University of Southern California Dissertations and Theses

Nanoelectronics based on gallium oxide and carbon nanotubes

(USC Thesis Other)

Nanoelectronics based on gallium oxide and carbon nanotubes

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content NANOELECTRONICS BASED ON GALLIUM OXIDE AND CARBON
NANOTUBES
By
Zhen Li
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Physics)
December 2021
Copyright 2021 Zhen Li
ii

Acknowledgements

First of all, I would like to express sincere gratitude to my advisor, Dr. Chongwu
Zhou, who offered me the opportunity and research platform to explore
nanomaterials and nanotechnology. Dr. Zhou encouraged and inspired me to pursue
my interest in nanoscience and nanoelectronics. His guidance armed me with critical
thinking and professional research knowledge, which are invaluable skills I gained
during the Ph.D. research period.
I am grateful to my dissertation committee members, Professor Stephan Haas,
Professor Aiichiro Nakano, Professor Wei Wu and Professor Moh El-Naggar for
their helpful discussion and instructive suggestions in my qualifying exam and
dissertation defense.
I would like to thank Prof. Michael S. Arnold, Dr. Katherine Jinkins and Sean
M. Foradori from University of Wisconsin-Madison for their research collaboration.
Also, I would like to thank Dr. Yu Cao, Dr. Xuan Cao, Dr. Yuqiang Ma, Dr.
Liang Chen, Dr. Sen Cong, Dr. Chenfei Shen, Dr. Anyi Zhang, Dr. Yihang Liu, Dr.
Qingzhou Liu, and Dr. Fanqi Wu for passing down research experience and
iii

technique skills to me. I would like to thank their guidance and help in my junior
years.
In addition, I would like to thank my colleagues and friends at University of
Southern California: Dr. Jihan Chen. Dr. Haotian Shi, Chi Xu, Mingrui Chen,
Dingzhou Cui, Zhiyuan Zhao, Zhi Cai, Bofan Zhao, Yu Wang, Dr. Donghai Zhu and
Jenny Lin, for all of the good will and help during my Ph.D. journey.
Finally, and most importantly, I would like to thank my beloved parents for
their unconditional care and support. It is your endless love that helping me to carry
on and make my accomplishments.

iv

Table of Contents

Acknowledgements .......................................................................................................................... ii
List of Figures ................................................................................................................................ vi
Abstract ............................................................................................................................................ x
Chapter 1 Introduction ................................................................................................................... 1
1.1 Introduction to Gallium Oxide ........................................................................................ 1
1.2 Introduction to Carbon Nanotubes ................................................................................. 3
1.3 Nanomaterials based Electronics .................................................................................... 7
1.4 References ........................................................................................................................ 9
Chapter 2 Quasi-two-dimensional β-Ga 2O 3 field effect transistors with large drain current and
low contact resistance via controlled formation of interfacial oxygen vacancies ...................... 15
2.1 Introduction ................................................................................................................... 15
2.2 Gallium oxide field-effect transistor device fabrication and characterization ............ 19
2.3 Systematically studied the effectiveness of argon annealing in improving the electronic
performance of β-Ga 2O 3 FETs ................................................................................................. 25
2.4 Summary ........................................................................................................................ 36
2.5 References ...................................................................................................................... 37
Chapter 3 Air-stable n-type transistors based on assembled aligned carbon nanotube arrays and
their application in complementary metal-oxide-semiconductor electronics ............................. 48
3.1 Introduction ................................................................................................................... 48
3.2 Preparation process of aligned carbon arrays .............................................................. 52
3.3 Characterization of aligned carbon nanotube arrays ................................................... 53
3.4 Fabrication and measurement of carbon nanotube devices ........................................ 55
3.5 Summary ........................................................................................................................ 73
3.6 Reference ........................................................................................................................ 74
Chapter 4 Carbon nanotube radio frequency electronics ............................................................ 84
4.1 Introduction ................................................................................................................... 84
4.2 Fabrication and measurement of CNT devices ............................................................ 89
4.3 Carbon nanotube integrated circuits............................................................................. 95
4.4 Summary ...................................................................................................................... 103
4.5 Reference ...................................................................................................................... 104
Chapter 5 Conclusions and future work .................................................................................... 108
5.1 Conclusions .................................................................................................................. 108
v

5.2 Future direction of nanoelectronics based on gallium oxide and carbon nanotube 109
5.3 References .................................................................................................................... 111
Bibliography ................................................................................................................................ 113

vi

List of Figures

Figure 1.1 Scheme of the crystal structure of β-Ga2O3. Reprinted from ref. [4]. --------------------2
Figure 1.2 (a) Breakdown electrical field versus bandgap of representative semiconductors. (b)
Theoretical limits of on-resistance versus breakdown voltage of major semiconductors. Reprinted
from ref. [6]. ---------------------------------------------------------------------------------------------------2
Figure 1.3 Scheme of the structure of carbon nanotubes. -----------------------------------------------4
Figure 1.4 Structure of single-wall carbon nanotubes. (a) Chirality map of SWNT, showing key
parameters such as the chiral index, chiral angle, and diameter (b, c) Atomic structure of (6,5) and
(9,1) SWNTs, respectively. Reprinted from ref. [18]. ---------------------------------------------------5
Figure 1.5 Schematic diagrams of armchair (a), zigzag (b), and chiral nanotubes (c). Reprinted
from ref. [20]. --------------------------------------------------------------------------------------------------6
Figure 2.1 Characteristics of β-Ga2O3 and typical β-Ga2O3 FETs. (a) AFM image and height
profile of the selected β-Ga2O3 flake. (b) TEM characterization of β-Ga2O3. (c) Scheme of a typical
β-Ga2O3 FET with Ti/Au contacts. (d) Optical image of a β-Ga2O3 FET with Ti/Au contacts. Inset
is the SEM image of the channel area. (e) Ids-Vds curves of β-Ga2O3 FET before (black curve) and
after (red curve) annealing at 300 ℃ in argon. Inset image is the enlarged Ids-Vds curve before
annealing to illustrate the Schottky behavior. (f) Ids-Vds family curves of a β-Ga2O3 FET under
high Vds after annealing. ------------------------------------------------------------------------------------23
Figure 2.2. Raman spectra of β-Ga2O3. -------------------------------------------------------------------24
Figure 2.3 Ids-Vds family curves of β-Ga2O3 FET (a), and device after leaving in air for two months
(b). -------------------------------------------------------------------------------------------------------------24
Figure 2.4 Electronic properties of β-Ga2O3 FETs with different annealing time. (a) Typical Ids-
Vg curves of a β-Ga2O3 FET with different annealing time. (b) Ids-Vg curves of a β-Ga2O3 FET
shown in logarithmic scale. Ids-Vds family curves of a β-Ga2O3 FET at different Vg before annealing
(c) and after annealing (d). Changes of on-state current (e) and corresponding changes of field
effect mobility (f) of β-Ga2O3 FETs versus different annealing time for seven devices. -----------28
Figure 2.5 Ti/Au contacts with different annealing temperature and annealing time. (a) Optical
images of Ti/Au pads before annealing and degraded Ti/Au pads after annealing. The SEM images
of Ti/Au contacts before annealing (b), after annealing at 300 ℃ in argon for 180 min (c), after
annealing at 500 ℃ in argon for 180 min (d), and after annealing at 300 ℃ in argon for 240 min
(e). (f) Ids-Vg curves of three β-Ga2O3 FETs with different annealing temperature and annealing
time. -----------------------------------------------------------------------------------------------------------29
Figure 2.6 Contact resistance of β-Ga2O3 FETs. (a) Optical image of a β-Ga2O3 FET with TLM
structure. (b) Total resistance versus channel length at different annealing time of a β-Ga2O3 FET.
vii

(c) Contact resistance versus annealing time of a β-Ga2O3 FET. (d) Contact resistance (logarithmic
scale) versus annealing time of three as-made β-Ga2O3 FETs with TLM structures. ---------------30
Figure 2.7 XPS spectra of pure β-Ga2O3, β-Ga2O3 after annealing in argon at 300 ℃ for 180
minutes, and Ti-coated (1 nm) β-Ga2O3 after annealing in argon at 300 ℃ for 180 minutes. -----32
Figure 2.8 XPS study of β-Ga2O3. Normalized Ga 2p3/2 XPS spectra and Ga 3d XPS spectra of
pure β-Ga2O3 (a), β-Ga2O3 after annealing in argon at 300 ℃ for 180 minutes (b), and Ti-coated
(1 nm) β-Ga2O3 after annealing in argon at 300 ℃ for 180 minutes (c). Black dots are experimental
data, red curves are simulated fitting curves. (d) Scheme of free energy of different metal oxides.
(e) Schematic diagram of proposed model for oxygen vacancies at the Ti/β-Ga2O3 interface. ----35
Figure 2.9 Ti 2p core-level XPS spectra for the Ti-coated (1 nm) β-Ga2O3 after annealing in argon
at 300 ℃ for 180 minutes. Black curves are experimental data; red curves are simulated fitting
curves. We can only observe three peaks: one at the binding energy of ~ 458.9 eV corresponds to
Ti
4+
, one at the binding energy of ~ 457.5 eV corresponds to Ti
3+
, and one at the binding energy
of ~ 456 eV corresponds to Ti
2+
. These peaks confirmed the existence of Magneli Phases in Ti-O
system. --------------------------------------------------------------------------------------------------------36
Figure 3.1 Characteristics of aligned carbon nanotube arrays. (a, b) SEM images of the aligned
tubes of a larger area (a) and aligned region (b). Inset is the Raman spectra of aligned carbon
nanotube arrays. (c) SEM image of the channel area of a CNT FET (Lch = 100 nm). (d, e) AFM
image and height profile of aligned nanotube array. The range of the color scalebar to the right of
the AFM image in (d) is 7 nm. (f) Optical image of the as-made CMOS inverter based on aligned
carbon nanotube arrays. -------------------------------------------------------------------------------------54
Figure 3.2 Electronic properties of CNT FETs based on aligned nanotube arrays. (a) Schematic
diagram of a back-gated carbon nanotube transistor. (b) Ids–Vgs curves of a typical short channel
nanotube transistor (Lch = 100 nm) in both linear (red) and log (blue) scales. (c) Ids–Vds curves of
the same transistor in (b). (d, e) Ids–Vgs curves and Ids–Vds curves of a typical long channel nanotube
transistor (Lch = 10 μm). (f) The averaged on-state current density and on/off ratio for carbon
nanotube transistors with different channel lengths. Error bars represent standard deviations from
10 FETs for each channel length. --------------------------------------------------------------------------59
Figure 3.3 Extracted carrier mobility for CNT field effect transistors (FETs) with different channel
lengths from 100 nm to 10 μm. The carrier mobility was calculated using the following equation:
μ =
𝐿 𝑊
1
𝐶 𝑜𝑥
𝑉 𝑑𝑠

𝑑 𝐼 𝑑𝑠
𝑑 𝑉 𝑔𝑠

where L and W are the channel length and width in the FET, respectively. Ids is the drain current,
Vds is the source-drain voltage, Vgs is the gate voltage, and Cox is the gate capacitance per unit area.
------------------------------------------------------------------------------------------------------------------60
Figure 3.4 Electronic properties of CNT FETs under ALD passivation. (a) Schematic diagram of
a top-gated CNT transistor. (b, c) Ids–Vgs curves of typical long channel (Lch = 10 μm, b) and short
channel (Lch = 100 nm, c) CNT transistors in both linear (red curve) and log (blue curve) scales
viii

after 40 nm Al2O3 ALD passivation. (d, e) Ids–Vgs curves of typical long channel (Lch = 10 μm, d)
and short channel (Lch = 100 nm, e) CNT transistors in both linear (red curve) and log (blue curve)
scales after 1 nm Al layer evaporation/oxidation and 40 nm Al2O3 ALD passivation. (f) Schematic
illustration of n-type conversion process. -----------------------------------------------------------------62
Figure 3.5 Air-stable performance of CNT FET. (a-b) Ids-Vds curves of an n-type behavior CNT
FET (Al + Al2O3 treatments, Pd source/drain metal contacts, Lch = 10 μm) (a), and FET after
leaving in air for three months (b). (c-d) Ids-Vds curves of an n-type behavior CNT FET (Al + Al2O3
treatments, Ti source/drain metal contacts, Lch = 100 nm) (c), and FET after leaving in air for three
months (d). ---------------------------------------------------------------------------------------------------66
Figure 3.6 Ids–Vgs transfer curves at Vds= 1 V of CNT FETs (Pd source/drain metal contacts) with
various channel lengths after 1 nm Al evaporation/oxidation and Al2O3 passivation. --------------67
Figure 3.7 Electronic properties of CNT FETs with Ti source/drain contacts. (a) Schematic
diagram of a top-gated CNT transistors with Ti source/drain contacts. (b) Ids–Vgs curves of a typical
short channel back-gated CNT transistors (Lch = 100 nm) in both linear (red) and log (blue) scales.
(c) Ids–Vgs curves of the transistor in (b) after 40 nm Al2O3 ALD passivation. (d) Ids–Vgs curves of
a typical short channel CNT FET (Lch = 100 nm) after 1 nm Al layer evaporation/oxidation and 40
nm Al2O3 ALD passivation. --------------------------------------------------------------------------------71
Figure 3.8 (a) Ids-Vgs curves of 10 n-type behavior CNT FETs (Al2O3 treatments, Ti source/drain
metal contacts, Lch = 100 nm). (b) Ids-Vgs curves of 10 n-type behavior CNT FETs (Al + Al2O3
treatments, Ti source/drain metal contacts, Lch = 100 nm). --------------------------------------------72
Figure 3.9 CMOS inverter circuit using the as-made p-type and n-type CNT FETs. (a) Ids–Vds
curves of a p-type behavior CNT FET (Al2O3 treatments, Pd source/drain metal contacts, Lch = 100
nm) and an n-type behavior CNT FET (Al + Al2O3 treatments, Ti source/drain metal contacts, L ch
= 100 nm). (b) Voltage transfer characteristics of the CMOS inverter based on aligned CNT array.
(c) Plot of inverter gain versus input voltage. ------------------------------------------------------------73
Figure 4.1 (a) Schematic diagram of a carbon nanotube FET. The fringe electrical fields from the
gate to the source and drain give rise to the parasitic capacitance. (b) A small-signal equivalent
circuit for a nanotube-based FET where gm is the transconductance, Cgs the intrinsic gate
capacitance, gd is the conductance, Cp,gs and Cp,gd are the gate–source and gate–drain parasitic
capacitances, Rp,s, Rp,d and Rgate are parasitic resistances for the source and drain and gate electrode.
Reprinted from ref. [2]. -------------------------------------------------------------------------------------86
Figure 4.2 A typical wafer map created using the AFM image to evaluate CNT density and
alignment. ----------------------------------------------------------------------------------------------------90
Figure 4.3 (a) Schematic diagram of self-aligned T-gate structure. (b) Optical image of the T-gate
channel area. (c) Cross-sectional SEM image of the T-gate structure. -------------------------------92
Figure 4.4 (a) Transfer curve and transconductance of self-aligned T-gate CNT devices. (b)
Current-gain and power-gain frequency response of the transistor in (a). (c) Output power versus
input power of the fundamental and the third order frequency. ---------------------------------------92
ix

Figure 4.5 Schematic diagram of as-made on-chip passive components and CNT RF devices. -96
Figure 4.6 (a) Optical image of the as-made passive components. (b), (c) are the impedance
frequency response of as-made passive components. ---------------------------------------------------97
Figure 4.7 Schematic diagram of AM. -------------------------------------------------------------------98
Figure 4.8 Ids-Vds relationship of two port CNT devices with different metal contacts and channel
lengths. ------------------------------------------------------------------------------------------------------101
Figure 4.9 (a) Optical image of two port CNT device. (b) Measured waveforms of input AM wave
and output information signal. ----------------------------------------------------------------------------101
Figure 4.10 (a) Circuit schematic of the designed amplifier. (b) Simulated power gain frequency
response. ----------------------------------------------------------------------------------------------------103

x

Abstract

In this dissertation, I present my work on the studying of the interfacial
reactions between metal contacts/β-Ga
2
O
3
nanomembranes to get reliable Ohmic
contacts, and the development of carbon nanotube-based field-effect transistors for
complementary metal-oxide-semiconductor and radio frequency electronic
applications. Silicon-based technology is approaching its scaling and performance
limits, but the demand for the more energy-efficient and more powerful computing
units still remains. Hence, it is required to study the physics of β-Ga
2
O
3
and carbon
nanotubes to fully realize their potential in the next-generation electronics.
Chapter 1 is an introduction of β-Ga
2
O
3
and carbon nanotubes, including their
nanostructure and electronic properties. Chapter 2 focuses on the mechanism of
interfacial reactions between metal contacts and β-Ga
2
O
3
. Chapter 3 focuses on the
carrier transport mechanism in carbon nanotube field-effect transistors. Chapter 4
reports my research achievements on carbon nanotube radio frequency electronics.
Chapter 5 summarizes my work and discusses the future research directions.
Chapter 2 demonstrates a clean method to obtain Ohmic contacts in β-Ga
2
O
3

transistors. Obtaining low contact resistance on β-Ga
2
O
3
field-effect transistors
(FETs) is difficult since reactions between β-Ga
2
O
3
and metal contacts are not fully
xi

understood. Herein, we show the importance of reactions at the metal/β-Ga
2
O
3

interface and the corresponding effects of these reactions on β-Ga
2
O
3
FET
performance. When titanium (Ti) is employed as the metal contact, annealing of β-
Ga
2
O
3
FETs in argon can effectively transform Schottky contacts into Ohmic
contacts. X-ray photoelectron spectroscopy (XPS) confirmed the formation of
oxygen vacancies at the Ti/β-Ga
2
O
3
interface after annealing.
In chapter 3, a scalable process to make n-type behavior FETs based on
assembled aligned CNT arrays is demonstrated. Fabrication of n-type field effect
transistors (FETs) based on assembled aligned CNT arrays has been challenging due
to the p-type nature of carbon nanotubes in air. Here we systematically study the
contribution of metal contacts and atomic layer deposition passivation in
determining the transistor polarity. Based on these experimental results, we report
the successful demonstration of complementary metal-oxide-semiconductor
inverters with good performance, which paves the way for realizing the promising
future of carbon nanotube nanoelectronics.
In chapter 4, we report the high-performance radio frequency transistors based
on continuous films of highly aligned CNTs. Further, we demonstrate the
implementation of a CNT integrated circuit for radio frequency applications. This
integrated circuit operates at GHz frequency regime, which successfully performs
demodulation of modulated waveform. The demonstrated circuit performance paves
xii

the way to realize complex CNT integrated circuit that can operate at practical
wireless communication systems.
The last chapter, chapter 5, is the summary and the discussion of future research
of nanoelectronics based on β-Ga
2
O
3
and carbon nanotubes.

1

Chapter 1 Introduction
1.1 Introduction to Gallium Oxide

Gallium oxide (Ga
2
O
3
) was first studied by R. Roy at 1952, who found
the existence of Ga
2
O
3
and its stability relations [1]. Till now, it is fond that
Ga
2
O
3
can occur in five different phases, α, β, γ, δ, and ε [1-3]. The monoclinic
phase β-Ga
2
O
3
is the most stable form (melting point ~ 1820
o
C) in air and
have been unveiled as a promising candidate for the next-generation
electronics [4-5]. Figure 1.1 shows the crystal structure of monoclinic β-
Ga
2
O
3
with the lattice parameters and angles (a = 12.2 Å, b = 3.0 Å, c = 5.8
Å, and β = 103.8
o
) [4]. The large value of a indicates that it is possible to
exfoliate thin-film β-Ga
2
O
3
along the (100) direction. β-Ga
2
O
3
crystal contains
two different Ga positions, one with octahedral (GaO
6
) and the other one with
tetrahedral (GaO
4
) [6], which yields to diverse anisotropic properties. With a
wide band gap ~ 4.7-4.9 eV, β-Ga
2
O
3
is predicted to exhibit large theoretical
breakdown electrical field ~ 8 MV/cm and decent Baliga’s figure of merit
>3000 (Figure 1.2). Baliga’s figure of merit (BFOM) is defined as: 𝐵𝐹𝑂𝑀 =
𝘀𝜇 𝐸 𝑐 2
, where ε, μ, E
c
are the dielectric constant, carrier mobility and critical
breakdown field. The excellent material properties of β-Ga
2
O
3
have led to
increased interest in its applications in power electronics.
2

Figure 1.1 Scheme of the crystal structure of β-Ga2O3. Reprinted from ref. [4].

Figure 1.2 (a) Breakdown electrical field versus bandgap of representative semiconductors.
(b) Theoretical limits of on-resistance versus breakdown voltage of major semiconductors.
Reprinted from ref. [6].

Moreover, power electronics requires large length dimensional devices
that can survive under extremely high voltage. But traditional materials such
as silicon suffers from small bandgap (~ 1.1 eV). Producing wide bandgap (~
3

3.3-3.4 eV) materials such as SiC or GaN still remains challenging. On the
contrary, mass production of β-Ga
2
O
3
can be achieved via multiple low-cost
growth techniques, such as edge-defined film-fed growth (EFG), floating zone
(FZ) and Czochralski methods [7-9]. Molecular beam epitaxy (MBE) and
metal organic vapor phase deposition (MOCVD) techniques are also popular
in high quality epitaxial growth of β-Ga
2
O
3
thin film for power devices [10-
11]. Owing to these great advantages

over other semiconductors, β-Ga
2
O
3
devices with large on-state currents, high breakdown voltage, decent on/off
ratios, and the potential of operating in radio frequency regime were reported
[12-14]. Therefore, β-Ga
2
O
3
is a competitive candidate which can find
applications in power electronics, radio frequency electronics, deep-
ultraviolet photodetectors, and light-emitting diodes (LEDs) [12-16].

1.2 Introduction to Carbon Nanotubes

Carbon nanotubes (CNTs) have gained intensively interest since they
were first observed by Sumio Ijima with transmission electron microscopy
(TEM) in 1991 [17]. Tremendous efforts have been devoted into the research
fields of CNTs in both fundamental physics and engineering applications.
CNTs can be considered as seamless, hollow cylinders made with carbon
atoms (Figure 1.3). As a promising one-dimensional (1D) material, it shows
4

large aspect ratio with the diameter in the order of nanometers and the length
in the order of micrometers.

Figure 1.3 Scheme of the structure of carbon nanotubes.

CNTs are categorized into single-walled CNTs (SWNTs) and multi-
walled CNTs (MWNTs). SWNTs are single layer of graphene sheet rolled up
into a honeycomb structure, and MWNTs can be considered as a coaxial
assembly of multiple SWNTs. The structure of the CNTs can be described the
chiral vector Ch, which is defined as:
𝑪 𝒉 = 𝑛 𝒂 𝟏 + 𝑚 𝒂 𝟐
𝒂 𝟏 = a(√3,0)
𝒂 𝟐 = 𝑎 (
√3
2
,
3
2
)
The a1 and a2 are the unit-based vectors of the hexagonal graphene lattice,
where a = 0.142 nm is the bond length of carbon-carbon bond. The chirality,
or chiral index (n, m) indicated the three-dimensional structure of SWNTs in
terms of carbon atoms arrangement. Figure 1.4 shows the atomic structure of
5

(6, 5) and (9, 1) SWNTs [18]. Noticeably, the electronic band structures and
many properties of SWNTs are determined by their chirality. For instance,
SWNTs will show metallic behavior if n - m = 3k, where k is an integer.
Otherwise SWNTs will behave as semiconductors. And the band gaps of
semiconducting SWNTs are inversely proportional to the diameter of
nanotubes [19].

Figure 1.4 Structure of single-wall carbon nanotubes. (a) Chirality map of SWNT, showing
key parameters such as the chiral index, chiral angle, and diameter (b, c) Atomic structure
of (6,5) and (9,1) SWNTs, respectively. Reprinted from ref. [18].

The chiral angle θ is the angle between the chiral vector Ch and unit-
based vector a1:
𝜃 = tan
−1
√3𝑚 𝑚 + 2𝑛
And the diameter d of SWNTs is defined as:
𝑑 =
√3𝑎 𝜋 √
𝑛 2
+ 𝑚𝑛 + 𝑚 2

6

Depending on the value of chiral angle θ, SWNTs can be classified into
armchair (θ = 30° ), zigzag (θ = 0° ) and chiral (0° < θ < 30° ) structures. Figure
1.5 shows the schematic diagrams of armchair, zigzag and chiral SWNTs.

Figure 1.5 Schematic diagrams of armchair (a), zigzag (b), and chiral nanotubes (c).
Reprinted from ref. [20].

