Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Semiconducting metal oxide nanostructures for scalable sensing applications
(USC Thesis Other)
Semiconducting metal oxide nanostructures for scalable sensing applications
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
SEMICONDUCTING METAL OXIDE NANOSTRUCTURES
FOR SCALABLE SENSING APPLICATIONS
by
Xiaoli Wang
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 Xiaoli Wang
ii
Dedication
To my parents, Don, and Josh
iii
Acknowledgements
I would first like to thank my advisor, Professor Chongwu Zhou, for giving me the
opportunity to pursue my interest in joining nanoelectronics with biotechnology. The
knowledge that I’ve gained in the field and the skills I’ve learned in problem solving
under his guidance will always be invaluable to me. Professor Zhou’s generosity with his
time, research advice, and resources has made many difficult goals achievable. Without
his academic and financial support, this dissertation would not be possible.
I would like to thank my dissertation committee: Professor Martin Gundersen and
Professor Edward Goo, as well as the rest of my qualification committee: Professor Alice
Parker and Professor Andrea Armani, for taking the time to help me through this process.
Professor Gundersen has been very encouraging to me the several years that I’ve be en his
TA. I appreciate his vote of confidence.
I would also like to thank Professor Mark Thompson, who has been co-advising the
biosensing project as long as I have been part of the team. His insights and advice had
given us new energy many times when things seemed impossible. Thanks also to Drs.
Ram Datar and Richard Cote for their collaboration.
I’m thankful to my biosensing teammates. Noppadol Arooynadet has always been there
to lend a helping hand since day one. Thank you for making the most frustrating
experiments a little more enjoyable. I would like to thank Yan Song for all the helpful
iv
discussions and my former teammates Rui Zhang, Hsiaokang Chang, Fumi Ishikawa, and
Marco Curreli for passing down their experience and making our lives easier.
Our experimental research couldn’t be completed without the use of many types of
equipments and facilities. I would like to express my gratitude to the staff at the
following facilities: USC cleanroom, USC CEMMA, UCLA Nanolab, and Hatachi SEM.
I’m also gratefu l for all the administrative support from Jenny Lin and Kim Reid.
Finally, I would like to thank all of my former and current Nanolab labmates. Not only
have you given me countless help in research throughout the years, but you’ve also made
my PhD study enjoyable and memorable.
v
Table of Contents
Dedication..... ...................................................................................................................... ii
Acknowledgements ............................................................................................................ iii
List of Figures ................................................................................................................... vii
Abstract......... .....................................................................................................................xv
Chapter 1 Introduction to Nanostructure Semiconducting Metal Oxide Sensors ..............1
1.1 Overview .............................................................................................................. 1
1.2 Background on semiconducting metal oxide sensors .......................................... 2
1.3 Mechanism of chemical sensing .......................................................................... 3
1.4 Mechanism of biosensing .................................................................................... 6
1.5 Effect of nano-dimensions on sensitivity ............................................................. 8
1.6 Chapter References ............................................................................................ 10
Chapter 2 Aligned Epitaxial SnO
2
Nanowires on Sapphire: Growth and Device
Applications ........................................................................................................12
2.1 Introduction ........................................................................................................ 12
2.2 Aligned nanowire growth and characterization ................................................. 13
2.2.1 Saphire substrate surface study ................................................................... 13
2.2.2 Nanowire synthesis ..................................................................................... 15
2.2.3 X-ray diffraction study ................................................................................ 19
2.2.4 Nanowire growth orientation ...................................................................... 19
2.2.5 TEM confirmation of nanowire growth orientation .................................... 21
2.2.6 Effect of partial pressure on alignment ....................................................... 24
2.3 Field-effect transistor device fabrication and characterization .......................... 25
2.3.1 Device fabrication ....................................................................................... 25
2.3.2 Device characterization ............................................................................... 26
2.4 Organic light-emitting diode (OLED) control circuit ........................................ 29
2.5 Photodetector and polarizer ............................................................................... 31
2.6 Aligned SnO
2
nanowires for NO
2
sensing ......................................................... 36
2.7 Summary ............................................................................................................ 38
2.8 Chapter References ............................................................................................ 40
Chapter 3 Highly Scalable, Uniform, and Sensitive Biosensors Based on Top-down
Indium Oxide Nanoribbon Field-Effect Transistor.............................................43
3.1 Introduction ........................................................................................................ 43
3.2 Nanoribbon biosensor fabrication ...................................................................... 45
3.1 Transistor behavior of various metal oxide nanoribbons ................................... 48
3.2 Uniformity of In
2
O
3
nanoribbon FET ................................................................ 51
3.3 In
2
O
3
nanoribbon stability in ionic solution....................................................... 55
3.4 pH sensing .......................................................................................................... 57
3.5 Surface chemistry for molecular binding ........................................................... 62
3.6 Biomolecule detection in buffer ......................................................................... 66
3.7 Summary ............................................................................................................ 68
vi
3.8 Chapter References ............................................................................................ 69
Chapter 4 Top-down Polysilicon Nanoribbons: a Comparison to In
2
O
3
Nanoribbon
Biosensors ...........................................................................................................71
4.1 Introduction ........................................................................................................ 71
4.2 Device Fabrication ............................................................................................. 73
4.3 Electrical characteristics of polysilicon nanoribbon FETs ................................ 78
4.3.1 Dependence of performance on the dopant solution concentration ............ 78
4.3.2 Statistical data of electrical performance .................................................... 80
4.3.3 Comparison with In
2
O
3
nanoribbon FET sensor ....................................... 81
4.4 pH sensing .......................................................................................................... 82
4.5 Biomarker detection ........................................................................................... 85
4.6 Summary ............................................................................................................ 88
4.7 Chapter References ............................................................................................ 90
Chapter 5 Application of In
2
O
3
Nanoribbon Biosensors and Electronic Enzymen-
Linked Immunosorbent Assay for Chest Pain Diagnosis ...................................91
5.1 Introduction ........................................................................................................ 91
5.2 Electronic enzyme-linked immunosorbent assay (ELISA)................................ 94
5.3 Shortened incubation time of Troponin ............................................................. 96
5.4 Precision adjustment of kinase–MB (CK-MB) concentration ........................... 99
5.5 Detection of B-type natriuretic peptide (BNP) in whole blood ....................... 102
5.6 Summary .......................................................................................................... 106
5.7 Chapter References .......................................................................................... 107
Chapter 6 Conclusion and Future Directions .................................................................109
6.1 Conclusion ....................................................................................................... 109
6.2 Future directions for aligned epitaxial nanowires ............................................ 110
6.2.1 Transfer by laser lift-off ............................................................................ 111
6.2.2 Transfer from quartz substrate .................................................................. 112
6.3 Future directions for In
2
O
3
nanoribbon biosensor ........................................... 113
6.3.1 On-chip liquid gate ................................................................................... 114
6.3.2 Microfluidic integration ............................................................................ 115
6.4 Chapter References .......................................................................................... 117
Bibliography. ..................................................................................................................118
vii
List of Figures
Figure 1.1. (a) Energy band diagram of a metal oxide interface after chemical
reaction with oxygen ions.
15, 16
5
Figure 1.2. (a) Energy band diagram of a carbon nanotube field effect transistor
in the off state of the device. (b) Energy band diagram of the CNT
FET in on-state.
19
6
Figure 1.3. Simulation of drain current (I
ds
) against gate voltage (V
lg
) curves for
before (black) and after (red) protein attachment on an ambipolar
carbon nanotube FET sensor due to (a) mobility change, (b)
dielectric change, and (c) gate bias.
20
8
Figure 1.4. Conduction path and corresponding band diagram for polycrystalline
SMO when (a) grain size is larger than 2Λ and when (b) grain size is
smaller than 2Λ.
15, 24
9
Figure 2.1 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. 15
Figure 2.2 SEM images of aligned SnO
2
nanowires grown on (a) A-plane, (b)
M-plane, and (c) R-plane sapphires. (d) HRSEM of aligned SnO
2
nanowires on R-plane sapphire. 17
Figure 2.3 Histograms of aligned SnO
2
nanowire assembly parameters. (a - c)
Aligned SnO
2
nanowire density on annealed A- plane (a), annealed
M-plane (b), and R-plane (c) sapphire 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 (f) sapphire substrates
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 substrates (i) 18
Figure 2.4 XRD data for aligned SnO
2
nanowires grown on (a) A-plane, (b) M-
plane, and (c) R-plane sapphires show all three planes tend to
interface the SnO
2
(101) plane. 19
viii
Figure 2.5 Diagrams of atomic arrangement for (a) A-plane sapphire, (b) (101)
plane SnO
2
, and (c) R-plane sapphire. 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. 21
Figure 2.6 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) Electron diffraction pattern of SnO
2
nanowire
cross-section taken from the FIB cut sample. (c) Electron diffraction
pattern of cross-section of R-plane sapphire. (d) Enlarged image of
(b) and overlaid with black dots representing software simulated
diffraction pattern of SnO
2
looking into the [1 ¯ 01] lattice vector
direction. (e) Enlarged image of (c) and overlaid with back dots
representing software simulated diffraction pattern of sapphire
looking into the [1 ¯ 101] lattice vector direction. The faint rings are
diffraction from the protective Pt layer. 22
Figure 2.7 High resolution transmission electron microscopy (HRTEM) imaging
of aligned SnO
2
nanowire cross section on R-plane sapphire
substrate. Inset shows spacing between the SnO
2
(010) planes.
Normal direction and plane index for both the nanowire and the
sapphire substrate are labeled. Circled “b” and “c” respectively
correspond to spots where the diffraction pattern in Figure 2.5 for
SnO
2
and sapphire are taken. 23
Figure 2.8. 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. 25
Figure 2.9 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. 28
ix
Figure 2.10 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). 29
Figure 2.11 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. 31
Figure 2.12 Photoconduction (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 inset, 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. (c)
Real-time detection of 2 different wavelengths. (d) Real-time
response from 254 nm illumination shows photoconduction
decreased and reached steady state after 60 minutes. Left inset shows
2% variation in current from 50 and 100 minutes. Right inset shows
stability test of photoconduction: each data point is the average of
photoconduction response over 200 s when exposed to 254 nm UV
illumination. 34
Figure 2.13. 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. 35
Figure 2.14 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
. 38
Figure 3.1 Fabrication processes of In
2
O
3
nanoribbons biosensors. (a) Low
pressure chemical vapor deposition of 500nm Si
3
N
4
on Si wafer
followed by spin-coating of photoresist. (b) First photolithography
step defining source and drain metal electrodes. (c) Deposition of
5nm Ti and 50nm Au electrodes by electron beam evaporation. (d)
Photoresist lift-off to expose electrodes. (e) Spin-coating of
photoresist for transistor channel. (f) Second photolithography step
x
defines the nanoribbon active channel. (g) In
2
O
3
deposition by radio
frequency sputtering. (h) Photoresist lift-off to expose In
2
O
3
nanoribbon channel. (i) An optical image In
2
O
3
nanoribbon
biosensors on a 3 inch wafer. Inset shows a magnified nanoribbon
chip composed of 4 subgroups of 6 nanoribbon devices. (f) An SEM
micrograph of two identical nanoribbon devices. 47
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. 48
Figure 3.3 Transistor family I
DS
-V
DS
and I
DS
-V
GS
curves for various metal oxide
nanoribbon devices. (a) I
DS
-V
DS
and (b) I
DS
-V
GS
for an InGaZnO
nanoribbon device. (c) I
DS
-V
DS
and (d) I
DS
-V
GS
for a SnO
2
nanoribbon device. (e) I
DS
-V
DS
and (f) I
DS
-V
GS
for an ITO
nanoribbon device. (g-h) Optical images of 6 ZnO nanoribbon
devices before (g) and after (h) 14 hour incubation in PBS show that
ZnO nanoribbons dissolved completely in PBS after 14 hours. 50
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. (e) On-state current reproduced
from (a) and plotted in logarithmic scale. Inset SEM images of two
representative nanoribbon devices show identical channel features.
(f) On-state current measured from 50 devices In
2
O
3
nanowire FET
devices. Inset SEM images taken from two representative devices
show non-uniformity in the channel. 53
Figure 3.5 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. 55
Figure 3.6 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. 57
Figure 3.7 (a) Schematic diagram of set-up for pH sensing experiments using
In
2
O
3
nanoribbon devices. Commercial pH buffer solution is
xi
confined in a Teflon electrochemical chamber. Liquid gate voltage is
applied through a Ag/AgCl electrode. (b) Real-time, normalized
current obtained from a 20 nm In
2
O
3
nanoribbon device exposed to
commercial buffer solutions with pH 4 to 9. 58
Figure 3.8 pH sensing in physiological pH range (a) Real-time sensing
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. (b) Plot of pH versus response for one of the devices. 59
Figure 3.9 Real-time pH sensing responses from In
2
O
3
nanoribbon devices with
ribbon thickness ranging from 10 to 50 nm. The pH of the
commercial buffer solutions used for sensing varies from pH 4 to 9.
Error bar indicates lowest and highest response among sensors tested
for each data point. 61
Figure 3.10. Comparison of pH sensing response from annealed nanoribbon to
that of non-annealed nanoribbon sensors. (a) X-ray diffraction
spectroscopy on 20 nm In
2
O
3
film shows sputtered In
2
O
3
is
amorphous before annealing (bottom) and polycrystalline after
annealing in low vacuum at 300 ºC for 30 minutes (top). (b)
Comparison of pH sensing responses obtained from the average of
three as-sputtered and three annealed devices exposed to commercial
pH buffer solutions with pH in a range of 4 to 9. 62
Figure 3.11 (a) A schematic diagram of a nanoribbon functionalized with amine
biotin to immobilize streptavidin conjugated with fluorescent red
dyes. (b) A negative fluorescence control for (a): nanoribbon is
functionalized with amine PEG, which cannot bind with streptavidin.
(c) A schematic diagram of the sample for fluorescence experiment:
In
2
O
3
ribbons on Au metal electrodes, deposited on a dielectric
substrate. (d) An optical image of (c). (e) Fluorescent image of the
sample with biotin probe and 500 nm SiO
2
substrate. (f) Negative
control of (e): sample with PEG probe and 500 nm SiO
2
substrate.
(g) Fluorescent images of the sample with biotin probe and 500 nm
Si
3
N
4
substrate. (h) Negative control of (g): sample with PEG probe
and500 nm Si
3
N
4
substrate. 65
Figure 3.12 CA125 sensing (a) Real-time sensing of CA125 antigen with limit
of detection at 0.1U/ml. Plot shows response from three sensors
performed simultaneously (b) Normalized current response versus
CA125 concentration for the corresponding sensors. Dotted data are
fitted to simulated Langmuir isotherm model. 67
xii
Figure 4.1 Polysilicon nanoribbon FET sensor fabrication (a) Polysilicon is
deposited via LPCVD. Inset shows Si/SiO
2
substrate (b) Conduction
is aided by spin-on dopants that are thermally diffused into the
polysilicon. (c) Active mesas are patterns are transferred onto
photoresist by the first photolithographic step. (d) Dry etching
transfers the nanoribbon pattern onto the polysilicon. (e) The second
photolithography step transfers the metal electrode patterns onto the
photoresist. (f) Ti/Au is deposited into source and drain electrode
areas by metal evaporation. (g) Lift-off of photoresist exposes
electrodes in the final step of the FET sensor fabrication. (h)
Photographic image of nanosensor arrays on a 3” wafer, and inset
shows zoomed-in SEM image of two identical polysilicon
nanoribbons. 77
Figure 4.2 (a) I
ds
versus V
ds
under various V
g
for a device doped with spin-on
boron solution of C
0
= 1×10
17
(step: 10V). (b) I
ds
versus V
g
under
various V
ds
for a device with 1×10
17
doping (step: 1V). (c) I
ds
versus
V
ds
under various V
g
for a device with 5×10
17
doping (step: 10V).
(d) I
ds
versus V
g
under various V
ds
for a device with 5×10
17
doping
(step: 1V). (e) I
ds
versus V
ds
under various V
g
for a device with
1×10
18
doping (step: 10V). (f) I
ds
versus V
g
under various V
ds
for a
device with 1×10
18
doping (step: 1V). 79
Figure 4.3 Polysilicon FET statistical characteristics show good uniformity
among 28 devices. (a) Device on-current at V
gs
= -40V and V
ds
= -1
V. (b) Drain current on/ off ratio. (c) Device transconductance. 80
Figure 4.4 Comparison of Ids-Vgs between In
2
O
3
and polysilicon nanoribbon
FET sensors. 82
Figure 4.5 pH sensing using polysilicon nanoribbon FET sensors (a) Sensing
set-up using Ag/AgCl liquid gate and Teflon cell for enclosing fluids.
(b) Signal (normalized current) for pH range from pH 2 to pH 10 in
steps of 2. (b) Signal for detecting pH in the physiological range in
steps of 0.2. 83
Figure 4.6 pH sensing using polysilicon nanoribbon FET sensors (a) Sensing
set-up using Ag/AgCl liquid gate and Teflon cell for enclosing fluids.
(b) Signal (normalized current) for pH range from pH 2 to pH 10 in
steps of 2. (b) Signal for detecting pH in the physiological range in
steps of 0.2. 85
Figure 4.7 CA-125 sensing. (a) Real time sensing of CA-125 antigen using
polysilicon nanoribbon sensors. (b) Comparison of CA125
xiii
concentration versus response from polysilicon and In
2
O
3
nanoribbon sensors. (c) Surface chemistry for attaching capture
molecules to polysilicon nanoribbon. 88
Figure 5.1 Schematic diagram of experimental set up and electronic ELISA for
cardiac biomarker. 95
Figure 5.2 Troponin sensing (a) Schematic diagram of electronic ELISA for
tronponin sensing. (b) Real time eELISA signal for 3 In
2
O
3
nanoribbon sensors performing detection of 300 pg/ml troponin,
which was incubated with the nanoribbon for 1 hour. (c) same
experiment as (b) with troponin incubation time reduced to 15
minuts. (d) Troponin biomarker concentration versus signal for 4
concentrations of troponin incubated 2 hours with nanoribbon sensor
and 3 concentrations with the shorter assay. 99
Figure 5.3 CK-MB sensing (a) Biotinylation of secondary CKMB antibody
targeting full biotinylation. (b) Partial biotinylation of secondary
CKMB antibodies. (c) Real time electronic ELISA sensing of 1
ng/ml CK-MB biomarker in PBS buffer. (d) Average sensing
response of 3 nanoribbon devices at 1 ng/ml, 3 ng/ml, and 5 ng/ml of
CK-MB markers in buffer with standard deviation among the 3
sensors as error bar. 101
Figure 5.4 Real time electronic ELISA assay signal for detecting (a) 0.1 pg/ml,
(b) 1 pg/ml, and (c) 10 pg/ml of BNP in buffer using In
2
O
3
nanoribbon biosensor. (d) Average sensing response of (a) – (c) with
standard deviation among the sensors as error bar. Logarithmic
fitting curve with R-squared value of 0.9936 shows good fit for BNP
calibration curve. 104
Figure 5.5 (a) Real time electronic ELISA assay signal for detecting 5 pg/ml of
BNP in 100x diluted whole blood In
2
O
3
nanoribbon biosensor. (d)
Average sensing response of (a) placed on the concentration
calibration curve for BNP. 105
Figure 6.1 Transfer of GaN membrane. (a) Starting materials are GaN
membrane on sapphire and adhesive on Si. (b) GaN is attached to Si
by the adhesive. (c) Laser heats the interface of GaN and sapphire,
decomposing GaN into Ga and N2 gas. (d) Annealing melts the Ga
and removes the saspphire. (e) adhesive is etched away, leaving free
standing GaN film. 112
Figure 6.2 SnO
2
nanowires grown on ST cut quartz. 113
xiv
Figure 6.3 On-chip liquid gate. (a) Example design of In
2
O
3
nanoribbon chips
fabricated on 3 inch wafer, with electrodes that can accommodate 24
devices; (b) Expanded optical image of one nanoribbon chip design.
(c) Expanded view of the layout that contains an on-chip solution
gate before the sensor. 114
Figure 6.4 Example microfluidic design. (a) A simple microfluidic chip design
can carry samples to the nanoribbon sensors through the micro-
channels. (b)Integrated platform combines the microfluidic chip with
the nanosensor array, allowing (c) Streptavidin with different colored
dye to flow independently through each channel. 116
xv
Abstract
Semiconducting metal oxides (SMO) are widely known to respond strongly to changes in
their environment through their surface chemistry. Sensors built from field-effect
transistors (FET) that use nanostructured SMOs as the active channel have the combined
advantages of having an ultra sensitive surface with large surface-to-volume ratio and
being equipped with an electronic read-out that can be label-free, fast-responding,
portable, and accessible. However, the obstacle of applying metal oxide nanostructures to
FET sensing technology in a scalable and controllable process that’s necessary for
practical applications must first be tackled. In this thesis two approaches are
demonstrated to overcome this problem.
In chapter one, background information for semiconducting metal oxides, the effect of
nanoscale materials on sensitivity, and the mechanism of chemical and biological sensing
are presented.
In chapter two, the growth of epitaxially aligned SnO
2
nanowires are demonstrated as a
“bottom-up” approach that can be used to fabricate FETs, photo-detectors, polarizers, and
chemical sensors in a salable process. Nanowire growth study shows good alignment
guided by substrate lattice, and the electronic quality of the aligned SnO
2
nanowires is
demonstrated through its ultra sensitive 0.2ppb detection level of NO
2
.
xvi
In chapter three, a “top -down” process is developed to fabricate In
2
O
3
nanoribbon
biosensors that are highly uniform and sensitive. The fabrication process requires only 2
conventional photolithography steps that are scalable for different wafer sizes and can use
a versatile range of substrates. The sensors are demonstrated to have strong and fast
response to pH, and a stable surface chemistry is used to apply the nanoribbon sensors to
specific and selective biomarker detection.
In chapter four, polysilicon is investigated as the optimal silicon-based material for
developing “top -down” nanoribbon biosensors. By comparing its fabrication process, pH
and biomarker sensitivity with the In
2
O
3
nanoribbon, it is concluded that although poly-
silicon nanoribbon fabrication is a scalable way to achieve silicon-based “top -down”
nanobiosensors, In
2
O
3
nanoribbons are more advantageous in terms of sensitivity,
uniformity, and versatility.
In chapter five, In
2
O
3
nanoribbon biosensors are used to perform an electronic enzyme-
linked immuno assay (ELISA) for the detection of multiple chest pain biomarkers.
Quantitative detection of the cardiac biomarkers troponin, creatine kinase – MB, and B-
type natriuretic peptide (BNP) were achieved within clinically relevant concentrations.
Detection of BNP in whole blood was also achieved with concentration response within
3% of prediction.
Finally, chapter 6 presents a summary of the thesis and future directions.
1
Chapter 1 Introduction to Nanostructure Semiconducting
Metal Oxide Sensors
1.1 Overview
Sensor technology has been an important part of many sectors of society ranging from
agricultural and energy to transportation security and medicine. One example of a
nineteenth century chemical sensor technology is the “hand -held” canary that stopped
singing to warn of too much combustible gases in the mines.
1
The explosion of
nanotechnology within the last twenty years has pushed the boundary of response times,
detection limits, sensitivity, portability, etc… for sensor technology, particularly for
chemical and biological sensors. This is partly due to the fact that nanostructures that
have at least one dimension in the range of 1 to 100nm have comparable sizes as many of
the chemical and biological species of interest, and are thus better for probing the
molecules. Another important feature of the nanostructures is their large surface-to-
volume ratio that allows their material properties to be strongly affected by their
environment. One obstacle in producing nanosensors for practical, everyday use is the
difficulty in assembling nanostructures in a controllable, repeatable, and scalable fashion
to allow the development of a systematic technology. This thesis will explore scalable
2
metal oxide nanostructure sensor technology from both a “bottom -up” and a “top -down”
perspective and dive into the applications that can best benefit from each process.
1.2 Background on semiconducting metal oxide sensors
Semiconducting metal oxides (SMOs) are compounds of oxygen and certain transition
and post-transition metals. Examples include ZnO, SnO
2
, In
2
O
3
, Ga
2
O
3
, TiO
2
, and VO
2
.
Most SMOs are n-type semiconductors due to the low-energy oxygen 2p orbitals that
make up their valence bands.
2
Furthermore, impurities
3
and native defects such as metal
interstitials
4
and oxygen vacancies
4, 5
are also likely to contribute to electron conduction.
The surface of these oxides can readily form physical and chemical reactions with other
molecules in its surroundings, often resulting in exchange of free electrons between the
molecules and the metal oxide. Depending on the nature and the concentration of the
surrounding molecules, their injection or withdrawal of free electrons can change the
conduction of the metal oxide significantly, resulting in a highly sensitive response from
the metal oxide. The ability for chemical bonds also facilitates receptor molecule binding,
contributing to good specificity in SMO sensors. Furthermore, many SMOs have a large,
direct bandgap. This enables photo-detection as well as molecular detection. Besides
having versatile sensing properties, SMOs can be flexible, optically transparent, resistant
to high temperatures, and compatible with microfabrication. For these reasons, SMOs
have become a front runner for solid state sensor electronics.
