E .IHUNUHHHWlfiiUHlHllUlHIHH.IIHHIWHIIWI

546

   

 

THESlt

rn'

 

 

plhhamx,

Michigan State
University

This is to certify that the
thesis entitled

FUNDAMENTAL ELECTRONIC AND STRUCTURAL
PROPERTIES OF CARBON ONIONS IN EXTREME
ENVIRONMENTS

presented by
Raied A. Al-Duhileb

has been accepted towards fulfillment
of the requirements for the

M. S. degree In Electrical Jineering

 

////%/«

VMajor Professor’ 5 Signature

¥//7/20/O

Date

MSU is an Affirmative Action/Equal Opportunity Employer

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 KIProi/AchrelelRC/DateDueJndd

 

F U?

FUNDAMENTAL ELECTRONIC AND STRUCTURAL PROPERTIES OF CARBON
ONIONS IN EXTREME ENVIRONMENTS
By

Raied A. Al-Duhileb

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE
Electrical Engineering

2010

 

Spi
(Sl

bal

lubr
ShOI

hybr

ABSTRACT
FUNDAMENTAL ELECTRONIC AND STRUCTURAL PROPERTIES OF CARBON
ONIONS IN EXTREME ENVIRONMENTS
By
Raied A. Al-Duhileb
The purpose of this research is to investigate the fundamental tribological
(frictional) and stability characteristics of the fullerene-related nano-materials, carbon
onions, which are nested fullerenes of typically 5-10 Shells. Due to their distinctive
structures, the fictional performance of carbon onions in different surroundings seems to
be enhanced compared to those of other nano-materials. The objectives of this research
are to: I) investigate the nanoscopic structural properties of carbon onions using high
resolution electron microscopy (HRTEM) including the known structural change from
Spherical to polygonal as a function of synthesis temperature; 2) investigate the change in
the spz/sp3 ratio as a function of increasing synthesis temperature using electron energy
loss spectroscopy (EELS); 3) investigate their oxygen functionalities using Raman
spectroscopy; 4) investigate their film stability using scanning electron microscopy
(SEM); and 5) assess their tribological performance in different environments using a
ball-on-disc tribometer.
The practical merit of this research is to develop an environmentally benign solid

lubricant that can be widely used in various applications. Experimental results have
shown that these nano-materials have unique structures and an increasing spz/sp3 bond

hybridization ratio as a function of increasing synthesis temperature.

iii

To my parents and lovely wife,
who provided all the support

to make it becomes true

 

Li

3 Car
3.1
3.2
3.3
3.4 .

4 Wm
4.1 (

4.2 C
4.2..
4.2.2
4.2.3

4.3 E}
4.3.]
4.3.2

5 Quanfit‘

TABLE OF CONTENTS

 

 

 

 

 

 

List of Tables - - - u -- - 4 vi
List of Figures - - - 4 ...... - - -- - vii
1 Introduction - -- - l
1.1 Literature Review of Carbon Nano-Materials ........................................................ 3
1.1.1 Carbon Onions .................................................................................................. 5
1.1.2 C60 Fullerenes (Buckyballs) ........................................................................... 10

1.1.3 Nanocrystalline Diamond (NCD) ................................................................... l3

2 Characterization Instruments and Methods - - - 21
2.1 Transmission Electron Microscopy ...................................................................... 21
2.1.1 TEM Operation .............................................................................................. 24

_ 2.2 Scanning Electron Microscopy ............................................................................. 27
2.3 Electron Energy Loss Spectroscopy ..................................................................... 30
2.4 Raman Spectroscopy ............................................................................................ 35
2.5 Sample Preparation Methods ................................................................................ 37
2.5.1 TEM, SEM, and EELS Samples .................................................................... 37

3 Carbon Atom Hybridization - -- - - - _- 40
3.1 Hybrid Orbital Introduction .................................................................................. 40
3.2 Sp Hybrid Orbitals ................................................................................................. 41
3.3 sp2 Hybrid Orbitals ............................................................................................... 44
3.4 sp3 Hybrid Orbitals ............................................................................................... 46

4 Quantitative EELS Investigation of Carbon Materials ........... 50
4.1 Quantitative Method: EELS ................................................................................. 51
4.2 Quantitative EELS: Research Issues .................................................................... 52
4.2.1 Near-Edge Segregation .................................................................................. 52
4.2.2 sp2 Reference Samples ................................................................................... 54
4.2.3 Inclusion of the 287.0 eV Peak ...................................................................... 56

4.3 EELS Deconvolution Techniques ......................................................................... 56
4.3.1 Two-Energy Windows Method ...................................................................... 56
4.3.2 Functional Fitting Method .............................................................................. 56

5 Quantitative EELS Investigation of Carbon Onions - ...... 60

 

iv

came

6.

Ha

 

5.1 Structural Characterization Using HRTEM .......................................................... 60

 

5.2 EELS Deconvolution: Gaussian Function ............................................................ 64
5.3 EELS Deconvolution: Lorentzian Function ......................................................... 68
5.4 EELS Deconvolution: F-Variance Function ......................................................... 72
5.4.1 Investigation of Synthesis Uniformity by SEM ............................................. 77
Conclusion and Future Work - 79
6.1 Discussed Research Conclusion ........................................................................... 79
6.2 Continuing Fundamental Studies: Raman Spectroscopy ...................................... 80
6.3 Continuing Application Studies: Irradiated Carbon Onions ................................. 82
6.4 Continuing Application Studies: Tribological Performance of Carbon Onions in
Harsh Environments ...................................................................................................... 87

 

it

T:
ter
wit

Tal

Tab
temp
with

TablI
tempe
and 2-

Table

LIST OF TABLES
Table 1.1 Physical properties of C60 fullerenes (Buckyballs). (Table: adapted fi'om
Hirata et al. [3], and www.sesres.corn/PhysicalProperties.asp [34]) ................................ 10

Table 5.1 spz/sp3 ratio of carbon onion samples prepared at increasing synthesis
temperature. The results were obtained using Gaussian deconvolution and 2-peak fit.... 65

Table 5.2 spz/sp3 ratio of carbon onion samples prepared at increasing synthesis
temperature. The results were obtained using Gaussian deconvolution and 3-peak fit.... 67

Table 5.3 sz/sp3 ratio of carbon onion samples prepared at increasing synthesis
temperature. The results were obtained using symmetric Lorentzian area deconvolution
with 3-peak fit and fixed starting & end points ................................................................ 69

Table 5.4 First and second inflection points (IPS) and their corresponding differences. 71

Table 5.5 First spz/sp3 ratio of carbon onion samples prepared at increasing synthesis
temperature. The results were obtained using symmetric Lorentzian area deconvolution
with 3-peak fit and exact inflection points ........................................................................ 72

Table 5.6 spz/sp3 ratio of carbon onion samples prepared at increasing synthesis
temperature. The results were obtained using asymmetric f-variance area deconvolution
and 2-peak fit .................................................................................................................... 75

Table 6.1 Raman spectroscopy of C60 and Carbon Onions at wavelength 532 nm ........ 82

vi

LIST OF FIGURES

Figure 1.1 An HRTEM image showing the multi-layer structure of carbon onions.
(Image: Raied A. Al-Duhileb) ............................................................................................ 6

Figure 1.2 EELS core loss spectra of (a) NCD, (b) spherical carbon onions, and (c)
polyhedral carbon onions. (Image: adapted from Tomita et al. [22]) ................................. 8

Figure 1.3 Raman spectra of carbon onion samples prepared at increasing synthesis
temperature. (Image: adapted from Tomita et al. [20]) ...................................................... 9

Figure 1.4 An HRTEM image of a single-walled carbon nanotube filled with a linear
chain of C60 particles. (Image: Benjamin W. Jacobs) ....................................................... 12

Figure 1.5 Raman spectra of a film of purified C60. (Image: adapted from Bethune et al.
[39]) ................................................................................................................................... 13

Figure 1.6 An HRTEM HRTEM image of NCD particles having an average size of 5
nm. (Image: Raied A. Al-Duhileb) ................................................................................... 14

Figure 1.7 EELS core loss spectra of purified NCD particles. Results revealed a Single
peak at 289 eV and were fi'ee of sp2 bonded carbon. (Image: adapted from Hansen et al.
[36]) ................................................................................................................................... 15

Figure 1.8 A Raman spectrum of NCD. (Image: adapted from Hansen et al. [36]) ....... 16

Figure 2.1 The TEM anatomy. The thermionic or field emission electron gun emits a
beam of focused electrons that penetrate the thin sample. The magnetic condenser,
objective, diffraction, intermediate, and projector lenses control the electron beam and
bring it to a focal point. Electrons that pass the sample are collected and the image is
shown on the viewing screen or camera. (Image: Raied A. AI-Duhileb) ......................... 23

Figure 2.2 Typical diffraction pattern (single crystal pattern). (Image: Raied A. A1-
Duhileb) ............................................................................................................................ 25

vii

Figure 2.3 Determining the atomic lattice Spacing (d) in diffi'action mode using wave
length of the electron (A), camera length (L), and relative distance between the diffraction
spots (R). (Image: Raied A. Al-Duhileb) .......................................................................... 26

Figure 2.4 A schematic diagram illustrating the basic components and operation of an
SEM. An electron gun focuses an electron beam on the surface of the sample. This results
in producing secondary electrons that are collected by special electron detector and
displayed as a mapped image on a screen. (Image: adapted from www.chm.bris.ac.uk [3])
........................................................................................................................................... 28

Figure 2.5 Energy level diagram showing excitation from the inner shell (valance band)
into the empty state higher shell. (Image: adapted from Brundle et al. [1]) ..................... 31

Figure 2.6 The magnetic prism is used to record the intensity of scattered electrons at
different position to obtain the energy loss spectrum. (Image: Raied A. Al-Duhileb) ..... 31

Figure 2.7 A schematic diagram illustrating the shape of the inner-shell ionization edge
of an element in the periodic table. (Image: Raied A. Al-Duhileb) .................................. 32

Figure 2.8 Excitation of an inner-Shell electron by incident light beam photons. The
energy of the scattered light beam photons is less than the energy of the incident laser
beam photons. (Image: Raied A. Al-Duhileb) .................................................................. 36

Figure 2.9 A schematic Raman scattering spectrum showing Rayleigh line, Stokes
scattering, and anti-Stokes scattering. (Image: adapted from Denk [7]) .......................... 37

Figure 3.1 A hybridization model of Sp orbitals. An electron is promoted to a higher
empty state due to the small energy differences between the lower 2s and the upper 2p
energy levels. (Image: Raied A. Al-Duhileb) ................................................................... 42

Figure 3.2 New atomic orbitals (i.e. spa and spb) resulting from mixing the 2s orbital
with one of the 2p orbitals (2px, 2py, or 2pz). The resulting energy states are located in
opposite directions, and centered on the atom. (Image: adapted from
www.chem.umass.edu [5]) ............................................................................................... 43

Figure 3.3 A hybridization model of sp2 orbitals. (Image: Raied A. Al-Duhileb) .......... 45

viii

 

 

 

Fig
orb
Im

Fig

Fig
25 0
WWW

Figu
trans

Figu
a purl

Figur

Figui
carbo
(lmag

Figur
hole.

Figur
I7000
Gauss

Figur
Gauss

Figure 3.4 New atomic orbitals (i.e., spfi, spg, and spfi) resulting from mixing the 25

orbital with two of the 2p orbitals (2px, 2py, or 2pZ). (Image: adapted from
www.chem.umass.edu [5]) ............................................................................................... 46

Figure 3.5 A hybridization model of sp3 orbitals. (Image: Raied A. Al-Duhileb) .......... 47

Figure 3.6 New atomic orbitals (i.e., spg, spg, sp2, and spfi) resulting from mixing the

23 orbital with the three 2p orbitals (2px, 2py, and 2pz). (Image: adapted from
www.chem.umass.edu [5]) ............................................................................................... 47

Figure 4.1 A typical EELS Spectrum from a largely Sp2 carbon sample with the
transition edges marked .................................................................................................... 52

Figure 4.2 A theoretical calculation of the energy-loss-near—edge structure (ELNES) for
a purely sp2 graphite .......................................................................................................... 53

Figure 5.1 An HRTEM image of the largely sp3 NCD. (Image: Raied A. Al-Duhileb). 61

Figure 5.2 HRTEM images of the structural transition fi'om spherical to polygonal for
carbon onions synthesized at temperatures (a) 1700°C (b) 2000°C and (c) 2300°C.
(Image: Raied A. Al-Duhileb) .......................................................................................... 62

Figure 5.3 A TEM image showing aggregate C60 particles suspended over a lacey film
hole. (Image: Raied A. Al-Duhileb) ................................................................................. 63

Figure 5.4 Deconvolution of the core loss EELS spectrum of carbon onions, prepared at
1700°C. The original spectrum was deconvolved using 2-peak fit and the symmetric
Gaussian deconvolution. (Image: produced using the OriginPro software package) ....... 65

Figure 5.5 A plot of the spz/sp3 ratio in carbon onions as a function of synthesis
temperature. The plot is based on results obtained using 2-peak fit and the symmetric
Gaussian deconvolution. (Image: produced using the OriginPro software package) ....... 66

ix

 

F igi
fiinc
Sl’ntj

Figu
I 700

Figure 5.6 Deconvolution of the core loss EELS spectrum of carbon onions, prepared at
1700°C. The original spectrum was deconvolved using 3-peak fit and the symmetric
Gaussian deconvolution. (Image: produced using the OriginPro software package) ....... 67

Figure 5.7 A plot of the spz/sp3 ratio in carbon onions as a function of synthesis
temperature. The plot is based on results obtained using 3-peak fit and the symmetric
Gaussian deconvolution. (Image: produced using the OriginPro software package) ....... 68

Figure 5.8 Deconvolution of the core loss EELS spectrum of carbon onions, prepared at
1700°C. The original spectrum was deconvolved using 3-peak fit and the symmetric
Lorentzian area deconvolutio. (Image: produced using the OriginPro software package)70

Figure 5.9 A plot of the sz/Sp3 ratio in carbon onions as a function of synthesis
temperature. The plot is based on results obtained using approach. (Image: produced
using the OriginPro software package) ............................................................................. 70

Figure 5.10 A plot of the sp2/sp3 ratio in carbon onions as a function of synthesis
temperature. The plot is based on results obtained using approach 2. (Image: produced
using the OriginPro software package) ............................................................................. 72

Figure 5.11 Theoretical near edge It* and 6* spectra from an orientation resolved
calculation, in which a graphite sample and the electron beam were at the “magic angle”
........................................................................................................................................... 74

Figure 5.12 Deconvolution of the core loss EELS spectrum of carbon onions, prepared
at 1700°C. The original spectrum was deconvolved using 2-peak fit and the asymmetric f-
variance function. (Image: produced using the OriginPro software package) .................. 76

Figure 5.13 A plot showing an increase in the spz/sp3 ratio in carbon onions as a
function of synthesis temperature. The results from the 2000°C were variable due to
synthesis uniformity issues. (Image: produced using the OriginPro software package) .. 76

Figure 5.14 SEM surface images of carbon onions synthesized at temperatures (a)
1700°C (b) 2000°C and (c) 2300°C. Images by: Raied A. Al-Duhileb ............................. 77

Figure 6.1 Experimental Raman spectra of C60 and carbon onion samples prepared at
increasing synthesis temperature. (Image: produced using the OriginPro software
package) ............................................................................................................................ 81

Figure 6.2 HRTEM images of pre-radiation carbon onion samples prepared at1700°C.
Typical spherical morphologies are observed in (a). A fused onions feature is observed in
(b). (Images by: Xudong Fan, with Kaylee McElroy, and Virginia M. Ayres). ............... 84

Figure 6.3 HRTEM images of carbon onion samples prepared at 1700°C following
irradiation with fully stripped Calcium-48 heavy ions at energies (a-c) 70 MeV/nucleon
and (d-e) 140 MeV/nucleon. (Images by: Xudong Fan, with Kaylee McElroy, and
Virginia M. Ayres). ........................................................................................................... 86

xi

 

Solid.

