to prevent this is to produce a notched W wire
support [17] which is more readily wet by the Au than a wire without the Vee
shape. Apparently gravity and surface tension effects combine to hold the Au
29.2 Gold Evaporation Source 313
Fig. 29.2 Au Evaporation Source. a Au evaporation source. b Evaporation versus time. c Power
consumption (I × V) versus time
droplet in the notch region. In Fig. 29.2a, the evaporation source is shown. It
consists of a W wire [18] with the notch, welded between two Ta degassing loops.
These loops can be independently degassed at higher temperatures than employed
for the Au evaporation and eliminate outgassing effects, as described in Sect. 38.1.
Current and voltage measurements (Fig. 29.2b) are made as shown, allowing one to
measure the watts expended and to keep this quantity constant throughout the
evaporation. Measuring the voltage drop across the active W evaporation filament is
superior to measurements made across the entire device, since one wishes to control
the resistance and temperature of the active evaporation element. Figure 29.2c
shows a series of Au deposition rate measurements made over several months. The
reproducibly is better than ±20 %, and allows Au to be deposited far down into the
submonolayer range at a slow rate. The thickness measurements for the Au film are
made by measuring the Auger attenuation of the substrate, which in this case is a
TiO2(110) single crystal [19].
29.3 Ionization Gauge Measurement of Film Evaporation

Rate
A convenient and inexpensive secondary standard for the measurement of the rate
of metal evaporation may be established using a pair of ionization gauges with
matched geometry. The flux of metal atoms through one gauge is measured by an
increase in ion current owing to the metal atom steady state gas density during
evaporation. A second gauge is used to make a small correction for the presence of
residual gases produced during evaporation. The calibration of the ion gauge is
made against a quartz thickness monitor (QTM) [20] (see p. 309).
Figure 29.3a shows the design of the apparatus, where a translatable QTM is
located line-of-sight to a metal evaporation source that contains two separate
sources housed in a shield. It is desired to evaporate known coverages of metal onto
the crystal, and the QTM is located very near the crystal location. After outgassing
of the source, the crystal is moved to a line-of-sight position and measurements of
the mass received by the QTM are made. Simultaneously, measurements of the
difference between the two ionization gauge readings are integrated versus time,
and this difference is related to the metal vapor steady state gas density inside the
ionization gauge assembly, which is located line-of-sight between the source and
the crystal and the QTM.
Figure 29.3b shows a cross calibration obtained for Sn evaporation. The linear
correlation over a wide Sn thickness range indicates that both measurement
methods are behaving as expected. However, the sensitivity of the ion gauge is high
compared to the commercial QTM, so that small film thicknesses can be measured
with the ionization gauge conveniently.
The dual-ionization gauge arrangement could also be calibrated against other
thickness standards such as RHEED intensity measurements made during
layer-by-layer growth [21].
29.4 Reusable Quartz Crystals for Film Thickness

Measurement
Quartz crystals used in thin film measurement are able to hold only so much
deposited material before becoming unusable. Typical thickness limitations are
approximately 105 Å of metal deposit. The thin films can sometimes be etched
away to recover the clean quartz oscillator, but strong chemical etchants will often
destroy the metallic electrodes that activate the crystal, requiring redeposition of the
electrodes [22].
A simple solution has been found to allow thin film overlayers on quartz crystals
to be removed without harming the electrodes [23]. A photoresist coating is
29.4 Reusable Quartz Crystals for Film Thickness Measurement 315
Fig. 29.3 Ionization Gauge Measurements of film Evaporation Rate. a Ionization gauge geometry.
b Correlation of QTM and BAG
deposited on the clean quartz crystal before use. Later, using acetone as a solvent
for the photoresist, the thin film to be removed is stripped away by dissolving the
underlying photoresist layer as shown in Fig. 29.4. It was found that a photoresist
layer of at least 3000 Å thickness is needed to permit stripping of all overlayer
Fig. 29.4 Reusable quartz crystal oscillatorcoating tube inner walls
deposits cleanly. This loads the quartz crystal only to about 1 % of its mass
capacity. The photoresist [24] is deposited using a spin-coating method, baked at
120 °C for 20 min, and then patterned to cover the area on the crystal used for
deposition. The photoresist coatings made in this manner can be stored in air but
should not be baked at temperatures approaching 200 °C because this makes the
photoresist insoluble in acetone.
29.5 Hybrid Thin Film Deposition Processes—Stable and Metastable Films 317
29.5 Hybrid Thin Film Deposition Processes—Stable

and Metastable Films
The deposition of thin films with desirable chemical and physical properties is
widely used for the production of electronic materials, tribological coatings, and
optical coatings. The processes used may be divided into those that produce ther-
modynamically stable compound films and those that produce metastable materials.
It is beyond the scope of this book to describe the wide range of film deposition
methods and materials that have been deposited in thin film form. Books and
reviews are widely available. Below, a short description of the some of the hybrid
film deposition methods is given [25], and figures schematically describing each
method are shown. Figure 29.5a–d illustrate ion bombardment from an external
plasma. Figure 29.5e illustrates the use of ions extracted from a plasma near the
substrate, and this method is now used widely in plasma assisted vapor deposition
processes.
Figure 29.5a—Ion Beam Deposition: Ions are extracted from the source where
they are produced by plasma excitation of the appropriate gas. The method has the
advantage that the energy of deposition of the ions can be controlled rather than
working with a mixture of ions, neutrals, and excited neutrals.
Figure 29.5b—Ion Beam Sputter Deposition: Ions are produced in an ion source
by a plasma. These are used as a means for sputttering a target, and collection
occurs on the substrate.
Figure 29.5c—Dual Beam Sputtering: Two ion beams are used. One sputters a
target; the second bombards the growing film on the substrate.
Figure 29.5d—Ion Bombardment Evaporation: Material is evaporated thermally
using electron beam or ohmic heating. The growing film is bombarded by ions
extracted from an ion source.
Figure 29.5e—Reactive Sputtering with Dissociable Reactive Gas: Metals are
sputter-deposited in a reactive gas such as silane to produce metal silicide films.
Many variations on the schematic methods shown in Fig. 29.5a–e are employed
and a wide variety of preparation methods are described in the literature [26, 27].
29.6 Sputter-Coating Inner Walls of Tubes
The need to coat the inside of a metal or insulator tube with a metal film can be met
by both electrochemical and chemical vapor deposition methods. It is also possible
to deposit pure metal films inside a tube using a special sputtering procedure
described here [28].
As shown in Fig. 29.6, a conical metal target, placed on the axis of a polished
steel tube to be coated, is exposed to a beam of ions that are directed down the tube.
The conical metal target is moved along the tube toward the ion source by an
electrical drive mechanism. In the particular case studied here, the inner diameter of
Fig. 29.5 Hybrid thin film deposition processes. a Ion beam deposition. b Ion beam sputtering.
c Dual beam sputtering. d Ion bombardment—evaporative deposition. e Reactive sputtering—
dissociable reactive gases
the tube was 6 mm, and a conical Ag target of 5 mm diameter was employed.
The target was pushed through the tube at a rate of 0.05 mm/s, while a 2 keV Ar+
beam was incident. A sputter yield of about 7 Ag/Ar+ was measured for a 60°
incidence angle.
29.6 Sputter-Coating Inner Walls of Tubes 319
Fig. 29.6 Sputter-coating inner walls of tubes
The conical target was chosen so that the maximum sputter yield could be
achieved using a 60° angle of incidence for the ions on the target surface. This
optimum Ar+ incidence angle for Ag sputter deposition was determined experi-
mentally using a separate apparatus that permitted the collection of the Ag metal
sputter deposit and its measurement by chemical analysis.
To determine that the Ag film deposited was uniform in thickness, the tube was
cut along its axis in four sections. Visual inspection showed that a complete Ag film
was deposited along the tube length, and sputter-depth profile measurements
showed that depth of the film was uniform to within about 10 %.
Experiments with Pt deposition on Ta tubes were also carried out and subsequent
mechanical embrittlement tests showed that the Pt deposit was very effective in
preventing hydrogen-induced embrittlement of the Ta tubes.
References
1. A. Linsebigler, G. Lu, J.T. Yates Jr, Surf. Sci. 294, 284 (1993)
2. S. Tanuma, C.J. Powell, D.R. Penn, W.F. Egelhoff Jr, J. Vac. Sci. Technol A8, 2213 (1990)
3. G. Sauerbrey, Zeit. Phys. 155, 206 (1959)
4. L.J. Slutsky, W.H. Wade, J. Chem. Phys. 36, 2688 (1962)
5. B. Kasemo, E. Tornqvist, Surf. Sci. 77, 209 (1978)
6. C. Lu and A. W. Czanderna, in Applications of Piezoelectric Quartz Crystal Microbalances
(Elsevier, New York, 1983), Chapter 2
7. A. Wajid, Rev. Sci. Instrum. 62, 2026 (1991)

8. G. Kremer, L.A. Moraga, Rev. Sci. Instrum. 56, 1467 (1985)
9. B. Ramadan, K. Piyakis, J.R. Kos, Rev. Sci. Instrum. 50, 867 (1979)
10. R.M. Mueller, W. White, Rev. Sci. Instrum. 39, 291 (1968)
11. L.L. Levenson, Supple. al Nuovo Cimento, Serie I, 5(2), 321 (1967)
12. L.L. Levenson, J. Vac. Sci. Technol. 8, 629 (1971)
13. C. Wang, R. Gomer, Surf. Sci. 74, 389 (1978)
14. A. Winkler, K.D. Rendulic, K. Wendl, Appl. Surf. Sci. 14, 209 (1982–83)
15. A. Winkler, Fresenius. Z. Anal. Chem. 319, 635 (1984)
16. A.J. Slavin, Rev. Sci. Instrum. 81, 1 (2010)
17. D.W. Goodman, private communication
18. H. Luth, Surfaces and Interfaces of Solid Materials, 3rd edn. (Springer, Berlin, 1995), p. 39
19. Z. Zhang, W. Tang, M. Neurock, J.T. Yates Jr, J. Phys. Chem. C 115, 23848–23853 (2011)
20. D.S. Campbell, in Physics of Nonmetallic Thin Films, ed. by C.H. Dupuy, A. Cachard
(Plenum, NY, 1976), p. 179. I first learned of this technique from Dr. W.F. Egelhoff, Jr. of the
National Institute of Standards and Technology (NIST). The work shown here is from P.
J. Chen, M.L. Colaianni, J.T. Yates, Jr., and M. Arbab, J. of Non-Crystalline Solids 155, 131
(1993)
21. D.A. Steigerwald, I. Jacob, W.F. Egelhoff Jr, Surf. Sci. 202, 472 (1988)
22. Recoating services are provided by Lebow Company, 5960 Mandarin Avenue, Goleta, CA
93117
23. W.K. Chao, H.K. Wong, J. Vac. Sci. Technol. A8, 150 (1990)
24. Photoresist AZ-1300-31 purchased from Shipley Company, Inc., 2300 Washington Street,
Newton, MA 02162
25. R.F. Bunshah, Miner. Process. Extr. Metall. Rev. 9, 147 (1992)
26. R.F. Bunshah, in Evaporation in Deposition Technologies for Films and Coatings, ed by R.F.
Bunshah, Second edn. (Noyes Publications, Park Ridge, 1994) p. 131
27. C.V. Deshpandey, R.F. Bunshah, Thin Solid Films 163, 131 (1988)
28. W. Ensinger, Rev. Sci. Instrum. 67, 318 (1996)
Chapter 30
Infrared Spectroscopy
30.1 Infrared Reflection Absorption Spectrometer
The use of infrared reflection-absorption spectroscopy (IRAS) has grown in

popularity since the publication of the classic papers by Greenler on this subject.
While the first IRAS measurements in ultrahigh vacuum were made with grating
spectrometers [1–3] the measurement technique has been revolutionized by Fourier
transform (FTIR) spectrometers, which possess greatly enhanced signal-to-noise
capabilities at high spectral resolution [4]. Various technical methods have been
used to improve the performance of FTIR spectrometers operating as IRAS
instruments [5–11]. The field has been well reviewed [12, 13].
The instrument shown in Fig. 30.1a, b has special characteristics which are
designed to improve the performance beyond that of conventional single beam
IRAS instruments. A collimated IR beam from the FTIR spectrometer is reflected
into an external optical bench by a flat mirror. The beam is split into two parts by a
right-angled aluminized beam splitter (made from a 90° glass prism) in the external
optical bench. One beam enters the IRAS cell containing the crystal, and the second
beam enters the reference cell, which is at ultrahigh vacuum. The beam splitter is
able to be moved laterally by a micrometer drive, a novel feature which allows the
ratio of the two beam intensities to be accurately and reproducibly adjusted. Both
beams are focused with KBr lenses. A beam collector, which is identical to the
beam splitter, collects both infrared beams from the IRAS cell and the reference cell
and reflects them onto the focussing mirror to the mercury cadmium telluride
(MCT) detector. Two electrically driven shutters in the infrared beams enable the
detection of these beams independently. For study of adsorbates on metal single
crystals, the use of a p-polarizer in the incident beam, as shown, will improve the
spectroscopic signal-to-noise ratio, since the polarization selection rule is strongly
obeyed in this case, and s-polarized light is ineffective in detecting the adsorbate.
The collection of infrared spectra is interfered with by the presence of H2O and
CO2 in the atmosphere within the optical path outside the ultrahigh vacuum system.

DOI 10.1007/978-3-319-17668-0_30
322 30 Infrared Spectroscopy
Fig. 30.1 Infrared reflection-absorption spectroscopy—IRAS. a IRAS-FTIR spectrometer for

surface studies (N2-Purged)—top view. b Vacuum system—side view. c Comparison of two data
acquisition methods in IRAS (6 h after baseline)
The “double-beam” arrangement shown here reduces this contribution significantly.

As an example, if a H2O band is 5 % of the total transmission, a 1 % drift of the
H2O partial pressure will yield, in a direct ratio measurement, a 0.05 % H2O
absorption band which would be greater in intensity than many weak surface
species’ absorption bands. Therefore, the optical path is purged extensively with
very dry nitrogen gas from boil off of a liquid nitrogen dewar. With the
double-beam method, small variations in the purged gas background spectrum can
be reduced as shown next.
In single-beam measurements, if the background spectrum signal is S0 and the
adsorbate spectrum signal is S1, then the desired adsorbate spectrum is log(S1/S0),
which includes any possible changes during the time period between the mea-
surements of S0 and S1. By the double beam method, we record the background
absorbance spectrum as log(S0/R0) and the adsorbate absorbance spectrum as log
(S1/R1). Then the difference absorbance spectrum is obtained by subtracting the
background absorbance spectrum from the adsorbate absorbance spectrum,
logðS1 =R1 Þ logðS0 =R0 Þ ¼ logðS1 =S0 ÞðR0 =R1 Þ ð30:1Þ
In (30.1), the term R0/R1 produces a cancellation of all changes in environment

during the period of taking the two spectra. This could include drifts of composition
of the purging gas, drifts in the source and detector, etc. In doing measurements on
30.1 Infrared Reflection Absorption Spectrometer 323
the crystal in the optical cell with a high pressure of gas also present, the spectral
contribution from the gas phase can be canceled almost completely.
As shown in Fig. 30.1b, the single crystal is moved by a long throw manipulator
from the surface analytical positions in the ultrahigh vacuum chamber (Level 2 and
Level 3) up into the IRAS cell (Level 1). The crystal can be rotated through 360° by
means of a doubly differentially pumped Teflon rotary seal (see p. 6) for precise and
reproducible angular orientation to the IR beam, which typically has an angle of
incidence of about 86°. The windows of the IR cell can be made of single crystal
CaF2 mounted in Conflat flanges [4] or can be made of KBr single crystals mounted
by means of a differentially pumped double Viton seal assembly (see p. 521). The
small optical chamber is sealed by a metal-seal gate valve that does not have the
problem of Viton O-ring outgassing under compression nor the problem of per-
meation. For a sealed-off static cell, this is important. For high pressure experi-
ments, the composition and pressure of the gas phase may be determined by
spectroscopic measurements through the reference cell for infrared active gases.
Figure 30.1c shows a comparison of operation of the IRAS apparatus in single
beam mode and in double-beam mode. The optical path is purged with dry nitrogen.
After 6 h, a severe intensity drift in atmospheric H2O (about 0.4 % T) and atmo-
spheric CO2 (about 0.2 % T) can be seen in the spectrum obtained by the single
beam method. In the spectrum obtained by the double-beam method, this effect is
eliminated. The comparison on Fig. 30.1c shows the ability of the double-beam
method to eliminate background effects owing to slow drifts during long periods.
For 900 scans for both S and R infrared measurements, a noise level of about
1 × 10−4 absorbance unit may be routinely obtained. The amount of light delivered
to the detector is significant in determining the noise level. Therefore, large single
crystals of the order of 15 mm diameter are favored in this work.
A FTIR spectrometer mounted on a precision optical mount has been employed
for IRAS measurements in which the path length between the spectrometer and the
external detector has been minimized by using only KBr lenses for focusing the
radiation [14]. Both the spectrometer and the detector are coupled to the UHV
chamber by small bellows, permitting each unit to be independently adjusted with
precision XYZ optical mountings.
Recently, an instrument designed to measure the mass of adsorbate on a thin film
in addition to the infrared spectrum has been described [15]. The working tem-
perature range of the quartz crystal oscillator (see p. 309) used for adsorbate mass
measurements is 80–350 K [15].
The instrument shown in Fig. 30.1 is described in detail in [16].
30.2 IRAS Measurements at High Pressure
The IRAS apparatus described on p. 321 has the special capability if working at
elevated pressures. At these pressures, electron spectroscopies fail because of the
reduction of the mean free path of electrons to lengths comparable to or less than
the electron traveling distances inside the apparatus. Fortunately, no such limita-
tions exist for optical techniques, assuming the gas phase at high density is optically
transparent in the required spectral regions.
An example is shown of the study of N2 adsorption on the Ni(111) surface.
Figure 30.2a shows that as the N2 pressure is raised over 4 orders of magnitude from
1 × 10−8 to 2.9 × 10−4 mbar, the equilibrium coverage of chemisorbed N2 at a crystal
temperature of 89 K increases as indicated by the increase in the absorbance of the
chemisorbed N2. Data of this sort were collected as a function of crystal temperature
in the range 89–115 K, and the results are shown in Fig. 30.2b, where the vertical
scale is the integrated IRAS absorbance. It is clear that as the temperature is
increased, higher pressures of N2 must be employed to produce a given integrated
absorbance for the chemisorbed N2. If one assumes that the integrated absorbance is
proportional to surface coverage (an approximation of unknown accuracy) then the
“isoabsorbance” plots versus reciprocal absolute temperature can be made, as shown
in Fig. 30.2c, and from the slope of these plots the enthalpy of adsorption may be
deduced at particular integrated absorbances. The inset to Fig. 30.2c shows this plot,
indicating that the enthalpy of adsorption is almost constant, or rises very slightly
with increasing absorbance (or coverage within the approximation used above).
Work of this type, in which an initially clean single crystal is exposed to high
pressures of a gas, must be carefully done with highly purified gas in order to avoid
contamination. For the example of about 10−4 mbar of gas, impurity levels of only
10 ppm will give sufficient impurity partial pressure that in about 1000 s of
exposure, one monolayer of impurity could be adsorbed if the impurity has a
sticking coefficient of unity over its coverage range in the presence of the dominant
gas. Effects of this type can often be discovered by examining the reversibility of
the spectral changes, or by desorption of the dominant adsorbate, followed by
readsorption. The purification of gases is discussed on pp. 505–514.
The slight changes in the vibrational spectrum of the chemisorbed N2 are
thought to be due to various island structures that develop at various equilibrium
coverages. The fact that so little shift occurs in the N–N stretching mode (a 24 cm−1
range) suggests that the electronic character of the chemisorbed N2 is not strongly
affected by changes in coverage, and suggests that the approximation that integrated
absorbance is proportional to coverage may be valid [17].
30.3 Internal Reflection IR Spectroscopy—Silicon
The use of infrared spectroscopy to probe the vibrational spectrum of adsorbates on

semiconductor surfaces has opened up a wide range of new physical understanding
of surface species structure and dynamics. In contrast to the reflection IR method
used on metallic surfaces (see p. 321), it is most appropriate to use internal
reflection IR spectroscopy for semiconductor surfaces. This permits the IR beam to
make many reflections at the interface, increasing the signal-to-noise ratio
significantly.
30.3 Internal Reflection IR Spectroscopy—Silicon 325
Fig. 30.2 IRAS measurements at high pressure. a Development of IRAS Spectra of N2/Ni(111).
b N2 Adsorption Isotherms on Ni(111) by IRAS. c Isosteric heat of adsorption N2/Ni(111) Infrared
reflection-absorption spectroscopy (IRAS)
Fig. 30.3 Internal refection IR spectroscopy—semiconductors. a U-shaped Si single crystal plus

vector diagram for p- and s-polarization. b UHV optical system
Figure 30.3a shows the geometrical arrangement used for holding a silicon
single crystal on a manipulator and for making connections for electrical heating
[18]. Here, Ta clips are attached to the top of the U-shaped crystal and the crystal
may be heated by the passage of current. Conductive cooling to the clips causes a
temperature gradient across the crystal (100 K at 1000 K), but in the portion probed
by the IR beam, only a small gradient is observed. The edges of the crystal are
beveled and polished for admittance of the focused IR beam. The vector diagram
for p- and s-polarization is also shown in Fig. 30.3a. Oxidation of the top surface
deactivates this surface for adsorption.
30.3 Internal Reflection IR Spectroscopy—Silicon 327
Figure 30.3b shows the UHV system and the associated optical system for use in
the FTIR internal reflection measurements [19]. The U-shaped silicon crystal is
mounted on a manipulator and may be rotated to an Auger spectrometer or to a
LEED apparatus. In its position for internal reflection spectroscopy, the focused IR
beam from the FTIR spectrometer passes through a CsI IR window (see p. 521) and
enters the silicon sample at the proper incident angle. After many internal reflec-
tions, the exiting radiation passes through another CsI window and is focused by a
mirror through a polarizer and onto the detector.
For silicon, the absorption of infrared light below about 1450 cm−1 prevents
transmission with typical sample lengths of 3–5 cm.
Excellent reviews of this field, giving many other optical arrangements for doing
IR surface spectroscopy on semiconductors, are available [20–25].
30.4 Internal Reflection IR Spectroscopy—Gallium

Arsenide
The use of internal reflection IR spectroscopy for silicon single crystals has been
discussed on p. 324. Different problems exist when the same measurement tech-
nique is applied to gallium arsenide. In order to permit IR transmission through
GaAs, the carrier concentration needs to be kept below about 1016/cm3 [26, 27].
GaAs with this level of carriers is commercially obtainable from one supplier,
which produces 1 Ω-cm material [27]. Normally, wafer GaAs is sold with con-
ductivities of about 106 Ω cm (semi-insulating), or about 10−2 Ω cm. While the
semi-insulating material would be satisfactory for IR measurements, it cannot be
conveniently heated by resistive heating; alternatively, the highly doped material
can be heated electrically, but its resistance is so low that large currents are needed
and temperature nonuniformities are produced [26]. This material is also opaque in
the infrared [26].
The mounting of 1 Ω-cm material for internal reflection infrared spectroscopy is
shown in Fig. 30.4. The infrared beam enters and exits through polished beveled
edges. Heating current flows perpendicular to the IR beam direction, using Ta
mounting clips on the top and bottom of the crystal. A thermocouple is conve-
niently glued to the crystal using a refractory cement, used also for thermocouple
attachment to other semiconductor crystals such as SiC [28]. Welding the ther-
mocouple first to a Ta foil square (1.2 mm × 1.2 mm) and then gluing this to the
GaAs provides good agreement between the thermocouple and pyrometry [26]. The
advantages of this mounting arrangement are that either surface can be studied, or if
both are studied together a two-fold improvement in the signal strength in the FTIR
measurement is obtained. The two sides of the crystal, if terminated by the same
element, exhibit orthogonal orientation of the dangling bonds. Thus, using s- and p-
polarized IR radiation, one can probe the adsorbate orientation [26]. A disadvantage
of working with the 1 Ω-cm material is that the resistance becomes too great for
Fig. 30.4 GaAs—mounting for internal reflection IR spectroscopy gallium arsenide
electrical heating when the sample is cooled to 120 K. This problem can be alle-
viated by warming the sample slightly with a projector lamp (see p. 371) [26].
30.5 Infrared Spectroscopy—Continuous Pumping

of Detector Dewar
InSb and HgCdTe infrared detectors are used in the mid-IR range (4000–500 cm−1)
for the study of adsorbed species on surfaces. The detector element is mounted on
the end of a small metal Dewar which is kept filled with liquid nitrogen during
operation, as shown in Fig. 30.5. The Dewar is often pumped on a separate vacuum
system and then sealed with a Viton-sealed valve. In addition, the detector window
is often sealed with a Viton O-ring. Thus, over time, slow permeation, outgassing,
and leakage can be expected (see p. 108), and the condensation of water and other
materials may occur on the detector surface, causing interference with infrared
measurements. These problems can be prevented by using a small appendage ion
pump to continuously pump the detector [29].
The IR detector is connected to a metal bellows-sealed valve and thence to a tee
as shown in Fig. 30.5. One branch of the tee is connected to the ion pump via a
1.5-cm-diameter bellows that extends up to about a 30-cm length as it makes the
transition from the detector to the ion pump. The second branch is temporarily
connected to an oil diffusion pump (liquid nitrogen trapped) for preliminary
pumping of the Dewar and the appendage pump.
30.5 Infrared Spectroscopy—Continuous Pumping of Detector Dewar 329
Fig. 30.5 Ion-pumped IR detector
Initially the entire system is pumped out by the diffusion pump. The valve on the
detector is then closed and the ion pump (not operating) and the metal tubing are
baked out with heating tape. When the pressure is low enough that the ion pump
can be operated, the valve to the diffusion pump is closed and the diffusion pump is
removed from the vacuum system. The heating is discontinued and when the line is
cooled, the valve to the detector is opened. Under these conditions, the IR detector
will be continuously pumped and condensation effects will be eliminated on the
detector. It is wise to place a conical shield just below the top of the Dewar entrance
to prevent liquid nitrogen from accidentally pouring over the detector window and
the vacuum seal for the amplifier wires, since these vacuum seals may be broken by
such accidental cooling [30].
30.6 Rapid IR Spectrometer Purge After Cell

Transfer—Using Rubber Sleeves
Infrared spectrometers require that the optical path length be free of absorbing
atmospheric gases. Both CO2 and H2O vapors must be removed by using either
N2(g) or specially-dried and CO2-scrubbed compressed air. Some IR spectrometers
have a sample compartment which is separately purged from the spectrometer
Fig. 30.6 Rapid IR

spectrometer purge after cell
transfer—using rubber
sleeves. a Sample cell in IR
spectrometer. b Purge
performance—CO2 and H2O
absorbance
30.6 Rapid IR Spectrometer Purge After Cell Transfer—Using Rubber Sleeves 331
source and detector compartments. The sample compartment is usually rather large
to admit large IR cells. It usually contains a sample compartment lid which is
opened and closed as the cell is transferred in and out. The cell and its connecting
tubing and wiring will cause problems with compartment sealing, requiring cutting
of the lid and unsatisfactory sealing with tape, etc. In Fig. 30.6a, an arrangement for
a portable cell is shown in which custom-cut rubber purge gas sleeves are arranged
to slide into the sample compartment under slight compression, thereby making a
gas tight seal with the inner walls of the sample compartment and the inserted IR
cell. The seals are held on the cell window flanges with band clamps. With this
arrangement, the IR spectrometer remains purged always, and the small volume
within the rubber sleeves is separately purged with great rapidity. Figure 30.6b
shows a comparison of the purge rate for the sleeves, compared to the entire
spectrometer, for CO2 and H2O. It is seen that the small-volume rubber sleeves are
purged about 25 times more rapidly than the entire spectrometer sample compart-
ment. In addition, complex gas-tight seals involving the compartment lid are
avoided, and the lid may be removed permanently. Soft rubber sleeves may be
obtained in several sizes from the plumbing department of a local hardware store,
where they are sold as flexible couplings for sewer systems. To cut the length to fit,
a mandrel is used on a lathe, and a smooth cut suitable for purge gas sealing at the
wall is produced with a sharp blade. This apparatus was produced in [31].
References
1. J. Pritchard, M.L. Sims, Trans. Far. Soc. 66, 427 (1970)

2. M.L. Kottke, R.G. Greenler, H.G. Tompkins, Surf. Sci. 32, 231 (1972)
3. J.T. Yates Jr, R.G. Greenler, I. Ratajczykowa, D.A. King, Surf. Sci. 36, 739 (1973)
4. B.E. Hayden, in Vibrational Spectroscopy of Molecules on Surfaces, eds. by J.T. Yates, Jr., T.
E. Madey (Plenum, New York, 1987)
5. E. Schweizer, B.N.J. Persson, M. Tüshaus, D. Höge, A.M. Bradshaw, Surf. Sci. 213, 49
(1989)
6. W. Erley, J. Electron Spectrosc. Related Phenom. 44, 65 (1987)
7. W. Erley, Surf. Sci. 205, L771 (1988)
8. D. Höge, M. Tüshaus, A.M. Bradshaw, Surf. Sci. 207, L935 (1988)
9. T. Buffeteau, B. Desbat, J.-M. Turlet, Mikrochim. Acta (Wien) II, 23 (1988)
10. W.G. Golden, in Fourier Transform Infrared Spectroscopy, vol. 4, eds. by J.R. Ferraro, L.
J. Basile (Academic, New York, 1985), p. 315
11. A.E. Dowrey, C. Marcott, Appl. Spectrosc. 36, 414 (1982)
12. F.M. Hoffmann, Surf. Sci. Rep. 3, 107 (1983)
13. Y.J. Chabal, Surf. Sci. Rep. 8, 211 (1988)
14. Q. Wu, L. Hanley, J. Phys. Chem. 97, 2677 (1993)
15. G. Zhuang, G. Chottiner, I.T. Bae, E. Hwang, D.A. Scherson, Rev. Sci. Instrum. 65, 2494
(1994)
16. Z. Xu, J.T. Yates Jr, J. Vac. Sci. Technol. A8, 3666 (1990)
17. J. Yoshinobu, R. Zenobi, J. Xu, Z. Xu, J.T. Yates Jr, J. Chem. Phys. 95, 9393 (1991)
18. Y.J. Chabal, Bell Laboratories, Lucent Technologies (personal communication). See Y.
J. Chabal, Phys. Rev. Lett. 50, 1850 (1983a).
19. Y.J. Chabal, Lucent, AT&T (personal communication). See Y.J. Chabal, E.E. Chaban, S.B.
Christman, J. Electron Spect. Rel. Phenom. 29, 35 (1983b)
20. N.J. Harrick, Internal Reflection Spectroscopy (Wiley, New York, 1967)
21. R.M. Mirabella, N.J. Harrick, Internal Reflection Spectroscopy: Review and Supplement
(Harrick Scientific Corporation, Ossining, NY, 1985)
22. Y.J. Chabal, in Semiconductor Interfaces: Formation and Properties, eds. by G. LeLay,
J. Derien, N. Boccara, Springer Proc. Phys. 22, 301 (1987)
23. Y.J. Chabal, Surf. Sci. Rep. 8, 211 (1988)
24. Y.J. Chabal, in Internal Reflection Spectroscopy: Theory and Applications, ed. by R.M.
Mirabella, Jr. (Marcel Dekker, New York, 1993), printed in Practical Spectroscopy, 15, 191
(1993)
25. M. Katiyar, J.R. Abelson, J. Vac. Sci. Technol. A13, 2005 (1995)
26. Dr. J.R. Creighton, Dept. 1126, MS 0601, P.O. Box 5800, Sandia National Laboratories,
Albuquerque, NM 87185-0601 (personal communication). J.R. Creighton, Appl. Surf. Sci.
82/83, 171 (1994)
27. GaAs with no intentional dopants (HB material, about 1 Ω-cm) is available from Sumitomo
Electric, 551 Madison Avenue, New York, NY 10022
28. F. Bozso, L. Muehlhoff, M. Trenary, W.J. Choyke, and J.T. Yates, Jr., J. Vac. Sci. Technol.
A2, 1271 (1984). Aremco 516 ceramic cement is used. Obtainable from Aremco Products,
Inc., P.O. Box 429, 23 Snowden Avenue, Ossining, NY 10562-0429; the cement should be
preserved in the refrigerator and has a limited shelf life
29. Contributed by Dr. Victor M. Bermudez, Naval Research Laboratory, Code 6863,
Washington, DC 20375-5347
30. Developed in author’s laboratory
31. A. Stevanovic, M. Büttner, Z. Zhang, J.T. Yares Jr, J. Am. Chem. Soc. 134, 324 (2012)
Chapter 31
Calorimetric Heats of Adsorption—Single
Crystals
31.1 Measurement of the Calorimetric Heat of Adsorption

on Ultrathin Metal Single Crystals
There are many methods for measuring the energy required for desorption of a
molecule from a single crystal surface using either isosteric heat measurements
under equilibrium conditions, or using desorption kinetic methods. The interpre-
tation of the isosteric measurements requires perfect reversibility in adsorption and
desorption, a condition not met when compound formation or dissolution of the
adsorbate in the crystal occurs. The interpretation of the kinetic measurements are
beset with difficulties having to do with the inability to write the desorption rate
laws that apply with accuracy. The calorimetric method avoids these difficulties by
directly measuring the heat evolved upon adsorption and by using only the first law
of thermodynamics.
In order to employ a calorimetric device, the heat capacity of the calorimeter
must be made very small in order for a measurable temperature rise to be produced.
This has been cleverly done, using single crystal films of metals which are only
2000 Å in thickness [1], using apparatus as shown in Fig. 31.1. These films are
grown epitaxially on an alkali halide crystal surface, and the continuous metal film
is then lifted off of the salt crystal as it dissolves in water. The metal crystal film is
transferred and attached to a ring, made of the same metal, and may be placed at the
focus of a pulsed molecular beam of the adsorbate. The thin crystal may be cleaned
in UHV by ion bombardment and by heating in vacuum or in reactive gases used
for cleaning. Standard surface analysis and LEED studies are routinely performed
on the thin crystal after cleaning. As the gas pulses arrive and are adsorbed on a
2-mm-diameter spot, the temperature of the thin crystal rises, and the pulsed
emission of infrared radiation is measured from the rear of the crystal by focusing
the radiation on a mercury cadmium telluride detector held at liquid nitrogen
temperature. Temperature rises on the crystal per pulse are below 0.1 K for
chemisorbing gases, where each pulse delivers about 0.01 mL of chemisorbed

DOI 10.1007/978-3-319-17668-0_31
334 31 Calorimetric Heats of Adsorption—Single Crystals
Fig. 31.1 Calorimetric adsorption heat—ultrathin metal films
species to the surface. The emissivity of the backside of the crystal is significantly
enhanced by initially coating it with amorphous carbon from a candle flame. For
calibration, a pulsed laser beam of known energy per pulse is used, and its
reflectivity at near normal incidence is measured. The temporal character of the
laser beam is matched to that of the pulsed molecular beam in the calibration
measurements. The energy losses after each pulse are found to be due to radiation
losses only.
In addition to the errors in measuring the small pulses of heat, errors in the
coverage measurement (and the sticking coefficient) currently limit this method [1].
The sticking coefficient and absolute coverage are measured using the method of
King and Wells [2], which measures the intensity of the reflected gas from the
crystal during each pulse. A gold flag is employed in these measurements.
A stagnation detector (see p. 429), which is a Bayard-Alpert gauge with a small
aperture, is employed for measurements of the absolute beam flux, as discussed on
p. 314.
31.2 Improved Adsorption Calorimeter for Single

Crystals at 100–350 K
The pioneering development of single crystal adsorption calorimetry (SCAC) by

the King group as documented in Sect. 31.1 has been markedly improved. A new
heat detector principle has been developed by the Campbell group, and the latest
stage of development is shown here [3]. A metal-coated pyroelectric polymer
ribbon (uniaxially oriented β-polyvinylidene fluoride, β-PVDF) of 9 µm thickness
and 4 mm width is employed as the detector, and electrical contact is made with the
front and back surfaces using Al film contacts. These films, with Al coatings, are
31.2 Improved Adsorption Calorimeter for Single Crystals at 100–350 K 335
available from Goodfellow. The PVDF detector is bent into an arch and gently
pressed to the back of the crystal which has been thermally cleaned in UHV before
contact. Pulses of adsorbent gas molecules, arriving at 0.5 Hz, cause temperature
bursts due to the evolution of the adsorption energy. The transient rise of crystal
temperature causes part of the PVDF detector to change its degree of polarization,
releasing an electrical charge through an electronic circuit. The ease of using the
PVDF detector compared to others allows the crystal to be cleaned at high tem-
peratures and also allows calorimeter operation below 300 K. This permits a wide
range of heat-of-adsorption experiments to be performed on thicker (*75 µm)
single crystals than previously possible.
The latest advances in the technique are sketched in Fig. 31.2A–D. It has been
found that noise, caused possibly by electrical leakage at the PVDF edges, can be
reduced by etching the Al film coating away from the ribbon edge. Figure 31.2 A
shows this: (a) indicates the aluminized PVDF film; (b) shows masking of the Al
Fig. 31.2 Improved adsorption calorimeter for single crystals at 100–350 K. A Electrodes on
PDVF front and back surfaces. B Bent PDVF sensor. C contact of sensor with thin sample.
D Experimental adsorption measurement
336 31 Calorimetric Heats of Adsorption—Single Crystals
film on front and back faces of the ribbon with nail polish; (c) shows the
electrochemically-etched ribbon with preservation of Al only under the nail polish
covering; (d) shows the final assembly with the removal of the nail polish with
methanol.
Figure 31.2B shows a schematic of the assembly for holding the bent PVDF
detector and for making electrical contact to both faces. A copper bar of thickness t,
of the order of a few mm, is used to define the width of the arch and the projection
of the arch. For contact with the crystal, the arch is compressed reproducibly by
0.5–1 mm.
Figure 31.2C shows a schematic of the design to compress the arch of the PVDF
detector against the back of the sample. Self-aligning conical holders are driven into
contact, and care in design prevents appreciable bending of the thin crystal sample.
Figure 31.2D shows a series of three experimental measurements of the heat
evolution during the adsorption of cyclohexene on Pt(111). Pulses of gas, involving
1.1 % of a monolayer are incident. As the coverage increases for the chemisorbed
cyclohexene, the heat of adsorption per pulse drops due to site blocking as well as
other effects. At 0.25 mL, the monolayer is complete and second layer adsorption,
with a constant heat of adsorption, takes place. The raw calorimetry data are shown
in the inset, and repeated experiments show a pulse-to-pulse standard deviation of
about 1.3 kJ/mol. The thermal energy released per pulse is 0.24 µJ which corre-
sponds to a temperature rise in the adsorption spot on the crystal of *7–8 mK. The
detector temperature rises in each pulse by 0.3 mK. The use of a pulsed He–Ne
laser on the crystal allows a calibration of the PVDF detector to give energy
liberated at the crystal surface. The improvements shown here have yielded a * 40
fold improvement in the pulse-pulse standard deviation over earlier PVDF detec-
tors. The paper [3] describes future developments to further enhance detection
sensitivity.
References
1. C.E. Borroni-Bird, D.A. King, Rev. Sci. Instrum. 62, 2177 (1991)
2. D.A. King, M.G. Wells, Surf. Sci. 29, 454 (1972)
3. W. Lew, O. Lytken, J.A. Farmer, M.C. Crowe, C.T. Campbell, Rev. Sci. Instrum. 81, 024102
(2010)
Chapter 32
Surface Debye Temperature
32.1 Measuring the Surface Debye Temperature

of a Single Crystal
The Debye temperature of a bulk solid is a measure of the stiffness of the bonds
within the crystal and is related to the Debye characteristic frequency, νD. This
frequency corresponds to the highest allowed vibrational mode in the crystal in the
distribution of frequencies following a mode frequency distribution function with a
magnitude which is proportional to the square of the mode frequency. The Debye
temperature is HD ¼ hmD =k, where k is Boltzmann’s constant.
The Debye-Waller factor, 2M, is connected to the Debye temperature through
(32.1), and determines the intensity of scattering of an electron beam incident on the
crystal through (32.2).

2M ¼ 24me E cos2 a þ V0 T=ma kðHD Þ2 ð32:1Þ
where ma = mass of the atoms in the crystal, me = mass of the electron, α = the angle
of incidence, V = the inner potential, and E is electron energy.
For an electron beam incident on a crystal, in the kinematic approximation,
I ¼ I0 e2M ð32:2Þ
where I0 is the incident intensity and I is the diffracted intensity.

Figure 32.1a shows a four-grid LEED apparatus with a movable Faraday cup to
intercept a selected diffraction beam [1]. The crystal surface in inclined by angle α
from the incident beam, and the diffracted 0,0 beam is being collected in the
drawing. As the crystal cools, the intensity of the diffracted beam increases as a
result of the decrease in the vibrational motion of the atoms in the region of the
solid surface being sampled by the incident electrons. As the energy of the incident
electron beam is decreased in successive experiments, a lower sampling depth is

DOI 10.1007/978-3-319-17668-0_32
338 32 Surface Debye Temperature
Fig. 32.1 Measuring the

surface debye temperature of
a single crystal. a LEED
Apparatus with Faraday Cup.
b 0,0 Beam Intensity, Rh
(111). c Debye temperature,
ΘD(K)
32.1 Measuring the Surface Debye Temperature of a Single Crystal 339
involved in the diffraction process, and the measured Debye temperature decreases
as a result of the looser binding of the crystal atoms near to the surface.
Figure 32.1b shows a Debye-Waller plot of lnðI Ibk Þ=I0 versus temperature for
various values of ΔK, the perpendicular component of the momentum transfer
between the electron and the metal atoms in the crystal. Here, Ibk is a background
scattering contribution which can either be measured in the vicinity of the 0,0 beam,
or treated parametrically as a variable. In this case, for low angles of incidence, one
can only measure an effective perpendicular Debye temperature as ΔK is varied [2].
Figure 32.1c shows typical behavior of the Debye temperature measurements for
three angles of incidence on a Rh(111) crystal as a function of ΔK. It may be seen
that all of the measurements show that as ΔK increases, and the depth of sampling
of the crystal by the electron beam increases, the measured value of ΘD increases to
a plateau at ΘD = 370 K, which is characteristic of the bulk of the crystal. The lower
limit of ΘD is near 200 K, and is characteristic of the surface metal atoms.
These measurements were made with a LEED apparatus mounted between two
pairs of Helmholtz coils to cancel the earth’s magnetic field. The values of the
diffracted beam intensity were measured as the crystal slowly cooled to avoid the
serious influence of the magnetic field associated with the electrical heating current.
The Faraday cup aperture subtended an angle of 4.6 × 10−3 steradians, and was
large enough to collect the entire beam, I.
References
1. C.M. Chan, P.A. Thiel, J.T. Yates Jr., W.H. Weinberg, Surf. Sci. 76, 296 (1978)
2. G.A. Somorjai, H.H. Farrell, Adv. Chem. Phys. 20, 215 (1971)
Chapter 33
Friction and Fracture
33.1 UHV Tribometer for Measuring the Coefficient

of Friction
Many technologies depend upon knowledge of the coefficient of friction between a

surface and a probe that slides across the surface. Since the coefficient of friction
μ = Fs/Fn, where Fs is the shear force, and Fn is the normal force, a device that
simultaneously measures both components of force during sliding is needed. In
addition, such a device should be able to be applied to surfaces that are initially
atomically clean, followed by studies on the surface which has a known coverage of
lubricant. These research demands have resulted in the development of the UHV
tribometer [1].
Figure 33.1a shows a schematic picture of the tribometer. A single crystal is
mounted on two support wires that connect to a rigid frame. Behind the crystal is a
coiled thermionic emitter used for heating the crystal to achieve a condition of
atomic cleanliness.
The frame surrounding the sample is connected by means of two L-shaped
Cu/Be alloy springs (0.006-in. thick by 0.25-in. wide) and each arm of the L is 0.5
in. in length. Glued to each of these springs are strain gauges (see p. 343) thermally
matched to the Cu/Be alloy. A total of eight strain gauges are employed, connected
to two Wheatstone bridges (normal; shear).
Measurements are made by bringing the sample in contact with the slider (also a
single crystal) which is mounted on a manipulator which can be translated at
constant speed, under an applied normal load. Figure 33.1b shows the displacement
components, δ, produced by applying a normal load to the sample, and it can be
seen that both normal displacements and shear displacements are sensed by the
strain gauges. However, for a normal force applied, the shear displacements
counterbalance, and are not measured by the Wheatstone bridges. In the same way,
for shear forces the normal displacement is counterbalanced and not measured by

DOI 10.1007/978-3-319-17668-0_33
342 33 Friction and Fracture
Fig. 33.1 UHV tribometer for measuring the coefficient of friction. a Strain gauge arrangement.
b Forces at slider/sample interface and displacement components
the bridges. This arrangement therefore permits the coefficient of friction to be

continuously measured during the sliding experiment.
It has been found on single crystal metal surfaces that a single monolayer of a
low molecular weight molecule is effective in reducing the coefficient of friction.
Three frictional regimes of increasing coverage of ethanol on an interface, a
S-covered Ni(100) crystal surface, have been observed. Below a monolayer,
sticking is observed; near a monolayer, stick-slip motion is observed; and above a
monolayer, slip motion is observed [1].
Another tribometer design [2] is shown on p. 343.
33.2 Pin-on-Disk Measurement—UHV 343
33.2 Pin-on-Disk Measurement—UHV
The measurement of the coefficient of friction (see p. 341) is of importance in

understanding lubrication. To do this without the influence of adsorption processes
from the atmosphere, tribometers operating in ultrahigh vacuum have been devised.
The device shown here has the ability to make friction coefficient measurements, as
well as measuring the electrical resistivity of the interface between the pin and the
disk. In addition, Auger spectroscopy measurements of the surface composition of
the wear track can be made, using a scanning instrument [2].
Figure 33.2a shows an exploded view of the tribometer. The disk sample is
mounted on a rotary platform that may be turned in UHV by an external motor.
A pin sample bent in the form of a loop maybe pressed onto the disk sample by
means of a spring tensioner that attaches to a cantilever beam. The cantilever beam
is free to move in the z-direction on bearings in a trunnion that can be moved in the
z-direction. Electrical contact is made to the pin for current and voltage measure-
ments. Two brushes (Cu/Be alloy) make electrical contact with a commutator
electrically connected to the sample; one of these is used for current, I, measure-
ments, and the second for voltage, V, measurements. The lateral force, F, is mea-
sured by two semiconductor strain gauges [3] that are accurately and squarely
mounted on a thinned section of the cantilever beam, using strain gauge cement [4].
The normal force, N, is adjusted by means of a screw adjuster on the spring
tensioner. Figure 33.2b shows the vector forces and the electrical measurement
method employed as the disk sample rotates. The entire pin-on-disk apparatus may
be moved inside the vacuum system by means of internal x-, y-, and z-bearings; and
rotary motion is provided by a rotary manipulator shaft.
The apparatus has been used to measure the effect of atmospheric gases on
contact electrical alloys under wear conditions [5, 6].
33.3 UHV Fracture Stage for Surface Analysis After

Fracture
The analysis of fracture surfaces is a significant feature of materials science, where

often the weakening of a solid can be related to the accumulation of impurities at an
internal interface. It is desirable to carry out the fracture and the surface analysis
without exposure to the atmosphere to avoid surface contamination of the fresh
fracture surface. Figure 33.3a shows an ultrahigh vacuum apparatus designed for this
purpose. A specimen to be fractured is clamped in the jaws of a fracture stage which
had been previously calibrated with a spring of known force constant. To measure
the load under tension, a load cell (Fig. 33.3b) was used on a lower flange as shown,
connected to the lower jaw by a stiff rod. This permits the force at the specimen to be
measured as a function of the displacement, which is automatically measured at the
stepper motor on top. To assure that the load cell does not experience a pressure
344 33 Friction and Fracture
Fig. 33.2 Pin-on-disk measurements—UHV. a Exploded view. b Forces and electrical contacts
differential due to atmospheric pressure on the outside, it is capped with a Conflat

flange which has a machined recess which is evacuated into the UHV system. Thus
both the upper and lower diaphragms are uninfluenced by gas pressure. The load cell
recorded force at 1 s intervals as the linear drive ran at a constant speed of 4 steps per
second (2.5 × 10−6 ms−1), giving force/displacement curves as shown in Fig. 33.3c.
It is noted in the comparison that atmosphere influences the force/displacement
curve. More force is required for fracture in the air, compared to fracture in UHV.
After fracture the freshly exposed surfaces are subjected to surface analysis by
vacuum transfer to a TOF-SIMS spectrometer [7, 8].
References 345
Fig. 33.3 UHV fracture stage for surface analysis after fracture. a UHV fracture device. b Load
cell for accurate operation in UHV. c Force/displacement plots for polyester resin
References
1. A.J. Gellman, J. Vac. Sci. Technol. A10, 180 (1992)

2. T.G. McDonald, D.E. Peebles, L.E. Pope, H.C. Peebles, Rev. Sci. Instrum. 58, 593 (1987)
3. P-doped U configuration, 350 Ω, 130 gauge factor, obtained from Kulite Semiconductor
Products, Inc., 1 Willow Tree Road, Leonia, NJ 07605
4. M-Bond 610 Epoxy Adhesive
5. L.E. Pope, D.E. Peebles, J. Vac. Sci. Technol. A4, 1659 (1986)
6. L.E. Pope, D.E. Peebles, Electrical Contacts-1986 (IEEE, New York, 1986), p. 65
7. A.R. Wood, N. Benedetto, N. Hooker, E. Scullion, P.A. Smith, J.F. Watts, Surf. Interface Anal.
40, 1409 (2008)
8. A.R. Wood, M.-L. Abel, P.A. Smith, J.F. Watts, J. Adhes. Sci. Technol. 23, 689 (2009)
Part IV
Thermal Control
Chapter 34
Heating Samples
34.1 Electronic Temperature Programmer—Metal

Crystals
The accurate and reproducible programming of the temperature of metal single

crystals is used for studies of temperature-programmed desorption (see Chap. 25, p.
273) as well as in many other applications. A review of the literature in this area to
1984 may be found in [1].
Figure 34.1a shows a block diagram of a temperature programmer design that
has been thoroughly tested in many applications [1]. An input signal from the
thermocouple, welded to a metal crystal, is fed into a thermocouple amplifier [2].
The input amplifier adds gain and offset to this signal, and delivers a 0 to +10 V
signal to the process controller. The thermocouple amplifier protects the single
crystal from being destroyed in the case of an open thermocouple circuit, which
drives the output of the amplifier to a high positive level. An electronic thermo-
couple reference junction eliminates the need for an external reference and is
routinely checked against a reference junction maintained at 77.3 or 273.2 K and
connected temporarily to the input for calibration.
The ramp generator is programmed by the laboratory worker, as shown sche-
matically in Fig. 34.1a. A more detailed block diagram of the ramp generator is
shown in Fig. 34.1b. A fixed frequency oscillator drives the divider circuit that
feeds a 12-bit binary counter. The 12-bit binary word is the input to a
digital-to-analog converter, which with scaling and offset, delivers a 1–5-V signal to
the process controller set point. The start point input presets the counter to a binary
count specified by the operator. The end point input determines the temperature at
which the ramp stops and holds constant temperature.
The power driver delivers up to 50 A with a burden voltage of 4 V. In addition,
the crystal may be isolated and a DC voltage may be applied. A commercial 5-V
DC supply was used, with a current regulator added to the output with appropriate

DOI 10.1007/978-3-319-17668-0_34
350 34 Heating Samples
Fig. 34.1 Electronic

temperature programmer—
metal crystals. a Block
diagram—temperature
programmer. b Block diagram
of ramp generator.
c Performance of programmer
34.1 Electronic Temperature Programmer—Metal Crystals 351
scaling and isolation to permit the process controller to control the power output.
An input of 1–5-V DC produces a 0–50-A DC output.
The process controller [3] (Fig. 34.1a), was selected as the primary control
device for the following reasons: (1) the ability to program the readouts in arbitrary
units; (2) accuracy (±0.05 %, full scale); (3) speed (updates 3 times/s); (4) digital
input of tuning parameters.
An illustration of the performance of the programmer is shown in Fig. 34.1c.
The programmer tuning constants were adjusted to produce the linear temperature
versus time performance from about 95–520 K as shown. The temperature stability
at 272 K is shown, where fluctuations of only ±0.35 K were observed. Scan rates
from about 2–8 K/s may be programmed into the controller.
In order to achieve optimum performance, appropriate values of the process
controller constants (gain, rate, and integral time) need to be established. These
three control constants are dependent on sample properties such as mass and
resistivity and on the sample-mounting configuration. Initially the rate and integral
time can be set to zero while an optimum gain is being established. An optimum
gain value is one which minimizes oscillations in the sample temperature, both
during the heating ramp and while the crystal is being maintained at constant
temperature, and the optimum gain may differ slightly for the two operating con-
ditions. It is also necessary to optimize the gain such that the difference between the
ramp voltage and the measured thermocouple voltage is minimized during the
heating ramp. The other control constants can then be adjusted to fine tune the
operation of the controller. Tuning of the integral time is essential, particularly
when trying to avoid an overshoot at the end of temperature ramp.
Another electronic controller, designed to correspond to the system’s thermal
response and capable of programming at 20 K/s without overshoot, has been
described [4].
34.2 Temperature Programming of Metal

and Semiconductor Crystals
Almost all of the studies of adsorption, desorption, and reactivity at surfaces involve
control of the temperature of the surface as well as control of temperature pro-
gramming. A traditional solution involves specialized controlling units, one of which
is mentioned on p. 349. These controllers provide a technically simple but rigid
solution for the temperature control. An alternative solution employs an acquisition
board in a personal computer controlled by a program written in any programming
language. This method provides great flexibility and can be used for any type of
temperature sensor and any desired temperature program. A graphical programming
language, LabVIEW [5], is especially appropriate for these tasks. It has a powerful
graphical interface, a large library of subroutines for data acquisition, and a variety of
mathematical operations. An example of a virtual instrument created by LabVIEW
for temperature measurement and control of a silicon crystal is described below.
Figure 34.2a shows a block diagram of the measurement setup for a silicon
crystal having a room temperature resistivity up to 100 Ω cm. The crystal tem-
perature is measured by a thermocouple that has a known relationship to the true
temperature (measured by pyrometer) as a function of temperature. A cold junction
compensator [6] is used to establish, electronically, a reference voltage corre-
sponding to the ice temperature, 273.2 K. The signal is fed to a separating amplifier
and amplified by 50–100 times, which minimizes the noise level received at the
acquisition board. A second purpose of the separating amplifier is the rejection of
the DC offset caused by the potential applied to the crystal for heating. The
acquisition board can only handle 10–20 V offsets, which would be normal for
metal crystals because of their higher conductivity compared to a semiconductor
crystal. The signal from the separating amplifier is fed to the acquisition board [7],
which has an analog input/output with a maximum input and output voltage of
10 V. Two DC power supplies are needed, one to power the separating amplifier
(±5 V) and a heating supply (60 W, 120 V, and a maximum current of 50 A).
Figure 34.2b shows a block diagram of the LabVIEW controlling program that
controls the heating power supply according to a predetermined program. A sample
of a complex heating/cooling program and the result is shown in Fig. 34.2c. This
program provides the desired heating ramp from −200 up to 850 °C, and then
cooling by various ramp functions, as shown on the front panel readout in
Fig. 34.2c. The lower right-hand side of the front panel readout shows the actual
program achieved compared to that desired.
Figure 34.2d shows a detailed view of the linearity of the heating ramp and the
error that was measured. Here, P = 3 K/s, the linear ramp function. The lower
difference plot shows the accuracy of this particular ramp and all of the data points
fell in a 1.5 K noise range up to about 1000 K, where the mechanical contact (see
p. 356) between the thermocouple and the Si crystal began to fail.
The ability to control the isothermal temperature of the crystal for long periods
of time is also excellent, with better than 1 K temperature constancy and error
spread. There is often an initial overshoot of the order of several K as one ramps up
to a desired temperature, and this can be minimized by adjusting the proportional
integrating differentiating (PID) parameters in the program, Fig. 34.2b.
The advantages of the LabVIEW approach to temperature programming are
listed below:
1. Flexibility—the temperature profile can be changed as desired and nonlinear
programs are possible. Any type of temperature sensor with a non-linear output
can be used (see p. 405). Any type of correction for errors, such as the error in a
thermocouple because of inefficient thermal contact (see p. 356), can be made.
2. High speed.
3. Universal interface—the acquisition board with multiple inputs and outputs cou-
pled to the LabVIEW program, can be used for computer control of other instru-
ments (e.g., a mass spectrometer). However, the interface cannot be used at the
same time with another acquisition board driven by a program working in DOS.
4. Convenient graphical programming with a large library of subroutines.
34.2 Temperature Programming of Metal and Semiconductor Crystals 353
Fig. 34.2 Temperature programming of metal and semiconductor crystals. a Block diagram—
resistive heating control. b Block diagram—LabVIEW controller logic. c Control parameters for
temperature program. d Linearity and error in heating ramp
34.3 Mounting Nonweldable Crystals for Resistive

Heating and Cooling
Crystals such as Ag, Al, Au, and Cu possess very high thermal conductivity and
cannot easily be spot-welded to other metals because of the lack of melting at the
low temperatures achieved at the junction during welding. The mounting of such
crystals to heating/cooling wires requires a special method that will provide good
mechanical, thermal, and electrical contact over many cycles of heating and cool-
ing, and a design for this is shown in Fig. 34.3.
The crystal holder is fabricated out of OFHC copper, sapphire, and tungsten.
These materials all possess optimum thermal conductivity properties below 100 K.
Fig. 34.3 Mounting nonweldable crystals for resistive heating and cooling
34.3 Mounting Nonweldable Crystals for Resistive Heating and Cooling 355
Figure 34.3 shows a scheme to mount such crystals. A slot is carefully cut into
two opposite edges of the crystal. The slot makes a tight sliding fit for the tungsten
wire heating lead, and the crystal should not be distorted to make this fit.
The tungsten heating/cooling leads are fabricated with a pair of U-loops that
remove the effects of linear expansion that otherwise would push the wire from the
slot. Thus, the crystal, incorrectly mounted as shown in the right-hand insert, is
subject to variation in the cooling and electrical contact as the wire is heated and
cooled owing to bowing of the wire in the slot. The looped wires on the other hand
undergo thermal expansion effects which are almost counterbalanced on either arm
of the U-loops.
Make the assembly as follows using gloved fingers:
1. Mount the preshaped W wire (0.38-mm diameter) on the lower set of Ta screws
and tighten.
2. Fix the preshaped upper W wire on the right hand screw and do not tighten the
screw.
3. Fix the crystal with its thermocouple (hole-mounted—details not shown) on the
lower wire.
4. Rotate the upper wire counter clockwise until it fits into the slot and then tighten
both screws carefully.
This mounting procedure produces a stable junction over many cycles that
permits cooling of the crystal to about 32 K when liquid He is used in the reservoir
intended for liquid nitrogen [8].
34.4 Strain-Free Single Crystal Mounting Azimuthal

Motion
In certain applications, single crystals must be mounted in such a manner that

during heating and cooling they are free from strain owing to expansion and
contraction effects in the metal to which they are welded. This is particularly true in
atom scattering experiments, where tilting of the crystal can cause significant
problems.
The design shown in Fig. 34.4 eliminates this problem by allowing the con-
duction leads welded to the crystal to freely expand and contract [9]. This is
facilitated by having one end of each lead free to slide. The square cross-section Ir
rods (0.7 × 0.7 mm2) are welded to Ta pads on the right-hand side, but are free to
slide inside of sapphire bearing supports on the other end. The sapphire support
bearings are held in place by Ta foils welded to the Ta support plates. Ta conductor
leads (about 30-A capacity) are connected in pairs to the Ta support plates on the
left, and current passes through these and the support plates to the Ir rods and thence
through the crystal.
Fig. 34.4 Strain free single crystal mounting—azimuthal motion
With a Ru crystal, temperatures above 1600 K may be reached, and cooling to

95 or 50 K was achieved using liquid nitrogen or liquid helium, respectively. The
Ru crystal was spot welded to the square Ir rods through a Ta sandwich foil.
34.5 Heating Silicon Crystals and Temperature

Measurements
One of the most difficult tasks in the laboratory is uniformly heating silicon single
crystals by resistive heating and measuring the temperature of the crystal. A variety
of methods of making electrical contact to silicon crystals have been devised that
34.5 Heating Silicon Crystals and Temperature Measurements 357
involve attachment to metal springs [10, 11] or elastic metal sheets [12, 13]. These
contact methods often result in the formation of hot spots at certain regions of contact
and can lead to nonuniform heating. Ta and Mo are used in these applications, since
other metals can lead to metal contamination of the silicon crystal, either through
metal diffusion through the bulk (Ni; Fe) or by transport by vaporization.
The attachment of thermocouples is also a difficult task owing to metal diffusion
or metal vaporization. Here, too, Ta-protective envelopes must be used, as in the
design of Bozack et al. [13], which employed Ta-protected W26 %Re/W5 %Re
thermocouples inserted into slots cut into the crystal edge [13]. We have found that
chromel-alumel (type K) thermocouples are unsatisfactory owing to the high vapor
pressure of Ni (10−5–10−4 Torr at 1450–1500 K) [14]. A few atomic percent of Ni
produces, on Si(100), the well-known split-off dimer defects [15]. Thus, W26 %
Re/W5 %Re thermocouples, shielded by Ta, have been used successfully in our
own work.
Adhesive thermocouple contacts, such as zirconia-based Ultratem (Aremco),
have been found to diffuse through Si above 1100 K. A Ta thermocouple tab
contact, joined by a layer of Ge that alloys with both Ta and Si was proposed [11],
but adhesion was lost on heating to high temperatures [16].
Figure 34.5a shows a design [17] that addresses the clamping of Si crystals to
provide rather uniform electrical contact and temperature distribution. A Ta sheet
(20 × 9 × 0.125 mm3) is bent as shown in the diagram to form a bracket shape,
which is then welded to a W rod (2-mm dia.), using a copper plate as an alignment
jig, producing the structure shown in Fig. 34.5a(d). Four slips of Si wafer material
are prepared (4.5 × 8.5 × 0.4 mm3) and used to sandwich a Si wafer slice
(23 × 8 × 0.4 mm3) as shown in Fig. 34.5a(e). The entire sandwich structure is
compressed by a flat U-shaped Ta spring that is wedged into the slot between the
top Si slip and the top of the bracket, Fig. 34.5a(f). The fully assembled mounting is
shown in Fig. 34.5a(g).
Figure 34.5b shows a method for attaching a W26 %Re/W5 %Re thermocouple
(0.08-mm dia.) to the crystal by a U-shaped clip design [17]. To prevent direct
contact of the ther-mocouple material to the crystal, it is shielded with a Ta tube
(0.19-mm inner dia., 0.06-mm wall thickness, 5-mm length). The prewelded ther-
mocouple is inserted into the Ta tube at the end, which has been previously cut off
at an angle [Fig. 34.5b(a, b)], and the tube is flattened [Fig. 34.5b(b)], and formed
into a bracket shape that tightly clips over the crystal edge [Fig. 34.5b(c)]. The
thermocouple exits from the angled end and is bent upward as shown in Fig. 34.5b
(c). The thermocouple voltage is converted to temperature using a calibration
method similar to that of Smentkowski et al. [18], as discussed on p. 405. A cold
junction compensator is used for the reference junction at 273 K.
Two designs for thermocouple attachment to Si crystals have been compared
using a calibrated optical pyrometer. Figure 34.5c shows the two designs where the
slotted crystal design from Bozack et al. [13] and the design shown here [17] are
compared. The deviation for the slotted design is from 50 to 150 K, whereas the
deviation of the U-shaped clip design is from 50 to 90 K over the same temperature
range.
Fig. 34.5 Heating silicon

crystals and temperature
measurements. a Mounting
design—Si sample.
b Thermocouple attachment.
c Temperature deviations—
two thermocouple attachment
methods
34.5 Heating Silicon Crystals and Temperature Measurements 359
The uniformity of heating of the Si crystal with this mounting design is found to
be ±15 K at 1400 K, as measured by an optical pyrometer. At temperatures above
about 1100 K, temperature oscillations are sometimes observed with the U-shaped
clip thermocouple.
Great care must be exercised in making the mount and the thermocouple clip. All
Ta parts must be boiled in concentrated HCl solution to remove Ni contaminants
[15]. All tools should be made of plastic to avoid metal contamination, and fabri-
cation should be done under clean conditions without contact to metals other than
cleaned Ta. In addition, Mo parts, which are more springy, may be substituted for Ta
parts, although the HCl cleaning procedure followed by STM studies has not been
done for crystals with Mo/Si contacts, as has been done for W/Ta contacts [15].
The use of interferometry for measuring the temperature changes in semicon-
ductor crystals has been reported [19]. Here, both temperature dependent changes in
the refractive index and in the thickness of the crystal wafer cause the optical path
in the wafer to change in comparison to λ/2n, where λ = the laser radiation
wavelength, and n = the refractive index. Temperature changes as small as 0.2 K
have been detected, but one must start at some known temperature in order to make
temperature measurements at other temperatures.
34.6 Avoiding Extraneous Electrical Effects Owing

to Heating Crystals
Single crystals of metals may be heated in several ways for studies at constant
temperatures. Unfortunately the presence of large AC or DC currents used for
heating and the concomitant presence of stray electric and magnetic fields will be a
severe disturbance to measurements involving electron currents to or from the
crystal. A general method used to avoid this problem involves pulsing the heating
current and making current measurements from the crystal during the time when the
heating pulse is off.
Figure 34.6a shows a typical mounting for a metal single crystal on a manipu-
lator using ohmic heating for the crystal. Large-diameter (1.0–1.5 mm) tungsten
support leads are connected to current and cooling leads on the manipulator and
conduct current to a pair of tungsten wires (0.3-mm diameter) welded to the back of
the crystal. Tungsten is preferred to tantalum because of its higher coefficient of
thermal conductivity at low temperatures. In this configuration the crystal is heated
by thermal conduction from the hot wires rather than by ohmic heating of the
crystal itself, and at incandescence, it will be seen that the parallel wires are
comparatively bright in the region between the crystal and the support leads. The
uniformity of welding to the crystal and the support leads can be judged by the
relative brightness of the four intermediate regions. Poor welds will yield higher
brightness and will possibly cause premature burnout. The thermocouple welded to
the crystal edge is used to measure the crystal temperature.
Fig. 34.6 Avoiding extraneous electrical effects owing to heating crystals. a. Interrupted ohmic
heating. b Interrupted electron bombardment heating
Figure 34.6b shows a typical arrangement for heating by electron bombardment.