3
One of the first demonstrations of controlling a metal oxide’s conductivity was the Nernst
lamp that was developed in 1897.
1, 6
It showed that yittria (Y
2
O
3
)-stabilized zirconia
(ZrO
2
), or YSZ, changed from an insulator to a conductor when heated over 700K. The
YSZ rod filament was stable enough in air that it did not need to be encapsulated in a
non-oxidizing environment. The first use of SMO for molecular detection occurred later
in 1956, when Heiland showed that the surface of a ZnO crystal changed conductivity
when exposed to oxygen.
7
Metal oxide thin film sensors quickly followed, and was used
by Seiyama to detect inflammable gasses in 1962.
8
Since then, academic and industrial
interest in SMO sensors intensified, and much effort have been directed at improving the
sensitivity and selectivity of SMO sensors. In 1991, Yamozoe showed that the sensitivity
of SMO sensors increased with decreasing grain size.
9
Soon after, a myriad of
publications began to introduce sensors composed of various 1D SMO structures such as
nanowires
10, 11
and nanobelts
12
. With the successful detection of environmental gasses at
the ppb level
10
, and bio-molecules at clinically relevant levels,
13
1D SMO detectors hold
great promise as industrial and medical sensing devices. However, due to the nature of
the growth of the 1D SMO structures, controlled, large-scale assemble of sensor devices
have become an obstacle for the oxide sensors to be practical.
1.3 Mechanism of chemical sensing
The detection mechanism of gasious molecules by SMO material is still a topic of
research. Two widely accepted models are the ionosorption model and oxygen-vacancy
4
models. In the ionosoption model, oxygen molecules from air are readily attached to the
metal oxide surface. The adsorbed oxygen molecule traps free electrons from the SMO’s
conduction band and forms the following possible oxygen species:
14
The oxygen ions negatively charge the oxide surface and induce a space charge layer in
metal oxide just beneath its surface. For an n-type SMO, the space charge layer is
positively charged and is partially or fully depleted of free electrons. The resultant
upward energy band bending with respect to the Fermi level is shown in Figure 1.1. The
reduced free electron concentration, as well as the barrier created by the band bending
lowers the metal oxide’ s conductivity. When target gas molecules adsorbs on the SMO
surface, they react with the oxygen ions and induces a signal by either increasing or
further decreasing the SMO conduction. Gas molecules such as H
2
S, CO, NH
3
, CH
4
and
SO
2
are reducing gases, and they increase SMO conduction by freeing electrons from the
oxygen ions. A characteristic reduction path is illustrated using CO:
14
Conversely, gases such as NO
2
, NO, N
2
O and CO
2
are oxidizing gasses that further
withdraw electrons, causing a decrease in the oxide conduction. Typical species involved
in the reaction are shown below using NO
2
as an example:
14
5
The change in surface charge caused by either reducing or oxidizing gases also helps to
lower or raise, respectively, the potential barrier at the surface. This affects the ease of
electron flow across grain barriers, contributing to change in conduction.
Figure 1.1. (a) Energy band diagram of a metal oxide interface after chemical reaction with oxygen ions.
15
,
16
In the oxygen vacancy model, the metal oxide may lose its lattice oxygen, resulting in an
oxygen vacancy,
, according to the equation
17
The vacancy defect is easily ionized to
by emmitting free electrons to the
conduction band. Adsorbed oxygen from the environment can re-fill these vacancies and
reduce free electron concentration. Reducing gases like CO
2
can increase SMO
conduction because they react to remove more surface oxygen species, which supports
6
oxygen vacancy formation. On the other hand, oxidizing gases like NO
2
free up more
oxygen species and decrease the number of oxygen vacancies.
1.4 Mechanism of biosensing
The operation principles of a field-effect transistor (FET) nanobiosensor are illustrated in
Figure 2 and explained in literature.
18
The metal oxide nanostructure is used as the
channel of the FET, bridging source-drain (S-D) electrodes. The surface of the channel is
chemically binded to receptor molecules (Figure 1.2(a)). These receptors have a high
binding affinity and selectivity toward a particular target analyte. When the target
molecule approaches the sensor surface, it is captured by the receptor molecule (Figure
1.2(b)). This binding interaction changes the environment at the channel surface—the
analyte is usually electrically charged—causing a change in the conductance of the
channel (1.2(b) inset) at binding. This electrical signal read out enables for label-free
detection of analytes, without complex optical equipments.
Figure 1.2. (a) Energy band diagram of a carbon nanotube field effect transistor in the off state of the
device. (b) Energy band diagram of the CNT FET in on-state.
19
S
D
Receptors
Nanostructure
Channel
(a)
Current
Time
(b)
Current
Time
Captured
Target
7
The mechanism of the conduction changing during molecular binding is also a debated
topic.
20-23
According to the ideal transistor linear (region often used for biosensing)
current equation
(10)
While the transistor dimenions (A, d, and L) and the drain voltage (V
ds
) are constant, a
change in conduction current (I
ds
) can be caused by either a change in mobility (μ), a
change in capacitance due to the difference in the dielectric constant (ε
r
) of the sensing
environment versus the binding molecule, or a gating effect (V
g
) caused by charges from
the binding molecule. These three situations are illustrated in Figure 1.3 by comparing
the I
ds
–V
gs
curves of an ambipolar FET device before and after protein binding. Figure
1.3(a) shows that a decrease in the slope of the I
ds
–V
gs
curve after protein binding also
decrease the I
ds
at fixed V
gs
. A change in I
ds
due to the slope indicates a reduction in
mobility and transconductance inside the channel, possibly due to an uneven electrostatic
field distribution caused by random binding with charged biomolecules. In Figure 1.3 (b)
the gate bias is shown to be less effective at inducing I
ds
. The current reduction in this
case can be attributed to a reduced gate capacitance caused by the low permittivity of the
bound biomolecule. Finally, Figure 1.3(c) shows a I
ds
change due to electrostatic gating
of the FET channel by charged target biomolecules. This type of change causes a
threshold voltage (V
T
) shift seeing in the figure.
8
Figure 1.3. Simulation of drain current (I
ds
) against gate voltage (V
lg
) curves for before (black) and after
(red) protein attachment on an ambipolar carbon nanotube FET sensor due to (a) mobility change, (b)
dielectric change, and (c) gate bias.
20
1.5 Effect of nano-dimensions on sensitivity
Biomolecule detection using nanostructures have two distinct advantageous. First,
nanostructures are the best probes for biomolecules due to the similarty in their sizes. For
example, nanowires and proteins are both in the range of tens of nanometers, making
nanowires a good transducer for protein activities. Secondly, nanomaterials and
nanostructures have high surface-to-volume ratio. This exposes a large proportion of the
material to its surroundings, enabling its electronic and chemical proterties to be easily
affect by any disturbance in the vicinity. These properties create great potentials for
highly sensitive biosensors.
Furthermore, nanowires are often single crystalline, which was shown to significantly
increase gas sensing performance by Yamazoe in 1991.
9
Adsorbed oxygen ions create a
charge depletion layer with a thickness of Λ at each grain surface according to reactions
outlined in Equations 1-4. In Figure 1.4 CO gas molecules that bind to the surface of the
(a) Mobility
(b) Capacitance
(c) Gating
9
SMO frees up surface oxygen ions and decreases Λ by releasing electrons back into the
metal oxide. The depleted areas of each grain are shown in white while the non-depleted
areas are shown in grey. Figure 1.4(a-b) illustrates the case where the grain size is large
compared to 2Λ. The increase in the conduction from (a) before to (b) after the CO
binding is mainly controlled by the energy barrier reduction, as seeing in the
corresponding band diagram. This change in Λ only affects the surface of the grains,
which is a very small part in comparision to the rest of the grain, and therefore, the
conduction change is small. In Figure 1.4(c-d), the grain size is smaller than 2Λ , thus
adsorbed oxygen ions can deplete the entire grain of carriers (1.4(c)). When CO binding
occurs (1.4(d)), the change in oxygen ion concentration lowers the energy for free
electron everywhere in the material, and the conduction signal is much more pronounced.
This explains why a single crystal nanowire is a much effctive chemical sensor then thin
film sensors.
Figure 1.4. Conduction path and corresponding band diagram for polycrystalline SMO when (a) grain size
is larger than 2Λ and when (b) grain size is smaller than 2Λ .
15, 24
(c)
(b)
(a)
(d)
10
1.6 Chapter References
1. Carpenter, M.A.; Mathur, S.; Kolmakov, A., Metal oxide nanomaterials for
chemical sensors. Springer: New York, 2013; p xiii-xx.
2. Svensson, B.G.; Pearton, S.J.; Jagadish, C., et al., Preface. In Oxide
Semiconductors, Elsevier Academic Press Inc: San Diego, CA, 2013; Vol. 88, pp
Xi-Xiii.
3. Lyons, J.L.; Janotti, A.; Van de Walle, C.G., Theory and Modeling of Oxide
Semiconductors. In Oxide Semiconductors, Svensson, B. G.; Pearton, S. J.;
Jagadish, C., Eds. Elsevier Academic Press Inc: San Diego, CA, 2013; Vol. 88, pp
1-38.
4. Kilic, C.; Zunger, A. Origins of coexistence of conductivity and transparency in
SnO(2). Physical review letters 2002, 88 (9), 095501.
5. Pati, S.; Majumder, S.B.; Banerji, P. Role of oxygen vacancy in optical and gas
sensing characteristics of ZnO thin films. J Alloy Compd 2012, 541, 376-379.
6. SMITH, H.M. THE NERNST LAMP. Science 1898, 8 (203), 689-690.
7. Heiland, G.; Mollwo, E.; Stockmann, F. Electronic Processes in Zinc Oxide. Solid
State Phys 1959, 8, 191-323.
8. Seiyama, T.; Kato, A.; Fujiishi, K., et al. A New Detector for Gaseous
Components Using Semiconductive Thin Films. Anal Chem 1962, 34 (11), 1502-
1503.
9. Yamazoe, N. New Approaches for Improving Semiconductor Gas Sensors. Sensor
Actuat B-Chem 1991, 5 (1-4), 7-19.
10. Zhang, D.H.; Liu, Z.Q.; Li, C., et al. Detection of NO2 down to ppb levels using
individual and multiple In2O3 nanowire devices. Nano Lett 2004, 4 (10), 1919-
1924.
11. Kolmakov, A.; Zhang, Y.X.; Cheng, G.S., et al. Detection of CO and O-2 using
tin oxide nanowire sensors. Adv Mater 2003, 15 (12), 997-+.
12. Comini, E.; Faglia, G.; Sberveglieri, G., et al. Stable and highly sensitive gas
sensors based on semiconducting oxide nanobelts. Appl Phys Lett 2002, 81 (10),
1869-1871.
13. Chang, H.K.; Ishikawa, F.N.; Zhang, R., et al. Rapid, label-free, electrical whole
blood bioassay based on nanobiosensor systems. ACS nano 2011, 5 (12), 9883-91.
14. Wetchakun, K.; Samerjai, T.; Tamaekong, N., et al. Semiconducting metal oxides
as sensors for environmentally hazardous gases. Sensor Actuat B-Chem 2011, 160
(1), 580-591.
15. Franke, M.E.; Koplin, T.J.; Simon, U. Metal and metal oxide nanoparticles in
chemiresistors: does the nanoscale matter? Small 2006, 2 (1), 36-50.
16. Barsan, N.; Weimar, U. Conduction model of metal oxide gas sensors. J
Electroceram 2001, 7 (3), 143-167.
11
17. Gurlo, A., Insights into the Mechanism of Gas Sensor Operation. In Metal oxide
nanomaterials for chemical sensors, Carpenter, M. A.; Mathur, S.; Kolmakov, A.,
Eds. Springer: New York, 2013; pp 3-34.
18. Curreli, M.; Zhang, R.; Ishikawa, F.N., et al. Real-Time, Label-Free Detection of
Biological Entities Using Nanowire-Based FETs. Ieee T Nanotechnol 2008, 7 (6),
651-667.
19. Ishikawa, F.N. Applications of one-dimensional structured nanomaterials as
biosensors and transparent electronics. University of Southern California, Los
Angeles, CA, 2009.
20. Heller, I.; Janssens, A.M.; Mannik, J., et al. Identifying the mechanism of
biosensing with carbon nanotube transistors. Nano Lett 2008, 8 (2), 591-595.
21. Nair, P.R.; Alam, M.A. Screening-limited response of nanobiosensors. Nano Lett
2008, 8 (5), 1281-1285.
22. Stern, E.; Wagner, R.; Sigworth, F.J., et al. Importance of the debye screening
length on nanowire field effect transistor sensors. Nano Lett 2007, 7 (11), 3405-
3409.
23. Tang, X.; Bansaruntip, S.; Nakayama, N., et al. Carbon nanotube DNA sensor and
sensing mechanism. Nano Lett 2006, 6 (8), 1632-6.
24. Schierbaum, K.D.; Weimar, U.; Gopel, W., et al. Conductance, Work Function
and Catalytic Activity of Sno2-Based Gas Sensors. Sensor Actuat B-Chem 1991,
3 (3), 205-214.
12
Chapter 2 Aligned Epitaxial SnO
2
Nanowires on Sapphire:
Growth and Device Applications
2.1 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,
1-3
pattern
transfer,
4
mechanical shear,
5
fluid flow in microchannels,
6
and orientation by an electric
field.
7
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.
8
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,
9
ZnO nanowires on sapphire,
10, 11
InAs
nanowires on InAs,
12
and InP nanowires on InP.
13
In this work, we show that synthesized
13
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)
14
material with high
electrical conductivity, optical transparency, and sensitivity to adsorbed molecules.
Successful applications of SnO
2
for field-effect transistors,
15, 16
transparent devices,
17
and
gas sensors
18-21
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 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.
2.2 Aligned nanowire growth and characterization
2.2.1 Saphire substrate surface study
We chose sapphire as the growth substrates for the aligned SnO
2
nanowires because of its
widely demonstrated ability to guide nanowire
9-11
and nanotube
22, 23
growth. We first
investigated the effect of annealing A (112 ¯ 0), M (101 ¯ 0), and R (11 ¯ 02) plane sapphire on
14
optimizing the growth conditions for the rutile structured SnO
2
nanowires. The
orientations for each of the three sapphire planes are shown in Figures 2.1 (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 2.1 (d)), M plane (Figure 2.1 (f)), and R
plane (Figure 2.1 (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 2.1 (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,
9, 11
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 [11
1] direction.
11
A- and R- plane sapphire, on the other hand,
retained their planar surface structure after annealing, as can be seen in Figures 2.1 (e)
and 2.1 (i), respectively . The lack of surface features on the A-plane and R-plane
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.
11
15
Figure 2.1 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.
2.2.2 Nanowire synthesis
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. 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
11
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.
23
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
[0001]
Al
2
O
3
(1102)
[0001] [0001] [0001]
a b
c
[2110]
[1210]
[1210]
[1210]
[1120]
Al
2
O
3
(1120)
A plane
Al
2
O
3
(1010)
M plane
Al
2
O
3
(1102)
R plane
Al 2O 3 (1120)
Al 2O 3 (1120) Al 2O 3 (1010) Al 2O 3 (1010) Al 2O 3 (1102)
[2110] [2110]
d e f g h i
[0001] [0001]
Al 2O 3 (1102)
16
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 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.
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, 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
17
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
as shown in Figure 2.2 (a). Conversely, Figure 2.2 (b) shows that SnO
2
nanowires aligned
on the annealed M-plane sapphire grew perpendicularly to the growth direction on A
plane, preferring to parallel the [112 ¯ 0] direction instead. Figure 2.2 (c) shows SnO
2
nanowires synthesized on the R-plane sapphire with a clear alignment in the [1 ¯ 101]
direction. A high resolution SEM image of an aligned nanowire on R plane in Figure 2.2
(d) shows smooth nanowire surface and a width around 123 nm. Further characterization
using SEM and AFM revealed that most of the wires are between 50 nm to 75 nm in
diameter and between 50 µm and 100 µm in length, sufficient for simple patterning of
source, drain and gate electrodes using standard photolithography techniques.
Figure 2.2 SEM images of aligned SnO
2
nanowires grown on (a) A-plane, (b) M-plane, and (c) R-plane
sapphires. (d) HRSEM of aligned SnO
2
nanowires on R-plane sapphire.
[1120]
[1102]
a b
c
SnO
2
NW
R-plane Al
2
O
3
100 nm
123 nm
d
18
We have also measured the nanowire density, alignment, and defect density (crossing or
crooked nanowires) using SEM images of 20 SnO
2
samples for each plane. As shown in
Figure 2.3, SnO
2
nanowires on the 20 R-plane sapphire samples 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 nanowires on A-plane
and M-plane are 2.79 ± 0.75 and 1.32 ± 0.46 nanowires/µm, respectively. Nanowires on
A-plane sapphire showed higher alignment defect density (1.38±0.65 nanowires/µm) than
nanowires on M-plane (0.53±0.26 nanowires/µm), but they have comparable alignment
angles (91% of nanowires are within ±1º).
Figure 2.3 Histograms of aligned SnO
2
nanowire assembly parameters. (a - c) Aligned SnO
2
nanowire
density on annealed A- plane (a), annealed M-plane (b), and R-plane (c) sapphire 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 (f) sapphire substrates 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 substrates (i)
0 1 2 3 4 5 6 7 8 9 10
0
5
10
15
20
Annealed M-plane
Number of Samples
Density (NW/ m)
Average density = 1.32 NW/ m
SD of density = 0.46 NW/ m
0 1 2 3 4 5 6 7 8 9 10
0
5
10
15
Number of Samples
Density (NW/ m)
Average density = 4.11 NW/ m
SD of density = 0.11 NW/ m
R-plane
0 1 2 3 4 5 6 7 8 9 10
0
2
4
6
8
10
Number of Samples
Density (NW/ m)
Average density = 2.79 NW/ m
SD of density = 0.75 NW/ m
A-plane
0 1 2 3 4
0
5
10
15
Annealed M-plane
Number of Samples
Defect Density (NW/ m)
Average of defect density = 0.53 NW/ m
SD of defect density = 0.26 NW/ m
0 1 2 3 4
0
2
4
6
8
10
A-plane
Number of Samples
Defect Density (NW/ m)
Average of defect density = 1.38 NW/ m
SD of defect density = 0.65 NW/ m
0 1 2 3 4
0
5
10
15
20
R-plane
Number of Samples
Defect Density (NW/ m)
Average of defect density = 0.35 NW/ m
SD of defect density = 0.34 NW/ m
a b c
d e f
g h
-5 -4 -3 -2 -1 0 1 2 3 4 5
0.1
1
10
100
Annealed M-plane
Number of Nanowires
Misalignment Angle (
O
)
i
-5 -4 -3 -2 -1 0 1 2 3 4 5
0.1
1
10
100
R-plane
Number of Nanowires
Misalignment Angle (
o
)
-5 -4 -3 -2 -1 0 1 2 3 4 5
0.1
1
10
100
A-plane
Number of Nanowires
Misalignment Angle (
O
)
19
2.2.3 X-ray diffraction study
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.
Figures 2.4 (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 2.4 XRD data for aligned SnO
2
nanowires grown on (a) A-plane, (b) M-plane, and (c) R-plane
sapphires show all three planes tend to interface the SnO
2
(101) plane.
2.2.4 Nanowire growth orientation
The preference of interfacing the SnO
2
(101) plane with sapphire is supported by several
previous reports,
24-26
from which the relative orientation between the aligned SnO
2
nanowires and the sapphire planes can also be predicted. For example, the interface
20 30 40 50 60
0
50
100
150
200
Sample
Stage
SnO
2
(200)
SnO
2
(101)
-
Al
2
O
3
(2204)
Intensity (a.u.)
2 Degrees
Al
2
O
3
(1102)
-
20 30 40 50 60
0
20
40
60 Sample
Stage
SnO
2
(200)
SnO
2
(101)
Intensity (a.u.)
2 Degrees
20 30 40 50 60
0
10
20
30
40
50
-
Al
2
O
3
(1120)
SnO
2
(101)
Sample
Stage
Intensity (a.u.)
2 Degrees
a
b c
20
between tilted SnO
2
nanowires and A-plane sapphire was reported as SnO
2
(101)
[
01]||Al
2
O
3
(11
0) [1
00],
24
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
2.5 (a) and (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 2.5 (a) aligns with the dashed axis in Figure 2.5 (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
[
01].
Additionally, the (101) surface of SnO
2
films have been observed to interface R-plane
sapphire
25, 26
with the two materials orientated such that SnO
2
(101) [010]||Al
2
O
3
(11 ¯ 02)
[112 ¯ 0]. This also agrees with the orientation for our aligned SnO
2
nanowires on R-plane
sapphire, and it is illustrated through the atomic arrangements of the SnO
2
(101) and
Al
2
O
3
(11 ¯ 02) surfaces shown in Figures 2.5 (b) and (c), respectively. The aligning axes
are drawn with the same line type again. Because the SEM image in Figure 2.2 (c)
revealed that nanowires grow along the Al
2
O
3
[11 ¯ 01] direction on R-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-plane sapphire, on the other hand, are aligned
along the nanogrooves, in which the exposed surface is mainly R plane.
11
Thus the
interface relationship on annealed M-plane sapphire is also SnO
2
(101) [010]||Al
2
O
3
(11 ¯
21
02) [11 2 ¯ 0]. However, because the nanogrooves are oriented in the Al
2
O
3
[11 2 ¯ 0]
direction, the nanowire growth direction on annealed M plane becomes SnO
2
[010].
Figure 2.5 Diagrams of atomic arrangement for (a) A-plane sapphire, (b) (101) plane SnO
2
, and (c) R-
plane sapphire. 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.
2.2.5 TEM confirmation of nanowire growth orientation
The aligned SnO
2
nanowire orientation on R-plane sapphire was further confirmed using
high resolution transmission electron microscopy (HRTEM) imaging of a cross-sectional
sample prepared by JEOL 4500 Focused Ion Beam (FIB). The schematic in Figure 2.6 (a)
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
nanowire growth. From the sample’s diffraction patterns, the growth direction of the
nanowire can be confirmed to be SnO
2
[1 ¯ 01] as shown in Figure 2.6(b), where two
normal planes are indicated to be SnO
2
(101) and SnO
2
(010) by the arrows. Similarly,
we can confirm that the nanowires grow along the sapphire [11 ¯ 01] direction from Figure
2.6(c), where the two normal planes are Al
2
O
3
(11 ¯ 02) and Al
2
O
3
(112 ¯ 0). The indexing of
the diffraction spots is confirmed using CrystalMaker® simulations for diffraction into
4.759Å
5.128Å
Al
2
O
3
(1102)
[1101]
[1120]
4.737Å
5.709Å
SnO
2
(101)
[010]
[101]
Al
2
O
3
(1120)
5.721Å
[1100]
4.336Å
[0001]
a b
c
22
the sapphire [1 ¯ 101] axis and the SnO
2
[1 ¯ 01] axis, which are overlaid in Figure 2.6 (d) and
(e), respectively.
Figure 2.6 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) Electron diffraction pattern of SnO
2
nanowire cross-section taken from the FIB cut sample. (c) Electron
diffraction pattern of cross-section of R-plane sapphire. (d) Enlarged image of (b) and overlaid with black
dots representing software simulated diffraction pattern of SnO
2
looking into the [1 ¯ 01] lattice vector
direction. (e) Enlarged image of (c) and overlaid with back dots representing software simulated diffraction
pattern of sapphire looking into the [1 ¯ 101] lattice vector direction. The faint rings are diffraction from the
protective Pt layer.
The HRTEM image of the aligned SnO
2
nanowire sample prepared according to Figure
2.6 (a) is shown in Figure 2.7. The inset image of Figure 2.7 shows a clearer view of the
lattice spacing of SnO
2
along [010] direction. This spacing and the interfacing plane
spacing of Al
2
O
3
in the [112
0] direction are measured to be approximately 4.8 Å, which
is close to lattice spacing specified in Figure 2.5(b) and (c). The locations from the cross-
c
Al
2
O
3
[1101]
1120
1102
SnO 2 (101)
SnO 2 [101]
SnO 2 (010)
Al 2 O 3 [1101]
a
Pt and C deposition
101
010
b
d
101
020
020
101
121
121
e
1120
1102
1102
1120
3124
1324
SnO
2
[101]
SnO
2
[101]
Al
2
O
3
[1101]
23
section used to obtain the diffraction patterns for the SnO
2
nanowire and sapphire are
indicated by the circled “ b” and “ c”, respecti vely. The orientations and interface planes
from TEM analysis show agreement with the XRD data and orientation data from
literature.
