Archaeological Conservation Using Polymers

Texas A&M University Anthropology Series
D. Gentry Steele, General Editor

Series Advisory Board

William Irons
Conrad Kottak
James F. OConnell
Harry J. Shafer
Erik Trinkaus
Michael R. Waters
Patty Jo Watson
ii
 iii

Archaeological
Conservation
Using Polymers
Practical Applications
for Organic Artifact
Stabilization

Foreword by J. M. Klosowski

Copyright
by C. Wayne Smith
Manufactured in the
United States of America
All rights reserved
First edition

The paper used in this book

meets the minimum requirements
of the American National Standard
for Permanence of Paper for
Printed Library Materials, .-.
Binding materials have been
chosen for durability.

Library of Congress Cataloging-in-Publication Data

Smith, C. Wayne.
Archaeological conservation using polymers : prac-
tical applicatons for organic artifact stabilization /
C. Wayne Smith.st ed.
p. cm.(Texas A&M University anthropology
series ; no. )
Includes index.
--- (alk. paper)
--- (pbk. : alk. paper)
. AntiquitiesCollecton and preservation.
. ArchaeologyMethodology. . Archaeological
chemistry. . Polymers. I. Title. II. Series.
.
.'dc

The author and publisher gratefully acknowledge

a generous grant from Bill Caruth in support of
publication of this volume.
 v

To Helen,

Thanks for sharing this

and all my adventures.
vi
 Contents

Foreword by J. M . Klosowski, ix
Acknowledgments, xm
Introduction, 3

CHAPTER
1 Laboratory Setup, 7
Major Instrumentation, 7
Small Necessities in the Laboratory, 8
Chemicals, 9
2 Baseline Mechanisms, 13
Dowel Experiment, 13
Mass Spectrographic Analysis of Out-Gases Created from the
Dehydration of Archaeological Wood Samples, 17
3 Archaeological Wood, 21
The Challenge of Conserving Waterlogged Wood, 21
Degradation and Shrinkage, 22
Waterlogged Wood from Saltwater Environments, 22
Case Study: Waterlogged Wooden Buttons with and without
Associated Thread, 26
Dry-Site ArtifactsDry and Desiccated Wood, 28
Reprocessing and Stabilization of PEG-Treated Wood, 30
Tongue Depressor Experiment, 31
Case Study: Re-treatment of Two PEG-Treated Sabots, 40
Re-treatment of PEG-Treated Waterlogged Wood, 43
Case Study: Treatment of Waterlogged Wood Using Hydrolyzable,
Multifunctional Alkoxysilane Polymers, 45
4 Leather Preservation, 60
Archaeological Leather, 60
Cleaning, 62
Chemical Cleaning, 62
Treatment of Leather, 62
PEG/Air-Drying Treatments, 63
Freeze-Drying PEG-Treated Artifacts, 64
PEG and Other Polymers, 65
Passivation Polymer Processes, 66
Case Study: A Successful Treatment Strategy for a Waterlogged Shoe, 66
Passivation Polymer Treatment for Dessicated Leather, 69
An Effective Treatment for Dry Leather, 70
Suggestions for Treating Leather between Sheets of Glass, 72
Storage and Display of Leather Artifacts, 72
 5 Composite Artifacts, 74
Case Study: Preservation of a Composite Artifact Containing Basketry
viii and Iron Shot, 74
6 Cordage and Textiles, 81
CONTENTS New Techniques for the Preservation of Waterlogged Rope, 81
Silicone Treatment Strategies, 82
Frankfurter Method of Rope Preservation, 82
Treating Waterlogged Rope in a Nonpolar Suspension Medium, 82
Incorporating the Use of Nonpolar Suspension Mediums and Elements
of the Frankfurter Method into Traditional Silicone Treatment Strategies, 83
Case Study: La Belle Rope, 83
Case Study: Preservation of Waterlogged Canvas from Port Royal, 90
7 Glass Conservation, 93
Devitrification, 96
Removal of Sulfide Stains from Lead Crystal, 96
Consolidating Waterlogged Glass Using Passivation Polymers, 96
An Effective Silicone Oil Treatment Strategy, 96
Reconstruction, 98
Case Study: Preservation of Seventeenth-Century Glass Using Polymers, 100
Case Study: Preserving Waterlogged Glass and Cork, 108
8 Ivory and Bone, 112
Basic Structural Differences, 112
Equipment Setup for Very Fragile Bone and Ivory, 113
Case Study: Consolidating Friable Bone, 114
Case Study: Ivory from Tantura-B Excavations in Israel, 115
Case Study: Waterlogged Tusks from Western Australia, 116
9 Expanding the Conservation Tool Kit, 119
Computerized Tomography and the Stereolithographic Process, 120
Case Study: Scanning an Encrusted ArtifactCT Scanning Used as a
Diagnostic Tool, 121
New ToolsNew Directions in Research, 121

Notes, 125
Index, 127
 Foreword
ix

Preservation with reactive chemicals can be mers and cross-linkers to penetrate cell walls
called by many names but the most promi- and then adding catalyst. Each technique has
nent is von Hagens term, plastination. Other its advantages.
terms are chemical bulking and reactive ll- Both techniques often require specimen
ing. They are all used to describe the process preparation, or preservation, which might
of coating and impregnating cells with ma- require rinsing them clean and setting the
terials that react with the walls to prevent tissues with formalin solution. Next is the
water from attaching and reacting. In many medium exchange, where water in cells and
cases, we are also talking about lling the cell cavities is exchanged for a more compatible
or opening to prevent collapse. and less reactive medium, usually acetone,
The goal is to place a substance on the but sometimes alcohols. The exchange is con-
cell that reacts to the carbonols (COH) tinued until most of the water is eliminated.
present on much of the cell surface. Reac- The next step is penetration with reactive
tive silanes (cross-linkers) are particularly materials. The classic von Hagen procedure
adept at this and can be combined with poly- uses a mixture of a catalyst (typically dibutyl-
mers (carbonol or silanol [SiOH] ended tindilaurate) and a silanol-ended polymer.
polymers) that also react with the silane This technique involves immersing the ac-
cross-linkers. Thus we have silane cross-link- etone-saturated specimen in a bath of this
ers that react to the cell wall, to each other, mixture under a very slight vacuum and boil-
and to the polymer. Almost always, this re- ing o the acetone with the vacuum to leave
quires one or more catalysts to hasten and little voids that are then lled with polymer/
complete the reactions. catalyst mixture. When the acetone is gone,
We can use very reactive cross-linkers like the cavities are drained of excess solution and
methyltriacetoxysilane MeSi(OCOCH3)3 or a cross- linker, typically tetraethoxysilane
less reactive cross-linkers like methyl- Si(OEt)4, is added. As this penetrates the cavi-
trimethoxysilane MeSi(OCH3)3. Typically, ties, the polymer, cross-linker, and catalyst
when deep penetration into cell walls or body converge and react. The reaction is permit-
cavities is desired, slower-to-react materials ted to continue until it is complete (typically
are used to postpone the reaction until pen- in some heated chamber since often the
etration is fairly complete. Alkoxy functional cross-linker is put into the system from a
(alcohol releasing) cross-linkers are pre- vapor).
ferred, as their reaction to the cell wall or to Our technique uses formalin (if desired)
the polymer usually requires a catalyst. The and acetone to displace water. A mixture of
von Hagen procedure uses a polymer and a polymer and cross-linker is then added to
catalyst to impregnate and coat the cells fol- accomplish cell penetration. The polymer
lowed by the introduction of the cross- and cross-linker mixture is preferable be-
linker. Our technique involves using poly- cause it is more stable than the polymer and
 catalyst mixture. Our mixture thus has a . Air exposure yields partial hydrolysis
longer shelf life, and material that is left over followed by condensation:
x or drained from the cavities can be stored and a. SiX + HOH
SiOH + HX
reused. The lack of reactivity of the rst two b. SiX + SiOH
SiOSi + HX
ingredients allows most of the work to be
done at ambient temperature and not in the The work done by C. Wayne Smith, oth-
freezer, providing both an equipment and a ers, and me concerns the investigation of
time advantage. dierent combinations of cross-linkers, poly-
The next step is the addition of the third mers, and catalysts, along with variations in
ingredient, the catalyst. Dibutyltindiacetate the exposure techniques. For instance, in
is preferred to dilaureate because the dia- most of Dr. Smiths work we use methyl-
cetate, being smaller, penetrates cells faster. trimethoxysilane MeSi(OMe)3 as the cross-
Typically, catalyst is painted on the surface linker instead of the tetramethoxysilane
and rinsed through open cavities, putting a Si(OEt) 4 favored by von Hagen. The
maximum catalyst dose at the surfaces to MeSi(OMe)3 is only trifunctional while the
enhance penetration. The specimen is then ethylorthosilicate is tetrafunctional: using
put out in the air to cure. The cure in the von MeSi(OMe)3 thus yields a lesser cross-link
Hagen procedure takes place when the cross- density and a more pliable specimen. The
linker penetrates the system. The von Hagen remainder of the procedure is like cooking:
procedure could also use the painting or dip- ingredients are mixed and matched to achieve
ping technique to expose the specimen to this a desired eect. To get a more pliable, leath-
third ingredient (the cross-linker), but that ery touch, use a trifunctional cross-linker and
might require a posttreatment procedure of a higher molecular weight polymer. To get a
exposing the specimen to air. Our procedure less leathery, stier touch, several procedures
exposes the specimen to air to bring it into can be performed: add more cross-linker to
contact with the water in the air to allow ex- the polymer/cross-linker mix to get some in-
cess cross-linker to react. Our method uses situ formed resin in the system; use a
an excess of cross-linker whereas von Hagens tetrafunctional cross-linker; use a lower mo-
does not; the water in the air completes the lecular weight polymer; or use some combi-
reaction. Our mostly room temperature pro- nation of these procedures.
cedure adds cross-linkers to polymers in the In addition to the variations in the func-
initial step and postpones the use of the tionality of the cross-linker and length of the
smaller catalyst to the nal step. The proce- polymer chain, there are also variations in the
dure is relatively quick. The specimens can catalysts that can be used. The workhorse
usually be allowed to cure in the air after one catalyst for our purposes is dibutyl-
or two weeks. tindiacetate. However, this is a forever active
The following equations illustrate the cure catalyst, and the system is forever equilibrat-
reactions: ing; this also happens with the dibutyl-
tindilaurate generally used with the von
. The reaction to the cell wall is SiX + Hagen procedure. Divalent tin catalysts yield
HOC
SiOC + HX X = alkoxy a nonequilibrating system because they even-
. The reaction to the polymers is SiX + tually become noncatalytic tin oxide, but they
HOSi
SiOSi + HX are dicult to use because they hydrolyze and
. The reaction with an excess of cross- lose activity if excess water is present or if
linker and insucient water is SiX + SiX no the catalyst is prematurely introduced. They
reaction are best reserved for special situations.
 Titanate catalysts are another possibility, to give good penetration of cells and cavities
but again their high reactivity makes for di- and to suciently complete reactions to pre-
culty in use, so they are also best reserved for vent degradation. Science also provides some xi
special situations. The polymers, cross-link- order of additions that favor penetration and
ers, and catalysts used with silicone sealants the various combinations to give the tough-
are the types of materials that can be used in ness when needed. Art is needed to choose
preservation. However, in preservation, we acceptable combinations to give the desired
can and do sometimes use no polymer, just eects to all manner of things one might wish
catalysts and cross-linker. It is an especially to preserve, such as skin, hair, glass, wood,
eective way to preserve old glass retrieved paper, or leather.
from the sea. Only short polymers can be Preservation with reactive chemicals is like
used with glass systems since a leathery touch cooking. The science of cooking provides
is undesirable in this application. This is all nutritious combinations to sustain health,
part of the science of preservation with reac- but the art of cooking provides the combi-
tive chemicals. With glass, polymeric cross- nations that taste good. The science of pres-
linkers are also quite useful. With hair, and ervation yields many combinations that will
hide with hair, one often wants to rinse out penetrate and preserve, but artistry deter-
the preservation chemicals before the cure mines which combination gives the result
proceeds. This is again part of the art of pres- desired in a specic specimen. That art is
ervation with reactive chemicals. what is primarily depicted in this book.
There is an art and a science to preserva-
tion with reactive chemicals, both of which J. M. Klosowski
must be practiced to achieve a desired result. Independent Consultant
Keep this in mind as you read through this and Scientist Emeritus for the
book. Science provides proper combinations Dow Corning Corporation
xii
 Acknowledgments
xiii

As a graduate student, I had the great privi- of artifacts, this is irresponsible archaeology.
lege of working with Donny Hamilton at the Delving into their comments more deeply,
Archaeological Preservation Research Labo- however, I was not surprised to learn that lack
ratory at Texas A&M University. In the eld of funding was not the sole reason for the
and in the laboratory, I was given the oppor- backlog of artifacts waiting to be conserved.
tunity to excavate and conserve a wide range Lack of knowledge and lack of faith in exist-
of artifacts from nautical excavations at Port ing technologies were key reasons expressed
Royal, Jamaica. It did not take long to see that, for holding o on the preservation of some
to some degree, our conservation eorts were artifacts.
in vain. Since there were no facilities for long- All my mentors have taught me the same
term controlled curation, many artifacts con- valuable lesson: knowledge is not static. Ar-
tinued to deteriorate even after appropriate chaeological chemistry and the development
conservation strategies had been applied. of better conservation solutions have been a
The long-held tenet that all conservation serendipitous adventure, resulting in the
methods had to be reversible also proved trou- awarding of three patentsthe rst ever
bling to me. As a graduate student, I had ob- awarded in the Department of Anthropol-
served rsthand that, in many situations, at- ogy. While Donny Hamilton, head of the
tempts at rst removing bulking agents from Nautical Archaeology Program at Texas A&M
fragile organic artifacts and retreatment us- University, and Jerome Klosowski, senior sci-
ing new conservation methods and materials entist at Dow Corning Corporation, are
resulted in additional damage to the artifacts coinventors of the Passivation Polymer tech-
that I could not justify. I soon found that con- nologies, their contributions have been more
servators from many countries held similar than scientic. I thank them both for their
doubts about the true reversibility of conven- encouragement, scientic visions, and deep-
tional treatment strategies. The most impor- seated belief in this research. I would also like
tant early lesson learned was that there were to thank Ole Crumlin-Pedersen, Senior Re-
very few conventions in the conservation of searcher at the Center for Maritime Archae-
organic artifacts; each artifact presented a ology in Roskilde, Denmark, for bringing the
unique challenge. It became apparent that if problems of long-term stability of PEG-
we were to truly address the conservation edict treated artifacts to my attention. His insights
of long-term well-being of the artifact, have had a tremendous impact on my re-
dierent treatment methods and materials had search and on the science of maritime archae-
to be devised to ensure long-term survival in ology. Many thanks, also, to Research Coor-
a noncurated environment. dinator Mark Gilberg of the National Center
Lack of funding was also an issue. Archae- for Preservation Technology and Training, for
ologists everywhere said they could not aord supporting initial research in deveoping Pas-
to conserve all the artifacts recovered from sivation Polymer technologies.
their excavations. Apart from the tragic loss The contributions of the extended research
 team at Dow Corning Corporation have been forms the unique alchemy of art and chem-
innumerable. I thank Leon Crossman and istry that is the basis of my research.
xiv Arthur Rathjen for championing the cause It is impossible to personally acknowledge
of advancing archaeological chemistry and all the conservators who have contributed to
joint research with Texas A&M University. my research. Your questions, concerns, and
Their generosity and commitment to provid- requests for assistance in the preservation of
ing educational resources for schools and artifacts have been the driving force behind
universities are inspirational. this research. I hope this book answers some
Without the nancial support and encour- of your questions. My greater hope, of course,
agement of Woodrow Jones and Ben Crouch, is that this work serves to generate new ques-
both of the College of Liberal Arts, develop- tions and directions in conservation research.
ment of Passivation Polymer technologies Where new insights abound, there is no stasis.
would not have been possible. Their vision
for advancing archaeological chemistry re- C. Wayne Smith
search within the College of Liberal Arts Texas A&M University
Archaeological Conservation Using Polymers

 Introduction
3

T
he archaeological conservation of highly with which many of these techniques can be
perishable organic matter has always used makes them viable for various conser-
been problematic. In , for example, vation challenges and for laboratories work-
the Texas Historical Commission began ex- ing with limited budgets.
cavating Ren-Robert Cavelier, Sieur de La In addition to discussing the advantages
Salles vessel, La Belle, which sank o the oered using Passivation Polymer technolo-
Texas coast in . The recovery of this ves- gies and materials, I consider a concept sel-
sel has presented a host of unique challenges dom addressed in conservation: artistry.
in conserving wood, basketry, leather, and an Variance in equipment, relative humidity,
assortment of supplies necessary to start a laboratory layout, intended end results, and
colony in the New World. Entombed in the level of expertise are a few variables that play
mud that sealed it from decay for more than an important role in the development of
three centuries, the waterlogged hull and preservation strategies and protocols that
hundreds of thousands of fragile artifacts, work for each technician. The key to consis-
including brain matter in the skull of one tent and aesthetically correct samples using
unfortunate sailor, would have been a futile Passivation Polymers, and indeed all conser-
conservation eort without new preservation vation materials, is a willingness to explore
technologies. treatment parameters and combinations of
The aim of this book is to help conserva- polymers.
tors develop successful treatment strategies In chapter the Archaeological Preser-
for organic materials using Passivation Poly- vation Research Laboratory at Texas A&M
mers. These polymers are part of a series of University is used as a model to illustrate
chemistries and techniques developed at an eective layout for day-to-day conserva-
Texas A&M University to provide ecient tion of small organic artifacts using poly-
and cost-eective conservation strategies for mers. Too often, the entire working budget
the purpose of archaeological preservation. of a conservation facility is spent in the ac-
Museums and historical societies have ex- quisition of elaborate equipment. This is a
pressed the need for durable artifacts used shame since well-trained conservators
in traveling exhibits and interactive displays. should be the rst consideration of any labo-
In many cases, traditionally treated artifacts ratory. Accordingly, a short list of essential
cannot withstand the rigors of extended trav- instrumentation is discussed in chapter .
eling and handling. Passivation Polymer The distinction is made between equipment
treated artifacts are more stable than their needed for conservation of artifacts and
counterparts preserved using older technolo- equipment and necessities for the purpose
gies and can withstand sensible handling ex- of research and development; small neces-
tremely well. These preservation processes are sities and chemicals are also discussed.
not a panacea for all artifacts, but the ease While conducting postdoctoral research at
 Dow Corning Corporation, I learned from ing polyethylene glycol (PEG) methods. Con-
professionals how to work eectively and eco- served during a time when PEG technology
4 nomically in the laboratory. Since it seems was new, these artifacts have not fared as well
that all archaeological projects are as artifacts preserved using the well- dened
underfunded and understaed, these cost- PEG treatment strategies developed in the
saving suggestions can remove some of the latter half of the twentieth century by Per
economic stresses involved in outtting a Homann, David Grattan, and others. Re-
conservation facility. treatment strategies developed to stabilize
Chapter examines some of the mechani- these dicult artifacts are discussed in a sec-
cal processes used for Passivation Polymer tion entitled Reprocessing and Stabilization
processing of organic materials from marine of PEG-Treated Wood and in the case study,
and land sites. Two experiments are provided The Re-treatment of Two PEG-Treated
to demonstrate water/solvent and solvent/ Sabots.
polymer exchange processes that are funda- New wood treatment strategies using self-
mental to many of the experiments and case condensing alkoxysilane polymers are cov-
studies throughout the book. In Dowel Ex- ered in another case study, Treatment of
periment, penetration depth of polymers Waterlogged Wood Using Hydrolyzable,
and the presence of water in organic materi- Multifunctional Alkoxysilane Polymers.
als are discussed. The experiment illustrates These treatment strategies enable the conser-
the relationship between solvent dehydration vator to successfully preserve waterlogged
and time in the treatment of wood. A second wood by introducing an alkoxysilane that, in
experiment, Mass Spectrographic Analysis the presence of atmospheric moisture, self-
of Out-Gases Created from the Dehydration condenses to form polymers that accurately
of Archaeological Wood Samples, illustrates maintain the diagnostic attributes of water-
the dierences in exchange potential of or- logged wood.
ganic artifacts that have been dehydrated us- Numerous approaches and suggestions for
ing a single solvent dehydration process as the preservation of waterlogged and desic-
compared to combined solvent processing. cated leather are outlined in chapter . Case
Chapter discusses the preservation of studies are oered to demonstrate the adapt-
wooden artifacts using Passivation Polymers ability of the baseline treatment process to
and is divided into two sections: waterlogged accommodate even badly fragmented and
wood and dry-site artifacts. Bacteria, fungi, desiccated artifacts. Important consider-
and the loss of structural integrity can be ations for the storage and display of leather
caused by waterlogging or desiccation from artifacts are also covered.
extreme dryness. Regardless of the objects One area of great concern for all conser-
archaeological provenance, stabilizing ar- vators is the treatment of composite arti-
chaeological wood is a great challenge for the facts. Often, it is necessary to disassemble
conservator. Mechanical destruction caused an artifact into its component parts prior to
by freeze-thaw cycles in lakes, swamps, and treatment. However, this can damage frag-
bogs may act to accelerate the degradation ile portions of the artifact. Additionally,
of wood. Regardless of what type of degra- structural changes in one or more compo-
dation has occurred in a wooden artifact, nents of the artifact often make the reassem-
Passivation Polymer technologies have proved bly stages of conservation dicult. In chapter
eective in preserving diagnostic attributes. the case study Preservation of a Composite
Conservators face another challenge: mu- Artifact Containing Basketry and Iron Shot
seums around the world are lled with arti- details how wicker basketry was stabilized
facts, big and small, that were preserved us- prior to removing heavy layers of concretion
encasing the basket of shot. After the basket preserve sections of ivory tusk from a ship-
was processed in stages, the iron shot that had wreck o the coast of Australia.
been removed from the basket were preserved In chapter I explore some of the new 5
using electrolytic reduction. The ease with tools and technologies that can help conser-
which calcareous material was removed from vators devise more eective conservation
the surfaces of the iron shot and the result- strategies. Until recently, the conservation
ant treatment of the iron suggest that expo- process has been hindered by the inability of
sure to Passivation Polymers may have acted traditional radiographs to supply sucient
to preserve the iron, arresting surface exfo- data to give the conservator a complete view
liation and pitting up to, and during, treat- of an artifact concealed in a concretion.
ment. Through computerized tomography (CT)
Preservation of cordage and textiles from and computer-aided design (CAD), however,
marine and land sites is discussed in chapter three-dimensional models of artifacts can be
. Numerous conventional treatment strate- created using laser-driven polymerization of
gies adapted for use with polymers are illus- each slice of CT data. Stereolithography is
trated in a series of case studies. another technology that is discussed.
It has been argued that glass is not an or- Many conservation materials, considered
ganic substance. While this is true, uxes used part of the conservation tool kit, are poly-
in the making of ancient glass are organic mers. Research has shown that these materi-
materials, which accounts, in part, for the als, once thought to be reversible, have a ten-
unique chemical structure of ancient glass. dency to cross-link with the organic substrate
Treatment methods, accelerated weathering of the artifacts they were intended to stabi-
testing, and scanning electron microscopic lize. While most of us tend to visualize chem-
analysis of polymer-treated glass are pre- ists in white lab coats when we think of cross-
sented in chapter . Several case studies illus- linked polymers, the process of cross-linking
trate the abilities of alkoxysilane materials to is a natural process. As archaeological chem-
stabilize badly deteriorated glass. Treatment istry plays an ever-increasing role in our dis-
techniques used for preserving extremely thin cipline, research scientists and conservators
sections of onion bottle glass are discussed alike will see the expanding role of polymers
to illustrate the degree of creativity that can in archaeological conservation. The polymer
be employed in the treatment of friable ob- chemists who helped dene the initial
jects. Like polyvinyl acetate and other tradi- mechanisms at work in Passivation Polymer
tional treatment strategies, reversibility of processes found the task of adapting their
these treatment methods remains an issue. thinking to t archaeological perspectives
Data from accelerated weathering and challenging. Though new terms will inevita-
nuclear magnetic resonance imaging experi- bly nd their way into the conservation lexi-
ments indicate that such glass remains re- con, I have tried to keep the use of new ter-
markably stable much longer than glass minology to a minimum in this book.
treated using conventional methods. Experience at the Archaeological Preser-
A discussion of ongoing research in the vation Research Laboratory (APRL) at Texas
preservation of badly deteriorated ivory and A&M University has taught us that while
bone from marine and terrestrial sites is the complete reversibility of conservation treat-
focus of chapter . In one case study treat- ments is desirable, it is a state that is some-
ment strategies applied to the preservation times easier philosophized than realized. Re-
of nely carved ivory from the Tantura La- treatability may be a more realistic end goal
goon excavations in Israel are discussed. An- when determining suitable conservation
other case study illustrates methods used to methods for any given artifact.
 The experiments and case studies oered Invariably, success in using any conserva-
in this book have proved successful for the tion strategy requires patience and a marriage
6 preservation of a wide range of artifacts. I of technical and artistic know-how. There-
hope these materials give the reader a broad fore, time should be spent experimenting
understanding of how Passivation Polymers, with baseline processes before attempting
and other silicone oils, may be used for the conservation of artifacts. There is no substi-
preservation of organic artifacts. Because no tute for hands-on experience, which allows
two pieces of wood are exactly alike, selected the conservator to better understand how to
treatment strategies must be adapted to treat use polymers in the preservation of organic
artifacts on a case-by-case basis. Polymer- material culture and how these treatments
based materials oer the conservator innite can be adapted to suit each artifact. This type
combinations of material, so that the preser- of experimentation will also help the conser-
vation needs of individual artifacts can be vator understand and better utilize the equip-
addressed. ment used for silicone oil preservation.
Anyone experienced in the complexities of The goal of this research has been to fuel
eld conservation would agree that artifact new dialogues in conservation and to en-
triage and stabilization play a major role in courage research along new avenues. Most
the eventual state of an artifact. Passivation important, this research encourages the
Polymers allow the conservator to stabilize reader to gain a new understanding of the
friable materials in the eld so they can be material culture we strive so hard to pre-
more easily transported to the laboratory, serve. The development of polymer chem-
where continued preservation can be under- istry for use in archaeological conservation
taken. Polymers can help stabilize artifacts, is in its infancy. To date, development of Pas-
even if more traditional conservation meth- sivation Polymers has helped bring new per-
ods are used once the artifact is secured in spectives to bear on the treatment of organic
the laboratory. artifacts.
 CHAPTER 1
7

Laboratory

Setup

E
xperimentation conducted on the de proof freezer system is essential for develop-
velopment of Passivation Polymers re mental research. For routine conservation Fig. .. A functional
quired the use of sophisticated labora- work, such equipment is not necessary. researchworking
tory equipment. This book, however, oers laboratory: (A) double-
some basic processes that can be imple- sided workbench; (B) sink;
(C) eye wash system;
mented without the need for elaborate tech- Major Instrumentation
(D) under-cabinet
nology. While there is no one ideal setup for
ammable liquid storage;
a functional lab, layout and location of equip- The investment in elaborate equipment to
(E) vented fume hood;
ment and materials should always adhere to outt a laboratory for polymer research and (F) vented warming oven;
guidelines established by reputable agencies artifact conservation is minimal. When (G) explosion-proof
such as the Occupational Safety and Health outtting a laboratory, safety should be the freezer; (H) air-drying
Administration (OSHA). Worldwide, a num- primary concern. High-ow ventilation and shelves; (I) computer with
ber of agencies act in the interest of promot- an eye-washing system are two of the most digital imaging camera;
ing laboratory safety. Many of these have important considerations. (J) electronic scales.
Web-based resources available, free of charge.
Although safety was a major consideration
in choosing material for the development of
Passivation Polymer technologies, many of
the materials suggested for use in the labora-
tory require proper handling and storage. All
technicians working in a chemistry lab need
to be familiar with proper procedures for han-
dling and disposing of chemicals. A complete
inventory of chemicals should always be kept
in case of an emergency, and material safety
data sheets (MSDS) for chemicals used in the
facility must be available for immediate con-
sultation. The MSDS catalog contains vital
contact information as well as important in-
structions if medical assistance is needed.
Figure . shows the oor plan of the Ar-
chaeological Preservation Research Labora-
tory at Texas A&M University. Much of our
developmental research was conducted in
this facility. Equipment such as the explosion-
 Because solvents are used in conjunction beaker with a vacuum plate acting as its
with many preservation processes, selecting lid. All major suppliers sell these beakers
8 equipment that will not break down when and most sell vacuum plates and gaskets
exposed to vapors is essential. Often, stain- that t perfectly. Many restaurant supply
less steel beakers and small supplies can be houses carry the same name brand
found at a restaurant supply store for a frac- beakers at much lower cost.
tion of the price of identical equipment sold Fume Hood. For many laboratories,
through a chemical supply company. Gener- acquisition of a fume hood is a major
ally, however, it is best to select equipment expenditure. While building codes vary,
certied for laboratory use. OSHA and other agencies have estab-
lished standards that should be followed.
Vacuum Pump. Older piston-type pumps These agencies have Web sites and
work best in the laboratory. Modern personnel willing to assist with design
pumps are more ecient, but they are questions.
lightweight in construction, and many Electronic Scales. Electronic scales, capable
companies use plastic internal parts. of measuring in milligrams, are also an
These tend to break down rapidly in the essential tool. Many of these instru-
presence of acetone or acetone vapors. ments can be zeroed after each chemical
Vacuum Gauge. A simple manometer is addition, making the measurement
essential, preferably one with an all- process easy and more accurate. Be sure
metal construction. In more inexpensive to account for the full range and volume
manometers, gradation marks are of materials that will be used in the
generally painted onto the back plate. laboratory when choosing a scale.
During water/acetone exchange pro- Explosion-Proof Freezer. Not all the
cesses, any solvent fumes passing experiments in this book require the use
through the unit will cause the painted of an explosion-proof freezer. The use of
surfaces to crack and fade. a freezer may be benecial, however,
Dewar Flask and Gas Traps. The best way when a treatment strategy calls for
to trap harmful out-gases that contami- mixing even a small amount of catalyst
nate the vacuum gauge and vacuum into a Passivation Polymer solution. The
pump oil is to use a Dewar ask (g. catalyzation process is slowed apprecia-
.). These asks are available through bly when artifact processing is con-
any major chemical supply company. A ducted in a freezer. A freezer unit is also
Dewar ask should be well insulated helpful when a conservation strategy
and large enough in volume to hold requires a slow, controlled acetone/
two cold traps hooked in parallel, along Passivation Polymer exchange.
with a generous quantity of dry ice.
With the Dewar ask and cold trap
assembly placed in-line between the Small Necessities in the Laboratory
vacuum chamber and the vacuum
pump, the cold environment should Disposable Cups. Disposable containers are
condense solvent fumes and prevent practical when working with Passivation
their escape into the manometer and Polymer solutions. Before purchasing a
vacuum pump oil. large supply of any particular brand,
Beaker and Vacuum Plate. A small, however, test the container to be sure it
countertop vacuum unit can be con- does not react with solvents. While it is
structed from a one-liter stainless steel possible to clean glassware, paper cups
 and plastic containers that can be
discarded after use are time savers.
Stir Rods. Plastic or glass stir rods work 9
best with Passivation Polymers. Never
use wooden stir sticks or tongue depres-
sors to mix polymer/cross-linker
solutions. The wood will absorb the
chemicals dierentially and interfere
with the stoichiometry of the solution.
Tongs and Tweezers. Because artifacts come
in all shapes and sizes, it is a good idea
to have a variety of big and small tongs
and tweezers in the lab. Restaurant
supply companies are an excellent
Fig. .. Dewar ask with
source for large stainless steel utensils.
cold traps and vacuum
For specialty tweezers and implements
lines.
for handling fragile materials, hobby
shops and chemical supply companies
generally have a better selection. precautions are taken. After treatment,
Acetone. Many processes outlined involve artifacts are generally allowed to drain of
an acetone/Passivation Polymer ex- excess free-owing oils while sitting on
change process. Acetone and other screen material. Since these materials are
solvents are necessary for expedient, often reusable, it is advisable to save
thorough exchange of water with runo Passivation Polymers. After
polymer preservatives. Proper handling, draining, however, the surfaces of the
storage, and waste management proce- artifact need to be wiped of excess
dures are mandatory. Stainless steel or Passivation Polymer solution. Layers of
high-density polyethylene safety cans lint-free paper and, in some cases,
outtted with spring-loaded safety caps newspaper can facilitate cleanup.
are the safest and easiest way to handle
solvents in the laboratory. These are
available in a variety of sizes from Chemicals
chemical and safety supply houses.
Metal Screens and Vegetable Bags. For Conservators working with Passivation Poly-
smaller artifacts, aluminum window mers to preserve organic artifacts need ac-
screen can be friction-tted on top of cess to a variety of chemicals, but they must
the artifact in solution to prevent them keep in mind important considerations in
from oating during acetone/polymer selecting and safely storing chemicals for use
exchange processes. Mesh bags, such as in a laboratory. Hundreds of silicone oils,
those used for bagging perishable foods cross-linkers, and catalysts are available for
in a grocery store, can help keep small use in an archaeological chemistry labora-
objects together during treatment. tory. The most important consideration in
Because of their open mesh design, these selecting polymers is the safety of laboratory
bags do not impede treatment. sta and students. Passivation Polymers were
Absorbent Materials. Working with selected for developmental research at Texas
Passivation Polymertreated artifacts A&M University because they are relatively
can be messy unless some simple safe to use and because research has proven
 Table . Cross-Linkers

10 Catalyst Noted Attributes

CR-20 Flexible stoichiometry
Generally 35% addition by weight of the chosen polymer
High % additions of CR-20 result in slightly brittle polymer structures
CR-22 Less flexible stoichiometry
Generally 3% addition by weight of the chosen polymer
Tends to create more brittle polymers

that they do not have detrimental eects able for all materials being used in the labo-
when applied to organic artifacts. ratory. These sheets contain the necessary in-
Regardless of the relative safety of the formation for lab personnel to work safely
chemicals chosen, the conservator should al- with chemicals. In the event of a spill, these
ways work in a well-ventilated environment. sheets also contain guidelines for safe and
Accordingly, a fume hood large enough to controlled cleanup. Medical treatment infor-
conduct day-to-day work is essential. While mation and emergency contact numbers
a fume hood may be the biggest investment should also be listed on the sheets. MSDS in-
when outtting a laboratory, it is essential formation is available on the Web. If the labo-
equipment. If used properly, the airow cre- ratory has a computer, it is advisable to have
ated by a fume hood will ensure that contami- a hot key or posted instructions so that per-
nants and harmful vapors are not a problem. sonnel can easily access safety information
Material safety data sheets must be avail- and emergency assistance.

Table . Catalysts
Material Application Method Chemical Reactivity Results

CT-32 Wiped, painted Long reaction time Clear reaction

(DBTDA)
Vapor deposition Long reaction time Clear reaction
containment chamber at (18 hrs)
elevated temperatures
Containment chamber at Longer catalyst life Clear reaction
room temperature
CT-30 Wiped, painted Shorter catalyst life Clear reaction
than CT-32
Vapor deposition Shorter catalyst life Clear reaction
containment chamber at than CT-32
room temperature
Containment chamber Catalyst life not as good Clear reaction
at room temperature as CT-32
CT-34 Wiped, painted Very fast Clear reaction
Vapor deposition Very fastoften slightly Often a yellow
containment chamber at yellow coloration coloration; often a
elevated temperatures powdery residue
that can be brushed
clean
Containment chamber Oxidation occurs quickly Often oxidation occurs
not recommended without canalization
taking place
 CR- is the main cross-linker used for ever, to determine stoichiometric consider-
processing archaeological materials.1 CR- is ations for use with organic artifacts. Table .
a hydrolyzable, multifunctional alkoxysilane provides an overview of CR- and CR-. 11
capable of tying two or more polymer chains CT-, a dibutyltin diacetate compound,
together. It has a specic gravity of . at C is the most commonly used catalyst for pre-
and a viscosity of . centistoke (CST). Added serving organic artifacts.2 This catalyst can be
to PR-, KP, PA Fluids, and -, CR- applied either as a topical application or
is an eective cross-linking agent. through vapor deposition to cure the polymer-
CR- cross-linker is similar to CR-. It impregnated artifact. CT- is a tin-based
is a tetra-functional alkoxysilane that has a catalyst that has a working life of approxi-
specic gravity of . at C and a viscos- mately hours. It is a good general pur-
ity of approximately . CTS. Observations pose catalyst for use in the laboratory, with
suggest that artifacts treated with polymer a specic gravity of . and a molecular
solutions including CR- are slightly more weight of .. Its chemical formulation is
malleable after treatment than those treated C 12 H 24 O 4 Sn. Exposure to CT- fumes
with CR-. More testing is required, how- should be avoided as should direct contact

Table . Silicone Oils

Material Viscosity Expected Result Additive Catalyst Approximate

Reaction Time

PR-10 Thin Rigid, easy CR-20 CT-32 24 hr

penetration
CR-22 CT-30 15 hr
CR-20 or CR-22 CT-34 20 min
PR-12 Thin Rigid, easy CR-20 CT-32 24 hr
penetration
CR-22 CT-30 15 hr
CR-20 or CR-22 CT-34 20 min
PR-14 Medium Less rigid, easy CR-20 CT-32 24 hr
penetration
CR-22 CT-30 15 hr
CR-20 or CR-22 CT-34 20 min
KP80 Very thin Very rigid CR-20 CT-32 24 hr
depending on
density of organic
substrate
CR-22 CT-30 15 hr
CR-20 or CR-22 CT-34 20 min
PA Fluid Very thin Not as rigid as CR-20 CT-32 24 hr
KP80 but firmer
than PR-10
CR-22 CT-30 15 hr
CR-20 or CR-22 CT-34 20 min
4-7041 Thin Similar to PA fluids CR-20 CT-32 24 hr
but may be more
brittle
CR-22 CT-30 15 hr
CR-20 or CR-22 CT-34 20 min
 with skin. When used in artifact conserva- frigerator during catalyzation. Table . pro-
tion, the catalyst remains chemically stable. vides an overview of these catalysts.
12 CT- is also a tin-based catalyst. CT- The main silicone oils used for preserv-
diers from CT- in that it has a working ing organic materials in the Archaeological
life of approximately hours in a contained Preservation Research Laboratory are dim-
environment. CT- has a specic gravity of ethyl siloxane, hydroxy-terminated polymers
., a molecular weight of ., and a with up to % dimethyl cyclosiloxane added
chemical designation of Sn(C8H15O2)2. (see table .). PR-, the medium viscosity
The fastest acting of the three catalysts of the readily available silicone oils, is an
used for experimental research is CT-. Also odorless polymer with a specic gravity of
known as Tyzor TPT Titanate, it is a mix- . at C and a viscosity of ,. CST.
ture of tetraisopropyl titanate (%) and iso- PR- is slightly more viscous than PR-. It
propyl alcohol (%). CT- has a working has a specic gravity of . at C and a
time of approximately hours in a contained viscosity of ,. CST. KP (. CST)
environment. It is pale yellow in color, which is less viscous than PR- silicone oil and has
is problematic when the catalyst is applied a specic gravity of . at C. PA Fluid is
using vapor deposition. Through the process also a dimethyl siloxane, hydroxy-terminated
of catalyzation, a yellow powder can form on silicone oil with octamethylcyclotetrasiloxane
the surfaces of the artifact. In most cases, the and decamethylcyclopentasiloxane added. PA
thin layer of powder is easily removed with a Fluid has a viscosity of . CST and a specic
soft bristle brush. Chemical reactivity of the gravity of . at C. Another silicone oil,
catalyst can be accelerated by warming the -, is a polydimethylsiloxane, hydroxy-
solution. Reaction time can be slowed by terminated polymer, with a viscosity of
keeping the artifact cooled in a freezer or re- CST and a specic gravity of . at C.
 CHAPTER 2
13

