FABRICATION AND CHARACTERIZATION OF COMPOUND SEMICONDUCTOR

SENSORS FOR PRESSURE, GAS, CHEMICAL, AND BIOMATERIAL SENSING

By

BYOUNG SAM KANG

A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT

OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2005

Copyright 2005

by

Byoung Sam Kang

To our LORD, Jesus Christ

iv

ACKNOWLEDGMENTS

I would like to thank God, our Father, for all of the graces He gave me while I was

studying here in Gainesville. The past years have been among the most crucial years in

my life, because He led me to meet invaluable people I could not see in other places, and

they served as Barnabas changing Gainesville into Gilgal. I am sorry I can not list all of

them here.

First of all, I really appreciate my supervisory committee chair, Dr. Fan Ren for

technical and financial support throughout this study. He generously gave me lots of

heartwarming advice professionally and also personally. He willingly guided me with

better ideas when I had difficulty in solving the problems. I am also indebted to my other

supervisory committee members: Steve Pearton, David Norton, and Jason Weaver. Their

significant contributions to this work are greatly appreciated. During collaboration with

Dr. Pearton’s and Dr. Norton’s groups, I had the opportunity to develop interpersonal

skills, emphasize efforts, and pursue goals. I am honored to have been associated with

such an eminent committee. I would also like to thank Dr. S.N.G. Chu (Multiplex Inc.

South Plainfield, New Jersey) for helping me with stress analysis of the hetero interfaces.

Many thanks go to my colleagues and coworkers in the Department of Chemical

Engineering: Jeff LaRoche, Ben Luo, Suku Kim, Jihyun Kim, Rishabh Mehandru,

Soohwan Jang, Hungta Wang, Traivis Anderson, Jau Jiun. Special thanks go to Kwang

Hyun Baik, Rohit, Sang Yoon Han, and Huck Soo Yang in Materials Science and

Engineering department for their various discussion and helping in dry etch. System

v

maintenance and good friendship of Brent Gila are gratefully acknowledged. I had many

happy discussions with him about work. His optimistic attitude could make him a good

and kind professor someday soon.

I thank my family for their love and support. I especially thank our parents, I really

appreciate the enormous love they have shown me. I hope I can be as good a parent for

my daughter as they have been. I hope this dissertation, in some small way, repays them

for their love which I can never forget. I thank my wife, Jinah, who helps me to

concentrate this work by taking care of all other works including our daughter,

Dahyoung. This dissertation would not easy without their consistent supports. Half of this

work is her achievement.

Finally, I thank valuable pastors, Hee-Young Sohn, and Joong-Soo Lee for guiding

me to learn the immeasurable depth of love in Jesus Christ. I pray that their consistent

love and whole dedication to our God Father will not be changed forever.

vi

TABLE OF CONTENTS page

ACKNOWLEDGMENTS ................................................................................................. iv

LIST OF FIGURES ......................................................................................................... viii

ABSTRACT..................................................................................................................... xiii

CHAPTER

1 INTRODUCTION ........................................................................................................1

2 BACKGROUND AND LITERATURE REVIEW ......................................................4

2.1 Historical Review ...................................................................................................4 2.2 Background.............................................................................................................7

2.2.1 AlGaN/GaN High Electron Mobility Transistors (HEMTs) ........................7 2.2.2 ZnO based Chemical Sensor ......................................................................11

3 PRESSURE SENSOR USING PIEZOELECTRIC POLARIZATION .....................15

3.1 Introduction...........................................................................................................15 3.2 Effect on External Strain on the Conductivity of AlGaN/GaN HEMTs ..............16 3.2 Pressure induced Changes in the Conductivity of AlGaN/GaN HEMTs .............24 3.3 Capacitance Pressure Sensor Based on GaN HEMT on Si Membrane ................31

4 CATALYST BASED GAS SENSOR FOR HYDROCARBON GASES..................38

4.1 Introduction...........................................................................................................38 4.2 AlGaN/GaN based MOS Diode Hydrogen Gas Sensor .......................................39 4.3 Hydrogen Induced Reversible Changes in Drain Current in

Sc2O3/AlGaN/GaN HEMTs...................................................................................45 4.4 Comparison of MOS and Schottky W/Pt-GaN Diodes for Hydrogen Detection .52 4.5 Detection of C2H4 Using Wide Bandgap Semiconductor Sensors .......................59 4.6 Hydrogen and Ozone Gas Sensing Using Multiple ZnO Nanorods .....................68

5 CHEMICAL SENSOR FOR POLYMERS AND POLAR LIQUIDS .......................75

5.1 Introduction...........................................................................................................75 5.2 Gateless AlGaN/GaN HEMT Response to Block Co-Polymers ..........................76

vii

5.3 pH Measurements with Single ZnO Nanorod Integrated with a Microchannel ...82

6 BIOSENSORS FOR BIOMATERIALS AND CELL GROWTH .............................89

6.1 Introduction...........................................................................................................89 6.2 Detection of Halide Ions with AlGaN/GaN HEMTs............................................91 6.3 Electrical Detection of Immobilized Proteins with Ungated AlGaN/GaN

HEMTs...................................................................................................................98 6.4 Use of 370 nm Light for Selective Area Fibroblast Cell Growth.......................103

7 SUMMARY AND FUTURE WORKS....................................................................114

7.1 Pressure Sensor...................................................................................................114 7.2 Gas Sensor ..........................................................................................................115 7.3 Chemical Sensor .................................................................................................117 7.4 Bio Sensor...........................................................................................................118

APPENDIX

A STRAIN CALCULATION OF A CIRCULAR MEMBRANE OF ALGAN/GAN UNDER DIFFERENTIAL PRESSURE...................................................................121

A.1 Strains in the Films ............................................................................................121 A.2 Electrical Polarization and Piezoelectric Effects ...............................................123 A.3 2DEG and Conductance of HEMT....................................................................124

LIST OF REFERENCES.................................................................................................126

BIOGRAPHICAL SKETCH ...........................................................................................138

viii

LIST OF FIGURES

Figure page 1-1 Semilog plot of intrinsic carrier concentration versus inverse temperature for Si,

GaAs, GaN (left). Performance of AlGaN/GaN based power amplifier as compared to GaAs and SiC based devices (right)......................................................2

2-1 Sheet carrier concentration in the 2DEG channel of AlGaN/GaN HEMT induced by the piezoelectric polarization as a function of Al concentration ...........................9

2-2 Piezoelectric (PE) and spontaneous (SP) polarization effects in Ga face or N-face AlGaN/GaN heterostructures. ..........................................................................10

2-3 Crystal structure of wurtzite ZnO.............................................................................12

2-4 Energy bandgap of several, III-V, and II-IV compound semiconductors as a function of lattice constant. ......................................................................................13

3-1 Layer structures of two-terminal non-mesa(top) and mesa(middle) devices, and top view photo-micrograph of fabricated two-terminal devices with different channel lengths(bottom). ..........................................................................................16

3-2 Pressure sensor package: experimental setup to detect I-V characteristics connected to the BNC cable according to various mechanical stresses (top) and mechanical stressor with cantilever (bottom)...........................................................18

3-3 Strain induced by AlGaN on GaN for the un-relaxed and partially relaxed AlGaN layer as a function of Al concentration (top) and sheet carrier concentration induced by the piezoelectric polarization as a function of Al concentration (bottom) .............................................................................................19

3-4 Effect of tensile or compressive stress on the conductivity of the AlGaN/GaN HEMT with mesa etching (top) and without mesa etching (bottom).......................22

3-5 Circular membrane of AlGaN/GaN on a Si substrate fabricated by etching a circular hole in the substrate (left). A deflection of the membrane away from the substrate due to differential pressure on the two sides of the membrane produces a tensile strain in the membrane (right)....................................................................25

3-6 Device structure with a finger patterned device on the HEMT membrane..............27

ix

3-7 SEM micrographs of via through the Si wafer(left)and cross sectional view of a via hole(right) ...........................................................................................................28

3-8 IDS-VDS characteristics at 25C from AlGaN/GaN HEMT membrane as a function of applied pressure. ..................................................................................................28

3-9 Channel conductivity of the AlGaN/GaN HEMT membrane as a function of differential pressure..................................................................................................29

3-10 Device structure (top) and SEM micrograph of AlGaN/GaN circular membrane on a Si substrate fabricated by etching a circular hole in the substrate(bottom)......32

3-11 Top view of HEMT capacitance pressure sensor(top) and capacitance as a function of pressure for different diaphragm radii (bottom). ...................................34

3-12 Capacitance change as a function of radius of the AlGaN/GaN HEMT membrane over the pressure range from -1 to +9.5 bar. ..........................................35

3-13 Capacitance change as a function of radius of the AlGaN/GaN HEMT membrane over the pressure range from -1 to +9.5 bar ...........................................36

4-1 Cross-sectional schematic of completed MOS diode on AlGaN/GaN HEMT layer structure (top) and plan-view photograph of device(bottom). ........................40

4-2 Forward I-V characteristics of MOS-HEMT based diode sensors of two different dimensions at 25°C measured under pure N2 or 10%H2 /90%N2 ambient. .............42

4-3 Time response at 25°C of MOS-HEMT based diode forward current at a fixed bias of 2V when switching the ambient from N2 to 10%H2 /90%N2 for periods of 10, 20 or 30 seconds and then back to pure N2 ........................................................43

4-4 Time response at 25°C of MOS-HEMT based diode forward current at a fixed bias of 2V for three cycles of switching the ambient from N2 to 10%H2 /90%N2 for periods of 10 (top) or 30 (bottom) ......................................................................44

4-5 Photograph of MOS HEMT hydrogen sensor..........................................................47

4-6 IDS-VDS characteristics of MOS-HEMT measured at 25oC under pure N2 ambient or in 10% H2/90%N2 ambient. .................................................................................48

4-7 Change in drain-source current for measurement in N2 versus 10%H2 /90%N2 ambient, as a function of gate voltage (top) and corresponding transconductance at a fixed drain-source voltage of 3V(bottom). ........................................................49

4-8 Time dependence of drain-source current when switching from N2 to 1%H2 /99% N2 ambient and back again. The top shows different injection times of the H2/N2, while the bottom shows the reversibility of the current change. ..................51

x

4-9 The W/Pt Schottky diode (top) and MOS diode (bottom). ......................................53

4-10 Photograph of packaged gas sensor .........................................................................53

4-11 Forward I-V characteristics at 300 oC(top) or 500 oC(bottom) from the Schottky and MOS diodes in pure N2 and 10% H2 /90% N2...................................................54

4-12 Measurement temperature dependence of forward I-V characteristics of the Schottky (top) and MOS (bottom) . .........................................................................55

4-13 Temperature dependence of turn-on voltage(top) and on-state resistance(bottom). ...................................................................................................57

4-14 Change in forward current when measuring in 10% H2 /90% N2 relative to pure N2 at 3 or 3.5 V in both the Schottky and MOS diodes ...........................................58

4-15 Schematic of AlGaN/GaN MOS diode (top) and bulk ZnO Schottky diode structure (bottom).....................................................................................................61

4-16 Forward I-V characteristics of MOS-HEMT based diode sensor at 400°C measured under pure N2 or 10% C2H4 /90% N2 ambients .......................................62

4-17 Change in MOS diode forward current at fixed forward bias of 2.5V(top) or at fixed current(bottom)................................................................................................64

4-18 I-V characteristics at 50°C (top) or 150 °C (bottom) of Pt/ZnO diodes measured in different ambients.................................................................................................65

4-19 Change in current at a fixed bias (top) or change in voltage at fixed current (bottom)....................................................................................................................67

4-20 TEM of ZnO nanorod...............................................................................................69

4-21 SEM image of ZnO multiple nanorods (top) and the pattern contacted by Al/Pt/Au electrodes (bottom) ...................................................................................70

4-22 I-V characteristics at different temperatures of ZnO multiple nanorods measured in either N2 or 10 % H2 in N2 ambient .....................................................................71

4-23 Change in current measured at 0.1 V for measurement in either N2 or 10%H2 in N2 ambients ..............................................................................................................72

4-24 I-V characteristics at 25C of ZnO multiple nanorods measured in either N2 OR 3% O3 in N2 ..............................................................................................................73

4-25 Time dependence of current at 1V bias when switching back and forth from N2 to 3% O3 in N2 ambients...........................................................................................74

5-1 Layout of gate HEMT structure (top) and device cross-section (bottom). ..............82

xi

5-2 Structure of block co-polymer, composed of different portions of PS and PEO (top) and chemical formula for PS and PEO (bottom).............................................83

5-3 Drain I-V characteristics for the air, PS and PEO....................................................84

5-4 Drain I-V characteristics for the different concentration of PS-PEO block copolymer.................................................................................................................85

5-5 Drain IV characteristics of copolymers with different composition. .......................86

5-6 Schematic (top) and scanning electron micrograph (bottom) of ZnO nanorod with integrated microchannel (4µm width)..............................................................89

5-7 I-V characteristics of ZnO nanorod after wire-bonding, measured either with or without UV (365nm) illumination............................................................................90

5-8 Change in current (top) or conductance (bottom) with pH (from 2-12) at V = 0.5V. .........................................................................................................................91

5-9 Relation between pH and conductance of ZnO nanorod either with or without UV (365nm) illumination.........................................................................................92

6-1 Schematic of HEMT structure for detection of halide ions (top) and plan view photomicrograph of completed device using a 20 nm Au film in the gate region (bottom). ...................................................................................................................92

6-2 Time dependence of current change in Au-gated HEMTs upon exposure to NaF (top) or NaCl (bottom) solutions with different concentration.. ..............................93

6-3 Current change as a function of concentration for HEMTs with or without the Au gate .....................................................................................................................95

6-4 Change in current as a function of time as a Au-gated HEMT is exposed to water. In the case of thiol modification of the gate, no response was observed. .....96

6-5 AC measurement of change in current in gateless HEMT exposed to water. The applied bias of 500 mV was modulated at 11 Hz.....................................................97

6-6 Structure of APS (3-Aminopropyl) triethoxysilane (top left), surface functionalization before chemical modification (top right). A schematic diagram of the gateless HEMT whose surface is functionalized by chemical modification in the gate region (bottom). ......................................................................................99

6-7 Fluorescent photographs of biotinylated probes on chemically treated GaN surface (left) and non-biotinated surface (right).....................................................100

xii

6-8 SEM micrographs of amine-functionalized GaN surfaces reacted with biotin NHS and then blocked and exposed to either BSA-coated (left) or streptavidin (right) nanoparticles. ..............................................................................................101

6-9 Change in HEMT drain-source current as a result of interaction between Biotin (Sulfo-NHS-Biotin) and streptavidine introduced to the gateless HEMT surface.102

6-10 Patterns formed on glass slides coated with 50 nm TiO2.. .....................................104

6-11 Optical spectrum of a GaN light emitting diode at an operating current of 4 mA (top). The current-voltage characteristics of the GaN based LED (bottom). .........105

6-12 Effect of UV light on growth of Fibroblast cells on glass slices and TiO2 coated glass slides..............................................................................................................107

6-13 Simulated reflectivity of 370 nm UV light with different substrates .....................108

6-14 Fibroblast cell growth on patterned TiO2 coated glass slide with the UV light illumination (top) and without UV light illumination (bottom). ............................110

6-15 A set-up of fibroblast cell on patterned TiO2 coated glass slides under different UV intensity illumination.......................................................................................111

6-16 Simulated reflectivity of 370 nm UV light with six TiO2/SiO2 dielectric mirror stacks ......................................................................................................................112

xiii

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

FABRICATION AND CHARACTERIZATION OF COMPOUND SEMICONDUCTOR SENSORS FOR PRESSURE, GAS, CHEMICAL, AND BIOMATERIAL SENSING

By

Byoung Sam Kang

December 2005

Chair: Fan Ren Major Department: Chemical Engineering

GaN-based diodes and high electron mobility transistors (HEMTs) for pressure,

gas, liquid and biological sensing were fabricated and characterized. A novel metal oxide,

ZnO was also evaluated as a potential electronic device for chemical sensing

applications. The devices presented herein showed high sensitivity and capability of

operation in harsh environmental conditions such as high temperature and pressure.

The AlGaN/GaN HEMTs grown on (0001) sapphire substrates show polarization-

induced two dimensional electron gas (2DEG) at the AlGaN/GaN hetero-interface.

Linear dependence of the 2DEG channel conductance on external strain was observed

using a cantilever beam in a bending configuration. To overcome the rigidity of sapphire

substrates in applying external stress, a micro pressure sensor using a 150 µm diameter

thin flexible AlGaN/GaN circular membrane with an interdigitated-finger device on a

(111) Si substrate was demonstrated. The measured pressure sensitivity was 7.1 ×10-2

xiv

mS/bar, which was two orders of magnitude larger than that of a cantilever beam pressure

sensor.

Pt gated AlGaN/GaN HEMT-based metal-oxide semiconductor (MOS) diodes and

field effect transistors (FETs) were demonstrated for detecting hydrocarbon gases,

followed by a comparison between MOS and W/Pt Schottky-based GaN diodes for

hydrogen sensing. Changes in current, when sensors exposed to hydrogen on the Pt-gated

AlGaN/GaN HEMT were approximately an order of magnitude larger than that of Pt

/GaN Schottky diodes and 5 times larger than Sc2O3/AlGaN/GaN MOS diodes.

For the liquid sensors, gateless AlGaN/GaN HEMTs showed large changes in

source-drain current on exposing the gate region to polar liquids and block copolymers.

The polar nature of these chemicals leads to a change of surface charge in the gate region

on the HEMT, producing a change in surface potential at the semiconductor/liquid

interface. For biomaterials detection, the gate region was chemically modified with

aminopropyl silane. As streptavidin was introduced to the biotin-functionalized gate

region, the drain-source current showed a clear decrease of 4 µA, which shows

interaction between antibody and antigen.

A Schottky diode was fabricated on a ZnO thin film and showed higher sensitivity

to hydrogen (5 ppm). A single ZnO nanorod FET-based sensor was also demonstrated.

Conductivity of the single nanorod sensor decreased linearly when the pH value of the

solution varied from 2 to 12.The measured sensitivity was 8.5 nS/pH in the dark and 20

nS/pH under UV (365 nm) illumination, showing tremendous potential for sensing

applications.

1

CHAPTER 1 INTRODUCTION

The sensor industry has grown rapidly in recent years. The non-military world

market for sensors exceeded expectations with US $42.2 billion in 2003 and it is

expected to reach US $54 billion by 2008 [Mar04]. This rapid growth of the sensor

market increases the need to develop chemical sensor technology in applications such as

chemical reactor processing, factory automation, industrial monitoring, automotive

industry, computers, robotics, and telecommunications [Wil05].

In practice, the sensing elements must be relatively small in size, robust, and should

not require a large sensing sample volume [Hun01, Liu04]. Due to the development of

fabrication and processing techniques, a Si-based chemical sensing microsystem has the

advantage of producing microsize structures in a highly uniform and geometrically well

defined manner [Mad97]. While Si has proven to be the primary contestant in the micro-

sensor market, there is an ever-growing need for devices operating at conditions beyond

the limits of silicon. Silicon based micro-sensors can not be operated in harsh

environments such as in high temperature, pressure, and chemically corrosive ambients.

Wide bandgap electronics and sensors based on GaN can be operated at elevated

temperatures (600°C) where conventional Si-based devices cannot function, being limited

to < 350°C. This is because of the low intrinsic carrier concentration of wide-bandgap

energy semiconductors at high temperature, as shown in Figure 1-1(left).

2

2

Figure 1-1. Semilog plot of intrinsic carrier concentration versus inverse temperature for Si, GaAs, GaN (left). Performance of AlGaN/GaN based power amplifier as compared to GaAs and SiC based devices (right).

The ability of GaN-based materials to function in high temperature, high power and

high flux/energy radiation conditions will enable large performance enhancements in a

wide variety of spacecraft, satellite, mining, automobile, nuclear power, and radar

applications. One additional attractive attribute of GaN is that sensors based on these

materials could be integrated with high-temperature electronic devices on the same chip.

AlGaN/GaN heterojunction based high electron mobility transistors (HEMTs) have

demonstrated extremely promising results for the use as power devices in many analog

applications due to the high sheet carrier concentration, electron mobility in the two

dimensional electron gas (2DEG) channel and high saturation velocity, as illustrated in

Figure 1-1 (right). Without any surface passivation, the sheet carrier concentration of the

polarization-induced 2DEGs confined at interfaces of AlGaN/GaN HEMT becomes

sensitive to any manipulation of surface charge. However, the nature of ambient

sensitivity for the unpassivated nitride used to build micro-sensors able to detect applied

strain and surface polarity change by polar liquids or toxic gas and harmful cancer cell

3

3

exposure to the surface of HEMTs. In addition, sensors fabricated from these wide

bandgap semiconductors could be readily integrated with solar blind UV detectors or

high temperature, high power electronics on the same chip [Kim00a, Ris94, Lut99].

Another wide bandgap semiconductor, ZnO proposed for the sensing applications

has several fundamental advantages. It has higher free-exciton binding energy (60 meV),

and more resistance to radiation damage, piezoelectricity and transparency. It has been

used effectively as sensing material based on near surface modification of charge

distribution with certain surface absorbed species. Another benefit of oxide-based

semiconductors is that they do not rely on specific catalytic metals for chemical detection

[Yun05]. Instead, they exploit the change in near surface conductivity most likely due to

the adsorption of chemical species. Most of the resistive gas sensors that employ surface

conductivity change have been metal oxide semiconductors (MOS) such as ZnO, SnO2,

In2O3, MoO3, WO3, and titanium substrated chromium oxide (CTO). The ZnO-based

MOS-based diodes and field effect transistors have been demonstrated and they can be

used with a wide range of chemicals using fractional surface conductivity by adsorption

and desorption.

CHAPTER 2 BACKGROUND AND LITERATURE REVIEW

2.1 Historical Review

The AlGaN/GaN high electron mobility transistors (HEMTs) have demonstrated

extremely promising results for use in broad band power amplifiers in wireless base

station applications, due to the high sheet carrier concentration, high electron mobility in

the two dimensional electron gas (2DEG) channel, and high saturation velocity [Mor99,

Eas02, Tar02, Zha00, Zha01, Joh02, Kou02, Pea99]. The high electron sheet carrier

concentration of nitride HEMTs is induced by piezoelectric polarization of the strained

AlGaN layer and spontaneous polarization [Kou02, Kan03, Che95, Ned98, Amb00].

Polarization induced piezoelectric properties play an important role in AlGaN/GaN

heterostructures. The high electron sheet carrier concentration in the strained

AlGaN/GaN layer suggests the possibility that nitride HEMTs may be excellent

candidates for sensing applications including pressure, chemical, gas, and biological

sensors.

For pressure sensor applications, only a few basic studies reported piezo effect

related cantilever beams [Str03, Dav04, Wu05]. Strittmatter et al. [Str03] reported that

capacitive strain can be sensed with GaN metal insulator semiconductor (MIS) diode but

for better performance, high quality surface oxide/GaN interface was needed. Davies et

al. [Dav04] first demonstrated the feasibility of free standing GaN cantilevers on Si

substrates. Both dry and wet etch processes were used to remove the Si substrate under

the GaN but any measured strain data using this device was not shown. Wu and Singh

5

5

[Wu05] examined potential for the strain sensor using a BaTiO3 piezoelectric

semiconductor field effect transistor. Two classes (Si and GaN) of heterostructures for

stress sensing were used and high sensitivity could be acquired using a very thin

piezoelectric layer. However, direct application of AlGaN/GaN HEMTs structure for

pressure sensors is not widely studied, especially for high pressure sensing.

Gas sensors have been fabricated on a number of semiconductors using catalytic

metals as the gate in the metal insulator semiconductor (MIS) or as the metal contact in

Schottky diodes [You82, Lun86, Rye87]. Various field effect transistors based on silicon

have been developed for hydrogen gas sensing [Lun89]. But silicon based sensors are

limited to operation in environments of below 250oC, prohibiting them from being used

as hydrocarbon detectors or for other applications requiring high temperature operation.

Because hydrocarbon gases should be decomposed by the catalytic metals and hydrogen

atoms diffuse to the device interface, it is presumed that a dipole forms, lowering the

effective work function of the metal and changing electrical characteristics of the devices.

Baranzanhi et al. [Bar95] demonstrated gas sensitivity of Pt gated SiC transistors

operating up to 500oC but the SiC Schottky diodes have displayed poor thermal stability.

Pd silicides were observed at temperatures as low as 425oC when Pd was used as

Schottky metal [Hun95, Che96]. Luther et al. [Lut99] first demonstrated Pt-GaN gas

sensor for hydrogen and propane at high temperature (200-400oC). The Pt–GaN gas

sensor showed faster response for hydrocarbons and enhanced sensitivity at higher

temperatures (500oC). After that, Schalwig et al. [Sch01] showed gas sensors for the

exhausted lean burn engines using Pt-GaN and Pt-HEMTs. The device performance was

investigated at high temperatures (200-600oC) and it was shown that a HEMT based gas

6

6

sensor was more sensitive than GaN diodes but detailed analysis of sensitivity differences

for GaN diodes vs HEMTs was left as future work.