TESear

Chapter 1

Introduction

Current conventional lubricants that are introduced between moving surfaces to
reduce frictional forces and improve efficiency of machinery applications are often
exposed to extreme environments. The performance of the existing lubricant approaches
in severe environments like space, chemical, and petroleum mediums faces many
challenges that might cause degradation and failure of these substances. For example,
liquid lubricants require frequent maintenance in terrestrial industries, and are not
practicable in space applications due to the exposure to vacuum.

To overcome the deficiencies of the presently available liquid lubricants,
researchers in tribology (science of friction) have been examining solid-based lubricant
alternativesl’z’3. Although graphite is the most familiar solid lubricant, experiments with
graphite films have revealed friction coefficients of about 0.21 in air but only 0.5 to 0.6 in
vacuum. The increase of the friction coefficient in vacuum is a well known effect due to
the absence of assisting agents like water vapor and oxygen4. This is usually compensated
by adding heavy metals, such as molybdenum or tungsten disulfide. These additives
introduce an environmentally hazardous moiety that is not present in typical graphitic
solid lubricants used in air3.

Nam-materials such as carbon nanotubes (CNTS)2, C60, and carbon onions3 are
solid-based substances with new properties enabled by their nanoscopic size. Recent

research indicates that nano-materials could exhibit better frictional performance in such

extreme environments compared to conventional materials”. Preliminary studies have
shown that these materials have low (< 1.0) fiiction coefficients in air and vacuum, and
maintain consistent behavior when exposed to heat and radiation1’2’3. Single-walled CNTS
(SWCNTS) exhibit friction coefficients of (~0.6) in air and an improved tribological
performance in vacuum (~0.2)2. C60 particles have friction coefficients of around 0.5 to
0.8 in both air and vacuum3. These results were due to infinitesimal size of these nano-
materials, which tend to clump and compress into high shear strength layerss. Carbon
onions, on the other hand, have shown the lowest fiiction coefficients of about 0.03 in
both air and vacuum3. These nano-materials do not require potentially hazardous
additives to attain their desirable tribological properties in vacuum; therefore, they Show
great promise as an environmentally benign solid lubricant in both ambient and vacuum
environments3 .

Carbon onions, consisting of multiple layers of concentric graphitic shells, are
fullerene—related nano-materials. Daniel U garte first observed them as a result of electron
beam irradiation of soot or diamond clusters in high resolution transmission electron
microscopy (HRTEM)3’6’7. They have been recognized as a new form of carbon with
unique structure3; however, their fimdamental properties have been far less studied than
those of C60 and other single-layer fullerenes such as C70. This research primarily focuses
on investigating the tribological and stability characteristics of carbon onions in air and
vacuum because they are expected to be effective and safe solid lubricant for many

potential applications.

 

tn
iOl
qu.
of-
con
nanl
Goo
mate
Scan
wear
macrt

tribon

1.1.

maIOF Sl

amOTphc

To fillly exploit the great potential of carbon onions as a solid lubricant, one needs
to study their physical structures, investigate their electronic structures and chemical
composition, and assess their tribological performance in different environments.

In this research, the surfaces and internal structures of carbon onions prepared at
increasing synthesis temperature are extensively studied using high resolution
transmission electron microscopy (HRTEM). The integrated areas under the carbon K-
ionization edges obtained using electron energy loss spectroscopy (EELS) are also
quantified to precisely measure the change in the spz/sp3 ratio. In addition, several state-
of-the—art material characterization methodologies like Raman Spectroscopy are used to
confirm the electronic structure of these nano-materials. C60 particles, which are pure Sp2
nano-materials, are used as a control to characterize the properties of carbon onions.
Good quality nanocrystalline diamond (NCD) powders, which are nearly pure Sp3 nano-
materials, are used as the starting material for preparing the carbon onion samples.
Scanning electron microscopy (SEM) is used to assess carbon onion film stability on
wear surfaces. These nanoscopic experimental results are then correlated with the
macroscopic fiictional performance of carbon onions assessed using a ball-on-disk

tribometer.

1.1. Literature Review of Carbon Nano—Materials

For a long period of time, the element carbon has been thought to have three
major solid forms or allotropes under ambient conditions: diamond, graphite and

amorphous carbon. Its physical properties depend on its chemical bonding types.

Diamond, which has sp3 hybridized bonds, is a transparent and hard material, whereas
graphite, which has sp2 hybridized bonds, is opaque and softs.

In 1985, a new form of carbon that was later called Buckminsterfullerene or
F ullerenes was discovered by Kroto et a19. Due to their remarkable discovery of this new
carbon allotrope, H. Kroto, R. Curl, and R. Smalley were awarded the 1996 Nobel Prize
for Chemistry. This new carbon form may have several shapes such as hollow sphere or
ellipsoid, and could contain both sp3 and Sp2 bonding sites"). In 1991, research by S.
Iijima extended the fullerene family to include single-walled and multi-walled CNTSl 1.
The discovery of fullerenes has significantly increased the number of known carbon
allotropes and resulted in starting a new era in science forming the beginning of
nanotechnology.

Scientists have been investigating the properties of these firllerene-related nano-
materials since their discovery. It has been found that the crystal Size of these nano-
materials in general is an important parameter that requires further consideration when
studying their characteristics. The stability of nano-structures differs from that of bulk
materials”. For instance, carbon clusters having less than 104 atoms were discovered to
continue exhibiting liquid-like behavior at very low temperatures that are far below the
melting point of the corresponding bulk materials, and are subject to surface
reconstructions. However, C60, and carbon nanotubes and their multi-shell versions are
highly stable crystals of nano-scale size 13.

Although graphite is the most known stable form of carbon crystals having a
lattice constant greater than 10 nm, it is believed that filllerene-related nano-materials are

more stable and preferred form for tiny clusters having lattice constant less than 2 nm. It

was found that, as the crystal dimensions increase, the stability of the fullerene-related
nano-materials decreases”. C60 fullerenes (also known as buckyballs) are the most stable
fullerenes identified to date. These consist of 60 carbon atoms that are arranged in 20
hexagons and 12 pentagons, and have an average diameter of 1 nm. Carbon onions are
composed of several concentric fullerene shells with a C60 inner shell. The outer Shells
would be individually less stable than C60 due to their larger Sizes (~ 5-10 nm)”;
however, the inter-layer interactions of the multi-shell carbon onion system provide an
additional source of stability. Also, stability is affected by whether the individual shells
are perfectly crystalline or defective. One critical parameter explored in this thesis is the
defect concentration of the individual shells and how this affects the carbon onion
properties.

Depending on the local energetic conditions, the small differences in the stability
of fullerene-related nano-materials may cause phase transformations between different
allotropes”. For instance, carbon onions can transform into other carbon forms such as
(N CD) and vice versan’ls. They can also transform into other cage structures like CNTS,

which are allotropes of carbon with a hollow cylindrical nanostructurelg.

1.1.1. Carbon Onions

Carbon onions are hollow nanometric fullerene-related materials that typically
have 7 to 10 fullerene shells with outer diameters ranging from 5 to 10 nm as shown in
Figure 1.120. The largest carbon onion molecules, which were synthesized by Daniel
Ugarte of the Federal Polytechnic School in Switzerland, have around 70 concentric

fullerene shells”. These nano—materials can be produced in small quantities using several

methods such as heat treatment of amorphous carbon soot or NCD particles as in the
present research, and ion implantation of carbon into metal substrates. The former
preparation method has the advantage of producing carbon onions with uniform size and
varied morphologyzz. Industrial scale-up of combustion methods with carefully controlled

operation and gas mixture parameters is enabling commercial quantity production”.

 

Figure 1.1 An HRTEM image showing the multi-layer structure of carbon onions.
Image by: Raied A. Al-Duhileb.

The physical properties of these nano-materials strongly depend on the growth
conditions. HRTEM images have shown a structural evolution from spherical to
polyhedral multi-graphitic shells as the synthesis temperature is increased. This structural
evolution has been associated with a change in the spZ/sp3 ratio, and assumed to be due to
a reduction in the potential sp3-defective hybridized bonds that are seen as broken shells

under the I-IRTEM24’25. A contribution of this thesis is that the assumed correlation of

 

SP

qu;
car
as I
like

ll€Xl

evolt

since

spheril
from ll
spectra
those 01
II Was g
Spectrum
are belie
amount 0
temperatu

Th

spherical to polyhedral multi-graphitic shells with a change in the spz/sp3 ratio is
quantitatively investigated using EELS. It is essential for the synthesis of optimum
carbon onion lubricating particles to precisely determine the change in their spz/sp3 ratio
as the growth temperature increases. Researchers have been utilizing several approaches
like HRTEM, EELS, and Raman Spectroscopy to investigate this phenomenon3’ 22. We
next review the results reported to date from each technique.

Early work by the Butenko group26 identified the spherical to polyhedral
evolution of the carbon onions as a fimction of the synthesis temperature. This result has
since been established by several groupszo’25 including ours.

Early literature for EELS reported the core loss spectra (i.e. K-ionization edges) of
spherical carbon onions prepared at lower annealing temperatures are slightly different
fi'om those of polyhedral carbon onions prepared at higher synthesis temperatures. EELS
spectra of spherical carbon onions are more similar to NCD K-ionization edges, whereas
those of polyhedral carbon onions Show some dissimilarities as shown in Figure 1.22234.
It was generally concluded that the similarity between spherical carbon onions and NCD
spectrum is due to the presence of sp3 bonds in both materials. Spherical carbon onions
are believed to have such hybridized bonding because either they contain a small residual
amount of diamond particles, or sp3 bonds remain as structural defects at lower grth
temperatures”.

The tribological performance of carbon onions was also reported in the literature.
These nano-materials reveal stable friction coefficients (< 0.1) both in air and vacuum3. It

has been noticed that fiictional performance of these nano-materials improves as the

carbon onion size gets larger.

 

TH

 

 

 
     
 

(c) Polyhedral
onions

      
 

:i
3; (b) Spherical
3: onions
'3
=
3
E 4
fl
(a) NCD

 

 

LLlLLlllAlAAIIAIIIIJAIEALJIIJILI

280 285 290 295 300 305 310
Energy Loss (eV)

 

Figure 1.2 EELS core loss spectra of (a) NCD, (b) spherical carbon onions, and (c)
polyhedral carbon onions”.

Raman spectroscopy, which is widely utilized in material sciences to examine the
phonon modes of different allotropes and carbon phases, has also been used to
characterize the nano-structure of carbon onions”. Raman spectra of carbon onion
samples, synthesized at different increasing temperatures, have revealed two broad
Raman bands at about 1350 and 1580 cm'1 as shown in Figure 1.3. The former band at
about 1350 cm'] was identified as a D band or disorder band due to the presence of finite-
sized graphite crystallites, dangling alkanes, etc. The second band at about 1580 cm'1 was
reported as a G band. In planar graphite, the G peak at 1580 cm'l corresponds to two very
close modes, which are the stretch mode of a single graphene layer and the shear mode of
two graphene layers. In Ref. [20], it was reported that the hill width at half maximum

(FWHM) of the G band decreased as the synthesis temperature increased.

The Ref. [20] results were acquired at room temperature, and are reminiscent of
the differences between the Raman spectra of single-walled and multi-walled carbon
nanotubes. Single-walled carbon nanotubes show clear peaks that correspond to
identified breathing (A) and tangential (E) modes”. Multi-walled carbon nanotubes show
two broad peaks that are very Similar to those shown in Figure 1.328. Similarly, the
single-walled C50 Raman spectrum shows clear peaks that correspond to identified

breathing and tangential (H and T) modeszg'30

, while at room temperature, the multi-
walled carbon onion Raman spectrum shows broad peaks. One of the contributions of this
thesis is the first, as far as we know, low temperature (77K) Raman spectroscopy of

multi-walled carbon onions, which is resolved into a series of clear peaks.

 

T'UI'U'IVI'TTDIUU'IIIUU'

Annealing
Temp. (“C

 
    
  

2200

 

 

Intensity (a.u.)

 

 

M.

1200 1300 1400 1500 16001700
Raman Shift (cm'l)

Figure 1.3 Raman spectra of carbon onion samples prepared at increasing synthesis
temperature”.

TH

 

 

 

1.1.2. C60 Fullerenes (Buckyballs)

To characterize the properties of carbon onions, pure sp2 C60 particles are used in
this research as a control nano-material. Although C60, which is the most stable form of
fullerenes, was first produced in isolable quantities in 1990 by causing an are between
two graphite rods to burn in a helium atmosphere, some of its chemical and physical
properties remain unclear, and are under investigation. It is known to consist of 20
hexagons and 12 pentagons that are the basis of a symmetrical closed cage structure.
Each carbon atom is tied to three other carbon atoms, and is sp2 hybridized9. C60
fullerenes have two bond lengths. The 6:6 ring bonds (~l .40 A), which are double bonds,
are shorter than the 6:5 bonds (~1.458 A)“. They tend not to form double bonds in the
pentagons, causing poor electron delocalization in these areas. Consequently, C60
molecules act as electron-deficient alkenes and react with electron-rich Species”.
Experiments have shown that C60 molecules are soluble in some solvents, such as

benzene, toluene and chloroform”. Some physical properties of C60 are listed in Table

1.1.

