FABRICATION OF GRAPHITE AND GRAPHENE-BASED MATERIALS … · 2018-01-10 · fabrication of graphite...

FABRICATION OF GRAPHITE AND GRAPHENE-BASED

MATERIALS FOR FIELD EMISSION APPLICATIONS

KOH TING TING ANGEL

(B. Appl. Sci. (Hons.) National University of Singapore)

A THESIS SUBMITTED

FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

DEPARTMENT OF MATERIALS SCIENCE &

ENGINEERING

NATIONAL UNIVERSITY OFSINGAPORE

2012

Declaration

i

DECLARATION

I hereby declare that this thesis is my original work and it has been written by me in its

entirety. I have duly acknowledged all the sources of information which have been used

in the thesis.

This thesis has also not been submitted for any degree in any university previously.

Koh Ting Ting Angel

28th

August 2012

Acknowledgements

ii

ACKNOWLEDGEMENTS

Embarking on the PhD journey has indeed been the most challenging task I have

taken in the 26 years of my life. Through this journey, I have gotten to know a handful of

wonderful people that have made an impact in my graduate life. It has been a privilege and an

honor to work with these brilliant and helpful people and they will always remain dear to me.

First, I would like to express my deepest gratitude to my supervisor Dr. Daniel Chua

who is my first teacher of research when I first started research work five years ago. He has

not only provided me with invaluable guidance and advice necessary to see me through this

program but at the same time gave me the freedom to pursue independent work. I sincerely

thank him for his patience, understanding and consolation during the bad times when nothing

seems to go right. I am also grateful to him for the numerous overseas opportunities that he

has given me, from which I have gained many priceless experiences from. I would also like

to thank him for putting up with my complaints and for tolerating the many lame jokes that I

uncontrollably crack all the time during our serious discussions. Through him, I have learnt

many lessons about research and life.

I am also grateful to Prof. Li Yi, Prof. Ding Jun, Dr. Chiu Cheng-Hsin and Dr. Stefan

Adams from whom I have acquired fundamental knowledge that I used as a basis for my

research ideas.

I would like to express my sincerest thanks to Mr. Tan Choon Wah from Physics

workshop. I am grateful for his help and support all these years in realizing my amateurish

technical design drawings into physical objects and for the many valuable life lessons that he

has enlightened me with. Without his help to fix broken equipment, my experiments would

Acknowledgements

iii

not have been possible. I would also like to thank Mr. Chen Gin Seng from the Department

of Physics for his help and guidance. In addition, I would like to thank Mr. Wong How

Kwong from the Department of Physics for his guidance and for providing new furniture

when the old one breaks for my lab in Science faculty. Despite the tiny size of my lab, he has

not forgotten about us and welcomed me as part of the Physics family all these years.

I would also like to extend my appreciation to the lab officers of DMSE especially,

Ms. Agnes Lim, Mr. Henche Kuan, Mr. Chan Yew Weng, Ms. Serene Chooi, Ms. He Jian

and Mr. Chen Qun their help and teachings all these years. Special thanks go out to Dr.

Zhang Jixuan for her patience and helpfulness.

My graduate student life would be less fulfilling if it was without the ups and downs I

shared with the following people. I am deeply grateful to my research group senior, Yuan

Mei. I would like to thank her for providing me with exceedingly useful discussions where

we bounced ideas off each other and for sharing my joy and horror moments upon the

revelation of a new discovery or some mistakes. I would also like to thank Jovan Hsieh, my

first mentor in research who has undoubtedly played an important part in my embarkation

towards a graduate research career. She has my thanks for giving me her moral support and

for being my listening ear since day one. I will not forget her kindness when she gave her

experiment slot to me when I was pressed for time. Special thanks go to Su Ru, Wang

Hongyu, Yu Jun, Tang Zhe, Le Quang Tri and Dr. Chen Ting, who have help me in one way

or another during my arduous PhD course of study.

Last but not least, I would like to express my heartfelt gratitude to my family and

especially my mum, for their unconditioned love and support in my research endeavor.

August 2012 Angel Koh

Table of Contents

iv

Table of Contents

DECLARATION i

ACKNOWLEDGEMENTS ii

TABLE OF CONTENTS iv

SUMMARY viii

LIST OF TABLES x

LIST OF FIGURES xi

LIST OF ABBREVATIONS xv

LIST OF PUBLICATIONS xviii

CHAPTER 1 INTRODUCTION 1

1.1 Graphite and monolayer graphite 1

1.2 Properties of graphene 3

1.2.1 Electrical properties 3

1.2.2 Mechanical properties 5

1.2.3 Optical properties 6

1.2.4 Thermal properties 7

1.3 Methods of fabricating graphene 7

1.4 Field emission theory 11

1.4.1 Model and principles 11

1.4.2 Field emission of carbon materials 13

1.5 Motivation and objectives 18

References 20

Table of Contents

v

CHAPTER 2 EXPERIMENTAL TECHNIQUES 25

2.1 Thin film deposition techniques 25

2.1.1 Magnetron sputtering (MS) 25

2.1.2 Pulsed laser deposition (PLD) 27

2.2 Graphene/graphene films Fabrication Methods 30

2.2.1 Modified Hummers method 30

2.2.2 Electrophoretic deposition (EPD) 31

2.3 Materials characterization 32

2.3.1 Scanning electron microscope (SEM) 32

2.3.2 Transmission electron microscope (TEM) 33

2.3.3 Raman spectroscopy 34

2.3.4 X-ray photoelectron spectroscopy (XPS) 36

2.3.5 Atomic force microscopy (AFM) 37

2.3.6 Field emission testing (FE) 39

References 41

CHAPTER 3 FABRICATION OF NANOGRAPHITE IN AMORPHOUS CARBON

(a-C) MATRIX FIELD EMITTERS 44

3.1 Introduction 44

3.2 Experimental method 47

3.3 Microstructural characterization of nanographite clusters in a-C matrix 49

3.3.1 Effect of temperature on growth of nanoclusters 49

3.3.2 Discussion 57

3.4 Field emission study of nanographite clusters in a-C matrix 59

3.4.1 Field emission properties 59

3.4.2 Emission mechanism of nanographite clusters in a-C matrix 65

3.5 Summary 69

References 70

Table of Contents

vi

CHAPTER 4 PULSED LASER FABRICATION OF FEW-LAYER GRAPHENE

USING METAL SUBSTRATES 73

4.1 Introduction 73

4.2 Few-layer graphene from Nickel substrates 80

4.2.1 Experimental method 80

4.2.2 Cooling rate dependence of graphene growth 81

4.2.3 Laser energy effects on graphene growth 85

4.3 Few-layer graphene segregation using other metal substrates 87


4.3.2 Segregation of graphene on different metal substrates 87

4.3.3 Effect of cooling rate when using different metals 90

4.3.4 Discussion 94

4.4 Few-layer graphene segregation using different pulsed laser wavelengths 98


4.4.2 Segregation with Nd:YAG and KrF pulsed laser 98

4.4.3 Discussion 102

4.5 Fabrication of few-layer graphene field emitters using PLD 108


4.5.2 Microstructural characterization of cones 108

4.5.3 Field emission study of graphene coated metal cones 112


4.6 Summary 117

References 120

CHAPTER 5 GRAPHENE-BASED FIELD EMITTERS BY

ELECTROPHORETIC DEPOSITION (EPD) 125

5.1 Introduction 125

5.2 Experimental method 129

5.3 Large area high density graphene for field emission 130

5.3.1 Microstructural characterization 130

Table of Contents

vii



5.4 Graphene/carbon nanotube hybrid materials for field emission 139

5.4.1 Microstructural characterization 139



5.5 Summary 146

References 147

CHAPTER 6 CONCLUSIONS AND FUTURE WORK 150

6.1 Conclusions 150

6.2 Future work 154

Summary

viii

Summary

Graphene has attracted a tremendous amount of research interest in the recent

decade due to its one-of-a-kind properties. With its excellent electrical, mechanical and

optical properties, graphene has the potential to outperform current available materials in

a wide range of applications. However, the fabrication and understanding of the growth

mechanism of graphene need to be established before realizing this material for practical

applications. While techniques such as chemical vapor deposition (CVD) and mechanical

exfoliation have been extensively investigated, studies on solid-state transformation of

carbon into graphene through physical vapor deposition (PVD) techniques are limited.

The aims of this dissertation were to understand the growth mechanism of

graphene segregation using pulsed laser deposition (PLD) with a solid carbon source and

to explore alternative methods to fabricate high sp2 and graphene-based materials for

field emission applications. To achieve this objective, investigations on growth and field

emission properties of nanographite in an amorphous carbon (a-C) matrix was first

conducted. Its field enhancement originated from the nanographite clusters within the

films that formed internal conducting channels in the largely insulating a-C matrix and

turn-on voltages decreased with an increase in nanographite density within the deposition

temperature range of 100–700 oC.

Subsequently, segregation of graphene was studied with PLD and the effects of

various parameters such as cooling rate, substrate, laser energy and wavelengths on the

growth of graphene were investigated. Briefly, low laser energy was favored for graphene

formation and different metals required different cooling profiles for graphene

Summary

ix

segregation. Using the knowledge acquired from these experiments, graphene-on-metal

nanocones field emitters with different densities were fabricated. It was observed that the

graphene-on-metal nanocones gave enhanced field emission properties as compared to

the bare metal nanocones due to the reduction of effective field emission tunneling

barrier, which was a result of graphene-metal charge transfer interactions. Controlling the

metal nanocones density was also an important factor, as electron screening from

neighboring cones should be minimized.

Lastly, graphene-based field emitters were fabricated with electrophoretic

deposition (EPD) and this presented a low cost, environmentally safe and highly

upscalable method to produce large area graphene field emitters. Graphene-only emitters

were deposited with different EPD durations and when the graphene density was too high,

screening effect resulted, which increased the turn-on electric field. To improve the

reliability of the graphene-only emitters, hybrid carbon nanotubes (CNT) with graphene

emitters were also fabricated. The CNT was envisioned to hold down the graphene flakes,

like a safety belt, at high voltages to prevent an early short circuit at relatively low

voltages. In contrast to pure graphene emitters, these hybrid emitters performed better

when the EPD duration was increased, as more graphene flakes were present when the

time was increased. These graphene flakes can help to improve the electrical conduction

of the material, as well as provide additional emission sites on the surface.

Hence, through the information gained from these experiments, graphene growth

segregation from solid carbon sources using PLD can be better understood and graphene

field emitter materials can be optimized during fabrication to achieve better performances.

List of Tables

x

List of Tables

Table 3.1 D and G peaks positions, amplitude, relative intensity ratio and estimated

sp2 content of the films deposited at different temperatures ......................51

Table 3.2 Comparison of field enhancement factors of various carbon-based emitters

....................................................................................................................66

Table 4.1 Carbon ion energies for step 1 and 2 during laser ablation of carbon target

for KrF and Nd:YAG systems .................................................................107

Table 4.2 Comparison of work function of free-standing graphene and clean metal

surfaces with the respective graphene on Ni or Co metal systems from ref.

60..............................................................................................................115

Table 4.3 Comparison of turn-on voltages of the graphene on metal cones samples

with the nanographite in amorphous carbon samples in Chapter 3 .........117

List of Figures

xi

List of Figures

Fig. 1.1 A graphene sheet is the building block of all other graphitic structures

ranging from fullerenes (1D), carbon nanotubes (2D) to graphite (3D) ......2

Fig. 1.2 Single layer and bi-layer graphene visible on 300 nm SiO2 under an

optical microscope but hardly visible on a 200 nm SiO2 substrate ..............3

Fig. 1.3 GO produced by Hummers method is yellow-brown in color while

reduced graphene is black in color .............................................................10

Fig. 1.4 Difference in emission barrier between field emission and thermionic

emission .....................................................................................................11

Fig. 1.5 Vertically oriented graphene flakes fabricated by spin coating onto

conductive silicon substrates in ref. 16 ......................................................16

Fig. 2.1 Schematic of RF magnetron sputtering showing the in-chamber

configurations ............................................................................................26

Fig. 2.2 Schematic of PLD used for nanographite in amorphous carbon deposition

and graphene segregation ...........................................................................28

Fig. 2.3 Plume seen during the ablation of target....................................................28

Fig. 2.4 Schematic of EPD setup with planar electrodes ........................................31

Fig. 2.5 Three possible outcomes during the interaction of photon and matter in

Raman spectroscopy ..................................................................................35

Fig. 2.6 Schematic of AFM .....................................................................................38

Fig. 2.7 Parallel plate electron emission tester setup ..............................................39

Fig. 2.8 Probe tip electron emission tester setup .....................................................40

Fig. 3.1 Raman spectrum for films deposited at room temperature, 100, 400 and

700 C ........................................................................................................50

Fig. 3.2 Raman ID/IG ratio and sp2 content changes of samples deposited at

temperature from 100-700 oC ....................................................................52

Fig. 3.3 Raman D and G peak positions of samples deposited at temperature from

100-700 oC .................................................................................................52

List of Figures

xii

Fig. 3.4 XPS spectra for films deposited at room temperature, 100, 400 and 700

C ...............................................................................................................53

Fig. 3.5 TEM micrograph for films deposited at (a) 100 C (b) 400 C (c) 700 C.

Inset shows the low magnification area view and SAED of crystalline

cluster with lattice spacing 0.33 nm corresponding to carbon ...................54

Fig. 3.6 Cluster density vs. temperature of 100, 400 and 700 oC samples ..............55

Fig. 3.7 AFM surface morphologies of (a) 100 C (b) 400 C (c) 700 C

samples........ ...............................................................................................56

Fig. 3.8 I-V characteristics of films deposited at 100, 400 and 700 oC ...................60

Fig. 3.9 Fowler-Nordheim plots at 50 µm for films deposited at 100, 400 and 700 oC................................................................................................................62

Fig. 3.10 Voltage-distance (V-d) curves for nanographite in a-C films deposited at

100, 400 and 700 oC ...................................................................................62

Fig. 3.11 Turn-on field versus anode-cathode distances of films deposited at 100,

400 and 700 oC. Inset (a) shows typical line shapes with 2 different height

approximations determined from simulations in Ref. 37. Inset (b) shows

line details at small distances .....................................................................64

Fig. 4.1 Raman spectra of samples cooled at different rates. The laser energy per

pulse was 50 mJ .........................................................................................81

Fig. 4.2 Cross-section TEM showing (a) the graphene layers above the Ni. (b)

Ni/Nickel Silicide/Si layered structure. Inset (c) is an overview optical

image showing rather uniform coverage ....................................................84

Fig. 4.3 Cross-section SEM showing delamination at Ni film-Si substrate interface

for medium and fast cooling samples ........................................................84

Fig. 4.4 Raman spectra of samples in which different pulsed laser energies were

used. Cooling rate was kept constant .........................................................86

Fig. 4.5 Raman spectra of carbon deposited on various metal substrates. Few-layer

graphene was formed only on Ni metal substrates when cooled with 1

oC/min to 550

oC followed by 20

oC/min...................................................89

Fig. 4.6 Cross-section TEM showing (a) the graphene layers above Ni (b) a more

disordered form of graphite formed when cooling conditions were too

fast............................................................................................................. .89

Fig. 4.7 Comparison of Raman spectra between Cu and Ni samples in which

different cooling rates 50 (top) and 100 oC/min (bottom) were used ........90

List of Figures

xiii

Fig. 4.8 Comparison of FLG grown on Co (top) and Ni (bottom). Full coverage

was obtained on Co substrate, while areas with graphene on Ni are

indicated by the black arrows (inset) .........................................................91

Fig. 4.9 Cross-section TEM image of graphene on top of Co showing 2-3

layers.......................................................................................................... 92

Fig. 4.10 Raman spectra of (a) as grown few-layer graphene on Co (top). (b)

Transferred few-layer graphene on Si, which has a significantly reduced D

peak indicating relatively few defects in the graphene (bottom) ...............92

Fig. 4.11 Raman spectra of areas covered and uncovered during carbon deposition

and segregation ..........................................................................................93

Fig. 4.12 Illustration of carbon atom concentration throughout the Co film as a

result of slow and fast cooling profiles. Carbon atoms are diffused deeper

into the Co film and harder to precipitate when cooling is slow. During

fast cooling, solubility is exceeded in the near-surface regions and

precipitation results ....................................................................................97

Fig. 4.13 Raman spectra of graphene on Ni segregated from carbon deposited by

Nd:YAG laser at 750 oC ..........................................................................100


KrF laser at 750 oC ...................................................................................100


Nd:YAG laser at 5 different spots. Optical images (right) show the spot

that was scanned ......................................................................................101


Nd:YAG laser when cooling rate was adjusted to a faster rate of 20 oC/min

..................................................................................................................102

Fig. 4.17 SEM images of original as-fabricated Ni cones (a) Type A, (b) Type B and

Co cones (c) Type A, (d) Type B .............................................................110

Fig. 4.18 SEM images of heat-only (a) Ni (b) Co metal nanocones .......................110

Fig. 4.19 SEM images of (a) Ni (b) Co metal nanocones after graphene

segregation............................................................................................... 111

Fig. 4.20 Raman spectra for Ni and Co metal cones ...............................................111

Fig. 4.21 Field emission I-V plots for graphene on cone samples ..........................112

Fig. 4.22 Field emission F-N plots for graphene on cone samples .........................113

List of Figures

xiv

Fig. 4.23 Schematic of work function shift when graphene is chemisorbed onto

metal ( G-on-M) and the position of EF in intrinsic versus doped

graphene........................................................................................... ........115

Fig. 5.1 SEM images of EPD deposited graphene field emitters with deposition

time of (a, b) 5, (c, d) 10 and (e, f) 15 min. Images on the left column were

taken at 1000X while those on the right were at 20,000X

magnification...........................................................................................132

Fig. 5.2 Cross-section SEM image of EPD deposited graphene showing near

vertical orientation of microflakes ...........................................................133

Fig. 5.3 (a) TEM images showing two as-prepared graphene flakes, one folded and

one flat. SAED (inset) confirmed that the graphene flakes were of high

crystalline quality. (b) Lattice fringes of scrolled or folded areas at high

magnification ...........................................................................................133

Fig. 5.4 Field emission J-E curves of the 5, 10, 15 min samples ..........................134

Fig. 5.5 F-N plots for 5 and 10 min samples. β can be calculated from the slope of

the straight line fit ....................................................................................135

Fig. 5.6 Schematic diagram showing the possible scenarios when deposition time

is (a) short or (b) extended. When flake density is high, a flatter overall

surface morphology gave poorer field enhancing properties due to

screening effect. .......................................................................................138

Fig. 5.7 SEM images of 2.5, 5 and 10 min graphene/CNT hybrid samples by EPD.

Left column shows magnification at 1000X while right column shows

20,000X ....................................................................................................140

Fig. 5.8 (a) Cross-section SEM images showing morphology with multiple

protrusions and a vertically aligned CNT-strapped graphene sheet (white

arrow) (b) Hybrid graphene/CNT film by filtration at 50% graphene in ref.

20 shows a flatter overall surface morphology ........................................141

Fig. 5.9 TEM images showing (a) high crystalline quality graphene (b) CNT used

in the experiments ....................................................................................141

Fig. 5.10 I-V plots of graphene/CNT hybrid emitters .............................................142

Fig. 5.11 F-N plots of graphene/CNT hybrid emitters ............................................143

List of Abbreviations

xv


0D Zero dimensional

1D One dimensional

2D Two dimensional

3D Three dimensional

a-C Amorphous carbon

AFM Atomic force microscopy

CCD Charge-couple device

CNT Carbon nanotubes

CVD Chemical vapor deposition

DC Direct current

dia. Diameter

DLC Diamond-like carbon

EPD Electrophoretic deposition

FE Field emission

FEG Field emission gun


xvi

FET Field effect transistors

FLG Few-layer graphene

F-N Fowler-Nordheim

FWHM Full width half maximum

GO Graphene oxide

HR High resolution

ITO Indium-tin oxide

KE Kinetic energy

LCD Liquid crystal displays

min Minutes

NEA Negative electron affinity

OLED Organic light emitting devices

PLD Pulsed laser deposition

PVD Physical vapor deposition

RF Radio frequency

RMS Root mean square

rpm rounds per minute


xvii

sec Seconds

SEM Scanning electron microscope

SLG Single layer graphene

TEM Transmission electron microscopy

WKB Wentzel–Kramers–Brillouin

XPS X-ray photoelectron spectroscopy

List of Publications

xviii

LIST OF PUBLICATIONS

1. A. T. T. Koh, Y. M. Foong, Likun Pan, Zhuo Sun and D. H. C. Chua, “Effective

large-area free-standing graphene field emitters by electrophoretic deposition”

Applied Physics Letters 101, 183107 (2012).

2. A. T. T. Koh, Y. M. Foong, and D. H. C. Chua, “Comparison of the mechanism

of low defect few- layer graphene fabricated on different metals by pulsed laser

deposition” Diamond and Related Materials 25, 98-102 (2012).

3. A. T. T. Koh, Y. M. Foong, B. Z. Phang, Daniel H. C. Chua, and M. Tanemura.

“Structural Studies of Aluminium Nitride Embedded in Amorphous Carbon”

Journal of Nanoscience and Nanotechnology 12, 6526-6530 (2012).

4. A. T. T. Koh, Y. M. Foong, J. Yu, D. H. C. Chua, A. T. S. Wee, Y. Kudo, K.

Okano, “Understanding Tube-like Electron Emission from Nanographite

Clustered Films” Journal of Applied Physics 110, 034903 (2011).

5. A. T. T. Koh, Y. M. Foong, and D. H. C. Chua, “Cooling rate and energy

dependence of pulsed laser fabricated graphene on nickel at low temperature”


6. A. T. T. Koh, J. Hsieh and D.H.C. Chua, “Structural characterization of Dual

Metal containing Diamond-like carbon Nanocomposite films by pulsed laser

deposition” Diamond and related Materials 19, 637 – 642 (2010).

7. A. T. T. Koh, J. Hsieh, Daniel H.C. Chua “Electron emission studies of CNTs

grown on Ti and Ni containing amorphous carbon nanocomposite films” Applied

Surface Science 256,178–182 (2009).


xix

8. A. T. T. Koh, Y. M. Foong, K. Mastubara, Z. Yusop, M. Tanemura and D. H. C.

Chua, “Graphene-coated metal nanocones for field emission”

9. A. T. T. Koh, Likun Pan, Zhuo Sun and D. H. C. Chua, “Hybrid

graphene/carbon nanotubes field emitters by electrophoretic deposition”

10. A. T. T. Koh, Y. M. Foong, K. Mastubara, Z. Yusop, M. Tanemura and D. H. C.

Chua, “A study on Nd:YAG and KrF pulsed laser effects on solid state

segregation of few-layer graphene”

11. Y. M. Foong, A. T. T. Koh, J. Hsieh, S. R. Lim, and D. H. C. Chua. “Materials

properties of ZnO/Diamond-like carbon (DLC) nanocomposite fabricated with

different source of targets” Diamond and Related Materials 25, 103-110 (2012).

12. J. Yu, P. Anetab, Angel T. T. Koh, Daniel H.C. Chua, J. Wei. “Enhanced

electron emission from tetrahedral amorphous carbon capped carbon nanotube

core-shelled structure” Diamond and Related Materials. 21, 37-41 (2012).

13. Jie Sun, Niclas Lindvall, Matthew T. Cole, Koh T. T. Angel, Teng Wang, Ken B.

K. Teo, Daniel H. C. Chua, Johan Liu and August Yurgens. “Low Partial

Pressure Chemical Vapor Deposition of Graphene on Copper”. IEEE

Transactions on Nanotechnology 11, 255-260 (2012).

14. Y. M. Foong, A. T. T. Koh, H. Y. Ng and D. H. C. Chua. “Mechanism behind the

surface evolution and microstructure changes of laser fabricated nanostructured

carbon composite” Journal of Applied Physics 110, 054904 (2011).


xx

15. Y. M. Foong, A. T. T. Koh, S. R. Lim, H. Y. Ng and D. H. C. Chua. “Properties

of laser fabricated nanostructured Cu/diamond-like carbon composite”. Journal

of Material Research 26, 2761-2771 (2011).

16. Y. M. Foong, A. T. T. Koh, and D. H. C. Chua. “Experimental and theoretical

study on the energy-dependent surface evolution and microstructure changes in

copper nanostructured composites” Journal of Physics D: Applied Physics 44,

385401 (2011).

17. Y. M. Foong, A. T. T. Koh, L. Niu, and D. H. C. Chua. “Mechanism behind the

formation of self-assembled nano-sized clusters in Diamond-like carbon

nanocomposite”. Journal of Nanoscience and Nanotechnology 11, 10511 (2011)

18. J. Yu, Y. M. Foong, A. T. T. Koh, and D. H. C. Chua. “Field emission

characteristics of tetrahedral amorphous carbon coated carbon nanotubes with

and without hydrogenation treatment”. Journal of Physical Chemistry C 115,

11336 (2011).

19. I. Saito, W. Miyazaki, M. Onishi, Y. Kudo, T. Masuzawa, T. Yamada, A. Koh, D.

Chua, K. Soga, M. Overend, M. Aono, G. A. J. Amaratunga, and K. Okano


20. T. Masuzawa, Y. Sato, Y. Kudo, I. Saito, T. Yamada, A. T. T. Koh, D. H. C.

Chua, T. Yoshino, W. J. Chun, S. Yamasaki, and K. Okano. “Correlation between

low threshold emission and C--N bond in nitrogen-doped diamond films” Journal

of Vacuum Science and Technology. B. 29, 02B119 (2011).

21. Y. M. Foong, A. T. T. Koh, J. Hsieh, and D. H. C. Chua. “A comparative study

on as-deposited and in situ oxidized ZnO/diamond-like carbon (DLC)


xxi

nanocomposite by pulsed laser deposition technique” Journal of Material

Research 25, 899 (2010).

22. J. Hsieh, A. T. T. Koh, D. H. C. Chua. “Surface studies of carbon-based ZnO

nanocomposite films” International Journal of Nanotechnology 6, 661 (2009)

Conference presentations:

1. Poster presentation “Segregation of Few-layer Graphene with Low Defects by

Pulsed Laser Deposition” 2012 Materials Research Society Spring Meeting, San

Francisco, California, USA, 9th

-13th

April 2012.

2. Poster presentation “Experimental and Theoretical Study on the Energy

Dependent Surface Evolution and Microstructure Changes in Copper

Nanostructured Composite” 2012 Materials Research Society Spring Meeting,

San Francisco, California, USA, 9th

-13th

April 2012.

