<%BANNER%>

Growth, Fabrication, and Characterization of Carbon Nanotubes, Nanotube Films, and Nanowires

Permanent Link: http://ufdc.ufl.edu/UFE0022789/00001

Material Information

Title: Growth, Fabrication, and Characterization of Carbon Nanotubes, Nanotube Films, and Nanowires
Physical Description: 1 online resource (122 p.)
Language: english
Creator: Choi, Yongho
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: cvd, nanotube, nanowire
Electrical and Computer Engineering -- Dissertations, Academic -- UF
Genre: Electrical and Computer Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: We explored nucleating the growth of nanomaterials by the ion implantation technique. We present experimental evidence that single-walled carbon nanotubes (SWNT) and SiOX and GaN nanowires can be grown by Fe ion implantation into SiO2/Si substrates, subsequent annealing, and chemical vapor deposition (CVD) growth. Moreover, we show that there is a dose and energy window of ion implantation in which SWNT and nanowire growth are observed for a given growth condition. For nanomaterials growth, catalyst is usually spun on or drop-dried from a liquid solution containing iron nanoparticles or deposited as solid thin film layers by evaporation or sputtering. However, it is not possible to pattern the liquid solution-based catalyst into very small dimensions or the thin film catalyst into nonplanar three-dimensional (3D) device structures, such as the sidewalls of high aspect ratio trenches. By adopting the ion implantation technique for nucleating nanomaterials growth, this thesis opens up the possibility of controlling the origin of nanomaterials at the nanometer scale and of integrating nanomaterials into nonplanar 3D device structures with precise dose control. We also fabricate micromachined Si Transmission Electron Microscopy (TEM) grids for direct TEM, as well as Atomic Force Microscopy (AFM), Scanning Electron Microscopy (SEM), and Raman characterization of as-grown nanomaterials. As a result, these micromachined TEM grids offer fast, easy, and reliable structural characterization. Furthermore these grids provide a low cost, mass producible, efficient, reliable, and versatile platform for direct TEM, AFM, SEM, and Raman analysis of as-grown nanomaterials, eliminating the need for any post-processing growth. We also explored fabrication and characterization of single-walled carbon nanotube films which are three-dimensional films of tens of nanometers thickness, consisting of an interwoven mesh of single-walled carbon nanotubes. We demonstrate, for the first time, patterning of SWNT films down to submicron lateral dimensions as small as 50 nm using e-beam lithography and inductively coupled plasma (ICP) etching. This simple and efficient 'top-down' patterning capability developed could open up tremendous opportunities for integrating single-walled nanotube films into a wide range of electronic and optoelectronic devices. Furthermore, we fabricate and characterize the effect of device geometry on the dark current of metal-semiconductor-metal (MSM) photodetectors based on SWNT film-GaAs Schottky contacts. We observed that dark currents of the MSM devices scale rationally with device geometry, such as the device active area, finger width, and finger spacing. These results open up the possibility of integrating SWNT films as transparent and conductive Schottky electrodes in conventional semiconductor electronic and optoelectronic devices.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Yongho Choi.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Ural, Ant.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022789:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022789/00001

Material Information

Title: Growth, Fabrication, and Characterization of Carbon Nanotubes, Nanotube Films, and Nanowires
Physical Description: 1 online resource (122 p.)
Language: english
Creator: Choi, Yongho
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: cvd, nanotube, nanowire
Electrical and Computer Engineering -- Dissertations, Academic -- UF
Genre: Electrical and Computer Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: We explored nucleating the growth of nanomaterials by the ion implantation technique. We present experimental evidence that single-walled carbon nanotubes (SWNT) and SiOX and GaN nanowires can be grown by Fe ion implantation into SiO2/Si substrates, subsequent annealing, and chemical vapor deposition (CVD) growth. Moreover, we show that there is a dose and energy window of ion implantation in which SWNT and nanowire growth are observed for a given growth condition. For nanomaterials growth, catalyst is usually spun on or drop-dried from a liquid solution containing iron nanoparticles or deposited as solid thin film layers by evaporation or sputtering. However, it is not possible to pattern the liquid solution-based catalyst into very small dimensions or the thin film catalyst into nonplanar three-dimensional (3D) device structures, such as the sidewalls of high aspect ratio trenches. By adopting the ion implantation technique for nucleating nanomaterials growth, this thesis opens up the possibility of controlling the origin of nanomaterials at the nanometer scale and of integrating nanomaterials into nonplanar 3D device structures with precise dose control. We also fabricate micromachined Si Transmission Electron Microscopy (TEM) grids for direct TEM, as well as Atomic Force Microscopy (AFM), Scanning Electron Microscopy (SEM), and Raman characterization of as-grown nanomaterials. As a result, these micromachined TEM grids offer fast, easy, and reliable structural characterization. Furthermore these grids provide a low cost, mass producible, efficient, reliable, and versatile platform for direct TEM, AFM, SEM, and Raman analysis of as-grown nanomaterials, eliminating the need for any post-processing growth. We also explored fabrication and characterization of single-walled carbon nanotube films which are three-dimensional films of tens of nanometers thickness, consisting of an interwoven mesh of single-walled carbon nanotubes. We demonstrate, for the first time, patterning of SWNT films down to submicron lateral dimensions as small as 50 nm using e-beam lithography and inductively coupled plasma (ICP) etching. This simple and efficient 'top-down' patterning capability developed could open up tremendous opportunities for integrating single-walled nanotube films into a wide range of electronic and optoelectronic devices. Furthermore, we fabricate and characterize the effect of device geometry on the dark current of metal-semiconductor-metal (MSM) photodetectors based on SWNT film-GaAs Schottky contacts. We observed that dark currents of the MSM devices scale rationally with device geometry, such as the device active area, finger width, and finger spacing. These results open up the possibility of integrating SWNT films as transparent and conductive Schottky electrodes in conventional semiconductor electronic and optoelectronic devices.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Yongho Choi.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Ural, Ant.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022789:00001


This item has the following downloads:


Full Text

PAGE 1

1 GROWTH, FABRICATION, AND CHARACTE RIZATION OF CARBON NANOTUBES, NANOTUBE FILMS, AND NANOWIRES By YONGHO CHOI A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Yongho Choi

PAGE 3

3 To my parents, Heeyual Park and Hyungjin Choi; my wife, Sungeun Kim; my son, Andrew Minjun Choi, with love.

PAGE 4

4 ACKNOWLEDGMENTS First and foremost, I express my sincere gratit ude to my advisor, Professor Ant Ural. It was one of the great fortunes in my life to work for him and to start my research career under his guidance. His greatest guidance helped me to learn knowledge and also helped me to improve skills of my logical thinking and handling of research projects. With his patience, encouragement, and invaluable advice, I improve d my research abilities and accomplish this work. I would like to express my sincere appreciati on to my committee members, Professor Gijs Bosman, Professor Jing Guo, and Professor Kirk Ziegler, for their invaluable discussions, suggestions and kind supports. My appreciation also goes to Ashkan Behnam, Jason Johnson, Leila Noriega, Joe Portillo, and all other former and current nanotech gr oup members. We discussed a lot of issues together and made a lot of good times as a friend. I give sincere appreciation and love to my family. Their support with belief is the strongest energy for me to stand up and to walk ahead.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4LIST OF TABLES................................................................................................................. ..........8LIST OF FIGURES................................................................................................................ .........9ABSTRACT....................................................................................................................... ............11 CHAPTER 1 INTRODUCTION..................................................................................................................13Single-Walled Carbon Nanotubes..........................................................................................15Single-Walled Carbon Nanotube Geomet rical and Electronic Structure........................15Single-Walled Carbon Nanotube Growth Methods........................................................17GaN Nanowires.................................................................................................................. ....182 SINGLE-WALLED CARBON NANOTUBE GR OWTH FROM ION IMPLANTED IRON CATALYST.................................................................................................................27Introduction................................................................................................................... ..........27Ion Implantation............................................................................................................... .......28Liquid-Based Catalyst.......................................................................................................... ..28Experimental Procedure......................................................................................................... .29Results and Discussion......................................................................................................... ..29AFM Characterization.....................................................................................................29TEM Characterization.....................................................................................................32Raman Spectroscopy Characte rization and Discussion..................................................33Electrical Characterization..............................................................................................36Conclusions.................................................................................................................... .........373 MICROMACHINED SILICON TRANS MISSION ELECTRON MICROSCOPY GRIDS FOR DIRECT CHARACTERIZATI ON OF AS-GROWN NANOTUBES AND NANOWIRES...................................................................................................................... ..47Introduction................................................................................................................... ..........47Design of the TEM Grids.......................................................................................................48Microfabrication of the TEM Grids........................................................................................49Nanotube Growth on the TEM Grids.....................................................................................50Characterization of Nanotubes Grown on the TEM Grids.....................................................50Conclusions.................................................................................................................... .........52

PAGE 6

6 4 SILICON OXIDE NANOWIRE GROWTH FROM IRON ION IMPLANTED SIO2 SUBSTRATES..................................................................................................................... ..58Introduction................................................................................................................... ..........58Experimental Method............................................................................................................ .59Results and Discussion......................................................................................................... ..60Conclusions.................................................................................................................... .........635 GaN NANOWIRE GROWTH FROM ION IMPLANTED IRON CATALYST..................71Introduction................................................................................................................... ..........71Experimental Method............................................................................................................ .72Results and Discussion......................................................................................................... ..73Conclusions.................................................................................................................... .........756 NANOLITHOGRAPHIC PATTERNING OF TRANSPARENT, CONDUCTIVE SINGLE-WALLED CARBON NANOTUBE FI LMS BY INDUCTIVELY COUPLED PLASMA REACTIVE ION ETCHING.................................................................................82Introduction................................................................................................................... ..........82SWNT Film Deposition..........................................................................................................83Lithography.................................................................................................................... .........83O2 Plasma Etching................................................................................................................ ..84Results and Discussion......................................................................................................... ..85Effect of Substrate Bias Power........................................................................................88Effect of Chamber Pressure.............................................................................................89Effect of Substrate Cooling.............................................................................................89Comparison to Parallel-Plate RIE System.......................................................................90Conclusions.................................................................................................................... .........907 FABRICATION AND DARK CURRENT CHARACTERIZATION OF METALSEMICONDUCTOR-METAL (MSM) PHOT ODETECTORS WITH TRANSPARENT AND CONDUCTIVE CARBON NANOTUBE FILM SCHOTTKY ELECTRODES.........98Introduction................................................................................................................... ..........98Experimental Method............................................................................................................ .99Result and Discussion.......................................................................................................... .100Conclusions.................................................................................................................... .......1028 CONCLUSIONS AND FUTURE WORK...........................................................................108Conclusions.................................................................................................................... .......108Suggestions for Future Work................................................................................................109APPENDIX A MASK LAYOUT.................................................................................................................111

PAGE 7

7 LIST OF REFERENCES.............................................................................................................113BIOGRAPHICAL SKETCH.......................................................................................................122

PAGE 8

8 LIST OF TABLES Table page 2-1 Average height and density of catalyst nanoparticles formed after annealing and the average diameter and density of nanotubes grown for each Fe ion implantation condition...................................................................................................................... ......462-2 Density of catalyst nanoparticles formed after growth for each Fe ion implantation condition...................................................................................................................... ......466-1 Etch rates of th e SWNT film and three different resists....................................................97

PAGE 9

9 LIST OF FIGURES Figure page 1-1 Unrolled graphene sheet making up SWNT and its honeycomb lattice............................231-2 Reciprocal lattice of gra phene (unrolled carbon nanotube)...............................................231-3 Energy dispersion relations for grap hene shown throughout whole region of Brillouin zone................................................................................................................. ....241-4 Schematic of CVD system used to grow SWNTs.............................................................251-5 Lattice constant of III-V nitride semi conductor materials as a function of bandgap energy......................................................................................................................... ........262-1 Distribution profiles of vari ous ions in crystalline silic on after implantation at energy of 200 keV..................................................................................................................... .....382-2 AFM images of liquid catalyst and SWNTs grown from the catalyst...............................382-3 AFM images of Fe catalyst nanopartic les formed after annealing and Carbon nanotubes grown by CVD..................................................................................................392-4 AFM images of Fe catalyst nanopartic les formed after annealing and Carbon nanotubes grown by CVD..................................................................................................402-5 Optical microscope image of micromach ined Si TEM grids and HRTEM images of SWNTs.......................................................................................................................... .....412-6 Micro Raman spectra of grown nanotubes........................................................................422-7 Kataura plot showing the electronic tran sition energies for all possible SWNTs as a function of diameter...........................................................................................................432-8 Cross-sectional schematic of pr ocess flow for SWNT transistor......................................442-9 Optical microscope image of fabricated nanotube device sample.....................................452-10 Cross-sectional schematic view of SWNT transistor device and IDS-VGS curve...............453-1 Images of micromachined Si TEM grid.............................................................................533-2 Process flow for micromachined Si TEM grids.................................................................543-3 HRTEM images of SWNTs taken using micromachined Si TEM grids...........................553-4 SEM images of nanotubes and AFM images of ion implanted catalyst and solutionbased catalyst................................................................................................................. ....56

PAGE 10

10 3-5 Micro-Raman spectra of SWNTs grown from ion implanted catalyst..............................574-1 Schematic of Fe ion implantation co ndition and depth profile of Fe atoms......................654-2 SEM images of silicon oxide nanowires............................................................................664-3 TEM images of silicon oxide nanowirese..........................................................................674-4 EDS spectrum of individual silicon oxide nanowire.........................................................684-5 AFM images of SiO2 sample surface.................................................................................694-6 Solid-liquid-solid (SLS) growth model of silicon oxide nanowires..................................705-1 Depth profile of Fe atoms in SiO2. and schematic of CVD setup......................................775-2 AFM images of Fe catalyst nanopa rticles formed after annealing....................................785-3 SEM image and XRD pattern of as-grown GaN nanowires..............................................795-4 HRTEM image, SAED pattern, and ED S spectrum of individual GaN nanowire............805-5 Vapor-liquid-solid (VLS) growth mechanism for GaN nanowires...................................816-1 AFM image and transmittan ce spectra of SWNT films.....................................................926-2 Schematic of ICP-RIE system...........................................................................................936-3 AFM image of etched SWNT film....................................................................................946-4 Text characters printed in SWNT film...............................................................................956-5 AFM image of a series of nanotube film lines...................................................................967-1 MSM photodetector.........................................................................................................1037-2 Process flow for MSM photodetect ors with SWNT film electrodes...............................1047-3 Dark current of SWNT film-GaAs MSM device.............................................................1057-4 Dark current of SWNT film-GaAs MSM device.............................................................106A-1 Mask layout for nanotube device fabrication...................................................................111A-2 Mask layout for single nanotube device..........................................................................112

PAGE 11

11 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy GROWTH, FABRICATION, AND CHARACTE RIZATION OF CARBON NANOTUBES, NANOTUBE FILMS, AND NANOWIRES By Yongho Choi December 2008 Chair: Ant Ural Major: Electrical and Computer Engineering We explored nucleating the gr owth of nanomaterials by the ion implantation technique. We present experimental evidence that si ngle-walled carbon nanot ubes (SWNT) and SiOX and GaN nanowires can be grown by Fe ion implantation into SiO2/Si substrates, subsequent annealing, and chemical vapor deposition (CVD) gr owth. Moreover, we show that there is a dose and energy window of ion im plantation in which SWNT and nanowire growth are observed for a given growth condition. For nanomaterial s growth, catalyst is usually spun on or dropdried from a liquid solution contai ning iron nanoparticles or deposit ed as solid thin film layers by evaporation or sputtering. However, it is no t possible to pattern the liquid solution-based catalyst into very small dimensions or the th in film catalyst into nonplanar three-dimensional (3D) device structures, such as the sidewalls of high aspect ratio trenches. By adopting the ion implantation technique for nucleating nanomaterials growth, this thesis opens up the possibility of controlling the origin of nanomaterials at th e nanometer scale and of integrating nanomaterials into nonplanar 3D device structures with pr ecise dose control. We also fabricate micromachined Si Transmission Electron Micros copy (TEM) grids for direct TEM, as well as Atomic Force Microscopy (AFM), Scanni ng Electron Microscopy (SEM), and Raman characterization of as-grown nanomaterials. As a result, these micromachined TEM grids offer

PAGE 12

12 fast, easy, and reliable structural characterizatio n. Furthermore these grids provide a low cost, mass producible, efficient, reliable, and versat ile platform for direct TEM, AFM, SEM, and Raman analysis of as-grown nanomaterials, elim inating the need for any post-processing growth. We also explored fabrication and character ization of single-walle d carbon nanotube films which are three-dimensional films of tens of na nometers thickness, consisting of an interwoven mesh of single-walled carbon nanotubes. We de monstrate, for the first time, patterning of SWNT films down to submicron lateral dimensions as small as 50 nm using e-beam lithography and inductively coupled plasma (ICP) etching. This simple a nd efficient top-down patterning capability developed could open up tremendous opportunities for integrating single-walled nanotube films into a wide range of electronic and optoelectronic devices. Furthermore, we fabricate and characterize the effect of de vice geometry on the dark current of metalsemiconductor-metal (MSM) photodetectors base d on SWNT film-GaAs Schottky contacts. We observed that dark currents of the MSM devi ces scale rationally with device geometry, such as the device active area, finger width, and finge r spacing. These results open up the possibility of integrating SWNT films as transparent and conductive Scho ttky electrodes in conventional semiconductor electronic and optoelectronic devices.

