7'

v: . l.‘LR-L“n

... "“"€Il ”a...

sun;

 

.;“.. "1“
~ MN

I 4’35

 

D‘!'::’;‘\) :7
’3 r . 3
li‘ijsJﬂy‘yﬁm If:
Hikﬁgiﬁ” 2 I EII. '3 :1?ng 2 (I
31 213%- J: '39:!» "z - 2 -
"£252! Might} . . . 9.32.5; !
i‘gﬁﬂ,’ J41.“ . . 2 A‘Li‘L—‘E ‘zifr 'izt:

\.
.2

' 2 I , 91:.“
”HM V ‘ 3:42.?

74S
2.212

«,1.
WI}: WW
:6‘1-3. J

22::

2.
h.
h"

319‘ ("‘5‘

‘12: .2432?" ”:23

Nit

 

'2'
‘gy a 2.0.9 ’ 'f ‘1
:: .‘2;.;2.‘v".,‘:.
° Hg." t 't‘ 2 '3'
22.2: “‘42,! '1.“ 2:

r11"

.2:

“a“? {332’ 22mg? =3
um

5‘11”";
.13" .39”.
: cu ;.

, 553
{"495”,
~A-‘V.
gm:

 

v—

Q‘ J“.

-3
5‘
«v
" ‘ 4.3.

 

THESIS
Z.

VESIR SIIBTYL BISRARE

IIIIIIIIII IIII III

 

 

 

 

 

 

 

|IIIIIIIIIIIIIIIIIIIIIIIII

3 12930

 

LIBRARY

Michigan State
University

 

 

 

This is to certify that the

dissertation entitled

TECHNOLOGY OF CVD DIAMOND THERMISTOR AND
THIN FILM HEATER

presented by

Gwo-Shii Yang

has been accepted towards fulﬁllment
of the requirements for

Ph. D. degreeinElectrical Engineering

 

 

WWW

Major professor

 

?/1Y/76

Date

 

MSU is an Afﬁrmative Action/Equal Opportuniry Institution 0-12771

 

._—__._ _,. _ __.— _

PLACE ll RETURN BOX to remove We checkom from your record.
TO AVOID FINES rotum on or before dete due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU le An Nﬂnnettve Action/Bond Opportunlty lnetItqun
Wm:

 

TECHNOLOGY OF CVD DIAMOND THEMISTOR
AND THIN FILM HEATER

By

Gwo-Shii Yang

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Electrical Engineering

1 996

ABSTRACT

TECHNOLOGY OF CVD DIAMOND THEMISTOR
AND THIN FILM HEATER

By

Gwo-Shii Yang

Due to the extended operational environments of modern sensors, it becomes
necessary to investigate new materials for optimal sensing performance, especially under
harsh environments. The unique intrinsic properties of diamond make it an excellent
material for modern sensors.

In the present work, an improved nucleation technique was developed to enhance the

diamond nucleation density up to 1.1 x 1011 cm'2 on Si and oxidized Si wafers. 1 pm thick
ﬁlms were obtained with a mean surface roughness of 30 nm. The ultrahigh nucleation
density technique is helpful to reduce the deposition time for obtaining continuous films.
The temperature sensing properties of diamond were studied on single crystal
substrates and polycrystalline samples with different grain sizes in the temperature range
of 300 - 1000 K. An average difference between temperatures determined by diamond
thermistors and those measured by thermocouples was in the range of 0.3289 - 4.8759%
(1.4270 - 25.8909 K). The sensitivity factor B of diamond thermistors ranged from 350 -
2200 K with a corresponding room temperature resistance of 0.977 - 509.5 KS2. It has
been found that the effect of doping level on the thermistor sensitivity ([3) is less
prominent in the case of polycrystalline ﬁlms. A better linearity of log(p) vs. 1000/T
relation was obtained for polycrystalline samples. The results suggest that polycrystalline

ﬁlms are preferable to single crystal samples for thermistor application. The maximum

heating temperature and powder density of 950 K and 93 watt/cmz, respectively, were
achieved for heating application. Due to its high thermal conductivity, the substrate
temperature was isothermal.

By using a single structure diamond resistor, heat generation and temperature sensing
were achieved simultaneously. The average difference between temperatures determined
by self-sensing of diamond thin ﬁlm heaters and reference temperatures measured by
thermocouples was in the range of 1.05 - 3.6%. The single structure diamond sensor/

heater device is reported for the ﬁrst time.

Copyright by
Gwo-Shii Yang
1996

to my family

ACKNOWLEDGMENTS

I would like to express my deep gratitude to my advisor, Prof. D. M. Aslam, for his
guidance and support throughout the present work. I am grateful for the opportunity to
learn from his example as both a research and an engineer.

I would like to thank Prof. D. Reinhard, Prof. J. Asmussen, and K. P. Kuo for their
supply of specimen in the study of nucleation density effects. I also would like to thank
Prof. J. V. Beck, Prof. J. J. McGrath, and M. White for their expertise in thermodynamics
and optical temperature measurement. Many thanks go to Prof. T. Grotjohn and J.
Mossbrucker for the assistance in Raman analysis and IL. Hoffman, Prof. B. Golding, and
A. Engebretson for their assistance in metal evaporation.

I would also like to thank the members of my research group, S. Sahli, D. Hong, I.
Taher, Dr. Kwon, and U. S. Kim, for their assistance and helpful discussion.

Last but not least, I would like to thank my parents for providing me with the
opportunity to complete this study. A special thank goes to my wife, Yuh-pyng, for
everything I can possibly think of.

TABLE OF CONTENTS

LIST OF TABLES x
LIST OF FIGURES xi
1 RESEARCH MOTIVATION AND GOALS .................. 1
1.1 Introduction ................................ 1

1.2 Objective of This Work .......................... 3

1.3 Dissertation Organization ......................... 6

2 BACKGROUND ................................. 7
2.1 Introduction ................................ 7

2.2 Temperature Sensors ............................ 7
2.2.1 Liquid-in-Glass Thermometers ................... 9

2.2.2 Resistance Therrnometry ...................... 10

2.2.2.1 Resistance Temperature Detector ............. 10

2.2.2.2 Semiconductor Therrnistor ................ 11

2.2.3 Thermoelectric Thermometry ................... 17

2.2.4 Optical Pyrometry ......................... 17

2.3 Thick/Thin Film Heaters .......................... 20

2.4 Diamond Properties ............................ 21

2.5 CVD Diamond Technology ........................ 25
2.5.1 Nucleation ............................. 25

2.5.2 Deposition Methods ........................ 27

2.5.3 Diamond film Structure Characterization ............. 32

2.5.4 Patterning .............................. 35

2.5.5 Doping ............................... 36

2.5.6 Metallization ............................ 37

vii

2.6 Diamond Sensors .............................. 38
2.6.1 Diamond Temperature Sensors .................. 38

2.6.2 Radiation Sensors ......................... 41

2.6.3 Piezoresistive Sensors ....................... 42

2.6.4 Gas Sensor ............................. 44

2.7 Summary .................................. 44
DIAMOND DEVICE INFRA-STRUCTURE --------------- 45
3. 1 Introduction ................................ 45
3.2 Thin Film Deposition Technology ..................... 45
3.2.1 Film Growth ............................ 46

3.2.1.1 Hot Filament Chemical Vapour Deposition System . . . 46

3.2.1.2 Film Quality Analysis .................. 50

3.2.2 Microwave Plasma Assisted Chemical Vapor Deposition Reactor 50

3.3 Nucleation Enhancement Technology ................... 56
3.3.1 Diamond Powder Loaded Fluids (DPLF) Method ......... 59

3.3.2 Ultrahigh Nucleation Density ................... 59

3.3.3 Effects of Nucleation Density ................... 61

3.4 Diamond Device Fabrication Technologies ................ 73
3.4.1 Patterning of Diamond Films ................... 73

3.4.1.1 Photolithographic Methods ................ 74

3.4.1.2 Direct-write Patterning Method ............. 77

3.4.1.3 Construction of Direct-write Seeding/Patteming System 77

3.4.2 Doping ............................... 82

3.4.3 Metallization ............................ 84

3.5 Summary .................................. 85

SINGLE ELEMENT DIAMOND SENSOR/HEATER STRUCTURE . . . . 89

4.1
4.2

4.3

Introduction ................................ 89
Experimental ................................ 90
4.2.1 Sample Descriptions ........................ 90
4.2.2 Measurement Setups ........................ 94
Temperature Sensors ........................... 102
4.3.1 Contact Resistance ........................ 102

viii

4.3.2 Effect of High Temperature Annealing on Resistivity ...... 102

4.3.3 Effects of Crystal Structure and Grain Size on Resistivity . . . 104

4.3.4 Temperature Dependence of Resistivity ............. 108

4.3.5 Temperature Sensing ....................... 112

4.3.6 Discussions ............................ 119

4.4 Thin Film Heaters ............................ 128
4.4.1 Heating Temperature and Power Density ............ 129

4.4.2 Temperature Proﬁle of Diamond Thin Film Heaters ...... 133

4.5 Heater/Sensor Structures ......................... 135
4.5.1 Diamond Resistors for Temperature Sensing .......... 139

4.5.2 Self-sensing Diamond Heater .................. 139

4.6 Summary ................................. 139

5 CONCLUSIONS AND FUTURE RESEARCH ................ 145
5. 1 Introduction ............................... 145

5.2 Summary of Contributions ........................ 145

5.3 Future Research ............................. 147
BIBLIOGRAPHY 148

ix

2.1
2.2
2.3
2.4
2.5

2.6

3.1
3.2
3.3

4. 1
4.2
4.3
4.4
4.5
4.6

4.7

LIST OF TABLES

Material properties used to measure temperature ................. 8
The characteristics of resistance temperature detectors ............. 12
Summary of important parameters of thermocouples .............. 18
Properties of semiconductor materials ...................... 23
Deposition Parameters Used in the Growth of Diamond Films by Hot Filament
Assisted CVD ................................... 29
Deposition Parameters used in the Growth of Diamond Films by Plasma enhanced
CVD ....................................... 31
Technical details of DPLF seeding processes .................. 60
Sample preparation parameters. ......................... 62
Deposition parameters of HFCVD and MPCVD reactors ............ 65
List of important process parameters of test structures .............. 91
Details of measurement setups .......................... 95
The effect of thermal annealing on resistance ................. 105
Comparison of resistivity for samples deposited at the same batch ....... 107
The calibration results of selected diamond therrnistors ............ 116
List of applied voltage and power at 480 and 680 K for diamond heaters with
different resistivity ................................ 133
The results of heating/sensing study ....................... 140

1.1

2. 1
2.2
2.3
2.4

2.5

2.6
2.7

3.1
3.2
3.3
3.4
3.5
3.6
3.7

3.8

3.9

3.10

LIST OF FIGURES

Realization of diamond sensor/heater structure ................... 5
Comparison of R-T characteristics for thennistors and Pt RTD .......... 16
Seebeck voltage of various thermocouples listed in Table 2.3 .......... 19
Band diagram of diamond ............................. 24
Schematic diagram of a hot ﬁlament-assisted CVD deposition system for growth

of diamond thin ﬁlms. .............................. 29
Schematic diagram of an oxygen-acetylene apparatus for the growth of diamond

ﬁlms. (From Hanssen et a1. (1991)) ....................... 33
The temperature dependence of properties for different diamond structures. . . 40
Results of gauge factors study by Taher et a1 ................... 43
The simpliﬁed schematic diagram of HFCVD system. ............. 48
The cross sectional view of microwave cavity ................... 52
The design of water cooled sample stage. .................... 53
The conﬁguration of microwave power supply .................. 55
Schematic of MPCVD ﬂow control and vacuum system ............. 57
Block diagram of the computer monitor system .................. 58

Comparison of (a), (b) DPLF] and (c), (d) DPLF3 samples. Samples (a) & (c) are
deposited by MPCVD at 470 °C for 6 hours. Samples (b) & (d) are deposited by

HFCVD at 900 °C for 4 hours. The DPLF 1 samples are continuous ....... 64
Surface roughness comparison of 1 um thickness ﬁlm; (a) DPLFl, (b) DPLF3, (c)
closer view of DPLFl, (d) closer view of DPLF3 ................. 65

Different growth stages of diamond ﬁlms deposited by MPCVD at 850 °C for
15, 25 and 60 minutes using DPLFl (a, b, c), DPLF2 (d, e, f) and DPLF3 (g, h, i),

respectively ..................................... 67
Different growth stages of DPLFl ﬁlms deposited by HFCVD at 900 °C for (a)
2.5, (b) 5, (c) 7.5 and (d) 10 hours ......................... 69

xi

3.11

3.12

3.13
3.14

3.15
3.16

3.17
3.18
3.19

3.20
3.21

3.22

4. l
4.2
4.3
4.4
4.5
4.6
4.7
4.8

4.9

4.10

4.11

Average surface roughness versus ﬁlm thickness for samples prepared by both

HFCVD and MPCVD ............................... 70
Film thickness and grain size versus deposition time for DPLF2 and DPLF3
samples prepared by MPCVD at 850 °C ...................... 70
Surface roughness versus grain density for all samples .............. 72
Rarnan spectra of three different seeding samples after 1 hour deposition by
MPCVD at 850 °C ................................. 72
Patterning of diamond film by DPLFs method (not to scale) ........... 75
Patterning results of (a) & (c) DPLF 1 and (b) & (d) DPLF3 samples by
photolithography. ................................ 76
Diamond patterns by direct write method ..................... 78
The schematic diagram of computer controlled direct-write system. ...... 80

The circuit diagram of pressure controlled spray nozzle (a) and assemblies of

dispense valve (b), & (c) .............................. 81
Temperature controlled sample holder ....................... 83
SEM pictures of metal contacts after annealing. (a) Al, (b) Cr on poly-diamond
and Si02 substrates. ............................... 86
The I-V characterization of diamond testchips with different grain sizes and
contact metals. .................................. 88
The schematic diagrams of diamond test structures ................ 92

The experimental setup configurations of temperature response characterization. 97

The experimental setup for heating/sensing characterization ........... 99
Schematic of temperature profile characterization ................ 101
The I-V characterization of three different metal contact structures. ..... 103
The grain size effect on resistivity before and after thermal annealing ...... 106
Schematic of typical placement of samples and doping source ......... 107

The representative measured data of (a) small grain polycrystalline diamond

(S2 & S7), (b) large grain polycrystalline (S3) and single crystal (S4) samples. 110
The resistivity (a) and normalized resistivity (b) versus temperature for selected
diamond thermistors in sample S7 ........................ 113
The resistivity (a) and normalized resistivity (b) versus temperature for selected
diamond thermistors in sample S2 ......................... 114
The long term stability of diamond resistivity .................. 119

xii

4.12

4.13
4.14
4.15
4.16
4.17
4.18

4. 19
4.20

The comparison of temperature response of single crystal and polycrystalline

diamond samples with compatible room temperature resistivity values. . . . 126
The B versus resistivity. ............................ 128
Temperature versus power density under different measurement conﬁgurations. 130
The temperature proﬁles of sample S3-1. ................... 136
The temperature proﬁles of sample S2-6 before the black paint coating. . . . 137
The temperature proﬁles of sample S2-6 with black paint coating ....... 138
The temperature sensing characterization of all sensor/heater samples. The solid

lines represent ﬁtting curves by Steinhart—Hart equation. ........... 140
Results of self-sensing diamond heaters ..................... 141

The difference of heating temperature determined by self-sensing (T cal.) and that
by thermocouple (Tmea) .............................. 142

xiii

CHAPTER].

FE$EARCHBMIHVATHMWAND
GOALS

1.1 Introduction

The great progress of microelectronics technology has changed many aspects of the
life of humanity. Due to advanced integrated circuit technologies, microprocessors and
related memory units have become available at relatively low cost not only for
professional usages, but also for consumer or household usages. Widespread applications
of microprocessors and related memory units, which are able to handle and store
information, have created the need for low cost and high performance sensors and
actuators in large quantities. Thus, the application areas of sensors and actuators have
shifted from the conventional purposes, such as measurement, testing, and/or automation,
to more consumer oriented ones (e. g., domestic and automotive electronics, household,
safety, comfort, and pleasure) and highly advanced scientiﬁc areas (e.g., medicine,
environment protection, research, etc.). In order to incorporate modern sensors and
actuators into microprocessor-based systems for a variety of applications, the
requirements for the sensor and actuator materials are now focused on long-term
reliability, compatibility with IC batch fabrication, environmental stability, and capability

of micromachining. Materials that are widely used or under investigation for modern

sensor/actuator applications include semiconductors, ceramics, glasses, optical ﬁbers, and
polymeric materials [1,2].

Temperature sensors are widely utilized in a variety of areas, such as aerospace,
manufacturing industries, medical and biological applications, and household usages [3].
Due to a strong demand in modern automation of machinery and process control, great
scientiﬁc and technological efforts are necessary to produce temperature sensors with a
wide range of temperatures, good measurement accuracy, excellent sensitivity, good
stability under various environments, and fast dynamic response. Common types of
temperature sensors include thermocouples, resistance temperature detectors (RTD),
thermistors, integrated circuit (IC) sensors, diodes and transistors [4,5,6]. Among those
different types of temperature sensors, thermistors are not only used to measure
temperature but also suitable for a wide range of applications. In the use of temperature

sensing, therrrristors are reported to be very sensitive but non-linear in the range of ~100 to

500 °C [5]. Other applications of thermistors include ﬂowmeters, liquid level sensors,
pressure and vacuum gauges, which are mainly based on the measurements of heat
dissipation [3]. For these applications, thernristors with self-heating capability are
essential. At present, metal-oxide based materials are commonly used in the commercially

available thermistors. The ﬁrst commercial thermistor was made of 002 in Germany in
1932 [3]. The typical maximum operation temperature of those metal-oxide therrrristors is

somewhat limited to 250 - 300 °C.

Heat generation is also required in a wide variety of areas. Depending upon the
applications, heaters with different element materials, shapes, and conﬁgurations are
developed to enhance the performance in certain ways. Different heater elements include
strip heaters, cartridge heaters, tubular heaters, immersion heaters, and so on. Among all
of them, the thick/thin ﬁlm heater offers the advantages of small sizes and direct
deposition on the object to be heated to achieve precise heating. The thick/thin ﬁlm

heaters commonly available today are based on nickel, aluminum, or carbon steel alloy

[7,8]. The maximum operation temperature and power density for those heaters are in the

range of 300 - 1300 °c and 0.8 - 40 watt/cmz, respectively [9.10.11].

In most real life heating applications, temperature sensing and control are essential.
For liquid level sensors, mass ﬂow meters, and vacuum and pressure gauges which are
based on measurements of heat dissipation, the capabilities of heat generation and
temperature sensing are required simultaneously. Since the materials commonly used as
heating elements lack high sensitivity to the temperature change, different materials are
utilized for heating and temperature sensing. In such a conﬁguration, thermal and
chemical properties of materials involved, if not carefully considered, may cause
problems. It is also important to minimize the response time and uncertainties in the
measurements associated with heat dissipation and placement of sensing and heating
elements. The use of a single element as a heater and a temperature sensor may help
eliminate such problems. The desired material properties for such an element include high
thermal conductivity, ability to be used as an electrical conductor and insulator, high
sensitivity to temperature, rnicromachining capability, resistance to chemical attack, and

mechanical stability.

1.2 Objective of This Work

Due to a unique combination of its mechanical, physical, chemical, and electrical
properties, diamond has been recognized as an excellent material for high temperature
electronics applications [12]. The properties of diamond are well-understood, but the use
for active electronic device applications is not feasible at this time because of the
difﬁculties associated with its economical growth and n-type doping. Therefore, the

current prognosis for diamond is primarily as a protective coating [13,14], a thermal

management ﬁlm [15], and a material for electron-emitting cathodes [16,17]. The
temperature sensors and thin ﬁlm heaters, which do not need n-type doping and single
crystal diamond, seem to be suitable applications to fully beneﬁt from the excellent
properties of diamond. Recent progress in the technology of chemical vapour deposited
(CVD) diamond has led to inexpensive high-quality intrinsic or p—type polycrystalline
ﬁlms on various non-diarnond substrates. Due to its large band gap, high thermal
conductivity, and chemical and mechanic stability, diamond is an excellent material for
high temperature sensing and heating applications, especially under harsh environments.
Early work on diamond thermistors has successfully demonstrated the excellent

sensing properties of diamond thermistors [18,19]. However, this also revealed some

problems mainly associated with the relatively low nucleation density of 108 cm’z. There
are also some limitations due to the seeding and patterning techniques used today, such as
poor selectivity and reproducibility, and restriction to two dimensional structures. In order
to utilize diamond for real world temperature sensor/heater applications, it is necessary to
develop mature device infrastructure and characterization technologies. Figure 1.1 shows
a number of essential techniques needed to be developed. First, a diamond device
infrastructure technology is required to successfully fabricate the sensor/heater structures.
Then a development in the characterization methods for an emerging material is also

irrrportant to fully calibrate and evaluate the devices.

The present research mainly focuses on the following issues:

(1) Enhancement in the diamond nucleation density to effectively reduce the grain

size, surface roughness, and deposition time for continuous ﬁlms.

(2) Development of a seeding and patterning technique, which is compatible with

standard IC fabrication processes, capable to enhance the selectivity, and

 

 

Diamond Sensor/
Heater Structure

 

 

 

 

 

 

 

->

 

 

Device Infrastructure

 

 

|~

 

 

 

V

 

->“ Characterization ”—

Figure 1.1 Realization of diamond sensor/heater structure.

Nucleation

Patteming

Deposition

Doping

 

Metallization

Film Quality

Electrical

 

 

Temperature
Sensing

Fr:

Heating

 

Reliability

adaptable to 3-D patterning.

(3) Fabrication of diamond sensor/heater test structures to optimize the device

performance.

(4) Study of the sensing and heating properties of diamond.

1.3 Dissertation Organization

In addition to research motivation and goals (chapter 1) and conclusions and future
research (chapter 5), this dissertation consists of three major parts, namely, background,
diamond device infrastructure, and single element diamond sensor/heater structures. They
are described in the chapter 2, 3, and 4, respectively.

In chapter 2, overviews of the current technologies of temperature sensors and thin
ﬁlm heaters are presented. The diamond properties, thin ﬁlm deposition techniques, and
diarnond-based sensors are also described in this chapter. Chapter 3 deals with the
development of infrastructure technologies for fabrication of diamond sensor/heater
devices, which includes thin ﬁlm deposition, ultrahigh nucleation density, and device
fabrication. Chapter 4 focuses on the characterization of diamond sensor/heater test
structures. The temperature sensing and heating properties of CVD diamond are
discussed.

CHAPTERZ

BACKGROUND

2.1 Introduction

This chapter presents an overview of materials related to the present research. First, a
report of current technologies of temperature measurement and thin ﬁlm heater is
presented. The properties of diamond are then examined to illustrate the advantages of
diamond for electronic applications. This follows a comprehensive description of the
CVD diamond technologies. Toward the end of this chapter, the current status of various

diamond sensors are reported.

2.2 Temperature Sensors

Temperature has been deﬁned in a qualitative way as the level of thermal energy, and
it has been identiﬁed as a state function, one that must be speciﬁed in order to characterize
the state of a material or system. A variety of material properties have been utilized to
determine the temperature and are summarized in Table 2.1. In this section, an overview

of the most commonly used thermometers is presented.

TABLE 2.1 Material properties used to measure temperature

 

Property Example

Thermal expansion

Liquid Mercury thermometer
Solid Piezoelectric crystal resonant frequency
Gas Gas thermometer

Electrical resistance
Metal Platinum resistance temperature detector
Semiconductor Therrrristor, transistor

Thermoelectric Thermocouple

Vapor pressure Liquid hydrogen

Light intensity Optical pyrometer

Heat emission Total radiation or infrared pyrometer

 

2.2.1 Liquid-in-Glass Thermometers

The liquid-in—glass thermometer was the first true temperature measuring instrument
[6]. It generally consists of a thin walled glass bulb attached to a glass stem with the bulb
and stem system sealed against its environment. The portion of the bulb-stem space that is
not occupied with the thermometer liquid usually is ﬁlled with a dry inert gas under
sufﬁcient pressure to prevent separation of the thermometer liquid. A scale is provided to
indicate the height to which the liquid column rises in the stem, and this reading is made to
indicate closely the temperature of the bulb. A contraction chamber (an enlargement of the
capillary) often is provided below the mail capillary to avoid the need for a long length of
capillary or to prevent contraction of the liquid column into bulb. An expansion chamber
(an enlargement of the capillary) often is provided above the main capillary to protect the
thermometer in the case of overheating. A reference point (usually ice point) should be
provided as a check against changes in the bulb volume [4,5,6].

The operation of a liquid-in-glass thermometer thus depends on the coefficient of
expansion of the liquid being greater than that of the containing glass bulb. Any increase
in the bulb temperature cause the liquid to expand and rise in the stem, with the difference
in volume between the bulb and stem serving to magnify the change in volume of the

liquid. The most commonly used thermometer liquid is mercury, though mercury-thallium

and organic liquids are also used under the circumstance of temperature below 0 °C. The

accuracies of this type of thermometers were reported to be in the range of 0.01 to l or 0.1
to 2 °C. depending on the type of immersion, in the operation temperature range of -328 to

500 0C [5,6].

10

2.2.2 Resistance Thermometry

A resistance thermometer generally consists of four major portions as follows [6]:

(l) A sensor, an electric circuit element whose resistance varies with temperature;

(2) A framework on which to support the sensor;

(3) A sheath by which the sensor is protected;

(4) Wires by which the sensor is connected to a measuring instrument (usually a

bridge), which is used to indicate the effect of variation in the sensor resistance.

This type of thermometer can be divided conveniently into two basic groups,
resistance temperature detectors (RTD) and thermistors, based on the construction
materials of metals and semiconductors, respectively. Resistance thermometers provide
temperature directly in the sense that no reference junction are involved, and no special
extension wires are needed to connect the sensor with measuring instrument. The
advantages of resistance thermometers are simplicity of circuits, sensitivity of

measurement, and stability of sensors.

2.2.2.1 Resistance Temperature Detector

The RTDs are generally made of platinum (Pt), nickel (Ni), and copper (Cu). The Pt
RTD has a quite linear and stable resistance-temperature relation over a wide temperature
range and is used exclusively for precise temperature measurement. However, it is
extremely sensitive to contamination and to strains. Thus a complex mounting and careful
packaging are essential for Pt RTDs. The RTDs made of Cu or Ni are much cheaper than

those made of Pt, but the operation temperature ranges are limited.

11

The dependence of resistance on temperature for most metals above the Debye

temperature can be written as a power series [5],

R = Ro(1+At+Br2+Ct3) .............. (2.1)

Here A, B, C are constants of material. The Callendar equation deﬁned as

r = 100(ﬂ) + 86-30(61) - I) ............... (2.2)

is most commonly used for relating platinum resistance to temperature. Here t is degrees
Celsius, R is resistance, and 8 is a constant representing the small departure from linearity
that is appropriate to the particular thermometer. The subscripts on R indicate the
resistance at 0 °C. at 100 °C. and at the measured temperature [5,6].

Generally speaking, the RTDs possess a positive temperature coefﬁcient and offer

precise temperature measurement over a fairly large temperature range (an accuracy of :1:

0.001 °C is routinely achieved using Callendar equation in laboratory usage [5]). Since the
resistance change in temperature of RTDs is small, very accurate measuring instruments

are required. Table 2.2 summarizes the characteristics of RTDs made of Pt, Cu, and Ni.