3. Poster presentation “Field Emission Mechanism of Nitrogen-doped Diamond with

Different C-N Concentration” 2011 Materials Research Society Spring Meeting,

San Francisco, California, USA, 25th

-29th

April 2011.

4. Oral presentation “Electron Emission from Nanographite clusters embedded in a-

C films” 2010 8th International Vacuum Electron Sources Conference (IVESC),

Nanjing, China, 14th

–16th

October 2010.

5. Poster presentation ” XPS and TEM analysis of Multi-elemental Nanocomposites

in Diamond-like amorphous carbon films” Ecoss 2008:The 25th European

Conference on Surface Science, Liverpool, England, 27th

July–1st August 2008.


xxii

6. Poster presentation “Ti & Ni in Diamond Like Carbon as Growth Substrates for

Carbon Nanotubes”. The 2008 Asian Conference on Nanoscience and

Nanotechnology (AsiaNano 2008), Biopolis, Singapore 3rd–7th November 2008.

Chapter 1 Introduction

1

Chapter 1

Introduction

Over the last few years, extensive research efforts have been devoted to graphene

due to its unique and attractive material properties. In this chapter, the discovery of

graphene and its properties are reviewed in section 1.1 and 1.2, respectively. Fabrication

methods of graphene will be explained in section 1.3. The principles and equations used

in field emission (FE) studies, as well as the application of graphene as a potential field

emitter will be reviewed in Section 1.4. Lastly, the motivations and objectives of this

project are detailed in section 1.5.

1.1 Graphite and monolayer graphite

Graphite, an allotrope of carbon, has been long known since 1789. Being the

thermodynamically most stable form of carbon at standard conditions, it was believed

that a single graphite layer cannot exist freely. It was till the year 2004 when Novoselov

et al.1 demonstrated the isolation of single layer graphite planes through mechanical

exfoliation of a highly ordered pyrolytic graphite target (often referred to as the scotch-

tape technique) that refuted this belief. Graphene is the name given to the monolayer

graphite sheets, which are the building blocks of all graphitic materials.

The physical structure of graphene resembles that of a soft membrane and a sheet

of graphene is not flat but rippled.2 These ripples are caused by the pre-existing strains in


2

Fig. 1.1 A graphene sheet is the building block of all other graphitic structures ranging

from fullerenes (0D), carbon nanotubes (1D) to graphite (3D).2

graphene and can cause charge distribution inhomogeneities, which are not favorable

towards electronic quality. It can be rolled up to form 0D fullerenes, 1D carbon

nanotubes (CNT) or stacked to form 3D graphite (Fig. 1.1).3 In fact, when a pencil is

pressed on paper during writing, there could be many single layer graphene sheets in the

numerous graphite stacks produced on paper. However, due to the lack of inspection

tools, graphene cannot be easily differentiated from graphite, until the method of viewing

graphene flakes placed on SiO2 substrate with a specific thickness (300 nm) that makes

these elusive flakes visible (Fig. 1.2) was discovered.4 Since then, there have been

increased interests in these single and few-layer planar aromatic carbon films due to the

special properties that it exhibit.


3

Fig. 1.2 Single layer and bi-layer graphene visible on 300 nm SiO2 under an optical

microscope but hardly visible on a 200 nm SiO2 substrate.3

To clarify the nomenclature, graphene, when strictly defined, refers to a single

layer (SLG), bilayer graphene refers to 2 layers while few-layer graphene (FLG) has 3–

10 layers. Above 10 layers, it is generally referred to as graphite. While there are

similarities and differences between the properties of graphene and FLG, FLG grown on

4H-SiC has been shown to behave like SLG due to a different (from A-B stacking)

stacking structure, giving rise to rotational faults in every graphene layer, thus decoupling

the individual sheets and resulting in isolated graphene layers.5 In the next section, the

properties of graphene will be detailed.

1.2 Properties of graphene

1.2.1 Electronic properties

Graphene is a zero bandgap semiconductor that exhibits unique electronic

properties such as massless Dirac fermions (zero rest mass and effective speed of light),6

ambipolar electric field effect7 and anomalous quantum hall effect at room temperature,

which is a characteristic behavior of Dirac fermions.8, 9

The origin of the massless


4

fermions is due to the linear energy relationship between electron energy, E, and

wavevector, k, in the form of

22

yxf kkE (1.1)

near the vertices of the hexagonal Brillouin zone of graphene.6, 10

The six corners of the

2D Brillouin zones are therefore known at Dirac points. Ambipolar electric field effect

occurs when the charge carriers can be changed continuously between electron and holes

in concentrations as high as 1013

cm-2

by changing the externally applied voltage.1, 7

The

charge carriers of freely suspended graphene have an intrinsic mobility limit of 2 × 105

cm2/V.s at low temperatures, which is the highest known of any semiconductor. However,

the mobility is limited to 4 × 104 cm

2/V.s for graphene on SiO2 substrate, due to extrinsic

scattering by surface phonons, showing that substrate interactions does affect the

properties of graphene.11

As a result of the high mobility, graphene charge carriers are

able to exhibit ballistic transport on a submicron scale, in which the electrons transport in

a medium with negligible resistivity and have mean free path much larger than the

dimensions of the box that holds the medium.12

However, the wavy or rippled surface of

a graphene sheet results in breaking of the translational invariance and hence limits

ballistic transport. Nevertheless, these electronic properties make graphene a highly

suitable material for electronic applications and it has been shown to be a promising

material for transparent electrodes,13

ultracapacitors,14

field emission15, 16

and transistor

applications.17

As a result of not possessing a band gap, pristine SLG cannot be used as field

effect transistors (FET) as it cannot be switched off. However, this problem has been


5

overcame with different methods that can be divided into those that disrupt and break the

hexagonal structure of graphene to induce a band gap and those that preserve the 6 fold

symmetry of the structure.18

For the former, this includes slicing graphene into

nanoribbons,19

chemical doping20

and creating graphene nanomesh21

structures with

lithography. The use of bilayer graphene for applications that require a band gap is also

possible without having to destroy the hexagonal rings of carbon. For instance, Xia et

al.22

has demonstrated a band gap in bilayer graphene and an on/off current ratio of 100

(compared to 4 for SLG) at room temperature when a biasing condition was applied for

their dual gate graphene FET design. Zhang et al.23

have also demonstrated bilayer

graphene FET with a tunable band gap up to 0.25 eV, using infrared micro-spectroscopy,

while Quhe et al.18

made use of graphene-substrate interaction between single graphene

and hexagonal boron nitride to induce a band gap. These methods preserve the high

carrier mobility of graphene which is desirable in electronic applications.

1.2.2 Mechanical properties

The Young’s modulus of defect-free graphene using atomic force microscopy

(AFM) nanoindentation on graphene suspended over holes was measured to be 1.0 TPa at

a thickness of 0.335 nm, which makes graphene one of the strongest materials ever.24

As

graphene is a two dimension (2D) material, the measured values were normalized with

the area instead of volume and the Young’s modulus is equivalent to the in-plane

stiffness (E2D

). The same value of 1 TPa was obtained for single wall CNT, which is

essentially a rolled up graphene sheet. Given its excellent mechanical properties, it can

value add the use of graphene in many applications that may not be primarily mechanical


6

in nature. The use of graphene in composites has also been reported to greatly improve

tensile strength by 150 % and increase the Young’s modulus by 10 times at a low

graphene loading of 1.8 vol.%25

and has outperformed multiwall CNT as a reinforcement

additive.26

Possessing excellent mechanical properties as an electronic material can

increase the robustness of the fabricated device by allowing it to withstand harsh

conditions with lower chances of breakage.

1.2.3 Optical properties

A single layer or few-layer graphene is extremely thin and thus optically

transparent, with a transparency of 97.7 % in the visible light range. The transmittance

value of graphene decreases linearly with an increase in the number of layers.27

The inter-

band optical transitions in SLG and bilayer graphene can be varied dramatically by

electrical gating. This is because the low density of states near the Dirac point of

graphene causes the Fermi level, EF, to change significantly when the carrier density

changes.28

The combination of high transparency, high conductivity and ultrafast

photoresponse makes graphene extremely suitable for optoelectronic applications such as

sensors, communication devices and transparent electrodes found in liquid crystal

displays (LCD), touchscreens and for organic light emitting devices (OLED), replacing

the expensive indium-tin oxide (ITO) which is hard to recycle and natural indium sources

are depleting.13


7

1.2.4 Thermal properties

The thermal conductivity of graphene, κ, is dominated by phonon transport,

specifically diffusive conduction at high temperatures and ballistic conduction at

sufficiently low temperatures.27, 29

The measured value of κ for a suspended mechanically

exfoliated graphene is ~5000 Wm−1

K−1

, which is higher than that of bulk graphite at

room temperature (2000 Wm− 1

K−1

) and twice that of diamond (2500 Wm− 1

K−1

). This

value was measured by observing the red-shift of the Raman G peak which depends

linearly on the sample’s temperature. Through the slope of the excitation power versus G

peak frequency, κ was estimated.30

The high thermal conductivity of graphene is useful

and beneficial when efficient heat dissipation is required in electronic applications.

It has also been shown that by making use of the negative thermal expansion

coefficient and orientation of the ripples on graphene surface, wavelength and amplitude

of the ripples in graphene can be controlled by thermal manipulation. These

morphological changes in the ripples with temperature may be the key to explaining

some of the unusual properties of graphene as well as opening a new area for strain based

devices.31

1.3 Methods of fabrication

Graphene is seen as the material of the next generation given its excellent

properties, but the bigger issue that needs to be addressed is in the area of graphene

fabrication. Conventional techniques to fabricate graphene are expensive while some are


8

tedious. The most popular and easiest method is the mechanical exfoliation method

which involves rubbing a highly pyrolytic graphite target against a piece of sticky

cellophane tape and depositing the graphite flakes on SiO2 of carefully chosen thickness

(typically 300 nm). By looking under the microscope, single and FLG flakes can be

distinguished from the color observed (Fig. 1.2). Flakes produced by this method can be

up to 10 m in size but there is a large variation in the thickness of the flakes produced.

While any one can do this with the correct materials and without specialized equipment,

this method is tedious, time consuming and above all not scalable, thus making it

unsuitable for industries.32

The second technique used in fabricating graphene is by growing graphene

epitaxially on SiC. This method involves high process temperatures (>1000 oC and up to

1600 oC) and also ultra-high vacuum conditions of 10

-9 Torr. While this method is not

complicated and graphene grown on SiC can be patterned with standard nanolithography

techniques,33

the large lattice mismatch between epitaxial graphene and the underlying

SiC substrate makes the quality incomparable to films obtained with the mechanical

exfoliation method.34

It is also difficult to transfer the epitaxial graphene on SiC onto

other substrates due to the high chemical stability of SiC. Substrate bonding may also

affect the electronic properties of graphene grown on SiC as mentioned earlier.11

The third growth method of graphene is carried out on metal substrates. The

recent interests in free-standing graphene have led to the boom in research on graphene

adsorbed on metal surfaces. After graphene is absorbed on the metal, the metal support

layer can be easily etched away using an acid solution. The remaining graphene layers

can either be free-standing and dredged up from the solution, or transferred onto another


9

substrate of choice. Graphene on metal substrates can be prepared by both solid and

gaseous carbon sources. On one hand, segregation of dissolved carbon on the substrate

surface through high temperature annealing and cooling of a carbon-containing metal can

be carried out either by solid or gaseous sources. If the carbon source is gaseous, for e.g.

methane or ethylene, decomposition of the gas first occurs on the metal surface to release

carbon. On the other hand, chemical vapor deposition (CVD) is one technique that uses

only gaseous carbon sources. When copper (Cu) catalyst is used with CVD, the growth of

graphene is reported to stop once the metal layer is covered with a SLG, but continues to

give FLG if nickel (Ni) catalyst is used instead. This is due to the difference in their

growth mechanisms, which is surface adsorption and segregation respectively for Cu and

Ni.35

Thus, single layer growth can be achieved by controlling the temperature or catalyst,

which give CVD an edge over mechanical exfoliation that results in a wide range of

thicknesses.36

Additionally, CVD is a highly scalable method that can be used to fabricate

large area graphene (~cm2).

37, 38 Metals such as ruthenium,

39 platinum,

40 copper and

nickel37, 41

have been used in the synthesis of graphene. In the case of nickel, to prevent

the growth of thick layers of graphite, thin layers of Ni in the order of a few hundred

nanometers deposited on SiO2/Si substrates can be used.37

Chemical exfoliation is also frequently used to obtain a large amount of free-

standing graphene. This technique first began with the Hummers method, which was

first demonstrated by W. S. Hummers in 1958 to prepare graphite oxide.42

The graphite

was first oxidized to form graphite oxide using potassium permanganate (KMnO4) and

hydrosulfuric acid (H2SO4), and subsequently exfoliated to form graphene oxide (GO).

The actual chemicals used today may differ from the original Hummers method and thus


10

it is called modified Hummers method. GO, which is yellow-brown as shown in Fig. 1.3,

can be reduced to graphene through treatment with hydrazine or more recently,

microwave assisted reduction.43

The end product can be easily transferred onto other

substrates by filtration, spray coating and other solution processing methods.44

Chemical

exfoliation provides a low cost and high yield method to prepare graphene and is suitable

for large scale production.

Fig. 1.3 GO produced by Hummers method is yellow-brown in color while reduced graphene is

black in color.45


11

1.4 Field emission theory

1.4.1 Model and principles

Field emission is the extraction of electrons from a metal or semiconductor under

a strong electric field into vacuum by tunneling through the surface triangular potential

barrier.46

When an electric field is applied, the surface potential is deformed such that the

shape of the surface potential barrier changes from square (no electric field) to

triangular.47

This is in contrast to thermionic emission where heat is applied and electrons

gain sufficient energy to surmount the barrier instead of tunneling through the barrier.

This difference in barriers between these two types of emission is depicted in Fig. 1.4.

Fig. 1.4 Difference in emission barrier between field emission and thermionic emission.


12

Fowler and Nordheim proposed the first model for field emission in 1928 for

metal surfaces at low temperatures.48

Nonetheless, this model is widely used in field

emission studies of graphitic materials till today. In this model, the emission current

density and macroscopic applied field can be described by the Fowler-Nordheim (F-N)

relation according to the following expression,49

E

BEAJ

23

2

exp)(

(1.2)

Where J is the macroscopic current density, E is the applied field, β is the field

enhancement factor, is the work function and constants A=1.54×10-6

AeV.V-2

,

B=6.83×107 V.eV

-3/2cm

-1.

By taking natural logarithm, equation 1.2 can be rewritten into the following

linear equation 1.3,

E

BA

E

J

23

2

2lnln

(1.3)

Field enhancement factor (β), which is used as a qualitative measure of field

emission property of the material, can be calculated from the slope (∆) of the straight

line plot of ln (J/E2) versus 1/E for a sample of known value using,

23

B (1.4)

Likewise, if β is known, can be calculated from the slope of the F-N plot instead.


13

The local electric field, Elocal, experienced on a microscopic scale at the emitting

site is thus related to the macroscopic applied field, Eappl, through β by the following

expression.

.appllocal EE (1.5)

1.4.2 Field emission of carbon materials

Field emission from carbon-based materials has been studied intensively for

decades and some of the carbon materials used as field emission cold cathodes are

diamond, doped or undoped DLC, and the well-known CNT owing to its high aspect ratio

and increased geometrical enhancement factor.

In field emission applications, diamond has attracted a lot of attention as cold

cathode emitters due to advantageous properties like high thermal conductivity and

negative electron affinity (NEA). NEA means that the minimum energy of electrons in

the vacuum is below the minimum energy of electrons in the conduction band.50

Phosphorous-doped diamond and nitrogen-doped diamond has been extensively studied

for emission along with different surface terminations such as oxygen and hydrogen.51

It

has been reported that field emission of heavily N-doped diamond is not limited by its

resistance and thus follows a metal-insulator-vacuum mechanism, whereby the electrons

are injected from the back contact into the conduction band of diamond and subsequently

flew into vacuum without a potential barrier due to NEA.52

It should be noted that NEA


14

is not a prerequisite to low applied fields as there are materials without NEA that still

exhibit good emission properties e.g., diamond-like carbon (DLC) and CNT.

Generally, low macroscopic threshold fields (<1–40 V/m) observed in carbon

materials are attributed to either geometrical field enhancement factors (external field

enhancement) or field enhancement due to internal conductive elements (internal field

enhancement). For CNTs, they have a high aspect ratio and are fully sp2 bonded; hence

electron transport is not a limiting factor within the tubes. Instead, the density of the

CNTs affects the emission threshold due to shielding effects. In addition, though CNTs

show remarkable performance as field emitters, their thermal stability and reliability are

often inferior, i.e. easily degraded under high applied voltages.

In the case of DLC, which possesses extremely smooth surfaces, the field

enhancement factors are largely contributed by internal field enhancement. The electric

field lines will concentrate at surface conductive elements, which are the sp2 bonded

regions in this case, and lead to enhancement of the applied field. Moreover, owing to the

presence of sp2 in combination with sp

3 bonding, DLC films have a certain edge over

fully sp3 bonded diamond since the sp

2 can provide electron conduction paths within the

films. Furthermore, sp2 clusters have different dielectric constants from sp

3 carbon,

giving rise to enhancement due to dielectric inhomogenity.53

The ability to grow large

area DLC at room temperature through economical methods is an added bonus to its

excellent properties, thus making it a potential candidate for cold cathode emitters.54

DLC

is known to emit electrons with threshold fields in the order of 10 Vµm-1

.55

There have

been various attempts to lower the emission of DLC films by doping metals such as Ag,56


15

Ti57

and Ni.58

In such cases, the increase in sp2 percentage and improved conductivity in

the films by metal incorporation plays a large part in the reduction of threshold field.

Graphene is essentially an unrolled single walled CNT and possesses excellent

conductivity. However, a single graphene sheet may not have good field enhancement

due to its rather flat profile. This is further aggravated by the fact that most graphene

fabrication methods including mechanical exfoliation, decomposition of SiC, segregation

from metals and CVD result in graphene that has a planar morphology parallel to the

substrate. To fully exploit the capabilities of graphene as a field emitter, the graphene

sheets should be perpendicular to the substrate such that the planes of highest

conductivity are along the direction of emission and electrons can thereby emit from the

sharp graphene edges. There are several studies that strive to achieve this in current

literature. Some early works include Eda et al.16

who fabricated graphene cathodes by

spin coating chemically exfoliated graphene onto Si substrate to gain a certain degree of

vertical alignment. The orientation of the sheets was related to the spin coating speeds

and best results were obtained from films spin coated at 600 rpm. Even though the turn-

on field (Eon) was 4 V/m, the maximum current density obtained was 1 mA/cm2, which

was below the requirement for actual high current applications. The authors have recently

carried out the same method of spin coating graphene sheets onto Si microtips and

lowered the Eon to 2.3 V/m.59

Another research group, Malesevic et al.60

managed to

fabricate vertically aligned FLG on titanium and silicon substrates by microwave plasma

enhanced CVD (MW-PECVD). The quality of FLG was optimized by controlling the H2

and CH2 precursor gas ratios. Using this method, the Eon was lowered to 1 V/m and

current density was increased to 14 mA/cm2. Although this method has some advantages


16

Fig. 1.5 Vertically oriented graphene flakes fabricated by spin coating onto conductive silicon

substrates in ref. 16.

such as the direct fabrication of graphene without extra steps, there is no simple way to

control the FLG density, which gives rise to screening effect (similar to the case of a

dense forest of vertically aligned CNT).

Chemically exfoliated graphene has also been studied as field emitters by using

screen printing techniques61

, electrophoretic deposition62

and coating on conductive

tape62

with turn-ons field ranging from 1–5.2 V/μm. Graphene, first exfoliated by

modified Hummers method and reduced by hydrazine, were compared against those

reduced by microwave assistance and it was reported that microwave-reduced graphene

(Eon = 0.39 V/μm) performed better than hydrazine-reduced graphene (Eon = 0.94 V/μm).

The authors attributed it to the larger size of graphene flakes produced by microwave

reduction leading to more protrusions when stacked, thus improving field enhancement.45

Another example of achieving protrusions was recently demonstrated by Pandey et al.

who generated morphological defects (such as edges, discontinuity and ripples) due to the

poor transfer and adhesion of CVD-grown graphene onto pristine Si (100) substrates.63


17

However, this method does not give reliable control over the density of defects and it can

be difficult to reproduce the results. Growth of graphene on nanotube or wire structures

have also been another area explored lately. For instance, Deng et al. reported the direct

growth of FLG on the tips of CNT arrays using a RF H plasma sputtering system.64

The

fabricated structure had sharp graphene edges extending outwards on the tips of a well

aligned CNT array and gave slightly better field emission results compared to a pure

CNT array. Other methods to create graphene-based cathodes with a field enhancing

morphology include the draping of graphene oxide on Ni nanotips and the Eon was

reported 5 × 105 V/m at 1 μA, which is a relatively high field and this is probably due to

the much lower conductivity of graphene oxide.65

Furthermore, Yang et al. reported the

large area transfer of CVD fabricated monolayer graphene onto a well aligned array ZnO

nanotips to produce 100–200 nm graphene-covered protrusions.66

These emitters were

capable of very high current density (~ 500 µA/cm2), clearly demonstrating the potential

of graphene for FE applications. Due to the unique properties of graphene, it is not

surprising that it has received considerable attention for field emission; however the

current graphene field emission technology is still at its infancy and will require more

time before reaching actual applications.


18

1.5 Motivations and objectives

From the above literature, it can be seen that there are still gaps in the area of

graphene fabrication. Since the first report on the formation of monolayer graphene by

Novoselov et al., extensive efforts have been devoted to the fabrication of graphene using

CVD, mechanical and chemical exfoliated methods. However, limited information is

available for physical vapor deposition (PVD) fabrication techniques, such as pulsed laser

deposition (PLD). The use of PLD could reduce carbon dissolution time through the

generation of charged energetic species that penetrate into the substrate. Despite

numerous studies on the growth mechanism of CVD deposited graphene, information on

the mechanism of solid-state transformation, segregation and growth mechanism of

graphene using PLD technique is lacking. Furthermore, even though carbon-based

composites such as metal-doped DLC and metal-coated CNT have been explored as new

generation field emitters, high sp2 carbon-based composites of graphene/CNT have not

been investigated.

As such, my research work was devoted to the following objectives:

1) Fabricate high sp2 carbon/graphene/few-layer graphene using physical deposition

methods i.e. PLD and understand the growth mechanism of graphene using PLD

segregation methods.

2) Study the field emission properties of the fabricated high sp2 and graphene-based

emitter materials.

3) Fabricate and explore high sp2 hybrid materials such as graphene/CNT for field

emission applications.


19

By achieving the objectives in this thesis, this work presented alternative methods

to the fabrication of graphene and has shed some light to understanding the growth

mechanism of graphene using PLD method. Through the fabrication and study of the

emission properties of graphene and graphene/CNT hybrid materials, this work will be

helpful towards contributing to the future development of low cost field emitters for

actual industrial applications.


20

References

1. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I.

V. Grigorieva and A. A. Firsov, Science 306, 666 (2004).

2. A. Fasolino, J. H. Los and M. I. Katsnelson, Nature Mater. 6, 858-861 (2007).

3. A. K. Geim and K. S. Novoselov, Nature Mater. 6, 183-191 (2007).

4. P. Blake, E. W. Hill, A. H. C. Neto, K. S. Novoselov, D. Jiang, R. Yang, T. J. Booth

and A. K. Geim, Appl. Phy. Lett. 91, 063124 (2007).

5. J. Hass, F. Varchon, J. E. Millán-Otoya, M. Sprinkle, N. Sharma, W. A. de Heer, C.

Berger, P. N. First, L. Magaud and E. H. Conrad, Phys. Rev. Lett. 100, 125504

(2008).

6. L. S. Panchakarla, K. S. Subrahmanyam, S. K. Saha, A. Govindaraj, H. R.

Krishnamurthy, U. V. Waghmare and C. N. R. Rao, Adv. Mater. 21, 4726 (2009).

7. K. S. Novoselov, S. V. Morozov, T. M. G. Mohinddin, L. A. Ponomarenko, D. C.

Elias, R. Yang, I. I. Barbolina, P. Blake, T. J. Booth, D. Jiang, J. Giesbers, E. W. Hill

and A. K. Geim, phys. status solidi (b) 244, 4106-4111 (2007).

8. K. S. Novoselov, Z. Jiang, Y. Zhang, S. V. Morozov, H. L. Stormer, U. Zeitler, J. C.

Maan, G. S. Boebinger, P. Kim and A. K. Geim, Science 315, 1379 (2007).

9. A. H. Castro Neto, F. Guinea, N. M. R. Peres, K. S. Novoselov and A. K. Geim, Rev.

Mod. Phys. 81, 109-162 (2009).

10. M. I. Katsnelson, Materials Today 10, 20-27 (2007).

11. J. H. Chen, C. Jang, S. Xiao, M. Ishigami and M. S. Fuhrer, Nature Nanotech. 3, 206-

209 (2008).

12. K. S. Novoselov, E. McCann, S. V. Morozov, V. I. Falko, M. I. Katsnelson, U.

Zeitler, D. Jiang, F. Schedin and A. K. Geim, Nature Phys. 2, 177 (2006).


21

13. J. Wu, M. Agrawal, H. c. A. Becerril, Z. Bao, Z. Liu, Y. Chen and P. Peumans, ACS

Nano 4, 43-48 (2009).

14. M. D. Stoller, S. Park, Y. Zhu, J. An and R. S. Ruoff, Nano Lett. 8, 3498-3502 (2008).

15. M. Araidai, Y. Nakamura and K. Watanabe, Phys. Rev. B 70, 245410 (2004).

16. G. Eda, H. Emrah Unalan, N. Rupesinghe, G. A. J. Amaratunga and M. Chhowalla,

Appl. Phys. Lett. 93, 233502 (2008).

17. M. Freitag, Nature Nanotech. 3, 455-457 (2008).

18. R. Quhe, J. Zheng, G. Luo, Q. Liu, R. Qin, J. Zhou, D. Yu, S. Nagase, W.-N. Mei, Z.

Gao and J. Lu, NPG Asia Mater. 4, e6 (2012).

19. X. Wang, Y. Ouyang, X. Li, H. Wang, J. Guo and H. Dai, Phys. Rev. Lett. 100,

206803 (2008).

20. F. Withers, M. Dubois and A. K. Savchenko, Phys. Rev. B 82, 073403 (2010).

21. J. Bai, X. Zhong, S. Jiang, Y. Huang and X. Duan, Nature Nanotech. 5, 190-194

(2010).