PAGE 13

13 CHAPTER 1 INTRODUCTION The continuous drive toward in creasingly smaller length scales has lead to many advances in science and technology over the past century. One of the most remarkable examples of this drive towards miniaturization is the microelect ronics industry, which has constantly relied on scaling to create denser, faster, and cheaper devices since the inve ntion of the first transistor. However, the scaling of silicon integrated circ uit technology is predicted to slow down and reach the end of the roadmap in the next few decades [1-4]. Instead of trying to push this topdown scaling paradigm further, an alternative is to use a bottom-up fabrication approach to achieve even smaller length scales. This bo ttom-up approach involve s using nanomaterials, which naturally have very small size, as the building blocks of new electrical, mechanical, chemical, and biological devices. Single-wal led carbon nanotubes (SWNTs) and semiconductor nanowires, which have attracted a significant amount of research a ttention in recent years, are promising nanoscale materials as building blocks for bottom-up assembly of nanoelectronics [5-27]. SWNTs have also remarkable physical a nd electronic properties, such as high mobility and current density. Based on these characteristic s, device applications such as sensors and transistors have been demonstr ated experimentally [28-30]. There are several growth met hods developed to synthesize SW NTs, namely arc-discharge, laser ablation and chemical vapo r deposition (CVD). However, controlled growth has remained a big challenge for manufacturing SWNTs for device applications. An essential component of the CVD growth process is the catalyst material placed on the substrate for nucleating the growth of carbon nanotubes. In this dissertation, we demonstrate that i on implantation, a wellestablished technique in silicon microfabrication, and subsequent annealing can be used as an alternative method to create cata lyst nanoparticles for single-wa lled carbon nanotube growth, as

PAGE 14

14 presented in Chapter 2. To characterize the structure of the as-grown SWNTs directly after CVD growth, a silicon based micromachined tr ansmission electron microscopy (TEM) grid is designed and fabricated, as described in Chapter 3. In Chapters 4 and 5, we further extend the use of ion implanted catalyst to nanowire growth. In Chapter 4, we show that SiOX nanowires can be grown from ion implanted catalyst by th e solid-liquid-solid (SLS) growth mechanism. In Chapter 5, we show the vapor-liquid-soild (VLS) growth of GaN nanowires from the ion implanted catalyst. An alternative way to overcome the manufactur ability problems of individual SWNTs is using single-walled carb on nanotube films, which are three-di mensional networks of interwoven SWNTs. In the nanotube film, individual variati ons in diameter and chirality of nanotubes are averaged, and this results in uniform physical and electronic properties. Furthermore, what makes SWNT films even more at tractive for device applications is that they are flexible, transparent, and conductive. A ny potential device application utilizing SWNT films [31-43] requires the capability to efficiently pattern th em. In Chapter 6, we present an efficient technique that we have deve loped to pattern SWNT films by photolithography or e-beam lithography, and subsequent O2 plasma etching using an indu ctively coupled plasma (ICP). Using the patterning technique developed and shown in Chapter 6, we fabricate MetalSemiconductor-Metal (MSM) photodetectors with SW NT film electrodes on GaAs substrates. In Chapter 7, we present the fabrication pr ocess of SWNT film-GaA s MSM photodetectors and the characterization of their dark current. Th e conclusions of this di ssertation and suggested future works are given in Chapter 8.

PAGE 15

15 Single-Walled Carbon Nanotubes Single-Walled Carbon Nanotube Ge ometrical and Electronic Structure A single walled carbon nanotube is a graphene sheet rolled in to a cylindrical shape of diameter ranging from 0.7 nm to 10 nm. Three types of SWNT structures namely, armchair, zigzag, and chiral, are possible depending on the chiral vector, Ch, which is defined below. Figure 1-1 shows the unrolled graphene sh eet and its honeycomb lattice. The vector OB is in the direction of the longitudinal nanotube axis, and the vector OA corresponds to a crosssection of the nanotube perpendicular to the nanotube axis. The chiral vector Ch and the the translation vector T can be defined by those two vectors, OA and OB respectively, and the nanotube can be formed when the point O is co nnected to A and the point B is connected to B The chiral vector Ch is defined as Ch = n a1 + m a2 = (n m), (1) where a1 and a2 are the real space unit vector s of graphene, as shown in Figure 1-1, and any point on the hexagonal graphene la ttice can be expressed by them. Three types of nanotubes can be classified by the chiral vector Ch. The zigzag nanotube is the case when Ch =(n 0), the armchair nanotube is the case when n=m, that is Ch =(n n), and all other combinations of n and m are chiral nanotubes. The diameter of the nanotube, dt, can be calculated by dt = L / = | Ch | / = h hC C / = a nm m n 2 2 / (2) where L is the circumference of the nanotube and a = 2.49 A is the lattice constant of the honeycomb graphene lattice. Fu rthermore, the chiral angle represents the angle between the chiral vector Ch and the vector a1, and it is in the range from 0 to 30. The chiral angle = 0 corresponds to zigzag nanotubes, 30 corresponds to armchair nanot ubes, and all other angles

PAGE 16

16 between 0 and 30 correspond to chiral nanotube s. The translational vector, T, is defined as the vector which is parallel to the nanotube axis and is normal to the chiral vector, Ch, as shown in Figure 1-1. Similar to the vector Ch, the vector T can be expressed in terms of the two basis vectors a1 and a2, as ) ,t (t t t2 1 2 1 2 1a a T (3) By using the relationship Ch 0 T, t1 and t2 can be written in terms of n and m as t1 = Rd n m 2 and t2 = Rd m n 2 (4) where dR is the greatest common divisor of (2 m + n ) and (2 n + m ). Introducing d as the greatest common divisor of n and m then we can get dR as dR = d if n m is not a multiple of 3 d = 3 d if n m is a multiple of 3 d (5) The electrical properties of the nanotube is determined by its geometrical structure. Figure 1-2 shows the reciprocal latti ce of graphene. The SWNT r eciprocal lattice vectors are K1 and K2. The reciprocal lattice vector K1 corresponds to the circumferential direction and K2 corresponds to the longitudinal direction. As a result, the follow ing relations between Ch, T, K1, and K2 exist: Ch 2 1K 0 1K T Ch.0 2K, 2 2K T (6) From these relations, we can get the expressions for K1 and K2 as K1 = ) ( 11 2 2 1b b t t N K2 = ) ( 12 1b b n m N (7) where b1 and b2 are the reciprocal lattice vectors of graphene given by

PAGE 17

17 b1 = a a 2 3 2 b2 = a a 2 3 2 (8) and N is the number of hexagons in a unit cell of the nanotube given by N = R Rd a L d nm n m2 2 2 22 ) ( 2 | | | | 2 1 ha a T C (9) Figure 1-3A shows the Brillouin zone of graphene and and are high symmetry points which correspond to the center, the corner, and the center of the edge, respectively. The energy dispersion relations for graphene throughout the whole region of the Brillouin zone are shown in Figure 1-3B. The allowed wavev ectors in the direction of K2 in Figure 1-2 are continuous for an infinitely long nanotube. In th e circumferential direction, however, there are N discrete allowed wavevectors. When the allowed wavevectors in the circumferential direction include the K point which is shown in Figure 1-3, the nanotube is metallic. In other words, the valence band and conduction bands meet each ot her. This corresponds to the condition, ( n m ) is a multiple of 3 (10) For all other cases, the nanotube is semiconducting. Furthermore, the bandgap of semiconducting nanotubes is inversely pr oportional to their diameter [7, 44]. Single-Walled Carbon Nanotube Growth Methods Several growth methods have been developed to synthesize carbon nanotubes, namely arc discharge, laser ablation, and chemical vapor deposition (CVD). In the arc discharge method, a voltage around 25 V is applied to two carbon rod electrodes with a separation ~1 mm. A high current flows between the two electrodes in a helium atmosp here. The carbon at the anode evaporates and condenses at the cathode and fo rms nanotubes. To synthesize SWNTs, transition metals such as Co, Ni, and Fe are used as catalyst material on the electrode. The as-grown SWNTs have few defects because of th e high temperature during growth (~3,000C) and their

PAGE 18

18 diameters are usually small and their diameter distribution is narrow. However, this method produces many by-products such as fullerenes and amorphous carbon, so a pur ification process is required after growth [6, 9]. The laser ablation method uses a laser to s ynthesize nanotubes. A pulsed laser vaporizes the graphite target insi de a tube furnace (~1,200C). To obtain SWNTs, transition metals are added to the graphite target. Ar gon or other inert gases flow to the tube to carry the grown nanotubes to the water-cooled c opper collector. This growth me thod produces many by-products similar to the arc discharge method. Nanotubes grown by both arc-discharge and laser ablation methods are usually bundled due to van der Walls forces. To get individu al nanotubes for device applications, post processing such as sonication is needed to break up the bundles [10, 45]. Chemical vapor deposition (CVD), however, does not suffer from these problems. The schematic and the picture of the CVD system are shown in Figure 1-4. CVD needs transition metals as catalysts to nucleate nanotubes. The tr ansition metals are typically dissolved in liquid solution and deposited on a sample. Then the sa mple is put inside a furnace and heated up to ~900C. The feedstock gases C2H4 and CH4 are flown at ~900C to supply the necessary carbon. The diameter of the as-grown nanotubes are typically determined by the catalyst size. Therefore, the catalyst condition is very impor tant during SWNT growth. The chemical vapor deposition method can produce pristin e and isolated nanotubes [12]. GaN Nanowires GaN-based III-V nitride semiconductor materi als have attracted significant research attention due to their unique properties, such as high breakdown field, large mobility, high drift velocity, good thermal conductivity, and good chemi cal and physical stability [30, 46, 47]. More importantly, AlGaInN family of III-V nitr ide semiconductor alloys have a direct energy

PAGE 19

19 bandgap, which varies between 6.2 eV and 1.95 eV depending on their composition, as shown in Figure 1-5. As a result, GaN-based III-V nitride semiconductors cover the wavelengths from the green into UV, and have applications in blue/green/UV light-em itting diodes (LEDs) in full-color displays, traffic lights, automotive ligh ting, and in white LEDs for room lighting [46, 48-61]. Furthermore, blue/green laser diodes [6 2-69] can be used in high storage-capacity DVDs. In addition, AlGaN-based photodetector s [70-76] can be us ed in solar-blind UV detection and have potential for a wi de range of applications, such as flame sensors for control of gas turbines and for detection of missiles. In addition to these well-established applica tions, there are also many newer applications of III-V nitride semiconductor materials. For example, UV LEDs or laser diodes based on IIIV nitrides can be employed as UV optical sources for use in airborne chemical and biological sensing systems, allowing direct multi-wavelength spectroscopic identification and monitoring of UV-induced reactions [48, 77]. The UV wave lengths are necessary in order to cause fluorescence in many of the targeted chemicals a nd biological agents. This would enable the manufacturing of compact biologi cal or chemical agent warning sensors with fast response and high detection sensitivity. Furthermore, simple Schottky diod e or field-effect transistor structures fabricated in GaN are sensitive to a number of gases, includi ng hydrogen and hydrocarbons. As a result, GaNbased III-V nitride materials can also be used as wide bandgap semiconductor sensors for gas detection. These detectors would have use in automobiles, aircraft, fire detectors, and in diagnosis of exhaust and emissions from industrial processes. The nitrides are also well-suited to high temperature applications because of th eir wide bandgaps and low intrinsic carrier concentrations. As a result, wide bandgap elec tronics and sensors based on GaN can operate at

PAGE 20

20 elevated temperatures (600 C, or 1112 F) where conventional silicon (Si)-based devices cannot function. The ability of these materials to f unction in high temperature, high power, and high energy radiation conditions will enable large performance enhancements in a wide variety of sensing applications where uncool ed operation is essential. Uncooled GaN electronics and sensors will reduce sensor package weights and co st and increase functiona l capabilities [78]. Given the high cost per weight of transporting sensors to remote environments, the weight savings gained by using wide ba ndgap devices could have large economic implications in the energy and space industries. Furthermore, III-V nitride materials have significant advantages for high power device applications due to their very high breakdown field, allowing the devices to support large voltages for high power operation [49, 79-81]. For example, they can be used in power amplifiers and monolithic microwave integrated ci rcuits (MMICs) as a part of high performance radar units and wireless broa dband communication links, and as ultra high power switches for the control of distribution on el ectricity grid networks. Although GaN and related III-V nitride material s in thin film form have been wellestablished for many of the appl ications listed above, they have some significant shortcomings: 1) It is very expensive to gr ow thin film GaN-based III-V nitride materials since it requires sophisticated growth equipment such as meta lorganic chemical vapor deposition (MOCVD) [8285] or molecular beam epitaxy (MBE) [86-89]. 2) It is very expensive to fabricate GaN-based III-V nitrid e devices since it requires advanced lithography and fabricat ion techniques to pattern thes e materials into very small nanoscale dimensions [90].

PAGE 21

21 3) The material quality of thin film GaN is still far-from-ideal. Most commercially available III-V nitride materials are currently grown heteroepitaxial ly on foreign substrates, such as sapphire and SiC, suffering from a high densit y of dislocations and microstructural defects due to lattice mismatch (See Figure 1-5. for the lattice consta nts) [61, 91, 92]. These defects act as leakage paths, nonradiative recombina tion centers, and impurity and metal migration channels, decreasing device performance. For example, the presence of a high density of dislocations and large residual st rain in GaN has been proven to be the limiting factor for LED efficiency and reliability [93]. Although small diameter free-standing bu lk GaN substrates are now commercially available, sapphire and SiC will likely remain the common substrates for nitride LEDs before large area low-cost GaN wa fers become available. Furthermore, the growth of defect-free III-V n itride semiconductor all oys, such as InGaN and AlGaN, present problems even on free-standing bulk GaN substrates, due to lattice mismatch. Consequently, the challenge for III-V nitride se miconductors is to find a cheap, large-scale, nonplanar, non-lithographic, and defect-free nanomanufacturing technique which would solve these shortcomings. GaN-based III-V nitrid e semiconductor nanowires offer a potential solution to this challenge. Semiconductor nanowires have re cently attracted much atten tion as promising candidates as building blocks for active electronic and photonic devices, such as photodetectors, LEDs, lasers, and biochemical sensors [27, 94-117]. Na nowires of III-V nitride materials do not suffer from the problems listed above for pl anar III-V nitride thin films. First of all, nanowires can be grown cheaply and in large-sc ale using a very simple ch emical vapor deposition (CVD) approach, as explained in detail in Chapter 5. Secondly, since these nanomaterials naturally have submicron, nanoscale dimensions, there is no need to do lithographic patterning and

PAGE 22

22 microfabrication after growth. Third, because nanowire synthesis is substrate-free, it should prevent formation of dislocations and defects originating from lattice mismatch [118]. This should enable the growth of def ect-free III-V nitride alloy nanowi res, such as InGaN and AlGaN.

PAGE 23

23 Figure 1-1. Unrolled graphene sheet making up SWNT and its honeycomb lattice. Ch is the chiral vector and T is the translational vector. Nanotube is formed when O is connected to A and B is connected to B This structure corresponds to a (4,2) nanotube. Real space unit vectors of graphene, a1 and a2, are shown in the inset on the right. Figure 1-2. Reciprocal latti ce of graphene (unrolled carbon nanotube). Reciprocal lattice vectors of SWNT are K1 and K2. The reciprocal lattice vector K1 corresponds to the circumferential direction and K2 corresponds to the longitudinal direction. 2K 1K 1b2b Ch O A B T B a1a2

PAGE 24

24 Figure 1-3. Energy dispersion relations for graphen. A) Br illouin zone of graphene. and are high symmetry points as the center the corner, and the center of the edge, respectively. B) Energy dispersion relations for graphene shown throughout whole region of the Brillouin zone. K M A B

PAGE 25

25 Figure 1-4. Images of CVD system. A) Photogr aph and B) Schematic of the CVD system used to grow SWNTs. The sample is placed in the 1 inch quartz tube and heated up to 900C and annealed for 2 min in Ar and H2 atmosphere. Feedstock gases C2H4 and CH4 are then flown at 900C to grow nanotubes on the sample.C2H4 or CH4 900oC Furnace A B

PAGE 26

26 Figure 1-5. Lattice constant of III-V nitride semiconductor materials as a function of their bandgap energy. Also shown for reference are the lattice consta nts of the commonly used substrates (SiC, MgAl2O4, and sapphire) for the growth of III-V nitride based materials (from Ref. [46]).

PAGE 27

27 CHAPTER 2 SINGLE-WALLED CARBON NANOTUBE GROW TH FROM ION IMPLANTED IRON CATALYST Introduction An essential component of the CVD growth proc ess is the catalyst ma terial placed on the substrate for nucleating the growth of car bon nanotubes. Typically, transition metal nanoparticles, such as nickel (Ni), iron (Fe), a nd cobalt (Co), are used as catalyst. For SWNT growth, catalyst is usually spun on or drop-dr ied from a liquid solution containing iron nanoparticles [11]. More recently, solid thin film layers depos ited by evaporation or sputtering have also been used as catalyst [119]. In or der to control the origin of nanotubes during CVD growth, the catalyst is typically patterned by lithography into sma ll islands [11]. However, it is not possible to pattern the liqui d solution-based catalyst into very small dimensions or the thin film catalyst into nonplanar three-dimensional (3D) device structures, such as the sidewalls of high aspect ratio trenches. In this chapter, we demonstrate that ion im plantation, a well-established technique in silicon microfabrication, and subseq uent annealing can be used as an alternative method to create catalyst nanoparticles. Ion-implanted catalyst is much easier to pattern into very small features and over high aspect ratio topography compared to other types of catalyst, and it offers extremely accurate control of the number of atoms introdu ced into the substrate (the dose). Since ion implantation is very reproducible, easily scal able, and compatible w ith standard silicon microfabrication, it could offer significant technol ogical advantages as a method to form catalyst nanoparticles. In this chapter, we first introduce ion impl antation and liquid-based catalysts for SWNT growth. We then present experimental ev idence that single-wa lled carbon nanotubes can indeed be produced by the process of Fe ion implantation into thermally grown SiO2 layers,

PAGE 28

28 subsequent annealing, and CVD growth. Finally, we systematically characterize the effect of implantation dose and energy on the structural properties of the catalyst nanoparticles and SWNTs that are formed. Ion Implantation Ion implantation is one of the well-establishe d techniques in silicon de vice fabrication. In the implantation process, ions or molecules are accelerated to hi gh energy and directly introduced into the substrate. The first benefit of this technique is the ability to control the depth of ions implanted into the substrate. To get i ons implanted deeper, the accelerating voltage can be increased. The second benefit is the ability to control the quantity of ions implanted. For higher number of ions, the implantation time can be increased. The number of implanted ions can be counted precisely. The ions implanted follow a random trajector y and get scattered by other atoms before losing their energy, which results in a Gaussian-like distribution. The average depth of the implanted i ons is called the projected range, Rp. The total number of ions implanted per unit area is called the dose. As an example, Figure 2-1 shows the distribution of several different ion species implanted into crystalline silicon at an energy of 200 kV. Liquid-Based Catalyst The most commonly used catalysts for th e CVD growth of SWNTs are liquid-based catalysts. As an example of such a catalyst, Figure 2-2A shows the AFM image of iron nitrate/IPA liquid catalyst after dispersion on a SiO2 substrate and Figure 2-2B shows the AFM image of the nanotubes grown from this cataly st by CVD. As mentioned previously, one significant problem with this catalyst is that it is not possible to pattern it into very small dimensions and over high aspect ratio topography.