Baseline

Mechanisms

S
ilicone oils do not mix with water. This, Removing water from any waterlogged
then, is the rst challenge a conserva- artifact can be harmful, especially if water-
tor faces after deciding to preserve logging has caused severe cellular damage. In
waterlogged materials using polymers. Passi- many cases, water in the cells is all that is
vation Polymer artifact treatment involves maintaining the diagnostic attributes of the
three main steps: water/acetone displacement artifact. The conservator must therefore un-
(WAD) or dehydration, acetone/polymer (AP) derstand the importance of proper dehydra-
displacement or Passivation Polymerization, tion as a precursor to treatment with any
and catalyzation of the polymer-impregnated bulking agent, including Passivation Poly-
artifact. mers. Pretreatment dehydration is generally
Two experiments in this chapter demon- not required for dry and desiccated organic
strate the relationship between dehydration materials. If wood or basketry feels damp due
of water from organic materials and the abil- to absorption of atmospheric moisture, ei-
ity of silicone oils to penetrate into organic ther dehydration of the samples or storage
substrates. In Dowel Experiment, the im- in a low humidity environment for a few days
portance of WAE displacement is illustrated. may be necessary to ensure the rapid uptake
Dehydration processes and length of time in of polymers during treatment.
dehydration inuence penetration depth and
rate of AP exchange in wooden dowels. The
second experiment, Mass Spectrographic Dowel Experiment
Analysis of Out-Gases Created from the De-
hydration of Archaeological Wood Samples, The purpose of this experiment was to de-
discusses single- and multiple-solvent dehy- termine rates of acetone and silicone dis-
dration techniques for organic materials and placement into the matrix of groups of
how multiple-solvent dehydration may en- structurally solid wooden dowels that had
hance penetration depth. undergone water/acetone displacement for
Experimentation is ongoing to develop one, two, three, or four days before treat-
processes for determining nondestructive ment. Because of the tight, undamaged cel-
percentage dehydration of organic materi- lular structure of the dowels, this experiment
als. Many processes are available for deter- represents a dicult conservation scenario
mining absolute dehydration in organic for wood. The cellular structure of most wa-
substrates. Unfortunately, most require test- terlogged and desiccated timbers facilitates
ing methods that either damage the sample the AP process and shortens the time re-
or are potentially stressful to the artifact in quired to impregnate the wood. This ex-
general. periment showed that () the penetration
 depth of acetone is dependent on time and into the core, than did day- dowels. (Day-
vacuum processing, and () Passivation Poly- dowels received hours dehydration; day-
14 mers will only penetrate the dowels to the , hours; day-, hours; day-, hours.)
depth of acetone within the samples. Each day, the dowels were immersed in
Forty-eight dowels, each four inches long PR- and then treated to continuous ac-
and one inch wide, were prepared so that their etone/Passivation Polymer displacement for
cut ends could be sealed by dipping them in hours at a reduced pressure of . Pa
a tray of ve-minute epoxy. Sealing the end ( Torr). During this -hour period, two
surfaces ensured that all penetration of ac- samples were removed from the Passivation
etone and Passivation Polymers into the core Polymer solution after ., , , , , and
of the dowels was through the sides of wood hours. Hence, the time of dehydration and
cells and not the tangential, porous ends of the time of treatment were the only variables
the dowels. After the epoxy had dried to in the experiment. At the designated times,
form hard end caps, the dowels were placed dowels were removed from the beaker and
into a ve-gallon pail and immersed in fresh quickly surface-wiped with a paper towel.
industrial-grade acetone. A section of alu- Each dowel was labeled and then promptly
minum screen and a ve-pound weight were cut in half on a band saw. The wet line of
placed over the dowels to ensure that they Passivation Polymer penetration could be
remained submerged throughout treatment. seen and marked with a ballpoint pen on
The pail was then placed into a large vacuum the crosscut surfaces of each half of the
chamber. dowel. It was important to mark this pen-
All the samples were dehydrated for etration depth (PD) as quickly as possible
hours at room temperature and a reduced because, in many cases, this line of penetra-
pressure of . Pa ( Torr). Twelve tion diused quickly after the samples were
Fig. .. Cross-sectional
samples were removed from the pail daily and cut. Figure . shows a typical cross-sectional
view of dowel and
penetration depth
placed into a beaker of PR- silicone oil so- view of Passivation Polymer penetration of
measurement: (A) dry lution. While the rst samples were being a dowel.
core; (B) Passivation analyzed, the remainder of the dowels con- In this experiment, it was reasoned that
Polymer wet zone; tinued in dehydration so that day- dowels as long as all samples were processed in the
(C) measured mean had four times as much dehydration and, same manner and within the same time
penetration depth. presumably, deeper penetration of acetone frame, the data should accurately represent
the PD for each sample and the dierences
in PD between samples. The mean PD was
then calculated for each collection time. Af-
ter four days, four series of dowels had un-
dergone the same processes of acetone/Pas-
sivation Polymer displacement and the same
assessments for PD. Data for PD among all
the dowels for all four days could then be
compared (see tables ..). Dehydration
time was expected to correlate with PD of
acetone into the matrix of the samples. These
same samples should also have a greater PD
of Passivation Polymers into their matrix as
compared to samples treated with a shorter
period of acetone dehydration.
Table . One-Day Dehydration Data

Sample Number Treatment Time Penetration Depth Mean Penetration Depth 15

(in hrs) (in mm) (in mm)
8.1.1 0.5 1.75 2.875
8.1.2 0.5 4.00
8.2.1 1.0 4.00 4.000
8.2.2 1.0 4.00
8.3.1 2.0 5.00 5.000
8.3.2 2.0 5.00
8.4.1 4.0 6.50 6.750
8.4.2 4.0 7.00
8.5.1 8.0 6.50 6.500
8.5.2 8.0 6.50
9.6.1 24.0 7.25 7.375
9.6.2 24.0 7.50

Table . Two-Day Dehydration Data

Sample Number Treatment Time Penetration Depth Mean Penetration Depth

(in hrs) (in mm) (in mm)
9.1.1 0.5 5.50 5.500
9.1.2 0.5 5.50
9.2.1 1.0 5.50 5.750
9.2.2 1.0 6.00
9.3.1 2.0 6.50 6.750
9.3.2 2.0 7.00
9.4.1 4.0 7.00 6.750
9.4.2 4.0 6.50
9.5.1 8.0 7.00 7.250
9.5.2 8.0 7.50
9.6.1 24.0 8.00 8.375
9.6.2 24.0 8.75

Table . Three-Day Dehydration Data

Sample No. Treatment Time Penetration Depth Mean Penetration Depth

(in hrs) (in mm) (in mm)

10.1.1 0.5 5.50 5.750

10.1.2 0.5 6.00
10.2.1 1.0 7.00 7.000
10.2.2 1.0 7.00
10.3.1 2.0 7.50 7.500
10.3.2 2.0 7.50
10.4.1 4.0 7.75 7.875
10.4.2 4.0 8.00
10.5.1 8.0 8.00 8.500
10.5.2 8.0 9.00
10.6.1 24.0 8.50 8.500
10.6.2 24.0 8.50
 Table . Four-Day Dehydration Data

16 Sample No. Treatment Time Penetration Depth Mean Penetration Depth

(in hrs) (in mm) (in mm)
11.1.1 0.5 5.00 5.000
11.1.2 0.5 5.75
11.2.1 1.0 6.50 6.500
11.2.2 1.0 6.75
11.3.1 2.0 7.00 7.250
11.3.2 2.0 7.50
11.4.1 4.0 8.00 8.250
11.4.2 4.0 8.50
11.5.1 8.0 8.50 8.500
11.5.2 8.0 8.50
11.6.1 24.0 8.75 8.750
11.6.2 24.0 8.75

Data for Dowel Experiment dehydration and acetone pretreatment of

samples that appear to be dry play an im-
The data were analyzed by multiple linear portant role in the bulking process of organic
regression. As shown in gure . PD bore a materials. Acetone and other water miscible
logarithmic relationship to the hours the solvents act as a conduit for the exchange of
samples were in Passivation Polymer solu- Passivation Polymer solutions into the core
tion; the data were transformed by taking the of an organic specimen.
logarithms of the number of treatment In all cases, the degree of silicone depends
hours. The analysis showed PD depended upon the amount of water/acetone exchange
upon the days of dehydration (p < .) and that has occurred throughout the specimen
the log-transformed hours of treatment (p being treated. This certainly dictates that, for
< .). One outlier was seen; its exclusion most materials, the time spent in water/ac-
did not aect the results signicance. Hence, etone exchanges cannot be bypassed. Indeed,
Passivation Polymer penetration into the a good, simple test is needed to indicate when
matrix of these wooden dowels was directly acetone has completely penetrated the struc-
linked to the penetration depth of acetone ture being treated and as much water as pos-
in the samples. Water/acetone exchange sible has been removed from the specimen.

Fig. .. Penetration
depth vs. treatment time
stratied by days of
dehydration.
Mass Spectrographic Analysis of Out- hydration processes at the Preservation Re-
Gases Created from the Dehydration search Laboratory are represented in table ..
of Archaeological Wood Samples 17
Electron Ionization Mass Spectral Analysis
During development of Passivation Polymer-
ization processes, the degree of dehydration Electron ionization (E.I.) mass spectra of the
and the methods used for dehydration of out-gases from the ve waterlogged wood
organic materials were discovered to aect samples were acquired on a VG Analytical S
treatment results. Artifacts treated using wa- high-resolution, double-focusing magnetic
ter/ethanol exchange (WEE) and ethanol/ sector mass spectrometer. The VG S was
polymer exchange (EPE) processes produced equipped with a VG /J data system that
very dierent results from similar materials allowed computer control of the instrument,
preserved using WAD/AP processes. Mass data recording, and data processing. The E.I.
spectrographic analysis was used to study the source was maintained at C during the
out-gases from similar size pieces of water- analyses, while eV positive analyte ions
logged wood that were dehydrated using were extracted and accelerated to keV and
WEE and WAD exchange processes. then mass analyzed.
Two factors are important for conserv- Gaseous samples were collected in a one-
ing organic materials with silicone oils. liter glass bulb and allowed to leak into the
While it is not necessary to eliminate all mass spectrometer through a regulated in-
water from an artifact, it is essential to re- let. Liquid samples were analyzed by inject-
move as much of the free-moving water as ing IL of the sample into a C heated
possible. Acetone can easily be driven from reservoir. The sample was then allowed to leak
samples in a vacuum chamber, so a dehy- into the mass spectrometer through a glass
dration process that eectively causes a jet. Approximately mass spectra were taken
higher output of volatile vapors should also and averaged per sample; then the mean ion
facilitate a more eective exchange of sili- abundance per sample was reported. Back-
cone oils to replace the acetone. To evaluate ground measurements of water (m/z+)
which dehydration processes are best suited were determined to be approximately mV
for conservation practices utilizing silicone- at the amplication used to collect the analyte
bulking techniques, a test was devised to mass spectra.
study collected gases from dehydrated The rst method of dehydration was to
samples. Electron ionization mass spectro- simply immerse the waterlogged wood in a
graphic analysis was performed on the out- series of fresh acetone baths, allowing ad-
gases from waterlogged archaeological equate time for each bath to draw free-ow-
wood. The ve most commonly used de- ing water from the sample. The acetone used

Table . Commonly Used Dehydration Processes

Dehydration Process General Procedure

Acetone Series of baths of fresh acetone to displace water

Alcohol/acetone Series of alcohol baths, followed by a series of baths in acetone
Acetone/vacuum in freezer Series of acetone baths under vacuum at 0C
Alcohol/acetone/vacuum Alcohol and acetone baths at 0C
in freezer
Acetone/warm out-gassing Series of acetone baths, followed by warming the artifact to 62C
to drive off water and acetone
 was an industrial-grade solvent, which had tents within the gas trap were collected and
been stored in a -gallon drum for approxi- placed into sterile specimen vials for trans-
18 mately three months prior to use in this ex- port to the mass spectrometer.
periment. Because of the small sample size, The second waterlogged wood sample,
sample A was placed into a beaker contain- sample B, was dehydrated in much the same
ing ml of fresh acetone and allowed to fashion as sample A, except that, for initial
remain at room temperature for hours. dehydration, it was placed into a beaker con-
The sample was then placed into an identi- taining ml of ethanol and allowed to
cal volume of fresh acetone for an additional stand for hours at room temperature. The
hours. alcohol was then replaced with ml of
After this period of dehydration, the sample fresh acetone. After being capped to prevent
was removed from the acetone bath and lightly evaporation, the sample was allowed to sit at
surface-tamped with a dry paper towel to re- room temperature for an additional hours.
move excess acetone from the surface. Addi- Subsequently, sample B was removed from
Fig. .. Beaker and gas tional acetone was allowed to out-gas by plac- the bath and treated identically to sample A.
trap assembly mounted in ing the sample on a dry paper towel for two Waterlogged wood sample C was dehy-
dry ice: (A) vacuum line
minutes. The sample was then placed into a drated using a series of acetone baths that
from vacuum chamber
clean beaker containing ml of PR- sili- were placed in a freezer-mounted vacuum
holding artifact; (B) two
cone oil. The beaker was immediately con- chamber as a means of determining if the
gas traps in series between
the vacuum chamber and
nected to an in-line gas trap assembly with combination of low temperature and vacuum
the vacuum pump; an attached vacuum pump (g. .). Dry ice had a signicant eect on the dehydration
(C) vacuum line leading surrounding the trap assembly was used to process. Sample C was placed into ml of
to vacuum pump; (D) freeze the entire range of gases being given fresh acetone and placed into the freezer-
Dewar ask, holding the o by the sample under a vacuum of Torr. mounted vacuum chamber, where it re-
gas traps and dry ice. After minutes of applied vacuum, the con- mained under a vacuum of Torr for
hours. After hours, the wood sample was
quickly transferred to a clean beaker contain-
ing ml of fresh acetone and then returned
to the freezer vacuum chamber for an addi-
tional hours of dehydration using the same
vacuum pressure as before. Upon completion
of the second acetone bath, the sample was
removed from the freezer and surface-patted
using dry paper towels. Subsequently, sample
C was treated identically to samples A and B.
A fourth sample, D, received the same
treatment as sample C except that, instead of
two acetone baths under vacuum, initial de-
hydration was started using ethanol so that,
after hours, the sample had been dehy-
drated in one bath of alcohol and one bath
of acetone. After collecting out-gases from
the sample using the same apparatus and
methodology used for the other samples, the
vial was ready for analysis. (If a sample could
not be immediately analyzed, it was stored
in a freezer.)
 Because silicone oil is nonvolatile and
chemically pure, it seemed to be an ideal in-
ert substance in which to immerse the wood 19
samples during the out-gassing process. To
test this hypothesis, an additional wood
sample, E, was dehydrated in two successive
baths of ml of fresh acetone at room tem-
perature. The sample was then placed into a
clean beaker, where it was warmed by a heat
lamp at C until all the surfaces of the wood
were light brown in color (g. .). At this
point, simply touching the outer surfaces of
the sample indicated that it was dry, although molecular acetone in this sample was ap- Fig. .. Basic beaker and
the weight indicated that the interior was still proximately equal to samples B, C, and D. The gas trap setup with heat
heavy with liquid. As with the other samples, levels of molecular acetone were nearly lamp and temperature
out-gases from sample E were collected in a double in sample E, which was warmed dur- probe added: (A) beaker
with wood sampleno
dry icecooled gas trap for a period of ing out-gassing.
oil; (B) vacuum line to gas
minutes. After the products of out-gassing Sample B was initially dehydrated at room
trap; (C) gas trap
from sample E were placed into a sterile, temperature using a -hour alcohol bath,
mounted in dry ice;
sealed vial, it, too, was analyzed. followed by an acetone bath. Recorded levels (D) container with dry ice;
of nitrogen were approximately % less than (E) vacuum line to
Results those recorded for sample A, while the levels vacuum pump; (F) heat
of methyl and ethyl were reduced by approxi- lampadjustable height;
No attempt was made to evaluate the physi- mately % and %, respectively. Acetone (G) temperature probe.
cal appearance of the ve samples used in this levels in sample B were only slightly lower
experiment. Variability among samples is an- than those recorded for sample A.
other uncontrollable aspect of wood research. Cold dehydration using successive baths
Although samples were all cut to the same size of acetone and an applied vacuum to dehy-
for testing (. cm long, . cm wide, and . drate sample C produced some interesting
cm thick), dierences in wood growth, degree results. Analysis of the out-gases indicated a
of waterlogging, fungal and bacterial degra- nearly twofold increase in methyl levels, while
dation, and a host of other factors can expand nitrogen levels for this sample were the low-
the degree of variability among samples. Data est of all samples tested, recording an inten-
from mass spectrographic analysis for each sity of ..
sample, however, answered the basic ques- Sample D, which was dehydrated in a se-
tions of water replacement and mobility of quence of cold alcohol and acetone baths
volatile solvents within the samples. The data under vacuum, registered increased levels of
showed signicant dierences among the rela- methyl and nitrogen (. and . intensities,
tive intensities of elements collected during respectively). Notably, the levels of acetone
experimentation. out-gassing were substantially greater (.
Sample A was dehydrated using a se- intensity).
quence of two room temperature acetone Sample E was warmed during the out-gas-
baths, each lasting for hours. Relative in- sing process. Silicone oil was not used to sup-
tensities of water, nitrogen, methyl, and oxy- port the sample during the gas collection
gen in this sample were the highest for all process. Levels of volatile elements such as
samples tested, with amounts recorded as ., methyl, ethyl, and acetone were predictably
., , and ., respectively. The intensity of higher in sample E, although water out-
 Table . Data from Mass Spectrographic Analysis of Waterlogged Wood

20 Sample Nitrogen Methyl H2O C2H3 Ethyl

A 6.50 25.00 2.50 7.50 2.50
B 3.00 8.00 2.00 13.00 14.00
C 2.50 14.80 2.00 12.00 2.50
D 3.80 19.60 2.00 6.00 2.50
E 4.00 22.50 2.00 8.00 3.80

Sample Oxygen C3H3 Acetone Acetone Acetone

A 2.00 4.50 100.00 2.00 28.00

B 2.00 4.50 100.00 3.00 25.00
C 4.50 7.50 100.00 2.40 24.30
D 0.20 5.00 9.00 100.00 32.00
E 1.80 3.50 12.50 100.00 53.00

gassing from the sample was comparable to gases from samples C and D is interesting
samples B, C, and D. A lack of media to sup- because the higher levels of methyl and ac-
port the degraded cell structure resulted in etone gases suggest that alcohol-acetone
greater damage to this section of waterlogged process dehydration may have been more
wood. Not surprisingly, the levels of out- eective in raising the overall levels of vola-
gassing methyl and ethyl were the highest in tile vapor exchange. Thus this process may
this sample, and molecular acetone levels have a higher potential for exchanging sili-
were % (. intensity) higher than those cone oils, or other bulking agents, into water-
recorded for sample D. logged materials.
With the exception of sample A, which reg-
Trends istered water levels of ., samples registered
water levels of ., which may suggest that,
Sample E, which acted as a control for this from the standpoint of the removal of free
experiment, served to indicate the maximum water from waterlogged wood, all processes
out-gassing potential that might be expected were relatively equal. Dierences in bulking
for this group of wood samples. Because heat potentials, then, lie in the heightened levels
was used to accentuate the eects of out- of volatile gases that result in cold process,
gassing, methyl, ethyl, and acetone levels for vacuum assisted alcohol-acetone dehydra-
the sample were the highest recorded. While tion. Table . includes the out-gassing data
less than the levels recorded for sample E, the for all ve samples. The data suggest that the
levels for sample D were appreciably higher rise in molecular acetone and methyl levels
than out-gassing levels recorded for samples resulting from multiple-solvent dehydration
processed at room temperature. makes two-stage cold dehydration potentially
A comparison of the relative levels of more appropriate for silicone bulking.
 CHAPTER 3
21

Archaeological

Wood

O
ne of the biggest problems conserva of wood from the denser interior. Passiva-
tors may encounter when developing tion Polymers seem to eectively prevent this
a conservation strategy for waterlogged problem.
wood is determining the degree of deteriora- The use of silicone preservation technolo-
tion that has occurred within the matrix of gies also allows the conservator to better con-
the artifact. Whether from a bog, a river, or a trol the strength, exibility, exterior nish,
swamp, a piece of wood will suer structural and color when treating waterlogged or se-
deterioration as a result of bacterial decay and verely desiccated wood. As with polyethylene
fungal and microbial attack. Eects of water glycol, Passivation Polymers can be blended
ow and natural abrasion from sands and silt to accentuate structural and aesthetic quali-
also compromise the physical integrity of ties that the conservator may wish to apply
wood if deposition has occurred in shallow to archaeological wood.
water. Because silicone oils do not mix with wa-
ter, all free-owing water must be removed
to allow impregnation of the wood. This can
The Challenge of Conserving be easily accomplished with solvent baths. To
Waterlogged Wood prevent stress, a series of baths should be
used, starting with alcohol dehydration and
While it is safe to assume that it is easier to working through to two or more acetone
penetrate the matrix of heavily waterlogged baths. For small artifacts, reduced pressure
wood with a bulking agent, to use the same applied to the wood in solution will facili-
bulking material for all artifacts would be tate water removal. Care must be taken in
too simplistic. Indeed, in the case of the large applying a vacuum to acetone as vapors will
oak timbers in the hull of La Belle, the pen- contaminate the vacuum pump oil, and this
etration rate of a bulking agent into the can hinder performance of the equipment.
outer, heavily waterlogged timbers, will be Acetone will also damage a pump with plas-
fast. The inner core of wood, however, is ex- tic components. Older vacuum pumps are
tremely solid, exhibiting nearly the same invaluable in a conservation laboratory be-
strength and density as freshly sawn timbers. cause their all-steel construction is impervi-
Penetration of a bulking agent into these ous to contamination. A Dewar ask with dry
timbers is much more dicult. As formi- ice should always be used to house two in-
dable as the task may be, however, the intro- line gas traps between the pump and the
duction of a stabilizing and consolidating vacuum chamber. This is the safest way to
agent is necessary to prevent exfoliation that prevent acetone vapors from entering the
occurs with the separation of the outer zone pump.
 Degradation and Shrinkage Table . Classications of
Waterlogged Wood
22 After long periods in wet soil, peat bogs, or
marine sites, the cell wall components in all Classification Water
wood will be degraded by bacterial action. 1 More than 400%
In general, water-soluble substances such as 2 185400%
starch and sugar are the rst to be leached 3 Less than 185%
from waterlogged wood, along with mineral
salts, coloring agents, tanning matters, and even % water. Waterlogged wood is of-
other bonding materials. In time, through ten classed according to the amount of water
hydrolysis, cellulose in the cell walls disinte- it contains. The most commonly used clas-
grates, leaving only a lignin network to sup- sications are shown in table ..
port the wood. Even the lignin will break All waterlogged timbers must remain com-
down over a long period of time. pletely wet up to, and throughout, the preser-
As a result of the disintegration of cellu- vation process. Wood recovered from saltwa-
lose and lignin, spaces between the cells and ter environments requires extensive rinsing in
molecules increase, and the wood becomes freshwater to eliminate as many soluble salts
more porous and permeable to water. All the as possible. Failure to eliminate salts in the
deteriorated elements of the wood, includ- matrix of the wood will cause problems once
ing cell cavities and intermolecular spaces, are the artifact is removed from treatment.
lled with water. The remaining lignin struc- Trapped salts will slowly crystallize, resulting
ture of wood cells and the absorbed water in possible distortion and exfoliation of large
act to preserve the shape of the wood. The layers of the artifact. Titration tests should be
loss of the ner cellulose tissue does not cause used to determine the point at which the bulk
much alteration in the gross volume of wood, of soluble salts have been rinsed from the
but the porosity is increased and the wood wood. As the results from the Dowel Experi-
absorbs water like a sponge. ment in chapter suggest, Passivation Poly-
A waterlogged wooden object will retain mers penetrate easily into an acetone-rich
its shape as long as it is kept wet. If the wood environment. Accordingly, prior to treat-
is exposed to air, the excess water evaporates ment, cleaned wooden artifacts should be
and the resulting surface tension forces of stored in an alcohol/acetone (:) environ-
the evaporating water cause the weakened ment. In many cases, artifacts can be stored
cell walls to collapse, creating considerable in % acetonethe longer the better.
shrinkage and distortion. The amount of
shrinkage is dependent upon the degree of
disintegration and the amount of water Waterlogged Wood from Saltwater
present. The amount of water in waterlogged Environments
wood can be determined by the following
equation: This procedure deals with any wood with an
internal structure that has been degraded by
the waterlogging process.
%H2O = weight of wet wood weight of oven dried wood
X 100
weight of oven- dried wood
Pretreatment

Wood containing more than % water is All waterlogged timbers must remain wet
considered degraded; it is not uncommon to prior to treatment. Regardless of whether
nd wood that contains more than % or they were recovered from saline or freshwater
environments, the water in the cell structures Calcareous deposits can often be removed
of these artifacts is preventing collapse and using dental picks or soft wooden dowels.
distortion. Any amount of air-drying or de- Small pneumatic chisels are useful for 23
hydration will cause potentially irrevers- removing slightly larger calcareous deposits.
ible cellular collapse. Care should be taken, however, to ensure that
There are additional challenges in the vibrations from the chisel do not shake the
treatment of waterlogged timbers from salt- wood to the point of causing additional
water environments. With evaporation, salts stresses. To minimize damage caused by
remaining in the wood will crystallize. Ex- vibrations, the artifact should be held in one
pansion and crystallization as the artifact hand or placed on soft vermiculite-lled
dries may result in extensive cracking and, cloth bags while removing concreted
possibly, exfoliation of large sections of wood. materials.
Even trace amounts of salts can damage . To enable deep penetration of polymers
structurally deteriorated wood. Because they throughout the wood, water must rst be
remain soluble, salts within the cell structure eliminated. For small artifacts, this can be
of the artifact act to absorb atmospheric accomplished using a series of dehydration
moisture and are not completely stabilized. baths, starting with ethanol and working
Soluble salts and minerals must, therefore, through to several baths of fresh acetone.
be ushed from waterlogged wood prior Water/acetone exchange will take several
to treatment. The best means of removing days, depending upon the artifacts size and
soluble salts is through freshwater rinses and, density. This phase of treatment cannot be
when possible, running water rinses. Over hurried. Incomplete dehydration or dehydra-
time, salts will dissolve into and be diluted in tion carried out too quickly may damage the
the volume of water in which the wood is artifact.
stored. Accordingly, large volumes of water Vacuum-assisted dehydration for the nal
are best for long-term storage. Water should acetone baths is often an eective way to
be changed frequently. Titration readings and ensure complete dehydration. During
simple silver nitrate tests can be used to de- dehydration, the vat containing the artifact in
termine parts per million (ppm) salt content fresh acetone is placed into a vacuum
prior to changing rinse water. chamber. Two gas traps, immersed in dry ice
Vats containing waterlogged wood should or liquid nitrogen, should be placed in-line
be kept covered. Exposure to sunlight will between the vacuum chamber and the
encourage the growth of algae, which may vacuum pump. Escaping acetone fumes and
be detrimental to the wood. Covering the vats moisture in the wood will be condensed and
will also help eliminate deposition of air- trapped in the gas traps, preventing contami-
borne contaminants, as well as mosquito and nation and possible deterioration of the
other insect larvae. How biocides aect the internal components of the pump. Generally,
bulking agents used to stabilize waterlogged reduced pressure using a slight vacuum of
wood or the cell wall structures of the wood . Pa ( Torr) is sucient to ensure
itself is not known. Accordingly, biocides thorough dehydration. Once dehydration is
should be used sparingly. complete, the artifact must be transferred to
the polymer solution while thoroughly wet
Methodology with acetone.
. A suitable volume of Passivation
. The surfaces of the wood should be Polymers (silicone oil mixed with a cross-
thoroughly cleaned using soft brushes, lint- linker) should be prepared in advance to
free cloths, and other appropriate tools. ensure quick transfer after dehydration. A
 Passivation Polymer can be selected from immersed in the polymer solution during
table .. Because waterlogged wood is fragile, treatment. Throughout acetone/polymer
24 care must be taken to ensure that the chosen solution displacement, the artifact may have a
polymer/cross-linker solution is not so tendency to oat as acetone vapors escape
viscous as to prevent rapid acetone/polymer from the wood. It may be necessary to place
exchange. The use of viscous polymer aluminum screen and sucient weight on
solutions will have the same eect as using top of the screen to keep the wood sub-
too high an initial percentage mixture of merged (g. .).
polyethylene glycol in water. If the chosen . To ensure the role displacement of
bulking agent cannot readily displace acetone acetone with a silicone oil solution, artifacts
in the cell structure of the wood, cellular the size of a sewing thimble may require
collapse and distortion of the artifact may treatment for one week at ambient pressure.
occur. A certain degree of experience is needed to
For wood that has a soft exterior zone determine the best treatment strategy for
surrounding a harder and possibly structur- small artifacts. If the artifact is in but not
ally sound core, it may be necessary to use a severly damaged by the waterlogging process,
blend of Passivation Polymers. Lower vacuum pressure may be used to assist the
viscosity Passivation Polymers will penetrate acetone/silicone oil solution displacement
the tight, solid matrix of less structurally process. If however, the artifact is more
damaged core wood, while more viscous severely damaged, it may be necessary to
Passivation Polymers will easily penetrate the conduct the acetone/silicone oil solution
softer exterior. Because higher viscosity displacement process at ambient pressure. At
polymers, such as PR- and PR-, are longer ambient pressure, acetone will vaporize more
polymer chains, they are excellent consolida- slowly from the cells of the wood, allowing
tors, binding the outer soft areas of water- the more viscous polymer solution slowly
logged wood to the more solid core. If the into the cells.
PEG is too viscous, cell walls will be collapsed For larger artifacts such as musket stocks
as the solution tries to enter the cells. and tool handles the conservator may choose
Similarly, an overly viscous solution of to use vacuum-assisted acetone/polymer
silicone oil will cause cell collapse. Acetone/ displacement. Low viscosity polymers must
silicone oils solution displacement will have be used for low-pressure treatment of wood.
the same damaging eects as PEG if the Accordingly, PR- is generally the best
solution tries to displace acetone, which is in choice. To a given volume of PR- polymer,
this case dispersed rapidly as water. use a % (by weight) addition of CR- and
Once a suitable volume of polymers or mix thoroughly. The artifact should then be
blended polymers, sucient for total carefully transferred and totally immersed in
immersion of the artifact, has been selected, a the vat of polymer solution. Aluminum mesh
cross-linker must be added. CR- is the best or some other mechanism is required to
choice of cross-linking agents to use with prevent the artifact from oating during
PR-, PR-, and PR- for preserving acetone/polymer displacement. Once the
waterlogged wood. A % (by weight) addition artifact is secure in the solution, place the vat
of CR- should be mixed in with the into a vacuum chamber. A slight reduction in
polymers. pressure is sucient to cause small bubbles to
. After thoroughly mixing the polymer/ ow from the artifact. When slight bubbling
cross-linker solution, quickly and carefully occurs, lock o the vacuum chamber and
transfer the acetone-laden wood to the turn o the pump. Never allow extensive
polymer solution. The artifact must remain bubbling to occur as this means acetone is
probably being driven from the wood faster drain until its surfaces are nearly dry to the
than the polymer solution can displace it. touch. This may take one or two days. To
Vacuum-assisted treatments work best when remove surface-pooled polymers from the 25
pressure is reduced slowly over several days. wood, use lint-free cloths and soft brushes. A
Generally, a reduced pressure of . Pa Q-tip dipped in CR- is often an ideal tool.
( Torr) is sucient for the treatment of any To remove excess amounts of bulking agent
small wooden artifacts if the conservator has from cracks and crevices, it may be necessary
elected to use a reduced pressure environ- to immerse the artifact in a small vat of
ment to speed up the acetone/polymer CR- and then wipe it with soft cloths and
displacement process. brushes.
Experience and good observations dictate . A list of catalysts is supplied in table .,
the amount of vacuum pressure that can be which also lists the attributes of each catalyst
applied to an artifact. If it is obvious that the and its intended end result. The best general-
diagnostic attributes of the artifact are being purpose catalyst is CT-, which has a working
maintained, it can be held at pressure in time of hours in a closed environment.
solution from to hours. With badly Catalysts can be applied either topically or as
degraded wood, acetone/polymer solution vapors to the surfaces of artifacts. In either
exchange is usually fast since the woods case, an applied catalyst will continue working
deteriorated structure allows rapid and safe to polymerize the polymer-impregnated
exchange of acetone for the more viscous artifact.
polymer solution. For less degraded wood, Topical applications are eective ways to
exchange of acetone for polymers must be initiate polymerization and attain total
conducted at a slower rate to prevent distor- stabilization of the artifact. Use a soft brush
tion and cellular collapse. or cloth to apply a thin, even layer of catalyst
. After bubbling has ceased, the artifact to the artifacts surfaces. After allowing the
should remain in a reduced pressure environ- catalyst to coat the surfaces of the wood for
ment for several hours, if not days. Care two or three minutes, use soft cloths to wipe
should be taken to ensure that the reduced o the solution. Place the artifact into a
pressure in the vacuum chamber is returned Ziploc bag and seal it shut. After hours,
to ambient pressure slowly. After the artifact remove the artifact and inspect it to ensure
has been allowed to sit in solution at ambient that it is dry to the touch. If it is not dry, it
pressure for at least hours, it can be may be necessary to return it to the Ziploc
removed from the Passivation Polymer bag, along with a small piece of cloth
solution. The artifact should then be carefully containing several drops of fresh catalyst.
placed on a screen suspended over a clean Seal the bag again and allow fresh catalyst
container to allow the polymer solution fumes to complete the catalyzation process.
owing from its surfaces to collect. Recovered A more aggressive form of vapor catalyst
solution can be reused almost indenitely, as deposition can be accomplished by placing
long as it is periodically placed for hour in the artifact into a Ziploc bag or plastic
a warming oven set at C. This will ensure container, along with a dish containing
that remaining acetone and water are several grams of CT- catalyst (g. .).
evaporated from the polymer solution. With Seal the container and place it in a vented
each reuse, a new % addition of CR- is warming oven set at C. Warming the
required. catalyst will accelerate its reactivity with the
The conservator need not be concerned polymer solution in the artifact. Generally,
about the safety of the wood while it is the artifact will be completely stable after
draining. It is best, in fact, to let the wood hours. When in doubt, you can store wood in
 second day. To complete the desalination
process, the buttons were rinsed in ve baths
Fig. .. Artifact in bulking
26 of deionized (DI) water, which were also
agent with friction-t
changed every second day. Titration testing
screen in place: (A) open
of the rinse water was conducted periodi-
container; (B) sucient

bulking agent to submerge cally throughout the desalination process.

artifact; (C) friction-t Generally, a level of ppm or lower is con-
screen to ensure immer- sidered adequate for starting the next phase
sion during treatment; of treatment.
(D) artifact. Dehydration of the buttons was con-
ducted using a series of organic solvent baths.
a Ziploc bag for several days, as long as fresh Starting with a solution of % ethanol in
catalyst is added daily. After catalyzation, DI water, each successive bath included a %
allow the artifact to sit in a fume hood for increase in ethanol until the buttons were
one or two days. The smell of the catalyst will dehydrated in % ethanol. Each phase of
dissipate quickly. alcohol dehydration lasted days. The sol-
vent baths continued, with each successive
bath receiving a % increment of acetone.
Case Study: Waterlogged Wooden Each bath lasted days. The nal stages of
Buttons with and without Associated dehydration included three successive baths
Thread of % acetone, each lasting days.
A solution of PR- with % CR- added
After their recovery from the shipwreck La by weight, sucient to immerse the buttons,
Belle, wooden buttons with and without as- was prepared. To ensure that the buttons and
sociated thread were stored in freshwater. attached threads did not oat during treat-
Initially tap water, with a chloride level of ment, small sections of aluminum window
approximately ppm, was used to desali- screen were friction-t over them in the con-
nate the buttons. Over the course of one tainer (g. .). During acetone/polymer dis-
Fig. .. The artifact in a month, the bath water was changed every placement, the solvent vaporizes and escapes
Ziploc bag with a dish two days. For the next two weeks, the but- from the wood, creating voids; continuous
containing several grams tons were rinsed in baths of rainwater, which immersion of the buttons is therefore needed
of CT- catalyst. like the tap water baths were changed every to ensure that these voids are rapidly lled
with polymer solution. The buttons re-
mained immersed in the polymer solution
for days.
After days, the buttons were removed
from the solution and placed onto a screen
so that runo polymers could be collected
for reuse (g. .). Each button was then in-
dividually wiped, rst with paper towels, to
remove the bulk of polymers, then with soft
cloths to remove any remaining solution. As
each button was dried, it was immediately
dropped into a bath of % CR- for a
period of no longer than two minutes so that
additional polymers could be washed from
its surfaces. Each button was then removed
from the CR- and carefully surface-dried, need for additional cleaning. After hours
taking care not to dislodge any associated of vapor deposition catalyzation, the buttons
thread. felt both dry and nonwaxy. They were then 27
The buttons were then placed on an el- removed from the Ziploc bag and placed on
evated window-screening tray in a Ziploc a tray for continued curing in a fume hood.
bag. A paper towel was crumpled into a ball The nal appraisal of the treatment pro-
and placed into the bag, along with the arti- cess for each button was subjective. If the
facts. The towel served to form an airspace buttons had felt waxy or had appeared to
around the artifacts. A second piece of be wet, they could have been immersed in
crumpled paper, containing cc of CT-, CR- to surface-clean the slight polymer
was then placed into the bag (g. .). Once lm that may have formed during treatment.
sealed, the bag was placed into a warming All the buttons with attached threads were
oven preset at C. After hours of dipped in CR- and then carefully surface-
catalyzation, the paper containing the used dried. This made the threads more exible.
catalyst was replaced with a new ball of pa-
per containing cc of fresh CT-. The bag
was resealed and returned to the warming
oven for an additional hours.
Care must be taken to ensure that paper
containing the CT- catalyst does not come
into direct contact with any artifacts during
catalyzation. Direct contact with the catalyst
will result in the formation of white particu-
late at the point of contact. This is polymer-
ized silicone oil, which can easily be removed
using a soft cloth once treatment is complete.
Avoiding contact, however, will eliminate the
Fig. .. Buttons draining
on a screen after polymer/
cross-linker treatment.

