USF Libraries
USF Digital Collections

Synthesis of nanostructures in single crystal silicon carbide by electron beam lithography

MISSING IMAGE

Material Information

Title:
Synthesis of nanostructures in single crystal silicon carbide by electron beam lithography
Physical Description:
Book
Language:
English
Creator:
Bieber, Jay A
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla.
Publication Date:

Subjects

Subjects / Keywords:
SiCNDs
low dimensional structures
nanodots
nanowires
NEMS
scanning electron microscopy
nanotechnology
Dissertations, Academic -- Engineering Science -- Masters -- USF   ( lcsh )
Genre:
government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Summary:
ABSTRACT: Nanostructures were formed on diced specimens of several silicon carbide polytypes and silicon using electron beam lithography. A general introduction to nanostructure synthesis and electron beam lithography,are presented. A scanning electron microscope was retrofitted with a commercially available electron beam lithography package and an electrostatic beam blanker to permit nanoscale lithography to be performed. A process was first developed and optimized on silicon substrates to expose, poly-methyl-methacrylate (PMMA) resist with an electron beam to make nanoscale nickel masks for reactive ion etching. The masks consist of an array of nickel dots that range in size from 20 to 100 nm in diameter. Several nanoscale structures were then fabricated in silicon carbide using electron beam lithography. The structures produced are characterized by field emission Scanning Electron Microscopy.
Thesis:
Thesis (M.S.E.S.)--University of South Florida, 2004.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Jay A. Bieber.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 79 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001469383
oclc - 55644219
notis - AJR1137
usfldc doi - E14-SFE0000284
usfldc handle - e14.284
System ID:
SFS0024979:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001469383
003 fts
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 040524s2004 flua sbm s000|0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0000284
035
(OCoLC)55644219
9
AJR1137
b SE
SFE0000284
040
FHM
c FHM
090
TA145
1 100
Bieber, Jay A.
0 245
Synthesis of nanostructures in single crystal silicon carbide by electron beam lithography
h [electronic resource] /
by Jay A. Bieber.
260
[Tampa, Fla.] :
University of South Florida,
2004.
502
Thesis (M.S.E.S.)--University of South Florida, 2004.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 79 pages.
520
ABSTRACT: Nanostructures were formed on diced specimens of several silicon carbide polytypes and silicon using electron beam lithography. A general introduction to nanostructure synthesis and electron beam lithography,are presented. A scanning electron microscope was retrofitted with a commercially available electron beam lithography package and an electrostatic beam blanker to permit nanoscale lithography to be performed. A process was first developed and optimized on silicon substrates to expose, poly-methyl-methacrylate (PMMA) resist with an electron beam to make nanoscale nickel masks for reactive ion etching. The masks consist of an array of nickel dots that range in size from 20 to 100 nm in diameter. Several nanoscale structures were then fabricated in silicon carbide using electron beam lithography. The structures produced are characterized by field emission Scanning Electron Microscopy.
590
Adviser: Saddow, Stephen E.
653
SiCNDs.
low dimensional structures.
nanodots.
nanowires.
NEMS.
scanning electron microscopy.
nanotechnology.
690
Dissertations, Academic
z USF
x Engineering Science
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.284



PAGE 1

Synthesis of Nanoscale Structures in Single Crystal Silicon Carbide by Electron Beam Lithography by Jay A. Bieber A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Engineering Science Department of Electrical Engineering College of Engineering University of South Florida Major Professor: Stephen E. Saddow, Ph.D. Wifrido A. Moreno, Ph.D. John T. Wolan, Ph.D. Date of Approval: March 22, 2004 Keywords: nanotechnology, scanning el ectron microscopy, NEMS, SiCNDs, low dimensional structures, nanodots, nanowires Copyright 2004 Jay A. Bieber

PAGE 2

Acknowledgements I would like to thank my committee members Dr. Stephen E. Saddow, Dr. Wilfrido A. Moreno, and Dr. John T. Wola n, for their continuous support, and encouragement to perform a nd complete this research. I would also like to thank the NNRC engineers Robert Tufts and Richard Everly for their support in maintaining the laborator y facilities in good operating condition on a daily basis. Without the support of the NN RC engineers this rese arch could not have been performed in a timely manner. I w ould also like to thank the NNRC laboratory assistants. In particular, the efforts of Matthew Salews ki and Gabriel Oliphant, who assisted with the characterization and the day to day opera tion of the electron microscopes and the electron beam lithogra phy system, proved invaluable. I also benefited from many fruitful hours of support, with respect to the equipment, from Dr. Joseph Nabity of JC Nabity Lithography Systems, Earl Weltmer of Scanservice Corporation and Jerry Tanner of JEOL. Many thanks go to Arati Lal for the patie nce and support exhibited while helping with proof reading and format checking of this thesis. I would also like thank several members of the silicon carbide group. They are Rachael Myers for the 3C-SiC growth on silicon, as well as Jeremy Walker and Shaila ja Rao for the nickel masking and optical lithography for these substrates.

PAGE 3

I would also like to thank Dr. Michael G. Kovac, Director of the Nanomaterials and Nanomanufacturing Research Center, (NNRC), at the Univ ersity of South Florida. Tampa Florida. Dr. Kovac supported the use of the laboratory equipment for this research, and played a key role in facil itating the purchase of the Nanometer Pattern Generation System (NPGS) which was used in this research. This system has given the NNRC the capability to produce nanoscale struct ures on a routine basis. I would also like to thank Dr. Kovac for supporting the flexib ility required in my work schedule that allowed me to attend classes and complete my research, while working as a NNRC engineer. This work was supported in part by th e DURINT program administered by the Office of Naval Research (Dr. C. Wood) under Grant N00014-0110715.

PAGE 4

i Table of Contents List of Tables................................................................................................................. .iii List of Figures................................................................................................................ ..iv Abstract....................................................................................................................... ...vii Chapter 1 Introduction.....................................................................................................1 1.1 Nanotechnology...................................................................................................1 1.2 Research Objective..............................................................................................2 1.3 Nanosynthesis of Silicon Carbide........................................................................3 1.3.1 Silicon Carbide Overview........................................................................4 1.3.2 Silicon Carbide Nanotechnology .............................................................5 1.3.3 Emerging Silicon Carbide Nanotechnology and Devices........................7 1.4 Summary..............................................................................................................8 Chapter 2 Nanopatterning of Semiconductors.................................................................9 2.1 Nanopatterning of Semiconductors.....................................................................9 2.2 Optical Lithography...........................................................................................10 2.3 Electron beam Lithography...............................................................................13 2.3.1 Converting an SEM into an Elect ron Beam Lithography System.........18 2.4 The Scanning Electron Microscope...................................................................19 2.4.1 Electron Sources....................................................................................21 2.5 Summary............................................................................................................27 Chapter 3 Synthesis of Nanoscale Structures by Electron Beam Lithography.............28 3.1 Introduction........................................................................................................28 3.2 Sample Preparation............................................................................................29 3.2.1 Resist Spinning……………………………..........................................30 3.3 Sample Holder...................................................................................................31 3.4 Electron Dose Array..........................................................................................32 3.5 Resist Processing...............................................................................................35 3.6 Reactive Ion Etching..........................................................................................35 3.7 Summary............................................................................................................40 Chapter 4 Synthesis and Characterization of Nanos cale Structures in Silicon Carbide42 4.1 Introduction........................................................................................................42

PAGE 5

ii 4.2 Electron Dose Array on Silicon.........................................................................45 4.3 PMMA Nanodots on Silicon..............................................................................46 4.4 Electron Dose Array on 6H-SiC........................................................................49 4.5 Reactive Ion Etching of 6H-SiC Wheel Pattern................................................52 4.6 Electron Beam Lithography of 4H-SiC.............................................................55 4.7 Summary............................................................................................................59 Chapter 5 Conclusion....................................................................................................61 5.1 Conclusion.........................................................................................................61 5.2 Future Work.......................................................................................................63 References..................................................................................................................... ..66

PAGE 6

iii List of Tables Table 1.1 Comparison of properties of severa l common semiconductors with SiC.........5 Table 2.1 MEMs processing techniques.........................................................................10 Table 2.2 Commonly used electron sources...................................................................21

PAGE 7

iv List of Figures Figure 1.1 SEM micrograph of the cross-sect ion of a thin SiC film deposited on a Si cantilever beam at 800 C...........................................................................6 Figure 1.2 Optical images of (a) SiCcoated and (b) uncoated poly Silicon micromachines following immersion in hot KOH solution. Note the SiC resonator is still intact..............................................................................7 Figure 2.1 Schematic of a lens aperture for a optical system........................................12 Figure 2.2 Block diagram of a t ypical commercial EBL system...................................15 Figure 2.3 A commercial electron beam writer, th e JOEL model JBX-5000LS/E (Photo courtesy of JEOL).............................................................................16 Figure 2.4 EM micrograph of 5n m lines formed by EBL on PMMA...........................17 Figure 2.5 Photograph of the JSM-840 afte r EBL conversion. The beam blanker, ion gauge, ion pump and NPGS com puter were added to the SEM for conversion to an EBL system.......................................................................18 Figure 2.6 Cross section of JS M 840 thermionic electron gun......................................24 Figure 2.7 Cross section of JSM 840 el ectron optics and specimen chamber...............25 Figure 3.1 Plot of film thicknesses vs. spinning speed for 950 MW PMMA resist......31 Figure 3.2 Aluminum sample holder develope d for the EBL described in this work...32 Figure 3.3 Wheel array test pattern for beam diagnostics, created using DesignCad...34 Figure 3.4 FE-SEM image of nickel RIE mask pattern formed on epitaxial 3CSiC grown on a (001) Si substrate................................................................36 Figure 3.5 Plot of RF power vs. etch de pth for 3C-SiC using RIE and a nickel mask..............................................................................................................37 Figure 3.6 Cross-section FE-SEM image of the 100 watt etch profile. The etched mesa depth is ~ 130 nm................................................................................38

PAGE 8

v Figure 3.7 Cross-section FE-SEM image of 400 watt etch profile. The etched mesa depth is ~1000 nm, or 1 m.................................................................38 Figure 3.8 FE-SEM image after 200 watt et ch. A corner of the nickel mask peeled up in this image to reve al the unetched 3C-SiC underneath. ...........39 Figure 3.9 Higher magnification image of Figure 7, showing a closer view of the etched and unetched areas of the 3C-SiC. Enhanced micromasking is observed near the cleaved edge of the silicon substrate...............................40 Figure 4.1 DesignCad TM drawing of na nodot test array used to measure EBL resolution in PMMA. Each of the four columns of dots receives a specified dose, which can be varied to determine the optimal dose for dot synthesis..................................................................................................43 Figure 4.2 Spot burning tec hnique in which the beam profile is improved, (from left to right), resulting in the well defined beam spot on the far right of the Figure......................................................................................................45 Figure 4.3 SEM micrograph of first succes sful liftoff of nickel wheels on a silicon substrate. As can be seen here the 452 C/cm2 produced the most complete wheel pattern and wa s therefore chosen as the most effective dose for resist exposure..................................................................46 Figure 4.4 SEM micrograph of nanodot pa ttern created in PMMA on silicon. This image was taken just after resi st development. The area electron dosage from left to right were 300, 400, 500 and 600 C/cm2 for each column of dots, respectively. .......................................................................47 Figure 4.5 High resolution SEM collage of individual dots from the pattern in Figure 4.4, after sputter coating with 10 nm of gold. The gold sputter coat is visible at this resolution and may interfere with dot metrology. ......48 Figure 4.6 Electron dose array for determ ining the exposure dose on 6H-SiC. The dose for each wheel is shown in units of C/cm2. The SEM image was taken just after resist development. ......................................................50 Figure 4.7 Wheel pattern from Figure 4.6, after deposition of 450 nm nickel and liftoff processing. This shows poor pattern fidelity and incomplete liftoff due to the thick nickel mask. .............................................................50 Figure 4.8 SEM micrograph of one of the wheels from the pattern in Figure 4.6, showing the effect of a slight tilt in the SEM sample stage and an out of focus beam................................................................................................51