Since the first demonstration of a fluid monitoring sensor based on AlGaN/GaN

hetero structures by Neuberger [Neu01], the application of AlGaN/GaN HEMTs as liquid

sensors has been a subject of intense research. Neuberger et al. suggested that the sensing

mechanism for chemical absorbates originated from compensation of the polarization

induced bound surface charge by interaction with polar molecules in the fluids. The time

dependence of changes in source-drain current of gateless HEMTs exposed to polar

liquids (isopropanol, acetone, methanol) with different dipole moments using

GaN/AlGaN hetero-interfaces was reported. In particular, it was shown that it is possible

to distinguish liquids with different polarities. Steinhoff et al. [Ste03a] suggested that the

native oxide on the nitride surface was responsible for the pH sensitivity of the response

of gateless GaN based heterostructure transistors to electrolyte solutions. It was shown

that the linear response of a nonmetallized GaN gate region using different pH valued

electrolyte solutions and sensitivity with a resolution better than 0.05 pH from pH = 2 to

pH = 12. Chaniotakis et al. [Cha05] showed that the GaN surface interacts selectively

with Lewis acids, such as sulphate (SO42-) and hydroxide (OH-) ions using impedance

spectra. It was also shown that gallium face GaN was considerably reactive with many

Lewis bases, from water to thiols and organic alcohols without any metal oxide and

nitrides

A novel metal oxide, ZnO has numerous attractive characteristics for gas and

chemical sensors [Kan05a, Kan05b, Heo04, Loo01, Nor04]. The bandgap can be

increased by Mg doping. The ZnO has been used effectively as gas sensor material based

7

7

on near surface modification of charge distribution with certain surface absorbed species.

In addition, it is attractive for biosensors given that Zn and Mg are essential elements for

neurotransmitter production and enzyme functioning [Ste05, Gur98]. The ZnO is

attractive for forming various types of nanorods, nanowires, and nanotubes [Hua01, Li04,

Kee04, Kin02, Liu03, Par03a, Ng03, Hu03, Par03b, Heo02, He03, Zhe01, Lyu03, Zha03,

Pan01, Lao03]. The large surface area of the nanorods makes them attractive for gas and

chemical sensing, and the ability to control their nucleation sites makes them candidates

for high density sensor arrays.

2.2 Background

2.2.1 AlGaN/GaN High Electron Mobility Transistors (HEMTs)

One of the most outstanding advantages of the GaN is the availability of

AlGaN/GaN heterostructures. The type I band alignment between AlGaN and GaN has

been shown to form a potential well and a 2-dimensional electron gas (2DEG) at the

heterointerface [Hen95]. When these materials are brought into contact, thermal

equilibrium requires alignment of their respective Fermi levels (EF). This induces

conduction (Ec) and valence (Ev) band bending in both the AlGaN and GaN layers and

can cause the GaN conduction band at the interface to drop below EF, as illustrated in

Figure 2-1.

Since the Fermi level can be viewed as an electrochemical potential for electrons,

majority electrons will accumulate in the narrow gap material just below the

heterointerface to fill the quasi triangular potential well between EC and EF. With the

heterointerface on one side and a potential barrier on the other, electrons in the 2DEG are

only free to move in along the plane of the interface. Modulation doped field effect

transistors (MODFETs) are a class of heterostructures FET that use selective barrier

8

8

doping to spatially separate ionized donors from the electrons in the 2DEG, leading to an

increase in channel mobility.

Figure 2-1. Simplified view of modulation doping with an enlarged view of energy band diagram illustrating formation of 2-dimensional electron gas at AlGaN/GaN heterointerface.

For this reason, these devices are also know as high electron mobility transistors, or

HEMTs.Unlike conventional III-V based HEMTs, such as AlGaAs/GaAs HEMTs, there

is no dopant in the typical nitride based HEMT structure and all the layers are undoped.

The carriers in the two dimensional electron (2DEG) gas channel is induced by

piezoelectric polarization of the strained AlGaN layer and spontaneous polarization,

which are very large in wurtzite III-nitrides. Carrier concentration > 1013 cm-3 in the

2DEG, which is 5 times larger than that in AlGaAs/GaAs material system, can be

routinely obtained. The portion of carrier concentration induced by the piezoelectric

effect is around 45-50%. This makes nitride HEMTs are excellent candidates for pressure

sensor and piezoelectric-related applications.

9

9

The piezoelectric polarization induced sheet carrier concentration of undoped Ga-

face AlGaN/GaN can be calculated by using Equation 2-1.

( ))()()()()()( 20 xExExe

edx

exxn CFb

ds ∆−+⎟⎟

⎠

⎞⎜⎜⎝

⎛−= φ

εεσ (2-1)

where σ(x) is piezoelectric polarization, x is the Al concentration in AlxGa1-xN, ε(x) is the

dielectric constant, dd is the AlGaN layer thickness, eφb is the Schottky barrier of the gate

contact on AlGaN, EF is the Fermi level and ∆Ec is the conduction band discontinuity

between AlGaN and GaN.

1E11 1E12 1E130.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

No relaxation Partial relaxation

X :

Al m

ole

fract

ion

Carrier concentration induced by PE(cm-2)

Figure 2-2. Sheet carrier concentration in the 2DEG channel of AlGaN/GaN HEMT induced by the piezoelectric polarization as a function of Al concentration

As shown in Figure 2-2, the sheet carrier concentration induced by the piezoelectric

polarization is a strong function of Al concentration. In the case of a partially relaxed

AlGaN strain layer, there is also a maximum sheet carrier concentration around Al =

0.35. If an external stress can be applied to AlGaN/GaN material system, the sheet carrier

10

10

concentration can be changed significantly and devices fabricated in this fashion could be

used in sensor-related applications.

The changes in the two dimensional (2D) channel of AlGaN/GaN HEMTs are

induced by spontaneous and piezoelectric polarization, which are balanced with positive

charges on the surface. Figure 2-3 shows schematic diagrams of the direction of the

spontaneous and piezoelectric polarization in both Ga and N face wurtzite GaN crystals

[Amb00].

For the Ga-face AlGaN/GaN HEMTs structure, at the surface of a relaxed GaN

buffer layer or a strained AlxGa1-xN barrier as well as at the interfaces of a AlxGa1-

xN/GaN heterostructure, the total polarization changes abruptly, causing a fixed two

dimensional polarization sheet charge σ, given by

σ AlGaN AlGaN AlGaNSP

AlGaNpzP P P= = + (2-2)

where, ),1(019.0)1(034.0090.0)( xxxxxP SPAlGaN −+−−−= (2-3)

)1(0402.00583.0)( xxxxP PZAlGaN −+−= . (2-4)

Figure 2-3. Piezoelectric (PE) and spontaneous (SP) polarization effects in Ga face or N-

face AlGaN/GaN heterostructures.

+σ

Ga face N face

-σ

PSP+ PPE

PS PS

PSP+ PPE

11

11

Therefore, the sheet charge density in the 2D channel of AlGaN/GaN HEMT is

extremely sensitive to its ambient. Numerous groups have demonstrated the feasibility of

AlGaN/GaN hetero-structures based hydrogen detectors with extremely fast time

response and capable of operating at high temperature (500-800oC), eliminating bulky

and expensive cooling systems [Amb02, Sch01, Eic01, Sch01a, Stu02, Eic03, Kim03a,

Kim03b, Kim03c, Amb03]. In addition, gateless AlGaN/GaN HEMTs show a strong

dependence of source/drain current on the polarity and concentration of polar solutions

[Ste03b]. There have also been recent reports of the investigation of the effect of external

strain on the conductivity of an AlGaN/GaN high electron mobility transistor [Kan03].

2.2.2 ZnO based Chemical Sensor

ZnO has numerous attractive characteristics for gas and chemical sensors [Kan05a,

Kan05b, Heo04, Loo01, Nor04] ZnO is a direct bandgap semiconductor normally form in

hexagonal (wurtzite) crystal structure like GaN, with lattice parameters a = 3.25 Å and c

= 5.12 Å. The Zn atoms are tetrahedrally coordinated with four O atoms, where the Zn –

d electrons hybridize with the O p-electrons. Alternating Zn and O layers form the crystal

structure shown in Figure 2-4.

Compared with GaN in Table 2-1 [Str00], it has direct bandgap energy of 3.37 eV,

which makes it transparent in visible light and operates in the UV blue wavelengths. The

exciton binding energy ~60 meV for ZnO, as compared to GaN ~25meV; the higher

exciton binding energy enhances the luminescence efficiency of light emission.

In the past, ZnO has been used in its polycrystalline form in applications ranging

from piezoelectric transducers [Kad92] to varistors [Lou80, Ver00], and as transparent

conducting electrodes [Pet79]. Recent improvements in the growth of high quality, single

crystalline ZnO in both bulk and epitaxial forms has revived interest in this material

12

12

[Pet79]. Especially, ZnO is a piezoelectric, transparent wide bandgap semiconductor used

in surface acoustic wave devices.

Figure 2-4. Crystal structure of wurtzite ZnO.

Table 2.1 Physical properties of GaN and ZnO. GaN ZnO

Bandgap 3.39 eV, direct 3.37 eV, direct

Crystal structure (Wurtzite)

a = 3.189 Å, c = 5.206 Å (c/a = 1.632)

a = 3.250 Å, c = 5.205 Å (c/a = 1.602)

Mobility 1000 cm2/V·s (e) 30 cm2/V·s (h)

200 cm2/V·s (e) 5-50 cm2/V·s (h)

Effective mass m* = 0.20m0 (e) m* = 0.80m0 (h)

m* = 0.24m0 (e) m* = 0.59m0 (h)

Saturation velocity 2.5×107 cm/s 3.2×107 cm/s

Exciton binding energy 28 meV 60 meV

Energy gaps of several common semiconductors are given in Figure 2-4 as a

function of lattice constant. The approximate boundaries of the visible spectrum are

shown as is the nature of the energy transition for each material. From the figure, it is

c

a

13

13

noted that group III-nitrides and II oxides have bandgap from red to shorter UV

wavelengths. The bandgap can be increased by Mg doping. ZnO has been effectively

used as a gas sensor material based on the near-surface modification of charge

distribution with certain surface-absorbed species [Kan93].

2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.01

2

3

4

5

6

7

8

red

violet

6H

-SiC

sapphire(O

substrate

)MgO

ZnO

CdO

InN

AlN

GaN

Ene

rgy

band

gap

(eV

)

Lattice constant a(A)

Figure 2-4. Energy bandgap of several, III-V, and II-IV compound semiconductors as a function of lattice constant.

In addition, it is attractive for biosensors given that Zn and Mg are essential

elements for neurotransmitter production and enzyme functioning [Mil98, Oga04]. ZnO

is attractive for forming various types of nanorods, nanowires and nanotubes [Hua01,

Li04, Kin02, Liu03, Par03a, Ng03, Hu03b, Par03b, Heo02, Nor04, Poo03, He03, Wu00,

Zhe01, Lyu01, Zha03b, Par03c, Yao02, Pan01, Lao03]. Compared with bulk materials, a

significant characteristic of nanostructures is their high surface to volume ratio. ZnO

nanowires have the same crystal structures as bulk structure, confirmed by X-ray

diffraction (XRD) and transmission electron microscopy (TEM) [Pan01, Roy03, Li02].

o

14

14

Therefore, all the bulk properties are still preserved for the nanowires and ZnO nanorods

are very promising for a wide variety of sensor applications.

15

CHAPTER 3 PRESSURE SENSOR USING PIEZOELECTRIC POLARIZATION

3.1 Introduction

There are a number of applications in the automotive, aerospace, and industrial

fields for robust miniaturized pressure sensors. A number of different semiconductors

systems have been used to make piezoresistive sensors [Ko99, Ned98, Wu97, You04,

Dad94]. Especially, polarization induced piezoelectric properties play a very important

role in strained AlGaN/GaN heterostructures. The high electron sheet carrier

concentration in the AlGaN/GaN layer suggests that nitride HEMTs can be used as

excellent pressure sensors.

Several research groups have reported piezo effect related GaN pressure sensors

[Str03, Dav04, Wu05]. Strittmatter et al. [Str03] reported that capacitive strain can be

sensed with GaN metal insulator semiconductor (MIS) diode but for better performance,

high quality surface oxide film on the GaN layer was needed. Davies et al. [Dav04] first

demonstrated the feasibility of the free standing GaN cantilevers on Si substrates but any

measured strain data using this device was not shown. Wu and Singh [Wu05] examined

the potential for the strain sensor using BaTiO3 piezoelectric semiconductor field effect

transistor. Two classes (Si and GaN) of heterostructures were used for stress sensing and

showed that high sensitivity can be acquired using a very thin piezoelectric BaTiO3 layer.

However, the direct application of AlGaN/GaN HEMTs structure for pressure sensors is

not widely studied, especially for high pressure sensing.

16

In this chapter, a thorough discussion of pressure sensor using AlGaN/GaN HEMT

structure will be given. In Section, 3.2, the effect of external strain on the sheet resistance

of the two dimensional electron gas channel in AlGaN/GaN HEMTs grown on sapphire

will be presented for the cantilever beam. In Section 3.3, AlGaN/GaN membrane

pressure sensor fabricated on Si substrate will be analyzed, which can overcome the

rigidity of sapphire substrate in applying external stress. In Section 3.4, the capacitive

pressure sensor which is less sensitive to variations in contact resistance will be

discussed.

3.2 Effect on External Strain on the Conductivity of AlGaN/GaN HEMTs

Two terminal high electron mobility Al0.25Ga0.75N/GaN devices used with simple

bonding test were used to to study the effect of external strain on the conductivity on the

sheet resistance of the two-dimensional electron gas channel in an AlGaN/GaN HEMT.

The HEMT structures consisted of a 3µm thick undoped GaN buffer, 30Å thick

Al0.25Ga0.75N spacer, 270Å thick Si-doped Al0.3Ga0.7N cap layer. The epi-layers were

grown on sapphire substrate by metal organic chemical vapor deposition (MOCVD). Two

sets of samples were fabricated; one with mesa definition and the other without the mesa

as shown in Figure 3-1(top). Mesa isolation was performed with an Inductively Coupled

Plasma (ICP) etching with Cl2/Ar based discharges at –90 V dc self-bias, ICP power of

300 W at 2 MHz and a process pressure of 5 mTorr. 100 × 100 µm2 ohmic contacts

separated with gaps of 5, 10, 20, 50 and 100 µm like a transmission-line-method (TLM)

pattern consisted of e-beam deposited Ti/Al/Pt/Au patterned by lift-off and annealed at

850 ºC, 45 sec under flowing N2. Plated Au was subsequently deposited on the ohmic

metal pads for wire bonding on the samples as shown in Figure 3-1(bottom).

17

Figure 3-1. Schematic diagrams of layer structures of two-terminal non-mesa(top) and mesa(middle) devices, and top view photo-micrograph of fabricated two-terminal devices with different channel lengths(bottom).

The devices were fabricated on half of 2” wafer, sawed into 2 mm wide stripes and

wire bonded on the test feature. The dc characteristics were obtained from measurements

on an Agilent 4156C parameter analyzer. Figure 3-2 illustrates the setup for measuring

the effect of external strain on the conductivity of 2DEG channel of the nitride HEMT.

18

Lucite blocks secure the sample and PCB board for testing. The contact pads were

connected to the PCB board, which had BNC connectors on the end for signal outputs,

with 1 mil thick gold wire. A high precision single axis traverse was used to bend the

sample.

Sheet charges in the AlGaN/GaN high electron mobility transistors (HEMTs) are

induced by spontaneous polarization and piezoelectric polarization [Kou02, Ras02,

Amb00, Che95]. Wurtzite GaN and AlGaN are tetrahedral semiconductors with a

hexagonal Bravais lattice with four atoms per unit cell. The misfit strain inside a film is

measured against its relaxed state. In the misfit strain calculation, the strain is calculated

against the relaxed films, ao, i.e.

)(

)()(xa

xaxa

o

omisfit

−=ε (3-1)

where a(x) is the lattice constants of AlxGa1-xN and ao(x) = (aGaN – aAlNx)Å = (3.189 -

0.077x)Å [Amb99]. Here a(x) - ao(x) is chosen because the AlGaN film is always under

tension due to a(x) ≤ aGaN and misfit will always be positive. Having defined the elastic

strain in the film, a partially relaxed film parameter should be defined. For a perfectly

coherent film, a(x) = aGaN or

xxxxa

xaa

o

oGaNmisfit 024.0

189.3077.0

077.0189.3077.0

)()(

.max =≈−

=−

=ε (3-2)

For a partially relaxed film, the ratio of strain comparing to the un-relaxed state

defines the degree of strain in the film,

( ))()()(

.max xaaxaxaxS

oGaN

o

misfit

misft

−−

==εε

(3-3)

Hence, the degree of relaxation is then given by

19

( ))()()(1

xaaxaaxSxr

oGaN

GaN

−−

=−= (3-4)

For around 300 Å of AlGaN layer on the top of GaN, r(x) was measured by

Ambacker [Amb00].

0 0 ≤ x ≤ 0.38

r(x) = 3.5x-1.33 0.38 ≤ x ≤0.67 (3-5)

1 0.67 ≤x ≤1

Lucite

To BNC

Lucite

PCB

Cantilever

HEMT

High precisionsingle axis traverse

Cantilever

Figure 3-2. Schematic diagram of a pressure sensor package: experimental setup to detect I-V characteristics connected to the BNC cable according to various mechanical stresses (top) and mechanical stressor with cantilever (bottom).

20

0 4 8 12 160.0

0.2

0.4

0.6

X : A

l mol

e fr

actio

n

Strain(x103)

No relaxation Partial relaxation

1011 1012 10130.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

X :

Al m

ole

frac

tion

Carrier concentration induced by PE(cm-2)

No relaxation Partial relaxation

Figure 3-3 Strain induced by AlGaN on GaN for the un-relaxed and partially relaxed AlGaN layer as a function of Al concentration (top) and sheet carrier concentration induced by the piezoelectric polarization as a function of Al concentration (bottom)

21

The piezoelectric polarization, for partially relaxed strained layer can be expressed

by modifying the Equation 3-3 subtracting the portion of relaxation

( ) ( ) ⎟⎟⎠

⎞⎜⎜⎝

⎛−⎟⎟

⎠

⎞⎜⎜⎝

⎛ −−=

33

133331

0

0)(12CCee

aaaxrx GaNσ (3-6)

where e31 and e33 are the piezoelectric coefficients and C13 and C33 are the elastic

constants.

Figure 3-3 (left) plots the relationship between Al concentration of AlGaN and

strain induced by AlGaN on GaN for the un-relaxed and partially relaxed conditions.

Unlike the linear model for the un-relaxed condition, there is a maximum strain around

Al concentration of 0.35. The piezoelectric polarization induced sheet carrier

concentration of undoped Ga-face AlGaN/GaN can be calculated by following equation

[Asb97]:

( ))()()()()()( 20 xExExe

edx

exxn CFb

ds ∆−+⎟⎟

⎠

⎞⎜⎜⎝

⎛−= φ

εεσ (3-7)

where ε(x) is the dielectric constant, dd is the AlGaN layer thickness, eφb is the Schottky

barrier of the gate contact on AlGaN, EF is the Fermi level and ∆Ec is the conduction

band discontinuity between AlGaN and GaN. In the case of applying external tensile

strain on the HEMT sample, a increase of conductivity was observed.

The mesa depth was around 500 Å, which is below the AlGaN/GaN interface (300

Å AlGaN layer on 3 µm GaN layer). As illustrated in Figure 3-3 (top), the sheet carrier

concentration induced by the piezoelectric polarization is a strong function of Al

concentration. In the case of a partially relaxed AlGaN strain layer, there is also a

maximum sheet carrier concentration around Al = 0.35. If an external stress can be

applied to AlGaN/GaN material system, the sheet carrier concentration can be changed

22

significantly and devices fabricated in this fashion could be used in sensor-related

applications.

In order to study the effect of external strain on the conductivity of the HEMT

material systems, transmission line patterns were fabricated with a mesa of 500 Å.

Elastic bending of an analogous system, GaAs grown on Si wafers, has been carefully

studied using Stony’s equation [Sto09, Chu98]. The curvature and wafer bowing for

different thickness of GaAs grown on Si was modeled [Chu98]. A pure bending of a

single beam was used in this work to estimate the strain, since the HEMT structure,

around 3 µm, is much thinner that that of sapphire substrate, 200 µm, and the degree of

deflection (maximum deflection is around 2.2 mm) is much shorter than the length of the

beam, 27 mm. The strain, εxx, of the bending can be estimated from the single beam with

thickness of t and unit width. The tensile strain near the top surface of the beam is simply

given by

εxx = td/ L2 (3-8)

where t is the sample thickness, d is the deflection and L is the length of the beam.

Figure 3-4 (top) shows the effects of external tensile and compressive strain on the

conductivity of AlGaN/GaN HEMT sample with mesa. Therefore, the AlGaN layer sits

above the beam and the external stress applied on the beam should not change the strain

of the AlGaN layer.

As a result, applying a tensile stress on the beam would pull apart the GaN atoms

only and not affect the AlGaN. The total strain on the AlGaN/GaN interface and

piezoelectric should increase, as observed in Figure 3-4 (top). In the case of applying a

23

compressive stress on the beam, the GaN atoms were pushed together and reduced the

total strain at the AlGaN/GaN interface and the resultant conductivity.

0 2 4 6 8 10 12 147.75

7.80

7.85

7.90

7.95

8.00

8.05

Con

duct

ivity

(mS)

Compressive or tensile stress(x106)

Tensile Compressive

0 1 2 3 4 5

8.50

8.55

8.60

8.65

Con

duct

ivity

(mS)

Compressive or tensile stress(x106)

Tensile Compressive

Figure 3-4. The effect of tensile or compressive stress on the conductivity of the AlGaN/GaN HEMT with mesa etching (top) and without mesa etching (bottom).

24

For the non-mesa devices fabricated on the same wafer in a different area, Figure 3-

4(bottom) shows the strain dependence of with a channel length of 10 µm. A reversal of

strain sensitivity indeed has observed indicating a larger strain relaxation in the

AlGaN/GaN layer.

The changes in conductance of the channel of Al0.25Ga0.75N/GaN high-electron-

mobility transistor structures during application of both tensile and compressive strain

were measured relatively large. For fixed Al mole fraction, the changes in conductance

were roughly linear over the range up to 1.4 x 107 N.m-2, with coefficients for planar

devices of + 9.8 x 10-12 S-N–1-m2 for tensile stress and – 1.05 x 10 -11 S-N-1-m2 for

compressive stress. For mesa-isolated structures, the coefficients were smaller due to the

reduced effect of the AlGaN strain, with values of –1.2 x10 –12 S-N-1-m2 for tensile stress

and + 1.97 x10 –12 S-N-1-m2 for compressive stress. The large changes in conductance

demonstrate that simple AlGaN/GaN heterostructures are promising for pressure and

strain sensor applications.

In summary, we have demonstrated the effect of the external strain on the

piezoelectric polarization of AlGaN/GaN material systems. This relatively large effect

may have application in strain and pressure sensors.

3.2 Pressure induced Changes in the Conductivity of AlGaN/GaN HEMTs

AlGaN/GaN high electron mobility transistors (HEMTs) show a strong dependence

of source/drain current on the piezoelectric polarization induced two dimensional electron

gas (2DEG). The spontaneous and piezoelectric polarization induced surface and

interface charges can be used to develop very sensitive but robust sensors for the

detection of pressure changes. The changes in the conductance of the channel of a

25

AlGaN/GaN High Electron Mobility Transistor (HEMT) membrane structure fabricated

on a Si substrate were measured during the application of both tensile and compressive

strain through changes in the ambient pressure.

The piezoelectric polarization induced sheet carrier concentration of undoped Ga-

face AlGaN/GaN can be calculated by following equation [Amb00, Amb99, Lu01]:

( ))()()()()()( 2

0 xExExeedx

exxn CFb

ds ∆−+⎟⎟

⎠

⎞⎜⎜⎝

⎛−= φ

εεσ (3-9)

where ε(x) is the dielectric constant, dd is the AlGaN layer thickness, eφb is the Schottky

barrier of the gate contact on AlGaN, EF is the Fermi level and ∆Ec is the conduction

band discontinuity between AlGaN and GaN. The sheet carrier concentration induced by

the piezoelectric polarization is a strong function of Al concentration. If an external

stress can be applied to AlGaN/GaN material system, the sheet carrier concentration can

be changed significantly and devices fabricated in this fashion could be used in sensor-

related applications.

Figure 3-5. Circular membrane of AlGaN/GaN on a Si substrate fabricated by etching a circular hole in the substrate (left). A deflection of the membrane away from the substrate due to differential pressure on the two sides of the membrane produces a tensile strain in the membrane (right).

26

The sensors used to monitor the differential pressure are made of a circular

membrane of AlGaN/GaN on a Si substrate by etching a circular hole in the substrate, as

illustrated in Figure 3-5(left). A deflection of the membrane away from the substrate due

to differential pressure on the two sides of the membrane produces a tensile strain in the

membrane, as shown in Figure 3-5(right).

The differential piezoelectric responses of AlGaN and GaN layers creates a space

charge which induces 2DEG at the AlGaN/GaN interface. The concentration of 2DEG is

expected to be directly correlated with the tensile strain in the membrane and hence with

the differential pressure. The radial strain is given by [Sto09, Chu98],

εr = (S-D)/D = (2θR – 2Rsinθ)/(2Rsinθ) = θ/sinθ - 1 (3-10)

The total tensile force around the edge of the circular membrane is,

T = πDtGaNσr = πD tGaN [EGaN/ (1- ν)] εr (3-11)

where tGaN is the film of thickness, D is the diameter of the via hole, σr = [EGaN/ (1- ν)] εr,

EGaN is the Young’s modulus and ν is the Poisson’s ratio of the GaN film. The

component of T along z direction is balanced by the force on the membrane due to a

differential pressure Pi - P0, where Pi and P0 are the inside and outside pressure

respectively. Hence Tsinθ = (Pi - P0 ) πD2/4 and the radial strain, εr, in the nitride film can

be expressed as a function of the differential pressure Pi - P0.