Fig. .. Button immersed in polymer/cross-

linker solution: (A) beaker; (B) polymer/
cross-linker solution; (C) aluminum screen; Fig. .. Buttons being prepared for the vapor
(D) button. catalyst deposition process.
 Inspection revealed the individual strands of lignum, starches, polysaccharides, poly-
the threads were well preserved and not phenolics, and water within the matrix of an
28 clumped together. Any threads that appeared artifact are lost through desiccation, leach-
to be slightly clumped were easily separated ing, oxidation, or attack by fungus and dry
while pat-drying the CR-. Over time, the rot. Excessive handling in this compromised
threads have not become sti and have de- state is dicult, and achieving long-term sta-
lighted conservators visiting the lab (g. .). bility may be impossible.
The results of Passivation Polymer treat- The use of polymers for the treatment of
ment of wooden buttons, with and without dried wood is a simple process since dehy-
associated threads, have been very successful. dration is seldom required. Completely
The artifacts retain their diagnostic features eliminating water from wood prior to treat-
with only ..% shrinkage of overall size. ment using Passivation Polymers is not nec-
The associated threads have retained their essary. Artifacts that have been exposed to a
color, yarn memory, exibility, texture, and dry environment or that have been stored
some elasticity. Over time, the artifacts have in an environment of less than % humid-
shown no change in their original posttreat- ity seldom need posttreatment dehydration.
ment appearance or condition. For small artifacts, such as buttons, the fol-
lowing processes are eective conservation
strategies.
Dry-Site ArtifactsDry and
Desiccated Wood Methodology

This procedure deals with weathered wood Where possible, the surfaces of all artifacts
and architectural materials, and with organic should be cleaned of loose debris. Most of
artifacts from dry cave sites. the time, this can be accomplished with a soft
brush or lint-free cloths. In situations where
Pretreatment dirt is heavily impacted on an artifact, how-
ever, it may be necessary to dampen the sur-
Usually, dry and desiccated wood samples faces before attempting any cleaning. After
recovered from land sites are fragile and light cleaning, sucient time should be allowed
in weight due to water loss. Over time, the for the artifact to dry before continuing treat-
Fig. .. Buttons after ment.
completion of the While the artifact is air-drying, prepare a
Passivation Polymer suitable volume of silicone oil/cross-linker
treatment. solution to ensure total immersion of the as-
semblage throughout treatment. Be sure to
consider the potential volume of polymer
solution that will be absorbed during treat-
ment. Because dry and desiccated wood tends
to be friable, choosing a suitable polymer/
cross-linker combination is important. Ab-
sorption of the polymer will be rapid in dry
wood. Choosing a short-backbone polymer,
which is also a low viscosity polymer, will
accomplish two things. The low viscosity of
the solution will enable deep and thorough
penetration of the wood; and once polymer-
ized, the smaller polymer will ensure that the observations dictate the amount of vacuum
wood is stronger than in its pretreatment pressure that can be applied. If the diagnostic
state. attributes of the artifact are not being 29
For dry cave-site wood, PR- with a % compromised, it can remain in reduced
addition of CR- cross-linker, added by pressure treatment for several hours. Because
weight, is a perfect combination of polymers of their porosity, dried wooden artifacts
for treatment. Other polymer choices are averaging in thickness up to . inches are
listed in table .. well penetrated after this type of processing.
. If vacuum-assisted impregnation has
. Pour sucient PR- into a container. been used, always make sure the artifact is
To this solution, add % CR- cross-linker returned to ambient pressure slowly. The
by weight and mix thoroughly. Always use a artifact should then be taken out of the
nonporous stir rod when mixing polymer vacuum chamber and, after careful removal
solutions. This will ensure that equal propor- from the Passivation Polymers solution,
tions and stoichiometry are maintained. placed onto an aluminum screen so that free-
Unless thoroughly mixed, the less viscous owing materials can be recovered. There is
material, in this case CR-, could be no hurry to continue treatment on desiccated
absorbed by a porous stir rod, resulting in wood. The artifact can safely drain for several
an improperly prepared polymer solution. hours and, if necessary, be stored indenitely
. Carefully place the artifact into the in solution until continued treatment can
PR-/CR- solution. Regardless of which occur.
viscosity polymer has been chosen, be careful . Before a catalyst is applied, excess
when immersing an organic substance. bulking agent needs to be removed from the
Delicate artifacts can be damaged if forced artifacts surfaces. While many of the polymer
recklessly into a viscous bulking agent. Place solutions are sticky to the touch, they are
a piece of aluminum screen over the artifact, easily wiped from the wood surfaces using
making sure that the screen friction-ts Q-tips or clean cloths. To remove excess
against the sides of the container. This will amounts of bulking agent from cracks and
ensure that the artifact remains immersed in crevices, the artifact can be wiped with a rag
the polymer solution throughout treatment. soaked in CR-. For complex surface-
(See g. ..) cleaning, or when all surface traces of bulking
. Dry wood needs little help to absorb a solution need to be removed, the artifact can
low viscosity polymer solution. If the be immersed in a bath of CR- then wiped
conservator feels the need for pressure with Q-tips and cloths.
treatment, proceed with caution. Very little . After the artifact has been wiped, a
vacuum will be required. During vacuum- suitable catalyst should be applied to all
assisted impregnation, use as little vacuum as surfaces as soon as possible. A list of catalysts
possible. As long as small bubbles ow slowly is supplied in table .. Catalysts attributes
from the wood, there is sucient pressure to and application methods are also provided in
ensure that the polymer will penetrate the that table. CT- is the best general-purpose
artifact. If bubbling is too vigorous, reduce catalyst for treating dried wood. After
the vacuum pressure immediately. Excessive applying CT- and allowing it to sit on the
bubbling may damage delicate artifacts. Once surfaces of the wood for no more than two
bubbling appears to slow in frequency, or minutes, wipe thoroughly and quickly. Using
stop, increase the vacuum pressure and the same procedures as noted for waterlogged
maintain pressure at a point where stream- wood, continue vapor deposition by placing
type bubbling occurs. Experience and good the artifact into a Ziploc bag. Place a small
 dish containing a few grams of CT- into the conservation strategies was still in its infancy,
bag and seal it shut. The catalyzation process have not fared as well.
30 will continue for up to hours. PEG was used in the treatment of wood,
If the surfaces of the wood feel damp, re- leather, paper, and other organic materials,
place the used catalyst with a dish contain- and was added to the water in which the ar-
ing fresh catalyst. CT- has a maximum tifacts were stored. Over time, the percent-
working time of hours. Once the artifact age of PEG was increased and the aqueous
feels dry, remove it from the Ziploc bag and solution permeated into the cell structure of
allow it to cure in fresh air for as long as pos- the artifact, depositing the carbowax bulk-
sible. While the artifact can remain in cataly- ing agent, which prevented cellular collapse.
zation as long as desired, the nal stage of Because this preservation process was still in
fresh air curing will allow nal curing and its developmental stages, conservators tended
ensure that catalyst fumes dissipate. If the to use diering percentages of PEG in treat-
wood requires spot treatment to remove ment, and in many cases, variance in post-
white particulate formed on the surfaces dur- treatment handling and curation of artifacts
ing catalyzation, use a soft cloth containing was substantial. Indeed, the pioneers of PEG
a few drops of CR-. conservation used freeze-drying in conjunc-
tion with PEG treatments. This combined
treatment strategy yielded superior nished
Reprocessing and Stabilization of artifacts that were more stable and aestheti-
PEG-Treated Wood cally correct.
But what about the hundreds of thou-
Re-treatment strategies have been of particu- sands of artifacts that were conserved with-
lar interest to conservators at the Archaeologi- out the benet of advancements such as
cal Preservation Research Laboratory. For sev- freeze-drying and blended PEG treatment
eral decades, preservation of waterlogged strategies? Some artifacts have been lost. The
wood using polyethylene glycol has been the majority of artifacts, however, have been care-
standard treatment preferred by most conser- fully curated and re-treated over time. Early
vators. A carbowax that is miscible in water, in Ole Crumlin-Pedersen of the Danish
PEG is available in an almost innite variety Viking Ship Museum traveled to Texas A&M
of molecular weights. A skillful conservator University to discuss the vessels in the mu-
can blend two or more molecular weights seum that had been preserved using PEG. He
of the bulking agent to attain specic and believed, as do I, that vessels preserved using
desirable traits for organic artifact preser- traditional PEG treatment methods require
vation. Like all conservation strategies, PEG additional treatment to make them as stable
treatments are always being rened; recent as possible. At Crumlin-Pedersens encour-
advances in archaeological chemistry have agement, re-treatment of PEG-treated wood
expanded our knowledge of PEG chemistry became a new challenge for conservators at
and its applications in archaeology. Texas A&M University. The purpose of early
Re-treatment of PEG-treated artifacts is a experimentation was not to discredit the
concept that has grown out of necessity. value of PEG in the preservation of water-
When preserved using standardized treat- logged wood. On the contrary, research ini-
ment strategies developed by Per Homann, tiatives were directed at improving and
David Grattan, and others in the s and strengthening the knowledge base for using
s, artifacts treated with PEG are stable PEG in archaeological preservation. While
and well preserved. Artifacts treated in the PEG is by far the most widely used bulking
mid-twentieth century, when development of agent for the preservation of waterlogged
wood, its basic characteristics oer some chal- a saturated solution of acetone rosin, which
lenges for the long-term well-being of pre- achieves a result similar to PVA emulsion.
served wood. In the Tongue Depressor Experiment, 31
The posttreatment hygroscopic nature of described below, CR-, a hydrolyzable, mul-
PEG is a concern when the carbowax com- tifunctional alkoxysilane was used to expe-
pound is used for the long-term curation of dite the removal of unbound PEG from the
structurally damaged, waterlogged wooden wood. CR- self-condenses in the aqueous
artifacts. When introduced into waterlogged environment of the wood, forming a methyl
wood, some of the PEG reacts with polysac- silane polymer that is less susceptible to wa-
charides, polyphenolics, and water to form ter absorption. Results of this experiment
complex bonds that are hydrophobic in na- have shown that while PEG is an excellent
ture. The unbound PEG, located within the bulking agent for archaeological preserva-
open structures of the wood, however, re- tion, it is not a reversible process. When un-
mains hydrophilic. Accordingly, uctuations bound PEG is removed from wood in the
in temperature and humidity result in the presence of an alkoxysilane solution, the re-
uptake of atmospheric moisture and inter- maining PEG is hydrophobic and less prone
cellular migration of the unbound PEG. In to dimensional shifting as the result of envi-
extreme conditions, the intercellular move- ronmental factors. The use of an alkoxysilane
ment of PEG stresses the weak walls of the polymer to reprocess PEG-treated wood
wood, causing shrinkage, distortion, and loss oers many advantages for the stabilization
of diagnostic artifact attributes. of waterlogged wooden artifacts. The process
Damage caused by waterlogging is sel- reduces, and sometimes eliminates, swelling
dom uniform throughout a timber. Often, that has occurred from introducing PEG into
the inner core remains in excellent condi- the waterlogged wood. After treatment, the
tion while its outer surfaces become soft and surfaces of the wood are stabilized and do
severely damaged. PEG treatments success- not require additional consolidation. All the
fully displace water in the cell structure of wood samples regained a natural wood tex-
waterlogged wood, but the bulking agent ture and color, and in all cases, the weights
alone is usually not capable of stabilizing and of the reprocessed samples were reduced.
protecting the surfaces. To stabilize badly de-
teriorated or surface-checked wood, an ap-
plication of higher molecular weight PEG Tongue Depressor Experiment
may act to consolidate the surfaces of the
artifact from abrasive damage caused by dust Nuclear magnetic resonance (NMR) analy-
and other airborne particles in the museum sis was used to investigate the chemical trans-
environment. formations that occurred when PEG was
At the conservation facilities of Parks introduced into samples of waterlogged
Canada in Ottawa, Canada, - Jade ad- wood and to characterize chemical reactions
hesive, a diluted polyvinyl acetate (PVA) resulting from the reprocessing of this wood
emulsion, is used to consolidate the checked using CR-. During reprocessing in a vat of
surfaces of PEG-treated wood.1 The substance CR-, unbound, free-owing PEG was re-
also strengthens the surfaces of PEG-treated moved from the wood. In the aqueous envi-
timbers.2 To ensure that badly deteriorated ronment of the PEG-treated wood, CR-
wooden artifacts have resilient outer surfaces, displaced free-owing PEG within the cell
conservators at the Archaeological Preserva- voids of the wood, hydrolyzing to form me-
tion Research Laboratory at Texas A&M Uni- thyl silane polymers. Evaluation of the repro-
versity have immersed PEG-treated wood in cessed wood indicated that swelling normally
 associated with traditional PEG preservation surfaces of the silicone oilimpregnated ar-
was eliminated. In many cases, samples of tifact or as a vapor. Vapor deposition accel-
32 wood were returned to their pretreatment erates the chemical reactivity of the catalyst
dimensions. Unlike PEG-treated wood and ensures an even application to all sur-
samples, which became darker in color and faces of the artifact.
waxy in surface appearance, the reprocessed Other materials included a Pyrex gradu-
wood appeared natural in coloration and ated cylinder ( ml), a piece of aluminum
surface texture. foil to form a loose-tting cap for the gradu-
ated cylinder, and a Despatch LFD series
Materials warming oven with computerized tempera-
ture and humidity controls.
Tongue depressors were chosen for experi-
mentation because they are plentiful and Methodology
uniform in dimensions, color, genus, and
species. Several control tongue depressors Twenty waterlogged tongue depressors were
were randomly collected to assist in evaluat- randomly selected, thoroughly rinsed in run-
ing PEG-treated wood. To create a supply of ning tap water, and stored in a large jar of
waterlogged wood, hundreds of white birch freshwater. The jar was then placed into a
(Betula papyrifera) tongue depressors were vented warming oven set at a constant tem-
placed into a four-liter stainless steel con- perature of C. Small amounts of water
tainer and boiled in freshwater for days. were added to the container each day to
Water was added at regular intervals to en- maintain a constant level throughout treat-
sure that the water level remained high ment. Each week, one % increment of
enough to keep all the wood pieces saturated. warmed PEG (by volume) was added,
The tongue depressors were then stored forming an aqueous solution of PEG. After
at room temperature in a large jar of water seven weeks, the warming oven was turned
with a tight-tting screw cap. One tongue o; the tongue depressors were allowed to sit
depressor, labeled TD, was stored separately in the PEG solution for another week.
in freshwater for comparison with the con- One tongue depressor, labeled TD, was
trol sample, TD, and the PEG re-treated removed from the PEG solution, placed be-
wood, TD. Several pieces of wood were al- tween paper towels to blot free-owing PEG
lowed to air-dry at room temperature for from its surfaces, and allowed to sit in fresh
hours. After air-drying, all samples exhibited air for hours before additional experimen-
extensive warping and shrinking. tation was conducted. The length, width,
At room temperature, PEG is a pow- thickness, and weight of TD were then re-
der. To make the task of incremental addi- corded, and NMR analysis was performed.
tions easier, a large jar of PEG was placed (The waterlogged sample, TD, and the un-
into a warming oven and warmed to a tem- treated sample, TD, were also subjected to
perature of C. At this temperature, it re- NMR analysis.) TD was then immersed in
mains in a liquid state. CR- is a clear mono- ml of CR- in a graduated cylinder. To
mer that reacts with water to form silanetriol prevent rapid evaporation of the silane, the
and methanol. The silane condenses with graduated cylinder was loosely capped with
available hydroxyl groups or other silanol a three-inch square of aluminum foil. The
monomers to form siloxane resins. CT- is cylinder was placed into a C vented warm-
a tin- based catalyst that remains active for ing oven for hours (g. .). The cylinder
approximately hours in a controlled envi- was then removed from the oven and allowed
ronment. It can be applied topically to the to cool to room temperature and sit for an
additional hours. When TD was removed
from the CR- solution, it was covered with
a thin white coating of PEG. This was re- 33
moved by wiping the tongue depressor with
a lint- free cloth. The remaining CR- solu-
tion in the graduated cylinder was milky
white due to PEG extracted from the tongue
Fig. .. Graduated
depressor. This solution was poured into a
cylinder setup for repro-
at pan and allowed to sit in a fume hood cessing of PEG-treated
for several days. wood using CR-:
Vapor deposition was chosen as the best (A) vented warming oven;
method to ensure that all surfaces of the (B) loose-tting aluminum
wood were evenly exposed to catalyst. To cap; (C) level of CR- in
form an eective environment in which cylinder; (D) CR-
warmed vapor fumes would remain in close solution in which TD was
contact with the surfaces of TD, a tall poly- immersed; (E) TD
ethylene container with a tight-tting lid was immersed in CR-.
used to create a containment chamber. In an
inverted position, the lid of the container
became the base of the unit. to temperatures. When immersed in warm
A Fisherbrand aluminum weighing dish CR-, PEG melts. Because it is not chemi-
with g of DBTDA catalyst was placed in cally bound to the cell walls of the wood, it
the middle of the base of the chamber. A piece migrates. This is the same reaction that would
of aluminum screen was placed over the dish be observed if the PEG-treated wood were
to act as a platform on which the tongue de- placed into warm water.
pressor could be placed during the cataly- Of interest, then, is the chemical reactiv-
zation process. Paper towels were placed on ity of CR- within the cell structure of the
top of the aluminum screen to prevent free- wood. PEG was introduced into the wood as
owing PEG/CR- solution from contami- an aqueous solution. Accordingly, water was Fig. .. Conguration of
nating the catalyst during treatment (g. .). bound to the PEG. It appears that H2O asso- the containment chamber
When the body of the chamber was snapped used for vapor deposition:
into place over the tight-tting base, an ideal (A) warming oven;
environment for vapor deposition of the (B) upper body of
catalyst was created. TD was left in catalyst polyethylene chamber,
vapor deposition for hours in a warming with a small hole in its
oven set at C. It was then removed from top; (C) CT- vapor
the oven and exposed to fresh air in a fume created by warming the
hood. After hours in air, TD was evalu- catalyst to _C; (D) TD
sitting on top of the
ated and samples of wood were collected for
source of catalyst;
NMR analysis.
(E) aluminum screen and
NMR spectrographic analysis showed that
paper towel, creating the
CR- was not polymerizing PEG within the platform on which TD
cell voids of the wood, as had been expected. sits; (F) aluminum
Instead, the warm solution of CR- was re- weighing dish holding g
moving the unbound PEG solution from the of CT- catalyst;
cells. PEG remains water miscible within the (G) tight-tting base of
cells of wood. It, therefore, remains sensitive the containment chamber.
 ciated with PEG had reacted with CR-, act- The preparation of TD by boiling in water
ing as part of the self-condensing reaction has been shown to alter the chemical consti-
34 that had formed new polymers within the cell tution of the tongue depressor, as evidenced
structure of the wooden tongue depressor. by the loss of the acetate resonances at and
Further NMR analysis was conducted to un- ppm in comparison with the spectrum
derstand this chemical reaction. on the control (TD). The changes in these
13
C spectra of the two birch wood tongue
1
H, 13C, and 29Si Nuclear Magnetic depressors are quite similar to those reported
Resonance Characterization by Wilson et al. for oak woods taken from
shipwrecks.4 The 13C spectral signature, along
The solid state NMR spectra were acquired with the macroscopic observations of exten-
with a Bruker MSL spectrometer oper- sive warping and shrinking upon drying in
ating at a magnetic eld of . Tesla. The air, suggest that this preparation provides a
high resolution solid state 1H (CRAMPS) suitable model for the analysis of waterlogged
spectra were acquired with an MREV- pulse wood.
sequence using a microsecond -degree The 13C CP/MAS spectrum of the water-
pulse and a tau value of microseconds. The logged tongue depressor (g. ., top) is nearly
1
H-13C cross-polarization (CP/MAS) spectra identical to that of the waterlogged tongue
with high power H decoupling and magic depressor (TD) after treatment with PEG
angle spinning were acquired with a -de- (g. ., bottom). The spectral dierence
gree pulse of microseconds, a contact time for the waterlogged tongue depressor treated
of . milliseconds, and a recycle delay of with PEG shows up in the 13C Bloch de-
seconds. cay spectrum (g. ., top). A large resonance
Additionally, 13C spectra (Bloch decay) at . ppm is present, in good agreement
were acquired with a -degree excitation with the sharp single resonance of . ppm
pulse with high power 1H decoupling and observed in the 13C solution state spectrum
magic angle spinning using a recycle delay for an aqueous solution of PEG and with
of seconds. The 29Si spectra (Bloch decay) the broader 13C resonance at ppm for PEG
were acquired using a -degree excitation in the solid state.
pulse with high power 1H decoupling and The absence of signal or small signal in-
magic angle spinning using a recycle delay tensity observed under cross-polarization is
of seconds. consistent with signicant molecular mobil-
The 13C CP/MAS spectrum for the un- ity of the PEG . Such mobility would at-
treated control tongue depressor (TD) is tenuate or average the 1H-13C dipolar inter-
shown at the bottom of gure .. The spec- action used for magnetization transfer in the
trum is consistent with a composition of experiment. This observation of the 13C reso-
cellulose, hemicellulose, and lignin.3 Both cel- nance of the PEG in the Bloch decay
lulose and hemicellulose are carbohydrates. spectrum and not in the cross-polarization
The 13C resonances of cellulose are between spectrum indicates that the PEG is still
and ppm. Hemicellulose contains ac- a very molecularly mobile species, even when
etate functional groups with resonances at absorbed within the wood. While molecular
and ppm. Lignin is a polymer of phenolic mobility sucient to attenuate the dipolar
cinnamyl alcohols. No signicant spectral interaction necessary for cross-polarization
dierences were noted in comparison with need not include translational mobility, only
the Bloch decay spectrum. The 13C CP/MAS slight temperature increases above ambient
spectrum for the waterlogged tongue depres- are necessary for the PEG to migrate out
sor (TD) is shown at the top of gure .. of the wood.
 TD was then treated with CR-. The 13C bond on the silicon sites with a methyl and
CP/MAS spectrum of the waterlogged two siloxy groups. The presence of hydroxyl
tongue depressor treated with both PEG groups as this fourth bond is conrmed in 35
and CR- (g. ., bottom) is nearly identi- the solid state 1H spectrum of the polymer
cal with that of the waterlogged tongue de- (g. .). In this spectrum, the peak at -.
pressor (TD) shown in gure . (top). The ppm shows the methyl protons on the silicon
dierence is a new 13C resonance, which ap- while the smaller peak at . shows the hy-
pears at -. ppm. It should be noted that droxyl protons on the silicon. The structure
the 13C Bloch decay spectrum (not shown)
of this sample still indicates the presence of
some PEG . Inside the tongue depres-
sor, a condensation reaction has occurred.
The CR- appears to have been hydrolyzed
in the aqueous environment to the triol,
giving o methanol. The triol then self-
condensed to form a silicone polymer. The
13
C CP/MAS spectrum of the solid polymer
formed under identical conditions in the
absence of the wood is shown in the top of
gure .. The major resonance is at -. ppm
with small resonances at . and . ppm.
In aqueous solution, the 13C resonance of
PEG is at . ppm.
Though CR- is not miscible in aqueous
solution, the 13C solution state spectrum ac-
quired in deuterated chloroform shows a
Fig. .. 13C CP/MAS spectra of the control tongue depressor TD (bottom)
methoxy (SiOCH3) resonance at . ppm
and the waterlogged tongue depressor TD (top).
and a methyl (SiCH3) resonance at .
ppm. This indicates that only small amounts
of PEG and methoxy groups are present with
a change in the methyl group resonance as
the polymer is formed.
No 29Si resonances were observed for the
control tongue depressor (TD). However,
gure . shows the 29Si spectra of the wa-
terlogged tongue depressor (TD) treated
with both PEG and CR- (bottom) and
of the solid polymer formed from the CR-
(top). The spectra are consistent with two types
of silicon environments, one with silicons
bearing two siloxy {H3CSi(OSiO)2OR} groups
( ppm) and one with silicons bearing
three siloxy groups ( ppm).
Since the 13C spectrum of the solid poly-
mer formed from the CR- (g. ., top)
showed only a trace of a methoxy resonance, Fig. .. 13C CP/MAS (bottom) and Bloch decay (top) spectra of the water-
it is clear that the methoxy is not the fourth logged tongue depressor TD treated with PEG .
36

Fig. .. 13C CP/MAS

spectra of the waterlogged
tongue depressor TD
treated with both PEG
and CR- (bottom)
and of the solid polymer
formed from the CR-
(top).

Fig. .. 29Si/MAS spectra

of the waterlogged tongue
depressor TD treated
with both PEG and
CR- (bottom) and of the
solid polymer formed
from the CR- (top).
 37

Fig. .. 1H CRAMPS
spectrum of the solid
polymer formed from
the CR-.

of the silicone polymer formed in the absence While the length of the wood remained un-
of the wood as well as within the wood is changed at . cm, its center width swelled
shown in gure .. to . cm, a .% increase. After removal
of PEG from the wood and treatment in CR-
Observations , swelling was almost completely elimi-
nated from TD. The length of TD re-
The PEG-treated tongue depressor TD was mained constant at . cm, and the width
a translucent, strong brown color (Munsell was reduced to . cm, a decrease of .%.
.YR .), and the surface texture of the The weight of TD, the control tongue de-
wood was waxy and slightly sticky. After treat- pressor, was . g. After treatment in PEG,
ment in CR-, however, the wood changed TD weighed . g, a .% increase. After
to a gray brown color (Munsell YR .). The
presence of heat and moisture during immer-
sion in CR- encouraged the ow of PEG
from the cells of the wood so that, after
the treatment, the surfaces of TD felt dry.
Additionally, the surface texture and wood
grain of TD were natural in appearance,
even though waterlogging had changed the
color.
Measurements of the control tongue de-
pressor (TD) were used to monitor dimen-
sional changes for TD at various stages of
treatment. TD measured . cm long and
. cm wide at the center. After preserva- Fig. .. Structure of
tion in PEG , swelling was noted in TD. the silicone polymer.
 treatment in CR-, the weight of TD was in a fume hood for several days. Once ex-
. g, a .% decrease. posed to fresh air, the solution formed a hard,
38 Prior to treatment, temperature variations clear polymer.
aected TD. At slightly elevated tempera- The solution of CR-, PEG, and water
tures of C, the surfaces of the wood were that resulted from the extraction of free-ow-
sticky; at C, its surfaces were wet. After ing PEG from wood formed hard polymers
treatment in CR-, however, stickiness and when the solution was exposed to fresh air
pooling of PEG were not noted when TD for three days. There was some 13C, NMR
was exposed to elevated temperatures. The evidence to indicate that the CR- also re-
CR--treated wood remained dry without acted with some of the hydroxyl end groups
signs of warping. of the PEG, aiding in the formation of new
The surfaces of TD were slick with a thin polymers within the cell wall structures of
layer of PEG when the wood was removed the wood. Scanning electron microscopy was
from the CR- solution. After wiping with conducted on samples of wood prior to and
lint-free cloths, the wood looked more like after treatment in CR-. In gure ., the
the original color and texture of the control cell structure of the wood is not evident due
wood sample. After a few minutes of expo- to the heavy concentration of PEG. After
sure to fresh air, the surfaces of the wood treatment in warmed CR-, however, PEG
were dry, without additional color changes. that had not bonded to cell walls was elimi-
When the CR- was cooled to room tem- nated. At higher magnication, the bond be-
perature, a suspension of PEG was noted in tween PEG and interior cell walls is evident
the solution (g. .). This solution was (g. .).
poured into a shallow pan and allowed to sit
Conclusions

Silicone polymers are formed when PEG-

treated wood is re-treated with CR-
monomers. These polymers are hydropho-
bic in nature, and their presence appears to
make the wood less sensitive to atmospheric
moisture. This is a potential benet for the
treatment of composite artifacts when the
stability of closely associated metal compo-
nents is an issue. Silicone oils can be used
for spot treating organic materials. The sili-
cone oil/CR- solution can be topically ap-
plied using Q-tips or small brushes. After
allowing the solution to permeate into the
organic substrate, a catalyst can be applied
topically or as a vapor. This allows the con-
servator to specially treat dicult areas of a
PEG-treated artifact or areas that may be in
contact with metals.
Little is known about the various by-
Fig. .. PEG from TD products created from the decomposition of
suspended in CR- PEG over time and the potential contami-
solution. nants resulting from long-term deterioration.
 39

Fig. .. PEG in
waterlogged wood.

One potential by-product is formic acid.5 due to heat-induced migration. After treat-
Polymers formed in solution with CR- are ment in CR-, PEG bonded to cell walls is
more rigid in structure, and in this state, for- stable and resistant to environmental inu-
mic acid is not formed. Methanol, however, ences. One advantage to the CR- treatment
is formed in the hydrolysis of CR-. Subse- process is that, after treatment, the diagnos- Fig. .. After free-owing
quent testing has shown that catalyzation of tic attributes and general aesthetics of the PEG is removed from the
PEG-impregnated artifacts is not necessary. wooden artifact are well preserved. Due to wood, PEG bound to the
CR- hydrolyzes in the aqueous environ- the process of degradation caused by immer- cell walls is evident.
ment to the triol. It then self-condenses to
form the silicone polymer, forming two dis-
tinct silicons illustrated in the 29Si spectra (see
g. .). The silicon environment has a me-
thyl group and three siloxy bonds (i.e., So-
O-Si). The second silicon environment has
only two bonds. The third is formed to the
methyl group while the fourth bond is to the
hydroxy group as in the 1H spectra (see g.
.). Polymerization can be accomplished
without catalyzation. The use of catalysts
should not be ruled out, however. Additional
experimentation is needed to determine if
catalyzation acts to strengthen or enhance the
polymers created by combining PEG with
CR-.
Repeated testing has shown that immer-
sion in heated CR- removes unbound PEG
from the cells of waterlogged wood, probably
 sion in water, CR-- treated wood does not with large canons is the sabot, a wooden, at-
have the same coloration as the nonwater- based attachment that allowed better posi-
40 logged control wood samples. The wood, tioning of the ammunition and fuse within
however, is dimensionally stable after treat- the breech of a gun. In this case study, I ex-
ment in the silane. Because unbound PEG amine an eective process developed at the
has been removed from the cell structure, the Archaeological Preservation Research Labo-
wood is lighter in weight. ratory to stabilize the wooden components
After several days exposure to fresh air, of a composite artifact initially preserved
the CR-/PEG solution resulting from ex- using polyethylene glycol.
traction of PEG from TD had formed into a The two sabots delivered to the lab were
hard, clear polymer. CR- self-condenses to part of an assemblage of Civil War spheri-
form this polymer, with or without PEG. At- cal shot cases recovered from Mirror Lake
tempts at dissolving the polymer in water, al- near North Calais, Vermont. These shot cases
cohol, or acetone have not been successful, and were intact when recovered, but the at-based
the polymer is not aected by uctuations in wooden sabots had been separated from the
temperature or humidity. Additional experi- cast iron components for conservation pur-
mentation is needed to determine if the CR- poses. First, each of the shells was considered
polymer, which had bonded to the cell walls live, necessitating that each be defused. Sec-
of the wood, strengthens the artifact. Prelimi- ond, it was determined that the component
nary results indicate this may be the case. parts required separate conservation and that
Because unbound PEG ows freely from the each part would need close monitoring in wet
artifact during treatment in CR-, cell inte- storage. Figure . is a more modern ex-
riors and voids within the wood remain un- ample of an assembled spherical shot.
clogged. Removal of free-owing PEG results Inspection of the two wooden artifacts
in a reduction in weight of the wooden arti- indicated that, even three years after treat-
fact. This reduces physical stress on the wood. ment, both remained wet to the touch and
The resultant polymers do not ll cell struc- surprisingly spongy. Because both sabots
tures and voids within the wood, making ad- were very wet, it was impossible to section
ditional re-treatment possible, if needed.
Because of the self-condensing nature of
CR-, resulting in the formation of poly-
mers, catalyzation is not necessary. In the case
of the waterlogged tongue depressors, the use
of topical and vapor deposition catalyzation
does not appear to adversely aect the end
results of CR- treatment.

Case Study: Re-treatment of Two

PEG-Treated Sabots

Numerous books have been written on the

history of the machinery of war, and exhaus-
tive studies have investigated technological
advances prior to and during the American
Civil War. One interesting adaptation to
spherical case ammunition often associated Fig. .. Sabot and shot.
 41

Fig. .. Sabots A and B.

and visually identify the genus and species to a / water/PEG mixture. Following
of the wood used for construction. Accord- treatment, the sabots were surface-wiped and
ingly, the decision was made to attempt this then freeze-dried to remove excess moisture.
identication after the wood had been re- Upon completion of the preservation pro-
treated. Sabot B was in better condition than cess using PEG, it was evident that it would
sabot A, which had numerous cracks around be impossible to reassemble the wooden
the upper edge of its concave surfaces and a bases with their conserved components, due
central hole and crack running through the to the hygroscopicity of the wood. Sabots A
base. In both cases, raised wooden grain along and B are illustrated in gure ., and the
the tangential plane suggested that, at some exterior surface texture of the sabots is illus- Fig. .. The exterior
point, the specimens had been partially de- trated in gure .. surface of one sabot.
hydrated, causing shrinkage and slight dis-
tortion of the surface wood. As E. E. Astrup
has noted, the problem of shrinkage is ex-
pected because of evaporation of residual
water within a treated specimen and the re-
sultant decrease in the volume of PEG.6
Conservation notes for the two artifacts
indicated that both sabots had been mechani-
cally cleaned with dental instruments. Each
artifact was then rinsed in a % bath of hy-
drochloric acid to remove iron stains prior
to rinsing in a % solution of hydrogen per-
oxide with % ammonia in deionized water
to remove organic stains. After additional
rinsing, the sabots were treated in a blend of
and PEG, which was slowly increased
 Prior to re-treatment, each of the sabots aluminum foil cap was placed over the top
was weighed, measured, and photographed of the cylinder. Then the TD, immersed in
42 to record all aspects of the artifacts for com- solution, was placed into a vented warming
parative analysis following re-treatment. Be- oven set at .C for hours (g. .). The
cause neither sabot appeared to be perfectly sample was then removed from the oven and
round, two sets of diameter measurements allowed to stand at room temperature for
were recorded. The position of the rst mea- hours before it was taken out of the CR-
surement was recorded, and a second diam- and surface-wiped with a paper towel. In-
eter reading was taken at degrees to the spection of the cylinder revealed that a large
rst. Tracings of the top and bottom surfaces volume of free-owing PEG had been ex-
of the sabots were made to assist in monitor- tracted from the tongue depressor. As the CR-
ing potential dimensional changes that may solution cooled, PEG remained in suspen-
have resulted from re-treatment. Dimen- sion, appearing as a uy white substance.
sional data appear in table .. The last step in the process of re-treating
the PEG-treated TD was to initiate vapor
Test Process in Preparation for catalyst deposition, using a tin-based cata-
Treating the Sabots lyst to complete the polymerization process.
To accomplish this, a closed environment was
An initial experiment was conducted to de- needed to contain catalyst fumes in close as-
termine the ecacy of an alkyl silane for re- sociation with the tongue depressor. A con-
moving free-owing PEG from the cellular tainment chamber was constructed out of a
structure of waterlogged wood. This type of one-quart polyethylene pail placed in an in-
cross-linker is commonly used in the Re- verted position so that its tight-tting cover
search Laboratory as a cross-linking agent for served as the at base of the unit. A catalyst
the purpose of polymerizing dimethyl silox- tray was created using an aluminum sample
ane, hydroxyl-terminated polymers. It has tray with mesh aluminum screen placed over
proved potentially useful for treating chemi- it. The screen formed a platform on which
cally bound PEG within the cell wall struc- the sample rested and was positioned above
tures of wood, reducing the hygroscopicity the catalyst tray during catalyst vapor deposi-
of the remaining PEG. tion processing. Several paper towels were
To illustrate the general results from re- placed on top of the screen to prevent free-
peated experimentation, a single waterlogged owing PEG from dripping into, and contami-
tongue depressor (TD), treated with PEG nating, the catalyst. Fifteen grams of CT-
, was taken from its PEG solution and were added to the catalyst tray, and the TD was
wiped with a paper towel to remove pooled positioned on the screen with the body of the
PEG from all its surfaces. The sample was containment chamber properly tted over it
then placed into a graduated cylinder con- (g. .).
taining ml of fresh CR-. A loose-tting Once assembled, the unit was placed into

Table . Postre-treatment Measurements for Sabots A and B

Sample Top Width (in cm) Base Width of Rim (in cm) Height (in cm) Weight (in g)
Sabot A 11.505 (1st reading) 9.72 (1st reading) 4.882 2.662
11.026 (2d reading) 9.75 (2d reading)
Sabot B 10.986 (1st reading) 10.00 (1st reading) 4.65 2.622
10.984 (2d reading) 9.73 (2d reading)
 solution into a milky colored solution as a
result of the PEG in suspension, which had
been removed from the tongue depressor. 43
Instead of remaining dark brown, the tongue
depressor was a light gray brown. The re-
treatment color was very similar to the color
of the control samples. Prior to re-treatment,
the PEG-treated tongue depressor had
swelled to a size slightly wider than the con-
trol tongue depressor. After re-treatment, the
tongue depressor had returned to its origi-
nal width.

Re-treatment of PEG-Treated
Waterlogged Wood

Fig. .. Conguration of the containment The methods used to preserve sabots A and
chamber used for catalyst vapor deposition: B were the same as those outlined for the
(A) vented oven; (B) loose-tting cap; tongue depressor above. Sabot B was placed
(C) graduated cylinder; (D) CR- solution; into a large plastic container and immersed
(E) tongue depressor. in a sucient volume of CR- to ensure that
there was approximately inch of CR-
covering the uppermost portion of the arti-
a vented warming oven set at C. The fact. A loose-tting top was positioned over
sample remained in catalyst vapor deposition the container, and the artifact in solution was
for hours. Following treatment, the sample placed into a well-vented warming oven set
was removed from the oven and transported at C. The artifact was left in treatment for
to a vented fume hood and opened. After the . hours.
TD was allowed to sit in fresh air for several Sabot A had extensive cracks that extended
days, it was placed into a vented warming through its sides and base. To determine if
oven and heated to F for a few hours. the presence of a catalyst in the cross-linker
Throughout the heating process, no dimen-
sional changes or sweating were noted.

Observations of the Test Process

After bulking with the PEG and subse-

Fig. .. Conguration of
quently being wiped with paper towels, the
the containment chamber
treated TD was dark brown. The edges of the
used for catalyst vapor
TD were translucent, and its surfaces felt deposition:
waxy and smooth. After re-treatment to re- (A) warming oven;
move free-owing PEG , the tongue de- (B) glass containment
pressor was removed from the oven, gently chamber; (C) catalyst
surface-wiped, and allowed to sit in fresh air. fumes; (D) tongue
After returning to room temperature, the depressor; (E) screen with
CR- had changed from a clear, watery paper towel; (F) catalyst.
 Observations

44 After hours of exposure to fresh air, both

sabots were measured, weighed, and photo-
graphed. The most notable change was in
coloration. In both cases, the overall color of
the artifacts had turned from a dark brown-
black to a lighter, sandy brown. Neither sabot
was greasy to the touch, and in both cases,
the surfaces were dry and natural in appear-
ance. Prior to re-treatment, the surfaces of
both artifacts looked as though they had ex-
perienced some dehydration, either in the
eld during excavation or through too rapid
treatment in the lab using PEG. The re-treat-
Fig. .. The computer- ment process to remove free-owing PEG
controlled vented warm- from the artifacts did not appear to have ex-
ing oven used for Sabots A acerbated the problem.
and B. Small nails had been used to fasten the
iron bands that held the spherical shot in the
solution would deter the development of sabot when the ammunition was manufac-
cracks in severely cracked wood, sabot A was tured. In both artifacts, small cracks centered
placed into a container with a similar vol- around these nail holes were noted prior to
ume of CR- mixed with CT- catalyst, in re-treatment. In all cases, the cracks in sabot
a : ratio. Like sabot B, sabot A was placed A were more visible after re- treatment, with
into a vented warming oven set at C, and one crack appearing to be slightly larger. The
treated for . hours (g. .). cracks in sabot B, however, reacted dierently
After treatment in their respective solu- to treatment. In most cases, the cracks re-
tions, sabots A and B were removed from the mained small and dicult to see. One small
oven and, still in their solutions, placed into crack, which ran from a nail hole upward to
a ventilated fume hood. Both artifacts were the thin edge of the sabot, was not visible af-
allowed to slowly return to room tempera- ter re-treatment, suggesting that the wood
ture over a period of approximately hours. may have swelled slightly, closing the crack.
The sabots were then removed from their Pre- and post-re-treatment data for both ar-
solutions and surface-wiped with soft, lint- tifacts are shown in table ..
free cloths. Inspection of the artifacts indi-
cated that, just as before re-treatment, the Conclusions
surfaces of both sabots were mushy in tex-
ture; the cell structure of the wood was also Clearly, less shrinkage occurred in sabot B
badly deteriorated. Each sabot was then than in sabot A. Minimal changes in cracks
placed on a tray in the fume hood and al- associated with nail holes running through
lowed to air-dry overnight in owing, fresh the sides of sabot B suggest that the presence
air. Changes in coloration were noticed im- of catalyst mixed with CR- assisted in sta-
mediately. During the rst minutes of air- bilizing the wood. The presence of catalyst
drying, the sabots became lighter in color and also appears to have prevented the slightly
more natural in appearance. higher rates of shrinkage noted in sabot A.
Table . Pre and Postre-treatment Data for Sabots A and B

Sample Pretreatment Posttreatment % Change 45

Sabot A
Top diameter (in cm) 11.050 (1) 11.01 0.36
Top diameter (in cm) 11.026 (2) 11.00 0.20
Base diameter (in cm) 9.72 9.70 0.21
Height (in cm) 4.88 4.67 4.31
Weight (in g) 266.20 238.40 9.58
Sabot B
Top diameter (in cm) 10.986 (1) 11.04 0.49
Top diameter (in cm) 10.984 (2) 11.01 0.24
Base diameter (in cm) 10.00 10.00 0.00
Height (in cm) 4.65 4.65 0.00
Weight (in g) 262.20 245.00 6.56

On the basis of this experiment, however, it dants to help stabilize polymer-bulking

is not evident what role the presence of CR- media.
played in accounting for the higher per- Initial tongue depressor experiments in-
centage of weight reduction for sabot A. dicate that Passivation chemistries allow the
Bilz et al. have noted that PEG degrades conservator to remove free-owing PEG
even at temperatures as low as C.7 Nu- from wood, and where applicable, eectively
merous museums have approached the Ar- reduce treatment-associated swelling that can
chaeological Preservation Research Labora- occur in PEG-treated artifacts. Passivation
tory asking for directed research to prevent chemistries appear to hold some potential for
the obvious deterioration of PEG-treated removing free-owing PEG from organic
artifacts that have been curated, even in the materials and for stabilizing artifacts that
best of controlled environments at room have been treated using PEG. Research at the
temperature. Data from recent experimen- Archaeological Preservation Research Labo-
tation suggest that PEG extraction and poly- ratory continues to dene experiments that
merization can be successfully conducted at may help us understand new methods of
room temperature. The process takes longer, working with PEG and other polymer pro-
but the room temperature treatments help cesses.
to minimize cracking and surface-checking.
Additional experiments should be conducted
to take into account the notable factors and Case Study: Treatment of Waterlogged
results included in the Bilz et al. study. Wood Using Hydrolyzable, Multifunc-
As Bilz et al. and others have noted, PEG tional Alkoxysilane Polymers
does degrade by random chain scission, es-
pecially in an anaerobic environment. Ex- Three groups of waterlogged wood samples
periments conducted on similarly unstable were used in this experiment. Group
polymers (silicone oils) have indicated that samples consisted of waterlogged tongue de-
the use of antioxidants such as hydroxy ani- pressors. Group samples were sectioned
sole may inhibit aspects of oxidation and sili- from a large piece of archaeological wood
cone oil degradation. Substantial research taken from marine excavations of the
needs to be conducted on the use of antioxi- community of Port Royal, Jamaica. Group
 samples consisted of treenails (wooden NMR analysis was also conducted to de-
dowels) extracted from the frames and large termine whether the waterlogged tongue
46 timbers of the seventeenth-century ship- depressors were suciently degraded to
wreck La Belle. Each of these hand-carved provide a valid substitute for archaeologi-
pieces of wood was similar in diameter (mean cal wood in such an experiment. Spectra of
. mm) and length (mean . mm). the group , , and wood samples were
Many of their surfaces bore diagnostic tool nearly identical to spectra reported by
marks. Michael Wilson et al. in The Degradation of
Nuclear magnetic resonance (NMR) spec- Wood in Old Indian Ocean Shipwrecks.8 To
tra of waterlogged wood treated with alko- determine the physical integrity of the
xysilane polymers indicates that, in an aque- group wood samples, control samples were
ous environment, CR- hydrolyzes to form oven-dried over a -hour period. In all
a triol. The triol self-condenses to form a cases, the degree of warping and shrinkage
range of polymers in the 29Si spectra. The indicated that they responded similarly to
primary silicon has a methyl group and three waterlogged archaeological wood.
siloxy bonds (i.e., So-O-Si). The second sili- Classroom conservation experimentation
con environment has only two bonds. The demonstrated that waterlogged tongue de-
third is formed to the methyl group, while pressors proved to be good indicators of the
the fourth bond is to the hydroxy group as in eectiveness of traditional treatment meth-
the 1H spectrum (g. .). One goal of this ods, such as polyethylene glycol, acetone
experiment was to determine whether these rosin, and sucrose. This same type of water-
polymers were sucient to maintain the logged wood appeared to work equally well
physical attributes, cell structure, and aesthet- as indicators of the eectiveness of polymer
ics of group , , and wood samples. preservation treatment strategies.