Table 1.1 Physical properties of C60 fullerenes (Buckyballs)34

 

 

 

Average C-C . . .

distance 1.44 (A) Resrstmty 1014 (ohm.m)
Average Friction Thermal

Coefficient3 0'5 conductivity 0'4 (W/mK)
Density 1.72 (g.cm'3) Boiling point Sublimes at 800 K

 

Several experiments aiming to investigate the electronic properties and chemical

structures of C60 molecules have been carried out using different material characterization

10

 

 

 

methodologies, such as HRTEM and EELS. A group of C50 fiillerenes at the edge of a
lacey film hole cannot be resolved into individual cages, and conclusively identified due
to their low contrast and small size (~1 nm diameter). Therefore, different methods were
reported to view the structure of C50 molecules. One way, used in this thesis, is
investigate a small-diameter carbon nanotube (CN T) or Boron Nitride nanotube (BNT)
filled with a linear chain of C50 called a “peapod”35. HRTEM images obtained using
peapod or other methods confirmed the predicted theoretical spherical structure of C50

molecules as shown in Figure 1.4.

 

Figure 1.4 An HRTEM image of a single-walled carbon nanotube filled with a linear
chain of C60 particles. Image by: Benjamin W. Jacob.

The core loss EELS spectra of C60 fullerenes Showed two major sharp K-edges

around 284 and 293 eV that correspond to a pure sp2 carbon material36’37. It is noted that

TH

 

 

 

special care must be taken during EELS investigation of C60 samples to ensure that the
carbon lacey film support is not included in the spectrum.

The tribological performance of C60 molecules was also reported in the literature.
Pure C60 films reveal relatively high fiiction coefficients (0.55 - 0.8) when tested under
different loads and materials. It is believed that these high fiiction coefficients are due to
the tendency of the C60 molecules to clump and compress into high shear strength layers
rather than due to the impurities in the fullerenes. However, some experiments have
shown lower fiiction coefficients for C60 particles when they are dissolved in some
chemical solvents, such as benzene-solvated C60. This improved frictional performance is
believed to be due to the lowered shear strength of the hexagonal close-packed structure
of the benzene-solvated C60 molecular crystals”.

A typical Raman spectrum of C60 fullerenes is shown in Figure 1.5 (a-b)”. It is
difficult to avoid the production of a Cso/Cm mixture during synthesis and therefore the
sample was purified using a column chromatography. The absence of an intense C70
Raman band at 1569 cm'1 confirmed the absence of C70, and the C60 sample purity.
Among the eight clear C60 Raman bands, three strong ones were observed. The results
revealed eight clear Raman bands and several weaker ones between 200 and 1600 cm].

The first strong Raman band was observed at 273 cm", which corresponds to the

predicted Hg squashing mode”. The two other strong Raman bands that appear at 496

and 1470 cm'1 correspond to totally symmetric breathing Ag modes”.

12

 

Intensity (a.u.)

400 800 1200 1600
Wavelength (cm'l)

Figure 1.5 Raman Spectra of a film of purified C60 3’.

1.1.3. N anocrystalline Diamond (N CD)

In addition to C60 particles, nanocrystalline diamond (NCD), which was
discovered in the early 19608 in the former Soviet Union“), was used as the starting
material for preparing the carbon onions. This nanomaterial, whose core is purely sp3
bonded carbon“, can be grown using different techniques, such as chemical vapor
deposition (CVD)42 and detonation of carbon-containing explosives”. The latter
technique produces NCD as a power. In this research, the term NCD refers to the NCD
powders produced using the detonation method. The powder is composed of particles
having an average size of 5 nm as Shown in Figure 1.6. Each particle consists of a
diamond core that is partially or completely surrounded by layers of graphitic and/or

amorphous carbon. This carbon nanomaterial, which can be produced in large

commercial quantities, is also anticipated to combine an active surface (large surface to

13

volume ratio) having similar properties to those of macroscopic diamond, such as the
diamond’s hardness and Young’s modulus, chemical stability, high thermal conductivity,
and electrical resistivity“.

NCD is used in composite materials like lubricants“, and it has many potential

medical, chemical, and electrical applications42’45.

The performance of its applications is

affected by the inability to provide particles having well-controlled surface chemistry.
The detonation synthesis method, used to produce NCD, results in raw soot

containing NCD particles and other non—diamond carbons. To date, producing pure sp3

NCD is a challenge due to the absence of an easy and cost effective purification treatment

method“.

 

Figure 1.6 An HRTEM image of NCD particles having an average size of 5 nm.
Image by: Raied A. Al-Duhileb.

 

 

 

 

 

The electronic properties and chemical structures of purified NCD particles has
been reported in the literature”. EELS core loss spectra showed results that had a single

peak at 289 eV, and were free of Sp2 bonded carbon as shown in Figure 1.7.

 

Intensity (a.u.)

 

 

 

 

Energy Loss (eV)

Figure 1.7 EELS core loss spectra of purified NCD particles. Results revealed a
Single peak at 289 eV and were free of Sp bonded carbon .

Purified NCD revealed vibrational Raman spectra as Shown in Figure 1.8. The
results Showed four (4) clear Raman bands and several weaker ones at a range between
500 and 1650 cm". The first strong Raman band, observed at 500 cm", corresponds to
amorphous sp3 bonded carbon. The second peak, observed at 1090 cm], corresponds to
diamond surface phonons. The third peak, observed at 1325 cm'l, corresponds to sp3
carbon. The last strong peak, observed at 1620 cm", was assigned to localized <1 0 0>

sp2 bonded pairs of carbon.

15

 

rH--~

 

 

 

Intensity

 

j j A J I. L j A l 1;J__L|j

500 1000 1500
Wavelength (cm'l)

Figure 1.8 A Raman spectrum of NCD“.

The tribological performance of the NCD films (~0.01-0.05)42’45 has been widely
reported in the literature, whereas NCD particles are less studied, and they showed higher

friction coefficients (~0.8)3.

l6

References

 

l.

10.

ll.

12.

13.

J. Yang and K. Groh, “Materials issues in the space environment”, Material Research
Society Bulletin 34: 12-16 (2010).

A. Hirata and N. Yoshioka, “Sliding fiiction properties of carbon nanotubes coatings
deposited by microwave plasma chemical vapor deposition”, Tribology International
37: 893—898 (2004).

A. Hirata, M. lgarashi, and T. Kaito “Study on solid lubricant properties of carbon

onions produced by heat treatment of diamond clusters or particles”, Tribology
International 37 : 899—905 (2004).

B. Yen, “Influence of water vapor and oxygen on the tribology of carbon materials
with Sp2 valence configuration”, Wear 192: 208-215 (1996).

W, Zhao, J. Tang, A. Puri, R. Sweany, Y. Li, and L. Chen , “Tribological properties
of fiillerenes C60 and C70 microparticles”, Journal of Materials Research 11: 2749-
2756.

D. Ugarte, “Curling and closure of graphitic networks under electron—beam
irradiation”, Nature 359: 707-715 (1992).

D. Ugarte, “Formation mechanism of quasi-spherical carbon particles induced by
electron bombardment”, Chemical Physics Letters 207: 473-481 (1993).

C. Mantel], Carbon and Graphite Handbook, (Interscience Publishers, New York,
1968)

H. Kroto, J. Heath, S. Obrien, R. Curl, and R. Smalley, “C60: Buckminsterfullerene”,
Nature 318: 162 (1985).

S. Mraz, “A new buckyball bounces into town”, machinedesigncom, (2005).
S. Iijima, “Helical microtubules of graphitic carbon”, Nature 354: 56 (1991).

J. Viecelli, and F. Ree, “Carbon particle phase transformation kinetics in detonation
waves”, Journal of Applied Physics 88: 683 (2000).

J. Viecelli, S. Bastea, J. Glosli, and F. Ree, “Phase transformations of nanometer size
carbon particles in shocked hydrocarbons and explosives”, The Journal of Chemical
Physics 115: 2730 (2001).

17

 

14. D. Tomanek and M. Schluter, “Growth regimes of carboncClusters”, Physical Review
Letters 67: 2331 (1991).

15. M. Zwanger and F. Banhart, “The structure of concentric-shell carbon onions as
determined by high-resolution electron microscopy”, Philosophical Magazine B-
Physics of Condensed Matter Statistical Mechanics Electronic Optical and Magnetic
Properties 72: 149-157 (1995).

16. A. Barnard, “Theory and modeling of nanocarbon phase stability”, Diamond and
Related Materials 15: 285-291 (2006).

17. A. Barnard and P. Zapol, “A model for the phase stability of arbitrary nanoparticles
as a function of size and shape”, The Journal of Chemical Physics 121: 4276 (2004).

18. V. Kuznetsov, A. Chuvilin, Y. Butenko, l. Mal'kov, and V. Titov, “Onion-like carbon
from ultra-disperse diamond”, Chemical Physics Letters 222: 343-348 (1994).

19. N. Park, K. Lee, S. Han, J. Yu, and J. Il'lrn, “Energetics of large carbon clusters:
crossover from fiillerenes to nanotubes”, Physical Review B 65: 121405 (2002).

20. S. Tomita, T. Sakurai, H. Ohta, M. Fujii, and S. Hayashi, “Structure and electronic
properties of carbon onions”, The Journal of Chemical Physics 114: 7477 (2001).

21. P. Schewe and B. Stein, “Carbon onions”, Nature 22: (1992).

22. S. Tomita, M. F ujii, S. Hayashi, and K. Yamamoto, “Electron energy-loss
spectroscopy of carbon onions”, Chemical Physics Letters 305: 225-229 (1999).

23. H. Richter, T. Lada, V. Vejins, and J. Howard, “Large-scale production of fullerenes,
carbon nanotubes and fullerenic Materials”, Nano-C, Inc., 33 Southwest Park,
Westwood, MA 02090, USA.

24. O. Mykhaylyka, Y. Solonin, D. Batchelder, and R. Brydson, “Transformation of
nanodiamond into carbon onions: A comparative study by high-resolution
transmission electron microscopy, electron energy-loss spectroscopy, x-ray

diffraction, small-angle x-ray scattering, and ultraviolet Raman spectroscopy”,
Journal of Applied Physics 97: 074302 (2005).

25. S. Osswald, G. Yushin, V. Mochalin, S. O. Kucheyev, and Y. Gogotsi, “Control of
sp3/sp2 carbon ratio and surface chemistry of nanodiamond powders by selective

oxidation in air”, Journal of American Chemical Society, 128: 11635-11642 (2006).

26. E. Obraztsova, M. F ujii, S. Hayashi, V. Kuznetsov, Y. Butenko, and A. Chuvilin,
“Raman identification of onion-like carbon”, Carbon 36: 821—826 (1998).

18

 

27.

28.

M. Dresselhaus, G. Dresselhaus, R. Saito, and A. Jorio, “Raman spectroscopy of
carbon nanotubes”, Physics Reports 409: 47-99 (2005).

H. Zhang, , G. Lin, Z. Zhou, X. Dong, and T. Chen, “Raman spectra of MWCNTS
and MWCNT-based Hz-adsorbing system”, Carbon 40: 2429-2436 (2002).

29. P. Zhou, K. Wang, Y. Wang, P. Eklund, M. Dresselhaus, G. Dresselhaus, and R.

Jishi, “Raman scattering in C60 and alkali-metal-saturated C60”, Physical Review B
46: 2595—2605 (1992).

30. G. Adams, J. Page, 0. Sankey, K. Sinha, J. Menendez, and D. Huffman, “First-

31.

32.

33.

34.

35.

36.

37.

38.

39.

principles quantum molecular-dynamics study of the vibrations of icosahedral C60”,
Physical Review B 44: 4052—4055 (1991).

K. Hedberg, L. Hedberg, D. Bethune, C. Brown, H. Dorn, R. Johnson, and M. Vries,
“Bond lengths in free molecules of buckminsterfullerene, C60, from gas-phase
electron diffraction”, Science 254: 410 — 412 (1991).

B. Yadav and R. Kumar, “Structure, properties and applications of fullerenes”,
International Journal of Nanotechnology and Applications 2: 15-24 (2008).

V. Bezmel'nitsyn, A. Eletskii, and M. Okun', “F ullerenes in solutions”, Physics—
Uspekhi 41: 1091(1998).

www.sesres.com/PhysicalProperties.asp, March 2010.

A. Zettl, J. Cumings, Wei-qiang Han, and W. Mickelson, “Boron nitride nanotube
peapods”, The American Institute of Physics, 140-144 (2002).

P. Hansen, P.Fallon and W. Kritschmer, “An EELS study of filllerite - Ciao/C70”,
Chemical Physics Letters 181: 367-372 (1991).

A. Papworth, C. Kiely, A. Burden, S. Silva, and G. Amaratunga, “Electron-energy-
loss spectroscopy characterization of the Sp2 bonding fraction within carbon thin
films”, Physical Review B 62: 12628—12631 (2000).

W. Zhao, J. Tang, A. Puri, R. Sweany, Y. Li, and L. Chen, “Tribological properties of
fullerenes C60 and C70 microparticles”, Material Research Society 11: 2749-2756
(1996)

D. Bethune, G. Meijer, W. Tang, H. Rosen, W. Golden, H. Seki, C. Brown, and M. de

Vries, “Vibrational Raman and infrared spectra of chromatographically separated C60
and C70 fullerene clusters”, Chemical Physics Letters 179: 181-186 (1991).

19

n

 

 

 

40.

41.

42.

43.

44.

45.

46.

V. Danilenko, “On the history of the discovery of nanodiamond synthesis”, Physics
of the Solid State 46: 595-599 (2004).

S. Prawer, K. Nugent, D. Jamieson, J. Orwa, L. Bursill, and J. Peng, “The Raman
spectrum of nanocrystalline diamond”, Chemical Physics Letters 332: 93-97 (2000).

A. Sumant, O. Auciello, R. Carpick, S. Srinivasan, and J. Butler,
“Ultrananocrystalline and nanocrystalline diamond thin films for MEMS/NEMS
Applications”. Material Research Society Bulletin 35: 281-288 (2010).

V. Danilenko, Synthesizing and sintering of diamond by explosion,
(Energoatomizdat: Moscow, 2003).

N. Red'kin, “Lubricants with ultradisperse diamond—graphite powder”, Chemistry
and Technology of Fuels and Oils 40: 164-170 (2004).