(θ - sin θ) = (Pi - P0 )[(1- ν)D]/(4EGaNtGaN) (3-13)

where EGaN is the Young’s modulus, D is the diameter of the via hole, tGaN is the film of

thickness and ν is the Poisson’s ratio of the GaN film. If θ is measured, the differential

pressure Pi - P0. can be estimated with Equation 3-13. We also derive the relationship

between conductance, σ, of AlGaN/GaN HEMT and radial strain, εr.

27

σ = α( rAlGaN) + β( fGaN ) × εr (3-14)

where α( rAlGaN) = µs1/[1 + (ε0ε(x)/tAlGaN e2)h2/4πm*(x)]- |∆PSP| –|eeff |∆εAlGaN rAlGaN – [ε0ε(x)/tAlGaN e] [eφb(x) – ∆Ec(x)] (3-15)

and β( fGaN ) = µs1/[1 + (ε0ε(x)/tAlGaN e2 )h2/4πm*(x)] |eeff (GaN)| - |eeff (AlGaN)| (3-16)

where µs is the mobility of 2DEG, ε0 is the electric permitivity, ε(x) = 9.5 – 0.5x is the

relative permitivity, eφb(x) = 0.84 + 1.3x (eV) is the Schottky barrier height, eeff = (e31-

e33)C13/ C33, h is Plank constant, e is the electron charge, m*(x) ~ 0.228me. By

monitoring the conductance of the HEMT on membrane, the pressure difference, Pi - P0,

can be obtained. The detail deivations of above equations are listed in the Appendix.

Figure 3-6. Schematic diagram of device structure with a finger patterned device on the HEMT membrane.

The HEMTs were grown by metalorganic chemical vapor deposition on 100 mm

(111) Si substrates at Nitronex Corporation. The structures consisted of an (Al,Ga)N-

based transition layer, ~0.8 µm undoped GaN buffer, and 300Å undoped AlGaN barrier

layer. Mesa isolation was performed with an Inductively Coupled Plasma (ICP) etching

with Cl2/Ar based discharges at –90 V dc self-bias, ICP power of 300 W at 2 MHz and a

28

process pressure of 5 mTorr. Ti/Al/Pt/Au based inter-digitated finger pattern separated by

4 µm was formed with e-beam deposition and standard lift-off shown in Figure 3-6.

The fingers were annealed at 850 ºC, 45 sec under flowing N2. Plated Au was

subsequently deposited on the ohmic metal pads for wire bonding on the samples. Via

holes were fabricated from the back side of the Si substrate and stopping on the GaN

layer using ICP etching with SF6/Ar. The etch selectivity is more than 1000:1. 2000Å of

AuSn was deposited on the backside of the sample and a glass slice. A RD automation

flip-chip bonder was used to bond the glass slice and the sample at 400 °C to seal off the

via holes.

Figure 3-7 shows scanning electron microscopy (SEM) photos of the via through

the Si wafer(left) and cross sectional view of the via(right).

Figure 3-7. SEM micrographs of via through the Si wafer(left)and cross sectional view of a via hole(right)

The dc current-voltage(I-V) characteristics were obtained from measurements on

an Agilent 4156C parameter analyzer while the device was measured at 25oC under either

Si Substrate

GaN

29

vacuum(10mTorr) or pressure (40-200psi) conditions. Figure 3-8 shows the drain-source

I-V characteristics from the membrane HEMT structure as a function of the ambient

pressure. This current increases with increasing pressure and decreases under vacuum

conditions.

-3 -2 -1 0 1 2 3-12

-8

-4

0

4

8

12

-0.4 -0.2 0.0 0.2 0.4-3-2-10123

I DS(

mA)

VDS(V)

I D

S(mA)

VDS(V)

10 mT atmosphere 40 psi 100 psi 200 psi

Figure 3-8. IDS-VDS characteristics at 25C from AlGaN/GaN HEMT membrane as a function of applied pressure.

The resulting channel conductance derived from this data is shown as a function of

differential pressure in Figure 3-9. In the case of applied positive pressure, which

corresponds to compressive strain induced in the HEMT layers, the conductivity

decreases with a coefficient of -7.1x10-2 mS/bar. For the case of applied negative

pressure (vacuum), the conductivity shows a positive coefficient of the same value within

experimental error, given the limited data for vacuum conditions. These trends are similar

to those observed with actual bending of HEMT samples on a cantilever beam to produce

tensile or compressive strain [Kan03], but exhibit sensitivities to the induced tensile or

compressive strain of almost two orders of magnitude larger. This is due to the absence of

30

the thick sapphire substrate that is present in the cantilever structures. The new membrane

structures are particularly sensitive to changes in differential pressure.

0 2 4 6 8 10 12 14

4.6

4.8

5.0

5.2

5.4

5.6

Con

duct

ivity

(mS

)

Pressure difference, |Po-Pi|(bar)

Vacuum(tensile strain) Pressure(compressive strain)

Figure 3-9.Channel conductivity of the AlGaN/GaN HEMT membrane as a function of differential pressure.

AlGaN/GaN high electron mobility transistors (HEMTs) show a strong dependence

of source/drain current on the piezoelectric polarization induced two dimensional electron

gas (2DEG). The spontaneous and piezoelectric polarization induced surface and

interface charges can be used to develop very sensitive but robust sensors for the

detection of pressure changes. The changes in the conductance of the channel of a

AlGaN/GaN High Electron Mobility Transistor (HEMT) membrane structure fabricated

on a Si substrate were measured during the application of both tensile and compressive

strain through changes in the ambient pressure. The conductivity of the channel shows a

linear change of –(+)7.1x10-2 mS/bar for application of compressive(tensile) strain. The

AlGaN/GaN HEMT membrane-based sensors appear to be promising for pressure

sensing applications.

31

In summary, an AlGaN/GaN HEMT membrane on Si shows large changes in

channel conductivity as a result of changes in ambient pressure. These structures appear

promising for use in integrated sensors in which the HEMTs can also be used for gas,

chemical and biological detection combined with on-chip transmission of the data.

3.3 Capacitance Pressure Sensor Based on GaN HEMT on Si Membrane

The AlGaN/GaN high-electron-mobility transistors (HEMTs) show a strong

dependence of the conductance of the channel when a membrane structure fabricated on a

Si substrate was measured during changes in the ambient pressure [Kan04c, Kan03,

Pea04a]. However, one drawback of piezoresistive sensors is that contact resistance

changes significantly with temperature and may mask the changes in sensor signal from

actual pressure changes [You04]. By sharp contrast, capacitive pressure sensors are less

sensitive to variations in contact resistance and in addition, sensors based on AlGaN/GaN

HEMTs could be readily integrated with off-chip wireless communication chips that

eliminate additional wiring capacitance. AlGaN/GaN high electron mobility transistors

(HEMTs) have demonstrated extremely promising results for use in broad-band power

amplifiers in wireless base station applications [Zha01, Tar02, Zha03a, Shu98]. The high

electron sheet carrier concentration of nitride HEMTs is induced by piezoelectric

polarization of the strained AlGaN layer and spontaneous polarization [Amb00, Amb99,

Asb97], suggesting that nitride HEMTs are excellent candidates for robust pressure

sensing.

In this part, Circular AlGaN/GaN diaphragms fabricated with radii 200-600 µm on

Si substrates show linear changes in capacitance over a range of applied pressure and that

the sign of the capacitance change is reversed when vacuum is applied to the diaphragm.

32

The sensors used to monitor the differential pressure are made of a circular

membrane of AlGaN/GaN HEMT on a Si substrate. The membrane is fabricated by

etching a circular hole in the substrate, as shown schematically in Figure 3-10 (top). A

scanning electron microscope (SEM) cross-sectional view of an actual device is shown at

the bottom of Figure 3-10

Figure 3-10. Schematic diagram of device structure (top) and SEM micrograph of AlGaN/GaN circular membrane on a Si substrate fabricated by etching a circular hole in the substrate(bottom).

Si Substrate

GaN

Si

Teflon

33

A deflection of the membrane away from the substrate due to differential pressure

on the two sides of the membrane produces a tensile strain in the membrane. This leads

to a change in the piezo-induced two dimensional electron gas (2DEG) density at the

AlGaN/GaN interface.

This in turn affects the capacitance of the HEMT diaphragm. The carrier density is

therefore directly correlated with the tensile strain in the membrane and hence with the

differential pressure. We have previously calculated the radial strain, εr, in the membrane

as a function of the differential pressure Pi - P0 [Kan04c] by employing a modified Stoney

analysis [Sto09, Chu98]. By monitoring the conductance of the HEMT on membrane,

the pressure difference, Pi - P0, can be obtained.

The HEMTs were grown by metal-organic chemical vapor deposition on 100 mm

(111) Si substrates at Nitronex Corporation. The structures consisted of an (Al,Ga)N-

based transition layer, ~0.8 µm undoped GaN buffer, and 300Å undoped AlGaN barrier

layer. Mesa isolation was performed with an Inductively Coupled Plasma (ICP) etching

with Cl2/Ar based discharges at –90 V dc self-bias, ICP power of 300 W at 2 MHz and a

process pressure of 5 mTorr. Ti/Al/Pt/Au based inter-digitated finger pattern separated

by 4 µm was formed with e-beam deposition and standard lift-off. The fingers were

annealed at 850 ºC, 45 sec under flowing N2. Plated Au was subsequently deposited on

the ohmic metal pads for wire bonding on the samples. Ohmic contact for the silicon

used 1000Å Al deposited by using sputter and annealed in nitrogen at 300°C. Via holes

were fabricated from the back side of the Si substrate and stopping on the GaN layer

using ICP etching with SF6/Ar. The etch selectivity is more than 1000:1. 2000Å of AuSn

was deposited on the backside of the sample and a glass slice.

34

0 2 4 6 8 1012

16

20

24

28

32

36

Cap

acita

nce(

pF)

Pressure(bar)

R=600 µm R=400 µm R=280 µm R=200 µm

Figure 3-11. Top view of HEMT capacitance pressure sensor(top) and capacitance as a function of pressure for different diaphragm radii (bottom).

The teflon bonding spin coating on the silicon wafer employed liquid Teflon

(CYTOP CTL-809M, from Bellex International Corp.) at 5000 rpm to a thickness of ~

5000Å and then flipchip bonding (RD automation flip-chip bonder) the fabricated device

35

using a mechanical force of 1000g between the upper and lower chucks for 10 minutes

and finally baking for 1 hour at 200°C to solidify the Teflon to seal off the via holes.

Figure 3-11(top) shows a top view of a completed HEMT capacitive pressure sensor.

The capacitance of the HEMT diaphragm structures were obtained from

measurements on an Agilent 4156C parameter analyzer while the device was measured at

25oC under either vacuum (-1 bar) or pressure (+9.5 bar) conditions. Figure 3-11(bottom)

shows the capacitance from the membrane HEMT structure as a function of the ambient

pressure, for different membrane radii. This capacitance increases with increasing

pressure and decreases under vacuum conditions, due to corresponding changes in the

carrier density in the 2DEG.

-2 0 2 4 6 8 10

-1

0

1

2

3

Cap

acita

nce

chan

ge(p

F)

Pressure(bar)

R=600 µm R=400 µm R=280 µm R=200 µm

Figure 3-12. Capacitance change as a function of radius of the AlGaN/GaN HEMT membrane over the pressure range from -1 to +9.5 bar.

36

The resulting capacitance change derived from this data is shown as a function of

pressure in Figure 3-12. In the case of applied positive pressure, which corresponds to

compressive strain induced in the HEMT layers, the capacitance increases in a linear

fashion over the range between 0 and +1 bar with a sensitivity of 0.86pF/bar for a 600

µm radius membrane. For the case of applied negative pressure (vacuum), the

conductivity shows a sensitivity of the same value within experimental error to a vacuum

of -0.5 bar. Within the linear range, the devices exhibited a hysteresis of <0.4%. Outside

these pressure limits the sensor has reduced sensitivity due to the device geometry. The

sensitivity could be increased by having a shallower via depth, obtained by thinning the

Si substrate.

Figure 3-13. Capacitance change as a function of radius of the AlGaN/GaN HEMT membrane over the pressure range from -1 to +9.5 bar

These trends are similar to those observed with actual bending of HEMT samples

on a cantilever beam to produce tensile or compressive strain [Kan03], but exhibit much

150 300 450 600

1

2

3

Cap

acita

nce

chan

ge(p

F)

Radius( m)µ

37

larger sensitivities to the induced tensile or compressive strain. This is due to the absence

of the thick sapphire substrate that is present in the cantilever structures, as we also

reported for the piezo-conductance membrane sensors previously [Kan04c].

Figure 3-13 shows the capacitance change as a function of radius of the

AlGaN/GaN HEMT membrane at a fixed pressure of +9.5 bar. The capacitance of the

channel displays a change of 7.19 +/-0.45x10-3 pF/µm. The sensor characteristics

measured at this same pressure on several different days showed a maximum capacitance

variation of 0.07pF, corresponding to a sensing repeatability of ~0.15 bar. The high

temperature characteristics still need to be established, but in this case will be limited by

the thermal stability of the ohmic contacts. Recent reports have shown that some contact

metallization on GaN HEMTs are stable for extended periods at 500oC [Sel04].

In summary, an AlGaN/GaN HEMT membrane on Si shows large changes in

capacitance as a result of changes in ambient pressure. This approach is less sensitive to

contact resistance variations with temperature than the previous conductance sensors. The

sensors can also be readily integrated with conventional HEMTs or Si circuitry to provide

off-chip wireless transmission of pressure data.

38

CHAPTER 4 CATALYST BASED GAS SENSOR FOR HYDROCARBON GASES

4.1 Introduction

Gas sensors have been fabricated on a number of semiconductors using catalytic

metals as the gate in the metal insulator semiconductor (MIS) or as the metal contact in

Schottky diodes [You82, Lun86, Rye87]. Various field effect transistors based on silicon

have been developed by several groups for hydrogen gas sensing [Lun89]. But the silicon

based sensors are limited to operation in environments of below 250oC, prohibiting them

from being used as hydrocarbon detectors or for other applications requiring high

temperature operation. Because hydrocarbon gases should be decomposed by the

catalytic metals and hydrogen atoms diffuse to the device interface, it is presumed that a

dipole forms, lowering the effective work function of the metal and changing electrical

characteristics of the devices.

Baranzanhi et al. [Bar95] demonstrated gas sensitive Pt gated SiC transistors

operating up to 500oC but the SiC Schottky diodes have displayed poor thermal stability

and formation of Pd silicides has been observed at temperatures as low as 425oC when Pd

was used as Schottky metal [Hun95, Che96]. Luther et al. [Lut99] first demonstrated Pt-

GaN gas sensor for hydrogen and propane at high temperature (200-400oC). It was also

exhibited that Pt–GaN gas sensor showed faster response for hydrocarbons and enhanced

sensitivity at higher temperatures (500oC) After that, Schalwig et al. [Sch01] showed gas

sensors for the exhausted lean burn engines using Pt-GaN and Pt-HEMTs. The device

performance at high temperature (200-600oC) was investigated. It was also shown that a

39

HEMT based gas sensor was more sensitive than GaN diodes but the detailed analysis of

sensitivity difference between GaN diodes and HEMTs was left as future work.

In this chapter, extensive discussion of all the processes including fabrication,

measurement, and characterization of gas sensor devices using wide bandgap

semiconductor diodes and transistors will be given. In Section 4.2, the higher

performance of AlGaN/GaN MOS diode for hydrogen gas sensor will be given compared

with Schottky GaN diode. In Section 4.3, hydrogen reversible changes in drain and

source current in the transistor based gas sensor will be given. In addition, the reason for

higher sensitivity of this structure operated with gain will also be discussed. In Section

4.4, a direct comparison of MOS and Schottky W/Pt-GaN diode for hydrogen detection

will be given. In Section 4.5, the method for detecting ethylene (C2H4), which causes

problems because of its strong double bond s and hence the difficulty in dissociating it at

modest temperature will be investigated using wide band gap semiconductor. In Section

4.6, nanotechnology driven sensor for hydrogen and ozone detection will be discussed

using multiple ZnO nanorods.

4.2 AlGaN/GaN based MOS Diode Hydrogen Gas Sensor

Simple GaN Schottky diodes exhibit strong changes in current upon exposure to

hydrogen containing ambients [Ste03a, Neu01, Sch01, Eic01, Sch02a, Stu02, Eic03,

Kim03a]. The effect is thought to be due to a lowering of the effective barrier height as

molecular hydrogen catalytically cracks on the metal gate and atomic hydrogen diffuses

to the interface between the metal and GaN, altering interfacial charge. Steinhoff et al.

[Ste03a] founded that it was necessary to have a native oxide present between

semiconductor and the gate metal in order to see significant current changes. Thus, it is

40

desirable to specifically incorporate an oxide into GaN-based diodes or HEMTs in order

to maximize the hydrogen detection response.

Figure 4-1.Cross-sectional schematic of completed MOS diode on AlGaN/GaN HEMT layer structure (top) and plan-view photograph of device(bottom).

Gas sensors based on MOS diode on AlGaN/GaN high electron mobility

transistor(HEMT) layer structure are of interest, because HEMTs are expected to be the

first GaN electronic device that is commercialized, as part of next generation radar and

wireless communication systems. These structures have much higher sensitivity than

Schottky diodes on GaN layer, because they are true transistors and therefore operate

41

with gain. In addition, the MOS-gate version of the HEMT has significantly better

thermal stability than a metal-gate structure [Kha01, Pal00, Sim00, Kou02, Sim02] and is

well-suited to gas sensing. When exposed to changes in ambient, changes in the surface

potential will lead to large changes in channel current.

HEMT layer structures were grown on C-plane Al2O3 substrates by Metal Organic

Chemical Vapor Deposition (MOCVD). The layer structure included an initial 2µm thick

undoped GaN buffer followed by a 35nm thick unintentionally doped Al0.28Ga0.72N layer.

The sheet carrier concentration was ~1×1013 cm-2 with a mobility of 980 cm2/V-s at room

temperature

Mesa isolation was achieved with 2000 Å plasma enhanced chemical vapor

deposited SiNx. The ohmic contacts was formed by lift-off of e-beam deposited

Ti(200Å)/Al(1000Å)/Pt(400Å)/Au(800Å). The contacts were annealed at 850 °C for 45

sec under a flowing N2 ambient in a Heatpulse 610T system. 400Å Sc2O3 was deposited

as a gate dielectric through a contact window of SiNx layer. Before oxide deposition, the

wafer was exposed to ozone for 25 minutes. It was then heat in-situ at 300 °C cleaning

for 10mins inside the growth chamber. 100 Å Sc2O3 was deposited on AlGaN/GaN by rf

plasma-activated MBE at 100 °C using elemental Sc evaporated from a standard effusion

all at 1130 °C and O2 derived from an Oxford RF plasma source [Gil01, Kim00b,

Kim00c]. 200 Å Pt Schottky contact was deposited on the top of Sc2O3. Then, final

metal of e-beam deposited Ti/Au (300Å/1200Å) interconnection contacts was employed

on the MOS-HEMT diodes. Figure 4-1 shows a schematic (top) and photograph (bottom)

of the completed device. The devices were bonded to electrical feed-through and exposed

to different gas ambient in an environmental chamber [Kim03a, Kim03b, Kim03c].

42

0.0 0.5 1.0 1.5 2.0 2.5 3.0

0

5

10

15

20

Cur

rent

(mA

)

Biased Voltage(V)

R30 10% H2 R10 N2 R30 10% H2 R10 N2

Figure 4-2. Forward I-V characteristics of MOS-HEMT based diode sensors of two different dimensions at 25°C measured under pure N2 or 10%H2 /90%N2 ambient ;R30-diode with 30 µm radius, R10 - diode with 10 µm radius.

Figure 4-2 shows the forward current-voltage (I-V) characteristics at 25 °C of the

MOS-HEMT diode both in pure N2 and in a 10%H2 / 90%H2 atmosphere. At a given

forward bias, the current increases upon introduction of the H2, through a lowering of the

effective barrier height. The H2 catalytically decomposes on the Pt metallization and

diffuses rapidly though the underlying oxide to the interface where it forms a dipole layer

[Eic03].

At 2.5V forward bias the change in forward current upon introduction of the

hydrogen into the ambient is ~6mA or equivalently 0.4V at a fixed current of 10mA. This

is roughly double the detection sensitivity of comparable GaN Schottky gas sensors

tested under the same conditions [Kim03c], confirming that the MOS-HEMT based diode

43

has advantages for applications requiring the ability to detect combustion gases even at

room temperature.

0 50 100 150 2006.7

6.8

6.9

7.0

7.1

7.2

10sec 20sec 30sec

Cur

rent

(mA

)

Time(sec)

Figure 4-3. Time response at 25°C of MOS-HEMT based diode forward current at a fixed bias of 2V when switching the ambient from N2 to 10%H2 /90%N2 for periods of 10, 20 or 30 seconds and then back to pure N2

As the detection temperature is increased, the response of the MOS-HEMT diodes

increases due to more efficient cracking of the hydrogen on the metal contact. The

threshold voltage for a MOSFET is given by [Shu90]

5.0)4

(2i

SBDBFBT C

eNVV

εφφ ++= (4-1)

where, VFB is the voltage required for flat band conditions, фB the barrier height, e the

electronic charge and Ci the Sc2O3 capacitance per unit area.

In analogy with results for MOS gas sensors in other materials systems [Ste03a],

the effect of the introduction of the atomic hydrogen into the oxide is to create a dipole

layer at the oxide/semiconductor interface that will screen some of the piezo-induced

44

charge in the HEMT channel. To test the time response of the MOS diode sensors, the

10%H2/90%N2 ambient was switched into the chamber through a mass flow controller for

periods of 10, 20 or 30 seconds and then switched back to pure N2.

0 50 100 150 200 250 30010.2

10.4

10.6

Current(mA)

Time(sec)

0 50 100 150 200 250 30010.2

10.4

10.6

10.8

Current(mA)

Time(sec)

Figure 4-4. Time response at 25°C of MOS-HEMT based diode forward current at a fixed bias of 2V for three cycles of switching the ambient from N2 to 10%H2 /90%N2 for periods of 10 (top) or 30 (bottom) seconds and then back to pure N2

45

Figure 4-3 shows the time dependence of forward current at a fixed bias of 2V

under these conditions. The response of the sensor is rapid(<1 sec), with saturation

taking almost the full 30 seconds. Upon switching out of the hydrogen–containing

ambient, the forward current decays exponentially back to its initial value. This time

constant is determined by the volume of the test chamber and the flow rate of the input

gases and is not limited by the response of the MOS diode itself. Figure 4-4 shows the

time response of the forward current at fixed bias to a series of gas injections into the

chamber, of duration 10secs each (top) or 30 secs each (bottom). The MOS diode shows

good repeatability in its changes of current and the ability to cycle this current in

response to repeated introductions of hydrogen into the ambient. Once again, the

response appears to be limited by the mass transport of gas into and out of the chamber

and not to the diffusion of hydrogen through the Pt/Sc2O3 stack.

In conclusion, AlGaN/GaN MOS-HEMT diodes appear well-suited to combustion

gas sensing applications. The changes in forward current are approximately double those

of simple GaN Schottky diode gas sensors tested under similar conditions and suggest

that integrated chips involving gas sensors and HEMT-based circuitry for off-chip

communication are feasible in the AlGaN/GaN system.

4.3 Hydrogen Induced Reversible Changes in Drain Current in Sc2O3/AlGaN/GaN HEMTs

It has been observed that AlGaN/GaN metal oxide semiconductor (MOS) diodes

utilizing Sc2O3 as the gate dielectric have approximately double the sensitivity for

hydrogen detection than Pt/GaN Schottky diodes [Kan04a]. The use of a true MOS

transistor should be even more effective because of the current gain in the 3-terminal

device. The reversible hydrogen-induced changes in drain-source current were

46

investigated using Sc2O3/AlGaN/GaN High Electron Mobility Transistors (HEMTs). The

current changes are significantly larger (a factor of 5) than observed for

Sc2O3/AlGaN/GaN MOS diodes exposed under the same conditions. The response time

of the HEMTs is limited by the mass transfer characteristics of the hydrogen-containing

gas ambient into the test chamber. These devices can be used as sensitive combustion gas

sensors, but the results also point out the susceptibility of the HEMTs to changes in

current depending on the composition of the ambient in which they are being operated.

The HEMT layer structures were grown on C-plane Al2O3 substrates by Metal

Organic Chemical Vapor Deposition (MOCVD). The layer structure included an initial

2µm thick undoped GaN buffer followed by a 35nm thick unintentionally doped

Al0.28Ga0.72N layer. The sheet carrier concentration was ~1×1013 cm-2 with a mobility of

980 cm2/V-s at room temperature. Mesa isolation was achieved by using an inductive

coupled plasma system with Ar/Cl2 based discharges. The ohmic contacts was formed by

lift-off of e-beam deposited Ti(200Å)/Al(1000Å)/Pt(400Å)/Au(800Å). The contacts

were annealed at 850 °C for 45 sec under a flowing N2 ambient in a Heatpulse 610T

system. 100Å Sc2O3 was deposited as a gate dielectric through a contact window of SiNx

layer. Before oxide deposition, the wafer was exposed to ozone for 25 minutes. It was

then heat in-situ at 300 °C cleaning for 10mins inside the growth chamber. The Sc2O3

was deposited by rf plasma-activated MBE at 100 °C using elemental Sc evaporated from

a standard effusion all at 1130 °C and O2 derived from an Oxford RF plasma source

[Gil01, Kim00b]. 200 Å Pt Schottky contact was deposited on the top of Sc2O3. Then,

final metal of e-beam deposited Ti/Au (300Å/1200Å) interconnection contacts was

employed on the MOS-HEMTs.