PAGE 9

vi Figure 4.9 SEM image of the 425 C/cm2 dose wheel from Figure 4.6 after liftoff. This image was taken at a 45 degree angle to reveal the thickness of the nickel mask in relation to the wheel. .................................52 Figure 4.10 SEM micrograph of the wheel hub and 6H-SiC mesa from the wheel in Figure 4.9 before RIE..............................................................................53 Figure 4.11 SEM micrograph of the wheel hub and 6H-SiC mesa from the wheel in Figure 4.9 after RIE, showing extensive micromasking.........................53 Figure 4.12 Close up SEM micrograph of the wheel hub and 6H-SiC mesa from Figure 4.9 after RIE, showing the etch depth of 151 nm, as measured at 45 degrees. This corresponds to an actual depth of 213 nm. From this the etch rate was calculated to be 42.6 nm/min for 6H-SiC. ..............54 Figure 4.13 SEM micrograph of 4H-SiC wh eel dose array after EBL and liftoff, showing the best wheel at a dose of 600 C/cm2. ......................................56 Figure 4.14 SEM micrograph of 4H-SiC na nodot array after EBL and liftoff. The 600 C/cm2 dose gave good dot formation for only the 40 and 50 nm dots. ............................................................................................................56 Figure 4.15 SEM micrograph of 600 C/cm2 wheel spokes after EBL and liftoff, before RIE, showing a the thic kness of the nickel mask.............................57 Figure 4.16 SEM micrograph of 600 C/cm2 wheel hub after a 2min RIE at 100 watts, showing the thickness of the nickel mask and the etch depth in the 4H-SiC...................................................................................................57 Figure 4.17 SEM micrograph of the 600 C/cm2 40 nm nickel dots before RIE. As measured from the Figure the dot height is 36 nm................................58 Figure 4.18 SEM micrograph of the same nickel dots after RIE, showing a dot height of 35nm and base width of 55nm.....................................................59

PAGE 10

vii Synthesis of Nanoscale Structures in Single Crystal Silicon Carbide by Electron Beam Lithography Jay A. Bieber ABSTRACT Nanostructures were formed on diced specimens of several silicon carbide polytypes and silicon using electron beam lithography. A general introduction to nanostructure synthesis and electron b eam lithography, are presented. A scanning electron microscope was retrofitted with a commercially available electron beam lithography package and an electr ostatic beam blanker to permit nanoscale lithography to be performed. A process was first developed and optimized on silicon substrates to expose, polymethyl-methacrylate (PMMA) resist with an electron beam to make nanoscale nickel masks for reactive ion etching. The masks consist of an array of nickel dots that range in size from 20 to 100 nm in diameter. Several na noscale structures were then fabricated in silicon carbide using electron beam l ithography. The structures produced are characterized by field emissi on Scanning Electron Microscopy.

PAGE 11

1 Chapter 1 Introduction 1.1 Nanotechnology The prefix nano comes from the Greek word for dwarf. The term is commonly used in mathematics and is a short notation used when we want to divide a unit by one billion. In terms of a powe r of ten it is denoted by 10-9 and is a number so small it is very difficult to conceptualize. As an exampl e, lets assume a technology where discovered which allowed a telescope to be built with e nough resolving power to see a coin the size of a quarter dollar on the surface of the m oon. The moon, which has a diameter of 3,474 kilometers, subtends an arc of 0.5 degrees in the sky as observed from the surface of the earth. A quarter, which has a diameter of about 2.5 cm, on the moons surface will then subtend an angle of 3.6 x 10-9 degrees, or 3.6 “nano” degrees as seen from earth. This is the scale of nanometers (nm). Nanotechnology, th en, as its name implies, is the science of making and manipulating structures, which are measured on a scale of 10-9 meters. These “nanostructures”, are defined as groups of atoms or molecules, having dimensions between 1 and 100 nm in size. On the nanoscale, the smallest structural units of matter are atoms and molecules. As we approach the nanoscale, entirely new physical properties emerge which are governed by quantum mechanics. Once we are able to control matter on the nanoscale to take advantage of these new properties, the commercial applications are limitless. One example of a device, which is currently in large-s cale production, is the giant

PAGE 12

2 magnetoresistive (GMR) sensors used in mode rn computer hard disks. GMR sensors were developed from the field of magnetoelect ronics, which is the study of how the spin of electrons can influence elec tronic properties and transport. GMR sensors, also known as spin valves, are comprised of alte rnating layers of ferromagnetic and nonferromagnetic layers several nanometers thick. When exposed to a magnetic field these layers exhibit an increase in resistance as mu ch as 10% due to a split in the density of states available for spin up and spin down electrons [1]. In the last few years, the U. S. gove rnment has recognized the potential of nanotechnology to help the economy and for national security, by supporting the National Nanotechnology Initiative (NNI). This is a federal R&D program established to coordinate the multi-agency efforts in nanos cale science, engineering and technology. The 2005 budget for this initiative provides over $1 billion dollars which is double the amount in 2001, which was the first year of its inception. Miniaturi zation of electronic components also continues to be a major driver of work in this area. Since the 1960s, much of the work toward the development of tools and techniques to build small structures has been done by the semiconduc tor industry during the development of integrated circuit technology. As long as the demand for smaller personnel computers and electronic gadgets such as digital cameras and cell phones continues to increase, so will funding in nanoscale R&D. 1.2 Research Objective In order to manipulate and build structur es with nanometer dimensions, we must first have a device which can image objects on this size scale. The scanning electron

PAGE 13

3 microscope, or SEM, is one of the first tool s developed that is cap able of viewing and measuring nanometer size structures. Earl y work describing the construction of the scanning electron microscope was performed in Germany (Knoll, 1935; von Ardenne in 1938) as referenced by P.J Breton in “From Mi crons to Nanometres: Early Landmarks in the science of Scanning Elect ron Microscope imaging”, [2]. Modern SEMs using aberration corrected electron optics are cap able of producing images magnified to 2,000,000 times, (2,000,000X) and have a resolution of 0.6 nm [3]. It is the subject of this thesis resear ch to develop an el ectron beam lithography (EBL) process to fabricate stru ctures with dimensions less than 100 nm using an electron beam from a modern SEM. A JEOL m odel JSM-840 SEM was retrofitted with a commercially available electron beam lithogr aphy package and an electrostatic beam blanker to permit nanoscale lithography to be performed. A process was developed and optimized first on silicon substrates to e xpose poly-methyl-methacrylate (PMMA) resist with an electron beam to make nanoscale nickel masks for reactive ion etching of semiconductor crystals. The masks consist of an array of dots, which range in size from 20 to 100nm in diameter. Several of these nanos cale structures were then fabricated in single-crystal silicon carbide substrates. Th e structures produced were characterized by field emission scanning electron microscopy (FE-SEM). 1.3 Nanosynthesis of Silicon Carbide Recently there has been in tense interest in car bon nanotubes (CNTs), silicon nanowires (SiNWs) and silicon nanodots (Si NDs) because of their unique properties. Prototype nanodevices such as transistors, di odes, switches, light-emitting diodes, lasers,

PAGE 14

4 chemical and biological sensors, etc. have been fabricated from SiNWs and SiNDs [4]. Silicon nanotechnology is of particular inte rest since these structures would be compatible with and take advantage of, the large existing silicon microelectronics knowledge base. One of the main reasons silicon has dominated as a semiconductor material, is its ability to fo rm a high quality oxide of SiO2. Since SiC also forms this stable oxide, many of the processes and t echniques, which have been developed for nanofabrication in silicon, can in principle be applied to SiC. Thus, as SiC matures as a semiconductor material, it is expected it will also begin to be used more in the manufacturing of microelectr onic and nanoscale devices. 1.3.1 Silicon Carbide Overview Silicon carbide (SiC) is an emerging semi conductor material that possesses high thermal, chemical and mechanical stabilit y. SiC is a indirect wide band gap (WBG) (EG > 2eV) semiconductor, with a range of singl e crystal polytypes which are now commonly available in 2 and 3 inch wafers. The polytypes include fully cubic (3CSiC), fully hexagonal (Wurtzite ) (2H-SiC) forms, aa well as free-standing substrates of 4Had 6H-SiC which have mixed crystal symmetry. These different polytypes arise from a number of combinations of stacked Si-C layers, with partially hexagonal and cubic structure, defined with a hexagonal unit cell (such as 4H, 6H, 8H etc.), or with a Rhombohedral unit cell (such as 15R, 21R, 33R etc.) wi th a defined degree of hexagonality [5]. Except for the cubic form (3C-SiC) all the remaining polytypes have uniaxial properties. Table 1.1 compares the properties of SiC with silicon and other WBG materials [6]. The SiC nanodots for this work were fabricated on die cut from a 2

PAGE 15

5 inch 4H-SiC wafer manufactured by CREE Inc. [7]. This material has a band gap of 3.26 eV, breakdown voltage of 2.2 x 106 V/cm, thermal conductivity of 3-3.8 Watt/cm K, and a saturation electron dr ift velocity of 2 x 107 cm/sec [8]. Because of the WBG it is possible to operate SiC devices at temperatures as high as 6 50 C (glowing red hot!), as opposed to 350 C for silicon without degradation in electrical performance. Table 1.1 Comparison of prope rties of several common semi conductors with SiC [6]. Property 3C–SiC 6H–SiC Si GaAs Diamond Band g a p ( eV ) 2.22.91.11.4 5.5 Max. Temperature (C) 873 1240 300 460 1100 Melting point (C) 1800 1800 1420 1240 ? Physical stability excellent excellentgood fair very good Electron Mobility (cm2 /Vs) 1000 600 1400 8500 2200 Hole Mobility (cm2 /Vs) 40 40 600 400 1600 Breakdown voltage (105 V/cm) 4.0 4.0 0.3 0.4 10 Thermal cond. (W/cm C) 5.0 5.0 1.5 0.5 20 Sat. velocity (107 cm /s) 2.5 2.0 1.0 2.0 2.7 Dielectric constant 9.7 10.0 11.8 12.8 5.5 1.3.2 Silicon Carbide Nanotechnology As of the writing of this thesis, the work presented here on the synthesis of silicon carbide nanodots (SiCNDs) by EBL is the first reported. Most of the work involving SiC has been in making microelectromechan ical (MEM) systems in the form of micromechanical resonators for RF devi ces, SiC high power electronics and high temperature gas sensors [9]. Silicon carbide (S iC) is currently under investigation as both

PAGE 16

6 a coating and a structural material for MEMS in harsh environments, due to its superior mechanical strength, chemical stability and excellent performance in high-temperature, high-power electronic component s [10]. Figure 1.1 shows an SEM micrograph of a thin polycrystalline SiC coating on silicon usi ng chemical vapor deposition (CVD). These coatings can be used to coat existing MEMS and NEMS devices for use in high temperature applications at 800–1000 C, or ha rsh chemical environments. An example of the use of this coating to protect a silicon based mechan ical resonator is shown in Figure 1.2. Images a and b show a silicon re sonator coated and uncoated, respectively, after being dipped in KOH. As can be seen here the uncoated silicon resonator has been dissolved away while the SiC co ated device remains intact. Figure 1.1 SEM micrograph of the cross-sectio n of a thin SiC film deposited on a Si cantilever beam at 800 C [10].