The tungsten coil emitter located about 3–5 mm behind the crystal is kept at ground
potential, and the crystal is biased positively (several hundred volts to 2 kV) (see
p. 361). Thermionically emitted electrons are used for heating the crystal and any
crystal can easily be melted accidentally with this method. In addition to electron
bombardment heating, it is also possible to use radiation heating to achieve crystal
temperatures approaching 1000 K. The upper temperature limits for radiation
34.6 Avoiding Extraneous Electrical Effects Owing to Heating Crystals 361
heating depend upon the magnitude of the heat leakage to the support wires through
the thinner wires welded to the crystal.
Three circuits for interrupted heating of single crystals have been reported in the
literature [20–22]. For ohmic heating, at 1000 K, the temperature controller [20]
required a duty cycle of 2/3. The switching rate was approximately 0.3 Hz, and the
temperature regulation was ±15 K. Better control would be possible at higher
switching rates [20]. For electron bombardment heating, a second circuit has been
published in which a mechanical relay is used [21]. Here, bounce within the relay
limits the time period for data acquisition to about 40 ms at 10 Hz. It is necessary to
switch both the filament heater current and the bias voltage and since the power
supplies are slow to recover from the switching, dummy loads are used in the off
cycle. A difficulty reported for switching large currents and high voltages is that
transients are produced that disturb electronic measuring equipment being used.
Proper shielding of the elements of the circuit and the use of shielded cables
eliminates these effects. A third circuit, which has shown favorable utilization in
conjunction with LEED, Auger, and HREELS measurements is reported in [22].
The circuits shown in [20, 22] employ older electronic components, and the user
should consult with an electronic specialist for designing similar circuits using
modern components.
34.7 Electron Bombardment Crystal Heating
Electron bombardment heating of a metal crystal can often be used to bring the
crystal to higher temperature than can be done conveniently by ohmic heating.
Thus, for the highly refractory metals such as W, Ta, Mo, and Re, all of which
require very high temperatures to remove impurities, electron bombardment is often
used (see also p. 363.)
A typical design for electron bombardment heating, adapted from [23], is shown
in Fig. 34.7. The crystal is supported on a manipulator by one or more welded wires
that connect to a stub on the end of the manipulator. This stub is electrically isolated
from the manipulator shaft by a high-voltage ceramic spacer that will work at
positive bias potentials of several kilovolts. The high-voltage connection to the stub
is provided by a well-insulated wire that passes through a ceramic sheath, and this
sheathed wire is firmly attached to the manipulator shaft by means of a ceramic
support designed to hold it tightly, and in isolation from the rotary manipulator
shaft or any other grounded parts.
The very simple electron bombardment gun is made of a stainless steel tube with
an aperture cap placed over the open end, and a custom made ceramic base to hold
the heavy filament leads. Filament current is supplied through copper braid that is
connected to the ceramic support, and thence to flexible power leads that connect to
feedthroughs on the manipulator flange. The electron gun is mechanically con-
nected to the ceramic support by a rod, as shown.
Fig. 34.7 Electron bombardment crystal heating
The filament for the electron gun can be made of pure W wire, or thoriated W
wire (see p. 133). Emission currents of 10–50 mA with a positive bias of 1 kV on
the crystal will supply 10–50 W, usually enough for heating to the desired tem-
perature, depending on the thermal conductivity of the welded leads to the crystal
and the radiation loss at the desired crystal temperature. For gentle heating, only
radiation heating from the filament can be used, and in this case a non-thoriated W
filament would be preferable. While the winding of W filaments is easily done,
34.7 Electron Bombardment Crystal Heating 363
professionally wound filaments of uniform shape may be purchased [24]. A very

compact and tightly wound W filament can be obtained from a slide projector
quartz-halogen lamp, and the quartz envelope can be cut off by a glassblower,
leaving an accurately mounted W filament attached to the quartz base. The filament
should not be removed from the lamp base as it will be difficult to reweld it to the
leads without strain and subsequent distortion upon heating.
It is emphasized that the coiled W filament (typically wire diameter = 0.2 or
0.3 mm) should be welded to its supporting leads under conditions where there is
no strain left in the coil after the second weld. Strain will result in extreme filament
distortion upon heating to incandescence.
The temperature of the crystal can be measured with an optical pyrometer for
adjustment (see p. 409). However, usually a thermocouple is employed for this, and
one must remember that the thermocouple will also be biased at kilovolts when the
crystal bias is on. Since these voltages are lethal, proper care must be taken in the
insulation of the thermocouple leads, and in the isolation of the readout device
connected to these leads.
One might consider reversing the bias and making the electron gun operate at a
negative bias of a kilovolt or so, and grounding the crystal. This is not advised,
since then electron bombardment will occur broadly to all regions at ground
potential in the vicinity of the crystal, resulting in excessive outgassing and damage
to devices and screens.
Even simplier electron guns may be constructed, the simplest being just an
emitter filament supported a few mm from the back of the crystal.
34.8 Indirect Sample Heating
The use of indirect heating of samples is especially important for materials that are
nonconductive and therefore cannot be heated by ohmic heating. However, indirect
heating of conductor materials is also a favored method of heating, since excellent
control of temperature can be achieved by indirect heating methods. Both heating
by radiation and heating by electron bombardment are termed indirect heating,
since neither method depends on the electrical resistance of the sample, as does
direct ohmic heating.
Figure 34.8a shows a typical method for indirectly heating a sample using a
button heater [25] that is supplied commercially [26]. The sample is held onto the
top surface of the Mo cylindrical shell that surrounds the W heater wire. The W
heater is immersed (potted) in an aluminum oxide powder. In the case shown here,
the sample is held in place by means of a Mo sleeve that fits snugly over the Mo
cylinder and has a lip that grips the front surface of the sample. This arrangement
may be used up to about 1500 K without damage to the heater assembly. Since the
heater wire is noninductively wound, electron spectroscopic measurements may be
made with heating power on [25].
Fig. 34.8 Indirect sample heating. a Conduction heating (T ≤ 1500 K). b Radiation and
e-bombardment heating (T ≥ 1500 K). c Static heating—config. B d Programmed heating and
cooling—config. B. e Ohmic support heating
Figure 34.8b shows another arrangement that maybe used at temperatures above
1500 K, but which also is useful at lower temperatures [27]. A Mo cylindrical can is
polished flat on its top surface, and the sample is placed in contact with the upper
surface. In the case shown here, a single crystal of diamond is held by two tungsten
or tungsten alloy wires by means of parallel slots in two opposing edges. The wires
are spot-welded to the edges of the can and tightly bind the diamond to the polished
top of the can. Inside the can is a flat W pancake spiral made by winding a W wire
onto a mandrel into a watch spring configuration as a tight coil. No aluminum oxide
potting powder is present, so that the Mo can is heated only by radiation from the W
spiral, or by radiation plus electron bombardment if more power is needed.
A reflective radiation shield behind the Mo can assists in increasing the efficiency of
heating the Mo support by the incandescent filament. A more complex electron
bombardment heating device involving a shielded thermionic emitter has been
reported [28].
The performance of the heating assembly in Fig. 34.8b has been measured for a
diamond single crystal [27], and the results are shown in Fig. 34.8c, d. In Fig. 34.8c
the temperatures of two thermocouples, buried in laser-drilled holes in the edge of
34.8 Indirect Sample Heating 365
the diamond crystal, are shown for static heating experiments, where the assembly
is brought to steady temperature in ultrahigh vacuum. The two thermocouples differ
by only 8 K in these experiments, indicating that the true temperature of the
diamond is being measured. However, at a diamond temperature of 1400 K, the
difference between the Mo support temperature and the diamond temperature
amounts to 140 K. This difference decreases at lower temperatures. Since diamond
is an excellent thermal conductor, it is likely that heat transfer from the Mo surface
to the diamond back surface is the limiting factor in these experiments. Similar
temperature differences between the temperature of a heated supporting surface and
a contacting sample have been measured by others at about 10−3 Torr pressure. The
sample was contacted to the supporting surface using a silver paint that worked up
to 1073 K, where a 22 K temperature difference was measured [29].
Figure 34.8d shows the difference between the heating behavior of the diamond
crystal and the Mo support can under conditions of temperature programming at a
rate of 1.3 K/s. At a diamond temperature of 1150 K, a difference of temperature
between the Mo support and the diamond of 225 K was noted (compared to only
90 K under static conditions as shown in Fig. 34.8c). The Mo can cools faster than
the diamond when the heating is interrupted, as would be expected for cooling
controlled by thermal conduction from the assembly.
The maximum filament heating power used to heat the diamond may be reduced
by employing a mixture of radiation heating combined with electron bombardment
heating. Thus, for example, a positive bias of 250 V produces thermionic heating of
12 W at a diamond temperature of 1273 K, whereas 47 W of filament heating power
was needed under conditions of zero bias to achieve the same temperature.
A successful variation of this method, involving ohmic heating of the support
assembly [29] is shown in Fig. 34.8e. Here, a polished Ta block of the same size as
a TiO2 single crystal is heated by means of W wires which pass through the body of
the Ta. The temperature of the crystal is measured with an embedded thermocouple
cemented [30–32] into a diagonal slot in the crystal corner, and programming of the
heating power is controlled by this thermocouple which acts to supply a feedback
signal. Somewhat similar methods for heating and cooling alkali halide crystals,
and using an In metal film as a heat transfer medium have been reported [33].
34.9 Indirect Heating of Compound Semiconductors
A variety of indirect heating methods for semiconductors which cannot be heated

by electrical conduction have been discussed on p. 363. GaAs is a special material
because of its fragile mechanical properties and because many samples are of the
incorrect conductivity for direct resistive heating (see p. 327). The method shown
here is especially good for this type of material [34].
As shown in Fig. 34.9 the GaAs crystal is bonded to a sapphire plate using an
indium film bond. The sapphire plate is coated on its back with a thin Ta film that is
Fig. 34.9 GaAs—mounting for indirect heating
used for resistive heating. A thermocouple is attached to the Ta film for the mea-
surement of the temperature.
The indium film is applied conservatively so as to not cause a short circuit
between the Ta clip leads that support the sapphire plate on 0.75-mm Ta support
leads. The indium film application may be easily done by using the Ta film resistor
to raise the sapphire temperature above the melting point of In (157 °C). The Ta
resistor film requires 1.5–2 A at 8–10 V to heat the unit to about 900 K. The GaAs
sample is grounded by using a Ta spring contact that is welded to a Ta clip.
The disadvantages of this technique are that the In reacts slowly with GaAs,
causing failure after long heating. This effect can be minimized by using only small
amount of In. In addition, at high temperatures, the In has a significant vapor
pressure (10−7 Torr at 833 K; 10−6 Torr at 895 K).
34.10 Heating Design—Insulator Crystals
A mounting design for electrically nonconducting crystals must provide a method

for uniformly heating the crystal. When ohmic heating is to be used, various
methods of placing the crystal on a heatable metal substrate have been described in
other sections.
A method for making intimate contact between an insulator crystal and a metal
[35] is shown in Fig. 34.10. The method involves first making a sputtered film of
the metal on the back of the crystal, as shown in Fig. 34.10a. In the case of a MgO
34.10 Heating Design—Insulator Crystals 367
Fig. 34.10 Heating design—insulator crystals
crystal described here, about 1 μm of Ni was sputter deposited in a period of

30 min. This film provides an intimate contact to the back of the crystal while
permitting the electrodeposition of a thicker metal film. Figure 34.10b shows the
crystal with the sputter-deposited Ni film and with a Ni heating wire lying across
the back. This assembly is placed into a vessel (Fig. 34.10c), and a thick coating of
Ni is electrodeposited, making good contact with the underlying sputter-deposited
Ni film as well as holding the heating wire in place. Between steps A and B, the
cement used is removed and the crystal is placed on a new glass slide (No. 2) to
remove the extraneous Ni film deposited on slide No. 1, which would also be
electroplated.
The cement used in this procedure to mask the crystal edges is a commercial
UHV leak sealer, Vacseal [36], and is acetone soluble. It is important to electroplate
the thick Ni film without stress, and special procedures for doing this must be
followed. A nickel sulfamate plating solution (1 L volume) was employed con-

taining nickel sulfamate (600 g/L), nickel bromide (45 g/L), boric acid (2 g/L),
saccharin (0.5 g/L), and a surfactant X-100 (2 drops). The bath temperature is 40 °C,
and its pH is 4.0 ± 0.1, which can be adjusted using sodium hydroxide solution. The
approximate current density is 3.4 A/dm2. The use of six to eight electrolytic grade
Ni anodes [37] is important, only one of which is shown in Fig. 34.10c. Before use
the electrolytic bath should be treated with 30 g of activated charcoal, filtered, and
then plated onto a dummy anode for 2 h. Deposition rates on the back of the insulator
crystal are typically 50 μm/h. The bath should be stirred with a magnetic stirring bar
and final attachment of the Ni heating wire takes place as shown in Fig. 34.10d. To
check on the stress in the Ni film being deposited, a thin strip of Cu
(1 × 10 × 0.05 cm3) is used to receive the Ni film as a test. After plating for 30 min,
the test strip should bend 1.5–2 mm toward the copper backside after the protective
coating of lacquer is removed, indicating that the Ni film adheres well and is
under slight compressive stress. The thermal expansion coefficient of the Ni,
13.3 × 10−6 K−1, should match that of the crystal, as in the case of MgO
(13.5 × 10−6 K−1). For a MgO crystal the Ni film could be cycled in vacuum many
times to 1073 K without failure.
34.11 Cement Mounting—Semiconductor Crystals
A number of methods for mounting semiconductor crystals in ways where they can
be electrically heated and cooled are discussed throughout this chapter beginning on
p. 349. In the case of crystals which cannot be heated by electrical conduction,
either because of unfavorable conductivity or because of unfavorable shape, it is
possible to use a refractory cement that holds the crystal to a tungsten disk that is
heated through welded contacts to refractory metal heating wires [38, 39].
In Fig. 34.11, an irregularly shaped SiC crystal has been cemented to a W disk
using Aremco Ultratem 516 cement, which has a ZrO2 base [40]. The W disk has
previously been welded to a pair of Ta heating wires that pass current from the
heavy W (1.5-mm-diameter) support leads. The cement is made electrically con-
ductive by mixing with the appropriate amount of powdered graphite. This permits
the electron spectroscopies to be used without sample charging.
A chromel-alumel thermocouple is connected to the back side of the SiC crystal
by means of an alloy solder (90 % Au/10 % Ta) that anchors a tiny Ta tab to the
crystal. The connection to the SiC crystal must be done in a preparative vacuum
chamber to avoid gross oxidation of the crystal during induction heating.
Measurements with AES and XPS indicate that no migration of the ceramic
cement occurs to the front of the SiC crystal up to 1400 K. The cement is adequate
for mounting light crystals, but it is by no means robust.
At 1000 K, the temperature indicated by the thermocouple may be about 100 K
too high compared to the true SiC temperature. The temperature range from 85 to
1400 K was employed, and minimal outgassing occurs at 1400 K [38].
34.12 Low-Stress Mount for Fragile Semiconductor Crystals 369
Fig. 34.11 Cement mounting—semiconductor crystals
34.12 Low-Stress Mount for Fragile Semiconductor

Crystals
A system for mounting thin semiconductor crystals in an ultrahigh vacuum system

for heating and cooling is described [41]. The method avoids stresses on the heated
crystals, allowing them to remain flat at high temperatures. The method also allows
a thermocouple to be mounted between two semiconductor samples, as shown in
Fig. 34.12.
Two L-shaped copper electrical conductor bars are attached to coolable electrical
feedthroughs that are in contact with a refrigerated Dewar tube. The Cu conductors
are separated and aligned by a Macor spacer. On the front of the L-shaped con-
ductor bars are tapped holes accommodating 0–80 stainless steel machine screws
that have had their threads removed under their heads to avoid snagging on the
semiconductor wafer samples. Chromel wire springs (0.016-in.-diameter wire,
formed around an 0.0625-in.-diameter rod with 38 turns/in.) provide compression
Fig. 34.12 Thin semiconductor crystal mounting—heating and cooling
force in the sandwich structure to be fabricated. A sandwich structure, consisting of

two identical semiconductor samples, is placed between Ta clips (which have
elongated holes allowing expansion to occur). Thermocouple wires are welded to
the 0.005-in.-thick Ta foil that is sandwiched between the semiconductor samples.
A hole drilled in the back semiconductor sample allows the thermocouple to extend
back to its terminals on the manipulator feedthrough. The entire sandwich structure
is compressed by turning and locking the nuts on the stainless steel machine screws.
For two 0.014-in.-thick Ge samples, the compression in the chromel springs is
typically 0.1 in.
34.12 Low-Stress Mount for Fragile Semiconductor Crystals 371
This arrangement heats samples of Ge up to their melting point (1211 K) with

virtually no temperature gradient across the samples. The thermocouple readings
are within 2 K of pyrometer readings. By cooling the Dewar with liquid nitrogen,
temperatures near 200 K are reached. The chromel springs lose their ability to
compress the sample when long annealing near 1273 K is carried out.
The reader is warned that the stainless steel and type K thermocouple material
used in this configuration is likely not to work for Si, which must be heated to much
higher temperatures for cleaning. Both stainless steel and type K alloys will transfer
Ni impurities to Si, giving unsatisfactory modifications of the Si surface structure
and its surface properties (see p. 548).
34.13 Radiation Heating of Crystals Through Glass

Windows
Single crystals may be heated by radiation from hot filaments mounted internally in
the ultrahigh vacuum system, and this method possesses the best efficiency of
coupling the radiant source (usually a hot W wire) to the crystal. For special cases,
however, it may be preferable to employ radiation heating from external tungsten
halogen lamp sources, which typically have maximum power limits in the 250–
1000-W range. Usually the maximum temperature that can be reached with an
external lamp is insufficient for cleaning the crystal and electron bombardment or
ohmic heating methods must be employed for initial cleaning.
A very early use of external radiation heating for programming the temperature
of semiconductor crystals can be found in the work of Palmberg and Peria [42]. In
Fig. 34.13a, their glass ultrahigh vacuum apparatus is shown. A track allows the
translation of the crystal, using a magnetic device, from location to location in the
linear vacuum system. Shown is a 200-W halogen lamp inside a spherical reflector
(silvered glass flask section). Figure 34.13b shows a variation on the design of
Fig. 34.13a, where a commercial 750-W lamp (operating at reduced power) is
housed in an ellipsoidal reflector, which has reflective aluminum focusing surfaces
behind and in front of the lamp [43]. This arrangement permitted crystal temper-
atures near 1000 K to be achieved using the lamp alone, and cooling of the
hemispherical glass surface was necessary to prevent damage to it by the
lamp. A nonlinear heating rate was observed. The crystal was mounted on a
0.25-mm-diameter W support wire, which minimized cooling by conduction to the
liquid-nitrogen-cooled support lead. A tungsten thermionic emitter is shown behind
the crystal and this is used for electron bombardment cleaning at 2500 K, assisted
by magnetic collimation. Figure 34.13c shows a lens arrangement for focusing light
from a 900-W lamp [44]. The single-glass double-convex lens, f/1.32, has a
magnification of unity. Lenses of this type are used as condenser lenses in pho-
tographic enlargers. The window is Pyrex glass with 95 % transmission in the near
infrared, where the spectral radiancy is greatest for tungsten. A silicon crystal could
Fig. 34.13 Radiation heating of crystals through glass windows. a Hemispherical reflector.
b Ellipsoidal reflector. c Lens + hemispherical reflector. d Temperature program for C
be raised to 900 K with this arrangement. The temperature program using the lamp
in Fig. 34.13c is shown in Fig. 34.13d, and the reproducibility over a period of
several days is within ±5 K.
34.14 Pyrolytic Graphite Heating Element for UHV
Heavy power deposition in an ultrahigh vacuum system can be achieved using

electron bombardment, with the inherent possibility of burning out the emitter
filament. The resistive heater described below avoids this problem.
The heater shown in Fig. 34.14a was constructed of chemical-vapor-deposited
(CVD) pyrolytic graphite, which is highly anisotropic in its electrical conductivity
[45]. The heater is fabricated from a slab of the graphite so that the current flow is in
the C direction, which has a resistivity about 103 times higher than in the A or B
directions. The width of the device in the C direction is governed by the thickness
of the pyrolytic graphite supplied by the manufacturer [46], which is about 25 mm.
The slab is cleaved along the A/B direction into slices 4–5 mm in thickness.
34.14 Pyrolytic Graphite Heating Element for UHV 373
Fig. 34.14 Pyrolytic graphite heating elements. a Heater design—current flow in C direction.
b Design for internal heating element—metal. c Resistivity in C direction
Resistive sections are separated from each other using a band saw or preferably a
diamond saw wheel. Care must be taken not to fracture the graphite during sawing.
The heater is mounted on threaded Mo connection rods, as shown in Fig. 34.14a.
These rods are not heated too much since the thermal conductivity along the
C-direction is small compared to the other directions. When the temperature in the
central region is in the 2000 K range ±30 K, typical power dissipation is about
2.0 kW for 155 V applied with 13.5-A current.
Figure 34.14b shows the heater used as in insertion device inside of a block to be
heated to about 1200 K. The pyrolytic graphite heater was mounted in a rectangular
hole and was insulated from the metal by an alumina spacer. Electrical contact was
made with a tungsten foil spring attached to the bottom of the heater as shown.
Although the heater is mechanically fragile, once in operation it may be
repeatedly cycled in temperature without damage. The authors report that this
design is compatible with ultrahigh vacuum.
The resistivity of the pyrolytic graphite heater was measured versus temperature,
as shown in Fig. 34.14c. Relatively poor agreement was obtained compared to the
manufacturer’s specifications. This level of variation is typical for CVD pyrolytic
graphite measured in the C-direction.
34.15 Heater—Paint-on Type
A simple resistance heater which can be painted on an insulating surface [47] may
be useful in indirectly heating components in an ultrahigh vacuum system (as
shown in Fig. 34.15). A thin layer of Aquadag [48], which is a colloidal graphite
suspension, is painted onto a roughened quartz surface and outgassed for 0.5 h by
baking in an oven at 300 °C, which is reported to be satisfactory for complete
outgassing [48]. The graphite layer can then be smoothed with fine abrasive paper.
Contact to the graphite layer is made using Ni or Pt strips, 0.1 mm thick, which are
sandwiched between the graphite layer and a second quartz plate that is clamped to
the first to assure good contact.
For a 27-mm-diameter plate, typical room temperature resistances are about
15 Ω. With a variable transformer, up to 20-V maximum, the heater could be used
continuously for long periods at temperatures of 500 °C, and operation at higher
temperatures may be possible. The application of these ideas to heating irregularly
shaped surfaces is possible.
Fig. 34.15 Heater—paint-on type

34.16 Mounting High-Temperature Tungsten Filaments 375
Fig. 34.16 Mounting of high-temperature tungsten filaments
34.16 Mounting High-Temperature Tungsten Filaments
The proper mounting of tungsten filaments so that operation at high temperatures

does not melt or vaporize the filament mount is sometimes a problem in instrument
design. A useful method of doing this without using complex screwing or clamping
arrangements is shown in Fig. 34.16. A Mo rod, threaded to fit into an electrical
mount, is drilled with an axial hole somewhat larger in diameter than the tungsten
wire. The drilling process should be done slowly (1000 rpm) to avoid splitting the
Mo rod [49]. The ends of the tungsten wire are then coiled by winding around a
mandrel, followed by stretching, such that the helix formed makes a tight sliding fit
into the hole in the Mo rod.
34.17 Rotatable Single Crystal
A cylindrical single crystal exposes different single crystal planes at different angles
around its perimeter. Thus experiments which sample small areas may be carried
out at different azimuthal angles, sampling different single crystal facets as well as
at different atomic step geometries and densities on vicinal planes. The crystal
geometry shown in Fig. 34.17a features a crystal shell with inside diameter of
14 mm and outside diameter of 20 mm, and involves a crystal length of 14 mm. The
crystal is clamped to the bottom of a cryogenic reservoir containing liquid N2,
which with long cooling times can bring the crystal down to 89 K. The crystal may
be heated for cleaning to above 900 K by radiation from the encased tungsten
filament or by electron bombardment. Cooling occurs at an initial rate of
*45 K min−1, with the lowest temperature being achieved in about 60 min cooling
Fig. 34.17 Rotatable single crystal. a Structure of cylindrical single crystal. b H2 dissociation
probability versus atomic step density
34.17 Rotatable Single Crystal 377
Fig. 34.18 Radiation-heated evaporator
time. The crystal may be programmed upwards in temperature at 1 K s−1 and

thermal desorption is measured by an apertured and differentially-pumped mass
spectrometer through a skimmer which is 6 mm wide and 1 mm high. Desorption
spectra are obtained with an azimuthal angular resolution of 2°. The crystal is dosed
by a molecular beam which produces a stripe on the crystal which is 6 mm wide and
0.3 mm high [50].
An example of the use of this approach involves measurements of the sticking
coefficient of H2 on the crystal at various positions involving different atomic step
densities. Figure 34.17b shows this measurement, plotted versus the inverse dis-
tance between atomic steps which depends on the azimuthal angle setting of the
crystal compared to the axis of the H2 molecular beam. The sticking coefficient rises
as the step density increases [50].
34.18 Radiation-Heated Evaporator
Tungsten lamps are excellent clean radiation sources for the interior of UHV sys-
tems. They may be of the quartz-halogen or quartz-Xenon type where the added gas
prevents W evaporation on the inner surface of the quartz envelope. UHV com-
patibility is related to having only quartz and metal exposed to the vacuum system.
Such bulbs are used in home lighting systems, auto headlamps, and projector lamps.
They have the advantage of being able to operate at high power without filament
burnout or W evaporation into the vacuum system. Also, there is no thermionic
electron emission. Gases in the bulb will not escape into the vacuum system. These
bulbs have been used also for passing focused radiation through windows for
heating of samples in UHV (see Sect. 34.13).
A particular use of a lamp for heating an evaporant is shown in Fig. 34.18. An Al
or Cu body, containing a glass vial with the evaporant, is heated from behind with a
lamp encased in the body as shown. The lamp is carefully shielded by aluminum
metal shields to prevent radiation heating of an insulating PTFE support or other
parts of the support assembly. Thermocouple wires may be attached to the evap-
orator, but electrical pin contacts to thermocouple wires should be shielded from
radiation to prevent heating of junctions which will supply spurious voltages. The
design shown has been used up to 300 °C [51].
References
1. R.J. Muha, S.M. Gates, P. Basu, J.T. Yates Jr., Rev. Sci. Instrum. 56, 613 (1985)
2. Model 2B50B, obtainable from Analog Devices, One Technology Way, Norwood, MA
02062-9106
3. Model UDC 500, obtainable from Honeywell, Micro Switch Division, 11 W. Spring Street,
Freeport, IL 61032-4353. More advanced controllers are now available from Honeywell, such
as the UDC 5000 ultrapro
4. M.A. DeAngelis, A.B. Anton, J. Vac. Sci. Technol. A10, 3507 (1992)
5. LabVIEW is available from National Instruments, 6504 Bridge Point Parkway, Austin, TX
78730-5039, USA
6. Type MJC-C, obtainable from Omega Engineering, Inc., P.O. Box 4047, Stamford, CT 06970,
USA
7. Type Lab-PC+, obtainable from National Instruments, see Ref. [5]
8. J. Ahner, D. Mocuta, R.D. Ramsier, J.T. Yates Jr., J. Chem. Phys. 105, 6553 (1996)
9. K.L. Rostov, Institute of General and Inorganic Chemistry, Bulgarian Academy of Sciences,
1113 Sofia, Bulgaria (private communication). This design is used in the laboratory of
Professor Peter Toennies, Max Planck Institut, University of Göttingen, Stroemungsf,
Bunsenstr. 20, D-37037 Göttingen, Germany
10. R.C. Henderson, R.F. Helm, Surf. Sci. 30, 310 (1972)
11. J.A. Stroscio, W. Ho, Rev. Sci. Instrum. 55, 1672 (1984)
12. K. Jacobi, W. Ranke, J. Electron. Spectrosc. Relat. Phenom. 8, 225 (1976)
13. M.J. Bozack, L. Muehlhoff, J.N. Russell Jr., W.J. Choyke, J.T. Yates Jr., J. Vac. Sci. Technol.
A5, 1 (1987)
References 379
14. Metal vapor pressures are listed and graphed, in Handbook of Electron Tube and Vacuum
Techniques, ed. by F. Rosebury (American Institute of Physics, New York, NY, 1993),
pp. 142–148
15. V.A. Ukraintsev, J.T. Yates Jr., Surf. Sci. 346, 31 (1996)
16. W. Ho (private communication)
17. H. Nishino, W. Yang, Z. Dohnalek, V. Ukraintsev, J.T. Yates Jr., J. Vac. Sci. Technol. A15,
182 (1997)
18. V. Smentkowski, J.T. Yates Jr., J. Vac. Sci. Technol. A14, 260 (1996)
19. V.M. Donnelly, J.A. McCaulley, J. Vac. Sci. Technol. A8, 84 (1990)
20. G.E. Thomas, W.H. Weinberg, J. Vac. Sci. Technol. 16, 87 (1979)
21. G.B. Hoflund, R.E. Gilbert, J. Vac. Sci. Technol. 18, 56 (1981)
22. J.A. Stroscio, L.J. Richter, W. Ho, Rev. Sci. Instrum. 55, 732 (1984)
23. J. Lecante, R. Riwan, C. Guillot, Surf. Sci. 35, 271 (1973)
24. Standard and custom wound W filaments may be obtained from C.M. Furnaces, Inc., 103
Dewey Street, Bloomfield, NJ 07003
25. J. Klebanoff, V.H. Ritz, R.E. Thomas, J. Vac. Sci. Technol. A2, 1396 (1984)
26. Spectra-Mat Inc., 100 Westgate Drive, Watsonville, CA 95076
27. V.S. Smentkowski, J.T. Yates Jr., J. Vac. Sci. Technol. A11, 3002 (1993)
28. P.H. Holloway, J.B. Hudson, Rev. Sci. Instrum. 43, 1045 (1972)
29. T.E. Jones, W.C. McGinnis, J.S. Briggs, Rev. Sci. Instrum. 65, 977 (1994)
30. J.C.S. Wong, A. Linsebigler, G. Lu, J.-F. Fan, J.T. Yates Jr., J. Phys. Chem. 99, 335 (1995)
31. F. Bozso, L. Muehlhoff, M. Trenary, W.J. Choyke, J.T. Yates Jr., J. Vac. Sci. Technol. A2,
1271 (1984). (Here the cement used was mixed with graphite in order to permit electrical
conductivity for electron spectroscopy of a SiC single crystal)
32. A zirconia-based refractory cement was employed which matched the thermal expansion
coefficient of TiO2. This particular cement is Aremco 571. The address is Aremco, P.O. Box
429, Ossining, NY 10562-0429
33. E.P. Marsh, F.L. Tabares, G.A. Bach, J.P. Cowin, Rev. Sci. Instrum. 57, 3134 (1986)
34. J.R. Creighton, Dept 1126, MS 0601, P.O. Box 5800, Sandia National Laboratories,
Albuquerque, NM 87185-0601 (personal communication)
35. J.C. Burns, P.C. Stair, R. Haidle, J. Vac. Sci. Technol. A8, 3886 (1990)
36. Vacseal is made by Space Environment Laboratories, RO. Box 1061, Boulder, CO 80306
37. Metal Finishing Guidebook and Directory (Metals and Plastics, Hackensack, NJ, 1989),
pp. 244–265
38. F. Bozso, L. Muehlhoff, M. Trenary, W.J. Choyke, J.T. Yates Jr., J. Vac. Sci. Technol. A2,
1271 (1984)
A5, 1 (1987)
40. Obtain from Aremco Products, Inc., P.O. Box 429, 23 Snowden Avenue, Ossining, NY
10562-0429; the cement should be preserved in the refrigerator and has a limited shelf life
41. K.A. Schultz, E.G. Seebauer, Rev. Sci. Instrum. 63, 218 (1992)
42. R.W. Palmberg, W.T. Peria, Surf. Sci. 6, 57 (1967)
43. T.E. Madey, J.T. Yates Jr., in Structure et Proprietes des Surfaces des Solides,
No. 187 (Editions de Centre National de la Recherche Scientifique, Paris, 7–11 July 1969),
p. 155
A5, 1 (1987)
45. C.S. Pitcher, P.C. Stangeby, G.E. Austin, G.M. McCracken, J. Vac. Sci. Technol. A5, 383
(1987)
46. Pfizer, Inc., 235 E. 42nd Street, New York, NY 10017
47. P. Winsemius, H.P. Lengkeek, Rev. Sci. Instrum. 44, 229 (1973)
48. Obtain “Dag 660” from Acheson Colloids, 1607 Washington Avenue, P.O. Drawer 611747,
Port Huron, MI 48061-1747
49. J.R. Conrad, S.F. Home, D.R. Pirkle, R.P. Torti, Rev. Sci. Instrum. 51, 144 (1980)
50. C.L. Juurlink, C. Hahn (personal communication)
51. R.G. Jones (private communication)
Chapter 35
Cooling Samples
35.1 6 K—Cryogenic Crystal Holder
Cooling single crystal samples in ultrahigh vacuum below 77 K was at one time
considered to be technically difficult because of the perceived need for cooled
radiation shields for the structure holding the crystal. Recently, excellent cooling
properties have been achieved without a radiation shield using a number of
refinements that have permitted rapid cooling to 6 K, control of temperature to
30–70 mK at 10 K, and measurements with absolute temperature errors of about
0.7 K at 10 K and 1.5 K at 100 K [1].
Figure 35.1 shows the design for efficient cooling [1]. A liquid He cryostat [2]
containing concentric ducts for liquid admission and gas transport away is mounted
on a standard UHV manipulator. While the publication [1] shows the use of OFHC
Cu cooling members, recent refinements have substituted highly polished Ag
members for Cu primarily because the surface of Cu oxidizes slowly in use, giving
a nonreflective surface that is not as efficient as the shiny Ag surface in reducing
warming from the radiation fromthe surroundings [3]. All heat-conducting inter-
faces are made by pressing smooth surfaces together and not by welding, which
produces poor thermal contact. A 10 × 10 × 55 mm3 block of Ag is bolted to the
bottom of the cryostat and holds two 1-mm-thick sapphire plates that are separated
from each other by a gap. Three 2-mm-diameter screws, with ceramic isolation
washers, apply force to hold two Ag electrical conduction leads in good contact
with the sapphire plates. The screws pass through the lower Ag bar in large
clearance holes that assure that no short circuits can occur. Each Ag bar is covered
with a stainless steel force plate to avoid distortion of the Ag owing to the force of
the screws. The two Ag conduction leads have 4-mm-diameter stainless steel power
leads exiting from their back and passing up the cooled cryostat surface. These
power leads are heat-stationed twice each, in series, to the cooled cryostat wall to
reduce heat flow downward to the crystal assembly. The heat stations involve
sapphire electrical isolation junctions, mounted 50 mm apart on the cryostat shield.

DOI 10.1007/978-3-319-17668-0_35
382 35 Cooling Samples
Fig. 35.1 6 K—cryogenic crystal holder radiation
The crystal is mounted on the Ag power leads using tungsten rods (5 mm

long × 2 mm in diameter). These snugly fit into holes in the end of the Ag leads and
are tightly connected, but not welded, to Ta stubs of about 2-mm length at right
angles, as shown in Fig. 35.1. The crystal is mounted on two different contact leads
(6 mm long) which consist of pairs of W/25 % Re wires of 0.5-mm diameter and two
Ir wires of square cross section, 0.45 mm × 0.45 mm. The logic for this mounting
structure is that at low temperatures pure metals do not have the needed resistivity for
ohmic heating, and an alloy must be used for this purpose. Conversely, the Ir rod has
good thermal conductivity at low temperatures. Unfortunately, the thermal expan-
sion coefficient is different for the two wires, so that heating the crystal puts torque
on the welds. To reduce this stress, the flexible Ta leads are incorporated, and the
spacing between the W/Re and the Ir leads is kept small.
35.1 6 K—Cryogenic Crystal Holder 383
The thermocouple (in this case, type K) is welded to the back of the crystal and
the leads pass upward after also being heat-stationed on the cryostat with isolation
by sapphire. Heating rates from 0.01 to 100 K/s are possible. A heating current of
70 A DC at about 8 V is sufficient to bring the crystal to 1600 K. The cooling time
between 650 and 6 K is 3 min.; from 1600 to 6.5 K the time is 5 min. At steady
state, the consumption of liquid He is 4 L/hr. With the cryostat cool, the base
pressure is below 10−11 mbar.
Reference [1] is highly instructive about the pitfalls of using cryogenic methods
in ultrahigh vacuum applications.
Others have employed closed-loop He cryostats for cooling of crystals in UHV
that do not achieve the very low temperatures described here [4–6].
35.2 Cooling Crystals by Contact with a Thermal

Reservoir
One of the best ways to employ a low-temperature, closed-cycle cryogenic source

with limited cooling capacity is to use a heat absorption reservoir that is continu-
ously cooled, but only used periodically for cooling the sample. A design using this
principle in shown in Fig. 35.2 [7], based on an earlier design [8].
The crystal and its mounting assembly, shown in Fig. 35.2a, is held on the end of
a conventional manipulator, and it may be heated either by radiation from the W
filament or by electron bombardment. The entire crystal and the heating assembly is
electrically isolated from the manipulator by means of a machinable ceramic sup-
port. The sample may be contacted to the thermal reservoir by means of a conical
thermal switch in order to initiate cooling. The stainless steel spring placed under
compression when this contact is made to yield the best thermal contact possible
between the two OFHC Cu components.
The cooling rate using this device is very impressive, and cooling from 1023 K
to about 20 K occurs in 300 s, as may be seen in Fig. 35.2b.
A second design in which contact between the sample and the closed cycle
refrigerator is maintained continuously is given in [9].
35.3 Cooling Samples on Manipulators—Thermal

Conductivity Issues
A variety of designs for holding samples on manipulators is available in the liter-

ature, and most of these have a requirement that the crystal be isolated electrically
from the manipulator, while being in contact thermally with a cryogenically cooled
reservoir, usually made of oxygen-free high conductivity copper (OFHC Cu).
Comparisons of cooling rates for different methods of electrical isolation have been
made and are presented below [10].
Fig. 35.2 Spring-loaded cooling by contact. a Side view of manipulator + thermal contact,
b cooling curve
Figure 35.3a shows the experimental apparatus used in this work. A Cu reservoir
with polished planar faces holds two Cu support bars that are clamped with equal
force to the reservoir. The total contact area is 3.13 cm2 for both support bars. The
temperature of each bar (T1 and T2) can be measured during cooling, using type K
35.3 Cooling Samples on Manipulators … 385
Fig. 35.3 Cooling samples on manipulators. a Manipulator cooling section. b Comparative

cooling rates. c Comparison of cooling rates—various insulators. d Rate of crystal cooling versus
support cooling
thermocouples that are wedged into small holes in the support bars. In addition, the
crystal temperature can also be measured as cooling occurs to determine the
influence of the 0.025-mm-diameter tungsten wire (and the associated spot welds)
on the crystal cooling rate, compared to the support bar cooling rate.
Figure 35.3b shows cooling experiments comparing three different insulator
materials. The slowest cooling rate of the support bars is achieved using Kapton
films, 0.037 mm in thickness. Faster cooling of the support bars is achieved using
0.500-mm sapphire single crystal insulators. The fastest cooling is achieved when
an indium film is used to join the sapphire to the cooling reservoir and the support
bars. For cooling times up to 90 s, cooling of the central reservoir to 77 K is taking
place, and the data in this time period are disregarded in the measurements.
Figure 35.3c shows semilogarithmic plots of the cooling rate after the reservoir
has reached constant temperature. The slope of these plots is a measure of the rate
of cooling through each insulator junction. It may be seen that the Kapton exhibits
cooling at 81 % of the rate of the sapphire, and that the indium coating on the
sapphire enhances its cooling rate by about 6 %.
Figure 35.3d shows a comparison of the cooling rate of the sapphire-insulated

support, compared to the cooling rate of a metal single crystal. This comparison was
carried out in order to test the effect of the thermal conductivity of the tungsten wire
(and its spot-welds and screw pressure contact). The crystal was found to cool at a
rate of 93 % of the cooling rate of the copper bar supports, indicating that this
crystal support design is certainly satisfactory.
The differences in the sapphire- and Kapton-controlled cooling rates are in
accordance with the measured thermal conductivity of these materials1. Thus,
sapphire, or indium-coated sapphire, exhibit superior qualities compared to Kapton
of the thickness employed here. Thinner Kapton is likely to be mechanically
damaged when it is placed under pressure or baked out and may therefore exhibit
electrical leakage.
35.4 Vacuum-Jacketed Cryogenic Manipulator Rod
Long manipulator rods filled with liquid nitrogen can act as cryogenic surfaces that
can physisorb or condense gases used in surface science studies. As the liquid
nitrogen level drops, reevolution of these gases can cause problems in experiments
conducted under ultrahigh vacuum. The design shown in the Fig. 35.4a eliminates
this problem by using a vacuum jacket to isolate the outer wall of the manipulator
rod from the inner tube, which is cooled with liquid nitrogen [11].
A copper reservoir is supplied with liquid nitrogen through a 1/8-in. diameter
stainless steel tube. The excess liquid and vapor exits through a 5/16-in.-diameter
outer tube, which is concentric with a l/2-in.-diameter stainless tube welded to a
Conflat flange. The outer tube is pumped and contains an adsorbent material for
maintenance of the vacuum, as is sometimes used in metal Dewar flask vacuum
jackets. The liquid nitrogen inlet and outlet tubes are able to rotate inside a com-
pression fitting that contains a spring-loaded Teflon seal that allows the sample to
be rotated while the nitrogen delivery tube and the outlet tube remain fixed. Liquid
nitrogen is drawn through a latex rubber tube into the assembly from an unpres-
surized storage Dewar by means of a small carbon vane vacuum pump [12] attached
to the outlet tube. A valve at the inlet of the pump is used to adjust the flow of liquid
nitrogen. This pump is isolated from the liquid nitrogen reservoir by a length of
copper tubing that permits the liquid nitrogen to convert to gaseous nitrogen before
entering the pump. The length of the manipulator is 1.3 m below the Conflat flange.
Owing to thermal contraction, the inner tube will decrease in length upon
cooling. Since this is fixed at only one end, it may contract and expand freely. The
differential expansion of the outer two tubes is taken up by a formed bellows that is
1
Tabulated values for thermal conductivity are sapphire, 41.90 J/s km at 273 K; 100 HN Kapton,
0.10-0.35 J/s km at 276 K. Goodfellow Catalog (Goodfellow Metals, 1995), p. 433; p. 501. In
addition reference [10] gives information and references on thermal conductivities as a function of
temperature. Sapphire exhibits a greatly enhanced thermal conductivity at 80 K of 960 J/sKm.
35.4 Vacuum-Jacketed Cryogenic … 387
Fig. 35.4 Vacuum-jacketed cryogenic manipulator rod. a Vacuum-jacketed sample support tube.
b Sample holder cryogenically coupled to copper reservoir
welded to the compression fitting. Some contraction and expansion of the outer tube
still occurs as the inner tube changes temperature.
The sample holder shown in Fig. 35.4b is made of OFHC copper and is held on
the base of the copper reservoir by four screws that pass through a 0.5-mm-thick
sapphire plate that permits efficient cooling (see p. 383) as well as electrical iso-
lation. A Ta support structure consisting of two arms for the sample is bolted to the
copper support arms. Heating leads and thermocouple leads to the sample pass on
the outside of the vacuum-jacketed manipulator rod and are heat-stationed on the
lower cooled sections of the assembly.
In operation, the sample is cooled to 90 K when liquid nitrogen is the refrigerant.
Measurements indicate that the outer tube slowly cools to about 200 K owing to
cooling from its lower end, where it contacts the copper reservoir. This could be
prevented by the installation of a small electrical heater on the outer tube.
Because of the length of the manipulator rod, a lower bearing is supplied in
which the manipulator rod rotates. This prevents vibration of the sample inside the
apparatus. Rotation and translation of the sample are supplied by a standard dif-
ferentially pumped rotary seal (see p. 6) and a standard x-y-manipulator stage.
35.5 Manipulator Component with Cryogenic Azimuthal

Rotation
A simple design [13], involving indirect heating of a sample that can be cooled to
cryogenic temperatures and undergo limited azimuthal rotation, is shown in
Fig. 35.5. The sample is mounted on a metal cap fixed to a sapphire cylinder that
makes thermal contact with a Cu-Be cup through a novel ring of Cu-Be finger stock
[14]. The Cu-Be cup is held by a hollow shaft that rotates in the Cu-Be block through
Fig. 35.5 Azimuthal rotation—cryogenic

35.5 Manipulator Component with Cryogenic Azimuthal Rotation 389
stainless steel ball bearings that permit azimuthal rotation of the cup and the asso-
ciated parts connected to the sample. The sapphire is a good thermal conductor but is
an electrical insulator. In order to provide more cooling of the sample, four Cu braids
are connected between the sample mount and the Cu-Be block, which contains
internal pathways for flowing liquid nitrogen refrigerant. Azimuthal rotation is
achieved using a rack and pinion connection allowing linear translation of the rack to
produce rotation of the Cu-Be cup and the sample. An internal heater may be used
for radiation heating or electron bombardment of the sample and its mount. It is
electrically isolated and consists of a two-hole ceramic tube fixed in position with a
ceramic collar so it does not rotate as the Cu-Be cup is rotated. The azimuthal
freedom is ±45°, limited by the length of the copper cooling braids. The temperature
range of the sample is from 130 to 1000 K, with cooling times of about 45 min.
35.6 Cooling Using Direct Contact
The use of cooling braids to supply flexibility in cooling of samples also causes a
limitation in cooling rate owing to the low thermal conductivity associated with
braids having a desired flexibility. Therefore, methods to cool movable samples by
direct contact with a cold reservoir have been developed.
The first of the methods [15] to be described is shown in Fig. 35.6. Here a 200-g
Cu tank is connected directly to a closed-cycle He refrigerator. The cold second
stage is shielded from external radiation by a first stage shield as shown. The Cu
tank contains a slot for the insertion of the sample holder containing a freshly
cleaved semiconductor sample in this case. This sample holder makes contact with
a droplet of molten Ga for the production of a metallic joint of high thermal
conductivity. The Ga can be melted by radiation heating from an attached heater as
shown in Fig. 35.6. The Cu tank can then be cooled to 20 K over a 140 min cooling
period after starting the refrigerator. The long cooling time is probably a result of
the large heat capacity of the massive Cu tank in this design. Temperature
adjustment can be achieved by using the radiative heater in opposition to the He
refrigeration unit. During heating, outgassing of the assembly with bursts of
pressure is minimal since cooling is being maintained and slow heating occurs.
The closed cycle refrigerator is mounted on a bottom-entry manipulator per-
mitting rotation and x, y, z-translation. Translation in the z-direction is especially
important because of thermal contraction of the device on cooling.
A thermal switch arrangement involving a high area Cu-Cu contact between a
liquid-He cooled reservoir and a crystal support assembly has produced a fast cooling
performance and a crystal temperature of 20 K after long equilibration times [16].
A design of a highly efficient attachment of a removable closed-cycle refrigerator
to a crystal mount has been described, with a limiting low temperature of 11 K [17].
Other arrangements for direct contact cooling are shown on p. 383 and 392.
Fig. 35.6 Cooling without thermal braid
35.7 Cooling Reservoir—Horizontal Manipulator
A horizontal manipulator is inherently difficult to cool with liquid nitrogen since

gravity may cause the refrigerant to drain from the tubing as it is being supplied. In
addition, rubber seals at the inlet end of the manipulator may cause difficulty when
the manipulator is rotated about its axis because of interference with the electrical
leads passing into the center of the manipulator as well as because of the inherent
difficulty of retention of refrigerant at all rotational angles.
35.7 Cooling Reservoir—Horizontal Manipulator 391
Fig. 35.7 Cooling reservoir—horizontal manipulator
The design shown in Fig. 35.7 uses two copper disks as heat transfer devices,
and at all rotational angles the disks are immersed in liquid nitrogen which is
confined by gravity in the expanded end of the manipulator [18]. The two current
carrying leads are brazed into holes in alternate copper disks so that efficient
thermal conduction will occur to each lead at all rotational angles. The use of two
Cu disks assures electrical isolation between the current carrying leads, which are
insulated up to the braze points with Teflon sleeves. The thermocouple leads pass
through an elongated hole in each disk and are insulated from the disks and each
other by Teflon sleeves.
The design shown in Fig. 35.7 is coupled to the manipulator by VCR fittings,
and the reader is referred to the original article [18] for details of this arrangement
and the alignment procedures for assembly. Tests with the design shown have
reached sample temperatures in the range 103–123 K.
35.8 Cryogenic Connections
The design of efficient heat exchange methods that also permit motion of a cooled
sample is made difficult by two factors: (1) the provision of sufficient thermal
conductivity in the materials making the thermal connection, and (2) the provision
of excellent thermal contacts between interfaces along the conduction pathway.
The best method for making a thermal contact between interfaces along the
conduction path is by welding [19, 20]. Brazing, while better than most pressed
mechanical connections, provides an alloy that is inferior to its metal components in
thermal conductivity. Wherever possible, welded thermal contacts between the
same materials (such as Cu or Ag) should be made.
Figure 35.8 shows a braid assembly made of OFHC Cu that is connected to a
cryostat and to a Cu sample holder by means of welds at two of the three contact
points. Because the Cu flexible braid work hardens upon extended usage, it should
be considered an expendable component. The Cu heat conductivity can be
improved by hydrogen firing at 1 bar pressure at a temperature of 600–800 °C for
3–5 min. Longer hydrogen firing may cause hydrogen embrittlement of the braid.
The effect of limited hydrogen treatment is probably to remove copper oxide films
that impede heat transfer at interfaces in the braid and on its connectors. Hydrogen
firing should be done only by experienced persons!
35.9 Sample Cooling Using Thermal Braid
Many experiments require sample cooling below room temperature and below the
temperature of liquid nitrogen. The use of copper thermal braid for this purpose
provides flexibility in sample motion, and with proper design copper braid can be
employed for efficient cooling.
In Fig. 35.9a, a copper braid is employed with a detachable thermal switch to
provide intermittent cooling of a sample mounted on a movable manipulator [21].
Fig. 35.8 Cryogenic connections

35.9 Sample Cooling Using Thermal Braid 393
Fig. 35.9 Cooling using thermal braid. a Detachable cooling assembly. b Continuously Coupled
cooling assembly
The braid is firmly connected to the sample mount by means of clamps made of Cu.
The other end is connected to a solid Cu cylinder, weighing about 100 g which
makes contact with a Cu cup. The Cu cup is cooled by flowing liquid nitrogen
through a coil wrapped and brazed in place. The use of brazing is highly recom-
mended, since the thermal conduction through a brazed joint is far superior to any
other mechanical contact except for a weld between identical metals. A reversible
metal contact between the Cu cylinder and the Cu cup is made using a liquid
gallium droplet (mp = 32 °C at a vapor pressure of 10−14 Torr). Thus, to melt the
Ga, hot air is passed through the coil on the cup; the weight is lowered into the cup
by moving the sample holder to which it is attached, and subsequent cooling freezes
the Ga, making a good thermal and mechanical contact. Sample cooling to
93 ± 10 K then occurs within about 200 s.
In Fig. 35.9b, a thermal braid is permanently attached to the sample holder [22].
The manipulator swings about a radius of 15 cm, and various degrees of freedom
are present. The sample holder is thermally isolated from the manipulator shaft and
most of the arm. The thermal braid is attached to a closed-cycle He refrigerator [23],
which has two stages of cooling to 80 K (15 W) and to about 10 K (1 W). The 80-K
stage provides cooling for a radiation shield that surrounds the low-temperature
stage, as shown schematically in the Figure. Modeling of the energy transfer was
done in order to compute the optimal dimensions for the sample holder and the
braid. The minimization of thermal conduction to the sample holder makes thermal
radiation the primary source of unwanted heat. Since the ratio of the surface area of
the sample holder to that of the thermal braid is about 1:12, a cooled radiation shield
around the sample is unnecessary in this design. It was calculated that the
cross-sectional area of the Cu braid was 0.075 cm2 on the basis of thermal con-
duction and thermal radiation considerations. It is of interest to note that this design
achieved sample temperatures of 14–15 K routinely with only 1 W of cooling in the
second stage of the closed-cycle refrigerator.
Copper braid should be made of oxygen-free, high-conductivity (OFHC) Cu for

maximum efficiency. UHV-compatible closed-cycle refrigerators are commercially
available [3]. It should be noted that efficient thermal connections to copper braid
can be made by brazing the ends with silver solder, which can then be deformed by
clamping. Procedures for achieving the same objective without brazing have been
described [24].
35.10 Flexible Cooling Reservoir
The use of copper braid for cooling samples on a manipulator is widespread. This
method of cooling avoids complex crystal motion devices that must also conduct
heat efficiently, and mountings with six degrees of motional freedom are often built
with braid cooling.
As the crystal and the manipulator shaft move, there will be the need for relief of
the strain in the copper cooling braid. While a long braid may be used to give a
sufficient degree of motion, this decreases the cooling efficiency of the system. It is
also possible, however, to mount the cooling reservoir itself on a flexible mount as
shown in Fig. 35.10 [25], thereby permitting a shorter braid to be used.
A mini-Conflat flange, mounted on the manipulator flange, is welded to a
standard Swagelock Tee. A stainless tube enters the vacuum system through this
flange and is connected to a very flexible bellows and then to a double kovar/glass
seal, which acts as an insulating break, electrically isolating the cooled reservoir
from the vacuum system and ground. An OFHC Cu reservoir is brazed to the end of
the tube assembly, and contains a polished flat to which the copper braid and its
copper lug are mounted (see p. 392). An indium foil gasket makes good thermal
contact when compressed by a bolt that passes through the flat and is drawn down
by a nut on the back side of the flat. The alternate use of threads in the soft copper
flat is not recommended because of the possibility of stripping threads.
Liquid nitrogen is supplied to the copper reservoir by suction from a styrofoam
cup or small Dewar through 1/8-in.-diameter Teflon tubing that can be sealed inside
the Swagelock fittings. Sample temperatures below 100 K can easily be achieved.
35.11 Enhanced Cooling Using Liquid Nitrogen
A simple method [26] has been devised and tested for cooling below the temper-
ature limit established by liquid nitrogen (77.3 K). The method involves bubbling
He gas through the liquid nitrogen (transpiration); evaporation of liquid nitrogen
35.11 Enhanced Cooling Using Liquid Nitrogen 395
Fig. 35.10 Flexible cooling reservoir

requires heat absorption from the fluid itself, and the temperature falls as a
consequence of the evaporation until the triple point (63.1 K) is reached. Thus, for
work requiring temperatures slightly below that achieved using liquid nitrogen, this
method is convenient. Compared to pumping liquid nitrogen to reduce its equi-
librium vapor pressure and to achieve lower temperatures, this method offers the
convenience of not requiring high-capacity pumps connected to manipulators by
large flexible tubing, a procedure that is often mechanically complex and
inconvenient.
A He bubble in equilibrium with N2(l) will contain both He and N2, so that the
partial pressure of N2 will be less than the total pressure within the bubble. Thus as
N2 vaporizes within the bubbles, cooling will occur to reduce the temperature below
the normal boiling point of N2(l). Figure 35.11a shows a demonstration apparatus
where this experiment was carried out. A fritted disk bubbler was used in the
right-hand Dewar flask, while the thermocouple reference junction was maintained
at 77.3 K in the left-hand Dewar. After about 80 min of transpiration, solid nitrogen
crystals were observed, and the temperature reached that of the triple point. When
the He flow was stopped, the temperature remained stable for some minutes as the
solid nitrogen melted at the triple point; heat leakage into the Dewar was respon-
sible for this as well as for the gradual warming beyond 120 min.
Figure 35.11b shows a similar experiment in a 1.42-m column of liquid nitrogen
that comprises the cooling reservoir of a reentrant stainless steel manipulator on an
ultrahigh vacuum system. The temperature of a metal single crystal was measured
as He bubbled into the bottom of the reservoir from the end of a 0.097-cm-diameter
tube. The crystal temperature fell from 86.5 to 74 K during He flow. Stability of
±0.5 K was achieved, and excellent cooling control maybe maintained in this
temperature range by adjustment of the He flow rate. Because of the limited amount
of N2(s) produced at the bottom of the reservoir, interruption of the He flow
produced an almost immediate rise in the crystal temperature as shown. Because of
the length of the column of liquid N2, additional liquid N2 may be added without
disturbance, as shown at the time indicated by the dots in Fig. 35.11b. This is a
distinct advantage of using a long column of refrigerant where transpiration cooling
occurs primarily at the bottom.
35.12 Level Alarm-Liquid Nitrogen
There are many designs for level alarms and controllers for liquid-nitrogen Dewars
and Fig. 35.12 shows one of the simplest alarm devices for keeping a small supply
Dewar filled [27]. A cork float senses the level of liquid nitrogen in the Dewar, and
when the level drops to its lower limit, contact is made between a metal plate and an
inverted metal cup at the top of the sealing cork lid. This completes the circuit,
causing both an audio and a light signal to be emitted, alerting the operator to refill
the Dewar through the refill funnel. The audio oscillator normally used to drive an
Auger spectrometer employing a lock-in amplifier can be used.
35.12 Level Alarm-Liquid Nitrogen 397
Fig. 35.11 Cryogenic trick for enhanced cooling with liquid N2, a Attainment of N2 triple point
by He transpiration, b Cooling of metal crystal by he transpiration
The Dewar is sealed to the top of the entrance tube using cotton plus ice to make
the seal. Simply wet the cotton from a plastic squeeze bottle, admit a little liquid
nitrogen to freeze the water, and repair any leaks with more water on the cotton.
This type of seal does not put any stress on the glass Dewar during warm-up in
contrast to compression seals, which will allow the expanding metal to strain the
glass Dewar bottom, sometimes resulting in breakage.
Fig. 35.12 Level alarm—liquid nitrogen
35.13 Liquid Nitrogen Reservoir—Non-leaking
For long term maintenance of liquid nitrogen in a cooled manipulator an external

reservoir may be used. Shown in Fig. 35.13 is a design to attach a silvered Dewar
flask to a central tube with a non-leaking seal. The central tube is evacuated through
its connectivity to the vacuum system, and hence exhibits low thermal conductance
to the outer wall. The Dewar is custom made to have a hole in its bottom, a few
millimeters larger in diameter than the outside diameter of the central tube.
Ordinary plumber’s Teflon tape is generously wrapped in a conical shape around
35.13 Liquid Nitrogen Reservoir—Non-leaking 399
Fig. 35.13 Liquid nitrogen

reservoir—non-leaking
the central tube and above the platform. The conical shape permits the Dewar to be
tightly sealed to the central tube by forcing it to slightly compress the many layers
of slippery Teflon tape. This makes a good seal for liquid-N2 which normally has a
low viscosity and tends to leak. Rubber seals become brittle in this application and
will fail. The Dewar is fitted with a loose-fitting styrofoam cover to prevent
excessive ice from forming. There may be condensation near the Teflon seal and
this can be eliminated by using a small fan for drying.
35.14 Cryogenic Cooling to Below 4 K—Photon

Stimulated Desorption of Adsorbed He
For studies of adsorbed He, it is necessary to cool below the boiling point of He.
This is especially difficult because of the need to reduce radiation heating at the
crystal to minimum values. The manipulator shown here achieves this with minimal
use of liquid-He. It is designed to also raise the crystal temperature to more than
1570 K, and this has been demonstrated with Ru single crystal samples. The
manipulator is immersed in a radiation shield in a liquid-N2 bath which itself is
cooled to below 77 K by pumping with a Roots blower. After heating the crystal to
high temperature in UHV to clean it, liquid He is admitted to the central Dewar
assembly and about 0.5 L accumulates. The inner He cryostat is then pumped with a
Roots blower until the final temperature of *1 K is reached, in an amazingly short
time of about 20 min after the high temperature cleaning procedure. About 0.3 L of
liquid He is consumed leaving 0.2 L for operation at the base temperature for about
Fig. 35.14 Low temperature cryostat for He adsorption/photon desorption
4 h. Synchrotron radiation enters the cryoshield through a small aperture. Because

adsorbed He on the crystal is removed by radiation from 300 K surfaces, it was
necessary to design the TOF spectrometer for liquid-N2 cooling also. The TOF
spectrometer uses a pair of microchannel plates for amplification of the signal from
desorbing ions and metastables, as well as for electrons emitted. It also contains a
gas dosing tube for exposing the crystal to adsorbates [28] (Fig. 35.14).
35.15 Static Temperature Control by Mixing Gas Streams 401
35.15 Static Temperature Control by Mixing Gas Streams
Occasionally it is desirable to have accurate and steady temperature control below

room temperature in an ultrahigh vacuum system, where bucking a cryogenic
temperature against ohmic heating (as on a normal manipulator) is impossible. This
objective can be achieved easily by having two temperature controlled gas steams
which are mixed to achieve the desired intermediate temperature [29]. Using Dewar
flasks filled with cryogenic fluids like liquid N2 (77.3 K) and other systems such as
dry ice/acetone (194.8 K), stable bath temperatures can be achieved. Flowing N2
gas is brought to temperature using an immersed coiled tube in each temperature
bath, T1 and T2. By adjusting the mixture using the two flow control valves, V1 and
V2, stable intermediate temperatures may be achieved in a massive OFHC Cu
thermal reservoir inside the UHV system through which the flowing gas circulates.
To have the highest degree of fine control, gas leak valves like those used in UHV
systems can be used to achieve stable gas flow conditions. Although it would be
best to use cryogenic systems undergoing a phase transition for each bath to achieve
the highest temperature stability, there are a number of fluid/refrigerant mixtures
which are used for low temperature baths in the absence of phase transition [30]
(Fig. 35.15).
35.16 Gas Thermal Switch
The ability to turn on and turn off cooling to a sample and its support may be useful
in ultrahigh vacuum operations where condensation on the sample/support system
must be avoided. This could be the case when the sample and its support are to be
immersed in a high-pressure gas in a high-pressure vessel inside the UHV system.
The design of a gas thermal switch to vary the cooling rate to a sample and its
support is shown here [31].
The principle of the gas thermal switch is to use a heat transfer gas (He(g) is
ideal) to conduct heat when cooling is needed, and then to evacuate the He(g) when
cooling is not required. As shown in the exploded view of Fig. 35.16a, the OFHC
copper cooling stage is maintained at the cryogenic temperature desired by means
of contact with a reservoir of the cryogenic fluid (l-N2 in this case). Brazed to this
heat station is a concentric assembly of copper tubes. The cooling tubes nest into a
set of concentric tubes connected to the OFHC copper sample holder that is ther-
mally connected to the sample. Cooling is achieved by introducing about 0.5 Torr
of He(g) into the 10-cm3 volume containing the thermal switch. The cooling is
interrupted by pumping the He(g) away. Given the geometry of the thermal switch
(not exactly specified in the paper), the thermal conductivity varies between about
Fig. 35.15 Static temperature control by mixing gas streams
0.8 and 1.4 W/K with He(g) as a heat transfer medium to about 0.03 W/K with a
vacuum of <5 × 10−3 Torr, which is achieved in 10-s pumping time.
Figure 35.16b shows the performance of the gas thermal switch. He(g) is
admitted with the sample at 300 K and the cooling station at 100 K. Cooling starts
immediately with a time constant of about 2 min; after some cooling, the He(g) is
35.16 Gas Thermal Switch 403
Fig. 35.16 Gas thermal switch. a Thermal switch—exploded view. b Cooling performance using
vacuum and He(g)
evacuated, and cooling slows. Then He(g) is reintroduced, and cooling occurs with
a time constant of about 2.6 min. The decrease in cooling rate in the second stage of
cooling may be due to the decrease in the thermal conductivity of He as its tem-
perature falls. Cooling to about 102 K is achieved in less than 10 min.
References
1. H. Schlichting, D. Menzel, Rev. Sci. Instrum. 64, 2013 (1993)

