characterization and applications of multiwalled carbon
TRANSCRIPT
University of Kentucky University of Kentucky
UKnowledge UKnowledge
University of Kentucky Doctoral Dissertations Graduate School
2004
Characterization and Applications of Multiwalled Carbon Characterization and Applications of Multiwalled Carbon
Nanotubes Nanotubes
Jenny Marie Hilding University of Kentucky, [email protected]
Right click to open a feedback form in a new tab to let us know how this document benefits you. Right click to open a feedback form in a new tab to let us know how this document benefits you.
Recommended Citation Recommended Citation Hilding, Jenny Marie, "Characterization and Applications of Multiwalled Carbon Nanotubes" (2004). University of Kentucky Doctoral Dissertations. 303. https://uknowledge.uky.edu/gradschool_diss/303
This Dissertation is brought to you for free and open access by the Graduate School at UKnowledge. It has been accepted for inclusion in University of Kentucky Doctoral Dissertations by an authorized administrator of UKnowledge. For more information, please contact [email protected].
ABSTRACT OF DISSERTATION
Jenny Marie Hilding
The Graduate School
University of Kentucky
2004
Characterization and Applications of Multiwalled Carbon Nanotubes
___________________________________
ABSTRACT OF DISSERTATION ___________________________________
A dissertation submitted in partial fulfillment of the
requirements for the degree of Doctor of Philosophy in the College of Engineering
at the University of Kentucky
By Jenny Marie Hilding
Lexington, Kentucky
Director: Dr. Eric A. Grulke, Professor of Chemical Engineering
Lexington, Kentucky
2004
Copyright © Jenny Marie Hilding 2004
ABSTRACT OF DISSERTATION
CHARACTERIZATION AND APPLICATIONS OF MULTIWALLED CARBON NANOTUBES
Multiwalled carbon nanotubes (MWNTs) have attracted great interest during the
last decade due to their possession of a unique set of properties. In addition to
their strength, MWNTs have well defined morphologies, with large aspect ratios
and pores in the meso range, and intriguing transport properties, such as high
electrical and thermal conductivity.
We are interested in how variations in the MWNT morphology affect areas of
possible engineering applications. We have identified morphology as a critical
element for the performance of MWNTs in engineering applications. Specific
areas studied and reported here spans from surface adsorption and capillary
condensation, to dispersion and dispersion processes, and transport properties
in relation to MWNT aspect ratio. This wide range of exploration is typically
needed for evaluating opportunities for new materials.
MWNTs can be used in different types of adsorption systems and it should be
possible to tailor the MWNT morphology to suit a specific adsorption process.
We found that the major part of butane, our model gas, adsorbs on the external
MWNT and only a small fraction ends up in the pores.
The unusually large aspect ratio makes MWNTs ideal as fillers in polymer
matrixes. Since MWNTs are electrically conductive, it is possible to align the
MWNTs in the matrix before curing. We investigated the effect of AC-fields on
aqueous MWNT dispersions and the possibility to align MWNTs in an electrical
field.
It is also necessary to develop suitable dispersion methods, to enable the
production of homogeneous dispersions and composites. We studied a number
of different mechanical dispersion methods and their effect on the MWNT
morphology.
KEYWORDS: Carbon Multiwalled Nanotubes, Morphology Modeling, Isosteric Heat of Adsorption, Mechanical Dispersion Methods, Alignment in Electrical Field
Jenny Marie Hilding
January 22nd, 2004
CHARACTERIZATION AND APPLICATIONS OF MULTIWALLED CARBON NANOTUBES
By
Jenny Marie Hilding
Eric A. Grulke Director of Dissertation
Barbara L. Knutson
Director of Graduate Studies
January 22nd, 2004
RULES FOR THE USE OF DISSERTATIONS
Unpublished dissertations submitted for the Doctor’s degree and deposited in the University of Kentucky Library are as a rule open for inspection, but are to be used only with due regard to the rights of the authors. Bibliographical references may be noted, but quotations or summaries of parts may be published only with the permission of the author, and with the usual scholarly acknowledgments. Extensive copying or publication of the dissertation in whole or in part also requires the consent of the Dean of the Graduate School of the University of Kentucky.
DISSERTATION
Jenny Marie Hilding
The Graduate School
University of Kentucky
2004
CHARACTERIZATION AND APPLICATIONS OF MULTIWALLED CARBON NANOTUBES
___________________________________
DISSERTATION ___________________________________
A dissertation submitted in partial fulfillment of the
requirements for the degree of Doctor of Philosophy in the College of Engineering
at the University of Kentucky
By Jenny Marie Hilding
Lexington, Kentucky
Director: Dr. Eric A. Grulke, Professor of Chemical Engineering
Lexington, Kentucky
2004
Copyright © Jenny Marie Hilding 2004
To My Family
iii
ACKNOWLEDGMENTS
First of all, I thank my husband, Andy Berens, with all of my heart. You are
always there to share my life and you are the most supportive person I have ever
met. Your ability to make me laugh and relax makes life easy and wonderful
every day.
I thank my parents, Christer and Peggy Hilding. Both of you have always
encouraged me to continue my studies, though without pressure. You have
always been supportive in any decision I made; even when it meant that we
would have to live on different continents.
I thank my advisor, Professor Eric Grulke, for teaching me to believe in myself
and to be independent.
I also thank
- My advisory board, including Professor Anderson, Bhattacharyya, Knutson,
Meier, and Selegue.
- The NSF-MRSEC program and the S.B.S. NASA Ames Research Center for
project funding.
- Alan Dozier of the Electron Microscopy Center for interesting discussions and
for providing some TEM photomicrographs.
- Among other undergraduate students Alexandra Teleki, David Hindenlang,
Mo'atasem Mohamed Abu Haleeqa, and May Hong for their help in the
laboratory.
- Mr. Jing Li and Professor Janet Lumpp, from the Electrical Engineering
Department at University of Kentucky, for design and manufacture of electrodes.
- Everyone else who made my time at University of Kentucky a pleasant
experience.
iv
TABLE OF CONTENTS
Acknowledgments . . . . . . . . . . . . . . . . . . . . . .iii List of Tables . . . . . . . . . . . . . . . . . . . . . . . viii List of Figures . . . . . . . . . . . . . . . . . . . . . . . .ix List of Files. . . . . . . . . . . . . . . . . . . . . . . . . .xi Chapter One: Introduction . . . . . . . . . . . . . . . . . . . 1 Chapter Two: Background. . . . . . . . . . . . . . . . . . . . 5
2.1 Introduction . . . . . . . . . . . . . . . . . . . . 5
2.2 Carbon Nanotube Properties . . . . . . . . . . . . . . 6
2.2.1 The Carbon Nanotube Structure . . . . . . . . . . . 6 2.2.2 Mechanical Properties. . . . . . . . . . . . . . . 8 2.2.3 Transport Properties of Carbon Nanotubes. . . . . . 12
2.3 Properties of Nanotubes Dispersions and Composites . . . . 16
2.3.1 Electrical Properties. . . . . . . . . . . . . . . 16 2.3.2 Mechanical Properties . . . . . . . . . . . . . . 17 2.3.3 Thermal Properties. . . . . . . . . . . . . . . 20
2.4 Structured Carbons for Gas Separations . . . . . . . . . 20
2.5 Nanotube Morphology in Relation to Production Method. . . . 21
2.5.1 Vaporization of a Carbon Target . . . . . . . . . . 22 2.5.2 Solar Energy Vaporization. . . . . . . . . . . . 23 2.5.3 Electric Arc Discharge . . . . . . . . . . . . . 23 2.5.4 Electrolysis . . . . . . . . . . . . . . . . . 24 2.5.5 Chemical Vapor Deposition . . . . . . . . . . . 24 2.5.6 Sonochemical Production of Carbon Nanotubes. . . . 25 2.5.7 Low Temperature Solid Pyrolysis. . . . . . . . . . 25 2.5.8 Catalyst Arrays. . . . . . . . . . . . . . . . 25 2.5.9 Synthesis from Polymers . . . . . . . . . . . . 26
2.6 Dispersion Challenges. . . . . . . . . . . . . . . . 26
2.6.1 Physical and Chemical Entanglements. . . . . . . 27 2.6.2 Physical Contacts between Nanotubes in Fluids. . . . 27
2.7 Dispersion Methods. . . . . . . . . . . . . . . . . 29
2.7.1 Chemical Dispersion Methods. . . . . . . . . . . 29
v
Acidic Treatments. . . . . . . . . . . . . . 30 Fluorination . . . . . . . . . . . . . . . . 32 Other Types of Functionalization. . . . . . . . . 33 Surfactants and Dispersion Agents. . . . . . . . 33
2.7.2 Technical Dispersion Approaches. . . . . . . . . 35 In-situ Production of CNTs . . . . . . . . . . . 35 Production Technical Solutions. . . . . . . . . . 35
2.7.3 High-Impact Mixing – Ball-Milling. . . . . . . . . 36 2.7.4 Electrical Dispersion Mechanisms. . . . . . . . . 37
Chapter Three: Morphology Measurements and Modeling. . . . . . . 38
3.1 Introduction. . . . . . . . . . . . . . . . . . . . 38
3.2 Visual Techniques for Morphology Measurements. . . . . 39
3.2.1 Transmission Electron Microscopy, TEM. . . . . . . 39 3.2.2 Scanning Electron Microscopy, SEM. . . . . . . . 40 3.2.3 Optical Microscopy . . . . . . . . . . . . . . 44
3.3 Multiwalled Carbon Nanotube Samples. . . . . . . . . . 44
3.3.1 Sample Preparation and Diameter Measurements . . . 45 3.3.2 Morphology Results for CVD-Produced MWNTs. . . . 45
3.4 Morphology Model . . . . . . . . . . . . . . . . . 46
3.5 Geometrical Calculations. . . . . . . . . . . . . . . 54
3.6 Summary and Conclusions . . . . . . . . . . . . . . 58
Chapter Four: Isothermal Adsorption. . . . . . . . . . . . . . . 59
4.1 Introduction. . . . . . . . . . . . . . . . . . . . 59
4.1.1 Hysteresis. . . . . . . . . . . . . . . . . . 60 4.1.2 Stepwise Isotherms. . . . . . . . . . . . . . 63 4.1.3 Porosity. . . . . . . . . . . . . . . . . . . 64 4.1.4 Other Surface Curvature Effects. . . . . . . . . . 65 4.1.5 Adsorbate Order and Orientation. . . . . . . . . . 65
4.2 Experimental Procedure. . . . . . . . . . . . . . . 67
4.2.1 Adsorption Isotherms. . . . . . . . . . . . . . 68 4.2.2 Adsorption Model. . . . . . . . . . . . . . . 69 4.2.3 Monolayer Calculations. . . . . . . . . . . . . 75 4.2.4 Adsorbate Packing Density Calculations . . . . . . . 78
vi
4.2.5 Multilayer Analysis. . . . . . . . . . . . . . . 81 4.3 Conclusions . . . . . . . . . . . . . . . . . . . 82
Chapter Five: Isosteric Heat of Adsorption . . . . . . . . . . . . 83
5.1 Introduction. . . . . . . . . . . . . . . . . . . . 83
5.1.1 Isosteric Heat of Adsorption. . . . . . . . . . . . 83 5.1.2 The Clausius-Clapeyron Equation. . . . . . . . . 84 5.1.3 Homogeneous and Heterogeneous Adsorbents. . . . 88 5.1.4 Adsorbent Porosity . . . . . . . . . . . . . . 89 5.1.5 Adsorption on a Highly Bent Surface. . . . . . . . 90
5.2 Experimental Section. . . . . . . . . . . . . . . . 91
5.2.1 Adsorption Isotherms. . . . . . . . . . . . . . 91 5.2.2 Isosteric Heat of Adsorption. . . . . . . . . . . 92
5.3 Summary, Discussion, and Conclusions. . . . . . . . . 99 Chapter Six: Dispersion of Carbon Nanotubes. . . . . . . . . . . 101
6.1 Introduction. . . . . . . . . . . . . . . . . . . . 101
6.2 Modeling Typical Particle Fragmentation Distributions . . . . . 102
6.3 Ultrasonication Methods . . . . . . . . . . . . . . . 103
6.3.1 Ultrasonication Bath . . . . . . . . . . . . . . 104 Experimental Procedure. . . . . . . . . . . . 105 Results and Discussion. . . . . . . . . . . . . 106
6.3.2 Ultrasonication Horn or Wand. . . . . . . . . . . 106 Experimental Procedure. . . . . . . . . . . . 110 Results and Discussion. . . . . . . . . . . . . 110
6.4 Stretching of MWNTs in PAN Spinning Dope. . . . . . . . 114
6.5 Grinding. . . . . . . . . . . . . . . . . . . . . 117
Experimental Procedure. . . . . . . . . . . . 117 Results and Discussion. . . . . . . . . . . . . 117
6.5.1 Percolation Theory. . . . . . . . . . . . . . . 126 Experimental Procedure. . . . . . . . . . . . 126 Results and Discussion. . . . . . . . . . . . . 127
6.6 Summary and Conclusions . . . . . . . . . . . . . . 130
vii
Chapter Seven: Alignment of Carbon Nanotubes in an Electrical Field. . 132
7.1 Introduction . . . . . . . . . . . . . . . . . . . 132
7.2 Alignment of Dispersed MWNTs . . . . . . . . . . . . 134
7.2.1 Results. . . . . . . . . . . . . . . . . . . 134 Flow Alignment at Low Frequency . . . . . . . . 136 Alignment Transition . . . . . . . . . . . . . 136
7.2.2 Discussion . . . . . . . . . . . . . . . . . 136
7.3 Alignment of Flocculated MWNTs. . . . . . . . . . . . 137
7.3.1 Experimental Procedure and Results. . . . . . . . 138
7.4 Summary and Conclusions . . . . . . . . . . . . . . 140 Chapter Eight: Summary and Conclusions . . . . . . . . . . . . 141
MWNT Morphology Analysis. . . . . . . . . . . . . . . . 141
Adsorption on MWNTs . . . . . . . . . . . . . . . . . . 142
MWNT Dispersion. . . . . . . . . . . . . . . . . . . . 144
MWNT Alignment in Electrical Fields. . . . . . . . . . . . . 146
Appendices. . . . . . . . . . . . . . . . . . . . . . . . . 147
Appendix A: List of Symbols and Nomenclature. . . . . . . . . 147
Appendix B: Derivation of Equation 5.2. . . . . . . . . . . . 150
Appendix C: Maple Program. . . . . . . . . . . . . . . . 152
Bibliography. . . . . . . . . . . . . . . . . . . . . . . . 156 Vita . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
viii
LIST OF TABLES
Table 2.1 Mechanical properties of MWNTS. . . . . . . . . . . . 9
Table 2.2 Mechanical properties of SWNTs. . . . . . . . . . . . 11
Table 2.3 Mechanical properties of high and low aspect ratio solids. . . 13
Table 2.4 Transport properties of SWNTs and MWNTs. . . . . . . . 14
Table 2.5 Transport properties of diamond and graphite.. . . . . . . 15
Table 3.1 Log-normal probability distribution parameters; inner diameter distribution fits, samples 1-3. . . . . . . . . . . . . 50
Table 3.2 Log-normal probability distribution parameters, outer diameter distribution fits, samples 1-5. . . . . . . . . . . . . 51
Table 3.3 Geometrical calculations for samples 1-5. . . . . . . . . 57
Table 4.1 Presence of hysteresis observed on MWNTs and SWNTs during the last decade. . . . . . . . . . . . . . . . . 62
Table 4.2 The modified BET equation parameters fitted to sample 1-5 adsorption data. . . . . . . . . . . . . . . . . 74
Table 4.3 Measured and calculated monolayer capacities for samples 1-5. . . . . . . . . . . . . . . . . 76
Table 4.4 Estimated molecular butane adsorption area. . . . . . . . 80
Table 5.1.a Isosteric heats of adsorption for hydrogen, oxygen and nitrogen on different types of carbon. . . . . . . . . . . . . 85
Table 5.1.b Isosteric heats of adsorption for hydrocarbons on different types o f carbon. . . . . . . . . . . . . . . . . . . . 86
Table 5.1.c Isosteric heats of adsorption for noble gases on different types of carbon. . . . . . . . . . . . . . . . . . . . 87
Table 5.2 The modified BET fitting parameters at different temperatures for samples 2, 4 and 5. . . . . . . . . . . . . . . . 95
Table 6.1 The MWNT mean length and standard deviation for MWNTs treated in ultrasonication bath. . . . . . . . . . . . 108
Table 6.2 The MWNT mean length and standard deviation for MWNTs treated with ultrasonication horn. . . . . . . . . . . . . . 112
Table 6.3 Comparison of MWNT lengths of spun and drawn MWNT/PAN composite fibers. . . . . . . . . . . . . . . . . 116
ix
LIST OF FIGURES
Figure 3.1 TEM picture of a multiwalled carbon nanotube. . . . . . . 41
Figure 3.2 SEM picture of as-produced MWNTs. . . . . . . . . . . 42
Figure 3.3 SEM picture showing MWNT diameter distribution. . . . . . 43
Figure 3.4 The inner diameters distribution, sample 1-3. . . . . . . . 48
Figure 3.5 The outer diameter distribution, samples 1-5. . . . . . . . 49
Figure 3.6 Log-normal probability distribution function fit, inner diameters. . 52
Figure 3.7 Log-normal probability distribution function fit, outer diameters. 53
Figure 3.8 The difference between the log-normal, the Poisson and the Gaussian cumulative distribution models. . . . . . . . 55
Figure 4.1 Adsorption isotherms for samples 1-5. . . . . . . . . . 70
Figure 4.2 Lack of hysteresis in an adsorption-desorption cycle. . . . . 71
Figure 4.3 Capillary condensation in MWNTs according to the Kelvin equation. . . . . . . . . . . . . . . . 72
Figure 4.4 Predicted nm plotted vs. experimental nm. . . . . . . . . 77
Figure 4.5 Estimated butane packing radius. . . . . . . . . . . . 79
Figure 5.1 Adsorption-desorption isotherms at five different temperatures. No hysteresis is observed. . . . . . . 93
Figure 5.2 A set of adsorption isotherms and modified BET fits. . . . . 94
Figure 5.3 Isosteric heat of adsorption plotted vs. relative mass uptake. . 97
Figure 5.4 TEM pictures showing different MWNT defects that may be present in the samples. . . . . . . . . . . . . 98
Figure 6.1 Cumulative MWNT length distribution as a function of processing time in ultrasonication bath. . . . . . . 107
Figure 6.2 The time derivative of MWNT length in US-bath. . . . . . . 109
Figure 6.3 Cumulative MWNT length distribution as a function of processing time with ultrasonication wand. . . . . . . . 111
Figure 6.4 The time derivative of MWNT length during the course of treatment with ultrasonication wand. . . . . . . . . 113
Figure 6.5 Cumulative distribution of MWNTs from spun and drawn MWNT/PAN fibers. . . . . . . . . . . . . . . . 115
Figure 6.6 MWNT lengths before mechanical grinding.. . . . . . . . 118
Figure 6.7 MWNT lengths after mechanical grinding. . . . . . . . . 119
x
Figure 6.8 SEM microphotographs showing the diameter decrease for MWNT agglomerates after mechanical grinding. . . . 120
Figure 6.9 Cumulative distribution of aspect ratios for ground and unground MWNTs. . . . . . . . . . . . 122
Figure 6.10 Cumulative distribution of MWNT diameters for ground and unground MWNTs. . . . . . . . . . . . 123
Figure 6.11 Cumulative distribution of MWNT lengths for ground and unground MWNTs. . . . . . . . . . . . 124
Figure 6.12 Cumulative distribution of particle sizes for ground and unground MWNTs. . . . . . . . . . . . 125
Figure 6.13 Predicted percolation threshold for MWNTs. . . . . . . . 128
Figure 6.14 Resistance in epoxy as a function of MWNT loading. . . . . 129
Figure 7.1 Light microscope picture series showing a MWNT-water dispersion before and after alignment.. . . . . . . 135
Figure 7.2 Formation of aligned MWNT clusters connecting the electrodes. . . . . . . . . . . . . . 139
xi
LIST OF FILES
MWNT2004.pdf. . . . . . . . . . . . . . . . . . . . .7.34 MB
1
Chapter One
Introduction
Carbon nanotubes (CNTs) have generated increasing interest during the recent
decade, thanks to their unique properties, such as high tensile strength, electrical
conductivity, and relative inertness. The properties of multiwalled carbon
nanotubes position them for a wide variety of potential applications, for example
in adsorption processes as storage media or separation tools, or in novel
materials research to develop materials with enhanced mechanical and/or
transport properties.
A major part of the research in the nanotube area up to date has been
concentrated to theoretical calculations, such as computer modeling and
simulations. The downside to these types of investigations is that the system
parameters are generally based on the assumption that the model system is
ideal, i.e. the system is homogeneous and lacks impurities, defects,
morphological differences etc. It is therefore of great importance to complement
these investigations with actual laboratory work and experimental data and to
generate a more realistic picture of CNT systems.
Since the large-scale production of carbon nanotubes is under constant
development, and may result in moderate costs, the possibilities to incorporate
carbon nanotubes in several engineering areas should be investigated.
However, there is a severe lack of practical understanding of how the CNT
morphology affects different chemical and materials engineering systems in
reality. No time or effort has been invested in researching how different
treatment processes affect the morphology. This is very surprising since one of
the interesting properties the CNTs possess is the unique geometry.
Singlewalled nanotubes (SWNTs) and multiwalled nanotubes (MWNTs) have
regular and controllable geometries with uniform porosities and large aspect
ratios. The MWNT size scales, ~30-250 nm in diameter and ~5-50 microns in
2
length, permit MWNT systems to be investigated using transmission electron
microscopy, scanning electron microscopy, and in some cases, light microscopy.
Determining the physical properties of CNT systems as a function of their
morphology should be one of the basic initial steps taken when investigating
different types of engineering processes or developing applications and materials
utilizing the CNT. It is therefore of great importance to look at the physical
properties of MWNTs as a function of their morphology.
We have been researching MWNTs produced at the Center for Applied Energy
Research (CAER), a research center affiliated with the University of Kentucky.
These MWNTs are produced through a chemical vapor deposition process
(CVD). We are interested in how variations in the MWNT morphology affect
areas of possible engineering applications. We have identified morphology as a
critical element for the performance of MWNTs in engineering applications.
Specific areas studied and reported here spans from surface adsorption and
capillary condensation, to dispersion and dispersion processes, and transport
properties in relation to MWNT aspect ratio. This wide range of exploration is
typically needed for evaluating opportunities for new materials.
Chapter Two shows prior findings of other research teams. We review the
physical properties of both SWNTs and MWNTs to inform the reader about the
unique nature of the CNT and to provide a solid background for the research
presented in the chapters that follow. A short initial section regarding the CNT
structure is provided, but more specifically the mechanical properties and
transport properties of single CNTs are listed. The reviewed data has been
generated both from theoretical calculations and through experiments. To show
the wide range of possible CNT utilization in the area of novel materials
development, a number of CNT/polymer composite properties are listed as well.
A short section provides information about CNTs and other types of carbon
structures as gas separation tools. Today, several different CNT production
methods are available, each producing MWNT samples of different
characteristics. The products may vary in any property, from yield, purity, length,
3
and diameter to self-assembly structures; hence it is of interest to know what
type of samples each process generates. We provide a short presentation of the
most common production methods and we also introduce a few less well-known
production methods that may generate increasing interest in the future. In the
last part we review literature reports, methods, and specific problems associated
with CNT dispersion processes.
The new material founded on our research begins in Chapter Three. Here we
present a standardized method developed in our laboratory, which is used to
accurately measure, model, and describe the morphology of any type of MWNT
sample. The method is relatively quick and simple, but crucial as a tool to
understand and analyze a majority of the research carried out in the laboratory in
other scientific areas. This type of characterization research has not been
carried out previously in any area of MWNT science, but is most important for
future research. With the aid of this method, large areas previously overlooked
or even ignored in CNT research science and engineering have been covered.
The method has lead to a new level of understanding and many interesting
insights to CNT systems from a chemical-, materials -, and process engineering
aspects. We have consistently used this method throughout all of our research
and it has proven to be an invaluable tool. It has helped us analyze our research
in a methodical and novel manner.
One of the areas this method has been very useful to us is the area of surface
adsorption science. We analyzed several different samples with varying
morphology and could, with the aid of our morphology measurement method,
develop a new understanding for the butane-MWNT adsorption system, an
important first step towards understanding adsorption processes and
mechanisms involved in adsorption on a MWNT system. This knowledge is
essential for designing engineering systems utilizing MWNTs as a gas storage or
separation tool. If the sample morphology is known, it is easy to measure,
calculate, and even predict surface adsorption, capillary condensation, and
monolayer adsorption. The results of this detailed investigation are reported in
4
Chapter Four. It has also enabled us to investigate how the isosteric heat of
adsorption is related to the MWNT morphology as presented in Chapter Five.
As one very exciting area in the CNT field is new materials development, it is of
importance to investigate the impact different dispersion methods and processes
have on the CNT morphology and quality. Even though it is likely that most of
the CNT dispersion methods available and widely used today (as well as other
CNT processing tools) will affect the CNT morphology, the area has never yet
been investigated. We acknowledge the need for observing the MWNT sample
morphology as a function of different mechanical dispersion processes.
Mechanical methods have the advantage of not changing the chemistry of the
MWNT or introducing additional chemicals in the MWNT system. We have
investigated two different types of ultrasonication methods, together with manual
grinding of MWNT paste and stretching of MWNTs incorporated in a matrix. The
results, presented in Chapter Six, are of great importance to many CNT research
groups, both from a theoretical and a practical point of view. The lack of
awareness of the severe impact these types of CNT dispersion processes do
have on the CNT morphology may cause major problems in many CNT
investigations. The possibility to use shortened and severely damaged MWNTs
to create electrically conductive networks in an epoxy matrix was investigated as
well. The system was analyzed and compared to percolation theory with
unexpected results.
Since MWNTs are electrically conductive it should be possible to align them in an
applied electrical field. The ability to align MWNTs in various media enables the
development of special materials and also engenders opportunities to create
practical model systems in a laboratory setting. We designed a simple system to
measure and analyze how an applied electrical field’s frequency and magnitude
impacts MWNTs dispersed in water. We were initially interested in investigating
MWNT behavior and interaction with the surroundings MWNT in the field mainly
as a function of size. However, the manipulation of MWNTs in electrical fields
generated some unexpected and intriguing results described in Chapter Seven.
5
Chapter Two
Background
2.1 Introduction
Single-walled and multiwalled carbon nanotubes possess a set of unique
properties that opens the door to several applications in the fields of science and
engineering. In addition to their exceptional strength, CNTs are thermally stable
up to high temperatures, both in atmospheric environments and in vacuum. The
morphology, with both well-defined, straight pores and large external surface
areas, is easily modeled and can even be treated as 2D geometries. The tube
wall consists of curved graphite sheets forming tubes with a diameter and length
on the nano- and micro-scale respectively, so that the aspect ratio is unusually
large. This, in combination with a good thermal and electrical conductivity, may
create endless possibilities in the fields of polymer and composite science.
Single-walled nanotubes consist of only one carbon layer, while multiwalled
carbon nanotubes contain at least two layers. SWNTs normally associate into
bundles with hexagonal packing that have interstitial sites capable of adsorbing
small molecules. MWNTs do usually not form ordered solids with interstitial
sites. SWNTs have pores in the microporous range while MWNTs have pores in
the mesoporous range. Internal pore volumes and external surface areas can be
quantitatively determined from measurements of SEM- or TEM-
microphotographs.
Depending on what synthesis method is used, the different products will display
differences in purity, yield, aspect ratio, nanotube diameters, defects types, and
physical entanglements. The nanotube characteristics can be manipulated by
changing processing and reactor conditions. A long-term scientific objective is to
be able to fabricate nanotubes to specific diameters and lengths at precise
locations; however, this is not yet fully accomplished. Only in a few
6
circumstances is the reactor product of the desired morphology for the
application. It is therefore necessary to investigate what impact CNT
manipulation and processing may have on the CNT morphology and properties.
In section 2.2 an extensive CNT property review is presented to display the unique
nature of both SWNTs and MWNTs. While there are many unusual properties of
carbon nanotubes, the focus is aimed towards mechanical and transport properties
because of their potential engineering and product applications. Tensile,
compressive and flexural moduli should relate to carbon nanotube performance in
structural composites. In section 2.3 CNTs are reviewed as fillers in different types
of materials. This presentation illustrates how the CNT filler changes the matrix’
mechanical properties and transport properties. This section also reviews CNTs
in dispersion systems. Section 2.4 reports how it is possible to use carbons of
different geometries in various gas separation processes, while section 2.5 gives
several examples of different CNT production methods and presents each
method’s resulting product in detail.
2.2 Carbon Nanotube Properties
In this section several properties of MWNTs and SWNTs are reviewed and
discussed. The reviewed data is taken both from theoretical calculations and
models, as well as from experimental results and more practical investigations.
2.2.1 The Carbon Nanotube Structure
SWNT properties can be modeled using molecular dynamics simulations, and
there are many predictions on their physical properties. Measuring these
properties can be difficult, as it is difficult to isolate single SWNTs for study. The
unusual properties of carbon nanotubes could lead to many bulk and surface
applications. The thermal conductivity of carbon nanotubes is two times larger
than that of a diamond and CNTs are thermally inert up to around 700oC in
atmospheric conditions or 2800oC in vacuum. The electrical carrying capacity is a
thousand times larger than that of copper wires (Collins, Hersam et al. 2001).
7
When SWNTs are pressurized up to 55 GPa, the tubes do not collapse, but a
superhard phase can be created (SPSWNT) having a hardness in the range of 62
to 150 GPa (Popov, Kyotani et al. 2002). SWNTs can have different helicity,
depending on to what angle the graphene sheet is rolled up. The helical structure
is denoted by two integers, (m,n), that indicate the number of lattice vectors in the
graphite plane (Garg and Sinnott 1998). For chiral structure, the integers are
(2n,n); for zig-zag, (n,0); and for armchair, (n,n). The physical properties of the
tube, especially the electrical properties, are dependent on the helicity.
Molecular dynamics studies show that axial deformation of SWNTs at 0 K is
strongly dependent on helicity. Armchair and zig-zag are the stiffest structures
(Ozaki, Iwasa et al. 2000). Armchair and chiral CNTs display metallic character,
while zig-zag CNTs are semi-conducting. Bending the tubes changes their
electrical properties. Topological defects increase the electrical resistance of
metallic NTs. For the metallic chiral tube, bending introduces a metal-
semiconductor transition manifesting itself in the occurrence of effective barriers for
transmission. Zig-zag nanotubes remain semi-conducting (d<1.5 nm) and the
armchair configuration keeps the metallic character during bending (d>0.7 nm)
(Buongiorno Nardelli, Fattebert et al. 2000). Bending and twisting a MWNT
changes its electronic transport properties. No effects were seen for low strains,
but when high strains were applied, the MWNT morphology changed resulting in
drastic effects in the local electron structure of the tube (Tekleab, Carroll et al.
2001). It has also been seen that the transmission function decreases for twist
>4o. When the tube is twisted, the electrical resistance increases, especially for
angles larger than 45o. This is strain enough to result in kinks, introducing a-ß
hybridization (Rochefort, Avouris et al. 1999). Also, for strains greater than 5% (at
high temperature), pentagon or heptagon defects are spontaneously formed, since
they are energetically favorable. This can lead to an onset of plastic deformation
of the nanotube (Hooke’s law or Stone-Wales transformation) (Bernholc, Brabec et
al. 1998). This type of configuration deformation can be seen as a singularity in
the stress-strain curve, due to energy release.
8
2.2.2 Mechanical Properties
Multiwalled carbon nanotubes consist of several seamlessly rolled up graphene
sheets, with typical shell or wall separations, dwall, of 0.34 nm (Muster, Burghard et
al. 1998). In the absence of mechanical entanglements, MWNTs tend to exist as
individual particles in suspensions with many organic liquids. Carbon nanotubes
are extremely strong, 30-100 times stronger than steel, but with only one sixth of
its weight.
Table 2.1 shows measured values of the mechanical properties of MWNTs. The
measured values of their Young’s moduli are in the range of 103 GPa. Tensile
strengths may be in the range of 50 GPa, with compressive strengths perhaps a
factor of two larger. Individual tubes are susceptible to bending, and can recover
their original shape after deformation.
Mechanical properties can be measured in a few different ways. The tip of an
atomic force microscope (AFM) can be used to measure the mechanical properties
of carbon nanotubes in many ways. One way is to place the tube over an alumina
ultrafiltration membrane and push on the side of the tube with the AFM tip. The
deflection is recorded as a function of the force (Salvetat, Bonard et al. 1999;
Salvetat, Briggs et al. 1999). This type of experiment has also been carried out
using a scanning probe microscope (SPM) tip (Shen, Jiang et al. 2000). If a
carbon nanotube is mounted between two opposing AFM tips, the outer shell
strength can be measured when the tube is pulled apart (Yu, Lourie et al. 2000).
One experiment has even been done with a specially designed stress-strain puller
to measure Young’s modulus of thin ropes consisting of aligned MWNTs (Xie, Li et
al. 2000). The tubes can also be attached between a surface and a AFM tip and
pulled (Yu, Files et al. 2000). TEM can be used to measure the intrinsic thermal
vibrations of tubes. From these measurements, the modulus can be calculated
(Treacy, Ebbesen et al. 1996; Krishnan, Dujardin et al. 1998).
9
Table 2.1 Mechanical properties of MWNTS. Physical measurements.
Measured values [ GPa ]
Property Value Reference
Outer shell strength 11-63 (Yu, Lourie et al. 2000)
Young’s modulus 810±410 (Salvetat, Bonard et al. 1999)
Young’s modulus 450 (Xie, Li et al. 2000)
Young’s modulus 1.8×103 (Treacy, Ebbesen et al. 1996)
Young’s modulus 1.3×103 (Wong, Sheehan et al. 1997)
Young’s modulus 270-950 (Yu, Lourie et al. 2000)
Elastic modulus 1.28×103±0.59×103 (Wong, Sheehan et al. 1997)
Young’s modulus, radial compression
9.7-80.8 (Shen, Jiang et al. 2000)
Axial tensile strength ~10-60 (Yu, Lourie et al. 2000)
Tensile strength, MWNT rope 3.6 (Xie, Li et al. 2000)
Axial elasticity ~200-400 (Yu, Lourie et al. 2000)
Axial compressive strength ~100 (Yu, Lourie et al. 2000)
Radial compressive strength 5.3 (Shen, Jiang et al. 2000)
Bending modulus, MWNTs with bamboo-type structure
D < 8 nm
D > 30 nm
23-32
1.2×103
0.2×103
(Wang, Gao et al. 2001)
Average bending strength 14.2 (Wong, Sheehan et al. 1997)
Average bending strength ~14 (Yu, Lourie et al. 2000)
Average bending strength 14.2±8.0 (Wang, Gao et al. 2001)
Calculated values [ GPa ]
Young’s modulus 1.0×103 (Muster, Burghard et al. 1998)
Copyright © Jenny Marie Hilding 2004
10
An oscillating voltage of the right amplitude applied on the nanotube can also
induce a mechanical resonance (Wang, Gao et al. 2001). One of the carbon
nanotube ends can also be pinned to a surface and pushed from the side (Wong,
Sheehan et al. 1997).
Measurements of the elastic modulus (Young’s modulus) are usually done by
atomic force microscopy (Wong, Sheehan et al. 1997), and give results in the
range of 103 GPa, giving reasonable agreement with theoretical estimates
(Muster, Burghard et al. 1998). The radial compression shows an interesting
nonlinearity with the applied stress, resulting in an elastic modulus that increases
with compression (Shen, Jiang et al. 2000). The bending modulus of MWNTs is
dependent on the tube diameter. The bending modulus decreases as the tube
diameter increases, an effect attributed to wall wrinkling during the bending
process. This effect almost vanishes when the tubes diameter is smaller than 12
nm (Wang, Gao et al. 2001). Most high strength fibers are brittle, but carbon
nanotubes can have very large strains at their yield point, perhaps as much as
0.30.
SWNTs are more flexible than MWNTs and generally self-assemble into bundles
as a way to minimize surface energy. These bundles can be very difficult to
disperse. The mechanical properties of SWNTs are more difficult to measure
due to their small size. Furthermore, the SWNTs are usually arranged in ropes,
so it is hard to measure the physical properties for one single tube. Table 2.2
gives some measured and calculated properties of SWNTs and SWNT bundles
or ropes.
Surprisingly, some measured values for the Young’s moduli of SWNT ropes are
similar to the calculated values of single SWNTs. Also, the bending moduli of
SWNTs with different helicities are quite similar.
11
Table 2.2 Mechanical properties of SWNTs.
Measured values [ GPa ]
Property Value Reference
Young’s modulus 1.25×103 (Krishnan, Dujardin et al. 1998)
Young’s modulus, SWNT ropes
320-1470 (Yu, Files et al. 2000)
Young’s modulus, SWNT ropes
6.5 ± 4.10 (Salvetat, Briggs et al. 1999)
Tensile load 13-1470 (Yu, Files et al. 2000)
Average bending strength 14.2 ± 8 (Salvetat, Briggs et al. 1999)
Shear modulus 6.5 ± 4.10 (Salvetat, Briggs et al. 1999)
Calculated values [ GPa ]
Young’s modulus 764 (Zhou, Duan et al. 2001)
Young’s modulus 1×103 (Popov, Van Doren et al. 2000)
Young’s modulus, Armchair
Zig-zag
Chiral
674
641
648
(Gao, Cagin et al. 1998)
Young’s modulus 0.97×103 (Lu 1997)
Young’s modulus,
Variation with radius
variation with helix angle
0.95×103-1.2×103
1.02×103-1.08×103
(Yao, Zhu et al. 2001)
Young’s modulus, SWNT ropes 320-1470 (Yu, Files et al. 2000)
Young’s modulus 1.22×103 (Goze, Vaccarini et al. 1999)
Tensile strength 6.249 (Zhou, Duan et al. 2001)
Breaking strength 13-52 (Yu, Files et al. 2000)
Bending modulus, Armchair
Zig-zag
Chiral
963.44
911.64
935.48
(Gao, Cagin et al. 1998)
Bending modulus (10,0),(6,6)
(10,10)
(10,5), (15,15)
(20,0)
1.22×103
1.24×103
1.25 × 103
1.26 × 103
(Goze, Vaccarini et al. 1999)
Copyright © Jenny Marie Hilding 2004
12
Table 2.3 provides mechanical properties of several other carbon-related
materials with high and low aspect ratios. Diamond exists as a three-
dimensional structure of carbon with sp3 bonds. It is the hardest material known
(hardness=10,900 kg/mm2), and has excellent tensile (3.5 GPa) and
compressive strength (110 GPa). Graphite has low mechanical strength and is
often used as a low friction coefficient material since the graphitic planes can
easily slip past each other. Carbon nanotubes, graphite and diamond have
similar Young’s moduli. Generally, all are stronger than vapor-grown fibers.