PAGE 17

7 Figure 1.2 Optical images of (a) SiCcoated and (b) uncoated poly Silicon micromachines following immersion in hot KOH solution [10]. Note the SiC resonator is still intact. 1.3.3 Emerging Silicon Carbide Nanotechnology and Devices SiC based devices are currently being developed which take advantage of the many properties which make it superior to si licon based technologies. Because of its high strength and high acoustic velocity it is being developed for used in high frequency resonators [9]. These include the growth of nanowires or nanotube s for ultra-sensitive detection of chemical and biological species Adding a coating, which is sensitive to a particular biological or chemical substance, may functionalize cantilevers made from these materials. These substances can be de tected by a change in resonator frequency as a result of the mass added to the cantileve r when they react with the coating. The anisotropic properties of low-dimensional nanomate rials such as nanowires, nanorods, nanowhiskers, nanotubes, etc. are hi ghly desirable attributes in the design and fabrication of nanodevices. Nanoelectronic de vices such as transistors, logic gates, tweezers, sensors, etc. have already been built from CNTs [4]. Once SiC nanotechnology

PAGE 18

8 matures, it is expected that these same devi ces will be fabricated in SiC. Work is currently underway to produce SiC nanotubes [11], [12]. SiC nanotubes can also be functionalized and may find use in hydrogen storage applications. 1.4 Summary As discussed above, to build a materi als base for nanotechnology, one needs to control or manipulate the proper ties, size and shape of materials at the nanometer level. This will be the first topic discusse d in Chapter 2 on "Nanopatterning of Semiconductors". A brief summary of the va rious techniques used in nanopatterning of semiconductors is presented. Optical Lithograp hy is presented in order to provide the background for discussion of EBL. An overv iew of the SEM and the conversion of the SEM into an EBL system, is also presented in Chapter 2. In Chapter 3," Synthesis of Nanoscale St ructures by Electron Beam Lithography", the details of the nanopatterni ng technique used in this wo rk are presented. In this chapter the process used to prepare samples fo r EBL is presented. Work done to develop the reactive ion etch (RIE) pr ocess for SiC which is used in this research, is also described. Because of the highly desirable properties of SiC, as discussed above, this was the material of choice for this work. Chapter 4 "Synthesis and Charac terization of Nanoscale Structures in Silicon Carbide" will then focus on work performed at the University of South Florida and will cover the experiment s which were performed to optimize the process on silicon, followed by EBL work on SiC to produce SiCNDs. Chapter 5 will present conclusions and future work.

PAGE 19

9 Chapter 2 Nanopatterning of Semiconductors 2.1 Nanopatterning of Semiconductors In this chapter some of the various me thods, which are used to create nanometer size structures, will be reviewed. Most of the techniques outlined here have come about through developments in microelectronics, pa rticularly in silicon integrated circuit technology, which have taken place since the early 1960's. Recent advancement in the areas of microelectromechanical systems (MEMS) and nanoelectromechanical systems (NEMS), have spawned the development of new micromachining and patterning techniques. Some of these emer ging techniques are discussed in the sections to follow. MEMS and NEMS are now used to describe almost any miniaturized device ranging in size from millimeters to the nanomet er size scale. These devices may be made from a variety of chemical, electrical and mechanical processes. While most MEMS devices contain some type of micron size me chanical actuator, some may only have a moving charge, such as in a chemical sens or. The structures are made by repeated application of one or more of the basic processes listed in Table 2.1 and, in the order necessary to yield th e desired device. Electron beam lithography (EBL) and reac tive ion etching (R IE) are the main processes used in this work. At the presen t time EBL has proven to be one of the best methods to rapidly build nanoscale device prototypes in semiconductors.

PAGE 20

10 Before we begin, a brief introduction to opt ical lithography will be presented to outline the motivation for the development of higher resolution technologies In addition, many standard techniques and techni cal terms have been carried over to EBL from optical lithography, after many years of development fr om optical lithography, so this review serves a second important function. Table 2.1 MEMs processing techniques. Oxidation [13], [14] Op tical lithography [23] Wet and Dry Etching [15], [16] Ch emical Mechanical Polishing [24] Diffusion [17] Ion beam milling [25] Physical vapor deposition [18] Electrochemical deposition [26] Chemical vapor deposition [19] [20] Nanotemplating [27] Ion implantation [21] Elect ron beam lithography [28] Epitaxy [22] Nanoimprinting [29] 2.2 Optical Lithography Optical Lithography is a process very si milar to photography in which light is used to copy a pattern or image onto anothe r medium. It involves coating a substrate with a liquid “resist”, called photoresist, wh ich is a photoactive pol ymer. Light is then passed through a mask onto the coated s ubstrate. The substrate then undergoes a development process to produce a copy of the mask image onto the substrate. Photoresists have been used for more than a cen tury to make printing plates for reproducing

PAGE 21

11 text and illustrations. Their current use in the semiconduc tor industry evolved from their use in the printed circuit boa rd industry since the 1920's. The most widely available optical lithogra phy system is the contact printer. Many of these instruments have been donated to universities by local industry as they upgrade their equipment to the latest microelectronic fabrication tech nologies. In contact printing the mask is in direct contact with the s ubstrate during exposure. Contact printing in semiconductor and NEMS fabrication can produce structures down to about 1 m. First the substrate on which the image is to be c opied is placed on a turn table and spun at several thousand revolutions per minute. Resi st is dripped on to the wafer and wets the substrate as it is spun out to th e edge, forming an even layer, typically on the order of 1 to 2 m thick. The actual thickness of the resist de pends on its viscosity, which is a function of the amount of solvent it contains and is i nversely proportional to th e square root of the spinning speed. The spinning speed and time for a given thickness is specified by the manufacturer of the resist [30]. After spinning, the substrate is then bake d to remove the solvent, thus hardening the resist. The term resist applies because the polymer is resistan t to chemical reaction when the wafer is put through the subsequent chemical processing steps necessary to produce device structures. When the resist is ir radiated with a certain wavelength of light, it becomes soluble in a solvent called a “d eveloper.” The deve loper is commonly an aqueous base such as tetra-methyl-ammoni um hydroxide (TMAH). We call the resist “positive” when areas exposed to light en ergy rinse away to produce openings in the resist. A resist in which all areas except the exposed area ri nse away in the developer are termed “negative“.

PAGE 22

12 The resist is used to form a pattern wh ich is used to create complex electrical circuits, micro-mechanical devices and struct ures. The pattern is created by passing light of a particular wavelength through a template or “mask”, which has the desired pattern etched into it. The openings etched in the ma sk material delineate where the resist is irradiated thus producing a copy of the circ uit or microstructure desired. The mask material is typically made from a thin meta l film such as chromium deposited on glass, but any material opaque to the light may be used. In optical lithography the minimum feature size, or resolution, is determined by the diffraction limit and is given by the equation; W = k( )/NA Where and NA are the exposure wavelength and numer ical aperture of the optical lithography tool, respectively a nd k is a process dependant factor that incorporates variables in the system that are not a f unction of the numerical aperture or the wavelength. The numerical aperture is equal to nsinA, where n is the refractive index, usually 1.0 for air and 2A is the acceptance a ngle of the mask-imaging lens, as shown in Figure 2.1. Figure 2.1 Schematic of a lens aperture for a optical system. A lens aperture 2A

PAGE 23

13 The latest optical lithography systems us e proximity printing in which the mask pattern is optically projected onto the substrate. The latest proximity systems also use a step and scan approach in which only a small area of the mask is exposed at one time [31]. These systems are capable of resoluti on just under 200nm. Most integrated circuit manufacturing is currently done in the range of 250 to 350 nm. Optical lithography is considered a high through-put technique sin ce many devices can be made in parallel on the same substrate. The main disadvantage lies in the difficulty of focusing light smaller than the optical wavelengths us ed due to diffraction limitations. 2.3 Electron Beam Lithography One way to get resolution below the lim its of optical lithography is to use electrons instead of light to expose the resist. Electr on beam lithography (EBL) was derived from the Scanning Electron Microsc ope (SEM), which has been used for many years to define surface features below the li mits of optical techniques. The first SEM based EBL systems were developed in the 1960s, shortly after the discovery that the common polymer poly(methyl-methcrylate) (P MMA), or common Lucite Plexiglas TM made an excellent electron beam resist. The spot size of an electron beam can be less than 5 angstroms, but the resolution is limited due to scattering of the electrons in the resist. Sub-100nm features are routinely produced using this technique. Th e main problem, however, is that it is a serial process, which limits it use in manufacturing. To write an entire integrat ed circuit using a scanning electron beam could take hours, as opposed to a few minutes of exposure using mask-based optical lithography systems. Si milar to optical lithography systems, EBL

PAGE 24

14 systems are also very complicated and expe nsive to maintain. EBL systems require almost constant maintenance to keep them operating within specifications. The main advantage of EBL is in its ability to do rapi d prototyping with a higher resolution than is currently available with optical lithography systems. This allows research to continue and gives the opportunity to foresee new problems, which might arise as devices approach the nanoscale. Once practical prot otypes have been produced by EBL, they may be mass-produced by future high-resoluti on parallel processing techniques such as nanoscale optical lithography when they become commercially available. Figure 2.2 shows a block diagram of a t ypical commercial EBL system [28]. The column is responsible for forming and cont rolling the electron beam. Underneath the column is a vacuum chamber containing a st age for positioning the sample to the area of interest and it allows for sa mple loading and unloading. A vacuum has to be maintained from the electron gun down through the column and within the sample chamber that holds the substrate to be pattern ed. Electrons are a form of ra diation called beta particles, which are strongly attenuated by air. A vacuum system capable of maintaining a pressure of at least 10E-5 Torr is required. A set of control electronics supplies power to the beam forming lenses and the electron beam scanning and positioning deflectors. The computer system handles such diverse functions as setting up an exposure sequence, loading and unloading the sample, aligning and focusing th e electron beam and se nding pattern data to the pattern generator.