47

Figure 4-5. Photograph of MOS HEMT hydrogen sensor

Figure 4-5 shows photograph (top) and a cross sectional schematic (bottom) of the

completed device. The gate dimension of the device is 1 × 50 µm2. The devices were

bonded to electrical feed-through and exposed to either pure N2 or 10%H2/90% N2

ambients in an environmental chamber in which the gases were introduced through

electronic mass flow controllers.

Figure 4-6 shows the MOS-HEMT drain-source current voltage (IDS-VDS)

characteristics at 25oC measured in both the pure N2 or 10%H2/90% N2 ambients. The

current is measurably larger in the latter case as would be expected if the hydrogen

48

catalytically dissociates on the Pt contact and diffuses through the Sc2O3 to the interface

where it screens some of the piezo-induced channel charge [Eic03].

0 1 2 3 4 5

0

5

10

15

20

25

30

35

40

I DS(

mA

)

VDS(V)

N2(VG=0 to -6V) 10% H2(VG=0 to -6V)

Figure 4-6. IDS-VDS characteristics of MOS-HEMT measured at 25oC under pure N2 ambient or in 10% H2/90%N2 ambient.

This is a clear demonstration of the sensitivity of AlGaN/GaN HEMT dc

characteristics to the presence of hydrogen in the ambient in which they are being

measured. The use of less efficient catalytic metals as the gate metallization would reduce

this sensitivity, but operation at elevated temperatures would increase the effect of the

hydrogen because of more efficient dissociation on the metal contact.

Figure 4-7 shows the measured change in drain-source current for measurement in

the two different ambients as a function of gate voltage for different drain-source

voltages. The maximum change in this current is ~3 mA(150 mA/mm), which is

approximately a factor of 5 larger than obtained for the same bias conditions in

Sc2O3/AlGaN/GaN MOS diodes exposed under the same conditions [Kan04a] and an

49

order of magnitude larger than the changes in forward current in simple Pt/GaN Schottky

diodes in the same test chamber.

-6 -5 -4 -3 -2 -1 0

0.5

1.0

1.5

2.0

2.5

3.0

I DS-I 0(m

A)

VG(V)

VDS=3V VDS=4V VDS=5V

-6 -5 -4 -3 -2 -1 00

20

40

60

80

100

120

140

gm(m

S/m

m)

VG(V)

N2 10% H2

VDS = 3V

Figure 4-7. Change in drain-source current for measurement in N2 versus 10%H2 /90%N2 ambient, as a function of gate voltage (top) and corresponding transconductance at a fixed drain-source voltage of 3V.

50

This shows the advantage of using the 3-terminal device structure, with its

attendant current gain. This approach is particularly advantageous when small

concentrations of hydrogen must be detected over a broad range of temperatures. The

MOS-HEMT is more thermally stable than a conventional metal gate HEMT and will be

usable at higher temperatures. The change in drain-source current tracked the device

transconductance, as shown at the bottom of Figure 4-7. The shift in peak

transconductance when hydrogen is present in the ambient is consistent with an increase

in the total channel charge.

Figure 4-8 shows some of the recovery characteristics of the MOS-HEMTs upon

cycling the ambient from N2 to 1%H2/ 99% N2 . While the change in drain-source current

is almost instantaneous (< 1 sec),the recovery back to the N2 ambient value is of the order

of 20 secs.

This is controlled by the mass transport characteristics of the gas out of the test

chamber, as demonstrated by changing the total flow rate upon switching the gas into the

chamber. Given that the current change upon introduction of the hydrogen is rapid, the

effective diffusivity of the atomic hydrogen through the Sc2O3 must be greater than 4

x10-12 cm2/V.s at 25C. Note the complete reversibility of the drain-source current for

repeated cycling of the ambient.

In conclusion, Sc2O3/AlGaN/GaN MOS-HEMTs show a marked sensitivity of their

drain-source current to the presence of hydrogen in the measurement ambient. This effect

is due to the dissociation of the molecular hydrogen on the Pt gate contact, followed by

diffusion of the atomic species to the oxide/semiconductor interface where it changes the

piezo-induced channel charge. The MOS-HEMTs show larger changes in current than

51

their corresponding MOS-diode or Schottky diode counterparts and show promise as

sensitive hydrogen detectors.

0 5 10 15 20 25 30 3516.2

16.3

16.4

16.5

16.6

16.7

16.8

I DS(

mA

)

Time(sec)

7 sec 5 sec 3 sec 1 sec

1% H2 in N2

0 20 40 60 80 100 120 140 16016.2

16.3

16.4

16.5

16.6

16.7

16.81% H2 in N2

I DS(

mA

)

Time(sec) at VG = -3 V

Figure 4-8. Time dependence of drain-source current when switching from N2 to 1%H2 /99% N2 ambient and back again. The top shows different injection times of the H2/N2, while the bottom shows the reversibility of the current change.

52

4.4 Comparison of MOS and Schottky W/Pt-GaN Diodes for Hydrogen Detection

For GaN Schottky diodes, there is evidence that the presence of an interfacial oxide

increases the magnitude of the change in barrier height [Sch02b] and hence the current

flowing at a given bias on the device. In support of that finding, we have previously

observed that AlGaN/GaN metal-oxide semiconductor(MOS) diodes utilizing Sc2O3 as

the gate dielectric have approximately double the sensitivity for hydrogen detection than

Pt/GaN Schottky diodes [Kan04a]. An additional key requirement in some sensor

applications such as long-term spaceflight is the need for very good reliability of the

contact metal on the semiconductor. We have found that W shows little reaction with

GaN to temperatures in excess of 700oC [Col96, Col97, Cao98] and when used as a bi-

layer with Pt, can also provide measurable sensitivity for H2 detection. The response time

of both types of sensor is limited by the mass transfer characteristics of the hydrogen-

containing gas ambient into the test chamber.

Approximately 6µm of n-GaN was grown on sapphire substrates by Metal Organic

Chemical Vapor Deposition. Ohmic contacts was formed by lift-off of Ti/Al/Pt/Au,

annealed at 500°C. In some cases, MOS structures were formed by deposition of 100Å

Sc2O as a gate dielectric through a contact window of SiNx .Before oxide deposition, the

wafer was exposed to ozone for 25 minutes. It was then heat in-situ at 300 °C cleaning

for 10mins inside the growth chamber. The Sc2O3 was deposited by rf plasma-activated

MBE at 100 °C using elemental Sc evaporated from a standard effusion all at 1130 °C

and O2 derived from an Oxford RF plasma source [Gil01, Kim00b]. 200 Å of W was

deposited on both types of samples by sputtering, followed by e-beam evaporation of

150Å of Pt. The Pt Schottky contacts were formed by lift-off.

53

Figure 4-9.Schematic of both the W/Pt Schottky diode (top) and MOS diode (bottom).

Figure 4-10. Photograph of packaged gas sensor

54

0 1 2 3 4 5

0.1

1

10

Cur

rent

(mA

)

Voltage(V) at 300oC

N2 - GaN 10% H2 - GaN N2 - GaN with oxide 10% H2 - GaN with oxide

0 1 2 3 4 5

0.1

1

10

Cur

rent

(mA

)

Voltage(V) at 600 oC

N2 - GaN 10% H2 - GaN N2 - GaN with oxide 10% H2 - GaN with oxide

Figure 4-11.Forward I-V characteristics at 300 oC(top) or 500 oC(bottom) from the Schottky and MOS diodes in pure N2 and 10% H2 /90% N2

Figure 4-9 shows a cross-sectional schematic of the completed devices. The

devices were bonded to electrical feed-through and exposed to either pure N2 or

55

10%H2/90% N2 ambients in an environmental chamber in which the gases were

introduced through electronic mass flow controllers.

0 1 2 3 4 5

0.1

1

10

Cur

rent

(mA

)

Voltage(V)

T25N T25H T100N T100H T200N T200H T300N T300H T400N T400H T500N T500H T600N T600H

0 1 2 3 4 5

0.1

1

10

Cur

rent

(mA

)

Voltage(V)

T25N T25H T100N T100H T200N T200H T300N T300H T400N T400H T500N T500H T600N T600H

Figure 4-12. Measurement temperature dependence of forward I-V characteristics of the Schottky (top) and MOS (bottom) diodes in both pure N2 and 10% H2 /90% N2 .

56

Figure 4-10 shows a photograph of a typical packaged device. Figure 4-11 shows

the forward I-V characteristics of both the Schottky and MOS diodes at both 300(top) and

600 ºC (bottom) measured in both the pure N2 or 10%H2/90% N2 ambients. The current is

measurably larger in the latter case as would be expected if the hydrogen catalytically

dissociates on the Pt contact and diffuses through the W (and the through the Sc2O3 in the

case of the MOS diodes) to the interface where it screens some of the piezo-induced

channel charge [Sve99]. The decrease in effective barrier height was obtained from

fitting the forward I-V characteristics to the relation for the thermionic emission over a

barrier

)exp()exp(. 2*

nkTeV

kTeTAJ b

Fφ

−= (4-2)

where J is the current density, A* is the Richardson’s constant for n-GaN, T the absolute

temperature, e the electronic charge, φb the barrier height, k Boltzmann’s constant , n the

ideality factor and V the applied voltage. From the data, the decrease in φb in the presence

of 10%H2 in the ambient was 30-50mV over the range of temperatures investigated here.

This is a clear demonstration of the sensitivity of GaN diode dc characteristics to the

presence of hydrogen in the ambient in which they are being measured. Operation at

elevated temperatures should increase the effect of the hydrogen because of more

efficient dissociation on the metal contact.

Figure 4-12 shows the forward I-V characteristics over the entire temperature range

investigated, measured in both pure N2 and 10%H2 /90%H2 .The turn-on voltage, VF, of

the diodes, defined as the voltage at which the forward current is 1A.cm-2, and given by

the relation

FONBF

F JRnΦTA

Je

nkTV ⋅++= )ln( 2** (4.3)

57

decreases with temperature for both types of diodes due to the lowering of the effective

energy barrier for electrons

0 100 200 300 400 500 600

0.25

0.50

0.75

1.00

1.25

VT(

V)

Temperature(oC)

GaN GaN with oxide

0 100 200 300 400 500 6000.005

0.010

0.015

0.020

0.025

Ron

(Wcm

2 )

Temperature(oC)

GaN GaN with oxide

Figure 4-13.Temperature dependence of turn-on voltage(top) and on-state resistance(bottom) for the GaN Schottky and MOS diodes.

The on-state resistance ,RON, derived from the relation

58

CSMBON RWsEVR +⋅⋅+⋅⋅= ρµε )/4( 32 (4.4)

where, ε is the GaN permittivity, µ the carrier mobility, S and WS are substrate resistivity

and thickness, and RC is the contact resistance ,increased with measurement temperature

for both types of diodes, as shown in Figure 4-13.The MOS-HEMT is more thermally

stable than a conventional metal gate HEMT and should be usable at higher temperatures.

0 100 200 300 400 500 600

0.0

0.5

1.0

1.5

2.0

Cur

rent

(mA

)

Temperature(oC)

GaN at 3V GaN at 3.5V GaN with oxide at 3V GaN with oxide at 3.5V

Figure 4-14.Change in forward current when measuring in 10% H2 /90% N2 relative to pure N2 at 3 or 3.5 V in both the Schottky and MOS diodes ,as a function of the measurement temperature

Figure 4-14 shows the change in forward current at fixed bias of 3 V for both types

of diodes when 10% H2 is introduced into the N2 ambient, as a function of temperature.

The use of the MOS diode structure provides a much broader temperature window in

which the detected change in current is above 1mA.This clearly an advantage in terms of

using the sensors over a wide range of temperatures without the need for an on-chip

heater.

59

The recovery characteristics of both types of diodes upon cycling the ambient from

N2 to 1%H2/ 99% N2 showed a rapid (< 1 sec) change of current upon introduction of the

hydrogen into the ambient, while the recovery back to the N2 ambient value took of the

order of 20 secs. This is controlled by the mass transport characteristics of the gas out of

the test chamber, as demonstrated by changing the total flow rate upon switching the gas

into the chamber. Given that the current change upon introduction of the hydrogen is

rapid, the effective diffusivity of the atomic hydrogen through the Sc2O3 must be greater

than 4 x10-12 cm2/V.s at 90oC. There was complete reversibility of the current for

repeated cycling of the ambient.

In conclusion, Pt/W/GaN MOS and Schottky diodes show a marked sensitivity of

their forward current to the presence of hydrogen in the measurement ambient. This

effect is due to the dissociation of the molecular hydrogen on the Pt gate contact,

followed by diffusion of the atomic species to the oxide/semiconductor interface where it

changes the piezo-induced channel charge. The MOS-diode shows a wider range of

temperatures in which it shows large changes in current than the Schottky diode

counterparts and shows promise as sensitive hydrogen detectors.

4.5 Detection of C2H4 Using Wide Bandgap Semiconductor Sensors

Currently, there is a strong interest in the development of wide bandgap

semiconductor gas sensors for applications including detection of combustion gases for

fuel leak detection in spacecraft, automobiles and aircraft, fire detectors, exhaust

diagnosis and emissions from industrial processes [Vas98, Sav00, Lol00, Con02, Arb93,

Hun02, Che96, Eke98, Sve99, Hun01, Che98, Tob97, Bar95, Cas98]. Of particular

interest are methods for detecting ethylene (C2H4), which offers problems because of its

strong double bonds and hence the difficulty in dissociating it at modest temperatures.

60

Wide bandgap semiconductors such as GaN and ZnO are capable of operating at much

higher temperatures than more conventional semiconductors such as Si. Diode or field-

effect transistor structures fabricated in these materials are sensitive to gases such as

hydrogen and hydrocarbons [Tom03, Mit03, Wol03, Ju03a, Lin01, Rao00, Mit98,

Cha02]. Ideal sensors have the ability to discriminate between different gases and arrays

that contain different metal oxides (eg.SnO2, ZnO, CuO, WO3) on the same chip can be

used to obtain this result [Tom03]. The gas sensing mechanism suggested include the

desorption of adsorbed surface oxygen and grain boundaries in poly-ZnO [Mit03],

exchange of charges between adsorbed gas species and the ZnO surface leading to

changes in depletion depth [Wol03] and changes in surface or grain boundary conduction

by gas adsorption/desorption [Kim03c]. Another prime focus should be the thermal

stability of the detectors, since they are expected to operate for long periods at elevated

temperature. MOS diode-based sensors have significantly better thermal stability than a

metal-gate structure and also sensitivity than Schottky diodes on GaN. In this work, we

show that both AlGaN/GaN MOS diodes and Pt/ZnO bulk Schottky diodes are capable of

detection of low concentrations(10%) of ethylene at temperatures between 50-300

°C(ZnO) or 25-400°C(GaN). AlGaN/GaN layer structures were grown on C-plane Al2O3

substrates by Metal Organic Chemical Vapor Deposition (MOCVD). The layer structure

included an initial 2µm thick undoped GaN buffer followed by a 35nm thick

unintentionally doped Al0.28Ga0.72N layer. The sheet carrier concentration was ~1×1013

cm-2 with a mobility of 980 cm2/V-s at room temperature. Device isolation was achieved

with 2000 Å plasma enhanced chemical vapor deposited SiNx. The ohmic contacts was

formed by lift-off of e-beam deposited Ti(200Å)/Al(1000Å)/Pt(400Å)/Au(800Å). The

61

contacts were annealed at 850 °C for 45 sec under a flowing N2 ambient in a Heatpulse

610T system. 400Å Sc2O3 was deposited as a gate dielectric through a contact window of

SiNx layer. Before oxide deposition, the wafer was exposed to ozone for 25 minutes. It

was then heat in-situ at 300 °C cleaning for 10mins inside the growth chamber. 100 Å

Sc2O3 was deposited on AlGaN/GaN by rf plasma-activated MBE at 100 °C using

elemental Sc evaporated from a standard effusion all at 1130 °C and O2 derived from an

Oxford RF plasma source [Gil01, Kim00b].

Figure 4-15. Schematic of AlGaN/GaN MOS diode (top) and bulk ZnO Schottky diode structure (bottom)

62

200 Å Pt Schottky contact was deposited on the top of Sc2O3. Then, final metal of

e-beam deposited Ti/Au (300Å/1200Å) interconnection contacts was employed on the

MOS-HEMT diodes. Figure 4-15 (top)shows a schematic of the completed device.

The bulk ZnO crystals from Cermet, Inc. showed electron concentration of 9 ×

1016 cm-3 and the electron mobility of 200 cm2/V.s. at room temperature from van der

Pauw measurements. The back(O-face) of the substrates were deposited with full area Ti

(200 Å)/Al (800 Å)/Pt (400 Å)/Au (800 Å) by e-beam evaporation. After metal

deposition, the samples were annealed in a Heatpulse 610 T system at 200 °C for 1 min

in N2 ambient. The front face was deposited with plasma-enhanced chemical vapor

deposited SiNx at 100°C and windows opened by wet etching so that a thin (20nm) layer

of Pt could be deposited by e-beam evaporation. After final metal of e-beam deposited

0.0 0.5 1.0 1.5 2.0 2.5 3.0

0

2

4

6

8

10

Diode C

urrent(mA)

Diode Voltage(V)

10% C2H4 N2

Measured at 400oC

Figure 4-16. Forward I-V characteristics of MOS-HEMT based diode sensor at 400°C measured under pure N2 or 10% C2H4 /90% N2 ambients

63

Ti/Au (300Å/1200Å) interconnection contacts was deposited, the devices were

bonded to electrical feed-throughs and exposed to different gas ambients in an

environmental chamber while the diode current-voltage (I-V) characteristics were

monitored. Figure 4-15 (bottom) shows a schematic of the completed device. Figure 4-16

shows the forward diode current-voltage (I-V) characteristics at 400°C of the

Pt/Sc2O3/AlGaN/GaN MOS-HEMT diode both in pure N2 and in a 10% C2H4/90%N2

atmosphere.

At a given forward bias, the current increases upon introduction of the C2H4.

Hydrogen either decomposed from C2H4 in the gas phase or chemisorbed on the Pt

Schottky contacts then decomposed and released hydrogen. The hydrogen diffused

rapidly though the Pt metallization and the underlying oxide to the interface where it

forms a dipole layer [Amb03] and lowered the effective barrier height.

Figure 4-17 shows both the change in current at fixed bias (top) and change in

voltage at fixed current (bottom) as a function of temperature for the MOS diodes when

switching from a 100 % N2 ambient to 10% C2H4/90% N2.

As the detection temperature is increased, the response of the MOS-HEMT diodes

increases due to more efficient cracking of the hydrogen on the metal contact. Note that

the changes in both current and voltage are quite large and readily detected.

In analogy with results for MOS gas sensors in other materials systems [Eic03], the

effect of the introduction of the atomic hydrogen into the oxide is to create a dipole layer

at the oxide/semiconductor interface that will screen some of the piezo-induced charge in

the HEMT channel.

64

0 100 200 300 4000

100

200

300

400

Cur

rent

Cha

nge

at 2

.5 V

(uA

)

Temperature(oC)

100 200 300 400

0

30

60

90

120

Vol

tage

cha

nge(

mV

)

Temperature(oC)

5 mA 10 mA 15 mA

Figure 4-17. Change in MOS diode forward current at fixed forward bias of 2.5V(top) or at fixed current(bottom) as a function of temperature for measurement in 100%N2 or 10% C2H4/90% N2.

65

-0.3 0.0 0.3 0.6 0.9 1.2 1.5

0

10

20

30

40

At 50 oC N2 10 % Ethylene 20 % Ethylene

Cur

rent

(mA

)

Voltage(V)

-0.3 0.0 0.3 0.6 0.9 1.2-10

0

10

20

30

40

At 150 oC N2 10 % Ethylene 20 % Ethylene

Cur

rent

(mA

)

Voltage(V)

Figure 4-18. I-V characteristics at 50°C (top) or 150 °C (bottom) of Pt/ZnO diodes measured in different ambients.

The time constant for response of the diodes was determined by the mass transport

characteristics of the gas into the volume of the test chamber and was not limited by the

response of the MOS diode itself.

66

Figure 4-18 shows the I-V characteristics at 50 and 150°C of the Pt/ZnO diode both

in pure N2 and in ambients containing various concentrations of C2H4. At a given forward

or reverse bias, the current increases upon introduction of the C2H4, through a lowering of

the effective barrier height. One of the main mechanisms is once again the catalytic

decomposition of the C2H4 on the Pt metallization, followed by diffusion to the

underlying interface with the ZnO. In conventional semiconductor gas sensors, the

hydrogen forms an interfacial dipole layer that can collapse the Schottky barrier and

produce more ohmic-like behavior for the Pt contact. The recovery of the rectifying

nature of the Pt contact was many orders of magnitude longer than for Pt/GaN or Pt/SiC

diodes measured under the same conditions in the same chamber. For measurements over

the temperature range 50-150°C, the activation energy for recovery of the rectification of

the contact was estimated from the change in forward current at a fixed bias of 1.5V. This

was thermally activated through a relation of the type IF =IO exp(-Ea/kT) with a value for

Ea of ~0.25 eV, comparable for the value of 0.17 eV obtained for the diffusivity of atomic

deuterium in plasma exposed bulk ZnO. This indicates suggests that at least some part of

the change in current upon hydrogen gas exposure is due to in-diffusion of hydrogen

shallow donors that increase the effective doping density in the near-surface region and

reduce the effective barrier height.

The changes in current at fixed bias or bias at fixed current were larger for the ZnO

diodes than for the AlGaN/GaN MOS diodes because of this additional detection

mechanism, as shown in Figure 4-19. Note that the changes in these parameters are

approximately an order of magnitude larger at 150°C. However the ZnO diodes wee not

thermally stable above ~300°C due to direct reaction of the Pt with the ZnO surface.

67

25 50 75 100 125 150

0

3

6

9

12

15

10 % Ethylene at 1.2 V 10 % Ethylene at -0.5 V 20 % Ethylene at 1.2 V 20 % Ethylene at -0.5 V

Cur

rent

cha

nge(

mA

)

Temperature(oC)

25 50 75 100 125 1500

100

200

300

400

10 % Ethylene at 10 mA 10 % Ethylene at 20 mA 20 % Ethylene at 10 mA 20 % Ethylene at 20 mA

Vol

tage

cha

nge(

mV

)

Temperature(oC)

Figure 4-19.Change in current at a fixed bias (top) or change in voltage at fixed current (bottom) as a function of measurement temperature in different percentages of C2H4/N2 ambients

In conclusion, AlGaN/GaN MOS-HEMT diodes and bulk ZnO Schottky diodes

appear well-suited to detection of C2H4. The former have a larger temperature range of

sensitivity, but the absolute changes in voltage or current are larger with the ZnO diodes.

68

The introduction of hydrogen shallow donors into the near-surface region of the ZnO is a

plausible mechanism for the non-recovery of the I-V characteristics at room temperature.

4.6 Hydrogen and Ozone Gas Sensing Using Multiple ZnO Nanorods

ZnO nanowires and nanorods are attracting attention for use in gas, humidity and

ultra-violet (UV) detectors [Wan04a, Kee04]. ZnO is attractive for a broad range of

applications in thin film form [Loo01, Wra99, gor99, Kri02, Lia01, Kuc02, Pea04b], but

the ability to make arrays of nanorods with large surface area which has been

demonstrated with a number of different growth methods has great potential for new

types of sensors that operate with low power requirements [Hua01, Li04, Kin02, Liu03,

Par03a, Ng03, Hu03b, Par03b, Heo02, Nor04, Poo03, Heo03, Wu00, Zhe01, Lyu01,

Zhe03b, Par03c, Yao02, Pan01, Lao03, Sun03]. A large variety of ZnO one-dimensional

structures have been demonstrated [Wan04b] The large surface area of the nanorods and

bio-safe characteristics of ZnO makes them attractive for gas and chemical sensing and

biomedical applications, and the ability to control their nucleation sites makes them

candidates for micro-lasers or memory arrays. To date, most of the work on ZnO nano-

structures has focused on the synthesis methods [Wan04b] and there have been only a

few reports of the electrical characteristics [Lia01, Kuc02, Pea04b, Hua01, Li04]. The

initial reports show a pronounced sensitivity of the nanowire conductivity to ultraviolet

illumination and the presence of oxygen in the measurement ambient [Wan04a, Kee04].

In this chapter, the gas ambient dependence of current-voltage characteristics of

multiple ZnO nanorods prepared by site-selective Molecular Beam Epitaxy (MBE) will

be investigated. These structures can readily detect a few percent of ozone in N2 at room

temperature and are able to detect hydrogen beginning at around 112oC. Over a limited

range of partial pressures of O3(POZONE) in the ambient gas, the conductance G of the

69

sensor at fixed bias voltage decreased according to the relation G=(GO + A(POZONE)0.5)-

1,where A is a constant and GO the resistance in N2.The nanorods begin to show

detectable sensitivity to hydrogen in the measurement ambient at around 112oC.

Figure 4-20. TEM of ZnO nanorod

The growth of the nanorods has been described in detail [Heo02, Nor04].