D. Gruen, O. Shenderova, and A. Vul’, Synthesis, properties and applications of
ultrananocrystalline diamond, (Springer: Dordrecht, Netherland, 2005).

V. Dolmatov, “Detonation synthesis ultradispersed diamonds: properties and
applications”, Russian Chemical Reviews 70: 607-626 (2001).

20

 

TH

 

 

Chapter 2

Characterization Instruments and Methods

Many instruments and methods are currently being used to investigate the
fundamental properties of novel nano-materials. These instruments and methods have
recently been developed in response to the expanding need to filrther explore and
characterize these newly discovered materials at the nanoscale. In this research, several
material characterization instruments were used to analyze the structural, chemical and
electronic properties of carbon onions, C60, and NCD. This chapter explains the
filndamental principles of the instruments utilized to characterize the nano-materials in

this research.

2.1. Transmission Electron MicroscOpyl’2

Transmission Electron Microscopy (TEM) is an important technique that is
widely used for nanoscale characterization. It can achieve very high spatial resolution
(~0.2 nm), and is used for obtaining highly magnified images, material chemical
compositions, and electronic structures.

In this technique, a beam of highly accelerated electrons is focused on a very thin
(< 100 nm) specimen. The deflected and undeflected electrons that penetrate the sample
are collected to generate the TEM signal. Because electrons tend to bend in the TEM

column, a series of magnetic lenses are positioned above and below the sample to bring a

21

parallel electron beam to a focal point. The generated TEM signal can be magnified as
little as 50 times to as much as a million times. This superior magnification is achieved
due to the small wavelength of electrons (~0.0037 nm) that result in better resolution

capabilities. The wave nature of an electron (A) is given by de Broglie’s relation:

 

 

 

A — E — h ' 1 21
p ,/2mqu 1+ ZqV ’ ( ' )
InoC2

where h is Plank’s constant 6.6 x 10'34 J .s, p is the electron momentum, mo is the electron
rest mass 9.1 x 10'31 kg, q is the fundamental electron charge 1.6 x 10'19 C, and V is the
TEM operating acceleration voltage. The relationship, given in Equation 2.1, includes the
relativistic correction required by the high velocities that the electrons achieve during
acceleration in the TEM column.

The TEM system anatomy is illustrated in Figure 2.1, showing the location of the
thin sample, electromagnetic lenses, the electron gun and the imaging and recording
system within the TEM. The diffraction-limited spatial resolution of two individual

points in the TEM system is given by Abbe’s Equation:

0.61 A.

Resolution Power = —,— , (2.2)
n 8111 a

where A is the wavelength of the electron, n is the refractive index and a is the acceptance

angle. The 0.61 factor describes Resolution Power in a TEM system, and it includes the

22

TH

 

 

theoretical diffraction limit and the additional limits introduced by the objective and
condenser lenses. The small wavelength of electrons results in a smaller and better
resolution capability. This simply suggests that TEM instruments operating at higher
accelerating voltages or energies provide a significant reduction in the minimum
resolvable spacing since the electron wavelength is inversely proportional to the

momentum, which is directly proportional to the amount of energy.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Gun U
Condenser lens 1_’ Condenser aperture ' Lenses
|—4 ////A
Objective lens E: Sample I I
. Objective aperture
Diffraction lens , _
Intermediate lens/v \ Selected area aperture
Projector lens /
Viewing screen
, Camera/film

 

 

 

 

Figure 2.1 The TEM anatomy. The thermionic or field emission electron gun emits a
beam of focused electrons that penetrate the thin sample. The magnetic condenser,
objective, diffraction, intermediate, and projector lenses control the electron beam and
bring it to a focal point. Electrons that pass the sample are collected and the image is
shown on the viewing screen or camera.

TEM instruments that operate on high accelerating voltages have improved
penetration ability because highly energetic electrons interact less with matter than low-

energy electrons. Therefore, analyzing thicker samples is achievable using high-voltage

TEM instruments. The quality of electron penetration is obtained by measuring the mean

23

 

 

d
II

5P
or]

the

diff

distance between scattering events. The fewer elastic and inelastic scatterings yield a
farther electron penetration into the sample.

Despite its great advantages, TEM has limited depth resolution, causing images to
be projected onto a two-dimensional detector. The structural information along the beam
direction of multi-structural features is superimposed at the image plane, which results in
a convolution of the scattering contrast. The image contrast fi'om a sample region must be
deconvolved in order to determine overlapping microstructural features of a given object.
It addition, TEM does not have an inherent ability to differentiate between atomic
species. In order to overcome this drawback, TEM instruments are often equipped with
other tools and techniques that distinguish between different atomic species by measuring

the deflection angles of scattered electrons.

2.1.1. TEM Operation“2

TEM instruments have two methods to view the sample: image mode and
diffraction mode.

In the image mode, the scattered electrons are utilized to create the image details
and atomic structure. Conventional TEM instruments have three different image modes:
bright-field microscopy, dark-field microscopy, and high resolution transmission electron
microscopy (HRTEM). In the bright-field mode, the diffracted electrons are blocked
using a small objective aperture, which only passes the un-diffracted (forward) electrons.
This mode provides higher intensity, lower contrast, and better focusing. In the dark-field
mode, in contrast, the diffracted electrons are allowed to pass beside the forward

electrons through the objective aperture. The dark-field image is not always a direct

24

reversal of the bright-field mode, and sometimes reveals more structural details. In the
HRTEM mode, a larger objective aperture is used to pass all electrons, which recombine
in the image forming process to create high resolution images. This mode enables the
user to thoroughly investigate the local structural details of different materials.

In the diffraction mode, a diffraction pattern is formed on the back focal plane of
the objective lens. The diffracting volume (i.e. diffi‘acted electrons) is limited by using a
selected-area aperture, which results in many spots or reflections as shown in Figure 2.2.
This technique is called selected-area diffraction (SAD) or selected-area electron

diffraction (SAED).

 

Figure 2.2 Typical diffraction pattern (single crystal pattern).

By measuring the relative distances and angles of the diffraction spots the crystal
structure of a particular area of the sample can be determined. The atomic lattice spacing

(d) is given by the relation:

25

 

TH

fl

 

(2.3)

where ll. is the electron wavelength, L is the camera length, and R is the relative distance

between the diffraction spots as shown in Figure 2.3. The results of the diffraction mode

analysis reveal one of the three typical diffraction patterns: single crystal, poly-crystal, or

 

 

  
    
 
  
  

 

 

 

 

 

 

amorphous.
Incident beam
J‘iifi‘ii‘ij specimen
(hkl) Reflecting planes
Rd = XL
L
Diffracted beam
Direct beam
R
O
'\

Diffraction spots

Figure 2.3 Determining the atomic lattice spacing (d) in diffraction mode using wave

length of the electron (A), camera length (L), and relative distance between the diffraction

spots (R).

26

It-

 

 

In this research, high resolution images of the internal structures of carbon onions,
C60, and NCD were obtained using a JEOL 2200FS field emission TEM operated at 200

kV at the Center for Advanced Microscopy at Michigan State University.

2.2. Scanning Electron Microscopyl’2

The scanning electron microscope (SEM) is another important tool for
investigating the structural properties of materials. It has an improved resolution that can
approach a few nm, and high magnification capabilities that can reach up to 300,000x. It
obtains images of different samples using an electron beam that can be accelerated at 15
kV. Although SEM has some operational features that are similar to TEM, the SEM
electron beam can interact with the sample surface (~ 5 nm), whereas electrons penetrate
and pass through the sample in TEM. Therefore, the SEM can investigate the surfaces of
samples that are thicker (> 100 nm) than those required by the TEM (< 100 nm).

The SEM electron beam, which is produced in a vacuum using either a thermionic
or a field emission electron gun, is focused on a particular area of the sample surface. The
SEM produces three types of images, namely secondary electron images, backseattered
electron images, and X-ray mapping. These different images are due to the complex
interactions that take place when an incident-beam electrons strike the surface of a
sample.

In secondary electron imaging, the incident-beam electrons transfer part of their
energy to the electrons of the sample. This creates low energy (average energy of about 3
eV ) secondary electrons that are capable of escaping the sample surface. The emitted

secondary electrons are partially collected using an electron detector at an angle to the

27

sample, and used to produce mapped images on a cathode ray tube (CRT) as shown in

 

 

 

   
  
  

 

 

 

 

 

 

 

 

 

 

Figure 2.43 .
Electron Gun
Lens 3 / Aperture
J Scan Generator
Scan Coils @qqfi
Lens . E ,__, CRT Display!
— Camera
Sample
Electron — Amplifier
Collector —l.._""—"'" High Vacuum Pump

 

 

 

Figure 2.4 A schematic diagram illustrating the basic components and operation of an
SEM. An electron gun focuses an electron beam on the surface 'of the sample. This results
in producing secondary electrons that are collected by special electron detector and
displayed as a mapped image on a screen’.

In the second imaging mode, backseattered electron imaging, a considerable
fi'action of the beam electrons undergoes elastic interactions with the electrons of the
sample. These interactions produce higher energy backseattered electrons (average
energy of about 50 eV). Backseattered electrons are produced from the whole beam-
sample interaction area, and therefore electrons from greater depths have sufficient
energy to escape the sample. The maximum escape depth of the backseattered electrons is

inversely proportional to the average atomic number (Z) of the sample. More

backseattered electrons can escape samples having small atomic numbers, whereas fewer

28

backscattered electrons escape samples having large atomic numbers. Due to their large
escape width (~ fraction of a micrometer), backscattered electrons result in a resolution
lower than that produced using secondary electrons. A scintillator or semiconductor
detector is positioned above the sample, and utilized to detect the high energy
backscattered electrons.

The third imaging mode is energy dispersive X-ray spectroscopy (EDS). In
addition to backscattered and secondary electrons, an inelastic interaction between the
beam electrons and the sample electrons excites primary sample electrons at lower
ground state atomic energy levels to higher energy shells in the atoms of the sample.
When the excited electrons decay to their ground states, they emit X—ray photons, which
have different energies and wavelengths (“energy dispersive”), from the entire beam-
sample interaction area. Because each element in the periodic table requires a unique
energy to excite the primary electrons to higher energy Shells, the emitted X-ray signal is
utilized to create elemental maps of the sample showing the spatial distribution of
particular elements in the field of view. The X-ray photons are detected by a photon
detector that is attached to the column of the SEM.

In this research, the sample uniformity of carbon onion, C60, and NCD samples
was investigated using Hitachi S-47OOII field emission SEM (FESEM), which is
equipped with an EDAX Phoenix energy dispersive X-ray spectrometer system for
microanalysis, a chamber scope for direct chamber viewing and a backscattered electron
detector for Z contrast, at the W. M. Keck Microfabrication Facility (KMF) at Michigan

State University.

29

2.3. Electron Energy Loss Spectroscopyl’2’4

Electron energy loss spectroscopy (EELS) is a method for analyzing the energy
distribution of elastic and inelastic scattered electrons that pass through a thin sample
characterized by TEM or scanning transmission electron microscope (STEM). This
useful technique provides elemental quantifications and electronic structures of different
samples. Compared to other material analytical techniques such as energy dispersive
spectroscopy (EDS), EELS has very high sensitivity to light elements and spatial energy
resolution of about 1 eV.

The spectral profile or signature of EELS is characterized by the amount of
energy that an electron loses when passing through a thin specimen. Electrons within the
sample atoms accept specific values of energy and are promoted to higher empty state
levels (conduction bands). This causes electrons of the TEM accelerated beam to lose
part of their energy. The lost energy is equivalent to the energy transferred to the sample
electrons as shown in Figure 2.5.

To obtain EELS spectra, inelastic scattered electrons that interact with the inner
shell electrons of the sample and lose part of their energy, and elastic scattered electrons
that interact with the atomic nuclei and do not lose energy are analyzed using a magnetic
prism as shown in Figure 2.6.

In this process, the scattered electrons. pass through a uniform magnetic field
generated by the magnetic prism. Inelastic scattered electrons exhibit smaller bending
radii, whereas elastic scattered electrons reveal lager bending radii corresponding to their
higher energy and velocity. The bending radius of an accelerated electron is given by

Equation (2.4):

30

 

 

 

 

 

 

 

Evac —
CB*
EF— ------ edge of CB
VB
M _
L__
K _
v
1s atomic orbital

 

 

 

 

Figure 2.5 Energy level diagram showing excitation from the inner shell (valance
band) into the empty state higher shell (conduction band)‘.

 

Electron with energy E,

 

O.‘l.‘l.‘l.’l.'l.’l,‘|.’l.’..‘I.°
x-a-x-x-x-x-x-xgx-x-x-
.o.'o"u.'o.'l.:o.'t.'0 .0..0.'0.

flux-\Ifi-xu .lfilfll\nfi¢fil

   

Electron with '

Electron with

energy E°__. l/’energy Eo-AE
I

' Radius = r

 

 

7

 

 

 

___-.... lilies

Figure 2.6 The magnetic prism is used to record the intensity of scattered electrons at
different position to obtain the energy loss spectrum.

 

 

Magnetic field = B

 

31

THI

 

'Y.m0.V
q.B’

 

radius (r) = (2.4)

where m0 is the rest mass of an electron, v is the velocity of an accelerated electron, q is
the unit electronic charge 1.6 X 10'19 C, and B is the magnetic field . The intensity of the
scattered electrons is recorded at different positions, which are then used to obtain the
EELS spectra.

Each element in the periodic table has a unique inner-shell ionization edge that
corresponds to the energy required for prompting an inner-shell electron to empty states
above the Fermi level as shown in Figure 2.7. On the experimental spectra, the edge
energy is determined at the point where the first derivative is the maximum or the second

derivative is zero.

 

Lower Probability

 

flunnnunnun

   
  

 

 

250 300 {350 400 450
Ex
Energy (eV)

Figure 2.7 A schematic diagram illustrating the shape of the inner-shell ionization
edge of an element in the periodic table.

 

 

 

32

Tl-r

 

The theory of elemental quantification is used in EELS to determine the precise
chemical compositions of a sample based on the differences between the inner-shell

binding energies of chemical elements. To quantify an inner-edge (K, L or M) in a given

sample, it is assumed that the inner-shell single scattering intensity (linner) is related to

the total intensity (IT) by Equation 2.5:

Iinner = I’inner IT, (25)

where Pinner is the probability of transition, and given by Equation 2.6:

P inner = N 0 inner [Ta (2-6)

where (Sinner is the inner-shell ionization cross-section and N is the areal density of

atoms (atoms / unit area).