Considering the electrical property, Field effect transistors (FETS) that
using semiconducting CNTS as the channel material have experimentally
demonstrated intrinsic mobility > 100,000 cm
2
V
-1
s
-1
[21-22], which surpass
the state-of-art silicon technology. Such high mobility is resulted from the
long mean free path (in the orders of hundred nanometers) and the scatter-free
ballistic transport in short channel CNT FETs. Moreover, the extremely small
dimension of CNTs offers small intrinsic capacitance, which enables CNT
devices to operate at radio-frequency regime. In regarding to metallic CNTs,
7

they show large current carrying density > 10
9
A/cm
2
, which is the key
advantage of interconnects in integrated circuits. [23-24] Besides, CNTs
exhibit excellent mechanical property (Young’s module over 1 TPa and
tensile strength over 200 Gpa) and thermal conductivity [25], which make
CNTs suitable for the next-generation flexible and stretchable electronics.

1.3 Nanomaterials based Electronics
The development of power electronics demands high-performance
semiconductor devices. A lot of studies on the β-Ga
2
O
3
have been carried out
in the past few years. And many progresses, such as high performance
Schottky barrier diodes, FETs showing large on state current and high
breakdown electrical field, and radio frequency power transistors were
achieved [4, 14]. However, the relevant performance of β-Ga
2
O
3
still leaves
much to be desired when compared with SiC and GaN. The research on β-
Ga
2
O
3
is at the early stage, which requires more attention and efforts from
researchers to fully realize its potential step by step.
As silicon-based complementary metal-oxide-semiconductor (CMOS)
electronics is approaching its scaling limits, carbon nanotubes (CNTs) have
been shown the potential to become the next-generation energy-efficient
electronic. Complex integrated carbon nanotube circuits such as computers
8

have been predicted to show power performance ten time better than silicon-
based technology [26]. Over the past decade, researchers pushed forward the
progress on CNT technology and successfully demonstrade individual digital
logic gates, a mininature computer operting one bit data, and even a 16-bit
modern micprocessor based on CNT [27-29]. However, the realization of
stable high-performance nanoscale FETs based on CNT is still very
challenging, which demands fabrication process that are compatible with
standard industry process.
When nanoelectronics falls into ballistic regime, the quantum resistance
is of order h/e
2
(~ 25 kΩ). However, nanoscale devices also offer small
intrinsic capacitance, which leads to short transist times. CNTs-based FETs
have been predicted to operated at THz regime due to their small size (~ nm)
and high carrier mobility [30]. CNT FETs have been experimentally
demonstated extrinsic current-gain cut off frequecy > 80 GHz and power-gain
cut off frequency > 100 GHz [31]. It is expected that CNT FETs will surpass
the state-of-art RF-CMOS in the near future. However, integrated CNT
circuits still leave much to be desired. The lack in passive components that are
compatible with back-end-of-line process limits the development of CNT
radio frequency electronics.

9

1.4 References
[1] Roy, R.; Hill, V. G.; Osborn, E. F. Polymorphism of Ga
2
O
3
and the System
Ga
2
O
3
-H
2
O. J. Am. Chem. Soc. 1952, 74, 3, 719-722.
[2] Kroll, P; Dronskowski, R.; Matrin, M. Formation of spinel-type gallium
oxynitrides: a density-functional study of binary and ternary phases in the
system Ga–O–N. J. Mater. Chem. 2005, 15, 3296-3302.
[3] Ahman, J.; Svensson, G.; Albertsson, J. A Reinvestigation of β-Ga
2
O
3
.
Acta Cryst. 1996, C52, 1336-1338.
[4] Higashiwaki, M.; Sasaki, K.; Murakami, H.; Kumagai, Y.; Koukitu, A.;
Kuramata, A.; Masui, T.; Tamakoshi, S. Recent progress in Ga
2
O
3
power
devices. Semocond. Sci. Technol. 2016, 31 034001.
[5] Ma, N.; Tanen, N.; Verma, A.; Guo, Z.; Luo, T. F.; Xing, H. L.; Jena, D.
Intrinsic electron mobility limits in β-Ga
2
O
3
. Appl. Phys. Lett. 2016, 109,
212101.
[6] Higashiwaki, M.; Sasaki, K.; Kuramata, A.; Masui, T.; Yamakoshi, S.
Gallium oxide (Ga
2
O
3
) metal-semiconductor field-effect transistors on
signle-crystal β-Ga
2
O
3
(010) substrates. Appl. Phys. Lett. 2012, 100,
013504
10

[7] Kuramata, A.; Koshi, K.; Watanabe, S.; Yamaoka, Y.; Masui, T.;
Yamakoshi, S. High-quality β-Ga
2
O
3
single crystals grown by edge-
defined film-fed growth. Jpn. J. Appl. Phys. 2016, 55 1202A2.
[8] Zhang, J.; Li, B.; Xia, C.; Pei, G.; Deng, Q.; Yang, Z.; Xu, W.; Shi, H.;
Wu, F.; Wu, Y.; Xu, J. Growth and spectral characterization of β-Ga
2
O
3

single crystals. J. Phys. Chem. Solids 2006, 67, 12, 2448-2451.
[9] Galazka, Z.; Irmscher, K.; Uecker, R.; Bertam, R.; Pietsch, M.;
Kwasniewski, A.; Naumann, M.; Schulz, T.; Schewski, R.; Klimm, D.;
Bickermann, M. On the bulk β-Ga
2
O
3
single crystals grown by the
Czochralski method. J. Cryst. Growth 2014, 404, 15, 184-191.
[10] Ví llora, E. G.; Shimamura, K.; Kitamura, K.; Aoki, K. Rf-plasma-
assisted molecular-beam-epitaxy of β-Ga
2
O
3
. Appl. Phys. Lett. 2006, 88,
031105.
[11] Kim, H. W.; Kim, N. H. Synthesis of β-Ga
2
O
3
nanowires by an
MOCVD approach. Appl. Phys. A 2005, 81, 763–765.
[12] Higashiwaki, M.; Sasaki, K.; Kamimura, T.; Hoi Wong, M.;
Krishnamurthy, D.; Kuramata, A.; Masui, T.; Yamakoshi, S. Depletion-
mode Ga
2
O
3
metaloxide-semiconductor field-effect transistors on β-
Ga
2
O
3
(010) substrates and temperature dependence of their device
characteristics. Appl. Phys. Lett. 2013, 103, 123511.
11

[13] Zhou, H.; Maize, K.; Noh, J.; Shakouri, A.; Ye, P. D. Thermodynamic
studies of β-Ga
2
O
3
nanomembrane field-effect transistors on a sapphire
substrate. ACS Omega 2017, 2, 7723–7729
[14] Green, A. J.; Chabak, K. D.; Baldini, M.; Moser, N.; Gilbert, R.; Fitch,
R. C.; Wagner, G.; Galazka, Z.; McCandless, J.; Crespo, A. et al. β-Ga
2
O
3

MOSFETs for radio frequency operation. IEEE Electron Device Lett.
2017, 38, 790–793.
[15] Oshima, T.; Okuno, T.; Fujita, S. Ga
2
O
3
Thin Film Growth on c-Plane
Sapphire Substrates by Molecular Beam Epitaxy for Deep-Ultraviolet
Photodetectors. Jpn. J. Appl. Phys. 2007, 46 7217.
[16] Kim, C. J.; Kang, D.; Song, I.; Park, J. C.; Lim, H.; Kim, S.; Lee, E.;
Chung, R.; Lee, J. C.; Park, Y. International Electron Devices Meeting
2006, 11-13.
[17] Iijima, S. Helical microtubules of graphitic carbon. Nature 1991, 354,
56-58
[18] Liu, B.; Wu, F.; Gui, H.; Zheng, M.; Zhou, C. Chirality-Controlled
Synthesis and Applications of Single-Wall Carbon Nanotubes. Acs Nano
2017, 11, 1, 31-53.
12

[19] Kataura, H.; Kumazawa, Y.; Maniwa, Y.; Umezu, I.; Suzuki, S.;
Ohtsuka, Y.; Achiba, Y. Optical properties of single-wall carbon
nanotubes. Synthetic Met. 1999, 103, 2555-2558.
[20] Saito, R.; Dresselhaus, G.; Dresselhaus, M. S., Physical properties of
carbon nanotubes. Imperial College Press, London, 1998.
[21] Durkop, T.; Getty, S. A.; Cobas, E.; Fuhrer, M. S. Extraordinary
Mobility in Semiconducting Carbon Nanotubes. Nano Lett. 2004, 4, 35-
39.
[22] Zhou, X. J.; Park, J. Y.; Huang, S. M.; Liu, J.; McEuen, P. L. Band
Structure, Phonon Scattering, and the Performance Limit of Single-
Walled Carbon Nanotube Transistors. Phys. Rev. Lett. 2005, 95.
[23] Pop, E.; Mann, D. A.; Goodson, K. E.; Dai, H. J. Electrical and thermal
transport in metallic single-wall carbon nanotubes on insulating substrates.
J. Appl. Phys. 2007, 101.
[24] Subramaniam, C.; Yamada, T.; Kobashi, K.; Sekiguchi, A.; Futaba, D.
N.; Yumura, M.; Hata, K. One hundred fold increase in current carrying
capacity in a carbon nanotubecopper composite. Nat. Commun. 2013, 4.
[25] Treacy, M. M. J.; Ebbesen, T. W.; Gibson, J. M. Exceptionally High
Young's Modulus Observed for Individual Carbon Nanotubes. Nature
1996, 381, 678-680.
13

[26] George, S.; Franklin, A. D.; Frank, D.; Lobez, J. M.; Cao, Q.; Park, H.,
Afzali, A.; Han, S. J.; Hannon, J. B.; Haensch, W. Toward high-
performance digital logic technology with carbon nanotubes. ACS Nano
2014, 8, 8730–8745.
[27] Zhang, J.; Wang, C.; Fu, Y.; Che, Y.; Zhou, C. Air-stable conversion
of separated carbon nanotube thin-film transistors from p-type to n-type
using atomic layer deposition of high-κ oxide and its application in CMOS
logic circuits. ACS Nano 2011, 5, 3284–3292.
[28] Shulaker, M.; Hills, G.; Patil, N.; Wei, H.; Chen, H. Y.; Wong, H. S. P.;
Mitra, S. Carbon nanotube computer. Nature 2013, 501, 526–530.
[29] Hills, G.; Lau, C.; Wright, A.; Fuller, S.; Bishop, M. D.; Srimani, T.;
Kanhaiya, P.; Ho, R.; Amer, A.; Stein, Y.; Murphy, D.; Arvind,
Chandrakasan, A.; Shulaker, M. Modern microprocessor built from
complementary carbon nanotube transistors. Nature 2019, 572, 595-602.
[30] Kienle, D.; Leonard, F. Terahertz Response of Carbon Nanotube
Transistors. Phys. Rev. Lett 2009, 103, 206601.
[31] Rutherglen, C.; Kane, A. A.; Marsh, P. F.; Marsh, P. F.; Cain. T. A.;
Hassan, B. I.; Alshareef, M. R.; Zhou, C.; Galatsis, K. Wafer-scalable,
aligned carbon nanotube transistors operating at frequencies of over
100 GHz. Nat. Electron. 2019, 2, 530–539.
14

15

Chapter 2 Quasi-two-dimensional β-
Ga
2
O
3
field effect transistors with large
drain current and low contact resistance
via controlled formation of interfacial
oxygen vacancies
2.1 Introduction
β-Ga
2
O
3
has been recently unveiled as a promising wide bandgap quasi-
two-dimensional (2D) semiconducting material with interesting optical and
electrical properties [1-10]. Advancements in the bulk single crystal β-Ga
2
O
3
growth methods and β-Ga
2
O
3
-based field-effect transistor (FETs)

have
attracted much attention into this regard and ignited substantial scientific
interests [6, 7, 11]. β-Ga
2
O
3
is not strictly a 2D material. However, owing to
the large lattice constant of 12.33 Å along the [100] direction of monoclinic
β-Ga
2
O
3
crystal [7, 8], quasi-2D β-Ga
2
O
3
nanomembranes

can be
mechanically exfoliated from the cleavage planes of β-Ga
2
O
3
crystal, similar
to the exfoliation of 2D materials [7-15]. Previously, Hwang et al. reported
they got 20 nm β-Ga
2
O
3
nanomembranes after exfoliation [6]. Zhou et al.
reported that by reducing the thickness of β-Ga
2
O
3
nanomembranes from 100
16

nm to 50 nm, the threshold voltage would shift from negative values in
depletion mode to positive values in enhance mode, but the drain current
would also decrease a lot [9]. FETs based on exfoliated β-Ga
2
O
3
have been
reported to show large on-state current [7, 9, 16], high breakdown voltage [17,
18], excellent on/off ratios [7, 9, 17], and the potential of operating in radio
frequency regime [19]. Although β-Ga
2
O
3
exhibits the properties as a
competitive candidate for the next-generation electronics [6, 7, 17], FETs
made with β-Ga
2
O
3
usually possess significant Schottky barriers [6, 7, 20-22],
which result in high contact resistance and small on-state current, and thus
hamper the FET performance [23-26]. For semiconducting metal-oxides, the
metal contacts often react with metal-oxides at the interface to produce
complex compounds, making the problem even more complicated [27-34].
Unfortunately, these reactions have not been well documented for β-Ga
2
O
3
,
leaving a gap in understanding the interface reactions and corresponding
effects on the electronic properties of β-Ga
2
O
3
devices [21, 35]. Hence, to fully
develop the potential of β-Ga
2
O
3
, reliable Ohmic contacts and investigation
of reactions at metal/β-Ga
2
O
3
interface are needed.
To improve the metal/semiconductor interface contacts, several methods
have been demonstrated, such as gas adsorption, doping techniques, and
utilizing the interface reactions [4, 23-26, 36-41]. Even though metal/β-Ga
2
O
3
17

contacts have not been fully studied, significant research has been carried out
for some other metal-oxides, such as titanium oxide (TiO
2
) and indium oxide
(In
2
O
3
) [26-33, 40-44], which can lend us ideas about how to improve the
metal/β-Ga
2
O
3
contacts. As a well-known metal-oxide, TiO
2
can be readily
reduced by most electrode metals [27-34, 42, 43]. Furthermore, by comparing
the electronic performance of TiO
2
-based memristors with different contact
metals, it was shown that the oxygen vacancies created by reactions at the
metal/TiO
2
interfaces dominate other factors in determining the device
performance [27-34, 43]. Previously, we reported that In
2
O
3
nanowire FETs
exhibited a pronounced increase in conductance after baking in vacuum [44,
45]. In contrast, baking the FETs in ambient air led to a reduction in carrier
concentration and caused drastic suppression in device performance [44]. The
underlying mechanism is that baking In
2
O
3
nanowire FETs in vacuum led to
oxygen vacancy formation in In
2
O
3
, resulting in increased doping of and
higher conductance [44]. On the other hand, baking In
2
O
3
nanowire FETs in
air led to reduced oxygen vacancy concentration, lower n-type doping and
hence lower conductance [43, 44]. The above phenomena allow us to tune the
Schottky barriers formed between the metal contacts and the β-Ga
2
O
3
, which
would improve the efficiency of charge injection at the metal/β-Ga
2
O
3
interface and lower the contact resistance [27-34, 43-47].
18

To suppress the contact resistance, different dopants and contact
materials were investigated for n-type and p-type behavior β-Ga
2
O
3
FETs [7,
20, 48]. N-type device performance has been reported by using indium tin
oxide (ITO) or aluminum zinc oxide (AZO) to achieve low contact resistance
of ~ 0.5-0.6 Ω·mm [49, 50]. Meanwhile, argon plasma bombardment has been
applied to achieve a large on-state current density of ~ 0.65 mA/μm, but no
current saturation was observed [9, 16]. P-type behavior β-Ga
2
O
3
FETs have
also been demonstrated by applying the zinc (Zn) or magnesium (Mg)-doping
methods, but the hole mobility and on-state current of p-type β-Ga
2
O
3
FETs
still left much to be desired [20, 51]. Until now, the effects of interactions
between the contacts and β-Ga
2
O
3
still need a lot more investigation [15, 21-
35]. Therefore, to determine the value of the contact resistance and
systematically investigate the interface reaction, transfer line measurements
(TLM) and X-ray photoelectron spectroscopy (XPS) are needed [23, 27, 46].
In this study, we report that n-type β-Ga
2
O
3
FETs made with
titanium/gold (Ti/Au) contacts could achieve a large saturation drain current
density of ~ 3.1 mA/μm and low contact resistance of ~ 0.387 Ω·mm after
annealing in argon atmosphere. Here we demonstrate that such a simple yet
effective argon annealing process can significantly improve the performance
of β-Ga
2
O
3
FETs with Ti/Au contacts. After the annealing process, the β-
19

Ga
2
O
3
FETs with Ti/Au contacts can be switched from Schottky to Ohmic
behavior. More interestingly, annealing in argon atmosphere led to the
formation of oxygen vacancies at the Ti/β-Ga
2
O
3
interface [27, 30-34, 43, 52],
effectively reducing the Schottky barriers at the interface and providing
reliable Ohmic contacts. All the electronic measurements and XPS results
suggest that the annealing method is a reliable and generic approach to
improve the performance of β-Ga
2
O
3
FETs.

2.2 Gallium oxide field-effect transistor device
fabrication and characterization
The β-Ga
2
O
3
material was first exfoliated to membranes with thickness
in tens of nanometer scale by vacuum tape. After exfoliation, the
nanomembranes were dry transferred onto a silicon (Si) wafer with silicon
oxide (SiO
2
) thickness of 300 nm. An atomic force microscopy (AFM,
Dimensional, 3100 Digital Instruments, taping mode) image of a typical β-
Ga
2
O
3
nanomembrane is shown in Figure 2.1a, revealing that the thickness of
nanomembrane is ~ 80 nm. To further explore the structural properties of
quasi-layered β-Ga
2
O
3
, high-resolution transmission electron microscopy
(HRTEM) study was performed. The TEM image of a β-Ga
2
O
3

20

nanomembrane is shown on the left part of Figure 2.1b. It clearly
demonstrated the β-Ga
2
O
3
(002) atomic planes with a lattice spacing of ~ 0.28
nm, which is in good agreement with the literatures [11, 12]. The right part of
Figure 2.1b shows the pattern of fast Fourier transformation (FFT), revealing
the symmetry and parameters of lattice groups [11, 12]. Raman spectroscopy
(Renishaw Raman with a 532 nm excitation laser and a laser spot size of ~
1μm) was also performed to investigate the quality of exfoliated β-Ga
2
O
3
nanomembranes (Figure 2.2). The phonon positions of exfoliated
nanomembranes keep the same as the bulk crystal of β-Ga
2
O
3
, indicating the
structure of β-Ga
2
O
3
nanomembranes did not degrade during the mechanical
exfoliation [52]. The schematic diagram of a typical β-Ga
2
O
3
backgated FET
is depicted in Figure 2.1c. Heavily p-type doped Si was used as the common
back gate, and the gate dielectric was a thermally grown 300 nm thick SiO
2

layer. The β-Ga
2
O
3
material (Sn-doped) was bought from MTI Corporation.
The β-Ga
2
O
3
was first exfoliated to nanometer scale by using vacuum tape
(Kapton Tapes). After exfoliation, the sample was dry-transferred onto
Si/SiO
2
substrates with alignment markers (the thickness of SiO
2
is 300 nm),
followed by rinsing and residue-removing procedures. A bilayer Poly (methyl
methacrylate) (PMMA) consisting of PMMA A6 and PMMA A2 was then
spin-coated onto the Si/SiO
2
surface. After spin-coating, electron beam
21

lithography was conducted to pattern the source/drain electrodes, followed by
development, e-beam metal evaporation and lift-off processes. The Ti (10 nm)
/Au (150 nm) metal stacks were deposited as source and drain contacts at the
pressure of 10
-7
Torr using an e-beam evaporator [26], the bottom silicon
substrate served as a global back gate. The electronic properties were
measured using Agilent 4156B.
Figure 2.1d shows the optical image of a β-Ga
2
O
3
FET. Scanning
electron microscopy (SEM) was employed to visualize the channel area. The
inset SEM image in Figure 2.1d shows the channel area of the β-Ga
2
O
3
FET
with channel width of ~ 5.1 μm and channel length of ~ 550 nm. Next, we
studied the electronic performance of this β-Ga
2
O
3
FET before and after
annealing. A strong asymmetry Schottky behavior can be clearly observed
from the nonlinear behavior of I
ds
-V
ds
curve before annealing (black curve)
from Figure 2.1e. To perform the annealing treatment, we put the as-made β-
Ga
2
O
3
FET into a quartz tube which was flushed with pure argon to clear the
system from any residual gas species. After annealing at 300 ℃ for 180
minutes, an Ohmic contact behavior of this β-Ga
2
O
3
FET is observed from the
red curve in Figure 2.1e. Surprisingly, the enhanced on-state drain current is
3 orders of magnitude larger than before, suggesting that the contacts of β-
Ga
2
O
3
FET were significantly improved after annealing in argon atmosphere.
22

The improved device performance reveals the transition from Schottky to
Ohmic contacts at the Ti/β-Ga
2
O
3
interface, which also gave rise to current
saturation. Figure 2.1f shows the output characteristics of as-made β-Ga
2
O
3

FET. The saturation behavior of the drain current of this device was observed
when high drain voltage was applied. The device sustained to work with
increasing drain voltage, reaching a large drain current density of ~ 3.1
mA/μm at V
ds
=100 V. The Ohmic contacts after annealing enable the
outstanding performance of β-Ga
2
O
3
FET, which can survive at high voltages
and switch under large drain currents without avalanche or ionization
breakdown. Compared to other recent studies reporting the current density of
β-Ga
2
O
3
FETs [7, 9, 16], our result has the highest saturation drain current
density reported to date. Considering its quasi-2D properties, β-Ga
2
O
3
FETs
can be easily integrated in various platforms for power and radio frequency
electronics, which are desirable for microscale on-chip management
applications. Moreover, the device performance almost remained the same
even after it was exposed in ambient air at room temperature for two months,
which suggests the air stability of β-Ga
2
O
3
FETs (Figure 2.3).
23

Figure 2.1 Characteristics of β-Ga2O3 and typical β-Ga2O3 FETs. (a) AFM image and
height profile of the selected β-Ga2O3 flake. (b) TEM characterization of β-Ga2O3. (c)
Scheme of a typical β-Ga2O3 FET with Ti/Au contacts. (d) Optical image of a β-Ga2O3
FET with Ti/Au contacts. Inset is the SEM image of the channel area. (e) Ids-Vds curves of
β-Ga2O3 FET before (black curve) and after (red curve) annealing at 300 ℃ in argon. Inset
image is the enlarged Ids-Vds curve before annealing to illustrate the Schottky behavior. (f)
Ids-Vds family curves of a β-Ga2O3 FET under high Vds after annealing.

a
b
c
SiO
2
dielectric
β-Ga
2
O
3
Si backgate
Ti/Au contacts
d
-10 -5 0 5 10
-600
-300
0
300
600
Before Annealing
After Annealing
V
g
= 0 V

Drain Current Density (A/m)
Drain Voltage (V)
2 μ
-10 -5 0 5 10
-0.2
-0.1
0.0
0.1

0 20 40 60 80 100
0.0
1.0
2.0
3.0
V
g
= 20 V
V
g
= -20 V
V
g
= -60 V
V
g
= -100 V

Drain Current Density (mA/m)
Drain Voltage (V)
e f
μ
0 3 6
0
40
80
Height (nm)
Distance (m)
80 nm
β-Ga
2
O
3
Ti /Au contacts
2 μ
1 n
[001]
d(002)=0.28 nm
(002)
(1

1

0)
(1

1

1)
24

Figure 2.2. Raman spectra of β-Ga2O3

Figure 2.3 Ids-Vds family curves of β-Ga2O3 FET (a), and device after leaving in air for two
months (b).