PAGE 29

29 Experimental Procedure 500 nm thick SiO2 layers were first thermally grown on silicon (100) substrates. Fe+ ions were implanted into these layers at an en ergy of 60 keV with three different doses (1014, 1015, and 1016 cm-2) and at a dose of 1015 cm-2 with three different energies (25, 60, and 130 keV). The projected range Rp of the 25, 60, and 130 keV implants in SiO2 are 23.9, 49.9, and 103.4 nm, respectively, based on SRIM [120] calculations. The as-implanted samples were placed in a 1inch quartz tube furnace and annealed at 900C under 300 sccm Ar and 200 sccm H2 flow for 2 minutes to form the catalyst nanoparticles. (T he schematic and photograph of the furnace is shown in Figure 1-4). After annealing, nanot ubes were grown at th e same temperature by discontinuing the Ar flow and introducing CH4 at 200 sccm. The co-flow of H2 remained constant at 200 sccm, and the growth lasted for 10 min. The catal yst nanoparticles and nanotubes obtained were characterized by a Digital Instruments Nanoscope III Atomic Force Microscope (AFM), a JEOL 2010F High Reso lution Transmission Electron Microscope (HRTEM) operating at 100 kV, and a Renishaw Raman spectroscopy system. Results and Discussion AFM Characterization The AFM images of Figure 2-3A, B, and C show the Fe catalyst nanoparticles formed on the SiO2 surface from implants of three different doses at 60 keV, after annealing the asimplanted substrates at 900oC for 2 minutes. The average height and density of these nanoparticles are listed in Table 2-1. Figure 2-3 shows that the low (1014 cm-2) implant dose results in very low density of catalyst particles, whereas the high (1016 cm-2) implant dose results in very large particle size. For the medium (1015 cm-2) dose sample, on the other hand, both small height (~2 nm) and high density Fe nanoparticles are obtained. Using a si mple model, the flux of Fe atoms diffusing to

PAGE 30

30 the oxide surface under high temperature is propor tional to the dose of the implant. As a consequence too little flux (e.g. 1014 cm-2 dose sample) results in too few particles, and too much flux (e.g. 1016 cm-2 dose sample) results in the aggregation of Fe atoms into larger clusters. Furthermore, we have also studied the effect of implant en ergy at a dose of 1015 cm-2 on the catalyst particle size and density, as presented in Table 2-1 and Figure 24. The low (25 keV) implant energy results in a very hi gh density of catalyst particles with slightly larger average size than those obtained from the medium (60 keV) implant energy, whereas the high (130 keV) implant energy results in very larg e particle size and low density. The AFM images of Figure 2-3D,E, and F show the carbon nanotubes grown by CVD at 900oC from the catalyst nanoparticles implanted with 60 keV and 1014, 1015 and 1016 cm-2 dose, respectively. As seen in Figure 2-3D, no nanotubes were grown on the low dose sample, consistent with the very low density of catal yst nanoparticles on that sample. The high and medium dose samples, on the other hand, produced nanotubes with a similar average diameter (2.2 nm), although the density of nanotubes was mu ch less for the high dose sample compared to the medium dose one (See Figure 2-3 and Table 2-1). This is due to the presence of a few small (~2 nm) catalyst nanoparticles on th e high dose sample, despite the f act that the average catalyst size is much larger (15 nm). We have also studied the effect of implant energy at a dose of 1015 cm-2 on the nanotube diameter and density. As shown in Figure 24, the high implant ener gy did not produce any nanotubes due to the very large particle size, and the low implant energy resulted in nanotubes which have a slightly larger average diamet er (~3 nm) than the medium energy sample, consistent with the slightly larger catalyst si ze [121]. We can define th e nanotube yield as the number of nanotubes grown per unit area divide d by the number of cat alyst nanoparticles per

PAGE 31

31 unit area. By this definition, the yield for the 60 keV and 25 keV implants are 0.035 and 0.009, respectively. The yield for the 60 keV implant is about 4 times higher than the 25 keV implant. This could be due to two reasons: The first reas on could be that the nano tube yield is a function of catalyst size, and the yield decreases as the average catalyst height increases. Previous studies have observed that smaller Fe catalyst nanoparticles tend to be more active in producing single-walled carbon nanotubes (SWNTs) compared to larger ones [122, 123]. They suggest that this could be due to small nanoparticles allowing carbon supersatur ation and facilitating nanotube growth more readily. The larger average cat alyst nanoparticle he ight could partly explain why the yield is lower for the 25 keV implant. Secondly, the catalyst nanoparticle height and density as shown in Table 2-1 is measured after the 2 min anneal at 900C, but before the nanotube growth step. During the 10 min nanotube growth step (also at 900 C), the implanted Fe atoms con tinue to out-diffuse towards the surface, and as a result, the catalyst nanoparticle density and height continue to evolve as the growth proceeds. This is in contrast to solu tion-based catalyst, where the catalyst dose on the surface does not change during growth. To s ee how much the catalyst nanoparticle density changes during the 10 min nanotube growth, we ha ve measured by AFM, the average density of catalyst nanoparticles after gr owth for both the 25 keV and 60 keV implants, as shown in Table 2-2. We can clearly see from Table 2-2 that the catalyst de nsity decreases for the 25 keV implant, whereas it increases for the 60 keV implant as the growth proceeds, becoming comparable at the end of the growth step. The density increase for the 60 keV case could be explained by more Fe out-diffusing to the oxid e surface. The density decrease in the 25 keV

PAGE 32

32 case can be explained by Fe nanopa rticles aggregating into larger and fewer clusters due to a high Fe surface concentration resulting from out-diffusion. If we calculate the yield based on the nanot ube density after growth, we get 0.015 and 0.013 for the 60 keV and 25 keV implants, respec tively, resulting in yield values which are almost identical within experimental error. Al ong with catalyst height dependent yield, the continued out-diffusion of Fe during the na notube growth could explain the observed discrepancy in yield between the 25 and 60 keV im plants. These experimental findings suggest that a lower implant energy does not necessa rily result in a higher nanotube density. TEM Characterization In order to verify by HRTEM that the nanot ubes grown from ion implanted catalyst are single-walled, we have fabricat ed special micromachined silicon substrates with narrow open slits on them, as shown in Figure 2-5A and B.[124 -126] The fabrication of the TEM grid will be discussed in detail in Chapter 3. Briefly, 350 nm oxide was grown on both the frond and the back side of a (100) Si wafe r and thin membranes of 1 mm2 area were wet-etched using a buffered oxide etch (BOE) solution. 1 mm2 windows etched from the back side silicon using a TMAH solution until ~15 m thick silicon membranes remain. Narrow open slits of width ranging from 1.5 to 4.5 m are dry etched on the top side oxide and open slits of the same dimension are dry etched in silicon from the top using a deep silicon etcher. Following the etch, the wafer is implanted with 60 keV, 1015 cm-2 Fe+ ions. After dicing the wafer, the resulting SiO2/Si substrates are annealed and grown by CVD, following the same procedure given above. The nanotubes grown on these substrates are comp letely suspended over the width of the slits, enabling direct TEM characterization of as-gro wn SWNTs from ion implanted catalyst. The HRTEM images we have obtained show clear ev idence that the as-grown nanotubes are single-

PAGE 33

33 walled with a diameter range in agreement w ith the AFM data [See Figure 2-5C and D]. The reason that only SWNTs are obtained is due to th e particular growth condition used. Previous work in the literature by severa l groups has also obtained only SWNTs using growth conditions similar to the one used in this work [121-123]. For example, Liu et al. have grown nanotubes (using monodispersed Fe-Mo nanopart icles as catalyst) by CVD at 900oC under hydrogen and methane flow. SWNT growth was observed when the average catalyst size was 3.6 nm, but no growth occurred when the average catalyst size wa s 8.5 nm [123]. Furthermore, using identical growth conditions as in our case (i.e. 900C, 200 sccm meth ane and 200 sccm hydrogen) and discrete iron oxide nanoparticles as catalyst, Dai et al. found that no nanotubes (including multiwalled) were produced for large na noparticles (> 5 6 nm) [122]. In addition, Lieber et al. has found that a low supply of carbon precursor gas produces only small diameter SWNTs even when the catalyst also contains large iron nanoparticle s [121]. Furthermore, theoretical work by Kanzow et al. has suggested that for MWNT grow th, one requires large catalyst particle size, high carbon supply, and lower temperatures [ 127]. The condition of high carbon supply and lower temperatures is not met in our growth cond ition. In fact, the growth conditions used in this work were chosen based on the literature and our previous experiments with solution-based Fe catalyst in order to be able to obtain onl y SWNTs. In short, the fact that multi-walled nanotubes and larger single-walled nanotubes are not formed in our growth can be explained by the particular growth condition used and this obser vation is consistent with previous work in the literature using similar growth conditions. Raman Spectroscopy Characterization and Discussion We have also characterized the nanotubes grown from the ion implanted catalyst (60 keV,1015 cm-2 Fe+) using micro-Raman spectroscopy, as shown in Figure 2-6. The Raman spectra were obtained by a Kr ion laser with an excitation wavelength of 647.1 nm and a spot

PAGE 34

34 size of about 1 m. The observation of characteristic multi-peak tangential mode (G band) features around 1580 cm-1 provides a signature of single-walle d nanotubes [128]. A Lorentzian fit of the G-band spectrum in Figure 2-6 yields peaks at 1570, 1579, and 1593 cm-1. Furthermore, a Radial Breathi ng Mode (RBM) peak at 200 cm-1 is observed, as shown in the inset of Figure 2-6, which corresponds to a SWNT diameter of around 1.2 nm [128], falling within the diameter range of 0.8-3.5 nm obtained by AFM. Raman measurements performed at a fixed laser energy give an idea of nanotubes that are in resonance with the particular laser line used, but do not give a complete characterization of the diameter distribution in a sample, since resonant Raman experiments only detect nanot ubes whose electronic en ergy spacings between van Hove singularities match the laser excitation energy [128, 129]. AFM and HRTEM analysis, on the other hand, give a more complete characterization of the di ameter distribution of nanotubes in a sample. Nonetheless, Raman data provides an independent confirmation of the existence of SWNTs in a sample and remains a very useful analyti cal tool for nanotube characterization. The radial breathing mode (RBM) Raman features, typically appearing between 120-250 cm-1 for SWNTs with diameters between 1-2 nm correspond to the atomic vibration of the carbon atoms in the radial direction. For isolated SWNTs on an oxidized silicon subs trate, the following relationship has been found between the RBM peak frequency RBM and the SWNT diameter dt, RBM = 248 / dt, (11) where RBM is in units of cm-1 and dt is in units of nm [128, 129]. However, for dt < 1, this relationship is not expected to hold due to nanot ube lattice distortions leading to a chirality dependence of RBM and for large diameter SWNTs with dt > 2nm, the intensity of the RBM feature is weak and is hardly observable [128, 129]. As a result, diameter characterization

PAGE 35

35 using the RBM features is typically limited to 1< dt <2 nm. Furthermore, as mentioned above, an observable Raman signal from a SWNT can on ly be obtained when the energy of the laser line used is equal to the ener gy separation between van Hove si ngularities in the valence and conduction bands [128, 129]. As a result, for a given laser energy, Raman analysis can only observe SWNTs with certain diameters. A plot of the optically allowed transition energies versus the nanotube diameter, known as the Kataur a plot [128-130], is shown in Figure 2-7. By drawing a line at Eii = 1.92 eV in Figure 2-7 (which corresponds to the = 647.1 wavelength Kr ion laser that we have used), we can find out which nanotube diam eters are in resonance with our laser. In the diameter range 1< dt <2 nm, only diameters between approximately 1.2-1.35 nm and 1.6-1.8 nm would be observable with the particul ar laser energy we are using. The inset of Figure 2-6 shows a peak at RBM =200 cm-1, which using the equation (12) corresponds to 2 1 td nm. This is indeed in the diameter range 1.2-1.35 nm observable by our laser energy. We found during Raman characterization that in most cases the RBM features were buried under the noise. As a result, we could only char acterize the diameters of a few nanotubes using the RBM Raman features. (In contrast, we ha ve characterized the diameters of about 200 nanotubes by AFM.) In the other few cases, we have also observed RBM peaks at frequencies corresponding to diameters of ar ound 1.2-1.3 nm. As a result, the significance of the 1.2 nm diameter value is not that it is the modal value, but it is the value detectable by our Raman laser energy. The Raman RBM data we have obtaine d is not sufficient to give a statistical characterization of the nanotube diameters, bu t provides an independent confirmation of the existence of SWNTs in our samples. For the 60keV/1015cm-2 implant condition, the diameter range of nanotubes measured by AFM was found to be between 0.8 and 3.5 nm, with a mean of 2.2 nm (as tabulated in Table 2-1) and a standard deviation of 0.6 nm. As a result, the 1.2 nm

PAGE 36

36 nanotubes observed by Raman RBM features fall wi thin the diameter range observed by AFM. Although the nanotubes shown in the HRTEM images of Figure 2-5 have diameters around 2-2.5 nm, we have also observed smaller and la rger diameter SWNTs by HRTEM. Our HRTEM characterization yielded a diameter range similar to the one obtained by AFM (i.e. 0.8-3.5 nm). Electrical Characterization In addition to structural ch aracterization, we have also measured the electrical characteristics of the as-grown nanotubes from ionimplanted catalyst. For this purpose, the asgrown nanotube sample (implantation condition of 1015 cm-2 dose and 1016 keV energy) was patterned by a lithographic process and Au was ev aporated on the sample to form electrodes. Figure 2-8 shows the cross-sectiona l schematic of the process flow for SWNT device fabrication. The as-grown sample has nanotubes random ly distributed on thermally grown SiO2 on a silicon substrate. The next stage is the spin coating of the lift-off resist (LOR) (Microchem LOR3B), and the photoresist (Shipley S1813) on the sample as shown in Figure 2-8B. Using LOR prevents sidewall coverage and ai ds in the lift-off process. Th e next stage is to expose the sample by ultraviolet light thr ough a mask. The mask was designe d to have several different sizes of electrode widths and sp acings in order to increase the possibility of having individual nanotube connections between the electrodes. The mask layout is shown in Appendix A. Briefly, the mask has four different electrode widths (7, 14, 28 and 100 m) and each width has four different electrode spacings (2, 4, 6, and 10 m). It has about 100 devices in ~1 cm2 area. After exposure, the sample is immersed in th e developer (Microchem MF-319) and the exposed photoresist and LOR under that exposed region is removed (Figure 2-8C). The next step is the Au deposition on the sample by e-beam evaporation, as shown in Figure 2-8D. After the deposition, the sample is covered with 100 nm of Au. The final patterni ng step is the lift-off

PAGE 37

37 process. The sample is immersed in the PG re mover (Microchem) and so nicated briefly at low power. Figure 2-8E shows the resulting structure. Figure 2-9 is an optical microscope image showing the top view of the fabricated sample. The width of electrode is 14 m and the spacing between two electrodes is 6 m, in this case. Figure 2-10 shows the cross-sectional schematic of the nanotube device and its IDS vs. VGS curve. The silicon substrate plays the role of a back gate and the drain voltage is applied to the Au electrode on one end of the nanotube, and the Au electrode on the other end (source) is grounded. As the IDS vs. VGS data shows, the device functions as a p -type transistor. Conclusions In conclusion, we have shown experimental evidence that single-walled carbon nanotubes can be grown by Fe ion implantation into SiO2/Si substrates, subseque nt annealing, and CVD growth. For a given growth condition, there is a dose and energy window in which nanotube growth is observed. This work opens up the possib ility of controlling the origin of SWNTs at the nanometer scale using advanced lithography tec hniques, such as electron-beam lithography, and of integrating nanotubes in to nonplanar 3D device structures with precise dose control. Ion implantation could offer significant technological advantages as a method to form catalyst nanoparticles for a wide range of nanoscale device applications.

PAGE 38

38 Figure 2-1. Distribution profiles of various ions in cr ystalline silicon after implantation at an energy of 200 keV. Heavy ions travel sh allower and have a narrower distribution than light ions. RP is the standard deviation and CP is the peak concentration of the implant. (Adapted from Ref. [131]) Figure 2-2. AFM image of A) iron nitrat e/IPA liquid catalyst dispersed on a SiO2 substrate and B) the SWNTs grown from the catalyst of part A. Catalyst Nanotube 500 nm 500 nm B A

PAGE 39

39 Figure 2-3. AFM images of A-C) Fe catalyst nanopartic les formed after annealing as-implanted SiO2 substrates, and D-F) Carbon nanot ubes grown by CVD from the catalyst nanoparticles of A-C, respectively. The Fe+ implant doses are A, D 1014, B, E 1015, and C, F 1016 cm-2 at an energy of 60 keV. A B D C F E

PAGE 40

40 Figure 2-4. AFM images of A-C) Fe catalyst na noparticles formed after annealing as-implanted SiO2 substrates, and D-F) Carbon nanot ubes grown by CVD from the catalyst nanoparticles of A-C, respectively. The Fe+ implant energy are A, D 25, B, E 60, and C, F 130 keV with a dose of 1015 cm-2. C D E F B A

PAGE 41

41 Figure 2-5. Images of as-grown SWNTs. A) Cr oss-sectional schematic (not to scale) and B) top view optical microscope image of the micromachined Si TEM grids that we have fabricated (not to scale). C), D) HR TEM images obtained from the 60 keV, 1015 cm-2 Fe+ implanted micromachined TEM grids, showing SWNTs. All of the 85 nanotubes imaged by HRTEM were singlewalled, and no multi-walled nanotubes were observed. A B D C

PAGE 42

42 Figure 2-6. Tangential mode (G band) micro Raman spectra of nanotubes grown from 60 keV, 1015 cm-2 Fe+ implanted catalyst. A Lorentzian fit of the G-band spectrum yields peaks at 1570, 1579, and 1593 cm-1. Inset shows Radial Breathing Mode (RBM) Raman spectrum with a peak at 200 cm-1, which corresponds to a SWNT diameter of around 1.2 nm [128].

PAGE 43

43 Figure 2-7. Kataura plot showing the electronic transition energies for all possible SWNTs as a function of diameter. Each data point deno tes a particular SWNT. The red line is drawn at the laser energy we have used in our Raman experiments. (Adapted from Ref. [128]).

