Bio N MWCNT

Bioapplications of Nitrogen-doped Carbon
Nanotubes
Hilary Jane Burch
Bionanotechnology IRC, Department of Physics

University of Oxford
Thesis submitted for the degree of

Doctor of Philosophy
Linacre College
Trinity Term 2006
Abstract
Applications such as biosensors are important for medical advances in the detection of molecules
such as antibodies, and carbon nanotubes (CNTs) hold the promise to improve detection to the single
(or very few) molecular level. However, un-doped CNT biosensor devices can have disadvantages
of low conductivity and high contact resistance. As a result of this work, a novel biosensor based on
nitrogen-doped multi-walled carbon nanotubes (CNx MWNTs) has been developed, using a design
which minimizes these problems.
CNTs are long, thin cylinders of carbon that have been the subject of much research since
their discovery in 1991. Their small diameter, high aspect ratio and strong carbon-carbon bonds,
give CNTs unique physical and electronic properties, all of which may be altered by nitrogendoping This thesis reports measurement of the electrical and thermal conductance of CNx MWNTs
using Atomic Force Microscopy (AFM). The CNx MWNTs were found to exhibit high conductance
consistent with theoretical predictions. Thermal conductance measurements were taken in vacuum
and the CNx MWNTs were found to have reduced thermal conductance compared to un-doped
MWNTs. The effect of acid oxidation on these properties was also investigated.
The potential of CNx MWNTs for bioapplications was investigated by functionalisation of the
tubes with the metalloproteins ferritin, ferredoxin, azurin and cytochrome c. The effect on the
biomolecules was monitored using spectroscopy and antibody binding assays, and proteins were
found to retain their biologically relevant conformation. These proteins were then sensed using a
biosensing device based on CNx MWNTs.
Acknowledgements
I would like to thank the following people:
Prof John Ryan for introducing me to a very interesting and ever expanding area of science,
and for giving me the opportunity to make a contribution to it.
Dr Sonia Antoranz Contera for her patient supervision and AFM training.
Dr Maurits de Planque for watching over my various pieces of biochemistry.
Dr Julia Davies for supervising my research at NPL during January, July and August 2006.
Dr Lizzie Brown for the use of her AFM conductance set-up at NPL and for her help with the
measurements. Also Dr Ling Hao for helpful discussions.
Dr Kislon Votchovsky for building the Keithley measurement system without which none of
my biosensor measurements would have been possible.
Dr David Cox for SEM pictures taken at the University of Surrey.

Dr Nicole Grobert for providing the raw nanotube material.
My parents for their support.
EPSRC and the National Physical Laboratory (NPL) for funding my research.
For my twin sister Gillian
Contents
1
Introduction
1.1
Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.1
Electrical Transport in Nitrogen-doped CNTs . . . . . . . . . . . . . . . .
1.1.2
Potential of Nitrogen-doped CNTs for Biological Applications . . . . . . .
1.1.3
Thermal Properties of Nitrogen-doped CNTs . . . . . . . . . . . . . . . .
1.1.4
Effect of Acid Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.5
Biosensor Device . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Publications and Presentations . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.1
Posters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.2
Oral Presentations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2.3
Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2
1.3
2
Carbon Nanotubes
2.1
Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2
Nitrogen-doping of CNTs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.1
Doping Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.2
Defects caused by Nitrogen-doping . . . . . . . . . . . . . . . . . . . . .
10
2.2.3
Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11
Synthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11
2.3.1
Synthesis of CNx MWNTs . . . . . . . . . . . . . . . . . . . . . . . . . .
12
Growth Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13
2.4.1
Growth of Un-doped CNTs . . . . . . . . . . . . . . . . . . . . . . . . .
13
2.4.2
Growth Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14
2.4.3
Growth of CNx MWNTs . . . . . . . . . . . . . . . . . . . . . . . . . . .
14
Electronic Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15
2.5.1
SWNTs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15
2.5.2
Chirality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16
2.3
2.4
2.5
2.6
Thermal Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3 Experimental Techniques
3.1
18
25
Atomic Force Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
25
3.1.1
Tip-sample interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . .
25
3.1.2
The Microscope Tip . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
29
3.1.3
Modes of Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
30
3.1.4
Optical Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
30
3.1.5
Feedback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
30
UV-vis Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
32
3.2.1
UV-vis of CNx MWNTs . . . . . . . . . . . . . . . . . . . . . . . . . . .
33
3.3
Circular Dichroism Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . .
34
3.4
Lyophilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
35
3.2
4 Electrical Conductance of CNx MWNTs

4.1
4.2
4.3
4.4
Electrical Transport in un-doped CNTs . . . . . . . . . . . . . . . . . . . . . . . .
37
4.1.1
Quantized Conductance . . . . . . . . . . . . . . . . . . . . . . . . . . .
37
4.1.2
SWNTs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
38
4.1.3
MWNTs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
38
Electronic Transport in CNx MWNTs . . . . . . . . . . . . . . . . . . . . . . . .
40
4.2.1
Density of States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
40
4.2.2
Previous Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
41
Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
42
4.3.1
AFM Set-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
42
Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
44
4.4.1
Annealing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
44
4.4.2
Conductance Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
44
4.4.3
G-V and I-V curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
46
4.4.4
Electrical Breakdown . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
47
5 Biofunctionalisation of CNx MWNTs

5.1
5.2
5.3
37
52
Previous Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
52
5.1.1
Biofunctionalisation of un-doped CNTs with Proteins . . . . . . . . . . . .
52
5.1.2
Functionalisation of Boron Nitride NTs with Proteins . . . . . . . . . . . .
53
5.1.3
Functionalisation of CNx MWNTs with Proteins . . . . . . . . . . . . . .
53
Effect on protein activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
53
5.2.1
Antibody Binding Assays . . . . . . . . . . . . . . . . . . . . . . . . . .
53
Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
54
5.4
5.5
5.3.1
Functionalisation with metalloproteins . . . . . . . . . . . . . . . . . . . .
55
5.3.2
Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
55
5.3.3
Apo proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
56
Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
57
5.4.1
Ferredoxin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
57
5.4.2
Ferritin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
59
5.4.3
Azurin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
61
5.4.4
Cytochrome c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
65
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
69
5.5.1
Protein Binding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
69
5.5.2
Changes in the conformation of cyt c . . . . . . . . . . . . . . . . . . . .
71
5.5.3
Mechanism of apoazurin preparation . . . . . . . . . . . . . . . . . . . . .
71
5.5.4
Future Bioapplications . . . . . . . . . . . . . . . . . . . . . . . . . . . .
72
Thermal conductance of CNx MWNTs
78
6.1
Previous Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
78
6.2
Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
78
6.3
Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
79
6.4
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
81
The Effect of Acid Treatment on CNx MWNTs
83
7.1
Previous Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
83
7.1.1
Solubility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
83
Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
84
7.2.1
Acid Oxdiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
84
7.2.2
Lyophilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
85
7.2.3
AFM Imaging and Dimension Data . . . . . . . . . . . . . . . . . . . . .
85
7.2.4
Acid-base titrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
85
7.2.5
Measurement of Electronic Transport Properties . . . . . . . . . . . . . .
86
7.2.6
Measurement of Thermal Transport Properties . . . . . . . . . . . . . . .
87
Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
88
7.3.1
Length Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
88
7.3.2
Diameter Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
88
7.3.3
Acid-base Titrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
88
7.3.4
Electrical Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
90
7.3.5
Thermal Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
92
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
94
7.4.1
94
7.2
7.3
7.4
Solubility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7.4.2
Oxidation Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . .
94
7.4.3
Electrical Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
96
8 Development of a CNx MWNT Biosensing Device

8.1
8.2
8.3
8.4
Previous Nanotube Sensing Work . . . . . . . . . . . . . . . . . . . . . . . . . . 101

8.1.1
Small Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
8.1.2
Sensing of Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
8.1.3
Sensing with CNx MWNTs . . . . . . . . . . . . . . . . . . . . . . . . . 102
Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
8.2.1
Electroless Plating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
8.2.2
Electrical Measurement Set-up . . . . . . . . . . . . . . . . . . . . . . . . 104
Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
8.3.1
Sensing of Ferritin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
8.3.2
Sensing of cyt c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
8.3.3
Sensing of azurin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
8.4.1
Sensing Mechansim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
9 Conclusions and Further Work

9.1
9.2
101
117
Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
9.1.1
Electrical Conductance of CNx MWNTs . . . . . . . . . . . . . . . . . . . 117
9.1.2
Biofunctionalisation of CNx MWNTs . . . . . . . . . . . . . . . . . . . . 117
9.1.3
Thermal Conductance of CNx MWNTs . . . . . . . . . . . . . . . . . . . 118
9.1.4
The Effect of Acid Treatment on CNx MWNTs . . . . . . . . . . . . . . . 118
9.1.5
Development of a CNx MWNT Biosensing Device . . . . . . . . . . . . . 119
Further Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

9.2.1
Control Experiments with un-doped MWNTs . . . . . . . . . . . . . . . . 119
9.2.2
Further Thermal Measurements . . . . . . . . . . . . . . . . . . . . . . . 119
9.2.3
Atomic Resolution of CNx MWNTs . . . . . . . . . . . . . . . . . . . . . 120
9.2.4
Improved Biosensor Device Fabrication . . . . . . . . . . . . . . . . . . . 120
9.2.5
Further Biomolecule Functionalisation
9.2.6
Inter-digitated Electrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
10 Appendix
124
. . . . . . . . . . . . . . . . . . . 121
10.1 Acid-base Titration Calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124

10.2 CNx MWNT Length and Diameter Data . . . . . . . . . . . . . . . . . . . . . . . 126
10.3 Hybridisation in Carbon Materials . . . . . . . . . . . . . . . . . . . . . . . . . . 131
10.3.1 Atomic Orbitals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
10.3.2 The Hybrid Orbital Model . . . . . . . . . . . . . . . . . . . . . . . . . . 132

10.4 Colloid Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
10.4.1 DLVO Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
10.5 Amino Acids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
10.5.1 Amidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
Chapter 1
Introduction
1.1 Motivation
Carbon nanotubes (CNTs) are unusual objects with interesting electrical and physical properties,
which make them valuable components for devices such as field-effect transistors, 1 gas sensors, 2
and interconnects 3 in nano-circuits.
1.1.1 Electrical Transport in Nitrogen-doped CNTs

As-grown CNTs can be either semiconducting or metallic. 4 Although efforts are underway to find
a method of separation according to electronic properties, 5 atomic doping (similar to silicon semiconductor technology) provides an alternative way to generate nanotubes with consistent electrical
properties 6 through n-doping. Whilst extensive measurements of the electrical transport properties
of un-doped CNTs have been made, 714 the electrical transport properties of nitrogen-doped CNTs
have been subject to very little investigation. 1517 Furthermore, the results of these investigations
suggest that nitrogen-doped CNTs exhibit very poor electrical transport, despite theoretical predictions to the contrary. 18 It is this apparent contradiction that motivated the first aim of this work,
which was to study the electrical transport properties of individual nitrogen-doped CNTs. Electrical
transport measurements were carried out at the National Physical Laboratory.
1.1.2 Potential of Nitrogen-doped CNTs for Biological Applications

The integration of CNTs with biology has also produced many advances. The successful functionalisation of CNTs with biomolecules has attracted significant interest for applications such as
biosensors 19,20 and drug delivery. 21 Nitrogen-doped CNTs have several advantages over un-doped
n
here refers to electron doping and not nitrogen doping. Throughout this thesis doping with nitrogen will be written
as N or nitrogen.
Introduction
CNTs for bioapplications. In addition to their interesting electrical transport properties, the structure of nitrogen-doped CNTs is rich in hydrophilic (N) defects. Similar nitrogen groups anchored
to gold substrates, have been shown to facilitate electron transfer to and from metalloproteins. 22,23
Nitrogen-doped CNTs have also recently been found to have reduced toxicity compared to un-doped
CNTs. 24 Despite these advantages, there is only one report of biofunctionalisation of nitrogendoped CNTs in the literature. 25 In the second part of this project the suitability of nitrogen-doped
CNTs for bioapplications was investigated, by functionalisation with the soluble metalloproteins
azurin, cytochrome c, ferritin and ferredoxin.
1.1.3 Thermal Properties of Nitrogen-doped CNTs

The thermal properties of CNTs play a critical role in controlling the performance and stability of
devices. 26,27 The thermal conductivity () of un-doped CNTs has been measured 2729 but the effect
of nitrogen-doping on has not been investigated. Thermal transport measurements were carried
out at the National Physical Laboratory using an AFM method. 30
1.1.4 Effect of Acid Treatment

Acid oxidation of CNTs improves their solubility in water, and allows further chemical derivitisation. Whilst the effect and mechanism of acid oxidation of un-doped CNTs is well understood, the
effect of atomic doping is unknown. In this work, nitrogen-doped CNTs were treated with concentrated acids and the effect of the oxidation on length, diameter, percentage of surface functional
groups and electrical and thermal transport properties was investigated.
1.1.5 Biosensor Device

Un-doped CNTs have been incorporated into sensors for biological molecules, 19,3136 which exhibit
a characteristic response of a decrease in electrical conductance upon protein binding. The electrical
transport properties of nitrogen-doped CNTs were found to be ideal for device fabrication, and the
final stage of the project was to develop a novel biosensor for the detection of metalloproteins. These
are interesting biological molecules many of which have redox properties that may be detectable in
an electronic device. The metalloproteins chosen for this work play roles in important biological
processes such as photosynthesis (ferredoxin), iron-storage in mammals (ferritin) and respiration
(cytochrome c). The response of the biosensor is consistent with results for un-doped CNTs; a
decline in conductance is observed for non-specific binding of proteins. The devices also sense
the specific antigen-antibody binding of anti-cytochrome c and anti-ferritin. Using apo-proteins,
the mechanism of protein sensing was investigated, and the metal groups within the proteins were
found not to play a significant role. Electrical measurements were carried out using a probe system
developed by Kislon Votchovsky.
2
Introduction
1.2 Publications and Presentations

A list of the presentations and publications that have been produced during this work is given below:
1.2.1 Posters
Electrical conductance and breakdown in individual CNx multiwall nanotubes by Hilary J.

Burch, Julia A. Davies, Elisabetta Brown, Ling Hao, Sonia Antoranz Contera, Nicole Grobert
and John F. Ryan presented at NT06 - Nagano, Japan, 18th - 23rd June 2006.
1.2.2 Oral Presentations
Nitrogen-doped carbon nanotubes as biosensors by Hilary Burch, Characterization Seminar, Department of Materials University of Oxford on 17th October 2005.
Developing a novel biosensor based on nitrogen-doped carbon nanotubes by Hilary Burch,

3rd Year Research Talk, Department of Physics, University of Oxford on 23rd May 2006.
Functionalisation of CNx multi-walled nanotubes with metalloproteins: Effect of adsorption

on protein conformation by Hilary Burch, Biophysics Meeting, Department of Physics, University of Oxford on 28th June 2006.
Conference Proceedings
Biosensing with CNx multi-walled carbon nanotubes by H. J. Burch, S. A. Contera, C. N.

Toledo, M. R. R. dePlanque, N. Grobert, K. Votchovsky and J. F. Ryan Proceedings of the
International Conference on Solid State Devices and Materials (SSDM) - Yokohama, Japan,
12th - 15th September 2006.
1.2.3 Publications
Electrical conductance and breakdown in individual CNx multiwall nanotubes by Hilary J.

Burch, Julia A. Davies, Elisabetta Brown, Ling Hao, Sonia Antoranz Contera, Nicole Grobert
and J. F. Ryan Appl. Phys. Lett. (2006) 89 pg. 143110.
Nitrogen doping of multiwalled carbon nanotubes promotes extensive adsorption of intact

metalloproteins through enhanced carboxylation by Hilary J. Burch, Sonia Antoranz Contera,
Maurits R. R. de Planque, Nicole Grobert and J. F. Ryan submitted to Adv. Mater. (2007).
Acid cutting of CNx multiwalled nanotubes and modified electrical conductance by Hilary J.
Burch, Sonia Antoranz Contera, Elisabetta Brown, Julia A. Davies, Nashville C. Toledo, Ling
Hao, Nicole Grobert and J. F. Ryan submitted to J. Phys. Chem. C. (2007).
3
Introduction
Sensing of metalloproteins with a CNx multiwalled nanotube device by Hilary J. Burch, Sonia
Antoranz Contera, Kislon Votchovsky, Lydia M. Harriss, Nicole Grobert and J. F. Ryan
submitted to Appl. Phys. Lett. (2007).
1.3 Thesis Outline

This thesis contains 8 further chapters
Chapter 2 is an introduction to the physical and electronic properties of CNTs. The theoretical
origins of measured electrical transport properties such as quantized electrical conductance
are also given.
Chapter 3 provides a description of the experimental techniques used in this work. It includes
a section on the principles of Atomic Force Microscopy, and the theoretical principles of the
spectroscopic techniques used to study the metalloproteins.
Chapter 4 contains measurements of the electrical conductance and breakdown of nitrogendoped CNTs.
Chapter 5 is an investigation into the biocompatibility of nitrogen-doped CNTs. Detailed in

this chapter are AFM images of nitrogen-doped CNT/protein conjugates and the experimental
procedures used to obtain successful functionalisation. This chapter also contains characteristic spectra for each metalloprotein.
Chapter 6 contains measurements of the thermal conductance of nitrogen-doped CNTs. An

estimate of the thermal conductivity is also given.
Chapter 7 investigates the effect of acid treatment on nitrogen-doped CNTs in particular with
respect to some of the properties measured in earlier chapters.
Chapter 8 describes the development of a novel biosensor which responds to the binding of
metalloproteins, and the further specific binding of antibodies.
Chapter 9 concludes the thesis and incorporates a discussion of possible future experiments
to improve and consolidate this work.
Introduction
References
[1] A. Javey, J. Guo, Q. Wang, M. Lundstrom, and H. Dai. Ballistic carbon nanotube field-effect
transistors. Nature, 424:654657, 2003.
[2] J. Kong, N. R. Franklin, C. Zhou, M. G. Chapline, S. Peng, K. Cho, and H. Dai. Nanotube
molecular wires as channels sensors. Science, 287:622625, 2000.
[3] S. J. Tans, M. H. Devoret, H. Dai, A. Thess, R. E. Smalley, L. J. Geerligs, and C. Dekker.
Individual single-walled carbon nanotubes as quantum wires. Nature, 386:474477, 1997.
[4] J. M. Wildoer, L. C. Venema, A. G. Rinzler, R. E. Smalley, and C. Dekker. Electronic strucure
and electronic of atomically resolved carbon nanotubes. Nature, 391:5961, 1998.
[5] R. Krupke, F. Hennrich, H. von Lohneysen, and M. M. Kappes. Separation of metallic from
semiconducting single-walled nanotubes. Science, 301:344347, 2003.
[6] R. Czerw, M. Terrones, J.-C. Charlier, X. Blase, B. Foley, R. Kamalakaran, N. Grobert, H. Terrones, P. M. Ajayan, W. Blau, D. Tekleab, M. Ruhle, and D. L. Carroll. Identification of
electron donor states in N-doped carbon nanotubes. Nano Lett., 1(9):457460, 2001.
[7] S. Frank, P. Poncharal, Z. L. Wang, and W. A. de Heer. Carbon nanotube quantum resistors.
Science, 280:17441746, 1998.
[8] P. Poncharal, C. Berger, Y. Yi, L. Wang, and W. A. de Heer. Room temperature ballistic
conduction in carbon nanotubes. J. Phys. Chem. B., 106:1210412118, 2002.
[9] C. Berger, Y. Yi, J. Gezo, P. Poncharal, and W. A. de Heer. Contacts, non-linear transport
effects and failure in multi-walled carbon nanotubes. New J. Phys., 5:158.1, 2003.
[10] H. Dai, E. W. Wong, and C. M. Lieber. Probing electrical transport in nanomaterials: conductivity of individual carbon nanotubes. Science, 272:523526, 1996.
[11] T. W. Ebbesen, H. J. Lezec, H. Hiura, J. W. Bennett, H. F. Ghaemi, and T. Thio. Electrical
conductivity of individual carbon nanotubes. Nature, 382:5456, 1996.
[12] J. Kong, H. T. Soh, A. M. Cassell, C. F. Quate, and H. Dai. Synthesis of individual singlewalled carbon nanotubes on patterned silicon wafers. Nature, 395:878881, 1998.
[13] C. Zhou, J. Kong, and H. Dai. Electrical measurements of individual semiconducting singlewalled nanotubes of various diameters. Appl. Phys. Lett., 76(12):15971599, 2000.
[14] A. Hassanien, M. Tokumoto, Y. Kumazawa, H. Kataura, Y. Maniwa, S. Suzuki, and Y. Achiba.
Atomic structure and electronic properties of single-wall carbon nanotubes probed by scanning
tunneling microscope at room temperature. Appl. Phys. Lett., 73(26):38393841, 1998.
[15] K. Xiao, Y. Liu, P. Hu, G. Yu, Y. Sun, and D. Zhu. n-type field-effect transistors made of an
individual nitrogen-doped multiwalled carbon nanotube. J. Am. Chem. Soc., 127:86148617,
2005.
5
Introduction
[16] K. Xiao, Y. Liu, P. Huang, G. Yu, L. Fu, and D. Zhu. High performance field-effect transistors
made of a multiwall CNx /C nanotube intermolecular junction. Appl. Phys. Lett., 83(23):4824
4826, 2003.
[17] D. Golberg, P. S. Dorozhkin, Y. Bando, Z.-C. Dong, C. C. Tang, Y. Uemura, N. Grobert,
M. Reyes-Reyes, H. Terrones, and M. Terrones. Structure, transport and field-emission properties of compound nanotubes: CNx vs BNCx (x 0.1). Appl. Phys. A, 76:499507, 2003.
[18] H. J. Choi, J. Ihm, S. G. Louie, and M. L. Cohen. Defects, quasibound states and quantum
conductance in metallic carbon nanotubes. Phys. Rev. Lett., 84(13):29172920, 2000.
[19] R. J. Chen, S. Bangsaruntip, K. A. Drouvalakis, N. Wong Shi Kam, M. Shim, Y. Li, W. Kim,
P. J. Utz, and H. Dai. Noncovalent functionalization of carbon nanotubes for highy specific
electronic biosensors. Proc. Natl. Acad. Sci. USA, 100(9):49844989, 2003.
[20] P. W. Barone, S. Baik, D. A. Heller, and M. S. Strano. Near-infrared optical sensors based on
single-walled carbon nanotubes. Nature Materials, 4:8692, 2005.
[21] N. Wong Shi Kam, T. C. Jessop, P. A. Wender, and H. Dai. Nanotube molecular transporters:
Internalization of carbon nanotube-protein conjugates in to mammalian cells. J. Am. Chem.
Soc., 126:68506851, 2004.
[22] F. A. Armstrong, H. A. O. Hill, and N. J. Walton. Direct electrochemistry of redox proteins.
Acc. Chem. Res., 21:407413, 1988.
[23] M. J. Eddowes and H. A. O. Hill. Novel method for the investigation of the electrochemisty
of metalloproteins: Cytochrome c. J. C. S. Chem. Comm., pages 771772, 1977.
[24] J. C. Carrero-Sanchez, A. L. Elas, R. Mancilla, G. Arrelln, H. Terrones, J. P. Laclette, and
M. Terrones. Biocompatibility and toxicological studies of carbon nanotubes doped with nitrogen. Nano Lett., 6(8):16091616, 2006.
[25] K. Jiang, L. S. Schadler, R. W. Siegel, X. Zhang, H. Zhang, and M. Terrones. Protein immoblisation on carbon nanotubes via a two-step process of diimide activated amidation. J. Mater.
Chem., 14:3739, 2004.
[26] M. S. Dresselhaus, G. Dresselhaus, and P. C. Eklund. Science of fullerenes and carbon nanotubes. Academic Press Inc., New York, 1st edition, 1996.
[27] Z. Yao, J.-S. Wang, B. Li, and G.-R. Liu. Thermal conductance of carbon nanotubes using
molecular dynamics. Phys. Rev. B, 71:085417, 2005.
[28] J. Hone, M. Whitney, C. Piskoti, and A. Zettl. Thermal conductivity of single-walled carbon
nanotubes. Phys. Rev. B, 59(4):25142516, 1999.
[29] W. Yi, L. Lu, Z. Dian-lin, Z. W. Pan, and S. S. Xie. Linear specific heat of carbon nanotubes.
Phys. Rev. B, 59(14):90159018, 1999.
[30] E. Brown, L. Hao, J. C. Gallop, and J. C. Macfarlane. Ballistic thermal and electrical conductance measurements on individual multiwall carbon nanotubes. Appl. Phys. Lett., 87:023107,
2005.
6
Introduction
[31] S. Sotiropoulou and N. A. Chaniotakis.

Bioanal. Chem., 375:103105, 2003.
Carbon nanotube array-based biosensor.
Anal.
[32] G. Gruner. Carbon nanotube transistors for biosensing applications. Anal. Bioanal. Chem.,
384:322335, 2006.
[33] K. Bradley, M. Briman, A. Star, and G. Gruner. Charge transfer from adsorbed proteins. Nano
Lett., 4(2):253256, 2004.
[34] K. Besteman, J.-O. Lee, F. G. M. Wiertz, H. A. Heering, and C. Dekker. Enzyme-coated
carbon nanotubes as single-molecule biosensors. Nano Lett., 3(6):727730, 2003.
[35] A. Star, J.-C. P. Gabriel, K. Bradley, and G. Gruner. Electronic detection of specific protein
binding using nanotube FET devices. Nano Lett., 3(4):459463, 2003.
[36] S. Boussaad, N. J. Tao, R. Zhang, T. Hopson, and L. A. Nagahara. In situ detection of cytochrome c adsorption with single walled carbon nanotube device. Chem. Commun., pages
15021503, 2003.
Chapter 2
Carbon Nanotubes
CNTs were first identified by Oberlin and Endo 1 in 1976 and were illustrated as cylindrical fullerenes, 2
capped with a buckyball hemisphere by Smalley 3 in 1990. The first reported synthesis of CNTs was
by Iijima 4 in 1991, which started the field of research into their many interesting physical and electrical properties.
2.1 Structure
Iijimas CNTs consisted of several coaxial graphite layers wound around a common axis, with an
inter-layer separation of
They were called multiwalled nanotubes (MWNTs).