2.2.3 Transport Properties of Carbon Nanotubes
The transport properties of carbon nanotubes are high compared to many solids,
Table 2.4. Percolation theory (discussed further in Chapter 7) suggests that
highly conductive, high-aspect-ratio solids could produce three-dimensional
networks with high transport properties. The thermal and electrical conductivities
of single nanotubes are thought to be higher than graphite, although the more
important number may be the conductivity of three-dimensional networks of
these solids. For example, defect densities may have a larger effect on transport
properties than on mechanical properties.
The transport properties of graphite and diamond are shown in Table 2.5 for
comparison. The thermal conductivity of diamond is one of the largest known
(2000 W/m·K). The thermal conductivity of graphite along its planar direction is
also very high (1000 W/m·K) while the value perpendicular to the plane is fairly
low. Measurements and computations of the thermal conductivity of an individual
single-walled nanotube exceed those for diamond. However, when the
computations are performed for bundles of SWNTs, the expected morphology in
most systems, the values resemble those of in-plane and cross-plane graphite.
Measured and computed values for the thermal conductivities of MWNTs are
also similar to those of graphite.
13
Table 2.3 Mechanical properties of high and low aspect ratio solids.
Mechanical properties for misc. material
Measured values [ GPa ]
Material Property Result Ref
Carbon fibers Young’s modulus 900 (Dresselhaus 1988)
Carbon fibers Elastic modulus
100-800 (Peebles 1995)
Carbon fibers 695 (Jacobsen, Tritt et al. 1995)
Graphite Young’s modulus 1.06×103 (Kelly 1981)
Graphite (Compression-Annealed Pyrolytic Graphite)
Young’s modulus 1.02×103 (Blakslee, Proctor et al. 1970)
SiC-nanorods Average bending strength
53.4 (Wang, Gao et al. 2001)
SiC nanowires Young’s modulus 20-32 (Wang, Gao et al. 2001)
SiC-SiO nanowire
SiC in wire core, SiO in sheath
Young’s modulus 46-81 (Wang, Gao et al. 2001)
Calculated values [ GPa ]
SiC-SiO nanowire
SiC in wire core, SiO in sheath
Young’s modulus 73-109 (Wang, Gao et al. 2001)
Graphite Young’s modulus 0.97×103 (Lu 1997)
Graphite Young’s modulus 1060 (Overney, Zhong et al. 1993)
Graphite Young’s modulus 1.02×103 (Pierson 1993)
Vapor-grown fibers Young’s modulus 0.15×103 (Tibbetts and Beetz 1987)
Diamond Young’s modulus 1.2×103 (Thostenson, Ren et al. 2001)
Vapor-grown fibers Tensile strength 2.6 (Endo, Kim et al. 2001)
Graphite Tensile strength 20 (Endo, Kim et al. 2001)
Copyright © Jenny Marie Hilding 2004
14
Table 2.4 Transport properties of SWNTs and MWNTs.
Conductivity properties for SWNTs and MWNTs Property Value Unit Ref
SWNT
Thermal conductivity 6000 W/m-K (Hone, Whitney et al. 1999; Hone, Llaguno et al. 2000)
Thermal conductivity* 2400-2900 W/m-K (Che, Cagin et al. 2000)
Thermal conductivity, Axial, bundles*
950 W/m-K (Che, Cagin et al. 2000)
Thermal conductivity, Radial, bundles*
5.6 W/m-K (Che, Cagin et al. 2000)
Critical burn-out current 109 A/cm2 (Radosavljevic, Lefebvre et al. 2001)
Electrical resistance
Fiber spun from 200 SWNTs
Fiber spun from 100 SWNTs
0.065
10
KΩ (Reulet, Kasumov et al. 2000)
MWNT
Thermal conductivity (300 K) 900 W/m-K (Kim, Shi et al. 2001)
Thermal conductivity (300 K)* 1980 W/m-K (Berber, Kwon et al. 2000)
Thermal conductivity
300 K high defect ratio.
25 W/m-K (Xie, Li et al. 2000)
Current density 1×103 A/cm2 (Kaneto, Tsuruta et al. 1999)
Critical burn-out current 100-600×10-6 A (Kaneto, Tsuruta et al. 1999)
Critical burn-out current 200×10-3 A (Zettl and Cumings 2001)
Specific heat
300K
~0.430 kJ/kg-K (Xie, Li et al. 2000)
Electrical conductivity 1000-2000 S/cm (Kaneto, Tsuruta et al. 1999)
Electrical resistance 6-120 kΩ (Zettl and Cumings 2001)
Electrical resistance (T=10-300)
2.0-3.5 kΩ (Xie, Li et al. 2000)
* Theoretical value
Copyright © Jenny Marie Hilding 2004
15
Table 2.5 Transport properties of diamond and graphite .
Conductivity properties for diamond and graphite
Property Value Unit Ref
Graphite
Thermal conductivity x-plane*
1000 W/m-K (Che, Cagin et al. 2000)
Thermal conductivity y-plane*
5.5 W/m-K (Che, Cagin et al. 2000)
Electrical resistance, in-plane
3.7×10-3 Ω-cm (Weast 1986)
Electrical resistance, perpendicular to in-plane
1.7×10-6 Ω-cm (Weast 1986)
Diamond
Thermal conductivity 2000 W/m-K (http://www.zyvex.com/nanotech/talks/XeroxPARC980312/sld034.ht
m)
Electrical resistivity 1013 W-m (http://www.zyvex.com/nanotech/talks/XeroxPARC980312/sld034.ht
m)
* Theoretical values
Copyright © Jenny Marie Hilding 2004
16
At least one measurement is several orders of magnitude lower, suggesting that
defects in the MWNTs significantly affect their transport properties. The electrical
resistance of MWNTs appears to be higher than that of SWNTs, and similar to
the values for graphite. Diamond is an excellent thermal conductor but acts like
a semiconductor in terms of electronic properties.
2.3 Properties of Nanotubes Dispersions and Composites
Electrical and thermal conductivities are important in developing conducting
suspensions and polymeric solids. Changes in these transport properties when
the nanotube is under mechanical strain could be important for electrical and
nano-mechanical devices.
It is tempting to apply the rule of mixtures to estimate the properties of carbon
nanotube suspensions and composites. As with other filled systems, this rubric
is only a first approximation and could very well provide a poor estimate since
dispersion, orientation, and the interphase linking the nanotubes to the
continuous phase all affect the bulk properties of the material. Percolation theory
may apply to transport properties of suspensions and composites, but also may
have significant limitations. Despite our limitations in predicting suspension and
composite properties, the physical property data of this section can be used to
compare carbon nanotubes to other potential additives, to develop estimates of
mixture properties, and to illustrate the potential variation in physical properties of
these materials.
2.3.1 Electrical Properties
The electrical properties of a composite are measured in terms of resistance.
Well-dispersed electrical fillers create a three-dimensional network, which
provides a conductive path through the composite. This is a commercial way to
turn an insulating material into an electrical conductor. The loading limit is called
percolation threshold and can be detected as a sharp drop in electrical
resistance. Single-walled carbon nanotubes in epoxy have percolation
17
thresholds in the range of 0.1-0.2 wt% or 0.04 vol% (Sandler, Shaffer et al. 1999;
Biercuk, Llaguno et al. 2002), while it has been determined that the percolation
threshold for nanotubes in PMPV is a as high as 8.5 wt% (Coleman, Curran et al.
1999). The investigation of electrical conductivity change of MWNT-filled epoxy
as a function of nanotube weight fraction is discussed in Chapter 7 (Hilding,
Grulke et al. 2003). We found that at a MWNT loading of about 1 wt%, there is a
large reduction in the material’s electrical resistance. Low levels of carbon
nanotubes could result in large increases in the thermal and electrical
conductivities of liquid suspensions and polymer composites. As a comparison,
some typical percolation thresholds for other fillers are 9-18 wt% for vapor-grown
carbon fibers in polypropylene (Lozano, Bonilla-Rios et al. 2001), 15 wt % for Cu-
powder (Garza 1993), 35 wt% for Al-powder (Dutta, Misra et al. 1992), and 20-
40 wt% for carbon black in epoxy (Dutta, Misra et al. 1992; Sandler, Shaffer et al.
1999). Such high concentrations will most likely compromise the physical
properties of the matrix and do not result in multifunctional materials. The
positive conductive effect of carbon nanotube filler is dramatic. It has been
reported that the electrical conductivity of PPV increases eight times by
introduction of SWNTs in the matrix (Curran, Ajayan et al. 1998). Adding
MWNTs to PMPV increases the electrical currency from 2×10-10 S/m to 3 S/m
(Coleman, Curran et al. 1999) and the electric conductivity increased by 340%
for fiber spun from petroleum pitch containing 5 wt% purified SWNTs (Andrews,
Jacques et al. 1999). As little as 0.1 vol% nanotubes increases the matrix
conductivity of the electrical insulator epoxy to 1 mS/m (Sandler, Shaffer et al.
1999). Metal oxides can also be converted from insulators to conductors by hot-
pressing the MeO powder, an insulating material, with nanotubes at moderate
temperatures (1200oC), resulting in an electrical conductivity between 0.2-4.0
S/cm (Flahaut, Peigney et al. 2000). To avoid electrical charging of an insulation
material, a matrix conductivity of 1 µS/m is necessary. At hot-pressing
temperatures above 1500oC, the nanotubes get destroyed and the composites
turn into insulators again.
18
The nanotube alignment in the matrix is important and can improve the electrical
properties even further. It has been observed that the electrical conductivity is
higher in, as supposed to perpendicular to, the flow-direction (0.1-12 vs. 0.08-7
S/m) for solvent cast and hot-press melt-mixed poly(methyl methacrylate)
(PMMA) with SWNTs (Haggenmueller, Gommans et al. 2000). When the
SWNTs are ground prior to film-casting, the emission current is improved. Films
made with ground tubes show a higher amount of free nanotube ends sticking
out perpendicular to the surface, which might be an explanation of this effect (Ito,
Konuma et al. 2001).
2.3.2 Mechanical Properties
Another way to determine whether the carbon nanotubes are well dispersed in a
matrix is to investigate the mechanical properties of the composite. If the
mechanical properties improve, the dispersion is most likely uniform throughout
the matrix. If the dispersion is poor, the mechanical properties will degrade
relative to the pure polymer matrix.
The Vicker’s hardness is 3.5 times higher for epoxy with a SWNT loading of 2
wt%, compared to pure epoxy. This result indicates good dispersion (Biercuk,
Llaguno et al. 2002). When loading the epoxy with 5 wt% MWNT, the tensile
strength increases from 3.1 GPa to 3.7 GPa and the compressive modulus
increases from 3.6 GPa to 4.5 GPa (Schadler, Giannaris et al. 1998).
Introduction of 1 wt% MWNT in a polystyrene matrix increases the elastic
modulus by 36-42% (to 450 MPa) and the break stress by 25% (Qian, Dickey et
al. 2000). There are quite a few reports on nanotube-fillers in PMMA. The
effects of fairly low amounts of nanotubes in PMMA are generally positive, as
only 1 wt% NT in PMMA increased the elastic modulus by 30% (Gong, Liu et al.
2000). One method developed to incorporate MWNTs in PMMA is to use the
free radical-initiator AIBN (2,2-azobisisobutyronitrile) to polymerize the PMMA.
When as-grown tubes (1-3 wt%) are used as a filler, a slight improvement of
mechanical properties is observed. If purified MWNTs are used (98% pure) the
effect is more dramatic; toughness, tensile strength and hardness increase 34%,
19
31% and 48%, respectively, for a loading of 1-7 wt%. When the MWNT loading
exceed 7 wt%, the tubes are not completely wrapped in the polymer and the
positive effects are lost, due to the residual stress in the PMMA matrix. If CNTs
get to 10 wt.%, the composite becomes very brittle (Jia, Wang et al. 1999).
Another mixing method is melt-blending MWNTs in PMMA at low temperatures
(200oC). The storage modulus increases, especially at high temperatures. By
adding 30 wt% MWNTs, the storage modulus is increased by a factor of 1.4 at
40oC and a factor of 6.0 at 120oC (Jin, Pramoda et al. 2001). Furthermore,
adding a small amount of poly(vinylidene fluoride) (PVDF) increases the storage
modulus dramatically. For example, adding 0.5 wt% PVDF to the matrix doubles
the storage modulus compared to a pure MWNT/PMMA composite. The
improvement is temperature-dependent, since the glass transition temperature,
Tg , decreases for the mix when PVDF has a lower Tg than PMMA. The lower Tg
means that the composite gets softer, or has a lower storage modulus, at higher
temperatures. The percentage of PVDF has to be kept low in the composite, not
to weaken the matrix (Jin, Pramoda et al. 2002). The mechanical improvements
on adding nanotubes to an amorphous polymer, as supposed to a semi-
crystalline polymer, can be greater since the semi-crystalline already is stiff.
The mechanical properties of fibers produced with carbon nanotubes have been
investigated as well. When 5 wt% purified SWNTs were spun with isotropic
petroleum pitch, the tensile strength increased by 90% (from 500 to 800 GPa)
and the elastic modulus increased by 150% (from 34 to 78 GPa) (Andrews,
Jacques et al. 1999). Nanotube ribbons and fibers have also been produced by
recondensing SWNT-surfactant dispersions in a poly(vinyl alcohol) solution. The
elastic modulus was measured at 9-15 GPa, which is 10 times higher than the
modulus of high-quality bucky paper (Baughman, Cui et al. 1999). The fibers
contain mainly aligned SWNT bundles and can be tied into a tight knot (Vigolo,
Penicaud et al. 2000; Vigolo, Penicaud et al. 2001).
Finally, carbon nanotubes have also been used to reinforce metallic compounds,
with various results. The most popular blending method in these cases is hot-
20
pressing at temperatures around 1200-1300oC. When nanotubes are hot-
pressed with metal oxides, the fracture strength and toughness decrease with
loading (Flahaut, Peigney et al. 2000), while when hot-pressed with nanophase
alumina powder, the fracture toughness increase from 3.4 to 4.2 MPa/m1/2
(Siegel, Chang et al. 2001) for a loading as high as 10 wt%.
2.3.3 Thermal Properties
It is also possible to improve the thermal stability of a composite by adding
carbon nanotubes. By adding 1 wt% SWNTs to an epoxy matrix, the thermal
conductivity increases 120% to 0.5 W/m·K at room temperature and 40% at 40K.
The thermal conductivity increases 60% for a 0.5 wt% SWNT load at RT
(Biercuk, Llaguno et al. 2002). By adding 1 wt% nanotubes in PMMA, the glass-
transition temperature, Tg, increases from 66 to 88oC (Gong, Liu et al. 2000) and
the heat-deflection temperature increases as well (Jia, Wang et al. 1999) and
the degradation temperature increases with 30oC for a MWNT loading of 26 wt%
(Jin, Pramoda et al. 2001).
Clearly, the transport properties of nanotube suspensions and composites can be
altered by the addition of carbon nanotubes. Control of the dispersion methods
is important to developing uniform and reproducible properties in these
applications.
2.4 Structured Carbons for Gas Separations
CClF2CF3 and CHF2CF3 can be successfully separated using SWNTs (Corbin
and Reutter 1997). The geometrical pore properties of the SWNT sample, such
as size or shape, affect separations as do the energetics of the adsorbate
surface (Nicholson and Gubbins 1996). One molecular dynamics study suggests
that SWNTs can expand to hold xylene (all 3 isomers), a relatively large molecule
(Takaba, Katagiri et al. 1995). The separation of H2 and CO on SWNTs has
been investigated using a grand canonical Monte Carlo simulation (GCMC),
predicting a possible CO enrichment up to 98% (Gu, Gao et al. 2002). Another
21
simulation predicts that a transition in selectivity of binary mixtures of
hydrocarbons on SWNTs can be achieved simply by changing the temperature
(Ayappa 1998). An adsorbed gas can be displaced by another, e.g. the
displacement between CH4 and CCl4 in an adsorbed monolayer when the
temperatures varies between 77 and 112 K (Weber and Goodstein 1999).
In structured carbons such as graphite with slit-like pores, GCMC studies of
binary CH-mixtures suggest that the selectivity is a function of the orientation of
molecules in the spacing between the pore walls (Cracknell and Nicholson 1994;
Cracknell, Nicholson et al. 1995; Nicholson 1996). Fluids adsorbed in
microporous materials can exhibit quantum sieving, where heavy isotopes are
preferentially adsorbed over light isotopes (H2-T2, 3He-4He, CH4-CD4, and H2-HD
(Wang, Challa et al. 1999; Challa, Sholl et al. 2001)). Predictions of isotherms
and selectivity can be evaluated using the enthalpy of vaporization of the
components and their heats of adsorption (Aranovich and Donohue 1999).
2.5 Nanotube Morphology in Relation to Production Method
When novel carbon nanotube materials or applications are investigated and
developed the morphology is crucial. Depending on what product property is
desirable it is possible to choose the most suitable nanotube geometry from the
beginning simply by picking right production method. While carbon nanotube
purification and separation of undesirable byproducts may be needed today, the
future of lower nanotube costs and better product quality lies mainly in learning to
control the synthesis chemistry and the reactor system. Work on the catalyst
morphology, control of the nanotube growth rate by control of the local
environment at the catalyst, and understanding the fouling of the catalyst should
lead to reactor systems that provide better control over the product quality. This
is why it is of importance to be familiar with the different carbon nanotube
production methods.
22
This section briefly reviews a number of different carbon nanotube production
methods and their respective product morphologies. Each production method is
described briefly and accompanied by information on the morphologies of the
carbon products and their purity. The carbon nanotube morphology can be
described relatively simply, with the inner diameter (D i), outer diameter (Do), and
the tube length (L). Generally, the carbon nanotube aspect ratio, L/Do, is >>1.
The external diameter and the length can be easily measured with SEM and in
some cases even a Reflective Light Microscope (RLM). The inner diameter can
be measured by using TEM.
Carbon nanotube purity can be difficult to assess, as both residual catalyst and
amorphous carbon may be present. Some researchers report purities based on
microscopy analyses alone, while others use digestion methods to solubilized
non-graphitic carbons and metals. There can be considerable variations in the
carbon nanotube morphologies developed using a single synthesis technique.
2.5.1 Vaporization of a Carbon Target
The laser ablation method is based on vaporization of carbon in an inert
atmosphere and can produce either SWNTs or MWNTs. The pulsed laser
(Guo, Nikolaev et al. 1995; Guo, Nikolaev et al. 1995; Yudasaka, Komatsu et al.
1997) evaporates carbon from the graphite target (~1500 °C) and the products
are transported by an inert gas flow (He or Ar) from the high-temperature zone to
a water-cooled copper collector. In presence of a transition metal catalyst, the
method yields between 70-90% SWNTs (Guo, Nikolaev et al. 1995) and
bimetallic catalysts are more efficient (Guo, Nikolaev et al. 1995; Guo, Nikolaev
et al. 1995; Thess, Lee et al. 1996; Sun, Mao et al. 1999; Yudasaka, Sensui et al.
1999; Yudasaka, Yamada et al. 1999; Hernandez, Ordejon et al. 2000). The
presence of boron in the system increases the length of the SWNTs, possibly by
preventing the closure of the CNT tip and promoting further growth (Blase,
Charlier et al. 1999). When a porous graphite target is used, the SWNT
production is enhanced (Yudasaka, Zhang et al. 2000).
23
2.5.2 Solar Energy Vaporization
A solar furnace is used to focus sunlight onto a graphite sample, giving
temperatures up to 3000 K (Laplaze, Bernier et al. 1994). The method can be
used to produce fullerenes (Chibante, Thess et al. 1993; Fields, Pitts et al. 1993;
Bernier, Laplaze et al. 1995; Laplaze, Bernier et al. 1996; Laplaze, Bernier et al.
1997; Laplaze, Bernier et al. 1997) in the absence of added catalysts, or either
bamboo-shaped MWNTs (Co/Ni, at 250×10-3 bar) or SWNTs (Co/Ni at 450×10-3
bar, Co or Ni/Y at 7×10-8 bar) (Fields 1996; Heben 1996; Laplaze 1996; Laplaze,
Bernier et al. 1997). The nanotube yield is generally low.
2.5.3 Electric Arc Discharge
Originally, the electrical arc method was used to produce fullerenes, and MWNTs
were found as a by-product (Iijima 1991). An electric arc is generated between
two opposing graphitic electrodes in a reactor filled with He or Ar (Ebbesen and
Ajayan 1992). A temperature of approximately 4000 K is reached between the
rods. When one of the electrodes is movable, the reaction is semi-batch. The
evaporation of pure graphite (Iijima 1991) produces fullerenes, amorphous
carbon, and graphitic sheets on the reactor walls, and fused MWNTs and
nanoparticles on the electrode (a typical product is (Ebbesen and Ajayan 1992;
Ebbesen 1994) Di=1-3 nm, Do=2-25 nm, and L=1 µm). The co-evaporation of
graphite with a metal or metal salts gives MWNTs on the cathode (Guerret-
Piecourt, Le Bouar et al. 1994; Ata, Hudson et al. 1995; Loiseau and Pascard
1996), (Kim, Rodriguez et al. 1991; Ajayan, Colliex et al. 1994; Lin, Wang et al.
1994; Seraphin, Zhou et al. 1994; Saito, Kawabata et al. 1995; Loiseau and
Pascard 1996). In some cases, SWNTs can be found as well (Subramoney
1993; Kiang, Goddard et al. 1994; Saito, Kawabata et al. 1995). The yield and
purity of SWNTs differs depending on the catalyst (Bethune 1996; Kiang and
Goddard 1996; Kiang, Goddard et al. 1996; Setlur, Lauerhaas et al. 1996;
Journet, Maser et al. 1997; Takizawa, Bandow et al. 1999; Yamashita, Hirayama
et al. 1999; Takizawa, Bandow et al. 2000) and the presence of H2 in the reactor
24
(Wang, Lin et al. 1995). It is possible to produce large quantities of MWNTs per
run, but the purity is ~20-50%.
2.5.4 Electrolysis
A wide range of nanomaterials can also be generated through electrolysis in
molten alkali halide salts using carbon electrodes. The process takes place in a
furnace either in air or under inert conditions. A current (3-5 A for nanotube
production) is applied through the graphite rod, the cathode erodes and particles
can be found dispersed in the melt after 3-4 minutes (Hsu, Hare et al. 1995; Hsu,
Terrones et al. 1996). The solid product is transferred into a toluene phase.
Spiral and curled CNTs are produced with L=5 µm and Do=2-20 nm.
2.5.5 Chemical Vapor Deposition
Chemical vapor deposition synthesis depends on metal catalysts that can be
deposited on a substrate, either in situ or prior to nanotube synthesis.
Nanotubes grow as a gaseous carbon source, usually a hydrocarbon,
decomposes on the catalyst particles and forms graphitic carbons.
Temperatures range from 873 to 1273 °C in most systems. The produced
MWNTs have high purity and are relatively long, with lengths reported even on
the millimeterscale (Pan, Xie et al. 1998). A wide range of metals have been
investigated as catalysts (Andrews, Jacques et al. 1999; Cui, Lu et al. 1999;
Iwasaki, Motoi et al. 1999; Gao, Wang et al. 2000; Grobert, Terrones et al. 2000;
Kamalakaran, Terrones et al. 2000; Lee, Park et al. 2000; Willems, Konya et al.
2000; Cassell, Verma et al. 2001; Mayne, Grobert et al. 2001). The morphology
of the tubes varies with the chosen catalyst (Gao, Wang et al. 2000; Grobert,
Terrones et al. 2000; Lee, Park et al. 2000; Willems, Konya et al. 2000) and by
using a feed-gas containing nitrogen (Grobert, Terrones et al. 2000; Nath,
Satishkumar et al. 2000), it is possible to incorporate significant amounts of
nitrogen in the tubes. CVD has also been used to produce SWNTs (Hafner,
Bronikowski et al. 1998; Cassell, Raymakers et al. 1999; Franklin and Dai 2000).
Bimetallic catalysts are predominately used to produce SWNTs, but can catalyze
the growth of MWNTs as well.
25
2.5.6 Sonochemical Production of Carbon Nanotubes
Nanotubes can be produced through a homogeneous sonochemistry process
(Katoh, Tasaka et al. 1999). The reaction is very fast and takes place at the ‘hot
spot’ created right at the tip of the sonication probe, where the temperature is
thought to reach over 5000 K. The molecules in the hot spot get pyrolyzed; with
a liquid-solid mix (ex. Benzene-metal particles), the reaction takes place on the
liquid-solid interface. However, the organic liquid can decompose and/or form
polymers. o-Dichlorobenzene in combination with ZnCl2 produces highly
crystalline nanotubes.
2.5.7 Low Temperature Solid Pyrolysis
This method is used to produce capped MWNTs in situ (Li, Yu et al. 1997), with a
morphology of L=0.1-1 µm, Di≥1.02 nm, and Do=10-25 nm. A refractory
metastable compound, such as nano-sized silicon carbonitride, is used as the
carbon source. The powder is pyrolyzed for 1 hour in an N2–filled graphite
furnace at temperatures between 1500-2200oC (Toptimum = 1700oC). The tubes
are formed on the powder surface and SCN is found in the hollow core.
2.5.8 Catalyst Arrays
A porous anodic aluminum oxide template (thickness of 1 µm, hole diameter of
80 nm) was used to create MWNTs of uniform length and diameter (Jeong, Lee
et al. 2002). The produced nanotubes have uniform thickness, but uneven
lengths. The nanotube products could be cut by sonicating the template. After
the production is complete the template is etched away, leaving nanotubes of
even morphology and good crystallinity.
SWNTs and MWNTs are produced in-situ by reducing a composite metal oxide
powder in a H2-HC atmosphere (Peigney, Laurent et al. 1997). A nanotube net is
created consisting of both SWNTs bundles (Dbundle<100 nm, Lbundle up to 100 µm)
and MWNTs (D=1.5-15 nm).
26
2.5.9 Synthesis from Polymers
Carbon nanotubes can be synthesized from the carbonization of emulsion
copolymers of polyacrylonitrile, PAN, and poly(methyl methacrylate), PMMA.
Micro-spheres with an outer shell of PAN and a center of PMMA are blended with
a PMMA matrix and spun into fibers. Heat-treatment for 30 minutes at 900oC in
inert atmosphere results in complete disappearance of PMMA and carbonization
of PAN to MWNTs (Hulicova, Hosoi et al. 2002). The micro-sphere sizes
determine the size of the produced nanotubes and the PAN-shell thickness
determines the nanotube wall-thickness.
Another version is chemical treatment of a polymer primarily consisting of carbon
to promote polymerization, followed by a 400oC heating session in an air-filled
furnace for 8 hrs (Cho, Hamada et al. 1996). The product consists of MWNTs
with D=1-20 nm and L<1µm.
2.6 Dispersion Challenges
The production methods for carbon nanotubes often result in products that have
varying diameters and lengths, and may have impurities that should be removed
prior to use as have been discussed previously. The CNT product may also
contain CNTs that are physically or chemically entangled.
One of the areas to which such entanglements may introduce problems, is in the
generation of different types of homogeneous CNT dispersions. Highly
entangled products, which are difficult to dispersion uniformly in fluids and melts,
result in fluid suspensions and composites with only modest improvements in
mechanical or transport properties. Aggregates are generally found to be an
impediment to most carbon nanotube applications. The aggregates may not
provide three-dimensional networks that efficiently carry mechanical loads or
transport properties for the mix or composite, hence losing the desired effects. In
this section we present a few of the problems caused by the nature of CNTs.
27
2.6.1 Physically and Chemically Entangled Carbon Nanotubes
The extremely high aspect ratios in combination with the high flexibilities
dramatically increase the possibilities for entanglements. Entangled aggregates
can be very hard to break up without damaging the nanotubes in different ways.
Entanglements can be divided into two subcategories; physical entanglements
and chemical entanglements. Physical entanglement of the nanotubes is defined
here as individual particles that are entwined, interwoven, or form loops around
other nanoparticles.
“Chemical” entanglements can also occur, and can be interpreted as surface-to-
surface attraction, adhesion, or self-assembly. Attractive van der Waals forces
between carbon surfaces increase the dispersion difficulty. Carbon surfaces
tend to be attracted to each other. For example, studies have shown that
dispersion of carbon black in diglycidyl ether-4,4-diaminodiphenyl sulfone
copolymer becomes more difficult as the surface area of the carbon black
increases, due to attractive forces between the aggregates (Kumar and Chandra
1989). As mentioned earlier, SWNTs have sp2/sp3 hybridization, which may
explain the pronounced attraction between the tubes. Furthermore, SWNTs from
research reactors are harvested in bundles, which are very hard to break up.
Interstitial channels are formed between the tubes in the bundles. These
channels adsorb a larger quantity of gas than the actual nanotube cores do, so
for an adsorption type application it might not be of interest to break up the
bundles. However, when using SWNTs as a reinforcing agent or conductive
filler, dispersing the bundles would increase uniformity and keep production costs
down.
2.6.2 Physical Contacts between Nanotubes in Fluids
The high aspect ratio of carbon nanotubes can lead to physical contacts between
separate particles dispersed in fluids, leading to three-dimensional networks that
can increase transport properties, such as electrical and thermal conductivity, but
can also increase suspension viscosity. The transport property changes can be
related to percolation theory predictions. However, there are several physical
28
attributes of carbon nanotube suspensions that are not well described by
percolation models, including the flexibility of the nanotubes and their orientation
under shearing condition.
Several methods are already demonstrated for control of nanotube orientation
once the tubes have been dispersed individually in fluids. These include shear
flows, elongational flows, electric fields, and magnetic fields. The preferred
orientation for a given application may not mean alignment in one direction:
several important applications depend on achieving three-dimensional networks
of high-aspect-ratio nanotubes to modify the transport properties of the fluid or
solid.
In some cases, it is desirable to align the nanotubes in a composite, a process
that requires the tubes to be well separated. It is possible to align SWNTs in
urethane or acrylate, either by shear flow prior to polymerization or by stretching
the already cured matrix (Wood, Zhao et al. 2001). There have also been
attempts to align MWNTs in a thermoplastic polymer by means of stretching (Jin,
Bower et al. 1998; Bower, Rosen et al. 1999). The films were stretched up to
500% and no broken tubes were detected, even though some tube buckling was
observed. In addition, it is possible to align carbon nanotubes by taking
advantage of their electrical conductivity. One research group dispersed carbon
nanotubes in ethanol in a DC-field (250 V/mm) for half a minute. Tube alignment
was observed and moreover longer tubes were collected at the cathode, which
means that the method can be used for purification and size-grading of
nanotubes (Bubke, Gnewuch et al. 1997). SWNTs shortened by ultrasonication
(l=20-100 nm) were aligned along a HOPG-lattice (highly oriented pyrolytic
graphite). The film showed a semi-conductive behavior (Yanagi, Sawada et al.
2001). Solvent casting and melt mixing of SWNTs and PMMA in a hot-press has
also been used to promote nanotube alignment. Higher electrical conductivity
was detected in the flow direction (0.118-11.5 S/m) than that in the perpendicular
direction (0.078-7 S/m) (Haggenmueller, Gommans et al. 2000).
29
2.7 Dispersion Methods
There are two different approaches to nanotube dispersion: mechanical (or
physical) methods and chemical methods. Mechanical dispersion methods, such
as ultrasonication, separate nanotubes from each other, but can also fragment
the nanotubes, decreasing their aspect ratio during processing. Chemical
methods use surfactants or functionalization to change the surface energy of the
nanotubes, improving their wetting or adhesion characteristics and reducing their
tendency to agglomerate in the continuous phase solvent. However, aggressive
chemical functionalization, such as using neat acids at high temperatures, can
digest the nanotubes. Both mechanical and chemical methods can alter the
aspect ratio distribution of the nanotubes, resulting in changes in the properties
of their dispersions.
Te sections below review literature reports and discussions regarding chemical
dispersion methods, surfactant treatment, and production technical solutions. In
section 2.7.4 we also discuss ball-milling, a popular mechanical dispersion
method. Another mechanical method commonly used is ultrasonication. We
have investigated the method thoroughly, together with mechanical grinding of
MWNT paste, and we will present and discuss these mechanical dispersion
methods together with our research data and conclusions in Chapter 6.
Finally, in section 2.7.5 we briefly review electrical methods and in Chapter 7 we
present the research we have been doing in this field.
2.7.1 Chemical Dispersion Methods
If the carbon nanotube surface is modified through functionalization, the
interaction between the tube and the surroundings is affected. By choosing a
specific type of functional group, it is possible to influence the interaction in
different ways. This treatment may enhance nanotube dispersion in different
solvent and polymer systems, allowing flexibility in creating novel hybrid
composites. For example, by cova lently attaching alkene groups to SWNT side-
walls, the solubility in various organic solvents (Chen, Hamon et al. 1998;
30
Holzinger, Vostrowsky et al. 2001; Chen, Liu et al. 2002; Georgakilas, Kordatos
et al. 2002), such as THF, chloroform, methylene chloride (Boul, Liu et al. 1999)
and DMF (Banerjee and Wong 2002), increases. The nanotubes can also be co-
polymerized in a polymer matrix as a block, graft or cross-linked polymer
(Varadan 2001; Tour, Bahr et al. 2002). Below follows a brief review of
functionalization methods investigated.
Acidic Treatments
Acid treatment in combination with thermal oxidation or decomposition, is a very
common way to purify both MWNTs and SWNTs (Moon, An et al. 2001; Huang,
Kajiura et al. 2002). Usually the nanotubes are boiled or soaked in acid for time
periods of hours or even days and thereafter burned in oxygen or air, pure or
mixed with other gases. In the first step, the catalytic metal particles that are left
in the sample are eliminated, and in the second step unwanted carbon, such as
amorphous carbon and carbon particles, are consumed. Oxidation may separate
SWNTs into individual fibers (Esumi, Ishigami et al. 1996; Chen, Rao et al.
2001). Furthermore, oxidation also cuts the nanotubes in sites with high
structure damage or defect density, which leads to the production of shorter
nanotubes (Colomer, Piedigrosso et al. 1998; Stepanek, De Menorval et al.
1999; Lee and Lee 2002) with a higher morphological quality. Also, thinner tubes
are usually more reactive than tubes with a larger diameter, due to the greater
strain on the bonds in the thinner tubes (Nagasawa, Yudasaka et al. 2000; Bahr,
Yang et al. 2001). When MWNTs are boiled in acid, their diameters decrease.
This mechanism is pronounced for tubes with stronger curvature. A molecular
dynamics study suggests that there might be negative effects from chemical
modifications of SWNT. For small tubes, the strain can cause the functional
groups to dissociate. The helical conformation has little impact. Furthermore,
introduction of sp3-hybridized carbon due to chemical functionalization does
degrade the strength of SWNT by ~15% (Garg and Sinnott 1998). This also
explains why the nanotube active sites are theoretically concentrated on the end
caps (Mawhinney, Naumenko et al. 2000), where the tube geometry presents its
31
largest curvature. In one investigation, Raman spectra show that functional
groups of acidic nature, such as –C=O, -COOH, and –OH, were attached to the
tube ends. After a sufficient period of boiling (2 hours) the sample could not
reach pH=6 when washed in de-ionized water, since the nanotubes cannot
precipitate due to the attractive forces between the acidic groups and H2O.
Furthermore, 18% of the sample had been consumed at this time. If nanotubes
are ultrasonicated in HNO3 or HNO3/H2SO4, acidic functional groups are created
as well. Also in this case, the functional-group quantity is time-dependent. The
oxygen concentration is eight times higher for treated tubes than for non-treated
tubes (Esumi, Ishigami et al. 1996).
Successful opening of nanotube ends have been performed using a variety of
reagents, such as HF/BF3 and OsO4. In addition, the tubes were filled with
metals and the ends can be closed again, or covered, by treatment with benzene
vapor, or with reagents such as ethylene glycol, in a reducing atmosphere of
argon and hydrogen at high temperature (Satishkumar, Govindaraj et al. 1996).
For SWNTs, it has been reported that it is possible to achieve a purity of >95%
with a yield of 20-50% (Jeong, Kim et al. 2001; Jeong, Kim et al. 2002). The
purification yield was reported at 54% for MWNTs (Jeong, Kim et al. 2002)
Oxidation can be carried out both in a liquid-phase and in a gas-phase.
Investigations suggest that primarily hydroxyl and carbonyl groups are formed
during gas-phase treatment of MWNTs, while liquid-phase treatment forms
carboxylic acid groups (Ago, Kugler et al. 1999). It has also been seen that
SWNTs treated with HNO3/H2SO4 or H2O2/H2SO4 mixtures produces both
carbonyl and ether functional groups (Kuznetsova, Popova et al. 2001). It is
possible to remove the groups by thermal decomposition in vacuum. Ethers and
quinones are harder to remove than carboxylic acid groups.
The acidic surface sites that are created on the nanotube surface are reactive
and act as nucleation sites functionalizing the tubes in other ways. One example
is the decoration of nanotubes with nanoscale metal clusters (Au, Pt, and Ag)
(Satishkumar, Vogl et al. 1996; Yu, Chen et al. 1998). Mercury-modified
32
nanotubes have also been prepared, but the mercury was introduced as a salt
solution (Bond, Miao et al. 2000).
