r... g
57;:

.e .1; .2112...
2.35.3143...
52:].

x.
..5.x.r7r

$3....2iv3rfz
1.39%?! 5.1....
n. n.

s , a .a.
4. ”Katie“? a? .V

IL... 5!.
i... x

 

lES!S
2
am

This is to certify that the

dissertation entitled

INVESTIGATION OF THE FILM PROPERTIES AND DEPOSITION
PROCESS OF a-C:H FILMS DEPOSITED WITH A MICROWAVE
ECR PLASMA REACTOR

presented by

Bo Keun Kim

has been accepted towards fulfillment
of the requirements for

Doctor of Philosophy degree in Electrical Engineering

 

 

WW

Major professor

Date May 1L._2Q0_0__

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

LIBRARY
Michigan State
University

 

PLACE lN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

DATE DUE DATE DUE DATE DUE

4

 

11/00 c10|FiCJDaieDuap65~pJ4

 

INVESTIGATION OF THE FILM PROPERTIES
AND DEPOSITION PROCESS OF a-CzH FILMS
DEPOSITED WITH A MICROWAVE ECR
PLASMA REACTOR

By

Bo Keun Kim

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Department of Electrical and Computer Engineering

2000

 

 

ABSTRACT

INVESTIGATION OF THE FILM PROPERTIES AND DEPOSITION
PROCESS OF a—C:H FILMS DEPOSITED WITH A NHCROWAVE ECR
PLASMA REACTOR

By

Bo Keun Kim

Hydrogenated amorphous carbon (a-CzH) films are deposited from acetylene gas
at pressures in the submillitorr range (0.2-0.6 mTorr), and methane-argon and acetylene-
argon gas mixtures at pressures in the millitorr range (1-5 mTorr) in a microwave ECR
plasma reactor operated with If biased substrate holder. The films deposited at pressures
in the submillitorr range showed a strong influence of ion energy and ion flux to neutral
flux ratio on the deposition process and film properties. The films showed a peak value of
optical bandgap when deposited at -200 V of rf induced substrate bias revealing the ion
energy effect. The effect of ion flux to neutral flux ratio was seen in the depositions done
with varied substrate positions from the discharge region and the threshold ratio of ion
flux to neutral flux for deposition of films with the peak is found to be in the range of
006-01. Maintaining a low deposition pressure is found to be critical to obtain fihns of
high optical bandgaps. The deposition rate (~90 nm/min) at 7.0 sccm of acetylene flow
rate is much higher than the filtered ion beam and plasma beam deposition systems used
for tetrahedral (hydrogenated) amorphous carbon film depositions in the literature.

The films deposited at pressures in the millitorr range showed variation of the

film properties dependent on the deposition condition. The films deposited with the two

 

 

 

different gas mixtures including argon-methane and argon-acetylene under similar input
variable conditions have substantially different properties including deposition rate, mass
density, optical absorption coefficient, refractive index, optical bandgap and hydrogen
content. The deposition variables varied included rf induced dc substrate bias voltage (0
to -100 V), argon/hydrocarbon gas flow ratio (0-1.0) and pressure (1-5 mTorr).

The discharge properties including electron temperature, ion saturation current,
and residual gas composition of the exit gas flow for the various gas mixtures were
measured to help explain the different deposition results from the acetylene-based and
methane-based gas mixtures. From the discharge properties, the ion flux to neutral flux
ratio is estimated and the carbon flux in the input gas flow is shown to be the rate-
limiting process of deposition. The variation of film property is attributed to the hydrogen
content in the film composition and the hydrogen content is controlled by the ion
bombardment effect in the film deposition process. For the films deposited from
acetylene-argon discharges the use of lower pressures to obtain an increased ion flux to
neutral flux ratio to the substrate was found to be critical for obtaining dense, low
hydrogen content films. For the films deposited from methane-argon discharge the
addition of argon to the discharge increased the fihn's mass density and lowered the
hydrogen content. In both methane-based and acetylene-based deposition processes the rf
induced bias was also a critical determining factor of film properties.

The variation of film properties in the film deposition at millitorr pressures can be
mainly explained with the consideration of the hydrogen content of the films. In contrast,
the variation of film properties in the film deposition at submillitorr pressures is mainly

attributed to sp3 to sp2 carbon bonding ratio changes in the film composition.

'——_____4

 

ACKNOWLEDGEMENTS

The author is indebted to many individuals for the successful completion of this
dissertation and his degree program The author wishes to thank his advisor, Dr. Timothy
A. Grotjohn, for his thoughtful and inspiring guidance and support, and also for his
painstaking review of this manuscript. He also wishes to extend his thanks to Dr. Donnie
K. Reinhard, Dr. Jes Asmussen, Department of Electrical and Computer Engineering, and
Dr. Brage Golding, Department of Physics and Astronomy, for serving on author’s
guidance committee, and for their great lectures and academic advices. A special thanks
is given to the author’s wife, Kyungsim Yoon, and to his family for their patience,

understanding and sacrifice during the course of this program.

 

 

 

 

TABLE OF CONTENTS

List of Tables ..................................................................................... vii
List of Figures .................................................................................... viii
Chapter 1
1. Introduction
1.1 Motivation ...................................................................................... l
1.1 Research Objectives ........................................................................... 3
1.2 Research Methods ............................................................................. 4
1.3 Research Outline ............................................................................ 5
Chapter 2
2 Hydrogenated Amorphous Carbon F ilrns (a-CzH): A Review
2.1 Composition of a—C:H Films ................................................................ 7
2.2 Electronic Structure of a—C: H Films ...................................................... 13
2.3 Deposition Mechanism of Diamond-like a-C and a-C:H Films ....................... 15
2.3.1 Microscopic Process .................................................................... 15
2.3.2 Macroscopic Process ................................................................... 22
2.4 Deposition Systems ......................................................................... 24
Chapter 3
3 The Deposition System and Film Characterization
3.1 Introduction .................................................................................. 34
3.2 Electron Cyclotron Resonance ............................................................ 34
3.3 Deposition System and Conditions ....................................................... 37
3.3.1 Description of Deposition System .................................................... 37
3.3.2 Deposition Conditions ................................................................. 43
3.3.3 Sample Preparation ..................................................................... 44
3.4 Characterization of Discharge Properties ............................................... 44
3.4.1 Double Langmuir Probe Measurement .............................................. 45
3.4.2 Determination of Ion Energy and Ion Flux .......................................... 45
3.4.3 Partial Pressure Analysis of Exit Gas and Temperature Measurement. . . . . .....48
3.5 Characterization of a-C:H Films .......................................................... 49
V

 

 

Chapter 4

4 Fihns Deposited from Acetylene Discharge at Pressure in the Submillitorr Range

4.1 Introduction ..................................................................................... 64
4.2 Discharge Properties at Pressures in the Submillitorr Range ............................. 66
4.3 The Effect of Ion Energy (RF Induced Substrate Bias) on Film Properties ............ 72
4.4 The Effect of Ion Flux Ratio to Neutral Flux ............................................... 78
4.4.1 The Effect of Pressure ..................................................................... 82
4.4.2 The Effect of Microwave Power ......................................................... 87
4.4.3 The Effect of Substrate Position .......................................................... 91
4.5 The Effect of Deposition Temperature ...................................................... 97
4.6 The Effect of Addition of Helium Gas ....................................................... 98
4.7 The Deposition Rate as a Function of the Acetylene Flow Rate ......................... 99
4.8 Summary ....................................................................................... 103
Chapter 5

5 Fihns Deposited fiom Acetylene-Argon, and Methane-Argon Discharges
at Pressures in the Millitorr Range

5.1 Introduction .................................................................................... 106
5.2 Discharge Properties at Pressures in the Millitorr Range ................................. 108
5.3 Film Properties at Pressures in the Millitorr Range ...................................... 117
5.3.1 Absorption Coefficients .................................................................. 117
5.3.2 The Effect of rf Induced Substrate Bias ................................................ 121
5.3.3 The Effect of Pressure .................................................................... 135
5.3.4 The Effect of Argon Flow Rate ......................................................... 142
5.4 Summary ....................................................................................... 147
Chapter 6
6 Conclusions .................................................................................... 151

List of References

vi

 

 

 

LIST OF TABLES

Table 2 — 1: Energies of various processes for carbon .......................................... 16
Table 2 - 2: Comparison of deposition methods ................................................ 33
Table 4 - 1: The deposition variable space ....................................................... 66

Table 4 - 2: Double Langmuir probe measurements for electron temperature and
plasma density ......................................................................... 67

Table 4 - 3: The sheath thickness and acetylene ion energy at variations
of rf induced substrate bias .......................................................... 68

Table 4 - 4: The effect of temperature effect on optical bandgaps (E.auc and E04)
and index of refraction (n) ........................................................... 97

Table 5 - 1: The input variable space ............................................................ 107

Table 5 - 2: Langmuir probe measurement of argon, methane-argon, and
acetylene-argon discharges. The argon flow rate is constant
at 8 sccm .............................................................................. 109

Table 5 - 3: Mass of various species ............................................................ 110

Table 5 - 4: The dependence of momentum on mass. The momentum of several
ion types is normalized by that of an atomic hydrogen ion
under the condition of the same ion energy and it is designated
by Mx/MHW ......................................................................... 137

Table 5-5: Comparison of film properties fiom argon (50 %)-methane (50 %)
and argon (50 %)-acetylene (50 %) discharges, and from two different
rf induced substrate biases of 0 and ~60 V ........................................ 148

vii

 

 

 

 

LIST OF FIGURES

Fig. 2 - 1: Schematic representation of hydrogenated amorphous carbon (a-CzH)

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.
Fig.
Fig.
Fig.

Fig.

Fig

Fig

Fig

Fig

Fig

. _—#4

films. The network is comprised of hydrogen, sp2 carbon atoms and sp3
carbon atoms. The lines represent the bonds and the arrows

represent the dangling bonds ........................................................... 8
2 - 2: sp2 carbon atoms in a-C:H films in the form of a cluster

of six-membered aromatic rings ......................................................... 9
2 - 3: Ternary phase diagram of hydrocarbon films ....................................... 1 1

2 - 4: These plots show the trends of properties of ta-C:H films with
the variation of sp3 fraction and sp fraction ........................................ 12

2 - 5: A schematic diagram of density of state (DOS) of a-C:H films,
which shows 0' and it states, and defect states ....................................... 14

2 - 6: How subplanted ions increase local density. A fraction n penetrates
the surface of the fihn while the fraction (1 -n) fails to penetrate and
increases film thickness ................................................................ l7
2 - 7: An example of penetration probability of C+ ions into a-C ....................... 19
2 - 8: An example of calculated dependence of density on ion energy ................. 20
2 - 9: Experimental setup of the rf plasma deposition system ............................ 25
2 - 10: Schematic diagram of the plasma beam source ................................... 27
2 - ll: Schematic diagram of filtered carbon ion beam system ......................... 28
. 2 - 12: Schematic diagram of one type of ECR-CVD system ............................ 30
. 3 - 1: Principle of ECR heating. The electron gains microwave
energy continuously ...................................................................... 36
. 3 - 2: The microwave ECR plasma source with the rf biased substrate holder ........ 38
. 3 - 3: The side view of the microwave cavity, the baseplate, and
the deposition chamber of the system ................................................ 39
. 3 - 4: The cross section (top View) of the baseplate of the system ...................... 42

viii

 

Fig.

Fig.

Fig.

Fig.

Fig.
Fig.
Fig.
Fig.
Fig.
Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

Fig

 

 

3 - 5: (a) The spectrometer used to measure the transmittance
and reflectacnce of the films and (b) the tilted angle needed
to measure the reflected beam ........................................................ 50

3 - 6: The measurement of transmittance and reflectance of light
for an a-C:H film on glass substrate ................................................. 51

3 - 7: The transmittance and reflectance of an a-C:H film versus wavelength.
The modeled reflectance data is simulated for the determination of

thickness and index of refraction of the film ........................................ 54
3 - 8: An example of a SEM cross-section for determination of thickness

of the film ................................................................................. 56
3 - 9: Absorption coefficient of an a—C:H film versus photon energy ................... 57
3 - 10: Refractive index of an a-C:H film versus photon energy ........................ 58
3 - 11: Tauc plot ofan a-C:H film ............................................................ 61
3 — 12: An example IR absorption spectra .................................................. 62
4 - 1: Partial pressure analysis for acetylene gas with the system off ................... 70
4 - 2: Partial pressure analysis for acetylene gas with the system discharge on... . ....71

4 -3: Optical bandgap (Em, and E04) versus rf induced substrate bias
for films deposited from acetylene gas feed at 0.2 mTorr
discharge pressure ....................................................................... 73

4 - 4: Hydrogen content versus rf induced substrate bias for films from
acetylene gas feed at 0.2 mTorr discharge pressure ................................ 76

4 - 5: Index of refraction at 523 A° versus rf induced substrate bias for
films from acetylene gas feed at 0.2 mTorr discharge pressure ................... 77

4 - 6: Deposition rate versus rf induced substrate bias for the films
deposited from acetylene discharges ................................................. 79

4 - 7: Current density on the substrate versus dc bias on the substrate
holder for acetylene discharge at pressure of 0.2 mTorr and substrate
position of 3.5 cm ........................................................................ 81

4 - 8: Optical bandgap (E.mic and E04) versus pressure for films deposited
with -200 V of rf induced substrate bias from acetylene gas feed ................. 83

4 — 9: Index of refraction versus pressure for fihns deposited
with -200 V of rf induced substrate bias from acetylene gas feed ................. 84

ix

 

 

 

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

Fig.