2.2.2.2 Semiconductor Thermistor

A thermistor is a special type of semiconductor with a predictable variation in
resistance as temperature is changed. Therrnistors can have either negative or positive

temperature coefﬁcients of resistance. The former are particular useful for temperature

12

TABLE 2.2 The characteristics of resistance temperature detectors

 

Platinum Copper Nickel
m

Average temperature coefﬁcient 0.00385 - 0.0042 0.0067
(01) of resistance over 0 to 100 °C 0903927
(°C')
Resistivity (0cm) 9.81 x 10" 1.529 x 10"5 5.91 x 10-6
Linearity of resistance versus tem- Excellent Excellent Poor
perature

 

compensation of electronic circuits and are widely used for this purpose [5]. Most
thermistors are compound semiconductors which are usually ceramic compositions of
transition metal oxides, such as CuZO, C0203, SnO, TiOz, fie/203, NiO, U203, and Mn203
[5]. In this case of thermistors, the exactly electrical conduction mechanism is complex
and poorly understood. At present, two models are used to provide an explanation for the
resistivity-temperature behavior of metal oxide thermistors. One model for electrical
conduction theory involves the happing mechanism which is observed in ferrites and
manganites that have a spine] crystal structure. In this model, conduction occurs when
charge carriers hop from one ionic site in the spine] lattice to an adjacent site. The other
model is based upon a consideration of energy bands. According to this model, the charge
carriers are the results of physical impurities rather than chemical impurities, and occur

from either an excess or a deﬁciency of oxygen atoms in the crystal lattice. Regardless of

13

the mechanism, the resistivity can be described by

 

) .......... (2.3)

where 71;, 1:, q, and [.1 are intrinsic carrier concentration, Boltzmann's constant, electronic
charge, and carrier mobility, respectively, and AE=E,- - Ef and AE=Ef- E,- for p—type and n-
type senriconductors, respectively. Because the exponential term usually dominates, the

resistivity of a thermistor can be approximated by

P 903 .......... (2.4)

where po is the resistivity at inﬁnite temperature and B is a material constant for the

device. The resistance R of a resistor is given by

R = p; .......... (25)

where L is the length and A is the cross-sectional area of the resistor. Using Eq. (2.4) and

Eq. (2.5), the resistance can be expressed as

1 1
R = RoeB(T-1;) .......... (2.6)

In Eq. (2.6), R0 is the resistance measured at T0. In semiconductor thermistor technology,

14

To is often taken at 25 0C. If temperature is the parameter to be determined from resistance

measurement, then Eq. (2.6) can be rewritten as

T(K) = {[13(ln(;3))+ 71-5}.l ........... (2.7)

Due to the non-linearity of ln(R) versus 1/1‘ characteristic of semiconductor thermistors,
Eq. (2.7) is valid only for small temperature spans for which the ln(R) versus llT
characteristic approximates a straight line. Various empirical equations were used to
describe the relation between resistivity (or resistance) and temperature of thernristors,
which is better than Eq. (2.7). Among those equations, the third order Steinhart-Hart
equation [20], deﬁned by

1
T = “+b(1“9)+‘(1“9)2+d(1“9)3 . .......... (2.8)

is determined to be the best mathematical expression of resistivity-temperature relation for
thermistors with negative temperature coefficient [5,6,20].

The sensitivity of temperature response of thermistor is usually characterized by the
constant B [3,5,20], deﬁned as

ln(Ro/R)
B = (171071715 .......... (2.9)

The typical value of B for conventional metal-oxide thermistors ranges from 2000 to 4000

K over a temperature range of 25 to 400 °C [5,20]. Another commonly used parameter of

15

sensitivity of temperature sensors is the temperature coefﬁcient or, which is deﬁned as

ldR
01 = R31 .......... (2.10)

It is usually expressed in terms of percentage change in resistance per degree. In the case
of thermistor, at is a strong function of temperature. From Eq. (2.6) and (2.10), or can be

represented by

H)
4114110.: °

or "' r 1
R0850 1;)

 

= .314 .......... (2.11)

Therrnistors offer a variety of applications. For temperature measurement, thermistors
are extremely sensitive to temperature change. They may be 1000 to 1 million times more
sensitive than a platinum RTD. Figure 2.1 shows the comparison of typical resistance-
temperature relation of Pt RTD and thermistors [6]. However, they are less stable and
nonlinear as conrpared to platinum RTD. In addition, the current flowing through a
thermistor should be small to minimize Joule heating, which introduces error for
temperature reading. For other applications, such as switching, temperature control, and
ﬂow measurement, sufﬁcient current to cause Joule heating is desirable. The nonlinear
nature of thermistors is undesirable in many applications. For electronic circuitry as well
as temperature measurement, the resistance is often modiﬁed by combining one or more
thermistors with other resistors to enhance the measurement resolution, sensitivity, or
linearity. Widely used circuits for correction of non-linearity of thermistors are

Wheatstone bridge, Callendar-Grifﬁths bridge, Smith bridge, and Mueller bridge [3,4,20].

 

 

 

 

 

 

 

 

 

 

100.0
. ﬂ 1 r r I
5
.1-
2 —I
10.0 .
a —
6 c:
‘ —I
2 Platinum RTD
(1000le
1.0 _
8 -
6 —
a? 2 _
.9
‘5 0.1 ..
8- -
6 ‘ Therrnistors
E " ‘ Ratio Typical
2 ,_ Ratio _ Ras'c resistance
R1351: (0)8!25°C
0.01_ am a a ——
a- REE 11.8 (100)
f 15.7 (500)
" 11.8 2930 -3.3 19.8 (2000)
2_ 15.7 3270 -3.7 _
0001 19.3 354° ’4'0 30.6 (30.000)
' 3C 30-5 4055 "I; _ 35.7 (100.000)
5- 35-7 ‘2” "- 39.2 (200.000)
9— 39'2 435° "'9 51.5 (1.000.000)
51.5 4710 -5.3
2» 75.0 5120 -5.7 - 75.0 (5.000.000)
0.0001 1 1 1 1 1 1
-50 0 so 100 150 200 250 300
Temperature(°C)

Figure 2.1 Comparison of R-T characteristics for thermistors and Pt RTD.

17

2.2.3 Thermoelectric Thermometry

Thermoelectric effects have been employed for over 100 years as a practical tool for
the measurement of temperature [21]. Thermoelectric temperature measurements are
made with thermocouples. Thermocouples are usually made from two dissimilar metal
wires connected so that one junction is held at a reference temperature and the other
junction serves as the temperature sensing device. The operation principle of
thermocouples are known as Seebeck effect, which a temperature difference is converted
into an electromotive force called Seebeck voltage. A thermoelectric temperature
measurement system usually consists of a sensing element, electrical lead wires, and a
voltage measuring instrument.

Different combination of metals or metal alloys are reported to manufacture
thermocouples to enhance the performance in a certain temperature range. In general,
thermocouples [22] offer a large measuring temperature range and linear response but are
less sensitive. Table 2.3 summarizes the compositions, operation temperature range, and
limits of error of some commercially available thermocouples [5.6.21]. A plot of Seebeck
voltage versus temperature for some common types of thermocouples is shown in Figure

2.2 [5,23].

2.2.4 Optical Pyrometry

The radiation from a hot body may be used as a measure of its temperature.
Theoretically either the total radiation emitted or that emitted at a single wavelength can
be used. In most cases, it is difﬁcult to ensure the measurement of total radiation due to the
strong absorption of infrared band in the atmosphere. In addition, the single wavelength is
also difﬁcult to achieved. In the practical application, a narrow band of wavelength is

established by a red ﬁlter so that calculation of temperature can be made on the basis of an

18

TABLE 2.3 Summary of important parameters of thermocouples

 

Limits of Error
T tu
Type Positive Wire Negative Wire empera re (°C or %)
”“8" (°C) Standard Special
B P17010130 99411116 871 to 1705 0.5%
E Nigocrlo N145C1155 0 to 316 1.667 1.25
316 to 871 0.5% 0.375%
I Fe N145C1155 0 10 277 2.222 1.1 1 1
277 to 760 0.75% 0.375%
K 19190010 N195Mn2A128i1 0 to 277 2.222 1.111
277 to 1260 0.75% 0.375%
N N1845C1'142811'4 N195.5814.4Mg0.1 0 to 277 2.222 1.1 1 1
277 to 1260 0.75% 0.375%
R P137Rh13 Pt 0 to 538 1.389
538 to 1482 0.25%
s PtgoRhlo Pt 0 to 538 1.389
538 to 1482 0.25%
T C“ Ni4SCu55 '184 to '101 1%
-101 to -59 2% 1%
-59 to 93 0.833 0.417
93 to 371 0.75% 0.375%

 

l9

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

” I E
/
70
,0 X
K
50 J— N
40 e--d
,, /l/ "
/'
20 :6
10 43%
l
0 1000 2000 3000 4000 50m
Temperatureﬂiahrenhelt) A

Figure 2.2 Seebeck voltage of various thermocouples listed in Table 2.3.

effective wavelength. The particular wavelength is usually taken as 0.655 urn [4,5].
However, the emissivity of the object is a function of wavelength and temperature. A
careful consideration of emissivity is essential to obtain an accurate temperature reading
of non-black body. An alternate design of two-color pyrometer, which employs two
wavelengths, offers a convenient measurement of temperature without the need of
emissivity correction. The operation principle is that energy radiated at one color increases
with temperature at a different rate from that at another color [6]. Generally speaking,
temperature obtained by single wavelength pyrometer with careful emissivity correction is

closer to the actual temperature than that by two-color pyrometer. The temperature

20

measured by two-color pyrometer is accurate only when the emissivity at two

wavelengths is equal. Accuracies obtainable with commercial automatic optical
pyrometers are on the order of 3; 3.89 °C for 816 to 1232 °C and of 1 6.67 °C for 1232 to

1760 °C [5.6].

2.3 Thick/Thin Film Heaters

In order to fulﬁll a wide variety of applications, heaters with different heating
elements, shapes and conﬁgurations were developed, such as strip heaters, cartridge
heaters, tubular heaters, immersion heaters, and thick/thin ﬁlm heaters. The thick/thin ﬁlm
heaters offer the advantages of small sizes and direct deposition/attachment on the heated

objects. In general, the thick ﬁlm technique involves screen printing pastes onto substrates

and ﬁring procedure in the temperature range of 850 - 1000 °C to remove the organic
materials in the pastes. The thin ﬁlm technique is referred to those produced by chemical
vapor deposition methods.

The performance of most commercially available thick/thin ﬁlm heaters based on
nickel, aluminum, or carbon steel alloy [7,8] are heavily dependent on the substrate

materials used. The alumina (A1203) and boron nitride (BN) are commonly used as a

substrate when high temperature heating (>600 °C) is required. Silicone rubber is the most
widely used substrate material in ﬂexible heaters. BeO. high density mulcorit (HDM).
porcelain enamelled steel (PBS). Kapton, and Mylar are also used as heater substrates

under different occasions [8,9,10,11]. The maximum operation temperature and power
density for those heaters are in the range of 300 - 1300 °C and 0.8 - 40 watt/cmz,
respectively.

In some special applications. heaters with special materials and structures are designed

21

and fabricated. For example, a polysilicon heater structure was realized using standard
CMOS technology to supply temperatures necessary for the operation of most

semiconductor gas sensors [24]. A doped tungsten wire was used as a heater operated at

2500 °C to enhance the energy of exhaust gases of thrusters. This was reported to extend a
satellite's life by 30% [25].

2.4 Diamond Properties

Diamond is well known as a tetrahedral network of carbon atoms in which each atom
is at the center of four other atoms. The carbon building unit is ordered in a cubic lattice
and interconnected by strong bonds. This makes diamond unique in many of its physical
properties.

The best known and most utilized property is the fact that diamond with a hardness of
10‘ rtg/rnrn2 is number one in all materials [26]. In addition, the combination of low
coefﬁcient of friction (when terminated with hydrogen). highest Young's modulus, low
thermal expansion, and high thermal conductivity make it an excellent material for
numbers of mechanical applications, such as cutting. bearing, heat spreading, and abrasion
applications [26.27.28.29]. Its optical index of refraction (2.41 at 591 nm) and
transnrissivity (225 from nm to far IR) make it an ideal anti-reflection coating for infrared
detectors [26.30.31].

Diamond is commonly regarded as an inert material that cannot be etched by boiling
acids or bases [32,33]. Its immunity of chemical attack makes utilization under chemical

harsh environments become possible. It has been known that diamond oxidizes in an

oxygen environment at temperatures over 600 °C [19,34]. Molten potassium and sodium

nitrides are reported to attack the diamond surface at a temperature in the range of 427 -

827 °C [30.35].

22

The electrical properties of large crystals found in nature or synthesized by high-
temperature high-pressure anvil machines were well understood [36,37]. The
understanding of CVD diamond, especially polycrystalline diamond, however is inferior.
Table 2.4 compares some important properties of commonly used semiconductor
materials for electronics applications. Diamond is known as an indirect-gap
semiconductor, with an energy gap of 5.5 eV at room temperature [36,37 .38]. A calculated
band diagram is shown in Figure 2.3 [33.39].

The properties of carrier mobility and dopant ionization energy are major factors as to
the ultimate practicality of diamond devices. A lot of efforts were made to determine the

carrier mobility for different types of diamonds. Williams et al. (1970) determined the

hole mobility in natural semiconducting diamond in the range of 700 - 2000 cmles at

room temperature by Hall effect measurement. In similar measurement as Williams, the

carrier mobility of synthetic type lIb diamond was determined lower ~300 cmst by
Collins at 1989. Vavilov et al. (1974) and Fujimori et al. (1990) measured hole mobility in

boron doped diamond of 200 - 700 cmst and 850 cm2/Vs, respectively, which are
comparable with 1300 cmst calculated for doping at 1016 cm '3 by Osmarn et al. (1989).

For ion implanted diamonds. values in the range of 200-700 cm2/Vs were observed
(Vavilov et al. 1989). The highest value measured in boron doped CVD single crystal

homo-epitaxial diamond was 1000 - 1430 cmles [40]. For the highly oriented diamond

ﬁlms. a mobility of 165 cmst was reported by Stoner et al. [41]. It is generally expected
that mobility in polycrystalline diamond ﬁlms will be less than that in single crystal ﬁlms
[42]. The Hall mobility of boron-doped polycrystalline diamond ﬁlms are reported in the

ranges of less than 1 up to ~100 cmst [43.44.45]. The combined electron-hole mobility
of undoped polycrystalline CVD diamond ﬁlms, as measured by transient

photoconductivity at low carrier density. are in the range of 50 - 4000 cmst,

23

TABLE 2.4 Properties of semiconductor materials

 

PROPERTIES

Lattice Constant (A)

Thermal Expansion (x10'6 °C)

Density (glcm3)

Melting Point (°C)

Band Gap (eV)

Saturated Electron Velocity

(x107 cm/s)

Carrier Mobility (cm2/Vs)

Electron
Hole

Breakdown (x105 V/cm)
Dielectric Constant
Resistivity (Q cm)

Thermal Conductivity
(chm K)

Absorption Edge (pm)
Refractive Index
Hardness (Kg/mmz)

Johnson Figure of Merit
(x1023 WQ/sz)

Keyes Figure of Merit
(x102 W°Clcm s)

Diamond B-SiC
3.567 4.358
1 .1 4.7
3.51 5 3.21 6
4000 2540
5.45 3.0
2.7 2.5
2200 400
1 600 50
1 00 40
5.5 9.7
1 013 1 50
20 5
0.2 0.4
2.42 2.65
1 0,000 3500
73,856 1 0,240
444 90.3

GaAs

5.65
5.9

1238

1.43

1.0

8500
400

60

12.5

0.46

3.4

62.5

6.3

Si
5.430
2.6
2.328
1420

1.1

1.0

1500
600

11.8

103

1.5

1.4

3.5
1000

9.0

13.8

 

24

 

 

 

Emmy (eV)

  

 

 

 

 

 

 

Figure 2.3 Band diagram of diamond.

respectively [46,47], which is comparable to that of the best single-crystal 11a natural
diamonds [48].
The activation energies for boron, computed from the slopes of measured p versus 111‘

curves assuming no compensation, are in the range of 0.22 - 0.34 eV for samples without
annealing. These values conespond to hole concentration values in the range of 10'8 - 2.0

x 1015 cm'3 [18]. In the case of highly doped diamond ﬁlms with a boron concentration on

25

the order of 10 2° cm'3, the resistivity and activation energies of less than 0.01 0cm and

0.02 eV were observed [49].

2.5 CVD Diamond Technology

This section reports the current CVD diamond technologies needed for electronic
devices and/or sensor applications. First, diamond ﬁlm deposition technologies. such as
nucleation enhancement methods, deposition reactors. and ﬁlm structure characterization,
are presented. To achieve semiconductor sensor devices in diamond, processing
capabilities such as patterning, doping. and metal contacts are critical. They are also
described in this section.

2.5.1 Nucleation

Diamond has been shown to nucleate on a wide variety of materials. Due to the low

nucleation density of diamond on non-diamond materials (~ 10 4 cm'z). substrate surfaces
are, in most cases, treated to enhance nucleation density. Diamond powder [50.51]. c-BN
powder [52] or other types of abrasive polishing, as well as ultrasonic agitation with
diamond powder suspension [53.54.55.561 are the most commonly utilized pre-treatment
techniques. Bias enhanced nucleation (BEN) has also been reported for effectively
creating nucleation sites on silicon substrates [57] and preparing highly oriented diamond
(HOD) ﬁlms [58,59]. Although those methods successfully enhanced the nucleation
density into the range of 107 to 1010 cm’z, an improvement over those existing techniques

is still desired in terms of the reproducibility, maximum nucleation density, and

efﬁciency. A brief description of the nucleation pre-treatment methods is given here.

26
0 Abrasive polishing

It is well known that diamond nucleation density may be increased several orders of
magnitude simply by scratching or abrading the substrate surface prior to placing it into
the growth chamber. Diamond powder [50,51]. c-BN, and SiC [52] powders of various
grain sizes were commonly utilized to polish the substrates for nucleation enhancement. A
rinse of organic solvents or cleaning by ultrasonic bath were sometimes followed to
remove the redundant powder sticking on substrates. It was reported that a nucleation
density of 3 x 1010 cm'2 is achieved by polishing with 10 nm diamond powder [50].
However, abrasive polish will cause damage of substrate surface and has a poor uniformity

and reproducibility.

° Ultrasonic treatment

In this metirod. diamond particles with a wide range of sizes were suspended into
hydrocarbon ﬂuids. With substrates placed inside, the diamond powder solution was
agitated by ultrasonic bath for a certain period of time [53.54]. The substrates were usually
followed by an organic solvent cleaning procedure. It is believed that the residual of

diamond powder bombarding on the substrate surface contributes to the nucleation of
diamond. By utilizing ultrasonic treatment, the nucleation density in the range of 107 to

1010 cm"2 was reported [54,60]. A non-uniformity of growth nucleation was also observed
[52].

0 Bias enhanced nucleation (BEN)

In stead of ex-situ pretreatment for nucleation enhancement. the in situ negative DC
biasing potential of -100 to -300 V [57.58.59] relative to the plasma is performed while

the substrate is immersed in a methane-hydrogen plasma. A thin layer of amorphous

27

carbon found to deposit during the biasing treatment is believed to contribute to the

enhancement of nucleation. A high density of up to 5 X 10 10 cm '2 was reported [41].

However. a long treatment period of 2 hours was required for a high nucleation density.

0 Diamond powder loaded photoresist

This method to enhance nucleation density for diamond growth is based on the idea of
spreading diamond seed crystals. suspended in photoresist, on the substrate surface.
Diamond powder loaded photoresist (DPR) is spin-coated on the substrate surface before
diamond deposition. The sample is subsequently placed in the diamond deposition reactor
and the photoresist evaporates at high deposition temperature leaving behind the diamond
powder particles which act as seeds for diamond growth. The DPR was prepared by

suspending ﬁne diamond powder particles with a mean size of 0.101 um into photoresist

[18.61]. The uniform and reproducible nucleation density of ~10 8 cm'2 was achieved. The
advantages of DPR method include simplicity of application, compatibility with standard
IC fabrication processes. good reproducibility, and no damage of substrates.

2.5.2 Deposition methods

The motivation and interest for the diamond synthesis started since the discovery by
Tennant in late 18 century [62] that it is simply a crystalline form of carbon. The early
attempts which were mostly based on unscientiﬁc approaches were unsuccessful. The real
breakthrough came in 1955 at General Electric by using a high temperature high pressure
(HPHT') process to synthesize diamond from graphite [63]. Simultaneously methods for
synthesizing diamond at low pressure have been under development since early 19505.
Since then. lots of efforts were made to improve the growth rate, ﬁlm quality. and

reproducibility due to the potential applications and advantages of diamond. In the

28

following section. a brief review of various chemical vapor deposition (CVD) methods for

growth of diamond ﬁlms is given.

0 Thermally enhanced CVD methods

Among numbers of thermal enhanced CVD rrrethods, the hot ﬁlament assisted CVD
process is the most commonly utilized one. It was ﬁrst successfully achieved by
Matsumoto and coworkers in 1982 [64,65]. In this method, the diamond ﬁlm is deposited
on a heated substrate from a gas mixture of methane and hydrogen [66] activated by a hot
ﬁlament made of tungsten (W), tantalum (Ta) or rhenium (Re) placed close to the
substrate. A schematic diagram of this deposition reactor is shown in Figure 2.4. Table 2.5
outlines the deposition parameters reported by various workers.

The role of hot ﬁlament is to thermally dissociate the gas mixture of methane and
hydrogen to form hydrocarbon radicals, for example CH. CH2, QHZ. CH3. and atomic
hydrogen. These radicals lead to preferential growth of diamond against graphite on
substrate. The main advantages of HFCVD method are its relative simplicity,
inexpensiveness and easy to scale up to large substrate areas. However. various
components in the system can add impurity into the ﬁlms due to increased vapor pressure
in the presence of hot ﬁlament at very high temperature. Some variety of thermal
enhanced CVD methods are also reported, such as, electron-assisted and laser assisted
thermal CVD [67].

0 Plasma enhanced CVD methods

Another widely utilized method of diamond growth is plasma enhanced CVD of
carbon containing species mixed in low concentration with hydrogen. The role of plasma
is to generate atomic hydrogen and to produce the proper carbon precursors for the growth

of diamond ﬁlms [66]. In such processes, atomic hydrogen is produced by electron impact

29

CH4 '1' H2
((12:83:11);ng Filament
Hydrogen) NWVWW

Sample

Heater Substrate

   

 

I
1

Figure 2.4 Schematic diagram of a hot filament-assisted CVD deposition system
for growth of diamond thin ﬁlm.

TABLE 2.5 Deposition Parameters Used in the Growth of Diamond Films by Hot

 

Filament-Assisted CVD
Gas Mixture Operating Substrate Filament
(CH4/Hz) Pressure Temperature Temperature

0.5 - 2% 10 - 100 tort 700 - 1000 °c 1900 - 2300 °c

 

30

dissociation of molecular hydrogen. Similarly. electron impact dissociation processes are
responsible for the formation of carbon-containing neutral or ionic radicals. The absolute
concentration of atomic hydrogen and neutral radicals depends on the pressure of the
plasma. In low pressure plasma, the electrons acquire high kinetic energy from the electric
ﬁeld. However. due to the high mean free path. they do not transfer much energy to
molecular species. As a result, the gas temperature is relatively low, and thus atomic
hydrogen and neutral radicals are produced in low concentrations by collisions with high
energy electrons only. In the case of high pressure, due to the small mean free path, the gas
and electron temperature are about the sarrre. Thus. the concentrations of atomic hydrogen
and neutral radicals are much higher. because both electron and molecular collisions
contribute to their formation. The accounts for the signiﬁcantly higher growth rates
reported in high pressure plasma.

Plasma generated either by various forms of electrical discharges or induction heating
have been used. Table 2.6 summarizes typical deposition condition reported by works for
three most commonly used plasma sources.

In DC-plasma assisted CVD method, a plasma in a Hz-hydrocarbon mixture is excited

by applying a DC bias across two parallel plates. one of which is the substrate. A growth
rate of 20 pm per hour was reported by Suzuki et al. [68]. Diamond ﬁlms produced by DC
plasma were reported to be under high stress and to contain high concentrations of
hydrogen. Additionally, they may contain impurities resulting from plasma erosion of the
electrodes [67]. As implied by the name, a RF discharge at 13.56 MHz is used to excite the
plasma in RF-plasma-assisted CVD. This method was reported to prepare diamond ﬁlms
with small crystal grains and well adhered to a variety of substrates [69]. In this method,
high power in the discharge was found necessary for efﬁcient diamond growth. which
frequently caused contamination of ﬁlms due to sputtering from the walls of the reactor

[67]. Microwave-plasma—assisted CVD methods have been used much more extensively

31

TABLE 2.6 Deposition Parameters used in the Growth of Diamond Films by
PlasmaenhancedCVD

 

Gas Mixture Operating Substrate

Method (CHJHz) Pressure Temperature Input
DC-plasma 0.5 - 5% 200 torr 600-800 °C Discharge (VII):
1000/4(VlA-cm2)
RF-plasma 0.5 - 2% l - 30 torr 800—1000 °C RF power:
500-1000 W
Microwave— 0.5 - 2% 5 - 100 torr 700-1000 °C 100-700 W
plasma

 

than any other methods and have a number of advantages over any other methods for
growth of diamond ﬁlms, such as no contamination due to lack of electrode erosion.
higher concentration of atomic hydrogen and hydrocarbon radicals due to higher plasma
density. An additional advantage is that the plasma is conﬁned in the center of deposition

chamber thus prevents carbon deposition on the walls of reactors [67].

9 Combustion ﬂame-assisted CVD

Diamond deposition by combustion flame assisted CVD was ﬁrst reported at 1988 by
Hirose and Kondo and triggered a worldwide activity partly because of the simplicity and
low cost of the experimental apparatus and partly because of the high growth rates of 150
[.1111 per hour [70].

A typical set up of combustion flame-assisted CVD reactor is shown in Figure 2.5. It

32

basically consists of an oxygen acetylene torch supplied with oxygen and acetylene with a
mass ﬂow control system. The main combustion reaction leads to the formation of CO and
H2 with a number of reactive intermediates (H. OH, 02 and 02H) [67]. Hydrogen addition
to the oxygen-acetylene ﬂame was found to reduce the amount of amorphous carbon in

diamond ﬁlms. A torch operated in an argon ﬁlled chamber was found to improve the

purity of diamond ﬁlms by Doverspike et al. [67].

2.5 .3 Diamond ﬁlm structure characterization

In order to use the CVD diamond in the practical applications, a careful
characterization is necessary to control and enhance the quality of diamond ﬁlms
deposited by CVD methods. However, the uality of diamond can be deﬁned in terms
of numbers of ways, such as. elemental composition, structure, bonding state, surface
morphology. and defect density [71], depending upon the applications. Secondary ion
mass spectrometry (SIMS) [60], Auger electron spectroscopy (AE8) [72], and X-ray
photoelectron spectroscopy (XPS) [58] can be used to determine the elemental
composition. Transmission electron diffraction (TED) and X-ray diffraction (XRD)
[50,59] allow the determination of crystal structure and lattice spacing. Raman
spectroscopy is very popular for determining true diamond bonding throughout the bulk of
the ﬁlms. Scanning electron microscopy (SEM), scanning tunnelling microscopy (STM)
[54], and atomic force microscopy (AFM) [50,73] are utilized to exanrine the surface
morphology and growth nucleation of diamond films. Transmission electron microscopy
(TEM) [71] can be used to observe the defect rrricrostructure in the bulk ﬁlms and to
determine the defect density. Furthermore, the localized bonding can be observed via
electron energy loss spectroscopy (EELS) of TEM [57]. A brief review of commonly used

tools is given below.