2. Constant flow cryostat, Air Products, APD-Cryogenics, P.O. Box 2802, Allentown, PA 18105
3. H. Schlichting (private communication)
4. K. Wandelt, S. Daiser, R. Miranda, H.-J. Forth, J. Phys. E17, 22 (1984)
5. A. Winkler, M. Hausenblas, M. Leisch, K. Rendulic, Vacuum 40, 39 (1990)
6. M. Milun, P. Pervan, T. Valla, P. Dukic, Rev. Sci. Instrum. 62, 3116 (1991)
7. J. Yoshinobu, H. Ogasawara, M. Kawai, Surf. Sci. 360, 234 (1996)
8. S. Masuda, Department of Chemistry, College of Arts and Sciences, The University of Tokyo,
Komaba, Meguro-ku, Tokyo 153, Japan
9. A. Winkler, M. Hausenblas, M. Leisch, K.D. Rendulic, Vacuum 40, 39 (1990)
10. C. Rusu, J.T. Yates Jr, J. Vac. Sci. Technol. A15, 436 (1997)
11. D.J. Frankel, B. Fruhberger, R.H. Jackson, D.J. Dwyer, Rev. Sci. Instrum. 64, 2368 (1993)
12. Model 0522, Gast Manufacturing Corp., P.O. Box 97, Benton Harbor, MI 49023-0097
13. P.G. Strupp, P.C. Stair, J. Vac. Sci. Technol. A9, 2410 (1991)
14. Catalog No. 97-216. The springs are obtainable from Instrument Specialties Co., P.O. Box A,
Delaware Water Gap, PA 18327. These springs make good mechanical and thermal contact
and permit dimensional changes due to differing thermal expansion coefficients of the sapphire
and the Cu-Be alloy
15. I.M. Vitomirov, C.M. Aldao, G.D. Waddill, J.H. Weaver, J. Vac. Sci. Technol. A8, 3368
(1990)
16. E.E. Chaban, Y.J. Chabal, Rev. Sci. Instrum. 54, 1031 (1983)
18. A. Winkler, M. Hausenblas, M. Leisch, K.D. Rendulic, Vacuum 40, 39 (1990)
Jülich GmbH, Postfach 1913, D-52425 Jülich, Germany (private communication). See also R.
David et al, Rev. of Sci. Instrum. 57, 2771 (1986)
20. Detailed procedures for the preparation and welding of Cu braid are presented in the literature.
See G.L. Fowler, J. Vac. Sci. Technol. A5, 2976 (1987)
21. E.E. Chaban, R.H. Citrin, F. Sette, J. Vac. Sci. Technol. A6, 3018 (1988)
22. V.K.F. Chia, D.K. Veirs, G.M. Rosenblatt, J. Vac. Sci. Technol. A7, 108 (1989)
23. Cryomech GB-220: Cryomech, Inc., 1630 Erie Boulevard E., Syracuse, NY 13210
25. Dr. L.H. Dubois, DARPA/DSO, 3701 N. Fairfax Dr., Arlington, VA 22203-1714 (private
communication)
26. J. Xu, H.J. Jansch, J.T. Yates Jr, J. Vac. Sci. Technol. A11, 726 (1993)
27. Dr. X.-C. Guo, Department of Chemical Engineering, Stanford University, Stanford, CA
28. P. Feulner, private communication from thesis of Sarah Kossler, U. Munich
29. W. Lew, O. Lytken, J.A. Farmer, M.C. Crowe, C.T. Campbell, Rev. Sci. Instrum. 81, 024102
(2010)
30. A.J. Gordon, R.A. Ford, The Chemist’s Companion (Wiley, New York, 1972), pp. 451–452
31. O. Sneh, S.M. George, J. Vac. Sci. Technol. A13, 493 (1995)
Chapter 36
Temperature Measurements
36.1 Tungsten-Rhenium Thermocouples—Calibration

Over a Wide Temperature Range
Tungsten/rhenium thermocouples are often used in ultrahigh vacuum because this

class of thermocouples can go to higher temperatures than all other thermocouple
alloys [1]. Such thermocouples avoid sample contamination effects owing to metal
vaporization, which is common for other thermocouples near the high-temperature
end of their operating range. Unfortunately, a factory calibration of
tungsten/rhenium thermocouples is not available below 255 K [1], although two
calibrations in the range 77–300 K [2], and in the range 4–77 K [3] are published
for W/5 % Re versus W/26 % Re thermocouples. The calibration data from the
three sources have been brought together into a single master calibration that
extends from either 4 or 32 to 2588 K [4]. The reader is referred to the original
paper [4] for calibration tables and for full consideration of the errors present in the
master calibration [4]. It should be noted that the variation of thermocouple alloy
composition from batch to batch may cause larger errors in the temperature regions
not covered by the factory calibration than in those temperature regions which are
reported by the manufacturer.
Figure 36.1a shows the master calibration curve from 4 to 2588 K, using data
from the three sources. This curve is not plotted at high enough resolution to be
useful. A tenth-order polynomial function has been fitted to the data, and the
function is given in [4], as well as at the end of this discussion. The difference
between the calibration data and the tenth order function is plotted in Fig. 36.1b,
showing that differences are always less than 7 K except for the very low

DOI 10.1007/978-3-319-17668-0_36
406 36 Temperature Measurements
Fig. 36.1 Tungsten-rhenium thermocouples—calibration over a wide temperature range.

a Temperature versus potential, W/26 % Re versus W/5 % Re Thermocouple. b Difference plot
(Measured-Calculated) for thermocouple—4 to 2588 K
temperature region. The maximum difference above 800 K is near 2 K. For tem-
peratures below 32 K, it is advisable to use the calibration in [3], which is exclu-
sively concerned with this temperature range. The fitting curve of [3] cannot be
extrapolated with accuracy above 32 K.
The power series for the range 4–2588 K is given below, and tabular solutions to
this equation are given in [4] at intervals of 0.05 or 0.10 mV.
36.1 Tungsten-Rhenium Thermocouples—Calibration Over a Wide Temperature Range 407
Table 36.1 Power series for the range 4–2588 K and solutions
T = ΣCnBn
C0 = 276.5327901 C6 = −4.147285358 × 10−3
C1 = 83.31475475 C7 = 1.529502993 × 10−4
C2 = −15.19381797 C8 = −3.422732361 × 10−6
C3 = 4.431847913 C9 = 4.253729476 × 10−8
C4 = −0.7258516658 C10 = −2.351967817 × 10−10
C5 = 0.06999557967
An extension of the calibration up to 3273 K has been reported [5].
36.2 Measuring Temperatures of Hot Filaments
The temperature of an electrically heated hot filament may be accurately measured

by using it as a resistance thermometer. By determining its resistance compared to
its resistance at a known temperature, quite accurate temperatures may be measured
and maintained.
Figure 36.2a shows a metal filament containing potential leads across the central
third of the filament. The filament is mounted on degassing loops to prevent out-
gassing effects during heating as described on p. 437. A known current, I, measured
to about four significant figures with a digital ammeter, is passed through the
filament using a current-regulated DC power supply. The electrical potential drop,
Vc, across the central third of the filament, is simultaneously recorded to about four
significant figures, using a digital voltmeter. The resistance of the central section of
the filament is calculated from Ohm’s law, Rc = Vc/I, and also the power dissipation
from the central section of the filament is calculated, Pc = IVc. If Rc is known at
some standard temperature, Ts, and if the resistivity, ρ, of the filament material is
known as a function of temperature (Tables for tungsten are readily available in
handbooks), then the temperature of the center section of the filament maybe cal-
culated from the ratio Rc(T)/Rc(Ts) = ρ(T)/p(Ts), where T is the unknown. In
practice, more accurate maintenance of temperature maybe achieved if Pc is
monitored continuously rather than Rc. Typically, from day to day, Pc may be
readjusted to match its previous values very accurately, assuring that the original
temperature is reestablished. However, in practice, because of slow warming effects
in the support system that have a minor effect on Tc, the measured value of Pc will
drift slightly for some minutes until stability of the hot filament temperature is
achieved.
There is one special point that must be emphasized for the most accurate
measurements. The potential leads should conduct as little heat as possible away
from the center section of the filament. To assure that this is the case, it is wise to
etch the potential leads electrochemically to a diameter that is as small as possible,
thereby limiting their thermal conductivity. For a well-made system, heating at
Fig. 36.2 Measuring temperatures of hot filaments. a Inner-potential leads. b Outer-potential

leads. c Nomograph for design of tungsten wire heaters
incandescence will display no optically detectable temperature gradient in the

vicinity of the welds of the thinned potential leads to the central section of the
filament. However, there will be a gradient of temperature at the welds connecting
the filament to the current leads, but the temperature measurement based on the
central section of the filament is made far from this region.
For work that does not require extremely high accuracy or reproducibility, one
may connect the potential leads outside the vacuum system, as in Fig. 36.2b,
making similar measurements as described above. Here, the resistivity ratios will
not accurately correspond to the filament temperature ratios because the filament
temperature gradient, as well as the resistance of the mounting leads, is being
sampled. In practice, measurements made in this fashion are best used to establish
the same filament temperature from day to day, and this may be done with sur-
prisingly high accuracy. The potential leads should be hard-soldered to the input
current leads in order to avoid the effect of having variable contact points as would
occur using clip leads.
36.2 Measuring Temperatures of Hot Filaments 409
One problem that can arise in the use of high-temperature filaments is the
evaporation of the filament material, causing the diameter of the filament to
decrease, upsetting the temperature calibrations. In addition, filaments containing
attached pieces of evaporant material can change resistance over time due to this
effect as well as to alloying effects.
A useful nomograph for estimating the power requirements needed to heat
tungsten filaments of various diameters is shown in Fig. 36.2c. To determine
heating parameters, a straight edge is used across the chart. The example shown in
the center of Fig. 36.2c is a guide to tungsten wire heater design, including cor-
rections for tungsten wire cooling at the ends welded to the feedthrough leads.
36.3 Pyrometric Measurement of Temperatures
The measurement of the brightness temperature of a hot object can be carried out
with an optical pyrometer working at a particular wavelength, λ. Often optical filters
in the pyrometer are used to define λ. Planck’s law for a perfect blackbody emitting
radiation energy, Wb(λ, T ), is

Wb ðk; T Þ ¼ 2phc2 =k5 ½expðhc=kkT Þ 1 ð36:1Þ
where T is the true temperature of the blackbody, c = speed of light,

k = Boltzmann’s constant, and h = Planck’s constant. Since in (36.1), hc/λkT ≫ 1,

Wb ðk; T Þ ¼ 2phc2 =k5 expðhc=kkT Þ ð36:2Þ
When an object is not a blackbody, the surface of the object is a partial inward
reflector of the radiant energy being emitted from the object, and the energy
emitted, Wn(λ, T ), is
Wn ðk; T Þ ¼ Wb ðk; T Þ eðk; T Þ ð36:3Þ
where ε(λ, T ) is the emissivity, a number less than unity.

In pyrometers of the disappearing-filament type, a tungsten filament, acting as a
black-body, is compared with that of the hot sample by observing a superposition of
the hot filament on the hot sample and then finding the pyrometer filament tem-
perature where it becomes invisible against the image of the hot sample. This
method is very sensitive, and a skilled operator, working in a darkened room with
dark-adapted eyes, can make comparisons to within a few degrees K. Figure 36.3a
shows the image seen through the eyepiece of a disappearing filament pyrometer.
The equal-temperature point should be approached alternately from the
low-temperature side and from the high-temperature side, and the readings from
both directions should be averaged to determine the brightness temperature of the

sample.
In the case where the filament and sample brightness temperatures are identical,
then

Wb k; Tpyro ¼ Wb k; Tsample eðk; T Þ ð1 RÞ2 ð36:4Þ
where R is the reflectivity of each surface of the window material through which the
measurement is being made. R is given by
R ¼ ðn 1Þ2 =ðn þ 1Þ2 ð36:5Þ
where n = the refractive index of the window material at λ.

Using (36.2) and (36.3), and solving for Tpyro, one obtains
.n h io
Tpyro ¼ Tsample 1 kkTsample =hc ln eðl RÞ2 ð36:6Þ
Both the emissivity, ε, and the window surface reflectivity, R, may be deter-
mined as correction factors for Tsample, as shown below.
For the emissivity, an experimentally determined number that is a specific
property of the material being measured must be used. Tabulations of the emis-
sivities for various metals are given in [6], as well as convenient correction plots to
directly convert the brightness temperature to the true temperature for various
emissivities.
For the particular case of Si at λ = 0.65 μm, the emissivity was measured
separately by Allen [7] and by Sato [8]. Their data are fit by the following equation
[9] in the temperature range 540–1700 K.

eð0:65 lm; T Þ ¼ 0:4826 2:09 105 T þ 0:1583 f1 þ exp½ðT 1410Þ=112:9g
ð36:7Þ
Figure 36.3b shows the fit of the empirical (36.7) to the data. The relationship
between Tpyro and Tsilicon can be calculated using (36.6) with values of R and
ε(0.65 μm, T) calculated from (36.5) and (36.7), respectively. In the case of glass
windows, the refractive index for 0.65 μm light is 1.46, and R = 0.035.
Figure 36.3c shows the deviation between the measured brightness temperature
with the pyrometer at 0.65 μm, and the true temperature for Si, viewing through a
glass window.
Similar considerations apply for pyrometric measurements made by electronic
methods in the infrared region, as is common for studies of semiconductor mate-
rials, except that a specific emissivity will apply for the particular material and the
wavelength of light employed.
36.4 Pyrometer Measurements Through Windows Coated … 411
Fig. 36.3 Pyrometric measurement of temperatures. a Disappearing filament pyrometer—visual.

b Si emissivity versus temperature. c Emissivity corrections to brightness temperature
36.4 Pyrometer Measurements Through Windows Coated

Inadvertently with Variable-Thickness Thin Films
The deposition of a thin film on an optical window can cause a negative error in
optical pyrometer readings, which will increase with the thickness of the deposited
film. In order to avoid this effect, an empirical method has been developed and
tested against thermocouple readings on a hot sample [1].
Figure 36.4a shows the basic principle of the method. A movable ionization
gauge having an emission regulated thermionic emitter is used as a standard. The
temperature of the emitter should remain constant in good vacuum for a standard
constant emission current that is set into the controller. When the window is first
installed, and when it is free of any thin film deposit, the brightness temperature of
the thermionic emitter is measured as a function of the emissivity setting on the
Fig. 36.4 Use of standard filament to calibrate pyrometer through film-coated windows. a Ion
Gauge pyrometry standard. b Typical temperature versus emissivity setting for various window
film deposits
pyrometer, and a curve like that of Fig. 36.4b (clear) is constructed. As the window
becomes coated with an inadvertent thin film deposit, the brightness temperature of
the thermionic emitter filament versus the emissivity setting will decrease, as shown
for various time periods in Fig. 36.4b after the thin film has grown to various
thicknesses. For any temperature, an emissivity setting returning the filament
brightness temperature to the same value can be determined, as shown for example
36.4 Pyrometer Measurements Through Windows Coated … 413
for a brightness temperature of 973 K (dotted line), where emissivity settings of 0.6,
0.5, and 0.3 are needed to bring the pyrometer back to the standard temperature of
973 K. This same emissivity setting is then used when the sample temperature is
being studied in the position of the thermionic emitter.
The pyrometer must, of course, be calibrated accurately in the absence of any
thin film deposit. Both thermocouple and phase transformation information may be
used for this purpose, or in the case of materials with known emissivities, tables of
emissivity versus temperature may be used to make the correction from brightness
temperature to true temperature. If the sample temperature is considerably lower
than the ion gauge filament temperature, it will be necessary to employ an inter-
posed neutral density filter for the thermionic emitter measurements, as shown in
Fig. 36.4a. Because the thermionic emitter operates at about 1500 K, a neutral
density filter is needed if measurements of sample temperatures considerably below
this temperature are needed. In the example given, a neutral filter of optical density
2.0 (10−2 transmission) is needed to reduce the apparent filament temperature to
about 1000 K for an emissivity setting of 0.6.
Separate calibrations using a clean window that did not become coated with the
thin film verified that this procedure was accurate and that the thermionic emitter
temperature was constant over a long time [1]. In the case shown here, an IRCON
Model 2000 pyrometer, working in the range 0.7 mm–1.05 μm, was employed [10].
36.5 Calibration Point for Optical Pyrometer

Used for Silicon
The provision of a calibration point for silicon at high temperatures for optical
pyrometers may be useful in certain situations. A method developed for doing this
in rapid wafer thermal processors [11] can be extended to other situations in the
laboratory.
In Fig. 36.5, a silicon wafer, 0.3 mm thick, and covered by a 5000-Å-thick SiO2
film, is coated with a 4000-Å Ge film, and then overcoated with a protective 1000-
Å Si3N4 film. This sandwich structure serves as a useful calibration standard. The
optical reflectivity of Ge at 0.6328 μm changes from 49 % just before semicon-
ducting Ge melts to 72 % for melted Ge, which behaves with metallic properties.
The melting point of Ge is 938 °C, and melting can be viewed through the Si3N4
overlayer. As shown in Fig. 36.5a, the sandwich test structure contains a K type
thermocouple embedded in the backside of the silicon wafer, and an optical
pyrometer to be calibrated at the melting point of Ge. A chopped laser beam is
reflected off of the Ge sandwich face, and the reflectivity of the surface is measured
as the temperature is raised, using external radiation for heating the wafer.
Figure 36.5b shows the result. At a thermocouple temperature of 938 °C, a sharp
Fig. 36.5 Pyrometer calibration at fixed point for Silicon. a Ge/Si test structure. b Reflectivity
changes at 0.6328 μm
36.5 Calibration Point for Optical Pyrometer Used for Silicon 415
increase in the reflectivity is observed, followed by a gradual decrease as the

temperature continues to rise. On slowly cooling the wafer, the freezing point of the
Ge is observed to be depressed to about 904 °C as a result of supercooling of the
liquid Ge. The melting point of Ge was found to be reproducible to ±1° at 938 °C in
ten measurements. It was shown that the thermal gradient through the Si crystal was
about 2.5 °C/mm at the melting point of the Ge. These Ge sandwich structures can
therefore be used periodically to calibrate the optical pyrometer emissivity setting
(see p. 409).
36.6 Control of Silicon Temperature Using Resistivity
Both thermocouples and pyrometers are used to measure and control the temper-
ature of silicon crystals during electrical heating. As discussed elsewhere (p. 548),
the use of thermocouples for this purpose often produces metal contamination of the
silicon crystal. A method is described here involving the measurement of the
electrical resistance of the crystal during heating and cooling in order to determine
and control the temperature. For purposes of evaluation of the resistance method,
the crystal used in these measurements contains a W/26 % Re:W/5 % Re ther-
mocouple which has a ball end buried inside a laser-drilled hole in the crystal and
held in place with a MgO cement (Aremco 571P) [12].
The crystal is mounted on a Ta holder as shown in Fig. 34.5a, p. 358. The crystal
dimensions are 9.5 × 3 × 0.3 mm3. The silicon is P type, 100 Ω cm, boron doped.
The resistance of the crystal was calculated using voltage and current measurements
made during heating. Comparisons to the thermocouple and to an optical and an IR
pyrometer are described in the original paper [12]. Figure 36.6a shows the
dependence of the logarithm of the measured resistance versus 1/T. Above about
600 K, the expected linear relationship is observed, where the dominant intrinsic
conductivity of the silicon exists. Below 600 K, the dominating conductivity is due
to the dopants for this particular crystal. The slope in the linear region corresponds
to the quantity EG/2kBT, and the best value of EG (the silicon band gap) for this
crystal was measured to be in the range 1.22 to 1.25 ± 0.05 eV, in good agreement
with literature values of 1.170–1.205 eV [13–15].
Figure 36.6b shows the type of temperature control that may be achieved using
the resistance and comparing to the temperature measured directly with the ther-
mocouple. A programmed temperature ramp of 3 K/s, interrupted by a constant
temperature period, and then a cooling ramp of −3 K/s are shown. The solid line
represents the temperature measured with the thermocouple, and the dotted line
with the silicon resistance. Good agreement is found with a maximum deviation of
20 at 1150 K. Control was carried out using LabVIEW [16], which is described on
p. 351.
Fig. 36.6 Control of Silicon

temperature using resistivity.
a Dependence of Log R
versus inverse temperature of
Si(100) crystal.
b Temperature programmed
heating and cooling of Si
(100)
The method must be calibrated against a standard at one or more points. This
may be done best using a pyrometer and comparing the resistance with the cor-
rected pyrometer reading (see p. 409). It would be better to do the calibration
against the pyrometer at several points in the intrinsic region in order to make the
best calibration, although the knowledge of the bandgap permits single point cal-
ibration [12].
References 417
References
1. The Omega Complete Temperature Measurement Handbook and Encyclopedia (Omega

Engineering, Inc., Stamford, CT, 1989), pp. z8–z28
2. D.R. Sandstrom, S.P. Withrow, J. Vac. Sci. Technol. 14, 748 (1977)
3. W.J. Mitchell, J. Xie, W.H. Weinberg, J. Vac. Sci. Technol. A11, 3133 (1993)
4. V.S. Smentkowski, J.T. Yates Jr, J. Vac. Sci. Technol. A14, 260 (1996)
5. R.R. Asamoto, P.E. Novak, Rev. Sci. Instrum. 38, 1047 (1967)
6. C.J. Smithells, E.A. Brandes (eds.), Metals Reference Book, 7th edn (Butterworths, London
and Boston, 1991)
7. E.G. Allen, J. Appl. Phys. 28, 1510 (1957)
8. T. Sato, Jap. J. Appl. Phys. 6, 339 (1967)
9. H. Nishino, W. Yang, Z. Dohnalek, V.A. Ukraintsev, W.J. Choyke, J.T. Yates Jr, J. Vac. Sci.
Technol. A15, 182 (1997)
10. S.L. Wright, R.F. Marks, W.I. Wang, J. Vac. Sci. Technol. B4, 505 (1986)
11. J.-M. Dilhac, C. Ganibal, N. Nolhier, B. Rousset, Rev. Sci. Instrum. 63, 188 (1992)
12. Z. Dohnálek, S. Mezhenny, I. Lyubinetsky, W.J. Choyke, J.T. Yates Jr, J. Vac. Sci. Technol.
A15, 2766 (1997)
13. W. Bludau, J. Appl. Phys. 45, 1846 (1974)
14. K.L. Shaklee, R.E. Nahory, Phys. Rev. Lett. 24, 942 (1970)
15. H.D. Barber, Solid State Electron. 10, 1039 (1967)
16. LabVIEW is sold by Analog Devices, One Technology Way, Norwood, MA 02062-9106
Part V
Delivery of Adsorbates to Surfaces
Chapter 37
Gases
37.1 Design and Performance of Microcapillary

Array Beam Doser
Beam dosers for quantitative exposure of surfaces to gases have become widely
used in surface science research. This is because of the disadvantages in filling an
ultrahigh vacuum system with gas in order to study adsorption processes. These
disadvantages relate to the large gas loads needed for system dosing as compared to
beam dosing. Such large gas loads result in adsorption on the walls of the vacuum
chamber, which then contributes to slow achievement of base pressure after
adsorption by the sample surface is complete. In addition, large gas loads in a
vacuum system can cause displacement effects from the system walls leading to
impurity adsorption, as well as regurgitation effects from ion pumps in which
impurity gases are produced (see p. 72). Also, the use of quantitative and absolute
calibration procedures for the gas flux coming from a collimated doser is much
superior to dosing with a system dose using an ionization gauge or a mass spec-
trometer, both of which have serious problems with absolute calibration.
Figure 37.1a shows a design for a microcapillary array beam doser [1].
A pinhole aperture made of stainless steel serves as a deformable gasket inside a
Cajon vacuum fitting [2]. Typical aperture diameters are about 5 μm for convenient
dosing fluxes for adsorption studies, as shown on p. 425, where calibration methods
are described. The flux of gas issuing from the pinhole aperture travels into the
doser body, where it strikes a baffle made of metal shim stock, held into place by
springy arms that bend against the inside wall of the doser body. This baffle
randomizes the gas molecules’ motion inside the doser body and behind the mi-
crocapillary collimator plate. Collimation of the gas occurs in the high-density
capillary array where parallel capillaries of 10 μm diameter and 500–1000 μm
length are arranged in a hexagonal close packing arrangement [3]. The microcap-
illary array disk has a radius, RS and should be purchased with the capillary axes
perpendicular to the array surface, rather than inclined as in the chevron design

DOI 10.1007/978-3-319-17668-0_37
422 37 Gases
Fig. 37.1 Design and performance of microcapillary array doser, a Design of doser (R = radius),
b Fractional interception by target
commonly manufactured [3]. The capillary array is held in place by a metal cap
that, with a clearance of several micrometer, confines the edge of the array in a
close-fitting groove on the end of the doser body. Care should be taken not to
compress mechanically the very fragile glass capillary array during installation. The
capillary array has a relatively high conductance for gases and acts mainly as a
collimator in the gas doser.
Figure 37.1b shows the intercepted flux fraction from the array by the target of
radius RT for various geometries [4, 23]. The most uniform dosing over the target
will take place with larger values of RS /RT but of course the fraction of the beam
intercepted will decrease as shown in Fig. 37.1b. The assumptions used in deriving
the functions shown in Fig. 37.1b assume that diffuse scattering occurs on the inside
walls of the capillaries under molecular flow conditions where the mean free path is
considerably longer than the diameter of the capillaries.
If the capillary array is too close to the target, reflected molecules from the target
will have a finite probability of incidence on the crystal for a second encounter
when they are reflected back by the capillary array surface.
37.1 Design and Performance of Microcapillary … 423
There are precautionary measures to be considered with glass microcapillary

array dosers. They are made of soft glass and therefore contain Si–OH groups on
their surface. The hydrogen atoms are labile and will exchange with –OD groups in
D2O or in organic molecules such as CH3OD [5]. In this case, isotopic exchange
overnight can be carried out until the isotopic purity of the transmitted gas is not
influenced by isotopic exchange on the inner walls of the capillaries. In addition,
reactions of some molecules such as fluorinated ethers with acid sites on the silica
surface have been reported [6]. Finally, certain unstable molecules such as Fe(CO)5
have been shown to be reactive with the glass surface [7].
Recently, less finely packed array dosers made of metal have been used with
success for uniform gas dosing in directions off the crystal normal [8].
For reactive gases, the use of an inert diluent gas has been described in [9] and
also on p. 89.
Historically, collimated array gas dosers have been used since the 1970s, and
some examples are given in [10–14].
A recent theoretical treatment of the transport of gases through microcapillary
array dosers and other tubular devices is found in [15], which also contains a
bibliography of earlier papers on this subject. A recent analysis of the flux distri-
bution from a tube doser and a capillary array doser has been reported [16].
37.2 Gas Handling System for Array Beam Doser
The effective use of the microcapillary array beam doser shown on p. 421 requires a
gas-handling system of appropriate design, as shown in Fig. 37.2 [17]. A stainless
steel system, pumped by a turbopump, is designed to attach to several glass storage
gas bulbs, using Cajon VCR fittings [17].
The storage of gases in glass bulbs at pressures near 1 atm prevents accidental
over-pressurizing of the gas-handling system by compressed gases in cylinders.
Also, during gas transfer to prebaked storage bulbs (for removal of adsorbed H2O
from the glass walls), one has the opportunity to use purification steps for gases
such as the freeze-pump-thaw method commonly used to remove noncondensable
impurities.
For calibration of the gas doser system, a test gas, such as N2, is admitted to the
gas-handling system to a few Torr pressure. The ballast bulb, of exactly known
volume, is also filled with this gas. As will be shown on p. 425, the conductance of
the pinhole aperture may be measured by measuring the time rate of change of
pressure in the gas-handling system as the N2 is slowly pumped through the pinhole
aperture. To interrupt the flow from the array doser, Valve-1 is closed, and the
region behind the doser is rapidly evacuated through Valve-2, keeping the gate
valve to the UHV system closed.
424 37 Gases
Fig. 37.2 Typical gas handling system—beam doser
Stainless steel systems such as this can be baked with heating tape to 150–200 °C
while pumping with the turbopump. Because of the low pumping speed through the
narrow tubing and the relatively high internal surface area compared to the volume,
limiting pressures in the 10−8 Torr range are usually measured. For gases added to
the Torr level, this background gas represents much less than 1 ppm of added
impurity, which is usually superior to the initial purity of the gas being studied.
The internal surface of the stainless steel gas-handling system has a significant
capacity to react with active gases such as Cl2. Gases of this sort may often be used
themselves to passify the stainless walls. This is done by storing the active gas at
pressures of several Torr pressure overnight, and then pumping the gas away before
adding fresh gas to the system [18]. It may be necessary to do this passivation treatment
in several cycles. In other cases, treatment of the gas-handling system with oxygen gas
to passivate the inner walls by oxidation has been found to be effective [19].
37.3 Calibration and Use of Array Beam Doser 425
37.3 Calibration and Use of Array Beam Doser
Figure 37.3a shows a typical design for an ultrahigh vacuum system equipped to
measure the kinetics of adsorption of a gas on an atomically clean crystal surface.
A quadrupole mass spectrometer (QMS detector), shielded to receive only a random
flux of background gases, and the microcapillary array doser are the principal
components of the apparatus for this measurement.
The calibration of the doser was performed with N2 over a period of about 105 s,
as shown in Fig. 37.3b. The data, plotted logarithmically, against time, t, provide a
linear curve whose slope is related to the conductance of the pinhole by (37.1).
dn=Pdt ¼ V=kT ½d ðInP=P0 Þ=dt ð37:1Þ
where n number of molecules, P = backing pressure, V = storage volume,

k = Boltzmann’s constant, and T = temperature in Kelvin. For the most accurate
measurement of the drop in pressure, the temperature-sensitive capacitance
manometer in the gas storage region (see p. 423) must be thermally shielded to
avoid temperature changes larger than a few tenths of a Kelvin.
Usually measurements of this type have a relative accuracy of better than 5 %
when gases of various molecular weights are compared to each other in the doser
Fig. 37.3 Use of beam doser for kinetics and coverage measurements. a Typical apparatus—
absolute coverage and sticking probability measurement, b Typical calibration curve—pinhole
conductance. c Adsorption uptake curve. d Sticking probability versus absolute coverage
426 37 Gases
[20]. For molecular flow conditions, where the mean free path of the gas on the
high-pressure side of the pinhole is larger than the diameter of the pinhole, the rate
of effusion through the pinhole will be proportional to M−1/2, where M = the
molecular mass of the gas. This condition continues to be approximately true as one
enters the transition region in gas density, at higher pressures, where viscous flow is
beginning to occur in the pinhole aperture which has finite wall length.
Figure 37.3c shows measurements of the adsorption of O2 on a Fe(110) surface at
90 K [21]. For this measurement, the shutter is placed between the doser and the
crystal, and a steady base pressure of the order of 1 × 10−10 Torr or less is estab-
lished, using the mass spectrometer tuned in this case to 32 amu. After steady
conditions are achieved, the doser is turned on with the shutter still in place and the
pressure measured by the mass spectrometer rises. Ideally the pressure should level
out prior to opening the shutter to allow adsorption from the beam to occur, but since
a small flux of random gas scattered from the back of the shutter is incident on the
crystal, causing a slow increase in coverage, the waiting time is limited and a
compromise must sometimes be reached in which “almost steady state” conditions
for Pback are established, as shown here. At this point, the shutter is opened, and
adsorption occurs until asymptotic behavior of the scattered flux of gas is observed.
Then, when the sticking probability has fallen to a low value, placing the shutter into
the beam causes no discontinuity to be observed in the mass spectrometer signal.
The data shown in Fig. 37.3c can be converted into sticking coefficient vs.
absolute coverage, as shown in Fig. 37.3d, if the fractional interception of the doser
beam is calculated (see p. 421) (or measured [20]). It is of interest to note that the
entire measurement of sticking probability and absolute coverage are determined
solely by the accuracy of the flux calibration and the accuracy of the fractional
interception of the collimated gas beam by the crystal [22, 23]. The mass spec-
trometer does not have to be calibrated in any absolute sense.
This method is based upon earlier work with variations on the features of the
apparatus and the measuring conditions [24–27].
37.4 Sticking Coefficient and Surface Reactivity
Measurements of the sticking coefficient and similar measurements of surface

reactivity, all based on the original work of King and Wells[28], form a central core
of modern studies of adsorption and reactivity on single crystals. This work is
discussed in various other chapters of this book [see Sects. 24.7, 25.5, 37.1–37.3, 37.
5]. The apparatus [29], shown in Fig. 37.4a, is very similar to that in Fig. 37.3.
A collimated molecular beam of adsorbate is incident on a single crystal (CO + Pt
(111)) maintained at constant temperature. A fraction of the incident beam is
adsorbed with a particular coverage-dependent sticking coefficient and the remainder
is scattered into the chamber, returning to the mass spectrometer after having been
randomized in scattering direction by collisions with the chamber walls. By knowing
the fractional interception of the crystal (see Sect. 37.1), the measurements permit
37.4 Sticking Coefficient and Surface Reactivity 427
Fig. 37.4 Sticking coefficient and surface reactivity. a Sticking coefficient and surface reaction
apparatus. b Typical uptake kinetics data
both the sticking coefficient to be measured as a function of absolute coverage and

the absolute coverage to be measured as discussed in Sect. 37.3. Figure 37.4b shows
the sticking coeffficient behavior of CO on Pt(111)at 300 K [2].
428 37 Gases
For a molecule which decomposes into fragments producing desorbing species,

it is also possible with this apparatus to measure something of the chemical pro-
cesses at work during adsorption under isothermal conditions. This is called
scanning kinetic spectroscopy (SKS) and is shown in Sect. 25.5. The present work
[28] also employs this method to observe the reaction of C2H4 on Pt(111), where
ethane desorption and hydrogen desorption accompany C2H4 adsorption after a
short delay time.
37.5 Beam Doser Shutoff—Condensable Gases
Very simple measures can sometimes be used to interrupt the flow of gases from a
gas doser. As indicated on p. 430, the use of a small deadspace valve of special
design is helpful in defining reproducible and absolute doses of gases into a colli-
mated gas doser system. In some cases, where gases condensable in liquid nitrogen
are being used with a gas doser, it is possible to interrupt the flow rapidly by using a
cold trap as a pump. In Fig. 37.5, a cold trap made of Pyrex glass is immersed in a
Dewar flask containing liquid nitrogen [30]. This cold trap is a very rapid pump for
all gases that are condensable at 77 K. Flowing gas through the doser (see p. 423)
may be turned off rapidly by closing the admission valve and then quickly opening
Fig. 37.5 Beam shutoff—condensable gases

37.5 Beam Doser Shutoff—Condensable Gases 429
the dump valve. This lowers the gas pressure very rapidly behind the conductance
limiting aperture, and turns the doser off. Many doses of gas, and different con-
densable gases, may be trapped without cleaning the collector tube. For cleaning,
the tube is disconnected from the doser line using the VCR fitting, cleaned with
solvents, baked briefly in an oven, reconnected, and evacuated using the pump in
the gas-handling system. This method is actually preferable to pumping corrosive
gases from the trap, after each dose, through a turbopump because of possible
failure of the turbopump after exposure over time to corrosive conditions.
37.6 Neutral Beam Intensity Measurement Using

a Stagnation Detector
A stagnation detector may be used to determine the absolute intensity of a

molecular beam. In Fig. 37.6, a portion of the beam (having a diameter larger than
the orifice), enters a chamber containing an ionization gauge and a calibrated leak.
After suitable corrections, the signal indicated by the ionization gauge can be
related to the flux of gas in the beam [31].
Under conditions where the beam is shut off, the steady-state pressure, P, is
Ps ¼ Qo =Corf ð37:2Þ
where Qo is the net outgassing rate and Corf is the conductance of the orifice.
When the valve to the calibrated leak is opened, the pressure increase will be
DPleak ¼ QL =Corf ð37:3Þ
where QL is the calibrated leak rate. The conductance of the orifice, which should
be sharp-edged is, under molecular flow conditions
Corf ¼ 2:86ðT=M Þ1=2 D2 L=s ð37:4Þ
where D is the orifice diameter in cm, T is the absolute temperature, and M is the
molecular mass of the gas.
When the molecular beam passes through the calibrated orifice it will cause a
pressure rise in the stagnation chamber given by
DPbeam ¼ QB =Corf ð37:5Þ
The impingement rate, IB, is related to ΔPbeam according to (37.6),
IB ¼ QB NL =pD2 ¼ 0:91 NL ðT=M Þ1=2 DPbeam ð37:6Þ
where NL = 3.237 × 1019 molecules/Torr-L at 298.15 K.

430 37 Gases
Fig. 37.6 Stagnation detector
It is important that the beam diameter be larger than the orifice and that the baffle
inside the stagnation detector be arranged to prevent direct entry of the gas from the
calibrated leak into the ionization gauge. For reactive gases such as oxygen, the
operating temperature of the ionization gauge filament and the emission current
should be reduced as much as possible to prevent pumping effects.
The intensity at a sample position that differs from the orifice position is related
to the intensity at the orifice by the inverse square law for the distances involved.
A good article detailing the many experimental aspects of the use of supersonic
molecular beams in surface chemistry studies may be found in [32].
37.7 Low Deadspace Aperture Valve for Repetitive Gas

Dosing
The use of gas dosers for adsorption experiments is highly recommended in order to
avoid impurity effects from wall displacement and ion pump regurgitation processes
as well as to have quantitative gas dosing capabilities, as discussed on pp. 421–428.
37.7 Low Deadspace Aperture Valve for Repetitive Gas Dosing 431
In the conventional gas dosing design, a rather large volume of gas is held in a
storage section of the gas line, and to interrupt dosing, one must pump out this gas
or some fraction of it held in a section of the gas storage line. For repetitive gas
dosing, this is inconvenient, and in addition for expensive gases it is wasteful. The
design shown here [33] avoids this effect by minimizing the deadspace volume so
that only a few seconds of gas dosing occurs following closure of the valve. In
addition, when a conductance-limiting aperture at the outlet to the small deadspace
volume is used, quantitative gas dosing can be done when the valve is opened.
Figure 37.7a shows a view of the aperture valve assembly. The valve body of a
stainless steel diaphragm ultrahigh vacuum valve (Nupro DL Series) is drilled and
Fig. 37.7 Low deadspace aperture dosing valve. a Aperture valve assembly. b Pumping behavior—
40 μm aperture. c Duty cycle—open and close
432 37 Gases
tapped with 1/16 NPT pipe threads to receive a stainless steel rod also threaded on the
upper end. This is done at the correct angle inside the inlet side of the valve. This rod
has a central hole, 2 mm in width, bored its length, except for an 0.003-in.-thick
membrane at the inner end. The membrane is then laser-drilled with a nominal
60-μm-diameter conical hole [34]. The upper end of the insert contains a screwdriver
slot for convenient insertion into the valve; Teflon tape is used to make a gas tight seal.
The inlet to the valved assembly (originally the outlet) is connected to a conventional
gas-handling system, containing pumping capabilities, a capacitance manometer, and
a ballast volume, V, used for conductance calibration of the laser-drilled aperture. The
outlet of the valve assembly is connected to a small flange that connects to a colli-
mating gas doser (see p. 421) inside the ultrahigh vacuum system. With this
arrangement, the deadspace when the valve is closed is about 0.2 cm3. Thus, when the
dosing is completed, and the valve is closed, evacuation of the deadspace into the
ultrahigh vacuum system by way of the doser does not involve the flow of large
amounts of gas, even when the backing pressure is near 0.1 Torr. Under these con-
ditions, the number of molecules in the deadspace is about 4 × 1014. If 20 % of these
molecules strike the crystal and the sticking coefficient is 0.1, about 2 % of a
monolayer will adsorb from the exhaust of the deadspace after the valve is closed.
Figure 37.7b shows a comparison of the pumping speed through the opened
valve for both Ar and H2, and may be used to measure the conductance of the
laser-drilled aperture. It is noted that the effusion rate follows a semilogarithmic
behavior from backing pressures of 10 Torr and lower, indicating that slight
departures from molecular flow conditions do not sensibly disturb the conductance
of the aperture from that predicted by the kinetic theory of gases for molecular flow.
The superposition of points for the two gases on the curve gives further confidence
in the notion that the behavior can be predicted from the molecular flow equations.
Figure 37.7c shows the response of a mass spectrometer in the vacuum system as
the valve is opened and closed for one cycle. A total rise time of about 3 s is seen as
the valve is opened. Constant Ar flux occurs beyond this time until the valve is
closed, where the pumping out of the deadspace volume exhibits a time constant of
1.3 s. Even though these on-and-off periods give relatively slow response times, the
doser may still be calibrated to give reproducible doses containing the on-off
periods as well as the constant flux period. The use of a smaller laser-drilled
aperture is possible, but the time constant for pumping will increase to unacceptable
levels. A major feature of the apparatus is that the unit can deliver rather large
fluxes at fairly low backing pressures and that it can be quickly cut off without
wasting gas. Multiple exposures that are reproducible can be achieved.
A variation on this idea, involving a trapping volume of about 0.4 cm3, has been
employed to dump known quantities of gas onto a crystal by way of an effusive
doser [35–37]. The trapping volume contains about 1.4 × 1016 molecules/Torr; and
the repeatability of the dose, using a capacitance manometer for pressure mea-
surements in the gas line feeding the doser volume, is estimated to be 0.5 %
standard deviation.
37.8 Gas Flow Regulation—Squeeze Valve 433
37.8 Gas Flow Regulation—Squeeze Valve
A very simple method for regulating the flow of gases into an ultrahigh vacuum
systern with a minimum of contaminants involves the use of a brass squeezing
device to restrict the flow through a thick-walled copper tube, as shown in
Fig. 37.8a [38]. The copper tubing (3 mm OD × 1 mm ID) is annealed in a flame,
then cleaned internally with nitric acid and rinsed with water and alcohol. The brass
squeeze device is clamped to the tube, and the flow rate may be adjusted over a
Fig. 37.8 Gas flow regulation—squeeze valve. a Squeeze valve. b Gas-handling system
434 37 Gases
wide range using the tightening screws. This action is reversible tens of times. The
valve cannot be closed completely. To adjust the flow rate after the valve is set, the
pressure in the high-pressure region is varied using the pressure gauge for mea-
surement. To shut down the flow, the upper UHV valve is closed and the region
between this valve and the squeeze valve is evacuated to the pump, as may be seen
in Fig. 37.8b.
37.9 Retractable Beam Doser with Convenient

On-Off Control
A convenient molecular beam doser which turns itself on or off upon retraction is
shown in Fig. 37.9. A leak valve or a calibrated leak can be used to control gas
flow. It is convenient to turn these dosers on or off by quickly pumping out the gas
supply to the doser. In this case, a translational manipulator is used to move the
doser in and out. When in the in position, the 4 holes in the doser tube between the
Fig. 37.9 Retractable molecular beam

37.9 Retractable Beam Doser with Convenient … 435
two O-ring seals are isolated from differential pumping and the doser delivers gas at
the established flow rate. When the doser is in the out position, the four pumping
holes are exposed to differential pumping, turning the doser off. The O-rings are
cemented into grooves machined in the on-off assembly. They slide into a polished
cylindrical region with a slight taper on the left side for easy admission. The doser
is guided by a Cu guide with a loose-fitting alignment hole. The doser contains a
microchannel plate for gas collimation and a baffle to prevent beaming of the gas to
the plate from the delivery tube, spoiling the collimation [39].
References
1. M. Bozack, L. Muehlhoff, J.N. Russell Jr, W.J. Choyke, J.T. Yates Jr, J. Vac. Sci. Technol.
A5, 1 (1987)
2. P.L. Hagans, B.M. DeKoven, J.L. Womack, J. Vac. Sci. Technol. A7, 3375 (1989).
Laser-drilled micron-size pinholes are produced in disks fitting VCR fittings by Precision
Laser Services, 314 E. Wallace St., Ft. Wayne, IN 46803
3. Capillary array blanks, made without being coated with active surfaces for use as electron
multipliers, are available from Galileo Electro-Optics Corporation, Galileo Park, P.O. Box
550, Sturbridge, MA 01518. Sizes up to 25-mm diameter are available
4. A. Winkler, J.T. Yates, Jr., J. Vac. Sci. and Technol. A6, 2929 (1988)
6. P.K. Leavitt, P.A. Thiel, J. Vac. Sci. Technol. A8, 148 (1990)
7. M. Henderson, R.D. Ramsier, J.T. Yates Jr, J. Vac. Sci. Technol. A9, 2785 (1991)
8. A.M. Glines, R.N. Carter, A.B. Anton, Rev. Sci. Instrum. 63, 1826 (1992)
9. E.C. Henn, M.E. Bussell, C.T. Campbell, J. Vac. Sci. Technol. A9, 10 (1991)
10. D.W. Goodman, T.E. Madey, M. Ono, J.T. Yates Jr, J. Catalysis 50, 279 (1977)
11. D.W. Goodman, J.T. Yates Jr, T.E. Madey, Chem. Phys. Lett. 53, 479 (1978)
13. D.E. Ibbotson, T.S. Wittrig, W.H. Weinberg, Surf. Sci. 110, 294 (1981)
14. J.E. Deffeyes, A.H. Smith, P.C. Stair, Surf. Sci. 163, 79 (1985)
15. D.M. Murphy, J. Vac. Sci. Technol. A7, 3075 (1989)
16. D.E. Kuhl, R.G. Tobin, Rev. Sci. Instrum. 66, 3016 (1995)
17. M.J. Bozack, L. Muehlhoff, J.N. Russell Jr, W.J. Choyke, J.T. Yates Jr, J. Vac. Sci. Technol.
A5, 1 (1987)
18. A.L. Linsebigler, V.S. Smentkowski, M.D. Ellison, J.T. Yates Jr, J. Am. Chem. Soc. 114, 465
(1992)
19. M.A. Henderson, R.D. Ramsier, J.T. Yates Jr, J. Vac. Sci. Technol. A9, 2785 (1991)
20. C.C. Cheng, R.M. Wallace, P.A. Taylor, W.J. Choyke, J.T. Yates Jr, J. Appl. Phys. 67, 3693
(1990)
21. V.S. Smentkowski, J.T. Yates, Jr., J. Vac. Sci. Technol. A7, 3325 (1989); Surf. Sci. 232, 113
(1990)
23. C.T. Campbell, S.M. Valone, J. Vac. Sci. Technol. 13, 408 (1985)
24. D.A. King, M.B. Wells, Surf. Sci. 29, 454 (1972)
25. T.E. Madey, Surf. Sci. 33, 355 (1972)
26. A.A. Bell, R. Gomer, J. Chem. Phys. 44, 1065 (1966)
27. C. Wang, R. Gomer, Surf. Sci. 84, 329 (1979)
28. D.A. King, M.B. Wells, Surf. Sci. 33, 355 (1972)
29. F. Zaera, Int. Rev. Phys. Chem. 21, 433 (2002)
436 37 Gases
30. Author’s personal experience

31. W.G. Dorfeld, J.B. Hudson, R. Zuhr, Surf. Sci. 57, 460 (1976)
32. S.T. Ceyer, D.J. Gladstone, M. McGonigal, M.T. Schulberg in Investigations of Surfaces and
Interfaces—Part A, Physical Methods of Chemistry Series, 2nd Ed., Vol. IXA, edited by B.W.
Rossiter and R.C. Baetzold (Wiley, New York, 1993)
33. S.M. Cohen, M.P. D’Evelyn, J. Vac. Sci. Technol. A9, 2414 (1991)
34. Lenox Laser, 1 Green Glade Ct., Phoenix, MD 21131
35. Dr. V.A. Ukrainsev, Texas Instruments, P.O. Box 655012, MS 461, Dallas, TX 75265 (private
communication)
36. Y.A. Gelman, B.S. Podol’skii, V.A. Ukraintsev, Pribory I Tekhnika Experimenta 6, 191
(1990)
37. I. Harrison, V.A. Ukraintsev, A.N., in Artsyukhovich, SPIE Proceedings, Laser Techniques for
Surface Science, ed. by H.-L. Dai, S.J. Sibener, vol 2125, p. 285 (1994)
38. G. Scoles, Chemistry Department, Princeton University, Princeton, NJ 08544 (private
communication). This method was devised at the FOM Laboratories in Amsterdam 40 years
ago and is still in use
39. Michael Schmidt and Ulli Diebold, private communication
Chapter 38
Evaporation Sources
38.1 Degassing Evaporation Sources
The deposition of thin films by evaporation constitutes one of the main experi-
mental procedures in surface and materials science. Indeed, the study of chemi-
sorption on clean surfaces has as one of its origins the use of thin metal films as
substrates in early studies done in the 1940s.
A major problem in making clean thin films originates from the outgassing of
heated surfaces during evaporation, leading to contamination of the thin film by
chemisorption from the background vacuum. This may be avoided by careful
design in which prior out-gassing is achieved for components that will be heated
during evaporation. Figure 38.1 illustrates how this is done, using two principles.
The first principle involves the use of degassing loops to support all filaments
that will be heated [1]. In the center of the evaporation shield, the evaporation loop,
containing welded pieces of the evaporant at its apex, is mounted on two degassing
loops. Knowing the current (I1) required to bring the evaporation loop to its desired
temperature for evaporation of the evaporant, one passes current I2 through the
degassing loops prior to the evaporation step. The current I2 should be somewhat
larger than I1/2 so as to bring the degassing loops to a higher temperature than they
will ever attain during evaporation, and the two loops should be outgassed sepa-
rately so as to avoid passing current through the evaporation loop during
outgassing.
The second principle illustrated in Fig. 38.1 is the use of a shield-outgassing
filament for heating the shield to temperatures higher than will be achieved while
evaporation occurs from the evaporation loop [2]. This is achieved by radiation
from the outer shield outgassing filament, where the shield thermocouple may be
used to monitor the temperature achieved during outgassing compared to that
during evaporation. Note that this shield outgassing filament is also mounted on
degassing loops.

DOI 10.1007/978-3-319-17668-0_38
438 38 Evaporation Sources
Fig. 38.1 Degassing evaporation sources
Schematic performance curves are shown in Fig. 38.1 to illustrate that extremely
rapid employment of a properly outgassed evaporation source is possible compared
to a poorly designed source that may outgas over days prior to being useful for
evaporation under good vacuum conditions. The shutter should not be opened until
an appropriate limiting pressure is achieved during evaporation.
Some evaporation sources will be subject to continual outgassing because of
design limitations preventing the use of methods shown here. In that case, reduction
of the unwanted gas fluence at the substrate may be achieved by differentially
pumping the source and the substrate chambers that are connected by an aperture [3].
38.1 Degassing Evaporation Sources 439
The monitoring of the temperature of the evaporation filament maybe achieved

quite simply by the use of current and voltage drop measurements as described on
p. 430.
A helpful word about the construction of welded assemblies involving refractory
wires such as tungsten and tantalum will be useful. When these wires are fabricated
from approximately 0.3-mm-diameter wire for the purposes described here, it is
often found that upon heating to high temperatures, the wires will distort geo-
metrically. This distortion is caused by mechanical strain present initially in the
welded assembly if the welds and the wire are placed under strain during attempts
to make additional welds on the same filament. To avoid this problem, all wires
should be formed by bending over gently curved mandrels, and the welding of the
preformed filaments should be carried out in such a way that no strain is placed on a
weld when subsequent welds are made and that no distortion of the preformed
filament shape occurs during final assembly by welding.
38.2 Metal Evaporation Sources for Downward

Evaporation
Two oven designs for the downward evaporation of metals using radiative heating
have been described that operate at very low outgassing rates [4]. The problem with
most downward evaporation sources that involve melting of the evaporating metal
is that gravity may cause molten metal to fall from the evaporator to the substrate.
The cylindrical oven body is made of tantalum foil, 0.05 mm thick, spot-welded
into shape and containing end caps made of the same size material. The oven is then
wrapped with several layers of 0.01-mm-thick Ta foil to provide substantial thermal
shielding. This heat shielding is essential to permit good thermal efficiency and
therefore low outgassing rates. Two 3-mm-diameter Mo rods are used to bring in
power to a W coil filament, 1 mm in diameter and 250 mm long. The longer of the
two rods is spot-welded to the inside of the Mo oven body to provide mechanical
support for the assembly. The second passes through a hole in the top cap and is
insulated by a thin tube of pyrolytic boron nitride, pBN. When the oven is operating
the resistance of the W coil is about 0.5 Ω. A W25 %Re/W3 %Re thermocouple
monitors the temperature of the top end cap, but does not necessarily read the
correct temperature of the metal source.
Figure 38.2a shows the design for a nonmelting metal (3-mm slab) that is sup-
ported with thin W wires on the bottom of the oven. A W shield, placed on the top of
the source material reduces upward evaporation, extending the life of the oven. This
source has been used for Fe film deposition rates of between 5 and 30 Å/min at
power levels of 90 W and above without melting. The oven-to-substrate distance is
10 cm in this case. Mn has also been evaporated with this design.
Figure 38.2b shows a modified design that permits evaporation of metals that
must be melted to achieve the desired rate of film growth. A set of Ta rings
Fig. 38.2 Downward evaporation source, a nonmelting metal, b melting metal
(0.01 mm thick) alternate with rings of the evaporating metal (0.1 mm thick). The
volume of the metal charge may approach 0.4 cm3. A thicker Ta shield disk
(0.05 mm thick) is placed above the stacked rings to direct the metal deposition
downward and to shield the substrate from the radiation of the hot W filament. For
Cu deposition, a power level of 95 W will give a deposition rate of 10 Å/min, and
145 W gives a rate of about 45 Å/min. The melted metal is held between the thin Ta
rings by surface tension. Silver has also been evaporated successfully with this oven
design. Metals that alloy with Ta may self-destruct the oven.
The ovens are surrounded by tightly coiled water-cooled Cu tubing that does not
touch the oven and that extracts much radiant heat. It is reported that with this
design, the total pressure rise is <5 × 10−11 Torr [4].
An excellent review of metal evaporation sources and materials compatibility at
high temperatures is given in [5].
38.3 Ultralow Coverage Metal Evaporation
The use of a quartz thickness monitor (QTM) for absolute measurement of the
thickness of deposited films has been discussed elsewhere (p. 309). The extension
of this technique to very low coverages of deposited material has been made by
Chin et al. [6]. Essentially they arrange for the rate of evaporation to be determined
by a QTM placed rather close to the evaporation source (7.6 cm), while collecting
38.3 Ultralow Coverage Metal Evaporation 441
Fig. 38.3 Ultralow coverage metal evaporator, a multiple evaporator source, b sources mounted
at long distance
the deposit on the crystal substrate at a much longer distance (dcrystal = 112 cm).
The apparatus shown in Fig. 38.3a consists of a set of multiple evaporation sources
of similar geometry mounted at a distance, dQTM, from the QTM, which is centrally
located. The crystal substrate is mounted in an ultrahigh vacuum chamber facing the
evaporation source, which is located far from the substrate in a long flight tube as
shown in Fig. 38.3b. Considering both the distances and the angles between the
normal to the QTM and the crystal, αQTM and αcryst., the equation for the relative
rates of deposition is
Rcryst: =RQTM ¼ ðdQTM Þ2 cosacryst: =ðdcryst:Þ2 cos aQTM

In the design and testing of this long-path evaporator, the effect of gravity on the
evaporating beam was shown to be negligible as was the effect of beam scattering
by residual gases at ultrahigh vacuum. For the geometry shown, it was calculated
that because of the lower pumping speed at the evaporation source compared to that
at the crystal, the pressure in the evaporator region would be about 40 % greater
than in the ultrahigh vacuum chamber.
In order to obtain a calibration factor for the apparatus, a second QTM was
placed at the crystal position and the ratio of the signals received during evaporation
from each of the individual sources was measured. It was found that each of the
evaporators gave the designed ratio of about 200 within a 20 % error range, whether
the evaporator source material is in the form of a bead or a coil of wire. It was also
shown that shadows cast by the QTM and its support structure were unimportant at
the flight distances employed. Minimum deposition levels of the order of 10−4 ML
are possible with this apparatus. An added advantage of the geometry used is that
heating of the crystal substrate by radiation from the sources is negligible.
An excellent review of metal atom evaporation sources and materials compati-
bility at high temperatures is given in [7].
38.4 Congruent Evaporation of Multicomponent

Materials
The evaporation of alloys and mixtures will almost always result in composition
changes during evaporation because of the differences in vapor pressures of the
different components. Thus, flash evaporation is often employed where the entire
alloy or mixture is rapidly evaporated to yield films of more uniform composition,
equivalent to that of the starting material. Often, the starting material is dispensed as
a powder onto a heated surface where evaporation occurs rapidly. This procedure
has certain deficiencies, such as the need to powder the material, which might
introduce impurities, and the problem with blockage of tubes, etc., with powdered
material. Often, too, the powdered material is mechanically ejected, causing film
inhomogeneities. The device shown in Fig. 38.4 overcomes these problems by use
of a liquid material for the dispenser device [8]. Films of InSb and As/Sb/Se have
been made by a predecessor of the method to be described, and films of doped Se
have been made by the new method [8].
A quartz capillary tube is partially immersed in a molten alloy of the desired
composition, and the liquid is kept slightly above the melting temperature. The
molten alloy is forced through the capillary tube by means of an inert gas with a
pressure of the order of 10 Torr, and this pressure is dynamically regulated by
means of pumping and a gas emission needle valve. The molten source material
streams out of the capillary tube into an evaporator heated to a temperature that
produces rapid evaporation from its inner surfaces, and is above the melting point
of any congruent compound that might form. The evaporator is baffled so that direct
38.4 Congruent Evaporation of Multicomponent Materials 443
Fig. 38.4 Congruent evaporation source
line of sight between the capillary end and the substrate is not present. In this
arrangement, the molten source material can be thoroughly outgassed prior to film
deposition. It is advantageous to heat the top of the capillary and the evaporator to
temperatures where the source material already starts to evaporate before leaving
the capillary.
38.5 Evaporator for Eutectic-Forming Metals
Often, eutectic-forming metals are combined to form evaporation sources for the
higher vapor pressure metal. However, at the temperature of vaporization, a
lower-melting eutectic alloy may form, causing the mechanical failure of the
evaporator device by melting.
A method to avoid this problem has been developed that prevents the formation
of the eutectic in sufficient quantities to cause mechanical failure. In the case of a
zirconium evaporator, this method has proved satisfactory [9]. Figure 38.5a shows
the basic principle of the method. A tungsten rod serves as a mandrel for the
evaporator coil, which consists of smaller-diameter W wire wound together with Zr
wire. Ohmic heating of the assembly by passing current through the large Ta
connector and support plates raises the Zr and W to the correct evaporation tem-
perature (about 2300 K), using 38–40 A, and permits Zr decomposition at a rate of
0.25–0.35 ML/min on the sample located 3 cm from the evaporation source.
A special outgassing procedure was employed for this source [9], which is gassy
given the high operating temperature and the method of construction.
The principle of operation may be seen from considerations of Fig. 38.5b. Here,
a lower melting eutectic mixture of Zr and W is produced at intermediate atomic
fractions of W. In the case of the source described above, one avoids arriving at the
eutectic region for a substantial part of the assembly by limiting the amount of Zr
available compared to the amount of W. Thus, even if the eutectic forms in the
immediate region of a Zr/W interface, diffusion of Zr from this region will of
necessity bring one to the left-hand region of the phase diagram and into the
composition range where extreme depression of the W melting point does not
occur. This principle is often applied in designs for W/Ni and W/Al [10] evapo-
rators where the same kind eutectic formation occurs, causing difficulties.
An excellent article describing metal evaporation methods and materials com-
patibility at high temperatures is available [11].
38.6 Simple Lithium Metal Evaporation Source
The use of alkali metals as adsorbed species and as modifiers of surfaces is well
documented in the literature [12], and several alkali metal evaporation sources are
used, as indicated on p. 483. A particularly nice source for Li has been described
and is presented here [13, 14].
A binary alloy of Al and Li, containing 6.5 % Li, was used in the form of a disk,
10 mm in diameter and 1.5 mm thick [15]. The material was rinsed in acetone
before use and is known to be covered with a thick oxide layer. Two
0.25-mm-diameter Ta wires were looped through two pairs of holes, and a fifth hole
was used for attachment of a type K thermocouple, as shown in Fig. 38.6a.
Although not used by the developers of this Li source [13, 14], the Ta heating wires
38.6 Simple Lithium Metal Evaporation Source 445
Fig. 38.5 Evaporator for eutectic-forming metals, a evaporator, b phase diagram
should be mounted on degassing loops (see p. 407 and 437) for maximum effi-
ciency in outgassing the support. The evaporation source is surrounded by a
stainless steel shield.
Li evolution from the source was monitored with a quadrupole mass spec-
trometer, and the temperature dependence of Li evolution is shown in Fig. 38.6b.
For actual film deposition, a temperature of 695 K was employed, and Li was
Fig. 38.6 Li thin film evaporation source, a Li evaporation source, b Li+ mass spectrometer signal—
T scan rate = 0.5 K/s
collected on an atomically clean Al(111) surface on which work-function and X-ray

photoelectron spectroscopy measurements were made. From these measurements, it
was estimated that approximately 1 ML/min was being deposited with the Al(111)
face approximately 1 cm from the source. A 25-ML layer of Li showed no evidence
for impurities by X-ray photoelectron spectroscopy. These results indicate that even
with the oxide coating, the alloy can supply clean Li by evaporation. The pressure
during deposition increased from 1.5 × 10−10 Torr to 3.5 × 10−10 Torr, and the
evolved gas was mainly H2, CO2, and H2O.
A more complex source involving Li/Al alloys is described in [16], and the
spatial, time-dependent, and temperature-dependent characteristics of this source
have been carefully evaluated.
38.7 Aluminum Evaporation Sources—Long Lived
The evaporation of aluminum in ultrahigh vacuum to yield a clean film requires

special expertise. The most common design for an evaporator involves the use of W
wire as a heater, with Al being evaporated while in contact with the hot W. This is a
short-lived source because of alloy formation, producing a low-melting alloy that
causes mechanical failure of the W wire. A clever method to avoid alloy formation
has been devised, in which a film of Al is oxidized on the W wire to form a barrier
38.7 Aluminum Evaporation Sources—Long Lived 447
that retards alloy effects [17]. A variety of the most recent methods for evaporating
Al have been reviewed in a paper which also presents a different solution to the
lifetime problem [18].
The evaporation source is made of zero-porosity alumina tubing (DIN 40 685
type AL23-710) [19], 1.1 mm inner diameter and 1.7 mm outer diameter, 6 mm
long. Into this was inserted a 6-mm piece of high-purity (99.999 %) Al wire, with a
1-mm diameter. A tungsten coil, made of 0.25-mm wire and wrapped around a
1.37-mm-diameter mandrel prior to use, was fastened tightly around the alumina
tube by virtue of the size difference between the coil and the tube. This assembly
was then spot-welded to two Mo rods of 2-mm diameter. The assembly was rigidly
supported between the Mo rods using ceramic locating blocks, not shown in
Fig. 38.7. The evaporation source was surrounded by a shield that was supported on
the mounting flange, and that contained an 8-mm-diameter hole on the axis of the
device. A rotatable shutter was used in front of the aperture to control the Al
deposition time and to permit outgassing of the source without deposition.
The source was degassed for hours at a 60 % power level compared to the
operating conditions used for evaporation. Typical operating conditions were
achieved at 2.5 V, 1.4 A, and 3.5 W, under which conditions the alumina tube
assembly glowed with a bright orange-yellow color. Final outgassing, with the
shutter closed, was carried out for 30 s at the operating temperature. Operation
could occur for 5–10 min before the pressure rose to 5 × 10−10 Torr. Under these
conditions, deposition occurs at a rate of about 0.5 ML/min. Auger spectroscopy
and X-ray photoelectron spectroscopy both indicated that impurities on the evap-
orated Al film were below the detection limit. The source could be operated
reproducibly for several months before failure. Since Al melts at the operating
temperature, the source should be mounted horizontally.
Fig. 38.7 Al evaporation source—long-lived