3.4 A.
Single-walled nanotubes (SWNTs) were synthesized in 1993, 5,6 and introduced the use of Co
catalysts during synthesis (see section 2.3) in addition to Fe. SWNTs can be envisaged as a single
graphene sheet wrapped into a seamless cylinder (fig. 2.1), the CNT has axial symmetry and exhibits
a spiral conformation called chirality. 7
2.2 Nitrogen-doping of CNTs

After the successful production of nitrogenous C60 and C70 fullerenes by Pradeep et al. 8 the possibility of stable carbon-nitride tubes was proposed. 911
8
Carbon Nanotubes
Figure 2.1: A single walled CNT
Figure 2.2: Possible bonding configurations for N in graphitic networks: (a) pyridine-like N, (b)
pyrrole-like N, (c) substitutional N, (d) nitrile -C N, (e) amine - NH2 , (f) single N pyridinic vacancy, (g) triple N pyridinic vacancy, and (h) interstitial N, from Ewels et al. 26
The first synthesis of nitrogen-doped multiwalled nanotubes (CNx MWNTs, x = 2-5%) was reported by Terrones et al. 12 in 1997, via synthesis of MWNTs in a nitrogen-rich atmosphere. Since
then, various synthetic routes and nitrogen concentrations have been reported for CNx MWNTs 1319
which have interesting electronic, 20,21 and physical 11,22 properties. The synthesis of the CNx
MWNTs used in this work was reported by Reyes-Reyes et al. 23 in 2004. More recently, CNx
SWNTs have been produced. 24,25
2.2.1 Doping Configurations

Nitrogen is incorporated into the CNT lattice during growth. The possible bonding configurations
for N in graphitic networks are shown in fig. 2.2. The three most common bonding configurations
9
Carbon Nanotubes
are:
pyridine-like sp2 - hybridized 6-fold ring arrangement (a).
pyrrole-like sp2 - hybridized 5-fold ring arrangement (b).
substitutional sp3 - hybridized 6-fold ring arrangement (c).
However, only the third arrangement, where a N atom is directly substituted for a C atom in
the hexagonal network, generates an electrical donor state 27 (where electrons from N are donated
into the graphitic network). In this arrangement, N uses three of its valence electrons to form 3
bonds, one electron fills the state and the fifth electron enters the state. In the case of the
pyridine-like arrangement, the lone pair (two electrons on N that not are required for bonding)
enters a non-bonding (px ) orbital and pyrrole-like N donates its remaining two electrons to a
orbital, completing the aromatic ring. As a result these two configurations are non-doping. 27 More
information on hybridization can be found in the Appendix (section 10.3).
Experimental evidence for the pyridine-like configuration can be seen in the intensity ratio of
the G to D bands in the Raman spectrum of CNx MWNTs, 28 also, x-ray photoelectron spectroscopy
(XPS) reveals both pyridine-like and substitutional nitrogen atoms in the hexagonal structure. 14
Nitrogen may also bond as a nitrile -C N [fig. 2.2(d)] or an amine -NH2 [fig. 2.2(e)] or exist as
N2 intercalated between the graphite layers. 23,2931
2.2.2 Defects caused by Nitrogen-doping

Two of the bonding possibilities above cause vacancies in the lattice [fig. 2.2(f) and (g)] or defects.
Srivastava et al. 32 found (via molecular dynamics simulations) that the type of defect formed is
dependent on the diameter of the CNT, and hence the strain energy induced by curvature of the
graphene sheet. For a CNT with a diameter less than 8 nm, when a N defect occurs, a nearby C
atom moves towards the location of the vacancy and forms two pentagons sharing a vertex. The
reconstruction for diameters greater than 8 nm is more simple, none of the C atoms around the
vacancy move, they remain equivalent forming three incomplete pentagons. 32
10
Carbon Nanotubes
Figure 2.3: (a) The bamboo-like structure of CNx MWNTs, from Jang et al. 33
2.2.3 Structure
The effect of nitrogen incorporation in the CNx MWNTs is to produce a unique corrugated hollow
nanotube structure, that is very different to the perfect structure of SWNTs. The CNx MWNTs are
also thicker than un-doped MWNTs, being 10-100 nm in diameter and up to several m in length.
The unique structure is due to the distortion and bending of the graphene layers around nitrogen
defects, which gives a bamboo-like appearence (fig. 2.3). Sections of this disordered material
(100-150 nm) are connected by straight segments of the familiar MWNT structure. 14 AFM and
Lateral FM measurements have shown that the wall of the nanotube is thicker in the straight region
than in the disordered region. This causes the characteristic surface roughness of bamboo CNTs,
which are curved because of the imperfect linkage of the bamboo-like compartments. 33
The characteristic bamboo structure can also be considered as polymerised carbon nitride nanobells 18
or nanocages, 34 where each nano-compartment is linked together to form the tube. Investigations
suggest that the overall morphology of the tubes, depends strongly on the nitrogen concentration. 13
2.3 Synthesis
There are several methods for the synthesis of CNTs including arc discharge, laser-vaporisation
and chemical vapour deposition (CVD), the latter being used for the synthesis of the CNx MWNTs
reported in this work. During CVD, a metal catalyst particle is deposited on a substrate, etched in
11
Carbon Nanotubes
Figure 2.4: TEM images of CNx MWNTs (Nicole Grobert)
dilute HF and then placed in a quartz boat. The boat is positioned in a CVD reaction furnace, and
after additional etching of the catalytic metal film, using NH3 gas at 750-1050 C, nano-size fine
catalytic metal particles are formed. The CNTs are grown on these fine catalytic metal sites.
2.3.1 Synthesis of CNx MWNTs

Gram quantities of pure and well-aligned CNx MWNTs were prepared by pyrolysing, at 950 C,
homogeneously dispersed aerosols generated from benzylamine/ferrocene solutions using an ultrasonic spraying device. Transmission electron microscopy (TEM) images (shown in fig. 2.4) reveal
that the products are generally arranged in carpet-like flakes and are almost free of any by-products,
such as polyhedral particles or amorphous carbon. 23
12
Carbon Nanotubes
Figure 2.5: The position of pentagons in a CNT cap
2.4 Growth Mechanism

2.4.1 Growth of Un-doped CNTs
Un-doped MWNTs consist of hollow carbon hexagonal networks, concentrically arranged around
one another, and terminated by end caps in which the surfaces meet under specific angles. 35 These
caps must include at least six pentagons in the hexagonal network (fig. 2.5) in order to satisfy
Eulers theorem. Most CNTs are closed, but in rare cases the tubes are open at either end. This
can be explained in terms of the growth mechanism. For the open tubes, the growth mechanism
seems to have been terminated by a sudden change in conditions, so that the species in the plasma
are unable to form the hexagonal or pentagonal rings that lead to closure of the tube ends. 36
Eulers Theorem
Like any simple polyhedron, a SWNT satisfies Eulers theorem relating the number of vertices (C
atoms), edges (covalent bonds), and faces: V - E + F = 2. If the number of pentagons is p, and
the other (F - p) faces are all hexagonal, then the doubled number of edges (each edge belongs to
two faces) is 5p + 6(F - p), which also equals the tripled number of vertices (each trivalent carbon
is shared by three adjacent faces). A simple accounting then yields p = 12, and therefore a perfect
defectless SWNT must have exactly 12 pentagons. 37
These pentagons are positive surface disinclinations in the hexagonal lattice which provide posi13
Carbon Nanotubes
tive curvature (+60 ) for the formation of a cap to close the tube. Some CNTs also contain heptagons
which are negative surface disinclinations which provide negative curvature (-60 ), opening up the
structure.
2.4.2 Growth Model

High-resolution TEM observations have been used to propose an open-ended growth model in
which carbon atoms and small clusters from the discharge plasma add on to the reactive dangling
bonds at the edges of the open-ended CNTs. 36 In the case of MWNTs, the walls are concentrically
thickened layer-by-layer from the inside to the outside along the same tubular axis.
During the growth of CNTs, the diffusion of carbon into the catalyst metal substrate, is believed
to be the rate-determining step. This rate can be described by an Arrhenius equation in which the
activation energy is the diffusion energy of carbon in the metal. 38
AeEa
RT
(2.1)
Where k is the rate constant, R is the gas constant, A is the pre-exponential factor and Ea is
the activation energy. Provided that the activation energy is positive, the rate constant will increase
with increasing temperature (T). Support for this model comes from experiments on the kinetics
of growth of CNTs from acetylene on -Fe metal catalysts at 1000 C. This yields an activation
energy 39 of 142 kJmol1 , which compares favourably to the activation energy of diffusion for iron 40
of 148 kJmol1 .
2.4.3 Growth of CNx MWNTs

Conventional CNT growth models cannot explain the regular internal bamboo cavities of nitrogendoped CNTs, so an alternative model has been proposed. 23,28,41 During synthesis, CNx species are
generated and deposited on the metal catalyst particles, where they react exothermically. Diffusion
of CNx species through the metal catalyst particle is responsible for CNT growth, and encapsulation
of N2 is caused by the decomposition of unstable reactive intermediates (fig. 2.6). The compart14
Carbon Nanotubes
Figure 2.6: The growth mechanism for CNx MWNTs from Reyes-Reyes et al. 23
mentalised structure is a result of the different precipitation rates of different CNx species; when
precipitation is slow, and there is a lack of material to maintain inner CNT growth, layers will
suddenly close.
2.5 Electronic Properties

2.5.1 SWNTs
Scanning Tunneling Microscopy (STM) studies of SWNTs have shown that tubes with certain
atomic configurations possess a finite density of states (DOS) at the Fermi Energy (E f ), while others
have none. 42,43 These two types of SWNTs are interpreted as having metallic and semiconducting
electrical transport properties respectively. 44
In the DOS at higher energies, sharp peaks are observed, referred to as van Hove singularities,
which are characteristic of low-dimensional conducting materials. These features are normally
symmetric about E f and their separation increases with decreasing tube diameter. 45 For an un-doped
metallic tube, E f is located at the charge neutrality point (CNP) where the bonding and antibonding
bands cross. 46
15
Carbon Nanotubes
Figure 2.7: Wrapping vector of a single walled CNT from Wildoer et al. 42
2.5.2 Chirality
As mentioned in section 2.1, SWNTs exhibit a spiral conformation called chirality. This strongly
affects their electronic properties, 47 via the two parameters of the atomic lattice, diameter (d) and
chiral angle (). 42,43,48 The chiral angle is the angle between the hexagon rows and the tube axis.
Consider the two-dimensional graphene sheet in fig. 2.7. In the circumferential direction (along
C), periodic boundary conditions apply, that is:
Ck
2q
(2.2)
Where k is the wave vector and q is an integer. This condition leads to a set of allowed values
for k which can be substituted into the energy dispersion relation for a graphene sheet to yield the
dispersion relations for the tube, with k labelling the various one-dimensional modes.
16
Carbon Nanotubes
Applying the wrapping vector

If we focus on the wrapping process in more detail, the overlapping of one of the hexagons on top
of another introduces a number of lattice vectors in both the x and y directions (a1 ) and (a2 ). The
overall wrapping vector which defines the relative location of the two overlapping hexagon sites is
C, and it is specified by a pair of integers (n,m). These integers relate C to the two unit vectors (a1 )
and (a2 ) such that:
C
na1 ma2
(2.3)
A metal can be defined as having filled bands at E f . In the case of CNTs, the way in which
we fold the graphene sheet can result in either occupied or un-occupied states at E f . These two
possibilities would make the tube either metallic or semiconductor respectively. 49
The condition for occupied states at E f is:
n-m
3p q
(2.4)
Thus:
A CNT is metallic if n - m is 3 or a multiple of 3.
Where the extra q is required the CNT is semiconducting.
The unit cell of the SWNT strongly depends on the choice of n and m. Dresselhaus et al. 50
proposed that three types of CNTs are possible depending on the values of (n,m):
Armchair tubes are defined by n = m and are metallic.
Zigzag tubes have m = 0 and are metallic or semiconducting depending on equation 2.4. These
and armchair tubes have the smallest unit cell.
All other tubes are known as chiral and have a finite wrapping angle with 0
The diameter of the CNT is related to the (n,m) values by the Hamada vector: 51
17
30 .
Carbon Nanotubes
a n2 m2 nm
(2.5)
for a rolled up CNT cylinder. 52 The energy

Where a is the lattice constant, taken to be 2.49 A
gap for a semiconducting tube is inversely proportional to the tube diameter and is of the order of 1
eV for a small diameter (
CNT. 53
7 A)
2.6 Thermal Properties

In macroscopic systems Fouriers Law describes the thermal conduction. This states that the time
rate of heat flow Q through a portion of material is proportional to the gradient of temperature
difference:
Q
T
A
t
L
(2.6)
where t is time, A is area, T is the temperature difference through which the heat is being
transferred, and L is the length of the body of matter through which the heat is passing. The constant
of proportionality for equation 2.6 is the thermal conductivity () defined as:
Q
t
L
AT
(2.7)
where heat is transmitted in a direction normal to a surface. In Chapter 6 this expression is

used to calculate for CNx MWNTs, but it cannot account for all thermal phenomena observed in
nanoscale systems. 54 Although disagreement still exists in the literature on the thermal properties
of CNTs, it is generally accepted that depends on T and possibly also on current and vacancy
concentration. 55
In conventional metals, electrical and thermal conductivity are closely related, however in CNTs
(and also graphite and diamond) phonons, not electrons are the primary carriers of heat, 56 so the
relationship between thermal and electrical conductance is not clear. The phonon free mean path in
CNTs is estimated to be very long 57 (100 nm - 1 m) 54 thus some CNTs can have ballistic thermal
18
Carbon Nanotubes
transport properties. 58,59

The thermal properties of CNx MWNTs have not been previously investigated, but as individual
CNx MWNTs are extremely long (up to 10 m) ballistic thermal transport is unlikely. Similar to
electrical conductance, the thermal conductance of CNTs is also quantised 60 in units of Gth where:
Gth
2 kB2 T
3h
(2.8)
In practice however, phonon scattering reduces the thermal conductivity, making it difficult to observe quantum thermal transport except at ultra low temperatures. 58
19
Carbon Nanotubes
References
[1] A. Oberlin, M. Endo, and T. Koyama. Filamentous growth of carbon through decomposition
of benzene. J. Cryst. Growth, 32:335349, 1976.
[2] H. W. Kroto. Semistable molecules in the laboratory and in space. Chem. Soc. Rev., 11:453
449, 1982.
[3] R. E. Smalley. DoD workshop. Washington DC, 1990.
[4] S. Iijima. Helical microtubules of graphitic carbon. Nature, 354:5658, 1991.
[5] S. Iijima and T. Ichihashi. Single-shell carbon nanotubes of 1-nm diameter. Nature, 363:603
605, 1993.
[6] D. S. Bethune, C. H. Kiang, M. S. DeVries, G. Gorman, R. Savoy, and R. Beyers. Cobalt
catalysed growth of carbon nanotubes with single-atomic-layer walls. Nature, 363:605607,
1993.
[7] J.-P. Salvetat, J.-M. Bonard, N. H. Thomson, A. J. Kulik, L. Forro, W. Benoit, and L. Zuppiroli.
Mechanical properties of carbon nanotubes. Appl. Phys. A, 69:255260, 1999.
[8] T. Pradeep, V. Vijayakrishnan, A. K. Santra, and C. N. R. Rao. Interaction of nitrogen with
fullerenes: Nitrogen derivatives of C60 and C70 . J. Phys. Chem., 95:1056410565, 1991.
[9] Y. Miyamoto, M. L. Cohen, and S. G. Louie. Theoretical investigation of graphitic carbon
nitride and possible tubule forms. Solid State Commun., 102(8):605608, 1997.
[10] K. Suenaga, M. P. Johansson, N. Hellgren, E. Broitman, L. R. Wallenberg, C. Colliex, J.-E.
Sundgren, and L. Hultman. Carbon nitride nanotubulite - densely-packed and well-aligned
tubular structures. Chem. Phys. Lett., 300:695700, 1999.
[11] A. Y. Liu and R. M. Wentzcovitch.
50(14):1036210365, 1994.
Stability of carbon nitride solids.
Phys. Rev. B,
[12] M. Terrones, N. Grobert, J. Olivares, J. P. Zhang, H. Terrones, K. Kordatos, W. K. Hsu, J. P.

Hare, P. D. Townsend, K. Prassides, A. K. Cheetham, H. W. Kroto, and D. R. M. Walton.
Controlled production of aligned-nanotube bundles. Nature, 388:5255, 1997.
[13] X. Wang, Y. Liu, D. Zhu, L. Zhang, H. Ma, M. Yao, and B. Zhang. Controllable growth,
structure and low field emission of well-aligned CNx nanotubes. J. Phys. Chem. B., 106:2186
2190, 2002.
[14] M. Nath, B. C. Satishkumar, A. Govindaraj, C. P. Vinod, and C. N. R. Rao. Production of
bundles of aligned carbon and carbon-nitrogen nanotubes by the pyrolysis of precursors on
silica-supported iron and cobalt catalysts. Chem. Phys. Lett., 322:333340, 2000.
[15] M. Glerup, M. Castignolles, M. Holzinger, G. Hug, A. Loiseau, and P. Bernier. Synthesis of
highly nitrogen-doped multi-walled carbon nanotubes. Chem. Commun., pages 25422543,
2003.
20
Carbon Nanotubes
[16] M. Terrones, P. Redlich, N. Grobert, S. Trasobares, W. K. Hsu, H. Terrones, Y. Q. Zhu, J. P.

Hare, C. L. Reeves, A. K. Cheetham, M. Ruhle, H. W. Kroto, and D. R. M. Walton. Carbon
nitride nanocomposites: formation of aligned Cx Ny nanofibres. Adv. Mater., 11(8):655658,
1999.
[17] M. Terrones, H. Terrones, N. Grobert, W. K. Hsu, Y. Q. Zhu, J. P. Hare, H. W. Kroto, D. R. M.
Walton, Ph. Kohler-Redlich, M. Ruhle, Z. P. Zhang, and A. K. Cheetham. Effiecient route
to large arrays of CNx nanofibres by pyrolysis of ferrocene/melamine mixtures. Appl. Phys.
Lett., 75(25):39323934, 1999.
[18] S. H. Lim, H. I. Elim, X. Y. Gao, A. T. S. Wee, W. Ji, J. Y. Lee, and J. Lin. Electronic and
optical properties of nitrogen-doped multiwall carbon nanotubes. Phys. Rev. B, 73:045402,
2006.
[19] R. Droppa Jr., C. T. M. Ribeiro, A. R. Zanatta, M. C. dos Santos, and F. Alvarez. Comprehensive spectroscopic study of nitrogenated carbon nanotubes. Phys. Rev. B, 69:045405, 2004.
[21] H. J. Burch, J. A. Davies, E. Brown, L. Hao, S. Antoranz Contera, N. Grobert, and J. F. Ryan.
Electrical conductance and breakdown in individual CNx multiwall nanotubes. Appl. Phys.
Lett., 89:143110, 2006.
[22] D. M. Teter and R. J. Hemley. Low-compressibility carbon nitrides. Science, 271(5245):53
55, 1996.
[23] M. Reyes-Reyes, N. Grobert, R. Kamalakaran, T. Seeger, D. Golberg, M. Ruhle, Y. Bando,
H. Terrones, and M. Terrones. Efficient encapsulation of gaseous nitrogen inside carbon nanotubes with bamboo-like structure using aerosol thermolysis. Chem. Phys. Lett., 396:167173,
2004.
[24] J. G. Wiltshire, L.-J. Li, L. M. Herz, R. J. Nicholas, M. Glerup, J.-L. Sauvajol, and A. N.
Khlobystov. Chirality-dependent boron-mediated growth of nitrogen-doped single-walled carbon nanotubes. Phys. Rev. B, 72:205431, 2005.
[25] V. Krstic, G. L. J. A. Rikken, P. Bernier, S. Roth, and M. Glerup. Nitrogen doping of metallic single-walled carbon nanotubes: n-type conduction and dipole scattering.
http://arxiv.org/abs/cond-mat/0601513.
[26] C. P. Ewels and M. Glerup. Nitrogen doping in carbon nanotubes. J. Nanosci. Nanotech.,
5(9):13451363, 2005.
[27] J. Robertson and C. A. Davis. Nitrogen doping in tetrahedral amorphous carbon. Diamond
relat. mater., 4:441444, 1995.
[28] J. H. Yang, B. J. Kim, Y. H. Kim, Y. J. Lee, B. H. Ha, Y. S. Shin, S.-Y. Park, H. S. Kim, C.-Y.
Park, C.-W. Yang, J.-B. Yoo, M. H. Kwon, K. Ihm, H.-J. Song, T.-H. Kang, H.-J. Shin, J.-Y.
21
Carbon Nanotubes
Park, and J.-M. Kim. Nitrogen-incorporated multiwalled carbon nanotubes grown by direct
current plasma-enhanced chemical vapor deposition. J. Vac. Sci. Technol. B, 23(3):930933,
2005.
[29] K. Suenaga, M. Yudasaka, C. Colliex, and S. Iijima. Radially modulated nitrogen distribution
in CNx nanotubular structures by CVD using Ni phthalocyanine. Chem. Phys. Lett., 316:365
372, 2000.
[30] J. Kotakoski, A. V. Krasheninnikov, Y. Ma, A. S. Foster, K. Nordlund, and R. M. Nieminen. B
and N ion implantation into carbon nanotubes: Insight from atomistic simulations. Phys. Rev.
B, 71:205408, 2005.
[31] H. C. Choi, S. Y. Bae, J. Park, K. Seo, C. Kim, B. Kim, H. J. Song, and H.-J. Shin. Experimental and theoretical studies on the structure on the structure of N-doped carbon nanotubes:
Possibility of intercalated molecular N2 . Appl. Phys. Lett., 85(23):57425744, 2004.
[32] D. Srivastava, M. Menon, C. Daraio, S. Jin, B. Sadanadan, and A. M. Rao. Vacancy-mediated
mechanism of nitrogen substitution in carbon nanotubes. Phys. Rev. B, 69:153414, 2004.
[33] J. W. Jang, C. E. Lee, T. J. Lee, S. C. Lyu, and C. J. Lee. Lateral force microscopy of bambooshaped multiwalled carbon nanotubes. Curr. Appl. Phys., 6:141144, 2006.
[34] M. Lin, J. P. Y. Tan, C. Boothroyd, K. P. Loh, E. S. Tok, and Y.-L. Foo. Direct observation of
single-walled carbon nanotube growth at the atomistic scale. Nano Lett., 6(3):449452, 2006.
[35] D. Zhou and L. Chow. Complex structure of carbon nanotubes and their implications for
formation mechanism. J. Appl. Phys., 93(12):99729976, 2003.
[36] S. Iijima, P. M. Ajayan, and T. Ichihashi. Growth model for carbon nanotubes. Phys. Rev.
Lett., 69(21):31013103, 1992.
[37] http://www.americanscientist.org. Fullerene nanotubes: C1 000 000 and beyond.
[38] C. L. Lee, J. H. Park, and J. Park. Synthesis of bamboo-shaped multiwalled carbon nanotubes
using thermal chemical vapor deposition. Chem. Phys. Lett., 323:560565, 2000.
[39] R. T. K. Baker, J. J. Chludzinski, and C. R. F. Lund. Activation energy of carbon nanotubes.
In Extended abstracts of the 18th Biennial Conference on Carbon, page 155, Worcester Polytechnic Institution, 1987.
[40] C. J. Smithells, E. A. Brandes, and G. B. Brook, editors. Smithells Metals Reference Book.
Butterworth-Heinemann, 7th edition, 1992.
[41] M. Terrones, P. M. Ajayan, F. Banhardt, X. Blase, D. L. Carroll, J.-C. Charlier, R. Czerw,
B. Foley, N. Grobert, R. Kamalakaran, P. Kohler-Redlich, T. Seeger, and H. Terrones. Ndoping and coalescence of carbon nanotubes: synthesis and electronic properties. Appl. Phys.
A, 74:355361, 2002.
[42] J. M. Wildoer, L. C. Venema, A. G. Rinzler, R. E. Smalley, and C. Dekker. Electronic strucure
and electronic of atomically resolved carbon nanotubes. Nature, 391:5961, 1998.
22
Carbon Nanotubes
[43] T. W. Odom, J.-L. Heung, P. Kim, and C. M. Lieber. Atomic structure and electronic properties
of single-walled carbon nanotubes. Nature, 391:6264, 1998.
[44] H. Ago, T. Kugler, F. Cacialli, W. R. Salaneck, M. S. P. Shaffer, A. H. Windle, and R. H.
Friend. Work functions and surface functional groups of multiwall carbon nanotubes. J. Phys.
Chem. B., 103:81168121, 1999.
[45] M. A. Pimenta, A. Marucci, S. D. M. Brown, and M. J. Matthews. Resonant raman effect in
single-wall carbon nanotubes. J. Mater. Res., 13(9):23962404, 1998.
[46] B. Stojetz, C. Miko, L. Forro, and C. Strunk. Effect of band structure on quantum intereference
in multiwall carbon nanotubes. Phys. Rev. Lett., 94:186802, 2005.
[47] R. Saito, M. Fujita, G. Dresselhaus, and M. S. Dresselhaus. Electronic structure of chiral
graphene tubules. Appl. Phys. Lett., 60:22042206, 1992.
[48] A. Bezryadin, A. R. M. Verschueren, S. J. Tans, and C. Dekker. Multiprobe transport experiments on single-wall carbon nanotubes. Phys. Rev. Lett., 80(18):40364039, 1998.
[49] R. J. Nicholas. Optical Properties of Carbon Nanotubes, Department of Physics Seminar,
2004.
[51] X. Liu, T. Pichler, M. Knupfer, M. S. Golden, J. Fink, H. Kataura, and Y. Achiba. Detailed
analysis of the mean diameter and diameter distribution of single-wall carbon nanotubes from
their optical response. Phys. Rev. B, 66:045411, 2002.
[52] R. Saito, M. S. Dresselhaus, and G. Dresselhaus. Physical Properties of Carbon Nanotubes.
Imperial College Press, 1998.
[53] R. Saito, G. Dresselhaus, and M. S. Dresselhaus. Tunneling conductance of connected carbon
Phys. Rev. B, 59(14):90159018, 1999.
[57] D. G. Cahill, W. K. Ford, K. E. Goodson, G. D. Mahan, A. Majumdar, H. J. Maris, R. Merlin,
and S. R. Phillpot. Nanoscale thermal transport. Appl. Phys. Rev., 93(2):793818, 2003.
[58] H.-Y. Chiu, V. V. Deshpande, H. W. Ch. Postma, C. N. Lau, C. Miko, L. Forro, and M. Bockrath. Ballistic phonon thermal transport in multiwalled carbon nanotubes. Phys. Rev. Lett.,
95:226101, 2005.
23
Carbon Nanotubes
2005.
[60] T. Yamamoto, S. Watanabe, and K. Watanabe. Universal features of quantized thermal conductance of carbon nanotubes. Phys. Rev. Lett., 92(7):075502, 2004.
24
Chapter 3
Experimental Techniques
3.1 Atomic Force Microscopy
AFM is one of a family of microscopy techniques that comes under the umbrella term Scanning
Probe Microscopy. 1 The first example of this type of microscopy was Scanning Tunelling Microscopy (STM), which generated an image by measuring a tunneling current between a conducting
surface and a conducting tip. AFM does not rely on tunneling to probe local properties of a sample,
but rather on the tip-sample force interaction.
3.1.1 Tip-sample interaction