33

Mass Flow

 

 

      
   

Controller
.. W
Nozzle
Flame

Substrate WW

1

Adjustable Stage

Figure 2.5 Schematic diagram of an oxygen-acetylene apparatus for the growth of
diamond ﬁlms. (From Hanssen et al. (1991))

ORamanspectroscopy

In Raman effect, discovered by physicist Sir Chandrasekhara Venkata Raman of India,
light is scattered by an atom or a molecule that changes its state. There is a corresponding
discrete decrease in the frequency of scattered radiation. Raman spectroscopy is widely

used in the analysis of materials, and the identiﬁcation of trace elements. In the diamond
related researches, Raman spectroscopy is a powerful tool to identify the presence of sp2
and sp3 bonding in diamond ﬁlms [74]. However, a great difference in the scattering
efﬁciency for graphite (sp2) and diamond (sp3) needs to be considered to make a valuable

evaluation. The wave number shift of 1332 corresponds to sp3 diamond peak.

0 Scanning electron microscopy (SEM)

SEM consists of an electron gun in a vacuum chamber column and images by
collecting the secondary electrons emitted from samples due to the incident electron beam.
The image then is displayed on a cathode ray tube (CRT) for directly vision [75]. Because
of the large magniﬁcation range of 10 X to 300,000 X [76] and great depth of ﬁeld, SEM
is a very powerful tool for studying the morphology and visual analysis of diamond ﬁlms.
It is widely utilized to inspect the surface morphology, crystal orientation. the grain sizes
and ﬁlm thickness, and to monitor the nucleation densities and patterning. The need for a
conducting specimen somewhat limits its utility for undoped ﬁlms on insulating

substrates.

- Atomic force microscopy (AFM)

Atomic force microscopy is another useful tool for studying the nucleation density,
crystal structure and surface morphology of CVD diamond ﬁlms. In an AFM, 2 very ﬁne

tip, mounted on a cantilever, is scanned through the sample to obtain the surface proﬁle.

35

The sample is mounted on a piezoelectric device which provides the movement of the
sample during scanning. The advantages of AFM are higher resolution of 0.1 nm or less
and greater sensitivity to deﬁne proﬁle differences of vertical variations in sample [75].
The vertical resolution of AFM is in the sub-A range. better than that of other
proﬁlometers [77], which make AFM ideal for study of surface roughness of diamond
ﬁlms. In addition, no vacuum is needed for the operation of AFM, and it can be used on

non-conducting surfaces.

2.5.4 Patterning

In order to fulﬁl a variety of applications and desirable structures, various techniques
have been developed to pattern diamond thin ﬁlms. Patterning techniques are. in general,
distinguished into two categories. selective deposition and selective etching. In addition,
the ArF excimer lasers with a photon energy of 6.4 eV and wave length of 193 nm was
reported to prepare free-standing single crystal diamond microstructures from
homoepitaxial ﬁlms [78]. The feature sizes prepared by this method are in the range of
200 - 400 pm with a thickness of ~15 11m.

0 Selective deposition

The selective deposition is achieved by pre-deposition nucleation on the desired area,
or by masking the undesirable area during the diamond deposition. Hirabayashi et al. used
Ar sputtering in undesired growth regions to suppress nucleation after substrates were

preheated by ultrasonic method [56]. Ma et al. reported selective deposition onto SiOz

stripes patterned on a Si substrate [53]. The Ar+ ion beam was used to suppress the
nucleation on both Si and SiOz except in the shadows cast by the downwind edge of the

8102 stripes. The selective nucleation was successfully achieved by standard

36

photolithography in DPR method, developed at Michigan State University. or by masking
substrates during the ultrasonic treatment by Masood et al. [61]. Leksono and coworkers
reported successfully the patterning of diamond ﬁlms by lift off process on Si wafer. Thin
ZnO ﬁlm and amorphous silicon were used as sacriﬁcial layers for lift off processes [79].

Si02 or Si3N4 as a masking layer during the diamond deposition was successfully reported

by Masood et al. [61] and Roppel and coworkers [80].

- Selective etching

The patterning of continuous diamond ﬁlms by selective etching was reported by
Masood et al. [19,61]. Using a rapid thermal processor (RTP), diamond ﬁlms were

successfully etched in oxygen environment at 700 °C. The Si02 or Si3N4 were used as

masks to block the undesirable area during etching process.

2.5.5 Doping

As a wide band gap semiconductor material, 1.3ij eV. diamond offers advantages for

electronic applications under extreme environmental conditions. Realization of those
applications requires a control over the ability to dope it p- and n-type. CVD diamond
ﬁlms deposited without intentional doping are usually good insulators. To achieve the
semiconducting diamond ﬁlms, one should be able to incorporate controlled amounts of
impurities into diamond ﬁlms. In general, doping of semiconductors could be achieved by
diffusion. ion implantation, or in situ doping. Because of the low diffusibility of most
elements in diamond, diffusion is impractical. Although ion implantation can achieve
active doping. the damages of crystal structure were introduced [81]. Therefore, in situ
doping was commonly utilized in preparing semiconducting diamond ﬁlms. Boron (B),

aluminum (Al), phosphor (P). lithium (Li), and nitrogen (N) are potential dopants of

37

diamond. P-type diamond ﬁlms were successfully prepared by using boron as an in situ
dopant. However, all attempts to obtain n-type doping were unsuccessful until now
[82,83].

A number of in situ doping techniques involving gaseous. liquid and solid sources to
incorporate boron into CVD diamond ﬁlms have been successfully used to produce the p-
type semiconducting diamond ﬁlms. Solid sources (B203 [84,85], boron powder [18])
have the advantages of chemical stability and simplicity. However, non-uniforrrrity of
doping are observed and it easily contaminates deposition chamber. Gaseous source
(B2H6 [41]) has a better controllability, but is poisonous and explosive. Liquid source
(B(CH3O)3 [84,86]) is also used to prepare p-type diamond and is reported to be

nonpoisonous. stable, and controllable.

2.5 .6 Metallization

Metallic contacts are essential for achieving electronic devices. For thermistor
applications. it is desirable that the resistance of devices be determined mainly by the
resistivity of the bulk material and that the inﬂuence of contact resistance be minimized. In
addition, the good adhesion and high temperature stability are essential for developing the
contact structure for high temperature device applications.

It is found that Ti and Mo form carbides with diamond at high temperatures which

provide good ohmic contact with diamond [87]. The contact resistance is found to
decrease with an increase in the doping level [88]. A contact resistivity (pc=Rc2/RD) of ~
1 x 10 '6 9 cm2 was reported upon annealing at 600 °C for highly doped (~ 10 20 cm '3)
diamond ﬁlms [89], where Rc is the contact resistance measured by transmission line

method (TML) and R3 is the sheet resistance of bulk diamond ﬁlm. TilAu, Ta/Au, Ti/Pt

38

and TilMo/Au were reported by several workers to provide good contacts with diamond
ﬁlms [18.89.90.9l]. With the presence of passivation layer. the TilAu and TilMo/Au

contacts were reported stable for extended periods at 500 and 670 °C. respectively. in
ambient environment [91,92]. It is also found that the Ti/Pt bi-layer provided a good

contact up to 1000 °C under high vacuum of 10'7 torr and a good adhesion between
diamond and SiOz [18.19].

2.6 Diamond Sensors

According to physical quantities to be measured, sensors can be roughly grouped into
temperature sensors, mechanical sensors (pressure sensors, accelerometers, vibrometers,
etc.). acoustic sensors (rrricrophones, ultrasonic sensors, hydrophones), radiation sensors.
humidity sensors, gas sensors, ion-selective sensors, sensors in medicine and biology, and
others (liquid component sensors, material identiﬁcation, etc.). Many known physical
phenomena are used in the design of those sensors [93].

Diamond with an excellent combination of properties has been studied by many
investigators for sensing applications. This section reports the current status of various

diamond sensors found in literatures.

2.6.1 Diamond temperature sensors

In order to utilize diamond for real world temperature sensor applications,
characterization of the temperature response, the sensitivity. the operation temperature
range, the reproducibility, and reliability are important. The temperature dependence of a
diamond temperature sensor is generally characterized in terms of resistivity (or

resistance) by two-point, four-point probe, and Van der Pauw [94] methods. The canier

39

concentration and mobility determined by Hall effect measurement [81,95] or by transient
photoconductivity measurement [96] are found important to further understand the
conduction mechanisms of diamond at different temperatures.

The resistivity of diamond was studied as a function of temperature for different
structures. doping levels. and geometry by a number of researchers. It is found that the

CVD polycrystalline diamond has a negative temperature coefﬁcient (NTC) of resistivity

and temperature coefﬁcient 01 of -0.005 to -0.02 K‘1 over the temperature range of -73 to
1273 K in vacuum [18]. The typical beta (B) factors of polycrystalline diamond
thermistors are in the range of ~1000 to 5500 K [97.98] as compared to 2000 to 4000 K of
conventional metal-oxide thermistors over 298 - 673 K [20]. However, the high sensitivity
(B value) is achieved at low doping level where the corresponding resistance values are
unacceptably high (i.e. greater than 1 M0) for practical applications. As reported by Fox
and coworkers. the polycrystalline ﬁlms are preferable to highly oriented and homo-
epitaxial diarnond ﬁlms for thermistor application due to their improved linearity with
inverse temperature [91]. The Figure 2.6 shows the temperature dependence of resistivity,
carrier concentration, and mobility for homoepitaxial, highly oriented, and polycrystalline
diamond reported by Fox et al. [91].

The effects of high temperature annealing on the diamond ﬁlms and metal contacts

were studied to make reliable and reproducible sensors. It is well known that diamond

ﬁlms oxidize at temperatures > 600 °C in oxygen presented environments. Oxidation of
diamond by oxygen plasma was observed even at lower temperatures. In oxygen ambient,

the maximum operation temperature, tirerefore, is limited by the oxidation temperature of

600 °C. The use of diamond thermistors in oxidizing ambient at high temperatures may be
possible if the diamond ﬁlm is passivated, which may. however, reduce the sensitivity of

the thermistor. Chalker et al. reported that non-passivated diamond ﬁlms becanre non-

conducting at 610 °C in air, while silica coated devices operated up to marginally higher

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E 17 D l-lonroepltaadel _ Q1313
3 10 "' A Wtyorlented A
10"q —° W” — ‘c‘ 100 We '3
2
o o
‘0“- 00° A '- v 10 04 U
‘18 a A E o
‘0 "' 6A$ '- I- A
10"- D - g 1 °
11 O D I (:1 Homeepitexiel
10 - D " A Highlyoriented
8 ‘010‘ 0:130 01 _ o Polycrystalline
I I l I I l ' *1 i r l l I r
2 3 4 5 6 7 8 2 s 4 5 e 7 8
1
room «(‘1 1000"“)
10000 _ P
g [a Homoepitexlal
o Highlyoriented
9 1000- 6 -
I 0 Po c etalllne
a g lyry , 6 (:1
V o
100d 0 -
a. 8 D
a: 88 a
Z 10- o x D "
'5 0°15 0 -
.. goal) (:1
' 9°44 o D
O 1--1 °AA DOD '
I: 8110
0.1- i r i l i
2.0 2.5 3.0 ‘ 3.5 4.0
1000/T (K)

Figure 2.6 The temperature dependence of properties for different diamond

structures.

41

temperature of 670 °C due to the crack of coating material [92]. A degradation of metal

contact (Ti/Mo/Au) was also observed via a combination of oxidation and carburization.
The long term stability at 500 °C for 200 hours was reported by Fox et al. It is also found

that presence of a silicon nitride passivation layer allowed stable operation in 800 °C [91].

2.6.2 Radiation sensors

Sensors for electromagnetic radiation may be classiﬁed as quantum sensors and
thermal sensors. Quantum sensors are based on photoenrission, photoconduction, or
photovoltaic effects. Thermal sensors operate in two steps: frrst, an incident photon ﬂux is
converted into a temperature variation, which then can be determined [93].

Diamond has been used for various radiation detection purposes. The diamond
radiation detectors reported to date are mainly simple metal-insulator-metal (MIM)
devices based on photoconductive effect. Electrons that absorbed a photon with energy
greater than bandgap (5.5 eV) raised from the valence band to conduction band. The
presence of the photon-induced electron in the conduction band gives an increase in the
conductivity. This type of diamond photodetectors are reported to detect the ultraviolet.
X-rays, 'y-rays, neutrons, high-energy electrons, and charged particle with high sensitivity
[99,100,101,102,103]. The MIM structure has been successfully used as a switching
device which is activated by electron beam bombardment [104]. The diamond diodes are
reported to be used as UV detectors by Marchywka et. al. [105]. By characterizing the
photocapacitance, diamond surface channel p-type MIS capacitors are reported to be
utilized as an UV radiation detector with low leakage [106].

Due to the excellent combination of properties, diamond detectors can be operated

under extremely high radiation ﬁelds with a large dynamic range in both response and

speed.

42

2.6.3 Piezoresistive Sensors

The piezoresistive effect of HPHT synthetic diamond was reported by Larsa et al. at
early eighties [107 .108]. The resistance was characterized as a function of stress in <100>

and <111> directions with excellent sensitivity over the pressure range of 0 to 2000

kg.cm’2. The results suggest that diamond is an excellent material for micromechanical
sensors. However the diamond sensors have never been promoted mainly due to the high
cost.

The observation of signiﬁcant piezoresistive effect was ﬁrst reported in both
polycrystalline and homoepitaxial CVD p—type diamond layers of 2 pm thick by Aslam et
al. in 1992 [109]. In addition, the ability to nricromachine diamond ﬁlms into miniature
beams and membranes, the applications of polycrystalline diamond ﬁhns in the strain
gauges, pressure sensors, and accelerometers then were studied by various researchers
[110,111,112].

The diamond cantilever beam and free standing diaphragm structure are commonly
used to study the gauge factor of diamond ﬁlms. The room temperature piezoresistivity
gauge factors of homoepitaxial and polycrystalline diamond ﬁlms are reported in the
range of 200 - 550 and 5 - 100, respectively [110,111]. The diamond ﬁlm gauge factors
are reported to increase with the temperature [110], grain size [111]. and resistivity
[110,111]. The electrical ﬁeld dependence of polycrystalline diamond gauge factors are
also investigated by Sahli et al. [112]. Figure 2.7 (a) shows the gauge factors of diamond
and silicon as a function of strain and (b) shows the doping dependence of gauge factors as

a function of strain reported by Taher et al. [110].

43

 

 

d
or

- + (20 chin-ca)

+(5 Olin-ca)
+(30 ODD-CI)

 

‘
O

GAUGE FACTOR

drool
TIT

W
W

W . . . .(b)

I
d
C

I

 

I
d
U

 

 

§

400 -200 0 200 400
smarmxro-G)

 

 

 

 

 

 

10‘
evrmiem 11111
I W
E 10’ evummc [1001
If \moemmn [10010050111101
‘“ 10’
0
3 310.001 omen
3 POLY-$200911!“
10‘
POLYCRYSTALUNE 20 one-ca
10° . 1 + e . .
0 2” 400 000
STRAIN ( x 10-6)

Figure 2.7 Results of gauge factors study by Taher et al.

2.6.4 Gas sensor

A Pdfr-diamond/p-diamond (MIS) structure was used for hydrogen sensing by Y.
Gurbuz et al. [113]. The I-V characteristics of diamond based MIS structure shifts from

near ohmic to rectifying behavior after exposure to 0.01 torr of H2 gas at 27 °C. The
capacitance is reported to decrease upon 11; absorption. This structure is reported to have a

high sensitivity, good reproducibility, and fast response.

2.7 Summary

A comprehensive report on the temperature measuring instruments and thin ﬁlm
heaters is presented in this chapter. The diamond properties. CVD diamond technologies.

and current status of diamond sensors are also discussed.

CHAPTER3

DLMMONDIHHHCEINFRA-
STRUCTURE

3.1 Introduction

To realize reliable and reproducible diamond micro-sensors and thin ﬁlm heaters. it is
necessary to develop diamond infrastructure technologies. The technologies developed or
used in the present study can be grouped into thin ﬁlm deposition, nucleation density
enhancement, and device fabrication. In this chapter, current efforts of development,
improvement and implementation of those technologies for realizing the diamond devices

are described in detail.

3.2 Thin Film Deposition Technology

The ability to grow high quality ﬁlms is essential for the realization of diamond
sensor/heater application. With the exception of MPCVD samples used in the nucleation
enhancement study, all diamond ﬁlms used in the present research were prepared by a hot
ﬁlament chemical vapour deposition system (HFCVD) [65]. The MPCVD samples were
supplied by professors Asmussen and Reinhard [1 14,116]. In addition to the description of
HFCVD. the design and construction of a new MPCVD system are described in this

45

46

section.

3.2.1 Film Growth

3.2.1.1 Hot Filament Chemical Vapour Deposition System

0 Description of reactor

The HFCVD reactor used in this work was designed and constructed at Michigan State
university by A. Masood, S. Sahli, and D. 8. Hong. Some modiﬁcations and optimizations
were incorporated during the course of this work to improve the reproducibility,
deposition uniformity, grth rate, and ﬁlm quality.

A simpliﬁed schematic diagram of this system is shown in Figure 3.1. The HFCVD
reactor consists of an 18-inch diameter stainless steel vacuum chamber with a 10 inch

access door. Ten 5 inch long tantalum (Ta) ﬁlaments in a series conﬁguration are

horizontally mounted on a 4.5 X 5 inch2 frame via 20 molybdenum (Mo) books. The

supporting frame consists of the combination of 2 BN bars and 2 Mo bars which also serve

as electrodes. The 4.5 X 5 inch2 ﬁlament array is utilized to ensure a uniform deposition

over a 4 inch wafer. The ﬁlament typically draws 20 - 23 Amp current at nominal
temperature of 2300 °C. The ﬁlament temperature is monitored by an optical pyrometer
through a 6 inch glass view port.

The reactant gas mixture consists of ultra-pure grade (99.995%) methane (CH4) and
hydrogen (H7). The ﬂow rates of these gases are independently controlled by MKS type
11598 mass ﬂow controllers. The operation pressure is monitored through a MKS type
122A baratron pressure gauge and is controlled by a MKS type 250 pressure controller via

a type 248A upstream valve. The base pressure is measured by a MDC thermocouple

47

vacuum gauge with a range of l - 1000 mtorr. A Varian SD-200 rotary pump is used for
evacuation. In addition, nitrogen is used to purge and backﬁll the chamber.

Either 4 inch Si wafer or graphite plate is used as sample holder. The supporting frame
is specially designed with a vertical movement mechanism to adjust the separation of
ﬁlament array and substrate in the range of 0.5 - 1 cm. Type K thermocouples are used to
monitor the both top and back side of substrate temperatures during the deposition
process. Since enough amount of heat is generated by a relatively large ﬁlament array, no

additional substrate heating is required to maintain the nominal substrate temperature of

900 °C for diamond deposition.

- Deposition Parameters

Since CVD growth of diamond is essentially in metastable phase, the selection and
consistency of processing parameters are very important to achieve reproducible high
quality diamond ﬁlms [62]. The ranges and typical values of deposition conditions used
throughout this research are summarized as follows:

(1) Gas composition: CH4 : H2 = 0.5% - 1% ( 1%)
(2) Gas Flow rate: CH4 : 1 - 5 sccm (4 seem)

H2 : 200 - 500 sccm (400 sccm)
(3) Filament temperature: 2200 - 2350 °C (2300 °C)

(4) Substrate temperature]: 850 - 950 °C (900 °C)

 

l. A temperature proﬁle of 4 inch substrate was characterized by measuring temperatures at several

points of substrate using thermocouples during a test deposition. A ~100 °C difference was found
between the temperatures measured at the center and the edge of the top side of substrate. The
substrate temperature then was monitored at the edge in the actual deposition to minimize the
blocking area by thermocouples.

47

vacuum gauge with a range of l - 1000 mtorr. A Varian SD-200 rotary pump is used for
evacuation. In addition, nitrogen is used to purge and backﬁll the chamber.

Either 4 inch Si wafer or graphite plate is used as sample holder. The supporting frame
is specially designed with a vertical movement mechanism to adjust the separation of
ﬁlament array and substrate in the range of 0.5 - 1 cm. Type K thermocouples are used to
monitor the both top and back side of substrate temperatures during the deposition
process. Since enough amount of heat is generated by a relatively large ﬁlament array. no

additional substrate heating is required to maintain the nominal substrate temperature of

900 °C for diamond deposition.

- Deposition Parameters

Since CVD growth of diamond is essentially in metastable phase, the selection and
consistency of processing parameters are very important to achieve reproducible high
quality diamond ﬁlms [62]. The ranges and typical values of deposition conditions used

throughout this research are summarized as follows:
(1) Gas composition: CH4 : H2 = 0.5% - 1% (1%)
(2) Gas Flow rate: CH4 : l - 5 seem (4 seem)
H2 : 200 - 500 sccm (400 sccm)
(3) Filament temperature: 2200 - 2350 °C (2300 °C)

(4) Substrate temperature]: 850 - 950 °c (900 °C)

 

l. A temperature proﬁle of 4 inch substrate was characterized by measuring temperatures at several

points of substrate using thermocouples during a test deposition. A ~100 °C difference was found
between the temperatures measured at the center and the edge of the top side of substrate. The
substrate temperature then was monitored at the edge in the actual deposition to minimize the
blocking area by thermocouples.

 

 

 

Mass ﬂow
controllers
H2 —~'.‘
9
. Li‘s
[ Pressure *
|
- controller N2 ’1‘

 

 

Boron powder
holder

my

~-

_'

48

_-|

Q

'.

 

Thermocouple
Sample
_

Sample holder
(4 inch wafer)

Filament

power
supply

 

 

 

 

—* Temperature
controller

 

 

 

Figure 3.1 The simpliﬁed schematic diagram of HFCVD system.

49

(5) Operation pressure: 40 - 50 torr (50 torr)

(6) Substrate to ﬁlament distance: 0.5 - 1.5 cm (1 cm)

(7) Doping source: Boron powder or 8203 plane boron source
° Operation Procedure

The operation sequence of HFCVD is reported to have a great inﬂuence on the
deposited ﬁlm quality [19]. The operation steps used through out this research are
described as follows.

Start up procedures:

(1) Samples to be deposited and doping sources (if any) are loaded on sample holder.
The distance of ﬁlaments and substrate is adjusted to be ~l cm. (The distance

varies with the sag of ﬁlaments.)

(2) Process chamber is evacuated down to a base pressure of less than 10 mtorr.

(3) Hydrogen is then introduced into the process chamber.

(4) When the chamber pressure reaches > 20 torr, the ﬁlament temperature is

brought up slowly to the nominal temperature of 2300 °C. The slow adjustment

of ﬁlament current is desired to avoid any damage.

(5) C11,; is switched on only after the substrate temperature reaches desired

temperature of 900 °C.

Shut down procedures:

(1) The CH4 is turned off to terminate the deposition processes. while the H2 remains

50

on for another 5 to 10 nrinutes. During this period, the hydrogen is the
preferential etchant to remove a thin conducting carbonaceous layer on the
surface of CVD diamond ﬁlms [19]. No conduction is found in the case of

undoped ﬁlms after this treatment.

(2) The applied ﬁlament current is decreased slowly down to zero. The chamber is

then evacuated. Samples are unloaded after the system is cool down.

3.2.1.2 Film Quality Analysis

The quality of both doped and undoped diamond ﬁlms were monitored by secondary
electron microscopy (SEM) and Raman spectroscopy. The SEM provided a direct vision
of ﬁlm surface morphology. Only samples with well faceted diamond crystals were used

for electrical measurement in this research. The Raman spectroscopy was also used to

ensure a diamond characteristic peak of 1332 cm".

3.2.2 Microwave Plasma Assisted Chemical Vapor Deposition Reactor

In addition to the HFCVD reactor, a microwave plasma assisted chemical vapor
deposition (MPCVD) reactor is designed and constructed using a microwave cavity
donated by professor Asmussen. This reactor. which became operational in June 1996.
could not be used for researches reported in this dissertation.

Based on the operational function. a MPCVD reactor can be categorized into the

following ﬁve major subsystems.

(1) Microwave cavity

51

(2) Process chamber

(3) Microwave power supply

(4) Gas ﬂow and vacuum system

(5) Computer monitor system

0 Microwave cavity

The cross sectional view of the cavity is shown in Figure 3.2. As shown. the cavity
wall is formed from a 7 inch I.D., water cooled, copper cylinder. The top end is formed
from a copper sliding short, which is electrically connected to the cavity walls via the
silver ﬁnger stock. By moving the sliding short vertically up and down. the electrical and
physical height of the cavity can be changed to ﬁne tune the microwave operating
electromagnetic mode. The cavity sits on a stainless steel. water-cooled base-plate.
Reactive gases ﬂow into the discharge zone, which is conﬁned by a 4 inch I.D. quartz
dome at the lower part of cavity. via the gas input tunnel inside the base plate. A 3 inch
high, stainless steel. water-cooled sample stage is attached to the bottom of the base plate
to hold samples up to 2 inch diameter. The water cooling capability is incorporated for
deposition at low temperature. As shown in Figure 3.3, three circular blades are designed
to ensure the uniformity of cooling effect. Microwave power is coupled into the cavity
through an adjustable coupling probe from the side of cavity. By adjusting the length of
the coupling probe. the microwave reﬂected power can be minimized for efﬁcient
deposition. The cavity used in this research is developed by J. Asmussen et al.
[1 15].

- Proceee chamber

5

N

 
  
  

  
   
 
 
 
   

 

Cavrty walls § \\\ §
\ e e \\
§ \ Slrdrng short \
\ s a
3 \ Power
Finger “091‘ coupling
Quartz dome .3 g /
E \\\\\\\\\\\\\\\\\\\\\\\\\\
—

    
   
 

Cooling water
tunnel Sample holder

////////////

/

y

 

_‘-‘_.'. .........

Gas input tunnel
O—ring

Cooling water tunnel

To process chamber

Figure 3.2 The cross sectional view of microwave cavity.

53

 
 
 
  
   
 
 
 
 
 
 
 
 
   

four 3/8" ID holes

5 3/8" apart from center
to center

(a) Top view

3/16" plate

1/16" tube

2" ID
3" high
1I12' holes
along 2.7” & 3.2”
6" dia circles
1/8" plate

 

 

 

 

 

 

3/8" Hole
Water outlet
3" long ___> - Water inlet
1’4" ID 3" long
1/4" ID

(b) Cross sectional view

Figure 3.3 The design of water cooled sample stage.

54

MPCVD diamond deposition reactor consists of a 17 inch high. 18 inch O.D. stainless
steel process chamber. The samples can be loaded and unloaded through a 10 inch access
door. The sample stage attaches to the base plate through a 7 inch opening at the top of the
chamber. Cooling water of the sample stage, vacuum gauges, and all electrical
feedthroughs are assembled through 2.75 inch Del-Seal ports.

0 Microwave power supply

An ASTeX model S-1000 power supply is used to generate up to 1 KW of 2.45 GHz
microwave power for diamond deposition. The conﬁguration of power supply circuit is
shown in Figure 3.4. A rack mount control unit contains the compact switching power
supply and control electronics which operates a separate magnetron head. Directional

coupler and circulator are used to monitor the reﬂected power.

0 Gas flow and vacuum system

Figure 3.5 shows the schematic drawing of the ﬂow control and vacuum system. The
reacting gases consist of ultra pure grade (99.995% purity) of H2 and CH4 supplied by
cylinder tanks. The MKS type 247C 4 channel ﬂow controller along with two MKS type
1159B mass ﬂow controllers are used to control the CH4 and H2 ﬂow rates in the ranges of
0 - 20 and 0 - 1000 sccm, respectively. Source gases are mixed before reaching the inlet on

the base-plate.