Fig. .. 1H CRAMPS
spectrum of the solid
polymer formed from
the CR-.
 47

Fig. .. 13C CP/MAS

spectra of an untreated,
nonwaterlogged control
tongue depressor (bottom)
and an untreated, water-
logged tongue depressor
(top).

MaterialsGroup 1 Waterlogged sample is illustrated at the top of gure ..

Tongue Depressors Long-term saturation in tap water altered the
chemical structure of the wood, as evidenced
Waterlogged tongue depressors were chosen by the loss of the acetate resonances at and
for group samples because they are easily ppm in comparison to the spectrum of a
obtained and relatively uniform in dimen- control tongue depressor that had not been
sion, grain, and color. To create a supply of waterlogged (g. ., bottom). The changes
waterlogged wood, thousands of white birch in these 13C spectra of the birch wood tongue
(Betula papyrifera) tongue depressors were depressors are quite similar to those reported
placed into one-gallon glass jars lled with by Wilson et al. for oak wood from ship-
tap water. The jars were then sealed and wrecks.9 The 13C spectral signature and the
stored in a cabinet. macroscopic observations of extensive warp-
For this experiment, hundreds of tongue ing and shrinking following air-drying sug-
depressors, which had been immersed in tap gest that these samples provide a suitable
water for a period of nine years, were emp- model for the analysis of waterlogged wood.
tied into a plastic vat and rinsed in running
tap water for hours. Ten tongue depressors MaterialsGroup 2 Waterlogged
were selected randomly from the rinse vat, Archaeological Wood
designated as air-dry samples, and used to
determine the mean water content of the A small plank of wood recovered during ar-
group samples. Eighteen additional tongue chaeological excavations at the submerged
depressors were randomly selected for treat- site of seventeenth-century Port Royal,
ment using acetone dehydration followed by Jamaica, was selected for group wood
acetone/CR- displacement. samples. The plank, measuring . cm wide,
The 13C CP/MAS spectrum for one group . cm long, and . cm thick, was cut into
48

Fig. .. Surface charac-

teristics, sections, and
dimensions of the
waterlogged plank used
for the group water-
logged wood samples. Top
surface of the plank (left);
obverse surface (right).

four sections. Due to its waterlogged state, shipwreck, La Belle. Each piece of wood was
the plank was very fragile. Sectioning was roughly carved and slightly tapered in shape.
accomplished using a long scalpel blade. Fig- After desalination in freshwater baths for
ure . illustrates the dimensions of the months, treenails were surface-dried with
group wood samples. paper towels and then weighed, measured,
and photographed (gs. . and .). Six-
MaterialsGroup 3 Waterlogged Treenails teen of the treenails were then placed into a
series of four ethanol baths, followed by a
Fig. .. Waterlogged Group samples were treenails extracted series of four acetone baths, each lasting two
treenail before air-drying. from the timbers of a seventeenth-century weeks. Eight of the treenails were picked ran-
domly and immersed in CR-. Eight were
immersed in Q-. The remaining two
treenails were air-dried for hours in a
vented warming oven to determine percent-
age of water content.

MaterialsChemicals

CR- is a chemical monomer that reacts

with water to form silanetriol and methanol.
The silane in turn condenses with available
hydroxyl groups or other silanol monomers
to form siloxane resins. CR- consists of
% methyltrimethoxysilane, % methyl al-
cohol, and % dimethyldimethoxysilane. The
 condensation product of CR- is a resin
with a molecular weight of . Like CR-,
Q- is generally a clear liquid. However, 49
because of the lower percentage of CR- and
higher percentage of alcohols, evaporation
during treatment is greater, and there are
fewer hydroxyl groups and other silanol
monomers available to form resins. The Q-
solution is complex, and consists of %
CR- (by weight), mixed with % methyl
alcohol, % isopropyl alcohol, % ethyl alco-
hol, and % dimethyldimethoxysilane. Indus-
trial-grade acetone, certied to be .% free
of water, was used for all dehydration pro-
cessing. mass of . g. W was placed into a fume Fig. .. Treenail follow-
hood and allowed to air-dry for hours at ing air-drying. Note that
Water Content in Group , , ambient pressure and a constant room tem- this treenail splintered
and Wood Samples perature of F (.C). After drying, it into three sections.
weighed . g, a weight loss of .%, with
Percentage of water content was calculated a calculated water content of .%.
for each group of samples using the follow-
ing equation:
Two waterlogged treenails, weighing . g
and . g wet weight, were allowed to air-
%H2O = weight of wet wood weight of oven-dried wood
X 100 dry in a well-ventilated fume hood for
weight of oven-dried wood
hours. After air-drying, the treenails weighed
. g and . g, a weight loss of .%
(.% water content) and .%
Ten group waterlogged tongue depressors, (.% water content), respectively.
designated as air-dry samples, were placed
into a ventilated warming oven, set at C, Treatment Methods
for hours. The mean water content of the
group samples air-dried in the oven was
.%. Eighteen group waterlogged tongue depres-
sors were surface-dried with paper towels and
labeled incrementally with a felt tip pen. Each
After hours of air-drying in a vented samples weight, length, width, and thickness
warming oven, set at C, W wood sec- were recorded.
tion weighed . g, a weight loss of .%. The samples were placed into a large bea-
Water content was calculated to be .%. ker containing one liter of fresh, industrial-
Because of the uniform thickness and con- grade acetone. After hours, they were
dition of the wood, the water content cal- transferred to a second beaker, containing
culation for W was assumed to reect the one liter of fresh acetone, and dehydrated for
general state of degradation of the other an additional hours. After hours of de-
wood sections. hydration, the group wood was transferred
Prior to air-drying, W, the second air- into a beaker containing one liter of CR-.
dried archaeological wood sample, had a The beaker was placed into a desiccator
 vacuum chamber, where the wood was held and dehydrated at ambient pressure and
in a reduced pressure environment of . room temperature for a total of hours. The
50 Pa ( Torr) for hours. The wood was then wood was then transferred to a beaker con-
allowed to sit in the solution at ambient pres- taining one liter of Q- and placed into a
sure and room temperature for hours be- desiccator vacuum chamber. A reduced pres-
fore being removed from the CR- solution sure of . Pa ( Torr) was applied for
and placed on paper towels in a vented fume hours. The wood was then left in solution
hood to air-dry for hours. at ambient pressure and room temperature
The samples were then placed in close for an additional hours. After acetone/Q-
proximity to an aluminum weighing dish displacement, W was removed from the
containing g of tap water and enclosed in Q- solution and placed in a Ziploc bag
a large Ziploc bag, thus creating a closed, with an aluminum weighing dish containing
humid environment. After hours of ex- g of tap water. The bag was sealed and the
posure to water vapor, the wood was removed wood was allowed to sit for hours. The
from the bag and allowed to air-dry in a fume wood was then removed from the bag and
hood before measurements and assessment placed in a ventilated fume hood for an ad-
of the treatment process were made. ditional hours of exposure to fresh air.
After air-drying, the wood was weighed and
measured.
Two sections of the plank, W and W, were
chosen for treatment with alkoxysilanes. W
was placed into a beaker containing one liter Sixteen treenails were rst dehydrated in a
of industrial-grade acetone and dehydrated series of three ethanol baths, each lasting one
at ambient pressure and room temperature week. Dehydration was then continued in a
for hours. The wood was then placed in series of three acetone baths, changed at two-
fresh acetone for an additional hours of week intervals. For the last hours of dehy-
dehydration. After a total of hours of de- dration, the samples were placed into a large
hydration, W was transferred to a beaker vacuum chamber and treated at a reduced
containing one liter of CR- and placed into pressure of . Pa ( Torr). During this
a desiccator vacuum chamber. A reduced phase of dehydration, the samples were
pressure of . Pa ( Torr) was applied monitored closely to ensure that they re-
for hours. The wood was left in the solu- mained immersed in acetone.
tion at ambient pressure and room tempera- Eight of the treenails were then randomly
ture for an additional hours. After acetone/ selected and carefully transferred to a large
CR- displacement, W was removed from beaker containing one liter of CR-. The
the CR- and placed into a Ziploc bag. An remaining treenails were transferred into a
aluminum weighing dish, containing g of beaker containing one liter of Q-. Both
tap water, was placed inside the bag, in close beakers were placed into a vacuum chamber
proximity to the CR--treated wood. The and treated at a reduced pressure of .
bag was sealed and the wood was allowed to Pa ( Torr) for hours. The treenails were
sit for hours. The wood was then removed then stored at ambient pressure in their re-
from the bag and placed in a ventilated fume spective polymer solutions for an additional
hood for an additional hours of exposure seven days. All the treenails were removed
to fresh air. After air-drying, the wood was from their solutions, surface-dried with pa-
weighed and measured. per towels, and then placed into a fume hood,
Like W, W was placed into a beaker con- where they were allowed to air-dry for
taining one liter of industrial-grade acetone hours.
Observations: Group 1 Wood of an untreated birch (Betula papyrifera) con-
trol tongue depressor showing uniformly
Prior to any treatment, the mean wet weight shaped, thick-walled tracheids. In contrast, 51
of the group CR--treated wood samples the tracheids in the waterlogged wood sample
was . g. After treatment, the mean weight (right) are irregular in shape, with deteriora-
was reduced to . g, a mean reduction of tion of the middle lamella. Figure . shows
.% (table .). two views of the microstructure of a group
The mean width of the waterlogged CR--treated sample of wood. Cell wall col-
tongue depressor samples that were simply lapse is negligible, and there is very little dis-
air-dried was reduced from . mm to . tortion and structural loss of middle lamella.
mm, a .% reduction (table .). The mean In gure ., two cross-sectional views indi-
length of these samples was reduced from cate that, after air-drying, the cell structure
. mm to . mm, a .% reduction of the group waterlogged wood samples
(table .). The mean thickness of these collapsed, causing extreme shrinkage and
samples decreased from . cm to . cm, a distortion of the wood.
reduction of .%. One control tongue de-
pressor and several air-dried group pieces
of wood are illustrated in gure .. Shrink-
age, distortion, and color change are evident
in these examples. After air-drying, the color
of the waterlogged tongue depressors had
shifted from their natural light yellow brown
( YR// Munsell) to a light gray brown
color (.Y// Munsell) or a darker brown-
ish gray ( YR// Munsell).
In contrast, the dimensions and aesthetic Fig. .. Control tongue
attributes of all the group CR--treated depressor (top); three air-
wood samples were well maintained after dried waterlogged tongue
treatment. Changes in length were minimal, depressors (below).
decreasing only .% after CR- treatment
(table .). Changes in width were noticeably
higher, with an average posttreatment reduc-
tion of .% (table .). One control tongue
depressor and nine CR--treated wood
samples are illustrated in gure ..
Cross section samples of control, un-
treated waterlogged wood, air-dried wood,
and CR-- treated tongue depressors were
analyzed using an environmental scanning
electron microscope (ESEM). For analytical
consistency, photographs of each sample
were recorded at -X magnication. Cell
shape retention, cell wall integrity, and gen- Fig. .. Control tongue
eral appearance were used to assess the e- depressor (top); nine
cacy of the treatments. CR--treated, water-
In gure ., the left image is a -X logged tongue depressors
magnication of the cross-sectional surface (below).
 Table . Weight Changes for Group CR--Treated
Tongue Depressor Samples
52
TD Wet Weight Posttreatment % Change
(in g) Weight (in g)

1 5.0 2.7 46.00

2 5.8 2.6 55.17
3 5.3 2.4 54.72
4 5.5 2.0 63.64
5 5.7 2.3 59.65
6 5.9 3.0 49.15
7 5.4 2.5 53.70
8 5.8 2.6 55.17
9 5.7 3.0 47.37
10 5.7 2.8 50.88
11 5.5 2.5 54.55
12 5.5 2.5 54.55
13 5.8 2.7 53.45
14 5.7 2.6 54.39
15 5.8 2.6 55.17
16 5.7 2.6 54.39
17 5.6 2.5 55.36
18 5.8 2.8 51.72
Mean Wet Mean Post- Mean %
Weight treatment Weight Change
5.63 2.59 53.83

Table . Width Change of Air-Dried Waterlogged Tongue Depressors

TD Wet Width Air-Dry Width % Change

(in mm) (in mm)
1 17.90 13.47 24.75
2 17.87 14.32 19.87
3 17.87 13.24 25.91
4 17.91 13.43 25.01
5 17.90 14.60 18.44
6 17.88 12.90 27.85
7 17.87 14.36 19.64
8 17.89 13.53 24.37
9 17.90 14.25 20.39
10 17.91 12.61 29.59
Mean Wet Mean Air-Dry Mean %
Width Width Change
17.89 13.67 23.58
Table . Change in Length of Air-Dried Waterlogged Tongue Depressors

TD Wet Length Air-Dry Length % Change 53

(in mm) (in mm)

1 152.50 151.08 0.93
2 152.22 150.88 0.88
3 152.47 150.86 1.06
4 152.26 151.03 0.81
5 152.48 150.97 0.99
6 152.51 151.09 0.93
7 152.59 151.77 0.54
8 152.48 151.11 0.90
9 152.35 151.76 0.39
10 152.34 151.62 0.47
Mean Wet Mean Air-Dry Mean %
Length Length Change
152.37 151.74 0.71

Table . Changes in Length for Group CR--Treated

Tongue Depressor Samples

TD Wet Length Preserved Length % Change

(in mm) (in mm)
1 152.49 151.75 0.49
2 152.46 151.60 0.56
3 152.34 152.04 0.20
4 151.97 151.17 0.53
5 152.48 152.03 0.30
6 152.58 152.10 0.31
7 151.82 151.33 0.32
8 152.48 151.94 0.35
9 152.68 151.96 0.47
10 152.50 151.97 0.35
11 152.59 151.55 0.68
12 152.50 151.65 0.56
13 152.46 151.97 0.32
14 152.21 151.58 0.41
15 152.58 151.99 0.39
16 152.56 151.85 0.47
17 152.51 152.14 0.24
18 151.48 150.78 0.46
Mean Wet Mean Post- Mean %
Length treatment Length Change
152.37 151.74 0.41
 Table . Changes in Width for Group CR--Treated
Tongue Depressor Samples
54
TD Wet Width Post Treatment % Change
(in mm) Width (in mm)

1 19.71 17.59 10.76

2 19.78 17.95 09.25
3 19.66 17.35 11.75
4 19.70 17.95 08.88
5 19.80 17.70 10.61
6 19.00 17.79 11.05
7 19.72 17.93 09.08
8 19.80 17.33 12.48
9 19.77 17.95 09.21
10 19.75 17.94 09.17
11 19.79 17.62 12.32
12 19.64 17.24 12.22
13 19.70 17.05 13.45
14 19.87 17.09 13.99
15 19.59 17.79 10.12
16 19.71 17.61 10.65
17 19.70 17.18 12.79
18 19.90 17.48 12.16
Mean Wet Mean Post- Mean %
Width treatment Width Change
19.76 17.86 11.11

Fig. .. Control group

tongue depressor (left) Observations: Group Wood
and an image of a water-
logged group tongue Prior to air-drying in a warming oven, W
depressor (right), both weighed . g and measured . cm wide,
viewed at -X . cm thick, and . cm long. Because W
magnication. was cut thinner than W, W, and W, it was
very exible and delicate. After hours of
air-drying, W was reduced from its wet
Fig. .. Two cross- weight of . g to . g, a decrease of .%.
sectional microscopic The water content of this sample was .%.
views (-X magnica- When removed from the warming oven, the
tion) of a group tongue wood had completely collapsed and frag-
depressor treated with mented into ve signicant sections, which
CR-. precluded length and width measurements.
Prior to treatment, W was . cm thick. Mean
Fig. .. Two cross-
thickness of W wood fragments after treat-
sectional views of group ,
air-dried wood samples.
ment was . cm, a reduction of .%.
In both images, cellular W was sectioned from the plank adjacent
distortion and collapse are to W. This section of wood had a wet weight
apparent, resulting in of . g and measured . cm wide, . cm
extreme warping and thick, and . cm long. W was designated
distortion. for treatment in CR-, after dehydration in
acetone. After treatment in CR-, W ap- long, a loss of .%. The pre- and posttreat-
peared uniformly dry and light in color. Only ment measurements of all four group
slight dimensional changes were noted. The samples are summarized in table .. 55
wet weight of W was reduced from . g to After hours of treatment, the W sample
. g, a decrease of .%. W measured . of wood had deteriorated into a pile of splin-
cm wide prior to treatment. After treatment, ters, precluding a comparison of pre- and
it measured . cm, a reduction of .%. posttreatment dimensions and aesthetic
Thickness of W prior to treatment was . qualities. The wood section was computed
cm. After treatment, it measured . cm thick, to have a water content of .%.
indicating that little or no signicant change Prior to treatment, the W wood was dark
had occurred in thickness. Before treatment, brown ( YR// Munsell). After treatment,
W measured . cm long. After treatment, it was a light gray brown ( YR// Munsell).
it measured . cm long, a loss of .%. The surfaces showed no signs of checking and
W had a wet weight of . g. Its wet mea- the wood looked very natural. After treatment,
surements were . cm wide, . cm thick, however, W was very light in weight, but the
and . cm long. W was designated for preserved wood withstood extensive handling
treatment in Q-, after an initial dehy- with no signs of deterioration or wear.
dration in acetone. After treatment in Q- After treatment, W was also aesthetically
, the weight of W decreased from . g. pleasing. Prior to treatment, W was dark
to . g, a loss of .%. The width of the brown ( YR// Munsell). After treatment,
sample decreased from . cm to . cm, a the wood was slightly darker ( YR//
reduction of .%. The thickness of W was Munsell) than W. Both W and W wood
reduced from . to . cm, a decrease of samples were natural in appearance, and the
.%. The wet length of W was . cm. surfaces of both pieces of wood felt dry.
After treatment, the wood measured . cm, Shrinkage of W, however, was more of a
a reduction in length of .%. problem.
The remaining section of wood, W, was Unlike W, which was oven-dried for
designated for air-drying. Prior to treatment, hours, W was allowed to air-dry in a vented
its wet weight was . g. Its wet measure- fume hood. W splintered into six large sec-
ments were . cm wide, . cm thick, and tions. Fragmentation of the wood was less
. cm long. W was the largest of two air- severe than that of W. Each of the W splin-
dried samples from the original plank of wa- ters of wood was distorted, making posttreat-
terlogged wood. Unlike W, W was similar ment measurement of dimensions dicult.
in size to the sections of wood treated with Because shrinkage and fragmentation were
alkoxysilane polymers. Prior to treatment, its extensive, it was impossible to t the distorted
wet weight was . g. After treatment, it sections of wood into a form that could be
weighed . g, a loss of .%. Its width de- accurately measured. The posttreatment con-
creased from . cm to . cm, a reduction dition of the four wood samples is illustrated
of .%. The thickness of the sample fell in gure ..
from . cm to approximately . cm, a re-
duction of .%. The posttreatment length Observations: Group 3 Wood
of W was dicult to determine. During
treatment, the sample splintered into six large Prior to treatment, the mean percentage wa-
sections, each suering gross distortion of its ter content in the group waterlogged tree-
edges. Prior to air-drying, the sample mea- nails was .%, appreciably lower than
sured . cm at its longest point. After treat- the water content of .% calculated for
ment, W measured approximately . cm the group wooden plank samples. During
 Table . Pre- and Posttreatment Measurement for Group Wood Samples

56 Sample Pretreatment Posttreatment % Change

W11
Weight (in g) 39.1 4.4 88.75

Length (in cm) 12.20

Width (in cm) 2.00
Thickness (in cm) 1.87 0.42 77.54
W2
Weight (in g) 115.7 28.7 75.20
Length (in cm) 14.68 14.63 0.34
Width (in cm) 4.13 3.96 4.12
Thickness (in cm) 1.86 1.85 0.54
W3
Weight (in g) 91.3 26.0 71.52
Length (in cm) 14.73 14.54 1.29
Width (in cm) 3.14 2.98 5.10
Thickness (in cm) 1.84 1.62 11.96
W42
Weight (in g) 104.3 10.3 90.13
Length (in cm) 12.27 10.41 15.16
Width (in cm) 3.05 2.55 16.39
Thickness (in cm) 1.87 0.72 61.50

1
Fragmentation precluded posttreatment length and width measurements.
2
Fragmentation precluded exact posttreatment length measurement; the value is an
approximation.

initial cleaning and desalination, the treenails Group wood samples treated with CR-
were found to be less spongy than the sec- alkoxysilane polymers experienced only
tions of plank, and the wood was noticeably slight changes in posttreatment weight,
harder. Many of the ends of the treenails had length, and diameter compared to the group
either been broken or splayed under the force Q--treated wood. After treatment, the
of being removed from the ships timbers. mean weight of CR--treated treenails was
. g, a reduction of .%. Treenails pre-
served using Q- polymer had a mean
weight of . g after treatment, a reduction
of .% (table .). Group treenails pre-
served using CR- had a mean reduction in
length of .%. The change in length was
greater for Q--treated treenails, with a
mean posttreatment length of . mm.
Fig. .. Group wood
This represents a shrinkage of .% shrink-
after treatment. W1, oven-
age (table .). Similarly, the mean change
dried wood; W2, CR--
in diameters for the group , CR--treated
treated wood; W3,
Q--treated wood; nails was substantially less than the mean
W4, air-dried wood in change in diameter for Q--treated tree-
vented fume hood. Note nails. The posttreatment mean diameter for
the comparatively lighter CR--treated wood was . mm, a reduc-
color of W2. tion of .%. Q--treated treenails had
a mean loss in posttreatment diameter of as well as the availability of nonwaterlogged
.% (table .). control samples, make the group wood
The most noticeable dierence between samples invaluable for experimentation. Be- 57
the CR-- and Q--treated treenails was cause of inherent inconsistencies of water-
in coloration. In all cases, the CR--treated logged archaeological wood, similar compara-
wood was much lighter in color. Q-- tive data cannot be derived for archaeological
treated wood tended to be darker, with fewer samples.
wood grains and surface features visible. Air-dried group wood samples experi-
enced a mean weight loss of .%, a mean
Discussion reduction in sample width of .%, and a
mean reduction in length of .%. All the
NMR spectral analysis, ESEM analysis, and samples were warped and twisted after air-
empirical data indicate that the structural drying.
integrity of the group tongue depressors CR--treated group wood samples were
was suciently degraded that the wood generally well preserved. Mean reduction in
samples can be used to evaluate preservation length was .% after treatment. Mean re-
treatments for waterlogged timbers from duction in width was .%. These gures
shipwrecks. Group wood samples also pro- may seem high, but a comparison of the post-
vide a reasonably homogenous source of treatment dimensions of the CR--treated
wood that allows quantiable and qualita- samples shows that they are nearly identical
tive analysis of the ecacy of consolidants to the untreated, control depressors, indicat-
being tested for use in conserving water- ing that the treated wood was restored to
logged wood. Regularity of size and species, dimensions nearly identical to those of the

Table . Changes in Weight for Group Treenail Artifacts

Treenail Treatment Wet Weight Posttreatment % Change

(in g) Weight (in g)
1 Air-dry 68.2 17.80 73.90
2 Air-dry 82.1 17.85 78.26
Mean (Air-dry) 75.15 17.83 76.08
3 CR-20 83.8 44.13 47.34
4 CR-20 59.8 32.42 45.79
7 CR-20 66.2 37.99 42.61
9 CR-20 65.7 35.36 46.18
10 CR-20 91.1 50.75 44.29
12 CR-20 67.4 38.32 43.15
15 CR-20 75.5 38.67 48.78
18 CR-20 84.2 44.00 47.74
Mean (CR-20) 74.21 40.48 45.74
5 Q-9-1315 69.7 37.83 45.72
6 Q-9-1315 62.2 32.86 47.17
8 Q-9-1315 77.8 40.85 47.49
11 Q-9-1315 64.4 36.59 43.18
13 Q-9-1315 82.1 46.94 42.83
14 Q-9-1315 55.5 33.43 39.77
16 Q-9-1315 91.5 53.90 41.09
17 Q-9-1315 63.5 36.32 42.80
Mean (Q- 9-1315) 70.84 39.75 43.76
 Table . Changes in Length for Group Treenail Artifacts

58 Treenail Treatment Pretreatment Length Posttreatment Length % Change

(in mm) (in mm)

1 Air-dry 106.5 92.53 13.12

2 Air-dry 145.39 122.69 15.61
Mean (Air-dry) 125.95 107.61 14.37
3 CR-20 140.84 140.30 0.38
4 CR-20 115.43 115.06 0.32
7 CR-20 148.55 148.18 0.25
9 CR-20 111.36 110.74 0.56
10 CR-20 122.47 122.05 0.34
12 CR-20 108.61 108.11 0.46
15 CR-20 144.59 144.29 0.21
18 CR-20 137.75 137.27 0.35
Mean (CR-20) 128.70 128.24 0.36
5 Q-9-1315 128.73 127.89 0.65
6 Q-9-1315 113.28 112.76 0.46
8 Q-9-1315 121.91 121.32 0.48
11 Q-9-1315 115.53 114.86 0.58
13 Q-9-1315 130.18 129.62 0.43
14 Q-9-1315 124.13 123.62 0.41
16 Q-9-1315 142.42 141.68 0.52
17 Q-9-1315 102.68 102.29 0.38
Mean (Q-9-1315) 122.36 121.76 0.49

Table . Changes in Diameter for Group Treenail Artifacts

Treenail Treatment Pretreatment Posttreatment % Change

Diameter (in mm) Diameter (in mm)
1 Air-dry 26.85 22.32 16.87
2 Air-dry 24.87 20.93 15.84
Mean (Air-dry) 25.86 21.63 16.36
3 CR-20 27.45 26.29 4.23
4 CR-20 25.05 24.14 3.63
7 CR-20 26.11 25.44 2.57
9 CR-20 27.92 27.14 2.79
10 CR-20 27.38 26.39 3.62
12 CR-20 26.75 25.65 4.12
15 CR-20 25.20 24.36 3.33
18 CR-20 27.39 26.56 3.03
Mean (CR-20) 26.66 25.75 3.42
5 Q-9-1315 27.01 23.81 11.85
6 Q-9-1315 26.83 24.27 9.54
8 Q-9-1315 29.30 26.60 9.22
11 Q-9-1315 26.77 24.82 7.28
13 Q-9-1315 27.02 24.15 10.63
14 Q-9-1315 22.31 20.68 7.31
16 Q-9-1315 27.55 25.26 8.31
17 Q-9-1315 27.58 24.62 10.74
Mean (Q-9-1315) 26.80 24.28 9.36
control wood samples. Swelling that occurred evident in the slightly higher rates of shrink-
during the waterlogging process had been age for W.
greatly reduced after the wood was treated The diagnostic attributes of W and W 59
in CR-. ESEM evaluation of these samples were preserved because sucient resins were
conrms that cell dimensions and shapes formed as the result of condensation to pre-
were similar to those of the control wood vent cellular collapse of the wood. The water-
samples. Slight shrinkage of the middle logged tongue depressors and the archaeologi-
lamella was noted after treatment. cal wood samples preserved with the CR-
After treatment, all the group tongue solution, which has a higher percentage of hy-
depressors were slightly gray brown in color drolyzable, multifunctional polymers, were the
(YR / Munsell) as compared to the color best-preserved specimens. In contrast, the Q-
of the control tongue depressors (YR / solution, containing a lower percentage
Munsell). This color shift is the result of solution of the same multifunctional
changes in the wood caused by long-term alkoxysilane polymers, was insucient to pre-
immersion in water. serve the diagnostic attributes of the wood.
Group wood samples were highly de- The waterlogged tongue depressors, calcu-
graded. The computed water content of the lated to have a moisture content of .%,
samples was between .% and .%, were very well preserved with CR-.
suggesting that water content was reasonably Wood treated using a higher percentage
uniform throughout the plank. Oven-dried of CR- solution looks natural in color and
W and air-dried W were much darker in texture following treatment. No surface-
color than the polymer-treated sections, W checking was noted in either the group
and W. This is the result of extreme cellular wood samples or the archaeological wood.
collapse and warpage that occurred as a re- During the waterlogging process, the natu-
sult of air-drying. Both W and W disinte- ral color of the wood was altered, resulting
grated to the point that accurate physical in a slightly grayish cast. The wood, however,
measurements were impossible to obtain. was dimensionally stable after treatment in
The conservation of W and W was the silane. Experimentation indicates that
considered successful as their physical dimen- resins created as the result of condensation
sions, surface textures, and individual char- can preserve even badly waterlogged wood
acteristics were accurately maintained. Lower very well. As C. V. Horie and others have sug-
rates of shrinkage were observed in the gested, conservation strategies using silicone
CR--treated W wood sample, and it was oils are not reversible.10 However, over time,
lighter in color than W. The W sample was the solubility of many adhesives and con-
darker due to the slightly higher rate of solidants currently in common use in con-
shrinkage that occurred as a result of treat- servation are also aected, rendering these
ment in Q-. techniques nonreversible as well.11
CR- is % pure, with a % addition of More experimentation is needed to deter-
alcohols. Q- is a less-rened solution mine if the addition of small percentages of
containing approximately % CR- mixed silicone oils might increase the bulking abil-
with organic solvents and trace amounts of ity of CR- and Q-, eectively reducing
dimethyldimethoxysilane. Accordingly, the shrinkage. Adding a small percentage of a low
resin-forming capabilities of the Q- so- viscosity silicone oil to Q- may increase
lution are insucient to preserve the dimen- its bulking ability despite its high alcohol con-
sional characteristics of the wood. This is tent, making it an eective treatment.
 CHAPTER 4
60

Leather

Preservation

L
eather recovered from archaeological lubricates collagen bers and prevents shrink-
sites presents a variety of challenges for age. Accordingly, PEG treatments based on
the conservator. Regardless of their the incremental addition of PEG to an aque-
condition, leather artifacts reveal something ous solution in which the artifact is immersed
about the technology and craftsmanship used have been the most common means of pre-
in their creation. Tool marks may indicate the serving waterlogged leather. In combination
level of skill of the cobbler. Wear marks may with freeze-drying technology, PEG treat-
help characterize the occupation and auence ments have been very successful in preserv-
of the owner. ing the diagnostic attributes of waterlogged
leather artifacts.
Mold, mildew, fungi, and exposure to ul-
Archaeological Leather traviolet light and airborne contaminants
such as sulphur dioxide all act to undermine
Whether from a terrestrial or marine site, ar- the integrity of leather. Singly, or in combi-
chaeological leather is aected by numerous nation, these factors complicate the process
environmental factors that act to compromise of stabilization. The often delicate condition
its general stability. Leather recovered from of archaeological leather necessitates that all
land sites may be brittle if the relative humid- destructive mechanisms be neutralized prior
ity (RH) is less than %. Often, embrittled to treatment. In most cases, archaeological
leather will crack and shrink excessively. In leather requires some degree of cleaning be-
dry and desiccated leather, collagen bers, fore conservation procedures are undertaken.
which are polymers made up of amino acids, Badly deteriorated leather, however, may be
may become rigid due to a lack of moisture. too friable to clean, and it may therefore be
These artifacts need careful rehydration to impossible to remove stains without causing
prevent continued cracking and deteriora- extensive damage. In such cases, it may be
tion. Dressings and topical applications of best to focus on stabilizing the leather.
lanolin and cedarwood oils have been used
to lubricate leather. To regain exibility in dry Variables to Consider
leather, conservators have used glycerol, sor-
bitol, or a wide range of molecular weights Many factors need to be considered before
of PEG. attempting to conserve leather. These factors
Wet leather also requires careful treatment may be beyond the abilities of the conserva-
to prevent surface drying and hardening of tor to determine but nonetheless play a role
collagen bers. Excessive hydration may lead in the state of deterioration aecting archaeo-
to the loss of tannins. In their absence, water logical leather. Determining genus and spe-
cies of the animal may be important when treated hide. Vegetable tanning processes pro-
trying to assess the degree of exibility and duce polyphenols, which readily form hydro-
color desirable in a preserved piece of leather. gen bonds with carbon atoms associated with 61
The degree of use wear may also be an im- available hydroxyl groups within the leather.
portant consideration. Apart from being di- For this reason, vegetable tanned leather, in
agnostic features of the artifact, extensive many cases, deteriorates rapidly as hydro-
scu marks and surface abrasion may help lyzation interrupts the chemical equilibrium
determine a suitable treatment strategy. of the leather. Bark from oak trees is still com-
Artifact association or proximity is an- monly used for noncommercial tanning of
other consideration when developing a con- hides; indeed, the bark and roots of many
servation strategy for leather. Archaeological plants have tannins that are eective in sta-
leather closely associated with large deposits bilizing proteins.
of organic materials may be heavily degraded For the most part, twentieth-century
due to fungus and decay, whereas leather in leather was tanned using chromium tanning
close association with iron artifacts may be processes; it is sometimes called mineral
heavily concreted or impregnated with min- tanned leather. Chromium salts, which act
erals and oxides. to reinforce collagen molecular chains, also
Initial means of tanning used to preserve give leather a chemical resistance to many
the leather may be a factor in an artifacts state agents of decomposition including molds
of deterioration. This may also help deter- and fungi, as well as airborne contaminants
mine viable treatment strategies for preserv- producing sulfuric acid. In comparison, veg-
ing the artifact. The earliest suspected means etable tanned hides are more susceptible to
of preparing animal hides were concerned airborne contaminants since the cross-linking
with slowing the processes of putrefaction. of proteins is not as complete as the chemi-
Besides hides, other organs including the cal transformation and stabilization induced
stomach and bladder were utilized. Brain tan- by chromium processing, which makes hides
ning was also common. In many cases, the somewhat hydrophobic.
brains of the animal were considered su- Regardless of the method used for tanning
cient to tan the whole useable hide of the leather, hides share one common character-
animal. Smoking and close skinning tech- istic. Collagen bers, which are brous pro-
niques, which removed thick layers of fat, teins that give leather its exibility, require
were also commonly used to prepare hides. natural lubrication for the hide to remain
By removing the thick layers of subcutane- supple. These same bers require chemical
ous tissue or using smoke to cure tissues, restructuring to prevent breakdown of the
hides could be eectively preserved. bers and putrefaction. Hydrolyzation acts
Historic period hides and leather were also to break down collagen chains, thus reduc-
commonly prepared using vegetable tanning ing a skins ability to remain supple once
methods. This incorporated the use of natu- water has been removed from the hide. Wa-
ral tannins found in the bark and roots of terlogged leather usually becomes increas-
numerous plants. Hides were soaked in baths ingly soft as the process leaches tannins and
of bark and water in order to infuse them fats from its matrix. Accordingly, determina-
with tannins that combine easily with col- tion of the physical state of deterioration can
lagen and proteins. Tannins replace water in be dicult. High pH levels may indicate the
the hide and reinforce side chain structures, presence of oxides in degraded leather. Nu-
essentially cross-linking protein molecules, merous tests can be made on wet leather.
making them stable. They are eective anti- Generally, these tests measure characteristics
oxidants and help maintain exibility of the or qualitative aspects. Regardless of the state
 of the artifact, leather removed from a wet Many inexperienced conservators hold the
environment must remain wet until treated. tip of the air scribe at too sharp an angle,
62 which directs the force of the stylus directly
into the artifact. By working at a shallow
Cleaning angle to the line of the artifact, the force of
the stylus is deected, reducing stress. With
Waterlogged leather recovered from ship- practice, an air scribe or pneumatic chisel can
wreck sites can pose many challenges. Leather be operated with nesse, allowing very ne
from these sites may be covered with con- and detailed mechanical removal of con-
creted deposits, which are the product of creted materials from leather and other deli-
oxidation of metals. Concreted material and cate artifacts.
impacted soil on leather from land sites
should be removed carefully using soft
wooden dowels and brushes. In some cases, Chemical Cleaning
leather may be too soft and fragile to clean. It
is best to treat these artifacts with a minimum Numerous organic acids can be used to re-
of cleaning and handling. Often, additional move corrosion and stains on leather recov-
cleaning can be successfully accomplished ered from marine environments. Citric acid,
after conservation, when the leather is in a oxalic acid, disodium EDTA, tetrasodium
more stable state. EDTA, ammonium citrate, and dilute hydro-
To prevent structural damage to badly chloric acid (%) are a few of the most
deteriorated leather, cleaning should be con- common treatments. Citric acid is a slow and
ducted using a soft brush, while the artifact gentle acid; hydrochloric acid tends to work
is submerged in freshwater. Leather in good more quickly. Whatever means are employed
condition can be cleaned by placing it on a to remove corrosion, concretion, and stains
clean working surface and using a gentle, di- from leather, it is essential to extensively rinse
verted ow of water to keep it wet. This al- the artifact in fresh running water baths af-
lows the conservator to inspect the artifact ter treatment. If chemicals are not completely
more closely. A direct ow of water on leather ushed from the matrix of the leather, there
should never be used since this may cause is a risk of chemical reactivity among tannins,
irreversible damage due to erosion and pit- fats, collagen bers, and the acids and neu-
ting of the artifacts surfaces. tralizing agents used to treat the artifact. Ac-
While it is best to remove concretion cordingly, chemical cleaning of waterlogged
from the surfaces of leather using soft or desiccated leather is often not the rst
wooden tools, heavier areas of concretion choice of many conservators.
can be removed using an air scribe. With a
gentle touch, using the vibration of the sty-
lus to break apart hard deposits, this instru- Treatment of Leather
ment can be more eective for removing
thick, concreted materials than wooden It is dicult and risky to attempt conserving
dowels. Care must be taken, however, since archaeological leather by controlled drying
these same vibrations can undermine the or freeze-drying methods without the aid of
artifacts integrity. It is often better to hold a a replacement bulking agent to maintain the
small piece of leather or place it on a soft artifacts physical attributes. As water is re-
bedding of cloth to absorb vibrations from moved from the artifact, collagen bers are
the air scribe. This will substantially reduce drawn together, causing hardening, shrink-
stress on the artifact. age, and disgurement. To prevent these
problems, water within the matrix must be bacteria. Use of organic solvents, however,
displaced with a suitable bulking agent that displaces water within the matrix of the
will enable the artifact to withstand the dry- leather. Loss of water as a lubricant in the 63
ing process. Over the last few decades, PEG leather results in stiness. In some circum-
has been widely used for the treatment of stances, this is useful during treatment since
leather. Through incremental additions of it adds a degree of rigidity to the artifact, pre-
PEG to water, the cells and voids in the ma- venting damage or loss of information due
trix of leather can be suciently supported to disarticulation. The state of deterioration
and lubricated to prevent the shrinkage and and the need for internal stabilization are
hardening that occur as the result of drying. prime considerations when deciding on a
Anyone who has ever worked with wet treatment strategy for leather.
leather has observed that the artifact remains Treating waterlogged leather with PEG
soft and supple while immersed in water. followed by freeze-drying to remove water
When moved to a bath of organic solvent, from the artifact has proved to be a highly
such as ethanol, the leather becomes stier successful conservation strategy. Because
and harder to manipulate. This change in PEG is soluble in water, it is an attractive
malleability occurs as water is replaced by the treatment method for leather. To a lesser de-
solvent. Progressive dehydration has been gree, PEG is soluble in methanol. Because of
successfully used to conserve some wet leather this, badly deteriorated leather can be stabi-
artifacts. By taking the artifact through a se- lized in a solution of water and alcohol, which
ries of baths, starting with ethanol and work- helps to give the leather some degree of ri-
ing up to fresh baths of acetone, it is possible gidity. At the Preservation Research Labora-
to dehydrate leather into a relatively stable tory, treatments usually involve incremental
state. additions of PEG to water to form an aque-
The problem with such treatment strate- ous solution in which the leather remains
gies, as Cronyn notes, is that there is a chance immersed until a high enough percentage of
of shrinkage and distortion in the artifact.1 PEG is introduced into the matrix of the
Degree of degradation, biological activity, leather to allow controlled, slow air-drying.
rate of water-solvent exchange, variance in Lower molecular weight PEGs (PEG
surface tension, and chemical reactivity with to ) are clear and somewhat viscous at
insoluble deposition on the surfaces of the room temperature. These products are ideal
artifact will aect dehydration and prepara- for use in room temperature treatment of
tion of the leather for continuing conserva- leather. After reaching % concentration
tion work. of PEG in water, the artifact can be removed
from the solution, surface-dried of excess
solution, and allowed to slowly air-dry in a
PEG/Air-Drying Treatments controlled environment. Controlled drying
will prevent excess pooling of PEG on the
While immersed in water, the bers of leather surfaces of the artifact. Surface deposition
are supple and exible. This works to the of PEG can be removed by daubing the
advantage of the conservator during treat- aected areas with a damp, lint-free cloth.
ment since the leather object is relatively Surface pooling, however, may be an indica-
easyto work with as compared to leather tion of more serious problems aecting the
stored in alcohol or water/alcohol solutions. stability of the artifact. If the leather is al-
Often a mixture of alcohol and water are used lowed to dry too quickly or if temperature
for the long-term storage of leather since such and humidity are not regulated, PEG will
solutions inhibit the growth of fungi and migrate to the surfaces of the artifact and,
 over a short period of time, shrinkage may long-term controlled storage in a standard
occur. Controlled, slow air-drying will pre- freezer. Humidity control is limited using a
64 vent most of these complications. Generally, conventional freezer system; nonetheless, ice
however, as much as % shrinkage has been crystals sublimate as gases, resulting in a sta-
associated with the PEG treatment of water- bilized artifact. With more elaborate freeze-
logged leather. drying systems, the vacuum, temperature,
and relative humidity can all be monitored
and regulated. Controlled environment pro-
Freeze-Drying PEG-Treated Artifacts cessing and monitoring of humidity allow the
conservator to determine when the treatment
Generally, there is no steadfast rule for de- process is complete.
termining the percentage of PEG that should The use of larger and more elaborate
be added to water to pretreat leather before freeze-dryers allows the conservator more
freeze-drying. Experience is the best guide control over the processing of artifacts (g.
since state of degradation, possible fungal .). Artifacts are usually frozen for a period
problems, and presence of oxides all aect the of time at very low temperatures (C)
potential for preserving leather. Conservators prior to placing them into a freeze-dryer.
use a wide range of pretreatment strategies Drying can then be accomplished in many
in preparation for freeze-drying. In the case ways. Low-temperature freeze-drying takes
of badly deteriorated leather, a % or higher a longer period of time, while near room tem-
aqueous solution of PEG may be necessary perature processing will be completed com-
to stabilize the artifact. Leather in good con- paratively quickly. Many successes have been
dition may require a lower percentage pre- reported using ultralow and near room tem-
treatment. perature methods, so it is best to experiment
Aqueous solutions of PEG may require the with a particular assemblage of equipment
use of biocides to control biological activity, to determine how to reliably obtain consis-
especially at lower percentages. Because it is tent results. Regardless of which strategy has
impossible to accurately assess the ramica- been prescribed for an assemblage of arti-
tions of their use to control microbial activ- facts, low vacuum is maintained throughout
ity, they should be closely monitored and processing.
applied sparingly. Many biocides are mildly Other processes have been successful for
toxic, and some people have a mild reaction the treatment of wet leather. J. M. Cronyn
when they come into contact with these reported success in using freeze-drying pro-
agents. Ortho-phenylphenol, commonly cessing with PEG in -methylpropanol.2
known as Dowicide , has been widely used Several nonpolar solvents, such as toluene
since it appears to be an eective deterrent to and kerosene, have also been reported to give
microbial activity. Tests indicate that it is less good results. Because many of these solvents
of a biological hazard than other commer- are ammable, care must be taken in their
cially available biocides. use. Be aware that changes in pressure and
After pretreatment, pooled PEG must be dispersion of fumes will alter the explosion
carefully removed from the surfaces of the point of some solvents.
leather before freeze-drying. This will ensure In recent years, PEG has fallen into disfa-
that the freeze-drying process is uniform. All vor with many conservators because of
conservators seem to have a freeze-drying changes in environmental laws, which vary
strategy that works well for their range of from one country to another. Disposal of
artifacts and levels of expertise. Many arti- PEG has become an issue for many labora-
facts have been successfully preserved using tories since the introduction of too much
PEG solution into a community septic sys-
tem can stop microbial activity, thereby eec-
tively halting the processing of waste water. 65
In the early stages of experimentation with
PEG, the material was considered environ-
mentally friendly and safe for laboratory use.
Today, conservators realize that, in large
quantities and at the higher level of concen-
tration used for the treatment of artifacts,
precautions need to be taken when handling
PEG. Certainly, protective clothing, goggles,
and limited exposure to PEG fumes are man-
datory.