PAGE 44

44 Figure 2-8. Cross-sectional schematic of proces s flow for SWNT transist or. Figures are not to scale. A) As-grown nanotube sample wh ich has nanotubes randomly distributed on thermally grown SiO2. Nanotubes are not shown in the figure. B) Spin coating of the sample with the lift-off resist (LOR) and the photoresist (PR). C) Sample is exposed by ultraviolet light through mask a nd immersed in developer and the exposed photoresist and LOR under that region is removed. D) Au deposition by e-beam evaporation. 100 nm of Au covers the whole surface of the sample. E) Sample is immersed in PG remover and sonicated brie fly at low power. The metal on top of the resist is removed. A B D E C

PAGE 45

45 Figure 2-9. Top view optical microscope imag e of the fabricated na notube device sample. The yellow area is Au electrode s and the gray area is SiO2. The SWNTs are randomly distributed and the number of nanotube connections for each device is random. Figure 2-10. Electrical characte rization of SWNT transistor. A) Cross-sectional schematic view of the SWNT transistor device. Back gate voltage is applied to the silicon substrate which is below the nanotube channel. B) IDSVGS curve of the SWNT transistor device and showing p -type behavior. 200 m A B

PAGE 46

46 Table 2-1. Average height and density of catalyst na noparticles formed after annealing and the average diameter and density of nanotubes grown for each Fe ion implantation condition. Fe+ ion implantation energy/dose (keV/cm-2) Catalyst nanoparticle height (nm) Catalyst nanoparticle density (m-2) Nanotube diameter (nm) Nanotube density (m-2) 60 / 10142.2< 160 / 10152.01702.2 6 60 / 10161542.2 1 25 / 10152.78002.8 7 130 / 1015216Note: Between 60-400 catalyst nanoparticles and 130-200 nanotubes from several different samples were characterized to obtain the va lues listed for each implantation condition. Table 2-2 Density of catalyst nanoparticles fo rmed after growth for each Fe ion implantation condition. Catalyst nanoparticle density (m-2) Fe+ ion implantation energy/dose (keV/cm-2) Before growth After growth 25 / 1015 800525 60 / 1015 170400

PAGE 47

47 CHAPTER 3 MICROMACHINED SILICON TRANSMISSION ELECTRON MICROSCOPY GRIDS FOR DIRECT CHARACTERIZATION OF AS-G ROWN NANOTUBES AND NANOWIRES Introduction Structural characterization of carbon nanotubes is typically performed by atomic force microscopy (AFM), scanning electron micros copy (SEM), micro-Raman spectroscopy, and transmission electron microscopy (TEM). AF M and SEM can only image the outer wall of a nanotube lying on a substrate. As a result, th ey cannot determine whether a carbon nanotube is single-walled or multi-walled, or give informa tion about the number of walls of a multi-walled carbon nanotube (MWNT). Furthermore, there ha s been recent interest in nanotube peapods, which are single-walled carbon nanotubes filled with C60 [132, 133]. Structural characterization of peapods is not possible with AFM and SEM, since they are incapable of determining whether a nanotube is filled or not [132]. High resolution TEM (HRTEM), on the other hand, is ideal for such nanostructure ch aracterization. However, TEM requires very thin samples, which are electron tran sparent. The best images are obtained if the nanotubes are suspended freely over a gap. Therefore, for TEM analysis, nanotubes are typically sonicated off the substrate in a solv ent (such as methanol, N,N-dimethylformamide (DMF), or 1,2-dichloroethane) and a few drops of the resulting suspension is deposited on commercially available holey-ca rbon TEM grids [12, 134]. This procedure has two potential problems: It can damage the nanotubes or alter their structural properties, and it does not work efficiently and reliably if the nanotube density is too low. This becomes a problem particularly for single-walled carbon nanotubes grown by ch emical vapor deposition (CVD) at high temperatures on bulk substrates, such as SiO2/Si. Typically, the dens ity of such nanotubes, used mainly for nanoelectronic device applicat ions [25, 26], is very low. As a result, characterization and optimization of CVD gr owth conditions becomes a tedious task.

PAGE 48

48 The ability to do TEM directly on as-grown nanotubes lying on sili con substrates would solve these problems. For this purpose, in th is chapter, we fabricate micromachined Si TEM grids with narrow open slits on them, and demonstrat e that they can be used for direct TEM, as well as AFM, SEM, and Raman characterization of as-grown nanotubes. The application of these grids is not limited to nanotubes, and can easily be extended to a wide range of nanomaterials, such as peapods, nanowires, na nofibers, and nanoribbons. Furthermore, these substrates are compatible with further microfab rication processes, such as lithography and thin film deposition. As a result, these micromach ined Si substrates provide a low cost, mass producible, efficient, and reliable platform for direct structural characterization of as-grown nanotubes, nanowires, and other nanomaterials. In recent work, nanotube-based nanoel ectronic devices compatible with TEM characterization have been presented. For exam ple, devices attached to metal electrodes after cleaving and etching a Si wafer [135] have been fabricated. In addition, devices have been suspended over windows etched in self-suppor ting silicon nitride membranes by e-beam lithography [132] or by a focused ion beam (FIB) [136, 137], or over windows etched in a silicon-on-insulator (SOI) wafe r [138]. The micromachined Si TEM grids fabricated in this chapter offer a simpler, faster, and more effi cient method for nanomater ials characterization. Design of the TEM Grids In our experimental design, each Si TEM grid occupies an area of 2 mm by 2.6 mm on a ~500 m thick Si substrate, and the open slits are located in the cen ter 1 mm by 1 mm windows where the Si substrate is thinned dow n to a membrane of approximately 15 m, as shown in Figure 3-1B. The size of the grid was chosen su ch that it fits into a standard TEM sample holder, as shown in Figure 3-1A. Each TEM grid contains 1440 slits with widths of 1.5, 2.5, 3.5

PAGE 49

49 or 4.5 m and lengths of 50 m [see Figure 3-1C and D]. On a 4 inch silicon wafer, 743 such Si TEM grids are fabricated. As a result, taking advantage of standard silicon microfabrication, these grids can be easily mass produced at a lo w cost in a massively parallel fashion. Microfabrication of the TEM Grids To fabricate the TEM grids, a 350 nm oxide was first thermally grown on both the front and the back side of a ~500 m thick 4-inch (100) Si wafer, as shown in Figure 3-2A. Following the oxidation, a 1.6 m thick Shipley SPR 3612 photores ist was spun on both the back and front side of the wafer. Photolithography was then performed on the back side. After exposure and development, 1 mm by 1 mm windows were etched on the back side oxide using a buffered oxide etch (BOE) solution until the windows were hydrophobic [see Figure 3-2B]. After stripping the photoresist, 1 mm by 1 mm windows were etched from the back side Si using a 25% tetramethylammonium hydr oxide (TMAH) solution in H2O in a beaker at 95oC until approximately 15 m thick Si membranes remained. The etch rate of the TMAH solution is approximately 60 m/hr, and the etch end point was dete rmined by periodic visual inspection [Figure 3-2C]. After the TMAH etch, top side lithography was performed aligned to the back side membranes using a 3 m thick Shipley SPR 220-3 photoresist. After exposure and development, slits of width ranging from 1.5 to 4.5 m and lengths of 50 m were dry etched on the top side oxide using an A pplied Materials P5000 magnetically -enhanced reactive ion etcher (MERIE) with a 35 sccm CHF3, 15 sccm CF4, and 100 sccm Ar chemistry at 200 mTorr pressure, 420 W rf power, and 60 Ga uss magnetic field [Figure 3-2D]. Following the oxide etch, slits of the same dimension were dry etched in the silicon membranes from the top using an STS Multiplex Inductively Coupled Plasma (ICP) Deep Reactive Ion Etcher (RIE) until punchthrough. After resist strip, the wafers were then coated with more phot oresist to protect the surface and

PAGE 50

50 diced into 2 mm by 2.6 mm grids using a wafersaw The final cross-se ctional structure is shown in Figure 3-2E. Nanotube Growth on the TEM Grids After the TEM grids were fabricated, na notubes were deposited on them using three different methods, to illustrate th e versatility of these grids for different applications. In the first case, commercial SWNT powder purchased from Iljin Nanotech Co. Ltd. (Seoul, Korea) was dispersed in DMF and spin coated on these TEM grids. In the second case, SWNTs were grown by CVD using a 20 g/mL solution of iron(III) nitr ate nonahydrate in 2-propanol as catalyst [24]. The grids were then placed in a 1-inch quartz tube furnace and grown at 940C under 1,000 sccm CH4, 20 sccm C2H4, and 500 sccm H2 flow for 10 minutes. In the third case, the Si TEM grids were ion implanted with Fe+ at an energy of 60 keV and a dose of 1015 cm-2. The as-implanted grids were again placed in a 1inch quartz tube furnace and heated up to 900C under 300 sccm Ar and 200 sccm H2 flow, which lasted 11 min. When the temperature reached 900C, the grids were annealed under the same argon and hydrogen flow rates for 2 min. As demonstrated in Chapter 2, this annealing step forms ~2 nm average size catalyst nanoparticles on the SiO2 surface, and these nanoparticles later act as catalyst for SWNT growth [139-141]. After annealing, nanotubes were grown at th e same temperature, without removing the ion implanted TEM grids from the tube furnace, under 200 sccm CH4 and 200 sccm H2 flow for 10 min. Characterization of Nanotubes Grown on the TEM Grids SWNTs obtained were characterized by a Dig ital Instruments Nanoscope III Atomic Force Microscope (AFM) operated in the tapping mo de, a JEOL 2010F high re solution transmission electron microscope (HRTEM) operating at 100 kV, a JEOL JSM-6335F field emission gun

PAGE 51

51 scanning electron microscope (FEG-SEM) ope rating at 15 kV, and a Renishaw Raman spectroscopy system. Figure 3-3A, B and C, and D and F show HRTEM images of SWNTs deposited from the commercial powder, grown by CVD using the solu tion-based catalyst, an d grown by CVD using the ion implanted catalyst, respectively. All images were taken using the micromachined Si TEM grids. Note that very high quality images of single-walled carbon nanotubes of a few nm diameter are very easily obtained using these grid s. Furthermore, since there are many slits on the TEM grid, and all slits have multiple nanotubes suspended across them, these micromachined grids reduce the TEM characterization time significantly. Figure 3-4A and B show SEM images of nanotubes freely suspended across the narrow slits of the micromachined TEM grids. Thes e nanotubes were grown by CVD using the ion implanted catalyst, and the images were taken fr om the same sample that was used for the TEM images of Figure 3-3D, E, and F. Furthermore, Figure 3-4C and D show AFM images of SWNTs on the TEM grid samples CVD-grown from the ion implanted catalyst and the solutionbased catalyst, respectively. AFM imaging wa s performed on the areas of the grids which do not have the slits. The grids are strong enough to support AFM imaging. These SEM and AFM results clearly show that TEM, AFM, and SE M characterization can be performed directly on the same as-grown sample. This is not possible with commercial TEM grids. We have also characterized the SWNTs grown from the ion implanted catalyst on these grids by micro-Raman spectroscopy. The Raman sp ectra were obtained by a Kr ion laser with an excitation wavelength of 647.1 nm and a spot size of about 1 m. In order to increase the signal to noise ratio, the Raman spectra were accumulated many times over the same SWNT. Figure 3-5 shows micro-Raman spectra of SWNT s suspended freely over the slits on the TEM

PAGE 52

52 grid. Characteristic multi-peak features around 1580 cm-1 (called the tangential mode or the G band) are observed [128, 129], as shown in Figure 3-5A. Figure 3-5B shows the radial breathing mode (RBM) spectra characteristic of SWNTs obtained from the same nanotube. The RBM peak occurring at 189 cm-1 corresponds to a SWNT diameter of 1.31 nm [128, 129]. The peak at 310 cm-1 comes from the Si substrate. Since nanotubes were freely suspended, the background signal from the substrate was mini mized, and the signal-to-noise ratio was significantly improved. As a result, in addi tion to TEM, the micromachined Si grids also improve the characterization of car bon nanotubes by Raman spectroscopy. Conclusions In conclusion, we have fabricated micromach ined Si TEM grids with narrow open slits on them, and demonstrated that they can be us ed for direct TEM, as well as AFM, SEM, and Raman characterization of as-grown nanotubes. As a result, these micromachined TEM grids offer fast, easy, and reliable structural ch aracterization of as-grown carbon nanotubes. Although we present data here only on SWNT charac terization, the application of these grids is not limited to nanotubes, and can be used for char acterizing a wide range of nanomaterials, such as peapods, nanowires, nanofibers, and nanoribbons Furthermore, depending on the application, smaller or larger slit widths can be designed, or other dielectric material s, besides oxide can be used. As a result, these substrates provide a low cost, mass producible, and efficient, reliable, versatile platform for direct TEM, AFM, SEM, a nd Raman analysis of as-grown nanomaterials, eliminating the need for any pos t-processing after CVD growth.

PAGE 53

53 Figure 3-1. Images of fabricated Si TEM grid. A) Digital photograph of TEM sample holder with micromachined Si TEM grid clamped in the groove on the right. The size of the grid was designed such that it fits easil y into a standard TEM sample holder. B) Top view schematic of the micromachined Si TEM grid. It occupies a 2 mm by 2.6 mm area with a thickness of ~500 m. The open slits are located in the center 1 mm by 1 mm area, where the Si substrate is thinned down to ~15 m to form a membrane. This center area is divided into 4 quadrants, each labeled by a number (1.5, 2.5, 3.5 or 4.5) corresponding to the widt h of the narrow open slits in that quadrant in units of m. C) Optical microscope image of the center area of the TEM grid in B, showing the 4 quadrants with the narrow open slits. Each TEM grid contains 1440 slits, and a four inch wafer contains 743 such grids. D) Close up optical microscope image of an array of slits. D C B A

PAGE 54

54 Figure 3-2. Cross-sectional schematic of process flow for the micromachined Si TEM grids. Some intermediate steps are not sh own. Figures are not to scale. A) Thermally grown 350 nm oxide on both the front and the back si de of a (100) Si wafer. B) 1 mm by 1 mm windows etched on the back side oxide using a buffered oxide etch (BOE) solution. C) 1 mm by 1 mm windows etched from the back side Si using a TMAH solution until ~15 m thick Si membranes remain. TMAH etches the 100 planes faster than the 111 planes. As a result, the side walls formed during the etch are not vertical as sketched in the figure, but sloped. D) Narrow open slits of width ranging from 1.5 to 4.5 m dry etched on the top side oxide. E) Open slits of the same dimension dry etched in silicon from the top using a deep silicon etcher. E D C B A

PAGE 55

55 Figure 3-3. HRTEM images of SWNTs taken using the micromachined Si TEM grids. A) SWNTs deposited from commercial powder B and C) SWNTs grown by CVD using the solution-based catalyst, and D-F) SW NTs grown by CVD using the ion implanted catalyst. The details of the catalyst and growth conditions are given in the text. Since there are many slits on the TEM grid, and all slits have multiple nanotubes suspended across them, these micromachined grids reduce the TEM characterization time significantly. F E D C B A

PAGE 56

56 Figure 3-4. Images of SWNTs grown on micromachined Si TEM grid. A and B) SEM images of nanotubes suspended across the slits of the micromachined TEM grids. These nanotubes were grown by CVD using the ion implanted catalyst, and the images were from the same sample as the TEM images of Figure 3-3 D-F). Nanotubes suspended across the slits is clear ly observed. C and D) show from ion implanted catalyst and solution-based catal yst, respectively. AFM imaging was AFM images of SWNTs on the solid parts of the TEM grid samples for CVD grown nanotubes performed on the solid areas of the grid which do not have the slits. D B C A

PAGE 57

57 Figure 3-5. Micro-Raman spectra of SWNTs grown from the ion implanted catalyst, suspended freely over the slits on the TEM grid. A) Tangential mode (G band) spectra of nanotubes showing multi-peaks characteristic of a metallic SWNT [128, 129] at 1553 and 1591 cm-1. B) Radial Breathing Mode (RBM) Raman spectra with a peak at 189 cm-1, which corresponds to a SWNT diameter of 1.31 nm [128, 129]. The signal-tonoise ratio was significan tly reduced by performing Raman on suspended nanotubes. A B

PAGE 58

58 CHAPTER 4 SILICON OXIDE NANOWIRE GROWTH FROM IRON ION IMPLANTED SIO2 SUBSTRATES Introduction As mentioned in the introducti on, one-dimensional (1D) nano structures, such as nanotubes and nanowires, have attracted significant research attention in recent years due to their unique structural and electronic properties. SiO2 is a material which is of great technological importance in silicon VLSI technology. Nanowires of silicon oxide have a great potential in applications such as low dimens ional waveguides, scanning near-f ield optical microscopy, blue light emitters, nanoscale optical devices and sensor s, sacrificial templates, and biosensors [142144]. Several methods have been used to grow sili con oxide nanowires, su ch as laser ablation [143], thermal evaporation [145], and chemical va por deposition (CVD) [146]. In most of these cases, a growth model based on the vapor-liquidsolid (VLS) growth mechanism [147] has been used to explain the observed results. An esse ntial component of the VLS growth process is the nanoscale catalyst particles require d to nucleate the growth of na nowires. For example, several recent studies have demonstrated the growth of silicon oxide nanowires from a variety of different catalyst nanoparticles, in cluding sputtered or evaporated metal thin films (such as Pt [142, 148], Au [149], Pd [149], and Ni [150]), molten Ga [146, 151-153], and Sn powder [154, 155]. In some of these studies, Si was supplied as a powder [143, 145] or in gaseous phase as silane (SiH4) [146]. In a number of other studies, on th e other hand, the catalyst material was deposited directly on the Si substrate [142, 148155]. In these studies, the Si reactant is supplied by the solid Si substrat e during growth, and the oxygen is supplied in gaseous form either intentionally by introducing oxygen or air into the chamber or unintentionally as a residual gas due to leakage or impurities in the carrier gases used.