Discontinuous thin films (~100 Å) of e-beam evaporated Ag were deposited on p-Si

(100) wafers terminated with native oxide. ZnO nanorods were deposited by MBE with a

base pressure of 5x10-8 mbar using high purity (99.9999%)Zn metal and an O3/O2 plasma

discharge as the source chemicals. The Zn pressure was varied between 5x10-7 and 5x10-8

mbar, while the beam pressure of the O3/O2 mixture was varied between 5 x 10 -6 and

5x10-4 mbar .The growth time was ~2 h at 400 ºC. The typical length of the resultant

nanorods was ~ 2 µm, with typical diameters in the range of 15 – 30 nm. Selected area

diffraction patterns showed the nanorods to be single-crystal. Figure 4-20 shows a

transmission electron micrograph of a single ZnO nanorod. The nanorods were heated in

hydrogen at 300oC to ensure they were conducting. They were released from the substrate

by dissolution of the Ag catalyst and then transferred to SiO2-coated Si substrates. E-

beam lithography was used to pattern sputtered Al/Ti/Au electrodes contacting both ends

20 nm

70

of multiple nanorods. The separation of the electrodes was ~3 um. A scanning electron

micrograph of the completed device is shown in Figure 4-21.

Figure 4-21. SEM image of ZnO multiple nanorods (top) and the pattern contacted by Al/Pt/Au electrodes (bottom)

Au wires were bonded to the contact pad for current –voltage (I-V) measurements

performed over the range 25-150 °C in a range of different ambients (,N2,3 % O3 in N2 or

10%H2 in N2).

Figure 4-22 shows the I-V characteristics from the multiple nanorods measured in

either N2 or 10H2/90% N2 ambients at different temperatures. At room temperature there

71

is no detectable change in current but the presence of the hydrogen in the ambient can be

detected beginning at ~112oC.The reversible chemisorption of reactive gases at the

surface of these metal oxides can produce a large and reversible variation in the

conductance of the material [Tom03].

-0.4 -0.2 0.0 0.2 0.4

-8.0x10-8

-4.0x10-8

0.0

4.0x10-8

8.0x10-8

Cur

rent

(A)

Voltage(V)

22oC N2

22oC H2

112oC N2

112oC H2

194oC N2

194oC H2

Figure 4-22. I-V characteristics at different temperatures of ZnO multiple nanorods measured in either N2 or 10 % H2 in N2 ambient

The gas sensing mechanism suggested include the desorption of adsorbed surface

oxygen and grain boundaries in poly-ZnO [Mit03], exchange of charges between

adsorbed gas species and the ZnO surface leading to changes in depletion depth [Mit98]

and changes in surface or grain boundary conduction by gas adsorption/desorption

[Cha02]. The detection mechanism is still not firmly established in these devices and

needs further study. When detecting hydrogen with the same types of rectifiers, we have

observed changes in current consistent with changes in the near -surface doping of the

ZnO, in which hydrogen introduces a shallow donor state. The nanorods also showed a

72

strong photoresponse to above bandgap UV light(366nm).The photoresponse was fast

and indicates that the changes in conductivity due to injection of carriers is bulk-related

and not due to surface effects [Kee04].

0 50 100 150 200

0.0

4.0x10-9

8.0x10-9

1.2x10-8

1.6x10-8

C

urre

nt d

iffer

ence

(A)

Temperature(oC)

Figure 4-23. Change in current measured at 0.1 V for measurement in either N2 or 10H2 in N2 ambients

Figure 4-23 shows the difference in current at fixed bias of 0.1V for measurement

in N2 versus 10%H2/90% N2 as a function of the measurement temperature. The change

in current is still in the nA range at 112oC but increases with temperature and is about

16nA at 200oC.These changes are readily detected by conventional ammeters. However,

the inability to detect hydrogen at room temperature means that an on-chip heater would

still be needed for any practical application of ZnO nanorods for detection of combustion

gases. The nanorods were more sensitive to the presence of ozone in the measurement

ambient. Figure 4-24 shows the room temperature I-V characteristics from the multiple

nanorods measured in pure N2 or 3% O3 in N2.

73

-3 -2 -1 0 1 2 3

-3

-2

-1

0

1

2

3

Cur

rent

(µA

)

Voltage(V)

N2 3% Ozone

Figure 4-24. I-V characteristics at 25C of ZnO multiple nanorods measured in either N2 OR 3% O3 in N2

The changes in current are much larger than for the case of hydrogen detection.

Over a relatively limited range of O3 partial pressures in the N2 ambient, the conductance

increased according to G =(GO + A(PO3)0.5)-1 , where A is a constant and GO the resistance

in pure N2 ambient. This is a similar dependence to the case of CO detection by SnO2

conduction sensors, in which the effective conductance increased as the square root of

CO partial pressure. The gas sensitivity can be calculated from the difference in

conductance in O3-containing ambients, divided by the conductance in pure N2, i.e.(GN2 -

GO)/GN2.At 25 °C ,the gas sensitivity was 18 % for 3%O3 in N2 .

Figure 4-25 shows the time dependence of change in current at a fixed voltage of

1V when switching from back and forth from N2 to 3% O3 in N2 ambients. The recovery

time constant is long(>10 mins) so that the nanorods are best suited to initial detection of

ozone rather than to determining the actual time dependence of change in concentration.

In the latter case a much faster recovery time would be needed. The gas sweep-out times

74

in our test chamber are relatively short (~a few secs) and therefore the long recovery time

is intrinsic to the nanorods.

0 200 400 600 800 10000.804

0.808

0.812

0.816

0.820

0.824

Ozone exposed

Cur

rent

Cha

nge(µA

)

Time(sec)

Current

Figure 4-25.Time dependence of current at 1V bias when switching back and forth from N2 to 3% O3 in N2 ambients

ZnO nanorods appear well-suited to detection of O3. They are sensitive at

temperatures as low as 25°C for percent levels of O3 in N2. The recovery characteristics

are quite slow at room temperature, indicating that the nanorods can be used only for

initial detection of ozone. The nanorods are also sensitive to detection of hydrogen at

more elevated temperatures(>~100oC) but are not sensitive at room temperature .The

ZnO nanorods can be placed on cheap transparent substrates such as glass, making them

attractive for low-cost sensing applications.

75

CHAPTER 5 CHEMICAL SENSOR FOR POLYMERS AND POLAR LIQUIDS

5.1 Introduction

Since the first demonstration of a fluid monitoring sensor based on AlGaN/GaN

hetero structures by Neuberger [Neu01], the application of AlGaN/GaN HEMTs as liquid

sensors has been a subject of intense research. Neuberger et al. have suggested that the

sensing mechanism for chemical absorbates originated from the compensation of the

polarization induced bound surface charge by interaction with polar molecules in the

fluids. The time dependence of changes in source-drain current of gateless HEMTs

exposed to polar liquids (isopropanol, acetone, methanol) with different dipole moments

using GaN/AlGaN hetero-interfaces was also reported. In particular, it was possible to

distinguish liquids with different polarities.

Steinhoff et al. suggested that the native oxide on the nitride surface was

responsible for the pH sensitivity of the response of gateless GaN based heterostructure

transistors to electrolyte solutions [Ste03a]. The linear response of nonmetallized GaN

gate region using different pH valued electrolyte solutions and sensitivity with a

resolution better than 0.05 pH from pH = 2 to pH = 12 were shown. Chaniotakis et al.

[Cha05] showed that the GaN surface interacts selectively with Lewis acids, such as

sulphate (SO42-) and hydroxide (OH-) ions using impedance spectra. It was also shown

that gallium face GaN was considerably reactive with many Lewis bases, from water to

thiols and organic alcohols without any metal oxide and nitrides.

76

A novel metal oxide, ZnO has numerous attractive characteristics for gas and

chemical sensors [Kan05a, Kan05b, Heo04, Loo01, Nor04]. ZnO is also attractive for

forming various types of nanorods, nanowires, and nanotubes [Hua01, Li04, Kee04,

Kin02, Liu03, Par03a, Ng03, Hu03, Par03b, Heo02, He03, Zhe01, Lyu03, Zha03, Pan01,

Lao03]. The large surface area of the nanorods makes them attractive for gas and

chemical sensing, and the ability to control their nucleation sites makes them candidates

for high density sensor arrays.

In this chapter, a study of two different gateless transistors will be given. In Section

5.2, the response of block copolymers on the gate area of AlGaN/GaN HEMT will be

will be discussed. In Section 5.3, the details of pH measurements with single ZnO

nanorods integrated with a micro-channel will be given.

5.2 Gateless AlGaN/GaN HEMT Response to Block Co-Polymers

The epi structure and processing sequence have been described in detail previously

[Meh04]. In brief, The HEMT structure was grown by metal organic chemical vapor

deposition at 1040 °C on c-plane Al2O3 substrates. The layer structure consisted of a low

temperature GaN nucleation layer, a 3 µm undoped GaN buffer and a 30 nm Al0.3Ga0.7N

undoped layer. The sheet carrier density in the channel was ~8×1012 cm-2 with a 300 K

electron mobility of 900 cm2/Vs. The ungated HEMT was fabricated using Cl2/Ar

inductively coupled plasma etching for mesa isolation and lift-off of e-beam deposited

Ti/Al/Pt/Au subsequently annealed at 850 °C for 30 s to lower the Ohmic contact

resistance. Silicon Nitride was used to encapsulate the source/drain regions, with only

the gate region open to allow the polar liquids to reach across the surface. A schematic of

the device is shown in Figure 5-1(top), plan view layout with a cross-sectional view at the

bottom of Figure 5-1. The source-drain current-voltage characteristics were measured at

77

25 °C using an Agilent 4156C parameter analyzer with the gate region exposed either to

air or various concentrations of the block co-polymers and individual polymers

Figure 5-1. Schematic layout of gate HEMT structure (top) and device cross-section (bottom).

. The block copolymers are composed different proportions of the individual

polymers, PS and PEO. The configuration and chemical symbols of the block copolymer

are illustrated in Figure 5-2. The block co-polymers and individual polymers used in this

work were dissolved in the benzyl alcohol.

78

The charges in the two-dimensional (2D) channel of AlGaN/GaN HEMTs are

induced by spontaneous and piezoelectric polarization, which are balanced with positive

charges on the surface.

Figure 5-2. Structure of block co-polymer, composed of different portions of PS and PEO (top) and chemical formula for PS and PEO (bottom).

The induced sheet carrier concentration of undoped Ga-face AlGaN/GaN can be

calculated by following equation [Asb97]:

( ))()()()()()( 20 xExExe

edx

exxn CFb

ds ∆−+⎟⎟

⎠

⎞⎜⎜⎝

⎛−= φ

εεσ (5-1)

where ε0 is the electric permittivity, ε(x) = 9.5-0.5x is the relative permittivity, x is the Al

mole fraction of AlxGa1-xN, dd is the AlGaN layer thickness, eφb is the Schottky barrier of

the gate contact on AlGaN (eфb(x) = 0.84 + 1.3x(eV)), EF is the Fermi level (EF(x) =

79

[9he2ns(x)/16ε0ε(x)√(8m*(x))]2/3+h2ns(x)/4πm*(x)), Ec is the conduction band (Eg(x) =

6.13x+3.42(1-x)-x(1-x)(eV)), and ∆Ec is the conduction band discontinuity between

AlGN and GaN (∆Ec(x) = 0.7[Eg(x)-Eg(0)]). Note that m*(x) is 0.228Therefore the sheet

charge density in the 2D channel of AlGaN/GaN HEMT is extremely sensitive to its

ambient. The adsorption of polar molecules on the surface of GaN affects in the surface

potential and device characteristics. As illustrated in Figure 5-3, nitride HEMT exposed

to different ambients and drain I-V characteristics were affected significantly. At 40 V of

drain bias voltage, drain current reduced for 25 and 50% for devices exposed to PE and

PS solution, respectively.

0 10 20 30 40

0

2

4

6

8

10

I DS(

mA

)

VDS(V)

air PEO PS

Figure 5-3. Drain I-V characteristics for the air, PS and PEO(weight concentration of the polymers are CPEO = 0.0387mg/ml and CPS = 0.3781mg/ml).

Due to large gate dimension (20 × 150 µm2) induced parasitic resistance, the drain

current did not reach the saturation at 40. If the device dimension reduced, larger

changes of drain current should be expected.

80

0 10 20 30 40

0

2

4

6

8

10

I DS(

mA

)

VDS(V)

air solvent C4 C3 C2 C1

0.0 0.2 0.4 0.6 0.8 1.0 1.22.0

2.5

3.0

3.5

4.0

Cur

rent

(mA

)

Concentration of R11(mg/mL) at 20V

Figure 5-4. Drain I-V characteristics for the different concentration of PS-PEO block copolymer (molecular weight: 58.8 kg/mol, 71% of PS and 29% PEO) and concentration of the copolymer solutions are following C1= 1.0917 mg/ml, C2 = 0.7612 mg/ml, C3 = 0.5504 mg/ml, C4 = 0.08734 mg/ml (top). The drain current reduction as a function of copolymer concentration (bottom).

81

The dipole moments of ethylene and styrene monomer are 1.89 and 0.62,

respectively [CRC97]. The dipole moment of ethylene monomer is three times larger

than that of styrene monomer; however, the effect of PEO solutions on drain current of

nitride HEMT is only half of the PS styrene.

0 10 20 30 40

0

2

4

6

8

10

I D

S(m

A)

VDS(V)

air R5 R11

Figure 5-5. Drain IV characteristics of copolymers with different composition(R5 : MW=66.6kg/mol, 53% of PS and 47% of PEO and R11 : MW 58.8 kg/mol, 71% of PS and 29% of PEO).

This could be due to PEO is extremely linear and dipole is along the polymer chain.

The 20 × 150 µm2 gate opening is much larger than the individual monomer in the PEO

chain. Therefore the dipole effect on the device is very locally within the big gate

opening and some of the net surface charges induced by the PEO were cancelled out. In

the case of PS, the polymer chain is very bulky and not very linear, the net net dipole

induce charge may be higher than that of PEO, therefore, larger drain current changes

were observed. If the gate dimension reduces to the size of individual monomer, the

complete opposite results may be obtained.

82

The concentration of the block copolymer also affected the changes of the drain

current, as showed in Figure 5-4. The molecular weight of the copolymer is 58.8kg/mole

and it contains 71% of PS and 29% PEO. The concentration of the copolymer was varied

from 1.0917 mg/ml to 0.08734 mg/ml. As the copolymer concentration increased, the

drain current reduced more. The current reduction verses the copolymer concentration is

not quite linear as showed in Figure 5-4 (bottom), this could be due to the high parasitic

resistance of the devices and the saturation currents were not reached. Copolymer with

similar molecular weight, but different compositions also had significant impact on the

drain IV characteristics. As shown in Figure 5-5, the copolymer with large percentage of

PS reduced more drain current, which was consistent with results in Figure 5-3.

Gateless AlGaN/GaN HEMTs show a strong dependence of source/drain current on

the polarity and concentration of polymer solutions. This suggests the possibility of

functionalizing the surface for application as biosensors, especially given the excellent

stability of the GaN and AlGaN surfaces which should minimize degradation of adsorbed

cells [Eic03].

5.3 pH Measurements with Single ZnO Nanorod Integrated with a Microchannel

The electrical response of ZnO nanorod surfaces to variations of the pH in

electrolyte solutions introduced via an integrated microchannel was analyzed. The ion-

induced changes in surface potential are readily measured as a change in conductance of

the single ZnO nanorods and suggest that these structures are very promising for a wide

variety of sensor applications.

The preparation of the nanorods has been described in detail previously [Heo02] .In

brief, discontinuous Au droplets were used as the catalyst for ZnO nanorods growth and

formed by annealing e-beam evaporated Au thin films (~100 Å) on p-Si (100) wafers at

83

700 °C. ZnO nanorods were deposited by MBE with a base pressure of 5x10-8 mbar

using high purity (99.9999%) Zn metal and an O3/O2 plasma discharge as the source

chemicals. The Zn pressure was varied between 4x10-6 and 2x10-7 mbar, while the beam

pressure of the O3/O2 mixture was varied between 5x10 -6 and 5x10-4 mbar The growth

time was ~2 h at 400 ~ 600 ºC. The typical length of the resultant nanorods was 2 ~ 10

µm, with typical diameters in the range of 30 – 150 nm. Selected area diffraction patterns

showed the nanorods to be single-crystal. They were released from the substrate by

sonication in ethanol and then transferred to SiO2-coated Si substrates. E-beam

lithography was used to pattern sputtered Al/Pt/Au electrodes contacting both ends of a

single nanorod. The separation of the electrodes was ~3.7 µm. Au wires were bonded to

the contact pad for current –voltage (I-V) measurements to be performed .An integrated

microchannel was made from SYLGARD@ 184 polymer from DOW CORNING. After

mixing this silicone elastomer with curing agent using a weight ratio of 10 : 1 and mixing

with for 5 min ,the solution was evacuated for 30 min to remove residual air bubbles. It

was then applied to the already etched Si wafer(channel length, 10-100um) in a cleaned

and degreased container to make a molding pattern. Another vacuum de-airing for 5 min

was used to remove air bubbles, followed by curing for 2 hours at 90 °C. After taking the

sample from the oven, the film was peeled from the bottom of the container. Entry and

exit holes in the ends of the channel were made with a small puncher(diameter less than

1mm) and the film immediately applied to the nanorod sensor. The pH solution was

applied using a syringe auto pipette (2-20ul). A scanning electron micrograph of the

completed device is shown in Figure 5-6.

84

Prior to the pH measurements, we used pH 4, 7, 10 buffer solutions from Fisher

Scientific to calibrate the electrode and the measurements at 25 °C were carried out in the

dark or under UV illumination from 365 nm light using an Agilent 4156C parameter

analyzer to avoid parasitic effects. The pH solution made by the titration method using

HNO3, NaOH and distilled water. The electrode was a conventional Acumet standard

Ag/AgCl electrode.

Figure 5-6. Schematic (top) and scanning electron micrograph (bottom) of ZnO nanorod with integrated microchannel (4µm width)

85

-0.4 -0.2 0.0 0.2 0.4-8.0x10-8

-4.0x10-8

0.0

4.0x10-8

8.0x10-8

1.2x10-7

I DS(A

)

VDS(V)

non UV UV(365nm)

Figure 5-7. I-V characteristics of ZnO nanorod after wire-bonding, measured either with or without UV (365nm) illumination.

Figure 5-7 shows the I-V characteristics from the single ZnO nanorod after wire

bonding, both in the dark and under UV illumination. The nanorods show a very strong

photoresponse. The conductivity is greatly increased as a result of the illumination, as

evidenced by the higher current. No effect was observed for illumination with below

bandgap light. The photoconduction appears predominantly to originate in bulk

conduction processes with only a minor surface trapping component [Kan05e].

The adsorption of polar molecules on the surface of ZnO affects the surface

potential and device characteristics. Figure 5-8(top) shows the current at a bias of 0.5V

as a function of time from nanorods exposed for 60s to a series of solutions whose pH

was varied from 2-12.

The current is significantly reduced upon exposure to these polar liquids as the pH

is increased. The corresponding nanorod conductance during exposure to the solutions is

86

shown at the bottom of Figure 5-8.The data in Figure 3 shows that the HEMT sensor is

sensitive to the concentration of the polar liquid and therefore could be used to

differentiate between liquids into which a small amount of leakage of another substance

has occurred.

0 100 200 300 400 500 6000.0

4.0x10-8

8.0x10-8

1.2x10-7

1.6x10-7

12111098765432

I D

S(A)

Time(sec)

non UV UV(365nm)

0 100 200 300 400 500 6000

50

100

150

200

250

300

12111098765432

Con

duct

ance

(nS

)

Time(sec)

non UV UV(365nm)

Figure 5-8. Change in current (top) or conductance (bottom) with pH (from 2-12) at V = 0.5V.

87

2 3 4 5 6 7 8 9 10 11 12

0

50

100

150

200

250

300

Con

duct

ance

(nS

)

pH

non UV UV(365nm)

Figure 5-9. Relation between pH and conductance of ZnO nanorod either with or without UV (365nm) illumination.

Figure 5-9 shows the conductance of the nanorods in either the dark or during UV

illumination at a bias of 0.5V for different pH values. The nanorods exhibit a linear

change in conductance between pH 2-12 of 8.5nS/pH in the dark and 20nS/pH when

illuminated with UV (365nm) light .The nanorods show stable operation with a resolution

of ~0.1 pH over the entire pH range ,showing the remarkable sensitivity of the HEMT to

relatively small changes in concentration of the liquid.

There is still much to understand about the mechanism of the current reduction in

relation to the adsorption of the polar liquid molecules on the ZnO surface. It is clear that

these molecules are bonded by van der-Waals type interactions and that they screen

surface change that is induced by polarization in the ZnO. Different chemicals are likely

to exhibit degrees of interaction with the ZnO surface. Steinhoff et. al. [Ste03a] found a

linear response to changes in the pH range 2-12 for ungated GaN-based transistor

88

structures and suggested that the native metal oxide on the semiconductor surface is

responsible. ZnO has the advantage that it is already a metal oxide and the nanorods can

be produced cheaply and transferred to any substrate.

In summary, ZnO nanorods show dramatic changes in conductance upon exposure

to polar liquids. Bonding of polar liquid molecules appear to alters the polarization-

induced positive surface change, leading to changes in the effective carrier density and

hence the drain-source current in biased nanorods. This suggests the possibility of

functionalizing the surface for application as biosensors, especially given the excellent

biocompatibility of the ZnO surfaces which should minimize degradation of adsorbed

cells.

89

CHAPTER 6 BIOSENSORS FOR BIOMATERIALS AND CELL GROWTH

6.1 Introduction

Biologically modified field effect transistors (BioFETs), either at conventional or

nano-dimensions, have the potential to directly detect biochemical interactions in

aqueous solutions for a wide variety of biosensing applications [Fri02, Ing05, Fen05,

Kan05f, Ste05]. To enhance the practicality of bioFETs, the device must be sensitive to

biochemical interactions on their surface, the surface must be functionalized to probe

specific biochemical interactions and the device must be stable in aqueous solutions

across a range of pH and salt concentrations. The AlGaN/GaN high electron mobility

transistors (HEMTs) are attractive for these applications, since it forms a high electron

sheet carrier concentration channel induced by piezoelectric polarization of the strained

AlGaN layer [Amb02, Sch01, Eic01, Sch02a, Stu02, Ste03b, Ste05].

The conducting channel of AlGaN/GaN based HEMT is very close to the surface

(< 35 nm) and extremely sensitive to the ambient, which should enhance detection

sensitivity. It has been shown that it is possible to distinguish liquids with different

polarities and to quantitatively measure pH over a broad range [Ste03a]. Steinhoff et al.

[Ste05]. recently reported the recording of cell action potentials from a layer of rat heart

muscle cells cultivated directly on the gate region of a AlGaN/GaN HEMT. The AlGaN

surface is chemically inert in electrolyte solutions, leading to a low linear drift and has a

sensitivity of ~57mV/pH (cf. SiO2 with 32-40 mV/pH). In addition there is the possibility

90

of integration with blue nitride light-emitting diodes for directed cell growth [Yao04] and

off-chip communications circuitry.

Controlled pattern cell growths, which can be used to spatially control the

development of cells and neurons may provide new approaches to the study of surface-

directed growth, intercellular communication, and organogenesis. Self-assembled

monolayer of organic functionalities has been used previously to pattern neuron cell

growth [Spa94, Ste92, Spa89]. UV and blue light have been used in inhibiting plant cell

growth and photo-catalyzed oxidation reactions to kill micro-organisms in the air and

water [Blo97, Gos97, Ono98]. Recent developments in high intensity blue and UV light

emitting diodes provide a potentially improved light source for controlled environment

plant growth applications such as in vitro micropropagation and biologically based

advanced life support for space missions and other application for inhibiting cell growths

[Kat90]. Blue LEDs, but not green or red LEDs, inhibited growth of retinal pigment

epithelial cells, aortic endothelial cells and fibroblasts in vitro [Oha02]. A wavelength of

470 nm blue light was used to inhibit the growth of B16 Melanoma cells, which showed

the potential for the application of UV radiation for cancer treatment.

In this chapter, a detail study for the transistor with chemically modified gate, i.e.

Au deposited gate for halide ions, antigen attached for antibody will be given. In

addition, selective cell growth method using UV LED also will be given. In Section 6.2,

the method of detecting halide ions using Au deposited gate AlGaN/GaN HEMT will be

analyzed. In Section 6.3, electrical detection of immobilized proteins with chemically

modified gate on AlGaN/GaN HEMT will be investigated. In Section 6.4 the effect of

UV illumination for selective area patterned cell growth will be discussed.

91

6.2 Detection of Halide Ions with AlGaN/GaN HEMTs

The most common application for AlGaN/GaN High Electron Mobility Transistors

(HEMTs) is expected to be in broad-band power amplifiers in base station applications

[Tar02, Gre00, Mis02, Zha03c, Lee02a, Bin02, Moo01, Wei03]. The high electron sheet

carrier concentration of nitride HEMTs is induced by piezoelectric polarization of the

strained AlGaN layer and spontaneous polarization is very large in wurtzite III-nitrides

[Neu01, Amb02, Sch01]. This suggests that nitride HEMTs are also excellent candidates

for sensor and piezoelectric-related applications. Gated and gateless HEMT structures

have demonstrated the ability to perform as combustion gas sensors, strain sensors and

also chemical detectors [Sch02a, Stu02, Kan04b, Ste03b, Ste05, Hil96]. In most cases,

the application of some external change in surface conditions changes the piezoelectric-

induced carrier density in the channel of the HEMT, which in turn alters the drain-source

or gate current. Several groups have shown the strong sensitivity of AlGaN/GaN

heterostructures to ions, polar liquids, hydrogen gas and even biological materials

[Kan04c, Pea04a, Ste03b, Ste05, Hil96]. In particular it has been shown that it is

possible to distinguish liquid with different polarities and to quantitatively measure pH

over a broad range [Ste03a]. The HEMT surface is chemically inert in electrolyte

solutions, leading to a low linear drift and has a sensitivity of ~57mV/pH (cf. SiO2 with

32-40 mV/pH). Moreover there is the prospect of ready integration with blue nitride

light-emitting diodes for directed cell growth and off-chip communications circuitry.