An improved design for this type of source has recently been reported [20]. Two
improvements are incorporated to outgas the source and its supports more thor-
oughly. First, the heating leads wrapped around the Al2O3 are mounted on
degassing loops as described on p. 407 and 437; secondly, the entire source is
surrounded by a Ta radiation shield that can be separately degassed by means of
electron bombardment from a W filament mounted inside the shield, somewhat as
shown on p. 407. After degassing by these methods, the source could be operated
for hours in the 10−11 mbar range [20]. In addition, the outside of the Al2O3 tube
was modified by grinding a helical channel along its length. This reduced the mass
of the tube and increased the thermal contact between the heater filament and the
tube. Finally a Chromel-Alumel thermocouple was connected to a Pt sphere that
itself was attached to the tube by melting a Pt wire with a hydrogen flame and
observing that a portion of the molten Pt drops away while another portion wets the
alumina [20]. The thermocouple reads a temperature hundreds of degrees Kelvin
below the Al temperature but this may still be used to control the Al evaporation
rate even when the thermal contact between the heating wire and the alumina tube
varies [20].
38.8 Cr Evaporator Sources of Compact Design
Often the methods used for metal film evaporation require large, high-current
evaporation devices that are incompatible with experimental apparatus designed to
study the films themselves. Both heating and large outgassing effects are associated
with many of the high-power evaporation devices. Therefore, very simple designs
will often give better results in research experiments that do not require massive
film deposition capabilities. The Cr evaporator discussed here is one example of
ultimate simplicity which meets these requirements [21].
Two sizes of W wire loops, 0.25 and 0.5 mm in diameter, are used. These wires
are cleaned with 600 grit sandpaper followed by ultrasonic cleaning in acetone and
methanol, and then are bent into a hairpin shape as shown in Fig. 38.8. The hairpin
filaments are suspended in an electroplating bath with about 2-mm submersion. The
electroplating bath is of the following composition [22]:
100 ml H2O, deionized
30 g CrO3
5 g NaOH (add slowly)
0.6 g H2SO4
0.1 ml ethanol
Electroplating is carried out with a Pb anode. Currents of about 45 and 90 mA
are used for the small and large filaments, respectively. The plating voltage is about
3 V. Electroplating for 18 h typically deposited 75 and 250 mg of Cr on the small
and large loops, respectively. Before use, the filaments are rinsed in deionized water
38.8 Cr Evaporator Sources of Compact Design 449
Fig. 38.8 Electrodeposited

Cr on W wires for Cr thin film
deposition
and methanol. After installation, degassing is done before bakeout at 5–6 A and
15–16 A for the small and large filaments, respectively. The filaments are outgassed
a second time during chamber cooling after bakeout, and may then be used in
ultrahigh vacuum without appreciable gas evolution in the 6 × 10−10 Torr range.
The small evaporators yielded Cr film growth rates of 4 Å/min at a current of 7 A
for a distance of 5 cm. The larger evaporator yielded 4 Å/min at a 10 cm distance at
20-A heating current. These evaporation rates could be increased by a factor of
10–20 by working at higher filament currents. Both filaments produced about
4000 Å of total Cr film thickness before depletion. Auger spectroscopy showed that
impurities were not present at the limit of Auger detection if the sources were
thoroughly outgassed before use. For early stages of evaporation from a poorly
outgassed filament, oxygen impurities were seen. The films produced are of suffi-
cient purity to be used in studies of the magnetic properties of multilayer structures
[23].
38.9 Electron Beam Evaporator for Refractory Materials
While many substances can be evaporated using ohmic heating of a filament or a

crucible, the successful evaporation of highly refractory materials such as tungsten
or molybdenum requires a more intense source of heat, which can be provided by
using a high-energy electron beam. A serious problem with conventional electron
beam evaporation methods is the emission of both high- and low-energy stray
electrons, which cause effects such as:
1. Electron stimulated desorption from walls and the cold shroud which may lead
to film contamination.
2. Film damage.
3. Interference with current measuring instruments, such as ionization gauges,
mass spectrometers and electron diffraction apparatus.
Of the two types of electron gun evaporators (magnetically and electrostatically
focused instruments), the magnetically focused instruments are more susceptible to
the difficulties mentioned above [24]. Figure 38.9a shows a schematic diagram of
an electrostatically focused electron gun. A thoriated W filament of circular
geometry is located between a shield and a cage, and both are biased at between −5
and −10 kV (−HV). The filament is heated by a 12-V, 30-A power supply.
Electrons are emitted and follow trajectories to the grounded hearth containing the
sample to be evaporated. A focusing mechanism (not shown) can change the rel-
ative position between the hearth and the rest of the components of the gun,
permitting an exact focusing of the electron beam on the top of the molten metal in
the hearth. A second degas mode involves keeping the filament at −3 kV and
grounding the remaining components of the electron gun evaporator. Out-gassing
was reduced by making all components out of refractory materials, such as W, Mo,
and Ta. The shield and hearth were made of Mo; the cage was constructed of Ta
mesh on a Mo frame with a W top cover containing a central hole. Using a Ta mesh
cage prevents heat buildup, and radiation-induced heating of large surfaces leading
to outgassing.
The secondary electrons produced in the gun cannot escape from the cage, which
is biased at −HV. The high-energy electrons have a potential V in the range
−HV < V < −HV + Vf, where Vf is the potential drop across the emitter filament due
to its resistance. In addition, there are high-energy backscattered electrons, which
mostly are at a potential of V or a slightly lower potential owing to energy loss upon
scattering. Since the potential at the hole in the cage can be positive relative to −HV
by a few hundred volts, backscattered electrons will escape from it. To eliminate
this problem, a tungsten mesh grid with 2-mm spacings was placed over the hole to
retard electrons there. In addition, by biasing the filament at −9.9 kV, excellent
focusing is achieved and stray electron currents are minimized, as depicted in
Fig. 38.9b, where calculated trajectories for cylindrical symmetry are shown.
This evaporation source can operate for refractory metal evaporation at pressures
<2 × 10−9 Torr [24]. A cryogenically shrouded source of somewhat similar design
has been described [25]. A highly stable electron beam evaporation source has been
described in [26]. A method involving the measurement of the ion current generated
by ionization of evaporating metal atoms and relating this measurement to the rate
of metal evaporation is described in [27].
An excellent review of metal evaporation methods and materials compatibility at
high temperatures is available in [28].
38.10 Small Electron Beam Evaporator for Refractory Metals 451
Fig. 38.9 Electron beam evaporation, a evaporator, b electron trajectories
38.10 Small Electron Beam Evaporator for Refractory

Metals
As discussed previously (p. 449), electron beam evaporation is a desirable method

for the deposition of thin films of refractory metals that does not involve the
problems of ohmic heating of a filament to very high temperatures. The design
shown in Fig. 38.10 incorporates simple features [29], combined with low mass
components, eliminating problems with large surfaces outgassing during operation.
Fig. 38.10 Electron beam evaporation, a evaporator design, b circuit
The device is simple to build, does not require cooling water for its components,
operates in the 10−9 Torr range, and produces clean films of refractory metals.
The evaporation device is mounted on a 2.75-in.-diameter flange containing
high-voltage feedthroughs. A tungsten emitter filament, about 1 cm in length and
0.012 cm in diameter, is spot-welded on 0.1-cm-diameter Ni rods as shown. It is
positioned 1–2 cm from the evaporation source. The source is a 6-cm wire
(0.025-cm diameter) made of the refractory metal to be evaporated that is fashioned
38.10 Small Electron Beam Evaporator for Refractory Metals 453
into a loop on the outer end; the base of this wire is welded to a 0.1-cm-diameter Ni
rod connected to a feedthrough pin by a set-screw contact for easy replacement. The
source wire is placed on the axis of a Mo wire helix with a diameter of about
1.5 cm, made of 0.025-cm-diameter Mo wire.
For operation, the source wire is set at about +6 kV = Va, and the grid is set at
about −3 kV = Vg. The filament voltage, Vf, is slowly raised using a current supply
until the collected current at the evaporator source is in the 5 to 20-mA range. Next
the negative voltage on the grid is empirically adjusted until electron bombardment
heating occurs only on the loop of the source wire. Overheating the loop can result
in melting which is undesirable. Films with greater than 1000-Å thickness have
been evaporated from W, Ta, and Ir at rates in the range 1–10 Å/min.
A variation on this source in which a small Mo crucible replaces the pendant
droplet source described here has been designed and tested [30]. The mass of the
crucible is kept to a minimum, and the emitter coil is wound so that it encircles the
crucible edge and is slightly below the edge [30].
An excellent review of metal evaporation sources and materials compatibility at
high temperatures is given in [31].
The cover of this second edition features this simple metal evaporator source
because of its wide application.
38.11 Electron Beam Evaporator
A design [32] differing somewhat from those shown elsewhere in this section, which
focuses a high-energy electron beam at a point on the end of a source wire for the
purpose of evaporating a metal is shown in Fig. 38.11. The metal source wire is
1–2 mm in diameter, mounted on a high-voltage Cu feedthrough that can be
translated in the vertical direction. In normal operation, a molten droplet of metal is
produced on the end of the wire by electron bombardment. The circular electron
emitter filament, made of 0.25-mm-diameter W wire, is looped over itself in a
15-mm diameter circle, and welded to two Ta wire leads (1-mm diameter), which are
connected to vacuum feedthroughs as shown. The distance between the plane of the
loop and the source tip is typically 6–9 mm. Nominal operating conditions are shown
in Fig. 38.11. As the source wire is evaporated, the beam current decreases and may
be restored by advancing the source to a position where the current returns to its
original value. With a source-to-sample distance of 15 cm, several hundred Å of film
may be deposited before source motion is necessary. The filament is shielded from
the substrate receiving the evaporated film by the aperture plate, and this reduces
radiative warming of the substrate. The source is readily outgassed behind a closed
shutter by running with the cooling water removed and at a reduced voltage (0.5–
1 kV). During normal operation, with water cooling, the pressure in the UHV system
remains in the 1–4 × 10−10 Torr range. Impurity transport to the film has been
measured for Fe and Co source wires using both Auger spectroscopy and X-ray
Fig. 38.11 Electron beam evaporator

38.11 Electron Beam Evaporator 455
photoelectron spectroscopy, and impurity levels below the detection limit were
found.
A design similar to this, but using a very small Mo crucible in the position of the
end of the source wire, has been described [33] for the evaporation of low-melting
metals under conditions where the low outgassing qualities of this evaporator are
retained.
A review of metal evaporation sources is contained in [34]. In addition, an
excellent review of metal evaporation sources and materials compatibility at high
temperatures is given in [35].
38.12 Break-Seal Ampoule Doser
Often low-vapor pressure compounds are used as precursors for the growth of
compound semiconductor films on a substrate. Many of these compounds react with
stainless steel and therefore cannot be transferred through heated tubing. This
design places a glass ampoule of the compound close to the substrate so that the
vapors do not come into contact with metal prior to deposition [36].
A glass ampoule is filled with the compound in a dry box under an inert
atmosphere and the ampoule is sealed. The ampoule is rigidly attached to the end of
a transfer rod inside a copper carrier assembly that fits the ampoule (Fig. 38.12).
The copper assembly is wound with W wire, sheathed in fiberglass or quartz fiber
sleeving, and a thermocouple is placed in contact with the copper carrier for
temperature measurements during vaporization from the ampoule. The use of a
bellows transfer device is superior to a magnetically driven device, since the heater
Fig. 38.12 Break-seal ampoule doser

and thermocouple wires are fixed along the transfer rod during motion and do not
have to flex each time the ampoule is moved.
A linear motion device with a stainless steel blade is used to break off the neck of
the ampoule when it is desired to start the vaporization process. Once the neck is
broken, the ampoule can be moved into the UHV chamber through a gate valve,
and the temperature of the doser can be adjusted to produce the desired rate of
precursor vaporization.
A cryopump is chosen for the first chamber because it pumps water vapor
efficiently and it is not impeded by glass fragments.
38.13 Cadmium Sulfide Evaporation Source
Thin films of CdS may be thermally evaporated from CdS powder. The apparatus
shown in Fig. 38.13 is made entirely of tantalum, tungsten, and boron nitride [37].
A large area baffled aperture produces a uniform CdS deposit. In addition the baffle
shields the substrate from radiation from the indirect heater. The baffle has a
3-mm-wide gap between its outer edge and the Ta shield.
The sublimation source is made of a Ta cup welded to a Ta tube. This slips over
a cylindrical shield made of BN, which contains a rigid coil of fairly heavy wound
W wire in three strands. The entire assembly may be unloaded by pulling upward
on the handle, leaving the W wire exposed. In operation, the CdS nearest to the
central heater sublimes first. The source temperature is measured with a Pt/PtRh 13
thermocouple welded to the baffle. At a temperature of 550 °C, CdS is deposited at
a rate of 7 Å s−1 on the substrate at 120 °C and 24 cm distant. The operation time at
this rate of evaporation is about 4 h before refilling is needed.
By choosing the correct evaporation and source temperatures, the resistivity of
the deposited CdS film may be controlled.
A CdS evaporation source using elemental evaporation has been described [38].
38.14 Arsenic Atom Source
The use of elements from Groups II, III, IV, V, and VI in the production of semi-
conductor devices is widespread. Many of the elements sublime as atoms, but in
Groups V and VI, sublimation as dimers and tetramers is commonly observed. It is
advantageous—for example, in the growth of GaAs surfaces—to employ atomic As,
which forms surfaces with fewer defect sites than surfaces grown from As2 or As4
species. The device shown in Fig. 38.14 permits the substantial atomization of As2
and As4 by use of a final heated atomizer stage [39]. The principles embodied here
can also be used with other elements where atomization is desired. It should be
remembered that the equilibrium between atomic species and higher molecular
38.14 Arsenic Atom Source 457
Fig. 38.13 Cadmium sulfide evaporation source

Fig. 38.14 Arsenic atom source
weight species is favored toward the atomic species by high temperatures and by low
pressures.
Figure 38.14 shows, schematically, the device used to make atomic beams of As
containing small fractions of As2 and As4. A sublimation source, loaded with As(s),
is connected to a first-stage cracking furnace through an adjustable aperture.
Adjustment of this aperture allows the beam pressure of As species to be controlled
throughout the apparatus. The subliming species are conducted through the first
stage cracking furnace and then exit to a heated plate that serves as the second-stage
atomizer. The products from the second stage are significantly enriched in atomic
As, and are emitted from a second aperture toward the substrate. The atomizer can
be operated in the range above about 1200 K when the local pressure is below about
10−3 Torr; it can also be operated in the range above 1400 K when the local
pressure is about 10−2 Torr. Many materials can be used for the atomizer plate with
various elements to be atomized, including refractory metals such as W, La, Nb, Re,
Mo, and Ta, and stainless steel, Ni, graphite, Pt, B, Si, ceramics, and BN.
Compound formation between the material of the atomizer plate and the beam is to
be avoided.
The preferred range of operation for various materials is given in Table 38.1, and
substantial but not complete atomization will be produced under the listed
conditions.
38.15 Portable Microevaporator for Depositing Low Coverages of Single Atoms 459
Table 38.1 Preferred range Substance Pressure at atomizer Temperature range (K)
of operation—atom sources (Torr)
As <10−2 >1400
<10−3 >1200
P <10−3 >2000
<10−5 >1800
Sb <10−2 >1400
<10−4 >1200
S <10−2 >2000
<10−4 >1700
Se <10−2 >1600
<10−4 >1300
Te <10−2 >1300
<10−4 >1100
38.15 Portable Microevaporator for Depositing Low

Coverages of Single Atoms
Temperature programmed desorption from an adsorbed layer may be used to

deposit small coverages of atoms or molecules onto surfaces and provides a method
for the control of very small coverages on a collector surface [40, 41]. To do this a
tungsten filament is first cleaned in ultrahigh vacuum by heating to high temper-
ature, and is then covered with the desired adsorbate, such as an alkali metal (see
Sect. 40.4) or another metal which is known to evaporate at a convenient filament
temperature. The tungsten evaporator can be conveniently made by removing the
quartz bulb of a halogen lamp and mounting the fused lamp base on a convenient
support, here shown in Fig. 38.15 as molybdenum bars connected through a sap-
phire baseplate. Electrical contact to the filament can be made using two tungsten
helices welded to the lamp leads (one shown) which permit the adjustment of the
filament position up or down. The diameter of the wire in the helices should be
large enough that the current needed for evaporation from the filament does not
cause desorption from of residual gases from the helices (see Sect. 38.1). An
evaporator of this type could be loaded with adsorbate at one position in the UHV
chamber and then swung around to the position needed for deposition. It has been
used to deposit Li, Pd and Au onto single crystals, and the observation of atomic
dispersion of the evaporated material has been made with the STM.
Fig. 38.15 Portable microevaporator for depositing low coverages of single atoms
38.16 Nanoscale Templating of Close-Packed C60

Nanostructures
Linear chains of fullerene molecules are of interest due to their one-dimensional

architecture. The C60 molecule is a lively candidate for the building block for such
linear structures. Figure 38.16a shows an STM image of partially oxidized Cu(110),
where striped structures of ···Cu–O–Cu··· assemblies form during oxidation at
250 °C with a few monolayer dose of O2. Oxidized stripes a few hundred nm in
length form, separated by 0.76 nm wide gaps. Shown in Fig. 38.16b is a ball model
of a portion of the striped structure with clean Cu regions of width of 3a0, where a0
is the 0.25 nm interatomic spacing for Cu in the lateral direction. This 0.76 nm wide
gap fits well a C60 molecule with a diameter of 0.73 nm.
The deposition of C60 from an evaporator source in UHV leads to the formation
of long chains of C60 molecules as shown in Fig. 38.16c. which indicates that
regular structures can be obtained by annealing at 250 °C. Wire lengths of C60
longer than 100 nm can be obtained by this procedure. In Fig. 38.16d a higher
resolution image shows that the C60 chains consist of an ordered-pentamer and a
disordered-hexamer units. Further details of the structures achieved and their
electronic properties are given in [42].
38.17 Quantitative Molecular Beam Dosing of Slightly

Volatile Organics
The field of surface science is increasing concerned with the study of larger mol-
ecules with low vapor pressures. While a number of types of heated dosers are
described in this book none have the capability to deposit an accurately known flux
38.17 Quantitative Molecular Beam Dosing of Slightly Volatile Organics 461
Fig. 38.16 Nanoscale

templating of closed-packed
C60 nanowires
of molecules. In the method described in Fig. 38.17, an oven containing an internal

circulating fan is constructed around the source (sample) bulb to adjust the vapor
pressure accurately and a heated delivery system, maintained above the vaporiza-
tion temperature, conducts the vapor through a heated leak valve and into a heated
doser tube [43].
Figure 38.17a shows the construction of the heated doser. All surfaces which
contact the vapor are heated above the sublimation temperature, ensuring that
condensation does not occur. To do this simply, the doser is made of OFHC copper
to ensure maximum thermal conductivity, and a glass sleeve containing a wound
tungsten heating wire is clamped on the doser just under the 2¾″ flange. Heating at
this point, as measured by a thermocouple, can be used to adjust the temperature of
the doser guide tube which faces the crystal to receive the organic adsorbate.
In Fig. 38.17b, a schematic of the high temperature vapor source and conduction
system is given. Inside of the oven, maintained at a constant temperature to produce
the desired vapor pressure, a Baratron capacitance manometer indicates the vapor
pressure of the organic. The heated source may be pumped by a small turbopump
which is connected through a cold trap to the sample bulb, 4. The heated source
Fig. 38.17 Quantitative molecular beam dosing of slightly volatile organics, a heated gas doser,
b vacuum system—heated, c reproducibility of vapor pressure measurements of naphthalene,
d enthalpy of sublimation for naphthalene, e mass spectrometer for response in UHV chamber
during dosing
38.17 Quantitative Molecular Beam Dosing of Slightly Volatile Organics 463
region is connected to the doser in the UHV system by a gas line which is heated
about 20 K above the source temperature in the oven. Flow to the heated doser can
be interrupted by pumping the source through the turbopump.
Figure 38.17c shows the measured vapor pressure of naphthalene in the tem-
perature range 294–346 K. Three separate measurements agree well. Figure 38.17d
shows a Clausius-Clapeyron plot of the vapor pressure; the slope yields the sub-
limation enthalpy of 71.0 ± 0.2 kJ/mole, which agrees well with the literature.
Figure 38.17e shows the behavior of the mass spectrometer in the vacuum
system as naphthalene vapor is admitted at various vapor pressures through the
heated doser system. The walls of the UHV system initially adsorb naphthalene
molecules, and the mass spectrometer observes this wall-adsorption during the
approach to steady state system pressure, which is achieved after about 400 s
dosing. However, a crystal placed in the molecular beam will see an immediate
steady flux of naphthalene and it was shown that the coverage increased in linear
proportion to the pressure of naphthalene as measured in the oven source multiplied
by the time of exposure [43].
A critical feature in the apparatus is a heated adjustable leak valve, 1, which is a
Granville-Phillips (series 203) variable leak valve. In the different experiments, the
leak valve is not adjusted to open or close it, since these valves are not reproducible
upon cycling open and closed. It serves therefore only to adjust the fraction of the
incident vapor which passes. It can be adjusted initially to select the time required to
achieve a desired coverage on the single crystal but is not adjusted further in
quantitative dosing measurements. Its leak rate during our measurements was about
2 × 1014 molecules naphthalene/Torr s.
This doser method has been shown to work well using other low vapor pressure
organics [43, 44].
38.18 Organic Doser
A simple doser for subliming volatile organic materials is made from a 1 in. ceramic
standoff with vented screws mounted in both ends. The top screw covers a small
reservoir containing the powdered material to be sublimed. The bottom reservoir is
empty and also contains a vented screw through which a thermocouple is forced
and spot welded. Heating is accomplished by passing current through 0.010″ W
(3 %Re) wire, although ordinary W wire would certainly serve too. A shutter may
be used to interrupt evaporation [45] (Fig. 38.18).
38.19 Vapor Deposition of Organic Thin Films
The evaporation of organic materials and higher-vapor-pressure inorganic materials

into an ultrahigh vacuum system may be done with various heated sources.
A difficulty arises with large Knudsen cells and metal boat evaporators in that they
have a large heat capacity and are difficult therefore to bring to temperature rapidly;
also, difficulties in controlling the temperature may be present. A small evaporation
source eliminating these problems is shown in Fig. 38.19 [46]. While the device
described in the literature was made from the cathode components of a commercial
electron gun [46], the basic design principles shown here may be useful in devices
built in many ways.
An indirectly heated cathode assembly in a ceramic base, and containing an
alumina-coated W spiral heater assembly, is provided with a second closely fitting
Ni tube to hold the organic material. In some cases, the organic material is held on a
plug of fine quartz wool to prevent it from falling out or bumping during evapo-
ration. A thermocouple is welded to the source tube to monitor the evaporation
temperature. A few mm3 of organic material is added, and the source is installed in
the vacuum system. One must be careful to determine that the vapor pressure of the
material at the base temperature of the vacuum system will not severely degrade the
vacuum. Materials such as long chain alkanes, phthalocyanines, and some metal
halides have been handled successfully at temperatures up to 500 °C. The film
deposition rate, controlled by the temperature, ranged from 0.01 to 5 nm/s, typically
0.1 nm/s. The gun can be used for several years.
38.20 Evaporator for Organic Film Deposition 465
Fig. 38.18 Organic doser
38.20 Evaporator for Organic Film Deposition
Many evaporators for organic materials are made of ceramic crucibles (e.g. boron
nitride) with ohmic heaters and thermocouples for achieving and measuring the
evaporation temperature. At least 3 electric feedthroughs are necessary for each
crucible if one leg of the ohmic heater is connected to ground. In this design, only a
single feedthrough is needed as the ohmic heater is also used as a temperature
Fig. 38.19 Vapor deposition of organic thin films
sensor by means of measuring its electrical resistance during operation. A Pt heater

wire is tightly wrapped around a grooved BN crucible and pressed into place by a
mica sheet insulator and a Mo shield which is in two sections for clamping tightly.
For measuring the temperature, a small current of order 0.1 A (I) is passed through
38.20 Evaporator for Organic Film Deposition 467
Fig. 38.20 Evaporator for organic film deposition
the Pt wire and the voltage drop (V) is measured. For the I/V ratio (R), the resistance
is obtained and used as a measure of the temperature by using the well-documented
polynomial of the Pt resistance versus T. For pulsed evaporation, the current is
increased using simple electronics, keeping the crucible at preset temperatures
before and after evaporation. The accuracy of the measurement of R and T can be
increased by using larger diameter Pt lead wires welded to the end of the resistance
wire, minimizing effects on the total circuit resistance by heating of the Pt lead
wires [47] (Fig. 38.20).
References
1. The use of degassing loops originates from the thin film community but the historical origin of
this useful method is not known
2. Professor R.J. Behm (personal communication)
3. W.F. Leonard, Rev. Sci. Instrum. 40, 187 (1969)
4. D.A. Steigerwald, W.F. Egelhoff Jr, J. Vac. Sci. Technol. A7, 3123 (1989)
5. K.J. Ross, B. Sonntag, Rev. Sci. Instrum. 66, 4409 (1995)
6. K.K. Chin, P. McKernan, I. Lindau, J. Vac. Sci. Technol. A4, 1949 (1986). This paper
contains references to tests of this device as well as to related devices
8. F. Jansen, J. Vac. Sci. Technol. 21, 106 (1982)
9. P.R. Davis, H.R. Poppa, J. Vac. Sci. Technol. 15, 1771 (1978)
10. P.J. Chen, M.L. Colaianni, J.T. Yates Jr, Phys. Rev. B 41, 8025 (1990)
12. H.P. Bonzel, A.M. Bradshaw, G. Ertl (ed.), Physics and Chemistry of Alkali Adsorption
(Amsterdam, Elsevier, 1989)
13. F.J. Esposto, K. Griffiths, P.R. Norton, R.S. Timsit, J. Vac. Sci. Technol. A12, 3245 (1994);
for details of the surface behavior of Li/Al alloys
14. F.J. Esposto, C.-S. Zhang, P.R. Norton, R.S. Timsit, Surf. Sci. 290, 93 (1993)
15. Supplied by Alcan, International, Ltd. It is rumored that Canadian beer cans are made of this
alloy (personal communication, P.R. Norton)
16. R.P. Schorn, E. Hintz, S. Musso, B. Schweer, Rev. Sci. Instrum. 60, 3275 (1989)
17. G. Polanski, J.P. Toennies, Surf. Sci. 260, 250 (1992)
18. W.J. Wytenburg, R.M. Lambert, J. Vac. Sci. Technol. A10, 3597 (1992)
19. Degussa Limited, Wilmslow, Cheshire, England
20. H.-G. Keck, E. Kopatzki, and R.J. Behm (private communication)
21. J.J. McClelland, J. Unguris, R.E. Scholten, D.T. Pierce, J. Vac. Sci. Technol. A11, 2863
(1993)
22. H.F. Mar et al. (eds.), Encyclopedia of Chemical Technology (Wiley, New York, 1979), vol. 6,
p. 104
23. J. Unguris, R.J. Celotta, D.T. Pierce, Phys. Rev. Lett. 67, 140 (1991)
24. M. Heilblum, J. Bloch, J.J. O’Sullivan, J. Vac. Sci. Technol. A3, 1885 (1985)
25. L. Ramberg, E. Flemming, T.G. Andersson, J. Vac. Sci. Technol. A4, 141 (1986)
26. Y.W. Beag, K. Egawa, R. Shimizu, Rev. Sci. Instrum. 64, 3647 (1993)
27. T. Jones, J. Sawler, D. Venus, Rev. Sci. Instrum. 64, 2008 (1993)
29. J.R. Waldrop, R.W. Grant, J. Vac. Sci. Technol. A1, 1553 (1983)
30. J. Araya-Pochet, G.A. Mulhollan, J.L. Erskine, Rev. Sci. Instrum. 62, 2288 (1991)
32. B.T. Jonker, J. Vac. Sci. Technol. A8, 3883 (1990)
33. J. Araya-Pochet, G.A. Mulhollan, J.L. Erskine, Rev. Sci. Instrum. 62, 2288 (1991)
34. L.I. Maissel, R. Glang, Handbook of Thin Film Technology (McGraw Hill, NY, 1970), pp. 1–51
36. Dr. M.S. Jackson, J.G. Ekerdt, Department of Chemical Engineering (The University of Texas
at Austin, Austin, TX 78712-1062) (private communication)
37. R.T. Galla, Rev. Sci. Instrum. 36, 403 (1965)
38. D. Beecham, Rev. Sci. Instrum. 41, 1654 (1970)
39. Professor Stephen R. Leone, University of Colorado at Boulder, Joint Institute for Laboratory
Astrophysics, Boulder, CO 80309-0440 (private communication). U.S. Patent No. 5,541,407
Arsenic Atom Source, S.R. Leone, A.L. Alstrin, P.G. Strupp, A.K. Kunz, S.M. Casey, and
R.V. Smilgys
References 469
40. H.-P. Rust, T. Konig, G.H. Simon, M. Nowicki, V. Simic-Milosevic, G. Thielsch, M. Heyde,
H.J. Freund, Rev. Sci. Instrum. 80, 113705 (2009)
41. K. Lammle, A. Schwarz, R. Wiesendanger, Rev. Sci. Instrum. 81, 053902 (2010)
42. M. Feng, J. Lee, J. Zhao, J.T. Yates Jr, H. Petek, J. Am. Chem. Soc. 129, 12394 (2007)
43. L. Thompson, J.-G. Lee, P. Maksymovych, J. Ahner, J.T. Yates Jr, J. Vac. Sci. Technol. A21,
491 (2003)
44. J. Lee, O. Kuzmych, J.T. Yates Jr, Surf. Sci. 582, 117 (2005)
45. K. Vijalakshmi, R.A. Rosenberg, private communication
46. N. Sato, K. Seki, H. Inokuchi, Rev. Sci. Instrum. 58, 1112 (1987)
47. P. Feulner, private communication
Chapter 39
Ions
39.1 Solid State Cesium Ion Gun Source
The use of Cs as an adsorbate of interest on metals and semiconductors extends

back to the early work of Langmuir in this area. Several types of sources have been
used, including sources that use elemental liquid Cs [1–4], and Cs+ surface ioni-
zation sources [5, 6]. One type of Cs+ source uses Cs vapor that feeds a porous W
plug from behind, producing high current density and low beam temperature. This
type of source suffers from the need for differential pumping, and the porous W can
also clog when the system is vented to air [5, 6]. A second type of surface ionization
source consists of an aluminosilicate powder deposited on a tungsten base. Cs+ ions
are stored in the aluminosilicate and are thermionically emitted [7]. This source
requires no differential pumping, but because of the variation in surface potential
across its surface, there is a large energy spread of the emitted ions. The source to
be described here combines the best features of both of these types of surface
ionization sources [7]. Depending upon the energy of the ions desired, different
extraction geometries are employed [7].
Figure 39.1a shows a cross-sectional view of the Cs+ source. A pellet of Cs-M
zeolite (cesium mordenite) is made by ion exchange with Na-M zeolite in aqueous
solution [8]. Successive exchanges of the Na+ for Cs+ must be carried out, and one
must use specially purified CsCl salt (reduced Rb impurity levels) to avoid the
incorporation of Rb into the source [9, 10]. By heating the exchanged mordenite
zeolite to 1050 °C, a phase transition from Na-M to Cs-M structure occurs. The
powder is then ground to a particle size of 1 μm, placed in a cylindrical die, and
heated to 1350 °C for 3 h, producing a hard pellet that may be machined slightly
with a grinding wheel. The emitting surface of the pellet is then coated with a W
film, which provides the surface ionization, as well as a surface of constant
potential. The back side of the pellet is coated with a Pt film for electrical contact.
When the pellet is heated, the flux of Cs toward the grounded W film can be
controlled by the applied bias potential across the pellet. A total charge of 10 C can

DOI 10.1007/978-3-319-17668-0_39
472 39 Ions
Fig. 39.1 Solid state Cs+ ion gun + other alkalis. a Ion gun design. b Performance. c Zeolite
source Performance for various alkalis
be extracted from a pellet 0.5 cm in diameter and 1 cm long. Current densities up to

30 mA/cm2 have been extracted.
The time needed to degas the source for operation at less than 10−10 Torr is
about 1 day at 1100 °C. When not being used, the source is kept at about 500 °C to
avoid the need for further outgassing. The advantage of this source is that only Cs+
is emitted, so that measurements of the collected charge can be converted into
absolute coverage [7].
In Fig. 39.1a, the W film on the pellet is pressed against an aperture by a spring,
and a bias electrode presses against the Pt film on the back of the pellet. Heating by
radiation occurs from a 0.25 mm diameter W coil filament, which is coated with
alumina powder. Coaxial Ta foils are used as internal shields to reduce the power
loss, and 15–18 W brings the pellet to operating temperature (about 1000 °C). In
this particular design, a Pierce electrode arrangement is used to control the ion
focusing for ions in the range 50–500 eV [7]. For ion energies in the range 5–50 eV,
a planar diode extraction geometry is used [7], and for ion energies above 500 eV,
the Pierce gun is mated to a column containing two Einzel lenses [7]. The per-
formance of the different gun designs is given in the original reference [7], and one
example of the Pierce gun performance is shown in Fig. 39.1b. Here, the space
39.1 Solid State Cesium Ion Gun Source 473
charge limitation governs the V3/2-dependence of emitted Cs+ current, where V is

the ion energy. However, at 920 °C, the supply of Cs+ falls near an ion energy of
100 eV, and a larger bias across the pellet restores the performance to that of a
space charge limited source.
It has been reported [11] that alkali ion sources involving aluminosilicates in
contact with tungsten metal must be “burned in” in ultrahigh vacuum for hours in
order to remove the impurity alkalis. Thus, alkalis with low ionization potential will
dominate in the ionization process at first, and trace quantities of these impurity
alkali atoms will result in a large fractional contamination during first operation of
the source. If the source is burned in and then allowed to stand for months in UHV,
the impurity ions may reappear in substantial quantity [11]. Thus, it is always wise
to use a mass spectrometer for analysis of the alkali ion product from these sources
[11]. The application of SIMION ion optical calculations [12] to a large-capacity
alkali ion gun has been reported [13].
A careful study of the relative efficiency of zeolite A ion sources for Li+, Na+,
K , and Cs+ has been published [14] and this report contains details of the prep-
+
aration and evaluation of the various sources. Figure 39.1c shows the saturation
current density for the four zeolite emitters coated on Pt surfaces as a function of
emitter temperature, measured in a simple diode arrangement in which the ion
emission extrapolated to zero bias was evaluated. An approximately 0.5-cm2 area of
Pt mesh emitter was used. A Rb+ ion source, operating with a mixture of Rb+-
substituted aluminosilicate zeolite and powdered Mo has been described. Here too a
Pierce-type electrostatic focusing structure was used [15].
The investigation of the Cs+ purity from various composition aluminosilicates
has been reported [16].
The design of a high-quality ion gun for delivering Li+ ions in the energy range
1–100 eV with an energy spread of only 0.22 eV has been published [17].
39.2 Alkaline Earth Metal Ion Sources
Although high-temperature sources for production of the alkalis are well known, the
use of similar alkaline earth sources is less well established. Recently, a study of the
oxides MgO, CaO, SrO, and BaO has been made, and the purity of the emitted
Mg+, Ca+, Sr+, and Ba+ ions have been investigated [18].
Filaments (0.25-mm diameter) of either W or Ta were used to support the oxides
in front of a quadrupole mass spectrometer as shown in Fig. 39.2a. This was done
by making a slurry of the powdered oxide in distilled water and transferring powder
to the metal coil. Tungsten wire was used for CaO and MgO, and Ta wire was used
for the other oxides, including MgO in later studies, since it was found that MgO
reacted with W. Measurements were made of the total ion emission as well as the
purity of the emitted ions as a function of the temperature. Temperature measure-
ments were made with a pyrometer.
474 39 Ions
Fig. 39.2 Alkaline earth ion sources. a Ion detection by mass spectrometer. b Ion current versus
T. c BaO/Ta. d SrO/Ta. e CaO/W
Figure 39.2b shows the ion current measured for three of the filaments as the
temperature is raised. Currents as high as 1 μA are observed at elevated tempera-
tures, but failure of the filaments occurs rather rapidly at extreme temperatures
owing to the reaction of the metal oxide with the refractory metal filament.
Figure 39.2c–e show the impurity behavior for BaO, SrO, and CaO as a function of
the temperature. In all cases, alkali metal ion impurities are observed, even though
the oxides were of purity of 99.9 % or better. This is due to the highly preferred
39.2 Alkaline Earth Metal Ion Sources 475
surface ionization of the alkalis as a result of their low first ionization energy. As the
filaments are heated to higher temperatures, the alkali ion impurities decrease in
yield and rather pure Ba+, Sr+, and Ca+ ions are liberated. The purity of Mg+ from
MgO coatings was less than 50 % at all temperatures, and this would therefore be
an unsuitable source for Mg+.
References
1. J.-L. Desplat, C.A. Papageorgopoulos, Surf. Sci. 92, 97 (1980)

2. G.S. Tompa, J.L. Lopes, G. Wohlrab, Rev. Sci. Instrum. 58, 1536 (1987)
3. T. Castro, R. Reifenberger, Rev. Sci. Instrum. 58, 289 (1987)
4. R.D. Swenumson, U. Even, Rev. Sci. Instrum. 52, 559 (1981)
5. D.G. Welkie, in Secondary Ion Mass Spectrometry SIMS V, ed. by A. Benninghoven et al.
(Springer, New York, 1986), p. 146
6. G.R. Brewer, Ion Propulsion (Gordon and Breach, New York, 1970), Chap. 4
7. A.E. Souzis, W.E. Carr, S.I. Kim, and M. Seidl, Rev. Sci, Instrum. 61, 788 (1990)
8. D.W. Breck, Zeolite Molecular Sieves (Wiley, New York, 1977), Chap. 7
9. R.E. Weber, L.F. Cordes, Rev. Sci. Instrum. 37, 112 (1966)
10. T.E. Madey, J.T. Yates Jr., J. Vac. Sci. Technol. 8, 39 (1971)
11. M.J. Hogan, P.P. Ong, Rev. Sci. Instrum. 61, 1338 (1990)
12. D.A. Dahl, J.E. Delmore, The SIMION PC/PS2 User’s Manual, Version 3.1, Informal Report,
EGG-CS-7233, Rev. 2 (November 1987). Program available from Idaho National Engineering
Laboratory, EG&G Idaho, Inc., P.O. Box 1625, Idaho Falls, ID 83415
13. S.T. Beck, D.W. Warner, B.A. Garland, F.V. Wells, Rev. Sci. Instrum. 60, 2653 (1989)
14. Y. Satoh, M. Takebe, K. Linuma, Rev. Sci. Instrum. 58, 138 (1987)
15. D.W. Hughes, R.K. Feeney, D.N. Hill, Rev. Sci. Instrum. 51, 1471 (1980)
16. P.P. Ong, T.L. Tan, Rev. Sci. Instrum. 65, 3729 (1994)
17. H.B. Haskell, O. Heinz, D.C. Lorents, Rev. Sci. Instrum. 37, 607 (1966)
18. T.L. Tan, Z.L. Zhou, P.P. Ong, Rev. Sci. Instrum. 66, 2879 (1995)
Chapter 40
Active Gases and Species
40.1 High Flux Atomic Hydrogen Sources—Thermal
Atomic hydrogen is employed for research in the surface science field as a reagent
for rapidly producing chemisorbed hydrogen on semiconductor surfaces and also on
metal surfaces that do not readily dissociate molecular H2. The surface chemistry of
atomic hydrogen has been investigated as early as the work of Langmuir [1, 2] who
first produced it on a hot tungsten filament. The production of atomic hydrogen has
been investigated by a number of workers [3–5]. It has been found that the effi-
ciency of H atom production is proportional to (PH2)1/2 at H2 pressures above about
10−6 Torr where pseudoequilibrium conditions exist. At lower H2 pressures, the rate
is proportional to PH2. Under these conditions, the reported efficiency of H2 dis-
sociation is about 0.3 per H2 collision with the tungsten surface at temperatures
above 2500 K [3]. It has been observed that the efficiency of dissociation is
decreased by impurities often present on a tungsten filament [4].
Two sources that place a relatively high H2 gas density near a heated tungsten
surface have recently been described [6, 7]. Both of these sources produce large
fluxes of atomic hydrogen compared to a bare tungsten filament operating in the
vacuum system, if comparisons are made at the same gas throughput to the system’s
pumps. In Fig. 40.1a a tungsten capillary (0.6-mm ID, 1.6-mm OD) is electron
bombarded from a tungsten filament so that the end of the capillary is heated to
1800–1815 K over a 4-mm length. An acceleration voltage of +560 V on the
capillary (HV) and a 30 mA emission current achieve this condition. A Cu shield,
with a conical hole drilled on axis, serves as a radiation shield. It is cooled by a
water circulation tube that is butt-brazed to the solid Cu shield. The sample is
placed 10 mm in front of the shield, or 20 mm from the end of the W capillary. The
back end of the W capillary is snugly fitted into a ceramic sleeve, which also
connects to a 4-mm-diameter stainless tube that delivers H2 gas to the atomizer.
Since the gas density is low during operation, no need exists for vacuum tight seals,
and the adapter uses only friction fits to the capillary and the tube. The working H2

DOI 10.1007/978-3-319-17668-0_40
478 40 Active Gases and Species
pressure at the entrance to the W capillary is about 10−2 Torr. In order to change the
position of the H atom doser, it is mounted on a translational device. It is reported
that a total atomic H flow of about 3 × 1014/s may be achieved under these
conditions. The degree of dissociation approaches 50 %. Therefore the total gas
influx is small, minimizing the background H2 pressure. It is estimated that about
20 % of the H atoms are emitted in a central portion of the beam with a cone angle
of ±15° [8, 9].
In Fig. 40.1b, a spiral W filament (1.5-cm diameter, 0.025-cm-diameter wire,
14.5-cm length) [10] is located within a cup-shaped stainless steel doser having a
perforated array in its central portion [7]. Hydrogen gas flows into this doser
through a tube that is connected to a calibrated conductance aperture (see pp. 89–
91; 421–428), and the flow rate is adjusted using a capacitance manometer on the
high-pressure side of the aperture. A typical flow rate of about 2 × 1016 H2/s is
employed, at a system pressure of 2 × 10−7 Torr. The incoming H2 encounters a
deflector baffle that produces a random flux on the back of the perforated array. The
sample is shielded from radiation by the wall of the doser body. In order to reflect
the H atoms to the sample (avoiding radiation heating of the sample as well as
possible impurities from the hot filament) a cooled Pyrex plate at 120 K is located
close to the outlet of the atomic hydrogen source. This reflects atomic H with high
efficiency [11, 12]. The Pyrex is held onto a stainless steel cooling reservoir with
metal clips and is thermally connected to the l–N2 reservoir with a film of In metal.
In studies of the efficiency of the doser [7], it was found that a reduction in
efficiency of a factor of 4.8 occurred using the reflector source compared to a
line-of-sight geometry. However, experiments working at the same system gas
throughput indicate that the doser produces about 50 times more atomic hydrogen
than a bare W filament, operating without a doser geometry. Thus, even the factor
of 4.8 reduction in efficiency owing to the reflector is compensated by the higher
atomic H production rate achieved at the high gas densities in the doser. This source
is therefore about a factor of 10 more efficient than a bare W wire in delivering
atomic H to the sample for identical H2 gas loads. A monolayer of H atoms (about
4 × 1014/cm2) was adsorbed by a Si(100) surface at 400 K when 4 × 1018 H2
molecules had passed through the doser with the W filament operating at 2000 K.
Figure 40.1c shows the relative efficiency of atomic H production plotted versus
the reciprocal temperature (K−1). The source produces atomic H at a rate which
reaches a maximum near 2000 K. Similar results were reported by George et al.
[13].
An interesting study, comparing the pumping speed of atomic H with that of H2
during adsorption on engineering surfaces in the range 2.5–4.3 K, has been pub-
lished [14]. Interestingly, the pumping speed for H2 is between a factor of 3 and 40
above that of atomic H. This difference in either condensation or adsorption is
attributed to the higher internal energy of the atomic H compared to H2.
A hydrogen atom source involving a design similar to that of [6] has been
carefully evaluated to determine the efficiency of atomic H production [15].
A second similar source has also been evaluated by mass spectrometric methods
40.1 High Flux Atomic Hydrogen Sources—Thermal 479
Fig. 40.1 High flux atomic H sources. a H atom source—hot tube design. b H atom source—
spiral filament design. c Efficiency of design B for H-atom formation
[16]. In addition, atomic H beams, cooled by collision with surfaces near 6 K, have
been produced [17].
A number of detectors for atomic H have been devised and are reviewed briefly
in [18]. One of these [18] involves the use of a thermistor (primarily NiO) for
measurement of an atomic H flux. Electrical conductivity changes, measured with a
Wheatstone bridge and phase-sensitive detection, were used for the evaluation of
the flux of atomic H made in a microwave discharge. Control experiments elimi-
nated the possibility of effects owing to the influence of heating and radiation on the
detector. The mode of operation of the detector is unknown and could be due to
local heating as a result of H atom recombination or to electronic surface effects.
The lower limit of sensitivity is estimated to be 1011 H/cm2 s.
An atomic H source involving the use of a W/26 % Re alloy tube ohmically
heated to 2400 K has been described. The W/26 % Re alloy does not have the brittle
properties of pure W [19].
Many interesting uses for atomic H have been described in the literature and are
beyond the scope of this book.
40.2 Atomic Hydrogen Beam Source—Radio

Frequency Driven
The use of thermal sources for the production of atomic H has been described on
p. 477. It is also possible to produce beams of atomic H using microwave discharge
sources. These sources are convenient to utilize but one must be cautious that
impurities and radiation from the discharge region do not affect the experiments.
There are many different RF discharge sources available, as described in [20].
A typical source not unlike others that have been made [20] is described below.
Figure 40.2 shows a diagram of a Pyrex glass source of simple design. It is
capable of producing atomic H with up to 95 % dissociation of H2(g). The density
of atomic H ranges from 5 × 1011 to 2 × 1013 H/cm3. The tube is 18-mm inside
diameter and 240 mm long. It is terminated at each end by a 1-mm-ID capillary to
prevent the discharge from spreading outside the source, and it is cooled by a water
Fig. 40.2 H atom source

40.2 Atomic Hydrogen Beam Source—Radio Frequency Driven 481
jacket to reduce the temperature and to reduce H atom recombination on the inner
walls [21]. This may be responsible for the long life and stability of this design.
The discharge is excited by about 25 W of RF power into a coaxial cavity
resonant at 35 MHz. The cavity consists of a copper cylinder, surrounding the
discharge tube, and a 12-turn helix of constant pitch. Power is coupled via a
single-turn link coupling. Details of cavity design and cavity dimensions have been
published [22]. The coaxial cavity is properly impedance-matched to the generator
to insure 90 % power absorption into the cavity. The RF generator is properly
designed to prevent the propagation of unwanted RF signals. The RF oscillator and
driver stages are housed within individual screened compartments. The amplifying
stages are narrow band with sharply tuned filters in each stage, eliminating har-
monics. Notch filters are used in the power input to prevent RF propagation into the
laboratory AC supply. Low noise was achieved in auxiliary instruments when the
RF supply was operating.
A hydrogen pressure of about 0.15 Torr was used in the source. Great care was
taken to clean the discharge tube before use, successively washing with hot chromic
acid (in sulfuric acid), water, acetone, HF(aq), and distilled water. A palladium leak
(see p. 511) was used to purify the H2 before use and back diffusion of contami-
nants was avoided. The tube operated for about 5000 h without trouble.
40.3 Atomic Fluorine Source
The delivery of reactive atomic species in high purity to surfaces poses one of the
difficult experimental problems in adsorption and surface reaction studies. Species
such at atomic F are particularly difficult to prepare and transfer without secondary
reactions on the walls of the source. An excellent design [23], probably applicable
in general to producing a number of reactive species, is shown in Fig. 40.3.
A low pressure 10 %F2/90 % He mixture is flowed through an alumina tube
which passes through an Evanson-type microwave discharge cavity, and the
products are then pumped through the apparatus using a rotary pump which is
corrosion resistant. The microwave discharge products are admitted to the prepa-
ration chamber through a 25-μm-diameter orifice mounted on the end of the outer
stainless steel tube. The orifice was laser-drilled (see p. 421) in a stainless steel
sheet of 25 μm thickness. The alumina tubing inside the source is thought to
become passivated with the formation of a nonvolatile layer of AlF3, and is superior
to a tube made of silica [23, 24]. The microwave discharge source [25] is operated
at 30–60 W of power in the total pressure range of 0.05-10 Torr.
The species emitted from the orifice were studied with a modulated quadrupole
mass spectrometer, which measured the consumption of F2 (38 amu). The results
show that essentially complete dissociation occurs below a total pressure of 2 Torr,
as shown in Fig. 40.3b. When working in the molecular flow regime below a
pressure of 5.7 Torr, the angular distribution is described as having a cos2 distri-
bution, and for sample sizes of the order of 0.5 cm at source-to-sample distances of
Fig. 40.3 Microwave source of F atoms. a Source, b performance versus pressure
about 1 cm, the fluence is rather uniform. It is calculated that at a total pressure of
0.7 Torr, the source delivers F atoms at a flux of about 1 × 1014/cm2 s in the beam
with a random flux from the source and the chamber being at about 5 % of this
beam flux.
Microwave sources are widely used in molecular beam studies [24, 26], but this
source is one of the first such sources to be successfully employed in surface studies
[27–29]. It is also useful for generating chlorine atoms. For chlorine, a quartz tube
may be used instead of an alumina tube [30, 31].
It should be noted that large gas loads accompany the use of this source, and that
a rapidly pumped preparation chamber, isolated from the ultrahigh vacuum surface
analysis chamber, is one way to reduce the gas load in the analysis chamber when
the source is being used.
40.4 Alkali Metal Sources 483
40.4 Alkali Metal Sources
The alkali metals form one of the most interesting class of adsorbates on the
surfaces of both metals and semiconductors [30]. As a result they have been widely
studied since the pioneering work of Langmuir in the 1930s [32]. The alkali metals
are also used in the production of low-work-function photocathode surfaces, and in
other applications in the field of physical electronics. As a result, excellent com-
mercial sources of alkalis for research are available from SAES getters [33].
Figure 40.4a shows a view of a typical SAES alkali getter source; these are
available to deliver Li, Na, K, Rb, and Cs. A powdered mixture of the alkali metal
chromate, M2CrO4, and a reducing agent (St 101 alloy—84 % Zr and 16 % Al) is
held in a nichrome boat with a very narrow slit running along its length, as shown.
A nichrome retaining wire is squeezed along the backside of the slit to prevent
powdered material from being ejected from the slit. Butting pieces of two designs
are described, which block the ends of the boat and which serve as connectors for
electrical heating. As current is passed, the temperature rises into the 550—850 °C
region, as indicated in Fig. 40.4b, and the solid state reduction reaction liberates the
alkali metal which escapes from the slit. The Zr also acts as a getter for impurity
gases such as H2 and CO. The typical operating temperature for deposition must be
adjusted on the basis of experience, but it is usually preferable to use lower rates for
controlled deposition experiments. The typical amount of H2 emitted from one of
the sources is about 1 Torr-cm3 per cm of getter length, and degassing is achieved
by initially working at 500 °C. Various geometrical shapes and sizes are available.
The solid state oxidation-reduction reaction between the reducing agents, Zr and
Al in the metal alloy reactant can be written as shown in Fig. 40.4c, but the exact
details of the reaction are probably not well worked out. An excess of the Zr-Al St
101 alloy is employed so as to have material for the gettering of active gases during
the reaction.
In addition to alkali metal getters, SAES also makes alkaline-earth Ba getter
sources of a different configuration from the alkali metal sources (Type St
15A/20HY/400 and St 15/20KLS/200). The reactant consists of 45 % BaAl4 alloy
mixed with 55 % Ni powder. This composition, on being heated to about 850 °C,
reacts exothermically instantaneously releasing about 25–30 % of its Ba. Continued
heating by RF methods releases the remainder of the Ba. These sources are used to
process vacuum tubes by throwing a heavy getter film of Ba, which is active for CO
[34], H2 and other gas adsorption. An SAES source of Ba geometrically similar to
the alkali sources is the type St 14/Fr, which can be ohmically heated as is done
with the alkali sources.
A description of an alternate design for a Cs ion source, involving diffusion of
Cs through a heated porous W plug, is given in [35].
Alkali metal sources may also be based on the direct evaporation of the metal,
and many such sources have been described. A novel source, which delivers large
quantities of Na as an atomic beam, has been described; it employs a metal mesh
wick to return excess Na to the evaporator section. This source also works under
Fig. 40.4 Alkali metal sources, a Structure of SAES alkali metal dispenser. b Typical
temperature-current performance. c Solid state reaction—using Zr or Al
40.4 Alkali Metal Sources 485
ultrahigh vacuum conditions [36]. In addition, zeolite sources containing alkalis are
often employed, and one of these is discussed on p. 471.
40.5 Thermal Source of Methyl Free Radicals
The involvement of methyl radicals in the surface chemistry of hydrocarbon

reactions is often encountered. However, because of the difficulty of producing
clean sources of CH3, this radical species has been neglected in adsorption studies.
Pyrolysis of molecules containing methyl groups has been the preferred method for
the generation of methyl radicals [37–41], but most methods also generate other
reactive species that can cause interference in studies of the methyl radical surface
chemistry. The source described here avoids this problem by using dimethyldiazine
as the source of the CH3 and by working at molecular flow conditions where high
pressure gas phase processes to produce other highly reactive molecules are avoi-
ded to a high degree [42]. The pyrolysis reaction
CH3 NNCH3 ! 2CH3 þ N2 ð40:1Þ
is a fairly clean reaction, and if gas phase reactions of the methyl radicals are
avoided, then only surface reactions in the source can contribute impurities.
Figure 40.5 shows a schematic design of the pyrolysis apparatus.
Dimethyldiazine is allowed to pass through a quartz tube that has its central section
heated by means of a closely wound Ta wire, 0.25-mm diameter, which is cemented
to the outside of the pyrolysis tube using a ceramic cement (Aremco Type 576). The
quartz tube is connected to the source of dimethyldiazine using a graphite ferrule
[43] in a swagelock reducing union. The thermocouple embedded into the heating
section allows the temperature to be adjusted to the optimum one for complete
conversion of dimethyldiazine to pyrolysis products, which occurs near 1200 K.
A cooling section beyond the pyrolysis region allows the CH3 radicals to become
accommodated at a lower temperature than their production temperature. Various
lengths of the three zones shown in Fig. 40.5b were studied [42]. A typical mass
spectrum for dimethyldiazine and for the completely pyrolyzed product is shown in
Fig. 40.5c. In addition to CH3, measurable fractions of CH4, C2H6, and H2 are
produced along with N2 at 1173 K and a chamber pressure of 5 × 10−7 Torr.
The dimethyldiazine was synthesized using the method of Renaud and Leitch
[44] and was stored in a stainless steel container over a period of several months
without decomposition. Ar was mixed with the dimethyldiazine as an internal
standard for quantitative flow measurements [42]. A quadrupole mass spectrometer
was used to acquire the mass spectra shown, working inside a stainless steel
chamber pumped by a turbomolecular pump [42].
Fig. 40.5 Methyl radical sources. a Methyl radical source. b Pyrolysis tube. c Mass spectrum of
gas from source
40.6 Production, Storage, and Use of Ozone
The ozone molecule, O3, is a powerful oxidizer, useful in situations where oxida-
tion of surfaces is needed under mild conditions. It has been used, for example, in
the growth of Ba2YCu3O7-δ thin films [45, 46]. However, when stored as the liquid
at cryogenic temperatures, O3 is dangerous and may explode. The apparatus
described here avoids this problem by storing O3 by adsorption in the pores of silica
gel at temperatures slightly above its boiling point [47, 48]. The rate of decom-
position under these storage conditions is about 5 % per week.
40.6 Production, Storage, and Use of Ozone 487
Ozone is toxic at a concentration of around 1 ppm by volume. The odor threshold

in air is about 0.01–0.02 ppm. The threshold limit value (TLV) is 0.1 ppm, which is
above the odor threshold. The use of ozone must involve safe storage, use of inert
surfaces such as passivated stainless steel, quartz, Pyrex glass, Teflon, or passivated
aluminum, and constant monitoring with an ozone detector [49].
Figure 40.6 shows an oxygen cylinder containing dry oxygen that feeds a com-
mercial ozone generator (which produces a corona discharge). The rate of production
of O3 is such that a concentration of O3 of 8 weight % is produced steadily at 1 atm
with an O2 feed rate of 1.7 L atm/min. This corresponds to the production of about
11 g of O3/h. The O3 is trapped at 195 K on 300 g of silica gel [50] up to levels of
about 7 weight %, so that in 2 h of operation this limit will be reached. Excess
oxygen is vented to the atmosphere with Valve A open and all other valves (B-D)
closed. The gel should display a deep blue color at this loading level. To transfer
small quantities of O3 into the vacuum chamber, Valve A is closed, and Valves B
and C are opened. Under these conditions, a steady pressure of O3 will be evolved
and pumped into the rotary pump, with a pressure of about 10−2 Torr, as monitored
by the capacitance manometer. Valve D may then be opened to admit O3 to the
vacuum chamber. This procedure has the advantage that the flowing O3 is contin-
uously produced by vaporization from the silica gel, maintaining purity during the
exposure of samples in the vacuum system to the tiny fraction of O3 that arrives
through the doser tube, compared to that which is exhausted.
Fig. 40.6 System for production, storage, and use of ozone

The system is equipped with O3 destruct tubes containing copper sponge that
catalytically decomposes O3. These are also available commercially. The rotary
pump oil should be non-hydrocarbon-based, using either Fomblin [51] or Krytox
[52]. Teflon O-rings should be used in all locations, and not rubber O-rings, since
ozone is well known to destroy rubber by attacking carbon-carbon double bonds.
Extreme care should be taken to keep the dry ice + acetone trap cold at all times
when O3 is trapped on the silica gel.
An alternate method [53] employs a cryopump controlled by a thermocouple
feedback to boil off O3 from liquid O3 at a precisely controlled temperature. The O3
is condensed in a vacuum-jacketed reservoir whose temperature is precisely
adjustable through the operation of a cryopump. These authors provide a rupture
disk for this O3 storage system in case of an explosion, which is a possibility for
liquid O3 stored in this manner.
A simple electrical discharge apparatus for producing 100 % pure O3 in a
nonflow system, and for continued repurification, has been described. Less than
100 cm3 of pure oxygen at atmospheric pressure is converted in a batch operation,
and purification occurs by pumping traces of residual O2 away at 77 K [54].
The calibration of a mass spectrometer for O3 and the measurement of its
cracking pattern in a magnetic mass spectrometer has been carried out carefully,
using a 5.3-L storage volume (Teflon coated) to receive the O3 from a cold trap after
pumping off impurity O2, etc. The storage volume was the source for a molecular
beam of O3 that entered the ionizer of the mass spectrometer. It was found that the
half-life of O3 at a pressure of 0.1 Torr in the storage volume was about 12 h at
room temperature [55].
40.7 Ozone Generator and Molecular Beam Doser Source
Ozone is widely used in surface chemistry and materials chemistry as a high power
oxidizing agent. In Sect 40.6, we show how to make relatively large quantities of
very pure laboratory ozone and how to quantitatively dose it into an ultrahigh
vacuum system. For occasions where small quantities of ozone are needed in an
ultrahigh vacuum system, a simpler doser working on a different principle [56] may
suffice as shown in Fig. 40.7.
Oxygen gas at a pressure less than about 100 Torr is admitted into a glass tube
specially constructed for the preparation of O3 and for the removal of excess O2.
The glass reservoir is immersed deeply in a liquid nitrogen bath and a Tesla coil is
used to convert O2 to O3. A dark blue liquid O3 will be seen to form on the inner
walls and to flow to the bottom of the reservoir. Several doses of O2 may be used to
build up the quantity of collected O3. The conversion process is aided by the
presence of a small amount of Cu wool in the neck of the preparation/storage
reservoir and the discharge from the Tesla discharge is directed into the gas by the
Cu. The discharge should be done through the liquid nitrogen in the bath.
Following partial conversion of O2 to O3, the excess of O2 is distilled away at liquid
40.7 Ozone Generator and Molecular Beam Doser Source 489
Fig. 40.7 Ozone generator and molecular beam doser source
nitrogen temperature being aware that liquid O3 has an appreciable vapor pressure
of a few Torr at 77 K and will also be removed slowly. The pump used for
distillation should not contain any oxidizable oil.
Admission of the O3 to the doser is accomplished by using a stainless steel
metering valve which allows the stored O3 to slowly leak into the glass beam doser
assembly and into the UHV system. The use of glass between the storage reservoir
and the Pyrex transfer tube assures minimal wall decomposition of the O3 by
collision with metal surfaces. Passivation by exposure to O3 will deactivate the
small amount of metal exposed. It will be found that even the Pyrex walls act
initially to decompose the O3 to some degree (see Sect. 40.6), and only by pas-
sivation in O3 will the rate of wall decomposition approach zero. To shut off the
doser beam at the UHV system, a small gate valve is placed into the middle of the
two section-glass doser assembly. Since it will be useful to have a way to monitor
the rate of flow of O3 into the ultrahigh vacuum system it is suggested that one uses
the response of the ion gauge in the UHV system to the flow of O3 to be monitored.
The O3 which is not adsorbed on the sample becomes randomized by collision with
the walls of the stainless steel system and will all decompose to O2 and the ion
gauge therefore measures an increase in O2 pressure which is proportional to the
beam flux of O3.
40.8 Detection of Ozone or Atomic Oxygen by Oxidation

of Silver Films
Silver reacts only slowly with molecular oxygen, while its reactivity with O3 is very
rapid. In the apparatus shown in Fig. 40.8a, O3, originating from liquid O3
vaporization in an O3 cryogenic storage system, is incident on a Ag film (500 Å
thickness) evaporated onto a quartz crystal microbalance [57] (see p. 486 and 309).
The rate of increase of mass is proportional to the O3 flux in the initial stages of
oxidation, and then falls off to a lower rate as bulk diffusion effects begin to control
the growth kinetics as shown in Fig. 40.8b. The O3 angular distribution from the
delivery tube (stainless steel) was also measured by moving the detector off axis,
maintaining a constant tube-to-detector distance of 6.3 cm, and a constant O3 flow
rate of 1.9 standard cm3 s−1. The profile of O3 is Gaussian (Fig. 40.8c), and when
Ar diluent was used, the flux of O3 decreased and the angular profile was broadened
somewhat. It was estimated that the detection efficiency of the Ag film was 26 % for
O3, assuming that one O atom from O3 oxidizes the Ag, and that molecular O2 is
evolved. This efficiency assumes that the O3 is 100 % pure when it exits the
delivery tube, and if this is not so, then the detection efficiency is higher. The near
100 % purity of the distilled O3 was chemically checked using reaction with KI
solution [58], and also by its vapor pressure. O3 fluxes in the range 1014 to
1016 cm−2 s−1 were measured.
Fig. 40.8 Detection of ozone or atomic oxygen. a Geometry for O3 detection. b Mass change
during O3 + Ag reaction. c Angular distribution of O3
40.8 Detection of Ozone or Atomic Oxygen by Oxidation of Silver Films 491
The fate of the O3 in the film growth system was studied by placing O3 detectors
in various positions outside the beam. It was found that almost no O3 was detected,
suggesting that efficient wall decomposition reactions occur, converting O3 to O2.
A similar study, involving atomic O made in an RF discharge, was made using
Ag films on a quartz crystal microbalance [59].
40.9 Purity of Ir and Pt Hot Filament Sources

for Atomic Oxygen Production
A number of reports have been made of the use of hot Ir and Pt filaments, operating
in an oxygen atmosphere, for the production of atomic oxygen, . Usually the atomic
oxygen is delivered directly to a surface by line-of-sight transmission from the
filament to the surface. The work to be described here has used mass spectrometry
of the gas phase above hot Ir and Pt filaments and Auger spectroscopy measure-
ments of the composition of an atomically clean Au surface to investigate exactly
what is evolved from the Ir and Pt surfaces [56, 60]. Unfortunately, it is found that
atomic oxygen is not cleanly produced and that the transport of Ir and Pt oxides also
occurs making suspect any use of this method to produce pure atomic O.
Figure 40.9a shows line-of-sight mass spectroscopy of an Ir filament operating in
1.8 × 10−5 Torr of O2. The detection of the Ir oxides is shown as the Ir filament
temperature is raised. It was shown that the Ir signal corresponds mainly to Ir(g)
production which occurs also in the oxygen atmosphere. Figure 40.9b shows the
Auger spectrum of a Au surface after long exposure to a hot Ir filament at 2096 K,
operating in ultrahigh vacuum. It is seen that even at this low temperature (compare
to Fig. 40.9a), over long times, the Au Auger spectrum is modified by the depo-
sition of Ir. An Auger ratio, A/B, may be used to deduce the magnitude of the Ir
coverage. In Fig. 40.9c, this ratio maximizes after about 100 h of exposure to Ir at
2096 K, meaning that the Au Auger signal has been obscured by several mono-
layers of Ir which have been deposited. In Fig. 40.9d, the same experiment, done in
O2(g), shows that Ir is detected by Auger spectroscopy on the Au substrate even at
*1600 K, whereas atomic O (or Ir oxides) are not detected until 1760 K. Thus
atomic oxygen, transport to Au does not occur in the absence of Ir transport
meaning that Ir should not be used as a source of pure atomic O [60].
A similar set of experiments were performed with Pt. As shown in Fig. 40.9e, the
low energy Auger spectrum of the Au is altered by the deposition of Pt at 1683 K in
ultrahigh vacuum. In 137 min of deposition the, Au 69 eV feature is extinguished,
meaning that several ML of Pt have deposited. In Fig. 40.9f, the development of the
O-Auger intensity and the Pt coverage, as measured by an Auger ratio, are com-
pared. It may be seen by a careful comparison of the two panels that Pt transport is
observed prior to atomic O (or Pt oxides) transport to the Au surface [61].
Fig. 40.9 Purity of Ir and Pt hot filament sources for atomic O production. a Line-of-sight mass
spectrometry—Ir + O2, b Auger ratios—Ir deposition in vacuum on Au (2096 K = TIr),
c Ir/Vacuum—2096 K—on Au, d Deposition of O and Ir in O2 on Au versus T, e Auger ratios—
Pt in vacuum on Au (1683 K = TPt), f Deposition of O and Pt in O2 on Au versus T
40.9 Purity of Ir and Pt Hot Filament Sources for Atomic Oxygen Production 493
Both of these experiments indicate that neither Ir nor Pt hot filaments, operating
in an oxygen atmosphere, can be trusted to deliver pure atomic O to an exposed
surface.
40.10 Pure Hydrogen Peroxide Doser

for Ultrahigh Vacuum
Hydrogen peroxide, H2O2, is a strong oxidizer, often used in aqueous solution for
oxidizing inorganic or organic substances. It works by a chemical reduction process
involving the oxygen, where the oxidation number of the peroxide ion (O1−)
changes to O2− in water product. A balanced equation for its use as an oxidizer is
2 þ 2e ! 2O
O2 2
reduction of peroxide
2A ! 2A 1þ
þ 2e oxidation of A
Usually H2O2 is sold by chemical suppliers in a 30 % aqueous solution, which,

when diluted 10 fold becomes the 3 % solution used as a medical antibiotic
solution. Pure H2O2 may be made by vacuum evaporation [62] of 30 % H2O2
solution, where the water component of the solution exhibits a higher vapor
pressure than the H2O2. Pure hydrogen peroxide is a faint blue syrupy liquid which
freezes at −0.4 °C [62].
The purification of H2O2 by vacuum distillation has been described in a careful
study [63], where the reflection IR spectrum of a thin film of H2O2 was measured as
it evaporated in ultrahigh vacuum from a reflective quartz microbalance surface,
and where in the final stages of evaporation a hydrogen peroxide dehydrate phase
was removed, leaving the least volatile pure H2O2 phase.
An apparatus to purify 30 % H2O2 and then to deliver pure H2O2 vapor into an
ultrahigh vacuum system is shown in Fig. 40.10. The micrometer leak valve con-
nects the purification system with the UHV system. The H2O2 30 % solution is
added through a PTFE (Teflon)-glass stopcock (lubricated only with a halocarbon
grease on the O-ring seals) and this solution is vapor distilled into a scroll pump
containing no organics. Following removal of the major fraction of the liquid by
distillation, pure liquid H2O2 is left in the bulb. The vapor pressure of the H2O2 is
about 0.76 Torr at 13 °C to 7.6 Torr at 45 °C [64]. At these pressures, H2O2(g) can
be easily transported to the UHV chamber using the leak valve.
Because H2O2 readily oxidizes organic compounds, the entire apparatus shown
in Fig. 40.10 should be cleaned initially with the boiling 30 % solution for complete
organic removal [63]. H2O2 taken from a 30 % aqueous solution without vacuum
distillation has been used to grow high quality ZnO epitaxial films [65].
The 30 % H2O2 solution is a strong oxidizer which will cause burns to the skin,
the respiratory tract and to the eyes. The heat of reaction with oxidizable materials
may cause ignition.
Fig. 40.10 Hydrogen peroxide purifier and gas source
References
1. I. Langmuir, J. Am. Chem. Soc. 34, 1310 (1912)