A few one-step procedures have also been developed. In one process the
sample is allowed to react with an acidic gas mixture at elevated temperatures
(Lee and Yoo 2000). Carbon dust and transition metals are simultaneously
removed during this treatment, which is relatively short, 10-30 minutes. Another
process combines air treatment and acid microwave digestion to remove a high
percentage of the metal and decreasing the purification operation time (Martinez,
Callejas et al.). There have also been reports claiming that it is possible to
selectively create surface oxygen groups on the inner nanotube wall, by HNO3
oxidation (Kyotani, Nakazaki et al. 2001).
The reports on surface area change differ greatly. Some groups claim that the
surface area and H2 storage capacity increase considerably (Eswaramoorthy,
Sen et al. 1999; Zhu, Chen et al. 2000). Other groups report the complete
opposite result, due to material compaction (Stepanek, De Menorval et al. 1999).
Compared to ball-milling, acidic boiling is more destructive. According to this
particular study, the same quantity of –OH, -COOH, and –C=O were found after
20 minutes of ball-milling, but only 3.5 % of the sample was consumed (Jia,
Wang et al. 1999).
Fluorination
Fluorination is a common first step in the attempt to functionalize SWNTs and
MWNTs. The nanotube is reacted with F2 gas and thereafter some of the F
substituents can be exchanged by nucleophilic substitution, leading to sidewall
functionalization. After the reaction, the remaining F can be partially or
completely eliminated (Margrave, Mickelson et al. 2000). Hydrazine (Margrave,
Mickelson et al. 1999) and LiBH4/LiAlH4 can be used for defluorination of tubes
(Chiang, Michelson et al. 2000). Preferred nucleophiles include alkyl-lithium
species (Margrave, Mickelson et al. 2000). It has been established that the
electrical properties of MWNTs change after fluorination, leading to a wide range
of electrical structures, from insulating to semi-conducting and metallic-like
33
behavior (Seifert, Kohler et al. 2000; Mukhopadhyay, Yokoyama et al. 2001; Gu,
Peng et al. 2002).
Other Types of Functionalization
It is possible to attach different types of functional groups to an individual SWNT
by applying a potential field (Kooi, Schlecht et al. 2002), so-called
nanolithography or electrochemistry. Furthermore, it is also possible to control
the thickness of the deposit layer by varying the treatment duration. It is
assumed that the reaction takes place primarily on defect sites on the tubes.
MWNTs, SWNTs, and graphite can also be functionalized by hydrogen. A way to
functionalize the tubes without creating covalent bonds is to immobilize
molecules on the nanotube surface by π-π-interaction (Wang, Liu et al. 2002).
This way the structure of the nanotube is not disturbed, keeping the electrical
properties intact. Nanotubes have also been modified in order to investigate bio-
systems, such as the adsorption behavior of proteins, taking place on a
functionalized sidewall. For example, SWNTs have also been modified by co-
functionalization of protein-resistant polymers and biotin, creating a specific
binding of streptavidin onto the SWNT (Shim, Kam et al. 2002). Amino-dextran
and periodate-oxidized dextran chains have also been immobilized onto plasma-
treated nanotubes. The resulting polysaccharide grafted nanotubes are very
hydrophilic (Chen, Dai et al. 2001).
Surfactants and Dispersion Agents
The role of a surfactant is to produce an efficient coating and induce electrostatic
repulsions that could counterbalance van der Waals attractions (Hunter 1986). In
our case, the electrostatic repulsion provided by adsorbed surfactants stabilizes
the nanotubes against the strong van der Waals interaction between the tubes,
hence preventing agglomeration. Polymer-coated objects experience a
reversible force. The polymer adsorbs onto the nanotube and repulsive forces
dominate over attractive van der Waals forces between the carbon nanotubes.
This balance of repulsive and attractive forces creates a thermodynamically
34
stable dispersion, which might even result in separation of SWNTs from the
bundles into individual nanotubes.
Surfactants have been used to disperse carbon nanotubes in several cases
(Duesberg, Muster et al. 1998; Hwang 2000; Riggs, Walker et al. 2000; Sun,
Riggs et al. 2000). Some examples of commonly used surfactants are sodium
dodecyl sulfate (SDS), both for SWNTs (Duesberg, Burghard et al. 1998; Liu,
Rinzler et al. 1998; Vigolo, Penicaud et al. 2001) and MWNTs (Burghard,
Duesberg et al. 1998; Lewenstein, Burgin et al. 2002), and Triton X-100 (TX-100)
(Liu, Rinzler et al. 1998; Riggs, Walker et al. 2000; Lewenstein, Burgin et al.
2002) for SWNTs. SWNTs have also been dispersed using lithium dodecyl
sulfate (LDS) as surfactant (Krstic, Duesberg et al. 1998; Krstic, Muster et al.
2000).
In contrast, poly(vinyl alcohol) (PVA) is not efficient at stabilizing the tubes
(Vigolo, Penicaud et al. 2001). Negatively charged SDS, positively charged
cetyltrimethyl ammonium chloride (CTAC) and dodecyltrimethyl ammonium
bromide (DTAB), nonionic pentaoxoethylenedodecyl ether (C12E5),
polysaccharide (Dextrin), or long chain synthetic polymer poly(ethylene oxide)
(PEO) could not act as efficient dispersing agents for SWNTs in aqueous
solutions (Bandyopadhyaya, Nativ-Roth et al. 2002). However, Gum Arabic (GA)
is reported to be an excellent surfactant for SWNTs in water. Both SDS (Vaeck
and Merken 1981; Asakawa, Kaneko et al. 1986; Ma and Xia 1992) and Triton X-
100 (Sandstrom 1983; Zhao and Gu 1987; Garamus and Pedersen 1998;
Gonzalez-Garcia, Gonzalez-Martin et al. 2000) have previously been used as
surfactants for carbon black. A problem with surfactant-induced dispersions is
finding a feasible way to remove the surfactant from the final product (Nagahara,
Burgin et al. 2001).
35
2.7.2 Technical Dispersion Approaches
In-situ Production of CNTs
One way to solve the dispersion problem is to try to produce the carbon
nanotube directly into the matrix, so called in-situ production. Attempts have
been made to produce MWNT in several different M-Al2O3 composites, where M
can be Fe, Co, Ni, or their alloys. The production step was followed by hot-
pressing. Both MWNT and SWNT were found in the composite. However, the
composites showed no improvement in physical properties (Peigney, Laurent et
al. 1997; Peigney, Laurent et al. 1999; Peigney, Laurent et al. 2000). As
mentioned earlier, carbon nanotubes have been hot-pressed in various metal
oxides. The tubes were destroyed when the temperature in the hot-press
exceeded 1200oC which results in the absence of property improvements
(Flahaut, Peigney et al. 2000).
Production Technical Solutions
SiO2 glass rods of micrometer sizes (l=10-15 µm, d=0.5-1.5 µm)), containing 6
wt% nanotubes, were prepared with the aid of a surfactant. The micro-rods were
incorporated in SiO2 tablets with a loading of 60 wt%, to promote nanotube
dispersion in the tablet. The Vicker’s hardness increased in both cases, but the
effect was larger for micro-rods with nanotubes (~100% increase) than for rods
without. The micro-rods were not broken at the tablet fracture surface (Hwang
and Hwang 2001).
36
2.7.3 High Impact Mixing - Ball-Milling
Ball-milling has been used to narrow the length and diameter distributions (Kim,
Hayashi et al. 2002) and opens the nanotubes (Li, Wei et al. 1999; Pierard,
Fonseca et al. 2001) for improved sorption capacity for gases. However, a large
amount of amorphous carbon is created (Li, Wei et al. 1999; Haluska, Hulman et
al. 2001), which clearly indicates that the tubes are damaged in different ways
and that ball-milling is a destructive method. The creation of amorphous carbon
introduces a high surface area, which is a more likely explanation to the
increased storage capacity than the introduction of open tube ends would be
(Haluska, Hulman et al. 2001; Niesz, Nagy et al. 2001).
Bending, defects and tube-tube contacts strongly modify the electrical behavior
of carbon nanotubes. Struc tural topological defects always increase the
resistance of metallic nanotubes, making it dependent on the defect density per
unit length (Jia, Wang et al. 1999).
Some groups have tried to produce boron nitride nanotubes and carbon
nanotubes from graphite through ball-milling. The iron from the mill balls
functions as a catalyst and the heat becomes elevated through the mechanical
impact, so the process parameters are in the right region, but the results have
not been impressive (Chen, Fitz Gerald et al. 1999; Chen, Fitz Gerald et al. 1999;
Chen, Chadderton et al. 2000; Awasthi, Kamalakaran et al. 2002). In addition,
the created nanotubes are destroyed after a longer period of milling.
Ball-milling has also been used in an attempt to intercalate lithium in SWNTs,
creating compounds to be used in batteries (Gao, Bower et al. 2000). Li-
intercalated graphite and carbonaceous materials are commercially used in Li-ion
batteries (Pistoia and Editor 1994; Kumar 1995). The intercalation involves
electron donation from the alkali metal to the nanotube. The same type of
experiments have also been carried out with K, Rb, and Cs on both SWNTs and
MWNTs (Zhou, Fleming et al. 1994; Suzuki and Tomita 1996; Lee, Kim et al.
1997; Bower, Suzuki et al. 1998; Andreoni and Editor 2000; Slanina 2001).
37
2.7.4 Electrical Dispersion Mechanisms
It is possible to break carbon nanotubes by applying an applied electrical field
externally. The CNTs can then emit electrons and if the applied voltage exceeds
the binding energy of a carbon atom, the atoms can be split off the tip. As a
result, external graphite layers get broken and may peel off (Wang, Gao et al.
2001).
Open-ended NTs placed close to each other can even bond to form conducting
electrical contacts (Buongiorno Nardelli, Fattebert et al. 2000). Bending, defects,
and tube-tube contacts strongly modify the electrical behavior of NT. Structural
topological defects always increase the resistance of metallic NTs to an extent
that is strongly dependent on the defect density per unit length. Fusion has also
been observed in other cases, not involving electrical fields. In one investigation
MWNTs were extended 2-12% lengthwise before failure. The breakage creates
two open tube-ends, but they fuse together internally in a matter of tenths of
seconds, resulting in two new, closed tubes (Hwang and Hwang 2001).
Copyright © Jenny Marie Hilding 2004
38
Chapter Three
Morphology Measurements and Modeling
3.1 Introduction
As can be seen in Chapter Two, a large number of property evaluations of
carbon nanotubes and composites containing CNTs have been measured and
presented during the past decade. Surprisingly, there is a severe lack of
property analysis in relation to CNT morphology. One major goal of the research
presented in this thesis is to investigate the MWNT morphology impact in
different areas of chemical engineering. The ability to determine the nanotube
morphology is crucial if we wish to fully understand and scientifically explain the
relation between nanotube morphology and nanotube properties. It was
therefore an absolute necessity for us to develop and formulate a standardized
method to measure and characterize the morphology of any carbon nanotube
sample.
This chapter presents such a method, which has been developed in our lab to
permit us to relatively quickly and easily measure the MWNT morphology and
describe it with accuracy and in detail. The method is a useful aid in the analysis
of wide range of different systems and processes involving MWNTs and has had
a significant impact on all the MWNT research we have done. The ability to
accurately determine the sample morphology is one of the contributing factors to
what makes our research truly unique. Such systems and processes will be
thoroughly described and discussed further in later chapters. This type of
approach has previously never been introduced or applied in the field of CNT
research before.
39
3.2 Visual Techniques for Morphology Measurements
The unusually large CNT aspect ratio introduces additional challenges in
measuring a MWNT sample’s morphology. For systems where the particle’s
dimensions are fairly uniform it is possible to use different computerized
measurement and analysis tools as an aid in determining the particle distribution.
This is not possible for a CNT system. It is therefore necessary to manually
measure enough particles and thereby generate sufficient information of the
sample morphology. The process is slow and presents a wide range of
challenges, which probably explains why so few CNT research groups
investigate the morphology of the samples they are basing their research on.
A few different visual instruments can be used to determine the MWNT
morphology. Below is a short presentation of the three instruments that have
frequently been used in our laboratory to investigate different MWNT systems.
3.2.1 Transmission Electron Microscopy, TEM
A few samples have been analyzed with the aid of Transmission Electron
Microscopy, TEM (JEOL, JEM-2000 FX Electron Microscope). Generally, a
hundred tubes from each sample were measured. The TEM has a very high
magnification and can give detailed information about the structure of a single
nanotube. The nanotubes appear ‘transparent’ in the TEM pictures, which
enables the measurement of both inner and outer diameter. An example is given
in Figure 3.1. As can be seen in Figure 3.1, the pore is centered at the core of
the MWNT and the diameter of the pore is large compared to the distance
between the multiple graphite -layers. It is also possible to determine the number
of layers in a MWNT tube wall and whether it contains any type of structural
damage or irregularities. One other type of interesting observation permitted by
the TEM is the shape and location of any residual catalyst that may be
incorporated in the tube.
One disadvantage with the TEM is that it has to be manually calibrated by taking
pictures of special calibration particles (latex gold shadowed resolution
40
calibration particles with d=0.261 µm from Electron Microscopy Services) several
times during each experimental session. The beads have a known size, which is
used to compare and calculate the measurements of the particles in the sample.
It can also be argued that the TEM is too powerful to generate a fair
interpretation of a MWNT sample. It is far too tempting to concentrate on details
rather than on the general picture. Therefore, a SEM or an optical microscope is
likely to give a more accurate overall view than a TEM, except when very short
(<100 nm) tubes are present.
3.2.2 Scanning Electron Microscopy, SEM
Most of the samples were imaged using Scanning Electron Microscopy, SEM
(Hitachi S-3200 Scanning Electron Microscope). Generally, at least 10 pictures,
from different areas of the sample holder, were taken of each sample. Every
single tube was measured from each picture, resulting in data based on few
hundred tubes from each sample.
The advantage of the SEM is that each picture generally shows several MWNTs.
The magnification can be varied to either concentrate on sample aggregates or
single tube diameters. This enables the measurement of a number of different
geometrical properties during one single experimental session. The SEM comes
with a built-in scale, so the machine is already calibrated. This scale is
automatically printed with each picture, which greatly simplifies data
measurement and conversion.
In general, the SEM is powerful enough to implement MWNT surface
observations, which enables the detection of external impurities or structural
damage. Figures 3.2 and 3.3 are examples of SEM pictures.
41
Figure 3.1 TEM picture showing the transparent appearance of a multiwalled
carbon nanotube. From a TEM picture it is possible to determine how many
graphene layers the tube wall consists of and it is easy to measure both the outer
and the inner tube diameters. The pictures also clearly show surface impurities
and structural damage.
42
Figure 3.2 This SEM picture shows as-produced MWNTs, which have been
harvested from the CVD substrate. The MWNTs grow in a layer, much like
grass, and it is easy to see that the lengths are very similar within the sample.
43
Figure 3.3 This SEM picture clearly shows the difference between the MWNT
diameters within one sample.
44
3.2.3 Optical Microscopy
When the MWNTs are sufficiently large, it is possible to use optical microscopy.
This method significantly shortens the time required for sample preparation and
analysis. Both liquid-dispersed as well as dried samples can be analyzed, which
is a great advantage compared to SEM or TEM, both of which operate under
vacuum. The ability to analyze liquid dispersions can be crucial, especially in
cases when time-dependent samples are analyzed.
The disadvantage with optical microscopy is that it is easy to overlook smaller
MWNTs. It is also hard to focus on dispersions, since the liquid film generally is
thicker than the actual focus range. Practically this means that MWNTs may
move in and out of focus during an analysis session.
The scale is built-in and can unfortunately not be transferred to the actual printed
pictures. It is therefore necessary to compare each sample picture to a picture of
a calibration grid taken at the same magnification.
3.3 Multiwalled Carbon Nanotube Samples
The MWNTs used to measure the morphology were produced at CAER (Center
for Applied Energy Research) through chemical vapor deposition, CVD
(Andrews, Jacques et al. 1999; Jacques, Villain et al. 2000). As can be seen in
Section 2.5.5, this technique produces aligned nanotube mats with no bundle
formation, very few nanotube contact points, homogeneous lengths, and well-
defined diameters. The MWNTs were cut from their substrate, so that all
nanotubes were open to the atmosphere. Typical SEM pictures showing MWNT
length and diameters can be seen in Figure 3.2 and Figure 3.3 respectively. The
MWNTs have a purity of >95% (the non-nanotube component is the remnant of
the iron catalyst used to grow the nanotubes). TEMs show essentially no
amorphous carbon within the mats or adhering to the nanotube exteriors.
45
3.3.1 Sample Preparation and Diameter Measurements
To measure and evaluate the diameter distributions, TEM or SEM were used.
The samples were dispersed in toluene or acetone to enable quick and complete
evaporation from the sample holder. Evaporated substances can create a film
on the electron beam, shortening its life span dramatically.
It is desirable to keep the MWNT concentration low, to avoid multiple layers or
entangled tubes on the sample holder. To separate the MWNTs, the dispersions
were treated in an ultrasonication bath. The time in the bath was kept to a
minimum, on the order of 5-10 seconds, to avoid breakage of the tubes.Five
different samples were measured to determine which size-distribution model fits
the MWNT diameter data best. The CVD production parameters had been
slightly altered between the runs to cause variations in the sample morphologies.
The data were collected from the pictures and transferred into Excel®
spreadsheets and converted according to the photomicrograph scale.
Dr Dali Qian measured diameters for samples 1-3 from TEM pictures (Dali 2001).
One hundred different MWNTs were measured from each sample. Samples 4-5
were measured from SEM pictures. Measurements for 481 and 359 MWNTs
were recorded for sample 4 and sample 5 respectively.
3.3.2 Morphology Results for CVD-Produced MWNTs
The MWNT length was measured from SEM pictures and established to be
uniform in the range of 50-55 µm. TEM photomicrographs showed that the
MWNTs are open at the ends and that both inner and outer diameters vary quite
a bit from tube to tube within a sample.
Since samples 4-5 were analyzed using SEM, their inner diameter distributions
could not be measured. However, it was assumed that the pores were in the
same range as samples 1-3 (Dali 2001), since previous workers have shown that
MWNT diameters are directly related to the size of the iron catalyst nanoparticle
(Andrews, Jacques et al. 1999; Jacques, Villain et al. 2000).
46
The collected inner and outer diameter data are plotted in Figures 3.4 and 3.5,
respectively. From Figure 3.4 it can be seen that the samples display pore
diameters between 2 and 16 nm, so the MWNT cores are strictly within the
mesoporous range (radius 1-25 nm by definition). The outer diameter varies
from ~16 up to almost 400 nm. Sample 1 seems to have a more narrow
distribution with a smaller average diameter, followed by samples 4, 5, 2, and 3.
3.4 Morphology Model
A probability density function needed to be fitted to the collected diameter
distribution data in order to obtain the distribution’s moments. Fitting a
differential distribution accurately might require as many as one thousand data
points and is practical when digital imaging software can be applied to the
problem. An alternative method is to determine the probability density function
coefficients of cumulative particle -size distributions when data in the order of 50–
500 nanotubes are available. This approach reduces the instrument time for
length measurements and often provides sufficient accuracy for engineering
models of the fragmentation process.
Gaussian distributions are often assumed to represent particle -size distributions.
However, we found that several different MWNT geometrical distributions
(diameters, the length of fractured MWNT, MWNT agglomerate sizes, etc) are
better described by log normal distributions, which have greater fractions of
higher length scales than do Gaussian distributions. The log-normal distribution
accounts for the long “tail” of larger sized nanotube geometries. If we call a
general geometrical property of interest L, we can fit a log-normal probability
model according to (Hilding, Grulke et al. 2001; Hilding, Grulke et al. 2003)
2)ln(
21
21
))(ln(
−−
⋅= σ
µ
πσ
L
eLf Equation 3.1
47
The MWNT geometrical measurements were fitted using either Excel® or
Systat®. The differential probability density function has two fitting parameters,
the logarithmic mean, µ, and the standard deviation of the distribution, σ.
The cumulative probability distribution function of this distribution is the integral
from negative infinity to ln(L) of Equation 3.1.
)ln(2
1))(ln(
)ln( )ln(21 2
LdeLFL L
∫∞−
−
−
⋅
⋅= σ
µ
πσ Equation 3.2
The interpretations of all carbon nanotube geometry data presented in this thesis
are based on log-normal size distributions.
Models for all five samples’ outer diameter distributions were fitted with the log-
normal cumulative distribution. The inner diameter data was fit in the same
manner for samples 1-3.
The log-normal cumulative distribution fits are plotted in Figures 3.4 and 3.5
together with the recorded diameters. The model readily fits the data collected
from each sample. Tables 3.1 and 3.2 list the log-normal probability distribution
parameters for inner and outer diameters, respectively. These parameters can
be interpreted to give an understanding as to how the morphological distribution
looks.
The difference between the sample-distributions may be easier to detect in a log-
normal probability as supposed to a cumulative distribution plot. For example, the
average diameter for each sample, d=eµ, also presented in Tables 3.4-5, can be
found at the peak maximum. The broadness of (and hence the height of) the
peak correlates to the width of the distribution range. The inner diameter
probability curve can be seen in Figure 3.6 and the outer in Figure 3.7.
48
Inner Diameter Data & Cumulative Distribution Fit
0
0.2
0.4
0.6
0.8
1
0 5 10 15 20
Diameter [ nm ]
Cum
ulat
ive
sum
Sample 1
Sample 2
Sample 3
Figure 3.4 The inner diameters are distributed within a fairly narrow range.
Sample 2 displays a few larger diameters than the other two samples. The pore
distributions are found to be on the range of 2 -16 nm (Hilding, Grulke et al.
2001).
49
Outer Diameter Data & Cumulative Distribution Fit
0
0.2
0.4
0.6
0.8
1
0 100 200 300 400 500
Diameter [ nm ]
Cum
ulat
ive
sum
Sample 1
Sample 2
Sample 3
Sample 4
Sample 5
Figure 3.5 The outer diameters are spread over a wide range, from ~12 up to
almost 400 nm. Samples 1 and 4 display the narrowest ranges and the smallest
average diameters followed by samples 5, 2, and 3.
50
Table 3.1 The table presents the log-normal probability distribution
parameters, µ and σ, for the inner diameter distribution fits of samples 1-3. The
distribution data was measured from a hundred nanotubes from TEM pictures.
The log-normal average, µ, is also presented as average diameters where, d=eµ.
The R2 values for all these fits are greater than 0.999.
Sample 1 2 3
µ 0.971 0.975 0.875
σ 0.33 0.60 0.60
Diameter [ nm ]
2.64 2.65 2.40
51
Table 3.2 Log-normal probability distribution parameters, µ and s, for the
outer diameter distribution fits of samples 1-3. The distribution data was
measured from TEM (samples 1-3) or SEM pictures (samples 4 and 5). The log-
normal average, µ, is also presented as average diameters, where d=eµ. Again,
the R2 values for these fits are greater than 0.999.
Sample 1 2 3 4 5
µ 3.19 3.91 3.97 3.66 3.78
σ 0.57 0.777 0.898 0.408 0.483
Diameter [ nm ]
24.3 49.9 53.2 39.1 43.8
52
Differential distribution fit
0
0.2
0.4
0.6
0.8
1
1.2
0 2 4 6 8 10 12 14 16 18 20 Diameter, D [ nm ]
Pro
babi
lity
dens
ity
Sample 1 Sample 2 Sample 3
Figure 3.6 The log-normal probability distribution function fit. The curves show
that samples 2 and 3 have a larger diameter probability range than sample 1. It
is also easy to se that the logarithmic mean, µ, is almost exactly the same for all
of the three samples.
53
Differential distribution for outer diameter
0
0.2
0.4
0.6
0.8
1
0 50 100 150 200 250 300 350 400
Diameter [ nm ]
Pro
babi
lity Sample 1
Sample 2Sample 3Sample 4Sample 5
Figure 3.7 The probability plots show that sample 1 has a significantly smaller
average diameter than the rest of the samples, followed by samples 4 and 5. It is
also clear that samples 1 and 4 display more narrow outer diameter ranges, with
sample 5 closely after. Samples 2 and 3 have comparatively wide distribution
ranges.
54
The fits confirm that the mean inner diameters are very similar, but sample 1
displays a more narrow distribution of this dimension as well. Samples 2 and 3
have almost identical inner diameter morphologies. As expected, sample 1 also
has a significantly smaller outer average diameter together with a narrower
distribution.
The log-normal cumulative distribution fit is plotted together with the Poisson and
the Gaussian cumulative fits in Figure 3.8 to show the differences between the
models.
3.5 Geometrical Calculations
Both the relative pore volume and surface area are expected to increase with
decreasing average outside diameter. It is a simple task to calculate geometrical
properties of the sample from its estimated diameter distributions.
If µ and s are known for a specific sample’s diameter distribution, it is possible to
calculate the density, pore volume, and surface area of that sample by
integrating over the log-normal probability function for each sample. However,
since we only recorded a moderate number of data points for each sample, it is
simpler to calculate sums over the entire range instead.
Pore volume and area per sample weight:
2
1 2∑
⋅⋅⋅=iD
i
MWNTpore
DLVρ
π Equation 3.3
∑⋅⋅=D
MWNT
DL
A1ρ
π Equation 3.4
55
Statistical distribution models
0
0.2
0.4
0.6
0.8
1
0 20 40 60 80 100 120 140
Diameter, d [ nm ]
Cum
ulat
ive
dist
ribut
ion
LognormPoissonGaussSample 1
Figure 3.8 The difference between the log-normal, the Poisson and the
Gaussian cumulative distribution models. The outer diameter data for sample 1
has been used to demonstrate that the log-normal model is superior.
56
If we assume that the density of the wall is the same as the density of pure
graphite, ρ=2 045 kg/m3, the sample weight, wMWNT, can be calculated as
−
⋅⋅⋅= ∑∑
2
1
2
1 22
iDi
D
MWNT
DDLw πρ Equation 3.5
Table 3.3 shows the calculated properties for the five samples. As expected, the
samples with smaller average diameters and narrower distributions possess
larger surface areas and pore volumes per gram. The surface areas are in the
range of ~14-55 m2/g and the pore volume is between 0.7 and 5.4 mm3/g.
By comparison, commercial activated carbons have typical surface areas in the
range 1000 m2/g (microporous), 10-100 m2/g (mesoporous), and 1 m2/g
(macroporous). The activated carbon pore volume is generally well above 300
mm3/g.
57
Table 3.3 The geometrical calculations presented here are based on
morphology data from TEM (samples 1-3) and SEM (samples 4 and 5) pictures.
Since length and outer diameter are the only data measured for samples 4 and 5,
it was assumed that the samples had the same average diameter as samples 1-
3. As expected, samples with smaller average diameters and more narrow
distributions have a larger surface area and pore volume per gram of sample.
Sample
1 2 3 4 5
Pore volume [ mm3/g ]
5.40 1.64 0.68 2.82 2.34
Internal surface area [ m2/g ]
5.76 1.11 0.59 3.01 2.06
External surface area [ m2/g ]
50.5 18.3 13.6 38.9 30.2
58
3.6 Summary and Conclusions
Carbon MWNTs have well-defined morphology that can be measured and
modeled accurately. Five different samples were analyzed; three samples from
TEM pictures and two samples from SEM pictures. The inner diameter
distributions were found to be essentially identical between the three samples
analyzed through TEM.
It was established that the data could be fitted well with a log-normal probability
model. The morphology can be analyzed from the two fitting parameters, the
logarithmic mean diameter µ and the standard deviation s.
The pore volume, inner and outer surface areas were calculated from the
collected data for each sample. As expected, the samples with a smaller mean
outer diameter had larger pore volumes and surface areas on a per-weight basis.
For all MWNT samples, the pore volume is only a small fraction of the solid
volume. The surface area differs significantly between the samples.
MWNT morphology data collected from many samples used in the various
applications of this dissertation have convinced us that there can be significant
variation in length, diameters, defects, entanglements, and impurities. The
following MWNT production parameters have been found to contribute to these
variations: CVD reactor temperature, carbon source, catalyst particle loading and
size, position in the CVD reactor, and CVD substrate wear due to usage.
Therefore, we conclude that at this stage of the MWNT production development,
it is necessary to always verify the MWNT morphology before analyzing research
data as a function of morphology.
Copyright © Jenny Marie Hilding 2004
59
Chapter Four
Isothermal Adsorption
4.1 Introduction
As mentioned in the introduction and described more in detail in Chapter Two,
CNTs have created interest both as a gas separation medium and also as a
possible gas storage material. CNTs have been proposed as adsorbents for a
variety of gases, from hydrogen (Dillon, Jones et al. 1997; Liu, Fan et al. 1999;
Stepanek, De Menorval et al. 1999; Wu, Chen et al. 1999), alkanes (Jagtoyen,
Derbyshire et al. 2000), noble gases (Stan, Crespi et al. 1998; Teizer, Hallock et
al. 1999; Teizer, Hallock et al. 2000), nitrogen (Inoue, Ichikuni et al. 1998), to
fluorocarbons (Corbin and Reutter 1997) thanks to their well defined geometry
and uniform pores. A great advantage is that temperature swing adsorption is
possible with MWNTs. Since the MWNTs are electrically conducting, controlled
currents can be used to heat the nanotube bed, resulting in very rapid desorption
and low cycle times in fixed bed operations, an advantage compared to activated
carbons, which usually are non-conductive. However, the small size of the
MWNTs is a potential disadvantage, since they would have to be immobilized in
a practical flow process.
Adsorption and desorption mechanisms should be easily analyzed and computed
from heats of adsorption and morphology information, facilitating the design of
adsorption systems. The study in this chapter was made to thoroughly
investigate the adsorption mechanism for butane on MWNTs in relation to the
MWNT morphology, as a first step toward investigating a hydrocarbon separation
process. To provide a solid theoretical basis to discuss and analyze the research
presented in this chapter from, this introduction illuminates a number of different
adsorption and surface phenomena that are of interest to a MWNT system from a
morphological point of view. Adsorption of gases on graphite might be
60
considered as a limiting case of adsorption on MWNTs in the limit of zero
curvature of the surface.
4.1.1 Hysteresis
Hysteresis is the distinguishing feature of a type IV isotherm, which corresponds
to the adsorption on a meso-porous solid (Avery and Ramsay 1973). The BET
equation is generally applicable to adsorbents with pores in the mesoporous
range. Pores in this range can be produced by compression of powdered solids
(Gregg and Sing 1982). TEM and SEM analyses show that our MWNTs have
few particle-to-particle contacts, hence the sample porosity is strictly due to
accessible nanotube pores. As mentioned in Chapter 3, our MWNT samples are
produced with at least one open end and are a mesoporous material. Butane is
small enough to adsorb through capillary condensation in a pore of this
dimension, but too large to adsorb between the tube walls, which have an
estimated spacing of 0.34 nm (Ajayan 1999).
If the gas phase relative pressure is increased, capillary condensation will occur
in the smaller pores initially and continue to the larger pores as the pressure
increase continues. The Kelvin equation (Equation 4.1) predicts that the vapor
pressure over a concave meniscus is lower than the saturation vapor pressure,
po, over a flat surface (Aveyard 1973; Gregg and Sing 1982). For cylindrical
pores:
−
=
m
L
RTrV
adso e
pp
γ2
Equation 4.1
where p/po, γ, and VL are adsorbate relative pressure, surface tension, and liquid
molar volume, respectively. Equation 4.1 is thought to give good prediction of
vapor pressures in pores down to 2 nm in diameter, which is the lower limit of
pore diameters for our MWNTs. For a cylinder open at both ends, the meniscus
is cylindrical and formed over the cylinder surface (2γrL), so that rm equals 2r,
where r is the pore radius. The capillary condensation has to be nucleated by
61
the monolayer film on the pore wall. This process is spontaneous, since the pore
radius decreases as the condensation progresses, which in turn further
decreases the vapor pressure. If the cylinder is only accessible at one end, the
meniscus is hemispherical over the closed end, rather than cylindrical over the
tube walls, and rm is now equal to r.
The course of events are slightly different for a desorption process. The
evaporation takes place from a hemispherical meniscus for both types of tube.
Compared to MWNTs, SWNTs have higher pore volume densities. Hysteresis is
generally present in systems using SWNTs as the adsorbent, but not common for
MWNTs. The SWNT diameter range is basically a Poisson distribution in the
microporous range, which means that the capillary condensation occurs in a
narrow relative pressure region for SWNTs (p/po~0.1-0.3) compared to MWNTs.
Activated carbons generally show condensed material that cannot be recovered
during desorption. Adsorption hysteresis has been observed for hydrocarbon
adsorbates on other graphitic materials; methane on graphite (Inaba and
Morrison 1986), methane on graphite foam and graphoil (Hamilton and
Goodstein 1983), and ethylene on graphite (Lysek, LaMadrid et al. 1992).
In Table 4.1 a few reports regarding hysteresis on a nanotube adsorbent are
reviewed. Mackie and coworkers (Mackie, Wolfson et al. 1997) observed no
hysteresis for the adsorption of methane on MWNTs. However, the same, but
activated MWNTs, do show hysteresis on adsorption of methane and nitrogen.
The activation process is most likely opening up the nanotube ends (the most
reactive site due to the high local curvature), hence creating cylindrical menisci.
62
Table 4.1 Hysteresis, or lack thereof, observed on MWNTs and SWNTs
during the last decade.
Adsorbent Adsorbate Hysteresis Temp.
[ K ]
Diameter
[ nm ]
Ref.
MWNT N2 None 77 4 ± 0.8 (Inoue, Ichikuni et al. 1998)
MWNT CH4 None 85 10-100 (Mackie, Wolfson et al. 1997)
MWNT CH4 None 85 1,000-2,000
(Mackie, Wolfson et al. 1997)
MWNT activated
CH4 Present 92 10-100 (Mackie, Wolfson et al. 1997)
MWNT Ar, N2 None Not given
Not given (Mackie, Lafdi et al. 1997)
MWNT activated
Ar, N2 Present Not given
Not given (Mackie, Lafdi et al. 1997)
SWNT N2 Present 77 Not given (Eswaramoorthy, Sen et al. 1999)
SWNT N2 Present 77 Not given (Yang 2000)
MWNT N2 Present 77 Not given (Yang 2000)
Copyright © Jenny Marie Hilding 2004
63
4.1.2 Stepwise Isotherms
Stepwise adsorption isotherms, or type VI isotherms, are often interpreted as
either the occurrence of phase transitions or a layer-by-layer adsorption
mechanism. The latter interpretation is the classical one (Gregg and Sing 1982).
The steps are results of the rapid decrease in interaction energy between the
surface and the adsorbate as the adsorbate builds up in different layers.
Isotherm steps are generally more pronounced at low temperatures (~75 K for Kr
and Ar) and the feature disappears as the temperature increases. The
ethylene/graphite adsorption isotherm has been found to be stepwise at low
temperatures (104.5 K; the melting point of ethylene is 104 K at atmospheric
pressure) (Meichel, Dawson et al. 1990). This adsorption system displays a
temperature-dependent transition from complete adsorbent wetting (type I
behavior) to incomplete wetting (type II behavior) for temperatures above, but
close to, the triple point (104 K).
The slope of each step is among other things a function of the homogeneity of
the adsorbent. Homogeneous materials have vertical isotherm steps. Stepwise
adsorption isotherms have been recorded for methane multilayer adsorption on
graphite near its melting point (64 < T < 105 K; the melting point of methane is 91
K at atmospheric pressure) (Hamilton and Goodstein 1983; Lysek, LaMadrid et
al. 1993).
For a heterogeneous material, each step will be replaced by a range of steps
corresponding to the occupation of various types of active site. With a sufficient
number of steps, the isotherm curve will smooth out completely and the result is
a type II isotherm. For example, increased graphitization may change
the isotherm from a type II to a type VI (Beebe and Young 1954).
64
The adsorption of Kr on closed MWNTs at 77 K has two steps while its
adsorption curve on Kr on graphite displayed five steps (Bougrine, Dupont-
Pavlovsky et al. 2002). Two-step isotherms have also been recorded for
methane and Kr adsorption on SWNTs (Muris, Dufau et al. 2000). It was
suggested that the steps represent adsorption on two different sites (the external
surface and the interstitial channel sites) but the interstitial channels are smaller
than the atomic diameter of Kr.
4.1.3 Porosity
The adsorbent porosity has a major impact on the adsorption process. Solids
with micropores (diameters <2 nm) show hysteresis for butane (Jagtoyen and
Derbyshire 1993). Mesoporous solids generally give rise to a type IV isotherm,
which is characterized by a hysteresis-loop.
Mesopores can be generated by compacting powdered solids, in which the many
particle-to-particle contacts result in hysteresis, while the same material prior to
compression show none (Jagtoyen and Derbyshire 1993). This result is most
likely due to the fact that the external surface area would decrease in relation to
the internal surface area during a compression process, which rearranges the
material from a ‘non-porous’ solid into a mesoporous solid. For example,
adsorption isotherms for methane, Kr, and Xe on exfoliated graphite and a
compressed product, Paypex, are slightly different due to increased surface
heterogeneity and formation of pores by mechanical means (Bockel, Coulomb et
al. 1982). Carbon surfaces tend to be attracted to each other and compression
of particulate carbons to generate higher bulk density solids can lead to new
pore-like structures. The molecular forces between carbon nanotubes are
influenced by both chirality and surface curvature (Kleiner and Eggert 2001;
Satoh 2001)
SWNTs have several different adsorption sites with different adsorption
potentials because SWNTs often self-assemble in small bundles. The adsorbate
can either condense in the nanotube pores, adsorb on the external wall surface,
or, in the case of SWNTs, on two additional adsorption sites between the tubes
65
(interstitial channels or external bundle grooves created by the hexagonal
packing). Small adsorbates tend to adsorb in the interstitial channels generated
between SWNTs in the bundles (Kiang, Goddard et al. 1995; Thess, Lee et al.
1996; Journet, Maser et al. 1997) and the effect is more of a capillary
condensation type than of a surface adsorption type.
Butane’s radius of gyration, rg, is 2.89 Å, while the distance between the carbon
multiwall layers is 3.35 Å. For slit shaped pores, adsorption will occur when the
diameter of the pore is up to 1.5σ, where σ is the diameter of the adsorbate
molecule. Therefore, butane is not likely to be adsorbing between the MWNT
layers.