PAGE 59

59 In Chapter 2, we have demonstrated that single-walled carbon nanotubes (SWNTs) can be produced by the process of iron (Fe) ion implantation into thermally grown SiO2 layers, subsequent annealing, and CVD gr owth [139]. In this chapter, we experimentally demonstrate a similar approach for silicon oxid e nanowire growth by implanting Fe+ ions directly into thermally grown SiO2 layers on Si wafers to nucleate sili con oxide nanowires during subsequent annealing in argon and hydrogen. In contrast to the previous work [144], both reactants (Si and O) come from the SiO2 substrate, which acts as a solid source. As a result, this work integrates and simultaneously demonstrates three important re sults: (1) The use of ion implantation as a versatile method to create catalys t nanoparticles for silicon oxid e nanowire growth, (2) The use of Fe, a commonly used catalyst for single-wa lled carbon nanotube growth, as an efficient catalyst also for silicon oxide nanow ire growth, and (3) The use of SiO2 layers as a solid source in a solid-liquid-solid (SLS) growth mechanis m to achieve a Si:O ratio close to 1:2. Furthermore, we study the e ffect of temperature, H2 gas flow, and growth time on the silicon oxide nanowire growth and explai n the results in term s of a simple physical growth model based on the SLS mechanism. Experimental Method In our experiment, 500 nm thick SiO2 layers were first thermally grown on silicon (100) substrates. Fe+ ions were ion implanted into the SiO2 layers at energy of 60 keV with a dose of 1015 cm-2, as illustrated in Figure 4-1A. The implantation de pth profile calculated by SRIM simulations [120] is shown in Figure 4-1B, where the projected range Rp is 50 nm, and the peak concentration Cp is ~1.51018 cm-3. For the growth of silicon oxide nanowires, the asimplanted samples were placed in a one inch qua rtz tube furnace. The quartz tube was then purged at room temperature with 350 sccm flow rate of Ar and 200 sccm flow rate of H2 for 15 min. After purging the tube, the temperature was increased to 1100oC and the samples were

PAGE 60

60 annealed for 30 min under the same gas fl ow rates (350 sccm Ar and 200 sccm H2) to grow the silicon oxide nanowires. The system was then cooled down to room temperature under the same gas flow rates and the samples were taken out of the furnace and characterized. The as-grown samples were characteri zed by a JEOL JSM-6335F field emission gun scanning electron microscope (SEM) operati ng at 10 kV, a JEOL 2010F high resolution transmission electron microscope (HRTEM) opera ting at 100 kV equipped with selected area electron diffraction (SAED) and ener gy dispersive X-ray spectroscopy ( EDS) capability, and a Digital Instruments Nanoscope III Atomic Fo rce Microscope (AFM) operated in the tapping mode. For the HRTEM characterization, the as-g rown nanowires were dispersed in 2-propanol, the solution was sonicated, and a few drops of the resulting suspension was deposited on commercially available Cu TEM grids co ated with a holey carbon film [156]. Results and Discussion Figure 4-2 shows SEM images of the as -grown silicon oxide nanowires at 1100oC under 350 sccm Ar and 200 sccm H2 flow rates. Based on SEM observations, the density of nanowires is high and most nanowires are longer than 10 m. Furthermore, based on analysis of SEM and TEM images, the diameter of the si licon oxide nanowires were found to be in the range between 10 and 40 nm. Figure 4-3A and B show TEM images of the as-grown silicon oxide nanowires. The inset of Figure 4-3B shows the SAED patte rn of an individual silicon oxide nanowire. The high resolution TEM image [Figure 4-3B] and the SAED pattern confirm that the as-grown silicon oxide na nowires are amorphous. In addition, Figure 4-4 shows the EDS spectrum of a silicon oxide nanowire. The copper and carbon peaks in the spectrum come from the TEM grid [156] that the nanowires are deposited on. The only other visible peaks are oxygen and silicon, confirming that the as-gro wn nanowires are indeed silicon oxide. Furthermore, quantitative EDS analysis revealed that the atomic ratio of Si to O is 1:2.6,

PAGE 61

61 implying that the as-grown nanowires are close to SiO2 nanowires. The use of a thermally grown SiO2 layer as the source material could have play ed an important role in getting close to a 1:2 ratio. Three other growth conditions were investigated in order to gain a further understanding of the effect of growth parameters on the ion implanted catalyst nanoparticle formation and nanowire growth. In the first case, the growth temperature was lowered to 1000oC, keeping all the other parameters the same. Due to the dense growth at 1100oC (Figure 4-2), however, it is not possible to obtain exact information about the catalyst nanoparticle density and height for that high temperature growth condition for a direct comparison. Secondly, as it will be mentioned later, the proposed solidliquid-solid (SLS) growth model involves the fo rmation of a liquid all oy droplet containing Fe, Si, and O, which then nucleate s the silicon oxid e nanowire. Figure 4-5A shows the AFM image of the sample after growth at 1000oC. It is clear from the AFM image that no nanowires have been grown at this lower temperature. The average height of th e catalyst nanoparticles on the surface of the substrate extracted from cr oss-sectional AFM analysis is 8.2 nm. The absence of nanowires at 1000oC could be explained by two factors: First, the diffusion rate of implanted Fe atoms towards the surface depends exponentially on temperature. As a result, the density and height of catalyst nanoparticles formed on the SiO2 surface by the out-diffusion of Fe atoms could be too small to nucleate nanowire grow th at lower temperatures. If the temperature drops below the melting point of this alloy, nano wires cannot be nucleated. As a result of these two reasons, temperature is found to be a critical factor for determining nanowire growth. This observation is in agreement with previous work which also found that high temperatures were required for the growth of silicon oxide nanowires [142, 144]. In the second growth condition investigated, the growth te mperature was kept at 1100oC, but no H2 gas was supplied during the

PAGE 62

62 growth process, keeping all th e other parameters the same. Figure 4-5B shows the AFM image of the sample surface after growth without any H2. As clearly seen in the figure, no nanowires have been grown in the absence of H2 flow. The average catalyst nanoparticle height measured by cross-sectional AFM analysis is about 1 nm in this case. It has been well demonstrated that the diameter of nanowires is correlated with th e size of the catalyst nanoparticles [157]. In this case, it is clear from the AFM image that the ca talyst nanoparticle size (~ 1nm) is too small to nucleate silicon oxide nanowires, which typically have diameters of tens of nanometers. Hydrogen is known to enhance th e diffusion of impurities in SiO2 [158, 159]. The observed increase in the catalyst nanoparticle size when H2 is flown during the growth can be explained by a similar enhancement of Fe diffusion in SiO2 in H2 ambient. These observations confirm that the presence of H2 gas flow is also a critical factor for the catalyst nanoparticle formation. In the third growth condition, the gr owth time was reduced to 10 min at 1100oC (instead of 30 min), keeping all the other parameters the same. Figure 4-5C shows the AFM image of the sample surface after growth for 10 min. It is clear from the figure that no nanowires have been grown when the growth time is short. The average catalyst nanopartic le height measured by cross-sectional AFM analysis is 6.2 nm in this ca se, which (similar to the previous cases) is too small to nucleate silicon oxide nanowires. As a result, growth time is also an important parameter, since the catalyst nanoparticle size increases as a function of time due to the continuous out-diffusion of implanted Fe atoms to the sample surface. We have also performed growth on a control SiO2 substrate with no Fe ion implantation. The growth temperature was 1100oC and all the other growth pa rameters were the same as before. No nanowire growth took place in the ab sence of Fe ion implantation. As a result,

PAGE 63

63 this control experiment provided di rect evidence that the presence of Fe nanoparticles is essential for nucleating the growth of silicon oxide nanowires. The vapor-liquid-solid (VLS) growth mech anism [147] has been commonly used to explain the growth of nanowires from catalyst nanoparticles. In VLS growth, the precursor is initially provided in gas-phase and forms a liqui d alloy with the catalyst nanoparticles. In contrast, in our silicon oxide na nowire growth process, silicon and oxygen are provided from the solid SiO2 substrate, resulting in a solid-liquid-solid (SLS) growth mechanism. In analogy to the VLS case, the physical model for the SLS growth of silicon oxide nanowires can be proposed as follows: First, the as-implanted Fe atoms [Figure 4-6A] out-diffuse to form catalyst nanoparticles on the SiO2 surface [Figure 4-6B]. The catalys t nanoparticles from which nanowire growth has not yet nucleated could also get larger by a surface diffusion and Ostwald ripening process. Liquid all oy droplets containing Fe, Si, and O form from these catalyst nanoparticles [Figure 4-6C]. When these liquid alloy dr oplets become supersaturated in silicon and oxygen, solid silicon oxide nanowires nucleat e and precipitate out of these droplets [Figure 4-6D]. EDS characterization over the whole ar ea of the nanowire TEM specimen showed no Fe peak and further investigation of the nanowires by SEM and TEM revealed no iron catalysts at the tip of the nanowires. These results indica te that the iron catalysts remain on the SiO2 substrate during the growth of nanowires, which implie s a base-growth mechanism as indicated in Figure 4-6D. Conclusions In conclusion, we have experimentally demons trated the growth of silicon oxide nanowires from Fe+ ions implanted into thermally grown SiO2 layers on Si. We explained the growth results in terms of a physical model based on th e solid-liquid-solid (SLS ) growth mechanism. We also showed that the nanowires are silicon oxid e with a Si:O ratio of 1:2.6. Furthermore,

PAGE 64

64 high temperature, hydrogen gas fl ow, and growth time were found to be critical parameters for silicon oxide nanowire growth. This works op ens up the possibility of growing nanowires directly from solid substrates, controlling the or igin of nanowires at the nanometer scale, and integrating them into nonplanar three-dimensional device structures with precise dose control. This method of nucleating nanowire growth is no t limited to silicon oxide nanowires; it could also be generally applied to the growth of ot her types of nanowires fo r potential applications.

PAGE 65

65 Figure 4-1. Conditions of ion implantation. A) Schematic of SiO2/Si substrate and Fe ion implantation condition used for catalyst na noparticle formation for silicon oxide growth. B) Depth profile of Fe atoms ion-implanted into thermally grown SiO2 at an energy of 60 keV and a dose of 1015 cm-2, calculated by SRIM simulations, giving a projected range of Rp = 50 nm and a peak concentration of Cp ~ 1.51018 cm-3, as labeled. A B

PAGE 66

66 Figure 4-2. Scanning electron microscopy (SEM) images of silicon oxide nanowires grown from Fe catalyst ion implanted (60 keV energy and 1015 cm-2 dose) into thermally grown SiO2 layers. The diameters of the na nowires range between 10 and 40 nm. A B

PAGE 67

67 Figure 4-3. Images of silicon oxide nanowir es. A) Low resolution and B) high resolution transmission electron microscopy (TEM) images of silicon oxide nanowires. The inset of B shows the selected area electron diffraction (SAED) of an individual nanowire. The high resolution TEM image and the SAED pattern confirm that the silicon oxide nanowires are amorphous. A B

PAGE 68

68 Figure 4-4. Energy dispersive X-ray spectroscopy ( EDS) spectrum of an individual silicon oxide nanowire. The Cu and C peaks in the spectrum come from the TEM grid that the nanowires are deposited on. The only othe r visible peaks are Si and O, confirming that the as-grown nanowires are silicon oxide. Quantitative EDS analysis revealed that the Si:O atomic ratio is 1:2.6, im plying that the nanowires are close to SiO2 nanowires.

PAGE 69

69 Figure 4-5. AFM images of the SiO2 sample surface A) grown at 1000oC, B) grown at 1100oC, but with no H2 gas supplied during the growth process, and C) grown at 1100oC, but for only 10 min, keeping all the other parameters the same as in Figure 4-2. No silicon oxide nanowire growth was observed in any of these cases. A B C

PAGE 70

70 Figure 4-6. Schematic diagram of the proposed solid-liquid-solid (S LS) growth model of silicon oxide nanowires: A) First, Fe atoms are ion-implanted into the SiO2 substrate. B) At high temperature, Fe atoms out-diffuse to form catalyst nanoparticles on the SiO2 surface. C) Liquid alloy droplets c ontaining Fe, Si, and O form on the SiO2 surface. The source of the silicon and oxygen is the SiO2 substrate. D) Silicon oxide nanowires nucleate and precipitate out of the liquid alloy droplets. A B C D

PAGE 71

71 CHAPTER 5 GAN NANOWIRE GROWTH FROM I ON IMPLANTED IRON CATALYST Introduction One-dimensional (1D) nanowires of GaN have been grown in recent years using a variety of techniques including laser ab lation [102, 160], arc discharge [161], template-assisted growth [162], thermal evaporation [116], metal-organi c chemical vapor deposition (MOCVD) [163], molecular beam epitaxy (MBE) [164], and chem ical vapor deposition (CVD) [165-168]. For most of the synthesis using CVD, a gr owth model based on the vapor-liquid-solid (VLS) growth mechanism [147] has been used to explain the observed results. As explained in Chapter 4, an essential component of the VLS gr owth process is the nano scale catalyst particles required to nucleate the growth of nanowir es. For example, for GaN nanowire growth, nanoparticles of metals such as In [167], Ni [ 167, 169], Co [167], Au [1 70], or Fe [167, 171] have been used as catalyst. In particular, Fe has been shown to be a good catalyst for GaN nanowire growth since it dissolves both Ga and N and does not form a more stable compound than GaN [102, 171]. In this chapter, similar to SWNTs and SiOX nanowires, we experimentally demonstrate a simple and efficient approach for nucleating the growth of GaN nanowires by ion implantation of Fe+ ions directly into thermally grown SiO2 layers and subsequent annealing to form catalyst nanoparticles. This work experimentally shows that ion implantation can be used as a versatile method to create catalyst nanoparticles for wi de bandgap semiconductor nanowire growth, as well. Furthermore, we systematically charac terize the as-grown nanomaterials, discuss the effect of growth parameters on the nanostructures grown, and explain the gr owth results in terms of simple physical models based on the VLS growth mechanism.

PAGE 72

72 Experimental Method For our experiment, a 500 nm thick SiO2 layer was thermally grown on silicon (100) substrates. Fe+ ions were implanted into these substrat es at an energy of 60 keV with a dose of 1015 or 1016 cm-2. The implantation dept h profile calculated by SRIM simulations [120] for the 1016 cm-2 dose is shown in Figure 5-1A, where the projected range Rp is 50 nm, and the peak concentration Cp is ~ 1910 5 1 cm-3. The as-implanted samples were then placed into a 1-inch atmospheric quartz tube furnace and annealed at 900C under 300 sccm Ar and 200 sccm H2 flow for 30 minutes to form iron catalyst na noparticles on the oxide surface. After this annealing step, the galli um metal source (5N purity, Alfa Aesa r) was poured into a quartz boat and placed in the tube furnace, as shown in Figure 5-1B. Next, the ion implanted and annealed substrate was placed ~3 cm downstream of the ga llium metal source. The substrate was heated up to 850oC and annealed at 850oC for 15 min under 500 sccm Ar flow. After this step, the growth was performed at 850oC for ~2 hours with 15 sccm flow rate of NH3 and 300 sccm flow rate of H2. As grown nanotubes were characterized the as-grown nanowire samples were characterized by a JEOL JSM-6335F field emissi on gun scanning electron microscope (SEM), a JEOL 2010F high resolution transmission electr on microscope (HRTEM ) operating at 200 kV equipped with selected area electron diffraction (SAED) a nd energy dispersive X-ray spectroscopy (EDS) capability, a Phillips MRD XPert X-ray Diffraction (XRD) spectrometer with Cu K radiation, and a Digital Instruments Nanoscope III Atomic Force Microscope (AFM) operated in the tapping mode. For th e HRTEM characterization, the as-grown nanowires were dispersed in 2-propanol, the so lution was sonicated, and a few drops of the resulting suspension was deposited on commercially available Cu TEM grids coated with a holey carbon film.

PAGE 73

73 Results and Discussion Figure 5-2A and B show the AFM images of the SiO2 substrate after the i on implantation of iron and subsequent 900oC, 30 min annealing step mentioned above for the 1015 and 1016 cm-2 dose implants, respectively. We observed that, under the same anneal and growth conditions, the higher implant dose substrates (1016 cm-2) result in a higher nanowire density compared to the lower dose ones (1015 cm-2). This could be explained by th e higher density of larger catalyst nanoparticles formed on the higher implan t dose substrate, as evident from Figure 5-2. The flux of Fe atoms diffusing to the oxide surface is proportional to the dos e of the implant; as a consequence, a higher dose results in larger nanoparticles. It is also worth noting that the AFM images in Figure 5-2 were taken after th e 30 min anneal step at 900oC. During the 2 hr growth step at 850oC, the implanted Fe atoms continue to out-d iffuse towards the surface, and as a result, the density and size of catalyst nanoparticles conti nue to evolve as the growth proceeds. This is in contrast to solution-based or thin-film catal yst, where the catalyst dose on the surface does not change during growth. Next, we present the growth results of GaN na nowires from these catalyst nanoparticles for the higher implant dose substrate (1016 cm-2), which was found to result in a higher density of nanowires. Figure 5-3A shows an SEM image of GaN nanowires grown from the 1016 cm-2 dose ion implanted iron catalyst. The diameters of th e as-grown nanowires range from 15-60 nm and their lengths are between 1-20 m based on SEM and TEM analysis. XRD measurements were performed on bulk samples to determine overall crystal structure and phase purity of the asgrown GaN nanowires. Figure 5-3B shows the XRD pattern of the as-grown GaN nanowires where the diffraction peaks labeled by their Mille r indices are indexed to wurtzite GaN with lattice constants a = 0.3186 nm and c = 0.5178 nm.

PAGE 74

74 We have also characterized the detailed structure and composition of individual GaN nanowires by HRTEM, SAED and EDS as shown in Figure 5-4A and B, providing further experimental evidence that the as-grown nanowires are single-cr ystal hexagonal wurtzite GaN. The effect of the various CVD parameters on the growth of GaN nanowires from ion implanted Fe catalyst was also investigated. The overall reaction of ga llium metal vapor and ammonia to form GaN nanowires can be expressed as 2 3H 3 GaN 2 NH 2 Ga 2 (12) First, by varying the distance between the Ga source and the substrate [see Figure 5-1B], we observed that the density of nanowires is reduced as the distance increases above 3 cm. Furthermore, we found that increasing th e growth time increases the average GaN nanowire length. Increasing th e growth temperature above 900oC was found to reduce the density of GaN nanowires, most likely due to th e evaporation of too much Ga from the source material at higher temperatures. In addition, having either a high (~30 sccm) or low (~5 sccm) NH3 flow rate was observed to significantly re duce the density and length of as-grown GaN nanowires. These findings indi cate that too much or too little supply of Ga or NH3 impede GaN nanowire growth. The presence of H2 co-flow was found to be a crucial factor for GaN nanowire growth, as well. Any residual oxygen in the CVD chambe r easily oxidizes the na nowires, yielding Ga2O3 nanowires instead of GaN. The hydrogen co-flo w prevents the oxidation of GaN nanowires, as well as reducing the decompos ition rate of ammonia. As mentioned previously in the experiment al procedure, we have performed a 15 min anneal in Ar atmosphere after the furnace temp erature has reached the growth temperature of 850oC. We found that this anneal step increases the GaN nanowire length and density. This

PAGE 75

75 could be due to the fact that Ga begins to eva porate from the source material during this anneal step, providing a sufficient supply of Ga vapor as soon as the NH3 is introduced into the chamber in the growth step. The growth of the GaN nanowires can be explained by the vapor-liquid-solid (VLS) growth model [147] as illustrated in Figure 5-5: First, the SiO2 substrate is implanted with Fe+. Then, during the subsequent anneal step, the as -implanted Fe atoms out-diffuse to the surface and aggregate to form catalyst nanoparticles on the SiO2 surface. The catalyst nanoparticles from which nanowire growth has not yet nucleated could also get larger by a surface diffusion and Ostwald ripening process. During the CVD growth, liquid alloy dropl ets containing Fe, Ga, and N form from these catalyst nanoparticle s. When these liquid alloy droplets become supersaturated in Ga and N, solid GaN nanowires nucleate and precipitate out of these droplets. Conclusions In conclusion, we have experimentally dem onstrated the catalytic CVD growth of GaN nanowires from Fe catalyst nanoparticles formed by ion im plantation into thermally grown SiO2 layers and subsequent annealing. This wo rk provides experimental evidence that ion implantation can also be used as a versatile me thod to create catalyst nanoparticles for wide bandgap semiconductor nanowire growth. Furthermore, we have systematically characterized the structural properties of the as-grown nanowires We have found that the distance between the Ga source and the substrate, growth temperature, growth time, and gas flow rates are all critical parameters for nanowire growth. The growth of GaN nanowires can be explained by the catalytic VLS growth mechanisms. This work opens up the possibility of controlling the origin of wide bandgap semiconductor nanowires at the nanometer scale us ing the technique of ca talyst ion implantation

PAGE 76

76 through a lithographically defined mask, of integr ating nanowires into non-planar 3D device structures.