22. F. Xia, D. B. Farmer, Y.-m. Lin and P. Avouris, Nano Lett. 10, 715-718 (2010).

23. Y. Zhang, T.-T. Tang, C. Girit, Z. Hao, M. C. Martin, A. Zettl, M. F. Crommie, Y. R.

Shen and F. Wang, Nature 459, 820-823 (2009).

24. C. Lee, X. Wei, J. W. Kysar and J. Hone, Science 321, 385-388 (2008).

25. X. Zhao, Q. Zhang, D. Chen and P. Lu, Macromolecules 43, 2357-2363 (2010).

26. M. A. Rafiee, J. Rafiee, Z. Wang, H. Song, Z.-Z. Yu and N. Koratkar, ACS Nano 3,

3884-3890 (2009).

27. Y. Zhu, S. Murali, M. D. Stoller, A. Velamakanni, R. D. Piner and R. S. Ruoff,

Carbon 48, 2118-2122 (2010).

28. F. Wang, Y. Zhang, C. Tian, C. Girit, A. Zettl, M. Crommie and Y. R. Shen, Science

320, 206-209 (2008).


22

29. C. Yu, L. Shi, Z. Yao, D. Li and A. Majumdar, Nano Lett. 5, 1842-1846 (2005).

30. A. A. Balandin, S. Ghosh, W. Bao, I. Calizo, D. Teweldebrhan, F. Miao and C. N.

Lau, Nano Lett. 8, 902-907 (2008).

31. W. Bao, F. Miao, Z. Chen, H. Zhang, W. Jang, C. Dames and C. N. Lau, Nature

Nanotech. 4, 562-566 (2009).

32. A. K. Geim, Science 324, 1530-1534 (2009).

33. C. Berger, Z. Song, X. Li, X. Wu, N. Brown, C. Naud, D. Mayou, T. Li, J. Hass, A.

N. Marchenkov, E. H. Conrad, P. N. First and W. A. de Heer, Science 312, 1191-

1196 (2006).

34. H. Huang, W. Chen, S. Chen and A. T. S. Wee, ACS Nano 2, 2513-2518 (2008).

35. X. Li, W. Cai, L. Colombo and R. S. Ruoff, Nano Lett. 9, 4268-4272 (2009).

36. J. Wintterlin and M. L. Bocquet, Surf. Sci. 603, 1841-1852 (2009).

37. K. S. Kim, Y. Zhao, H. Jang, S. Y. Lee, J. M. Kim, K. S. Kim, J.-H. Ahn, P. Kim, J.-

Y. Choi and B. H. Hong, Nature 457, 706-710 (2009).

38. X. Li, W. Cai, J. An, S. Kim, J. Nah, D. Yang, R. Piner, A. Velamakanni, I. Jung, E.

Tutuc, S. K. Banerjee, L. Colombo and R. S. Ruoff, Science 324, 1312-1314 (2009).

39. P. W. Sutter, J.-I. Flege and E. A. Sutter, Nature Mater. 7, 406-411 (2008).

40. H. Ueta, M. Saida, C. Nakai, Y. Yamada, M. Sasaki and S. Yamamoto, Surf. Sci. 560,

183-190 (2004).

41. A. Reina, X. Jia, J. Ho, D. Nezich, H. Son, V. Bulovic, M. S. Dresselhaus and J.

Kong, Nano Lett. 9, 30-35 (2008).

42. W. S. Hummers and R. E. Offeman, J. Am. Chem. Soc. 80, 1339-1339 (1958).

43. W. Chen, L. Yan and P. R. Bangal, Carbon 48, 1146-1152 (2010).

44. S. Pei, J. Zhao, J. Du, W. Ren and H.-M. Cheng, Carbon 48, 4466-4474 (2010).


23

45. K. Wang, T. Feng, M. Qian, H. Ding, Y. Chen and Z. Sun, Appl. Surf. Sci. 257, 5808

(2011).

46. R. G. Forbes, Solid-State Electron. 45, 779-808 (2001).

47. J.-M. Bonard, M. Croci, I. Arfaoui, O. Noury, D. Sarangi and A. Châtelain, Diam.

Relat. Mater. 11, 763-768 (2002).

48. R. H. Fowler and L. Nordheim, Proc. R. Soc. A 119, 173-181 (1928).

49. C. X. Xu, X. W. Sun and B. J. Chen, Appl. Phy. Lett. 84, 1540-1542 (2004).

50. M. W. Geis, N. N. Efremow, K. E. Krohn, J. C. Twichell, T. M. Lyszczarz, R. Kalish,

J. A. Greer and M. D. Tabat, Nature 393, 431-435 (1998).

51. T. Yamada, H. Kato, S. Shikata, C. Nebel, H. Yamaguchi, Y. Kudo and K. Okano, J.

Vac. Sci. Technol. B Microelectro.n Nanometer. Struct. Process. Meas. Phenom. 24,

967 (2006).

52. K. Okano, T. Yamada, A. Sawabe, S. Koizumi, J. Itoh and G. A. J. Amaratunga,

Appl. Phy. Lett. 79, 275 (2001).

53. J. D. Carey, R. D. Forrest, R. U. A. Khan and S. R. P. Silva, Appl. Phy. Lett. 77,

2006 (2000).

54. J. Robertson, Carbon 37, 759 (1999).

55. J. Robertson, Mater. Sci. Eng. R 37, 129 (2002).

56. S. Ahmed, M. Moon and K. Lee, Appl. Phy. Lett. 92, 193502 (2008).

57. X. Z. Ding, Y. J. Li, Z. Sun, B. K. Tay, S. P. Lau, G. Y. Chen, W. Y. Cheung and S.

P. Wong, J. Appl. Phys. 88, 3842 (2000).

58. A. T. T. Koh, J. Hsieh and D. H. C. Chua, Diam. Relat. Mater. 19, 637 (2010).

59. E. Stratakis, G. Eda, H. Yamaguchi, E. Kymakis, C. Fotakis and M. Chhowalla,

Nanoscale 4, 3069-3074 (2012).


24

60. A. Malesevic, R. Kemps, A. Vanhulsel, M. P. Chowdhury, A. Volodin and C. Van

Haesendonck, J. Appl. Phys. 104, 084301 (2008).

61. Q. Min, F. Tao, D. Hui, L. Lifeng, L. Haibo, C. Yiwei and S. Zhuo, Nanotechnol. 20,

425702 (2009).

62. Z.-S. Wu, S. Pei, W. Ren, D. Tang, L. Gao, B. Liu, F. Li, C. Liu and H.-M. Cheng,

Adv. Mater. 21, 1756-1760 (2009).

63. S. Pandey, P. Rai, S. Patole, F. Gunes, G.-D. Kwon, J.-B. Yoo, P. Nikolaev and S.

Arepalli, Appl. Phy. Lett. 100, 043104 (2012).

64. J.-h. Deng, R.-t. Zheng, Y.-m. Yang, Y. Zhao and G.-a. Cheng, Carbon 50, 4732

(2012).

65. D. Ye, S. Moussa, J. D. Ferguson, A. A. Baski and M. S. El-Shall, Nano Lett. 12,

1265-1268 (2012).

66. Z. Yang, Q. Zhao, Y. Ou, W. Wang, H. Li and D. Yu, Appl. Phys. Lett. 101, 173107

(2012).

Chapter 2 Experimental techniques

25

Chapter 2

Experimental techniques

In this chapter, the principles behind each experimental technique used will be

described. The techniques used for thin film deposition will be detailed in section 2.1,

while section 2.2 describes the experimental methods used to fabricate nanographite,

graphene and graphene emitters. Lastly, the material characterization techniques will be

detailed in section 2.3.

2.1 Thin film deposition techniques

2.1.1 Magnetron sputtering

Magnetron sputtering is a powerful and flexible technique that can be used to

make coatings of solid metal, alloys and other compounds. It can be used with a direct

current (DC) or radio frequency (RF) source. DC sputtering is not suitable for insulators

as it causes positive charge buildup on the target, which in turn repels the positive argon

(Ar) ions and stops the sputtering process.1 On the other hand, RF allows the sputtering

of insulators and operation at lower Ar pressures, leading to less collisions and better

line-of-sight deposition.2 In this technique, a target, which is the cathode, is bombarded

by energetic ions. These ions, typically Ar ions, are produced in a glow discharge plasma

between the cathode and anode in the vacuum chamber in the range of 10-6

Torr. Upon

ion bombardment, some of the target atoms are dislodged or sputtered and subsequently

condensed onto the substrate. Secondary electrons produced as part of the bombardment


26

process also plays a part in the continuation of the plasma. The difference between

conventional sputtering and magnetron sputtering is that in order to increase the

sputtering rate, a magnetic field is applied. The magnetic field, which is applied

perpendicularly to the electric field, constraints the motion of the electrons causing them

to move in a helical manner near to the target surface region. By limiting the motion of

the electrons near the target, higher sputtering rates can be achieved due to an increased

probability of electron-atom collision that results in higher ionization ratios and a dense

plasma is produced near the target surface. Subsequently, the increase in ionization rates

results in higher deposition rates.3 A schematic of magnetron sputtering is shown in Fig.

2.1.4 In this project, magnetron sputtering was used to deposit nickel (Ni), cobalt (Co),

iron (Fe) and copper (Cu) thin films onto highly doped (n++

) silicon substrates. The metal

thin films were then used as substrates for further experiments.

Fig. 2.1 Schematic of RF magnetron sputtering showing the in-chamber configurations.4


27

2.1.2 Pulsed laser deposition (PLD)

Pulsed laser deposition is a physical vapor deposition technique, which has been

effectively used to fabricate thin films material such as DLC, metal multilayers, ceramics,

superconductive and nanostructured materials.5-9

The advantages of using PLD for thin

film deposition include the ability to grow high quality films, to ablate any material and

to obtain a stoichiometric transfer of target material onto the substrate, which is

especially useful in the case of composite materials.10-12

PLD systems come in different

laser wavelengths depending on the source of laser. Some common wavelengths used are

ArF (193 nm), KrF (248 nm) and ND:YAG (1064, 532, 355 and 266 nm) lasers from the

first to fourth harmonic.13-15 ArF and KrF lasers are also known as excimer lasers that

radiate energy in the ultra-violet range and they make use of a gaseous mixture as laser

source.

The PLD system used in the experiments in this work consisted of a KrF (λ=

248 nm) Lambda Physik laser, an optics system to guide the laser and a homebrewed

deposition chamber as shown in Fig. 2.2. The system consists of a box where the pulsed

laser is produced and a vacuum chamber where the substrate and target are placed. The

substrate and the target are positioned 5 cm apart and facing each other in the chamber,

which is evacuated to a pressure of 10-6

Tor by a turbomolecular pump and a roughing

pump. Short pulses of laser are guided by various focusing optic lens before it enters the

vacuum chamber and strikes the target. The target then absorbs the focused laser and

when the energy density used is sufficiently high, ablation of the target occurs, thereby

expelling a highly energetic and forward directed flux of material, known as the plume,


28

Fig. 2.2 Schematic of PLD used for nanographite in amorphous carbon deposition and graphene

segregation.

Fig. 2.3 Plume seen during the ablation of target.


29

orthogonally from the target surface. The plume, which condenses on the substrate

surface placed directly in front of the target, consists of ions, atoms, excited states,

electrons and macroparticles or fragments of expelled material.6 Figure 2.3 shows the

bluish green plume seen during deposition and also the positions of the laser, substrate

and target in the chamber.

The deposited film properties are affected by several parameters of the PLD

deposition. Ion energy, which is determined by laser fluence, is one of the critical factors

known to affect the type of carbon film formed, whether diamond-like or graphitic

films.16

PLD has the widest range of acceptable energies without requiring any external

biasing compared to various other techniques.9 Furthermore, in addition to the laser

fluence (J/cm2), the repetition rate of the laser pulses, substrate temperature, target to

substrate distance and deposition environment (such as the use of reactive gases) can be

varied, thus making PLD a versatile technique. The typical pulse duration of such lasers

is in the nanosecond range. However in the recent decade, the use of femtosecond pulse

(10-15

sec) duration lasers, which have a 800 nm wavelength have been explored.17-19

The

extremely fast pulse duration increases the energy of the ejected particles up to a few keV

and highly wear resistant DLC materials can be more easily obtained. A detailed review

on PLD and its mechanisms related to film growth can be found in the review by

Willmott and Hubber in reference 20.20


30

2.2 Graphene/graphene films fabrication methods

2.2.1 Modified Hummers method

Generally, modified Hummers method first involves the oxidation of graphite into

graphite oxide through the intercalation and expansion of the graphite lattice. Graphite

oxide is then exfoliated into graphene oxide (GO) through ultrasonication or stirring in

water and GO is reduced to form graphene. The reduction of GO is an important step as

GO is an insulator and has different properties from graphene. This method uses natural

graphite powder, potassium permanganate (KMnO4), sodium nitrate (NaNO3),

hydrosulfuric acid (H2SO4), hydrogen peroxide (H2O2) aqueous solution and ultrapure

water. Graphite powder and NaNO3 was added to H2SO4 with the jar in an ice bath.

While stirring, KMnO4 was slowly added to the mixture to prevent the temperature from

exceeding 20 oC. Upon completion of this step, the ice bath was removed and the

temperature was maintained at 35 oC for 30 min. Water was then stirred in, causing

effervescence and raising the temperature to 98 oC. The brown colored suspension was

kept at this temperature for 15 min and more water was added. H2O2 was then added to

reduce the residual KMnO4. After the reduction process, the suspension consisting of

graphite oxide is a bright yellow color. This graphite oxide mixture was then filtered and

washed with water and remaining metal ions were treated and removed. Exfoliation of

graphite oxide into GO can be done by ultra-sonication. GO was reduced to graphene and

this was achieved by microwave assisted reduction of its aqueous suspension for the

graphene flakes used in this work.21


31

2.2.2 Electrophoretic deposition (EPD)

Electrophoretic deposition is a widely used technique in the industry for the

fabrication of phosphors for high resolution displays as well as coatings.22, 23

Its

advantages include high deposition rate, low cost, ability to deposit on substrates with

complex shapes, ease of up-scaling due to its simplicity and the flexibility to deposit

many materials. It is also capable of producing homogeneous films with a high packing

density.24

In this technique, two electrodes are placed in a colloidal suspension of the

material. In our setup, the cathode is the substrate and the anode is a graphite electrode.

During the preparation of the suspension, a charger material, such as Mg(NO3)2,

La(NO3)3, MgCl2, or Al(NO3)3, is added to the solution. The metal ions are preferentially

Fig. 2.4 Schematic of EPD setup with planar electrodes.


32

adsorbed onto the powder particles, charging it in the process. EPD is then achieved by

two steps. First, when an electric field is applied, charged particles in the suspension

move towards the electrode. Second, the particles are accumulated at the respective

electrode to form a coherent deposit.24

A schematic of the EPD setup is shown in Fig. 2.4.

2.3 Materials characterization

2.3.1 Scanning electron microscope (SEM)

Scanning electron microscope is a useful and flexible imaging technique for

surface and subsurface analysis of solid materials. The probe in SEM is an electron

source from a tungsten hairpin (thermionic source), LaB6 crystal or field emission gun

(FEG).25

To obtain surface images, Philips XL-30 and Zeiss Supra SEM, both with FEG

sources, were employed. SEM allows the viewing of specimens too small to be examined

by a light microscope. Since the wavelength of electrons is much smaller than that of

visible light, it has a higher resolving power. Over the years, the resolution for SEM has

improved from 50 nm to 1–5 nm today and from a simple tool it has now become an

equipment widely used in many areas of such as biology, chemistry and metallurgy.26

In its primary mode, the secondary electrons are detected in SEM to give us

images of the sample.27

The images appear three dimensional due to the small angular

aperture of the electron probe giving the large depth of field that the SEM possesses .As a

result, rough surfaces such as pollen grains stay in focus throughout the whole sample

image.


33

No lenses are involved in SEM image magnification. Instead, the electrons raster

across the surface of the specimen and at the same time, a spot of the cathode ray tube

(controlled by current from the detector) scans across a screen in a similar rectangular

fashion except that the area is far larger. SEM imaging is always carried out in a vacuum

environment as the presence of gas molecules may cause the instability of the electron

beam from the source. Also, samples are required to be conductive to avoid charging,

which may cause a distorted image to form. Thus, samples that are non-conductive have

to be sputtered with a layer of gold first. In addition, a SEM equipped with energy

dispersive spectroscopy capabilities can also give us the elemental composition of the

sample by analyzing the characteristic x-rays given off by the sample.27

2.3.2 Transmission electron microscope (TEM)

Transmission electron microscopy is an imaging technique whereby a beam of

electrons, focused by magnetic lens, passes through the thickness of the sample,

interacting with the specimen as this happens and forms a two dimensional projected

image of a three dimensional specimen.28

Some basic instrumentation in TEM are an

electron gun to illuminate the sample, magnetic lens to demagnify and control the size of

the beam hitting the sample, fluorescent viewing screen, a charge-coupled device (CCD)

camera. A detailed setup of TEM can be found in ref. 27.27

The point to point resolution

of TEM is reported to be 0.1 nm or better. Such a resolution allows viewing of the

arrangement of individual atoms, thus allowing the periodic lineup of atoms in crystalline

structures to be seen.


34

TEM viewing is also carried out in a vacuum environment. Also, another

requirement is that the sample must be thin enough to be transparent to electrons. For

nanoparticles, specially designed TEM copper (Cu) grids are used to support the film

which is deposited on top of the grids and such samples can be viewed directly without

polishing or grinding.27

2.3.3 Raman spectroscopy

Raman spectroscopy is a vibrational spectroscopy technique that can obtain

information about molecular vibrational motion and acquire fingerprints to help identify

and quantify samples.29

It is a flexible technique that can be used on solids, liquids,

powders and films. It can also be used on mixtures and composites as well and the

resulting spectrum is just a superposition of the individual components. It does not

require a vacuum environment to operate in and the spectra can be acquired relatively

quickly. No sample preparation is needed and the spectra are characteristic of each type

of material. Raman spectroscopy is especially useful in the study of carbon materials as it

is highly sensitive to symmetric covalent bonds that have little or no dipole moment

which is the case of carbon-carbon bonds.30


35

Fig. 2.5 Three possible outcomes during the interaction of photon and matter in Raman

spectroscopy.

In this technique, a monochromatic light or laser, in the visible, infrared or

ultraviolet range is focused by a microscope onto the sample. The laser interacts with the

sample and causes excitation of the molecules to a virtual energy state. Depending on the

nature of the interaction of the photon with the sample, three outcomes, as depicted in Fig.

2.5, can occur when the molecule relaxes and the scattered photon is collected.31

Firstly,

elastic scattering or Rayleigh scattering is said to have occurred if the scattered photon is

of the same frequency as the incident photon, and this happens to the majority of photons.

If the molecules relax into a different vibrational state, the energies of the scattered

photons are shifted either up or down and this is also known as the Raman Effect. The

difference in the energy between the incident photon and scattered photon provides

information on the microstructure of the sample. In the case where the final vibrational

state of the molecule is higher, the scattered photon is of a lower frequency in order to

maintain the energy balance of the system and this is called a Stokes shift. Conversely, if

the final vibrational state of the molecule is lower and the scattered photon is of a higher

frequency, this is known as an Anti-Stokes shift. A Renishaw Raman spectrometer 2000


36

with 514.51 nm green argon laser was used in this project. The laser beam was focused

onto the sample surface using an optical microscope with a magnification of 50 times at

spot size of 1 μm. The Raman spectra were acquired in the range of 1000–3000 cm-1

.

2.3.4 X-ray photoelectron spectroscopy (XPS)

X-ray photoelectron spectroscopy is a surface analysis technique that provides

chemical state and compositional information from an outer surface thickness of 1–10 nm.

It is capable of detecting all elements except helium and hydrogen.32, 33

Comparing

various surface analysis instruments like Auger Electron Spectroscopy (AES) and

Secondary Ion Mass Spectroscopy (SIMS), it is the easiest to obtain chemical state

information from XPS data. Thus, XPS is the preferred choice when information in this

area is required.32

For our purpose, a monochromatic XPS (Al Kα (1486.6 eV)) operated

at 275 W with a pass energy of 10 eV (narrow scan) was used.

Instrumentation in XPS consists of several components. They include a

preparation chamber for sample cleaning and experiments, analytical chamber where the

photon source is present, a hemispherical electron analyzer to measure the kinetic energy

(KE) of the photoelectrons and a detector at the end. X-ray photons bombard the surface

atoms of the sample. With sufficient energy, the photons can excite the core electrons of

the atoms and cause the ejection of a photoelectron from an atom. The KE of the

photoelectron ejected is then measured and converted into binding energy. This binding

energy is characteristic of the element and its oxidation state. Binding energy, EB, can be

calculated from equation 2.1,34


37

EB = hυ – Ek – (2.1)

where, EB is the binding energy of the photoelectron, h is Planck’s constant, υ is the

frequency of the x-ray source, Ek is the kinetic energy of the photoelectron and is the

work function of the spectrometer.

From the equation, it can be seen that the ejected photoelectron must reach the

detector with an undisrupted path or else analytical information will be lost. As such, an

ultra-high vacuum environment is required in XPS, to fulfill this condition as well as to

prevent surface contamination. The sample surface can also be cleaned by argon sputter

to remove any oxidized layer or contaminants before the information is acquired.

2.3.5 Atomic force microscopy (AFM)

Atomic force microscopy is a type of scanning probe microscope that has superior

resolving power. It consists of a scanning tip rastering across the surface to obtain a

morphology image of atomic resolution. A Veeco, Nanoscope 3, Multimode AFM was

used to characterize the films produced.

Unlike SEM and scanning tunneling microscopy, AFM works for both conductive

and non-conductive samples.35

AFM can also operate in an ambient environment. The

probe in AFM is also known as a cantilever, which is essentially a probe with an

extremely sharp tip scanning across the surface to produce a three dimensional surface


38

Fig. 2.6 Schematic of AFM.36

map. When the tip is brought close to the sample surface, the cantilever may be deflected

according to surface changes. The deflection of the laser off the back of the cantilever is

monitored by the photodiode as seen in the schematic diagram of AFM shown in Fig. 2.6.

The diameter of the tip is less than 40 nm and usually made from silicon or silicon

nitride.36

AFM can be operated in a few modes including contact and tapping mode. For

our thin films sample, tapping mode is used. In tapping mode, the tip only contacts the

sample intermittently, for a much shorter time as compared to contact mode. The

cantilever in tapping mode oscillates at its resonant frequency and minimizes the chances

of damage especially to soft samples.36


39

2.3.6 Field emission testing (FE)

1) Parallel Plate Setup

Electron emission, which is a quantum tunneling effect, was tested with a self-

made electron emission tester. The setup used was in the form of a parallel plate

configuration with the cathode being the sample and the anode was an ITO-coated glass

slide. The cathode and anode are separated by a 100 µm polymer spacer layer with a cut

hole in the center of the film to allow electron emission only from that area. During

testing, the voltage was incrementally increased until significant current was produced.

Turn-on field with this setup is defined as the voltage needed to produce a current density

of 10 µA/cm2. The readings were taken with a LabView 8 software Keithley 2410

voltage source and current measurement unit. The measurements were taken under

vacuum conditions at a pressure of 1×10-6

Torr maintained by a turbomolecular pump. A

schematic cross-sectional diagram of the setup is shown below in Fig. 2.7.

Fig. 2.7 Parallel plate electron emission tester setup.

e-


40

2) Probe tip setup

The difference between the probe tip and parallel plate setup is that in the former,

the anode is a tip (tungsten in our case) and the cathode-anode distance can be varied

from 0–100 µm. The self-made setup was specially fitted with a micrometer screw

gauges outside the vacuum chamber that allowed adjustments in the X-Y-Z axis without

breaking the vacuum. Turn-on voltage with this setup is defined as the voltage needed to

produce a current of 1 nA. The self-made probe tip emission setup used in our

experiments is shown in Fig. 2.8.

Fig. 2.8 Probe tip electron emission tester setup.


41

References

1. P. D. Davidse, Vacuum 17, 139-145 (1967).

2. J. L. Vossen, J.Vac. Sci. Technol. 8, S12-S30 (1971).

3. P. J. Kelly and R. D. Arnell, Vacuum 56, 159-172 (2000).

4. D. Song, Appl. Surf.Sci. 254, 4171-4178 (2008).

5. N. Benchikh, F. Garrelie, C. Donnet, B. Bouchet-Fabre, K. Wolski, F. Rogemond, A.

S. Loir and J. L. Subtil, Thin Solid Films 482, 287 (2005).

6. F. Bonaccorso, C. Bongiorno, B. Fazio, P. G. Gucciardi, O. M. Maragò, A. Morone

and C. Spinella, Appl. Surf.Sci. 254, 1260-1263 (2007).

7. R. Janmohamed, J. J. Steele, C. Scurtescu and Y. Y. Tsui., Appl. Surf.Sci. 253, 7964

(2007).

8. A. H. Jayatissa, F. Sato, N. Saito, Y. Hirano and K. Takizawa, Carbon 38, 1145

(2000).

9. A. A. Voevodin and M. S. Donley, Surf. Coat. Techol. 82, 199 (1996).

10. P. Murray and D. Peeler, J. Electron. Mater. 23, 855-859 (1994).

11. H.-U. Krebs, M. Weisheit, J. Faupel, E. Süske, T. Scharf, C. Fuhse, M. Störmer, K.

Sturm, M. Seibt, H. Kijewski, D. Nelke, E. Panchenko and M. Buback, edited by B.

Kramer (Springer Berlin / Heidelberg, 2003), Vol. 43, pp. 101-107.

12. H. M. Christen and G. Eres, J. Phys. Condens. Matter 20, 264005 (2008).

13. F. Xiong, Y. Y. Wang, V. Leppert and R. P. H. Chang, J. Mater. Res. 8, 2265-2272

(1993).

14. J. J. Gaumet, A. Wakisaka, Y. Shimizu and Y. Tamori, J. Chem. Soc., Faraday Trans.

89, 1667-1670 (1993).

15. A. T. T. Koh, J. Hsieh and D. H. C. Chua, Diam. Relat. Mater. 19, 637-642 (2010).


42

16. G. Reisse, B. Keiper, S. Weissmantel and U. Falke, Appl. Surf.Sci. 127–129, 500-506

(1998).