Figure 2.7 High resolution transmission electron microscopy (HRTEM) imaging of aligned SnO
2
nanowire
cross section on R-plane sapphire substrate. Inset shows spacing between the SnO
2
(010) planes. Normal
direction and plane index for both the nanowire and the sapphire substrate are labeled. Circled “ b” and “ c”
respectively correspond to spots where the diffraction pattern in Figure 2.5 for SnO
2
and sapphire are taken.
Pt
coating
4.8 Å
b
c
24
2.2.6 Effect of partial pressure on alignment
We also observed that synthesis pressure significantly affected whether the resultant
nanowires are parallel, planar SnO
2
nanowires or randomly oriented, free-standing
nanowires. By increasing the pressure inside the furnace to atmospheric pressure during
the synthesis, we obtained a higher density of SnO
2
nanowires on top of the Au catalyst.
Figure S3 (c) shows a SEM image of SnO
2
nanowires grown on R-plane sapphire under
such an atmospheric condition. The densely interlaced nanowires appear similar to
freestanding nanowire forests grown on Si substrates.
16
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
2
nanowire forest revealed no significant 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 conditions was
unsuccessful. This comparison suggests that synthesis at atmospheric pressure, where the
Sn vapor partial pressure is large, favors growth of dense nanowires that are forest-like
and un-aligned, while lower Sn partial pressure allows the nanowires to grow in
alignment on the substrate surface. Similar effect of partial pressure on nanowire growth
was also observed for InAs nanowires
12
, 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.
25
Figure 2.8. 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.
2.3 Field-effect transistor device fabrication and characterization
2.3.1 Device fabrication
Scalable device fabrication is an important step for practical integration of metal oxide
nanowires in applications like display and memory technology
15, 27
and various types of
sensors.
28, 29
After synthesis, aligned SnO
2
nanowires grown on sapphire were fabricated
as field effect transistors (FETs) using standard photolithography technology. 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
26
thickness of 5 nm and 75 nm respectively. Afterwards, a 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.
2.3.2 Device characterization
The finished device is represented in the schematic in Figure 2.9 (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 2.9 (b), and the alignment of SnO
2
nanowires across the
source and drain electrodes can be seen in the SEM image of Figure 2.9 (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 2.9 (d) and (e). In the drain current (I
D
)
versus the drain-to-source voltage (V
D
) plot of Figure 2.9 (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 2.9 (e) also
confirms that the current is turned off with V
G
< -4 V. The curve is measured at V
D
= 200
27
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 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.
15
Assuming that the aluminum oxide
dielectric constant is around 9,
30
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.
15, 31, 32
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.
16
This confirms the
high quality of our aligned SnO
2
nanowires and 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.
28
Figure 2.9 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.
Statistical data is also analyzed using 20 devices. Figure 2.10 (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 2.10 (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 2.10 (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 2.10 (d), where the average V
TH
is
29
-4.22V with a standard deviation of 0.81V. The histogram of electron mobility is shown
in Figure 2.10 (e) with an average of 71.68 cm
2
/V•s.
Figure 2.10 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).
2.4 Organic light-emitting diode (OLED) control circuit
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.
33, 34
To fully develop OLED as a low-cost, large-scale product,
0 -1 -2 -3 -4 -5 -6 -7
0
5
10
15
Number of Devices
Threshold Voltage (V
TH
,V)
Average V
TH
= -4.22 V
S.D. of V
TH
= 0.81 V
a
0 50 100 150 200
0
5
10
Average = 71.68 cm
2
/V.s
S.D. of = 43.80 cm
2
/V.s
Number of Devices
Mobility ( ,cm
2
/V.s)
e
c
d
b
0 100 200 300
0
5
10
15
Average g
m
= 157 nS
S.D. of g
m
= 37.23 nS
Number of Devices
Transconductance (g
m
,nS)
0 1 2 3 4
0
5
10
15
20
Average I
D
= 1.48 A
S.D. of I
D
= 0.33 A
Number of Devices
On State Current (I
D
, A)
1E4-1E5 1E5-1E6 1E6-1E7 1E7-1E8
0
5
10
15
20
Number of Devices
On/Off Ratio
30
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 2.11 (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 2.11 (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 2.11 (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 2.11 (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
OLED is very bright when V
G
= 10 V, gets dimmer as V
G
decreases toward negative
31
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 2.11 (b), where
the curve enters the cut-off region around -10 V.
Figure 2.11 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.
2.5 Photodetector and polarizer
SnO
2
has also been documented to have excellent photoconductive properties, whether as
a thin film
14
or as a nanowire,
16
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
-5 0 5 10 15 20 25
0
2
4
6
8
10 VG = 10V
Step = -2V
I
OLED
( A)
V
DD
(V)
-20 -15 -10 -5 0 5 10 15
0.0
0.1
0.2
0.3
0.4
0.5
0.6
VD = 15V
Step = -2V
I
OLED
( A)
Gate Voltage (V)
a b c
V
G
= 10V V
G
= 6V V
G
= 1V V
G
= -6V
V
G
= -10V
d
VG
VDD
I
OLED
VG
VDD
I
OLED
V
G
V
DD
=13V
I
OLED
32
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
measurements were taken in air, at room temperature, and under indoor incandescent
light. From the I
D
-V
D
data presented in Figure 2.12 (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.
35
The
mechanisms of photoconduction in metal oxides and the difference in conduction due to
the two wavelengths are well documented.
16, 36
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 2.12 (b). In this experiment, the V
D
of the device was
33
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 2.12 (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.
Moreover, the stability of photocurrent has been investigated using the same
photodetector that is used for experiments in Figure 2.12 (a-c). The aligned SnO
2
nanowire detector was illuminated by the 254 nm UV lamp for 100 minutes, and the
current is plotted in Figure 2.12 (d). Photocurrent rise sharply in the response upon UV
illumination, similar to Figure 2.12 (b), and reaches isteady state after 50 minutes with a
variation of 2 % between 50 and 100 minutes (Figure 2.12 (d) left inset). The long term
stability of this photodetector has also been investigated. Six months after we performed
measurements shown in Figure 2.12 (b) and (c), we performed similar photocurrent
measurement once every day for 15 days, and the right inset in Figure 2.12 (d) 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 an average
34
photocurrent of 123.05 ± 16.15 nA, which is in the same order of the magnitude as data
shown in Figure 2.12 (b) and (c) collected at the beginning of the 6 months.
Figure 2.12 Photoconduction (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 inset, 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. (c) Real-time
detection of 2 different wavelengths. (d) Real-time response from 254 nm illumination shows
photoconduction decreased and reached steady state after 60 minutes. Left inset shows 2% variation in
current from 50 and 100 minutes. Right inset shows stability test of photoconduction: each data point is the
average of photoconduction response over 200 s when exposed to 254 nm UV illumination.
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.
16, 35
SnO
2
nanowires possess both a one-dimentional structure
-1.0 -0.5 0.0 0.5 1.0
-0.4
-0.3
-0.2
-0.1
0.0
0.1
0.2
0.3
0.4
Before UV illumination
= 365 nm
= 254 nm
Drain Current ( A)
Gate Voltage (V)
-1.0 -0.5 0.0 0.5 1.0
-10
-5
0
5
10
Drain Current (nA)
Gate Voltage (V)
0 500 1000 1500
0
20
40
60
80
100
120
140
= 254 nm
Drain Current (nA)
Time (s)
a
b
c
0 200 400 600 800
0
25
50
75
100
125
150
175
= 365 nm
Drain Current (nA)
Time (s)
= 254 nm
0 20 40 60 80 100
0
30
60
90
120
150
180
210
240
Photocurrent (I
DS
, nA)
Time (minutes)
Turn 254 nm UV on
60 80 100
0.0
0.5
1.0
1.5
I/I0
Time (minutes)
d
0 5 10 15
0
50
100
150
Current (nA)
Time (Days)
Mean = 123.05 nA
S.D. = 16.15 nA
35
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 2.13 shows a 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 =0.3 was calculated according to = (G
||
-G⊥) /(G
||
+G⊥). This
polarization ratio is higher than previously observed values for that of GaN
35
and laser-
ablation grown SnO
2
16
nanowire devices, and equal to the observed value for that of
carbon nanotube devices.
37
This improvement further illustrates the significance of
controlled orientation of nanostructures for electronic applications.
Figure 2.13. 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.
-45 0 45 90 135 180 225
6
8
10
12
14
16
18
Measured ID
Fitted ID
Drain Current (nA)
Polarization Angle (
o
)
UV lamp
Linear
polarized
light
Polarizer
Device
36
2.6 Aligned SnO
2
nanowires for NO
2
sensing
Metal oxide nanowires have stimulated significant interest for chemical sensing and
biosensing applications, which have also been discussed in two recent review papers.
38, 39
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.
40-42
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 2.14 (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
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
37
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,
43
which is one of the most sensitive NO
2
nanosensors reported to date. This detection limit
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
20
and the 0.2 ppm limit from
SnO
2
nanowire sensors enhanced with additional resistance modulation.
21
Although many
other one-dimensional metal oxide NO
2
nanosensors have been reported
21, 44-49
, their
detection limits typically range from 1 ppm (as in the case of CuO nanowire sensors)
48
to
1 ppb (as statistically extrapolated for TiO
2
nanowires).
49
Normalized current change from the real-time sensing is plotted against the NO
2
concentration in Figure 2.14 (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 2.14 (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.
38
Figure 2.14 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
.
2.7 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
0.0 0.5 1.0 1.5 2.0
0.5
0.6
0.7
0.8
Device 1
| I/I
0
|
NO
2
Concentration (ppb)
0 1 2 3 4 5
1.2
1.4
1.6
1.8
2.0
| I
0
/ I |
1/Concentration (1/ppb)
-1.0 -0.5 0.0 0.5 1.0
-15
-10
-5
0
5
10
15
20
0 ppb
0.2 ppb
0.5 ppb
1 ppb
2 ppb
Drain Current ( A)
Drain Voltage (V)
Device 1
0 20 40 60 80 100 120 140 160
-1.2
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
Device 1
Device 2
I/I
0
Time (minutes)
2 ppb 1ppb
0.5ppb
b
a
0.2ppb NO
2
a
c b
39
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
nanowire FET for
polarized UV light detection, where the polarization ratio is higher than that from laser-
ablation synthesized nanowire FETs. And 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.
40
2.8 Chapter References
1. Jin, S.; Whang, D.; McAlpine, M.C., et al. Scalable Interconnection and
Integration of Nanowire Devices without Registration. Nano Letters 2004, 4 (5),
915-919.
2. Whang, D.; Jin, S.; Wu, Y., et al. Large-Scale Hierarchical Organization of
Nanowire Arrays for Integrated Nanosystems. Nano Letters 2003, 3 (9), 1255-
1259.
3. Kim, F.; Kwan, S.; Akana, J., et al. Langmuir-Blodgett nanorod assembly.
Journal of the American Chemical Society 2001, 123 (18), 4360-1.
4. Melosh, N.A.; Boukai, A.; Diana, F., et al. Ultrahigh-density nanowire lattices
and circuits. Science (New York, N.Y.) 2003, 300 (5616), 112-5.
5. Yu, G.; Cao, A.; Lieber, C.M. Large-area blown bubble films of aligned
nanowires and carbon nanotubes. Nature nanotechnology 2007, 2 (6), 372-7.
6. Huang, Y.; Duan, X.; Wei, Q., et al. Directed assembly of one-dimensional
nanostructures into functional networks. Science (New York, N.Y.) 2001, 291
(5504), 630-3.
7. Smith, P.A.; Nordquist, C.D.; Jackson, T.N., et al. Electric-field assisted assembly
and alignment of metallic nanowires. Applied Physics Letters 2000, 77 (9), 1399-
1401.
8. Yao, J.; Yan, H.; Lieber, C.M. A nanoscale combing technique for the large-scale
assembly of highly aligned nanowires. Nature nanotechnology 2013, 8 (5), 329-
335.
9. Tsivion, D.; Schvartzman, M.; Popovitz-Biro, R., et al. Guided growth of
millimeter-long horizontal nanowires with controlled orientations. Science (New
York, N.Y.) 2011, 333 (6045), 1003-7.
10. Nikoobakht, B.; Michaels, C.A.; Stranick, S.J., et al. Horizontal growth and in
situ assembly of oriented zinc oxide nanowires. Applied Physics Letters 2004, 85
(15), 3244-3246.
11. Tsivion, D.; Schvartzman, M.; Popovitz-Biro, R., et al. Guided growth of
horizontal ZnO nanowires with controlled orientations on flat and faceted
sapphire surfaces. ACS nano 2012, 6 (7), 6433-45.
12. Zi, Y.; Jung, K.; Zakharov, D., et al. Understanding Self-Aligned Planar Growth
of InAs Nanowires. Nano Lett 2013.
13. Mattila, M.; Hakkarainen, T.; Jiang, H., et al. Effect of substrate orientation on
the catalyst-free growth of InP nanowires. Nanotechnology 2007, 18 (15).
14. Camagni, P.; Faglia, G.; Galinetto, P., et al. Photosensitivity activation of SnO2
thin film gas sensors at room temperature. Sensor Actuat B-Chem 1996, 31 (1-2),
99-103.
15. Dattoli, E.N.; Wan, Q.; Guo, W., et al. Fully transparent thin-film transistor
devices based on SnO2 nanowires. Nano Lett 2007, 7 (8), 2463-9.
16. Liu, Z.Q.; Zhang, D.H.; Han, S., et al. Laser ablation synthesis and electron
transport studies of tin oxide nanowires. Adv Mater 2003, 15 (20), 1754-+.
41
17. Amin, N.; Isaka, T.; Yamada, A., et al. Highly efficient μm thick CdTe
solar cells with textured TCOs. Solar Energy Materials and Solar Cells 2001, 67
(1–4), 195-201.
18. Comini, E.; Faglia, G.; Sberveglieri, G. UV light activation of tin oxide thin films
for NO2 sensing at low temperatures. Sensor Actuat B-Chem 2001, 78 (1-3), 73-
77.
19. Maiti, A.; Rodriguez, J.A.; Law, M., et al. SnO2 nanoribbons as NO2 sensors:
Insights from first principles calculations. Nano Letters 2003, 3 (8), 1025-1028.
20. Law, M.; Kind, H.; Messer, B., et al. Photochemical sensing of NO2 with SnO2
nanoribbon nanosensors at room temperature. Angew Chem Int Edit 2002, 41
(13), 2405-2408.
21. Choi, Y.J.; Hwang, I.S.; Park, J.G., et al. Novel fabrication of an SnO(2)
nanowire gas sensor with high sensitivity. Nanotechnology 2008, 19 (9).
22. Ago, H.; Nakamura, K.; Ikeda, K., et al. Aligned growth of isolated single-walled
carbon nanotubes programmed by atomic arrangement of substrate surface. Chem
Phys Lett 2005, 408 (4-6), 433-438.
23. Ryu, K.; Badmaev, A.; Wang, C., et al. CMOS-Analogous Wafer-Scale
Nanotube-on-Insulator Approach for Submicrometer Devices and Integrated
Circuits Using Aligned Nanotubes. Nano Letters 2009, 9 (1), 189-197.
24. Leonardy, A.; Hung, W.-Z.; Tsai, D.-S., et al. Structural Features of SnO2
Nanowires and Raman Spectroscopy Analysis. Crystal Growth & Design 2009, 9
(9), 3958-3963.
25. Semancik, S.; Cavicchi, R.E. The growth of thin, epitaxial SnO2 films for gas
sensing applications. Thin Solid Films 1991, 206 (1–2), 81-87.
26. Dominguez, J.E.; Pan, X.Q.; Fu, L., et al. Epitaxial SnO[sub 2] thin films grown
on (1-bar 012) sapphire by femtosecond pulsed laser deposition. Journal of
Applied Physics 2002, 91 (3), 1060-1065.
27. Lei, B.; Li, C.; Zhang, D., et al. Nanowire transistors with ferroelectric gate
dielectrics: Enhanced performance and memory effects. Applied Physics Letters
2004, 84 (22), 4553-4555.
28. Ishikawa, F.N.; Chang, H.-K.; Curreli, M., et al. Label-Free, Electrical Detection
of the SARS Virus N-Protein with Nanowire Biosensors Utilizing Antibody
Mimics as Capture Probes. ACS nano 2009, 3 (5), 1219-1224.
29. Kolmakov, A.; Zhang, Y.; Cheng, G., et al. Detection of CO and O2 Using Tin
Oxide Nanowire Sensors. Adv Mater 2003, 15 (12), 997-1000.
30. Kordos, P.; Gregusova, D.; Stoklas, R., et al. Improved transport properties of
Al2O3/AlGaN/GaN metal-oxide-semiconductor heterostructure field-effect
transistor. Applied Physics Letters 2007, 90 (12).
31. Martel, R.; Schmidt, T.; Shea, H.R., et al. Single- and multi-wall carbon nanotube
field-effect transistors. Applied Physics Letters 1998, 73 (17), 2447-2449.
32. Zhou, C.W.; Kong, J.; Yenilmez, E., et al. Modulated chemical doping of
individual carbon nanotubes. Science (New York, N.Y.) 2000, 290 (5496), 1552-
1555.
42
33. Forrest, S.R.; Burrows, P.E.; Thompson, M.E. Organic Emitters Promise a New-
Generation of Displays. Laser Focus World 1995, 31 (2), 99-&.
34. Sheats, J.R.; Antoniadis, H.; Hueschen, M., et al. Organic electroluminescent
devices. Science (New York, N.Y.) 1996, 273 (5277), 884-888.
35. Han, S.; Jin, W.; Zhang, D.H., et al. Photoconduction studies on GaN nanowire
transistors under UV and polarized UV illumination. Chem Phys Lett 2004, 389
(1-3), 176-180.
36. Zhang, D.; Li, C.; Han, S., et al. Ultraviolet photodetection properties of indium
oxide nanowires. Appl Phys a-Mater 2003, 77 (1), 163-166.
37. Freitag, M.; Martin, Y.; Misewich, J.A., et al. Photoconductivity of single carbon
nanotubes. Nano Letters 2003, 3 (8), 1067-1071.
38. Po-Chiang, C.; Guozhen, S.; Chongwu, Z. Chemical Sensors and Electronic
Noses Based on 1-D Metal Oxide Nanostructures. Nanotechnology, IEEE
Transactions on 2008, 7 (6), 668-682.
39. Curreli, M.; Rui, Z.; Ishikawa, F.N., et al. Real-Time, Label-Free Detection of
Biological Entities Using Nanowire-Based FETs. Nanotechnology, IEEE
Transactions on 2008, 7 (6), 651-667.
40. Kolmakov, A.; Klenov, D.O.; Lilach, Y., et al. Enhanced gas sensing by
individual SnO2 nanowires and nanobelts functionalized with Pd catalyst
particles. Nano Letters 2005, 5 (4), 667-673.
41. Fryberger, T.B.; Semancik, S. Conductance Response of Pd/Sno2 (110) Model
Gas Sensors to H2 and O2. Sensor Actuat B-Chem 1990, 2 (4), 305-309.
42. Li, C.; Lei, B.; Zhang, D., et al. Chemical gating of In2O3 nanowires by organic
and biomolecules. Applied Physics Letters 2003, 83 (19), 4014-4016.
43. Pengfei, Q.F.; Vermesh, O.; Grecu, M., et al. Toward large arrays of multiplex
functionalized carbon nanotube sensors for highly sensitive and selective
molecular detection. Nano Letters 2003, 3 (3), 347-351.
44. Cao, B.B.; Chen, J.J.; Tang, X.J., et al. Growth of monoclinic WO3 nanowire
array for highly sensitive NO2 detection. J Mater Chem 2009, 19 (16), 2323-
2327.
45. Rout, C.S.; Ganesh, K.; Govindaraj, A., et al. Sensors for the nitrogen oxides,
NO2, NO and N2O, based on In2O3 and WO3 nanowires. Appl Phys a-Mater
2006, 85 (3), 241-246.
46. Zhang, D.H.; Liu, Z.Q.; Li, C., et al. Detection of NO2 down to ppb levels using
individual and multiple In2O3 nanowire devices. Nano Letters 2004, 4 (10),
1919-1924.
47. Ahn, M.W.; Park, K.S.; Heo, J.H., et al. On-chip fabrication of ZnO-nanowire gas
sensor with high gas sensitivity. Sensor Actuat B-Chem 2009, 138 (1), 168-173.
48. Kim, Y.S.; Hwang, I.S.; Kim, S.J., et al. CuO nanowire gas sensors for air quality
control in automotive cabin. Sensor Actuat B-Chem 2008, 135 (1), 298-303.
49. Kim, I.D.; Rothschild, A.; Lee, B.H., et al. Ultrasensitive chemiresistors based on
electrospun TiO2 nanofibers. Nano Letters 2006, 6 (9), 2009-2013.
43
Chapter 3 Highly Scalable, Uniform, and Sensitive
Biosensors Based on Top-down Indium Oxide Nanoribbon
Field-Effect Transistor
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
1-5
, nucleic acids
6, 7
, viruses
8
, and other small molecules
9
.
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.
1, 10
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.
2, 11-13
Among these, electron-beam lithography
14
, imprint
lithography
15
, spacer technique
12
, and photolithography
2
have been explored to optimize
dimensional control and scalability with promising results. Specifically, top-down
44
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.
11, 16
Metal oxides have been used traditionally as sensor materials because their surfaces are
very sensitive to changes in the environment.
17-19
The 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,
18, 20
biosensors,
3-5
optical detectors,
21
thin film transistors
(TFTs),
22
and other electronic applications.
23
In this work, we investigate sputtered metal
oxide
nanoribbons as the active channel in field-effect transistors (FETs) for biosensor
applications. The nanoribbon thickness is well controlled by the highly scalable radio
frequency (RF) sputtering technique, and the entire sensor is fabricated by a top-down, 2-
mask, conventional photolithography process. Oxide sputtering occurs at room
temperature, which eliminates any high-heat post-processing commonly required for
dopant diffusion in silicon platforms. Furthermore, no toxic gasses are needed during the
fabrication steps, and the lift-off process after sputtering circumvents the need for harsh
plasma etching. Moreover, sputtered metal oxide films are amorphous, and this lack of
grain boundaries can further improve the device-to-device variations in electrical signals.
45
3.2 Nanoribbon biosensor fabrication
The top-down fabrication process of In
2
O
3
nanoribbon biosensors is compatible with
complementary metal-oxide-semiconductor technology. We have chosen a Si substrate
with either a 50nm SiO
2
dielectric layer or a 500nm Si
3
N
4
dielectric layer for our
experiments, but the fabrication is easily transferred onto other types of substrates such as
flexible plastics or transparent glass. The thin 50nm SiO
2
dielectric layer is used for back-
gate device measurement as it provides a strong back-gate coupling. For sensing in fluid
with Ag/AgCl liquid gate, a thicker dielectric can be used. Furthermore, when the surface
chemistry needs to be contained on the nanoribbon surface, Si
3
N
4
is preferred over SiO
2
as the passivation layer because Si
3
N
4
is more resistant to our metal oxide surface
chemistry. The Si
3
N
4
layer is deposited on Si substrates by low pressure chemical vapor
deposition (LPCVD).
The completion of the nanoribbon biosensor requires two photolithography steps. The
first step defines the source and the drain metal electrodes of the transistor. The detailed
process is shown in Figure 3.1a-d. First we spin photoresist on the dielectric layer (a), and
the geometry and location of the electrodes are transferred onto the photoresist with good
precision and control through photolithography (b). Next, 5nm Ti and 50nm Au are
evaporated onto the exposed substrate in an electron beam evaporator (c), and a lift-off
process is used to wash away the photoresist to reveal the electrodes (d). A second
photolithography mask is used to fabricate the metal oxide channel according Figures
46
3.1e-h. The photoresist spin-coating (e) and pattern transfer (f) for the nanoribbon
geometry repeats the procedure used for metal electrode deposition. A layer of In
2
O
3
is
then deposited onto the substrate and part of the electrodes by RF sputtering with
thickness targeted at 10 nm to 50 nm (g). Nanoribbons are formed after the lift-off
process (h), which we leave till 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 3.1i 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 1j 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.