A Varian SD-300 mechanical pump with a base pressure of 1 x 10'4 torr is used for
pumping. Two MKS type 622A Baratron pressure transducers along with a MKS type
PDR-C-2C 2 channel read out are used to monitor the chamber pressure at the range of 1 -
1000 torr and under 1 torr, respectively. During the deposition process. the operation
pressure are measured and controlled by a MKS type 622A baratron gauge with full scale
of 100 torr, a MKS type 651 exhaust valve controller, and a MKS type 653A throttle

55

Magnetron head Directional coupler

ml.

1 ->
/ r i To cavity
Circulator

Water cooled dummy load

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

C
1:) 1:) 1:) El

 

 

 

 

 

 

 

 

 

Control and power unit

Figure 3.4 The conﬁguration of microwave power supply.

56

valve. Nitrogen is used for system vent and exhaust purge.

- Computer monitor system

A computer monitor system is incorporated in the design of MPCVD system based on
both safety and ﬁlm quality control considerations. The primary goals of this monitor
system are to monitor the operation parameters during the deposition process and to
control the shut down procedure. A block diagram of this system is shown in Figure 3.6.
During the deposition process, the operation pressure of main chamber. ﬂow rates of
reactant gases, input and reﬂected power of microwave. and temperatures of system at
several critical points are monitored by a personal computer along with read-out units of
according parameters. The proper procedure to turn off the gases ﬂow, microwave power,
and throttle valve will take place at the end of deposition process or whenever a fault is
detected.

3.3 Nucleation Enhancement Technology

Diamond has been shown to nucleate on a wide variety of materials. Due to the low
nucleation density on non-diamond materials (~ 104 cm'z), substrates are, in most cases.

treated to enhance nucleation density into the range of 107 - 1010 cm'2 [117]. The
commonly used pie-treatment techniques are described in previous chapter (2.5.1).
Because of low nucleation density. long deposition time is required to prepare continuous
diamond films. Consequently. rougher ﬁlm surface is resulted due to large grain sizes. The
surface roughness of 0.5 - 2 11m was observed for 1.5 to 4 tlm thick ﬁlms prepared using

an initial nucleation density of ~ 108 cm'z. The rough surface problems become more
pronounced for sensor/device application, since highly smooth ﬁlm surfaces are essential

in post-deposition lithography and etching of metal contacts. There are also some

57

 

 

 

 

 

 

 

 

 

Mass flow
controller
H2 Cavity
/ Discharges
CH4 I Base-plate
5‘ 1"
Pressure gauges ms °mb“

 

 

 

 

 

 

E
76

 

Y

N: vent :®:

 

 

 

 

 

Pressure controller

 

 

 

 

I / Throttle valve
Manual valve

Rotary
pump

 

 

 

 

Figure 3.5 Schematic of MPCVD ﬂow control and vacuum system.

58

 

Operating pressure
(pressure controller)

 

 

f

Input and reﬂected microwave power
(Microwave control and power unit)

 

 

Gas ﬂow rate
(ﬂow controller)

 

Local system temperature
(Temperature controller)

 

H2 gas ﬂow onloff

 

CH4 gas ﬂow on/off

 

Microwave power on/off

Computer Monitor System

 

Throttle valve open/close

V V V V

 

 

 

 

Figure 3.6 Block diagram of the computer monitor system.

 

Experimental System

 

 

59

limitation of seeding and patterning techniques used today. such as poor selectivity, ill
reproducibility. and restriction to two dimensional structures. In this section. a

development of ultrahigh density seeding technique, which enhances the nucleation up to

1.1 x 1011 cm'z. is ﬁrst described. The effect of nucleation density on the surface

roughness and growth rate of CVD diamond is also evaluated at different deposition
stages.

3.3.1 Diamond Powder Loaded Fluids (DPLF) Method

Diamond Powder Loaded Fluids (DPLF) with different carrier ﬂuids. mean powder
sizes and densities are coated on substrates to enhance the nucleation density. The idea of
this method is to spread diamond crystals, suspended in carrier ﬂuids, on the substrate
surface. During the diamond deposition process, the carrier ﬂuids are evaporated at initial
stages leaving behind the diamond particles which act as seeds for diamond growth.
DPLFs are applied to substrates through direct writing, spinning, spraying or brush-
painting. The technical details of different DPLFs are summarized in Table 3.1. The
DPLF] and DPLF2 are commercially available diamond powder suspensions, 1/40 SQG
and N-6, from Du Pont Chernicals-Repauno Plant, NJ. The DPLF3 is prepared by
suspending diamond powders into photoresist [19,61]. In contrast to other nucleation
procedures, DPLF method does not cause any damage to the substrate surface. This
method is highly compatible with standard integrated circuit (IC) processes and amenable
to patterning. doping, and coating of 3-D samples.

3.3.2 Ultrahigh Nucleation Density

Silicon wafers, cleaned by either the standard RCA or piranha solution (112804 : H202

TABLE 3.1 Technical details of DPLF seeding processes.

 

 

DPLFl DPLF2 DPLF3

Carrier Fluid Water Water Photoresist
Mean Powder Size 0.038 urn 0.10111m 0.10111m
Density 40 carats/liter 100 carats/liter 12 carats/liter
Nucleation Density 21011 cm'2 -1010 cm'2 -108 cm'2
Application Spray, brushing, Spray. brushing. Spray or spain
Methods or direct-write or direct-write
Patterning Photolithography Photolithography Photolithography

or direct-write or direct-write

 

= 60% : 40%). were coated with different DPLFS to study the primary nucleation
densities. Before the primary nucleation densities were measured. samples were treated in

hydrogen environment at 900 °C for 10 minutes to evaporate the carrier materials. which
were photoresist or pure water with a pH value of 7, and to etch any non-diamond carbon
impurities. The nucleation densities were determined both by counting the number of
particles appearing on atomic force microscopy (AFM) [118] rnicrographs and by the
cross-section analysis function associated with AFM.

An extremely high primary nucleation density of 1.1 x 10 ‘1 cm ' 2 was achieved on Si
substrates by coating with the 0.038 nricrometer powder (DPLFl). By assuming that the
diamond particles are closely and uniformly spread on the substrate surface and that each
particle occupies a square or an equilateral triangle area with a diagonal or a side of 0.038

61

am, the computed particle densities are 1.385 X 1011 cm‘2 or 1.599 X 1011 cm'z,
respectively. which are consistent with our measurements. Moreover, this density is

substantially higher than those found in previous reports [SO-56.72]. The primary
nucleation densities of DPLF2 and DPLF3 are in the ranges of 1.36 - 1.5 x 1010 cm'2 and

1.321 - 3.40 x 108 cm'2. respectively.

3.3.3 Effects of Nucleation Density

Samples coated with DPLF]. DPLF2, and DPLF3 are used to study the effect of

nucleation density on diamond ﬁlm deposition. The diamond ﬁlms are synthesized by

either HFCVD or MPCVD under temperature range of 470 - 900 °C for various periods of
time. The detail sample preparation parameters are summarized in Table 3.2. The diamond
ﬁlm quality is monitored by SEM, AFM, and Raman spectroscopy. The grain size.
nucleation density. and surface roughness were characterized directly by AFM. The ﬁlm
thickness was measured by either Dektak surface proﬁlometery for patterned samples or
SEM for un-pattemed samples. In the case of DPLF3 ﬁlms prepared by MPCVD at 850

°C for 5 and 15 minutes while the ﬁlms were not yet continuous, the thickness is
determined by averaging the thickness of those areas where most of grains were

coalescent.

0 Deposition Time for Continuous Film

The ultrahigh nucleation density tremendously reduces the deposition time needed for
obtaining continuous ﬁlms. Samples in group G], coated by DPLFl (G-la & b) and

DPLF3 (G-lc & d). are deposited by both MPCVD and HFCVD at 470 °C and 900 °C.

respectively, to study the deposition time needed for continuous ﬁlms. In contrast to the

62

TABLE 3.2 Sample preparation parameters.

 

 

 

 

Group Seeding Deposition Deposition
I. D. Fluid Parameters Time For study
G-l
G-la DPLF] MPCVD, 470 °c 1 6 hours deposition time for
G-lb HFCVD, 900 °C 2 - 6 hours continuous ﬁlm.
G-lc DPLF3 MPCVD, 470 °c 1 6 hours
G-ld HFCVD, 900 °C 2 - 6 hours
G-2
G-2a DPLFl HFCVD, 900 °C vary, to maintain surface roughness
G-2b DPLF3 HFCVD, 900 °C ﬁlm thickness comparison.
within 0.95 - 1.1
11m
G-3
G-3a DPLFl MPCVD. 850 °C 5 - 60 rrrin. ﬁlm characteriza-
G-3b DPLF2 MPCVD, 850 °C 5 - 60 rrrin. tion at different
G-3c DPLF3 MPCVD, 850 °C 5 - 60 min. deposition stages
64 DPLF] HFCVD. 900 °C 2.5 - 10 hours as above

 

1.

The 470 °C MPCVD ﬁlms were grown by Prof. Reinhard using machine developed in M. J.
Ulczynski's dissertation research.

63

DPLF3 seeded samples (G-lc & d). the high seeding density samples (G-la & b) formed
continuous ﬁlms of sub-micron thickness after the deposition time of 6 hours for MPCVD
and 2.5 hours for HFCVD even under a low growth rate of ~ 0.05 11111 per hour and ~ 0.1 -
0.15 11m per hour. Figure 3.7 shows AFM micrographs of representative samples. Various

colors in ﬁlm appearance were observed due to light interference effect.

0 Surface Roughness Comparison

The surface morphology are strongly inﬂuenced by the process parameters [67] which
can affect the surface roughness through twining, growth competition, and texture
evolution. To focus on the effect of nucleation density on surface roughness for different
growth rates and temperatures, the deposition parameters are maintained in a comparable
range for both HFCVD and MPCVD reactors.

Samples in group G—2 are prepared to compare the surface roughness of continuous
ﬁlms with the same thickness. The deposition was carried out by HFCVD under the
deposition conditions listed in Table 3.3. The deposition time is varied to maintain the
ﬁlm thickness within the range of 0.95 - 1.1 pm. The average surface roughness of G-2a
and G-2b samples, with a scan area of 3 pm square. are 29.311 and 124 nm, respectively.
Figure 3.8 shows the surface proﬁles taken by AFM. The results indicate that in spite of
the similarities in thickness and deposition conditions, ﬁlms with smaller grain sizes and

higher nucleation density are smoother than the other set of samples at early deposition

stage.

0 Film characterization at different deposition stages

To study the effect of nucleation density on ﬁlm quality, surface roughness, and
growth rate, samples in group G-3 pre-treated by three different seeding ﬂuids were

deposited by MPCVD at 850 °C at the same batch to minimize the inﬂuences of ﬁlm

,. r ,0. 0455's.)»;
0

I
(
’ a

((70

.9 ..

,

r5. .. e...
:_

s
can

:5. r.

F—* I
a

f.

(I...
. I...
. .

 

I..
a...»

f (a), (b) DPLF] and (c), (d) DPLF3 samples. Samples (11)

arison 0

Figure 3.7 Comp

& (c) are deposited by MPCVD at 470 0C for 6 hours. Samples (b) & (11)

°C for 4 hours. The DPLF] samples are

are deposited by HFCVD at 900

continuous.

65

TABLE 3.3 Deposition parameters of HFCVD and MPCVD reactors.

 

Operating Substrate
pressure temperature

Reactor Gas mixture

HFCVD 50 torr 850 - 950 °C H2: 200 sccm
CH4: 1 seem

MPCVD 90 torr 850 °C H2: 500 sccm

CH4: 15 sccm

Note

Filament Temperature
2000 - 2200 °C
Microwave power

2KW

 

 

Figure 3.8 Surface roughness comparison of l tun thickness film; (a) DPLFl. (b)

DPLF3, (c) closer view of DPLFl, (d) closer view of DPLF3.

66

morphology by different deposition conditions. Samples were characterized at several
deposition stages by AFM, and Raman spectroscopy. The representative AFM surface
proﬁles of ﬁlms after 15, 25, and 60 minutes deposition are shown in Figure 3.9.

As observed in Figure 3.9, all ﬁlms have similar surface texture and morphology in
every characterization stages as a result of identical deposition condition which Simplify
the study of surface roughness. In the meantime, samples of G-3a with the highest primary
nucleation density have the most smooth ﬁlm surfaces as compared with those of G-3b
and G-30 in all deposition stages. The well-faceted crystal structures became more
pronounced after a longer deposition time for all three sets of samples. Similar results
were observed in the case of samples in group G-4 prepared by HFCVD reactor. as shown
in Figure 3.10. It is noteworthy that the samples (G-4) prepared by HFCVD have
smoother surface than those by MPCVD regardless of the similarity in seeding density
and ﬁlm thickness. as indicated in Figure 3.11. which plots the average surface roughness
versus ﬁlm thickness for samples in both G-3 and G4. More twinning was observed in
HFCVD samples directly resulting in smaller surface grain sizes and smoother ﬁlms. This
suggests that growth conditions of different reactors play an important role on the
exhibition of ﬁlm morphology. Differences in the HFCVD samples and MPCVD samples
include the operating pressure, temperature. and ratio of gas mixture, as summarized in
Table 3.3. all of which may provide major contributions to the differences in ﬁlm
roughness [67.119].

Figure 3.11 shows the plot of average surface roughness versus ﬁlm thickness for all
samples prepared by both MPCVD and HFCVD (groups G3 and G4). It is clearly
demonstrated that the surfaces of DPLF] and DPLF2 samples become rougher gradually
as the ﬁlms become thicker. Because of the discontinuity of the lowest seeding density
ﬁlms (DPLF3) in early stage of deposition, the surface roughness does not show any

major increase. However, the surface roughness of DPLF3 samples is expected to increase

67

 

Figure 3.9 Different growth stages of diamond ﬁlms deposited by MPCVD at 850

°C for 15, 25 and 60 minutes using DPLF] (a, b, c), DPLF2 (d, e, f) and
DPLF3 (g, h, i), respectively.

68

1.000 rue/div
LIED I-ldlv

 

Figure 3.9 (cont'd)

69

 

Figure 3.10 Different growth stages of DPLF] films deposited by HFCVD at 900 °C

for (a) 2.5, (b) 5, (c) 7.5 and (d) 10 hours.

70

 

 

 

 

 

 

 

 

 

 

 

 

180 T I I I
160 - ..
E140 . ........................... issue—i.
5 _ My 1...... --------------
33120 - ~
g ’ 1
8100 "'
3 80 " ,M" q
0 ......
1— . l—,-.-A—l ...... .
o - .......... "'.’ .
8‘”. .............. .
5 40 _ ............ ,m" —o— DPLF1,HFCVD
> _ hitti- -------- a DPLF1.MPC ‘
< tar" -e-- DPLF2. MP0
2° ' g ----- --+-- 091531190 1
o 1 1 1 1 l 1 1 1 1 l 1 1 r 1 l 1 1 1 1 l 1 1 1 _1
0.0 0.5 1.0 1.5 2.0 2.5
Film thickness (um)
Figure 3.11 Average surface roughness versus film thickness for samples prepared
by both HFCVD and MPCVD.
2.5 1 - 2.5
: + DPLFZFilmThidtneee ]
. m...‘ 0911:2111». eminSiae ,. .
A . ---l--- DPLF3.FtlmThiclmeee ,w" ~
$2.0 ~ m... DPLF3.Ave.GralnSize ,w’ . 2,0

 

 

 

 

E
5*.-
v s
m .—
10 . ‘0
01.5 '_' " 1.5 c
.3 E ‘ 8.
51.0 : - 1.0 g
i L >
0.5 i (0'5 <

i l

0.0 I 1 r 1 l r 1 1 4 1 #1 l r 1 1 l 1 1 1 1 1 1 J_L r 1 1 0.0
O 10 20 30 40 50 60 70

Deposition time (min)

Figure 3.12 Film thickness and grain Size versus deposition time for DPLF2 and

DPLF3 samples prepared by MPCVD at 850 °C

71

after longer deposition times.

In Figure 3.12, the surface grain size and ﬁlm thickness of DPLF2 (G-3b) and DPLF3

(G-3c) seeded diamond ﬁlms prepared by MPCVD at 850 °C are plotted as a function of
deposition time. Note that the DPLF2 and DPLF3 have the same powder size, but
different nucleation densities. A large difference in surface grain sizes of two seeding
samples at early deposition Stages is observed and is mainly due to the discontinuity of
DPLF3 ﬁlms. In the case of DPLF2, the grains merged together after only ﬁve nrinutes of
deposition time leading to continuous ﬁlms which limit the increase in grain sizes. In the
case of DPLF3, the ﬁlms were not continuous until after 25 nrinutes of deposition. This
allows individual crystals to grow out three dimensionally to a larger size. A reduced
difference in grain sizes is observed for longer deposition times when both ﬁlms are
continuous. The results also suggest that there is no appreciable effect of nucleation
density on growth rate (in terms of ﬁlm thickness). The thicknesses of both samples are
increased almost linearly with respect to deposition time. The small difference in
thickness between two samples may be caused by the difference of initial seeding
thickness or non-uniformity in growth of diamond ﬁlms.

The surface roughness is plotted as a function of surface grain density for all samples
in Figure 3.13. It is found that the surface roughness is strongly dependent on the surface
grain density as well as primary nucleation density for all samples. Despite the differences
in deposition time, reactors. and ﬁlm thickness, ﬁlms with higher surface grain density (or
smaller grain size) Show smoother surfaces. For the same surface grain density, the surface
roughness tend to be larger for DPLF3 ﬁlms which have a far lower primary nucleation
density than the other two groups of samples.

The Raman spectra of group G—3 ﬁlms after one hour of deposition are shown in
Figure 3.14. The Raman spectra were excited from a HeNe laser with a wavelength of 633

nm and power of 25 mW. A large spot size of 250 pm was used to ensure a coverage of

72

 

 

 

 

 

m H ' "l"'l"'11 Illlr' IYI'I‘T'VITIIIII T "l"'l"'l I 11111 ' "l

100 “00 i

- o DPLFlbyHFCVDmOC ~

160 - a DPLFtbyMPCVD47OC .

A . A A I DPLF2 by MPCVD 470 c .

3‘40 _ AA 0 DPLF3byMPCVD4TOC y

F A v DPLF1byMPcvoasoc .

A 8 o DPLF2byMPcvo850c

‘20 ‘ A enrabywcvoesoc ‘

§jm .- .1

. o o .

80 - .

§ . v v .

60 - .

‘1‘: . ‘ - a
3 co

m 40 i- O '0 1

~ 60 0 ‘ .

20 i- &)b o % -t

. 00 v°v0 .

o .1.1.1.r.1111111 .1.111.1.i.1 111111 11.1.91...Plrn1-._._._|_

 

 

 

10" 2 3 4 109 2 3 4 1010 2 3 4 1011 2
Grain density (cm‘z)

Figure 3.13 Surface roughness versus grain density for all samples.

 

 

 

i

 

ity

i

 

Relatiée intens

I

:1.
ll

 

 

l a l 1 L 1 L 1 1 1 I L l 1 1

1250 1270 1290 1310 1330 1350 1370 1390 1410
Wave number shift (cm‘1)

 

Figure 3.14 Raman spectra of three different seeding samples after 1 hour

deposition by MPCVD at 850 °C.

73

fairly large number of grains. The full width at half maximum (FWHM) of Raman signal

is usually an indication of the quality of the ﬁlm. The broader Raman peak (~25 cm") of
DPLF 1 sample may be due to a higher density of grain boundaries which contain a larger

density of impurities and structure defects.

3.4 Diamond Device Fabrication Technologies

3.4.1 Patterning of Diamond Films

For many of the potential microelectronics and microsensor applications, the patterned
polycrystalline diamond ﬁlms are required. Various patterning techniques, as described in
previous chapter (2.5.4), have been developed to fulﬁl the need of desirable structures in
those applications. However, most of those techniques, which involve selective etching or
selective deposition, either cause damages on substrate surfaces or have poor
reproducibility. Recently, a novel patterning process, diamond power loaded photoresist
(DPR) patterning technique, has been developed at Michigan State University to
selectively deposit high quality polycrystalline diamond ﬁlms through standard
photolithographic processes [61]. This technique not only achieves excellent selectivity
without surface damages, but is also compatible with the existing integrated circuit
fabrication technology. This technique is adopted in the present research to pattern CVD
diamond ﬁlms. This section deals with the modiﬁcation of DPR technique to incorporate
patterning of ultrahigh nucleation density samples. In addition, problems, difﬁculties and

proposed solution are also included in this section.

74

3.4.1.1 Photolithographic Methods

The DPR patterning technique is used in this research to prepare samples coated with
DPLF3. In this technique, DPLF3 is spin-coated on substrates, Si or oxidized Si wafer in
most cases, and is patterned by standard photolithography. The spin rate was in the range
of 2000 - 4000 rpm. An enhanced diamond nucleation density was observed at lower spin
rate, however, the seeding uniformity seems inferior. An optimal spin speed of 3000 rpm

was found to possess a nucleation density on the order of 108 cm’2 with good uniformity.

The ratio of nucleation density in the undesired area and that in the desired area is in the

range of 7.57 x 10'3 - 3.38 x 10“. In the case of DPLF] and DPLF2, a combination of
seeding ﬂuids and photoresist are used for seeding/patteming. As shown in Figure 3.15,
ﬁrst, a sacriﬁcial layer of photoresist is uniformly spin—coated on the substrate with a high
spin rate of 4500 rpm. The seeding ﬂuid then is manually brushed over. The second layer
photoresist is spin-coated at a speed of 4000 - 4500 rpm to prevent the seeding material
from being removed during the developing process. Due to the poor transparency of
DPLF] and DPLF2, a full exposure of the ﬁrst photoresist layer prior to the coating of
seeding ﬂuid is helpful to improve the selectivity. However, it may cause the problem of
over developing resulting a poorly deﬁned pattern edge. The schematic of DPLFs
patterning procedure by standard photolithography is shown in Figure 3.15. SEM pictures
of diamond structures1 prepared by both patterning procedures are shown in Figure 3.16
for comparison. The wavy edge observed on DPLF 1 pattern is mainly caused by the over
developing of the ﬁrst layer photoresist and the non-uniformity of DPLFl seeding ﬂuid
coated by hand brushing. A uniform and well controlled application technique is essential

to improve the pattern quality of DPLF l and DPLF2 seeded samples.

 

l. The lst mask of second generation testchip by Izzat Taher is used to deﬁne the diamond patterns.

75

 

Si02(3 um)
Si

 

 

 

 

Oxidized Si wafer.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

DPLF3(~1ttm)
sro2 PR 5102
Si Si
PR/DPLF l/PR layers coating. DPLF3 coating.
PR DPLF3(~1ttm)
- - DPLF‘ — _
‘ ‘1 ' 3102 PR 8103 .
Si Si
Lithographic patterning. Lithographic patterning.
Diamond powders
l. . ‘ ,Sioz
l Si

 

 

After heating to 900 °C.

i

Diamond Film(l-3um)

 

 

 

 

 

- , _
7 . SiOz
Si
After diamond deposition.

Figure 3.15 Patterning of diamond film by DPLFs method (not to scale).

76

1 0011 m

2011 m

 

Figure 3.16 Patterning results of (a) & (c) DPLF] and (b) & (d) DPLF3 samples by

photolithography.

77

3.4.1.2 Direct-write Patterning Method

The nucleation enhancement and patterning techniques used today are mostly
restricted to ﬂat samples. In the case of real world micro—sensor or -heater applications,
the objects are most likely to have arbitrary shapes. In order to take full advantages of
CVD diamond devices, which can be deposited directly on the objects to minimize the
uncertainties of measurements associated with the heat dissipation and the placement of
sensing elements, there is a need to develop a seeding and patterning technique which is
compatible with standard IC fabrication processes, amenable to ultrahigh nucleation
density, capable to enhance the selectivity, and adaptable to three dimensional patterning.

In direct-write patterning technique, DPLFs are used as seeding materials and are
applied to substrate surfaces via a ﬁne spray nozzle. The minimum pattern size can be
varied by selecting the opening sizes of nozzles. In initial experiments, a capillary of wire
bonder, with an opening of ~40 11111 (1.5 mils) was ﬁxed in front of syringe needle and was
used as a spray nozzle. The width of lines manually created on Si or oxidized Si wafers are
in the range of 125 - 600 pm. In the case of DPLFl and DPLF2, a substrate temperature of

65 - 75 °C is found optimal to dry the seeding ﬂuids before it spreads out. Figure 3.17

shows a picture of diamond lines and patterns created in preliminary experiments.

3.4.1.3 Construction of Direct-write Seeding/Patterning System

A computer controlled direct-write system is therefore designed and constructed to
optimize the performance of direct—write pattern technique. The simpliﬁed schematic
diagram of this system is shown in Figure 3.18. This system consists of three major sub-
systems, namely, a three dimensional linear position stage, a air pressure controlled spray
nozzle, and a temperature controlled sample holder. The major assemblies sit inside a 24

inch I.D., 10 inch high stainless steel base chamber with a 20 inch high, 24 inch O.D.,

78

 

Figure 3.17 Diamond patterns by direct write method.

79

stainless steel bell jar cover. A Varian SD-200 rotary pump is used to evacuate the

chamber.

0 3-D Position Stage

Three AEROTECH model ATS-100 positional stages with a ﬁne linear resolution of 1
pm along with a model U12R 4-channel motor controller are used to control the
movement of the sample holder in X-Y axes and the nozzle in Z axis. In order to adopt 4
inch wafer technology, the linear movements of X, Y, and Z axes are 100, 100, and 50
mm, respectively. The 3 linear positional stages are prepared to be vacuum compatible to
minimize the outgasing and contamination. The control unit of positional stages is

interfaced with a personal computer via EBB-488 interface.

0 Spray Nozzle

A EFDl pressure controlled dispense valve system is used to spray the seeding
material of DPLFs. The system consists of a model 740V-SS ultra precise needle valve, a
2.5 oz. cartridge reservoir, two air pressure regulators, and associated assembly. The
amount of dispense ﬂuid through the 740V-SS is determined by the valve open time, ﬂuid
pressure, ﬂow control adjustment, dispense tip output size, and ﬂuid viscosity. The
operation of 740V-SS is triggered by a input gas, N2, with a optimal operation pressure of
70 - 90 psi in a rated cycle frequency greater than 400 per minute. A solenoid valve, which
is interfaced with a personal computer via an electrical relay, is chosen to control the on/

off of N2 gas, thus determine the open time of 740V-SS. Figure 3.19 shows the circuit

diagram of dispense valve system and the picture of 740V-SS needle valve.

0 Temperature Controlled Sample Holder

 

l. EFD Dispense Valve Systems Group, East Providence, RI, U. S. A.

80

Bell jar chamber

Flow
controller

IEEE-488
interface

Z-axis

4 channel
motor controller

Power supply /
Temperature controller

 

Figure 3.18 The schematic diagram of computer controlled direct-write system.

81

(a) Regulators

W

         
  

Solenoid
valve

Spray valve

Computer

 

 

 

 

Piston and

Mounting Hole
”4.28 UNF
W09 SW09
Needie Postings
VIM Outﬁtter

Fluldlnlet
1/5 NPT

 

Figure 3.19 The circuit diagram of pressure controlled spray nozzle (a) and

assemblies of dispense valve (b), & (c).