4.1.4 Other Surface Curvature Effects
A highly bent graphite sheet, such as the wall in a carbon nanotube, has
strained double bonds, resulting in sp2/sp3 hybridization. Recent EELS-
experiments show that the curvature has to be extensive to lead to a measurable
hybridization (Stephan, Ajayan et al. 1996; Yase, Horiuchi et al. 1996; Botton,
Burnell et al. 1997; Knupfer, Pichler et al. 1999; Reed, Sarikaya et al. 2001;
Stephan, Kociak et al. 2001); thus the effect has only been seen in SWNTs, not
in MWNTs. SWNTs not only display a high surface potential inside the pore
relative to a flat graphite surface (Stan, Gatica et al. 1999), but also on the outer
wall of the tube. Orbital or potential overlaps between surfaces that are
sufficiently close to each other, such as the graphite sheets in graphite (Gregg
and Sing 1982), can result in slit condensation. Large pore volumes and strained
surface bonds give two possible explanations to why gases adsorb more quickly
on SWNTs than on planar graphite.
4.1.5 Adsorbate Order and Orientation
The adsorption of polar molecules with a tetrahedral or a pyramidal shape (CH4,
CF4, CHF3, and CHCl3 (Hammond and Mahanti 1990)) on a graphite substrate
has been modeled based on the Lennard-Jones 6-12 and Coulomb pair
potentials. The adsorbate molecular orientation shifts, relative to the adsorbent
66
surface, as a function of adsorbate polarity. Butane adsorption on graphite has
been measured near its melting temperature to study local order at the surface
(Alkhafaji and Migone 1996). A submonolayer melting was found to take place at
112.6 K (the normal melting point of butane is 135 K). Computer simulations
suggest that melting for butane monolayers is associated with slight out-of plane
motions (tilting) of the molecules that serve as a vacancy formation mechanism
(Hansen and Taub 1992). However, neutron diffraction studies show that, at low
temperatures (11 K), no molecular tilt is observed for the butane molecules on a
graphite lattice. On the other hand, when butane is adsorbed on the tightly
packed butane monolayer, the molecules tilt out-of-plane by 22o, indicating the
existence of a crystalline bilayer. In addition, a stepwise butane adsorption
growth sequence was recorded (Herwig, Newton et al. 1994).
Several studies suggest that the preferred equilibrium orientation for a
hydrocarbon physisorbed on a graphitic surface is parallel to the wall (Hoory and
Prausnitz 1967; Battezzati, Pisani et al. 1975; Andrews, Jacques et al. 1999).
Surface motion of physisorbed molecules across the surface is possible
(Battezzati, Pisani et al. 1975). Vacancy creation mechanisms for adsorbate
molecules depend on molecular shape and polarity, and the adsorbate’s ability to
wet the adsorbent surface. For example, the dominant mechanism for N2
submonolayer vacancy production is migration of molecules from the edges of
2D solid islands. As the monolayer is approaching completion, this mechanism
becomes less important and the dominating process becomes promotion of N2
molecules to the second layer (Etters, Roth et al. 1990; Etters, Kuchta et al.
1993).
The principal mechanism for vacancy creation for butane in the already adsorbed
monolayer is tilting of the molecule axis away from the basal plane of graphite
(Hansen and Taub 1992; Hansen, Newton et al. 1993). This mechanism is
though to be less affected by layer completion. This theory is proven by a
smaller increase in melting temperature for butane than for N2, a consequence of
67
the fact that less energy is required to tilt a molecule than to promote it (Alkhafaji
and Migone 1996).
Butane is organized in a rectangular unit cell with a two-sublattice herringbone
arrangement of the molecules, where one lattice vector is commensurate and the
other is not. When the angle θ approaches 1, the monolayer goes toward a
complete commensurability, exhibiting a rectangular commensurate herringbone
phase on graphite (Herwig, Newton et al. 1994; Alkhafaji and Migone 1996).
4.2 Experimental Procedure
Adsorption isotherms were determined in a Hiden instrument (IGA-002). Each
sample (~45 mg of samples 1-5) was hung in a small quartz bulb connected to a
microbalance. Adsorbed water and other gases were removed by heating the
system to 473.15 K at atmospheric pressure. At this temperature, the vapor
pressure of water in the capillary pores exceeded the saturation conditions, as
calculated from the Kelvin equation using vapor pressure, liquid density, and
surface tension from the DIPPR data set (Daubert 1989).
After degassing, the system was cooled down to 298.15 K for sorption studies.
The pressure of the system was decreased to 0.2 mbar (corresponding to a
relative pressure of 8×10-5 for butane) and the system was left until the sample
weight was stabilized and measured. High purity butane was allowed to flow
over the sample and the new weight was noted after stabilization. The
microbalance/microbalance controller of the Hiden instrument has a sensitivity
with an intrinsic accuracy equivalent to ± 0.006% of the selected weighing range
(100 mg in this case). The pressure control system can typically maintain a
pressure set-point to within 0.02% of the operating range. The thermo-regulator
has an accuracy of at least ± 0.1ºC and the intrinsic resolution of the temperature
sensor is in the range of 0.1-0.25 K.
The pressure was increased in small increments of 0.05% of the pressure span
and the sample was allowed to reach equilibrium after each step. The minimum
68
hold time was set at 5 minutes. Typical pressure changes take about 15
seconds to be achieved. Generally, it would take less than 30 seconds to reach
95% of the maximum uptake for one pressure step, which is nearly the amount of
time needed for the instrument to set the new pressure. When the mass or
pressure fluctuations are less than 0.02 % of the span, the sample is deemed to
be at equilibrium and the next step is taken. This procedure continued until the
system reached the maximum cycle pressure of 2250 mbar, which is less than
the saturated vapor pressure, po, for butane at 298.15 K (2437 mbar).
Desorption curves were obtained permitting pressure reductions to low final
pressures.
4.2.1 Adsorption Isotherms
Figure 4.1 shows the raw data of the adsorption isotherms of butane on the five
MWNT samples at 298.15 K and relative pressures from zero to ~0.9. The
adsorption isotherms were of type II, which corresponds to adsorption on a
nonporous solid, and similar to isotherms previously reported for N2 and CH4 on
carbon nanotubes (Mackie, Wolfson et al. 1997; Inoue, Ichikuni et al. 1998). The
type II isotherm was expected, since the majority of the mass-uptake occurs on
the outer surface. Only between 0.03-0.3% of the adsorbed butane condenses
in the actual pores.
The adsorption isotherms were reversible upon desorption: the weight uptake
data matched to within 1% or less for the two curves, as can be seen in Figure
4.2. The adsorption/desorption isotherms show no indication of hysteresis for
either of the samples. Figure 4.3 shows the relative pressure ranges over which
capillary adsorption should theoretically occur for our system according to the
Kelvin equation (Eq. 4.1). We investigated two scenarios; in the first case it was
assumed that only one tube end is open, in the second case both tube ends were
considered accessible. Since capillary condensation is a function of pore
diameter size, we calculated the relative pressure both for the minimum and
maximum inner diameter, 2 and 12 nm respectively. The relative pressure range
over which capillary adsorption should occur was found to be surprisingly wide;
69
0.3<p/po<0.8 for MWNTs with one end open and 0.5<p/po<0.9 for MWNTs with
both ends accessible, for a system at room temperature (298 K). The wide
relative pressure range may be one explanation to why we are not detecting any
hysteresis in our adsorption isotherms. In theory, condensation and evaporation
take place at different relative pressures for a tubes open at two ends. A
hysteresis loop in the adsorption/desorption cycle could be the result of two
accessible ends. The MWNTs of this study had only one open end, and no
bottlenecks as may be present in adsorbents such as zeolites and activated
carbon. This may be another possible explanation to why our MWNT/butane
adsorption systems lack hysteresis.
4.2.2 Adsorption Model
A type II isotherm can be successfully fitted with a modified BET model
(Brunauer, Skalny et al. 1969):
( )
−+
−
=
oo
o
m
pp
ckpp
k
pp
cknn
111 g/g Equation 4.1
The model predicts mass adsorption, n, as a function of relative pressure, p/po,
where p is the system pressure and po is the adsorbate vapor pressure at system
temperature. The DIPPR data set (Daubert 1989) may be used to estimate
physical properties as functions of temperature. The modified BET equation
uses the parameters nm, c, and k, where nm corresponds to the monolayer
capacity, and c and k to the adsorption capacity and can be found at the point of
inflection of the isotherm.
70
Relative Butane Mass Uptake
0
10
20
30
40
50
60
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Relative pressure, P/Po
Mas
s up
take
, [ m
g/g
]
Sample 1 Sample 2 Sample 3 Sample 4 Sample 5
Figure 4.1 Adsorption isotherms for samples 1-5. It is clear that sample 1 has
a much larger adsorption capacity that the other samples, thanks to its
significantly larger surface area. The modified BET equation gives excellent fits
to our system.
71
Figure 4.2 No proof of hysteresis can be observed in the adsorption-
desorption cycles. The figure presents a full cyc le for Sample 1 at 298 K as an
example.
Adsorption - Desorption Cycle for Sample 1
0
5
10
15
20
25
30
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Relative pressure, P/Po [ - ]
Mas
s up
take
[ m
ol b
utan
e/g
]
Adsorption
Desorption
72
Relative vapor pressure as a function of pore diameter
0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10 12
Pore diameter, d [ nm ]
Rel
ativ
e va
por p
ress
ure,
p/p
o [
- ]
1 end accessible
2 ends
278 K
308 K
Figure 4.3 Capillary condensation in MWNTs as a function of MWNT pore
diameter and temperature according to the Kelvin equation. The monolayer
coverage is complete at a relative pressure of approximately 0.05, so the
capillary condensation occurs after a full monolayer adsorption.
73
If k set to 1, the resulting function is the classical BET equation, which normally
gives a good fit for a relative pressure in the range of 0-0.3. (Brunauer, Emmett
et al. 1938). Several other adsorbate/carbon adsorbent data sets, C1-C7 on
activated carbon (Holland, Al-Muhtaseb et al. 2001), hydrocarbons and N2 on
activated carbon fibers (Mangun, Daley et al. 1998), and hydrogen on graphitic
herringbone-structured carbon fibers (at 77 & 300 K) (Ahn, Ye et al. 1998) have
type II isotherms that can be fitted with the BET model. The parameter c was
found to be approximately 100 for all the systems, which is expected (Beebe,
Polley et al. 1947). The modified BET model is relatively insensitive to parameter
c.
The modified BET equation is used specifically for multilayer sorption with the
number of layers less than 10 and gave excellent fits, as can be seen as full lines
in Figure 4.3. The parameters for the modified BET fits are presented in Table
4.2.
74
Table 4.2 The modified BET equation parameters fitted to the sample 1-5
butane adsorption isotherms at 298 K. The error values, presented in the same
unit as the monolayer capacity, nm, and the R2 values show that the modified
BET fits are excellent.
Sample
BET parameter
1 2 3 4 5
nm [ mg/g ] 13.5 6.91 4.57 6.71 5.18
k 0.75 0.73 0.69 0.76 0.8
c 161 132 212 100 116
Error [ mg/g ] 2×10-8 9×10-10 5×10-11 6×10-9 3×10-9
R2 0.999 1.000 1.000 1.000 1.000
Copyright © Jenny Marie Hilding 2004
75
4.2.3 Monolayer Calculations
If the adsorption of butane on MWNTs is well-described by the BET equation, we
should be able to use butane adsorption data to infer the nanotube morphology.
We have previously discussed that our MWNT inner and outer MWNT diameters
could be fitted using a two-parameter log-normal probability distribution model
given in Equation 3.1
It has been previously established that the inner diameter distribution is
independent of outside diameter. As a correlative calculation, nm was calculated
from the distribution model, using a simple geometrically based equation (see
Appendix B);
( )
( ) ∫
∫
∑
∑∞
∞
=
=
−⋅
+⋅=
−
+⋅=
0
22
0
1
22
1,
)(),),(ln(
)2(),),(ln(.
2'
dDDDDf
dDRDDfk
DD
RDkn
c
g
n
nc
n
ng
predm
σµ
σµ g/g Equation 4.2
The parameter k’ is a constant and stands for
2'gCA
but
rNM
k⋅⋅
=ρ m Equation 4.3
D and Dc are the MWNT external and core diameter respectively, Mbut and rg are
the butane molar mass and radius of gyration, NA is Avogadro’s constant, and ρc
is the density of graphite. The function, f, is given by Equation 3.1.
Table 4.3 compares the calculated monolayer coverage to corresponding
modified BET model coefficient. The predicted monolayer capacities are also
plotted versus the experimental nm values in Figure 4.4.
76
Table 4.3 Monolayer capacity from the BET fits (Exp. nm) and calculated from
morphology data (Calc. nm) together with the difference in percent. The
experimental nm is also presented in a per-area unit to show the difference in
packing density between the samples.
Sample 1 2 3 4 5
Exp. nm [ mg/g ] 13.5 6.91 4.56 6.71 5.18
Calc. nm [ mg/g ] 14.2 4.89 3.59 4.75 3.85
Difference in % 4.9 -29 -21 -34 -27
nm [ mg/m2 ] 0.28 0.41 0.37 0.48 0.39
77
Predicted BET parameter nm compared to data
0
2
4
6
8
10
12
14
16
0 2 4 6 8 10 12 14 16
Experimental nm [ mg/g ]
Pre
dict
ed n
m
[ mg/
g ]
y=x
Sample 1
Sample 2
Sample 3
Sample 4
Sample 5
Figure 4.4 The predicted nm values for the four samples with larger diameters
are about 30% lower than the experimental value for nm.
78
Clearly, these independently determined values agree fairly well for samples 1-5,
with a difference of about 30% for samples 2-5 and as low as 5% for sample 1.
To compare the samples’ monolayer capacity further, the measured nm was
converted to a per area basis as can also be seen in Table 4.3. It is clear that
the packing density is fairly close between the larger samples. However, the
adsorption density is lower for sample 1 compared to the other samples. This
result is in agreement with what is predicted from the Kelvin equation.
4.2.4 Adsorbate Packing Density Calculations
We calculated the butane packing radius by iterating Equation 4.1 to fit the
measured monolayer capacity, nm. The procedure was carried out for all five
samples and the result is plotted in Figure 4.5 . The dotted line in Figure 4.5 is
the butane radius of gyration, 2.89 Å.
The packing radius ranges between ~2.4-2.6 Å for the larger tubes. For the
smallest MWNTs, the butane radius is closer to 3 Å. If we assume a square
packing, the surface areas are ~23-27 Å2/butane molecule for the larger tubes.
The sample with the smallest diameter is occupied by butane molecules each
taking up an area of 36 Å2. The latter number is very close to values already
reported in literature (~30-40 Å2), as shown in Table 4.4.
From nm it is possible to determine the area occupied by each butane molecule
on the MWNT external surface. Assuming a square butane packing lattice, the
area per molecule is calculated to lie between 26 and 32 Å2. For five of the six
samples analyzed, the estimated areas were found to be smaller than values
previously reported in the literature. These five samples had larger average
diameters. If the calculated areas are plotted vs. corresponding nm, a minimum
seems to occur for the two samples with average diameter of 48.5 and 48.7 nm
respectively.
79
Butane packing radius vs. monolayer capacity
2.0
2.2
2.4
2.6
2.8
3.0
0.0 2.5 5.0 7.5 10.0 12.5 15.0Monolayer capacity nm [ mg/g ]
But
ane
pack
ing
radi
us, r
[ Å
]
rg
Sample 1
Sample 2
Sample 3
Sample 4
Sample 5
Figure 4.5 The estimated butane packing radius, calculated by iterating nm.
The dotted line across the graph indicates the standard butane radius of gyration,
which is 2.89 Å. The curve going through the line is just added to lead the eye.
80
Table 4.4 Estimated adsorption areas per molecular for a full butane
monolayer adsorbed on different carbon surfaces.
Adsorbate Adsorbent Area/molecule
[ Å2 ] Reference
Carbon black Butane, 273 K 32-32.2 (Beebe, Polley et
al. 1947)
Graphite Butane 30.1 (experimental)
38.3 (calculated)
(Hoory and
Prausnitz 1967)
Graphite Butane 32.7 (Alkhafaji and
Migone 1996)
MWNT Butane 23-36 Here
81
If we assume that butane is 7.01 Å long and 3.06 Å wide, the packing would have
to be very tight to enable a horizontal butane orientation relative to the MWNT
surface. It is probably more likely that the molecules are slightly tilted from the
surface, but more detailed studies are necessary to clarify this issue. The
volume for a solid butane phase is about 120-125 Å3 (DIPPR), which is a 5
Å/side cube, or a 6 Å-diameter sphere, so the estimated areas are not
unreasonable.
4.2.5 Multilayer Analysis
Various studies on adsorption of rod-like molecules (such as butane or nitrogen)
on flat graphite surfaces (Herwig, Newton et al. 1994; Herwig, Wu et al. 1997)
show that the effective packing of the monolayer and the multilayers differ. The
monolayer has a highly ordered structure, while the multilayers are more loosely
packed. Longer CH-chains (C>3) can switch between gauche- and trans-
conformation while adsorbed on graphite, thereby changing the effective packing
as well. Also, butane molecules might be tilted relative to the graphite surface,
adsorbing primarily with only one end/carbon attached to the surface. These
three mechanisms could result in different multilayer densities with different
MWNT radii.
The result from a recent computer simulation has also suggested that an
increase in the grade of conformational deformation of a surface also increases
the local reactivity, due to local deformations of the reactive sites (Weedon,
Haddon et al. 1999). The experiment simulates chemisorption of hydrogen
atoms onto a bent and strained SWNT. Assuming that one type of surface
deformation is a highly graded curvature, one would expect tubes with increasing
curvature to show a higher uptake of butane. Hydrogen (Srivastava, Brenner et
al. 1999) is computed to chemisorb preferentially onto folded surfaces of
deformed nanotubes. Therefore, surfaces with more curvature could exhibit
increased attraction energies for adsorbed molecules, resulting in higher surface
loadings and possibly different local packing densities.
82
4.3 Conclusions
We have established that butane adsorption at room temperature on CVD-
produced MWNTs gives rise to an adsorption isotherm of type II despite the
slight negative curvature of the MWNT surface. Type II adsorption is defined as
physisorption on a nonporous solid and can be fitted with a modified BET model.
The experiments are reproducible and all the butane is readily desorbed. No
hysteresis was observed, which would be of great advantage in a process
involving adsorption of valuable or toxic compounds, where activated carbons
would irreversibly adsorb a percentage of the condensed matter.
The monolayer capacity is greater for samples with smaller diameters, but the
monolayer packing density was lowest for the sample with the smallest average
tube diameter. The adsorption capacity was also larger for the samples with
smaller diameters.
Pore condensation accounts for less than 0.5% of the adsorption capacity for all
the samples, a finding consistent with the relatively small inner pore volume. In
addition to its low relative mass, capillary condensation may not be detectable
since, according to the Kelvin equation, the pores are expected to be filled at
wide relative pressure range; between 0.3 and 0.9 depending on tube diameter
and accessibility. The remainder of the butane adsorption is assumed to be
surface condensation, which leads us to conclude that multilayer adsorption
occurs. It is clear that the sample diameter distribution has a major impact on the
adsorption capacity.
The BET monolayer coefficients were fairly consistent with those calculated
directly from the adsorbate properties and the MWNT morphology, which leads
us to conclude that a number as small as one hundred MWNTs, is sufficient to
analyze a sample accurately. BET isotherms and the model suggest that butane
molecules are probably tilted away from the surface perpendicular. This result
leads us to conclude that pore volume and surface area distributions can be used
in combination with the Kelvin and BET equations to predict the magnitudes of
pore condensation and surface condensation mechanisms.
Copyright © Jenny Marie Hilding 2004
83
Chapter Five
Isosteric Heat of Adsorption
5.1 Introduction
In this chapter we report the isosteric heats of adsorption, qst, for butane on
MWNT samples with different morphology. In the sections below we introduce
and describe a number of different adsorption phenomena, theories, and reviews
relevant to our experiments and useful in the evaluations of our data. The qst
data generated by us and presented here has been calculated over a range of
surface loadings at temperatures below the normal boiling point and analyzed
form a morphological point of view. Isosteric heat of adsorption measurements
on carbon nanotubes can provide the basis for interpreting adsorption
mechanisms and the ability to learn more about MWNTs as an adsorbent.
5.1.1 Isosteric Heat of Adsorption
The isosteric heat of adsorption, qst, or the negative differential molar enthalpy of
adsorption -∆Hads, is the amount of energy released when a molecule adsorbs
onto a surface. The isosteric heat of adsorption in the limit of low coverage is
due only to adsorbate-surface interaction, hence closely related to the binding
energy between adsorbate and adsorbent:
TkHq BBadsst 2+=∆−= ε eV Equation 5.1
where εB is the binding energy (generally given in eV), kB is Boltzmann’s
constant, and T is the temperature. Experimental adsorption of hydrogen,
oxygen, nitrogen, methane, butane, and the noble gases on structured carbons is
collected in Table 5.1.a-c. Any binding energy data has been recalculated to
isosteric heat of adsorption to enable quick and easy comparison. The data in
Table will be discussed through the sections following below.
84
Tables 5.1.a-c present isosteric heats of adsorption for different adsorbate
molecules on CNTs compared to other forms of carbon at various temperatures
and surface coverage. CC stands for methods involving the application of the
Clausius-Clapeyron equation on a set of isotherms. Please observe that many of
the qst values are not exact values, since they are read from graphs. The letters
a-f in the tables correspond to the footnotes listed below. The list gives a more
detailed description of some of the adsorbents:
a. Carbon black with a model pore diameter of 28.5 nm
b. Graphitized Spheron. No large change in pore diameter (30 nm), but a reduction in
surface area from 110 to 89.4 m2/g
c. Carbon black with pore diameter 60 nm
d. Carbon black with pore diameter 50 nm
e. Carbon black with pore diameter 22.5 nm
f. Carbon black with pore diameter 56 nm
Oscillations in qst can suggest a system displaying stepwise multilayer adsorption
while peaks in the qst curve may indicate co-existing phases on the surface. The
stepwise isotherm is theoretically predicted to occur on homogeneous
adsorbents and will be discussed more in a special section.
5.1.2 The Clausius-Clapeyron Equation
Commonly qst is generated through the Clausius-Clapeyron equation by applying
the equation to a set of adsorption isotherms. The equation is based on the
assumption that the adsorbate is an ideal gas and sorbs with negligible volume.
It estimates qst as a function of pressure and temperature at a fixed mass uptake.
anadsst T
pRTHq
∂∂=∆−= ln2
J/mol Equation 5.2
Ln(p) is simply plotted vs. T and lines are fitted for several different mass uptakes
to generate an adequate number of data points. The resulting slopes are
multiplied by RT2 to get qst. A break in the slope of a linear fit indicates that a
phase transition is occurring. One example is the phase transition for methane
adsorbed on SWNTs at 88±2 K (Muris, Dufau et al. 2000).
85
Table 5.1.a Isosteric heats of adsorption for hydrogen, oxygen and nitrogen on
different types of carbon.
Adsor bate
Adsorbent
Initial qst [kJ/mol]
qst
[kJ/mol] (var. θ)
qcond [kJ/mol] (Daubert
1989)
Temp [K]
Ref
H2 SWNT - 11.6±0.6
θ not given - 77-87
(Pradhan, Sumanasekera
et al. 2002)
H2 Graphon 3.8 1.3 θ~1
0.9 20 (Pace and
Siebert 1959)
H2 Graphitized
carbon black 5.41
(5.59 for D2) - - 90-138
(Constabaris, Sams et al.
1961)
N2 SWNT 19.6 - - 300-343 (Wei, Hsu et al. 2003)
N2 Close ended
SWNT 9.0 3.5 θ=1
5.43-4.84 81-94 (Yoo, Rue et al.
2002)
N2 Open ended
SWNTs 18.0
9.0 θ=1
3.03- 117-130 (Yoo, Rue et al.
2003)
N2
Carbon Black: Spherona
Graphonb Sterling Sc
Sterling Ld
18.1 12.1 16.7 16.7
θ=1 12.5 12.1 12.0 11.4
5.54
78
(Beebe, Biscoe et al. 1947)
N2 Grafoil 15.0 7.6 θ=1
5.50 79.3 (Piper, Morrison
et al. 1983)
N2
Carbon Black: Spherona
Graphonb
Sterling FT-Ge
Sterling MT-Gf
-
θ>0 16.7 11.3 10.9 10.0
-
242-301 142-222 142-178 142-189
(Gale and Beebe 1964)
O2 SWNT 20.3 - - 300-343 (Wei, Hsu et al. 2003)
O2
Carbon Black: Spherona
Graphonb
Sterling FT-Ge
Sterling MT-Gf
-
θ>0 16.7 11.3 10.9 10.5
-
3.78— 3.78— 3.78--
242-301 142-222 142-178 142-189
(Gale and Beebe 1964)
O2 Spherona 16.7 8.1 θ=1
7.15 78 (Beebe, Biscoe
et al. 1947)
a. Carbon black with a model pore diameter of 28.5 nm
b. Graphitized Spheron. No large change in pore diameter (30 nm),
but a reduction in surface area from 110 to 89.4 m2/g
c. Carbon black with pore diameter 60 nm
d. Carbon black with pore diameter 50 nm
e. Carbon black with pore diameter 22.5 nm
f. Carbon black with pore diameter 56 nm
Copyright © Jenny Marie Hilding 2004
86
Table 5.1.b Isosteric heats of adsorption for various CH gases on different
types of carbon.
Adsor bate
Adsorbent
Initial qst [kJ/mol]
qst
[kJ/mol] (var. θ)
qcond [kJ/mol] (Daubert
1989)
Temp [K]
Ref
CH4 Closed SWNTs 28.9
9.65 θ=0.7 8.72 ~90
(Talapatra and Migone 2002)
CH4 Closed SWNTs 24.4 - 5.96-- 159.88-
194.68 (Talapatra, Zambano et
al. 2000)
CH4 Graphitized
carbon 13.3 - - 253-348 (Constabaris, Sams et al.
1962)
CH4 Graphitized
carbon
12.69 (12.57 for
CD4) - - 224-297
(Constabaris, Sams et al.
1961)
CH4 Graphite 15.1 - - - (Vidali, Ihm et al. 1991)
CH4 Grafoil 17.1 19.0 θ=1 - 84.5
(Piper and Morrison
1984)
C2H4 MWNT - 11.6±0.8
θ=0.5 - 77.5-81.4 (Masenelli-
Varlot, McRae et al. 2002)
C2H4 Graphite - 16.5 - - (Masenelli-
Varlot, McRae et al. 2002)
C2H4 Graphite 22.7-24.8 16.7-18.8
θ=1 --15.3 98-120 (Inaba and Morrison
1986)
C4H10 Graphite 34.0±0.1 - 19.8-6.84 318-418 (Chirnside and Pope 1964)
C4H10 Graphite 33.9 - 18.1-- 341-433 (Ross,
Saelens et al. 1962)
C4H10 Graphite - 36.4, 34.8
θ=0.5 - - (Kiselev 1961)
C4H10 Graphitized
carbon
31.2 35.6 θ=0.5 23.4-17.5 253-348
(Hoory and Prausnitz
1967)
C4H10 Graphitized
carbon fibers
14-22 (Mangun,
Daley et al. 1998)
Copyright © Jenny Marie Hilding 2004
87
Table 5.1.c Isosteric heats of adsorption for noble gases on different types of
carbon. Copyright © Jenny Marie Hilding 2004
Adsor- bate
Adsorbent
Initial qst
[kJ/mol]
qst
[kJ/mol] (var. θ)
qcond [kJ/mol] (Daubert
1989)
Temp [K]
Ref
He SWNT - 1.9 - 14-16 (Teizer, Hallock et al. 1999; Teizer, Hallock et al.
2000) He Graphoil 1.50
1.45 0.46 0.29
0.048-- 5 12
(Elgin and Goodstein
1974)
Ne Closed SWNTs 5.81 - 1.23-- 37.66-57.61
(Talapatra, Zambano et
al. 2000)
Ne Graphite 3.69 - - - (Vidali, Ihm et al. 1991)
Ar SWNT 17.6 θ=1 5.2
4.96-4.11
120.5-133 (Yoo, Rue et al. 2002)
Ar SWNT 19.0 - - 300-343 (Wei, Hsu et al. 2003)
Ar Carbon Black: Spherona
Graphonb
Sterling FT-Ge
Sterling MT-Gf
-
θ>0 16.3 11.3 10.9 10.5
-
3.21— 3.21— 3.21--
242-301 142-222 142-178 142-189
(Gale and Beebe 1964)
Ar Carbon Black: Spherona
Graphonb
16.7 10.9
θ=1 10.0 12.1
- 78 (Beebe, Biscoe et al. 1947; Beebe, Millard et al.
1953)
Ar Graphonb 1273 K 1773 K 2273 K 2973 K
12.8 11.9 10.7 10.7
θ=1 7.7 7.7 7.7 7.7
- 78 (Beebe and Young 1954)
Ar Graphitized carbon black:
9.84 8.79 θ=1
6.33-3.76
90-137 (Sams, Constabaris et
al. 1962)
Kr MWNTs - 11.6 θ=1
- 77.5-83.7 (Bougrine, Varlot et al.
2001)
Kr Graphitized carbon black
12.61 15.0 θ=1
- 95-105 (Putnam and Fort 1975)
Xe SWNT 31.6 - 10.98-- 210-295 (Zambano, Talapatra et
al. 2001)
Xe Closed SWNTs 31.5 - 10.46-- 220-295 (Talapatra, Zambano et
al. 2000)
Xe Graphite 19.9 - - - (Vidali, Ihm et al. 1991)
88
5.1.3 Homogeneous and Heterogeneous Adsorbents
In reality, there are no perfectly homogeneous adsorbents, but graphitized
carbon black consists mainly of (001) planes and a low percentage of sites with
higher energy. For a homogeneous surface, qst generally increases steadily up
to a maximum right before the monolayer completion, due to adsorbate-
adsorbate interactions. Heterogeneous sites give qst curves that are high at low
fractional coverage and then quickly drop as the more highly energetic active
sites are occupied. This behavior has been observed for ethylene (Meichel,
Dawson et al. 1990), methane, Xe, and N2 adsorption on graphite (Piper and
Morrison 1984). For example, the ethylene isosteric heat of adsorption on
graphite (98 < T < 120 K) is 20.4 kJ/mol, as can be seen in Table 5.1 (Inaba and
Morrison 1986). The qst curve displays a sharp submonolayer minimum, likely
caused by the energetic heterogeneity of the adsorbate surface. After the drop,
qst increases as a result of the interactions between the adsorbate molecules.
When the first layer is complete, qst drops dramatically again.
The isosteric heat of adsorption for the Kr-MWNT system was measured to be
11.6 kJ/mol at θ=1, which is lower than 15.0 kJ/mol for the Kr-graphite system
(Putnam and Fort 1975; Bougrine, Varlot et al. 2001). Moreover, the monolayer
condensation pressures over the MWNT surface are higher than those observed
with graphite, which is consistent with what is predicted by the Kelvin equation.
The higher condensation pressure probably leads to an incomplete wetting of the
surface by limiting the numbers of Kr monolayers adsorbed before its bulk
condensation (Bougrine, Varlot et al. 2001).
For a heterogeneous material, each step will be replaced by a range of steps
corresponding to the occupation of various types of active site. With a sufficient
number of steps, the isotherm curve will smooth out completely and the result is
a type II isotherm. As mentioned earlier, increased graphitization may change
the isotherm from a type II to a type VI (Beebe and Young 1954). Another
observed effect is that the initial heat of adsorption decreased and the drop in qst
at low surface coverage disappeared, also an effect of increased homogeneity.
89
The nature of the adsorbate may also affect the shape of the steps. For
example, the steps are more well defined for Kr than for Ar; a result of krypton’s
higher polarizability (αKr = 2.48×10-24 cm3, αAr = 1.63×10-24 cm3) (Gregg and Sing
1982).
5.1.4 Adsorbent Porosity
The adsorbent porosity has a major impact on the adsorption process. For
example, adsorption isotherms for methane, Kr, and Xe on exfoliated graphite
and a compressed product, Paypex, are slightly different due to increased
surface heterogeneity and formation of pores by mechanical means (Bockel,
Coulomb et al. 1982). Carbon surfaces tend to be attracted to each other and
compression of particulate carbons to generate higher bulk density solids can
lead to new pore-like structures. The molecular forces between carbon
nanotubes are influenced by both chirality and surface curvature (Hilding, Grulke
et al. 2001; Kleiner and Eggert 2001; Satoh 2001). As mentioned in a previous
chapter, SWNTs have several different adsorption sites with different adsorption
potentials because SWNTs often self-assemble in small bundles. The adsorbate
can either condense in the nanotube pores, adsorb on the external wall surface,
or, in the case of SWNTs, on two additional adsorption sites in between the tubes
(interstitial channels or external bundle grooves created by the hexagonal
packing). Small adsorbates tends to adsorb in the interstitial channels generated
between SWNTs in the bundles (Kiang, Goddard et al. 1995; Thess, Lee et al.
1996; Journet, Maser et al. 1997) and the effect is more of a capillary
condensation type than a surface adsorption type.
qst does not differ much between external and internal simulated adsorption of Xe
on isolated SWNTs (Kuznetsova, Yates et al. 2000; Simonyan, Johnson et al.
2001). These results are consistent with experimental data presented in Table
5.1 (Vidali, Ihm et al. 1991; Talapatra, Zambano et al. 2000; Zambano, Talapatra
et al. 2001). The same lack of difference in qst has been recorded for methane
adsorbed on open-ended and close-ended SWNTs (Stan, Gatica et al. 1999;
Kuznetsova, Yates et al. 2000; Stephan, Kociak et al. 2001). However, N2
90
adsorbed on open-ended SWNTs gives an initial qst double that for the N2/close-
ended SWNT system (Beebe, Biscoe et al. 1947; Yoo, Rue et al. 2002; Yoo, Rue
et al. 2003). The result suggests that N2 is small enough to actually adsorb into
the SWNTs, while methane and Xe are not. As mentioned in Chapter 4, one
molecular dynamics study suggests that SWNTs can expand to hold xylene, a
relatively large molecule. By contrast, MWNTs that are open at one end have
only two types of surfaces for adsorption, the inner wall of the pore and the
exterior cylindrical surface.
5.1.5 Adsorption on a Highly Bent Surface
As discussed earlier (section 4.2), a highly bent graphite sheet may contain
strained double bonds, resulting in sp2/sp3 hybridization. This may provide an
explanation to why methane adsorption on close-ended SWNTs displays a
binding energy, eB, 76% larger than that of methane adsorbed on planar graphite
(Weber, Talapatra et al. 2000; Weber, Talapatra et al. 2000). The adsorption
capacity has also been found to be higher, possibly due to a larger number of
carbon atoms neighboring each adsorbed molecule in the interstitial channels as
compared to adsorption on a flat graphite surface. Another investigation shows
that εB is ~60 % larger for the He/SWNT system compared to the He/graphite
system (Teizer, Hallock et al. 1999). The same type of result has also been
found for H2 or Ne in SWNTs, where the binding energy was calculated to be
~150% larger than for H2 or Ne on graphite (Stan and Cole 1998; Stan and Cole
1998; Stan, Crespi et al. 1998; Stan, Gatica et al. 1999).
As can be seen in Table 5.1, qst for gas adsorption on MWNTs is consistently
lower than qst on graphite, confirming that the surface strain is not sufficient to
create orbital hybridization and that the pore density is too low to have an impact
on the isosteric heat of adsorption.
91
5.2 Experimental Section
The MWNT samples used in the experiments were samples 2, 4 and 5 presented
in Chapters 3 and 4. The adsorption experiments were carried out in the same
fashion as described in detail in Chapter 4.3.
After each completed pressure cycle, the system temperature was increased 5 or
10 K and the system was allowed to reach equilibrium before measuring a new
isotherm. This procedure was repeated 3-4 times for each sample, giving
between four and five different isotherms. The temperatures were kept in a
range of 278.15-308.15 K.
5.2.1 Adsorption Isotherms
The adsorption isotherms were of type II, which is consistent with our previous
experiments. Adsorption should be highest for the solids with the highest
external surface area (a direct result of the sample’s average diameter). Since
sample 2 has the largest adsorption capacity, it is expected to possess the
highest surface area per unit weight. Sample 4 has a slightly larger average
diameter, closely followed by sample 5.
Consistent with previous experiments, no hysteresis was detected and the
experiments were reproducible. As an example, the full adsorption-desorption
cycles for sample 5 at five different temperatures can be seen in Figure 5.1.
The modified BET fits were excellent as shown for sample 4 in Figure 5.2. Model
coefficients were fitted using Maple. The parameters for each experiment are
presented in Table 5.2, together with the errors and R2 values. The monolayer
capacity, nm, seems to increase slightly with system temperature, while k
decreases with T. The parameter c seems to be changing inconsistently with T,
but is found to be close to 100 as expected. The modified BET fit for each
isotherm is included in the graph as a full line. The data collapse to one curve
when the isotherms are plotted versus relative pressure.
92
5.2.2 Isosteric Heat of Adsorption
The calculated isosteric heats of adsorption of the three samples are shown in
Figure 5.3 as a function of monolayer coverage (θ). The calculations were based
on the Clausius-Clapeyron equation, Equation 5.2. Approximately 30 data points
were generated for each sample and R2 for the linear fits were generally >0.95.
The data points cover the entire measured adsorption range and show that the
isosteric heat of adsorption is high at very low coverage, decreases steadily and
displays a minimum when the monolayer is complete (θ=1). Finally qst
approaches the heat of condensation for butane (Daubert 1989) when θ~2.5.
The fact that the isosteric heat of adsorption is highest initially and starts to
decrease immediately, suggests that the MWNT samples are heterogeneous
adsorbents. The samples are numbered in order of their average diameters, so
qst at low surface coverage does not correlate to nanotube morphology.