17. F. Garrelie, A. S. Loir, C. Donnet, F. Rogemond, R. Le Harzic, M. Belin, E.

Audouard and P. Laporte, Surf. Coat. Techol. 163–164, 306-312 (2003).

18. N. Benchikh, F. Garrelie, C. Donnet, B. Bouchet-Fabre, K. Wolski, F. Rogemond, A.

S. Loir and J. L. Subtil, Thin Solid Films 482, 287-292 (2005).

19. F. Qian, V. Craciun, R. K. Singh, S. D. Dutta and P. P. Pronko, J. Appl. Phys. 86,

2281-2290 (1999).

20. P. R. Willmott and J. R. Huber, Rev. Mod. Phys. 72, 315 (2000).

21. W. Chen, L. Yan and P. R. Bangal, Carbon 48, 1146-1152 (2010).

22. K. Y. Sasaki and J. B. Talbot, Adv. Mater. 11, 91-105 (1999).

23. B. J. C. Thomas, A. R. Boccaccini and M. S. P. Shaffer, J. Am. Ceram. Soc. 88, 980-

982 (2005).

24. A. R. Boccaccini, J. Cho, J. A. Roether, B. J. C. Thomas, E. Jane Minay and M. S. P.

Shaffer, Carbon 44, 3149-3160 (2006).

25. P. J. Goodhew, F. J. Humphreys and R. Beanland, Electron Microscopy and Analysis.

(Taylor & Francis, UK, 2001).

26. A. McPherson, A. J. Malkin, Y. G. Kuznetsov and M. Plomp, Acta Crystallogr. Sect.

D 57, 1053-1060 (2001).

27. J. Goldstein, Scanning Electron Microscopy and X-Ray Microanalysis. (Kluwer

Academic/Plenum Publishers, New York, 2003).

28. X. F. Zhang and Z. Zhang, Progress in Transmission Electron Microscopy. (Springer,

New York, 2001).


43

29. J. R. Ferraro, K. Nakamoto and C. W. Brown, Introductory Raman Spectroscopy.

(Academic Press, Amsterdam, 2003).

30. Y. Wang, D. C. Alsmeyer and R. L. McCreery, Chem. Mater. 2, 557-563 (1990).

31. E. Smith and G. Dent, Modern Raman Spectroscopy: A Practical Approach. (John

Wiley & Sons, Chichester, 2005).

32. P. E. J. Flewitt and R. K. Wild, Physical Methods for Materials Characterisation.

(Institute of Physics Pub., Bristol, 2003).

33. A. J. Garratt-Reed and D. C. Bell, Energy-dispersive X-ray Analysis in the Electron

Microscope. (BIOS, Oxford, 2003).

34. H. S. Nalwa, Advances in Surface Science. (Academic Press, San Diego, 2001).

35. D. G. Seiler, Characterization and metrology for ULSI technology: 2003

International Conference on Characterization and Metrology for ULSI Technology,

Austin, Texas, 24-28 March 2003. (American Institute of Physics, USA, 2003).

36. Uppsala University, Department of Pharmacy,

(http://www.farmfak.uu.se/farm/farmfyskem-web/instrumentation/afm.shtml, 2012)

http://www.farmfak.uu.se/farm/farmfyskem-web/instrumentation/afm.shtml

Chapter 3 Fabrication of Nano-graphite in a-C matrix field emitters

44

Chapter 3

Fabrication of nano-graphite in amorphous

carbon (a-C) matrix field emitters

In this chapter, carbon was deposited on Si substrates at various temperatures

using the PLD technique. The effect of temperature on the microstructure, surface and

field emission properties of the resulted films were studied. By relating the field emission

characteristics and the microstructure of the films, the emission mechanism of

nanographite in a-C film was investigated.

3.1 Introduction

Carbon materials for field emission applications have been a subject of great

interest for decades. The various polymorphic forms of carbon such as carbon nanotubes

(CNT), diamond and amorphous carbon (a-C) and diamond-like carbon (DLC) have been

demonstrated as potential field emitters.1-3

On one hand, while it is rather clear that the good emission properties of CNT

stems out from its high aspect ratio morphology and good conductivity from orientated

sp2 bonded graphene sheets, electron emission from planar amorphous carbon films on

the other hand, tends to be a topic of much debate. CNT has been favored for its very low

turn-on fields, in fact the lowest reported for the carbon family, but it suffers from


45

stability issues and tends to be burnt out easily.4 As the focus shifts towards flat emitters

which are mechanically stable, easily deposited on a variety of substrates and possess a

low work function (4–5 eV), planar amorphous carbon films, whether doped, undoped,

hydrogenated (a-C:H) or unhydrogenated, has been given a fair share of attention ever

since it demonstrated potential to be a cold cathode material during the 1990s.

Over the years, various models have been proposed for electron emission from

amorphous carbon. Several popular and widely accepted models include the Amaratunga

and Silva model5 on nitrogen doped a-C:H films, which has been used as a reference for

other amorphous carbon films. Nitrogen doped a-C:H was a popular choice of doped

films due to nitrogen being a shallow donor, thus allowing for the easy formation of a

space-charged layer that leads to sharper band bending. In this model, the n++

silicon

substrate is the true cathode, while the nitrogen doped a-C:H film is regarded as a space

charged layer through which electrons coming from the substrate is accelerated. These

electrons, termed “hot electrons”, subsequently acquire energies near to the vacuum level

and are thus able to emit at lower turn-on voltages. While models that involve these space

charged controlled mechanism may very well explain nitrogen doped a-C:H films, they

may not be apt for undoped nanostructured a-C or DLC films which are non-

hydrogenated. It was also briefly suggested that DLC films with the widest bandgap and

highest sp3 would be best for field emission.

6 This was later disproved in an in-depth

study which led to the consensus that the sp2 phase in a-C films plays an important role in

the field emission process.5 The conductive sp

2 elements facilitate conduction through the

film by forming electron paths or channels to deliver electrons to the surface to be

emitted. Sp2 bonded carbon atoms in planar films can come in the form of short olefinic


46

chains as in DLC (low sp2) to nanoclustered graphite in a-C matrix (intermediate sp

2

content) and finally fully sp2 film made of graphite sheets in the case of graphene or

nanographite. In the case of flat films in which the field enhancement factor () is unity,

it was documented that a field in the order of 500 V/µm is required for emission,7

However, it only takes less than a tenth of that value for a-C films.

Since DLC is known for its smoothness in the range of root mean square (RMS)

0.1 nm, there have been doubts about its field enhancing ability. In several studies of

DLC, β has been given the value of 1 which translates to no enhancement and lowered

emission turn-on fields are a result of lowered work function or electron affinity values.8,9

In addition, there have been various reports on improved emission with doped DLC films

with silver10

, nickel11, 12

or titanium11

which formed conductive metal nanoclusters. The

clusters in turn formed conductive paths within the largely insulating diamond-like

matrix and led to relatively lower emission turn-on voltages. However, these electron

channels have not been directly proven or seen.

In this chapter, it was experimentally showed that the electron channels present in

a-C films gave it a CNT-like or tube-like emission behavior when the emission testing

distance was varied. Changes in emission characteristics with deposition temperature

were also studied. It is worth noting that no conditioning was done on the films as there

have been doubts about the nature of emission from a-C films whether it is intrinsic or

caused by surface and/or internal damage due to a conditioning process which involves

the ramping of voltages in cycles from low to high and vice versa prior to testing. 7, 13


47

3.2 Experimental method

The films were deposited using a carbon target (99.9% purity, 325 mesh) with a

KrF (λ = 248 nm) Lambda Physik excimer pulsed laser system with a laser fluence of 10

J/cm2, pulse duration of 25 ns and frequency of 20 Hz. The target was rotated with a

speed of 60 rpm and fixed at a distance of 5 cm from the highly n doped (n++

) Si (100)

substrate. The sp2 content was controlled by varying the temperature from 100 to 700

oC

in increasing steps of 100 oC using a heater stage in vacuum. The base pressure was kept

below 5 × 10-6

Torr during deposition.

Raman spectra were recorded at room temperature with a Renishaw system using

an excitation source of 514.5 nm lines from Ar+ laser. The spectra were acquired in the

range of 1000–2000 cm-1

and accumulated for 10 times. The laser power used was 25

mW to avoid sample damage. Surface morphology and surface roughness were obtained

with a Digital Instruments, Nanoscope III, Multimode AFM operated in tapping mode.

AFM images acquired were flattened using the software provided by the manufacturer

and not modified by any filtering. The RMS roughness was evaluated over an area of 1 ×

1 m. XPS was carried out with a Kratos axis ultra DLD system monochromatic XPS (Al

Kα at 1486.6 eV source). The XPS spectrometer was operated at 225 W to obtain the

spectra of the samples. By deconvoluting the carbon peaks, chemical bonding and sp2

content of the films were determined. The spectra were obtained in a concentric

hemispherical analyzer with a pass energy (Ep) of 20 eV in narrow scan mode. A JEOL

JEM 3010F HRTEM (LaB6) operated at 300 kV was used for TEM imaging to observe

the nanoclusters directly.


48

Field emission current versus anode voltage (I–V) was measured with a probe tip

setup. The anode was an electrochemically etched tungsten needle with a tip diameter of

200 µm, while the cathode was the sample. Testing was conducted in at a base pressure

of ~4 × 10-6

Pa. The zero-distance point was determined by first moving the anode with 7

V of applied voltage in contact with the cathode and subsequently moving the anode

away from the cathode to a point where the current diminishes to zero. On the same spot

of the sample, the distance between the anode tip and substrate was adjusted using a

micrometer stage with an accuracy of 0.5 µm. The anode-substrate distances were varied

from 1 to 100 µm without breaking vacuum. The turn-on voltage (Von), at each distance

was the voltage needed to produce a current of 1 nA and the current limit was set at 1 µA

to prevent damage to the anode.


49

3.3 Microstructural characterization of nano-graphite

clusters in a-C matrix

3.3.1 Effect of temperature on microstructure

The changes in microstructures and the degree of sp2 clustering in films with

increasing deposition temperature were examined using Raman Spectroscopy. As shown

in Fig. 3.1, a broad peak centered on 1560 cm-1

, which is the typical spectrum for

amorphous carbon, was observed for the room temperature and 100 oC samples, while

two distinct peaks were observed for 400 and 700 oC samples.

14, 15 These peaks can be

deconvoluted into two Gaussian peaks known as the disorder and graphitic peak, i.e. the

D and G peak, which corresponded to the breathing mode of the six fold aromatic rings

and the in-plane C–C stretching mode of sp2 hybridized carbon atoms, respectively.

Unlike the D peak which only occurs due to the hexagonal rings, the G peak occurs for

all sp2 sites inclusive of sp

2 chains and rings.

14 The full width half maximum (FWHM),

the intensities of D (ID) and G (IG) peaks, and the ID/IG ratio of all the samples are

summarized in Table 3.1. Different studies have shown that the ratio of D to G peak

intensity, i.e. the ID/IG ratio, can be used to estimate the sp3 bonding in a-C, in which a

higher ID/IG ratio corresponds to a lower fraction of sp3 bonding present in the films.

16, 17

It was observed that samples deposited at room temperature and those at lower

temperatures (100–300 oC) had lower ID/IG ratio and possessed a high percentage of sp

3

bonding, akin to diamond type bonding. The sp2 component here mainly exists in the

form of olefinic chains intertwined with a sp3 carbon network. As the temperature

increased, the ID/IG ratio of the samples increased and the separation between D and G


50

peaks increased. The changes in ID/IG ratio and D and G peak positions in relation with

temperature were illustrated in Fig. 3.2 and 3.3, respectively. The D peak became more

obvious and shifted towards a lower 1350 cm-1

and G peak shifted towards a higher 1600

cm-1

, which are peaks observed in a commercial graphite spectrum.12

Above 400 oC, a

high degree of aromatic clustering was observed from the distinctly separated D and G

peaks, which indicated that certain elements in the films were becoming graphite-like.

Fig. 3.1 Raman spectrum for films deposited at room temperature, 100, 400 and 700 C.


51

Table 3.1 D and G peaks positions, amplitude, relative intensity ratio and estimated sp2 content of

the films deposited at different temperatures.

Sample

D peak

position

(cm-1

)

D peak

amplitude

(ID)

G peak

position

(cm-1

)

G peak

amplitude

(IG)

ID/IG sp

2 content

(%)

0 1432 71 1574 180 0.39 32

100 1424 552 1574 861 0.64 49

200 1422 905 1575 956 0.95 69

300 1413 569 1578 558 1.02 74

400 1385 398 1576 369 1.08 78

500 1382 373 1580 340 1.10 79

600 1383 591 1583 526 1.12 81

700 1376 473 1585 411 1.15 83


52

Fig. 3.2 Raman ID/IG ratio and sp2 content changes of samples deposited at temperature from

100–700 oC.

Fig. 3.3 Raman D and G peak positions of samples deposited at temperature from 100–700 oC.


53

XPS was employed on all the samples in order to confirm the estimated sp2

percentages. The XPS spectra for selected samples are as shown in Fig. 3.4 and were

deconvoluted into 5 Gaussian peaks. The binding energy at 284.5, 285.2, 286.5 and

288.0 eV are corresponding to sp2, sp

3, C–contaminated and C=O, respectively.

18 By

calculating the area under the peak sp2/(sp

2+sp

3), the sp

2 content were determined and are

summarized in Table 3.1. From the ratios of the peak contributions in XPS and Raman,

the sp2 percentage calculations ranged from 32 to 83 % from room temperature to

deposition at 700 oC. Thus, this confirmed the effects of heating in graphitizing the sp

3

bonding of DLC.

Fig. 3.4 XPS spectra for films deposited at room temperature, 100, 400 and 700 C.


54

Fig. 3.5 TEM micrograph for films deposited at (a) 100 C (b) 400 C (c) 700 C. Inset

shows the low magnification area view and SAED of crystalline cluster with lattice spacing 0.33

nm corresponding to carbon.


55

The clusters in the samples can be directly observed in TEM images shown in Fig.

3.5. The TEM image of 100 ºC sample in Fig. 3.5(a) showed no discrete nanocrystalline

clusters. On the other hand, clusters were observed in Fig. 3.5(b) and 3.5(c), which

corresponded to the 400 and 700 ºC samples respectively. A higher cluster density was

also observed for the 700 ºC sample in lower magnification images as seen in figure

insert 3.5c(ii). These results related well with the Raman results, where increased

clustering was seen from the separated D and G peaks. A plot of the cluster density

versus temperature is shown in Fig. 3.6, where the density at 100, 400 and 700 oC is 0,

1.27 x 1011

, and 3.65 x 1011

clusters/cm2, respectively. In addition, selected area electron

diffraction (SAED) (inset 3.5c(i)) showed that the clusters are crystalline with a lattice

spacing of 0.333 nm, thus confirming that the clusters are indeed made up of graphite and

thus fully sp2 bonded. This is in great contrast to a largely amorphous carbon film

Fig. 3.6 Cluster density vs. temperature of 100, 400 and 700 C samples.


56

Fig. 3.7 AFM surface morphologies of (a) 100 C (b) 400 C (c) 700 C samples.

deposited with minimal external heating where only short olefinic sp2 chains are present

in such films as seen in Fig. 3.5(a) for the 100 oC sample. These short chains in the

carbon network cannot be directly observed from the images and thus appear to be

featureless with no lattice fringes (lines) when viewed under TEM as the ordering of the

atoms are random with no particular order.

AFM topological images in Fig. 3.7 show the surface morphology of the films.

There was a slight increase in RMS roughness when the temperature was increased with

the surface roughness values 0.25 and 0.35 and 0.39 nm for the 100, 400 and 700 ºC

samples, respectively.


57

3.3.2 Discussion

Experimental results showed that sp2 bonding increased with increasing

deposition temperature. Raman spectroscopy revealed that the ID/IG ratio also increased

with increasing deposition temperature and ID/IG ratio has been reported to be

proportional to the square of correlation length (La) or size of nanocluster.19

As such, the

increase in ID/IG ratio with increasing deposition temperatures corresponded to an

increase in La in diamond-like films and increase in cluster size in the case of

nanographite films, as evidenced by TEM.

This formation of clusters is a temperature driven phenomenon.19, 20

At room

temperature deposition, energetic carbon atoms that penetrate into the growing film on

the substrate possess poor mobility, and carbon adatom mobility is also reduced due to a

low substrate temperature. As a result, high sp3 films are formed by a subsurface growth

process otherwise known as subplantation.21

However, with an increase in temperature,

the conversion of randomly distributed sp2 and sp

3 a-C into ordered clusters and aromatic

rings occurs and is facilitated by the increase in thermal energy.22

With higher deposition

temperatures, the subplanted carbon atoms have improved mobility and can diffuse

within the growing film, hence aromatic cluster aggregation is enhanced.23

The formation

of sp2 bonding can be enhanced by 1) heating during deposition or 2) annealing after

deposition and growth is completed. The microstructure differences between the two do

not differ greatly, but the transition temperature from a highly sp3 to sp

2 film is

significantly lower for heating during deposition. This transition temperature from

amorphous DLC to nanographite clustered a-C films for post-deposition annealing occurs


58

near 800 oC

24 whereas in our case, the transition occurred below 400

oC. This lowered

transition temperature can be explained through the activation of infrequent events

between pulses of impinging carbon ions onto the substrate, during which a period of 1 ps

of elevated temperature or thermal spike occurred, as shown by Molecular Dynamics

simulation results.25

As such, the heating during deposition is a more effective method to

achieve nanostructuring through heat. It was also observed that through sufficient heating,

substantial clustering of sp2 occurs and a preferential direction may be adopted by the sp

2

phase, resulting in a parallel or perpendicular orientation with respect to the substrate.26


59

3.4 Field emission study of nanographite clusters in a-

C matrix

3.4.1 Field emission properties

In order to investigate the field emission properties of the films, I-V characteristic

curves were obtained at various distances. The I-V characteristics of the films deposited

at 100, 400 and 700 oC is shown in Fig. 3.8. The spread of the curves at various distances

reflected the ease of emission with the 700 oC sample being the easiest emitter and 100

oC being the most difficult. To illustrate the differences, Von at 50 µm for the 3 samples

were 1760, 1420 and 1110 V, respectively for the samples in increasing temperatures.

Fowler-Nordheim (F-N) plots were also obtained using data from a single anode-cathode

distance of 50 µm in Fig. 3.9, which showed a linear trend between ln (I/V2) and (1/V),

indicating electron emission was of a metallic nature. By taking ratio of the slopes of the

plots according to equation 1.4, we can obtain a relation as follows,

(3.1)

In the temperature range tested, the change in work function of the films tested

was 0.4 eV.27

As such, when taking ratio, the term φ1/φ2 can be approximated to 1. The

enhancement factors of such flat film emitters can be divided into two parts; the internal

enhancement (βin) and external geometrical enhancement factor (βex) and the total β is

composed of the product of the two.26

Factors that affect the external enhancement factor

1

22/3

2

1

2

1 )(


60

Fig. 3.8 I-V characteristics of films deposited at 100, 400 and 700 oC.


61

are geometrical in nature and are related to the physical dimensions of the emitter feature

such as a tube or needle. In this case, external enhancement is usually taken to be unity.

Internal field enhancement, on the other hand, is typically used for planar emitters whose

surfaces usually appear featureless. The combination of microstructure and composition

of such materials thus determines the internal field enhancement factor.26

In this work,

the surface morphology in AFM images showed a slight roughness increment when

deposition was carried out at higher temperatures. However, such small changes in

surface topology are insufficient to qualify as geometrical enhancement sites for efficient

field emission. The best example of enhancement due to physical or geometrical

structuring is CNT. They have a very high aspect ratio28

ranging from 20 to ~10000 in

extremely high aspect ratio29

CNT. In contrast, the aspect ratio of the protrusions found in

our samples was less than 0.1, hence geometrical enhancement was not considered for

such flat films. Thus, βex was taken to be unity and equation 3.1 becomes,

(3.2)

This meant that the ratio of the slopes corresponded to a ratio of internal enhancement

factors (βin) such that β100: β400: β700 = 1: 1.48: 1.90. This ratio is expected not to

change with different inter-electrode distances as it represents the difference between the

samples. However, the values of the individual β of the samples at different distances will

change.

Using the data obtained from varied anode-cathode distances, the Von

versus distance (Von-D) curves with an inverted y-axis was plotted to illustrate the

voltage drop in vacuum as in Fig. 3.10. In obtaining data for these plots, anode-cathode

1

2

2

1

i n

i n


62

Fig. 3.9 Fowler-Nordheim plots at 50 µm for films deposited at 100, 400 and 700 oC.

Fig. 3.10 Voltage-distance (V-D) curves for nanographite in a-C films deposited at 100, 400 and

700 oC.


63

distances were varied in vacuum from 1 to 100 µm on the same spot of the sample. By

inverting the axis, the graph is the potential profile of electrons in vacuum. A detailed

illustration of the graph conversion can be found in ref. 30.30

From this, the electric field

in vacuum during emission was obtained by determining the slopes (Λ) of the Von-D

curves and the majority of the voltage drop was seen in vacuum. The electric fields in

vacuum for samples deposited at 100, 400 and 700 oC were 33.1, 23.9 and 16.4 V/µm

respectively. It can also been seen that for the 700 oC sample, the bending of the vacuum

level was the least required for emission. Thus, at any tested distance the barrier for

emission was lowest for the 700 oC sample.

Subsequently, Wentzel–Kramers–Brillouin (WKB) approximation, which has been

used to describe the tunneling probability of a triangular barrier formed when the vacuum

level potential bends under a strong electric field, was used to obtain a relationship

between β and electric field in vacuum (Λ).31

In this approximation, the equation for

transmission coefficient is,

D= (3.3)

where is the barrier height, m is the electron effective mass, and Λ is the electric field

in vacuum, which was determined from the slopes of the Von-D curves. From the WKB

equation, the barrier height ratio was related to the electric field in the form β 1/ β 2 = Λ 2/

Λ1, identical to the relationship derived from the F-N equation. Calculating the barrier

height ratios then gave β100: β400: β700 = 1: 1.46: 2.02, agreeing well with the barrier

height ratios obtained independently from the F-N relationship.

2

3

3

28exp

he

m


64

From the typical I-V curves, the turn-on field, defined as the voltage divided by the

anode-cathode separation, was plotted as a function of the separation for samples

deposited at 100, 400 and 700 oC. The applied turn-on fields (Eappl) increased

exponentially with decreasing anode-cathode distance as seen from Fig. 3.11. In opposite,

with the increase in anode-cathode distance, Eappl for the samples decreased and settled to

a saturated value after the distance of about 10 µm. Such turn-on fields versus distance

line shapes are typically observed for CNT and other tip emitters, thus hinting a possible

tube-like emission mechanism from the nanostructured graphite films.

Fig. 3.11 Turn-on field versus anode-cathode distances of films deposited at 100, 400 and 700 oC.

Inset (a) shows typical line shapes with 2 different height approximations determined from

simulations in Ref. 37. Inset (b) shows line details at small distances.


65

3.4.2 Emission mechanism of nanographite clusters in a-C

matrix

The tube-like emission from atomic smooth DLC can be revealed by considering

the parameters and emission mechanism in materials. In field emission studies, emission

properties are commonly summarized to a parameter that serves as a measure of field

enhancing ability of the sample, otherwise known as the field enhancement factor, β.

However, comparison of β values should be made with caution as β depends strongly on

experimental setup and even more so on anode-cathode distances.32

As such, it is more

meaningful to present the series of β values as a ratio across our samples since anode-

cathode distance was varied. In addition, a comparison of β for samples in this work

against various carbon-based field emitters can be found in Table 3.2.

The process of field emission occurs in 3 steps; first, electrons are injected from

the back contact into the film. Second, the electrons will have to travel within the film

and lastly be transported to the surface where the electrons emit into vacuum in this last

step. To improve field emission properties, either of the steps should be enhanced, or

ideally all steps should be enhanced, either through use of different substrates, dopants or

surface treatments. For instance, using a semiconductor as substrate can allow only

electrons with sufficient energy to tunnel through the substrate-film heterojunction and

give a higher energy electron population transversing the film.33

Additionally, the

limiting step differs for various films such as for polymeric films like nitrogen doped a-

C:H films, electron injection is the crucial factor while for diamond-like carbon, the

barrier mainly lies at the film front. Sp2 bonded graphite is highly conductive while sp

3


66

Table 3.2 Comparison of field enhancement factors of various carbon-based emitters.

Sample Type of Anode Inter-electrode

distance

Field

enhancement

factor, β

Ref.

Nanographite in a-C

(100 oC)

200 μm etched W

needle 50 μm 160 This work

Nanographite in a-C

(400 oC)

200 μm etched W


Nanographite in a-C

(700 oC)

200 μm etched W


Oriented short sp2

planes dispersed in

amorphous matrix

2 mm dia. stainless

steel ball 20-40 μm 215-680 Shpilman et al.

26

N-containing DLC 4 mm dia. stainless

steel ball - 171 Gröning et al.

34

Ag Doped DLC ITO Parallel Plate 200 μm 2081–5876 Ahmed et al.10

Multi-wall CNT film 3 mm dia. stainless

steel cylinder 125 μm 1400 Bonard et al.

32

Single multi-wall

CNT

3 mm dia.stainless

steel cylinder 1 mm 30, 000 Bonard et al.

32


67

bonded diamond has negative electron affinity. Both of which are properties that can

facilitate field emission.

Conduction can occurs through 2 ways; i) directly through overlapping graphite

clusters or ii) by a hopping mechanism that occurs both between the nanographite

clusters and sp2 short chains in the a-C matrix. An increased amount of nanographite

clusters can increase the conductivity of the films as well as improve electron hopping,

which is the main conduction mechanism in a-C films.35

As such, the nanographite

clusters can form conductive channels through the matrix, transporting electrons to the

emission sites in vacuum. These conductive channels are estimated to be about 100–600

nm but physical features of such length have not been observed directly in literature. On

the surface, field lines will concentrate at the more conductive elements which are the

surface clusters. Dielectric differences between the sp2 and sp

3 elements would also mean

different local fields experienced, such that the emission can be seen to occur from freely

suspended conducting channels in vacuum (as the channels are encased in an insulating

matrix), mimicking the behavior of CNT. This emission enhancement mechanism for

diamond-like carbon and its related composites was first proposed by Gröning.36

This CNT or tube-like emission is further ascertained as we observed an

exponential increase in turn-on field when the anode-cathode distance was reduced below

10 um, while a saturated value was obtained for distances above 10 um as shown in Fig.