2. I. Langmuir, J. Am. Chem. Soc. 37, 417 (1915)
3. J.N. Smith Jr, W.L. Fite, J. Chem. Phys. 37, 898 (1962)
4. D. Brennan, P.C. Fletcher, Proc. Roy. Soc. London Ser. A 250, 389 (1959)
5. T.W. Hickmott, J. Chem. Phys. 32, 810 (1960)
6. U. Bischler and E. Bertel, J. Vac. Sci. Technol. A11, 458 (1993). W capillary for this source
may be obtained from Metallwerke Plansee GmbH, A-6600 Reutte/Tirol, Austria, Tele: 43
5672 70 0; FAX: 43 5672 70 500. The entire source is available from Physikalisches Buro
Stein-muller GmbH, Rinnerstr. 16d, A-6074 Rinn/Aldrans, Austria
7. K.H. Bornscheuer, S.R. Lucas, W.J. Choyke, W.D. Partlow, J.T. Yates Jr, J. Vac. Sci.
Technol. A11, 2822 (1993)
8. B.A. Von Gottwald, Vakuum Tech. 22, 106 (1973)
9. C.T. Campbell, S.M. Valone, J. Vac. Sci. Technol. A3, 408 (1985)
10. Spiral W filaments in standard or special dimensions may be obtained from CM Furnaces Inc.,
103 Dewey Street, Bloomfield, NJ 07003
11. H. Wise, B.J. Wood, Adv. At. Mol. Phys. 3, 291 (1967)
12. J.P. Toennies, W. Welz, G. Wolf, J. Chem. Phys. 71, 614 (1979)
References 495
13. B.E. Koehler, C.H. Mak, D.A. Arthur, P.A. Coon, S.M. George, J. Chem. Phys. 89, 1709
(1988)
14. V.G. Luppov, M. Mertig, T. Roser, B.S. van Guilder, B. Vuaridel, YuM Melnik, A.F.
Prudkoglyad, Rev. Sci. Instrum. 62, 2738 (1991)
15. B. Van Zyl, M.W. Gealy, Rev. Sci. Instrum. 57, 359 (1986). It is interesting to note that in
these studies, the maximum in H2 dissociation probability is not reached until tungsten
temperatures above 2400 K are employed. This may be due to the presence of gas phase
collisions here, where collisionless conditions generally apply in sources used for surface
studies in ultrahigh vacuum
16. H. Koschmieder, V. Raible, Rev. Sci. Instrum. 46, 536 (1975)
17. A. Hershcovitch, A. Kponou, and TO. Niinikoski, Rev. Sci. Instrum. 58, 547 (1987)
18. K.C. Harvey, C. Fehrenbach Jr, Rev. Sci. Instrum. 54, 1117 (1983)
19. M. Eisenstadt, Rev. Sci. Instrum. 36, 1878 (1965)
20. J. Slevin, W. Stirling, Rev. Sci. Instrum. 52, 1780 (1981)
21. B.J. Wood, H. Wise, J. Phys. Chem. 66, 1049 (1962)
22. W.W. MacAlpine, R.O. Schildknecht, in Proceedings of IRE, 2099 (1959)
23. C.D. Stinespring, A. Freedman, C.E. Colb, J. Vac. Sci. Technol. A4, 1946 (1986)
24. C.E. Kolb, M. Kaufman, J. Phys. Chem. 76, 947 (1972)
25. Microwave sources are available from Opthos Instruments, Inc., 17805 Caddy Drive,
Rockville, MD 20855
26. S. Sibener, R.J. Buss, C.Y. Ng, Y.T. Lee, Rev. Sci. Instrum. 51, 167 (1980)
27. C.D. Stinespring, A. Freedman, Appl. Phys. Lett. 48, 718 (1986)
28. A. Freedman, J. Appl. Phys. 75, 3112 (1994)
29. G.N. Robinson, A. Freedman, R.L. Graham, Langmuir 11, 2600 (1995)
30. A. Freedman, C.D. Stinespring, J. Phys. Chem. 96, 2253 (1992)
31. H. Bonzel, A.M. Bradshaw, G. Ertl, Physics and Chemistry of Alkali Metal Adsorption
(Elsevier, Amsterdam, 1989)
32. J.B. Taylor, I. Langmuir, Phys. Rev. 44, 423 (1933)
33. Getter sources may be obtained from SAES Getters S.p.A., Head Office, Via Gallarate, 215,
20151 Milano, Italy. USA Offices: SAES Getters/U.S.A. Inc., 1122 E. Cheyenne Mtn. Blvd.,
Colorado Springs, CO 80906
34. F. Ricca, P. Delia, Porta. Vacuum 10, 215 (1960)
35. D.L. Schaefer, Rev. Sci. Instrum. 41, 274 (1970)
36. L.V. Hau, J.A. Golovchenko, M.M. Burns, Rev. Sci. Instrum. 65, 3746 (1994)
37. F. Kalos, A.E. Grosser, Rev. Sci. Instrum. 40, 804 (1969)
38. K.N. Walzl, C.F. Koerting, I.M. Xavier, A. Kuppermann, J. Chem. Phys. 86, 88 (1987)
39. F.A. Elder, C. Giese, B. Steiner, M. Inghram, J. Chem. Phys. 36, 3292 (1962)
40. P. Chen, S.D. Colson, W.A. Chupta, J.A. Berson, J. Phys. Chem. 90, 2319 (1986)
41. J. Dyke, N. Jonathan, E. Lee, A. Morris, J. Chem. Soc., Faraday II 72,1385 (1976)
42. X.-D. Peng, R. Viswanathan, G.H. Smudde Jr, P.C. Stair, Rev. Sci. Instrum. 63, 3930 (1992)
43. Alltech Assoc., Inc. 2051 Waukegan Rd., Deerfield, IL 60015
44. R. Renaud, L.C. Leitch, Can. J. Chem. 32, 545 (1954). Proper caution should be employed
working with dimethyldiazine as it is a carcinogenic agent
45. D.D. Berkley, A.M. Goldman, I. Maps, B.R. Johnson, J. Morton, T. Wang, Rev. Sci. Instrum.
60, 3769 (1989)
46. D.D. Berkley, B.R. Johnson, N. Anand, K.M. Beauchamp, L.E. Conroy, A.M. Goldman, K.
Mauersberger, M.L. Mecartney, J. Morton, M. Tuominen, Y.-J. Zhang, Appl. Phys. Lett. 53,
1973 (1988)
47. E. Coleman, T. Siegrist, D.A. Mixon, P.L. Trevor, D.J. Trevor, J. Vac. Sci. Technol. A9, 2408
(1991)
48. G.A. Cook, A.D. Kiffer, C.V. Klumpp, A.H. Malik, L.A. Spence, Advances in Chemistry
Series 21 (American Chemical Society, Washington, DC, 1959), p. 44
49. Matheson Gas Data Book, 6th ed. (Matheson Gas Products, Secaucus, NJ, 1980), pp. 574–577
50. Aldrich #24,982-3; CAS (7631-86-9). Sold by Aldrich Chemical Co., P.O. Box 2060,
Milwaukee, WI 53201
51. Fomblin is a registered trademark of Montedison Specialty Chemicals, Milano, Italy
52. Krytox is a registered trademark of E.I. dupont De Nemours and Co., Inc., Wilmington, DE
53. S. Hosokawa, S. Ichimura, Rev. Sci. Instrum. 62, 1614 (1991)
54. M. Griggs, S. Kaye, Rev. Sci. Instrum. 39, 1685 (1968)
55. S. Anderson, K. Mauersberger, Rev. Sci. Instrum. 52, 1025 (1981)
56. L. Andrews, R.C. Spiker Jr, J. Phys. Chem. 76, 3208 (1972)
57. J.T. Kucera, J.D. Perkins, K. Uwai, J.M. Graybeal, T.P. Orlando, Rev. Sci. Instrum. 62, 1630
(1991)
58. D.H. Byers, B.E. Saltzman, Advances in Chemistry Series, vol. 21 (American Chemical
Society, Washington, DC, 1959), p. 93
59. V. Matijasevic, E.L. Garwin, R.H. Hammond, Rev. Sci. Instrum. 61, 1747 (1990)
60. V.S. Smentkowski, J.T. Yates Jr, J. Vac. Sci. Technol., A 12, 219 (1994)
61. V.S. Smentkowski, J.T. Yates Jr, J. Vac. Sci. Technol., A 12, 224 (1994)
62. W.T. Foley, P.A. Giguere, Hydrogen Peroxide and its analogues II. Phase equilibrium in the
system hydrogen peroxide-water. Can. J. Chem. 29, 123 (1951)
63. M.J. Loeffler, B.D. Teolis, R.A. Baragiola, J. Phys. Chem. B 110, 6911 (2006)
64. CRC Handbook of Chemistry and Physics, 84th edn. (CRC Press, Boca Raton, 2003–2004),
Section 6-71
65. N. Izyumskaya, V. Avrutin, W. Schoch, A. El-Shaer, F. Reuβ, Th. Gruber, A. Waag, J. Crystal
Growth, 269, 356 (2004)
Chapter 41
Electrochemical Sources and Devices
41.1 Electrochemical Sulfur Source
A solid-state electrochemical cell is especially convenient as a source for sulfur

vapor deposition in surface science experiments. The cell was originally designed
by Wagner [1] in the 1950s, and variations on its design have been in use since
[2–7]. In Fig. 41.1a, a cross-sectional view of the cell is shown. It consists of a glass
tube, 2.6 mm in inside diameter at the outlet on the top, which has been made by a
glassblower to have a slightly larger diameter on its bottom end. The waist joint in
the center serves as a stop to hold the contents under compressive force, as will be
described. A pellet of AgI fits tightly into the glass tube and is contacted by a silver
foil cathode electrode on its bottom that is welded to a Pt wire. The Pt wire is
directed through the center of a glass capillary piston and through a spring device
that exerts a compression force on the piston from the bottom. Above the AgI pellet
is a Ag2S pellet, compressed against a Pt ring that is welded to a Pt wire anode that
passes up through the collimating tube. The assembly is contained inside a metal
tube that may be heated by ohmic heating from a resistance wire insulated from the
metal tube and wrapped around it. At a cell temperature of 485 K and a potential
drop of 0.15 V, a maximum flux of 4–5 × 1012 S2 molecules/s was generated. The
literature indicates that S2 is the major gas species generated under these conditions
[5, 6].
The measured angular distribution of S on a Pt crystal is shown in Fig. 41.1b, but
unfortunately the doser-crystal distance was not reported for these measurements
[2]. A Gaussian distribution is observed with Auger spectroscopy, where data
measured in both the x- and y-direction are superimposed on the same graph.
The electrochemical cell operates in ultrahigh vacuum with minimal outgassing
and may be operated with pressures being maintained in the 10−10-Torr range.

DOI 10.1007/978-3-319-17668-0_41
498 41 Electrochemical Sources and Devices
Fig. 41.1 Electrochemical S source. a Electrochemical cell. b Spatial S distribution
41.2 Electrochemical Halogen Sources
Halogens are important surface species in surface chemistry and their delivery to
surfaces in a controlled manner forms an important research technology. This
discussion concerns the delivery of iodine gas by means of electrochemical
41.2 Electrochemical Halogen Sources 499
generation [8], but others have used similar methods for both chlorine and bromine
[7, 9, 10].
The source shown in Fig. 41.2a is a silver iodide pellet. At moderate tempera-
tures, an electrochemical reaction, 2AgI(s) → 2Ag(s) + I2(g), occurs in the pellet as
a result of a potential applied between two electrodes, relying on the ionic mobility
in the heated pellet. The mobile defect concentration may be increased by mixing
about 5 % CdI2 with the AgI before casting the pellet.
The current flow through the heated pellet can be regulated using a constant
current DC source, and a voltage of about 25 V will cause a current of about 25 μA
to flow without external heating. The cells display extremely non-ohmic behavior,
especially during start-up. In order to increase the efficiency of the source, it is
housed in a Pyrex heater tube coated on the outside with a semiconducting SnO2
thin film (see p. 123) having a resistance of about 10–100 Ω/cm of tube.
Alternatively, the Pyrex tube can be heated using a spiral nichrome wire wrapped
around the tube [11]. The temperature of the tube and the pellet is adjusted by
passing current through the SnO2 film, which has a band of Aquadag colloidal
graphite on each end to act as a contact between the conductive film and two
clamps. At a temperature of 390 K, an applied voltage of 20 V produces a stable
cell current of about 100 μA. The cell walls may adsorb I2 during initial operation,
but then stable operating conditions are achieved. Faraday’s laws may be used to
roughly estimate the flux of iodine from the source.
Figure 41.2b shows the output of the electrochemical cell as a function of the
current as measured with a quadrupole mass spectrometer in line of sight, and on
axis with the source. The system pressure was simultaneously measured with an
ionization gauge that did not have line of sight to the source, and it is believed that
it measures mainly I2 gas. The mass spectrum indicates that I+ at 127 amu and I2+ at
254 amu are the main ions generated in the mass spectrometer from I2(g), although
some H2 and Cl2 are also produced. If the pellet is heated to higher temperatures,
near 430 K, the conductivity may be increased to give higher I2 fluxes, but
deposition of CdI2 also occurs slowly. It is also reported that a baffle at room
temperature may be used to avoid deposition of undesirable substances, and that
gaseous iodine can still be deposited on the sample with this arrangement.
To make the AgI/CdI2 button, 7 g of the mixed salt is melted and poured onto
the Pt/Rh mesh that is held on a glass plate under a cylindrical mold made of Pyrex.
The Ag coil is inserted into the top of the melt before solidification. If the Pt/Rh
mesh is heated to redness in a Bunsen burner flame prior to being covered with the
liquid melt, and if it is held flat on the plate, the melt will cover only the exposed
wire and will not submerge the electrode. Alternatively, red-hot Pt/Rh can be fused
into the cold pellet [7]. The pellet can easily be removed from the Pyrex mold after
cooling for insertion into the long support-heater tubing, which acts as a beam
collimator in use.
Similar devices have been made for producing all of the halogens [7], and their
performance is similar to that reported here. It is reported that after many hours of
operation, a dead short often occurs through the pellet. This indicates that a metallic
Ag dendrite has grown inside the pellet; brief passage of a high current through the
Fig. 41.2 Typical electrochemical halogen source. a Cell. b Mass spectrum at 150 °C
41.2 Electrochemical Halogen Sources 501
heated pellet may break the dendrite. In addition, for the Cl2 cell used at excessive
temperatures (>420 K) significant AgCl(g) was evolved. The halogen sources may
be voltage modulated in the frequency range of 0.5–22.5 Hz, and the modulated
halogen beam produced can then be used in kinetic studies on single crystal sur-
faces [9], and useful waveforms may be employed.
The production of pure F2 by electrochemical means has been partially suc-
cessful [10], although the high reactivity of fluorine with materials has caused
problems. A CaF2 pellet, held by springy contacts between a Pt plate and a Pt
gauze, was employed for electrochemical generation at temperatures of 420–450 K.
The gas evolved at the anode was focused onto the sample with a short tube, coated
with CaF2. Small amounts of SiF4 were detected if glass tubing was used. The
source delivers small amounts of platinum fluorides and HF. Studies with this
source of the adsorption of F2 on Pt showed that PtF2 is evolved above about 500 K
from Pt at low coverages, and that PtF4 is evolved at high coverages, in addition to
F(g) at 890 K [12]. The HF yield from the electrochemical cell may be reduced to a
few percent of the F2 yield by outgassing and electrolysis. Currents between 2 and
50 μA were used. During F2 dosing, the total pressure of the vacuum system was
below 1 × 10−9 mbar.
Halogen atom sources have been described and one example is a pulsed source
involving Cl2 dissociation in a microwave cavity. One interesting feature of this
work is the study of 2P3/2- and 2P1/2-Cl atom losses on Teflon walls [13].
A source of halogen ions produced by surface ionization on a LaB6 surface of
low work function has been described [14].
41.3 Electrochemical I2 Doser
Halogen dosers, based on electrochemical generation are popular and have been
treated in Sect. 41.2 of this book. Recently an I2 doser working at low temperature
has been developed which adds to the technology described previously (Fig. 41.3).
Previous dosers required high temperatures to achieve electrical conductivity in the
solid state electrolyte. By using the compound Ag4RbI5, high ionic conductivity is
achieved near room temperature. Ag4RbI5 is prepared and about 0.5 g is com-
pressed into a disk of 2 mm thickness and 7.5 mm width using a hydraulic press at
600 MPa, for 5 min. The disk is made with Pt mesh on the anode side. The disk is
placed between two stainless electrodes and the application of a voltage causes
current to flow with the production of I2(g). The cell had an operating resistance of
about 2.5 kΩ and operates at about 200 µA when dosing. The collimator directs the
I2 flow to the sample. The Ag4RbI5 is stable up to 500 K allowing mild vacuum
system bakeout to be used. The design cannot be extended to similar compounds of
Cl and Br because these mixed salts have insufficient conductivity. Outgassing at
100 µA is recommended prior to opening the shutter for dosing [15].
Fig. 41.3 I2 electrochemical doser
41.4 Electrochemical Device Working in UHV
The application of UHV surface analysis methods to working solid state electro-
chemical devices is of widespread interest. To study working devices, it is necessary
to carry out electrochemical redox processes involving gases, and to control elec-
trical potentials in the device. In addition, since ion mobilities govern device per-
formance, one must be able to control the device temperature during operation.
Figure 41.4a shows a top view of a working device, with a Ni-film deposited
working electrode. The device is mounted on a base plate which is transferrable to
one or more manipulators and which connects to an external potentiostat/galvanostat
used for impedance spectroscopy. Three springy electrodes mounted around the
device exterior make contact at various points, as shown in Fig. 41.4b. The small
contact area of the probes minimizes heat transfer from the heated device. Device
41.4 Electrochemical Device Working in UHV 503
Fig. 41.4 Electrochemical device working in UHV. a Top view. b Side view. c Thermal behavior
heating is achieved by means of ceramic heater button with an embedded resistor. In

Fig. 41.4c, the real component of the device impedance is plotted versus temperature
measured with a pyrometer, and the relative resistance is also shown during heating.
These plots show that good contact is being made over the temperature range studied
and the temperature dependence of the resistance measured is consistent with that of
a Ni film used in the device as the upper electrode, indicating that most of the
impedance is in the film [16].
References
1. C. Wagner, J. Chem. Phys. 21, 1819 (1953)

2. W. Heegemann, K.H. Meister, E. Bechtold, K. Hayek, Surf. Sci. 49, 161 (1975)
3. T.E. Gallon, I.G. Higginbotham, M. Prutton, H. Tokutaka, Surf. Sci. 21, 224 (1970)
4. H. Tokutaka, M. Prutton, I.G. Higginbotham, T.E. Gallon, Surf. Sci. 21, 233 (1970)
5. J.C. Greenbank, B.B. Argent, Trans. Far. Soc. 61, 655 (1965)
6. D. Detry, J. Drowart, P. Goldfinger, H. Keller, H. Rickert, Z. Physik, Chem. NF 55, 314
(1967)
7. N.D. Spencer, P.J. Goddard, P.W. Davies, M. Kitson, R.M. Lambert, J. Vac. Sci. Technol. A1,
1554 (1983). (This paper contains references to a number of arrangements to the
electrochemical generation of halogens, etc)
8. A. Mowbray, R.G. Jones, J. Vac. Sci. Technol. A7, 3373 (1989)
9. W.T. Tysoe, N.D. Spencer, R.M. Lambert, Surf. Sci. 120, 413 (1982)
10. R.G. Jones, Prog. Surf. Sci. 27, 25 (1988). (A general review of halogen adsorption on
surfaces)
11. Professor Philip R. Watson, Department of Chemistry, Oregon State University, Corvallis,
Oregon 97331-4003 (Private communication)
12. E. Bechtold, Appl. Surf. Sci. 7, 231 (1981)
13. W. Müller-Markgraf, M.J. Rossi, Rev. Sci. Instrum. 61, 1217 (1990)
14. N. Kashihira, E. Vietzke, G. Zellermann, Rev. Sci. Instrum. 48, 171 (1977)
15. S.A. Furman, D.A. Harrington, J. Vac. Sci. Technol. A 14, 256 (1996)
16. J.A. Whaley, A.H. McDaniel, F. El Gabaly, R.L. Farrow, M.E. Grass, Z. Hussain, Z. Liu,
M.A. Linne, H. Bluhm, K.F. McCarty, Rev. Sci. Instrum. 81, 086104 (2010)
Chapter 42
Gas Purification
42.1 Gas Purification for Experiments at High Pressures
The use of high-pressure gases for experiments involving single crystal surfaces
always involves the problem of the preferential adsorption on the crystal of trace
impurities present in the gas. At high pressures, even 1 ppm of adsorbable impurity
can result in the selective adsorption of a monolayer or more of impurity. This
problem maybe eliminated by several methods, depending upon the gases and the
impurities involved.
For gases with low condensation temperatures, storage over previously baked
(573 K) zeolite pellets [1] at liquid nitrogen temperature is a favored method for the
adsorption of trace impurities with high condensation temperatures [2]. This is a
favored method for the additional purification of 99.9999 % pure methane, for
example [2]. In addition, carbon monoxide almost always contains traces of Ni
(CO)4 and Fe(CO)5 as impurities produced within the storage cylinder or in the
stainless steel gas handling system, and these metal carbonyls can be removed by
storage of the CO over a wide pore zeolite at liquid nitrogen temperature. See
p. 509 for a discussion of the use of cryogenic storage and distillation at liquid
hydrogen temperature. Alternatively, metal carbonyls can be removed by passing
CO through a copper coil heated to 100 °C, which is sufficient to decompose the
carbonyls, and then through a trap cooled with liquid nitrogen. It is important to
avoid the use of stainless steel tubing following the purification of the CO [3].
Adsorption of active chemisorbing gases in the presence of inactive gas is often
effectively carried out by means of a Ti getter film which is freshly deposited inside
a storage vessel under vacuum [3–5]. See p. 507 for the use of a Ti getter film for
the final stages of purification of He(g) [5]. An apparatus involving the use of
zeolite trapping followed by storage over a Ti film is shown in Fig. 42.1a.
A commercial titanium sublimation pump cartridge is employed in a stainless steel
storage vessel, and gas storage takes place at 77 K.

DOI 10.1007/978-3-319-17668-0_42
506 42 Gas Purification
Fig. 42.1 Gas purification for experiments at high pressure. a Gas purification by adsorption.
b Gas purification over catalyst (high area)
42.1 Gas Purification for Experiments at High Pressures 507
Catalytic chemists have for years used a method involving gas exposure to a
high area catalyst prior to admission to the reactor chamber. This method has also
been applied to Ni single crystals being studied under high pressures of methane [6,
7]. Figure 42.1b shows the configuration of the gas purification apparatus used in
these studies. A U-tube is filled with a Ni catalyst (in this case), which consists of
Ni particles supported on a high surface area oxide support. The particle size ranges
from 0.3 to 0.8 mm diameter. Both ends of the U-tube are plugged with quartz wool
to prevent transfer of dust away from the catalyst. The catalyst is activated by
reduction in flowing H2 at 800 K, and then is evacuated at 800 K to remove H2.
Either methane or nitrogen may be cleaned of impurities such as CO, O2, CO2, and
hydrocarbons (excluding CH4) using this apparatus.
The cleanup of H2 maybe achieved using a supported Cu catalyst in the same
manner. The chemisorption of H2 by Cu is suppressed by working at liquid nitrogen
temperature. The activation of the Cu catalyst is first carried out using H2, and one
must be careful to initially use a low pressure of H2, since the exothermic reduction
of copper oxides may cause a runaway condition. Reduction temperatures of 650 K
are recommended [8, 9].
42.2 Production and Purity Measurement for Extremely

Pure He
The production of He with extremely high purity for use in ultrahigh vacuum
systems is a challenge requiring special gas handling techniques. It must be
remembered that hydrogen is a major impurity in stainless steel ultrahigh vacuum
systems, and that to avoid hydrogen impurities, glass vacuum systems must
therefore be employed. The bakeout treatment of a glass vacuum system differs
from a metal system, and bakeout temperatures of the order of 600 K must be used
[10].
A number of methods for producing pure He have been reported in the literature.
The use of superfluid He leaks [11] and diffusion through Vycor glass [12, 13] have
been used by others, but the diffusion rate for H2 through hot Vycor is about 0.1 of
that for He [11], so that other methods must be used to achieve large factor
reductions in the hydrogen impurity level.
The apparatus [14] shown in Fig. 42.2 was mainly constructed of Pyrex glass
with bakeable metal valves, and it was processed by bakeout to achieve a base
pressure of 1 × 10−10 Torr. The pumping speed in this apparatus is only about 2 L/s.
Operation of the purification system begins with the deposition of Ti getter films on
the inner walls of the two bulbs, G1 and G2, by heating W loops to which Ti wire
segments have been spot welded. The surface area of these films is each about
300 cm2. The Vycor leak is outgassed at 1025 K until the outgassing cannot be
detected using the Bayard-Alpert gauge at 1 × 10−10 Torr pressure. Valve V4 is then
closed and He is effused through the Vycor leak at 975 K. The He used on the high
Fig. 42.2 Production and purity measurement for extremely pure He
pressure side of the Vycor leak is specified to be 99.999 % pure, and it is admitted
to approximately 1 atm pressure, into a gas handling system initially at 10−7 or
10−8 Torr. Typical He effusion rates are approximately 6 × 1014 He atoms/s.
Storage of He takes place overnight at several Torr pressure in the closed system
over the freshly evaporated Ti getter film, G1.
The Rh(111) single crystal was used as an integrating detector of H2 in the
purified He, using temperature-programmed desorption (TPD) with the mass
spectrometer (QMS) as the detector. Purified He was admitted to less than 1 Torr
pressure, with Valve V2 open to the system, and with a freshly deposited Ti film in
getter bulb, G2, in order to continue getter purification of the He during exposure to
the clean Rh(111) single crystal. An exposure of 1000 s was used in the He purity
test, leading to exposures of from 100 to 900 Torr s to the He. Following pumping
with the Hg diffusion pump, followed by use of the ion pump, TPD of both H2 and
CO was carried out, and compared to blank experiments without He. A least
squares fit to the data gave 5.4 × 1010 H/cm2-Torr s (He) for the contamination rate
with the CO contamination rate about an order of magnitude smaller. It was shown
that the rate of H2 adsorption was not limited by gas diffusion kinetics at the He
pressures used. Assuming a sticking coefficient of unity for the H2, these mea-
surements indicate that P(H2)/P(He) = 1.9 × 10−11, which is a factor of about 106
improvement over the original high purity He [14]. Flowing He has been similarly
purified using hot Ti sponge in the flow lines [15].
Superleaks for the control of He flow into a vacuum system have been described
[16]. They are made of porous Vycor glass [17] with approximately 40-Å-diameter
42.2 Production and Purity Measurement for Extremely Pure He 509
pores occupying about 30 % of the volume. By using various diameters of the

Vycor, the He leak rate can be adjusted. A leak made of 1-mm-diameter porous
Vycor, 8 mm long, exhibits a leak rate for He of 4 × 10−6 std. cm3/s at 300 K at
1 atm presure. The leaks may be used also by immersion in liquid He [16].
42.3 Purification of Gases by Cryogenic or Gettering

Methods
The production of high-purity gases for laboratory use in ultrahigh vacuum

applications has been discussed previously in the case of He (see p. 507). Here we
present other methods for the purification of gases such as He, H2, Ar, N2, and O2
[18].
As shown in Fig. 42.3a, a combination of zeolite adsorption and cryogenic
impurity condensation was employed for both He and H2 purification. The appa-
ratus, made of glass, consists of a conventional gas handling system connected to a
cryogenic purification unit. The zeolite trap and the glass wool trap, along with the
collection tube and associated glass lines, were baked out to achieve ultrahigh
vacuum. The cryostat was then filled with either l-H2 or l-He, which causes the
pressure in the tube to drop to <10−12 Torr. Gas from the gas handling system was
then slowly admitted to the cold system, and condensation of highly pure gas was
observed to occur, drop by drop, off of a tungsten field emission tip located in the
collection tube, which was also a field emission microscope. This indicates that the
glass wool section of the apparatus was condensing impurities preferentially from
the incoming H2 or He gas, and that gas phase transport of purified gas occurred
from the glass wool section to the collection tube. The purity of the collected gas
was measured by evaporation of the liquified gas, which filled the collection tube,
and observing that the W field emission tip was still clean after immersion and
reevacuation. Concentrations of chemisorbing impurities were estimated to be
<5 × 1012/cm3 of liquid, considering the diffusion lengths in the liquid. This cor-
responds to an atomic fraction of impurities of the order of 10−14 in the purified gas.
In the case of Ar, N2, and O2 that cannot be purified by the method shown in
Fig. 42.3a, because of the high freezing point of these gases, a second method of
purification was devised, as shown in Fig. 42.3b [18]. Here, the glass vacuum
system was baked to 430 °C to achieve ultrahigh vacuum, and a Ta or Ni film was
evaporated on the walls of the 2-L vessel, after seal-off from the high vacuum line
using a torch seal-off. Ar or N2 was then admitted into the system, and condensed
into the trap at l-H2 temperature, after which the connection to the gas source was
sealed off. Then the condensed gas was allowed to evaporate and condense several
times to facilitate gas contact with the getter film, since diffusion through the gas
will be slow at the * 1 atm pressure employed. The gas could be condensed into
the collection tube using liquid N2-O2 mixtures whose temperature could be
adjusted by using different compositions, and monitored with an immersed
Fig. 42.3 Gas purification

apparatus. a He and H2
purification. b Ar and N2
purification
42.3 Purification of Gases by Cryogenic or Gettering Methods 511
thermocouple. By reestablishing ultrahigh vacuum by condensing the gas into the

trap within the cryostat, field emission measurements showed that no detectable
adsorption had occurred on the field emission tip for the case of Ar. For N2
purification, a Ni getter film was employed, but irreversible chemisorption of N2
prevented the use of field emission to detect impurities. For O2, admission of tank
gas through l-N2 traps was used without a getter film, and field emission mea-
surements were also not possible because of oxygen chemisorptions.
The power of metal getter films for maintaining very low vacuum conditions in
sealed-off glass vacuum systems is clear from observations made with field emis-
sion tips sealed off in the presence of getter films. In one instance, almost no
contamination of the tip was observed in one year, indicating a pressure of the order
of 10−16 Torr or lower [19].
42.4 Gas Purification by Permeation Through Metals
For some applications the supply of extremely pure gases into a vacuum system is
needed, and conventional cylinder sources are inadequately pure for this purpose.
Thus, for years, chemists have used permeation through high-temperature mem-
branes for the removal of impurities from the gas. In addition to supplying pure gas,
the permeation technique also affords easy control of flow rate by the adjustment of
the temperature.
In Fig. 42.4a, either hydrogen or oxygen permeation sources are shown [20]. For
hydrogen, pure Ni permeation tubing is used [21]; for oxygen, pure silver perme-
ation tubing is used [21]. The permeation tubes are 10–13 cm in length. The thin
permeation tubes are silver brazed (in hydrogen) to the ends of 0.25-in.-diameter
copper tubes mounted on ceramic insulator feedthroughs welded into a flange.
Passing current through the device provides the heating necessary to allow gas
permeation to occur. The purity of gases produced in this manner has been studied
by Whetten and Young [22], Young [23], and comparisons are made in the book by
Redhead [24]. The worker should remember that while high gas purity can be
achieved by permeation methods, it is also likely that the gas in a UHV system will
become impure owing to displacement from the walls of the chamber or by
interaction with hot filaments [24]. Studies of Ag permeation sources for the pro-
duction of pure O2 have shown that long periods of operation of the source must
occur before gases like H2S, SO2, CO, and CO2 are eliminated to the percent level
[25, 26]. It may be best to condition a Ag source on a separate vacuum system to
avoid the evolution of the gases into the main system [25, 26].
In Fig. 42.4b, a glass device involving a permeation thimble is shown. For
hydrogen, both pure Ni thimbles and pure Pd thimbles have been used. For oxygen,
pure silver thimbles are employed. The thimble, brazed to a Kovar seal, is centrally
mounted in a glass envelope containing a helical refractory metal heating filament
that heats the thimble by radiation. Permeation of purified gas from the outer
Fig. 42.4 Gas purification by permeation through metals. a H2 or O2 permeation source—ohmic

heating. b permeation thimble—radiation heating
envelope to the inside of the thimble takes place. In both experimental arrange-
ments, shown in Fig. 42.4a, b, it is appropriate to flow the source gas slowly
through the apparatus to replenish gas at the source.
A solid state source of oxygen, followed by permeation through a Ag tube wall,
is described on p. 513.
The use of a pure iron permeation source for purification of hydrogen and
deuterium has been reported in [27–29].
42.5 O2 Source—Solid State 513
42.5 O2 Source—Solid State
Various permeation sources, in which gaseous O2 is purified by diffusion through a

selective metal, are described on p. 511. These sources all involve the use of oxygen
gas, supplied as a gas, to drive the diffusion/purification process. A clever modi-
fication of this idea involves the use of a solid that liberates O2 at the
diffusion/purification temperature, supplying a long-lived source of purified oxygen
without the need for a vacuum system to introduce O2 gas into the permeation
source [30].
Figure 42.5 shows a Ag tube device packed with MnO2 powder. The tube is
sealed by electron beam welding and then welded to two Ag tabs as shown. The
electron beam welding, being done under vacuum, automatically removes most of
the entrained air in the powdered MnO2.
Passing electrical current through the Ag tube raises its temperature first to a
degassing temperature of about 450 °C; when the source is resistively heated to
about 550 °C, the MnO2 begins to decompose, liberating O2 inside the tube that
then diffuses through the Ag, becoming highly purified. The rate of production of
purified O2 increases as the temperature is raised, and gas pressures from about
10−9 to 10−2 Torr can be sustained in pumped UHV systems using this source.
42.6 Reusable Sample Holder for Hydrogen Permeation

Studies
The permeation of hydrogen gas through pure metals and alloys is of technological
and scientific interest. One of the methods used to make permeation measurements
is to electron-beam-weld the metal sample diaphragm to a cylindrical tube end [31].
Another method is to use a sample with machined knife edges that contact a
deformable gasket [32]. Elastomer gasket or O-ring sealing is only useful at room
Fig. 42.5 O2 source—solid state

Fig. 42.6 Reusable sample holder—hydrogen permeation
temperature. The method to be described here differs from the earlier methods, is
simple, and the apparatus is reusable [33].
Two standard Conflat flanges with a standard Cu gasket are used. The sample, in
the form of a circular disk, is mounted inside the Cu gasket, and two Al gaskets are
placed on either side of the sample. The Cu gasket, when bolted in place, protects
the sample from the outside environment. The Al gaskets are plastically deformed
between the sample and the inner region of the flanges, and their thickness should
be properly adjusted to a dimension leading to compression and sealing. In cases
where an incomplete seal at the Al gasket was found, this was attributed to work
hardening of the gasket during manufacture. This could be corrected by annealing
the assembled sample holder for 1–3 h at 573 K, then retightening the bolts. The
sample holder was used up to 623 K, and is shown in Fig. 42.6.
An entirely different method for measuring hydrogen diffusivity in metals,
involving electrical measurements as hydrogen diffuses along the length of a foil, is
described in [33].
References
1. One may use Varian Vacsorb molecular sieve, catalog No. 944-0000, obtainable from Varian
Vacuum Products, 121 Hartwell Avenue, Lexington, MA 02173
2. T.P. Beebe, D.W. Goodman, B.D. Kay, J.T. Yates Jr, J. Chem. Phys. 87, 2305 (1987)
3. Dr. J. Wambach, Danish Technical University, Department of Physics, Building 307,
DK-2800 Lyngby, Denmark (private communication)
4. L. Hanley, Z. Xu, J.T. Yates Jr, Surf. Sci. 248, 265 (1991)
5. J.T. Yates Jr, J.J. Zinck, W.H. Weinberg, Rev. Sci. Inst. 50, 132 (1979)
6. I. Chorkendorff, Technical University of Denmark, Building 307, DK-2800, Lyngby,
Denmark (private communication)
7. I. Chorkendorff, I. Alstrup, S. Ullmann, Surf. Sci. 227, 291 (1990)
8. P.A. Taylor, P.B. Rasmussen, C.V. Ovesen, P. Stoltze, I. Chorkendorff, Surf. Sci. 261, 191
(1992)
References 515
9. P.B. Rasmussen, M. Kazuta, I. Chorkendorff, Surf. Sci. 318, 267 (1994)

(American Institute of Physics, New York, NY, 1993), pp. 417–430
11. M.A. Biondi, Rev. Sci. Instrum. 22, 535 (1951)
12. F.J. Norton, J. Appl. Phys. 28, 34 (1957)
13. J.R. Young, N.R. Whetten, in Transactions of 8th Vacuum Symposium and 2nd International
Congress (American Vacuum Society, New York, NY, 1961), p. 625
14. J.T. Yates Jr, J.J. Zinck, W.H. Weinberg, Rev. Sci. Instrum. 50, 132 (1979)
15. G.B. Hoflund, R.P. Merrill, J. Vac. Sci. Technol. A1, 1560 (1983)
16. M.F. Wilson, D.O. Edwards, J.T. Tough, Rev. Sci. Instrum. 39, 134 (1968)
17. Vycor glass is made by Leaching Pyrex glass in alkali solution. It is available from major
scientific glass manufacturers
18. B. Halpern, R. Gomer, J. Chem. Phys. 51, 1031 (1969)
19. Personal observations
20. L.C. Burton, J. Vac. Sci. Technol. 8, 758 (1971)
21. Type “A” Ni tubing was used; fine (99.99 %) silver tubing was used. Uniform Tubes, Inc., 7th
Avenue, Collegeville, PA 19426
22. N.R. Whetten, J.R. Young, Rev. Sci. Instrum. 30, 472 (1959)
23. J.R. Young, Rev. Sci. Instrum. 34, 891 (1963)
(originally published by Chapman and Hall, 1968; reprinted American Institute of Physics,
New York, 1993)
25. L.C. Beavis, Rev. Sci. Instrum. 43, 122 (1972)
26. E.G. Burroughs, Rev. Sci. Instrum. 40, 33 (1969)
27. M. Arbab, J.B. Hudson, Appl. Surf. Sci. 29, 1 (1987)
28. M. Arbab, J.B. Hudson, Surf. Sci. 206, 317 (1988)
29. M. Arbab, J.B. Hudson, Surf. Sci. 209, 183 (1989)
30. D.L. Schaefer, Rev. Sci. Instrum. 41, 274 (1970)
31. M. Kanno, M. Yamawaki, T. Namba, in Proceedings of the 2nd Japan Institute of Metals
International Symposium, Hydrogen in Metals, Supplement to Trans. JIM 21,105 (1980)
32. E.H. Van Deventer, V.A. Maroni, J. Nucl. Mater. 113, 65 (1983)
33. D.J. Pine, R.M. Cotts, Rev. Sci. Instrum. 55, 614 (1984)
Chapter 43
Liquid Handling in UHV
43.1 Adding Liquid to Atomically Clean Surfaces
As surface chemistry deals with more and more complex systems it has become
necessary to work at high gas pressures as described elsewhere in this book. In
addition, examples are available where work under liquids has been done while
maintaining clean surfaces. The simple addition of a drop of liquid water to an
atomically clean TiO2 single crystal was found to be very difficult when pure water
injection with a syringe through a rubber septum was attempted. Large quantities of
silicon were found on the crystal following injection and evaporation of the water
droplet, and this was thought to be due to silicones present in the rubber seal.
Therefore, a syringe-injection procedure using only a tiny stainless guide tube for
the syringe needle was employed as shown in Fig. 43.1a. The wetting chamber was
baked to achieve ultrahigh vacuum with the syringe seal closed. Upon establishing
an atmospheric pressure of very pure flowing O2 gas (<0.05 ppm hydrocarbons)
(plus a few extra Torr to cause outward atmospheric flow), it was found that
ultrapure water could be injected into the stainless guide tube while O2 gas con-
stantly flowed from the inside to the outside along the loose fit between the
hypodermic syringe and the guide tube wall. Following droplet evaporation on
the atomically clean TiO2 crystal inside the chamber (achieved by reducing the
relative humidity of the flowing gas), and evacuation, the crystal was withdrawn
into the analytical chamber and an Auger analysis was performed. Figure 43.1b
shows the Auger spectrum after a wetting/evaporation cycle. The C Auger signal,
compared to the TiO2 Auger features, showed that less than 0.05 ML of carbon
remained after evaporation. The lower limit of carbon for the cleaned crystal was
about 0.005 ML. The syringe needle and the syringe were treated in boiling H2O2
(aq) for several hours to remove hydrocarbons prior to the water deposition
experiments [1].

DOI 10.1007/978-3-319-17668-0_43
518 43 Liquid Handling in UHV
Fig. 43.1 Adding liquid to atomically clean surfaces. a Apparatus for wetting studies of
ultrahigh-vacuum-prepared TiO2 surfaces. b Proof of surface cleanliness by auger spectroscopy
Reference
1. T. Zubkov, D. Stahl, T.L. Thompson, D. Panayotov, O. Diwald, J.T. Yates Jr, J. Phys. Chem.
B 109, 15454 (2005)
Part VI
UHV Windows
Chapter 44
Spectroscopic Windows
44.1 Mounting Infrared Windows on UHV Systems
For the transmission of infrared radiation to and from an ultrahigh vacuum system
for the purpose of carrying out infrared spectroscopic measurements, it is essential
to employ windows that are highly transparent in the desired spectral region. Often,
alkali halide crystals are used for this purpose. Alkali halides have a small linear
thermal expansion coefficient relative to metals (about 0.1 of the metal) and thus
undergo high stresses under elevated temperature conditions when rigid contact is
made between the window material and the metal. Several window-mounting
designs that place elastomer seals between the window and the metal have been
reported [1–5]. In each of these designs the elastomer yields by means of shear
strains during bakeout relieving the strain in the alkali halide window and thereby
avoiding fracture. The common elastomer in all of the designs is Viton (see p. 108
for studies of gas diffusion through Viton), and this material will stand repeated
bakeout at 200 °C. Two of the designs, shown in Fig. 44.1a, b, involve differential
pumping to avoid leakage and/or diffusion of atmospheric gases through the Viton
seals. The designs shown in Fig. 44.1c, d do not employ differential pumping. The
design of Fig. 44.1c places an Al gasket between two Viton O-rings, eliminating
leakage or diffusion through the outer O-ring. The design of Fig. 44.1d eliminates
differential pumping or seals between the Viton gaskets.
Figure 44.1a is the original differentially pumped design by Hollins and
Pritchard [1], and has proven effective in many different laboratories [2]. The NaCl
or KBr window is retained in a stainless steel flange by an outer clamping flange.
The inner space between the Viton O-rings is differentially pumped to pressures of
the order of 0.1 Torr by a rotary pump, and it is reported that this pumping is
essential for the best system pressure. On the ultrahigh vacuum side of the window,
the O-ring is compressed outside of a loosely fitting retainer ring that keeps it from

DOI 10.1007/978-3-319-17668-0_44
522 44 Spectroscopic Windows
Fig. 44.1 Infrared

spectroscopy Windows
Mounting. a Original design
—outer clamping flange.
b Modified flange design—
zero profile. c Modified
design—no differential
pumping. d Viton gasket
design—no differential
Pumping
44.1 Mounting Infrared Windows on UHV Systems 523
moving inward and also avoids local trapping of gas introduced during assembly. If
the differentially pumped space is let up to air before bakeout, the system pressure
responds immediately; however, after bakeout, the O-ring seals mate more effec-
tively to the metal and window, and only a small degradation of vacuum occurs if
air is introduced into the differential volume. The indicated base pressure after
bakeout is about 5 × 10−11 Torr.
Figure 44.1b shows a modification of the original design in which no extension
beyond the system outer flange occurs (zero profile) [3]. Here, a modified Conflat
flange, equipped with differential pumping, receives an O-ring plus a metal ring
retainer to hold the O-ring in a circular geometry. An IR window is seated on this
O-ring, and a double O-ring plus retainer ring is compressed on the outside of the
window by a threaded thrust ring that screws into the flange opening. The inner
O-ring (1/16-in.) of the pair seals against the window, while a slightly larger
diameter 1/16-in. O-ring seals against a ledge in the flange. The atmospheric
pressure forces the inner O-ring against the outer rim of its groove in the retainer
ring. This device has been employed with KRS-5 and Csl window material, and is
routinely baked at 100–110 °C. Once the seal is made, no effect of pumping out the
differential space is observed, and pressures of the order of 5 × 10−11 Torr are
reported.
Figure 44.1c shows a third method for sealing IR windows to an ultrahigh
vacuum system [4]. The components are constructed of KF 50 flanges, which are
commercially available. The outer section has its groove widened in order to hold
the O-ring. The inner section has been modified by welding tubulation containing
an O-ring groove into a hole cut into the KF 50 flange. A specially designed Al
gasket is placed between the inner and outer flanges before assembly. An IR
window is then pressed between the inner and outer O-ring seals and compression
by three separate clamps results in gasket deformation, leading to a vacuum-tight
seal between the two O-rings, and removing the need for differential pumping.
Bakeout of the window was limited to 150 °C and the O-rings were observed not to
stick to the sealing surfaces. The leak rate was estimated to be smaller than
1 × 10−12 Pa m3 s−1.
Figure 44.1d shows a double seal made of two flat Viton gaskets that seal against
the IR window [5]. A special thrust ring is bolted onto a modified Conflat flange
using six bolts and a small volume at the edge of the IR window is captured by both
gaskets. Bake-out temperatures up to 150 °C have been used, and the base pressure
is reported to be 2 × 10−10 mbar.
All four of these designs must be carefully made to avoid metal contact with the
brittle window material upon compression of the O-ring seals.
A design similar to that of Fig. 44.1a, b has been reported [6] for sealing IR
windows to a glass system.
For some applications, CaF2 single crystal windows1 [7–10] and MgO windows2
[11, 12] may be attached to flanges for IR studies, and these seals require neither
differential pumping nor O-ring-sealing technology. A review of this method may
be found in [13], and an RF heating method for making these seals is given in [14].
Irtran 4 windows (ZnSe, polycrystalline) have been sealed to Voscovit 501 alloy
using Pyroceram 7575 sealant [15].
The use of a lead gasket in conjunction with sealing a NaCl window to a
stainless steel flange has been reported to be successful for ultrahigh vacuum
applications. A specially shaped flange sealing surface and a series of conical
washers acting as compression springs are involved to compensate for differential
thermal expansion. These windows assemblies are bakable to 275 °C, and exhibit a
leak plus outgas rate of <3 × 10−10 atm cm3/s [16].
A simple method involving the sealing of Irtran 2 and Irtran 6 [17] into rotatable
Conflat flanges has been given in [18]. Here, a silicone cement (Gevac [19]) is used
to seal the window to the rotatable insert, and a ring of Al is placed between the
outer flange and the insert in order to apply pressure directly over the underlying
knife edge to prevent flexure of the flange during tightening. This design is superior
to that of [20], which involves flexure of the nonrotatable flange employed.
44.2 Windows for Vacuum UV Transmission

into Ultrahigh Vacuum Systems
The Viton O-ring technology for sealing windows such as KBr, CaF2, and MgO to
ultrahigh vacuum systems involve severe limitations on the bakeout temperatures
that may be employed (see p. 521). Other sealing methods involving, for example,
fused AgCl cements, are susceptible to leak formation upon exposure to ultraviolet
light, and also are incompatible with many metals because of corrosion effects.
A new method, which works for both CaF2 and MgF2 windows, is described
below [21].
Figure 44.2 shows a stainless steel window flange welded to a tubulation that
connects to the ultrahigh vacuum system. A 0.65-cm-wide step is provided to
support the window, which is 1 mm smaller in diameter than the outer diameter of
1
CaF2 windows, sealed directly to Conflat flanges through a thin silver/silver chloride intermediate
sealing assembly, may be obtained commercially from Harshaw Crystal and Electronic Products,
Solon Technologies, 6801 Cochran Road, Solon, OH 44139. These windows are unsatisfactory for
use with ultraviolet light because of the sensitivity of the AgCl cement to photodecomposition.
A description of the sealing process is given in [7–10].
2
MgO windows have been sealed to Pyrex glass sleeves using gold/silver chloride sealing with the
Pyrex being sealed by a Kovar graded seal to stainless steel flanges. For a diagram of this method,
see [11, 12].
44.2 Windows for Vacuum UV Transmission into Ultrahigh Vacuum Systems 525
Fig. 44.2 Vacuum UV Window—UHV
the step. An outer window retaining ring presses the window against the seal, and
the pressure at the seal can be adjusted using bolts which compress springs around
the perimeter. A soft aluminum foil cushion is used between the retaining ring and
the outer part of the window.
The vacuum seal is made by building up a Vacseal [22] bead on both the
window and the step surface of the flange as shown in Fig. 44.2. The Vacseal bead
is not considered to be a leak sealer between the window and the flange, but a
deformable gasket material. The following recipe is recommended for the pro-
duction of a high temperature seal which is He leak tight:
1. Clean and degrease the window and the flange.
2. Apply Vacseal as shown to the step and window surfaces as shown, keeping
both surfaces horizontal. A 1.5-2.5-mm-thick bead is formed on each surface.
3. Place flange and window into an oven at room temperature and heat to 125 °C
for 5 h. Then, following this precuring procedure, allow to cool to room tem-
perature. There should be no bubbles in the beads.
4. Invert window and carefully place on the flange. The two beads should coalesce
into a single seal about 2–2.5 mm in thickness without voids or bubbles.
5. Gently tighten the retainer ring but do not displace the Vacseal sealing material.
6. Cure at 150 °C for 15 h. Cool to room temperature and inspect for cracks or
voids. The sealant should be amber in color. The spring-loaded bolts may be
tightened slightly in this step.
7. Place the seal into the oven at 200 °C for 8 h. Cool and leak check.
If voids or cracks appear during the above procedure, start over. Methyl ethyl
ketone (MEK) may be used to dissolve the cured Vacseal. Ethyl alcohol will not
dissolve the seal and may be used to clean the window without causing a leak to
form. One must be careful to not apply mechanical stresses to the cured window
assembly. The seals are bakeable to 330–400 °C and are leak-tight for periods in
excess of 6 months [21]. This sealing technique has been used on windows of
7.5-cm diameter [21]. The optical distortion of the widows is small, and thermal
cycling may be carried out.
Another method, useful for mounting quartz windows on Al flanges, has been
reported [23]. Here, an indium seal is employed and the window may be baked to
120 °C, which is the bakeout temperature needed for Al ultrahigh vacuum systems.
44.3 Construction of Be Windows for Ultrahigh

Vacuum Use
Thin Be windows are often employed for transmission of X-rays into UHV sys-
tems. Although electron-beam-welded windows are commercially available [24],
brazing onto stainless steel Conflat flanges is sometimes used [25, 26]. The pro-
cedure for doing this is reported here [26].
Figure 44.3a schematically shows the mechanical procedure employed for
making an ultrahigh vacuum brazed seal to type 304 stainless steel. Beryllium foil,
0.25–0.50 mm thick, must be specified to be of vacuum quality [27]. Such foils may
be used for spans up to 20 cm in length. The stainless steel surface is abraded with
250 grit sandpaper, and all sharp edges on the vacuum side are rounded. A brazed
region 6–9 mm wide is employed. The Be foil is shaped in exterior dimensions by
etching, using common fingernail polish to protect the regions to be preserved. The
protective layer can later be removed using acetone solvent. The acid etching
mixture is specified in Table 44.1. The braze material is 99.99 % Ag foil of
0.075-mm thickness, cut into a ring. The silver should be cleaned with silver polish
or by means of a chemical bath specified in Table 44.2, in which metallic Al is used
to reduce oxidized and sulfided Ag layers. The cleaned components are loaded into
a stainless steel cylinder and the lower flange is tack welded to the cylinder. The
stacked assembly is covered with quartz cloth and loaded with a weight as shown in
Fig. 44.3a. The assembly is fired in a vacuum furnace according to the procedure
outlined in Table 44.3.
He leak detection indicates no leaks at the sensitivity of the detector,
10−11 mbar L s−1.
The transmission of X-rays through Be foils of different thicknesses is shown in
Fig. 44.3b.
44.3 Construction of Be Windows for Ultrahigh Vacuum Use 527
Fig. 44.3 Beryllium Window Fabrication. a Brazing Be windows to stainless steel. b X-ray
transmission through Be
Table 44.1 Be etching HNO3 2 % by volume

solution (0.5 mm/min)
H2SO4 2 % by volume
HF (aq) 2 % by volume
H2O Balance
Table 44.2 Silver cleaning H2O 1 L, boiling in beaker

NaHCO3 15 cm3 of powder
NaCl 15 cm3 of powder
Add large piece of Al foil
Add Ag ring—boil 1–2 min
Remove Ag ring—clean in H2O
Table 44.3 Brazing Evacuate to 10−4 Torr

procedure
Preheat to 850 °C, 45 min
Stabilize at 900 °C, 25 min
Braze at 960 °C, 10 min
Cool
Vent furnace, leak check Be seal
44.4 Bakeable Aluminum Ultrahigh Vacuum Window
Low atomic number elements are often used for windows to transmit X-rays or
high-energy particle beams. A very simple bakeable aluminum foil window
mounting design is shown in Fig. 44.4 [28]. The Al foil window, cut into a circle, is
mounted between two Cu gaskets that are used to make vacuum seals to two
Conflat flanges. The window material is Alcoa 3003 aluminum with an H19 temper.
The only special machining required is to put a radius, R, onto the inside edge of the
Cu gasket placed on the vacuum side of the window. The radius R should equal the
thickness of the window material, T.
Fig. 44.4 Conflat Flange seal for Al foil window

44.4 Bakeable Aluminum Ultrahigh Vacuum Window 529
Both 0.05- and 0.13-mm Al foil of 3.81-cm diameter and 10-cm diameter exhibit
a leak rate for He of <5 × 10−10 std. cm3 s−1. Foils of 0.13-mm thickness have been
baked at 260 °C overnight with one side at atmospheric pressure without devel-
oping leaks; foils of 0.05-mm thickness have been baked at 180 °C overnight
without developing leaks. This sealing technique has been shown to be unsuc-
cessful with Be windows.
44.5 Metal Thin Foil Windows—Welding
Very thin metal foils may be welded to a stainless steel flange using electron beam
welding in vacuum. However, there is danger of damaging the foil because of the
lack of heat conduction away from the fusion region. A method [29] for successful
welding is shown in Fig. 44.5.
The foil, which may be of Cu, Ni, or stainless steel, is several microns in
thickness. It is placed on the flange and a backup washer of 304 stainless steel
(0.05–0.13 mm thick, made by photoetching from sheet material) is placed on top
of the foil, as shown. A hold-down fixture and clamps are used. The electron beam
(120 kV, 250 μA, 0.13-mm diameter) is directed as shown in Fig. 44.5, and the foil
is fused to the flange, after which the washer may be removed. A rotating table,
with a velocity of 76 cm/min at the weld point is used.
If the foil initially contains wrinkles, it will flatten as a result of welding at the
circumference. This is thought to be due to radial forces exerted at the circumfer-
ence during cooling of the weld puddle.
44.6 Preparation of Large-Area Si Single Crystal

Windows
The preparation of large windows of uniform thickness made of single crystal

silicon could be useful in many ways, for example, in X-ray lithography, in
energy-loss particle detectors, in Schottky diodes, and Josephson junctions, in ion
beam analysis, and in in situ transmission electron microscopy. A variety of thin-
ning methods have been used [30], and the description here is concerned with
preparing such windows in a thickness range from 1000 to 20,000 Å [30]. The
method has been demonstrated for crystals of (100), (111), and (110) orientation
[30].
Figure 44.6a shows the basic sequence of steps employed. Device-grade silicon
wafers with a high-quality surface finish (e.g., Syton polish), of either p-type or
n-type with 1–10 Ω-cm resistivity are subjected to boron diffusion in a furnace
operating at 1050 °C with a planar BN diffusion source and nitrogen gas as the
carrier gas. Roughly speaking, a diffusion time of 15 min will give a B penetration
Fig. 44.5 Metal thin film window—welding
depth that is appropriate for making a window of 2000–2500-Å thickness. The

diffusion time increases quadratically as the desired window thickness. The boron
glass deposited on the surface is later removed by dipping into boiling HNO3
(aq) followed by HF (aq), causing controlled crystal thinning. The silicon wafer is
then cleaved into smaller pieces by a diamond scriber. Flat quartz discs are then
used to support the wafer sections, using a paraffin dam around the perimeter of the
Si sample. The wafer with the polished side down is then placed on the dam, and
sealed on the edge by pressing down a piece of Teflon tape [31] onto the top of the
wafer. The tape has a hole cut into it for the window etch. This protects the shielded
regions of the wafer from etch solutions, as shown in Fig. 44.6b.
Two etch procedures are used. A preetch, which is fast, removes the majority of
the silicon, at a rate of 25 μm/min. This consists of a mixture of the following
44.6 Preparation of Large-Area Si Single Crystal Windows 531
Fig. 44.6 Making thin Si

windows by etching.
a Sequence of steps—
schematic. b Structure before
preetch. c Preetch. d Selective
etch
solutions: HF(aq) (48 % concentration); HNO3 (70.4 % concentration), and glacial

acetic acid, CH3COOH (100 % concentration) in the volumetric ratio of 2:5:2. The
etch rate can be increased or decreased by adjusting the HF concentration.
For the preetch procedure, as shown in Fig. 44.6c, the quartz-Si assembly is
placed in a Teflon or polypropylene beaker (250 ml), which may be rotated and
which contains a gas jet bubbler (plastic) supplying oxygen or nitrogen gas for
agitation. The rotation of the beaker plus agitation by bubbling often gives uniform
etching conditions, and also detaches bubbles that may form on the Si surface.
Thinning to about 0.0025-cm thickness leads to a transparent film that will appear
to be reddish by light transmission, indicating that the preetching should be
immediately stopped. The Teflon tape is then carefully removed using a razor blade
working from the outside, or using boiling trichloroethylene to dissolve the adhe-
sive on the tape. Care must be taken not to disturb the fragile Si window.
The selective, slow etching procedure is done with a mixture of pyrocatechol
(C6H4(OH)2), ethylene diamine (NH2(CH2)2NH2)-(EDA), and water in the ratio
3 g:17 ml:8 ml. This etching solution does not attack Si if the B doping exceeds
7 × 1019/cm3. Etching rates for Si(111), Si(100), and Si(110) are found to be 3, 50,
and 30 μm/h, respectively. The selective etch is carried out as shown in Fig. 44.6d,
by leaning the quartz-Si assembly against a glass slide and operating at 110 °C.
A curved glass cover (watch glass) should be placed over the boiling solution to
condense the distillate and return it to the main etch solution. The EDA is the
etching agent.
With Si(111) crystals, the EDA etching rate is slow, and lateral etching around
the edge of the crystal will dominate, giving a noncircular boundary. This can be
prevented by arranging in the preetch to have nonuniform etching by lowering the
gas flow rate which leads to preferential etching in the middle of the window. Under
these conditions, the EDA etching will start at the thinnest central region and will
proceed laterally toward the edge, etching the non-(l 11) planes which exist there.
The etching is stopped before the etching reaches the window boundary. This
process leads to oval/round etched shapes which are mechanically more stable than
etch configurations with sharp corners.
Modification of the highly doped windows is possible after etching. The B
concentration may be lowered [32], and the window may probably be used as a
seed to grow Si layers by CVD processes [30]. The thin Si windows can be made
amorphous by ion implantation up to 1/3 of the thickness [30]. The window can be
heated to 1000 °C without breaking. Layers of SiO2 and metal silicides can also be
grown at temperatures up to 850 °C [33, 34].
44.7 Strain-Free Optical Window Mounting
The effect of birefringence in strained optical windows can be a significant per-

turbation in ellipsometric measurements made inside of an ultrahigh vacuum sys-
tem. This strain can be caused by the distortion of a flange during tightening of the
44.7 Strain-Free Optical Window Mounting 533
Fig. 44.7 Strain-free optical window mounting-section
mounting bolts that can be transmitted to the window. A re-entrant design such as
that shown in Fig. 44.7 can eliminate this type of strain [35].
To construct this mounting, the glassblower will start with a Kovar-to-glass seal
that has been welded into the flange. This seal first has glass matched to the thermal
expansion coefficient of Kovar alloy, and then a graded seal leading to Pyrex on the
outer end of the seal. The Pyrex is fashioned into an S-shaped reentrant seal, with
the window mounting flared and polished flat to receive the window and its sealant.
The length of glass in the S-section is sufficient to eliminate the transmission of
stress to the window when the flange is tightened.
Glass-to-metal seals other than Kovar are sometimes used in these applications3
[36, 37].
References
1. P. Hollins, J. Pritchard, J. Vac. Sci. Technol. 17, 665 (1980) (this report also contains
references to earlier window sealing methods)
2. Z. Xu, J.T. Yates Jr, J. Vac. Sci. Technol. A8, 3666 (1990)
3. R.G. Tobin, C. Chung, J.S. Luo, J. Vac. Sci. Technol. A12, 264 (1994)
4. C. Damerow, W. Erley, J. Vac. Sci. Technol. A5, 2974 (1987)
5. A. Bradshaw, private communication. The gaskets may be obtained from Busak and
Sham-ban GmbH & Co., Handwerkstr. 5–7, 70565 Stuttgart, Germany
6. R. Raval, S. Munro, A. Coats, G. Beange, J. Vac. Sci. Technol. A10, 3593 (1992)
7. D.D. Berger, Rev. Sci. Instrum. 53, 1619 (1982)
3
Glass-to-metal sealing methods are described in [36, 37].
8. R.W. Roberts, J.F. Harrod, H.A. Poran, Rev. Sci. Instrum. 38, 1105 (1967)
9. B. Caras, Rev. Sci. Instrum. 39, 1441 (1968)
10. Y.-C. Chiu, Rev. Sci. Instrum. 41, 639 (1970)
11. J.T. Yates Jr, D.A. King, Surf. Sci. 30, 601 (1972)
12. R.W. Roberts, J.F. Harrod, H.A. Poran, Rev. Sci. Instrum. 38, 1105 (1967)
13. Y.-C. Chiu, Rev. Sci. Instrum. 41, 639 (1970)
14. M. Aslam, Rev. Sci. Instrum. 42, 1892 (1971)
15. P.E. Wierenga, G.J. Mollenhorst, A.T.B. Wikkerink, Rev. Sci. Instrum. 49, 408 (1978)
16. T.J. Manuccia, J.R. Peele, C.E. Geosling, Rev. Sci. Instrum. 52, 1857 (1981)
17. Irtran optics are available from Eastman Kodak Co., 5/81 RC MC 02010, Rochester, NY
14650-2010
18. M. Kottke, R.G. Greenler, Rev. Sci. Instrum. 42, 1235 (1971)
19. A silicon resin, type KL5, like that used in Ref. [14] can be obtained from Kurt J. Lesker
Company, 1515 Worthington Avenue, Clairton, PA 15025-9911
20. L.J. Schkolnick, Rev. Sci. Instrum. 39, 122 (1968)
21. C.H. Muller III, M.W. Barrett, D.D. Lowenthal, Rev. Sci. Instrum. 59, 1425 (1988)
22. Vacseal is sold as a viscous liquid in 3 oz. bottles. Space Environment Laboratories, Box
1061, Boulder, CO 80306
23. H. Saeki, J. Ikeda, H. Ishimaru, Vacuum 39, 563 (1989)
24. Electrofusion Corp., 34325 Ardenwood Blvd., Fremont, CA 94536
25. P.L. Hartman, Rev. Sci. Instrum. 57, 2564 (1986)
26. G. Navrotski, J. Vac. Sci. Technol A12, 262 (1994)
27. Brush-Wellman, Elmore, OH 43416
28. J. Edgecumbe, Rev. Sci. Instrum. 37, 1419 (1966)
29. P. Fowler, J.V. Minges, Rev. Sci. Instrum. 40, 841 (1969)
30. N.W. Cheung, Rev. Sci. Instrum. 51, 1212 (1980)
31. Obtained from Connecticut Hard Rubber Co., P.O. Box 2560, Meriden, CT 06450
32. C.L. Huang, T. van Duzer, IEEE Trans. Electron Devices ED-23, 579 (1976)
33. N.W. Cheung, L.C. Feldman, P.J. Silverman, I. Stensgaard, Appl. Phys. Lett. 35, 859 (1979)
34. N.W. Cheung, M.-A. Nicolet, J.W. Mayer, in Proceedings of the Symposium on Thin Film
Phenomena- Interfaces and Interactions, vol. 80.2, ed. by J.E.E. Baglin, J.M. Poare (The
Electrochemical Society, Princeton, 1980)
35. Professor David Waldeck, Department of Chemistry, University of Pittsburgh, Pittsburgh, PA
15260 (private communication). The design originated with Professor C.B. Harris’ group at
the University of California, Berkeley, CA
36. E.L. Wheeler, Scientific Glassblowing (Wiley Interscience, NY, 1958), p. 157
37. W.E. Barr, V.J. Anhorn, Scientific and Industrial Glass Blowing and Laboratory Techniques
(Instrument Publishing Co., Pittsburgh, 1947), p. 116
Chapter 45
Observation Windows
45.1 Shielded Observation Window
In ultrahigh vacuum evaporation of materials to produce thin films, the accumu-

lation of deposits on observation windows can be detrimental to temperature
measurements using an optical pyrometer. A calibration procedure using a constant
temperature internal source has been described earlier (see p. 411), and this method
will work for thin film deposits. However, a preferable procedure involves shut-
tering the window so that short observations can be made; then the window can be
protected during most of the evaporation period. An internal rotatable shutter
provides one way to do this, and the design shown in Fig. 45.1 shows another
method [1].
A Pyrex window is sealed onto a glass-metal seal that involves either Kovar
alloy or stainless steel tubing mated to the glass. The tubing is welded to a formed
metal bellows connected to a flange as shown. A retention mechanism, held in place
by a wing nut, allows the bellows to be bent so that the window is out of line of
sight from the evaporation source during most of the evaporation period. For
observation, the retention mechanism is removed and the window then aligns with
the source. One should use formed bellows, not welded bellows, for constructing
this device.
Alternatively, a disposable window, inserted into the chamber, can be used to
protect observation windows [2].