Consider a tip and sample about 100 nm apart. At this separation, any attractive interactions are too
small to exert a significant force between the atoms at the end of the tip and the substrate. As the
separation is reduced, the force between tip and sample rises rapidly. This occurs even if both the
tip and sample are electrically neutral, the forces being exerted are van der Waals forces.
Van der Waals Forces

When two electrically neutral and non-magnetic bodies are held at a distance of 1 - 10 nm, van
der Waals (VDWs) forces will act between them. On a molecular scale, three separate interactions
contribute to VDWs forces: 2
25
1. Orientation Forces
2. Induction Forces
3. Dispersion Forces
Orientation forces arise from the interaction of two polar molecules. The permanent dipole moments on each molecule interact to generate an attractive force known as dipole-dipole interactions.
For two anti-parallel dipoles (A and B) in their ground states, interacting in adjacent bundles, the
interaction energy is given by: 2
UAB
A B
40 r3
(3.1)
where A and B are the corresponding dipole moments, 0 is the permittivity of free space,
the permittivity of the material and r is the separation between them.
Induction forces arise from the interaction of a polar molecule with a non-polar molecule. The
polar molecule induces a dipole on a nearby non-polar molecule, generating an attractive force
known as a dipole-induced dipole interaction.
Dispersion forces are the most important form of interaction to consider in the context of understanding AFM. Their existence was proposed by London 3 in 1937, who suggested that the electrons
in a neutral molecule are in continuous motion. Thus for very short time intervals, these molecules
will possess finite fluctuating dipole and higher multipole moments. On adjacent molecules, the
interaction of simultaneously created dipoles gives rise to an attractive force. Mathematically, the
dispersion energy cannot be completely described in classical terms, as its origins are quantum
mechanical. It takes the form:
Udisp
32 I
440 2 r6
(3.2)
where r is the distance between two interacting molecules, is the polarisability and I is the
ionisation potential. Thus the dispersion force is proportional to r7 , demonstrating that dispersion
forces are long-range and rapidly increase as molecules are brought together, mutually aligning
26
Figure 3.1: Scheme of the tip interacting with the sample. The little circles represent atoms (Kislon
Votchovsky).
neighbouring molecules. VDWs forces are important in the understanding of the interactions within
CNT bundles, 4 and will be discussed further in section 7.4.1.
Electrostatic Forces
Although VDWs forces play a significant role in the tip-sample interaction. When the sample and
tip are charged, the largest physical influence between them will be the electrostatic or Coulombic
force. For two oppositely charged ions q1 and q2 in a vacuum, with a separation r the force between
them is given by:
q1 q2
40 r2
(3.3)
As the ions are brought closer together the attractive force between them rises sharply until
eventually the outer electrons around each ion interact, and repel each other.
Lennard-Jones Potential
Repulsive forces also result when a tip and sample interact at a small distance, counteracting the
attractive dispersion forces. Imagine that an atom at the very apex of the AFM tip, interacts with a
particle at the surface of the sample (see fig. 3.1).
The behaviour of the potential energy as the distance r between them changes (fig. 3.2). It can
be described by the pair-potential energy function Epair r. In the case of hard spheres, as we are
27
Figure 3.2: The Lennard-Jones Potential

approximating here, the pair-potential takes the special case of the Mie pair-potential, the LennardJones function:
E pair r
12
(3.4)
where is the hard sphere diameter.
Capillary Forces
Additional forces also contribute to the potential between the tip and sample, particularly when
imaging in different media. In air for example, the tip resting on a surface is an ideal nucleation site
for the condensation of water vapour present in the air. This means that the tip will be pulled down
towards the sample by a strong liquid meniscus, giving rise to a so-called capillary force. 1
28
Figure 3.3: A silicon nitride AFM cantilever
3.1.2 The Microscope Tip

An AFM tip consists of an extremely sharp (Si or Si3 N4 ) micro-fabricated spike, mounted on the end
of a cantilever (fig. 3.3). The cantilever allows the tip to move up and down as it tracks the sample.
AFM cantilevers can be fabricated with a force constant between 0.01 - 50 N/m and the cantilevers
used in this work have a force constant of 2 N/m. The force contribution from the bending of the
cantilever is determined by Hookes Law:
kd
(3.5)
where k is the force constant and d is its displacement. The resonant frequency (f ) of the
cantilever is given by:
1
2
k
m
(3.6)
where m is the effective mass of the cantilever. k increases with lever thickness but decreases
with cantilever length.
AFM uses the tip-sample force interaction to generate an image consisting of many rows or
lines of information placed one above the other. 5
AFM tips are not infinitely sharp and so every image is a convolution of the actual surface
topography. The image obtained also depends upon the shape of the tip and the interaction potential.
29
When imaging small objects such as DNA (single molecules are
2 nm in diameter) an effect
known as probe broadening can occur. Only the hemispherical region at the tip apex touches the
molecule, and the width of the molecule is overestimated even with very accurate calibration of the
instrument.
3.1.3 Modes of Operation

An AFM has three modes of operation; contact mode, alternate-contact (AC) mode and non-contact
mode. However non contact mode [fig. 3.2(b)] was not used in this work. In contact mode, as the
name suggests, the tip and sample remain in close contact i.e. in the repulsive regime of the intermolecular force curve in fig. 3.2 as the scanning proceeds. 1
In AC or tapping mode [fig. 3.2(c)] the tip operates near its resonant frequency (
100 kHz)
at the end of the cantilever, deriving topographic images from measurement of the attractive or
repulsive forces between the tip and sample, at an intermittent distance (fig. 3.4). There is still
contact as defined above [fig. 3.2(a)], but the cantilever taps the surface for only a very small fraction
of its oscillation period.
3.1.4 Optical Detection

In contact mode as a raster-scan drags the tip over the sample, the vertical deflection of the cantilever
is measured, this indicates the local sample height. Vertical deflection is measured using an optical
lever. The optical lever operates by reflecting a laser beam off the cantilever. Angular deflection
of the cantilever causes an approximately twofold larger angular deflection of the laser beam. The
reflected laser beam strikes a position-sensitive photodetector consisting of photodiodes (fig. 3.5).
The difference between the photodiode signals indicates the position of the laser spot on the detector,
and thus the angular deflection of the cantilever.
3.1.5 Feedback
AFMs use feedback to regulate the force on the sample. The AFM not only measures the force on
the sample but also regulates it, allowing acquisition of images at very low forces.
30
Figure 3.4: Tip engagement procedure in tapping mode. It is first necessary to tune the cantilever
to its resonance frequency and to assess a setpoint amplitude smaller than the free resonance amplitude. The engagement procedure is complete when the oscillation amplitude of the cantilever has
decreased to the setpoint amplitude. (Kislon Votchovsky)
Figure 3.5: Laser deflection off the cantilever striking the photodetector
31
In contact mode the feedback loop consists of a tube scanner which controls the height of the
entire sample, the cantilever and optical lever, which measures the local height of the sample, and a
feedback circuit that attempts to keep the cantilever deflection constant. In AC mode the feedback
system keeps the amplitude of the tip constant, rather than its deflection. However, both systems
work by adjusting the voltage applied to a piezo, a tube-shaped material that expands or contracts in
the presence of an electric field or conversely, creates a potential difference when forced to expand
or contract. 6 AFM feedback loops tend to have a bandwidth of about 10 kHz, resulting in image
acquisition times of about 1 min. 5
For this work a Dimension 3100 scanning probe microscope system (Veeco Instruments) with
the NanoScope IV controller was used. The scanning head was operated using the NanoScope v 6.2
r1 software and AC images were taken using OMCL - AC 240 TS cantilevers from Olympus. AFM
was used in this work to image and measure the length and diameter of CNx MWNTs before and
after acid treatment, and to characterize the CNx MWNT/protein conjugates. Electrical transport
measurements were also taken using an AFM method.
3.2 UV-vis Spectroscopy

The absorption of light by a molecule results in the conversion of radiant energy to the energy of
rotation, vibration or altered electronic configuration. 7 The visible region of the spectrum comprises
photon wavelengths between
300 - 800 nm. Photons within this range possess sufficient energy
to promote or excite a molecular electron to a higher energy orbital. This molecule is known as
excited and has an energy Ee , whereas the ground state molecule has the energy Eg . The energy of
the radiation absorbed must equal the difference between these two energies:
Ee Eg
(3.7)
Where h is Plancks constant and is the frequency of the photon. When sample molecules
are exposed to light which has an energy that matches a possible electronic transition within the
molecule, some of the light energy will be absorbed as the electron is promoted to a higher energy
32
Figure 3.6: UV-vis spectrum of CNx MWNTs

orbital. An optical spectrophotometer records the wavelengths at which absorption occurs, together
with the degree of absorption at each wavelength. The resulting spectrum is presented as a graph of
absorbance (A) versus wavelength ().
The energetically most favorable excitation occurs from the highest occupied molecular orbital
(HOMO) to the lowest unoccupied molecular orbital (LUMO). The energy E required to effect an
electron promotion is given by:
c
E
(3.8)
where is the wavelength of the UV-vis spectroscopy was used in this work to ascertain the effect of adsorption to CNx MWNTs on metalloproteins by monitoring their characteristic absorbance
spectra.
3.2.1 UV-vis of CNx MWNTs

CNx MWNTs are black, and so absorb in the region of interest (300 - 800 nm) for the proteins used
in this work. The UV-vis spectrum of CNx MWNTs (fig. 3.6) is smooth and featureless, similar to
that of un-doped CNTs. 8,9
All spectra were taken using a Heios spectrophotometer (Thermo Electron Corporation) and
corrected for the contribution of CNx MWNTs.
33
Figure 3.7: The amide bond (N-C=O) and nearest C neighbours (C and C ). The arrows indicate
possible bond rotations
3.3 Circular Dichroism Spectroscopy

Plane polarised light can be regarded as the sum of left and right circularly polarised light. When
plane polarised light passes through a solution containing an optically active substance, the left and
right circularly polarised components of the plane polarised light are absorbed by different amounts.
When these components are recombined they appear as elliptically polarised light.
The secondary structures of proteins used in this work each impart a distinct CD signal. Therefore the alpha helix (-helix) of cytochrome c has a different CD spectral signature to the beta turn
(-turn) of azurin. Proteins exhibit a CD signal because their building blocks (peptides) contain a
chiral bond known as an amide bond shown in fig. 3.7. The carbon (C ) is bonded to both the
carbonyl (C=O) and amino (-NH) functional groups. CD spectra were recorded on a J-710 spectropolarimeter (JASCO) with the parameters: 1 mm path length, 0.5 nm resolution, 20 nm/min scan
rate, 1 s response time, 1 nm bandwidth, at 25 C. CNx MWNTs were found not to exhibit a CD
spectrum in the region of interest, although recently un-doped SWNTs have been shown to exhibit a
CD spectrum when functionalised with DNA. 10 All spectra given in this thesis have been smoothed
and corrected for the presence of phosphate buffer.
The units of CD here are the mean residue ellipticity (degrees). In order to convert the
measured ellipticiy (obs ) into the correct units (degcm2 /dmol), the concentration (c) of protein,
calculated from UV-vis spectra, is required.
34
Figure 3.8: Phase diagram showing the process of lyophilization
obs
10 c l N
(3.9)
where l is the length of the liquid cell and N is the number of amino acid residues in the protein.
3.4 Lyophilization
Lyophilization (or freeze-drying) is a process which extracts the water from biological and other
molecules so that they remain stable and are easier to store at room temperature. Lyophilization
is carried out using sublimation which is the transition of a substance from the solid to the vapour
phase, without first passing through the intermediate liquid phase. Under a vacuum, the ice is
sublimed directly into water vapour which is then drawn off. A cold condenser chamber and/or
condenser plates provide a surface for the vapour to re-solidify on. The surface must be colder than
the temperature of the surface of the material being dried, or the vapour will not migrate to the
collector.
In a typical phase diagram (fig. 3.8), the boundary between gas and liquid runs from the triple
point to the critical point. Lyophilization (blue arrow) brings the system around the triple point,
avoiding the direct liquid-gas transition seen in ordinary drying (green arrow). A VirTis 2K Lyophilizer
was used in this work to obtain solid CNx MWNTs from solution, after acid oxidation.
35
References
[1] V. J. Morris, A. J. Kirby, and A. P. Gunning. Atomic Force Microscopy for Biologists. Imperial
College Press, 1st edition, 1999.
[2] N. Israelachivli. Intermolecular and Surface Forces. Academic Press, 2nd edition, 1992.
[3] F. London. The general theory of molecular forces. Transactions of the Faraday Society, pages
826, 1937.
[4] Z. Chen, K. Kobashi, U. Rauwald, R. Booker, H. Fan, W.-F. Hwang, and J. M. Tour. Soluble
ultra-short single-walled carbon nanotubes. J. Am. Chem. Soc., 128:1056810571, 2006.
[5] D. Baselt. The tip-sample interaction in atomic force microscopy and its implications for
biological applications. PhD, California Institute of Technology, 1993.
[6] J. A. Gallego-Juarez. Piezoelectric ceramics and ultrasonic transducers. J. Phys. E: Sci. Instrum., 22:804816, 1989.
[7] A. P. Demchenko. Ultraviolet Spectroscopy of Proteins. Springer-Verlag, Kiev, 1986.
[8] K. D. Ausman, R. Piner, O. Lourie, R. S. Ruoff, and M. Korobov. Organic solvent dispersions
of single-walled carbon nanotubes: Toward solutions of pristine nanotubes. J. Phys. Chem. B.,
104:89118915, 2000.
[9] O. Jost, A. A. Gorbunov, W. Pompe, T. Pichler, R. Friedlein, M. Knupfer, M. Reibold,
H.-D. Bauer, L. Dunsch, M. S. Golden, and J. Fink. Diameter grouping in bulk samples
of single-walled carbon nanotubes from optical absorption spectroscopy. Appl. Phys. Lett.,
75(15):22172219, 1999.
[10] G. Dukovic, M. Balaz, P. Doak, N. Berova, M. Zheng, R. S. McLean, and L. Brus. Racemic
single-walled carbon nanotubes exhibit circular dichroism when warpped with DNA. J. Am.
Chem. Soc., 128:90049005, 2006.
36
Chapter 4
Electrical Conductance of CNx MWNTs

4.1 Electrical Transport in un-doped CNTs
4.1.1 Quantized Conductance
The motion of electrons around the circumference of a CNT is quantized. This phenomenon occurs
because in quantum theory electrons can be viewed as waves. If the electron wavelength (e ) is
not a multiple of the circumference of the tube, it will destructively interfere with itself. Therefore
only values of e that are integer multiples of the circumference of the tubes are allowed, severely
limiting the number of energy states available for conduction around the cylinder. The dominant
remaining conduction path is along the axis of the tube making CNTs function as one-dimensional
quantum wires. 1
Because of the circumferential quantization, the electronic states in the CNTs do not form a
single wide-energy band. Instead they split into a number of equally spaced, one-dimensional subbands. 2 In a perfect metallic nanotube two 1D subbands carry the electric current. 3
From the Landauer formulism we expect the conductance of a nanotube to be determined by the
number of conducting channels at E f . 4
Ti
i 1
2e2
h
37
TiG0
i 1
(4.1)
where G0 = 2e2 /h = 1/(12.9 k) and the sum runs over the M available quantum channels.
Ti are the transmission factors of each channel with 0
Ti
1. For the important (n,n) armchair
nanotubes which are metallic, there are two pz derived bands at E f one even and one odd under the
n axial mirror symmetries of the tube. As a result, one expects a single armchair SWNT to have a
conductance of Te To G0 where Te and To are the transmission coefficients of the odd and even
bands respectively. If the transmission coefficients are unity, we then get a conductance of 2 G0 at
Ef.
4.1.2 SWNTs
Individual SWNTs between two electrodes can behave as molecular quantum wires 5 showing ballistic conductance. A ballistic conductor is one in which an electron injected at one end will emerge at
the other with certainty. Therefore, there is no backscattering of electrons, which yields an intrinsic
electric resistance. 6 Observable at low temperature, sections of SWNT may act as small islands with
very low capacitance (1018 F). These islands have charging energy sufficient to produce a Coulomb
blockade effect. Conductance peaks are then observed as gate voltage (Vg ) is varied, giving a field
effect transistor (FET). 7
4.1.3 MWNTs
Ballistic electrical transport has also been observed in MWNTs, 8 but some debate exists, 9 with
Schonenberger et al. 10 labelling their MWNTs quasi-ballistic.
These discrepancies arise because most investigated MWNTs cannot be described as perfect
sets of coaxial cylinders, but exhibit defects which can affect the transport mechanisms. 11 Electron
microscopy investigations show that MWNTs generally consist of shells of different helicity, such
that one atomic layer can be a metal and another a semiconductor. 12 From section 2.5.2, depending
on diameter, we can predict that
1
3
of the shells in a MWNT are metallic.
Langer et al. 13 claimed that the electric current flows in the outermost (metallic) nanotube wall
only, even if successive inner layers are metallic. However the results of Bourlon et al. 14 revealed
a substantial contribution from conduction in the second shell through tunelling. The ratio between
38
the contribution of the outermost shell and the (possible) contributions of internal shells can be
described by the ratio: 12
Rext
Rt
(4.2)
where Rext is the resistance of the external shell and Rt is the resistance connecting this shell
to internal shells ( 10 km1 ). 14 is expected to be temperature dependent, but for metallic
nanotubes
1.
Observation of 1 G0
Frank et al. 8 observed a conductance close to the quantized value 1 G0 for MWNTs. This is half the
expected value of 2 G0 , and several explanations have been proposed to rationalize this observation.
The first possibility is that in MWNTs an additional effect of inter-tube interaction appears. The
hopping of electrons between the tubes leads to a further splitting of otherwise degenerate energy
levels, the separation of which is governed by the probability of electron tunneling between adjacent
tubes. A metallic nanotube has two extended electron bands crossing E f , but when moving away
from E f more bands are able to contribute to the current, an effect that gives a corresponding increase in G. In reality, since a nanotube is never perfect, the propagating electrons will be scattered
by lattice defects, phonons or contacts. 15,16 Delaney et al. 4 considered the effect of inter-tube interactions at E f . These interactions perturb the band structure of the nanotubes significantly and cause
existing even and odd bands to repel each other, creating a depletion of states at E f . In graphite,
similar inter-layer interactions reduce the local density of states at E f by a factor of 12 . As long as
this repulsion is greater than the small bias voltages used in the experiment, these other states will
not be available for transport.
A second possibility, revealed by ab initio calculations is that of scattering. For the two contributing subbands at E f for an (n, n) CNT, one of the channels is (bonding) and the other
(antibonding). In addition, whilst the channel carries electrons through the nanotube without
scattering, the channel experiences resonant scattering by coupling to the metal contact. As a
39
result the conductance of the channel becomes negligible. 17

These scattering and inter-tube effects reduce the transmission (equation 4.1) at E f by one unit
of conductance, resulting in a measurement of 1 G0 for an individual MWNT. 8,18,19
4.2 Electronic Transport in CNx MWNTs

Nitrogen is a natural choice for a dopant in CNTs as it has a similar atomic radius as carbon (rN =
75 pm c.f rC = 77 pm) while possessing one electron more. Since the preparation of nitrogen-doped
nanotubes, various theoretical studies have predicted they will be metallic independent of tubule
diameter or chirality. 2023 This is highly desirable because at present, it is not possible to synthesize
un-doped CNTs with the same, pre-determined chirality, and therefore the same, known, electronic
properties.
4.2.1 Density of States

The effect of substitutional N on the local density of states of CNx MWNTs at E f has been determined theoretically. Tunneling spectra taken of the CNx MWNTs, were converted to the equivalent
LDOS using the accepted Feenstra algorithm of numerical differentiation and normalisation for tip
height: 20
dI
dV
(4.3)
Fig. 4.1 shows the calculated local density of states (LDOS) for CNx MWNTs (black) for both
armchair (a) and zigzag (b) tubes. The arrow indicates an additional donor feature occuring at
0.8 eV. By comparing these data with un-doped CNTs (red) it can be seen that nitrogen-doping is
responsible for the prominent donor like features close to E f . As a direct result of doping with N,
the semiconducting (zigzag) nanotube becomes metallic. This occurs because two frontier bands
overlap and become partially filled, producing a new band with zero band gap. 20
Yang et al. also determined the electronic structure of nitrogen-doped nanotubes and found that
the density of states at E f decreased with increasing concentration of pyridinic nitrogen. 24
40
Figure 4.1: Theoretical LDOS for (a) an armchair and (b) a zigzag CNT. The LDOS of the un-doped
CNT is shown in red, and the LDOS for the CNx CNT is shown in black. Image is taken from Czerw
et al. 20
4.2.2 Previous Work
The tunneling spectroscopy measurements and theoretical calculations of Czerw et al. 20 suggest
that CNx MWNTs are metallic. In addition Golberg et al. 25 reported resistivities of 10-100 k for
ropes of CNx MWNTs. An individual CNx MWNT FET has also been reported, 26 which had a conductance of
0.01 G0 for zero voltage gate bias. This conductance is much lower than the intrinsic
value of 2 G0 at E f predicted by theory for a (10,10) metallic SWNT doped with nitrogen. 11 Hence
experimental measurements of the intrinsic conductance of individual CNx MWNTs, which are not
limited by high contact resistances, are required in order to gain an understanding of their electrical
transport properties and potential applications. Furthermore, un-doped MWNTs are known to have
an extremely high power threshold for breakdown, 27 and can withstand currents up to 0.5 mA, 28
but the effect of N-doping on these properties is unknown.
41
Figure 4.2: Experimental setup from Brown et al. 3
4.3 Experimental
Interfacing a single CNT to a conducting substrate, in order to take conductance measurements
is normally hampered by very large contact resistances, caused by scattering. By measuring the
properties of individual CNT protruding from a larger flake and using a graphite surface (with
similar crystal structure and lattice spacing) we can approximate an ideal system. 3 In addition,
as the CNTs are free-standing, the possibility of influence or current leakage from the underlying
substrate is reduced. 29
The flakes of CNx MWNTs used are densely packed and typically 100 m in diameter and up
to 2 mm long. The CNx MWNTs show extraordinary alignment [fig. 4.3(b)] and narrower bundles
protrude from the flake which in turn taper towards individual tubes at their ends [fig. 4.3(c)].
4.3.1 AFM Set-up

Using a modified version of the setup of Brown et al. 3 (fig. 4.2) a flake of CNx MWNTs was placed
between the arms of a Wollaston wire thermal probe (Veeco). The flake was clamped into place
with the aid of a micro-manipulator stage.
In a typical experiment, the z-piezo of the AFM was used to move the flake towards the HOPG
42
Figure 4.3: (a) AC-AFM image of an individual CNx MWNT (Sonia Antoranz Contera) (b) SEM
image of the flake between the arms of the probe (Jian Li) (c) SEM image of an individual CNx
MWNT protruding from the flake (Jian Li) (d) circuit diagram of the set-up used in the experiment.
43
(Highly Ordered Pyrolytic Graphite) until a stable mechanical contact was realized. The force
applied to the probe was kept constant during subsequent electrical measurements. The probe was
held at system ground whilst an HP 3245A Universal Source was used to apply a voltage (VDC )
across the flake and a 1 k resistor in series [fig. 4.3(d)]. The resistor was used to limit the current
across the CNx MWNT. The voltage drop across it (VNT ) and the voltage drop across the resistor
(VR ) were measured simultaneously using Fluke 8842A multimeters, allowing the conductance G
to be calculated. Measurements were conducted on the individual CNx MWNTs protruding from
the end of the flake [fig. 4.3(c)]. Several different flakes were used, which were regularly re-oriented
to expose new individual CNx MWNTs.
4.4 Results
4.4.1 Annealing
Fig. 4.4 is a conductance versus time plot showing mechanical contact at 40 s, after which a conductance of 1 G0 was measured for an applied voltage of 1.0 V (VDC ). Between 60 and 210 s, VDC was
swept between -3V and +3V whilst the conductance was measured, shown in the G-V plot (inset).
In the first part of the voltage sweep, from 0 - 2.6 V (VNT ), the slope of the G-V curve was
steeper and the conductance values smaller than in subsequent sweeps. We have attributed this to
an annealing process at the contact, and hence a reduction in the contact resistance. This was
probably achieved by a cleaning procedure 30 involving the burning away of poorly conducting
material (e.g. water or trapped hydrocarbons) between the CNx MWNT and the HOPG or a change
in the contact geometry. 19 After the initial voltage sweep, a reproducible G-V curve was observed.
For times between 210 and 240 s [fig. 4.4(a)] a constant potential (VDC ) of 1.0 V was applied and a
post-annealing contact of 1.2 G0 was measured.
4.4.2 Conductance Values

Conductance values for CNx MWNTs at low bias (VNT
0.1 V) were measured for 77 new and
different contacts made between the CNx MWNT flake and the HOPG. Statistical analysis of the
44
Figure 4.4: (a) Conductance measurements made during the initial approach of the flake, followed
by a stable mechanical contact between the CNx MWNT and the HOPG. Inset shows the annealing
of the CNx MWNT/HOPG contact during the initial voltage sweep. (b) Statistical analysis of all
77 contacts made after annealing at VNT 2.0 V. The histogram shows that 70% of the contacts
were in the range 0 - 2 G0 . (c) Histogram for conductance values between 0 and 2 G0 only.
45
results, [figs. 4.4(b) and (c)], indicates two different systems; individual CNx MWNTs with conductance values close to 1 G0 , and bundles of CNx MWNTs exhibiting larger values of conductance,
between 2 and 17 G0 . For those contacts (
70%) giving a conductance between 0 and 2 G0
[fig. 4.4(c)] a mean value of 1.0 0.3 G0 has been calculated.