47
Figure 3.1 Fabrication processes of In
2
O
3
nanoribbons biosensors. (a) Low pressure chemical vapor
deposition of 500nm Si
3
N
4
on Si wafer followed by spin-coating of photoresist. (b) First photolithography
step defining source and drain metal electrodes. (c) Deposition of 5nm Ti and 50nm Au electrodes by
electron beam evaporation. (d) Photoresist lift-off to expose electrodes. (e) Spin-coating of photoresist for
transistor channel. (f) Second photolithography step defines the nanoribbon active channel. (g) In
2
O
3
deposition by radio frequency sputtering. (h) Photoresist lift-off to expose In
2
O
3
nanoribbon channel. (i) An
optical image In
2
O
3
nanoribbon biosensors on a 3 inch wafer. Inset shows a magnified nanoribbon chip
composed of 4 subgroups of 6 nanoribbon devices. (f) An SEM micrograph of two identical nanoribbon
devices.
c
b
d
a
In
2
O
3
j
50 µm
i
1.5 mm
1 cm
g
f
h
e
In
2
O
3
48
3.1 Transistor behavior of various metal oxide nanoribbons
After device fabrication, In
2
O
3
nanoribbon FETs were characterized by an Agilent
semiconductor analyzer 4156B using a back gate. Figures 2a and 2b 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 2a exemplifies a good metal oxide semiconductor field effect 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 2a and 2b 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. In Figure S2a – f in the Supporting Information, we have included family I
DS
-
0 10 20 30 40 50
0
1
2
3
V
GS
= 100 V, step = 10 V
Drain Current ( A)
Drain Voltage (V)
a
b
-100 -50 0 50 100
0
50
100
150
200
250
V
DS
= 1, step = 0.2 V
Drain Current (nA)
Back Gate Voltage (V)
49
V
DS
and I
DS
-V
GS
plots for sputtered InGaZnO, SnO
2
and ITO nanoribbon devices. Both
InGaZnO (Figures S2a and b) and SnO
2
(Figures S2c and d) show good MOSFET
behavior, but they are 10 and 100 times more resistive than the In
2
O
3
nanoribbon device,
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 S2e and f in the Supporting
Information) 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 S2g and h in the Supporting Information show optical
images of as-fabricated ZnO nanoribbons in air and 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.
50
Figure 3.3 Transistor family I
DS
-V
DS
and I
DS
-V
GS
curves for various metal oxide nanoribbon devices. (a)
I
DS
-V
DS
and (b) I
DS
-V
GS
for an InGaZnO nanoribbon device. (c) I
DS
-V
DS
and (d) I
DS
-V
GS
for a SnO
2
nanoribbon device. (e) I
DS
-V
DS
and (f) I
DS
-V
GS
for an ITO nanoribbon device. (g-h) Optical images of 6
ZnO nanoribbon devices before (g) and after (h) 14 hour incubation in PBS show that ZnO nanoribbons
dissolved completely in PBS after 14 hours.
InGaZnO Nanoribbon
SnO
2
Nanoribbon
0 20 40 60 80 100
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Drain Current (nA)
Back Gate Voltage (V)
V
DS
= 1.2 V, step = 0.3 V
a b
c
d
0 5 10 15 20
0
5
10
15
V
GS
= 80 V, step = 10V
Drain Current (nA)
Drain Voltage (V)
0 10 20 30 40 50
0
100
200
300
400
V
GS
=100 V, step = 10 V
Drain Current (nA)
Drain Voltage (V)
0 20 40 60 80 100
0
10
20
30
40
50
Drain Current (nA)
Back Gate Voltage (V)
V
DS
= 1 V, Step = 0.2 V
0 10 20 30 40 50
0
1
2
3
4
V
GS
=100 V, Step = 10 V
Drain Current (mA)
Drain Voltage (V)
ITO Nanoribbon
ZnO Nanoribbon
g h
e f
Before incubation
in PBS
80µm
-100 -50 0 50 100
0
20
40
60
80
100
V
DS
= 1 V, Step = 0.2 V
Drain Current ( A)
Back Gate Voltage (V)
80µm
After incubation
in 1xPBS for 14 hr
ZnO NR
51
3.2 Uniformity of In
2
O
3
nanoribbon FET
Statistical analyses of key electrical properties for 50 In
2
O
3
nanoribbon FETs are plotted
in Figure 2c to f. The dielectric used in these studies is 50 nm SiO
2
. The fifty devices
were taken across a 3” wafer to demonstrate that uniformity across a wafer is very high.
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
2c. 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 2d 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
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 2e. 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 2f 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
. To bench mark the uniformity of our “top -down” nanoribon devices,
the on-state current of 50 In
2
O
3
nanoribbon FETs (Figure 2g) is compared to that of 50
In
2
O
3
nanowire FETs shown in Figure 2h. In
2
O
3
nanoribbon devices show more uniform
52
on-state current than nanowire devices due to their high dimensional control. The SEM
insets in Figure 2g 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 2h. This variation is inherent in the “bottom -up” fabrication process.
53
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. (e) On-state current reproduced from (a) and plotted in logarithmic
scale. Inset SEM images of two representative nanoribbon devices show identical channel features. (f) On-
state current measured from 50 devices In
2
O
3
nanowire FET devices. Inset SEM images taken from two
representative devices show non-uniformity in the channel.
0 10 20 30 40 50
10
-9
10
-8
10
-7
10
-6
10
-5
10
-4
10
-3
10
-2
I
ON
(A)
Device Index
0 10 20 30 40 50
10
-11
10
-10
10
-9
10
-8
10
-7
10
-6
10
-5
10
-4
I
ON
(A)
Device Index
0 10 20 30 40 50
0
5
10
15
20
Average = 15.38 V
S.D. = 0.78 V
V
DS
= 0.6 V
V
TH
(V)
Device Index
0 10 20 30 40 50
0.0
0.5
1.0
1.5
I
ON
( A)
Device Index
Average = 919 nA
S.D. = 36 nA
V
DS
= 0.6 V, V
GS
=30 V
a
b
0 10 20 30 40 50
0
10
20
30
40
Average = 23.38 cm
2
/V.s
S.D. = 1.42 cm
2
/V.s
V
DS
= 0.6 V
cm
2
/V.s)
Device Index
0 10 20 30 40 50
10
0
10
1
10
2
10
3
10
4
10
5
10
6
10
7
10
8
On/off Ratio
Device Index
c d
e
f
In
2
O
3
nanowire FETs
2 µm 2 µm
i ii
In
2
O
3
nanoribbonFETs
50 µm
50 µm
ii
i
54
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 to investigate sensor uniformity in
the biosensing environment. Figure 3.5a shows that the average on-state current (I
ON
)
measured at V
DS
= 200 mV with a liquid gate voltage (V
LG
) of 1 V is 1.39 µA. The I
ON
standard deviation is 0.11 µA, or 8% of the average. The transconductance of these 30
devices (Figure 3.5b) falls within a narrow distribution between 3.5 µS to 4.5 µS, with
the average at 3.76 µS and a standard variation of 0.30 µS (8% of average). Figure 3.5c
shows that the variation in threshold voltage (V
TH
) is also small, with the average at 0.53
V and the standard variation at 0.01 V, which is only 2% of the average. Lastly, on-state
to off-state current ratios fall between 10
4
and 10
5
(Figure 3.5d). The small device-to-
device variations in liquid environment show good potential for biosensing.
55
Figure 3.5 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.
3.3 In
2
O
3
nanoribbon stability in ionic solution
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
0 5 10 15 20 25 30
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Average = 0.53 V
S.D. = 0.01 V
V
TH
(V)
Device Index
0 5 10 15 20 25 30
0.0
0.5
1.0
1.5
2.0
2.5
3.0
I
ON
( A)
Device Index
Average I
ON
= 1.39 A
S.D. of I
ON
= 0.11 A
0 5 10 15 20 25 30
10
0
10
1
10
2
10
3
10
4
10
5
10
6
On/off ratio
Device Index
0 5 10 15 20 25 30
0
1
2
3
4
5
6
g
m
( S)
Device Index
Average = 3.76 S
S.D. = 0.30 S
a
b
c
d
56
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 3.6 in the Supporting Information. In Figure S5a, 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 S5b 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 S5d, 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.
24
This advantage makes the In
2
O
3
nanoribbon platform a
promising candidate for in-vivo and in-vitro applications.
57
Figure 3.6 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.
3.4 pH sensing
Although the fabricated In
2
O
3
nanoribbon devices have demonstrated FET behavior, as
seeing in the previous sections, they must be sensitive to changes in ionic concentration
in order to perform as biosensors. To confirm this, we have conducted a series of pH
detection experiments with varying objectives to investigate the sensitivity of the
nanoribbons under different conditions. A general set up for pH sensing that is used for
0 20 40 60 80 100
0
2
4
6
8
10
Averager = 3.29 A
S.D. = 0.87 A
V
DS
= 0.2 V, V
LG
= 1 V
I
ON
( A)
Time (Days)
0 20 40 60 80 100
0.0
0.2
0.4
0.6
0.8
1.0
Averager = 0.53 V
S.D. = 0.07
V, V
DS
= 0.2 V
V
TH
(V)
Time (Days)
10 20 30 40 50 80 110
10
0
10
1
10
2
10
3
10
4
10
5
10
6
10
7
10
8
Time (Days)
on/off ratio
a
b
c d
0 20 40 60 80 100
0
5
10
15
20
Averager = 9.26 S
S.D.= 1.88 S, V
DS
= 0.2V
g
m
( S)
Time (Days)
58
all the pH experiments is shown in the schematic diagram of Figure 3.7a. 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 of varying pH in and out of the sensing chamber using
pipettes. Figure 3.7b shows the real-time sensing response, in the form of normalized
current, of a 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.
Figure 3.7 (a) Schematic diagram of set-up for pH sensing experiments using In
2
O
3
nanoribbon devices.
Commercial pH buffer solution is confined in a Teflon electrochemical chamber. Liquid gate voltage is
applied through a Ag/AgCl electrode. (b) Real-time, normalized current obtained from a 20 nm In
2
O
3
nanoribbon device exposed to commercial buffer solutions with pH 4 to 9.
a b
0 100 200 300 400 500 600
0
10
20
30
40
50
60
pH 4
pH 5
pH 6
pH 7
pH 8
I/I
0
Time (s)
pH 9
59
To test the nanoribbon device sensitivity to pH changes in a range that is 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 3.8a shows the
real-time response of three 20nm thick nanoribbon devices to pH change in this range.
These devices show a 5 times decrease in conduction with a pH change of 1.5. Figure
3.8b shows that the decrease of the normalized current is close to exponential as a
function of pH. Similar exponential conduction change in response to pH variation has
been observed from the unfunctionalized Si nanowire FET platform.
1
Figure 3.8 pH sensing in physiological pH range (a) Real-time sensing 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. (b) Plot of pH versus response for one of the devices.
For the nanoribbon conduction to show a strong change in response to charges from the
sensing environment, the optimal nanoribbon thickness needs to be within the transistor
Debye length, λ
D
. This parameter determines the distance into the semiconducting
nanoribbon at which point surface charges are no longer felt, and it is calculated by
6.5 7.0 7.5 8.0 8.5
0.0
0.2
0.4
0.6
0.8
1.0
I/I
0
pH
a b
200 400 600 800
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Device 1
Device 2
Device 3
I/I
0
Time (s)
pH 7.85
pH 8.18
pH 6.98
pH 6.72
pH 7.56
pH 7.32
60
,
11
where , k
B
, T, q, and n stand for the permittivity (7.97x10
-13
F/cm
for In
2
O
3
)
25
, the Boltzmann's constant, temperature, charge constant, and charge density,
respectively. The charge density is calculated as . The mobility, μ = 23.38
cm
2
/V-s, is taken from the average nanoribbon transistor mobility calculated in Figure
3.4c. The resistivity is calculated be
Ω-cm, where R is estimated as a
40MΩ channel resistance obt ained from I
DS
-V
DS
measurements and the area-to-length
ratio of the nanoribbon is estimated to be
. From these values, we
find that n = 2.48x10
16
cm
-3
, which gives a Debye length of λ
D
= 23nm.
Based on the calculated Debye length, we investigate the dependence of our sensor
performance on the nanoribbon thickness by comparing the response to pH change from
sensors with a ribbon thickness around λ
D
= 23nm. The different thicknesses are prepared
by sputtering 10nm, 20nm, 30nm, 40nm, and 50nm of In
2
O
3.
As shown in Figure 3.9,
conduction of all devices decreases exponentially when the pH increases from pH 4 to 9,
regardless of the nanoribbon depth. As expected, the 10 nm and 20 nm In
2
O
3
nanoribbon
devices are the most sensitive to the ionic changes 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 can be contributed from low film uniformity during In
2
O
3
deposition. For the rest
of the experiment, we target a 20 nm ribbon thickness, which gives us good stability and
sensitivity.
61
Figure 3.9 Real-time pH sensing responses from In
2
O
3
nanoribbon devices with ribbon thickness ranging
from 10 to 50 nm. The pH of the commercial buffer solutions used for sensing varies from pH 4 to 9. Error
bar indicates lowest and highest response among sensors tested for each data point.
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 300º C for 30 minutes in low vacuum
condition to obtain polycrystalline In
2
O
3
. Figure 3.10a shows X-ray diffraction (XRD)
patterns of as-sputtered (bottom) and annealed In
2
O
3
films (top). No In
2
O
3
peaks were
observed from the as-sputtered film, which confirms the amorphous nature of the as-
sputtered In
2
O
3
nanoribbons. In contrast, the XRD pattern of the annealed In
2
O
3
film
shows peaks for the (222), (400), (440) and (622) In
2
O
3
planes. These multiple peaks are
indicators of poly-crystallinity and were also observed by other researchers.
26
Next, we
performed pH sensing experiments to compare the results from both the polycrystalline
and the amorphous In
2
O
3
nanoribbon devices. Figure 3.10b shows average pH sensing
responses from three devices of annealed (red) and non-annealed (black) In
2
O
3
nanoribbon FETs. Both types of devices showed comparable performance in detecting
4 5 6 7 8 9
0
10
20
30
40
50
60
70
80
90
100
10nm
20nm
30nm
40nm
50nm
I/I
0
pH
62
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 3.10. Comparison of pH sensing response from annealed nanoribbon to that of non-annealed
nanoribbon sensors. (a) X-ray diffraction spectroscopy on 20 nm In
2
O
3
film shows sputtered In
2
O
3
is
amorphous before annealing (bottom) and polycrystalline after annealing in low vacuum at 300 ºC for 30
minutes (top). (b) Comparison of pH sensing responses obtained from the average of three as-sputtered and
three annealed devices exposed to commercial pH buffer solutions with pH in a range of 4 to 9.
3.5 Surface chemistry for molecular binding
A useful biosensor is not only sensitive to changes in its vicinity, but must also be
sensitive specifically to the target analyte. Specificity of the In
2
O
3
nanoribon biosensor is
achieved by chemically binding to the sensor surface a capture molecule that is chosen to
specifically recognize the analyte of interest. The surface chemistry for covalently
attaching the capture molecule to the In
2
O
3
nanoribbon surface is facilitated by a
phosphonic acid functional group linked with N-(3-Dimethylaminopropyl)-N′-
ethylcarbodiimide hydrochloride (EDC) and N-Hydroxysuccinimide (NHS). This
20 30 40 50 60
2 (Degrees)
As-Sputtered In
2
O
3
nanoribbon
Si
(622)
(440)
(400)
(222) Si
Annealed In
2
O
3
nanoribbon
Si
4 5 6 7 8 9
0
10
20
30
40
50
60
Annealed
As-Sputtered
I/I
0
pH
a b
63
coupling has been described with detail in previous studies on In
2
O
3
nanowire-based
devices.
3, 5
Mainly, a phosphonic linker molecule with a COOH terminal is bound to the
In
2
O
3
channel surface by incubating the sensor in the phosphonic acid solution.
EDC/NHS chemistry is then used to link the COOH group of the linker molecule to the
corresponding amine group of the capture probe protein.
Figures 3.11a shows a schematic diagram of how the phosphonic linker is applied on the
nanoribbon surface for the binding of biotin. A series of fluorescence experiments using
biotin and fluorescent dye-tagged streptavidin as the model probe-analyte system were
then conducted to confirm the success of this chemistry when applied to the nanoribbon.
As shown in Figure 3.11b, the negative controls are anchored with amine polyethylene
glycol (PEG) as the probe instead of biotin because PEG repeals the streptavidin so that
no fluorescence should be seen. Figures 3.11c and d show the schematic diagram and the
optical image, respectively, of the nanoribbon sample layout. We except to see
fluorescence on the In
2
O
3
part of the sample when the probe is biotin, and no
fluorescence when the probe is PEG. For the samples in Figures 3.11e and f, the metal
electrodes and the In
2
O
3
nanoribbons are deposited on SiO
2
substrates. As expected, we
see strong fluorescence on the In
2
O
3
ribbon when biotin probes are used (e) while little
fluorescence is seen on the In
2
O
3
when PEG probes are used (f). Additionally, the dark
electrodes in (e) confirm that the chemistry does not bind probes to the Au electrodes.
However, fluorescence on the SiO
2
substrate in both (e) and (d) is equivalent to that of
the In
2
O
3
, suggesting that the phosphonic linker also attaches well on SiO
2
. Although this
64
may be advantageous in cases where the density of the linker is more important than the
linker selectivity to In
2
O
3
, it can be undesirable when molecules need to bind exclusively
to the In
2
O
3
surface. For example, target analytes can be captured on both the In
2
O
3
nanoribbon and the SiO
2
substrate. This competition binding dilutes the analyte
concentration on the nanoribbon and can weaken the signal, if the signal is caused by
direct coupling of the analyte to the nanoribbon. In another experiment, Si
3
N
4
substrate
are used instead of SiO
2
, as shown in Figures 3.11g and h. Similar to Figures 3.11e and f,
biotin is the probe in (g) and PEG is the probe in (h). The background fluorescence on the
Si
3
N
4
substrates are significantly weaker than that on SiO
2
substrate, indicating that Si
3
N
4
substrates can successfully suppress competition binding.
65
Figure 3.11 (a) A schematic diagram of a nanoribbon functionalized with amine biotin to immobilize
streptavidin conjugated with fluorescent red dyes. (b) A negative fluorescence control for (a): nanoribbon is
functionalized with amine PEG, which cannot bind with streptavidin. (c) A schematic diagram of the
sample for fluorescence experiment: In
2
O
3
ribbons on Au metal electrodes, deposited on a dielectric
substrate. (d) An optical image of (c). (e) Fluorescent image of the sample with biotin probe and 500 nm
SiO
2
substrate. (f) Negative control of (e): sample with PEG probe and 500 nm SiO
2
substrate. (g)
Fluorescent images of the sample with biotin probe and 500 nm Si
3
N
4
substrate. (h) Negative control of (g):
sample with PEG probe and500 nm Si
3
N
4
substrate.
a b
c d
e f
g h
SiO
2
100 µm
100 µm
Si
3
N
4
100 µm
Si
3
N
4
100 µm
SiO
2
100 µm
66
3.6 Biomolecule detection in buffer
As a proof-of-concept demonstration of biological sensing, we show that the In
2
O
3
nanoribbon sensors can be used to detect the ovarian cancer biomarker 125 (CA-125) in
PBS buffer. Using the chemistry described in the previous section, we first conjugate the
surface of the In
2
O
3
nanoribbon channel with CA-125 antibody. The sensor is then
attached to a mixing cell and submerged in 500µl of PBS buffer. During the sensing
experiment, a 300mV source-drain voltage is used to obtain the current measurement, and
a 300mV liquid gate voltage is applied to the devices to bias the sensor in the desired
region of operation. To perform the detection, the pure PBS buffer solution was drawn
out and replaced with 500µl of buffer containing an initial concentration of the CA125
antigen. This solution exchange process is continued for PBS solutions containing
increasing concentrations of CA125. Real time sensing data is shown in Figure 3.12a,
where the normalized current versus time is plotted for three devices. In this plot, all
three devices showed uniform response to the introduction of the different concentrations
of CA-125. The detection limit is shown to be 0.1U/ml. This limit is comparable to the
detection limit reported for In
2
O
3
nanowire sensing of CA125 in blood serum
5
, and it is
two orders of magnitude more sensitive than the clinically relevant level for diagnosis.
27
Furthermore, the plot shows that the normalized signal is very uniform among the three
sensors that were tested simultaneously. This level of uniformity even without any further
calibration process reflects the advantage of the top-down nanoribbon structure over
67
bottom-up structures. This small variation is beneficial for the integration of a large array
of the nanosensors and for quantitative analysis of multiplex sensing data.
The CA125 antigen concentrations versus normalized current responses for all three
sensors are plotted in Figure 3.11b, along with their simulated Langmuir isotherm model
fitting. This model can be described by the following equation:
where ∆I/I
0
is the normalized response, n is the CA125 antigen concentration, and A and
α are simulated constants. Using this model, we can estimate the dissociation constant of
CA125 by calculating the concentration of the antigen when the percentage of occupied
probe binding sites, θ is one half. In other words, from the equation
the dissociation constant is 8 U/ml, within an order of magnitude of previous values
5
.
Figure 3.12 CA125 sensing (a) Real-time sensing of CA125 antigen with limit of detection at 0.1U/ml.
Plot shows response from three sensors performed simultaneously (b) Normalized current response versus
CA125 concentration for the corresponding sensors. Dotted data are fitted to simulated Langmuir isotherm
model.
1400 1600 1800 2000
0.92
0.94
0.96
0.98
1.00
1.02
100 U/ml
10 U/ml
1 U/ml
0.1 U/ml
Device A
Device B
Device C
I/I
0
Time (s)
0 20 40 60 80 100
0
1
2
3
4
5
6
Device A
Device B
Device C
Device A Simulated
Device B Simulated
Device C Simulated
I/I
0
CA125 concentration, U/ml
a b
68
3.7 Summary
In conclusion, we have demonstrated a top-down biosensor platform that is based on the
In
2
O
3
nanoribbon structure as the active channel of the FET device. We have shown that
not only does this type of sensor retain the label-free, fast-responding, and highly
sensitive qualities of other 1D electronic biosensor platforms, but it can also be fabricated
through a completely CMOS compatible process that uses conventional photolithography
and room temperature sputtering. We have demonstrated that these sensors show very
low device-to-device variations in multiple pH and CA125 antigen sensing experiments,
and that its sensitivity to antigens more than meets the clinically relevant limits. Such
qualities demonstrates that this is a platform with great potentials for developing
diagnostic tools that are low-cost, reliable, and compatible with many substrates, such as
glass or plastic, that are convenient for disposable, point-of-care diagnostic tools.
69
3.8 Chapter References
1. Cui, Y.; Wei, Q.; Park, H., et al. Nanowire nanosensors for highly sensitive and
selective detection of biological and chemical species. Science 2001, 293 (5533),
1289-92.
2. Stern, E.; Klemic, J.F.; Routenberg, D.A., et al. Label-free immunodetection with
CMOS-compatible semiconducting nanowires. Nature 2007, 445 (7127), 519-22.
3. Li, C.; Curreli, M.; Lin, H., et al. Complementary detection of prostate-specific
antigen using In2O3 nanowires and carbon nanotubes. Journal of the American
Chemical Society 2005, 127 (36), 12484-5.
4. Ishikawa, F.N.; Chang, H.K.; Curreli, M., et al. Label-free, electrical detection of
the SARS virus N-protein with nanowire biosensors utilizing antibody mimics as
capture probes. ACS nano 2009, 3 (5), 1219-24.
5. Chang, H.K.; Ishikawa, F.N.; Zhang, R., et al. Rapid, label-free, electrical whole
blood bioassay based on nanobiosensor systems. ACS nano 2011, 5 (12), 9883-91.
6. Hahm, J.; Lieber, C.M. Direct ultrasensitive electrical detection of DNA and
DNA sequence variations using nanowire nanosensors. Nano Lett 2004, 4 (1), 51-
54.
7. Gao, A.; Lu, N.; Dai, P., et al. Silicon-nanowire-based CMOS-compatible field-
effect transistor nanosensors for ultrasensitive electrical detection of nucleic acids.
Nano Lett 2011, 11 (9), 3974-8.
8. Patolsky, F.; Zheng, G.F.; Hayden, O., et al. Electrical detection of single viruses.
P Natl Acad Sci USA 2004, 101 (39), 14017-14022.
9. Lin, C.H.; Hsiao, C.Y.; Hung, C.H., et al. Ultrasensitive detection of dopamine
using a polysilicon nanowire field-effect transistor. Chem Commun (Camb) 2008,
(44), 5749-51.
10. Cui, Y.; Lauhon, L.J.; Gudiksen, M.S., et al. Diameter-controlled synthesis of
single-crystal silicon nanowires. Appl Phys Lett 2001, 78 (15), 2214-2216.
11. Elfström, N.; Karlström, A.E.; Linnros, J. Silicon Nanoribbons for Electrical
Detection of Biomolecules. Nano Lett 2008, 8 (3), 945-949.
12. Hakim, M.M.A.; Lombardini, M.; Sun, K., et al. Thin Film Polycrystalline
Silicon Nanowire Biosensors. Nano Lett 2012, 12 (4), 1868-1872.
13. Chang, H.-K.; Wang, X.; Aroonyadet, N., et al. Top-down Fabricated Polysilicon
Nanoribbon Biosensor Chips for Cancer Diagnosis. MRS Online Proceedings
Library 2013, 1569, null-null.