PAGE 25

15 Figure 2.2 Block diagram of a typi cal commercial EBL system [28].

PAGE 26

16 Figure 2.3 shows a picture of a typical commercial EBL system including the column, chamber and control electronics. It is designed for various applications in the research field, including GaAs FETs, optical elements, X-ray masks, Si devices and quantum effect devices that requ ire ultra fine pattern exposure. Figure 2.3 A commercial electron beam writer, the JOEL model JBX-5000LS/E (Photo courtesy of JEOL) Figure 2.4 shows a striking example of the small resolution that can be achieved with EBL [32]. Long ~ 5nm wide openings were made in positive PMMA by EBL. Then a granular Au film was thermally evaporated such that only one Au island formed across the width of each line and liftoff produced the results seen in the image. This work was done on a 200 kV Transmission Electron Microscope (TEM)/SEM which is not commonly available in most laboratories. Computer Power Supplies Control Console Electron Column Airlock Specimen Chamber / Stage

PAGE 27

17 Figure 2.4 SEM micrograph of 5nm lin es formed by EBL on PMMA [32]. An instrument such as this, operated at high accelerating voltage, is necessary to achieve this low resolution. Although this is a very specialized system, it does demonstrate the practical resolution EBL is capable of. On a more commonly available modern thermal field emission lithography system, such as the one shown in Fig. 2.3, the resolution is approximately 30nm.

PAGE 28

18 2.3.1 Converting an SEM into an Electron Beam Lithography System The EBL work in this research was pe rformed on a Japan Electro Optical Ltd. model JSM 840 Scanning Electron Microscope that was converte d into an EBL system. Figure 2.5 shows a photograph of the syst em. The conversion was done using a commercial package called the nanometer patte rn generation system (NPGS). The NPGS system is available from JC Nabity L ithography systems in Bozeman Montana. Figure 2.5 Photograph of the JSM-840 after EBL conversion. The beam blanker, ion gauge, ion pump and NPGS computer were added to the SEM for conversion to an EBL system. Ion Pump Picoammeter NPGS Computer Electron Column Airlock Control Console Ion Gauge Beam Blanker

PAGE 29

19 The beam blanker, ion gauge, ion pump and NPGS computer system were added during the conversion of the SEM into an EBL sy stem. The package consists of a desktop computer, with a scanning control board mount ed in an available PCI expansion slot on the computer motherboard. The user inte rfaces to the SEM through the scan control board via a computer program called NPGS. The original JSM 840 came with an elec tromagnetic beam blanker. This was replaced with an electrostatic beam blanke r manufactured by Scan Se rvice Corporation of Tustin California. This was done to reduce th e noise inherent in a magnetic blanker and to increase the blanking speed. An ion pump was added to the electron gun chamber to improve the vacuum. This helps to increas e the life of the elec tron gun filament and reduces contamination, which can lead to arci ng when the acceleration voltage is applied. A cold cathode ion gauge was also added to the gun chamber to allow the vacuum to be monitored. A record of the vacuum level helps to ensure the vacuum system is functioning properly. This is important to get reproducible EB L results, since the electron beam characteristics can change with fluxuations in vacuum level. To understand the complications one can encounter while attempting to write nanoscale patterns a good understanding of the SE M is required. Thus, the next sections will discuss the scanning electron microscope and electron optics. 2.4 The Scanning Electron Microscope The SEM, as its name implies, uses a beam of electrons, typically focused to below 5nm diameter, to scan the topography of a surface and produce a magnified image. The electrons in the beam are accelerated by an electrostatic potential with an energy

PAGE 30

20 range typically from ~ 0.5 to 200 kV. This gi ves the electrons the ve locity they need to be directed to an area of interest where th ey cause emission of s econdary electrons (SE) from the surface. Secondary electrons have energy in the range of 5 to 50 electron volts (eV) and can only escape from the first 10-20 atomic layers of the surface, depending on the material being studied. A secondary electr on detector, or SED plac ed in the specimen chamber collects these electrons. The intensity of the electrons is measured by the SED and displayed on a Cathode Ray Tube (CRT ). Topography is produced in an SEM image by raised areas such as small ridges or the edges of small pores that are present on the sample surface. These edges have a larger surface area and allow more secondary electrons to be emitted. Recessed areas, such as a hole, attenuate or absorb secondary electron emission, reducing the signal output from the SED. The topography thus modulates the SED output, which is then sent directly to the CRT. As the electron beam in the CRT display is scanned over its surface, the SED modulates the intensity of the electron beam in the CRT, causing areas of attenuated SE output to appear dark. The SEM electron beam is then scanned over the surface to be imaged in synchronization with the CRT beam as it scans the surface of the display. Every line and pixel scanned in the CRT ha s a one to one mapping to a spot on the surface being imaged. Any areas of high SE em ission appear as bright areas on the CRT. Magnification is simply a matter of reducing th e scanning area so it maps the area to be imaged to a large CRT screen. Assuming a CRT dimension of 10x10 cm, a 100,000x magnification would require a scan area of 0.1 meter/10E5, or 1x1 m. A typical CRT has 525 scan lines. To scan this many lines on the sample surface requires a beam small enough to accommodate 525 diameters within a 1x1 micron scan area.

PAGE 31

21 2.4.1 Electron Sources Generation of such a small beam size st arts with choosing a suitable electron source. The size of the source is important since the length of the electron column determines the final spot size. Table 2.2 lists some of the most commonly available electron sources [28]. The standard thermioni c electron source uses joule heating of a V shaped tungsten wire 5-100 um in radius heated to ~2500 to 3000 K. Lanthanum hexaboride (LaB6) has been used for ma ny years and has now become an industry standard, due to its long life and high brightness. It is typically a rod of sintered powder about 1mm in diameter with a tip machined to a few microns in radius It has a very low work function and high brightness is obtaine d by operating it at a temperature of ~ 1900K. However it requir es a vacuum of 10-8 Torr, which is usually achieved by the addition of an ion pump to the SEM system. Table 2.2 Commonly used electron sources [28]. Source type Temp. (K) Brightness (A/cm2/sr) Source size Energy spread (eV) Vacuum requirement (Torr) Work function Tungsten thermionic 2700 ~105 25 um 2-3 10-6 4.5 LaB6 1900 ~106 10 um 2-3 10-8 2.4 Thermal (Schottky) ZrO field emitter 1800 ~108 20 nm 0.9 10-9 2.7 Tungsten field emitter 273 ~109 5 nm 0.22 10-10 4.5

PAGE 32

22 The emission current density, Jc, delivered by thermionic sources depends on the temperature, T and is expr essed by the Richardson law. Jc = AcT2exp(-Ew/kT) A/cm2 (2.1) Where Ac is a constant of the material (referr ed to as the Richardson constant) and Ew is the work function. Higher temperatures deli ver greater beam current, but the tradeoff is an exponentially decreasing lifet ime due to thermal evaporation of the cathode material. Field emission sources typically consist of a tungsten rod sharpened to a point, typically 5-100 nm in radius. The sharp tip helps provide the very high electric fields needed to pull electrons out of the metal. Si ngle crystal tungsten is used because it is a mechanically strong material. In order to get the desired brightne ss in electron current the electron extraction potential is held as high as possible. In fact, the fields are held so high that the tungsten tip is at the threshold of self destruction due to mechanical stress induced by the electric field [33]. The emission current density, Jc, delive red by Field emission, depends on the electric field, E and follows th e Fowler-Nordheim equation [34]. Jc= BE2exp ( 6.8 x 107 3/2 /E) (2.2) Where B is a field-independent constant of dimensions (A/V2) and E is the applied field (V/cm).

PAGE 33

23 Although cold field emission sources have become the sources of choice in electron microscopes, they have seen little use in EBL due to their instability in output, which is more of a problem for lithography th an microscopy. The instability is caused by atoms that adsorb onto the surface of the tip [35]. This changes the work function, which results in large changes in the emission current To minimize the cu rrent fluctuations, the electron source must be operated with ion pumps in an extreme ultra high vacuum (UHV) environment, 10-10 Torr, or better, which comes with a significant increase in cost. The latest development in electron sources is the thermal field emission source. It is available in many commercial EBL system s and electron microscopes. This source combines the tungsten tip of the field emi ssion source and the heating of the thermal source. The tip is operated at a temperat ure of ~ 1000 to 1800 K, which makes it less sensitive to gas adsorption. Although "thermal field emitter" is its common name, it is more properly called a "Schottky emitter" since the electrons escape over the work function barrier by both thermal excitation and field emission. Its brightness is almost as high as a cold field emission source, with a slightly larger tip size of 20 nm and an intermediate energy spread. The tungsten is usually coated with zirconium oxide to reduce the work function. It requ ires a vacuum in the range of 10-9 Torr, which means a more expensive pump than is required for a thermal emitter. Figure 2.6 shows a cross section of a t ypical thermionic electron gun assembly which houses the source filament [36]. A ne gative bias voltage applied to the Wehnelt cylinder causes the electrons to come to a focu s just beyond a circular aperture located at its center.

PAGE 34

24 Figure 2.6 Cross section of JSM 840 thermionic electron gun [36]. Once the beam leaves the Wehnelt of the el ectron gun they are accelerated by voltage applied to the anode. After leaving the anode the beam then passes through alignment coils, which are used to reduce any astigmatis m in the beam, as shown in Figure 2.7. A large amount of astigmatism causes the beam to be elliptical rather than cylindrical and the net result is a poorly fo cused beam. The astigmatism is reduced by adjusting the X and Y stigmators on the JSM 840 control console.

PAGE 35

25 Figure 2.7 Cross section of JSM 840 elec tron optics and specimen chamber [36].

PAGE 36

26 The beam then passes through a two stage magnetic condenser lens system which demagnifies the beam and determines the final beam current. The beam then impinges upon the objective aperture, which only allows the center of the demagnified beam to pass through. As the user turns up the curr ent in the condenser lens the beam is defocused and expands. This in turn reduces the beam current to the target because a smaller fraction of the total beam current is allowed thr ough the objective aperture. The beam then goes through the coils, which scan the beam and through the objective lens, which focuses the beam on the target. The four user determined parameters which are most important for high resolution Lithography and SEM imaging are as follows; 1) High acceleration voltages yield the smallest beam size, 2) High condenser lens setting decreases the beam current and thus decreases the beam size as discussed above, 3) Smaller objective lens aperture di ameter also decreases beam size and, 4) Smallest possible target to ob jective lens working distance (WD). High acceleration voltages make the beam mo re monochromatic in energy and this reduces chromatic aberration allowing fo r a more focused beam. Choosing a small working distance by adjusting the height or z-axis of the specimen stage, reduces the distance the electron beam must travel before reaching the target. Th is is important to reduce the interaction of the beam with any stray electric or magne tic fields, which may be present near the SEM.

PAGE 37

27 2.5 Summary The limitations and benefits of both optical lithography and EBL have been presented. Both systems are high cost and expensive to maintain. The main disadvantage for EBL is that it is a serial processing technique. However, the main advantage of EBL techniques is its ability to rapidly prototype de vices at resolutions currently unattainable by op tical lithography. Commercial systems have been used to fabricate metal line widths as small as 5nm which is well into the nanoscale and far below that which can be expected to be de veloped by optical lithography, using currently available technology. As part of this work, a JOEL model JSM-840 SEM was retrofitted with a nanometer pattern generation system (NPGS) procured from J.C. Nabity lithography systems [37]. An electrostatic beam blanke r, ion pump and ion gauge were added to improve the performance of the system. A re view of SEM operation and electron sources was also presented. The details of how this system can be used to make nanoscale patterns in semiconductor materials will be presented in Chapter 3.