PEG and Other Polymers

Fig. .. Modern freeze-
Polyethylene glycol is a polymer. At this point dryer used for PEG-freeze-
in the history of conservation chemistry, a drying large artifacts.
great deal of experimentation is required to
understand the complex relationships among
polymers, organic materials, and the chemi- disgurement of the artifact. Because water-
cal reactivity of resins and organic materials. logged leather is relatively porous, however,
Many processes once thought to be revers- almost all molecular weights of PEG can be
ible are really not reversible at all. PEG treat- used.
ments t into this category. Another example Dry and desiccated leather, on the other
is the treatment of organic materials with ac- hand, may be more receptive to treatment
etone rosin. When completed, this treatment using lower molecular weight materials,
essentially encapsulates the entire artifact, which tend to be less viscous. Higher molecu-
resulting in the restructuring of the artifact lar weight PEG tends to be white in color
into what has been described as a single resin and has the consistency of paste at room tem-
molecule. Experimentation, advances in perature, whereas PEG remains hard.
chemistry, and the advantage of several de- The elevated temperatures required to melt
cades of accumulated data suggest that it higher molecular weights of PEG make them
would be better if conservators dispense with more dicult to use with delicate leather.
reversibility as a core issue in selecting con- Conservators have successfully treated
servation strategies and, instead, focus on the thin strips of leather using topical applica-
issue of re-treatability of artifacts. tions of PEG. Many vessels have been spray-
Initially, many conservators attempted to treated over long periods of time to facilitate
treat waterlogged leather with high concen- the absorption of PEG into large timbers.
tration aqueous solutions of PEG. This pro- Because of their generally smaller size, how-
cess has been largely abandoned since, like ever, the conservation of leather artifacts is
wood, leather tends to absorb the bulking usually accomplished by immersing them in
agent more eectively when applied incre- water and using incremental additions
mentally. It is possible that high concentra- of PEG.
tions of PEG do not permit a balanced dis- Much of the popularity of PEG stems from
placement of water for PEG, resulting in its ease of use and relatively low cost. Because
 PEG dissolves in water, certain considerations Passivation Polymer Processes
must be kept in mind. Artifacts require proper
66 curation to prevent damage. Once treated, Passivation Polymers are part of a series of
they remain hygroscopic and require a tem- chemistries and techniques developed at
perature- and humidity-controlled environ- Texas A&M University for the purpose of
ment to remain stable. Over time, the artifact archaeological preservation. These processes
will become acclimatized to a museum envi- were developed in response to the need for
ronment by either absorbing or giving o artifacts that were more durable and stable
moisture to equalize with environmental after conservation. Passivation Polymers can
conditions in the museum. The absorption be used for the treatment of both waterlogged
of humidity from the environment results in and desiccated leather artifacts. Treatment
a loss or alteration of the original bulking procedures dier for artifacts recovered from
characteristics of the PEG. This, in turn, com- marine sites since it is essential to displace
promises the stability of the artifact. In some the water before introducing a suitable Pas-
cases, PEG-treated artifacts remain greasy sivation Polymer and cross-linker. The case
after treatment. Many conservators have study presented below concerns a leather ar-
avoided this condition by treating leather tifact that was recovered from La Belle, which
with blends of PEG, resulting in slightly less sank o the Texas coast in .
supple leather.
It is also important to control exposure
of organic materials to ultraviolet light. Over Case Study: A Successful Treatment
long periods of time, this range of the light Strategy for a Waterlogged Shoe
spectrum causes damage to leather and acts
to chemically alter the nature of the PEG. The Upon arrival at the Preservation Research
chemistry of leather treatment is complex, Laboratory, one of the shoes recovered from
and because some degree of bonding occurs the Texas Historical Commissions excava-
with the introduction of hydroxil-rich PEG tion of the vessel La Belle was placed into
to the leather, this process cannot be com- freshwater for extensive rinsing. For several
pletely reversed. This does not preclude the months, the shoe was allowed to desalinate
Fig. .. A nearly complete fact that artifacts treated with PEG can often along with numerous other artifacts in a vat
waterlogged shoe. be successfully re-treated. of freshwater that was changed on a regular
basis. Before conservation was begun, the
shoe was photographed and thoroughly
documented.
The most common methods of preserv-
ing waterlogged leather include variations of
impregnation of the artifact with PEG. Once
completed, the PEG will prevent the struc-
turally deteriorated matrix of the leather
from collapsing during the preservation pro-
cess. While these processes are eective con-
servation strategies, they are time-consuming
and may produce artifacts that require spe-
cial curation and handling.
PEG dissolves in water, making it a con-
venient agent to ll the damaged cellular
structure and lubricate the ber bundles of
the artifact. Because PEG remains miscible slight vacuum ( Torr) applied to the arti-
in water after an artifact has been treated, fact to ensure maximum removal of free-
however, long-term stabilization can pose owing water. Throughout this last step of 67
problems. Changes in humidity, tempera- water/acetone exchange (WAE), a piece of alu-
ture, or both can cause the PEG to become minum screen was friction-tted over the
unstable. In severe environmental condi- shoe, pressing against the sides of the con-
tions, intercellular movement of PEG within tainer housing the artifact. This screen
and between the cellular structures of an ensured that the artifact remained fully im-
artifact can cause signicant structural dam- mersed in acetone throughout the dehydra-
age. Intracellular migration of PEG can re- tion process. (See g. ..)
sult in surface pooling and, eventually, the loss After the WAE process was complete, the
of diagnostic attributes associated with an shoe was placed into another container and
artifact. immediately immersed in a PR- Passivation
Research at Texas A&M University has Polymer solution with a % addition of
culminated in the development of eective CR- by weight. As before, a piece of fric-
new treatment processes for preserving wa- tion-t aluminum screen was placed over the
terlogged and desiccated leather artifacts. artifact in solution to ensure that it remained
These processes are easily implemented us- completely immersed throughout the ac-
ing basic equipment commonly found in a etone/silicone oil exchange (ASE) process.
modestly equipped conservation facility. While in this solution, the shoe was placed
Most important, these treatment strategies into a bell jar vacuum chamber, and at room
stabilize objects, requiring little controlled temperature, a vacuum of Torr pressure
curation. was applied. When small bubbles were
To demonstrate a basic process of silicone steadily streaming from the artifact, the
oil preservation, a waterlogged shoe dating vacuum pump was turned o and the arti-
from the mid-seventeenth century was pre- fact was left under pressure for hours. (See
served (g. .). An initial inspection of the g. ..)
shoe indicated that much of the upper struc- After hours of ASE at pressure, valves
ture and heel were missing. The insole, sole, were opened, returning the solution to am-
and toe portions were in excellent but friable bient pressure. The artifact was left in solu-
condition. Prior to treatment, there was a tion for hours at room temperature and
slight separation between the insole and sole ambient pressure. It was then removed from
sections. Much of the stitching, however, was
in good condition and appeared to be hold-
ing the sections together quite well. Prior to
treatment, the shoe was placed into a large
container and rinsed in running fresh tap
water for several days to remove soluble salts
and other debris. Titration testing was used
Fig. .. Apparatus used
to determine the length of the rinsing treat-
for dehydrating water-
ment.
logged leather:
The shoe was then immersed in fresh ac- (A) vacuum chamber;
etone. After three days of passive water/ac- (B) container with fresh
etone exchange at room temperature, it was industrial grade acetone;
placed into a new container of fresh acetone. (C) aluminum screen over
While immersed in this acetone bath, the shoe shoe; (D) shoe immersed
was placed into a vacuum chamber with a in acetone.
 to lightly pat the surfaces of the leather. This
helped to remove pooled areas on the sur-
68 faces. Lint-free cloths with several drops of
CT- catalyst were then used to wipe the
surfaces of the leather. A thin, even coating
of catalyst was applied to these surfaces. Af-
ter approximately minutes, clean cloths
were used to wipe any remaining catalyst and
silicone oils from the shoes surfaces. The ar-
tifact was lightly polished using dry, soft, lint-
free cloths. The shoe was then placed into a
large Ziploc bag. A small piece of wadded
cloth, dampened with several drops of cata-
lyst, was placed in the bag with the shoe. Af-
ter the bag was sealed, the artifact was left in
close contact with catalyst fumes for an ad-
ditional hours. Exposing an artifact to
catalyst fumes ensures that polymerization
will continue into the deep matrix of the
leather such as the thick sole and heel.
After treatment, the shoe was removed
from the bag and allowed to sit in a well-
ventilated fume hood for several hours. Be-
Fig. .. Shoe in vacuum the silicone oil solution and placed on an alu- cause we were not sure how friable the arti-
chamber, housed in a minum screen; free-owing silicone oil so- fact was during treatment, we did not wipe
container with a venti- lution was allowed to drain from it. Note that, the surfaces of the leather as thoroughly as
lated cap to prevent at this point, exposure to air is no longer a we could have. Posttreatment was conducted
splashing while in
concern and the conservator can take time using wooden dowels and a very soft brush
treatment.
to situate or arrange the artifact before con- to remove small areas of pooled polymer
tinuing to the last stage of polymerization. along the line of stitching on the insole of
Fig. .. Bottom view of After allowing the polymer solution to the shoe. Posttreatment evaluation revealed
a nearly complete shoe. drain from the shoe, lint-free cloths were used that no measurable shrinkage occurred as the
result of treatment. The diagnostic features
of the shoe have been maintained to this day,
and the artifact does not require special
curation or handling. (See g. ..)
Accelerated aging tests have indicated that
archaeological leather samples preserved us-
ing these technologies are stable and not
damaged by uctuations in temperature,
humidity, or ultraviolet radiation. Data col-
lected from extensive testing of this specic
Passivation Polymer indicate that the artifact
will remain unchanged and completely stable
for at least years. The polymer will re-
main supple beyond this time frame and, if
needed, the artifact can be re-treated.
Passivation Polymer Treatment for present, soft wooden dowels or dental tools
Desiccated Leather may be required to clean these areas. Stains
are a common problem associated with 69
Dry and desiccated leather artifacts are often leather artifacts from land sites due in part
fragile, making excessive handling a problem. to their close association with other artifacts,
In many respects, waterlogged leather is easier absorption of minerals in the soil, and bac-
to work with since water acts to lubricate the terial and fungal activity. All these problems
collagen bers. Leather recovered from cave must be addressed to ensure long-term sta-
sites and other dry environments is often bilization.
more dicult to restore since the lack of Flat pieces of dry leather can be treated
moisture causes collagen bers to shrink. using minimal equipment; and since it is not
Once this has occurred, it is sometimes im- necessary to dehydrate the leather before con-
possible to soften the leather and reshape it. tinuing treatment, processing can be accom-
Dry and desiccated artifacts, nonetheless, re- plished in a short period of time. Because dry
quire conservation to prevent further dete- leather is often brittle, it is helpful to place
rioration. the artifact between two sheets of glass to
Desiccation and dry salty conditions may prevent stresses caused by handling. Small
actually help preserve some forms of leather. elastic bands or binder clips can be used to
Cronyn suggests that rawhide, tanned and sandwich the leather between sheets of glass
semitanned materials, and parchment may (g. .).
be better preserved due to desiccation or Leather shoes and artifacts that are not
close association with a salty environment.3 inherently at cannot be placed between
Salt, he points out, acts to desiccate leather sheets of glass prior to treatment. A number
and inhibit damage caused by microorgan- of precautions can be taken, however, to en-
isms. In many situations, disintegration of sure that these artifacts will be correctly
the leather may be advanced to the point that shaped and aesthetically accurate. Loosely
it must be consolidated to prevent further crumpled gauze placed inside a shoe works
loss of surface details. Polymers have proved well to maintain its shape during treatment.
to be an eective agent for preservation of The presence of such an open-structured
dry leather and for the consolidation of material does not aect the ability of a Passi-
leather that has become fragile over time. vation Polymer to penetrate into the leather.
In many cases, the wide variety of Passiva- Expandable polyethylene mesh bags, similar
tion Polymers being used for artifact con- to those used in the produce section of a su-
servation enables the conservator to select permarket to hold fruits and vegetables, are
polymers with suciently long structures an ideal way to keep small and fragmented
that will increase the exibility of dry, sti artifacts together during treatment. Because
artifacts. of their open-weave design and inherent ex-
Unlike artifacts recovered from marine ibility, these bags have proved to be extremely
environments, much of the leather from ter- useful in handling fragile artifacts.
restrial sites does not need dehydration prior
to treatment with polymers. Leather recov- Water/Acetone Exchange
ered from these sites, however, usually needs
surface-cleaning. Often gentle wiping with a Prior to treatment with polymers, water must
lint-free, damp cloth will remove encrusted be removed from the leather. To prevent
surface dirt. Fortunately, leather from terres- damage, the leather must be slowly dehy-
trial sites is seldom encased in harder con- drated, starting rst with a series of ethanol
creted materials. When slight encrustation is baths and working up to at least two baths of
 acetone. This will appreciably reduce shrink- Rate of evaporation of acetone is only one
age and problems of hardening related to the factor that can be controlled. Viscosity, the
70 locking of bers as water is displaced within length of the polymer backbone or centis-
the leathers matrix. While immersed in wa- toke of polymers, and the temperature at
ter, waterlogged leather remains supple. The which acetone/polymer exchange is con-
initial stages of treatment are therefore the ducted are important in controlling the end
most opportune times to adjust and shape product. Generally, shorter chain polymers,
the artifact. In successive baths of ethanol such as PR- and PR-, are easily exchanged
and acetone, the artifact will become more into the matrix of leather because of two fac-
rigid. Once the artifact has been dehydrated tors. First, their relatively small size makes it
in acetone, it should remain in the nal bath easier for them to penetrate into the open and
until a solution of polymers is prepared. Re- porous structure of waterlogged leather. Sec-
moving it from acetone and allowing it to ond, because of their small size, these poly-
quickly air-dry will cause irreversible damage. mers tend to permeate more quickly than
Dry and desiccated leather, on the other larger centistoke polymers. Leather treated
hand, needs only to be surface-cleaned prior with smaller centistoke polymers tends to be
to treatment in polymers. more rigid than leather treated with larger
centistoke polymers.
Choosing a Suitable Polymer or PR- has a relatively long chemical back-
Combination of Polymers bone compared to PR- and PR-. Artifacts
treated with PR- tend to be more supple
Polymer technology has progressed to the after treatment. By blending polymers, the
point that conservators have a wide range of conservator can design a treatment strategy
viscosities and controllable attributes to specically geared to treating problem as-
choose from for the treatment of artifacts. pects of an individual artifact. For example,
Just like PEG, polymers can be mixed in in- in a badly deteriorated, waterlogged shoe,
nite combinations to allow the conservator blends can be mixed to ensure that the sur-
control over such attributes as exibility, ten- face integrity of the leather is preserved as
sile strength, texture, and degree of consoli- well as the thicker, more stable areas of the
dation that will be imparted to an artifact. artifact.
At present, acetone/polymer exchange is
the mechanism most commonly used to im-
pregnate an organic substance with polymer/ An Effective Treatment for Dry Leather
cross-linker solutions. Normally, acetone in
the matrix of an artifact will evaporate quickly Dry leather does not require dehydration in
at room temperature and ambient pressure. acetone before treatment. If the artifact has
Unless the artifact is immersed in a suitable been surface-cleaned using water, air-drying
replacement media, the acetone will vapor- in a well-ventilated fume hood for at least
ize and, with nothing to displace the acetone, hours is necessary. The movement of air will
shrinkage and distortion will occur. In a ensure that sucient moisture has been re-
vacuum, the boiling point of acetone is re- moved. If the artifact is being treated between
duced, resulting in faster evaporation. Con- sheets of glass, the time required to ensure
trolling the rate of evaporation and the rate dryness will be longer.
of displacement of acetone with a suitable Once the artifact is cleaned and prepared
bulking agent, then, is the key to successfully for treatment, it can be carefully immersed
maintaining the diagnostic attributes of an into one of the PR-series polymers to which
artifact. CR- has been added in the amount of %
by weight. Normally, it is best to friction-t a
piece of aluminum screen over the top of the
artifact in solution to ensure that it remains 71
immersed throughout treatment. Because
acetone is not being used as the driving
mechanism to transfer polymers into the
leather, a slight vacuum can assist the pro-
cess. If the leather is too fragile or deemed to
be unsuitable for withstanding vacuum pres-
sure, the process of impregnating the artifact
can be accomplished at room temperature
and ambient pressure. This process, naturally,
will take longer.
For sections of leather ranging from to
mm thick, it takes approximately min-
utes for the polymer solution to be adequately
absorbed. The use of vacuum will speed up
absorption. The process must be closely
monitored, however, because if air is drawn
Fig. .. Glass plates used
from the artifact faster than polymers can ll
to support fragile leather
the matrix of the leather, some distortion may Once the artifact is aesthetically correct during treatment with
occur. To avoid these problems, use vacuum and free of pooled polymers on its surfaces, silicone oils: (A) rubber
cautiously. Over a short period of time, all catalyst can be applied. This phase of treat- bands; (B) leather; (C)
the PR-series polymers will successfully pen- ment should be conducted in a well-venti- clear glass plates.
etrate dry and desiccated leather. lated area, preferably in a fume hood. Tran-
After treatment in a polymer solution, the sient vapors may contaminate open vats of
artifact should be allowed to drain for at least polymer/cross-linker solutions that are in
hours. Because the polymer solution can close proximity to the workspace. The con-
be reused several times, place the artifact on servator must also avoid personal exposure
aluminum mesh screen over a large con- for health and safety reasons.
tainer, so that runo polymers can be recov- CT- is a good catalyst for leather artifacts
ered. During this phase of treatment, it is because it can easily be applied to the surfaces
wise to spend as much time as needed to of the leather using soft cloths or cotton swabs.
position and pose the artifact and to ensure Regardless of the means of application, the ar-
that everything is done to make the leather tifact should be covered in a thin, even layer of
natural looking. Using soft lint-free cloths catalyst. After to minutes, use a soft cloth to
or gauze to gently daub the surfaces of the remove the catalyst and any polymer solution
artifact will remove pooled polymer solution that may have pooled on the surface of the ar-
and prepare the surface for catalyzation. tifact as a result of chemical reactivity with the
Once catalyst has been applied to the catalyst. Successive soft cloths can be used to
leathers surfaces, cleanup will be a little more gently wipe the surfaces of the artifact. In a
labor intensive. In situations where the short period of time, the leather will have a
leather is very fragmented or fragile, it may natural texture and feel. Even though most of
be better to do minimal cleanup before the catalyst has been removed from the sur-
catalyzation with the intention of working faces of the artifact, a chemical reaction with
more with the artifact after it has been stabi- the polymer/cross-linker-rich matrix of the
lized and is capable of being handled. leather will continue for approximately
 hours. After topically applying catalyst to Once suciently drained, the elastic bands
leather, it is often advisable to place it into an or binder clips holding the artifact in place
72 airtight environment with a small dish of should be removed and the top glass plate gen-
CT- and allow the catalyst vapors to con- tly slid away. Blotter material or soft gauze can
tinue the process of catalyzation. The pro- then be used to gently daub the surfaces of
cess is the same as that suggested for delicate the leather and surrounding surfaces of glass.
artifacts that cannot withstand handling. Once the top surface is dry, a clean glass plate
In situations where the leather is too frag- should be placed over the top surface of the
ile to withstand surface-wiping, catalyzation artifact and the entire artifact turned upside
can be accomplished using vapor deposition. down so that the bottom surface of the arti-
The simplest means of initiating catalyzation fact and covering glass plate are on top. This
is to place the artifact and a small dish con- glass is then removed and the process of clean-
taining a few grams of CT- into an airtight ing the bottom surface of the artifact is com-
environment and allow the artifact to remain pleted. Once cleaned of free-owing polymers,
in treatment for at least hours. After the artifact should be transferred so that it is
hours, the dish of catalyst will need to be re- sandwiched between thin sheets of lint-free
placed with a fresh one. To increase the vol- cloth between two sheets of clean glass. After
ume of vapors, the containment chamber can rubber bands are applied to hold the encased
be placed into a ventilated warming oven set artifact together, vapor catalyst deposition can
at C. If additional treatment is deemed be completed. The process sounds involved
necessary, new catalyst needs to be placed into but is actually quickly accomplished with
the containment chamber every hours. minimal stress to the artifact.
After vapor deposition, all leather should
be placed in a well-ventilated area for at least
hours. This will ensure that the residual Storage and Display of Leather Artifacts
smell of catalyst disappears. Exposure to air
will also ensure that the polymer is set and The excavation, documentation, recovery,
the artifact is completely stable. and conservation of artifacts are costly and
time-consuming. All the labor and research
invested in an assemblage can be lost in a rela-
Suggestions for Treating Leather tively short period of time if the artifacts are
between Sheets of Glass not properly stored and displayed. It is, there-
fore, important to have a detailed plan for
Glass sheets make it easy to work with fragile the long-term storage of artifacts before treat-
leather and, with a few precautions, the pro- ment is initiated. Chromium tanned leather
cess is highly successful. Because the surfaces is, in part, more resilient to some atmospheric
of the leather are obstructed by the glass pollutants than vegetable tanned leather. But
sheets, the process of acetone/polymer ex- other factors play a role in the safe storage of
change will be slower; the use of vacuum will leather. Treatment methods are also impor-
speed up the process considerably. Once poly- tant. Because of the water miscible nature of
mers have been introduced into the leather, PEG, artifacts treated using this method re-
the artifact can be removed from the poly- quire a more controlled environment than
mer solution and allowed to drain of free- leather treated with Passivation Polymers.
owing solution. This process, too, will be Other considerations are relative humidity,
slower since the viscosity of the polymer so- exposure to natural light, microorganisms,
lution and the available free space between airborne pollutants, and handling by re-
the sheets of glass will work to slow drainage. searchers and museum personnel.
Relative Humidity Temperature

Most organic artifacts are hygroscopic. Many chemical reactions are slowed or in- 73
Waterlogging processes act to break down the hibited by keeping organic materials at lower
brous structures of leather as well as remove temperatures. Temperatures must never be so
natural lubricants that protect bers from low, however, that changes in humidity cause
hydrolysis. Hydrolysis increases the amount problems with condensation. It is important
of water in the matrix of the leather, which to keep artifacts at a constant, cool tempera-
ultimately must be replaced with an agent ture below the melting point of the bulking
that can stabilize the leather without its los- agents and consolidants used to treat the as-
ing physical integrity. Relative humidity is an semblage. PEG-treated artifacts are most vul-
important consideration for the storage of nerable to temperature changes because as
leather since elevations in humidity can pro- temperatures are decreased, the relative hu-
mote microbial activity and new problems midity increases, causing potential problems
in the stabilization of the artifact. in the stability of the artifact. Artifacts pre-
Maintaining a constant relative humidity served using Passivation Polymer technologies
in a display environment will prevent ab- are substantially more stable and are unaected
sorption and release of moisture from by changes in temperature. Although stability
leather. In many cases, leather treated with is increased, however, good practice dictates
PEG can swell if relative humidity is too high that all artifacts be treated with the same con-
(% or more). The miscibility of PEG- cerns for safety and stability.
treated leather encourages beading and sur-
face pooling of the bulking agent as mois- Light
ture in the surrounding environment is
absorbed into the leather. Ideally, maintain- Exposure to light is one of the most poten-
ing a constant relative humidity is best for tially dangerous problems for artifacts in
artifacts, and it is a costly environmental fac- museums. Ultraviolet light causes oxidation
tor to ignore. RH is computed using the fol- and sets a number of chemical reactions in
lowing simple equation: motion that are not fully understood. Many
of the consolidants and adhesives used in the
Amount of H2O in air X 100 = preservation of leather are sensitive to light
Relative Humidity Maximum H2O air and, over a short period of time, become
can hold at a specific temperature brittle or lose their desired characteristics.
 CHAPTER 5
74

Composite

Artifacts

C
omposite artifacts are generally made of wicker basket, heavily concreted with iron
two or more contrasting materials. Ex- oxides and iron shot, which was recovered
amples include iron cannon balls out- from La Belle.
tted with iron straps and a wooden base
called a sabot, and knives recovered from the
shipwreck La Belle consisting of hard wooden Case Study: Preservation of a
handles that contain sections of iron knife Composite Artifact Containing
blade. Basketry and Iron Shot
Often each component in a composite ar-
tifact requires a dierent treatment strategy. The silty bottom of Matagorda Bay has cre-
If this is the case, disassembling the artifact ated the ideal anaerobic environment in
into component parts may be essential for which the large assemblage of artifacts from
preservation. Two complications arise when La Belle have been well preserved. When re-
disassembly is required. Waterlogged artifacts covered, the wicker basket and iron shot were
are usually structurally compromised. Even indistinguishable, since oxides from the cast
the most precise job of disassembly may re- iron balls had combined with minerals and
sult in breakage or damage to one of the com- sediments to create a large formless block,
ponents. Also, when components are treated commonly referred to as a concretion. Apart
separately, they are exposed to treatment- from the general outline of a few of the larger
related structural changes that may inhibit shot, the only distinguishable feature was a
posttreatment reassembly. Wooden compo- large wooden barrel stave encrusted to the
nents, for instance, are more prone to swell- attest surface of the concretion.
ing during preservation. During treatment, Once the stave and layers of concreted
badly deteriorated iron may be prone to sur- material were removed, the attened outer
face spalling, resulting in the loss of metal. surface of a wicker basket was visible. Addi-
Even the most gifted conservator will en- tional removal of encrustation revealed that
counter problems in treating some compos- a bar shot was also concreted to the top sur-
ite artifacts. face of the block. Before an initial evaluation
One cannot overstress the point that sili- could be made, a pneumatic chisel was used
cone oils preservation techniques are not a to remove the bar shot from the outer sur-
cure-all for the conservation of all artifacts. faces of the artifact. Radiographs and me-
These technologies, however, do oer some chanical cleaning of a small area of the
benets for the preservation of delicate arti- concretions attest surface indicated the
facts concreted together with heavy iron ar- outline of one side and the base of a basket.
tifacts. Such is the case of a waterlogged From the X-rays, the basket appeared to be
intact but heavily permeated with oxides and lease of very low molecular weight silicone
concreted material. compounds during setting.1 In creating a
Inside the basket, small shot were clus- successful facing compound for the treat- 75
tered, held in place by impacted sediments ment of painted plasters, Pulga concluded
and concreted materials (g. .). The shot there was a direct link between the amount
ranged in size from . to . cm in diameter. of staining that occurred and the porosity
Because the outer surface of the basket ap- of the organic material being treated.2 Pulga
peared to be intact, the decision was made to and others have noted that shorter setting
clean the thin layer of encrustation encasing times reduce the release time of siliceous
this outer surface. To minimize damage to compounds and, therefore, reduce the risk
the fragile wicker basketry, a rubber cast of of staining. Since the basketry was thoroughly
the bottom was made. During treatment, this stained with minerals and surface concretion,
cast would act to support and protect the potential staining caused by siliceous com-
basketry while the shot were mechanically pounds during treatment was determined to
cleaned and removed from the interior of the be minimal.
basket.
Displacement Mechanism
Materials and Chemical Sequence

Because of the weight of the overlying con- In order to introduce a polymer/cross-linker

creted cast iron shot, high tensional strength solution into the basketry, it was rst neces-
and dimensional stability were two criteria sary to eliminate water in the wicker and sur-
used in selecting molding materials. HSIII rounding concretion and iron. Baths of indus-
RTV silicone elastomer compound (,. trial-grade acetone were used to thoroughly
MPA.S) was chosen. This silicone rubber is dehydrate the concreted artifact. Because of
ideal for capturing surface details. It is o- its volatility, acetone could easily permeate
white in color and not prone to hazardous or thin layers of encrustation to ensure that all
unplanned polymerization due to chemical surfaces of the basket were dehydrated prior
reactivity with oxides or other materials. to the application of the polymer. After dehy-
HSIII RTV has been widely used for restora- dration, a polymer/cross-linker solution could
tion castings because of its ability to capture be introduced into the matrix of the basketry
ne surface details and its long-term stability by a process of passive displacement of evapo-
and ease of use. rating acetone.
The weight of the shot and the fragility of For this project, PR-, a hydroxyl-ended,
the wicker necessitated the creation of a functional silanol was combined with Q-
stable, solid support for the basket that would cross-linker (% by weight) to form a low
not cause stains to the organic component viscosity polymer capable of displacing
of the artifact or cause additional cleaning
problems in the crevices of the basketry or
iron shot. During initial testing, HSIII RTV
did not react with other resins and casting
materials that might potentially be needed
as treatment progressed. Signicantly, this
casting material did not interact adversely
with the wet surfaces of the wicker and iron. Fig. .. Concretion with
Concerns have been raised that staining shot visible on upper
can occur to porous surfaces, due to the re- surface.
 acetone in the matrix of the basketry. Q- polymerization. The basic mechanisms, how-
is a hydrolyzable, multifunctional silane ever, are not disputed. Alkoxysilanes undergo
76 capable of tying two or more polymer chains a process of hydrolysis/condensation to form
together. CR- and Q- cross-linker, a cured, cross-linked matrix. The alkoxy
which is methyltrimethoxysilane diluted groups are rst hydrolyzed with atmospheric
with methanol, are similar in structure to water, evolving an alcohol and thus forming
tetramethyloxysilane, with one of the silanol groups (g. ., ). This results in con-
methoxy functional groups replaced with a densation of the silanol groups, either by al-
methyl group. The methyl group increases cohol or water condensation. This reaction
the electron density of the silicon atom, caus- can be extrapolated to explain the curing of
ing the reaction to be completed more a hydroxy-terminated polydimethylsiloxane
quickly.3 using alkoxysilane as a cross-linker.4
After acetone was displaced with the poly- While there are diering theories about
mer solution, the basketry was cured using a the mechanism, the most probable process
topical application of a dibutyltin diacetate is the cleavage of the Sn-O-Si bond, which
catalyst followed by several days of vapor links the alkoxysilane with the polydimethyl-
catalyst deposition in a controlled environ- siloxane (g. ., 3). The reaction continues
ment. The mechanism of preservation is a with the remaining methoxy and hydroxy
combination of surface consolidation and groups until all have reacted or until the re-
thorough penetration and polymerization of sulting cross-linked polymer is too xed for
polymers throughout the basketry. the remaining methoxy groups to nd re-
maining hydroxy groups. In this reaction,
Chemical Reaction alkoxysilane is considered to be a cross-linker
because it cross-links the linear poly-
Many variations have been oered regarding dimethylsiloxane into a three-dimensional
the mechanisms involved in the process of matrix.

Fig. .. Formula depicting

the process of polymeriza-
tion.
MethodsRemoval of Concreted Material where it was allowed to drain of free-owing
and Initial Treatment of the Basket polymers. Soft cloths were used to wipe ad-
ditional polymer solution from the surfaces 77
Prior to treatment, extensive radiographs of of the concretion and from the irregular sur-
the artifact were taken. Traces of the outline faces and recesses of the basket.
of the basketry and shot were transferred
onto Mylar. These drawings were necessary Molding the Bottom of the Basket
to assist in the removal of the shot with mini-
mal disturbance to the underlying basketry. Once the thin layer of concretion covering
The layer of concretion covering the base the attest surface of the basket had been
of the basket was very thin (.mm) due in cleaned, dehydrated, and impregnated with
part to a barrel stave covering a section of the a polymer/cross-linker solution, a cast of the
base. A pneumatic chisel was used to remove surface of the basket was made, in order to
the stave and the underlying layer of calcare- create a form-tted, supportive base that
ous material. Dental tools were used to re- would protect the exposed section of basket
move some of the calcareous material be- while the upper surfaces of the artifact were
tween the strands of wicker. Care was taken cleaned and the individual shot removed.
to ensure that the basket remained wet Modeling clay was used to create a deep-sided
throughout this phase of mechanical clean- dam around the exposed edge of the basket.
ing. The entire concretion was then placed A mold of the bottom surface of the basket
into a large vat of industrial-grade acetone. was made using HSIII RTV silicone elastomer
After hours of room temperature dehydra- compound. After the cast was allowed to cure
tion, it was removed and rapidly transferred for hours, the clay dam was removed and
into a vat containing a silicone oil solution the entire artifact was turned over so that it
consisting of PR- silicone oil to which a was resting in the newly created base (g. .).
% addition of Q- cross-linker had The HSIII RTV rubber mold was rm enough
been added, by weight, and thoroughly to support the exposed surfaces of the bas-
mixed. The artifact was allowed to sit in the ket throughout treatment but soft enough to
polymer solution for seven days. The artifact dissipate potentially damaging vibrations
was then removed from the solution and from pneumatic chiseling and manual re-
placed onto several layers of newspaper, moval of the iron shot.