14. Park, I.; Li, Z.; Pisano, A.P., et al. Top-down fabricated silicon nanowire sensors
for real-time chemical detection. Nanotechnology 2010, 21 (1), 015501.
15. Wang, D.; Sheriff, B.A.; Heath, J.R. Silicon p-FETs from ultrahigh density
nanowire arrays. Nano Lett 2006, 6 (6), 1096-100.
16. Vacic, A.; Criscione, J.M.; Stern, E., et al. Multiplexed SOI BioFETs. Biosensors
& bioelectronics 2011, 28 (1), 239-42.
70
17. Wang, X.; Aroonyadet, N.; Zhang, Y., et al. Aligned Epitaxial SnO Nanowires on
Sapphire: Growth and Device Applications. Nano Lett 2014.
18. Zhang, D.H.; Liu, Z.Q.; Li, C., et al. Detection of NO2 down to ppb levels using
individual and multiple In2O3 nanowire devices. Nano Lett 2004, 4 (10), 1919-
1924.
19. Chen, P.C.; Sukcharoenchoke, S.; Ryu, K., et al. 2,4,6-Trinitrotoluene (TNT)
Chemical Sensing Based on Aligned Single-Walled Carbon Nanotubes and ZnO
Nanowires. Adv Mater 2010, 22 (17), 1900-+.
20. Li, C.; Zhang, D.; Han, S., et al. Synthesis, electronic properties, and applications
of indium oxide nanowires. Ann Ny Acad Sci 2003, 1006, 104-121.
21. Zhang, D.; Li, C.; Han, S., et al. Ultraviolet photodetection properties of indium
oxide nanowires. Appl Phys a-Mater 2003, 77 (1), 163-166.
22. Ju, S.; Facchetti, A.; Xuan, Y., et al. Fabrication of fully transparent nanowire
transistors for transparent and flexible electronics. Nature nanotechnology 2007, 2
(6), 378-84.
23. Chen, P.C.; Shen, G.Z.; Chen, H.T., et al. High-Performance Single-Crystalline
Arsenic-Doped Indium Oxide Nanowires for Transparent Thin-Film Transistors
and Active Matrix Organic Light-Emitting Diode Displays. ACS nano 2009, 3
(11), 3383-3390.
24. Zhou, W.; Dai, X.C.; Fu, T.M., et al. Long Term Stability of Nanowire
Nanoelectronics in Physiological Environments. Nano Lett 2014, 14 (3), 1614-
1619.
25. Bellingham, J.R.; Phillips, W.A.; Adkins, C.J. Electrical and Optical-Properties of
Amorphous Indium Oxide. J Phys-Condens Mat 1990, 2 (28), 6207-6221.
26. Yuan, Z.J.; Zhu, X.M.; Wang, X.O., et al. Annealing effects of In2O3 thin films
on electrical properties and application in thin film transistors. Thin Solid Films
2011, 519 (10), 3254-3258.
27. Niloff, J.M.; Bast, R.C., Jr.; Schaetzl, E.M., et al. Predictive value of CA 125
antigen levels in second-look procedures for ovarian cancer. American Journal of
Obstetrics & Gynecology 151 (7), 981-986.
71
Chapter 4 Top-down Polysilicon Nanoribbons: a
Comparison to In
2
O
3
Nanoribbon Biosensors
4.1 Introduction
Although we have demonstrated that metal oxide nanoribbon biosensor is a good
platform for combining the high sensitivity of one dimensional sensors with the
uniformity and scalability of “top -down” electronics, it is still worthwhile to explore a
scalable nanobiosensor platform based on silicon due to use of silicon as the standard for
high yield electronics in the industry. Several recent works have emerged in support of
using “top -down” nanoribbons instead of “bottom -up” na nostructures for field-effect
transistor biosensor platforms using single-crystalline silicon.
1-3
Despite the promising
result, single-crystalline silicon on insulator (SOI) wafers with extremely thin (~50nm)
active layer are required to produce the nanobiosensors, but SOI with such thin active
layers are expensive and not easily available. Also, precise oxidation and wet etching are
needed to achieve the desired thickness in single crystalline SOI, and thus further limit
the time and cost efficiency of such platforms. To find an optimal silicon-based material
for scalable biosensing applications, we investigate polysilicon for the fabrication of
“top-down” nanoribbon FET biosensors because polysilicon films under 50nm can be
achieved with good control and scalability. When they are deposited on silicon nitride or
72
silicon dioxide, the polysilicon-on-insulator substrate is cheaper than single crystalline
SOI substrates. We believe the benefits of a polysilicon nanoribbon sensor would be
more comparable to the In
2
O
3
nanoribbon sensors while its silicon characteristics can be
used to evaluate the performance of In
2
O
3
nanoribbon sensors.
In this chapter, we have fabricated polysilicon nanoribbon FET biosensors 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. Furthermore biomarker detection is
performed with clinically relevant sensitivity, and thus confirms the practical value of
polysilicon nanoribbon biosensors. Finally, we have compared the performance of
polysilicon nanoribbon biosensors with that of In
2
O
3
described in the previous chapter,
and differences of the two sensor types are outlined.
73
4.2 Device Fabrication
Several key differences between the polysilicon nanoribbon and the In
2
O
3
nanoribbon
biosensors emerge during the device fabrication process. Detailed fabrication process for
polysilicon nanoribbon transistor is illustrated in Figure 4.1. Although we have fabricated
both types of nanoribbon devices on a substrate consist of a silicon wafer with a dielectric
layer, as shown in the inset of Figure 4.1a, In
2
O
3
nanoribbons can be sputtered on other
versatile substrates such as flexible plastic. On the other hand, to achieve quality
polysilicon for the transistor channel, the thin layer of polysilicon seen in Figure 4.1a
needs to be deposited via low-pressure chemical vapor deposition (LPCVD) at a
temperature of around 600°C. At this high temperature, substrates for the polysilicon
deposition are limited to glass materials. This prevents the applications of silicon-based
nanobiosensors in areas such as flexible sensing, creating a significant advantage for
metal oxide sensors over silicon-based sensors. Polysilicon thin-film transistors have
been demonstrated on plastic substrates using pulse laser processing,
4
but this method is
incompatible with scalable technology. The LPCVD process also requires silane, SiH
4
,
which is a highly combustible gas that’s been linked to industrial accidents.
5
Sputtered
deposition of metal oxide uses argon and avoids any toxic gases. Aside from these
differences in the material deposition process, we are able to obtain uniform layers of
polysilicon between 35nm to 80nm through LPCVD. This is comparable to the sputtered
thickness of metal oxide and is a distinct advantage over nanobiosensors fabricated from
74
SOI, whose single crystalline silicon top layer is difficult to be thinned to this scale
without careful oxidation and etching processes.
A second difference between silicon-based material and metal oxide is that silicon-based
devices must be doped with extrinsic impurities to achieve useful conduction. Because
the polysilicon needs only light doping to achieve good semiconducting performance as
the active transistor channel, our investigation concludes that spin-on doping is the most
controllable and cost-effect method to lightly dope the channel in a scalable way, in
comparison to other doping methods such as in-situ doping during deposition and ion
implantation. In Figure 4.1b, spin-on boron dopants (purchased from Emulsitone
cooperation) of desired doping concentration is applied to the polysilicon via spin coating.
The diffusion is performed at 1100°C for 15 minutes in nitrogen environment. This post-
annealing stop is common for many methods of doping silicon-based semiconductor,
such as ion implantation,
6
and the high annealing temperature also excludes a variety of
substrates available to metal oxide FET biosensors. At the end of the diffusion process, a
buffered HF solution is used to remove spin-on dopant glass.
The first photolithography step finally begins after the preparation of the polysilicon layer,
as shown in Figure 4.1c. This photolithography step is the same process used for metal
oxide nanoribbons described in the previous chapter, and is a scalable process used in
standard CMOS fabrication. However, polysilicon and other silicon nanoribbons with a
nano-meter thickness necessary for biosensing cannot be shaped by a simple lift-off
75
process, and an etching step must be used. For the polysilicon, we find that a CF
4
reactive-ion etch (RIE) is effective for defining a 30nm to 80nm thick nanoribbon, and
this process is used to etch the contact lead and nanoribbon area, as seen in Figure 4.1d.
We note here that metal oxide nanoribbons, in contrast, are defined by a lift-off process
and therefore avoids toxic etching gases like CF
4
by eliminating the etching step.
After the polysilicon nanoribbon is etched down, the remaining fabrication (Figure 4.1e-g)
of source and drain metal electrodes is the same as that used for the In
2
O
3
nanoribbon
sensor electrodes. Mainly, a second photolithography step is performed to pattern the
metal contacts (e). Then HF dipping is required before the metal deposition to remove the
native oxide. Immediately following the HF etch, 5nm Ti and 45nm Au are deposited as
electrodes via electron beam evaporation (f). And finally, the electrodes are exposed by
lifting off the photoresist (g).
Although the fabrication of the electrodes is the same for the In
2
O
3
and the polysilicon
nanoribbon sensors, there is a difference in the sequence of the fabrication. In
2
O
3
nanoribbons are deposited during the last step of the sensor fabrication process, leaving
the In
2
O
3
surface free from being in contact with any photoresist. We believe this clean
surface is an advantage for the chemical binding of linker molecules to the nanoribbon.
The deposition of polysilicon makes it impossible to be left to the last step, and the
polysilicon nanoribbons are exposed to photoresist during the patterning of the metal
76
electrodes. This can make surface chemistry less reliable, depending on the cleaning
process.
Despite the disadvantages that the polysilicon nanoribbon FET fabrication faces when
compared to metal oxide nanoribbon sensors, the overall process is still efficient and
scalable with only two masks required. Shown in Figure 4.1(h) and 1(f) is a photographic
image of hundreds of polysilicon nanoribbon sensor arrays fabricated on a whole 3”
wafer. The yield is ~100%, and the structure of the devices exhibit very little device-to-
device variation due to the top-down process, as can be seen in the inset of two
nanoribbon device SEM images.
77
Figure 4.1 Polysilicon nanoribbon FET sensor fabrication (a) Polysilicon is deposited via LPCVD. Inset
shows Si/SiO
2
substrate (b) Conduction is aided by spin-on dopants that are thermally diffused into the
polysilicon. (c) Active mesas are patterns are transferred onto photoresist by the first photolithographic step.
(d) Dry etching transfers the nanoribbon pattern onto the polysilicon. (e) The second photolithography step
transfers the metal electrode patterns onto the photoresist. (f) Ti/Au is deposited into source and drain
electrode areas by metal evaporation. (g) Lift-off of photoresist exposes electrodes in the final step of the
FET sensor fabrication. (h) Photographic image of nanosensor arrays on a 3” wafer, and inset shows
zoomed-in SEM image of two identical polysilicon nanoribbons.
Boron doping
Spin-on dopant
Thermal diffusion
PR
Channel pattern on PR
Polysi channel etch by RIE
Electrode pattern on PR
Electrode deposition by
evaporation
Metal electrode lift-off
Polysi deposition by LPCVD
Channel :
2 X 5 μm
Wafer-scale polysi nanoribbon FETs
a
c
e
g
b
d
f
h
78
4.3 Electrical characteristics of polysilicon nanoribbon FETs
4.3.1 Dependence of performance on the dopant solution concentration
During the fabrication, spin-on boron dopant solutions of three different concentrations,
(C
0
= 1×10
17
, 5×10
17
and 1×10
18
) are applied to the polysilicon. We have characterized
the electrical properties for devices of all three doping concentrations with an Agilent
semiconductor analyzer. During the measurement, back gate voltage is applied to the
silicon substrate with silicon oxide as dielectric, and the result is shown in Figure 2.
Plotted in Figure 2 (a) and (b) is the source-drain current (I
ds
) versus source-drain voltage
(V
ds
) under various back gate voltage (V
g
) and I
ds
versus V
g
under various V
ds
for a
device with C
0
= 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. The I
ds
versus V
ds
under various V
g
and I
ds
versus
V
g
under various V
ds
for a device with C
0
= 5×10
17
doping concentration are shown in
Figure 2 (c) and (d), 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
g
and I
ds
versus V
g
under various V
ds
for a device with C
0
= 1×10
18
doping
concentration are shown in Figure 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.
79
Figure 4.2 (a) I
ds
versus V
ds
under various V
g
for a device doped with spin-on boron solution of C
0
=
1×10
17
(step: 10V). (b) I
ds
versus V
g
under various V
ds
for a device with 1×10
17
doping (step: 1V). (c) I
ds
versus V
ds
under various V
g
for a device with 5×10
17
doping (step: 10V). (d) I
ds
versus V
g
under various V
ds
for a device with 5×10
17
doping (step: 1V). (e) I
ds
versus V
ds
under various V
g
for a device with 1×10
18
doping (step: 10V). (f) I
ds
versus V
g
under various V
ds
for a device with 1×10
18
doping (step: 1V).
80
4.3.2 Statistical data of electrical performance
Top-down polysilicon nanoribbons also show good uniformity among 28 FET devices
tested. In Figure 4.3a, the on-current among the devices is within -10nA to -30nA when
taken at V
gs
= -40V and V
ds
= -1V. In Figure 4.3b, the on/off current ratio for the 28
devices fall within the same magnitude between 100 and 400. The transconductance data
is plotted in Figure 4.3c, where the range is shown to be between 4 nS and 10 nS. The
uniformity of the polysilicon nanoribbon FETs is further support that top-down
fabrication with relaxed lateral dimensions are advantageous in large-scale sensor
production.
Figure 4.3 Polysilicon FET statistical characteristics show good uniformity among 28 devices. (a) Device
on-current at V
gs
= -40V and V
ds
= -1 V. (b) Drain current on/ off ratio. (c) Device transconductance.
0 5 10 15 20 25
0
4
8
12
g
m
(nS)
Device Index
0 5 10 15 20 25
0
-10
-20
-30
-40
I
DS
(nA)
0 5 10 15 20 25
0
200
400
600
on/off
a
b
c
81
4.3.3 Comparison with In
2
O
3
nanoribbon FET sensor
A comparison between the electrical performance of the In
2
O
3
nanoribbon FET and the
polysilicon nanoribbon FET is demonstrated through two I
ds
-V
gs
curves typical of each
device type. Both of the nanoribbons shown in Figure 4.3 have a thickness around 50 to
55nm. The I
ds
-V
gs
plots are obtained using an Ag/AgCl liquid gate in the Teflon cell set-
up with 0.01xPBS solution as the dielectric. Liquid gate instead of back-gate
performances are used for comparison because the liquid set-up provides identical
dielectric conditions. Absolute values are taken for the current of the polysilicon
nanoribbon FET to make better comparison with the n-type In
2
O
3
nanoribbon device. The
liquid gate voltage is swept from a point with minimum drain current to Vgs = 0.6 V.
Figure 4.4 shows that for nanoribbons of similar thickness, In
2
O
3
nanoribbon FETs have
an on-current about one order of magnitude high than that of polysilicon nanoribbons,
while the on/off current ratio of In
2
O
3
nanoribbon FETs are about 2 orders of magnitude
higher. Higher current is advantageous for more stable sensing signals and more reliable
readout. The higher on/off ratio can provide In
2
O
3
nanoribbon FETs with a high
sensitivity than the polysilicon device.
82
Figure 4.4 Comparison of Ids-Vgs between In
2
O
3
and polysilicon nanoribbon FET sensors.
4.4 pH sensing
To demonstrate that the polysilicon FETs are ion-sensitive, we preform pH sensing with
various different parameters. The set-up for the sensing is similar to that used for In
2
O
3
nanoribbon sensing experiments, and is illustrated again in Figure 4.5a. During the
sensing experiment, a 200mV source-drain voltage and a 200mV liquid gate voltage is
applied to the devices. The source-drain current is constantly monitored by a
semiconductor analyzer. The devices were first soaked in a pH 2 buffer, and then the
buffer was exchanged to one with pH 4. The device in Figure 4.5b shows significant
increase in conductance after the buffer exchange, and such increase in conductance can
be explained by the p-type transistor behavior of the polysilicon devices, where the
decrease of the positive H
+
ion in more basic buffer solutions will create more channel
-0.2 0.0 0.2 0.4 0.6
1E-10
1E-9
1E-8
1E-7
1E-6
1E-5
55 nm Polysi NR
Drain Current (A)
Liquid Gate Voltage (V)
50 nm In
2
O
3
NR
83
carriers for conductions. Further buffer exchanges were performed to change the pH to 6,
6, 8, and 10, respectively. The variation in conductivity from pH 2 to pH 10 is around
~600%. Further pH sensing experiment is carried out in buffers of physiological range
with a pH step of 0.2, and the result is shown in Figure 4.5c. From pH 7.2 to pH 8, the
device shows a conduction increase of 50% with high signal-to-noise ratio, suggesting
the polysilicon nanosensor is highly sensitive to ions.
Figure 4.5 pH sensing using polysilicon nanoribbon FET sensors (a) Sensing set-up using Ag/AgCl liquid
gate and Teflon cell for enclosing fluids. (b) Signal (normalized current) for pH range from pH 2 to pH 10
in steps of 2. (b) Signal for detecting pH in the physiological range in steps of 0.2.
Reponses to pH from polysilicon nanoribbons of two different thicknesses are compared
in Figure 4.6a. As with the In
2
O
3
nanoribbons, sensitivity increases with decreased
thickness from 80 nm to 55 nm, suggesting that the same ionic concentration can deplete
higher percentage of carriers inside the polysilicon of thinner ribbons due to the Debye
0 200 400 600 800 1000 1200
0
1
2
3
4
5
6
7
I/I
0
Time (s)
pH10
pH8
pH6
pH2
pH4
100 200 300 400 500 600 700
1.0
1.1
1.2
1.3
1.4
1.5
pH8
pH7.8
pH7.6
pH7.4
pH7.2
I/I
0
Time (s)
a
c
b
Ag/AgCl Solution Gate
Electrode
Buffers of various pH
Polysi
84
length screening, which is estimated to be around 10 nm to 40 nm
2
with the dopant
concentration used in these experiments.
In Figure 4.6b, the dependence of the polysilicon nanoribbons’ sensitivity on doping
concentration is investigated. By increasing the spin-on boron dopant solution
concentration by an order of magnitude from the 10
17
to the 10
18
, the conduction change
from pH 2 to pH 10 decreased by about a half. This is expected because the Debye length
decreases as the carrier concentration inside the semiconductor increases. Lower the
doping concentration further in polysilicon resulted in a decrease in current, which
lowered the stability of the signal and made sensing read-out difficult. We find the 10
17
dopant solution to be the best compromise between a stable current signal and sensitivity.
Furthermore, decreasing the carrier concentration below this point by cost-effective spin-
on dopant is difficult to control. Therefore, even though lowering the doping may achieve
more sensitive detection, we believe this is the optimal doping condition that doesn’t
compromise other device and yield parameters.
A comparison between the pH sensing of a 55 nm polysilicon nanoribbon device and a 50
nm In
2
O
3
nanoribbon device is shown in Figure 4.6c. Between nanoribbons of similar
thickness, the conduction change from pH 4 to pH 9 is 3 times higher in the In
2
O
3
nanoribbon. This can be contributed from the higher on/off ratio as seen in the previous
section.
85
Figure 4.6 pH sensing using polysilicon nanoribbon FET sensors (a) Sensing set-up using Ag/AgCl liquid
gate and Teflon cell for enclosing fluids. (b) Signal (normalized current) for pH range from pH 2 to pH 10
in steps of 2. (b) Signal for detecting pH in the physiological range in steps of 0.2.
4.5 Biomarker detection
To test the performance of polysilicon nanoribbon sensors in molecular detection, CA125
antigen is chosen to be the model biomarker because of its association with several types
of cancers.
7
The polysilicon devices are first conjugated with CA125 antibody using the
surface chemistry developed by Thompson’s group at USC, and shown in the inset of
Figure 4.7a. During the sensing experiment, a 200mV source-drain voltage and liquid
3 4 5 6 7 8 9
1
2
3
4
5
6
55 nm polysi NR
80 nm polysi NR
Exponential fitting
Exponential fitting
I/I
0
pH
0 2 4 6 8 10
0
1
2
3
4
5
[dopant]=10
17
[dopant]=10
18
Fitting of 10
17
Fitting 10
18
I/I
0
pH
4 5 6 7 8 9
0
5
10
15
20
Polysi NR
Exponential fitting
I/I
0
pH
9 8 7 6 5 4
In2O3 NR
Exponential fitting
a
c
b
86
gate voltage are applied to the device. Higher concentrations of CA125 are progressively
added to the sensor surface.
The normalized current versus time for one polysilicon nanoribbon device is plotted in
Figure 4.7a. The current begins to respond to the introduction of CA125 solutions at a
concentration of 1U/ml (5pM). This limit of detection is two orders of magnitude lower
than the clinically relevant level for diagnosing ovarian cancer.
8
The device shows larger
responses to CA125 at the higher concentrations of 10 U/ml and 100 U/ml. In
comparison, the polysilicon sensor did not show a significant response to 250nM bovine
serum albumin (BSA), as shown around t = 760s in Figure 4.7a, even though the BSA
concentration is several hundred times higher than the target analyte. This is good
indication that the detection is specific to CA125, and the response to BSA can be further
suppressed using passivation methods. The response versus CA125 concentration is
plotted in Figure 4.7b along with the simulated Langmuir isotherm model fitting. From
this model, we obtain a dissociation constant of 22 U/ml for CA125, which is between
the value predicted by our In
2
O
3
nanoribbon sensors from the previous chapter and that
predicted using In
2
O
3
nanowires.
9
The results suggest that polysilicon nanoribbon
biosensors have great potential as a platform for clinical diagnosis of diseases such as
cancers.
We use CA125 sensing as a metric for comparing the performance of polysilicon and
In
2
O
3
nanoribbon sensors in their detection of biomolecules. In Figure 4.7c, we note that
87
the In
2
O
3
nanoribbon sensor has a detection limit that is one order of magnitude lower
than that of the polysilicon nanoribbon. Although both detection limits meets the
clinically relevant limit for ovarian cancer, a lower detection limit is desired for early
detection of various diseases when the biomarker level is still very low. The sensitivity of
the In
2
O
3
nanoribbon sensor at each concentration is also about twice that of the
polysilicon nanoribbon. Several factors can contribute to the superior detection limit and
sensitivity of the In
2
O
3
nanoribbon sensors. The higher on/off ratio of the In
2
O
3
sensor
was previously described in section 3. Another contributing factor is that silicon-based
materials have an insulating native oxide layer that acts as an additional dielectric layer
between the molecules and the sensor. This is not the case for metal oxides as the entire
material is semiconducting. The oxide layer can be etched away for molecules to directly
bind to the silicon or the polysilicon.
10
However, this extra chemistry adds complexity to
the sensing and user interface.
88
Figure 4.7 CA-125 sensing. (a) Real time sensing of CA-125 antigen using polysilicon nanoribbon sensors.
(b) Comparison of CA125 concentration versus response from polysilicon and In
2
O
3
nanoribbon sensors. (c)
Surface chemistry for attaching capture molecules to polysilicon nanoribbon.
4.6 Summary
In conclusion, we have developed a label-free, electrical biosensor platform based on
polysilicon nanoribbons that can be fabricated by conventional photolithography with
only easily available materials and equipments required, thus results in great time and
cost efficiency as well as scalability. The devices show great response to pH changes
with a wide dynamic range and high sensitivity. Biomarker detection is demonstrated
0.1 1 10 100
0
1
2
3
4
5
6
Polysi NR
In2O3 NR
Linear fitting
Linear fitting
I/I0 (%)
CA125 Concentration, U/ml
b c
400 500 600 700 800
1.00
1.02
1.04
1.06
I/I
0
Time (s)
1U/ml 10U/ml
100U/ml
250nM BSA
a
0 20 40 60 80 100
0
1
2
3
4
Polysi nanoribbon
Langmuir-Isotherm fitting
I/I0 (%)
CA125 Concentration, U/ml
89
using CA-125, an ovarian cancer biomarker, as model with clinically relevant sensitivity.
Conductivity of multiple devices is monitored simultaneously during pH sensing
measurements, and the data confirms the uniformity of device performance. Such results
suggest that polysilicon nanoribbon devices exhibit good potential to act as a reliable
sensitive platform for future nanobiosensors.
In comparison to In
2
O
3
nanoribbon sensors, however, polysilicon nanoribbon sensors
have several key disadvantages, and they must be addressed for silicon based nanoribbon
biosensors to become a preferred platform for practical applications. Substrate versatility
is important as biosensors become more integrated with everyday life, and this challenge
is faced by all silicon-based biosensors. Obtaining a very thin active layer of silicon must
also become more accessible, controllable, and cost-effective. Although this can be
greatly improved by LPCVD of polysilicon instead of using single crystalline silicon, a
thickness below 30nm is still difficult. In contrast, sputtered metal oxides can easily reach
10nm, which is advantageous for pushing the detection limit of biomarker disease for
early diagnosis, without complex fabrication. Overall, although both types of sensors are
sufficient for clinical applications, In
2
O
3
nanoribbon biosensors may give more reliable
sensing with more straightforward fabrication.