4 - 10: Current density to the substrate versus pressure for the acetylene
discharge .................................................................................. 85

4 - 11: Optical bandgap (Etauc and E04) versus absorbed microwave power
for films deposited with 200 V of rf induced substrate bias fiom acetylene
gas feed at 0.2 mTorr discharge pressure ........................................... 88

4 - 12: Index of refraction versus absorbed microwave power for films
deposited with 200 V of rf induced substrate bias from acetylene
gas feed at 0.2 mTorr discharge pressure ............................................ 89

4 - 13: Current density to the substrate holder versus absorbed microwave
power for an acetylene discharge at 0.2 mTorr pressure ........................ 90

4 - 14: Optical bandgap (Etauc and E04) versus rf induced substrate bias
for films deposited from acetylene gas feed at 0.2 mTorr discharge
pressure with substrate positions (s.p.) of 3.5 cm and 6.0 cm ................... 92

4 - 15: Index of refiaction versus rf induced substrate bias for films deposited
from acetylene gas feed at 0.2 mTorr discharge pressure with
substrate positions (s.p) of 3.5 cm and 6.0 cm ...................................... 93

4 - 16: Current density on the substrate versus dc bias on the substrate

holder for acetylene discharge at 0.2 mTorr pressure at substrate

positions (s.p) of 3.5 cm and 6.0 cm ................................................ 96
4 - 17: Optical bandgap as a fiinction of ion flux to neutral flux ratio .................. 96

4 - 18: Optical bandgap (E,auc and E04) versus flow rate of helium for films
deposited from acetylene and helium gas feed with 200 V of rf
induced substrate bias ................................................................ 100

4 - 19: Index of refiaction versus flow rate of helium for films
deposited fiom acetylene and helium gas feed with 200 V of
rf induced substrate bias .............................................................. 101

4 — 20: The deposition rate of a-C:H films versus the flowrate of acetylene
gas into the discharge. The pressure of the discharges varied from
0.2 mTorr to 0.45 mTorr as the acetylene flow rates increased from
4 sccm to 35 sccm .................................................................... 102

5 - 1: Electron temperature for argon discharges versus pressure in the
ECR-CVD system ..................................................................... 111

5 - 2: Plasma density, np, for argon discharges versus pressure in the
ECR-CVD system ..................................................................... 112

 

 

 

Fig. 5 - 3: Partial pressure analysis for the methane-argon gas mixture with
the discharge off ........................................................................ 113

Fig. 5 - 4: Partial pressure analysis for the methane-argon gas mixture with the
discharge on ............................................................................. 114

Fig. 5 - 5: Partial pressure analysis for acetylene—argon gas mixture with
discharge off ............................................................................ 1 15

Fig. 5 - 6: Partial pressure analysis for acetylene-argon gas mixture with
discharge on ............................................................................ 116

Fig. 5 - 7: Optical absorption coefficients of films deposited in methane-argon
discharges. Data is plotted versus photon energy at various rf induced
substrate biases .......................................................................... 118

Fig. 5 - 8: Optical absorption coefficients of films deposited in methane-argon
discharges. Data is plotted versus photon energy at various argon
flow ratios ................................................................................ 1 19

Fig. 5 -9: Optical absorption coefficients of films deposited in
acetylene-argon discharges. Data is plotted versus photon energy

at various rf induced substrate biases ................................................ 120

Fig. 5 - 10: Deposition rate versus rf induced substrate bias for methane-based
and acetylene-based films ........................................................... 123

Fig. 5 - 11: Mass density versus rf induced substrate bias for methane-based
and acetylene-based films ........................................................... 124

Fig. 5 - 12: Hydrogen content (at. %) versus rf induced substrate-bias for
methane-based and acetylene-based films ......................................... 125

Fig. 5 - 13: Index of refraction versus rf induced substrate bias for methane-based
and acetylene-based films ........................................................... 126

Fig. 5 - 14: Optical bandgap (Euluc and E04) versus rf induced substrate bias
for methane-based and acetylene-based films ................................... 127

Fig. 5 - 15: Variation of optical bandgap versus hydrogen content for
acetylene-based films and methane-based films .................................. 133

Fig. 5 - 16: Variation of optical bandgap versus mass density for acetylene-based
films and methane-based films deposited in this study and in the study
of Ref. [3] ............................................................................. 134

Fig. 5 - l7: Deposition rate of methane and acetylene-based films versus

xi

 

 

deposition pressure .................................................................... 13 8

Fig. 5 - 18: Index of refraction of methane and acetylene deposited films
versus deposition pressure ........................................................... 139

Fig. 5 - 19: Optical bandgap (Emuc and E04) of methane and acetylene deposited
films versus deposition pressure .................................................... 140

Fig. 5 - 20: Deposition rate of methane and acetylene deposited films versus argon
flow ratio ............................................................................... 143

Fig. 5 - 21: Index of refraction of methane and acetylene deposited films
versus argon flow ratio ............................................................... 144

Fig. 5 - 22: Optical bandgap (E,auc and E04) of methane and acetylene deposited
films versus argon flow ratio ....................................................... 145

 

 

Chapter 1
1. Introduction

1.1 Motivation

Hydrogenated amorphous carbon (a-CzH) films are amorphous materials
containing a mixture of sp3 and sp2 hybridized carbon and hydrogen. The films do not
have long range order in their spatial structure unlike diamond which is an sp3 hybridized
carbon crystal and graphite which is a sp2 hybridized carbon crystal. They do have short
range order and possibly medium range order. a—C:H films contain lower levels of
hydrogen as compared to hydrocarbon polymers.

The properties of a—C:H films are mainly determined by the sp3/sp2 ratio and
hydrogen content. The content of sp3 hybridization sites determines the mechanical
properties of the films like density, hardness, stress, etc. and the content of sp2 sites
primarily determines optical and electrical properties like optical bandgap and
conductivity. Hydrogen in the films passivates the dangling bonds and influences the
mechanical, optical and electrical properties of the films. The films having high sp3 sites,
low sp2 sites, and low hydrogen content show extreme hardness and high density, and
they are called diamond-like carbon (DLC) films or tetrahedral hydrogenated amorphous
carbon films (ta-CzH). Graphite-er films have relatively high sp2 hybridization sites and
are soft, and polymer-like films contain high levels of hydrogen and are very soft. Thus,
a-C:H films can have a wide range of properties including those of diamond, graphite and

polymers depending on the deposition method and deposition conditions.

—_;4

 

 

 

The interesting properties of DLC films are characterized as extreme hardness,
extreme smoothness, low friction coefficient, high optical transparency over a wide
spectral range of photon energies, high electrical resistivity and high chemical inertness.
Some of film properties fiom the literatures [1-3] include sp3 fraction (0.2 - 0.8),
hydrogen-content (25 - 65 at.%), optical gap (0.8 - 3.0 eV), index of refraction (1.5 - 2.3),
density (1.3 -3.0 g/cm3), and resistivity (106- 1015 Gem). Thus, the films have
applications as protective coatings, and as optical coatings for anti-reflection and infrared
filters, etc. [4, 5]. Additionally, possible applications for electronic device materials are
under investigation [6-11].

Hydrogenated amorphous carbon (a-CzH) and amorphous carbon (a-CzH) films
have been deposited by a wide range of techniques[12] including dc plasma
deposition[13], rf plasma deposition [14], plasma beam source deposition [3, 15], filtered
ion beam deposition [16—19], and microwave electron cyclotron resonance (ECR) plasma
deposition [20-30]. The feed gases for the chemical vapor deposition (CVD) method are
usually methane or acetylene as the hydrocarbon gas with or without argon gas or
hydrogen gas. The properties of the films change with deposition method, type of feed
gases, and deposition conditions. The fihns are deposited at low temperature, which
inhibits growth of crystals and gives an amorphous film structure. The C-C sp3/sp2 ratio
of the films is mamly determined by ion bombardment energy and carbon ion flux to
neutral flux ratio to the substrate during the deposition process. The hydrogen content in
the films is also strongly dependent on ion bombardment energy.

For the deposition of films with high sp3 hybridization sites (ta-OH), the carbon

ion flux to neutral flux ratio onto the substrate is high and the ion bombardment energy

 

 

 

must be at a certain appropriate value which is about lOOeV per carbon. The systems
used for the deposition of ta-C:H films are the plasma beam source system and filtered
ion beam systems, which provide high ion flux to neutral flux ratio onto the substrate.
The systems usually have low deposition rates of ta-C:H films.

In this investigation, a microwave ECR-CVD reactor with rf biased substrate
holder is used as the deposition system. The microwave ECR-CVD system creates a high
density of charged and excited species at low deposition pressure (< 10’3 Torr) and it has
a low deposition temperature. The high density of charged and excited species gives a
high deposition rate and the lack of electrodes in the system inhibits contamination of
depositing films giving high quality films. This investigation studies and applies the
ECR-CVD system to deposit a-C:H films with a range of properties. Specifically, the
deposition of a-C:H films with high sp3 carbon-carbon bonding percentages at rates

exceeding previous investigations and methods will be explored.

1.2 Research Objectives

The objective of this project is to investigate the deposition process and the film
properties of a-C:H films deposited using a low-pressure, high-density microwave plasma
source. The investigation will characterize a-C:H film properties including thickness,
density, optical gap, refiactive index and hydrogen content. The investigation will also
establish the deposition conditions such as rf induced substrate bias, deposition pressure,
argon flow ratio to hydrocarbon feed gas, substrate temperature, position of the substrate

and microwave input power that produce desired a-CzH film properties. Further, it will

establish the effects of deposition conditions on fihn properties in terms of discharge

 

 

 

properties such as ion energy, ion flux to neutral flux ratio, ion type and deposition
temperature and possible fihn deposition mechanisms. A specific technological goal is to
deposit a-CzH films with a high sp3 carbon bonding percentage at a high deposition rate

exceeding 50 nm/min.

1.3 Research Methods

The methodology to be followed divides the deposition process into three sets of
variables including input deposition reactor variables, plasma deposition internal
variables and output variables / film properties. In the deposition of the a-C:H fihns using
the microwave ECR-CVD system in this investigation, the input variables are rf induced
substrate bias voltage, input microwave power, chamber pressure, substrate
heating/cooling, feed gas type and flow rate of feed gases. The input variables then
determine the internal variables such as ion density and type, ion energy, neutral / radical
concentration and type, ion flux to neutral flux ratio of species onto the substrate, and
deposition surface temperature with which films are deposited. Lastly, the internal
variables determine the outputs that include the film compositions, i.e., percent sp3,
percent spz, hydrogen content and the film's deposition rate. The film properties
associated with the outputs are density, internal stress, hardness, index of refraction,
optical band gap and deposition rate.

The research plan has two major components.

1) Characterize and quantify for the a-C:H deposition process using Csz, CH4-Ar and
Csz-Ar gas feeds the relationships of (a) input variables to internal variables and (b)

internal variables to outputs.

 

 

 

 

 

2) Compare the deposition results measured in 1) above to the prediction of a model

found in the literature.

1.4 Research Outline

Chapter 2 reviews hydrogenated amorphous carbon films in the literature. The
composition and electronic structure of amorphous carbon films are reviewed to explain
a-C:H films. Deposition mechanism of amorphous carbon films is also reviewed to show
how the fihns are deposited and what determines the film's properties. Next the various
deposition systems of amorphous carbon films are introduced and their film properties
are shown. In Chapter 3 the deposition system of this investigation and the
characterization methods of discharge properties and fihn' properties are described. F ilrn
properties versus variation of discharge conditions or properties are presented and
explained in Chapter 4 and 5 . In Chapter 4 films are deposited from acetylene feed gas at
pressures in the submillitorr range and the film properties are compared to the deposition
models in the literature. The objectives in Chapter 4 are to deposit high sp3 carbon-carbon
bonded ta-C:H or diamond-like films using a microwave ECR plasma reactor at high
deposition rate and to understand the deposition process of the films by investigating the
effects of ion energy, ion flux to neutral flux ratio, deposition temperature and
hydrocarbon flow rate. In Chapter 5, films are deposited from acetylene-argon and
methane-argon gas feeds at pressures in the millitorr range to produce a range of film
properties and to compare the films from each gas feeds. The objectives in Chapter 5 are

to establish the variation of film properties possible by depositing the films at different

 

 

 

 

 

deposition conditions and to understand the deposition process of the films by
investigating the effects of rf induced substrate bias, pressure and argon flow ratio. In
Chapter 6, the results of this investigation are summarized and the conclusions are

presented.

 

 

Chapter 2

2. Hydrogenated Amorphous Carbon Films: A Review

2.1 Composition of a-C:H Films

In a-C:H films, each carbon atom forms sp3 or sp2 hybridization bonds with other
carbon atoms and hydrogen atoms. sp3 sites form four tetrahedral C-C or C-H 0 bonds
and sp2 sites form three C-C trigonal 6 bonds and one C-C or OH 1: bonds. The
schematic compositional structure is represented in Fig. 2-1[3l]. One group labeled A is
for sp3 bonded carbon atoms, and a group labeled B is for sp2 carbon atoms in a six-
membered aromatic ring. Two sp2 olefinic carbon atoms in a double bond are labeled C,
and one sp2 carbon atom in an isolated fiee radical site with a dangling bond (represented
by an arrow) is labeled D. The isolated free radical sites is believed to form it bonded
carbon pairs to lower the energy of the system. The 1: bonded carbon pairs ultimately
form an aromatic ring and nearby aromatic rings are further condensed to graphite
clusters of sp2 aromatic rings. An example of the graphite cluster with 5 aromatic rings is
shown in Fig. 2-2 [31]. Thus the sp2 sites are embedded in a sp3 bonded matrix as sp2
clusters and spatially localized in the structure of the a-C:H films. The sp2 carbon atoms

can also form five-membered rings. Some unbonded hydrogen atoms may also reside in

the amorphous film structure.

 

 

 

‘0'. "n.
0
O

 

  
 

Fig. 2 - 1: Schematic representation of hydrogenated amorphous carbon films
(a-C:H). The network is comprised of hydrogen ( O ), sp2 carbon

atoms ( Q) and sp3 carbon atoms (.). The lines represent the bonds

and the arrows represent the dangling bonds.

 

 

 

 

/\ /

\

Fig. 2 - 2: sp2 carbon atoms in a-C:H films in the form of a cluster
of six-membered aromatic rings.

 

Various forms of hydrocarbon films can be distinguished by the content of sp3

and sp2 carbon and hydrogen. The ternary phase diagram of the content of sp3 and sp2
carbon and hydrogen for hydrocarbon films [32, 33] is shown in Fig. 2-3. The area close
to the hydrogen-rich comer marks the region where no stable films can be formed. The
top sp3 corner corresponds to fully diamond-like carbon films, the bottom left corner
corresponds to fully graphite-like carbon films, and lastly the bottom right part near the
triangle of no film corresponds to polymer-like carbon films. The rigidity boundary
indicates the boundary between rigid and floppy, polymer-like networks. The position of
this boundary line depends on the number of aromatic rings in the sp2 graphitic clusters.
The diamond-like quality is proportional to the perpendicular distance above this line [3].
The tetrahedral hydrogenated amorphous carbon films (ta-CzH) are a-C:H films, that have
a high ratio of sp3 carbon sites, are shown in the shaded region above the a-C:H region it
the figure.

The properties of a-C:H films depends on the composition of the films [31]. The
diamond-like properties of films come fiom a high sp3 content, which makes the filn
structure over constrainted. On the other hand, a high hydrogen atom content in the filn

yields many monovalent C-H bonds. These bonds make the film structure

underconstrainted, which makes the film floppy. One sp2 carbon site forms 3 strong c

bondings in a plane and one weak 7! bonding perpendicular to the plane, thus it alsr
makes the films soft. Thus density, hardness and Young’s modulus are nearlj

proportional to sp3 C-C content of the films (Fig. 2-4) assuming a fixed hydrogen content

10

Sp3

     
   
  

ta-C:H

(Filtered 10“ (Plasma Beam Source)

Beam Source)
ta-C

Rigidity
Boundary
Smaller Optical

Mean N0 film

 

a-C:H
(Plasma Deposition from Hydrocarbon Gas)

Fig. 2 - 3: Ternary phase diagram of hydrocarbon films.