PAGE 38

28 Chapter 3 Synthesis of Nanoscale Structures by Electron Beam Lithography 3.1 Introduction In this chapter a general overview of th e techniques and methods used in Electron Beam Lithography (EBL) will be presented. Very subtle changes in the environment surrounding a lithography system can lead to problems, which may affect the quality of the patterns produced. Some of the genera l precautions that should be observed for successful lithography are detailed below. Fo llowing this, the general procedures used for the EBL work on silicon and silicon carbide will be presented. The first step in producing nanoscale struct ures is to make sure all sources of electromagnetic interference and mechanical vibrations in the proximity of the lithography system are reduced to a minimu m. A handheld digital gauss meter was purchased to ensure the fiel ds surrounding the electron colu mn were at or below 0.5 milligauss (mG) as specified in the JEOL JSM-840 manual. Noise from an old model CRT screen about 3-4 feet from the column measured about 3 mG prevented the first attempts at high-resolution work. After rem oving this, the field measured approximately 0.08 mG next to the column. Even if the us er simply rotates around in a swivel chair used in the lithography laboratory produced significant shifts in the SEM image at magnifications greater than 100,000X. This wa s due to the magnetism in the iron frame

PAGE 39

29 of the chair. These chairs were replaced with chairs constructed with plastic and aluminum and the problem was solved. Since no method was available to measur e mechanical vibration, pre-emptive measures were taken to reduce possible source s of vibration, including replacement of the rubber bushing connecting the turbo pump to the specimen chamber and placing the rubber feet under the column frame. The JS M-840 column has an integrated air shock system, which requires compressed air and was deemed to be in good working order. The operator must also check the instrument to be sure there is nothing touching the column, or bridging the small gap between th e column and control console which will conduct vibration. Transmitted sources of sound were also a concern. An electron microscope is a very sensitive microphone. It has been observe d that casual talking near the EBL system can deflect the beam approximately +/10nm. In order to reduce the possibility of interference from talking or vibration when the room door is opened and closed, quiet signs were placed around the microscope and on the access door to the room to indicate lithography in progress. Most of the succe ssful lithography patterns were produced after hours when the traffic around the lith ography facility was at a minimum. 3.2 Sample Preparation The samples used for initial trials of the lithography system were taken from polished 2-inch single crystal silicon wafers These were chosen for initial studies because of the high cost of silicon carbide wafers. Approximately 10 runs were done on silicon before the process was optimized enough to produce successful patterns with a

PAGE 40

30 feature size of ~20 nm in the PMMA resist To produce substrates for lithography the wafers were first cleaned by ultrasonic bath in acetone and then in 2-propanol. The SiC substrates used in the final lithography work were received as 1cm squares and were also prepared in the same manner as the silicon samples. 3.2.1 Resist Spinning The silicon wafers and SiC die were coat ed with PMMA resist, using a resist spinner. The resist manuf acturer recommended the use of a standard 950,000 molecular weight resist because it has been shown to perform well in high resolution EBL work [38]. The resist used was a 950 MW resist with a 3% Anisole solvent obtained from MicroChem Inc.{39}. Generally the higher the molecular weight resists result in higher contrast, which result in sharpe r edges. Figure 3.1 shows a sp in curve for 950MW resist. For this work a spin speed of 4000 RPM was chosen which resulted in a resist thickness of approximately 195 nm. The highest spin speed was chosen from this chart to produce the thinnest resist possible. When the electron beam strikes the surface of the resist, the electrons immediat ely begin scattering laterall y. The thin resist thickness reduces this effect and allows higher resolu tion lithography. The wafers were then baked in a convection oven at 170oC for 1 hour and cleaved into 1cm square die. The resist thickness was measured using a stylus prof ilometer, by scanning over an area where the resist was removed.

PAGE 41

31 Figure 3.1 Plot of film th icknesses vs. spinning speed fo r 950 MW PMMA resist [38] 3.3 Sample Holder After the resist has been cured, the coat ed dies are then scribed with a single stroke of a diamond scribe, with a scratch 23 mm long. The scratch is used as a marker to make it easier to locate pa tterns on the die after EBL processing. The scratch is oriented as shown in Figure 3.2. The die is then fastened with copper clips to an aluminum SEM sample holder that was made specifically for EBL work. A Faraday cup was made by drilling a small 2 mm hole into the aluminum. A 3 mm TEM grid with a 0 1000 2000 3000 4000 5000 010002000300040005000 Spin Speed (rpm)Film Thickness (nm )

PAGE 42

32 50-micron hole was then centered over the hole and glued in place with carbon paint. The Faraday cup is used to make accurate m easurements of the electron beam current. A small piece of a carbon, stub sputter coated wi th gold, was fixed in place using silver paint. The small sputtered gold particles ar e useful for focusing and adjusting the beam without exposing the resist. This gold surface then serves as a gold standard during EBL processing. Figure 3.2 Aluminum sample holder develope d for the EBL described in this work. 3.4 Electron Dose Array After the die has been prepared and lo aded into the SEM, it is ready for patterning. The beam is first carefully aligne d and focused using the gold standard on the sample holder. A beam current measurement is then made using the Faraday cup. This reading is input into the nanometer pattern generation system (NPGS) scan control software which then adjusts the scanning speed to deliver a specified electron dose [37]. Aluminum Sample Holder Copper Clips Substrate Scratch Gold Sputtered Carbon Stub. TEM Aperature Faraday Cup Carbon Paint

PAGE 43

33 The NPGS software drives the pattern genera tor card. Patterns may be input into the NPGS software from any common CAD p ackage, which supports DWG, DXF, GDSII, CIF and IGES file formats. The patterns for this work were create d using DesignCAD, which is a commercial computer-aided-design program that is integrat ed into the NPGS software [40]. Pattern designs can consist of lines of arbitrary sl ope, as well as circles, circular arcs and arbitrary filled polygons. Text, Bezier curves, cubic spline curves and elliptical arcs can also be generated and written as a series of s hort lines. Pattern elements that are to have different exposure parameters such as dose, exposure point spacing, microscope beam current, microscope magnification etc., are desi gned in different draw ing layers or with different line colors. This gives an almost unlimited number of exposure conditions within a single pattern. Good EBL practice requires running a test pattern with the same beam conditions on the die just before, or just after, a device patte rn has been written. The test pattern typically consists of an arra y of circular patterns with ra dial spokes as shown Figure 3.3 In this work an array of nine wheels we re drawn using DesignCad, with each wheel programmed to receive a specified dose. The doses were incremented on each wheel with lowest dose on the upper left wheel and the largest on the lower right. The dose is identified in the DesignCad drawing by usi ng a different color for each wheel. The wheels in Figure 3.3 are all black and shown he re just for illustration. The drawing is then imported into the NPGS software, whic h recognizes each color is to receive a different dose. The dose for each color is specified by the user in the NPGS software.

PAGE 44

34 The black dots to the lower right of each wheel are a dump point to place the beam after writing each wheel. The Wheel arra y is a useful means of determining the dose that produces the best pattern. It is al so useful for diagnosis of exposure problems such as beam stigmation, defocused beam, tilted stage, line frequency noise, etc. A series of wheel pictures which dem onstrate most of the common problems is available on the J. C. Nabity website [41]. It is necessary to place the dump point several microns away from the patterns to prevent electrons from scattering up from the substrate and therefore over exposing the pattern. The wheel diameter designed for this pattern was 10 microns and the border outline was 95 x 95 micron. Thus, the entire width of this matrix is less than the diameter of a human hair and is too small to be seen with the naked eye. Figure 3.3 Wheel array test pa ttern for beam diagnostics, cr eated using DesignCad [40].

PAGE 45

35 3.5 Resist Processing After exposing the resist with this wheel array the die is removed from the SEM and processed to develop the resist. The de veloper solution consists of 1 part methyl isobutyl ketone (MIBK), to 3 parts 2-propanol. The exposed die is rinsed for 70 seconds in the developer and then rinsed in 2-pr opanol for 20 seconds, followed by a 20 second rinse in de-ionized (DI) water. After de velopment the exposed lines delineating the patterns are dissolved away. The next step is to inspect the patterns first with optical microscopy and then with the SEM. Even though the pattern lines are usually much too small to be resolved by optical lithography, one can still see if the patterns are there or not. This also gives an initial assessment of the quality of the patt erns and can save many hours of SEM time searching for patterns that do not exist. Af ter observation in the optical microscope, the dies are then viewed in the SEM for pattern characterization and metrology. It is necessary to sputter a thin gold film on the pattern to prevent charging before SEM observation. 3.6 Reactive Ion Etching The next step in the process is to de posit a thin layer of metal through the resist pattern to form a metal mask, which is then used to transfer the pattern onto the substrate. The areas of metal covering the remaining resi st are removed by dissolving the resist in acetone by the liftoff process. In this work Reactive Ion Etching, or RIE, was used to etch th e patterns into the silicon carbide substrates. Nickel was chosen as the mask material due to its resistance to etching in the RIE process [ 42]. The thickness of the nickel can vary depending on the

PAGE 46

36 process used to transfer the pattern onto the substrate. For the be st liftoff results, the nickel should be thinner than th e resist. The fidelity of the pa ttern after liftoff degrades as the nickel becomes much thicker than the resist. RIE experiments were carried out on 3C-SiC epitaxial films grown in house on silicon substrates [43] using a Plasma Th erm PT700 RIE system. A mask designed for broad area laser diodes was patterned onto the 3C film using optical lithography [44]. An Electron beam evaporator was then used to deposit 1000 angstroms of nickel to form the mask which was used for RIE etch rate experiments. Figure 3.4 shows a low magnification FE-SEM image of the nickel pattern after liftoff processing. Figure 3.4 FE-SEM image of nickel RIE ma sk pattern formed on epitaxial 3C-SiC grown on a (001) Si substrate. The etch rate experiments were conducted at a power of 100, 200, 300 and 400 watts for 5 minutes each. The etch runs were done at a pressure of 300 mTorr, with gas

PAGE 47

37 flows of 40 sccm SF6, 20 sccm O2 and 10 sccm H2. Figure 3.5 shows a plot of the etch depth vs. power. The curve shown in this plot was produced using a polynomial fit through the four data points. The etch depths were measured in cross section using FESEM. y = 0.0005x2 + 0.3029x 2 0 50 100 150 200 250 0100200300400500 RF power (Watts)Etch rate (nm/min) Figure 3.5 Plot of RF power vs. etch dept h for 3C-SiC using RIE and a nickel mask. Figures 3.6 and 3.7 show cross sectiona l views of the 100 and 400 watt etch profiles for reference. When SiC is etched on an aluminum platen using oxygen and a fluorinated gas, columnar residues form due to microscopic masking, also known as micromasking [45]. The micromasks are formed from small amounts of aluminum, which are removed from the bottom electrode and deposited on the SiC surface.