A conductance of 1 G0 is half the expected contribution for a CNx SWNT 11 or for a CNx
MWNT, in which only the outermost metallic shell conducts. Explanations for the observation of 1
quantum of conductance rather than 2, are detailed in section 4.1.3, however scattering at physical
defects and N-doping sites may also reduce the conductance in the present work.
The absence of any low conductances (
0.3 G0 ) in our measurements suggests that all the
CNx MWNTs show high conductances and that the contact resistance is low. The high conductances could either indicate that the CNx MWNTs are intrinsically metallic or that they are doped
semiconductors with narrow or vanishing bandgaps. High conductance throughout an entire sample of MWNTs is potentially advantageous for many applications such as field emission devices,
biosensors and nanoelectronic components including interconnects.
The low conductances measured in the work of others 25,26 may be due to physical and chemical
differences in the raw CNx MWNTs, damage to the sample caused by processing, or high contact
resistances due to Schottky barriers or poor physical contact at the NT-electrode interfaces.
4.4.3 G-V and I-V curves

Fig. 4.5 (a) and (b) show typical G-V and I-V curves for an individual CNx MWNT after annealing.
The corresponding G-V and I-V curves for a bundle of CNx MWNTs are shown in figs. 4.5 (c) and
(d). The conductance increases linearly with voltage and the measurements were found to remain
stable over several hours. The V-shaped G-V characteristics recorded are similar to those reported
for un-doped MWNTs. 30,31 For those plots with a conductance of 1.0 G0 at low bias, the average
gradient is 0.7
0.2 G0/V which is larger than previous measurements for un-doped MWNTs of
0.27 G0 /V 31 and 0.3 G0 /V. 18 The difference probably indicates that in our CNx MWNTs, additional
subbands become available at high voltages. 18
46
Figure 4.5: (a) G-V curve and (b) corresponding I-V curve for an individual CNx MWNT, (c) G-V
curve and (d) corresponding I-V curve for a bundle of CNx MWNTs.
4.4.4 Electrical Breakdown
Individual CNx MWNT contacts were then subjected to electrical breakdown by increasing VDC
gradually to a maximum of 3 V. Fig. 4.6 (a) shows the breakdown of an individual CNx MWNT
which begins with a steady decline in current down to 60 A, followed by a series of sharp steps
probably associated with the removal of individual layers. 32 Fig. 4.6 (b) shows both increasing and
decreasing current steps, similar to previous observations for un-doped MWNTs, 33 which were
assigned to processes involving the breakdown of inner layers or a combination of inner and outer
layers. The more common observation of decreasing steps only [fig. 4.6(c)] has been associated with
the breakdown of outer layers only. For the 20 electrical breakdown measurements of CNx MWNTs
made, 8 resulted in steps after significant electrical breakdown of the tube (G = 0.3 G0 ). Step sizes
were in the range 6-40 A with an average size of 20 10 A. The irregular step sizes presented
in this work may be attributed to varying degrees of N-doping in different layers, differences in
inter-layer coupling or the breakdown of several layers simultaneously.
47
Figure 4.6: Electrical breakdown of individual CNx MWNTs showing; (a) gradual decline in current
followed by discrete steps (b) both increasing and decreasing current step and (c) another example
of steps during breakdown. The numbers adjacent to the arrows indicate the magnitudes of the steps
in A.
Power Threshold values
The threshold power (Pthresh ), the highest power reached during the breakdown, was also calculated
and found to be 490
100 W. This is larger than the value of 300 W previously reported for
un-doped MWNTs in air. 32
48
References
[1] C. P. Poole Jr. and F. J. Owens. Carbon Nanotubes. John Wiley and Sons Inc., 2003.
[2] S. Sapmaz, P. Jarillo-Herrero, J. Kong, C. Dekker, L. P. Kouwenhoven, and H. S. J. Van der
Zwan. Electronic excitation of metallic carbon nanotubes. Phys. Rev. B, 71:153402, 2005.
2005.
[4] P. Delaney, M. Di Ventra, and S. T. Pantelides. Quantized conductance of multiwalled carbon
nanotubes. Appl. Phys. Lett., 75(24):37873789, 1999.
[5] R. Saito, M. S. Dresselhaus, and G. Dresselhaus. Physical Properties of Carbon Nanotubes.
Imperial College Press, 1998.
[6] http://www.nd.edu/ techrev/Archive/Winter2002/a6.html. Carbon nanotubes, wunderkind materials, 2002.
[7] V. Skakalova, A. B. Kaiser, Y.-S. Woo, and S. Roth. Electronic transport in carbon nanotubes:
From individual nanotubes to thin and thick networks. Phys. Rev. B, 74:085403, 2006.
[8] S. Frank, P. Poncharal, Z. L. Wang, and W. A. de Heer. Carbon nanotube quantum resistors.
Science, 280:17441746, 1998.
[9] M. Zamkov, N. Woody, B. Shan, Z. Chang, and P. Richard. Lifetime of charge carriers in
multiwalled nanotubes. Phys. Rev. Lett., 94:056803, 2005.
[10] C. Schonenberger, A. Bachtold, C. Strunk, J.-P. Salvetat, and L. Forro. Interference and interaction in multi-wall carbon nanotubes. Appl. Phys. A, 69:283295, 1999.
[12] A. Y. Kasumov, H. Bouchiat, B. Reulet, O. Stephan, I. I. Khodos, Y. B. Gorbatov, and C. Colliex. Conductivity and atomic structure of isolated multiwalled carbon nanotubes. Europhys.
Lett., 43(1):88994, 1998.
[13] L. Langer, E. Bayot, E. Grivei, J.-P. Issi, J. P. Heremans, C. H. Olk, L. Stockman, C. van
Haesendonck, and Y. Bruynseraede. Quantum transport in a multiwalled carbon nanotube.
Phys. Rev. Lett., 76(3):479482, 1996.
[14] B. Bourlon, C. Miko, L. Forro, D. C. Glattli, and A. Bachtold. Determination of the intershell
conductance in the multiwalled carbon nanotubes. Phys. Rev. Lett., 93(17):176806, 2004.
[15] M. P. Anantram and T. R. Govindan. Conductance of carbon nanotubes with disorder: A
numerical study. Phys. Rev. B., 58(8):48824887, 1998.
[16] V. H. Crespi, M. L. Cohen, and A. Rubio. In situ band gap engineering of carbon nanotubes.
Phys. Rev. Lett., 79(11):20932096, 1997.
49
[17] H. J. Choi, J. Ihm, Y.-G. Yoon, and S. G. Louie. Possible explanation for the conductance of a
single quantum unit in metallic carbon nanotubes. Phys. Rev. B, 60(20):1400914011, 1999.
[18] P. Poncharal, C. Berger, Y. Yi, L. Wang, and W. A. de Heer. Room temperature ballistic
conduction in carbon nanotubes. J. Phys. Chem. B., 106:1210412118, 2002.
[19] C. Berger, Y. Yi, J. Gezo, P. Poncharal, and W. A. de Heer. Contacts, non-linear transport
effects and failure in multi-walled carbon nanotubes. New J. Phys., 5:158.1, 2003.
[21] Y. Huang, J. Gao, and R. Liu. Structure and electronic properties of nitrogen-containing carbon
nanotubes. Synth. Met., 113:251255, 2000.
[22] R. Sen, B. C. Satishkumar, A. Govindaraj, K. R. Harikumar, G. Raina, J.-P. Zhang, A. K.
Cheetham, and C. N. R. Rao. B-C-N, C-N and B-N nanotubes produced by the pyrolysis of
precursor molecules over Co catalysts. Chem. Phys. Lett., 287:671676, 1998.
[23] S. H. Lim, H. I. Elim, X. Y. Gao, A. T. S. Wee, W. Ji, J. Y. Lee, and J. Lin. Electronic and
optical properties of nitrogen-doped multiwall carbon nanotubes. Phys. Rev. B, 73:045402,
2006.
[24] J. H. Yang, B. J. Kim, Y. H. Kim, Y. J. Lee, B. H. Ha, Y. S. Shin, S.-Y. Park, H. S. Kim, C.-Y.
Park, C.-W. Yang, J.-B. Yoo, M. H. Kwon, K. Ihm, H.-J. Song, T.-H. Kang, H.-J. Shin, J.-Y.
Park, and J.-M. Kim. Nitrogen-incorporated multiwalled carbon nanotubes grown by direct
current plasma-enhanced chemical vapor deposition. J. Vac. Sci. Technol. B, 23(3):930933,
2005.
2005.
[27] P. G. Collins, M. S. Arnold, and P. Avouris. Engineering carbon nanotubes and nanotube
circuits using electrical breakdown. Science, 292:706, 2001.
[28] B. Bourlon, D. C. Glattli, B. Placais, J. M. Berroir, C. Miko, L. Forro, and A. Bachtold.
Geometrical dependence of high-bias current in multiwalled carbon nanotubes. Phys. Rev.
Lett., 92(2):026804, 2004.
[29] P. Dorozhkin, H. Nejoh, and D. Fujita. Low-energy electron point source microscope as a tool
for transport measurements of free-standing nanometer-scale objects: Application to carbon
nanotubes. J. Vac. Sci. Technol. B, 20(3):10441047, 2002.
50
[30] T. D. Yuzvinsky, W. Mickleson, S. Aloni, S. L. Konsek, A. M. Fennimore, G. E. Begtrup,

A Kis, B. C. Regan, and A. Zettl. Imaging the life story of nanotube devices. Appl. Phys.
Lett., 87:083103, 2005.
[31] M. Tsutsui, S. Mitsuya, S. Kurokawa, and A. Sakai. Conductance versus bias voltage characteristics of multi-walled carbon nanotubes. Nanotechnology, 16:18631867, 2005.
[32] P. G. Collins, M. Hersam, M. Arnold, R. Martel, and Ph. Avouris. Current saturation and
electrical breakdown im multiwalled carbon nanotubes. Phys. Rev. Lett., 86(14):31283131,
2001.
[33] J. Y. Huang, S. Chen, S. H. Jo, Z. Wang, D. X. Han, G. Chen, M. S. Dresselhaus, and Z. F. Ren.
Atomic-scale imaging of wall-by-wall breakdown and concurrent transport measurements in
multiwall carbon nanotubes. Phys. Rev. Lett., 94:236802, 2005.
51
Chapter 5
Biofunctionalisation of CNx MWNTs

Carbon nanotubes are a useful support material for the immobilisation of biomolecules because of
their strength, stability, and electrical conductivity. 1 A wide range of biological molecules have
been adsorbed onto CNTs, 2 as the first step towards the development of biosensors, 3 drug delivery
systems 4 and anti-cancer agents. 5
5.1 Previous Work

5.1.1 Biofunctionalisation of un-doped CNTs with Proteins
Proteins are naturally occurring nanosystems endowed with diverse properties and functions. 6 Much
progress has been made in the functionalisation of CNTs with proteins, which adsorb individually,
strongly and noncovalently along the lengths of CNTs. 7 Proteins including streptavidin, 810 ferritin, 3,11,12 cytochrome c, 13,14 and bovine serum albumin (BSA) 15,16 and polysaccharides such as
starch, 17,18 have all been successfully immobilised onto un-doped CNTs. Antibodies have also been
bound independently 19 and as part of a sensing process. 20,21
In addition to sidewall functionalisation, CNTs have a prominent central cavity into which small
proteins have been introduced, thus forming natural nano test tubes. 1,2224
52
5.1.2 Functionalisation of Boron Nitride NTs with Proteins

BN nanotubes have been functionalised with ferritin, cytochrome c, streptavidin and glucose oxidase without coupling agents. 25 However the immobilisation process was extremely slow, with 120
h stirring time required to obtain successful functionalisation of the NTs.
5.1.3 Functionalisation of CNx MWNTs with Proteins

CNx MWNTs have several distinct advantages over un-doped CNTs for bioapplications; they exhibit high conductances independent of chirality, 26 they are less toxic, 27 and their large-diameter
structure is rich in defects, particularly hydrophilic nitrogen sites. 28 Similar nitrogen groups have
been shown to facilitate electron transfer between metalloproteins and Au electrodes. 29,30 Despite
these advantages, examples of the integration of CNx MWNTs with biological systems are rare.
Jiang et al. have covalently immobilised ferritin and BSA on CNx MWNTs. They used a twostep process of diimide-activated amidation, to covalently attach the proteins to the sidewalls of the
tubes. 31 In this work we show that covalent linkages are not necessary for the immobilisation of
ferritin onto CNx MWNTs.
5.2 Effect on protein activity

A successful bio-CNT interface requires not only good adhesion of the biomolecule, but also conservation of its structure and functionality. 9 Previous protein functionalisation work with CNTs, has
demonstrated successful protein coverage of individual CNTs in solution, but the effect of adsorption on protein conformation and activity has only been examined for films of CNTs. 13,3236 By
using CD spectroscopy, UV-vis spectroscopy and antibody binding assays, the conformation and
activity of adsorbed proteins can be investigated.
5.2.1 Antibody Binding Assays

Antibodies are proteins synthesized by an animal in response to the presence of a foreign substance.
The basic principle of any immunochemical technique, is that a specific antibody will bind to its
53
Figure 5.1: The shape of an IgG antibody

specific antigen to give an exclusive antibody-antigen complex. This binding is mediated by the
sum of many weak interactions between the antigen and antibody. These weak interactions include
hydrogen bonds, VDWs forces, and ionic and/or hydrophobic interactions. These interactions can
only take place if the antigen and antibody molecules are close enough for some atoms of the
antibody to fit into complementary recesses on the antigen known as the epitope. 37
Structure of antibodies
Antibodies consists of four polypeptides; two heavy chains and two light chains, which join to form
a Y shaped molecule (fig. 5.1). The amino acid sequence in the tips of the Y, varies greatly among
different antibodies. This variable region, composed of 110-130 amino acids, give the antibody its
specificity for binding its antigen. Immunoglobulins (Ig) are divided into five major classes, IgM,
IgG, IgA, IgD and IgE, based on their structure and immune function. 37 In this work, the antibodies
used are IgG immunoglobins.
5.3 Experimental
In this work CNx MWNTs were functionalised with metalloproteins (proteins containing at least
one metal ion, usually a transition metal with a variety of accessible oxidation states) using a noncovalent scheme. Both raw and acid-treated (see Chapter 7) CNx MWNTs were functionalised
54
with the Fe-containing proteins ferritin, ferredoxin and cytochrome c and the Cu-containing protein
azurin. It has been proposed that the acid groups generated during oxidation may assist in the
anchoring of biological species to the CNT. 22,23,34 In the following work this proposal is addressed
systematically for CNx MWNTs.
5.3.1 Functionalisation with metalloproteins

Metalloproteins were chosen for this work because much is already known about their structure
and functionality. 38 Ferredoxin from Spinacia Oleracea, azurin from Pseudomonas aeruginosa and
cytochrome c (cyt c) from rabbit heart were incubated as nM solutions in 5 mM phosphate buffer
(pH 6.01 containing 5 mM NaH2 PO4 and 5 mM NaCl) with the CNx MWNTs for 2 h in an ice bath
with magnetic stirring. In the case of ferritin from horse spleen, the functionalisation scheme of Lin
et al. 14 was followed.
Commercially available anti-cyt c (produced in rabbit) and anti-ferritin (produced in sheep)
were used in the antibody binding assays. After the 2 h incubation period described above, unadsorbed protein was removed by centrifugation for 10 min at 18,000 g and the pellet re-suspended
in buffer. AC-AFM imaging of a small sample, was used to verify that all unadsorbed protein had
been removed. Antibodies were then diluted 20 times with MilliQ water, added to the protein/CNx
MWNT solution and incubated with stirring for a further 1 h in an ice bath. All proteins and
antibodies were purchased from Sigma.
The CNx MWNT bioconjugate solutions were dispersed onto glass cover slips, dried and then
washed with MilliQ water before AC-AFM imaging in air.
5.3.2 Spectroscopy
CD and UV-vis spectra of free cyt c (concentrated cyt c in solution in the absence of CNx MWNTs)
and free azurin were taken against a 5 mM phosphate buffer background. CNx MWNTs/protein
conjugates were then added prepared as above, and any un-adsorbed protein removed. Protein
concentrations before and after this procedure were calculated from absorbance spectra. Extensive
control experiments were carried out to monitor the effect of environmental conditions such as
55
temperature and light on the incubation.

The protein molecule may be considered as complex system of chromophores 39 (a group of
atoms in a molecule that includes a conjugated bond capable of absorbing UV light). The chromopores of cyt c and azurin (fig. 5.10 and fig. 5.8 respectively) are responsible for the colours of the
proteins and their characteristic UV-vis spectra.
5.3.3 Apo proteins

Apo proteins are essentially the empty shells of metalloproteins. In the biosensing part of this
project, the metalloproteins are seen to interact with the CNx MWNTs, producing a response. A
simple way to ascertain if this response is due to the metal centre or to the external protein structure,
is to repeat the experiments with the corresponding apo proteins. Apo proteins are unfortunately
not widely commercially available, although apoferritin is an exception.
Preparation of Apoazurin
The copper atom in azurin can be eliminated to create apoazurin without change in the overall
structure of the protein. 40 Apoazurin is not available commercially but can be prepared by a series
of dialysis experiments to reduce the Cu ion and then complex it out of the protein. 41,42
In this work apoazurin was prepared using the method of Blaszak et al. 41 Native azurin was
reduced with ascorbate in MiliQ water to give azurin which has a colourless Cu centre. Azurin
solution was then dialysed against 0.05 M sodium phosphate buffer containing 0.25 M NaCl and 0.1
M thiourea (a Cu complexing agent) for at least 12 h. The complexing agent was then removed by
a second dialysis against 0.1 M phosphate buffer (adjusted to pH 7 using NaOH) for at least 12 h.
All dialyses were carried out at 5-10 C using an ice bath. The absence of azurin was determined
by UV-vis spectroscopy.
Crystallization of Apoferritin
Apoferritin readily forms crystals in sodium acetate solution containing certain amounts of cadmium
sulphate. 43 Apoferritin was prepared in 0.2 M sodium acetate (NaOOCH3 ) buffer, with pH 5.0
56
Figure 5.2: Ferredoxin

(adjusted using glacial acetic acid). Crystallization using CdSO4 commenced from a 1 mM solution
at room temperature on a glass slide. The crystals were imaged in liquid by AFM.
5.4 Results
5.4.1 Ferredoxin
Ferredoxin is a small (12 kDa) water-soluble protein containing at least one ion-sulphur cluster. 37
It is acidic, and found in various organisms where it acts as a multifunctional electron carrier. In
particular, ferredoxin (fig. 5.2) is a carrier molecule in the electron transport chain in photosynthesis,
passing electrons to NADP reductase. It also has a role in nitrogen fixation. 44
Structure
The iron-sulphur clusters (brown) contain one or more Fe ions usually chelated by cysteine residues
(see section 10.5 for amino acid structures), whose S atoms coordinate to the metal. The clusters
can be of the type [4Fe - 4S], [3Fe - 3S] or [2Fe - 4S] with a charge from 0
+3.
Fig. 5.3 shows individual raw [(a) and (b)] and acid-treated [(c) and (d)] CNx MWNTs function57
Figure 5.3: AC-AFM images of individual raw [(a) and (c)] and acid-treated [(b) and (d)] CNx
MWNTs functionalised with ferredoxin
58
alised with ferredoxin. In these AFM images, the individual CNx MWNTs are distinguishable from
the globular structures which represent the ferredoxin molecules. For the raw tubes [fig. 5.3(a) and
(b)] coverage is poor, with only a few proteins attached to the surface. The ferredoxin molecules
bind to the raw CNx MWNT in the form of large clusters
100 - 200 nm in height [indicated by
arrows in (a)]. In comparison, the acid-treated tubes [fig. 5.3(c) and (d)] are more evenly covered in
smaller ferredoxin clusters
20 nm in height. The large clusters of ferredoxin indicate a preference
for adsorption to neighbouring (hydrophilic) proteins over adsorption to the mainly hydrophobic
surface of the raw CNx MWNT. The areas where ferredoxin does bind on the raw CNx MWNT
may indicate the location of pyridinic nitrogen sites on the CNx MWNT surface. The activity of
adsorbed ferredoxin could not be investigated by spectroscopy as it does not exhibit a characteristic
UV-vis spectrum. In addition, anti-ferredoxin is not available commercially.
The improved adsorption of ferredoxin to acid-treated CNx MWNTs is due to additional interactions between the protein, and oxygenated groups on the acid-treated CNx MWNTs 45 (see
Chapter 7). These interactions are most likely to be hydrogen bonds, but the possibility of covalent
amide linkages (section 10.5.1) between protein -NH2 (amine) moieties and -COOH (carboxylic
acid) groups on the CNx MWNTs cannot be excluded.
5.4.2 Ferritin
Ferritin (fig. 5.4) stores Fe (red) intracellularly in the liver in mammals and is also found in plants,
fungi and bacteria. Ferritins belong to the Class II diiron-carboxylate proteins and have the capacity
to remove Fe2 ions from solution in the presence of O2 , and to deposit up to 4500 Fe atoms into the
protein interior in a mineral form (Fe2 O3 .H2 O). Fe is stored as its ferric hydroxide rather than as an
ion, because the solution chemistry of Fe at physiological pH values is dominated by the tendency
to oxidise to an insoluble form. 46
Structure
Ferritin (440 kDa) consists of a spherical shell composed of 24 subunits. Each subunit is an individual molecule that joins its neighbouring subunits through noncovalent interactions. There are
59
Figure 5.4: Ferritin

two types of subunit (H and L) which occur in variable proportions and are folded into ellipsoids
approximately 12 nm in diameter. In vitro studies have shown that the H-subunit facilitates iron
oxidation, while the L-subunit expediates nucleation and core formation. 47 These subunits are four
-helices of 172 amino acid residues which are arranged to give a hollow, symmetrical shell with
an inner cavity diameter of 9 nm.
Fig. 5.5(a) shows very good coverage of ferritin molecules on an individual acid-treated CNx
MWNT. Ferritin has a very robust structure and is able to withstand biological extremes of temperature and pH. It is automatically stabilized against agglomeration in aqueous solution by functional
groups on the proteins, which allows the deposition of isolated ferritin molecules on CNTs. 48 The
conjugate in fig. 5.5 (a) has an average height of
60 nm [fig. 5.5(c)]. An antibody binding assay
was then used to demonstrate that the bound ferritin retains its immunoactivity.
The conjugate sample was washed by centrifugation to remove any un-adsorbed ferritin, and
then incubated with its specific IgG antibody; anti-ferritin. Fig. 5.5(b) shows a ferritin-CNx MWNT
conjugate which has been exposed to anti-ferritin. This new conjugate (fig. 5.6) has an average
height of
200 nm and the images show a clear difference between simple ferritin functionalisation
of the CNx MWNT, and the additional anti-ferritin binding. During control experiments, binding of
the anti-ferritin direct to CNx MWNT was insignificant compared to its binding to the ferritin-CNx
conjugate (fig. 5.7). This result has also been seen for un-doped CNTs with a peptide antibody. 21
60
Figure 5.5: AC-AFM images of (a) an acid-treated CNx MWNT functionalised with ferritin, (b)
a ferritin-CNx MWNT conjugate functionalised with anti-ferritin, (c) and (d) represent the z-range
height profiles for each image
The binding of anti-ferritin to adsorbed ferritin demonstrates that the structural characteristics of the
epitope (the part of the antigen to which the antibody specifically binds) remain unchanged when
adsorbed to CNx MWNTs. 21
5.4.3 Azurin
Azurin (14 kDa) is a bacterial protein which is believed to function as an electron transferase, 41 it
belongs to the cupredoxin family and is a 128 residue Cu containing protein, which appears vivid
blue in solution.
The physiological function of azurin is that of an electron carrier, 49 specifically as an electron
donor to nitride reductase which is a denitrifying bacteria 50 (a bacteria involved in the process of
removing nitrogen from soil). Electrons are donated from azurin to the bacteria, where they reduce
nitrite to nitrous oxide. Azurin (fig. 5.8) is widely studied as a model electron-transfer protein. 51
61
Figure 5.6: Representation of a CNT functionalised with ferritin and then IgG anti-ferritin (not to
scale)
Figure 5.7: AC-AFM image of an acid-treated CNx MWNT after incubation with anti-ferritin
62
Figure 5.8: Azurin and its chromophore (inset)
Structure
Azurin is a single-domain protein with a -turn structure composed of eight -turns, arranged in a
double-wound Greek-key topology (named after a pattern common on Greek pottery, it represents
three strands connected by hairpins, followed by a larger connection to the fourth strand, which
lies adjacent to the first). The redox-active Cu site is coordinated by cysteine, histidine and methionine ligands 51 allowing for electron transfer activity. 40 Crystal structures show that the redox centre
(fig. 5.8 inset) consists of a very irregular tetrahedral coordination with two S from methionine
(Met) and cysteine and two histidine N atoms. The bond between the Cu atom and the S of the
cysteine is unusually short (
210 pm), this is because the Cu geometry represents a compromise
between several possible conformations.