25

2.3 Systematically studied the effectiveness of argon
annealing in improving the electronic
performance of β-Ga
2
O
3
FETs
We have systematically studied the effectiveness of argon annealing in
improving the electronic performance of β-Ga
2
O
3
FETs at elevated annealing
time. The typical transfer characteristics (I
ds
-V
g
curves) of a β-Ga
2
O
3
FET
before and after annealing at 10, 30, 60, 180 minutes are shown in Figure 2.4a.
It was discerned that annealing in argon at 300 ℃ for only 10 minutes can
increase the on-state current density a lot. Figure 2.4b shows the I
ds
-V
g
curves
of this β-Ga
2
O
3
FET in logarithmic scale. Noticeably, the on-state current
density continued to increase when the FET was sequentially annealed and
reached a maximum of ~ 620 μA/μm at V
ds
=12 V, which was almost 3000
times larger than the drain current density before annealing. In addition, we
can clearly observe that the device showed a low off-state current density of
~ 10
-7
μA/μm even after annealing in argon for 180 minutes, resulting in a
very high on/off ratio of ~ 6× 10
9
. Output characteristics of the electronic
properties of the β-Ga
2
O
3
FET are shown in Figure 2.4c and Figure 2.4d. The
I
ds
-V
ds
family curves of the β-Ga
2
O
3
FET before annealing are shown in Figure
2c with a nonlinear behavior at small V
ds
values. A strong Schottky barrier
26

with a rather small drain current density ~ 0.18 μA/μm at V
ds
=12 V can be
discerned for the device. After annealing in argon for 180 minutes, a more
Ohmic contact behavior of this device is observed from Figure 2.4d. The
substantially improved device performance after annealing shows that high
temperature treatment can transform the Schottky contacts between Ti and β-
Ga
2
O
3
into Ohmic contacts. It’s also worth to mention that the improvements
after annealing are highly reproducible. We have fabricated multiple β-Ga
2
O
3
FETs to test their performance and examine the generality of the annealing
method in improving the device performance. The statistics of the key figures
of merit are shown in Figure 2.4e and 2.4f. The significantly improvements
of FET performance after annealing were observed for all devices. All devices
showed substantial increases of the on-state current density at different
annealing time. In Figure 2.4e, we note that all the devices show a large on-
state current of several hundred microampere after annealing for 180 minutes.
Similar improvement of field effect mobility after annealing in argon was also
observed from Figure 2.4f. These effective mobility values were estimated
using the standard FET model:
𝛍 =
𝑳 𝑾 𝟏 𝑪 𝒐𝒙
𝑽 𝒅𝒔
𝒅 𝑰 𝒅𝒔
𝒅 𝑽 𝒈 ,
27

where L and W are the channel length and width of the FET, respectively. I
ds

is the drain current, V
ds
is the source-drain voltage, V
g
is the gate voltage, and
C
ox
is the gate capacitance per unit are.
Again, β-Ga
2
O
3
FET after annealing possess much higher mobility. The
maximum field-effect mobility reaches ~ 65 cm
2
/V· s after annealing the
device in argon for 180 minutes. Such large on-state current, high on/off ratio
and electron mobility offer the promise of β-Ga
2
O
3
for the power electronic
devices and radio frequency applications. Furthermore, we notice that
annealing the devices at higher temperature (500 ℃) or annealing at 300 ℃
for longer time (240 min) would deteriorate the device performance. The SEM
images in Figure 2.5 show that the Ti/Au contacts had a continuous and
smooth morphology before annealing, and the Ti/Au contacts remained good
after optimum annealing at 300 ℃ in argon for 180 min, but the contacts
degraded significantly with the further increase in temperature and time,
which led to poor electronic performance.
28

Figure 2.4 Electronic properties of β-Ga2O3 FETs with different annealing time. (a)
Typical Ids-Vg curves of a β-Ga2O3 FET with different annealing time. (b) Ids-Vg curves of
a β-Ga2O3 FET shown in logarithmic scale. Ids-Vds family curves of a β-Ga2O3 FET at
different Vg before annealing (c) and after annealing (d). Changes of on-state current (e)
and corresponding changes of field effect mobility (f) of β-Ga2O3 FETs versus different
annealing time for seven devices

0 4 8 12
0.00
0.05
0.10
0.15
0.20
V
g
= 40 V
V
g
= 10 V
V
g
= -20 V
V
g
= -50 V
V
g
= -80 V

Drain Current Density (A/m)
Drain Voltage (V)
0 4 8 12
0
200
400
600 V
g
= 40 V
V
g
= 10 V
V
g
= -20 V
V
g
= -50 V
V
g
= -80 V

Drain Current Density (A/m)
Drain Voltage (V)
a
c
b
d
-100 -50 0 50
0
200
400
600
V
ds
= 12 V
180 min
60 min
30 min
10 min
0 min

Drain Current Density (A/m)
Gate Voltage (V)
-100 -50 0 50
10
-7
10
-5
10
-3
10
-1
10
1
10
3
180 min
60 min
30 min
10 min
0 min

Drain Current Density (A/m)
Gate Voltage (V)
V
ds
= 12 V
on / off ~6× 10
9
e f
0 30 60 90 120 150 180
10
-2
10
-1
10
0
10
1
10
2
10
3
Annealing Time (minute)
Annealing Time (mintue)
Annealing Time (mintue) Annealing Time (mintue)
Device 1
Device 2
Device 3
Device 4
Device 5
Device 6

On Current Density (A/m)
V
ds
=12 V
V
g
= 40 V
0 30 60 90 120 150 180
0
20
40
60
V
ds
=12 V
V
g
= 40 V
Device 1
Device 2
Device 3
Device 4
Device 5
Device 6

Mobility(cm
2
/V  s)
Annealing Time (minute)
29

Figure 2.5 Ti/Au contacts with different annealing temperature and annealing time. (a)
Optical images of Ti/Au pads before annealing and degraded Ti/Au pads after annealing.
The SEM images of Ti/Au contacts before annealing (b), after annealing at 300 ℃ in argon
for 180 min (c), after annealing at 500 ℃ in argon for 180 min (d), and after annealing at
300 ℃ in argon for 240 min (e). (f) Ids-Vg curves of three β-Ga2O3 FETs with different
annealing temperature and annealing time.

Ti/Au contacts
Before Annealing Degraded contacts
Ti/Au contacts
Ti/Au
contacts
Ti/Au
contacts
1 μ
1 μ 1 μ
1 μ
Ti/Au
contacts
-100 -50 0
10
-7
10
-5
10
-3
10
-1
10
1
10
3
V
ds
= 12 V

Drain Current Density (A/m)
Gate Voltage (V)
Annealing at 300
o
C for 180 min
Annealing at 300
o
C for 240 min
Annealing at 500
o
C for 180 min
e
c
a
b
d
f
30

Figure 2.6 Contact resistance of β-Ga2O3 FETs. (a) Optical image of a β-Ga2O3 FET with
TLM structure. (b) Total resistance versus channel length at different annealing time of a
β-Ga2O3 FET. (c) Contact resistance versus annealing time of a β-Ga2O3 FET. (d) Contact
resistance (logarithmic scale) versus annealing time of three as-made β-Ga2O3 FETs with
TLM structures.

To further study the contacts at the Ti/β-Ga
2
O
3
interface, we conducted
the TLM measurement with different channel lengths. Figure 2.6a is the
optical image of a typical β-Ga
2
O
3
FET with TLM structure. Channel lengths
of this TLM structure are measured to be ~ 0.74, ~ 1.04 and ~ 2.74 μm. The
fitted total resistance versus channel length curves of an as-made β-Ga
2
O
3
FET measured under different annealing time are shown in Figure 2.6b. Based
c d
b
a
0 1 2 3
0
30
60
4000
6000
0 min
10 min
30 min
60 min
180 min

Resistance (  mm)
Channel Length (m)
V
g
= 40 V
0 30 60 90 120 150 180
10
-1
10
0
10
1
10
2
10
3
10
4
Device 1
Device 2
Device 3

Contact Resistance (  mm)
Annealing Time (minute)
V
g
= 40 V
0 30 60 90 120 150 180
0
5
10
430
0.387   mm

Contact Resistance (  mm)
Annealing Time (minute)
V
g
= 40 V
μ
β-Ga
2
O
3
Ti/Au contacts
31

on the equation R
total
=2× R
contact
+R
channel
, the total resistance would be twice as
much as the contact resistance when the channel length approaches zero. Thus,
R
contact
can be determined by calculating one half of the y-intercept of the fitted
total resistance curves. The contact resistance values of this β-Ga
2
O
3
FET at
different annealing time are extracted and summarized in Figure 2.6c. With
the increase of annealing time to 180 minutes, the contact resistance
normalized by the channel width decreases from ~ 430 to ~ 0.387 Ω·mm. The
changes in the contact resistance reveal that the annealing in argon can
effectively improve the device contact and bring the R
contact
down to an
extremely small level. Figure 2.6d summarizes the improvement of contact
resistance with different annealing time of three β-Ga
2
O
3
FETs. Again, we can
observe that the argon-annealed β-Ga
2
O
3
FETs possess contact resistance
much lower than those of the as-made β-Ga
2
O
3
FETs. The results suggest the
generality of the argon annealing method to improve the device performance,
facilitate the electron injection from the metal contacts to the conduction band
of β-Ga
2
O
3
, resulting in the low-resistance state at interfaces. Such low
contact resistance is very essential in power electronics. Our results shed light
on the importance of annealing, which can be combined with other contact
engineering techniques to achieve scaled fabrication of high-performance β-
Ga
2
O
3
FETs.
32

Figure 2.7 XPS spectra of pure β-Ga2O3, β-Ga2O3 after annealing in argon at 300 ℃ for
180 minutes, and Ti-coated (1 nm) β-Ga2O3 after annealing in argon at 300 ℃ for 180
minutes.

It is important to understand what happened during annealing and the
mechanism behind the improvement of device performance. Therefore, X-ray
spectroscopy (XPS) was employed to study the science involved during the
annealing process (Figure 2.7 in ESM). Detailed XPS spectra of pure β-Ga
2
O
3
and its control samples are shown in Figure 2.8. In Figure 2.8a, XPS
measurements of pure β-Ga
2
O
3
shows a major peak at the binding energy of ~
33

1118.0 eV corresponds to Ga 2p
3/2
from β-Ga
2
O
3
[47, 53, 55]. The β-Ga
2
O
3
3d
signal is composed of two separate peaks: one at the binding energy of ~ 20.95
eV corresponds to Ga 3d
3/2
from β-Ga
2
O
3
, and the other one at ~ 20.5 eV
corresponds to Ga 3d
5/2
from β-Ga
2
O
3
[47, 53-55].

Interestingly, after
annealing pure β-Ga
2
O
3
in argon at 300 ℃ for 180 minutes, new peaks with
small intensities are observed from XPS measurements. In Figure 2.8b, the
new peaks are at lower binding energy compared with the nearby major peaks
from pure β-Ga
2
O
3
. Considering the fact that the pure gallium (Ga) materials
have a Ga 2p
3/2
peak at 1116.5 eV and a Ga 3d
5/2
peak at 18.7 eV [55], these
new peaks presented in Figure 2.8b are in the intermediate area between peaks
of β-Ga
2
O
3
and gallium element, correlating to gallium with small oxidation
numbers and indicating the existence of oxygen vacancies in β-Ga
2
O
3
after
argon annealing [47, 54]. It is hard to generate enough oxygen vacancies in β-
Ga
2
O
3
to be detected by XPS under ambient atmosphere [46]. While in the
oxygen-poor atmosphere such as argon, it is much easier to form the oxygen
vacancies because of the negative formation energy [46]. The pressure from
argon atmosphere leads the diffusion of oxygen atoms from normal lattice
sites to the gaseous state, generating oxygen vacancies at the surface of β-
Ga
2
O
3
[44, 47]. These oxygen vacancies would change the chemical states of
surface β-Ga
2
O
3
and create new peaks in XPS. In Figure 2.8c, similar new
34

peaks are also observed for the

Ti-coated (1 nm) β-Ga
2
O
3
which was annealed
in argon at 300 ℃ for 180 minutes. Surprisingly, the XPS data of the annealed
sample exhibited two shoulders near the major peaks from pure β-Ga
2
O
3
.
Figure 2.8c shows one shoulder around the binding energy of ~ 1116.6 eV,
and the other one ~ 19.3 eV. These two shoulders can also be explained by
the presence of oxygen vacancies at the Ti/β-Ga
2
O
3
interface. There are many
Magné li phases with the formula Ti
n
O
2n-1
in the Ti-O system, where Magné li
phases from Ti
2
O
3
up to Ti
20
O
39
have already been discovered [42]. Figure
2.8d lists the free energy of Ti
n
O
2n-1
and β-Ga
2
O
3
formation [12, 42]. And
Figure 2.9 confirmed the existence of Magnéli phase at the Ti/β-Ga
2
O
3
interface after annealing.

Since Magné li phases of Ti-O system usually have
much lower formation energy than β-Ga
2
O
3
, the Ti metal contacts can easily
reduce Ga ions in β-Ga
2
O
3
at high temperatures [26-35, 42-48]. Therefore, the
two shoulders observed in Figure 2.8c can be assigned to the high
concentration of oxygen vacancies at the Ti/β-Ga
2
O
3
interface [47, 53-55]. At
high temperature in argon atmosphere, Ti layer served as a chemically reactive
contact to reduce the β-Ga
2
O
3
and generated significant amount of oxygen
vacancies locally distributed at the interface [29, 33, 46-50]. These oxygen
vacancies can act as effective electron donors in the n-type

semiconductors.
Thus, the heavily oxygen-vacancy “self-doped” surface will narrow the
35

depletion layer, resulting in Ohmic contacts behavior and the low contact
resistance of β-Ga
2
O
3
FETs.

Figure 2.8 XPS study of β-Ga2O3. Normalized Ga 2p3/2 XPS spectra and Ga 3d XPS spectra
of pure β-Ga2O3 (a), β-Ga2O3 after annealing in argon at 300 ℃ for 180 minutes (b), and
Ti-coated (1 nm) β-Ga2O3 after annealing in argon at 300 ℃ for 180 minutes (c). Black
-4000
-3500
-3000
-2500
-2000
-1500
-1000
-500
.........
Ti
4
O
7
Ti
3
O
5
Ti
2
O
3
Ga
2
O
3
Standard enthalpy of formation (KJ· mol
-1
)
TiO
2
Ti
n
O
2n-1
Mangeli Phase
a b
c d
1120 1118 1116 22 20 18
Ga 3d
5/2
for Ga Ga 2p
3/2
for Ga
Ga 3d
5/2
for Ga
2
O
3
Ga 3d
3/2
for Ga
2
O
3
Ga 2p
3/2
for Ga
2
O
3
Binding Energy (eV)

Normalized Intensity
1120 1118 1116 22 20 18
Ga 3d
5/2
for Ga Ga 2p
3/2
for Ga
Ga 3d
5/2
for Ga
2
O
3
Ga 2p
3/2
for Ga
2
O
3
Ga 3d
3/2
for Ga
2
O
3
Binding Energy (eV)

Normalized Intensity
Ga signals for Ga with
small oxidation number
1120 1118 1116 22 20 18
Ga 3d
5/2
for Ga Ga 2p
3/2
for Ga
Binding Energy (eV)

Normalized Intensity
Ga 2p
3/2
for Ga
2
O
3 Ga 3d
5/2
for Ga
2
O
3
Ga 3d
3/2
for Ga
2
O
3
Ga signals for Ga with
small oxidation number
Ti contact
β-phase Ga
2
O
3
Oxygen vacancy
e
36

dots are experimental data, red curves are simulated fitting curves. (d) Scheme of free
energy of different metal oxides. (e) Schematic diagram of proposed model for oxygen
vacancies at the Ti/β-Ga2O3 interface.

Figure 2.9 Ti 2p core-level XPS spectra for the Ti-coated (1 nm) β-Ga2O3 after annealing
in argon at 300 ℃ for 180 minutes. Black curves are experimental data; red curves are
simulated fitting curves. We can only observe three peaks: one at the binding energy of ~
458.9 eV corresponds to Ti
4+
, one at the binding energy of ~ 457.5 eV corresponds to Ti
3+
,
and one at the binding energy of ~ 456 eV corresponds to Ti
2+
. These peaks confirmed the
existence of Magneli Phases in Ti-O system.

2.4 Summary
In summary, we have fabricated β-Ga
2
O
3
FETs with Ti/Au contacts and
demonstrated that the argon annealing method would significantly improve
37

the device performance. With different annealing time, the devices were
switched from a rectifying behavior to an Ohmic behavior. After annealing in
argon at 300 ℃ for 180 minutes, the β-Ga
2
O
3
FETs showed a large saturation
drain current density of ~ 3.1 mA/μm and a high on/off ratio of ~ 6× 10
9
. We
further showed that the FET contact resistances were reduced to ~ 0.387
Ω·mm after argon annealing, which can be explained by the generation of
oxygen vacancies at the Ti/β-Ga
2
O
3
interface. XPS study showed that Ti can
reduce β-Ga
2
O
3
to create a large amount of oxygen vacancies, and in turn
resulted in good device contacts. Our results shown in this study illuminated
the important role of metal/β-Ga
2
O
3
interface in boosting the performance of
β-Ga
2
O
3
FETs and can further benefit devices made with other metal-oxides
and quasi-2D materials

2.5 References

[1] Novoselov, K. S.; Geim, A. K.; Morozov, S. V .; Jiang, D.; Zhang, Y .;
Dubonos, S. V .; Grigorieva, I. V .; Firsov, A. A., Electric Field Effect in
Atomically Thin Carbon Films. Science 2004, 306 (5696), 666.
[2] Wang, Q. H.; Kalantar-Zadeh, K.; Kis, A.; Coleman, J. N.; Strano, M. S.,
Electronics and optoelectronics of two-dimensional transition metal
38

dichalcogenides. Nat. Nanotechnol. 2012, 7, 699.
[3] Carey, B. J.; Ou, J. Z.; Clark, R. M.; Berean, K. J.; Zavabeti, A.; Chesman,
A. S. R.; Russo, S. P.; Lau, D. W. M.; Xu, Z.-Q.; Bao, Q. et al. Wafer-scale
two-dimensional semiconductors from printed oxide skin of liquid metals.
Nat. Commun. 2017, 8, 14482.
[4] Liu, B.; Abbas, A.; Zhou, C., Two‐Dimensional Semiconductors: From
Materials Preparation to Electronic Applications. Adv. Electron. Mater.
2017, 3 (7), 1700045.
[5] Kalantar-zadeh, K.; Ou, J. Z.; Daeneke, T.; Mitchell, A.; Sasaki, T.; Fuhrer,
M. S., Two dimensional and layered transition metal oxides. Appl. Mater.
Today 2016, 5, 73-89.
[6] Hwang, W. S.; Verma, A.; Peelaers, H.; Protasenko, V .; Rouvimov, S.;
Xing, H.; Seabaugh, A.; Haensch, W.; de Walle, C. V .; Galazka, Z.;
Albrecht, M.; Fornari, R.; Jena, D., High-voltage field effect transistors
with wide-bandgap β-Ga2O3 nanomembranes. Appl. Phys. Lett. 2014,
104 (20), 203111.
[7] Higashiwaki, M.; Sasaki, K.; Kamimura, T.; Hoi Wong, M.;
Krishnamurthy, D.; Kuramata, A.; Masui, T.; Yamakoshi, S., Depletion-
mode Ga
2
O
3
metal-oxide-semiconductor field-effect transistors on β-
39

Ga
2
O
3
(010) substrates and temperature dependence of their device
characteristics. Appl. Phys. Lett. 2013, 103 (12), 123511.
[8] Irmscher, K.; Galazka, Z.; Pietsch, M.; Uecker, R.; Fornari, R., Electrical
properties of β-Ga
2
O
3
single crystals grown by the Czochralski method. J.
Appl. Phys. 2011, 110 (6), 063720.
[9] Zhou, H.; Si, M.; Alghamdi, S.; Qiu, G.; Yang, L.; Ye, P. D., High-
Performance Depletion/Enhancement-ode β-Ga
2
O
3
on Insulator (GOOI)
Field-Effect Transistors with Record Drain Currents of 600/450 mA/mm.
IEEE Electron Device Lett. 2017, 38 (1), 103-106.
[10] Wu, F.; Chen, L.; Zhang, A.; Hong, Y .-L.; Shih, N.-Y .; Cho, S.-Y .;
Drake, G. A.; Fleetham, T.; Cong, S.; Cao, X. et al. High-Performance
Sub-Micrometer Channel WSe
2
Field-Effect Transistors Prepared Using a
Flood–Dike Printing Method. ACS Nano 2017, 11 (12), 12536-12546.
[11] Galazka, Z.; Uecker, R.; Irmscher, K.; Albrecht, M.; Klimm, D.; Pietsch,
M.; Brützam, M.; Bertram, R.; Ganschow, S.; Fornari, R., Czochralski
growth and characterization of β‐Ga
2
O
3
single crystals. Cryst. Res.
Technol. 2010, 45 (12), 1229-1236.
[12] Ahman, J.; Svensson, G.; Albertsson, J., A Reinvestigation of β-
Gallium Oxide. Acta Cryst. C 1996, 52 (6), 1336-1338.
40

[13] Lovejoy, T. C.; Yitamben, E. N.; Shamir, N.; Morales, J.; Villora, E. G.;
Shimamura, K.; Zheng, S.; Ohuchi, F. S.; Olmstead, M. A., Surface
morphology and electronic structure of bulk single crystal β-Ga
2
O
3
(100).
Appl. Phys. Lett. 2009, 94 (8), 081906.
[14] Ma, N.; Tanen, N.; Verma, A.; Guo, Z.; Luo, T.; Xing, H.; Jena, D.,
Intrinsic electron mobility limits in β-Ga
2
O
3
. Appl. Phys. Lett. 2016, 109
(21), 212101.
[15] Ramana, C. V .; Rubio, E. J.; Barraza, C. D.; Miranda Gallardo, A.;
McPeak, S.; Kotru, S.; Grant, J. T., Chemical bonding, optical constants,
and electrical resistivity of sputter-deposited gallium oxide thin films. J.
Appl. Phys. 2014, 115 (4), 043508.
[16] Zhou, H.; Maize, K.; Noh, J.; Shakouri, A.; Ye, P. D., Thermodynamic
Studies of β-Ga
2
O
3
Nanomembrane Field-Effect Transistors on a Sapphire
Substrate. ACS Omega 2017, 2 (11), 7723-7729.
[17] Green, A. J.; Chabak, K. D.; Heller, E. R.; Fitch, R. C.; Baldini, M.;
Fiedler, A.; Irmscher, K.; Wagner, G.; Galazka, Z.; Tetlak, S. E. et al. 3.8-
MV/cm Breakdown Strength of MOVPE-Grown Sn-Doped β-Ga
2
O
3

MOSFETs. IEEE Electron Device Lett. 2016, 37 (7), 902-905.
[18] Wong, M. H.; Sasaki, K.; Kuramata, A.; Yamakoshi, S.; Higashiwaki,
41

M., Field-Plated Ga
2
O
3
MOSFETs With a Breakdown V oltage of Over 750
V . IEEE Electron Device Lett. 2016, 37 (2), 212-215.
[19] Green, A. J.; Chabak, K. D.; Baldini, M.; Moser, N.; Gilbert, R.; Fitch,
R. C.; Wagner, G.; Galazka, Z.; Mccandless, J.; Crespo, A.; Leedy, K.;
Jessen, G. H., β- Ga
2
O
3
MOSFETs for Radio Frequency Operation. IEEE
Electron Device Lett. 2017, 38 (6), 790-793.
[20] Chang, P.-C.; Fan, Z.; Tseng, W.-Y .; Rajagopal, A.; Lu, J. G., β-Ga
2
O
3

nanowires: Synthesis, characterization, and p-channel field-effect
transistor. Appl. Phys. Lett. 2005, 87 (22), 222102.
[21] Farzana, E.; Zhang, Z.; Paul, P. K.; Arehart, A. R.; Ringel, S. A.,
Influence of metal choice on (010) β-Ga
2
O
3
Schottky diode properties.
Appl. Phys. Lett. 2017, 110 (20), 202102.
[22] Yao, Y .; Davis, R. F.; Porter, L. M., Investigation of Different Metals as
Ohmic Contacts to β-Ga
2
O
3
: Comparison and Analysis of Electrical
Behavior, Morphology, and Other Physical Properties. J. Electron. Mater.
2017, 46 (4), 2053-2060.
[23] Ma, Y .; Shen, C.; Zhang, A.; Chen, L.; Liu, Y .; Chen, J.; Liu, Q.; Li, Z.;
Amer, M. R.; Nilges, T.; Abbas, A. N.; Zhou, C., Black Phosphorus Field-
Effect Transistors with Work Function Tunable Contacts. ACS Nano 2017,
42

11 (7), 7126-7133.
[24] Mann, D.; Javey, A.; Kong, J.; Wang, Q.; Dai, H., Ballistic Transport in
Metallic Nanotubes with Reliable Pd Ohmic Contacts. Nano Lett. 2003, 3
(11), 1541-1544.
[25] Ma, Y .; Liu, B.; Zhang, A.; Chen, L.; Fathi, M.; Shen, C.; Abbas, A. N.;
Ge, M.; Mecklenburg, M.; Zhou, C., Reversible Semiconducting-to-
Metallic Phase Transition in Chemical Vapor Deposition Grown
Monolayer WSe
2
and Applications for Devices. ACS Nano 2015, 9 (7),
7383-7391.
[26] English, C. D.; Shine, G.; Dorgan, V . E.; Saraswat, K. C.; Pop, E.,
Improved Contacts to MoS
2
Transistors by Ultra-High Vacuum Metal
Deposition. Nano Lett. 2016, 16 (6), 3824-3830.
[27] Brillson, L. J., Chemical reaction and charge redistribution at metal–
semiconductor interfaces. J. Vac. Sci. Technol. 1978, 15 (4), 1378-1383.
[28] Guo, Y .; Zhou, J.; Liu, Y .; Zhou, X.; Yao, F.; Tan, C.; Wu, J.; Lin, L.;
Liu, K.; Liu, Z.; Peng, H., Chemical Intercalation of Topological Insulator
Grid Nanostructures for High‐Performance Transparent Electrodes. Adv.
Mater. 2017, 29 (44), 1703424.
[29] Mosbacker, H. L.; Strzhemechny, Y . M.; White, B. D.; Smith, P. E.;
43