82

A circular stainless steel sample holder is built to hold samples up to 4 inch diameter.
A uniform heating over the whole sample area is achieved using four tubular heaters
connected in a parallel conﬁguration. An Omegalux CN371 temperature controller along
with a type K thermocouple are utilized to maintain a constant substrate temperature
during seeding process. In order to avoid any damage of linear positional stage caused by
excess heat, a 0.5 inch thick thermal insulating epoxy material is inserted between the
heating plate and the linear stage. A drawing of the sample holder and a schematic of

heating circuit are given in Figure 3.20.

3.4.2 Doping

CVD diamond ﬁlms deposited without intentional dopin'g are usually good insulators.
For electronic device applications, it is necessary to produce good quality semiconductive
diamond ﬁlms. Although the NH4H2PO4 and (CH3O)3P were reported to produce n-type
semiconductive CVD diamond ﬁlms, the conductivity is too low for electronics
application [120,121]. At present, boron is the only dopant successfully used to fabricate
thin ﬁlm diamond based electronic devices using CVD techniques. A number of in situ
doping techniques involving gaseous (B2H6, B(CH3O)3) and solid (8203, B powders)
sources to incorporate boron into CVD diamond ﬁlms have been successfully devised.
Since gaseous sources are highly poisonous, only solid sources are used in this research.

High purity (5N8) amorphous boron powders and B2031 plana boron sources are used
to fabricate semiconductive diamond ﬁlms via in situ doping technique in this work. Two
specially designed holders, as shown in Figure 3.1, made of 1 mm thick Mo plate were
used to introduce boron powder into process chamber. The doping sources were placed on

substrate holder during the deposition. In the earlier work, a uniform doping proﬁle was

 

l. The boron doping source is available ﬁom Techneglas Co., Perrysburg, OH.

83

 

0
0
0
0
0

0000000000

 

Epoxy thermal insulator

00000

Four holes to ﬁt 1/4" dia. tubular heaters.

 

(a) Top view (b) Side View

Thermocouple

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Temperature _
controller
I
Relay I Sample holder with
1 N 4 tubular heaters
AC power
(c) Circuit diagram

Figure 3.20 Temperature controlled sample holder.

84

conﬁrmed by secondary ion mass spectroscopy (SIMS) measurement [19]. A layer of
undoped diamond under the p—type layer was found to enhance the quality of p-type

diamond ﬁlms, measured in terms of sp3lsp2 ratio by Raman spectroscopy, even for very
high doping levels [19,122].

3.4.3 Metallization

Metallization is an important process in the fabrication of electronic devices. TilAu,
Ti/Pt, and Ti/MolAu were reported by several workers to provide good contacts with both
single crystal and polycrystalline diamonds. For the diamond sensor/heater application,
the choice of contact metal was based on (i) ohmic contacts, (ii) good adhesion to both
diamond and silicon dioxide and (iii) resistance to high temperatures.

Due to the limitation of single target metal evaporation facility, Al and Cr were tested

on the samples which diamond ﬁlms were synthesized on SiOz substrates. Both Al and Cr
with a thickness in the range of 0.6 - 1 11m were thermally evaporated under a pressure less
than 10'6 torr. Different patterns, used in the present research, were deﬁned either by
stainless steel shadow masks1 or by photolithographic processes and wet etching. Due to
the small sizes of single crystal samples and large grain polycrystalline diamond
substrates, double layers of Pthi (5000 A1500 A) contacts were thermally deposited
through copper shadow masksz.

After the metallization, samples with contact metals of Al, Cr, and Ptfl‘i were

thermally annealed at 500, 500, and 800 °C for 20, 20, and 5 minutes, respectively, in N2

by tubular furnace or rapid thermal processor (RTP). Both Al and Pt/Ti contact layers

 

1. The masks were fabricated at the Ford photographic facility.

2. The M1 double layer was deposited using the two targets thermal evaporator in Department of
Chemistry. The maximum sample size is limited to 1 inch diameter.

85

show good adhesion on diamond and Si02 (if any). The Cr contacts, especially on

diamond area, peel off after heat treatment. It is believe that Cr oxidizes immediately in

atmosphere ambient which results poor adhesion on diamond and SiOz. A non-oxidized

protective layer, such as Au, is suggested to form a reliable contact. Figure 3.21 shows
SEM pictures of three metal contacts on single crystal and polycrystalline diamond
samples after annealing. The ohmic contacts of A1 with polycrystalline diamond ﬁlms and
Pt/I‘i with both polycrystalline and single crystal diamond substrates are conﬁrmed by I-V
measurements at room temperature, as shown in Figure 3.22. A Keithley 224
programmable current source along with a Fluke 8840A multimeter are used for I-V

characterization. The voltage reading was taken right after current was applied to

minimize the self-heating effect especially at current > 10'4 A.

3.5 Summary

In this chapter, the development, modiﬁcation and implementation of diamond
infrastructure techniques are described in terms of thin ﬁlm deposition, enhancement of
nucleation density, and device fabrication.

In addition to the design and construction of a MPCVD reactor, the HFCVD
deposition technique used for preparing CVD diamond samples in the present work is
described in 3.2. In 3.3, an improved seeding technique for diamond thin ﬁlm deposition
is demonstrated. An extremely high nucleation density on the order of 1.1 x 1011 cm“2 was
achieved by coating 0.038 um diamond powders on Si substrates. One micrometer thick
ﬁlms with the mean surface roughness of 30 nm, as measured by AFM, were obtained
[73,117]. The effects of nucleation density on growth rate, surface roughness and grain

size are studied in the deposition temperature range of 470 - 950 °C using both HFCVD

and MPCVD. It is found that the surface roughness is strongly dependent on nucleation

86

20 11111

 

Figure 3.21 SEM pictures of metal contacts after annealing. (a) Al, (b) Cr on poly-
diamond and SiOz substrates.

87

100 um

i?

m

"1 I),
a ~ ,

._.\._..'.M_ —- —-_

 

Figure 3.21 (Cont'd)

88

 

 

 

 

 

 

0.03 t r 1 1 1 1 a r Y ' rﬁ 1 1 I
0.02 - .. .
0.01 r- .. . q
5:: . .
g
c O-WI -————. :.., ,2 - —.——_-.
O u
t
3 .
0
.000‘ " A 0 up q
’ -- —O— Alonsmalwm 7
—e— PWonlarpegraln
‘0'02 - 7' —D— Ptl'l'lonsiule ‘
T
.0003 1 1 4% l J l ’1 l 1 I

 

 

-100 -80 -60 ~40 -20 0 20 40 60 80 100

Voltage (V)

Figure 3.22 The I-V characterization of diamond testchips with different grain sizes
and contact metals.

density at the early stage of deposition. This enhanced nucleation density is helpful to
obtain continuous and smooth thin ﬁlms for short deposition periods, especially atilow
temperatures. The techniques of diamond electronic device fabrication, such as patterning,
doping, and metallization, are presented at 3.4. A direct-write technique is proposed in this
chapter to enhance the selectivity of patterned polycrystalline diamond ﬁlms. The

preliminary experiments show encouraging results. A computer controlled system is

constructed to optimize the performance.

CHAPTER 4

SINGLE ELEMENT DIAMOND
SENSOR/HEATER STRUCTURE

4.1 Introduction

Heat generation and temperature sensing are required for controlled heating
applications and for liquid level sensors, mass ﬂow meters, and vacuum and pressure
gauges, which operate based on the variations of heat dissipation [3]. As the materials
commonly used as heating elements lack high sensitivity to the temperature change,
different materials are utilized for heating and temperature sensing. In such a
conﬁguration, thermal and chemicai properties of materials involved, if not carefully
considered, may cause problems. It is also important to minimize the response time and
uncertainties of measurements associated with heat dissipation and placement of sensing
and heating elements. The use of a single element as a heater and a temperature sensor
may help eliminate such problems.

Due to its large band gap, high thermal conductivity, high sensitivity to temperature
change, micro-machining capability, the ability to be used as an electrical conductor and
insulator, and chemical and mechanical stability, diamond is an excellent material for high
temperature sensing and heating applications, especially under harsh environments. This

chapter describes the temperature sensing and heating properties of single-element-

89

diamond sensor/heater structures.

4.2 Experimental

4.2.1 Sample Descriptions

P-type boron doped diamond resistors with different dimensions and structures are
synthesized on oxidized Si wafers, polycrystalline and single crystal (100) type II-a
diamond substrates by HFCVD. High purity boron powder or B203 planar boron wafer
are used as sources for in-situ doping. The fabrication processes are described in previous
chapters in detail. Figure 4.1 presents the schematic designs or photograph of all seven test
structures. Table 4.1 summarizes some important parameters of samples of all test
structures. The metal contacts, both Al and Pt/I‘i double layer, are thermally deposited on
p-type diamond as electrical connections. In the case of test structure #7 samples (group
S7), the thermal annealing is done inside a thermal furnace with N2 ﬂow. For the rest of
the samples, the annealing is performed by using a rapid thermal processor (RTP) in
nitrogen environment. With the exception of test structures #3, #4, and #7 (groups S3, S4,

and S7), samples are annealed before the metallization. Al contact then is annealed at 400
- 450 °C for 5 - 10 minutes in N2.

The sample notation in the form of "SX-Y" are used through out this chapter to
describe a single sample. The "X" speciﬁes the group number as well as the test structure
of the sample. The "Y" denotes sample number with the same structure. In the case of test
structure #1 and #7 (groups 81 and S7), which contain more than one resistor per test chip,
the form "SX—RZ" is used to indicate a certain resistor by number "Z". For example, the
83-1 denotes the sample No. 1 with #3 test structure and S l-R3 refers to the number three

resistor of test structure #1 sample.

91

TABLE 4.1 List of important process parameters of test structures

 

#1

#2

#3

Test

Sample
structure group

81

82

83-1
83-2
83-3

84-1
84-2

85-1
85-2

86-1
86-2

87

Grain size

0"")

1-2

1-2

15-50
15-50
150-350

None

03-05
1-2

Doping source

Boron powder
Boron powder
Boron powder
B203
B203

Boron powder
B203

Boron powder
B203

Boron powder

Boron powder

Metal
contact

Al

Al

Mi

Mi

Annealing

650 °C. 10 min

750 °C, 10 min

800 °C, 5 min

800 °C, 5 min

800 °C, 5 min

700 °C. 10 min

500 °C. 20 min

 

Note :

1. Group 81, 82, 85, 86-2, and 87 samples are prepared by DPLF3 nucleation method.

2. 86-1 sample is prepared using DPLF] nucleation method.

3. Group 83 substrates with large grain sizes are commercially available from Diamonex
Incorporated, Allentown, PA.

4. Group 84 substrates are single crystal (100) type Ila diamond supplied by Dubbeldee
Harris Diamond Corp.,Mount Arlinton, NJ.

92

(a) (b)

 

‘ \
H

Test structure 1 (~15 cmz) Test structure 2 (~2 cmz)
Diamond resistor (1.5 cm2)

Diamond resistors (0.3x1 cm2)
(C) (d) I

Test structure 4 (0.3 cm2)

      

   

Test structure 3 (1 cm2)

\W\\\\V \
“em \ W

\\\\\\\\\\\\

Test structure 5 (chip size 3 x 6.5 cm2)
- P-type Diamond (1-1.5|.tm) * Undoped Polycgystallisne Diamond
(structure 1, 2, : 1-1. um)
Metal Contact (~0.5 um) (structure 3: 200 Ill“)
ggdéfueg 811.1% nm) - Undoped Single crystal Diamond
(structure 2: 300 um) (200 um)

\

 

Figure 4.1 The schematic diagrams of diamond test structures.

93

 

(f) ’I‘ENIPERATURE‘ TENH’ERA'" 'RE
SENSORS SENSORS
(:AmlgflL‘Fiz'rck
I 4 if‘r u.
HALL SENSOR '
PRESSURE SENSOR
ACCELEROAIETER .\l()SFli'l‘
. 3
(g) Test structure 6 (chip size 1 cm2)

10

a
u
m
a
a
as
a
%
a

m
9

a 3?!
w

a

ﬁ

g

m .
wt?

lmm

 

Test structure 7

(on the right shows the close view of diamond thermistors array)

Figure 4.1 (Cont'd)

94

4.2.2 Measurement Setups

A development in the characterization methods is important for new materials.
Initially, performance of fundamental property characterizations is essential to understand
the material and to identify applications. Once a prototype is developed, the
characterization can then focus more on device performance to optimize device
conﬁguration and structure. The characterization of single element diamond sensor/heater
structures in this research consists of temperature response characterization, heating
properties characterization, and temperature proﬁle characterization. In addition, the
effects of grain size on electrical properties and resistivity stability are studied by Hall and
R vs. Time measurements, respectively. This section describes the experimental
measurement setups used in these characterizations in detail. With the exception of the

Hall and infrared camera measurements, all measurements are performed inside a 18 inch

stainless steel vacuum chamber which maintains a vacuum of 10'3 to 10'7 torr or a static

atmosphere ambient to stabilize the temperature reading.

0 Temperature Response Characterization

It is mainly an I-V measurement under different temperatures. The resistivity of
diamond ﬁlms is determined as a function of temperature by applying a constant current
of 1 11A. Temperature is measured by type K thermocouples. The details of this setup can
be found in Table 4.2. Since temperature of samples is varied via an external 1.75 inch
diameter boralectric heating plate, a BN probe station, based on the considerations of high
temperature stability and high thermal conductivity, is custom made for R vs. Temp.
measurement. The electrical connection with remote equipment is provided through

tungsten (W) probes along with stainless steel clamping bases. Oxidized Si wafer with the

95

TABLE 4.2 Details of measurement setups

 

Characterization Equipment list

Temperature a. Keithley 224 programmable current
response source
b. HP 3457A multimeter and/or Fluke
8840A multimeter
c. Omega Digicator and/or CN9000
temperature read-out with type-K
thermocouples

Heating a. HP 712B high voltage DC power
properties supply
b. HP 3457A multimeter and/or Fluke
8840A multimeter
c. Omega Digicator and/or CN9000
temperature read-out with type-K
thermocouples
d. Boralectric heating plate

Temperature a. Variac AC power supply
proﬁle b. Fluke 77 portable multimeters
c. Inframetrics Model 600L infrared
imaging radiometer
d. Thermoteknix ThermoGram image
processing card
e. FOR.A VTG33 video timer
f. Video cassette recorder

Hall MMR technology Low temperature Hall
measurement system:
a. H-50 Hall, Van Der Pauw controller
b. R2500-2 LTHS refrigerator,
c. K-20 programmable temperature
controller

Sample placement

a. BN holder
(5.08cm x 3.81cm
x 0.635cm)

b. Oxidized Si wafer
(2.54cm x 2.54cm

x 500nm)

a. Al holder
(5.08cm x 3.81cm
x 0.635cm)

b. Al clamp
(0.9525cm x 3.81
cm x 0.3175cm)

c. Mica glass holder
(5.08cm x 3.81cm
x 0.635cm)

Hanging in the air
via two copper
clamps.

Substrate holder of
LTHS.

 

96

thickness of 500 11m is also used as an alternation for measurements with maximum
temperature higher than 900 K. A schematic diagram of measurement setup and different

conﬁgurations of sample holders are shown in Figure 4.2.

0 Heating Properties Characterization

Supplying a DC voltage up to 287 V, the temperature, applied voltage and current of
diamond sensor/heater test structures are recorded. The applied power density and
resistivity are then calculated from the collected data as functions of temperature for
power density vs. heating temperature and simultaneous heating/sensing characterizations,
respectively. Diamond samples were tested under different conﬁgurations using Al holder,
mica glass holder, and Al clamp. A1 holder, with high thermal conductivity, is used as a
heating load for diamond thin ﬁlm heater in a temperature range of 300 ~ 500 K. Al clamp
is used to construct a cantilever conﬁguration. Based on the consideration of high
temperature capability and thermal insulation, a mica glass holder is chosen to reduce the
amount of heat generated by diamond thin ﬁlm heaters loss through conduction.
Therefore, a high heating temperature of 950 K can be reached. A saw-toothed shape mica
glass structure, placed under the sample, was found to further improve the thermal
isolation. The temperatures are determined by type K thermocouples. Tungsten (W)
probes along with stainless steel clamping bases are used to provide electrical contacts
with the remote equipment. The details of equipment used in this setup are listed in Table
4.2. Figure 4.3 shows a schematic diagram of measurement setup and different sample
holder conﬁgurations.

This setup is also adopted for resistivity stability evaluation.

0 Temperature Profile Characterization

The temperature proﬁle characterization of diamond thin ﬁlm heaters were achieved

97

 

   
   
 
   
   

Thermocouple

/

 

Temperature
read-out

 

Current Voltage
source meter

 

Sample holder

 

Heatrng plate

 

 

Chamber

(a) Temperature response characterization

Heating plate $102 holder

Metal shield Test chip

   
   

Probes

(b) Top view of 8i02 holder conﬁguration

Figure 4.2 The experimental setup conﬁgurations of temperature response
characterization.

98

 

(c) BN holder

Figure 4.2 (Cont'd)

 

  

Thermocouple

 

 

 

 

Temperature
read-out

 

 

 

 

 

 

Chamber

 

 

 

Voltage
source

 

 

 

Current
meter

 

(a) Sensing/heating characterization

 

 

(c) Al holder/clamp

I

(b) Mica holder

Figure 4.3 The experimental setup for heating/sensing characterization.

 

100

utilizing an infrared camera system, consisting of (i) Inframetrics Model 600L infrared
imaging radiometer, (ii) thermal image processing system, (iii) FOR.A VTG33 video
timer, and (iv) a video cassette recorder. The ﬁrst two components, the infrared radiometer
and the image processor are used for temperature measurement. The video timer was used
to superimpose digital indications of time and date onto the transient output from
radiometer. This output was recorded in standard RS-170 format on a video cassette
recorder. A diagram of the data acquisition system is displayed in Figure 4.4.

Transient temperature measurements were made utilizing the Inframetrics 600L
infrared radiometer in conjunction with the Thermoteknix ThermoGRAM image
processing card. The system allows real-time thermal imaging of static and dynamic
events with variable image averaging, temperature ranges, emissivity settings, and ﬁelds
of view. Individual thermal images, consisting of 365x280 individual pixels can be
acquired and saved. Temperatures are calculated from the radiant intensity of each pixel
using calibration curves unique to the spectral response of the system. Over 102,000
surface temperature measurements are accessible in each thermal image.

The 8 bit infrared system can operate in a variety of temperature span settings. In its
smallest setting of 5 °C, a maximum thermal sensitivity of 0.02 °C can be obtained. The
special external optics used, consisting of a 6" close - up lens attached to the end of a 3X
telescopic lens, allows a spatial resolution of better than 1 pixel/10 1.1m in the smallest

ﬁeld of view. The smallest area that can be fully imaged is approximately 4 mm x 3 mm.

- Hall Measurement

A MMR Technology low temperature Hall Van der Pauw system is used to measured
the Hall concentration and mobility at 300 K.

101

 

 

 

Video Timer I

 

 

 

 

 

l VCFﬂ '2

lllILKII

 

 

 

 

 

Infrared Camera
Control Unit

 

 

 

| AC Power

 

 

 

 

:l Supply
I

Current
Meter

 

 

 

 

 

 

 

Sample I

 

Infrared Camera

Figure 4.4 Schematic of temperature proﬁle characterization.

102

4.3 Temperature Sensors

Due to its negative temperature coefﬁcient of resistivity, diamond can be used as a
temperature sensor [18]. In this section, the temperature response of diamond is studied
over a wide variety of sample structures (homoepitaxial and polycrystalline), grain sizes,
doping levels, and annealing temperatures. A qualitative description of results is provided

at the end of this section.

4.3.1 Contact Resistance

With the exception of the study of annealing effect on resistivity, the electrical
contacts were provided by thermally evaporated Al (on small grain ﬁlms) or PtlTi (on
both large grain and homoepitaxial diamonds). Annealing prior to electrical measurements
was performed to enhance the contact quality. The representative I-V curves of three
different metal-diamond interface structures, shown in Figure 4.5, indicate an ohmic
behavior over a large applied current range after annealing. Plot (a) and (b) present the
applied currents in both forward and reverse directions. Low contact resistivity on the
order of 10’6 Gem2 was reported using carbide forming metal (Ti) on both single crystal
and polycrystalline diamond [88,89,90]. A small applied current of 1 11A was used for the
resistivity measurements in the present study. As the contact resistance is negligible, two
point-contact conﬁguration was used unless noticed otherwise. The electrical contacts

used in annealing effect study will be discussed in next section.

4.3.2 Effect of High Temperature Annealing on Resistivity

In order to produce reliable diamond temperature sensors, the effect of high

103

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

(a)
0.03 I T I T I l I l T r l I I I 1 I I I
0.02 - " "
0.01 - T ' . -
g P O a +
ﬂ _...‘-.3- I _I_— I
E0 00| .————.— -;-__- ~- — I
h . L
a
O
-0.01 - . ° -- -
T —O— Alonsmallgraln
-a— Pt/Tlon large grain
'0-02 ' " —D-— Pt/Tlonslnglecrys ‘
0.03 1 1 1 l 1 l 1 l 1 1 I 1 I 1 l 1 l 1
-100 -8 -60 -40 -20 0 20 40 60 80 100
Voltage (V)
(b)
5E-5 - ..
4E-5 - -
3E-5 - -
2E-5 - '1
$155 - -
E ' . . ‘
g OEO _ v ' .
31E-5 - _
O - .
-2E-5 - -
-3E-5 - —0— Alon smallgraln 7 -
~ —-A— Pt/Tlon large grain ‘
-4E-5 : —1:}— Pt/Tlonslnglecrys l '
-5E-5 *- -
-1.5 -1 .0 -0.5 0.0 0.5 1 .0 1.5
Voltage (V)

Figure 4.5 The I-V characterization of three different metal contact structures.

104

temperature annealing on resistivity of diamond ﬁlms was monitored before and after a
heat treatment using RTP (for both groups 83 and 84 samples) or tubular furnace (for
group 87 samples). A 4-point probe station was utilized to measure the resistivity of large

grain samples in group 83. Due to the small size of single crystal samples in group 84 (0.3

cmz), only two inner probes of the 4-point probe station could form press contacts on the
surface of tested ﬁlms. In the case of small grain samples (group 87), evaporated Al was
used in 2-point contact conﬁguration. The details of sample description can be found in
Table 4.1 and Figure 4.1. Table 4.3 summarizes the change in resistivity of these samples

after a thermal annealing, where p, and p2 represent the resistivity before- and after-

annealing, respectively. A 460 - 985% increase in resistivity was observed for
homoepitaxial diamond samples in group 84. No appreciable change in resistivity was
observed in the case of large grain sizes polycrystalline diamond samples (group 83). The
ratio of resistivity after and before annealing of 83 samples was in the range of 0.963 -
1.415. The resistivity of small grain samples (group 87) showed a 50 - 80% decrease after

thermal annealing at 500 0C. Figure 4.6 plots the ratio of pzlpl versus grain sizes. The

present results show that the overall increase in resistivity after thermal annealing is more
prominent for single crystal diamonds. The thermal annealing was found helpful to
stabilize the resistivity-temperature characteristics of diamond ﬁlms within the maximum

annealing temperature.

4.3.3 Effects of Crystal Structure and Grain Size on Resistivity

Using single crystal substrates and polycrystalline ﬁlms with grain sizes in the range
of 0.2 - 300 pm, the effects of crystal structure and grain size on resistivity were studied to
control the doping of CVD diamond. Table 4.4 summarizes the representative results.

Several observations can be made based on the data in Table 4.4:

105

TABLE 4.3 The effect of thermal annealing on resistance

 

Sample Grain size

No. p1 (9cm) pzlpl (1101) Annealing Contact method
83-1 0.40675 1.185 15-50 300 °C, 5 min 4 point probe
83-2 2.78530 1.415 15-50 800 °C, 5 min 4 point probe
S3-3 0.46898 0.919 150-350 300 °C, 5 min 4 point probe
84-1 0.36939 10.849 Single 800 °C, 5 min 2 point probe

crystal
84-2 1.81588 5.605 Single 300 °C, 5 min 2 point probe
crystal
S7-Rl 61.7801 0.202 1-2 500 °C, 20 min Evaporated Al
S7-R3 52.7231 0.222 1-2 500 °C, 20 min Evaporated A1
S7-R6 110.568 0.480 1-2 500 °C, 20 min Evaporated Al
S7-R9 174.904 0.394 1-2 500 °C, 20 min Evaporated Al
S7-R10 167.075 0.392 1-2 500 °C, 20 min Evaporated Al

 

106

 

 

 

 

 

2 _1111 ‘I'I'I'I'I'l TIIITI 1I'l'I'l'I'lTlIlll 'I'l‘I‘l'I'l [1111‘ “'l"'
i-

F 1
101 : A:
” i * Z
- O saoargegrah) -
’ A S4(slnglecrys '
d
‘3 C] S7(smallgrah) .3
3 '7'— _ 7 .1
v 1
2 P .4
v- » 1
Q - .
\N r .
Q1oo I- .1
Z I O I Z
+- 1
l- -i
4 :_- EB: E
D 4
3 :- 1
2 :. 1:9: .‘
.1
10.1 1111 111111111111 1 11L11 1 11111111111 1 11111 1 1 11.1.1111! 111111 1 111111111.

 

100 234 101234 102 234 103 234
GrainSizeutm)

Figure4.6 Thegrainsizeeffectonredstivitybeforeandafterthermalannealing.

(1) The resistivity strongly depends on the relative position of samples with respect
to the doping source during the deposition. In deposition batch #4, samples in
group 82 with the same test structure and grain size were placed along the same
direction but different distances from the boron source. The resistivity of these
samples were found to increase as the distance from the doping source increased.
The non-uniform doping is due mainly to the concentration gradient of boron

atom in gas phase during deposition process.

107

TABLE 4.4 Comparison of resistivity for samples deposited at the same batch.

 

 

 

 

 

 

 

 

 

Batch Sample . . Distance from . . .
No. No. Gram Size (1.1m) doping source Resrstrvrty (chn)
#1 86—1 0.3 - 0.5 2 cm 6.9461
86-2 1 - 2 2 cm 27.0538
#2 S3-l 15 - 50 1.5 cm 0.4823
841 Single crystal 1.5 cm 4.0075
85'] 1 '2 1.5 cm on the ”0 3.%19
direction
#3 85-2 1 -2 2.5 cm 1.5691
S3-2 15 - 50 2.5 cm 3.9412
83-3 150 - 350 1 cm 0.4312
84-2 Single crystal 1 cm 10.1780
#4 82-7 1 - 2 1 cm 0.25395
82-1 1—2 3cm 1.1339
S2-5 1 - 2 5 cm 3.83849
TOP view Side view
Gas inlet
Doping source 1
Doping
source E Sample
Sample area f
Sample holder 3 } Pumping hose

Figure 4.7 Schematic of typical placement of samples and doping source.

(2)

(3)

108

The resistivity is also affected by the crystal structure and grain size of diamond
ﬁlms. Maintaining the same relative position to boron source in the same
deposition batch, polycrystalline diamond samples with small grain sizes tend to
have lower resistivity and higher Hall concentration1 than those of samples with
large grain sizes, as evidenced in deposition batch #1 (86-1 vs. 86-2) and in batch
#3 (SS-2 vs. S3-2). Similar results were obtained in the comparison of
polycrystalline ﬁlms versus homoepitaxial ﬁlms. In the case of both batch #2
(83-1 vs. 84-1) and batch #3 (83-3 vs. S4-2), the polycrystalline samples have

lower resistivity than the single crystal diamonds do.