In this chapter, we studied the effect of exposing the gate region to halide ions

(NaF and NaCl) both with and without a thin Au layer present on the gate. Changes in

the source-drain current in the tens of micro-Amp range are observed relative to the value

92

measured in air or water. The drain-source current change observed is a logarithmic

function of the concentration and reinforces the notion that AlGaN/GaN HEMT

structures are very promising for a wide variety of chemical sensor applications.

Figure 6-1.Schematic of HEMT structure for detection of halide ions (top) and plan view photomicrograph of completed device using a 20 nm Au film in the gate region (bottom).

gate area exposed(Au:2nm)

ohmic contact metal

final metal

SiNx

10 µm

93

0 20 40 60 80 100 1201.113

1.116

1.119

1.122

1.125

1.128

Cur

rent

(mA

)

Time(sec)

water 10 uM 100 uM 1 mM 10 mM 100 mM 1 M

0 20 40 60 80 100 1201.115

1.120

1.125

1.130

1.135

1.140

1.145

Cur

rent

(mA

)

Time(sec)

water 10 uM 100 uM 1 mM 10 mM 100 mM 1 M

Figure 6-2.Time dependence of current change in Au-gated HEMTs upon exposure to NaF (top) or NaCl (bottom) solutions with different concentration. Water was also used as a reference. In each case, the aqueous solutions were introduced to the HEMT surface at ~20 secs.

94

The HEMT structures consisted of a 3µm thick undoped GaN buffer, 30Å thick

Al0.3Ga0.7N spacer, 220Å thick Si-doped Al0.3Ga0.7N cap layer. The epi-layers were

grown by rf plasma-assisted Molecular Beam Epitaxy on the thick GaN buffers produced

on sapphire substrates by metal organic chemical vapor deposition (MOCVD). Mesa

isolation was performed with an Inductively Coupled Plasma (ICP) etching with Cl2/Ar

based discharges at –90 V dc self-bias, ICP power of 300 W at 2 MHz and a process

pressure of 5 mTorr. 100 × 100 µm2 Ohmic contacts separated with gaps of 10 µm

consisted of e-beam deposited Ti/Al/Pt/Au patterned by lift-off and annealed at 850 ºC,

45 sec under flowing N2. Silicon nitride was used to encapsulate the source/drain regions,

with only the gate region open to allow the polar liquids to across the surface. In some

cases, we deposited a thin (~2 nm) Au film on the gate region, to facilitate chemical

binding of the different solutions we investigated. A schematic cross-section of the

device is shown at the top of Figure 6-1, while a scanning electron micrograph plan view

of the completed device using the Au film on the gate is shown at the bottom of Figure 6-

1. The source-drain current-voltage characteristics were measured at 25°C using an

Agilent 4156C parameter analyzer with the gate region exposed either to air, ~3mm2 of

water, or 10µM-1M solutions of NaF or NaCl. Both DC and AC measurements were

performed, with the latter approach used to prevent side electrochemical reactions.

Figure 6-2 shows the time dependence of current change in Au-gated HEMTs upon

exposure to NaF (top) or NaCl (bottom) solutions with different concentration. Water

was also used as a reference. The addition of the Au gate increased the detection

sensitivity and the consistency for detection of the halide ions. The NaCl leads to the

opposite polarity of current change relative to NaF due to the highly adsorbed Cl- ions,

95

which creates postitive image charge on the Au film for charge neutrality, induces a

higher positive charge on the AlGaN surface, and finally increases the piezo-induced

charge density in the HEMT channel. Past measurements on adsorption of halides onto

Au surfaces have shown Cl- to be rapidly and strongly adsorbed, while F- is only weakly

attached [Ign97, Bon99].

1x10-5 1x10-4 10-3 10-2 10-1 1001.11

1.12

1.13

1.18

1.19

1.20

Cur

rent

(mA

)

Concentration(M)

gateless NaF gateless NaCl Au gate NaF Au gate NaCl

Figure 6-3.Current change as a function of concentration for HEMTs with or without the Au gate

Figure 6-3 shows the current change as a function of the concentration for HEMTs

with or without the Au gate. The presence of the Au gate leads to a logarithmic

dependence of current (measured at an applied bias of 500 mV) on the concentration for

both NaCl and NaF. By sharp contrast, the gateless devices show lower changes, with no

clear dependence on the concentration, suggesting non-specific adsorption of the halide

ions.

96

Another advantage of the Au is that it has a chemical affinity with thiol, forming a

self-assembled monolayer [Kim04, Chu01]. Long chain alkanethiols, adsorb onto the

gold films and form ordered monolayer films due to specific interaction of gold with

sulfur group in thiol [Chu01].

0 20 40 60 80 100

1.115

1.120

1.125

1.130WaterAir

C

urre

nt(m

A)

Time(sec)

thiol modification w/o modification

Figure 6-4.Change in current as a function of time as a Au-gated HEMT is exposed to water. In the case of thiol modification of the gate, no response was observed.

In this study, 1-octadecanethiol (HS(CH2)17CH3 , ODT) was used. The Au film

coated with ODT is hydrophobic due to the hydrophocity of the thiol. The ODT coated

Au gated devices no longer showed a response to the addition of water to the gate region,

indicating that the entire gate area was totally covered with the hydrophobic thiols which

blocked the adsorption of ions from water (Figure 6-4). This is an important first step

towards the goal of using AlGaN/GaN HEMTs as fast, robust electrical-based bio

sensors. Well established thiol chemistry can be employed to immobilize various

97

functional groups on the Au film, which can recognize chemicals and biomaterials such

as DNA, protein, and virus [Zha02, Koh04, Lee02b, Smi03, Cui01].

Figure 6-5.AC measurement of change in current in gateless HEMT exposed to water. The applied bias of 500 mV was modulated at 11 Hz.

Finally to confirm that side electrochemical reactions were not playing a role

during the DC measurements of current, we used AC measurements of current to confirm

the changes observed during introduction of the aqueous solutions. Figure 6-5(left)

shows an example for introduction of water in to the gate region, while the 500mV bias

was modulated at 11Hz. The resulting peak current is plotted at the right of Figure 6-5

and confirms the result obtained from the DC data.

In conclusion, AlGaN/GaN ungated HEMTs show dramatic changes in drain-

source current upon exposure to polar liquids in the gate region. Bonding of polar liquid

molecules appear to alters the polarization-induced positive surface change, leading to

changes in the channel carrier density and hence the drain-source current. The results

show the potential of AlGaN/GaN transistor structures for a variety of chemical and

biological sensing applications.

0 20 40 60-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

Cur

rent

(mA

)

Time(min)

-5 0 5 10 15 20 25 30 35 40 45 50 551.18

1.19

1.20

1.21

1.22WaterAir

Cur

rent

(mA

)

Time(sec)

gateless

98

6.3 Electrical Detection of Immobilized Proteins with Ungated AlGaN/GaN HEMTs

The HEMT surface with receptors to sense the binding of target molecules was

functionalized in aqueous solutions. Conjugation or hydridization events of solution

molecules with the molecules attached to the gate region in the HEMT change the

conductivity of the HEMT channel by changing the charge distribution in the attached

molecules and consequently the charge on the nitride surface. First we treat the AlGaN

with aminopropyl silane (APS). The AlGaN forms a thin oxide layer when exposed to

oxygen, which can be reacted with the APS to functionalize the surface with amine

groups. By covalently attaching the antigen biotin to the surface of the HMET gate

region, the device can be made sensitive to the conjugation of the antibody protein

streptavidin. The immobilization of the protein on these sites was detected as a change in

HEMT drain-source current.

The HEMT structures consisted of a 3 µm thick undoped GaN buffer, 30Å thick

Al0.3Ga0.7N spacer, 220Å thick Si-doped Al0.3Ga0.7N cap layer. The epi-layers were

grown by rf plasma-assisted Molecular Beam Epitaxy on the thick GaN buffers produced

on sapphire substrates by metal organic chemical vapor deposition (MOCVD). Mesa

isolation was performed with an Inductively Coupled Plasma (ICP) etching with Cl2/Ar

based discharges at –90 V dc self-bias, ICP power of 300 W at 2 MHz and a process

pressure of 5 mTorr. 100 × 100 µm2 Ohmic contacts separated with gaps of 10 µm

consisted of e-beam deposited Ti/Al/Pt/Au patterned by lift-off and annealed at 850 ºC,

45 sec under flowing N2. Silicon nitride was used to encapsulate the source/drain

regions, with only the gate region open to allow the liquid solutions to cross the surface.

A schematic cross-section of the device is shown at the bottom of Figure 6-6. The source-

99

drain current-voltage characteristics were measured at 25°C using an Agilent 4156C

parameter analyzer with the gate region exposed. In addition, we used GaN epi films on

sapphire substrates grown in a similar fashion to the HEMT structures as templates for

studying the attachment of the amine groups.

Figure 6-6. Structure of APS (3-Aminopropyl) triethoxysilane (top left), surface functionalization before chemical modification (top right). A schematic diagram of the gateless HEMT whose surface is functionalized by chemical modification in the gate region (bottom).

The GaN thin films on sapphire were treated with 1% APS (3-Aminopropyl)

triethoxysilane in water for 24 hours to functionalize the surface with amine groups and

then washed with ethanol and water. The chemical structure of the APS is shown at the

top of Figure 6-6, along with a schematic of the functionalized surface. To confirm the

modification worked, we first reacted the amine surface with amine-reactive TMR

GaN

AlGaN

gate regionexposed

ohmic contact SiNx

2DEG

Sapphire

100

(Tetramethylrhodamine) and Fluorescein [Dra04, Hic94]. There was a strong signal with

repeating dark lines observed. However, the GaN surface reflects enough light that the

excitation laser light was reflected and read as emission, making accurate characterization

of the surface impossible. The excitation and emission of organic dyes are close together

and we could not find filters sharp enough to block all the excitation.

Figure 6-7. Fluorescent photographs of biotinylated probes on chemically treated GaN surface (left) and non-biotinated surface (right).

To overcome the problem of GaN reflectivity, the organometallic fluorophore

Rubpy was selected because of its large Stokes shift which separates excitation and

emission. Since Rubpy does not have functionalized versions, we used Rubpy-doped

silica nanoparticles labeled with streptavidin to confirm the nitride surface

functionalization. A control test was developed by labeling the fluorescent nanoparticles

with the protein bovine serum albumin (BSA), which as no affinity for biotin or

biotinated surfaces. The amine-functionalized GaN was further reacted with N-

Hydroxysulfosuccinimidobiotin (sulfo-NHS-biotin) and then blocked and exposed to the

Rubpy nanoparticles. Sulfo-NHS-biotin enables simple and efficient biotin labeling of

antibodies, proteins and any other primary amine-containing macromolecules in solution

101

or on a surface. Specific labeling of cell surface proteins is another common application

for these uniquely water-soluble and membrane impermeable reagents. Biotin is a small

naturally occurring vitamin that binds with high affinity to avidin and streptavidin

proteins. Because it is so small (244 Da), biotin can be conjugated to many proteins

without altering their biological activities. The difference in the emission of the

biotinated GaN with streptavidin nanoparticles and biotinated GaN with BSA coated

nanoparticles is large. Non-biotinated surfaces showed almost no emission, as illustrated

in Figure 6-7.

Figure 6-8.SEM micrographs of amine-functionalized GaN surfaces reacted with biotin NHS and then blocked and exposed to either BSA-coated (left) or streptavidin (right) nanoparticles.

When the same area was viewed in the scanning electron microscope (SEM), the

difference in the nanoparticle concentration was significant, as shown in Figure 6-8.

There were still a large number of non-specifically bound nanoparticles after washing.

We attribute this to the adhesion of the negatively charged silica particles with the

positively charged amine groups that did not react with NHS-biotin. Nevertheless, the

102

contrast in concentration is significant, so we believe that we are modifying the surface

successfully.

0 200 400 600 800

1.732

1.736

1.740

1.744

Streptavidineadded

Sulfo NHS Biotin added

I D

S(m

A)

Time(sec)

Figure 6-9. Change in HEMT drain-source current as a result of interaction between Biotin (Sulfo-NHS-Biotin) and streptavidine introduced to the gateless HEMT surface.

To detect the immobilization of proteins on the surface of the HEMT, we used a

similar surface modification procedure, with attachment of amine groups, followed by

exposure to biotin solution and then streptavidin protein. Labeled proteins were purified

from unlabeled proteins using immobilized streptavidin and avidin affinity gels . After

24 hours of APS chemical modification by immersing the device into a 1% aqueous APS

solution and washing with water and ethanol, the HEMT was dipped into a 5mg/2ml

biotin solution at pH 7.4 (sodium phosphate buffer). After 65 seconds the source-drain

current of the nitride HEMT decreased 5 µA due to the bonding between the amine group

and biotin. After waiting for 10 minutes, the nitride HEMT was removed from the biotin

103

solution, followed by a three times rinse with buffer solution and dipped into the 5mg/ml

streptavidin solution around 645 sec. As shown in Figure 6-9, the HEMT showed another

reduction (~4 µA) in drain-source current, suggesting that the protein has been

immobilized on the functionalized surface, which explains that the current in the gate

channel can be depleted by the interaction between biotin and streptavidin molecules on

the gate area. We also found that current change has not been detected for the HEMT

without chemical modification on the gate surface. This change in current is easily

detectable and was reproducible upon cleaning the surface and remeasuring.

In summary, we have shown that through a chemical modification sequence, the

gate region of an AlGaN/GaN HEMT structure can be functionalized for detection of

streptavidin proteins. This electronic detection approach is a significant step towards

integration with microfluidic channels specially designed for the detection of DNA

hybridization.

6.4 Use of 370 nm Light for Selective Area Fibroblast Cell Growth

There is a strong interest in controlled pattern cell growths, which can be used to

spatially control the development of cells and neurons. Such methods may provide new

approaches to the study of surface-directed growth, intercellular communication, and

organogenesis or be used to control the alignment of individual cells with transducer

elements in biosensors and implant. Self-assembled monolayers of organic

functionalities have been used previously to pattern neuron cell growth [Spa94, Ste92,

Spa89].

UV and blue light have been used in inhibiting plant cell growth and photo-

catalyzed oxidation reactions to kill micro-organisms in the air and water [Blo97, Gos97,

Ono98]. Recent developments in high intensity blue and UV light emitting diodes provide

104

a potentially improved light source for controlled environment plant growth applications

such as in vitro micropropagation and biologically based advanced life support for space

missions and other application for inhibiting cell growths [Kat90]. Blue LEDs, but not

green or red LEDs, inhibited growth of retinal pigment epithelial cells, aortic endothelial

cells and fibroblasts in vitro [Oha02]. A wavelength of 470 nm blue light was used to

inhibit the growth of B16 Melanoma cells, which showed the potential for the application

of UV radiation for cancer treatment.

Figure 6-10. Patterns formed on glass slides coated with 50 nm TiO2. Patterns were generated by etching off TiO2 in the patterns; squares with the dimension of 100 × 100 µm2 and circles with diameter of 20 and 40 µm.

In this chapter, we present a novel technique for selective-area patterned cell

growth using UV illumination combining with a TiO2 thin film based reflection coating.

The effects of UV dose and different growth substrates on the fibroblast cell growth were

investigated. Simulations of the reflected 370 nm UV light from the glass slice or TiO2

coated glass slice were performed. Simulation of more complicated TiO2/SiO2 mirror

stacks was also carried out to obtain higher UV light reflectivity to provide even higher

105

resolution patterned growth and to lower the intensity of the light needed to initiate

growth.

300 350 400 450 500

0

20

40

60

80

Rel

ativ

e po

wer

inte

nsity

Wave length(nm)

-1 0 1 2 3 4

0

10

20

30

40

50

Cur

rent

(mA

)

Voltage(V)

Figure 6-11. Optical spectrum of a GaN light emitting diode at an operating current of 4 mA (top). The current-voltage characteristics of the GaN based LED (bottom).

106

The TiO2 film growth on glass slices was performed in a reactive radio frequency

(RF) magnetron sputter deposition system equipped with a load-lock for substrate

exchange. A quartz lamp heater provided for substrate heating up to 750ºC. Two-inch

diameter Ti sputtering targets were used. The target-to-substrate distance was

approximately 15 cm. The base pressure of the deposition system was on the order of 5 ×

10-8 Torr. The substrates were cleaned in trichloroethylene, acetone, and ethanol prior to

mounting on the sample platen with Ag paint. The electronic grade oxygen gas was

provided through a mass flow control valve. The structure of the film was determined

using x-ray diffraction via a Cu K-alpha source.

A 55 nm thick blanket layer of TiO2 was deposited on glass slides for all the TiO2

coated samples. Circular and square patterns on the TiO2 coated samples were generated

with standard photolithography and etched with Ar/Cl2 based discharges in an inductively

coupled plasma system using resist as the etch mask. As shown in Figure 6-10, the TiO2

was readily etched off and created faithful replication of the circular and square patterns

from the mask; square with the dimension of 100 × 100 µm2 and circles with diameter of

40 and 20 µm.

Prior to the cell growth, the glass slides with patterned and without TiO2 films were

cleaned and sterilized with the following procedure: washed with DI water, sprayed with

acetone for 3 mins under a pressure of 40 psi, exposed to O2 plasma in a Techtronic

parallel plate reactive ion etching system for 1 min to remove the carbon residue, dipped

into a 5% HCl solution for 1 min, rinsed with DI water, ultrasonic cleaned for 10 min. in

DI water and sterilization of the samples at 300 °C for 2 hours.

107

0.0 0.2 0.4 0.6 0.8 1.00

20

40

60

80

100

120

glass TiO2/glass

Cel

l Num

ber

(% o

f Con

trol

)

Power Density of UV Illumination (mW/cm2)

Figure 6-12. Effect of UV light on growth of Fibroblast cells on glass slices and TiO2 coated glass slides

Ledtronics high bright GaN based light emitting diodes (LEDs) were used as the

UV exposure source during some of the cell growths. Six of the LEDs were packaged in

a Teflon disk spaced 1 cm from each other. The LEDs were electrically connected

together and clamped with a 500 Ω serial resistor to limit the output power. An HP5550

power supply was used to bias the diodes. The spectral distribution and current-voltage

characteristics of the UV LEDs are shown in Figure 6-11. The UV spectral distribution

of the LED was measured with a Verity optical emission spectrometer and the LED peak

emission wavelength of 382 nm was obtained. The LED turn-on voltage was 1.2 V and

was biased at 3.6 V during some of the cell growths. The power densities of the LEDs

were measured with Newport 840-C optical power meter.

The Chinese hamster lung fibroblast V79 cells (CCL-93) from the American Type

Culture Collection (ATCC Rockville, MD) were used as the test bed for these

experiments. The cells were cultured in Dulbecco's Modified Eagle's Medium (DMEM;

108

Gibco BRL, Grand Island, NY) supplemented with 5% fetal bovine serum(FBS, Sigma)

and 1% of penicillin (Sigma), in a humidified chamber of 5% CO2 and 95% air at 36.5

oC.

After 48 hours incubation, the DMSO enzyme was injected to make the cell float

from the bottom of the flask. The cell suspension was centrifuged for 5 min in 300 earth

gravity force for the separation. 0.125 mL of cell suspension was taken out to count cell

density with a hemocytometer.

300 350 400 450 5000

20

40

60

80

100

glass TiO2/glass

Ref

lect

ance

Wavelength (nm)

Figure 6-13. Simulated reflectivity of 370 nm UV light with different substrates

The effect of UV light on the fibroblast cell growth is shown in Figure 6-12. For

this set of experiments, the cells were grown on petri dishes under constant UV exposure

at different power densities and the dishes were not coated with TiO2. Adjustment of the

power density was achieved by moving the dish away from the LED source and the UV

power density for each location of the dish were measured with Newport 840-C optical

power meter. The number of cells for the UV exposed samples was expressed as a

109

percentage of the number of cells in the dark control in Figure 6-12. The fibroblast cell

growth in blue light was significantly reduced as the light intensity increased. The cell

was barely grown for the power > 0.8 mW/cm2 of UV exposure in this set of

experiments. The effect of UV illumination for the cell growth on the glass slides with

TiO2 coating is also shown in Figure 6-12.

In this experiment, glass slides coated with 55 nm TiO2 were placed in petri dishes

along with glass slices without TiO2 coating. The reference samples without the TiO2

coating showed the same trend as the first experiment, as illustrated in Figure 6-12 for the

control slices. However, the samples with TiO2 coating showed much higher percentages

of cell growth under the same UV exposures. We attribute this phenomenon to the fact

that most of the UV light shining on the TiO2 coated sample is actually reflected back due

to the refractive index difference between TiO2 and glass.

As illustrated in Figure 6-13, we performed simulations of the reflectivities of 382

nm UV light from both a glass slide and a glass slide coated with 55 nm TiO2 using the

TFcal simulator. For the glass slide, there was only around 8% of the 382 nm light

reflected while there was 68% of the 382 nm UV light reflected from the TiO2 coated

glass slide. Therefore the fibroblast cells could grow on the TiO2/glass substrates under

the conditions of UV exposure, while they would not grow on the bare glass slide. We

used this TiO2 coating technique to control the cell growth patterns. As shown in Figure

6-14 (top), a patterned TiO2/glass substrate with 100 × 100 µm2 squares and circles with

diameters of 40 and 20 µm was used for cell growth. No cell growth was observed on the

glass (square and circular dark areas), which received the higher dose of UV illumination.

The cell grew well on the top of TiO2 (gray area), which reflected most of the UV light.

110

As is clearly shown in Figure 6-14 (top), extremely sharp edge definitions were obtained

down to < 4 µm. The gaps between squares are 16, 8, 4 and 2 µm. With this TiO2

coating on glass technique, the cells can be controllably grown on the desired areas or

specific patterns. Same substrate without the UV light exposure the cells grew well on

both glass and TiO2 surface, as illustrated in Figure 6-14 (bottom).

Figure 6-14. Fibroblast cell growth on patterned TiO2 coated glass slide with the UV light illumination (top) and without UV light illumination (bottom).

We have also confirmed the patterned growth under UV illumination with a set of

four patterned TiO2 coated substrates, as illustrated in Figure 6-15. Four glass slices

111

coated with patterned TiO2 were placed in a petri dish. The LEDs are 1 cm above the

petri dish. The estimated power density of the LEDs for the glass slice 1, 2, 3 and 4 are

2, 0.8, 0.8, and 0.05 mW/cm2.

There was no cell growth on slice 1 for both TiO2 and glass area due to high dose

of UV illumination. For Slice, almost no effect of UV was observed on the cell growth

and the cells grew on both TiO2 and glass area. The cell growths on the slice 2 and 3

were very similar to the patterned growth as illustrated in Figure 6-15. The cells were

selectively grown on the TiO2 coated area but not on the glass area. With, this TiO2

coating on glass technique, the cells can be controllably grown on the desired areas or

specific patterns.

Figure 6-15. A set-up of fibroblast cell on patterned TiO2 coated glass slides under different UV intensity illumination.

TiO2/SiO2 dielectric stacks have been widely used in 850 nm vertical cavity surface

emitting lasers(VCSELs). The cavity length of the VCSEL is much shorter that of ridge

112

waveguide lasers. Therefore the reflectivities of mirrors for VCSEL have to be better

than 99.5%. As shown in Figure 6-13, there is a reflectivity of 68% for 55 nm of TiO2

coating used in our cell growth experiment, which is not optimized. If the TiO2/SiO2

dielectric stacks are used in the experiment the effect of the UV light on the cell growth

will be further reduced. With 5 TiO2/SiO2 dielectric stacks, the reflectivity can reach

98% and as shown in Figure 6-16, a reflectivity of >99% can be obtained with 6

TiO2/SiO2 dielectric stacks.

200 300 400 500 600

0

20

40

60

80

100

Ref

lect

ance

Wave length(nm)

Figure 6-16. Simulated reflectivity of 370 nm UV light with six TiO2/SiO2 dielectric mirror stacks

We have studied the effect of 370 nm UV light on fibroblast cell growth. The UV

light inhibited the cell growth. By employing a thin TiO2 film on the glass substrate for

cell growth, most of the UV light was reflected from the substrate and allowed the cells

to grow. Combining the techniques of TiO2 coating, standard photolithographic

patterning and etching of TiO2, produces a novel method for selective area cell growth.

113

A 20 µm resolution of pattern cell growth was achieved. With optimization of TiO2

thickness and number of TiO2/SiO2 stacks, better resolution and less intensity of UV light

for selective area cell growth can be expected.

114

CHAPTER 7 SUMMARY AND FUTURE WORKS

7.1 Pressure Sensor

AlGaN/GaN HEMTs grown on (0001) sapphire were studied in terms of the

changes in the conductance of two dimensional electron gas (2DEG) with changes in

external strain. By using a two terminal non mesa device on a sapphire beam, a simple

beam bending experiment confirmed the predicted strain dependence of the channel

conductance. The changes in the conductance of the channel of Al0.25Ga0.75N/GaN

HEMTs were roughly linear over the range up to 1.4 x 107 Nm-2, with coefficients for

planar devices of + 9.8 x 10-12 S N–1 m2 for tensile stress and – 1.05 x 10 -11 S N-1 m2 for

compressive stress. For mesa-isolated structures, the coefficients were smaller due to the

reduced effect of the AlGaN layer strain, with values of –1.2 x10 –12 S N-1 m2 for tensile

stress and + 1.97 x10 –12 S N-1 m2 for compressive stress.