Densnty
Hardness

 

 

D
sp3 Fraction
A
a.
a
co
'5
s:
a
an
'3
.2
H
a.
O
>

 

 

sp2 Fraction

Fig. 2 - 4: These plots show the trends of properties of ta-C:H films
with the variation of sp3 fraction and sp2 fraction.

   

Other properties like the optical bandgap are determined primarily by the sp2 carbc
bondings as shown in Fig. 2-4 for ta-C:H films and discussed in more detail in the ne:

section.

2.2 Electronic Structure of a-C:H Films [34]

The density of states (DOS) is schematically shown in Fig. 2-5.[34]. A sp3 si'
forms 4 6 bonds and a Sp2 site forms three 0 bonds and one 71: bond. 0* and 7r* represe:
antibondings of o and 1: bonds. The 0' and 0* states form deep valence and conductic
band states and 7t and n* states form band edge states. Photoemission spectra shows
states are at the top of the valence band and their density can be used to extract an S]
bonding fraction [3 5]. The conduction band DOS has been probed by electron energy 10
spectroscopy (EELS), which shows a prepeak for the n* states of sp2 sites, which can 1

used to measure the sp2 bonding fraction in DLC films [3, 16, 36].

As seen in the previous section, sp2 sites form embedded clusters in a sp3 matr
thus the it states are localized. The 0 states are not localized except possibility for tl
states at the edge of the 0' and 0* bands, i.e. the tail states. a-C:H films show hi;
resistivity because the it states are localized and the gap between 0' and 0* is large. T]
optical bandgap is determined by the 1: and n* states. The optical bandgap is control].
by the distortion of sp2 rings or chains, not by the size of the clusters of sp2 states [3
38].

There are also defect states deep in the gap. The defect states in a-C:(H) filr

come fi'om isolated sp2 sites and dangling bonds which are not paired up as it bonds [39

13

Density of States

 

Valence Band Bandgap Conduction Band

 

 

 

 

‘5 i

 

 

\Defect State

Energy

Fig. 2 - 5: A schematic diagram of density of state (DOS) of a-C:H films,
which shows 0' and it states, and defect states.

14

and are strongly localized. These defect states generally make the a-C:(H) films show p-

type behavior [34].

2.3 Deposition Mechanism of Diamond-like a-C and a-C:H Films

2.3.1 Microscopic Process [2, 40—42]

The local bonding of a—C and a-C:H can be defined principally in terms of two
parameters, the hydrogen atom fraction and the sp3 bonding fraction or the analogous
macroscopic parameters hydrogen content and mass density. A model of the deposition
processes should be able to account for the variation of these parameters with the
deposition conditions.

In the model developed by Robertson [40], the sp3 bonding occurs due to the ion
flux into subsurface positions causing a metastable increase in density. In the highly
energetic conditions of ion bombardment, atomic hybridizations are expected to adjust
readily to the local density, becoming more sp2 if the density is low and more sp3 if the
density is high. The density will increase if an incident ion penetrates the first atomic
layer of the film and enters an interstitial, subsurface position, where it dissipates energy
to the neighboring atoms and acquires bulk bonding of the appropriate hybridization.
Lower energy ions do not penetrate but just stick to the surface. Higher energy ions
penetrate further and increase the density in deeper layers. However, the ion uses only
part of its energy in penetrating the surface. The excess energy dissipates quite rapidly in

a thermal spike, during which the excess density can relax. Hence, a maximum density

occurs at an optimum ion energy that maximizes the penetrative yield but minimizes the

 

relaxation of the density increment. Typical energies of various processes in carbon

deposition are listed in Table 2 - 1.

Table 2 - 1: Energies of various processes for carbon [43].

 

 

 

 

 

 

 

 

 

 

Item Energy (eV)
Sputtering yield by C+ ions : 0.15 500
Sputtering yield by Ar+ ions : 0.07 500
Displacement energy of carbon atoms in diamond 80
Displacement energy of carbon atoms in graphite 25
Bond energy of diamond 7.41
Intraplanar bond energy in graphite 7.43
i
Interplanar bond energy in graphite 0.86 i
C --H bond energy 3.5

 

 

The model by Robertson [40] considers film growth from a beam of flux F

containing a fraction ¢of fast ions of energy E. In steady state, the fraction of ions at

interstitial sites, n, is given by the difference between the penetration flux and the

relaxation flux. The expression is,

nF = f¢F — ,B¢Fn (2 - 1)

where f is the fraction of ions which penetrate the surface and ,6 is the number of

relaxation atoms per each impact ion. Then,

 

= f 2-2
" 1/¢+fl ( )

Ion Flux F Ion Flux F

v tr
Growth — _— Implanted —
A: I

 

 

 

Fraction (I-n) Fraction n

E
<—>

 

Fig. 2 - 6: How subplanted ions increase local density. A fraction n
penetrates the surface of the film while the fraction (I-n)
fails to penetrate and increases film thickness.

 

n is related to the density increment as follows and as shown in Fig. 2-6. During a time
AT, the deposition of non-penetrating atoms and ions adds a layer of density p0 and
thickness Ax on the top of the film,

Ax=F(l—n)At/p0 (2-3)

 

The interstitials formed by penetrating ions produce additional density of zip,
Ap=FnAt/Ax (2-4)

Thus the density increment, Ap, is

 

fl = n (2 - 5)
Po 1 _ n

Combining equations of (2-2) and (2—5) gives

_A_p = _f— (2 _ 6)
p l/¢ - f + l3

The penetration fraction or penetration probability, f, of ions is dependent on the
displacement threshold energy, E4, of target nuclei and the surface binding energy, E3. The
energy increases the kinetic energy of ions by E; as they enter the solid. Thus, surface
binding the net penetration threshold for free ions is EmhFEd-EB, Here, Ed and E3 are 25
eV and 4.5 eV, respectively [44]. The penetration probability, f can be calculated as a
function of ion energy E, with a simulation code such as TRIM [45]. Fig. 2-7 shows an
example of calculated penetration probability. The figure was redrawn after the figure in
Ref. [40]

The relaxation of density is described by the thermal spike model [46]. Each

incident ion produces a thermal spike. The excess ion energy in the lattice dissipates by

 

 

1.0 -

2= 2:
a so
I I

Penetration Probability, f
c
In

0.2 -

 

 

 

0.0

 

10 100 1000
Ion Energy (eV)

Fig. 2 - 7: An example of penetration probability of C+ ions into a-C.

Density (g/cms)

3.6

 

3.4 -

 

 

 

 

10 100 1000
Ion Energy (eV)

Fig. 2 - 8: An example of calculated dependence of density on ion energy.

20

thermal diffusivity, which, in turn, anneals the film structure. With the thermal spike

model, fl is calculated as,

 

Ei
134.0164on (2- 7)

where E,- is ion energy, p=vao/D, v is a typical phonon frequency, a0 is the bond length, D
is thermal diffusivity and E0 is activation energy of annealing. E0 is estimated from the
thermal stability of mass selected ion beam (MSIB) a-C films, whose sp3 bondings
transform at T,=750 C to sp2 bondings. The value of E0 is estimated to be 3.1 eV as
determined in [40]. Combining equations of (2-6) and (2-7) gives

A_P = f _
p0 1/¢ — f + 0.016p(Ei/E0) (2 8)

Thus, the density increment depends on two terms f and fl, and two parameters, net
penetration threshold Emh) which is embedded in the penetration probability, f, and
activation energy of annealing, E0. In general, Em) controls onset of densification at low
ion energies and E0 controls the decrease of density at higher ion energy. An example of
a typical trend of calculated dependence of density on ion energy from the above
equation (2-8) is shown in Fig. 2-8. This figure was redrawn after the figure in Ref. [40].
The relaxation process is also modeled in other literature [47].

In the processes of deposition of a-C:H films, the a—C:H surface layer must be
dehydrogenated to form a solid film. Thermal dehydrogenation would leave undesired sp2
sites because the reaction to form sp2 bondings is more favored than sp3 bondings at
higher temperatures. Ion bombardment can also dehydrogenate and leave sp3 sites. The
incident ions dehydrogenate a—C:H by the preferential displacement of hydrogen atoms.

Hydrogen atoms are preferentially displaced as its displacement threshold is much lower

21

 

than C (3.5 eV versus 25 eV in Table 2-1]), basically because a hydrogen atom is
monovalent whereas a carbon atom is bonded to four other atoms. The liberated H atoms

recombine into hydrogen molecules that then effuse through micropores from the film.

2.3.2 Macroscopic Process

Properties of a-C:H films are mainly determined by the film's sp3 fraction and
hydrogen content, which are strongly dependent on plasma discharge properties and
deposition conditions. The main factors that determine the sp3 fraction and hydrogen
content are substrate temperature, energy of ions impacting the substrate and ion flux to
neutral flux ratio onto the substrate. Generally to obtain a-C:H films with high mass
density and low hydrogen concentration, the desired conditions include a highly ionized
plasma (high ion flux to neutral flux ratio), a low substrate temperature and a controlled
ion energy (typically about 100 eV per carbon [3]).

Ion bombardment energy onto the substrate can be readily controlled by the
magnitude of a negative substrate bias. The kinetic energy of ions impacts or bombards
the growing a—C:H films. As described above, ion species containing carbon atoms with
sufficient kinetic energy penetrate the growing films to a certain depth, dissociate
themselves expelling hydrogen atoms, relax their impact energy to surrounding bonds
and are eventually incorporated into the growing films. The magnitude of the impacting
ion bombardment energy determines whether the newly incorporating carbon atoms form
bondings of sp3 or sp2 at low deposition temperatures. In general the ion bombardment
energy should not be too low, which forms polymer-like films, and not too high, which
forms sp2 graphite-like films. The bombarding ions may include ions of inert gases. The

inert ions only bombard the fihns giving energy to dissociate the adsorbed or

22

 

 

incorporated hydrocarbon species and to form new bondings of the hydrocarbon species.
Most of the inert ions will not be incorporated into the growing films.

From the above brief discussion of effects of ion bombardment energy, it can be
inferred that the flux of ions onto the substrate should be high. And the ions in the flux
should have the same charge and similar masses so that they give an appropriate uniform
ion impacting effect that is needed to form the high sp3 bonding ratio of diamond-like
carbon films. The neutral radicals or molecules are not accelerated by the substrate bias,
thus, they cannot give the film impacting energy like the ions. Rather, the neutral radicals
and molecules are just adsorbed on the surface of growing films where they may
dissociate, but they do not affect the bonds below the surface. The neutral adsorbed
species on the surface of growing films can however disturb the ion impacting effect,
which results in lower sp3 C-C ratios. The presence of ions of different energies and
masses can also disturb the desired deposition process. Thus it is important to have a high
ratio of ion fluxes to neutral flux and a uniform ion energy to deposit hard diamond—like
carbon films containing a high sp3 fraction.

a-C:H films are usually deposited significantly below 350 C for hard diamond-
like carbon films. The low temperature (often less than 100 C) inhibits the formation of
sp2 sites in a-C:H films. A high ratio of sp2 graphite sites, generally resulting from films
deposited at high temperature, makes the films soft. The films are then graphite-like
carbon films. The substrate temperature is affected by the pressure of the plasma and the
power deposited by the bias on the substrate and it can be controlled by heating or

cooling the substrate holder.

23

 

 

2.4 Deposition Systems

Hydrogenated amorphous carbon (a-C:H) and amorphous carbon (a—C:H) films
have been deposited by a wide range of techniques. RF capacitively-coupled plasma
systems are widely used for a-C:H film deposition. The systems require a pressure of 10
mTorr-1 Torr to maintain the plasma [48]. One of the rf power electrodes is used as the
substrate holder and the electric field is normal to the substrate holder. The rf induced
substrate bias is created by the rf power provided to excite the plasma. Thus the rf
induced substrate bias is a function of the rf power. The high pressure induces collisions
in the sheath between the plasma and the substrate that reduces the ion energy onto the
growing film. And, the higher pressure produces a small ion flux to neutral flux ratio.
Because of above reasons, a much higher substrate bias is usually used in rf plasma
deposition of a-C:H films than other systems.

Zou et a1. [14] used a rf plasma system to deposit a-C:H films. The schematic
diagram of the experimental setup is shown in Fig. 2-9. In the rf plasma deposition
system the source gas was methane, the pressure range was 1mTorr to 100 mTorr, and
the substrate bias range was 0 to 1400V. In general, as the substrate bias voltage was
raised above -200 V the deposition rate (020 nm/min) increased, and the density (1.5-2.2
g/cm3), atomic ratio of hydrogen (0.13-0.33), stress (0.42-4 GPa) and hardness (1.5-5x103
kp/mmz) all decreased. The hydrogen content usually decreases with the increasing
substrate bias voltage as in other deposition methods. The sp3/sp2 ratio (1.5-3) increased

first and reached a maximum value and then decreased with the increasing substrate bias.

24

 

 

Pump

 

 

 

 

Substrate Gas Inlet
RF Power Supply
13.56 MHz _
( 0~-1400 V )
I Cooling Water

Fig. 2 - 9: Experimental setup of the rf plasma deposition system.

25

 

The results in the deposition were that soft polymer-like carbon films were deposited for
the range of 0~-100 V of substrate bias voltage, hard diamond-like films for —100~-600 V
and soft graphite-like films for 600-1400V.

Weiler et a1. [3] used a plasma beam source that provided a highly ionized
monoenergetic plasma beam of C2H2+ ions to deposit tetrahedral hydrogenated
amorphous carbon (ta-C:H) films, which have high sp3 bonding (up to 80 %) and which
are very hard. The schematic diagram of the plasma beam source is shown in Fig. 2-10.
This system works at 13.6 MHz rf power and uses acetylene gas at a pressure of 0.38
mTorr. Acetylene gas was selected because it forms mostly C2H2+ ions in low pressure
plasmas [49]. The gas flow rate was kept constant at 10 sccm. The pressure in the
background deposition chamber was 3.8 x10’2 mTorr. The ion energy distribution was
quite sharp (within 5%). The ion flux to neutral flux ratio on the substrate was estimated
to be 0.95. The deposited film density (2.2-2.9 g/cm3), sp3 fi'action (0.2-0.8), stress (29
GPa), Tauc optical gap (1.0-2.3 eV), Young's modulus (170-290 GPa) and hardness (23-
60 GPa) increased first, then reached a maximum (90-100 eV) and finally decreased (>
100 eV) with the variation of ion energy. The hydrogen content (22-28 %) showed a
decreasing trend with increasing ion energy.