In batch #2, the small grain sample 85-1 possesses higher resistivity than that of
large grain S3-1, although the distances between each sample and boron source
are the same. This may be attributed to the inhomogenerous boron distribution in

the gas phase due to the directional gas ﬂow inside the deposition chamber.

4.3.4 Temperature Dependence of Resistivity

In the practical temperature sensing application, the zero-power resistance of a

thermistor is measured to determine the temperature using a given characteristic equation
or a listed table of resistance versus temperature values associated with this particular
thermistor. The zero-power resistance is deﬁned as a DC resistance value of a thermistor
measured at a speciﬁed temperature with a power dissipation by the thermistor low
enough that any further decrease in power will result in not more than 0.1% (or 1/10 of the

speciﬁed measurement tolerance, whichever is smaller) change in resistance [123].

 

1.Due to the pattern of metal contacts, room temperature Hall measurement is only performed for

group 86 samples.

109

In the present study, the resistivity not considering the effect of resistor dimensions on
resistance is used to compare diamond resistors having different dimensions and
structures. The zero-power resistivity, determined from the applied voltage, constant
current of 1 itA and the resistor dimensions, is calibrated as a function of temperature in
the temperature range of 300 - 1000 K for samples in groups 81, 82, 83, S4, and 87. The
important parameters and schematic diagrams of samples used in this study are described
in Table 4.1 and Figure 4.1, respectively. It is found that the R versus T characteristic of
the diamond thermistor is reproducible after a thermal annealing of temperature higher
than the maximum characterization temperature. Representative curves of resistivity
versus reciprocal of temperature for samples with different substrate structures are
presented in Figure 4.8. The plot (a) shows that the initial measurement results of small
grain samples from groups 82 and 87 over a resistivity range of 0.156 - 65.493 0cm. The
initial measurement data of samples from groups 83 (large grain) and 84 (single crystal)
are presented in the plot (b). A signiﬁcant reduction in resistivity ranged from 50% to two
orders of magnitude was observed over the temperature range of 300 K to 673 K. It
suggests that boron doped CVD diamond ﬁlms are an excellent material for temperature
sensing.

Non-linear relations of the log(p) versus 1000/T were observed for both single crystal
and polycrystalline diamonds. For the small grain polycrystalline samples in groups 82
and 87, the linearity of log(p) vs. 1000/T relation seems to decrease with increasing
resistivity, as shown in the Figure 4.8 (a). A merge of log(p) versus 1000/T curves of
samples 824 and 87-Rl at temperature higher than 670 K suggests a slightly different
temperature dependence between two sample groups. Due to the different doping sources

(boron powder vs. 3203 wafer) were used to prepare the p-type diamond ﬁlms in the case

of single crystal and large grain samples, different temperature behaviors were observed

for these samples, as indicated in Figure 4.8 (b). The log(p) versus 1000/T curves were

110

 

 

(a)
Temperature (K)
1000 500 333
1 l I I T
102 E 0 82-1 + i
: x 82-2 + + ++ :
.. D 32.3 + .ﬂu VW W _
3 E 0 82 4 + V b T:
'- + $7 n10 -'
21 A 87 111 $57 5
101 V S7 F14

 

 

 

I 111111"
%E I I I I
<l<1-
<l+
r>
1>
1>
B

(E) AA o oo
c: 3 0
a 8 so
> .
a- : O O
3 ‘ 0° 0 0
(I 3 0 00C)
2:; (90 1:1 131:1 1:113

 

—I.
9..
11111111
E
XE]

11.1.1 1111111

alalsl 111111] 11

 

”Al 1 Llllll 1 l

 

d 1"

1000rr (K")

b

Figure 4.8 The representative measured data of (a) small grain polycrystalline

diamond (82 & 87), (b) large grain polycrystalline (83) and single

crystal (84) samples.

(b)

—l
uh NG¥ °-‘

Resistivity (9cm)
%

MOD

5‘3.

Figure 4.8

111

Temperature (K)

 

 

 

 

 

 

 

 

1000 500 333
: l l I l 1
t X 834 (large min) ”-1113" i
: A $2 (large grain) :l: .
~ I: 8+1 (stream it ~
' + 8+2 (esteem 1
700%”
3 Ch *9 a?“ a
- [P A 1
L Ucb I &A q
; .355 '3 ,+ g g
3 ﬁts: ”fly 3
_ + q
_ A x )8 W _
E 32” "MW
E a

1 2 3
1000/r (K1)
(Cont'd)

112

merged at 670 K for large grain samples (83-1 and 83-2). In the case of single crystal
samples (84-1 and 84-2), the merge of log(p) versus 1000/T curves was occurred at 380
K.

A large variation in the room temperature resistivity was observed for small grain
polycrystalline diamond thermistors in each sample groups of 82 and 87, despite the
thermistors in each group were deposited in the same batches. The non-uniform resistivity
distribution of resistors in the same deposition batch is discussed in 4.3.3. It is noteworthy
that the percentage change of resistivity at different temperatures is found to be consistent
for thermistors in group 87. However, no consistence of R-T behaviors was found for
group 82 samples. Figure 4.9 (a) shows the resistivity versus temperature plots for
selected resistors in the group 87. The corresponding temperature dependence of
normalized resistivity (RT/RTD), which represents a percentage change in resistivity due to
temperature, is shown in Figure 4.9 (b). The R-p/R-I-o vs. T curves of thermistors on sample

87 were consistent. The R vs. T and RT/RTO vs. T plots of samples in group 82 appear in

Figure 4.10 (a) and (b) for comparison. In contrast to Figure 4.9 (b), no consistence
behavior of normalized resistivity was found in Figure 4.10 (b).

4.3.5 Temperature Sensing

0 Characteristic equation

Due to the non-linearity of log(p) and inverse temperature relation of diamond

thermistors, the third order Steinhart-Hart equation [3,20], deﬁned by

113

\l
O

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(a) I I j r l I I I
. + -
+ O S7-Fl1
6° ' + A S7-Fl3 ‘
El S7-R4 .
50 .. w + 0 $7416 ..
v
E ’ EV + 374310]
@40- ea ~
"E 30 - $193+ 4
'5
.22 V +
8 20 ' a T]- -l
r: ”tum
+
1 n % d
0 . 0 0A O 23 AA [WEEl u- q
0 _ ‘1‘- 0 A o A o o A [1% -
300 400 500 600 700
Temperature (K)
(b) 1.1 I I l f I I T
1.0 - o sr-m -
Q - A 374-13 -
°- 0.9 - 1:1 S7-R4 1
g - o sane -
a: 0-8 " v S7419 -
E ' + 874110 4
v 0.7 " .1
, .
g We .
"9, 0.5 ~ Oct -
a t a -
IO 0.4 '- ﬂ .
o . _
g 0.3 — +<> -
g * a H
0.2 . -
2 01 Z O 8% j
- _ CDO+ cm .
0.0 L 1 l 1 1 1 1 1
300 400 500 600 700
Temperature (K)

Figure 4.9 The resistivity (a) and normalized resistivity (b) versus temperature for
selected diamond thermistors in sample 87.

114

 

 

 

 

 

 

 

 

 

 

 

 

 

 

oo
row
50
O
O
O
O
O

i
3
3

(a) 6 I I I I I I I I I I I I r I
5 l- .-
. X 82-1 .
B + 82-2
E 4 ' A V 52-3 '*
. A [:1 82-4 ‘
i A3 A $25
3 - A A 0 82-3 .
g _ m o s... _
‘13
'17) 2 r ..
G
E " -1
1 " .1
O - -
200
(b) A 1.0 r- ?) -
\° ' (>0 X 82-1
°.§ 0.9 - Xﬁg; O O .1. 32.2 a
' V 82-3
m 0'8 : x A+-l1- <> [:1 324 j
E 07 _ XXII] 19Cg + o A $25 .
' _ x O + 0 82-6 .
:E 0.6 - X V V 43331-320 0 82-7 4
== ' 0A +1 + o ‘
.2 05 ' X D 00 V 4.4-+4? d
m VVV
0 b 9 va “'58 °
h- 0.4 - Do 67 41- ~
E
E
O
2

0° 0
O

I 1

l

 

 

 

200 300 400 500 600 700 800 900
Temperature (K)

Figure 4.10 The resistivity (a) and normalized resistivity (b) versus temperature for
selected diamond thermistors in sample 82.

115

;.= a + b(lnp) +c(lnp)2 +d(lnp)3.

is adopted to describe the resistivity-temperature relation for temperature sensing
applications. The Steinhart-Hart equation is determined to be an excellent mathematical
expression of thermistors with negative temperature coefﬁcients of resistivity over a large
temperature range [3,20]. The average percentage difference between temperatures
computed from Steinhart-Hart equation and those measured directly by thermocouples is
in the range of 0.3289% - 4.8759% (1.4270 K - 25.8909 K) over the entire
characterization temperature range. Table 4.5 lists the calibration results of selected
diamond thermistors. It is worth pointing out that the reference temperature measured by
thermocouples has a certain degree of uncertainty. The limits of error for type K
thermocouples are rated as 2.2 K in the range of 73 to 273 K or 0.75% above 273 K
[21,23]. According to its speciﬁcation, the average error of type K thermocouples over the
present characterization temperature range is 4.875 K. Therefore, a precise temperature
measurement and a tightly controlled calibration temperature, which are difﬁcult to
achieve using currently available equipment, are essential to fully characterize the

accuracy of diamond thermistors.

0 Sensitivity of diamond temperature sensor

The sensitivity of a semiconductor thermistor is usually characterized by the factor B

[5] which is deﬁned by

116

TABLE 4.5 The calibration results of selected diamond thermistors.

 

Resistivity

Sample @ 300K Grain size AT AT/Tmea B
No. (am) (pm) (K) (93) (K)
Sl-Rl 0.4924 1 - 2 2.4806 0.5028 985.042
81-R2 0.5887 1 - 2 3.8809 0.7338 1015.97
81-R3 0.7925 1 - 2 1.8464 0.3747 977.467
82-1 1.1339 1 - 2 11.4208 1.9979 963.570
82-2 0.1478 1 - 2 15.2163 2.7111 384.808
82—3 0.2243 1 - 2 5.7503 0.9396 457.117
82-4 5.0337 1 - 2 8.0798 1.6527 1163.38
83-1 0.4823 15 - 50 17.7358 2.5722 471.025
83-2 3.9412 15 - 50 20.6596 3.0826 1591.50
84-1 4.0075 Single 19.3886 2.9830 846.442
crystal
84-2 10.1780 Single 17.1291 2.8256 2208.45
crystal
S7-R1 12.3600 1 - 2 1.4270 0.3289 1530.50
S7-R4 45.3008 1 - 2 3.0807 0.6433 1525.52
S7-R10 65.4931 1 - 2 4.0216 0.7974 1586.13

 

Note: 1. AT denotes the average difference of temperature calculated from S-H equation and tem-
perature measured by thermocouple.

117

R
1
ln(

(1-
TI

EL,

B:

) .

where R1 and R2 are resistance at temperatures T1 and T2, respectively. Another commonly

1:1"

used indicator of sensitivity of temperature sensors is the temperature coefficient 01 which

is deﬁned as

ldR
= 1127* = 4377

It is usually expressed in terms of the percentage change in resistance per degree. Since a
is a strong function of temperature for semiconductor thermistors and it doesn't represent
any more information than B factor, B is chosen in the present study to evaluate the
sensitivity of diamond thermistors. Based on the deﬁnition of B factor, it is also a strong
indicator of temperature behavior of diamond thermistor in a speciﬁed temperature range.
The room temperature resistivity and B values calculated over 300 - 683 K of selected
diamond thermistors are listed in Table 4.5. The results suggest that B, as well as
temperature behavior, strongly depends on the resistivity and crystal structures. In the
present study, B values of all thermistors range from 349.334 to 2208.452 K. The
corresponding room temperature resistivity and resistance values are in the ranges of
0.1507 - 65.493 9cm and 0.977 - 509.5 KQ, respectively. The reported B value of
diamond thermistors is in the range of 940 - 5500 K [97,124] as compared to 2000 - 4000

K for conventional metal-oxide thermistors [3] for the temperature range of 300 - 67 3 K.

118

However, high sensitivity (B = 5500 K) is achieved at lower doping level, the
corresponding resistance values are unacceptably high (> IMO) for practical applications.
The low resistivity is helpful to maximize applied power and heating temperature for

heating application.

0 Long term stability

Due to the difﬁculty of maintaining a constant environment temperature other than
room temperature, the long term (48 hours) stability of resistivity are studied only under
room temperature for selected samples in group 83 (large grain) and 84 (single crystal).
The attempt to maintain a stable environmental temperature for calibration was achieved

by performing the experiments inside a 18" stainless steel vacuum chamber under a

pressure range of 10" torr. The sample holder was arranged to have the maximum thermal
isolation in such a conﬁguration that a mica glass holder was suspended by two ceramic
bars to minimize the contact area with chamber body. An average resistivity variation in

the range of 0.1367 - 0.4778% is observed which corresponds to an ambient temperature
change of 1 - 2 °C. The substrate temperature monitored by thermocouples during each of

the characterization periods also showed a small change in 1 - 2 °C with time. It is
reasonable to believe that the resistivity of diamond thermistors maintains a constant value
at a speciﬁed temperature close to room temperature. Figure 4.11 shows the representative
results of long term stability characterization. Due to the potential possibility of property
changes or failures in sensor material, electrical lead, and contact interface, a similar study
at high temperature range is essential to ensure the long term reliability of diamond

thermistors.

119

 

 

 

 

 

 

 

' T I I I I I ﬁ
60 - -
' 1
50 - ..
A . _°'- 834M”)
g 40 . —0— 84-1 (01mm .
V —&'— “2(Wm)
o
2 30 - -
ﬂ .
g 20 -
a: .
10 - .
" .1
0 - -
1 1 l 1 l 1 l 1 1 1
0 10 20 3O 40 50
Time (hours)

Figure 4.11 The long term stability of diamond resistivity.

4.3.6 Discussions

A better understanding of conduction mechanisms in CVD diamond is essential to
properly interpret the temperature response of diamond ﬁlms and to maximize the
performance of diamond thermistors. Due to its large bandgap of 5.45 eV, the temperature
ranges of semiconducting conductions in diamond are higher than those in Si. The
existence of relatively high density of defects and impurities in CVD diamond results in

increased charge carrier scattering and possibly the impurity/defect band conduction. This

120

makes the understanding of charge carrier transport mechanism in diamond difﬁcult. In
the case of polycrystalline diamond, the presence of grains and grain boundaries makes
the conduction even more complicated.

For p-type semiconductive diamond, valence band conduction, impurity band
conduction, and hoping conduction are reported [95,125]. The extrinsic conductivity,
under the conditions of partial compensation and nondegeneracy, may be expressed as

sum of three terms

El E2 E3
0' = olexp(-n)+ozexp(-ET)+03exp(- )

where 01, 02, and 03 are constants and E1, 1'52, and E3 are the activation energies
associated with the valence band conduction, the impurity band conduction, and the
hopping conduction.

A comprehensive grain-boundary trapping theory, originated from the study of
polycrystalline silicon [95,126], may be helpful for understanding the grain boundary

conduction of CVD diamond. This theory states:

(1) For p-type semiconductor;

(i) Dopant (boron) is evenly distributed throughout the ﬁlm but holes are trapped
at the grain boundaries which contain high densities of traps.

(ii) Due to the presence of trapped holes, a region of width (0 into the grain is

depleted and the ﬁee carrier concentration in the ﬁlms is reduced.

(iii) The buildup of charge at the grain boundaries leads to the formation of a
potential barrier, which reduces the mobility of free carriers crossing the grain
boundaries.

121

(2) In the case of n-type material, dopant atoms, such as P, segregate in the grain
boundary region.

In the case of CVD polycrystalline diamond, non-diamond carbon present at the grain
boundaries may have higher conduction than that of grains. Most studies tend to suggest
that the charge transport in p-type CVD diamond results from valence band conduction,

and impurity/defect band conduction at temperature above 300 K.

0 Effect of high temperature annealing on resistivity

In our study, the thermal annealing is found helpful to stabilize the resistivity-
temperature response of CVD diamond. A similar observation was also reported by
several researchers and was attributed to dissociation of hydrogen from the diamond
surface which cause anomalous or unstable conduction [127]. The hydrogen is also
believed to exist in CVD diamond ﬁlms and to passivate the deep traps [128]. The thermal
annealing causes hydrogen to outdiffuse resulting in an increase of trapping density
leading to increased capture of holes. An increase in resistivity of diamond ﬁlms was
observed for both undoped and B-doped polycrystalline diamonds after annealing [129].
This effect is reported to be more prominent for low doping level and high quality (in term
of grain boundary density) diamonds [130].

In the present study, a tremendous increase in resistivity was found for single crystal
diamonds (group S4) after annealing. The increase of resistivity is believed to be due
mainly to the reduction in the concentration of free carriers due to traps vacated by the
hydrogen outdiffusion after annealing. The small change of resistivity after annealing
found in the large grain samples (group 83) suggests that the conduction of polycrystalline
ﬁlms may be dominated by the conduction through grain boundaries, which may not be
affected by the outdiffusion of hydrogen [130]. In addition, the reduced annealing effects
on the resistivity of highly doped samples (SB-1 and 833-3) were observed as compared to

122

that of lightly doped sample (SB-2). one may assume that the average density of traps
generated by hydrogen outdiffusion is same for all doping levels [130]. Therefore, the
annealing should affect the resistivity of highly doped samples to less extent. In the case of
small grain samples (group 87), two thermally evaporated Al strips were used as electrical
contacts. Therefore, the observation of decreased resistivity after annealing may be
attributed to the reduction of contact resistance of Al pads after annealing. In order to
quantitatively explain the observations in the current research, it is important to know the

conduction mechanism at grain boundary and its contribution.

0 Effects of crystal structure and grain size on resistivity

As described in section 4.3.3, the resistivity of diamond resistors prepared within the
same deposition batch strongly depends on the grain size and crystal structure. The
resistivity of boron doped diamond ﬁlms is determined by both carrier concentration and
mobility. The Hall mobility of homoexpitaxial diamond is reported to be an order of
magnitude higher than that of polycrystalline diamond (~1ooo vs. >100 em2V'1s")
[42,95,129]. Therefore, the relative low resistivity of polycrystalline samples found in the
present research may be explained by the higher activated carrier concentration due to the
dominant impurity/defect band conduction.

The impurity/defect band conduction mechanism was evident for both boron doped

polycrystalline and homoepitaxial ﬁlms with doping levels of 1.5 x 1020 cm'3 and 4 x 1019

cm'3

, as measured by SIMS, respectively [95]. It is more prominent in polycrystalline
ﬁlms due to the high incorporated boron concentration. It has been reported that the
polycrystalline ﬁlms incorporated 2 - 4 times more boron than homoepitaxial ﬁlms within
the same growth batch [95]. The corresponding room temperature resistivity and Hall

concentration for polycrystalline and homoepitaxial ﬁlms were reported to be ~O.2 Gem

and 3.7 x 1019 cm‘3, and ~60 0cm and 6.8 x 1015 cm’3, respectively. The room

123

temperature resistivity of both polycrystalline and homoepitaxial ﬁlms used in the present
study are in the comparable ranges with the reported values. It is likely that the impurity
conduction actually takes place, especially for polycrystalline ﬁlms. The extremely high
activated carrier concentration of polycrystalline ﬁlms caused by the reduction of
activation energy due to impurity/defect band conduction is therefore expected resulting
in low resistivity as compared to that of homoepitaxial ﬁlms.

The boron incorporated rate for (111) growth planes has been reported to be higher
than that of (100) growth planes [129]. The boron-doped homoepitaxial diamond ﬁlms
used in this research were grown on (100) oriented diamond crystals. The B-doped
polycrystalline diamond ﬁlms had dominant grain orientation of (111). Whether the lower
resistivity of polycrystalline ﬁlms can be attributed to the orientation effect is not clear
and need further study. Enhanced boron concentration along the grain boundary may also
be responsible for low resistivity. For p—type polycrystalline diamond ﬁlms, the
segregation of boron at crystal defects and grain boundaries was suggested resulting in
high local concentration [95].

Non-diamond carbon presented at the polycrystalline diamond grain boundary may be
electrically conducting [131,132]. Therefore, depending upon the level of intra-grain
conduction versus grain boundary conduction, the conduction of polycrystalline diamond
ﬁlms can be dominated at grain boundary region. The results of even lower resistivity for
smaller grain size polycrystalline samples, which consist of larger proportion of grain

boundaries, than that of large grain samples further support the above conclusion.

- Temperature dependence of resistivity

The non-linear log(p) vs. 1000/T relation observed in the present study also suggests
the existence of multiple conduction processes in CVD diamond ﬁlms. For boron-doped

CVD diamonds, the valence band conduction mechanism has been identiﬁed for

124

homoepitaxial ﬁlms in the temperature range of 300 - 600 K with an activation energy of

~03? eV [95]. The impurity band conduction has been suggested to dominate the
conduction in polycrystalline diamonds at the boron concentration higher than 2 x 10 2°

cm‘3, as measured by SIMS [95,125]. The hopping mechanism became evident at
temperature below 280 K [95]. In addition, the presence of defects, impurities,
dislocations, and grain boundaries, in the case of polycrystalline ﬁlms, can introduce
different scattering mechanisms which may strongly alter the conduction process. With
the exception of hopping conduction which becomes more prominent at temperature range
lower than the current characterization temperature range, all other conduction processes
may play a role in our samples.

In the mono- and poly- crystalline films, different temperature behaviors were
observed in Figure 4.8 which may be attributed to the presence of both valence band
conduction and impurity band conduction. In order to further illustrate the differences in
resistivity-temperature behaviors of single crystal and polycrystalline samples, selected R
vs. T curves of both mono- and poly-crystalline samples in Figure 4.8 with comparable
room temperature resistivity values of ~4.5 and ~10.5 Qcm are re-plotted in Figure 4.12.
With comparable grain size, polycrystalline samples with low resistivity values possess
smaller slopes of log(p) vs. 1000/1' curves as compared to those of high resistivity
samples in the low temperature ranges. For the small grain samples, the slope of low
resistivity sample (82-4) is smaller than that of high resistivity sample (S7-R3), as shown
in Figure 4.12. The similar result was also found in the case of large grain samples of S3—l
and S3-2 at temperature less than 625 K. This suggests that the impurity band conduction,
whose activation energy is less than that of valence band conduction, plays a more
important role in the conduction of sample with lower resistivity (thus higher dopant
concentration). The change of slope of S3-1 sample at temperature higher than 625 K may

be due to the increased valence band conduction at high temperature. In the case of single

125

crystal diamonds, similar result was observed for samples S4-1 and S4-2. The transition
occurred for sample S4-l at ~570 K may be due to the increased contribution of valence
band conduction at high temperature range. With the exception of samples S4-1, the
activation energies, as indicated by the slope of log(p) vs. 1000/T curves, of all samples in
Figure 4.12 follow the order of single crystal, large grain, then small grain ﬁlms. It seems
to agree with the conclusion that impurity/defect band conduction is more prominent in
polycrystalline ﬁlms due to the high incorporated boron concentration. In spite of the
nonlinear relation of log(p) vs. inverse temperature, a better linearity is observed for
polycrystalline samples (82-4, 83-2, & S7-R3) as compared to single crystal diamonds
(S4-1 & S4-2) in Figure 4.12. The superior linearity of R-T dependence suggests that
polycrystalline diamond is preferable to single crystal samples for thermistor
application.

The different R vs. T behaviors found for the small grain polycrystalline samples in
groups 82 and 87, as shown in Figure 4.9 and 4.10, suggest that the ﬁlms quality of
diamond ﬁlms alter the R-T behaviors of diamond thermistors. Due to the natural sag of

4.5 x 5 inch2 horizontally mounted ﬁlament array equipped in HFCVD reactor, a non-
uniform temperature distribution of substrate holder was resulted which introduced a
perturbation of deposition condition. This caused differences in surface morphology and
crystal structure of ﬁlms deposited in the same batch. In the case of group S7 samples, the

thermistor array, which contains 10 resistors, is occupied within a relatively small spread

area of ~1 x 3.5 mm2 on each test chip (see Figure 4.1 for details). Only a small variation
of deposition condition in this case is expected thus the similar diamond quality is resulted

for every thermistors on the S7 test chip. In contrast, each thermistor on group 82 samples

occupied an area of 1.5 x 1.5 cm2. The wide spread deposition position of nine samples in
the same batch may introduce difference in ﬁlm quality for each thermistor. Different

resistivity-temperature behaviors observed for each samples in group 82 may then be

126

Temperature (K)

 

 

 

 

 

 

 

 

 

1000 500 333
I l I I l r | I l I I

10‘ : X :
E 3 3
_ o q
4 g + j.
E O ‘3
o 2 E 1
c: t. s
v b . 4
2‘ - '6 E13 -
.516) _, gt 3.0 -
5 : 0°43 ' 2pm 1
m '- .,(V “In” ~ ‘ g. I A
a) 4 ‘ Q“ 5E I n “ ‘ ‘ ‘1
m a’x >. ’ Lumi‘ A “A": - 3
3? :' 99‘2" “ :
. O ‘ , A? .
2 _' I lytfjll -

~ I Trail“ '

- ”it”:
101 - A" -
— l l L l l l l r l r I -
1 O 1 5 2.0 2.5 3.0 3.5

Figure 4.12 The comparison of temperature response of single crystal and
polycrystalline diamond samples with compatible room temperature

resistivity values.

127

attributed to the difference in ﬁlm quality. The important parameters of ﬁlm quality
include crystal structure, surface morphology, grain size, grain boundary volume, defects,
and impurity concentration. However, further study is necessary to identify the parameters
of ﬁlm quality that dominate the temperature behavior of diamond thermistors. By tightly
controlling the deposition conditions and thus diamond ﬁlm quality, a universal equation

can be used to describe the RT’RTo vs. T relation of thermistors fabricated in the same

batch despite the difference in room temperature resistivity. It is helpful in the practical
applications, as no individual characterization is needed for every thermistors. It is
noteworthy that the difference in the annealing temperature and procedure of sample

groups S2 and S7 may also have an impact on the R-T behavior. In the case of 82 samples,

a 10 minutes annealing at 750 °C in N2 was performed prior to the metallization. The Al
contacts were then annealed at 400 - 450 °C for 10 minutes in N2. Only a 20 minutes 500

°C annealing was performed for S7 samples with Al contacts deposited. However, the
effect of annealing temperature and procedure is not clear at this point.

The B values of thermistors in groups 81, 82, S3, S4 and S7 are plotted as function of
resistivity in Figure 4.13. Basically the B value indicates the sensitivity of thermistor to
temperature change. It is also a indication of activation energy. With the exception of
groups 81 and S7, the B values increase with the room temperature resistivity of
thermistors for groups 82, S3, and S4, as indicated in Figure 4.13. For heavily doped
samples, the low B value which relates to low activation energy seems to indicate that the
impurity/defect band conduction may be dominant. In addition, the degree of B value
increment seems increasing with grain size, i.e., the effect of doping level on the
thermistor sensitivity is less pronounced for polycrystalline ﬁlms than for single crystal
diamonds. The result suggests that with a speciﬁc sensitivity requirement, the resistivity
of polycrystalline diamond thermistors can be varied in a large range to fulﬁl the

requirement of a particular application. The similar B values of SI and 87 samples

128

represent the similarity in temperature behavior as described previously.

 

2m II """"”'""| I U111 ' ‘ 'U"'I"'I"'l I 11" ' 111"‘l"'l"'l ITIIT

 

 

 

2200 t » 0 ‘

- o 81(smallgraln) -

2000 - A 82(de grain) ':

'_ X 83(lar90 Drain) .