These mathematical relationships can be extended to a Spectrum with multiple
edges corresponding to different chemical compositions. For example, the expression for
analyzing the k-ionization edges of a sample having two elements A and B is given by

Equation 2.7:

33

TH

 

— = fl. (2.7)

This expression is also applicable for L- and M-ionization edges, and combinations of
edges. Therefore, to measure the ratio of two elements or allotropes that are present in a
sample, the areal intensities under EELS spectra of each element or allotrope should be
obtained using special de—convolution (fitting) techniques, and analyzed using Equation
2.7.

Users may experience some difficulties in obtaining EELS spectra if the sample is
too thick and/or the TEM beam is unstable. Care should be taken in preparing samples for
EELS analysis, and the beam should be in the standby mode for at least eight hours to
maximize its stability.

The powerful EELS capability is employed in this research to analyze the bond
hybridization (i.e., the sz/Sp3 ratio) of the carbon onion samples as a function of
synthesis temperature. Because EELS can achieve very high spatial resolution, and is
highly sensitive to the different bond structures within a sample, it can accurately
quantify of the nature of the bonding present in the carbon systems.

In this research, EELS experiments were performed using an integrated Omega
filter in a JEOL 2200FS. The energy resolution was around 1.5 eV, evaluated using the
fill] width half max (F WHM) of the zero loss peak. EELS spectra were first corrected for
instrument background and plural scattering using Gatan analysis software. The corrected
Spectra were then analyzed using a deconvolution technique based on asymmetric f-

variance area (Peak F itTM by SigmaPlot, version 4.12). The use of the asymmetric f-

34

variance area to describe the K-edge versus a symmetric Gaussian description is an

original contribution of this thesiss.

2.4. Raman Spectroscopy“6

The bonding nature of solid samples can also be studied using Raman
spectroscopy. This technique sets the atoms of a solid sample into vibration modes
(phonons), which depend on the atomic masses and bond force constants of the sample
atoms, using a laser beam (photons). The laser beam photons are often elastically
scattered when they interact with electrons of the sample. This elastic interaction between
the laser beam and the sample’s electrons is called Rayleigh scattering. Less fi'equently,
the photons of the laser beam scatter inelastically and lose part of their initial energy
when interacting with the sample’s electrons. This inelastic interaction between the
photons of the laser beam and the electrons of the sample takes place by one of two
methods, namely Stokes scattering and anti-Stokes scattering. In Stokes scattering, the
sample’s electrons are promoted to higher energy states because the incident laser beam
photons transfer part of their energy to them. As a result, the scattered laser beam photons
will have less energy and a longer wavelength than the incident beam laser photons, as
shown in Figure 2.8. The difference in wavelengths of the incident and scattered laser
beam photons, which is called Raman shift, is used to determine the elemental
composition of the sample because each element in the periodic table exhibits a unique

spectrum and Raman shift.

35

 

Incident photon Ei

Promoted e'

 

0
@ Ei>E,—>Ai<).s

Figure 2.8 Excitation of an inner-shell electron by incident light beam photons. The
energy of the scattered light beam photons is less than the energy of the incident laser
beam photons.

 

 

Scattered photon E,

 

In anti-Stokes scattering, scattered laser beam photons gain some energy when the
excited electrons of the sample, which are at higher energy states, drop down to lower
vibration energy states and release part of their energy. This causes the wavelength of the
scattered laser beam photons to be shorter, and their energy to be higher. This process is a
mirror of the Stokes scattering on the opposite side of the Rayleigh line as shown in
Figure 2.97. Because this process strongly depends on temperature, anti-Stokes scattering
is usually not measured, and does not contribute to a Stokes Raman Spectrum.

Each element exhibits a unique Raman spectrum, which corresponds to a
distinctive phonon mode, when the sample interacts with laser beam photons. Therefore,
Raman spectroscopy is used in this research to further analyze the different allotropes of
carbon: carbon onions, C60, and NCD. Room temperature Raman spectroscopy
experiments were performed using the Kaiser Optical micro Raman Spectrograph that

uses a frequency-doubled yttrium aluminum garnet (YAG) laser with a wavelength of

36

532 nm, and is available at the W. M. Keck Microfabrication Facility (KMF) at Michigan

State University.

 

Rayleigh
Band
D j i

Stokes Anti-Stokes
Band Band

.L_A__.

V0 — Vvib V0 V0 + Vvib

 

 

 

 

 

 

 

Figure 2.9 A schematic Raman scattering spectrum showing Rayleigh line, Stokes
scattering, and anti-Stokes scattering7.

2.5. Sample Preparation Methods

The accuracy of the material characterization instruments strongly depends on the
preparation quality of the analyzed specimens. The preparation procedures are crucial and
require special care to obtain high-quality results. This section outlines the different
sample preparation methods that were followed to perform the experimental part of the

research.

2.5.1. TEM, SEM, and EELS Samples

37

,4.

I711

 

 

Carbon onions were prepared from crystalline diamond nanoparticles having an
average diameter of 5 nm. The diamond nanoparticles were heated in inert ambience in
an infrared gold image fumace. A graphite holder filled with 10 mg of diamond
nanoparticles was placed inside a furnace evacuated to approximately 1.3 Pa with a rotary
pump, and slowly heated in argon gas flow at 1.5><105 Pa to 1700°C, 2000°C and
2300°C, respectively. The furnace temperature was held for 60 s and then gradually
cooled to room temperature in argon flow. The C60 and NCD (Cheap Tubes Inc.,
Brattleboro, VT) samples were acquired commercially, and were 99wt% pure.

To prepare the samples for TEM, SEM, and EELS analysis, these carbon nano-
materials were suspended in ethyl alcohol, sonicated, and then dispersed onto carbon
lacey film 200 mesh copper grids (SPI Supplies, West Chester, PA). Special care was
taken to acquire an EELS spectrum of only the carbon onions and not the carbon lacey
film. Large holes in the lacey film were utilized to ensure the EELS spectra contained

only carbon onions.

38

References

 

1. C. Brundle, C. Evans, and S. Wilson, Encyclopedia of Materials Characterization,
(Butterworth-Heinemann, Stoneham, MA, 1992).

2. S. Flegler, J. Heckman, and K. Klomparens, Scanning and Transmission Electron
Microscopy: An Introduction, (Oxford University Press, New York, NY, 1993).

3. http://www.chm.bris.ac.uk/pt/diamond/stuthesis/chapter2.htrn, January 2010.

4. R. Egerton, Electron Energy Loss Spectroscopy in the Electron Microscope, (Plenum
Press, New York, NY, 1986).

5. R. Alduhaileb, V. Ayres, E. Jacobs, X. Fan, K. McElroy, A. Hirata, M. Horikoshi, N.
Lehnert, and M. Galinato, “Carbon onion films-molecular interactions of multi-layer

firllerenes”, Materials Research Society Fall Meeting, Boston, MA, 2009.

6. R. McCreery, Raman Spectroscopy for Chemical Analysis, (A Wiley-Interscience
Publication, New York, NY, 2000).

7. M. Denk, “Vibrational spectroscopy”, CHEM 2070 - Structure and Spectroscopy,
Guelph, Ontario N1G 2W1, Canada (2005).

39

Chapter 3

Carbon Atom Hybridization

The element carbon exists in several forms or allotropes in nature, and may form
an enormous number of chemical compounds. This diversity is mainly due to the unusual
number of bond types that this element can form with other carbon atoms and elements in
the periodic table. Understanding the different bond types in carbon is essential for
investigating the structural and frictional properties of carbon nano-materials, such as
carbon onions, C60, and NCD. Single-shell fullerene C60 and multi-shell fullerene carbon
onions are almost purely Sp2 bonded carbon materials. On the other hand, the starting
NCD particles used to prepare the carbon onion samples are almost purely Sp3 bonded
carbon material. The original work in this research quantifies the ratio of the different
hybridized bonding present in the carbon onions, and investigates the effect of bond
hybridization on their tribological performance. This chapter introduces the basic
chemical principles and bond hybridization types existing in nano-materials used in this

research.

3.1. Hybrid Orbital Introductionl’2’3”4

The concept of orbital hybridization was introduced to explain the atomic orbital
structures of some chemical compounds and molecules that exhibited unpredicted

chemical shapes4. For instance, the early predicted shapes of the water molecule (H20),

40

ammonia (NH3), and methane (CH4), based on strict s- and p- orbital filling, did not
coincide with the experimental outcomes.

The element carbon, the sixth element in the periodic table, has six electrons
around the nucleus in an electron configuration (152 252 2p2). The two core electrons in
the innermost orbital (15) are strongly tied to the nucleus, and require high energy to
promote them to higher energy levels. The original assumption was that the two electrons
in the 25 orbital were similarly in a closed shell and therefore inert. Carbon was expected
to form bonds with two hydrogen (H) atoms using its remaining two electrons in two
orthogonal 2p orbitals, resulting in two C—H bonds with a bond angle of 90° between
them. However, the practical chemical experiments revealed several other hydrocarbon
compounds including methane CH4 with its four-fold tetrahedral C-H configuration. This
phenomenon also took place in other chemical compounds such as BF3 and BC13. To
explain these unexpected results, chemists suggested that the 2s electrons must contribute
in the formation of bonds between carbon and boron atoms, and other elements.

In reality, the small energy difference between the upper 2p and the lower 25
energy levels results in an opportunity to change the valence electron configuration of
carbon from strict [252] 2p2 to a series of S-p orbital combinations by mixing the
electronic wave functions of the 25 and 2p electrons (hybridization). The possible bond

hybridizations for carbon are: 5p, sz and S133.

3.2. sp Hybrid Orbitalsl’z’3

In the Sp bond hybridization model, the 25 orbital is hybridized or mixed with one

of the 2p orbitals (2px, 2py, or 2pz). This mixture results in two of the valence electrons

41

being located in hybridized Sp orbitals with two of the valence electrons located in
unchanged p orbitals, as Shown in Figure 3.1. The mixture of the 25 and 2p atomic orbital
wave functions |25) and |2p) can be mathematically represented by a linear combination

as shown in Equations 3.1 and 3.2:

 

Original electron configuration of carbon

C...l___I_ LII l

152 252 2px 2py 2pz

 

 

sp orbital hybridization of carbon

C411 l_l l 1

Is2 Sp. Spb p p

 

 

 

 

 

Figure 3.1 A hybridization model of Sp orbitals. An electron is promoted to a higher
empty state due to the small energy differences between the lower 25 and the upper 2p
energy levels.

|8pa> = C1 |28> + C2 |2px> (3.1)

ha.) = C3 l28> + C4 |2px). ' (3.2)

where C1, C2, C3, and C4 are the mixing or linear combination coefficientsz. These

coefficients can be calculated using the orthonormality conditions: (Spalspb) = 0,

42

(Spalspa) = 1r (Spblspb) = 1: and ((Spalspal + (Spblspblhs component = 1-
Solving for the unknown coefficients using these conditions yields the following results:

1
C1 = C2 = C3 =5;, and C4 = -$. Equations 3.1 and 3.2 can therefore be rewritten as:

Ispa> = f [|2s> + I2px>l (3.3)
Ispb> = :15 112s>-12p.>1. (3.4)

The atomic orbitals of the new energy states (i.e., Spa and spb) are centered on the

atom, separated by 180°, and located in opposite directions as Shown in Figure 3.2. These

two orbitals have high electron probability, and two electrons of the same energy.

 

Arrangement of Hybrid Orbitals Geometric Figures

alt-uh. for 3.4.x... 1800
S 7' ‘ ‘ .
' . H
. .
.‘r
.
,.

v.4
. at} C .4 p . ;. ' .
. ; - — _- ~.~_ . ~ Llnear
. ~.'. ‘. u
._ ,

 

 

 

 

 

Figure 3.2 New atomic orbitals (i.e. Spa and spb) resulting from mixing the 25 orbital
with one of the 2p orbitals (2px, 2py, or 2pz). The resulting energy states are located in
opposite directions, and centered on the atom. Source: www.chem.umass.edu5.

The triple bond between the two carbon atoms consists of one sigma (a) bond,

which is formed by the covalent overlap of 5p hybrid orbitals, and two pi (11') bonds,

43

Tuss

which are also covalent bonds but weaker than 0 bonds because of the lesser degree of

overlap]. The two It bonds are formed by the overlap of parallel 2py and 2pz orbitals.

3.3. 5p2 Hybrid Orbitalsl’2’3

In the 5p2 orbital hybridization model the 2s orbital is hybridized or mixed with

two of the 2p orbitals (2px, 2py, or 2pz). This results in three Sp2 orbitals and one

unchanged p orbital as shown in Figure 3.3. The linear combination of the 2s and 2p

atomic orbitals can be mathematically represented by Equations 3.5, 3.6, and 3.7:

lsp§> = C1128) - J 1 - Cf|2py> (3.5)
\f-
Ispt> = c212s> + 71 - c;- [731% +§12py>] (3.6)
J—
Isp§> = Csl25> + 71 - Ci {-33 lsz> + i |2py> ] (3.7)

where C1, C2, and C3 are the mixture or linear combination coefficients. These

coefficients are determined using the same orthonormal requirements applied to the

(Sp?) and (25 components), which are given Equations 3.8, 3.9, and 3.10.

Cf + C52 + C3 = 1 (3.8)

C1C2—%./1—Cf 1—C;2 =0 (3.9)

44

 

c1c3+§,/1—cf,/1—c§ = 0. (3.10)

 

Original electron configuration of carbon

C*l_i LI LT

152 252 2px 2py 2pz

5p2 orbital hybridization of carbon

Csl__I Li 1 1

1s2 sz SP2 8112 p

 

 

 

Figure 3.3 A hybridization model of Sp2 orbitals.

Solving for the unknown coefficients using these conditions yields the following

1 1
results: C1 = C2 = E and C3 = - 7:3"

The atomic orbitals of the new energy states (i.e., spg, Spg, and Spg) are

centered on the atom, separated by 120°, and directed toward the corners of an equilateral
triangle as shown in Figure 3.4. These three orbitals have high electron probability, and

three electrons of the same energy.

45

 

Arrangement of Hybrid Orbitals Geometric Figures

         

Trigonal-
Planar

 

 

 

Figure 3.4 New atomic orbitals (i.e., Spg, Spa, and 81),?) resulting from mixing the
25 orbital with two of the 2p orbitals (2px, 2py, or 2pz). Source: www.chem.umass.edu’.