93
Adsorption-Desorption Isotherms
0
5
10
15
20
25
0.E+00 5.E+04 1.E+05 2.E+05 2.E+05 3.E+05
Pressure [ Pa ]
Mas
supt
ake
[ m
g/g
]
278.15 K293.15 K298.15 K303.15 K308.15 K
Sample 5
Figure 5.1 Adsorption-desorption isotherms for sample 5 at five different
temperatures. No hysteresis is observed.
94
Adsorption Isotherms with Modified BET fits
0
5
10
15
20
25
30
0.0E+00 5.0E+04 1.0E+05 1.5E+05 2.0E+05 2.5E+05
Pressure, p [ Pa ]
Mas
supt
ake,
n
[ mg/
g ]
278.15K288.15 K298.15 K308.15 KM. BET model
Sample 4
Figure 5.2 A set of adsorption isotherms for sample 4, at four different
temperatures between 278 and 308 K. The modified BET fit for each isotherm is
included in the graph as a full line. The modified BET coefficients are presented
in Table 6.2 below, together with the error and R2 values of the fits.
95
Table 5.2 The modified BET fitting parameters at different temperatures for
samples 2, 4 and 5. As shown in Chapter 5, the modified BET fits are excellent.
The errors are given in the same unit as the monolayer capacity and the R2
values are ≥0.99.
Temperature [ K ] Sample Parameter
278 283 288 293 298 308
nm
[ mg/g ] 6.003 - 6.1 - 6.3 6.6
k 0.78 - 0.80 - 0.75 0.76
c 117 - 128 - 92 94
Error
[ mg/g ] 9×10-9 - 6×10-9 - 6×10-9 1×10-9
2
R2 1.000 1.000 1.000 1.000
nm
[ mg/g ] 6.71 - 6.64 - 7.20 7.32
k 0.78 - 0.78 - 0.76 0.72
c 104 - 109 - 100 89
Error
[ mg/g ] 2×10-9 - 8×10-9 - 7×10-10 3×10-10
4
R2 1.000 0.999 1.000 1.000
nm
[ mg/g ] 5.18 5.05 5.10 5.35 5.29 -
k 0.82 0.82 0.82 0.81 0.80 -
c 134 100 105 128 116 -
Error
[ mg/g ] 8×10-8 2×10-9 2×10-9 4×10-9 3×10-9 -
5
R2 0.996 1.000 1.000 0.999 1.000
96
Prior work on butane adsorption on graphitized carbon fibers showed that the
isosteric heat of adsorption increased with decreasing fiber diameter, however,
these fibers were porous (Mangun, Daley et al. 1998).
It is likely that the initial heat of adsorption depends more on surface structure
than on the nanotube diameter; the defects on the MWNTs are most likely
dominating the initial stages of the adsorption. Several studies suggest that
defect sites are of great importance to the adsorption process. The impact can
be seen not only in systems where nanotubes (Kim, No et al. 2001; Lee, Kang et
al. 2002; Mann and Halls 2002; Zhang, Lin et al. 2002; Zhou and Shi 2003) or
other carbon materials (Sattler 1992; Sibrell and Miller 1992) are used as the
adsorbate, but also for metallic crystalline materials (Au, Hirsch et al. 1988;
Jakob, Gsell et al. 2001). It is not likely that the defects are reactive to butane,
as chemisorption might cause hysteresis or irreversible adsorption. Figure 5.4
shows a few different types of defects found in our samples. These defects may
affect the isosteric heat of adsorption for our system.
At low surface coverage, qst is higher than the heat of condensation. Lennard-
Jones 6-12 potential calculations show that the attractive force between butane
molecules is much stronger than it is between butane and carbon. Also, butane
incompletely wets graphite (Herwig, Newton et al. 1994). These results suggest
that the structure in the monolayer region is different from the structure in the
multilayer region. The result is consistent with previous work with butane
adsorption near it melting point; multilayer films formed at θ ~2.65 on graphite
foam (Alkhafaji and Migone 1996) and θ > 2.0 on Paypex (Herwig, Newton et al.
1994). The first two layers were found to be highly ordered, crystalline phases.
Bilayer crystallization accompanied by some molecular reorientation is believed
to occur in the N2-graphite system at low temperatures also (Wang, Newton et al.
1989). At much higher temperatures, the first two layers may not be as highly
ordered. For θ ~2.5, the isosteric heats of adsorption are essentially the same for
the three samples (Figure 5.3), and are similar to the heat of condensation of
butane.
97
Heat of adsorption vs. relative uptake
0
5
10
15
20
25
30
35
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Massuptake, n/nm [ - ]
Isos
teric
Hea
t of A
dsor
ptio
n, q
st
[ kJ/
mol
]
Sample 4
Sample 2
Sample 5
qcond for butane at 278-308 K
Figure 5.3 The isosteric heat of adsorption for sample 2, 4, and 5 is plotted vs.
the relative mass uptake. The shaded field running across the graph represents
the heat of condensation of butane between 278 and 308 K.
98
Figure 5.4 An example of different MWNT structural defects that may be
present in a MWNT sample. Counter-clockwise from the top: a broken and
exfoliated layer, a strongly bent MWNT, and a torn MWNT.
99
This is a likely explanation for the increase in the heat of adsorption after the
monolayer is complete, since in the upper layers a butane molecule will be
surrounded only by butane.
In Table 5.1, the MWNT-butane system is compared with other types of carbon
adsorbent-adsorbate systems. The qst at multilayer coverage for the graphite-
adsorbate system is generally higher than for the MWNT-adsorbate system.
According to the Kelvin equation, the vapor pressure over a convex surface
should be higher than it would be over a flat surface, which is a possible
explanation to the lower isosteric heats of adsorption. Surface impurities in the
form of amorphous carbon have been detected in TEM photos; this type of
impurity will also decrease the heat of adsorption.
5.3 Summary, Discussion, and Conclusions
The isosteric heat of adsorption, qst, was calculated from each set of isotherms
by applying the Clausius-Clapeyron equation via the modified BET model. The
BET fits and the Clausius-Clapeyron linear fits generally displayed R2 values
>0.999 and >0.95 respectively.
The magnitude of qst confirms that the adsorption mechanism is physisorption.
Adsorption of various alkanes on carbon black systems (Beebe, Polley et al.
1947) has been proven to display initial heats of adsorption approximately 3
times larger than their isosteric heat of condensation. In our case, the initial heat
of adsorption is maximum 1.3 times the heat of condensation. The heats of
adsorption are highest at very low levels of surface coverage and decrease
steadily until a minima is reached at θ=1. These results indicate that the MWNT
samples act as heterogeneous adsorbents.
For θ>2.5, the isosteric heat of adsorption is approaching the butane heat of
condensation. This is consistent with previous investigations near the melting
point of butane, which reported a bilayer crystal and a bulk phase build -up in the
region of θ=2-2.65. The initial heats of adsorption were found to lie between 22-
100
26 kJ/mol, which is around 2/3 of reported qst values for butane adsorption on
graphite. Two phenomena could contribute to this effect: 1) MWNT surface
impurities, such as amorphous carbon, which can lower the surface potential,
and 2) the concave meniscus on the external surface that would lead to a lower
surface potential. It seems likely that the butane packing differs between each
adsorbed layer on the MWNT wall.
Adsorption on MWNT is less energetically favorable compared to adsorption on
flat graphite. On the other hand, SWNT surfaces seem to be more attractive
than graphite, but there is lack of agreement between groups as to why. Our
results demonstrate that the key to understanding adsorption on nanotubes is to
carefully research the adsorbent to determine surface area, pore volume, defect
fraction, and impurity fraction. These properties have a major impact on the
adsorbent’s interaction with the adsorbate.
The isosteric heat of adsorption for butane adsorption on MWNTs can be related
to their morphologies. The higher heats of adsorption at low coverage are
probably related to the presence of surface defects. The minima in qst, near the
full monolayer capacity, relate to the high self-association of butane. The
adsorption of other gases (the noble gases from He to Xe, hydrogen, oxygen,
nitrogen, methane, and ethylene) on other ordered carbons (different forms of
graphite and single-walled nanotubes) show hysteresis, multiple step adsorption,
and sites having different heats of adsorption. This investigation, in combination
with reported data from other sources, shows the importance of a careful
investigation and interpretation of adsorption as a function of morphology.
From the findings presented in Chapters 4 and 5 we conclude that MWNTs are
not interesting as a gas storage material due to the low relative pore volume. It is
possible to use MWNTs as a separation tool, but the adsorption capacity is low
compared to many other carbon based adsorption systems. The advantage is
the lack of hysteresis, which makes MWNTs interesting for adsorption processes
specifically involving toxic or valuable gases.
Copyright © Jenny Marie Hilding 2004
101
Chapter Six
Dispersion of Carbon Nanotubes
6.1 Introduction
Carbon nanotubes are examples of nanoparticles with very high aspect ratios,
unusual mechanical properties, and intriguing transport properties. As such, they
constitute a useful model system for evaluating the potential of nanoparticle
additives to liquids and polymeric solids for achieving significant improvements in
bulk properties at low volume loadings.
Production processes for carbon nanotubes often produce mixtures of solid
morphologies that are mechanically entangled or that self-associate into
aggregates. Entangled or aggregated nanoparticles often need to be dispersed
into fluid suspensions in order to develop materials that have unique mechanical
characteristics or transport properties. Important challenges to developing
applications for these unique materials include: 1) purification and separation of
nanotubes by chemistry and morphology, 2) uniform and reproducible dispersion,
and 3) orientation of these solids in liquid and melt phases.
Many of the conventional dispersion methods cause fragmentation of the
nanotubes, which can be modeled using the moments of the length distribution.
As with other solids, the breakage rate of carbon nanotubes depends on their
lengths, with the longest particles experiencing the highest breakage rate.
This chapter focuses primarily on the challenge of developing reproducible
dispersions of carbon nanotubes in liquid phases. Two phenomena affect carbon
nanotube dispersions: nanotube morphology and attractive forces between the
tubes. One factor, which directly affects the dispersion homogeneity, is the
presence of agglomerates, which in turn is a function of MWNT length and
structure. We present here the effects of ultrasonication bath, ultrasonication
with wand, elongational flow, and grinding on the morphology of MWNTs.
102
6.2 Modeling Typical Particle Fragmentation Distributions
Many mathematical functions have been used to model the particle size
distributions of comminution processes. These functions include empirical
models as well as probability density functions. As particles are fragmented and
broken during processing, their length distributions change as do the model
coefficients that describe the distributions. Probability density functions are
particularly useful if they provide a good fit to particle size distributions since the
moments of the distributions can be used in kinetic models that predict the
change in the length distributions with time. These models have been used to
interpret the reduction of chain lengths during processes such as catalytic or
thermally induced polymer degradation and polymer ultrasonication. Gel
permeation chromatography of polymers yields complete differential distributions
that provide rich information on the fragmentation process. A key assumption of
these models is that one fragmentation event occurs in the chain (or particle or
nanotube) at a time, an assumption that would be met by most mechanical
processes of nanotubes.
The rates of particle fragmentation for most minerals, ceramics, metals and
polymers generally decrease as the characteristic particle size decreases. As
the materials fragment to smaller sizes, less of the applied energy is transformed
into fragmentation and more in particle motion, particle compression, particle
flexing and other mechanisms. A simple kinetic model based on binary
fragmentation that can describe these phenomena is (McCoy 1994; Wang 1995;
Rangarajan, Bhattacharyya et al. 1998):
bb Lk
dtdL
⋅=− Equation 6.1
where L is a characteristic material length, t is time, k is a rate constant, and b is
the exponent that describes the change in fragmentation rate with length. When
b=1, Equation 6.1 describes a first order fracture process that does not depend
103
on chain length. We anticipate that nanotube fragmentation will depend on
length. The previous references demonstrate how continuous distributions can
be analyzed to determine the coefficients, k and b, and the new product
distributions at any time, t. This elegant approach is beyond the scope of this
section as it assumes that each nanotube will be exposed to similar processing
conditions of energy per unit volume, simple shear, elongational flow or
mechanical force. Except for suspensions with low concentrations of nanotubes,
these conditions may not be met for many of the methods. Some of the practical
problems of assuring uniformity of dispersion forces throughout the fluid are
discussed for each method. As an approximation, we have analyzed the
changes in the average length as a function of time for several different
fragmentation methods and reported apparent coefficients. More rigorous
analyses are available if distribution of forces is well known within the processing
volume.
6.3 Ultrasonication Methods
Ultrasonication of carbon nanotubes in solvents such as alcohols is a common
technique for dispersing samples for electron microscopy. One way to improve
the dispersion of nanotubes is to shorten the tubes. The shorter tubes are less
likely to entangle and arrange into aggregates. However, there are some serious
disadvantages with breaking the tubes into smaller pieces. When the tube-walls
are broken in order to create a cut, the wall may become damaged in other ways
as well. “Worm-eaten” or “ragged” tube walls (Koshio, Yudasaka et al. 2001;
Koshio, Yudasaka et al. 2001; Zhang, Yudasaka et al. 2002) and walls with cuts,
buckles and irreversible bends (Esumi, Ishigami et al. 1996; Shelimov, Esenaliev
et al. 1998) are consequences of chemical processing, ultrasonication treatment
or a combination of both methods.
SWNT lengths decrease only after the bundle size has gotten smaller (Yudasaka
2002). SWNTs rearrange into super-ropes after the bundles are broken up and
the SWNTs are shortened (Shelimov, Esenaliev et al. 1998; Ausman, O'Connell
104
et al. 2001). These super-ropes have diameters more than twenty times the
initial bundle diameter. There have been attempts to develop less destructive
ultrasonication methods. One example is ultrasonication with diamond crystals,
a method that reportedly destroys the SWNT bundles but not the tubes (Haluska,
Hulman et al. 2001). Raman-spectra show typical SWNT peaks even after 10
hours of treatment with this method.
Ultrasonication creates expansion and peeling or fraction of MWNT graphene
layers. The destruction of multiwalled nanotubes seems to initiate on the
external layers and travel towards the center. The nanotube layers seem quite
independent, so MWNTs would not only get shorter, but actually thinner with time
(Lu, lago et al. 1996).
Ultrasonication is an extremely common tool used to break up nanotube
aggregates during purification, mixing, and other types of solution processing
techniques. Therefore, the nanotube morphology in the suspension or solid is
that developed during processing, and not that of the original nanoparticle
additive. In some case, ultrasonication can be used to remove impurities.
SWNTs have been purified from ~ 70% to ~90% by ultrasonication-assisted
filtration. About 30-70% of the starting material was not recovered in this process
(Shelimov, Esenaliev et al. 1998). Ultrasonication also may lower nanotube
quality. For example, ultrasonication lowers the oxidation onset temperature
from 600oC to 500oC for MWNTs. This onset temperature is lower than for
graphite, which is approximately 540oC (Lu, lago et al. 1996).
6.3.1 Ultrasonication Bath
There are two major methods for delivering ultrasonic energy into liquids, the
ultrasonic bath and the ultrasonic horn or wand. Ultrasonication disperses solids
primarily through a bubble nucleation and collapse sequence.
The ultrasonication bath has a higher frequency (40-50 kHz) than cell
dismembrator horns (25 kHz). Ultrasonication of fluids leads to three physical
mechanisms: cavitation of the fluid, localized heating, and the formation of free
radicals. Cavitation, the formation and implosion of bubbles, can cause
105
dispersion and fracture of solids. The frequency of the ultrasound determines the
maximum bubble size in the fluid. Low frequencies (about 20 kHz) produce large
bubbles and high energy forces occur as they collapse. Increasing the frequency
reduces bubble size and nucleation, so that cavitation is reduced. Cavitation
does not occur in many liquids at frequencies higher than 2.5 MHz. The
ultrasonication bath does not produce a defined cavitation zone as does a horn
and the energy seems to be more uniformly dispersed through the liquid phase.
Systems with low frequencies (20 – 100 kHz) and high power (100 – 5000 W) are
used to modify materials (Povey 1998).
Bubbles nucleating at solid surfaces and rapidly expanding can push particles
apart. Solid particles can remain separated after bubble collapse if they are
wetted by the fluid phase and if the volume fraction of nanoparticles in the fluid
phase is low enough so that solid movement is possible.
Experimental Procedure
Multiwalled carbon nanotubes (wall material) were dispersed in toluene by using
an ultrasonication bath (a frequency of ~55 Hz). The MWNTs had initial
dimensions of L = 50 µm, Di = 2.9 nm and Do = 25 nm, respectively. Water in the
ultrasonication bath was kept at the same level as the toluene in the 200 mL
glass beaker, thus promoting a uniform energy distribution. The MWNT loading
was 0.1 wt%, and samples were taken from the bath after 5, 10, 15, 20, and 25
minutes. Each sample was analyzed by using Hitachi 3200N Variable-Pressure
SEM. Ten SEM microphotographs were taken of each sample and from these a
minimum of a hundred MWNT lengths were measured.
All MWNT lengths or agglomerate sizes reported in this chapter were fitted to the
log-normal distribution model discussed in Chapter 3 using Systat®. The
measured MWNT lengths were plotted cumulatively and modeled using the
integral of Equation 3.1.
106
Results and Discussion
Figure 6.1 compares the distributions of the time sequence samples. The model
parameters are presented in Table 6.1. Equation 6.1 was fitted to the mean
average length data (Figure 6.2). The change in average length during the first
few minutes of ultrasonication is dramatic. After the first five minutes, the
average length is reduced from 50 to 17 µm, a decrease of more than 65%.
According to the fraction rate model, the tubes are reduced by 27% after only 30
s of treatment. After processing for an additional 25 minutes, the MWNT average
length decreased to 6.5 µm, which is about 12% of the initial mean length. It is
not unexpected that the fraction speed slows down after a longer period of
treatment, since it takes more energy to break shorter, more stable, tubes. The
average length data are well described by a model that is cubic in length (kb =
3.55x10-4, b = 3). Parameter b is actually 2.9 ± 0.12, but sat to 3, since we can
then use the integer for simple analysis via McCoy (McCoy 1994). We have
done limited comparisons of the moments for b = 3 and we found excellent
correspondence between theory and actual distribution changes.
6.3.2 Ultrasonication Horn or Wand
The tips of an ultrasonic horn or wand oscillate at a fixed frequency with variable
power being applied to the fluid phase. The rapid oscillation of the wand tip
produces a conical field of high energy in the fluid. The solvent within this conical
field undergoes nucleate boiling and bubble collapse, which is the primary
mechanism by which ultrasonic energy disperses materials. The wand tip
vibrations, along with the rapid generation and collapse of bubbles, induce a flow
that moves away from the wand tip and then recirculates through the conical
zone again. The size of the zone and the local velocity fields depend on the
boiling point of the solvent, the fluid phase viscosity, the energy applied, and the
geometry of the vessel and the wand placement.
107
Cumulative length distributions for MWNTs in US bath
0.0
0.2
0.4
0.6
0.8
1.0
0.0 10.0 20.0 30.0 40.0 50.0
Length [µm]
Cum
ulat
ive
Sum
5min10min15min20min25min
Figure 6.1 Cumulative MWNT length distribution as a function of processing
time in ultrasonication bath. The initial distribution changes are quite dramatic,
while the changes are more moderate after a longer treatment period. As can be
seen, the curves seem to take a wave form. This may be an artifact of the
particles measured. We have observed possible multimodal distributions and we
may need ~ 1000 data points to verify if this is the case here. Based on
experience we determine that the data presented in the graph above is sufficient
to interpret the length breakage rate for ultrasonication.
108
Table 6.1 MWNTs treated in ultrasonication bath. The MWNT mean length,
eµ, and standard deviation, σ, are the model parameters used to fit the log-
normal model to the data. R2 values are above 0.99 for the fits. The initial
MWNT average length was 50 µm.
Time [ min ]
Mean length eµ
[ µm ]
Standard deviation
σ
0 50 -
5 16.2 0.41
10 10.4 0.54
15 8.6 0.52
20 6.6 0.50
25 6.5 0.43
Copyright © Jenny Marie Hilding 2004
109
MWNT length as a function of time in ultrasonication bath
0
10
20
30
40
50
0 5 10 15 20 25
Time, t [ min ]
Leng
th, L
[
µm ]
Figure 6.2 The time derivative of MWNT length in US-bath. The data is fitted
with a power model where the frequency factor is 3.55 x 10-4 and the exponential
parameter is 3. R2 = 0.999.
110
The volume fraction of nanotubes in the suspension affects the solid surface per
fluid volume. Fluids with high continuous phase viscosities, for example, polymer
solutions and spinning dopes, may not give rapid recirculation of the process
liquid through the sonication zone. Since suspensions of MWNTs are shear
thinning (Karicherla 2001), the flow field near the wand tip may be only a small
volume and may have low recirculating velocities through the sonication zone,
leading to low dispersion efficiencies. MWNT suspensions with polymer
solutions as the continuous phase may also have greatly reduced fluid
circulations near the wand tip. Not all the bubbles may collapse immediately,
particularly if the solvent does not wet the nanotubes well or if a polymer solution
continuous phase of high viscosity reduces the rate of bubble coalescence. At
high solids loadings, the nanotubes can trap gas bubbles and create a rigid
network that prevents fluid flow.
Experimental Procedure
A 0.1 wt% dispersion of MWNTs in toluene was treated with an ultrasonication
horn. Data were collected and modeled in the same manner as above.
Results and Discussion
Figure 6.3 shows the cumulative size distribution data sets with fits. The data is
presented in Table 6.2. Figure 6.4 shows the fit of Equation 6.1 to these data (kb
= 4.06x10-4, b = 3, actual value 3.1 ± 0.7 (McCoy 1994)). The data set is well-
described by a cubic model, but with a slightly different rate coefficient compared
to the US-bath model parameter kb. The difference between the bath and wand
sonication frequencies may account for some of this change. Another variable
may be the MWNT initial length, which differed greatly between the bath (Lo = 50
µm) and the wand experiments (Lo = 25 µm). The length decreased 54% during
the first 5 minutes and according to the fracture rate model, the tubes are
reduced by 11% after only 30 s of treatment. After 25 minutes the average
length was decreased to 6.9 µm, which is 28% of the initial length. Clearly, the
treatment has a major impact on the MWNT morphology.
111
Cumulative length frequency for MWNTs with US horn
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 10 20 30 40 50 60
Length [ µm ]
Cum
ulat
ive
Sum
5 Min
10 Min
20 Min
25 min
Model
Figure 6.3 Plot showing the cumulative MWNT length distribution as a function
of processing time with ultrasonication wand.
112
Table 6.2 MWNTs treated with ultrasonication wand. The MWNT mean
length, eµ, and standard deviation, σ, are used as model parameters to fit the log-
normal model to the data. R2 values are above 0.99 for the fits. The initial
MWNT average length was 25 µm.
Time [ min ]
Mean length eµ
[ µm ]
Standard deviation
σ
0 25 -
5 11.4 0.58
10 10.8 0.55
20 8.2 0.66
25 6.9 0.59
Copyright © Jenny Marie Hilding 2004
113
MWNT length as a function of time with ultrasonication wand
0
5
10
15
20
25
30
0 5 10 15 20 25Time, t [ min ]
Leng
th, L
[ µ
m ]
Figure 6.4 The time derivative of MWNT length in US-bath. The data are fitted
with a power model where the frequency factor is 4.06 x 10-4 and the exponential
parameter is 3. R2 = 0.997.
114
6.4 Stretching of MWNTs in PAN Spinning Dope
It is important to develop dispersion methods and processes that generate
homogeneous MWNT dispersions, since the dispersion quality most likely will
directly affect the product quality. One example is the development of
continuous fibers containing carbon nanotubes. Such systems could be used as
high-strength or high-toughness fibers. The use of high-aspect-ratio solids in
spun fibers will depend on uniform, independent particle dispersions, proper
orientation of the fibers in the spinneret or die, drawing to develop fiber strength.
Poor dispersion will result in unspinnable polymer dopes.
Weisenberger (Weisenberger 2002) studied the spinning of polyacrylonitrile
(PAN) dopes loaded with MWNTs for producing MWNT/PAN fiber composites. A
cell dismembrator (horn ultrasonication at 25 kHz) was used to disperse MWNTs
into a solvent (dimethyl acetamide) before adding polymer to form the spinning
dope. The sample was cooled in an ice bath to remove excess heat. The fluid
was sonicated at 300 W of power for ten second intervals over various time
periods to produce well-dispersed suspensions for spinning. The quality of the
dispersion was evaluated by determining whether the dope prepared from the
suspension could be spun into continuous fibers.
We measured Weisenberger’s stretched MWNTs from SEM microphotographs,
to investigate the impact of fiber spinning on the MWNT morphology. The
MWNT lengths, before and after fiber spinning, can be compared in Figure 6.5
and Table 6.3. No length decrease was observed. This result may not be
surprising, since it is expected that a MWNT incorporated in a solid matrix will be
subjected to ‘sword-in-sheath’ failure mechanism. This type of mechanical failure
will cause damage to and peel off the MWNT wall outermost layer only.
Therefore, the only MWNT morphology change in the system should be a non-
measurable diameter decrease (recall that dwall ~0.3 nm but DO = 400 nm). It
also means that the MWNT length should be the same after the stretching
process as it was before.
115
Comparison between stretched and un-stretched MWNTs
0
0.2
0.4
0.6
0.8
1
0 1 2 3 4 5 6 7 8
Length, L [ µm ]
Cum
ulat
ive
Sum
Un-stretchedStretchedModel1Model2
Figure 6.5 Cumulative distribution of MWNTs from spun and drawn
MWNT/PAN fibers.
116
Table 6.3 Comparison of MWNT lengths of spun and drawn samples.
MWNTs in PAN.
Sample
Mean length, L
[ µm ]
Standard deviation
σ Un-stretched 3.2 0.36
Stretched 3.0 0.56
117
6.5 Grinding
There are few reports on the subject of rubbing or grinding carbon nanotubes to
decrease the size. Rubbing is more destructive than anything else. The process
introduces cuts and bends in SWNTs, but no change in storage capacity is
observed (Haluska, Hulman et al. 2001). A less damaging method is chemically
cutting SWNTs by grinding them in a fluid (α- or ß-cyclodextrin) using mortar and
pestle. Both tube lengths and bundle diameters were noticeably reduced. Other
grinding agents were used as well, but did not give as good results; samples
contained mostly long tubes (Chen, Dyer et al. 2001).
Experimental Procedure
We chose to research how the MWNT morphology changes when MWNTs are
hand-ground with a mortar and pestle. The MWNTs were mixed with a small
amount of toluene, creating a thick paste. The paste was ground for
approximately an hour, with no further addition of toluene. The MWNT length
and diameter were measured before and after the grinding, using a Hitachi
3200N Variable-Pressure SEM. In addition, the MWNT agglomerate particle
sizes were measured as well (MWNTs from the CVD process are entangled as
harvested from the reactor and are associated into agglomerates). The
diameters, tube length, aspect ratio, and agglomerate data sets were fitted using
the log-normal model discussed in Chapter 3.
Results and Discussion
Both the lengths and agglomerate sizes decrease significantly, as can be seen
from SEM photographs. Figure 6.6 shows MWNT diameter and length before
grinding and Figure 6.7 shows the MWNTs after the grinding process is
completed. Figure 6.8 compares the agglomerate size before and after grinding.
118
Figure 6.6 MWNT lengths before mechanical grinding with mortar and pestle
119
Figure 6.7 MWNT lengths after mechanical grinding with mortar and pestle.
120
Figure 6.8 SEM microphotographs showing the diameter decrease for MWNT
agglomerates after approximately one hour of mechanical grinding with mortar
and pestle. Observe the different scales of the different halves.
121
The MWNT inverse aspect ratio is plotted, together with the log normal fit, in
Figure 6.9. The diameter, length, and agglomerate size distributions before and
after grinding, accompanied by respective fitted log-normal model curve, are
presented in Figures 6.10, 6.11, and 6.12, respectively. The MWNT diameter
distributions, before and after grinding, are in essence identical, with an average
of approximately 18 µm. The length distribution changes from nearly
monodisperse (L = 55 microns) to a log-normal distribution with an average
length of 3 microns. The typical agglomerate size was reduced from 150 microns
to 50 microns, an 80% decrease.
The grinding process produced significant defects in the MWNTs. An example of
different types of defects introduced to the MWNT sample by grinding can be
seen in Figure 5.4.
Copyright © Jenny Marie Hilding 2004
122
Modeling the inverse aspect ratio for nontreated and ground MWNTs
0
0.2
0.4
0.6
0.8
1
0 0.05 0.1 0.15 0.2 0.25 0.3
Inverse aspect ratio, Af (d/L) [ - ]
Cum
ulat
ive
Sum
Nontreated MWNTs
Ground MWNTs
Model 1
Model 2
Figure 6.9 Cumulative distribution of aspect ratios for ground and unground
MWNTs. Parameters are d/L = 64×10-3, σ = 0.48 for Model 1, and d/L = 31×10-3,
σ = 0.48 for Model 2.
Copyright © Jenny Marie Hilding 2004
123
Diameter distribution - Nontreated and ground MWNTs
0
0.2
0.4
0.6
0.8
1
0 0.1 0.2 0.3 0.4 0.5
Diameter, d [ µm ]
Cum
ulat
ive
Sum
Nontreated MWNTs
Ground MWNTs
Model 1
Model 2
Figure 6.10 Cumulative distribution of MWNT diameters for ground and
unground MWNTs. Model parameters are d = 0.17 µm, σ = 0.48 for Model 1,
and d = 0.20 µm, σ = 0.34 for Model 2. R2 is above 0.9 for both models.
Copyright © Jenny Marie Hilding 2004
124
MWNT length for nontreated and ground MWNTs
0
0.2
0.4
0.6
0.8
1
0 10 20 30 40 50 60 70
Length, L [ µm ]
Cum
ulat
ive
Sum
Nontreated MWNTs
Ground MWNTs
Model
Figure 6.11 Cumulative distribution of MWNT lengths for ground and unground
MWNTs. Model parameters for ground nanotubes are l = 30.3 µm, σ = 0.53 with
R2 above 0.999.
Copyright © Jenny Marie Hilding 2004
125
Particle size distributionComparison between nontreated and ground MWNTs
0
0.2
0.4
0.6
0.8
1
0 200 400 600 800 1000
Particle size, L [ µm ]
Cum
ulat
ive
Sum
Nontreated MWNTs
Ground MWNTs
Model 1
Model 2
Figure 6.12 Cumulative distribution of particle sizes for ground and unground
MWNTs. Model parameters are L = 40.6 µm, σ = 0.71 for Model 1, and d =
169.3 µm, σ = 0.71 for Model 2. R2 is above 0.99 for both models.
Copyright © Jenny Marie Hilding 2004
126
6.5.1 Percolation Theory
Percolation theory (Stauffer and Aharony 1992) suggests that solids can produce
three-dimensional networks, if the solid is properly dispersed in the dispersant.
The theory predicts the network to have the same transport properties as the
pure solid. When such a network is completely connected in a matrix, a sudden
drop in transport resistance can be measured. This drop occurs at the
percolation threshold, pc. Equation 6.2 shows an empirical formula developed to
predict the pc as a function of the aspect ratio, A f (Stauffer and Aharony 1992)
325.1
2
658.1763.133.1261.14742.7
875.9
ffff
ffc AAAA
AAp
⋅+⋅+⋅+⋅+
+⋅= Equation 6.2
To investigate how the grinding process affects the MWNT electrical structure,
we decided to measure the electrical resistance as a function of wt% ground
MWNT filler in epoxy. If the MWNT structure is severely damaged it is
reasonable to expect that no drop in electrical resistance will occur, even if the
MWNT is homogeneously dispersed in the matrix. If the MWNTs are still
conductive, pc can be used to determine if the MWNTs are homogeneously
dispersed in the epoxy and/or the amount of MWNTs necessary to create the
network.
Experimental Procedure
Samples of five different compositions were generated and measured; 0.04, 0.1,
0.4, 1.0, and 4.0 wt% ground MWNTs in epoxy. For each composition, three
different samples were individually prepared to ensure accurate sample
generation, measurements, and analysis. The non-transparent epoxy-MWNT
blends were analyzed using a light microscope, to confirm homogeneity of the
sample surface. Each blend was thereafter used as contact cement between
copper-covered strips of circuit board. After the blend had been cured under
atmospheric conditions, each sample’s thickness, length, and width were
measured. Each sample’s electrical resistance was measured five times at tree
different occasions with two different instruments. The measurement were made
127
both in the x - and the y -directions, thereafter multiplied by the area and divided
by the thickness to enable comparison. The collected data was analyzed
according to percolation theory.
Results and Discussion
We used percolation theory (Equation 6.2) to predict the percolation threshold for
MWNT filler, both before and after grinding. The predicted values are plotted
together with the function pc in Figure 6.13. In reality, the percolation threshold
for our system was reached for epoxy containing between 0.4 and 1.0 wt%
MWNT, as can be seen in Figure 6.14. However, the drop in electrical resistance
was not as steep as it should be in an ideal system, according to theory. It is
also curious that the electrical resistance seemed to increase for the epoxy
containing 0.1 wt% MWNTs. We have no single explanation for this result. It is
possible that the MWNTs are not perforating the epoxy surface sufficiently, which
leads to poor electrical contacts between the MWNTs. It has previously been
suggested that variable agglomeration may occur, particularly for volume
fractions near the percolation threshold, but no rigorous proof has been
published. The experimental MWNT loading is approximately 10 times higher
than the predicted value, 0.05 wt% (Figure 6.14). However, typical percolation
thresholds for carbon black in epoxy are found in the range of 20-40 wt% (Dutta,
Misra et al. 1992; Sandler, Shaffer et al. 1999), which means that we only need
MWNTs of 2.5-5% the mass of carbon black necessary to reach the percolation
threshold in epoxy. Measured percolation thresholds for other systems have
been discussed in detail in Chapter 2 and show that systems using CNT as filler
reach the percolation threshold at significantly lower loadings.
128
Figure 6.13 The percolation threshold, pc, calculated from an empirical
equation (Equation 7.2) and plotted as a function of aspect ratio, Af (Stauffer and
Aharony 1992). The calculated values for ground and non-ground MWNTs are
plotted in the curve to show where their respective thresholds are expected to
occur.
129
Resistance in Epoxy as a function of wt% MWNT loading
1.E-03
1.E-02
1.E-01
1.E+00
1.E+01
1.E+02
1.E-02 1.E-01 1.E+00 1.E+01
MWNT loading [ wt% ]
Res
ista
nce,
ρ
[ kΩ
*m ]
Figure 6.14 The percolation threshold, here demonstrated as a drop in
electrical resistance, plotted vs. MWNT loading (in wt%) in epoxy. The used
MWNTs have been pretreated through grinding and have an aspect ratio of
approximately 3. The percolation threshold can be found around a loading of 0.4
wt% MWNTs for this particular system. The error-bars in the graph represent the
standard deviations between the different samples and the different
measurements for each sample.
130
From our result it is possible to conclude that the MWNT dispersion is
homogeneous in the matrix, since a threshold is reached. It also shows us that
short and damaged MWNTs still can be used to create a conductive network in
epoxy. The discrepancy from the predicted threshold may be a result of
structural defects introduced to the MWNTs through the grinding process. These
defects may affect the MWNT electrical structure in a negative fashion. A MWNT
is electrically conductive as long as at least one of the walls is conducting.
However, if we assume that the outer tube walls are most damaged, it may give
a reasonable explanation to why the percolation threshold occurs at larger
loadings than expected, since the outermost walls are likely to present the
contact points between the tubes in the matrix. It is also of interest to recall that
Equation 6.2 was developed before MWNTs were available. It is likely that the
equation is not applicable for systems containing this unique type of material.
6.6 Summary and Conclusions
Dispersing carbon nanotubes into liquids can be achieved through either
mechanical (physical) or chemical methods (see Chapter 2), which could be a
very challenging task pending on the quality and surface morphology of the
nanotubes used.
We have researched a number of different mechanical dispersion methods;
ultrasonication with bath and wand, stretching and grinding. It was discovered
that the mechanical dispersion methods are highly destructive to MWNTs; that is
the processed sample’s morphology is subjected to drastic changes.
The stretching of MWNTs encapsulated in a polymer matrix does not affect the
MWNT morphology. It is likely that the MWNTs remain essentially intact thanks
to the sword-in-sheath mechanism; only the outermost wall is broken and the rest
of the MWNT remains intact.
131
Ultrasonication shortens MWNTs very quickly. Since ultrasonication is the most
common dispersion and processing tool in the field of CNTs technology today,
the results of these investigations are crucial. Ultrasonication produces shorter
tubes with clean breaks. We propose that bubbles are first and foremost
nucleated at defect sites. When a bubble collapses at the defect site, a rapid
shock wave is generated perpendicular to the MWNT axis, resulting in a fairly
complete break due to a very high strain rate.
Grinding a MWNT paste is extremely destructive to the tubes. The method also
decreases the size of MWNT agglomerates, but does not eliminate them
completely. A number of different defects are introduced to the MWNT system.
Theoretically, these defects should lead to a severe degradation of the MWNT
properties. This seemed to be confirmed through a simple study in percolation
theory, even though the short and damaged MWNTs still are superior compared
to most other commercially used fillers. The grinding process subjects the
MWNTs to high strain but the strain rate is low in comparison to the
ultrasonication process.
Finally, the studies accounted for in this chapter establish that the most common
and most widely used mechanical dispersion methods introduce severe defects
in MWNT dispersions. With the results from our investigations presented in this
chapter we have shown that it is necessary to analyze the nanotubes after any
mechanical mixing process is completed, not before. This is true also for
relatively short periods of processing time, even on the order of a few seconds.
It would be of interest to further investigate the breaking mechanisms involved in
the ultrasonication process as well as the grinding process.
Copyright © Jenny Marie Hilding 2004
132
Chapter Seven
Alignment of Carbon Nanotubes in an Electrical Field
7.1 Introduction
As discussed in Chapter 6, the high aspect ratio of carbon nanotubes helps to
significantly decrease the amount of CNT filler necessary to reach the percolation
threshold in a matrix. Practically, this means that the CNT filler could change a
composite’s transport properties without severely interfering with the mechanical
strength in the matrix, thereby weakening the material (see Chapter 2.3.1).
Through incorporation of CNT in a material matrix, the properties of the CNTs are
transferred to the matrix producing a novel, hopefully dramatically improved,
material.