Although some other metal ions e.g. zinc 52 can bind in this site, only copper exhibits a reduction
potential in the physiologically useful range.
Fig. 5.9(a) shows an acid-treated CNx MWNT functionalised with azurin. Again functionalisation is more efficient than for a raw CNx MWNT (not shown). Azurin exhibits a characteristic CD
spectrum and fig. 5.9(b) shows both its free (i) and adsorbed (ii) CD spectra. Information on the
secondary structure of azurin can be deciphered from fig. 5.9(b)(i) by comparison with theroretical
CD reference spectra. 53 The CD spectrum of free azurin is characterized by a minimum at 219 nm
63
Figure 5.9: (a) AC-AFM image of an acid-treated CNx MWNT functionalised with azurin, (b) CD
spectra for (i) free and (ii) adsorbed azurin, (c) UV-vis spectrum for free azurin and (d) UV-vis
spectrum for adsorbed azurin
and cross-over point of 204 nm and is typical for a protein with secondary structure consisting of
80% -turn and 20% -helix. 53
After incubation with CNx MWNTs any unadsorbed azurin was removed from solution by washing and the spectrum re-taken (ii). The concentration of adsorbed protein remaining after this process was found to be much lower than the concentration of free azurin (72 M c.f 280 M) and the
reduced demonstrates that fewer protein molecules are present. However the characteristic shape
of the CD spectrum remains, indicating that adsorption to CNx MWNTs does not significantly alter
the secondary structure of azurin through denaturing or decoiling.
Azurin also absorbs in the visible with a peak at max
600 nm [fig. 5.9(c)] due to a (Cys)S
Cu2 dx2 y2 ligand-to-metal charge transfer. 54 After washing, this peak is still present [fig. 5.9(d)]
although reduced in intensity, indicating that the electronic structure of the Cu chromophore is
conserved during adsorption to CNx MWNTs.
Anti-azurin is not available commercially, therefore an antibody binding assay could not be used
to determine the effect of adsorption to CNx MWNTs on the immunoactivity of azurin.
64
Figure 5.10: Cytochrome c and its chromophore (inset)
5.4.4 Cytochrome c
Cytochromes are a group of red heme proteins that undergo reversible oxidation and reduction
through their complexed Fe atoms. Cyt c is one of the three proton-pumping assemblies of the
respiratory chain which catalyzes the reduction of O2 to H2 O via cytochrome oxidase. 37 During
redox processes, the changes in charge may be considered formally to occur at the Fe atoms, but in
reality there is considerable electron delocalisation over the whole cluster.
4cytc2 4H O2
4cytc3 2H2 O
(5.1)
The cyt c heme group (fig. 5.10 inset), is bound with tetragonal symmetry 55 to the protein by
the thioether bonds of cysteine residues.
Structure
Cyt c (12 kDa) is a water soluble protein that consists of a single polypeptide chain of 104 amino
acid residues (fig. 5.10). In its native state, cyt c contains three major helices and two helical
65
Figure 5.11: (a) AC-AFM image of an acid-treated CNx MWNT functionalised with cyt c. (b)
CD spectra for (i) free and (ii) adsorbed cyt c. (c) UV-vis spectrum for free cyt c and (d) UV-vis
spectrum for adsorbed cyt c
and
elements inter-connected by loops. 56 The protein is roughly spherical with a diameter of 34 A
folds to form a hydrophobic interior and a hydrophilic exterior.
Fig. 5.11(a) shows even coverage of cyt c on an acid-treated CNx MWNT. The height of the
CNx MWNT is
c is
20 nm and the average height of this conjugate is
40 nm. Given the size of cyt
3 nm, this indicates that cyt c also binds to the CNx MWNT in the form of aggregates.
Free cyt c also exhibits a characteristic CD spectrum, shown in fig. 5.11(b)(i). The spectrum
is characterized by two minima near 208 and 220 nm and a cross-over point of 202 nm, and is
typical for a protein with a secondary structure consisting of 95% -helix and 5% -turn. 53 The CD
spectrum of adsorbed cyt c after washing, is shown in fig. 5.11(b)(ii). It is very similar to that of the
free protein, with both minima remaining, suggesting that cyt c retains its secondary structure when
adsorbed to CNx MWNTs. This is in comparison to the CD spectrum of denatured cyt c (in urea
solution) which is reduced to a single positive peak at about 258 nm. 57 Rearrangement or loosening
of the helical structure of cyt c can be indicated by changes in the 220 208 ratio. 56,58 However the
ratio of the two minima in the free cyt c CD spectrum (i) is 1.2, and this ratio is the same in the
66
CD spectrum for adsorbed cyt c (ii) suggesting that the overall secondary structure of the protein is
retained.
Cyt c also absorbs strongly in the visible region, exhibiting characteristic peaks between 200 800 nm. These absorption bands arise because of
electronic transitions 57 within the highly
conjugated heme group (fig. 5.10 inset). Fig. 5.11(c) shows the UV-vis spectrum of concentrated
free cyt c with the bands at
410 nm (Soret band) and
530 nm ( band) that are characteristic
of cyt c in its oxidized form. 38 Fig. 5.11(d) shows the spectrum for adsorbed cyt c. The signal is
far less intense, because any free cyt c remaining in solution has been removed by centrifugation.
However the Soret and peaks are still evident, suggesting that the stable attachment 32 of cyt c to
the CNx MWNT does not lead to expulsion of the heme group.
An antibody binding assay was then used to demonstrate the continuing immunoactivity of cyt
c. The CNx MWNT-cyt c sample was washed by centrifugation to remove any un-adsorbed protein.
It was then incubated with its specific IgG antibody; anti-cyt c. Fig. 5.12(b) shows a cyt c-CNx
MWNT conjugate which has been exposed to anti-cyt c. This new conjugate has an average height
of
140 nm [fig. 5.12(d)] compared to the average height of the cyt c only conjugate of
40 nm
[fig. 5.12(c)].
During control experiments, binding of the anti-cyt c direct to CNx MWNTs was insignificant
compared to its binding to the cyt c-CNx MWNT conjugate (not shown).
The binding of anti-cyt c to adsorbed cyt c demonstrates that the structural characteristics of the
epitope of cyt c remain unchanged when adsorbed to CNx MWNTs.
Apoferritin
Apoferritin (fig. 5.13) has the same external structure as ferritin, 59 but is colourless because the red
colour in ferritin originates from electronic transitions within the Fe group. 46
Apoferritin is abundant in the -cells of the pancreas where it acts as an anti-oxidant. It is
produced in the mucosal cells of the small intestine and binds Fe to form its parent protein ferritin.
Both ferritin and apoferritin crystallize in the same structure in the presence of certain metal ions.
Apoferritin crystals (fig. 5.14) were formed by surface nucleation 60 into a face-centred cubic
67
Figure 5.12: AC-AFM images of (a) an acid-treated CNx MWNT functionalised with cyt c, (b) a
cyt c-CNx MWNT conjugate functionalised with anti-cyt c, (c) and (d) represent the z-range height
profiles for each image
Figure 5.13: Apoferritin, showing the hollow internal core
68
Figure 5.14: Liquid AC-AFM image of apoferritin crystals

crystal structure. The crystals form upon the addition of Cd2 ions, which change the repulsive
interactions to attractive. 43
Apoazurin
Apoazurin does not occur naturally but can be prepared by a series of dialyses (see section 5.3.3).
Whilst the secondary structure of apoazurin is practically indistinguishable from that of azurin, 61
observations during folding and unfolding, indicate that the Cu atom stabilizes the native structure. 52
Apoferritin and apoazurin were conjugated with CNx MWNTs (not shown) to ensure successful
adhesion. Following this, the apoproteins were used during biosensor measurements (Chapter 8) to
elucidate the contribution of metal ions to the sensing mechanism.
5.5 Discussion
5.5.1 Protein Binding
The metalloproteins undergo non-specific binding (NSB) with the CNx MWNTs, a result that
has been observed previously for un-doped CNTs. 3,62 The interactions between protein and CNx
MWNT are likely to be hydrophobic interactions, 19,63 or - interactions. 62,64 The adsorption cov69
Figure 5.15: Apoazurin

erage is dependent on the protein concentration, incubation time and the CNT suspension. The
possibility of -stacking as a feasible alternative to covalent attachment for the CNT/cyt c system
has already been proposed. It is possible that -stacking provides more freedom for the heme group
to interact with the CNT reducing the necessity for changes in the protein secondary structure. 65
Effect of acid oxidation

Functional acid groups on the CNx MWNTs introduce another interaction between protein and CNx
MWNT; hydrogen bonding. There is also the possibility of amidation of NH2 groups on the protein
with COOH groups on acid-treated CNx MWNTs, although this is unlikely; Basiuk et al. 66 heated
amines to 170 C with acid-treated un-doped SWNTs for 1 h, and failed to see evidence of amide
bands in the IR spectra of the product. Some C-N stretches were observed, but these were assigned
to physisorbed amines.
Functional acid groups also improve dispersion of CNx MWNTs in solution, which maximizes
the CNT surface area available for protein binding.
Improved dispersion of CNx MWNTs can also be achieved through the use of a surfactant
such as sodium dodecyl sulfate (SDS) (see section 7.1.1). However, it was found that the CNx
MWNTs-SDS solution was not compatible with the metalloproteins due to denaturation. SDS is an
amphipathic detergent, which binds non-covalently to proteins, with a stoichiometry of around one
70
SDS molecule per two amino acids. It causes proteins to denature and disassociate from each other
by forming a chemical bridge between hydrophobic and hydrophilic environments, thus disrupting
the hydrophobic forces needed to retain native protein structure. Detergent induced denaturation is
usually irreversible. 37
5.5.2 Changes in the conformation of cyt c

UV-vis spectroscopy is a useful tool for monitoring the possible changes in the Soret band, which
is not specific to cyt c, but also occurs in the UV-vis spectra of the other heme proteins; myoglobin
and hemoglobin. 67 The effect of adsorption on cyt c conformation could also be followed using
XPS, however DSouza et al. found that the signals corresponding to the Fe and S atoms of the
heme group were very weak, apparently indicating that the heme moiety was buried too deep inside
the polypeptide chain of cyt c to be probed using XPS. 36 In general, the unfolding of cyt c can be
pictured as the progressive breaking of hydrogen bonds in the protein interior and gradual exposure
of amino acid residues normally buried within the structure to solvent molecules. The investigation
of both the secondary structure of the protein, and the electronic environment of the heme group is
important, as Xu et al. 56 observed changes within the heme group between native and guanidine
denatured cyt c, despite apparent similarities in their secondary structures.
5.5.3 Mechanism of apoazurin preparation

The procedure for preparing apoazurin follows two main steps. The first is the reduction of Cu2
to Cu using ascorbic acid (fig. 5.16), a water-soluble reagent that is easily oxidized to docosahexaenoic acid (DHA).
During this process the azurin solution changes colour from its characteristic blue to colourless.
The colour exhibited by azurin (and cyt c) is due to the transition metal within the protein. The
electronic stucture of the metal ion when coordinated by ligands such as amino acids, contains
an energy gap E in the electronic orbitals. Unpaired d electrons can be promoted over this gap,
absorbing energy which corresponds to wavelengths in the visible region of the electromagnetic
spectrum. Cu ions in solution are colourless because its electronic configuation ([Ar] 4s0 3d10 )
71
Figure 5.16: Ascorbic acid and thiourea

contains no unpaired d electrons for promotion.
The second step is the removal of the Cu ions by chelation with thiourea (fig. 5.16). In the
presence of ascorbate, thiourea forms a colourless, redox-inactive complex with Cu . Although
the precise configuration of this complex is not known, the thiol group is critical in the chelation
process, 68 and a single Cu ion can bind up to four thiourea molecules. 69
5.5.4 Future Bioapplications

The proteins successfully bound to CNx MWNTs all contain a metal ion core which may have interesting effects on the electronic properties of the CNx MWNT. Whereas some previous work has
involved a covalent 31 or non-covalent 62 adhesion layer between protein and CNT, this does not allow for direct measurement of the contribution made by the proteins to the electronic properties 7 of
the CNx MWNTs. In Chapter 8 a biosensor device is described for the detection of metalloproteins,
which relies on NSB only.
72
References
[1] S. C. Tsang, Z. Guo, Y. K. Chen, M. L. H. Green, H. A. O. Hill, T. W. Hambley, and P. J.
Sadler. Immobilization of platinated and iodinated oligonucleotides on carbon nanotubes.
Angew. Chem. Int. Ed., 36(20):21972200, 1997.
[2] Y. Lin, S. Taylor, H. Li, K. A. Shiral Fernando, L. Qu, W. Wang, L. Gu, B. Zhou, and Y.-P.
Sun. Advances toward bioapplications of carbon nanotubes. J. Mater. Chem., 14:527541,
2004.
[4] N. Wong Shi Kam, T. C. Jessop, P. A. Wender, and H. Dai. Nanotube molecular transporters:
Internalization of carbon nanotube-protein conjugates in to mammalian cells. J. Am. Chem.
Soc., 126:68506851, 2004.
[5] N. Wong Shi Kam, M. J. OConnell, J. A. Wisdom, and H. Dai. Carbon nanotubes as mulitfunctional biological transporters and near-infrared agents for selective cancer cell destruction.
Proc. Nat. Acad. Sci., 102(33):116005, 2005.
[6] V. Renugopalakrishnan, A. Strzelczyk, P. Li, A. A. Mokhnatyuk, S. H. Gursahani, M. Nagaraju, M. Prabhakaran, H. Arjomandi, and S. L. Lakka. Retroengineering bacteriorhodopsins:
Design of small proteins by bionanotechnology. Int. J. Quantum Chem., 95:627631, 2003.
[7] J. J. Davis, K. S. Coleman, B. R Azamian, C. B. Bagshaw, and M. L. H. Green. Chemical and
biochemical sensing with modified single wall carbon nanotubes. Chem. Eur. J., 9:37323739,
2003.
[8] F. Balavoine, P. Shultz, C. Richard, V. Mallouh, T. W. Ebbesen, and C. Mioskowski. Helical
crystallization of proteins on carbon nanotubes: a first step towards the development of new
biosensors. Angew. Chem. Int. Ed., 38(13/14):19121915, 1999.
Lett., 4(2):253256, 2004.
[10] M. Shim, N. Wong Shi Kam, R. J. Chen, Y. Li, and H. Dai. Functionalization of carbon
nanotubes for biocompatibility and biomolecular recognition. Nano Lett., 2(4):285288, 2002.
[11] S. Bhattacharyya, C. Sinturel, J.-P. Salvetat, and M.-L. Saboungi. Protein-functionalized carbon nanotube-polymer composites. Appl. Phys. Lett., 86:113104, 2005.
[12] R. Prakash, R. Superfine, S. Washburn, and M. R. Falvo. Functionalization of carbon
nanotubes with proteins and quantum dots in aqueous buffer solutions. Appl. Phys. Lett.,
88:063102, 2006.
[13] B. R Azamian, J. J. Davis, K. S. Coleman, C. B. Bagshaw, and M. L. H. Green. Bioelectrochemical single-walled carbon nanotubes. J. Am. Chem. Soc., 124:1266412665, 2002.
73
[14] Y. Lin, L. F. Allard, and Y.-P. Sun. Protein-affinity of single-walled carbon nanotubes in water.
J. Phys. Chem. B., 108:37603764, 2004.
[15] W. Huang, S. Taylor, K. Fu, Y. Lin, D. Zhang, T. W. Hanks, A. M. Rao, and Y.-P. Sun. Attaching proteins to carbon nanotubes via diimide-activated amidation. Nano Lett., 2(4):311314,
2002.
[16] S. S. Karajanagi, H. B. Yang, P. Asuri, E. Sellitto, J. S. Dordick, and R. S. Kane. Proteinassisted solubilzation of single-walled carbon nanotubes. Langmuir, 22:13921395, 2006.
[17] A. Star, D. W. Steuerman, J. R. Heath, and J. F. Stoddart. Starched carbon nanotubes. Angew.
Chem. Int. Ed., 41(14):25082512, 2002.
[18] O.-K. Kim, J. Je, J. W. Baldwin, S. Kooi, P. E. Pehrsson, and L. J. Buckley. Solubilzation
of single-wll carbon nanotubes by supramolecular encapsulation of helical amylose. J. Am.
Chem. Soc., 125:44264427, 2003.
[19] B. F. Erlanger, B.-X. Chen, M. Zhu, and L. Brus. Binding of an anti-fullerene IgG monoclonal
antibody to SWNTs. Nano Lett., 1(9):465467, 2001.
[20] R. J. Chen, H. C. Choi, S. Bangsaruntip, E. Yenilmez, X. Tang, Q. Wang, Y.-L. Chang, and
H. Dai. An investigation of the mechanisms of electronic sensing of protein adsorption on
carbon nanotube devices. J. Am. Chem. Soc., 126:15631568, 2004.
[21] A. Bianco and M. Prato. Can carbon nanotubes be considered useful tools for biological
applications? Adv. Mater., 15(20):17651768, 2003.
[22] S. C. Tsang, J. J. Davis, M. L. H. Green, H. A. O. Hill, Y. C. Leung, and P. J. Sadler. Immobilization of small proteins in carbon nanotubes: High-resolution transmission electron
microscopy study and catalytic activity. Chem. Commun., pages 18031805, 1995.
[23] J. J. Davis, M. L. H. Green, H. A. O. Hill, Y. C. Leung, P. J. Sadler, J. Sloan, A. V. Xavider,
and S. C. Tsang. The immobilisation of proteins in carbon nanotubes. Inorg. Chim. Acta,
272:261266, 1998.
[24] Z. Guo, P. J. Sadler, and S. C. Tsang. Immobilization of DNA and proteins on carbon nanotubes. Adv. Mater., 10(9):701702, 1998.
[25] C. Zhi, Y. Bando, C. Tang, and D. Golberg. Immobilization of proteins on Boron Nitride
nanotubes. J. Am. Chem. Soc., 127:1714417145, 2005.
Lett., 89:143110, 2006.
[27] J. C. Carrero-Sanchez, A. L. Elas, R. Mancilla, G. Arrelln, H. Terrones, J. P. Laclette, and
M. Terrones. Biocompatibility and toxicological studies of carbon nanotubes doped with nitrogen. Nano Lett., 6(8):16091616, 2006.
74
[28] F. Villalpando-Paez, A. H. Romero, E. Munoz Sandoval, L. M. Martnez, H. Terrones, and

M. Terrones. Fabrication of vapor and gas sensors using films of aligned CNx sensors. Chem.
Phys. Lett., 386:137143, 2004.
Acc. Chem. Res., 21:407413, 1988.
[31] K. Jiang, L. S. Schadler, R. W. Siegel, X. Zhang, H. Zhang, and M. Terrones. Protein immoblisation on carbon nanotubes via a two-step process of diimide activated amidation. J. Mater.
Chem., 14:3739, 2004.
[32] X. Jiang, Z. Zhang, H. Bai, Z. Qu, J. Jiang, E. Wang, and S. Dong. Effect of electrode surface
microstructure an electron transfer induced conformation changes in cytochrome c monitored
by in situ UV and CD spectroelectrochemistry. Spectrochim. Acta, Part A, 61:943951, 2005.
[33] J.-Z. Xu, J.-J. Zhu, Q. Wu, Z. Hu, and H-Y. Chen. An amperometric biosensor based on
the coimmobilization of horseradish peroxidase and methylene blue on a carbon nanotubes
modified electrode. Electroanalysis, 15(3), 2003.
[34] L. Zhao, H. Lin, and N. Hu. Electroactive films of protein-coated multiwalled carbon nanotubes. J. Colloid Interface Sci., 296:204211, 2006.
[35] J. J. Davis, R. J. Coles, and H. A. O. Hill. Protein electrochemistry at carbon nanotube electrodes. J. Electroanal. Chem., 440:279282, 1997.
[36] F. D. DSouza, L. M. Rogers, E. S. ODell, A. Kochman, and W. Kutner. Immoblization
and electrochemical redox behaviour of cytochrome c on fullerene film-modified electrodes.
Bioelectrochemistry, 66:3540, 2005.
[37] L. Stryer. Biochemistry. W.H. Freeman and Co., 4th edition, 1995.
[38] M. Collinson and E. F. Bowden. UV-visible spectroscopy of adsorbed cytochrome c on tin
oxide electrodes. Anal. Chem., 64:14701476, 1992.
[39] A. P. Demchenko. Ultraviolet Spectroscopy of Proteins. Springer-Verlag, Kiev, 1986.
[40] C. J. Wilson and P. Wittung-Stafshede. Role of structural determinants in folding of the
sandwich-like protein Pseudomonas aeruginosa azurin. Proc. Nat. Acad. Sci., 102(11):3984
3987, 2005.
[41] J. A. Blaszak, D. R. McMillin, A. T. Thornton, and D. L. Tennent. Kinetics of Copper(II)
uptake by apoazurin on complexing media. J. Biol. Chem., 258(16):98869892, 1983.
[42] L. Fuentes, J. Oyola, M. Fernandez, and E. Quinones. Conformational changes in azurin
from Pseudomona aeruginosa induced through chemical and physical protocols. Biophys. J.,
87:18731880, 2004.
75
[43] D. N. Petsev, B. R. Thomas, S.-T. Yau, and P. G. Vekilov. Interactions and aggregation of
apoferritin molecules in solution: Effects of added electrolytes. Biophys. J., 78:20602069,
2000.
[44] A Dictionary of Biology. Oxford University Press, Oxford, 4th edition, 2000.
[45] K. Jiang, A. Eitan, L. S. Schadler, P. M. Ajayan, R. W. Siegel, N. Grobert, M. Mayne,
M. Reyes-Reyes, H. Terrones, and M. Terrones. Selective attachment of gold nanoparticles to
nitrogen-doped carbon nanotubes. Nano Lett., 3(3):275277, 2003.
[46] M. Heusterspreute and R. R. Crichton. Amino acid sequence of horse spleen apoferritin. FEBS
Lett., 129(2):322327, 1981.
[47] X. Yang and N. D. Chasteen. Ferroxidase activity of ferritin: effects of pH, buffer and Fe(II)
and Fe(III) concentrations on Fe(II) autoxidation and ferroxidation. Biochem. J., 338:615618,
1999.
[48] Y. Li, W. Kim, Y. Zhang, M. Rolandi, D. Wang, and H. Dai. Growth of single-walled carbon nanotubes from discrete catalytic nanoparticles of various sizes. J. Phys. Chem. B.,
105:1142411431, 2001.
[49] J. Leckner. Folding and Structure of Azurin - The influence of a Metal. PhD thesis, Chalmers
University of Technology, 2001.
[50] H. Nar, R. Huber, A. C. Messerschmidt, M. Barth, M. Jaquinod, M. van der Kamp, and G. W.
Canters. Characterization and crystal structure of zinc azurin, a by-product of heterologous
expression in Escherichia coli of Pseudomonas aeruginosa copper azurin. Eur. J. Biochem.,
205:11231129, 1992.
[51] N. Borovok, A. B. Kotlyar, I. Pecht, L. K. Skov, and O. Farver. Photoinduced electron transfer
in singly labeled thiouredopyrenetrisulfonate azurin derivatives. FEBS, 457:277282, 1999.
[52] J. E. Hansen, M. K. McBrayer, M. Robbins, and Y. Suh. A pH dependence study on the
unfolding and refolding of apoazurin: comparison with Zn(II) azurin. Cell Biochem. Biophys.,
36(1):1940, 2002.
[53] N. Greenfield and G. D. Fasman. Computed circular dichroism spectra for the evaluation of
protein conformation. Biochemistry, 8:41084116, 1969.
[54] L. Jeuken. The structure-function relation of small copper-containing proteins. PhD theis,
Leiden University, 2000.
[55] V. sMilitello, M. Leone, A. Cupane, R. Santucci, and A. Desideri. Local dynamic properties
of the heme pocket in native and solvent-induced molten-globule-like states of cytochrome c.
Biophys. Chem., 97:121128, 2002.
[56] Q. Xu and T. A. Keiderling. Optical spectroscopic differentiation of various equilibrium denatured states of horse cytochrome c. Biopolymers, 73:716726, 2004.
76
[57] M. Fedurco, J. Augustynski, C. Indiani, G. Smulevich, M. Antalk, M. Bano, E. Sedlak, M. C.

Glascock, and J. H. Dawson. The heme iron coordination of unfolded ferric and ferrous
cytochrome c in neutral and acidic urea solutions. Spectroscopic and electrochemical studies.
Biochimica et Biophysica Acta, 1703:3141, 2004.
[58] A. Muga, H. H. Mantsch, and W. K. Surewicz. Membrane binding induced destalibilization
of cytochrome c. Biochemistry, 30(29):72197224, 1991.
[59] R. R. Crichton, R. Eason, A. Barclay, and C. F. A. Bryce. The subunit structure of horse spleen
apoferritin: the molecular weight of the oligomer and its stability to dissociation by dilution.
Biochem. J., 131:855857, 1973.
[60] S.-T. Yau, B. R. Thomas, and P. G. Vekilov. Molecular mechanisms of crystallization and
defect formation. Phys. Rev. Lett., 85(2):353356, 2000.
[61] W. K. Surewicz, G. Szabo, and H. H. Mantsch. Conformational properties of azurin in solution
as determined from resolution-enhanced Fourier-transform infrared spectra. Eur. J. Biochem.,
167:519523, 1987.
[62] R. J. Chen, Y. Zhang, D. Wang, and H. Dai. Noncovalent sidewall functionalization of SWNTs
for protein immobilization. J. Am. Chem. Soc., 123:38383839, 2001.
[63] S. S. Karajanagi, A. A. Vertegel, R. S. Kane, and J. S. Dordick. Structure and function of
enzymes adsorbed onto single-walled carbon nanotubes. Langmuir, 20:1159411599, 2004.
[64] G. R. Diekmann, A. B. Dalton, P. A. Johnson, J. Razal, J. Chen, G. M. Giordano, E. Munoz,
I. H. Musselman, R. H. Baughman, and R. K. Draper. Controlled assembly of carbon nanotubes by designed amphiphillic peptide helices. J. Am. Chem. Soc., 125:17701777, 2003.
[65] S. S. Tatke, V. Renugopalakrishnan, and M. Prabhakaran. Interfacing biological macromolecules with carbon nanotubes and silicon surfaces: a computer modelling and dynamic
simulation study. Nanotechnology, 15:684690, 2004.
[66] E. V. Basiuk, V. A. Basiuk, J.-G. Banuelos, J.-M. Saniger-Blesa, V. A. Pokrovskiy, Y. T. Gromovoy, A. V. Mischanchuk, and B. G. Mischanchuk. Interaction of oxidized single-walled
carbon nanotubes with vaporous amines. J. Phys. Chem. B., 106:15881597, 2002.
[67] J.-J. Feng, G. Zhao, J.-J. Xu, and H-Y. Chen. Direct electrochemistry and electrocatalysis
of heme proteins immoblized on gold nanoparticles stabilized by chitosan. Anal. Biochem.,
342:280286, 2005.
[68] B.-Z. Zhu, W. E. Antholine, and B. Frei. Thiourea protects against copper-induced oxidative
damage by formation of a redox-inactive thiourea-copper complex. Free Radical Biology &
Medicine, 32(12):13331338, 2002.
[69] C. J. Doona and D. M. Stanbury. Equilibrium and redox kinetics of copper(II) - thiourea
complexes. Inorg. Chem., 35:32102116, 1996.
77
Chapter 6

6.1 Previous Work
The thermal properties of CNTs are of fundamental interest, and also play a critical role in controlling the performance and stability of devices. 1,2 CNTs have been predicted to have very high
thermal conductance, 3 and high values for the thermal conductivity, of un-doped CNTs (3000
W/mK at room temperature 4 ) have been determined. 2,5,6
The range of values reported is across several orders of magnitude, probably because of differences in the method used (e.g. 3 vs molecular dynamics) or differences in the CNT size and type,
which give rise to different mean free paths of energy carriers, 7 .
6.2 Experimental
Using the set-up of Brown et al. 8 detailed in Chapter 4, simultaneous electrical and thermal measurements were taken. This required independent temperature and voltage circuits. A DC resistance
thermometer was used to record the cantilever tip temperature, and the electrical conductance of
the tip was measured with a lock-in amplifier, which detected an AC voltage drop across the tube.
A flake of CNx MWNTs was mounted in a Veeco thermal probe, a short exposed section of the
Pt-Rh core at the probes tip (known as a Wollaston Wire) acted as the resistance thermometer and
78
measured the temperature at the top of the CNx MWNT flake.