3.11. This corresponded to a variation of β as the anode-cathode distance changes in

order to preserve a constant local electric field (Eloc) at all distances. This is contrary to

previous beliefs that β does not change with distance and emission data for various

materials were compared across inconsistent distances.32

This decreasing Eappl trend with


68

distance has been observed only for CNT37

and iron tip emitters38

and the expected thin

film behavior line shape (dashed) is illustrated in inset 3.11(a). This proves that

nanoclustered graphite films may not be as “flat” in terms of field emission as we think it

might be. As such, a new understanding for field emission from graphitic planar films is

needed.

The shape of the plot in Fig. 3.11 can be explained by the empirical equation below,

VhD

Elocal

)(

1 (3.4)

where E is the field, V is the applied voltage, D is the distance between anode and

cathode and h is the height of the feature or conduction channel. A typical way of

defining distance is to start the measurement from the bottom of the substrate as

illustrated in inset 3.11(a) while it is in fact more appropriate to start the measurement

from the tip of the emitter feature. The curve obtained can be divided into 2 regions,

namely the near-anode region and far anode region. In the near anode region, the term

1/(D-h) dominates the equation and small changes in D can greatly affect the shape of the

plot, giving a sharp slope for small distances. When D is very much larger than h in the

far anode regions, the distance related term approximates to 1/D regardless of the changes

in h. This approximation agrees well with simulation results from CNT.37

It was also observed that for distances below 10 µm (Inset 3.11(b)), turn-on fields

for the 700 oC sample was larger than the 100

oC sample while the trend was reversed at

distances greater than 10 µm. This can be explained by considering the feature height and

dominating factors in the 2 regions. First, given that h700>h100, this would imply that 1/(D-


69

h700) >1/(D-h100) for the near anode region and this would result in a higher Eon for 700 oC

sample. Likewise, at the far anode region, feature height is no longer a dominating factor

and distance terms for both cases are 1/D. Thus, in this region, β dominates the equation

instead and given that the β700 is twice that of β100 as determined earlier, Eappl-700 is

smaller than Eappl-100 in accordance to equation 1.5.

3.5 Summary

In summary, in-situ heating during deposition led to the formation of

nanographite clusters within an a-C matrix. Sp2 content and aromatic ring formation

within the films increased with temperature. With the formation of nanostructures within

the film and sp2/sp

3 inhomogenities in the microstructure, field emission was enhanced.

Internal enhancement factor ratios were both determined with F-N relationship and WBK

approximation showing that the enhancement was doubled with increasing temperatures.

Through analysis of the turn-on field at various anode-cathode distances of field emission,

a CNT-like emission mechanism was observed for the nanographite clusters embedded a-

C films.


70

References

1. O. M. Küttel, O. Gröning, C. Emmenegger, L. Nilsson, E. Maillard, L. Diederich

and L. Schlapbach, Carbon 37, 745-752 (1999).

2. A. N. Obraztsov, A. P. Volkov and I. Pavlovsky, Diam. Relat. Mater. 9, 1190-

1195 (2000).

3. Y. Saito and S. Uemura, Carbon 38, 169-182 (2000).

4. S. Berber and D. Tománek, Phys. Rev. B 77, 073407 (2008).

5. G. A. J. Amaratunga and S. R. P. Silva, Appl. Phys. Lett. 68, 2529 (1996).

6. J. Robertson and M. Rutter, Diam. Relat. Mater. 7, 620-625 (1998).

7. J. Carey, R. Forrest and S. Silva, Appl. Phys. Lett. 78, 2339 (2001).

8. S. Kundoo, P. Saha and K. K. Chattopadhyay, Mater. Lett. 58, 3920 (2004).

9. Y. Umehara, S. Murai, Y. Koide and M. Murakami, Diam. Relat. Mater. 11, 1429

(2002).

10. S. F. Ahmed, M.-W. Moon and K.-R. Lee, Appl. Phys. Lett. 92, 193502 (2008).

11. A. T. T. Koh, J. Hsieh and D. H. C. Chua, Diam. Relat. Mater. 19, 637-642.

12. R. Narula, R. Panknin and S. Reich, Phys. Rev. B 82, 045418 (2010).

13. W. T. Diamond, J. Vac. Sci. & Technol. A 16, 707-719 (1998).

14. A. C. Ferrari, Diam. Relat. Mater. 11, 1053-1061 (2002).

15. F. Atchison, T. Brys, M. Daum, P. Fierlinger, A. Foelske, M. Gupta, R. Henneck,

S. Heule, M. Kasprzak, K. Kirch, R. Kötz, M. Kuzniak, T. Lippert, C. F. Meyer, F.

Nolting, A. Pichlmaier, D. Schneider, B. Schultrich, P. Siemroth and U.

Straumann, Diam. Relat. Mater. 16, 334-341 (2007).


71

16. S. Xu, B. K. Tay, H. S. Tan, L. Zhong, Y. Q. Tu, S. R. P. Silva and W. I. Milne, J.

Appl. Phys. 79, 7234 (1996).

17. C. Casiraghi, A. C. Ferrari and J. Robertson, Phys. Rev. B 72, 085401 (2005).

18. K. Wu, W. Chen, J. Hwang, H. Wei, C. Kou, C. Y. Lee, Y. L. Liu and H. Huang,

Appl. Phys. A. Mater. 95, 707-712 (2009).

19. A. Ilie, A. Ferrari, T. Yagi, S. Rodil, J. Robertson, E. Barborini and P. Milani, J.

Appl. Phys. 90, 2024 (2001).

20. E. Cappelli, S. Orlando, V. Morandi, M. Servidori and C. Scilletta, J. Phys.: Conf.

Ser. 59, 616 (2007).

21. Y. Lifshitz, Diam. Relat. Mater. 5, 388 (1996).

22. J. Robertson, Mater. Sci. Eng. R 37, 129 (2002).

23. B. K. Tay, X. Shi, H. S. Tan and D. H. C. Chua, Surf. Interface Anal. 28, 231

(1999).

24. S. Anders, J. Diaz, J. W. Ager, R. Y. Lo and D. B. Bogy, Appl. Phys. Lett. 71,

3367 (1997).

25. N. A. Marks, M. F. Cover and C. Kocer, Appl. Phys. Lett. 89, 131924 (2006).

26. Z. Shpilman, B. Philosoph, R. Kalish, S. Michaelson and A. Hoffman, Appl. Phys.

Lett. 89, 252114 (2006).

27. A. Ilie, A. C. Ferrari, T. Yagi and J. Robertson, Appl. Phys. Lett. 76, 2627 (2000).

28. M. Chhowalla, K. B. K. Teo, C. Ducati, N. L. Rupesinghe, G. A. J. Amaratunga,

A. C. Ferrari, D. Roy, J. Robertson and W. I. Milne, J. Appl. Phys. 90, 5308-5317

(2001).


72

29. D. N. Hutchison, N. B. Morrill, Q. Aten, B. W. Turner, B. D. Jensen, L. L.

Howell, R. R. Vanfleet and R. C. Davis, J. Microelectromech. S. 19, 75-82 (2010).

30. Y. Kudo, Y. Sato, T. Masuzawa, T. Yamada, I. Saito, T. Yoshino, W. J. Chun, S.

Yamasaki and K. Okano, J.Vac. Sci. Technol. B 28, 506-510 (2010).

31. Y. Kudo, T. Yamada, H. Yamaguchi, T. Masuzawa, I. Saito, S. Shikata, C. E.

Nebel and K. Okano, Jpn. J. Appl. Phys. 47, 8921 (2008).

32. J. Bonard, M. Croci, I. Arfaoui, O. Noury, D. Sarangi and A. Chātelain, Diam.

Relat. Mater. 11, 763-768 (2002).

33. S. Silva and J. Carey, Diam. Relat. Mater. 12, 151-158 (2003).

34. O. Groning, O. M. Kuttel, P. Groning and L. Schlapbach, App.Phys. Lett. 71,

2253 (1997).

35. J. D. Carey and S. R. P. Silva, Phys. Rev. B 70, 235417 (2004).

36. O. Gröning, O. M. Küttel, P. Gröning and L. Schlapbach, Appl. Surf. Sci. 111,

135-139 (1997).

37. R. C. Smith and S. R. P. Silva, J. Appl. Phys. 106, 014314 (2009).

38. R. Smith, R. Forrest, J. Carey, W. Hsu and S. Silva, Appl. Phys. Lett. 87, 013111

(2005).

Chapter 4 Pulsed laser fabrication of few-layer graphene using metal substrates

73

Chapter 4

Pulsed laser fabrication of few-layer

graphene using metal substrates

In this chapter, pulsed laser deposition (PLD) was used to fabricate few-layer

graphene (FLG). The effects of cooling rate and laser energy on graphene growth on Ni

metal substrates were studied in section 4.2. In section 4.3, the feasibility of Ni, Co, Fe

and Cu metal as substrates for graphene growth with PLD was explored. The use of

different laser wavelength PLD systems was studied in relation to graphene growth in

section 4.4. Lastly, FLG field emitters were fabricated and studied in section 4.5.

4.1 Introduction

In recent years, increased attention has been directed towards exploring

alternative methods to fabricate graphene, in particular, graphene through the deposition

of carbon on metals. Metals like ruthenium1, platinum

2 and nickel

3, 4 have been used in

the synthesis of graphene. Using metals to catalyst the growth of graphene or growing

adsorb graphene on metal surfaces is the third most used method of graphene synthesis

after mechanical exfoliation and epitaxial growth on SiC. High temperature requirements

(1000–1600 oC), ultra-high vacuum condition (10

-9 Torr), as well as the difficulties in

removing graphene from chemically stable SiC often complicate the fabrication process

and escalate costs.5 For graphene on metal substrate, free-standing graphene can be


74

obtained by etching with acid and transferring the graphene onto another substrate of

choice and can thus be a more effective method of fabrication. The range of graphene

thicknesses obtained from graphene grown on metals is also smaller as compared to

methods such as mechanical exfoliation.5 Hence, growing graphene on metals presents

several advantages including reduced temperatures, less stringent vacuum conditions and

ease of transferring graphene onto other arbitrary substrates by dissolving the metal

layer.6

Currently there are two main techniques that make use of metal substrates for

graphene growth; they are (i) chemical vapor deposition (CVD) and (ii) segregation.

CVD graphene has been proven to be a successful technique for growing large area

graphene (~cm2) by using Ni

3, 4 and Cu

7 as the main catalysts and hydrocarbon gas

sources. For instance, Reina et al.4 reported the fabrication of 1–12 graphene layers over

large area on Ni substrates using atmospheric pressure CVD. Li et al.7 also fabricated

large area graphene on Cu using CVD with methane and hydrogen gas source. In addition,

it was observed that the thickness of graphene layers did not increase with the thickness

of Cu as the growth process was self-limiting.8 Although CVD graphene is very

successful on Cu metals, the self-limiting formation of monolayer graphene on Cu may

be a deterrent for certain electronic applications. This is because monolayer graphene

lacks a bandgap, which is required for on/off switching in devices.9 Conversely, FLG is a

2D semimetal, which has been used to fabricate devices such as field effect transistors

(FET). 10

Studies have also suggested that FLG may be better for graphene FET as it

gives better manufacturing reliability.11


75

For segregation, the carbon source can be in the form of a hydrocarbon gas or in

the form of a solid-state carbon target, with the former being an unlimited source of

carbon and the latter being a limited source. Generally, this method makes use of the

carbon solubility of metal where carbon is absorbed by the metal at a higher temperature.

When the temperature is reduced, the carbon solubility is exceeded and precipitation of

the excess carbon occurs at the surface of the metal. However, it should be noted that

carbon uptake by the metals occurs at temperatures above 600 oC, thus this is the

minimum temperature at which such methods can succeed.12

Single or few-layer growth

can be achieved by controlling the cooling profile during segregation. Previously, Yu et

al.13

have fabricated graphene on Ni by segregation using a CH4: H2: Ar gas source at

1000 ºC.

Methods that make use of gaseous carbon sources whether in CVD or segregation

techniques, require the Ni-catalyzed decomposition of the hydrocarbon gas source and

subsequent diffusion into the metal, thus needing a longer time (>10 times) for the carbon

dissolution process. Here, the uptake of carbon by the metal can only be increased by an

increase in temperature. In this chapter, FLG was fabricated using pulsed laser ablation of

solid carbon targets at a relatively lower temperature of 750 oC, whereby the carbon

source was limited to that supplied within 1.5 minutes of target ablation (~7 nm of

carbon). PLD has been a popular method for fabricating a-C materials such as DLC and

its related nanostructured composites.14

This is because PLD provides carbon species that

possess energy up to a few hundred eV15

and particularly the energy of carbon species for

short wavelength lasers fall within the favored energy range for carbon sub-

implantation,16

a mechanism required for DLC formation.


76

From the point of view of segregation, the charged energetic carbon species

generated in the PLD plume has an added advantage of reduced carbon uptake duration.

The time taken for the absorption of carbon atoms into the metal layer is reduced for PLD

deposited carbon as solid solution formation of carbon in metal is facilitated by the

charged carbon ions that are able to penetrate the surface i.e. implantation of carbon ions.

Neutral carbon atoms on the other hand are unable to penetrate the surface and will be

deposited on the surface and absorbed by diffusion thereafter. In the case of solid-state

carbon segregation, the metal substrate was solely used as a medium to absorb incoming

carbon, instead of a catalyst to break down hydrocarbon gases.

Graphene was first fabricated using PLD in 2010 by Zhang et al17

with deposition

temperatures ranging from 1000 to 1300 oC and subsequently a study on FLG at reduced

temperature using PLD was carried out by Koh et al.18

While extensive research has been

carried out with CVD or segregation with gaseous carbon sources, extremely limited

information and studies are available on the parameters that affect segregation using

energetic carbon sources (e.g. from PLD) and solid-state metal-carbon diffusion methods.

Factors such as temperature, cooling rate, amount of carbon and laser energy (when using

a pulsed laser system) has to be controlled well in accordance with the type of method

used to achieve the desired results.

In section 4.2, changes in the quality and thickness of graphene formed on Ni

substrates using different cooling rates as well as different energies of the pulsed laser

beam were studied. Subsequently, in section 4.3, the viability of various metals to form

FLG from a solid carbon source provided by PLD was investigated. The metals chosen as

substrates were Ni, Cu, Co and Fe and the lattice constants of graphene, Ni, Cu, Co and


77

Fe are 0.357, 0.352, 0.361, 0.251 and 0.287 nm respectively, with Ni and Cu having the

smallest lattice mismatches. Ni has been largely successful in fabricating graphene in

several techniques ranging from common methods such as CVD and segregation, to less

popular methods such as graphene from metal melts.19

It was also interesting to find out

if the self-limiting behavior of Cu can be extended to other forms of carbon sources

besides gaseous sources used in CVD. Co and Fe are common catalysts for CNT

fabrication and Co has the highest solubility for carbon among the metals tested.

Although epitaxial growth of graphene on single crystal Pt,2 Ir and Ru

1 is well known, it

was neither included nor the focus of this study as the use of such substrates would

increase fabrication costs tremendously.

In section 4.4, the effects of using different laser sources were investigated. The

two lasers compared were a shorter wavelength excimer KrF laser (248 nm) and a longer

wavelength Nd:YAG laser (266 nm). Different laser wavelengths are known to produce

carbon films with different amounts of sp2 and sp

3 bonding and thus a different

microstructure entirely. Using experimental results and theoretical calculations, the

microstructure differences of segregated graphene using these two different lasers, as

well as the reasons behind the observed phenomenon were studied.

It has been established in earlier chapters that graphene possesses the ideal

combination of electronic, mechanical and optical properties as robust materials for

electronic applications such as field emission or even transparent emitters. Materials used

as cold cathode emitters are required to exhibit field enhancing effects whether internal or

external and the absence of such features can lead to extremely high turn-on fields (Eon)

or even no possible electron emission at all. The high aspect ratio (area to thickness ratio)


78

of graphene could potentially give dramatic field enhancing abilities when combined with

its good electron transport within the graphene sheet. However, many fabrication

methods produce graphene sheets that are morphologically flat or planar as graphene

growth proceeds laterally unlike in the case of CNT where 100 % vertical alignment can

be easily obtainable. Planar surfaces with no field enhancement can require turn-on

electric fields of 500–1000 V/µm20, 21

and such high electric fields are undesirable as it

may cause vacuum breakdown or electrical discharge.22

Thus, numerous research efforts

have been put into achieving vertically aligned graphene through mainly three methods; 1)

control the CVD growth process to get vertical petal-like FLG nanowalls,23-25

2) transfer

planar graphene sheets grown on metal substrates and process them thereafter to obtain

the desired alignment on a different substrate24, 26

and 3) use chemical processes to obtain

large quantities of free-standing exfoliated graphene from graphite and deposit graphene

films with solution processing methods such as filtration,27 spin coating28

or screen

printing.29

However, graphene nanowalls (GNW) by CVD in method 1 leads to a high

density of closely packed GNW that can produce a certain degree of electron screening

effect, give a higher Eon, depending on the control of density during fabrication. Method

2 which uses polymer support layers to transfer planar graphene sheets grown on metals

onto other substrates, can lead to extra defects generated during the transfer and also

incomplete transfer leads to non-uniform emitter surfaces.7 The third method will be

covered in Chapter 5 and thus not discussed in detail in this chapter.

Besides aligning graphene, a pre-shaped substrate, in the form of sharp tip

structures can also be used to help graphene acquire the required protrusions for field

emission. For instance, Ye et al.30

has reported on the fabrication of an emitter made


79

from graphene oxide sheets draped onto nickel nanotip arrays such that the flat GO sheets

acquired sharp protrusions throughout and thereby effectively reducing the Eon. However,

the conductivity of GO is significantly lesser as compared to graphene and it is possible

that the authors used GO instead of graphene to demonstrate a large area emitter (35 mm2)

as it is relatively easier to obtain large pieces of free standing GO.

In order to preserve the conductivity of as-grown graphene, it is best to directly

align graphene during growth and thus avoid transfer methods altogether. This can be

achieved by growing graphene onto tip or cone-shaped substrates. However, due to the

high temperatures used for some methods (>1000 oC) to fabricate graphene, metal cones

will undergo significant diffusion and reflow during the heating process and lose their

shapes as a result. The method of decomposition of SiC to form graphene is also not

suitable as it is very difficult to fabricate SiC cones due to its extremely high hardness.31

With the recent knowledge and understanding gained in the area of solid-state

transformation of carbon to graphene using a PLD deposited carbon source at

comparatively reduced temperatures, fabrication of FLG covered Ni and Co metal

nanocones structures that can be used for field emission applications were demonstrated

in section 4.5. The FLG was grown directly on metal cones and this eliminated problems

encountered during the transfer of graphene using support layers that in the process may

result in additional graphene breakage, low surface coverage fraction and incomplete

transfer that can affect the electron transport properties of graphene. Field emission

studies were also carried out on the fabricated FLG on the metal nanocones and their

field enhancing mechanisms were studied.


80

4.2 Few-layer graphene from Nickel substrates

4.2.1 Experimental method

Ni films of about 600 nm thickness were sputtered onto highly doped (n++

) Si

substrates. Subsequently, the Ni thin film substrates were transferred to the KrF (λ=

248nm) Lambda Physik excimer PLD system, which was evacuated to 5 x 10-6

Torr. A

carbon target (99.9% purity) was ablated with a laser energy of 50 mJ, a pulse duration

of 25 ns and a frequency of 10 Hz. The laser spot size was 1 mm2 and the target was

rotated at 6 rpm with the laser ablating a circular outline of 2 cm in radius. Substrate

temperature was kept at 750 oC during ablation. Subsequently, the Ni substrates were

cooled with three different controlled cooling processes 1, 50 and 100 oC/min. In

comparing the effects of laser energy, the cooling rate used was that of 1 oC/min and

energies 50, 100, 200 mJ were used for the comparison.

A Reinshaw Raman spectroscopy with 514.2 nm green laser beam was used to

examine the microstructure of the fabricated carbon. The green laser at 50 mW was

focused onto the sample with an optical microscope at 50X magnification and the spot

size was 1 µm. Spectra were obtained in the range of 1000–3000 cm-1

. A JEOL JEM

2010F high resolution TEM (with FEG source) operated at 200 kV was used to obtain

bright field cross-section images of the sample so as to view the graphene layers on these

samples.


81

4.2.2 Cooling rate dependence of graphene growth

Visible Raman spectroscopy (514.5 nm) is a suitable technique to probe the

structure of sp2 rich carbon as visible excitation is highly sensitive towards sp

2 bonding

and its photons resonate with the π states.32

Figure 4.1 shows the Raman spectra of the

samples cooled at different rates. The spectra of graphene consist of 3 peaks at 1350 cm-1

,

1590 cm-1

and 2700 cm-1

. The D (disorder) peak at 1350 cm-1

depends on the breathing

mode of the six fold aromatic rings, while it is independent of the thickness of graphene

and is activated purely by disorder. The D peak is thus absent in perfectly crystalline

graphite. The G (graphitic) peak situated around 1590 cm-1

depends on the in-plane

stretching motion and occurs for all sp2 sites inclusive of sp

2 chains and rings. The peak

at 2700 cm-1

(at a wavenumber exactly double that of D peak) is the second order D peak,

Fig. 4.1 Raman spectra of samples cooled at different rates. The laser energy per pulse was 50 mJ.


82

which is generally referred to as the 2D peak, is present even in the absence of defect.

There is a significant difference in the shape of the 2D peak between graphite and

graphene. In graphite, the 2D peak is not symmetrical and has a shoulder peak at a

slightly lower wavenumber. However, for graphene, the 2D peak is symmetrical. The

non-symmetrical peak with a shoulder peak attached at lower wavenumber can be clearly

distinguished when the graphene layers increase to 5 and beyond.32

As such, looking at

the peaks in the Raman spectra, it can be deduced that there were less than 5 layers in the

FLG samples. The intensity of the Raman peaks also suggested that the graphene layers

were not single layers and a higher 2D to G peak ratio would mean fewer layers were

present. Monolayer graphene has I(2D)/I(G) ratio of about 2 and the intensity of G peak

is known to increase almost linearly as the number of layers increases.33

By this

approximation, the I(2D)/I(G) ratio for the 1 oC/min and 50

oC/min cooling rates were

about 0.30 and 0.25 respectively which translated to 3–4 layers, clearly showing features

of FLG.

The 3–4 layers of graphene of the 1 oC/min sample can also be directly observed

in the TEM image (Fig. 4.2(a)), in which the carbon lattice fringes on the surface of the

Ni films had a lattice spacing of 0.33 nm, confirming the presence of carbon. Figure 4.2(b)

shows the layered structure of the samples and the Ni layer was seen to react with Si to

form nickel silicide at the interface, reducing the Ni layer to ~ 400–480 nm. In using a

cooling rate of 50 oC/min, the D peak was observed to reduce such that the I(D)/I(G) ratio

was 0.75 compared to a 1.15 ratio seen in the 1 oC/min cooling rate sample. This

indicated that a lower amount of disorder or defects were present in the sample cooled

with 50 oC/min. The samples cooled at 50 and 100

oC/min were also largely delaminated


83

at the Ni film-Si substrate interface when cooling was completed (Fig. 4.3), while

samples cooled at 1 oC/min remained attached to the underlying Si substrate. The lattice

constants of the materials used are as follow; Ni = 0.352 nm, C = 0.356 nm and Si =

0.543 nm. Referring to the Raman spectra in Fig 4.1, stress and strain can affect the G

peak position.33-35

As a result of the compressive stress experienced by the layered

structure, a slightly blue shifted G peak is seen in the Raman spectra in Fig 4.1 for the

sample cooled at 1 oC/min initially while for the samples cooled at 50 and 100

oC/min,

stress was released through delamination.

The differences in the Raman spectrum can be explained in terms of the

deposition process by pulsed laser and cooling process thereafter. Firstly, when the laser

ablates the target, carbon atoms with a certain amount of energy are deposited onto the Ni

substrate. These carbon atoms can either be 1) deposited onto the surface of Ni before

diffusing into the Ni thin film due to the heat present or 2) be implanted slightly into

near-surface regions of the Ni film. As such, the majority of the carbon will be

concentrated at near-surface regions, while the rest of nickel remained largely carbon free.

When cooled at 50 oC/min, the solubility of carbon in Ni dropped rapidly and the amount

of carbon atoms that exceeded the solubility limit are precipitated onto the Ni surface. As

the temperature decreased, carbon atoms continue to precipitate out and heat was

maintained such that the carbon atoms had sufficient mobility to rearrange themselves to

form a more ordered crystalline structure. This less defective ordering was translated to a

lower I(D)/I (G) ratio in the Raman spectra.


84

Fig. 4.2 Cross-section TEM showing (a) the graphene layers above the Ni. (b) Ni/Nickel

Silicide/Si layered structure. Inset (c) is an overview optical image showing rather

uniform coverage.

Fig. 4.3 Cross-section SEM showing delamination at Ni film-Si substrate interface for

medium and fast cooling samples.

Si substrate

Ni film (with graphene)


85

For samples that were cooled at 1 oC/min, the temperature (750

oC) was almost

maintained right after laser ablation was stopped and close to equilibrium conditions were

experienced. As such, large amounts of carbon atoms remained in the Ni film and had

sufficient time to diffuse from near-surface regions deeper into the bulk of the Ni film

where a larger amount of carbon can be dissolved in Ni without reaching its solubility

limit. As such, a minimal amount of carbon segregated on the surface as the temperature

was decreased slowly. Owing to the competition between segregation and diffusion

deeper into the bulk Ni, the larger amount of un-precipitated carbon that remained in the

Ni film after slow cooling (1 oC/min) resulted in a higher D peak (Fig. 4.1) as compared

to the sample with medium cooling (50 oC/min), which had less carbon remaining in the

Ni film after segregation. No 2D peak was detected for sample cooled at 100 oC/min as

the large amount of carbon that was precipitated onto the surface lacked the mobility to

form crystalline carbon within a much shorter cooling time and instead a more disordered

form of carbon was formed.

4.2.3 Laser energy effects on graphene growth

It is well understood that the laser energy affects the penetration depth of atoms

onto the substrate and this affects the carbon precipitation process.36, 37

The Raman

spectra shown in Fig. 4.4 are samples fabricated with laser energies, 50, 100 and 200 mJ.

It can be seen that using 50 and 100 mJ yielded similar results, while using 200 mJ

produced no 2D peaks. In the pulsed laser process, a higher laser energy would give rise

to more energetic carbon atoms and subsequently lead to a deeper implantation of carbon

atoms into the Ni film bulk which has not exceeded the solubility limit. While 100 mJ


86

and 50 mJ ablations would result in a shallower implantation in regions nearer to the

surface, making the precipitation process easier for the amount of carbon that was

supplied.