PAGE 77

77 Figure 5-1. Preparation of GaN nanowire grow th from ion implanted catalyst. A) Depth profile of Fe atoms ion-implan ted into thermally grown SiO2 at an energy of 60 keV and a dose of 1016 cm-2 calculated by SRIM simulations, giving a projected range of Rp = 50 nm and a peak concentration of Cp ~ 1910 5 1 cm-3, as labeled. B) Schematic of the chemical vapor deposition (CVD) setup used for GaN nanowire growth, indicating the Ga source and th e ion implanted substrate. B A

PAGE 78

78 Figure 5-2. AFM images of Fe catalyst nanopar ticles formed after annealing the as-implanted SiO2 substrates for 30 min at 900C under Ar and H2 flow. The ion implantation energy is 60 keV and the implant doses are A) 1015 and B) 1016 cm-2. A B

PAGE 79

79 Figure 5-3 Characterizations of grown GaN nanowires. A) SEM image and B) XRD pattern of the as-grown GaN nanowires (from the 1016 cm-2 dose ion implanted Fe catalyst) where the diffraction peaks labeled by thei r Miller indices are indexed to wurtzite GaN with lattice constants a = 0.3186 nm and c = 0.5178 nm. A B

PAGE 80

80 Figure 5-4. Characterizations of grown GaN nanowires. A) HRTEM image showing the selected area electron diffraction (SAED) pa ttern (<001> zone axis) in the inset and B) Energy dispersive X-ray spectroscopy (EDS) spectrum of an individual GaN nanowire, providing further experimental ev idence that the as-grown nanowires are single-crystal wurtzite GaN. The Si peak is from the substrate. A B

PAGE 81

81 Figure 5-5. Schematic illustration of the vapor-liquid-solid (VLS) growth mechanism for GaN nanowires. First, Fe atoms are ion-implanted into the SiO2 substrate. During the subsequent anneal step, the as-implanted Fe atoms out-diffuse to the surface and aggregate to form catalyst nanoparticles on the SiO2 surface. Next, during the CVD growth, liquid alloy droplets containing Fe, Ga, and N form from these catalyst nanoparticles. Finally, when these liquid alloy droplets become supersaturated in Ga and N, solid GaN nanowires nucleate and precipitate out of these droplets.

PAGE 82

82 CHAPTER 6 NANOLITHOGRAPHIC PATTERNING OF TRANSPARENT, CONDUCTIVE SINGLEWALLED CARBON NANOTUBE FILMS BY INDUCTIVELY COUPLED PLASMA REACTIVE ION ETCHING Introduction A single-walled nanotube film (SWNT film) is a three-dimensional film of tens of nanometers thickness, consisting of an interwoven mesh of single-walled nanotubes, as shown in Figure 6-1A. In the case of SWNT films, indi vidual variations in diameter and chirality are ensemble averaged to yield uniform physical an d electronic properties [172-175]. As a result, the reproducibility and re liability problems found in indi vidual nanotubes are solved, and carbon nanotube film based devices can easily be ma ss produced in a cost effective manner. Furthermore, what makes SWNT films even more attractive for device applications is that they are flexible, transparent, and conductive. SWNT films have a resistivity on the order of 10-4 cm, and it is also optically transparent over the visible and near-infrare d (near-IR) portions of the spectrum. For example, Rinzler et al. have demonstrated that an as-prepared nanotube film of 50 nm thickness has a transmittance greater than 70% over the visible part of the spectrum, and this transmittance increases to more than 90% in the near-IR at a wavelength of 2 m [172], as shown in Figure 6-1B. This transmittance is comparable to ITO [176]. These outstanding properties have established SWNT films as a new class of optically transparent and electrica lly conducting materials that can be us ed in applications such as thin film transistors [31, 32], flexible microelect ronics [33-35], chemical sensors [36-39], and optoelectronic devices [40-43]. Any potential device application utilizing SWNT films requires the capability to efficiently pattern them. In this chapter, we use photol ithography and e-beam lithography, and subsequent O2 plasma etching in an ICP-RIE system to pa ttern nanotube films down to submicron lateral

PAGE 83

83 dimensions. We experimentally show that f eatures with linewidths less than 100 nm can be successfully patterned using this technique with good sele ctivity and directionality. In addition, we systematically study the effect of ICP-RIE etch parameters, such as the substrate bias power and chamber pressure, on the SWNT film etch ra te and etch selectivity. We also compare O2 plasma etching of SWNT films in an ICP-RIE syst em to that in a conventional parallel plate RIE system. We find that using an ICP-RIE system significantly increases the SWNT film etch rate and the etch selectivity between the SW NT film and polymethylmethacrylate (PMMA) compared to a conventional RIE system, making it pos sible to pattern lateral features as small as 100 nm. The simple and efficient top-down patterning capability opens up tremendous opportunities for integrating SWNT films into a wide range of electronic and optoelectronic devices. SWNT Film Deposition SWNT films were deposited by a vacuum f iltration method as described in detail previously [37, 40, 172]. In summary, a di lute suspension of purified SWNTs was vacuumfiltered onto a filtration membrane. The nanotubes deposit as a thin film on the membrane with the thickness of the film controlled by the con centration of nanotubes in the suspension and the volume of the suspension filtered. The film can then be transferred onto a desired substrate by placing the film side against the substrate, appl ying pressure, and drying the film. To complete the process, the filtration membrane is dissolv ed in a solvent, leavi ng only the nanotube film adhered to the substrate. The substrates used in this work were (100) silicon with a 500 nm layer of thermally grown SiO2 on top. Lithography After the deposition step, the SWNT film was patterned e ither by photolithography or ebeam lithography. For photolithography, three differe nt types of resist processes were used as

PAGE 84

84 the mask. The first process used a 1.3 m thick layer of Shipley Mi croposit S1813 photoresist. It was found by extensive AFM imaging that when the S1813 resist is deposited directly on top of individual nanotubes, it leav es a residue [134]. On the ot her hand, it was observed that Microchem LOR3B lift-off resist and PMMA do not contaminate nanotubes [134]. To protect the SWNT film from potential contamination due to the S1813 resist, for the second process, a dual layer resist structure consisting of a 1.3 m thick S1813 layer on top of a 250 nm thick LOR3B layer was used. Finally, for the third proce ss, a dual layer resist process consisting of a 1.3 m thick Shipley S1813 layer on top of a 250 nm thick PMMA layer (950K, 4% in anisole) was used. Since the S1813 resist is not in dire ct contact with the na notubes in the second and third processes, no residue is left on the nanot ube film during fabrication. However, PMMA cannot be exposed by the 365 nm light source avai lable in the Karl Suss MA-6 contact mask aligner that was used for photolit hography. As a result, for the thir d process, the S1813 layer was first exposed by the mask aligner and developed. Subsequently, the PMMA layer was patterned by O2 plasma etching with the S1813 layer acting as the mask. In all three resist processes, Shipley Microposit MF319 was used as the developer. For e-beam lithography, a single layer of PMMA (950K, 2 or 4% in anisole depe nding on the feature size patterned and PMMA thickness desired) was used as the masking laye r, and a Raith 150 e-beam writer was used for exposure. O2 Plasma Etching After exposure and development, the nanotube film not protected by the resist mask was etched using an O2 chemistry in a Unaxis Shuttlelock IC P-RIE system. The schematic of the ICP etcher is shown in Figure 6-2. The ICP-RIE system d ecouples plasma density (controlled by the ICP power supply) and ion energy (contr olled by the substrate power supply). As a

PAGE 85

85 result, compared to conventional diode RIE systems, very high plasma densities (>1011 ions cm3) can be achieved at lower pressures, resulting in more anisotropic etch profiles and significantly higher etch rates [131, 177]. Etching in ICP systems has a large physical component due to the ions combined with a chemical component. O2 plasma is commonly used for removing organic materials such as photoresist, and has also been used to etch carbon nanotubes [31, 178]. The reaction between oxygen and organic materials pr oduces volatile species such as CO and CO2, which are pumped out during the etch process [1 31]. The etch parameters for our initial SWNT etch recipe were 300 W power on the 2 MHz ICP rf supply, 100 W power on the 13.56 MHz substrate rf supply, 45 mTorr ch amber pressure, and a 20 sccm O2 flow rate. In addition, a helium flow rate of 10 sccm was used to cool down the substrate. We will discuss below how changing various etch parameters a ffects the etch rates of the SWNT film and resists. After the ICP-RIE etch, the resist mask la yers were stripped in acetone when S1813 or PMMA were used as the mask, and in Microchem Nanoremover PG when LOR3B was included in the mask, since LOR3B does not dissolve in acetone. Results and Discussion Resulting SWNT film etch profiles were char acterized by a Digital Instruments Nanoscope III AFM. Figure 6-3A shows an AFM image of a ~3 m line etched in a ~20 nm thick nanotube film using the LOR3B/S1813 dual re sist photolithography process (i.e. the second process) and the initial ICP etch recipe given ab ove. The cross-sectiona l height profile for the same AFM image is plotted in Figure 6-3B, showing clearly the tr ansition between the film and the etched regions. Similar etch profiles were obtained using the other two resist processes described above. Figure 6-3C shows an AFM image of a seri es of lines with nanotube film width and spacing of about 200 nm patterned by e-beam lithography and ICP etching of a SWNT film

PAGE 86

86 of about 14 nm thickness. The cross-sectional he ight profile for the same AFM image is plotted in Figure 6-3D, showing a clear tran sition between the film and the etched regions even at these submicron lateral dimensions. Furthermore, Figure 6-4A and B show AFM and SEM images, respectively, of letters printed on nanotube film using e-beam lithography and ICP-RIE etching. The width of the text characters are on the order of 200-300 nm and th e SWNT film thickness is 20 nm, demonstrating that the nanotube film can indeed be patterned into nanometer size structures of arbitrary shape by this fabrication method. In order to characterize quantitat ively the SWNT film and resist etch rates using the initial ICP-RIE etch recipe given above, a series of lines with equal width and spacing were partially etched in 50-100 nm thick nanotube films, such that some nanotube film still remained in the etched areas. Figure 6-5A shows an AFM image of such a series of lines with ~200 nm width and spacing partially etched in a 75 nm thick SWNT film. Unlike the lines in Figure 6-3C, the lines in Figure 6-5A have not been etched all the way down to the substrate, as evident from the texture of the remaining film visible in the et ched areas. By measuring the height difference between the partially etched and not-etched film lines using cross-sectional AFM analysis, the average etch depth, and as a result, the et ch rate can be calculated. For example, Figure 6-5B shows the cross-sectional AFM pr ofile for the lines shown in Figure 6-5A, giving an average etch depth of about 19 nm for this particular sample. Dividing this depth by the etch time of 8 s, a nanotube film etch rate of ~2.4 nm/s is obtained. Furthermore, the S1813, LOR3B, and PMMA etch rates were determined by measuring the initial and final resist thicknesses using a Nanometrics Nanospec spectrometer, and dividing by the etch time. Using the initial recipe (i.e. 300 W ICP power, 100 W substrate bias power 45 mTorr chamber pressure, 20 sccm O2 flow

PAGE 87

87 rate, and 10 sccm helium flow rate for substrate cooling), etch ra tes of 2.37, 4.59, 4.58, and 6.65 nm/s were observed for SWNT film, S1813, LOR3B, and PMMA, respectively, as listed in the first column of Table 6-1. The error bar on thes e etch rates is approximately % 10 The SWNT film etch rate is similar in magnit ude to the ~4 nm/s observed in recent work using another ICP system [178]. For SWNT films of tens of nm thickness, such as those used in this work, the 2.37 nm/s etch rate of the initial recipe provides bot h reasonably short etch times and a good control of the etch uniformity. The selectivity S of the etch between the nanotube film and the resist mask is defined by RESIST SWNTr r S: where where SWNTr is the etch rate of the SWNT film and RESISTr is the etch rate of the partic ular resist used as the mask. Using this definition, selectivity values of 1:1.94, 1:1.93, a nd 1:2.81 are obtained for S1813, LOR3B, and PMMA masking layers, respectively. Carbon nanotubes are much harder to etch compared to photoresists since they are chemically resistant and structural ly stable [28]. As a result, the etch rate of the SWNT film is slower than that of resists in an O2 plasma, and the selectivity values are less than unity. Since the resists are used as the etch mask, they need to be thick enough to withstand the nanotube film etch. The minimum resist thickness required for a given SWNT etch process is determined by the selectivity S of the etch process. Typical S1813 and LOR3B resist thicknesses used for photolithography are larger than 1 m; as a result, based on the selectivity values given above, hundr eds of nm thick SWNT films can easily be patterned by photolithography. More importantly, since th e PMMA etch rate is not significantly higher than the nanot ube film etch rate, typical PMMA thicknesses necessary for ebeam lithography (100-300 nm) can be used to pa ttern thin SWNT films (i.e. less than 100 nm) down to very small (<100 nm) lateral features. In short, although the et ch selectivity between the SWNT film and PMMA is less than unity (S = 0.36), it is still large enough to allow for e-

PAGE 88

88 beam patterning of SWNT films. We show below that this selectivity value is much smaller (S = 0.09) for a conventional parallel plate diode RIE system. We also show below that the SWNT film etch rate is much fa ster in a ICP-RIE system compared to that in a conventional RIE system. The high density plasma and low pressu re achievable in an ICP-RIE system provides a large physical etch component, which results in si gnificantly higher etch rates. As a result, the use of an ICP-RIE system is crucial in the ability to pattern SWNT films down to ~100 nm lateral dimensions by e-beam lithography. Aspect ratio dependent etching has been observe d in some etch processes, such as silicon trench etching, resulting in a lower etch rate for smaller wi dth trenches [131]. Using the approach described in the preceding paragraph, we systematically studied the effect of the line width on the nanotube film etch rate for widths ranging from 50 m all the way down to 100 nm. The spacing between the etched lines was set equal to the width of the lines in all cases. The etch rate was found to be almost constant at 2.37 .3 nm/s, independent of the line width etched. This is most likely due to the fact that all our samples have a SWNT f ilm thickness t < 100 nm. The aspect ratio AR of the nanotube film etched, defined as t/w, where w is the width of the line etched, always satisfies 1 AR for all the samples. In other words, the plasma density is high enough and the aspect ratio is small enough so that reactant species are able to make it to the bottom of the etched lines even fo r the smallest (100 nm) linewidths. Effect of Substrate Bias Power Furthermore, we systematically studied the eff ect of changing various ICP etch parameters on the etch rates of the SWNT film and different resists, as listed in Table 6-1, using the procedure described above. To investigate the effect of the substrate bias power on the etch rate, we decreased the substrate power from 100 W to 15 W, keeping all the other etch

PAGE 89

89 parameters constant as in the initial recipe. Table 6-1 shows that the na notube film and resist etch rates are decreased by about a factor of 10 compared to those of the initial recipe. By reducing the substrate bias, the ion energy is reduced resulting in a substantially slower etch rate. A slow etch rate could be usef ul in applications where the SW NT film thickness is very small and the etch rate and uniformity n eeds to be precisely controlled. Effect of Chamber Pressure To investigate the effect of chamber pressure on the etch rate, we decreased the chamber pressure from 45 mTorr to 10 mTorr, keeping all the other etch parameters constant as in the initial recipe. Table 6-1 shows that the nanotube film and resist etch rates increase by a factor between 1.7 and 3.5 compared to those of the in itial recipe. A lower chamber pressure results in a more directed etch because of fewer gas-phase collisions in the sheath, and also increases the etch rates due to a larger physical etch compone nt. A faster etch rate could be useful in applications where the SWNT film thickness is la rge. Furthermore, by taking the ratio of the etch rates listed in Table 6-1, selectivity va lues of 1:0.95, 1:1.21, and 1:1.59 are obtained for S1813, LOR3B, and PMMA masking layers, respectiv ely. These selectivity values are higher than those of the initial recipe. This is likely due to an in crease in the physic al etch component, which etches the nanotube film and resists at si milar rates. In addition, increasing the chamber pressure from 45 mTorr to 100 mTorr (maximum pressure achievable in our system) was found not to change the etch rates of the nanotube film and re sists significantly. Effect of Substrate Cooling Furthermore, we have investigated the effect of substrate cooling on the etch rates of the SWNT film and resists. Increasing the helium flow rate (which actively cools the substrate) from 10 sccm to 40 sccm, keeping all other etch pa rameters constant as in the initial recipe, did not change the etch rates of the SWNT film and resists compared to those of the initial recipe.

PAGE 90

90 Comparison to Parallel-Plate RIE System To compare the etch rates of SWNT film and the three resists in an ICP-RIE system to those in a conventional parallel plate RIE system, we have al so etched the SWNT film and resists using a Plasma Sciences RIE 200W etcher, in which there is only one rf power source of 13.56 MHz frequency, and as a result, the pl asma density and the ion energy are no longer decoupled. Using an rf power of 30 W, O2 flow rate of 12.5 sccm, and a chamber pressure of 140 mTorr, we have observed that the etch ra tes of both the SWNT film and resists are substantially lower in this system, as listed in the last column of Table 6-1. The etch rate of the SWNT film in the conventional RIE system was 0.05 nm/s, which is about 5 times slower than that in the ICP-RIE system even with a low substrate bias power of 15 W (See Table 6-1). This is due to a lower plasma density in the c onventional RIE system. Furthermore, for the conventional RIE system, the etch selectivity be tween the SWNT film and the resist mask has decreased to 1:2.8, 1:3.8, and 1:11 for S1813, LO R3B, and PMMA, respectively. This is due to a reduction in the physical etching component using the conventional RIE system. These results demonstrate that the use of an ICP etcher provides significant advantages, such as faster etch rates and better sele ctivity, over conventional parallel plate plasma systems in order to be able pattern submicron feat ures in nanotube films. Conclusions In conclusion, in this chapter, we have de monstrated the ability to efficiently pattern SWNT films with good sele ctivity and directionali ty down to submicron lateral dimensions by photolithography or e-beam lithography and O2 plasma etching using an ICP-RIE system. We systematically studied the effect of ICP-RIE etch parameters on the nanotube film etch rate and etch selectivity. Decreasing th e substrate power from 100 W to 15 W, decreased the nanotube film and resist etch rates by about a factor of 10. Decreasing the chamber pressure from 45

PAGE 91

91 mTorr to 10 mTorr increased the nanotube film a nd resist etch rates by a factor between 1.7 and 3.5. It also increased the etch selectivity between the nanotube film and the resist masks. On the other hand, increasing the chamber pressure from 45 mTorr to 100 mTorr did not change the etch rates of the nanotube film and resists signifi cantly. Similarly, increasing the helium flow rate (which actively cools the substrate) from 10 sccm to 40 sccm did not produce a significant change on the etch rates of the SWNT film and the three resists. Furthermore, the SWNT film etch rate was found to be independent of the line width etched for linewidths ranging from 50 m down to 100 nm. In addition, by comparing the etch rates of SWNT film and the three resists in an ICP-RIE system to those in a c onventional parallel plat e RIE system, we have demonstrated that using an ICPRIE system provides significant advantages, such as faster etch rates and better etch selectiv ity, over conventional pa rallel plate RIE plasma systems, making it possible to pattern lateral features as small as 100 nm in nanotube films. In short, the simple and efficient top-dow n patterning capability developed in this chapter could open up tremendous opportunities for integrating si ngle-walled nanotube films into a wide range of electronic and optoelectronic devices.