‘°°° . o mangle .
1500 _ Y 87(emallgraln) X V V y «1

g 1400 - ‘
m. r "
1200 - A '

. A i

1000 f 00 o A j
800 ' A O ‘.

600 - x j

200 - j

 

 

 

104 2 345 100 z 345 101 2 345 102

ResistIVity (mm)

Figure 4.13 The B versus resistivity

4.4 Thin Film Heaters

The characterization of diamond thin ﬁlm heater includes heating temperature, power

density and heating uniformity. Group 81, 82, and S3 samples are used for heating study

in this section.

129

4.4.1 Heating Temperature and Power Density

Supplying DC voltage up to 287 V, the temperature of the samples, subjected to Joule

heating, was measured as a function of applied power density by thermocouples at

different points of resistors. Representative curves of data collected for temperature

measured at the center of each resistor are shown in Figure 4.14 for three test structures

under different conﬁgurations. The details of sample structure and measurement

conﬁguration can be found in Table 4.1, Figure 4.1, and Figure 4.3, respectively. Several

observations and explanations can be made based on the heating temperature data of these

diamond heater structures:

(1)

(2)

The heating temperature shows a linear increase with applied power density up to
550 K in all data sets. A non-linear relationship between power density and
resulting temperature was observed at high heating temperature range. Based on
the Stefan-Boltzmann law, the energy emitted from an emitting surface is

proportion to 4th power of object temperature as described by [133],

w=eor“

Where W is the rate of energy emission per unit area, T is the absolute
temperature of the body, a is a universal physical constant, and 0’ is the
emissivity of the material. Therefore, the inferior heating efficiency at high
temperature range can be explained by an increased amount of heat loss through

radiation.

The heating efﬁciency mainly depends on the size of heat capacity and the

amount of heat loss through conduction. Sample SB-l shows the best heating

130

 

 

 

 

 

 

 

 

1” j V U T 1 ﬂ U ‘ I 1 1 I ' ﬁ I 1 T 1
' f A st-n1.Nduttp.mlr
900 - X 0 81-81.Ndarrlp.hvaeurm -
. 0 St-R1.onAl,lndr
+ 82-1.onuica.lnvaeurm

9800 ' x C] 82-2.onrriea.lnvaum '
V . V 824.th\~:”!| ,
2 Q- x 834.0”th

700 - O A -
:2: e°°
a ' 0° .
E 600 - o 0&3
9 ' 0° ‘
0500 " O '
r: 0 0 .
'= ' mg“ 0
8 400 - -
I . 8 00° .

300 - 86"” -

m 1 l l r l r l r l J l 1 l r l i l r l

010 20 30 40 50 60 70 80 90100
Power density (watt/cmz)

Figure 4.14 Temperature versus power density under different measurement
conﬁgurations.

efﬁciency due to the smallest heat capacity of the sample substrate and the least
conduction and convection losses of mica glass holder and vacuum environment,
respectively. Under the same experiment conﬁguration as sample 83-1, a
relatively good heating efﬁciency of test structure #2 samples (82-1, -2 & -3) was
also observed. The slight different results of 82-1, -2, and -3 samples are believed
due mainly to the small variation of Si substrate dimensions. In the case of test
structure #1 sample (S 1), since heat is generated by one of the three resistors (R1)
only, the relatively large Si substrate and both Al clamp and Al holder represent

large heat capacity to the diamond heater. The poor heating efﬁciency then is

(3)

131

resulted. Due to the smaller contact area with Al clamp in the cantilever
conﬁguration than that of on Al holder conﬁguration, higher heating efﬁciency

of Al clamp conﬁguration is also observed.

A slightly less effective heating is observed when the measurements of 81
sample is performed in the air as compared to that under vacuum in the
temperature range of 300 - 700 K. This result suggests that the heat loss during
the characterization is dominated by the conduction through the Al clamp and the
heat loss via the natural convection of the static air inside the chamber is
relatively small. The convection heat loss into the air can be simply described by

the Newton's law of cooling [133]:

q = hAAt (Btu/hr)

Where q, A, and At denote the rate of heat flow, contact area, and temperature
difference of heated object and surrounding air, respectively. The h in the above
equation is known as the heat transfer coefficient, ﬁlm coefﬁcient, or ﬁlm
conductance. In the case of horizontal plates with heated upper surface, the h of
free convection for air can be approximated by [133]

h = 027(95’) (Btthr-ftZ-F)

The rate of heat flow by conduction across an area of A is given by Fourier's law

as [134]

132

At
4 = kAKL (Btu/hr)

The constant k is the thermal conductivity of material through which the heat
transmits. By a simple calculation, the convection and conduction heat losses of
81 sample ﬁxed on the Al clamp at 673 K are 7.96 and 4382.4 Btu/hr,

respectively, which agree well with the experimental observation.

By varying the doping level and resistivity, the applied voltage required to reach
certain heating temperature can be optimized for a particular application. Table 4.6
summarizes the resistivity at room temperature and applied voltage and power needed to
reach 480 and 680 K of all samples for comparison. In the case of samples from test
structure #2 group (82-1, -2, & -3), relatively low voltage is required to drive the sample
with low resistivity to a certain temperature, in spite of the applied powers are in the
compatible range. The ”~" sign denotes the approximated value computed from the
experimental data.

A heating temperature in excess of 950 K was measured at the center of the test

structure #3 sample while a power of 8.6 watts was supplied under vacuum environment

(<10‘5 torr). The maximum heating temperature of test structure #1 and #2 is limited by
the diamond oxidation temperature of ~ 873 K and the Al melting point of 933 K. If
diamond is heated to temperatures above 873 - 973 K in oxygen, it oxidizes [34] and the
ﬁlm may be completely etched in a few minutes. An A] contact failure was observed for
heating over 850 K even under vacuum. Thus, a higher maximum heating temperature is
possible if (1) a high melting point metal contact, such as: Pt/I‘i or Aqu‘i, is used, and (2)

a passivation layer, such as Si02 is present or if the heater is operated under non-oxidizing

133

TABLE4.6 UstofappliedvoltageandpoweratMdeSOKfordiamond

 

 

heaters with different resistivity.
Sim?“ R3313? 480K "wins WW 680K
No.
‘0‘“) Voltage (V) Power(W) Voltage (V) Power(W)

W
32-1 1.1339 103.5 4.59 ~140 ~17

s2-2 0.1478 46.55 3.31 73.05 11.65

823 0.2243 59.45 4.16 93.3 16.99

S3-1 0.4823 ~47 ~2.27 ~66 ~6

 

environment. A relatively high power density of 93 watt/cm2 was used in the case of test
structure #1 sample due to a large thickness of Si substrate (500 tun) as compared to that
of diamond ﬁlm (2 - 3 urn). The maximum power and operating temperature of
commercially available thick ﬁlm heaters (nickel, Al, or carbon steel alloy) [8,11] are in
the ranges of 0.775 - 38.75 watt/cm2 and 573 - 1473 K, respectively. The high breakdown

voltage of diamond heater allows a high voltage and a large power density to be applied
for fast heating speed even under a large load.

4.4.2 Temperature Proﬁle of Diamond Thin Film Heaters

It is of advantage to characterize the temperature proﬁle of whole heater substrate in

134

stead of only single point temperature, especially when uniform heating and precisely
controlled temperature are required. In this section, the temperature proﬁle of diamond
heater substrates is studied by using an infrared camera1 system. The information of
experimental setup is described in 4.2.2. The samples were suspended vertically using two
copper clamps for support as well as electrical contacts. With this arrangement, heat was
dissipated by conduction through the electrical contacts and by natural convection. While
applying a AC voltage of 100 V, the substrate transient temperature proﬁles of samples
S2-6 and S3-l are recorded.

The thermal imaging radiometer responds to the sum of the emitted, reﬂected, and
transmitted energies emanating from the object of interest. This combination of energies is
called the target radiosity. To obtain the target temperature, the emitted energy must be
determined from the total energy of the incoming radiosity. The result is then scaled up by
an emissivity correction factor to obtain a blackbody equivalent value. This value is then
converted to a temperature using the calibration curves within the imaging software.
Therefore, before accurate thermal images can be acquired, the emissivity of the diamond
samples must be determined.

For the sample S3-1, the surface of the diamond substrate was such that uniform
emissivity could be assumed. The emissivity was determined by comparing the energy of
two radiating bodies transmitted through and reflected by the surface. Using this method
the emissivity of sample 83-1 was found to be 0.64. The maximum heating temperatures
in excess of 673 K was reached after 65 second of heating with a applied current value of
75 mA. The one dimensional temperature proﬁles taken at the center of substrates
perpendicular to the direction of the current ﬂow are plotted in Figure 4.15 (a) for every 10
to 15 second of heating period. Plot (b) shows two dimensional temperature distribution of

SB-l sample after 65 seconds of heating. Due to the high thermal conductivity of diamond

 

l. The infrared system is available in Dept. of ME. All the measurements are performed by Matt
White under the supervision of Dr. John McGrath.

135

substrate, the substrate was relatively isothermal. The low temperature region appeared at
the lower portion of the infrared image was due to the presence of one copper clamps.
In the case of sample S2-6, the diamond pattern of 15 x 15 x 0.003 mm was deposited

on 300 um thick Si wafer with a 3 pm layer of Si02 in between. Two Al contact regions,

deposited directly on the substrate, created non-uniformity in the surface emissivity. The
initial experiment shows a relatively low temperature reading on the metal contact areas as
displayed in Figure 4.16 (a) and (b). Plot (a) shows the substrate temperature proﬁle along
the current flow direction. The images of two A] contact areas and copper clamps are
clearly illustrated in 2-D surface temperature proﬁle of plot (b). These artiﬁcial results are
introduced by the difference in the emissivity of diamond and Al contacts.

A thin coat of ﬂat black paint was then applied on the sample surface to reduce this
non-uniform surface emissivity. The emissivity of this black paint was assumed to be
0.95. Figure 4.17 (a) and (b) show the one dimensional time varying substrate temperature
perpendicular to the current ﬂow direction and two dimensional temperature plot,
respectively. As we can see in Figure 4.17 (b), with the exception of the presence of two
copper clamps, the surface temperature was relatively uniformly distributed. The heating
temperatures saturated at 440 K after about 65 second of heating and the corresponding

current values were in the range of 49 to 52.2 mA.

4.5 Heater/Sensor Structures

Samples with test structure #1, #2, and #3 are used in this section to study the

simultaneous sensing performance of diamond heater structures.

136

 

 

 

 

 

 

peratld're (K)

 

Tern
%
‘3

00
8

 

lrrI rrrrrrrrrrrrr

 

 

 

 

 

Position
0))

mm: W743 311.002 $21.95 38138 342.001 332.314 W 37234

 

Figure 4.15 The temperature proﬁles of sample S3-1.

137

 

 

 

 

 

 

 

 

 

Position
0))

mm: 75.602 80.5549 97M 103.500 119.272 1&1” 141.035 151.909

 

Figure 4.16 The temperature profiles of sample 82-6 before the black paint coating.

138

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

(a) 450 , . I v I r r f I ' I '
430- meal
1' I‘
410 - 4; . .. .. ; ""'""W 4:
WWW “ “ “ W '
A390 V30 -
2370: mulla-
m A A
“350- ‘ " W
s _W s
333011” -
03 - 0 58» W
rawew ; :2: a
I 5090c ‘
270- + 105:» '
250 l l J I J 1 1_L L
Position

(1))

mm 19.015 19.010 147322 15.07 1w

 

Figure 4.17 The temperature proﬁles of sample 82-6 with black paint coating.

139

4.5.1 Diamond Resistors for Temperature Sensing

The resistivity-temperature relation of every sensor/heater test structures used in this
section is ﬁrst calibrated followed by the procedures described in section 4.3.4 and 4.3.5.
Figure 4.18 shows both experimental data and the ﬁtted curves of resistivity-temperature
characteristics for all samples. The ﬁtting results will be used to determine the temperate
during the heating application. The evaluated coefﬁcients (a, b, c, d) of 3rd order
Steinhart-Hart equations are listed in Table 4.7.

4.5 .2 Self-sensing diamond heater

During a heating process, the resistivity, determined from resistor dimensions and
applied voltage and current, is computed to determine the heating temperature via
Steinhart-Hart equation with pre-deterrnined coefﬁcients (as described in the previous
section). The results of such a heating/sensing experiment are shown in Figure 4.19. The
average differences of the temperature determined by self-sensing of diamond thin ﬁlm
heaters and the temperature measured by thermocouple are in the range of 1.05 - 3.6%
over the entire characterization temperature range. The details are listed in Table 4.7. It
may be noted that the sensitivity of diamond sensors decreases as temperature increases.
This results in an increase in the temperature difference at high temperature range as
shown in Figure 4.20 (a). However, the percentage of temperature difference seems to
remain in a comparable range over the entire heating temperature range as illustrated in

Figure 4.20 (b).

4.6 Summary

A single CVD diamond resistor is used as a thin ﬁlm heater and a temperature sensor

10°

8

A 5

g 4

9i 3
a?

.2 2
3.7.5
(I)
o

m 1

10'

6

5

4

2

140

 

 

 

 

 

 

 

 

r , l l I l r ﬁ r
'-
v
I- V at
I- -t
_. v 0 814-11 4
'- v V 82.1 1
r , A $22 :-
3 ~‘. , E] 32.3 1.
:- t's) v v x 83-1 ;
L- 91 ~ 1
I— . “‘ "_ \ v —I
e . ‘ 4
\
b J‘ \‘ ‘v’v '1
p .I: ‘ “ 4
" n i‘ .e‘ p "‘\'\v d
b s "o‘ ‘ ‘v ': ‘ ‘ 4
5-. = s . vv ,. .,
‘\ ‘ \‘ V VVV
'A- :A 5 3..., X
‘ h e . ;““l ‘
‘K' A - =\:=‘ . V
d
: A?‘ ‘ A ‘ ';:JL‘ -: = s: a d
__ 5‘ A ".“.\ _‘
'a- A

_ A p _
_ /A A ‘ A _

D A; ‘ ‘
L' .1 "1-- a
y— -1
._ 1
t 1 L 1 1 1 L 1 1 1 '

00 300 400 500 600 700 800 900 1000
Temperature (K)

Figure 4.18 The temperature sensing characterization of all tremor/heater samples.
The solid lines represent ﬁtting curves by Steinhart-Hart equation.

TABLE 4.7 The results of heating/sensing study.

 

 

sag? (x 1' 03) (x 11,03) (x1c05) (x 10’) M (K) Km“ (7°)

Sl-Rl 4.84816 2.75084 94.0121 15.047 5.4007 1.0530

$2-1 3.41489 1.35979 4.27009 4.37662 4.7714 1.1136

82-2 37.2139 36.2285 1253.31 149.24 5.4537 1.4044

823 16.9735 17.2692 703.979 106.368 17.5895 3.5756

s3-1 8.37893 10.2236 522.01 95.2057 19.0476 2.7982
Note: 1. [a, b, c, d] are the coefﬁcients of Steinhart-Hart ﬁtting equation.

2. AT denotes the average difference between the heating temperature determined by self-
sensing and that by thermocouple.

141

 

 

 

 

 

 

 

 

"I'I'l"'l"'l"'l ' r r r ' r * 1 ' T ' r ' 1 ' I‘I'T‘VT'I'IW'I‘VI ‘
1000 ~ '1
. $ #
900 " v
7 + Sl-Rt 1
8m 1- ' A $2" .-
g I- “ \ D 82.2
V ' .\ 0 82-3
8 700 '— H‘s-n .. \ v 83" d
3 1- ’0, .\ .
‘ \
e 600 "' a A \\ "
8. L- t: 0) v
g 500 " n 3‘ "
I- - a e.‘
m I. I v A d
' 1| ‘ '4. ‘
300 - ‘11 o, - -_ ‘
ninlarnlalrlrrrlrrrl 1 l 1 l 1 a l a e l irlrrnlrnlnralrlrl
2005 5 7 5 910-1 2 s 4 5 5 7 5 e100
Resistivity (0cm) ~

Figure 4.19 Results of self-aching diamond heaters.

in the present study. The temperature sensing properties of diamond are studied over a
variety of sample substrates and conﬁgurations in the temperature range of 300 - 1000 K.
An excellent sensitivity is observed under 673 K. Due to its resistivity-temperature
characteristics, diamond thermistors are less sensitive at high temperature range. The
average error of sensing temperature as compared to that measured by thermocouple is

found in the range of 0.3289 - 4.8759%.

Due to its high thermal conductivity, diamond thin ﬁlm heater promises a uniformly

142

 

 

 

 

 

 

 

 

 

 

 

 

 

(a)

+100 1 I l 1 I l l l r I l l l l I
0 W .
. .1
-100 - -
+100 - I S1-R1 -
" A 82-1 '
0 ———111|-a——H A A ‘5 i— I 322 -
' 0 82-3 '
A -100 - .
¥3 +100 - V 83.1 '
3 0 W .
E . .
'1- 1s .
8+ ' e ‘1
|-— - q
0 w—H-O : ‘+— ‘
-100 ~ -
+100 - -
- -1
0‘ ‘ h V V y A
t ' ‘ v 77 -

_100' 1 l 1 l 1 l 1 l 1 J 1 I 1 l 1

300 500 700 900

Measured temperature (K)

Figure 4.20 The difference of heating temperature determined by self-sensing (Tu...)
and that by thermocouple (Tma).

143

 

 

 

 

 

 

 

 

 

 

 

 

(1))
+20 __ ' l ' l ' I I T ' l ' I l I .,
_ -1
0 W '1
-20 I e St-Rt j
_ A $1 .
+20 _ I W _
Q 0 ——11t-e—‘—t—1 s is a 4 323 t
2., - v 331 7
. -20 t -
3 +20 - ‘
E _ -
I; 0 W -
he? -20 ~ -
.7. +20 : ‘ . :
E, 0 T—‘Q—V—A“ ‘+— j
b -20 - ~
+20 ' 1
0 _———*-" y" ' ' ' TW 3
-20 '- 1 l 1 l 1 l 1 l 1 l L l l l .-

300 500 700 900

Measured temperature (K)

Figure 4.20 (Cont'd)

144

distributed temperature proﬁle under ideal boundary conditions. The power density and

heating temperature are in the ranges of 0 - 93 watt/cm2 and 300 - 950 K, respectively.
Because of the negative temperature coefﬁcient of resistance, diamond heater possesses a
high heating efﬁciency in the high temperature range. The average error of self-
temperature-sensing during heating application is in the range of 1.0530 - 3.5756%. By
varying the doping level, these parameters can be optimized for various applications,
particularly in harsh environments. In addition, the diamond resistors could be directly
deposited on the object to be heated and/or sensed with an exact and accurate custom

designed pattern.

CHAPTER 5

CONCLUSIONS AND FUTURE
RESEARCH

5.1 Introduction

The primary objective of this research is to realize the usage of CVD diamond thin
ﬁlms for temperature sensing and heating applications. In order to achieve this goal, ﬁrst,
the development of diamond device infrastructure technologies were addressed to
successfully fabricate the diamond sensor and/or heater test structures. The temperature
sensing and heating properties of diamond were then individually characterized on a
variety of test structures. Finally, the performance of a single element diamond
temperature sensor and heater structure was evaluated. The current research is excepted to
extend the applications of CVD diamond on a variety of sensing areas, including gas
sensor, liquid level sensors, mass flow meters, and vacuum and pressure gauges, and to
help commercialize the diamond temperature sensors and thin ﬁlm heaters in the near

term.

5.2 Summary of Contributions

1' Ultrahigh nucleation density

145

146

An improved nucleation technique for diamond thin ﬁlm deposition was

demonstrated. An ultrahigh nucleation density on the order of 1011 cm'z, achieved using
diamond powders with an average particle size of 0.038 pm, is the highest reported value.
Using this technique, a mean surface roughness of 30 nm, as measured by AFM, was
achieved for 1 pm thick ﬁlms. The ultrahigh nucleation density technique can provide

submicron continuous ﬁlms in a short deposition period.

0 Temperature Sensing

The annealing effect was studied using single crystal diamonds and polycrystalline
ﬁlms with different grain sizes. The increase in resistivity after annealing was found more
prominent for single crystal diamonds. The third order Steinhart-Hart equation was used
to calculate the temperature from measured resistivity with an average difference in the
range of 0.3289% - 4.8759% (1.4270 - 25.8909 K) as compared to reference temperature
measured by thermocouples. The sensitivity measurement of B factor in the range of

349.334 - 2208.452 K was found to increase with the room temperature resistivity.

- Heating

The heating properties of CVD diamond thin ﬁlm heaters were studied for the ﬁrst

time. The applied power density and maximum heating temperature were in the range of 0

- 93 watt/cm2 and 950 K, respectively. Due to the high thermal conductivity of diamond,

the heater substrates were found relatively isothermal.

0 Self-sensing diamond heater

A single element diamond resistor was used as heater as well as a temperature sensor

simultaneously. The heating temperature determined from the R-T characterization of the

147

diamond resistor was found to possess an average difference in the range of 1.0530% -
3.5756% (4.7714 - 17.5895 K) as compared to that directly measured by thermocouple.
By varying the doping level, the performance of such a sensor/heater element can be

optimized for particular application.

5.3 Future Research

Although in the present research, a diamond sensor/heater structure is characterized
for the ﬁrst time, there is a need in the following areas for realization of the diamond

sensor/heater applications.

(1) The understanding of electrical transport mechanisms of semiconducting CVD
diamonds is important to optimize the performance of diamond sensor/heater

Stl'llCIUI'CS.

(ii) Different ﬁlament conﬁguration should be tested to improve the consistency of

ﬁlm quality over a large deposition area.

(iii) The uniformity of doping need to be enhanced by gaseous boron source in
HFCVD reactor.

(iv) A passivation material should be carefully chosen to overcome the oxidation of

diamond for high temperature operation.

(v) The potential applications of the diamond sensor/heat structure, include gas
sensors, liquid level sensors, mass flow meters, and vacuum and pressure

gauges, need to be developed.

BIBLIOGRAPHY

10.

ll.

12.

BIBLIOGRAPHY

Ed. N. Eror et al., Expanding the Vision of Sensor Materials, National Academy
Press, Washington, D. C., 1995.

G. Harsanyi, Polymer Films in Sensors Applications, Technomic Publishing
Company, Lancaster, PA, 1995.

H. B. Sachse, Semiconducting Temperature Sensors and their Applications, John
Wiley & Sons, New York, 1975.

J. A. Hall, The Measurement of Temperature, Chapman and Hall, London, 1966.

T. D. McGee, Principles and Methods of Temperature Measurements, John Wiley &
Sons, New York, 1988.

R. P. Benedict, Fundamentals of Temperature, Pressure, and Flow Measurements,
John Wiley & Sons, New York, 1984.

R. B. Tait, R. Humphries, and J. Lorenz, "Thick Film Heater Elements and
Temperature Sensors in Modern Domestic Appliances," IEEE Transactions on

Industry Applications, vol. 30, no. 3, pp. 573-577, 1994.

A. Sale, R. Tait, J. Whitrnarsh, and M. A. Stein, "Thick Film Heater Elements and
Heat Sensors," Hybrid Circuit Technology, pp. 37-43, 1991.

Flexible Heating Elements, a product catalogue of The Ohmwave Company, Inc.,
Niantic, CT.

Product Bulletin of The Electra-Flex Heat, Inc., Bloomﬁeld, CT, 1995.
The Electric Heaters Handbook, vol. 28, Omega Engineering, Inc., Stamford, CT.

I. M. Buckley-Golder, P. R. Chalker, C. Johnston, S. Romani, and M. Werner, "High
Temperature Electronic Applications of Diamond," Advances in New Diamond

148

13.

14.

15.

l6.

17.

18.

19.

20.

21.

22.

23.

24.

25.

149

Science and Technology, Ed. S. Saito, N. Fujimori, O. Fukunaga, M. Kamo, K.
Kobashi and M. Yoshikawa, MYU, Tokyo, pp. 669-678, 1994.

D. G. Bhat, "Diamond-coated Cutting Tools: a Status Report," Mat. Tech., vol. 9, no.
11/12. PP. 253-261, 1994.

R. Hay, "A Hard Tool is Good to Find," Cutting Tool Engineering, vol. 45, no. 1,
1993.

T. J. Moravec and A. Partha, "Diamond Takes the Heat," Advanced Packaging,
October, pp. 8-11, 1993.

M. W. Geis, N. N. Efremow, J. D. Woodhouse, M. D. Aleese, M. Marchywka, D. G.
Socket, and J. F. Hochedez, "Diamond Cold Cathode," IEEE Elec. Dev. Lett., vol.
12. 139.456-459, 1991.

D. Hong, and M. Aslam, "Field Emission from P-type Polycrystalline Diamond
Films," J. Vac. Sci. Technol. B, vol. 13, no. 2, pp. 427-430, 1995.

M. Aslam, G. S. Yang, and A. Masood, "Boron-doped Vapor-deposited Diamond
Temperature Microsensors," Sensors and Actuators: A., vol. 45, no. 2, pp. 131-137,
1994.

A. Masood, "Technology and Electronic Properties of Diamond Film Microsensors
for Thermal Signals," Ph. D. dissertation, Michigan State University, 1991.

R. Pallas-Areny and J. G. Webster, Sensors and Signal Conditioning, John Wiley &
Sons, Inc., New York, 1991.

D. I. Finch, General Principles of Thermoelectric Thermometry, Reinhold
Publishing Corporation, Philadelphia, PA, 1962.

A. Boyer, E. Cisse, Y. Azzouz, and J. P. Ceron, "Narrow-bandgap Semiconductor
Based Thermal Sensors," Sensors and Actuators: A., vol. 25-27, pp. 637-640, 1991.

The Temperature Handbook, vol. 28, Omega Engineering, Inc., Stamford, CT.
S. Wessel, M. Pararneswaran, S. R. Morrison, and R. F. Frindt, "CMOS Thermally-
isolated Heater Structure as a Substrate for Semiconductor Gas Sensors,"

Microelectronics Journal, vol. 23, no. 6, pp. 451-456, 1992.

M. Gottschalk, "Tungsten Heater Boosts Rocket Fuel Efﬁciency," Design News, vol.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

150

49, no. 23, pp. 95-97, 1993.

M. N. Yoder, "The Vision of Diamond as an Engineered Material," Synthetic
Diamond: Emerging C VD Science and Technology, Ed. K. E. Spear and J. P.
Dismukes, pp. 3-17, John Wiley & Sons, New York, 1994.

E. M. Wilks and J. Wilks, "The Hardness and Wear of Diamond During Grinding
and Polishing," Physical Properties of Diamond, Ed. R. Berman, pp. 221-250,
Clarendon Press, Oxford, 1965.

F. P. Bowden and D. Tabor, "Deformation, Friction, and Wear of Diamond,"
Physical Properties of Diamond, Ed. R. Berman, pp. 184-220, Clarendon Press,
Oxford, 1965.

T. H. Huang, C. T. Kuo, T. S. Lin, and C. S. Chang, "Wear Behaviour of Various
Diamond-coated Cutting Tools under Different Deposition Conditions," Diamond
and Related Materials, vol. 2, pp. 928-932, 1993.

S. Tolansky, "Optical Studies on Diamond," Physical Properties of Diamond, Ed. R.
Berman, pp. 135-173, Clarendon Press, Oxford, 1965.

H. W. Ko, S. E. Hsu, S. J. Yang, M. S. Tsai, and Y. H. Lee, "Characterization of
Diamond Films for Optical Coatings," Diamond and Related Materials, vol. 2, pp.
694-698, 1993.