90
4.7 Chapter References
1. Stern, E.; Klemic, J.F.; Routenberg, D.A., et al. Label-free immunodetection with
CMOS-compatible semiconducting nanowires. Nature 2007, 445 (7127), 519-22.
2. Elfström, N.; Karlström, A.E.; Linnros, J. Silicon Nanoribbons for Electrical
Detection of Biomolecules. Nano Lett 2008, 8 (3), 945-949.
3. Hakim, M.M.A.; Lombardini, M.; Sun, K., et al. Thin Film Polycrystalline
Silicon Nanowire Biosensors. Nano Lett 2012, 12 (4), 1868-1872.
4. Carey, P.G.; Smith, P.M.; Theiss, S.D., et al. Polysilicon thin film transistors
fabricated on low temperature plastic substrates. J Vac Sci Technol A 1999, 17 (4),
1946-1949.
5. Chen, J.R. Characteristics of fire and explosion in semiconductor fabrication
processes. Process Safety Progress 2002, 21 (1), 19-25.
6. Vacic, A.; Criscione, J.M.; Stern, E., et al. Multiplexed SOI BioFETs. Biosensors
& bioelectronics 2011, 28 (1), 239-242.
7. Bast, R.C., Jr.; Xu, F.J.; Yu, Y.H., et al. CA 125: the past and the future. The
International journal of biological markers 1998, 13 (4), 179-87.
8. Visintin, I.; Feng, Z.; Longton, G., et al. Diagnostic markers for early detection of
ovarian cancer. Clin Cancer Res 2008, 14 (4), 1065-1072.
9. Chang, H.-K.; Ishikawa, F.N.; Zhang, R., et al. Rapid, Label-Free, Electrical
Whole Blood Bioassay Based on Nanobiosensor Systems. ACS Nano 2011, 5 (12),
9883-9891.
10. Bunimovich, Y.L.; Shin, Y.S.; Yeo, W.S., et al. Quantitative real-time
measurements of DNA hybridization with alkylated nonoxidized silicon
nanowires in electrolyte solution. Journal of the American Chemical Society 2006,
128 (50), 16323-16331.
91
Chapter 5 Application of In
2
O
3
Nanoribbon Biosensors and
Electronic Enzymen-Linked Immunosorbent Assay for Chest
Pain Diagnosis
5.1 Introduction
Every year about 5 million patients visit the emergency department (ED) because of chest
pain symptoms.
1
If an initial electrocardiogram (ECG) assessment at the ED reveals a
ST-segment deviation, the patient is placed at high risk for acute myocardial infarction
(AMI), or heart attack, and the established medical procedures are administered to the
patient. However, ECG sensitivity may be low as 50%,
2-5
and patients who show no ST
elevation can still be at high risk for unstable angina or non-ST segment elevation AMI.
For these cases, cardiac biomarkers have become increasingly important for swiftly risk
stratifying and diagnosing patients who may still need immediate treatment.
The effectiveness of the biomarkers to properly diagnose and triage chest pain patients is
based on several factors. First, test turnaround time should be fast because early treatment
of myocardial infarction is crucial to recovery. The American Heart Association has
stated a recommended 60 minute, and a preferred 30 minute, turnaround time from
sample collection to result reporting.
6
Second, obtaining the trend in the cardiac
92
biomarker concentration in the hours after patient’s arrival is a crucial addition to the
initial cardiac biomarker reading for an accurate diagnosis. Current biomarker trends are
collected through serial biomarker readings, such as testing at 0, 30, 60, and 90 minutes
after patient arrival at the ED.
7
Such fast turnaround times are difficult to achieve in a
central laboratory setting and is often aided by a point-of-care (POC) device.
8
Additionally, multiple cardiac biomarkers testing may improve the diagnosis process of
heart attack over single biomarker testing.
9
The National Academy of Clinical
Biochemistry has recommended testing for an early biomarker that elevates within the
first 6 hours of chest pain in conjuncture with an AMI specific biomarker that is
increased in the blood even after 6 to 9 hours.
10
Point-of-care platforms are ideal for
multiple cardiac biomarker testing with fast turnaround times, but current POC devices
lack the sensitivity and specificity of central laboratory biomarker testing.
11
For POC
devices to more effectively aid rapid decision making in both the ED and on the field,
there’s a need for further investigation of emerging sensor technology in order to bridge
the performance gap between POC device and central laboratory testing for cardiac
biomarkers.
As seen in the CA125 detection in previous chapters, the In
2
O
3
nanoribbon sensor
response to target analytes reaches a stable response within 2 minutes. This can provide a
shorter sample- collection-to-result time than the 17 minutes accomplished by the POC
device currently used in the ED.
12
The quick response time makes the In
2
O
3
nanoribbon
sensor especially advantageous for analyzing the first blood-draw sample, from which
93
rapid decisions are made for the patients’ treatment. The small device -to-device variation
previously demonstrated among the In
2
O
3
nanoribbons can provide good statistical
confidence for calibrating cardiac biomarker concentrations. Furthermore, In
2
O
3
nanoribbon sensors can provide quantitative analysis for a large detectable concentration
range spanning at least 3 orders of magnitude and a detection limit in the femtogram per
mililiter range. This sensitivity can help to differentiate biomarker changes at each serial
reading. Because the sensing is electronic, the completed product supports simple
interface and compactness while having the capability to integrate with other microfluidic
and electronic functional groups, such as wireless data output. These properties make
In
2
O
3
nanoribbon sensors well suited for analyzing medical conditions such as heart
attack that need urgent, point-of-care (POC) medical attention.
In this chapter, we demonstrate how top-down In
2
O
3
nanoribbon biosensors can be
optimized for the quantitative detection of cardiac biomarkers using 3 protein biomarkers
commonly associated with heart attack and heart failure as model. First, troponin is used
to investigate a shortened protein incubation time from 2 hours to 15 minutes. Then
creatine kinase –MB (CK-MB) is used to demonstrate how to optimize the detection
range by creating fewer binding sites for the signal protein. Lastly, we demonstrate how
an electronic enzyme-linked immunosorbent assay (ELISA) is used to process whole
blood in a straightforward manner for BNP detection.
94
5.2 Electronic enzyme-linked immunosorbent assay (ELISA)
Direct electrical detection of biomolecules in their physiological environment is often
impeded by Debye screening from the high salt concentration in the sample solutions.
Sandwich ELISA,
13
on the other hand, detects signals associated with the reactions
between a test solution and the conjugated enzymes on secondary antibodies instead of
the biomarker. The sandwiched structure not only overcomes the Debye screening from
salt in the fluid 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 apply an electronic ELISA
technique that uses pH change due to urease enzyme activity as the amplification signal.
The details of the assay are shown in Figure 5.1, and the main sequence is as follows. The
nanoribbon sensor surfaces are first chemically modified by a phosphonic acid linker
group in the same way as described in literature
14, 15
and Chapter 3. The nanoribbon
sensor is enclosed by the mixing cell described in Chapter 3. A solution of capture
antibodies specific to the cardiac biomarker is first introduced to the surface of the sensor
and become immobilized by the surface modification. After incubation the mixing cell is
rinsed with buffer. Then the cardiac biomarkers are introduced to the sensor surface by
incubation. The biomarkers are contained either within the physiological fluid sample of
the patient or in a solution of buffer for experimental purposes. The biomarkers are
95
subsequently captured by the antibody, and any unbound excess is washed off. Next a
solution of biotinylated secondary antibody also specific to the cardiac biomarker is
introduced to sensor by incubation and binds itself to the biomarkers. After rinsing out
un-bound biotinylated antibody, streptavidin solution in PBS is incubated next. The
biotin end of the secondary antibody group is used to bind to a streptavidin, which in turn
is bound to a biotinylated urease, which is the last solution to incubate the sensor.
Figure 5.1 Schematic diagram of experimental set up and electronic ELISA for cardiac biomarker.
When a solution of urea is introduced to nanoribbon sensor surface with this sandwich
structure, the urea causes an increase in the pH of the solution due to consumption of
hydrogen ions according to the reaction
16
:
Cardiac Biomarker
(CB)
CB
Dielectric
96
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
nanoribbon sensors because the amount of charges released during the pH increase is
very high. It is much higher than the amount of charge transferred during the direct
binding between the analyte and the capture antibody. This amplifies the detection signal
and allows the sensor to detect very low concentrations of the analyte. Furthermore, the
solution for the pH detection step is independent of the fluid containing the biomarker,
since the solutions are rinsed out after each step. This allows cardiac biomarkers to be
collected in a physiological samples such as whole blood without the limitation of the
Debye screening effect.
16
5.3 Shortened incubation time of Troponin
Troponin is the biomarker of choice for evaluating chest pain patients for possible heart
attack.
8
Elevated blood troponin levels have a positive correlation to the risk of death in
the heart disease patients,
17, 18
and the biomarker is a good guide for identifying patients
for certain types of treatment.
8
From a study of 1818 patients presented with chest pain,
the interquartile range (IQR) for non-AMI patients is between 5 to 9 pg/ml while the
range for AMI patients is 59 to 1918 pg/ml.
19
In the first biomarker detection experiment,
97
we use troponin as the model cardiac biomarker to demonstrate how the In
2
O
3
nanoribbon biosensor can be used to optimize the electronic ELISA assay turnaround
time by shortening the incubation of the cardiac biomarker.
The recommended incubation time for protein biomarkers to bind to capture antibodies in
ELISA essays is around 1-3 hours at 37°C, or longer in room temperature.
20
However,
previous In
2
O
3
nanoribbon sensing experiments such as the CA125 detection in Chapter 3
have shown that binding between the protein and the antibody on the nanoribbon can be
around 2 minutes. In Figure 5.2 we explore the sensitivity when the troponin biomarker
incubation time is reduced from 2 hours to 15 minutes, both in room temperature. Figure
5.2a repeats the sensing set-up for the electronic ELISA experiment, with antibodies
specific to troponin as the capture probe and the biotinylated secondary detection probe.
Voltage to the Ag/AgCl liquid electrode and the drain is supplied by Agilent B1500 as
before, which also monitors the current change. Before sensing begins, the device is
linked with all the molecules as shown in Figure 5.2a. At time t = 0 in Figure 5.2b, the
device is rinsed with and submerged in 0.01x diluted PBS buffer when the baseline
current is taken. The buffer is then replaced with 10 mM urea in 0.01xPBS around 500s
as indicated by the arrow. Figure 4b shows the real-time responses when the urea solution
is introduced into the sensing chamber that was previously incubated in 300 pg/ml of
troponin in PBS buffer. The urease-urea interaction drastically reduces the device
conductance by 88.5% of the baseline signal. The pH change between the buffer solution
used for the baseline and the final solution in the sensing chamber is measured to be 1.45
98
by a commercial Mettler Toledo pH meter. This increase in pH is consistent with the
decrease in conduction of the In
2
O
3
nanoribbon device.
The 300pg/ml troponin in buffer used for Figure 5.2b was incubated for 2 hour in room
temperature. In Figure 5.2c, we repeat the experiment in 5.2b, but the troponin incubation
time is 15 minutes. This is an 8 time decrease in incubation time, and the current decrease
in the shortened assay time is still quite high at 82%. Next we performed the same
experiment for troponin concentrations covering the beginning of the 2
nd
quartile for non-
AMI patients to the median of AMI patients. In Figure 5.2d, we show the average
sensitivity for 3 sensors at several troponin concentrations within this range with a
biomarker incubation time of 2 hours (black) and 15 minutes (red). We see that the 15
minute biomarker incubation time reduces the saturation of detection signal at troponin
concentrations higher than 10 pg/ml, without significantly compromising sensitivity. This
holds great advantage for reducing turnaround time of cardiac biomarker analysis in the
clinic.
99
Figure 5.2 Troponin sensing (a) Schematic diagram of electronic ELISA for tronponin sensing. (b) Real
time eELISA signal for 3 In2O3 nanoribbon sensors performing detection of 300 pg/ml troponin, which
was incubated with the nanoribbon for 1 hour. (c) same experiment as (b) with troponin incubation time
reduced to 15 minuts. (d) Troponin biomarker concentration versus signal for 4 concentrations of troponin
incubated 2 hours with nanoribbon sensor and 3 concentrations with the shorter assay.
5.4 Precision adjustment of kinase –MB (CK-MB) concentration
The blood biomarker creatine kinase-MB (CK-MB) has long been used for AMI
detection. However CK-MB levels in the blood of healthy persons is relatively high
compared to AMI patients, with an interquartile range level of non-AMI patients at 0.6
ng/ml to 1.7 ng/ml, and that of AMI patients from 1.5 ng/ml to 10.5 ng/ml.
19
The
0.1 1 10 100
0
20
40
60
80
100
2 hours
15 minutes
(%)
Troponin I Concentration (pg/ml)
0 500 1000 1500 2000
0.0
0.3
0.6
0.9
1.2
1.5
Device A
Device B
Device C
I/I
0
Time (s)
10mM Urea
in 0.01xPBS
[Troponin I] = 300pg/ml
I/I
0
=81.97%
a
d
b
c
0 500 1000 1500 2000 2500
0.0
0.3
0.6
0.9
1.2
1.5
Device 1
Device 2
Device 3
I/I
0
Time (s)
10mM Urea
in 0.01xPBS
[Troponin] = 300 pg/ml
ΔI/I
0
= 88.48%
ΔpH =1.45
2 hour incubation
15 minute incubation
100
detection of CK-MB must be able to distinguish this less than an order of magnitude
difference in order to effectively diagnose using the CK-MB biomarker. In this project,
we use CK-MB as a model biomarker to investigate how to optimize the precision of
In
2
O
3
nanoribbon biosensors towards quantifying CK-MB biomarkers in this range.
In the troponin sensing experiment of the previous section, we showed that the current of
the In
2
O
3
nanoribbon sensor drops to about 65% of the baseline at a troponin
concentration of 10 pg/ml and 55% at a concentration of 1 pg/ml. This sensitivity is about
10% conduction change per decade of biomarker concentration change. While this is
beneficial for covering a large range of concentrations for markers like troponin whose
elevation in AMI patients is high, 10% per decade may not be sensitive enough for CK-
MB, whose concentration of interest can be a small range within the one decade between
1 ng/ml to 10 ng/ml. To obtain a higher percentage of current change per concentration so
that we can more precisely differentiate between the CK-MB concentrations of interest,
we biotinylated only a fraction of the secondary CK-MB antibodies. This scheme is done
by Thompson’s lab at USC and is shown in Figure 5.3a and 5.3b. Most of the secondary
CK-MB antibodies in (a) are covalently tagged with biotin, so streptavidin and
biotinylated urease can be subsequently immobilized onto the nanoribbon surface with an
urease-to-CKMB ratio of nearly 1:1. By targeting a biotin-to-antibody binding at a
smaller efficiency, a large fraction of the secondary CK-MB antibodies are left without
biotinylation, as shown in (b). Streptavidin can no longer be immobilized to un-
biotinylated secondary antibodies, which in turn reduces the urease to CK-MB ratio. In
101
Figure 5.3c, we demonstrate through real time sensing that CK-MB detection with the
lower urease-to-marker ratio is still stable for 1 ng/ml of CK-MB with a sensitivity of
43% conduction change. The sensing is repeated for 3 ng/ml and 5 ng/ml, at which
concentration heart failure is strongly indicated. The average of 3 sensor data for each
concentration is plotted in Figure 5.3d with standard deviation amongst plotted as the
error bar. At 5 ng/ml, the current is 83% of baseline, giving a 40% sensitivity range for a
concentration difference of 4 ng/ml. This large range can enable detection of minute
changes in concentration, such as from 2.5 ng/ml to 3 ng/ml. In addition, the small
device-to-device signal standard deviation makes readout at this precision possible with
the In
2
O
3
nanoribbon sensor platform.
Figure 5.3 CK-MB sensing (a) Biotinylation of secondary CKMB antibody targeting full biotinylation. (b)
Partial biotinylation of secondary CKMB antibodies. (c) Real time electronic ELISA sensing of 1 ng/ml
CK-MB biomarker in PBS buffer. (d) Average sensing response of 3 nanoribbon devices at 1 ng/ml, 3
ng/ml, and 5 ng/ml of CK-MB markers in buffer with standard deviation among the 3 sensors as error bar.
0 1000 2000 3000
0.0
0.2
0.4
0.6
0.8
1.0
Device 1
Device 2
Device 3
I/I
0
Time (s)
10mM
Urea
[CK-MB] = 1 ng/ml
I/I
0
= 43.36%
c d
In
2
O
3
Nanoribbon
Urease
Streptavidin
biotin
CKMB antibody
CKMB
In
2
O
3
Nanoribbon
a
b
0 1 2 3 4 5 6
40
50
60
70
80
90
I/I
0
(%)
CK-MB Concentration (ng/ml)
102
5.5 Detection of B-type natriuretic peptide (BNP) in whole blood
Detection of cardiac biomarkers in whole blood is essential to POC sensor platforms used
for situations where complicated patient blood processing is not possible and defeats the
purpose of fast, cheap, and convenient disease testing. The main problems for FET
nanosensor detection caused by whole blood are unspecific binding of non-target proteins
and Debye length screening from salts, as previously described. Recent efforts to process
whole blood for FET nanosensors have been demonstrated using a microfluidic chip,
21
desalting columns,
22
and filtration.
15
However, we demonstrate in the next experiment
that by applying electronic ELISA assay on In
2
O
3
nanoribbon sensors, we can detect the
cardiac biomarker B-type natriuretic peptide (BNP) in whole blood whithout any sample
processing at all.
BNP is associated with heart failure and have been shown to substantially improve AMI
diagnosis when included in a multiple cardiac biomarker panel.
23
More importantly, for
blood samples taken when chest pain patients first arrive at the emergency department,
BNP is shown to be more sensitive for AMI diagnosis than other cardiac biomarkers such
as CKMB and troponin,
24
which do not elevate until at least 2 hours after the onset of
AMI symptoms. In fact, even when a patient’s troponin level is normal, a BNP
concentration greater than 100 pg/ml is a good indicator of AMI.
24
BNP higher than 900
pg/ml is considered severe heart failure.
25
In order to simulate quantitative detection of
BNP in patients’ blood, we first obtained a standard calibration curve with known
103
concentrations of BNP in buffer. As shown in Figure 5.4, we have targeted BNP
concentrations of 100 fg/ml, 1 pg/ml, and 10 pg/ml. For each of the 3 concentration, 3
In
2
O
3
nanoribbon sensors were used in the electronic ELISA assay as described in the
previous sections. Figure 5.4a shows the real time sensing curve for the smallest
concentration of 100 fg/ml. The current drops to 59% of the baseline after the
introduction of 10mM Urea solution. For the 1 pg/ml (Figure 5.4b) and the 10 pg/ml
(Figure 5.4c) detection, the current drops to 72% and 83% of baseline, respectively. The
average and the standard deviation for each of the three concentrations are plotted in
Figure 5.4d. The largest device-to-device signal deviation is 3.8% and comes from the
detection of 0.1 pg/ml BNP. At 1 and 10 pg/ml, the standard deviations are both 0.31%.
In logarithmic scale, the linear fitting has an R-squared value of 0.9936, suggesting a
good fit for the BNP concentration calibration curve.
104
Figure 5.4 Real time electronic ELISA assay signal for detecting (a) 0.1 pg/ml, (b) 1 pg/ml, and (c) 10
pg/ml of BNP in buffer using In
2
O
3
nanoribbon biosensor. (d) Average sensing response of (a) – (c) with
standard deviation among the sensors as error bar. Logarithmic fitting curve with R-squared value of
0.9936 shows good fit for BNP calibration curve.
Next, BNP is spiked with healthy whole blood to simulate an AMI patient sample with a
BNP concentration of 500 pg/ml, indicating mild heart failure.
25
Because the nanoribbon
sensor is more sensitive than the simulated concentration, a real patient sample at 500
pg/ml would be first diluted 100 times to be detected by the nanoribbon sensor. We note
that this sample dilution is not due to difficulties in ionic screening and does not filter out
any non-specific proteins or blood cells. To simulate this sample preparation, 2 μl of
healthy whole blood is first diluted 00 times with xPBS to 200 μl. Then 2 μl of 5 pg/ml
a b
0 1000 2000 3000
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Device A
Device B
Device C
I/I
0
Time (s)
[BNP] = 10 pg/ml
I//I
0
=83.31%
10mM
Urea
0 500 1000 1500
0.0
0.2
0.4
0.6
0.8
1.0
Device A
Device B
Device C
I/I
0
Time (s)
10mM Urea
0.01x
PBS
[BNP] = 100 fg/ml
I/I
0
= 58.77%
0 500 1000
0.2
0.4
0.6
0.8
1.0
1.2
[BNP] = 1 pg/ml
I/I
0
= 72.74%
Device A
Device B
I/I
0
Time (s)
10 mM
Urea
d
c
0.1 1 10
0
20
40
60
80
100
(%)
BNP Concentration (pg/ml)
105
BNP in 1xPBS is added to the diluted whole blood to simulate 500 pg/ml BNP in whole
blood that’s diluted 00 times. This sample is used to incubate the nanoribbon sensor
prepared with BNP capture antibodies. The remaining steps of the electronic ELISA
assay follows those described in previous sections. Figure 5.5a shows the real time signal
when 10mM of urea is introduced to the sensor. The current decreased to 77% of the
baseline, and the standard deviation among the 3 sensors used for the detection is 3.15%.
In Figure 5.5b, the average response of the 3 In
2
O
3
nanoribbon sensors for this detection
is placed on the calibration curve of Figure 5.4d as a red circle. We expect it to fall on the
calibration curve for BNP = 5 ng/ml. The graph shows that the deviation of the detection
signal from the calibration curve is only 3%. This falls within the device-to-device
variation and is expected for the experiment.
Figure 5.5 (a) Real time electronic ELISA assay signal for detecting 5 pg/ml of BNP in 100x diluted whole
blood In
2
O
3
nanoribbon biosensor. (d) Average sensing response of (a) placed on the concentration
calibration curve for BNP.
0 1500 3000 4500
0.00
0.25
0.50
0.75
1.00
1.25
1.50
Device A
Device B
Device C
I/I
0
Time (s)
10mM
Urea
[BNP] = 5 pg/ml in whole blood
I/I
0
= 77.15%
SD = 3.15%
a b
0.1 1 10
55
60
65
70
75
80
85
90
BNP in PBS
BNP in Whole Blood
(%)
BNP Concentration (pg/ml)
Δ =3%
106
5.6 Summary
In conclusion, we’ve demonstrated that In
2
O
3
nanoribbon biosensors can be used to
quantitatively detect 3 cardiac biomarkers within the concentrations relevant to clinical
diagnosis. Through all the sensing experiments, we have achieved device-to-device signal
variation of less than 4% without additional calibration methods necessary for nanowire
biosensors.
26
This uniformity is a characteristic of the top-down nanosensor and is an
advantage for qualitatively interpreting biomarker concentrations. In a production sitting,
further improvements can be made toward the uniformity by monitoring the nanoribbon
film thickness after sputtering and chemical modification to reduce the device-to-device
variation down a fraction of a percentage. Such highly uniform batches of sensors can
give good statistical confidence for the biomarker concentration that they report. This
confidence level combined with a turnaround time on the order of minutes is a good basis
for improving current POC devices for cardiac marker detection in an emergency
situation. Moreover, the platform can be integrated with other electronic components for
better data analysis.
107
5.7 Chapter References
1. Nourjah, P. National Hospital Ambulatory Medical Care Survey: 1997 emergency
department summary. Advance data 1999, (304), 1-24.
2. Brush, J.E.; Brand, D.A.; Acampora, D., et al. Use of the Initial
Electrocardiogram to Predict in-Hospital Complications of Acute Myocardial-
Infarction. New Engl J Med 1985, 312 (18), 1137-1141.
3. Gibler, W.B.; Young, G.P.; Hedges, J.R., et al. Acute myocardial infarction in
chest pain patients with nondiagnostic ECGs: serial CK-MB sampling in the
emergency department. The Emergency Medicine Cardiac Research Group.
Annals of emergency medicine 1992, 21 (5), 504-12.
4. Lee, T.H.; Rouan, G.W.; Weisberg, M.C., et al. Sensitivity of routine clinical
criteria for diagnosing myocardial infarction within 24 hours of hospitalization.
Annals of internal medicine 1987, 106 (2), 181-6.