PAGE 92

92 Figure 6-1. Image and optical properties of SW NT films. A) Atomic force microscope (AFM) image of a 150 nm thick film and B) transmittance spectra for two films of thickness 50 nm and 240 nm. The curves with higher tr ansmittance shown in the upper left are for the 50 nm film. The gray curve shows the transmittance of unbaked film and the black curve represents the transmittance of baked (dedoped) film. (Taken from Ref. [172]) B A

PAGE 93

93 Figure 6-2. Schematic of ICP-RIE system show ing separated ICP and substrate power supplies.

PAGE 94

94 Figure 6-3. Images of patterned SWNT films. A) Top view AFM image of a ~3 m line etched in a ~20 nm thick SWNT film using the LOR3B/S1813 dual layer resist photolithography process and the initial ICP etch recipe de scribed in the text. The SiO2 substrate is exposed in the center area, where the film is completely removed. The SWNT mesh making up the nanotube film is clearly visible at the left and right of the AFM image. B) Cross-sectional height data for the AFM image of part A. C) AFM image of a series of nanotube film lines having equal widths and spacings of ~200 nm, patterned on SiO2 by e-beam lithography and ICP-RIE etching, as described in the text. The film thickness is ~14 nm. D) Cross-sectional height data for the AFM image of part C. A B D C

PAGE 95

95 Figure 6-4. Text characters printed in SWNT film by e-beam lithograp hy and ICP-RIE etching, A) AFM image of UF and B) SEM image of GATORS. The width of the text characters are on the order of 200-300 nm and the SWNT film thickness is 20 nm. A B

PAGE 96

96 Figure 6-5. Images of patterned SWNT films. A) AFM image of a series of nanotube film lines having equal widths (and spacings) of ~200 nm half-way etched in a 75 nm thick SWNT film by e-beam lithography and ICP-RIE etching, as described in the text. Unlike the lines shown in the AFM image of Figure 6-3C, the lines in this AFM image have not been etched all the way down to the substrate, as evident from the texture of the remaining film mesh visibl e in the etched areas. The scale bar is 200 nm. B) Cross-sectional height data fo r the AFM image of part A, showing an average etch depth of about 19 nm for this particular sample. The etch rate can be calculated by dividing this etch depth by the total etch time. A B

PAGE 97

97 Table 6-1 Etch rates of SWNT film and th ree different resists (S 1813, LOR3B, and PMMA) under different plasma etch conditions usi ng Unaxis Shuttlelock ICP-RIE system and Plasma Sciences RIE 200W system. ICP-RIE System Etch Rates (nm/s) Material Initial recipe Low Substrate Bias Power (15 W) Low Chamber Pressure (10 mTorr) Parallel Plate RIE System Etch Rate (nm/s) SWNT film 2.37 0.238.280.05 S1813 4.59 0.397.850.14 LOR3B 4.58 0.4410.00.19 PMMA 6.65 0.6313.20.55 Note: The initial ICP-RIE recipe in column I corresponds to an ICP power of 300 W, substrate bias power of 100 W, chamber pressure of 45 mTorr, and O2 flow rate of 20 sccm. In addition, a Helium flow rate of 10 sccm was used to cool down the substrate. The headings of the other columns indicate the parameters changed compared to the initial recipe, with all the other parameters kept c onstant. The parallel plate RIE system etch parameters were rf power of 30 W, O2 flow rate of 12.5 sccm, and a chamber pressure of 140 mTorr.

PAGE 98

98 CHAPTER 7 FABRICATION AND DARK CURRENT CHARACTERIZATION OF METALSEMICONDUCTOR-METAL (MSM) PHOTOD ETECTORS WITH TRANSPARENT AND CONDUCTIVE CARBON NANOTUBE FILM SCHOTTKY ELECTRODES Introduction Among various photodetector structures, metal-semiconductor-metal (MSM) photodetectors (PDs) are one of the most promis ing candidates for monolithic integration in optoelectronic integrated circu its (OEICs). An MSM photodetec tor is an optoelectronic device that absorbs optical energy and c onverts it to electrical energy in the form of a photocurrent. Photodetectors are widely used in optical co mmunication systems. An MSM photodetector is fabricated by forming two Schottky contacts on an undoped semiconductor layer. The typical configuration is metal contacts in the form of interdigitated finger electrodes placed on a semiconductor substrate, as shown in Figure 7-1. The semiconducting material is reverse biased to form a depletion layer and create an electric field fo r photogenerated carriers to flow through. This produces a small leakage curr ent called the dark cu rrent, and should be minimized to reduce power dissipation of the de vice, and increase its sensitivity. When the device is exposed to light, its absorption in the semiconductor layer cr eates electron/hole pairs that are swept out of the absorbing semiconduc ting layer by the electric field, creating the photocurrent. The advantages of MSM PDs compared to other commonly used detectors, such as PIN diodes and avalanche photodiod es, are simplicity of fabrica tion, low cost, opportunity for monolithic integration, potent ially large bandwidth, and very low capacitance [179]. The low capacitance can be extremely usef ul in reducing the amplifier noi se and increasing signal to noise ratio of an optical link [180]. However, the main disadvantage limiting the widespread use of MSM PDs is their low responsivity due to the reflection of light by the metal electrodes on

PAGE 99

99 top of the semiconductor. This results in a low photocurrent-to-dark current ratio defined asdark photoI I PDR / In this chapter, we successfully integrate transparent and conduc tive single-walled carbon nanotube film (SWNT film) electro des with GaAs substrates to make hybrid Schottky MSM PDs. We demonstrate the Schottky behavior of SW NT film contacts on GaAs by fabricating and characterizing Metal-Semiconductor-Metal (MSM ) devices with SWNT film electrodes. Furthermore, we study the effect of device geom etry on the dark curren t of these SWNT filmGaAs MSM devices. In recent previous work, SWNT films have only been demonstrated as ohmic contacts in optoelectronic devices, such as GaN light-emitting diodes (LEDs) [40], organic solar cells [41, 43], a nd organic LEDs [42]. Furthermor e, unlike applications based on individual SWNTs, SWNT films do not suffer from reliability and reproducibility problems, since individual variations in di ameter, chirality, location, and di rection are ensemble averaged to yield uniform physical a nd electronic properties. Experimental Method MSM photodetectors with SWNT film electrodes were fabr icated on nominally undoped (~810 1 2 cm) (100) GaAs substrates. After the GaAs substrate was clean ed by solvents, a ~100 nm thick silicon nitride (S iN) isolation layer was deposited on the substrate using plasma enhanced chemical vapor deposition (PECVD) [Figure 7-2A]. Subsequently, active area windows of various dimensions we re opened in the SiN film usi ng plasma etching. This was followed by the deposition of ~40 nm thick SWNT film, prepared by vacuum-filtration as explained in detail previously [40, 172, 181] [Figure 7-2B]. The deposited SWNT film had a resistivity of about 10-4 .cm. The SWNT film was then patterned into interdigitated finger electrodes by photolithography and inductively c oupled plasma etching (ICP), using a method

PAGE 100

100 we have described previously [181, 182] [Figure 7-2C]. Finally, for ease of electrical probing, a Chromium/Palladium (7 nm/43 nm) metal stack was patterned on the nanotube film contact pads using photolithography, e-beam eva poration, and subsequent lift-off [Figure 7-2D]. The SWNT film contact pads end up on top of the SiN isolation, which eliminates parasitic leakage paths, and therefore reduces the dark current [183]. Figure 7-2E shows the optical microscope image of a finished MSM device and Figure 7-2F shows the Atomic Force Microscope (AFM) image of the area between two SW NT film electrode fingers. In order to study the effect of device geometry on the dark current, MSM devices with different active area width FW finger length FL finger width W and finger spacing S were fabricated. Thes e dimensions are labeled in Figure 7-2E. Finger widths smaller than 5 m were not used, since it has been shown by 4point-probe measurements and Mont e Carlo simulations that the SWNT film resistivity increases strongly at smaller widths because the film approaches the percolation threshold [181, 184]. Result and Discussion The Figure 7-3 shows the dark I-V characteristics of this device at room temperature (294 K) in linear scale. The data clearly exhibit the characteristic I V curves of two back-toback Schottky diodes making up the MSM photodet ector [185-187], demonstrating that the SWNT film indeed makes a Schottky contact to GaAs. Furthermore, the symmetry of the I V curve in the inset demonstrates that the two metal-semiconductor (M-S) contacts are identical, implying that the SWNT film acts as a uniform ma terial. In an MSM structure, at high applied voltage, one of the M-S contacts is reverse bias ed and the other one forward biased; and as a result, the reverse biased contac t limits the current a nd results in current saturation [185, 188]. However, Figure 7-3 shows that after the first steep rise, the current does not completely saturate at high voltages, but slowly increases with incr easing voltage. This slow increase in the current

PAGE 101

101 at higher voltages can be explained by Schottky barrier lowering due to charge accumulation at surface states, and image force lowering at the ed ges of the electrodes where there is a strong electric field, as previously repor ted for MSM photodetectors [189, 190]. Next, we study the effect of the device geometry on the dark I-V characteristics at room temperature (294 K). Figure 7-4A shows the dark I V curves for devices with identical W FL and FW but with S ranging from 10 to 20 m. Increasing the spacing in these devices (with the same area) decreases the number of finger pairs n which is given by n = FW / 2( W + S ). This decrease in turn reduces the am ount of dark current in the devi ce, in agreement with the trend observed in Figure 7-4A. Figure 7-4B shows the dark I V curves for devices with identical FL and FW but with W = S ranging from 15 to 30 m. The dark current is found to monotonically decrease with an increase in W. It has been observed that in MSM detectors, beyond a certain finger width, the dark current beco mes roughly independe nt of the width W and is proportional to the product of FL and n [191]. This is due to current crow ding at the edges of the electrodes as illustrated explicitly in Figure 7-4C, which shows a MEDICI [192] simulation of the crosssectional current density distri bution in the GaAs substrate be tween two electrode fingers ( W = S = 20 m) at V = 3 V bias. It is evident from this simulation that current crowding occurs at the electrode edges, which results in the effective device area to be weakly dependent on the width of the electrode. Therefore, since FL and FW are constant and W = S for all the devices shown in Figure 7-4B, n thus the dark current shoul d be inversely proportional to W in agreement with the observed trend in the figure. Finally, Figure 7-4D shows the dark I V curves for devices with identical W and S but with FL = FW ranging from 200 to 400 m. In this case, both FL and n (which depends linearly on FW ) vary, resulting in a strong cha nge in the amount of dark current, in agreement with the trend observed in Figure 7-4D. These results show that the dark

PAGE 102

102 current in SWNT film-GaAs MSM devices scal es rationally with device geometry. Ongoing work by other students in our gr oup include extracting the barrier height of SW NT film-GaAs Schottky contacts using temperature-dependent I-V measurements and characterizing the photocurrent of these MSM devices. Conclusions In summary, we have fabricated and charac terized the effect of device geometry on the dark current of MSM photodetectors based on SW NT film-GaAs Schottky contacts. We have observed that dark currents of the MSM devices scale rationally with device geometry, such as the device active area, finger width, and finger sp acing. The results open up the possibility of integrating SWNT films as transparent and c onductive Schottky electr odes in conventional semiconductor electronic and optoelectronic devices.

PAGE 103

103 Figure 7-1. Schematic of MSM photodetector w ith interdigitated metal finger electrodes on top of semiconducting layer. When the phot odetector is illumi nated with light, electron/hole pairs are creat ed in the semiconductor, which are swept out of the absorbing semiconducting layer by the elec tric field between the metal finger electrodes, creati ng a photocurrent. / h / / / h s e m i c o n d u c t o r M e t a l / h / / / h s e m i c o n d u c t o r M e t a l(a) / h / / / h s e m i c o n d u c t o r M e t a l / h / / / h s e m i c o n d u c t o r M e t a l(a)

PAGE 104

104 Figure 7-2. Schematic of process flow for MSM phot odetectors with SWNT film electrodes along the dashed line AB shown in part E: A) SiN isolation layer deposited on a GaAs substrate, B) SWNT film prepared by vacuum-filtration deposited on the substrate after opening the active windows in the SiN layer, C) SWNT film patterned into interdigita ted electrode fingers by photolith ography and inductively coupled plasma (ICP) etching, and D) Cr/Pd meta l contacts patterned on the nanotube film contact pads using photolithography, e-beam ev aporation, and subse quent lift-off. E) Optical microscope image of the finish ed MSM photodetector, showing the various device dimensions. F) Atomic Force Mi croscope (AFM) image showing the area between two SWNT film electrode finge rs of the MSM device of part E. A B C D E F

PAGE 105

105 Figure 7-3. Dark current versus applied voltage measured at room temperature (294 K) for a SWNT film-GaAs MSM device with W = S = 15 m and FL = FW = 300 m. The data clearly exhibit the characteristic I V curves of two back-toback Schottky diodes making up the MSM photodetector, demonstrati ng that the SWNT film indeed makes a Schottky contact to GaAs.

PAGE 106

106 Figure 7-4. Dark current of fabricated MSM devi ce. A) Dark current vers us applied voltage at room temperature (294 K) for SW NT film-GaAs MSM devices with W = 5 m and FL = FW = 400 m, but with spacing S ranging from 10 to 20 m, as labeled in the figure. B) Dark current versus applied voltage for SWNT film-GaAs MSM devices with FL = FW = 300 m, but with W = S ranging from 15 to 30 m, as labeled in the figure. C) MEDICI simulation of the cross-sectional curren t density distribution in the GaAs substrate between two SW NT film electrode fingers ( W = S = 20 m) at V = 3 V bias. Darker colors correspond to hi gher current density (courtesy of Ashkan Behnam). D) Dark current versus app lied voltage for SWNT film-GaAs MSM devices with W = 5 m and S = 15 m, but with FL = FW ranging from 200 to 400 m, as labeled. B A

PAGE 107

107 Figure 7-4. Continued C D

PAGE 108

108 CHAPTER 8 CONCLUSIONS AND FUTURE WORK Conclusions This dissertation is mainly divided into two subjects. The first subject is nucleating the growth of nanomaterials by the ion implantation technique. We have shown experimental evidence that single-walled car bon nanotubes (SWNT) and SiOX and GaN nanowires can be grown by Fe ion implantation into SiO2/Si substrates, subsequent annealing, and CVD growth. Moreover, we have shown that there is a dose and energy window of ion implantation in which SWNT and nanowire growth are observed for a given growth condition. For nanomaterials growth, catalyst is usually spun on or drop-dr ied from a liquid solution containing iron nanoparticles or deposited as solid thin film layers by evaporation or sputtering. However, it is not possible to pattern the liquid solution-based catalyst into very small dimensions or the thin film catalyst into nonplanar three-dimensional (3D) device structures, such as the sidewalls of high aspect ratio trenches. By adopting ion im plantation technique for nanomaterials growth, this thesis opens up the possibili ty of controlling the origin of nanomaterials at the nanometer scale and of integrating nanomaterials into nonpl anar 3D device structur es with precise dose control. We also have fabricated micromachined Si TEM grids for direct TEM, as well as AFM, SEM, and Raman characterization of as-gro wn nanomaterials. As a result, these micromachined TEM grids offer fast, easy, and reliable structural characterization. Furthermore these grids provide a low cost, mass producible, efficient, reliable, and versatile platform for direct TEM, AFM, SEM, and Ra man analysis of as-g rown nanomaterials, eliminating the need for any post-processing growth. The second subject is the fabr ication and characterization of single-walled carbon nanotube films which are three-dimensional films of tens of nanometers thickness, consisting of an

PAGE 109

109 interwoven mesh of single-walled carbon nanotubes We have demonstrated, for the first time, patterning of SWNT films down to submicron lateral dimensions as small as 50 nm using ebeam lithography and inductively coupled plasma (IC P) etching. This simple and efficient top-down patterning capability develope d could open up tremendous opportunities for integrating single-walled nanotube films into a wide range of electronic and optoelectronic devices. Furthermore, we have fabricated and characterized the effect of device geometry on the dark current of MSM photode tectors based on SWNT film-G aAs Schottky contacts. We have observed that dark currents of the MSM devi ces scale rationally with device geometry, such as the device active area, finger width, and finge r spacing. These results open up the possibility of integrating SWNT films as transparent and conductive Scho ttky electrodes in conventional semiconductor electronic and optoelectronic devices. Suggestions for Future Work Future work could include fabricating nanotube and nanowire devices such as field effect transistors (FETs), sensors, and light-emitting diodes (LEDs) by using ion implanted catalyst. As mentioned in the previous chapters, ion impl anted catalyst could make it possible to grow nanomaterials at desired locations and integrat e them into 3D structures. Moreover, it is possible to control density of nanomaterials during growth by adjusting the dose and energy. With these capabilities, more efficient and advanced nanotube and na nowire devices can be designed and fabricated in a controllable way. SWNT films can be integrated with conven tional semiconductors for the microfabrication of electronic and optoelectronic devices such as tr ansistors, lasers, solar cells, and photodiodes. Future work is needed to measure the photocur rent of SWNT film-GaAs MSM photodetectors to extract and compare the responsivity with conve ntional MSM devices. Moreover, future work is needed to measure the I-V characteristics of SWNT film-GaA s Schottky contacts as a function

PAGE 110

110 of temperature and their C-V characteristics in order to ex perimentally extract the Schottky barrier height. Furthermore, MSM devices with various geometries using SWNT film electrodes on other substrates such as s ilicon and germanium could be fabricated.