Fig. 4.4 Raman spectra of samples in which different pulsed laser energies were used. Cooling

rate was kept constant.


87

4.3 Few-layer graphene segregation using other metal

substrates


Metal films of approximately 500–600 nm thickness were sputter-deposited onto

highly doped (n++

) Si substrates. The thin metal film substrates were subsequently

transferred to the PLD chamber. The laser energy used in this section was 50 mJ. The

other laser parameters and chamber conditions used remained the same as those detailed

in section 4.2.1. The metal substrates were then cooled under controlled cooling at an

initial rate of 1 oC/min until the temperature was 550

oC, followed by a faster cooling rate

of 20 oC/min to room temperature.

4.3.2 Segregation of graphene on different metal substrates

Figure 4.5 shows the Raman spectra of the various metal substrates that were used

during deposition. As observed, there were two peaks present at ~1350 and ~1590 cm-1

across all four metals. In the case of Ni, there was an additional peak at ~2700 cm-1

.

These peaks at 1350, 1590 and 2700 cm-1

corresponded to the disorder (D), graphitic (G)

and second-order disorder (2D) peak respectively.33

The G peak, which is a non-resonant

scattering process (does not involve real electronic states), occurs due to the in-plane

vibrations of the sp2 carbon whether in chain or ring form. This peak is caused by the

phonons near the Brillouin zone centre (otherwise known as Γ point) whose momenta are

small. The D and 2D peaks are due to double resonant scattering processes, which


88

involve two continuous electronic transitions between two real states with different

energies and momenta, and are caused by phonons near the K point of the Brillouin

zone.38

Its shape and intensity can be used to determine the presence of graphene or

graphite and the thickness of graphene respectively. Based on the Raman spectra, it

appeared that only Ni was successful in forming FLG while carbon on the other metal

substrates was more disordered even though there was some evidence of clustering.

The shape of the 2D peak was symmetrical and the absence of the shoulder peak

at lower wavenumber suggested that FLG was present.39

Based on the peak shape in our

Raman spectra, it can be seen that there were less than 5 layers in the FLG on Ni samples.

For a more detailed estimation from the I(2D)/I(G) ratio of ~0.30, the graphene on Ni

corresponded to 3–4 layers.7

TEM images of the graphene on Ni when the cooling conditions are (a) suitable and

(b) too fast are shown in Fig. 4.6(a) and (b) respectively. In Fig. 4.6(a), few-layer

graphene which was imaged as lattice fringes was seen on top of Ni and it agreed well

with the Raman results of 3–4 layers. Fringes on the surface of the Ni films had a lattice

spacing of 0.33 nm, thereby confirming the presence of carbon. On the other hand, Fig.

4.6(b) shows the carbon lattice fringes that appeared to be more disordered and were

oriented in various directions.


89

Fig. 4.5 Raman spectra of carbon deposited on various metal substrates. Few-layer

graphene was formed only on Ni metal substrates when cooled with 1 oC/min to 550

oC followed

by 20 oC/min.

Fig. 4.6 Cross-section TEM showing (a) the graphene layers above Ni (b) a more disordered form

of graphite formed when cooling conditions were too fast.


90

4.3.3 Effect of cooling rate when using different metals

Figure 4.7 shows the Raman spectra of Ni and Cu substrate samples that were

subjected to different cooling profiles after the deposition of carbon. For the 50 oC/min

cooling rate it was seen that only Ni was observed to be successful in forming graphene

as shown by the 2D peak it exhibited, whereas the absence of the 2D peak indicated that

there was no graphene on Cu. For the 100 oC/min cooling rate, both Ni and Cu failed to

yield any positive results for graphene. This showed that cooling rate is critical in the

formation of graphene or polycrystalline graphite.

Fig. 4.7 Comparison of Raman spectra between Cu and Ni samples in which different

cooling rates 50 (top) and 100 oC/min (bottom) were used.


91

Fig. 4.8 Comparison of FLG grown on Co (top) and Ni (bottom). Full coverage was obtained on

Co substrate, while areas with graphene on Ni are indicated by the black arrows (inset).

Few-layer graphene was also formed on Co when a faster cooling profile

(20 oC/min) was adjusted to cater to Co. The Raman spectrum of FLG on Co is shown in

Fig. 4.8 where D, G and 2D peaks were present. This was also confirmed by TEM images

in Fig 4.9 which showed 2–3 layers of FLG. Under the optical microscope, it was

observed that the FLG grown on Co gave a higher surface coverage as compared to the

FLG on Ni sample. The as-grown FLG on Co was also transferred to Si substrate by

dissolving the Co metal in dilute HCl and it was observed that the D peak in the Raman

spectrum decreased significantly (Fig. 4.10). This showed that a large part of the disorder

contributing to the D peak was attributed to the amorphous carbon that was embedded

deep within the metal substrate during the pulsed laser ablation and that the segregated

FLG was relatively free of defects by itself.


92

Fig. 4.9 Cross-section TEM image of graphene on top of Co showing 2-3 layers.

Fig. 4.10 Raman spectra of (a) as grown few-layer graphene on Co (top). (b) Transferred few-

layer graphene on Si, which has a significantly reduced D peak indicating relatively few defects

in the graphene (bottom).


93

This phenomenon was also observed on areas where the Ni and Co metals were

covered during carbon deposition using PLD. With the use of stainless steel clips, certain

areas of the Ni and Co metals were blocked from the forward directed carbon plume. It

was observed that areas with carbon directly deposited onto, gave Raman spectra that

were a superposition of disordered carbon and graphene peaks, while areas that were

covered by the clips gave graphene with significantly reduced defects as seen from the

Raman spectra of the respective areas in Fig. 4.11. Graphene was able to segregate on the

covered area by diffusion of the carbon atoms nearby. This confirmed that most of the D

peak counts belonged to the underlying carbon dissolved in metal.

Fig. 4.11 Raman spectra of areas covered and uncovered during carbon deposition and

segregation.


94

4.3.4 Discussion

Graphene on metal substrates can be fabricated by two methods. The first method

is by segregating dissolved carbon in metal onto the surface under controlled temperature

annealing after carbon deposition (e.g. with PLD) and the second method is by

decomposing a precursor gas such as ethylene or methane on the metal surface, as in the

case of CVD. The former is a physical process that involves properties such as the

solubility of carbon in metal while the latter is a chemical reaction that makes use of the

metal layer as a catalyst to break down the gaseous carbon source. Another difference

between a CVD process and a PLD segregation process lies in the carbon source. In CVD,

the carbon source is in the form of a gaseous gas mixture, which way surpasses the

amount of carbon needed to form FLG. Conversely, PLD uses a solid carbon target,

which limits the carbon source during segregation to that supplied during target ablation.

Lastly, the carbon supplied by PLD is an energetic source of carbon ions, which can

penetrate deeper into the metal layer rather than stay on the metal film surface.

Across the four metals tested, graphene was formed only on top of Ni but not on

Cu, Co, Fe. To better understand this, the solubilities of carbon in these metals should be

examined. The carbon solubilities in Ni, Cu, Co and Fe at about 700 ºC are 1, 0.05, 1.9

and 0.3 at.% respectively.40

The relatively high solubility of carbon in Ni at 700 ºC

supported the formation of graphene on Ni. On the other hand, given the extremely low

carbon solubility in Cu, the precipitation process could not produce enough carbon to

give crystalline graphene as insufficient carbon was absorbed into Cu. Instead, the as-

deposited carbon remained largely on top of Cu and as heating proceeded and the initially


95

amorphous carbon on top of Cu began to exhibit signs of clustering as observed by

separate D and G peaks. This is in stark contrast to copper’s success when using CVD,

which gave self-limiting growth of single layer graphene. This is due to the non-chemical

nature of the segregation method, which does not rely on the catalytic ability of metals

and depends solely on its carbon solubility. As such, a metal with sufficient carbon

solubility would be desired over its catalytic ability when PLD deposition followed by

segregation is employed.

It is however interesting to note that there was no positive result for graphene on

Co in spite of its high carbon solubility when the same cooling recipe (1 oC/min followed

by 20 oC/min) for Ni was used. To understand this phenomenon, the rates of diffusion of

carbon into Co and Ni are considered. The rate of diffusion is characterized by a constant

known as the diffusion coefficient, D, which can be expressed as:

)exp(

RT

QAD (4.1)

where A is cm2s

-1, Q is in kJmol

-1, R= 8.31441 Jmol

-1K

-1.

The A and Q values for Ni are 0.12 and 137.3 respectively while the A and Q

values for Co are 0.31 and 153.7, respectively.41

As the activation energy, Q, is in the

exponent factor, it dominates the equation, making diffusion of carbon in Co slower than

that of Ni. The diffusion coefficients of carbon at 1000 ºC in Ni and Co are calculated to

be 2.79 × 10-11

and 1.53 × 10-11

m2s

-1, respectively.

42 This implies that Co required a

different cooling recipe from Ni due to differences in their physico-chemical properties

and this can be explained through the formation process. During the cooling process,


96

there is a competition between two processes; a) carbon atoms diffusing in towards the

metal substrate closer to the Si side that contains low concentration of carbon and b) the

precipitation process that results as the carbon solubility limit in metal is exceeded.

In the case of Co, which has a higher carbon solubility and lower diffusion

coefficient, when the cooling profile is slower, carbon atoms can make use of the longer

time to diffuse further into the metal film. As a result, it is more difficult for near surface

carbon atoms to exceed the solubility limit when a higher temperature is constantly held

for an extended period of time since carbon with low D can have more time to

redistribute. Conversely, a faster cooling rate (at 20 oC/min from 750

oC to room

temperature), would maintain a carbon saturated near-surface region to allow for

adequate precipitation, which led to the formation of FLG (Fig. 4.12). On the other hand,

for Ni, which has almost half the carbon solubility compared to Co, can easily exceed the

solubility limit when the temperature gradually decreases. From the distinct D, G and 2D

peaks observed in the Raman spectra for Co, it was suggested that the fabrication FLG on

Co was crystalline or nanocrystalline. As such, the Tuinstra-Koenig (T-K) relation43

can

be used to provide a size estimate (La) of the graphitic domains where La= C

()[I(D)/I(G)]-1

and C(514.2nm) is ~ 4.4nm. The La value for FLG on Co was calculated

to be 19.5 nm. However, it should be noted that the T-K relation has a tendency to

underestimate the domain size due to the dominant effect of small crystallites and the

equation assumes a uniform domain size throughout the sample.44

Therefore, larger

domains should be present in the fabricated FLG.


97

Fig. 4.12 Illustration of carbon atom concentration throughout the Co film as a result of slow and

fast cooling profiles. Carbon atoms are diffused deeper into the Co film and harder to precipitate

when cooling is slow. During fast cooling, solubility is exceeded in the near-surface regions and

precipitation results.


98

4.4 Few-layer graphene segregation using different

pulsed laser wavelengths


Nickel films of approximately 500–600 nm thickness were sputter-deposited onto

highly doped (n++

) Si substrates. The thin metal film substrates were subsequently

transferred to the pulsed laser systems which were evacuated to 5 x 10-4

Pa prior to

deposition. The pulsed laser systems used were 1) krypton fluoride (KrF, λ= 248nm)

Lambda Physik excimer PLD system with a pulsed duration of 25 ns and 2) neodymium-

doped yttrium aluminum garnet (Nd:YAG, λ= 266nm) with a pulse duration of 10–18 ns.

The frequency and laser energy was 10 Hz and 50 mJ for both lasers. A carbon target

(99.9% purity) was ablated with the respective lasers. Substrate temperature was kept at

750 oC during ablation. Deposition was carried out to yield an estimated thickness of 7

nm of carbon film if it had been deposited on silicon. The metal substrates were then

cooled under controlled cooling at an initial rate of 4 oC/min until the temperature was

550 oC, followed by a faster cooling rate of 20

oC/min to room temperature, unless stated

otherwise.

4.4.2 Segregation with Nd:YAG and KrF pulsed laser

The Raman spectra of graphene segregated using carbon from the Nd: YAG laser

seen in Fig. 4.13 showed 3 distinctive peaks at ~1347 (D peak) and ~1582 cm-1

(G peak)

and ~2687 cm-1

(2D peak) which are peaks characteristic of graphene layers. The peak at

2900 cm-1

is a combination peak of G and D peak. Similarly, the three characteristic


99

peaks were also observed in the Raman spectra (Fig. 4.14) obtained from the graphene

segregated from carbon from the KrF laser under similar conditions. The D, G and 2D

peak for KrF fabricated sample was at 1338, 1577, 2698 cm-1

respectively. The D peak,

indicates the presence of defects or numerous graphene edges. As the shape and position

of the 2D peak depends on the electronic properties of the carbon structure, the

symmetrically shaped 2D peak (with no shoulder peak) suggested that the peak belonged

to graphene layers and not bulk graphite.

The position of the 2D peak was used to determine the thickness of the graphene

layers. The peak wavenumber is also known to increase with the number of layers.45

As

the Raman spectra of KrF showed a 2D peak at 2698 cm-1

as compared to 2689 cm-1

in

the Nd:YAG sample, the Nd:YAG sample can thus be deduced to consist of relatively

thinner graphene layers.46

The G peak, which is also known as the graphitic peak has

intensity that increases with the number of graphene layers while peak position

wavenumber decreases with thickness.45

The position of G peak in the KrF sample was

1576 cm-1

as compared to the 1582 cm-1

peak in Nd:YAG sample, which pointed to a

thinner graphene stack in the Nd:YAG fabricated graphene once again. Furthermore,

through the relative peak intensities of the 2D and G peak, the thickness of the samples

can be further verified. As a reference, the I(2D)/I(G) ratio of a single layer graphene

sample is 2.33, 47

From Fig.4.13, I(2D)/I(G) ratio is 1.57, which translated to ~2 layers of

graphene for the Nd:YAG fabricated sample. In addition, the I(2D)/I(G) ratio is 0.30 in

Fig. 4.14, which gave ~4 layers of graphene for the KrF fabricated sample. Raman

spectra in Fig 4.15 obtained from various spots on the sample deposited with Nd:YAG

laser showed that the graphene thickness at these spots were rather uniform within a


100

Fig. 4.13 Raman spectra of graphene on Ni segregated from carbon deposited by Nd:YAG laser

at 750 oC.

Fig. 4.14 Raman spectra of graphene on Ni segregated from carbon deposited by KrF laser at

750 oC.


101

range from 2–3 layers. However, it should be noted that not 100% coverage was observed

through optical microscope images as seen in Fig 4.15.

Additionally, segregation was also carried out using Nd:YAG laser at a different

cooling rate of 20 oC/min and it was observed that few-layer graphene was also formed at

this cooling rate as shown in Fig. 4.16. However, the resultant graphene was thicker as

seen from the lower I(2D)/I(G) ratio of 0.78. The faster cooling rate (20 oC/min) allowed

for more graphene layers to precipitate during cooling.


at 5 different spots. Optical images (right) show the spot that was scanned.


102


when cooling rate was adjusted to a faster rate of 20 oC/min.

4.4.3 Discussion

It is known that the type of laser (i.e. wavelength) and laser fluence (J/cm2) will

affect the characteristics of the ablated plume components such as energies of the ions

and type of ions and neutrals present within the plume.15

These will in turn affect the

microstructure and resultant type of film obtained.48

As such, it is imperative to

understand the type of lasers used in this experiment. On one hand, Nd:YAG is a solid-

state laser that makes use of a crystal as the laser source. The wavelength of such lasers

naturally occurs at 1064 nm. However, through the use of Q-switching, output powers

can be increased and pulsed duration reduced to the nanosecond range. The laser

frequency can thus be doubled, tripled or quadrupled to give 532, 355 and 266 nm

wavelengths respectively, which are higher harmonics of laser radiation.49

On the other


103

hand, KrF is a gas based laser that makes use of the reaction between krypton and

fluorine gases to emit a laser at wavelength 248 nm, which is in the near ultraviolet range.

The process of plasma formation from a target when a pulsed laser ablates the

target can be divided into three main steps. In the first step, the incident laser radiation is

absorbed and interacts with the surface atoms of the target. This electronic absorption is

followed by extremely rapid heating of the target surface and leads to the formation of a

plume that consists of positive carbon ions, electrons and neutral particles. When the

laser fluence used is high enough for the surface temperature to reach beyond the critical

temperature instantly (as is the case of this study), thermal evaporation by phase

explosion occurs. The density of excited atoms increases with the laser fluence to a point

where the target surface is no longer a bound solid but should be regarded as a dense gas

in which strong repulsive forces of the like-charged ions are present. This repulsive force

in combination with the evaporation of material leads to the expulsion of energetic

particles.50

A shorter wavelength laser, thus a higher photon energy, will have a lower

optical penetration depth and result in a smaller ablated area and hence give particles with

higher energy. For carbon target which has a high melting point and thermal conductivity,

droplet emission is reduced.15

In the second step, the incident laser beam interacts with the expelled target

material leading to the formation and expansion of an isothermal plume. As the expulsion

of target material through the initial interaction between incident laser and target occurs

in the range of picoseconds and the pulse duration of the laser is in the range of

nanoseconds, the plume will be irradiated by the incident laser beam unless a pulse

duration smaller than the picosecond range is used.51

Hence, this laser interaction with


104

the expelled plume is a significant process in the ablation process. At this stage,

absorption of the laser by the plume leads to the attenuation of the laser intensity incident

on the target and also leads to additional excitation, fragmentation and ionization of

plume species.

Lastly, in the third step, the plasma expands adiabatically in an anisotropic three

dimensional manner with an overall forward direction that is characteristic of pulsed laser

ablation.50, 52

The first two steps occur during the laser beam irradiation period (i.e. pulse

duration), while the third step occurs after the laser pulse irradiation ceases. The

microstructure of the resultant film depends on temperature of the substrate and energies

of the plume species which is dependent on the wavelength, pulse duration, intensity and

energy of the laser source.

In laser plume studies of Nd:YAG ablated carbon, it was observed using time-of-

flight mass spectroscopy data that the size of carbon clusters decreases with increasing

laser power for 1024 and 532 nm. But in the 266 nm laser ablated plume, carbon ion

clusters were below C7+ even when the laser power density was at the lowest. This

showed that photoelectronic excitations played an important role in shorter wavelength

ablation as large carbon clusters (Cn+ where n≥7) are dissociated photoelectronically by

absorbing the 266 nm radiation. The dominant carbon ion species in the carbon plume

ablated by 266 nm wavelength is C+

at a fluence of ≥ 5 Jcm-2

together with some C3+ and

small amounts of C5+.49, 53

In the laser plume studies of KrF laser, it was observed that C+,

C2+ and C3

+ ion species were present.

48, 53, 54


105

It has been reported that the ablation rate and energetics of the plume are affected

by the absorption of laser fluence by the target, which is dependent on the optical

properties of laser-target interactions, rather than the thermal properties of the target

material. As such, the energy absorbed by atoms from the incident laser pulse was

calculated for each laser system by dividing the laser fluence by the number of neutrals,

nn. As mentioned earlier, the laser pulse will interact with the expelled plume as well,

since the pulse duration is greater than the plume expansion duration. The absorption of

the laser by the plume subsequently occurs by an inverse bremsstrahlung process and its

absorption, αp can be estimated by,

kT

h

T

nZ i

p

exp11069.3

32/1

23

8 (4.2)

where Z, ni, T, h, k and υ are the average ion charge, ion density and temperature of the

plasma, Planck constant, Boltzmann constant and frequency of the laser light,

respectively.52, 55

The ion density, ni, in cm-3

can be approximated by using Saha

equation55

, which can be expressed as,

2/1

2/315 exp104.2

kT

UnTn i

ni (4.3)

where nn is the density of neutrals in cm-3

and Ui is the first ionization potential of carbon

atom which is 11.26 eV.56

The temperature was taken to be 3600 K, which is the

sublimation temperature of carbon. As αp is directly proportional to ni2, absorption mainly

occurs very near to the target surface where ni is high.


106

The laser energy absorbed by electronic absorption, Q, can be estimated by the

product of αp and fluence of the laser I, in J/cm-3

(equation 4.4). From these calculations,

the energy absorbed per carbon target atom due to the incident laser (first step of

ablation), E1, and the additional energy absorbed due to laser-plume interactions (second

step of ablation), E2, can be estimated by,57

IQ p (4.4)

From the estimations of energies absorbed by a carbon target atom/ion by the

absorption and interaction with different laser wavelengths in Table 4.1, it can be seen

that the energy of the carbon species in the Nd:YAG laser were ~1.26 eV lesser as

compared to the KrF ablated carbon species. Despite the differences in wavelength and

energies of depositing carbon species, both types of laser were viable for graphene

segregation. The lower energy of carbon species from the Nd:YAG laser could also give

a higher sp2 near surface carbon concentration that favored the formation of more

graphitic rather than amorphous carbon bonds. Thus, it can be seen that the D and G

peaks of the Nd:YAG graphene was distinctly separated while the KrF graphene spectra

gave a superposition of graphene and amorphous carbon peaks with a more conjoint D

and G peaks. In addition, it is expected that similar results will be obtained with the

Nd:YAG laser as with the KrF system with Ni, Co, Cu and Fe substrates. That is, Ni and

Co will be suitable but Cu and Fe will not be due to their poor carbon solubilities.


107

Table 4.1 Carbon ion energies for step 1 and 2 during laser ablation of carbon target for KrF and

Nd:YAG systems.

Laser Type E1

(eV/ion)

E2

(eV/ion)

KrF 56.34 2.72 × 10-5

Nd:YAG 55.08 3.36 × 10-5


108

4.5 Fabrication of few-layer graphene field emitters

using PLD


Ni and Co metal nanocones were fabricated from sputter-deposited metal films

using ion irradiation technique by bombardment with Ar+ ions using a Kaufman type ion

gun (ION TECH. INC. Ltd., model 3-1500-100FC). The incidence angle was 45 oC to the

surface, and irradiation was performed at room temperature for 7 min. The diameter and

the energy of the ion beam employed were 6 cm and 1000 eV, respectively. The basal

and working pressures on the ion beam chamber were 10−5

and 5 × 10−2

Pa, respectively.

Two types of Ni and Co metal nanocones were prepared for each metal, one with distinct

taller cones with some small cones in-between and one with a high density of uniform

shorter cones. From here forth, the two types of cones will be referred to as Type A and

Type B cones respectively. The prepared metal nanocones were then transferred to the

KrF excimer PLD chamber and graphene segregation was carried out at cooling rates

suitable for each metal according to the results obtained earlier and with the same laser

parameters as detailed in section 4.2 and 4.3. The metal nanocones were examined before

and after the graphene segregation using a Zeiss Supra 40 FEG SEM.

4.5.2 Microstructural characterization of cones

Figure 4.17 shows the surface morphology of the as-fabricated Type A (Fig.

4.17(a, c)) and Type B (Fig. 4.17(b, d)) Ni and Co metal cones. It can be seen that for


109

both Ni and Co metals, high density cones arrays resulted from the ion irradiation and

Type A consisted of tall cones separated by smaller cones while Type B consisted of

small cone tips with uniform height. For Ni nanocones, the heights of the cones range

from 400–700 nm for Type A and 200–300 nm for Type B. For Co, the sizes of the

nanocones range from 400–840 nm for Type A and 200–300 nm for Type B. The apex

radius of the cones were ~40 nm for the cones in Type A and slightly smaller than that

for the small cones in Type B.

The effects of heating on the metal nanocones were studied by carrying out the

heating and cooling process in the chamber without the deposition of carbon. The

corresponding SEM images are shown in Fig. 4.18. It can be seen that the heating process

did not destroy the conical structures entirely and the conical shapes were largely

maintained. After graphene was segregated using PLD-deposited carbon, the morphology

of the surface was examined again. Figure 4.19 shows the respective SEM images of both

Type A and B Ni and Co cones after carbon deposition and graphene segregation.

Comparing the images to the heat-only Type A cones, it was observed that the energetic

species did a small amount of additional damage to the cone structures in addition to the

heat effect. Thus, this made it more critical for the laser energy to be kept to a minimum

to not cause significant sputtering damage on the cones. The Raman spectra of Ni and Co

cones after segregation are shown in Fig. 4.20. The three characteristic peaks, D, G and

2D peaks confirmed the presence of FLG on the metal cones.


110

Fig. 4.17 SEM images of original as-fabricated Ni cones (a) Type A, (b) Type B and Co cones (c)

Type A, (d) Type B.

Fig. 4.18 SEM images of heat-only (a) Ni (b) Co metal nanocones.


111

Fig. 4.19 SEM images of (a) Ni (b) Co metal nanocones after graphene segregation.

Fig. 4.20 Raman spectra for Ni and Co metal cones.


112

4.5.3 Field emission study of graphene coated metal cones

The field emission I-V plots and F-N plots are shown in Fig. 4.21 and 4.22 for the

various cone samples. The turn-on voltage (Von) for Type A cones is 70 V and 105 V for

Co and Ni respectively. Also, the Von for Type B cones is 470 V and 580 V for Co and Ni

respectively. From the slopes of the F-N plots for the samples, β was estimated to be 275,

65 for Co and Ni Type A cones, 8 and 4 for Co and Ni Type B cones respectively, when

the work function of the material was assumed to be 5 eV.

Fig. 4.21 Field emission I-V plots for graphene on cone samples.


113

Fig. 4.22 Field emission F-N plots for graphene on cone samples.

4.5.4 Discussion

Spindt tips made of metals such as molybdenum (Mo) were the first cold cathode

material that was used for field emission displays. Electrons are extracted from these

sharp tips under a high electric field and used to bombard phosphor screens. However,

they were susceptible to poisoning by certain elements (e.g. sulfur) that are given off the

sulfur-based phosphor screen and ion erosion by ion bombardment.58, 59

This causes

blunting of the tips and field emission performance is compromised. As a result, Spindt

tips are plagued with stability and reliability issues that led to the rise of carbon-based

field emitters. Given that graphene has excellent mechanical properties and chemically

stability, growing FLG directly onto the metal cones can help to provide a layer of

mechanical and chemical protection for the metal cones.