P.J. Fallon et a1. [16] deposited a highly tetrahedrally bonded form of
nonhydrogenated amorphous carbon films (ta-C) with a filtered beam of C+ ions
produced by a cathodic carbon arc. A schematic figure of the system is shown in Fig. 2-
11. The ions produced by cathodic carbon are are filtered by a magnetic field filter which

selects only species of a fixed mass and ion charge. Therefore, the system provides the

26

 

 

 

= 4— Substrate
Holder

Extraction
Apertu re

 

 

 

 

 

 

 

Tungsten _
Grid
Plasma
Magnets
! RF Electrode
I l \ Movable Ceramic Pipe
RF Power Supply T
( 13.6 MHz) Gas Inlet
( Acetylene )

Fig. 2 - 10: Schematic diagram of the plasma beam source.

27

 

Substrate ( n-Si )

 

 

 

 

 

 

Magnetic , I
Field Filter +
a "" "'> C
i
I
Ion Source

( Cathodic Carbon Arc)

 

 

 

Fig. 2 - 11: Schematic diagram of filtered carbon ion beam system.

 

substrate with almost 100 % of carbon ions and films grow with mostly carbon ions. The
system, however, has a low deposition rate and needs a high vacuum level. The pressure
during deposition was 0.01 mTorr. The incident ion energy was varied by applying a
negative bias voltage to the substrate. sp3 fraction, and compressive stress passed through
a broad peak at an ion energy of about 120 eV. The density was roughly linear to the
percentage of sp3 bonding.
A mass selected ion beam deposition (MSIBD) method was used to deposit ta-C
films by Ronning and his coworkers [19]. The C+ ions were deposited with energies
between 20 eV and 1000 eV in an UHV-deposition chamber in which the pressure was
1e SS than 7.5x1045 mTorr. The sp3 fraction increased rapidly with increasing ion energy at
low ion energies and reached a broad maximum with 85 % sp3 bonded C atoms between
1 ()0 and 300 eV. For further increasing ion energies the sp3 fraction decreased slowly to
5 5 % for the ta—C films prepared with an ion energy of 1000 eV. They also pointed out
that the fraction of sp3 bonded C atoms of ta-C films deposited with vacuum are [16]
decreased dramatically for ion energies above about 200 eV resulting in sp2 bonded a-C
fillins. They attributed this sp3 to sp2 transition seen for the vacuum are deposited films to
local heating of the films due to a much higher ion flux as compared to MSIBD.
The ECR—CVD (electron cyclotron resonance-chemical vapor deposition) method
With a rf powered substrate holder is also used for deposition of a—C:H films. The ECR
plasma system excites the plasma through ECR heating which permits operation at
re l‘ltively low pressure as compared to rf capacitively-coupled discharges. The ECR-
CVD method with a rf powered substrate holder has a number of features making it an

ttl‘active method [50]. First, microwave ECR plasmas create a high densny of ion

29

 

 

 

 

 

 

 

2.45 GHz
Gas Inlet {— Microwave
(Hydrogen, —— Power
Methane)
E 1St Magnets
g 2“d Magnets
Substrate

’
13.56 MHz l /
RF Power _

Su , Substrate
pply Holder

Q

 

 

 

 

Fig. 2 - 12: Schematic diagram of one type of ECR-CVD system.

30

A

 

species (typically 1011-1012 cm'3) at low pressures in the submillitorr to a few millitorr
range. This combination of a high ion density with a low pressure (i.e., low neutral
density), can produce fluxes of species to the substrate that have a higher ion flux to
neutral ratio than capacitively rf coupled systems. The high ion plasma density also yields
a faster deposition rate, and the low pressure reduces substrate heating by the neutral gas
that assists in maintaining low substrate temperatures during deposition. The substrate
bias voltage is readily controlled in ECR-CVD systems by applying a bias to the
substrate. Both dc and rf biases have been used [20]. The most versatile approach that
permits the use of both conducting and insulating substrates is rf power applied to the
Substrate holder. Because of the difference in electron current flow and positive ion
entrent flow to the substrate, an induced dc self-bias is produced making the substrate
Surface negative, which attracts positive ions through the sheath formed between the
plasma and the substrate. At the low pressures used in ECR-CVD depositions, the ions
cross the sheath and arrive at the film without suffering significant collisions and with a
uniform energy produced by the difference in potential between the plasma and the dc
induced substrate bias. This use of a microwave ECR plasma source for the plasma
generation and a rf power supply for inducing the bias on the substrate allows the ion
density and the ion energy to be independently controlled. This control is useful for
O btaining desired film properties and for understanding the deposition process.
ECR-CVD systems have used a number of different precursor gases or gas
miXtures for the creation of hydrocarbon ions for the deposition of a-C:H films. Pure
lile‘ihane discharges have been used by Zarrabian et. a1. [21] at a pressure of 2.6 mTorr,

by Kuo, Kunhardt and Srivastsa [22] at pressures of 0.1-0.5 mTorr, by Fujita and

 

 

Matsumoto[23] at a pressure of 0.3 mTorr, and by Zeinert et. al.[24] at a pressure of 2.6
mTorr. Mixed methane/hydrogen discharges have been used by Pastel and Varhue [25] at
a pressure of 3 mTorr, and by Yoon and coworkers [20, 26, 27] at pressures of 6-15
mTorr. The work by Pool and Shing [28] used a hydrogen gas flow into the plasma
generation region and a downstream injection of methane at pressures of 5-55 mTorr.
Work by Andry, Pastel and Varhue [29] used both pure benzene and pure methane
discharges operating at pressures of 0.3-3.0 mTorr. Another gas mixture utilized was an
argon/methane mixture by Kuramoto et. a1. [30] in which the argon was injected in the
plasma source region and the methane was injected downstream at a pressure of 0.7
mTorr. Yoon and coworkers [20, 26, 27] used a system as shown in Fig. 2-12. They
Showed that the film hardness increased to a maximum value of ~17 GPa at a substrate
bias voltage of -50 V and then decreased thereafter in the range of -50 to -200V substrate
bias voltage for the films deposited at 7 mTorr. The optical gaps, E04, of the films
decreased rapidly to a minimum (~2.4 eV) at -50 V and then increased slightly thereafter.
The IR spectra suggested that the intensity of the OH absorption peak (bonded hydrogen
Content) decreased as the induced substrate bias increased. In general, a clear
uIilderstanding of the ECR-CVD deposition process for a-C:H films is lacking and
diIii‘erent researchers using different ECR systems often get different results.
Table 2-2 compares selected deposition parameters of several deposition methods.
RF plasma assisted techniques are operated at high pressure and have low ion density,
and these give a low ion flux to neutral flux ratio onto the substrate. The advantage of the
rf Plasma assisted methods is that relatively high deposition rates on a large area are

130 Ssible. And the disadvantage is that it is hard to control the substrate bias and input

32

 

 

power independently [12]. The ECR plasma methods have higher ion density and are
operated at lower pressure, which give a higher ion flux to neutral flux ratio onto the
substrate. The advantage of ECR plasma methods is that they have high deposition rates
and good control of the substrate bias and the substrate position relative to plasma.
Filtered ion beam and plasma beam methods have very high ion flux ratios onto the

substrate, but the ion flux to the substrate is low resulting in low deposition rates.

Table 2 - 2: Comparison of deposition methods

 

 

 

 

 

Ion Pressure Neutral Ion Electron Dep. Ref.
Method Density (mTorr) Density Flux/ Temp. Rate
(cm'3) (cm'3) Neutral (eV) (nm/ min)
flux

RI 109~10“ 10~1000 3x10l4 ~10“ 1~5 ~20 [14,48]
CVD ~3x10'6
ECR 10’°~1012 0.5~15 1x1012 ~102 2~7 ~50 [20,48]
P lasma ~5x10'4

F i ltered 0.01 ~1 low [16]

I o n

B earn

Plasma 0.037 0.95 15 [3]

B earn

 

 

 

 

 

 

 

 

 

33

 

 

Chapter 3
3. The Film Deposition System and Film Characterization Methods

3.1 Introduction

The hydrogenated amorphous carbon films of this investigation were deposited

with a multipolar electron cyclotron resonance (ECR) microwave-cavity discharge
system. In this chapter, the basic principle of electron cyclotron heating is explained and
the multipolar electron cyclotron resonance microwave-cavity discharge deposition
System is described. This system includes a microwave cavity, a baseplate, a microwave
power unit, a deposition chamber, a substrate holder biased with a rf power supply unit,
pressure gauges, etc. The methods used to characterize the properties of discharges in the
reactor are also described. This chapter then describes pre-deposition preparation of
substrates, and the treatment of samples after deposition. Next, characterization methods
of the properties of the films are described and explained. The properties of films
Characterized include thickness of films, mass density, hydrogen content, index of

refraction and optical bandgap.

3-2 Electron Cyclotron Resonance

Microwave ECR discharge systems can generate high densities of plasma at low
p r eSsure with low neutral gas temperature. This capability of microwave ECR systems
yie Ids applications in etching and thin film deposition. In this section, the basic principles
of an ECR discharge are explained. The microwave power source provides the deposition

ystem With the energy to maintain the discharge in the microwave ECR system The

34

L

 

 

electrons are heated by the microwave energy through the ECR effect in the system. Then
the heated electrons ionize, dissociate and excite the species of gases maintaining the
discharge inside the discharge chamber. The ECR heating of electrons occurs when the

frequency of the microwave energy, a), is equal to the electron cyclotron frequency 0),,

somewhere in discharge region. That is,
a) = (0,, (3'1)

The electron cyclotron fiequency is expressed as

9—3 (3-2)

Where q is the electron charge, B is an externally applied magnetic field strength, and me

is electron mass.

The principle of ECR heating of electrons is illustrated in Fig. 3-1 [51]. The
electric field vector of the microwave fields can be decomposed into the sum of a right
hand polarized (RHP) vector and a left hand polarized (LI-1P) vector. As shown in Fig. 3-

1 , a steady state RHP electric field E is directed in the xy plane and an uniform magnetic
field B is applied externally along the z direction. Then an electron in the magnetic field
gyrates in a right hand direction at the fiequency of are. Each figure, (a), (b), (c) and (d)
in Fig. 3-1 represents electron motion and direction change of the right hand polarized

electric field of the microwave in every one quarter period. With the fiequency of the
111i<>rowave given by, co= race, for the right hand polarized wave and an electron, the force,
‘95 continuously accelerates the electron in the direction of the motion of the electron.

Thus, the electron gains energy continuously. For the left hand polarized electric field,

the force, -qE is parallel to the motion of electron for the first one quarter period and the

35

 

 

 

Fig. 3 - 1: Principle of ECR heating. The electron gains microwave
energy continuously.

36

A—_____—__

third one quarter period and opposite for the second one quarter period and the fourth
one quarter period. Thus, the electron energy oscillates with no time average energy gain
for the left hand polarized electric field of the microwave. From the figure, it can be seen
that the efficiency of ECR heating is maximized when the electric field of the microwave,
E, and the externally applied magnetic field, B, are perpendicular. The microwave
frequency is usually 2.45 GHz. Thus, the strength of the magnetic field is obtained to be

875 Gauss for the ECR heating from equations (3-1) and (3-2).

3.3 Deposition System and Conditions
3.3.1 Description of the Deposition System

Hydrogenated amorphous carbon (a-C:H) films in this investigation were
deposited using a microwave ECR deposition system as shown in Fig. 3-2 [52-59]. The
main parts of the system consist of a cylindrical microwave cavity, a deposition chamber,
a baseplate between microwave cavity and deposition chamber, a substrate holder, a
pumping system, a microwave power supply unit, an rf power supply unit, and a gas
supply unit [60, 61]. The side view of the microwave cavity, baseplate, and deposition
chamber parts is shown in detail in Fig. 3-3. The diameter of the cylindrical microwave
cavity (LC) is fixed at 17.6 cm and its height (L5) is changeable to tune the microwave
cavity. The cylindrical quartz dome inside the microwave cavity has dimensions of 9 cm
diameter (Lq) and 5 cm height (L1,). The baseplate consists of the upper and the lower
parts. The heights of upper part (L...) and lower part (Lb) are 2.0 cm and 2.8 cm,
respectively. s.p. in the figure designates the distance between the bottom of the baseplate

and the substrate and is varied to give ion flux variation onto the substrate.

37

 

Microwave
Power Supply

9 G

 

 

 

 

Microwave Cavity

Deposition Chamber

 

 

 

 

 

RF
Power Supply

9 G

 

Gate Valve

Conductance alve

 

 

 

 

 

Diffusion

 

 

j

“I:

 
      
 

g  _+

Flow Meters

t

I ®-El-®-

Pressure
Meters

 

 

 

Gas Tanks

ll

Roughing

 

 

 

Pump

 

 

 

Fig. 3 - 2: The microwave ECR plasma source with the rf biased substrate holder.

38

 

Sliding Short

 

Microwave
/ Cavity Body
4— Lc
2'45 GHZ L LS Quartz Dome

/

   
    
  

Permanent
hdagnet

Microwave Power pr“— Lq __>|

 

 

 

 

aseplate

......
.......

oooooo

 

 

I" 3’ I52 ' l
Deposition l 7 A Gas Inlet
Chamber s.p.
P A ‘___ __ Substrate Holder

 

 

 

 

 

 

13.56 MHz J EFI ”’4 7
RF Power (M _ '1
lTurbopump r:

 

 

 

 

i
— I

Sate Valve

._<8>__

 

 

Fig. 3 - 3: The side view of the microwave cavity, the baseplate, and the
deposition chamber of the system.

39

The length of the microwave antenna probe (LP) is changeable to match the impedance
between the microwave source and the microwave cavity. The substrate holder is a
rectangular and has dimensions of 10.5 cm x 5.0 cm.

The cylindrical microwave cavity intensifies a specific microwave mode in it and the
intensified microwave energy produces the plasma of the feed gases in a cylindrical
quartz dome inside the cavity via the ECR heating effect. The microwaves are introduced
into the cavity through a microwave antenna that is the core part of the coaxial
transmission line and is inserted through the side of the microwave cavity. The
impedance of the microwave source unit is matched to that of the microwave cavity by
adjusting the height of the microwave cavity (Ls) and the length of the microwave
antenna (LP). The microwave source unit generates microwave energy, controls the
power of the microwaves and sends the microwaves to the microwave antenna via
waveguides. The microwave source unit consists of a microwave power supply (Model
4074, Therrnex INC.), a microwave power source controller (Model 4006, Thermex
INC.) and two microwave power meters (Model 432A, Hewlett Packard). The microwave
power meters measure the incident microwave power to the microwave cavity and the
reflected microwave power fiom the microwave cavity. The net power that generates the
discharge inside the quartz dome is the incident power minus the reflected power.