The matrix can either be solidified from a liquid state, such as in the case of a
ceramic, metallic, or polymeric material or the matrix can continue to exist in a
liquid form, for example as an oil. The ability to align CNTs in a matrix would
facilitate the development of novel engineered materials. Since a MWNT will
conduct if at least one of its layers conducts (Saito, Dresselhaus et al. 1993) we
conclude that electrical alignment of MWNTs should be a simple, cheap, and
efficient method for orienting these solids in a liquid medium.
A number of different electrical alignment methods have been designed and
evaluated (Chernozatonskii, Fedorov et al. 1993; Ajayan, Stephan et al. 1994;
Ajayan 1995; de Heer, Bacsa et al. 1995), but there are only a couple of non-
destructive methods that enable the alignment of CNTs in a non-limited volume
size. One example of such a method is the alignment of CNTs in an applied
electrical field.
The advantages of alignment are many. First, alignment may enhance the ability
to move, separate, and hence manipulate single CNTs. This opens the door to
the evolvement of greatly improved and simplified CNT property analysis
techniques. A simple example is that it is significantly easier to measure
133
morphological distributions (diameter, length, etc.) if the CNTs are disentangled.
Secondly, it should be a simple task to use electrical fields to separate CNTs
from non-conductive sample impurities or badly damaged CNTs, either through
electrical migration of the conductive CNTs towards the electrodes or through
gravitational force on non-conductive particles. Recent experimental data has
even suggested that it is possible to sort and separate CNTs according to size in
an electrical field (Bubke, Gnewuch et al. 1997). Alignment could also lead to
production of ‘1D’ materials; that is, it should be possible to produce composites
that are conductive (Haggenmueller, Gommans et al. 2000) or very strong in only
one dimension. Finally, alignment could enable a simple production process of,
for example, membranes with narrow pore distributions simply by slicing an
already made composite perpendicular to the aligned CNTs.
A major part of the CNT research to date has concentrated upon theoretical
models and predictions (Fishbine 1996), as opposed to more practically oriented
investigations. At present, only a handful of articles are available reporting actual
experimental data and observations regarding the behavior of nanotubes in
electrical fields (Yamamoto, Akita et al. 1996; Bubke, Gnewuch et al. 1997;
Chen, Saito et al. 2001).
The experiments and results reported here were made as an introductory step in
investigating the possibilities to align MWNTs in electrical fields. We were
primarily interested in determining how the different system parameters would
affect rate and quality of MWNT alignment. Parameters particularly interesting
are MWNT morphology, electrical field strength and frequency.
134
7.2 Alignment of Dispersed MWNTs
The MWNT samples used in the experiments were CVD-produced at CAER.
The nanotubes were dispersed in water with the aid of a non-ionic surfactant
(Igepal CO-630 Rhodia, or nonyl phenol ethoxylate) followed by a short period
(~10 s) of ultrasonication in a water bath. Igepal has been shown to give stable
MWNT dispersions over time periods of about two weeks in aqueous systems.
No attempt to optimize the stability of the dispersion has been done at this time.
An EG&G galvanostat generator (model 5208) was used to generate the
electrical field. An oscilloscope was used to measure the voltage and frequency
emitted. The Ag/AgCl electrodes were printed directly onto the glass slides. The
generator and oscilloscope were connected with wires and alligator clips to the
electrodes on the glass slides. As the electrodes can be easily scraped off the
glass surface when in contact with the crocodile clips, aluminum foil was
wrapped around the electrodes for protection. This method worked well as it
provided minimum electrical resistance. An optical microscope (Axiotech 25 HD,
Zeiss) with a video camera was used to capture movies of the experiment.
After the glass slide was placed on the optical microscope, the nanotube-water
suspension was carefully dropped between the electrodes. The alternating
current (AC) electrical field was then applied to the electrodes. The experiment
was done with field strengths ranging from 15 V/cm to 140 V/cm at frequencies of
0.10, 20, 100 and 500 Hz. Movies between the lengths of 15-45 seconds were
captured to observe the reaction of the dispersed nanotubes to the electrical
field. The movies were later studied to observe the rate of alignment of the
nanotubes, dispersion, connections and size dependence.
7.2.1 Results
The alignment quality was found to be a function of both frequency and AC field
strength. The highest quality alignment occurs when all CNTs are parallel to the
field direction. A comparison of alignment quality between dispersed MWNTs
before and after alignment can be seen in Figure 7.1.
135
Figure 7.1 MWNT-water dispersions as photographed through an optical
microscope. The upper picture shows the dispersion before the electrical field
has been applied and the lower picture shows the dispersion after the alignment
equilibrium has occurred. Nanotubes in the lower picture appear blurred due to
their motion.
136
Flow Alignment at Low Frequency
At the lowest frequency, 0.1 Hz, the dispersion was forced into a pumping wave
motion, which slowly altered directions between the electrodes. The nanotubes
would orient themselves (and move) parallel to the electrical field lines, but as a
result of the pumping motion the MWNTs could not stay aligned in the applied
field. Instead, the MWNTs were forced to align with the dispersion streamlines,
hence forced to rearrange every time the electrical field direction changed. This
motion pattern was not found to be size dependent, but MWNTs of all sizes
followed the streamlines with identical velocity. No such movements were
observed for frequencies above 0.1 Hz (≥ 20 Hz), but here the alignment seemed
to be a function of the magnitude of the applied field alone.
Alignment Transition
No alignment was observed for field strengths of 15 and 30 V/cm within the time
span of the experiments. At 15 V/cm, the MWNTs seemed to exhibit Brownian
motion in the water. At 46.5 V/cm and 20 Hz, no alignment was observed, but at
an elevated frequency (100-500 Hz) some alignment occurred in the time-span of
5-15 s. For field strengths of 70 V/cm and above (72 and 140 V/cm), all the
dispersed MWNTs aligned quickly and completely. The time periods were
sufficiently shorter and in the order of 1 s.
7.2.2 Discussion
The MWNTs moved and aligned independently of each other and the electrodes.
Previously, the CNT motion towards the cathode in a isopropyl alcohol dispersion
has been estimated to ~5×10-5 cm2V-1s-1 (the velocity is proportional to the DC
field strength) (Yamamoto, Akita et al. 1996). In a DC field, our MWNTs would
be expected to move 2.1 mm closer to the cathode during 30 s in an applied field
of 140 V/cm. Such a movement should have been easily detectable but did not
occur. Because we used an AC field, the integrated net force is zero, so no net
gradient should be observed. Another report claims that the estimated mobility of
5×10-5 cm2V-1s-1 is much higher than what is required to account for observed
effects in ethanol (Bubke, Gnewuch et al. 1997), even though accumulation on
137
the cathode was observed. This electrophoresis effect is no doubt dependent on
the type of electrical field applied; an AC-field is expected to reduce particle
motion compared to motions induced by a DC-field. However, anisotropic effects
have been observed for CNTs in AC-fields but over time periods of minutes
(Bubke, Gnewuch et al. 1997). No motion was observed when a DC-field was
applied to the system for SWNTs in ethanol (Chen, Saito et al. 2001).
No DC-field experiments were carried out in our system. Furthermore, no
systematic motion other than rotation in a fixed location was observed, probably
a result of the AC. Rotation occurred both in the MWNT x-plane and θ-plane, so
the MWNTs were oriented parallel with the field lines and rotate around their own
core axis.
The few non-dispersed MWNT clusters present showed no signs of breaking up.
It has been suggested that higher frequencies than those used here are needed
to sufficiently break up aggregates (Chen, Saito et al. 2001). The tubes were not
observed to flex in any way, but stayed rigid during the course of each
experiment.
7.3 Alignment of Flocculated MWNTs
After leaving the dispersion to rest for approximately 1.5 months, we observed
that a phase separation had occurred. The dispersed sample had been stored in
a small, transparent glass bottle, which was sealed with a plastic lid. The bottle
was about half-full and the MWNTs were flocculated close to the bottom of the
bottle. The dispersant system was obviously not as effective as we initially
thought. There are a few possible explanations to the sudden change in
dispersion properties. First of all, a phase separation may have resulted in a
phase with a high surfactant concentration and one phase with MWNTs. This
type of phase separation is quite common in surfactant systems and is generally
both concentration and temperature dependent. It was confirmed that the
microscope lamp heated up the dispersion. However, the cloud point for a 1wt%
nonyl phenol ethoxylate solution is found between 51-56oC. It is unlikely that the
138
microscope would heat up the dispersion to temperatures of that magnitude. The
dispersion did not appear cloudy. The second explanation is that the system
may only slowly reach equilibrium. The surfactant is most likely adsorbing onto
the MWNT wall, but first it must diffuse to the surface through the water phase.
The MWNT surface also has to compete with the surfactant itself, which is most
likely self-assembling into micelles. For our system, micelle formation may be
energetically more favorable than MWNT-surfactant adsorption.
Experimental Procedure and Results
The flock could not be broken up simply by shaking the bottle, but after about 15
seconds of ultrasonication treatment in US-bath, the MWNTs were re-dispersed.
The re-dispersed sample was placed on the sample holder in the optical
microscope as described before. An electrical field of 140 V/cm and 500 Hz was
applied. Initially, no behavioral differences were observed, the MWNTs aligned
exactly as before. After about 5 minutes a rapid change occurred. The MWNTs
started to quickly move towards each other and the electrodes, see Figure 7.2.
In just a matter of seconds clusters of MWNTs were formed and continued to
branch out until the electrodes were connected. After the connection was
complete, the MWNT movements in the system quickly slowed down.
139
Figure 7.2 The MWNTs are quickly forming clusters that connect the
electrodes. The MWNTs are moving and swaying violently. One example of the
movement can be seen from small, circled, lump in the lower right hand corner of
the first picture. In the second picture the lump has moved and re-oriented itself
before getting stuck to the electrode surface.
Copyright © Jenny Marie Hilding 2004
2
3 4
1
140
7.4 Summary and Conclusions
CNTs have been reported to align and spontaneously connect to each other
during alignment processes in systems similar to ours, both in DC (Yamamoto,
Akita et al. 1996) and in AC (Chen, Saito et al. 2001). No such movements were
observed initially in our system, which may have been an effect of the non-ionic
surfactant.
Freshly prepared MWNT dispersions can be aligned through electrostatic dipole
moments induced in the nanotube by an applied electrostatic AC field. The
MWNT alignment rate is a function of both field strength and frequency. No
efficient alignment was observed for low frequency AC-systems (0.1 Hz) or for
field strengths below 70 V/cm. The MWNTs moved independently of the
surrounding system. The only movements observed were MWNT orientation
parallel with the applied field and rotation around their own axes. The MWNTs
were fully aligned in less than 1 second in an electrical field ≥70 V/cm with a
frequency ≥20 Hz. Poorly dispersed MWNTs seem to easily attract each other.
We observed MWNTs quickly assemble to clusters, which were built up to form
connecting bridges between the electrodes. Both results are of interest from an
engineering point of view and suggest that it is possible to control network
formation in this type of system.
We conclude that the alignment of CNTs in an applied electrical field may be an
efficient, quick, and cheap method with many advantages. The method opens
doors to many interesting and exciting possibilities for production of novel
materials with unique properties. In the future it would be of interest to
investigate monomer systems to continue the research of alignment of CNT filler
in polymeric matrixes. It is also necessary to investigate MWNT compatible
surfactants and dispersion transport phenomena in different MWNT dispersion
systems. Chemical dispersants and suspending agents may play important roles
in the alignment of CNTs in electrical fields. It is clear that there is a wide range
of issues and phenomena that need to be fur ther addressed before we can reach
a practical understanding for this type of system.
141
Chapter Eight
Summary and Conclusions
The main focus of the research reported here has been aimed toward the MWNT
surface interactions in a number of different science and technology fields.
MWNTs have exceptional mechanical and transport properties, which in
combination with the unusually large aspect ratio place MWNTs as a very
interesting material from an engineering point of view. We have investigated how it
is possible to utilize the unique nature of MWNTs in different engineering
applications and processes.
MWNT Morphology Analysis
To better interpret our data and research findings we have developed a
standardized method, which enables us to accurately measure, model, and
describe the morphology of any type of MWNT sample. We have established that
MWNT inner and outer diameter, length, and agglomerate size can be well
described by a two-parameter log-normal probability model (Equation 3.1). The
two fitting parameters are the logarithmic mean diameter µ and the standard
deviation s. We have also shown that a relatively limited number of data points is
sufficient to successfully describe the MWNT morphology with fit a log-normal
model fit. This type of characterization research has never before been carried out
in any area of CNT science. By analyzing five different as-produced MWNT
samples in accordance with our method, we have shown that the inner diameter is
found in a very narrow range (DI = 2-12 nm), so it is possible and practical to
assume a constant inner diameter for modeling purposes. The MWNT samples we
have used in our research can be defined as strictly mesoporous materials. The
outer diameter shows larger variations and is found in the range of ~20-400 nm.
Our method shows that the outer diameter varies both within each sample but that
also µ varies significantly between samples.
142
Our model has enabled us to easily calculate pore volume, inner surface area,
and outer surface area (Equations 3.3, 3.4, and 3.5) of any MWNT sample. This
proved essential for the interpretation of our data in the research of MWNTs
adsorption systems. As expected, samples with smaller mean outer diameter had
larger pore volumes and surface areas on a per-weight basis. The pore volume
was calculated to be between 0.7 and 5.4 mm3/g. This is very low for a porous
material and has never been calculated nor discussed before. However, the
pore volume is only a small fraction of the solid volume. We found that the outer
surface areas lie in the range of ~14-55 m2/g.
Adsorption on MWNTs
One of the areas this method has been very useful to us is in surface adsorption
science. We used several different samples with varying morphology and could,
with the aid of the morphology measurement method, develop a new
understanding for the butane-MWNT adsorption systems. We have shown that
butane adsorption at room temperature on CVD-produced MWNTs give rise to
an adsorption isotherm of type II, which in essence is physisorption on a
nonporous solid. The isotherms were successfully fitted with a modified BET
model, with excellent results. All the experiments were reproducible and all the
butane is readily desorbed. No hysteresis was observed, which is of great
advantage in a process involving adsorption of valuable or toxic compounds,
where activated carbons would irreversibly adsorb a percentage of the
condensed matter. Pore condensation accounts for less than 0.5% of the
adsorption capacity for all the samples. In addition to its low relative mass,
capillary condensation may not be detectable in the form of hysteresis since
according to the Kelvin equation, the pores are expected to be filled at wide
relative pressure range; between 0.3 and 0.9 depending on tube diameter and
accessibility. The remainder of the butane adsorption is assumed to be surface
condensation, which leads us to conclude that multilayer adsorption occurs. The
monolayer capacity is greater for samples with smaller diameters, but the
monolayer packing density is lowest for the sample with the smallest average
143
tube diameter. We were able to predict the BET monolayer capacity, nm, fairly
accurately as a function of our morphology models by applying an equation we
derived from simple geometrical relations. The results confirm that a number as
small as one hundred MWNTs is sufficient to analyze a sample’s morphology
accurately. The BET isotherms together with the generated monolayer capacity
model indicate that the butane molecules probably are tilted away from the
surface perpendicular. he adsorption capacity is larger for samples with smaller
diameters. It is clear that the sample diameter distribution has a major impact on
the adsorption capacity. As we have shown that the absolute majority of butane
is adsorbed on the MWNT outer surface and not in the pore, we conclude that
MWNTs are not primarily interesting as a gas storage tool, but future research
should be aimed towards the utilization of MWNTs in gas separation processes
instead. Processes involving toxic or valuable gases may be of specific interest
since hysteresis is non-present in the MWNT adsorption-desorption cycle.
Samples with smaller diameters are preferable as they can adsorb a greater
amount of gas.
We calculated the isosteric heat of adsorption, qst, from sets of isotherms by
applying the Clausius-Clapeyron equation. The initial heat of adsorption was
found to be between 22-26 kJ/mol. This is maximum 1.3 times the heat of
condensation and around 2/3 of reported qst values for butane adsorption on
graphite. The magnitude of qst did confirm that the adsorption mechanism is
physisorption. The high initial heat of adsorption indicates that the MWNT
samples act as heterogeneous adsorbents, most likely an effect of surface
defects. The minimum in qst, near the full monolayer capacity, relates to the high
self-association of butane. For θ>2.5, the isosteric heat of adsorption is
approaching the butane heat of condensation, suggesting a bilayer crystal
followed by a bulk phase build-up. We did not observe any relation between
sample morphology and isosteric adsorption. We conclude that the isosteric heat
of adsorption is most likely affected primarily by surface heterogeneity and not by
a sample’s diameter distribution.
144
MWNT Dispersion
We have investigated a number of mechanical methods, since they have the
advantage of not changing the chemistry of the MWNT or introducing additional
chemicals in the MWNT system. Two different types of ultrasonication were
investigated, together with manual grinding and stretching of MWNTs
incorporated in a matrix. The ultrasonication processes were investigated by
modeling the rate of length decrease as a function of MWNT breakage. The
stretching and grinding processes were evaluated by comparing the MWNT
morphology before and after treatment completion. In all cases the morphology
evaluation method we developed were used.
We determined that stretching of MWNTs in a polymer matrix has no impact on
MWNT morphology. This is most likely an effect of a “sword-in-sheath” failure
mechanism, which limits the damage to the outermost tube wall; hence no
change in diameter or length occurs.
Ultrasonication shortens MWNTs very quickly. The change in average length
during the first few minutes of ultrasonication is dramatic. Both methods
investigated could be described with a cubic rate model and have remarkable
similar rate constants even though the frequency differed between the US bath
and US wand experiments. According to the rate models, the initial MWNT
length is decreased by ~30% for the US bath process and ~50% for the US wand
process during the first 30 seconds. Since ultrasonication is the most common
dispersion and processing tool in the field of CNTs technology today, the results
of these investigations are crucial. We established that ultrasonication produces
short tubes with clean breaks through SEM analysis. We propose a breakage
mechanism where bubbles are nucleated at defect sites located at the MWNT
outer surface. When the bubble collapses at the already weak defect site, a
rapid shock wave is generated perpendicular to the MWNT axis, resulting in a
fairly complete break due to a very high strain rate. We conclude that the most
common and most widely used mechanical dispersion methods drastically
change the MWNT morphology and that the main mechanism behind dispersion
145
through ultrasonication is breakage of the tube. It is therefore necessary to
analyze the nanotubes after the mixing process is completed, not before. The
results are of great importance to many CNT research groups. The lack of
awareness of the severity of these types of CNT dispersion processes may
cause major problems in many CNT investigations.
Grinding efficiently shortens the MWNTs, but in addition several types of extreme
defects are introduced to the tube structure. The grinding process subjects the
MWNTs to high strain but the strain rate is low in comparison to the
ultrasonication process, which explains why the breaks are not clean. The
method also decreases the size of MWNT agglomerates, but does not eliminate
them completely. Theoretically, the structural defects should lead to a severe
degradation of the MWNT transport properties. This seemed to be confirmed
through a simple study in percolation theory; the experimental MWNT loading
was found to be approximately 10 times higher than the predicted value.
However, the percolation theory was developed long before MWNTs were
available. It is likely that the percolation model is not applicable to systems
containing this unique type of material. The short, damaged MWNTs are still
superior compared to other commercially used fillers; the percolation threshold
for our system was reached for epoxy containing between 0.4 and 1.0 wt%,
which is only 2.5-5% the amount of carbon black necessary to reach the
percolation threshold in epoxy. The occurrence of a percolation threshold
confirms that the MWNTs are homogeneously dispersed throughout the epoxy
matrix and that the structural damage is not sufficient to completely damage the
MWNT electrical structure. We conclude that it is possible to create a conductive
epoxy-based composite, without severely interrupting the epoxy matrix, thus
avoiding compromising the epoxy’s strength. This investigation shows that
MWNTs are very interesting as filler material in the area of novel composite
development.
We suggest further investigations and research to determine the breaking
mechanisms involved in the ultrasonication process as well as the grinding
process.
146
MWNT Alignment in Electrical Fields
We have shown that it is possible to align MWNTs in aqueous dispersions
through electrostatic dipole moments induced in the nanotube by an applied
electrostatic AC field. The MWNT alignment rate and quality is a function of
both field strength and frequency. We found that the frequency has to be above
0.1 Hz or alignment is not accomplished. The MWNTs were fully aligned in less
than 1 second in an electrical field of no less than 70 V/cm. MWNTs orientate
themselves parallel with the applied field and may also rotate around their own
axes. Poorly dispersed MWNTs seem to easily attract each other. MWNTs
quickly assemble to clusters, forming a network between the electrodes. We
found no relation between alignment speed or quality and MWNT morphology.
We conclude that the alignment of CNTs in an applied electrical field could be an
efficient, quick, and cheap method with many advantages. The alignment of both
well-dispersed and poorly dispersed MWNTs in electrical AC-fields gave
intriguing results. It seems obvious that there are many opportunities to find a
vide range of new MWNTs applications in the fields of composite development
and novel materials engineering. In the future it would be of interest to
investigate monomer systems to continue the research of alignment of CNT filler
in polymeric matrixes.
It is also necessary to investigate MWNT compatible surfactants and dispersion
transport phenomena in different MWNT dispersion systems. Chemical
dispersants and suspending agents may play important roles in the alignment of
CNTs in electrical fields. It is clear that there is a wide range of issues and
phenomena that need to be further addressed before we can reach a practical
understanding for this type of system.
Copyright © Jenny Marie Hilding 2004
147
APPENDIX A
LISTS OF SYMBOLS AND ABBREVIATIONS
Symbol:
Af aspect ratio ( - )
b rate exponent ( - )
c BET constant ( - )
D, Do MWNT external or outer diameter ( nm )
Dc , Di MWNT core or inner diameter ( nm )
dwall MWNT wall separation ( nm )
k BET constant ( - )
kB Boltzmann’s constant ( mol/molecule )
kb rate constant ( s-1 )
L length ( µm )
Lo initial length ( µm )
Mbut the butane molar mass ( g/mol )
n adsorbed quantity ( mol )
NA Avogadro’s constant ( molecules/mol )
nm BET const., monolayer adsorption capacity (mol)
p pressure ( atm )
p/po relative pressure ( - )
pc percolation threshold ( wt% )
po saturated vapor pressure ( atm )
148
qst isosteric heat of adsorption ( kJ/mol )
R the universal gas constant ( kJ/mol,K )
r pore radius ( nm )
rg butane radius of gyration ( Å )
rm meniscus radius ( nm )
T temperature (K)
t time ( s )
VL liquid molar volume ( m3 )
Vwall MWNT wall volume ( m3 )
α polarizability (cm3 )
γ surface tension ( N/m )
∆Ha adsorption enthalpy ( kJ/mol )
εB binding energy ( kJ/mol )
θ relative surface coverage, n/nm ( - )
µ mean of ln(x) ( - )
ρc density of graphite ( kg/m3 )
σ standard deviation of ln(x) ( - )
Abbreviation:
AC alternating current
AFM atomic force microscope
BET Brunauer, Emmett, and Teller
CAER Center for Applied Energy Research
CTAC cetyltrimethyl ammonium chloride
149
CVD chemical vapor deposition process
DC direct current
DIPPR Design Institute for Physical Properties
DTAB dodecyltrimethyl ammonium bromide
GCMC grand canonical Monte Carlo simulations
HOPG highly oriented pyrolytic graphite
LDS lithium dodecyl sulfate
MWNT multiwalled carbon nanotube
NT nanotube
PAN polyacrylonitrile
PEO poly(ethylene oxide)
PMMA poly(methyl methacrylate)
PMPV poly(1,3-phenylenediphenylvinylene)
PPV poly(p-phenylenevinylene)
PVA poly(vinyl alcohol)
RLM reflective light microscope
SDS sodium dodecyl sulfate
SEM scanning electron microscope
SPM scanning probe microscope
SPSWNT superhard phase single-walled nanotube
SWNT single-walled carbon nanotube
TEM transmission electron microscope
US ultrasonication
150
APPENDIX B
DERIVATION OF EQUATION 4.2
Number of butane molecules with a
radius of gyration, rg (2.89Å), over a
MWNT with the length L:
grL
A⋅
=2
Number of butane molecules with
radius of gyration rg around the
circumference of a MWNT with outer
diameter D:
g
g
r
rDB
⋅
⋅+=
2
)2(π
Total number, N, of butane molecules over a MWNT:
24
)2(
g
g
r
rDLBAN
⋅
⋅+⋅⋅=⋅=
π
Total numbers of butane molecules over n measured MWNTs:
∑∑== ⋅
⋅+⋅⋅=⋅=
n
n g
gn
n r
rDLBAN
12
1 4
)2(π
151
Assuming that the length, L, is the same for every MWNT:
∑∑==
⋅+⋅⋅⋅
=⋅=n
ng
g
n
n
rDrL
BAN1
21
)2(4π
The total weight of butane on the MWNTs:
∑=
⋅+⋅⋅⋅
⋅=n
ng
gA
but rDrL
NM
M1
2 )2(4π
where Mbut is the molar weight of butane (58.1 g/mol), and NA is Avogadro’s
constant.
The weight, w, for one MWNT with the wall volume Vwall, which is assumed to
have the same density as graphite, ρC, and an inner core diameter of DC:
−⋅⋅⋅=
⋅⋅−
⋅⋅⋅=⋅=
4)(
44
2222c
Cc
CwallC
DDLDLDLVw
πρ
ππρρ ( g )
Total weight, W, of n measured MWNTs:
∑∑==
−⋅⋅
=⋅=1
22
1
)(4 n
CC
n
nwallC DDn
LVW
πρρ ( g )
Predicted monolayer coverage, nm, pred:
−
⋅+⋅=
−
⋅+⋅
⋅⋅==
∑
∑
∑
∑
=
=
=
=n
nC
n
ng
n
nc
n
ng
gA
butmpredm
DD
rDk
DD
rD
rwNM
WM
n
1
22
1
1
22
12,
)(
)2('
)(
)2( ( g/g )
k’ is constant, so nm, pred is a function of D only, if we assume that DC is constant.
152
APPENDIX C
MAPLE PROGRAM
The Maple program below was written to calculate the monolayer capacity for
Samples 1-5 from the integrated log-normal diameter distribution. The program
works well, but interestingly, summation over the density distribution gives nm
values that are closer to experimental data, especially for the samples with a
wider distribution.
The program was also used to calculate the area occupies by each butane
molecule, simply by changing ‘rg1’ until nm1 is the same as our experimental
value.
$An attempt to calculate nm for a sample with a determined morphology distribution through integration. December 2002. Calculating nm as a function of D. Here, D=exp(x) is used, to assure that the distribution fits are correct. All geometrical measurements are given in units of nanometers. Constants for this particular sample, VGB61 - old morphology -(Sample 1)-. These constants are parameters from the normal probability distribution fit:$ > mu1:=ln(24.3312714023456) ; $log-normal mean of sample
1 morphology$ >sigma1:=5.69E-01; $log-normal standard
deviation of sample 1 morphology$
>rg1:=0.289; $butane radius of gyration, sample 1$
>Na:=6.023e23; $Avogadro’s constant$ >rho:=2045000e-27; $density graphite $ >m:=58.1222; $molar mass butane$ >k1:=m/(rho*Na*rg1^2); $constant k’, sample 1$ >di1:=2.63939832; $MWNT inner diameter,
assumed to be constant$
:= µ1 3.191762412
:= σ1 0.569
:= rg1 0.288
:= Na 0.6023 1024
153
:= ρ 0.2045000 10 -20
:= m 58.1222
:= k1 0.5689195967
:= di1 2.63939832
$The normal probability distribution fit:$ >f1:=(1/(sqrt(2*Pi)*sigma1))*exp(-(x-mu1)^2/(2*sigma1^2)); >plot(f1,x=0..10);
:= f1 0.87873462212 e
( )−1.544349072 ( ) − x 3.191762412 2
π
$The cumulative distribution fit:$ >F1:=Int(f1,x=0..N);plot(F1,N=0..10);value(Int(f1,x=0..10));
:= F1 d⌠
⌡
0
N
0.87873462212 e
( )−1.544349072( ) − x 3.1917624122
πx
154
0.9999999898
$Here does the geometrical calculations of nm begin. The equations are based on the theory that each butane is a spherical molecule, which takes up a space corresponding to butane's radius of gyration. From that assumption in combination with the known morphology of the sample, it is easy to calculate the exact number of moles of butane covering the sample surface - i.e. nm. The total diameter:$ > value(Int(f1*x,x=0..100));
3.191762413
$The area part:$ >narea1:=Int(f1*exp(x),x=0..100);area1:=(value(narea1))+2*rg1;
:= narea1 d⌠
⌡
0
100
0.87873462212 e
( )−1.544349072( ) − x 3.1917624122e x
πx
:= area1 29.18275593
155
$The volume:$ >nvolume1:=Int(f1*(exp(x)^2),x=0..100);volume1:=value(nvolume1)-di1^2;
:= nvolume1 d⌠
⌡
0
100
0.87873462212 e
( )−1.544349072( ) − x 3.1917624122( )e x
2
πx
:= volume1 1124.247750
> nm1:=k1*(area1)/(volume1);
:= nm1 0.01476777848
156
BIBLIOGRAPHY:
Ago, H., T. Kugler, F. Cacialli, W. R. Salaneck, M. S. P. Shaffer, A. H. Windle
and R. H. Friend (1999). "Work Functions and Surface Functional Groups of
Multiwall Carbon Nanotubes." Journal of Physical Chemistry B 103(38): 8116-
8121.
Ahn, C. C., Y. Ye, B. V. Ratnakumar, C. Witham, R. C. Bowman, Jr. and B. Fultz
(1998). Hydrogen desorption and adsorption measurements on graphite
nanofibers. Applied Physics Letters. 73: 3378-3380.
Ajayan, P. M. (1995). Aligned carbon nanotubes in a thin polymer film. Advanced
Materials (Weinheim, Germany). 7: 489-91.
Ajayan, P. M. (1999). Nanotubes from Carbon. Chemical Reviews (Washington,
D. C.). 99: 1787-1799.
Ajayan, P. M., C. Colliex, J. M. Lambert, P. Bernier, L. Barbedette, M. Tence and
O. Stephan (1994). Growth of manganese filled carbon nanofibers in the vapor
phase. Phys. Rev. Lett. 72: 1722-5.
Ajayan, P. M., O. Stephan, C. Colliex and D. Trauth (1994). Aligned carbon
nanotube arrays formed by cutting a polymer resin-nanotube composite. Science
(Washington, DC, United States). 265: 1212-14.
Alkhafaji, M. T. and A. D. Migone (1996). Heat-capacity study of butane on
graphite. Physical Review B: Condensed Matter. 53: 11152-8.
Andreoni, W. and Editor (2000). The Physics of Fullerene-Based and Fullerene-
Related Materials. [In: Phys. Chem. Mater. Low-Dimens. Struct., 2000; 23]: 445
pp.
Andrews, R., D. Jacques, A. M. Rao, F. Derbyshire, D. Qian, X. Fan, E. C.
Dickey and J. Chen (1999). Continuous production of aligned carbon nanotubes:
a step closer to commercial realization. Chemical Physics Letters. 303: 467-474.
157
Andrews, R., D. Jacques, A. M. Rao, T. Rantell, F. Derbyshire, Y. Chen, J. Chen
and R. C. Haddon (1999). Nanotube composite carbon fibers. Applied Physics
Letters. 75: 1329-1331.
Aranovich, G. L. and M. D. Donohue (1999). Adsorption of Chain Molecules.
Journal of Colloid and Interface Science. 213: 457-464.
Asakawa, T., Y. Kaneko and K. Ogino (1986). "Removal of trace amounts of
multicomponent organic compounds dissolved in water. II. Adsorption on carbon
adsorbents for phenol and anionic surfactant mixtures. I." Yukagaku 35(5): 367-
72.
Ata, M., A. J. Hudson, K. Yamura and K. i. Kurihara (1995). Carbon nanotubes
filled with gadolinium and hafnium carbides. Jpn. J. Appl. Phys., Part 1. 34: 4207-
12.
Au, C. T., W. Hirsch and W. Hirschwald (1988). Adsorption of carbon monoxide
and carbon dioxide on annealed and defect zinc oxide (000.hivin.1) surfaces
studied by photoelectron spectroscopy (XPS and UPS). Surface Science. 197:
391-401.
Ausman, K. D., M. J. O'Connell, P. Boul, L. M. Ericson, M. J. Casavant, D. A.
Walters, C. Huffman, R. Saini, Y. Wang, E. Haroz, E. W. Billups and R. E.
Smalley (2001). Roping and wrapping carbon nanotubes. AIP Conference
Proceedings. 591: 226-230.
Avery, R. G. and J. D. F. Ramsay (1973). Sorption of nitrogen in porous
compacts of silica and zirconia powders. Journal of Colloid and Interface
Science. 42: 597-606.
Aveyard, R. H., D. A. (1973). An introduction to the principles of surface
chemistry. London, UK, Syndics of the Cambridge University Press.
Awasthi, K., R. Kamalakaran, A. K. Singh and O. N. Srivastava (2002). Ball-
milled carbon and hydrogen storage. International Journal of Hydrogen Energy.
27: 425-432.
158
Ayappa, K. G. (1998). Simulations of Binary Mixture Adsorption in Carbon
Nanotubes: Transitions in Adsorbed Fluid Composition. Langmuir. 14: 880-890.
Bahr, J. L., J. Yang, D. V. Kosynkin, M. J. Bronikowski, R. E. Smalley and J. M.
Tour (2001). "Functionalization of Carbon Nanotubes by Electrochemical
Reduction of Aryl Diazonium Salts: A Bucky Paper Electrode." Journal of the
American Chemical Society 123(27): 6536-6542.
Bandyopadhyaya, R., E. Nativ-Roth, O. Regev and R. Yerushalmi-Rozen (2002).
"Stabilization of individual carbon nanotubes in aqueous solutions." Nano Letters
2(1): 25-28.
Banerjee, S. and S. S. Wong (2002). "Functionalization of Carbon Nanotubes
with a Metal-Containing Molecular Complex." Nano Letters 2(1): 49-53.
Battezzati, L., C. Pisani and F. Ricca (1975). Equilibrium conformation and
surface motion of hydrocarbon molecules physisorbed on graphite. Journal of the
Chemical Society, Faraday Transactions 2: Molecular and Chemical Physics. 71:
1629-39.
Baughman, R. H., C. Cui, A. A. Zakhidov, Z. Iqbal, J. N. Barisci, G. M. Spinks, G.
G. Wallace, A. Mazzoldi, d. De Rossi, A. G. Rinzler, O. Jaschinski, S. Roth and
M. Kertesz (1999). Carbon nanotube actuators. Science (Washington, D. C.).
284: 1340-1344.
Beebe, R. A., J. Biscoe, W. R. Smith and C. B. Wendell (1947). Heats of
adsorption on carbon black. I. J. Am. Chem. Soc. 69: 95-101.
Beebe, R. A., B. Millard and J. Cynarski (1953). Heats of adsorption of nitrogen
and argon on porous and nonporous carbon adsorbents at -195.degree. Journal
of the American Chemical Society. 75: 839-45.
Beebe, R. A., M. H. Polley, W. R. Smith and C. B. Wendell (1947). Heats of
adsorption on carbon black. II. J. Am. Chem. Soc. 69: 2294-9.
159
Beebe, R. A. and D. M. Young (1954). Heats of adsorption of argon on a series
of carbon blacks graphitized at successively higher temperatures. Journal of
Physical Chemistry. 58: 93-6.
Berber, S., Y.-K. Kwon and D. Tomanek (2000). Unusually High Thermal
Conductivity of Carbon Nanotubes. Physical Review Letters. 84: 4613-4616.
Bernholc, J., C. Brabec, M. B. Nardelli, A. Maiti, C. Roland and B. I. Yakobson
(1998). Theory of growth and mechanical properties of nanotubes. Applied
Physics A: Materials Science & Processing . A67: 39-46.
Bernier, P., D. Laplaze, J. Auriol, L. Barbedette, G. Flamant, M. Lebrun, A.
Brunelle and S. Della-Negra (1995). Production of fullerenes from solar energy.
Synth. Met. 70: 1455-6.
Bethune, D. S. (1996). Adding metal to carbon: production and properties of
metallofullerenes and single-layer nanotubes. NATO ASI Ser., Ser. E. 316: 165-
181.
Biercuk, M. J., M. C. Llaguno, M. Radosavljevic, J. K. Hyun, A. T. Johnson and J.
E. Fischer (2002). Carbon nanotube composites for thermal management. Los
Alamos National Laboratory, Preprint Archive, Condensed Matter: 1-12,
arXiv:cond-mat/0205418.
Blakslee, O. L., D. G. Proctor, E. J. Seldin and G. B. Spence (1970). Elastic
constants of compression-annealed pyrolytic graphite. J. Appl. Phys. 41: 3373-
82.
Blase, X., J. C. Charlier, A. De Vita, R. Car, P. Redlich, M. Terrones, W. K. Hsu,
H. Terrones, D. L. Carroll and P. M. Ajayan (1999). Boron-Mediated Growth of
Long Helicity-Selected Carbon Nanotubes. Physical Review Letters. 83: 5078-
5081.
Bockel, C., J. P. Coulomb and N. Dupont-Pavlovsky (1982). Comparison of the
adsorptive properties of Papyex and uncompressed exfoliated graphite. Surface
Science. 116: 369-79.
160
Bond, A. M., W. Miao and C. L. Raston (2000). "Mercury(II) Immobilized on
Carbon Nanotubes: Synthesis, Characterization, and Redox Properties."
Langmuir 16(14): 6004-6012.
Botton, G. A., G. Burnell, C. J. Humphreys, T. Yadav and J. C. Withers (1997).
Microstructural and electron spectroscopic characterization of carbon
nanostructures and nanotubes produced using multimetal catalysts. Journal of
Physics and Chemistry of Solids. 58: 1091-1102.
Bougrine, A., N. Dupont-Pavlovsky, J. Ghanbaja, D. Billaud and F. Beguin
(2002). Adsorption studies of a krypton film adsorbed on catalytically synthesized
multiwalled carbon nanotubes. Dependence on the nanotube morphology.
Surface Science. 506: 137-144.