5. Rude, R.E.; Poole, W.K.; Muller, J.E., et al. Electrocardiographic and Clinical-
Criteria for Recognition of Acute Myocardial-Infarction Based on Analysis of
3,697 Patients. Am J Cardiol 1983, 52 (8), 936-941.
6. Braunwald, E.; Antman, E.M.; Beasley, J.W., et al. ACC/AHA guidelines for the
management of patients with unstable angina and non-ST-segment elevation
myocardial infarction: executive summary and recommendations. A report of the
American College of Cardiology/American Heart Association task force on
practice guidelines (committee on the management of patients with unstable
angina). Circulation 2000, 102 (10), 1193-209.
7. Ng, S.M.; Krishnaswamy, P.; Morissey, R., et al. Ninety-minute accelerated
critical pathway for chest pain evaluation. Am J Cardiol 2001, 88 (6), 611-617.
8. Lewandrowski, K.; Chen, A.; Januzzi, J. Cardiac markers for myocardial
infarction. A brief review. American journal of clinical pathology 2002, 118
Suppl, S93-9.
9. Han, J.H.G., W.B., Biomarkers in the Emergency Department: Rapid Diagnosis
and Triage. In Cardiovascular Biomarkers: pathophysiology and disease
management, Morrow, D. A., Ed. Humana Press: 2006.
10. Wu, A.H.B.; Apple, F.S.; Gibler, W.B., et al. National Academy of Clinical
Biochemistry standards of laboratory practice: Recommendations for the use of
cardiac markers in coronary artery diseases. Clin Chem 1999, 45 (7), 1104-1121.
11. Lee-Lewandrowski, E.; Januzzi, J.L., Jr.; Grisson, R., et al. Evaluation of first-
draw whole blood, point-of-care cardiac markers in the context of the universal
definition of myocardial infarction: a comparison of a multimarker panel to
troponin alone and to testing in the central laboratory. Archives of pathology &
laboratory medicine 2011, 135 (4), 459-63.
12. Lee-Lewandrowski, E.; Benzer, T.; Corboy, D., et al. Cardiac Marker Testing As
Part Of An Emergency Department Point ‐of ‐Care Satellite Laboratory In A Large
Academic Medical Center: Practical Issues Concerning Implementation. Point of
Care 2002, 1 (3), 145-154 10.1097/01.POC.0000023109.92641.C7.
108
13. Butler, J.E. Enzyme-linked immunosorbent assay (Reprinted from
Immunochemistry, pg 759-803, 1994). J Immunoassay 2000, 21 (2-3), 165-209.
14. Li, C.; Curreli, M.; Lin, H., et al. Complementary detection of prostate-specific
antigen using ln(2)O(3) nanowires and carbon nanotubes. J Am Chem Soc 2005,
127 (36), 12484-12485.
15. Chang, H.K.; Ishikawa, F.N.; Zhang, R., et al. Rapid, Label-Free, Electrical
Whole Blood Bioassay Based on Nanobiosensor Systems. Acs Nano 2011, 5 (12),
9883-9891.
16. Stern, E.; Vacic, A.; Li, C., et al. A nanoelectronic enzyme-linked immunosorbent
assay for detection of proteins in physiological solutions. Small 2010, 6 (2), 232-
8.
17. Newby, L.K.; Christenson, R.H.; Ohman, E.M., et al. Value of serial troponin T
measures for early and late risk stratification in patients with acute coronary
syndromes. Circulation 1998, 98 (18), 1853-1859.
18. Ohman, E.M.; Armstrong, P.W.; Christenson, R.H., et al. Cardiac troponin T
levels for risk stratification in acute myocardial ischemia. New Engl J Med 1996,
335 (18), 1333-1341.
19. Keller, T.; Zeller, T.; Ojeda, F., et al. Serial Changes in Highly Sensitive
Troponin I Assay and Early Diagnosis of Myocardial Infarction. Jama-J Am Med
Assoc 2011, 306 (24), 2684-2693.
20. Crowther, J.R., ELISA: Theory and Practice. Humana Press Inc.: Totowa, New
Jersey 1995; Vol. 42.
21. Stern, E.; Vacic, A.; Rajan, N.K., et al. Label-free biomarker detection from
whole blood. Nat Nanotechnol 2010, 5 (2), 138-142.
22. Zheng, G.F.; Patolsky, F.; Cui, Y., et al. Multiplexed electrical detection of cancer
markers with nanowire sensor arrays. Nat Biotechnol 2005, 23 (10), 1294-1301.
23. Christenson, R.H.A., H., Biomarkers of Myocardial Necrosis. In Cardiovascular
biomarkers: pathophysiology and disease management, Morrow, D. A., Ed.
Humana Press Inc.: Totowa, New Jersy, 2006.
24. Bassan, R.; Potsch, A.; Maisel, A., et al. B-type natriuretic peptide: a novel early
blood marker of acute myocardial infarction in patients with chest pain and no ST
segment elevation. Eur Heart J 2005, 26 (3), 234-240.
25. B-type Natriuretic Peptide (BNP) Blood Test.
http://my.clevelandclinic.org/services/heart/diagnostics-testing/laboratory-tests/b-
type-natriuretic-peptide-bnp-bloodtest (January 2, 2015),
26. Ishikawa, F.N.; Curreli, M.; Chang, H.K., et al. A Calibration Method for
Nanowire Biosensors to Suppress Device-to-Device Variation. Acs Nano 2009, 3
(12), 3969-3976.
109
Chapter 6 Conclusion and Future Directions
6.1 Conclusion
In this thesis, nanostructure, semiconducting metal oxides as the active channel material
in field-effect transistors (FET) have been investigated for various sensing applications
ranging from photo detection to chemical and biological sensing. The overarching
objective of the thesis is to overcome the obstacle of applying metal oxide nanostructures
to FET sensing technology in a scalable and controllable process that’s necessary for
practical applications. To this end, scalable processing was explored from both a
“bottom-up” approach and a “top -down” approach. Chapter 2 demonstrated a “bottom -up”
synthesis of epitaxial SnO
2
nanowires with controlled alignment, guided by the lattice of
the substrate. The easy of integrating the well-aligned nanowires with straightforward
photolithography was demonstrated in chemical sensing applications, where the quality
of the nanowires stemming from their alignment is believed to have contributed to the
ultra sensitive detection NO
2
down to 0.2 ppb. In Chapter 3, a “top -down” fabrication of
In
2
O
3
nanoribbon was demonstrated to produce uniform and sensitive FET biosensors.
However, because silicon-based electronics is still the standard for high yield electronics,
comparison of fabrication and performance was made between In
2
O
3
nanoribbon and
silicon based FET sensors. In chapter 4, polysilicon nanoribbon sensors were investigated
as the comparable silicon-based device because polysilicon films 50nm or thinner can be
110
readily achieved with good precision, unlike single-crystalline silicon-on-insulator film
that use less precise oxidation and wet etching cycles for thinning. Additionally
polysilicon is lower cost. By comparing fabrication, electronic performance, pH sensing,
and cancer biomarker sensing between the polysillicon nanoribbon and the In
2
O
3
nanoribbon sensors, it was concluded that the In
2
O
3
nanoribbon platform is more
advantageous for biosensing due to its compatibility with versatile substrates, higher
electronic stability, lower device-to-device variation, and lower detection limit for
biomolecules. In chapter 6, electronic ELISA was performed using In
2
O
3
nanoribbon
biosensors to detect 3 model cardiac biomarkers. The detection of troponin, CK-MB, and
BNP were demonstrated at clinically relevant concentrations with responses
corresponding to quantitative concentration levels. Finally, using electronic ELISA, the
detection of BNP in whole blood was demonstrated on the In
2
O
3
nanoribbon sensor, with
a 3% error in interpreting the concentration.
6.2 Future directions for aligned epitaxial nanowires
In chapter 2, aligned nanowires have demonstrated the ability to form ordered, bottom-up
assembly that allows high quality SnO
2
nanowires to be patterned at controlled locations.
Although this is a significant step toward large-scale fabrication of high-quality nanowire
devices and sensors, the nanowires must be transfered from the sapphire wafers to a less
costly substrate such as silicon, or to more transparent and flexible substrates such as
111
polyethylene terephthalate (PET) or polydimethylsiloxane (PDMS). To this end, transfer
of aligned epitaxial SnO
2
nanowires will be an important direction for electronic device
studies and applications. Epitaxially aligned SnO
2
nanowires form covalent bonds with
the underneath substrates, and removing the SnO
2
wires from the sapphire substrate by
mechanical transfer has so far not been successful. Below are two methods to consider
that can break the chemical bond between the sapphire and the wires for aligned transfer.
6.2.1 Transfer by laser lift-off
One possible transfer method is the laser lift-off method proposed by Wong et al. for
GaN thin films.
1, 2
In this process, as shown in Figure 6.1, GaN film was originally
epitaxially grown on sapphire substrate. To transfer the film onto a silicon wafer, a layer
of adhesive is first applied to the silicon (a). Next, the GaN film is attached to the silicon
by the adhesive while still on the original sapphire substrate (b). A pulsed KrF laser is
then used to decompose the GaN into Ga metal and N
2
gas at the sapphire interface (c).
Because GaN has a short optical absorption length, the pulsed ultraviolet-laser irradiation
can be confined to the interface. A second annealing at 40°C is then used to melt the Ga-
rich surface in order to separate the film from the sapphire substrate (d). Finally, the
adhesive on the silicon can be dissolved in solvent to free the film (e). If SnO
2
can be
decomposed by irradiation in a process similar to GaN decomposition, then it is possible
to transfer aligned SnO
2
nanowires using this method. Further research will focus on the
laser wavelength that can be used to confine the decomposition of SnO
2
to the sapphire
112
interface. Localizing the annealing to a very thin layer may be especially important
because the diameters of the nanowires are around 00nm, where as GaN film was 3 μm.
Figure 6.1 Transfer of GaN membrane. (a) Starting materials are GaN membrane on sapphire and adhesive
on Si. (b) GaN is attached to Si by the adhesive. (c) Laser heats the interface of GaN and sapphire,
decomposing GaN into Ga and N2 gas. (d) Annealing melts the Ga and removes the saspphire. (e) adhesive
is etched away, leaving free standing GaN film.
6.2.2 Transfer from quartz substrate
Different planes of quartz substrate have been reported to guide aligned CNT growth
through van der Waals interactions.
3
Because quartz is more easily etched chemically
than sapphire without damaging the nanowires, quartz substrates can be used instead of
sapphire to guide SnO
2
naowire alignment. Synthesized nanowire can then be transferred
to different substrates after etching away the quartz substrate. ST cut quartz substrate is
initially explored for epitaxial alignment of SnO
2
nanowires, and the preliminary result is
shown in Figure 6.2. In the SEM image, SnO
2
nanowires are seen to spread outward from
the catalyst strip on the quartz substrate. Although the alignment is weaker than
nanowires grown on sapphire, the nanowires appear to grow both perpendicular and
parallel to the catalyst strip. Further research can experiment on the strength of the
(e) (d) (c) (b) (a)
113
alignment on different cuts of quartz. For example, GaN nanowires have been reported to
align on X and Y cut quartz. The quartz was then selectively etched away to transfer GaN
nanowires onto silicon. Devices with thus transferred GaN nanowires have shown
transfer behavior.
4
Figure 6.2 SnO
2
nanowires grown on ST cut quartz.
6.3 Future directions for In
2
O
3
nanoribbon biosensor
For biosensors based on In
2
O
3
nanoribbons, the platform is currently robust and ready for
more clinically relevant functionalities. One goal of the nanoribbon sensor platform is a
compact point-of-care (POC) device that can be used by patients and emergency personal
outside of the central laboratory facilities. Toward that goal, focus will be on serveral
functionalities that is expected to facilitate the interface of the sensor platform for
clinicians in a medical lab environment. A second goal of the nanoribbon sensor is be
used as a tool for biochemical reactions. Due to the ultra sensitivity of the device, it can
be beneficial for studying the electrochemical processes happening between molecules.
114
6.3.1 On-chip liquid gate
For the sensing platform to become an independent unit, the Ag/AgCl liquid gate must be
replaced with an on-chip liquid gate. The fabrication of the on-chip gate can be
compatible with the 2-mask photolithography process. The idea is that one of the metal
electrodes can be disconnected from a nanoribbon device and act as the liquid gate when
voltage is applied. One version of the design is shown in Figure 6.3 as an example.
Figure 6.3a shows an image of a 3” waf er containing the design of 5 whole nanoribbon
chips. In this particular design, each chip (Figure 6.3b) contains 8 rows of nanoribbon
devices, as shown by the outlined box. The enlarged image in Figure 6.3c shows source
and drain electrodes, represented by the blue lines running into the circle, contact the
nanoribbon to form a FET. The on-chip electrode, on the other hand, is a single electrode
placed in close proximity to the device (circled).
Figure 6.3 On-chip liquid gate. (a) Example design of In
2
O
3
nanoribbon chips fabricated on 3 inch wafer,
with electrodes that can accommodate 24 devices; (b) Expanded optical image of one nanoribbon chip
design. (c) Expanded view of the layout that contains an on-chip solution gate before the sensor.
a
b
c
Nanoribbon sensor
On-chip solution gate
115
6.3.2 Microfluidic integration
Integration of nanobiosensors with microfluidics can help to make the sensors practical
for clinical usage. Microfluidics can minimizes patient blood sample volume to a few µL,
such as the amount obtained from a simple finger prick. The microfluidic chip distributes
a very small but sufficient amount of the sample to each sensor. The channels confine the
distributed sample volume close enough to each sensor for the entire distributed volume
to interface with the sensor surface. The Polydimethylsiloxane (PDMS) material of the
microfluidic ship also prevents evaporation of the sample so that such small amounts are
sufficient. Furthermore, the sensors are designed with high density to enable multiple
disease detection on a compact device. This demands that biomarker receptor solutions,
or the reagents, be delivered in very narrow passages to ensure that it is accurately
delivered to the right sensor.
The microfluidic chip fabrication is already an ongoing part of this project. The initial
test design and integration with the sensor chip is shown below. Figure 6.4a shows a
possible design for the microfluidic chip, with 4 simple horizontal channels for fluid flow.
When integrated with the sensor chip (Figure 6.4b), each channel can be used for one
type of biomarker sensing so that 4 biomarkers can be detected simultaneously. This can
reduce sample size and also turnaround time when detection of multiple biomarkers is
needed. In Figure 6.4c, fluorescent images show the precise delivery of streptavidin
tagged with different colored dyes. Different nanoribbon and microfluidic chips can be
designed to accommodate the biomarker panel as needed. In the final version, the
116
location and the number of the on-chip liquid gates may need to be optimized for sensor
biasing.
Figure 6.4 Example microfluidic design. (a) A simple microfluidic chip design can carry samples to the
nanoribbon sensors through the micro-channels. (b)Integrated platform combines the microfluidic chip with
the nanosensor array, allowing (c) Streptavidin with different colored dye to flow independently through
each channel.
Integrated Platform
Blue streptavidin
Red streptavidin
Negative control
Microfluidc PDMS chip design
a c b
117
6.4 Chapter References
1. Wong, W.S.; Sands, T.; Cheung, N.W. Damage-free separation of GaN thin films
from sapphire substrates. Appl Phys Lett 1998, 72 (5), 599-601.
2. Wong, W.S.; Sands, T.; Cheung, N.W., et al. Fabrication of thin-film InGaN
light-emitting diode membranes by laser lift-off. Appl Phys Lett 1999, 75 (10),
1360-1362.
3. Xiao, J.L.; Dunham, S.; Liu, P., et al. Alignment Controlled Growth of Single-
Walled Carbon Nanotubes on Quartz Substrates. Nano Lett 2009, 9 (12), 4311-
4319.
4. Goren-Ruck, L.; Tsivion, D.; Schvartzman, M., et al. Guided Growth of
Horizontal GaN Nanowires on Quartz and Their Transfer to Other Substrates. Acs
Nano 2014, 8 (3), 2838-2847.
118
Bibliography
1. B-type Natriuretic Peptide (BNP) Blood Test.
http://my.clevelandclinic.org/services/heart/diagnostics-testing/laboratory-tests/b-
type-natriuretic-peptide-bnp-bloodtest (January 2, 2015),
2. Ago, H.; Nakamura, K.; Ikeda, K., et al. Aligned growth of isolated single-walled
carbon nanotubes programmed by atomic arrangement of substrate surface. Chem
Phys Lett 2005, 408 (4-6), 433-438.
3. Ahn, M.W.; Park, K.S.; Heo, J.H., et al. On-chip fabrication of ZnO-nanowire gas
sensor with high gas sensitivity. Sensor Actuat B-Chem 2009, 138 (1), 168-173.
4. Amin, N.; Isaka, T.; Yamada, A., et al. Highly efficient μm thick CdTe
solar cells with textured TCOs. Solar Energy Materials and Solar Cells 2001, 67
(1–4), 195-201.
5. Barsan, N.; Weimar, U. Conduction model of metal oxide gas sensors. J
Electroceram 2001, 7 (3), 143-167.
6. Bassan, R.; Potsch, A.; Maisel, A., et al. B-type natriuretic peptide: a novel early
blood marker of acute myocardial infarction in patients with chest pain and no ST
segment elevation. Eur Heart J 2005, 26 (3), 234-240.
7. Bast, R.C., Jr.; Xu, F.J.; Yu, Y.H., et al. CA 125: the past and the future. The
International journal of biological markers 1998, 13 (4), 179-87.
8. Bellingham, J.R.; Phillips, W.A.; Adkins, C.J. Electrical and Optical-Properties of
Amorphous Indium Oxide. J Phys-Condens Mat 1990, 2 (28), 6207-6221.
9. Braunwald, E.; Antman, E.M.; Beasley, J.W., et al. ACC/AHA guidelines for the
management of patients with unstable angina and non-ST-segment elevation
myocardial infarction: executive summary and recommendations. A report of the
American College of Cardiology/American Heart Association task force on
practice guidelines (committee on the management of patients with unstable
angina). Circulation 2000, 102 (10), 1193-209.
10. Brush, J.E.; Brand, D.A.; Acampora, D., et al. Use of the Initial
Electrocardiogram to Predict in-Hospital Complications of Acute Myocardial-
Infarction. New Engl J Med 1985, 312 (18), 1137-1141.
119
11. Bunimovich, Y.L.; Shin, Y.S.; Yeo, W.S., et al. Quantitative real-time
measurements of DNA hybridization with alkylated nonoxidized silicon
nanowires in electrolyte solution. Journal of the American Chemical Society
2006, 128 (50), 16323-16331.
12. Butler, J.E. Enzyme-linked immunosorbent assay (Reprinted from
Immunochemistry, pg 759-803, 1994). J Immunoassay 2000, 21 (2-3), 165-209.
13. Camagni, P.; Faglia, G.; Galinetto, P., et al. Photosensitivity activation of SnO2
thin film gas sensors at room temperature. Sensor Actuat B-Chem 1996, 31 (1-2),
99-103.
14. Cao, B.B.; Chen, J.J.; Tang, X.J., et al. Growth of monoclinic WO3 nanowire
array for highly sensitive NO2 detection. J Mater Chem 2009, 19 (16), 2323-
2327.
15. Carey, P.G.; Smith, P.M.; Theiss, S.D., et al. Polysilicon thin film transistors
fabricated on low temperature plastic substrates. J Vac Sci Technol A 1999, 17
(4), 1946-1949.
16. Carpenter, M.A.; Mathur, S.; Kolmakov, A., Metal oxide nanomaterials for
chemical sensors. Springer: New York, 2013; p xiii-xx.
17. Chang, H.K.; Ishikawa, F.N.; Zhang, R., et al. Rapid, label-free, electrical whole
blood bioassay based on nanobiosensor systems. ACS nano 2011, 5 (12), 9883-91.
18. Chang, H.K.; Ishikawa, F.N.; Zhang, R., et al. Rapid, Label-Free, Electrical
Whole Blood Bioassay Based on Nanobiosensor Systems. ACS nano 2011, 5 (12),
9883-9891.
19. Chang, H.-K.; Ishikawa, F.N.; Zhang, R., et al. Rapid, Label-Free, Electrical
Whole Blood Bioassay Based on Nanobiosensor Systems. ACS nano 2011, 5 (12),
9883-9891.
20. Chang, H.-K.; Wang, X.; Aroonyadet, N., et al. Top-down Fabricated Polysilicon
Nanoribbon Biosensor Chips for Cancer Diagnosis. MRS Online Proceedings
Library 2013, 1569, null-null.
21. Chen, J.R. Characteristics of fire and explosion in semiconductor fabrication
processes. Process Safety Progress 2002, 21 (1), 19-25.
120
22. Chen, P.C.; Shen, G.Z.; Chen, H.T., et al. High-Performance Single-Crystalline
Arsenic-Doped Indium Oxide Nanowires for Transparent Thin-Film Transistors
and Active Matrix Organic Light-Emitting Diode Displays. ACS nano 2009, 3
(11), 3383-3390.
23. Chen, P.C.; Sukcharoenchoke, S.; Ryu, K., et al. 2,4,6-Trinitrotoluene (TNT)
Chemical Sensing Based on Aligned Single-Walled Carbon Nanotubes and ZnO
Nanowires. Adv Mater 2010, 22 (17), 1900-+.
24. Choi, Y.J.; Hwang, I.S.; Park, J.G., et al. Novel fabrication of an SnO(2)
nanowire gas sensor with high sensitivity. Nanotechnology 2008, 19 (9).
25. Christenson, R.H.A., H., Biomarkers of Myocardial Necrosis. In Cardiovascular
biomarkers: pathophysiology and disease management, Morrow, D. A., Ed.
Humana Press Inc.: Totowa, New Jersy, 2006.
26. Comini, E.; Faglia, G.; Sberveglieri, G. UV light activation of tin oxide thin films
for NO2 sensing at low temperatures. Sensor Actuat B-Chem 2001, 78 (1-3), 73-
77.
27. Comini, E.; Faglia, G.; Sberveglieri, G., et al. Stable and highly sensitive gas
sensors based on semiconducting oxide nanobelts. Appl Phys Lett 2002, 81 (10),
1869-1871.
28. Crowther, J.R., ELISA: Theory and Practice. Humana Press Inc.: Totowa, New
Jersey 1995; Vol. 42.
29. Cui, Y.; Lauhon, L.J.; Gudiksen, M.S., et al. Diameter-controlled synthesis of
single-crystal silicon nanowires. Appl Phys Lett 2001, 78 (15), 2214-2216.
30. Cui, Y.; Wei, Q.; Park, H., et al. Nanowire nanosensors for highly sensitive and
selective detection of biological and chemical species. Science 2001, 293 (5533),
1289-92.
31. Curreli, M.; Rui, Z.; Ishikawa, F.N., et al. Real-Time, Label-Free Detection of
Biological Entities Using Nanowire-Based FETs. Nanotechnology, IEEE
Transactions on 2008, 7 (6), 651-667.
32. Curreli, M.; Zhang, R.; Ishikawa, F.N., et al. Real-Time, Label-Free Detection of
Biological Entities Using Nanowire-Based FETs. Ieee T Nanotechnol 2008, 7 (6),
651-667.
121
33. Dattoli, E.N.; Wan, Q.; Guo, W., et al. Fully transparent thin-film transistor
devices based on SnO2 nanowires. Nano Lett 2007, 7 (8), 2463-9.
34. Dominguez, J.E.; Pan, X.Q.; Fu, L., et al. Epitaxial SnO[sub 2] thin films grown
on (1-bar 012) sapphire by femtosecond pulsed laser deposition. Journal of
Applied Physics 2002, 91 (3), 1060-1065.
35. Elfström, N.; Karlström, A.E.; Linnros, J. Silicon Nanoribbons for Electrical
Detection of Biomolecules. Nano Lett 2008, 8 (3), 945-949.
36. Forrest, S.R.; Burrows, P.E.; Thompson, M.E. Organic Emitters Promise a New-
Generation of Displays. Laser Focus World 1995, 31 (2), 99-&.
37. Franke, M.E.; Koplin, T.J.; Simon, U. Metal and metal oxide nanoparticles in
chemiresistors: does the nanoscale matter? Small 2006, 2 (1), 36-50.
38. Freitag, M.; Martin, Y.; Misewich, J.A., et al. Photoconductivity of single carbon
nanotubes. Nano Lett 2003, 3 (8), 1067-1071.
39. Fryberger, T.B.; Semancik, S. Conductance Response of Pd/Sno2 (110) Model
Gas Sensors to H2 and O2. Sensor Actuat B-Chem 1990, 2 (4), 305-309.
40. Gao, A.; Lu, N.; Dai, P., et al. Silicon-nanowire-based CMOS-compatible field-
effect transistor nanosensors for ultrasensitive electrical detection of nucleic acids.
Nano Lett 2011, 11 (9), 3974-8.