J. E. Field, The Properties of Diamonds, Academic Press, London, 1979.

Ed. W. J. Choyke et al., Materials for High-Temperature Semiconductor Devices,
National Academy Press, Washington, D. C., 1995.

W. Zhu, X. H. Wang, A. R. Badzian, and R. Messier, "An Investigation of the
Oxidation Behavior of CVD Diamond Films," New Diamond Sci. and Tech., MRS
Int. Conf Proc., pp. 821-826, 1991.

J. Wei and Y. Tzeng, "Reaction of CVD Diamond with Molten Iron Film," 2nd Int.
Conﬁ on Appl. of Diamond Films and Related Materials, Ed. Yoshikawa et al., vol.
2, pp. 609-612, Tokyo, Japan, 1993.

M. H. L. Pryce, "Electronic Structure of Diamond," Physical Properties of
Diamond, Ed. R. Berman, pp. 251-273, Clarendon Press, Oxford, 1965.

F. C. Champion, Electronic Properties of Diamonds, Butterworths & Co. Ltd.,

38.

39.

41.

42.

43.

45.

151

London, 1963.

P. J. Kemmey and P. T. Wedepohl, "Semiconducting Diamond," Physical
Properties of Diamond, Ed. R. Berman, pp. 325-355, Clarendon Press, Oxford,
1965.

R. Berrnan and M. Martinez, "Diamond Research," Industrial Diamond Review,
1976.

D. M. Malta, L. S. Plano, G. J. Tessmer, C. T. Kao, and B. A. Fox, "Improvements in
the Electronic Properties of CVD Diamond Films for Device Applications,"
Proceeding of 2nd International High Temperature Electronics Conference, vol. 2,
pp. P75-80, Charlotte, NC, 1994.

B. R. Stoner, C. T. Kao, D. M. Malta, and R. C. Glass, "Hall Effect Measurements
on Boron-doped, Highly Oriented Diamond Films Grown on Silicon Via Microwave
Plasma Chemical Vapor Deposition," Appl. Phys. Lett., vol. 62, pp. 2347-2349,
1993.

J. A. von Windheirn, V. Venkatesan, D. M. Malta, and K. Das, "Comparison of the
Electric Properties of Single-crystal and Polycrystalline Diamond by Hall Effect and
Capacitance-voltage Measurements," Diamond and Related Materials, vol. 2, pp.
841-846, 1993.

A. Masood, M. Aslam, M. A. Tamor, and T. J. Potter, "Synthesis and Electrical
Characterization of Boron-doped Thin Diamond Films," Appl. Phys. Lett., vol. 61,
pp. 1832-1834, 1992.

J. L. Davidson, "Diamond Electrical Properties and Electronic Device Behavior,"
Synthetic Diamond: Emerging C VD Science and Technology, Ed. K. E. Spear and J.
P. Dismukes, pp. 355-399, John Wiley & Sons, New York, 1994.

T. Sugino, K. Karasutani, H. Kataoka, K. Kobashi, and J. Shirafuji, "Electrical
Characterization of Boron-doped Polycrystalline Diamond Films Synthesized by
Hot Filament CVD," Advances in New Diamond Science and Technology, Ed. S.
Saito, N. Fujimori, O. Fukunaga, M. Kamo, K. Kobashi and M. Yoshikawa, MYU,
Tokyo, pp. 705-708, 1994.

L. S. Pan, D. R. Kania, S. Han, J. W. Agerii, M. Landstrass, L. O. Landen, and P.
Pianetta, "Electrical Transport Properties of Undoped CVD Diamond Films,"
Science, vol. 255, pp. 830-833, 1992.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

152

M. A. Plano, M. I. Landstrass, L. S. Pan, S. Han, D. R. Kania, S. McWilliam, and J.
W. Ager, "Polycrystalline CVD Diamond Films with High Electrical Mobility,"
Science, vol. 260, pp. 1310-1312, 1993.

L. S. Pan, S. Han, D. R. Kania, M. A. Plano, and M. I. Landstrass, "Electrial
Properties of High Quality Diamond Films," Diamond and Related Materials, vol. 2,
pp. 820-824, 1993.

N. Fujimori, T. Imai, and A. D0i, "Characterization of Conducting Diamond Films,"
Vacuum, vol. 36, pp. 99-102, 1986.

R. Erz, W. Dotter, K. Jung and H. Ehrhardt, "Preparation of Smooth and
Nanocrystalline Diamond Films," Diamond and Related Materials, v01. 2, pp. 449-
453, 1993.

Y. Hayashi, W. Drawl, R. W. Collins, and R. Messier, "In-process Ellipsometric
Monitoring of Diamond Film Growth by Microwave Plasma Enhanced Chemical
Vapor Deposition," Appl. Phys. Lett., vol. 60, no. 23, pp. 2868-2870, 1992.

E. J. Bienk and S. S. Eskildsen, "The Effect of Surface Preparation on the Nucleation
of Diamond on Silicon," Diamond and Related Materials, vol. 2, pp. 432-437, 1993.

J. S. Ma, H. Kawarada, T. Yonehara, J. I. Suzuki, J. Wei, Y. Yokota and A. Hiraki,
"Selective Nucleation and Growth of Diamond Particles by Plasma-assisted
Chemical Vapor Deposition," Appl. Phys. Lett., vol. 55, no. 11, pp. 1071-1073,
1989.

M. P. Everson and M. A. Tamor, "Studies of Nucleation and Growth Morphology of
Boron-doped Diamond Microcrystals by Scanning Tunnelling Microscopy," J. Vac.
Sci. Technol. B, vol. 9, no. 3, pp. 1570-1575, 1991.

S. Iijima, Y. Aikawa, and K. Baba, "Early Formation of Chemical Vapor Deposition
Diamond Films," Appl. Phys. Lett., vol. 57, no. 25, pp. 2646-2648, 1990.

K. Hirabayashi, Y. Taniguchi, O. Takamatsu, T. Ikeda, K. Ikoma, and N. Iwasaki-
Kurihara, "Selective Deposition of Diamond Crystals by Chemical Vapor
Deposition Using a Tungsten-ﬁlament Method," Appl. Phys. Lett., vol. 53, no. 19,
pp. 1815-1817, 1988.

B. W. Sheldon, R. Csencsits, J. Rankin, and R. E. Boekenhauer, "Bias-enhanced
Nucleation of Diamond During Microwave-assisted Chemical Vapor Deposition," J.
Appl. Phys, vol. 75, no. 10, pp. 5001-5008, 1994.

58.

59.

61.

62.

63.

65.

66.

67.

68.

69.

153

BR. Stoner, G. H. Ma, S. D. Wolter, and J. T. Glass, "Characterization of Bias-
enhanced Nucleation of Diamond on Silicon by in vacuo Surface Analysis and
Transmission Electron Microscopy," Phys. Rev. B, vol. 45, pp. 11067-11084, 1992.

X. Jiang, C. P. Klages, R. Zachai, M. Hartweg, and H. J. Fusser, "Epitaial Diamond
Thin Films on (001) Silicon Substrates," Appl. Phys. Lett., vol. 62, pp. 3438-3440,
1993.

M. P. Everson and M. A. Tamor, "Investigation of Growth Rates and Morphology
for Diamond Growth by Chemical Vapor Deposition," J. Mater. Res., vol. 7, no. 6,
pp. 1438-1443, 1992.

A. Masood, M. Aslam, M. A. Tamor and T. J. Potter, "Techniques for Patterning
CVD Diamond Films on Non-diamond Substrates," J. Electrochem Soc., vol. 138,
no. 11, pp. L67-L68, 1991.

J. C. Angus and CC. Hayman, "Low-temperature, Metastable Growth of Diamond
and Diamond-like Phases," Science, vol. 241, pp. 913-922, 1988.

F. P. Bundy, H. T. Hall, H. M. Stomg and R. H. Wentroff, "Man-made Diamond,"
Nature, vol. 176, pp. 51-55, 1955.

S. Matsumoto, Y. Sato, M. Kamo and N. Setaka, "Vapor Deposition of Diamond
Particles from Methane," Jpn. J. Appl. Phys., vol. 21, pp. L183-L185, 1982.

S. Matsumoto, Y. Sato, M. Tsutsumi and N. Setaka, "Growth of Diamond Particles
from Methane-hydrogen Gas," J. Materials Science, vol. 17, pp. 3106-3122, 1982.

J. C. Angus, Y. Wang, and M. Sunkara, "Metastable Growth of Diamond and
Diamond-like Phases," Annu. Rev. Mater. Sci., vol. 21, pp. 221-248, 1991.

T. D. Moustakas, "Growth of Diamond by CVD Methods and Effects of Process
Parameters," Synthetic Diamond: Emerging C VD Science and Technology, Ed. K. E.
Spear and J. P. Dismukes, pp. 145-192, John Wiley & Sons, New York, 1994.

K. Suzuki, A. Sawabe, H. Yasuda, and T. Inuzuka, "Growth of Diamond Thin Films
by DC Plasma Chemical Vapor Deposition," Appl. Phys. Lett., vol. 50, no. 12, pp.
728-729, 1987.

S. Matsumoto, "Chemical Vapour Deposition of Diamond in RF Glow Discharge,"
J. Mater. Sci. Lett., vol. 4, pp. 600-602, 1985.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

154

K. A. Snail, R. G. Vardiman, J. P. Estrera, J. W. Glesener, C. Merzbacher, C. J.
Craigie, C. M. Marks, R. Glosser, and J. A. Freitas, Jr., "Diamond Growth in
Turbulent Oxygen-acetylene Flames," J. Appl. Phys., vol. 74, no. 12, pp. 7561-7571,
1993.

G. M. Ma, B. E. Williams, J. T. Glass, and J. T. Prater, "Analysis Via Transmission
Electron Microscopy of the Quality of Diamond Films Deposited from the Vapor
Phase," Diamond and Related Materials, vol. 1, pp. 25-32, 1991.

J. J. Dubray, C. G. Pantano, M. Meloncelli, and E. Bertran, "Nucleation of Diamond
on Silicon, SiAlON, and Graphite Substrates Coated with an a-C:H Layer," J. Vac.
Sci. Technol. A, vol. 9, no. 6, pp. 3012-3018, 1991.

G. S. Yang and M. Aslam, "Ultrahigh Nucleation Density for Growth of
Smooth Diamond Films," Appl. Phys. Lett., vol. 66, no. 3, pp. 311-313, 1995.

Y. Namba, E. Heidrarpour, and M. Nakayarna, "Size Effects Appearing in the
Raman Spectra of Polycrystalline Diamond," J. Appl. Phys., vol. 72, no. 5, pp. 1748-
1751, 1992.

S. L. Flegler, J. W. Heckman, and K. L. Klomparens, Scanning and Transmission
Electron Microscopy, W. H. Freeman and Company, New York, 1993.

J. B. Bindell, "Scanning Electron Microscopy," Encyclopedia of Materials
Characterization, Ed. C. R. Brundle, C. A. Evans, Jr., and S. Wilson, pp. 70-84,
Butterworth-Heinemann, Boston, 1992.

R. S. Howland and M. D. Kirk, "Scanning Tunnelling Microscopy and Scanning
Force Microscopy," Encyclopedia of Materials Characterization, Ed. C. R. Brundle,
C. A. Evans, Jr., and S. Wilson, pp. 85-98, Butterworth-Heinemann, Boston, 1992.

J. D. Hunn and C. P. Christensen, "Ion Beam and Laser-assisted Micromachining of
Single-crystal Diamond," Solid State Technology, pp. 57-60, 1994.

M. W. Leksono and H. R. Shanks, "Patterning of Microwave Plasma Deposited
Diamond Films," Mat. Res. Soc. Symp. Proc., vol. 189, pp. 87-90, 1991.

T. Roppel, R. Ramesham, C. Ellis, and S. Y. Lee, "Thin Film Diamond
Microstructures," Thin Solid Films, vol. 212, pp. 56-62, 1992.

P. R. de la Houssaye, C. M. Penchina, C. A. Hewett, J. R. Zeidler, and R. G. Wilson,
"Hall Mobility and Carrier Concentration versus Temperature for Type IIa Natural

82.

83.

84.

85.

86.

87.

88.

89.

91.

155

Insulating Diamond Doped with Boron by Ion Implantation," J. Appl. Phys., vol. 71,
no.7, PP. 3220-3224, 1992.

J. Bemholc, S. A. Kajihara, and A. Antonelli, "Theory of Doping of Diamond," Mat.
Res. Soc. Symp. Proc., vol. 242, pp. 323-333, 1992.

G. Z. Cao, F. A. J. M. Driessen, G. J. Bauhuis, L. J. Giling, and P. F. A. Alkemade,
"Homoepitaxial Diamond Films Codoped with Phosphorus and Nitrogen by
Chemical Vapor Deposition," J. Appl. Phys., vol. 78, no. 5, pp. 3125-3131, 1995.

J. G. Ran, C. Q. Zheng, J. Ren, and S. M. Hong, "Properties and Texture of B-doped
Diamond Films as Thermal Sensor," Diamond and Related Materials, vol. 2, pp.
793-796, 1993.

G. Kawaguchi, J. Nakanishi, A. Otsuki, T. Oku, and M. Murakarni, "Dependence of
Contact Resistance on Metal Electronegativity for B-doped Diamond Films," J.
Appl. Phys., vol. 75, no. 10, pp. 5165-5170, 1994.

C. F. Chen and S. H. Chen, "Effects of Annealing on the Electrical Properties of
Boron—doped Diamond Films," Advances in New Diamond Science and Technology,
Ed. S. Saito, N. Fujimori, O. Fukunaga, M. Kamo, K. Kobashi and M. Yoshikawa,
MYU, Tokyo, pp. 709-712, 1994.

A. Otsuki, J. Nakanishi, G. Kawaguchi, T. Oku, and M. Murakami, "Formation of
Carbide Ohmic Contacts for p-type Diamond," Advances in New Diamond Science
and Technology, Ed. S. Saito, N. Fujimori, O. Fukunaga, M. Kamo, K. Kobashi, and
M. Yoshikawa, MYU, Tokyo, pp. 713-716, 1994.

C. A. Hewett, M. J. Taylor, J. R. Zeidler, and M. W. Geis, "Spectﬁc Contact
Resistance Measurements of Ohmic Contacts to Semiconducting Diamond," J. Appl.
Phys., vol. 77, no. 2, pp. 755-760, 1995.

J. Nakanishi, A. Otsuki, T. Oku, O. Ishiwata, and M. Murakami, "Formation of
Ohmic Contacts to p-type Diamond Using Carbide Forming Metals," J. Appl. Phys.,
vol. 76, no. 4, pp. 2293-2298, 1994.

Y. Nishibayashi, N. Toda, H. Shiomi, and S. Shikata, "Thermally Stable Ohmic
Contact to Boron Doped Diamond Films," Advances in New Diamond Science and
Technology, Ed. 8. Saito, N. Fujimori, O. Fukunaga, M. Kamo, K. Kobashi, and M.
Yoshikawa, MYU, Tokyo, pp. 717-720, 1994.

B. A. Fox, D. M. Malta, H. A. Wynands, S. R. Sahaida and J. A. von Windheim,

92.

93.

94.

95.

96.

97.

98.

100.

156

"High Temperature Electrical Measurements and Properties of Diamond,"

Proceeding of 2nd International High Temperature Electronics Conference, vol. 2,
pp. VI3-8, Charlotte, NC, 1994.

P. R. Chalker, C. Johnson, J. A. A. Crossley, J. Ambrose, C. F. Ayres, R. E. Harper,
I. M. Buckley-Golder, and K. Kobashi, "Degradation Mechanisms of Passivated and

Unpassivated Diamond Therrnistors," Diamond and related Materials, vol. 2, pp.
1100-1106, 1993.

A. D. Khazan, Transducers and Their Elements, PTR Prentice Hall, Englewood
Cliffs, NJ, 1994.

L. J. Van der Pauw, "A Method of Measuring Speciﬁc Resistivity and Hall Effect of
Discs of Arbitrary Shapes," Phillips Res. Repts, vol. 13, pp. 1-9, 1958.

D. M. Malta, J. A. von Windheim, H. A. Wynands, and B. A. Fox, "Comparison of
the Electrical Properties of Simultaneously Deposited Homoepitaxial and
Polycrystalline Diamond Films," J. Appl. Phys., vol. 77, no. 4, pp.1536-1545, 1995.

L. S. Pan, D. R. Kania, P. Pianetta, J. W. Ager, M. I. Landstrass, and S. Han,
"Temperature Dependent Mobility in Single-crystal and Chemical Vapor-deposited
Diamond," J. Appl. Phys., vol. 73, no. 6, pp. 2888-2894, 1993.

J. P. Bade, S. R. Sahaida, B. R. Stoner, J. A. von Windheim, J. T. Glass, K. Miyata,
K. Nishimura, and K. Kobashi, "Fabrication of Diamond Thin-ﬁlm Thermistors for
High-temperature Applications," Diamond and Related Materials, vol. 2, pp. 816-
819, 1993.

G. S. Yang and M. Aslam, "CVD Diamond Resistor as Heater and Temperature
Sensor," Applications of Diamond Films and Related Materials; Third International
Conference, Ed. A. Feldman, Y. Tzeng, W. A. Yarbrough, M. Yoshikawa, and M.
Murakawa, N IST Special Publication, Washington D. C., pp. 125-128, 1995.

D. R. Kania, M. I. Landstrass, M. A. Plano, L. S. Pan, and S. Han, "Diamond
radiation detectors," Diamond and related Materials, vol. 2, pp. 1012-1019, 1993.

V. V. Tokiy, V. I. Tirnchenko, V. A. Soroka, N. Tokiy, B. V. Spitsyn, and L. L.
Bouilov, "Diamond Film Metal-Semiconductor-Metal Photodetector," Applications
of Diamond Films and Related Materials; Third International Conference, Ed. A.
Feldman, Y. Tzeng, W. A. Yarbrough, M. Yoshikawa, and M. Murakawa, NIST
Special Publication, Washington D. C., pp. 145-148, 1995.

101.

102.

103.

104.

105.

106.

107.

108.

109.

110.

157

S. Han, R. S. Wagner, J. Joseph, M. A. Plano, and M. D. Moyer, "Chemical Vapor
Deposited Diamond Radiation Detectors for Ultrahigh Radiation Dose-rate
Measurements: Response to Subnanosecond, l6-MeV Electron Pulses," Rev. Sci.
Instrum, vol. 66, no. 12, pp. 5516-5521, 1995.

S. Han, S. G. Prussin, L. 8. Pan, D. R. Kania, S. M. Lane, G. Dennis, R. S. Wagner,
and K. Harris, "Temporally Resolved Response of a Natural Type IIA Diamond
Detector to Single-Particle Excitation," Diamond and Related Materials, vol. 2, pp.
835-840, 1993.

V. F. Dvoryankin, A. A. Kudryashov, Yu. Sh. Temirov, L. L. Bouilov, G. A.
Sokolina, and A. E. Alexenko, "CVD Diamond Films for X-ray Detectors,"
Applications of Diamond Films and Related Materials; Third International
Conference, Ed. A. Feldman, Y. Tzeng, W. A. Yarbrough, M. Yoshikawa, and M.
Murakawa, NIST Special Publication, Washington D. C., pp. 145-148, 1995.

S. H. Lin and L. H. Sverdrup, "Electron Beam Activated Diamond Devices,"
Applications of Diamond Films and Related Materials; Third International
Conference, Ed. A. Feldman, Y. Tzeng, W. A. Yarbrough, M. Yoshikawa, and M.
Murakawa, NIST Special Publication, Washington D. C., pp. 79-82, 1995.

M. Marchywka, J. Hochedez, M. W. Geis, D. G. Socker, D. Moses, and R. T.
Goldberg, "Ultraviolet Photoresponse Characteristics of Diamond Diodes," Appl.
Optics, vol. 30, no. 34, pp. 5011-5013, 1991.

M. Marchywka and D. Moses, "Optical Characterization of Diamond MIS
Capacitors," IEEE Trans. Elect. Devices, vol. 41, no. 7, pp. 1265 - 1272, 1994.

V. I. Latsa and Y. M. Rotner, "Effect of Axial Pressure on Some Properties of
Synthetic Semiconductor Diamonds (original in Russian)," Vys. Davleniya Svoita
Mater., pp. 55-59, 1980.

V. I. Latsa, I. N. Mikhailik, and G. V. Krishchuk, "Current-Voltage Characteristics
of Synthetic Diamond under Pressure (original in Russian)," Fiz Tekh. Vys.
Davlenii., vol. 9, pp. 62-65, 1982.

M. Aslam, I. Taher, A. Masood, M. A. Tamor, and T. J. Potter, "Piezoresistivity in
Vapor Deposited Diamond Films," Appl. Phys. Lett., vol. 60, pp. 2923-2925, 1992.

I. Taher, M. Aslam, M. A. Tamor, T. J. Potter, and R. C. Elder, "Piezoresistive
Microsensors Using P-type CVD Diamond Films," Sensors and Actuators: A, vol.
45. pp. 35-43, 1994.

158

111. D. R. Wur, J. L. Davidson, W. P. Kang, and D. L. Kinser, "Polycrystalline Diamond
Pressure Sensor," J. of Microelectromechanical System, vol. 4, no. 1, pp. 34-41,
1995.

112. S. Sahli, X. Hou, and M. Aslam, "Piezoresistive Gauge Factor of Polycrystalline
Diamond Measured at Different Fields and Temperatures," Applications of Diamond
Films and Related Materials; Third lntemational Conference, Ed. A. Feldman, Y.
Tzeng, W. A. Yarbrough, M. Yoshikawa, and M. Murakawa, NIST Special
Publication, Washington D. C., pp. 95-98, 1995.

113. Y. Gurbuz, W. P. Kang, J. L. Davidson, D. V. Kems, and B. Henderson, "High
Sensitivity and Wide Temperature Tolerance Hydrogen Gas Sensors Utilizing
PECVD Diamond Technology," Applications of Diamond Films and Related
Materials; Third International Conference, Ed. A. Feldman, Y. Tzeng, W. A.
Yarbrough, M. Yoshikawa, and M. Murakawa, NIST Special Publication,
Washington D. C., pp. 87-90, 1995.

114. J. Zhang, B. Huang, D. K. Reinhard, and J. Asmussen, "An Investigation of The
Effects of Electromagnetic Field Patterns on Microwave Plasma Diamond Thin Film
Deposition," J. Vac. Sci. and Technol. A, vol. 8, pp. 2124-2128, 1990.

115. J. Zhang, "Diamond Thin Film Deposition by Microwave Plasma Enhanced
Chemical Vapor Deposition," Ph. D. dissertation, Michigan State University, 1992.

116. M. J. Ulczynski, D. K. Reinhard, M. Prystajko, and J. Asmussen, "Low Temperature
Deposition of Thin Film Diamond," Advances in New Diamond Science and
Technology, Ed. S. Saito, N. Fujimori, O. Fukunaga, M. Kamo, K. Kobashi and M.
Yoshikawa, MYU, Tokyo, pp. 41-44, 1994.

117. G. S. Yang, M. Asalm, K. P. Kuo, D. K. Reinhard, and J. Asmussem, "Effect of
Ultrahigh Nucleation Density on Diamond Growth at Different Growth Rates and
Temperatures," J. Vac. Sci. and Technol. B, vol. 13, no. 3, pp. 1030-1036, 1995.

118. The atomic force microscope unit is a Nano Scope [II from Digital Instruments
Corporation.

119. M. A. Tamor and M. P. Everson, "On the Role of Penetration Twins in the
Morphological Development of Vapor-Grown Diamond Films," Proceeding of SPIE
Conference on Diamond-Film Semiconductors, pp. 14-25, Los Angeles, CA, 1994.

120. K. Bekku, Y. Mori, N. Eirnori, H. Makita, A. Hatta, T. Ito, and A. Hiraki, "Growth
and Characterization of CVD Diamond Films Doped with Phosphorous," Advances

121.

122.

123.

124.

125.

126.

127.

128.

129.

130.

159

in New Diamond Science and Technology, Ed. S. Saito, N. Fujimori, O. Fukunaga,
M. Kamo, K. Kobashi and M. Yoshikawa, MYU, Tokyo, pp. 701-704, 1994.

L. J. Giling, Y. S. Wang, G. Z. Cao, G. J. Bauhuis, and P. F. A. Alkemade,
"Diamond Films Co-doped with Phosphorus and Nitrogen," Advances in New
Diamond Science and Technology, Ed. S. Saito, N. Fujimori, O. Fukunaga, M.
Kamo, K. Kobashi and M. Yoshikawa, MYU, Tokyo, pp. 359-362, 1994.

M. A. Plano, S. Zhao, C. F. Gardinier, M. I. Landstrass, D. R. Kania, H. Kagan, K.
K. Gan, R. Kass, L. S. Pan, S. Han, S. Schnetzer, and R. Stone, "Thickness
Dependence of the Electrical Characteristics of Chemical Vapor Deposited Diamond
Films," Appl. Phys. Lett., vol. 64, no. 2, pp. 193-195, 1994.

The deﬁnition of zero-power resistance is reprinted from MIL-T-23648A, the US
Department of Defense general speciﬁcation for thermistors.

G. S. Yang and D. M. Aslam, "Single-structure Heater and Temperature Sensor

Using a p-type Polycrystalline Diamond Resistor," IEEE Electron Device Letters,
vol. 17, no. 5, pp. 250-252, 1996.

M. Werner, O. Dorsch, H. U. Baerwind, E. Obermeier, L. Haase, W. Seifert, A.
Ringhandt, C. Johnston, S. Romani, H. Bishop, and P. R. Chalker, "Charge Transport
in Heavily B-doped Polycrystalline Diamond Films," Appl. Phys. Lett., vol. 64, no.
5, pp. 595-597, 1994.

P. J. French and A. G. R. Evans, "Piezoresistance in Polysilicon and its Applications
to Strain Gauges," Solid State Electronics, vol. 32, no. 1, pp. 1-10, 1989.

M. I. Landstrass and K. V. Ravi, "Resistivity of Chemical Vapor Deposited
Diamond Films," Appl. Phys. Lett., vol. 55, no. 10, pp. 975-977, 1989.

S. Albin and L.Watkins, "Electrical Properties of Thin Film and Bulk Diamond
Treated in Hydrogen Plasma," Diamond, Silicon Carbide, and Related Wide
Bandgap Semiconductors, Ed. J. T. Glass, R. Messier, and N. Fujimori, Materials
Research Society Symposium Proceedings, vol. 162, pp. 303-308, 1990.

A. T. Collins, "The Optical and Electronic Properties of Semiconducting Diamond,"
Philos. Trans. R. Soc. London A, vol. 342, pp. 233-244, 1993.

S. Sahli and D. M. Aslam, "Effect of Post-deposition Anneal on the Resistivity of P-
type Polycrystalline Diamond Films," in review by Appl. Phys. Lett., 1996.

160

131. Y. Muto, T. 811ng, and J. Shirafuji, "Electrical Conduction in Undoped Diamond
Films Prepared by Chemical Vapor Deposition," Appl. Phys. Lett., vol. 59, no. 7, pp.
843-845, 1991.

132. H. Shiomi, K. Tanabe, Y. Nishibayashi, and N. Fujimori, "Epitaxial Growth of High
Quality Diamond Film by the Microwave Plasma-Assisted Chemical-Vapor-
Deposition Method," Jap. J. Appl. Phys., vol. 29, pp. 34-40, 1990.

133. A. J. Chapman, Heat Transfer, Macmillan Co., New York, 1960.

134. A. I. Brown and S. M. Marco, Introduction to Heat Transfer, McGraw Hill Co., New
York, 1958.

 

a '7‘“
l
l