Fig. .. Making the HSIII

RTV rubber base.
 Removing the Shot solution, most of the remaining concreted
material was easily removed. After more con-
78 Before removing the shot from the interior cretion was manually removed, the basket
of the basket, topical applications of the PR- was immersed in a vat containing PR-/Q-
/Q- were applied as iron shot were solution for approximately hours. It was
taken out and new areas of the inner surface then removed from the polymer solution and
of the basket were exposed. This ensured that placed on several layers of newspaper. Soft
these areas of the basket were impregnated cloths were used to wipe the surfaces of the
with the polymer solution. Rate of absorp- basket. Because the basketry was fully im-
tion indicated the degree of penetration of pregnated with the polymer solution, there
the polymer solution into the fabric of the was no concern that the wicker would dry
basket. The presence of pooled polymers in out or shrink. Accordingly, the basket was
areas topically treated with the polymer so- allowed to drain of free-owing silicone oil
lution indicated when absorption of the poly- solution for several hours (g. .). While the
mer into the basket was complete. Dental solution was draining, the basket surfaces
tools and pneumatic chisels were then used were wiped several times to remove pooled
to remove impacted sediments and oxides polymers.
from the baskets interior (g. .).
Spot applications of the polymer solution Catalyzation
continued as the remaining shot were re-
moved. Once the internal and external sur- CT- was used to polymerize the silicone oil
faces of the basket were cleaned, the focus of impregnated wicker. CT- can be applied
work shifted to additional cleaning of im- either topically or as a vapor. Because of the
pacted silt and oxide-laden concreted mate- complexities of this project, both methods
rial within the woven recesses of the wicker were used to initiate polymerization. Fragile
Fig. .. Removal of shot strands. Dental tools were the safest and most areas located at the upper rim of the basket
from the interior of the eective tools for this work. Because the were spot treated with CT- using Q-tips
basket. wicker was impregnated with the polymer and soft cloths. This was necessary to stabi-
lize these areas while additional cleaning of
the wicker was carried out. After ve min-
utes of chemical reactivity, areas applied with
spot treatments of CT- were thoroughly
wiped with soft cloths.
Once the basket had been cleaned, it was
placed into a large polyethylene bag so that
the entire artifact could be vapor treated with
CT- catalyst. Several lint-free cloths contain-
ing drops of CT- were wadded and placed
into the bag along with the basket. When
sealed, the polyethylene bag formed an ideal
containment chamber, allowing catalyst fumes
to remain in close proximity to the surfaces
of the basket. After hours of exposure to
catalyst vapors, the basket was removed. Soft
cloths dampened with Q- cross-linker
were used to remove small areas of surface-
pooled polymers from the basket. After a brief
cleanup, the basket was returned to the poly-
ethylene bag along with three wadded cloths,
each containing several drops of CT-. 79
After a total of hours of catalyzation,
the basket was removed from its containment
chamber and placed into a fume hood. Dur-
ing the polymerization process, a few small
areas of white particulate were noted on the
surface of the basket in areas that had been Fig. .. Polymer solution
spot treated with CT-. This particulate was draining from the surfaces
of the basket.
easily removed by wiping the aected areas
with a lint-free cloth dampened with Q-
cross-linker. of the basket. No chemical reactivity was
noted with the wet surfaces of the basket
Iron Shot during the casting process and, in spite of the
viscosity of the material, highly detailed casts
All the iron shot were conserved using elec- of the basket were obtained. The casting
trolytic reduction. Each shot was placed into material reproduced undercuts in the baskets
a mild steel vat. A form-tted mild steel surface due to the irregular shapes of the
anode was constructed using -gauge ex- strands of wicker and the variation in the
panded mild steel mesh with -inch open- voids between the strands. Because HSIII
ings. An all-steel construction C-clamp was RTV remains supple, casting of undercuts
used to hold the shot during treatment. A and irregularities in the outer surface of the
section of AWG, separation , V maxi- basket rendered an extremely accurate repro-
mum rating insulated copper wire and duction of the baskets surface detail.
Mueller zinc-plated alligator (number ) Numerous silicone oil/cross-linker com-
were used to connect the artifact to a regu- binations were considered for treatment of
lated DC power supply. An aqueous solution the wicker. Because of concerns about the
of % sodium hydroxide solution was used inability of a viscous polymer compound to
as an electrolyte solution. penetrate through oxides and surrounding
Mercuric nitrate testing was used to de- concretion, Q- was selected as a cross-
termine quantitative amounts of chlorides in linking agent. Q- contains a high per-
the electrolyte solution. Once salts were mini- centage of methanol, which acts to make the
mized in each shot, they were placed into resultant polymer solution less viscous. The
baths of deionized water and boiled, hours use of methanol-based materials, however,
per day, for three days. Each shot was then requires that work be conducted in a well-
coated with a solution of Bakers tannic acid, ventilated fume hood and that appropriate
dissolved in ethanol. Three separate coatings apparel be worn to minimize exposure.
of tannic acid were applied to each shot, al- Because the concretion had been thor-
lowing a ferric tannate barrier to form on the oughly dehydrated in baths of acetone, dis-
metals surface. Each shot was then treated placement of acetone with the polymer
in microcrystalline wax. solution was eective. As individual shot
were removed from the interior of the bas-
Discussion ket, the use of a topical application of the
polymer solution ensured that each newly
HSIII RTV casting materials worked well for exposed section of basket was thoroughly
creating a cast of the surface detail of the side treated.
 The strategy of casting the partially ex- fresh air. Organo-tin compounds undergo
posed at surface of the basket worked well. hydrolysis with atmospheric water, produc-
80 As the surrounding concretion was me- ing organo-tin hydroxide. These oxides are
chanically removed, the rubber mold of the the agents of catalysis.
surface and its voids and undercuts acted to For purposes of treating the cast iron shot,
ensure that the physical size and shape of the electrolytic reduction processing ensured re-
basket were being preserved as treatment moval of soluble salts from the iron and pro-
continued. Because of the friable state of the vided excellent surface-cleaning capabilities.
wicker, removal from the concretion prior to Use of tannic acid followed by treatment with
treatment would have compromised dimen- microcrystalline wax has a proven track
sional data and other physical attributes of record for treatment of iron artifacts.
the basket. The major benet of using silicone oils in
The task of cleaning compacted sediments the preservation of composite artifacts is that
from between the bers of the basket was a the polymer/cross-linker does not have a
time-consuming task. Because the individual negative eect on the iron artifact component
bers were treated with a silicone oil/cross- of the concretion. This allows the conserva-
linker solution, the process of cleaning was tor to treat the organic component of the
facilitated since the polymer solution acted as artifact before commencing the task of re-
a lubricant between the surfaces of the wicker moving the bulk of the concretion. The pres-
strands and the surrounding sediments. ence of the polymer solution in the matrix
CT- is a versatile catalyst. Because there of the organic component prevents it from
were numerous delicate strands of wicker dehydrating. Polymer-impregnated silts and
around the upper edge of the basket, spot concretion can usually be easily cleaned from
treatment using CT- allowed conservators the surfaces of the organic component. This
to preserve small areas that were in jeopardy is advantageous for small, fragile sections of
of being lost or destroyed, while allowing the basketry.
more robust strands of wicker to remain Spot treatment catalyzation of the thin
unpolymerized until the entire basket was strands of wicker forming the edge of the
ready for catalyzation. Acids and bases should basket is advisable when working with larger,
be avoided for purposes of artifact conser- cumbersome artifacts. The ability to localize
vation; therefore, organo-tin catalysts are treatment ensures that frail strands are well
desirable since they tend to rapidly oxidize preserved while work continues on other ar-
into a neutral substance when exposed to eas of the artifact.
 CHAPTER 6
81

Cordage

and Textiles

T
hroughout history, cordage in one form monoxide (CO) greatly aect cordage and
or another has been made from the textiles. Storage and display of artifacts in
natural bers of plants and animals. close proximity to parking garages should
Cotton, ax, jute, hemp, grasses, silk, wool, therefore be discouraged. Airborne soot cre-
hair, and strips of animal hides have been ated from incomplete combustion of fossil
used to make rope. These materials have been fuels may chemically react with fabrics. Par-
woven, braided, or spun together to create ticulate contaminants lodge between bers,
threads for sewing and weaving. causing wear and physical damage.
Artifacts made of cotton and ax are Singly, or in combination, these factors
greatly susceptible to fungi and bacteria in and ever-present microorganisms act to un-
humid conditions and are seldom recovered dermine an artifacts tensile strength and pli-
from archaeological excavations. Other arti- ability. Generally, textile conservation should
facts made using animal bers, which are be left in the hands of specially trained con-
primarily proteins, are better able to survive servators. Accordingly, most of this chapter
because they are less prone to destruction by is devoted to the preservation of waterlogged
bacteria. All cordage and textiles, however, are rope. One case study, however, discusses the
aected by exposure to light. Natural dyes will preservation of tar-coated canvas associated
fade quickly if an artifact is left in direct, with a gudgeon strap from a vessel that sank
bright sunlight. The ultraviolet spectrum of during the earthquake of in Kingston
normal light will accelerate deterioration of Harbor, Jamaica.
the artifact, often reducing its pliability. Air-
borne pollutants interact with dyes and natu-
ral bers. Care should be taken to ensure that New Techniques for the Preservation
artifacts are not stored in modern cabinetry of Waterlogged Rope
constructed using particleboard and urea-
based nishes. O-gases from these materials To help illustrate new techniques of preserv-
interact with the natural bers of the artifact ing waterlogged rope, I discuss rope recov-
and some chemicals commonly used in their ered from La Salles vessel, La Belle, which
preservation. sank o the Texas coast in . Among the
Other forms of air pollution can aect organic materials found on the vessel were
museum displays and long-term curation of lines from the rigging and coils of rope that
artifacts. Fumes from automobile emissions had been stored in her hold. Conservators
are dangerous to organic artifacts. Hydrocar- have been faced with the formidable task of
bons, nitrogen oxides (NOX), and carbon preserving the large amount of hemp rope
 recovered from the vessel, which includes lm; the lm is then heat-sealed to produce
segments more than . m long. Especially a formtting jacket in which each rope
82 challenging to the conservators is the anchor sample remains throughout treatment. The
line of the vessel, consisting of a continuous packaged rope is attached to a piece of
section of rope more than m long. The Masonite, which acts to support the rope, and
diameter of these sections of rope ranges then treated with PEG. After treatment, the
from . to . cm, after conservation. Al- rope is placed into a large freeze-drying unit
though most of the recovered rope appeared and freeze-dried at C with a % rela-
to be in pristine condition during excavation, tive humidity.
microscopic analysis indicated that most of After freeze-drying, the rope sample is re-
the bers were thin and visibly degraded; moved from the Masonite/polyethylene bag
microbial action had caused the bers to de- structure and allowed to sit in fresh air. Rope
cay while signicant water saturation had specimens treated with the Frankfurter
further weakened them. method often require additional treatment
with applications of polyurethane in ethyl-
acetate.2 This is necessary because the rope is
Silicone Treatment Strategies often extremely delicate after processing. Rope
treated with the Frankfurter method retains
The Texas A&M Archaeological Preservation its pretreatment color and the individual
Research Laboratory technique for stabiliz- bers, yarns, and strands that comprise the
ing waterlogged rope with silicone oils in- rope are well preserved. Like other success-
volves a displacement of the water trapped ful treatments for severely deteriorated wa-
in the rope bers with acetone, followed by terlogged rope, this process is generally not
the replacement of the acetone with a hy- reversible due to two factors. First, the appli-
droxyl-ended functional polymer and cross- cation of polyurethane in ethylacetate is gen-
linker. The polymer-impregnated rope is erally not reversible. More important, how-
then cured by exposure to a catalyst, which ever, is the fact that most treated rope samples
is applied either topically or as a vapor. I sug- are very friable and desiccated after treat-
gest that the preservation of the treated rope ment, making additional treatment dicult.
is a result of both surface consolidation and
penetration of hydroxyl-ended polymers
with cross-linking agents added. The poly- Treating Waterlogged Rope in a
mer-rich matrix of the artifact is then treated Nonpolar Suspension Medium
with a tin-based catalyst, which completes the
polymerization process. Experiments conducted by the National Mu-
seum Conservation Laboratories have also
demonstrated that treating waterlogged rope
Frankfurter Method of Rope with PEG in a volatile, nonpolar solution such
Preservation as ether or kerosene enables the individual
bers to oat during treatment, facilitating
Conservators at the National Museum Con- thorough impregnation of the PEG within the
servation Laboratories in Brede, Denmark, ropes matrix. The use of suspension medi-
routinely use a technique for conserving wa- ums in PEG treatments results in rope speci-
terlogged rope that they refer to as the mens that lack the matted appearance of rope
Frankfurter method.1 This process in- treated with PEG alone; the resulting rope,
volves encapsulating waterlogged rope be- however, is extremely fragile and very suscep-
tween sheets of perforated polypropylene tible to environmental changes.
Incorporating the Use of Nonpolar proposed hybrid silicone treatment process
Suspension Mediums and Elements of (Si- and Si-), while the third sample (WL)
the Frankfurter Method into Tradi- would be allowed to air-dry at room tempera- 83
tional Silicone Treatment Strategies ture for a -hour period.
The samples were rinsed in fresh running
Experiments conducted with silicone oil water for two days to ensure the removal of
treatments at the Archaeological Preservation soluble salts. The samples were then placed
Research Laboratory have demonstrated that on a sheet of glass for additional manual
treating waterlogged rope that has not been cleaning. During this process, the samples
enclosed in some form of permeable mate- were positioned beneath a constant, gentle
rial results in a specimen that tends to un- ow of tap water to keep the rope wet while
ravel slightly during treatment. We believed debris was ushed from its surfaces. Like the
the polypropylene jacket used in the Frank- majority of rope from the assemblage, the
furter method would provide a permeable samples were partially covered with black
membrane that would facilitate chemical and dark brown sulde stains. These stains
transfer and also serve to protect the physi- resulted from the fact that the cotton cloth
cal integrity of the artifact during treatment. in which the rope had been transported
Furthermore, after observing the results of from the site to the laboratory had decayed
experiments conducted by the National Mu- en route. Most of these stains were removed
seum Conservation Laboratories on the use by lightly rubbing the aected areas with a
of suspension mediums in the treatment of cotton swab. No attempts were made to re-
waterlogged rope with PEG, we anticipated move deeply set stains by chemical means,
that the use of a suspension medium during as it was feared that additional chemical ad-
the polymerization of waterlogged rope ditives would interfere with the conserva-
would alleviate the slightly matted appear- tion process.
ance commonly observed after silicone oil Each sample that was to undergo silicone
treatments that do not involve the use of a treatment was placed between two sheets of
nonpolar suspension medium. perforated polyethylene lm, which is scored
with uniform holes that allow water, acetone,
and silicone oil to freely diuse (g. .).
Case Study: La Belle Rope These sheets of polyethylene lm were then
heat-sealed, creating formtting, ventilated
The following procedure is but one example bags in which the ropes would remain
of the use of preservation polymers in rope throughout the initial stages of treatment.
conservation. For researchers in the eld of Ziploc vegetable bags are an ideal source of
conservation, exploration with other silicone perforated polyethylene lm; they are readily
preservation polymers and cross-linkers is available and easily sealed to form a pouch,
recommended in order to determine the re- using either a heat-sealing appliance or a
sultant attributes of varying combinations of small soldering iron and brown paper.
these invaluable materials. Each encased rope was placed into a bea-
The majority of rope recovered from La ker containing ml of fresh acetone, and
Belle was transported to the Texas A&M Uni- the beaker was set in a vacuum chamber. At
versity Preservation Research Laboratory, room temperature, a vacuum of . Pa
where it was stored in freshwater until treat- ( Torr) was applied to the samples to in-
ment. Three rope samples of similar length duce rapid displacement of the water with
were taken from a single continuous coil. Two acetone. The samples initially bubbled rap-
of these samples were to be treated by the idly as air and acetone were driven from the
 than PR-. Because of the porosity of the
waterlogged rope, larger molecular weight
84 polymers such as PR- are expected to eas-
ily permeate the matrix of the rope samples.
Due to its increased viscosity, PR- acts as
a consolidant by keeping loose strands
together; furthermore, rope that has been
treated with PR- tends to be more exible
after treatment than rope treated with smaller
molecular weight polymers. A blend of these
two silicone oils was used for this experiment
to ensure that the nished product main-
tained a degree of exibility as well as inter-
nal rigidity and physical strength. CR- %
by weight was added to the PR-/PR- sili-
cone oil solution. CR- is a highly ecient
Fig. .. Waterlogged rope internal structures of the rope. After approxi- cross-linker that has been shown to work well
encased between sheets of mately minutes, the rapid bubbling ceased, with silane polymers.
Ziploc vegetable bag and smaller, more infrequent bubbles were After placing the dehydrated ropes in clean
material prior to water/ observed escaping from the ropes. The beakers, a sucient amount of the silicone
acetone displacement. samples were then removed from the water- oil/cross-linker solution was added to each
laden acetone and placed into clean beakers beaker in order to immerse the samples in
containing ml of fresh acetone. Each bea- solution. Aluminum mesh was securely xed
ker was returned to the vacuum chamber and over the packaged ropes to prevent them
a vacuum of approximately . Pa ( from oating to the surface of this viscous
Torr) was applied. Once the bubbling ceased, mixture. A vacuum of . Pa ( Torr)
the ropes were removed from the vacuum was applied to the samples in solution for
chamber and allowed to sit at ambient pres- minutes to ensure that the acetone present
sure and room temperature. in the rope bers would vaporize rapidly, fa-
The next phase of treatment was to ex- cilitating a thorough penetration of silicone
change the acetone with an appropriate sili- oil solution throughout the artifacts. During
cone/cross-linker solution. The polymer and the initial stages of vacuum treatment, large
cross-linker were specically chosen to pro- bubbles were observed escaping from the
duce a desired texture and strength. To main- ropes. After minutes, this rapid bubbling
tain exibility in the treated rope samples, diminished and sporadic, small bubbles rose
two hydroxyl-ended silicone oils were from the artifacts.
blended together in a : solution, by The packaged ropes were then taken out
weight. The lighter of the two polymers was of the vacuum chamber and allowed to sit
PR-, which is a low viscosity hydroxyl- in solution at ambient pressure and room
ended uid.3 Repeated experimentation in- temperature. After hours the samples were
dicates that lighter molecular weight silicone removed from the silicone oil/cross-linker
oils such as PR- tend to penetrate easily solution and from their perforated polyeth-
into organic materials such as rope; once po- ylene bags. The samples were placed on an
lymerized, however, they tend to produce a aluminum screen to allow drainage of ex-
rigid artifact. cess free-owing silicone oil solution. After
PR- is a slightly more viscous hydroxyl- hour the surfaces of the ropes appeared to
ended uid with a larger molecular weight be reasonably dry, and the artifacts were
placed in beakers containing ml of fresh of deterioration caused by waterlogging, as
CR- cross-linker. Immersion in CR- af- well as to compare the results of the silicone-
ter bulking the samples with a silicone oil/ treated rope against an untreated specimen, 85
cross-linker solution is helpful in removing an additional sample of rope from the same
additional silicone oil solution from the ex- coil was weighed, measured, and allowed to
ternal surfaces of the rope. During the im- air-dry at room temperature in a vented fume
mersion process, the rope surfaces were hood for hours.
wiped with a soft brush to facilitate removal
of excess silicone oil solution. After ve min- Results and Discussion
utes of immersion and surface preparation,
polymerization was initiated by exposing the The silicone oiltreated rope samples felt
rope sample to a tin-based catalyst. slightly sti, but were stable and aesthetically
The samples were placed into loose, per- pleasing compared to the sample that was
forated polyethylene bags, and the bags were simply allowed to air-dry. The individual -
heat-sealed shut. Each package was sus- bers, strands, and ply of the silicone-treated
pended with two wooden clothespins from rope samples were easily distinguishable
wooden dowels. These dowels rested on the after treatment and did not become matted
top edges of a small vat containing kerosene or compressed. The high degree of visible
with CT-, % by weight (g. .). The open detail in the silicone-treated samples was sur-
structure of the mesh bag evenly exposed the prising because these features were indistin-
surfaces of the rope to the kerosene/catalyst guishable in a waterlogged state. And while
solution. With the samples suspended in the it may be impossible to determine the origi-
solution, the vat was placed into a vacuum nal color of the waterlogged rope, the post- Fig. .. Rope sample
suspended in kerosene/
chamber, and a vacuum of . Pa ( treatment coloration of the silicone-treated
CT- solution:
Torr) was applied. After minutes under samples was acceptable, ranging from a pale
(A) clothespins; (B) dowel
vacuum, the valves of the chamber were to medium brown. (See tables ., ., ., and
spanning the length of the
locked, and the rope was left suspended in ..) Figure . is a photograph of sample Si- vat; (C) perforated
the solution overnight. The following morn- . Prior to treatment, this sample was loosely polyethylene bag;
ing, the vacuum chamber was returned to twisted. Following treatment, no discernible (D) silicone-impregnated
ambient pressure, and the samples were re- changes were observed in the twist or physi- rope suspended in mesh;
moved from the kerosene/catalyst solution. cal dimensions of either silicone oiltreated (E) vat containing
The rope was removed from the perforated rope sample. kerosene/CT- solution.
polyethylene bags and placed on several pa-
per towels, which absorbed the kerosene/
cross-linker and silicone oil solutions from
the artifacts. Immediately following removal
from the kerosene/cross-linker solution, the
surfaces of the cordage were covered with a
thin, slippery coating of silicone. After the
rope was exposed to fresh air for a few min-
utes, droplets of fully cured polymer were
observed on one end of the samples. These
were easily removed using a soft, lint-free
cloth. After the rope was allowed to air-dry
in a vented fume hood for hours, the sur-
faces of the samples appeared dry and very
natural in texture. To determine the degree
 Posttreatment Strength ing. To prevent slippage with the other
samples, the second silicone-treated rope
86 After one week of air-drying, the silicone oil sample (Si-) and the untreated sample (WL)
treated rope and the comparably sized sample were mounted into the load frame using
of untreated rope (WL) were taken to the wedge grips only after being outtted with
Texas Engineering Experiment Station, Test- epoxy potted ends. This is a more complex
ing Machinery and Repair Laboratory at mounting process that requires that the ends
Texas A&M University for tensile strength of the rope be cemented into a cone-shaped
testing. We believed these tests would pro- epoxy base prior to mounting in the frame.
vide insight into the strength characteristics The use of these potted ends eliminated slip-
of polymer-treated rope. Tensile strength page and resulted in more reliable data.
testing was conducted with a kip ( kip = Tensile strength testing has demonstrated
lbs tensile strength) MTS servohydraulic that rope preserved in silicone oil is consider-
frame, which measures the maximum load- ably stronger than rope that has been allowed
breaking point of materials. To more accu- to air-dry. When tested, the rope section WL
rately measure the maximum load-breaking failed at . lbs tension, and the sections of
point of low-potential tensile strength mate- rope treated with silicone oils, Si- and Si-,
rials such as the fragile, treated La Belle rope failed at . lbs and . lbs, respectively.
bers, a kip load cell was mounted into the Table . lists the data acquired from this ten-
jaws of the kip machine. sion test. As a result of waterlogging, which
Data control and acquisition were re- deteriorated and weakened individual bers
corded using Gardner Systems software. of the rope, the strands that make up the rope
Time, distance, and pounds force were mea- failed at dierent times in each of the samples.
sured for each sample. In each test, tensile The sawtooth graphs in gures ., ., and
strength testing continued until the sample . represent the tension and breaking points
failed. Rope (Si-), treated with silicone oils, of the bers, strands, and plies of the rope
was mounted in the load frame using wedge samples. Although posttreatment strength
grips. This sample slipped once during test- may not be an important factor in the deci-

Table . Measurements and Percentage Changes for Air-Dried Rope Sample WL

Time Weight % Change Width % Change Length % Change

(in g) (in cm) (in cm)
Before air-drying 12.2 1.516 9.788
1 hour air-drying 3.10 74.590% 1.480 2.374% 8.761 10.492%
3 hours air-drying 2.82 76.885% 1.472 2.902% 8.751 10.594%
5 hours air-drying 2.70 77.868% 1.472 2.902% 8.642 11.708%
24 hours air-drying 2.70 77.868% 1.472 2.902% 8.531 12.842%

Table . Texture, Integral Strength, and Color of Air-Dried Rope Sample WL

Pretreatment Posttreatment
Texture Soft-mushy Brittle
Integral strength Fragile Friable
Color 10YR-2/2 very dark 5YR-6/2 5YR-5\2
brown 10YR-2.1 black pinkish gray
Table . Measurements and Percentage Changes for Silicone-
Treated Rope Samples
87
Measurement Mean % Change
Length of rope (in cm)
pretreatment 14.308 14.228 1.118
posttreatment 14.148
Width Si- (in cm)
pretreatment 0.9280 .922 1.293
posttreatment 0.9160
Width Si- (in cm)
pretreatment 0.9380 .932 1.386
posttreatment 0.9250
Weight of rope (in g)
pretreatment 16.597 11.499 61.438
posttreatment 6.400

Table . Color and Flexibility of Silicone-Treated Rope Samples

Pretreatment Posttreatment
Color 10YR-2/2 very dark brown 10YR-6/3 pale brown 10YR
10YR-2.1 black 4/3brown
Flexibility Limp, almost formless, individual Dry, individual strands visible, slightly stiff
strands indistinguishable

Note: Color determined using Munsell Soil Color Charts, Macbeth Division, Kollmorgen
Corporation, Baltimore, Md., 1975 ed., p. 26.

sion to conserve rope by a particular method, chemical transfer and also helped protect the
it is benecial to know that silicone oil physical integrity of the artifacts during treat-
treated cordage is more internally stable and ment. Rope treated without being enclosed
stronger than rope not treated with silicone in some form of permeable material, such as
oil. a perforated polyethylene bag, tends to un-
ravel slightly during treatment.
Effectiveness of Incorporating Nonpolar Silicone oiltreated rope treated without
Suspension Mediums and Elements of the a nonpolar suspension medium such as kero- Fig. .. Silicone oil
Frankfurter Method into Traditional sene usually results in an artifact with a treated rope Si-.
Silicone Treatment Strategies

The Danish process of using perforated

polypropylene lm to make a formtting
jacket within which archaeological rope is
treatedan approach most commonly uti-
lized in the Frankfurter methodworked
well when incorporated into our silicone
treatment. The polypropylene jacket pro-
vided a permeable membrane that facilitated
 Table . Tensile Strength Data for Rope Samples

88 File Sample Peak Load (in lbs) Comments

Rope 1 Si-1 36.5 Wedge grip mounted
Rope 2 Si-2 27.7 Epoxy potted ends
Rope 3 WL (waterlogged) 2.6 Epoxy potted ends

slightly matted appearance. Immersion in a nature of kerosene. An additional disadvan-

nonpolar solution enables the bers to oat tage in using a kerosene/catalyst medium for
and facilitates the polymerization of indi- polymerization is that it takes several days for
vidual bers. The use of kerosene with a tin- the faint odor of kerosene to be eliminated
Fig. .. Rope Si- load/ based catalyst works well as a medium for from the artifact. Other volatile solvents may
displacement data: polymerization, but the kerosene/catalyst work well in place of kerosene, but these ex-
Te increasing tension; polymerization medium is not ideal for rou- periments have not yet been conducted.
B break point of strand; tine laboratory use due to the ammable Experimentation has indicated that CR-
F failure of strand. cross-linker/CT- catalyst, % by weight,
solution is an eective and safe substitute for
the kerosene/catalyst mixture. After treated
samples were removed from the CR-/CT-
catalyst solution, the surfaces of the rope
were not slippery, suggesting that more com-
plete catalysis occurred while the sample was
in solution. Residual odors associated with
this catalyzation medium dissipated in a
matter of minutes once the sample was ex-
posed to fresh air.
After organic materials have been impreg-
nated with silicone oil/cross-linker solutions,
immersing them in CR- cross-linker and
Fig. .. Rope Si- load/ surface-wiping with a cotton swab or a lint-
displacement data: free cloth is an eective way to remove excess
Te increasing tension; polymer from the surfaces of an artifact. In
B break point of strand; some cases, excess silicone oils have been re-
F failure of strand.
moved from the surface of rope by immer-
sion in CR- under a slight vacuum. The
process appears to eliminate a great deal of
silicone oil solution from within the deep
crevices and voids on the ropes surface.
Obvious benets of using silicone oils for
conserving waterlogged rope include the
short treatment duration and the minimal
amount of laboratory equipment required
for the process; PEG/freeze-drying methods
of rope preservation require substantially
more time and labor. The silicone oiltreated
examples were conserved in less than
 89

Fig. .. Rope WL load/

displacement data:
Te increasing tension;
B break point of strand;
F failure of strand.

hours. In addition, artifacts treated with sili- Silicone oil treatments, like treatments
cone oils do not require special curation and that require applications of polyurethane in
the dimensional attributes of the artifact ethylacetate, are not reversible; this is not to
seem to be accurately preserved. As Vera de say, however, that rope treated with silicone Fig. .. Accelerated
la Cruz Baltizar has observed, silicone oil oil cannot be re-treated. In the past, several weathering machinery
treated samples appear to be dimensionally fragile leather and canvas artifacts preserved used for testing silicone
stable with good coloration.4 Repeated test- with lower centistoke silicone oils have been oiltreated artifacts.
ing of the hybrid silicone oil treatment
described above at the Archaeological Pres-
ervation Research Laboratory consistently
yields waterlogged rope specimens that are
aesthetically pleasing and dimensionally
stable.
Accelerated aging tests and data supplied
by Dow Corning Corporation continue to be
encouraging regarding the long-term stabil-
ity of silicone oiltreated artifacts. Eight sili-
cone oiltreated samples were subjected to
an extended test in an accelerated weather-
ing machine (g. .). The samples were
exposed to four months of continuous, al-
ternating cycles of hours at high humidity
(%) and high temperature (C) with a
UV light, and hours at a lower humid-
ity (%) and temperature (C) with no
light exposure. The tested sample data (in-
cluding overall dimensions, color, and sur-
face integrity) were comparable to data for
silicone oiltreated specimens that had not
undergone accelerated weathering.
 re-treated, using more viscous polymers to metal plate could be treated using appropri-
add additional strength and stability. More ate conservation strategies with the idea that,
90 important, waterlogged rope appears to re- once completed, the gudgeon assembly could
spond well to treatment using silicone oils. be documented and reconstructed on paper.
Although silicone oil processes will not ad- Microscopic analysis of the canvas indi-
dress all archaeological conservation issues, cated that the backing had been made from
the eld of archaeological conservation can cotton canvas material that had been crudely
benet from ongoing research into silicone cut to follow the shape of the gudgeon plate.
oil treatment techniques. Two squared holes were present in the sec-
Based on experimentation to date, I con- tion that was to be treated using silicone oils.
cur with Baltazars observations that silicone These holes lined up with the remains of two
oil preservation is a promising technique for large bolts that would have been used to fas-
the consolidation and preservation of many ten the gudgeon plate to the hull of the ves-
waterlogged materials.5 The added strength sel. The warp of the fabric was counted at
and elasticity characteristics associated with strands per inch while the weft averaged
the silicone oil process used for this experi- strands per inch. Variances in warp and weft
ment may have important implications for counts, as well as the generally uneven shape
the structural well-being of some artifacts. of the strands, suggest that this was probably
Continuing research at the Archaeological a common, though sturdy, material.
Preservation Research Laboratory is focused Once removed from the back of the gud-
upon these issues. geon, the canvas was placed into a vat of fresh
tap water, and the surfaces of the cloth were
lightly cleaned, using ngertips to remove
Case Study: Preservation of Waterlogged loose debris and concretion. The canvas was
Canvas from Port Royal initially placed into a vat of acetone, which
successfully removed traces of pitch embed-
Among the wide range of artifacts recovered ded in the weave of the fabric. After several
from underwater excavations at Port Royal, rinses of freshwater, which were used to re-
Jamaica, was an interesting section of a gud- move salts and any remaining acetone from
geon from a vessel. Because the Port Royal the fabric, the canvas was placed into a %
site is covered by a deep silt layer, many or- solution of hydrochloric acid in water to aid
ganic artifacts have been recovered that in the removal of rust stains and minute
would not have survived, or at least be as well specks of concretion. After treatment, the
preserved, at a land site. Artifact PR - cloth was immersed in several freshwater
is such an object. When recovered, the gud- baths to remove any remaining acid. The can-
geon plate was heavily concreted, looking vas was then briey placed into a % solu-
more like a long rock than a piece of ships tion of hydrogen peroxide to remove heavy
equipment. sulde stains. The fabric was again rinsed in
During initial phases of conservation, we freshwater.
were delighted to note that a long strip of The goal of the entire cleaning process was
canvas had been used as a backing behind to remove sucient pitch and staining to
the gudgeon plate, along with a thick layer of enable the polymer to penetrate the cloth but
pitch. After the outer layer of concreted ma- not enough to disturb their function as di-
terial had been removed from the artifact, it agnostic attributes of the canvas. The clean-
was a relatively simple task to slowly sepa- ing process complete, the fabric was placed
rate the canvas backing from the back sur- into an acetone bath to dehydrate the mate-
face of the plate. Once removed, the cloth and rial in preparation for silicone bulking. To aid
in the removal of any remaining water, the containing grams of catalyst was placed
container holding the fabric in acetone was into the containment chamber, and the pro-
placed into a vacuum chamber and a vacuum cess was continued for an additional 91
of Torr was applied until all bubbling hours.
ceased. At this point, all remaining water had After two days of polymerization, the fab-
been removed from the artifact. The canvas ric felt slightly damp, but had retained the
was then quickly placed between two sheets look and feel of cotton canvas material. The
of lint-free paper and blotted to remove some warming oven was turned o, and the fabric
of the acetone in the fabric. was left in the containment chamber at room
After blotting, the canvas was placed into temperature for an additional hours. Af-
a large beaker containing mg of PR- ter a total of three days, the fabric was re-
silicone oil. A mesh screen was placed on top moved and allowed to air-dry undisturbed
of the fabric in solution as a means of keep- on a sheet of white paper.
ing the cloth suspended in solution through-
out the bulking process. The beaker was then Observations
placed back into the vacuum chamber, and
as before, a vacuum was applied to the fabric The processes used for removing embedded
for hours. Vigorous bubbling ceased after concretion and stains from the cotton can-
minutes and no bubbles were noted after vas are conventional conservation processes
hour of applied vacuum. The fabric was al- that are widely used in the treatment of stable
lowed to sit in solution and slowly returned fabrics. Care was taken to thoroughly rinse
to ambient room temperature, where it re- the artifact prior to bulking with silicone oils;
mained for hours. the eects of interaction between these ma-
Fig. .. Containment
Before the fabric was removed from the terials are not known. The porous nature and chamber conguration
silicone solution, a warming oven was pre- loose weave of the fabric suggested there illustrating the canvas
heated to C. As in many of our other ex- would be no diculty in the dehydration and material sitting on an
periments, a containment chamber was cre- subsequent bulking of the material with sili- aluminum screen above
ated by placing an inverted polypropylene cone oils; this meant that using the longer a tray containing CT-
pail with a tight-tting lid into the warming process of applying a vacuum to the fabric catalyst.
oven. In the center of the lid (acting as the
base of the chamber), we placed a at dish
containing grams of CT- catalyst. A large
mesh screen covering the dish served as a
platform on which the fabric could be placed
for exposure to warm catalyst fumes (g. .).
After much of the free-owing silicone was
removed by suspending the canvas over a
stainless steel dish and allowing excess sili-
cone to drip from the material, the fabric was
placed between two sheets of newspaper and
lightly pressed to remove additional oils. The
fabric was then placed onto the screen in the
containment chamber, where it was exposed
to catalyst fumes for hours. After hours,
the catalyst had become hard and crusty as a
result of silicone oils dripping into the dish,
forming polymerized material. A new tray
 acetone baths prior to silicone bulking would
have removed more of the staining and pitch
92 from the fabric. These, however, are diagnos-
tic attributes of the cloth, indicating some-
thing about the mounting of the gudgeon
strap to the vessel. Accordingly, dehydration
time was closely monitored to ensure that
these stains were not removed.
After treatment, the fabric was both ex-
ible and stable. Indeed, ve years after treat-
ment, exibility and physical integrity of the
canvas remain unchanged (g. .). Placing
the treated canvas on sheets of white paper
Fig. .. Posttreatment in a freezer-mounted vacuum chamber to was the simplest means of determining the
view of the canvas after drive o acetone was not necessary. degree of polymerization that had occurred.
catalyzation. Pre-and posttreatment traces of the out- After it sat on paper for one week, no oil spots
line of the canvas material, as well as mea- were noted. If polymerization were incom-
surements taken from the edge of the mate- plete, the points of contact between the can-
rial to the edges of the nail holes, indicated vas and paper would be oil stained. Since no
that no determinable shrinkage occurred stains were noted, this was a good visual in-
in the fabric. More extensive treatment in dicator that polymerization was complete.
 CHAPTER 7
93

Glass

Conservation

I
n ancient times, artisans combined secret variety poses a considerable challenge for
ingredients, which they heated to high archaeological conservators.
temperatures to form a wide range of col- The basic component of glass is silica
ors of glass. They used ash and varieties of (SiO2), which, unlike quartz, is a randomly
plants to make ux, which helped in meld- arranged network of crystalline structures.
ing their complex chemical compositions. The presence of additional elements such as
The chemical processes in a marine environ- oxides (R2O) or potash (K2O) act as uxes,
ment that act to deteriorate glass are as com- which by their nature aect some silicon-
plex and varied as the processes of making oxygen bonds. Unattached oxygen atoms are
it. In this environment, organic components negatively charged and, as Cronyn has ob-
of ancient glass are easily leached out of its served, loosely hold monovalent cations
matrix through a series of complex reactions, within the spaces of the network.2 Because
rendering the glass unstable. these bonds are weak, the cations migrate
Despite the fact that glass has been made within and out of the network in close asso-
for more than years, scientists know little ciation with water. The proportion of silica,
about the development of the glassmaking ux, and other stabilizing agents within glass
process and the myriad combinations of ma- is innitely variable. These proportions aect
terials used by artisans to create glass the character and quality of glass. Addition-
throughout history. Archaeologists have re- ally, the combination of these elements will
covered an incredible variety of glass from determine the melting point of the glass. The
terrestrial and underwater excavations. From coloration of glass depends on the state of
analysis of these artifacts, we know that a oxidation-reduction of the artifact as well as
wide variety of raw materials can be com- mixtures of ions and the presence of addi-
bined to produce raw glass. tives, which can accentuate or diminish color.
Although there is signicant debate about The basic composition of nearly all an-
the fundamental nature of glass, Robert Brill cient glass consists of a soda-lime-silica mix-
of the Corning Glass Museum considers it to ture with substantial impurities added to the
be a state of matter that combines the rigid- mixture from raw materials. In medieval
ity of a crystal with the largely random mo- Europe, the source of silica was sand and, in
lecular structure of a liquid.1 He has observed some cases, crushed stones. Limestone and
that hundreds of thousands of forms of glass burned shells were a common source for the
have been made; the term glass therefore is lime component of glass. Mesopotamian tab-
a generic term used to loosely describe an lets also specify that ground red shells from
endless series of chemical variations. This the sea were a good source for lime.3 As the
 Roman Empire expanded, so did the spread minute amounts of manganese are added to
of technologies related to glassmaking. This make the glass clear, but potash glass is less
94 inuenced English glassmaking and the tech- clear than soda glass. Most early glass is green
nologies used during the colonial period of because of iron impurities in the materials.
Jamaica. The alkali lowers the melting point of the
Numerous factors complicate attempts to sand and the ux facilitates the mixture of
understand the processes of devitrication the components. As long as the original glass
that aect archaeological glass. First, the in- mixture is kept in balance, the resulting glass
teraction of water with glass, especially sea- will be stable. Problems arise when an ex-
water, causes chemical instability. Association cess of alkali and a deciency in lime (cal-
with other materials within an archaeologi- cium oxide is used as a stabilizer) are used
cal site can complicate the chemistry of glass, in the mixture, for the glass will be especially
since oxides and minerals may interact posi- susceptible to attack by moisture. A relative
tively or negatively with the degradation pro- humidity (RH) higher than % can be dan-
cesses of the glass. In addition, the sediments gerous. Old glass with % sodium or
in which the glass is buried profoundly aect potassium may begin to weep and break
the degradation process. These and other fac- down. This is often called glass disease.
tors account for the wide range of colors and In all glass, the sodium and potassium
degrees of degradation observed within the oxides are hygroscopic; therefore, the surface
Port Royal glass assemblage, consisting of of the glass absorbs moisture from the air.
hundreds of bottles, known as onion bottles; The absorbed moisture and exposure to car-
sections of window pane, often with lead bon dioxide causes the NaO2 or NaOH and
caming attached; and fragments of small the KO2 or KOH to convert to sodium or
ointment bottles. potassium carbonate. Both NaCO2 and KCO2
Glass is usually the most stable of archaeo- are extremely hygroscopic. At an RH of %
logical materials, but it can undergo some and above (and in some cases as low as %),
complex disintegrationespecially seven- drops of moisture appear on the glass sur-
teenth-century glass. Ideally, glass should face. In water, especially saltwater, the Na and
consist of % silica, % alkali or K carbonates in unstable glass can leach out,
soda ash (sodium carbonate) or potash (po- leaving only a fragile, porous hydrated silica
tassium carbonate, usually derived from (SiO2) network. This causes the glass to craze,
wood ash), and % ux lime (calcium crack, ake, and pit, giving it a frosty appear-
oxide). Soda-lime glass is the most common ance. In some cases, an actual separation of
glass throughout the history of glassmaking, layers of glass from the body occurs. Fortu-
and the modern equivalent is approximately nately, glass from the nineteenth century or
% silica (SiO2), % soda (Na2CO3), later seldom manifests such eects. Problems
and % lime added as a stabilizer. Soda are rarely encountered on glass found at sites
glass is characteristic of southern Europe, dating from this period. Pearson discusses
where it is made from crushed white pebbles glass deterioration and reviews the various
and soda ash derived from burnt marine veg- glass conservation procedures.4
etation. Soda glass, often used for the manu- Currently, the decomposition of glass is
facture of cheap glass, is twice as soluble in imperfectly understood, but most glass tech-
water as potash glass. nologists agree that glass decomposition is
Potash glass is more characteristic of in- due to preferential leaching and diusion of
land Europe, where it is made from local alkali ions (Na and K) across a hydrated po-
sands and potash derived from wood ash and rous silica network. Sodium ions are removed
burnt inland vegetation. A little salt and and replaced by hydrogen ions, which diuse
into the glass to preserve the electrical bal- simple rinsing and removal of incidental
ance. The silicates are converted into a hy- stains (especially lead sulde staining on any
drated silica network through which sodium lead crystal) or calcareous deposits is envi- 95
ions can diuse out. Glass retrieved from an sioned. The main problems will be related
acid environment often appears laminated, to gluing pieces together and possibly some
with an iridescent lm formed by the leached restorations. All the problems likely to be en-
silica layers. The alkali that leaches out is neu- countered are discussed thoroughly in Con-
tralized by the acid and fewer hydroxyl ions servation of Glass by Roy Newton and Sandra
are available to react with the silica. This Davison.5
causes the silica layer to thicken and become Submersion in seawater will cause glass to
gelatinized as the alkali leaches out. Glass become unstable. If left untreated, the arti-
excavated from an alkaline environment is fact will deteriorate. A technique described
less likely to have laminated layers because by Plenderleith and Werner is representative
an abundance of hydroxyl ions are available of a range of conservation strategies that are
to react with the silica network. Normally, a eective for stabilizing archaeological glass.6
protective layer does not form on glass ex-
posed to alkaline solutions. The dissolution . Wash the artifact thoroughly in running
of the glass proceeds at a constant rate. The tap water.
alkali ions are always extracted in excess of . Soak it in distilled water.
the silica, leaving an alkali-decient layer that . Immerse the artifact in two baths of
continually thickens as the deterioration alcohol to dry quickly. This treatment will
moves deeper into the glass. retard the disintegration and also improve the
There are considerable dierences of opin- appearance of the glass. It does not, however,
ion regarding the preservation of unstable always stop the breakdown of the glass.
glass. Some conservators believe the only . If applicable, apply an organic lacquer
treatment should be to keep the glass at low PVA, Acryloid B-to impede the disinte-
RH so that it cannot react with any moisture. gration.
While a RH environment of % is ideal for . Store the artifact in a dry environment
archaeological class, the range can be ex- with the relative humidity no higher than
tended to % if the glass is stable. The weep- %; others say % is ideal. The Corning
ing or sweaty condition is sometimes made Glass Museum keeps incipient crizzled glass
worse by the application of a surface lacquer stored at % RH. RH % is the critical
or sealant, which traps humidity in close as- point at which KCO3 becomes moist.
sociation with the aected glass. No resin seal-
ants are impervious to water vapor and the The above treatment does not attempt to re-
disintegration continues under the sealant move any of the glass corrosion products,
until the glass falls apart. Other glass conser- which often result in layers of opaque glass
vators try to remove the alkalinity from the and which can be removed with various acid
glass to halt the deterioration. treatments. The decision to remove surface
When it comes to eighteenth- and nine- corrosion products that often mask the color
teenth-century sites, we are on much more of the glass must be made on a case-by-case
secure ground. Most, if not all, of the glass basis. Removal of corrosion products can
found from this period was produced from a signicantly reduce the thickness of the walls
stable glass formulation, with few problems and sometimes weaken the glass object. In-
arising other than normal devitrication. discriminate removal of surface corrosion
Since the glass is impervious to salt contami- products can weaken, blur, or alter surface
nation, no conservation treatment other than details. The corrosion layers of a glass object
 can be deemed a part of the history of the Consolidating Waterlogged Glass Using
object and thus a diagnostic attribute. They Passivation Polymers
96 should not be removed without good reason.
Another useful method for halting devitri-
cation is treating waterlogged glass with
Devitrification silicone oil, which, when polymerized, forms
a complex layer and bond with the remain-
Devitrication is a natural process that oc- ing matrix of the glass. As noted in the
curs in siliceous material. In int and obsid- Plenderleith and Werner process, the rst
ian, this process provides the basis for obsid- steps must include extensive washing, rst in
ian hydration dating. The surface of any glass tap water and then distilled water, to remove
from any time period usually becomes hy- soluble salts and debris. After rinsing, the
drated, especially soda glass. Devitrication following steps have proved eective in con-
occurs when the surface of the glass becomes solidating small glass artifacts with silicone
partly crystalline as it absorbs moisture from oils.
the atmosphereand from being submerged First, it is essential to drive o water from
in water. As the glass becomes crystalline, the the artifact. This can be accomplished by us-
surface becomes crazed and akes from the ing two baths of alcohol and allowing the
body. Devitried glass has a frosty or cloudy, glass to remain immersed in each bath for
iridescent appearance. Pane glass is especially hours. For processes utilizing silicone oils,
susceptible to this process. To prevent further additional dehydration using fresh indus-
devitrication and to consolidate the crazed trial-grade acetone is benecial. To prepare
surface, a coating of PVT or Acryloid B- is glass for treatment with a silicone oil solu-
applied. Any of these surface adhesives will tion, the most eective means of water/ac-
smooth out the irregularities in the pitted, etone exchange is to use vacuum to assist in
crazed surface of the glass, making it appear driving o water in the waterlogged glass.
more transparent by lling in the small cracks Acetone is more volatile than alcohol, which
and forming optical bridges. Merely wetting assists in removing water and alcohol from
glass will cause it to appear clearer for the the glass object.
same reason. Used in conjunction with a vacuum
chamber, the process of acetone/silicone oil
displacement is much more ecient than al-
Removal of Sulfide Stains from Lead cohol/silicone oil displacement. When a
Crystal vacuum is applied to glass in treatment, the
boiling point of acetone within the matrix
Leaded glass, which includes a wide variety of the glass is lowered, causing remaining
of stemware and forms of lead crystal, can acetone to be rapidly driven from the glass.
become badly stained by lead sulde. Glass This creates a void, enabling silicone oils to
that is normally clear will emerge from ma- replace water and acetone that have been re-
rine and anaerobic sites with a dense black moved from the artifact.
lm on its surfaces. A % solution of
hydrogen peroxide is used, as with ceramics,
to remove these sulde stains. Besides stain An Effective Silicone Oil Treatment
removal, strengthening of glass artifacts with Strategy
a consolidating resin is often required. Frag-
ments can be reassembled with a good glue The following is a simple and eective pro-
or, if necessary, an epoxy such as Araldite. cess for treating small glass artifacts. It is
essential that most of this work be con- screen over the glass will prevent it from
ducted in a well-ventilated fume hood or moving during treatment. Place the glass in
work area. Rubber gloves should be worn to solution into a vacuum chamber and slowly 97
prevent long-term chemical contact with apply a vacuum. Almost immediately, small
bare skin. Older vacuum pumps, which have bubbles will begin emerging from the artifact.
all-metal components, are more desirable As acetone is driven from the artifact, silicone
than newer pumps when working with ac- oil will displace it. A vacuum of Torr
etone; vapors escaping into these pumps will (. Pa) is sucient to assist penetration
not damage any internal parts. If a modern of the consolidant into the matrix of the
pump is used, a Dewar ask, gas traps, and glass. After allowing the glass in solution to sit
dry ice must be used to prevent acetone at Torr (. Pa) vacuum for
fumes from damaging the plastic compo- minutes, slowly return the artifact in solution
nents in the pump. to ambient pressure. Allow the artifact to sit
in solution for an additional hour before
. After immersing the artifact in a handling (g. .).
container of fresh acetone, place the con- . For the last phase of treatment, two
tainer in a vacuum chamber. Slowly increase small beakers are needed. The rst beaker
the vacuum in the chamber. Initially, there should be lled with sucient CT- catalyst
will be a profusion of large bubbles (g. .). to immerse the artifact. An equal volume of
Over a short period of time, rapid bubbling CR- cross-linker should be poured into
will subside, though smaller bubbles will the second beaker. Once the cross-linker-
continue to be visible. At this point, acetone and catalyst-lled beakers have been
and remaining water are being driven from prepared, the artifact should be carefully
the artifact. The valves of the vacuum removed from the polymer/cross-linker
chamber should then be locked o and the solution and placed into the beaker contain-
artifact allowed to sit in the solution for at ing the CR- cross-linker. Allow the artifact
least hour. to sit in this solution for a few minutes. It
. While water/acetone exchange is taking should then be carefully removed and
place, a suitable consolidant should be allowed to sit on lint-free cloth for a short
prepared. For heavily crizzled glass, PR- has period of time. It is dicult to give an exact
proved to be an excellent polymer. Enough time frame for allowing the artifact to drain,
PR- should be poured into a clean, dry but usually two or three minutes is sucient.
container to ensure that the artifact will be After draining, the artifact should be placed
completely submerged during treatment. For in the CT- catalyst for a short period of
small artifacts, plastic-coated paper cups are time. Monitor this phase of treatment
useful and convenient. To this volume of carefully. After a few seconds, a zone of
PR-, add a % addition of CR- (by chemical reactivity will be noticed around
weight). This solution needs to be mixed the artifact. Generally, the artifact is left in
thoroughly, using a nonporous stir rod. Rapid the catalyst for the same period of time as it
mixing will result in bubbles in the solution. was treated in the CR- solution. It should
These can be removed, if desired, by placing then be removed, very lightly surface-wiped
the polymer solution into a vacuum chamber with lint-free cloths, and the processes of
and applying a slight vacuum. immersion in CR- and CT- should be
. Once the polymer solution is ready, repeated, at least one more time. For most
carefully and quickly remove the artifact from artifacts, two or three immersion cycles are
the acetone and immerse it in the PR-/CR- sucient to create a stable, consolidated end
solution. A small piece of aluminum product (g. .).
 . After treatment, the CT- catalyst needs
time to work. Some experimentation is
98 required but, generally, only a very light
surface-wiping of the artifact is required to
remove free-owing silicone oil solution
from its surfaces. The glass should then be
placed into a well-ventilated environment
and allowed to air-cure for several hours
before handling. It is possible to touch
completed artifacts within a few minutes, but
for a complete cure, the process may take as
long as hours. For larger artifacts, or in
cases where additional contact with an active Fig. .. The glass bead in solution during the
catalyst may be desirable, the nished artifact acetone/polymer displacement process.
and a cloth with several drops of CT- can
be placed into a glass jar with a tight-tting
lid. When the jar is tightly sealed, the artifact with a uniform, satin nish. If this doesnt
will react with the catalyst fumes, causing remove pooled silicone from the artifact, use
additional or more rapid polymerization. the edge of a soft wooden dowel or tooth-
pick to gently mechanically clean the poly-
Posttreatment Aesthetics mer from the artifacts surface.