114

The as-prepared Ni and Co cones did not reach the electron emission turn-on

current within the maximum limits of the voltage. But after graphene was segregated onto

the cones, field emission at a considerably lower voltage took place. This was attributed

to a change of the tunneling barrier at the interface between the cone tip surface and

vacuum. From first principles studies of graphene on Ni and Co substrates by

Khomyakov et al.,60

it has been shown that the chemisorption of graphene onto these

metals resulted in an electron transfer process from metal to graphene which n-doped

graphene. The work function of free-standing graphene and clean metal surfaces as

compared to the adsorbed graphene on metal systems are shown in Table 4.2. It can be

seen that the work function of pure metal cones was significantly reduced by 1.66–

1.81 eV after the adsorption of graphene. As the emission barrier height is affected by

work function of a material, a reduced emission barrier height is resulted.61

The shift of

EF has been illustrated in Fig. 4.23 where a schematic of the band structure of the metal

cone with and without a graphene coating is shown. Similar effects have been cited in the

cases of gold (Au)-graphene and aluminum (Al)-graphene. A higher reduction potential

(relative to standard hydrogen electrode) of Au at 1.5 V compared to 0.14 V of graphene

resulted in electron transfer from graphene to Au, while a negative reduction potential of

Al at -1.66 V compared to graphene led to electron transfer from Al to graphene.62

The

reduction potentials of Ni, Co and graphene are -0.25, -0.28 and 0.14 V respectively.62, 63

Hence, charge transfer from Ni and Co metals to graphene occurred, raising the Fermi

level, EF, of graphene and led to the reduction of work function (barrier height) at the

surface of the coated metal tips.


115

Table 4.2 Comparison of work function of free-standing graphene and clean metal surfaces with

the respective graphene on Ni or Co metal systems from ref. 60.

Work Function (eV) Graphene Ni Co

Free-standing 4.48 5.47 5.44

Graphene on metal system 3.66 3.78

Difference 1.81 1.66

Fig. 4.23 Schematic of work function shift when graphene is chemisorbed onto metal ( G-on-M)

and the position of EF in intrinsic versus doped graphene.


116

From the SEM images, the Type B Ni cones lost most of their definitive cone tip

outlines as compared to the Type B Co cones, where the small nanocones could still be

identified after graphene segregation. This was attributed to the slower cooling profile

used for the Ni cones (as mentioned in section 4.3), thus allowing a larger degree of

diffusion and reflow of the Ni conical structures to take place. Despite that, field emission

still occurred as the surface of the Ni remained rough, consisting of nano protrusions as a

reduced β factor. This proved that the change in tunneling barrier at the emitter surface by

applying a graphene coating onto the metal cones, instead of the physical aspect ratio, is

the dominant factor in reducing the turn-on voltages for the samples.

Type A samples consisted of nanocones that were of a range of heights, with a

mix of some very tall cones in the midst of some shorter cones, while Type B samples

consisted of high density nanocones that were of a relatively uniform height. This gave

the former an advantage of an improved field enhancement as a result of the spaced-out

tall cones. On the other hand, the numerous cones of the same height resulted in a large

degree of electron screening effect as evidenced by the large drop in β and higher field

emission turn-on voltage for these samples. In addition, when comparing the electrical

conductivities of Co and Ni, the significantly higher conductivity of Co at 17.9×106 Sm

-1,

as compared to 14.6×106 Sm

-1 for Ni also helped to provide more efficient electron

transport from the substrate to the surface, allowing for a lower turn on voltage.64

In

addition, a comparison of the turn-on voltages for the graphene on metal nanocones

samples with the nanographite in amorphous carbon samples in Chapter 3 is shown in

Table 4.3.


117

Table 4.3 Comparison of turn-on voltages of the graphene on metal cones samples with the

nanographite in amorphous carbon samples in Chapter 3.

Sample Von

(V)

Co Type A 70

Ni Type A 105

Co Type B 470

Ni Type B 580

Nano-graphite in a-C 420-500

4.6 Summary

In section 4.2, FLG was fabricated on Ni substrates using the PLD technique. The

effects of cooling rate and laser energy of the graphene product were studied with Raman

spectroscopy. It was observed that the cooling rate was critical in the production of

graphene, graphite (>10 layers) or a slightly more disordered carbon. On the other hand,

laser energy was less critical as long as the energy fell below 100 mJ. With a precise

control of temperature and amount of source carbon, the formation of graphene layers can

be controlled. The results have shown that it is possible to produce graphene using solid

sources and a physical deposition technique at relatively lower temperatures compared to

graphene epitaxial growth using SiC, as long as the temperature control and laser energy

are optimized.

Few-layer graphene was fabricated on Ni and Co metal substrates using PLD as

the required carbon source in section 4.3. Although graphene growth on Cu has been very


118

successful with a CVD process involving chemical reactions with precursors, the failure

to produce graphene when using a solid carbon source with Cu was because the catalytic

behavior of the metal was not utilized for such a setup. Instead, the metal’s solubility was

critical to the successful fabrication of graphene by segregation. When using a cooling

rate of 1 oC/min followed by 20

oC/min, FLG was only formed on Ni and not Co. With an

increase in cooling rate, largely defect-free FLG with 100 % substrate coverage was also

formed on Co. As such, besides carbon solubility of the substrate, the cooling recipe used

for different metals was also extremely important owing to the competing processes

during segregation.

In section 4.4, FLG was fabricated by both Nd:YAG and KrF pulsed lasers at

750 oC. The thickness of the resultant graphene layers was estimated from the relative

positions of characteristic Raman peaks of graphene as well the the relative I(2D)/I(G)

ratios. The corresponding ratio of graphene in the Nd:YAG sample was 1.57 which

translates to ~2 layers of graphene, while the ratio for the KrF sample was 0.30, which

gave ~4 layers of graphene. Through theoretical calculations, the carbon species in the

Nd:YAG laser were of a lower energy than the KrF laser. Even though the two laser

sources had different wavelengths and carbon ion energies, both laser systems were

capable of producing graphene when the same cooling profile was used.

Lastly, in section 4.5, cones of various density and heights from Ni and Co metals

were fabricated and the segregation technique studied in sections 4.2 and 4.3 was

subsequently applied to grow graphene directly onto the metal cones. The fabricated

graphene on metal cones were studied for their field emission properties. It was observed

that the mix of tall cones separated by small cones in-between, otherwise known as


119

Type A cones, provided a better field emission enhancement as compared to the array of

highly dense cones with uniform heights (i.e. Type B cones). By applying a chemisorbed

coating of graphene onto metals such as Ni and Co, field emission properties were

enhanced greatly by the reduction of tunneling barrier of the metal cones and lower turn-

on voltages were required for electron emission as a result.


120

References

1. P. W. Sutter, J.-I. Flege and E. A. Sutter, Nature Mater. 7, 406-411 (2008).

2. H. Ueta, M. Saida, C. Nakai, Y. Yamada, M. Sasaki and S. Yamamoto, Surf. Sci. 560,

183-190 (2004).




Kong, Nano Lett. 9, 30-35 (2008).

5. J. Wintterlin and M.-L. Bocquet, Surf. Sci. 603, 1841-1852 (2009).

6. C. He, W. Wang, S. Deng, N. Xu, Z. Li, G. Chen and J. Peng, J. Phys. Chem. A 113,

7048-7053 (2009).

7. X. Li, W. Cai, J. An, S. Kim, J. Nah and D. Yang, Science 324, 1312 (2009).

8. C. Mattevi, H. Kim and M. Chhowalla, J. Mater. Chem. 21, 3324 (2011).

9. J. B. Oostinga, H. B. Heersche, X. Liu, A. F. Morpurgo and L. M. K. Vandersypen,

Nature Mater. 7, 151-157 (2008).

10. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I.

V. Grigorieva and A. A. Firsov, Science 306, 666-669 (2004).

11. Z. Jin, Y. Su, J. Chen, X. Liu and D. Wu, Appl. Phys. Lett. 95, 233110 (2009).

12. J. A. odr guez-Manzo, C. Pham-Huu and F. Banhart, ACS Nano 5, 1529-1534

(2011).

13. Q. Yu, J. Lian, S. Siriponglert, H. Li, Y. Chen and S. Pei, Appl. Phys. Lett. 93,

113103 (2008).

14. A. T. T. Koh, J. Hsieh and D. H. C. Chua, Diam. Relat. Mater. 19, 637-642 (2010).

15. A. A. Voevodin and M. S. Donley, Surf. Coat. Technol. 82, 199-213 (1996).


121

16. J. Robertson, Mater. Sci. Eng. R 37, 129-281 (2002).

17. H. Zhang and P. X. Feng, Carbon 48, 359-364 (2010).

18. A. T. T. Koh, Y. M. Foong and D. H. C. Chua, Appl. Phy. Lett. 97, 114102 (2010).

19. S. Amini, J. Garay, G. Liu, A. A. Balandin and R. Abbaschian, J. Appl. Phys. 108,

094321 (2010).

20. J. Carey, R. Forrest and S. Silva, Appl. Phys. Lett. 78, 2339 (2001).

21. S. Bellucci, A. Malesevic, F. Micciulla, I. Sacco, R. Kemps, A. Vanhulsel and C. Van

Haesendonck, Nanosci. Nanotechnol. Lett. 3, 907-912 (2011).

22. R. V. Latham, High Voltage Vacuum Insulation: Basic Concepts and Technological

Practice. (Academic Press, London, 1995).



24. P. Hojati-Talemi and G. P. Simon, Carbon 49, 2875-2877 (2011).

25. H.-C. Hsu, C.-H. Wang, S. K. Nataraj, H.-C. Huang, H.-Y. Du, S.-T. Chang, L.-C.

Chen and K.-H. Chen, Diam. Relat. Mater. 25, 176-179 (2012).

26. S. Pandey, P. Rai, S. Patole, F. Gunes, G.-D. Kwon, J.-B. Yoo, P. Nikolaev and S.

Arepalli, Appl. Phy. Lett. 100, 043104 (2012).

27. K. Wang, T. Feng, M. Qian, H. Ding, Y. Chen and Z. Sun, Appl. Surf. Sci. 257,

5808-5812 (2011).

28. E. Stratakis, G. Eda, H. Yamaguchi, E. Kymakis, C. Fotakis and M. Chhowalla,

Nanoscale 4, 3069-3074 (2012).

29. Q. Min, F. Tao, D. Hui, L. Lifeng, L. Haibo, C. Yiwei and S. Zhuo, Nanotechnology

20, 425702 (2009).


122


1265-1268 (2012).

31. I. Yonenaga, Phys. B Condens. Matter 308–310, 1150-1152 (2001).

32. A. C. Ferrari, Solid State Commun. 143, 47 (2007).

33. D. Graf, F. Molitor, K. Ensslin, C. Stampfer, A. Jungen, C. Hierold and L. Wirtz,

Nano Lett. 7, 238-242 (2007).

34. X. Duan, H. Son, B. Gao, J. Zhang, T. Wu, G. G. Samsonidze, M. S. Dresselhaus, Z.

Liu and J. Kong, Nano Lett. 7, 2116-2121 (2007).

35. D. S. Lee, C. Riedl, B. Krauss, K. von Klitzing, U. Starke and J. H. Smet, Nano Lett.

8, 4320-4325 (2008).

36. C. Uzan-Saguy, C. Cytermann, R. Brener, V. Richter, M. Shaanan and R. Kalish,

Appl. Phy. Lett. 67, 1194 (1995).

37. J. W. Ager, S. Anders, A. Anders, B. Wei, X. Y. Yao, I. G. Brown, C. S. Bhatia and

K. Komvopoulos, Diam. Relat. Mater. 8, 451-456 (1999).

38. A. Obraztsov, E. Obraztsova, A. Zolotukhin and A. Tyurnina, J. Exp. Theor. Phys.

106, 569-574 (2008).

39. A. C. Ferrari, Solid State Commun. 143, 47 (2007).

40. W. F. Gale and T. C. Totemeier, (Butterworth-Heinemann, Great Britain, 1992), pp.

148- 154.

41. W. F. Gale and T. C. Totemeier, (Elsevier Butterworth-Heinemann, The Netherlands,

2004), pp. 19- 25.

42. A. Moisala, A. Nasibulin and E. Kauppinen, J. Phys.: Cond. Matt. 15, S3011-S3035

(2003).

43. F. Tuinstra and J. Koenig, J. Chem. Phys. 53, 1126 (1970).


123

44. A. C. Ferrari and J. Robertson, Phys. Rev. B 61, 14095 (2000).

45. Z. Ni, Y. Wang, T. Yu and Z. Shen, Nano Res. 1, 273-291 (2008).

46. S. J. Chae, F. Güneş, K. K. Kim, E. S. Kim, G. H. Han, S. M. Kim, H.-J. Shin, S.-M.

Yoon, J.-Y. Choi, M. H. Park, C. W. Yang, D. Pribat and Y. H. Lee, Adv. Mater. 21,

2328-2333 (2009).

47. S. He, K.-K. Liu, S. Su, J. Yan, X. Mao, D. Wang, Y. He, L.-J. Li, S. Song and C.

Fan, Anal. Chem. 84, 4622-4627 (2012).

48. K. Yamamoto, Y. Koga, S. Fujiwara and F. Kokai, Jpn. J. Appl. Phys. 36, L1333

(1997).

49. J. J. Gaumet, A. Wakisaka, Y. Shimizu and Y. Tamori, J. Chem. Soc., Faraday Trans.

89, 1667-1670 (1993).

50. K. J. Koivusaari, J. evoska and S. . vuori, J. Appl. hys. 85, 2915 (1999).

51. M. N. R. Ashfold, F. Claeyssens, G. M. Fuge and S. J. Henley, Chem. Soc. Rev. 33,

23-31 (2004).

52. R. K. Singh and J. Narayan, Phys. Rev. B 41, 8843-8859 (1990).

53. K. Yamamoto, Y. Koga, S. Fujiwara, F. Kokai and R. B. Heimann, Appl. Phys. A

Mater. Sci. Process. 66, 115-117 (1998).

54. P. Murray and D. Peeler, J. Electron. Mater. 23, 855-859 (1994).

55. P. R. Willmott and J. R. Huber, Rev. Mod. Phys. 72, 315 (2000).

56. D. R. Lide, CRC Handbook of Chemistry and Physics: A Ready-reference Book of

Chemical and Physical Data. (CRC Press, Boca Raton, Fla, 2008).

57. A. Peled, Photo-excited Processes, Diagnostics, and Applications: Fundamentals and

Advanced Topics. (Kluwer Academic, Norwell, MA, 2003).

58. S. R. P. Silva, Properties of Amorphous Carbon. (INSPEC, London, 2003).


124

59. J. Robertson, Thin Solid Films 296, 61-65 (1997).

60. P. A. Khomyakov, G. Giovannetti, P. C. Rusu, G. Brocks, J. van den Brink and P. J.

Kelly, Phys. Rev. B 79, 195425 (2009).

61. K. L. Jensen, J. Lebowitz, Y. Y. Lau and J. Luginsland, J. Appl. Phys. 111, 054917-

054919 (2012).

62. H. J. Jeong, H. D. Jeong, H. Y. Kim, S. H. Kim, J. S. Kim, S. Y. Jeong, J. T. Han and

G.-W. Lee, Small 8, 272-280 (2012).

63. G. Eda, H. Emrah Unalan, N. Rupesinghe, G. A. J. Amaratunga and M. Chhowalla,

Appl. Phy. Lett. 93, 233502 (2008).

64. J. Blitz, Electrical and Magnetic Methods of Non-Destructive Testing. (Chapman &

Hall, London, 1997).

Chapter 5 Graphene-based field emitters by EPD

125

Chapter 5

Graphene-based field emitters by

electrophoretic deposition

In this chapter, chemically exfoliated graphene was used to fabricate graphene-

based field emitters using electrophoretic deposition (EPD). Free-standing graphene

emitters were studied in section 5.3, while a hybrid of CNT and graphene emitters were

fabricated and investigated in section 5.4.

5.1 Introduction

While there are currently a number of methods for graphene fabrication, only

certain methods are suitable for the mass production of graphene layers. For instance,

while CVD on metals is widely used to form single to FLG, a transfer technique is

required to release the grown graphene from the underlying metal substrate to obtain

free-standing graphene.1-3

Thus, this may be time-consuming for applications where a

large amount of individual graphene flakes are required. This may especially be so in the

case of field emission, where the emitter is generally preferred to be of a high density and

a high aspect ratio.4 Both of these requirements can be fulfilled by chemically exfoliated

graphene. This is because the small size of resultant flakes can provide numerous

emission sites and the high aspect ratio of a graphene flake (lateral size to thickness ratio)

facilitates field enhancement. In chemical exfoliation, graphene flakes are prepared by

the oxidation of graphite to graphite oxide, which is then exfoliated into graphene oxide


126

(GO). Reduction of GO was typically achieved with hydrazine, which is highly toxic and

unstable.5 However, in recent years, a safer microwave assisted reduction method that

involves heating GO organic or aqueous suspension in a microwave has been developed.6

Thus, with the recent advances in chemical exfoliation of graphene, a high yield of

graphene can be obtained with safer methods. As a result, low-cost and environmentally

green field emitters can be achieved.7, 8

While field emission for other carbon materials such as carbon nanotubes,

diamond-like carbon and its composites has been intensively studied, the area of

graphene field emission applications is studied to a lesser extent. Some early works

include the fabrication of graphene composites by spin coating to lift the horizontal

graphene from the substrate by Goki et al.9 and the CVD preparation of vertically aligned

graphene nanowalls by Malesevic et al..10

Wu et al.11

had also prepared graphene field

emitters with EPD and compared it to an emitter made by directly coating graphene

powder onto conductive tape, thus proving the effectiveness of EPD for field emission

applications. In section 5.3, graphene field emitters were fabricated using EPD, which is

a low cost, high deposition rate and easily up-scaled method to deposit the graphene

emitters. The effects of density and deposition time, on field emission properties of the

samples were then studied.

In section 5.4, emitters made of CNT and graphene materials were fabricated and

investigated for their FE properties. Since the discovery of CNT in 1991, it has attracted

immense interest in various applications ranging from field emission,12

energy storage,13

composite reinforcement,14

and sensors.15

Even with the discovery of its carbon


127

counterpart, isolated graphene sheets, which is cited to be able to outperform CNT in

various applications, it may sometimes be beneficial to combine both the materials to

produce a synergistic effect, instead of replacing either one or the other entirely in the

fabrication of materials for applications.

Methods to combine these two materials such as layer-by-layer assembly,16

spin

coating17

and filtration18

have been studied by various groups of researchers. For example,

Yu et al. have proposed a layer-by-layer electrostatic self-assembly of CNT/graphene

hybrid films by immersing a substrate into aqueous solutions of complementarily

functionalized materials for supercapacitor applications.16

Furthermore, the synergistic

behavior of the hybrid material films has also been demonstrated in current literature. For

instance, Cai et al. fabricated conductive hybrid GO/CNT films on glass substrates, in

which CNT imparted conductivity to the otherwise insulator film and GO acted as a

carrier, which imparted the ability of layer-by-layer assembly and compatibility with the

glass substrate due to the hydrophilic oxygen groups on GO.19

In addition, CNT films have also been proposed for transparent electrode

applications as voids in the CNT network increases the transparency of the film. However,

a large amount of voids also reduces the overall conductance, giving rise to a tradeoff for

transparency and conductivity in such materials. Even though pristine graphene possess

excellent electronic and mechanical properties, the difficulties in obtaining high quality

graphene can lead to structural defects that compromise conductivity and mechanical

strength. As such, this leads to much lower actual values for the various graphene

properties as compared to theoretical predictions. Hence, this limits the graphene

replacement of CNT for applications. Instead, desirable results can be obtained if both


128

materials work complementarily as shown by Li et al.17

who fabricated free-standing

graphene/CNT hybrid films through the transfer of CNT films onto CVD fabricated

graphene on Cu and subsequent etching of Cu. The CNT voids were found to be filled

with nanosized graphene, resulting in a highly transparent film with improved

conductivity as compared to CNT or graphene alone.18

Hybrid materials fabricated using solution processing methods such as vacuum

filtration, have also been systematically studied to directly compare the electrical and

mechanical properties of pure CNT films with graphene/CNT hybrid samples. Khan et al.

reported that hybrid graphene/CNT films raised the stiffness of the CNT-only films from

2 to 4.8 GPa at 20 wt. % and the conductivity of the films increased until a 95 wt. %

loading of graphene.20

However, it can be seen from their SEM images that when such

filtration methods were used to prepare hybrid films, the surface morphology was flat and

may not be suitable for FE applications. The films by filtration can be easily transferred

to any substrate but this would also mean that the electrical connectivity with the

substrate may not be ideal for FE.

In section 5.3, free-standing graphene emitters with very low turn-on field were

fabricated for FE application. However, it was observed that when the voltage was raised

to a relatively high value, the graphene microflakes would be elevated towards the ITO

glass slide. The burnt-out voltage value decreased with an increase in EPD deposition

time. Thus in section 5.4, a hybrid graphene/CNT emitter where the CNT was envisioned

to act like a safety belt mechanism to strap down the graphene microflakes in order to

improve the reliability at higher voltages was proposed. The hybrid films were fabricated

by EPD and their field emission properties in relation with deposition time were studied.


129

5.2 Experimental method

The graphite powder, potassium permanganate (KMnO4) and potassium sulfate

(K2SO4) used in this experiment were purchased from Sinopharm Chemical Reagent Co.

Ltd. Other solutions used in the preparation process are 98 % hydrosulfuric acid (H2SO4),

30 % hydrogen peroxide (H2O2) aqueous solution and ultrapure water. Modified

Hummers method21

was used to oxidize the graphite power into graphite oxide. The

graphite oxide was exfoliated into hydrophilic GO by ultra-sonication. The prepared GO

was then added to ultrapure water. 0.2 ml of ammonia was then added to this suspension,

which adjusted the overall pH value of the dispersion to 10. The exfoliation and

reducibility of the GO has been reported to be greatly improved with the presence of

ammonia.22

This suspension was then placed in a microwave system (Microwave

Synthesis For Scale Up CEM Inc.) at a temperature of 150 oC for 30 min, with the RF

power adjusted to 100 W. The suspension was then cooled to room temperature and

centrifuged at 3000 rpm/min to get a homogenously dispersed graphene suspension.

After drying, 80 mg of graphene powder was added to 200 ml of ethanol to give a

0.4 mg/ml graphene suspension. 10 mg of Al(NO3)3 was added as charger material and

the suspension was ultrasonicated for 1 hr. The deposition durations used were 5, 10 and

15 min. For the hybrid emitters, graphene/CNT mixture solutions were prepared by

adding 40 mg of graphene and 40 mg of CNT into 200 ml of ethanol to give a 0.4 mg/ml

hybrid suspension with 50-50 wt.% composition. The deposition timings used were 2.5, 5

and 10 min.


130

A piece of conductive carbon paper was used as the positive electrode (anode),

while the substrate, highly doped Si (n++

) was used as the negative electrode (cathode).

The electrodes were then submerged into the graphene suspension under an applied

voltage of 20 V. The as-prepared size of the graphene field emitters for demonstration

was 2 × 1 cm each. The graphene layers on these samples were viewed with a JEOL JEM

2010F high resolution TEM (200KV FEG) operated at 200 kV. A Philips XL-30 FE-

SEM was used to obtain sub-surface images of the samples. For larger area FE testing, a

parallel plate configuration was used. The graphene cathode was separated from the

indium tin oxide (ITO) coated glass slide anode by a 100 µm thick polymer spacer. The

measurements were taken under high vacuum conditions (1×10-6

Torr), with a LabView 8

software and Keithley 2410 high voltage source and current measurement unit. The turn-

on voltage was defined as the voltage needed to produce a current density of 10 µA/cm2.

5.3 Large area high density graphene for field emission

5.3.1 Microstructural characterization

Figure 5.1 shows the low magnification SEM images of the graphene emitters

after 5, 10 and 15 minutes of deposition. It was observed that a uniform spread of

graphene flakes was obtained throughout the substrate. At higher magnifications, it can

be seen that the flakes were randomly stacked and oriented, with some flakes protruding

higher above others and some almost perpendicular to the substrate. This can be directly

observed with a cross-section SEM images in Fig. 5.2. A configuration like this is


131

beneficial towards field emission as it improves the field enhancement of the surface.

Each flake remained thin after the EPD deposition process and the distribution of the

flakes was largely homogenous and no particular aggregation in specific areas was

observed. With the increased deposition time, the density of graphene flakes increased

such that the flakes begin to stack onto one another without contact with the substrate,

while the orientation still remained random. Figure 5.3(a) shows the TEM image of two

as-prepared graphene flakes (before EPD deposition), one of which is folded into half

while the other is lying almost entirely flat. The flakes consisted of areas, which were

scrolled and crumpled. This is typical for graphene sheets due to their extreme thinness

and thermodynamic nature. Lattice fringes can also be observed at higher magnifications

of the scrolled or folded regions in Fig. 5.3(b), while unfolded regions were generally

featureless. The largest dimension of the flake is about 2 μm. The almost featureless and

highly transparent areas on the flakes were monolayer graphene areas. Selected area

electron diffraction (SAED) of the flake gave a six fold symmetry diffraction dots

showing the high crystalline quality of the individual flake.


132

Fig. 5.1 SEM images of EPD deposited graphene field emitters with deposition time of (a, b) 5, (c,

d) 10 and (e, f) 15 min. Images on the left column were taken at 1000X, while those on the right

were at 20,000X magnification.


133

Fig. 5.2 Cross-section SEM image of EPD deposited graphene showing near vertical orientation

of microflakes.

Fig. 5.3 (a) TEM images showing two as-prepared graphene flakes, one folded and one flat.

SAED (inset) confirmed that the graphene flakes were of high crystalline quality. (b) Lattice

fringes of scrolled or folded areas at high magnification.


134


The field emission characteristics of the samples deposited with 5, 10 and 15 min

are shown in Fig. 5.4. The turn-on field was 0.68 and 0.88 V/μm for the 5 and 10 min

samples respectively, while the sample deposited with 15 min burnt-out (short circuited)

before reaching the required current value.

From the Fowler-Nordheim (F-N) plots seen in Fig. 5.5, it can be seen that

electron emission did occur by tunneling as the data fitted well into a straight line when

ln (J/E2) versus 1/E. By taking the work function of graphene to be 5 eV (the same as

graphite),10

the field enhancement factors, β, were calculated from the slopes of the F-N

plots using equation 1.4 to be 39,500 and 17,300, for the 5 min and 10 min samples

respectively.

Fig. 5.4 Field emission J-E curves of the 5, 10, 15 min samples.


135

Fig. 5.5 F-N plots for 5 and 10 min samples. β can be calculated from the slope of the straight

line fit.