DOI 10.1007/978-3-319-17668-0_45
536 45 Observation Windows
Fig. 45.1 Shielded observation window
45.2 Heated Internal Window for Prevention of Film

Deposition
External windows used for observation of evaporation sources may be protected

from film deposition internally by using a heated window to intercept the evapo-
rating material. If the temperature of the internal window is sufficiently high to
cause evaporation, optical measurements maybe made through the two windows
without interference from film deposition.
Figure 45.2 shows schematically an arrangement for heating a quartz internal
window to temperatures sufficiently high the prevent As deposition during MBE
growth of GaAs, for example [3]. The inner quartz window is heated ohmically
through a Ti film deposited around the edge of the quartz disk on both faces to a
thickness of about 1.5 μm. The electrical resistance of the film, when contacted at
45.2 Heated Internal Window for Prevention of Film Deposition 537
Fig. 45.2 Windows heated windows
points opposite each other, is about 2 Ω at room temperature and about 10 Ω at the
operating temperature. Electrical contact to the Ti film is made using W clips made
from 0.25-mm-diameter wire and protective Ta foil pads prevent abrasion of the Ti
film by the W clips. Electrical heating at a power level of about 20 W maintains the
temperature around the edge of the quartz window at 550–600 °C, and the tem-
perature in the center is estimated to be 350–400 °C, based on a calculation for
radial heat transfer [4]. It was reported that evaporation of over 1000 g of As over
1 year was carried out using one protective quartz window, and it was found that
the optical transmission of the internal window only decreased by 5 %, corre-
sponding to a temperature error of only 2° for pyrometry measurements at 600 °C.
For GaAs deposition, it was found that the heated window did not introduce
impurities into the films produced by molecular beam epitaxy.
Alternatively, the use of disposable internal glass windows has been employed
successfully for this purpose [5].
45.3 UHV Glass Window Seal—Low Optical Distortion
Glass seals in Conflat flanges are generally made by a brazing technique that
distorts the window, limiting its usage for certain optical studies. In addition, when
optical coatings are applied to flange-sealed glass windows, the high temperature of
the coating process may further distort the window. A number of alternative win-
dow seals have been described [6–10] but often require considerable modification
of the standard flanges.
In Fig. 45.3, a modified Conflat flange and gasket are employed to seal BK 7
[11] glass into an ultrahigh vacuum chamber [12]. The window material can range
from high precision optical flats to window glass. These glass blanks can be
optically coated if desired. A commercial OFHC Cu gasket is machined to have its
own knife edge. The window seals the gasket by deforming the Cu knife edge. The
dimensions shown are somewhat arbitrary. Pressure is applied to the window seal
using a rotatable bolt ring which may be modified to increase the area which applies
force to the window. Twenty sheets of 0.001-in.-thick Al foil are used as a cushion
between the retaining ring and the glass. Alternatively either Cu or Pb wire may be
used for the cushion.
The bolts are tightened uniformly and slowly until no leak is observed with a He
leak detector. For a 4.5-in.-diameter flange, torques of about 3 ft. lb/bolt are
employed for sealing. The technique forms seals with less than 10−10 atm cm3 s−1
Fig. 45.3 Windows UHV glass window seal—low optical distortion

45.3 UHV Glass Window Seal—Low Optical Distortion 539
He leak rates. Baking to 250 °C may open leaks, and these can be eliminated by
retightening the flange bolts. A test for stress-induced birefringence showed that
only 0.01 % of 851-nm light passed through two crossed linear polarizers with a
window mounted between them.
It is important to keep the Cu knife edge at least 0.1 in. from the edge of the glass
to prevent breakage, and this may necessitate offsetting the Cu knife edge from the
underlying stainless steel knife edge, as shown in Fig. 45.3 [12].
45.4 Zero-Length Window Assembly
Several methods for mounting windows on ultrahigh vacuum systems are available,
but the method shown here is the most simple, requiring only standard components.
As shown in Fig. 45.4, a window is retained between two Viton O-ring seals, using
a rotatable outer flange without its normal insert [13]. The assembly is mounted
onto a normal Conflat flange attached to the vacuum system. The inner seal may be
Fig. 45.4 Zero length window

either a O-ring or a special Viton gasket manufactured by Varian for use in place of
copper gaskets. The upper bakeout temperature for the Viton is 150 °C. The general
design is also applicable to larger flanges.
References
1. A. Chambers, J. Vac. Sci. Technol. A7, 259 (1989)

2. Dr. W.A. Egelhoff, National Institute of Standards and Technology, Gaithersburg, MD
3. D.E. Mars, J.N. Miller, J. Vac. Sci. Technol. B7, 682 (1989)
4. D.E. Mars, J.N. Miller, J. Vac. Sci. Technol. B4, 571 (1986)
5. Dr. W.F. Egelhoff, National Institute of Standards and Technology, Gaithersburg, MD
6. T.J. Manuccia, J.R. Peele, C.E. Geosling, Rev. Sci. Instrum. 52, 1857 (1981)
7. C. Damerow, W. Erley, J. Vac. Sci. Technol. A5, 2974 (1987)
8. C.H. Muller III, M.W. Barrett, D.D. Lowenthal, Rev. Sci. Instrum. 59, 1425 (1988)
9. H. Saeki, J. Ikeda, H. Ishimru, Vacuum 6, 563 (1989)
10. M.T. Duignan, D.J. Garhardt, W.T. Whitney, Rev. Sci. Instrum. 60, 3537 (1989)
11. BK 7 blanks are supplied by Newport Optical Glass Inc., Costa Mesa, CA, and by Glassfab
Inc., P.O. Box 1880, Rochester, NY 14603
12. A. Noble, M. Kasevich, Rev. Sci. Instrum. 65, 3042 (1994)
13. L.A. Biagi, J. Vac. Sci. Technol. 8, 464 (1971)
Part VII
Surface Preparation Methods
Chapter 46
Cleaning Metal and Semiconductor
Crystals—Examples
46.1 Surface Segregation and the Cleaning of Metal

Single Crystals
The preparation of clean single crystals of the metals is fundamentally important in

surface science research, since surface properties are often highly dependent on
poisons and promoters that are present. It is commonly known that bulk impurities
present in zone-refined single crystals at the ppm level can often segregate to the
surface upon heating. Based on the phase diagram of two component alloys, Burton
and Machlin [1] gave an easy method to predict whether a solute would segregate to
the surface upon heating, as described below.
In Fig. 46.1a, two typical phase diagrams are shown for alloys, where the region
between the solidus (dashed curve) and the liquidus (solid curve) is a two-phase
region. In phase diagram (a), at some constant % solute, if the solid is warmed to
the melting point on the dashed curve, a liquid depleted in solute will be formed. In
contrast, in phase diagram (b), melting of the alloy solid will produce a liquid phase
that is enriched in solute. The authors point out that many of those qualities that
distinguish a liquid from a solid—namely, lower symmetry and lower coordination
—also distinguish the surface from the bulk, so that the tendency for the liquid to be
enriched or depleted of the solute on warming may apply also to the surface of an
alloy. This expectation was confirmed for 16 alloys, where alloys with almost
coincident liquidus and solidus curves were categorized as type (b′), and behaved as
type (a) systems, and exhibited no surface segregation.
If methods can be found to stabilize a solute on a surface by reaction to form a
stable compound or to remove the solute from the surface as a volatile species, then
it is possible to deplete the bulk phase of this solute. This is commonly achieved by
using oxygen as a reactant, producing a volatile oxide that enhances the concen-
tration gradient of the solute, allowing more solute to diffuse to the surface where it
is removed by reaction. In combination with sputtering, surface segregation is

DOI 10.1007/978-3-319-17668-0_46
544 46 Cleaning Metal and Semiconductor Crystals—Examples
Fig. 46.1 Surface segregation and cleaning of metal surfaces chemically. a Typical phase diagram
for alloys. b Si segregation on Pt. c Si segregation in vacuum and in O2 on Pt. d Silicon
segregation on Pd. e B-O vibrational modes on Rh(111) after reaction at 970 K in O2
46.1 Surface Segregation and the Cleaning of Metal Single Crystals 545
effective in purifying the bulk (or at least the near surface region) so that bulk
impurities are not available at the surface during an experiment.
A description of the surface cleaning methods employed for a total of 74 ele-
ments has been given in a very useful review by Musket et al. [2].
Several examples of the use of chemical methods for cleaning surfaces in this
manner are presented below to illustrate the general principles involved and the
types of measurements used to follow the cleaning process.
Platinum
Elements such as Si, Ca, Al, S, C, and P are often found in zone-refined Pt crystals,
and an excellent review of and insight into this problem has been provided in the
work of Bonzel et al. [3–5]1. Si segregates strongly to the surface of Pt, but this was
not appreciated until the early 1980s because the prominent Si (92 eV) feature is
almost coincident with a Pt(94 eV) feature. This may be seen in Auger spectra in
Fig. 46.1b, where heavy Si segregation is observed under equilibrium conditions at
873 K, and less heavy segregation is observed at higher temperatures. Spectrum
(a) is of the freshly sputtered surface (Si free) and the small Pt Auger feature at
about 94 eV is evident. Studies of the effect of O2 on the rate of Si segregation to
the surface show that at 10−7 mbar and at Pt temperatures of 873 K, no enhance-
ment of the Si segregation occurs in the presence of O2, as shown in Fig. 46.1c.
Studies at higher temperature [3–5] indicate that bulk SiO2 will form for
T > 1000 K. It is pointed out that the Auger intensity of the O(KLL) feature,
associated with Si atoms, is the best monitor of the amount of Si present after
oxidation. Platinum oxides are thermodynamically unstable in vacuum at the
temperatures discussed here. The properties of the oxide on silicon-contaminated Pt
first led investigators to postulate the presence of a “platinum oxide” with unex-
pected thermodynamic stability on Pt surfaces [3–5].
Palladium
Similar results to those found for Pt are reported for Pd [6]. Here, too, Si-stabilized
oxides are formed under conditions where PdO would be thermodynamically
unstable, as shown in Fig. 46.1d. The oxidation occurred at 1073–1273 K in
10−6 mbar O2 for spectra (a) and (b). Argon ion sputtering was used prior to
acquiring spectrum (c).
Rhodium
A common impurity in single crystal Rh is boron, which is difficult to remove by
sputtering [7, 8]. It has been found that heavy oxidation is effective in removing
boron by the formation of a boron oxide species that is best detected by
high-resolution electron spectroscopy (HREELS) [9]. The interference between Rh
and B Auger spectral features makes Auger spectroscopy impractical for seeing
small amounts of B on Rh. Figure 46.1e shows an HREELS spectrum of the B-O
vibrational modes obtained when Auger spectroscopy was ineffective in detecting B.
1
Earlier important studies of this problem include [4].
46.2 Chemical Method to Measure and Remove

Carbon from Pd
The high thermodynamic energy of NO compared to O2 may be used to oxidize

impurities from single crystal surfaces as well as from the walls of ultrahigh vac-
uum systems as was described on p. 87.
In a special case, involving carbon impurities purposely introduced to a Pd
single crystal, this effect was employed not only to clean the Pd crystal of carbon
but also to provide a rather sensitive method for showing that carbon had been
removed [10]. Carbon on Pd is difficult to measure by Auger spectroscopy because
of the overlap of the C(KLL) Auger line with the Pd(MNN) line. Figure 46.2a
shows a series of Auger spectra for increasing exposure to C2H4 using quantitative
exposure conditions. A careful examination of the Auger spectra shows that small
changes occur, and a ratio measurement shown in Fig. 46.2b indicates that the
(C + Pd)/Pd Auger ratio does increase slightly following C2H4 exposure, but the
poor signal-to-noise ratio on this number makes a quantitative measurement
impossible.
The carbon may be quantitatively reacted away using NO as an oxidizing agent
in much the same fashion as was done by others using O2 [11] to oxidize the C as
well as to measure the presence of C using measurements of the CO and CO2
formation in thermal desorption. In the case of the NO being employed as an
oxidizing agent, using isotopically labeled 15NO permits the discrimination of 15N2
(30 amu) product from CO (28 amu product) and hence to supply a quantitative
basis for measuring the coverage of carbon from the 15N2 formation. Figure 46.2c
shows the thermal desorption spectra for 30 amu species produced after treating
carbon-covered Pd with 15NO. Successive experiments were done with higher
levels of carbon, deposited by quantitative doses of C2H4. The cross-hatched
portion of the thermal desorption spectra is due to 15N2 from the reaction 2C
(a) + 2NO(a) → 2CO + N2. The sensitivity of the method for low carbon levels
may be judged by the fact that C2H4 exposures much less that 1 × 1014 cm−2 give
measurable 15N2 yields in the high-temperature peak. The low-temperature
desorption peak, also from 15N2, originates from the normal dissociation of NO
on the Pd surface.
Using this measurement method for determining the presence of small amounts
of carbon on Pd, a cleaning procedure involving exposure to 3 × 1016 NO cm−2 at
510 K followed by heating in vacuum to 850 K was devised. This method gave
reproducibly clean Pd, which yielded the three overlapping 15N2 desorption curves
shown at the bottom of Fig. 46.2c for three sequential 15NO adsorption experiments
on the carbon-free crystal.
46.2 Chemical Method to Measure and Remove Carbon from Pd 547
Fig. 46.2 Chemical method to measure and remove C from Pd. a Pd auger spectra for increasing
C coverage. b Auger ratio for increasing C coverage. c 2C + 2 15NO(a) → 15N2 + 2CO
46.3 Surface Cleaning Procedures—Silicon
The production of metal-free silicon surfaces is a common starting point in much

surface science work dealing with this element. The metals nickel and iron are
common impurities originating from the vacuum system environment owing to
various types of sources. Their presence causes significant reconstruction effects on
Si(100) [12–14] and on Si(111) [12, 15, 16]. The STM permits the observation of
reconstruction effects at high sensitivity [17], and recently STM has been used to
compare the statistical distribution of defects with the measured Auger intensity
from impurity Ni in the surface region [17]. The transport of Ni through the silicon
lattice is extremely rapid [18], causing Ni contamination effects to spread
throughout the crystal from their origin during ordinary thermal processing.
The fundamental unit of the reconstructed Si(100) surface owing to very low
levels of Ni contamination is the “split-off-dimer” structure, shown in Fig. 46.3a.
Four silicon dimer sites are affected and produce a characteristic STM image of two
neighboring dark dimer sites, a neighbor bright dimer site, and a neighbor dark
dimer site. From detailed Auger spectroscopy studies, about one-half of the
detected Ni atoms are responsible for the split-off dimer sites [17]. As the con-
centration of impurity Ni increases, the split-off dimer structures organize them-
selves to produce vacancy channels having (2 × n) overlayer periodicities.
Ultimately, high corrugation levels of Si(100) with 50-Å roughness are produced.
Figure 46.3b shows the correlation between the statistical counting of
split-off-dimer structures and other methods of surface analysis. A linear relation-
ship is observed between the density of the defects and the surface analysis for Ni.
The origin of the Ni impurities may be from several sources listed below:
1. Ni bulk contamination in Ta used in contact with silicon: The bulk analysis of
high-purity Ta may give 3 ppm of Ni. The typical amount of Ta in the vicinity of
the Si could therefore supply hundreds of monolayers of Ni to the Si.
2. Ni surface contamination of Ta following machining: The transfer of Ni from
alloy tools used to manufacture Ta parts may be a source of significant Ni levels.
3. Ni contamination from thermocouple alloys: Alloys such as chromel-alumel
(type K) are about 90 % Ni. At temperatures of 1450–1500 K, the vapor
pressure of Ni will be in the 10−4–10−5 Torr range, and even protection of the Si
by Ta foil envelopes surrounding the thermocouple junction will not prevent Ni
contamination from this source. It is recommended that Ta-protected
tungsten/rhenium thermocouples be used.
4. Metals deposited by sputtering. Almost all workers using the STM as a detector
of surface cleanliness and structure on Si(100) have discarded the use of ion
bombardment as a way of cleaning the surface. Metal atoms produced from
sputtering effects inside the ion gun, or from sputtering of metal structures in the
neighborhood of the Si sample, are probably transferred to the Si surface.
A simple method that seems to eliminate Ni contamination in Ta components
consists of boiling for 5–10 min in concentrated HC1 (<5 ppb Ni), and then boiling
46.3 Surface Cleaning Procedures—Silicon 549
Fig. 46.3 Surface cleaning procedure—Silicon. a STM image of Si(100) showing

“Split-Off-Dimer” defects. b Correlation of STM defect and AES and LEED measurements
in deionized water. This procedure has been shown to reduce the Ni concentration
in the Si surface to below the Auger sensitivity (signal to noise = 0.0003) level,
giving Si crystals that are long-lived and almost Ni-free as measured by STM [17].
The cleaning procedure for Si(100) wafer strips is then to heat to 900–950 K
overnight to outgas and then to anneal to 1420–1450 K for 30–60 s, cool quickly to
1000–1100 K, hold at temperature to 120 s, and finally cool down to 300 K at a
slow rate of 2 K/s. Alternating current heating is used.
Fig. 46.4 Preparation of clean tungsten single crystals. a Sample holder and flashing geometry.
b Thermal desorption spectra of W(110) crystal behavior at 6 × 10−8 mbar O2 pressure. Both CO
and O2 partial pressures are monitored
46.4 Preparation of Atomically Clean Tungsten Single Crystals 551
46.4 Preparation of Atomically Clean Tungsten

Single Crystals
Tungsten and some of the other refractory transition metals are easily contaminated
in ultrahigh vacuum as a result of hydrocarbon adsorption and decomposition to
make a metal carbide film. This film can be oxidized in vacuum to make a very
stable molecule, CO, which desorbs, leaving pure clean metal behind. The surface
cleaning procedure occurs in two steps: (1). Cycles of low power flashes
(T * 1200 K) in P(O2) = 6 × 10−8 mbar; and (2). a single high desorption pressure
flash (T = 2200 K) in vacuum to remove the oxide layer produced in the first stage
of treatment. The loss of oxide at high pressure leaves a clean surface. In the case of
tungsten wire, the loss of tungsten oxides has been carefully studied [19–21].
Figure 46.4a shows a typical W single crystal held in a molybdenum holder and a
spiral W heater filament. Figure 46.4b shows the typical oxygen, carbon monoxide
and power in the cleaning of the carbon from the crystal surface. It may be seen in
Fig. 46.4b that CO desorption reaches a constant behavior during heating more than
once. The first CO desorption differs from later stages, indicating that most CO is
lost in the first cycle under oxygen [22].
References
1. J.J. Burton, E.S. Machlin, Phys. Rev. Lett. 37, 1433 (1976)
2. R.G. Musket, W. McLean, C.A. Colmenares, D.M. Makowiecki, W.J. Siekhaus, Appl. Surf.
Sci. 10, 143 (1982)
3. H.P. Bonzel, A.M. Franken, G. Pirug, Surf. Sci. 104, 625 (1981)
4. M.J. Cardillo, G.E. Becker, Surf. Sci. 99, 269 (1980)
5. H. Niehus, G. Comsa, Surf. Sci. 102, L14 (1981)
6. S.D. Bader, L. Richter, T.W. Orent, Surf. Sci. 115, 501 (1982)
7. D.G. Castner, G.A. Somorjai, Appl. Surf. Sci. 6, 29 (1980)
8. P.A. Thiel, J.T. Yates Jr, W.H. Weinberg, Surf. Sci. 82, 22 (1979)
9. S. Semancik, G.L. Haller, J.T. Yates Jr, Appl. Surf. Sci. 10, 546 (1982)
10. R.D. Ramsier, K.-W. Lee, J.T. Yates Jr, J. Vac. Sci. Technol. A13, 188 (1995)
11. H. Conrad, G. Ertl, J. Küppers, E.E. Latta, Surf. Sci. 65, 235 (1977); ibid 245 (1977)
12. K. Kato, T. Ide, S. Miura, A. Tamura, T. Ichinokawa, Surf. Sci. 194, L87 (1988)
13. H. Niehus, U.K. Kohler, M. Copel, J.E. Demuth, J. Microscopy 152, 735 (1988)
14. X.F. Lin, K.J. Wan, J. Nogami, Phys. Rev. B 47, 13491 (1993)
15. A.E. Dolbak, B.Z. Olshanetsky, S.I. Stenin, S.A. Teys, T.A. Gavrilova, Surf. Sci. 218, 37
(1989)
16. R.M. Wallace, C.C. Cheng, P.A. Taylor, W.J. Choyke, J.T. Yates Jr, Appl. Surf. Sci. 45, 201
(1990)
17. V. Ukrainsev, J.T. Yates Jr, Surf. Sci. 346, 31 (1996)
18. E.R. Weber, N. Wiehl, Mat. Res. Soc. Symp. Proc. 14, 19 (1983)
19. D.A. King, T.E. Madey, J.T. Yates Jr, J. Chem. Phys. 55, 3236 (1971)
20. D.A. King, T.E. Madey, J.T. Yates Jr, J. Chem. Phys. 55, 3247 (1971)
21. D.A. King, T.E. Madey, J.T. Yates Jr, J. Chem. Soc. Faraday Trans. 1(68), 1347 (1972)
22. Kh Zakeri, T.R.F. Peixoto, Y. Zhang, J. Prokop, J. Kirschner, Surf. Sci. 604, L1 (2010)
Part VIII
High-Area Solids
Chapter 47
Infrared Cells
47.1 Infrared Cell for Adsorption Studies on Supported

Catalysts
The study of supported catalysts using transmission IR spectroscopy is a

well-established technique employed widely throughout the field. Over 50 cell
designs have been devised and reported in the literature [1]. The cell described in
Fig. 47.1 has special features of importance for some types of applications, espe-
cially where a simple design for use at cryogenic temperatures is desired. A unique
feature of this cell is the capability to study the adsorption of gases on the support
and on the support plus the active catalytic material in the same experiments [2].
The cell consists of two 2.75-in. diameter Conflat flanges containing IR trans-
parent (down to 1000 cm−1) windows made of CaF2 single crystals, and joined to
the flange through semiflexible Ag mounting rings, using fused AgCl cement.
These window flanges are sealed with Cu gaskets to a double-sided central flange
that contains the support for the sample. The sample support is a Cu ring (OFHC
Cu) that is brazed with Ag braze (no Cd alloy) to a thin stainless steel tube that exits
through two thin-walled support tubes at the edge of the central flange. A mixture of
liquid nitrogen and nitrogen gas is passed through the supporting tubes to cool the
sample to near 100 K, and the temperature of the support ring is monitored by a
thermocouple entering the central flange at another position. Temperature control
may be achieved in the low-temperature region by control of the flow rate of the
refrigerant, and at high temperatures electrical heating by an external resistive
heater may be controlled. The cell is evacuated through a stainless steel valve that
leads to the gas-handling and vacuum system. Temperature gradients across the
CaF2 sample holding disk (to be described below) that is mounted in the Cu ring are
negligible, as shown by experiments with multiple thermocouples cemented to the
CaF2 plate [3].
The sample of a high area oxide support containing metal particles is prepared
by spraying a slurry of the support material and a salt of the metal onto a CaF2

DOI 10.1007/978-3-319-17668-0_47
556 47 Infrared Cells
Fig. 47.1 IR cell for adsorption studies on supported catalysts. a IR cell. b. Transmission IR
spectra
single crystal disk held on a hot plate at about 70 °C. Here a 9:1 acetone:water
solution is the solvent for the metal salt and a carrier for the oxide support. Flash
evaporation of the solvent occurs during the spraying, and the oxide sticks to the
disk and retains the ions from the salt. In some cases it will be necessary to lightly
abrade the CaF2 disk to permit better adhesion. The disk is masked so that one-half
is sprayed with the catalyst/support mixture while the other half is sprayed with the
47.1 Infrared Cell for Adsorption Studies on Supported Catalysts 557
support alone. This permits adsorption processes on the support to be studied

independently under the identical conditions present at the supported catalyst.
Typical deposited surface densities are at the 0.01 g/cm2 level. The disk containing
the adjacent sprayed sections is transferred to the cell and clipped into the Cu ring,
and the cell is assembled, using annealed Cu gaskets (see p. 99) and gently tight-
ening alternate bolts a little at a time to prevent excessive flexure of the window
flanges and fracture of the CaF2 windows. The catalyst is then prepared for study by
appropriate heating and reduction conditions using hydrogen, and heating the cell in
a small oven that surrounds it. The CaF2 windows should be carefully protected
against thermal shock in use.
After reduction and evacuation to remove hydrogen, the sample is ready for
chemisorption experiments. Figure 47.1b shows an example of the chemisorption of
N2 on Rh metal that is supported on A12O3. At the temperature of the experiment,
N2 is also physically adsorbed on the A12O3. The spectrum on the left was mea-
sured on the right-hand side of the sample and shows only physisorbed N2. The
spectrum on the right, measured on the left-hand side of the cell, shows the presence
of both physisorbed N2 on the A12O3 support and chemisorbed N2 on the metallic
Rh catalyst surface. The cell is translated from side to side inside the spectrometer
using a simple linear bearing track arrangement in order to measure each half of
the sample.
This cell has been used extensively where a simple design was satisfactory.
A more complex design, having additional capabilities, is shown in [1] and dis-
cussed on p. 557. The CaF2 windows mounted in Conflat flanges are commercially
available [4].
47.2 Wide-Temperature-Range IR Cell for High-Area

Solids
Over 50 cell designs for use in the study of high area surfaces by infrared spec-
troscopy have been reported since 1967, as reviewed in [1]. The cell shown in
Fig. 47.2 possesses unique features, which, in combination, are not found in other
designs reported. These features include ultrahigh vacuum operation, wide tem-
perature range extending from about 100 K to above 1500 K without use of an
external furnace, accurate temperature control to 1 K, linear temperature pro-
grammability up to 8 K/s, and very rapid sample cooling rates [1].
The cell design in Fig. 47.2a comprises a six-way stainless cube to which
windows and gas delivery and pumping tubes may be attached by means of 2.75-in.
Conflat flanges. The sample itself is held on a tungsten grid that is clamped between
high thermal conductivity Ni clamps [5], mounted on an electrical feedthrough that
exits from the bottom of a reentrant Dewar made of stainless steel. The feedthrough
also contains matched thermocouple leads that are attached to a thermocouple
welded to the top center of the grid. The entire assembly is oriented using a
Fig. 47.2 Wide-temperature-range IR cell for high area solids. a Cell + sample holder. b Spraying
technique for sample preparation. c Hydraulic pressing technique. d Grid temperature distribution
at steady state. e Temperature programming grid
rotatable flange and is inserted into the six-way cube, along with other desired
attachments.
The grid material is made of thin tungsten foil that has precise square holes
photo-etched in 0.0125–0.05-mm-thick material and exhibiting 58 % optical
transparency in the IR [1]. Each of the square holes is 0.22 mm in width, sur-
rounded by 0.05-mm boundaries. The uniform geometry is advantageous in con-
trolling the uniformity of powder deposition, as will be shown below. Gold grids of
the same geometry have also been employed in this cell.
Two methods may be used for the deposition of high-area solids on the W grid.
As shown in Fig. 47.2b, Step 1, a spraying technique [1] like that described on
p. 123 may be used in which the powdered solid is suspended in a volatile solvent
and sprayed onto the heated grid. Heating to about 70 °C is achieved using tem-
porary power leads to the grid and its Ni clamp support, causing flash evaporation
of the solvent. The grid is rigidly held in a T-shaped Macor jig-mask that is
designed to maintain correct spacing between the Ni clamps so that they may be
slid over the power feedthrough leads on the bottom of the reentrant Dewar without
47.2 Wide-Temperature-Range IR Cell for High-Area Solids 559
flexure of the grid containing the powder deposit as shown in Step 2. The Macor jig
is then removed, as shown in Step 3. The powder first sticks to the boundaries of the
0.22-mm-wide openings and then as the deposition continues, the powder uni-
formly fills the holes. Oxide powder densities up to about 0.01 g/cm2 may be
achieved in the grid openings.
A second method for placing powder into the grid openings is shown in
Fig. 47.2c [6]. Here, the unmounted grid is held between two sheets of glossy
weighing paper and an excess of powder is sprinkled over the grid. Two stainless
steel dies are used to sandwich the grid and its protective paper, and hydraulic
pressure of about 12,000 psi is applied. This causes the powder to enter the grid
openings and to be rigidly held there. Upon removal of the glossy paper, a razor
blade is used to scrape off the excess powder from the top of the grid, and the grid is
then mounted in the Ni clamps for loading into the IR cell. The powder material in
the grid openings will withstand gentle bending of the grid. Powder densities of the
order of 0.004 g/cm2 may be achieved. The spectroscopic, gas transport, and
thermal behavior of the pressed powders is very similar to that of the sprayed
deposits [6].
The thermal properties of the grid during electrical heating to a stable temper-
ature by a feedback controller have been characterized as shown in Fig. 47.2d, e [1].
In Fig. 47.2d, four thermocouples were attached as shown in the diagram. The open
circle and black triangle points were used to measure temperature gradients along a
vertical line in the center of the grid. The three open point measurements were made
to determine the temperature gradient along a horizontal line through the center of
the grid. In the central 4-mm-wide region of the grid that is sampled by the IR
beam, no vertical gradient in temperature is observed. A small dropoff in temper-
ature is seen in the central region as one moves horizontally owing to heat con-
duction away from the grid by the Ni clamps. The ability to temperature-program
the grid and the supported powder to elevated temperatures is illustrated in
Fig. 47.2e. Here, almost linear heating rates between 1.9 and 8.4 K/s are demon-
strated, with the controller being programmed to stop the heating ramp at about
750 K in this case. Rapid cooling in 10–20 s from 900 K to the base temperature of
325 or 150 K maybe achieved [1].
Other cells for heating catalysts to high temperatures and cooling to low tem-
peratures have been described. One of these is used for EXAFS measurements and
works in a range from about 77 to 700 K [7]. Another, involving the use of a
pressed sample in a copper grid, has been described and characterized for thermal
behavior [8].
Cells in which the sample can be heated provide an ideal situation for the study
of infrared emission spectroscopy. While many IR emission studies have been done
on planar surfaces containing coatings, a few studies of high area surfaces as IR
emission sources have been carried out too [9, 10]. One must be careful to rule out
artifacts in this type of measurement [10].
47.3 Measurements of Diffusion Through Powders Using

Transmission Infrared Spectroscopy
Measuring the activated diffusion of adsorbed molecules through powdered

adsorbents may be carried out by using transmission IR spectroscopy to monitor
either the adsorbate average concentration or the surface coverage of markers on the
adsorbent surface which are modified by the arrival of an adsorbate molecule [11–
15]. These measurements are carried out isothermally in the type of IR cell shown
in Sect. 47.2. In this cell, a thin sample of the powder is held inside of the openings
in a tungsten support mesh, and as isothermal diffusion occurs, the absorbance is
used to follow the migration process. Two methods have been used successfully:
(1) An ice layer of the adsorbate is warmed and its inward diffusion is monitored by
its arrival at –OH sites which are modified upon hydrogen bonding to the adsorbate
molecule [11, 12]; (2) A fully covered layer is monitored during isothermal
desorption outward [13–15]. These two methods are schematically shown in
Fig. 47.3a, b, where cross sections through the powder depth are shown before and
after diffusion.
Figure 47.3c shows the schematic potential energy profile for a diffusing mol-
ecule. Initially, when localized by chemisorption on its adsorption site, the potential
energy will be minimized. Lateral motion occurs when the molecule is promoted to
the mobile precursor state [16], where it is held by weaker van der Waals forces and
is able to transport along the surface over low barriers. This process takes place
within the pore structure of the powder. Should the molecule be considered free of
the surface in the pore, the process would be called Knudsen diffusion [17, 18], but
it is believed that the mobile precursor model is more applicable since adsorption
processes on well defined single crystal surfaces often proceed by the mobile
precursor process [16]. At the end of the transport along a pore, desorption will
occur into vacuum from the mobile precursor state. Within the framework of the
mobile precursor state model, the activation energy for surface diffusion is given by
the height of the escape barrier as shown in Fig. 47.3c. As the molecule moves
laterally, it may frequently encounter an open chemisorption (trap) site and return to
a lower potential energy.
Figure 47.3d shows four surface diffusion experiments for an organic molecule
diffusing on an oxide surface from 260 to 290 K [15]. The curves are fits to a two
process model, where slower diffusion amongst strong binding sites is mixed with
faster diffusion on weaker binding sites. Figure 47.3e shows the Arrhenius plots for
the two processes. Molecules moving in the faster process deplete first from the
surface. Often, the IR spectrum of the diffusing molecule changes a little in moving
from the mixed diffusion process to the slower process as surface coverage is
depleted.
47.3 Measurements of Diffusion Through Powders Using … 561
Fig. 47.3 Measurement of diffusion through powders using transmission IR spectroscopy.

a Diffusion from condensed ice film. b Diffusion from adsorbed monolayer. c Potential energy
profile during diffusion. d Normalized integrated absorbance from adsorbed molecules diffusion
into vacuum. e Activation energy for parallel diffusion processes. f Effect of H2O(g) injection on
diffusion organic molecules
In Fig. 47.3e, a diffusion experiment was carried out in which a non-

hydroxylated oxide was employed [14]. As diffusion progressed, water vapor was
injected and then pumped away. This caused the diffusing organic molecule to more
quickly depopulate due to its weaker bonding to surface –OH groups compared to
clean oxide surface sites. The hydroxyl groups promote rapid surface diffusion.
47.4 Using Transmission Infrared Spectroscopy

to Witness Diffusion into Core-Shell Structures
Core-shell structures are sometimes able to transport adsorbed molecules from the
outside to the inside. Since the transporting of adsorbed molecules undergoes dif-
ferent interactions with the shell surfaces and the core surface, its frequency is an
excellent indicator of its progress through the shell to the core. Figure 47.4a shows
a core-shell structure which has been investigated in this manner. A *10 nm
diameter CoO shell has been prepared around a *2 nm diameter Pt particle [19].
Carbon monoxide has been added and the temperature has been raised while
recording transmission IR spectra. CO is postulated not to transport by a gas phase
route through pores in the CoO shell; however CO does transport as an adsorbed
species across CoO surfaces which surround the pores in the shell structure.
Eventually the diffusing CO reaches the Pt core and chemisorbs, giving a charac-
teristic vibrational spectrum for CO/Pt. Figure 47.4b shows the intensity of the
CO/CoO and CO/Pt spectral features as a function of temperature when the sample
is placed under 2 Torr of CO at 110 K and then warmed. Initially, no spectral
features are resolved, as the CO has not entered the shell structure. At about 150 K,
CO/CoO appears and increases in coverage as CO enters into the CoO pores. At
about 200 K, CO transport through the shell structure carries it to Pt sites on the
core, and CO/Pt spectral features begin to develop. Figure 47.4c shows the spectral
measurements as a function of temperature. This is a modification of the method
discussed in Sect. 47.3 where adsorbates are observed to transport through mac-
roscopic distances into the depth of a packed bed of a high surface area adsorbent—
but here the transport distances involved are in the nanometer regime.
47.5 Diffusion in Powders as Measured by Transmission

IR Spectroscopy
The understanding of transport processes by diffusion through powdered adsorbents

is a key to the production of efficient technological sorbents and for efficient
transport of reactants to catalytic sites dispersed on high surface area material.
Transmission IR spectroscopy offers a method to measure the rate of diffusion of
molecules through such materials. One has a choice of monitoring molecular dif-
fusion into a powder from an outer layer of the ice form of the molecule [12, 20] or
monitoring the diffusion outward from a saturated layer preadsorbed on the high
area material [13–15, 21]. A technique involving pressed powder samples, held
under temperature-controlled conditions in a W-grid in an IR cell, is used. The cell
and the methods of using it are described in Sects. 47.2–47.4.
In the case of inward diffusion from the condensed ice on the outer geometrical
surface of the powder, it is important to measure the arrival of molecules at interior
sites in the powder bed. One of the best ways to do this for oxide surfaces is to
47.5 Diffusion in Powders as Measured by Transmission IR Spectroscopy 563
Fig. 47.4 Using transmission

infrared spectroscopy to
witness diffusion into
core-shell structures.
a Schematic of CO diffusion
through a CoO shell in
Pt@CoO. b Integrated
absorbance during CO
diffusion in CoO@Pt.
c Difference spectra for CO
transport through Pt@CoO
monitor the free hydroxyl groups as they slowly become complexed with the
adsorbing molecules. Usually isolated M–OH groups will shift downward in fre-
quency when hydrogen bonded to adsorbed molecules, producing M–OH···X
species. This includes the Xe molecule [20] as well as many types of organic
molecules which sense the strong dipole of the M–OH group. Figure 47.5a shows
the cross section of a pressed powder involved in this inward-diffusion experiment.
It is important in such an experiment to be certain that the source of the diffusing
molecules, the ice, is not depleted as this would change the boundary condition for
the diffusion process. Figure 47.5b shows the drop in absorbance of isolated Al–OH
species and the production of complexed species as diffusion inward occurs. The
drop in normalized integrated Al–OH absorbance is shown versus time as diffusion
occurs and using Fick’s second law of diffusion, the diffusion coefficient can be
extracted from these data [12].
In the case of outward diffusion, shown schematically in Fig. 47.5c, molecules
become mobile inside the pore structure of the powdered adsorbent, which has been
saturated with adsorbate, and eventually desorb, causing the integrated absorbance
to fall. Figure 47.5d shows a fitted kinetic plot for the diffusion where two activated
diffusion processes were found, leading to slower and faster diffusion, as shown by
the Arrhenius plot [13]. Defect sites on the oxide surface are thought to be
responsible for the slow rate of diffusion over these sites. The value of the diffusion
coefficient is derivable from these data where the boundary condition of zero
coverage applies to the outer geometrical surface [13]. The outward diffusion is
sometimes accompanied by small changes in the IR spectrum as the adsorbed
molecules leave, since one preferentially samples the more strongly-bound species
as the diffusion/desorption process proceeds [13–15]. The outward diffusion
method more readily lends itself to studies of temperature dependence since in the
inward process one must remain below the melting temperature of the ice. Recently,
diffusion studies through a metal catalyst supported on an oxide were able to detect
the separate transport of an adsorbed molecule from metal- to- metal site as well as
transport across the oxide sites [21].
47.6 Transmission IR Spectroscopy Through Opaque

Materials
The adsorption of gases on high surface area solid surfaces is of interest techno-
logically. Transmission IR spectroscopy can be applied to such materials (see
Sect. 47.2) if they possess sufficient optical transparency. If the materials are
opaque, they may be studied by DRIFTS (Diffuse Reflection Infrared Fourier
Transform Spectroscopy) [22–25]. However, even opaque materials may be studied
by transmission IR spectroscopy, where transmission occurs by light scattering
through the powdered material and the adsorbed species are located along scattering
47.6 Transmission IR Spectroscopy Through Opaque Materials 565
Fig. 47.5 Diffusion in powders as measured by transmission IR spectroscopy. a Molecular

diffusion from ice overlayer—cross section through compresses sample. b Molecular diffusion
from saturated monolayer—cross section through compresses sample. c Typical spectroscopic data
showing molecular capture by Al–OH groups and kinetics of diffusion from ice. d Typical
spectroscopic data showing diffusion/desorption from MgO powder and arrhenius behavior for
two processes
pathways [26]. Much of the history of the use of transmission IR spectroscopy for
studying adsorption on powdered materials is contained in [27–29].
For the study of opaque materials it is well known that suspending the opaque
particles in an optically transparent matrix (the famous pressed disk method) is
useful. Figure 47.6a shows a hydraulic pressing die which may be used for making
6 different samples on a single tungsten grid support as described in Sect. 47.2.
Figure 47.6b shows a schematic of carbon particles suspended in pressed KBr,
Fig. 47.6 Transmission IR spectroscopy through opaque materials. a Hydraulic pressing die—6
samples. b Cross section through grid + KBr/C sample. c Transmission IR—CO/C/KBr.
d Transmission IR—CO/KBr
where the carbon particles are stored at KBr grain boundaries in the pressed sample.
The individual square openings in the grid are 0.22 mm2 in area, separated by W
grid elements. The pressed samples are 0.051 mm in thickness. Light scattered
through the KBr matrix interacts with the adsorbed species on the suspended carbon
particles.
Figure 47.6c shows the IR spectrum for a high area C/KBr sample when
adsorbed CO is in equilibrium with the gas at the cryogenic temperatures shown in
the figure legend. Two regions of IR absorption are seen. At 2145 and 2131 cm−1,
CO on the carbon surface is detected; at lower CO frequencies, adsorption on the
KBr surfaces is detected. For a KBr matrix alone, Fig. 47.6d shows only the low
frequency CO species as strong bands, accompanied by very low coverage species
with a frequency of 2145 cm−1, close to that of gas phase CO. It is noted that for
many molecules adsorbed on surfaces, there is a matrix shift of a few cm−1 due to
the dual interaction of the molecule with its adsorbent, coupled with interactions
with the surrounding KBr matrix [30].
References 567
References
1. P. Basu, T. Ballinger, J.T. Yates Jr, Rev. Sci. Instrum. 59, 1321 (1988)
2. H.P. Wang, J.T. Yates Jr, J. Phys. Chem. 88, 852 (1984)
3. T.P. Beebe, P. Gelin, J.T. Yates Jr, Surface Sci. 148, 526 (1984)
4. Windows of CaF2 Single Crystal Mounted in Conflat Flanges may be Obtained from Harshaw
Crystal and Electronic Products. Solon Technologies, 6801 Cochran Rd., Solon, OH, 44139
5. Custom Fabricated Grids made of W or Other Metals are Available from Buckbee Mears, 245
E. Sixth Street, St. Paul, MN 55101
6. T.H. Ballinger, J.C.S. Wong, J.T. Yates Jr, Langmuir 8, 1676 (1992)
7. F.W.H. Kampers, T.M.J. Maas, J. van Grondelle, P. Brinkgreve, D.C. Konigsberger, Rev. Sci.
Instrum. 60, 2635 (1989)
8. M. Hughes, L. Bourget, R.G. Greenler, J. Cat. 139, 691 (1933)
9. D. Kember, D.H. Chenery, N. Sheppard, J. Fell, Spectrochim. Acta 35A, 455 (1979)
10. D. Kember, N. Sheppard, J. Chem. Soc., Faraday Trans. 2(77), 1321 (1981)
11. D.B. Mawhinney, J.A. Rossin, K. Gerhart, J.T. Yates Jr, Langmuir 16, 2237 (2000)
12. S. Kim, O. Byl, J.-C. Liu, J.K. Johnson, J.T. Yates Jr, J. Phys. Chem. B 110, 9204 (2006)
13. S. Kim, X. Wang, C. Buda, M. Neurock, O.B. Koper, J.T. Yates Jr, J. Phys. Chem. C 113,
2219 (2009)
14. X. Wang, S. Kim, C. Buda, M. Neurock, O.B. Koper, J.T. Yates Jr, J. Phys. Chem. C 113,
2228 (2009)
15. I.X. Green, C. Buda, Z. Zhang, M. Neurock, J.T. Yates Jr, J. Phys. Chem. C 114, 16649
(2010)
16. W.H. Weinberg, in Kinetics of Interface Reactions, eds. by M. Grunze, H.-J. Kreuzer
(Springer, New York, London, 1986)
17. J.H. de Boer, The Dynamical Character of Adsorption (Oxford University Press, London,
1968)
18. D.D. Do, Adsorption Analysis Equilibria and Kinetics (Imperial College Press, London, 1998)
19. S. Kim, Y. Yin, A.P. Alivisatos, G.A. Somorjai, J.T. Yates Jr, J. Am. Chem. Soc. 129, 9510
(2007)
20. T.H. Ballinger, P. Basu, J.T. Yates Jr, J. Phys. Chem. 93, 6758 (1989)
21. I. Green, W. Tang, M. Neurock, J.T. Yates Jr, Science 333, 736 (2011)
22. J.T. Yates Jr, T.E. Madey (eds.), Vibrational Spectroscopy of Molecules on Surfaces (Plenum,
NY, 1987)
23. P.E. Fanning, M.A. Vannice, Carbon 31, 721 (1993)
24. B.J. Meldrum, C.H. Rochester, J. Chem. Soc. Faraday Trans. 86, 861 (1990)
25. M.B. Mitchell, V.N. Sheinker, E.A. Mintz, J. Phys. Chem. B 101, 11192 (1997)
26. A.T. Bell, M.L. Hair, Vibrational Spectroscopies for Adsorbed Species (American Chemical
Society, Washington, DC, 1980)
27. M.L. Hair, Infrared Spectroscopy in Surface Chemistry (Dekker, New York, 1967)
28. L.H. Little, Infrared Spectra of Adsorbed Species (Academic Press, New York, 1966)
29. A.V. Kiszelev, V.I. Lygin, Infrared Spectra of Surface Compounds (Wiley, New York, 1975)
30. D.B. Mawhinney, J.A. Rossin, K. Gerhart, J.T. Yates Jr, Langmuir 15, 4617 (1999)
Chapter 48
Adsorption/Desorption—Thermal
48.1 Adsorption and Desorption from High-Area Solids
The study of the surface behavior of high area solids can be carried out using
extensions of the fundamental measurement methods developed for studies of low
area single crystal surfaces. Among the most widely used methods in single crystal
studies is temperature-programmed desorption (TPD), often called thermal
desorption spectroscopy (TDS) [1]. The TPD method, with appropriate calibration,
can be used to measure absolute surface coverages as well as the kinetics of
desorption as a function of surface coverage on both single crystals and high-area
solids.
Figure 48.1a shows a thermal desorption spectrometer that was used for the
study of ethylene adsorption on a high-area (423 m2/g) NaX zeolite [2]. The
sample, of mass 10 μg, was deposited on a Ta disk by use of a micropipet that
deposited known volumes of a slurry containing the suspended zeolite in suspen-
sion. Upon evaporation of the solvent, it was found that the zeolite particles were
separated from each other as they adhered to the Ta surface. Thus, interparticle
diffusion effects during adsorption and desorption were eliminated by the high
particle dispersion, although pore diffusion in the zeolite particles was clearly
evident in the measurements [2].
The high-area sample is first cleaned by heating in vacuum using a temperature
programmer that operates in a feedback circuit involving the thermocouple. Gases
are adsorbed in the pumped system, or for high exposures, with the gate valve
closed.
The sample is held in front of a 2-μm sampling aperture that may be opened and
closed using an O-ring seal on the end of a linear motion device. With the gate
valve closed, this permits the sampling of gases near the sample surface at high
pressures as would be necessary for the study of catalytic reactions. However, with
the gate valve open after adsorption, sampling of desorbing gas from the powdered
sample is possible in the pumped system through direct transmission from the

DOI 10.1007/978-3-319-17668-0_48
570 48 Adsorption/Desorption—Thermal
Fig. 48.1 Adsorption/

Desorption on powdered
solids. a Thermal desorption
mass spectrometer. b Typical
Adsorptionuptake data
48.1 Adsorption and Desorption from High-Area Solids 571
sample to the mass spectrometer. The mass spectrometer and pumping speed of the
system may be separately calibrated using an absolute flow of gas into the pumped
system through a calibrated leak (see p. 425). Knowing the mass of the sample and
its surface area per unit mass, the absolute coverage of adsorbate may be deter-
mined by thermal desorption measurements, integrating the pressure signal over the
time duration of the desorption peak.
Figure 48.1b shows an example of the uptake of ethylene versus exposure on
NaX zeolite at 114 K. At low exposures, adsorption is postulated to occur on the
outer surfaces of the isolated zeolite particles [2, 3]; at exposures above about 200 L
(1 L = 1 × 10−6 Torr s), adsorption occurs into the pores of the zeolite, and the
efficiency of adsorption increases significantly as indicated by the lower curve of
Fig. 48.1b. Over the entire upper range of exposures, a typical sigmoid uptake curve
is found, as shown in the upper curve of Fig. 48.1b. The inset shows an analysis that
may be used to predict the saturation coverage of the adsorbate, which occurs in the
limit of 1/ε = 0, where ε is the gas exposure.
Extensive work on the analysis of TPD spectra from high-area solids has been
reported and appropriate references are listed in [2]. Noteworthy is a study by
Demmin and Gorte [4]. A major consideration is the effect of interparticle diffusion
on the observed desorption kinetics, and the method shown here, involving very
high particle dispersion, avoids this effect, while also offering sensitivity at the
10−4 ML level for powders with very high specific surface areas. Small corrections
for the adsorption on the Ta support may be easily applied.
Similar methods have been applied to larger samples (0.1 g) of powdered
material in the form of compressed pellets, but the influence of interparticle
diffusion is not eliminated in these studies [5].
A recent review of temperature-programmed desorption from high area solids
maybe found in [6].
48.2 Thermal Desorption from High Area Materials
The extension of temperature programmed desorption (TPD) to high area materials

has occurred. A common problem is that the surfaces studied have such a wide
range of adsorption sites that clear kinetic resolution of desorption from these sites
is not possible. Also, because high area powdered materials which have been
packed into a powder bed will mix gas desorption processes with surface diffusion
processes, the resolution of desorption states into separate processes is virtually
impossible. Early studies on zeolites were able to separate the external surface sites
from interior sites using temperature programmed desorption [7]. Transmission IR
studies (see Sects. 47.3–47.5) are also able to observe surface diffusion processes
[8–10] as well as the adsorption sites in core shell structures [11].
One class of high area material which displays well defined adsorption sites is
carbon single walled nanotubes (C-SWNTs) and the successful discrimination of
these sites has been achieved by TPD methods [12–16]. In Fig. 48.2a, the UHV
Fig. 48.2 Thermal desorption from high area materials. a Thermal desorption apparatus.
b. n-Heptane thermal desorption from SWNT. c Experimental detection of different SWNT
binding sites for several n-Alkanes. d Graphite versus single walled nanotubes—n-Heptane
desorption
48.2 Thermal Desorption from High Area Materials 573
apparatus, involving an apertured mass spectrometer is shown [16]. The SWNTs

are delivered in solvent to a Au platelet which can be programmed upwards in
temperature. Because of their high specific surface area a few µg are all that is
needed to obtain strong desorption signals. In the figure a molecular beam doser is
shown which is used for quantitative delivery of adsorbate to the sample. We find in
control experiments that desorption from the support is of order of 1 % of that from
the SWNT sample. Figure 48.2b shows the desorption spectra of n-heptane from an
opened SWNT sample [15]. The sequential depopulation of various desorption
states, assigned as shown in the schematic to various classes of sites, is easily
observed. The desorption begins, at high enough coverage, by depopulating mul-

tilayer n-heptane, followed by desorption from the external first layer sites, the
groove sites and finally the internal sites [14]. The desorption kinetics from these
sites have been modeled. Figure 48.2c shows the behavior for a series of linear
alkanes and it may be seen that the temperatures of desorption from the various
states monotonically increase with increasing chain length and adsorption energy
[17]. Finally, in Fig. 48.2d, we see desorption of n-heptane from graphite compared
to nanotubes. While graphite exhibits an adsorption state which evolves at higher
temperature than the multilayer n-heptane, the various sites on SWNTs are not
observed [18]. The method has also been applied for the study of confined chemical
reactions in the nanotubes [19], where surface defect sites have been found to be
important.
48.3 Microcalorimetric Studies of Adsorption Heat

on Powders
The direct measurement of the energy released during chemisorption provides a

measurement of the binding energy of adsorbed molecules that depends only of the
first law of thermodynamics. A method for making these measurements on thin
single crystal surfaces has been given on p. 333.
In Fig. 48.3a, the cell used for calorimetric measurement on high-area metal
catalysts is shown [20]. A metallic Ni catalyst, prepared from Ni(II)(NO3)2 · 6H2O
powder by reduction in H2 at 623 K, is stored under He in a sealed-off Pyrex NMR
tube, and this tube is located in one arm of the calorimeter cell. The metal weighs
between 0.5 and 1.0 g and has a surface area of 1.7 m2/g and a CO capacity of
22 μmol/g. The leak rate of the closed cell was measured to be about 10−6 μmol/min
in a 70-cm3 volume.
Calorimetric measurements are made on the catalyst by first breaking the glass
NMR tube using the linear motion feedthrough that presses down on the inner tube.
Then small amounts of CO (or other gases) are admitted and the thermal response
of the calorimeter is measured, comparing the sample cell to the reference cell in
two identical calorimeters. The uptake of gas is measured using standard volumetric
techniques and a capacitance manometer with a sensitivity of 0.5 × 10−4 Torr.
The calorimeter is a commercial device, a Setaram C-80 heat-flux calorimeter.
Typical results are shown in Fig. 48.3b, and are compared with measurements
made on a Ni(100) single crystal [21]. The coverage scales are normalized to the
saturation coverage in both cases. It may be seen that the agreement is excellent,
especially for the heat of adsorption at zero coverage. These results suggest that the
chemical reduction technique used for preparation of the powdered Ni results is a
clean surface that behaves in a fashion similar to atomically clean Ni surfaces.
In [21], the range of the initial adsorption heat was 8 %, comparing three low index
planes of Ni.
48.3 Microcalorimetric Studies of Adsorption Heat on Powders 575
Fig. 48.3 Microcalorimetric

studies microcalorimeter—
adsorption heats on powders.
a Calorimeter cells. b CO
chemisorption on Ni
A second example of the use of microcalorimetry for powdered samples may be

found in [22–24]. Here, the sample, weighing about 0.5 g, is spread over the flat
bottom of the calorimeter (6.3-cm2 geometrical area). This permits rapid equili-
bration between the gas and the adsorbed layer, and also rapid thermal equilibrium.
48.4 Chemical Vapor Synthesis of Oxide Nanoparticles
Chemical vapor synthesis (CVS) is well suited for the production of alkaline earth oxide
nanoparticles (MgO, CaO, SrO, and BaO) and mixed oxides such as CaxMg1−xO.
Due to the absence of solvents, the vapor synthesis can yield clean material and controlled
particle size. The reactor is shown in Fig. 48.4a, allowing controlled oxidation of metal
vapors under reduced pressure [25]. Concentric quartz tubes are placed in a furnace with
good temperature control. The inner tube contains two ceramic boats containing several
Fig. 48.4 Chemical vapor synthesis of oxide nanoparticles. a CVS reactor. b TEM image of MgO
nanocubes
48.4 Chemical Vapor Synthesis of Oxide Nanoparticles 577
grams of metal and the temperature is adjusted under Ar flow to produce sufficient metal
vapor pressure. The inert Ar carries the metal vapor to the end of the inner tube where it
meets flowing oxygen. The exothermic oxidation reaction produces a bright stable flame
at the end of the inner tube and alkaline earth oxide nanoparticles are produced as a result
of homogeneous nucleation in the gas phase. Due to continuous pumping, the residence
time of the oxide nuclei within the flame remains short enough to prevent substantial
coarsening and coalescence. A by-pass system (Valve V1) allows one to avoid particle
collection during uncontrolled process conditions during initial heating. Following par-
ticle synthesis and collection in the particle collection net, the furnace is rapidly cooled by
flowing air through the region surrounding the outer quartz tube. The oxide nanoparticles
may be transferred in air to other instruments for study, and cleaning procedures in
vacuum should be used to remove contaminants adsorbed on the nanoparticles during
transfer. An electron micrograph of nanometer sized MgO nanocubes, exposing MgO
(100) faces is shown in Fig. 48.4b.
References
1. J.T. Yates Jr, Meth. Exp. Phys. 22, 425 (1985)

2. P. Basu, J.T. Yates Jr., Surf. Sci. 177, 291 (1986). U.S. Patents 4,877,584 and 4,663,297
3. M. Kiskinova, G.L. Griffin, J.T. Yates Jr, J. Catal. 71, 278 (1981)
4. R.A. Demmin, R. Gorte, J. Catal. 90, 32 (1984)
5. G.D. Moggridge, J.P.-S. Badyal, R.M. Lambert, J. Vac. Sci. Technol. A8, 3874 (1990)
6. R.J. Gorte, Catal. Today 28, 405 (1996)
7. P. Basu, J.T. Yates Jr, Surf. Sci. 177, 291 (1986)
8. S. Kim, O. Byl, J.-C. Liu, J.K. Johnson, J.T. Yates Jr, J. Phys. Chem. B 110, 9204 (2006)
9. S. Kim, X. Wang, C. Buda, M. Neurock, O. Koper, J.T. Yates, Jr. J. Phys. Chem. C 113, 2219
(2009)
10. X. Wang, S. Kim, C. Buda, M. Neurock, O. Koper, J.T. Yates Jr, J. Phys. Chem. C 113, 2228
(2009)
11. S. Kim, Y. Yang, A.P. Alivisatos, G.A. Somorjai, J.T. Yates Jr, J. Am. Chem. Soc. 129, 9510
(2007)
12. P. Kondratyuk, J.T. Yates Jr, Chem. Phys. Lett. 383, 314 (2004)
13. P. Kondratyuk, J.T. Yates Jr, Chem. Phys. Lett. 410, 324 (2005)
14. P. Kondratyuk, Y. Wang, J.K. Johnson, J.T. Yates Jr, J. Phys. Chem. B 109, 20999 (2005)
15. P. Kondratyuk, Y. Wang, J. Liu, J.K. Johnson, J.T. Yates Jr, J. Phys. Chem. C 111, 4578
(2007)
16. P. Kondratyuk, J.T. Yates Jr, J. Am. Chem. Soc. 129, 8736 (2007)
17. P. Kondratyuk, J.T. Yates Jr, Acc. Chem. Res. 40, 995 (2007)
18. L. Mandeltort, J.T. Yates Jr (private communication)
19. L. Mandeltort, M. Buettner, J.T. Yates Jr, P. Choudhury, L. Xiao, J.K. Johnson, J. Phys.
Chem. C 114, 17148 (2010)
20. B.E. Spiewak, P. Livin, R.D. Cartright, J.A. Dumesic, J. Phys. Chem. 100, 17260 (1996)
21. J.T. Stuckless, N. Al-Sarraf, C. Wartnaby, D.A. King, J. Chem. Phys. 99, 2202 (1993)
22. D.J. Parrillo, C. Lee, R.J. Gorte, App. Catal. A: General 110, 67 (1994)
23. D.J. Parrillo, R.J. Gorte, Catal. Lett. 17, 16 (1992)
24. D.J. Parrillo, R.J. Gorte, J. Phys. Chem. 97, 8786 (1993)
25. O. Diwald, Institute of Particle Technology, University Erlangen-Nuremberg (private
communication)
Part IX
Safety
Chapter 49
Protecting the Vacuum System
49.1 Delay Circuit for Turbopumping Protection Against

Vacuum Loss from Power Interruptions
Power interruptions can play havoc with UHV systems operating with a turbopump,
and particularly when these systems are being baked out. A power failure has the
effect of venting a turbopump, using the commercial vent valve designed to prevent
oil back-streaming from the forepump into the turbopump and the vacuum chamber
during a power interruption. A partial solution to avoid massive air entry into the
UHV system is to employ a pneumatic gate valve between the turbopump and the
vacuum system. However, there are two defects in this protection system: (1) the
time of closure of the gate valve is too long to prevent at least partial loss of vacuum
in the chamber; and, (2) when the power returns, the gate valve will quickly open,
exposing the chamber to a turbopump which is at or near atmospheric pressure.
A circuit to prevent these effects is described below [1]. There is a commercial
device that will do almost the same thing electrically [2], and a pneumatic device
also for the same purpose [3].
The first problem mentioned above is solved by venting the turbopump only
after the gate valve is completely closed. This is done by supplying short-term
emergency power to the turbopump vent valve for the necessary period of time
(approximately 8 s is used). The second problem is solved by controlling the
pneumatic gate valve with a relay [4] that latches in the off position upon power
failure. This prevents the valve from reopening when the power returns until the
relay is overridden.
Figure 49.1 shows the circuit, costing less than $100, which accomplishes both
objectives. The nondelayed socket supplies power to the gate valve relay. A line
switch between the nondelayed socket and the gate valve supplies manual control
of the gate valve. The nondelayed socket power is designed to latch in the off
position in the case of a power interruption, and pushing the reset button reactivates
power at this socket when the main power is restored.