3.4. sp3 Hybrid Orbitalsl’z’3

In the Sp3 orbital hybridization model the 25 orbital is hybridized or mixed with all
three of the 2p orbitals (2px, 2py, and 2pz), which results in four equivalent Sp3 orbitals

as shown in Figure 3.5. Each 5p3 hybrid orbital consists of a large lobe, which points in
one direction, and a small lobe, which points in the opposite direction as shown in Figure

3.6.

46

 

II

 

Original electron configuration of carbon

o*l_1 LILT

152 252 2px 2py 2pz

sp3 orbital hybridization of carbon

C*T_IT__I ll

152 5p3 5p3 5p3 sp3

 

 

 

Figure 3.5 A hybridization model of 5p3 orbitals.

 

Arrangement of Hybrid Orbitals Geometric Figures

109.5o

   

Tetrahedral

 

 

 

Figure 3.6 New atomic orbitals (i.e., spg, spg, sp2, and spfi) resulting from
mixing the 25 orbital with the three 2p orbitals (2px, 2py, and 2pZ). Source:

www.chem.umass.edu5.

The tetrahedral bond directions from a carbon atom that is at the center can be
selected as (1, 1, 1), (-1, -1,1),(-1, l, -l), or (1, -1, -1). The combination ofthe 25 and the
three 2p atomic orbitals then can be mathematically represented by Equations 3.11, 3.12,

3.13, and 3.14.

47

lsp2> = §[12s> +12px> + 12py> +12pz>1 (3.11)

Isp§> = $1125 — 12px> - 12p,» +12pz>l (3.12)
Isp2> = -:-[I2s> — I2px> + |2py> - I2pz>l (3.13)
Isp3> = §[12s> + |ng —12py> - 12p.>] (3.14)

The atomic orbitals of the new energy states (i.e., Spg, Spg, Spg’, and Spa) are
centered on the atom, separated by 109.5°, and directed toward the comers of a regular
tetrahedron. These four orbitals have high electron probability, and four electrons of the

same energy.

In general for Sp" hybridization, there are (n+1) electrons belonging to the
hybridized 0' orbital of the carbon atom, and [4 - (n+1)] electrons belonging to the it
orbitals that originate from the unchanged p orbitals. In the case of Sp3 hybridization, four

valance electrons occupy four a bonding states. It is also possible for carbon atoms to

. . 2+ . . .
have a mixed sp2 and 5p3 character described as Sp n With 0 < n < 1. This 15 the case for

the pentagon rings found in fullerenes whose bonding is basically sp2 but whose three-
dimensional bending gives these carbon atoms a partially Sp3 character.
This can result in possible Sp2 hybridization of a planner carbon ring and

hybridization found in fullerenes.

48

 

Tl bl!-

 

 

References

 

H

. K. Vollhardt and N. Schore, Organic Chemistry: Structure and Function, (Oxford
University Press, New York, NY, 2002).

2. R. Saito, G. Dresselhaus, and M. Dresselhaus, Physical Properties of Carbon
Nanotubes, (Imperial College Press, Lomdpm, ON, Canada, 1999).

3. http://www.science.uwaterloo.ca/~cchieh/cact/c120/hybrid.html, February 2010.
4. A. Beiser, Perspectives of Modern physics, (McGraw-Hill, New York, NY, 1969).

5. http://www.chem.umass.edu, February 2010.

49

Chapter 4

Quantitative EELS Investigation of Carbon

Materials

The atomistic and/or nanoscale properties that enable improved tribological
performance of carbon onions have not yet been clearly identified. The potential
properties, which might influence the frictional performance of these nano-matelials, are:
l) the electronic interactions between individual nano-surfaces and/or with wear surfaces;
2) the individual nano-mechanical properties that affect collective load bearing; and 3)
the effect of defects on these properties.

The focus of the present work is on the electronic interactions between individual
nano-surfaces described by the sz or Sp3 bond hybridization of the carbon onion
surfaces. The spZ/sp3 bond hybridization ratio is an important fundamental tribological
property because spz-sp2 regions would be rt-electron rich and able to slide like graphene
layers, while 5p3-sp3 regions could form chemical bonds leading to atomistic stiction.

For carbon onions, there is a known topographical evolution, which iS observable
by HRTEM, from a spherical to a polyhedral multi-layer structure as a function of
increasing synthesis temperature]. It has been postulated that the structural evolution is
accompanied by a change of the spz/sp3 bond hybridization ratioz. However, this has not
been quantitatively investigated in the carbon onion nanomaterial system. In this

research, the postulated correlation of the sz/Sp3 bond hybridization ratio with the

50

 

 

 

spherical to polyhedral structural evolution is quantitatively investigated using EELS due
to its high spatial resolution (~l eV) and sensitivity to different bond structures within a

sample3 .

4.1. Quantitative Method: EELS

EELS is a sensitive analytical method that can measure a transition edge (inner
Shell to conduction band) rather than a Gaussian transition distribution3. AS discussed in
Chapter 2, there are several transition edges: K, L, and M, defined by the atomic Shell
that an ejected electron originates from. The present work utilizes the K-shell transition
of a carbon 15 electron to an available conduction band (*) energy level. As discussed in
Chapter 3, the available conduction band energy levels in an sz hybridized system are
the 1t* and the 0* levels. Therefore, the expected K-edges for a purely Sp2 carbon
nanomaterial are the 15 -> 1r* and the 15 —> 0* transition edges. Figure 4.1 shows a
typical EELS spectrum from a largely Sp2 carbon onion sample with the transition edges
marked. Each leading edge is actually a delta function, which is symmetrically broadened

by the energy resolution of the instrument.

A purely Sp2 material would have a ratio of intensities I 15 —) n4 / I 13 -) or =
1/3 because there are three 5p2 hybridized 6* energy levels for every one rt* unhybridized
energy level. A departure from this ratio means that the material is not purely sz. In

quantitative EELS, the integrated areas under the carbon K-ionization transition edges are

used to experimentally determine any departure of a test sample (U) from a known sz

51

reference sample (R) (i.e., % sz z

19/ 19
Iii/I.R

 

). However, there are several research

issues involved in the quantitative EELS measurement, which are discussed next.

 

L

Intensity (a.u.)

 

15 -—> 1t* 15 —+ 6*

K—edge at K-edge at

284.7 eV I292.0 eV
‘ I

 

 

 

290 300 310 320 330
Energy (CW

280

 

Figure 4.1 A typical EELS spectrum from a largely sz carbon onion sample,
synthesized at 1700°C, with the transition edges marked.

4.2. Quantitative EELS: Research Issues

4.2.1. Near-Edge Segregation

To precisely characterize the sz/Sp3 content using EELS, the 11:* feature and its

cross section must be isolated from the spectrum. Typical It“ and 6* peak segregation

methods have accuracy limitations due to the assumption that these features are well

52

 

 

 

separated in energy, and their cross sections can be independently obtained. However, the
spectrum edge intensities can be changed by the plural inelastic scattering and plasmon
(low-energy) losses (discussed in Chapter 2) that can shift the oscillator strengths to
higher or lower energies4. Figure 4.2, Shown below, is adapted from a theoretical
calculation of the energy-loss-near—edge structure (ELNES) for graphite, which is a
purely sp2 material". It makes the point that both the 15 —-) rt* and the 15 —r 6* transitions
have long tails and therefore significant overlap out to high energies. Although only two
transitions are involved, clear segregation of the 15 —) It“ and the 15 —) 6* components is
problematic if using an energy window or functional fitting method that does not

accurately take the higher energy contributions into account.

 

 

 

 

 

I"\
\\
.\
\ >l<

I ‘ O-

I \\

I '\ :1:
\i \‘\ 7T
ll \\

\I \\

 

 

 

Figure 4.2 A theoretical calculation of the energy-loss-near-edge structure (ELNES)
for a purely 5p2 graphite.

53

 

4.2.2. sp2 Reference Samples

Graphite and C60 are both used as sz reference (R) samples in quantitative EELS
of carbon materials or nanomaterials. Both have properties which cause them to deviate
from an ideal purely Sp2 standard. Graphite, typically highly oriented pyrolytic graphite
(HOPG), is a directional material with different electronic properties in its a-axis and c-
axis orientations. Previous research has demonstrated that this affects its interaction with
the TEM electron beams. The strongest effect is on the interaction of the n* electrons
with beam, which changes the integrated area under the 15 —> it“ Spectral curve. The
minimum integrated 15 —> 11* area is found when the electron beam is parallel to the
HOPG c-axis, a configuration referred to as the “magic angle”. The theoretical near edge
7t* and 6* spectra, shown in Figure 4.2, were from an orientation resolved calculation, in
which the graphite sample and the electron beam were at the “magic angle” during the

Monte Carlo investigation4’5.

A C50 fullerene reference sample does not have the same strong orientation
dependence as a graphite sample because the 1t* energy levels are Spatially randomized
by the spherical structure. C60 does have bond hybridization character of spz'28 rather than
sz due to its tight curvature4. However, the curvature influence affects the 15 —) it“ and
15 -) 6* integrated area about equally°, rather than preferentially affecting the 15 —-) 1t*

integrated area. A C60 fullerene reference sample is used in the present work.

4.2.3. Inclusion of the 287.0 eV Peak

54

THEé)

 

The most extensive literature on quantitative EELS of carbon materials is for the
amorphous tetrahedral carbon (a—C) system. The spz/sp3 ratio is an important fundamental
measurement in the a-C system, although by reason of its correlation with hardness rather
than low friction.

Investigation of the hybridized chemical bonding of a-C samples revealed the two
major sp2 carbon edges in the EELS spectra. The first edge of the carbon K-edge spectra
corresponds to the 15 —) 1t* transition at 284.7 eV, and the second prominent edge
corresponds to the 15 —> 0* transition at 292.0 eV. The inclusion of an intermediate peak
at about 287.0 eV has been a subject of discussion. The interpretation of the feature at
287.0 eV has been variously interpreted as a peak due to a C-H electronic transition7, a
peak due to an electronic transition within a “pseudo-molecular” carbon domain
embedded within an a-C matrix8, and as part of the 15 —-> 1t* edge transition itself‘. In the
present research, the temperature evolution of the carbon onion system was used to test

the need for a separate 287.0 eV peak.

4.3. EELS Deconvolution Techniques

4.3.1. Two-Energy Windows Method

The two-energy windows method is a simple technique used to characterize the
spz/sp3 content within a carbon sample. In this method, two-energy windows (AE,r and
ABC) are arbitrarily selected and placed at points in the EELS spectrum that appear to
represent the separation of the 15 —> rt* from the 15 -> 6* transition. For example, a

typical energy window selection is 282-288 eV for the 15 —> 1t* and 284-310 eV for the

55

True-n.

 

ls —) 6*. These choices are based on the assumption that the transition can be completely
separated from the 6* transition. Then, the integrated areas of both energy windows are
determined, and their ratio is compared with a reference ratio to deduce the Sp2 or sp3
percentage. This technique strongly depends on the selection of the widths of both energy

windows, and it can result in an uncertainty of about 10% in the sp2/sp3 ratio4.

4.3.2. Functional Fitting Method

The functional fitting method is another technique that uses different distribution
functions to deconvolve the EELS spectrum. In this technique, two or three peaks are
typically utilized to deconvolve the carbon EELS spectrum. The lowest fitted curve at
about 285.0 eV corresponds to the n* feature, and the highest fitted curve at about 8 eV
above the rt* onset corresponds to the 6* feature. The inclusion of a third intermediate
fitted curve at about 287.0-288.0 eV is an area of discussion. To determine the sp2
fraction in a carbon sample using this technique, the integrated areas of the fitted edges
are calculated and compared with a reference ratio to deduce the Sp2 percentage as given

by Equation 4.1”,

2 IU/ 1"
% Sp == "ii/T3 (4.1)
7! 0

where I}: is the integrated area under the fitted curve corresponding to the 1r* transition

of the unknown material, I}; is the integrated area under the fitted curve corresponding to

56

the 6* transition of the unknown material, IE is the integrated area under the fitted curve

corresponding to the 1r* transition of the reference material, and I5 is the integrated area

under the fitted curve corresponding to the 6* transition of the reference material. The
sp3 fraction in the sample can be determined by subtracting the sp2 percentage from 100%
(i.e., %sp3 = 100% — %sp2). The Sp2/8p3 ratio can be then calculated by dividing the Sp2

%sp2
%sp3

 

)9-

fi'action by the Sp3 fraction (i.e., Sp2/5p3 ratio =

The functional fitting method again makes the assumption that the 1r* and 6*
transitions can be completely separated. Also, all the examples reviewed in this research
used symmetric Gaussian or Lorentzian functions to fit the spectral data within a typical
282-310 eV total energy range. This results in over-representation of the near edge

contributions and under-representation of the higher-order tail.

4.4. Discussion of Issues and a New Approach

The two-energy window and the functional fitting methods both make the
assumption that the rt* transition can be separated from the 6* transition within the near-
edge 282-310 eV region. Functional fitting methods also rely on 2- or 3-peak
deconvolution using symmetric Gaussian and Lorentzian fiinctions over the same energy
range. While these functions accurately reproduce the instrumentally broadened delta
function edge, they do not effectively represent the higher energy oscillator strengths of

the 15 —> 1t* and 15 —> 6* transitions.

57

THE-.3

 

 

 

One solution is to represent the higher energy contributions with a large number
of additional Gaussian or Lorentzian peaks, then sum them appropriately to represent the
Is —) rt* and 15 —> 6* transitions. It would be difficult to identify which peaks belong
with which transition. A new approach in this thesis is to use a fitting firnction that can

more accurately reproduce the higher energy asymmetric transition characteristics.

58

 

That

if.)

References

 

l. S. Tomita, T. Sakurai, H. Ohta, M. F ujii, and S. Hayashi, “Structure and electronic
properties of carbon onions”, The Journal of Chemical Physics 114: 7477 (2001).

2. S. Tomita, M. Fujii, S. Hayashi, and K. Yamamoto, “Electron energy-loss
Spectroscopy of carbon onions”, Chemical Physics Letters 305: 225-229 (1999).

3. C. Brundle, C. Evans, and S. Wilson, “Encyclopedia of Materials Characterization”,
Butterworth-Heinemann, Stoneham, MA (1992).

4. J. Titantah, and D. Lamoen, “Technique for the sp2/Sp3 characterization of carbon
materials: Ab initio calculation of near-edge structure in electron-energy—loss
spectra”, Physical. Review B 70, 075115 (2004).