Look, D. C.; Reynolds, D. C.; Litton, C. W.; Brillson, L. J., Role of near-
surface states in ohmic-Schottky conversion of Au contacts to ZnO. Appl.
Phys. Lett. 2005, 87 (1), 012102.
[30] Sawa, A., Resistive switching in transition metal oxides. Mater. Today
2008, 11 (6), 28-36.
[31] Fortunato, E.; Barquinha, P.; Martins, R., Oxide Semiconductor Thin‐
Film Transistors: A Review of Recent Advances. Adv. Mater. 2012, 24 (22),
2945-2986.
[32] Waser, R.; Dittmann, R.; Staikov, G.; Szot, K., Redox‐Based Resistive
Switching Memories – Nanoionic Mechanisms, Prospects, and
Challenges. Adv. Mater. 2009, 21 (25‐26), 2632-2663.
[33] Yang, J. J.; Strachan John, P.; Xia, Q.; Ohlberg Douglas, A. A.; Kuekes
Philip, J.; Kelley Ronald, D.; Stickle William, F.; Stewart Duncan, R.;
Medeiros‐Ribeiro, G.; Williams, R. S., Diffusion of Adhesion Layer
Metals Controls Nanoscale Memristive Switching. Adv. Mater. 2010, 22
(36), 4034-4038.
[34] Yang, J. J.; Pickett, M. D.; Li, X.; Ohlberg, D. A. A.; Stewart, D. R.;
Williams, R. S., Memristive switching mechanism for metal/oxide/metal
nanodevices. Nat. Nanotechnol. 2008, 3, 429.
44

[35] Gao, X.; Xia, Y .; Ji, J.; Xu, H.; Su, Y .; Li, H.; Yang, C.; Guo, H.; Yin,
J.; Liu, Z., Effect of top electrode materials on bipolar resistive switching
behavior of gallium oxide films. Appl. Phys. Lett. 2010, 97 (19), 193501.
[36] Javey, A.; Guo, J.; Wang, Q.; Lundstrom, M.; Dai, H., Ballistic carbon
nanotube field-effect transistors. Nature 2003, 424, 654.
[37] Liu, B.; Ma, Y .; Zhang, A.; Chen, L.; Abbas, A. N.; Liu, Y .; Shen, C.;
Wan, H.; Zhou, C., High-Performance WSe2 Field-Effect Transistors via
Controlled Formation of In-Plane Heterojunctions. ACS Nano 2016, 10
(5), 5153-5160.
[38] Mak, K. F.; He, K.; Lee, C.; Lee, G. H.; Hone, J.; Heinz, T. F.; Shan, J.,
Tightly bound trions in monolayer MoS
2
. Nat. Mater. 2012, 12, 207.
[39] Wood, J. D.; Wells, S. A.; Jariwala, D.; Chen, K.-S.; Cho, E.; Sangwan,
V . K.; Liu, X.; Lauhon, L. J.; Marks, T. J.; Hersam, M. C., Effective
Passivation of Exfoliated Black Phosphorus Transistors against Ambient
Degradation. Nano Lett. 2014, 14 (12), 6964-6970.
[40] Desai, S. B.; Seol, G.; Kang, J. S.; Fang, H.; Battaglia, C.; Kapadia, R.;
Ager, J. W.; Guo, J.; Javey, A., Strain-Induced Indirect to Direct Bandgap
Transition in Multilayer WSe
2
. Nano Lett. 2014, 14 (8), 4592-4597.
[41] Liu, Y .; Guo, J.; Zhu, E.; Liao, L.; Lee, S.-J.; Ding, M.; Shakir, I.;
45

Gambin, V .; Huang, Y .; Duan, X., Approaching the Schottky–Mott limit
in van der Waals metal–semiconductor junctions. Nature 2 0 1 8 , 557 (7707),
696-700.
[42] Waldner, P.; Eriksson, G., Thermodynamic modelling of the system
titanium-oxygen. Calphad 1999, 23 (2), 189-218.
[43] Yang, J. J.; Strachan, J. P.; Miao, F.; Zhang, M.-X.; Pickett, M. D.; Yi,
W.; Ohlberg, D. A. A.; Medeiros-Ribeiro, G.; Williams, R. S., Metal/TiO
2

interfaces for memristive switches. Appl. Phys. A 2011, 102 (4), 785-789.
[44] Lei, B.; Li, C.; Zhang, D.; Tang, T.; Zhou, C., Tuning electronic
properties of In2O3 nanowires by doping control. Appl. Phys. A 2004, 79
(3), 439-442.
[45] Liu, Q.; Liu, Y .; Wu, F.; Cao, X.; Li, Z.; Alharbi, M.; Abbas, A. N.;
Amer, M. R.; Zhou, C., Highly Sensitive and Wearable In
2
O
3
Nanoribbon
Transistor Biosensors with Integrated On-Chip Gate for Glucose
Monitoring in Body Fluids. ACS Nano 2018, 12 (2), 1170-1178.
[46] Hollinger, G.; Skheyta-Kabbani, R.; Gendry, M., Oxides on GaAs and
InAs surfaces: An x-ray-photoelectron-spectroscopy study of reference
compounds and thin oxide layers. Phys. Rev. B 1994, 49 (16), 11159-
11167.
46

[47] Dong, L.; Jia, R.; Xin, B.; Peng, B.; Zhang, Y ., Effects of oxygen
vacancies on the structural and optical properties of β-Ga
2
O
3
. Sci. Rep.
2017, 7, 40160.
[48] Varley, J. B.; Weber, J. R.; Janotti, A.; Van de Walle, C. G., Oxygen
vacancies and donor impurities in β-Ga
2
O
3
. Appl. Phys. Lett. 2010, 97 (14),
142106.
[49] Carey, P. H.; Yang, J.; Ren, F.; Hays, D. C.; Pearton, S. J.; Jang, S.;
Kuramata, A.; Kravchenko, I. I., Ohmic contacts on n-type β-Ga
2
O
3
using
AZO/Ti/Au. AIP Adv. 2017, 7 (9), 095313.
[50] Carey, P. H.; Yang, J.; Ren, F.; Hays, D. C.; Pearton, S. J.; Kuramata,
A.; Kravchenko, I. I., Improvement of Ohmic contacts on Ga
2
O
3
through
use of ITO-interlayers. J. Vac. Sci. Technol. B 2017, 35 (6), 061201.
[51] Qian, Y . P.; Guo, D. Y .; Chu, X. L.; Shi, H. Z.; Zhu, W. K.; Wang, K.;
Huang, X. K.; Wang, H. et al. Mg-doped p-type β-Ga
2
O
3
thin film for
solar-blind ultraviolet photodetector. Mater. Lett. 2017, 209, 558-561.
[52] Strachan John, P.; Pickett Matthew, D.; Yang, J. J.; Aloni, S.; David
Kilcoyne, A. L.; Medeiros‐Ribeiro, G.; Stanley Williams, R., Direct
Identification of the Conducting Channels in a Functioning Memristive
Device. Adv. Mater. 2010, 22 (32), 3573-3577.
47

[53] Zhao, Y .; Frost Ray, L., Raman spectroscopy and characterisation of α‐
gallium oxyhydroxide and β‐gallium oxide nanorods. J. Raman Spectrosc.
2008, 39 (10), 1494-1501.
[54] Bourque, J. L.; Biesinger, M. C.; Baines, K. M., Chemical state
determination of molecular gallium compounds using XPS. Dalton Trans.
2016, 45 (18), 7678-7696.
[55] NIST X-ray Photoelectron Spectroscopy Database, NIST Standard
Reference Database Number 20, National Institute of Standards and
Technology, Gaithersburg MD, 20899 (2000)

48

Chapter 3 Air-stable n-type transistors
based on assembled aligned carbon
nanotube arrays and their application
in complementary metal-oxide-
semiconductor electronics
3.1 Introduction
In recent years, intense research and studies have been conducted on the
new generation of materials for energy-efficient and high-speed nano-
electronics after silicon transistors approach their physical and theoretical
limits [1–4]. Carbon nanotube (CNT), due to its one-dimensional nature and
excellent electronic properties, is a promising channel material that can enable
high-performance and multi-functional (such as flexible and wearable)
electronics [5–17]. Significant accomplishments have been made on both
performance and scaling for carbon nanotube field effect transistors (FETs)
[9–11, 18–22]. However, most of the previous work studied either individual
or carbon nanotube random networks. Recently, aligned nanotubes based on
assembling pre-separated semiconducting carbon nanotubes have emerged
and served as an important platform for advanced CNT electronics [23–25].
49

In 2016, Brady et al. reported the floating evaporative self-assembly (FESA)
technique to assemble high-semiconducting-purity nanotubes into aligned
high-density arrays on SiO
2
or quartz substrates [23]. Aligned nanotube arrays
enjoy the benefits of not only the high aligned nanotube density, but also fewer
defects in the transistor channel area. CNT array transistors can take
advantage of simultaneous contributions of the carrying currents passed by
multiple-nanotube channels to achieve high current densities [23, 24]. Quasi-
ballistic p-type behavior FESA carbon nanotube array transistors with a
saturated on-state current density exceeding silicon have been achieved when
compared at the same gate oxide thickness and off-state current density [24].
Besides, p-type behavior carbon nanotube array transistors were investigated
for radio frequency electronics and Rutherglen et al. observed excellent radio
frequency performance, which operated at over 100 gigahertz frequencies and
is close to that of GaAs technology [26–29].

Furthermore, Liu et al.
demonstrated high-density nanotube arrays (> 100 CNTs per micrometer) for
high-performance electronics [25].
As for now, the work on such high-density, aligned carbon nanotube
arrays have barely reported n-type behavior transistors. However, obtaining
both p-type and n-type transistors is of great importance in realizing the
promising future of complementary nano-electronic circuits with low steady-
50

state power dissipation and short stage delay [15, 22]. Even though significant
research work was devoted to producing n-type behavior transistors based on
individual nanotubes or nanotube networks, many scientific challenges still
need to be solved for reliable industrial-scale applications of aligned nanotube
arrays23, 30–34, 35]. Semiconducting CNTs are usually p-type materials in
atmosphere due to the adsorption of oxygen [36]. Many groups including our
own have researched techniques to convert CNT to n-type material, such as
chemical doping [33, 37], using metal contacts with small work functions (Gd,
Sc, or Y) [23, 33–34],

and passivating CNT with atomic layer deposition
(ALD) [22, 30–32]. However, whether these techniques can be extended to
assembled aligned CNT array devices is not clear so far. Importantly, the
carrier transport mechanisms can be different between long-channel
transistors (channel length >> nanotube length) and short-channel transistors
(channel length << nanotube length). The effects of metal contacts and ALD
passivation on assembled aligned CNT arrays need to be systematically
studied to provide a reliable and robust n-type conversion method. Long-term
stability and standard fabrication compatibility are needed for future
semiconductor electronics. It is thus necessary to investigate and develop an
n-type conversion method for aligned CNT array FETs with short channel
lengths, and to study ways to mitigate any adverse effects.
51

Here, by taking advantage of the FESA platform [24, 25], we have
successfully achieved aligned, high-density, and high-semiconducting-ratio
carbon nanotubes and fabricated FESA CNT FETs of various channel lengths
from 100 nm to 10 μm. The mechanism of carrier type has been systematically
studied to shed light on the importance of factors affecting the n-type
conversion process. We have found that for FESA CNT FETs (with ALD
passivation) with channel length ~ 100 nm, the transistor polarity is dominated
by the work function of source/drain metal contacts. We attribute this to the
direct contact carrier transport in short-channel transistors. We have then
combined the use of Ti contacts and ALD passivation to obtain air-stable n-
type behavior CNT FETs with high on-state current density ~ 130 μA/μm,
while maintaining large on/off ratio of > 10
6
. The use of Ti contacts is
particularly interesting as Ti is a metal with moderate work function (as
compared to low-work function metals such as Gd, Sc), and while Ti contacts
are widely used for III-V semiconductors, it has not been thoroughly studied
as nanotube contacts. A complementary metal-oxide-semiconductor (CMOS)
inverter is also demonstrated using the as-made p-type and n-type behavior
CNT FETs. Our experimental demonstration and understanding of the
transport behavior in aligned CNT array transistors can help the development
of future high-performance aligned CNT array transistor-based electronics.
52

3.2 Preparation process of aligned carbon arrays
Semiconducting carbon nanotube inks isolated using polyfluorene
derivative polymer wrapper poly[(9,9-dioctylfluorenyl-2,7-diyl)-alt-co-(6,6’-
{2,2’-bipyridine})] (PFO-BPy) are prepared according to published
procedures. Briefly, a 1:1 mass ratio of arc-discharge carbon nanotube soot
(Sigma-Aldrich, #698695) and PFO-BPy (American Dye Source, Inc.,
Quebec, Canada; #ADS153-UV) are combined at a concentration of 2 mg/mL
in toluene. The polymer and nanotube solution are sonicated using a horn-tip
ultrasonicator (Fisher Scientific, Waltham, MA; Sonic Dismembrator 500) at
40% amplitude for 10 min. After sonication, the dispersion is centrifuged, and
the supernatant is retained. The supernatants are concentrated and filtered
through a 5 µ m filter. The filtered solution is then centrifuged for 18–24 h.
After centrifugation, the pellets are retained and dispersed using horn-tip
ultrasonication in fresh toluene. These centrifugation and sonication steps are
repeated to remove excess PFO-BPy in solution until the nanotube: PFO-BPy
ratio is approximately 1:1. The resulting high semiconducting purity (>
99.98%) carbon nanotube (typical diameter ~ 1.5 nm, typical length ~ 500 nm)
pellets are dispersed at a concentration of 30 µ g/mL in ethanol stabilized
53

chloroform (Fisher Scientific, #C606SK-1) and the nanotube inks are aligned
using the previously developed FESA process [24, 25]. The CNT arrays were
boiled in toluene at 120 ℃ for 1 h to remove excess PFO-BPy. Upon
completion, CNT arrays were annealed in vacuum at a base pressure of 10
–5

torr and temperature of 400 ℃ for 1 h.

3.3 Characterization of aligned carbon nanotube
arrays
The high-density, well-aligned semiconducting nanotubes in this study
were deposited using the FESA technique. The scanning electron microscope
(SEM) images show the uniform aligned nanotube arrays reside in stripes of
width ~ 30 m (Figure 3.1a). Figure 3.1b shows a zoomed-in image of the
aligned region. The tubes are well-aligned with high densities. Inset is the
Raman spectra of FESA CNT. The Raman spectrum shows a G-band/D-band
peak ratio as high as 17, which verifies that our carbon nanotube arrays are of
high quality and with low defects. Figure 3.1c shows the SEM image of the
channel area of FESA CNT FET with channel length = 100 nm, where
individual carbon nanotubes can be clearly identified inside the channel area.
The nanotube packing density of the FESA CNT film in channel area is
54

determined as ~ 50 tubes/μm from SEM measurements. Figures 3.1d and 3.1e
depict the Atomic Force Microscopy (AFM) image and height profile of the
FESA CNT array. Figure 3.1f is the optical image of the as-made CMOS
inverter, which is constructed of p-type and n-type FETs based on the aligned
CNT array. With the highly aligned and dense nanotube array, excellent
transconductance and on-state current can be expected, which would advance
the electronic performance of CNT transistors and provide an excellent
platform for CNT electronics.

Figure 3.1 Characteristics of aligned carbon nanotube arrays. (a, b) SEM images of the
aligned tubes of a larger area (a) and aligned region (b). Inset is the Raman spectra of
aligned carbon nanotube arrays. (c) SEM image of the channel area of a CNT FET (Lch =
100 nm). (d, e) AFM image and height profile of aligned nanotube array. The range of the
50 μm
1 μm
200 nm
d
b c a
e f
500 nm
V
in
V
dd
V
out
Gnd
100 μm
55

color scalebar to the right of the AFM image in (d) is 7 nm. (f) Optical image of the as-
made CMOS inverter based on aligned carbon nanotube arrays.

3.4 Fabrication and measurement of carbon
nanotube devices
The schematic diagram of the back-gated device structure based on the
FESA CNTs is depicted in Figure 3.2a. The active channel areas were
selectively chosen to land in the aligned regions. Heavily p-type doped Si was
used as the common back gate. The gate dielectric is a thermally grown 15
nm thick SiO
2
layer. High-semiconducting-ratio, densely aligned carbon
nanotubes were deposited onto the SiO
2
surface via the FESA technique.
Alignment markers were patterned followed by identifying the aligned CNT
regions with SEM. The CNTs outside the device channel regions were etched
away by oxygen plasma. During the fabrication process, poly(methyl
methacrylate) (PMMA) layers were spin-coated onto the Si/SiO
2
surface, and
then electron beam lithography was used to define the interested region on the
wafer. All metal stacks were deposited at 10
−7
Torr using an e-beam
evaporator. Pure 30 nm palladium (Pd) formed the source and drain metal
contacts of the transistors Channel lengths were patterned with 100 nm, 200
nm, 500 nm, 1 μm, 2 μm, 5 μm, and 10 μm with the same channel width (W
56

= 2 μm) to study the length scaling behavior. The fabricated aligned CNT
array FETs were measured with a probe station and semiconductor parameter
analyzer to characterize the electrical performance. Holes are the majority
carriers injecting into CNTs from Pd electrodes either via tunneling or thermal
emission, which leads to typical p-type behavior transistors. The black arrows
in the following transfer curves show the forward and backward curves. The
electronic properties were measured using a Keithley 4200-SCS
semiconductor characterization system.
The electrical performance of CNT FET with channel length = 100 nm
is depicted in Figure 3.2b and 3.2c. In Figure 3.2b, the red curve is in the linear
scale while the blue curve is in the logarithm scale. We achieved high on-state
current density of ~ 520 μA/μm, peak transconductance normalized by width
g
m
of ~ 163.2 μS/μm, and on/off ratio of ~ 5,000 for the back-gated CNT
transistor. We studied further on FESA CNT FETs with longer channel
lengths. Figure 3.2d presents the I
ds
−V
gs
transfer curves of a representative
CNT FET with channel length = 10 μm. We can observe that the transistor
shows high on/off ratio of ~ 2 × 10
5
, peak transconductance normalized by
width g
m
~ 1.2 μS/μm, and high current density of ~ 2.5 μA/μm even under
relatively small drain bias. The I
ds
–V
ds
output curves of the same CNT FET
were plotted in Figure 3.2e, which shows output current saturation behavior.
57

A negatively biased drain voltage of 1 V is applied here to measure p-type
CNT FETs. In the subsequent measurements, V
ds
would be positively biased
to measure the n-type performance of FETs. We also quantified hysteresis by
characterizing the voltage difference (δV) when the drain current equaled half
of the maximum value in the measurement. Notably, all of the CNT FETs
show large hysteresis (δV = 5.9 V for channel length = 100 nm, δV = 4.3 V for
channel length = 10 μm) under double sweep DC measurement. The origin of
the prominent hysteresis is believed to be the adsorption of moisture and
oxygen at the CNT/SiO
2
interface [38–39]. Hydroxyl/water groups at the CNT
/ SiO
2
surface form the electron traps, which are the important factors of large
hysteresis in FETs. The electrons in the traps in the vicinity of the CNTs
would provide negative gating towards the on-state and lead to large
hysteresis. Therefore, ALD passivation is believed to be one route to minimize
the effects of oxygen and moisture and drastically reduce the hysteresis [30].
Channel length scaling behaviors of FESA CNT FETs are plotted in Fig 3.2f
to further investigate and understand the carrier transport mechanism and its
dependence on the channel lengths. It is found that the on/off ratio decreases
with decreasing channel length. There can be a combination of various factors
behind the decreasing on/off ratio phenomenon. One possible factor can be
the short channel effects, which include possible tunneling and barrier
58

lowering effects [40]. Another possible factor is the existence of metallic
tubes directly bridging the source and drain. Other possibilities include the
reduced gate control over the channel with short-channel length as well as the
tube-tube interaction due to the densely aligned nature (cross tube junctions
are not frequent but can sometimes happen in the active short channel area)
[41]. Improving the nanotube pitch uniformity and alignment with minimal
inter tube interaction would be necessary in order for CNT arrays to be the
candidate for beyond-silicon nano-electronics. As for the on-state current
density, it decreases with the increasing channel length, but the extracted field
effect mobility (estimated using the standard FET model, see Figure 3.3 in the
for more details) shows a relatively weak dependence on channel length
(Figure 3.3a) at long channel lengths, which is in agreement with the
percolation transport mechanism in carbon nanotube devices [42]. On the
contrary, the on-state current and the field effect mobility increase
dramatically with the decreasing channel length in the short-channel regime
where the channel length is scaled down to nanotube length of ~ 500 nm
(Figure 3.3b). This phenomenon can be attributed to the domination of direct
contact transport in short channel regime [25]. These experimental results
demonstrate the exceptional electrical properties of FESA CNT and can shed
light on the importance of charge transport in CNT electronics.
59

Figure 3.2 Electronic properties of CNT FETs based on aligned nanotube arrays. (a)
Schematic diagram of a back-gated carbon nanotube transistor. (b) Ids–Vgs curves of a
typical short channel nanotube transistor (Lch = 100 nm) in both linear (red) and log (blue)
scales. (c) Ids–Vds curves of the same transistor in (b). (d, e) Ids–Vgs curves and Ids–Vds curves
of a typical long channel nanotube transistor (L ch = 10 μm). (f) The averaged on-state
current density and on/off ratio for carbon nanotube transistors with different channel
lengths. Error bars represent standard deviations from 10 FETs for each channel length.

a c b
d e f
60

Figure 3.3 Extracted carrier mobility for CNT field effect transistors (FETs) with different
channel lengths from 100 nm to 10 μm. The carrier mobility was calculated using the
following equation:
μ =
𝐿 𝑊
1
𝐶 𝑜𝑥
𝑉 𝑑𝑠

𝑑 𝐼 𝑑𝑠
𝑑 𝑉 𝑔𝑠

where L and W are the channel length and width in the FET, respectively. Ids is the drain
current, Vds is the source-drain voltage, Vgs is the gate voltage, and Cox is the gate
capacitance per unit area.

Complementary electronics offers low static power consumption as
significant power is only drawn when the CMOS circuits are switching
between on and off states. However, CNT FETs often show p-type behavior
in the air. An effective n-type conversion method is needed for CNT CMOS
integrated circuit applications. We performed a systematic study of ALD
passivation on FESA CNT FETs with the different channel lengths, and the
experimental results are summarized in Figure 3.4. The schematic process
diagrams are illustrated in Figure 3.4a. For the ALD approach, a 40 nm Al
2
O
3

layer was deposited via ALD to form the gate oxide and 1 nm/50 nm Ti/Au
stack was utilized as the top gate electrode. Besides atomic layer deposition,
liquid metal reaction media such as Gallium-based liquid alloys can also serve
as a solvent for synthesizing Al
2
O
3
on the aligned CNT array [43]. In addition,
we tried an aluminum (Al) assisted ALD passivation approach following the
work pioneered by Tang et al. and Wei et al. [22, 31], in which a 1 nm thick
Al layer was first evaporated on the channel area. Devices were then heated
61

on the hot plate (120 ℃ for 1 h) to let the Al layer oxidize in air. We found
that the formed AlO
x
layer would aid the subsequent ALD process to ensure
a harmless and high-yield approach. I
ds
–V
gs
transfer curves (both linear and
logarithm scale) of typical long channel device (channel length = 10 μm) and
short channel device (channel length = 100 nm) with only ALD treatment are
shown in Figure 3.4b and 3.4c respectively. These devices both show a high
on/off ratio above 10
6
and nearly hysteresis-free behavior (δV < 0.5 V) as the
ALD layers can keep nanotubes passivated from oxygen molecules and
moisture in the air. The long channel device shows on-state current density ~
2.2 μA/μm, which are slightly lower than the current before the ALD process.
And the long channel devices were turned into ambipolar behavior with n-
branch current instead of predominant p-type behavior. The low temperature
(80 ℃) ALD process baked devices and drove away the oxygen molecules
adsorbed in the vicinity of nanotubes. Meanwhile, the deposited Al
2
O
3
layer
covered nanotubes and prevented the oxygen molecules from being absorbed
again. Besides, it is reported that the ALD Al
2
O
3
layer can be rich in positive
charges because of the deficiency of oxygen atoms, which gives rise to
electron conduction and leads to ambipolar behavior [30, 44-46]. The short
channel device shows on-state current density ~ 460 μA/μm, peak
transconductance normalized by width g
m
~ 210 μS/μm, subthreshold swing
62

~ 300 mV/dec, and on/off ratio > 4 × 10
6
. The on/off ratio increased from ~
5,000 for devices without ALD to ~ 10
6
for devices with ALD. We attribute
this increase in on/off ratio to the ALD passivation. It is known that the
adsorbed oxygen molecules and hydroxyl/water groups on the surface can
form electron traps. The electrons in the traps in the vicinity of the CNTs can
provide negative gating effect biasing the CNTs toward the on state, thus
leading to higher off-state current in the I
ds
-V
gs
measurements [38–39, 47].