Bougrine, A., K. Varlot, N. Dupont-Pavlovsky, J. Ghanbaja, D. Billaud and F.
Beguin (2001). Comparison of the adsorption properties of krypton on multi-
walled carbon nanotubes and on graphite. Materials Research Society
Symposium Proceedings. 633: A14.7.1-A14.7.6.
Boul, P. J., J. Liu, E. T. Mickelson, C. B. Huffman, L. M. Ericson, I. W. Chiang, K.
A. Smith, D. T. Colbert, R. H. Hauge, J. L. Margrave and R. E. Smalley (1999).
"Reversible sidewall functionalization of buckytubes." Chemical Physics Letters
310(3,4): 367-372.
Bower, C., R. Rosen, L. Jin, J. Han and O. Zhou (1999). Deformation of carbon
nanotubes in nanotube-polymer composites. Applied Physics Letters. 74: 3317-
3319.
Bower, C., S. Suzuki, K. Tanigaki and O. Zhou (1998). Synthesis and structure of
pristine and alkali-metal-intercalated single-walled carbon nanotubes. Applied
Physics A: Materials Science & Processing . A67: 47-52.
Brunauer, S., P. H. Emmett and E. Teller (1938). Adsorption of gases in
multimolecular layers. J. Am. Chem. Soc. 60: 309-19.
161
Brunauer, S., J. Skalny and E. E. Bodor (1969). Adsorption on nonporous solids.
Journal of Colloid and Interface Science. 30: 546-52.
Bubke, K., H. Gnewuch, M. Hempstead, J. Hammer and M. L. H. Green (1997).
Optical anisotropy of dispersed carbon nanotubes induced by an electric field.
Appl. Phys. Lett. 71: 1906-1908.
Buongiorno Nardelli, M., J. L. Fattebert, D. Orlikowski, C. Roland, Q. Zhao and J.
Bernholc (2000). Mechanical properties, defects and electronic behavior of
carbon nanotubes. Carbon. 38: 1703-1711.
Burghard, M., G. Duesberg, G. Philipp, J. Muster and S. Roth (1998). "Controlled
adsorption of carbon nanotubes on chemically modified electrode arrays."
Advanced Materials (Weinheim, Germany) 10(8): 584-588.
Cassell, A. M., J. A. Raymakers, J. Kong and H. Dai (1999). Large Scale CVD
Synthesis of Single-Walled Carbon Nanotubes. Journal of Physical Chemistry B.
103: 6484-6492.
Cassell, A. M., S. Verma, L. Delzeit, M. Meyyappan and J. Han (2001).
Combinatorial Optimization of Heterogeneous Catalysts Used in the Growth of
Carbon Nanotubes. Langmuir. 17: 260-264.
Challa, S. R., D. S. Sholl and J. K. Johnson (2001). Light isotope separation in
carbon nanotubes through quantum molecular sieving. Physical Review B:
Condensed Matter and Materials Physics. 63: 245419/1-245419/9.
Che, J., T. Cagin and W. A. Goddard, III (2000). Thermal conductivity of carbon
nanotubes. Nanotechnology. 11: 65-69.
Chen, J., M. J. Dyer and M.-F. Yu (2001). Cyclodextrin-mediated soft cutting of
single-walled carbon nanotubes. Journal of the American Chemical Society. 123:
6201-6202.
Chen, J., M. A. Hamon, H. Hu, Y. Chen, A. M. Rao, P. C. Eklund and R. C.
Haddon (1998). "Solution properties of single-walled carbon nanotubes." Science
(Washington, D. C.) 282(5386): 95-98.
162
Chen, J., H. Liu, W. A. Weimer, M. D. Halls, D. H. Waldeck and G. C. Walker
(2002). "Noncovalent Engineering of Carbon Nanotube Surfaces by Rigid,
Functional Conjugated Polymers." Journal of the American Chemical Society
124(31): 9034-9035.
Chen, J., A. M. Rao, S. Lyuksyutov, M. E. Itkis, M. A. Hamon, H. Hu, R. W. Cohn,
P. C. Eklund, D. T. Colbert, R. E. Smalley and R. C. Haddon (2001). "Dissolution
of Full-Length Single-Walled Carbon Nanotubes." Journal of Physical Chemistry
B 105(13): 2525-2528.
Chen, Q., L. Dai, M. Gao, S. Huang and A. Mau (2001). "Plasma activation of
carbon nanotubes for chemical modification." Journal of Physical Chemistry B
105(3): 618-622.
Chen, X. Q., T. Saito, H. Yamada and K. Matsushige (2001). Aligning single -wall
carbon nanotubes with an alternating-current electric field. Applied Physics
Letters. 78: 3714-3716.
Chen, Y., L. T. Chadderton, J. S. Williams and J. F. Gerald (2000). Solid-state
formation of carbon and boron nitride nanotubes. Materials Science Forum. 343-
346: 63-67.
Chen, Y., J. Fitz Gerald, L. T. Chadderton and L. Chaffron (1999). Investigation
of nanoporous carbon powders produced by high energy ball milling and
formation of carbon nanotubes during subsequent annealing. Materials Science
Forum. 312-314: 375-379.
Chen, Y., J. Fitz Gerald, J. S. Williams and S. Bulcock (1999). Synthesis of boron
nitride nanotubes at low temperatures using reactive ball milling. Chemical
Physics Letters. 299: 260-264.
Chernozatonskii, L. A., E. A. Fedorov, Z. J. Kosakovskaya, V. I. Panov and S. V.
Savinov (1993). STM evidence of smallest rod presence in nanofilament carbon
structure. Pis'ma v Zhurnal Eksperimental'noi i Teoreticheskoi Fiziki. 57: 35-8.
163
Chiang, I. W., E. T. Michelson, P. J. Boul, R. H. Hauge, R. E. Smalley and J. L.
Margrave (2000). "Fluorination, defluorination, and derivatization of single -wall
carbon nanotubes." Abstr. Pap. - Am. Chem. Soc. 220th: IEC-153.
Chibante, L. P. F., A. Thess, J. M. Alford, M. D. Diener and R. E. Smalley (1993).
Solar generation of the fullerenes. J. Phys. Chem. 97: 8696-700.
Chirnside, G. C. and C. G. Pope (1964). Limiting isosteric heats of adsorption of
n-butane, n-pentane, and n-hexane on graphitized carbon black. J. Phys. Chem.
68: 2377-9.
Cho, W.-S., E. Hamada, Y. Kondo and K. Takayanagi (1996). Synthesis of
carbon nanotubes from bulk polymer. Appl. Phys. Lett. 69: 278-279.
Coleman, J. N., S. Curran, A. B. Dalton, A. P. Davey, B. Mc Carthy, W. Blau and
R. C. Barklie (1999). Physical doping of a conjugated polymer with carbon
nanotubes. Synthetic Metals . 102: 1174-1175.
Collins, P. G., M. Hersam, M. Arnold, R. Martel and P. Avouris (2001). Current
Saturation and Electrical Breakdown in Multiwalled Carbon Nanotubes. Physical
Review Letters. 86: 3128-3131.
Colomer, J. F., P. Piedigrosso, K. Mukhopadhyay, Z. Konya, I. Willems, A.
Fonseca and J. B. Nagy (1998). "Purification of multi-wall carbon nanotubes
produced over supported catalysts." Proceedings - Electrochemical Society 98-
8(Recent Advances in the Chemistry and Physics of Fullerenes and Related
Materials): 830-842.
Constabaris, G., J. R. Sams, Jr. and G. D. Halsey, Jr. (1961). The interaction of
H2, D2, CH4, and CD4 with graphitized carbon black. Journal of Physical
Chemistry. 65: 367-9.
Constabaris, G., J. R. Sams, Jr. and G. D. Halsey, Jr. (1962). Adsorption
isotherms of neon, argon, krypton, xenon, hydrogen, deuterium, methane, and
tetradeuteriomethane on highly graphitized carbon black P33 (2700.degree.).
Journal of Chemical Physics. 37: 915.
164
Corbin, D. R. and D. W. Reutter (1997). Adsorptive purification of
pentafluoroethane from its mixtures with chloropentafluoroethane using carbon
nanotubes. U.S. US, (E. I. Du Pont De Nemours and Company, USA). 4 pp.
Cracknell, R. F. and D. Nicholson (1994). Grand canonical Monte Carlo study of
Lennard-Jones mixtures in slit pores. Journal of the Chemical Society, Faraday
Transactions. 90: 1487-93.
Cracknell, R. F., D. Nicholson and K. E. Gubbins (1995). Molecular dynamics
study of the self-diffusion of supercritical methane in silt-shaped graphitic
micropores. Journal of the Chemical Society, Faraday Transactions. 91: 1377-
83.
Cui, S., C. Z. Lu, Y. L. Qiao and L. Cui (1999). Large-scale preparation of carbon
nanotubes by nickel catalyzed decomposition of methane at 600.degree.C.
Carbon. 37: 2070-2073.
Curran, S. A., P. M. Ajayan, W. J. Blau, D. L. Carroll, J. N. Coleman, A. B.
Dalton, A. P. Davey, A. Drury, B. McCarthy, S. Maier and A. Strevens (1998). A
composite from poly(m-phenylenevinylene-co-2,5-dioctoxy-p-phenylenevinylene)
and carbon nanotubes. A novel material for molecular optoelectronics. Advanced
Materials (Weinheim, Germany). 10: 1091-1093.
Dali, Q. (2001). Multiwalled carbon nanotube CVD synthesis, modification, and
composite applications. Chemical and Materials Engineering. Lexington,
University of Kentucky: 182.
Daubert, T. E. D., R. P.; Sibul, H. M.; Stebbins, C. C., Ed. (1989). Design Institute
for Phys. Prop. Data A.I.Ch.E. DIPPR. New York, N.Y, Taylor & Francis.
de Heer, W. A., W. S. Bacsa, A. Chatelain, T. Gerfin, R. Humphrey-Baker, L.
Forro and D. Ugarte (1995). Aligned carbon nanotube films: production and
optical and electronic properties. Science (Washington, D. C.). 268: 845-7.
165
Dillon, A. C., K. M. Jones, T. A. Bekkedahl, C. H. Kiang, D. S. Bethune and M. J.
Heben (1997). Storage of hydrogen in single-walled carbon nanotubes. Nature
(London) . 386: 377-379.
Dresselhaus, M. S. e. a. (1988). Graphite Fibers and Filaments. New York,
Springer-Verlag.
Duesberg, G. S., M. Burghard, J. Muster, G. Philipp and S. Roth (1998).
"Separation of carbon nanotubes by size exclusion chromatography." Chemical
Communications (Cambridge)(3): 435-436.
Duesberg, G. S., J. Muster, V. Krstic, M. Burghard and S. Roth (1998).
"Chromatographic size separation of single-wall carbon nanotubes." Applied
Physics A: Materials Science & Processing A67(1): 117-119.
Dutta, A. L., A. K. Misra and R. P. Singh (1992). Electrical conduction and
ultrasound wave propagation in particulate and short fiber composites of PVC. II:
PVC-aluminum coated glass fiber composites. J. Vinyl Technol. 14: 93-100.
Ebbesen, T. W. (1994). Carbon nanotubes. Annu. Rev. Mater. Sci. 24: 235-64.
Ebbesen, T. W. and P. M. Ajayan (1992). Large-scale synthesis of carbon
nanotubes. Nature (London) . 358: 220-2.
Elgin, R. L. and D. L. Goodstein (1974). Thermodynamic functions for helium-4
submonolayers. Low Temp. Phys.-LT 13, Proc. Int. Conf. Low Temp. Phys., 13th.
2: 175-9.
Endo, M., Y. A. Kim, T. Matusita and T. Hayashi (2001). From vapor-grown
carbon fibers (VGCFs) to carbon nanotubes. NATO Science Series, Series E:
Applied Sciences. 372: 51-61.
Esumi, K., M. Ishigami, A. Nakajima, K. Sawada and H. Honda (1996). Chemical
treatment of carbon nanotubes. Carbon. 34: 279-81.
Esumi, K., M. Ishigami, A. Nakajima, K. Sawada and H. Honda (1996). "Chemical
treatment of carbon nanotubes." Carbon 34(2): 279-81.
166
Eswaramoorthy, M., R. Sen and C. N. R. Rao (1999). "A study of micropores in
single-walled carbon nanotubes by the adsorption of gases and vapors."
Chemical Physics Letters 304(3,4): 207-210.
Etters, R. D., B. Kuchta and J. Belak (1993). Vacancy-induced melting of
molecular nitrogen adlayers on graphite. Physical Review Letters. 70: 826-9.
Etters, R. D., M. W. Roth and B. Kuchta (1990). Calculated melting behavior of
partial and complete molecular nitrogen monolayers deposited on graphite.
Physical Review Letters. 65: 3140-3.
Fields, C. L., J. R. Pitts, M. J. Hale, C. Bingham, A. Lewandowski and D. E. King
(1993). Formation of fullerenes in highly concentrated solar flux. J. Phys. Chem.
97: 8701-2.
Fields, C. L. P., J.R.; Mischler, D.; Bingham, C.; Lewandowski, A.; Schultz, D.L.;
Bekkedahl, T.A.; Jones, K.M.; Heben, M.J. (1996). Proceedings of the 8th
International Symposium on Solar Thermal Concentrating Technologies, Köln,
Germany.
Fishbine, B. H. (1996). "Carbon Nanotube Alignment and Manipulation using
Electrostatic Fields." Fullerene Science & Technology 4(1): 87-100.
Flahaut, E., A. Peigney, C. Laurent, C. Marliere, F. Chastel and A. Rousset
(2000). Carbon nanotube-metal-oxide nanocomposites: microstructure, electrical
conductivity and mechanical properties. Acta Materialia. 48: 3803-3812.
Franklin, N. R. and H. Dai (2000). An enhanced CVD approach to extensive
nanotube networks with directionality. Advanced Materials (Weinheim,
Germany) . 12: 890-894.
Gale, R. L. and R. A. Beebe (1964). Determination of heats of adsorption on
carbon black and bone mineral by chromatography using the eluted pulse
technique. Journal of Physical Chemistry. 68: 555-67.
Gao, B., C. Bower, J. D. Lorentzen, L. Fleming, A. Kleinhammes, X. P. Tang, L.
E. McNeil, Y. Wu and O. Zhou (2000). Enhanced saturation lithium composition
167
in ball-milled single-walled carbon nanotubes. Chemical Physics Letters. 327: 69-
75.
Gao, G., T. Cagin and W. A. Goddard, III (1998). Energetics, structure,
mechanical and vibrational properties of single-walled carbon nanotubes.
Nanotechnology. 9: 184-191.
Gao, R., Z. L. Wang and S. Fan (2000). Kinetically Controlled Growth of Helical
and Zigzag Shapes of Carbon Nanotubes. Journal of Physical Chemistry B. 104:
1227-1234.
Garamus, V. M. and J. S. Pedersen (1998). "A small-angle neutron scattering
study of the structure of graphitized carbon black aggregates in Triton X-
100/water solutions." Colloids and Surfaces, A: Physicochemical and
Engineering Aspects 132(2-3): 203-212.
Garg, A. and S. B. Sinnott (1998). Effect of chemical functionalization on the
mechanical properties of carbon nanotubes. Chemical Physics Letters. 295: 273-
278.
Garza, R. L., K.; Guerrero, C. (1993). Mmemorias de Compositos 1993. Saltillo
Coahuila ,Mexico.
Georgakilas, V., K. Kordatos, M. Prato, D. M. Guldi, M. Holzinger and A. Hirsch
(2002). "Organic Functionalization of Carbon Nanotubes." Journal of the
American Chemical Society 124(5): 760-761.
Gong, X., J. Liu, S. Baskaran, R. D. Voise and J. S. Young (2000). Surfactant-
Assisted Processing of Carbon Nanotube/Polymer Composites. Chemistry of
Materials. 12: 1049-1052.
Gonzalez-Garcia, C. M., M. L. Gonzalez-Martin, V. Gomez-Serrano, J. M. Bruque
and L. Labajos-Broncano (2000). "Determination of the free energy of adsorption
on carbon blacks of a nonionic surfactant from aqueous solutions." Langmuir
16(8): 3950-3956.
168
Goze, C., L. Vaccarini, L. Henrard, P. Bernier, E. Hernandez and A. Rubio
(1999). Elastic and mechanical properties of carbon nanotubes. Synthetic Metals.
103: 2500-2501.
Gregg, S. J. and K. S. W. Sing (1982). Adsorption, Surface Area and Porosity.
New York, NY, Academic Press Inc.
Grobert, N., M. Terrones, S. Trasobares, K. Kordatos, H. Terrones, J. Olivares, J.
P. Zhang, P. Redlich, W. K. Hsu, C. L. Reeves, D. J. Wallis, Y. Q. Zhu, J. P.
Hare, A. J. Pidduck, H. W. Kroto and D. R. M. Walton (2000). A novel route to
aligned nanotubes and nanofibers using laser-patterned catalytic substrates.
Applied Physics A: Materials Science & Processing . 70: 175-183.
Gu, C., G.-H. Gao, Y.-X. Yu and T. Nitta (2002). Simulation for separation of
hydrogen and carbon monoxide by adsorption on single-walled carbon
nanotubes. Fluid Phase Equilibria. 194-197: 297-307.
Gu, Z., H. Peng, J. L. Zimmerman, I. W. Chiang, V. N. Khabashesku, R. H.
Hauge and J. L. Margrave (2002). "Fluorination of polymeric C60 and carbon
nanotubes: A starting point for various chemical modifications of nanostructured
carbon materials." Abstracts of Papers, 223rd ACS National Meeting, Orlando,
FL, United States, April 7-11, 2002: FLUO-012.
Guerret-Piecourt, C., Y. Le Bouar, A. Loiseau and H. Pascard (1994). Relation
between metal electronic structure and morphology of metal compounds inside
carbon nanotubes. Nature (London). 372: 761-5.
Guo, T., P. Nikolaev, A. G. Rinzler, D. Tomanek, D. T. Colbert and R. E. Smalley
(1995). Self-Assembly of Tubular Fullerenes. J. Phys. Chem. 99: 10694-7.
Guo, T., P. Nikolaev, A. Thess, D. T. Colbert and R. E. Smalley (1995). Catalytic
growth of single -walled nanotubes by laser vaporization. Chem. Phys. Lett. 243:
49-54.
169
Hafner, J. H., M. J. Bronikowski, B. R. Azamian, P. Nikolaev, A. G. Rinzler, D. T.
Colbert, K. A. Smith and R. E. Smalley (1998). Catalytic growth of single -wall
carbon nanotubes from metal particles. Chemical Physics Letters. 296: 195-202.
Haggenmueller, R., H. H. Gommans, A. G. Rinzler, J. E. Fischer and K. I. Winey
(2000). Aligned single -wall carbon nanotubes in composites by melt processing
methods. Chemical Physics Letters. 330: 219-225.
Haluska, M., M. Hulman, M. Hirscher, M. Becher, S. Roth, I. Stepanek and P.
Bernier (2001). Hydrogen storage in mechanically treated single wall carbon
nanotubes. AIP Conference Proceedings. 591: 603-608.
Hamilton, J. J. and D. L. Goodstein (1983). Thermodynamic study of methane
multilayers adsorbed on graphite. Physical Review B: Condensed Matter and
Materials Physics. 28: 3838-48.
Hammond, W. R. and S. D. Mahanti (1990). Polar pyramidal molecules on a
graphite substrate. Surface Science. 234: 308-14.
Hansen, F. Y., J. C. Newton and H. Taub (1993). Molecular-dynamics studies of
the melting of butane and hexane monolayers adsorbed on the basal-plane
surface of graphite. Journal of Chemical Physics. 98: 4128-41.
Hansen, F. Y. and H. Taub (1992). Melting mechanism in monolayers of flexible
rod-shaped molecules. Physical Review Letters. 69: 652-5.
Heben, M. J. B., T.A.; Schultz,D.L.; Jones,K.M.; Dillon,A.C.; Curtis,C.J.;
Bingham,C.; Pitts,J.R.; Lewandowski,A. ; Fields C.L. (1996). Proceedings of the
Symposium on Recent Advances in the Chemistry and Physics of Fullerenes and
Related Materials, The Electrochemical Society Inc.
Hernandez, E., P. Ordejon, I. Boustani, A. Rubio and J. A. Alonso (2000). Tight
binding molecular dynamics studies of boron assisted nanotube growth. Journal
of Chemical Physics. 113: 3814-3821.
170
Herwig, K. W., J. C. Newton and H. Taub (1994). Structure and growth of butane
films adsorbed on graphite. Physical Review B: Condensed Matter. 50: 15287-
97.
Herwig, K. W., Z. Wu, P. Dai, H. Taub and F. Y. Hansen (1997). "Quasielastic
neutron scattering and molecular dynamics simulation studies of the melting
transition in butane and hexane monolayers adsorbed on graphite." Journal of
Chemical Physics 107(13): 5186-5196.
Hilding, J., E. A. Grulke, S. B. Sinnott, D. Qian, R. Andrews and M. Jagtoyen
(2001). Sorption of butane on carbon multiwall nanotubes at room temperature.
Langmuir. 17: 7540-7544.
Hilding, J., E. A. Grulke, Z. G. Zhang and F. Lockwood (2003). Dispersion of
carbon nanotubes in liquids. Journal of Dispersion Science and Technology. 24:
1-41.
Holland, C. E., S. A. Al-Muhtaseb and J. A. Ritter (2001). Adsorption of C1-C7
Normal Alkanes on BAX Activated Carbon. 1. Potential Theory Correlation and
Adsorbent Characterization. Industrial & Engineering Chemistry Research. 40:
338-346.
Holzinger, M., O. Vostrowsky, A. Hirsch, F. Hennrich, M. Kappes, R. Weiss and
F. Jellen (2001). "Sidewall functionalization of carbon nanotubes." Angewandte
Chemie, International Edition 40(21): 4002-4005.
Hone, J., M. C. Llaguno, N. M. Nemes, A. T. Johnson, J. E. Fischer, D. A.
Walters, M. J. Casavant, J. Schmidt and R. E. Smalley (2000). Electrical and
thermal transport properties of magnetically aligned single wall carbon nanotube
films. Applied Physics Letters. 77: 666-668.
Hone, J., M. Whitney, C. Piskoti and A. Zettl (1999). Thermal conductivity of
single-walled carbon nanotubes. Physical Review B: Condensed Matter and
Materials Physics. 59: R2514-R2516.
171
Hoory, S. E. and J. M. Prausnitz (1967). Adsorption of hydrocarbons on
graphitized carbon. Transactions of the Faraday Society. 63: 455-60.
Hsu, W. K., J. P. Hare, M. Terrones, H. W. Kroto, D. R. M. Walton and P. J. F.
Harris (1995). Condensed-phase nanotubes. Nature (London). 377: 687.
Hsu, W. K., M. Terrones, J. P. Hare, H. Terrones, H. W. Kroto and D. R. M.
Walton (1996). Electrolytic formation of carbon nanostructures. Chemical Physics
Letters. 262: 161-166.
http://www.zyvex.com/nanotech/talks/XeroxPARC980312/sld034.htm.
Huang, H., H. Kajiura, A. Yamada and M. Ata (2002). "Purification and alignment
of arc-synthesis single-walled carbon nanotube bundles." Chemical Physics
Letters 356(5,6): 567-572.
Hulicova, D., K. Hosoi, S.-I. Kuroda, H. Abe and A. Oya (2002). Carbon
nanotubes prepared by spinning and carbonizing fine core-shell polymer
microspheres. Advanced Materials (Weinheim, Germany). 14: 452-455.
Hunter, R. J. (1986). Foundations of Colloid Science, Vol. 1.
Hwang, G. L. and K. C. Hwang (2001). Breakage, fusion, and healing of carbon
nanotubes. Nano Letters. 1: 435-438.
Hwang, K. C. (2000). "Synthesis of SiO2-carbon nanotubes micro rods and their
applications in reinforcements of ceramics." Abstr. Pap. - Am. Chem. Soc. 220th:
PHYS-387.
Iijima, S. (1991). Helical microtubules of graphitic carbon. Nature (London) . 354:
56-8.
Inaba, A. and J. A. Morrison (1986). Ethylene on graphite: heats of adsorption
and phase diagram. Physical Review B: Condensed Matter and Materials
Physics. 34: 3238-42.
172
Inaba, A. and J. A. Morrison (1986). The wetting transition and
adsorption/desorption hysteresis for the methane/graphite system. Chemical
Physics Letters. 124: 361-4.
Inoue, S., N. Ichikuni, T. Suzuki, T. Uematsu and K. Kaneko (1998). Capillary
Condensation of N2 on Multiwall Carbon Nanotubes. Journal of Physical
Chemistry B. 102: 4689-4692.
Ito, F., K. Konuma and A. Okamoto (2001). Electron emission from single-walled
carbon nanotubes with sharpened bundles. Journal of Applied Physics. 89: 8141-
8145.
Iwasaki, T., T. Motoi and T. Den (1999). Multiwalled carbon nanotubes growth in
anodic alumina nanoholes. Applied Physics Letters. 75: 2044-2046.
Jacobsen, R. L., T. M. Tritt, J. R. Guth, A. C. Ehrlich and D. J. Gillespie (1995).
Mechanical properties of vapor-grown carbon fiber. Carbon. 33: 1217-21.
Jacques, D., S. Villain, A. M. Rao, R. Andrews, F. Derbyshire, E. C. Dickey and
D. Qian (2000). Synthesis of multiwalled carbon nanotubes. Materials Research
Society Symposium Proceedings. 593: 15-20.
Jagtoyen, M. and F. Derbyshire (1993). "Some considerations of the origins of
porosity in carbons from chemically activated wood." Carbon 31(7): 1185-92.
Jagtoyen, M., F. J. Derbyshire and E. Grulke (2000). Separation of gas mixtures
using multiwalled nanotubes. Abstr. Pap. - Am. Chem. Soc. 220th: FUEL-106.
Jakob, P., M. Gsell and D. Menzel (2001). Interactions of adsorbates with locally
strained substrate lattices. Journal of Chemical Physics. 114: 10075-10085.
Jeong, S.-H., O.-J. Lee, K.-H. Lee, S. H. Oh and C.-G. Park (2002). Preparation
of Aligned Carbon Nanotubes with Prescribed Dimensions: Template Synthesis
and Sonication Cutting Approach. Chemistry of Materials . 14: 1859-1862.
173
Jeong, T., T. H. Kim, W.-Y. Kim, K.-H. Lee and Y.-B. Hahn (2002). "High yield
purification of carbon nanotubes with H2S-O2 mixture." Korean Journal of
Chemical Engineering 19(3): 519-523.
Jeong, T., W.-Y. Kim and Y.-B. Hahn (2001). "A new purification method of
single-wall carbon nanotubes using H2S and O2 mixture gas." Chemical Physics
Letters 344(1,2): 18-22.
Jia, Z., Z. Wang, J. Liang, B. Wei and D. Wu (1999). Production of short multi-
walled carbon nanotubes. Carbon. 37: 903-906.
Jia, Z., Z. Wang, C. Xu, J. Liang, B. Wei, D. Wu and S. Zhu (1999). Study on
poly(methyl methacrylate)/carbon nanotube composites. Materials Science &
Engineering, A: Structural Materials: Properties, Microstructure and Processing.
A271: 395-400.
Jin, L., C. Bower and O. Zhou (1998). Alignment of carbon nanotubes in a
polymer matrix by mechanical stretching. Applied Physics Letters. 73: 1197-
1199.
Jin, Z., K. P. Pramoda, S. H. Goh and G. Xu (2002). Poly(vinylidene fluoride)-
assisted melt-blending of multi-walled carbon nanotube/poly(methyl
methacrylate) composites. Materials Research Bulletin. 37: 271-278.
Jin, Z., K. P. Pramoda, G. Xu and S. H. Goh (2001). Dynamic mechanical
behavior of melt-processed multi-walled carbon nanotube/poly(methyl
methacrylate) composites. Chemical Physics Letters. 337: 43-47.
Journet, C., W. K. Maser, P. Bernier, A. Loiseau, M. Lamy de la Chapells, S.
Lefrant, P. Deniard, R. Lee and J. E. Fischer (1997). Large-scale production of
single-walled carbon nanotubes by the electric-arc technique. Nature (London).
388: 756-758.
Kamalakaran, R., M. Terrones, T. Seeger, P. Kohler-Redlich, M. Ruhle, Y. A.
Kim, T. Hayashi and M. Endo (2000). Synthesis of thick and crystalline nanotube
arrays by spray pyrolysis. Applied Physics Letters. 77: 3385-3387.
174
Kaneto, K., M. Tsuruta, G. Sakai, W. Y. Cho and Y. Ando (1999). Electrical
conductivities of multi-wall carbon nanotubes. Synthetic Metals. 103: 2543-2546.
Karicherla, A. (2001). Shear dependent and time dependent behavior of carbon
multiwall nanotube suspensions. Chemical & Materials Engineering. Lexington,
KY, University of Kentucky.
Katoh, R., Y. Tasaka, E. Sekreta, M. Yumura, F. Ikazaki, Y. Kakudate and S.
Fujiwara (1999). Sonochemical production of a carbon nanotube. Ultrasonics
Sonochemistry. 6: 185-187.
Kelly, B. T. (1981). Physics of Graphite: 475 pp.
Kiang, C.-H. and W. A. Goddard, III (1996). Polyyne ring nucleus growth model
for single-layer carbon nanotubes. Phys. Rev. Lett. 76: 2515-18.
Kiang, C.-H., W. A. Goddard, III, R. Beyers and D. S. Bethune (1995). Carbon
nanotubes with single-layer walls. Carbon. 33: 903-14.
Kiang, C.-H., W. A. Goddard, III, R. Beyers, J. R. Salem and D. S. Bethune
(1994). Catalytic Synthesis of Single-Layer Carbon Nanotubes with a Wide
Range of Diameters. J. Phys. Chem. 98: 6612-18.
Kiang, C.-H., W. A. Goddard, III, R. Beyers, J. R. Salem and D. S. Bethune
(1996). Catalytic effects of heavy metals on the growth of carbon nanotubes and
nanoparticles. J. Phys. Chem. Solids. 57: 35-9.
Kim, J., K. No and C. J. Lee (2001). Growth and field emission of carbon
nanotubes on electroplated Ni catalyst coated on glass substrates. Journal of
Applied Physics. 90: 2591-2594.
Kim, M. S., N. M. Rodriguez and R. T. K. Baker (1991). Interaction of
hydrocarbons with copper-nickel and nickel in the formation of carbon filaments.
J. Catal. 131: 60-73.
175
Kim, P., L. Shi, A. Majumdar and P. L. McEuen (2001). Thermal Transport
Measurements of Individual Multiwalled Nanotubes. Physical Review Letters. 87:
215502/1-215502/4.
Kim, Y. A., T. Hayashi, Y. Fukai, M. Endo, T. Yanagisawa and M. S. Dresselhaus
(2002). Effect of ball milling on morphology of cup-stacked carbon nanotubes.
Chemical Physics Letters. 355: 279-284.
Kiselev, A. V. (1961). Surface chemistry, adsorption energy, and adsorption
equilibriums. Quart. Rev. (London). 15: 99-124.
Kleiner, A. and S. Eggert (2001). Curvature, hybridization, and STM images of
carbon nanotubes. Physical Review B: Condensed Matter and Materials Physics.
64: 113402/1-113402/4.
Knupfer, M., T. Pichler, M. S. Golden, J. Fink, A. Rinzler and R. E. Smalley
(1999). Electron energy-loss spectroscopy studies of single wall carbon
nanotubes. Carbon. 37: 733-738.
Kooi, S. E., U. Schlecht, M. Burghard and K. Kern (2002). "Electrochemical
modification of single carbon nanotubes." Angewandte Chemie, International
Edition 41(8): 1353-1355.
Koshio, A., M. Yudasaka and S. Iijima (2001). Thermal degradation of ragged
single-wall carbon nanotubes produced by polymer-assisted ultrasonication.
Chemical Physics Letters. 341: 461-466.
Koshio, A., M. Yudasaka, M. Zhang and S. Iijima (2001). A Simple Way to
Chemically React Single-Wall Carbon Nanotubes with Organic Materials Using
Ultrasonication. Nano Letters. 1: 361-363.
Krishnan, A., E. Dujardin, T. W. Ebbesen, P. N. Yianilos and M. M. J. Treacy
(1998). Young's modulus of single-walled nanotubes. Physical Review B:
Condensed Matter and Materials Physics. 58: 14013-14019.
176
Krstic, V., G. S. Duesberg, J. Muster, M. Burghard and S. Roth (1998).
"Langmuir-Blodgett Films of Matrix-Diluted Single-Walled Carbon Nanotubes."
Chemistry of Materials 10(9): 2338-2340.
Krstic, V., J. Muster, G. S. Duesberg, G. Philipp, M. Burghard and S. Roth
(2000). "Electrical transport in single-walled carbon nanotube bundles embedded
in Langmuir-Blodgett monolayers." Synthetic Metals 110(3): 245-249.
Kumar, T. P. (1995). Lithium Batteries: New Materials, Developments and
Perspectives, Edited by G. Pistoia. Bull. Electrochem. 11: 167-8.
Kumar, V. and Y. Chandra (1989). Characterization of epoxy resin system and
dispersion of carbon black powder. J. Polym. Mater. 6: 51-6.
Kuznetsova, A., I. Popova, J. T. Yates, Jr., M. J. Bronikowski, C. B. Huffman, J.
Liu, R. E. Smalley, H. H. Hwu and J. G. Chen (2001). "Oxygen-Containing
Functional Groups on Single-Wall Carbon Nanotubes: NEXAFS and Vibrational
Spectroscopic Studies." Journal of the American Chemical Society 123(43):
10699-10704.
Kuznetsova, A., J. T. Yates, Jr., J. Liu and R. E. Smalley (2000). Physical
adsorption of xenon in open single walled carbon nanotubes. Observation of a
quasi-one-dimensional confined Xe phase. Journal of Chemical Physics. 112:
9590-9598.
Kyotani, T., S. Nakazaki, W.-H. Xu and A. Tomita (2001). "Chemical modification
of the inner walls of carbon nanotubes by HNO3 oxidation." Carbon 39(5): 782-
785.
Laplaze, D., P. Bernier, L. Barbedette, J. M. Lambert, G. Flamant, M. Lebrun, A.
Brunelle and S. Della-Negra (1994). Production of fullerenes from solar energy:
the Odeillo experiment. C. R. Acad. Sci., Ser. II: Mec., Phys., Chim., Astron. 318:
733-8.
Laplaze, D., P. Bernier, G. Flamant, M. Lebrun, A. Brunelle and S. Della -Negra
(1996). Preparation of fullerenes using a solar furnace. Synth. Met. 77: 67-71.
177
Laplaze, D., P. Bernier, C. Journet, J. L. Sauvajol, D. Bormann, G. Flamant and
M. Lebrun (1997). The use of solar energy for the production of fullerenes and
porous silicon. Journal de Physique III. 7: 463-472.
Laplaze, D., P. Bernier, C. Journet, V. Vie, G. Flamant and M. Lebrun (1997).
Carbon sublimation using a solar furnace. Synthetic Metals . 86: 2295-2296.
Laplaze, D. B. J., C.; Vie, V.; Flamant, G.; Philippot, E.; Lebrun, M. (1996).
Proceedings of the 8th International Symposium on Solar Thermal Concentrating
Technologies, Köln, Germany.
Lee, C. J., J. Park, J. M. Kim, Y. Huh, J. Y. Lee and K. S. No (2000). Low-
temperature growth of carbon nanotubes by thermal chemical vapor deposition
using Pd, Cr, and Pt as co-catalyst. Chemical Physics Letters. 327: 277-283.
Lee, C. J., J. H. Park and J. Park (2000). Synthesis of bamboo-shaped
multiwalled carbon nanotubes using thermal chemical vapor deposition.
Chemical Physics Letters. 323: 560-565.
Lee, C.-J. and J.-E. Yoo (2000). Acidic gas phase for purification of carbon
nanotubes by heat treatment in a tube furnace. Eur. Pat. Appl. Ep, (Iljin Nanotech
Co., Ltd., S. Korea). 8 pp.
Lee, H., Y.-S. Kang, S.-H. Kim and J.-Y. Lee (2002). Hydrogen desorption
properties of multiwall carbon nanotubes with closed and open structures.
Applied Physics Letters. 80: 577-579.
Lee, H.-J. and H. Lee (2002). "Vertical alignment of chemically modified single-
walled carbon nanotubes on self-assembled monolayers containing an amine
group." Abstracts of Papers, 223rd ACS National Meeting, Orlando, FL, United
States, April 7-11, 2002: COLL-317.
Lee, R. S., H. J. Kim, J. E. Fischer, A. Thess and R. E. Smalley (1997).
Conductivity-enhancement in single -walled carbon nanotube bundles doped with
K and Br. Nature (London). 388: 255-257.
178
Lewenstein, J. C., T. P. Burgin, A. Ribayrol, L. A. Nagahara and R. K. Tsui
(2002). "High-Yield Selective Placement of Carbon Nanotubes on Pre-Patterned
Electrodes." Nano Letters 2(5): 443-446.
Li, Y. B., B. Q. Wei, J. Liang, Q. Yu and D. H. Wu (1999). Transformation of
carbon nanotubes to nanoparticles by ball milling process. Carbon. 37: 493-497.
Li, Y. L., Y. D. Yu and Y. Liang (1997). A novel method for synthesis of carbon
nanotubes: low temp. solid pyrolysis. Journal of Materials Research. 12: 1678-
1680.
Lin, X., X. K. Wang, V. P. Dravid, R. P. H. Chang and J. B. Ketterson (1994).
Large scale synthesis of single-shell carbon nanotubes. Appl. Phys. Lett. 64:
181-3.
Liu, C., Y. Y. Fan, M. Liu, H. T. Cong, H. M. Cheng and M. S. Dresselhaus
(1999). Hydrogen storage in single-walled carbon nanotubes at room
temperature. Science (Washington, D. C.). 286: 1127-1129.
Liu, J., A. G. Rinzler, H. Dai, J. H. Hafner, R. K. Bradley, P. J. Boul, A. Lu, T.
Iverson, K. Shelimov, C. B. Huffman, F. Rodriguez-Macias, Y.-S. Shon, T. R.