5. A. Papworth, C. Kiely, A. Burden, S. Silva, and G. Amaratunga, “Electron-energy-
loss spectroscopy characterization of the sp2 bonding fraction within carbon thin
films”, Physical Review B 62: 12628—12631 (2000).

6. J. Martins, N. Troullier and J .Weaver, “Analysis of occupied and empty electronic
states of C60”, Chemical Physics letters 180: 457-460 (1991).

7. Mykhaylyka, Y. Solonin, D. Batchelder, and R. Brydson, “Transformation of
nanodiamond into carbon onions: A comparative study by high-resolution
transmission electron microscopy, electron energy-loss spectroscopy, x-ray
diffraction, small-angle x-ray scattering, and ultraviolet Raman spectroscopy”,

Journal of Applied Physics 97: 074302 (2005).

8. R. Brydson, Z. Zhili, and A. Brown, “Revisiting the determination of carbon sp2/Sp3
ratios of the EELS carbon K-edge”, EMC 2008 14th European Microscopy Congress
1: 357-358 (2008).

9 . S. Osswald, G. Yushin, V. Mochalin, S. O. Kucheyev, and Y. Gogotsi, “Control of

sp3/sp2 carbon ratio and surface chemistry of nanodiamond powders by selective
oxidation in air”, Journal of American Chemical Society, 128: 11635-11642 (2006).

59

Chapter 5

Quantitative EELS Investigation of Carbon

Onions

All of the considerations in the previous chapter had to be separately considered
in order to perform the quantitative EELS investigation of the new carbon onion system
that is one of the main contributions of this thesis. In this research, different functional
fitting techniques were utilized to precisely reproduce the characteristics of the leading
edge, which is a delta function symmetrically broadened by the energy resolution of the
instrument, and the higher energy states, which are the asymmetric contributions from all
available energy levels in the energy density of stateS'. The techniques, used in this thesis
to analyze the EELS spectrums of carbon onions, represent a new research approach for
the carbon onion system and for carbon nanomaterials.

The EELS investigation is set within the context of HRTEM and SEM
experiments that correlate the fundamental electronic sp2/5p3 investigations with the

synthesis conditions.

5.1. Structural Characterization Using HRTEM

HRTEM images confirmed that the previously reported2 Spherical to polygonal
structural transition occurs over a temperature range of 1700°C to 2300°C for the present

synthesis conditions. No evidence for a remnant NCD core was observed in the carbon

60

’HE:

onion samples. Therefore, the HRTEM images indicated that the NCD starting material,
shown in Figure 5.1, had been completely converted into the innermost carbon onion
layers. Representative images of the spherical to polygonal structural transition are shown

in Figure 5.2 (a-c).

 

Figure 5.1 An HRTEM image of the largely Sp3 NCD. Image by: Raied A. A1-
Duhileb.

61

10 nm

 

Figure 5.2 HRTEM images of the structural transition from spherical to polygonal for
carbon onions synthesized at temperatures (a) 1700°C (b) 2000°C and (c) 2300°C. Images
by: Raied A. Al-Duhileb.

62

 

 

THE:

 

For quantitative EELS analysis, C60 firllerenes were used as a non-directional3
reference standard having a theoretical intensity ratio (i.e., I1}: / lg) of one-third. A group
of C50 fullerenes suspended over the edge of a lacey film hole is shown in Figure 5.3. The
individual cages are difficult to distinguish due to light contrast and small size; therefore,

care was taken to investigate only groups suspended well over the edge of a lacey film

hole.

 

Figure 5.3 A TEM image showing aggregate C60 particles suspended over a lacey
film hole. Image by: Raied A. Al-Duhileb.

In the coming sections, we quantitatively measure the sp2/5p3 bond hybridization
ratio of different carbon onion samples prepared at an increasing synthesis temperature
using several EELS deconvolution techniques. We also investigate (a) whether inclusion
of a 287.0 eV feature is required in a 2-peak fit by a symmetrical deconvolution function;
and (b) whether inclusion of a 287.0 eV feature in a 3-peak fit by a symmetrical

deconvolution function produces a good fit.

63

5.2. EELS Deconvolution: Gaussian Function

Approach 1 (Gaussian [auction with Z-peak fitting):

The core loss EELS spectrums of carbon onions, prepared at increasing synthesis

temperature, were analyzed using symmetric Gaussian area deconvolution function:

 

_ 30 ex _l(X-31)2
y— ,iznaz p 2 a2 ’ (5'1)

where a0 is the area, a1 if the center, and a2 is the width.

The results, which were obtained using 2-peak fitting, are Shown in Table 5.1. A
typical deconvolution using this method is shown in Figure 5.4. No clear increasing or
decreasing sz/sp3 trend was observed using a Gaussian function with 2-peak fitting
(Figure 5.5). Furthermore, in at least one instance (row identified with bold border), a 3-
peak fit that included a peak at about 287.0 eV was required to obtain a 0.989 coefficient

of determination (COD).

64

 

 

 

Table 5.1

sz/Sp3 ratio of carbon onion samples prepared at increasing synthesis
temperature. The results were obtained using Gaussian deconvolution and 2-peak fit.

 

 

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

Temp- C013 I,r Io %sp2 %sp3 802/803
1700°C 0.998 28384.32 106570.00 0.64 0.36 1.76
1700°C 0.989 38044.33 160220.00 0.85 0.15 5.56
2000°C 0.992 64570.15 359260.00 0.46 0.54 0.86
2000°C 0.997 81883.80 484740.00 0.44 0.56 0.78
2000°C 0.998 288310.00 813370.00 0.99 0.01 99.00
2300°C 0.998 1 14530.00 419090.00 0.65 0.35 1.86
2300°C 0.993 142900.00 565360.00 0.61 0.39 1.57
Original EELS
, Spec rum
18000 -
16000 ..
r4000 - ,
A i Generated Flt
=3 12000 -
3 10000 -
g 8000 - lm Fit at~285 eV I 30d Fit “.492 ev
g, 6000 _
,5 4000 -
2000 4
0 1
-2000.,...,.,.,.,.,.,.,.
278 280 282 284 286 288 290 292 294 296
Energy (eV)
Figure 5.4 Deconvolution of the core loss EELS spectrum of carbon onions, prepared

at 1700°C. The original spectrum was deconvolved using 2-peak fit and the symmetric
Gaussian deconvolution.

65

 

 

 

 

1.0 -
0.8 —
G
E
0.6 —
a:
n
:1.
$ 0.4 ~
:1.
m
0.2 -
I
0.0 . I
. .
r700

, r
1800

1300‘ 2000' 21'00
Temperature (°C)

., . , .
2200 2300

 

 

Figure 5.5 A plot of the sp2/5p3 ratio in carbon onions as a function of synthesis
temperature. The plot is based on results obtained using 2-peak fit and the symmetric

Gaussian deconvolution.

Approach 2 (Gaussian [unction with 3-peak fitting]:

The results of 3-peak fitting using symmetric Gaussian deconvolution with

inclusion of a third feature at 287.0 eV (rows identified with bold borders) are shown in

Table 5.2. A typical deconvolution using this method is shown in Figure 5.6. Although all

samples had CoD of 0.97 or above using the 3-peak symmetric fit, no clear increasing or

decreasing sp2/5p3 trend was observed (Figure 5.7) for the Gaussian deconvolution with

3-peak fitting.

66

 

 

 

 

 

 

 

 

 

 

 

  

  

 
  
  

 

 

 

 

 

 

 

Table 5.2 sp2/5p3 ratio of carbon onion samples prepared at increasing synthesis
temperature. The results were obtained using Gaussian deconvolution and 3-peak fit.
Temp. COD 11: 10. "/osp2 "/osp3 sp2/5p3
1700°C 0.990 23974.79 32161.78 2.22 -1.22 -1.82
1700°C 0.981 45052.71 22508.18 2.27 -1.27 —1.79
2000°C 0.974 67397.28 81505.44 1.96 -0.96 -2.04
2000°C 0.995 109060.00 36756.27 2.42 -1.42 -1.70
2000°C 0.999 256560.00 58267.76 0.99 0.01 99.00
2300°C 1.000 92994.61 27554.33 2.47 -1.47 -1.68
2300°C 0.993 133940.00 38433.95 2.48 -1.48 -1.68
‘ OriginalEELS
30000 .. Specltrum
25000 4
G .
:3 20000 4
3 l Generated Fit
g 150"“ ‘ 1“ Fit at ~ 235 e 12"" Fit at ~ 292 eV
= 1 ‘
8 10000 -
.5 . I
5000 . \. (WV
o.‘ ~ ..
278 Y 280 f28'2 284 '286 12818 290 r 292 *29-4 I 296 -
Energy (eV)
Figure 5.6 Deconvolution of the core loss EELS spectrum of carbon onions, prepared

at 1700°C. The original spectrum was deconvolved using 3-peak fit and the symmetric

67

 

THES

 

 

 

,
1.0- I
0.8-
E
a 0.6-
:1:
”=- 04-
$ ' .
Q-
“ 0.2~
0.0“ I I I
I ' f T I Y r

 

I v v , r I 1
1700 1800 1900 2000 21 2200 2300
Temperature (°C)

 

 

 

Figure 5.7 A plot of the sz/sp3 ratio in carbon onions as a function of synthesis
temperature. The plot is based on results obtained using 3-peak fit and the symmetric
Gaussian deconvolution.

5.3. EELS Deconvolution: Lorentzian Function

Approach 1 (Lorentzian area [auction with Ed starting and end points 2:

To represent the experimental data, a 3-peak functional fitting was performed

using symmetric Lorentzian area function:

 

(5.2)

 

where a0 is the area, a; if the center, and a2 is the width.
Two appropriately placed peaks represented the 1r* and 6* edges, and a third peak

at 287.0 eV was included. The results of the three-peak Lorentzian deconvolution are

68

 

THC:

 

shown in table 5.3. A typical deconvolution using this approach is shown in Figure 5.8.

The results had a coefficient of determination (COD) of 0.95 or above. The data in each

temperature displayed a range of values, and no clear increasing or decreasing sp2/sp3

trend was observed (Figure 5.9).

Table 5.3

sp2/sp3 ratio of carbon onion samples prepared at increasing synthesis
temperature. The results were obtained using symmetric Lorentzian area deconvolution
with 3-peak fit and fixed starting & end points.

 

 

 

 

 

 

 

 

Temp. C0D In [0 %sp2 "/osp3 Sp2/8p3
1700°C 0.976 16163.17 1 14220.00 0.53 0.47 1.13
1700°C 0.970 26625.66 194910.00 0.45 0.55 0.81
2000°C 0.961 44847.17 379460.00 0.35 0.65 0.55
2000°C 0.952 75468.82 490150.00 0.42 0.58 0.72
2000°C 0.982 1 12210.00 904160.00 0.60 0.40 1.51
2300°C 0.954 57657.10 447680.00 0.50 0.50 0.99
2300°C 0.961 72822.98 616480.00 0.43 0.57 0.75

 

69

 

 

THE:

 

 

  
 
  
   
  
  

 

 

 

 

 

OriginalEELS
20000.1 Spectmm
13.999. -
1699.0. .
140001
A .
5' 12000 . Generated Flt
v 10000 ,
:3 8000 . 12"" Fit at~292 eV
§ 60001 1“Fitat~2sse / ’
r
5 4000 , I ) ’3rdFitat~287eV
20001 f [V \
-2000"T'i'I‘T‘V'I'I‘T'I‘
273 280 282 284 286 288 290 292 294 296
Energy (eV)

 

 

 

Figure 5.8 Deconvolution of the core loss EELS spectrum of carbon onions, prepared
at 1700°C. The original spectrum was deconvolved using 3-peak fit and the symmetric
Lorentzian area deconvolution.

 

2L6

la4

L2

1.0 .

sp2/sp3 Ratio

(L8 '

(L6

 

0'4 1 V 1 r I ' I l ' I * T ‘
1700 1800 1900 2000 2100 2200 2300

Temperature (°C)

 

v

 

 

 

Figure 5.9 A plot of the sp2/3p3 ratio in carbon onions as a fimction of synthesis
temperature. The plot is based on results obtained using approach 1.

70

Approach 2 (Lorentzian area tunction with exact inflection points 2:

The 11* and 6* edges were located within about :1:2 eV by using the inflection

 

points within the appropriate 71* and 6* data ranges to identify an edge. The start of the
functional fit for the II:* transition was consistently set at 5 eV below the 71* edge
identified by its inflection point. The finish of the fimctional fit for the 6* transition was
consistently set at 5 eV above the 0* edge identified by its inflection point. The goal was
to compensate for the observed overall shifting of the experimental data due to the zero-
loss peak being not precisely set at zero. The inflection points and their differences are
shown in Table 5.4. The differences fall in an expected 6-7 eV range. The results of the
three-peak Lorentzian deconvolution are shown in Table 5.5

The data in each temperature displayed a range of values, and no clear increasing or

decreasing sp2/Sp3 trend was observed as shown in Figure 5.10.

 

 

 

 

 

 

 

 

 

Table 5.4 First and second inflection points (IPs) and their corresponding

differences.
Temp. 1“t [P 2"d [P A (IPS) Fit Start Fit Finish
1700°C 283.90 eV 290.44 eV 6.54 eV 279 eV 295 eV
1700°C 283.96 eV 289.98 eV 6.02 eV 279 eV 295 eV
2000°C 278.94 eV 286.10 eV 7.16 eV 274 eV 291 eV
2000°C 283.42 eV 289.58 eV 6.16 eV 278 eV 295 eV
2000°C 283.16 eV 289.70 eV 6.54 eV 278 eV 295 eV
2300°C 280.23 eV 287.66 eV 7.43 eV 275 eV 293 eV
2300°C 283.86 eV 290.37 eV 6.51 eV 279 eV 295 eV

 

7l

THE!

Table 5.5

sp2/Sp3 ratio of carbon onion samples prepared at increasing synthesis
temperature. The results were obtained using symmetric Lorentzian area deconvolution
with 3-peak fit and exact inflection points.

 

 

 

 

 

 

 

 

 

 

 

Temp. COD 11: 10. %sp2 °/osp3 sp2/sp3
1700°C 0.957 14822.24 113550.00 0.46 0.54 0.86
1700°C 0.962 29665.54 190820.00 0.47 0.53 0.88
2000°C 0.961 46691.33 381880.00 0.38 0.62 0.61
2000°C 0.942 72458.07 485390.00 0.44 0.56 0.80
2000°C 0.975 217300.00 862270.00 0.72 0.28 2.54
2300°C 0.967 77822.95 438060.00 0.60 0.40 1.51
2300°C 0.961 89663.65 614300.00 0.47 0.53 0.89
2.5 , I
a 2.0 .
'5
a 4
a:
n 1.5-I I
3 .
"a.
m 1.0 «
I
‘ I
I
0.5 .
1200 . 1500-1930 . 2000 T 21100 - 2200 . 2300 ‘

 

Temperature (°C)

 

 

Figure 5.10 A plot of the sp2/5p3 ratio in carbon onions as a function of synthesis
temperature. The plot is based on results obtained using approach 2.