Regarding the devices with ALD coating, the deposited ALD layer can keep
nanotubes passivated from oxygen molecules and moisture in the air, thus
leading to low off-state current and high on/off ratio.

Figure 3.4 Electronic properties of CNT FETs under ALD passivation. (a) Schematic
diagram of a top-gated CNT transistor. (b, c) Ids–Vgs curves of typical long channel (Lch =
10 μm, b) and short channel (Lch = 100 nm, c) CNT transistors in both linear (red curve)
a
b
c
d e f
63

and log (blue curve) scales after 40 nm Al2O3 ALD passivation. (d, e) Ids–Vgs curves of
typical long channel (Lch = 10 μm, d) and short channel (Lch = 100 nm, e) CNT transistors
in both linear (red curve) and log (blue curve) scales after 1 nm Al layer
evaporation/oxidation and 40 nm Al2O3 ALD passivation. (f) Schematic illustration of n-
type conversion process.

However, CMOS electronics prefer separate p-type and n-type behavior
transistors instead of ambipolar behavior transistors. The simple ALD Al
2
O
3

passivation approach presented above is not sufficient and further
developments are needed to convert the FESA CNT devices into n-type
behavior. Channel doping techniques have been investigated a lot as an n-type
conversion approach for CNT FETs. But the traditional organic dopants often
bring poor stability and fabrication compatibility issues, which limit the
application range of doping techniques [33, 37]. Here we utilize a 1 nm Al
layer to assist the Al
2
O
3
ALD passivation layer and get the n-type behavior
CNT transistors. The electrical performance of the long channel device
(channel length = 10 μm) after Al doping and ALD is shown in Fig 3.4d.
Unipolar n-type FESA CNT FET was successfully achieved with hysteresis-
free behavior (δV < 0.5 V). The device showed an n-branch on-current density
~ 2 μA/μm and on/off ratio > 2 × 10
6
. This method also offers the benefit of
long-term stability in air. The electrical performance of the same device
almost kept the same even after being exposed in ambient condition for 3
64

months (Figure 3.5). The evaporated thin Al layer oxidized in air and served
as a seeding layer between FESA CNT surface and ALD Al
2
O
3
layer. The
formed native AlO
x
can introduce more positive fixed charges at the
nanotube/ALD interface [30, 45, 46]. These positive charges would generate
electrical fields and bend the nanotube energy band downward. In long
channel devices (channel length >> nanotube length), tube-tube junctions
affect charge transport mostly and conduction is in fact a percolation process.
Charge transport would be determined by the conduction that can surpass the
average tube-tube junction barrier height to provide a connected path. Under
the Al layer and ALD Al
2
O
3
treatment, the bulk of the nanotubes in the long-
channel FETs were converted to n-type, enabling the FETs to show n-type
transistor behavior. But the situation becomes different when it comes to short
channel devices (channel length < nanotube length). Figure 3.4e shows the I
ds
-
V
gs
curves in linear scale and logarithm scale for FESA CNT FET with
channel length equals 100 nm. It has a p-branch on-state current density ~ 420
μA/μm. The device shows ambipolar behavior with a stronger n-branch on-
state current density ~ 2 μA/μm. In the short channel regime, tube-tube
junctions hardly exist in the channel area and the metal/nanotube junctions
dominate the charge transport. Carriers overcome the metal/nanotube contacts
and transport in the nanotube directly. Palladium (Pd) has a large work
65

function, and usually the Fermi level aligns with the valence band of
nanotubes, which results in larger Schottky barriers for electrons even when
nanotubes are passivated. Therefore, hole conduction is still much stronger
than electron conduction, which lets the devices exhibit larger p-type current
than n-type current. I
ds
–V
gs
curves of typical FESA CNT FETs with various
channel lengths (from 100 nm to 10 μm) after Al layer evaporation and Al
2
O
3

passivation at the same V
ds
are presented Figure 3.6, which shows the device
changed from almost ambipolar to predominant n-type behavior. The
increased surface scattering in longer channel devices lead to less steep
subthreshold transition than shorter-channel devices [48–49].
66

Figure 3.5 Air-stable performance of CNT FET. (a-b) Ids-Vds curves of an n-type behavior
CNT FET (Al + Al2O3 treatments, Pd source/drain metal contacts, L ch = 10 μm) (a), and
FET after leaving in air for three months (b). (c-d) Ids-Vds curves of an n-type behavior
CNT FET (Al + Al2O3 treatments, Ti source/drain metal contacts, L ch = 100 nm) (c), and
FET after leaving in air for three months (d).

67

Figure 3.6 Ids–Vgs transfer curves at Vds= 1 V of CNT FETs (Pd source/drain metal contacts)
with various channel lengths after 1 nm Al evaporation/oxidation and Al2O3 passivation.

As channel length scaled down to less than nanotube length, the devices
approach contact-dominated transport regime and Al layer + ALD Al
2
O
3

method is not sufficient to convert short channel FESA CNT FETs to n-type
transistors. Alternatively, metal contact engineering has the possibility to
obtain n-type conduction from the nanotubes [23, 33]. Low-work-function
metals would allow the Fermi level of the metal contacts to align with the
conduction band of carbon nanotubes and lead to large Schottky barriers for
holes but ohmic contacts for electrons. However, this technique nevertheless
68

has some drawbacks. For example, low-work-function metals such as Gd or
Sc are very susceptible to air and moisture and can be easily oxidized, which
gives rise to high resistance and leads to the degraded output current.
Therefore, it is of great importance to develop a durable metal contact
engineering technique to convert FESA CNT FETs to n-type behavior. Figure
3.7a depicts the schematic diagram of n-type conversion techniques of FESA
CNT short channel FETs. Back-gated FESA CNT FETs with 30 nm Titanium
(Ti, with a work function of 4.3 eV) as source/drain metal contacts were
fabricated firstly. Figure 3.7b presents the transfer characteristics for a typical
back-gated FET (channel length = 100 nm) measured at V
ds
= 1 V. The device
still shows p-type behavior with a relatively small on-state current density ~
1 μA/μm and on/off ratio ~ 10,000. The transistor polarity of CNT FETs with
short channel devices (channel length < nanotube length), is dominated by the
work function of source/drain metal contacts. Palladium has a large work
function (5.3 eV), and usually the Fermi level aligns with the valence band of
nanotubes, which results in strong hole conduction and leads to large p-type
current. On the contrary, titanium forms low work function metal contacts (4.3
eV), and the Fermi level is far away from the nanotube valence band, which
results in a large barrier for holes conduction and leads to smaller p-branch
current. Nanotubes are heavily p-type due to the adsorption of oxygen
69

molecules, so even the devices with Ti contacts exhibit p-type behavior in air
rather than n-type behavior. To suppress the p-type conduction in devices,
ALD Al
2
O
3
passivation layer was deposited on top of the device as gate
dielectric. Low temperature is a crucial factor because Ti is likely to get
oxidized under elevated temperatures during the ALD process. A 1 nm/50 nm
Ti/Au stack was then deposited by e-beam evaporation as the gate electrodes.
The I
ds
–V
gs
transfer curves of the device after passivation are shown in Figure
3.7c. Interestingly, the figure indicates that the n-type conduction becomes
predominant and the p-type conduction is suppressed. The device shows n-
type on-state current density ~ 50 μA/μm, peak transconductance normalized
by width g
m
~ 20.1 μS/μm, and on/off ratio > 10
5
. After passivation, the Fermi
level of FESA CNTs shifts towards conduction band because of desorption of
oxygen, resulting in n-type behavior. We then further tried a 1 nm Al layer +
ALD Al
2
O
3
technique to improve the n-type performance. The I
ds
–V
gs
transfer
curves are shown in Figure 3.7d and the device shows an n-type on-state
current ~ 130 μA/μm, peak transconductance normalized by width g
m
~ 54.7
μS/μm, with on/off ratio > 2 × 10
6
. The formed AlO
x
at nanotube/Al
2
O
3
interface would bend the energy band downward, and together with the Ti
contacts make the barrier for electrons much smaller than before, resulting in
n-type behavior with decent performance. Figure 3.8 shows I
ds
– V
gs
transfer
70

curves for 10 n-type CNT FETs (channel length = 100 nm) without 1 nm Al
layer (a) and with 1 nm Al layer (b). Using thinner dielectric layer with higher
dielectric constant will to improve the gate control and lead to higher
subthreshold swing. We note that the p-type transistors showed more Schottky
behavior than the n-type transistors, while n-type transistors showed smaller
current density than the p-type devices. For nanotubes with diameters < ~ 2
nm, it was observed that Schottky barriers at the Pd contacts are small but
cannot be completely eliminated as the bandgap of single-wall nanotubes is
inversely proportional to the diameter [6]. The diameter of our carbon
nanotube arrays is ~ 1.5 nm, and hence a small Schottky barrier is expected.
Further engineering of the metal contacts by using different metals or by
annealing may lead to elimination of Schottky barriers for p-type transistors.
On the other hand, electron transport in our n-type devices may be affected by
scattering due to the fixed oxide charges, resulting in lower current density.
We also note that our long-channel devices show Ohmic behavior while the
short-channel transistors show slight Schottky behavior. The reason is that for
long-channel devices, the transport is dominated by scattering/hopping of
carriers in the channel instead of contacts, and thus the slight Schottky
behavior of contacts has become unnoticeable in the I
ds
–V
ds
curves.
71

Figure 3.7 Electronic properties of CNT FETs with Ti source/drain contacts. (a) Schematic
diagram of a top-gated CNT transistors with Ti source/drain contacts. (b) Ids–Vgs curves of
a typical short channel back-gated CNT transistors (Lch = 100 nm) in both linear (red) and
log (blue) scales. (c) Ids–Vgs curves of the transistor in (b) after 40 nm Al2O3 ALD
passivation. (d) Ids–Vgs curves of a typical short channel CNT FET (Lch = 100 nm) after 1
nm Al layer evaporation/oxidation and 40 nm Al2O3 ALD passivation.

72

Figure 3.8 (a) Ids-Vgs curves of 10 n-type behavior CNT FETs (Al2O3 treatments, Ti
source/drain metal contacts, Lch = 100 nm). (b) Ids-Vgs curves of 10 n-type behavior CNT
FETs (Al + Al2O3 treatments, Ti source/drain metal contacts, Lch = 100 nm).

Furthermore, an integrated CMOS inverter was demonstrated with the
as-made p-type and n-type FESA CNT FETs (channel length = 100 nm). FETs
with Ti source/drain contacts are selected and followed by 1 nm Al layer
evaporation and 80 ℃ Al
2
O
3
ALD to achieve the n-type behavior transistors.
The p-type transistors are the CNT FETs with Pd source/drain contacts after
ALD passivation treatment. The channel widths of the p-type behavior FET
and n-type behavior FET are designed to obtain similar drain current (W
p
= 2
μm, W
n
= 7 μm). The output characteristics of CNT-based CMOS inverter is
shown in Figure 3.9b. The inverter works with V
ds
= 1 V, and the input voltage
is swept from –2 to 2 V. The output voltage switched from 1 V to 0 V and
results in a maximum gain of 5, suggesting the successful operation of FESA
73

CNT inverter. We note that the input voltage of the inverter had wider range
than the output voltage because the 40 nm ALD dielectric we used required
higher voltage to turn and turn off the transistors. In the future, we can use
thinner dielectric layer to improve the gate control and get the inverter with
symmetric input and output and rail-to-rail operation.

Figure 3.9 CMOS inverter circuit using the as-made p-type and n-type CNT FETs. (a) Ids–
Vds curves of a p-type behavior CNT FET (Al2O3 treatments, Pd source/drain metal contacts,
Lch = 100 nm) and an n-type behavior CNT FET (Al + Al2O3 treatments, Ti source/drain
metal contacts, Lch = 100 nm). (b) Voltage transfer characteristics of the CMOS inverter
based on aligned CNT array. (c) Plot of inverter gain versus input voltage

3.5 Summary
Our work presents the significant advance in the fabrication of n-type densely
aligned FESA CNT FETs and the application of FESA CNTs in CMOS logic
circuits. We systematically studied the effects of ALD passivation and metal
contact engineering on devices with various channel lengths. For long channel
(channel length >> nanotube length) devices, a thin Al layer followed by
74

oxidation and subsequent ALD passivation are sufficient for the carrier type
conversion and resulted in n-type behavior. We then further revealed that for
FESA CNT FETs (with ALD passivation) with short channel devices (channel
length < nanotube length), the transistor polarity is dominated by the work
function of source/drain metal contacts. Air-stable n-type behavior CNT FETs
can be achieved using the Ti contacts and ALD passivation method and
showed high on-state current density ~ 130 μA/μm with on/off ratio of >10
6
.
Moreover, a CMOS inverter has been demonstrated using as-obtained p-type
and n-type FESA CNT FETs. These findings help the understanding of the
device physics and would benefit future high-performance air-stable CNT
FETs. Our achievements may open a door to future large-scale and energy-
efficient CNT based CMOS computing units.

3.6 Reference
[1] Tans, S.; Verschueren, A.; Dekker, C. Room-temperature transistor based
on a single carbon nanotube. Nature 1998, 393, 49-52.
[2] Sazonova, V.; Yaish, Y.; Üstü nel, H.; Roundy, D.; Arias, T. A.; McEuen,
P. L. A tunnable carbon nanotube electromechanical oscillator. Nature
2004, 431, 284-287.
75

[3] Avouris, P.; Chen, Z.; Perebeinos, V. Carbon-based electronics. Nat.
Nanotech. 2007, 2, 605-456.
[4] De Volder, M. F. L.; Tawfick, S. H.; Baughman, R. H.; Hart, A. J. Carbon
Nanotubes: present and future commercial applications. Science 2013,
339, 535–539.
[5] Zhou, C.; Kong, J.; Yenilmez, E.; Dai, H. Modulated Chemical Doping of
Individual Carbon Nanotubes Carbon nanotube computer. Science 2000,
290, 1552-1555.
[6] Javey, A.; Guo, J.; Wang, Q.; Lundstrom, M.; Dai, H. J. Ballistic carbon
nanotube field-effect transistors. Nature 2003, 424, 654-657.
[7] Kang, S.; Kocabas, C.; Ozel, T.; Shim, M.; Pimparkar, N.; Alam, M. A.;
Roktin, S. V.; Rogers, J. A. High-performance electronics using dense,
perfectly aligned arrays of single-walled carbon nanotubes. Nat. Nanotech.
2007, 2, 230–236.
[8] Shulaker, M. M.; Hills, G.; Patil, N.; Wei, H.; Chen, H. Y.; PhilipWong,
H. S.; Mitra, S. Carbon nanotube computer. Nature 2013, 501, 526-530.
[9] Geier, M.; McMorrow, J.; Xu, W.; Zhu, J.; Kim, C. H.; Marks, T. J.;
Hersam, M.C. Solution-processed carbon nanotube thin-film
complementary static random access memory. Nat. Nanotech. 2015, 10,
944–948.
76

[10] Qiu, C. G.; Zhang, Z. Y.; Xiao, M. M.; Yang, Y. J.; Zhong, D. L.; Peng,
L. M. Scaling carbon nanotube complementary transistors to 5-nm gate
lengths. Science 2017, 355, 271-276.
[11] Han, S. J.; Tang, J.; Kumar, B.; Falk, A.; Farmer, D.; Tulevski, G.;
Jenkins, K.; Afzali, A.; Oida, S.; Ott, J.; Hannon, J.; Haensch, W. High-
speed logic integrated circuits with solution-processed self-assembled
carbon nanotubes. Nat. Nanotech. 2017, 12, 861–865.
[12] Liu, J.; Wang, C.; Tu, X.; Liu, B.; Chen, L.; Zheng, M.; Zhou, C.
Chirality-controlled synthesis of single-wall carbon nanotubes using
vapour-phase epitaxy. Nat. Commun. 2012, 3, 1199.
[13] Yang, F.; Wang, X.; Zhang, D.; Yang, J.; Luo, D.; Xu, Z.; Wei, J.;
Wang, J. Q.; Xu, Z.; Peng, F.; Li, X.; Li, R.; Li, Y.; Li, M.; Bai, X.; Ding,
F.; Li, Y. Chirality-specific growth of single-walled carbon nanotubes on
solid alloy catalysts. Nature 2014, 510, 522-524.
[14] Zhang, S. C.; Kang, L. X.; Wang, X.; Tong, L. M.; Yang, L. W.; Wang,
Z. Q.; Qi, K.; Deng, S. B.; Li, Q. W.; Bai, X. D.; Ding, F.; Zhang, J. Arrays
of Horizontal Carbon Nanotubes of Controlled Chirality Grown Using
Designed Catalysts. Nature 2017, 543, 234-238.
[15] Lipomi, D.; Vosgueritchian, M.; Tee, BK.; Hellstrom, S. L.; Lee, J. A.;
Fox, C. H.; Bao, Z. Skin-like pressure and strain sensors based on
77

transparent elastic films of carbon nanotubes. Nat. Nanotech. 2011, 6,
788–792.
[16] Sun, Dm.; Timmermans, M.; Tian, Y.; Nasibulin, A. G.; Kauppinen, E.
I.; Kishimoto, S.; Mizutani, T.; Ohno, Y. Flexible high-performance
carbon nanotube integrated circuits. Nat. Nanotech. 2011, 6, 156–161.
[17] Park, S.; Vosguerichian, M. and Bao, Z. A review of fabrication and
applications of carbon nanotube film-based flexible electronics.
Nanoscale 2013, 5, 1727-1752.
[18] Peng, L. M.; Zhang, Z.; Qiu, C. Carbon nanotube digital electronics.
Nat. Electron. 2019, 2, 499-505.
[19] Xiang, L.; Zhang, H.; Dong, G.; Zhong, D.; Han, J.; Liang, X.; Zhang,
Z.; Peng, L. M.; Hu, Y. Low-power carbon nanotube-based integrated
circuits that can be transferred to biological surfaces. Nat. Electron. 2018,
1, 237–245.
[20] Franklin, A. D.; Chen, Z. H. Length scaling of carbon nanotube
transistors. Nat. Nanotechnol. 2010, 5, 858-862.
[21] Cao, Q.; Tersoff, J.; Farmer, D. B.; Zhu, Y.; Han, S. J. Carbon nanotube
transistors scaled to a 40-nanometer footprint. Science 2017, 356, 1369-
1372.
78

[22] Tang, J.; Cao, Q.; Tulevski, G.; Jenkins, K.; Nela, L.; Farmer, D. B.;
Han, S. J. Flexible CMOS integrated circuits based on carbon nanotubes
with sub-10 ns stage delays. Nat. Electron. 2018, 1, 191–196.
[23] Hills, G.; Lau, C.; Wright, A.; Fuller, S.; Bishop, M. D.; Srimani, T.;
Kanhaiya, P.; Ho, R.; Amer, A.; Stein, Yosi; Murphy, D.; Arvind;
Chandrakasan, A.; Shulaker, M. M. Modern microprocessor built from
complementary carbon nanotube transistors. Nature 2019, 572, 595–602.
[24] Joo, Y.; Brady, G. J.; Arnold, M. S.; Gopalan, P. Dose-Controlled,
Floating Evaporative Self-assembly and Alignment of Semiconducting
Carbon Nanotubes from Organic Solvents. Langmuir 2014, 30, 3460-
3466.
[25] Brady, G. J.; Way, A. J.; Safron, N. S.; Evensen, H. T.; Gopalan, P.;
Arnold, M. S. Quasi-ballistic carbon nanotube array transistors with
current density exceeding Si and GaAs. Sci. Adv. 2016, 2.
[26] Liu, L.; Han, J.; Xu, L.; Zhou, J.; Zhao, C.; Ding, S.; Shi, H.; Xiao, M.;
Ding, L.; Ma, Z.; Jin, C.; Zhang, Z.; Peng, L. M. Aligned, high-density
semiconducting carbon nanotube arrays for high-performance electronics.
Science, 2020, 368, 850-856.
[27] Cao, Y.; Brady, G. J.; Gui, H.; Rutherglen, C.; Arnold, M. S.; Zhou, C.
Radio Frequency Transistors Using Aligned Semiconducting Carbon
79

Nanotubes with Current-Gain Cutoff Frequency and Maximum
Oscillation Frequency Simultaneously Greater than 70 GHz. ACS Nano
2016, 10, 6782–6790.
[28] Rutherglen, C.; Kane, A. A.; Marsh, P. F.; Marsh, P. F.; Cain. T. A.;
Hassan, B. I.; Alshareef, M. R.; Zhou, C.; Galatsis, K. Wafer-scalable,
aligned carbon nanotube transistors operating at frequencies of over
100 GHz. Nat. Electron. 2019, 2, 530–539.
[29] Zhong, D.; Shi, H.; Ding, L.; Zhao, C.; Liu, J.; Zhou, J.; Zhang, Z.;
Peng, L. M. Carbon Nanotube Film-Based Radio Frequency Transistors
with Maximum Oscillation Frequency above 100 GHz. ACS Appl. Mater.
Interfaces 2019, 11, 42496-42503.
[30] Zhang, J.; Wang, C.; Fu, Y.; Che, Y.; Zhou, C. Air-stable conversion
of separated carbon nanotube thin-film transistors from p-type to n-type
using atomic layer deposition of high-κ oxide and its application in CMOS
logic circuits. ACS Nano 2011, 5, 3284–3292.
[31] Wei, H.; Chen, H. Y.; Liyanage, L.; Wong, H. S. P.; Mitra, S. Air-stable
technique for fabricating n-type carbon nanotube FETs. IEDM Tech. Dig.
2011, 23.2.1–23.2.4.
80

[32] Li, G.; Li, Q.; Jin, Y.; Zhao, Y.; Xiao, X.; Jiang, K.; Wang, J.; Fan, S.
Fabrication of air-stable n-type carbon thin-film transistors on flexible
substrates using bilayer dielectrics. Nanoscale 2015, 7, 17693-17701.
[33] Tang, J.; Farmer, D.; Bangsaruntip, S.; Chiu, K. C.; Kumar, B.; Han, S.
J. Contact engineering and channel doping for robust carbon nanotube
NFETs. 2017 Int. Symp. VLSI Tech. Syst. Appl.
[34] Yang, Y.; Ding, L.; Han, J.; Zhang, Z; Peng, L.M.; High-performance
complementary transistors and medium-scale integrated circuits based on
carbon nanotube thin films. ACS Nano 2017, 11, 4124–4132.
[35] Brady, G. J.; Joo, Y.; Roy, S. S.; Gopalan, P.; Arnold, M. S. High
performance transistors via aligned polyfluorene-sorted carbon nanotubes.
Appl. Phys. Lett. 2014, 104, 083107.
[36] Kang, D.; Park, N.; Ko, J. H.; Bae, E.; Park, W. Oxygen-induced p-type
doping of a long individual single-walled carbon nanotube.
Nanotechnology 2005, 16, 1048-1052.
[37] Avery, A.; Zhou, B.; Lee, J.; Lee, E. S.; Miller, E. M.; Ihly, R.;
Wesenberg, D.; Mistry, K. S.; Guillot, S. L.; Zink, B. L.; Kim, Y. H.;
Blackburn, J. L.; Ferguson, A. J. Tailored semiconducting carbon
nanotube networks with enhanced thermoelectric properties. Nat. Energy
2016, 1, 16033.
81