PAGE 48

38 Figure 3.6 Cross-section FE-SEM image of the 100 watt etch profile. The etched mesa depth is ~ 130 nm. Figure 3.7 Cross-section FE-SEM image of 400 watt etch profile. The etched mesa depth is ~1000 nm, or 1 m.

PAGE 49

39 To reduce this effect, the bottom electrode on the RIE system was covered with a graphite ;plate and 10 sccm of hydrogen was added to the gas mixture. The hydrogen is added because it reacts with aluminum fo rming compounds, which are then pumped out of the system. Despite these precautions micromasking was still observed, although at a much reduced level. Figure 3.8 shows an SE M image in which a corner of the nickel mask has peeled away to reveal the 3C film underneath. Figure 3.8 FE-SEM image after 20 0 watt etch. A corner of the nickel mask peeled up in this image to reveal the unetched 3C-SiC underneath. Figure 3.9 shows a higher magnification image of this area showing the interface between the etched and non-etch ed regions. The micromasking can be seen here as small spikes near the edge of the die.

PAGE 50

40 Figure 3.9 Higher magnification image of Figur e 7, showing a closer view of the etched and unetched areas of the 3C-SiC. E nhanced micromasking is observed near the cleaved edge of the silicon substrate. 3.7 Summary The first step in the synthesis of nanoscale structures by EBL is to take precautions to create an environment near the system which is free of electromagnetic interference and mechanical vibration. In a f acility, which is not designed for EBL, it is recommended that the work be done during hou rs of low traffic to minimize coupling of noise into the electron beam. Sample preparation for nanoscale EBL consists of cleaning and dicing of substrate material, which has been spin coate d, with a thin layer of PMMA resist. The dies were first cut into 1 cm squares whic h can fit onto a SEM sample holder designed for EBL work. Before mounting to the sample holder, the dies were diamond scribed to allow registration of the EB L patterns after processing. Electron dose arrays using

PAGE 51

41 spoked wheels were then patterned onto the resist to determine the opti mal electron dose. Following this, the resist was develope d and inspected by optical and electron microscopy. The dies were coated with nick el and liftoff used to form a mask for subsequent RIE processing. Using standard optical li thography and metal liftoff processing the etch rate for 3C-SiC was determined. This etch rate was then used to predict the etch rate on nanopatterned 4H-SiC structures, formed via EBL and discussed in Chapter 4. Once the die preparation and processing steps were in pla ce, the next step was to optimize the EBL process to produce nanoscal e patterns. This subject will be presented in more detail in the next chapter.

PAGE 52

42 Chapter 4 Synthesis and Characterization of Nano scale Structures in Silicon Carbide 4.1 Introduction In this chapter, the synthesis of nanoscal e structures in silicon carbide will be discussed. First, the procedures and lithog raphy patterns used to optimize the electron beam profile will first be presented. This will be followed by characterization using scanning electron microscopy (SEM) of the electron beam lithography (EBL) runs performed in this work. The EBL runs were first performed on s ilicon to optimize the process. These processes were then perf ormed on 4Hand 6H-SiC crystals and the pattern etched into the substrate using the reactive ion etching (RIE ) process outlined in Section 3.7. For all SEM characterization a Hitachi model S-800 field emission SEM was used at a working distance of ~5mm a nd beam voltage of 25kV, unless otherwise specified. The EBL runs were done at a beam energy of 35keV, at a working distance of ~ 5mm and the lowest beam current se tting available on the JOEL JSM-840, which measures ~ 6 pA in the faraday cup. The initial EBL runs, which were conduc ted on silicon substrates, produced a dot resolution of approximately 50 nm diameter in poly(methyl-methcrylate) PMMA. In some cases PMMA nanodots less than 10 nm in diameter were produced as a result of under exposure of the resist. Figure 4.1 shows a DesignCad drawing of the dot array,

PAGE 53

43 Figure 4.1 DesignCad TM drawing of nanodot te st array used to measure EBL resolution in PMMA. Each of the four columns of dots receives a specified dose, which can be varied to determine the optimal dose for dot synthesis. which was used to determine the smallest dot or pixel, which could be produced in the PMMA resist. This drawing consists of 4 co lumns of dot patterns. The NPGS software was programmed to deliver a different dose to each column. The SEM was then used to characterize the dots to see which of the four columns produced the most accurate patterns, thus determining the most effectiv e electron dose. Within each column, four 6x6 arrays were designed with dot sizes of 20, 30, 40 and 50 nm diameter. Rectangles were drawn around the entire array and labeling wa s added to aid identif ication of individual arrays during characterization.

PAGE 54

44 During the course of the EBL process op timization, a technique was developed to optimize the beam. This involved turning the SEM electron beam scanning system off, thus putting the SEM in spot mode. In spot mo de the beam is focused to the center of the scanning area which is ~100x100 microns at 1000X magnification. The scanning, or raster width, is found by divi ding the SEM CRT dimension, in this case 0.1 meter, by the magnification. The beam is allowed to dwell in spot mode for 30 seconds, which “burns” a spot in the resist. The lowest beam curre nt possible, ~ 5pA, was used to get the most well defined beam possible. The JSM-840 spec ifications indicate a resolution of 2nm for this SEM, which would indicate a beam diam eter several times smaller than this. Assuming a 0.5 nm beam diameter, the area of the beam is 0.196 nm2 which, for 5pA of beam current gives a beam current density of approximately 2500 A/cm2. Although the total power is very low, due to the high local current density the resist undergoes a change in polymerization, which is visible und er the SEM. This gives a direct means to observe the beam profile, which can then be adjusted to give a well-defined and symmetrical beam. Figure 4.2 shows a series of spots burned in the resist. After each 30second burn the SEM beam alignment, focus a nd astigmatism are adjusted until there is no further improvement of the beam profile. As can be seen in the Figure 4.2, the beam starts out slightly out of focus and after ad justing the focus and astigmatism the spot on the far right was produced, which is smaller and more symmetrical. Once the beam has been optimized an accu rate measurement of the beam current is obtained using the Faraday cup located near the substrate on the sample holder. The SEM stage micrometers are then moved to the end of the scratch made in the substrate

PAGE 55

45 during sample preparation and a final spot is burned at this lo cation and photographed just prior to pattern writing. The beam is then blanked and the micrometers are moved to Figure 4.2 Spot burning techni que in which the beam profile is improved, (from left to right), resulting in the we ll defined beam spot on the far right of the Figure. an area 100 microns from the end of the scratc h. The pattern design is then loaded into the NPGS software and the scanning system is switched to computer control. The proper magnification is selected and the pattern is run to expose the resist. The patterned substrate is finally removed and devel oped as discussed in section 3.6. 4.2 Electron Dose Array on Silicon Figure 4.3 shows an SEM image of one of the first successful wheel patterns produced on a silicon substrate after liftoff. This pattern was used to determine the best electron dose to use for subsequent experiments. The resist thickness for this test pattern was 150 nm, which was then coated with 50 nm of nickel after patterning and resist development. As can be seen here, the mo st complete wheel was produced at a dose of 452 C/cm2.

PAGE 56

46 Figure 4.3 SEM micrograph of first successful liftoff of nickel wheels on a silicon substrate. As can be seen here the 452 C/cm2 produced the most complete wheel pattern and was therefore chosen as the most effective dose for resist exposure. 4.3 PMMA Nanodots on Silicon Figure 4.4 shows an SEM image of the na nodot pattern of Figur e 4.1 after writing the pattern in PMMA on a silicon substrat e. This image was taken just after developingand after the resist was sputter coated with a 5 nm gold film to prevent charging during SEM characterization. Th e rectangle, which was drawn around the nanodots and the text labels for each dot array were all successfully developed. The area electron dosage from left to right was 300, 400, 500 and 600 C/cm2 for each column,

PAGE 57

47 respectively. This span was chosen to center around the areal dose of 452 C/cm2 which was determined from the silicon wheel experiment above. Figure 4.4 SEM micrograph of nanodot pattern created in P MMA on silicon. This image was taken just after resist development. The area electron dosage from left to right was 300, 400, 500 and 600 C/cm2 for each column of dots, respectively. Figure 4.5 shows a collage of SEM images of one dot from each of the dot arrays of Figure 4.4, taken at a magnification of 200,000X The collage is arranged in the same order as the larger dot arrays in Figure 4.1. In order for a dot array to be considered successful, all the dots from the entire 6x6 a rray for each size and column had to be visible. Since only a few of the 20 nm dot s appeared sporadically they were not

PAGE 58

48 considered reproducible and were not included for comparison in the collage. As can be observed in Figure 4.4, the 30, 40 and 50 nm dots were produced with some variability in accuracy. The best results appear to be in the third column from the left, i.e. the 500 C/cm2 dose. Figure 4.5 High resolution SEM collage of i ndividual dots from the pattern in Figure 4.4, after sputter coating with 10 nm of gold. The gold s putter coat is visible at this resolution and may in terfere with dot metrology. Interestingly, the under exposed 300 and 400 C/cm2 dose arrays for the 30 nm dots actually produced dots which were below 10 nm in size. All these dots are difficult to measure in the SEM due to the necessity of the gold sputter coat to prevent charging. The grain size of the sputtere d gold particles tends to inte rfere with the dot metrology. This is particularly eviden t with the 40 nm dot at 500 C/cm2, where the dot appears to have been deformed as a result of being straddled by a crack in the gold film. The sputtered gold film for this image was 15 nm thick. Dot Size (nm) 30 40 50 Dose C/cm2) 300 400 500 600

PAGE 59

49 4.4 Electron Dose Array on 6H-SiC The first experiments on silicon carbide were performed on 6H-SiC and prepared as described in Section 3.2. Figures 4.6 and 4.7 show SEM micrographs of the wheel array after patterning and metal liftoff, re spectively. The dosages used for each wheel were as labeled in th e Figure in units of C/cm2. In this experiment an error was made during electron beam deposition re sulting in a nickel film wh ich was 3 times thicker than the resist. As a general rule, for successful liftoff, the nickel thickness should be the inverse of this or 1/3 the thic kness of the resist. As can be seen in Figure 4.7, the nickel film did still liftoff in areas that had a proximity to the electron beam, i.e. around the edges and the wheels. Although the electron dose in these areas was not sufficien t to dissolve the resist during development, it was enough to allow it to make it more sensi tive to dissolution in the acetone used for liftoff. This may suggest a method to assist liftoff in cases where a process calls for a nickel mask, which is thicke r, then the resist. This would be a simple matter of including an additional step, i.e. using the SEM to image over the entire area of the pattern to deliver a dose just under that needed for resist development.