PAGE 111

111 APPENDIX A MASK LAYOUT Figure A-1. Mask layout for nanotube device fa brication. Mask has f our different electrode widths (7, 14, 28, and 100 m) and each width has four different electrode spacings (2, 4, 6, and 10 m).

PAGE 112

112 Figure A-2. Mask layout. A) area where electrodes have 7 m width and 2 m spacing. B) device which has 7 m width and 2 m spacing. Blue color areas are places for metals and red boxes are designed for patt erning catalysts on t op of electrodes. A B

PAGE 113

113 LIST OF REFERENCES [1] Extending the road beyond CMOS in IEEE Ci rcuits and Devices Magazine (2002), Vol. 18, p. 28. [2] Semiconductor Industry Association (SIA) International Roadmap for Semiconductors 2003 Edition, http://public.itrs.net [3] SIA girds for life after CMOS in EE Times (2004), p. 1. [4] R. Compano, Technology Roadmap for Nanoelectronics 2000). [5] S. Iijima, Nature 354, 56 (1991). [6] S. Iijima, and T. Ichihashi, Nature 363, 603 (1993). [7] R. Saito et al. Phys. Rev. B 46, 1804 (1992). [8] J. W. Mintmire, D. H. Robertson, and C. T. White, J. Phys. Chem. Solid. 54, 1835 (1993). [9] D. S. Bethune et al. Nature 363, 605 (1993). [10] A. Thess et al. Science 273, 483 (1996). [11] J. Kong et al. Nature 395, 878 (1998). [12] J. Kong, A. M. Cassell, and H. J. Dai, Chem. Phys. Lett. 292, 567 (1998). [13] M. S. Dresselhaus, G. Dresselhaus, and P. C. Eklund, Science of fullerenes and carbon nanotubes (Academic Press, San Diego, 1996). [14] R. Saito, G. Dresselhaus, and M. S. Dresselhaus, Physical properties of carbon nanotubes (Imperial College Press, London, 1998). [15] M. S. Dresselhaus, G. Dresselhaus, and P. Avouris, Carbon nanotubes : synthesis, structure, properties, and applications (Springer, New York, 2000). [16] A. Jorio et al. Phys. Rev. Lett. 86, 1118 (2001). [17] S. J. Tans, R. M. Verschueren, and C. Dekker, Nature 393, 49 (1998). [18] C. Dekker, Phys. Today 52, 22 (1999). [19] H. W. C. Postma et al. Science 293, 76 (2001). [20] P. L. McEuen, Phys. World 13, 31 (2000).

PAGE 114

114 [21] P. L. McEuen, M. S. Fuhrer, and P. Hongkun, Ieee Trans. Nanotechnol. 1, 78 (2002). [22] P. Kim, and C. M. Lieber, Science 286, 2148 (1999). [23] T. Rueckes et al. Science 289, 94 (2000). [24] J. H. Hafner et al. J. Phys. Chem. B 105, 743 (2001). [25] J. Appenzeller et al. Ieee Trans. Nanotechnol. 1, 184 (2002). [26] A. Javey et al. Nat. Mater. 1, 241 (2002). [27] C. M. Lieber, Mrs Bull. 28, 486 (2003). [28] M. Dresselhaus, G. Dr esselhaus, and P. Avouris, Carbon Nanotubes: Synthesis, Structure, Properties and Applications (Springer, 2001). [29] P. L. McEuen, and J. Y. Park, Mrs Bull. 29, 272 (2004). [30] S. J. Pearton, GaN and Related Materials (Gordon and Breach, New York, 1997). [31] T. Ozel et al. Nano Lett. 5, 905 (2005). [32] R. Seidel et al. Nano Lett. 4, 831 (2004). [33] K. Bradley, J. C. P. Gabriel, and G. Gruner, Nano Lett. 3, 1353 (2003). [34] E. S. Snow et al. Appl. Phys. Lett. 86, 033105 (2005). [35] G. B. Blanchet et al. Appl. Phys. Lett. 85, 828 (2004). [36] L. Valentini et al. Appl. Phys. Lett. 82, 961 (2003). [37] J. Oakley et al. Nanotechnology 16, 2218 (2005). [38] J. P. Novak et al. Appl. Phys. Lett. 83, 4026 (2003). [39] E. S. Snow et al. Science 307, 1942 (2005). [40] K. Lee et al. Nano Lett. 4, 911 (2004). [41] A. D. Pasquier et al. Appl. Phys. Lett. 87, 203511 (2005). [42] C. M. Aguirre et al. Appl. Phys. Lett. 88, 183104 (2006). [43] M. W. Rowell et al. Appl. Phys. Lett. 88, 233506 (2006).

PAGE 115

115 [44] J. W. Mintmire, B. I. Dunlap and C. T. White, Phys. Rev. Lett. 68, 631 (1992). [45] T. Guo et al. Chem. Phys. Lett. 243, 49 (1995). [46] S. Nakamura, S. J. Pearton, and G. Fasol, The Blue Laser Diode: The Complete Story (Springer-Verlad, Berlin, 2000). [47] H. Morkoc, Nitride Semiconductors and Devices (Springer-Verlag, Berlin, 1999). [48] J. K. Sheu et al. Ieee Photonic Tech. L. 15, 18 (2003). [49] J. H. Jang et al. Inst. Phys. Conf. Ser. 184, 295 (2005). [50] C. B. Xiong et al. J. Lumin. 122, 185 (2007). [51] W. Q. Fang et al. J. Lumin. 126, 636 (2007). [52] S. C. Hsu, and C. Y. Liu, Electrochem. Solid St. 9, G171 (2006). [53] P. Waltereit et al. Nature 406, 865 (2000). [54] S. Nakamura, and S. F. Chichibu, Introduction to Nitride Semiconductor Blue Lasers and Light-Emitting Diodes (Taylor and Francis, London, 2000). [55] F. A. Ponce, and D. P. Bour, Nature 386, 351 (1997). [56] H. Morkoc, and S. N. Mohammad, Science 267, 51 (1995). [57] E. J. Kang et al. Electrochem. Solid St. 8, G327 (2005). [58] Y. Taniyasu, M. Kasu, and T. Makimoto, Nature 441, 325 (2006). [59] J. Han et al. Appl. Phys. Lett. 73, 1688 (1998). [60] H. Hirayama et al. Appl. Phys. Lett. 80, 37 (2002). [61] T. Wang et al. J. Cryst. Growth 235, 177 (2002). [62] S. Ito et al. Phys. Status. Solidi. A 204, 2073 (2007). [63] Y. Ohba, T. Gotoda, and K. Kaneko, J. Cryst. Growth 298, 682 (2007). [64] J. T. Chu et al. Jpn. J. Appl. Phys. 1 45, 2556 (2006). [65] S. N. Lee et al. Appl. Phys. Lett. 88, 111101 (2006).

PAGE 116

116 [66] C. Skierbiszewski, Acta. Phys. Pol. A 108, 635 (2005). [67] Y. Ohba, and S. Iida, Physi ca Status Solidi a-Applied Research 200, 126 (2003). [68] T. Miyajima et al. J. Phys-Condens. Mat. 13, 7099 (2001). [69] S. Nakamura, Ieee J. Sel. Top. Quant. 3, 435 (1997). [70] C. K. Wang et al. Ieee T. Electron Dev. 53, 38 (2006). [71] D. W. Kim et al. Physica Status Solidi a-Applied Research 201, 2686 (2004). [72] C. Pernot et al. Jpn. J. Appl. Phys. 2 39, L387 (2000). [73] H. Jiang et al. Jpn. J. Appl. Phys. 2 41, L34 (2002). [74] H. Jiang et al. Jpn. J. Appl. Phys. 2 40, L505 (2001). [75] K. P. Korona et al. Acta. Phys. Pol. A 110, 211 (2006). [76] J. B. Limb et al. Appl. Phys. Lett. 89, 011112 (2006). [77] S. Nakamura, IEICE Transactions on Electronics E83C, 529 (2000). [78] S. J. Pearton et al. Mat. Sci. Eng. B-Solid. 82, 227 (2001). [79] A. P. Zhang et al. Appl. Phys. Lett. 78, 823 (2001). [80] J. W. Johnson et al. Ieee T. Electron Dev. 49, 32 (2002). [81] J. W. Johnson et al. Solid. State. Electron. 46, 911 (2002). [82] P. Kung et al. J. Eur. Ceram. Soc. 17, 1781 (1997). [83] S. Keller et al. Mater. Sci. Forum 264-2, 1157 (1998). [84] M. Losurdo, P. Capezzuto, and G. Bruno, Mrs Internet. J. N. S. R. 4 (1999). [85] D. W. Kum, and D. J. Byun, J. Electron. Mater. 26, 1098 (1997). [86] K. T. Liu et al. J. Cryst. Growth 263, 400 (2004). [87] V. V. Mamutin et al. J. Cryst. Growth 202, 346 (1999). [88] M. A. Sanchez-Garcia et al. J. Cryst. Growth 202, 415 (1999).

PAGE 117

117 [89] R. F. Davis et al. Solid. State. Electron. 41, 129 (1997). [90] T. Weimann et al. Microelectron. Eng. 57-8, 903 (2001). [91] R. S. Q. Fareed et al. Mater. Chem. Phys. 64, 260 (2000). [92] H. Amano, and I. Akasaki, Opt. Mater. 19, 219 (2002). [93] S. J. Pearton et al. J. Appl. Phys. 86, 1 (1999). [94] H. J. Choi et al. Adv. Mater. 17, 1351 (2005). [95] F. Qian et al. Nano Lett. 4, 1975 (2004). [96] S. Gradecak et al. Appl. Phys. Lett. 87 (2005). [97] G. S. Cheng et al. Appl. Phys. Lett. 83, 1578 (2003). [98] Y. Huang et al. Nano Lett. 2, 101 (2002). [99] Z. H. Zhong et al. Nano Lett. 3, 343 (2003). [100] S. Han et al. Chem. Phys. Lett. 389, 176 (2004). [101] V. Dobrokhotov et al. J. Appl. Phys. 99, 104302 (2006). [102] X. F. Duan, and C. M. Lieber, J. Am. Chem. Soc. 122, 188 (2000). [103] Y. Huang et al. Science 291, 630 (2001). [104] A. Javey et al. Nano Lett. 7, 773 (2007). [105] F. Qian et al. Nano Lett. 5, 2287 (2005). [106] Y. Li et al. Nano Lett. 6, 1468 (2006). [107] Y. Li et al. Materials Today 9, 18 (2006). [108] H. J. Choi et al. J. Phys. Chem. B 107, 8721 (2003). [109] J. C. Johnson et al. Nat. Mater. 1, 106 (2002). [110] T. Y. Kim et al. J. Cryst. Growth 257, 97 (2003). [111] T. Kuykendall et al. Nano Lett. 3, 1063 (2003).

PAGE 118

118 [112] S. K. Lee et al. Phys. Status. Solidi. B 241, 2775 (2004). [113] S. K. Lee et al. Phil. Mag. 87, 2105 (2007). [114] P. J. Pauzauskie et al. J. Am. Chem. Soc. 127, 17146 (2005). [115] P. J. Pauzauskie, and P. Yang, Materials Today 9, 36 (2006). [116] E. A. Stach et al. Nano Lett. 3, 867 (2003). [117] Y. Y. Wu, and P. D. Yang, J. Am. Chem. Soc. 123, 3165 (2001). [118] P. Deb et al. Nano Lett. 5, 1847 (2005). [119] L. Delzeit et al. Chem. Phys. Lett. 348, 368 (2001). [120] computer code SRIM (www.srim.org ) [121] C. L. Cheung et al. J. Phys. Chem. B 106, 2429 (2002). [122] Y. Li et al. J. Phys. Chem. B 105, 11424 (2001). [123] Y. Li et al. Chem. Mater. 13, 1008 (2001). [124] Y. Choi et al. Nanotechnology 17, 4635 (2006). [125] Y. Choi et al. in Materials Research Society (M RS) Fall Meeting (B oston, MA, 2006). [126] Y. Choi, and A. Ural, in P hotonics West (San Jose, CA, 2007). [127] H. Kanzow, and A. Ding, Phys. Rev. B 60, 11180 (1999). [128] A. Jorio et al. New J. Phys. 5, 139 (2003). [129] A. Jorio et al. Philos. T. R. Soc. A 362, 2311 (2004). [130] H. Kataura et al. Synthetic Met. 103, 2555 (1999). [131] J. D. Plummer, M. D. Deal, and P. B. Griffin, Silicon VLSI Technology: Fundamentals, Practice, and Modeling (Prentice Hall, Upper Saddle River, NJ, 2000). [132] S. B. Chikkannanavar et al. Nano Lett. 5, 151 (2005). [133] B. W. Smith, M. Monthioux, and D. E. Luzzi, Nature 396, 323 (1998). [134] J. S. Oakley, in Physics (University of Florida, Gainesville (FL), 2005), p. 97.

PAGE 119

119 [135] J. C. Meyer et al. Appl. Phys. Lett. 85, 2911 (2004). [136] H. B. Peng et al. Appl. Phys. Lett. 83, 4238 (2003). [137] H. B. Peng, and J. A. Golovchenko, Appl. Phys. Lett. 84, 5428 (2004). [138] T. Kim et al. Appl. Phys. Lett. 87, 173108 (2005). [139] Y. H. Choi, J. Sippel-Oakley and A. Ural, Appl. Phys. Lett. 89, 153130 (2006). [140] Y. Choi et al. in Optics East (Boston, MA, 2005). [141] Y. Choi et al. in Materials Research Society (M RS) Fall Meeting (B oston, MA, 2005). [142] C. Y. Wang et al. J. Vac. Sci. Technol. B 24, 613 (2006). [143] D. P. Yu et al. Appl. Phys. Lett. 73, 3076 (1998). [144] D. K. Sood, P. K. Sekhar, and S. Bhansali, Appl. Phys. Lett. 88, 143110 (2006). [145] C. H. Liang et al. J. Non-Cryst. Solids 277, 63 (2000). [146] Z. W. Pan et al. Nano Lett. 3, 1279 (2003). [147] R. S. Wagner, and W. C. Ellis, Appl. Phys. Lett. 4, 89 (1964). [148] P. K. Sekhar et al. Nanotechnology 17, 4606 (2006). [149] H. K. Park et al. Physica. E 37, 158 (2007). [150] J. H. Hsu et al. Nanotechnology 17, 170 (2006). [151] B. Zheng et al. Adv. Mater. 14, 122 (2002). [152] Z. W. Pan et al. J. Am. Chem. Soc. 124, 1817 (2002). [153] L. Dai et al. Phys. Lett. A 335, 304 (2005). [154] S. H. Sun et al. Solid. State. Commun. 128, 287 (2003). [155] P. Wu et al. Nanotechnology 18, 125601 (2007). [156] Lacey Carbon TEM grids were used. For more information, see http://www.tedpella.com/supflm_html/suptfilm.htm [157] Y. Cui et al. Appl. Phys. Lett. 78, 2214 (2001).

PAGE 120

120 [158] K. Shimakura, T. Suzuki, and Y. Yadoiwa, Solid. State. Electron. 18, 991 (1975). [159] A. S. Grove, C. T. Sah, and O. Leistiko, J. Appl. Phys. 35, 2695 (1964). [160] J. Q. Hu et al. J. Phys. Chem. B 106, 9536 (2002). [161] W. Q. Han et al. Appl. Phys. Lett. 76, 652 (2000). [162] G. S. Cheng et al. Appl. Phys. Lett. 75, 2455 (1999). [163] K. W. Chang, and J. J. Wu, Adv. Mater. 16, 545 (2004). [164] R. Calarco et al. Nano Lett. 7, 2248 (2007). [165] M. He et al. Appl. Phys. Lett. 77, 3731 (2000). [166] Y. P. Song et al. Phys. Rev. B 69, 075304 (2004). [167] C. C. Chen et al. J. Am. Chem. Soc. 123, 2791 (2001). [168] J. Li et al. Nano Lett. 6, 148 (2006). [169] J.-R. Kim et al. Appl. Phys. Lett. 80, 3548 (2002). [170] S. Han et al. Materials Research Society 18, 245 (2003). [171] W.-Q. Han, and A. Zettl, Appl. Phys. Lett. 80, 303 (2002). [172] Z. C. Wu et al. Science 305, 1273 (2004). [173] E. S. Snow et al. Appl. Phys. Lett. 82, 2145 (2003). [174] Y. X. Zhou, L. B. Hu, and G. Gruner, Appl. Phys. Lett. 88, 123109 (2006). [175] E. Bekyarova et al. J. Am. Chem. Soc. 127, 5990 (2005). [176] J. W. Seo et al. Ieee Photonic Tech. L. 4, 888 (1992). [177] H. K. Kim et al. J. Vac. Sci. Technol. B 21, 1273 (2003). [178] S. X. Lu, and B. Panchapakesan, Appl. Phys. Lett. 88, 253107 (2006). [179] A. K. Okyay, C. O. Chui, and K. C. Saraswat, Appl. Phys. Lett. 88, 063506 (2006). [180] P. Kapur, and K. C. Saraswat, in In ternational Interconnect Technology Conference (2002).

PAGE 121

121 [181] A. Behnam et al. Appl. Phys. Lett. 89, 093107 (2006). [182] A. Behnam et al. J. Vac. Sci. Technol. B 25, 348 (2007). [183] W. A. Wohlmuth, P. Fay, and I. Adesida, Ieee Photonic Tech. L. 8, 1061 (1996). [184] A. Behnam, and A. Ural, Phys. Rev. B 75 (2007). [185] S. M. Sze, D. J. Coleman, and A. Loya, Solid. State. Electron. 14, 1209 (1971). [186] A. Behnam et al. J. Appl. Phys. 103 (2008). [187] A. Behnam et al. Appl. Phys. Lett. 92 (2008). [188] M. Ito, and O. Wada, Ieee J. Quantum Elect. 22, 1073 (1986). [189] M. Klingenstein et al. Solid. State. Electron. 37, 333 (1994). [190] J. Burm, and L. F. Eastman, Ieee Photonic Tech. L. 8, 113 (1996). [191] M. Seto et al. Appl. Phys. Lett. 75, 1976 (1999). [192] MEDICI computer code version W-2004.09-0 from Synopsys Incorporation.

PAGE 122

122 BIOGRAPHICAL SKETCH Yongho Choi was born in 1974, in Daegu, Sout h Korea. He received his bachelors degree in electronic engineeri ng from Kumoh National Institute of Technology, South Korea, in 2000. He started his graduate program in electri cal and computer engineering at the University of Florida in fall 2003. He began working fo r Professor Ant Ural, studying nanotechnology. His research interests include nano-de vice fabrication and characterization.