The Wollaston wire can be heated to
600 K by ohmic dissipation. All measurements were
done in vacuum (105 mbar) to reduce extraneous heat flow. During the measurements the DC
resistance of the probe was monitored and a feedback loop controlled the probe bias current, to
maintain the probe temperature. The bias current and resistance were recorded such that on contact
between CNx MWNT and substrate (HOPG), the change in these parameters allowed the resulting
heat flow down the CNx MWNTs to be calculated. The temperature at both ends of the CNx MWNT
was known, allowing the thermal conductance to be calculated.
The CNx MWNT flake was brought into contact with the HOPG by lowering the probe (using
piezo control) until a pre-set force was applied by the individual CNx MWNT protruding from the
end of the flake. It was then moved repeatedly away from and towards the substrate (by means of a
force curve) at a constant speed (typically 20 nm/s) whilst thermal and electrical conductances were
recorded simultaneously.
6.3 Results
Individual CNx MWNTs have an electrical conductance of
1 G0 per tube. 9 When a bundle of
aligned CNx MWNTs in contact with the HOPG substrate is moved away from the surface during a
force curve, steps in units of 1 G0 are observed [fig. 6.1 (blue)]. As thermal conductance measurements are also being taken, a corresponding step in the power loss (W) is simultaneously observed
[fig. 6.1 (red)].
More than 60 simultaneous electrical and thermal steps were observed and in fig. 6.2 electrical
conductance (G0 ) step size is plotted against the corresponding step in thermal conductance (or
power loss per degree Kelvin). Although there is some scatter in the data points, there is also a
linear trend with a mean slope 0.1445. The units of the slope in fig. 6.2 are:
W
T G0
P
G0 T
Q
G0 T t
(6.1)
As shown in Chapter 4 an electrical conductance of 1 G0 corresponds to an individual CNx

79
Figure 6.1: Simultaneous electrical and thermal steps for an individual CNx MWNT
MWNT. Thus the slope can be simplified to:

Q
T t
(6.2)
for an individual tube. Recalling the equation for :
Q
L
t
AT
(6.3)
slope L
A
(6.4)
A simple substitution yields:
By considering a CNx MWNT as a solid cylinder 2 the cross-sectional area A will be r2 . Thus
using an average diameter of 21.7 nm (see section 7.3.2), A = 1.48 1015 m2 , and L = 2.44 m,
(see section 7.3.1) the thermal conductivity can be calculated to be:
240 W/mK at room temperature.

80
Figure 6.2: Electrical step size vs thermal step size for CNx MWNTs
6.4 Discussion
The measured value of is an estimate because of the ambiguity of the dimensions of CNx MWNTs,
which make it difficult to derive a value for the absolute magnitude of the thermal conductivity. 5 Despite any additional error, a value of 240 W/mK is still significantly lower than the value measured
for individual un-doped MWNTs using the same method (2.5104 W/mK). 8
More similar values of thermal conductivity have however been obtained for a mat of aligned
un-doped MWNTs. 6 The reason for low thermal conductivity in these mats is the presence of many
defects such as atomic vacancies. 10 Che et al. determined theoretically the influence of various
defects, and as expected, found that decreases as the vacancy concentration increases. 11 Given the
concentration of defects in CNx MWNTs (section 2.2.2) it is not surprising therefore that they too
exhibit poor thermal conductivity.
81
References
[3] S. Berber, Kwon. Y., and D. Tomanek. Unusually high thermal conductivity of carbon nanotubes. Phys. Rev. Lett., 84:46134616, 2000.
[4] P. Kim, L. Shi, A. Majumdar, and P. L. McEuen. Thermal transport measurements of individual
multiwalled nanotubes. Phys. Rev. Lett., 21:215502, 2001.
Phys. Rev. B, 59(14):90159018, 1999.
[7] T.-Y. Choi, D. Poulikakos, J. Tharian, and U. Sennhauser. Measurement of the thermal conductivity of individual carbon nanotubes by the four-point three- method. Nano Lett., 6(8):1589
1593, 2006.
2005.
Lett., 89:143110, 2006.
[10] X. H. Yan, Y. Xiao, and Z. M. Li. Effects of intertube coupling and the tube chirality on
thermal transport of carbon nanotubes. J. Appl. Phys., 99:124305, 2006.
[11] J. Che, T. C
agn, and W. A. Goddard III. Thermal conductivity of carbon nanotubes. Nanotechnology, 11:6569, 2000.
82
Chapter 7
The Effect of Acid Treatment on CNx

MWNTs
7.1 Previous Work
The sp2 hexagonal carbon network of CNTs (see Appendix section 10.3) is chemically inert, but
oxidation of nanotube caps and sidewalls results in rehybridisation, 1,2 allowing a wide range of
chemical derivitisations. 3,4 Oxidation of CNTs can be achieved by heating the CNTs above 700 C
in atmospheres of air, 5 carbon dioxide 6 or oxygen. 7 However, yields are very small and and better
results can be obtained using liquid-phase oxidants such as KMnO4 , OsO4 , or concentrated nitric
or hydrochloric acid. 6,811 Reaction of CNTs with nitric acid alone is extremely slow and weak 2,8
and a mixture of nitric and sulphuric acid is preferable. 12 In the following work CNx MWNTs are
oxidised using a mixture of HNO3 and H2 SO4 , and the effect of acid treatment on various physical
properties is investigated.
7.1.1 Solubility
Oxidation of CNTs is desirable because it improves their dispersion in water 3,8 due to the acid
groups created on the nanotube surface. 8,10,13 This is important because a major drawback of CNTs,
particularly relevant to their compatibility with biological systems, is their insolubility in all types
83
Figure 7.1: AC-AFM image of CNx MWNTs dispersed in water without (left) and with (right) 1%
SDS
of solvents. 14 All CNTs including CNx MWNTs, are very insoluble in water. They are slightly
more soluble in organic 15 solvents such as ethanol, methanol and acetone, but these solvents are
unsuitable for biological applications.
Un-doped SWNTs can be dissolved in water with the assistance of a surfactant such as 1% (by
weight) sodium dodecyl sulfate (SDS) accompanied by high power sonication. 16 This process has
also been successfully applied to CNx MWNTs. Fig. 7.1 shows AFM images of the raw material
after 30 min sonication in MilliQ water (left) and sonication in a 1% SDS/MilliQ water solution
(right).
7.2 Experimental
7.2.1 Acid Oxdiation
Less than 1 mg of raw nanotubes was suspended in a 3:1 (v:v) mixture of concentrated H2 SO4
and HNO3 . The solution was left at room temperature in a sonic bath for 2 h. 17 The resulting black
solution was filtered through a hydrophilized PTFE membrane (47 mm, pore size 1.2 m, Whatman)
using a 3-piece funnel set.
84
7.2.2 Lyophilization
Recovery of acid-treated tubes from the membrane was low. Sonication was used to re-suspend the
oxidized CNx MWNTs in MilliQ water, and lyophilization was used to recover the tubes as a solid
from this solution.
23 mg of raw CNx MWNTs were subject to acid oxidation at room temperature for 2 h, washed
and filtered. The filter paper and glassware were sonicated in a sonic bath to produce 200 mL
of black solution. The solution was divided between two large round-bottom flasks and frozen
by liquid N2 with swirling to form a thin layer of ice, with maximum surface area possible. The
solution was lyophilized at a pressure of 1 mTorr and temperature -70 C. The ice was removed by
sublimation overnight, leaving 21 mg black, fluffy powder.
7.2.3 AFM Imaging and Dimension Data

Acid-treated CNx MWNTs were suspended in MilliQ water with sonication, and then dispersed
on glass slides for imaging by AFM. During imaging, the Nanoscope software (Veeco) was used
to measure length and diameter data. Length and diameter data for raw CNx MWNTs was also
recorded for comparison.
7.2.4 Acid-base titrations

The percentage of acid groups on the CNx MWNTs after acid-oxidation was estimated via acid-base
titration. Two different methods 9,18 were used, both of which gave a similar result for percentage of
acid sites. In all cases the CNx MWNTs were washed thoroughly with MilliQ water before titration,
to remove any acidic residues.
Titration of basic CNx MWNTs with an indicator

Following the method of Satishkumar et al. 9 the indicator phenolphthalein (fig. 7.2) was used to
show the end-point of the titration. Phenolphthalein has a pink colourless transition at pH 8.2. A
known quantity of lyophilized acid-treated CNx MWNTs (measured using a SCIENTECH SM50
85
Figure 7.2: (a) Phenolphthalein in basic aqueous/EtOH solution and (b) molecular formula of phenolphthalein
0.01 mg balance) was briefly sonicated and then stirred in 0.005 M NaOH for 24 h. The solution
was titrated against 0.005 M potassium hydrogen phthalate, a weak acid.
Titration of basic CNx MWNTs with a pH Meter

Following the method of Hu et al. 18 a pH meter was used to indicate the end-point of a strong
acid-strong base titration. A known quantity of lyophilized, acid-treated CNx MWNTs was stirred
in 0.05 M NaOH for 48 h. Any excess NaOH was titrated against 0.05 M HCl until the neutral
point was reached, which in this case was pH 7.0. The end-point was indicated using a HANNA
Instruments 1230 pH meter.
The effect of NaOH on the acid-treated CNx MWNTs can be viewed as:
CNx COOH aq NaOH aq
CNx COONa aq NaOH excess H2 O l (7.1)
The number of moles of excess NaOH was then determined by titration with 0.05 M HCl:
NaOH aq HCl aq
H2 O l NaCl aq
(7.2)
7.2.5 Measurement of Electronic Transport Properties

The intrinsic conductances of acid-treated CNx MWNTs were measured using the same method as
Chapter 4 but using a flake of lyophilized acid-treated CNx MWNTs instead. SEM images of flakes
of both raw and acid-treated CNx MWNTs are shown in fig. 7.3. An indication of the physical
86
Figure 7.3: SEM images of flakes of (a) raw and (b) acid-treated CNx MWNTs mounted on Veeco
thermal probes (David Cox)
effect of acid oxidation on the CNx MWNTs can be seen in the image. The raw CNx MWNT flake
[fig. 7.3(a)] consists of many closely aligned individual and bundles of CNx MWNTs, whereas the
flake of acid-treated CNx MWNTs [fig. 7.3(b)] is highly disordered.
G-V and I-V curves, and electrical breakdown measurements were recorded for acid-treated
CNx MWNTs as per Chapter 4.
7.2.6 Measurement of Thermal Transport Properties
The thermal properties of acid-treated CNx MWNTs were measured using the same method as in
Chapter 6.
87
7.3 Results
7.3.1 Length Data
Length data are given in the Appendix (section 10.2). Acid-treatment of un-doped CNTs results in
shorter, open tubes. 13 The significant difference in reaction rates between the caps and sidewalls 7
is responsible for this shortening.
Shortening was also observed after acid treatment of CNx MWNTs. Fig. 7.4 shows length
histograms for (a) acid-treated and (b) raw CNx MWNTs. AC-AFM images (Nashville Toledo) of
dispersions of acid-treated and raw CNx MWNTs are shown in fig. 7.4 (c) and (d) respectively.
The average length before acid treatment was 2.44 1.59 m, and the average length after was
1.19
the
0.76 m. This represents a shortening speed of
600 nmhour1 , considerably faster than
200 nmhour1 reported for un-doped SWNTs under stronger conditions. 3 The AFM images
show that the CNx MWNTs are better dispersed 17 after acid treatment and that the acid-treated CNx
MWNTs appear to be all of a similar length, which was found to be
1 m. When the reaction
time at room temperature was increased (up to 66 h) no appreciable effect on length statistics was
observed; it appears that the acid oxidation process is self-limiting.
7.3.2 Diameter Data

In the case of MWNTs, previous studies have shown that during acid treatment, successive individual layers are peeled away resulting in a decrease in diameter or thinning. 1,5,6
The average diameter before acid treatment was 21.7 7.5 nm, and the average diameter after
was 13.7 4.4 nm.
7.3.3 Acid-base Titrations

An acid-base titration curve for titration with phenolphthalein is shown in fig. 7.6. The shape of the
graph is typical of a weak acid vs strong base titration.
The surface density of acid groups (the number of C atoms oxidized) present on CNx MWNTs
after 2 h acid oxidation was found to be in the range 30-35% for all titration methods. This is signif88
Figure 7.4: Histograms of (a) lengths of acid-treated CNx MWNTs and (b) lengths of raw CNx
MWNTs. Corresponding AC-AFM images (Nashville Toledo) are shown in (c) and (d)
Figure 7.5: Histogram of (a) acid-treated CNx MWNT and (b) raw diameter data.
89
Figure 7.6: Acid-base titration curve for CNx MWNTs and NaOH
icantly larger than the 10-15% obtained for un-doped acid-treated MWNTs. 8,10,19 No appreciable
change in the percentage of acid functional groups was observed after acid oxidation for 66 h. An
example calculation of the percentage of acid sites is given in the Appendix section 10.1.
7.3.4 Electrical Transport

The electrical transport properties of raw CNx MWNTs prior to acid treatment were reported in
Chapter 4. Fig. 7.7 shows histograms of intrinsic conductances for (a) acid-treated and (b) raw CNx
MWNTs at low bias voltage. Most of the data points (73% raw and 90% acid) fall in the range 0
- 2 G0 which is the range shown in fig. 7.7. Higher values of conductance were also observed, but
in both cases these are likely to correspond to bundles of CNx MWNTs. The histogram for acidtreated CNx MWNTs shows a propensity for lower values of intrinsic conductance. However since
no obvious peak is observed we can only say that the conductance observed is poorly defined, and
in the majority of cases low, i.e.
1.0 G0 . In comparison, the mean conductance for an individual
raw CNx MWNTs is well defined, as shown by the Gaussian distribution plotted in fig. 7.7(b) which
peaks at 1.0 0.3 G0 .
The reduced conductance of the acid-treated CNx MWNTs compared to raw CNx MWNTs is
also reflected in the current-voltage characteristics. The I-V and G-V curves shown in fig. 7.8(a) and
90
Figure 7.7: Histograms showing the intrinsic conductances for (a) acid-treated and (b) raw CNx
MWNTs
(c) for acid-treated CNx MWNTs are typical of the 150 different contacts measured. The average
slope of acid-treated G-V curves for the range 0 to 2 V is 0.19 G0 /V, significantly smaller than the
average of 0.70 G0 /V measured for raw CNx MWNTs. Different behaviour was also observed in
the electrical breakdown of the acid-treated CNx MWNTs. For both individual raw and acid-treated
CNx MWNTs, electrical breakdown occurs when the voltage across the nanotube (Vcnt )
3.0 V. An
example of the breakdown of an individual acid-treated CNx MWNT is shown in fig. 7.9(a). Once
the threshold voltage is reached, a stable current is maintained followed by two steps (each 40 A),
the current then stablises before two more steps (70 A and 80 A) with the final step taking the
current to zero. In contrast, during the breakdown of an individual raw CNx MWNT a gentle decline
in current is observed followed by a number of smaller sharp steps [fig. 7.9(b)]. Steps observed
during breakdown are normally associated with the breakdown of individual MWNT layers. The
large decline steps observed for acid-treated CNx MWNT may correspond to the breakdown of
many MWNT layers simultaneously. 20 The average step size for observed during breakdown was
60
30 A for individual acid-treated CNx MWNTs, and 20 10 A for individual raw CNx
MWNTs.
91
Figure 7.8: G-V curves for (a) acid-treated and (b) raw CNx MWNTs and (c) and (d) show the
corresponding I-V curves.
Power Threshold values

Acid-treated CNx MWNTs consistently reached higher values of Pthresh compared to raw CNx
MWNTs. The highest power value reached during breakdown of acid-treated CNx MWNTs was
1003 615 W compared to 490 100 W for raw CNx MWNTs.
7.3.5 Thermal Transport

Fig. 7.10 shows a graph of power loss step size vs electrical conductance for both acid-treated (red)
and raw (blue) CNx MWNTs. For the acid-treated CNx MWNTs both the thermal conductance and
the electrical conductance are reduced simultaneously, producing a graph with a very similar slope
and scatter to raw CNx MWNTs. The graph has slope 0.1606 W/K per G0 however as the electrical
conductance of acid-treated CNx MWNTs is poorly defined (section 7.3.4) it is not possible to
calculate a
Q
T t
per tube. Hence the thermal conductivity for acid-treated CNx MWNTs cannot be
calculated.
92
Figure 7.9: Breakdown of an individual (a) acid-treated CNx MWNT and (b) and an individual raw
CNx MWNT at 3 V followed in real time. Arrows indicate the size of current steps in A.
Figure 7.10: Electrical step size vs power loss step size for acid treated (red) and raw (blue) CNx
MWNTs
93
7.4 Discussion
7.4.1 Solubility
CNTs are highly insoluble due to the very strong intermolecular forces that cause them to aggregate
in solution. Un-doped CNTs are uncharged so the forces operating are VDWs (see section 3.1.1)
which have a corresponding energy of
500 eV per m of the tube-tube contact. 21
The forces are generated between adjacent -electron bonding networks on CNT sidewalls and
can be disrupted by sonication. However for CNx MWNTs, extremely strong sonication is required. 22 This is because of additional contribution to the intermolecular forces known as dipoledipole interactions (see section 3.1.1). As nitrogen is more electronegative than carbon (3.0 vs 2.5)
a small dipole is generated across the C-N bond which interacts attractively with dipoles on adjacent
nanotubes. Hydrogen-bonding is an example of a strong dipole-dipole interaction.
Sonication
Ultrasonication was used to disperse CNx MWNT aggregates and often a CNT ring or crop circle
was observed in AFM measurements. These rings are caused by bubble formation during sonication
of the CNT solution. In the simplest model of ring formation process, the CNT coils over itself to
form a loop inducing significant strain in the process. 23 When bubbles in the solution suddenly
collapse, the CNTs remain bent mechanically.
7.4.2 Oxidation Mechanism

The oxidation mechanism for un-doped nanotubes is well understood; In the absence of any other
defects, 24 oxidative reactivity is dominated by strain, 1 i.e. attack occurs selectively at regions of
greatest curvature such as non-six-membered rings, which occur at the CNT cap.
A commonly encountered tip configuration involves a conical structure where five pentagons
are clustered to form a sharp tip and a sixth pentagon occurs further away (fig. 7.11). The sixth
pentagon allows a seamless merging of the cone-like tip and the rest of the tube, but also puts the
structure under severe long-range strain. 25 This strain increases local reactivity by causing a shift in
94
Figure 7.11: A TEM image showing the effect of the sixth pentagon (arrow) on a MWNT. The sixth
pentagon produces a sharp ridge which extends to the tip from Iijima et al. 25
hybridisation of the atoms from the sp2 of graphite to something intermediate between it and sp3 .
The rehybridisation causes the CNT to buckle and together with the curvature, this will accelerate
the reaction. 1
In a MWNT, concentric layers do not all react at the same rate - since each shell has its own tip,
the inner layers persist longer than the outer shells. 1
Mechanism for CNx MWNTs

The effect of substitutional atomic doping 3,26 on oxidation is unknown, but any additional defects
will become sites of preferential acid etching, 1 reducing the significance of the cap regions. 27 In
CNx MWNTs nitrogen-doping can produce vacancies or defects in the hexagonal carbon lattice (see
section 2.2.2). 28 These doping sites are present in both the sidewalls and the caps of CNx MWNTs,
at a concentration of 2-5%. Given these additional defects, a different acid oxidation mechanism is
likely for CNx MWNTs.
Preferential etching at nitrogen-doping defects over heptagons or pentagons, may explain why
more gentle reaction conditions are required for CNx MWNTs (e.g. 2 h at room temperature 17
compared to reflux for 4.5 h at 140 C 6 ). If acid etching occurs at nitrogen-doping sites, a higher
surface density of functional acid groups (compared to un-doped CNTs) should exist on the CNx
95
MWNTs after acid treatment. This was found to be the case, as measured by acid-base titration.
The self-limiting nature of the oxidation, as shown in the limiting length of
1 m observed after
2 h at room temperature, may indicate either the reaching of a structural impasse or the completion
of attack on all available nitrogen-based defects. The bamboo segments of CNx MWNTs may also
be weak points in the structure that are vulnerable to cleaving during acid oxidation.
If the reaction of all available nitrogen-based defects is the limiting factor, further shortening
of the CNx MWNTs (by reaction at heptagon and pentagon defects) would then require stronger
conditions. In order to test this theory, after 2 h at room temperature the acid/CNx MWNTs solution
was heated to 60 C for a further 1 h. The average length after this further acid treatment was found
to be 660 230 nm and the average diameter 11.0 3.0 nm.
Functional Groups
The two concentrated acids have different functions in the oxidation process. The nitric acid has a
purification function (to remove the metal catalyst particles 29 ) whereas the sulphuric acid creates
acid functionalities. An oxygen incorporation mechanism for CNTs is not entirely known, but the
overall result of the oxidizing treatment is the incorporation of oxygen into a C-C bond forming
ethers (C-O-C) 30 , alcohols (C-OH), ketones (C=O), and by far the most common, carboxylic acids
(COOH, fig. 7.12) .
The multitude of functional groups formed on carbon surfaces is due to the different reactivities
of armchair and zig-zag edge-sites. 31 The presence of esters (COOR) can be excluded as esters are
hydrolysed in acid media. 32
Whilst the surface will remain largely hydrophobic in nature, the defects are oxidized, producing
hydrophilic areas. These make the CNTs soluble in water and improve protein functionalisation
(Chapter 5).
7.4.3 Electrical Transport