The gas supply unit sends and controls the flow rate of source gases of a-CzH
films from gas tanks into the quartz dome. It consists of three flow meters, three flow
controllers (MKS Instruments INC.) and gas tanks so that three kinds of gases can be
provided into the deposition system simultaneously. The rf power supply unit is

connected to the substrate holder inside the deposition chamber and provides the

40

 

substrate with the rf induced substrate negative bias. The rf induced substrate bias
provides ion with ion bombardment energy on the surface of growing films and its
determination will be discussed later in this section. It has the rf power supply (HFS-
500E, Plasma-Therm INC) and a matching network which matches the 13.56 MHz rf
power fi'om the rf power supply to the substrate holder. The substrate holder has a heater
so that the deposition temperature can be raised. However it has no active cooling, so the
temperature of the substrates increase some during the deposition even when the heater is
turned off. For example, the temperature increased 80-100 °C for 5 minutes of deposition
at 3 mTorr pressure. The height of the substrate holder can be varied so that the distance
of the substrate fi'om the region of plasma generation (or, from the bottom of base plate)
can be changed. The increasing distance will decrease the ionic flux of species to the
substrate. A typical value used is 3.5 cm below the bottom of the baseplate. To measure
the pressure in the deposition chamber, a capacitance manometer (Type 627, MKS
Instruments INC.) and a hot cathode pressure gauge with yttrium coated iridium
filaments (MKS Instruments INC.) are used. The capacitance manometer can determine
the pressure from 0.1 mTorr to 100 mTorr and is used to measure the discharge pressure
of the deposition chamber. The pressure controller (Type 651, MKS Instruments INC.)
reads the capacitance manometer and controls the gate conductance valve to achieve the
desired deposition pressure. The hot cathode pressure gauge, which can measure pressure
levels of microtorrs, is used to measure base vacuum level and is used to calibrate/zero
the capacitance manometer. The hot cathode controller (Type 919, MKS Instruments

INC) is used to read the hot cathode pressure gauge.

41

 

Cooling Water Line Gas Channel

  
  
   

N eodymium-Iron
-Boron Magnets

Pin Holes for
Gas Distribution

 

 

 

Polygon Shaped Iron ECR Surface

Retainer of Magnets

Fig. 3 - 4: The cross section (top view) of the baseplate of the system.

42

 

 

 

The plasma used to deposit the films is generated in the region of the baseplate
via ECR heating. The baseplate is illustrated in Fig. 3-3 and Fig. 3-4. The baseplate
consists of upper and lower parts. The upper part has eight neodymium-iron—boron-
magnets providing the magnetic field needed for the ECR coupling. The north and south
poles of magnets are alternated toward the discharge region and the ECR surface on
which ECR coupling occurs is drawn qualitatively in Fig. 3-4. The lower part has eight
pin holes for gas distribution. The holes are designed to eject the gas upward so the
ejected gas travels through the ECR zones and then diffuses downward to the substrate.
The plasma discharge can produce a lot of heat so the baseplate also has a water line to

cool the plate.

3.3.2 Deposition Conditions

The microwave ECR system generates a high density of plasma (typically 101°-
1012 cm’3) through ECR heating at low pressures in the submillitorr to a few millitorr
range. This combination of a high ion density with a low pressure (i.e., low neutral
density), can produce fluxes of species to the deposition substrate that have high ion flux
to neutral flux ratio. The system is operated with 200-400 W input microwave power to
create plasma inside the quartz dome. The range of rf induced substrate bias was varied
from 0 V to -3 00 V. The gases injected include primarily acetylene, methane-argon or
acetylene-argon mixtures with total flow rates ranging from 720 sccm. The argon flow
rate is varied from zero to 50% of the total flow rate. The plasma diffuses into the
deposition chamber where a rf biased substrate holder is located. The diffusion is

designated with the thick arrows in Fig. 3-3. The specific pressure range in this study is

43

0.2 to 5 mTorr. These pressure values insure that the ions moving through the sheath

experience no significant collisions.

3.3.3 Sample Preparation

In this study the substrates used are glass of either 76mm x 25 mm x 1 mm or 12
mm x 12 mm x 0.1 mm in size and pieces of silicon wafers. The glass substrates are
cleaned in methyl alcohol using an ultrasonic cleaner for 90 minutes and then rinsed in
deionized water for 30 minutes before deposition. For some cases, the glass substrate of
0.1 mm thickness is mounted on the substrate holder using a heat conducting paste on the
substrate holder side of the substrate. This helps to get a good thermal contact between
the substrate and the substrate holder keeping the temperature of a substrate low. The
sample with the heat conducting paste is cleaned again after deposition in methyl alcohol
for 90 minutes and then rinsed in deionized water for 30 minutes to remove the heat

conducting paste.

3.4 Characterization of Discharge Properties

The discharge properties such as electron temperature, plasma density and
saturation ion current of acetylene, methane-argon and acetylene-argon discharges
generated in the microwave ECR plasma reactor are characterized by using a double
Langmuir probe. The plasma sheath thickness above the substrate and ion energy onto the
substrate are estimated with the electron temperature, the plasma density and the rf
induced substrate bias using either the matrix sheath potential theory or Child Law sheath
potential theory. The gas composition of exit gas from the discharge chamber are

analyzed with a partial pressure analyzer (PPA), MKS 600A-PPT.

44

 

3.4.1 Double Langmuir Probe Measurement

The double Langmuir was used to measure the electron temperature and ion
saturation current for several deposition conditions. The double Langmuir probe has two
metal probes 0.5 mm in diameter and 5 mm long. The probes are separated by 3 mm. The
Langmuir probe measurements were taken at the substrate position, which was 8.5 cm
below the top of the quartz dome that con fines the plasma. The electron temperature of
discharge is determined from the I-V curve of double Langmuir probe measurement

using the below equation [51].

I. d] “
T =—-’91 —— _ 3-3
e 2 [dV V—OJ ( )

where 1.0,, is the ion saturation current of the I-V curve. The plasma density is obtained

from the ion saturation current and electron temperature using the following equation

[51].

1/2
nO = —-——I’°" M— (3—4)
0.61eA k7;

where e is the electron charge, A is surface area of the probe, k is Boltzman constant and
M is theion mass of the discharge. The substrate position was also 3.5 cm below the
plasma source opening (see Fig. 3-3). At this position the ECR static magnetic field is

low and does not interfere with the Langmuir probe measurements.

3.4.2 Determination of Ion Energy and Ion Flux
The ion bombardment energy onto the substrate and the ion flux to neutral flux

ratio are important factors for determination of fihn properties in the deposition process

45

 

as discussed in Chapter 2. The ion bombardment energy onto the surface of growing
films is determined by the potential difference between the plasma and the substrate. This
difference is determined primarily by the rf induced dc bias W of the substrate holder.
The induced dc bias on the substrate is measured with respect to the chamber wall
potential. The actual ion energy is the induced dc bias energy gain plus the potential
difference between the plasma and chamber walls. As a measure of this plasma-to-
chamber wall difference, the plasma potential with respect to the chamber wall potential
id estimated fiom the plasma sheath potential when the electron and positive ion fluxes to
the wall are equal. This condition of equal fluxes occurs when the surface is at the
floating potential. This potential between the plasma and an electrically floating wall is

kT, H;
+ .—

m=QMDe 28

(3-5)

 

where T e is the electron temperature, e is the electron charge, k is Boltzmann’s constant,
and C(MJ is a constant that depends on the ion mass [51]. For example, C=2.8 for the
hydrogen ion, 4.2 for the methane ion, 4.5 for the acetylene ion and 4.7 for the argon ion.
A potential difference may also exist between the floating potential of a surface in the
substrate holder region and the other chamber surfaces. This is due to: (l) ambipolar
diffusion effects producing spatial variations in the plasma potential, and (2) variations in
the electrical contact of the plasma to the various surfaces resulting fiom
insulating/partially conducting films on the surfaces that can change versus processing
history. Hence the energy of the ions is the energy gained due to the induced substrate
bias ¢rf plus the plasma sheath potential between the plasma and the chamber walls ¢w
plus the potential difference between the floating potential at the substrate location and

the chamber walls dim/y. The rf induced dc bias is measured by connecting a low pass filter

46

 

to the rf power supply output to provide a dc signal to a voltmeter. The sheath potential of
equation. (3-5) is determined from the electron temperature, whose measurement is
described in the next section. Finally, the potential my; at the substrate location is
measured by inserting a small sputter cleaned probe at the substrate location and
measuring the potential between the probe and the chamber walls. Typical values
measured for this potential ¢difl were 2-10 V. Its value depends on both the plasma
operating conditions and the history of previous depositions and cleanings of the chamber
walls. The rf induced substrate bias was varied from 0 to -3 00 V in this study.

The actual ion energy on the surface of the substrate will be smaller than the value
of |¢4j +¢W+¢d1fl discussed in the above paragraph because the insulating glass substrate
has finite thickness in the plasma sheath. The potential on the surface of plasma side of
the substrate is the potential of My] + ¢w ¢djfl minus the potential between the surface of
the substrate and the substrate holder. The potential on the surface of the substrate can be
obtained fiom the potential of IM+¢w+¢dfl the thickness of plasma sheath and the
thickness of the substrate. The thickness of plasma sheath and the potential at the surface
of the substrate are calculated using either the matrix sheath or Child Law sheath given in

the below equations [51]. For the matrix sheath,

2

 

<I>(x)=—"": x— (3-6)
80 2
and for Child Law sheath,

3 enSuB “2 23 4M 4/3 _
(D(X)_—[E[_:g:—] (M) X] (3 7)

47

 

 

 

where n, is the plasma density at the sheath edge (n,=0. 61:10), mg is Bohm velocity
(u3=(kTe/M”2), and M is the mass of the ion. The plasma density, no and the electron
temperature, T e were measured using a double Langmuir probe. The plasma sheath
thickness is calculated, then the potential at the suface of the substrate is calculated using
the above equations.

The ion flux to neutral flux ratio onto a substrate is estimated by calculating the
neutral flux using the discharge pressure and temperature and the ion flux from the
measurement of ion current density of the substrate holder. The neutral flux was

calculated with the equation:

F =§ng <v> (3-8)

)1

where n8 is the neutral density calculated from p=nkT, and <v> is an average neutral
velocity in the discharge given by <v>=(8kT/er/0”2. The ion current density of the
substrate holder was measured by applying DC bias on the substrate holder in a discharge
and measuring the DC current. The ion flux, I", was estimated by dividing the ion current

density by the electron charge. Then the ion flux to neutral flux ratio is Fi/l'n.

3.4.3 Partial Pressure Analysis of Exit Gas and Temperature Measurement

The partial pressure analyzer (PPA) is used to measure the composition of the exit
gas. The sampling point is near the bottom of the processing chamber where the residual
exit gas is pumped out as shown in Fig. 3-2. The chamber gas composition was sampled
by connecting a 6 mm diameter tube :from the chamber to a turbomolecular pump/PPA
unit. The partial pressures reported in this investigation are the pressures measured at the

PPA location; hence they are lower than the chamber pressure. The relative partial

48

 

pressures at the PPA location are assumed to be representative of the species
concentration in the chamber. The substrate temperature was measured with an F-type
thermocouple. The thermocouple was attached to the substrate using a heat conducting
paste to keep it a good thermal contact with the substrate and no direct contact with the

discharge gas.

3.5 Characterization of a-C:H Films

For the determination of the deposited film's thickness, optical absorption
coefficient and refractive index, the transmission and reflection data of light is measured
by using a visible-near infrared spectrometer for the wavelengths of 400-1600 nm. The
spectrometer is illustrated in ( a ) of Fig. 3-5. In the figure, the Xe-Hg arc lamp is used for
the light source. The arc lamp igniter (model 68705, Oriel Corporation) ignites the lamp
and the arc lamp supply (model 68700, Oriel Corporation) is used to give the lamp
power. The monochromator (model 77200, M: M — 2 nm, Oriel Corporation) selects a
specific light of wavelength. The interval of wavelengths was chosen to be 20 nm. The
photosensor (model 814-SL silicon detector for light of 400-1000 nm, and model 818-IR
germanium detector for the light of 800-1600 nm, Newport Corporation) is used to detect
the transmitted and reflected light from the film The optical power meter (model 835,
Newport Corporation) measures the power of the light that is detected with the
photosensors. The lens of focal length L1 (16 cm) focuses the light into the entrance slit
S1 (fixed at 600 pm) of the monochromator, the lens of focal length L; (1.5 cm) focuses
the light from the slit 8; (700 pm) on the substrate which has the film to be characterized
and the lens of focal length L3 (1.5 cm) collects the light into the sensor of the

photometer. The filter F1 preselects some spectral band of light to remove the second

49

Arc Lamp Arc Lamp Igniter Arc Lamp Supply

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Ll
F1
8‘ Film/Substrate
Monochromator r Photometer .
Optical
Sz <>> <.>- ' Power
Meter
L2 '_ L3
(3)
Film/Substrate
Incident Transmitted
G Reflected \ Photometer
Photometer
(b)

Fig. 3 - 5: (a) The spectrometer used to measure the transmittance and
reflectacnce of the films and (b) the tilted angle needed to
measure the reflected beam.

50

 

Reflected

Incident

1

 

 

 

 

air ’73 :1

A
a-C:H film 7 d2 71‘, = n2 — 1k,
I
it, — 1.54
substrate d'

7

air r70 = l

l

Transmitted

Fig. 3 - 6: The measurement of transmittance and reflectance of light

for an a-C:H film on glass substrate.

 

order wavelength of light from the monochromator. Three different filters are used and

their preselecting bands are 400—650 nm, 650-1000 nm and 850-1600 nm. The incident

light is not exactly normal to the surface of the film, but slightly tilted to measure the

reflected beam. The tilted angle, (1. between the incident and reflected beams is illustrated

in( b) of Fig. 3-5 and is about 15°.

The transmittance and reflectance of an a-C:H film deposited on a glass substrate

is shown schematically in Fig. 3-6. The equations [62] describing the transmittance T

and the reflectance R in terms of the quantities defined in Fig. 3-6 are

ital '21 1— r 10
Wfil
= Ir '2 + —————lt’ '4 r102
I " Irltlz’io
where
t 32 exPUflz )t 21

 

I _.