[38] Park, R.S.; Shulaker, M. M.; Hills, G.; Liyanage, L. S.; Lee, S.; Tang,
A.; Mitra, S.; PhilipWong, H. S. Hysteresis in Carbon Nanotube
Transistors: Measurement and Analysis of Trap Density, Energy Level,
and Spatial Distribution. ACS Nano 2016, 10, 4599-4608.
[39] Jin, S. H.; Islam, A. E.; Kim, T.; Kim, J.; Alam, M. A.; Rogers, J. A.
Sources of Hysteresis in Carbon Nanotube Field-Effect Transistors and
Their Elimination Via Methylsiloxane Encapsulants and Optimized
Growth Procedures. Adv. Funct. Mater. 2012, 22, 2276-2284.
[40] Franklin, A. D.; Farmer, D. B.; Haensch, W. Defining and Overcoming
the Contact Resistance Challenge in Scaled Carbon Nanotube Transistors.
Acs Nano 2014, 8, 7333-7339.
[41] Che, Y. C.; Wang, C.; Liu, J.; Liu, B. L.; Lin, X.; Parker, J.; Beasley,
C.; Wong, H. S. P.; Zhou, C. W. Selective Synthesis and Device
Applications of Semiconducting Single-Walled Carbon Nanotubes Using
Isopropyl Alcohol as Feedstock. Acs Nano 2012, 6, 7454-7462.
[42] Baranovskii, S. D.; Nenashev, A. V.; Oelerich, J. O.; Greiner, S. H. M.;
Dvurechenskii, A. V.; Gerbhard, F. Percolation description of charge
transport in the random barrier model applied to atmorphous oxide
semiconductors. Europhys. Lett. 2019, 127-57004.
82

[43] Zavabeti, A.; Zhang, B. Y.; Castro, I. A.; Ou, J. Z.; Carey, B. J.;
Mohiuddin, Md.; Datta, R. S.; Xu, C.; Mouritz, A. P.; McConville, C. F.;
O’Mullane, A. P.; Daeneke, T.; Kalantar-Zadeh, K. GREEN Synthesis of
Low-Dimensional Aluminum Oxide Hydroxide and Oxide Using Liquid
Metal Reaction Media: Ultrahigh Flux Membranes. Adv. Funct. Marer.
2018, 1804057.
[44] Jinkins, K. R.; Chan, J.; Jacobberger, R. M.; Berson, A; Arnold, M. S.
Substrate-Wide Confined Shear Alignment of Carbon Nanotubes for Thin
Film Transistors. Adv. Electron. Mater. 2019, 5, 1800593.
[45] Moriyama, N.; Ohno, Y.; Kitamura, T.; Kishimoto, S.; Mizutani, T.
Change in Carrier Type in High-k Gate Carbon Nanotube Field-E ffect
Transistors by Interface Fixed Charges. Nanotechnology 2010, 21,
165201.
[46] Rinkiol, M.; Johansson, A.; Zavodchikova, M.; Toppari, J.; Albert, G.;
Nasibulin, A.; Kauppinen, E.; Torma, P. HighYield of Memory Elements
from Carbon Nanotube Field- Effect Transistors with Atomic Layer
Deposited Gate Dielectric. New J. Phys. 2008, 10, 103019.
[47] Mudimela, P. R.; Grigoras, K.; Anoshkin, I. V.; Varpula, A.; Ermolov,
V.; Anisimov, A. S.; Nasibulin, A. G.; Novikov, S.; Kauppinen, E. I.
83

Single-Walled Carbon Nanotube Network Field Effect Transistor as a
Humidity Sensor. J Sensors 2012.
[48] Y. Cui, R. Xin, Z. Yu, Y. Pan, Z.-Y. Ong, X. Wei, J. Wang, H. Nan, Z.
Ni, Y. Wu, T. Chen, Y. Shi, B. Wang, G. Zhang, Y.-W. Zhang, X. Wang.
High-Performance Monolayer WS2 Field-Effect Transistors on High-κ
Dielectrics. Adv. Mater. 2015, 27, 5230.
[49] S. Kim, A. Konar, W.-S. Hwang, J. H. Lee, J. Lee, J. Yang, C. Jung, H.
Kim, J.-B. Yoo, J.-Y. Choi, Y. W. Jin, S. Y. Lee, D. Jena, W. Choi, K.
Kim. High-mobility and low-power thin-film transistors based on
multilayer MoS
2
crystals. Nat Commun. 2012, 3, 1011.

84

Chapter 4 Carbon nanotube radio
frequency electronics
4.1 Introduction
When nanoelectronics goes into ballistic transport regime, the resistance
is in the order of quantum resistance ~ h/e
2
(~ 25 kΩ). Such high resistance
seems to hamper the transit time and limit the operation frequency of nano
circuits. However, the quantum capacitance of nanoelectronics devices is very
small. For instance, given the finite quantum energy level spacing of electrons
in one-dimension (1D):
𝛿𝐸 =
𝑑𝐸
𝑑𝑘
𝛿𝑘 =
ℎ
2𝜋 𝑣 𝐹 2𝜋 𝐿
where L is the length of system, v
F
is the Fermi velocity. Carbon nanotube
(CNT) is a well know 1D material, and the fermi velocity of CNT is ~ 8 × 10
5

m/s. The Quantum capacitance per length of CNT can be derived by equating
the energy with e
2
:
𝐶 𝑄 =
𝑒 2
𝛿 𝐸 =
𝑒 2
ℎ𝑣 𝐹 ≈ 0. × 10
−16
𝐹 /𝜇𝑚
The electrostatic potential of CNT is also small:
85

𝐶 𝐸 =
2𝜋𝜖
ln (
ℎ
𝑑 )
≈ 0. × 10
−16
𝐹 /𝜇𝑚
where h is the distance to the ground and d is the diameter of nanotubes. For
a nanotube with 1μm length, the resistance can be ~ 6.25 kΩ [1]. The
theoretical limit of RC frequency is:
𝑓 =
1
2𝜋𝑅𝐶 ≈ 1 𝑇𝐻𝑧
Therefore, CNTs are considered one of the most promising materials for radio
frequency (RF) analogy electronics.
However, this estimation is only based on the property of carbon
nanotubes. The real performance of CNT devices is often limited by parasitic
impedance. The device structure of a typical CNT field-effect transistor (FET)
is shown in Figure 4.1 [2], the radio frequency performance is determined by
the intrinsic material property and the parasitic resistance/capacitance. The
most two important metrics to evaluate RF performance are current-gain cut
off frequency f
T
and power-gain cut-off frequency f
max
. The current/power
gain will decrease under high frequency and the frequency at which the gain
equals 1 is defined as the cut off frequency.
86

Figure 4.1 (a) Schematic diagram of a carbon nanotube FET. The fringe electrical fields
from the gate to the source and drain give rise to the parasitic capacitance. (b) A small-
signal equivalent circuit for a nanotube-based FET where gm is the transconductance, Cgs
the intrinsic gate capacitance, gd is the conductance, Cp,gs and Cp,gd are the gate–source and
gate–drain parasitic capacitances, Rp,s, Rp,d and Rgate are parasitic resistances for the source
and drain and gate electrode. Reprinted from ref. [2].

Based on the structure in Figure 4.1, f
T
and f
max
are defined as:
87

𝑓 𝑇 =
𝑔 𝑚 2𝜋 {(𝐶 𝑔𝑠
+ 𝐶 𝑝 ,𝑔𝑠
+ 𝐶 𝑝 ,𝑔𝑑
)[(𝑅 𝑝 ,𝑠 + 𝑅 𝑝 ,𝑑 )𝑔 𝑑 + 1] + 𝐶 𝑝 ,𝑔𝑑
𝑔 𝑚 (𝑅 𝑝 ,𝑠 + 𝑅 𝑝 ,𝑑 )}

𝑓 𝑚𝑎𝑥 ≈
𝑓 𝑇 2√𝑔 𝑑 (𝑅 𝑝𝑠
+ 𝑅 𝑔 ) + 2𝜋 𝑓 𝑇 𝐶 𝑝𝑑
𝑅 𝑔

Recently, researchers from Carbonics Inc. experimentally achieved
extrinsic f
T
> 80 GHz and f
max
> 100 GHz [3], showing the promising future
of carbon nanotube in the next-generation electronics. Achievements are
made from every aspect of material and device structure, including self-
aligned T-gate structure [4], high-density and high semiconducting ratio
aligned carbon nanotube arrays [5-6], and scaling of channel length and gate
length [7]. However, the lack of suitable back-end-of-line (BEOL) process
limits the progress of integrated CNT circuits. High quality passive
components are the key part of radio frequency applications. But the
development of BEOL process to fabricate compatible passive components is
time consuming and costly. So far, most CNT RF work mainly focused on
transistor itself [3-7]. Only on-chip ring oscillators have been demonstrated,
88

and complex integrated CNT circuits such as amplifiers, mixers, voltage
control oscillators were built via external passive device in discrete form [8-
11]. These issues raise the question about the possibility for CNT to realize its
potential and make a near-term entry in the RF market. To exceed the current
RF-CMOS technology soon, a scalable and low-cost method is required to
fabricate passive components and integrate CNT transistors and circuits on
the same chip.
Besides the development in passive devices fabrication, uniformly
distributed and high-density aligned carbon nanotube arrays are also needed
to realize the potential of CNT. Previous self-assemble process, such as FESA
(floating evaporative self-assembly) can deposit CNT inks on substrate and
form a high-density (~ 50 tubes/μm) aligned CNT arrays [12]. However,
aligned FESA CNTs normally reside in stripes of with ~ 30 μm. Between the
stripes are the transient regions with few micrometres wide that consist
misaligned nanotubes and random network [12]. Here, a solution-based shear
is applied to produce continuous films of high-density aligned CNT arrays. A
uniform deposition on a 10 × 10 cm
2
substrate was previously demonstrated,
which advance the development of on chip CNT integrated circuits [13].
By taking the advantage of flow-induced shear platform. We present the
design of on-ship radio receiver based on aligned carbon nanotube arrays. The
89

radio receiver is predicted to receive and demodulate information signal
carried on a GHz waveform. Device performance is fully preserved due to the
clean fabrication process. This novel integration scheme paves the way to
realize the promising future of CNT in large-scale digital circuits.

4.2 Fabrication and measurement of CNT devices
Semiconducting carbon nanotube inks isolated using polyfluorene
derivative polymer wrapper poly[(9,9-dioctylfluorenyl-2,7-diyl)-alt-co-(6,6’-
{2,2’-bipyridine})] (PFO-BPy) are prepared according to published
procedures [12-13]. Briefly, a 1:1 mass ratio of arc-discharge carbon nanotube
soot (Sigma-Aldrich, #698695) and PFO-BPy (American Dye Source, Inc.,
Quebec, Canada; #ADS153-UV) are combined at a concentration of 2 mg/mL
in toluene. The polymer and nanotube solution are sonicated using a horn-tip
ultrasonicator (Fisher Scientific, Waltham, MA; Sonic Dismembrator 500) at
40% amplitude for 10 min. After sonication, the dispersion is centrifuged in a
swing bucket rotor (Sorvall WX, TH-641, Thermo-Scientific) at 3 × 10
5
g for
10 min to remove soot and undispersed nanotubes. The upper 90% of the
supernatant was collected and concentrated to a volume of 60 mL. The
concentrated supernatant was centrifuged for 18–24 h and dispersed in toluene
90

via sonication utilizing the horn tip sonicator. The process of centrifugation
and sonication in toluene was repeated three times to remove most of the
excess PFO-BPy. The final solution was prepared by horn tip sonication of
the rinsed carbon nanotube pellet in chloroform (stabilized with ethanol from
Fisher Scientific, #C606SK-1). Figure 4.2 shows a wafer map created using
atomic force microscopy (AFM) image to evaluate CNT density and
alignment. From which we can observe the alignment direction and a uniform
packing density ~ 50 tubes/μm.

Figure 4.2 A typical wafer map created using the AFM image to evaluate CNT density
and alignment.

After aligned CNT arrays were deposited on the quartz wafer, alignment
markers were patterned followed by identifying the aligned regions with
91

scanning electron microscopy (SEM). The CNT arrays outside the designed
channel area were etched away by oxygen plasma. Source and drain electrodes
(1/100 nm thickness of titanium / pallium) were then patterned. Sequentially,
an aluminum T-shape gate (∼ 140 nm thickness) was formed by electron beam
lithography and metal deposition. Device will then be baked at 120
o
C to form
oxidized aluminum as gate dielectric. The thickness of oxidized aluminum is
~ 4 nm according to our previous publications [14-15]. Finally, 15 nm
thickness of palladium was deposited as the self-aligned source and drain
contacts. Figure 4.3a shows the schematic diagram of the self-aligned T-gate
structure. Figure 4.3b shows the zoomed in optical image of the T-gate
channel area. Figure 4.3c presents the cross-sectional view SEM image of a
T-gate structure after cleaving the substrate along the channel direction. The
T-gate structure provides aligned air gaps between source/drain metal contacts
and gate electrodes, which significantly reduce the parasitic capacitance and
lead to decent radio frequency performance. The channel length of self-
aligned T-gate transistor is the foot length of T-gate structure ~ 100 nm, and
the ungated region on each side of the T-gate is ~ 50 nm.
92

Figure 4.3 (a) Schematic diagram of self-aligned T-gate structure. (b) Optical image of the
T-gate channel area. (c) Cross-sectional SEM image of the T-gate structure.

Figure 4.4 (a) Transfer curve of typical self-aligned T-gate CNT devices. (b)
Transconductance and output conductance of the transistor in (a). (c) Current-gain and
power-gain frequency response of the transistor in (a). (d) Output power versus input power
of the fundamental and the third order frequency.

93

With the high-density assembled aligned nanotube arrays and self-
aligned T-gate structure, it is foreseen that the RF performance of transistor
should be much better. Figure 4.4 presents the direct current (DC) and RF
performance of CNT transistors based on combined use of flow-induced shear
CNT arrays and self-aligned T-gate structure. Figure 4.4a is the transfer
characteristic of a typical self-aligned T-gate CNT transistor, the red curve is
in the linear scale and the blue curve is in the logarithm scale. Transistor
shows an on-state current density ~330 μA/μm. This on-current density is
comparable with the best CNT RF transistors reported so far. The on/off ratio
is ~ 40 because of the ungated region on each side of T-gate. Fortunately, low
on/off ratio will not hamper the RF performance of CNT transistors. There is
not a direct relation between the on/off ratio and frequency response. Figure
4.4b presents the transconductance and output conductance respectively. At -
1 V drain bias, device exhibits a peak transconductance density ~ 300 μS/μm,
on-par with the state-of-art CNT RF FETs. Moreover, device shows a decent
output conductance density ~ 240 μS/μm. The RF performance is shown in
Figure 4.4c. PNA vector network analyzer and microprobes, together with
ground-signal-ground structure were utilized to measure the S-parameters. It
is well known that h
21
is the forward current gain, and h21 is given by:
94

h21 tells us the current gain of the transistor at a given operating frequency.
Current gain cut off frequency f
T
can be derived from h
21
. The maximum
available gain (MAG) is defined as:

where K is the stability factor. The frequency at which MAG = 0 dB is power
gain cut-off frequency f
max
. When the transistor was biased at V
ds
= -1 V and
V
gs
= -1.5 V, the f
T
is ~ 41 GHz and f
max
is ~ 30 GHz.
Linearity performance was presented in Figure 4.4d. Two RF signals
with frequencies of 1 GHz and 1.2 GHz were applied to the gate electrode
simultaneously, and a spectrum analyzer was used to measure the output
signal power level at drain metal pad. The output signal can be written in the
form:
95

where the fundamental signal increases with the slope ~ 10 dB/dec and the
third order signal increases with the slope ~ 30 dB/dec. The input third order
intercept point IIP
3
is determined to be ~ 24 dBm and the output third order
intercept point OIP
3
is determined to be ~ 8 dBm, which is comparable to 180
nm RF-CMOS technology [16].

4.3 Carbon nanotube integrated circuits
Figure 4.5 is the schematic diagram of as-made on-chip passive
components. Both capacitors and inductor are fabricated with two metal layers
and one dielectric layer on the quartz substrate. 1 nm/50 nm Ti/Au metal
stacks were utilized as the underlayer metal connection. Au was selected
because of its low resistivity, air stability and small Young’s modulus to
minimize fatigue [17]. 300 nm SiO
2
was deposited as the dielectric layer
because of its reliability and similar Young’s modulus with Au. Finally, 1
nm/500 nm Ti/Cu metal stacks were deposited as the upper layer metal pads.
96

Figure 4.5a, 4.5b and 4.5c shows the schematic image of as-made capacitors,
inductors, and CNT devices.

Figure 4.5 Schematic diagram of as-made on-chip passive components and CNT RF
devices.

Figure 4.6a shows the optical image of capacitors and inductors in GSG
structure for RF testing purpose. Figure 4.6b, 4.6c show the impedance versus
frequency. As expected, impedance of the as-made passive components
almost keeps the same when frequency varies from hundreds of MHz to
several tens of GHz.
97

Figure 4.6 (a) Optical image of the as-made passive components. (b), (c) are the impedance
frequency response of as-made passive components.

In modern electronic communication, modulation technique is often used
to transmit information signals with a radio carrier wave. Amplitude
modulation (AM) is a commonly used method for transmitting audio in radio
broadcasting, the amplitude the carrier wave is varied in proportion to that of
the information signal. In other words, a continuous radio wave has its
strength modulated by an information signal before transmission. The
information signal determines the envelope of the radio wave. Figure 4.7 is
the schematic diagram showing the mechanism of AM.
98

Figure 4.7 Schematic diagram of AM.

Carbon nanotube devices were reported to demodulate AM RF signal by
taking advantage of the nonlinear I
ds
versus V
ds
relationship. The I
ds
-V
ds
relationship can be expanded in Taylor series as follow:

where V
0
is the DC bias voltage, v is the small RF signal voltage, and the first
derivative is denoted as G
d
. The voltage across device can be express as:
99

Then the current will be:

The term reveals that device with strong Schottky behavior is capable
of demodulating AM radio wave.
Carbon nanotube devices were reported to demodulate AM RF signal by
taking advantage of the nonlinear I-V relationship [18]. We then made two
port CNT devices with different metal contacts and channel lengths to study
the nonlinear I
ds
-V
ds
relationship. Figure 4.8 presents the I
ds
versus V
ds
curves
for CNT devices different contacts and channel lengths. The black curve is
the I
ds
-V
ds
curves, and the red curve is the first derivative, from which we can
found that device with Ti/Pd metal stack shows the strongest Schottky
behavior at channel length = 1 μm. For nanotubes with diameters < ~ 2 nm, it
was observed that Schottky barriers at the Pd contacts are small but cannot be
completely eliminated as the bandgap of single-wall nanotubes is inversely
100

proportional to the diameter [19]. We also note that our long-channel devices
show Ohmic behavior while the short-channel transistors show slight
Schottky behavior. The reason is that for long-channel devices, the transport
is dominated by scattering/hopping of carriers in the channel instead of
contacts, and thus the slight Schottky behavior of contacts has become
unnoticeable in the I
ds
–V
ds
curves.
We then made demodulation circuits based on two port CNT devices
with Ti/Pd contacts (channel length = 1 μm). Figure 4.9a shows the optical
image of demodulation device. To better demonstrate the functionality of this
part, an amplitude modulated (100 Hz square wave) radio wave of 1 GHz was
sent to the input to mimic the real data transmission. Figure 4.9b shows the
measured input radio wave (yellow curve) and output information signal (blue
curve) on oscilloscope, suggesting the high-quality demodulation device was
achieved.

101

Figure 4.8 Ids-Vds relationship of two port CNT devices with different metal contacts and
channel lengths.

102

Figure 4.9 (a) Optical image of two port CNT device. (b) Measured waveforms of input
AM wave and output information signal.

Based on these excellent passive components and CNT RF transistors,
we then designed an on-chip band-pass amplifiers, providing amplification
and filtering of radio frequency signals. Figure 4.10a shows the circuit
schematic of the amplifier containing 5 active and passive components. The
circuit topologies are optimized to achieve the power gain function. There are
inductors at both the gate and drain side to provide a DC path to the gate and
drain of transistor. When the transistor is biased at Vds = -1 V and Vgs = -1.5
V, the amplifier is predicted to exhibit a power gain ~ 10.7 dB at 10 GHz
frequency with 50 Ω input and 50 Ω load.
103

Figure 4.10 (a) Circuit schematic of the designed amplifier. (b) Simulated power gain
frequency response.

4.4 Summary
In conclusion, we made CNT RF transistors with decent performance,
which is a key part of power amplifier in integrated circuits. We then studied
the metal contacts and channel length in determine the nonlinear I-V
relationship in two port CNT devices. Based on these finding, we made a
device that successfully demodulated RF radio wave and output information
signal.
104

4.5 Reference
[1] Saito, R.; Dresselhaus, G.; Dresselhaus, M. S., Physical properties of
carbon nanotubes. Imperial College Press, London, 1998.
[2] Rutherglen, C.; Jain, D.; Burke, P. Nanotube electronics for
radiofrequency applications. Nature Nanotechnology 2009, 4, 811-819.
[3] Rutherglen, C.; Kane, A. A.; Marsh, P. F.; Marsh, P. F.; Cain. T. A.;
Hassan, B. I.; Alshareef, M. R.; Zhou, C.; Galatsis, K. Wafer-scalable,
aligned carbon nanotube transistors operating at frequencies of over
100 GHz. Nat. Electron. 2019, 2, 530–539
[4] Cao, Y.; Brady, G. J.; Gui, H.; Rutherglen, C.; Arnold, M. S.; Zhou, C.
Radio Frequency Transistors Using Aligned Semiconducting Carbon
Nanotubes with Current-Gain Cutoff Frequency and Maximum
Oscillation Frequency Simultaneously Greater than 70 GHz. ACS Nano
2016, 10, 6782–6790.
[5] Brady, G. J.; Way, A. J.; Safron, N. S.; Evensen, H. T.; Gopalan, P.;
Arnold, M. S. Quasi-ballistic carbon nanotube array transistors with
current density exceeding Si and GaAs. Sci. Adv. 2016, 2.
[6] Liu, L.; Han, J.; Xu, L.; Zhou, J.; Zhao, C.; Ding, S.; Shi, H.; Xiao, M.;
Ding, L.; Ma, Z.; Jin, C.; Zhang, Z.; Peng, L. M. Aligned, high-density
105

semiconducting carbon nanotube arrays for high-performance electronics.
Science, 2020, 368, 850-856.
[7] Zhong, D.; Shi, H.; Ding, L.; Zhao, C.; Liu, J.; Zhou, J.; Zhang, Z.; Peng,
L. M. Carbon Nanotube Film-Based Radio Frequency Transistors with
Maximum Oscillation Frequency above 100 GHz. ACS Appl. Mater.
Interfaces 2019, 11, 45, 42496-42503.
[8] Yang, Y.; Ding, L.; Chen, H.; Han, J.; Zhang, Z.; Peng, L. M. Carbon
nanotube network film-based ring oscillators with sub 10-ns propagation
time and their applications in radio-frequency signal transmission. Nano
Res. 2018, 11, 1, 300–310.
[9] Rutherglen, C.; Burke, P. Carbon nanotube radio. Nano Lett. 2007, 7, 11,
3296–3299.
[10] Amlani, I.; Lewis, J.; Lee, K.; Zhang, R.; Deng, J.; Wong, H.-S. P. First
demonstration of AC gain from a single-walled carbon nanotube common-
source amplifier. Proc. Int. Electron Devices Meet. 2006, 1–4.
[11] Pliva, J.; Carta, C.; Claus, M.; Schroter, M.; Ellinger, F. On the design
of active downconversion mixers for wireless communications on a
carbon nanotube FET technology. Proc. SBMO/IEEE MTT-S Int. Microw.
Optoelectron. Conf. 2011, 984–988.
106

[12] Joo, Y.; Brady, G. J.; Arnold, M. S.; Gopalan, P. Dose-Controlled,
Floating Evaporative Self-assembly and Alignment of Semiconducting
Carbon Nanotubes from Organic Solvents. Langmuir 2014, 30, 3460-
3466.
[13] Jinkins, K. R.; Chan, J.; Jacobberger, R. M.; Berson, A.; Arnold, M. S.
Substrate-Wide Confined Shear Alignment of Carbon Nanotubes for Thin
Film Transistors. Adv. Electron. Mat. 2019, 5, 2.
[14] Che, Y. C.; Badmaev, A.; Jooyaie, A.; Wu, T.; Zhang, J. L.; Wang, C.;
Galatsis, K.; Enaya, H. A.; Zhou, C. W. Self-aligned T-gate high-purity
semiconducting carbon nanotube RF transistors operated in quasi-ballistic
transport and quantum capacitance regime. ACS Nano 2012, 6, 6936–6943.
[15] Badmaev, A.; Che, Y. C.; Li, Z.; Wang, C.; Zhou, C. W. Self-aligned
fabrication of graphene RF transistors with T-shaped gate. ACS Nano
2012, 6, 3371–3376.
[16] Che, Y.; Lin, Y. C.; Kim, P.; Zhou, C. T-Gate Aligned Nanotube Radio
Frequency Transistors and Circuits with Superior Performance. ACS Nano
2013, 7, 4343−4350.
[17] Holloway, P. H. Gold/chromium metallizations for electronic devices.
Gold Bulletin 1979,12, 99-106.
107

[18] Rutherglen, C.; Burke, P.; Carbon nanotube radio. Nano Lett. 2007, 7,
11, 3296–3299.
[19] Javey, A.; Guo, J.; Wang, Q.; Lundstrom, M.; Dai, H. J. Ballistic carbon
nanotube field-effect transistors. Nature 2003, 424, 654-657.