PAGE 60

50 Figure 4.6 Electron dose array for determini ng the exposure dose on 6H-SiC. The dose for each wheel is shown in units of C/cm2. The SEM image was taken just after resist development. Figure 4.7 Wheel pattern from Figure 4.6, afte r deposition of 450 nm nickel and liftoff processing. This shows poor pattern fide lity and incomplete liftoff due to the thick nickel mask. 200 275 350 425 500 575 650 725 800

PAGE 61

51 Figure 4.8 shows one of the wheels of Fi gure 4.6 at higher magnification. This demonstrates the utility of the wheel patterns in beam diagnostics. As can be seen in the Figure, the line width of 92 nm on the left si de of the wheel is th inner than the119 nm line on the right edge, indicat ing the sample holder may ha ve had a slight tilt during patterning which may have defocused the beam. The spokes in this wheel were also larger than expected at 66 nm, which also in dicates an out of focus beam. Even at the very high exposure dose during the spot bur ning shown in Figure 4.2, the beam was less then 50 nm. It would be expected that the spot size in the resi st during writing should have been less than this because of the decreased exposure time. Figure 4.8 SEM micrograph of one of the wheel s from the pattern in Figure 4.6, showing the effect of a slight tilt in the SEM sample stage and an out of focus beam.

PAGE 62

52 4.5 Reactive Ion Etching of 6H-SiC Wheel Pattern As a first test of the SiC RIE pro cess the wheel pattern in Figure 4.7 was processed using RIE on 6H-SiC. This wa s done using the process parameters from section 3.7 at a power of 100 watts. Figure 4.9 shows a close-up of wheel #4 from this pattern exposed at a dose of 425 C/cm2. This image was taken at an angle of 45 degrees to show the thickness of the nickel mask in relation to the wheel. Figure 4.9 SEM image of the 425 C/cm2 dose wheel from Figure 4.6 after liftoff. This image was taken at a 45 degree angle to reveal the thickness of the nickel mask in relation to the wheel. Figures 4.10 and 4.11 show a close-up of the wheel hub before and after etching, respectively. The 6H-SiC mesa between the wheel hubs shows extensive micromasking

PAGE 63

53 Figure 4.10 SEM micrograph of the wheel hub and 6H-SiC mesa from the wheel in Figure 4.9 before RIE. Figure 4.11 SEM micrograph of the wheel hub and 6H-SiC mesa from the wheel in Figure 4.9 after RIE, showing extensive micromasking.

PAGE 64

54 due to the incomplete removal of the nickel after liftoff. This is likely due to the redeposition of the nickel from adjacent areas. Figure 4.12 shows a higher magnification image of the wheel hub where the etch depth was measured to be 213 nm. Because this image was tilted at 45 degrees, the etch depth was found by dividing the vertical measurement of 151 nm by sin (45o) or 0.707. This gives an etch rate of 42.6 nm/min which is considerably higher than the 100 watt etch on 3C-SiC. The micromasking can be seen in greater detail here and appear as small spikes on the order 20 nm in diameter. Figure 4.12 Close up SEM micrograph of th e wheel hub and 6H-SiC mesa from Figure 4.9 after RIE, showing the etch depth of 151 nm, as measured at 45 degrees. This corresponds to an actual depth of 213 nm. From this the etch rate was calculated to be 42.6 nm/min for 6H-SiC.

PAGE 65

55 4.6 Electron Beam Lithography of 4H-SiC EBL was then performed on 4H-SiC die also prepared as presented in section 3.2. Figures 4.13 and 4.14, show SEM micrographs of the wheel dose array and nanodot arrays, respectively, after EBL and liftoff. The dosage used for each wheel and column were as labeled, in units of C/cm2. For this run the 600 C/cm2 wheel appeared to have performed the best, giving the most complete wheel after liftoff. Figures 4.15 and 4.16 show the wheel spokes magnified at 150,000X, be fore and after etching. These images were taken at a tilt of 85 degrees and show no signs of micromasking, which was present on the 6H-SiC sample. This is likely due to the fact that there is now only a very small amount of nickel in the vicinity of the pa tterns which could be redeposited. The etch power for this run was also 100 watts and was a pplied for 2 min. This resulted in an etch depth of approximately 42 nm as measured us ing SEM, giving an et ch rate of ~21nm/min for the 4H-SiC. The nickel th ickness also appears to have been reduced from 37 nm to 30 nm after the etch run. The error in measur ement using the SEM can be on the order of a few nanometers, given the resolution is specif ied at 2 nm on a gold standard from the manufacturer [36]. The top edges of the ni ckel lines appear smoother and were also likely etched a few nanometers after the 2 minute etch run. As was the case in the nanodot array in silicon, only the 30, 40 and 50 nm dots were developed in the re sist on 4H-SiC. The 600 C/cm2 dose was included as one of the nanodot column doses as labeled in Figur e 4.14 and also produced the most complete nickel dot patterns. Only the 40 and 50 nm dots were developed in this run on 4H-SiC. None of the 20 or 30 nm dots survived liftoff.

PAGE 66

56 Figure 4.13 SEM micrograph of 4H-SiC wheel dose array after EBL and liftoff, showing the best wheel at a dose of 600 C/cm2. Figure 4.14 SEM micrograph of 4H-SiC na nodot array after EBL and liftoff. The 600 C/cm2 dose gave good dot formation for only the 40 and 50 nm dots. 250 300 350 400 450 500 550 600 650 200 400 600 800

PAGE 67

57 Figure 4.15 SEM micrograph of 600 C/cm2 wheel spokes after EBL and liftoff, before RIE, showing a the thickness of the nickel mask. Figure 4.16 SEM micrograph of 600 C/cm2 wheel hub after a 2min RIE at 100 watts, showing the thickness of the nickel ma sk and the etch depth in the 4H-SiC.

PAGE 68

58 Figure 4.17 shows an SEM image of the 40 nm nickel dots before etching and again show a height of 36 nm as measured on the image, which agrees well with the 37nm measurement made on the wheel spoke in Figure 4.15. Figure 4.18 shows these same nickel dots after the 2-minute etch run. The dot in the foreground shows a height of approximately 35 nm here, with a width of 55 nm at the base. Both the nickel dot and the SiC under it appear to have some taper. The taper in the nickel dot was also evident in the image before etching in Figure 4.17. The noise in these images and a few nanometers in measurement error are expected due to th e fact that these pict ures were taken at 300,000X, which is the highest magnification av ailable on the Hitachi S-800 FE-SEM. Figure 4.17 SEM micrograph of the 600 C/cm2 40 nm nickel dots before RIE. As measured from the Figure the dot height is 36 nm.

PAGE 69

59 Figure 4.18 SEM micrograph of the same nickel dots after RIE, showing a dot height of 35nm and base width of 55nm. At this magnification the slightest vibrati on or electromagnetic in terference tends to cause noise, which makes accurate metrology very difficult. To measure features less than 50 nm, an SEM with higher magnificati on and resolution capability will be needed. 4.7 Summary In the first section of this chapter, the procedures a nd lithography patterns used to optimize the electron beam profile were pr esented. Wheel and nanodot patterns were designed and used to determine the optim um electron dose needed to produce the smallest dots in PMMA resist on silicon. A spot burning technique was then developed to optimize the electron beam profile befo re pattern writing, 30 nm nanodots were successfully produced in th e PMMA at a dose of 500 C/cm2.

PAGE 70

60 These techniques were then applied to produce nanostructures on single crystal 6Hand 4H-SiC substrates. The final EBL r uns yielded nanodots, which were measured to be 40 to 50 nm in diameter using a Hitachi FE-SEM.

PAGE 71

61 Chapter 5 Conclusion 5.1 Conclusion Electron Beam Lithography (EBL) has been used to pattern 40 nanometer diameter nanodots in single crystal silicon ca rbide. In Chapter 1, an introduction to nanotechnology and the general processes used to fabricat e nanoscale structures was presented. Silicon carbide is an emerging el ectronic material and was chosen for this work because it is a wide bandgap material a nd has excellent material properties, which allow it to be used in high power and/or high temperature applications. SiC is also biocompatible, hence processing SiC material surfaces to contain nanoscale features may find applications in chemical and biological sensing and medical technology. While the applications of nano-textured SiC surfaces have yet to be fully identified, clearly the first step is to create such surfaces. Only then can technologists study the interaction of nanoscale SiC features with chemical and biological systems and matter. Many of the process techniques and termi nologies used in EBL have come from optical lithography. Since optical lithography is also the main technique used in large scale patterning of semiconductors, a general review of this technique was given in Chapter 2. A general introduction to EBL a nd scanning electron microscopy SEM, were also presented in Chapter 2. In this work an EBL system was cons tructed by retrofitting a JEOL SEM with a commercially available nanometer pattern gene ration (NPGS) system. An electrostatic

PAGE 72

62 beam blanker was also added to the system to to avoid exposing the resist between patterns. An ion pump was added to the SEM electron gun chamber to improve the vacuum, thus decreasing contamination of the electron optics and allowing longer filament life in the electron source. An i on gauge was also added to the system to monitor the vacuum level during EBL patternin g. As discussed in Chapter 3, an EBL process was developed and optimized on sili con first and then on silicon carbide to produce nanoscale structures in poly-methyl -methacrylate, (PMMA) resist. An RIE process was then developed on 3C -SiC to etch these patterns into silicon carbide crystals, using an electron beam deposited nickel mask and liftoff processing. Initial EBL test runs on s ilicon were able to produce 50 nm diameter features in PMMA resist on silicon. Pattern arrays usi ng wheel and dot structures were designed to determine the optimum electron dose, as seen in Chapter 3. After process optimization, using a spot burning technique, this was reduced to approx imately 20 nm diameter in PMMA resist on silicon. Dots as small as 10 nm in diameter were produced in resist that was under exposed. As presented in Chapter 4, these processes and patterns were then applied to silicon carbide material. After pa tterning and resist development, nickel was deposited to form a mask layer for RIE to etch the patterns into the silicon carbide surface. The resulting structures were then characterized using FE-SEM. The FE-SEM results show the successful synthesis of 4H-SiC nanodots from the 40 and 50 nm diameter dot arrays. The nanodots from the 40 nm array had a tape red profile and were measured to be 55nm at the base.