1‘ r23r21 eXP("2ifl2)

= r32 +r21 eXP(2ifl2)
I 1" r23r21 exP(‘—2ifl2)

 

r t: ’12 +r23 exP(2iflz)
I 1 "”2351 CXP(—2ifl2)

 

and
r10=(fil—fiO)/(fil +fio)
r21=(;iZ _fii)/(fiz +fil)
r12 :_r21

r32 =(fi3 ‘fi2)/(fi3 +fi2)

52

(3-9)

(3-10)

r 23 = —r 32

I10 =2'r‘il/(Fil + 170)

In =271’2/(1’7i2 +171)

t32 =2fi3/(ii'3 +52)

,62 = 27;??de //1

The indices of refiaction are

170 =1 for air,

171 =1.54 for the glass substrate,

52 2 n2 — ilc2 for the a-C:H film and

7:} =1 for air
17, is the complex index of refraction of the deposited film where n2 is the conventional
real index of refraction and K2 is the extinction coefficient of the films. For the
measured transmittance, T and reflectance, R, M2 and K2 are found from the above

equations of T and R. The nonlinear equations of (3-9) and (3-10) are solved by using the

command of 'fsolve' in Matlab. Then the absorption coefficient, a, is also found from the

extinction coefficient, K2 , with the following equation,

47m2
a = 3-11
A ( )

 

An example of the measured transmittance and reflectance is shown in Fig. 3-7.
The thickness and index of refraction are determined by comparing the measured

reflectance of light with a modeled reflectance found by choosing appropriate film

53

 

1.0

 

     
 

0.8 - 00°
00
00
§ 00040— Transmittance
'3 00°
‘53 0.6 - o°°°
a ,o
a: F o
1:
= o
a O
8 0.4 -
g o
.35 ° Modeled Reflectance
E - o
3 Measured Reflectance
g; 0.2

 

 

 

400 600 800 1000 1200 1400 1600
Wavelength (nm)

Fig. 3 - 7: The transmittance and reflectance of an a-C:H film versus
wavelength. The modeled reflectance data is simulated for the
determination of thickness and index of refraction of the film.

54

 

 

 

thickness and index of refraction values. The thickness of films is mainly determined by
the periodicity of the reflectance and the index of refraction by the amplitude of the
reflectance At the shorter wavelengths the oscillating amplitude of reflection data is faded
due to the absorption of light by the a-C:H film. The modeled reflectance is not faded
because the absorption of light is not considered in the model. In the example of Fig. 3-7,
the modeled value of thickness and the index of refiaction are 310 nm and 2.4,
respectively. The thickness can be determined within 10 % error bound of its value.
Then, the thickness is returned to equations of (3-9) and (3-10) for the calculation of the
extinction coefficient, K2 and index of refraction, n; with the variation of the light
frequency. Thickness was also checked on several films using SEM cross- sections. An
example of a SEM cross-section is shown in Fig. 3-8. The thickness from the optical
method coincides with that of the SEM picture within 10 %. The absorption coefficient
versus wavelength is calculated from equation (3-11) for films whose thickness is less
than 300 nm. Fig. 3-9 and Fig. 3-10 show typical plots of the index of refraction and
absorption coefficient of the film versus photon energy which are found by using

equations (3-9) and (3-10).
However the above procedure to find 11 2 and K2 does not work for the films

whose thickness is more than 300 nm because the command 'fsolve' in Matlab has
difiiculty to find the convergence point of the solutions. Thus, for thicker films an
alternative simplifed equation is used to find the absorption coefficient, a, fiom the

measured transmittance, T, and reflectance, R. The simplified equation is expressed as

55

 

 

 

Fig. 3 - 8: An example of a SEM cross-section for determination
of thickness of the film.

56

 

-1

Absorption Coefficient (cm)

 

 

1.0E+05

r l I I

 

1.0E+04

ITTI

E04

1.0E+03
1.4 1.8 2.2 2.6 3.0

Eiergy (eV)

 

 

 

 

Fig. 3 - 9: Absorption coefficient of an a-C:H film versus photon energy.

57

 

Index of Refraction

2.4

2.2

2.0

1.8

1.6

 

 

1 1 I I I l 1 l

 

 

1.8 2.2 2.6 3.0
Energy (eV)

1.4

Fig. 3 - 10: Refractive index of an a-C:H film versus photon energy.

58

T=(1—R)exp(- ad) (3 - 12)
where d is film thickness. The index of refi'action is found using the amplitude of

reflectance in Fig. 3-7 and the below equation.

2
“(’54) (343)
n2+l

The index of refiaction, thus, is an average value over the light frequency of the

 

measurement.

The E04 bandgap is determined by reading the photon energy at 0L=104 cm'1 in the
absorption versus photon energy plot (see Fig. 3-9). Tauc optical gap is derived from the
absorption coefficient using a Tauc plot [62]. It is often observed in semiconducting
glasses that at high absorption levels (on>104 cm'l) the absorption constant, a, has the
following fiequency dependence:

hvat=A(hv—E,m)2 (3-14)
where A is a proportional constant, h is Plank's constant, v is angular light frequency, and
EM is the Tauc optical bandgap. From the above relation the Tauc plot can be drawn as
in Fig. 3-11. The Tauc optical bandgap is then the energy (1.3 eV in the Fig. 3-11
example) at the point that the linear fit meets the x-axis.

The density of the films was determined by measuring the mass difference of the
virgin substrate before deposition and the film deposited substrate after deposition using a
Denver Instrument M-220D balance. The balance has 31 g of maximum capacity, 1.0 mg
of minimum capacity and 0.01 mg of readibility. The glass substrate samples used for

mass measurements were 1.4 cm2 and 0.17 mm thick. The mass change was typically

59

 

 

 

0.15-0.3 mg for the films deposited at pressures in the millitorr range. The mass density

was determined from the measured mass of the film and its volume. Since both the
thickness measurement and the mass measurement have uncertainties of 5-10 %, the
uncertainty in the mass density measurement is 10-20 %.

For determination of the hydrogen content of a—C:H films, the infrared spectra for
the region from 2800 cm'1 to 3200 cm"1 was obtained by using a Beckman IR 4220
spectrometer. The infrared absorption coefficient is derived fiom the transmission spectra
and thickness data with equation (3-12). The typical absorption coefficient as fiinction of
wavenumber is shown Fig. 3-12. The peak C at 2956 cm'1 is due to spz-CHZ (olefinic)
bonds, the peak B at 2920 cm" is due to the sp3-CH2 asymmetric stretching mode bonds,
and the peak A at 2870 cm'1 is due to the sp3-CH3 symmetric stretching mode bonds [63].
The infrared absorption coefiicient is used to determine the bonded hydrogen content of

the a—C:H films using the equation [64],
n” =A[gda) (3-15)
(0

where n” is hydrogen content in cm'3, A is equal to (1.35 i 0.3) x 1021 cm"1 [65], or is the
absorption coefficient and (0 is the infrared light fiequency. The hydrogen content in
atomic percentage was obtained from the hydrogen content and the mass density of the
film as follows:

_XH+XC

3-16

p N. ( )

H=i—x100 (3-17)
X” +XC

60

 

600

500

400

(hva)l/2 (eV cm")“2
(a)
G
C

N
G
G

100 Etauc

 

 

 

0.0 1.0 2.0 3.0 4.0
Energy (eV)

Fig. 3 - 11: Tauc plot of an a-CzH film.

6]

 

Absorption coefficient (cm-1)

 

200

150 -

100 -

 

U!
G
I

 

   

 

 

Ill

2600 2800 3000 3200
Wavenumber (cm-1)

l 1 l

 

_50 1 I

Fig. 3 - 12: An example IR absorption spectra.

62

 

where XH and XC are atomic concentration of hydrogen and carbon, respectively, N0 is
Avogadro number, p is mass density and H is atomic percentage of hydrogen. X H and p
are determined by the measurement described previously, thus XC can be obtained with
equation (3-16) and atomic percentage of hydrogen H in a-C:H films is obtained with
equation (3-17). It should be noted that this method only detects the hydrogen bonded to

carbon in the film and not free hydrogen trapped in the film.

63

Chapter 4

4. Films Deposited from Acetylene Discharges at Pressurs in the
Submillitorr Range

4.1. Introduction
Hydrogenated amorphous carbon fihns were deposited at pressures in the
submillitorr range, and the dependence of the film properties on the ion impacting energy
and the ion flux to neutral flux ratio onto the substrate during deposition are investigated
in this chapter. According to the subplantation model of deposition as discussed in
Chapter 2, the film has a maximum sp3 ratio when it is deposited with approximately 100
eV of impacting ion energy per carbon and with high ion flux to neutral flux ratio onto
the substrate.
In this chapter, the objectives are to deposit high sp3 carbon-carbon bonded ta-
C:H or diamond-like films using a microwave ECR plasma reactor at high deposition rate
and to understand the deposition process of the films by investigating the effects of ion
energy, ion flux to neutral flux ratio, deposition temperature and hydrocarbon flow rate.
The rf induced substrate bias was varied from -80 V to -300 V to give variation of the ion
energy onto the substrate. The deposition pressure was reduced as low as the deposition
system could operate to maximize the ion flux to neutral flux ratio onto the substrate. The
films were mostly deposited at 0.2 mTorr for about 45 seconds of deposition time at a
substrate temperature near room temperature using acetylene gas in most cases.
Acetylene was selected as the source gas of the films because it forms mostly Csz+ ions
in low pressure plasmas [3]. Thus, the acetylene discharge at low pressure can give

relatively uniform impacting ion energy per carbon onto the substrate. The typical flow

64

 

 

rate of acetylene gas was 7 seem. The typical incident microwave power was 600 W. The
reflected microwave power was high (about 350 W) when compared to other gas
mixtures that contain inert gas under similar pressure conditions, and it varied with
deposition condition such as pressure, substrate distance below the base plate and gas
mixtures. The typical net absorbed microwave power for the acetylene gas discharges
was about 250 W. The above deposition conditions will be termed to be the nominal
deposition condition of low pressure deposition, hereafter in this chapter. The substrates
are 0.17 mm thick glass and are mounted on a substrate holder with a heat sink
compound, i.e. thermally conducting paste, to keep the temperature of the substrate near
room temperature during the deposition process. Before film deposition, the substrates
were cleaned in methanol using an ultrasonic cleaner and sputtered in an argon discharge
for 30 seconds to further clean the substrate surface. After the deposition the films were
again cleaned in methanol with a ultrasonic cleaner to remove the heat sink compound.

The input variables were varied to develop an understanding of the film properties
in relation to the deposition conditions. Some specific relationships studied and reported
in the following sections are input variable variations of feed gas rate, ion energy,
substrate temperature and ion flux to neutral flux ratio. A summary of the deposition
variables used is presented in Table 4— 1.

In the following sections, the discharge prOperties measured at pressures in the
submillitorr range are first presented and discussed. In particular they will be used to
provide some explanation to the film properties versus variation of the input variables.

Next, the effects of the ion energy and the ion flux to neutral flux ratio are shown and

65

discussed. Lastly, the effects of the deposition. temperature and the flow rate of source gas
are investigated.

Table 4-1: The deposition variable space.

 

 

 

 

 

 

 

 

 

Input Variables Nominal Value Variable Range
Acetylene Flow Rate (sccm) 7 4 - 35
Pressure (mTorr) 2.0 0.2 - 0.6
RF Induced Substrate Bias (-V) 200 80 - 300
Absorbed Microwave Power (W) 250 170 - 400
Substrate Position (cm) 3.5 3.5, 6.0

Substrate Thickness and Type 0.17 mm glass 0.17 or 1.0 mm glass

Helium Flow Rate (sccm) 0 0.0 - 2.5

 

 

 

 

4.2. Discharge Properties at Pressures in the Submillitorr Range

Selected discharge properties at pressures in the submillitorr range were measured
using a double Langmuir probe and partial pressure analyzer (PPA). Specifically, the
double Langmuir probe was used to measure the electron temperature and plasma
density. The plasma sheath thickness and ion energy onto the substrate are then estimated
using the result of the Langmuir probe measurements. In another set of experiments, a
PPA measured the partial pressures of exit gases out of the discharge chamber.

The electron temperature and plasma density of acetylene discharges at 0.2 mTorr
were measured for three different discharge conditions as indicated in Table 4-2. During
the measurements the discharges coated the Langmuir probe with a-C:H films very

quickly so the voltage on the probes was swept fast, i.e. within 30 seconds. The results of

66

the Langmuir probe measurements are presented in Table 4-2. The discharge pressure
was 0.2 mTorr and the flow rate of acetylene gas was 7 seem for the measurements. The
electron temperatures vary about 8 eV and the plasma densities are similar to one another
in Table 4-2. The fluctuation of the electron temperatures is considered due to the
unstable measurement caused by the coating effects of a-C:H films on the probes.

Table 4 - 2: Double Langmuir probe measurements for electron temperature and plasma
density.

 

 

 

 

 

Microwave Power (W) Saturation Ion Electron Plasma Density
: Incident/Reflected Current (mA) Temperature (eV) (x 1010 cm'3)
(Net Input Power)
600/400 (200) 0.12 8.1 2.1
600/350 (250) 0.12 8.8 2.2
600/350 (250) 0.12 7.7 2.2

 

 

 

 

Using the double Langmuir probe measurements, the plasma sheath thickness and
the energy of ions onto the substrate surface can be estimated as discussed in Section
3.4.2 of Chapter 3. The a-C:H films presented in this chapter were usually deposited with
250 W of net input microwave power. Thus, the averages of the electron temperatures
and plasma densities at 250 W of input microwave power are used for the calculation of
the plasma sheath thickness and the potential at the surface of the substrate. These
average values are 8.3 eV for the electron temperature and 2.2 x 1010 cm'3 for the plasma
density. The floating potential ((15,. in equation (3-5) in Section 3.4.2) is calculated to be
41 V. The potential difference (gidwrin Section 3.4.2) between the floating potential at the
substrate location and the chamber walls was measured to be 2-10 eV and is chosen to be

6 V as the representative value. Then the plasma sheath thickness and the acetylene ion

67

 

energy at the surface of the substrate can be calculated by using equations (3-6) and (3-7).
The ion energy can be influenced by the collisions in the plasma sheath thickness. But the
collisions are not considered significant in the discharge pressure regime of 0.2 mTorr.
The mean free path for argon is estimated with the equation of 7t=1/ 165p cm where p is in
Torr [51], giving the mean free path 1»: 30 cm at 0.2 mTorr. The mean free path is much
larger than the plasma sheath thickness. If the mean free path of acetylene is assumed to
be the same order as the argon case, the collisions of an acetylene ion in the sheath
thickness can be ignored and the ion energy is determined by only the potential difference
between plasma and the surface of a substrate. The results of ion energy calculations are
shown in Table 4—3 for several different rf induced substrate biases (¢,f). The sheath

thickness of plasma is larger than the thickness of the substrate (0.17 mm) in theses cases.