DOI 10.1007/978-3-319-17668-0_49
582 49 Protecting the Vacuum System
Fig. 49.1 Delay circuit for turbopumping protection against vacuum loss from power interruption
The delayed socket supplies emergency power to the turbopump vent valve. It is
powered by a battery-driven backup AC power supply [5]. When main power is
interrupted, emergency power is supplied and then shuts off, venting the turbo-
pump. For momentary main power interruptions lasting less than 8 s, the turbo-
pump vent valve will not open because power is supplied from the backup supply.
The control system does not assume that when power returns the turbopump
restarts and can be automatically opened to the vacuum system, unlike other sys-
tems [2, 3]. It is deemed best to leave the system shut down after power inter-
ruptions, awaiting manual restarting.
49.2 Fast-Closing Beam Valve for UHV Chamber Protection 583
49.2 Fast-Closing Beam Valve for UHV Chamber

Protection
It is often necessary to protect an ultrahigh vacuum system from vacuum failure in

an auxiliary system, such as an attached beam source system. In this particular case,
since the beam is of small cross section, the valve can be very small. The valve
must be electrically actuated by relays connected to pressure gauges in both
systems.
Figure 49.2 shows a modified solenoid actuated valve [6] that is inexpensive,
fast, and ultrahigh-vacuum-compatible. The solenoid driving the valve is encased in
a housing that is sealed by O-rings so that gas evolution from the heated coil will
Fig. 49.2 Fast closing beam valve—UHV chamber

not degrade the vacuum. The housing is tightened into place by using a nut that
engages a threaded rod at the top of the assembly.
When the current is on, the wedged-shaped Cu seal is drawn up into the sole-
noid; interruption of vacuum results in turning the solenoid off, and the Cu seal is
dropped into the sealing position, where it is held in place by a spring and by
gravity. The seal is made against a Viton O-ring in the lower portion of the valve.
Copper cooling coils are used to dissipate heat from the solenoid. Electrical con-
nections are brought out from the solenoid housing through a Conflat flange, not
shown.
The leak rate through the closed valve is 4 × 10−5 Torr L/s, and is too large to be
used to protect an ultrahigh vacuum system without a pumping section between the
valve and the system. In the case of the use of this valve, a differentially pumped
chopper region is present and serves this purpose [7].
A similar principle has been used to insert a large gate valve into an ultrahigh
vacuum system [8].
49.3 Safety System for Oil Diffusion-Pumped UHV

Systems
Modern oil diffusion pumps are now used to achieve ultrahigh vacuum, and these
pumps exhibit some advantages over more expensive ion pumps or turbomolecular
pumps. Among the advantages are lower cost, the ability to pump all gases, the
absence of magnetic fields, and the absence of vibrations. Among the disadvantages
are the possibilities of system contamination if failure of various components occurs
[9]. The failure items are loss of vacuum, loss of power, loss of cooling water, and
loss of liquid nitrogen cooling in the ultrahigh vacuum trap. A pneumatically
actuated valve between the pumps and the ultrahigh vacuum system is the key
element that closes to protect the system if some failure should occur.
A schematic drawing of an oil diffusion pumped ultrahigh vacuum system is
shown in Fig. 49.3a. It consists of a standard bakeable chamber with an ionization
gauge and a titanium sublimation pump (TSP). The pneumatically actuated gate
valve separates the system from the external diffusion pump and liquid nitrogen
trap. The oil diffusion pump contains special perfluoro polyether pump oil (Fomblin
type) or polyphenyl ether (Santovac-5) pump oil used for achieving the best
pressure. Silicone diffusion pump oils should be avoided in all UHV apparatus. The
oil diffusion pump is protected from the oils of the mechanical pump by a bakeable
molecular sieve trap, and the foreline pressure is monitored by a thermocouple
gauge or other pressure gauge.
The simple relay circuit shown in Fig. 49.3b is designed to protect the system
from the various failure possibilities [9]. Digital electronic components were
avoided to reduce the circuit sensitivity to electrical noise. The control items in the
vacuum system are: (1) foreline thermocouple gauge maximum pressure set point—
49.3 Safety System for Oil Diffusion-Pumped UHV Systems 585
Fig. 49.3 Safety system for oil-diffusion-pumped UHV systems. a Schematic oil diffusion
pumped UHV system. b Safety system schematic
to J1; (2) water flow sensor—to J2; (3) ionization gauge maximum pressure setpoint
—to J3; (4) liquid nitrogen filler with low level sensor—to J4; (5) pneumatically
actuated system valve—to J5. The relay, Rl, is a double-pole, single-throw relay
and is turned on by closing the power switch, SW1, and then depressing the
momentary contact switch, SW2. This sets R1 if the fore-line pressure is below the
Table 49.1 Interlock response to failure conditions

Failure Diffusion pump power System valve
Chamber pressure too high On Closed
Liquid N2 level too low On Closed
Foreline pressure too high Off Closed
Water failurea Off Closed
Electrical failure Off Closed
a
An additional bimetallic thermal switch control is available on most diffusion pumps that
interrupts the heating power if the water goes off
thermocouple gauge set point and if there is sufficient water flowing through the
diffusion pump, closing J2. The switch SW3, which controls the relay, R2, may
then be used to switch on the diffusion pump. The system valve, SW4, can then be
actuated if continuity exists through the right-hand side of the circuit, meaning that
the liquid nitrogen trap and the ion gauge are operating in the proper range. If one is
in the initial pumpdown period, the pressure will be too high at the ionization
gauge, and this gauge and the thermocouple gauge can be bypassed temporarily by
closing switch SW5.
Table 49.1 shows the various failure conditions and the components’ condition
when these failures occur.
49.4 Folding Linear Magnetic Translator
Rather long linear and rotary magnetic translators are often used in ultrahigh
vacuum systems, and these are always in danger of being hit by laboratory per-
sonnel as they pass by. The design shown in Fig. 49.4 prevents this by folding the
device housing containing the retracted magnetic translator/rotor.
Figure 49.4a shows a side view of the device. The welded bellows is folded
upward through a 90° angle by a hinge and locked into place. In the horizontal
position the bellows is held in a slightly compressed configuration by the gimbal
subassembly. The outer end of the outer cylinder of the translator has a threaded rod
attached on axis and this is retained in a “scissors” subassembly that permits
accurate vertical positioning of the end of the translator. The bench holding the
device is attached to the system lower flange by a rigid yoke assembly.
Figure 49.4b shows the gimbal subassembly and the two rotational adjustments
that are possible as well as a lateral translation for alignment. A double-sided flange
is used in order to be able to disassemble either the bellows or the translator section
independently. Washers, used as spacers under the gimbal, permit vertical
adjustment.
A critical design consideration is the appropriate size of the welded bellows.
Figure 49.4c shows the dimensional parameters associated with a welded
bellows-namely, the relaxed or free length (FL) at one atmosphere, the extended
49.4 Folding Linear Magnetic Translator 587
Fig. 49.4 Folding linear magnetic translator. a Side view. b Gimball subassembly. c Welded
bellows—dimensional definitions. d Bellows in bent configuration
length (EL), and the compressed length (CL). These parameters are specified by the
bellows manufacturer. In the bent condition, the inner radius of the bellows is
defined as the working length (WL) in Fig. 49.4d. The outer length of the bent
bellows is therefore WL + θr, and the condition in (49.1) must be met:
WL þ hr\EL ð49:1Þ
In order to achieve the condition in (49.1), the bellows must be compressed

during installation by using the gimbal’s subassembly, such that WL < FL. Note
that welded bellows must be used for this device, since preformed bellows will not
have the range of extension or the flexibility needed.
References
1. J.A. Polta, P.A. Thiel, J. Vac. Sci. Technol. A5, 386 (1987)
2. Balzers Part No. TCV 102. Obtained from Balzers, 8 Sagamore Park, Hudson, NH 03051
3. J.P. Saint-Germain, G. Abel, B.L. Stansfield, J. Vac. Sci. Technol. A4, 2391 (1986)
4. Time delay CNS-35-76 from Newark Electronics, 4801 N. Ravenswood Avenue, Chicago, IL
60640-4496
5. Available from Tripp Lite, 500 N. Orleans Street, Chicago, IL 60610
6. Vacoa model SOV-050-00-SS, Vacuum Accessories Corporation of America, 390 Central

Avenue, Bohemia, NY 11716
7. T. Engel, D.N. Braid, Rev. Sci. Instrum. 57, 301 (1986)
8. J.J. Zinck, W.H. Weinberg, Rev. Sci. Instrum. 55, 810 (1984)
9. L.H. Dubois, J. Vac. Sci. Technol. A6, 162 (1988)
10. J.T. Fitch, J.J. Sumakeris, G. Lucovsky, J. Vac. Sci. Technol. A11, 2871 (1993)
Chapter 50
Protecting Personnel
50.1 Electrical Shocks in the Laboratory
The quantitative effects of electrical current flowing through the human body are
listed in the Table 50.1 [1]. It is the current flow through the body that is the
fundamental parameter involved in electrical shock, and the level of danger with a
high-voltage power supply is related to the current rating of the supply. Safety with
electrical devices involves minimizing accidental current flow through the body,
especially through the heart. One may have to work with one hand in a pocket when
there is danger of shock from high-voltage wires, but of course this is not an
adequate protection against shock. Electrical shielding of leads and terminals is an
obvious precaution that must be taken.
The avoidance of electrical shocks in the laboratory is carried out with an
understanding of the meaning of the electrical ground and how to connect electrical
equipment to ground. Figure 50.1a shows a three-wire wall outlet and a polarized
male plug. The green ground wire is a non-current-carrying voltage reference line
that is connected to the earth somewhere near the laboratory building. The white
wire, which is neutral, is connected to the ground wire in the vicinity of the wall
outlet. At the point of connection, the electrical potential of the white wire will also
be at ground potential. However, as one moves from this connection point, there
will be a change in the potential of the white wire owing to the IR drop caused by
the current flow through the white wire. One should avoid connecting the white
wire to the ground wire in the laboratory. The black wire is the hot wire, and it
remains isolated from both the ground and the neutral wire. The wall outlet contains
three contacts—the large slot for the neutral wire, the small slot for the hot wire,
and the larger hole for the ground connection.

DOI 10.1007/978-3-319-17668-0_50
590 50 Protecting Personnel
Table 50.1 Quantitative effects of electric current on man

Effect DC AC—60 Hz AC—10,000 Hz
Men Women Men Women Men Women
(mA) (mA) (mA) (mA) (mA) (mA)
Slight sensation on hand 1 0.6 0.4 0.3 7 5
Perception threshold 5.2 3.5 1.1 0.7 12 8
Nonpainful shock— 9 6 1.8 1.2 17 11
muscular control not lost
Painful shock—muscular 62 41 9 6 55 37
control not lost
Painful shock—let-go 76 51 16 10.5 75 50
control threshold
Painful and severe shock 90 60 23 15 94 63
—muscular contractions;
breathing difficult
Possible ventricular fibrillation
3-s shocks 500 500 100 100
−1/2
T-s shocks 165T 165T−1/2
Figure 50.1b shows two electronic units connected to AC power through

three-wire power leads. Usually the chassis of each of the units will be grounded
internally to the green grounding wire. However, to make doubly sure that the
chassis of each is grounded, it is good practice to connect each chassis to a separate
external ground buss by means of the chassis ground connection on each unit. The
reason for this is that if the hot wire or any internal high-voltage source should
accidentally touch the chassis from the inside, the grounding of the chassis will
protect the laboratory worker from accidentally contacting a high electrical
potential.
50.2 Accidental Electrical Charging from Ionization

Gauge
Accidental activation of the electron bombardment degassing capability of an

ionization gauge in the presence of gas pressures that will support an electrical
discharge can lead to potentially lethal electrostatic potentials on isolated compo-
nents of an ultrahigh vacuum system. Figure 50.2 shows a schematic diagram of an
ionization gauge operating at pressures in the 10−3 Torr range. A plasma is pro-
duced when the gauge is accidentally switched to the degas mode as a positive
50.2 Accidental Electrical Charging from Ionization Gauge 591
Fig. 50.1 Electrical shocks in the laboratory. a 110 V AC wall outlet + polarized plug. b Coupled
chassis ground
potential in the kilovolt range is applied to the grid for normal electron bombard-
ment cleaning. An electrically isolated feedthrough inside the vacuum system is
immersed in the plasma and comes to the plasma potential which is near the
potential of the grid. Contact with this feedthrough electrode external to the vacuum
system can result in exposure to high electrical potentials, and one example of this
592 50 Protecting Personnel
Fig. 50.2 Accidental electrical charging from ionization gauge
effect leading to death has been reported. A safe operating procedure to avoid this
type of effect is never to attempt to outgas an ionization gauge when the pressure is
high, which would be a pointless thing to do anyway.
Reference
1. W.T. Peria, Department of Electrical Engineering, Institute of Technology, 4-174 EE/CSci

Building, 200 Union Street S.E., Minneapolis, MN 55455 (private communication). The table is
taken from the Physical Electronics Laboratory Manual, ca. 1965
Appendix A
Useful Books—Experimental Surface Science Methods

Anderson, R.L., Revealing Microstructures in Metals, Scientific Paper
425-C000-P2 (West-inghouse Research Laboratories, Pittsburgh, PA, 1961).
Baria, D.N. and Bautista, R.G., Compilation of Experimental Work on Thermal
Emittances of Materials, Part I—Metals; Part II—Non-Metals; Part III—Alloys
(Ames Laboratory, USAEC, Iowa State University, Ames, IA 50010, 1974).
Available from National Technical Information Service, Dept. A, Springfield, VA
22151.
Barr, W.E. and Anhorn, V.J., Scientific and Industrial Glassblowing and
Laboratory Techniques (Instrument Publishing Co., Pittsburgh, PA, 1947).
Beavis, L.C., Harwood, V.J. and Thomas, M.T., Vacuum Hazards Manual, AVS
Monograph Series M-l, 2nd Ed. (American Vacuum Society, New York, NY
10017, 1979).
Berman, A., Total Pressure Measurements in Vacuum Technology (Academic,
New York, 1985).
Braker, W. and Mossman, A.L., Matheson Gas Data Book (Matheson Gas
Products, Secau-cas, NJ 07094, 1980).
Breck, D.W., Zeolite Molecular Sieves (Wiley, New York, NY, 1977).
Bunshah, R.F., Ed., Handbook of Deposition Technologies for Films and Coatings,
2nd Ed. (Noyes, Park Ridge, NJ, 1994).
Dawson, P.H., Quadrupole Mass Spectrometry and Its Applications (American
Institute of Physics, New York, NY 10017-3483, 1995).
Drinkwine, M.S. and Lichtman, D., Partial Pressure Analyzers and Analysis, AVS
Monograph Series, M-3 (American Vacuum Society, New York, NY 10017,1979).
Dushman, S., Scientific Foundation of Vacuum Technique, 2nd Ed. (Wiley, New
York, NY, 1962).
Ertl, G. and Küppers, J., Low Energy Electrons and Surface Chemistry, 1st Ed.,
(Verlag Chemie, Weinheim, 1974); 2nd Ed., completely revised (VCH
Verlagsgesellschaft, mbH, Weinheim, 1985).
Fomenko, V.S., Handbook of Thermionic Properties (Translated), Edited by
Samsonov, G.V. (Plenum, New York, NY, 1966).

DOI 10.1007/978-3-319-17668-0
594 Appendix A
French, T.E. (Revised by Vierck, C.J.), A Manual of Engineering Drawing, 7th Ed.
(McGraw Hill, New York, NY, 1947).
Gomer, R., Field Emission and Field Ionization (American Institute of Physics,
New York, NY 10017-3483, 1993).
Gordon, A. J. and Ford, R.A., The Chemist’s Companion—A Handbook of Data,
Techniques, and References (Wiley, New York, NY, 1972).
Hablanian, M.H., Diffusion Pumps, AVS Monograph Series, M-5 (American
Vacuum Society, New York, NY 10017, 1983).
Kennard, E.H., Kinetic Theory of Gases (McGraw-Hill, New York, NY, 1938).
Kohl, W.H., Handbook of Materials and Techniques for Vacuum Devices
(American Institute of Physics, New York, NY 10017-3483, 1995).
Leco Corporation, Metal Etching and Polishing (St. Joseph, MI 49085, 1977).
Leybold-Heraeus, Vacuum Technology Its Foundations, Formulae, and Tables
(Leybold-Heraeus, Pittsburgh, PA 15632).
Loeb, L.B., The Kinetic Theory of Gases, 3rd Ed., (Dover, NY, 1961).
Lunn, G. and Sansone, E.B., Destruction of Hazardous Materials in the
Laboratory (Wiley, New York, NY, 1990).
Madey, T.E., Ed., Vacuum Science and Technology, History of Vacuum Science
and Technology, Vol. 1 (American Vacuum Society, AIP Press, New York, NY,
1984).
Maissei, L.I. and Glong, R., Handbook of Thin Film Technology (McGraw-Hill,
New York, NY, 1970).
Mark, H.F., Ed., et al., Encyclopedia of Chemical Technology (Wiley, New York,
NY, 1979).
Moore, J.H., Davis, C.C. and Coplan, M.A., Building Scientific Apparatus
(Addison-Wesley Reading, MA, 1983).
Müller, E.W. and Tsong, T.T., Field Ion Microscopy (Elsevier, New York, NY,
1969).
O'Hanlon, J.F., A User’s Guide to Vacuum Technology (Wiley, New York, NY,
1980).
The Omega Complete Temperature Measurement Handbook and Encyclopedia,
Omega Engineering, Inc., P.O. Box 4047, One Omega Drive, Stamford, CT 06907.
Pierce, J.R., Theory and Design of Electron Beams (D. van Nostrand Co., Inc.,
New York, NY, 1949); 2nd Ed. (1954).
Redhead, P.A., Hobson, J.P. and Kornelsen, E.V., The Physical Basis of Ultrahigh
Vacuum (American Institute of Physics, New York, NY 10017-3483, 1993).
Redhead, P.A., Ed., Vacuum Science and Technology, History of Vacuum Science
and Technology, Vol. 2 (American Vacuum Society, AIP Press, New York, NY,
1994).
Roberts, R.W. and Vanderslice, T.A., Ultrahigh Vacuum and Its Applications
(Prentice-Hall, Engelwood Cliffs, NJ, 1963).
Robinson, N.W., The Physical Principles of Ultrahigh Vacuum Systems and
Equipment (Chapman and Hall, London, 1968).
Rosebury, F., Handbook of Electron Tube and Vacuum Techniques (AIP,
American Institute of Physics, New York, NY, 1993).
Appendix A 595
Roth, A., Vacuum Sealing Techniques (Pergamon, Oxford, UK, 1966).

Roth, A., Vacuum Technology, 2nd Ed. (North Holland, Amsterdam, 1982).
Sala, A., Radiant Properties of Materials (Elsevier, Amsterdam, 1986).
Smithells, C.J. and Brandes, E.A., Eds., Metals Reference Book, 7th Ed.
(Butterworths, Lon-don and Boston, 1991).
Spangenberg, K.R., Vacuum Tubes (McGraw-Hill, New York, NY, 1948).
Strong, J., Neher, H.V., Whitford, A.E., Cartwright, CH. and Hayward, R.,
Procedures in Experimental Physics (Prentice-Hall, New York, NY, 1946).
Thompkins, H.G., The Fundamentals of Vacuum Technology, AVS Monograph
Series, M-6 (American Vacuum Society, New York, NY 10017,1991).
Thompkins, H.G., Pumps Used in Vacuum Technology, AVS Monograph Series,
M-9 (American Vacuum Society, New York, NY 10017,1991).
Thompkins, H.G., Vacuum Gauging and Control, AVS Monograph Series, M-12
(American Vacuum Society, New York, NY 10017,1994).
Wheeler, E.L., Scientific Glassblowing (Wiley, New York, NY, 1958).
Wilson, E.B., An Introduction to Scientific Research, 1st Ed. (McGraw-Hill, New
York, NY, 1952).
Wilson, N.G. and Beavis, L.C., Handbook of Vacuum Leak Detection, AVS
Monograph Series, M-2 (American Vacuum Society, New York, NY 10017, 1979).
Wood, E.A., Crystal Orientation Manual (Columbia University Press, New York,
NY, 1963).
Wutz, M., Adam, H. Walcher, W., Theory and Practice of Vacuum Technology
(Friedr. Vieweg und Sohn, Braunschweig/Wiesbaden, Germany, 1989).
Appendix B
Manufacturers
See Tables B.1, B.2 and B.3.
Table B.1 United States, UK, and Canadian manufacturers

Company Products
12R Tools for scientists
PO Box 159
Cheltenham, PA 19012
Tel: 215-379-3333
Fax: 215-663-8847
II-VI Infrared materials
375 Saxonburg Blvd.
Saxonburg, PA 16056
Tel: 412-352-4455
Fax: 412-352-4980
Accuratus Macor machineable glass; ceramics
RD4 Brass Castle Rd.
Washington, NJ 07882
Tel: 908-689-0880
Fax: 908-689-8796
Ace Glass, Inc. Glass products
1430 Northwest Blvd.
PO Box 688
Vineland, NJ 08360
Tel: 800-223-4524
Fax: 800-543-6752
or
63941 S. Hancock St.
PO Box 996
Louisville, KY 40201
Tel: 800-626-5381
Fax: 800-635-4698
(continued)

DOI 10.1007/978-3-319-17668-0
598 Appendix B
Table B.1 (continued)

Company Products
Ace Plastic Co. Plastic balls for crystal models
494 Wortman Ave.
Brooklyn, NY 11208-5428
Tel: 212-523-5500
Fax: 407-659-3816
ACF Metals Metal foils, carbon foils
2239 E. Kleindale Rd.
Tucson AZ 85719-2440
Tel: 520-325-9557
Fax: 520-32509493
Acheson Colloids Aquadag—colloidal graphite
1607 Washington Ave.
PO Drawer 611747
Port Huron, MI 48061-1747
Tel: 800-255-1908
Fax: 313-984-1446
Ad-Vance Magnetics, Inc. Magnetic shields
625 Monroe St.
PO Box 69
Rochester IN 46975
Tel: 219-223-3158
Fax: 219-223-2524
Advanced Ceramics Corp. Graphite monochromators
P.O. Box 94924
Cleveland, OH 44101-4924
or
11907 Madison Ave.
Lakewood,OH 44107-5026
Tel: 800-822-4322
Fax: 216-529-3954
Advent Associates, Ltd. Ion guns, vacuum components
6663 Woodwell St.
Pittsburgh, PA 15217
Tel: 412-422-8100
Fax: 412-422-8101
Advent Research Materials Metals, alloys, foils, sheets, wires, rods, tubes
LTD
Blyth Road Industrial Estate
Halesworth Suffolk
England 1P19 8DD
Tel: 44 1986 874555
Fax: 44 1986874557
AESAR Single crystals
Johnson Matthey, Inc.
PO Box 1087
Seabrook, NH 03874
Tel: 1-800-343-1990
(continued)
Appendix B 599

Company Products
AIRCO Specialty gases
Lansbrook Rd.
Acton, MA 01720
Tel: 508-263-7769
Fax: 508-263-6014
Aithaca Single crystals
50 Charles Lindbergh Blvd.
Suite 400
Uniondale, NY 11553
Tel: 516-229-2330
Fax: 516-227-2350
Alberox Feedthroughs
Industrial Park
New Bedford, MA
02745-1289
Tel: 508-995-1725
Fax: 508-995-6954
Alcatel Pumps, mechanical
Phil Danielson
Danielson Associates, Inc.
PO Box 157
51471/2 Main St.
Donners Grove, IL 60515
Tel: 612-370-1279
Aldrich Chemical Co. Chemicals
PO Box 2060
Milwaukee, WI 53201
Tel: 414-273-3850
Fax: 414-273-4979
Alfa Chemicals, high purity electronic grade
Johnson Matthey
30 Bond St.
Ward Hill, MA 01835-8099
Tel: 800-343-0660
Fax: 800-322-4757
Allen Datagraph, Inc. X-Y recorders/plotters
Industrial Way
Salem, NH 03079
Tel: 800-258-6360
Fax: 603-893-9042
Alltech Assoc. Inc. Graphite fittings for Swagelok fittings
2051 Waukegan Rd.
Deerfield, IL 60015
Tel: 800-255-8324
Fax: 708-948-1078
American Laboratory Annual buyer’s edition
PO Box 3220
Lowell, MA 01853-9878
(continued)
600 Appendix B

Company Products
Analog Devices Computer chips and boards
One Technology Way
Norwood, MA 02062-9106
Tel: 617-329-4700
Fax: 617-326-8703
APD Cryogenics, Inc. Cryopumps; Helitron He coolers
1833 Vultee St.
Allentown, PA 18103-4783
Tel: 610-791-6700
Fax: 610-791-0440
Aremco Single crystals
23 Snowden Ave.
PO Box 429
Ossining,NY 10562-0429
Tel: 914-762-0685
Fax: 914-762-1663
Automation Gages, Inc. Translation stages
850 Hudson Ave.
Rochester, NY 14621
Tel: 1-800-922-0329
Fax: 716-338-7370
Baikowski International Alumina polishing powder
Corp.
6006 Old Pineville Rd.
Charlotte, NC 28210
Tel: 704-523-1707
Baker Process Equipment Filters, dryers, purifiers
PO Box 898
308-Moon Clinton Rd.
Coraopolis, PA 15108
Tel: 412-264-8271
Fax: 412-264-8758
Balzers Mass spectrometers
8 Sayamore Park Rd.
Hudson, NH 03051
Tel: 603-595-3250
Fax: 603-595-3200
Barnes/Spectra Tech. IR spectrometers
652 Glenbrook Rd.
Stamford, CT 06906
Tel: 800-243-9186
Bearings, Inc. Ball bearing power transmission units
5536 Baum Blvd.
Tel: 412-681-4900
Berghof/America Teflon products
PO Box 6029
Concord, CA 94524
Tel: 800-544-5004
Fax: 510-827-1189
(continued)
Appendix B 601

Company Products
Bertan High-voltage power supplies
121 New South Rd.
Hicksville, NY 11801
Tel: 516-433-3110
Fax: 516-935-1766
Bid Service Surplus vacuum and semiconductor processing equipment
PO Box 128
Bradley Beach, NJ 07720
Tel: 908-775-8300
Fax: 908-774-1443
Blake Industries, Inc. X-ray instruments
660 Jerusalem Rd.
Scotch Plains, NJ 07076
Tel: 908-233-7240
Bomco Inc. Quartz-Mo seals
125 Gloucester Ave.
Gloucester, MA 01930-2294
Tel: 508-283-9000
Fax: 508-283-2882
Bristol Organics LTD Fluorine compounds (organic)
Sharpness Docks
Berkeley
Glos GL 13 9UG
UK
Tel: 0453-811400
Brush-Wellman Beryllium
Elmore, OH 43416
Tel: 419-862-2745
Buckbee Mears Corp. Photolithography W grid
245 East 6th St.
St. Paul, MN 55101-1959
Tel: 612-228-6400
Fax: 612-228-6572
Buehler Analyst Crystal cutting, polishing, fiber optics, optical instrumentation
41 Waukegan Rd.
PO Box 1
Lake Bluff, IL 60044
Tel: 847-295-6500
Fax: 847-295-7979
C&H Sales Surplus equipment
2176 E. Colorado Blvd.
PO Box 5356
Pasadena, CA 91117-9988
Tel: 800-325-9465
Fax: 818-796-4875
C.M. Furnaces, Inc. Evaporation sources, filaments
103 Dewey St.
Bloomfield, NJ 07003
Tel: 201-338-6500
Fax: 201-338-1625
(continued)
602 Appendix B

Company Products
California Fine Wire Co. Wire, thermocouple wire, fine wire
PO Box 446
Grover, CA 93483-0446
Tel: 805-489-5144
Fax: 805-489-5352
Cambridge Isotopes Isotopes, stable
50 Frontage Rd.
Andover, MA 01810-5413
Tel: 800-322-1174
Fax: 508-749-2768
Canberra Industries, Inc. Counting electronic
Nuclear Products Group
800 Research Pkwy.
Meriden, CT 06450
Tel: 203-238-2351
Fax: 203-235-1347
Ceramaseal Feedthroughs
PO Box 260, Rte. 20
New Lebanon, NY 12125
Tel: 518-794-7800
Fax: 518-794-8080
Cercoa Ceramic materials
ECS Co. 55 Webster Ave.
New Rochelle, NY 10801
Tel: 914-636-4325
Fax: 914-636-3701
Chemvac Associates Vacuum joints
808 High Point Ave.
Virginia Beach, VA 23451
Tel: 804-422-9823
Chromalox Heaters
641 Alpha Dr.
Tel: 800-443-2640
Fax: 412-967-5148
Cliftronics, Inc. Electronics equipment and supplies, electron guns
2 South St.
Clifton Springs, NY 14432
Tel: 315-462-9471
Cole-Palmer Laboratory supplies-general
7425 North Oak Park Ave.
Chicago, IL 60648
Tel: 800-323-4340
Comstock Electron energy analyzers, electron guns
1005 Alvin Weinberg Dr.
Oak Ridge, TN 37830
Tel: 615-483-7690
Fax: 615-481-3884
(continued)
Appendix B 603

Company Products
Connecticut Hard Rubber Teflon tape
Co.
PO Box 2560
Meriden, CT 06450
Tel: 203-634-0555
Connecticut Technology Langmuir-Blodgett film deposition
PO Box 524
Stratford, CT 06497
Tel: 203-375-5015
Fax: 203-375-0541
Cooke Vacuum Products Vacuum components
13 Merritt St.
South Norwalk, CT 06854
Tel: 203-853-9500
Fax: 203-858-9553
Copper & Brass Sales UHV soft aluminized bronze
17401 Ten Mile
East Detroit, MI 48021
Tel: 800-926-2600
Fax: 313-775-5078
Cory Lab Inc. Sapphire/metal feedthroughs
823 5th St.
Menominee, MI 49858
Tel: 906-863-9336
Cotronics Ceramic materials
3379 Shore Pkwy.
Brooklyn, NY 11235
Tel: 718-646-7996
Fax: 718-646-3028
CREE Silicon carbide
2810 Meridian Pkwy.
Durham, NC 27713
Tel: 919-361-5709
Fax: 919-361-4630
Cryofab Cryogenic equipment, dewars, etc.
PO Box 485
540 N. Michigan Ave.
Kenilworth, NJ 07033
Tel: 207-686-3636
Fax: 908-686-9536
Cryomech Inc. Cryogenic equipment
1630 Erie Blvd. E.
Syracuse, NY 13210
Tel: 315-475-9692
Fax: 315-422-1202
(continued)
604 Appendix B

Company Products
Custom Cable Corp. Eletric equipment and supplies, coaxial cable
242 Butler St.
PO Box 1050
Westbury NY 11590-3193
Tel: 800-242-2253
Fax: 516-334-3989
CVC Vacuum products
525 Lee Rd.
PO Box 1886
Rochester, NY 14603
Tel: 716-458-2550
CVT Engineering Ltd. Vacuum transfer devices
110 Shawmut Rd.
Canton, MA 02021
Dabilier Ion sources
207 Radley Rd.
Abingdon Oxon
OX14 3XA
England
Tel: 0235-28271
Daedal, Inc. Positioning equipment
1140 Sandy Hill Rd.
Harrison City, PA 15636
Tel: 800-245-6903
Fax: 412-861-3330
DEI Pulse generators (high voltage)
2301 Research Blvd.
Suite 105
Fort Collins, CO 80526
Tel: 970-493-1901
Fax: 970-493-1903
DoAll Northern IL. Co. Machine tools, precision gauges, cutting tools
236 Laurel Ave.
Des Plaines, IL 60016
Tel: 800-955-8191
Fax: 708-824-4340
Dow Corning Silicon products
PO Box 0994
Midland, MI 48640-0994
Tel: 517-496-4000
Duniway Stockroom Corp. Rebuilt ion pumps and vacuum components
1600 N. Shoreline Blvd.
Tel: 800-446-8811
Fax: 415-965-0764
Dupont Polymers
1007 Market St.
N-2426-A
Wilmington, DE 19898
Tel: 800-527-2601
Fax: 302-773-2291
(continued)
Appendix B 605

Company Products
E. McGrath, Inc. Used and new surplus equipment
35 Osborne St.
Salem, MA 01970-2599
Tel: 508-744-3546
Fax: 508-741-4020
Ealing Corp. Optical components
89 Doug Brown Way
Holliston, MA 01746
Tel: 800-343-4912
Fax: 508-429-7893
East Coast Sales Tools-semiconductors
11 Business Park Dr.
Armonk, NY 10504
Tel: 914-273-1400
Fax: 914-273-5866
Eastman Kodak Co. Irtran optics
5/81 RC MC 02010
Rochester, NY 14650-2010
Tel: 716-722-4385 ext. 1704
Fax: 716-477-4947
Edmond Scientific Optics, etc.
A-415 Edscorp. Bldg.
Barrington, NJ 08007
Tel: 609-573-6250
Fax: 609-573-6295
Edwards High Vacuum Vacuum components
301 Ballardvale St.
Wilmington, MA 01887
Tel: 508-658-5410
Fax: 508-658-7969
Electrofusion Corp. Beryllium windows
34325 Ardenwood Blvd.
Fremont, CA 94536
Tel: 415-795-1100
Electronic Space Products
International
1050 Bensen Way
Ashland, OR 97520
Tel: 800-638-2581
Energy Beam Sciences LaB6 cathodes
11 Bowles Rd.
PO Box 468
Agawam MA 01001
Tel: 800-992-9037
Fax: 413-789-2786
(continued)
606 Appendix B

Company Products
Energy Research & Aluminum foams
Generation, Inc.
900 Stanford Ave.
Oakland, CA 94608
Tel: 510-658-9785
Fax: 510-658-7428
Engelhard Industries Liquid bright gold
Division
Engelhard Corp.
Electronic Materials Systems
1 West Central Ave.
East Newark, NJ 07029
Tel: 201-268-7800
EPI MBE Products Group MBE products
1290 Hammond Rd.
St. Paul, MN 55110
Tel: 612-653-0488
Fax: 612-653-0725
Esco Optical materials-sapphire
171 Oakridge Rd.
Oakridge, NJ 07438
Tel: 800-327-1399
Fax: 201-697-3011
Extrel Products Quadrupole mass spectrometers
575 Epsilon Dr.
Box 11512
Pittsburgh, PA 15238-2838
Tel: 412-963-7530
Fax: 412-963-6578
Fairfield Synthesis Custom organic synthesis
PO Box 20
Blythewood, SC 29016
Tel: 803-754-3856
Fax: 803-754-8833
Farchan Organic chemicals
2603 NW 74th Place
Gainesville, FL 32653
Tel: 352-378-5864
FEI Electron guns
PO Box 654
McMinnville, OR 97128
Tel: 503-648-7075
Fil-Tech Vacuum supply filaments
1143 Diamond St.
San Francisco, CA 94114
(continued)
Appendix B 607

Company Products
Fine Science Tools Surgical tools
373 Vintage Park Dr.
Foster City, CA 94404
Tel: 800-521-2109
Fax: 415-349-3729
e-mail: info@finescience.
com
Fluorocarbon Corp. Teflon
10871 Kyle St.
PO Box 520
Los Alamitas, CA 90720
Tel: 213-594-0941
Fluorochem Ltd. Molecular formular index
Wesley St.
Old Glossop
Derbyshire SK13 9RY
England
Tel: 04574-68921
Fluoroware Teflon-semiconductor tools, handling equipment
102 Jonathan Blvd. N.
Chaska, MN 55318
Tel: 612-448-3131
Fax: 612-448-5576
Foametal, Inc. Metal foams
37645 Vine St.
Willoughby, OH 44904
Tel: 216-951-9081
Fax: 216-951-9360
Friedfich & Diownick, Inc. Glass capillary
2127 Wheaton Ave.
PO Box 230
Millville, NJ 08332
Tel: 609-825-0305
Fax: 609-327-4299
Furon Gaskets, seals
4402 Corporate Center Dr.
Los Alamitos, CA 90720
Tel: 714-995-1818
Fax: 714-761-1270
Fyrepel Aluminized fiberglass sheet
PO Box 518
Newark, OH 43055
Tel: 614-344-0391
Fax: 614-344-6025
Galileo Channeltron and chaneel plate multipliers
Galileo Park
PO Box 550
Sturbridge, MA 01566
Tel: 800-648-1800
Fax: 508-347-9191
(continued)
608 Appendix B

Company Products
Gallard/Schlesinger Chemicals, metals
584 Mineola Ave.
Carle Place, NY 11514
Tel: 516-333-5600
Fax: 516-333-5628
Gast Manufacturing Corp. Industrial equipment and supplies, carbon vane vacuum pump
2300 NM 63 #139
Benton Harbor, MI 49022
Tel: 616-926-6171
Glassfab, Inc.
PO Box 1880
Rochester, NY 14603
Tel: 716-262-4000
Fax: 716-454-4305
Goodfellow Metals Metal foils
800 Lancaster Ave.
Berwyn, PA 19302-1780
Tel: 610-640-1612
Fax: 800-283-2020
Hamamatsu Photomultipliers
420 South Ave.
Middlesex, NJ 08846
Tel: 908-469-6640
Hamilton Technology, Inc. HAVAR foil
101 N. Queen St.
PO Box 4787
Lancaster, PA 17604
Tel: 717-299-2581
Fax: 717-397-8510
Hanovia Division of Organometallic paints
Engelhard Industries
1 West Central Ave.
East Newark, NJ 07029
Tel: 201-268-7800
Hansen R.G. & Associates Laboratory cryogenic systems and components
631 Chapala St.
Santa Barbara, CA 93101
Tel: 805-564-3388
Fax: 805-963-0733
Harrick Scientific Corp. IR-VIS-UV accessories, spectrometers
Croton Dam Rd.
Box 867
Ossining, NY 10562
Tel: 914-762-0020
(continued)
Appendix B 609

Company Products
Harshaw Optical IR and UV windows
6801 Cochran Rd.
Solon, OH 44139
Tel: 800-472-5656
Fax: 216-349-6581
Helicoflex Vacuum fittings
400 Myrtle Ave.
Boonton, NJ 07005
Tel: 201-334-8100
Helicoflex Co. Seals
118W. 12th St.
Parkerburg, WV 26101
Tel: 304-428-1644
Fax: 304-428-3490
e-mail: [email protected]
Helicoflex Co. Seals
PO Box 9889
Columbia, SC 29290
Tel: 803-783-1880
Fax: 803-783-4279
Hex Single crystals
Metal Specialties Inc.
515 Commerce Dr.
Fairfield, CT 06430
Tel: 203-384-0331
Fax: 203-368-4082
High Vacuum Apparatus Gate valves
7055 Engle Rd.
Suite 6-607
Middleburg Hts., OH 44130
Tel: 216-826-0559
Fax: 216-826-1171
Honeywell Electronics
Industrial Automation &
Controls
1100 Virginia Dr.
Washington, PA 19034
Honeywell Micro Electronics
Switch Div.
11 W. Spring St.
Freeport, IL 61032-4353
Tel: 800-537-6945
Fax: 815-235-5574
IBM Instruments, Inc. FTIR spectrometers
Orchard Park
PO Box 332
Danbury, CT 06810
(continued)
610 Appendix B

Company Products
ICON Isotopes
19 Ox Bow Ln.
Summit, NJ 07901
Tel: 908-273-0449
Fax: 908-273-0449
Indium Corporation of Indium
America
PO Box 269
1676 Lincoln Ave.
Utica, NY 13503-0269
Tel: 800-448-9240
Fax: 800-221-5759
Infrared Laboratories, Inc. Dewars, He detectors
1808 E. 17th St.
Tucson, AZ 85719-6505
Tel: 520-622-7074
Fax: 520-623-0765
Insaco Sapphire fabrications
1365 Canary Rd.
PO Box 9006
Quakertown, PA
18951-9006
Tel: 215-536-3500
Fax: 215-536-7750
Instrument Specialties Co. Springs
PO Box A
Delaware Water Gap, PA
18327
Tel: 717-424-8510
Fax: 717-424-6213
Instruments SA, Inc. Monochromators, optical
3880 Park Ave.
Edison, NJ 08820-3097
Tel: 800-438-7739
Fax: 908-549-5125
Insulator Seal, Inc. Feedthroughs-ceramic/metal
23874-B Cabot Blvd.
Hayward, CA 94545-1661
Tel: 800-548-9509
Fax: 510-887-7475
Intek Services, Inc. Teflon suspension
7 McMillan Way
Suite 101
Newark, DE 19713
Tel: 302-366-8530
(continued)
Appendix B 611

Company Products
International Crystal Lab IR accessories
11 Erie St.
Garfield, NJ 07026
Tel: 201-478-8944
Fax: 201-478-4201
IonSpec Fourier transform mass spectrometry
One Longstreet
Irvine, CA 92714
Tel: 714-857-0793
Isotec, Inc. Isotopic gases
3858 Benner Rd.
Miamisburg, OH
45432-4304
Tel: 513-859-1808
Fax: 513-859-4878
Israelachvili Instrument Surface force intrument
2233 Foothill Ln.
Tel: 805-899-9292
Fax: 805-899-9190
J&H Berge, Inc. Ball bearings, ultrasonic (nalgene) cleaners
4111 S.Clinton Ave.
So. Plainfield, NJ
07080-0310
Tel: 1-800-684-1234
Jensen Tools Tools
7815 S. 46th St.
Phoenix, AZ 85044-5399
Tel: 800-426-1194
Fax: 800-366-9662
JML Direct Optical components
690 Portland Ave.
Tel: 716-342-9482
Fax: 716-342-6125
Keithley Electronics-measurement
28775 Aurora Rd.
Cleveland, OH 44139-1891
Tel: 216-248-0400
Fax: 216-248-6168
Kepco Power supplies
131-38 Sanford Ave.
Flushing, NY 11352
Tel: 718-461-7000
Fax: 718-767-1102
(continued)
612 Appendix B

Company Products
Kimball Physics Inc. Electron and ion guns, LaB6 emitters
311 Kimball Hill Rd.
Wilton, NH 03086
Tel: 603-878-1616
Fax: 603-878-3700
Kulite Semiconductor Strain gauge
Products, Inc.
1 Willow Tree Rd.
Leonia, NJ 07605
Tel: 201-461-0900
Fax: 201-461-0990
Lab Safety Supply Safety lab equipment
PO Box 1368
Janesville, WI 53547-1368
Tel: 800-356-0783
Fax: 800-543-9910
Lake Shore Cryotonics, Inc. Cryogenics,
64 Walnut St. cryogenic thermometry
Westerville, OH 43081-2399
Tel: 614-891-2243
Fax: 614-891-1392
Lamba Power supplies
515 Broad Hollow Rd.
Melville, NY 11747-3700
Tel: 516-694-4200
Lasertechnics Pulsed valve
5500 Wilshire Ave., NE
Albuquerque, NM 87113
Tel: 505-822-1123
Fax: 505-821-2213
LDS Vacuum Products, Inc. Vacuum supplies
980 Sunshine Ln.
Aitamonte Springs,
FL 32714
Tel: 407-862-4643
Fax: 407-862-8723
Lebow Co.
5960 Mandarin Ave.
Goleta, CA 93117
Tel: 805-964-7117
Fax: 805-964-7117
Leisk Flanges and gaskets
Albert Dr., Burgess Hill
West Sussex, RH15 9NX
England
Tel: 044-46-45821 2
(continued)
Appendix B 613

Company Products
Lenox Laser Laser beam machining, laser repair
1 Green Glade St.
Phoenix, MD 21131
Tel: 301-592-3106
Lesker, Kurt J. Co. Vacuum components
1515 Worthington Ave.
Clairton, PA 15025
Tel: 412-233-4200
Fax: 412-233-4275
Linseis X-Y recorders
PO Box 666
Princeton, NJ 08550
Tel: 609-799-6282
Fax: 209-799-7739
Love Controls Corp. Temperature controllers
475 Wheeling Rd.
Wheeling, IL 60090
Tel: 800-828-4588
Fax: 708-541-7366
M-G Industries Gases
175 Meister Ave.
PO Box 5328
North Branch, NJ 08876
Tel: 610-695-7400
Fax: 610-695-7600
Magnetic Shielding Corp. Magnetic shields
750 N. Thomas Dr.
Bensenville, IL 60106
Tel: 708-766-7800
Fax: 708-766-2813
Mark V Laboratory Metallographic equipment
18 Kripes Rd.
East Granby, CT 06026
Tel: 800-243-9776
Fax: 203-653-4087
Markson Lboratory supplies/general
PO Box 1359
Hillsboro, OR 97123-9981
Tel: 800-528-5114
Fax: 800-858-2243
Materials Research Corp. High-purity materials for thin film deposition
542 Rte. 303
Orangeburg, NY 10962
Tel: 914-359-4200
Fax: 914-359-3453
McAllister STM, Kelvin probe, vacuum components
W.280-T Prairie Ave.
Cover D’Alend, ID 83814
Tel: 208-772-9527
(continued)
614 Appendix B

Company Products
MDC Vacuum components
2384 Cabot Blvd.
Hayward, CA 94545-1651
Tel: 800-443-8817
Fax: 510-887-0626
Melles Grot Optical items
1770 Kettering St.
Irvine, CA 92714
Tel: 800-835-2626
Fax 714-261-7589
Metal Bellows Bellows, etc.
1075 Providence Highway
Sharon, MA 02067
Tel: 617-784-1400
Fax: 617-784-1405
Metal Crystals & Oxides, Metals (pure), metal foils, single crytstal metals
Ltd.
Button End, Harston
CB2 5WX
Cambridge, England
Tel: 0223-872022
Metaleurop Single crystals
111 Richmond St. West
Suite 418
Toronto, Ontario Canada
Tel: 416-366-3954
Fax: 416-366-0586
Meyer Tool & Mfg., Inc. Vacuum chambers-fabrication
4601 Southwest Highway
Oak Lawn, IL 60453
Tel: 708-425-9080
Fax: 708-425-2612
Micromeritics Particle size and surface area measurement instruments
One Micrometitics Dr.
Norcross, GA 30093-1877
Tel: 770-662-3633
Fax: 770-662-3696
Midac IR emission
17911 Fifth Ave.
Irvine, CA 92714
Tel: 714-660-8558
Fax: 714-660-9334
Miller-Stephenson Chemical Fluorocarbon lubricants
Co.
Beckus Ave.
Danbury, CT 06810
Tel: 203-743-4447
Fax: 203-791-8702
(continued)
Appendix B 615

Company Products
Minnesota Valley Vacuum insulated piping systems
Engineering, Inc.
407 Seventh St. NW
New Prague, MN 56071
Tel: 612-758-4484
MKS Capacitance manometers
Six Shattuch Rd.
Andover, MA 01810-2449
Tel: 800-227-8766
Fax: 508-975-0093
MMR Technologies Cryogenic-closed cycle refrigerators
1400 N. Shoreline Blvd. A5
Mountain View, CA
94043-1312
Tel: 415-962-9620
Fax: 415-962-9647
Monocrystals Single crystals
24400 Highland Rd.
Richmond Heights, OH
44143
Tel: 216-531-9820
Fax: 216-531-9789
MSA Air pollution sensors
PO Box 426
Tel: 800-672-2222
Fax: 412-967-3451
MSC Machinist’s tools
151 Sunnyside Blvd.
Plainview, NY 11803
Tel: 800-753-7937
Fax: 516-349-0265
MTS Systems Corp.
Box 24012
Minneapolis, MN 55424
Tel: 612-937-4000
Fax: 612-937-4515
Munns Mfg., Inc. Vacuum equipment (UHV)
5885 La Ribera St.
Livermore.CA 94550
Tel: 510-447-0310
Fax: 510-447-0110
MVE Vacuum-insulated products
8011 34th Ave., S.
Bloomington, MN
55425-1636
Tel: 800-247-4446
Fax: 612-853-9661
(continued)
616 Appendix B

Company Products
National Instruments LabVIEW, other instrumentation for data analyses and
6504 Bridge Point Pkwy. acquistion
Austin, TX 78730-5039
Tel: 512-794-0100
Fax: 512-794-8411
Website: www.natinst.com
Neslab Instruments Cooling units
PO Box 1178
Portsmouth, NH 03802-1178
Tel: 603-436-9444
Fax: 603-436-8411
New Focus, Inc. Optical components
340 Pioneer Way
Tel: 415-961-2108
Fax: 415-961-6072
Newark Electronics Electronics distributor
4801 N. Ravenswood Ave.
Chicago, IL 60640-4496
Tel: 312-784-5100
Fax: 312-907-5217
Newport Corp. Pulsed valves, laser optical supports
1791 Deere Ave.
Irvine, CA 92714
Tel: 800-222-6440
Fax: 714-253-1680
Newport Optical Optical components
Costa Mesa, CA
Nolan Supply Tools
PO Box 6289
111-115 Leo Ave.
Syracuse, NY 13217
Tel: 800-736-2204
Fax: 315-463-0316
Omega Thermocouples, etc.
PO Box 4047
One Omega Dr.
Stamford, CT 06907
Tel: 800-826-6342
Fax: 203-359-7700
Opthos Instruments Inc. RF sources
17805 Caddy Dr.
Rockville, MD 20855
Tel: 301-926-0589
Fax: 301-963-3060
(continued)
Appendix B 617

Company Products
Oriel Micrometer positioners
250 Long Beach Blvd.
PO Box 872
Stratford, CT 06497-0872
Tel: 203-377-8282
Fax: 203-378-2457
Ortec Electronics
PO Box 1397
Easley, SC 29640
Tel: 803-859-1471
Fax: 803-859-8580
Osaka Vacuum Turbo pumps
5547 Beacon St.
Tel: 412-422-8100
Fax: 412-422-8101
Park Scientific Instruments STM
1171 Borregas Ave.
Sunnyvale, CA 94089-1304
Tel: 800-776-1602
Fax: 408-747-1601
PCR Fluorinated compoundsm silicanes
PO Box 14166
Gainesville, FL 32802
Tel: 904-376-8246
Fax: 904-371-6246
Perkin Elmer Vacuum equipment and electron spectrometers
6509 Flying Cloud Dr.
Eden Prairie, MN 55344
Tel: 800-237-3603
Fax: 612-828-6322
Petrarch Systems Inc. Silicon compounds
Bartram Rd.
Bristol, PA 19007
Tel: 215-781-9255
Pfizer, Inc. Chemicals, pyrolytic graphite
235 E. 42nd St.
New York, NY 10017
Tel: 212-573-2323
Fax: 212-573-1166
Phoenix Equipment Co. Used equipment-scientific
100 Fernwood Ave.
Tel: 716-266-5550
Fax: 716-266-1598
(continued)
618 Appendix B

Company Products
PIC Design Gears
PO Box 1004
Benson Rd.
Middlebury, CT 06762
Tel: 800-243-6125
Fax: 203-758-8271
Pittsburgh Corning Insulation/fiberglass
800 Presque Isle Dr.
Tel: 412-327-6100
Fax: 412-327-5890
Polysciences Metals
400 Valley Rd.
Warrington, PA 18976-2590
Tel: 800-523-2575
Fax: 800-343-3291
e-mail: [email protected].
com
Polytec PI Inc. Piezoelectric translators
3001 Redhill Ave.
Costa Mesa, CA 92626
Tel: 714-850-1835
Fax: 714-850-1831
Praxair Specialty gases
3101 Preble Ave, NE
Tel: 412-766-6900
Fax: 412-766-9532
Precision Laser Services Inc. Drilling, welding, cutting
314 E. Wallace
Fort Wayne, IN 46803
Tel: 219-744-4375
Fax: 219-744-5666
Progressive Technology Inc. Ceramic tubing, washers, bolts, nuts, and balls
12122 Dry Creek Rd.
#1021
Auburn, CA 45602
Tel: 916-823-5047
Fax: 916-823-3275
Pyrometer Instrument Co. Optical pyrometers
234 Industrial Pkwy.
PO Box 70
Northvale, NJ 07647
Tel: 201-768-2000
Fax: 201-768-2570
e-mail:
[email protected]
(continued)
Appendix B 619

Company Products
Qualitech Gas Products, Inc. Gases
110 Nichol Ave.
McKees Rocks, PA 15136
Tel: 412-771-7660
Quantar Technology, Inc.
3004 Mission St.
Santa Cruz, CA 95060-5700
Tel: 408-429-5227
Fax: 408-429-5131
Queensgate Instruments, Inc. Nano positioning systems
Suite 600
90 Merrick Ave.
East Meadow, NY 11554
Tel: 516-357-3900
Fax: 516-357-9686
R.G. Hansen & Associates Cryogenic equipment
631 Chapala St.
Tel: 805-564-3388
Fax: 805-963-0733
R.M. Jordan Co. Reconditioned quadrupole mass spectrometers, etc.
990 Golden Gate Terrace
Grass Valley, CA 95945
Tel: 916-272-4580
Fax: 916-272-2955
Rama Corporation Heaters-bakeout
600 W. Esplanade Ave.
San Jacinto, CA 92583
Tel: 909-654-7351
Fax: 909-654-3748
Raychem Electronic equipment and supplies, SME alloys
300 Constitution Dr.
Menlo Park, CA 94025
Tel: 415-361-3860
RHK Technology High-resolution electron energy loss spectrometer
209 Voorheis Rd.
Pontiac, MI 48053
Tel: 313-858-8958
S&J Trading, Inc. Mica
PO Box 337
Glenoaks, NY 11004
SAES Getters, Inc. Alkali getters
1122 East Cheyenne
Mountain Blvd.
Tel: 719-576-3200
Fax: 719-576-5025
(continued)
620 Appendix B

Company Products
Sargeant-Welch Vacuum components
PO Box 5229
Buffalo Grove, IL
60089-5229
Tel: 800-727-4368
Fax: 800-676-2540
e-mail:
[email protected]
Saville & Co. Shim stock
414 Rodi Rd.
PO Box 27286
Tel: 412-731-6401
Fax: 412-731-4098
Scientific Gas Products Gases
2330 Hamilton Blvd.
S. Plainfield, NJ 07080-9986
Tel: 201-754-7700
Scientific Instrumentation
Services, Inc.
1027 Old York Rd.
Ringoes, NJ 08551-1039
Tel: 908-788-5550
Fax: 908-806-6631
Scientific Instruments Cryogenics, closed cycle refrigerators, low temperature
Rd. 3, Box 593 thermocouples (Cr/Fe:Au)
Ringoes, NJ 08551-9675
Tel: 908-788-5550
Scott Specialty Gases Gases
6141 Easton Rd.
PO Box 310
Plumsteadville, PA
18949-0310
Tel: 800-220-4436
Fax: 215-766-0320
Sensote, Inc.
1202 Chesapeake Ave.
Columbus, OH 43212
Tel: 800-848-6564 x880
Fax: 614-486-0506
Setra Pressure gauge, digital capacitance manometers
45 Nagog Park
Acton, MA 01720
Tel: 508-263-1400 or
800-257-3872
Fax: 508-264-0292
(continued)
Appendix B 621

Company Products
Shipley Company, Inc. Photo resist coating materials
2300 Washington St.
Newton, MA 02162
Tel: 617-969-5500
Sigmund Cohn Corp. Very fine wire, precision fabricated metals
121 S. Columbus Ave.
Mt. Vernon, NY 10553
Tel: 914-664-5300
Fax: 914-664-5377
Small Parts Parts, small nut, bolts, gears, washers, miscellaneous
13980 NW 58th Ct. parts-stainless steel PO mechanical components
Box 4650
Miami Lakes, FL
33014-0650
Tel: 800-220-4242
Fax: 800-423-9009
South Bay Technology Crystal cutting equipment
1120 Via Callejon
San Clemente, CA 92672
Tel: 714-492-2600
Fax: 714-492-1499
Space Environment Vacseal leak sealer
Laboratories
Box 1061
Boulder, CO 80306
Tel: 303-443-4090
Specialty Minerals, Inc. Pyrolytic graphite
640 N. 13th St.
Eastern, PA 18042
Tel: 610-250-3321
Fax: 610-250-3325
Spectra Tech Ion sources, heaters
2 Res. Dr.
PO Box 869
Shelton, CT 06484-0869
Tel: 800-243-9186
Fax: 203-926-8909
Spectra-Mat, Inc. Ion sources
100 Westgate Dr.
Watsonville, CA 95076
Tel: 408-722-4116
Fax: 408-722-4172
Starrett Tool chests, precision tools
29 McKean St.
Tel: 412-431-8700
Fax: 412-431-5150
(continued)
622 Appendix B

Company Products
Stavac Associates Ion sources
114 Black Thorn Dr.
Butler, PA 16001
Tel: 412-586-2100
Fax: 412-586-5927
Strem Chemicals, custom synthesis
7 Mulken Way
Dexter Industrial Park
Newburyport, MA
01950-4098
Tel: 508-462-3191
Fax: 508-465-3104
Sumitomo Electric GaAs
551 Madison Ave.
New York, NY 10022
Tel: 212-308-6444
Surface Interface, Inc. Manipulators, UHV banana plugs
110 Pioneer Way Suite
Tel: 415-965-8205
Fax: 415-965-8207
Swagelok Vacuum fittings
31400 Aurora Rd.
Solon, OH 44139
Tel: 216-439-5934
Fax: 216-349-5843
Swiss Jewel Co. Sapphire-fabricated barings, etc.
613 Lafayette Bldg.
Philadelphia, PA 19106
Tel: 215-925-2867
Fax: 215-922-3055
T.E. Brown & Co. Spark cutter
Button End
Harston
Cambridge England
Tel: 0223-872072
Technical Devices, Inc. Wire bonding tool
5721 Prescott Rd.
Prescott, AZ 86301
Tel: 602-772-7500
Technitool Tools, specialized
5 Apollo Rd.
Box 368
Plymouth Meeting, PA
19462
Tel: 610-941-2400
(continued)
Appendix B 623

Company Products
Fax: 610-828-5623
Terranova Pressure measurement, vacuum valves
5841 Bell Rd.
Auburn, CA 95602
Tel: 800-519-2634
Fax: 916-269-2877
Test Lab Co. Reconditioned test equipment
1066 Linda Vista Ave.
Tel: 800-42-5835
Thermionics UHV apparatus and accessories
22815 Sutro St.
PO Box 3711
Hayward, CA 94540
Tel: 510-538-3304
Fax: 510-538-2889
Thorlabs Inc. Optical components
435 Rte. 206
Newton, NJ 07860-0366
Tel: 201-579-7227
Fax: 201-383-8406
Time Motion Tools Tools
12778 Brookprinter PI.
Poway, CA 92064
Tel: 619-679-0303
Topac, Inc. Channel multipliers
99 Derby St.
Suite 303
Hingham, MA 02043
Tel: 617-740-8778
Fax: 617-740-8779
Tripp Lite Electronics
500 N. Orleans St.
Chicago, IL 60610
Tel: 312-329-1777
Fax: 312-644-6505
Tube Sales Stainless steel tubing
Interstate Industrial Park
Bellmaur, NJ 08031
Tel: 215-923-5176
Tucker Electronic Electronics
PO Box 551419
Dallas, TX 75355-1419
Tel: 800-527-4642
Fax: 214-348-0367
(continued)
624 Appendix B

Company Products
Tylan General Vacuum components
1110 Shepard Dr.
Blue Bell, PA 19422
Tel: 610-272-1933
Fax: 603-894-4909
U-C Components, Inc. Vented screws
410 Logue Ave.
Mountain View, CA
94043-4019
Tel: 415-964-3827
Fax: 415-964-0216
Uniform Tubes, Inc. Metal tubing, specialty
7th Ave.
Collegeville, PA 19426
Tel: 215-539-0700
Unimetrics Corp. Micro syringes and accessories
501 Earl Rd.
Shorewood, IL 60436
Tel: 800-854-6931
Fax: 815-741-0799
Unique Wire Weaving Co., Stainless mesh
Inc.
762 Ramsey Ave.
Hillside, NJ 07205-1094
Tel: 800-445-6912
Fax: 908-688-4601
United Chemical Silanes and silicones
Technologies, Inc.
2731 Bartram Rd.
Bristol, PA 19007
Tel: 215-781-9255
Fax: 215-785-1226
United Detector Electronic equipment and supplies
12515 Chadron Ave.
Hawthorne, CA 90250
Tel: 310-978-1150
Fax: 310-644-1727
UTI Mass spectrometers
1313 Fifth St. SE
Suite 130
Minneapolis, MN 55414
Tel: 612-378-1393
Fax: 612-378-2287
(continued)
Appendix B 625

Company Products
Vacoa Vacuum valves and accessories
10-7 Technology Dr.
E. Setauket, NY 11733
Tel: 516-751-7115
Fax: 516-751-7154
Vacuum General Capacitance gauges
9577 Chesapeake Dr.
San Diego, CA 92123
Tel: 619-571-1222
Fax: 619-576-1703
Vacuum Science Workshop, Electron spectrometers
Ltd.
5045 N. 30th St.
Tel: 719-531-7711
Fax: 719-531-7714
Varflex Glass braid
512 W.Court St.
Rome, NY 13440
Tel: 315-336-4400
Fax: 315-336-0005
Varian Turbopumps, vacuum products, feedthroughs
121 Hartwell Ave.
Lexington, MA 02173-3133
Tel: 1-800-882-7426
Fax: 617-860-5437
Vat Gate valves
600 West Cummings Park
Woburn, MA 01801
Tel: 617-935-1446
Fax: 617-935-3940
Velmex, Inc. Unislide assemblies
7550 State Routes 5 & 20
Bloomfield, NY 14469-9389
Tel: 800-642-6446
Fax: 716-657-6153
Victor Equipment Co. Gas apparatus
101 S. Hanley Rd.
St. Louis, MO 63105
Tel: 314-721-5573
Fax: 314-721-4822
Virginia Semiconductor, Inc. Silicon membranes, silicon crystals
1501 Powhatan St.
Fredericksburg, VA 22401
Tel: 540-373-2900
Fax: 540-371-0371
(continued)
626 Appendix B

Company Products
Voltaix Gases-semiconductor process
PO Box 5357
197 Meister Ave.
North Branch, NJ 08876
Tel: 908-231-9060
Fax: 908-231-9063
W.R. Grace Silica, powdered; silica, gel
7221 W. Parkland Court
Milwaukee, WI 53223
Tel: 414-354-4400
Fax: 414-354-5088
Watlow Heating elements-bakeout
2987 Babcock Blvd.
Tel: 412-311-5004
Fax: 412-322-1322
Wavemat Microwave plasma sources
44191 Plymouth Oaks Blvd.
Suite 100
Plymounth, MI 48170
Wiley Organic chemicals
1245 South Sixth St.
Coshocton, OH 43812
Tel: 614-622-0755
Fax: 614-622-3231
Wilmad Glass Co. Glass tubing-capillary
Rt. 40 & Oak Rd.
Buena, NJ 08310
Tel: 800-297-8165
Fax: 800-297-8174
Winfred M. Berg Inc. Precision mechanical components
499 Ocean Ave.
East Rockaway, NY 11518
Tel: 516-599-5010
YSI Temperature measurement
Yellow Springs Instrument
Co., Inc.
1725 Branum Ln.
Yellow Springs, OH 45387
Tel: 513-767-7241
Fax: 513-767-9353
Appendix B 627
Table B.2 German and European manufacturers