PAGE 73

63 5.2 Future Work To create reproducible structures smalle r than 40 nm diamet er, additional steps will have to be taken to reduce the interf erence and noise in the EBL facility. The Hitachi FE-SEM used to characterize the nanodots was used at its absolute maximum magnification. To accurately characterize stru ctures much smaller than this will require an electron microscope that can produce im ages with higher reso lution and less noise. Improvements can also be made in the JEOL JSM-840 SEM to increase the EBL resolution. This would involve replacement of the tungsten thermi onic emitter with a lanthnanum hexaboride (LaB6) emitter. As shown in Table 2.2 of Section 2.4.1, the brightness of a LaB6 emitter can be an order of magnitude higher than a tungsten emitter. In addition, as can be seen in Table 2.2, La B6 also has a smaller source size, which should improve the resolution for EBL Since successful 20 nm diameter dots were produced in the PMMA resist, it should be possible to synthesize structures of this size and possibly sm aller. The tapered profile seen in the silicon carbide nanodots Si CNDs in Chapter 4 may be due to the fact that the electrons immediately begin to diffuse laterally as they pass through the resist. For the EBL in this research a single 195 nm resist layer was used. As seen in the literature, it is possible a doubl e layered resist structure can yield higher resolution and could help mitigate this effect [46]. This would entail using a two layered resist using a more sensitive resist in the top layer and performing experiments in which the thickness of these layers are varied to comp ensate for electron scattering. One interesting outcome from the PMMA nanodot experiments in Section 4.3 were the 300 and 400 C/cm2 doses for the 30 nm dot array in Figure 4.5. These

PAGE 74

64 appeared to produce holes in the resist, which were less than 10 nm. These arrays were considered under exposed because they did not produce dots 30 nm in diameter as specified by design. Experiments designed to determine if this effect can be controlled to reproduce dots less than 10 nm should be conducted. Atomic force microscopy (AFM) or a higher resolution SEM than the one that was used for the metrology in this work would have to be used in this case to accura tely image and measure dots of this size. Extensive AFM studies could also be perf ormed to determine the effect of surface roughness on the resultant EBL. These studie s should be performed on the SiC surface as received, after coating with PMMA and after PMMA exposur e. AFM may be able to verify if the 10 nm dots seen in Figure 4.5 have been completely developed through to the substrate. Another area to be explored is a study of the aspect ratio possible in SiC using EBL. High aspect ratio dots or columns woul d be useful for their high surface area in variety of applications. Experiments shoul d also be performed to determine if low dimension hollow cylinders with high aspect ratios can be produced using EBL to form SiC nanotubes. These would also be useful fo r high surface area structures if they can be practically made with high aspe ct ratios and with diameters in the 10 nanometer range. As discussed in Section 1.3.3, once functionalized these high area surface area structures will be useful for such devices as chemical and biological sensors and in hydrogen storage applications. Experiments should also be performed to refine the RIE process used for etching the SiC. These would include experiments w ith a range of the RIE process parameters

PAGE 75

65 including the gas composition, pressure and RF power. Different mask materials may also be tested to increase the etch se lectivity and produce more defined edges. Other experiments will include a study of the effects of SiC polytype on low dimensional structures. The a lattice consta nt of Cubic 3C-SiC is 0.436 nm, while the a and c lattice dimensions of hexagonal 6H -SiC are 0.308 and 1.512 and for 4H-SiC, 0.308 and 1.008 nm, respectively [5]. It can be e xpected that as the dimensions of SiC nanostructures approach the nanometer range, the lattice constant size may become a factor. In the case of 6H-SiC a 10 nm diameter dot may only be 6 or 7 units across, if the c lattice dimension is pa rallel to the surface.

PAGE 76

66 References 1. Nanoelectronics and Information Technology edited by Rainer Waser, (WILEYVCH Verlag GmbH & Co. Kg aA, Weinheim, c2003) p 625. 2. P.J. Breton, “From Microns to Nanometer s: Early Landmarks in the science of Scanning Electron Microscope imaging”, Scanning Microscopy Vol. 13, No.1, 1999 (pages 1-6). 3. http://www.jeol.com/sem/semprods/jsm7700f.html. 4. Boon K. Teo, “Doing chemistry on low-dimensional silicon surfaces: silicon nanowires as platforms and templates”, Coordination Chemistry Reviews Volume 246, Issues 1-2, November 2003, Pages 229-246. 5. Process technology for si licon carbide devices, edited by Carl-Mikael Zetterling, London : INSPEC, c2002. 6. Peter Rback, Modeling of the Sublimation Growth of Silicon Carbide Crystals Ph.D. dissertation, Helsinki University of Technology, Helsinki Finland, June 1999. 7. Cree. Inc., 4600 Silicon Drive Durham, NC 27703. 8. Cree. Inc., 4600 Silicon Drive Durham, NC 27703. ( http://www.Cree.com/Pr oducts/sic_silicarb.asp ). 9. Mehran Mehregany and Christian A. Zorman, SiC MEMS: opportunities and challenges for applications in harsh environments, Thin Solid Films Volumes 355356, 1 November 1999, Pages 518-524. 10. Conrad R. Stoldt, Carlo Carraro, W. R obert Ashurst, Di Gao, Roger T. Howe and Roya Maboudian, A low-temperature CVD process for silicon carbide MEMS, Sensors and Actuators A: Physical Volumes 97-98, 1 April 2002, Pages 410-415. 11. Jean-Mario Nhut, Ricardo Vieira, Laurie Pesant, Jean-Philippe Tessonnier, Nicolas Keller, Gaby Ehret, Cuong Pham-Huu a nd Marc J. Ledoux, “Synthesis and catalytic uses of carbon and silico n carbide nanostructures ”, Catalysis Today Volume 76, Issue 1, 1 November 2002, Pages 11-32.

PAGE 77

67 12. B. Q. Wei, J. W. Ward, R. Vajtai, P. M. Ajayan, R. Ma and G. Ramanath, “Simultaneous growth of silicon ca rbide nanorods and carbon nanotubes by chemical vapor deposition ”, Chemical Physics Letters Volume 354, Issues 3-4, 12 March 2002, Pages 264-268. 13. A. Tibrewal, Oxidation of single crystal SiC in flowing plasma after-glow MS Thesis, University of South Florida, Tampa Fl, Nov. 2002. 14. A. G. Revesz and H. L. Hughes, “The structural aspects of non-crystalline SiO2 films on silicon: a review”, Journal of Non-Crystalline Solids, Volume 328, Issues 1-3, 15 October 2003, Pages 48-63. 15. K.E. Bean, “Anisotropic etching of silicon”, IEEE Trans. Electron Devices, ED-25 (1978), pp. 1185–1193. 16. V. Kandregula, Reactive Ion Etching of SiC in Fluorinated Plasmas MS Thesis, University of South Florida, Tampa Fl, Nov. 2002. 17. Silva K. Thesis, M. J. Caturla, M. D. Johnson, J. Zhu, T. Lenosky, B. Sadigh and T. Diaz de la Rubia, “Atomic scale models of ion implantation and dopant diffusion in silicon”, Thin Solid Films Volume 365, Issue 2, 17 April 2000, Pages 219-230. 18. T. M. Anklam, L. V. Berzins, D. G. Braun, C. Haynam, T. Meier and M. A. McClelland, “Evaporation rate and co mposition monitoring of electron beam physical vapor deposition processes”, Surface and Coatings Technology Volumes 76-77, Part 2, December 1995, Pages 681-686. 19. K. L. Choy, “Chemical vapour deposition of coatings”, Progress in Materials Science Volume 48, Issue 2, 2003, Pages 57-170. 20. Alain Dollet, “Multiscale modeling of CVD film growth--a review of recent works”, Surface and Coatings Technology Volumes 177-178, 30 January 2004, Pages 245-251. 21. J. S. Williams, “Ion implantation of semiconductors”, Materials Science and Engineering A Volume 253, Issues 1-2, 30 September 1998. 22. S. E. Saddow, M. Mynbaeva, M. C. D. Smith, A. N. Smirnov and V. Dimitriev, “Growth of SiC epitaxial layers on po rous surfaces of varying porosity”, Applied Surface Science Volume 184, Issues 1-4, 12 December 2001, Pages 72-78. 23. Bernard Fay, “Advanced optical lithography development, from UV to EUV”, Microelectronic Engineering Volumes 61-62, July 2002, Pages 11-24.

PAGE 78

68 24. Kuide Qin, Brij Moudgil and Chang-Won Pa rk, “A chemical mechanical polishing model incorporating both the chem ical and mechanical effects”, Thin Solid Films Volume 446, Issue 2, 15 January 2004, Pages 277-286. 25. N. P. Hung, Y. Q. Fu and M. Y. Ali, “Focused ion beam machining of silicon”, Journal of Materials Processing Technology Volume 127, Issue 2, 30 September 2002, Pages 256-260. 26. Thomas P. Niesen and Mark R. De Guire, “Review: deposition of ceramic thin films at low temperatures from aqueous solutions”, Solid State Ionics Volume 151, Issues 1-4, November 2002, Pages 61-68. 27. Liang, J., et. al., “Two-dimensional late ral superlattices of nanostructures: Nonlithographic formation by anodic membrane template“, Journal of Applied Physics v. 91 no. 4 (February 15 2002) p. 2544-6. 28. M.A. McCord, M.J. Rooks, Handbook of Microlithography, Micromachining and Microfabrication Chapter 2, Electron Beam Lithography, Vol. 1, SPIE, The International Society for Optical En gineering, Bellingham WA, 1997 pages 128252. 29. Frank Gottschalch, Thomas Hoffmann, Clivia M. Sotomayor Torres, Hubert Schulz and Hella-Christin Scheer, “Polymer i ssues in nanoimpri nting technique”, SolidState Electronics Volume 43, Issue 6, June 1999, Pages 1079-1083. 30. Shipley Company, 500 Nickerson Road Marlborough, MA 01752, USA. 31. ASML US Inc., 8555 South River Parkway Tempe, AZ 85284, USA. 32. Christopher Vieu, Franck Carcenac a nd Huguette Launois, "From Nanoto Macroscale Science and Technology”, Condensed Matter News Vol 6, Issue 3-4, p22-30, (1998). 33. Goldstein, et. al Scanning Electron Micros copy and X-ray Microanalysis Plenum Press, New York, 1981, p 30. 34. R.H. Fowler and L.W. Nordheim, Proc. R. Soc. London A 119 (1928), p. 173. 35. Todd and Rhodin, Surface Science vol. 42, 1974 p 109. 36. JEOL JSM-840 operating manual 37. J. C. Nabity Lithography systems PO Box 5354 Bozeman, MT 59717. (http://www.jcnabity.com).

PAGE 79

69 38. MicroChem Corp.1254 Chest nut Street Newton, MA 02464. ( http://www.microchem.com/products/pmma.htm ). 39. MicroChem Corp.1254 Chest nut Street Newton, MA 02464. 40. Upperspace (makers of DesignCAD, former ly ViaGrafix), Software Division, One American Way, Pryor, OK 74361, USA. 41. J. C. Nabity Lithography systems PO Box 5354 Bozeman, MT 59717. ( http://www.jcnabity.com/pictures.htm#Exposure%20Guide ). 42. A. L. Syrkin, J. M. Bluet, J. Camassel and R. Bonnot, “Reactive ion etching of 6HSiC in an ECR plasma of CF4-O2 mi xtures using both Ni and Al masks”, Materials Science and Engineering B Volume 46, Issues 1-3, April 1997, Pages 374-378. 43. Rachael Myers, CVD Growth of SiC on Novel Si Substrates Master's Thesis, University of South Florida, Tampa, FL, Oct, 2003. 44. Stephen E. Saddow, Optical Control of Microwave In tegrated Circuits Using HighSpeed Photoconductive Switches, Ph.D. Dissertation, University of Maryland, College Park MD, Dec., 1993. 45. W. Reichert, D. Stefan, et.al., “Fabri cation of Smooth, SiC Surfaces by Reactive Ion Etching using a Gr aphite Electrode”, Materials Science and Engineering B46 (1997) 190-194. 46. T. Kster, B. Hadam, J. Gondermann, B. Spangenberg, H. G. Roskos, H. Kurz, J. Brunner and G. Abstreiter, “Investigation of Si/SiGe heterostru ctures patterned by reactive ion etching”, Microelectronic Engineering Volume 30, Issues 1-4, January 1996, Pages 341-344.