Table 4 - 3: The sheath thickness and acetylene ion energy at variations of rf induced
substrate bias.

 

 

 

 

 

 

 

 

RF Induced Sheath Thickness (mm) Acetylene Ion Energy (eV)
Substrate Bias,
«in (-V) by Matrix by Child Law by Matrix by Child Law
Sheath Sheath Sheath Sheath
100 1.09 1.24 107 121
150 1.26 1.55 148 169
200 1.41 1.84 192 217
250 1.55 2.12 236 263
300 1.68 2.39 246 315

 

 

 

 

 

68

 

The values of ion energy are similar to those of the rf induced substrate bias. The ion
energy onto the surface of the substrate, for example, is approximately 200 eV at the
nominal rf induced substrate bias of -200 V.

Fig. 4-1 and Fig. 4-2 show the PPA data of acetylene gas with the discharge on and
discharge off. The discharge off data in Fig. 4-1 shows standard cracking patterns of
acetylene gas by electron impact ionization in the PPA unit. The acetylene gas in the
chamber exit flow decreases drastically with the plasma ignition in the deposition
chamber as seen in Fig. 4-2. The ratio of acetylene partial pressure (mass226 amu) with
discharge on to acetylene partial pressure with discharge off is 0.09. This fact suggests
that the species containing carbon atoms in the discharge on case are activated by
excitation, dissociation or ionization and come to have a high sticking coefficient, and
then most of them are adsorbed on the surface of the substrate or the walls of discharge
chamber. Another observation is that the hydrogen molecule concentration increases
significantly with the ignition of plasma. This is believed to occur due to the
fragmentation of acetylene gas. In the discharge case, the chamber pressure is largely
supported by hydrogen gas, and the partial pressure of hydrocarbon species is not high
compared to the other species. Thus the neutral carbon flux to the substrate will not be
high compared with fluxes of other species. The water vapor increased in the case of
discharge on and seems to be originating fiom the water molecules that had been
adsorbed on the chamber walls and have escaped away from the walls by ignition of the

discharge.

69

10

 

Acetylene (No Discharge)

 

 

 

Csz
5

A 1

but

but

o
[‘1

:3.
V

0

S

3 H o
94 (H2 2 [ILl
'5
'5

LI
3:
0 01

 

 

 

15 20 25 30 35 40 45

Nhs s (amu)

Fig. 4 - 1: Partial pressure analysis for acetylene gas with the system off

(no discharge).

70

 

Partial Pressure (”Torr)

 

 

 

 

 

 

10
Acetylene
1 H2
H20 Csz
0.1
0.01
0.001 1 l J I l

 

 

 

 

0 5 10 15 20 25 30 35 40 45

Mass (amu)

Fig. 4 - 2: Partial pressure analysis for acetylene gas with the system
discharge on.

71

4.3. The Effect of Ion Energy (RF Induced Substrate Bias) on Film Properties

The impacting ion energy per carbon atom onto the surface of growing films and
high values of ion flux as compared to neutral flux to the grong films are crucial
factors for the deposition of tetrahedral (hydrogenated) amorphous carbon film as
discussed in Section 2.3. In this section, the effect of ion energy on the properties of the
films will be investigated.

Fig. 4-3 shows the optical bandgap (E04 and EM) of the deposited films versus
variation of rf induced substrate bias. The optical bandgap is high at low magnitude of
induced substrate bias, and also has extreme values of 1.77 eV for E04 and 1.34 eV for

Emuc at -200 V of If induced substrate bias (¢,f in Table 4-3). The bars at the peak values

shows the range of the optical bandgaps obtained from repeated experiments. The
acetylene gas plasma has a relatively simple ionization and fragmentation pattern at low
pressure and the major ion species is the Csz+ ion in the plasma [3], giving a uniform
ion bombardment energy at a fixed rf induced substrate bias. The plasma sheath thickness
and the ion energy for several rf induced substrate bias as were shown earlier in Table 4-
3. The -200 V of rf induced substrate bias corresponds to approximately 200 eV, 192 eV
by matrix sheath theory and 217 eV by Child Law sheath theory. Thus the optimum
energy of the ion flux to the substrate is about 200 eV ion energies and the energy per
carbon atom is about 100 eV as other literature has cited [3, 40]. These other research
results were obtained using a filtered ion beam deposition method [16] for ta-C and a

plasma beam source with a tungsten ion extraction grid [3]. In our experiments, the peak

72

2.5

 

 

 

 

2.0 -
I I

E - L
a.
‘31: I I I I
E 1.5 ~
a
CD
E
a [ Etauc
O

1.0 -

0.5 1 l 1 I 1 I 1

0 100 200 300 400

RF Induced Substrate Bias (-V)

Fig. 4 -3: Optical bandgap (Em. and E04) versus rf induced substrate bias for the
films deposited from acetylene gas feed at 0.2 mTorr discharge pressure.
The bars at the peaks show the range of measurement values of the
optical bandgaps for the films deposited repeatedly with the same
deposition condition.

73

 

 

 

value of the optical bandgap at 100 eV per carbon atom is obtained with a deposition
system that does not have any structure of ion filtering or ion extraction grid.

The films deposited at near -200 V of rf induced substrate bias are considered to
be tetrahedral hydrogenated amorphous carbon films. Even if the film structure properties
like mass density or sp3 bonding ratio to sp2 bonding has not been characterized in this
investigation, the occurrence of a peak of optical bandgap at -200 V substrate bias
strongly suggest the films are ta-C:H when compared with other results published [3, 16,
40]. The optical bandgap of the films is dependent on the density of sp3 sites and the
distortion of sp2 rings or chains [37, 38]. The peak of optical bandgap at -200 V occurs
owing to high formation of sp3 sites at the corresponding carbon ion bombardment energy
and matches very well with the model by Robertson [40] and the results of others [3, 16].
The low values of optical bandgaps at ion energies just below the induced substrate
biases of -100 V are attributed to low carbon ion energy onto the growing films. The low
energy of these carbon ions have a lower penetration probability in the growing film that
results in a smaller increase of the local mass density and a low sp3 fraction. In the region
of ion energy exceeding the optimum rf induced substrate bias of -200 V, the excess ion
energy induces relaxation of the diamond-like film structure according to the
subplantation model of Robertson discussed in Section 3 of Chapter 2. The high values of
optical bandgap at lower magnitude of the rf induced substrate biases are considered to be
related with the high hydrogen content at these biases [66, 67]. The sp3 carbon bond is
characterized by a lower binding energy than sp2 bonds, while unpaired electrons in
dangling orbits of amorphous carbon create states in the energy gap between bonding and

antibonding states. Hydrogen removes these states from the gap by closing dangling

74

 

bonds and reduces the density of graphite states [68]. Low energy of the ions could not
dehydrogenated the films because the ions have insufficient energy to break carbon-
hydrogen bonds liberating hydrogen from the film structure. Thus, the fihn's optical
bandgap is high at the region of low ion energy. The higher hydrogen content can hinder
formation of sp2 sites by the preferential formation of sp3 carbon- hydrogen bonds. Thus
the sp2 carbon - carbon bondings will not be high for films deposited at low carbon ion
energies. The relationship between the hydrogen content and optical bandgap will be
discussed in more detail in the next chapter which deals with the films deposited at
pressures in the millitorr range.

The infrared active hydrogen content of the films was measured for the samples
deposited. The absorptance present in the FTIR spectra of films in the region of 2700 to
3100 cm’1 were measured as discussed in the previous chapter. The film’s substrates used
for the measurement of hydrogen content were thick pieces of glass (~ 1 mm) to remove
a coherent interference effect between the reflected IR beams and the transmitting IR
beams that occur for thin substrates. The result is shown in Fig. 4-4. The unit of hydrogen
content is not presented using atomic percentages (at.%) because the density of the films
has not been measured. Thus the hydrogen content with the unit of cm”3 does not
necessarily present the film’s percent hydrogen composition. The hydrogen content
shows a slight decrease with increasing magnitude of the rf induced substrate bias. The
trend is expected because the higher implanting ion energy will expel more hydrogen
atoms fiom the films resulting in less hydrogen content. The film thickness is very thin

(~60 nm) so slight surface contamination of the substrates can affect the hydrogen

75

2.0E+22

 

 

 

 

1.8E+22 -
'a
e.
*5 1.6E+22 r
a
a
O
o
:1
it 1.4E+22 -
8 .
'6
E
o . . .
1.2E+22 -
O
1.0E+22 ‘ ‘ ' l 1 1 L
0 100 200 300 400

RF induced substrate bias (-V)

Fig. 4 - 4: Hydrogen content versus rf induced substrate bias for films from
acetylene gas feed at 0.2 mTorr discharge pressure.

76

 

2.4

Index of refraction

 

 

 

1.6 l I .» 1 . 1 m
0 100 200 300 400

RF induced substrate bias (-V)

Fig. 4 - 5: Index of refraction at 633 nm versus rf induced substrate bias for
films from acetylene gas feed at 0.2 mTorr discharge pressure.
The bars present the standard deviation of the measurements for
the films.

77

content measured for the films, which will give partial explanation of the fluctuation of
the hydrogen content shown in the figure.

Fig. 4~—5 shows the index of refraction of the films versus variation of rf induced
substrate bias. The values do not show significant variation and are all about 2.2. The
fluctuation of the values is almost within the standard deviation of the measurements
which is about 0.05. The trend shows a slight decrease with increasing magnitude of the
rf induced substrate bias and is similar to that of hydrogen content and not to the that of
optical bandgap. Therefore the index of refraction is not considered as sensitive as optical
bangap to the sp3 fraction of the films.

Fig. 4-6 shows the deposition rate of the films versus variation of rf induced
substrate bias. The deposition time was in the range of 45-60 seconds and was measured
within i 5 seconds. The deposition rate is not significantly varying with the variation of
if induced substrate bias. The fluctuation of the deposition rates will be partially
explained by the measurement error of the deposition time. The average of the deposition
rates is 90 nm/min. The deposition rate at 7.0 sccm of acetylene flow rate in the nominal
deposition condition is much higher than the other filtered ion beam and plasma beam
deposition systems used for tetrahedral (hydrogenated) amorphous carbon film

depositions as in Table 2-1.

4.4. The Effect of Ion Flux to Neutral Flux Ratio

In section 4.3, the effect of the ion energy on the film properties is shown and

discussed. In this section, the effect of ion flux on the film properties is investigated. To

78

 

 

 

 

120
O
100 ~ . .
O

A 9 o .
a
E 80 —
E
5
0 O
‘5 60 —
m
=1
.2
a 40 ~
0
c:

20 —

0 I 1 L l 1 L

0 100 200 300 400

RF induced Substrate Bias (-V)

Fig. 4 — 6: Deposition rate versus rf induced substrate bias for the films
deposited from acetylene discharges.

79

see the effect of ion flux onto the surface of growing films in this investigation, several
kinds of experiments were performed including variation of pressure, absorbed
microwave power, and vertical position of the substrate. The variation of the deposition
conditions is expected to change the ion flux to neutral flux ratio. The ratio of ion flux to
neutral flux of species that contains carbon atoms onto the substrate is estimated for the
case of this ECR deposition process.

The fluxes of neutrals and ions are estimated first for the nominal deposition
condition to compare the ion flux to neutral flux ratios of the discharges for the various
off-nominal deposition conditions. The nominal deposition condition was previously
defined with a deposition pressure of 0.2 mTorr, an absorbed microwave power of 250
W, a 7 seem flow rate of acetylene gas and a substrate position of 3.5 cm. The radical and
neutral flux onto the surface of the substrate is estimated from the pressure and gas
temperature of a discharge. The ions and neutrals are all assumed to be those of acetylene
in the estimation. The gas temperature of the discharges is assumed to be 600 K. The
neutral flux is obtained using equation (3-8), Fn=(1/4)ng<v>. The equation, ng=p/kTg
gives a neutral flux of 3.22x1018 m'3 at 0.2 mTorr and 600 K. The mean velocity of
neutrals is 697 m/s at 600 K for acetylene molecules using the equation,
<v>=(8kTg/mn)I/2. Thus, the neutral flux is 5.61 x 1016 cm'zs'l at 0.2 mTorr and 600 K
for acetylene molecules. The ion flux of species that contains carbon atoms onto the
substrate is estimated by measuring the ion current density to the conducting substrate
holder with varying DC bias. The substrate holder area is 54 cm2 inside the acetylene gas
discharge. The current density is shown in Fig. 4-7. The current density is near linear

versus the variation of DC bias and is about 0.9 mA/cmz. This current density

80

 

 

 

1.0

 

 

 

 

. O
O
O
O

0.9 - o ’
«A O

a

C)
2 ’ ‘
E

O

a... .

5 c
c:
*5 .
0
i:
5

0.7 -

0.6 I I I I

50 100 150 200 250 300

DC Bias on Substrate Holder (-V)

Fig. 4 - 7: Current density on the substrate versus dc bias on the substrate holder

for acetylene discharge at pressure of 0.2 mTorr and substrate posion
of 3.5 cm.

81

 

corresponds to 6.0x1015 cm'Z-s'l of ion flux onto the substrate at -200V DC substrate bias.
In this estimation of neutral flux and ion flux, the neutrals and ions of hydrogen
molecules are all regarded as those of acetylene. Then, the rough estimation of the ratio
of ion flux to neutral flux onto the substrate holder is about 10 %, which is smaller than
the value of the filtered ion beam (~ 100%) [16] and plasma beam source with tungsten
ion extraction grid (95%) [3]. This fact shows the tetrahedral hydrogenated amorphous
carbon films can be deposited with a 0.1 ratio of ion flux to neutral flux onto a substrate

in a ECR plasma reactor that has a rf biased substrate holder operating at low pressure.

4.4.1 The Effect of Pressure

The molecular flux onto the surface of a container is proportional to the pressure
at a certain temperature so the discharge pressure of a deposition system determines the
fluxes of neutral species to the surface of the substrate in the deposition chamber. The
plasma density also increases as the discharge pressure increases but does not as quickly
as the density of neutral species. The effect of pressure variation in the deposition
process, therefore, can indicate the effect of the variation of ion flux to neutral flux ratio
onto the surface of the substrate. To investigate the effect of pressure variation, films
were deposited on pieces of thick glass with the nominal deposition condition cited above
in the introduction part of this chapter except for the pressure variation. The pressure was
varied from 0.22 mTorr to 0.6 mTorr by adjusting the gate conductance valve of the
deposition system. The pressure effects on the film properties are shown in Fig. 4-8 and

Fig. 4-9. The current density onto the surface of the substrate was measured with —200 V

82

 

 

 

2.5

 

 

 

 

 

2.0 -
E . E...
G:
a
N)
g 1.5 b
33 f a
.g b Etauc
‘5.
O
1.0 -
0.5 I l I
0 0.2 0.4 0.6 0.8

Pressure (mTorr)

Fig. 4 - 8: Optical bandgap (Etauc and E04) versus pressure for films deposited
with -200 V of rf induced substrate bias from acetylene gas feed.