After treatment, the glass may require addi- Analysis of Polymer-Treated Glass
tional cleanup to remove surface deposits of
silicone. In most cases, pooled silicone can Analysis of silicone-treated glass indicates
be easily removed with a soft cloth. For larger that a complex, consolidating coating has
surface pooling, wipe the artifact surface with been formed on the surface of treated glass.
Fig. .. Rapid bubbling a soft cloth containing a few drops of CR- Chemical analysis of the glass also indicates
during the acetone cross-linker. This is usually sucient to re- that silicone polymers are present in all the
dehydration process. move surface polymer and leave the artifact test sites within the core of the sample. Fig-
ure . is a cross section of a piece of treated
glass. Three distinct layers are visible in the
cross view.

Reconstruction

Glass can be repaired and reconstructed with

the same glues used for pottery. However,
optically clear epoxy resins are generally
used because they adhere to the smooth,
nonporous glass more readily; they dry
clearer and shrink less than the solvent res-
ins; and they are, therefore, less noticeable.
The bonds are also stronger. The epoxy res-
ins, however, are usually irreversible. Hysol
Epoxy with Hardener and Araldite
are the two brands most commonly used in
glass repair. The new super glues made of
cyanoacrylate are used quite often to piece 99
glass together quickly. After using the cy-
anoacrylate, epoxy is owed into the cracks
with an artists brush to permanently glue the
pieces. It is exceptionally dicult and time-
consuming to gap-ll glass. Considerable
work and experience are required. The prob-
lem of matching transparent glass colors is
equally dicult. All these problems are dis-
cussed in greater detail in Conservation of
Glass.7
As is the case in all conservation, the con-
servator must be able to recognize what the EDTA, tetra sodium. Iron stains are com- Fig. .. The glass bead
problems are and know what can be used to monly removed with % oxalic acid or % during treatment in the
counter them. In glass conservation, when EDTA, disodium. CR-/CT- cycle.
lead oxides are found on glass, they can be Realistically, few problems, other than re-
removed with % nitric acid; % sulfuric construction and restoration, are likely to be
acid can be used to remove iron oxide, to encountered on any of the glass objects found
neutralize the alkalinity of glass that is break- in archaeological sites dating from the mid-
ing down, and occasionally to remove calcar- eighteenth century to the present. Essentially,
eous deposits. Calcareous deposits are com- the same chemicals and equipment required
monly removed with % hydrochloric acid for treating ceramics are used for conserving
and, on some occasions, by immersion in % glass.

Fig. .. Posttreatment
cross-sectional, micro-
scopic (X ,) view of
the surface of the treated
glass. Three distinct
polymer layers are visible.
 Case Study: Preservation of Seven- CR- is a trifunctional monomer resis-
teenth-Century Glass Using Polymers tant to solvents and photo-oxidation.11 When
100 used as a pretreatment to the edges of glass
Long-term stabilization of waterlogged glass being repaired using an adhesive, silane
from Port Royal, Jamaica, has been a problem. groups in the monomers act to reduce water
Because of the lack of controlled-environment absorption into the adhesive joint, resulting
curation, many of the bottles preserved us- in a longer-lasting bond with better adhesion
ing conventional polyvinyl acetate (PVAC) qualities. This means that the junction of the
treatments have been lost due to exfoliation repaired pieces of glass is less prone to ab-
of surface layers and eventual breakdown of sorption of atmospheric moisture than the
the glass. Experimentation, aimed at devel- glass surrounding the site of repair. The abil-
oping conservation strategies to preserve this ity of silane to form strong Si-C bonds and
assemblage, has generated new tools and to reduce moisture absorption makes it po-
questions regarding artifact conservation tentially benecial for the stabilization of ar-
using polymers. Because of environmental chaeological glass.
factors, PVAC polymer processes are Glass made in England during the seven-
ineective in stabilizing the Port Royal glass teenth century consisted of approximately
assemblage. This has fueled our experimen- % sand (silica) and a uxing agent made
tal research into the use of siloxane polymers from plant ash. Fluxing agents were used to
for archaeological preservation. reduce ring temperatures below C.
While PVAC is considered a stable con- Glass formed at lower temperatures are sus-
solidant that is resistant to yellowing, it has a ceptible to hydration. Lime was added to the
low glass transition temperature (Tg ) (C) silica/ux mixture to make the resulting glass
and is prone to cold ow.8 Dust and dirt tend harder, more durable, and resistant to hy-
to accumulate on the surfaces and deep crev- drolysis. A two-stage reaction occurs on the
ices of glass treated with PVAC.9 Both these surface of the glass during extended periods
factors appear to aect the Port Royal glass of immersion in water. Sodium ions are ex-
assemblage. Over time, bottles exposed to changed for hydrogen ions, resulting in a re-
temperature variations, humidity, and ultra- duced volume of the surface glass. This ex-
violet light have developed thick layers of change produces a surface layer of material
oxidation on their surfaces. In some cases, that is more siliceous than the original glass.
the outer surfaces of the bottles have exfoli- Higher Si content in the crusted layers of glass
ated, resulting in the loss of diagnostic suggests that extensive ion exchange has oc-
attributes. curred.
One interesting avenue of research has All the onion bottles in the Port Royal as-
been the use of aminopropyltrimeth- semblage are dark grayish green. According
oxysilane, a common silane coupling agent, to Plenderleith and Werner, this color is due
which has been used to facilitate the bond- to high concentrations of iron oxides.12 To
ing of two pieces of glass when epoxy sys- remain chemically stable, glass should con-
tems are being used as adhesives to repair tain at least % silica and not more than %
broken artifacts. Coupling agents improve alkali metal oxides. Additionally, glass re-
the adhesion of polymers to glass.10 Because quires between % to % calcium oxide.
of its ability to retain strong Si-C bonds, CR- Glass containing higher concentrations of
was included with other polymers in an alkali metal oxides is potentially hygroscopic.
experiment to determine which viscosities Alkali ions will diuse and leach throughout
(centistoke) of polymers best preserve ar- a porous, silica network attracting other ions
chaeological glass. and mineral-rich sediments surrounding the
glass. This can potentially result in increased then subjected to accelerated weathering tests
calcium in hydrated glass. In turn, these el- to determine if each polymer protected the
evated calcium levels are associated with glass from environmental damage. 101
extensive surface crazing that reduces the
transparency of the glass and results in occa- Pretreatment Analysis of the Glass
sional spalling of the surface. Accordingly, the
outer surfaces of many bottles in the assem- Prior to treatment, scanning electron micro-
blage are pitted and frequently concreted scopic analysis and neutron activation analy-
with white carbonate deposits. sis were conducted on all glass shards.
Since the sinking of Port Royal in , the Microscopy revealed a distinct outer layer or
glass assemblage has been subjected to hy- outer crust on the samples (g. .). The in-
drolytic attack under pressure. Selective ner surface of the weathered crust, closely
leaching of soluble materials, including alkali associated with the solid surface of the bottle,
metal oxides, has resulted in a high silica resi- contained low proportions of calcium and
due in the weathered crust on the surface of high readings of magnesium and silica. The
the glass. Analysis of the outer crust indicates inner core of glass contained higher propor-
that alkaline ingredients, originally essential tions of calcium. This variability resulted
in the forming of the glass, are missing and from absorption of magnesium from seabed
have been replaced with water. sediments with consequent leaching of cal-
The purpose of this experiment was to cium from the core of the glass outward to
evaluate the ability of selected short-chain, the surface of the vessel. The greatest area of
silanol-ended polydimethyl siloxane poly- chemical exchange appeared to be between
mers to conserve similarly sized shards of the inner edge of the weathering crust and
devitried glass. The mechanism of impreg- the outer surface of the inner glass core. The
nation used to introduce these polymers into crusts silica content was higher than that of
each glass shard was acetone/polymer dis- the core glass. Distribution of elements in the
placement. This is a commonly applied pro- crust and core glass edges are compared in
cess using the rate of evaporation of acetone, table ..
at reduced or ambient pressure, to displace Side view micrographs of the surface of
acetone with a polymer solution.13 Cataly- glass samples from Port Royal indicated that
zation of the polymer-impregnated glass was the surfaces are generally coated with a thick
accomplished using vapor deposition of a crust of oxidized glass, soluble and insoluble
tin-based catalyst. All treated samples were salts, and concretion from close association

Table .. Sample Locations and Elements

Element (parts per million)

Na Mg Al Si K Ca Fe Cl

Location of Sample
Crust
Outer Edge .44 3.09 4.69 69.97 5.74 9.17 2.96 2.21
Inner Edge .41 4.88 1.96 73.50 4.07 6.11 7.13 1.07

Glass Core
Outer Core .65 1.33 .89 53.14 7.76 30.96 3.62 .74
Inner Core .53 1.15 .74 52.89 8.26 31.63 3.12 .97
102

Fig. .. Port Royal onion B

bottle OBG: (A) outer
surface; (B) inner surface
of siliceous outer layer of
glass. Silicone attachments
spread outward from the
core surface to the outer
layer of glass.

with other artifacts. The surface crust is eectively. Inspection of several bottles indi-
hydrophilic, containing amorphous silica cated the consolidant only partially pen-
compounds. The core of the glass is highly etrated the outer surfaces of the glass. Because
hydrated, resulting in a complex matrix that the consolidant did not form an eective bar-
is dicult to stabilize. rier on the surface of the core glass, deterio-
The general instability of the Port Royal ration was not halted. Additionally, the
onion bottles may be due, in part, to the vis- consolidant did not seem to act as an adhe-
cosity of PVAC V and its inability to bond sive between layers of aking glass, which
may have caused the high rate of surface de-
terioration. Lower viscosity consolidants may
penetrate throughout the matrix of an arti- 103
fact but fail to act as an adhesive between the
crust and core, resulting in poor consolida-
tion of the glass.

Overview: Conservation of the

Port Royal Glass Assemblage

Preservation of the Port Royal glass assem- Fig. .. Exfoliating

blage has been conducted using a process of surface layers on a Port
slow desalination, followed by dehydration Royal onion bottle
in successive baths of organic solvents, and preserved with PVAC V.
concluding with consolidation by impregnat-
ing the glass with polyvinyl acetate. Vinyl glass, resulting in a supercial skin that is
polymers, like PVAC, have a refractive index prone to spalling. Lower-viscosity consoli-
similar to glass and generally remain stable. dants may penetrate throughout the matrix
Accordingly, many conservators prefer to use of the glass but fail to act as an adhesive be-
vinyl polymers for conserving waterlogged tween the crust and core, resulting in poor
glass. While lower molecular weight solutions consolidation of the glass.
of PVAC are resistant to yellowing, they are In order to successfully conserve water-
prone to cold ow.14 This can be a signicant logged glass with silicone oils, water within
problem for the long-term stability of the its matrix must be displaced using a volatile
glass. solvent such as acetone. Acetone is then re-
Initially, Port Royal onion bottles pre- placed by a hydroxyl-ended, functional poly-
served with PVAC appeared to be stable. mer, mixed with a cross-linking agent. To
Unfortunately, inadequate facilities for facilitate displacement of acetone with poly-
curation have caused several of these artifacts mer, each sample was placed into a vacuum
to become unstable. For example, exposure chamber. At reduced pressure, the boiling
to high humidity and uctuating tempera- point of acetone is lowered, and thus acetone
tures has resulted in extensive exfoliation (g. is rapidly vaporized and driven from the
.) and discoloration of the outer surfaces glass. Rapid solvent vaporization allows the
of the glass. In a few cases, entire artifacts have polymer/cross-linker solution to permeate
deteriorated. the matrix of the glass more easily.
Chemical analysis of the Port Royal as-
semblage indicates that the surfaces of the Materials
glass samples are generally coated with a
thick crust of oxidized glass, soluble and in- The rst onion bottle shard, OBG, was
soluble salts, and concretion from close as- treated with PR- polydimethyl siloxane,
sociation with other artifacts. The surface which is a hydroxy-terminated silicone oil
crust of the glass is hydrophilic, consisting with up to % dimethyl cyclosiloxane added.
of amorphous silica compounds, whereas PR- was the longest chain silanol-ended
the core of the glass is highly hydrated. The polymer used for this experiment. It is a vis-
resulting complex matrix is dicult to sta- cous, odorless polymer ( cts). OBG was
bilize. Viscous materials such as PVA V do treated with Dow Corning PA Fluid (
not penetrate easily into the solid core of cts), a short-chain silanol-ended, poly-
 Table . Treatment Summary for Glass Fragment Samples

104 Sample Acetone Dehydration Silane

OBG1 Acetone PR-10
OBG2 Acetone PA fluid
OBG3 Acetone KP80
OBG4 Acetone 4-7041
OBG5 Acetone no silanol
OBG6 No acetone no silanol

Note: For all samples, cross-linker was 3% CR-20; catalyst was CT-32.

dimethyl siloxane, with octamethylcyclo- at room temperature. Cessation of rapid

tetrasiloxane and decamethylcyclopenta- bubbling was used as a visual indicator that
siloxane added. OBG was treated with KP free-owing water and air had been displaced
( cts), which is also a silanol-ended, by acetone during dehydration of all the
polydimethyl siloxane. OBG was treated shards.
with Dow-Corning -, a very short chain All the samples, were then placed into their
silanol-ended, polydimethyl siloxane ( respective polymer/cross-linker solutions
cts). No silicone oil was applied to OBG. In- and treated in reduced pressure environ-
stead, this shard was dehydrated in acetone ments for hours. Sample OBG was not
and then impregnated with CR- cross- dehydrated prior to treatment. Alternately,
linker. The samples were catalyzed using CT- OBG was immersed in ml of CR- and
catalyst. Sample OBG was treated with treated in a reduced pressure environment
CR- cross-linker without rst being dehy- of . Pa ( Torr) for hours. Data for
drated. The sample was then catalyzed with type of dehydration, silanol name, percent-
CT-. age of cross-linker applied by weight, and
CR- cross-linker is a hydrolyzable, mul- catalyst for the six glass samples are summa-
tifunctional alkoxysilane, capable of tying rized in table ..
two or more polymer chains together. CT- After hours of acetone/polymer solu-
catalyst was applied in vapor form to poly- tion displacement, each sample was placed
merize the shards. into a plastic container with a tight-tting
lid. For this experiment, the lid and body of
Experiment Procedures a liter plastic container were placed in an
inverted position so the lid formed the at
Prior to treatment, six similarly sized frag- base of a containment chamber. A -g ca-
ments of glass from a broken onion bottle pacity aluminum dish was placed in the cen-
were rinsed in a series of baths of deion- ter of the base of the containment chamber.
ized water to remove soluble salts. Each bath Fifteen grams of CT- catalyst were placed
was exchanged for fresh deionized water af- in this dish. A section of aluminum mesh was
ter hours. Each of ve samples, OBG, placed over the aluminum dish to form a
OBG, OBG, OBG, and OBG, were then platform on which a shard of glass could be
placed into a beaker containing ml of positioned during catalyzation. Several pa-
fresh, industrial-grade acetone. Immersed in per towels were placed on the screen to ab-
the solvent, each sample was placed into a sorb silicone oils owing from the glass dur-
vacuum chamber and a reduced pressure of ing the nal stages of treatment. With each
. Pa ( Torr) was applied for hours sample positioned in its containment cham-
ber, the body of the container was placed over
the lid and, with rm pressure, snapped into
place, sealing each piece of glass within the 105
container. A small hole was made in the up-
per surface of each chamber to prevent pres-
surization during the catalyzation process
(g. .).
Positioned in their containment cham-
bers, the six glass samples were placed into a
vented warming oven and exposed to a con-
stant temperature of _C for hours. The
samples were then removed from the oven
and from their containment chambers and
placed on lint-free cloths, exposed to air in a
fume hood. All samples were evaluated for
color, transparency, opacity, dryness, and sur-
face scaling. After hours in air, the samples
were reevaluated using the same criteria.
Accelerated Weathering Tests Fig. .. Containment
Observations of Polymer Stabilization chamber conguration;
All samples were subjected to ve months of (A) body of chamber;
Upon completion of the acetone/polymer accelerated weathering testing using a (B) warmed catalyst vapor
displacement and catalyzation processes, all Q-Panel, Q-U-V Accelerated Weathering in close proximity to
artifact; (C) glass artifact;
the glass shards were removed from their Tester. Weathering cycles consisted of alter-
(D) aluminum screen and
containment chambers and allowed to sit in nating -hour cycles of high humidity (%)
paper towel; (E) g of CT-
fresh air. Two evaluations of each shard were and high temperature (C) with a UV-A
catalyst in aluminum
conducted. After ve minutes of exposure to light, followed by hours at lower humidity dish; (F) lid acting as the
air, OBG, OBG, OBG, OBG, and OBG (%) and temperature (C) with no light base of the unit.
were very natural in appearance. OBG re- exposure. The UV-A spectrum, to na-
mained transparent after hours in the air. nometers, is the spectrum that causes the
After hours, no oxidation was noted on most damage for many polymers including
OBG; the shard retained its coloration. polyethylene, PVC, acrylic, and polyvinyl
OGBG appeared to darken slightly after acetate.
treatment. While the shard was aesthetically
pleasing during its rst evaluation, slight scal- ObservationsAccelerated Weathering
ing and oxidation were noted during the -
hour evaluation. No changes were noted over After weathering, the samples were taken to
the -hour period of evaluation for OBG. the Scanning Electron Microscopy Lab at
The sample retained its opacity, dryness, and Texas A&M University for analysis using elec-
dark green color. OBG was slightly opaque tron dispersal spectroscopy and visual analy-
when rst evaluated. After hours, the shard sis. All samples were weighed and measured
was heavily oxidized and extensive aking to determine what structural changes resulted
was evident. OBG was slightly cloudy in from accelerated weathering. (See g. ..)
appearance when rst evaluated. After Sample OBG remained constant in
hours, the sample had a slight iridescent shine weight and general appearance after acceler-
on its surfaces. Data for all samples are com- ated weathering. Width measurement, taken
pared in table .. at the widest point of the sample, indicated
 Table . Observations of Glass Samples Conserved with Polymers

106 Sample Color Opacity Dryness Scaling Appearance

Time in Air
OBG1

5 min Light green None Tacky None Good
24 hr Light green None Dry None Good
OBG2
5 min Light green None Dry None Good
24 hr Light None Dry None Very good
OBG3
5 min Darker Slight Dry None Good
24 hr Darker Increasing Dry Surface flakes Poor
OBG4
5 min Dark green None Dry None Excellent
24 hr Dark green None Dry None Excellent
OBG5
5 min White/green Opaque Dry Some noted Poor
24 hr White pink/green Opaque Dry Extensive flaking Very poor
OBG6
5 min Dark green Slightly opaque Dry None Poor
24 hr Dark green Slightly iridescent Dry None Poor

shrinkage of .%. The mass of OBG was Conclusions

reduced by .% after accelerated weather-
ing. Width measurement indicated shrinkage This experiment demonstrated a direct cor-
of .%. OBG measured . g after ac- relation among polymer chain length, pen-
celerated weathering, indicating a reduction etration into cracks and voids within the
in mass of .%. Posttreatment shrinkage matrix of the glass, and ability to create a
in this sample was .%, as measured at the protective barrier against oxidation and
widest point of the sample. OBG was re- environmental destruction. In general,
duced in mass by .%. The width of this short-chain polymers appear to preserve wa-
sample shrank by .% after accelerated terlogged glass better than longer-chain
weathering. OBG was reduced in mass by polymers. This suggests that viscosity (chain
.% after accelerated weathering. The length) is a factor determining the ability of
sample also shrank .% as the result of treat- polymers to displace acetone in degraded
ment. OBG experienced the greatest loss of glass.
mass, with a recorded reduction of .%. OBG was treated with PR-, the long-
Physical shrinkage for the sample was .%. est chain polymer used in this experiment.
Aesthetically and dimensionally, OBG Following processing, the shard was aes-
and OBG were the most resilient to acceler- thetically pleasing. After accelerated weath-
ated weathering tests. During microscopic ering, changes in the width of the sample
analysis, shard OBG was accidentally were noted. Results of shard OBG, treated
dropped on the oor and broke into two sec- with PA Fluid ( cts), were also very
tions. The new break oered an opportunity good. Evaluation of OBG, treated with
to view the interior core of glass. A slight band KP ( cts), indicated that the poly-
of discoloration was noted. Over time, the mer had not protected the glass from envi-
size and appearance of this discoloration did ronmental destruction. A small amount of
not change posttreatment opacity was noted on the
surfaces of the shard, and after accelerated
weathering, dimensional changes and
aking were noted. 107
OBG, treated with -, the shortest
chain polymer ( cts), was evaluated as the
best preserved of the glass shards. After ac-
celerated weathering tests, color, clarity, and
dimensional stability were maintained. Pres-
ervation results for OBG and OBG were
predictable. The use of acetone dehydration,
with no silanol to create a barrier, acceler-
ated oxidation in shard OBG. Since OBG
was not dehydrated, the application of a cata-
lyst may have bonded some of the water in
the sample, preventing the high degree of
oxidation and structural change noted in
OBG. Posttreatment aesthetics and dimen-
sional changes of the six glass samples, and
images of the shards are illustrated in gures
. and ., respectively.
Scanning electron microscopy analysis of Fig. . Posttreatment
evaluation of glass
the glass samples showed that a complex bar-
samples OBGOBG.
rier coat had been formed on the surfaces of
OBG was the most
OBG, OBG, OBGand OBG. Polymers aesthetically pleasing
had permeated deeply into the voids of these glass. The shard remained
shards, possibly forming an eective barrier dimensionally stable with
within the glass. The small band of discol- comparatively little weight
oration noted in OBG at the point of break- loss during accelerated
age was assumed to be slight oxidation re- weathering tests.
sulting from exposure of interior glass to air.
As noted, however, the amount and inten-
sity of the discoloration did not change over
time. Microscopic analysis of the broken edge OBG1 OBG2
indicated that stress fracture lines caused the
discoloration thought to be oxidation. The
presence of stress fracture lines also aected
the coloration of glass. Fracture lines aect
light refraction and glass density, resulting in
a change in coloration.
Analysis of the chemical structure of OBG OBG3 OBG4
indicated that the silicone (Si) peak was sub-
stantially larger than that of calcium (Ca). In
untreated control samples of glass, Si and Ca
peaks were similar in proportion, suggesting
that the presence of additional silicone was
partially masking the presence of calcium. Iron Fig. .. Photos of glass
was present in the treated glass samples. This OBG5 OBG6 samples OBGOBG.
 was probably due to the close association of removed from the concreted sediments on
these glass shards to iron tools and house- the interior of the bottle. To avoid additional
108 hold implements within the wreckage of the stress to the bottle caused by dierences in
submerged buildings at Port Royal. temperature, the bottle neck was placed in
an initial dehydration bath of acetone, with
great care being taken to ensure that both the
Case Study: Preserving Waterlogged acetone and rinse water were as close to the
Glass and Cork same temperature as possible.
After hours, the artifact was removed
After extensive evaluation of the conserva- from the bath and placed into a clean beaker
tion procedures used to preserve an entire with new acetone. After an additional
onion bottle from the excavations at Port hours, the artifact was removed from this
Royal, Jamaica, the neck section of another bath and placed into another clean beaker
bottle with its cork still in place was chosen with more new acetone. The beaker contain-
for conservation by silicone bulking using ing the artifact was then placed into a vacuum
PR- silicone, which is the most viscous chamber mounted in a chest freezer. With the
polymer generally used in our laboratory. temperature maintained at C, a vacuum of
Bulking the cell structure of waterlogged . Torr (. Pa) was applied to the ar-
corks with PR- silicone had been very suc- tifact in its acetone bath for hours. After
cessful, and we believed the potentially hours of applied vacuum, vigorous bubbling
greater bulking capacity and increased cohe- of the acetone had decreased substantially,
sive strength of this silicone were advanta- and for the remaining hours, smaller and
geous in conserving this artifact. Because less frequent bubbles were observed in the
successful bulking and stabilization of ar- acetone solution. At this point, the vacuum
chaeological waterlogged samples of glass was turned o and the artifact was allowed
and cork had been conducted, it was felt that to sit in the freezer for hours.
their combination as a composite artifact With dehydration completed, the artifact
posed no new problems or concerns for the was placed into a clean beaker, and PR- sili-
stabilization of this onion bottle neck. Prior cone was quickly poured into the container
to treatment, a large section along the length so that the artifact was completely immersed
of the bottle neck was noted to be loose to in the silicone solution. The PR- siloxane
the touch. had been stored alongside the sample inside
Before conservation, this artifact, along the freezer to ensure that the artifact was not
with numerous other onion bottles, had been subjected to additional stresses caused by
stored in a bath of fresh tap water for at least dierential temperatures during the transfer
two years. The bottle neck was carefully to the bulking solution.
cleaned in a fresh tub of water, using nger- After carefully decanting o the last bath
tips to remove any algae and debris on the of acetone, we quickly transferred the artifact
surfaces of the glass. The cork appeared to to a clean beaker and added PR- silicone
be in an advanced state of disintegration. Due until the bottle neck was completely sub-
to its fragile state, no attempts were made to merged in the uid. The beaker containing the
clean its deep cracks and surfaces. artifact and bulking agent was then placed
Because of the artifacts small size, the de- back into the freezer-mounted vacuum cham-
hydration process was accelerated in com- ber. As before, a vacuum of . Torr (.
parison to that used for the complete onion Pa) was applied for hours. The vacuum was
bottle, which required lengthy dehydration then turned o and the artifact was stored
to ensure that as much water as possible was overnight in the freezer.
 After approximately hours, methyl- the chamber and allowed to sit for hours
hydrocyclosiloxane (% by volume), a cross- (g. .). The artifact was then inspected to
linking agent, was added to the PR- solu- make sure that the excess silicone had not 109
tion and a vacuum of . Torr (. Pa) dripped into the catalyst. A thin layer of
was applied for hours as a means of draw- crusted silicone had formed on the surface
ing this agent into the matrix of the glass. of the catalyst, so the dish was removed mo-
Bulking as an additive process, in which mentarily, and two more ounces of CR-
cross-linkers are added to the bulking com- were added to the tray. With fresh catalyst in
pound, tends to leave voids in cells and within place, the warming oven was closed and the
the physical structure of the artifact more artifact was allowed to cure undisturbed for
completely bulked than when cross-linkers hours.
are applied as a separate bath. In a two-stage After curing, the surfaces of the bottle and
process, the act of bulking the artifact with cork were dry to the touch and the paper
cross-linking chemicals appeared to remove towel, which had absorbed excess silicone,
free-running silicone from structural voids had become sti with polymerization sili-
within the artifact. Because we wanted to ll cone. The bottle neck was removed from the
the voids of the glass in the hopes of sealing oven and allowed to sit at room temperature
it from air and other detrimental elements, for approximately hours. Two days after
the additive process seemed to be the better being removed from the warming oven, a
conservation approach. It would ensure that small sample of glass was removed from the
as much silicone as possible was left in con- base of the loose section of the bottle neck.
tact within the microssures and small voids Additionally, a small sample of cork was
of the glass. taken from the interior surface of the cork.
With the removal of free water and dis- These samples were set aside for electron
placement bulking with silicone and cross- microscopic analysis. As a nal step, the fresh-
linking agents complete, the bottle neck was break surface of the glass was sprayed with
removed from the solution and all excess u- several thin coats of Krylon clear spray
ids were allowed to drain o. After a couple of to ensure that the matrix of the glass was pro-
minutes, all free-owing silicone had drained tected from the environment.
from the surfaces of the artifact and the bottle
appeared to be thinly coated in silicone. A Microscopic Analysis
warming oven was preheated to C. To ex-
pose the artifact to the concentrated fumes of Scanning electron microscopy was conducted
the catalyst, a containment chamber was at the Electron Microscopy Center at Texas
placed in the warming oven. Consisting of an A&M University. To determine the degree of
inverted polypropylene pail with a tight-tting bulking that had occurred in the cork and
lid acting as the base of the chamber, this de- glass samples, scanning electron microscopy
vice is inexpensive and easily adaptable for was necessary to both view and then imple-
variations on the curing process. ment X-ray microanalysis as a means of iden-
Two ounces of CR- catalyst were placed tifying features in the cross sections and to
on a at dish in the center of the base of the determine basic chemical compositions. Us-
chamber. A piece of aluminum screen was ing a Joel JSMT A scanning microscope
placed over the top of the dish to form a plat- set at a working distance of mm, a series
form for the bottle neck during the curing of pictures were taken of the cross-sectional
process. A paper towel was placed on the views of both samples.
screen to absorb any additional free-owing Inspection of the cork sample revealed
silicone. The bottle neck was then placed into that heavy bulking had occurred. In gure
 ., the cross-sectional microscope view of Prior to analysis, all samples taken from the
the cork magnied at times, reveals that bottle neck and cork were coated in gold/pal-
110 nearly all surfaces of the cell structure are ladium and then mounted on stubs. Because
heavily coated with silicone and, in many this microscope was designed to allow the op-
cases, the cells appear to be completely lled. erator to take photographs of samples as well
SEM images of the glass neck remnant sur- as give nearly instantaneous charted analysis
rounding the cork indicate the formation of of the elements found within features of the
a complex layer of polymer throughout all samples, it was the ideal tool for analyzing the
the cracks and voids of the glass. cross sections of cork and glass.
The targeted area for analysis was high in
Fig. .. Curing chamber silicone with an intensity reading averaging
X-Ray Microanalysis
and warming oven setup:
over counts per second. Carbon, oxygen,
(A) warming oven;
To verify the observations made during mi- magnesium, and calcium were also recorded
(B) containment chamber;
(C) neck of bottle with
croscopy analysis, additional analysis of the in this sample although these elements, as
cork; (D) aluminum chemical composition of the elements identi- well as the gold and palladium coating on the
screen on which artifact ed in the cross sections of glass and cork was sample, were well below the intensity read-
sits during catalyst necessary. Using a Joel JSM scanning ings of silicone in the glass. Figure . is a
deposition; (E) aluminum microscope, X-ray microanalysis of cross sec- -times magnication of a cross-sectional
tray holding g of CR-. tion features of both samples was conducted. view of silicone-bulked archaeological glass.
Analysis of the tendrils, as indicated by the
target zone labeled A, revealed that they were
almost entirely silicone. Similar ndings were
noted on other areas of this sample.
Analysis of the silicone-bulked cork sample
was most impressive. When viewed at -
times magnication, the cell structure and
surface voids of the cork sample appear to
have been heavily lled with silicone. Target-
ing the large mass in the center of gure .,
analysis indicated that this area of the sample
was heavily bulked with silicone. Small
amounts of carbon, oxygen, gold, and palla-
dium were also noted in the target area but
in much smaller amounts in comparison to
silicone. Additional readings were taken along
the cell walls in several areas. With a slight
increase in the readings of carbon in these
areas as a result of cell wall structures, sili-
cone remained the dominant substance
identied.