41. Gibler, W.B.; Young, G.P.; Hedges, J.R., et al. Acute myocardial infarction in
chest pain patients with nondiagnostic ECGs: serial CK-MB sampling in the
emergency department. The Emergency Medicine Cardiac Research Group.
Annals of emergency medicine 1992, 21 (5), 504-12.
42. Goren-Ruck, L.; Tsivion, D.; Schvartzman, M., et al. Guided Growth of
Horizontal GaN Nanowires on Quartz and Their Transfer to Other Substrates.
ACS nano 2014, 8 (3), 2838-2847.
43. Gurlo, A., Insights into the Mechanism of Gas Sensor Operation. In Metal oxide
nanomaterials for chemical sensors, Carpenter, M. A.; Mathur, S.; Kolmakov, A.,
Eds. Springer: New York, 2013; pp 3-34.
44. Hahm, J.; Lieber, C.M. Direct ultrasensitive electrical detection of DNA and
DNA sequence variations using nanowire nanosensors. Nano Lett 2004, 4 (1), 51-
54.
122
45. Hakim, M.M.A.; Lombardini, M.; Sun, K., et al. Thin Film Polycrystalline
Silicon Nanowire Biosensors. Nano Lett 2012, 12 (4), 1868-1872.
46. Han, J.H.G., W.B., Biomarkers in the Emergency Department: Rapid Diagnosis
and Triage. In Cardiovascular Biomarkers: pathophysiology and disease
management, Morrow, D. A., Ed. Humana Press: 2006.
47. Han, S.; Jin, W.; Zhang, D.H., et al. Photoconduction studies on GaN nanowire
transistors under UV and polarized UV illumination. Chem Phys Lett 2004, 389
(1-3), 176-180.
48. Heiland, G.; Mollwo, E.; Stockmann, F. Electronic Processes in Zinc Oxide. Solid
State Phys 1959, 8, 191-323.
49. Heller, I.; Janssens, A.M.; Mannik, J., et al. Identifying the mechanism of
biosensing with carbon nanotube transistors. Nano Lett 2008, 8 (2), 591-595.
50. Huang, Y.; Duan, X.; Wei, Q., et al. Directed assembly of one-dimensional
nanostructures into functional networks. Science 2001, 291 (5504), 630-3.
51. Ishikawa, F.N. Applications of one-dimensional structured nanomaterials as
biosensors and transparent electronics. University of Southern California, Los
Angeles, CA, 2009.
52. Ishikawa, F.N.; Chang, H.K.; Curreli, M., et al. Label-free, electrical detection of
the SARS virus N-protein with nanowire biosensors utilizing antibody mimics as
capture probes. ACS nano 2009, 3 (5), 1219-24.
53. Ishikawa, F.N.; Chang, H.-K.; Curreli, M., et al. Label-Free, Electrical Detection
of the SARS Virus N-Protein with Nanowire Biosensors Utilizing Antibody
Mimics as Capture Probes. ACS nano 2009, 3 (5), 1219-1224.
54. Ishikawa, F.N.; Curreli, M.; Chang, H.K., et al. A Calibration Method for
Nanowire Biosensors to Suppress Device-to-Device Variation. ACS nano 2009, 3
(12), 3969-3976.
55. Jin, S.; Whang, D.; McAlpine, M.C., et al. Scalable Interconnection and
Integration of Nanowire Devices without Registration. Nano Lett 2004, 4 (5),
915-919.
123
56. Ju, S.; Facchetti, A.; Xuan, Y., et al. Fabrication of fully transparent nanowire
transistors for transparent and flexible electronics. Nature nanotechnology 2007, 2
(6), 378-84.
57. Keller, T.; Zeller, T.; Ojeda, F., et al. Serial Changes in Highly Sensitive
Troponin I Assay and Early Diagnosis of Myocardial Infarction. Jama-J Am Med
Assoc 2011, 306 (24), 2684-2693.
58. Kilic, C.; Zunger, A. Origins of coexistence of conductivity and transparency in
SnO(2). Physical review letters 2002, 88 (9), 095501.
59. Kim, F.; Kwan, S.; Akana, J., et al. Langmuir-Blodgett nanorod assembly.
Journal of the American Chemical Society 2001, 123 (18), 4360-1.
60. Kim, I.D.; Rothschild, A.; Lee, B.H., et al. Ultrasensitive chemiresistors based on
electrospun TiO2 nanofibers. Nano Lett 2006, 6 (9), 2009-2013.
61. Kim, Y.S.; Hwang, I.S.; Kim, S.J., et al. CuO nanowire gas sensors for air quality
control in automotive cabin. Sensor Actuat B-Chem 2008, 135 (1), 298-303.
62. Kolmakov, A.; Klenov, D.O.; Lilach, Y., et al. Enhanced gas sensing by
individual SnO2 nanowires and nanobelts functionalized with Pd catalyst
particles. Nano Lett 2005, 5 (4), 667-673.
63. Kolmakov, A.; Zhang, Y.; Cheng, G., et al. Detection of CO and O2 Using Tin
Oxide Nanowire Sensors. Adv Mater 2003, 15 (12), 997-1000.
64. Kolmakov, A.; Zhang, Y.X.; Cheng, G.S., et al. Detection of CO and O-2 using
tin oxide nanowire sensors. Adv Mater 2003, 15 (12), 997-+.
65. Kordos, P.; Gregusova, D.; Stoklas, R., et al. Improved transport properties of
Al2O3/AlGaN/GaN metal-oxide-semiconductor heterostructure field-effect
transistor. Appl Phys Lett 2007, 90 (12).
66. Law, M.; Kind, H.; Messer, B., et al. Photochemical sensing of NO2 with SnO2
nanoribbon nanosensors at room temperature. Angew Chem Int Edit 2002, 41
(13), 2405-2408.
67. Lee, T.H.; Rouan, G.W.; Weisberg, M.C., et al. Sensitivity of routine clinical
criteria for diagnosing myocardial infarction within 24 hours of hospitalization.
Annals of internal medicine 1987, 106 (2), 181-6.
124
68. Lee-Lewandrowski, E.; Benzer, T.; Corboy, D., et al. Cardiac Marker Testing As
Part Of An Emergency Department Point ‐of ‐Care Satellite Laboratory In A Large
Academic Medical Center: Practical Issues Concerning Implementation. Point of
Care 2002, 1 (3), 145-154 10.1097/01.POC.0000023109.92641.C7.
69. Lee-Lewandrowski, E.; Januzzi, J.L., Jr.; Grisson, R., et al. Evaluation of first-
draw whole blood, point-of-care cardiac markers in the context of the universal
definition of myocardial infarction: a comparison of a multimarker panel to
troponin alone and to testing in the central laboratory. Archives of pathology &
laboratory medicine 2011, 135 (4), 459-63.
70. Lei, B.; Li, C.; Zhang, D., et al. Nanowire transistors with ferroelectric gate
dielectrics: Enhanced performance and memory effects. Appl Phys Lett 2004, 84
(22), 4553-4555.
71. Leonardy, A.; Hung, W.-Z.; Tsai, D.-S., et al. Structural Features of SnO2
Nanowires and Raman Spectroscopy Analysis. Crystal Growth & Design 2009, 9
(9), 3958-3963.
72. Lewandrowski, K.; Chen, A.; Januzzi, J. Cardiac markers for myocardial
infarction. A brief review. American journal of clinical pathology 2002, 118
Suppl, S93-9.
73. Li, C.; Curreli, M.; Lin, H., et al. Complementary detection of prostate-specific
antigen using In2O3 nanowires and carbon nanotubes. Journal of the American
Chemical Society 2005, 127 (36), 12484-5.
74. Li, C.; Curreli, M.; Lin, H., et al. Complementary detection of prostate-specific
antigen using ln(2)O(3) nanowires and carbon nanotubes. Journal of the
American Chemical Society 2005, 127 (36), 12484-12485.
75. Li, C.; Lei, B.; Zhang, D., et al. Chemical gating of In2O3 nanowires by organic
and biomolecules. Appl Phys Lett 2003, 83 (19), 4014-4016.
76. Li, C.; Zhang, D.; Han, S., et al. Synthesis, electronic properties, and applications
of indium oxide nanowires. Ann Ny Acad Sci 2003, 1006, 104-121.
77. Lin, C.H.; Hsiao, C.Y.; Hung, C.H., et al. Ultrasensitive detection of dopamine
using a polysilicon nanowire field-effect transistor. Chem Commun (Camb) 2008,
(44), 5749-51.
125
78. Liu, Z.Q.; Zhang, D.H.; Han, S., et al. Laser ablation synthesis and electron
transport studies of tin oxide nanowires. Adv Mater 2003, 15 (20), 1754-+.
79. Lyons, J.L.; Janotti, A.; Van de Walle, C.G., Theory and Modeling of Oxide
Semiconductors. In Oxide Semiconductors, Svensson, B. G.; Pearton, S. J.;
Jagadish, C., Eds. Elsevier Academic Press Inc: San Diego, CA, 2013; Vol. 88, pp
1-38.
80. Maiti, A.; Rodriguez, J.A.; Law, M., et al. SnO2 nanoribbons as NO2 sensors:
Insights from first principles calculations. Nano Lett 2003, 3 (8), 1025-1028.
81. Martel, R.; Schmidt, T.; Shea, H.R., et al. Single- and multi-wall carbon nanotube
field-effect transistors. Appl Phys Lett 1998, 73 (17), 2447-2449.
82. Mattila, M.; Hakkarainen, T.; Jiang, H., et al. Effect of substrate orientation on
the catalyst-free growth of InP nanowires. Nanotechnology 2007, 18 (15).
83. Melosh, N.A.; Boukai, A.; Diana, F., et al. Ultrahigh-density nanowire lattices
and circuits. Science 2003, 300 (5616), 112-5.
84. Nair, P.R.; Alam, M.A. Screening-limited response of nanobiosensors. Nano Lett
2008, 8 (5), 1281-1285.
85. Newby, L.K.; Christenson, R.H.; Ohman, E.M., et al. Value of serial troponin T
measures for early and late risk stratification in patients with acute coronary
syndromes. Circulation 1998, 98 (18), 1853-1859.
86. Ng, S.M.; Krishnaswamy, P.; Morissey, R., et al. Ninety-minute accelerated
critical pathway for chest pain evaluation. Am J Cardiol 2001, 88 (6), 611-617.
87. Nikoobakht, B.; Michaels, C.A.; Stranick, S.J., et al. Horizontal growth and in
situ assembly of oriented zinc oxide nanowires. Appl Phys Lett 2004, 85 (15),
3244-3246.
88. Niloff, J.M.; Bast, R.C., Jr.; Schaetzl, E.M., et al. Predictive value of CA 125
antigen levels in second-look procedures for ovarian cancer. American Journal of
Obstetrics & Gynecology 151 (7), 981-986.
89. Nourjah, P. National Hospital Ambulatory Medical Care Survey: 1997 emergency
department summary. Advance data 1999, (304), 1-24.
126
90. Ohman, E.M.; Armstrong, P.W.; Christenson, R.H., et al. Cardiac troponin T
levels for risk stratification in acute myocardial ischemia. New Engl J Med 1996,
335 (18), 1333-1341.
91. Park, I.; Li, Z.; Pisano, A.P., et al. Top-down fabricated silicon nanowire sensors
for real-time chemical detection. Nanotechnology 2010, 21 (1), 015501.
92. Pati, S.; Majumder, S.B.; Banerji, P. Role of oxygen vacancy in optical and gas
sensing characteristics of ZnO thin films. J Alloy Compd 2012, 541, 376-379.
93. Patolsky, F.; Zheng, G.F.; Hayden, O., et al. Electrical detection of single viruses.
P Natl Acad Sci USA 2004, 101 (39), 14017-14022.
94. Pengfei, Q.F.; Vermesh, O.; Grecu, M., et al. Toward large arrays of multiplex
functionalized carbon nanotube sensors for highly sensitive and selective
molecular detection. Nano Lett 2003, 3 (3), 347-351.
95. Po-Chiang, C.; Guozhen, S.; Chongwu, Z. Chemical Sensors and Electronic
Noses Based on 1-D Metal Oxide Nanostructures. Nanotechnology, IEEE
Transactions on 2008, 7 (6), 668-682.
96. Rout, C.S.; Ganesh, K.; Govindaraj, A., et al. Sensors for the nitrogen oxides,
NO2, NO and N2O, based on In2O3 and WO3 nanowires. Appl Phys a-Mater
2006, 85 (3), 241-246.
97. Rude, R.E.; Poole, W.K.; Muller, J.E., et al. Electrocardiographic and Clinical-
Criteria for Recognition of Acute Myocardial-Infarction Based on Analysis of
3,697 Patients. Am J Cardiol 1983, 52 (8), 936-941.
98. Ryu, K.; Badmaev, A.; Wang, C., et al. CMOS-Analogous Wafer-Scale
Nanotube-on-Insulator Approach for Submicrometer Devices and Integrated
Circuits Using Aligned Nanotubes. Nano Lett 2009, 9 (1), 189-197.
99. Schierbaum, K.D.; Weimar, U.; Gopel, W., et al. Conductance, Work Function
and Catalytic Activity of Sno2-Based Gas Sensors. Sensor Actuat B-Chem 1991,
3 (3), 205-214.
100. Seiyama, T.; Kato, A.; Fujiishi, K., et al. A New Detector for Gaseous
Components Using Semiconductive Thin Films. Anal Chem 1962, 34 (11), 1502-
1503.
127
101. Semancik, S.; Cavicchi, R.E. The growth of thin, epitaxial SnO2 films for gas
sensing applications. Thin Solid Films 1991, 206 (1–2), 81-87.
102. Sheats, J.R.; Antoniadis, H.; Hueschen, M., et al. Organic electroluminescent
devices. Science 1996, 273 (5277), 884-888.
103. SMITH, H.M. THE NERNST LAMP. Science 1898, 8 (203), 689-690.
104. Smith, P.A.; Nordquist, C.D.; Jackson, T.N., et al. Electric-field assisted assembly
and alignment of metallic nanowires. Appl Phys Lett 2000, 77 (9), 1399-1401.
105. Stern, E.; Klemic, J.F.; Routenberg, D.A., et al. Label-free immunodetection with
CMOS-compatible semiconducting nanowires. Nature 2007, 445 (7127), 519-22.
106. Stern, E.; Vacic, A.; Li, C., et al. A nanoelectronic enzyme-linked immunosorbent
assay for detection of proteins in physiological solutions. Small 2010, 6 (2), 232-
8.
107. Stern, E.; Vacic, A.; Rajan, N.K., et al. Label-free biomarker detection from
whole blood. Nature nanotechnology 2010, 5 (2), 138-142.
108. Stern, E.; Wagner, R.; Sigworth, F.J., et al. Importance of the debye screening
length on nanowire field effect transistor sensors. Nano Lett 2007, 7 (11), 3405-
3409.
109. Svensson, B.G.; Pearton, S.J.; Jagadish, C., et al., Preface. In Oxide
Semiconductors, Elsevier Academic Press Inc: San Diego, CA, 2013; Vol. 88, pp
Xi-Xiii.
110. Tang, X.; Bansaruntip, S.; Nakayama, N., et al. Carbon nanotube DNA sensor and
sensing mechanism. Nano Lett 2006, 6 (8), 1632-6.
111. Tsivion, D.; Schvartzman, M.; Popovitz-Biro, R., et al. Guided growth of
horizontal ZnO nanowires with controlled orientations on flat and faceted
sapphire surfaces. ACS nano 2012, 6 (7), 6433-45.
112. Tsivion, D.; Schvartzman, M.; Popovitz-Biro, R., et al. Guided growth of
millimeter-long horizontal nanowires with controlled orientations. Science 2011,
333 (6045), 1003-7.
113. Vacic, A.; Criscione, J.M.; Stern, E., et al. Multiplexed SOI BioFETs. Biosensors
& bioelectronics 2011, 28 (1), 239-42.
128
114. Visintin, I.; Feng, Z.; Longton, G., et al. Diagnostic markers for early detection of
ovarian cancer. Clin Cancer Res 2008, 14 (4), 1065-1072.
115. Wang, D.; Sheriff, B.A.; Heath, J.R. Silicon p-FETs from ultrahigh density
nanowire arrays. Nano Lett 2006, 6 (6), 1096-100.
116. Wang, X.; Aroonyadet, N.; Zhang, Y., et al. Aligned Epitaxial SnO Nanowires on
Sapphire: Growth and Device Applications. Nano Lett 2014.
117. Wetchakun, K.; Samerjai, T.; Tamaekong, N., et al. Semiconducting metal oxides
as sensors for environmentally hazardous gases. Sensor Actuat B-Chem 2011, 160
(1), 580-591.
118. Whang, D.; Jin, S.; Wu, Y., et al. Large-Scale Hierarchical Organization of
Nanowire Arrays for Integrated Nanosystems. Nano Lett 2003, 3 (9), 1255-1259.
119. Wong, W.S.; Sands, T.; Cheung, N.W. Damage-free separation of GaN thin films
from sapphire substrates. Appl Phys Lett 1998, 72 (5), 599-601.
120. Wong, W.S.; Sands, T.; Cheung, N.W., et al. Fabrication of thin-film InGaN
light-emitting diode membranes by laser lift-off. Appl Phys Lett 1999, 75 (10),
1360-1362.
121. Wu, A.H.B.; Apple, F.S.; Gibler, W.B., et al. National Academy of Clinical
Biochemistry standards of laboratory practice: Recommendations for the use of
cardiac markers in coronary artery diseases. Clin Chem 1999, 45 (7), 1104-1121.
122. Xiao, J.L.; Dunham, S.; Liu, P., et al. Alignment Controlled Growth of Single-
Walled Carbon Nanotubes on Quartz Substrates. Nano Lett 2009, 9 (12), 4311-
4319.
123. Yamazoe, N. New Approaches for Improving Semiconductor Gas Sensors. Sensor
Actuat B-Chem 1991, 5 (1-4), 7-19.
124. Yao, J.; Yan, H.; Lieber, C.M. A nanoscale combing technique for the large-scale
assembly of highly aligned nanowires. Nature nanotechnology 2013, 8 (5), 329-
335.
125. Yu, G.; Cao, A.; Lieber, C.M. Large-area blown bubble films of aligned
nanowires and carbon nanotubes. Nature nanotechnology 2007, 2 (6), 372-7.
129
126. Yuan, Z.J.; Zhu, X.M.; Wang, X.O., et al. Annealing effects of In2O3 thin films
on electrical properties and application in thin film transistors. Thin Solid Films
2011, 519 (10), 3254-3258.
127. Zhang, D.; Li, C.; Han, S., et al. Ultraviolet photodetection properties of indium
oxide nanowires. Appl Phys a-Mater 2003, 77 (1), 163-166.
128. Zhang, D.H.; Liu, Z.Q.; Li, C., et al. Detection of NO2 down to ppb levels using
individual and multiple In2O3 nanowire devices. Nano Lett 2004, 4 (10), 1919-
1924.
129. Zheng, G.F.; Patolsky, F.; Cui, Y., et al. Multiplexed electrical detection of cancer
markers with nanowire sensor arrays. Nat Biotechnol 2005, 23 (10), 1294-1301.
130. Zhou, C.W.; Kong, J.; Yenilmez, E., et al. Modulated chemical doping of
individual carbon nanotubes. Science 2000, 290 (5496), 1552-1555.
131. Zhou, W.; Dai, X.C.; Fu, T.M., et al. Long Term Stability of Nanowire
Nanoelectronics in Physiological Environments. Nano Lett 2014, 14 (3), 1614-
1619.
132. Zi, Y.; Jung, K.; Zakharov, D., et al. Understanding Self-Aligned Planar Growth
of InAs Nanowires. Nano Lett 2013.
Abstract (if available)
Abstract
Semiconducting metal oxides (SMO) are widely known to respond strongly to changes in their environment through their surface chemistry. Sensors built from field‐effect transistors (FET) that use nanostructured SMOs as the active channel have the combined advantages of having an ultra sensitive surface with large surface‐to‐volume ratio and being equipped with an electronic read‐out that can be label‐free, fast‐responding, portable, and accessible. However, the obstacle of applying metal oxide nanostructures to FET sensing technology in a scalable and controllable process that’s necessary for practical applications must first be tackled. In this thesis two approaches are demonstrated to overcome this problem. ❧ In chapter one, background information for semiconducting metal oxides, the effect of nanoscale materials on sensitivity, and the mechanism of chemical and biological sensing are presented. ❧ In chapter two, the growth of epitaxially aligned SnO₂ nanowires are demonstrated as a “bottom‐up” approach that can be used to fabricate FETs, photo‐detectors, polarizers, and chemical sensors in a salable process. Nanowire growth study shows good alignment guided by substrate lattice, and the electronic quality of the aligned SnO₂ nanowires is demonstrated through its ultra sensitive 0.2ppb detection level of NO₂. ❧ In chapter three, a “top‐down” process is developed to fabricate In₂O₃ nanoribbon biosensors that are highly uniform and sensitive. The fabrication process requires only 2 conventional photolithography steps that are scalable for different wafer sizes and can use a versatile range of substrates. The sensors are demonstrated to have strong and fast response to pH, and a stable surface chemistry is used to apply the nanoribbon sensors to specific and selective biomarker detection. ❧ In chapter four, polysilicon is investigated as the optimal silicon-based material for developing “top‐down” nanoribbon biosensors. By comparing its fabrication process, pH and biomarker sensitivity with the In₂O₃ nanoribbon, it is concluded that although poly‐silicon nanoribbon fabrication is a scalable way to achieve silicon‐based “top‐down” nanobiosensors, In₂O₃ nanoribbons are more advantageous in terms of sensitivity, uniformity, and versatility. ❧ In chapter five, In₂O₃ nanoribbon biosensors are used to perform an electronic enzyme‐linked immuno assay (ELISA) for the detection of multiple chest pain biomarkers. Quantitative detection of the cardiac biomarkers troponin, creatine kinase—MB, and B-type natriuretic peptide (BNP) were achieved within clinically relevant concentrations. Detection of BNP in whole blood was also achieved with concentration response within 3% of prediction. ❧ Finally, chapter 6 presents a summary of the thesis and future directions.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
One-dimensional nanomaterials for electronic and sensing applications
PDF
Biosensing and biomimetic electronics
PDF
Surface functionalization of nanomaterials and the development of nanobiosensors
PDF
In2O3 COVID-19 biosensors and two-dimensional materials: synthesis and applications
PDF
Surface functionalization of nanomaterials and development of field effect transistor nanobiosensors
PDF
One-dimensional nanomaterials: synthesis and applications
PDF
Biophysical studies of nanosecond pulsed electric field induced cell membrane permeabilization
PDF
Design and characterization of metal and semiconducting nanostructures and nanodevices
PDF
Novel nanomaterials for electronics, optoelectronics and sensing applications
PDF
One-dimensional nanostructures for chemical sensing, transparent electronics, and energy conversion and storage devices
PDF
Carbon nanotube nanoelectronics and macroelectronics
PDF
Zero-dimensional and one-dimensional nanostructured materials for application in photovoltaic cells
PDF
Nanomaterials for macroelectronics and energy storage device
PDF
Growth, characterization of gallium arsenide based nanowires and application in photovoltaic cells
PDF
Modification of electrode materials for lithium ion batteries
PDF
GaAs nanowire optoelectronic and carbon nanotube electronic device applications
PDF
Carbon nanotube macroelectronics
PDF
Nanostructured silicon anode and sulfur cathode for lithium rechargeable batteries
PDF
Printed electronics based on carbon nanotubes and two-dimensional transition metal dichalcogenides
PDF
Substrate-induced effects in thermally cycled graphene & electron and thermoelectric transports across 2D material heterostructures
Asset Metadata
Creator
Wang, Xiaoli
(author)
Core Title
Semiconducting metal oxide nanostructures for scalable sensing applications
School
Viterbi School of Engineering
Degree
Doctor of Philosophy
Degree Program
Electrical Engineering
Publication Date
02/02/2015
Defense Date
01/12/2015
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
aligned nanowires,chemical sensor,In₂O₃ nanoribbon,metal oxide,nanoribbon biosensor,NO₂ nanowire sensor,OAI-PMH Harvest,SnO₂ nanowire
Format
application/pdf
(imt)
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Zhou, Chongwu (
committee chair
), Goo, Edward K. (
committee member
), Gundersen, Martin A. (
committee member
)
Creator Email
xiaoliw@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c3-528117
Unique identifier
UC11298269
Identifier
etd-WangXiaoli-3157.pdf (filename),usctheses-c3-528117 (legacy record id)
Legacy Identifier
etd-WangXiaoli-3157.pdf
Dmrecord
528117
Document Type
Dissertation
Format
application/pdf (imt)
Rights
Wang, Xiaoli
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
Tags
aligned nanowires
chemical sensor
In₂O₃ nanoribbon
metal oxide
nanoribbon biosensor
NO₂ nanowire sensor
SnO₂ nanowire