Company Products
Advent Research Materials Ltd Pure materials, metals, foils, wires, powders
Blyth Road Industrial Estate
Halesworth, Suffolk
England IP 19 8DD
Tel: 44-1986 874555
Fax: 44-1986-874557
Alcatel Hochvakuumtechnik GmbH Vacuum components, quadrupole mass spectrometer,
Am Kreuzeck 10 pumps
D-97877 Wertheim
Tel: 09342-8700
Fax: 09342-87030
Alfa Produkte Pure metals, metal products, noble metal technology,
Zeppelinstr. 7 single crystals, wires, foils, powder
D-76185 Karlsruhe
Tel: 0721-853061
Fax: 0721-849674
Araldit AtI
Ciba-Geigy GmbH
Oeflinger Str.
D-79664 Wehr (Boswn) Germany
Aurel GmbH Si-crystals
Malteser Str. 443C
D-86899 Landsberg
Balzers Hochvakuum GmbH Vacuum components
Otto-von-Guericke-Ring 7
D-65205 Wiesbaden
Tel: 06122-706-0
Fax: 06122-70610
Bender & Hobein GmbH “Surfrasi” silconizing agent, glass coating
Fraunhofer Str. 7
D-85737 Ismaning
Tel: 089-9965480
Fax: 089-9965490
Berger Lahr GmbH Synchronous and step motors of all sizes, positioning
Breslauer Str. 7 systems
D-77933 Lahr
Tel: 07821-5820
Fax: 07821-582313
or
Zuständiger Kundendienst
SIG Positec GmbH
Büro Mitte
Maybachstr. 14
D-63456 Hanau
Tel: 06181-66580
Fax: 06181-665818
(continued)
628 Appendix B

Company Products
Best Swagelok, Whitey, Cajon, Nupro, Snotrik, valves,
Ventil & Fitting GmbH glass-metal connections
Robert-Bosch-Str. 20
D-63477 Maintal-Dörnigheim
Tel: 06181-46041
Fax: 06181-432161
Busak & Shamban GmbH & Co. Gaskets
Handwerkstr. 5-7
D-70565 Stuttgart 80
Tel: 0711 7864-0
Fax: 0711-7803171
Caburn-MDC UHV components, feedthroughs, coaxial cables
Ostendstr. 1
D-12459 Berlin
Tel: 030-6953-9840
Fax: 030-635-3786
Caddock Europe High precision resistors, high voltage resistors up to
Leopoldstr. 28a/2 10000 MOhm
D-80802 München
Tel: 089-390799
Fax: 089-391230
or
Caddock Electronics Inc.
1717 Chicago Ave.
Riverside, CA 92507-2364
Tel: 001-714-788-1700
Fax: 001-714-369-1151
Carl Roth GmbH+Co Chemicals, laboratory equipment
Postfach 21 11 62
D-76161 Karlsruhe
Tel: 0721-56060
Fax: 0721-560649
CETEC Spezialarmaturen UHV components
Ernst H.F. Hiller
Vahrer Str. 248 d
D-28329 Bremen
Tel: 0421-461855
Fax: 0421-462385
CKL Vacuumtechnik Vertriebs UHV components, pumps, sputter guns
GmbH
Unterfeldring 17
D-85256 Vierkirchen
Tel: 08139-890
Fax: 08139-8943
CompoTherm High-speed pyrometers
Oberflächentechnik GmbH
Siemenstr. 16-18
D-28857 Syke
Tel: 04242-593-12
Fax: 04242-593-13
(continued)
Appendix B 629

Company Products
Conatex Technical equipment for scientists: thermometers,
Postfach 1220 thermocouples
D-66592 St. Wendel/Saar
Tel: 06851-2071
Fax: 06851-2076
Cryo Vac Cryo technology, cryo component and materials,
Dr. Johannes Klug, Handelsvertretung insulation, low temperature glue
Hasbacher Str. la
D-51503 Rösrath
Tel: 02205-85454
Fax: 02205-5797
Cryophysics Modulators, crystals, harmonic generating,
Landwehrstr. 48 polarizers, Q-Switches, lasermetrics, LN2 cryostats,
D-64293 Darmstadt super insulation
Tel: 06151-86281
Fax: 06151-84481
Cryoson GmbH Steel dewars, LN2-containers
Postfach 1104
D-63821 Schöllkrippen
Tel: 06024-1233
Fax: 06024-7879
Deutsche Air Liquide—Kryotechnik LN2 dewars, cryo technology, LHe containers
König-Adolf-Str. 3
D-65191 Wiesbaden
Tel: 0611-543048
Fax: 0611-547264
Dr. Sjuts Optotechnik Electro multipliers, channeltrons
Max-Planck-Str. 1
D-37191 Katlenburg-Lindau
Tel: 05556-1011
Fax: 05556-4587
EA-Elektro-Automatik High current power supplies
Helmholtzstr. 33-35
D-31747 Viersen
Tel: 02162-3785-0
Fax: 02162-16230
EBARA Germany GmbH Turbo and cryo pumps
Donaustr. 7
D-63452 Hanau
Tel: 06181-1876-0
Fax: 06181-1876-40
EMCO Elektronik GmbH RF power amplifiers, micro wave components
Einsteinstr. 35
D-82152 Martibsried
Fax: 089-85-97785
(continued)
630 Appendix B

Company Products
EMV Elektronische Meβgerӓte Electronic measuring instruments
Vertriebs GmbH
Wallbergstr. 7
D-82024 Taufkirchen
Tel: 089-6128054
Fax: 089-6128593
ENI Germany RF amplifiers, power amplifiers
A Division of Aster Germany GmbH
Holderӓckerstr. 15
D-70839 Gerlingen
Tel: 07156-21095
Fax: 07156-49372
Epoxy Produkte GmbH Solvent free epoxy products conducting cement (for
Guntherstr. 31 LiTaO3 contacts)
Postfach 1207
D-64658 Fürth/Odenwald
Tel: 06253-4546
Fax: 06253-21687
Farnell Advance Electronic GmbH Power supplies, measuring instruments
Dieselstr. 21
D-63533 Mainhausen
Tel: 06182-21091-93
Fax: 06182-28819
FAST Comtec GmbH Transient recorder, PC-boards, NIM boards,
Grünwalder Weg 28a multichannel scaler, discriminators, time-counters
D-82041 Obserhaching
Tel: 089-613-1081
Fax: 089-613-6171
FHR Anlagenbau, GmbH UHV, HV, sputter instruments and supplies, pure
Bergener Ring 41 gases, valves
D-01458 Ottendorf-Okrilla/Dresden
Tel: 089-613-1081
Fax: 089-613-6171
FISONS Instruments UHV components, surface science, analysis
Vertriebs-GmbH instruments
peter-Sander-Str. 43
D-55252 mainz-Kastel
Fax: 06134-289904
FOM Single crystal preparation
Kruislan 407
1098 SJ Amsterdam
The Netherlands
Tel: 31-20-6081301
Fax: 31-20-6684106
Friatec Ceramic parts, UHV feedthroughs
(ehemals Friedrichsfeld GmbH)
Postfach: 710261
D-68222 Mannheim
Tel: 0621-4860
Fax: 0621-477999
(continued)
Appendix B 631

Company Products
Fritz Eichenauer GmbH Heating Elements
Industriestr. 1
D-76770 Hatzenbühl
Tel: 07275-7020
Fax: 07275-702200
FuG Elektronik GmbH High-precision power supplies, high-voltage power
Florianstr. 2 supplies and resistors
D-83024 Rosenheim
Tel: 08031-81093
Fax: 08031-81099
General Scanning GmbH Galvo drives, scanners
Münchener Str. 2
D-82152 Planegg
Tel: 089-8595031
Fax: 089-8599487
Goodfellow GmbH Metals in all forms: foils, crystals, tubes, powder
Postfach 5427
D-65729 Eschborn
Tel: 06196-703441
Fax: 06196-482954
Gossen GmbH Mess- und General measuring instruments, controlling
Regeltechnik instruments
Postfach 1780
D-91007 Erlangen
Tel: 09131-8271
Fax: 09131-28895
Gräff High-precision thermometers,
Bonner Str. 54 thermocouple instruments
D-53842 Troisdorf
Tel: 02241-41031
Fax: 02241-404002
Hahn & Bischof UHV-, MBE-instruments, analysis instruments, nano
Donarweg 3 crystal fabrication
D-81739 München
Tel: 089-601-4409
Fax: 089-601-2748
Hameg Instruments Oscilloscopes, signal generators, multimeters
Kelsterbacher Str. 17
D-60528 Frankfurt
Tel: 069-67805-0
Fax: 069-67805-13
Hela GmbH Electrical wires, polymide cables, kapton wire
Postfach 411011
D-12120 Berlin
Tel: 030-7927080
(continued)
632 Appendix B

Company Products
Heraeus Sensor GmbH Temperature measuring program, thermocouples,
Im Stift 6-8 high-precision resistors, temperature probes
D-58119 Hagen
Tel: 02334-8040
Fax: 02334-80432
Heraeus Wittmann GmbH Silicon-heating elements, heating ribbons
Englerstr. 11
D-69126 Heidelberg
Tel: 06221-30430
Hiden Analytical Quadrupole mass spectrometer, ion sources for
Vertretung für Deutschland surface analysis
VACUA GmbH
Niedmannweg 13
D-82431 Ried/Kochel
Tel: 088-57693010
Fax: 088-579771
HMS-Electronik Hans M Strassner Extremely low-noise pre-amplifiers, lock-in
GmbH amplifiers
Am Arenzberg 42
D-51381 Leverkusen
Tel: 02171-3814
Fax: 02171-33852
Hoke GmbH Pure gases and gas handling supply
Hanauer Landstr. 11-13
D-60134 Frankfurt
Tel: 069-4050090
Fax: 069-4050094
Hositrad/Holland CVD, UHV parts, manipulators, loadlocks, Ti
P.O.B. 114 Pumps. Repair of ion pumps (Varian and Leybold)
3870 CC Hoevelaken
The Netherlands
Tel: 03495-37210
Fax: 03495-35274
Hositrad/Deutschland
Lindnergasse 2
D-93074 Regensburg
Tel: 0941-55827
Fax: 0941-562296
HP Hewlett-Packard GmbH Synthesizer, function generators, oscilloscopes, RF
Vertriebszentrum Bad Homburg components, multimeters, power supplies
Hewlett-Packard-Str.
D-61352 Bad Homburg v.d.H.
Tel: 06172-16-0
Fax: 06172-16-1309
HP Direct
Dornierstr. 7
D-71034 Böblingen
Tel: 07031-140
Fax: 07031-142100
(continued)
Appendix B 633

Company Products
Ina Wälzlager Schaeffler KG Needle bearings, linear bearings, linear motions,
Industriestr. 1-3 planetary gearings
D-91074
Herzogenaurach
Tel: 09132-820
Fax: 09242-5349
Intervac Henning GmbH Vacuum components supply
Gemeindedingerstr. 7
D-97892 Kreuzwertheim
Tel: 09342-22417
Fax: 09342-7058
Isopad GmbH Heating supplies: heating ribbons, collars, tubes, etc
Engelerstr. 11
D-69126 Heidelberg
Tel: 06221-30430
Fax: 06221-304343
Joachim Schwarz UHV components
Hindenburgstr. 12
D-76332 Bad Herrenalb
Tel: 07083-51833
Fax: 07083-51925
Joisten & Kettenbaum Tools for surface treatment, polishing supply,
Asselborner Weg 14-16 diamond paste
D-51429 Bergisch Gladbach
Tel: 02204-839-0
Fax: 02204-839-60
JUST Vakuum Technik Vacuum components, pumps, quadrupole mass
Lehnstr. 34 spectrometers
D-66869 Kusel
Tel: 06381-6060
Fax: 06381-6040
Karlsruher Glastechnisches Werk Glass instruments, cryostats for LN2 and LHe.
Schieder GmbH Specialize in insulation dewars
Spezialfabrik für Isoliergefäβe
Gablonzer Str. 6
D-76185 Karlsruhe
Tel: 0721-95897-0
Fax: 0721-95897-77
Keithley Instruments GmbH High precision measuring instruments
Landsberger Str. 65
D-82110 Germering
Tel: 089-8493070
Fax: 089-84930759
Klaus Schaefer GmbH Marketing of products: Crowcon, Extrel, General
Postfach 1660 Vacuum, Walker Scientific, Ceramaseal,
D-63206 Langen Cryomagnetics, Leak Sealant, IMAG
Tel: 06103-79085
Fax: 06103-71799
(continued)
634 Appendix B

Company Products
KMF Laborchemie GmbH “Surfrasil” siliconizing agent, glass coating
Postfach 900129
D-51111 Köln
Tel: 02241-96850
Fax: 02241-67353
Knuerr-Heinzinger Elecronic GmbH High-precision power supplies. High voltage and
Anto-Jakob-Str. 4 high-current power supplies
D-83026 Rosenheim
Tel: 08031-44040
Fax: 08031-440444
Kristallhandel Kelpin Metal single crystals, electro-optical and
Im Schilling 18 acusto-optical crystals
D-69181 Leimen
Tel: 06224-72558
Fax: 06224-77189
Kurt J. Lesker Company Vacuum components
Karl-Heinz-Beckurts-Str. 13
D-52428 Jülich
Tel: 0130-812843
Fax: 0130-812842
LeCroy GmbH Digital oscilloscopes, CAMAC modules, transient
Mannheimer Str. 177 recorders
D-69123 Heidelberg
Tel: 06221-831001
Fax: 06221-834655
Leybold Vakuum GmbH Ultra high vacuum parts and analysis instruments
Bonner Str. 498
D-50968 Köln
Tel: 0221-347-0
Fax: 0221-347-1250
or
Postfach 510760
D-50943 Köln
Linde AG Gases and gas handling supplies
Kostheimer Landstrae 25
D-55246 Mainz-Kostheim
Tel: 06134-208-31
Fax: 06134-208-25
M.K. Juchheim GmbH Thermometer, process controlling,
Moltkestr. 13-31 thermocouples, time switches
D-36039 Fulda
Tel: 0661-60030
Fax: 0661-6003-500
MaTecK Material-Technologie & Single crystals, wafers, pure materials and alloys,
Kristalle high-precision polishing service
Dr. H. Schlich
Karl-Heinz-Beckurts-Str. 13
D-52428 Jülich
Tel: 02461-690740
Fax: 02461-690749
(continued)
Appendix B 635

Company Products
Mawi-therm Thermometer, temperature controller infrared
Temperatur-Prozeβtechnik thermometer, pyrometer, thermoelements
Hofstr. 23
D-40789 Monheim
Tel: 02173-51094
Fax: 02173-58623
Melja Diagnostik GmbH “Surfrasil” siliconizing agent, glass coating
Lilienthalstr. 7-25
D-34123 Kassel
Tel: 0561-57-15-77
Fax: 0561-57-16-36
Menz & Könecke GmbH Electrical heating supplies for baking out
D-41379 Brüggen
Tel: 02163-5940
Fax: 02163-5210
Merck Chemicals and metals. Also inexpensive vacuum
Frankfurterstr. 250 parts and lubricants of different companies: Alcatel,
D-64271 Darmstadt Busch, Kinney, Tietschle, CTI, Edwards, Leybold,
Tel: 06151-723368 Welch, Sargent, Stokes, Balzers, etc. Precision Plus
Fax: 06151-722000 Vacuum Parts, Inc.
Messer Griesheim GmbH Gases and gas handling supply, liquid nitrogen and
Postfach 210553 helium
In der Steinwiese 5
D-57074 Siegen
Tel; 0271-69040
Fax: 0271-6904160
Metal Crystals & Oxides LTD Unit Single crystals, special materials
B5 Button End Industrial Estate
Harston
Cambridge
England CB2 5NX
Tel: 1223-872072
Fax: 1223-872517
Metrofunk Kabel-Union GmbH Electrical wires, polymide cables, kapton wire
Postfach 410109
D-12111 Berlin
Tel: 030-8314052
Fax: 030-7925343
MKS Instruments Deutschland GmbH Vacuum components, gas analyzer, capacitance
Schatzbogen 43 manometer, gas flow controllers
D-81829 München
Tel: 089-420008-0
Fax: 089-424106
(continued)
636 Appendix B

Company Products
Moorfield Associates VSW instruments, HREELS IB200 spectrometer.
Moorfield House Used UHV analysis instruments
Plumley Moor Rd.
Plumley
CheshierWA 16 9RS
UK
Tel: 0565-722609
Niederlassung Hanau
Wilhelm-Rohn-Str. 25
D-63405 Hanau
Tel: 06181-34-1202
Fax: 06181-34-1212
neoLAB General laboratory equipment
Rischerstr. 7
D-69123 Heidelberg
Tel: 06221-844219
Fax: 06221-844233
Normag Special glass instruments, glass grinding spindle
Labor- und Verfahrenstechnik GmbH valves
Postfach 1269
Feldstr. 1
D-65719 Hofheim a.Ts.
Tel: 06192-2099-0
Fax: 06192-209939
Omicron Vakuumphysik GmbH Rear view LEED, STM, AFM, UPS He-source,
Idsteiner Str. 78 evaporation sources
D-62232 Taunusstein
Tel: 06128-73044
Fax: 06128-73596
Oxford Instruments High precision superconducting magnets,
Kreuzberger Ring 36 cryotechnology, low temperature STM
D-65205 Wiesbaden
Tel: 0611-7640
Fax: 0611-764175
Paskovsky Meβelektronik GmbH Amplifier and linearization modules for
Postfach 1259 thermocouples
D-63502 Langenselbold
Tel: 06184-3042
Fax: 06184-62610
Perkin Elmer UHV surface analysis equipment: XPS, AES, ESCA,
Bruckmannring 40 UPS, ISS, SAM, SIMS, LEED, HEED, electron,
D-85764 Oberschleissheim photon, and ion sources, hemispherical or cylindrical
Tel: 089-315-717-0 analyzers, manipulators
Fax: 089-315-3117
Phytron-Elektronik GmbH Step motors, special clutches
Industriestr. 14
D-82194 Gröbenzell
Tel: 08142-503-0
Fax: 08142-503-190
(continued)
Appendix B 637

Company Products
Piezosystem Jena Piezo, high-precision positioning systems
Wildenbruchstr. 15
D-07745 Jena
Tel: 03641-675-390
Fax: 03641-675-393
Pink Vakuumtechnik Vacuum systems, special manufacturing
Am Kessler 6
Postfach 1509
D-97877 Wertheim
Tel: 09342-872-0
Fax: 09342-872-20
Polytec GmbH Inorganic high-temperature cements
Polytec Platz 5-7
D-76337 Waldbronn
Tel: 07243-6040
Fax: 07243-69944
Rohde und Schwarz Synthesizer, function generators, signal generators,
Vertriebs-GmbH RF components, oscilloscopes
Mühldofst.r 15
D-81671 München
Tel: 089-41290
Fax: 089-41293208
Rolf Müller Thermo electrical components, thermocouples,
Juethorn Str. 72 miniaturized welding tools
D-22043 Hamburg
Tel: 040-652-65602
Fax: 040-570-3000
Rosenkranz Elektronik Used electronic instruments, oscilloscopes, function
Gro-Gerauer Weg 55 generators multimeters
D-64295 Darmstadt
Tel: 06151-3998-0
Fax: 06151-399818
SAES GETTERS Alkali-metal dispensers, NEG pumps, getter material
Postfach 420552
Gerolsteiner Str. 1
D-50899 Köln
Tel: 021-443086
Fax: 0221-441095
SASKIA Hochvakuum—und Vacuum systems, pumps, gauges
Labortechnik
Am Vogelherd 3
D-98693 Ilmenau
Tel: 03677-6040
Fax: 03677-604110
(continued)
638 Appendix B

Company Products
Scandivac Thermionics vacuum parts, UHV-rotation tables for
Nordic Vacuum Engineering AB manipulators, electron guns, pumps, valves
Industrivägen 16
S-59041 Rimforsa
Sweden
Tel: 46-494-20530
Fax: 46-494-21197
Schlumberger Technologie GmbH CAMAC and NIM modules, crates, Si-detectors,
Ingolstädter Str. 62a digital voltmeters
D-80939-München
Tel: 089-318890
Fax: 089-35908
Schulz-Electronic GmbH Power supplies
Postfach 11 01 18
D-76486 Baden-Baden
Tel: 07223-58054
Fax: 07223-58093
SI Spectroscopy Instruments GmbH SRS Stanford Research Systems: lock-in amplifier,
Rudolf-Diesel-Str. 7a signal generators, timer counter, boxcar averager,
Postfach 1326 chopper
D-82239 Gilching
Tel: 08105-5011
Fax: 08105-5577
Sigma-Aldrich Chemie GmbH Chemical materials and equipment, glass instruments
Postfach
D-82039 Steinheim
Tel: 0130-5155
Fax: 0130-6490
Specs GmbH Surface science analysis instruments, electron,
Volastr. 5 photon and ion sources
D-13355 Berlin
Tel: 030-4633031
Fax: 030-4642083
Struck Transient recorders, CAMAC modules
Bäckerbarg 6
D-22889 Tangsted/Hamburg
Tel: 04109-55-0
Fax: 04109-5533
Struers GmbH Metallographic products, grinding and polishing
Albert Einstein Str. 5 machines, polishing supplies, diamond paste
D-40699 Erkrath
Tel: 0211-20030
Fax: 0211-200399
tectra GmbH Vacuum components, all kinds of surface science
Reuterweg 65 equipment and analysis instruments
D-60323 Frankfurt/M
Tel: 069-720040
Fax: 069-720400
(continued)
Appendix B 639

Company Products
Tektronix GmbH Digital and analog oscilloscopes, spectrum analyzers
Sedanstr. 13-17
D-50668 Köln
Tel: 0221-72220
Fax: 0221-722362
Telemeter Electronic RF amplifier, signal generators, microwave
Posthof 4 components
D-86609 Donauwörth
Tel: 0906-706-930
Fax: 0906-21706
Thermal Quarz-Schmeize GmbH Various products made of glass
Vor dem Haβel 8
D-21438 Brackel
Tel: 04185-4035
Fax: 04185-5309
VAB Vakuum Anlagen Bau GmBH UHV systems and components, Ti sublimation
Sandhöhe 11 pumps, valves, manipulators
D-25337 Elmshorn
Tel: 04121-73095
Vacom Vacuum components, pumps, valves, gauges, feed
Leutragraben 8a throughs
D-07743 Jena
Tel: 03641-443994
Fax: 03641-443993
Vakuumschmelze GmbH Permanent magnets
Grüner Weg 37
Postfach 2253
D-63456 Hanau
Tel: 06181-380
Fax: 06181-382645
Varian GmbH-Vacuum Products Vacuum components, electron guns, surface science
Kuehnstr. 71d analysis instruments
D-22045 Hamburg
Tel: 040-6696031
Fax: 040-6682282
VAT Valves, bellows, vacuum parts
Bretonischer Ring 7
D-85630 Grasbrunn-Neukeferloh
Tel: 089-465015
Fax: 089-463765
Victoreen Inc. High-precision, high-voltage resistors
6000 Cochran Rd.
Cleveland, OH 44139
Tel: 001-216-248-9300
Fax: 001-216-248-9301
or
Compotron
Geretsrieder Str. 4
D-84478 Waldkraiburg
Tel: 08638-3051
Fax: 08638-2781
(continued)
640 Appendix B

Company Products
VSE Vakuumtechnik GmbH Metal valves, leak valves, vacuum components
Rottmosserstr. 16
D-83543 Rott a. Inn
Tel: 08039-4454
Fax: 08039-4457
VSI Vakuum Science Instruments Analyzer for UPS, AES, ISS, HREELS-spectrometer
GmbH Video-LEED, electron guns, vacuum components
Bahnhofstr. 2
D-65307 Bad Schwalbach
Tel: 06124-70830
Fax: 06124-708333
Wacker-Chemie GmbH Si-single crystals
Hans-Seidel-Platz 4
D-81737 München
Tel: 089-627901
Fax: 089-62791770
Wallace & Tiernan Pennwalt manometer, aneroid-absolute manometer
Postfach 1563
D-89305 Günzburg
Tel: 08221-904-0
Fax: 08221-904-203
Wavetek Electronics GmbH Function generators, pulse generators, microwave
Freisinger Str. 34 components
D-85737 Ismaning
Tel: 089-9609490
Fax: 089-967170
Well-Diamantdrahtsägen Diamond wire saws, crystal cutting tools
Oppauerstr. 37
D-68305 Mannheim
Tel: 0621-741990
Fax: 0621-745897
Witzenmann GmbH Bellows, diaphragm bellows
Östliche Karl-Friedrich-Str. 134
D-75175 Pforzheim
Tel: 07231-5810
Fax: 07231-581820
Zinsser Analytical GmbH Pure metals, alloys, chemicals
Eschborner Landstr. 135
D-60489 Frankfurt
Tel: 069-7891060
Fax: 069-78910680
All addresses with D postal codes are located in Germany
Appendix B 641
Table B.3 Japanese manufacturers

Company Products
Anelva Gauges, ion pumps, mechanical pumps, vacuum flanges, valves, etc.
5-8-1 Yotsuya
Fuchu-city
Tokyo 183
Japan
Tel: 31423 34 0221
Fax: 81 423 60 2277
Daikin Cryopumps, cryo-refrigerator
2-4-12
Nakazaki-Nishi
Kita-ku
Osaka 530
Japan
Tel: 81 6 374 9306
Fax: 81 6 374 9296
Denka Ceramics, LaB6 cathodes
141 Yurakucho
Chiyoda-ku
Tokyo 100
Japan
Tel: 81 3 3507 5257
Fax: 81 3 3507 5078
Hamamatsu Photonics CCD camera, microchannel plate, multi-capillary plate, etc
8-21 Toranomon
Minato-ku
Tokyo 105
Japan
Tel: 81 3 3436 0491
Fax: 81 3 3433 6997
or
314-5 Toyooka-mura
Iwata-gun
Shizuoka 438-01
Japan
Tel: 81 539 62 5245
Fax: 81 539 62 2205
Irie Kiken Co., Ltd. Flexible tubes, quick-coupling (NW) components
11-7 Ginza
Chuou-ku
Tokyo 104
Japan
Tel: 81 3 3542 4692
Fax: 81 3 5565 7064
Index
Note: Page numbers followed by ‘f’ denotes Figure and ‘n’ denotes Footnote
A lubrication and, 343

Absolute adsorption Azimuthal sample motion, 3, 4f
coverage measurement, 253–255 cyrogenic manipulators for, 388–389
King and Wells method, 253, 334, 426 strain-free mounting for, 355, 356f
TPD method, 569, 571
Absoulte coverage, 253, 255, 309 B
Absolute sticking coefficient, 256 Bakeout
at desorption temperature, 256 eliminated, 45
measurement, 253–255 glass systems, 506
Adsorption heaters for, 83
gas purification and, 506f in metal UHV systems, 83–85
metal fatigue and, 95 to UHV condition, 69
powders, 576 viton seal and, 109
supported catalysts, 555 wall oxidation and, 87
Adsorption/desorption from high area solids, Bayard-Alpert gauge
569–571 contamination of, 135
Alkali metal source, 483–484 corrections for external radiation and
Aperture, adjustable iris, 30–31 external electron flux effects, 142–143
Aquadag, 11, 47, 120, 123, 196, 374, 499 filament replacement in, 145
Argon purification, 510f high sensitivity of, 262
Atomic arsenic, 456–458 ion gun based on, 215–216
Atomic fluorine, 481–482 Bearings
Atomic hydrogen high-speed rotary shafts, 29–30
detectors, 480 internal ball, 10
high-flux thermal source, 477–480 lifetime, 25
radiofrequency driven source, 480–481 rotation translation, 27f
Atomic oxygen detection, 490–491 self lubricating, 26–27
Atomic oxygen production and purity, steel ball, 29
491–492 teflon, 29
Atomization tungsten carbide ball, 27
arsenic, 456 vacuum ball, 11
fluorine, 481 Bellows, 109
hydrogen, 477, 481 folding, 587
Au hypalon/nylon, 52
evaporation source, 313 temporary leak sealing of, 108
single crystal, Au(111), 155 Beryllium window construction, 526
Au UPS spectrum, 236f Birefringence, 532
Auger electron spectroscopy (AES), 196 Bond-Barrel holder, 156, 158–159
calibration with quartz microbalance, 312 Books, experimental methods, 593

DOI 10.1007/978-3-319-17668-0
644 Index
Boron as impurity, 545 palladium, 546

Boron nitride silicon, 548
crucible, 467 tungsten, 551
diffusion, 529 CMA auger spectrometer, 310
lubricant, 26 photoelectron spectrum measurement, 236
shield, 456 Pt deposition measurement, 309, 310, 312
Brightness temperature, 409 CO, carbon monoxide
desorption form polycrystalline W, 259
C purification, 503
C60 evaporation, nanoscale templating, 460 tungsten single crystal, 551
Calibrated effusive molecular beam Coefficient of friction, 341–344
measurements, 253–255 Conductiver coatings, 123–124
Calorimetric heats of adsorption, 333–336, Conflat flange
574–575 aluminum, 101
Calorimter, Setaram C-80, 575f for Al foil window, 528f
Capacitance sexless, 99
crystal positioning using, 11 Cooling reservoir, 385, 390
measuring circuit, 12f flexible, 394, 395f
Capacitance manometer, 425, 432, 461, 478, for horizontal manipulator, 390, 391f
486, 574 Cooling samples
Capillaries, cleaning, 89–91 by mixing gas streams, 401
Carbon monoxide. See CO contact with thermal reservoir, 383
Cataphoretic desorption crystal holder for, 381–383
of LaB6, 136 gas thermal switch, 401–402
of ThO2, 133 on manipulators, thermal conductivity
Cathodes issues, 383, 385–386
ThO2/LaB6 coated, 133, 135, 187 using direct contact, 389
CeB6 coated, 139 variable rate, 401–403
ThO2coated Ir, 219 Cryogenic connections, 392
CEMA. See Chevron electron multichannel Crystal cleaving, 93–94
plate array Crystal heating
Cement electron bombardment, 361–363
AgCl, 525, 555 interrupted ohmic heating, 359–361
MgO (Armeco 517P), 415 Crystal manipulator, 3–5
semiconductor crystal mounting, 368, 369f Crystal mounting
vacseal, 367, 525 non-weldable, 354, 355f
ZrO2 (aremco Ultratem 516), 368 strain free for azimuthal motion, 355–356
Ceramic insulator deterioration, 64 Crystal positioning, 11–12
Cesium ion gun design, 471–473 Crystal temperature measurement, 362
Channeltron detector, 262–264 Cu, copper
Charged particle oscillator, He(I) and He(II) braid, 54, 305, 392
ultraviolet source, 236 C60 on Cu(110), 460
Chemical actinometer, 245–246 for cooling tank, 389
Chemisorption gasket removal device, 102–103
CO and Ni, 574, 575f improved heat conductivity, 392
CO and Pt, 426, 427f oxidation of, 381
cyclohexene on Pt, 335 reservoir, 386, 387f, 389
IR transmission spectra, 556f Cu, copper, 381. See also OFHC Cu
N2on Rh, 557 Cu-Be cup, 388
O2 on Pd, 238–242
Chevron electron multichannel plate array D
(CEMA) detector, 232 Dark current, 262
Cleaning crystals Debye temperature, 337–339
metal, 543 Desorption kinetics
Index 645
angular resolved, 273–275 Electron gun

from single crystal faces, 259 beam properties, 188
Detectors cathode, 141
InSb infrared, 328 design and behavior, 187–189
ion-pumped infrared, 329f electrostatically focused, 450, 451f
MCP-type, 227, 229 emission current regulation, 189
MgCdTe (MCT) infrared, 321, 328, 334 f low-energy design, 192f
PVDF thermal, 335 low energy/broad beam, 189–192
resistive anode, 225, 227 pulsed, 189, 227, 228f, 231, 231f
silicon surface barrier, 305, 306 work function measurement, 291–292
stagnation, 334, 429 Electron multipliers
Deuterium kinetic isotope effect, 282 channeltron-type, 262–264
Diamond, heating behavior, 364, 365f Cu/Be-type rejuvenation, 206
Diamondlike carbon coating, 101 Electron reflection coefficient measurement,
Differential pumping method, 6–11 293–294
Dimethyldiazine Electron scattering, single and multiple
carcinogenic agent, 495 processes, 205
methyl radical source, 485 Electron spectrometer calibration, 196–197
Diode ion pump performance in He pumping, Electrons, specularly reflected, 202
75–76 Electroplating, 120, 367, 448
Dynode multiplier, regeneration of, 263 Emissivity, 410
Dynode rejuvenation, 205–206 Epoxy adhesive, 179, 343
ESDIAD
E benzoate on Cu(110), 227, 228f
EDA. See Ethylene diamine pattern for CO on Cu(110), 225, 226f
Einzel lens, 187, 220, 472 planar grid apparatus, 227
Electrical conductivity, TiO2 single crystal, pulse counting analyzer, 225–227
200–201 see. LEED
Electrical connections time-of-flight, 226–228
band clamp, 67 Etching
cold-formed wire connector, 64 large thin window fabrication procedure,
leakage, 64 529, 531–532
phosphor bronze spring, 66 STM tips, 168–172
sliding metal contact, 66 Ethylene adsorption/desorption study,
tungsten-to-tungsten wrap, 61 zeolitesurface, 569–571
ultrahigh resistance vacuum feedthroughs, Ethylene dimaine (EDA) etching agent, 532
64–65 Eutectics, Zr and W, 444–445
Electrical isolation of UHV components, 151 Evanston-type microwave discharge cavity,
Electrical shock 480
avoidance, 591 Evaporation source
effect on man, 589 aluminum, 446–448
from ionization gauge, 590 cadmium sulfide, 456
Electrochemical device, UHV, 500, 501f congruent, 442, 444–446
Electrodeposition, Cr on W wire, 448–449 gold, 310–313
Electron beam lithium, 444–446
angular divergence measurement, 198–199 see. Metal evaporation
damage from stray electrons, 260 EXAFS measurement, 559
evaporator, 450f
point focused, 453 F
position monitor, 224 Faraday cup, 162
small evaporator design, 451–453 absolute detection efficiency using, 226
Electron bombardment heating, experimental modified for electron current measurement,
difficulties, 277 202–203
Electron energy analyzers, 191, 194–196 movable, 218–219
646 Index
multiple, 188, 198–199 OFHC Cu type, 104

Fe(CO)5 decomposition, 89 removal device, 102–103
Feedthrough silver-plated copper, 100
high-speed rotary, 25 titanium nitride-coated copper, 100
ultrahigh resistance sapphire-metal, 64 see. Seals
Fermi level determination, 197 Gas thermal switch, 401–403
Filament power supply tester, 141 Gd2(MoO)4. See GMO
Filament temperature measurements, 407–409 Ge, germanium, optical pyrometer calibration,
W filament nomograph, 408 413
Films Geiger Müller counter
BeO on dynodes, 205 for 14C on Pt(111), 305
electrically conductive SnO2, 123 for inverse photoemission intensity, 207
for O2 and O3 detection, 490 Getter films
Si3N4, 413 He storage, 505
thickness measurements, 315 maintaining vacuum and, 511
Flanges Ti, 273, 505
Conflat, 6 TSP pump, 72
KF, 523 GMO [Gd2(MoO4)3], in measuring ion-beam
Formic acid, isotopically labeled, 271 size, 223f
Fracture stage, 343, 345f Graphite, electron spectrometer calibration,
Friction measurement, 341–343 196
Grazing incidence X-ray photoemission
G spectroscopy (GIXPS), 247
GaAs, gallium arsenide Grids
indirect heating of, 365–366 conical grid fabrication, 121–122
internal reflection IR spectroscopy and, hemispherical, 119–120, 227
326–327 in electron gun, 189
MBE growth of, 536 IR cell sample support, 557–559
surface oxidation, 247, 249f large flat mesh, 120–121
Ga, gallium, metallic joint, 389, 393 tungsten, 244f, 557, 558f
Galling
prevention, 47 H
protection against, 29 Halogens
Gas chromatography sensitivity enhancement, electrochemical source, 497–498
287–288 microwave discharge source, 480, 501
Gas doser, 422 Heaters
Gas dosers, 89–90, 422–426, 428–430, 432 button type, 363–365
designs, 90f halogen lamp, 378
repetitive dosing, 430 paint-on type, 374, 374f
retractable, 434–435 pyrolytic graphite, 372–373
Gas flow regulation, 433 tungsten filaments, 375, 375f
Gas purification Heating
apparatus, 510f compound semiconductors , 365
at high pressure, 505–507 conduction, 364f
cryogenic or gettering method, 509–511 design for insulator crystals, 366–368
extremely pure He, 507–509 indirect, 363–365
metal permeation method, 511–512 irregular shapes paint-on type , 374
nitrogen, 509 Ohmic, 50, 55, 364f
Gaskets, 99, 101–104 radiation and electron bombardment, 277,
annealed copper, 99 362f, 364f, 377f
copper, 50, 56 silicon crystals, 356–359
cryogenic, 101–102 temperature programmed, 277, 349–353,
deformable, 513, 525 415–416
gold wire, 100, 101f He(I)/He(II)
Index 647
photoemission source, 235–236 opaque materials, 564–566

UV resonance line source, 236–238 sealing windows to glass, 523
He helium semiconductors and, 324
adsorbed He, 399–400 window mounting, 521–523
below 4K, 399–400 Insulators, electrical
cryocooler, 93, 399 Al2O3, 64
extremely pure, 507–509 kapton, 11, 151, 385
leak detection probe, 107 macor, 207
leaktight seal , 525 sapphire, 64, 381, 385
production and purity of, 507–509 Internal reflection IR spectroscopy, 328
pumping test , 75 gallium arsenide, 327
purification apparatus, 510f p and s polarization vector diagram, 326f
transpiration used for lowering l-N2 silicon and, 324–326
temperature, 394–396 UHV optical system, 326f
High area solids Inverse photoemission, 206–207
adsorption and desorption on, 569–572 Iodine (I2) source, 499–501
core-shell structures, 562–563 Ion beam
diffusion IR study, 560–566 angular distribution, 225, 226, 230
IR cell for, 557–559 polishing, 160
microcalorimetric studies , 574 STM tip sharpening, 173
opaque materials, 564–566 time-of-flight separation, 227, 228f
High-pressure reactor with kinetic energy and mass resolution,
gold-plated pan bronze , 41 229
sampling with gas chromatography , 287 Ion beam sources
High-pressure sample transfer cell, 50–51 axial magnetic field and, 218f
Hot filament resistance thermometer , 265, 407 Colutron GD-2, 220
HREELS high-resolution electron energy loss low energy, magnetic focused, 220–221
pectroscopy, 196, 205, 361 measuring beam dimensions, 223
for boron detection on Rh, 545 Ion bombardment
Hydrogen evaporation, 317–318
as impurity, 507 neutrals desorbed by, 221
atomic, 477–480 sources, 213–214
desorption form Ni(111), 273, 274f Ionization gauge
dissolved in stainless steel , 87 corrections for external radiation and
permeation studies, 513–514 external electron flux effects, 142
purification, 505–507 electrical shock from, 591
Hydrogen peroxide doser, 493–494 filament replacement, 145f, 145–146
geometry in evaporation rate measurement,
I 314
Infrared cells metal film evaporation rate measurement,
adsorption studies, 555 314–315
high surface area study, 558 movable, 411
wide temperature range, 558, 559f Ion gun
Infrared reflection-absorption spectroscopy based on Bayard-Alpert gauge, 214, 215
(IRAS), 321–325 cesium, 471
Fourier transform spectrometers (FTIR) cold cathode, 216–218
and, 321, 323, 327 design of, 472, 473
measurement at high pressure, 323–325 low energy, 220, 222f
Infrared spectroscopy low gas emission, 211–212
continuously pumped detector, 328–329 modified, 212, 220, 221
differential pumping widow seals and, 522f pierce, 472–473
diffusion into core-shell structures, 562–564 pulsed, 232
diffusion through powders, 560–561 Ion pumps
rapid purge using rubber sleeves, 329–330 cleaning by chemical etching, 78
648 Index
methane liberation form, 72 of translating Auger spectrometer, 16

Ion sources platform (LMP), 32
alkali, 473 solenoid driven, 15
alkaline earth, 474 Liquid adding to atomically clean
broad beam, 218–219 surfaces, 517
IRAS. See Infrared Liquid nitrogen
reflection-absorptionspectroscopy enhanced cooling of, 394
IrO2cathode insulating layer, 218 level alarm, 396, 398f
Isosteric heat measurements, 325f, 333 non-leaking, 398, 399f
trap, 79
J triple point, 396
JASON program linear Poisson equation, Low-energy electron diffraction. See LEED
magnetic shielding, 149 Low energy electron diffraction (LEED)
Joule-Thompson refrigerator, 115 angle of incidence measurement in,
162–163
K beam intensity measurement, 201
Kapton insulator, 11, 151 Kr isotherms using, 94f
Kapton/sapphire, comparative cooling rates, movable Faraday cup in, 337, 338f
385–386 pattern, for Cu(110), 225, 226f
Kaufman ion sources, 218 retarding diode method and, 289
Kelvin probe XLEED, 302
Au electrode, 299 Lubricants
effect of stray capacitance in, 297 aquadag, 123. See also Aquadag
electromagnetic driven, 298f boron nitride, 26
high-temperature flow reactor, 299, 300f degradation, 29
piezodriven, 296f, 297 fluorocarbon, 179
reproducibility, 299 for heavy sliding loads, 27–28
thermal desorption spectroscopy and, 297 graphite, 8, 100
work function in semiconductors, 295 molybdenum sulfide, 9, 31
Kinetic isotope effect, 282 Santovac 5 pump oil, 8
for deuterium, 280 TiC/Ag, 28
Kinetics of adsorption, 253, 255 Lubrication
Knudsen cell, 464 coefficient of friction and, 341
of metal/metal interface, 27
L
LaB6 thermionic emitter material, 137 M
LabVIEW program, 277, 351, 415 Macor insulator, 140, 206
controller logic, 353f Magnetic eddy current drive, 22
Laser-induced fluorescence (LIF), 231 Manipulator
Laser-ionized neutrals time of flight detection, cooling samples on, 383–386
231–232 horizontal cooling of, 389
Leaks multiple motion, 3–5, 18
calibrated, 571 rotary high precision, 13
check groove, 102, 103f vacuum-jacketed cryogenic, 386
detection probe, 107–108 Manufacturers, equipment and components,
liquid sealers, 108 597
temporary sealing of bellows, 108–109 Mass spectrometers
LEED/ESDIAD analyzer, 225–227 calibration, 270, 271f
LIF. See Laser-induced fluorescence differentially pumped, 72
Linear enhanced sensitivity, 267–268
of translating Auger spectrometer, 17 magic angle thermal desorption, 275–276
Linear motion sensitivity comparison with bare ion gauge,
actuator for, 15f 260
inexpensive device for, 15–16 sensitivity ratios for, 270
Index 649
Ta filament mass scale calibration, 266, Motional freedom

267f cooling braid and, 392, 393f
thermal desorption, 571f cryogenic manipulator, 386
unshielded, 275 shape memory effect and, 20–22
UTI-100C, 270 six degrees of, 3
see. Quadrupole mass spectrometers see. Rotary motion
Mass spectrometry, 259–271 Motion in vacuum, 3–36
MCP. See Microchannel plate detector Motors
Mechanical action on samples high speed, 25–26
crystal cleaver, 93–95 stepper motors, automated TPD, 278
fatigue of Al(111) crystal, 95 very high speed, 26f
mica cleavage device, 95 Mu metal, 297
piezoelectric fatigue apparatus, 95–96 Multicapillary array collimator, 211
tungsten carbide blade, 93
Metal evaporation N
alloys and mixtures, 440 N2, nitrogen, polarizability form work function
chromium evaporator, 448 measurement, 291
degassing, 437 Naphthalene dosing, 463, 464f
downward evaporation, 439 Neutral beam intensity measurement, 429
eutectic mixtures, 444 Ni permeation source, hydrogen purification,
flash evaporation, 442 511
low coverages of single atoms, 459 NO, nitric oxide
measurement of rate, 314 impurity oxidation, 546
multiple evaporator, 441f oxidative wall passivation, 87–88
portable microevaporator, 459
ultralow coverage, 440 O
Metal fatigue, 95 Observation windows
Metal foam, vaccum applications, 75–77 film deposition prevention, 536–537
Metal foil low optical distortion glass, 538
aluminum, 528 shielded, 535–536
beryllium, 526 strain-free mounting, 533
HAVAR, 179 zero length assembly, 539
stretching, 179–180 OFHC Cu, oxygen free high conductivity
tantalum/platinum sandwich weld, 175, copper, 104
176f cooling braid, 392, 497
wrinkle free, 179 cooling reservoir, 394
Metal single crystal cleaning, 543–551 crystal holder, 355
Methanol heated doser, 460–464
reducing agent in spotwelding, 175, 176f hollow bolt seal, 104f
reaction on Ni(111), 280–282 sample cooling, 383
Methyl chloride thermal desorption spectra, 72 thermal shield, 34
Methyl radical source, 485 Optical pyrometer
Microcalorimetric studies, 574 calibration point for silicon, 413, 415
Microcapillary array beam doser, 421–429 disappearing filament type, 409–411, 412f
calibration and use of, 425–426 measurement through coated window, 411
design, 421, 422f radiant energy equation, 409
fractional interception, 426 Optical windows. See Observation windows
gas handling for, 423, 424f Organic doser
glass, 89 BN crucible, 466
kinetic studies, 500, 501f ceramic standoff, 464
precautions, 423 volatile organics, 460–464
retractable, 435 Outgassing
shutoff with condensable gas, 428 degassing loops, 313, 407–409
Micorchannel plate detector (MCP), 225–227 from motors, 24
650 Index
Outgassing (cont.) Kelvin probe actuator, 299

prevention, 407 Tube, 34
studies, 110 Planck’s law, 409
Oxygen Polarized UV light for photochemistry,
atomic oxygen production, 491–493 240–242
chemisorbed, 238, 240, 241f Polarizer-compensator-sample-analyzer
detection of atomic oxygen, 490 (PCSA), 302, 303f
MnO2solid state source, 513 Polycrystalline metals
purification, 505 Ag, 95
Ozone Ni, 295, 297f
detection of, 490 W, 259
generator, 488–489, 489f Pulse counting, 225–229
mass spectrometer cracking pattern, 487, Pump oil
488 Fomblin, 488, 584
molecular beam doser, 488–489 Santovac-5, 584
production, storage, use, 487–488 Pumping
differential, 6, 7, 9, 41, 235, 237f, 435, 471,
P 521
Palladium getter film, 72, 273, 483, 505
carbon removal from, 546–547 parallel, 72
surface segregation and, 545 pressure equalizing device, 72–74
Palladium leak, hydrogen purification with, tests, 76f
481 two-stage differential, 237f
Palladium permeation source, hydrogen vibration-free ballast-type, 71
purification, 511 Pumps
Pancake radiation heater, 47 aspirator/sorption, 69–70
Passivation, surface diode ion, 72, 75–76
oxidative, 424 ion, 8, 56, 69, 71, 72, 75, 211, 236, 327
teflon coating, 91 oil contamination elimination, 69
Permeation oil diffusion, 237
atmospheric, 109 regurgitation effect, 72
gas purification by, 511–512 safety system, 584–586
reusable samples, 513–514 see. Titanium sublimation pump
Phosphor screen triode ion, 75
beam dimension measurement, 198 turbomolecular, 41, 237
CEMA detector and, 231f, 232 zeolite sorption, 69, 115
dusting method for coating, 129–130 Pyrolytic graphite
electrostatic deposition of, 129, 130 chemical vapor deposited, 372
LEED apparatus, 123, 226 paint-on heater, 374
sedimentation deposition of, 127–129 Pyrometer. See Optical pyrometer
spot photometer and, 202
Photochemical surface processes, 238–243 Q
Photochemistry QTM. See Quartz thickness monitor
of chemisorbed O2, 238, 239f, 241f Quadrupole mass spectrometers (QMS)
photoreduction of Ag, 242 channeltron detector, 262–264
polarized UV light in, 240–242 enhanced sensitivity, 267
Photoemission studies, hollow cathode enhanced signal-to-noise ratio in, 262–264
resonance lamp for, 235–236 floating grid, 265, 266f
Photon stimulated desorption, 232, 399 line-of-sight for thermal desorption,
Pierce ion gun, 472 269–270
Piezoceramic foil, in Kelvin probe, 295–297 see. Mass spectrometers
Piezoelctric actuators modulated ion source, 267–268
drive, 34, 95 ozone and atomic oxygen detection, 490
fatigue on polycrystalline Ag, 95, 96f shielded, 259–262, 425
Index 651
spurious electron emission elimination, longitudinal with disconnect, 53f

265–266 magnetic rod, 52
Quartz crystals mirror port, for enhanced sample transfer,
oscillator, 321 56, 57f
thin film measurement on, 314–315 movable crystal, 41f
see. Quartz thickness monitor rod-through-chamber, 42, 43f, 44 f
Quartz flat partial reflector, 238 rotation to multiple ports, 45
Quartz microbalance. See Quartz thickness transfer rod, 47
monitor translatable reactor cup, 40f
Quartz prism UV polarizer, 238, 240 vibration-free, 56
Quartz thickness monitor (QTM), 309–312, with trolley system, 33
314–315 Sample transfer with disconnect
calibration of, 309 inexpensive, 54, 55f
magnetic drive, 24 in standard UHV system, 52
very low coverage measurement, 440 in STM systems, 54
Sapphire
R crystal holder, 354
Radiation damage reduction, 243 electrical leakage prevention, 64–65
Radioactive adsorbates, 305–307 rod, 299, 300f
Rapid wafer thermal processors, 413 support bearing, 355
Re ribbon, ThO2 deposition on, 136 thermal conductivity of, 386
Resonance enhanced multiphoton ionization UV grade window, 238
(REMPI), 231 Scanning kinetic spectroscopy (SKS), 280–282
Retarding diode method, 289–291 Scanning tunneling microscopy (STM), 54,
RHEED intensity measurement, 313 155, 168, 299, 459, 548
Richardson equation, 133 repeatable sample positioning, 34–35
Rotary motion, 3–14 Scanning tunneling microscopy. See STM
arcsec resolution, 13–14 Scanning tunneling spectroscopy (STS),
axial, 6 299–300
bearings, 29 Sealants, use of, 24, 108, 523, 525, 533
cryogenic azimuthal, 388–389 Seals
electrical readout, 10 compression, 103–105
fine angular adjustment of, 14f doubly differentially pumped, 45, 322
heavy-duty turntable, 20 for infrared windows, 521
high-speed feedthrough, 25–26 gold wire cryogenic-type, 99–102
magnetic shutter, 23–24 helicoflex O-ring, 113
optimization of angular resolution, 20 high force, 43
rapid, 10 high temperature, 6, 524
through small angles, 18–19 indium, 526
turntable, 20 kovar glass, 394, 533
two-axis device, 19f metal to Teflon, 43
ultrahigh vacuum, 6–11 rotary design, 7f, 8f
sliding rod, 42f
S spring-loaded Teflon, 9, 52, 56
SAES alkali metal sources, 483 UHV glass window, strain-free, 533
Safety, 589–592 vacseal leak sealer, windows sealer, 367,
Sample cooling, 392–394 525
Sample transfer, 39–49, 51–57 Wilson type, 6
at low temperature, 45, 47, 48 see. Gaskets
from high pressure to UHV, 39–41, 44 Semiconductor crystals
from UHV to external cell, 45, 46 cement mounting for, 368–369
high-pressure cell, 49–52 cleaning, 548–549
internal cup high-pressure cell, 49–50 interferometry and temperature
Lebold-Heraeus device, 41 measurement, 359
652 Index
Semiconductor crystals (cont.) reconstruction effects, 548

low stress mount for heating/cooling, see. Metal single crystals
369–371 use as calorimeter, 240, 241f
SiC, 327 UV irradiation of, 238
Shape memory effect (SME) SKS. See Scanning kinetic spectroscopy
Clamp, 22f SME. See Shape memory effect
NiTi alloy, 21 Single crystal surface
shutter, 22f Au(111), fabrication, 155, 156f
small motions with, 21–22 chemisorption and reaction on, 305
Shelton diode for work function measurement, Kelvin probe desorption measurement, 295
293–294 kinetics on, 284
Shielding, 149–150 synthesis reactions on, 284
ad-mu 80, 150 tungsten, atomically clean, 550–551
electrical, 151–152 Sources
magnetic in UHV, 149–150 alkali metal, 483
Shutter device, rotatable, 22–23 alkaline earth ion, 474
Silica gel atomic arsenic, 456
Davisil, 131n atomic fluorine, 481–482
ozone storage, 486 BN diffusion, 529
Silicon cesium, 471–473
crystal heating, 361–363 electrochemical cell, 497, 498f
crystal mounting, 358f He(I), He(II), 238f
emissivity of, 409–410 high-flux atomic hydrogen, 477–480
etching, 530 hydrogen peroxide, 493–494
split-off-dimer defects, Ni induced, 549f I2, 499–501
surface cleaning, 548–549 ion bombardment, 213–214
temperature control using resistivity, 415 Kaufman ion source, 218
Silver films, for ozone and atomic oxygen Lyman-α, 245–247
detection, 490 methyl free radicals, 485
Silver permeation source, oxygen purification, pulsed vacuum UV, 247–248
511 UV from high-pressure Hg arc, 238
SIMION electrostatic lens and analysis see. Atomic hydrogen; Thermionicemitters;
program, 189 Vacuum ultraviolet sources
ion optical calculations, 473 Spot photometer construction and use,
Single crystals 201–202
alignment, 156–159 Spot-welding
anisotropy of electrical conductivity, difficult junctions, 175–177
200–201 see. Welding
Bond-Barrel holder, 156, 158f slotted crystals, 177
CaF2disk, 557 Springs
calorimetry, 333, 334 chromel wire, 369
curvature of, in grinding and polishing, 157 Cu-Be, 68, 388
cutting goniometer, 159 phosphor bronze, 66
cylindrical, 375–377 Ta wire, 67
desorption kinetics on, 259 Sputtering
electron beam angle of incidence on, 163, axial, to sharpen tips, 173
164f cleaning by, 213
fabrication, 155–160 coating tube inner walls, 314
ion beam polishing, 160 contamination effects, 213
orientation, grinding and polishing, crystal polishing, 160
156–159 deposition rates, 216
photochemical processes on, 238 for hybrid thin film deposition, 314–315
radiation heating of through glass windows, ion gun for, 211–221
371–372 neutral species, 275
Index 653
TiN, 99 Temperature-programmed reaction

Stagnation detector, 429 spectroscopy (TPRS), 280, 282–284
Sticking coefficient, 270, 377 Temperature programming
measurement, 330–331, 366 metal crystals, 349–351
Sticking probability versus absolute coverage, metal and semiconductor crystals, 351–353
426f Tesla coil, 91, 206, 488
equation, 255 Thermal braid cooling, 389, 392–394
STM tip, 171 Thermal desorption, 273–280
approach mechanism for, 31 angular distribution, 273–275
density of states, 299 apparatus for, 274f
etching for field ion microscopy, 169f apparatus for automated TPD, 278–280
fabrication, 167–173 azimuthal angular resolution on a
outgassing, tip flasher, 170 cylindrical single crystal, 376–377
reproducible etching, 168–169 CO on Ru(001), 298, 299f
sharpening, 170, 173 H2on Ni(111), 273, 274f
single crystal metal, 173 line-of-sight, 269–270
Strain gauges, 337, 341 Thermal desorption spectroscopy (TDS), 259,
Sulfur source, 497 267–268, 275–276. See also TPD
Supersonic Thermionic emitters
molecular beams, 429 brightness temperature, 409, 411
nozzles, 180–181 carburization, 136
Supported catalysts, 555 cathode design, 142f
Surface barrier detector, 302–303 electron gun, 187
Surface contaminants, 265 emitter evaporation rates, 138f, 139
Surface Debye temperature, 334–336 filament replacement in, 145–146
Surface poisoning, 280 high-temperature operation, 141–142
Surface reaction kinetics at high LaB6 coated, 136, 137f
pressure, 44 lanthanum hexaboride, 135–136
Surface reaction processes, 280 low work function coatings, 137
Surface segregation, 545 mounting, 139–140
Surface studies shielded, 215
electrochemistry apparatus, 115f shielded and open emitters, 140f
ellipsometry for studying adsorption, 300 spiral, 139
enhanced surface sensitivity, 247 thoriated, 133–135
radiotracer techniques for, 302–303 Thermocouple gauge, sealing, 104, 105f
Syringe, adding liquid, 517, 518f Thermocouples
attachment, 357, 358
T avoiding excess local heating, 275
Tantalum (Ta) plasma, 247 calibration, 349, 405–407
Tantalum tube, hydrogen embrittlement cemented, 365
prevention, 215 chromel-alumel, 357, 448
TDS. See Thermal desorption spectroscopy contamination, 357
Teflon welding to metal single crystal, 175
compression fitting, 115 Thin film deposition, 309–319
impurities from, 50 absolute surface coverages, 305–307
liners, 42 hybrid methods, 317–318
Temperature measurement measurement, thickness limitations, 313
hot filaments, 407–409 on tube inner walls, 314–315
pyrometric, 409, 410 organic vapor deposition cell, 464, 466f
thermocouples, 349–351 stable/metastable films, 314
W filament nomograph, 408 Tilt motion, 3, 5f
Temperature-programmed desorption (TPD), Titanium sublimation pump (TSP), 72, 84, 115,
259–262, 266, 278–280, 283, 284 211, 247, 259, 505, 584
654 Index
Topping model, 291, 292f V

Temperature-programmed desorption (TPD), Vacuum ultraviolet sources (VUV), 235–240,
259, 275–276, 508, 569–571 247–248
TPRS. See Temperature-programmed reaction Valves
Translator, folding linear magnetic, 586 fast closing, 583–584
Traps low deadspace aperture, 430, 431f
liquid nitrogen, 79–80 polyfluoroalkoxy (PFA), 89
metal foam, 75–77 small gate, 45
water-cooled Cu, 235 squeeze, 433–434
zeolite, 235 straight through, 47, 235
Tribology, review, 29 Variable aperture, flange mounted, 30–31
Tribometer, 341–343 Variable aperture, organics, 464
TSP. See Titanium sublimation pump Viton
Tungsten atmospheric permeation, 108, 110
coils, 62 bakeout limitation, 524
crystal holder, 354 outgassing, 40, 323
grid, 244f, 557, 558f Vycor permeation source, helium purification,
hairpin filament, 309 507
high-temperature operation, 375
single crystal, atomically clean, 550–551 W
wire, fabrication, 62 Wagner electrochemical cell, 497
wire heaters, nomograph for design, 409f Wall passivation, 87–91
Tungsten ribbon attachment to power leads, Wall reactions
61–62 minimization, 89–91
Turbopump power interrupt protection, reduction with gold film, 39
581–582 Wehnelt cylinder, 187
Welding
U for thermal contact, 392
Ultra high vacuum systems (UHV) metal foil window, 529
access port, 56–57 sandwich weld, 175, 176f
adding liquid to surfaces, 517–518 see. Spot-welding
aluminum, 113–114 Wien filter, 220
bakeout, 83 Wilson seal, 6
chamber protection, 583 Wilson differential pumping method, 6
common impurities in, 548 Windows
differential pumping in, 8 alkali halide, 521
electrochemical device, 502–503 bakeable aluminum, 528
grating rotation, 19–20 bakeout limitations, 525
fracture stage for surface analysis, 343–344 beryllium, 526, 527f
stainless steel wall passivation, 87–88 CaF2, 207, 323, 523, 524, 557
surface electrochemistry apparatus, disposable, 535
114–116 for vacuum UV transmission, 525–526
trolley transport system, 33 glass seals for, 537–538
Ultrathin metal single crystals, 333, 334f heated, 536–537
Ultraviolet light (UV) IR transparent, 321, 557
lamp for photochemistry on surfaces, 238 Itran, 524
Lyman-α, 245–247 large-area Si crystal preparation, 529
polarized, 240 metal thin foil welding, 529
Ultraviolet photoelectron spectroscopy (UPS), mounting, 521
196 SrF2 , 207
gold (Au) spectrum, 236 thin film coated, 411
Index 655
UV grade sapphire, 238 measurement, 297

UV transparent, 207 minimization of differences along a surface,
X-ray transmission, 526–529 189
zero length assembly, 539 SnO2, 123
see. Observation windows tungsten, 133
Wire
ceramic-bead insulated, 66 X
gold wire, 155 Xe/Ag(111), 302
polycrystalline, 174 X-ray photoemission spectroscopy, grazing
Pt heater wire, 466, 467f incidence, (GIXPS), 247
single crystal, 163 X-ray standing wave spectroscopy (XSW), 13
straightening and braiding, 182–183
tantalum, 50, 266 Z
tungsten, 50, 62, 408f Zeolite
Work function ethylene adsorption study on, 569
changes, retarding diode method, 287–288 for gas storage, 505
difference for two surfaces, 291 trap, 77f, 509
electron gun measurement, 289–291 Zone electropolishing, 167–168
electron reflection coefficient
measurements, 291–292

Experimental Innovations in Surface Science: John T. Yates JR

Uploaded by

Copyright:

Available Formats

Experimental Innovations in Surface Science: John T. Yates JR

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Experimental Innovations in Surface Science: John T. Yates JR

Uploaded by

Copyright:

Available Formats

John T. Yates Jr.

ISBN 978-3-319-17667-3 ISBN 978-3-319-17668-0 (eBook)

Library of Congress Control Number: 2015941351

Springer Cham Heidelberg New York Dordrecht London

Printed on acid-free paper

Springer International Publishing AG Switzerland is part of Springer Science+Business Media

A number of new experimental innovations have been added in over 50 new

Virginia John T. Yates Jr.

tunneling microscope (STM), complex molecular beam techniques, and complex

March 1997 John T. Yates, Jr.

Part I Vacuum System Technology

1 Motion in Vacuum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 3

2 Sample Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 39

4 Pumping and Trapping. . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 69

6 Behavior of UHV Systems . . . . . . . . . . . . . . . . . . . . . . ........ 87

7 Mechanical Action on Samples . . . . . . . . . . . . . . . . . . . . . . . . . . 93

9 Leak Repairs and Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107

10 Specialized UHV Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113

Part II Mechanical Fabrication Techniques

12 Conductive Coatings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123

13 Phosphor Screens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

14 Thermionic Emitters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 133

15 Shielding . . . . . ....................... . . . . . . . . . . . . . . 149

16 Single Crystal Fabrication/Orientation . . . . . . . . . . . . . . . . . . . .. 155

17 Tip Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167

17.4 Reverse Bias Method for Sharp Tip Etching . . . . . . . . . . . . . 170

19 Mechanical Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179

Part III Measurement Methods

20 Electrons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 187

21 Ions . ....................................... . . . . . . 211

21.6 Focusing of Low-Energy Ions by Magnetic Field . . . . . . . . . 220

22 Photons (UV) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 235

23 Kinetics of Adsorption/Adsorbate Coverages . . . . . ........... 253

24 Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 259

24.4 Ta Mass Spectrometer Filament . . . . . . . . . . . . . . . . . . ... 265

25 Temperature-Programmed Desorption/Reaction . . . . . . . . . . . . . . 273

26 Gas Chromatography—Enhanced Sensitivity . . . . . . . . . . . . . . . . 287

27 Work Function, Tunneling Spectroscopy and Ellipsometry . . . ... 289

28 Radioactive Adsorbates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305

29 Thin Film Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309

29.4 Reusable Quartz Crystals for Film

30 Infrared Spectroscopy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321

31 Calorimetric Heats of Adsorption—Single Crystals. . . . . . ...... 333

32 Surface Debye Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337

33 Friction and Fracture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 341

Part IV Thermal Control

34 Heating Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349

34.3 Mounting Nonweldable Crystals for Resistive Heating

35 Cooling Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..... 381

36 Temperature Measurements . . . . . . . . . . . . . . . . . . . . . . . . . ... 405

Part V Delivery of Adsorbates to Surfaces

37 Gases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 421

38 Evaporation Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437

38.15 Portable Microevaporator for Depositing Low Coverages

40 Active Gases and Species . . . . . . . . . . . . . . . . . . . . . . . . . ..... 477

41 Electrochemical Sources and Devices . . . . . . . . . . . . . . . . . . . . . . 497

42 Gas Purification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 505

42.4 Gas Purification by Permeation Through Metals .......... 511

43 Liquid Handling in UHV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 517