The low conductances observed for acid treated CNx MWNTs are likely to be due to scattering
at the defects created during acid oxidation. Individual defects (including nitrogen 33 and boron 34
96
Figure 7.12: A carboxylic acid group on an oxidized nanotube from Jiang et al. 17
sites) have been shown to reduce conductance due to scattering. 35,36 Whilst the effect of a single
-OH group on the electronic properties of a SWNT has been investigated, 37 the effect of multiple
defects produced during acid treatment, 38 is not known.
Given that multiple defects produced during CVD synthesis cause a significant increase in resistivity, 39,40 it is likely that multiple acid oxidation defects will have a similar effect. The contribution
of acid treatment to contact resistance could also explain the reduced conductances observed. Ago et
al. measured the work function of both raw and acid-treated un-doped MWNTs and saw an increase
from 4.3 eV to 5.1 eV after acid treatment. This increase was explained in terms of a reduction in
the -conjugation of the MWNTs. 19
97
References
[1] N. Yao, V. Lordi, S. X. C. Ma, E. Dujardin, A. Krishnan, M. M. J. Treacy, and T. W. Ebbesen.
Structure and oxidation patterns of carbon nanotubes. J. Mater. Res., 13(9):24322437, 1998.
[2] E. Dujardin, T. W. Ebbesen, A. Krishnan, and M. M. J. Treacy. Purification of single-shell
nanotubes. Adv. Mater., 10(8):611613, 1998.
[3] J. Liu, A. G. Rinzler, H. Dai, J. H. Hafner, K. Bradley, P. J. Bou, A. Lu, T. Iverson, K. Shelimor,
C. B. Huffman, F. Rodriguez-Macias, Y.-S. Shan, T. R. Lee, D. T. Colbert, and R. E. Smalley.
Fullerene pipes. Science, 280:12531256, 1998.
[4] D. Tasis, N. Tagamatarchis, V. Georgakilas, and M. Prato. Soluble carbon nanotubes. Chem.
Eur. J., 9:40004008, 2003.
[5] P. M. Ajayan, T. W. Ebbesen, T. Ichihashi, S. Iijima, K. Tanigaki, and H. Hiura. Opening
carbon nanotubes with oxygen and implications for filling. Nature, 362:522525, 1993.
[6] S. C. Tsang, Y. K. Chen, P. J. F. Harris, and M. L. H. Green. A simple chemical method of
opening and filling carbon nanotubes. Nature, 372:159162, 1994.
[7] T. W. Ebbesen, P. M. Ajayan, H. Hiura, and K. Tanigaki. Purification of nanotubes. Nature,
367:519, 1994.
[8] H. Hiura, T. W. Ebbesen, and K. Tanigaki. Opening and purification of carbon nanotubes in
high yields. Adv. Mater., 7(3):275276, 1995.
[9] B. C. Satishkumar, A. Govindaraj, J. Mofokeng, G. N. Subbanna, and C. N. R. Rao. Novel
experiments with carbon nanotubes: Opening, filling, closing and functionalizing nanotubes.
J. Phys. B: At. Mol. Opt. Phys., 29:49254934, 1996.
[10] C. Niu, E. K. Sichel, R. Hoch, D. Moy, and H. Tennent. High power electrochemical capacitors
based on carbon nanotube electrodes. Appl. Phys. Lett., 70(11):14801482, 1997.
[11] J. Chen, M. A. Hamon, H. Hu, Y. Chen, A. M. Rao, P.C. Eklund, and R. C. Haddon. Solution
properties of single-walled carbon nanotubes. Science, 282:9598, 1998.
[12] K. Esumi, M. Ishigami, A. Nakajima, K. Sawada, and H. Honda. Chemical treatment of carbon
nanotubes. Carbon, 34(2):279281, 1996.
[13] R. M. Lago, S. C. Tsang, K. L. Lu, Y. K. Chen, and M. L. H. Green. Filling carbon nanotubes
with small palladium metal crystallites: the effect of surface acid groups. J. Chem. Soc., Chem.
Commum., pages 13551356, 1995.
[14] A. Bianco, K. Kostarelos, C. D. Partidos, and M. Prato. Biomedical applications of functionalised carbon nanotubes. Chem. Commun., pages 571577, 2005.
[15] L. M. Ericson, H. Fan, H. Peng, V. A. Davis, W. Zhou, J. Sulpizio, Y. Wang, R. Booker,
J. Vavro, C. Guthy, A. N. G. Parra-Vasquez, M. J. Kim, S. Ramesh, R. K. Saini, C. Kittrell,
G. Lavin, H. Schmidt, W. Wade Adams, W. E. Billups, M. Pasquali, W.-F. Hwang, R. H.
98
Hauge, J. E. Fischer, and R. E. Smalley. Macroscopic, neat, single-walled carbon nanotube

fibers. Science, 205:14471450, 2004.
[16] J.-M. Bonard, T. Stora, J.-P. Salvetat, F. Maier, T. Stockli, C. Duschl, L. Forro, W. A. de Heer,
and A. Chatelain. Purification and size-selection of carbon nanotubes. Adv. Mater., 9(10):827
832, 1997.
[18] H. Hu, P. Bhowmik, B. Zhao, M. A. Hamon, M. E. Itkis, and R. C. Haddon. Determination of
the acidic sites of purified single-walled carbon nanotubes by acid-base titration. Chem. Phys.
Lett., 345:2528, 2001.
[19] H. Ago, T. Kugler, F. Cacialli, W. R. Salaneck, M. S. P. Shaffer, A. H. Windle, and R. H.
Friend. Work functions and surface functional groups of multiwall carbon nanotubes. J. Phys.
Chem. B., 103:81168121, 1999.
[20] P. G. Collins, M. Hersam, M. Arnold, R. Martel, and Ph. Avouris. Current saturation and
electrical breakdown im multiwalled carbon nanotubes. Phys. Rev. Lett., 86(14):31283131,
2001.
[21] M. J. OConnell, S. M. Bachilo, C. B. Huffman, V. C. Moore, M. S. Strano, E. H. Haroz,
K. L. Rialon, P. J. Boul, W. H. Noon, C. Kittrell, J. Ma, R. H. Hauge, R. B. Weisman, and
R. E. Smalley. Band gap fluorescence from individual single-wall carbon nanotubes. Science,
297:593596, 2002.
[23] R. Martel, R. Herbert, H. R. Shea, and P. Avouris. Rings of single-walled carbon nanotubes.
Nature, 398:299300, 1999.
[24] P. M. Ajayan and B. I. Yakobson. Oxygen breaks into carbon world. Nature, 414:818819,
2006.
[25] S. Iijima, P. M. Ajayan, and T. Ichihashi. Growth model for carbon nanotubes. Phys. Rev.
Lett., 69(21):31013103, 1992.
[26] M. Terrones, N. Grobert, J. Olivares, J. P. Zhang, H. Terrones, K. Kordatos, W. K. Hsu, J. P.
Hare, P. D. Townsend, K. Prassides, A. K. Cheetham, H. W. Kroto, and D. R. M. Walton.
Controlled production of aligned-nanotube bundles. Nature, 388:5255, 1997.
[27] M. S. P. Shaffer, X. Fang, and A. H. Windle. Dispersion and packing of carbon nanotubes.
Carbon, 36(11):16031612, 1998.
5(9):13451363, 2005.
99
[29] H. Hu, Y. Ni, V. Montana, R. C. Haddon, and V. Parpura. Chemically functionalized carbon
nanotubes as substrates for neuronal growth. Nano Lett., 4(3):507511, 2004.
[30] J.-L. Li, K. N. Kudin, M. J. McAllister, R. K. Prudhomme, I. A. Aksay, and R. Car. Oxygendriven unzipping of graphitic materials. Phys. Rev. Lett., 96:176101, 2006.
[31] R. Zacharia. Desorption of gases from graphitic and porous carbon surfaces. Dissertation,
Freie Universitat Berlin, 2004.
[32] E. V. Basiuk, V. A. Basiuk, J.-G. Banuelos, J.-M. Saniger-Blesa, V. A. Pokrovskiy, Y. T. Gromovoy, A. V. Mischanchuk, and B. G. Mischanchuk. Interaction of oxidized single-walled
carbon nanotubes with vaporous amines. J. Phys. Chem. B., 106:15881597, 2002.
[34] S. Latil, S. Roche, D. Mayou, and J.-C. Charlier. Mesoscopic transport in chemically doped
carbon nanotubes. Phys. Rev. Lett., 92(25):256805, 2004.
[35] M. P. Anantram and T. R. Govindan. Conductance of carbon nanotubes with disorder: A
numerical study. Phys. Rev. B., 58(8):48824887, 1998.
[36] V. H. Crespi, M. L. Cohen, and A. Rubio. In situ band gap engineering of carbon nanotubes.
Phys. Rev. Lett., 79(11):20932096, 1997.
[37] H. Pan, Y. P. Feng, and J. Y. Lin. Ab initio study of OH-functionalized single-wall carbon
nanotubes. Phys. Rev. B, 70:2454251, 2004.
[38] M. Valcarel, B. M. Simonet, S. Cadenas, and B. Suarez. Present and future applications of
carbon nanotubes to analytical science. Anal. Bioanal. Chem., 382:17831790, 2005.
[39] H. Dai, E. W. Wong, and C. M. Lieber. Probing electrical transport in nanomaterials: conductivity of individual carbon nanotubes. Science, 272:523526, 1996.
[40] K. Mlhave, Gudnason. S. J., A. T. Pederson, C. H. Clausen, A. Horsewell, and P. Bgglid.
Transmission electron microscopy study of individual carbon nanotube breakdown caused by
Joule heating in air. Nano Lett., 6(8):16631668, 2006.
100
Chapter 8
Development of a CNx MWNT

Biosensing Device
8.1 Previous Nanotube Sensing Work
8.1.1 Small Molecules
SWNTs have attracted much attention as chemical sensors for small gas molecules such as carbon
dioxide, 1,2 nitric oxide, 35 ammonia, 46 fluorine, 7 and oxygen. 8 The basis of the sensing, is that on
exposure to certain gas molecules, characteristic changes in the electrical properties of the SWNTs
occur.
8.1.2 Sensing of Proteins

The compatibility of CNTs with biological molecules, 9 has also led to the development of sensors
for biomolecules. The first report of a CNT biosensor was by Sotirpoulou et al. 10 who fabricated
a glucose sensing device via covalent functionalization of aligned MWNTs. Since then there have
been many reports of un-doped CNT devices for molecular recognition 1115 including a device for
the detection of metalloproteins. 16
101
Development of a CNx MWNT Biosensing Device
8.1.3 Sensing with CNx MWNTs

CNx MWNT films have been employed for electrochemical, 17 gas, and vapor sensing, 18 but the
potential of CNx MWNTs for biosensing 19 has not yet been explored.
The mechanism of sensing for CNx MWNTs is likely to be different to that for un-doped CNTs
due to inhomogeneously distributed nitrogen dopants within the tube wall. These dopants (pyridinelike with a lone pair which can bind to incoming species. 18,19 ) provide a way to activate regions
along the tube wall, 20 and have been shown to facilitate electron transfer between metalloproteins
and Au electrodes. 21,22
In Chapter 5 CNx MWNTs were functionalised with cyt c, ferritin and azurin without decoiling
or denaturation of the proteins, suggesting that CNx MWNTs are suitable components for bioapplications involving these proteins. The electronic transport properties of CNx MWNTs e.g. high
conductances and large current threshold, are also highly desirable for device fabrication (Chapter 4). The following work combines these two factors in the development of a CNx MWNT
biosensor device, which can sense ferritin, cyt c and azurin. Consistent with previous results for
un-doped CNTs, a decrease in electrical conductance was observed upon protein binding. Specific
sensing was achieved by adding a further layer of the antibodies anti-ferritin and anti-cyt c to their
respective antigens. Finally the apoproteins apoferritin and apoazurin were used to ascertain the
contribution of the metal groups to the sensing mechanism.
8.2 Experimental
The scheme for fabrication of the CNx MWNT biosensing device is shown in fig. 8.1. In the first
step, raw CNx MWNTs were sonicated in EtOH and then dispersed as a random network onto glass
substrates [fig. 8.1(a)]. An electrode gap was generated using a PTFE mask (b) and Au electrodes
were fabricated using electroless deposition (c).
A concentrated solution of CNx MWNTs was dispersed in EtOH by sonication. 100 L of this
solution was then deposited onto a glass substrate and dried on a hotplate. A 0.5 mm thick strip of
PTFE tape was placed over the deposit to form a gap and the substrate immersed in 5 mM HAuCl4
102
Figure 8.1: Scheme showing the fabrication steps for the biosensing device; (a) CNx MWNTs are
dispersed in ethanol on a glass substrate, (b) a PTFE mask is applied to create an electrode gap, (c)
after electroless Au plating the un-masked CNx MWNTs are coated in Au nanoparticles
solution for 60 s. After immersion, the substrates were rinsed thoroughly with MilliQ water. Plating
on CNx MWNTs and not in the gap was confirmed by AC-AFM imaging [fig. 8.2(a)]. HAuCl4 is a
yellow crystalline solid that forms a bright gold solution [fig. 8.2(b)] and during the process visible
Au plating was evident.
8.2.1 Electroless Plating

Au electrodes for CNT devices can be fabricated using various methods including metal evaporation, 11 attachment of nanoparticles after oxidation, 23 electron-beam-deposition, 14 electrolysis 24
and spontaneous electroless plating. 25
Electroless metal deposition was developed by Brenner and Riddell 26 and involves the continuous build up of metal on a substrate which has been immersed in a suitable electrolyte. Electrodes
and a current source are not required because a direct redox reaction occurs between the metal ions
and the substrate. However, only metal ions with a redox potential higher than that of the substrate
will plate out of the electrolyte solution. 25,27
Although the redox potential of CNx MWNTs has not been measured, the redox potential for
un-doped SWNTs is known to be + 0.500 V vs SHE (standard hydrogen electrode). 25 Thus only
metals with a redox potential more positive than + 0.500 V will spontaneously deposit e.g. Au and
Pt, from the table of E values.
In this case, the substrate is the CNx MWNTs and the electrolyte is 5 mM HAuCl4 solution. The
deposition can be regarded as two half-equations, one for the reduction of Au ions out of solution:
103
Figure 8.2: (a) AC-AFM phase image of gold nanoparticles plated on CNx MWNTs, (b) 5 mM
HAuCl4 solution
Au3 aq 3e
Au s
(8.1)
CNx MWNT aq e
(8.2)
and another for the oxidation of CNx MWNTs:

CNx MWNT aq
System
E (vs SHE)
SWNT
+0.500V
Cu(NO3 2 /Cu
+0.340 V
Ag(NH3 2 /Ag
+0.373 V
AuCl
4 /Au
+1.002 V
PtCl24 /Pt
+0.775 V
8.2.2 Electrical Measurement Set-up

Two-probe electrical conductivity measurements were performed using a Keithley 6340 sub-femtoamp
sourcemeter. The sourcemeter was connected to the probes (or needles) via the scheme shown in
fig. 8.3. The set-up was enclosed in a grounded Faraday cage [fig. 8.4(a)] with all the connection
to the outside through a second cage. The probes [fig. 8.4(b)] were illuminated with PHOTONIC
PL2000 fibre optics. An optical microscope (Leica) with a digital camera, operated with the IM50
104
Figure 8.3: The box connection scheme (Kislon Votchovsky)
software, was used to locate the probes, which were then positioned over the devices using micromanipulators. The I-t and I-V data acquired were recorded using a Labview program. Background
noise current measurements were taken before each device measurement, and were very low, approximately 1 fA [fig. 8.5(a)].
For biosensing measurements the devices were first immersed in phosphate buffer (containing
5 mM NaH2 PO4 and 5 mM NaCl, pH 6.01) inside a small plastic cell. When a stable current had
been established, 1 M protein solutions in the same buffer were then introduced to the solution and
Figure 8.4: (a) The set-up encased in a Faraday cage, (b) the probes (Kislon Votchovsky)
105
Figure 8.5: (a) A typical I-V background measurement and (b) control sensing I-t measurement
where buffer only is added (indicated by an arrow)
the response recorded.
8.3 Results
I-V curves were taken at various points during the plating process; the glass substrates had current
10 pA, the unplated CNx MWNTs a current of
1 nA, and the completed device a current of
1 mA.
When the devices were immersed in buffer an immediate drop in current was observed from
1 mA to
100 A. When this lower value was stable, protein solutions were added to the buffer
using a micropipette, and the subsequent decline in current observed. Control experiments where
buffer only was added to the solution [fig. 8.5(b)] showed no response.
8.3.1 Sensing of Ferritin

The change in I-V characteristics for a typical device exposed to ferritin is shown in fig. 8.6(a). The
dry devices exhibited linear I-V characteristics (red) and high conductances as expected for CNx
MWNTs. After immersion in phosphate buffer (blue) and exposure to ferritin (black), the current
carried by the device was significantly reduced. The response to ferritin was also observed in real106
Figure 8.6: (a) I-V curves for a dry device (red), after immersion in 5 mM phosphate buffer (blue),
and after the introduction of 1 M ferritin (black), (b) real time I-t measurement for a device exposed
to 1 M ferritin
time I-t measurements [fig. 8.6(b)]. A steady current of 90 A is observed until 1 M ferritin is added
to the solution (indicated by an arrow). Immediately after the protein (
20 L) is introduced, the
conductance of the device begins to decrease, before eventually reaching a stable value.
Control experiments with apoferritin

In order to elucidate the role of the internal Fe atoms in the response mechanism, sensing experiments were repeated with 1 M apoferritin. The I-V characteristics after exposure to apoferritin
[fig. 8.7 (a)] are very similar to those observed for ferritin [fig. 8.6(a)]. Real-time measurements
of the response of the devices to apoferritin were also recorded [fig. 8.7 (b)]. 1 M apoferritin was
introduced to the buffer at 250s (indicated by an arrow) at which point the current began to drop
from 100 A before stabilizing at 60 A.
The response of the device to apoferritin indicates that the metal group of ferritin does not
participate in the sensing mechanism.
Additional sensing of anti-ferritin

The adsorbed layer of ferritin on the CNx MWNT device was then used to detect a second protein.
As the second layer must have an affinity for the first, 28 the IgG antibody anti-ferritin (which was
shown to bind to adsorbed ferritin in Chaper 5) was chosen.
107
and after the introduction of 1 M apoferritin (black), (b) real time I-t measurement for a device
exposed 1 M apoferritin
As previously, 1 M ferritin was initially added to a device (fig. 8.8) producing a reduction in
current from 100 A to 50 A. Once this smaller current had stabilized, anti-ferritin was added,
whereupon a second signal change was observed (fig. 8.8), indicating successful binding of antiferritin to pre-adsorbed ferritin. 28
8.3.2 Sensing of cyt c

Similar behaviour is observed when the devices are exposed to cyt c. Fig. 8.9(a) shows typical I-V
characteristics for a device and its response to 1 M cyt c. The real-time response to 1 M is shown
in fig. 8.9(b). The device has a current of 120 A after immersion in buffer, then 1 M cyt c is added
at 150 s (indicated by an arrow) leading to decline in current to 60 A.
Experiments with apoproteins could not be applied to cyt c because apocyt c was not available.
Specific sensing of anti-cyt c

As was the case with ferritin, adsorbed cyt c was found to detect its IgG antibody (fig. 8.10). At 80
s a solution of 1 M cyt c was added to the device (indicated by an arrow) producing a decrease in
current from 120 A to 60 A. Then as a test for specificity, anti-ferritin was added to the device at
150 s. After no discernible response was observed, anti-cyt c was introduced at 450 s. This caused
a decrease in current from 50 A to 30 A.
108
Figure 8.8: Real time I-t response of a device to ferritin and then a second response to its specific
antibody antiferritin
and after the introduction of 1 M cyt c (black), (b) real time I-t measurement for a device exposed
to 1 M cyt c
109
Figure 8.10: Real time I-t response of a device to cyt c followed by a lack of response to anti-ferritin
and a second response to its specific antibody anti-cyt c
8.3.3 Sensing of azurin

The CNx MWNT device also showed a response to azurin (fig. 8.11 (a)) similar to that observed for
other metalloproteins. A stable current of 110 A was established in buffer before 1 M azurin was
added to the device (arrow). This caused a decline in current to 40 A.
Control experiments with apoazurin

Apoazurin was prepared as in section 5.3.3 and added to a device [fig. 8.11(b)] at 30 s. The response
of the device to apoazurin was very similar to its response to azurin, causing a decrease in current
from 110 A to 30 A.
8.4 Discussion
A protein binding induced decline in current was observed for
40 devices, but the initial current
(in buffer) and percentage decreases vary greatly across the devices. This is probably due to slight
differences in fabrication as the devices are made by hand.
Fig. 8.12(a) shows that initial currents as high as 4 mA were observed for biosensor devices.
However the most common initial current was in the range 0 - 500 A. Within this range (inset)
110
Figure 8.11: Real time I-t response of CNx MWNT devices to (a) 1 M azurin and (b) 1M
apoazurin
Figure 8.12: Histograms showing the range of (a) initial current values observed after the addition of
protein for the biosensor devices (inset 0 - 500 A range) and (b) the drop observed as a percentage
of the initial current
111
the majority of devices exhibited an initial current between 0 - 200 and these devices were used
to record the results shown on the previous pages. Fig. 8.12(b) demonstrates that the percentage
current drop was subject to more variation than the initial current, but that an average percentage
drop of 43.5 23.6 % was observed over all devices.
Binding of the proteins to the CNx MWNT devices was irreversible, as the device current did not
recover upon rinsing. The results shown that CNx MWNT devices can be used to detect clinically
and biologically important interactions between antibodies and antigens. Biosensing experiments
were not conducted with ferredoxin because in Chapter 5 it was not shown to remain correctly
folded when bound to CNx MWNTs.
8.4.1 Sensing Mechansim

The mechanism responsible for change in conductance of a SWNT device upon exposure to proteins
is subject to much debate. One suggestion is that similar to the mechanism for gas sensing, 4 the
response is due to charge-transfer. 15 Star et al. suggested that charge flows from adsorbed amine
groups producing a response in their devices. 13
Any charge transfer between the functionalising molecules (i.e. proteins) and the CNT is related
to the capacitance of the devices and the diameter and the length of the CNT. Peng et al. 19 suggested
that doped CNTs would have an increased charge-transfer capability as compared to that of an
intrinsic CNTs.
However Chen et al. sensed both positively and negatively charged proteins 11 and all produced
the same decrease in conductance. Another explanation could be that the binding of biomolecules
causes geometrical deformations which act as scattering sites on the CNT. 12,15 This would lead to a
decrease in charge mobility, thus suppressing the current of the device.
Effect of Dielectric Constant

A third possibility is that protein adsorption on the devices affects the dielectric constant of the
electrical double layer (see Appendix section 10.4) in aqueous solution. 28 In a systematic investigation by Chen et al. the proteins BSA, HSA and avidin were adsorbed to SWNT devices where
112
the electrodes had been functionalized with a protein-repelling polymer. In this case, no decrease in
conductance was observed indicating that protein adsorption at or near the metal-nanotube contacts
appears to be the key to conductance modulation. 28
Role of Metal Groups

Metalloproteins in solution are able to communicate with SWNTs in the form of an electrode. It
therefore follows that they may be able to communicate electronically (via a redox-active centre)
with the delocalised system of a CNT. 29 This hypothesis has not been tested until now and experiments with both apoazurin and apoferritin demonstrate that the metal group does not participate
in the sensing response. This result may be expected for ferritin as it is an iron-storage protein
only and not known to act as an electron donor. 30 Azurin however is known to act as an electron
donor in some situations, 31 so the likelihood of metal group participation in the sensing mechanism
is increased. From the results of experiments with apoproteins it appears that only the outer shell
participates in the response of the sensor.
Performance Limits
All nanobiosensors are subject to finite performance limits of the concentration of biomolecules that
can be detected under reasonable settling times. For a diffusion limited regime, when a sensor is
inserted into the analyte, the time required to capture a certain number of analyte molecules is given
by ts . This response time also depends on the dimensionality of the sensor such that:
0 ts
kD
(8.3)
where 0 is the analyte concentration and kD is a dimensionality dependent constant. 32 As such,

a one-dimensional sensor (such as a planar ion-sensitive FET) with ts = 100 s can detect down to
nM levels, whereas a two-dimensional sensor (such as a cylindrical CNT) device can detect down
to 100 fM concentration. 32
113
References
[1] A. Star, T.-R. Han, V. Joshi, J.-C. P. Gabriel, and G. Gruner. Nanoelectronic carbon dioxide
sensors. Adv. Mater., 16(22):20492052, 2004.
[2] A. Zribi, A. Knobloch, and R. Rao. CO2 detection using carbon nanotube networks and
micromachined resonant transducers. Appl. Phys. Lett., 86:203112, 2005.
[3] P. Qi, O. Vermesh, M. Grecu, A. Javey, Q. Wang, and H. Dai. Toward large arrays of multiplex
functionalized carbon nanotube sensors for highly sensitive and selective molecular dectection.
Nano Lett., 3(3):347351, 2003.
[4] J. Kong, N. R. Franklin, C. Zhou, M. G. Chapline, S. Peng, K. Cho, and H. Dai. Nanotube
molecular wires as channels sensors. Science, 287:622625, 2000.
[5] H. Chang, J. D. Lee, S. M. Lee, and Y. H. Lee. Adsorption of NH3 and NO2 molecules on
carbon nanotubes. Appl. Phys. Lett., 79(9):38633865, 2001.
[6] K. Bradley, J.-C. P. Gabriel, M. Briman, A. Star, and G. Gruner. Charge transfer from ammonia
physisorbed on nanotubes. Phys. Rev. Lett., 91(21):218301, 2003.
[7] E. T. Mickleson, C. B. Huffman, A. G. Rinzler, R. E. Smalley, R. H. Hauge, and J. L. Margave.
Flourination of single-walled carbon nanotubes. Chem. Phys. Lett., 296(1-2):188194, 1998.
[8] P. G. Collins, K. Bradley, M. Ishigami, and A. Zettl. Extreme oxygen sensitivity of electronic
properties of carbon nanotubes. Science, 287:18011803, 2000.
[9] Y. Lin, S. Taylor, H. Li, K. A. Shiral Fernando, L. Qu, W. Wang, L. Gu, B. Zhou, and Y.-P.
Sun. Advances toward bioapplications of carbon nanotubes. J. Mater. Chem., 14:527541,
2004.
[10] S. Sotiropoulou and N. A. Chaniotakis.
Bioanal. Chem., 375:103105, 2003.
Carbon nanotube array-based biosensor.
Anal.
[12] G. Gruner. Carbon nanotube transistors for biosensing applications. Anal. Bioanal. Chem.,
384:322335, 2006.
Lett., 4(2):253256, 2004.
[14] K. Besteman, J.-O. Lee, F. G. M. Wiertz, H. A. Heering, and C. Dekker. Enzyme-coated
carbon nanotubes as single-molecule biosensors. Nano Lett., 3(6):727730, 2003.
[15] A. Star, J.-C. P. Gabriel, K. Bradley, and G. Gruner. Electronic detection of specific protein
binding using nanotube FET devices. Nano Lett., 3(4):459463, 2003.
114
[16] S. Boussaad, N. J. Tao, R. Zhang, T. Hopson, and L. A. Nagahara. In situ detection of cytochrome c adsorption with single walled carbon nanotube device. Chem. Commun., pages
15021503, 2003.
[17] W.-C. Fang, C.-L. Sun, J.-H. Huang, L.-C. Chen, O. Chyan, K.-H. Chen, and P. Papakonstantinou. Enhanced electrochemical properties of arrayed CNx nanotubes directly grown on
Ti-buffered silicon substrates. Electrochem. Solid-State Lett., 9(3):175178, 2006.
[18] F. Villalpando-Paez, A. H. Romero, E. Munoz Sandoval, L. M. Martnez, H. Terrones, and
M. Terrones. Fabrication of vapor and gas sensors using films of aligned CNx sensors. Chem.
Phys. Lett., 386:137143, 2004.
[19] S. Peng and K. Cho. Ab initio study of carbon nanotube sensors. Nano Lett., 3(4):513517,
2003.
5(9):13451363, 2005.
Acc. Chem. Res., 21:407413, 1988.
[24] T. M. Day, P. R. Unwin, N. R. Wilson, and J. V. Macpherson. Electrochemical templating of
metal nanoparticles and nanowires on single-walled carbon nanotube networks. J. Am. Chem.
Soc., 127:1063910647, 2005.
[25] H. C. Choi, M. Shim, S. Bangsaruntip, and H. Dai. Spontaneous reduction of metal ions on
the sidewalls of carbon nanotubes. J. Am. Chem. Soc., 124:90589059, 2002.
[26] A. Brenner and G. E. Riddell. Electroless plating by a process of controlled self-continuing
reduction. Proc. Am. Electroplat. Soc., 33:1623, 1946.
[27] S. Down. CNTs receive support for metallisation. Chemistry World, 2(9):25, 2005.
[28] R. J. Chen, H. C. Choi, S. Bangsaruntip, E. Yenilmez, X. Tang, Q. Wang, Y.-L. Chang, and
H. Dai. An investigation of the mechanisms of electronic sensing of protein adsorption on
carbon nanotube devices. J. Am. Chem. Soc., 126:15631568, 2004.
[29] J. J. Davis, K. S. Coleman, B. R Azamian, C. B. Bagshaw, and M. L. H. Green. Chemical and
biochemical sensing with modified single wall carbon nanotubes. Chem. Eur. J., 9:37323739,
2003.
[30] G. C. King, S. H. Choi, S.-H. Chu, J.-W. Kim, Y. Park, P. Lillehei, G. D. Watt, R. Davis, and
J. N. Harb. Development of a bio-nanobattery for distributed power storage systems. Technical
report, NASA, 2004.
115
[31] H. Nar, R. Huber, A. C. Messerschmidt, M. Barth, M. Jaquinod, M. van der Kamp, and G. W.
Canters. Characterization and crystal structure of zinc azurin, a by-product of heterologous
expression in Escherichia coli of Pseudomonas aeruginosa copper azurin. Eur. J. Biochem.,
205:11231129, 1992.
[32] P. R. Nair and M. A. Alam.
88:233120, 2006.
Performance limits of nanobiosensors.
116
Appl. Phys. Lett.,
Chapter 9
Conclusions and Further Work