Observations

The assumption that the glass and cork in

this waterlogged onion bottle neck could be
conserved successfully as a compound arti-
fact appears to have been correct. This may
 111

have been possible in part because of the pounds may penetrate into the glass well but, Fig. .. Microscopic
highly degraded state of the cork. Micro- characteristic of their size, they should not (X ,) view of silicone
scopic observations and X-ray microanaly- oer as much mechanical strength and cohe- oiltreated waterlogged
sis determined that voids and cell structures sive bonding ability within the matrix of the cork.
of both materials were heavily bulked with devitried glass. Unlike the end surfaces of the
silicone. The solid state of the silicone within cork, which are visible and feel slightly rub-
these structures further indicates the eec- bery to the touch, the interior and exterior
tiveness of the processes of cross-linking and surfaces of the onion bottle glass feel dry and
catalyst polymerization. natural in texture. In both materials, PR- has
The process of bulking these artifacts us- acted to ll cellular voids, as well as cracks and
ing an additive process in which cross-linking cavities, and seems to be an excellent optical
agents were added to PR- ensured the most bridge since it is not shiny or unnatural in
complete bulking possible. It would be inter- appearance. Several thin coats of Krylon
esting to replicate this experiment using clear spray were applied to the outer surfaces
lighter molecular weight siloxanes such as PR- of the bottle neck, acting as an additional seal-
. Because of their smaller size, these com- ant against environmental exposure.
 CHAPTER 8
112

Ivory and Bone

I
vory and bone artifacts have been recov collagen), and the main inorganic materials
ered from nearly all excavations con are calcium phosphate associated with car-
ducted by Texas A&M University and the bonates and uoride. Viewed under a micro-
Institute of Nautical Archaeology. Excava- scope, bone has a coarse grain structure made
tions of the seventeenth-century submerged up of collagen bers with lacunae or voids
city of Port Royal, Jamaica, which perished throughout. A crystalline inorganic material
in a catastrophic earthquake in , have un- called hydroxyapatite surrounds these col-
covered bones in cooking pots from meals be- lagen brils, providing strength to the bone.
ing prepared at the time. Skeletal remains of The chemical structure of hydroxyapatite is
young children who perished when they were generally listed as Ca10(PO4)6(OH)2.
trapped indoors have also been recovered. Bone and ivory are anisotropic, meaning
Excavations carried out by the Texas His- that their collagen brils and associated
torical Commission on the vessel La Belle structures have directional properties. Bone
have led to the recovery of the bones of at and ivory artifacts have dierent mechanical
least two sailors. Buried in the anaerobic en- and strength characteristics along their
vironment of Matagorda Bay, o the south lengths, widths, and thicknesses. Accordingly,
Texas coast, these remains were so protected artifacts made of these materials are struc-
that a large volume of brain matter was well turally weakened when they are altered and
preserved in the cranium of one victim. Small used for manufacturing purposes. As the
pieces of cartilage were also recovered, and structural layers are removed or altered, as
generally, the skeletal remains were in excel- in the case of carvings, the support mecha-
lent condition. While skeletal materials of- nism of the bone is weakened.
ten survive well in marine environments, An important component of ivory is den-
bones recovered from land sites may be in tine, which is very hard and dense. When
poorer condition, depending on soil acidity, viewed through a microscope, this matrix
pH, and the presence of water. looks very compact in cell structure with very
few voids. Characteristically, ivory appears
to have a network of lenticular areas, which
Basic Structural Differences result from the intersection of a system of
striations radiating out from the center and
The structure and function of the components increase in number as the tusk grows. As ivory
of bone are complex. From an archaeological ages or becomes waterlogged, this radiating
perspective, however, the major organic com- pattern may appear similar to growth rings
ponent of skeletal materials is ossein (protein in a tree.
 Prolonged immersion in water will cause Equipment Setup for Very Fragile
hydrolysis of the ossein, resulting in poten- Bone and Ivory
tial warping and deterioration of an artifact. 113
Care must be taken with heavily degraded ar- In The Conservation of Antiquities and Works
tifacts to prevent loss of diagnostic attributes of Art Plenderleith and Werner described a
due to collapse of the cellular structure. Inor- simple vacuum chamber system (g. .).
ganic salts and hydroxyapatite making up the Instead of a vacuum pump, they suggested
rest of the structure are aected by acidic con- using a water pump system to supply slight
ditions. Over time, waterlogged bone and vacuum to achieve thorough impregnation
ivory absorb organic and inorganic salts. of a bulking agent into the porous structure
Desalination, therefore, is an important of an artifact.1 This system works well in situ-
phase in treatment of artifacts from saltwa- ations where the use of a high eciency
ter environments. If salts are not removed, vacuum pump may be too powerful and po-
artifacts may become distorted and, in some tentially hazardous for delicate artifacts. This
cases, suer from surface exfoliation as crys- form of water-driven vacuum treatment is
talline structures form during air-drying. slow but highly eective for the treatment of
Bone and ivory recovered from terrestrial bone and ivory artifacts that are small, thin
sites are aected by acidic soil. Hydroxyapa- walled, or both. Many of the artifacts pro-
tite dissolves in acidic conditions, leaving cessed at the Archaeological Preservation Fig. .. Conguration for
unsupported anisotropic collagen brils to Research Laboratory are more robust. These water ow vacuum
shrink and warp as the artifact is air-dried. artifacts can be safely processed using a sys- processing of delicate
In alkaline conditions, the organic collagen tem similar to that proposed by Plenderleith artifacts: (A) desiccator/
hydrolyzes and is often attacked by bacteria and Werner but substituting a high volume, vacuum jar with valve;
(B) beaker containing
and other microbial activity. This often re- electric vacuum pump (g. ., F). Because
artifact and either acetone
sults in the hydroxyapatite becoming brittle. of the presence of volatile vapors resulting
or polymer solution;
Thus bone recovered from alkaline soils may from solvent dehydration, a double gas trap
(C) in-line vacuum gauge;
be structurally unstable and very crumbly. assembly (C) immersed in a container of dry (D) in-line trap to catch
Passivation Polymers have proved to be ice or liquid nitrogen (D) acts to eciently solvent or polymer
useful for the preservation of bone and ivory freeze gases escaping the vacuum chamber. solution; (E) hose con-
from marine and terrestrial sites. Three case Note also that the vacuum gauge (E) should nected to running water
studies are presented in this chapter. The rst be placed in line after the gas trap assembly source; (F) sink with
is an example of consolidating friable bone to prevent damage from solvent vapors. drain.
from a land site. The second study illustrates
a successful process for preserving delicately
carved ivory from nautical excavations at
Tantura Lagoon in Israel. The last case study
illustrates the successful processes used to pre-
serve large sections of elephant tusk from ex-
cavations of a Dutch East Indiaman, Vergulde
Draeck, which sank o the coast of Western
Australia in .
Before discussing preservation strategies
for bone, however, I examine various equip-
ment congurations and how each may be
used for the treatment of archaeological bone
and ivory.
114

Fig. .. Equipment Case Study: Consolidating Friable Bone quickly rinsed in a gentle stream of tap wa-
conguration using high ter to remove additional dirt and dust from
eciency vacuum pump: Dry, desiccated bone is generally easy to con- its surfaces. After rinsing, it was left in a fume
(A) desiccator/vacuum jar solidate using Passivation Polymers. The hood for hours to air-dry. The bone and
with valve; (B) beaker bone sample brought to the research labora- its foil wrapper were then gently immersed
containing artifact and tory had been recovered from excavations of in a beaker of acetone for approximately
either acetone or polymer
the well-dened Clovis occupation strata at hour to ensure that dehydration was com-
solution; (C) in-line gas
the University of Texas Gault site during a plete. Because the bone fragments were too
traps embedded in dry ice;
(D) Dewar ask with dry eld school excavation conducted by Texas fragile to remove from the foil, sections of the
ice; (E) manometer; A&M University. After large clumps of soil bone were rst treated with a topical applica-
(F) vacuum pump. were removed with a soft brush, the bone tion of the polymer/cross-linker solution.
fragments were stored in aluminum foil. Fig- Because of its viscosity, CR-, was selected
ure . illustrates the bones fragmented sur- for consolidation purposes. A % addition of
faces. CR- by weight was thoroughly mixed with
Before treatment, the fragmented bone, CR- to create a suitable volume of polymer/
supported by the aluminum foil wrapper, was cross-linker solution to consolidate the bone.

Fig. .. Fragments of
bone wrapped in alumi-
num foil.
Q-tips were used to gently apply the solution. bone fragments felt dry. They were placed in
Because the bone was thoroughly dry, the a fume hood for approximately hours to
polymer solution was quickly absorbed. Re- dissipate the smell of catalyst before analysis 115
peated spot applications were made over the and labeling were completed.
entire top surface until the bone was wet. The
bone was then gently transferred and inverted
onto a second piece of aluminum foil, and Case Study: Ivory from Tantura-B
the process of applying the polymer solution Excavations in Israel
was repeated.
After polymer/cross-linker solution was Shelley Wachsman, of the Nautical Archae-
applied topically, the bone was more stable. ology Program at Texas A&M University, re-
The artifact was then carefully removed from covered a nely carved piece of ivory from a
the foil, immersed in the polymer/cross- shipwreck o the coast of Israel dated to the
linker solution, and placed into a glass desic- early ninth century C.E. The artifact mea-
cator/vacuum chamber, where it was treated sured approximately three inches wide, four
in a reduced atmosphere environment of inches long, and one-sixteenth-inch thick.
Torr (. Pa). For a short period of time, Inspection of the ivory indicated it had been
a rapid stream of bubbles emerged from the intricately carved and was in very good con-
bone. Care was taken to apply only sucient dition. Fortunately, the artifact had been
vacuum to displace air in the bone with the stored in seawater and transported in this
polymer solution. After approximately state to Texas A&M University. Prior to treat-
hours, only occasional bubbles were noted, ment, the artifact was desalinated in a series
and the bone was allowed to sit in the poly- of freshwater rinses. Initially, tap water was
mer solution at ambient pressure for an ad- used. Over time, however, baths of rainwater
ditional hours. The fragments were then and, nally, deionized water were used to
removed from the polymer solution and rinse salts from the artifact. Titration tests
placed on a screen over a plastic container were used to assess the rinse water. At ap-
where runo polymers were collected for re- proximately ppm, titration readings were
use (g. .). stable and the ivory was considered ready for
The bone fragments were allowed to drain treatment.
of polymer solution for several hours. Soft The artifact was placed into a beaker con-
cloths were used to blot the surfaces of the taining a : solution of alcohol and water
fragments to remove any remaining pooled
polymer solution. The fragments were then
placed into a large Ziploc bag. An aluminum
dish containing g of CT- catalyst was
placed alongside the bone fragments and the
bag was sealed, forming an airtight environ-
ment. Over the next hours, catalyst vapors
acted to polymerize the bone fragments.
CT- catalyst has a working life of approxi-
mately hours in a contained air environ-
ment. To ensure that the bone fragments were Fig. .. After vacuum-
completely cured, a fresh dish of catalyst was assisted impregnation, the
placed in the bag and the catalyzation phase bone fragments were
of treatment was repeated for an additional allowed to drain of excess
hours. After hours of catalyzation, the polymer solution.
 for initial dehydration. After days, the ivory gently wipe the surfaces of the artifact, care-
was moved to a second beaker containing fully removing any remaining polymer solu-
116 fresh isopropyl alcohol, where it was dehy- tion from carved crevices.
drated for an additional days. To complete Catalyzation was completed using a two-
dehydration, the ivory was then transferred stage process. First, CT- was topically ap-
to a beaker containing fresh acetone. It re- plied to the surfaces of the artifact. Soft cloths
mained in this solution for days. were used to wipe o the catalyst after allow-
Because the surfaces of the artifact ap- ing it to sit for two minutes. The surfaces of
peared to be very hard and smooth, there was the ivory felt smooth and had a uniform
concern that the use of a viscous polymer semigloss appearance. The artifact was then
might inhibit thorough penetration of the placed into a large Ziploc bag with an alumi-
polymer solution into the matrix of the arti- num dish containing g of CT-. The bag
fact. Accordingly, PR- Passivation Polymer was sealed, and the artifact was catalyzed for
was selected. To a suitable volume of PR-, hours. When removed from the bag, the
a % (by weight) addition of CR- cross- ivory was not glossy in appearance and did
linker was thoroughly mixed. The solution not feel rubbery to the touch. Most impor-
was then placed into a vacuum chamber and tant, no exfoliation or surface aking had
treated in a reduced atmosphere environ- occurred. The surfaces of the artifact were
ment for approximately ve minutes to de- lightly wiped with soft cloths, and no addi-
air the solution prior to treatment. The ivory tional treatment was required (g. .).
was quickly transferred to the container of
the polymer/cross-linker solution. The con-
tainer was placed into a desiccator/vacuum Case Study: Waterlogged Tusks from
chamber, and the artifact was treated in a Western Australia
reduced atmosphere environment for sev-
eral hours. Initially, only sucient vacuum Two sections of waterlogged and badly dete-
was applied so that a very slow stream of riorated tusk were shipped to the Archaeo-
bubbles was noted. The chamber was locked logical Preservation Research Laboratory
o at reduced pressure to allow ecient ac- from the Western Australia Maritime Mu-
etone-silicone oil solution exchange. After seum in Perth, Australia. These were sections
approximately hours, the vacuum pump taken from a cargo of long tusks that had
was turned on, further reducing the pres- been recovered from the shipwreck of a
sure in the chamber. After hours of treat- Dutch East Indiaman, Vergulde Draeck. After
ment no bubbles were observed emerging being recovered from the wreck, the tusks
from the surfaces of the ivory. The container had remained in storage in freshwater. The
was removed from the vacuum chamber, and samples sent to the lab (GDA and
the ivory was left in solution for an additional GTA) for treatment, however, had been
hours. wrapped in cotton gauze, which started to
After the artifact was carefully removed decay in transit. Apart from an odor prob-
from the polymer solution, it was placed di- lem, neither sample sustained physical dam-
rectly onto sheets of tissue paper. The goal age during shipment.
was to fully support the ivory while using the Both sections of tusk were unwrapped,
paper to blot polymers from its surfaces. immediately rinsed in running tap water for
Sheets of tissue paper were also placed on the hours, then treated for days in tap water
top surface of the artifact to remove the poly- rinse baths that were changed each day.
mer solution. After minutes, the ivory felt When examined, areas of the surface of
dry. Soft, lint-free cloths were then used to sample GTA were aking. Rinsing in
freshwater did not appear to exacerbate this
problem. This section of tusk measured .
cm in length and . cm in diameter at its 117
widest point. Its pretreatment wet weight was
. g. Artifact GTA ranged in color
from darker areas that registered Yr/
Munsell to lighter areas that registered R ./
Munsell in color. Near the base of the arti-
fact was a roughly triangular shaped section
that was bluish gray in color. This section of
the tusk was slightly spongy compared to the
rest of the artifact.
Prior to treatment, GTA was wrapped
in cotton string to ensure that no delami-
nation of the artifact occurred during treat-
ment (g. .). The ivory was then immersed
in industrial-grade acetone and placed into
a vacuum chamber, where it was treated in
a reduced atmosphere environment of
Torr (. Pa) for hours. After being linker (by weight) was prepared. The ivory Fig. .. Posttreatment
immersed in fresh acetone, the artifact was was then quickly transferred from the acetone view of the ivory.
treated for an additional hours in the bath into the polymer solution. Immersed in
same reduced atmosphere environment this solution, the ivory was placed into a
(g. .). vacuum chamber and treated at a reduced
After vacuum-assisted dehydration, the pressure of Torr (. Pa) for hours.
artifact was transferred to a fresh beaker of After acetone/polymer solution displace-
acetone, where it remained for an additional ment, the artifact was removed from the
hours at ambient pressure and room tem- polymer solution and allowed to drain of
perature. A solution of PR- ( centistoke) surface polymers for approximately hours.
mixed with a % addition of CR- cross- The string binding was then removed, and

Fig. . Tusk (artifact

GTA) during acetone/
Fig. .. Tusk (artifact GTA) bound in polymer solution dis-
cotton string, prior to catalyzation. placement.
 the artifact was immersed in a bath of CR-
to assist in washing additional polymer so-
118 lution from surface cracks and crevices. Af-
ter immersion cleaning for less than one
minute, the tusk was removed from the so-
lution and gently brushed with a soft tooth-
brush to help clean crevices on the surface
(g. .). The tusk was then gently wiped with
lint-free cloths, rebound with cotton string,
and placed into a large plastic container,
along with an aluminum dish containing Fig. .. End view of the tusks. The slightly
g of CT- catalyst. After hours of expo- blue gray discoloration of artifact GTA
(left) is visible on the right side of the artifact
sure to catalyst vapors at C, the tusk was
near its base. Artifact GDA (right) is
removed and inspected.
lighter in color. Note that the slight amount
Slight seepage of polymer solution was
of delamination visible on the base of both
noted on the base of the tusk and in the area tusks has not increased as the result of
of the softer, blue gray ivory. These areas treatment.
were rst wiped with a cloth dampened with
a few drops of CR- cross-linker, which
acted to remove surface-pooled polymer lyzed for an additional days, exposed to
solution. CT- was applied topically to the CT- catalyst vapors. Each day, an alumi-
base and discolored areas of the tusk, and num dish containing g of fresh CT- was
left on for four minutes; soft cloths were placed in the container to ensure maximum
then used to wipe the areas dry of remain- catalyzation.
ing catalyst. The artifact was placed back The second section of tusk, GDA, was
into a plastic container, where it was cata- conserved using the same procedures. Un-
like artifact GTA, however, this artifact
did not have any blue gray discoloration. Af-
ter treatment, both artifacts were photo-
graphed, weighed, and measured. In neither
case was shrinkage noted, and no cracks de-
veloped as the result of treatment. To some
degree, the polymer solution acted as an op-
tical bridge, eliminating the patch of surface
aking apparent in artifact GTA before
treatment (g. .).
Both sections of tusk have remained di-
mensionally stable after treatment. No seep-
age of polymer solution has been noted, and
both artifacts remain stable in coloration.
Preservation of these artifacts demanded a
quick and imaginative response in trying to
account for possible delamination of the
Fig. .. Surface-cleaning tusks during treatment. Because the tusks
the tusk (artifact were badly degraded, we felt it wise to tightly
GTA) with CR- and bind the artifacts with cotton string to pre-
a soft toothbrush. vent further delamination during treatment.
 CHAPTER 9
119

Expanding

the Conservation

Tool Kit

A
ll case studies and research discussed cavity of the concretion is. But devising ap-
to this point have been directed at the propriate conservation strategies using tradi-
preservation of organic substances by tional radiographs is dicult when dealing
displacing water and salts with polymer and with large, complex artifacts. More important,
resins that, once polymerized, ensure the it is extremely dicult to clean the cavities of
preservation of diagnostic attributes of or- intricate artifacts suciently to allow the ow
ganic artifacts. Silicone oil technologies are of resins used in the mold-making phase of
useful for other aspects of artifact preserva- conservation.
tion as well. Combining computerized to- Iron artifacts recovered from marine en-
mography (CT) data and computer-aided vironments fall into three basic categories.
design (CAD) modeling software, technicians The rst group is concreted artifacts in which
can create three-dimensional models of ar- no metal remains. The second group com-
tifacts using laser driven polymerization of prises solid metal artifacts deeply imbedded
each slice of CT data. The process permits in concretion due either to partial surface
nonintrusive investigation and recording of disintegration or to close proximity with
complex artifacts. other artifacts that have deteriorated. The last
CT and CAD technologies can be espe- group, consisting of concreted metal objects
cially useful in the preservation of iron ob- in which only some of the metal remains, are
jects from marine environments. When ex- the most challenging artifacts. In all three
posed to water, the natural tendency of iron cases, the conservation process is often hin-
is to revert to oxides. In a marine environ- dered by the inability of traditional radio-
ment, oxides, sediments, and salts combine graphs to supply sucient data to give the
to form a hard outer surface as deterioration conservator a complete view of the artifact
of the metal progresses. The buildup of con- concealed within the concretion. Delicate
cretion also acts to preserve the original form details of a three-dimensional artifact are also
of the artifact. In many cases, little to no iron often impossible to distinguish in two-
will remain while wood closely associated dimensional lm prints. During casting and
with the artifact may be preserved in perfect, removal of the outer layer of concretion with
waterlogged condition. pneumatic chisels, ne details of an artifact
Traditionally, conservators use radio- may be missed unless the conservator can
graphs to look inside the concreted mass. In distinguish surface features and act quickly
doing so, they can devise a conservation strat- enough to prevent destruction and loss of
egy for the artifact, based on the determina- internal features.
tion of what iron remains and how clean the The advent of stereolithographic (STL),
 rapid-prototype technologies represents a concreted artifacts that could be easily
new set of tools for better diagnosis and handled and x-rayed; and we found a com-
120 modeling of artifacts prior to commencing pany willing to work with us to create rapid
conservation. This technology aords many prototype reproductions of the artifact.
benets for the archaeological conservator. CT technology is a nondestructive means
Three-dimensional imaging of a complex of looking inside a concreted artifact. The
artifact allows the conservator to view the machinery, working with high energy X rays,
artifact from an innite variety of angles and produces two-dimensional slices of the cross
perspectives. Through the use of three- section of an artifact. Each slice is between
dimensional CAD programs, a radiographic . inches and . inches, depending on
view of an artifact can be manipulated and the resolution needed. Thinner slices produce
rotated until the conservator has the infor- higher resolution images. Naturally, produc-
mation necessary to commence conservation. ing a complete image requires many more
To attain similar views of the interior of a slices when done in high resolution. Scan-
concretion with conventional methods is time- ning time is therefore longer and all related
consuming and often less accurate, based on costs are higher. Creating a CT image requires
the limitations of the diagnostic machinery. compiling a series of slices, which when com-
CT-STL analysis of an artifact allows the con- plete allow the operator and conservator to
servator to create an exact replica of the arti- evaluate the integrity of the artifact inside the
fact in resin prior to treatment of the origi- concretion. Each slice is a discrete set of data
nal artifact. A three-dimensional model is an about one particular section of the artifact.
invaluable aid to anyone who must clean en- When the CT scanning process is complete,
crustation from delicate features of an arti- data that have been stored on disk during the
fact. Additionally, the ability to create a resin scanning must then be processed using spe-
cast replica of the artifact being treated is an cial software that renders the slices, creating
insurance policy. If something goes wrong a three-dimensional image. Programs such
during treatment, the presence of the replica as MIMICS (Materialises Interactive Medi-
may serve to retrieve data that might other- cal Image Control System) are widely used
wise be lost. for the purpose of aligning and correcting
individual scanned images. Once scanned
images are prepared, additional software is
Computerized Tomography and the required to transform the data into an inte-
Stereolithographic Process grated image that is usable on rapid proto-
typing instruments.
Stereolithography is literally three-dimensional The stereolithographic machinery is
printing. I have not attempted to explain the unique. It consists of a large vat of photo-
stereolithographic process and associated curable resin, mirrors, a UV laser assembly,
equipment, because for archaeological pur- and an elevator assembly that is precisely
poses, understanding potential applications driven through the vat (g. .). As image data
for the process is far more important than are processed, the laser assembly scans the
understanding the mechanics of the technol- layer of the part being modeled. As one layer
ogy. Essentially, three main components were is completed, the elevator assembly drops
brought together to do reproductive model- slightly to allow the laser to draw the next
ing of encrusted artifacts: a local hospital section of artifact information onto the rst
donated time and personnel to obtain com- layer (g. .). The process continues until
puterized tomographic (CT) images of what all the layers have been re-created. The cast
is inside an encrustation; we identied heavily is put into an oven for curing of the resins
before touch-up work can be completed.
Depending on the size of the artifact, the
rapid prototyping system can take several 121
hours.


Case Study: Scanning an Encrusted
ArtifactCT Scanning Used as a Diag-
nostic Tool

A small artifact, thought to be a heavily con-

creted sword hilt, was chosen for demon-
stration purposes. The artifact had been
recovered from a nautical excavation site and
had been kept wet prior to treatment. This
was an important step to take for two rea-
sons. In a wet state, contrasted densities of
the various materials in the concretion are
accentuated, resulting in a better artifact
image. Second, as a concretion dries out, it
becomes harder. This makes the task of re-
moving it more dicult, especially when it
is in close association to small and delicate Three-dimensional images are innitely Fig. .. A schematic
artifacts. scalable. An example will serve to illustrate drawing illustrating the
Prior to CT imaging, the artifact was vi- the importance of scalable models for lling basic stereolithographic
sually inspected for identiable surface fea- in the gaps in special collections. The nature process: (A) laser gun;
tures (g. .). The artifact was then mounted of the fossil record is such that some speci- (B) mirror; (C) elevator
into the CT unit, using the headrest assem- mens are rare. Other specimens may be more assembly; (D) reected
laser beam; (E) vat;
bly for placement (g. .). Throughout the plentiful, but incomplete. Stereolithographic
(F) replicated artifact;
scanning process, individual scans were vis- rapid prototyping can be used to ll in the
(G) photocurable resin.
ible on the computer monitor. Naturally, each gaps in museum collections. If, for example,
scanned image provided only a limited view the specimen in one museum is missing a
of the artifact (g. .). After compiling in single vertebra or series of vertebrae, appro-
graphic software, however, an excellent three- priately fashioned replica can be made using
dimensional image of a sword handle was samples from another museum or a more
created (g. .). intact specimen. Even if the fossil animals

New ToolsNew Directions in Research

The realm of digital media oers exciting

application opportunities for archaeological
investigations. At a time when museums are Fig. .. View of
adding interactive displays and mobile ex- the process of
hibits, rapid prototype technology has stereolithography as a
innite uses in the museum and research section of resin is solidi-
environment. ed using a laser.
122

Fig. .. A small concre-

tion is discussed prior to
CT scanning. Note that
the encrustation is
supported by rubber
bands and tape.

Fig. .. The concretion is

placed in the headrest of
the CT machinery.

dier in size, appropriately sized vertebra can Stereolithographic prototyping can also be
be manufactured to make the fossil remains used to ll in gaps of an artifact assemblage
complete for display purposes. Because the for display purposes. If for instance, a par-
fabricated vertebrae are made of resin, they ticular Clovis point is missing in an assem-
can be dissimilarly colored to illustrate that blage, a replica artifact can be made from a
they are replacement parts or, when desired, borrowed artifact. Depending on the resolu-
colored to look like real bone. tion of the original CT scans, reproductions
 123

Fig. .. As individual
scans are completed, they
are visible on the com-
puter screen.

can be extremely accurate in size and sur-

face detail. Rapid prototype models can also
be invaluable in the classroom. Instead of
working with rare and often fragile speci-
mens, instructors can utilize resilient resin
copies that can withstand years of student
handling.
Images of artifacts compiled from CT data
can be rendered into three-dimensional im-
ages in CAD software. These images, which
are created as part of the stereolithographic
process, are quite useful for display purposes.
These three-dimensional models can also be
used as the basic gures to create morphed
animation sequences for displays. Numerous
inexpensive software programs can be used
to turn start and end point images into ani-
mated sequences.

Fig. .. After scanning,

the individual slices of the
image are compiled to
create a three-dimensional
image of the artifact
inside the concretion.
124

 Notes
125

Chapter . E. E. Astrup, A Medieval Log House in

OsloConservation of Waterlogged
. CR- is also known as MTMS Softwoods with Polymer Glycol (PEG),
(methyltrimethoxysilane); its chemical Proceedings of the Fifth ICOM Group on Wet
formula is (CH3O)3 SiCH3. For consistency I Organic Archaeological Materials Conference,
have used the term CR- throughout this .
book. . Malcolm Bilz, Lesley Dean, David W.
. CT- is also known as DBTDA (dibutyltin Grattan, J. Cliord McCawley, and Lesley
diacetate); its chemical formulation is McMillen, A Study of the Thermal
C12H24O4Sn. For consistency I have used the Breakdown of Polyethylene Glycol,
term CT- throughout this book. Proceedings of the Fifth ICOM Group on Wet
Organic Archaeological Materials Conference,
Chapter .
. Michael A. Wilson, Ian M. Godfrey, John V.
. Charles D. Moore, Reassembly of a Hanna, Robinson A. Quezada, and Kim S.
Sixteenth-Century Basque Chalupa, Finnie, The Degradation of Wood in Old
Material History Review (fall ). Indian Ocean Shipwrecks, Geochem :
. Martha Carrier, personal communication, (): .
June , , regarding preservation of the . Ibid.
timbers of the Red Bay Chalupa. . C. V. Horie, Materials for Conservation:
. R. H. Atalla and D. L. Vanderhart, Native Organic Consolidants, Adhesives, and
Cellulose: A Composite of Two Distinct Coatings (Oxford: Butterworths, ), .
Crystalline Forms, Science (): ; . David Gilroy and Ian Godfrey, eds., A
Joseph B. Lambert, Catherine E. Shawl, and Practical Guide to the Conservation and Care
Jamie A. Sterns, Nuclear Magnetic of Collections (Perth, Australia: Western
Resonance in Archaeology, Chemical Society Australia Maritime Museum, ), .
Review (): ; Michael A. Wilson,
Ian M. Godfrey, John V. Hanna, Robinson A. Chapter 4
Quezada, and Kim S. Finnie, The
Degradation of Wood in Old Indian Ocean . J. M. Cronyn, The Elements of Archaeological
Shipwrecks, Geochem : (): . Conservation (New York: Routledge, ), .
. Michael A. Wilson, Ian M. Godfrey, John V. . Ibid.
Hanna, Robinson A. Quezada, and Kim S. . J. M. Cronyn, The Elements of Archaeological
Finnie, The Degradation of Wood in Old Conservation (New York: Routledge, ),
Indian Ocean Shipwrecks, Geochem : .
(): .
. Allen Brownstein, The Chemistry of Chapter 5
Polyethylene Glycol, in Proceedings of the
IOCM Waterlogged Working Group . J. P. Maish, Silicone Rubber Staining of
Conference, ed. D. Grattan (Ottawa: Terracotta Surfaces, Studies in Conservation
Canadian Conservation Institute, ). (): .
 . Stefano Pulga, A Note on the Use of . H. J. Plenderleith and A. E. A. Werner, The
Silicone Rubber Facings in the Reassembly Conservation of Antiquities and Works of
126 of Archaeological Painted Plasters, Studies Art: Treatment, Repair, and Restoration, d
in Conservation (): . ed. (Oxford: Oxford University Press,
. Elizabeth S. Goins, The Acid/Base ).

Characteristics of Sandstone, Limestone, . Roy Newton and Sandra Davison, Conser-

and Marble, and Its Eects upon the vation of Glass (London and Boston:
Polymerization of Tetraethoxysilane, paper Butterworths, ), .
from the Institute of Archaeology, . Ibid.
(London, ): . . G. Torrraca, Synthetic Materials Used in the
. Mason Miller, personal communication, Conservation of Cultural Property (Rome:
. Miller, a graduate student at Texas International Centre for the Study of the
A&M, directed studies in the universitys Preservation and the Restoration of
Archeological Preservation Research Cultural Property, ), .
Laboratory. . Roy Newton and Sandra Davison, Conser-
vation of Glass (London and Boston:
Chapter Butterworths, ), ; J. M.
Bettembourg, Protection des Verresde
. Inger Bojesen Koefoed, Ion Meyer, Poul Vitraux Contre les Agents Atmosphriques,
Jensen, Kristiane Straetkvern, Report 4.1: Etudes des Filmes de Rsines
Conservation of Wet Archaeological Rope Synthettiques, Verres et Rfract ():
(Brede, Denmark: National Museum of ; R. F. Erret, M. Flynn, and R. Brill,
Denmark, Section for Organic Materials, The Use of Silanes on Glass, Adhesives, and
). Consolidants, Proceedings of the Tenth IIC
. Ibid. Congress (Paris, ), .
. Corcoran Laboratories, Springknoll . C. V. Horie, Materials for Conservation:
Lane, Bay City, Mich. , --. Organic Consolidants, Adhesives, and
. Vera de la Cruz Baltizar, Plastination as a Coatings (Oxford: Butterworths, ), .
Consolidation Technique for Archaeological . H. J. Plenderleith and A. E. A. Werner. The
Bone, Waterlogged Leather, and Waterlogged Conservation of Antiquities and Works of Art:
Wood. Diploma thesis, Department of Art, Treatment, Repair, and Restoration, d ed.
Queens University, Kingston, Ontario, , (Oxford: Oxford University Press, .)
. . Jerome Melvin Klosowski, C. Wayne Smith,
. Ibid. and Donny Leon Hamilton, Conservation of
Organic and Inorganic Materials, U.S. Patent
Chapter and Trademark Oce, Washington, D.C.,
August , , patent ,,.
. Robert H. Brill, Ancient Glass, Scientific . G. Gueskens, M. Borsu, and C. David,
American (November ). Photosynthesis and Radiolysis of
. J. M. Cronyn, The Elements of Archaeological Polyvinylacetate. Eects of Temperature
Conservation (New York: Routledge, ). on the Photolysis, European Polymer
. Robert H. Brill, Ancient Glass, Scientific Journal (Oxford: Pergamon Press, ), .
American (November ).
. Colin Pearson, ed., Conservation of Marine Chapter
Archaeological Objects (Boston:
Butterworths, ). . H. J. Plenderleith and A. E. A. Werner, The
. Roy Newton and Sandra Davison, Conser- Conservation of Antiquities and Works of Art:
vation of Glass (London and Boston: Treatment, Repair, and Restoration, d ed.
Butterworths, ), . (Oxford: Oxford University Press, ).
 Page links created automatically - disregard ones not formed from page numbers

Index

Italic t indicates a table; f indicates a figure. CR-12, b o n e / i v o r y treatments, 11415

CR-20, described, lot, 11
4-7041: described, 11t, 12; glass treatments, 104, CR-20, application techniques: bone/ivory, 114
107 15,116,11718; glass, 97, 100,104; leather, 70
71; rope, 8485, 88; sabots, 4345; tongue
absorbent materials, selection guidelines, 9 depressor experiments, 3140, 4243;
acetone: safety guidelines, 9, 21; as storage w o o d e n buttons, 2628; w o o d from dry
m e d i u m , 23. See also ethanol/acetone sites, 29,30; w o o d i n general, 24,25
displacement/exchange; water/ethanol CR-20/Q9-1315, c o m p a r i s o n experiment:
displacement/exchange methodology, 4550; results, 5159
acetone/polymer ( A P ) displacement. See CR-22: described, lot, 11; glass bottle treat-
Passivation Polymerization ments, 109
alcohol as storage m e d i u m , 23, 63. See also C r o n y n , J. M . , 63, 64, 69
ethanol/acetone displacement/exchange; cross-linkers, overview, ixx, 911. See also CR-
water/ethanol displacement/exchange 20 entries; CR-22; Q9-1315, application
alkoxysilane polymers. See CR-20 entries procedures
AP. See Passivation Polymerization Crumlin-Pedersen, Ole, 30
Astrup, E . E., 41 CT-30, described, 10t,12
CT-32, described, 10t,1112
Baltizar, Vera de la C r u z , 89,90 CT-32, application techniques: bone/ivory, 115,
basket w i t h i r o n shot, 7480 116,118; canvas, 91; glass, 9798,104105;
beakers, selection guidelines, 8 leather, 68,7172; rope, 85, 88; sabots, 4445;
Bilz, M a l c o l m , 45 tongue depressor experiments, 32,4243;
biocides for leather, 64 wicker basket, 7879, 80; w o o d e n buttons,
bone/ivory artifacts, 11218 27; w o o d from d r y sites, 2930; w o o d in
general, 2526
C A D (computer-aided design), 11923 CT-34, described, 10t,12
canvas artifacts, 9092 C T (computerized tomography), 11923
catalysts, overview, x x i , 1012,2526. See also cups, selection guidelines, 89
CT-32
chemicals, overview, 912. See also specific D B T D A . See CT-32 entries
chemicals, e.g. CR-22; PR-12 The Degradation of Wood in Old Indian Ocean
C i v i l W a r sabots, 4045 Shipwrecks ( W i l s o n ) , 46
cold traps, selection guidelines/setup, 8,9f dehydration processes: described, 13,17t; dowel
composite artifacts: basket w i t h shot, 7480; experiment, 1316; spectrographic analysis
conservation challenges, 74; CR-20 topical study, 1720. See also water/acetone
benefits, 38; glass bottle w i t h cork, 10811; displacement/exchange; water/ethanol
sabots, 4042, 4345 displacement/exchange
computer-aided design ( C A D ) , 11923 desalination treatments: bone/ivory, 113,115,
computerized tomography ( C T ) , 11923 11617; o n i o n bottles, 104; rope, 83; treenails,
The Conservation of Antiquities and Works of 48; w o o d e n buttons, 26; w o o d i n general,
Art (Plenderleith & Werner), 113 2223
cork treatments, 10811 devitrification, 96
 Page links created automatically - disregard ones not formed from page numbers

Dewar flasks, selection/safety guidelines, 8,9f, laboratory equipment, selection/setup

21 guidelines, 712
disposable containers, selection guidelines, 89 leather artifacts: conservation challenges, 60
128 dowel experiment, 1316 62; preservation options, 6266, 6972;
D o w i c i d e 1 for leather, 64 shoe case study, 6668; storage/display
INDEX dry-site artifacts, treatments: bone, 11315; 7273
leather, 60, 6972; w o o d , 2831
manometers, selection guidelines, 8
electrolytic reduction, 79, 80 mass spectral analysis experiment, 1720
electronic scales, 8 material safety data sheets ( M S D S ) , 7,10
electron i n o n i z a t i o n mass spectral analysis, 17 mesh bags, selection guidelines, 9
20 metal screens, selection guidelines, 9
epoxy resins, 9899 M i r r o r Lake sabots, 4045
equipment recommendations, 712 m o u l d i n g materials, 75, 77, 79
ethanol/acetone displacement/exchange: glass, M S D S (material safety data sheets), 7,10
96; ivory, 11516; leather, 63, 6970; treenails, M T M S . See C R - 2 0 entries
48, 50; w o o d e n buttons, 26
ethanol/polymer displacement/exchange: nuclear magnetic resonance ( N M R ) analysis,
inefficiencies, 96; spectrographic analysis tongue depressor experiment, 3132,3440,
study, 1720 46
explosion-proof freezers, 8
eye-washing systems, 7 o n i o n bottles, 100108
ortho-phenylphenol for leather, 64
4-7041: described, 11t, 12; glass treatments, 104,
107 PA F l u i d : described, 11t, 12; o n i o n bottle
Frankfurter method, 82,83, 8788 treatments, 103104,106
freeze-drying techniques: leather, 6465; rope, Passivation Polymerization: overview, i x x i , 9
82 12; chemical theory of, y6; conservation
freezers, selection guidelines, 8 advantages, 3, 6,21, 80; dehydration
fume hoods, selection guidelines, 8,10 experiments, 1320. See also specific artifacts,
e.g. leather artifacts; w o o d , waterlogged
gas traps, selection guidelines, 8,9f P E G treatments: leather i n general, 60, 6366,
Gault site artifacts, 11415 73; rope, 8283; shoe, 6668. See also w o o d ,
glass artifacts: bottle w i t h cork, 10811; PEG-treated
conservation challenges, 9396; o n i o n Plenderleith, H . J., 95,100,113
bottles, 100108; preservation options, 95 polyethylene glycol. See P E G treatments
99 polyvinyl acetate emulsions: glass treatments,
glass plates for leather treatments, 71f, 72 100,102103; PGE-treated w o o d , 31
Grattan, D a v i d , 30 Port Royal artifacts, treatments: canvas, 9092;
o n i o n bottles, 100108; w o o d plank, 4749,
Hoffman, Per, 30 50,5455, 56t, 57, 59
HSIII R T V c o m p o u n d , 75, 77, 79 PR-10: described, 11t, 12, 24; dowel experiment,
14; spectrographic analysis study, 18
iron artifacts: basket w i t h shot, 7480; C T / C A D PR-10, application procedures: glass, 103104,
technologies, 11921; sabots, 4042, 4345 106; ivory, 116,117; leather, 70; rope, 84;
ivory/bone artifacts, 11218 wicker basket, 7576, 77, 78; w o o d e n
buttons, 26; w o o d from d r y sites, 29
kerosene treatments, 82,85, 8788 PR-12: canvas treatments, 91; described, 11t, 12,
K P 8 0 : described, 11t, 12; glass treatments, 104, 24; leather treatments, 67, 70; rope treat-
106107 ments, 84
PR-14: described, 11t, 24; glass treatments, 97,
La Belle, artifact treatments: basket w i t h iron 108109, 111; leather treatments, 70
shot, 7480; leather shoe, 6668; rope, 8182, Pulga, Stefano, 75
8390; timbers, 21; treenails, 4546,4849; P V A / P V A C emulsions: glass treatments, 100,
w o o d e n buttons, 2628 102103; PGE-treated w o o d , 31
 Page links created automatically - disregard ones not formed from page numbers

Q9-1315, application procedures: treenails, 48, tongue depressor experiment, 4950;

49,50,5659; wicker basket, 7576, 77, 78, 79; treenails/wood plank, 50; w o o d from d r y
w o o d plank, 50, 55,56t, 59 sites, 29; w o o d i n general, 21, 23, 2425
129
ventilation guidelines, 7, 8,10
reproductions o f artifacts, 12123 Vergulde Draeck cargo, 116
rope artifacts: conservation challenges, 8182; von Hagen procedure, i x x INDEX
preservation options, 8283; treatment
procedures, 8385; treatment results, 8590 Wachsman, Shelley, 115
water/acetone displacement/exchange:
sabots w i t h shot, 4045 overview, ix, 13; bone/ivory, 114,117; canvas,
safety guidelines, 712, 6465, 71 9091; dowel experiment, 1316; glass, 104,
salt water artifacts. See desalination treatments 107,108; leather, 67, 6970; rope, 8384;
scales, electronic, 8 spectrographic analysis study, 1720; tongue
screens, metal, 9 depressor experiments, 49; wicker basket
shoe, leather, 6668 w i t h shot, 75, 77; w o o d i n general, 21, 23;
silicone oils, overview, 11t, 12. See also 4-7041; w o o d plank, 50
KP80; P A F l u i d ; PR-10; PR-12; PR-14 water/ethanol displacement/exchange: glass, 96;
stereolithographic (STL) technology, 11923 ivory, 11516; leather, 63, 6970; spectro-
stir rods, selection guidelines, 9 graphic analysis study, 1720; treenails, 48,
superglues, 99 50; w o o d e n buttons, 26; w o o d i n general,
sword hilt, 121,123f 21,23
Tantura-B artifacts, 11516 weathering cycles, glass, 105106
tongs/tweezers, selection guidelines, 9 Werner, E. A . , 95,100,113
tongue depressor experiments: C R - 2 0 hydro- wicker basket w i t h i r o n shot, 7480
lyzing process, 4547, 4854,57,59; W i l s o n , M i c h a e l , 46
reprocessing o f P E G treatments, 3140, w o o d , dry-site, 2831
4243 w o o d , PEG-treated: sabots, 4042, 4345;
treenails, 4546,4849, 50, 5559 tongue depressor experiments, 3140,4243;
tusks, 11618 treatment challenges, 24,3031
tweezers/tongs, selection guidelines, 9 w o o d , waterlogged: overview o f treatment
Tyzor T P T Titanate. See CT-34, described procedures, 2326,3031; buttons, 2628;
conservation challenges, 2123; dehydration
v a c u u m procedures: bone/ivory, 113,114f, 115, experiments, 1320; plank, 4749, 50, 5455,
116,117; canvas, 91; dowel experiment, 14; 56t, 57, 59; sabots, 4042, 4345; tongue
equipment recommendations, 8, 9f; glass, depressor experiments, 3140,4243, 4547,
9697,103,104,108; leather, 67, 71; rope, 83 4854, 57, 59; treenails, 4546,4849,50,
84,85; spectrographic analysis study, 1820; 5559