5.3.3 Discussion

Many low dimensional carbon structures, such as CNT, carbon and metal

nanoclusters embedded in DLC have been studied for field emission applications. The

turn-on fields for CNT fabricated by EPD ranged from 0.83–3.4 V/um.23-26

As such, from

the above results, it can be seen that the turn-on fields (0.68 and 0.88 V/um) for EPD

graphene emitters were comparable, if not lower as compared to the turn-on fields for

CNT deposited by EPD technique. Graphene grown by CVD on Cu substrates have also

been studied previously, and field enhancement was achieved by patterning the planar

graphene surface with photolithography and wet etching to obtain a field emission array

with each element being 30 × 30 μm2 and 20 μm apart.

27 They obtained a turn-on field of

7.2 V/μm at a current of 100 nA/cm2. This may be attributed to the relatively large


136

individual emitting element as compared to the dimension of a graphene or a CNT,

leading to poorer field enhancement properties.

It was observed that the 15 min sample burnt out at higher voltages before

reaching the required turn-on current value for emission. This was attributed to the

deposition time for being excessively long, leading to poor adhesion of the top layer of

graphene flakes. This can be better understood by looking at the reactions occurring at

the electrodes when a voltage is applied. Since the voltage applied is high enough (>5 V),

electrolysis of water occurs at the cathode as described in the following equation 5.1. It

should be noted that if the current of the electrochemical cell is too high, excessive H2

will be produced and bubbling may be observed at the substrate surface. This is undesired

as the bubbles will affect the adhesion of depositing species onto the substrate.

2 H2O+2e- H2 (g) + 2OH

- (5.1)

The Al(NO3)3 charger also dissociates in water to give the following ions as

shown in equation 5.2 and Al3+

is preferentially adsorbed onto the graphene flakes.

Al(NO3)3 Al(NO3)+

+ 2NO3-

(5.2)

Using nitrates as the charger material can also provide additional OH- ions by the

following reduction reactions of the nitrate ion.

NO3- + H2O +2e

- NO2

- + 2OH

- (5.3)

NO3- + 6H2O + 8e

- NH3 + 9OH

- (5.4)

The Al3+

adsorbed onto the graphene flakes, then reacts with at the cathode where

there is a high concentration of OH- ions to form metal hydroxide bonds.

Al(NO3)+

+ 3OH- Al(OH)3 +NO3

- (5.5)


137

The adhesion of the depositing species to the substrate is thus achieved by the

formation of metal hydroxides from metal charger ions (adsorbed on graphene flakes) at

the cathode surface, which contributes to hydrogen bonding between graphene flakes and

substrate.28-32

However, once the substrate surface is covered uniformly with graphene

flakes, the subsequent incoming flakes will begin to pile on top of the existing flakes and

have no contact with the substrate. These piled up flakes have a weaker bonding and can

lead to a higher contact resistance between the top and bottom graphene flakes, which in

turn hinders electron emission from the surface.33

Flakes that were piled onto each other

were more easily attracted toward the anode (ITO glass slide) and cause a short circuit

before a satisfactory field emission current was achieved. On the other hand, the higher

turn-on field for the 10 min sample, as compared to the 5 min sample, was attributed to

the more pronounced screening effect due to the much denser graphene flakes

configuration present.

The low turn-on field for EPD graphene emitters was mainly attributed to (i) the

high conductivity of graphene, (ii) the high density of emitting elements and (iii) random

orientation of the individual flakes, leading to protrusions and increasing the local field

enhancement effect. Some flakes were oriented normally to the substrate and electron

emission can occur from the graphene flake edges. Additionally, when a field is applied,

the free-standing flakes could align to the direction of the electric field while pivoting on

their bond with the substrate, thus effectively enhancing field emission. Emission from

graphene edges has been well agreed by researchers to be the key contributing emitting

element.34-36

However, when the density of the emitting elements (graphene flakes in this

case) became too high, electron screening occurred and this reduced the overall


138

effectiveness of the field emitter as the entire surface appeared electronically flat with

poor local field enhancements. Moreover, the reliability of EPD fabricated emitters was

also reduced when the stacking of graphene flakes occurred with an increased deposition

duration, which results in the emitter being easily burnt out. On the other hand, if the

deposition time is too short, the graphene flakes do not cover the surface of the substrate

uniformly. The possible scenarios are depicted in Fig. 5.6 Deposition time should be

controlled to maximize the field enhancement effects of the overall emitter surface. As

such, with the correct optimizations during fabrication, graphene-based field emitters

could potentially overtake the performance of CNT as emitters.

Fig. 5.6 Schematic diagram showing the possible scenarios when deposition time is (a) short or (b)

extended. When flake density is high, a flatter overall surface morphology gave poorer field

enhancing properties due to screening effect.


139

5.4 Graphene/carbon nanotube hybrid materials for field

emission

5.4.1 Microstructural characterization

From the SEM images in Fig. 5.7, it can be seen that the microstructure of the

hybrid films consisted of the typical tangled spaghetti-like CNT network with some areas

filled with graphene microflakes. At higher magnifications, the graphene microflakes

were observed to be either on the surface or embedded within the network of CNT. This

morphology was consistent throughout the entire sample, as well as through all the

deposition timings tested. The surface morphology of the samples appeared to be rougher

and had more graphene microflakes protrusions when the deposition time was increased

from 2.5 to 5 and 10 min. In contrast to the 5 and 10 min samples, at the shortest duration

of 2.5 min, larger CNT-only areas between protrusions were observed from the low

magnification SEM images. Even though the surface graphene sheets were “strapped”

down by the CNT strands, some sheets maintained a vertical position as seen in the high

magnification SEM image of the 2.5 min sample in Fig. 5.7(b) and the cross section SEM

image in Fig. 5.8(a). When the surface morphologies of hybrid samples deposited by

EPD and vacuum filtration method were compared using a SEM image in Fig 5.8(b),20

it

was observed that vacuum filtration method gave a much flatter morphology as compared

to EPD.


140

The TEM images of pre-deposited graphene flakes and CNT use in preparing the

EPD solution are shown in Fig. 5.9. The inset shows the high crystalline quality of the

graphene flakes and the CNT were multiwalled with a diameter of 20–40 nm each.

Fig. 5.7 SEM images of 2.5, 5 and 10 min graphene/CNT hybrid samples by EPD. Left column

shows magnification at 1000X while right column shows 20,000X.


141

Fig. 5.8 (a) Cross-section SEM images showing morphology with multiple protrusions and a

vertically aligned CNT-strapped graphene sheet (white arrow) (b) Hybrid graphene/CNT film by

filtration at 50% graphene in ref. 20 shows a flatter overall surface morphology.20

Reprinted from

Carbon 48, 2825(2010) with permission from Elsevier.

Fig. 5.9 TEM images showing (a) high crystalline quality graphene (b) CNT used in the

experiments.


142


The field emission I-V curves of the hybrid samples in comparison to a CNT-only

sample were presented in Fig. 5.10. The turn-on fields of the 2.5, 5 and 10 min samples

were 2.84, 2.50, 1.30 V/µm respectively, as compared to 3.25 V/µm for a 5 min CNT-

only sample. The turn-on field was generally improved for the hybrid films as compared

to pure CNT films. Even though the turn-on field was still higher than pure graphene

films, better reliability and less burnt-out situations were observed for the hybrid samples

at higher voltages.

The respective F-N plots for the hybrid samples are shown in Fig. 5.11. Using a

work function of 5 eV for carbon, the β obtained from equation 1.4 was 1800, 3400,

19700 for the 2.5, 5 and 10 min hybrid samples respectively.

Fig. 5.10 I-V plots of graphene/CNT hybrid emitters.


143

Fig. 5.11 F-N plots of graphene/CNT hybrid emitters.

5.4.3 Discussion

It has been established that by combining both CNT and graphene, hybrid

materials that gave improved properties can be produced. The method of fabrication can

also affect the performance of the hybrid materials in certain applications. For instance,

samples deposited with vacuum filtration methods (Fig. 5.8) gave smoother surface

morphologies as compared to EPD and may not be ideal for FE applications. In contrast,

EPD gave hybrid samples with many surface protrusions that can aid the local field

enhancement. In these experiments, it was observed that there was more CNT than

graphene on the hybrid samples, despite the fact that the CNT and graphene hybrid

suspension was prepared to contain a 50-50 wt. % of each type of material. This became

less pronounced with increasing deposition times as the amount of graphene deposited


144

was observed to increase when the deposition duration was increased to 10 and 15 min.

The duration needed to fully cover the substrate macroscopically was also reduced for

hybrid suspensions as compared to pure graphene suspensions. This may be due to the

different drag forces experienced by the CNT and graphene flakes in the solution, when

moving towards the cathode. CNT have a more streamlined body (20–40 nm) as

compared to graphene microflakes, which can be of sizes up to 10 μm. Thus, the

graphene flakes faced a larger drag force when moving through the solution. Larger

flakes can also take a longer time to arrive. As such, larger flakes, in addition to small

flakes were deposited in the 10 min sample as compared to the 2.5 and 5 min samples.

However, the delay in arrival times of the two species was not significant enough to

cause phase separation. The microstructure of the hybrid was therefore well mixed

throughout when the deposition duration was set beyond the minimum time needed for

the graphene flakes to arrive on the substrate. It is worthy to note that this issue would not

be present in methods such as spin coating and vacuum filtration, which does not involve

the electrostatic attraction between species in the solution and substrate. As such, the 50-

50 wt. % of CNT and graphene would be maintained in these methods.

To achieve improved FE characteristics, it is ideal that the material has a good

conductivity within, as well as possessing areas of local field enhancement. These can be

in the form of (i) physical surface protrusions or (ii) areas where field lines are

concentrated at the surface of the emitter, due to different conductivities of phases in a

composite. An example of the former would be the samples studied in this chapter, while

the latter is the material fabricated and discussed in Chapter 3. Graphene and CNT each

played a slightly different role in this hybrid material even though both of them


145

ultimately contributed to field emission at the surface. Within the film, the microstructure

can be viewed as graphene embedded within the CNT matrix. However, at the surface

regions where FE occurs, the CNT strands coated the graphene flakes instead. As such,

graphene had at least two roles; 1) to act as connection particles between the CNT strands

and improve the conductivity within the porous CNT network when they were embedded

inside, 2) to act as additional local field enhancing spots when they were located at near

surface regions. On the other hand, the role of CNT was to strap the graphene flakes

down at high voltages such that they would not detach from the substrate and elevate

towards the ITO glass slide, thereby improving the reliability of the emitters.

From the results, it was observed that with an increase in deposition time, the

amount of graphene flakes on the surface and the thickness of the film increased. As the

conductivity of porous CNT films depend on the contact resistance of inter-tube junctions,

having more graphene flakes embedded would create extra connecting conduction paths

arising from sheet-tube junctions and thus increased film conductivity.20, 37

Also, having a

larger amount of graphene sheets on the surface, when the deposition time was increased,

served to provide more emission sites at the surface. This in turn led to lower emission

turn-on fields. However, the turn-on fields of the hybrid films were higher than pure

graphene films. This could be due to the inability of the strapped-down graphene sheets

to align to the direction of the applied field as in the case of the free-standing graphene-

only emitters. The obtained results from section 5.3 and 5.4 also suggested that with a

decrease in CNT fraction and more graphene flakes at the surface, turn-on field could be

further lowered for the hybrid samples.


146

5.5 Summary

In summary, uniform 2 × 1 cm2 graphene field emitters were fabricated by EPD in

section 5.3. The turn-on field was 0.68 and 0.88 V/um for the 5 min and 10 min samples

respectively. The higher turn-on field of the 10 min sample was attributed to the very

high density of graphene flakes, which led to electron screening and lowered the overall

field enhancement of the emitter. The random orientation of the graphene flake led to

local field enhancement due to surface protrusions, as well as numerous emission sites

which include those from the edges of graphene flakes oriented normal to the substrate.

Graphene field emitters fabricated by EPD, in combination with chemical exfoliation and

microwave assisted reduction, presented a low-cost, high yield and safe method for

graphene field emitters that can easily up-scaled.

Graphene/CNT hybrid films were prepared using EPD in section 5.4. The

fabricated films consisted of well-dispersed graphene sheets within a spaghetti CNT

network and possessed a surface morphology that was favorable for FE applications. The

results revealed that with an increase in deposition duration, the turn-on field decreased

for the hybrid materials. The turn-on fields of the 2.5, 5 and 10 min samples were 2.84,

2.50, 1.30 V/µm respectively. Lower turn-on fields for the hybrid samples as compared to

pure CNT samples were attributed to the increased conductivity within the samples by

intersheet-tube contacts and additional field enhancing graphene emission spots on the

surface. These hybrid materials offer a range of possibilities for new materials for

nanotechnology.


147

References

1. L. G. De Arco, Z. Yi, A. Kumar and Z. Chongwu, IEEE Trans. Nanotechnol. 8, 135-

138 (2009).




Kong, Nano Lett. 9, 30-35 (2008).

4. J. M. Bonard, J. P. Salvetat, T. Stöckli, L. Forró and A. Châtelain, Appl. Phys. A

Mater. Sci. Process. 69, 245-254 (1999).

5. S. Pei and H.-M. Cheng, Carbon 50, 3210-3228 (2012).

6. Y. Zhu, S. Murali, W. Cai, X. Li, J. W. Suk, J. R. Potts and R. S. Ruoff, Adv. Mater.

22, 3906-3924 (2010).

7. Y. Zhu, S. Murali, M. D. Stoller, A. Velamakanni, R. D. Piner and R. S. Ruoff,

Carbon 48, 2118-2122 (2011).

8. G. M. Morales, P. Schifani, G. Ellis, C. Ballesteros, M. Gerardo, C. Barbero and H. J.

Salavagione, Carbon 49, 2809-2816 (2011).

9. G. Eda, H. E. Unalan, N. Rupesinghe, G. A. J. Amaratunga and M. Chhowalla, Appl.

Phys. Lett. 93, 233502 (2008).



11. Z.-S. Wu, S. Pei, W. Ren, D. Tang, L. Gao, B. Liu, F. Li, C. Liu and H.-M. Cheng,

Adv. Mater. 21, 1756-1760 (2009).

12. Z. W. Pan, F. C. K. Au, H. L. Lai, W. Y. Zhou, L. F. Sun, Z. Q. Liu, D. S. Tang, C. S.

Lee, S. T. Lee and S. S. Xie, J. Phys. Chem. C 105, 1519-1522 (2001).


148

13. Z. Chen, Y. Qin, D. Weng, Q. Xiao, Y. Peng, X. Wang, H. Li, F. Wei and Y. Lu, Adv.

Funct. Mater. 19, 3420-3426 (2009).

14. W. Ma, L. Liu, Z. Zhang, R. Yang, G. Liu, T. Zhang, X. An, X. Yi, Y. Ren, Z. Niu, J.

Li, H. Dong, W. Zhou, P. M. Ajayan and S. Xie, Nano Lett. 9, 2855-2861 (2009).

15. M. Pumera, Chem. Rec. 9, 211-223 (2009).

16. D. Yu and L. Dai, J. Phys. Chem. C 1, 467-470 (2009).

17. V. C. Tung, L.-M. Chen, M. J. Allen, J. K. Wassei, K. Nelson, R. B. Kaner and Y.

Yang, Nano Lett. 9, 1949-1955 (2009).

18. C. Li, Z. Li, H. Zhu, K. Wang, J. Wei, X. Li, P. Sun, H. Zhang and D. Wu, J. Phys.

Chem. C 114, 14008-14012 (2010).

19. D. Cai, M. Song and C. Xu, Adv. Mater. 20, 1706-1709 (2008).

20. U. Khan, I. O’Connor, Y. K. Gun’ko and J. N. Coleman, Carbon 48, 2825-2830

(2010).

21. W. S. Hummers and R. E. Offeman, J. Am. Chem. Soc. 80, 1339-1339 (1958).

22. I. Janowska, K. Chizari, O. Ersen, S. Zafeiratos, D. Soubane, V. Costa, V. Speisser, C.

Boeglin, M. Houllé, D. Bégin, D. Plee, M.-J. Ledoux and C. Pham-Huu, Nano Res. 3,

126-137 (2010).

23. S. M. Jung, J. Hahn, H. Y. Jung and J. S. Suh, Nano Lett. 6, 1569-1573 (2006).

24. L. Wang, Y. Chen, T. Chen, W. Que and Z. Sun, Mater. Lett. 61, 1265-1269 (2007).

25. W. Y. Sung, S. M. Lee, W. J. Kim, J. G. Ok, H. Y. Lee and Y. H. Kim, Diam. Relat.

Mater. 17, 1003-1007 (2008).

26. Y. Chen, H. Jiang, D. Li, H. Song, Z. Li, X. Sun, G. Miao and H. Zhao, Nanoscale

Res. Lett. 6, 1-8 (2011).

27. C.-K. Huang, Y. Ou, Y. Bie, Q. Zhao and D. Yu, Appl. Phys. Lett. 98, 263104 (2011).


149

28. B. J. C. Thomas, A. R. Boccaccini and M. S. P. Shaffer, J. Am. Chem. Soc. 88, 980-

982 (2005).

29. O. O. Van der Biest and L. J. Vandeperre, Annu. Rev. Mater. Sci. 29, 327-352.

30. J. B. Talbot, E. Sluzky and S. K. Kurinec, J. Mater. Sci. 39, 771-778 (2004).

31. J. Cho, S. Schaab, J. Roether and A. Boccaccini, J. Nanopart. Res. 10, 99-105 (2008).

32. R. Hashaikeh and J. Szpunar, J. Coating. Tech. Res. 8, 161-169 (2011).

33. L. Gomez De Arco, Y. Zhang, C. W. Schlenker, K. Ryu, M. E. Thompson and C.

Zhou, ACS Nano 4, 2865-2873 (2010).

34. C. Wu, F. Li, Y. Zhang and T. Guo, Carbon 50, 3622-3626 (2012).

35. H. Yamaguchi, K. Murakami, G. Eda, T. Fujita, P. Guan, W. Wang, C. Gong, J.

Boisse, S. Miller, M. Acik, K. Cho, Y. J. Chabal, M. Chen, F. Wakaya, M. Takai and

M. Chhowalla, ACS Nano 5, 4945-4952 (2011).


1265-1268 (2012).

37. P. E. Lyons, S. De, F. Blighe, V. Nicolosi, L. F. C. Pereira, M. S. Ferreira and J. N.

Coleman, J. Appl. Phys. 104, 044302-044308 (2008).

Chapter 6 Conclusion and future work

150

Chapter 6

Conclusions and future work

6.1 Conclusions

The primary objective of this work was to fabricate sp2 and graphene-based

materials for field emission applications. The first step taken towards this aim was to

perform carbon deposition on Si substrates at 100, 400 and 700 oC with a PLD system to

grow nanographite in a-C matrix. It was shown that the sp2 content in the samples

increased with the increase in deposition temperature. Additionally, the sp2 content of the

film was mainly agglomerated into nanoclusters as seen in TEM observations. From

AFM surface morphology images, the surface of the samples were highly smooth and

external field enhancement factor, βex, was taken to be 1. The 700 oC sample had the

lowest turn-on voltage among the samples and this was attributed to the improved

internal enhancement by the high density of nanographite clusters present. The electrical

conduction within the films was improved through conduction channels formed either by

overlapping nanographite clusters or by electron hopping between the clusters. Hence,

the presence of such conduction channels gave the flat films a CNT-like emission

mechanism. The relative field emission barrier heights were also estimated for the 3

samples with β100: β400: β700 = 1: 1.46: 2.02.

By extending the method of carbon deposition with heating during pulsed laser

ablation of carbon to a transition metal substrate (e.g. Ni) in place of Si substrate, few-

layer graphene was fabricated from solid state transformation of PLD deposited carbon.


151

The slow cooling (1 oC/min) and medium cooling rates (50

oC/min) gave few-layer

graphene, while the fast cooling rate gave disordered carbon due to insufficient time for

the carbon atoms to arrange into ordered crystalline structures. The graphene fabricated

were 3–4 layers thick by estimation from the I(2D)/I(G) ratio of Raman peaks. The FLG

could also be directly observed under TEM. In addition, lower pulsed laser energies of 50

or 100 mJ gave FLG formation, while no FLG was detected for the pulsed laser energy of

200 mJ. This was due to a shallower implantation of carbon atoms into the metal

substrate at lower energies, which in turn facilitated the precipitation process.

Graphene segregation using PLD was subsequently studied with the use of

various metal substrates such as Ni, Cu, Co and Fe. By using the 1 oC/min cooling profile,

graphene was only formed on Ni but absent in the other three metals. On one hand, the

absence of FLG on Cu was due to its extremely low carbon solubility. On the other hand,

for a high carbon solubility metal like Co, a suitable cooling recipe is needed for FLG

formation. By increasing the cooling rate to suit Co, FLG was also formed on Co metal

with a high coverage. The fabricated graphene on Co was transferred onto a Si substrate

by etching the Co metal substrate with dilute acid. After the transfer, it was observed that

the D peak signal dropped significantly. This proved that a large part of the D peak signal

was contributed by the amorphous carbon that was still embedded within the metal

substrate after the completion of the segregation process. This was further confirmed by

the reduced D peak signal of Raman spectra taken from areas exposed and unexposed to

the carbon plume during deposition.

In addition, graphene segregation using PLD was also studied with the use of

different wavelength lasers such as Nd:YAG (λ=266 nm) laser as compared to the KrF


152

(λ=248 nm). Even though the energy of the carbon species in the Nd:YAG system was >1

eV smaller than those in the KrF system, FLG was formed on Ni substrates with the same

cooling profile for both lasers. On a side note, Nd:YAG might have a slight advantage

over KrF with its lower energy carbon species, which favors the formation of graphitic

sp2 carbon rather than amorphous carbon.

Using the information gained from PLD segregation of graphene on metals, Ni

and Co metal nanocones with varied densities and height distribution were fabricated for

field emission purposes. The field emission properties of the metal nanocones were

greatly enhanced with the graphene layer coating. This was attributed to the charge

transfer interactions between the metal nanocones and graphene, which resulted in a

reduced work function for the combined material. This in turn led to a reduced electron

tunneling barrier height that manifested as a lower turn-on voltage for the graphene-on-

metal nanocones. For both metals, a lower turn-on voltage was obtained from the taller

cones separated with shorter cones in-between them (Type A), while a higher turn-on

voltage was acquired from the high density uniform height nanocones (Type B), due to

more serious electron screening effects experienced in the latter configuration.

Large area graphene emitters were also fabricated using EPD. The fabricated

emitters consisted of randomly oriented graphene flakes with some flakes being vertically

oriented with respect to the substrate. This gave a favorable configuration for field

emission in contrast to methods such as vacuum filtration that generally gave a flatter

surface morphology. With the increase in deposition duration, the density of graphene

flakes on the substrate increased. However, the field emission turn-on voltage also


153

increased with deposition time. The sample that was deposited for 5 min gave the lowest

turn-on voltage followed by the sample that was deposited for 10 min. No considerable

emission was observed for the sample deposited at 15 min before a short circuit occurred.

The poorer emission characteristics of the 10 and 15 min samples were due to the

reduction in surface geometrical enhancement caused by overcrowding of graphene

flakes on the substrate. In addition, as the stacked graphene flakes had no contact with the

substrate, poor contact resistance between the flakes can also lead to higher overall

resistance for the thicker samples. Furthermore, when the deposition duration was

excessively long in the case of 15 min, the samples were easily short circuited at higher

voltages as the piled-up flakes elevated towards the ITO glass slide used as the anode.

To improve the reliability of the graphene EPD emitters, hybrid graphene/CNT

emitters were prepared by EPD. The CNT strands were envisioned to resemble a “safety

belt” mechanism that would hold down the graphene flakes at high voltages. Despite the

50-50 wt.% of CNT and graphene used in the prepared EPD source solution, the

fabricated emitters were predominately CNT with some graphene flakes embedded

within the CNT network. The hybrid materials showed improved field emission

properties as compared to pure CNT emitters fabricated with the same method. This was

because the graphene flakes embedded within the film provided additional conduction

paths for electron transport to the surface in the otherwise porous CNT film. The

presence of graphene flakes also increased the number of emitting spots on the surface of

the hybrid film.


154

From the studies carried out in this dissertation, various effective sp2 and

graphene-based field emitters can be fabricated from physical methods such as PLD,

microstructure enhancements by using sharp cones shapes and solution processing

methods such as EPD. With careful control of cooling profile, parameters affecting

emitter density and combination of materials, the performance of graphene-based field

emitters can be enhanced.

6.2 Future work

In this work, the fabrication of graphene through solid state transformation of

carbon provided by PLD at 750 oC was studied. It has also been shown that using this

technique, FLG can be grown directly onto nano-patterned substrates for various

applications. However, some loss in the structural outline of the metal cones, calls to

attention that the deposition temperature has to be further decreased. As such, it may be

beneficial to explore longer laser wavelengths lasers such as 532 or 1064 nm in future

that can allow for an easier sp2 formation as well as shallower carbon ion implantation

depth. The carbon ions in pulsed laser plumes have sufficient energy to implant carbon

atoms into the metal substrates and this can aid the absorption of sufficient carbon at a

lower temperature. However, if implantation occurs too deeply into the metal substrate,

precipitation can be difficult. Sufficient carbon has to be absorbed by the near-surface

metal regions to exceed the C solubility for segregation to occur and a shallower

implantation depth is preferred since saturation can occur at areas nearer to the surface

and cooling time can be reduced as well.


155

In the other section for large area emitter fabrication with EPD, hybrid

graphene/CNT emitter materials were studied. However, the 1:1 CNT to graphene ratio in

the source solution was not directly transferred onto the substrates during fabrication.

Hence, more control over the EPD process of hybrid materials is desired and the growth

mechanism of hybrid materials using EPD should be investigated. Furthermore, it would

be interesting to investigate the field emission properties of a hybrid emitter, whereby a

large amount of graphene flakes are strapped down by a minimal amount of CNT strands.

Once again, a better understanding of the EPD process is required to achieve such a

calculated configuration.

In closing, the development of graphene-based materials for field emission is in

the stage of infancy. As the field emission properties of graphene are highly affected by

its configuration, fabrication techniques play a critical role in realizing its full potential in

this area. There remain yet many unexplored methods and challenges towards graphene

field emitter fabrication before it can be employed in actual applications. Nonetheless,

with a suitable technique and control of the correct parameters involved, graphene can

prove to triumph other materials of its class and play a pivotal role in the advancement of

field emission devices.

Date post:	07-Jul-2020
Category:	Documents
Upload:	others
View:	5 times
Download:	0 times

FABRICATION OF GRAPHITE AND GRAPHENE-BASED MATERIALS … · 2018-01-10 · fabrication of graphite...

Documents