A Micro membrane pressure sensor used to monitor the differential pressure was

demonstrated with a circular membrane of AlGaN/GaN on a (111) Si substrate by etching

a circular hole in the Si substrate. The conductivity of the channel was measured in terms

of the changes in the conductance between two inter digit contact fingers with changes in

pressure. The dc current-voltage characteristics were measured at 25oC under either

vacuum (10 m Torr) or pressurized (200 psi) conditions. For the application of

compressive (tensile) stress, the changes in the conductance of the channel showed a

linear change of -(+)7.1 ×10-2 mS/bar (7.1 ×10-10 S N-1 m2), which was two orders of

magnitude larger than that of cantilever beam pressure sensors.

115

The circular AlGaN/GaN diaphragms were fabricated with radii of 200 – 600 µm

on Si substrates. The changes in capacitance over a range of applied pressure were linear

and were reversed when vacuum is applied to the diaphragm. The capacitance of the

channel displayed a change of 7.19±0.45×10-3 pF/µm as a function of the radius of the

membrane at fixed pressure of 9.5 bar and exhibited a linear characteristic response

between -0.5 and +1 bar with a sensitivity of 0.86 pF/bar for a 600 µm radius membrane.

The hysteresis was 0.4% in the linear range.

Future efforts in this area should involve different gas environments in the pressure

sensor system. The catalytic Schottky gate between the inter digit fingered source and

drain pressure sensor will make the previous pressure sensor work in different gas

ambients. Additional improvements in AlGaN/ GaN material growth, device design, and

processing techniques may extend the sensitivity range well beyond this.

7.2 Gas Sensor

The characteristics of Sc2O3/AlGaN/GaN metal-oxide semiconductor diodes as

hydrogen gas sensors were demonstrated. At 25°C, a change in forward current of ~ 6mA

at a bias of 2V was obtained in response to a change in ambient from pure N2 to 10% H2/

90% N2. This is approximately double the change in forward current obtained in Pt/GaN

Schottky diodes measured under the same conditions .The mechanism of the change in

forward gate current appears to be formation of a dipole layer at the oxide/GaN interface

that screens some of the piezo-induced channel charge. The MOS-diode response time

was limited by the mass transport of gas into the test chamber and not by the diffusion of

atomic hydrogen through the metal/oxide stack.

Pt contacted AlGaN/GaN HEMTs with Sc2O3 gate dielectrics showed reversible

changes in drain-source current upon exposure to H2-containing ambients, even at room

116

temperature. The changes in current (as high as 3mA for relatively low gate voltage and

drain-source voltage) were approximately an order of magnitude larger than for Pt /GaN

Schottky diodes and a factor of 5 larger than Sc2O3/AlGaN/GaN metal-oxide

semiconductor(MOS) diodes exposed under the same conditions. This showed the

advantage of using a transistor structure in which the gain produces larger current

changes upon exposure to hydrogen-containing ambients. The increase in current was the

result of a decrease in effective barrier height of the MOS gate of 30-50mV at 25C for

10%H2/90%N2 ambients relative to pure N2 and is due to catalytic dissociation of the H2

on the Pt contact, followed by diffusion to the Sc2O3 /AlGaN interface.

The characteristics of Sc2O3/AlGaN/GaN metal-oxide semiconductor (MOS)

diodes and Pt/ZnO Schottky diodes as detectors of C2H4 were analyzed. At 25°C, a

change in forward current of ~ 40 µA at a bias of 2.5V was obtained in response to a

change in ambient from pure N2 to 10% C2H4/ 90% N2. The current changes were almost

linearly proportional to the testing temperature and reached around 400 µA at 400 °C.

The mechanism of the change in forward gate current appears to be formation of a dipole

layer at the oxide/AlGaN interface that screens some of the piezo-induced channel charge

at the AlGaN/GaN interface. The ZnO diodes showed no detectable change in current

when exposed to ethylene at 25°C but exhibited large changes (up to 10 m A) at higher

temperatures. In these diodes the detection mechanism appeared to also involve

introduction of hydrogen donors into the near-surface region of the ZnO, increasing the

effective doping level under the rectifying contact.

ZnO nanorods grown by site selective Molecular Beam Epitaxy (MBE) showed

current-voltage characteristics that were sensitive to the presence of hydrogen or ozone in

117

the measurement ambient for temperatures as low as ~112 °C for H2 or room temperature

for O3. The sensitivity to hydrogen increased sharply with temperature, and multiple

nanorods contacted at both ends by ohmic electrodes showed currents of ~10-8A at 200

°C and a differential current change of ~18% when changing from a pure N2 ambient to

10% H2 in N2. The nanorods were able to detect small concentrations (3% by flow) of O3

in N2, with changes in current of ~10-7 A at 25C.The sensitivity was 18% for O3 at room

temperature.

All devices fabricated for this study employed Pt/Au or Pd/Au Schottky gate

metallization. Different Schottky metals can be tried for detecting different gases

including hydrogen with different sensitivities. Multiple transistors with different gate

metals in one array gas sensor platform may be an extremely sensitive monitor for a

mixture of different gases in ambient condition. The device in the gas sensor system

connected with Au wire bond could not stand more than 700oC in the heating furnace.

For high temperature measurement, extra development in design and fabrication of a

ceramic or quartz based device package will be needed.

7.3 Chemical Sensor

Gateless AlGaN/GaN HEMT structures exhibited large changes in source-drain

current upon exposing the gate region to various block co-polymer solutions. The polar

nature of some of these polymer chains leads to a change of surface charges in the gate

region on the HEMT, producing a change in surface potential at the semiconductor/liquid

interface. Different gate modulation in AlGaN/GaN HEMT was exhibited according to

the different kinds, concentration and structures of polymers based on solvent and air.

Single ZnO nanorods with ohmic contacts at either end exhibited large changes in

current upon exposing the surface region to polar liquids introduced through an integrated

118

microchannel .The polar nature of the electrolyte introduced led to a change of surface

charges on the nanorod, producing a change in surface potential at the semiconductor

/liquid interface. The nanorods exhibited a linear change in conductance between pH 2-

12 of 8.5nS/pH in the dark and 20nS/pH when illuminated with UV (365nm) light .The

nanorods showed stable operation with a resolution of ~0.1 pH over the entire pH range.

Compared with the gas sensor, which exhibited sharp reproducible response

without memory effects, gateless AlGaN/GaN HEMT for polar liquid needed some

settling time to remove residual effect for exact data acquisition. As shown in Chapter 5,

the laminar flowed buffer solutions (pH 2-12) could be easily exposed upon the nanorod

surface in tiny amounts through micro channels using a syringe pump without any

memory effects. Future works relating to exposing polar liquids on the gate area should

involve micro channels, but PDMS adhesion on the device and sealing at the end of

microchannel still should be improved for better performance of micro chemical sensing

devices. Even though the single ZnO nanorod device had higher sensitivity due to its

large surface to volume ratio, it required a number fabrication steps to be ready for

testing. Therefore, pre-aligned multiple ZnO nanorods grown in the MBE by fixing the

initial catalyst position may be more favorable for easy fabrication and multiple

productions in the future.

7.4 Bio Sensor

AlGaN/GaN HEMTs both with and without an Au gate were found to exhibit

significant changes in channel conductance upon exposing the gate region to various

halide ions. The polar nature of the halide ions lead to a change in surface charge in the

gate region of the HEMT, producing a change in surface potential at the

semiconductor/liquid interface. HEMTs with Au-gate electrode not only doubled the

119

sensitivity of the change in channel conductance as compared to gateless HEMT, but also

showed the opposite conductance behavior. When anions adsorbed on the Au, they

produced a counter charge for electrovalence. These anions dragged some counter ions

from the bulk solution or created an image positive charge on the metal for the required

neutrality. The HEMTs can be used as sensors for a range of chemicals through

appropriate modification with thiol chemistry on the Au surface.

Ungated AlGaN/GaN structures were functionalized in the gate region with

aminopropyl silane. This served as a binding layer to the AlGaN surface for attachment

of fluorescent biological probes. Fluorescence microscopy showed that the chemical

treatment creates sites for specific absorption of probes. Biotin was then added to the

functionalized surface to bind with high affinity to streptavidin proteins. The HEMT

drain-source current showed a clear decrease of 4 µA as this protein was introduced to the

surface, showing the promise of this all-electronic detection approach for biological

sensing

The effect of 370 nm UV light on fibroblast cell growth was exhibited. At this

wavelength, the UV light produced a strong inhibition of the cell growth. By employing

a thin TiO2 film on the glass template for cell growth, most of UV light was reflected

from the substrate and allowed the cells to grow. The TiO2 thin film could be patterned

with standard photolithography followed by etching. Employing this technique, a 4 µm

resolution of patterned cell growth was achieved. Based on the simulation results of

using TiO2/SiO2 mirror stacks, 99.5% of UV light can be reflected from the TiO2/SiO2

coated area and even higher resolution and a lower intensity of UV light for selective area

cell growth can be achieved.

120

Some studies of cell adhesion on GaN and AlN are underway. With a coating of

fibronectin below the cells, bio cells can be adhered very well to the III nitride surface

and survive in a culture medium for a couple of days. So far very little is known about the

interaction of living cell tissue with the III nitride surface, and a systematic investigation

of this topic is being performed currently. Eventually, it will be necessary to integrate a

lipid bilayer membrane onto the surface of the gateless HEMT devices in order to

measure single ion currents associated with high sensitivity ion channel detection.

121

APPENDIX A STRAIN CALCULATION OF A CIRCULAR MEMBRANE OF ALGAN/GAN

UNDER DIFFERENTIAL PRESSURE

The proposed pressure sensors will be made of a circular membrane of AlGaN/GaN

on a SiC substrate by etching a circular hole in the substrate. A deflection of the

membrane away from the substrate due to differential pressure on the two sides of the

membrane produces a tensile strain in the membrane. The differential piezoelectric

response of AlGaN and GaN layers creates a space charge which induces 2DEG at the

AlGaN/GaN interface. The concentration of 2DEG is expected to be directly correlated

with the tensile strain in the membrane and hence with the differential pressure. A high

electron mobility transistor at the center of the membrane senses the 2DEG such that a

change in the conductance of the transistor measures the differential pressure. To analyze

the transducer response, we need to calculate the piezoelectric induced polarization in

each film by taking into account of all the stresses in the AlGaN/GaN structure.

A.1 Strains in the Films

The radial strain is given by

εr = (S-D)/D = (2θR – 2Rsinθ)/(2Rsinθ) = θ/sinθ - 1 (A-1)

and the corresponding stress is

σr = [EGaN/ (1- ν)] εr (A-2)

where EGaN is the Youngs modulus and ν is the Poisson’s ratio of the GaN film. The

total tensile force around the edge of the circular membrane is,

T = πDtGaNσr = πD tGaN [EGaN/ (1- ν)] εr (A-3)

122

where tGaN is the film of thickness. The component of T along z direction is balance by

the force on the membrane due to a differential pressure Pi - P0, where Pi and P0 are the

inside and outside pressure respectively. Hence

Tsinθ = (Pi - P0 ) πD2/4 or πDtGaN [EGaN/ (1- ν)] (θ - sin θ) = (Pi - P0 ) πD2/4 (A-4)

which gives

(θ - sin θ) = (Pi - P0 )[(1- ν)D]/(4EGaNtGaN) (A-5)

Equations (A.1) and (A.5) gives the relationship between the radial strain in the

film εr and the differential pressure Pi - P0.

The total stress in a film can be separated into two parts, i.e.

τ = τm + τa (A-6)

where τm is the misfit stress from the adjacent layers and τa is the applied stress due to

bending of the structure. The corresponding strains in each layer is

ε = εm + εa (A-7)

These elastic strains lead to the piezoelectric induced polarization in the layers. A

difference in the polarizations of the adjacent layers results a space charge, which induces

a 2DEG at the AlGaN/GaN interface. To calculate the strain in each layer, we start with

the misfit strain components. Since the misfit strain in an epitaxial layer is calculated

against its relaxed lattice constant, if a(x) is the in-plane lattice constant of AlxGa1-xN, the

misfit strain is given by

ε(x) = [a(x) - a0(x)]/a0(x) (A-8)

where, a0(x) = 3.189 – 0.077x (Å) is the theoretical relaxed lattice constant. Note that ε(x)

> 0, the misfit strain in AlGaN is always tensile. Therefore the misfit strain in GaN is

compressive. From the force balance equation between the two layers,

123

EGaN|ε2|tGaN = EAlGaN|ε(x)|tAlGaN (A-9)

where EGaN and EAlGaN are the Youngs modulus of the films, the compressive strain in

GaN is given by

ε1 = - (EAlGaNtAlGaN/ EGaNtGaN) ε(x) (A-10)

The applied strain in a circular thin film membrane deflected by a differential

pressure from the two sides of the membrane is given by Equations (A-1) and (A-5), i.e.

εr = θ/sin θ – 1, and (θ - sin θ) = (Pi - P0 )[(1- ν)D]/(4EGaNtGaN) (A-11)

Here we only consider GaN as the membrane, we will refine the calculation later to

include the relatively thin AlGaN layer. For small deflection of the membrane where θ <

15 º, the following approximation can be used,

θ - sin θ ≈ θ3/3! (A-12)

Substituting Equation (A-12) into Equation (A-11), we have

θ = [3(1- ν)D(Pi - P0 )/(2EGaNtGaN)]1/3, εr = [3(1- ν)D(Pi - P0 )/(2EGaNtGaN)]1/3 / sin [3(1- ν)D(Pi - P0 )/(2EGaNtGaN)]1/3– 1 (A-13)

On the other hand, if we can measure the deflection of the film θ, Equation (A-11)

calculates the strain εr, which in turn gives the differential pressure.

A.2 Electrical Polarization and Piezoelectric Effects

The total electric polarization P in AlxGa1-xN and GaN consist of the spontaneous

polarization PSP and the piezoelectric polarization PPE, i.e.,

PAlGaN = PSP (AlGaN) + PPE (AlGaN) = - | PSP (AlGaN) | - |eeff (AlGaN)|[ε(x) + εr] (A.14)

where, eeff = e31 - e33 C13/ C33, and

PGaN = PSP (GaN) + PPE (GaN) = - |PSP(GaN)| - |eeff(GaN)|[-(EAlGaNtAlGaN/EGaNtGaN)ε(x)+εr] (A-15)

124

respectively. The difference in the electric polarization is then,

∆P = PAlGaN - PGaN = - |PSP(AlGaN)| - |eeff(AlGaN)|[ε(x)+εr] + |PSP(GaN)|

+ |eeff(GaN)| [-(EAlGaNtAlGaN/ EGaNtGaN)ε(x) + εr] (A-16)

Since the AlGaN/GaN interface usually contains misfit dislocations, the misfit

strain in AlGaN is partially relaxed. To take it into account, ε(x) can be expressed as

[a0(x) – aGaN] rAlGaN /a0(x) = ∆εAlGaN rAlGaN, where rAlGaN is the fraction of un-relaxed

mismatch in AlxGa1-xN layer. To simplify Equation (A-16), we have,

∆P = - |∆PSP| - |eeff|∆εAlGaNrAlGaN - |∆eeff|εr, (A-17)

where, |∆PSP| = | PSP (AlGaN) | - | PSP (GaN) | where | PSP (AlGaN) | > | PSP (GaN) |, |eeff | = |eeff (AlGaN) | + |eeff (GaN) | (EAlGaNtAlGaN/ EGaNtGaN), |∆eeff | = |eeff (AlGaN) | - |eeff (GaN) |.

A.3 2DEG and Conductance of HEMT

The corresponding 2DEG concentration at the AlxGa1-xN/GaN interface is given by

ns(x) = ∆P/e - [ε0ε(x)/tAlGaN e2][eφb(x) + EF(x) – ∆Ec(x)] (A-18)

where ε0 is the electric permitivity, ε(x) = 9.5 – 0.5x is the relative permitivity, eφb(x) =

0.84 + 1.3x (eV) is the Schottky barrier height,

EF(x) = [9 he2ns(x)/16ε0ε(x) √(8m*(x))]2/3 + h2ns(x)/4πm*(x) is the Fermi levels,

and ∆Ec = 0.7[Eg(x) – Eg (0)|, where Eg (x) = 6.13x+3.42(1-x)- x(1-x) (eV).

Note that m*(x) ~ 0.228me. Since EF(x) is a function of ns(x), to further simplify

equation (A-13), we can drop the 1st term in EF(x) for less than 5% error in ns(x), then we

have ns(x) = 1/[1 +(ε0ε(x)/tAlGaNe2)h2/4πm*(x)](1/e)- |∆PSP| - |eeff |∆εAlGaN rAlGaN

- [ε0ε(x)/tAlGaNe][eφb(x) – ∆Ec(x)] - |∆eeff |εr (A-19)

The conductance of the device is given by σ = eµsns, where µs is the mobility of

2DEG. Substituting equation (A.15) for ns, we obtain

125

σ = α( rAlGaN) + β( fGaN ) εr, where (A-20)

α( rAlGaN) = µs1/[1 + (ε0ε(x)/tAlGaN e2)h2/4πm*(x)]- |∆PSP| –|eeff |∆εAlGaN rAlGaN – [ε0ε(x)/tAlGaN e] [eφb(x) – ∆Ec(x)] (A-21)

and β( fGaN ) = µs1/[1 + (ε0ε(x)/tAlGaN e2 )h2/4πm*(x)] |eeff (GaN)| - |eeff (AlGaN)| (A-22)

Finally, this device involves strain in the radial direction, the difference in the

effective piezoelectric coefficients between GaN and AlGaN in the radial direction can be

different from using the published values. However, if the mobility of 2DEG is know, the

difference in effective piezoelectric coefficients between GaN and AlGaN in the radial

direction can be calculated from the measured conductance vs differential pressure of the

devices.

126

LIST OF REFERENCES

Amb99 O. Ambacher, J. Smart, J. R. Shealy, N.G. Weimann, K. Chu, M. Murphy, W. J. Schaff, L.E. Eastman, R. Dimitrov, L. Wittmer, M. Stutzmann, W. Rieger and J. Hilsenbeck, J. Appl. Phys., 85 3222 (1999).

Amb00 O. Ambacher, B. Foutz, J. Smart, J. R. Shealy, N.G. Weimann, K. Chu, M. Murphy, W. J. Schaff, L. Wittmer, L.E. Eastman, R. Dimitrov, A. Mitchell and M. Stutzmann, J. Appl. Phys., 87 334 (2000).

Amb02 O. Ambacher, M. Eickhoff, G. Steinhoff, M. Hermann, L. Gorgens, V. Werss, Proc ECS, 214 27 (2002).

Amb03 O. Ambacher, M. Eickhoff, A. Link, M. Hermann, M. Stutzmann, Bernardini, V. Tilak, and L. F. Eastmann, Phys. Stat. Solidi C 6 1878 (2003).

Arb93 A. Arbab, A. Spetz and I. Lundström, Sens. Actuators B, 15 19 (1993).

Asb97 P. M. Asbeck, E. T. Yu, S. S. Lau, G. J. Sullivan, J. Van Hove, and J. M. Redwing, Electron. Lett. 33 1230 (1997).

Bai89 C. D. Bain, E. B. Roughton, Y. T. Tao, J Evall, G. M. Whiteside, and R. G. Nuzzo, J. Am. Chem. Soc., 111 321 (1989).

Bar95 A. Baranzahi, A. Lloyd Spetz and I. Lundström, Appl. Phys. Lett., 67 3203 (1995).

Bin02 S. C. Binari, P. B. Klein and T. E. Kazior, Proc. IEEE, 90 1048 (2002).

Blo97 S. S. Block, V. P. Seng and D. W. Goswami, J. Solar Energy Eng., 119 85 (1997).

Bon99 M. Boncheva, L. Scheiber, P. Lincoln, H. Vogel and B. Ackerman, Langmuir, 15 4317 (1999).

Bou83 L. Bousse, N. F. DeRooij and P. Bergbeld, IEEE Trans. Electron, Dev., ED-30 1263 (1983).

Cao98 X. A. Cao, S. J. Pearton, F. Ren, and J. R. Lothian, Appl. Phys. Lett. 73 942 (1998).

Cas98 J. B. Casady, A. K. Agarwal, S. Seshadri, R. R. Siergiej, L. B. Rowland, M. F. MacMillan, D. C. Sheridan, P. A. Sanger and C. D. Brandt, Solid-State Electron,. 42 2165 (1998).

127

Cha02 J. F. Chang, H. H. Kuo, I. C. Leu and M. H. Hon, Sensors and Actuators B, 84 258 (2002).

Cha05, N. A. Chaniotakis, Y. Alifragis, A. Gerogakilas, and G. Konstantinidis, Appl. Phys. Lett. 86 164203 (2005).

Che95 W. Q. Chen and S. K. Hark, J. Appl. Phys., 77, 5747 (1995).

Che96 L. Y. Chen, G. W. Hunter, P. G. Neudeck, D. L. Knight, C. C. Liu, and Q. H. Wu, Proceeding of the Third International Symposium on Ceramic Sensors, H. U. Anderson, M. Liu, and N. Yamazoe, Editors, Electrochemical Society Inc. Pennington, NJ, 92-98, (1996).

Che98 L. Chen, G. W. Hunter, and P. G. Neudeck, J. Vac. Sci. Technol A 15, 1228 (1997); J. Vac. Sci. Techonol. A, 16 2890 (1998).

Chu98 S. N. G. Chu, J. Electrochem. Soc. 145, 3621 (1998).

Chu01 T. D. Chung, J. Park, J. Kim, H. Lim, M. J. Choi, J. R. Kim, J. R., S. K. Chang, H. Kim, Analytical Chemistry, 73 3975 (2001).

Chu05 S. N. G. Chu, F. Ren, S. J. Pearton, B. S. Kang, S. Kim, B. P. Gila, C. R. Abernathy, J.-I. Chyi, W. J. Johnson, and J. Lin, presented at TMS Annual Meeting, San Francisco, February (2005).

Col96 M. W. Cole, D. W. Eckart, W. Y. Han, R. L. Pfeffer, T. Monahan, F. Ren, C. Yuan, R. A. Stall, S. J. Pearton, Y. Li, and Y. Lu, J. Appl. Phys., 80 278 (1996).

Col97 M. W. Cole, F. Ren, and S. J. Pearton, J. Electrochem. Soc., 144 L275 (1997).

Con02 E. J. Connolly, G. M. O'Halloran, H. T. M. Pham, P. M. Sarro and P. J. French, Sens. Actuators A, 99 25 (2002).

CRC97 CRC handbook of chemistry and physics, 77th ed., New York, CRC press (1997).

Cui01 Y. Cui, Q. Wei, H. Park and C. M. Lieber, Science, 293 1289 (2001).

Dad94 J. L. Davidson, D. R. Wur, W. P. Kang, D. Kinser and D. V. Kerns, Proc. Symp. Microstructures and Microfabricated Systems, Electrochem.Soc., Pennington, NJ, 94, 30.(1994).

Dav04 S. Davis, T. S. Huang, M. H. Gass, A. J. Papworth, T. B. Joyce, P. R. Chalker, Appl. Phys. Lett., 84 2566 (2004).

Dra04 T J. Drake and Weihong Tan, Applied Spectroscopy, 58 269 (2004).

Eas02 L.F.Eastman and U.K.Mishra, IEEE Spectrum 39, 28 (2002).

128

Ebe96 D. Eberhardt, E. Santos and W. Schmickler, J. Electroanalytical Chemistry, 419 23 (1996).

Eic01 M. Eickhoff, R. Neuberger, G. Steinhoff, O. Ambacher, G. Muller, M. Stutzmann, Phys. Status Solidi B, 228 519 (2001).

Eic03 M. Eickhoff, J. Schalwig, G. Steinhoff, O. Weidmann, L. Gorgens, R. Neuberger, M. Hermann, B. Baur, G. Muller, O. Ambacher, and M. Stutzmann, Phys. Stat. Solidi C, 6 1908 (2003).

Eke98 L. G. Ekedahl, M. Eriksson and I. Lundstrom, Acc. Chem. Res., 31 249 (1998).

Fen05 G. Feng, F. Patolsky and C.M. Lieber, Mat. Res. Soc. Symp., 828 A2.2.1 (2005).

Fri02 J. Fritz, E. B. Cooper, S. Gaudet, P. K. Sorger and S. R. Manalis, Proc. National Academy of Sciences of the United States, 99 14142 (2002).

Fun86 C. D. Fung, P. W. Cheng and W. H. Ho, IEEE Trans. Electron, Dev., ED-33 8 (1986).

Gil01 B.P. Gila, J.W. Johnson, R. Mehandru, B. Luo, A.H. Onstine, V. Krishnamoorthy, S. Bates, C.R. Abernathy, F. Ren, S.J. Pearton, Phys. Stat. Solid A, 188 239 (2001).

Gor99 C.R. Gorla, N.W. Emanetoglu, S. Liang, W.E. Mayo, Y. Lu, M. Wraback and H. Shen, J. Appl. Phys. 85 2595 (1999).

Gos97 D. Y. Goseami, D. M. Trivedi and S. S. Block, Trans, ASME., 119 92(1997).

Gre00 B. M. Green, K. K. Chu, E. M. Chumbes, J. A. Smart, J. R. Shealy, and L. F. Eastman, IEEE Electron. Dev. Lett., 21 268 (2000).

Gur98 A. Gurlo, N. Barsan, M. Ivanovskaya, U. Weimar, and W. Gopel, Sens. Actuators B 47 92 (1998).

Har95 H. L. Hartnagel, A. L. Dawar, A. K. Jain and C. Jagadish, Semiconducting Transparent Thin Films, IOP Publishing, Bristol (1995).

He03 M. He, M. M. E. Fahmi, S. Noor Mohammad, R. N. Jacobs, L. Salamanca-Riba, F. Felt, M. Jah, A. Sharma, D. Lakins, Appl. Phys. Lett., 82 3749 (2003).