9.1 Conclusions
9.1.1 Electrical Conductance of CNx MWNTs
In Chapter 4 the electrical conductance of individual CNx MWNTs was measured and found to
have a mean value of 1.0 0.3 G0 consistent with experiments for un-doped MWNTs. This result
had not been achieved previously, as all measurements of the conductance of CNx MWNTs in the
literature 1,2 reported very low conductances. The I-V characteristics of the CNx MWNTs were
measured and found to be linear, and electrical breakdown of individual CNx MWNTs resulted in
a series of discrete steps corresponding to the breakdown of individual MWNT layers. The steps
observed were of unequal sizes, in contrast to the equal steps observed for un-doped MWNTs. 3
This is probably due to different concentrations of N in each layer. These measurement were taken
at NPL during January 2006.
9.1.2 Biofunctionalisation of CNx MWNTs

In Chapter 5 CNx MWNTs were functionalised with four different metalloproteins; ferritin, ferredoxin, cyt c and azurin. For the proteins with characteristic UV-vis and CD spectra (cyt c and
azurin) results before and after adsorption showed that the proteins remained correctly folded when
attached to CNx MWNTs. The immunoactivity of ferritin and cyt c was investigated with an an117
tibody binding assay for each protein. All bio-CNT conjugates were imaged using AC-AFM. The
apoproteins of ferritin and azurin were introduced for later biosensing measurements. An in-depth
investigation into conformational changes of metalloproteins bound to CNTs was given, which had
not previously been reported.
9.1.3 Thermal Conductance of CNx MWNTs

In Chapter 6 the thermal conductance of raw CNx MWNTs was measured using an AFM method.
This work was done at NPL during July and August 2006. Simultaneous electrical and thermal
conductance steps in the measurements were observed when a flake of CNx MWNTs was moved
away from, and toward the substrate. The values of these steps (for many measurements) were
plotted, and the graph showed a linear trend with mean slope 0.1445 W/K. From this measurement, the thermal conductivity () of raw CNx MWNTs was calculated to be
240 W/mK. This is
significantly lower than the values previously observed for individual un-doped MWNTs, but similar to values of for mats of aligned un-doped MWNTs containing many defects such as atomic
vacancies. A value for for CNx MWNTs has not been reported previously.
9.1.4 The Effect of Acid Treatment on CNx MWNTs

In Chapter 7 the effect of acid oxidation on CNx MWNTs was investigated. The length, solubility,
percentage of acid groups and electrical and thermal properties were investigated after acid oxidation and compared with measurements taken on raw CNx MWNTs. The CNx MWNTs were found
to shorten after 2 h acid oxidation at room temperature, and the percentage of acid sites was measured by titration to be 30-35%. The acid-treated CNx MWNTs had poor electrical conductivity
compared to raw CNx MWNTs, and an estimate for for acid-treated CNx MWNTs could not be
calculated. More gentle oxidation conditions were required for CNx MWNTs than those reported
previously 46 for un-doped MWNTs. This is due to the higher number of defects present in the
sidewalls of CNx MWNTs which act as preferential etching sites over pentagon and heptagons.
118
9.1.5 Development of a CNx MWNT Biosensing Device

Chapter 8 describes the development of a novel biosensor device base on CNx MWNTs. A network
of CNx MWNTs is contacted by Au electrodes deposited via an electroless plating method. The
devices can sense the binding of ferritin, azurin and cyt c and also the additional binding of antiferritin and anti-cyt c. The sensing of apoferritin and apoazurin demonstrate that the metal centres
of ferritin and azurin do not make a significant contribution to the sensing mechanism.
9.2 Further Work

9.2.1 Control Experiments with un-doped MWNTs
As a control for the electrical and thermal conductance measurements on CNx MWNTs, they were
repeated with a flake of un-doped MWNTs [fig. 9.1(a)] synthesised under the same experimental
conditions (section 2.3.1).
However closer inspection of the flake [fig. 9.1(b)] of un-doped MWNTs revealed the presence a
large number of Fe nanoparticles at the tips of the individual MWNTs, not present in CNx MWNTs.
The contribution of these Fe nanoparticles to the electrical and thermal conduction measurements
is unknown, therefore the data recorded cannot be taken as reliable control sample.
Due to the Fe nanoparticles, these un-doped MWNTs display remarkable magnetic properties
which may have future applications. In the meantime it would be interesting to image the un-doped
MWNTs with Magnetic Force Microscopy (available at NPL).
9.2.2 Further Thermal Measurements

In Chapter 6 the thermal conductivity of CNx MWNTs was calculated to be
240 W/mK at room
temperature. It would interesting to determine the dependence of on temperature, as it has been

shown to have a marked dependence for un-doped CNTs. 7
119
Figure 9.1: (a) SEM image of a flake of un-doped CNx MWNTs mounted on a thermal probe and
(b) high magnification SEM image of the same flake showing Fe nanoparticles incorporated in the
individual MWNTs (David Cox)
9.2.3 Atomic Resolution of CNx MWNTs

Some STM images, thought to show atomic resolution of CNx MWNTs (fig. 9.2) have been reported. 8 However, more detailed imaging is required to confirm the presence of multiple defects,
which are used to explain many of the properties observed in this thesis. Using high resolution,
ultra-high vacuum STM (at NPL) atomic resolution of the CNx MWNTs may be possible.
9.2.4 Improved Biosensor Device Fabrication

The biosensor devices in Chapter 8 respond to the adsorbance of metalloproteins. However, the
I-V characteristics and response of each individual device is slightly different because the devices
are currently made by hand. In order to overcome this, and achieve a reproducible response across
different devices, automated device fabrication is required.
The effect of ionic strength on biosensor performance will be investigated (Lydia Harriss) by
systematic experiments using different buffer concentrations. A series of control experiments using
120
Figure 9.2: High magnification STM image of an individual CNx MWNT from Czerw et al. 8
the same fabrication protocol but un-doped MWNTs will also be carried out.
9.2.5 Further Biomolecule Functionalisation

In the early part of this project additional (non-metalloprotein) biomolecules were combined with
CNx MWNTs, including bacteriorhodopsin (fig. 9.3) and DNA. However, whether these molecules
remain biologically active or not when bound, was never ascertained. In addition, the biosensor
device was not tested for a response to these molecules.
9.2.6 Inter-digitated Electrodes

CNx MWNTs were dispersed on inter-digitated electrodes fabricated via electron beam deposition
by David Cox (University of Surrey and NPL). The electrodes consist of six Au fingers [fig. 9.4(a)]
800 nm in width, each separated by a gap of
150 nm [fig. 9.4(b)] mounted on a Si chip.
Several CNx MWNTs from a 1 ng/mL EtOH solution were found (by SEM imaging) to lie across
the electrode gaps (fig. 9.5), with further AFM imaging revealing multiple CNx MWNTs on each
chip. Individual CNx MWNTs now need to be electrically isolated to allow probe measurements to
be taken.
121
Figure 9.3: AC-AFM amplitude image of an individual CNx MWNT on top of the hexagonal layers
of bR
Figure 9.4: SEM images of (a) electrodes fabricated by EBD and (b) close-up of the same electrodes
showing the dimensions of the electrode gap (David Cox)
122
Figure 9.5: SEM image of an individual CNx MWNT (between the green lines) lying across two
electrodes (David Cox)
References
2005.
[3] P. G. Collins, M. S. Arnold, and P. Avouris. Engineering carbon nanotubes and nanotube circuits
using electrical breakdown. Science, 292:706, 2001.
[4] H. Hiura, T. W. Ebbesen, and K. Tanigaki. Opening and purification of carbon nanotubes in
high yields. Adv. Mater., 7(3):275276, 1995.
[5] E. Dujardin, T. W. Ebbesen, A. Krishnan, and M. M. J. Treacy. Purification of single-shell
nanotubes. Adv. Mater., 10(8):611613, 1998.
[6] K. Esumi, M. Ishigami, A. Nakajima, K. Sawada, and H. Honda. Chemical treatment of carbon
nanotubes. Carbon, 34(2):279281, 1996.
[8] R. Czerw, M. Terrones, J.-C. Charlier, X. Blase, B. Foley, R. Kamalakaran, N. Grobert, H. Terrones, P. M. Ajayan, W. Blau, D. Tekleab, M. Ruhle, and D. L. Carroll. Identification of electron
donor states in N-doped carbon nanotubes. Nano Lett., 1(9):457460, 2001.
123
Chapter 10
Appendix
10.1 Acid-base Titration Calculation
In section 7.2.4 acid-treated CNx MWNTs were stirred with 0.05 M NaOH for 48 h and then titrated
against 0.05 M HCl.
The CNx MWNTs were stirred in 4 mL of 0.05 M NaOH solution. Calculate the total moles
of NaOH:
concentration
moles
volume L
moles
0 05
4
1000
2 104 mol
(10.1)
In this example, 1900 L of 0.05 HCl were required to neutralise any excess NaOH. Calculate
the moles of HCl required:
concentration
moles
volume L
moles
0 05 1900 106
9 5 105 mol
(10.2)
This value equals the number of moles of excess NaOH. Thus the number of moles of COOH
groups on the CNx MWNTs is:
Total moles of NaOH - excess moles of NaOH = 2104 - 9.5105 = 1.05104 mol
124
Appendix
Moles of CNx MWNTs

This sample contained 0.00381 g of lyophilized acid-treated CNx MWNTs, however the concentration of N in these CNx MWNTs is unknown (2-5%). An approximation of 3.5% N will be used for
the purpose of these calculations.
Calculate the molar mass of CN0 035 , molar mass = 12 + (0.035)15 = 12.525 gmol .
Calculate the number of moles of CNx MWNTs in 0.00381 g:
moles
mass
Mr
0 00381
12 525
3 04 104 mol
(10.3)
Atoms of CNx MWNTs

Using the Avogadro constant NA
atoms
6 02 1023 mol1 we can calculate the number of atoms:
3 04 104 6 022 1023
1 83 1020 atoms
(10.4)
Percentage of C atoms oxidized

The number of moles of COOH groups was calculated above to be 1.18104 calculate what percentage this is of the total number of C and N atoms.
The number of C atoms oxidised to COOH is:

1 05 104 6 022 1023
6 32 1019
(10.5)
Percentage of C atoms oxidised is:

6 32 1019
1 83 1020
100
125
34 5%
(10.6)
Appendix
10.2 CNx MWNT Length and Diameter Data

Length and diameter data were taken from AC-AFM images using Nanoscope software. Due to tip
convolution the height of the CNT was taken to be a more accurate measure of the diameter, than
its width.
Raw
acid-treated at 25C
acid-treated at 60C
l (nm)
d (nm)
l (nm)
d (nm)
l (nm)
d (nm)
325
18
237
15
203
10
367
25
342
10
304
10
388
32
345
14
376
12.5
418
24
386
12
389
10
422
25
391
20
401
12
545
40
403
20
426
10
558
35
435
457
13
580
23
439
22
471
603
19
448
12
477
16
633
32
453
19
482
10
685
25
454
10
491
12
732
10
469
23
517
13
764
11
470
10
536
12
830
21
495
20
540
860
35
508
11
559
880
14
516
10
590
15
895
530
12
611
15
922
19
545
20
626
12
929
22
550
13
654
8.5
950
12
559
19
674
11
126
Appendix
Raw
acid-treated at 25C
acid-treated at 60C
l (nm)
d (nm)
l (nm)
d (nm)
l (nm)
d (nm)
957
17
578
13
687
11
959
21
585
10
699
962
29
586
10
710
980
24
589
749
989
16
625
11
775
1010
20
626
12
788
1010
16
648
20
791
1030
20
657
13
813
1040
16
681
19
821
1040
31
699
13
999
1050
30
700
10
1000
1050
20
716
10
1007
1080
20
735
1017
1100
20
759
11
1040
1140
18
739
12
1050
1148
14
749
20
1100
1160
16
749
13
1167
10
751
16
1170
23
758
11
1180
23
781
1190
10
781
13
1200
10
785
1220
30
788
1280
12
796
1310
32
802
1380
17.5
806
127
Appendix
Raw
acid-treated at 25C
acid-treated at 60C
l (nm)
l (nm)
d (nm)
l (nm)
1380
25
815
1380
26
822
1427
29
822
1430
30
835
1460
25
851
1509
42
871
1600
18
881
1612
14
904
1670
911
1671
20
915
1682
24
926
1738
24
943
1780
949
1791
955
1820
962
1840
973
1893
979
1990
996
2014
1000
2020
1010
2054
1010
2058
1010
2112
1010
2170
1010
2272
1015
2280
1040
d (nm)
128
d (nm)
Appendix
Raw
l (nm)
d (nm)
acid-treated at 25 C
l (nm)
2310
1040
2387
1047
2448
1070
2450
1080
2560
1082
2621
1090
2700
1090
2716
1100
2720
1120
2752
1150
2780
1160
2870
1170
2950
1171
2977
1171
3148
1180
3264
1180
3343
1220
3350
1230
3364
1240
3400
1270
3440
1280
3450
1290
3450
1330
3552
1350
3585
1370
3622
1400
129
d (nm)
Appendix
Raw
l (nm)
d (nm)
l (nm)
3710
1420
3769
1430
3850
1539
3910
1580
3926
1620
4080
1680
4202
1756
4202
1768
4224
1820
4240
1830
4296
1847
4307
1880
4400
1970
4550
2120
4970
2155
5030
2155
5057
2209
5060
2463
5270
2561
5300
2570
5370
2744
5450
2770
5450
2960
5450
3070
5470
3078
5510
3080
130
d (nm)
Appendix
Raw
l (nm)
d (nm)
l (nm)
5860
3290
5873
3386
5950
3708
6219
3963
d (nm)
10.3 Hybridisation in Carbon Materials

10.3.1 Atomic Orbitals
An atomic orbital (AO) is the probability distribution of an electron in that atom. AOs are derived
from quantum mechanics and represent a three dimensional volume (i.e. the region in space) in
which an electron is most likely to be found around an atom. AOs cannot be observed experimentally.
The orbital is derived from the wavefunction of the atom , which can be used to define the
radial distribution function P(r):
Prdr
d
0
d r2 sin dr
(10.7)
This is the probability of finding an electron at a point (r, ) in a spherical shell of thickness dr.
Where r is the radius, is the colatitude, is the azimuth (known as spherical polar coordinates).
Solving this integral:
Prdr
4r2
(10.8)
AOs are denoted by the notation ns, np, nd, nf. For example, in atomic hydrogen, the only
occupied orbital is the 1s orbital which is spherical in shape. Higher energy orbitals consist of lobes
and rings in accordance with the quantum numbers that define them. These quantum numbers are:
n - the principal quantum number which denotes the energy of the orbital.
131
Appendix
Figure 10.1: Hybridisation of a spherical s and a px orbital to produce an sp orbital
l - the orbital quantum number which denotes the number of angular nodes the orbital will
have. For various historical reasons a value of l=0 is designated s, and a value of l=1 is
designated p.
Thus for the carbon 1s orbital n=1 and l=0 hence it contains no node, so is spherical in shape.
Conversely for the carbon 2p orbital n=2 and l = 1 so the orbital contains one node in its shape
(fig. 10.1).
10.3.2 The Hybrid Orbital Model

The hybrid orbital model, developed by Pauling, attempts to explain bonding in molecules, by using
a combination of orbitals on the individual atoms. Bonding results when AOs on adjacent atoms,
mix. In hybridisation, AOs of differing energies combine to form a new orbital whose energy is
determined by a weighted average. These new orbitals are known as hybrids.
Consider forming a hybrid by mixing the 2s orbital with the 2px orbital. The result is an sp
hybrid orbital (fig. 10.1) that has lower energy to the individual s and p orbitals but a different
spatial distribution.
In addition to the sp hybrid there are also two un-hybridized p orbitals which form two 132
Appendix
Figure 10.2: (a) Hybridisation of the s and the px and py orbitals to produce an sp2 orbital and (b)
the bonding in graphite
bonds. sp2 hybridisation [fig. 10.2(a)] combines one s and two p orbitals resulting in three equivalent
hybrids directed to the corners of an equilateral triangle. This leaves one p orbital perpendicular to
the plane of the hybrids, to give a planar geometry. This is the kind of hybridisation that occurs in
the hexagons of graphite [fig. 10.2(b)].
Finally, sp3 hybridisation, as found in methane, combines one s and three p orbitals to produce
a tetrahedral geometry.
10.4 Colloid Stability

A colloid is a dispersion of solid particles in a fluid. The dispersion (or London 1 ) force between
identical particles is always attractive, thus in order to produce a stable colloid there must also
be a repulsive interaction. Although formally CNTs are not charged, their suspension in water
can bring about surface charging. This will lead to two layers of opposite charges that are held
apart (despite attractive Coulombic interactions) because this separation is entropically favourable.
This configuration is know as a diffuse or electrical double-layer and it attempts to prevent the
coagulation of particles in solution. 2 The diffuse double-layer is sometimes known as an ionic
atmosphere with thickness the Debye Length. 3
133
Appendix
For acid-treated CNx MWNTs, dissociation of COOH groups on the nanotube surface (into
COO and H ) will increase both the double-layer repulsions and the Coulombic interactions.
10.4.1 DLVO Theory

The diffuse-double layer is obviously not always successful in keeping particles apart, as not all
particles will form a stable colloid in solution. CNTs in particular are keen to coagulate into bundles
(see section 7.1.1). According to Derjaguin, Landau, Verwey and Overbeek, for a stable colloid to
form there must be a balance between the repulsive interaction of the charges of the double layers
on neighbouring particles, and the attractive interactions arising from VDWs forces. 2 The stability
of a colloid as a function of inter-particle separation is shown in fig. 10.3 where the solid blue line
represents VDWs forces and the dotted line represents double-layer forces. Colloids are unstable
at either minimum in the VDWs forces on the graph. The primary minimum (a) occurs when the
inter-particle separation is zero so the particles are in contact. For some systems, on the other side
of the energy barrier (b) there will be a secondary minimum (c).
The behaviour of CNTs in solution indicates that dispersions of CNTs probably sit in the primary minimum (a) as coagulates. When the coagulates are agitated say by ultrasonication, they
will disperse (albeit briefly) as the energy from ultrasonication allows them to overcome the energy
barrier (b). For formally uncharged particles such as CNTs this energy barrier will be low, leading
to slow re-aggregation of particles.
For charged particles (such as acid-treated CNx MWNTs) the energy barrier (b) will be much
higher preventing re-aggregation, and allowing a stable colloid to remain after ultrasonication
The effect of surfactants

When a charged surfactant such as SDS (fig. 10.4) is added to raw CNx MWNTs in water, and
the solution sonicated, the dispersion improves considerably (section 7.1.1). This is because the
SDS molecules form micelles around the CNTs, which increase the surface charging and hence the
double-layer forces. This additional repulsive contribution will improve the stability of the colloid.
In addition, by keeping individual CNTs apart, the speed of sedimentation is reduced, according
134
Appendix
Figure 10.3: The balance of interactions in a colloid according to DLVO Theory
Figure 10.4: Sodium dodecyl sulphate
135
Appendix
Figure 10.5: The mechanism for amidation

to Stokes Law for the rate of sedimentation of a dispersed system:
dx
dt
d 2 1 2 g
18
(10.9)
where d is the diameter of the particle (21.7 nm), 1 and 2 are the densities of the solid and
liquid (C = 2267 kgm3 and H2 O = 1000 kgm3 ), g is the acceleration due to gravity and is
the viscosity of the liquid (0.89 for H2 O). Using these approximate values a sedimentation rate for
CNTs of:
3.6 x 1013 m/s was calculated.
DNA also has the effect of dispersing CNTs in solution. 4
10.5 Amino Acids

Amino acid structures and abbreviations are shown in fig. 10.6.
10.5.1 Amidation
Proteins were shown to adsorb more efficiently to acid-treated CNx MWNTs compared to raw CNx
MWNTs in Chapter 5. One of the possible explanations for this is the formation of amide linkages
between amines on the protein and carboxylic acid groups on the acid-treated CNx MWNTs.
The mechanism is shown in fig. 10.5 and involves attack by the N lone pair on the carbonyl
carbon. Overall the reaction is a hydrolysis as water is produced.
136
Appendix
Figure 10.6: Amino acid structures and their abbreviations, individual images from Wikipedia
137
Appendix
Figures of ferritin, azurin, ferredoxin, apoferritin and apoazurin were prepared using MOLMOL. 5
138
Appendix
References
[1] F. London. The general theory of molecular forces. Transactions of the Faraday Society, pages
826, 1937.
[2] N. Israelachivli. Intermolecular and Surface Forces. Academic Press, 2nd edition, 1992.
[3] V. J. Morris, A. J. Kirby, and A. P. Gunning. Atomic Force Microscopy for Biologists. Imperial
College Press, 1st edition, 1999.
[4] N. Nakashima, S. Okuzono, H. Murakami, T. Nakai, and K. Yoshikawa. DNA dissolves singlewalled carbon nanotubes in water. Chem. Lett., 32(5):465466, 2003.
[5] R. Koradi, M. Billeter, and K. Wuthrich. MOLMOL: a program for display and analysis of
macromolecular structures. J. Mol. Graphics, 14:5155, 1996.
139

Bio N MWCNT

Uploaded by

Copyright:

Available Formats

Bio N MWCNT

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bio N MWCNT

Uploaded by

Copyright:

Available Formats

Bioapplications of Nitrogen-doped Carbon

Hilary Jane Burch

Bionanotechnology IRC, Department of Physics

Thesis submitted for the degree of

Dr David Cox for SEM pictures taken at the University of Surrey.

For my twin sister Gillian

Electrical Transport in Nitrogen-doped CNTs . . . . . . . . . . . . . . . .

Potential of Nitrogen-doped CNTs for Biological Applications . . . . . . .

Thermal Properties of Nitrogen-doped CNTs . . . . . . . . . . . . . . . .

Effect of Acid Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . .

Publications and Presentations . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Defects caused by Nitrogen-doping . . . . . . . . . . . . . . . . . . . . .

Synthesis of CNx MWNTs . . . . . . . . . . . . . . . . . . . . . . . . . .

Growth of Un-doped CNTs . . . . . . . . . . . . . . . . . . . . . . . . .

Growth of CNx MWNTs . . . . . . . . . . . . . . . . . . . . . . . . . . .

Atomic Force Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

The Microscope Tip . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

UV-vis of CNx MWNTs . . . . . . . . . . . . . . . . . . . . . . . . . . .

Circular Dichroism Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . .

4 Electrical Conductance of CNx MWNTs

Electrical Transport in un-doped CNTs . . . . . . . . . . . . . . . . . . . . . . . .

Electronic Transport in CNx MWNTs . . . . . . . . . . . . . . . . . . . . . . . .

G-V and I-V curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5 Biofunctionalisation of CNx MWNTs

Biofunctionalisation of un-doped CNTs with Proteins . . . . . . . . . . . .

Functionalisation of Boron Nitride NTs with Proteins . . . . . . . . . . . .

Functionalisation of CNx MWNTs with Proteins . . . . . . . . . . . . . .

Effect on protein activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Antibody Binding Assays . . . . . . . . . . . . . . . . . . . . . . . . . .

Functionalisation with metalloproteins . . . . . . . . . . . . . . . . . . . .

Changes in the conformation of cyt c . . . . . . . . . . . . . . . . . . . .

Mechanism of apoazurin preparation . . . . . . . . . . . . . . . . . . . . .

Thermal conductance of CNx MWNTs

The Effect of Acid Treatment on CNx MWNTs

AFM Imaging and Dimension Data . . . . . . . . . . . . . . . . . . . . .

Measurement of Electronic Transport Properties . . . . . . . . . . . . . .

Measurement of Thermal Transport Properties . . . . . . . . . . . . . . .

8 Development of a CNx MWNT Biosensing Device

Previous Nanotube Sensing Work . . . . . . . . . . . . . . . . . . . . . . . . . . 101

Small Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

Sensing of Proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101

Sensing with CNx MWNTs . . . . . . . . . . . . . . . . . . . . . . . . . 102

Electroless Plating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103

Electrical Measurement Set-up . . . . . . . . . . . . . . . . . . . . . . . . 104

Sensing of Ferritin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

Sensing of cyt c . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

Sensing of azurin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

Sensing Mechansim . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112

9 Conclusions and Further Work

Electrical Conductance of CNx MWNTs . . . . . . . . . . . . . . . . . . . 117

Biofunctionalisation of CNx MWNTs . . . . . . . . . . . . . . . . . . . . 117

Thermal Conductance of CNx MWNTs . . . . . . . . . . . . . . . . . . . 118

The Effect of Acid Treatment on CNx MWNTs . . . . . . . . . . . . . . . 118

Development of a CNx MWNT Biosensing Device . . . . . . . . . . . . . 119

Further Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

Control Experiments with un-doped MWNTs . . . . . . . . . . . . . . . . 119

Further Thermal Measurements . . . . . . . . . . . . . . . . . . . . . . . 119

Atomic Resolution of CNx MWNTs . . . . . . . . . . . . . . . . . . . . . 120

Improved Biosensor Device Fabrication . . . . . . . . . . . . . . . . . . . 120