108

Chapter 5 Conclusions and future work
5.1 Conclusions
In this thesis, I concluded my research on the device physics of β-Ga
2
O
3

and carbon nanotubes. I explored the applications of β-Ga
2
O
3
and carbon
nanotube (CNT) in nanoelectronics, power electronics and radio frequency
(RF) electronics, and pushed the frontiers of these directions. Quasi-two-
dimensional (2D) β-Ga
2
O
3
is a rediscovered metal-oxide semiconductor with
an ultra-wide bandgap of 4.6–4.9 eV. It has been reported to be a promising
material for next-generation power and radio frequency electronics. In chapter
2, I discussed my research in forming reliable Ohmic contacts for β-Ga
2
O
3

field-effect transistors (FETs). Obtaining low contact resistance on β-Ga2O3
FETs is difficult since reactions between β-Ga2O3 and metal contacts are not
fully understood. We experimentally demonstrate that annealing β-Ga
2
O
3

FETs in argon led to the formation of oxygen vacancies at the Ti/β-Ga2O3
interface, effectively reducing the Schottky barriers and providing reliable
Ohmic contacts. Chapter 3 and chapter 4 talked about my progress in CNT
nanoelectronics and radio frequency electronics. It was revealed that for CNT
FETs at short channel regime (channel length < nanotube length), the
transistor polarity is dominated by the work function of source/drain metal
109

contacts. Atomic layer passivation method and metal contact engineering are
needed to achieve air-stable n-type transistors. I then showed the superior
electrical properties of CNT FETs in RF electronics. A suitable back-end-of-
line (BEOL) process that can economically manufacture CNT integrated
circuits are demonstrated. The decent performance of CNT demodulation unit
also paves the way in the realization of CNT integrated circuit in practical
wireless communication systems.

5.2 Future direction of nanoelectronics based on
gallium oxide and carbon nanotube
Layered materials oftern demonstrate different properties in 2D and 3D
situation [1]. Therefore, it offers a good platform for creation of
heterojunction with varied properties. The family of layered materials is
increasing dramatically, which allow a large number of combinations of
materials than traditional growth method [1]. We can start to combine some
classical 2D materials such as graphene with β-Ga
2
O
3
. Van der Waals forces
will hold these materials together to form a heterostructure. The limited
electrostatic control and lower impact ionization rate can help us achieve large
110

break down electric field strength, which are beneficial to future power
electronics [2].
As Moore’s law is nearing its end, three-dimensional (3D) integration
technique is being actively pursued. And is of particularly interest [3]. It can
achieve chip-stacking with multiple vertical layers, which increase functions
per unit area and improve the circuit performance. The current Silicon
technology is mature with available commercially design and foundry
services. Carbon nanotube-based devices can be carried out above the current
Si-based technology, which will maintain the excellent performance of both
CNT and Si electronics.
To improve the radio frequency performance of T-gate devices, other
high-k materials such as HfO
2
can be adopted as the gate dielectric. High-k
materials offer stronger gate control, smaller leakage current, and
sustainability under higher gate voltage, which will lead to better current-gain
and power-gain cut-off frequencies. And CNTs exhibit excellent mechanical
property (Young’s module over 1 TPa and tensile strength over 200 Gpa) [4],
which are suitable for flexible electronics. The as-made self-aligned T-gate
device can be transferred on the flexible substrate to characterize its
bendability performance. A wearable RF ID can be made based on flexible
CNT RF devices.
111

Neuromorphic architectures, inspired by the human brain, mimic
structure neural systems and are predicted to solve von-Neumann bottleneck
by offering time- and power- efficient benefits [5-6]. CNT is sensitive to the
charge-trapping scattering due to its small physical dimensions. Therefore, the
conductance of CNT devices will vary by the charge state, which can be
applied to emulate synaptic behavior and implement artificial neural networks
[7-8].

5.3 References

[1] Novoselov, K. S.; Mischenko, A.; Carvalho, A. Castro Neto, A. H. 2D
materials and van der Waals heterostructures. Science 2016, 353, 6298.
[2] Yan, X.; Esqueda, I. S.; Ma, J.; Tice, J.; Wang, H. High breakdown electric
field in β−Ga2O3/graphene vertical barristor heterostructure. Appl. Phys.
Lett. 2018, 112, 032101.
[3] Shulaker, M. M.; Wu, T. F.; Pal, A.; Zhao, L.; Nishi, Y.; Saraswat, K.;
Wong, H. S. P.; Mitra, S. Monolithic 3d integration of logic and memory:
Carbon nanotube fets, resistive ram, and silicon fets. Int El Devices Meet
2014.
112

[4] Treacy, M. M. J.; Ebbesen, T. W.; Gibson, J. M. Exceptionally High
Young's Modulus Observed for Individual Carbon Nanotubes. Nature
1996, 381, 678-680.
[5] Indiveri, B. G.; Liu, S. Memory and Information Processing in
Neuromorphic Systems. Proc. IEEE 2015, 103, 1379−1397.
[6] Wei, H.; Patil, N.; Lin, A.; Wong, H. S. P.; Mitra, S. Monolithic Three-
Dimensional Integrated Circuits Using Carbon Nanotube FETs and
Interconnects. Technol. Dig. - Int. Electron Devices Meet. IEDM 2009,
577−580.
[7] Kim, K.; Chen, C.; Truong, Q.; Shen, A. M.; Chen, Y. A Carbon Nanotube
Synapse with Dynamic Logic and Learning. Adv. Mater. 2013, 25,
1693−1698.
[8] Bushmaker, A. W.; Oklejas, V.; Walker, D.; Hopkins, A. R.; Chen, J.;
Cronin, S. B. Single-Ion Adsorption and Switching in Carbon Nanotubes.
Nat. Commun. 2016, 7, 1−8.

113

Bibliography

[1] Roy, R.; Hill, V. G.; Osborn, E. F. Polymorphism of Ga
2
O
3
and the System
Ga
2
O
3
-H
2
O. J. Am. Chem. Soc. 1952, 74, 3, 719-722.
[2] Kroll, P; Dronskowski, R.; Matrin, M. Formation of spinel-type gallium
oxynitrides: a density-functional study of binary and ternary phases in the
system Ga–O–N. J. Mater. Chem. 2005, 15, 3296-3302.
[3] Ahman, J.; Svensson, G.; Albertsson, J. A Reinvestigation of β-Ga
2
O
3
.
Acta Cryst. 1996, C52, 1336-1338.
[4] Higashiwaki, M.; Sasaki, K.; Murakami, H.; Kumagai, Y.; Koukitu, A.;
Kuramata, A.; Masui, T.; Tamakoshi, S. Recent progress in Ga
2
O
3
power
devices. Semocond. Sci. Technol. 2016, 31 034001.
[5] Ma, N.; Tanen, N.; Verma, A.; Guo, Z.; Luo, T. F.; Xing, H. L.; Jena, D.
Intrinsic electron mobility limits in β-Ga
2
O
3
. Appl. Phys. Lett. 2016, 109,
212101.
[6] Higashiwaki, M.; Sasaki, K.; Kuramata, A.; Masui, T.; Yamakoshi, S.
Gallium oxide (Ga
2
O
3
) metal-semiconductor field-effect transistors on
signle-crystal β-Ga
2
O
3
(010) substrates. Appl. Phys. Lett. 2012, 100,
013504
114

[7] Kuramata, A.; Koshi, K.; Watanabe, S.; Yamaoka, Y.; Masui, T.;
Yamakoshi, S. High-quality β-Ga
2
O
3
single crystals grown by edge-
defined film-fed growth. Jpn. J. Appl. Phys. 2016, 55 1202A2.
[8] Zhang, J.; Li, B.; Xia, C.; Pei, G.; Deng, Q.; Yang, Z.; Xu, W.; Shi, H.;
Wu, F.; Wu, Y.; Xu, J. Growth and spectral characterization of β-Ga
2
O
3

single crystals. J. Phys. Chem. Solids 2006, 67, 12, 2448-2451.
[9] Galazka, Z.; Irmscher, K.; Uecker, R.; Bertam, R.; Pietsch, M.;
Kwasniewski, A.; Naumann, M.; Schulz, T.; Schewski, R.; Klimm, D.;
Bickermann, M. On the bulk β-Ga
2
O
3
single crystals grown by the
Czochralski method. J. Cryst. Growth 2014, 404, 15, 184-191.
[10] Ví llora, E. G.; Shimamura, K.; Kitamura, K.; Aoki, K. Rf-plasma-
assisted molecular-beam-epitaxy of β-Ga
2
O
3
. Appl. Phys. Lett. 2006, 88,
031105.
[11] Kim, H. W.; Kim, N. H. Synthesis of β-Ga
2
O
3
nanowires by an
MOCVD approach. Appl. Phys. A 2005, 81, 763–765.
[12] Higashiwaki, M.; Sasaki, K.; Kamimura, T.; Hoi Wong, M.;
Krishnamurthy, D.; Kuramata, A.; Masui, T.; Yamakoshi, S. Depletion-
mode Ga
2
O
3
metaloxide-semiconductor field-effect transistors on β-
Ga
2
O
3
(010) substrates and temperature dependence of their device
characteristics. Appl. Phys. Lett. 2013, 103, 123511.
115

[13] Zhou, H.; Maize, K.; Noh, J.; Shakouri, A.; Ye, P. D. Thermodynamic
studies of β-Ga
2
O
3
nanomembrane field-effect transistors on a sapphire
substrate. ACS Omega 2017, 2, 7723–7729
[14] Green, A. J.; Chabak, K. D.; Baldini, M.; Moser, N.; Gilbert, R.; Fitch,
R. C.; Wagner, G.; Galazka, Z.; McCandless, J.; Crespo, A. et al. β-Ga
2
O
3

MOSFETs for radio frequency operation. IEEE Electron Device Lett.
2017, 38, 790–793.
[15] Oshima, T.; Okuno, T.; Fujita, S. Ga
2
O
3
Thin Film Growth on c-Plane
Sapphire Substrates by Molecular Beam Epitaxy for Deep-Ultraviolet
Photodetectors. Jpn. J. Appl. Phys. 2007, 46 7217.
[16] Kim, C. J.; Kang, D.; Song, I.; Park, J. C.; Lim, H.; Kim, S.; Lee, E.;
Chung, R.; Lee, J. C.; Park, Y. International Electron Devices Meeting
2006, 11-13.
[17] Iijima, S. Helical microtubules of graphitic carbon. Nature 1991, 354,
56-58
[18] Liu, B.; Wu, F.; Gui, H.; Zheng, M.; Zhou, C. Chirality-Controlled
Synthesis and Applications of Single-Wall Carbon Nanotubes. Acs Nano
2017, 11, 1, 31-53.
116

[19] Kataura, H.; Kumazawa, Y.; Maniwa, Y.; Umezu, I.; Suzuki, S.;
Ohtsuka, Y.; Achiba, Y. Optical properties of single-wall carbon
nanotubes. Synthetic Met. 1999, 103, 2555-2558.
[20] Saito, R.; Dresselhaus, G.; Dresselhaus, M. S., Physical properties of
carbon nanotubes. Imperial College Press, London, 1998.
[21] Durkop, T.; Getty, S. A.; Cobas, E.; Fuhrer, M. S. Extraordinary
Mobility in Semiconducting Carbon Nanotubes. Nano Lett. 2004, 4, 35-
39.
[22] Zhou, X. J.; Park, J. Y.; Huang, S. M.; Liu, J.; McEuen, P. L. Band
Structure, Phonon Scattering, and the Performance Limit of Single-
Walled Carbon Nanotube Transistors. Phys. Rev. Lett. 2005, 95.
[23] Pop, E.; Mann, D. A.; Goodson, K. E.; Dai, H. J. Electrical and thermal
transport in metallic single-wall carbon nanotubes on insulating substrates.
J. Appl. Phys. 2007, 101.
[24] Subramaniam, C.; Yamada, T.; Kobashi, K.; Sekiguchi, A.; Futaba, D.
N.; Yumura, M.; Hata, K. One hundred fold increase in current carrying
capacity in a carbon nanotubecopper composite. Nat. Commun. 2013, 4.
[25] Treacy, M. M. J.; Ebbesen, T. W.; Gibson, J. M. Exceptionally High
Young's Modulus Observed for Individual Carbon Nanotubes. Nature
1996, 381, 678-680.
117

[26] George, S.; Franklin, A. D.; Frank, D.; Lobez, J. M.; Cao, Q.; Park, H.,
Afzali, A.; Han, S. J.; Hannon, J. B.; Haensch, W. Toward high-
performance digital logic technology with carbon nanotubes. ACS Nano
2014, 8, 8730–8745.
[27] Zhang, J.; Wang, C.; Fu, Y.; Che, Y.; Zhou, C. Air-stable conversion
of separated carbon nanotube thin-film transistors from p-type to n-type
using atomic layer deposition of high-κ oxide and its application in CMOS
logic circuits. ACS Nano 2011, 5, 3284–3292.
[28] Shulaker, M.; Hills, G.; Patil, N.; Wei, H.; Chen, H. Y.; Wong, H. S. P.;
Mitra, S. Carbon nanotube computer. Nature 2013, 501, 526–530.
[29] Hills, G.; Lau, C.; Wright, A.; Fuller, S.; Bishop, M. D.; Srimani, T.;
Kanhaiya, P.; Ho, R.; Amer, A.; Stein, Y.; Murphy, D.; Arvind,
Chandrakasan, A.; Shulaker, M. Modern microprocessor built from
complementary carbon nanotube transistors. Nature 2019, 572, 595-602.
[30] Kienle, D.; Leonard, F. Terahertz Response of Carbon Nanotube
Transistors. Phys. Rev. Lett 2009, 103, 206601.
[31] Rutherglen, C.; Kane, A. A.; Marsh, P. F.; Marsh, P. F.; Cain. T. A.;
Hassan, B. I.; Alshareef, M. R.; Zhou, C.; Galatsis, K. Wafer-scalable,
aligned carbon nanotube transistors operating at frequencies of over
100 GHz. Nat. Electron. 2019, 2, 530–539.
118

[32] Novoselov, K. S.; Geim, A. K.; Morozov, S. V .; Jiang, D.; Zhang, Y .;
Dubonos, S. V .; Grigorieva, I. V .; Firsov, A. A., Electric Field Effect in
Atomically Thin Carbon Films. Science 2004, 306 (5696), 666.
[33] Wang, Q. H.; Kalantar-Zadeh, K.; Kis, A.; Coleman, J. N.; Strano, M.
S., Electronics and optoelectronics of two-dimensional transition metal
dichalcogenides. Nat. Nanotechnol. 2012, 7, 699.
[34] Carey, B. J.; Ou, J. Z.; Clark, R. M.; Berean, K. J.; Zavabeti, A.;
Chesman, A. S. R.; Russo, S. P.; Lau, D. W. M.; Xu, Z.-Q.; Bao, Q. et al.
Wafer-scale two-dimensional semiconductors from printed oxide skin of
liquid metals. Nat. Commun. 2017, 8, 14482.
[35] Liu, B.; Abbas, A.; Zhou, C., Two‐Dimensional Semiconductors: From
Materials Preparation to Electronic Applications. Adv. Electron. Mater.
2017, 3 (7), 1700045.
[36] Kalantar-zadeh, K.; Ou, J. Z.; Daeneke, T.; Mitchell, A.; Sasaki, T.;
Fuhrer, M. S., Two dimensional and layered transition metal oxides. Appl.
Mater. Today 2016, 5, 73-89.
[37] Hwang, W. S.; Verma, A.; Peelaers, H.; Protasenko, V .; Rouvimov, S.;
Xing, H.; Seabaugh, A.; Haensch, W.; de Walle, C. V .; Galazka, Z.;
Albrecht, M.; Fornari, R.; Jena, D., High-voltage field effect transistors
119

with wide-bandgap β-Ga2O3 nanomembranes. Appl. Phys. Lett. 2014,
104 (20), 203111.
[38] Higashiwaki, M.; Sasaki, K.; Kamimura, T.; Hoi Wong, M.;
Krishnamurthy, D.; Kuramata, A.; Masui, T.; Yamakoshi, S., Depletion-
mode Ga
2
O
3
metal-oxide-semiconductor field-effect transistors on β-
Ga
2
O
3
(010) substrates and temperature dependence of their device
characteristics. Appl. Phys. Lett. 2013, 103 (12), 123511.
[39] Irmscher, K.; Galazka, Z.; Pietsch, M.; Uecker, R.; Fornari, R.,
Electrical properties of β-Ga
2
O
3
single crystals grown by the Czochralski
method. J. Appl. Phys. 2011, 110 (6), 063720.
[40] Zhou, H.; Si, M.; Alghamdi, S.; Qiu, G.; Yang, L.; Ye, P. D., High-
Performance Depletion/Enhancement-ode β-Ga
2
O
3
on Insulator (GOOI)
Field-Effect Transistors with Record Drain Currents of 600/450 mA/mm.
IEEE Electron Device Lett. 2017, 38 (1), 103-106.
[41] Wu, F.; Chen, L.; Zhang, A.; Hong, Y .-L.; Shih, N.-Y .; Cho, S.-Y .;
Drake, G. A.; Fleetham, T.; Cong, S.; Cao, X. et al. High-Performance
Sub-Micrometer Channel WSe
2
Field-Effect Transistors Prepared Using a
Flood–Dike Printing Method. ACS Nano 2017, 11 (12), 12536-12546.
[42] Galazka, Z.; Uecker, R.; Irmscher, K.; Albrecht, M.; Klimm, D.; Pietsch,
120

M.; Brützam, M.; Bertram, R.; Ganschow, S.; Fornari, R., Czochralski
growth and characterization of β‐Ga
2
O
3
single crystals. Cryst. Res.
Technol. 2010, 45 (12), 1229-1236.
[43] Lovejoy, T. C.; Yitamben, E. N.; Shamir, N.; Morales, J.; Villora, E. G.;
Shimamura, K.; Zheng, S.; Ohuchi, F. S.; Olmstead, M. A., Surface
morphology and electronic structure of bulk single crystal β-Ga
2
O
3
(100).
Appl. Phys. Lett. 2009, 94 (8), 081906.
[44] Ramana, C. V .; Rubio, E. J.; Barraza, C. D.; Miranda Gallardo, A.;
McPeak, S.; Kotru, S.; Grant, J. T., Chemical bonding, optical constants,
and electrical resistivity of sputter-deposited gallium oxide thin films. J.
Appl. Phys. 2014, 115 (4), 043508.
[45] Green, A. J.; Chabak, K. D.; Heller, E. R.; Fitch, R. C.; Baldini, M.;
Fiedler, A.; Irmscher, K.; Wagner, G.; Galazka, Z.; Tetlak, S. E. et al. 3.8-
MV/cm Breakdown Strength of MOVPE-Grown Sn-Doped β-Ga
2
O
3

MOSFETs. IEEE Electron Device Lett. 2016, 37 (7), 902-905.
[46] Wong, M. H.; Sasaki, K.; Kuramata, A.; Yamakoshi, S.; Higashiwaki,
M., Field-Plated Ga
2
O
3
MOSFETs With a Breakdown V oltage of Over 750
V . IEEE Electron Device Lett. 2016, 37 (2), 212-215.
[47] Chang, P.-C.; Fan, Z.; Tseng, W.-Y .; Rajagopal, A.; Lu, J. G., β-Ga
2
O
3

121

nanowires: Synthesis, characterization, and p-channel field-effect
transistor. Appl. Phys. Lett. 2005, 87 (22), 222102.
[48] Farzana, E.; Zhang, Z.; Paul, P. K.; Arehart, A. R.; Ringel, S. A.,
Influence of metal choice on (010) β-Ga
2
O
3
Schottky diode properties.
Appl. Phys. Lett. 2017, 110 (20), 202102.
[49] Yao, Y .; Davis, R. F.; Porter, L. M., Investigation of Different Metals as
Ohmic Contacts to β-Ga
2
O
3
: Comparison and Analysis of Electrical
Behavior, Morphology, and Other Physical Properties. J. Electron. Mater.
2017, 46 (4), 2053-2060.
[50] Ma, Y .; Shen, C.; Zhang, A.; Chen, L.; Liu, Y .; Chen, J.; Liu, Q.; Li, Z.;
Amer, M. R.; Nilges, T.; Abbas, A. N.; Zhou, C., Black Phosphorus Field-
Effect Transistors with Work Function Tunable Contacts. ACS Nano 2017,
11 (7), 7126-7133.
[51] Mann, D.; Javey, A.; Kong, J.; Wang, Q.; Dai, H., Ballistic Transport in
Metallic Nanotubes with Reliable Pd Ohmic Contacts. Nano Lett. 2003, 3
(11), 1541-1544.
[52] Ma, Y .; Liu, B.; Zhang, A.; Chen, L.; Fathi, M.; Shen, C.; Abbas, A. N.;
Ge, M.; Mecklenburg, M.; Zhou, C., Reversible Semiconducting-to-
Metallic Phase Transition in Chemical Vapor Deposition Grown
122

Monolayer WSe
2
and Applications for Devices. ACS Nano 2015, 9 (7),
7383-7391.
[53] English, C. D.; Shine, G.; Dorgan, V . E.; Saraswat, K. C.; Pop, E.,
Improved Contacts to MoS
2
Transistors by Ultra-High Vacuum Metal
Deposition. Nano Lett. 2016, 16 (6), 3824-3830.
[54] Brillson, L. J., Chemical reaction and charge redistribution at metal–
semiconductor interfaces. J. Vac. Sci. Technol. 1978, 15 (4), 1378-1383.
[55] Guo, Y .; Zhou, J.; Liu, Y .; Zhou, X.; Yao, F.; Tan, C.; Wu, J.; Lin, L.;
Liu, K.; Liu, Z.; Peng, H., Chemical Intercalation of Topological Insulator
Grid Nanostructures for High‐Performance Transparent Electrodes. Adv.
Mater. 2017, 29 (44), 1703424.
[56] Mosbacker, H. L.; Strzhemechny, Y . M.; White, B. D.; Smith, P. E.;
Look, D. C.; Reynolds, D. C.; Litton, C. W.; Brillson, L. J., Role of near-
surface states in ohmic-Schottky conversion of Au contacts to ZnO. Appl.
Phys. Lett. 2005, 87 (1), 012102.
[57] Sawa, A., Resistive switching in transition metal oxides. Mater. Today
2008, 11 (6), 28-36.
[58] Fortunato, E.; Barquinha, P.; Martins, R., Oxide Semiconductor Thin‐
Film Transistors: A Review of Recent Advances. Adv. Mater. 2012, 24 (22),
123

2945-2986.
[59] Waser, R.; Dittmann, R.; Staikov, G.; Szot, K., Redox‐Based Resistive
Switching Memories – Nanoionic Mechanisms, Prospects, and
Challenges. Adv. Mater. 2009, 21 (25‐26), 2632-2663.
[60] Yang, J. J.; Strachan John, P.; Xia, Q.; Ohlberg Douglas, A. A.; Kuekes
Philip, J.; Kelley Ronald, D.; Stickle William, F.; Stewart Duncan, R.;
Medeiros‐Ribeiro, G.; Williams, R. S., Diffusion of Adhesion Layer
Metals Controls Nanoscale Memristive Switching. Adv. Mater. 2010, 22
(36), 4034-4038.
[61] Yang, J. J.; Pickett, M. D.; Li, X.; Ohlberg, D. A. A.; Stewart, D. R.;
Williams, R. S., Memristive switching mechanism for metal/oxide/metal
nanodevices. Nat. Nanotechnol. 2008, 3, 429.
[62] Gao, X.; Xia, Y .; Ji, J.; Xu, H.; Su, Y .; Li, H.; Yang, C.; Guo, H.; Yin,
J.; Liu, Z., Effect of top electrode materials on bipolar resistive switching
behavior of gallium oxide films. Appl. Phys. Lett. 2010, 97 (19), 193501.
[63] Javey, A.; Guo, J.; Wang, Q.; Lundstrom, M.; Dai, H., Ballistic carbon
nanotube field-effect transistors. Nature 2003, 424, 654.
[64] Liu, B.; Ma, Y .; Zhang, A.; Chen, L.; Abbas, A. N.; Liu, Y .; Shen, C.;
Wan, H.; Zhou, C., High-Performance WSe2 Field-Effect Transistors via
124