Hen95 G. S. Henshaw, D. H. Dawson, and D. E. Williams, J. Mater. Chem. 5 1791 (1995).

Heo02 Y. W. Heo, V. Varadarjan, M. Kaufman, K. Kim, D. P. Norton, F. Ren and P. H. Fleming, Appl. Phys. Lett., 81 3046 (2002).

Heo04 Y. W. Heo, D. P. Norton, L. C. Tien, Y. Kwon, B. S. Kang, F. Ren, S. J.Pearton, J. R. LaRoche, Mat.Sci. Eng., 47 1 (2004).

129

Hic94 J. J. Hickman, S. K. Bhatia, J. N. Quong, P. Shoen, D. A. Stringer, C. J. Pike and C. W. Cotman, J. Vac. Sci. Technol., A12 607(1994).

Hil96 A. C. Hillier, Sunghyun Kim, and Allen J. Bard, J. Phys. Chem, 100 18808 (1996).

Hu03a J. Q. Hu and Y. Bando, Appl. Phys. Lett., 82 14019(2003).

Hu03b Y.Hu, X. Zhou, Q. Han, Q. Cao and Y. Huang, Mat. Sci. Eng.B, 99 41 (2003).

Hua01 M. H. Huang, S, Mao, H. Feick, H. Yan, Y. Wu, H. Kind, E. Weber, R. Russo and P. Yang, Science, 292 1897 (2001).

Hun95 G. W. Hunter, P. G. Neudeck, L. Y. Chen, D. Knight, C. C. Liu, Q. H. Wu, 31st Joint Prop. Conf. San Diego, CA (1995).

Hun01 G. W. Hunter, C. C. Liu, and D. B. Makel, MEMS Handbook, GadelHak, Editor, CRC Press LLC, Boca Raton, FL (2001).

Hun02 G. W. Hunter, P. G. Neudeck, R. S. Okojie, G. M. Beheim, V. Thomas, L. Chen, D. Lukco, C. C. Liu, B. Ward and D. Makel, Proc. ECS 01 212 (2002).

Ign97 A. Ignaczak and J. A. N. Gomes, J. Electroanalytical Chemistry, 420 71 (1997).

Ing05 S. Ingebrandt, Y. Han, M. Reddy Sakkari, R. Stockmann, O. Belinsky and A. Offenhauser, Mat. Res. Soc. Symp., 828 A6.9.1 (2005).

Joh01 J. W. Johnson, A.G. Baca, R. D. Briggs, R. J. Shul, C. Monier, F. Ren, S. J. Pearton, A. M. Dabiran, A. M. Wowchack, C. J. Polley and P. P. Chow, Solid-State Electron., 45, 1979 (2001).

Joh02 J. W. Johnson, A. P. Zhang, W.B. Luo, F. Ren, S. J. Pearton, S. S. Park, Y. J. Park and J.I. Chyi, IEEE Trans Electron. Dev. 49 32 (2002).

Kad92 M. Kadota, T. Kasanami, and M. Miakata, Jpn. J. Appl. Phys. 31 3013 (1992).

Kan93 W. P. Kang and C. K. Kim, Sensors Actuators B, 14, 682 (1993).

Kan03 B. S. Kang, S. Kim, J. Kim, F. Ren, K. Baik, S. J. Pearton, B. Gila, C.R. Abernathy, C. Pan, G. Chen, J. I. Chyi, V. Chandrsekaran, M. Sheplak, T. Nishida and S.N.G. Chu, Appl. Phys. Lett., 83 4845 (2003).

Kan04a B. S. Kang, F. Ren, B. P. Gila, C. R. Abernathy and S. J. Pearton, Appl.Phys.Lett. 84 1123 (2004).

Kan04b B. S. Kang, G. Louche, R. S. Duran, Y. Gnanou, S. J. Pearton and F. Ren, Solid-State Electron., 48 851 (2004).

130

Kan04c B. S. Kang, S. Kim, F. Ren, J. W. Johnson, R. J. Therrien, P. Rajagopal, J. C. Roberts, E. L. Piner, K. J. Linthicum, S. N. G. Chu, K. Baik, B. P. Gila, C. R. Abernathy, and S. J. Pearton, Appl. Phys. Lett. 85 2962 (2004).

Kan05a B. S. Kang, F. Ren, B. S. Jeong, Y. W. Kwon, K. H. Baik, D. P. Norton, and S. J. Pearton, J. Vac. Sci. Technol., B23 57 (2005).

Kan05b B. S. Kang, F. Ren, L. Wang, C. Lofton, W. Tan, S. J. Pearton, A. Dabiran, A. Osinsky, and P. P. Chow, Appl. Phys. Lett., 87 023508 (2005).

Kan05c B. S. Kang, F. Ren, M. C. Kang, C. Lofton, Weihong Tan, S. J. Pearton, A. Dabiran, A. Osinsky and P.P. Chow, Appl. Phys. Lett., 86 173502 (2005)

Kan05d B. S. Kang, F. Ren, Y. W. Heo, L. C. Tien, B. P. Gila, D. P. Norton, S. J. Pearton, Appl. Phys. A., A80, 1029 (2005).

Kan05e B. S. Kang, S. Kim, F. Ren, Y. W. Heo, B. P. Gila, C. R. Abernathy, D. P. Norton, S. J. Pearton, Appl. Phys. A., A80, 259 (2005).

Kan05f B. S. Kang, Y. W. Heo, L. C. Tien, F. Ren, D. P. Norton and S. J. Pearton, Appl. Phys. A 80 1029 (2005).

Kat90 C. Kathleen, F. C. Delori and K. Akeo, Current Eye Res., 6 549 (1990).

Kee04 K. Keem, H. Kim, G. T. Kim, J. S. Lee, B. Min, K. Cho, M. Y. Sung and S. Kim, Appl. Phys. Lett., 84 4376 (2004).

Kha01 M. A. Khan, X. Hu, G. Simin,A. Lunev, J. Yang, R. Gaska, M.S. Shur, IEEE Elec. Dev. Lett., 21 63 (2001).

Kim00a Jihyun Kim, F. Ren, B. P. Gila, C. R. Abernathy, S. J. Pearton, Appl. Phys. Lett., 82 739 (2000).

Kim00b Jihyun Kim, R. Mehandru, B. Luo, F. Ren, B. P. Gila, A. H. Onstine, C. R. Abernathy, S. J. Pearton, and Y. Irokawa, Appl. Phys. Lett., 80 4555 (2000).

Kim00c Jihyun Kim, R. Mehandru, B. Luo, F. Ren, B. P. Gila, A. H. Onstine, C. R. Abernathy, S. J. Pearton, and Y. Irokawa, Appl. Phys. Lett., 81 373 (2000).

Kim03a J. Kim, B. Gila, C. R. Abernathy, G. Y. Chung, F. Ren, and S. J. Pearton, Solid State Electronics, 47 1487 (2003).

Kim03b J. Kim, B. Gila, G. Y. Chung, C. R. Abernathy, S. J. Pearton, and F. Ren, Solid State Electronics, 47 1069 (2003).

Kim03c J. Kim, F. Ren, B. Gila, C. R. Abernathy, and S. J. Pearton, Appl. Phys. Lett. 82 739 (2003).

131

Kim04 D. S. Kim, H. J. Park, H. M. Jung, J. K. Shin, P. Choi, J. H. Lee and G. Lim, Jpn. J. Appl. Phys., 43 3855 (2004).

Kin02 B. H. Kind, H. Yan, B. Messer, M. Law and P. Yang, Adv. Mater 14 158 (2002).

Ko99 W. H. Ko and Q. Wang, Sens. Actuators, 75,242 (1999).

Koh04 P. Kohli, C. C. Harrell, Z. Cao, R. Gasparac, W. Tan, C. R. Martin, Science, 305 984 (2004).

Kou02 A. Koudymov, X. Hu, K. Simin, G. Simin, M. Ali, J. Yang, M. A. Khan, IEEE Elec. Dev. Lett., 23 449 (2002)

Kri02 S. Krishnamoorthy, A. A. Iliadis, A. Inumpudi, S. Choopun, R. D. Vispute and T. Venkatesan, Solid-State Electron. 46 1631 (2002).

Kuc02 S. O. Kucheyev, J. E. Bradley, J. S. Williams, C. Jagadish and M. V. Swain, Appl. Phys. Lett., 80 956 (2002).

Lao03 J. Y. Lao, J. Y. Huang, D. Z. Wang, Z. F. Ren: Nano Lett. 3 235 (2003).

Lee02a C. Lee, H. Wang, J. Yang, L. Witkowski, M. Muir, M. Asif Khan and P. Saunier, Electron. Lett., 38, 924 (2002).

Lee02b K. B. Lee, S. J. Park, C. A. Mirkin, J. C. Smith, M, Mrksich, Science, 295 702 (2002).

Li02 Y. B. Li, Y. Bando, T. Sato, and K. Kurashima, Appl. Phys. Lett., 81 144 (2002).

Li04 Q. H. Li, Q. Wan, Y. X. Liang and T. H. Wang, Appl. Phys. Lett., 84 4556 (2004).

Lia01 S. Liang, H. Sheng, Y. Liu, Z. Huo, Y. Lu, and H. Shen, J. Cryst. Growth, 225 110 (2001).

Lin01 Z. Ling, C. Leach and R. Freer, J. European Ceramic Society, 21 1977(2001).

Liu03 C. H. Liu, W. C. Liu, F. C. K. Au, J. X. Ding, C. S. Lee and S.T.Lee, Appl. Phys. Lett., 83 3168 (2003).

Liu04 C. C. Liu, P. J. Hesketh, G. W. Hunter, The Electrochemical Society, Interface, Summer, 22-27 (2004).

Loo01 D.C. Look, Mater. Sci. Eng. B80 383 (2001).

Lou80 L. F. Lou, Appl. Phys. Lett., 36 570 (1980).

Lu01 W. Lu, J. Yang, M. A. Khan and I. Adesida, IEEE Trans. Electron Dev., ED48, 581 (2001).

132

Lun86 I. Lundstrom, M. Armgath, A. Spetz, F. Winquist, Sens. Act., 10 399 (1986).

Lun89 I. Lundstrom, M. Armgath, L. G. Petersson, Phys CRC Crit. Rev. Sol. State. Mater. Sci., 15 201 (1989).

Lut99 B. P. Luther, S. D. Wolter, S. E. Mohney, Sens. Actuators B, 56 164 (1999).

Lyu01 S.C. Lyu, Y. Zhang, H. Ruh, H.J. Lee, H.W. Shim, E.K. Suh, C.J. Lee: Chem. Phys. 104 233 (2001).

Mad97 M. Madou, Fundamentals of Microfabrication, CRC press, Boca Raton, FL (1997).

Mar04 Market Research Report: World Market for Sensors, Business Communications Co., November (2004).

Meh04 R. Mehandru, B. Luo, B. S. Kang, J. Kim, F. Ren, S. J. Pearton, C. C. Fan. G. T. Chen, and J. I Chy, Solid State Electronics, 48 351 (2004).

Mil98 B. G. Miler, T. W. Trautman and R. Wolfenden, J. Am. Chem. Soc., 120 2666 (1998).

Mis02 U. K. Mishra, P. Parikh and Y. F. Wu, Proc. IEEE 90 1022 (2002).

Mit03 K. D. Mitzner, J. Sternhagen and D. W. Galipeau, Sensors and Actuators B, 93 92 (2003).

Mit98 P. Mitra, A. P. Chatterjee and H. S. Maiti, Mater. Lett., 35 33(1998).

Moh95 S. N. Mohammad, A.A. Salvador, H. Morkoc, Proc. IEEE 83 1306 (1995).

Moo01 J. S. Moon, M. Micovic, P. Janke, P. Hashimoto, W. S. Wong, R. D. Widman, L. McCray, A. Kurdoghlian and C. Nguyen, Electron. Lett. 37 528 (2001).

Mor99 H. Morkoc,R. Cingolani and B. Gil, Solid-State Electron.43,1909 (1999).

Ned98 A. A. Ned, R. S. Okojie and A. D. Kurtz, Proc. High Temp. Electronics Conf, 14 257 (1998).

Neu01 R. Neuberger, G. Muller, O. Ambacher, M. Stutzmann, Phys Status Solidi A 185 85 (2001).

Ng03 H. T. Ng, J. Li, M. K. Smith, P. Nguygen, A. Cassell, J. Han and M. Meyyappan, Science, 300 1249 (2003).

Nic03 N. H. Nickel and Fleischer, Physics. Rev. Lett. 90 197402 (2003).

Nor04 D. P. Norton, Y. W. Heo, M. P. Ivill, K. Ip, S. J. Pearton, M. F. Chisholm and T. Steiner, Materials Today, June 34 (2004).

133

Oga04 K. Ogata, K. Koike, T. Tanite, T. Komuro, S. Sasa, M. Inoue and M. Yano, Sensors Actuators B, 100 209 (2004).

Oha02 M. Ohara, Y. Kawashima, O. Katoh and H. Watanabe, Jpn. J. Cancer Res., 95 551 (2002).

Ono98 E. Ono, J. L. Cuello and K. A. Jordan, Life Support & Bio Sphere Sci., 5 403 (1998).

Pal D. W. Palmer from http://www.semiconductors.co.uk.

Pal00 N. Pala, R. Gaska, S. Rumyantsev, M.S. Shur, M. Asif Khan, X. Hu, G. Simin, J. Yang, Electron. Lett., 36 268 (2000).

Pan01 Z. W. Pan, Z. R. Dai, and Z. L. Wnag, Science, 291 1947 (2001).

Par03a W. I. Park, G. C. Yi, J. W. Kim and S.M.Park, Appl. Phys. Lett., 82 4358(2003).

Par03b W. I. Park, G. C. Yi, M. Y. Kim and S. J. Pennycook, Adv.Mater. 15 526 (2003).

Par03c W. I. Park, Y. H. Jun, S. W. Jung, G. Yi: Appl. Phys. Lett. 82 964 (2003).

Pea04a S. J. Pearton, B. S. Kang, S. Kim, F. Ren, B.P. Gila, C.R. Abernathy, J. Lin and S.N.G. Chu, J. Phys: Condensed Matter, 16 R961 (2004).

Pea04b S. J. Pearton, D. P. Norton, K. Ip, Y. W. Heo and T. Steiner, J. Vac. Sci. Technol. B 22 932 (2004).

Pea99 S. J. Pearton, J.C. Zolper, R. J. Shul, and F. Ren, J. Appl. Phys., 86 1 (1999).

Pet79 P. Petrou, R. Singh and D. E. Brodie, Appl. Phys. Lett. 35 930 (1979)

Poo03 P. J. Poole, J. Lefebvre, J. Fraser, Appl. Phys. Lett. 83 2055 (2003).

Rao00 B. B. Rao, Materials Chem.Phys.,64 62(2000).

Ras02 H. Rashmi, A. Kranti, S. Haldar, R.S. Gupta, Microelectron. J., 33, 205 (2002).

Ris94 L. Ristic, Sensor Technology and Devices, Artech House Inc. Boston London (1994).

Roy03 V. A. L. Roy, A. B. Djurisc, W. K. Chan, J. Gao, H. F. Lui, and C. Surya, Appl. Phys. Lett., 83 141 (2003).

Rye87 R. R. Rye, A. J. Ricco, J. Appl. Phys., 52 1084 (1987).

Sav00 S. M. Savage, A. Konstantinov, A. M. Saroukan and C. Harris, Proc. ICSCRM 99 pp. 511-515 (2000).

134

Sch01 J. Schalwig, G. Muller, O. Ambacher, M. Stutzmann, Phys Status Solidi A, 185 39 (2001).

Sch02a J. Schalwig, G. Muller, M. Eickhoff, O. Ambacher, M. Stutzmann. Sens Actuat B, 81 425 (2002).

Sch02b J. Schalwig, G. Muller, U. Karrer, M. Eickhoff, O. Ambacher, M. Stutzmann, L. Gorgens and G. Dollinger, Appl. Phys. Lett. 80 1222 (2002).

Sel04 D. Selvanathan, F. M. Mohammed, A. Tesfayesus, and I. Adesida, J. Vac. Sci. Technol. B, 22 2409 (2004).

Shu90 M. S. Shur, Physics of Semiconductor Devices, Prentice Hall, Eaglewood Cliffs, NJ (1990).

Shu98 M. S. Shur, Solid-State Electron. 42 2131 (1998).

Sim00 G. Simin, X. Hu, N. Ilinskaya, A. Kumar, A. Koudymov, J. Zhang, M. Asif Khan, R. Gaska, M.S. Shur, Electron. Lett., 36 2043 (2000).

Sim02 G. Simin, A. Koudymov, H. Fatima, J. P. Zhang, J. Yang, M. A. Khan, X. Hu, A. Tarakji, R. Gaska, M. S. Shur, IEEE Elec. Dev. Lett., 23 458 (2002).

Siu79 W. M. Siu, R. S. C.Collold, IEEE Trans. Electron, Dev., ED 26 1805 (1979).

Smi03 J. C. Smith, K. B. Lee, Q. Wang, M. G. Finn, J. E. Johnson, M. Mrksich, C. A. Mirkin, Nano Letters, 3 883 (2003).

Spa89 E. P. Spalding and D. J. Cosgrove, Planta, 178 407 (1989).

Spa94 B. J. Spargo, M. A. Testoff, T. B. Nielesen, D. A. Stenger, J. J. Hickman and A. S. Rudolph, Proc. Natl. Acad. Sci., 91 11070 (1994).

Sta98 Identification of Defects in Semiconductors, ed M. Stavola (Academic Press, San Diego, CA 1998); Hydrogen in Semiconductors II, ed. N. Nickel (Academic Press, San Diego,1999).

Spe00 A. Lloyd Spetz, P. Tobias, L. Unéus, H. Svenningstorp, L.-G. Ekedahl and I. Lundström, Sens. Actuators B, 70 67 (2000).

Spe97 A. Lloyd Spetz, A. Baranzahi, P. Tobias, and I. Lundström, Phys. Stat. Sol. A, 162 493 (1997).

Ste03a G. Steinhoff, M Hermann, W. J. Schaff, L. F. Eastmann, M. Stutzmann, and M. Eickhoff, Appl. Phys. Lett., 83 177 (2003).

Ste03b G. Steinoff, O. Purrucker, M. Tanaka, M. Stutzmann and M. Eickoff, Adv. Funct. Mater., 13 841 (2003).

135

Ste05 G. Steinhoff, B. Baur, G. Wrobel, S. Ingebrandt, A. Offenhausser, A. Dadgar, A. Krost, M. Stutzmann, and M. Eickhoff, Appl. Phys. Lett., 86 033901 (2005).

Ste92 D. A. Stenger, J. H. Georger, C. S. Dulcey, J. J. Hickman, A. S. Rudolph, T. B. Nielsen, S. M. Mccort and J. M. Calvert, J. Am, Chem. Soc., 114 8435 (1992).

Sto09 G. G. Stoney, Proc. R. Soc., London, 82 172 (1909).

Str00 B. G. Streetman, Sanjay Banerjee, Solid State Electronic Devices, Fifth Ed. Prentice Hall (2000).

Str03 B. P. Strittmatter, R. A. Beach, G. S. Picus, T. C. McGill, J. Appl. Phys. 94 5958 (2003).

Stu02 M. Stutzmann, G. Steinhoff, M. Eickhoff, O. Ambacher, C. E. Nobel, J. Schalwig, Diamond Rel. Mater. 11 886 (2002).

Stu98 M. Stutzmann, O. Ambacher, H. Angerer, C. Deger, and R. Dimitrov, Diam. Relat. Mater., 7 123(1998).

Sun03 X. W. Sun, S. F. Yu, C. X. Xu, C. Yuen, B. J. Chen, S. Li: Jpn. J. Appl. Phys. 42 L1229 (2003).

Sve99 H. Svenningstorp, P. Tobias, I. Lundström, P. Salomonsson, P. Mårtensson, L.G. Ekedahl and A. L. Spetz, Sens. Actuators B, 57 159 (1999).

Tar02 A. Tarakji, X. Hu, A. Koudymov, G. Simin, J. Yang, M. A. Khan, M.S. Shur, R. Gaska, Solid-State Electronics, 46 1211 (2002).

Tob97 P. Tobias, A. Baranzahi, A. L. Spetz, O. Kordina, E. Janzen and I. Lundstrom, IEEE Electron. Dev. Lett., 18 287 (1997).

Tom03 A. A. Tomchenko, G. P. Harmer, B. T. Marquis and J. W. Allen, Sensors and Actuators B, 93 126 (2003).

Vas98 A. Vasiliev, W. Moritz, V. Fillipov, L. Bartholomäus, A. Terentjev and T. Gabusjan, Sens. Actuators B, 49 133 (1998).

Ver00 P. M. Verghese and D. R. Clarke, J. Appl. Phys., 87 4430 (2000).

Wan04a Q. Wan, Q. H. Li, Y. J. Chen, T. H. Wang, X. L. He,J. P. Li and C. L. Lin, Appl. Phys. Lett., 84 3654 (2004).

Wan04b Z. L. Wang, Materials Today, June, 26 (2004).

Wei03 N. G. Weimann, M. J. Manfra and T. Wachtler, IEEE Electron Dev. Lett., 24 57 (2003).

Wil99 D. E. Williams, Sens. Actuators B 57 1 (1999).

136

Wil05 J. S. Wilson, Sensor Technology Hand Book, Boston Elsevier e-book Newnes Publishing, Inc., (2005).

Wol03 J. Wollenstein, J. A. Plaza, C. Cane,Y. Min, H. Botttner and H. L. Tuller, Sensors and Actuators B, 93 350 (2003).

Wra99 M. Wraback, H. Shen, S. Liang, C.R. Gorla, and Y. Lu, Appl. Phys. Lett. 76 507 (1999).

Wu00 X. C. Wu, W. H. Song, W. D. Huang, M. H. Pu, B. Zhao, Y. P. Sun, J. J. Du: Chem. Phys. Lett., 328 5 (2000).

Wu97 C. H. Wu, S. Stefanescu, H. I. Kuo, C. A. Zorman and M. Mehregany, Proc. Int. Conf. Solid-State Sensors and Actuators, 514 (1997).

Wu04 Y. R. Wu, J. Singh, Appl. Phys. Lett. 85 1223 (2004).

Yal74 D. E. Yales, S. Levine, and T. W. Healy, J. Chem. for Faraday Trans., 70 1807 (1974).

Yao02 B. D. Yao, Y. F. Chan, N. Wang, Appl. Phys. Lett. 81 757 (2002).

Yao04 G. Yao and W. Tan, Anal. Biochem., 331 216 (2004).

You82 M. Yousuf, B. Kuliyev, B. Lalevic, T. L. Poteat, Solid Stat. Electron. 25 753 (1982)

You04 D. J. Young, J. Du, C.A. Zorman and W. H. Ko, IEEE Sensors J.4 464 (2004).

Yun05 F. Yun, S. Chevtchenko, Y. T. Moon, H. Morkoc, T. J. Fawcett, and J. T. Wolan, Appl. Phys. Lett., 87 073507 (2005).

Zha00 A. P. Zhang, G. Dang, F. Ren, J. Han, A. Y. Polyakov, N. B. Smirnov, N. V. Govorkov, J. M. Redwing, X. A. Cao and S. J. Pearton, Appl. Phys. Lett., 76 1767 (2000).

Zha01 A. P. Zhang, G. T. Dang, F. Ren, H. Cho, K. P. Lee, S. J. Pearton, J.-I. Chi, T. E. Nee, C. M. Lee and C. C Chuo, IEEE Trans Electron Dev., 48 407 (2001).

Zha02 H. Zhang, H., Z. Li, C. A. Mirkin, Advanced Materials,14 1472 (2002).

Zha03a A. P. Zhang, L. B. Rowland, E. B. Kaminsky, J. B. Tucker, J. W. Kretchmer, A. F. Allen, J. Cook, and B. J. Edward, Electron. Lett., 39 245 (2003).

Zha03b A.P. Zhang, L.B. Rowland, E.B. Kaminsky, V. Tilak, J.C. Grande, J. Teetsov, A. Vertiatchikh and L. F. Eastman, J.Electron.Mater.32 388 (2003).

Zha03c B. P. Zhang, N. T. Binh, Y. Segawa, K. Wakatsuki, N. Usami: Appl. Phys. Lett. 83 1635 (2003).

137

Zhe01 M. J. Zheng, L. D. Zhang, G. H. Li, X. Y. Zhang, X. F. Wang: Appl. Phys. Lett. 79 839 (2001).

138

BIOGRAPHICAL SKETCH

Byoung Sam Kang was born on August 2nd, 1971, in Gwang Ju, South Korea. He

graduated with a bachelor’s degree in Chemical Engineering from Po Hang Institute of

Science and Technology (POTECH) in December 1995 and then he earned Master of

Science in Chemical Engineering at POSTECH in December 1997.

From 1997 to 2002, he worked at fuel cell research group at Korea Electric Power

Research Institute (KEPRI) in Daejeon, South Korea, as technical staff for fuel cell

system design and analysis. On November 28th, 1998, he married his wife, Jinah and they

had a beautiful daughter named Dahyoung on January 23rd, 2001.

In August 2002, he enrolled at the University of Florida in the Department of

Chemical Engineering. From the spring of 2003, he has pursued a PhD degree under the

guidance of Dr. Fan Ren. His main research involved various sensors using piezoelectric

polarization based wide bandgap electronic devices. He is co-author of approximately 50

journal and conference papers dealing with compound semiconductor sensing devices.

He graduated from the University of Florida with a doctoral degree in Chemical

Engineering in December 2005.