5.4. EELS Deconvolution: F-Variance Function

72

The core loss EELS spectrums of carbon onion samples, prepared at increasing

synthesis temperature, were deconvolved using an asymmetric f-variance deconvolution

 

 

function:
iii- 1
a F [a__3+ a4] [: 3]a— 2 3[X_-__2+ 31+ a4 (33- 2)] 2
__ 0 a3 (34.-|- 2)
Y — —a3 — a4 , (5.2)

a (x-al a4(33-2)) 2
3 32 a3(a4+2)

32F i23111121] 1+ a4

 

 

where an is the area, a; if the center, a; is the width, a3 is shape 1, and a4 is shape 2. As
shown in Figure 5.11, the f-variance area function can accurately represent both the

Ileading edges and higher energy contributions of the ls ——) 11* and Is —> 0* transitions.

73

 

 

 

 

6*

 

 

 

 

)1

Figure 5.11 Theoretical near edge 11* and 6* spectra fiom an orientation resolved
calculation, in which a graphite sample and the electron beam were at the “magic angle”.

 

74

 

 

The results, which were obtained using this method with 2-peak fitting, are shown in

Table 5.6. A typical deconvolution using this method is shown in Figure 5.12. The results

had a 0.97 CoD or above, and showed a clear trend, which was an increase in the sp2/Sp3

bond hybridization ratio as shown in Figure 5.13.

Table 5.6

sp2/Sp3 ratio of carbon onion samples prepared at increasing synthesis

temperature. The results were obtained using asymmetric f-variance area deconvolution

and 2-peak fit.

 

 

 

 

 

 

 

Temp. C0D 111 10 "/..sp2 %sp3 Sp2/Sp3
1700°C 0.999 39547.39 95377.24 0.89 0.1 1 7.94
1700°C 0.993 65823.27 156490.00 0.90 0.10 8.73
2000°C 0.978 1 17590.00 295790.00 0.86 0.14 6.25
2000°C 0.998 125260.00 43 7990.00 0.67 0.33 2.07
2000°C 0.997 369690.00 728380.00 0.99 0.01 99.00
2300°C 0.999 166230.00 358740.00 0.96 0.04 23.71
2300°C 0.996 225710.00 479510.00 0.97 0.03 32.19

 

75

 

Original EELS
Spectrum

   
  

 

 

 

 

; 1400 ‘ Generated Fit
3 . 1st Fit at~285 eV
I
53’ 12"" Fit at ~ 292 eV
m 7000 ,. <
1:
O
H
5 1
0 “ ~ -
282 288 294
Energy (eV)

 

 

 

Figure 5.12 Deconvolution of the core loss EELS spectrum of carbon onions, prepared
at 1700°C. The original spectrum was deconvolved using 2-peak fit and the asymmetric f-
variance function.

 

 

 

 

 

1.0 r I
1
0.8 .
c .
'f’. 06
a: ' '
n 1
a.
“x” 0.4 1
N
D. u
m ----- a
0.2 - ----------------
. ‘ .............. .
0.0 . n
1700 1800 1900 2000 2100 2200 2300
Temperature (°C)

 

Figure 5.13 A plot showing an increase in the sp2/5p3 ratio in carbon onions as a
function of synthesis temperature. The results from the 2000°C were variable due to
synthesis uniformity issues.

76

THE:

Only carbon onion samples prepared at 2000°C continued to display a wide range
of values. Information from a group member4 who participated in the synthesis
experiments indicated that synthesis uniformity issues could be an issue at the 2000°C

synthesis temperature.

5.4.1. Investigation of Synthesis Uniformity by

SEM

SEM images shown in Figure 5.14 indicated that synthesis uniformity issues may
be present. Synthesis was fairly uniform for carbon onions grown at 1700°C. However,
carbon onions grown at 2000°C showed evidence of microcrystallite graphite formation
in addition to carbon onion synthesis. Carbon onions grown at 2300°C showed better
uniformity than those grown at 2000°C but less uniformity than samples grown 1700°C,

with some microcrystallite graphite formation observed.

 

Figure 5.14 SEM surface images of carbon onions synthesized at temperatures (a)
1700°C (b) 2000°C and (c) 2300°C. The inset in (b) is a TEM image. Images by: Raied A.
Al-Duhileb.

77

THE:

Y\\

 

 

References

 

1. R. Egerton, “Chapter 3: Electron Scattering Theory,” Electron Energy Loss
Spectroscopy in the Electron Microscope (Plenum Press, New York, NY, 1986) pp.
129-228.

2. S. Tomita, T. Sakurai, H. Ohta, M. Fujii, and S. Hayashi, “Structure and electronic
properties of carbon onions”, The Journal of Chemical Physics 114: 7477 (2001).

3. J. Titantah, and D. Lamoen, “Technique for the sp2/Sp3 characterization of carbon
materials: Ab initio calculation of near-edge structure in electron-energy-loss
spectra”, Physical. Review B 70, 075115 (2004).

4. Benjamin W. Jacobs, Ph.D. during NSF DMI-0631978, IREE: Nano-Mechanical and
Electronic Investigations with Tokyo Institute of Technology (2007).

78

Chapter 6

Conclusion and Future Work

6.1. Discussed Research Conclusion

Qualitative research on the fundamental electronic properties of carbon nano-
materials, such as a-C, carbon onions, C60, carbon nanotubes, and other related fullerenes
have been widely reported in the literature; however, the firndamental characteristics of
these nano-materials have been less quantitatively studied. In this research, we
quantitatively investigated the sp2/Sp3 bond hybridization ratio in different carbon onion
samples using EELS due to the anticipated influence of this electronic property on the
frictional performance of carbon onions. We also developed a new approach to accurately
investigate the sp2/Sp3 content within a carbon sample. The previous quantitative
investigations reported in the literature are based on symmetrical Gaussian or Lorentzian
deconvolution functions. In this thesis, EELS spectrums were analyzed using an
asymmetric f-variance deconvolution function to reproduce the characteristics of the
leading edges.

The quantitative experimental results, analyzed using the asymmetlic f-variance
deconvolution function, showed a clear increase in the sp2/Sp3 bond hybridization ratio in
carbon onions. Analysis by symmetric Gaussian and Lorentzian deconvolution functions

did not reveal clear increasing or decreasing trends.

79

Carbon onion samples prepared at 2000°C displayed a wide range of values.
These were traced to synthesis uniformity issues using TEM, SEM and, as discussed

below, micro Raman spectroscopy.

6.2. Continuing Fundamental Studies: Raman

Spectroscopy

The bonding nature of carbon onions can be studied using Raman spectroscopy.
This technique sets carbon onions into vibrational modes (phonons), which depend on the
atomic mass and bond force constants of carbon atoms, using a laser beam (photons).
Preliminary investigations, shown in Figure 6.1 and Table 6.1, provided an independent
confirmation of the quantitative EELS results. The carbon onion temperature series
samples were investigated by room temperature micro Raman spectroscopy at a
wavelength of 532 run. As carbon onions are multi-shell fullerenes, C60 was also
investigated as a control. A nearly 50% decrease in the FWHM of the G peak was
observed for carbon onion samples synthesized at 1700°C versus 2300°C. Also, the
intensities of both the D and G peaks were greatly reduced in the 2000°C sample,
providing further indication of sample uniformity issues. All three carbon onion samples
displayed a broad peak at about 985 cm". This is tentatively identified as a disordered
amorphous carbon peak. We note that the intensity of this peak was greatly reduced in the
2300°C sample.

These preliminary micro Raman investigations were conducted at room

temperature. Low temperature (77K) Raman spectroscopy of multi-walled carbon onions,

80

which would be resolved into a series of clear peaks, would be very beneficial for
examining their vibrational modes in extreme environment and correlating the results

with their tribological performance.

 

Carbon Onions
2300“C

 

Carbon Onions
2000°C

 

Carbon Onions
1700°C

Counts (a.u.)

"VV'U'V

 

 

IAIAJAIAIJIAJAI

0 250 500 750 1000 1250 1500 1750 2000
Raman Shift (cm")

Figure 6.1 Experimental Raman spectra of C50 and carbon onion samples prepared at
increasing synthesis temperature.

 

 

 

81

Table 6.1 Raman spectroscopy of C60 and carbon onions at wavelength 532 11111.
CO CO CO

 

 

 

 

C6“ 1700°C 2000°C 2300°C
Pea} 271.5
(“m ) 492.1 .
991.5 984.6 984.9
1340.3 1347.2 1348.0
1368.0
1462.7
1586.0 1579.2 1579.4
1593.1
F -1 45.9
(cm ) 16.7
1837.4 1720 1592b
46.5 24.1a 42.7
214.8
14.9
43.1 24.83 21.9
105.8
Tentative Hg(1)
11’ Ago)
D D D
Hg(l)+Hg(5)
Age) G G G
13(1)+H2(5)
ID Reference 1 2 2 2

 

a. Although listed in Table 6.1, the intensities of these two peaks were very weak.
b. The intensity of this peak was very weak.

6.3. Continuing Application Studies: Irradiated

Carbon Onions

82

Carbon onions have demonstrated excellent tribological performance in air and
vacuum environments. Lubricants that can function well in a vacuum environment are
greatly needed for moving mechanical parts in space, e. g., for solar panel and robotic arm
deployment and retraction. In addition to being an extreme vacuum environment, space is
a high radiation environment. High radiation environments are also found in other
application areas, such as particle colliders and nuclear reactors. Therefore, experimental
and theoretical investigations can be conducted to explore the radiation resilience of
carbon onions, pending their application in extreme radiation environments including
space, which is also a vacuum environment.

Experiments to irradiate carbon onion samples using highly ionized beams are
ongoing. The carbon onions are being exposed to heavy ion irradiation at the National
Superconducting Cyclotron Laboratory (N SCL) at Michigan State University and their
post-radiation properties are being characterized using the quantitative approaches
developed in this research. Experiments to date have been performed with primary beams
Oxygen-16, Argon-40 and Calcium-48. Irradiations for a cumulative dose of 10,000 Gray
(Joule/kg) are performed at 140 MeV/nucleon and 70 MeV/nucleon to test the influence
of charge-to-mass ratio on coupling between heavy ion species and carbon onions, while
maintaining energy conditions comparable to those encountered in space, nuclear
reactors and particle collider environments.

Results to date indicate that carbon onions may respond to heavy ion irradiation
with self—healing and self-annealing reactions that are parallel to those observed at
increased synthesis temperatures. HRTEM images of pre-radiation carbon onions

synthesized at 1700°C are shown in Figure 6.2. The image, shown in (a), displays carbon

83

onions with typical spherical morphologies, with some defects observable in the atomic
planes. The image, shown in (b), displays a feature of two fused onions. This feature

could be found at 1700°C but it was rarely observed.

 

Figure 6.2 HRTEM images of pre—radiation carbon onion samples prepared
at1700°C. Typical spherical morphologies are observed in (a). A fused onions feature is
observed in (b). Images by: Xudong Fan, with Kaylee McElroy, and Virginia M. Ayres.

HRTEM images of carbon onions synthesized at 1700°C and irradiated with a
10,000 Gray dose of fully stripped Calcium-48 heavy ion are shown in Figure 6.3. We
first note that the structural integrity of the carbon onions was largely maintained. The
results of irradiation at 70 MeV/nucleon (lefi hand side) included a greater frequency of
fused onions (a), and an increase in the polygonal nature of the atomic layers (b). Areas
such as (c) showed some evidence for conversion into another form of carbon such as
graphite or amorphous carbon. The results of irradiation at 140 MeV/nucleon (right hand

side) showed a similar increase infused and polygonal onions (d and e) but also showed

84

more evidence of defective atomic planes in both. Furthermore, evidence for conversion

of carbon onions into defective amorphous carbon in local areas was clearly observed (1).
When these results and other are quantified using the methods developed in this research,
it will provide a quantitative description of energy deposition in the carbon onion system

during irradiation by heavy ions.

85

   

5 11m
_

Figure6.3 HRTEM images of carbon onion sarnepls prepared at 1700°C following
irradiation with fully stripped Calcium-48 heavy ions at energies (a-c) 70 MeV/nucleon
and (d-e) 140 MeV/nucleon. Images by: Xudong Fan, with Kaylee McElroy, and Virginia
M. Ayres.

86

6.4. Continuing Application Studies: Tribological
Performance of Carbon Onions in Harsh

Environments

To eliminate the frequent operational failures and serious incidents associated
with the performance of the conventional lubricant approaches, new nano-particle solid
based lubricants are currently under investigation for use in both ordinary and harsh
environments. Nano-particulate molybdenum and tungsten disulfides have been
successfirlly used for improving the frictional performance of these nano—particle solid
based lubricants in terrestrial applications; however, the environmental impact of nano-
sized heavy metal derivatives is still under investigation. Nano-carbon materials such as
carbon onions, carbon nanotubes, C60, and related fullerenes are also under investigation
as environmentally benign alternatives. The best overall lubrication performance across
air, vacuum and polymer environments to date has been achieved with carbon onions3.

Although preliminary tribological studies of carbon onions showed an improved
frictional performance in air and vacuum, further investigations are desired to examine
their frictional behavior in harsh environments seen in terrestrial industries and space. In
addition, future experimental investigations should consider assessing the tribological
performance of different carbon onion samples prepared at increasing annealing
temperatures, and studying the effects on their outer structures post performing the

frictional assessment experiments.

87

 

 

In.

 

References

 

1.

Z. Dong, P. Zhou, J. Holden, P. Eklund, M. Dresselhaus, and G. Dresselhaus,
“Observation of higher-order Raman modes in C60 films”, Physical Review B 48:
2862—2865 (1993).

S. Tomita, T. Sakurai, H. Ohta, M. Fujii, and S. Hayashi, “Structure and electronic
properties of carbon onions”, The Journal of Chemical Physics 114: 7477 (2001).

A. Hirata, M. lgarashi, and T. Kaito “Study on solid lubricant properties of carbon
onions produced by heat treatment of diamond clusters or particles”, Tribology
International 37 : 899—905 (2004).

88

TI-

IIILI a