Lee, D. T. Colbert and R. E. Smalley (1998). "Fullerene pipes." Science
(Washington, D. C.) 280(5367): 1253-1256.
Loiseau, A. and H. Pascard (1996). Synthesis of long carbon nanotubes filled
with Se, S, Sb and Ge by the arc method. Chem. Phys. Lett. 256: 246-252.
Lozano, K., J. Bonilla-Rios and E. V. Barrera (2001). A study on nanofiber-
reinforced thermoplastic composites (II): investigation of the mixing rheology and
conduction properties. Journal of Applied Polymer Science. 80: 1162-1172.
Lu, J. P. (1997). Elastic Properties of Carbon Nanotubes and Nanoropes.
Physical Review Letters. 79: 1297-1300.
Lu, K. L., R. M. lago, Y. K. Chen, M. L. H. Green, P. J. F. Harris and S. C. Tsang
(1996). Mechanical damage of carbon nanotubes by ultrasound. Carbon. 34:
814-816.
179
Lysek, M. J., M. LaMadrid, P. Day and D. Goodstein (1992). Adsorption
hysteresis, capillary condensation and melting in multilayer methane films on
graphite foam. Langmuir. 8: 898-900.
Lysek, M. J., M. A. LaMadrid, P. K. Day and D. L. Goodstein (1993). Heat
capacity of multilayer methane on graphite: phase transitions in the first four
layers. Physical Review B: Condensed Matter and Materials Physics. 47: 7389-
400.
Ma, C. and Y. Xia (1992). "Mixed adsorption of sodium dodecyl sulfate and
ethoxylated nonylphenols on carbon black and the stability of carbon black
dispersions in mixed solutions of sodium dodecyl sulfate and ethoxylated
nonylphenols." Colloids Surf. 66(3): 215-21.
Mackie, E. B., K. Lafdi and A. D. Migone (1997). Adsorption Isotherm Studies on
Catalytic Carbon Nanotubes. APS March Meeting.
Mackie, E. B., R. A. Wolfson, L. M. Arnold, K. Lafdi and A. D. Migone (1997).
Adsorption Studies of Methane Films on Catalytic Carbon Nanotubes and on
Carbon Filaments. Langmuir. 13: 7197-7201.
Mangun, C. L., M. A. Daley, R. D. Braatz and J. Economy (1998). Effect of pore
size on adsorption of hydrocarbons in phenolic-based activated carbon fibers.
Carbon. 36: 123-129.
Mann, D. J. and M. D. Halls (2002). Ab initio simulations of oxygen atom insertion
and substitutional doping of carbon nanotubes. Journal of Chemical Physics.
116: 9014-9020.
Margrave, J. L., E. T. Mickelson, I. W. Chiang, J. L. Zimmerman, P. J. Boul, J.
Lozano, J. Liu, R. E. Smalley and R. H. Hauge (1999). "Solvation of fluorinated
single-wall carbon nanotubes in alcohol solvents." Book of Abstracts, 218th ACS
National Meeting, New Orleans, Aug. 22-26: FLUO-013.
Margrave, J. L., E. T. Mickelson, R. Hauge, P. Boul, C. B. Huffman, J. Liu, R. E.
Smalley, K. Smith and D. T. Colbert (2000). Chemical derivatization of single-wall
180
carbon nanotubes to facilitate solvation for carbon fiber synthesis. PCT Int. Appl.
Wo, (William Marsh Rice University, USA). 76 pp.
Martinez, M. T., M. A. Callejas, A. M. Benito, W. K. Maser, M. Cochet, J. M.
Andres, J. Schreiber, O. Chauvet and J. L. G. Fierro "Microwave single walled
carbon nanotubes purification." Chem Commun (Camb) FIELD Publication
Date:2002 May 7(9): 1000-1. FIELD Reference Number: FIELD Journal
Code:9610838 FIELD Call Number:.
Masenelli-Varlot, K., E. McRae and N. Dupont-Pavlovsky (2002). Comparative
adsorption of simple molecules on carbon nanotubes Dependence of the
adsorption properties on the nanotube morphology. Applied Surface Science.
196: 209-215.
Mawhinney, D. B., V. Naumenko, A. Kuznetsova, J. T. Yates, Jr., J. Liu and R. E.
Smalley (2000). "Infrared Spectral Evidence for the Etching of Carbon
Nanotubes: Ozone Oxidation at 298 K." Journal of the American Chemical
Society 122(10): 2383-2384.
Mayne, M., N. Grobert, M. Terrones, R. Kamalakaran, M. Ruhle, H. W. Kroto and
D. R. M. Walton (2001). Pyrolytic production of aligned carbon nanotubes from
homogeneously dispersed benzene-based aerosols. Chemical Physics Letters.
338: 101-107.
McCoy, B. J., M. Wang (1994). "Continuous mixture fragmentation kinetics:
particle size reduction and molecular cracking." Chem. Eng. Sci. 49(22): 3773.
Meichel, T., P. T. Dawson, J. A. Morrison, D. E. Sullivan and Y. Koga (1990).
Adsorption of ethylene on graphite near saturation: wetting transition/capillary
condensation. Langmuir. 6: 1579-84.
Moon, J.-M., K. H. An, Y. H. Lee, Y. S. Park, D. J. Bae and G.-S. Park (2001).
"High-Yield Purification Process of Singlewalled Carbon Nanotubes." Journal of
Physical Chemistry B 105(24): 5677-5681.
181
Mukhopadhyay, I., Y. Yokoyama, F. Okino, S. Kawasaki, H. Touhara and W. K.
Hsu (2001). "Effect of chemical modification on electrochemical Li insertion in
highly ordered multi-wall carbon nanotubes." Proceedings - Electrochemical
Society 2001-14(Reactive Intermediates in Organic and Biological
Electrochemistry): 37-40.
Muris, M., N. Dufau, M. Bienfait, N. Dupont-Pavlovsky, Y. Grillet and J. P.
Palmari (2000). Methane and krypton adsorption on single-walled carbon
nanotubes. Langmuir. 16: 7019-7022.
Muster, J., M. Burghard, S. Roth, G. S. Duesberg, E. Hernandez and A. Rubio
(1998). Scanning force microscopy characterization of individual carbon
nanotubes on electrode arrays. Journal of Vacuum Science & Technology, B:
Microelectronics and Nanometer Structures. 16: 2796-2801.
Nagahara, L., T. Burgin, J. Lewenstein, T. Harvey, III, C.-L. Shieh and R. Tsui
(2001). "Surfactant removal from carbon nanotubes." IP.com Journal 1(5): 65.
Nagasawa, S., M. Yudasaka, K. Hirahara, T. Ichihashi and S. Iijima (2000).
"Effect of oxidation on single-wall carbon nanotubes." Chemical Physics Letters
328(4,5,6): 374-380.
Nath, M., B. C. Satishkumar, A. Govindaraj, C. P. Vinod and C. N. R. Rao (2000).
Production of bundles of aligned carbon and carbon-nitrogen nanotubes by the
pyrolysis of precursors on silica-supported iron and cobalt catalysts. Chemical
Physics Letters. 322: 333-340.
Nicholson, D. (1996). Using computer simulation to study the properties of
molecules in micropores. Journal of the Chemical Society, Faraday Transactions.
92: 1-9.
Nicholson, D. and K. E. Gubbins (1996). Separation of carbon dioxide-methane
mixtures by adsorption: effects of geometry and energetics on selectivity. Journal
of Chemical Physics. 104: 8126-8134.
182
Niesz, K., J. B. Nagy, A. Fonseca, I. Willems, Z. Konya, I. Vesselenyi, D. Mehn,
G. Bister and I. Kiricsi (2001). Functional groups generated by mechanical and
chemical breaking of multiwall carbon nanotubes. AIP Conference Proceedings.
591: 345-348.
Overney, G., W. Zhong and D. Tomanek (1993). Structural rigidity and low
frequency vibrational modes of long carbon tubules. Z. Phys. D: At., Mol.
Clusters. 27: 93-6.
Ozaki, T., Y. Iwasa and T. Mitani (2000). Stiffness of Single-Walled Carbon
Nanotubes under Large Strain. Physical Review Letters. 84: 1712-1715.
Pace, E. L. and A. R. Siebert (1959). Heat of adsorption of parahydrogen and
orthodeuterium on Graphon. Journal of Physical Chemistry. 63: 1398-1400.
Pan, Z. W., S. S. Xie, B. H. Chang, C. Y. Wang, L. Lu, W. Liu, W. Y. Zhou, W. Z.
Li and L. X. Qian (1998). Very long carbon nanotubes. Nature (London). 394:
631-632.
Peebles, L. H., Jr. (1995). Carbon Fibers: Formation, Structure, and Properties:
224 pp.
Peigney, A., C. Laurent, F. Dobigeon and A. Rousset (1997). Carbon nanotubes
grown in situ by a novel catalytic method. Journal of Materials Research. 12:
613-615.
Peigney, A., C. Laurent, E. Flahaut and A. Rousset (1999). Carbon nanotubes as
a part of novel ceramic matrix nanocomposites. Advances in Science and
Technology (Faenza, Italy). 15: 593-604.
Peigney, A., C. Laurent, E. Flahaut and A. Rousset (2000). Carbon nanotubes in
novel ceramic matrix nanocomposites. Ceramics International. 26: 677-683.
Peigney, A., C. Laurent and A. Rousset (1997). Synthesis and characterization of
alumina matrix nanocomposites containing carbon nanotubes. Key Engineering
Materials. 132-136: 743-746.
183
Pierard, N., A. Fonseca, Z. Konya, I. Willems, G. Van Tendeloo and J. B. Nagy
(2001). Production of short carbon nanotubes with open tips by ball milling.
Chemical Physics Letters. 335: 1-8.
Pierson, H. O. (1993). Handbook of Carbon, Graphite, Diamond, and Fullerenes:
Properties, Processing, and Applications . Park Ridge, NJ, Noyes Publications.
Piper, J. and J. A. Morrison (1984). Heats of adsorption of methane multilayers
on graphite. Physical Review B: Condensed Matter and Materials Physics. 30:
3486-9.
Piper, J., J. A. Morrison, C. Peters and Y. Ozaki (1983). Heats and entropies of
adsorption of nitrogen on Grafoil at 79.3 K. Journal of the Chemical Society,
Faraday Transactions 1: Physical Chemistry in Condensed Phases. 79: 2863-74.
Pistoia, G. and Editor (1994). Lithium Batteries: New Materials, Developments
and Perspectives. [In: Ind. Chem. Libr., 1994; 5]: 482 pp.
Popov, M., M. Kyotani, R. J. Nemanich and Y. Koga (2002). Superhard phase
composed of single-wall carbon nanotubes. Physical Review B: Condensed
Matter and Materials Physics. 65: 033408/1-033408/4.
Popov, V. N., V. E. Van Doren and M. Balkanski (2000). Elastic properties of
single-walled carbon nanotubes. Physical Review B: Condensed Matter and
Materials Physics. 61: 3078-3084.
Povey, M. J. W., T. J. Mason (1998). Ultrasound in food processing.
Pradhan, B. K., G. U. Sumanasekera, K. W. Adu, H. E. Romero, K. A. Williams
and P. C. Eklund (2002). Experimental probes of the molecular hydrogen-carbon
nanotube interaction. Physica B: Condensed Matter (Amsterdam, Netherlands).
323: 115-121.
Putnam, F. A. and T. Fort, Jr. (1975). Physical adsorption of patchwise
heterogeneous surfaces. I. Heterogeneity, two-dimensional phase transitions,
and spreading pressure of the krypton-graphitized carbon black system near
100.deg.K. Journal of Physical Chemistry. 79: 459-67.
184
Qian, D., E. C. Dickey, R. Andrews and T. Rantell (2000). Load transfer and
deformation mechanisms in carbon nanotube-polystyrene composites. Applied
Physics Letters. 76: 2868-2870.
Radosavljevic, M., J. Lefebvre and A. T. Johnson (2001). High-field electrical
transport and breakdown in bundles of single -wall carbon nanotubes. Physical
Review B: Condensed Matter and Materials Physics. 64: 241307/1-241307/4.
Rangarajan, P., D. Bhattacharyya and E. Grulke (1998). HDPE liquefaction:
random chain scission model. Journal of Applied Polymer Science. 70: 1239-
1251.
Reed, B. W., M. Sarikaya, L. R. Dalton and G. F. Bertsch (2001). Transmission
electron energy-loss spectroscopy study of carbon nanotubes upon high
temperature treatment. Applied Physics Letters. 78: 3358-3360.
Reulet, B., A. Y. Kasumov, M. Kociak, R. Deblock, I. I. Khodos, Y. B. Gorbatov,
V. T. Volkov, C. Journet and H. Bouchiat (2000). Acoustoelectric Effects in
Carbon Nanotubes. Physical Review Letters. 85: 2829-2832.
Riggs, J. E., D. B. Walker, D. L. Carroll and Y.-P. Sun (2000). "Carbon nanotubes
as optical limiters." Proceedings - Electrochemical Society 2000-10(Fullerenes
2000--Volume 8: Electrochemistry and Photochemistry): 267-280.
Riggs, J. E., D. B. Walker, D. L. Carroll and Y.-P. Sun (2000). "Optical Limiting
Properties of Suspended and Solubilized Carbon Nanotubes." Journal of
Physical Chemistry B 104(30): 7071-7076.
Rochefort, A., P. Avouris, F. Lesage and D. R. Salahub (1999). Electrical and
mechanical properties of distorted carbon nanotubes. Physical Review B:
Condensed Matter and Materials Physics. 60: 13824-13830.
Ross, S., J. K. Saelens and J. P. Olivier (1962). Physical adsorption. XVIII.
Limiting isosteric heats of adsorption of gases on graphitized carbon by the
chromatographic method. J. Phys. Chem. 66: 696-700.
185
Saito, R., G. Dresselhaus and M. S. Dresselhaus (1993). Electronic structure of
double-layer graphene tubules. Journal of Applied Physics. 73: 494-500.
Saito, Y., K. Kawabata and M. Okuda (1995). Single-Layered Carbon Nanotubes
Synthesized by Catalytic Assistance of Rare-Earths in a Carbon Arc. J. Phys.
Chem. 99: 16076-9.
Salvetat, J. P., J. M. Bonard, N. H. Thomson, A. J. Kulik, L. Forro, W. Benoit and
L. Zuppiroli (1999). Mechanical properties of carbon nanotubes. Applied Physics
A: Materials Science & Processing. 69: 255-260.
Salvetat, J.-P., G. A. D. Briggs, J.-M. Bonard, R. R. Bacsa, A. J. Kulik, T. Stockli,
N. A. Burnham and L. Forro (1999). Elastic and Shear Moduli of Single-Walled
Carbon Nanotube Ropes. Physical Review Letters. 82: 944-947.
Sams, J. R., Jr., G. Constabaris and G. D. Halsey, Jr. (1962). Adsorption of
argon on graphitized carbon black. Surface area and heats and entropies of
adsorption. Journal of Physical Chemistry. 66: 2154-8.
Sandler, J., M. S. P. Shaffer, T. Prasse, W. Bauhofer, K. Schulte and A. H.
Windle (1999). Development of a dispersion process for carbon nanotubes in an
epoxy matrix and the resulting electrical properties. Polymer. 40: 5967-5971.
Sandstrom, P. H. (1983). Free-flowing SBR black masterbatch powder. U.S. Us,
(Goodyear Tire and Rubber Co., USA). 7 pp.
Satishkumar, B. C., A. Govindaraj, J. Mofokeng, G. N. Subbanna and C. N. R.
Rao (1996). "Novel experiments with carbon nanotubes: opening, filling, closing
and functionalizing nanotubes." Journal of Physics B: Atomic, Molecular and
Optical Physics 29(21): 4925-4934.
Satishkumar, B. C., E. M. Vogl, A. Govindaraj and C. N. R. Rao (1996). "The
decoration of carbon nanotubes by metal nanoparticles." Journal of Physics D:
Applied Physics 29(12): 3173-3176.
Satoh, A. (2001). Rheological properties and orientational distributions of
particles of a dilute colloidal dispersion composed of ferromagnetic
186
spherocylinder particles (approximate solutions by means of Galerkin's method).
Nippon Kikai Gakkai Ronbunshu, B-hen. 67: 112-119.
Sattler, K. (1992). STM of anomalous carbon structures. International Journal of
Modern Physics B: Condensed Matter Physics, Statistical Physics, Applied
Physics. 6: 3603-12.
Schadler, L. S., S. C. Giannaris and P. M. Ajayan (1998). Load transfer in carbon
nanotube epoxy composites. Applied Physics Letters. 73: 3842-3844.
Seifert, G., T. Kohler and T. Frauenheim (2000). "Molecular wires, solenoids, and
capacitors by sidewall functionalization of carbon nanotubes." Applied Physics
Letters 77(9): 1313-1315.
Seraphin, S., D. Zhou, J. Jiao, M. A. Minke, S. Wang, T. Yadav and J. C. Withers
(1994). Catalytic role of nickel, palladium, and platinum in the formation of carbon
nanoclusters. Chem. Phys. Lett. 217: 191-8.
Setlur, A. A., J. M. Lauerhaas, J. Y. Dai and R. P. H. Chang (1996). A method for
synthesizing large quantities of carbon nanotubes and encapsulated copper
nanowires. Appl. Phys. Lett. 69: 345-347.
Shelimov, K. B., R. O. Esenaliev, A. G. Rinzler, C. B. Huffman and R. E. Smalley
(1998). Purification of single-wall carbon nanotubes by ultrasonically assisted
filtration. Chemical Physics Letters. 282: 429-434.
Shen, W., B. Jiang, B. S. Han and S.-s. Xie (2000). Investigation of the Radial
Compression of Carbon Nanotubes with a Scanning Probe Microscope. Physical
Review Letters. 84: 3634-3637.
Shim, M., N. W. S. Kam, R. J. Chen, Y. Li and H. Dai (2002). "Functionalization
of carbon nanotubes for biocompatibility and biomolecular recognition." Nano
Letters 2(4): 285-288.
Sibrell, P. L. and J. D. Miller (1992). Significance of graphitic structural features in
gold adsorption by carbon. Minerals & Metallurgical Processing. 9: 189-95.
187
Siegel, R. W., S. K. Chang, B. J. Ash, J. Stone, P. M. Ajayan, R. W. Doremus
and L. S. Schadler (2001). Mechanical behavior of polymer and ceramic matrix
nanocomposites. Scripta Materialia. 44: 2061-2064.
Simonyan, V. V., J. K. Johnson, A. Kuznetsova and J. T. Yates, Jr. (2001).
Molecular simulation of xenon adsorption on single -walled carbon nanotubes.
Journal of Chemical Physics. 114: 4180-4185.
Slanina, Z. (2001). The physics of fullerene-based and fullerene-related
materials, Wanda Andreoni, Ed. Fullerene Sci. Technol. 9: 561.
Srivastava, D., D. W. Brenner, J. D. Schall, K. D. Ausman, M. Yu and R. S. Ruoff
(1999). "Predictions of Enhanced Chemical Reactivity at Regions of Local
Conformational Strain on Carbon Nanotubes: Kinky Chemistry." Journal of
Physical Chemistry B 103(21): 4330-4337.
Stan, G. and M. W. Cole (1998). Hydrogen adsorption in nanotubes. Journal of
Low Temperature Physics. 110: 539-544.
Stan, G. and M. W. Cole (1998). Low coverage adsorption in cylindrical pores.
Surface Science. 395: 280-291.
Stan, G., V. H. Crespi, M. W. Cole and M. Boninsegni (1998). Interstitial He and
Ne in nanotube bundles. Journal of Low Temperature Physics. 113: 447-452.
Stan, G., S. M. Gatica, M. Boninsegni, S. Curtarolo and M. W. Cole (1999).
Atoms in nanotubes: small dimensions and variable dimensionality. American
Journal of Physics. 67: 1170-1176.
Stauffer, D. and A. Aharony (1992). Introduction to Percolation Theory. London,
Taylor and Francis.
Stepanek, I., L. C. De Menorval, R. Edwards and P. Bernier (1999). "Carbon
nanotubes and gas adsorption." AIP Conference Proceedings 486(Electronic
Properties of Novel Materials--Science and Technology of Molecular
Nanostructures): 456-461.
188
Stephan, O., P. M. Ajayan, C. Colliex, F. Cyrot-Lackmann and E. Sandre (1996).
Curvature-induced bonding changes in carbon nanotubes investigated by
electron energy-loss spectrometry. Physical Review B: Condensed Matter. 53:
13824-13829.
Stephan, O., M. Kociak, L. Henrard, K. Suenaga, A. Gloter, M. Tence, E. Sandre
and C. Colliex (2001). Electron energy-loss spectroscopy on individual
nanotubes. Journal of Electron Spectroscopy and Related Phenomena. 114-116:
209-217.
Subramoney, S. R., R.S.; Lorents, D.C.; Malhotra, R. (1993). Nature 366(637).
Sun, L. F., J. M. Mao, Z. W. Pan, B. H. Chang, W. Y. Zhou, G. Wang, L. X. Qian
and S. S. Xie (1999). Growth of straight nanotubes with a cobalt-nickel catalyst
by chemical vapor deposition. Applied Physics Letters. 74: 644-646.
Sun, Y.-P., J. E. Riggs, B. Martin and D. L. Carroll (2000). "Optical limiting
properties of suspended and solubilized carbon nanotubes." Abstr. Pap. - Am.
Chem. Soc. 220th: IEC-152.
Suzuki, S. and M. Tomita (1996). Observation of potassium-intercalated carbon
nanotubes and their valence-band excitation spectra. J. Appl. Phys. 79: 3739-43.
Takaba, H., M. Katagiri, M. Kubo, R. Vetrivel and A. Miyamoto (1995). Molecular
design of carbon nanotubes for the separation of molecules. Microporous
Materials. 3: 449-55.
Takizawa, M., S. Bandow, T. Torii and S. Iijima (1999). Effect of environment
temperature for synthesizing single-wall carbon nanotubes by arc vaporization
method. Chemical Physics Letters. 302: 146-150.
Takizawa, M., S. Bandow, M. Yudasaka, Y. Ando, H. Shimoyama and S. Iijima
(2000). Change of tube diameter distribution of single -wall carbon nanotubes
induced by changing the bimetallic ratio of Ni and Y catalysts. Chemical Physics
Letters. 326: 351-357.
189
Talapatra, S. and A. D. Migone (2002). Adsorption of methane on bundles of
closed-ended single-wall carbon nanotubes. Physical Review B: Condensed
Matter and Materials Physics. 65: 045416/1-045416/6.
Talapatra, S., A. Z. Zambano, S. E. Weber and A. D. Migone (2000). Gases Do
Not Adsorb on the Interstitial Channels of Closed-Ended Single-Walled Carbon
Nanotube Bundles. Physical Review Letters. 85: 138-141.
Teizer, W., R. B. Hallock, E. Dujardin and T. W. Ebbesen (1999). 4He Desorption
from Single Wall Carbon Nanotube Bundles: A One-Dimensional Adsorbate.
Physical Review Letters. 82: 5305-5308.
Teizer, W., R. B. Hallock, E. Dujardin and T. W. Ebbesen (2000). 44e Desorption
from Single Wall Carbon Nanotube Bundles: A One-Dimensional Adsorbate.
[Erratum to document cited in CA131:161950]. Physical Review Letters. 84:
1844-1845.
Tekleab, D., D. L. Carroll, G. G. Samsonidze and B. I. Yakobson (2001). Strain-
induced electronic property heterogeneity of a carbon nanotube. Physical Review
B: Condensed Matter and Materials Physics. 64: 035419/1-035419/5.
Thess, A., R. Lee, P. Nikolaev, H. Dai, P. Petit, J. Robert, C. Xu, Y. H. Lee, S. G.
Kim and et al. (1996). Crystalline ropes of metallic carbon nanotubes. Science
(Washington, D. C.). 273: 483-487.
Thostenson, E. T., Z. Ren and T. W. Chou (2001). Advances in the science and
technology of carbon nanotubes and their composites: a review. Composites
Science and Technology. 61: 1899-1912.
Tibbetts, G. G. and C. P. Beetz, Jr. (1987). Mechanical properties of vapor-grown
carbon fibers. J. Phys. D: Appl. Phys. 20: 292-7.
Tour, J. M., J. L. Bahr and J. Yang (2002). Process for derivatizing carbon
nanotubes with diazonium species and compositions thereof. PCT Int. Appl. Wo,
(William Marsh Rice University, USA). 45 pp.
190
Treacy, M. M. J., T. W. Ebbesen and J. M. Gibson (1996). Exceptionally high
Young's modulus observed for individual carbon nanotubes. Nature (London).
381: 678-680.
Vaeck, S. V. and G. V. Merken (1981). "Some experiments on detergency in
aqueous and nonaqueous media. XII. Further studies on the interactions
between surfactants, electrolytes and pH in the deposition of hydrophobic
particulate soil." Tenside Deterg. 18(3): 126-9.
Varadan, V. K. (2001). MEMS- and NEMS-based smart devices and systems.
Proceedings of SPIE-The International Society for Optical Engineering. 4591: 28-
38.
Vidali, G., G. Ihm, H. Y. Kim and M. W. Cole (1991). Potentials of physical
adsorption. Surface Science Reports. 12: 133-181.
Vigolo, B., A. Penicaud, C. Coulon, C. Sauder, R. Pailler, C. Journet, P. Bernier
and P. Poulin (2000). Macroscopic fibers and ribbons of oriented carbon
nanotubes. Science (Washington, D. C.). 290: 1331-1334.
Vigolo, B., A. Penicaud, C. Coulon, C. Sauder, R. Pailler, C. Journet, P. Bernier
and P. Poulin (2001). "Dispersions and fibers of carbon nanotubes." Materials
Research Society Symposium Proceedings 633(Nanotubes and Related
Materials): A12.1.1-A12.1.9.
Wang, M., J. M. Smith, B. J. McCoy (1995). "Continuous mixture kinetics for
thermal degradation of poly(styrene-allyl alcohol) in solution." AIChE Journal
41(6): 1521.
Wang, Q., S. R. Challa, D. S. Sholl and J. K. Johnson (1999). Quantum Sieving
in Carbon Nanotubes and Zeolites. Physical Review Letters. 82: 956-959.
Wang, S. K., J. C. Newton, R. Wang, H. Taub, J. R. Dennison and H. Shechter
(1989). Multilayer structure of nitrogen adsorbed on graphite. Physical Review B:
Condensed Matter and Materials Physics. 39: 10331-41.
191
Wang, X., Y. Liu, W. Qiu and D. Zhu (2002). "Immobilization of tetra-tert-
butylphthalocyanines on carbon nanotubes: a first step towards the development
of new nanomaterials." Journal of Materials Chemistry 12(6): 1636-1639.
Wang, X. K., X. W. Lin, V. P. Dravid, J. B. Ketterson and R. P. H. Chang (1995).
Carbon nanotubes synthesized in a hydrogen arc discharge. Appl. Phys. Lett. 66:
2430-2.
Wang, Z. L., R. P. Gao, P. Poncharal, W. A. de Heer, Z. R. Dai and Z. W. Pan
(2001). Mechanical and electrostatic properties of carbon nanotubes and
nanowires. Materials Science & Engineering, C: Biomimetic and Supramolecular
Systems. C16: 3-10.
Weast, R. C. (1986). CRC HANDBOOK OF CHEMISTRY AND PHYSICS, CRC
Press, Inc.
Weber, S., S. Talapatra, C. Journet and A. Migone (2000). Direct measurement
of binding energy via adsorption of methane on SWNT. Science and Application
of Nanotubes, [Proceedings of Nanotube '99, an International Conference], East
Lansing, MI, United States, July 24-27, 1999: 215-221.
Weber, S. E., S. Talapatra, C. Journet, A. Zambano and A. D. Migone (2000).
Determination of the binding energy of methane on single-walled carbon
nanotube bundles. Physical Review B: Condensed Matter and Materials Physics.
61: 13150-13154.
Weber, W. J. and D. L. Goodstein (1999). Theory and Observation of
Displacement Phenomena in Coadsorbed Films. Physical Review Letters. 83:
3888-3891.
Weedon, B. R., R. C. Haddon, H. P. Spielmann and M. S. Meier (1999).
"Fulleroid Addition Regiochemistry Is Driven by .pi.-Orbital Misalignment."
Journal of the American Chemical Society 121(2): 335-340.
192
Wei, B.-Y., M.-C. Hsu, Y.-S. Yang, S.-H. Chien and H.-M. Lin (2003). Gases
adsorption on single-walled carbon nanotubes measured by piezoelectric quartz
crystal microbalance. Materials Chemistry and Physics. 81: 126-133.
Weisenberger, M. (2002). Synthesis and characterization of multiwalled carbon
nanotube/polyacrylonitrile composite fibers and resulting carbon fibers. Chemical
& Materials Engineering. Lexington, KY, University of Kentucky.
Willems, I., Z. Konya, J. F. Colomer, G. Van Tendeloo, N. Nagaraju, A. Fonseca
and J. B. Nagy (2000). Control of the outer diameter of thin carbon nanotubes
synthesized by catalytic decomposition of hydrocarbons. Chemical Physics
Letters. 317: 71-76.
Wong, E. W., P. E. Sheehan and C. M. Lieber (1997). Nanobeam mechanics:
elasticity, strength, and toughness of nanorods and nanotubes. Science
(Washington, D. C.). 277: 1971-1975.
Wood, J. R., Q. Zhao and H. D. Wagner (2001). Orientation of carbon nanotubes
in polymers and its detection by Raman spectroscopy. Composites, Part A:
Applied Science and Manufacturing. 32A: 391-399.
Wu, X. B., P. Chen, J. Lin and K. L. Tan (1999). Hydrogen uptake by carbon
nanotubes. International Journal of Hydrogen Energy. 25: 261-265.
Xie, S., W. Li, Z. Pan, B. Chang and L. Sun (2000). Mechanical and physical
properties on carbon nanotube. Journal of Physics and Chemistry of Solids. 61:
1153-1158.
Yamamoto, K., S. Akita and Y. Nakayama (1996). "Orientation of carbon
nanotubes using electrophoresis." Japanese Journal of Applied Physics, Part 2:
Letters 35(7B): L917-L918.
Yamashita, J., H. Hirayama, Y. Ohshima and K. Takayanagi (1999). Growth of a
single-wall carbon nanotube in the gap of scanning tunneling microscope.
Applied Physics Letters. 74: 2450-2452.
193
Yanagi, H., E. Sawada, A. Manivannan and L. A. Nagahara (2001). Self-
orientation of short single-walled carbon nanotubes deposited on graphite.
Applied Physics Letters. 78: 1355-1357.
Yang, Q. H. F., Y.Y.; Li, F.; Liu, C.; Liu, M.; Fan, Y.Z.; Cheng, H.M.; Wang, M.Z.
(2000). Analysis of pore structure of carbon nanotube. Eurocarbon 2000 1st
world conference on carbon, Berlin, Germany.
Yao, Z., C.-C. Zhu, M. Cheng and J. Liu (2001). Mechanical properties of carbon
nanotube by molecular dynamics simulation. Computational Materials Science.
22: 180-184.
Yase, K., S. Horiuchi, M. Kyotani, M. Yumura, K. Uchida, S. Ohshima, Y. Kuriki,
F. Ikazaki and N. Yamahira (1996). Angular-resolved EELS of a carbon
nanotube. Thin Solid Films. 273: 222-4.
Yoo, D.-H., G.-H. Rue, M. H. W. Chan, Y.-H. Hwang and H.-K. Kim (2003). Study
of Nitrogen Adsorbed on Open-Ended Nanotube Bundles. Journal of Physical
Chemistry B. 107: 1540-1542.
Yoo, D.-H., G.-H. Rue, Y.-H. Hwang and H.-K. Kim (2002). Study of nitrogen
adsorbed on single-walled carbon nanotube bundles. Journal of Physical
Chemistry B. 106: 3371-3374.
Yoo, D.-H., G.-H. Rue, J.-Y. Seo, Y.-H. Hwang, M. H. W. Chan and H.-K. Kim
(2002). Study of argon adsorbed on open-ended carbon nanotube bundles.
Journal of Physical Chemistry B. 106: 9000-9003.
Yu, M.-F., B. S. Files, S. Arepalli and R. S. Ruoff (2000). Tensile loading of ropes
of single wall carbon nanotubes and their mechanical properties. Physical
Review Letters. 84: 5552-5555.
Yu, M.-F., O. Lourie, M. J. Dyer, K. Moloni, T. F. Kelly and R. S. Ruoff (2000).
Strength and breaking mechanism of multiwalled carbon nanotubes under tensile
load. Science (Washington, D. C.). 287: 637-640.
194
Yu, R., L. Chen, Q. Liu, J. Lin, K.-L. Tan, S. C. Ng, H. S. O. Chan, G.-Q. Xu and
T. S. A. Hor (1998). "Platinum Deposition on Carbon Nanotubes via Chemical
Modification." Chemistry of Materials 10(3): 718-722.
Yudasaka, M. (2002). Carbon nanotubes, environmentally benign new materials.
Jidosha Gijutsukai Chubu Shibuho. 51: 48-51.
Yudasaka, M., T. Komatsu, T. Ichihashi and S. Iijima (1997). Single-wall carbon
nanotube formation by laser ablation using double-targets of carbon and metal.
Chemical Physics Letters. 278: 102-106.
Yudasaka, M., N. Sensui, M. Takizawa, S. Bandow, T. Ichihashi and S. Iijima
(1999). Formation of single -wall carbon nanotubes catalyzed by Ni separating
from Y in laser ablation or in arc discharge using a C target containing a NiY
catalyst. Chemical Physics Letters. 312: 155-160.
Yudasaka, M., R. Yamada, N. Sensui, T. Wilkins, T. Ichihashi and S. Iijima
(1999). Mechanism of the Effect of NiCo, Ni and Co Catalysts on the Yield of
Single-Wall Carbon Nanotubes Formed by Pulsed Nd:YAG Laser Ablation.
Journal of Physical Chemistry B. 103: 6224-6229.
Yudasaka, M., M. Zhang and S. Iijima (2000). Porous target enhances production
of single-wall carbon nanotubes by laser ablation. Chemical Physics Letters. 323:
549-553.
Zambano, A. J., S. Talapatra and A. D. Migone (2001). Binding energy and
monolayer capacity of Xe on single -wall carbon nanotube bundles. Physical
Review B: Condensed Matter and Materials Physics. 64: 075415/1-075415/6.
Zettl, A. and J. Cumings (2001). Electromechanical properties of multiwall carbon
nanotubes. AIP Conference Proceedings. 590: 107-112.
Zhang, H.-B., G.-D. Lin, Z.-H. Zhou, X. Dong and T. Chen (2002). Raman
spectra of MWCNTs and MWCNT-based H2-adsorbing system. Carbon. 40:
2429-2436.
195
Zhang, M., M. Yudasaka, A. Koshio, C. Jabs, T. Ichihashi and S. Iijima (2002).
Structure of single-wall carbon nanotubes purified and cut using polymer. Applied
Physics A: Materials Science & Processing . 74: 7-10.
Zhao, Z. and T. Gu (1987). "Adsorption and wetting. II. Adsorption of Triton X-
100 and Triton-305 onto carbon black from water and cyclohexane solutions."
Huaxue Xuebao 45(7): 645-50.
Zhou, G., W. Duan and B. Gu (2001). First-principles study on morphology and
mechanical properties of single-walled carbon nanotube. Chemical Physics
Letters. 333: 344-349.
Zhou, L. G. and S. Q. Shi (2003). Adsorption of foreign atoms on Stone-Wales
defects in carbon nanotube. Carbon. 41: 613-615.
Zhou, O., R. M. Fleming, D. W. Murphy, C. H. Chen, R. C. Haddon, A. P.
Ramirez and S. H. Glarum (1994). Defects in carbon nanostructures. Science
(Washington, D. C., 1883-). 263: 1744-7.
Zhu, H. W., A. Chen, Z. Q. Mao, C. L. Xu, X. Xiao, B. Q. Wei, J. Liang and D. H.
Wu (2000). "The effect of surface treatments on hydrogen storage of carbon
nanotubes." Journal of Materials Science Letters 19(14): 1237-1239.
196
VITA
Jenny Hilding was born on September 6, 1972, in Solna, Sweden. She attended
The Royal Institute of Technology in Stockholm, Sweden, where she received
her M.S. degree in Chemical Engineering in December 1998. She began her
Ph.D. studies at the University of Kentucky in June, 1999.
Publications:
Refereed Journal Articles/Presentation Records:
Hilding, Jenny and Eric Grulke, “Alignment of Dispersed Multiwalled Carbon Nanotubes in Low Strength AC Electrical Fields”, submitted to Nano Letters, 01/07/2004.
Hilding, Jenny and Eric Grulke, “Heat of Adsorption of Butane on Multiwalled Carbon Nanotubes”, submitted to Journal of Physical Chemistry B, 08/11/2003.
Hilding, Jenny; Eric A. Grulke; Z. George Zhang and Fran Lockwood, “Dispersion of Carbon Nanotubes in Liquids”, Review, Journal of Dispersion Science and Technology, 24(1) (2003) 1-41.
Hilding, Jenny; Eric A. Grulke; Susan B. Sinnott; Dali Qian; Rodney Andrews and Marit Jagtoyen, “Sorption of Butane on Carbon Multiwall Nanotubes at Room Temperature”, Langmuir, 17(24) (2001) 7540-7544
Other Publications:
Hilding, Jenny, Eric A. Grulke, Susan B. Sinnott, Dali Qian, Rodney Andrews and Marit Jagtoyen, “Adsorption of butane on multiwall carbon nanotubes” AIChE Annual Meeting 2001, paper 28u, 6 P. Publisher: New York, N.Y. ; American Institute of Chemical Engineers, 2001.
Kashiwagi, Takashi; Eric Grulke; Jenny Hilding; Richard Harris; Walid Awad and Jack Douglas, “Thermal Degradation and Flammability Properties of Polypropylene/Carbon Nanotube Composites”, Macromolecular Rapid Communications, 23(13) (2002) 761-765
Jenny Marie Hilding
January 22nd, 2004