Controlled Formation of In-Plane Heterojunctions. ACS Nano 2016, 10
(5), 5153-5160.
[65] Mak, K. F.; He, K.; Lee, C.; Lee, G. H.; Hone, J.; Heinz, T. F.; Shan, J.,
Tightly bound trions in monolayer MoS
2
. Nat. Mater. 2012, 12, 207.
[66] Wood, J. D.; Wells, S. A.; Jariwala, D.; Chen, K.-S.; Cho, E.; Sangwan,
V . K.; Liu, X.; Lauhon, L. J.; Marks, T. J.; Hersam, M. C., Effective
Passivation of Exfoliated Black Phosphorus Transistors against Ambient
Degradation. Nano Lett. 2014, 14 (12), 6964-6970.
[67] Desai, S. B.; Seol, G.; Kang, J. S.; Fang, H.; Battaglia, C.; Kapadia, R.;
Ager, J. W.; Guo, J.; Javey, A., Strain-Induced Indirect to Direct Bandgap
Transition in Multilayer WSe
2
. Nano Lett. 2014, 14 (8), 4592-4597.
[68] Liu, Y .; Guo, J.; Zhu, E.; Liao, L.; Lee, S.-J.; Ding, M.; Shakir, I.;
Gambin, V .; Huang, Y .; Duan, X., Approaching the Schottky–Mott limit
in van der Waals metal–semiconductor junctions. Nature 2 0 1 8 , 557 (7707),
696-700.
[69] Waldner, P.; Eriksson, G., Thermodynamic modelling of the system
titanium-oxygen. Calphad 1999, 23 (2), 189-218.
[70] Yang, J. J.; Strachan, J. P.; Miao, F.; Zhang, M.-X.; Pickett, M. D.; Yi,
W.; Ohlberg, D. A. A.; Medeiros-Ribeiro, G.; Williams, R. S., Metal/TiO
2

125

interfaces for memristive switches. Appl. Phys. A 2011, 102 (4), 785-789.
[71] Lei, B.; Li, C.; Zhang, D.; Tang, T.; Zhou, C., Tuning electronic
properties of In2O3 nanowires by doping control. Appl. Phys. A 2004, 79
(3), 439-442.
[72] Liu, Q.; Liu, Y .; Wu, F.; Cao, X.; Li, Z.; Alharbi, M.; Abbas, A. N.;
Amer, M. R.; Zhou, C., Highly Sensitive and Wearable In
2
O
3
Nanoribbon
Transistor Biosensors with Integrated On-Chip Gate for Glucose
Monitoring in Body Fluids. ACS Nano 2018, 12 (2), 1170-1178.
[73] Hollinger, G.; Skheyta-Kabbani, R.; Gendry, M., Oxides on GaAs and
InAs surfaces: An x-ray-photoelectron-spectroscopy study of reference
compounds and thin oxide layers. Phys. Rev. B 1994, 49 (16), 11159-
11167.
[74] Dong, L.; Jia, R.; Xin, B.; Peng, B.; Zhang, Y ., Effects of oxygen
vacancies on the structural and optical properties of β-Ga
2
O
3
. Sci. Rep.
2017, 7, 40160.
[75] Varley, J. B.; Weber, J. R.; Janotti, A.; Van de Walle, C. G., Oxygen
vacancies and donor impurities in β-Ga
2
O
3
. Appl. Phys. Lett. 2010, 97 (14),
142106.
[76] Carey, P. H.; Yang, J.; Ren, F.; Hays, D. C.; Pearton, S. J.; Jang, S.;
126

Kuramata, A.; Kravchenko, I. I., Ohmic contacts on n-type β-Ga
2
O
3
using
AZO/Ti/Au. AIP Adv. 2017, 7 (9), 095313.
[77] Carey, P. H.; Yang, J.; Ren, F.; Hays, D. C.; Pearton, S. J.; Kuramata,
A.; Kravchenko, I. I., Improvement of Ohmic contacts on Ga
2
O
3
through
use of ITO-interlayers. J. Vac. Sci. Technol. B 2017, 35 (6), 061201.
[78] Qian, Y . P.; Guo, D. Y .; Chu, X. L.; Shi, H. Z.; Zhu, W. K.; Wang, K.;
Huang, X. K.; Wang, H. et al. Mg-doped p-type β-Ga
2
O
3
thin film for
solar-blind ultraviolet photodetector. Mater. Lett. 2017, 209, 558-561.
[79] Strachan John, P.; Pickett Matthew, D.; Yang, J. J.; Aloni, S.; David
Kilcoyne, A. L.; Medeiros‐Ribeiro, G.; Stanley Williams, R., Direct
Identification of the Conducting Channels in a Functioning Memristive
Device. Adv. Mater. 2010, 22 (32), 3573-3577.
[80] Zhao, Y .; Frost Ray, L., Raman spectroscopy and characterisation of α‐
gallium oxyhydroxide and β‐gallium oxide nanorods. J. Raman Spectrosc.
2008, 39 (10), 1494-1501.
[81] Bourque, J. L.; Biesinger, M. C.; Baines, K. M., Chemical state
determination of molecular gallium compounds using XPS. Dalton Trans.
2016, 45 (18), 7678-7696.
[82] NIST X-ray Photoelectron Spectroscopy Database, NIST Standard
127

Reference Database Number 20, National Institute of Standards and
Technology, Gaithersburg MD, 20899 (2000)
[83] Tans, S.; Verschueren, A.; Dekker, C. Room-temperature transistor
based on a single carbon nanotube. Nature 1998, 393, 49-52.
[84] Sazonova, V.; Yaish, Y.; Üstü nel, H.; Roundy, D.; Arias, T. A.;
McEuen, P. L. A tunnable carbon nanotube electromechanical oscillator.
Nature 2004, 431, 284-287.
[85] Avouris, P.; Chen, Z.; Perebeinos, V. Carbon-based electronics. Nat.
Nanotech. 2007, 2, 605-456.
[86] De Volder, M. F. L.; Tawfick, S. H.; Baughman, R. H.; Hart, A. J.
Carbon Nanotubes: present and future commercial applications. Science
2013, 339, 535–539.
[87] Zhou, C.; Kong, J.; Yenilmez, E.; Dai, H. Modulated Chemical Doping
of Individual Carbon Nanotubes. Science 2000, 290, 1552-1555.
[88] Kang, S.; Kocabas, C.; Ozel, T.; Shim, M.; Pimparkar, N.; Alam, M.
A.; Roktin, S. V.; Rogers, J. A. High-performance electronics using dense,
perfectly aligned arrays of single-walled carbon nanotubes. Nat. Nanotech.
2007, 2, 230–236.
[89] Geier, M.; McMorrow, J.; Xu, W.; Zhu, J.; Kim, C. H.; Marks, T. J.;
Hersam, M.C. Solution-processed carbon nanotube thin-film
128

complementary static random access memory. Nat. Nanotech. 2015, 10,
944–948.
[90] Qiu, C. G.; Zhang, Z. Y.; Xiao, M. M.; Yang, Y. J.; Zhong, D. L.; Peng,
L. M. Scaling carbon nanotube complementary transistors to 5-nm gate
lengths. Science 2017, 355, 271-276.
[91] Han, S. J.; Tang, J.; Kumar, B.; Falk, A.; Farmer, D.; Tulevski, G.;
Jenkins, K.; Afzali, A.; Oida, S.; Ott, J.; Hannon, J.; Haensch, W. High-
speed logic integrated circuits with solution-processed self-assembled
carbon nanotubes. Nat. Nanotech. 2017, 12, 861–865.
[92] Yang, F.; Wang, X.; Zhang, D.; Yang, J.; Luo, D.; Xu, Z.; Wei, J.;
Wang, J. Q.; Xu, Z.; Peng, F.; Li, X.; Li, R.; Li, Y.; Li, M.; Bai, X.; Ding,
F.; Li, Y. Chirality-specific growth of single-walled carbon nanotubes on
solid alloy catalysts. Nature 2014, 510, 522-524.
[93] Zhang, S. C.; Kang, L. X.; Wang, X.; Tong, L. M.; Yang, L. W.; Wang,
Z. Q.; Qi, K.; Deng, S. B.; Li, Q. W.; Bai, X. D.; Ding, F.; Zhang, J. Arrays
of Horizontal Carbon Nanotubes of Controlled Chirality Grown Using
Designed Catalysts. Nature 2017, 543, 234-238.
[94] Lipomi, D.; Vosgueritchian, M.; Tee, BK.; Hellstrom, S. L.; Lee, J. A.;
Fox, C. H.; Bao, Z. Skin-like pressure and strain sensors based on
129

transparent elastic films of carbon nanotubes. Nat. Nanotech. 2011, 6,
788–792.
[95] Sun, Dm.; Timmermans, M.; Tian, Y.; Nasibulin, A. G.; Kauppinen, E.
I.; Kishimoto, S.; Mizutani, T.; Ohno, Y. Flexible high-performance
carbon nanotube integrated circuits. Nat. Nanotech. 2011, 6, 156–161.
[96] Park, S.; Vosguerichian, M. and Bao, Z. A review of fabrication and
applications of carbon nanotube film-based flexible electronics.
Nanoscale 2013, 5, 1727-1752.
[97] Peng, L. M.; Zhang, Z.; Qiu, C. Carbon nanotube digital electronics.
Nat. Electron. 2019, 2, 499-505.
[98] Xiang, L.; Zhang, H.; Dong, G.; Zhong, D.; Han, J.; Liang, X.; Zhang,
Z.; Peng, L. M.; Hu, Y. Low-power carbon nanotube-based integrated
circuits that can be transferred to biological surfaces. Nat. Electron. 2018,
1, 237–245.
[99] Franklin, A. D.; Chen, Z. H. Length scaling of carbon nanotube
transistors. Nat. Nanotechnol. 2010, 5, 858-862.
[100] Cao, Q.; Tersoff, J.; Farmer, D. B.; Zhu, Y.; Han, S. J. Carbon nanotube
transistors scaled to a 40-nanometer footprint. Science 2017, 356, 1369-
1372.
130

[101] Tang, J.; Cao, Q.; Tulevski, G.; Jenkins, K.; Nela, L.; Farmer, D. B.;
Han, S. J. Flexible CMOS integrated circuits based on carbon nanotubes
with sub-10 ns stage delays. Nat. Electron. 2018, 1, 191–196.
[102] Joo, Y.; Brady, G. J.; Arnold, M. S.; Gopalan, P. Dose-Controlled,
Floating Evaporative Self-assembly and Alignment of Semiconducting
Carbon Nanotubes from Organic Solvents. Langmuir 2014, 30, 3460-
3466.
[103] Brady, G. J.; Way, A. J.; Safron, N. S.; Evensen, H. T.; Gopalan, P.;
Arnold, M. S. Quasi-ballistic carbon nanotube array transistors with
current density exceeding Si and GaAs. Sci. Adv. 2016, 2.
[104] Liu, L.; Han, J.; Xu, L.; Zhou, J.; Zhao, C.; Ding, S.; Shi, H.; Xiao, M.;
Ding, L.; Ma, Z.; Jin, C.; Zhang, Z.; Peng, L. M. Aligned, high-density
semiconducting carbon nanotube arrays for high-performance electronics.
Science, 2020, 368, 850-856.
[105] Cao, Y.; Brady, G. J.; Gui, H.; Rutherglen, C.; Arnold, M. S.; Zhou, C.
Radio Frequency Transistors Using Aligned Semiconducting Carbon
Nanotubes with Current-Gain Cutoff Frequency and Maximum
Oscillation Frequency Simultaneously Greater than 70 GHz. ACS Nano
2016, 10, 6782–6790.
131

[106] Zhong, D.; Shi, H.; Ding, L.; Zhao, C.; Liu, J.; Zhou, J.; Zhang, Z.;
Peng, L. M. Carbon Nanotube Film-Based Radio Frequency Transistors
with Maximum Oscillation Frequency above 100 GHz. ACS Appl. Mater.
Interfaces 2019, 11, 42496-42503.
[107] Wei, H.; Chen, H. Y.; Liyanage, L.; Wong, H. S. P.; Mitra, S. Air-stable
technique for fabricating n-type carbon nanotube FETs. IEDM Tech. Dig.
2011, 23.2.1–23.2.4.
[108] Li, G.; Li, Q.; Jin, Y.; Zhao, Y.; Xiao, X.; Jiang, K.; Wang, J.; Fan, S.
Fabrication of air-stable n-type carbon thin-film transistors on flexible
substrates using bilayer dielectrics. Nanoscale 2015, 7, 17693-17701.
[109] Tang, J.; Farmer, D.; Bangsaruntip, S.; Chiu, K. C.; Kumar, B.; Han, S.
J. Contact engineering and channel doping for robust carbon nanotube
NFETs. 2017 Int. Symp. VLSI Tech. Syst. Appl.
[110] Yang, Y.; Ding, L.; Han, J.; Zhang, Z; Peng, L.M.; High-performance
complementary transistors and medium-scale integrated circuits based on
carbon nanotube thin films. ACS Nano 2017, 11, 4124–4132.
[111] Brady, G. J.; Joo, Y.; Roy, S. S.; Gopalan, P.; Arnold, M. S. High
performance transistors via aligned polyfluorene-sorted carbon nanotubes.
Appl. Phys. Lett. 2014, 104, 083107.
132

[112] Kang, D.; Park, N.; Ko, J. H.; Bae, E.; Park, W. Oxygen-induced p-type
doping of a long individual single-walled carbon nanotube.
Nanotechnology 2005, 16, 1048-1052.
[113] Avery, A.; Zhou, B.; Lee, J.; Lee, E. S.; Miller, E. M.; Ihly, R.;
Wesenberg, D.; Mistry, K. S.; Guillot, S. L.; Zink, B. L.; Kim, Y. H.;
Blackburn, J. L.; Ferguson, A. J. Tailored semiconducting carbon
nanotube networks with enhanced thermoelectric properties. Nat. Energy
2016, 1, 16033.
[114] Park, R.S.; Shulaker, M. M.; Hills, G.; Liyanage, L. S.; Lee, S.; Tang,
A.; Mitra, S.; PhilipWong, H. S. Hysteresis in Carbon Nanotube
Transistors: Measurement and Analysis of Trap Density, Energy Level,
and Spatial Distribution. ACS Nano 2016, 10, 4599-4608.
[115] Jin, S. H.; Islam, A. E.; Kim, T.; Kim, J.; Alam, M. A.; Rogers, J. A.
Sources of Hysteresis in Carbon Nanotube Field-Effect Transistors and
Their Elimination Via Methylsiloxane Encapsulants and Optimized
Growth Procedures. Adv. Funct. Mater. 2012, 22, 2276-2284.
[116] Franklin, A. D.; Farmer, D. B.; Haensch, W. Defining and Overcoming
the Contact Resistance Challenge in Scaled Carbon Nanotube Transistors.
Acs Nano 2014, 8, 7333-7339.
133

[117] Che, Y. C.; Wang, C.; Liu, J.; Liu, B. L.; Lin, X.; Parker, J.; Beasley,
C.; Wong, H. S. P.; Zhou, C. W. Selective Synthesis and Device
Applications of Semiconducting Single-Walled Carbon Nanotubes Using
Isopropyl Alcohol as Feedstock. ACS Nano 2012, 6, 7454-7462.
[118] Baranovskii, S. D.; Nenashev, A. V.; Oelerich, J. O.; Greiner, S. H. M.;
Dvurechenskii, A. V.; Gerbhard, F. Percolation description of charge
transport in the random barrier model applied to atmorphous oxide
semiconductors. Europhys. Lett. 2019, 127-57004.
[119] Zavabeti, A.; Zhang, B. Y.; Castro, I. A.; Ou, J. Z.; Carey, B. J.;
Mohiuddin, Md.; Datta, R. S.; Xu, C.; Mouritz, A. P.; McConville, C. F.;
O’Mullane, A. P.; Daeneke, T.; Kalantar-Zadeh, K. GREEN Synthesis of
Low-Dimensional Aluminum Oxide Hydroxide and Oxide Using Liquid
Metal Reaction Media: Ultrahigh Flux Membranes. Adv. Funct. Marer.
2018, 1804057.
[120] Jinkins, K. R.; Chan, J.; Jacobberger, R. M.; Berson, A; Arnold, M. S.
Substrate-Wide Confined Shear Alignment of Carbon Nanotubes for Thin
Film Transistors. Adv. Electron. Mater. 2019, 5, 1800593.
[121] Moriyama, N.; Ohno, Y.; Kitamura, T.; Kishimoto, S.; Mizutani, T.
Change in Carrier Type in High-k Gate Carbon Nanotube Field-Effect
134

Transistors by Interface Fixed Charges. Nanotechnology 2010, 21,
165201.
[122] Rinkiol, M.; Johansson, A.; Zavodchikova, M.; Toppari, J.; Albert, G.;
Nasibulin, A.; Kauppinen, E.; Torma, P. High Yield of Memory Elements
from Carbon Nanotube Field- Effect Transistors with Atomic Layer
Deposited Gate Dielectric. New J. Phys. 2008, 10, 103019.
[123] Mudimela, P. R.; Grigoras, K.; Anoshkin, I. V.; Varpula, A.; Ermolov,
V.; Anisimov, A. S.; Nasibulin, A. G.; Novikov, S.; Kauppinen, E. I.
Single-Walled Carbon Nanotube Network Field Effect Transistor as a
Humidity Sensor. J Sensors 2012.
[124] Y. Cui, R. Xin, Z. Yu, Y. Pan, Z.-Y. Ong, X. Wei, J. Wang, H. Nan, Z.
Ni, Y. Wu, T. Chen, Y. Shi, B. Wang, G. Zhang, Y.-W. Zhang, X. Wang.
High-Performance Monolayer WS
2
Field-Effect Transistors on High-κ
Dielectrics. Adv. Mater. 2015, 27, 5230.
[125] S. Kim, A. Konar, W.-S. Hwang, J. H. Lee, J. Lee, J. Yang, C. Jung, H.
Kim, J.-B. Yoo, J.-Y. Choi, Y. W. Jin, S. Y. Lee, D. Jena, W. Choi, K.
Kim. High-mobility and low-power thin-film transistors based on
multilayer MoS
2
crystals. Nat Commun. 2012, 3, 1011.
[126] Rutherglen, C.; Jain, D.; Burke, P. Nanotube electronics for
radiofrequency applications. Nature Nanotechnology 2009, 4, 811-819.
135

[127] Yang, Y.; Ding, L.; Chen, H.; Han, J.; Zhang, Z.; Peng, L. M. Carbon
nanotube network film-based ring oscillators with sub 10-ns propagation
time and their applications in radio-frequency signal transmission. Nano
Res. 2018, 11, 1, 300–310.
[128] Rutherglen, C.; Burke, P. Carbon nanotube radio. Nano Lett. 2007, 7,
11, 3296–3299.
[129] Amlani, I.; Lewis, J.; Lee, K.; Zhang, R.; Deng, J.; Wong, H.-S. P. First
demonstration of AC gain from a single-walled carbon nanotube common-
source amplifier. Proc. Int. Electron Devices Meet. 2006, 1–4.
[130] Pliva, J.; Carta, C.; Claus, M.; Schroter, M.; Ellinger, F. On the design
of active downconversion mixers for wireless communications on a
carbon nanotube FET technology. Proc. SBMO/IEEE MTT-S Int. Microw.
Optoelectron. Conf. 2011, 984–988.
[131] Che, Y. C.; Badmaev, A.; Jooyaie, A.; Wu, T.; Zhang, J. L.; Wang, C.;
Galatsis, K.; Enaya, H. A.; Zhou, C. W. Self-aligned T-gate high-purity
semiconducting carbon nanotube RF transistors operated in quasi-ballistic
transport and quantum capacitance regime. ACS Nano 2012, 6, 6936–6943.
[132] Badmaev, A.; Che, Y. C.; Li, Z.; Wang, C.; Zhou, C. W. Self-aligned
fabrication of graphene RF transistors with T-shaped gate. ACS Nano
2012, 6, 3371–3376.
136

[133] Che, Y.; Lin, Y. C.; Kim, P.; Zhou, C. T-Gate Aligned Nanotube Radio
Frequency Transistors and Circuits with Superior Performance. ACS Nano
2013, 7, 4343−4350.
[134] Holloway, P. H. Gold/chromium metallizations for electronic devices.
Gold Bulletin 1979,12, 99-106.
[135] Novoselov, K. S.; Mischenko, A.; Carvalho, A. Castro Neto, A. H. 2D
materials and van der Waals heterostructures. Science 2016, 353, 6298.
[136] Yan, X.; Esqueda, I. S.; Ma, J.; Tice, J.; Wang, H. High breakdown
electric field in β−Ga2O3/graphene vertical barristor heterostructure. Appl.
Phys. Lett. 2018, 112, 032101.
[137] Shulaker, M. M.; Wu, T. F.; Pal, A.; Zhao, L.; Nishi, Y.; Saraswat, K.;
Wong, H. S. P.; Mitra, S. Monolithic 3d integration of logic and memory:
Carbon nanotube fets, resistive ram, and silicon fets. Int El Devices Meet
2014.
[138] Indiveri, B. G.; Liu, S. Memory and Information Processing in
Neuromorphic Systems. Proc. IEEE 2015, 103, 1379−1397.
[139] Wei, H.; Patil, N.; Lin, A.; Wong, H. S. P.; Mitra, S. Monolithic Three-
Dimensional Integrated Circuits Using Carbon Nanotube FETs and
Interconnects. Technol. Dig. - Int. Electron Devices Meet. IEDM 2009,
577−580.
137

[140] Kim, K.; Chen, C.; Truong, Q.; Shen, A. M.; Chen, Y. A Carbon
Nanotube Synapse with Dynamic Logic and Learning. Adv. Mater. 2013,
25, 1693−1698.
[141] Bushmaker, A. W.; Oklejas, V.; Walker, D.; Hopkins, A. R.; Chen, J.;
Cronin, S. B. Single-Ion Adsorption and Switching in Carbon Nanotubes.
Nat. Commun. 2016, 7, 1−8.

Abstract (if available)

Abstract In this dissertation, I present my work on the studying of the interfacial reactions between metal contacts/β-Ga₂O₃ nanomembranes to get reliable Ohmic contacts, and the development of carbon nanotube-based field-effect transistors for complementary metal-oxide-semiconductor and radio frequency electronic applications. Silicon-based technology is approaching its scaling and performance limits, but the demand for the more energy-efficient and more powerful computing units still remains. Hence, it is required to study the physics of β-Ga₂O₃ and carbon nanotubes to fully realize their potential in the next-generation electronics.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Carbon material-based nanoelectronics

PDF

Carbon nanotube nanoelectronics and macroelectronics

PDF

The adsorption and selective deposition of molecular and nanocluster ions on carbon based devices

PDF

GaAs nanowire optoelectronic and carbon nanotube electronic device applications

PDF

Printed and flexible carbon nanotube macroelectronics

PDF

Light Emission from Carbon Nanotubes and Two-Dimensional Materials

PDF

Electronic and optoelectronic devices based on quasi-metallic carbon nanotubes

PDF

Printed electronics based on carbon nanotubes and two-dimensional transition metal dichalcogenides

PDF

Single-wall carbon nanotubes separation and their device study

PDF

Carbon nanotube macroelectronics

PDF

Optoelectronic properties and device physics of individual suspended carbon nanotubes

PDF

Optical, mechanical, and electrical properties of nano-structured materials

PDF

Graphene and carbon nanotubes: synthesis, characterization and applications for beyond silicon electronics

PDF

One-dimensional nanomaterials: synthesis and applications

PDF

Controlled synthesis, characterization and applications of carbon nanotubes

PDF

Raman spectroscopy of carbon nanotubes under axial strain and surface-enhanced Raman spectroscopy of individual carbon nanotubes

PDF

Entanglement parity effects in quantum spin chains

PDF

Synthesis, assembly, and applications of single-walled carbon nanotube

PDF

Memristor device engineering and memristor-based analog computers for mobile robotics

PDF

Raman spectroscopy and electrical transport in suspended carbon nanotube field effect transistors under applied bias and gate voltages

Asset Metadata

Creator Li, Zhen (author)

Core Title Nanoelectronics based on gallium oxide and carbon nanotubes

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Physics

Degree Conferral Date 2021-12

Publication Date 09/30/2021

Defense Date 06/29/2021

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag carbon nanotube,device physics,gallium oxide,nanoelectronics,OAI-PMH Harvest,transistor

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Haas, Stephan (committee chair), El-Naggar, Moh (committee member), Nakano, Aiichiro (committee member), Wu, Wei (committee member), Zhou, Chongwu (committee member)

Creator Email zhenli768@gmail.com,zli791@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC16011559

Unique identifier UC16011559

Legacy Identifier etd-LiZhen-10120

Document Type Dissertation

Format application/pdf (imt)

Rights Li, Zhen

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu