University of Southern California Dissertations and Theses

One-dimensional nanomaterials for electronic and sensing applications

(USC Thesis Other)

One-dimensional nanomaterials for electronic and sensing applications

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content ONE-DIMENSIONAL NANOMATERIALS FOR ELECTRONIC AND SENSING
APPLICATIONS

by

Noppadol Aroonyadet

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
(ELECTRICAL ENGINEERING)

May 2015

Copyright 2015 Noppadol Aroonyadet

ii
Epigraph

"There is nothing that perseverance cannot win"
iii
Dedication

This dissertation is dedicated to my beloved family.

iv
Acknowledgement
Firstly, I would like to thank Royal Thai Government for providing me the
financial support for 6 years of my graduate studies at University of Southern California.
I would like to sincerely and gratefully thank my adviser, Professor Chongwu
Zhou for offering me an opportunity to study in the Ph.D. program at University of
Southern California, giving me tremendous and valuable guidance and supports. Without
his guidance and supports, most of my works would have never accomplished. Moreover,
I would like to thank Professor Mark Thompson for his valuable suggestion and supports
in nanobiosensing projects. In addition, I would like to thank Professor Wei Wu for
serving as a committee for both of my qualifying exam and dissertation defense.
Furthermore, I would like to thank Professor Stephen Cronin and Professor Noah
Malmstadt for giving their suggestion during my qualifying exam and their time to serve
as my qualifying exam committees.
I would like to thank our collaborators: Professor Mark Thompson, Professor
Richard Cote, Professor Ram Datar and Professor Thomas Chen for their helpful
suggestion and fruitful discussion in nanobiosensing projects.
I would like to thank former and current members of our group and others: Dr.
Fumiaki Ishikawa, Dr. Hsiao-kang Chang, Dr. Marco Curreli, Dr. Rui Zhang, Dr.Xiaoli
Wang, Yan Song, Dr. Po-Chiang Chen, Dr. Chuan Wang, Dr. Anuj Madaria, Dr.
Alexander Badmaev, Dr Lewis Gomez, Dr. Jialu Zhang, Dr. Yi Zhang, Dr. Haitian Chen,
Dr. Maoqing Yao, Dr. Yuchi Che, Dr. Bilu Liu, Dr. Gang Liu, Dr. Jiepeng Rong, Dr.
Minyuan Ge, Xin Fang, Dr. Luyao Zhang, Ahmed Abbas, Sen Cong, Qingzhou Liu, Yu
Cao, Fanqi Wu, Anyi Zhang, Liang Zhen, Zhen Li, Yuqiang Ma, Xuan Cao, Hui Gui,
Yihang Liu, Chenfei Shen, Rohan Dhall, Pyojae Kim, Pattaramon Vuttipittayamongkol,
Saowalak Sukchareonchoke, and Yue Fu for their helps and friendship.
Finally, I would like to thank my family for their encouragement and consistent
supports throughout my entire life.
v
Table of Contents
Epigraph ..................................................................................................................... ii
Dedication .................................................................................................................... iii
Acknowledgement ............................................................................................................. iv
List of Figures .................................................................................................................... ix
Abstract .................................................................................................................. xxi
Chapter 1 Introduction ..................................................................................................1
1.1 Introduction to nanowires ............................................................................1
1.2 Synthesis of nanowires ................................................................................1
1.3 Sensing mechanism ......................................................................................2
1.4 Important factors to optimize biosensing sensitivity ...................................3
1.4.1 Debye screening length .......................................................................3
1.4.2 Liquid gate electrode...........................................................................5
1.4.3 Subthreshold regime for higher sensitivity .........................................5
Chapter 2 Top-Down Polysilicon Nanoribbon Biosensors ..........................................6
2.1 Introduction ..................................................................................................6
2.2 Top-down polysilicon nanoribbon field effect transistor (FET)
fabrication ....................................................................................................7
2.3 Electrical performance of polysilicon nanoribbon FETs .............................9
2.4 pH sensing for biosensor sensitivity test ....................................................12
2.5 Surface chemistry for biomolecular immobilization on polysilicon
nanoribbon .................................................................................................14
2.6 Cancer antigen-125 (CA-125) biomarker sensing in buffer ......................16
2.7 Summary ....................................................................................................18
vi
Chapter 3 Highly Scalable, Uniform, and Sensitive Biosensors Based on Top-
Down Indium Oxide Nanoribbons and Electronic Enzyme-Linked
Immunosorbent Assay for Detection of Human Immunodeficiency
Virus p24 Proteins......................................................................................19
3.1 Introduction ................................................................................................19
3.2 Development of top-down indium oxide nanoribbon field effect
transistors and their electrical performance ...............................................21
3.3 Statistical study on electrical performance of In
2
O
3
nanoribbon
biosensors ...................................................................................................24
3.4 Investigation on long-term stability of In
2
O
3
nanoribbon devices in
aqueous buffer ............................................................................................28
3.5 pH sensing on In
2
O
3
nanoribbon biosensors and nanoribbon
thickness study ...........................................................................................29
3.6 Surface chemistry for In
2
O
3
nanoribbons ..................................................33
3.7 Integrated In
2
O
3
nanoribbon biosensors with electronic ELISA for
detection of streptavidin as the study model ..............................................37
3.8 Detection of HIV1 p24 proteins by electronic ELISA on In
2
O
3

nanoribbon biosensors ...............................................................................42
3.9 Summary ....................................................................................................45
Chapter 4 Highly Scalable, Uniform and Sensitive In
2
O
3
Nanoribbon
Biosensors for Myocardial Infarction ........................................................46
4.1 Introduction ................................................................................................46
4.2 In
2
O
3
nanoribbon biosensor chip fabrication .............................................48
4.3 Electrical Characteristic performance of the In
2
O
3
nanoribbon
biosensor chip with on-chip gate electrodes ..............................................50
vii
4.4 Electronic enzyme-linked immunosorbent assay for diagnosis of
myocardial infarction .................................................................................53
4.5 Summary ....................................................................................................60
Chapter 5 Epitaxial Growth of Aligned SnO
2
Nanowires on Sapphire and
Their Device Applications .........................................................................62
5.1 Introduction ................................................................................................62
5.2 Aligned SnO
2
nanowire synthesis and characterization ............................63
5.3 Aligned SnO
2
nanowire field effect transistors (FETs) and their
electrical performance ................................................................................72
5.4 Aligned SnO
2
nanowire transistors for the control circuit in display
application ..................................................................................................75
5.5 Aligned SnO
2
nanowire transistors for photodetecting application ...........76
5.6 Aligned SnO
2
nanowire transistors for chemical sensing
application ..................................................................................................80
5.7 Summary ....................................................................................................82
Chapter 6 Summary and Future Work ........................................................................84
6.1 Summary ....................................................................................................84
6.2 Future Work ...............................................................................................86
6.2.1 Investigation for enhancement of In
2
O
3
nanoribbon biosensor
sensitivity ..........................................................................................86
6.2.2 Optimization of the assay time of In
2
O
3
nanoribbon
biosensors for diagnosis of myocardial infarction ............................87
6.2.3 Multiplex detection on In
2
O
3
nanoribbon biosensor chips for
diagnosis of myocardial infarction....................................................88
6.2.4 Highly scalable, uniform and sensitive In
2
O
3
nanoribbon
chemical sensors ...............................................................................88
viii
Bibliography ....................................................................................................................91
Appendices
Appendix I: Top-Down Polysilicon Nanoribbon Biosensors ......................................103
I.1 Materials ..................................................................................................103
I.2 Polysilicon nanoribbon field effect transistor fabrication ........................103
I.3 Surface functionalization for poly-Si nanoribbon biosensors ..................104
Appendix II: Highly Scalable, Uniform, and Sensitive Biosensors Based on Top
-Down Indium Oxide Nanoribbons and Electronic Enzyme-Linked
Immunosorbent Assay for Detection of Human Immunodeficiency
Virus p24 Proteins....................................................................................106
II.1 Materials ..................................................................................................106
II.2 Indium oxide nanoribbon field effect transistor fabrication ....................106
II.3 Surface functionalization for In
2
O
3
nanoribbon biosensors .....................107
II.4 In
2
O
3
nanoribbon transistor Debye length calculation.............................109
Appendix III: Highly Scalable, Uniform and Sensitive In
2
O
3
Nanoribbon
Biosensors for Myocardial Infarction ......................................................111
III.1 Materials ..................................................................................................111
Appendix IV: Epitaxial Growth of Aligned SnO
2
Nanowires on Sapphire and
Their Device Applications .......................................................................112
IV.1 Materials ..................................................................................................112
IV.2 Aligned SnO
2
nanowire synthesis ...........................................................112
IV.3 Aligned SnO
2
nanowire field effect transistor (FET) fabrication ...........113
Appendix V: References ................................................................................................114

ix
List of Figures
Figure 1.1 Simulated I-V
LG
curves before and after absorption of proteins on
the surface of the nanotube to explain biosensing mechanisms with
V
DS
= 10 mV (a) Electrostatic gating effect lowers semiconducting
bands by 50 mV (b) Change in Schottky barrier height between work
function of metal and CNT for 30 mV (c) Effect from capacitance
changes with 90 % protein coverage on the surface of CNT (d) 2 %
reduction in carrier mobility from the initial value before molecular
absorption.
20
....................................................................................................3
Figure 2.1 Fabrication processes of polysilicon nanoribbons biosensors. (a)
Polysilicon is deposited via LPCVD on Si
3
N
4
/Si substrate (inset) (b)
Dopant is spun-coated on poly-Si layer and high temperature
annealing is required for diffusion process (c) Active mesas defined
by photolithography and CF
4
dry etching. (d) Metal electrodes
defined by photography and deposited by electron beam evaporation.
Silicon nitride for passivation layer is deposited by PECVD before
lift-off process (e) An optical image of poly-Si NR FET devices on a
3” wafer. (f) A magnified pattern for 1 chip containing 4 groups of
FET sensors and 1 group of 6 FET sensors (g). (h) A SEM image of
active area of a sensor. ....................................................................................8
Figure 2.2 (a) I
DS
versus V
DS
under various V
GS
for a device with 1×10
17
doping
(step of 10 V). (b) I
DS
versus V
GS
under various V
DS
for a device
with 1×10
17
doping (step: 1V). (c) I
DS
versus V
DS
under various V
GS

for a device with 5×10
17
doping (step: 10V). (d) I
DS
versus V
GS

under various V
DS
for a device with 5×10
17
doping (step: 1V). (e) I
DS

versus V
DS
under various V
GS
for a device with 1×10
18
doping (step:
10V). (f) I
DS
versus V
GS
under various V
DS
for a device with 1×10
18

doping (step: 1V). .........................................................................................10
x
Figure 2.3 Statistical distributions of electrical performance from 20 poly-Si
nanoribbon FET devices at 1x10
17
doping concentration (a) On-state
current (I
ON
) at V
DS
= 1 V and V
GS
= - 40 V. (b) On/off current ratio
(c) Threshold voltage (V
TH
) (d) Transconductance (g
m
) (e) Electron
motilities (µ). ................................................................................................11
Figure 2.4 pH sensing of poly-Si nanoribbon FETs with pH 4 to pH 9 buffered
solutions (a) Normalized sampling current under pH 4 to pH 9 for
three devices measured simultaneously. (b) The relationship between
average normalized response from (a) and change in pH of media
solution (c) I
DS
–V
GS
curves when pH of the solution changed from
pH = 4 to pH = 9 (d) The difference of threshold voltage at each pH
relative to V
TH
at pH = 9 and change of buffered pH solution. ....................12
Figure 2.5 (a) Normalized real-time responses of poly-Si nanoribbon devices
versus time from change in pH 7.2 to pH 8.0 (b) Plot of average
normalized responses (I/I
0
) and change in pH of the media solution
extracted from (a) ..........................................................................................13
Figure 2.6 Fluorescent images of (a) 500 nm SiO
2
on Si substrate (b) 500nm
Si
3
N
4
on Si substrate (c) 55 nm poly-Si on Si
3
N
4
/Si substrate and (d)
55 nm poly-Si with 10 nm deposited SiO
2
layer were functionalized
to immobilized amine biotin and streptavidin with red fluorescent
dye (e) 500 nm Si
3
N
4
on Si substrate and (f) 55 nm poly-Si on
Si
3
N
4
/Si substrate were functionalized to anchor amine-PEG
molecules and incubated in streptavidin with fluorescent red dye. ..............16
Figure 2.7 CA-125 biomarker sensing with poly-Si nanoribbon biosensors (a)
Normalized response of one of three poly-Si nanoribbon sensors
monitored simultaneously (b) Normalized sensing response when
non-target molecules (BSA) introduced in the system at 150 times
higher concentration. (c) Normalized response versus CA-125
concentration with the Langmuir isotherm fitting. .......................................17
xi
Figure 3.1 Fabrication processes of In
2
O
3
nanoribbon biosensors. (a) The first
photolithography step defining metal electrodes on top of Si
3
N
4
on
Si wafer substrate (b) 5/45 nm Ti/Au metal electrode deposited by
evaporation followed by lift-off. (c) The second photolithography
step defining nanoribbon active channel. (d) In
2
O
3
was deposited by
RF sputtering and lift-off to expose In
2
O
3
nanoribbon channel. (e)
An optical image of a 3 inch wafer of In
2
O
3
nanoribbon biosensors.
Inset shows a magnified image of a nanoribbon chip composing of 4
subgroups of 6 nanoribbon devices. (f) An SEM micrograph of
nanoribbon devices in a subgroup. ................................................................22
Figure 3.2 (a) A family of I
DS
-V
DS
curves and (b) A family of I
DS
-V
GS
curves
from an In
2
O
3
nanoribbon FET. ....................................................................23
Figure 3.3 Families of I
DS
-V
DS
and I
DS
-V
GS
curves of (a, b) an InGaZnO
nanoribbon device (c, d) a SnO
2
nanoribbon device (d, e) an ITO
nanoribbon device at V
DS
= 200mV. Optical images of a ZnO
nanoribbon device (g) before and (h) after 14 hour incubation in
PBS. ZnO nanoribbons dissolved completely in PBS after 14 hours. ..........24
Figure 3.4 Plots of electrical performance from 50 In
2
O
3
nanoribbon transistors
(a) On-state current (I
ON
) at V
GS
= 30 V and V
DS
= 600 mV. (b)
Threshold voltage (V
TH
) (c) Mobilities (µ) and (d) On-state to off-
state current ratios at V
DS
= 600 mV. ............................................................25
Figure 3.5 (a) On-state current measured from 50 In
2
O
3
nanoribbon devices
from Figure 3.4 (a) in logarithmic scale with two identical SEM inset
images taken from different representative devices on the substrate.
(b) On-state current measured from 50 devices In
2
O
3
nanowire FET
devices with inset SEM images taken from two different devices to
show non-uniformity of nanowire devices ...................................................26
Figure 3.6 Distribution of electrical performance measured from 30 In
2
O
3

nanoribbon devices in 0.01x PBS solution using a Ag/AgCl gate
xii
electrode. (a) On-state drain current (I
ON
) at V
DS
= 200 mV and V
LG
= 1 V (b) Transconductance (g
m
) at V
DS
= 200 mV (c) Threshold
voltage (V
TH
) at V
DS
= 200 mV and (d) On state to off-state current
ratios at V
DS
= 200 mV..................................................................................27
Figure 3.7 Key electrical performance parameters measured from 18 In
2
O
3

nanoribbon devices immerged in 1x PBS at room temperature for the
aqueous stability test (a) On-state current (I
ON
) measured at V
DS
=
200 mV and V
LG
= 1 V. (b) Transconductance (g
m
) at V
DS
= 200
mV. (c) Threshold voltage (V
TH
) at V
DS
= 200 mV. (d) On-state to
off-state current ratio at V
DS
= 200mV. ........................................................29
Figure 3.8 (a) Schematic diagram of pH sensing experiment on In
2
O
3

nanoribbon devices. Commercial pH buffer solution was confined in
a Teflon electrochemical chamber. Liquid gate voltage was applied
through a Ag/AgCl electrode. (b) Real-time response obtained from
a 20 nm In
2
O
3
nanoribbon device exposed to commercial buffer
solutions with pH 4 to 9 (c) Real-time responses obtained from three
20 nm In
2
O
3
nanoribbon devices in buffer solutions with pH in the
physiological range 6.7 to 8.2 with step of 0.3. (d) Real-time
responses from In
2
O
3
nanoribbon devices at different thickness
ranging from 10 to 50 nm exposed to commercial buffer solutions
with pH 4 to 9. ..............................................................................................31
Figure 3.9 Plots from X-ray diffraction spectroscopy on 20 nm In
2
O
3
film (a)
Before annealing (b) After annealing in low vacuum at 300 ºC for 30
minutes. (c) Comparison of average pH sensing responses obtained
from three as-sputtered and annealed devices exposed to commercial
pH buffer solutions with pH in a range of 4 to 9. .........................................32
Figure 3.10 (a) A schematic diagram of a fluorescent sample functionalized with
amine biotin to immobilize streptavidin conjugated with fluorescent
red dyes. (b) A schematic diagram of fluorescent a negative control
xiii
functionalized with amine PEG molecules which cannot bind with
streptavidin conjugated with red fluorescent dyes. (c) A schematic
diagram of In
2
O
3
pads on Au metal electrodes for fluorescent study
(d) An optical image of In
2
O
3
pads on Au electrodes for fluorescent
experiment. Fluorescent images of (e) sample and (f) negative
control of In
2
O
3
pads fabricated on 500 nm SiO
2
on the Si substrate.
Fluorescent images of (g) sample and (h) negative control of In
2
O
3

pads fabricated on 500 nm Si
3
N
4
on Si substrate. .........................................34
Figure 3.11 SEM micrographs taken from In
2
O
3
nanoribbons functionalized with
phosphonic acid linker molecules and immobilized (a) amine-PEG3
biotin (b) amine-PEG on the surface of nanoribbons before
introduced 4 nM streptavidin conjugated with 20 nm Au
nanoparticles. ................................................................................................35
Figure 3.12 Fluorescent images of In
2
O
3
nanoribbons functionalized with 1 mM
3-phosphonopropionic acid for (a) 2 hours (b) 4 hours (c) 5.5 hours
(d) 7 hour. On the top row, amine-PEG3 biotin molecules were
immobilized on In
2
O
3
nanoribbons while in the bottom, row amine-
PEG molecules were anchored to In
2
O
3
nanoribbons before
introducing streptavidin with fluorescent dye. (e) Plot of thickness of
etched In
2
O
3
layer versus 3-phosphonopropionic acid incubation
time. ..............................................................................................................36
Figure 3.13 (a) Schematic diagram of streptavidin electronic ELISA. (b) An
optical micrograph of an In
2
O
3
nanoribbon (brown color) on Ti/Au
metal electrodes after incubating with 1 µM streptavidin conjugated
with red fluorescent dyes (c) A fluorescent image of the device in (b)
to confirm binding of streptavidin molecules on an In
2
O
3
nanoribbon
having biotin capture probes (d) Real-time responses measured from
three In
2
O
3
nanoribbon devices after incubating with streptavidin
with fluorescent dyes without presence of urease enzymes in the
xiv
sensing chamber. Devices showed increase in conduction due to
decrease in pH from 7.4 of 0.01xPBS to 6.61 of 100 mM urea in
0.01xPBS. (e) Normalized real-time responses of 1 µM streptavidin
electronic ELISA from 3 In
2
O
3
nanoribbon devices monitored
simultaneously. Introducing urea into the sensing chamber, hydroxyl
groups on the surface of nanoribbons are deprotonated due to urea-
urease enzyme reaction which consumes hydrogen ions in the
solution yielding more negatively gating effect and decrease in
electrical conduction of nanoribbon devices. ................................................39
Figure 3.14 (a) Plots of drain current versus liquid gate voltage (I
DS
-V
LG
)
measured from devices with and without In
2
O
3
nanoribbon in
0.01xPBS. Plots of I
DS
versus time measured from (b) an In
2
O
3

nanoribbon device and (c) a device without In
2
O
3
nanoribbon in
0.01xPBS. .....................................................................................................41
Figure 3.15 (a) A plot of average normalized current responses and streptavidin
concentration calculated from 3 devices monitored simultaneously in
each concentration. (b) A plot of pH changes in the sensing chamber
measured by a commercial pH meter and streptavidin concentration. .........42
Figure 3.16 (a) Schematic diagram of electronic ELISA for HIV1 p24 detection
(b) Real-time responses monitored from 3 In
2
O
3
nanoribbon devices
simultaneously at 20 fg/ml of p24 proteins in PBS. Conduction of all
devices decreased upon presence of urea in the sensing chamber. (c)
A plot of average normalized responses from 3 devices at each p24
concentration and p24 concentration in pg/ml. (d) A plot of change
in pH in the sensing chamber measured from a commercial pH meter
and p24 protein concentration in pg/ml. .......................................................44
Figure 4.1 (a) An optical image of 5 fabricated In
2
O
3
nanoribbon biosensor
chips of a 3 inch 500 nm SiO
2
on Si wafer (b) A magnified image of
a biosensor chip which consists of 4 subgroups for multiplex
xv
detection . (c) An photograph of a PDMS microwell stamp on top of
a biosensor chip for sensing experiment. (d) A magnified photo of
one subgroup in a chip comprising 5 devices and a gate electrode at
the 3rd position. (d) An image of two identical nanoribbon in a
subgroup. .......................................................................................................49
Figure 4.2 (a) Schematic diagram of electrochemical measurement to test
effectiveness of on-chip gate. Plots of I
DS
-V
GS
curves applied gate
voltage through (b) a Ag/AgCl electrode and (c) an on-chip Ti/Au/Ti
gate electrode on the same device. (d) A plot of liquid gate potential
(V
REF
) and drain current (I
DS
). ......................................................................50
Figure 4.3 Plots of key electrical performance from 20 In
2
O
3
nanoribbon
devices using on-chip gate electrodes (a) On-state current (I
ON
) at
V
DS
= 200 mV and V
LG
= 2 V (b) Transconductance (g
m
) (c)
Threshold voltage (V
TH
) (d) On-state to off-state current ratios at
V
DS
= 200 mV. ..............................................................................................52
Figure 4.4 (a) A schematic diagram of pH sensing experiment on an In
2
O
3

nanoribbon biosensor chip. Gate voltage was applied through the
Ti/Au/Ti on-chip gate electrode. A Teflon electrochemical cell was
mounted on a biosensor chip to confine commercial pH solution
during the experiment. (b) Average normalized real-time sensing
responses measured from 3 nanoribbon devices exposed to
commercial pH solutions from pH 4 to pH 9. ...............................................53
Figure 4.5 (a) A schematic diagram of electronic ELISA for troponin I
detection (b) Normalized real-time responses monitored from three
In
2
O
3
nanoribbon sensors simultaneously at 300 pg/ml of troponin I
proteins in 1xPBS without presence of urease enzymes in the sensing
chamber. Conduction of all sensors increased due to decrease in pH
of urea solution in the chamber. (c) A plot of real-time responses
from the same devices in (b) after immobilization of biotinylated
xvi
urease on nanoribbons. Conduction of all three devices decrease
upon increase in pH in the sensing chamber due to reaction between
urea and urease enzymes. (d) A plot of average responses from three
devices at each troponin I concentration in pg/ml. .......................................55
Figure 4.6 Electronic ELISA for detection of troponin I with short assay time
(a) Real-time sensing responses of three In
2
O
3
nanoribbon devices
monitored concurrently at troponin I concentration of 300 pg/ml in
PBS. Normalized conduction of nanoribbon devices decreased from
the base line due to increase in pH in the sensing chamber caused by
hydrolysis of urea from reaction with urease enzymes. (b) A plot of
changes in average normalized responses for three devices at each
troponin I. (c) Comparison between average responses of long assay
time and short assay times. ...........................................................................57
Figure 4.7 (a) A schematic diagram of electronic ELISA for detection of CK-
MB biomarkers (b) A plot of average normalized responses from
three devices at each CK-MB concentration. (c) A schematic
diagram of electronic ELISA for detection of BNP biomarkers (b) A
plot of average normalized responses from three devices at each
BNP concentration with a blind test of BNP in human whole blood. ..........59
Figure 5.1 Aligned SnO
2
nanowire growth study. (a-c) Orientation of A-plane
(a), M-plane (b), and R-plane (c) sapphire. Wafers are outlined on
each plane with the wafer flat bolded. (d-i) AFM scans of A(d, e),
M(f, g), and R(h, i) sapphire surfaces before(d, f, h) and after(e, g, i)
annealing. Scale bars are 30 nm. (j-l) SEM images of aligned SnO
2

nanowires grown on A-plane (j), M-plane (k), and R-plane (l)
sapphires. Sapphire orientations are included on the bottom left. ................64
Figure 5.2 Histograms of nanowire assembly parameters of aligned SnO
2

nanowires on annealed A-plane, annealed M-plane, and R-plane
sapphire. (a - c) aligned SnO
2
nanowire density on annealed A-
xvii
plane (a), annealed M-plane, (b) and R-plane sapphire substrates (c)
from 20 samples for each plane. (d-e) SnO
2
nanowire alignment
defect density on annealed A- plane (d), annealed M-plane (e), and
R-plane sapphire substrates (f) from 20 samples for each plane. (g-i)
nanowire misalignment angles of 100 nanowires from 20 samples on
annealed A- plane (g), annealed M-plane (h) and R-plane sapphire
substrates (i) ..................................................................................................66
Figure 5.3 XRD data for aligned SnO
2
nanowires grown on A-plane (a), M-
plane (b), and R-plane (c) sapphires show all three planes tend to
interface the SnO
2
(101) plane. .....................................................................67
Figure 5.4 Diagrams of atomic arrangement for A-plane sapphire (a), (101)
plane SnO
2
(b), and R-plane sapphire (c). Dashed vectors show
sapphire to SnO
2
lattice alignment in the y direction while solid
vectors show alignment in the x direction. Green circles are oxygen
atoms, pink circles are aluminum atoms, and gray circles are tin
atoms. ............................................................................................................68
Figure 5.5 Transmission electron microscopy (TEM) imaging of aligned SnO
2

nanowire on R-plane sapphire substrate. (a) Diagram of cross-
sectional sample preparation by FIB lift-out technique, from the as-
grown sample to protective platinum and carbon deposition, and to
finished cross-sectional TEM sample. Direction of the electron beam
is parallel to the nanowire growth direction as indicated by the
double arrow. (b) Bright spots are electron diffraction pattern of
sapphire taken from the cross-sectional sample. Black dots are
simulated diffraction pattern of sapphire looking into the [1 ‾ 101]
lattice vector direction. (c) Bright spots are electron diffraction
pattern of SnO
2
taken from the cross-sectional sample. Black dots
are simulated diffraction pattern of SnO
2
looking into the [1 ‾ 01]
lattice vector direction. The faint rings are diffraction from the
xviii
protective Pt layer. (d) TEM image showing the location on the
cross-sectional sample from where the SnO
2
electron beam
diffraction patterns in (b) and (c) are taken. Area enclosed by the
dashed circles with underscored b and c correspond to image (b) and
(c), respectively. ............................................................................................69
Figure 5.6 (a) TEM image of a cross-sectional view of aligned SnO
2
nanowire
on sapphire. (b) Electron diffraction pattern of SnO
2
nanowire taken
from similar cross-section locations. (c) Electron diffraction pattern
of cross-section of R-plane sapphire. ............................................................70
Figure 5.7 (a) High-resolution TEM image of the cross section of an aligned
SnO
2
nanowire on R-plane sapphire substrate. (b) High-resolution
SEM image of an aligned SnO
2
nanowire on R-plane sapphire. (c)
SEM image of SnO
2
nanowires grown on R-plane sapphire with the
synthesis pressure set to atmospheric pressure. Dense, un-aligned
nanowire forest can be seen with such synthesis condition. .........................71
Figure 5.8 Aligned SnO
2
nanowire transistor study. (a) Diagram of device
fabrication. (b) Top-view optical image of an aligned SnO
2
nanowire
transistor. (c) SEM image of two aligned SnO
2
nanowires bridging
the source and drain electrodes of a nanowire transistor. (d) I
D
-V
D

family plot of the transistor shown in (c). (e) I
D
-V
G
plot of the
transistor in (c), plotted with standard scale in black and logarithmic
scale in blue. The subplot shows the transconductance of the same
device. ...........................................................................................................73
Figure 5.9 Histograms of device performance from 20 aligned SnO
2
nanowire
transistors. (a) On state drain current (I
D
) at V
D
= 200 mV, V
G
= 10
V. (b) On/Off ratio of I
D
. (c) Transconductance (g
m
), (d) Threshold
voltage (V
TH
). (e) Mobility (µ). ...................................................................74
Figure 5.10 Aligned SnO
2
nanowire FET control circuit for OLED. (a) Circuit
diagram of OLED connection to FET. (b) I
OLED
-V
G
family curve (c)
xix
I
OLED
-V
DD
family curve (d) Optical images of OLED intensity as V
G

decreases. ......................................................................................................76
Figure 5.11 Photoconduction and polarization detection. (a) I
D
-V
D
curves of
aligned SnO
2
device before UV illumination (black), after 365 nm
UV illumination (red), and after 254 nm UV illumination (blue).
Expanded curves for before and 365 nm illumination are presented
in the subplot for clarity, with the same color legend. (b) Real-time
detection of 254 nm UV illumination on an aligned SnO
2
device as
the UV lamp is turned on and off for 5 cycles. (c) Real-time
detection of 2 different wavelengths using one aligned SnO
2
device.
(d) Polarized 254 nm UV detection. Black triangles are averaged
peak I
D
during the time that the UV lamp is turned on at the
corresponding angle. The red curve is fitted data showing cos2θ
dependence. ...................................................................................................78
Figure 5.12 (a) Stability test of real-time response from UV 254 nm illumination
on aligned SnO2 nanowire detector with V
DS
= 500 mV.
Photoconduction decreased and reached steady state after 60 minutes
(b) Long term stability test of photoconduction. Each data point is an
average photoconduction response over 200 s after 254 nm UV
illumination measured from the same sensor used in Figure 4.11. ...............79
Figure 5.13 NO
2
sensing (a) Real-time detection of NO
2
gas of various
concentrations by 2 different aligned SnO
2
nanowire devices. NO
2

gas is turned on at point “a” and turned off at point “b” (b) Plot of
normalized drain current change (ΔI/I
0
) against NO
2
concentration.
In the subplot, the inverse of normalized current and concentration
are shown. The black triangles represent measured data, and the red
line is a linear fit of the 4 concentrations data. (b) I
D
-V
D
plots of
device 1 after being introduced to increasing concentrations of NO
2
. .........81
xx
Figure 6.1 Average pH sensing response from three In
2
O
3
nanoribbon
biosensors after immobilization of urease enzymes with
normalization at pH 7.4................................................................................87
Figure 6.2 NO
2
sensing (a) Real-time detection of NO
2
gas with several
concentration of NO
2
by In
2
O
3
nanoribbon sensors (b) A plot of
normalized sensing response from three sensors and NO
2
concentration with Langmuir isotherm fitting having correlation
coefficient of 0.97. ........................................................................................90
Figure I.1 Schematic diagram of surface functionalization for poly-Si
nanoribbon biosensors (a) Hydroxyl groups were generated by O
2
plasma treatment (b) Amine termination was anchored by immersing
in mixture of 5% 3-APDMS in anhydrous toluene for 2 hours (c)
Carboxyl function group was generated by incubating in 5mg/ml of
succinic anhydride in anhydrous THF and 5% triethylamine for 4
hours (d) Antibody was immobilized by interacting with NHS ester
group after conversion of carboxylic group to NHS ester by
EDC/NHS. ..................................................................................................105
Figure II.1 Schematic diagram of In
2
O
3
surface chemistry to immobilize
antibodies or amine molecules ....................................................................108
Figure II.2 Schematic diagram of synthesis of biotinylated phosphonic acid
linker molecules. .........................................................................................109
Figure II.3 (a) Plot of Debye length versus charge density in the In
2
O
3

nanoribbon device (b) I
DS
-V
DS
in the narrow linear regime with V
GS

from 27 to 9 V with step of 3 V ..................................................................110

xxi
Abstract
Nanostructure based field effect transistors (FET) have drawn attention from
researchers around the globe due to their great electrical performance suitable for many
electronic applications. In addition, one-dimensional nanomaterials have large surface to
volume ratio which is a desirous characteristic for sensing applications. One-dimensional
nanomaterial based FET sensors have been demonstrated their ultrahigh sensitivity to
detect biomolecules for medical application and toxic gases for industrial applications.
However, control assembly of one-dimensional nanomaterials is still one of challenges to
hinder them from practical uses. In the first chapter of this dissertation, basic and theories
about nanowire synthesis and key components for the sensing application have been
addressed. In the chapter 2 to 4, we demonstrate scalable top-down approaches to
fabricate nanoribbon FETs using 2 photolithographic masks to define precisely
dimension and position of metal electrodes and nanoribbons. The relaxation in the lateral
dimension facilitates the simple and scalable fabrication process compatible with the
conventional microelectronic facilities yielding 100 % functional devices with uniform
electrical performance. Thickness of nanoribbons can be precisely controlled in the
deposition process.
In chapter 2, we select low pressure chemical vapor deposition poly-silicon
(LPCVD poly-Si) as the nanoribbon material. Spin-on dopant solution with thermal
annealing are chosen to tune doping concentration in the material without need of any
toxic gas or expensive ion implantation. Poly-Si nanoribbon devices exhibit good ionic
sensitivity both in wide pH range from pH 4 to 9 and in physiological solution range from
pH 7.2 to 8. To optimize number of capture probes on nanoribbons, thin layer of silicon
dioxide needs to be deposited on nanoribbon to improve surface chemistry. We
demonstrate detection of cancer antigen-125 (CA-125), a biomarker for ovarian cancer,
using our poly-Si nanoribbon devices with limit of detection 10 U/ml which is an order of
magnitude lower than the clinically relevant level.
xxii
In chapter 3, we change the nanoribbon material from poly-Si to indium oxide
(In
2
O
3
). In
2
O
3
is an inherently semiconducting material without requirement of doping
concentration and its deposition process is radio frequency sputtering at room
temperature enabling the use of low cost substrates instead of silicon. In
2
O
3
nanoribbon
devices exhibit better electrical performance and long-term stability in the aqueous
condition over other metal oxides. In addition, In
2
O
3
nanoribbon devices show excellent
ionic sensitivity in both wide pH range (pH 4 to 9) and physiological solution pH range
(pH 6.8 to 8.2). Combination between the In
2
O
3
nanoribbon platform and a signal
amplification technique, electronic enzyme-linked immunosorbent assay (ELISA), we
achieve high sensitivity to target analytes such as streptavidin and human
immunodeficiency virus type 1 (HIV-1) p24 proteins. This approach circumvents Debye
screening effect in ionic solution to bypass complicated sample preparation and
demonstrates detection of p24 protein at 20 fg/ml (about 250 viruses/ml) or 3 orders of
magnitude lower than commercial ELISA kits with 35 % conduction change in human
blood serum. With the demonstrated sensitivity, scalability and uniformity, the In
2
O
3

nanoribbon sensor platform makes a great progress toward clinical testing such as early
diagnosis of acquired immunodeficiency syndrome (AIDS).
In chapter 4, we have demonstrated an integration of the on-chip gate electrode to
the In
2
O
3
nanoribbon sensor chip to move our platform closer to the practical setting. The
on-chip gate exhibits similar gating performance to the traditional Ag/AgCl electrode at
the same liquid potential. The In
2
O
3
nanoribbon biosensor chip with the on-chip gate
shows excellent ionic sensitivity in the wide range of pH from 4 to 9. A biomarker panel
for diagnosis of acute myocardial infarction, namely troponin I, creatine kinase MB (CK-
MB), and b-type natriuretic peptide (BNP), is selected to demonstrate the point of care
application for the emergency setting. To fit with requirement of the medical emergency,
the assay time is reduced from 10 hours to be an hour without loss of significant
sensitivity. We demonstrate detection of troponin I at 100 fg/ml with 30% change in
conduction with the projected limit of detection about 15 ag/ml or about 5 orders of
magnitude lower than commercial troponin I ELISA kits. We have completed calibration
xxiii
curves for all 3 biomarkers and testified our calibration curves with spiked 500 pg/ml
BNP in human whole blood to simulate a sample from a patient. Average sensing
response from this blind test shows acceptable 3% lower than the calculated value and
still differentiate the patient's condition correctly. Further optimization and more
statistical study are required for the clinical use.
In chapter 5, we demonstrate control assembly synthesis of rutile semiconducting
aligned SnO
2
nanowires on A-plane, M-plane and R-plane sapphire substrates for
scalable and practical device applications. X-ray diffraction and transmission electron
microscopy were used to characterize aligned nanowires for their growth mechanism and
direction. Simple photolithography for patterning metal electrodes is required for
fabrication of nanowire FETs. Transistors exhibit excellent electrical performance with
on/off current ratio on the order of 10
6
, electron mobilities about 71.68 cm
2
/V.s and high
current to the external organic light emitting diode display. In addition to electronic
application, aligned nanowire devices can be utilized as photodetectors to differentiate
between 254 and 365 nm ultra violet wavelengths. Their alignment also helps to
segregate among different polarization angles with polarization ratio of
photoconductance (σ) of 0.3. Lastly, we demonstrate align SnO
2
nanowire devices as
scalable and ultrasensitive nitrogen dioxide (NO
2
) chemical sensors at concentration of
0.2 ppb.
In the last chapter, all topics mentioned in this dissertation are summarized and
the future work for the In
2
O
3
nanoribbon sensor is discussed.

1
Chapter 1 Introduction
1.1 Introduction to nanowires
Nanowires are one of the major one-dimensional nanomaterials which have drawn
interests from researchers more than a decade to investigate their synthesis mechanism
1
,
assemblies
2
, their electrical
3
, optical
4
, and magnetic properties, and prospective
applications
3, 5, 6
. They have been demonstrated for great potential in many applications
such as electronic circuits,
3, 4
energy harvesting
7, 8
, bio-and chemical sensing
6, 9, 10
. They
can be categorized by their electrical properties: metal (Ag)
11
, semiconducting (Si
6
, Ge
12
,
GaAs
13
, InAs
1
, In
2
O
3
14
, SnO
2
15
, ZnO
16
etc.) or polymer
8
.
1.2 Synthesis of nanowires
Many mechanisms have been used to synthesize nanowires such as chemical
vapor deposition (CVD)
17
, solution based synthesis
18
or etching from bulk materials
19
.
However, the most common approach is CVD with vapor-solid-liquid (VLS) mechanism.
Firstly metal catalysts (Au, Ni, Fe, etc.) are deposited on the grown substrate. For
example for SnO
2
nanowire synthesis, we can deposit Au film and anneal it at high
temperature to form Au nanoparticle droplets or synthesized Au nanoparticles directly on
to the grown substrate. After that, the catalyst deposited substrate is loaded in to a furnace
and heated up to certain temperature while Sn vapor is introduced into the system with
carrier gas (usually N
2
or Ar). Firstly, Sn vapor will diffuse to Au particle to form Sn-Au
alloy. When this alloy reaches supersaturated stage and Sn vapor is kept supplying into
the synthesis system, Sn will precipitate out and react with residual of oxygen inside the
system to form SnO
2
nanowires. As long as, Sn vapor is continued to feed into the
system, this process will contribute to growth of nanowires and the metal catalyst is
pushed up.

2
1.3 Sensing mechanism
In biosensing experiment, conduction will change when target analytes binds to
capture probes on the surface of biosensors. Many researchers have proposed possible
mechanisms which are electrostatic gating of the sensing channel, change in Schottky
barrier height, alteration in carrier mobility, and modification of dielectric constant to
explain interaction between molecules and behavior of the sensor. Heller et al. has
systematically studied and identified possible mechanisms for biosensing with simulation
and experimental results of ambipolar carbon nanotube devices.
20
Figure 1.1 (a)
demonstrated the electrostatic gating effect from binding of molecules on the sensing
channel. I-V
LG
was shifted when gate potential of the channel was changed, but
transconductance before and after molecular absorption remained constant. Figure 1.1 (b)
shows changes in Schottky barrier height due to absorption of molecules on source and
drain electrodes which altered metal work function and conduction in both p and n
branches inversely. For Schottky barrier mechanism, threshold voltage remained constant
while transconductance decreased in p-branch and increased in n branch when we
compared values from before and after the sensing experiment. Figure 1.1 (c) shows an
effect from reduction of capacitance due to lower value of dielectric constant due to
absorbed proteins with 90 % coverage on the surface of the nanotube device than one of
buffer solution. In this mechanism, transconductance in both branches decreased due to
weaker gate coupling. Figure 1.1 (d) shows effect from reduction in carrier mobility due
to non-uniform distribution of bound charged molecules. When mobility was decreased,
electrical conduction and transconductnace were reduced in both branches, but threshold
voltage was still constant before and after protein absorption. At the summary this study,
electrostatic gating and change in Schottky barrier height are influent in alteration of
sensor behaviors in biosensing experiment. If metal electrodes are passivated, it is most
likely that sensing responses are affected by electrostatic gating from molecular
interaction.

3

Figure 1.1 Simulated I-V
LG
curves before and after absorption of proteins on the surface of the
nanotube to explain biosensing mechanisms with V
DS
= 10 mV (a) Electrostatic gating effect lowers
semiconducting bands by 50 mV (b) Change in Schottky barrier height between work function of
metal and CNT for 30 mV (c) Effect from capacitance changes with 90 % protein coverage on the
surface of CNT (d) 2 % reduction in carrier mobility from the initial value before molecular
absorption.
20

1.4 Important factors to optimize biosensing sensitivity
To achieve high sensitivity for biosensing, we need to consider key parameters to
optimize our sensing performance.
1.4.1 Debye screening length
Debye screening length (λ
D
) is a distance where the number of opposite charges
from the electrolyte solution approach charged proteins which will suppress electrostatic
gating potential and reduce sensing responses significantly if charged proteins are located

4
beyond this length. It was demonstrated by Stern et al that Debye length plays an
important role in determining the magnitude of sensing responses.
21
The Debye length in
the thermally equilibrium system at room temperature can be calculated as following
equation:
λ

=

∑

(1.1)
where l
B
is the Bjerrum length = 0.7 nm, summation is for all ion species, and ρ
i
and z
i

are the density and the valence of the ion, respectively. From this equation, Debye
screening in standard 1x phosphate buffer saline (PBS) solution was calculated as 0.7 nm.
If we dilute this buffer 10 time to be 0.1xPBS, we will increase the Debye length to be
2.3 nm. For further dilution to 0.01xPBS, it yields the Debye length of 7.3 nm. To
calculate electrostatic potential from a charged molecule as a function of distance, we can
use following equation:
φr
Q
πε

e

(1.2)
where φ(r) is the potential change at distance r, Q is the charge, ε
0
is the dielectric
constant of the medium, and k
0
is 1/λ
D
. This is a solution for screened Poisson equation.
From equation (1.2), the effect from a charged molecule reduces significantly when the
distance is longer than the Debye screening length. This is evidence that if we want to
maximize our sensing performance, we have to pay more attention to ionic strength of
our medium. However, it was proposed that the Debye length can be increased from the
value calculated from (1.1) 10 times by injecting the external flow of ions.
22
This external
flow will cause imbalance between diffusive ions and drifting ions and the diffusive
component is not completely compensated the drift component which yield weaker
screening effect. From the biosensing perspective, we can extend the Debye screening
length by pressuring flow of target analytes in the microfluidic channel to simulate this
effect.

5
1.4.2 Liquid gate electrode
Another important component in the biosensing set-up is the liquid gate electrode
used to tune operational point to our sensors. It is essential to maintain constant applied
liquid gate potential throughout the experiment to avoid artifact responses from the
sensing setup which may lead to incorrect interpretation of sensing results. There are
several types of the gate electrode used in electrochemical experiment. A Pt electrode is
widely used as for liquid gate biasing due to its stability in electrolytes during
electrochemical experiment.
23-25
Recently, Minot et al reported that potential applied
through a Pt electrode in the biosensing experiment was unstable due to absorption of
bovine serum albumin (BSA) proteins on the surface of this Pt wire.
26
They proposed to
use an Ag/AgCl electrode to apply gate voltage to the sensor and results from their
experiment exhibited more stable result resulting from separation between the electrode
and analyte solution by the porous glass membrane which prevents large molecules to
reach surface of Ag/AgCl electrode and allows only ionic exchange.
1.4.3 Subthreshold regime for higher sensitivity
After selecting ionic strength of sensing medium and a proper type of gate
electrode, we have to select an operational point in subthreshold regime to maximize the
sensing response. Gao et al have demonstrated exponential enhancement of sensing
sensitivity in subthreshold regime.
27
They performed pH sensing on 3 different regimes:
subthreshold, near-threshold and linear regime. Changes in conduction from pH 4 to pH 9
showed 600% in the subthreshold regime while responses from linear and near-threshold
regimes were 50% and 150%, respectively.
27

6
Chapter 2 Top-Down Polysilicon Nanoribbon Biosensors
2.1 Introduction
In recent years label-free, electrical nanobiosensors have drawn lots of research
interests due to the potential of achieving superior time and cost efficiency to current
state-of-the-art biosensing platform such as ELISA. Among nanobiosensors studied by
various research teams, most of them are fabricated by the “bottom-up” technique,
5, 6, 9, 17,
28-32
which nanostructures are assembled to make devices. One of the major challenges
for nanosensors fabricated by the bottom-up technique is assembly, which can
significantly limit the yield and uniformity of such nanosensors.
30
The yield is highly
related to cost and throughput, and uniformity is essential to the reliability of
nanobiosensors. Although intensive research efforts have been made toward assembly of
nanostructures, most of the techniques still lack controllability, reproducibility and
scalability. The other school of process in nanotechnology is “top-down” fabrication
which seeks to create nanoscale devices by using larger, externally-controlled ones to
direct their assembly.
19, 33-44
Top-down fabrication is much more controllable than their
bottom-up counterpart, thus leads to more uniform device performance. As a result,
nanobiosensors fabricated using these top-down approaches can yield more reliable
diagnosis. One of the major challenges for top-down nanobiosensors, however, is to
achieve large surface-to-volume ratio, as surface-to-volume ratio is directly linked to
sensitivity. Research efforts have been made toward reducing the critical dimension of
nanostructures (hence increase surface-to-volume ratio) by applying techniques such as
electron beam lithography and directional etching.
35, 41, 45
However, such techniques are
extremely time-consuming, cost-inefficient, and have poor scalability. Those drawbacks
significantly limit the commercial impact the aforementioned nanobiosensors can
potentially generate. In this regard, a recent study demonstrates that highly sensitive
nanobiosensors can be achieved without pushing critical dimensions to a limit.
19, 33, 34, 43

By carefully limiting the active layer thickness, silicon nanoribbon based biosensors can

7
be fabricated by conventional photography (~µm critical dimension) with clinically
relevant sensitivity. Despite the promising result, single-crystalline silicon on insulator
(SOI) wafers with extremely thin (~50nm) active layer are required to produce the
nanobiosensors. SOI wafers with such thin active layers are not easily available, and thus
can be very expensive. Also, precise oxidation and wet etching are needed to achieve the
desired thickness, and thus further limit the time and cost efficiency of such platform.
In section, we describe a nanobiosensor platform based on polysilicon
nanoribbons. The devices are fabricated with ~100% yield and great uniformity due to
top-down process. The polysilicon is deposited using a highly scalable, precise and cost-
efficient low-pressure chemical vapor deposition (LPCVD). Unlike SOI wafers, the
thickness of polysilicon active layer and dielectric layer can be well controlled and easily
customized for different applications. The whole process is compatible with conventional
photolithography with only two masks required, thus are incredibly time and cost
efficient. Moreover, the fabrication can be performed on full wafers, which results in
great scalability. By performing pH sensing experiment with a wide dynamic range and
high sensitivity, we demonstrate that the polysilicon nanoribbon sensors are highly
sensitive to ions. Finally biomarker detection is performed with clinically relevant
sensitivity, and thus confirms the practical value of polysilicon nanoribbon biosensors.
Multiple devices are monitored simultaneously during all sensing experiments, and the
response shows great uniformity. We demonstrate that polysilicon nanoribbon can act as
a highly efficient, reliable and scalability platform for nanobiosensors. Such a platform
exhibits great potential toward label-free, electrical biosensor with enormous practical
impact.
2.2 Top-down polysilicon nanoribbon field effect transistor (FET)
fabrication
The device fabrication process is shown in Figure 2.1. Fabrication process starts
with a simple silicon nitride (Si
3
N
4
) on silicon wafer as in the inset of Figure 2.1 (a). The

8
silicon nitride thickness is determined by the desired dielectric thickness. A thin layer
(typically 50 nm) of polysilicon is deposited via LPCVD as shown in Figure 2.1 (a).
Spin-on dopant of desired doping concentration is applied to the polysilicon via spin
coating. The diffusion is performed under 1100°C for 15 minutes in nitrogen
environment. This process is shown in Figure 2.1 (b) A buffered HF solution with 7:1
ratio is used to remove spin-on dopant upon the completion of diffusion process.
Photolithography and a CF
4
dry etch are then performed to define the contact lead and
nanoribbon area as shown in Figure 2.1(c). A second step of photolithography is then
performed to define the metal contact. Finally, 5 nm Ti and 45 nm Au are deposited as

Figure 2.1 Fabrication processes of polysilicon nanoribbons biosensors. (a) Polysilicon is deposited
via LPCVD on Si
3
N
4
/Si substrate (inset) (b) Dopant is spun-coated on poly-Si layer and high
temperature annealing is required for diffusion process (c) Active mesas defined by photolithography
and CF
4
dry etching. (d) Metal electrodes defined by photography and deposited by electron beam
evaporation. Silicon nitride for passivation layer is deposited by PECVD before lift-off process (e) An
optical image of poly-Si NR FET devices on a 3” wafer. (f) A magnified pattern for 1 chip containing
4 groups of FET sensors and 1 group of 6 FET sensors (g). (h) A SEM image of active area of a
sensor.

9
electrodes via electron beam evaporation as shown in Figure 2.1 (d). A buffered HF
dipping is required to remove thin layer of native oxide before the metal evaporation. A
silicon nitride passivation layer can be deposited via plasma enhanced chemical vapor
deposition (PECVD) before metal lift-off process to reduce electrical current leakage
during sensing experiment. We note that the device fabrication is highly efficient and
scalable with only two masks required and all the process involved can be performed on
wafer scale as shown in Figures 2.1(e) which is photographic images of hundreds of
polysilicon nanoribbon sensor arrays fabricated on a whole 3” wafer. In Figure 2.1 (f),
each chip has 4 sub-groups of 6 FET devices. A magnified image if each sub-group is
shown in Figure 2.1 (g). A scanning electron microscopic (SEM) image of polysilicon
nanoribbons, respectively is shown in Figure 2.1 (h) with width of 2 µm.
2.3 Electrical performance of polysilicon nanoribbon FETs
After fabrication, devices were characterized with an Agilent semiconductor
analyzer 4156B and yield of functional devices are about 100 % which is higher than the
bottom up approach due to higher controllability. During the fabrication process, three
different doping concentrations, 1×10
17
, 5×10
17
and 1×10
18
, were applied to polysilicon
by spin-coating. We characterized the electrical properties for devices of all three doping
concentrations after device fabrication by back gat measurement. During the
measurement, voltage is applied to the silicon substrate with 500nm silicon oxide as
dielectric, and the result is shown in Figure 2.2. Plots in Figure 2.2 (a) and (b) are the
source-drain current (I
DS
) versus source-drain voltage (V
DS
) under various back gate
voltage (V
GS
) and I
DS
versus V
GS
under various V
DS
for a device with 1×10
17
doping,
respectively. The device exhibits ohmic contact behavior with strong gate dependence.
On/off ratio reaches 250 and such value is high enough for nanobiosensor application.
We note that the on/off ratio can be further improved by using dielectrics with higher
dielectric constant, or by applying the gate voltage via liquid gate which yields stronger
gating effect. The I
DS
versus V
DS
under various V
GS
and I
DS
versus V
GS
under various V
DS

for a device with 5×10
17
doping concentration are shown in Figure 2.2 (c) and (d),

10
respectively. The device is 10 times more conductive compared to the 1×10
17
device.
However, the on/off ratio is reduced to 62. The I
DS
versus V
DS
under various V
GS
and I
DS

versus V
GS
under various V
DS
for a device with 1×10
18
doping concentration are shown in
Figure 2.2 (e) and (f), respectively. The device exhibits higher on-state current and lower
on/off ratio compared to devices with lower doping concentrations. Further study
indicates that devices with low doping concentration yield better sensitivity as
nanobiosensors. Thus, the 1×10
17
doped devices are selected for further sensing
experiments in this chapter unless otherwise stated.

Figure 2.2 (a) I
DS
versus V
DS
under various V
GS
for a device with 1×10
17
doping (step of 10 V). (b)
I
DS
versus V
GS
under various V
DS
for a device with 1×10
17
doping (step: 1V). (c) I
DS
versus V
DS
under
various V
GS
for a device with 5×10
17
doping (step: 10V). (d) I
DS
versus V
GS
under various V
DS
for a
device with 5×10
17
doping (step: 1V). (e) I
DS
versus V
DS
under various V
GS
for a device with 1×10
18

doping (step: 10V). (f) I
DS
versus V
GS
under various V
DS
for a device with 1×10
18
doping (step: 1V).
To demonstrate uniformity of device performance that our top-down polysilicon
nanoribbon FETs can produce, we have plotted key figure-of-merits from 20 devices
fabricated on 500 nm Si
3
N
4
on Si substrate to suppress competition binding of

11
biomolecules at 1x10
17
doping concentration as shown in the Figure 2.3. Figure 2.3 (a)
shows that average I
DS
measured at V
DS
= -1 V and V
GS
= -40 is -20.68 nA with the
standard deviation of 2.76 nA. The on/off current ratios for all transistors fall in a range
of 200 and 1000 as shown in Figure 2.3 (b). Figure 2.3 (c) shows a tight variation of
threshold voltage with the average V
TH
of -13.89 V and the standard variation of 1.09 V.
The average transconductance from these 20 devices is 1.13 nS with the standard
variation of 0.15 nS shown in Figure 2.3 (d). The electron mobility was calculated from
the relationship with transconductance

⁄

where the gate capacitance (C)
is calculated from the parallel plate model ( !/# and channel length (L) of 80 µm.
From relative dielectric constant of Si
3
N
4
equal to 7.5, estimate plate area (A) of 460 µm
2

and Si
3
N
4
dielectric thickness (d) of 500 nm, the gate capacitance is calculated to be
6.11x10
-14
F. From these parameters, mobilities of all 20 devices are plotted in Figure 2.3
(e) with the average electron mobility of 1.21 cm
2
/V•s and the standard variation of 0.16
cm
2
/V•s.

Figure 2.3 Statistical distributions of electrical performance from 20 poly-Si nanoribbon FET devices
at 1x10
17
doping concentration (a) On-state current (I
ON
) at V
DS
= 1 V and V
GS
= - 40 V. (b) On/off
current ratio (c) Threshold voltage (V
TH
) (d) Transconductance (g
m
) (e) Electron motilities (µ).

12
2.4 pH sensing for biosensor sensitivity test
To demonstrate ionic sensitivity of the polysilicon nanoribbon FETs, we
preformed pH sensing with three unfunctionalized devices simultaneously shown in
Figure 2.4. During the sensing experiment, a -300 mV source-drain voltage and a -
300mV liquid gate voltage is applied to the devices. The source-drain current is
constantly monitored by Agilent semiconductor analyzer B1500A. In Figure 2.4 (a),
devices were first soaked in pH 9 buffered solution and then the buffer was exchanged to
pH 8. All three devices showed significant decrease in conductance after the buffer
exchange, and such decrease in conductance can be explained by the p-type transistor
behavior of the polysilicon devices. Further buffer exchanges were performed to change

Figure 2.4 pH sensing of poly-Si nanoribbon FETs with pH 4 to pH 9 buffered solutions (a)
Normalized sampling current under pH 4 to pH 9 for three devices measured simultaneously. (b) The
relationship between average normalized response from (a) and change in pH of media solution (c) I
DS
–V
GS
curves when pH of the solution changed from pH = 4 to pH = 9 (d) The difference of threshold
voltage at each pH relative to V
TH
at pH = 9 and change of buffered pH solution.

13
the pH to 7, 6, 5, 4, 5, 6, 7, 8 and 9, respectively. The normalized current versus time for
all 3 devices is plotted in Figure 2.4 (a). All three devices showed decreased conductivity
under lower pH solutions. The average normalized responses from all three devices at
each pH in Figure 2.4 (a) are plotted in Figure 2.4 (b). The variation in conductivity from
pH 4 to pH 9 is around 300% with exponentially increasing trend with correlation
coefficient of 99.79% which is similar to what observed from silicon nanowire FETs.
6

The devices show a wide dynamic range to pH variation. Figure 2.4 (c) shows I
DS
-V
GS
curves of a poly-Si nanoribbon FET measured in different pH solution. The linear shift in
threshold voltage toward negative direction was observed (shown in Figure 2.4 (d)) when
pH of the solution was decreased from pH 9 to pH 4 due to increase in H
+
ions in the
solution which yields more positively gating voltage through the p-type poly-Si active
area. Further pH sensing experiment (shown in Figure 2.5) is carried out with a pH step
of 0.2 in a range of human physiological solution between pH 7.2 to pH 8.0. Figure 2.5
(a) shows normalized responses from a nanoribbon device with 10% increase toward 0.2
change in pH with high signal-to-noise ratio. The average sensing responses from the
device exhibit a linear relationship to a narrow range of variation in pH with the
correlation coefficient of 99.55% as shown in Figure 2.5 (b). From these pH sensing
experiments, we can conclude that polysilicon nanoribbon sensors are highly sensitive to
ionic changes.

Figure 2.5 (a) Normalized real-time responses of poly-Si nanoribbon devices versus time from
change in pH 7.2 to pH 8.0 (b) Plot of average normalized responses (I/I
0
) and change in pH of the
media solution extracted from (a)

14
2.5 Surface chemistry for biomolecular immobilization on polysilicon
nanoribbon
In order to enable polysilicon nanoribbon FETs to function as biosensors, we need
to immobilize biomolecules as capture probes to enhance selectivity and specificity
towards the target molecules. Several methods have been proposed and utilized to
demonstrate highly sensitive Si based biosensors.
17, 19, 21, 28, 33-35, 37-40, 42, 43, 46
They can be
categorized to be two major different approaches that are functionalization with and
without SiO
2
on Si active materials.
35
Devices functionalized without native oxide on Si
need to be dipped in hydrofluoric acid (HF) to strip off native oxide and later to be
performed in the controlled inert environment to prevent a thin layer of native oxide to be
regenerated. However, they showed higher sensitivity than devices functionalized with
SiO
2
on top of the Si material.
35
To reduce complexity and a chance to fail to attach
biomolecules during surface functionalization, we selected to functionalize our poly-Si
nanoribbon FETs with SiO
2
on the surface of poly-Si nanoribbons. Detail of the surface
functionalization can be found in the Appendix I.3. In Figure 2.6, the functionalization
scheme mentioned in the Appendix I.3 has been treated on different substrates for
qualitative comparison of the effectiveness of the surface chemistry. Streptavidins (SA)
with red fluorescent dyes have been selected as our target molecules due to their high
contrast and high selectively binding between biotin and SA. Figure 2.6 (a) to (d) are
fluorescent images taken from Nikon Fluorescent microscope of commercial 500 nm
SiO
2
on Si, 500 nm LPCVD Si
3
N
4
deposited on Si substrate, 55 nm LPCVD poly-Si, and
10 nm evaporated SiO
2
on 55 nm LPCVD poly-Si, respectively. They were
functionalized with 3-Aminopropyldimethylethoxysilane (APDMS), attached amine-
polyethylene glycol (PEG)-3 biotins as the capture probes, and incubated in 2 µM SA
with red dyed for 2 hours before rinsed off with PBS and water. Si
3
N
4
substrate (Figure
2.6 (6) showed less SA binding than the SiO
2
substrate (Figure 2.6 (b)) which can be seen
from comparison of their brightness and contrast. From this finding, if we use Si
3
N
4
as
the substrate for poly-Si deposition, we can reduce competition binding of the target

15
molecules from the substrate significantly yielding to increase sensitivity of our
biosensors. From Figure 2.6 (c), when we deposited a thin SiO
2
about 10 nm on poly-Si,
this type of sample showed more SA with dye binding than poly-Si sample with the
native oxide (Figure 2.6 (d)). However, we need to trade of between number of binding
sites from the extra oxide layer and extended distance of target molecules from the poly-
Si nanoribbons which will affect the sensitivity of our sensors due Debye’s screening
effect.
21
There are several ways to increase the oxide layer, namely oxidation and
deposition. Dry oxidation is normally used and more appropriate to grow a 10 nm oxide
layer than the wet oxidation because of its slower growth rate to have better control,
better oxide quality and less defects. Dry oxidation is normally performed at high
temperature and will affect doping concentration of our nanoribbon because dopants can
be driven further during oxidation process which can cause poly-Si FET to be more
resistive and to increase leakage current through the substrate. The evaporation of the thin
SiO
2
layer is recommended because it is performed at room temperature and will not
change the electrical properties of poly-Si nanoribbon FETs, but it is required another
photolithographic step to pattern trenches to deposit SiO
2
only the nanoribbon area.
Figure 2.6 (e) and (f) are fluorescent images of negative controls on 500 nm Si
3
N
4
and 55
nm poly-Si, respectively. They were functionalized with APDMS, immobilized amine-
PEG molecules instead of amine-biotins and incubated in the same concentration of SA
with red dyes for 2 hours. They showed dark images because SA with dyes did not bind
with PEG groups and were washed away after SA incubation.

16

Figure 2.6 Fluorescent images of (a) 500 nm SiO
2
on Si substrate (b) 500nm Si
3
N
4
on Si substrate (c)
55 nm poly-Si on Si
3
N
4
/Si substrate and (d) 55 nm poly-Si with 10 nm deposited SiO
2
layer were
functionalized to immobilized amine biotin and streptavidin with red fluorescent dye (e) 500 nm Si
3
N
4

on Si substrate and (f) 55 nm poly-Si on Si
3
N
4
/Si substrate were functionalized to anchor amine-PEG
molecules and incubated in streptavidin with fluorescent red dye.
2.6 Cancer antigen-125 (CA-125) biomarker sensing in buffer
Biomarker detection was also performed. We chose cancer antigen 125 (CA-125),
an ovarian cancer biomarker, as the target of study. Ovarian cancer is one of the major
causes of death of women around the globe.
31
The early stage diagnosis is important
because patients can receive proper treatment on time to increase their survival rate. The
poly-Si nanoribbon devices were first conjugated with CA-125 antibodies as protocols
mentioned in the Appendix I so that the devices can bind to CA-125 selectively. During
the sensing experiment, a 300 mV source-drain voltage and a 300 mV liquid gate voltage
is applied to the devices. CA-125 was progressively added to the sensing environment.
The normalized current versus time of one of three devices is plotted in Figure 2.7 (a).
Note that three devices showed uniform response to the presence of CA-125. All three
devices started to show response at CA-125 concentration of 10U/ml (50pM).
31
Such

17
limit of detection is one order of magnitude lower than the clinically relevant level for
diagnosis.
31
The devices showed larger response to CA-125 of higher concentrations. In
comparison, none of the devices showed response to 150 nM bovine serum albumin as
shown in the Figure 2.7 (b), even the concentration is 300 times higher than the target
analyte. The Figure 2.7 (c) shows that sensing responses fit well with Langmuir isotherm
model. The sensing data shows that the polysilicon nanosensor can detect biomarkers
with clinically relevant sensitivity and great selectivity. The results suggest that
polysilicon nanoribbon biosensors have great potential as a platform for clinical diagnosis
of diseases such as cancers.

Figure 2.7 CA-125 biomarker sensing with poly-Si nanoribbon biosensors (a) Normalized response of
one of three poly-Si nanoribbon sensors monitored simultaneously (b) Normalized sensing response
when non-target molecules (BSA) introduced in the system at 150 times higher concentration. (c)
Normalized response versus CA-125 concentration with the Langmuir isotherm fitting.

18
2.7 Summary
In conclusion, we have developed a label-free, electrical biosensor platform based
on polysilicon nanoribbons that can be fabricated by conventional photolithography with
low cost, good uniformity, scalability and controllability of electrical performance. The
devices showed good sensing response to pH changes with a wide dynamic range and
with physiological range. To increase sensing sensitivity, Si
3
N
4
substrates were used to
reduce completion binding from the substrate and thin layer of 10 nm SiO
2
on poly-Si
helps to improve surface chemistry by increasing number of capture probes on sensing
surface. Biomarker detection is demonstrated using CA-125, an ovarian cancer
biomarker, as a study model with clinically relevant sensitivity. Conductivity of multiple
devices is monitored simultaneously during all the measurements, and the data confirms
the uniformity of device performance. Such results suggest that polysilicon nanoribbon
devices exhibit great potential to act as a highly efficient, reliable and sensitive platform
for future nanobiosensors.

19
Chapter 3 Highly Scalable, Uniform, and Sensitive Biosensors
Based on Top-Down Indium Oxide Nanoribbons and
Electronic Enzyme-Linked Immunosorbent Assay for
Detection of Human Immunodeficiency Virus p24 Proteins
3.1 Introduction
Nanobiosensors based on nanostructure field-effect-transistors (FETs) have
become an area of intense research because their detection of biomolecules has potential
applications ranging from health monitoring to drug discovery. Their real-time, highly
sensitive, and electrical sensing capabilities have been demonstrated in the detection of
biomolecules such as proteins
6, 9, 19, 31, 32
, nucleic acids
40, 47
, viruses
48
, and other small
molecules
46
. Nanobiosensors fabricated using bottom-up assembly, such as nanowire
biosensors, are highly sensitive because bottom-up synthesis yields high crystalline
quality and critical dimensions as small as a few nanometers.
6, 49
Top-down biosensors,
on the other hand, can add scalability and device uniformity, but must overcome
fabrication challenges to achieve similar sensing performance. This has propelled the
recent emergence of several top-down nanobiosensor fabrication techniques based on
silicon-on-insulator (SOI) and poly-silicon materials.
19, 33, 50, 51
Among these, electron-
beam lithography
52
, imprint lithography
53
, spacer technique
50
, and photolithography
19

have been explored to optimize dimensional control and scalability with promising
results. Specifically, top-down nanoribbon biosensors, with relaxed lateral dimensions,
have allowed straightforward photolithographic processing, resulting in high uniformity
and functional device yield. This relaxed dimensional requirement can be achieved
without losing sufficient sensitivity because the critical dimension, which is the channel
depth, remains in the nanoscale. Their larger area also provides more surface for analytes
to bind.
33, 43

20
Metal oxides have been traditionally used as sensor materials because their
surfaces are very sensitive to changes in the environment.
10, 15, 16
Metal oxides are
inherently semiconducting without impurity doping, and their electrical properties are
stable during sensing. As an example, In
2
O
3
nanowires have been successfully applied to
highly sensitive chemical sensors,
3, 10
biosensors,
9, 31, 32
optical detectors,
4
thin film
transistors (TFTs),
54
and other electronic applications.
55
In this work, we investigate the
sensing properties of In
2
O
3
nanoribbon FET biosensors for the detection of human
immunodeficiency virus type 1 (HIV-1) p24 antigen. The In
2
O
3
nanoribbon thickness is
well controlled by the highly scalable radio frequency (RF) sputtering technique, and the
entire sensor is fabricated by a top-down, two-mask conventional photolithographical
process.
Diagnosis of HIV-1 infection, the cause of acquired immune deficiency syndrome
(AIDS), relies on the detection of HIV-1 ribonucleic acid (RNA), capsid antigen p24, and
anti-HIV antibody.
56
Most common food and drug administration (FDA)-approved HIV-
1 diagnostic assays target HIV-1 RNA because of the ease of polymerase chain reaction
(PCR) mediated amplification. Although this method is very sensitive, it is expensive,
requires well-trained staff, and must be performed in well-equipped laboratories. RNA
testing in rural or remote settings is also a major challenge. Testing for the HIV-1 p24
antigen is a good alternative because it can be done in resource-limited situations.
Antigen p24 level is significantly high during the early, acute phase of infection and the
terminal stage of AIDS. It is also a useful marker for predicting CD4+ T cell count
decreases, disease progression for early detection of HIV-1 infection, and patient
prognosis. Early detection of HIV infection helps to prevent HIV transmission and to
prolong AIDS condition by receiving proper treatment. HIV-1 p24 antigen is usually
detected by enzyme-linked immunosorbent assay (ELISA). However, the detection
sensitivity of the conventional assay is less than desirable: 10–20 pg/ml.
57, 58
Many
research groups have tried to lower the limit of detection for early diagnosis of HIV using
several approaches such as modifying ELISA with a booster step
57
, sandwich p24
amperometric immunosensor based on a modified electrode
59
, nanoparticle based

21
fluorescent assay
58, 60
and atomic force microscopy of sandwich HIV p24 nanoarrays
61
to
the level close to HIV nucleic acid amplification test.
62, 63
From these attempts, the lowest
detectable level of HIV p24 based on modified ELISA techniques is 10 fg/ml with the
projected limit of detection to be 5 fg/ml.
60
As an alternative approach, we demonstrate
In
2
O
3
nanoribbon biosensors detection of p24 proteins at 20 fg/ml, or about 250
viruses/ml,
60
with a 35% conduction change from a baseline conduction measured in
human blood serum. We have projected limit of detection of our sensors to be about 200
ag/ml with 1% change in conduction. This detection limit can possibly diagnose HIV
infection about 7 to 10 days earlier than the detectable window enabled by conventional
ELISA.
60
In fact, this detection limit is much closer to the 40 HIV viruses/ml detection
limit from PCR.
62, 63
We are able to achieve this sensitive electrical detection in
physiological solutions because electronic ELISA circumvents Debye length screening of
electrical signals in high salt concentration and has amplification from the enzyme-
substrate reaction. Moreover, the scalable and low-cost nanoribbon platform can be
highly portable for point-of-care field-testing.
3.2 Development of top-down indium oxide nanoribbon field effect
transistors and their electrical performance
The fabrication of In
2
O
3
nanoribbon biosensors requires two photolithography
steps. First, 500 nm of Si
3
N
4
is deposited on Si substrates by low pressure chemical
vapor deposition (LPCVD). Silicon nitride is used instead of silicon dioxide as the
passivation layer because it is more resistant to our surface chemistry, resulting in better
suppression of competition binding of biomolecules to the surface of the substrate. In
Figure 1 (a), the first lithography step defines the source and drain electrodes. It is
followed by electron beam evaporation of 5 nm Ti and 45 nm Au, as shown in Figure 1
(b). After that, both the dimension and the position of the nanoribbons are defined by the
second mask, as shown in Figure 1 (c). The In
2
O
3
nanoribbons are then deposited by RF
sputtering with thickness targeted at 10 nm to 50 nm. Nanoribbons are formed after the

22
lift-off process which is the last step in the fabrication. Because nanoribbons are never
exposed to any additional layer or photoresist, this fabrication order leaves a pristine
nanoribbon surface for surface chemistry. Figure 1 (d) shows a wafer-scale photo of the
fabricated In
2
O
3
nanoribbon FETs with 100% yield, and its inset shows a magnified
optical image of one nanoribbon chip, which contains 4 subgroups of 6 nanoribbon FET
devices. A comparison between the scanning electron microscopy (SEM) images of two
nanoribbon FETs in Figure 1f shows that the 50 µm by 2 µm channels are identical. This
uniformity in fabrication is expected to extend to the electrical properties of the devices.

Figure 3.1 Fabrication processes of In
2
O
3
nanoribbon biosensors. (a) The first photolithography step
defining metal electrodes on top of Si
3
N
4
on Si wafer substrate (b) 5/45 nm Ti/Au metal electrode
deposited by evaporation followed by lift-off. (c) The second photolithography step defining
nanoribbon active channel. (d) In
2
O
3
was deposited by RF sputtering and lift-off to expose In
2
O
3

nanoribbon channel. (e) An optical image of a 3 inch wafer of In
2
O
3
nanoribbon biosensors. Inset
shows a magnified image of a nanoribbon chip composing of 4 subgroups of 6 nanoribbon devices. (f)
An SEM micrograph of nanoribbon devices in a subgroup.
After device fabrication, In
2
O
3
nanoribbon FETs were characterized by an Agilent
semiconductor analyzer 4156B using a back gate. Figures 2 (a) and 2 (b) respectively
show family plots of I
DS
-V
DS
and I
DS
-V
GS
curves measured from an In
2
O
3
nanoribbon
device. Figure 2 (a) exemplifies a good metal oxide semiconductor field effect

23
transistor(MOSFET) behavior where I
DS
varies linearly with V
DS
at low voltage range
and saturates at voltages higher than 20 V. In addition, both Figures 2 (a) and 2 (b) show
that I
DS
decreases with decreasing V
GS,
in accordance with n-channel transistor behavior.
From both the I
DS
-V
DS
and the I
DS
-V
GS
plots, the device is turned off at V
GS
below 0 V,
with on-state/off-state current ratios in the range of 10
5
to 10
6
.

Figure 3.2 (a) A family of I
DS
-V
DS
curves and (b) A family of I
DS
-V
GS
curves from an In
2
O
3
nanoribbon FET.
Besides In
2
O
3
, we have also studied other possible materials using this top-down
approach. From Figure 3.3 (a) to (f), we have included family I
DS
-V
DS
and I
DS
-V
GS
plots
for sputtered InGaZnO, SnO
2
and ITO nanoribbon devices. Both InGaZnO (Figures 3.3
(a) and (b)) and SnO
2
(Figures 3.3 (c) and (d)) show good MOSFET behavior, but they
are 10 and 100 times more resistive than the In
2
O
3
nanoribbon device in Figure 3.2 (a)
and (b), respectively. Furthermore, SnO
2
and InGaZnO nanoribbon devices have
threshold voltages between 50 V to 60 V and on/off current ratios in the range of 10
3
to
10
5
, making these two materials less efficient and less sensitive than In
2
O
3
as FET
sensors. Measurements of ITO nanoribbon devices (Figures 3.3 (e) and (f)) show even
less desirable MOSFET behavior. They exhibit poor gate voltage dependence and a low
on/off current ratio around one. Another metal oxide commonly used for transistors and
sensors, ZnO, also proves to be incompatible with biosensing in a liquid environment.
Figures 3.3 (g) and (h) show optical images of as-fabricated ZnO nanoribbons in air and

24
after 14 hours in 1x phosphate buffer saline (PBS), respectively. We observed that ZnO is
unstable in the PBS solution and dissolved completely after 14 hours. These results
support In
2
O
3
as the optimal nanoribbon material for our study.

Figure 3.3 Families of I
DS
-V
DS
and I
DS
-V
GS
curves of (a, b) an InGaZnO nanoribbon device (c, d) a
SnO
2
nanoribbon device (d, e) an ITO nanoribbon device at V
DS
= 200mV. Optical images of a ZnO
nanoribbon device (g) before and (h) after 14 hour incubation in PBS. ZnO nanoribbons dissolved
completely in PBS after 14 hours.
3.3 Statistical study on electrical performance of In
2
O
3
nanoribbon
biosensors
Statistical analyses of key electrical properties for 50 In
2
O
3
nanoribbon FETs are
plotted in Figure 3.4 (a) to (d). The dielectric used in these studies is 50 nm SiO
2
. The
average on-state current (I
ON
) measured at V
DS
= 600 mV and V
GS
= 30 V is 918.6 nA
with a standard deviation of 36 nA, or 4% from the average value, as shown in Figure 3.4
(a). The threshold voltage (V
TH
) is extracted from the I
DS
-V
GS
curve with V
DS
set to 600
mV, and the distribution is plotted in Figure 3.4 (b) with an average of 15.38 V and a
standard deviation of 0.78 V, or 5% of the average. The electron mobility (µ) was

25
calculated from the relationship g
m
= µ(C/L
2
)V
DS
where g
m
is taken as the maximum of
the derivative of the I
DS
-V
GS
curve. The gate capacitance (C) was calculated from the
parallel plate model (C= A/d), and the channel length (L) is 80 µm. With a SiO
2
relative
dielectric constant of 3.9, calculated plate area (A) of 460 µm
2
, and SiO
2
dielectric
thickness (d) of 50 nm, the gate capacitance is calculated to be 3.18x10
-13
F. From these
parameters, mobilities of all 50 devices are calculated and plotted in Figure 3.4 (c). The
average electron mobility is 23.38 cm
2
/V•s, and the standard deviation is 1.42 cm
2
/V•s or
6 % of the average. Lastly, Figure 3.4 (d) shows the on-state and off-state current ratios
from 50 devices. Most of the In
2
O
3
nanoribbon devices demonstrate a good on/off ratio
between 10
5
to 10
7
.

Figure 3.4 Plots of electrical performance from 50 In
2
O
3
nanoribbon transistors (a) On-state current
(I
ON
) at V
GS
= 30 V and V
DS
= 600 mV. (b) Threshold voltage (V
TH
) (c) Mobilities (µ) and (d) On-
state to off-state current ratios at V
DS
= 600 mV.

26
To bench mark the uniformity of our “top-down” nanoribbon devices, the on-
state current of 50 In
2
O
3
nanoribbon FETs (Figure 3.5 (a)) is compared to that of 50 In
2
O
3

nanowire FETs shown in Figure 3.5 (b). In
2
O
3
nanoribbon devices show more uniform
on-state current than nanowire devices due to their high dimensional control. The SEM
insets in Figure 3.5 (a) of two identical representative devices reflect the low device-to-
device variation of the “top-down” approach. On the other hand, the large variation in the
on-state current of In
2
O
3
nanowire devices are attributed to the difference in the number
of nanowires bridging between metal electrodes, as shown in the inset SEM images in
Figure 3.5 (b). This variation is inherent in the “bottom-up” fabrication process.

Figure 3.5 (a) On-state current measured from 50 In
2
O
3
nanoribbon devices from Figure 3.4 (a) in
logarithmic scale with two identical SEM inset images taken from different representative devices on
the substrate. (b) On-state current measured from 50 devices In
2
O
3
nanowire FET devices with inset
SEM images taken from two different devices to show non-uniformity of nanowire devices
In addition to demonstrating the electrical uniformity of In
2
O
3
nanoribbon devices
in air, we have also measured 30 devices in 0.01x PBS buffer solution to investigate
sensor uniformity in the biosensing environment. Key characteristics in solution are
shown in Figure 3.6. Figure 3.6 (a) shows that the average of the on-state current (I
ON
)
measured at V
DS
= 200 mV and liquid gate voltage (V
LG
) = 1 V is 1.39 µA, with a
standard deviation of 0.11 µA (8% of the average). Transconductance of these 30 devices
(Figure 3.6 (b)) falls in a narrow distribution between 3.5 µS to 4.5 µS, with the average

27
at 3.76 µS and a standard variation of 0.30 µS (8% of average). Figure 3.6 (c) shows that
variation in threshold voltage (V
TH
) is also small, with the average at 0.53 V and the
standard variation at 0.01 V, only 2% of the average. Lastly, on-state to off-state current
ratios fall between 10
4
and 10
5
as shown in Figure 3.6 (d). The small device-to-device
variations in liquid environment show good potential for biosensing application.

Figure 3.6 Distribution of electrical performance measured from 30 In
2
O
3
nanoribbon devices in
0.01x PBS solution using a Ag/AgCl gate electrode. (a) On-state drain current (I
ON
) at V
DS
= 200 mV
and V
LG
= 1V (b) Transconductance (g
m
) at V
DS
= 200 mV (c) Threshold voltage (V
TH
) at V
DS
= 200
mV and (d) On state to off-state current ratios at V
DS
= 200 mV.

28
3.4 Investigation on long-term stability of In
2
O
3
nanoribbon devices in
aqueous buffer
For biosensing applications, devices are regularly exposed to fluids with certain
ionic concentration and osmolarity. The time of exposure can even be lengthened for
monitoring cellular signals from live cells. The electrical stability of devices in such
media will play an important role towards biosensing applications because degraded
devices can give false signals. To test device stability in physiological fluids, we immerse
the devices in 1xPBS as the model fluid. Electrical measurements were taken once every
10 days for the first 50 days, then once every 30 days after the initial 50 day period. Key
electrical characteristics of 18 In
2
O
3
nanoribbon sensors were extracted from their I
DS
-
V
LG
curves and plotted in Figure 3.7. In Figure 3.7 (a), the average on-state current (I
ON
)
retained about 70% of the initial average current after 80 days and remained stable
around this level after 4 months. In Figures 3.7 (b) and (c), fluctuations in both the
average transconductance (g
m
) and the threshold voltage (V
TH
) are within around 30% of
the initial values. In Figure 3.7 (d), the median on/off current ratio remained within an
order of magnitude after 4 months. These trends show good long-term stability of the
In
2
O
3
nanoribbon sensors in physiological fluids, and this is achieved without requiring
additional protection layers to prevent the hydrolysis of native oxide as in the case of
silicon channels.
64
This advantage makes the In
2
O
3
nanoribbon platform a promising
candidate for in-vivo and in-vitro applications.

29

Figure 3.7 Key electrical performance parameters measured from 18 In
2
O
3
nanoribbon devices
immerged in 1x PBS at room temperature for the aqueous stability test (a) On-state current (I
ON
)
measured at V
DS
= 200 mV and V
LG
= 1 V. (b) Transconductance (g
m
) at V
DS
= 200 mV. (c)
Threshold voltage (V
TH
) at V
DS
= 200 mV. (d) On-state to off-state current ratio at V
DS
= 200mV.
3.5 pH sensing on In
2
O
3
nanoribbon biosensors and nanoribbon
thickness study
To test the sensitive of the In
2
O
3
nanoribbon FET biosensors to changes in ionic
concentration, we conducted a series of pH detection experiments under different
conditions, as shown in Figure 3.8. A schematic diagram of the setup for pH sensing is
shown in Figure 3.8 (a). We mechanically mount a Teflon electrochemical cell on top of
an In
2
O
3
nanoribbon chip in order to confine the sensing solution to the FET channel
area. Then we manually exchange different solutions varying in pH in and out of the
sensing chamber using pipettes. Figure 3.8 (b) shows the real-time sensing response of a

30
20 nm un-functionalized In
2
O
3
nanoribbon device to commercial pH buffer solutions
ranging from pH 4 to pH 9. The device shows an increase in conduction with decreased
pH. This trend agrees with the gate voltage modulation behavior of n-channel transistors:
a higher number of positive hydrogen ions in a lower pH solution yields higher current
due to a positive gating effect on the In
2
O
3
nanoribbon device. The normalized sensing
response shows excellent pH sensitivity over a wide range between pH 4 to 9. To test the
nanoribbon device sensitivity to pH changes in a range more relevant for biosensing
applications, we have performed pH sensing experiment in the physiological range from
pH 6.7 to pH 8.2, with steps of about 0.3. Figure 3.8 (c) shows the average of the
normalized current responses from three In
2
O
3
nanoribbon FETs with 20 nm thickness.
These devices show a 5 times decrease in conduction with a pH change of 1.5.
We have also investigated the dependence of sensor performance on the
nanoribbon depth by comparing pH sensing signals from sensors with varying
thicknesses. The distance into the semiconducting nanoribbon at which surface charges
are no longer felt is defined as the Debye length, $

%&
'
(/)

*, where %, k
B
, T, q,
and n stand for the permittivity (7.97x10
-13
F/cm for In
2
O
3
)
65
, the Boltzmann's constant,
temperature, charge constant, and charge density, respectively.
33
To achieve good
sensitivity, the optimal nanoribbon thickness needs to be within the transistor Debye
length. For the In
2
O
3
nanoribbons, this Debye length is calculated to be 23 nm using
methods described in the Appendix II. To test how this affects pH sensitivity, In
2
O
3

nanoribbons are fabricated with thicknesses ranging from 10 to 50 nm. As shown in
Figure 3.8 (d), conduction of all devices decreases exponentially when the pH increases
from pH 4 to 9, regardless of the nanoribbon depth. Similar exponential conduction
change in response to pH variation has been observed from the unfunctionalized Si
nanowire FET platform.
6
As expected, the 10 nm and 20 nm In
2
O
3
nanoribbon devices
are the most sensitive to the ionic change in the buffer solution because it is easier to
deplete carriers in the channel with thickness within the Debye length. However, the 10
nm In
2
O
3
nanoribbon FET also shows the highest fluctuation in sensing signals, which

31
can be contributed to lower film uniformity during In
2
O
3
deposition. For the rest of the
experiment, we targeted a 20 nm ribbon thickness, which gives us good stability and
sensitivity.

Figure 3.8 (a) Schematic diagram of pH sensing experiment on In
2
O
3
nanoribbon devices.
Commercial pH buffer solution was confined in a Teflon electrochemical chamber. Liquid gate
voltage was applied through a Ag/AgCl electrode. (b) Real-time response obtained from a 20 nm
In
2
O
3
nanoribbon device exposed to commercial buffer solutions with pH 4 to 9 (c) Real-time
responses obtained from three 20 nm In
2
O
3
nanoribbon devices in buffer solutions with pH in the
physiological range 6.7 to 8.2 with step of 0.3. (d) Real-time responses from In
2
O
3
nanoribbon devices
at different thickness ranging from 10 to 50 nm exposed to commercial buffer solutions with pH 4 to
9.
To study the dependence of In
2
O
3
nanoribbon sensitivity on crystallinity, we
annealed 20 nm In
2
O
3
films and nanoribbon devices at 300 ºC for 30 minutes in low
vacuum condition to obtain polycrystalline In
2
O
3
. Figures 3.9 (a) and (b) show X-ray

32
diffraction (XRD) patterns of as-sputtered and annealed In
2
O
3
films, respectively. In
Figure 3.9 (a), no In
2
O
3
peak was observed from the as-sputtered film, which confirms
the amorphous nature of as-sputtered In
2
O
3
nanoribbons. In contrast, the XRD pattern of
the annealed In
2
O
3
film in Figure 3.9 (b) shows peaks for the (222), (400), (440) and
(622) In
2
O
3
planes. These multiple peaks are indicators of polycrystallinity and were also
observed by other researchers.
66
Next, we performed a pH sensing experiment to compare
the results from both the polycrystalline and the amorphous In
2
O
3
nanoribbon devices.
Figure 3.9 (c) shows average pH sensing responses from three devices of annealed and
non-annealed In
2
O
3
nanoribbon FETs. Both types of devices show comparable
performance in detecting pH from 4 to 9. From this demonstration, we believe it is
unnecessary to perform additional post-annealing to obtain polycrystalline structure for
sensing.

Figure 3.9 Plots from X-ray diffraction spectroscopy on 20 nm In
2
O
3
film (a) Before annealing (b)
After annealing in low vacuum at 300 ºC for 30 minutes. (c) Comparison of average pH sensing
responses obtained from three as-sputtered and annealed devices exposed to commercial pH buffer
solutions with pH in a range of 4 to 9.

33
3.6 Surface chemistry for In
2
O
3
nanoribbons
After confirming strong sensitivity of In
2
O
3
nanoribbon devices to charges in
fluids, we applied a phosphonic acid functional group as the linker molecule to covalently
anchor probe biomolecules to the In
2
O
3
surface for specific sensing. The same
phosphonic acid to In
2
O
3
chemistry, along with the N-(3-Dimethylaminopropyl)-N′-
ethylcarbodiimide hydrochloride (EDC) / N-Hydroxysuccinimide (NHS) coupling to
biomolecules, has been demonstrated in previous studies on In
2
O
3
nanowire-based
devices.
9, 31
Details of the surface chemistry for the phosphonic acid functionalization can
be found in the Appendix II. Figures 3.10 (a) and (b) show schematic diagrams of how
fluorescent experiments were performed using biotin and fluorescent red dye-tagged
streptavidin as the model probe-analyte system. The negative controls are anchored with
amine polyethylene glycol (PEG) as the probe instead of biotin. Figures 3.10 (c) and (d)
show the substrate schematic diagram and its optical image, respectively. We deposited
large In
2
O
3
pads on metal electrodes for better optical visibility fabricated on 500 nm
SiO
2
and 500 nm Si
3
N
4
on Si substrates. Figure 3.10 (e) and (f) show fluorescent images
taken from In
2
O
3
fabricated on SiO
2
/Si substrates for sample and its negative control,
respectively. Figure 3.10 (g) and (h) show fluorescent images taken from In
2
O
3
fabricated
on Si
3
N
4
/Si substrates for sample and its negative control, respectively. We observed
strong red fluorescent signals from In
2
O
3
pad on the metal electrodes from both samples
on SiO
2
/Si and Si
3
N
4
/Si substrates confirming that we were successful to attach biotin as
probe molecules to capture steptavidin molecules as shown in Figure 3.10 (e) and (g),
respectively. The fluorescent signal from the SiO
2
substrate in Figure 3.10 (e) shows
comparable fluorescent brightness to ones from In
2
O
3
pads while the fluorescent signal
from the Si
3
N
4
substrate in Figure 3.10 (g) has much less fluorescent signal than one from
In
2
O
3
pads. Fluorescent images of their negative controls shown in Figure 3.10 (f) and
(h), which were absence of biotin molecules, exhibited less streptavidin binding than their
sample shown in Figure 3.10 (e) and (g), respectively. These four experiments also show

34
that the phosphonic acid chemistry will react with the SiO
2
substrate while the Si
3
N
4

substrate can be used to suppress such competition binding if necessary.

Figure 3.10 (a) A schematic diagram of a fluorescent sample functionalized with amine biotin to
immobilize streptavidin conjugated with fluorescent red dyes. (b) A schematic diagram of fluorescent
a negative control functionalized with amine PEG molecules which cannot bind with streptavidin
conjugated with red fluorescent dyes. (c) A schematic diagram of In
2
O
3
pads on Au metal electrodes
for fluorescent study (d) An optical image of In
2
O
3
pads on Au electrodes for fluorescent experiment.
Fluorescent images of (e) sample and (f) negative control of In
2
O
3
pads fabricated on 500 nm SiO
2
on
the Si substrate. Fluorescent images of (g) sample and (h) negative control of In
2
O
3
pads fabricated on
500 nm Si
3
N
4
on Si substrate.
To be more quantitative, In
2
O
3
nanoribbons were functionalized with phosphonic
acid linker and followed by incubation in 1 mM amine-PEG3-biotin for the sample and
amine-PEG for the negative control. After that, both sample and negative control were
incubated in 4 nM streptavidin conjugated with 20 nm Au nanoparticles. Scanning
electron microscopic (SEM) micrographs shown in Figure 3.11 were taken after
comprehensively rinsing with 1xPBS and deioninzed water before drying in nitrogen

35
stream. Figure 3.11 (a) shows an SEM image of 2 µm wide In
2
O
3
nanoribbon
immobilized with amine-PEG3 biotin has 120 nanoparticles/µm
2
while its negative
control (Figure 3.11 (b)) attached linker molecules and amine-PEG molecules has about 2
nanoparticles/µm
2
. From this experiment, it is an evidence to assure that surface
chemistry is successfully attached amine molecules on the surface of In
2
O
3
nanoribbons.

Figure 3.11 SEM micrographs taken from In
2
O
3
nanoribbons functionalized with phosphonic acid
linker molecules and immobilized (a) amine-PEG3 biotin (b) amine-PEG on the surface of
nanoribbons before introduced 4 nM streptavidin conjugated with 20 nm Au nanoparticles.
It has been reported an investigation of coverage of phosphonic acid linker on the
ITO substrate that coverage of ITO incubated in 0.1 mM 3-phosphonopropionic acid was
saturated about 60 % after 10 hours while another ITO sample immerged in 1 mM 3-
phosphonopropionic acid reached the steady state level about 90% after 5 hour
incubation.
67
To increase more coverage of probe molecules on the In
2
O
3
nanoribbon
surface meaning to enhance its sensitivity, we have performed electro-impedance
spectroscopy (EIS) measurement on 40 nm sputtered In
2
O
3
film on Si substrate incubated
in 1 mM 3-phosphonopropionic acid for 5.5 hours and a piece of bare In
2
O
3
film on Si
substrate to verify the coverage of phosphonic acid linker on the In
2
O
3
sample. From EIS
measurement, resistance from bare In
2
O
3
film (R
BARE
) was 61.26 kΩ while resistance of
In
2
O
3
film functionalized with 1 mM 3-phosphonopropionic acid for 5.5 hours (R
5.5HR
)
was 339.32 kΩ. Therefore, the linker molecule coverage is equal to 1-R
BARE
/R
5.5HR
or

36
81.95% which is close to 90% from the previous study on ITO. For visual verification,
we have performed fluorescent test on In
2
O
3
nanoribbon devices functionalized with 1
mM 3-phosphonopropionic acid at different incubation time (2, 4, 5.5 and 7 hours) as
shown in Figure 3.12 (a) to (d). After attachment of linker molecules, the top row
nanoribbons in Figure 3.12 were immerged in 10 mM amine-PEG3 biotin in 1xPBS
while the bottom row ribbons were soaked with 10 mM amine-PEG as the negative
control. To compare among different incubation times, both samples incubated for 5.5
and 7 hours had strong fluorescent signals throughout the nanoribbon while fluorescent
signal in the channel area from samples incubated for 2 and 4 hours were weaker than 2
other samples. After inspection with the fluorescent microscope, nanoribbons were
inspected their nanoribbon thickness using the atomic force microscope (AFM).
Nanoribbons were etched by 1 mM 3-phosphonopropionic acid for 8, 12, 17 and 24 nm
after 2, 4, 5.5 and 7 hour incubation, respectively. Figure 3.12 (e) shows plot of etched
nanoribbon versus incubation; it indicates that aqueous solution of phosphonic acid linker

Figure 3.12 Fluorescent images of In
2
O
3
nanoribbons functionalized with 1 mM 3-
phosphonopropionic acid for (a) 2 hours (b) 4 hours (c) 5.5 hours (d) 7 hour. On the top row, amine-
PEG3 biotin molecules were immobilized on In
2
O
3
nanoribbons while in the bottom, row amine-PEG
molecules were anchored to In
2
O
3
nanoribbons before introducing streptavidin with fluorescent dye.
(e) Plot of thickness of etched In
2
O
3
layer versus 3-phosphonopropionic acid incubation time.

37
can etch 40 nm In
2
O
3
nanoribbon completely after 12 hours. Even though fluorescent
signals from both 5.5 and 7 hour incubation time were comparable, electrical signals
from 7 hour incubated devices showed more fluctuating than ones from 5.5 hour
incubated devices. Therefore, we conclude that 5.5 hour incubation is good enough
surface coverage (82%) of linker molecules for binding of probe molecules.
3.7 Integrated In
2
O
3
nanoribbon biosensors with electronic ELISA for
detection of streptavidin as the study model
Direct electrical detection of biomolecules in their physiological environment are
often impeded by Debye screening from the high salt concentration in the sample
solution. Sandwich ELISA
68
, on the other hand, detects signals associated with the
reactions between the substrate solution and the conjugated enzymes on secondary
antibodies instead of the biomarker. The sandwich structure not only overcomes the
Debye screening issue but also incorporates an amplification scheme to lower the signal
to noise ratio (SNR), which can be much higher for direct analyte detection without
amplification, especially when the amount of analytes are small. In the following In
2
O
3

nanoribbon sensing experiments, we applied an electronic ELISA technique that uses pH
change due to urease enzyme activity as the amplification signal. The urease enzymes are
linked to the secondary antibody through biotin and streptavidin. When a solution of urea
is introduced to nanoribbon sensor surface with the sandwich structure, the urea causes an
increase in the pH of the solution due to consumption of hydrogen ions according to the
reaction
69
:
+,-. / 21

2 / 1
3
456786
9 : : : ; 2<1

3
/ 12
=

The urease deprotonates free hydroxyl groups on the surface of In
2
O
3
nanoribbon,
and the pH increases due to the reduction of positive hydrogen ions and surface potential.
The increase in negative surface charges is responsible for the decrease in conduction of
the n-type In
2
O
3
nanoribbon FETs. The pH change is easily measured by the In
2
O
3

38
nanoribbon sensors because more charges are released during the pH increase than from
the direct binding between an analyte and a probe antibody. This allows the sensor to
detect very low concentrations of the analyte in physiological samples without the
limitation of the Debye screening effect.
69

We performed electronic ELISA sensing on streptavidin first so that the electrical
signal can be confirmed by the fluorescence signal from the binding between biotin and
streptavidin tagged with fluorescent dye. Figure 3.13 (a) shows a schematic diagram of
the sensing setup and the sequence of molecule binding. Cleaned In
2
O
3
nanoribbon
devices were functionalized with 1 mM biotinylated phosphonic acid linkers (details of
biotinylated phosphonic linker synthesis and surface functionalization are described in
the Appendix II) in methanol for 5.5 hours. Devices were then immersed in a solution of
streptavidin conjugated with red fluorescent dye for 2 hours at room temperature before
rinsing off excess streptavidin. Fluorescence confirmation of streptavidin binding is
shown in Figure 3.13 (b) and (c). In Figure 3.13 (b), the optical photograph shows one
In
2
O
3
nanoribbon device after incubating in 1 µM of dye-tagged streptavidin. Figure 3.13
(c) shows the fluorescence image taken from the same device in Figure 3.13 (b). The
strong fluorescence on the In
2
O
3
suggests that binding between biotin and streptavidin
has occurred. Furthermore, a control experiment was performed by introducing 100mM
of the urea solution to the chamber before the incubation of biotinylated urease as shown
in Figure 3.13 (d). To begin with immerging devices in 0.01x PBS pH 7.4, when we
replaced this 0.01xPBS with 100 mM urea (pH 6.61) in 0.01xPBS, conduction of
nanoribbon devices increased due to higher number of positive ions exposed to the
sensing surface than ones in 0.01x PBS. Without urease enzymes, conduction of devices
maintained about 10% higher than the base line level as in Figure 3.13 (d).

39

Figure 3.13 (a) Schematic diagram of streptavidin electronic ELISA. (b) An optical micrograph of an
In
2
O
3
nanoribbon (brown color) on Ti/Au metal electrodes after incubating with 1 µM streptavidin
conjugated with red fluorescent dyes (c) A fluorescent image of the device in (b) to confirm binding of
streptavidin molecules on an In
2
O
3
nanoribbon having biotin capture probes (d) Real-time responses
measured from three In
2
O
3
nanoribbon devices after incubating with streptavidin with fluorescent
dyes without presence of urease enzymes in the sensing chamber. Devices showed increase in
conduction due to decrease in pH from 7.4 of 0.01xPBS to 6.61 of 100 mM urea in 0.01xPBS. (e)
Normalized real-time responses of 1 µM streptavidin electronic ELISA from 3 In
2
O
3
nanoribbon
devices monitored simultaneously. Introducing urea into the sensing chamber, hydroxyl groups on the
surface of nanoribbons are deprotonated due to urea-urease enzyme reaction which consumes
hydrogen ions in the solution yielding more negatively gating effect and decrease in electrical
conduction of nanoribbon devices.

40
After attaching streptavidin, the devices were incubated in 100 µg/ml biotinylated
urease enzymes in 1xPBS for 2 hours at room temperature before rinsing off any excess
protein. The amount of urease enzymes was determined by streptavidin concentration
anchored on the nanoribbon surface. To perform the sensing, we immersed the devices in
a baseline solution of 0.01x PBS with pH 7.4. This was then replaced with 100 mM urea
in 0.01xPBS (pH = 6.61) to detect the presence of streptavidin. Figure 3.13 (e) shows the
real-time responses when the urea solution was introduced into the sensing chamber. The
urease-urea interaction drastically reduced the conduction of devices to 11.2% of the
baseline signal. The pH of the final solution in the sensing chamber was measured to be
8.68 by a commercial Mettler Toledo pH meter. The 1.28 increase in pH is consistent
with the decrease in conduction of the In
2
O
3
nanoribbon device. The decrease in pH from
the baseline increased the sensor conduction, as shown in Figure 3.13 (d), further
suggesting that the conduction decrease in Figure 3.13 (e) is indeed due to urea
interaction with the urease enzyme that is part of the sandwich system containing the
target biomarker.
To ensure that the sensing response is generated from the In
2
O
3
nanoribbon
device instead of the surrounding fluids, we performed experiments to measure the
electrical current from a device with an In
2
O
3
nanoribbon and a device without any In
2
O
3

nanoribbon immersed in 0.01xPBS as shown in Figure 3.14. Figure 3.14 (a) shows plots
of I
DS
-V
LG
from devices with and without In
2
O
3
nanoribbon. We found that the In
2
O
3

nanoribbon device exhibits good gate dependence with I
DS
about 1.3 μA at V
LG
= 1 V and
V
DS
= 200 mV. The device without any nanoribbon bridging between source and drain
electrodes does not show any gate dependence and its conduction is negligible. Figure
3.14 (b) and (c) show sampling responses of devices with and without In
2
O
3
nanoribbon.
We observed about 320 nA from an In
2
O
3
nanoribbon device and about 4.5 pA from the
device having only electrodes without any In
2
O
3
nanoribbon, or about 5 orders of
magnitude different in conduction between these two devices. From these experiments,
we can conclude that device conduction is generated from In
2
O
3
nanoribbon biosensors

41
and not contributed from conduction between source and drain metal electrodes via the
buffer solution.

Figure 3.14 (a) Plots of drain current versus liquid gate voltage (I
DS
-V
LG
) measured from devices with
and without In
2
O
3
nanoribbon in 0.01xPBS. Plots of I
DS
versus time measured from (b) an In
2
O
3

nanoribbon device and (c) a device without In
2
O
3
nanoribbon in 0.01xPBS.
After testing with 1 µM streptavidin concentration, we repeated real-time sensing
for other streptavidin concentrations, namely 100 nM, 10 nM, 1 nM, 100 pM, 10 pM, and
1 pM. Figure 3.15 (a) shows normalized steady state responses for each of the above
streptavidin concentrations. Each data point is an average of three sensors. The response
increases exponentially with increasing streptavidin concentration. The change in pH
between the final urea solution in the sensing chamber and baseline 0.01XPBS buffer is
plotted against streptavidin concentration as shown in Figure 3.15 (b). This relationship
also shows the same exponential trend. With the pH amplification scheme enabled by the

42
electronic ELISA, our nanoribbon biosensors have detected a streptavidin concentration
that is 4 orders of magnitude lower than the 1 nM reported for In
2
O
3
nanowire sensors
with the same magnitude of the sensing response (about 2%).
31

Figure 3.15 (a) A plot of average normalized current responses and streptavidin concentration
calculated from 3 devices monitored simultaneously in each concentration. (b) A plot of pH changes
in the sensing chamber measured by a commercial pH meter and streptavidin concentration.
3.8 Detection of HIV1 p24 proteins by electronic ELISA on In
2
O
3

nanoribbon biosensors
The ultra low limit of detection demonstrated by our nanoribbon platform is
advantageous for detecting biomarkers like the HIV p24 protein, whose presence even at
an extreme low level can indicate early stage of HIV infection. To perform electronic
ELISA detection for the p24 protein, we functionalized our devices with the phosphonic
linker molecules and use EDC/NHS chemistry to anchor the HIV1 p24 antibodies on the
surface of nanoribbons, as mentioned in the Method Section. Known concentrations of
the p24 proteins in 1xPBS buffer were then introduced to the sensor for antigen-antibody
binding. Next, the secondary biotinylated HIV1-p24 antibodies were anchored on the
captured proteins by incubation at room temperature for 4 hours before rinsing
extensively with 1xPBS. After that, devices were incubated in 100 pM streptavidins in
1xPBS to provide binding sites for 0.1 mg/ml biotinylated urease enzymes, which is the

43
last step before the sensing. Urease enzymes can also be directly linked to the secondary
antibody to reduce the number of binding steps. Figure 3.16 (a) shows the schematic
diagram for the above sequence of molecules in the electronic ELISA for p24 protein
detection. Figure 3.16 (b) shows normalized real-time electronic-ELISA sensing
responses to HIV1 p24 at 20 fg/ml monitored from three In
2
O
3
nanoribbon devices
simultaneously. The conduction of the devices was reduced by about 35% when they
were exposed to 100 mM urea solution because of the increase in the pH of solution in
the sensing chamber that was induced by reaction between immobilized urease enzymes
and urea solution. Figure 3.16 (c) shows a plot of average normalized responses from
three devices monitored simultaneously at each different p24 protein concentration from
20 fg/ml to 20 pg/ml. Responses from electronic-ELISA show exponential relation with
p24 concentration as shown in Figure 3.16 (c). Figure 3.16 (d) shows a plot of pH
changes in the sensing chamber before and after sensing measured by a pH meter and p24
concentration. It also shows an exponential relationship, which agrees with the trend of
changes in conduction in Figure 3.16 (c).
In addition, we have spiked known concentrations of p24 proteins in human
serum as the target analytes in several electronic ELISA experiments in order to
demonstrate our device capability to selectively perform sensing in the physiological
solutions. We observed similar changes in electrical conduction and in the pH of the
sensing solution to what we had obtained from p24 sensing in PBS, despite the fact that
blood serum is composed of numerous competing proteins such as human serum
albumins and human serum globulins.
70
The PBS sensing data are shown as black
rectangles, and human serum data are shown as red triangles in Figures 3.16 (c) and (d),
respectively. The matching of the buffer and the serum data is a good indicator that the
signals from both media are attributed to mainly the p24 proteins and not the competing
proteins in the physiological fluid. These results serve as a good evidence for the
selectivity of our sensors in the complex media because our devices can selectively detect
p24 proteins in human serum. As a result, we can use this electronic ELISA approach in

44
different kinds of physiological solutions without complicated sample preparation steps,
as competing proteins/biomolecules in the fluids are washed off, leaving only target
analytes immobilized by capture probes. From our approach, we could detect HIV-1 p24
proteins about 3 orders of magnitude lower than limit of detection of the conventional
ELISA approach.
58,57

Figure 3.16 (a) Schematic diagram of electronic ELISA for HIV1 p24 detection (b) Real-time
responses monitored from 3 In
2
O
3
nanoribbon devices simultaneously at 20 fg/ml of p24 proteins in
PBS. Conduction of all devices decreased upon presence of urea in the sensing chamber. (c) A plot of
average normalized responses from 3 devices at each p24 concentration and p24 concentration in
pg/ml. (d) A plot of change in pH in the sensing chamber measured from a commercial pH meter and
p24 protein concentration in pg/ml.

45
3.9 Summary
In conclusion, we have demonstrated a top-down approach for In
2
O
3
nanoribbon
FET fabrication using two photolithographic masks to define the position and the
dimensions of the metal electrodes and the nanoribbons. Devices fabricated using this
approach show uniform and good electrical performance without requiring doping or
post-process annealing. The fabrication is highly scalable, low cost, and a low
temperature process that is compatible with the CMOS fabrication facilities. In
2
O
3
is
selected for the nanoribbon material because its electrical performance and long-term
stability in aqueous solution are better than other metal oxide materials. In
2
O
3
nanoribbon
devices exhibited good sensitivity in both wide range of pH solution from pH 4 to 9 and
physiological range between 6.7 and 8.2. Streptavidin-biotin has been chosen to
demonstrate signal amplifying electronic ELISA with picomolar sensitivity showing 15%
changes in normalized current. We demonstrated electronic ELISA for detection of HIV
p24 proteins at concentration about 20 fg/ml or 250 viruses/ml, which is about 3 orders of
magnitude lower than commercial ELISA kit on the market. Depending on choice of
capture probes, our uniform, scalable, sensitive, top-down In
2
O
3
nanoribbon biosensor
platform integrated with the electronic ELISA technique can be utilized for diagnosis of
other infectious diseases and cancers. We believe that our In
2
O
3
nanoribbon platform can
be applied to other biological and medical applications.

46
Chapter 4 Highly Scalable, Uniform and Sensitive In
2
O
3

Nanoribbon Biosensors for Myocardial Infarction
4.1 Introduction
The requirements of urgent care diagnosis system are real-time, sensitive, specific
and multiplex responses to several acute and fatal diseases which may need immediate
medical attention. The medical personnel needs to quick, correct and comprehensive
information in order to identify what kind of disease the patient has, to evaluate how
severity condition of the patient is, and to provide proper treatment for that patient.
Myocardial infarction is one of major emergency diseases in the United States and
around the globe. Every year over a million of American visit the emergency department
due to their chest pain or other symptoms which are sign of acute coronary syndrome
(ACS), but only 10% of these patients experience acute myocardial infarction (AMI).
71
In
medical setting, diagnosis of myocardial infarction relies on electrocardiogram (ECG) to
record electrical activities of heart via electrodes attached on the skin, blood test for
detection of cardiac biomarkers using enzyme-linked immunosorbent assay (ELISA) or
monitoring blood flow using dye or radioactive substances combined with the imaging
technique. ECG can identify the damaged area due to the abnormality of electrical signals
for quick assessment in the emergency department. However, ECG has significant
limitations that it is unable to quantify severity of heart damage or it shows entirely
normal electrical patterns on 50 % of ACS patients.
72
Therefore, patients are usually
diagnosed to be AMI based on 2 of 3 indicators: typical symptoms of AMI, Q wave
patterns from ECG and blood test results from ELISA.
71
Troponin, a Food and Drug
Administrative (FDA) approved biomarker for AMI, is a complex regulatory protein
found in cardiac muscle and has 3 isoforms: C (the calcium binding component), T (the
tropomyosin binding component) and I (the inhibitory component). Troponin I and T
have sufficient identity to allow production of their antibodies. Troponin I and T are
released to blood stream due death of cardiac muscle cells; therefore, troponin I and T are

47
not in the blood of healthy people.
71
Levels of both troponin isoforms increase within 4 to
9 hours and peak about 12 to 24 hours after AMI.
71
In addition to troponin, detection of
creatine kinase-MB (CK-MB), an isoenzyme found in heart muscle, can improve early
diagnosis of AMI.
73
Level of CK-MB increases within 2 to 4 hours after cardiac muscle
injury.
74
However, detection of only CK-MB biomarker can lead to false positive for
diagnosis of AMI because CK-MB can be released to blood stream due to muscle
disease, congestive heart failure, or pulmonary edema.
71
Another emerging biomarker for
diagnosis of AMI is b-type natriuretic peptide (BNP) which normally uses to diagnose
congestive heart failure.
75-78
BNP is secreted from cardiac ventricles because of change in
pressure in heart chamber due to heart failure. To incorporate these biomarkers as a panel
for AMI helps to improve specificity, to reduce false interpretation and to identify
severity of the patient's condition. However, detection of cardiac biomarkers using the
conventional ELISA approach requires costly instruments and well-trained staffs to
conduct the test and needs to be performed in the well-equipped medical setting. In
addition, limit of detection for commercial gold standard troponin ELISA kits is 6 pg/ml
which is undesirable for early prognosis for AMI.
79
Several approaches have been
proposed by many researchers to ease the blood test for diagnosis of AMI and try to
lower detectable of key biomarkers for early detection of AMI. Bottom-up polyaniline
nanowire biosensors have been demonstrated to detect myoglobin, troponin I, CK-MB
and BNP in diluted phosphate buffer saline (PBS).
80
Top-down silicon nanowire
biosensors were utilized to detect troponin T
81-83
, troponin I
79
, CK-MM
83
, and CK-MB.
83

However, these approaches need to deal with either complicated sample preparation
process
81, 82
to lower salt concentration to increase Debye screening length or data
processing to extract change in resistivity.
83

The top-down In
2
O
3
nanoribbon field effect transistor (FET) biosensor have been
demonstrated for high degree of scalability and uniformity due to its compatibility with
the microelectronic fabrication process in the previous chapter. In addition, its fabrication
process requires only 2 photolithographic masks to define dimension and position of

48
metal electrodes and nanoribbons. Moreover, the fabrication process of In
2
O
3
nanoribbon
biosensors is the low temperature process which enable to use other low cost substrates
instead of silicon substrates. Benefits of the In
2
O
3
nanoribbon are that In
2
O
3
is inherently
semiconducting material and its electrical property relies on oxygen vacancies inside
nanoribbon without requirement of any toxic gas as the dopant. Furthermore, In
2
O
3

nanoribbon devices show good long-term stability in PBS, which is beneficial for in-vivo
and in-vitro applications. Combined with electronic ELISA, In
2
O
3
nanoribbon biosensors
have projected sub fg/ml sensitivity for detection of p24 proteins in physiological
solution without complication of sample preparation.
From all mentioned advantages of our In
2
O
3
nanoribbon platform, we
demonstrate detection of a panel of cardiac biomarkers, namely troponin I, CK-MB and
BNP on In
2
O
3
nanoribbon biosensors with use of electronic ELISA technique to ease
detection in physiological solution. We fabricate the In
2
O
3
nanoribbon biosensor chip
which can accommodate multiplex detection of four biomarkers with integration of on-
chip gate electrodes using 2 photolithographic masks. We have reduced the assay time of
our approach from 10 hour to be less than a hour without loss of significant sensitivity.
We perform experiments for each biomarker in PBS to create calibration curves. Lastly,
we demonstrate a blind test using spiked BNP protein in human blood and average of
sensing response from this test fits with the calibration curve within the error range.
4.2 In
2
O
3
nanoribbon biosensor chip fabrication
We use simple two photolithographic masks for fabrication of In
2
O
3
nanoribbon
biosensor as mentioned in section 3.2 in Chapter 3 to define position and dimension of
metal electrodes and In
2
O
3
nanoribbons. Biosensor chips were fabricated on the 500 nm
silicon dioxide (SiO
2
) on Si substrate. After photolithography of the first mask to define
dimension and position of metal electrodes, three layers of 5 nm Ti, 50 nm Au and 5 nm
Ti electrodes were deposited by an electron beam evaporator. The second mask was used
to define dimension and position of In
2
O
3
nanoribbons before the scalable radio

49
frequency (RF) sputtering of In
2
O
3
nanoribbon with targeted thickness about 40 nm to
accommodate etching of nanoribbons from surface chemistry as mentioned in section 3.6
in Chapter 3. Figure 4.1 (a) shows a wafer-scale photograph of In
2
O
3
nanoribbon
biosensor chips fabricated on a 3" SiO
2
on Si substrates with 100% functional device
yield. Figure 4.1 (b) and (c) show photographs of an In
2
O
3
nanoribbon biosensor chip
composing of 4 groups of 5 nanoribbon devices and a gate electrode and an biosensor
chip attached with a four-well polydimethylsiloxane (PDMS) stamp to confine solutions
during the sensing experiment, respectively. Figure 4.1 (d) shows a magnified image of a
group of devices on the biosensor chip. Figure 4.1 (e) shows an image of two identical
In
2
O
3
nanoribbons as an evidence of device uniformity from the top-down fabrication
approach. The PDMS stamp was fabricated by casting on mixture of curing agent and
PDMS at ratio of 1:10 on a tapper-shaped Teflon mold before thermally cure at 80°C for
30 minutes. After microwell fabrication, the PDMS stamp can be attach with the
biosensor chip using either adhesive or mechanical clamp.

Figure 4.1 (a) An optical image of 5 fabricated In
2
O
3
nanoribbon biosensor chips of a 3 inch 500 nm
SiO
2
on Si wafer (b) A magnified image of a biosensor chip which consists of 4 subgroups for
multiplex detection . (c) An photograph of a PDMS microwell stamp on top of a biosensor chip for
sensing experiment. (d) A magnified photo of one subgroup in a chip comprising 5 devices and a gate
electrode at the 3rd position. (d) An image of two identical nanoribbon in a subgroup.

50
4.3 Electrical Characteristic performance of the In
2
O
3
nanoribbon
biosensor chip with on-chip gate electrodes
After fabrication, devices were characterized by an Agilent semiconductor
analyzer B1500A using liquid gating to ensure that we have functional devices and good
electrical performance. To perform biosensing experiment, gate voltage is applied
through a liquid gate electrode to tune the operational point of sensors to maximize their
sensitivity. The Ag/AgCl electrode is commonly used as the reference electrode in the
electrochemical experiment and biosensing experiment because it provides stable
potential due to separation between electrode and analyte solution by a glass frit.
26

However, integration of the Ag/AgCl electrode into the biosensor chip remains
challenging.
84, 85
We have compared between performance of the external Ag/AgCl
Figure 4.2 (a) Schematic diagram of electrochemical measurement to test effectiveness of on-chip
gate. Plots of I
DS
-V
GS
curves applied gate voltage through (b) a Ag/AgCl electrode and (c) an on-chip
Ti/Au/Ti gate electrode on the same device. (d) A plot of liquid gate potential (V
REF
) and drain current
(I
DS
).

51
electrode and the on-chip Ti/Au/Ti electrode as shown in Figure 4.2. Figure 4.2 (a) shows
a schematic diagram of our electrochemical experiment. We have set our source of the
nanoribbon device as the working electrode and the gate electrode as the counter
electrode. We used another Ag/AgCl electrode as the reference electrode to monitor
change in potential (V
REF
) in the electrochemical Teflon cell. Figure 4.2 (b) shows a plot
of I
DS
-V
GS
measured when gate voltage was biased through a Ag/AgCl electrode. It
shows that V
GS
is similar to V
REF
. While V
GS
applied through an on-chip gate electrode
shows weaker potential (V
REF
) than one applied through the Ag/AgCl electrode as shown
in Figure 4.2 (c). However, when we plotted drain current and liquid potential from both
types of electrodes. They show similar gating effect to the device as shown in Figure 4.2
(d). From results in Figure 4.2, the on-chip gate is a promising solution for integration of
the gate electrode on the biosensor chip and its performance for gating the transistor is
similar to one from the Ag/AgCl electrode.
After comparison gating performance from the on-chip gate and Ag/AgCl
electrodes, we have conducted a statistical study of key performance indices from 20
In
2
O
3
nanoribbon devices on a biosensor chip measured in 0.01xPBS and biased through
the on-chip gate. Figure 4.3 (a) shows that the average of the on-state current(I
ON
)
measured at drain voltage (V
DS
) = 200 mV and liquid gate voltage (V
LG
) is 618.95 nA
with the standard variation of 80.22 nA. Transconductance (g
m
) of these 20 transistors
fall in a range between 2 to 4 µS with an average value of 2.71 µS and their standard
variation of 0.60 µS as shown in Figure 4.3 (b). Figure 4.3 (c) shows that standard
variation in threshold voltage (V
TH
) of these devices is small about 0.06 V or 3.48 % with
the average at 1.72 V. The last figure of merit is on-state to off-state current ratio as
shown in Figure 4.3 (d). On/off ratios of these transistors are in a range of 10
4
to 10
6
.
From these figures of merit, we can conclude that the on-chip gate electrode has a good
control over nanoribbon transistors in the aqueous environment with a compact form
factor.

52

Figure 4.3 Plots of key electrical performance from 20 In
2
O
3
nanoribbon devices using on-chip gate
electrodes (a) On-state current (I
ON
) at V
DS
= 200 mV and V
LG
= 2 V (b) Transconductance (g
m
) (c)
Threshold voltage (V
TH
) (d) On-state to off-state current ratios at V
DS
= 200 mV.
After testing performance of the on-chip gate electrode over nanoribbon
transistors in buffer solution, we performed a pH sensing experiment using the on-chip
gate electrode instead of a Ag/AgCl electrode to test ionic sensitivity of the biosensor
chip using commercial pH solutions. Figure 4.4 (a) shows a schematic diagram of the pH
sensing experiment. We mechanically attached an electrochemical cell on a subgroup of a
biosensor chip and manually exchange solution in and out of the sensing chamber using
pipettes. Figure 4.4 (b) shows exponential decay relationship between average
normalized responses from three as-sputtered sensors monitored simultaneously and pH
in the sensing chamber. Devices show decrease in conduction when pH in the sensing
chamber increase because hydroxyl groups on the nanoribbon surface are deprotonated

53
due to more negative OH ions in the solution resulting in the negatively gating effect on
the channel area of the n-type In
2
O
3
nanoribbon transistor. Our In
2
O
3
nanoribbon sensors
combined with use of the on-chip gate shows 45 times change in responses when pH in
the sensing chamber increases from pH 4 to pH 9. From the pH sensing experiment, we
have more confidence that the on-chip gate electrode can be utilized in the biosensing
application.

Figure 4.4 (a) A schematic diagram of pH sensing experiment on an In
2
O
3
nanoribbon biosensor
chip. Gate voltage was applied through the Ti/Au/Ti on-chip gate electrode. A Teflon electrochemical
cell was mounted on a biosensor chip to confine commercial pH solution during the experiment. (b)
Average normalized real-time sensing responses measured from 3 nanoribbon devices exposed to
commercial pH solutions from pH 4 to pH 9.
4.4 Electronic enzyme-linked immunosorbent assay for diagnosis of
myocardial infarction
After device characterization to ensure that devices has good electrical
performance and ionic sensitivity, the biosensor chip was functionalized with 1 mM 3-
Phosphonopropionic acid for 5.5 hours to anchor linker with carboxylic functional groups
on the surface of In
2
O
3
nanoribbons according to methods mentioned in Appendix II.3.
The antibody with amine terminal group was coupled to the linker molecule through the
N - (3 - Dimethylaminopropyl) - N'- ethylcarbodiimide hydrochloride (EDC) / N -

54
Hydroxysuccinimide (NHS) chemistry. Figure 4.5 (a) shows a schematic diagram of
electronic ELISA for detection of troponin I. After immobilization of capture antibodies
on the sensor surface, devices were incubated in target molecules for 2 hours at room
temperature before rinsing off excess proteins. Next step is that devices were soaked in
solution of 100 μg/ml biotinylated secondary antibody for 4 hours before rinsing the
unbound antibody. After that, 100 pM streptavidin was introduced into the sensing
chamber to provide binding sites for biotinylated urease enzymes. Devices were
incubated in streptavidin solution for 2 hours at room temperature before rinsing off
excess proteins. Figure 4.5 (b) shows normalized real-time responses from three
nanoribbon devices after immobilization of streptavidin as a control experiment for
troponin I detection. In the beginning, we immersed devices in 0.01xPBS with pH 7.4
and replaced 0.01xPBS in the sensing chamber with 10 mM Urea in 0.01xPBS with pH
6.59. Conduction of devices increases due to increase in pH in the sensing chamber
which yields positively gating on n-type nanoribbon transistors as shown in Figure 4.5
(b). After the control experiment, devices were rinsed with 1xPBS before they were
incubated in 100 µg/ml biotinylated urease enzyme for 2 hours at room temperature to
immobilize urease enzymes on the sensor surface. After 2 hour incubation, sensors were
washed with 1xPBS to remove unbound enzymes from the surface before the electronic
ELISA experiment. The number of bound enzyme on the surface of nanoribbon varies
with concentration of captured target molecules. To perform the electronic ELISA, we
firstly immersed devices in 0.01xPBS with pH 7.4 before we exchanged from 0.01xPBS
to 10 mM Urea in 0.01xPBS. Figure 4.5 (c) shows normalized real-time sensing
responses from three nanoribbon sensors monitored simultaneously for detection of
troponin I protein at 300 pg/ml concentration. Average conduction of all three sensors
reduced about 88.47% from the base line current due to hydrolysis reaction between urea
and urease enzymes which causes increase in pH in the sensing chamber from 7.4 to 8.85
measured by a commercial pH meter. The hydrolysis reaction deprotonates hydrogen ions
out from hydroxyl groups on the surface of nanoribbons resulting in negatively gating
effect for n-type In
2
O
3
nanoribbon FETs. After detection of troponin I at 300 pg/ml, we

55

Figure 4.5 (a) A schematic diagram of electronic ELISA for troponin I detection (b) Normalized real-
time responses monitored from three In
2
O
3
nanoribbon sensors simultaneously at 300 pg/ml of
troponin I proteins in 1xPBS without presence of urease enzymes in the sensing chamber. Conduction
of all sensors increased due to decrease in pH of urea solution in the chamber. (c) A plot of real-time
responses from the same devices in (b) after immobilization of biotinylated urease on nanoribbons.
Conduction of all three devices decrease upon increase in pH in the sensing chamber due to reaction
between urea and urease enzymes. (d) A plot of average responses from three devices at each troponin
I concentration in pg/ml.
repeated similar sensing experiment for other concentrations, namely 10 pg/ml, 1 pg/ml
and 100 fg/ml. Figure 4.5 (d) shows average normalized responses for each troponin I
concentration mentioned earlier. Each data point was calculated from three sensors
monitored concurrently during the experiment. The sensing response decreases
exponentially upon decrease in concentration of troponin I target molecules. However,

56
the sensing response seems to saturate at 300 pg/ml or higher concentration. For patients
with acute myocardial infarction, the level of troponin I is elevated due to release of
troponin proteins from death of cardiac myocytes. The clinically relevant level for
troponin I is about 281 pg/ml.
86
However, the early diagnosis of troponin is critical to
reduce rate of mortality because the physician can provide the proper treatment to that
patient. Any detectable level is a sign of myocardial infarction and severity depends on
concentration of troponin protein because troponin is not present in healthy people.
71

To move our sensing platform toward practical use in diagnosis of myocardial
infarction, total assay time needs to be optimized and reduced from about 10 hours in the
previous experiments to match with medical need for the urgent care application. We
have conducted the series of experiments for optimization of assay time. We
functionalized our sensors with linker molecules and capture antibodies on the surface of
nanoribbons with similar procedures because these preparation steps can be completed in
advance. We have reduced incubation time of target analytes, biotinylated secondary
antibodies, streptavidin and biotinylated urease enzymes from 2, 4, 2 and 2 hours to be
only 10 minutes for each step, respectively. The total amount of time is reduced from 10
hours to about 40 minutes after the sensor preparation. Figure 4.6 (a) shows normalized
sensing responses from 3 sensors for detection of troponin I at concentration of 300
pg/ml. Devices were firstly submerged in 0.01x PBS before we replaced it with 10 mM
Urea. We observed reduction of current from the base line to be about 18.02 % of the
base line current or a change of 81.97%. We performed similar experiments for the short
assay time for other concentration, namely 10 pg/ml and 1 pg/ml. Figure 4.6 (b) shows
average sensing responses from three devices monitored simultaneously for each
concentrations. Responses show exponential correlation with troponin I concentration as
observed previously in Figure 4.5 (d). Figure 4.6 (c) show comparison between average
responses from long assay time and short assay at different troponin I concentrations.

57

Figure 4.6 Electronic ELISA for detection of troponin I with short assay time (a) Real-time sensing
responses of three In
2
O
3
nanoribbon devices monitored concurrently at troponin I concentration of 300
pg/ml in PBS. Normalized conduction of nanoribbon devices decreased from the base line due to
increase in pH in the sensing chamber caused by hydrolysis of urea from reaction with urease
enzymes. (b) A plot of changes in average normalized responses for three devices at each troponin I.
(c) Comparison between average responses of long assay time and short assay times.
Responses from the short assay time are lower than ones from the long assay time in the
range of 2.08% - 14.77 %. These discrepancies may generate from error of pipettes,
which have accuracy range about 3% to 0.5% and precision range about 1.5% to 0.2%,
during dilution of the protein from the original concentration. Another important factor is
binding kinetic of each protein because we assume that binding of each protein reaches
equilibrium for the long assay time, but this assumption might not be valid for the short
assay time. Further investigation about binding kinetic of each protein needs to be

58
optimized using Surface Plasmon Resonance (SPR) measurement for the optimal
sensitivity and the minimum assay time. With equation from its fitting curved in Figure
4.6 (b), the projected limit of detection with 1% response is 15 ag/ml which is about 5
orders of magnitude more sensitive than the commercial ELISA.
79

After detection of troponin I protein, we performed the similar short assay time
protocol for other cardiac biomarkers which are CK-MB and BNP for diagnosis of
myocardial infarction. Figure 4.7 (a) shows a schematic diagram for electronic ELISA to
detect CK-MB biomarker. We performed similar protocols to troponin electronic ELISA
sensing. Figure 4.7 (b) is a plot of average responses from three devices monitored
simultaneously at each CK-MB concentration in 1xPBS. The CK-MB sensing response
shows linear relationship with a narrow range of the analyte concentration from 1 to 5
ng/ml which covers clinically relevant threshold at 3 ng/ml.
86
We notice that CK-MB
sensing has lower responses that ones from troponin I sensing which tends to saturate at
the level higher than 300 pg/ml. The cause which lowers CK-MB sensing responses may
be lower degree of biotinylation for the secondary CK-MB antibody than the secondary
biotinylated antibody of the other biomarker; therefore, some of labeling CK-MB
antibody may not have biotin, which will reduce number of immobilized urease enzymes
on nanoribbons and the sensing response. From the CK-MB, our assay can be utilized to
segregate among patients with AMI or non coronary chest pain around the threshold.
After we completed the experiment for a calibration curve for the CK-MB biomarker, we
performed a series of experiments for detection of BNP which is a heart failure biomarker
and an emerging AMI biomarker. Figure 4.7 (c) shows a schematic diagram for detection
of BNP using In
2
O
3
nanoribbon biosensors. Figure 4.7 (d) shows a plot of average
responses from three sensors for each data point and BNP concentration in PBS (black
dot). When the BNP concentration decreases, its sensing signal also reduces
exponentially. We have tried to simulate a use of our sensors in human whole blood. We
spiked BNP protein with targeted BNP concentration about 500 pg/ml in human whole
blood before dilution about 100 times with 1xPBS to prevent saturation of the sensing

59
response. After that, we immersed sensors in this diluted solution of BNP in human
whole blood for 10 minutes and rinsed off unbound proteins with 1xPBS. We performed
similarly for other steps, which are binding of biotinylated secondary BNP antibody,
streptavidin and biotinylated urease enzyme. Average response from sensing experiment
for detection of 5 pg/ml BNP in whole blood is close to our BNP calibration curve with
difference of 3.03% from the calibration curve. If we calculate concentration from the
sensing response using the equation of the calibration curve, we predict the concentration
to be 2.83 pg/ml after 100 times dilution. This discrepancy may cause by accumulated

Figure 4.7 (a) A schematic diagram of electronic ELISA for detection of CK-MB biomarkers (b) A
plot of average normalized responses from three devices at each CK-MB concentration. (c) A
schematic diagram of electronic ELISA for detection of BNP biomarkers (b) A plot of average
normalized responses from three devices at each BNP concentration with a blind test of BNP in
human whole blood.

60
error in precision of the pipette resulting in possible concentration to be 5.27-4.54 pg/ml
and a small device to device variation. For the BNP assay, we can run two parallel tests
which are undiluted sample and 100 time diluted sample in order to fulfill the diagnostic
range of BNP biomarker which is 100 pg/ml for the clinically relevant threshold for heart
failure.
87, 88
If response from the undiluted sample shows saturated response at 90 % or
greater, we have to check the result from the assay used the diluted sample to quantify
concentration of BNP in the patient 's sample to identify the cause of heart discomfort.
However, if the test result from the undiluted sample shows a response less than 90%, it
means that patients is less like to have congestive heart failure due to the result from a
study suggesting that patients with than BNP level less than 50 pg/ml are less likely to
have congestive heart failure.
88
From the result of this blind test, we should be able to
apply all calibration curves for a panel of biomarkers for diagnosis of acute myocardial
infarction. However, statistical study needs to be perform with patients' samples to
investigate in accuracy and specificity of the assay including to quantify number of false
positive and false negative to move forward to the practical use of this assay.
4.5 Summary
In summary, we have demonstrated fabrication of highly scalable top-down In
2
O
3

nanoribbon biosensor chips with integration of on-chip gate using 2 simple
photolithographic masks to define position and dimension of metal electrodes and
nanoribbons. The Ti/Au/Ti on-chip gate electrode shows comparable gating performance
to the standard Ag/AgCl electrode at the same liquid gate potential. Furthermore, In
2
O
3

nanoribbon devices show uniform electrical performance in the aqueous condition when
gate voltage is applied through the on-chip gate electrode. In addition, In
2
O
3
nanoribbon
devices with the on-chip gate electrode shows good performance in pH sensing
experiment with 45 times change in conduction when pH is reduced from 4 to 9. We
demonstrated electronic ELISA detection of troponin I at 100 fg/ml with 30 % response
which is about 2 orders of magnitude lower than commercial ELISA kits. We have
reduced the assay time from 10 hours to 1 hour to move forward to the practical use

61
without significant loss of sensitivity. After demonstration detection of troponin I, we
conducted experiments to complete calibration curves for other selected cardiac
biomarkers for diagnosis of myocardial infarction, CK-MB and BNP. We covered all
clinically important ranges of all biomarkers to enable a practical use of our assay for
diagnosis and prognosis of myocardial infarction. Lastly, we demonstrated a test using
spiked BNP in diluted human blood. Our assay showed 3% less than the value from the
calibration curve which is within error range of the dilution error. To be used for the
point of care application, further optimization of assay time and more statistical study
with patients' samples or spiked samples need to be investigated.

62
Chapter 5 Epitaxial Growth of Aligned SnO
2
Nanowires on
Sapphire and Their Device Applications
5.1 Introduction
Research on aligned semiconducting nanowires has gained significant momentum
in recent years due to their importance in building scalable and dimensionally
controllable bottom-up devices for various applications, ranging from diodes and
transistors to chemical and biological sensors. Many methods have been developed to
manipulate these nanowires post-growth, including the Langmuir-Blodgett
compression,
89-91
pattern transfer,
92
mechanical shear,
93
fluid flow in microchannels,
94

and orientation by an electric field.
95
With a recently reported nanocombing method,
misalignment of nanowires can be controlled within ±1° with small cross defect density
of 0.04 nanowires per µm.
2
On the other hand, guiding the direction of nanowire growth
by using the epitaxial relation between the nanowire and the substrate combines synthesis
and assembly into one single step. This also provides control of nanowire
crystallographic orientation and direct fabrication of nanowire devices on growth
substrates without the need for a transfer process. Substrate guided growth of
semiconducting nanowires has been successfully demonstrated for GaN nanowires on
sapphire,
96
ZnO nanowires on sapphire,
97, 98
InAs nanowires on InAs,
99
and InP
nanowires on InP.
100
In this work, we show that synthesized SnO
2
nanowires not only
demonstrate guided growth on R-plane and annealed A- and M-plane sapphire, but also
provide good assembly for easy integration in various electronic applications.
The n-type semiconductor SnO
2
is a direct, wide bandgap (3.6 eV)
101
material
with high electrical conductivity, optical transparency, and sensitivity to adsorbed
molecules. Successful applications of SnO
2
for field-effect transistors,
102, 103
transparent
devices,
104
and gas sensors
105-108
have already been realized. SnO
2
nanowires can be
especially advantageous for these applications reliant on their dimensional compatibility
with nanoelectronics and their high surface to volume ratio that is important for

63
sensitivity. Here we show that aligned, planar SnO
2
nanowires can be used to achieve
better control of the transistor channel orientation, which creates an advantage for various
sensing applications. In addition, SnO
2
nanowires have a rutile structure not exhibited in
nanostructures in previous substrate guided growth studies, and their synthesis deserves
further investigation.
5.2 Aligned SnO
2
nanowire synthesis and characterization
We chose sapphire as the growth substrates for the aligned SnO
2
nanowires
because of its widely demonstrated ability to guide nanowire
96-98
and nanotube
109, 110

growth. We first investigated the effect of annealing A (112 ‾ 0), M (101 ‾ 0), and R (11 ‾ 02)
plane sapphire on optimizing the growth conditions for the rutile structured SnO
2

nanowires. The orientations for each of the three sapphire planes are shown in Figures 5.1
(a-c) along with the orientation of the wafer flat, which is indicated as the bold edge of
the outlined wafer in each plane. High temperature annealing of each sapphire plane was
done in ambient air, and atomic force microscopy (AFM) images were examined before
and after annealing. Pristine substrates of A plane (Figure 5.1 (d)), M plane (Figure 5.1
(f)), and R plane (Figure 5.1 (h)) all show planar surfaces. After annealing, however, we
observed clear V-shaped nanogroove structures in the M-plane sapphire that are
approximately 15 nm deep, shown as Figure 5.1 (g). The presence of the grooves is
consistent with observations that were used to explain the mechanism of graphoepitaxial
growth of aligned ZnO and GaN nanowire,
96, 98
and the same mechanism can be applied
to parallel SnO
2
nanowires on annealed M plane as well. It is due to the fact that the M-
plane sapphire is thermodynamically unstable during high temperature annealing. Its
crystal facet will be transformed from (101 ‾ 0) to nanostructure grooves composed of S-
plane and R-plane facets along [112
>
1] direction.
98
A- and R- plane sapphire, on the other
hand, retained their planar surface structure after annealing, as can be seen in Figures 5.1
(e) and 5.1 (i), respectively . The lack of surface features on the A-plane and R-plane

64
sapphire suggests that the alignment of SnO
2
nanowires on these planes are aligned by
lattice guided growth instead, in agreement with the mechanism of aligned ZnO growth.
98

Figure 5.1 Aligned SnO
2
nanowire growth study. (a-c) Orientation of A-plane (a), M-plane (b), and
R-plane (c) sapphire. Wafers are outlined on each plane with the wafer flat bolded. (d-i) AFM scans of
A(d, e), M(f, g), and R(h, i) sapphire surfaces before(d, f, h) and after(e, g, i) annealing. Scale bars are
30 nm. (j-l) SEM images of aligned SnO
2
nanowires grown on A-plane (j), M-plane (k), and R-plane
(l) sapphires. Sapphire orientations are included on the bottom left.
Before synthesis, catalysts were first deposited onto both annealed and non-
annealed sapphire substrates in the shape of Au stripes using standard photolithography
and electron beam evaporation. The guided SnO
2
nanowires were grown using a vapor-
liquid-solid (VLS) process in a low pressure chemical vapor deposition (CVD) system.
Details of the synthesis can be found in the Supporting Information. After synthesis,
aligned SnO
2
nanowire growth was confirmed and characterized by scanning electron
microscopy (SEM) imaging. We first observed that substrate annealing affected SnO
2

nanowire growth differently, depending on the plane orientation of the sapphire substrate.
On un-annealed M-plane sapphire, SnO
2
nanowires grew in two perpendicular directions,

65
crossing each other such that the length between each junction is under 5 µm. Parallel
SnO
2
nanowires on the M plane were observed only on annealed M planes. SnO
2

nanowires grown on un-annealed A-plane sapphire had short lengths under a few
microns, and poor parallel alignment. On R-plane sapphires, however, no significant
difference between SnO
2
nanowire growth on annealed and on un-annealed substrates
were found. Therefore, the following experiments and discussions were carried out with
aligned SnO
2
nanowires grown on annealed A-plane, annealed M-plane, and un-annealed
R-plane sapphires.
On the annealed A-plane sapphire, SnO
2
nanowires grew parallel to the [11 ‾ 00]
direction, or perpendicular to the direction of nanowires grown the annealed M-plane
sapphire shown in Figure 5.1 (j). Conversely, Figure 5.1 (k) shows that SnO
2
nanowires
aligned on the annealed M-plane sapphire grew parallel to the [112 ‾ 0] direction, which is
perpendicular to the A plane. Figure 5.1 (l) shows SnO
2
nanowires synthesized on the R-
plane sapphire with a clear alignment in the [1 ‾ 10 1] direction. The SEM images show the
nanowire lengths on all planes range from 10µm to 100 µm, with a large percentage of
wires between 50 µm and 100 µm in length. The long nanowires allow for simple
patterning of source, drain and gate electrodes using standard photolithography
techniques. Further characterization using SEM and AFM revealed that the average
diameters of wires on all planes to be between 50 nm to 75 nm.
We have quantitatively and statistically analyzed crucial nanowire assembly
parameters from SEM images of SnO
2
nanowires taken from 20 samples of these three
types of sapphire substrates, and the results are shown in Figure 5.2. From histograms of
nanowire density, alignment defect density (defined as crossing or crooked nanowires),
and misalignment angle of SnO
2
nanowires on three different types of substrates in
Figure 5.2, SnO
2
nanowires synthesized on un-annealed R-plane sapphire substrates
showed high density (4.11 ± 0.11 nanowires/µm), low alignment defect density
(0.35±0.34 nanowires/µm), and good alignment (98% within ±1º misalignment).
Densities of aligned SnO
2
nanowires on annealed A-plane and annealed M-plane
sapphire are 2.79 ± 0.75 and 1.32 ± 0.46 nanowires/µm, respectively, lower than the

66
density on R-plane sapphire. Even though nanowires on annealed A-plane sapphire
showed higher alignment defect density (1.38±0.65 nanowires/µm) than nanowires on
annealed M-plane sapphire (0.53±0.26 nanowires/µm), they have comparable distribution
of alignment angles (91%of nanowires on A plane and 90% of nanowires on M plane are
within ±1º).
Figure 5.2 Histograms of nanowire assembly parameters of aligned SnO
2
nanowires on annealed A-
plane, annealed M-plane, and R-plane sapphire. (a - c) aligned SnO
2
nanowire density on annealed A-
plane (a), annealed M-plane, (b) and R-plane sapphire substrates (c) from 20 samples for each plane.
(d-e) SnO
2
nanowire alignment defect density on annealed A- plane (d), annealed M-plane (e), and R-
plane sapphire substrates (f) from 20 samples for each plane. (g-i) nanowire misalignment angles of
100 nanowires from 20 samples on annealed A- plane (g), annealed M-plane (h) and R-plane sapphire
substrates (i)
X-ray diffraction (XRD) patterns of the nanowires were taken to investigate the
relationship between the rutile structured SnO
2
nanowires and the sapphire substrates.

67
Figures 5.3 (a-c) show the SnO
2
nanowire orientation on the A-plane, M-plane, and R-
plane substrates, respectively. The SnO
2
(101) peak appears as the dominant SnO
2
peak
on all three sapphire planes, suggesting SnO
2
(101) to be the interfacing plane in all three
cases. Minor peaks such as SnO
2
(200) may be caused by random, un-aligned SnO
2

nanowires grown around the catalyst stripe. Confirmation of A-plane and R-plane
sapphire substrates is evident in the Al
2
O
3
(112 ‾ 0), (11 ‾ 02), and (22 ‾ 04) peaks that appear
in the corresponding XRD plots. Although no M-plane peak is seen, it can be explained
by the irregular surface caused by the grooves found after annealing.
Figure 5.3 XRD data for aligned SnO
2
nanowires grown on A-plane (a), M-plane (b), and R-plane (c)
sapphires show all three planes tend to interface the SnO
2
(101) plane.
The preference of interfacing the SnO
2
(101) plane with sapphire is supported by
several previous reports,
111-113
from which the relative orientation between the aligned
SnO
2
nanowires and the sapphire planes can also be predicted. For example, the interface
between tilted SnO
2
nanowires and A-plane sapphire was reported as SnO
2
(101)
[1
>
01]||Al
2
O
3
(112
>
0) [11
>
00],
111
and is assumed to be the orientation between our aligned
SnO
2
nanowires and the A-plane sapphire substrate as well. This is illustrated in Figures
5.4 (a) and 5.4 (b), which respectively depict the atomic structure of the Al
2
O
3
(112 ‾ 0) and
SnO
2
(101) plane. The interface of the two planes overlapped in such a way that the
dashed axis in Figure 4.4 (a) aligns with the dashed axis in Figure 5.4 (b), while the solid
axis align with the corresponding solid axis. Since the nanowires grow epitaxially along
Al
2
O
3
[11 ‾ 00], it follows that the SnO
2
nanowire growth direction on A-plane sapphire is
SnO
2
[1
>
01].

Additionally, the (101) surface of SnO
plane sapphire
112, 113

[010]||Al
2
O
3
(11 ‾ 02) [112 ‾ 0]. This also agrees with the orientation for our aligned SnO
nanowires on R-plane sapphire, and it i
the SnO
2
(101) and Al
respectively. The aligning axes are drawn with the same line type again. B
SEM image in Figure 5
direction on R-plane sapphire, we can conclude that the SnO
also grow along the SnO
on the other hand, are aligned along the nanogrooves, in which the exposed surface is
mainly R plane.
98
Thus the interface relationship on annealed M
SnO
2
(101) [010]||Al
2
O
3
(1
in the Al
2
O
3
[112 ‾ 0] direction, the nanowire growth direction on annealed M plane
becomes SnO
2
[010].
Figure 5.4 Diagrams of atomic arrangement for A
plane sapphire (c). Dashed vectors show sapphire to SnO
solid vectors show alignment in the x direction. Green circles are oxygen atoms, pink circles are
aluminum atoms, and gray circles are tin atoms.
The aligned SnO
confirmed using transmission electron microscopy (TEM) imaging of a cross
sample prepared by JEOL 4500 Focused Ion Beam (FIB). The schem
illustrates that the cross-section was cut perpendicularly to the nanowire growth direction,
and the direction of the electron beam from the JEOL 2100F TEM is parallel to the
68
Additionally, the (101) surface of SnO
2
films have been observed to interface R
with the two materials orientated such that SnO
0]. This also agrees with the orientation for our aligned SnO
plane sapphire, and it is illustrated through the atomic arrangements of
(101) and Al
2
O
3
(11 ‾ 02) surfaces shown in Figures 5.4 (a) and 5.4 (b
respectively. The aligning axes are drawn with the same line type again. B
SEM image in Figure 5.1 (l) revealed that nanowires grow along the Al
plane sapphire, we can conclude that the SnO
2
nanowires on the R plane
also grow along the SnO
2
[1 ‾ 01] direction. Nanowires on the annealed M
ther hand, are aligned along the nanogrooves, in which the exposed surface is
Thus the interface relationship on annealed M-plane sapphire is also
(11 ‾ 02) [112 ‾ 0]. However, because the nanogrooves are oriented
0] direction, the nanowire growth direction on annealed M plane
Diagrams of atomic arrangement for A-plane sapphire (a), (101) plane SnO
plane sapphire (c). Dashed vectors show sapphire to SnO
2
lattice alignment in the y direction while
solid vectors show alignment in the x direction. Green circles are oxygen atoms, pink circles are
aluminum atoms, and gray circles are tin atoms.
igned SnO
2
nanowire orientation on R-plane sapphire was further
confirmed using transmission electron microscopy (TEM) imaging of a cross
sample prepared by JEOL 4500 Focused Ion Beam (FIB). The schematic in Figure 5
section was cut perpendicularly to the nanowire growth direction,
and the direction of the electron beam from the JEOL 2100F TEM is parallel to the
films have been observed to interface R-
with the two materials orientated such that SnO
2
(101)
0]. This also agrees with the orientation for our aligned SnO
2
s illustrated through the atomic arrangements of
02) surfaces shown in Figures 5.4 (a) and 5.4 (b),
respectively. The aligning axes are drawn with the same line type again. Because the
vealed that nanowires grow along the Al
2
O
3
[11 ‾ 01]
nanowires on the R plane
01] direction. Nanowires on the annealed M-plane sapphire,
ther hand, are aligned along the nanogrooves, in which the exposed surface is
plane sapphire is also
0]. However, because the nanogrooves are oriented
0] direction, the nanowire growth direction on annealed M plane

plane sapphire (a), (101) plane SnO
2
(b), and R-
lattice alignment in the y direction while
solid vectors show alignment in the x direction. Green circles are oxygen atoms, pink circles are
plane sapphire was further
confirmed using transmission electron microscopy (TEM) imaging of a cross-sectional
atic in Figure 5.5 (a)
section was cut perpendicularly to the nanowire growth direction,
and the direction of the electron beam from the JEOL 2100F TEM is parallel to the

69
nanowire growth. The resulting TEM image can be seen in Figure 5.6 (a). The growth
direction of the nanowire can be confirmed to be SnO
2
[1 ‾ 01] from the diffraction pattern
of Figure 5.6 (b), where two normal planes are indicated to be SnO
2
(101) and SnO
2

(010). Similarly, we can confirm that the nanowires grow along the sapphire [11 ‾ 01]

Figure 5.5 Transmission electron microscopy (TEM) imaging of aligned SnO
2
nanowire on R-plane
sapphire substrate. (a) Diagram of cross-sectional sample preparation by FIB lift-out technique, from
the as-grown sample to protective platinum and carbon deposition, and to finished cross-sectional
TEM sample. Direction of the electron beam is parallel to the nanowire growth direction as indicated
by the double arrow. (b) Bright spots are electron diffraction pattern of sapphire taken from the cross-
sectional sample. Black dots are simulated diffraction pattern of sapphire looking into the [1 ‾ 101]
lattice vector direction. (c) Bright spots are electron diffraction pattern of SnO
2
taken from the cross-
sectional sample. Black dots are simulated diffraction pattern of SnO
2
looking into the [1 ‾ 01] lattice
vector direction. The faint rings are diffraction from the protective Pt layer. (d) TEM image showing
the location on the cross-sectional sample from where the SnO
2
electron beam diffraction patterns in
(b) and (c) are taken. Area enclosed by the dashed circles with underscored b and c correspond to
image (b) and (c), respectively.
Al 2 O 3 [1101] SnO 2 [101]
SnO 2 (101)
SnO 2 [101]
SnO 2 (010)
Al 2 O 3 [1101]
a
d
b c
Pt and C deposition
SnO 2 [101]
Al 2 O 3 [1101]
Sputtered Pt
101
020
020
101
121
121
1120
1102
1102
1120
3124
1324
b
c

direction from Figure 5.6 (c), where the two normal planes are Al
(112 ‾ 0). The indexing of the diffraction spots is confirmed using CrystalMak
simulations for diffraction into the sapphire [
are overlaid in Figure 5.5 (b) and (c), respectively. The exact locations from the cross
section used to obtain the diffraction patterns
shown in Figure 5.5 (d). The orientations and interface planes from TEM analysis show
agreement with the XRD data and orientation data from literature.
Figure 5.6 (a) TEM image of a cross
Electron diffraction pattern of SnO
diffraction pattern of cross-section of R
Figure 5.7 (a) shows a high
taken from the same spot as Figure 5
and Al
2
O
3
in [112
>
0] are measured to be approximately 4.8 Å, which is close
specified in Figure 5.4 (b) and (c). Figure 5
a single aligned SnO
2
nanowire on an R
width about 123 nm. We also observed that synthesis pressure significantly affected
whether the resultant nanowires are parallel, planar SnO
oriented, free-standing nanowires. By increasing the pressure inside the furnace to
atmospheric pressure during the synthesis, we obtained a higher density of SnO
nanowires on top of the
nanowires grown on R-plane sapphire under such an atmospheric condition. The densely
interlaced nanowires appear similar to freestanding nanowire forests grown on Si
70
.6 (c), where the two normal planes are Al
2
O
3
(1
0). The indexing of the diffraction spots is confirmed using CrystalMak
simulations for diffraction into the sapphire [1 ‾ 101] axis and the SnO
2
[
.5 (b) and (c), respectively. The exact locations from the cross
section used to obtain the diffraction patterns for the sapphire and SnO
.5 (d). The orientations and interface planes from TEM analysis show
agreement with the XRD data and orientation data from literature.
TEM image of a cross-sectional view of aligned SnO
2
nanowire on sapphire. (b)
Electron diffraction pattern of SnO
2
nanowire taken from similar cross-section locations. (c) Electron
section of R-plane sapphire.
shows a high-resolution TEM image of an aligned SnO
ot as Figure 5.6 (a). Lattice spacing of SnO
2
along [010] direction
0] are measured to be approximately 4.8 Å, which is close
5.4 (b) and (c). Figure 5.7 (b) shows a high-resolution SEM image of
nanowire on an R-plane sapphire substrate with smooth surface and
width about 123 nm. We also observed that synthesis pressure significantly affected
sultant nanowires are parallel, planar SnO
2
nanowires or randomly
standing nanowires. By increasing the pressure inside the furnace to
atmospheric pressure during the synthesis, we obtained a higher density of SnO
top of the Au catalyst. Figure 5.7 (c) shows a SEM image of SnO
plane sapphire under such an atmospheric condition. The densely
interlaced nanowires appear similar to freestanding nanowire forests grown on Si
(11 ‾ 02) and Al
2
O
3

0). The indexing of the diffraction spots is confirmed using CrystalMaker®
1 ‾ 01] axis, which
.5 (b) and (c), respectively. The exact locations from the cross-
for the sapphire and SnO
2
nanowire are
.5 (d). The orientations and interface planes from TEM analysis show

nanowire on sapphire. (b)
section locations. (c) Electron
resolution TEM image of an aligned SnO
2
nanowire
along [010] direction
0] are measured to be approximately 4.8 Å, which is close to values
resolution SEM image of
plane sapphire substrate with smooth surface and
width about 123 nm. We also observed that synthesis pressure significantly affected
nanowires or randomly
standing nanowires. By increasing the pressure inside the furnace to
atmospheric pressure during the synthesis, we obtained a higher density of SnO
2

.7 (c) shows a SEM image of SnO
2

plane sapphire under such an atmospheric condition. The densely
interlaced nanowires appear similar to freestanding nanowire forests grown on Si

substrates.
103
They are connected to the substrates only at one end, and are not epitaxially
in-plane with the sapphire surface. Careful sonication of this dense SnO
revealed no significant layers of aligned SnO
hidden layer of aligned nanowires being removed during the sonication is ruled out
because purposeful sonication of visibly aligned SnO
conditions was unsuccessful. This compar
pressure, where the Sn vapor partial pressure is large, favors growth of dense
that are forest-like and un
to grow in alignment on the substrate surface. Similar effect of partial pressure on
Figure 5.7 (a) High-resolution TEM image of the cross section of an aligned SnO
plane sapphire substrate. (b) High
sapphire. (c) SEM image of SnO
set to atmospheric pressure. Dense, un
condition.
71
They are connected to the substrates only at one end, and are not epitaxially
plane with the sapphire surface. Careful sonication of this dense SnO
2
ignificant layers of aligned SnO
2
nanowire underneath. The possibility of a
hidden layer of aligned nanowires being removed during the sonication is ruled out
because purposeful sonication of visibly aligned SnO
2
nanowires under the same
uccessful. This comparison suggests that synthesis at
where the Sn vapor partial pressure is large, favors growth of dense
like and un-aligned, while lower Sn partial pressure allows the nanowires
alignment on the substrate surface. Similar effect of partial pressure on
resolution TEM image of the cross section of an aligned SnO
plane sapphire substrate. (b) High-resolution SEM image of an aligned SnO
2
nanowire on R
sapphire. (c) SEM image of SnO
2
nanowires grown on R-plane sapphire with the synthesis pressure
set to atmospheric pressure. Dense, un-aligned nanowire forest can be seen with such synthesis
They are connected to the substrates only at one end, and are not epitaxially
2
nanowire forest
nanowire underneath. The possibility of a
hidden layer of aligned nanowires being removed during the sonication is ruled out
nanowires under the same
ison suggests that synthesis at atmospheric
where the Sn vapor partial pressure is large, favors growth of dense nanowires
aligned, while lower Sn partial pressure allows the nanowires
alignment on the substrate surface. Similar effect of partial pressure on

resolution TEM image of the cross section of an aligned SnO
2
nanowire on R-
nanowire on R-plane
plane sapphire with the synthesis pressure
owire forest can be seen with such synthesis

72
nanowire growth was also observed for InAs nanowires
99
, where the Gibbs-Thomson
equation was used to show that higher precursor vapor pressure is required for free-
standing nanowires than that required for planar nanowires.
5.3 Aligned SnO
2
nanowire field effect transistors (FETs) and their
electrical performance
Scalable device fabrication is an important step for practical integration of metal
oxide nanowires in applications like display and memory technology
102, 114
and various
types of sensors.
32, 115
After synthesis, aligned SnO
2
nanowires grown on sapphire were
fabricated as FETs using standard photolithography technology. Details of the fabrication
are described in the Methods Section in the Appendix IV. The finished device is
represented in the schematic in Figure 5.8 (a), which shows how the electrodes can be
easily patterned perpendicular to the length of the SnO
2
nanowire without complex
techniques such as electron beam lithography. A top-view optical image of one device is
shown in Figure 5.8 (b), and the alignment of SnO
2
nanowires across the source and drain
electrodes can be seen in the SEM image of Figure 5.8 (c), where 2 nanowires that grew
parallel to each other are covered with metal electrodes oriented perpendicularly to the
length of nanowires. The FET characteristics of this device with the 2 parallel nanowires
are shown in Figures 5.8 (d) and (e). In the drain current (I
D
) versus the drain-to-source
voltage (V
D
) plot of Figure 5.8 (d), the device exemplifies a MOSFET behavior in which
the I
D
is linearly dependent on V
D
at low drain voltages but saturates when V
D
increases
beyond a few volts. The current is also shown to increase with the gate voltage (V
G
). At a
V
G
of 10 V the current can reach 40 A, while at a V
G
of -4 V the current is shown to be
turned off. The I
D
versus V
G
plot in Figure 5.8 (e) also confirms that the current is turned
off with V
G
< -4 V. The curve is measured at V
D
= 200 mV, where the transistor is in the
ohmic region. The plot of I
D
in log scale indicates a good on/off ratio of over 10
6
, and the

73

Figure 5.8 Aligned SnO
2
nanowire transistor study. (a) Diagram of device fabrication. (b) Top-view
optical image of an aligned SnO
2
nanowire transistor. (c) SEM image of two aligned SnO
2
nanowires
bridging the source and drain electrodes of a nanowire transistor. (d) I
D
-V
D
family plot of the
transistor shown in (c). (e) I
D
-V
G
plot of the transistor in (c), plotted with standard scale in black and
logarithmic scale in blue. The subplot shows the transconductance of the same device.
subplot shows the transconductance (dI
D
/dV
G
) of the device to be around 0.16 µS. This
transconductance value and the relationship dI
D
/dV
G
= (C/L
2
)V
D
can be used to calculate
the electron mobility, µ. The nanowire channel length (L) is 5 µm. The gate capacitance
(C) can be estimated by modeling the nanowires as cylinders on the sapphire plane.
102

Assuming that the aluminum oxide dielectric constant is around 9,
116
and using an
average nanowire diameter of 75 nm each, the gate capacitance is calculated to be 1.79 ×
10
-15
F for each nanowire.
102, 117, 118
From these values, the corresponding electron
mobility is 55 cm
2
/V·s. To benchmark its device performance, we note that the aligned
SnO
2
nanowire device has an on/off ratio and mobility that is higher than that of devices
based on individual spin-coated SnO
2
nanowires previously demonstrated using a laser
ablation technology.
103
This confirms the high quality of our aligned SnO
2
nanowires and

74
supports the idea that aligned nanowire transistors have the potential to be fabricated with
superior yield and scalability compared to individual nanowire device, and expected to
have superior performance compared to nanowire network devices.
To show the performance which the aligned SnO
2
nanowire transistors can
produce, we have plotted histograms of device performance from 20 devices in Figure
5.9. Figure 5.9 (a) shows that the I
D
measured at V
D
= 200 mV and V
G
= 10 V are
typically between 1 µA and 3 µA with the average I
D
being 1.48 µA and the standard
deviation being 0.33 µA. The on/off current ratios (Figure 5.9 (b)) for all the transistors
fall in the range of 10
4
to 10
8
with majority (13 out of 20 devices) showing on/off ratios
of 10
5
to 10
6
. In Figure 5.9 (c), the average transconductance of the 20 devices is 157 nS
with a standard deviation of 37.23 nS. The threshold voltage (V
TH
) variation is shown in
Figure 5.9 (d), where the average V
TH
is -4.22V with a standard deviation of 0.81V. The
histogram of electron mobility is shown in Figure 5.9 (e) with an average of 71.68 cm
2

/V•s.

Figure 5.9 Histograms of device performance from 20 aligned SnO
2
nanowire transistors. (a) On state
drain current (I
D
) at V
D
= 200 mV, V
G
= 10 V. (b) On/Off ratio of I
D
. (c) Transconductance (g
m
), (d)
Threshold voltage (V
TH
). (e) Mobility (µ).

75
5.4 Aligned SnO
2
nanowire transistors for the control circuit in display
application
The high I
on
, on/off ratio, and mobility of the aligned SnO
2
nanowire transistor are
highly desired traits in many micro- and nano- electronic applications such as organic
light emitting diode (OLED) control circuitry. OLEDs are a promising display
technology due to many of their superior characteristics such as light-weight, excellent
color purity, low power consumption, etc.
119, 120
To fully develop OLED as a low-cost,
large-scale product, much research is currently on-going to study the material and
fabrication of the OLED and its driving circuit. Aligned SnO
2
nanowire transistors are
compatible with flexible and transparent electronics, in addition to having good electronic
properties as mentioned above, and can be a good candidate for the driving circuit. As a
proof of concept, we demonstrate the application of a top-gated, aligned SnO
2
nanowire
FET as the control for an external OLED with the structure of 4,4’-bis[N-(1-naphthyl)-N-
phenylamino]biphenyl(NPD)/tris(8-hydroxyquinoline) aluminum(Alq3). This is a green
light OLED with indium tin oxide (ITO) as the anode and aluminum (Al) as the cathode.
The entire circuit is connected as shown in Figure 5.10 (a). A coaxial cable with a clamp
terminal is connected to V
DD
on one side while its clamp side is placed on the cathode of
the OLED. Another cable of the same type has its clamp side placed on the anode of the
OLED while the other side is connected to the drain of the aligned SnO
2
nanowire FET.
The relationship between the current through the diode (I
OLED
) and the power supply
(V
DD
) is plotted in Figure 5.10 (b), and the curves show good diode behavior with a clear
cut-off region and triode region under different SnO
2
nanowire FET gate voltage (V
G
),
showing good control from the FET over the OLED. The cut-off voltage of V
DD
is around
-13V in accordance with the threshold voltage of the OLED. From the I
OLED
-V
G
curves in
Figure 5.10 (c), the FET is capable of providing enough driving current for the OLED,
which requires approximately 0.2µA to have observable light emission. The optical
images in Figure 5.10 (d) show the OLED at various light intensities under different V
G

values of the SnO
2
nanowire FET, at fixed V
DD
= 13 V. From the optical images, the

76
OLED is very bright when V
G
= 10 V, gets dimmer as V
G
decreases toward negative
voltages, and is totally turned off when V
G
becomes more negative than -10 V,
corresponding to the I
OLED
-V
G
curve measured at V
DD
= 13 V in Figure 5.10 (b), where
the curve enters the cut-off region around -10 V.

Figure 5.10 Aligned SnO
2
nanowire FET control circuit for OLED. (a) Circuit diagram of OLED
connection to FET. (b) I
OLED
-V
G
family curve (c) I
OLED
-V
DD
family curve (d) Optical images of OLED
intensity as V
G
decreases.
5.5 Aligned SnO
2
nanowire transistors for photodetecting application
SnO
2
has also been documented to have excellent photoconductive properties,
whether as a thin film
101
or as a nanowire,
103
and we expect that the alignment of SnO
2

nanowires will further improve this response. The parallel alignment of the nanowires is
expected to be especially advantageous for having strong sensitivity to polarized light due
to the uniformity of the nanowire orientation between the source and drain electrodes of
the FET device. To test the photoconductive properties of the devices, two UV lamps
with wavelength of 254 nm and 365 nm, placed 2 cm above our aligned SnO
2
nanowire
FET, were used as photo source. No gate voltage was applied to the nanowire FET, and

77
measurements were taken in air, at room temperature, and under indoor incandescent
light. From the I
D
-V
D
data presented in Figure 5.11 (a), enhanced conduction was
observed for UV illumination of both wavelengths. However, the 254 nm light induced a
significantly larger magnitude of response than that elicited from the 365 nm light, whose
expanded curve is shown in the subplot. The zero-bias conductance before and after the
365 nm UV light exposure is calculated to be 1.1 and 8.8 nS, respectively, and the
illumination with 254 nm UV light induced a conductance value of 320 nS. This on/off
ratio is comparable to those reported for GaN nanowire FET photoconduction.
121
The
mechanisms of photoconduction in metal oxides and the difference in conduction due to
the two wavelengths are well documented.
103, 115
Because SnO
2
has a 3.6 eV bandgap, a
photon from the 254 nm UV light with an energy of 4.9 eV can sufficiently excite
electron-hole pair generation while a photon from the 365 nm light with an energy of 3.4
eV cannot. However, a small increase in conduction in response to the 365 nm light still
occurs due to the non-zero photon energy spectrum. The second photoconduction
mechanism occurs because UV light cleanses adsorbed species from the nanowire surface
and frees electrons from reduced molecules such as O
-
2
. Real time UV detection was also
performed by turning the 254 nm UV lamp on and off for five cycles, as shown in Figure
5.11 (b). In this experiment, the V
D
of the device was fixed at 500 mV. The conductance
of the nanowire showed a rapid increase from 1.3 nS to an averaged on-conduction of
193 nS upon exposure to UV light. Detailed data analysis revealed that the conductance
of the nanowire increased to about 50% of its average on-current within 5 s after the UV
lamp was turned on, and reached about 90% within 10 s. The aligned SnO
2
nanowire
sensor also exhibited a high recovery speed, as seen in the sharp current drop when the
light was turned off. The real-time current response to the 365 nm UV illumination is
shown in Figure 5.11 (c) as a comparison to the 254 nm light. The difference in response
reiterates the ability of the nanowire device to distinguish specific wavelengths, and the
FET shows good stability when detecting multiple light sources.

78
Figure 5.11 Photoconduction and polarization detection. (a) I
D
-V
D
curves of aligned SnO
2
device
before UV illumination (black), after 365 nm UV illumination (red), and after 254 nm UV
illumination (blue). Expanded curves for before and 365 nm illumination are presented in the subplot
for clarity, with the same color legend. (b) Real-time detection of 254 nm UV illumination on an
aligned SnO
2
device as the UV lamp is turned on and off for 5 cycles. (c) Real-time detection of 2
different wavelengths using one aligned SnO
2
device. (d) Polarized 254 nm UV detection. Black
triangles are averaged peak I
D
during the time that the UV lamp is turned on at the corresponding
angle. The red curve is fitted data showing cos2θ dependence.
Moreover, the stability of photocurrent has been investigated on the same
photodetector as shown in Figure 5.12. The aligned SnO
2
nanowire detector was
illuminated using the 254 nm UV lamp over 100 minutes shown in Figure 5.12 (a). The
photocurrent exhibited a sharp response similar to the result shown in Figure 5.12 (b)
upon UV illumination, and reached its steady state after 50 minutes as shown in Figure

79
5.12 (a) with only small variation of 2 % between 50 minutes and 100 minutes as shown
in Figure 5.12 (a) inset. The long term stability of this photodetector has also been
investigated. Six months after we performed measurements shown in Figure 5.11 (b) and
(c), we performed similar photocurrent measurement once every day for 15 days, and
Figure 5.12 (b) shows the photocurrent averaged between time = 0 s and 200 s upon 254
nm UV illumination. The aligned SnO
2
nanowire photodetector shows good long term
stability with the average photocurrent being 123.05 ± 16.15 nA, which is in the same
order of the magnitude as data shown in Figure 4.11 (b) and (c).

Figure 5.12 (a) Stability test of real-time response from UV 254 nm illumination on aligned SnO2
nanowire detector with V
DS
= 500 mV. Photoconduction decreased and reached steady state after 60
minutes (b) Long term stability test of photoconduction. Each data point is an average
photoconduction response over 200 s after 254 nm UV illumination measured from the same sensor
used in Figure 4.11.
Besides working as photodetectors, semiconducting nanowires are also expected
to discriminate between different polarization states of the incident radiation due to their
one-dimensional nature.
103, 121
SnO
2
nanowires possess both an one-dimentional structure
and a bandgap in the UV regime, and thus are excellent material for polarized UV
studies. The polarization detection measurement was carried out by mounting a Glan-
laser linear polarizer between the aligned SnO
2
nanowire device and the 254 nm
wavelength UV lamp. All UV illumination that reached the device was passed through
the polarizer as the polarizer was rotated through various angles. Figure 5.11 (d) shows a

80
plot of the aligned SnO
2
nanowire current as a function of the polarizer angle while the
voltage between source and drain electrodes of the nanowire device was set at 500 mV.
The nanowire conductance showed a periodic dependence (cos2θ) on the polarization
angle (θ), with a period of 180°. From the maximum and minimum conductance values
respectively observed under a parallel (G
||
= 32 nS) and a perpendicular (G ⊥=16 nS) field,
a polarization ratio of σ=0.3 was calculated according to A = (G
||
- G ⊥) /(G
||
+ G ⊥). This
polarization ratio is higher than previously observed values for that of GaN
121
and laser-
ablation grown SnO
2
103
nanowire devices, and equal to the observed value for that of
carbon nanotube devices.
122
This improvement further illustrates the significance of
controlled orientation of nanostructures for electronic applications.
5.6 Aligned SnO
2
nanowire transistors for chemical sensing application
Metal oxide nanowires have stimulated significant interest for chemical sensing
and biosensing applications, which have also been discussed in two recent review
papers.
5, 123
The mechanisms of conduction changes within SnO
2
and other metal oxides
due to oxidizing or reducing gas molecules adsorbed on the surface have been well
documented.
124-126
The scalability and control allowed by aligned SnO
2
nanowires can
help to further advance this material as a practical gas sensor. An example of an oxidizing
species with strong electron withdrawing capability is the environmental toxin NO
2
gas
molecule. We chose to detect NO
2
to show proof of concept because this gas is a
dangerous air pollutant that contributes to the formation of ozone and acid rain. To test
the performance of the aligned SnO
2
nanowires gas sensors, different concentrations of
NO
2
are diluted in argon and then introduced to the device surface. Figure 5.13 (a) shows
the real time response to NO
2
gas from the aligned SnO
2
nanowire sensor device. At time
0, the sensor is illuminated with 254 nm wavelength UV light to increase the device
conduction to a suitable level for sensing. Argon is also introduced as the ambient gas.
No gate bias was applied to the sensor, and V
D
is fixed at 500 mV during the duration of
the sensing. At point “a” in the plot, NO
2
is introduced to the sensor surface at a

81
concentration of 0.2 ppb. The effect of carrier reduction through the withdrawal of
electrons by adsorbed NO
2
is seen here. The sensor current immediately decreases about
50-60% and then begins to saturate around point “b”, at which point the NO
2
gas is
turned off while the argon and the UV light are kept on. This procedure is repeated for
NO
2
at concentrations of 0.5 ppb, 1 ppb, and 2 ppb. Concentrations beyond 2 ppb saturate
the sensor response by turning the sensor to a virtually off state. Our detection limit of 0.2
ppb is comparable to the 0.1 ppb NO
2
limit of a functionalized carbon nanotube sensor,
127

which is one of the most sensitive NO
2
nanosensors reported to date. This detection limit

Figure 5.13 NO
2
sensing (a) Real-time detection of NO
2
gas of various concentrations by 2 different
aligned SnO
2
nanowire devices. NO
2
gas is turned on at point “a” and turned off at point “b” (b) Plot
of normalized drain current change (ΔI/I
0
) against NO
2
concentration. In the subplot, the inverse of
normalized current and concentration are shown. The black triangles represent measured data, and the
red line is a linear fit of the 4 concentration data. (b) I
D
-V
D
plots of device 1 after being introduced to
increasing concentrations of NO
2
.

82
is more than sufficient for the environmental health standard of 53 ppb. It is also lower
than the 2 ppm limit reported from SnO
2
nanoribbon sensors
107
and the 0.2 ppm limit
from SnO
2
nanowire sensors enhanced with additional resistance modulation.
108

Although many other one-dimensional metal oxide NO
2
nanosensors have been
reported
10, 108, 128-132
, their detection limits typically range from 1 ppm (as in the case of
CuO nanowire sensors)
131
to 1 ppb (as statistically extrapolated for TiO
2
nanowires).
132

Normalized current change from the real-time sensing is plotted against the NO
2

concentration in Figure 5.13 (b), where the first three smaller concentrations are shown to
fit the Langmuir-Isotherm model very well, as can be seen from the linear fit in the
subplot of inverse responses. The data point for the 2 ppb concentration in the subplot
deviates slightly from the linear fit due to the saturation. The 0.2 ppb detection level is
confirmed again in a second experiment (Figure 5.13 (c)) where the aligned SnO
2

nanowire device current is plotted against the drain voltage at each NO
2
concentration.
Resistance is shown to decrease significantly, as the concentration is increased from 0
ppb to 2 ppb. The clear shift in conduction at all NO
2
concentration levels shows good
repeatability and stability of the sensor.
5.7 Summary
In conclusion, we have demonstrated the growth of parallel, planar SnO
2

nanowires guided by annealed A-plane, annealed M-plane, and R-plane sapphire. A
relatively lower synthesis pressure was shown to favor the growth of these guided, planar
nanowires while a higher pressure was shown to favor the growth of a non-planar and un-
aligned nanowire forest. The alignment orientation of the SnO
2
nanowire and the
sapphire substrates were explored using XRD measurements. A straightforward
photolithography process was demonstrated for patterning the aligned nanowires for
working FETs with high mobilities and on/off ratios sufficient for driving an external
OLED, with a clear distinction between on and off intensities. The electrical parameters
support the advantages of aligned SnO
2
nanowire transistors over network SnO
2

nanowire FETs. This is further demonstrated in the performance of the aligned SnO
2

83
nanowire FET for polarized UV light detection, where the polarization ratio is higher
than that from laser-ablation synthesized nanowire FETs. Finally, we demonstrated that
the SnO
2
nanowire FET can be used as NO
2
detectors with sensitivity below the ppb
range. As SnO
2
has been an important material for many of the applications investigated
in this paper, this demonstration of aligned SnO
2
nanowires is an important step toward a
more scalable and higher performance SnO
2
nanowire devices. Having established this
initial platform, we will further investigate the synthesis and fabrication process to
achieve FETs with low device-to-device variations for future applications such as
transparent and flexible electronics and sensors.

84
Chapter 6 Summary and Future Work
6.1 Summary
In this dissertation, we have demonstrated electronic and sensing applications of
one-dimensional nanomaterials, which are polysilicon (poly-Si) nanoribbons, metal oxide
nanoribbons and tin oxide (SnO
2
) nanowires. They exhibit promising performance and
have potential to be utilized in the practical aspect.
Firstly, we have developed a scalable and uniform top-down approach to fabricate
poly-Si nanoribbon biosensors using two photolithographic masks to define dimension
and position of metal electrodes and nanoribbons. All fabrication processes are
compatible with the conventional microelectronic facilities. Poly-Si on Si is deposited by
low pressure chemical deposition (LVCVD) with highly controllable and reproducible
thickness. Doping concentration can be easily controlled by choice of concentration of
dopant solution and thermal annealing without need of any toxic gas or ion implantation.
Devices show good uniformity of their electrical performance and good ionic sensitivity
on both the wide pH range, pH 4 to pH 9 and the physiological pH range, pH 7.2 to 8.
We investigate surface chemistry of poly-Si nanoribbons in order to optimize their
sensitivity. We found that with thin layer of deposited silicon dioxide, number of bound
target molecules is improved significantly. We have demonstrated detection of cancer
antigen 125 (CA-125), a biomarker for ovarian cancer, at 10 U/ml which is about an
order of magnitude lower than the clinically relevant level.
Secondly, we have demonstrated another highly scalable top-down fabrication
technique for sputtered indium oxide (In
2
O
3
) nanorribon field effect transistors using two
photolithographic masks to define dimension and position of metal electrodes and
nanoribbons. Devices fabricated using this approach show uniform and good electrical
performance without requiring doping or post-process annealing. The fabrication is
highly scalable, low cost, and a low temperature process that is compatible with the

85
CMOS fabrication facilities. In
2
O
3
is selected for the nanoribbon material because its
electrical performance and long-term stability in aqueous solution are better than other
metal oxide materials. In
2
O
3
nanoribbon devices exhibited good sensitivity in both wide
range of pH solution from pH 4 to 9 and physiological range between 6.7 and 8.2.
Streptavidin-biotin has been chosen to demonstrate signal amplifying electronic enzyme-
linked immonusorbent assay (ELISA) with picomolar sensitivity showing 15% changes
in normalized current. We demonstrated electronic ELISA for detection of HIV p24
proteins at concentration about 20 fg/ml or 250 viruses/ml, which is about 3 orders of
magnitude lower than commercial ELISA kit on the market, in both buffer and human
blood serum.
Thirdly, we have redesigned our In
2
O
3
naoribbon FET biosensor chips integrated
with the on-chip gate electrode to accommodate multiplex detection of biomarkers. The
on-chip gate electrode exhibits comparable electrical performance to the Ag/AgCl
electrode at the same liquid gate potential. In addition, In
2
O
3
nanoribbon biosensors using
the on-chip gate show uniform electrical performance and good ionic sensitivity for
change in wide range of pH solutions. To move forward for practical use of our assay, we
demonstrated detection of troponin I, a biomarker for diagnosis of acute myocardial
infarction (AMI), with reduction of assay time from 10 hours to 1 hours without
significant loss of sensitivity. The projected limit of detection is about 5 orders of
magnitude lower than commercial ELISA kits. We completed series of experiments to
generate calibration curves for a biomarker panel for AMI which includes troponin I,
creatine kinase MB (CK-MB) and b-type natriuretic peptide (BNP). To test accuracy of
our calibration curves, we ran an assay with spiked BNP in human whole blood. The
difference between the measured response and the calculated value from the calibration
curve is within an acceptable range of 3%.
The control assembly is one of major challenges for bottom-up nanometerials to
move to the practice use. Lastly, we have demonstrated the growth of parallel, planar
SnO
2
nanowires guided by annealed A-plane, annealed M-plane, and R-plane sapphire. In

86
our study, a relatively lower synthesis pressure was shown to favor the growth of these
guided and planar nanowires while a higher pressure was shown to favor the growth of a
non-planar and un-aligned nanowire forest. The confirmation of the alignment orientation
of the SnO
2
nanowire and the sapphire substrates was characterized by X-ray diffraction
(XRD) and transmission electron microscopy (TEM). Fabrication of aligned nanowire
FETs was demonstrated by simply patterning metal electrodes. Aligned nanowire
transistors shows better electrical performance than devices fabricated from network
nanowires. We demonstrated prospective applications of the aligned SnO
2
nanowires as
transistors for the driving circuit of organic light emitting diodes (OLED),
photodetectors, ultra violet (UV) polarization detector, and ultra sensitive NO
2
gas
sensors.
6.2 Future Work
6.2.1 Investigation for enhancement of In
2
O
3
nanoribbon biosensor sensitivity
Higher sensitivity is always better for early diagnosis of the acute myocardial
infarction because the patient can receive proper treatment on time and mortality rate can
be decreased. There are two options which can help to improve sensitivity of our
platform.
• We can change the substrate for device fabrication from silicon nitride to be
silicon dioxide because our phosphonic acid surface chemistry reacts with any
oxide surface to provide more binding sites for capture probe molecules. As a
result, there is more chance to have more targets immobilized on either
nanoribbons or the substrate yielding more urease enzymes and higher change in
conduction.
• We can change a pair of enzyme and substrate to lower pH in the sensing
chamber to increase magnitude of sensing response. To exemplify, if we have a
change of pH 1.8 in the sensing chamber from hydrolysis of urea, we will

87
measure response about 90% reduction in conduction. While the same change in
pH about 1.8 from another pair of enzyme and substrate such as acetylcholine and
acetylcholineterase, we expect to have increase in conduction about 300% due to
increase in hydrogen ions in the solution and more positively gating effect on the
n-type In
2
O
3
nanoribbon transistors.
4 5 6 7 8 9
0
1
2
3
4
5
6
7

I/I
0
pH

Figure 6.1 Average pH sensing response from three In
2
O
3
nanoribbon biosensors after
immobilization of urease enzymes with normalization at pH 7.4
6.2.2 Optimization of the assay time of In
2
O
3
nanoribbon biosensors for
diagnosis of myocardial infarction
To move our assay forward to the practical use, we need to optimize between the
assay time and its sensitivity. As mentioned in the chapter 4 that we have reduced the
assay time from 10 hours to be a hours, we possibly shorten the assay time to be less than
an hour.
• We may try to couple between secondary antibody and urease enzyme directly to
reduce number of incubation steps from 3 steps to 1 step.
• We can optimize incubation further by measuring binding kinetics of proteins in
each step using surface Plasmon resonance (SPR) for all three biomarkers, namely
troponin I, CK-MB and BNP. After obtaining binding kinetics of all steps for

88
each biomarker, we can determine the minimum incubation time of each step
when the binding rate in each step reaches equilibrium.
After reducing number of incubation steps and obtaining minimum incubation
time for each step, we need to perform a series of experiment to identify sensitivity of
each biomarker and to generate new calibration curves for all three biomarkers and
compare results with the gold standard conventional ELISA.
6.2.3 Multiplex detection on In
2
O
3
nanoribbon biosensor chips for diagnosis of
myocardial infarction
After maximizing sensitivity of our sensor platform, optimizing the assay time
and obtaining a new set of calibration curves, we need to testify our calibration curves by
spiked three proteins together into vials of physiological solution (serum or whole blood)
for a statistical study about accuracy of our assay including number of false positive and
false negative. Lastly, we can test our calibration curves with patients' samples and
compare results from our sensors with results from conventional ELISA for a statistical
study.
6.2.4 Highly scalable, uniform and sensitive In
2
O
3
nanoribbon chemical
sensors
In addition to biosensing application, In
2
O
3
is a well-know material for ultra
sensitive chemical sensing application. In
2
O
3
nanowire based FET sensors exhibit high
sensitivity on nitrogen dioxide (NO
2
), carbon monoxide (CO), hydrogen (H
2
), methane
(CH
4
), nitrous oxide (NO), ozone (O
3
), ammonia (NH
3
), acetone vapor, and ethanol
vapor.
123
Therefore, In
2
O
3
nanoribbon transistors should response the same gases similar
to In
2
O
3
nanowire sensors. As a preliminary study, we have performed chemical sensing
experiment with NO
2
which one of toxic gases generated from automobile combustion or
industrial process causing irritation to our respiratory trace and acid rain to
environment.
10
Figure 6.2 (a) shows real-time responses from three In
2
O
3
nanoribbon

89
sensors monitored simultaneously and V
DS
was applied at 200 mV to all sensors. Firstly,
argon (Ar) was introduced as the carrier gas until sensing responses show a stable
baseline. The 254 nm ultra violet (UV) lamp was turned on at 250 s for 3 s to clean from
contaminants from sensing surface. After a new base line has been set, 0.5 ppb NO
2

diluted in Ar was introduced to our sensors for 500 s before turning NO
2
off and turning
UV lamp on for 3 s. We have repeated this cycle with other NO
2
(1, 5, 10, and 100 ppb).
Sensing responses upon NO
2
exposure were gradually decreased to 1.3% at NO
2

concentration of 0.5 ppb. Devices exhibited the similar trend for higher NO
2

concentration. When we plotted these normalized responses with NO
2
concentration and
fitted with Langmuir isotherm model for molecular absorption with correlation
coefficient of 0.97 as shown in Figure 6.2 (b). NO
2
is generally known as electron
withdrawers; therefore, absorption of NO
2
molecules on In
2
O
3
nanoribbon surface yields
decrease in conduction.
133
When we compare our results with literatures on NO
2
sensing
on single In
2
O
3
nanowire transistor and multiple nanowire transistors, nanoribbon devices
had 50 % lower responses at the same concentration due to their lower surface per
volume ratio than one of nanowires as mentioned earlier, but nanoribbon chemical
sensors have better device scalability and uniformity.
10
However, our In
2
O
3
nanoribbon
devices showed their lowest detection level at 0.5 ppb which is 2 orders of magnitude
lower than 53 ppb set by the Occupational Safety & Health Administration (OSHA).
134

As this preliminary result, it is a proof that our In
2
O
3
nanoribbon platform can perform as
highly sensitive chemical sensors as well. From low temperature fabrication process of
our In
2
O
3
nanoribbon sensors, we can fabricate sensors on flexible or wearable substrates.

90

Figure 6.2 NO
2
sensing (a) Real-time detection of NO
2
gas with several concentration of NO
2
by
In
2
O
3
nanoribbon sensors (b) A plot of normalized sensing response from three sensors and NO
2
concentration with Langmuir isotherm fitting having correlation coefficient of 0.97.

91
Bibliography
Ago, H.; Nakamura, K.; Ikeda, K.; Uehara, N.; Ishigami, N.; Tsuji, M. Chem Phys Lett
2005, 408, (4-6), 433-438.
Ahn, M. W.; Park, K. S.; Heo, J. H.; Kim, D. W.; Choi, K. J.; Park, J. G. Sensor Actuat
B-Chem 2009, 138, (1), 168-173.
Amin, N.; Isaka, T.; Yamada, A.; Konagai, M. Solar Energy Materials and Solar Cells
2001, 67, (1–4), 195-201.
Arakawa, N.; Nakamura, M.; Aoki, H.; Hiramori, K. Cardiology 1994, 85, (5), 334-340.
Bellingham, J. R.; Phillips, W. A.; Adkins, C. J. J Phys-Condens Mat 1990, 2, (28),
6207-6221.
Bunimovich, Y. L.; Shin, Y. S.; Yeo, W. S.; Amori, M.; Kwong, G.; Heath, J. R. Journal
of the American Chemical Society 2006, 128, (50), 16323-16331.
Butler, J. E. Journal of immunoassay 2000, 21, (2-3), 165-209.
Byon, H. R.; Choi, H. C. Journal of the American Chemical Society 2006, 128, (7), 2188-
2189.
Camagni, P.; Faglia, G.; Galinetto, P.; Perego, C.; Samoggia, G.; Sberveglieri, G. Sensor
Actuat B-Chem 1996, 31, (1-2), 99-103.
Cao, B. B.; Chen, J. J.; Tang, X. J.; Zhou, W. L. J Mater Chem 2009, 19, (16), 2323-
2327.
Chang, H. K.; Ishikawa, F. N.; Zhang, R.; Datar, R.; Cote, R. J.; Thompson, M. E.; Zhou,
C. ACS nano 2011, 5, (12), 9883-9891.
Chang, H. K.; Wang, X.; Aroonyadet, N.; Zhang, R.; Song, Y.; Datar, R.; Cote, R.;
Thompson, M.; Zhou, C. MRS Online Proceedings Library 2013, 1569

92
Chang, L.; Song, L.; Fournier, D. R.; Kan, C. W.; Patel, P. P.; Ferrell, E. P.; Pink, B. A.;
Minnehan, K. A.; Hanlon, D. W.; Duffy, D. C.; Wilson, D. H. Journal of virological
methods 2013, 188, (1-2), 153-160.
Chen, P. C.; Shen, G. Z.; Chen, H. T.; Ha, Y. G.; Wu, C.; Sukcharoenchoke, S.; Fu, Y.;
Liu, J.; Facchetti, A.; Marks, T. J.; Thompson, M. E.; Zhou, C. W. ACS nano 2009, 3,
(11), 3383-3390.
Chen, P. C.; Sukcharoenchoke, S.; Ryu, K.; de Arco, L. G.; Badmaev, A.; Wang, C.;
Zhou, C. W. Advanced Materials 2010, 22, (17), 1900-1904.
Chen, P. C.; Shen, G.; Zhou, C. Nanotechnology, IEEE Transactions on 2008, 7, (6),
668-682.
Chen, R. J.; Bangsaruntip, S.; Drouvalakis, K. A.; Kam, N. W.; Shim, M.; Li, Y.; Kim,
W.; Utz, P. J.; Dai, H. Proceedings of the National Academy of Sciences of the United
States of America 2003, 100, (9), 4984-499.
Chiesa, M.; Cardenas, P. P.; Oton, F.; Martinez, J.; Mas-Torrent, M.; Garcia, F.; Alonso,
J. C.; Rovira, C.; Garcia, R. Nano Lett 2012, 12, (3), 1275-1281.
Choi, Y. J.; Hwang, I. S.; Park, J. G.; Choi, K. J.; Park, J. H.; Lee, J. H. Nanotechnology
2008, 19, (9).
Chua, J. H.; Chee, R. E.; Agarwal, A.; Wong, S. M.; Zhang, G. J. Analytical chemistry
2009, 81, (15), 6266-6271.
Comini, E.; Faglia, G.; Sberveglieri, G. Sensor Actuat B-Chem 2001, 78, (1-3), 73-77.
Cui, Y.; Lauhon, L. J.; Gudiksen, M. S.; Wang, J. F.; Lieber, C. M. Applied Physics
Letters 2001, 78, (15), 2214-2216.
Cui, Y.; Wei, Q.; Park, H.; Lieber, C. M. Science 2001, 293, (5533), 1289-1292.

93
Curreli, M.; Rui, Z.; Ishikawa, F. N.; Chang, H.-K.; Cote, R. J.; Chongwu, Z.; Thompson,
M. E. Nanotechnology, IEEE Transactions on 2008, 7, (6), 651-667.
Dattoli, E. N.; Wan, Q.; Guo, W.; Chen, Y.; Pan, X.; Lu, W. Nano letters 2007, 7, (8),
2463-2469.
Dominguez, J. E.; Pan, X. Q.; Fu, L.; Van Rompay, P. A.; Zhang, Z.; Nees, J. A.; Pronko,
P. P. Journal of Applied Physics 2002, 91, (3), 1060-1065.
Duan, X. X.; Li, Y.; Rajan, N. K.; Routenberg, D. A.; Modis, Y.; Reed, M. A. Nature
nanotechnology 2012, 7, (6), 401-407.
Duan, X.; Lieber, C. M. Advanced Materials 2000, 12, (4), 298-302.
Durak-Nalbantic, A.; Dzubur, A.; Dilic, M.; Pozderac, Z.; Mujanovic-Narancic, A.;
Kulic, M.; Hodzic, E.; Resic, N.; Brdjanovic, S.; Zvizdic, F. Bosnian journal of basic
medical sciences / Udruzenje basicnih mediciniskih znanosti = Association of Basic
Medical Sciences 2012, 12, (3), 164-168.
Elfström, N.; Karlström, A. E.; Linnros, J. Nano Lett 2008, 8, (3), 945-949.
Fiebig, E. W.; Wright, D. J.; Rawal, B. D.; Garrett, P. E.; Schumacher, R. T.; Peddada,
L.; Heldebrant, C.; Smith, R.; Conrad, A.; Kleinman, S. H.; Busch, M. P. AIDS 2003, 17,
(13), 1871-1879.
Forrest, S. R.; Burrows, P. E.; Thompson, M. E. Laser Focus World 1995, 31, (2), 99-107
Freitag, M.; Martin, Y.; Misewich, J. A.; Martel, R.; Avouris, P. H. Nano Lett 2003, 3,
(8), 1067-1071.
Fryberger, T. B.; Semancik, S. Sensor Actuat B-Chem 1990, 2, (4), 305-309.
Gao, A.; Lu, N.; Dai, P.; Li, T.; Pei, H.; Gao, X.; Gong, Y.; Wang, Y.; Fan, C. Nano Lett
2011, 11, (9), 3974-3978.

94
Gao, A.; Lu, N.; Wang, Y.; Dai, P.; Li, T.; Gao, X.; Fan, C. Nano Lett 2012, 12, (10),
5262-5268.
Gao, X. P.; Zheng, G.; Lieber, C. M. Nano letters 2010, 10, (2), 547-552.
Hahm, J.; Lieber, C. M. Nano letters 2004, 4, (1), 51-54.
Hakim, M. M. A.; Lombardini, M.; Sun, K.; Giustiniano, F.; Roach, P. L.; Davies, D. E.;
Howarth, P. H.; de Planque, M. R. R.; Morgan, H.; Ashburn, P. Nano Lett 2012, 12, (4),
1868-1872.
Han, S.; Jin, W.; Zhang, D. H.; Tang, T.; Li, C.; Liu, X. L.; Liu, Z. Q.; Lei, B.; Zhou, C.
W. Chem Phys Lett 2004, 389, (1-3), 176-180.
Heller, I.; Janssens, A. M.; Mannik, J.; Minot, E. D.; Lemay, S. G.; Dekker, C. Nano Lett
2008, 8, (2), 591-595.
Huang, I. Y.; Huang, R.-S.; Lo, L.-H. Sensors and Actuators B: Chemical 2003, 94, (1),
53-64.
Huang, Y.; Duan, X.; Wei, Q.; Lieber, C. M. Science 2001, 291, (5504), 630-633.
Ishikawa, F. N.; Chang, H.-K.; Curreli, M.; Liao, H.-I.; Olson, C. A.; Chen, P.-C.; Zhang,
R.; Roberts, R. W.; Sun, R.; Cote, R. J.; Thompson, M. E.; Zhou, C. ACS nano 2009, 3,
(5), 1219-1224.
Ishikawa, F. N.; Curreli, M.; Chang, H. K.; Chen, P. C.; Zhang, R.; Cote, R. J.;
Thompson, M. E.; Zhou, C. ACS nano 2009, 3, (12), 3969-3976.
Jin, S.; Whang, D.; McAlpine, M. C.; Friedman, R. S.; Wu, Y.; Lieber, C. M. Nano Lett
2004, 4, (5), 915-919.
Jo, K.; Yu, H. Z.; Yang, H. Electrochim Acta 2011, 56, (13), 4828-4833.

95
Ju, S.; Facchetti, A.; Xuan, Y.; Liu, J.; Ishikawa, F.; Ye, P.; Zhou, C.; Marks, T. J.; Janes,
D. B. Nature nanotechnology 2007, 2, (6), 378-384.
Juhasz, R.; Elfström, N.; Linnros, J. Nano Lett 2004, 5, (2), 275-280.
Keller, T.; Zeller, T.; Ojeda, F.; Tzikas, S.; Lillpopp, L.; Sinning, C.; Wild, P.; Genth-
Zotz, S.; Warnholtz, A.; Giannitsis, E.; Mockel, M.; Bickel, C.; Peetz, D.; Lackner, K.;
Baldus, S.; Munzel, T.; Blankenberg, S. JAMA : the journal of the American Medical
Association 2011, 306, (24), 2684-2693.
Kim, F.; Kwan, S.; Akana, J.; Yang, P. Journal of the American Chemical Society 2001,
123, (18), 4360-4361.
Kim, I. D.; Rothschild, A.; Lee, B. H.; Kim, D. Y.; Jo, S. M.; Tuller, H. L. Nano Lett
2006, 6, (9), 2009-2013.
Kim, Y. S.; Hwang, I. S.; Kim, S. J.; Lee, C. Y.; Lee, J. H. Sensor Actuat B-Chem 2008,
135, (1), 298-303.
Kolmakov, A.; Klenov, D. O.; Lilach, Y.; Stemmer, S.; Moskovits, M. Nano Lett 2005, 5,
(4), 667-673.
Kolmakov, A.; Zhang, Y.; Cheng, G.; Moskovits, M. Advanced Materials 2003, 15, (12),
997-1000.
Kong, J.; Franklin, N. R.; Zhou, C. W.; Chapline, M. G.; Peng, S.; Cho, K. J.; Dai, H. J.
Science 2000, 287, (5453), 622-625.
Kong, T.; Su, R.; Zhang, B.; Zhang, Q.; Cheng, G. Biosensors & bioelectronics 2012, 34,
(1), 267-272.
Kordos, P.; Gregusova, D.; Stoklas, R.; Cico, K.; Novak, J. Applied Physics Letters 2007,
90, (12).

96
Krebs, H. A. Annual Review of Biochemistry 1950, 19, (1), 409-430.
Law, M.; Kind, H.; Messer, B.; Kim, F.; Yang, P. D. Angew Chem Int Edit 2002, 41,
(13), 2405-2408.
Lee, I.; Luo, X.; Huang, J.; Cui, X. T.; Yun, M. Biosensors 2012, 2, (2), 205-220.
Lee, K. B.; Kim, E.-Y.; Mirkin, C. A.; Wolinsky, S. M. Nano Lett 2004, 4, (10), 1869-
1872.
Lei, B.; Li, C.; Zhang, D.; Zhou, Q. F.; Shung, K. K.; Zhou, C. Applied Physics Letters
2004, 84, (22), 4553-4555.
Leonardy, A.; Hung, W.-Z.; Tsai, D.-S.; Chou, C.-C.; Huang, Y.-S. Crystal Growth &
Design 2009, 9, (9), 3958-3963.
Lewandrowski, K.; Chen, A.; Januzzi, J. American journal of clinical pathology 2002,
118 Suppl, S93-99.
Li, C.; Curreli, M.; Lin, H.; Lei, B.; Ishikawa, F. N.; Datar, R.; Cote, R. J.; Thompson, M.
E.; Zhou, C. Journal of the American Chemical Society 2005, 127, (36), 12484-12485.
Li, C.; Lei, B.; Zhang, D.; Liu, X.; Han, S.; Tang, T.; Rouhanizadeh, M.; Hsiai, T.; Zhou,
C. Applied Physics Letters 2003, 83, (19), 4014-4016.
Li, C.; Zhang, D.; Han, S.; Liu, X. L.; Tang, T.; Zhou, C. W. Advanced Materials 2003,
15, (2), 143-146.
Li, C.; Zhang, D.; Han, S.; Liu, X.; Tang, T.; Lei, B.; Liu, Z.; Zhou, C. Ann Ny Acad Sci
2003, 1006, 104-121.
Lin, C. H.; Hsiao, C. Y.; Hung, C. H.; Lo, Y. R.; Lee, C. C.; Su, C. J.; Lin, H. C.; Ko, F.
H.; Huang, T. Y.; Yang, Y. S. Chem Commun (Camb) 2008, (44), 5749-5751.
Liu, Y.; Sauer, J.; Dutton, R. W. Journal of Applied Physics 2008, 103, (8), 084701-1-4.

97
Liu, Z. Q.; Zhang, D. H.; Han, S.; Li, C.; Tang, T.; Jin, W.; Liu, X. L.; Lei, B.; Zhou, C.
W. Advanced Materials 2003, 15, (20), 1754-1757.
Lu, N.; Gao, A.; Dai, P.; Li, T.; Wang, Y.; Gao, X.; Song, S.; Fan, C. Methods 2013, 63,
(3), 212-218.
Lu, N.; Gao, A.; Dai, P.; Song, S.; Fan, C.; Wang, Y.; Li, T. Small 2014, 10, (10), 2022-
2028.
Maisel, A. S.; Krishnaswamy, P.; Nowak, R. M.; McCord, J.; Hollander, J. E.; Duc, P.;
Omland, T.; Storrow, A. B.; Abraham, W. T.; Wu, A. H.; Clopton, P.; Steg, P. G.;
Westheim, A.; Knudsen, C. W.; Perez, A.; Kazanegra, R.; Herrmann, H. C.; McCullough,
P. A. The New England journal of medicine 2002, 347, (3), 161-167.
Maiti, A.; Rodriguez, J. A.; Law, M.; Kung, P.; McKinney, J. R.; Yang, P. D. Nano Lett
2003, 3, (8), 1025-1028.
Martel, R.; Schmidt, T.; Shea, H. R.; Hertel, T.; Avouris, P. Applied Physics Letters
1998, 73, (17), 2447-2449.
Matsumoto, T.; Ohashi, A.; Ito, N. Analytica Chimica Acta 2002, 462, (2), 253-259.
Mattila, M.; Hakkarainen, T.; Jiang, H.; Kauppinen, E. I.; Lipsanen, H. Nanotechnology
2007, 18, (15).
McCullough, P. A.; Nowak, R. M.; McCord, J.; Hollander, J. E.; Herrmann, H. C.; Steg,
P. G.; Duc, P.; Westheim, A.; Omland, T.; Knudsen, C. W.; Storrow, A. B.; Abraham, W.
T.; Lamba, S.; Wu, A. H.; Perez, A.; Clopton, P.; Krishnaswamy, P.; Kazanegra, R.;
Maisel, A. S. Circulation 2002, 106, (4), 416-422.
McLean, A. S.; Huang, S. J. Annals of intensive care 2012, 2, 8.
Melosh, N. A.; Boukai, A.; Diana, F.; Gerardot, B.; Badolato, A.; Petroff, P. M.; Heath, J.
R. Science 2003, 300, (5616), 112-115.

98
Minot, E. D.; Janssens, A. M.; Heller, I.; Heering, H. A.; Dekker, C.; Lemay, S. G.
Applied Physics Letters 2007, 91, (9), 093507 - 093507-3
Morales, A. M.; Lieber, C. M. Science 1998, 279, (5348), 208-211.
Morita, E.; Yasue, H.; Yoshimura, M.; Ogawa, H.; Jougasaki, M.; Matsumura, T.;
Mukoyama, M.; Nakao, K. Circulation 1993, 88, (1), 82-91.
Nikoobakht, B.; Michaels, C. A.; Stranick, S. J.; Vaudin, M. D. Applied Physics Letters
2004, 85, (15), 3244-3246.
Park, I.; Li, Z.; Pisano, A. P.; Williams, R. S. Nanotechnology 2010, 21, (1), 015501.
Patolsky, F.; Zheng, G. F.; Hayden, O.; Lakadamyali, M.; Zhuang, X. W.; Lieber, C. M.
Proceedings of the National Academy of Sciences of the United States of America 2004,
101, (39), 14017-14022.
Patolsky, F.; Zheng, G.; Lieber, C. M. Nanomedicine (Lond) 2006, 1, (1), 51-65.
Patolsky, F.; Zheng, G.; Lieber, C. M. Nature protocols 2006, 1, (4), 1711-1724.
Pengfei, Q. F.; Vermesh, O.; Grecu, M.; Javey, A.; Wang, O.; Dai, H. J.; Peng, S.; Cho,
K. J. Nano Lett 2003, 3, (3), 347-351.
Persson, A. I.; Larsson, M. W.; Stenstrom, S.; Ohlsson, B. J.; Samuelson, L.; Wallenberg,
L. R. Nature materials 2004, 3, (10), 677-681.
Pud, S.; Li, J.; Sibiliev, V.; Petrychuk, M.; Kovalenko, V.; Offenhausser, A.; Vitusevich,
S. Nano Lett 2014, 14, (2), 578-584.
Puleo, P. R.; Meyer, D.; Wathen, C.; Tawa, C. B.; Wheeler, S.; Hamburg, R. J.; Ali, N.;
Obermueller, S. D.; Triana, J. F.; Zimmerman, J. L.; et al. The New England journal of
medicine 1994, 331, (9), 561-566.

99
Qureshi, A.; Gurbuz, Y.; Niazi, J. H. Sensors and Actuators B: Chemical 2012, 171–172,
(0), 62-76.
Richards, A. M.; Nicholls, M. G.; Espiner, E. A.; Lainchbury, J. G.; Troughton, R. W.;
Elliott, J.; Frampton, C.; Turner, J.; Crozier, I. G.; Yandle, T. G. Circulation 2003, 107,
(22), 2786-2792.
Rout, C. S.; Ganesh, K.; Govindaraj, A.; Rao, C. N. R. Appl Phys a-Mater 2006, 85, (3),
241-246.
Rutstein, S. E.; Kamwendo, D.; Lugali, L.; Thengolose, I.; Tegha, G.; Fiscus, S. A.;
Nelson, J. A. E.; Hosseinipour, M. C.; Sarr, A.; Gupta, S.; Chimbwandira, F.; Mwenda,
R.; Mataya, R. J Clin Virol 2014, 60, (4), 392-398.
Ryu, K.; Badmaev, A.; Wang, C.; Lin, A.; Patil, N.; Gomez, L.; Kumar, A.; Mitra, S.;
Wong, H. S. P.; Zhou, C. W. Nano Lett 2009, 9, (1), 189-197.
Ryu, Y. K.; Chiesa, M.; Garcia, R. Nanotechnology 2013, 24, (31), 315205.
Schockmel, G. A.; Yerly, S.; Perrin, L. J Acq Immun Def Synd 1997, 14, (2), 179-183.
Semancik, S.; Cavicchi, R. E. Thin Solid Films 1991, 206, (1–2), 81-87.
Sheats, J. R.; Antoniadis, H.; Hueschen, M.; Leonard, W.; Miller, J.; Moon, R.; Roitman,
D.; Stocking, A. Science 1996, 273, (5277), 884-888.
Smith, P. A.; Nordquist, C. D.; Jackson, T. N.; Mayer, T. S.; Martin, B. R.; Mbindyo, J.;
Mallouk, T. E. Applied Physics Letters 2000, 77, (9), 1399-1401.
Stern, E.; Klemic, J. F.; Routenberg, D. A.; Wyrembak, P. N.; Turner-Evans, D. B.;
Hamilton, A. D.; LaVan, D. A.; Fahmy, T. M.; Reed, M. A. Nature 2007, 445, (7127),
519-522.

100
Stern, E.; Vacic, A.; Li, C.; Ishikawa, F. N.; Zhou, C.; Reed, M. A.; Fahmy, T. M. Small
2010, 6, (2), 232-238.
Stern, E.; Wagner, R.; Sigworth, F. J.; Breaker, R.; Fahmy, T. M.; Reed, M. A. Nano
letters 2007, 7, (11), 3405-3409.
Sutthent, R.; Gaudart, N.; Chokpaibulkit, K.; Tanliang, N.; Kanoksinsombath, C.;
Chaisilwatana, P. Journal of clinical microbiology 2003, 41, (3), 1016-1022.
Tang, S.; Hewlett, I. The Journal of infectious diseases 2010, 201 Suppl 1, S59-64.
Tang, X.; Bansaruntip, S.; Nakayama, N.; Yenilmez, E.; Chang, Y. L.; Wang, Q. Nano
letters 2006, 6, (8), 1632-1636.
Tarasov, A.; Wipf, M.; Stoop, R. L.; Bedner, K.; Fu, W.; Guzenko, V. A.; Knopfmacher,
O.; Calame, M.; Schonenberger, C. ACS nano 2012, 6, (10), 9291-9298.
Tsivion, D.; Schvartzman, M.; Popovitz-Biro, R.; Joselevich, E. ACS nano 2012, 6, (7),
6433-6445.
Tsivion, D.; Schvartzman, M.; Popovitz-Biro, R.; von Huth, P.; Joselevich, E. Science
2011, 333, (6045), 1003-1007.
Vacic, A.; Criscione, J. M.; Stern, E.; Rajan, N. K.; Fahmy, T.; Reed, M. A. Biosensors
& bioelectronics 2011, 28, (1), 239-242.
Vayssieres, L. Advanced Materials 2003, 15, (5), 464-466.
Wang, D.; Sheriff, B. A.; Heath, J. R. Nano letters 2006, 6, (6), 1096-1100.
Wang, X.; Aroonyadet, N.; Zhang, Y.; Mecklenburg, M.; Fang, X.; Chen, H.; Goo, E.;
Zhou, C. Nano letters 2014, 14, (6), 3014-3022.
Wang, Z. L.; Song, J. Science 2006, 312, (5771), 242-246.

101
Wang, Z.; Liu, J.; Chen, X.; Wan, J.; Qian, Y. Chemistry 2004, 11, (1), 160-163.
Whang, D.; Jin, S.; Wu, Y.; Lieber, C. M. Nano Lett 2003, 3, (9), 1255-1259.
Xin, H.; Kim, F. S.; Jenekhe, S. A. Journal of the American Chemical Society 2008, 130,
(16), 5424-5425.
Yao, J.; Yan, H.; Lieber, C. M. Nature nanotechnology 2013, 8, (5), 329-335.
Yu, G.; Cao, A.; Lieber, C. M. Nature nanotechnology 2007, 2, (6), 372-377.
Yuan, Z. J.; Zhu, X. M.; Wang, X. O.; Cai, X. K.; Zhang, B. P.; Qiu, D. J.; Wu, H. Z.
Thin Solid Films 2011, 519, (10), 3254-3258.
Zhang, D. H.; Liu, Z. Q.; Li, C.; Tang, T.; Liu, X. L.; Han, S.; Lei, B.; Zhou, C. W. Nano
letters 2004, 4, (10), 1919-1924.
Zhang, D.; Li, C.; Han, S.; Liu, X.; Tang, T.; Jin, W.; Zhou, C. Appl Phys a-Mater 2003,
77, (1), 163-166.
Zhang, D.; Liu, Z. Q.; Li, C.; Tang, T.; Liu, X. L.; Han, S.; Lei, B.; Zhou, C. W. Nano
Lett 2004, 4, (10), 1919-1924.
Zhang, G. J.; Chai, K. T. C.; Luo, H. Z. H.; Huang, J. M.; Tay, I. G. K.; Lim, A. E.-J.; Je,
M. Biosensors and Bioelectronics 2012, 35, (1), 218-223.
Zhang, G. J.; Luo, Z. H. H.; Huang, M. J.; Ang, J. A. J.; Kang, T. G.; Ji, H. Biosensors
and Bioelectronics 2011, 28, (1), 459-463.
Zheng, G.; Patolsky, F.; Cui, Y.; Wang, W. U.; Lieber, C. M. Nature biotechnology 2005,
23, (10), 1294-1301.
Zheng, L.; Jia, L.; Li, B.; Situ, B.; Liu, Q.; Wang, Q.; Gan, N. Molecules 2012, 17, (5),
5988-6000.

102
Zhou, C.; Kong, J.; Yenilmez, E.; Dai, H. J. Science 2000, 290, (5496), 1552-1555.
Zhou, W.; Dai, X. C.; Fu, T. M.; Xie, C.; Liu, J.; Lieber, C. M. Nano Lett 2014, 14, (3),
1614-1619.
Zi, Y.; Jung, K.; Zakharov, D.; Yang, C. Nano letters 2013.

103
Appendix I: Top-Down Polysilicon Nanoribbon Biosensors
I.1 Materials
3” 500 nm SiO
2
on Si wafers and 4” Si wafers were purchased from SQI. Au and
Ti for metal sources of electron beam evaporation were purchased from Plasmaterials.
Borosilicafilm solutions with doping concentration 1x10
17
, 5x10
17
, and 1x10
18
cm
-3
were
purchased from Emulsitone Chemicals. 3-Aminopropyldimethylethoxysilane (APDMS)
was purchased from Gelest. Succinic anhydride with purity 99%, triethylamine with
purity of 99%, N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC)
with purity of 98%, and N-Hydroxysuccinimide (NHS) with purity of 98 % were
purchased from Sigma Aldrich. Streptavidin, Alexa Fluor® 568 conjugate was purchased
from Life Technologies. Amine-PEG3-Biotin was purchased from Thermo Scientific.
Cancer antigen 125 (CA-125) proteins and monoclonal antibodies were purchased from
Fitzgerald Industries. Amine PEG was purchased from Nanonocs.
I.2 Polysilicon nanoribbon field effect transistor fabrication
500 nm Si
3
N
4
on Si wafers and 55 nm polysilicon (poly-Si) were deposited by
low-pressure chemical vapor deposition (LPCVD) system in Nanoelectronic Research
Facilities (NRF), University of California at Los Angeles. After that, the poly-Si on
Si
3
N
4
/Si wafer was rinsed with acetone and isopropyl alcohol before dried in nitrogen
stream. After solvent cleaning, the poly-Si on Si
3
N
4
/Si substrate was placed on a hot
place at 120 ºC for 5 minutes to repel all solvent residual and cool down in room
temperature. After cleaning process, the poly-Si film was drop-coated with desired
doping concentration of Borosilica solution and spun at 3000 rpm for 30s. To diffuse
Boron (B) dopants from the surface into the poly-Si film, thermal annealing was
performed at 1100ºC in N
2
environment for 15 minutes. To remove Borosilica glass on
top of the poly-Si layer, the poly-Si on substrate was dipped into 7:1 buffer oxide etchant
(BOE) for 4 minutes and followed by extensively rinsing with deionized water and

104
drying in N
2
stream. Active mesa was defined by photolithography before CF
4
dry
etching at 60 W, 100 mTorr for 1.1 minutes to remove unwanted area. After patterning
poly-Si nanoribbons, source and drain electrodes were defined by photolithography. O
2

discum was treated at 60 W 50 mTorr for 18 s to remove thin layer of photoresist residual
and followed by BOE dipping to remove a thin layer of native oxide before metal
deposition for better metal contact. The Ti/Au/Ti metals with 5/45/5 nm thickness were
deposited by electron beam evaporation. After device fabrication, thin layer of Si
3
N
4

about 40 nm was deposited by plasma enhanced chemical vapor deposition (PECVD) in
NRF for passivation layer if it is needed before the lift-off process. To open metal pads
for electrical contact, photoresist was coated on top of nanoribbon transistors before CF
4

dry etching at 100 W 100 mTorr for 1 minute. Lastly, coated photoresist was rinsed by
acetone and isopropyl alcohol and dried in N
2
stream. To stabilize and enhance surface
chemistry to immobilize bioprobes, thin layer of SiO
2
(10nm) was oxidized by dry
thermal oxidation, but device performance can be changed due to high temperature
during dry oxidation. An alternative approach is to pattern a trench on the nanoribbon
area and deposits 10 nm SiO
2
using electron beam evaporation. Lift-off process will be
applied to remove unwanted area of SiO
2

I.3 Surface functionalization for poly-Si nanoribbon biosensors
After device fabrication, poly-Si nanoribbon FETs were submerged in boiling
acetone and isopropyl alcohol for 5 minutes before dried in N
2
stream. To generate
hydroxyl groups on the surface of poly-Si to accommodate aminosilane linker molecules,
devices were treated by O
2
asher at 100 W 150 mTorr for 40 s as shown in Figure I.1 (a).
To anchor amine terminal (Figure I.1 (b)), devices were incubated in 5% 3-APDMS in
anhydrous toluene for 2 hours in N
2
environment. After 2 hours, devices were extensively
rinsed with toluene and tetrahydrofuran (THF) before they were annealed at 120 ºC in N
2

environment for 12 hours to dehydrate surface and to reinforce amino-silane groups on
the poly-Si. To generate carboxylic terminals, devices were immerged in mixture of 5
mg/ml of succinic anhydride in anhydrous THF and 5% triethylamine for 4 hours as

105
shown in Figure I.1 (c). Before incubating with amine terminal molecules, devices were
rinsed with THF and then deionized (DI) water. To convert from carboxylic acid
functional groups to NHS ester groups for covalently binding of amine molecules on the
poly-Si surface, devices were treated with mixture of 20 mM EDC and 5 mM NHS in DI
water for 1 hour before rinsing with DI water. In Figure I.1 (d), amine molecules (such as
antibody) in 10 mM phosphate buffer saline (PBS) solution pH 7.4 were reacted with
NHS ester functional groups and were bound on the poly-Si surface to increase
specificity and selectivity toward target molecules.

Figure I.1 Schematic diagram of surface functionalization for poly-Si nanoribbon biosensors (a)
Hydroxyl groups were generated by O
2
plasma treatment (b) Amine termination was anchored by
immersing in mixture of 5% 3-APDMS in anhydrous toluene for 2 hours (c) Carboxyl function group
was generated by incubating in 5mg/ml of succinic anhydride in anhydrous THF and 5% triethylamine
for 4 hours (d) Antibody was immobilized by interacting with NHS ester group after conversion of
carboxylic group to NHS ester by EDC/NHS.

106
Appendix II: Highly Scalable, Uniform, and Sensitive
Biosensors Based on Top-Down Indium Oxide Nanoribbons
and Electronic Enzyme-Linked Immunosorbent Assay for
Detection of Human Immunodeficiency Virus p24 Proteins
II.1 Materials
3” 500 nm SiO
2
on Si wafers and 4” Si wafers were purchased from SQI. LOL
2000 and 3612 photoresist were purchased from Shipley. Au and Ti for metal sources of
electron beam evaporation and an indium oxide (In
2
O
3
) sputtering target with purity of
99.99% were obtianed from Plasmaterials. 3-Phosphonopropioninc acid with purity of 94
%, N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC) with purity
of 98%, and N-Hydroxysuccinimide (NHS) with purity of 98 % were purchased from
Sigma Aldrich. Streptavidin, Alexa Fluor 568 conjugate was purchased from Life
Technologies. Amine-PEG3-Biotin was purchased from Thermo Scientific. Human
immunodeficiency virus 1 (HIV1) p24 antibodies, HIV1 p24 proteins, and biotinylated
HIV1 p24 antibodies were purchased from Fitzgerald Industries. Amine PEG was
purchased from Nanonocs. Streptavidin conjugated with 20 nm Au nanoparticles were
purchased from Ted Pella.
II.2 Indium oxide nanoribbon field effect transistor fabrication
500 nm Si
3
N
4
on Si wafers were deposited by low pressure chemical vapor
deposition (LPCVD) system in Nanoelectronic Research Facilities (NRF), University of
California at Los Angeles. After that, Si
3
N
4
/Si wafer was rinsed with acetone and
isopropyl alcohol before dried in nitrogen stream before the fabrication process. After
solvent cleaning, the Si
3
N
4
/Si substrate was placed on a hot place at 120 ºC for 5 minutes
to repel all solvent residual and cool down in room temperature. After cleaning process
and coating with bi-layered photoresist (LOL 2000 and SPR 3612 photoresist), source

107
and drain electrodes were patterned using and standard photolithography. O
2
discum was
treated at 60 W 50 mTorr for 18 s to remove thin layer of photoresist residual before
metal deposition for better metal contact. Then, Ti/Au metal with 5/45 nm thickness were
deposited by electron beam evaporation and lift-off process was used to remove
unwanted areas. After coating with bi-layered photoresist, the second mask was used to
define nanoribbons by traditional photolithographic process. After nanoribbon patterning,
O
2
discum was treated at 60 W 50 mTorr for 18 s to remove thin layer of photoresist
residual before In
2
O
3
sputtering. In
2
O
3
nanoribbons were sputtered by Denton II
sputtering system in NRF. After sputtering, unwanted In
2
O
3
was removed by the lift-off
process yielding the pristine surface.
II.3 Surface functionalization for In
2
O
3
nanoribbon biosensors
After device fabrication, In
2
O
3
nanoribbon FETs were submerged in boiling
acetone and isopropyl alcohol for 5 minutes before dried in N
2
stream. To generate
hydroxyl groups on the surface of In
2
O
3
nanoribbon to accommodate phosphonic acid
linker molecules, devices were treated by O
2
asher at 100 W 150 mTorr for 40 s as shown
in Figure II.2 (a). After cleaning, devices were incubated in 1 mM aqueous solution of 3-
Phosphonopropioninc acid for 5.5 hours. After liker molecule incubation, devices were
extensively rinsed with deionized (DI) water before they were annealed at 120 ºC in N
2

environment for 12 hours to dehydrate surface and to reinforce linkers on the In
2
O
3

nanoribbon surface as shown in Figure II.2 (b). To convert from carboxylic acid
functional groups to NHS ester groups for covalently binding of amine molecules on the
nanoribbon surface shown in Figure II.1 (c), devices were treated with mixture of 20 mM
EDC and 5 mM NHS in DI water for 1 hour before rinsing with DI water. In Figure II.2
(d), amine molecules (such as antibodies) in 10 mM phosphate buffer saline (PBS)
solution pH 7.4 were reacted with NHS ester functional groups and were bound on the
nanoribbon surface to increase specificity and selectivity toward target molecules.

108

Figure II.1 Schematic diagram of In
2
O
3
surface chemistry to immobilize antibodies or amine
molecules
II.4 Synthesis of biotinylated phosphonic acid linker
Figure II.3 shows a schematic diagram of process to synthesize biotinylated
phosphonic acid linker molecules. The process started with synthesis of 6-
(diethoxyphosphoryl)hexyl 5 - (2 - oxohexahydro - 1H-thieno [3, 4 - d] imidazol – 4 - yl)
pentanoate (product 1). A stirred solution of diethyl (6-hydroxyhexyl)phosphonate
(0.952g, 4 mmole) in dry Dimethylformamide (DMF) (30 mL) was mixed with biotin
powder (0.813g 3.3 mmole), EDC•HCl (766 mg, 4 mmole) and 4-
Dimethylaminopyridine (DMAP) (488 mg, 4 mmole). The reaction mixture was stirred at
room temperature under inert atmosphere for 2 hours and poured into water. The aqueous
layer was extracted with ethyl acetate. The organic layer was washed with water, 1 M
sodium hydroxide (NaOH) aqueous solution, 1 M hydrocholic acid (HCl) aqueous
solution and brine, dried over Na
2
SO
4
and concentrated in vacuo to give the crude oil.
The crude residue was purified by flash chromatography (Dichloromethane/Methanol
alcohol, DCM/MeOH, 10:1) to give the purified of product 1 (0.777g, 1.6 mmole, 50%).
Synthesis of (6-((5-(2-oxohexahydro-1H-thieno[3,4-d]imidazol-4-yl) pentanoyl)
oxy) hexyl) phosphonic acid (product 2) started with a stirred solution of product 1 (0.1
g, 0.2 mmole) in dry DCM (10 mL). Bromotrimethylsilane (0.113 g, 0.75 mmole) was

109
added into a stirred solution under inert atmosphere. The reaction mixture was stirred at
room temperature for overnight and volatiles were evaporated in vacuo. 10 mL methanol
was added into the residue and the mixture was stirred for 2 hours at room temperature.
The reaction mixture was evaporated under in vacuo to give product 2 (quantitative) as a
hygroscopic solid.

Figure II.2 Schematic diagram of synthesis of biotinylated phosphonic acid linker molecules.
II.4 In
2
O
3
nanoribbon transistor Debye length calculation
Sensing response is related to gating effect from binding of target molecules on
the nanoribbon surface. A distance from the surface to which charged molecules bind
affects electrical conduction of the nanoribbon device defined by Debye length ($

%&
'
(/)

<
'
) where ε, k
B
, T, q, and N
B
stand for permittivity of In
2
O
3
(7.97x10
-13

F/cm)
65
, Boltzmann's constant, temperature, charge constant and charge density,
respectively.
33
To have good sensitivity, the optimal nanoribbon thickness needs to be
within the transistor Debye length. Figure II.3 (a) shows a plot of Debye length and
charge density in the In
2
O
3
nanoribbon device. From I-V
DS
plot in Figure II.3 (b) at V
GS
=
12 V, total resistance (R) is approximate to be only channel resistance (R = 40 MΩ) due

110
to a small vale of contact resistance (R
C
). From the nanoribbon structure, we have
calculated (A/l)
eff
to be 2.67x10
-7
cm. Therefore, resistivity (ρ) is calculated from ρ=
R(A/l)
eff
= 10.8 Ω.cm. From average electron mobility (µ) of In
2
O
3
nanoribbon devices
shown in Figure 3.4 (c) about 23.38 cm
2
/V•s, electron density (n) is calculated from
* 1/-C = 2.48x10
16
cm
-3
where e is electron charge in C. This yields $

to be
approximate 23 nm.

Figure II.3 (a) Plot of Debye length versus charge density in the In
2
O
3
nanoribbon device (b) I
DS
-V
DS

in the narrow linear regime with V
GS
from 27 to 9 V with step of 3 V.

111
Appendix III: Highly Scalable, Uniform and Sensitive In
2
O
3

Nanoribbon Biosensors for Myocardial Infarction
III.1 Materials
3” 500 nm SiO
2
on Si wafers and 4” Si wafers were purchased from SQI. LOL
2000 and 3612 photoresist were purchased from Shipley. Au and Ti for metal sources of
electron beam evaporation and an indium oxide (In
2
O
3
) sputtering target with purity of
99.99% were obtianed from Plasmaterials. 3-Phosphonopropioninc acid with purity of 94
%, N-(3-Dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC) with purity
of 98%, and N-Hydroxysuccinimide (NHS) with purity of 98 % were purchased from
Sigma Aldrich. Streptavidin was purchased from Life Technologies. Amine-PEG3-Biotin
and biotinylation kit was purchased from Thermo Scientific. Troponin I antibody,
biotinylated troponin I antibody, troponin protein, creatine kinase MB (CK-MB) antibody
pair, CK-MB protein, Brain natriuretic peptide (BNP) and BNP antibody were purchase
from Fitzgerald. Biotinylated BNP antibody was purchased from Abcam.

112
Appendix IV: Epitaxial Growth of Aligned SnO
2
Nanowires
on Sapphire and Their Device Applications
IV.1 Materials
A-plane, M-plane, and R-plane sapphire wafers with 2-inch diameter were
purchased from University Wafers with the primary flats oriented in C plane, C plane,
and A plane, respectively. Sn powder precursor was purchased from Alfa Aesar, and
purities of both 99.99% and 99.995% were used. Au catalysts were deposited by
evaporating a metal source purchased from Plasmaterials. The Ar used as the carrier gas
during nanowire synthesis was purchased as ultra high purity Ar (99.999%) from
Gilmore Liquid and Gas Company. NO
2
gas used for chemical sensing was purchased as
1030 ppm NO
2
balanced with Ar from Airgas.
IV.2 Aligned SnO
2
nanowire synthesis
The sapphire substrates were cleaned by sonication in acetone, followed by
rinsing in acetone, isopropyl alcohol, and distilled water, then dried in a gas stream.
Ambient annealing of M-plane sapphire substrates was done at 1400°C
98
in a 1500°C
maximum tube furnace from Lindberg/Blue M with a ramping rate of 23°C/min. An
1100°C one-inch tube furnace from Lindberg/Blue M was used to anneal the A-plane
sapphire in air.
110
Annealing time is 1.5 hours. After annealing, stripes of 5 µm by 5-30
µm, and about 100 µm apart were perpendicularly patterned to the primary flat of each
substrate so that the catalyst stripes are perpendicular to the direction of aligned SnO
2

nanowire growth. An Au catalyst thickness of 3 nm to 6 nm was confirmed using an
atomic force microscope (AFM). Before nanowire synthesis, substrates with Au catalyst
were placed into the one-inch furnace to complete a thermal de-wetting step at 800°C for
15 minutes in ambient air. The synthesis of aligned SnO
2
nanowires was carried out in a
horizontal quartz tube in the one-inch furnace. The Sn powder precursor was placed
upstream in a quartz boat. The different planes of sapphire substrates with the de-wetted

113
Au catalysts were placed in the same quartz boat downstream of the precursor. The boat
was positioned in the center of the one-inch quartz furnace tube and then heated to 900°C
for one hour at a ramping rate of approximately 45 degrees per minute. Argon was used
as the carrier gas, and a mechanical pump with a pumping rate of 4 m
3
/h was used to
sustain the pressure inside the furnace at around 200 Torr for growing aligned SnO
2

nanowires.
IV.3 Aligned SnO
2
nanowire field effect transistor (FET) fabrication
Sapphire substrates containing aligned SnO
2
nanowires were sonicated in
isopropyl alcohol (IPA) to remove the small amounts of free-standing un-aligned SnO
2

nanowires on the surface. Following substrate cleaning, source and drain electrodes were
patterned on the substrates using bi-layered photoresist and standard photolithography.
The metals Ti and Au were then deposited into the electrode patterns through electron
beam evaporation at a thickness of 5 nm and 75 nm respectively. Afterwards, an uniform
layer of 40 nm aluminum oxide was deposited as gate dielectric through atomic layer
deposition (ALD). Next, photolithography was used again to pattern the gate electrode,
which was deposited using the same process as source and drain deposition. Lastly,
buffered oxide etch (BOE) was used to remove aluminum oxide over the source and drain
in order to form contacts to the electrodes.
Scanning electron microscopy (SEM) images of SnO
2
nanowires and devices
were taken using a Hitachi field-emission scanning electron microscope at a voltage of 1
kV. Surface scans of sapphire was made using tapping AFM in tapping mode. Current
measurements for FET, photoconduction, and chemical sensing experiments were done
using semiconductor analyzer Agilent 4156B.

114
Appendix V: References
1. Duan, X.; Lieber, C. M. Advanced Materials 2000, 12, (4), 298-302.
2. Yao, J.; Yan, H.; Lieber, C. M. Nature nanotechnology 2013, 8, (5), 329-335.
3. Li, C.; Zhang, D.; Han, S.; Liu, X.; Tang, T.; Lei, B.; Liu, Z.; Zhou, C. Ann Ny
Acad Sci 2003, 1006, 104-121.
4. Zhang, D.; Li, C.; Han, S.; Liu, X.; Tang, T.; Jin, W.; Zhou, C. Appl Phys a-Mater
2003, 77, (1), 163-166.
5. Curreli, M.; Rui, Z.; Ishikawa, F. N.; Chang, H.-K.; Cote, R. J.; Chongwu, Z.;
Thompson, M. E. Nanotechnology, IEEE Transactions on 2008, 7, (6), 651-667.
6. Cui, Y.; Wei, Q.; Park, H.; Lieber, C. M. Science 2001, 293, (5533), 1289-1292.
7. Wang, Z. L.; Song, J. Science 2006, 312, (5771), 242-246.
8. Xin, H.; Kim, F. S.; Jenekhe, S. A. Journal of the American Chemical Society
2008, 130, (16), 5424-5425.
9. Li, C.; Curreli, M.; Lin, H.; Lei, B.; Ishikawa, F. N.; Datar, R.; Cote, R. J.;
Thompson, M. E.; Zhou, C. Journal of the American Chemical Society 2005, 127,
(36), 12484-12485.
10. Zhang, D. ; Liu, Z. Q.; Li, C.; Tang, T.; Liu, X. L.; Han, S.; Lei, B.; Zhou, C. W.
Nano letters 2004, 4, (10), 1919-1924.
11. Wang, Z.; Liu, J.; Chen, X.; Wan, J.; Qian, Y. Chemistry 2004, 11, (1), 160-163.
12. Morales, A. M.; Lieber, C. M. Science 1998, 279, (5348), 208-211.
13. Persson, A. I.; Larsson, M. W.; Stenstrom, S.; Ohlsson, B. J.; Samuelson, L.;
Wallenberg, L. R. Nature materials 2004, 3, (10), 677-681.
14. Li, C.; Zhang, D.; Han, S.; Liu, X. L.; Tang, T.; Zhou, C. W. Advanced Materials
2003, 15, (2), 143-146.
15. Wang, X.; Aroonyadet, N.; Zhang, Y.; Mecklenburg, M.; Fang, X.; Chen, H.;
Goo, E.; Zhou, C. Nano letters 2014, 14, (6), 3014-3022.
16. Chen, P. C.; Sukcharoenchoke, S.; Ryu, K.; de Arco, L. G.; Badmaev, A.; Wang,
C.; Zhou, C. W. Advanced Materials 2010, 22, (17), 1900-1904.

115
17. Patolsky, F.; Zheng, G.; Lieber, C. M. Nature protocols 2006, 1, (4), 1711-24.
18. Vayssieres, L. Advanced Materials 2003, 15, (5), 464-466.
19. Stern, E.; Klemic, J. F.; Routenberg, D. A.; Wyrembak, P. N.; Turner-Evans, D.
B.; Hamilton, A. D.; LaVan, D. A.; Fahmy, T. M.; Reed, M. A. Nature 2007, 445,
(7127), 519-522.
20. Heller, I.; Janssens, A. M.; Mannik, J.; Minot, E. D.; Lemay, S. G.; Dekker, C.
Nano Lett 2008, 8, (2), 591-595.
21. Stern, E.; Wagner, R.; Sigworth, F. J.; Breaker, R.; Fahmy, T. M.; Reed, M. A.
Nano letters 2007, 7, (11), 3405-3409.
22. Liu, Y.; Sauer, J.; Dutton, R. W. Journal of Applied Physics 2008, 103, (8),
084701-1-4.
23. Byon, H. R.; Choi, H. C. Journal of the American Chemical Society 2006, 128,
(7), 2188-2189.
24. Chen, R. J.; Bangsaruntip, S.; Drouvalakis, K. A.; Kam, N. W.; Shim, M.; Li, Y.;
Kim, W.; Utz, P. J.; Dai, H. Proceedings of the National Academy of Sciences of
the United States of America 2003, 100, (9), 4984-4989.
25. Tang, X.; Bansaruntip, S.; Nakayama, N.; Yenilmez, E.; Chang, Y. L.; Wang, Q.
Nano letters 2006, 6, (8), 1632-1636.
26. Minot, E. D.; Janssens, A. M.; Heller, I.; Heering, H. A.; Dekker, C.; Lemay, S.
G. Applied Physics Letters 2007, 91, (9), 093507 - 093507-3
27. Gao, X. P.; Zheng, G.; Lieber, C. M. Nano letters 2010, 10, (2), 547-552.
28. Zheng, G.; Patolsky, F.; Cui, Y.; Wang, W. U.; Lieber, C. M. Nature
biotechnology 2005, 23, (10), 1294-1301.
29. Patolsky, F.; Zheng, G.; Lieber, C. M. Nanomedicine (Lond) 2006, 1, (1), 51-65.
30. Ishikawa, F. N.; Curreli, M.; Chang, H. K.; Chen, P. C.; Zhang, R.; Cote, R. J.;
Thompson, M. E.; Zhou, C. ACS nano 2009, 3, (12), 3969-3976.
31. Chang, H. K.; Ishikawa, F. N.; Zhang, R.; Datar, R.; Cote, R. J.; Thompson, M.
E.; Zhou, C. ACS nano 2011, 5, (12), 9883-9891.

116
32. Ishikawa, F. N.; Chang, H.-K.; Curreli, M.; Liao, H.-I.; Olson, C. A.; Chen, P.-C.;
Zhang, R.; Roberts, R. W.; Sun, R.; Cote, R. J.; Thompson, M. E.; Zhou, C. ACS
nano 2009, 3, (5), 1219-1224.
33. Elfström, N.; Karlström, A. E.; Linnros, J. Nano Lett 2008, 8, (3), 945-949.
34. Duan, X. X.; Li, Y.; Rajan, N. K.; Routenberg, D. A.; Modis, Y.; Reed, M. A.
Nature nanotechnology 2012, 7, (6), 401-407.
35. Bunimovich, Y. L.; Shin, Y. S.; Yeo, W. S.; Amori, M.; Kwong, G.; Heath, J. R.
Journal of the American Chemical Society 2006, 128, (50), 16323-16331.
36. Pud, S.; Li, J.; Sibiliev, V.; Petrychuk, M.; Kovalenko, V.; Offenhausser, A.;
Vitusevich, S. Nano Lett 2014, 14, (2), 578-584.
37. Lu, N.; Gao, A.; Dai, P.; Song, S.; Fan, C.; Wang, Y.; Li, T. Small 2014, 10, (10),
2022-2028.
38. Lu, N.; Gao, A.; Dai, P.; Li, T.; Wang, Y.; Gao, X.; Song, S.; Fan, C. Methods
2013, 63, (3), 212-218.
39. Gao, A.; Lu, N.; Wang, Y.; Dai, P.; Li, T.; Gao, X.; Fan, C. Nano Lett 2012, 12,
(10), 5262-5268.
40. Gao, A.; Lu, N.; Dai, P.; Li, T.; Pei, H.; Gao, X.; Gong, Y.; Wang, Y.; Fan, C.
Nano Lett 2011, 11, (9), 3974-3978.
41. Ryu, Y. K.; Chiesa, M.; Garcia, R. Nanotechnology 2013, 24, (31), 315205.
42. Chiesa, M.; Cardenas, P. P.; Oton, F.; Martinez, J.; Mas-Torrent, M.; Garcia, F.;
Alonso, J. C.; Rovira, C.; Garcia, R. Nano Lett 2012, 12, (3), 1275-1281.
43. Vacic, A.; Criscione, J. M.; Stern, E.; Rajan, N. K.; Fahmy, T.; Reed, M. A.
Biosensors & bioelectronics 2011, 28, (1), 239-242.
44. Tarasov, A.; Wipf, M.; Stoop, R. L.; Bedner, K.; Fu, W.; Guzenko, V. A.;
Knopfmacher, O.; Calame, M.; Schonenberger, C. ACS nano 2012, 6, (10), 9291-
9298.
45. Juhasz, R.; Elfström, N.; Linnros, J. Nano Lett 2004, 5, (2), 275-280.

117
46. Lin, C. H.; Hsiao, C. Y.; Hung, C. H.; Lo, Y. R.; Lee, C. C.; Su, C. J.; Lin, H. C.;
Ko, F. H.; Huang, T. Y.; Yang, Y. S. Chem Commun (Camb) 2008, (44), 5749-
5751.
47. Hahm, J.; Lieber, C. M. Nano letters 2004, 4, (1), 51-54.
48. Patolsky, F.; Zheng, G. F.; Hayden, O.; Lakadamyali, M.; Zhuang, X. W.; Lieber,
C. M. Proceedings of the National Academy of Sciences of the United States of
America 2004, 101, (39), 14017-14022.
49. Cui, Y.; Lauhon, L. J.; Gudiksen, M. S.; Wang, J. F.; Lieber, C. M. Applied
Physics Letters 2001, 78, (15), 2214-2216.
50. Hakim, M. M. A.; Lombardini, M.; Sun, K.; Giustiniano, F.; Roach, P. L.; Davies,
D. E.; Howarth, P. H.; de Planque, M. R. R.; Morgan, H.; Ashburn, P. Nano Lett
2012, 12, (4), 1868-1872.
51. Chang, H. K.; Wang, X.; Aroonyadet, N.; Zhang, R.; Song, Y.; Datar, R.; Cote,
R.; Thompson, M.; Zhou, C. MRS Online Proceedings Library 2013, 1569.
52. Park, I.; Li, Z.; Pisano, A. P.; Williams, R. S. Nanotechnology 2010, 21, (1),
015501.
53. Wang, D.; Sheriff, B. A.; Heath, J. R. Nano letters 2006, 6, (6), 1096-100.
54. Ju, S.; Facchetti, A.; Xuan, Y.; Liu, J.; Ishikawa, F.; Ye, P.; Zhou, C.; Marks, T.
J.; Janes, D. B. Nature nanotechnology 2007, 2, (6), 378-84.
55. Chen, P. C.; Shen, G. Z.; Chen, H. T.; Ha, Y. G.; Wu, C.; Sukcharoenchoke, S.;
Fu, Y.; Liu, J.; Facchetti, A.; Marks, T. J.; Thompson, M. E.; Zhou, C. W. ACS
nano 2009, 3, (11), 3383-3390.
56. Fiebig, E. W.; Wright, D. J.; Rawal, B. D.; Garrett, P. E.; Schumacher, R. T.;
Peddada, L.; Heldebrant, C.; Smith, R.; Conrad, A.; Kleinman, S. H.; Busch, M.
P. AIDS 2003, 17, (13), 1871-1879.
57. Sutthent, R.; Gaudart, N.; Chokpaibulkit, K.; Tanliang, N.; Kanoksinsombath, C.;
Chaisilwatana, P. Journal of clinical microbiology 2003, 41, (3), 1016-1022.
58. Tang, S.; Hewlett, I. The Journal of infectious diseases 2010, 201 Suppl 1, S59-
64.

118
59. Zheng, L.; Jia, L.; Li, B.; Situ, B.; Liu, Q.; Wang, Q.; Gan, N. Molecules 2012,
17, (5), 5988-6000.
60. Chang, L.; Song, L.; Fournier, D. R.; Kan, C. W.; Patel, P. P.; Ferrell, E. P.; Pink,
B. A.; Minnehan, K. A.; Hanlon, D. W.; Duffy, D. C.; Wilson, D. H. Journal of
virological methods 2013, 188, (1-2), 153-160.
61. Lee, K. B.; Kim, E.-Y.; Mirkin, C. A.; Wolinsky, S. M. Nano Lett 2004, 4, (10),
1869-1872.
62. Rutstein, S. E.; Kamwendo, D.; Lugali, L.; Thengolose, I.; Tegha, G.; Fiscus, S.
A.; Nelson, J. A. E.; Hosseinipour, M. C.; Sarr, A.; Gupta, S.; Chimbwandira, F.;
Mwenda, R.; Mataya, R. J Clin Virol 2014, 60, (4), 392-398.
63. Schockmel, G. A.; Yerly, S.; Perrin, L. J Acq Immun Def Synd 1997, 14, (2), 179-
183.
64. Zhou, W.; Dai, X. C.; Fu, T. M.; Xie, C.; Liu, J.; Lieber, C. M. Nano Lett 2014,
14, (3), 1614-1619.
65. Bellingham, J. R.; Phillips, W. A.; Adkins, C. J. J Phys-Condens Mat 1990, 2,
(28), 6207-6221.
66. Yuan, Z. J.; Zhu, X. M.; Wang, X. O.; Cai, X. K.; Zhang, B. P.; Qiu, D. J.; Wu,
H. Z. Thin Solid Films 2011, 519, (10), 3254-3258.
67. Jo, K.; Yu, H. Z.; Yang, H. Electrochim Acta 2011, 56, (13), 4828-4833.
68. Butler, J. E. Journal of immunoassay 2000, 21, (2-3), 165-209.
69. Stern, E.; Vacic, A.; Li, C.; Ishikawa, F. N.; Zhou, C.; Reed, M. A.; Fahmy, T. M.
Small 2010, 6, (2), 232-238.
70. Krebs, H. A. Annual Review of Biochemistry 1950, 19, (1), 409-430.
71. Lewandrowski, K.; Chen, A.; Januzzi, J. American journal of clinical pathology
2002, 118 Suppl, S93-99.
72. Qureshi, A.; Gurbuz, Y.; Niazi, J. H. Sensors and Actuators B: Chemical 2012,
171–172, (0), 62-76.

119
73. Puleo, P. R.; Meyer, D.; Wathen, C.; Tawa, C. B.; Wheeler, S.; Hamburg, R. J.;
Ali, N.; Obermueller, S. D.; Triana, J. F.; Zimmerman, J. L.; et al. The New
England journal of medicine 1994, 331, (9), 561-566.
74. McLean, A. S.; Huang, S. J. Annals of intensive care 2012, 2, 8.
75. Durak-Nalbantic, A.; Dzubur, A.; Dilic, M.; Pozderac, Z.; Mujanovic-Narancic,
A.; Kulic, M.; Hodzic, E.; Resic, N.; Brdjanovic, S.; Zvizdic, F. Bosnian journal
of basic medical sciences / Udruzenje basicnih mediciniskih znanosti =
Association of Basic Medical Sciences 2012, 12, (3), 164-168.
76. Arakawa, N.; Nakamura, M.; Aoki, H.; Hiramori, K. Cardiology 1994, 85, (5),
334-340.
77. Morita, E.; Yasue, H.; Yoshimura, M.; Ogawa, H.; Jougasaki, M.; Matsumura, T.;
Mukoyama, M.; Nakao, K. Circulation 1993, 88, (1), 82-91.
78. Richards, A. M.; Nicholls, M. G.; Espiner, E. A.; Lainchbury, J. G.; Troughton,
R. W.; Elliott, J.; Frampton, C.; Turner, J.; Crozier, I. G.; Yandle, T. G.
Circulation 2003, 107, (22), 2786-2792.
79. Kong, T.; Su, R.; Zhang, B.; Zhang, Q.; Cheng, G. Biosensors & bioelectronics
2012, 34, (1), 267-272.
80. Lee, I.; Luo, X.; Huang, J.; Cui, X. T.; Yun, M. Biosensors 2012, 2, (2), 205-220.
81. Zhang, G. J.; Luo, Z. H. H.; Huang, M. J.; Ang, J. A. J.; Kang, T. G.; Ji, H.
Biosensors and Bioelectronics 2011, 28, (1), 459-463.
82. Chua, J. H.; Chee, R. E.; Agarwal, A.; Wong, S. M.; Zhang, G. J. Analytical
chemistry 2009, 81, (15), 6266-6271.
83. Zhang, G. J.; Chai, K. T. C.; Luo, H. Z. H.; Huang, J. M.; Tay, I. G. K.; Lim, A.
E.-J.; Je, M. Biosensors and Bioelectronics 2012, 35, (1), 218-223.
84. Huang, I. Y.; Huang, R.-S.; Lo, L.-H. Sensors and Actuators B: Chemical 2003,
94, (1), 53-64.
85. Matsumoto, T.; Ohashi, A.; Ito, N. Analytica Chimica Acta 2002, 462, (2), 253-
259.

120
86. Keller, T.; Zeller, T.; Ojeda, F.; Tzikas, S.; Lillpopp, L.; Sinning, C.; Wild, P.;
Genth-Zotz, S.; Warnholtz, A.; Giannitsis, E.; Mockel, M.; Bickel, C.; Peetz, D.;
Lackner, K.; Baldus, S.; Munzel, T.; Blankenberg, S. JAMA : the journal of the
American Medical Association 2011, 306, (24), 2684-2693.
87. McCullough, P. A.; Nowak, R. M.; McCord, J.; Hollander, J. E.; Herrmann, H.
C.; Steg, P. G.; Duc, P.; Westheim, A.; Omland, T.; Knudsen, C. W.; Storrow, A.
B.; Abraham, W. T.; Lamba, S.; Wu, A. H.; Perez, A.; Clopton, P.;
Krishnaswamy, P.; Kazanegra, R.; Maisel, A. S. Circulation 2002, 106, (4), 416-
422.
88. Maisel, A. S.; Krishnaswamy, P.; Nowak, R. M.; McCord, J.; Hollander, J. E.;
Duc, P.; Omland, T.; Storrow, A. B.; Abraham, W. T.; Wu, A. H.; Clopton, P.;
Steg, P. G.; Westheim, A.; Knudsen, C. W.; Perez, A.; Kazanegra, R.; Herrmann,
H. C.; McCullough, P. A. The New England journal of medicine 2002, 347, (3),
161-167.
89. Jin, S.; Whang, D.; McAlpine, M. C.; Friedman, R. S.; Wu, Y.; Lieber, C. M.
Nano Lett 2004, 4, (5), 915-919.
90. Whang, D.; Jin, S.; Wu, Y.; Lieber, C. M. Nano Lett 2003, 3, (9), 1255-1259.
91. Kim, F.; Kwan, S.; Akana, J.; Yang, P. Journal of the American Chemical Society
2001, 123, (18), 4360-4361.
92. Melosh, N. A.; Boukai, A.; Diana, F.; Gerardot, B.; Badolato, A.; Petroff, P. M.;
Heath, J. R. Science 2003, 300, (5616), 112-115.
93. Yu, G.; Cao, A.; Lieber, C. M. Nature nanotechnology 2007, 2, (6), 372-377.
94. Huang, Y.; Duan, X.; Wei, Q.; Lieber, C. M. Science 2001, 291, (5504), 630-633.
95. Smith, P. A.; Nordquist, C. D.; Jackson, T. N.; Mayer, T. S.; Martin, B. R.;
Mbindyo, J.; Mallouk, T. E. Applied Physics Letters 2000, 77, (9), 1399-1401.
96. Tsivion, D.; Schvartzman, M.; Popovitz-Biro, R.; von Huth, P.; Joselevich, E.
Science 2011, 333, (6045), 1003-1007.
97. Nikoobakht, B.; Michaels, C. A.; Stranick, S. J.; Vaudin, M. D. Applied Physics
Letters 2004, 85, (15), 3244-3246.

121
98. Tsivion, D.; Schvartzman, M.; Popovitz-Biro, R.; Joselevich, E. ACS nano 2012,
6, (7), 6433-6445.
99. Zi, Y.; Jung, K.; Zakharov, D.; Yang, C. Nano letters 2013.
100. Mattila, M.; Hakkarainen, T.; Jiang, H.; Kauppinen, E. I.; Lipsanen, H.
Nanotechnology 2007, 18, (15).
101. Camagni, P.; Faglia, G.; Galinetto, P.; Perego, C.; Samoggia, G.; Sberveglieri, G.
Sensor Actuat B-Chem 1996, 31, (1-2), 99-103.
102. Dattoli, E. N.; Wan, Q.; Guo, W.; Chen, Y.; Pan, X.; Lu, W. Nano letters 2007, 7,
(8), 2463-2469.
103. Liu, Z. Q.; Zhang, D. H.; Han, S.; Li, C.; Tang, T.; Jin, W.; Liu, X. L.; Lei, B.;
Zhou, C. W. Advanced Materials 2003, 15, (20), 1754-1757.
104. Amin, N.; Isaka, T.; Yamada, A.; Konagai, M. Solar Energy Materials and Solar
Cells 2001, 67, (1–4), 195-201.
105. Comini, E.; Faglia, G.; Sberveglieri, G. Sensor Actuat B-Chem 2001, 78, (1-3),
73-77.
106. Maiti, A.; Rodriguez, J. A.; Law, M.; Kung, P.; McKinney, J. R.; Yang, P. D.
Nano Lett 2003, 3, (8), 1025-1028.
107. Law, M.; Kind, H.; Messer, B.; Kim, F.; Yang, P. D. Angew Chem Int Edit 2002,
41, (13), 2405-2408.
108. Choi, Y. J.; Hwang, I. S.; Park, J. G.; Choi, K. J.; Park, J. H.; Lee, J. H.
Nanotechnology 2008, 19, (9).
109. Ago, H.; Nakamura, K.; Ikeda, K.; Uehara, N.; Ishigami, N.; Tsuji, M. Chem Phys
Lett 2005, 408, (4-6), 433-438.
110. Ryu, K.; Badmaev, A.; Wang, C.; Lin, A.; Patil, N.; Gomez, L.; Kumar, A.;
Mitra, S.; Wong, H. S. P.; Zhou, C. W. Nano Lett 2009, 9, (1), 189-197.
111. Leonardy, A.; Hung, W.-Z.; Tsai, D.-S.; Chou, C.-C.; Huang, Y.-S. Crystal
Growth & Design 2009, 9, (9), 3958-3963.
112. Semancik, S.; Cavicchi, R. E. Thin Solid Films 1991, 206, (1–2), 81-87.

122
113. Dominguez, J. E.; Pan, X. Q.; Fu, L.; Van Rompay, P. A.; Zhang, Z.; Nees, J. A.;
Pronko, P. P. Journal of Applied Physics 2002, 91, (3), 1060-1065.
114. Lei, B.; Li, C.; Zhang, D.; Zhou, Q. F.; Shung, K. K.; Zhou, C. Applied Physics
Letters 2004, 84, (22), 4553-4555.
115. Kolmakov, A.; Zhang, Y.; Cheng, G.; Moskovits, M. Advanced Materials 2003,
15, (12), 997-1000.
116. Kordos, P.; Gregusova, D.; Stoklas, R.; Cico, K.; Novak, J. Applied Physics
Letters 2007, 90, (12).
117. Martel, R.; Schmidt, T.; Shea, H. R.; Hertel, T.; Avouris, P. Applied Physics
Letters 1998, 73, (17), 2447-2449.
118. Zhou, C.; Kong, J.; Yenilmez, E.; Dai, H. J. Science 2000, 290, (5496), 1552-
1555.
119. Forrest, S. R.; Burrows, P. E.; Thompson, M. E. Laser Focus World 1995, 31, (2),
99-107.
120. Sheats, J. R.; Antoniadis, H.; Hueschen, M.; Leonard, W.; Miller, J.; Moon, R.;
Roitman, D.; Stocking, A. Science 1996, 273, (5277), 884-888.
121. Han, S.; Jin, W.; Zhang, D. H.; Tang, T.; Li, C.; Liu, X. L.; Liu, Z. Q.; Lei, B.;
Zhou, C. W. Chem Phys Lett 2004, 389, (1-3), 176-180.
122. Freitag, M.; Martin, Y.; Misewich, J. A.; Martel, R.; Avouris, P. H. Nano Lett
2003, 3, (8), 1067-1071.
123. Chen, P. C.; Shen, G.; Zhou, C. Nanotechnology, IEEE Transactions on 2008, 7,
(6), 668-682.
124. Kolmakov, A.; Klenov, D. O.; Lilach, Y.; Stemmer, S.; Moskovits, M. Nano Lett
2005, 5, (4), 667-673.
125. Fryberger, T. B.; Semancik, S. Sensor Actuat B-Chem 1990, 2, (4), 305-309.
126. Li, C.; Lei, B.; Zhang, D.; Liu, X.; Han, S.; Tang, T.; Rouhanizadeh, M.; Hsiai,
T.; Zhou, C. Applied Physics Letters 2003, 83, (19), 4014-4016.
127. Pengfei, Q. F.; Vermesh, O.; Grecu, M.; Javey, A.; Wang, O.; Dai, H. J.; Peng, S.;
Cho, K. J. Nano Lett 2003, 3, (3), 347-351.

123
128. Cao, B. B.; Chen, J. J.; Tang, X. J.; Zhou, W. L. J Mater Chem 2009, 19, (16),
2323-2327.
129. Rout, C. S.; Ganesh, K.; Govindaraj, A.; Rao, C. N. R. Appl Phys a-Mater 2006,
85, (3), 241-246.
130. Ahn, M. W.; Park, K. S.; Heo, J. H.; Kim, D. W.; Choi, K. J.; Park, J. G. Sensor
Actuat B-Chem 2009, 138, (1), 168-173.
131. Kim, Y. S.; Hwang, I. S.; Kim, S. J.; Lee, C. Y.; Lee, J. H. Sensor Actuat B-Chem
2008, 135, (1), 298-303.
132. Kim, I. D.; Rothschild, A.; Lee, B. H.; Kim, D. Y.; Jo, S. M.; Tuller, H. L. Nano
Lett 2006, 6, (9), 2009-2013.
133. Kong, J.; Franklin, N. R.; Zhou, C. W.; Chapline, M. G.; Peng, S.; Cho, K. J.;
Dai, H. J. Science 2000, 287, (5453), 622-625.
134. Zhang, D.; Liu, Z. Q.; Li, C.; Tang, T.; Liu, X. L.; Han, S.; Lei, B.; Zhou, C. W.
Nano Lett 2004, 4, (10), 1919-1924.

Abstract (if available)

Abstract Nanostructure based field effect transistors (FET) have drawn attention from researchers around the globe due to their great electrical performance suitable for many electronic applications. In addition, one-dimensional nanomaterials have large surface to volume ratio which is a desirous characteristic for sensing applications. One-dimensional nanomaterial based FET sensors have been demonstrated their ultrahigh sensitivity to detect biomolecules for medical application and toxic gases for industrial applications. However, control assembly of one-dimensional nanomaterials is still one of challenges to hinder them from practical uses. In the first chapter of this dissertation, basic and theories about nanowire synthesis and key components for the sensing application have been addressed. In the chapter 2 to 4, we demonstrate scalable top-down approaches to fabricate nanoribbon FETs using 2 photolithographic masks to define precisely dimension and position of metal electrodes and nanoribbons. The relaxation in the lateral dimension facilitates the simple and scalable fabrication process compatible with the conventional microelectronic facilities yielding 100% functional devices with uniform electrical performance. Thickness of nanoribbons can be precisely controlled in the deposition process. ❧ In chapter 2, we select low pressure chemical vapor deposition poly-silicon (LPCVD poly-Si) as the nanoribbon material. Spin-on dopant solution with thermal annealing are chosen to tune doping concentration in the material without need of any toxic gas or expensive ion implantation. Poly-Si nanoribbon devices exhibit good ionic sensitivity both in wide pH range from pH 4 to 9 and in physiological solution range from pH 7.2 to 8. To optimize number of capture probes on nanoribbons, thin layer of silicon dioxide needs to be deposited on nanoribbon to improve surface chemistry. We demonstrate detection of cancer antigen-125 (CA-125), a biomarker for ovarian cancer, using our poly-Si nanoribbon devices with limit of detection 10 U/ml which is an order of magnitude lower than the clinically relevant level. ❧ In chapter 3, we change the nanoribbon material from poly-Si to indium oxide (In₂O₃). In₂O₃ is an inherently semiconducting material without requirement of doping concentration and its deposition process is radio frequency sputtering at room temperature enabling the use of low cost substrates instead of silicon. In₂O₃ nanoribbon devices exhibit better electrical performance and long-term stability in the aqueous condition over other metal oxides. In addition, In₂O₃ nanoribbon devices show excellent ionic sensitivity in both wide pH range (pH 4 to 9) and physiological solution pH range (pH 6.8 to 8.2). Combination between the In₂O₃ nanoribbon platform and a signal amplification technique, electronic enzyme-linked immunosorbent assay (ELISA), we achieve high sensitivity to target analytes such as streptavidin and human immunodeficiency virus type 1 (HIV-1) p24 proteins. This approach circumvents Debye screening effect in ionic solution to bypass complicated sample preparation and demonstrates detection of p24 protein at 20 fg/ml (about 250 viruses/ml) or 3 orders of magnitude lower than commercial ELISA kits with 35% conduction change in human blood serum. With the demonstrated sensitivity, scalability and uniformity, the In₂O₃ nanoribbon sensor platform makes a great progress toward clinical testing such as early diagnosis of acquired immunodeficiency syndrome (AIDS). ❧ In chapter 4, we have demonstrated an integration of the on-chip gate electrode to the In₂O₃ nanoribbon sensor chip to move our platform closer to the practical setting. The on-chip gate exhibits similar gating performance to the traditional Ag/AgCl electrode at the same liquid potential. The In₂O₃ nanoribbon biosensor chip with the on-chip gate shows excellent ionic sensitivity in the wide range of pH from 4 to 9. A biomarker panel for diagnosis of acute myocardial infarction, namely troponin I, creatine kinase MB (CK-MB), and b-type natriuretic peptide (BNP), is selected to demonstrate the point of care application for the emergency setting. To fit with requirement of the medical emergency, the assay time is reduced from 10 hours to be an hour without loss of significant sensitivity. We demonstrate detection of troponin I at 100 fg/ml with 30% change in conduction with the projected limit of detection about 15 ag/ml or about 5 orders of magnitude lower than commercial troponin I ELISA kits. We have completed calibration curves for all 3 biomarkers and testified our calibration curves with spiked 500 pg/ml BNP in human whole blood to simulate a sample from a patient. Average sensing response from this blind test shows acceptable 3% lower than the calculated value and still differentiate the patient's condition correctly. Further optimization and more statistical study are required for the clinical use. ❧ In chapter 5, we demonstrate control assembly synthesis of rutile semiconducting aligned SnO₂ nanowires on A-plane, M-plane and R-plane sapphire substrates for scalable and practical device applications. X-ray diffraction and transmission electron microscopy were used to characterize aligned nanowires for their growth mechanism and direction. Simple photolithography for patterning metal electrodes is required for fabrication of nanowire FETs. Transistors exhibit excellent electrical performance with on/off current ratio on the order of 10⁶, electron mobilities about 71.68 cm²/V.s and high current to the external organic light emitting diode display. In addition to electronic application, aligned nanowire devices can be utilized as photodetectors to differentiate between 254 and 365 nm ultra violet wavelengths. Their alignment also helps to segregate among different polarization angles with polarization ratio of photoconductance (σ) of 0.3. Lastly, we demonstrate align SnO₂ nanowire devices as scalable and ultrasensitive nitrogen dioxide (NO₂) chemical sensors at concentration of 0.2 ppb. ❧ In the last chapter, all topics mentioned in this dissertation are summarized and the future work for the In₂O₃ nanoribbon sensor is discussed.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Semiconducting metal oxide nanostructures for scalable sensing applications

PDF

Biosensing and biomimetic electronics

PDF

Surface functionalization of nanomaterials and development of field effect transistor nanobiosensors

PDF

In2O3 COVID-19 biosensors and two-dimensional materials: synthesis and applications

PDF

GaAs nanowire optoelectronic and carbon nanotube electronic device applications

PDF

Surface functionalization of nanomaterials and the development of nanobiosensors

PDF

Zero-dimensional and one-dimensional nanostructured materials for application in photovoltaic cells

PDF

Nanomaterials for macroelectronics and energy storage device

PDF

One-dimensional nanostructures for chemical sensing, transparent electronics, and energy conversion and storage devices

PDF

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

PDF

Nanostructure electronic band modeling for solar cell and lighting applications

PDF

Low-dimensional asymmetric crystals: fundamental properties and applications

PDF

Studies of nanosecond pulsed power for modifications of biomaterials and nanomaterials (SWCNT)

PDF

Integrated photonics assisted electron emission devices

PDF

Applications enabled by plasmonic nano-finger and analog computing based on memristive devices

PDF

Chemical vapor deposition of graphene: synthesis, characterization, and applications

Asset Metadata

Creator Aroonyadet, Noppadol (author)

Core Title One-dimensional nanomaterials for electronic and sensing applications

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Electrical Engineering

Publication Date 01/26/2015

Defense Date 01/07/2015

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

Tag aligned nanowire synthesis,CA-125 detection,chemical sensing,diagnosis of myocardial infarction,electronic ELISA,epitaxial growth on sapphire,HIV p24 detection,indium oxide nanoribbon,multiplex detection,nanowire electronics,OAI-PMH Harvest,OLED control circuit,photodetector,polarization detector,polysilicon nanoribbon,tin oxide nanowire,top-down nanoribbon biosensor

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Zhou, Chongwu (committee chair), Thompson, Mark E. (committee member), Wu, Wei (committee member)

Creator Email aroonyad@usc.edu,nop_ar@hotmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-525518

Unique identifier UC11298559

Identifier etd-Aroonyadet-3133.pdf (filename),usctheses-c3-525518 (legacy record id)

Legacy Identifier etd-Aroonyadet-3133.pdf

Dmrecord 525518

Document Type Dissertation

Format application/pdf (imt)

Rights Aroonyadet, Noppadol

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 a...

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