83

 

 

 

 

 

 

 

2.4
2.2 -
a \\
.2 "
‘5
E
s 20 -
m 0
94-:
¢
N
0 ..
'5
.5
1.8 r
1.6 L I I
0 0.2 0.4 0.6 0.8

Pressure (mTorr)

Fig. 4 - 9: Index of refraction versus pressure for films deposited with -200 V of rf
induced substrate bias from acetylene gas feed.

84

 

1.2

 

 

 

 

1.1 e
A O
E
3
<1
.5.
g) 10 L] .
m
5
a .. ’
E
SE
= 0.9 - O
U

0
0.8 ‘ i ‘
0 0.1 0.2 0.3 0.4 0.5

Pressure (mTorr)

0.6

Fig. 4 - 10: Current density to the substrate versus pressure for the acetylene

discharge.

85

 

 

of DC bias on the substrate holder and the same other deposition conditions, and it is
shown on Fig. 4-10. The optical bandgap has a slightly increasing trend and the index of
refraction shows a slightly decreasing trend versus increasing deposition pressure.

To interpret these results it is useful to estimate the ion flux to neutral flux ratio at various
pressures. The pressure variation increases the neutral flux up to 2.7 times from the value
at the nominal deposition condition because the neutral flux is directly proportional to
pressure. The ion flux onto the surface of the substrate can be estimated from the current
density of Fig. 4-10. The current density increased 1.3 times as it varied fiom 0.9
mA/cm2 to 1.2 mA/cm2 across the range of variation of the pressure, hence so did the ion
flux. Thus the ion flux did not increase as much as the neutral flux in the variation of the
pressure. Therefore, the nominal ion flux to neutral flux ratio decreased by 50 %
(=1.3/2.7) as the pressure increased in the range. The increase in the pressure can also be
expected to alter the composition of the species fluxes to the deposition surface. Based on
the PPA results presented earlier in Section 4-2, the dominant neutral flux is expected. to
be hydrogen. With a fixed flow rate of acetylene gas, the hydrocarbon species are
activated in the discharge and subsequently stick on the surface of the substrate or the
chamber walls, that is, are consumed in the chamber, thus do not contribute significantly
to the increase of total pressure in the deposition chamber. But the hydrogen gas is not
consumed in the chamber and the partial pressure of hydrogen gas is higher than the
hydrocarbon gas and hence hydrogen forms the large part of the total pressure of the
deposition chamber. Therefore, the ratio of the ion flux to the neutral flux of hydrocarbon
species does not change significantly with the variation of the deposition pressure with

the fixed feeding of acetylene gas. Thus, the variation of deposition pressure changes the

86

 

 

hydrogen flux primarily and therefore the pressure variation shows only a small affect on
the film properties across the 0.2-0.6 mTorr range. The slight increasing trend of optical
bandgap and decreasing trend of index of refraction are considered to be due to the effect
of increasing hydrogen content of the films with the increasing deposition pressure. In the
other literature, the pressure showed a big influence on the properties of the ta—C:H films
prepared by rf plasma deposition by increasing the plasma density using a magnetic field
[69]. It is because the ion flux to neutral flux ratio and the ion energy are affected

significantly with variation of the deposition pressure in rf discharges.

4.4.2. The Effect of Microwave Power

The absorbed microwave power also increases the plasma density of the discharge
according to the global model of plasma discharges [51], thus, it is expected that
microwave power changes will affect the ratio of ion flux to neutral flux onto the
substrate. The absorbed power was varied from 170 W to 410 W in the experiments to
investigate the effect of the absorbed microwave power on the properties of films. The
films were deposited on the pieces of 0.17mm thick glass attached to the substrate holder
using a heat sink compound in the nominal deposition condition. The properties of optical
bandgap and index of refraction did not show significant variation with the changing
absorbed microwave power as seen in Fig. 4-11 and Fig. 4-12. The current density with
the microwave power variation is presented in Fig. 4 - 13. The current density is 0.9

mA/cm2 at the nominal deposition condition (250W), 0.8 mA/cm2 at 170 W and 1.12

87

 

 

2.5

 

 

 

 

 

2.0 -
E04

5‘
3
2'
%D 1'5 . Etauc
3 f' I *-
m
'5:
t:
G.
O 1.0 .

0.5 ' L

100 200 300 400 500

Absorbed Microwave Power (W)

Fig. 4 - 11: Optical bandgap (Eu...c and E04) versus absorbed microwave power
for films deposited with 200 V of rf induced substrate bias from
acetylene gas feed at 0.2 mTorr discharge pressure.

88

 

 

2.4

 

 

 

 

O
2.2 - s
O
O

a
.2 -
E
E 2.0 -
Gnu-1
O
:3 -
'6
E.

1.8 -

1.6 1 I 1 I 1 I 1

100 200 300 400 500

Absorbed Microwave Power (W)

Fig. 4 - 12: Index of refraction versus absorbed microwave power for films
deposited with 200 V of rf induced substrate bias from acetylene gas
feed at 0.2 mTorr discharge pressure.

89

 

1.2

 

 

 

 

o
1.1 -
"E 1.0 -
s . '
E, r
E 0.9 .-
5
c: .
‘5
t 0.8 - ’9 ’
5
U
0.7 L-
0.6 1 I 1 I 1 I 1
100 200 300 400 500

Absorbed Microwave Power (W)

Fig. 4 - 13: Current density to the substrate holder versus absorbed microwave
power for an acetylene discharge at 0.2 mTorr pressure.

90

 

mA/cm2 at 410 W. Thus, the ion flux to neutral flux ratio decreased 11 % at 170 W and
increased 24 % at 410 W from the value at the nominal deposition condition. The optical
properties of the films did not show significant variation within the variation of the ion
flux to neutral flux ratio onto the surface of the substrate for —11 % and 24 % variation
about the nominal deposition condition. In rf deposition of ta-C:H films by increasing the
plasma density using a magnetic field, the increased rf power increased the optical

bandgap and the hardness [69].

4.4.3. The Effect of Substrate Position

The plasma density decreases exponentially in an argon discharge with increasing
distance from the center of the discharge region for the discharge reactor of this
investigation [70]. With the motivation from this fact, the a-C:H films were deposited at
two different positions of the substrate. The substrate positions were 3.5 cm or 6.0 cm
below the base-plate of the discharge chamber. One group of films was deposited at 3.5
cm and another group at 6.0 cm with a variation of rf induced substrate bias and with the
nominal deposition values for other input variables. The results of the optical properties
of the films are presented in Fig. 4-14 and Fig. 4-15. The current density was again
measured at the two different positions versus variation of the rf induced substrate bias to
estimate the ion flux to the surface of the substrate, and it is shown in Fig. 4-16. The
optical bandgap shows very different trends for the two different substrate positions. The

curve for 3.5 cm substrate position has a peak value of the optical bandgap at -200 V of rf

91

 

 

 

2.5

2.0 -

E04, S.p.=6.0 cm

 

Optical Bandgap (eV)
)—1
in

Eta“, s.p.=3.5 c H
1.0 ’

 

 

 

0.5 1 I 1 I 1 I 1 I 1
50 100 150 200 250 300

RF Induced Substrate Bias (-V)

Fig. 4 - 14: Optical bandgap (Etauc and E04) versus rf induced substrate bias for
films deposited from acetylene gas feed at 0.2 mTorr discharge
pressure with substrate positions (s.p.) of 3.5 cm and 6.0 cm. The
bars at the peaks show the range of measurement values of the
optical bandgaps for the films deposited repeatedly with the same
deposition condition.

92

 

 

 

 

2.4

 

 

 

. O I I I I
2.2 - o O
1:
.2 '
13 O :s.p.=6.0 cm
“3 20
m ' I :s.p.=3.5 cm
“-1
e
we
0 _
1:
.5
1.8 *-
1.6 L L 1 I 1 I 1 I 1
50 100 150 200 250 300

RF Induced Substrate Bias (-V)

Fig. 4 - 15: Index of refraction versus rf' induced substrate bias for films deposited
from acetylene gas feed at 0.2 mTorr discharge pressure with substrate
positions (s.p) of 3.5 cm and 6.0 cm.

93

 

1.0

 

 

 

 

O
O . .
09 - 9 O
O
. Scpo=305 cm
NA 0.8 " .
E O
U
2
5
g» 0.7 -
E’
Q
G
E
i: 0.6 " I I I
5 I I
U I
I I I s.p.=6.0 cm
I
0.5 -
0.4 ‘ l 1 1
50 100 150 200 250 300

DC Bias on Substrate Holder (-V)

Fig. 4 - 16: Current density on the substrate versus dc bias on the substrate
holder for acetylene discharge at 0.2 mTorr pressure at substrate
positions (s.p) of 3.5 cm and 6.0 cm.

94

 

 

 

induced substrate bias. The bars at the peak of optical bandgaps show the range of the
optical bandgaps obtained from the repeated experiments. This observation was already
shown earlier in Fig. 4-3 and it repeats here again The variation of optical bandgap near
the peak value is considered. to be due to the variation of sp3 fraction in the film
composition and the film deposited at —200 V rf induced substrate bias is considered to
be ta—C:H fihns as discussed in Section 4.3. On the other hand, the curve with 6.0 cm
substrate position is almost flat not showing the peak value. The index of refraction at 3.5
cm substrate position is slightly higher than the value at 6.0 cm substrate position. The
current density at 3.5 cm substrate position (0.9 mA/cmz) is significantly higher than that
at 6.0 cm substrate position (0.57 mA/cmz). Thus, the ion flux to neutral flux ratio at 6.0
cm substrate position decreased about 40 % from the value with the nominal deposition
condition. The facts that the films of 6.0 cm substrate position have relatively high and
flat values of optical bandgap and the ion flux is lower than the nominal case suggest that
the films are hydrogenated amorphous carbon films with lower carbon-carbon sp3
fiaction. The low value of ion flux onto the growing films at 6.0 cm substrate position
could not expel the hydrogen atoms in the films enough to lower the Optical bandgaps as
discussed in the effect of rf induced substrate bias in Section 4-3.

Fig. 4-17 shows the optical bandgap, Emuc, as a function of the ion flux to neutral
flux ratio. ‘I/N’ stands for the ion flux to neutral flux ratio. The data for I/N=0.95 is from
the literature [3] and the data for I/N=0.l and 0.06 is from Fig. 4-14 of this investigation.

The Optical bandgap shows a high peak value for the fihns deposited with I/N=0.95 and a

95

 

 

 

 

 

 

I/N=0.95 [3]
2.0 —
9
e
g; f/N=0.l
é” 1.5
g ‘
.§ mm
*5.
o
1.0 ,
0.5 m
50 100 15) 200 250 300
1011 m “TQH; (9V)

Fig. 4 - 17: Optical bandgap as a function of ion flux to neutral flux ratio

96

 

 

low peak value with I/N=0.1 but does not show any peak value with I/N=0.06. Therefore
the threshold ratio of ion flux to neutral flux for depositionof ta-C:H films is found to be

in the range of 0.06-0.1 in this investigation.

4.5. The Effect of Deposition Temperature

The deposition temperature is considered an important factor in the deposition of
tetrahedral amorphous (hydrogenated) carbon films as discussed in Section 2.4 [19]. A
high deposition temperature induces the relaxation of bonding structure fiom diamo nd-
like carbon films to graphite-like carbon films. In this study a comparison was done
between depositions at two different temperatures. The temperature difference was
achieved by using a thermally conducting paste for attaching the substrate to the substrate
holder, while other substrates were not thermally attached. The thermally conducting
paste (heat sink compound) makes a good thermal contact between the substrate and the
substrate holder so that heat arriving at the substrate can be quickly transferred to the
substrate holder keeping the temperature of the substrate near the temperature of the
substrate holder. The film with heat sink compound is the deposited at a lower deposition
temperature than the film without heat sink compound. The result is presented in Table 4-
4.

Table 4 - 4: The effect of temperature effect on optical bandgaps (Emu, and E04) and index
of refraction (n).

 

 

 

 

Sample Emuc (eV) E04 (eV) 11
Film with heat sink compound 1.45 1.80 2.27
Film without heat sink compound 1.19 1.59 2.27

 

 

 

 

 

97

 

 

 

The sample with the heat sink compound gave the higher value of optical bandgap
of 1.44 eV than the other's 1.19 eV of optical bandgap. The lower optical bandgap of the
sample deposited at the higher temperature (no thermally conducting paste) is likely due
to higher sp2 content in the film (i.e., more thermal relaxation of the bonding structure has
occurred in the film without heat sink compound). The index of refraction was the same
for the two samples at 2.27. Thus, the effect of deposition temperature can be
demonstrated by attaching the substrate to the substrate holder using the heat sink
compound. Because of this temperature influence, most experiments in this chapter were
performed with the substrates thermally attached to the substrate holder by using the heat

sink compound.

4.6 The Effect of Addition of Helium Gas

In the deposition process most of the hydrocarbon species are activated and
subsequently stick on the surface of the substrate or the walls of deposition chamber,
thus, they do not contribute significantly on the total pressure of the chamber as discussed
based on the PPA data in Section 4.2. The hydrogen gas has the biggest partial pressure
in the chamber of the acetylene discharge. In this section the effect of partial replacement
of the hydrogen gas by helium gas in the deposition chamber is investigated by adding
helium gas to the acetylene feed gas going into the chamber. The deposition condition
was the same as the nominal deposition condition except for the addition of helium gas
into the feed gas. The flow rate of helium gas was varied fiom 0.5 seem to 2.5 sccm The

films were deposited on the thin substrates that were just put on the substrate holder

98

 

 

 

without using the heat sink compound. The optical properties of the films are presented in
Fig. 4-18 and Fig. 4-19. The films show almost no or minimal influence of the variation
of helium flow rates on the optical bandgap and index of refraction. The index of
refraction does not show any significant variation with the variation of flow rate of
helium and has similar values to those of other films such as films in Table 4-4. However,
the optical bandgaps (1.25< Em, (eV) <1.45) of the films prepared with the addition of
helium are a little higher than the values (1.19 eV for Emuc as listed in Section 4.5) of
films deposited without either the heat sink compound and helium. One group of samples
done with helium addition was deposited on the substrates attached to the substrate
holder using heat sink compound. Unfortunately, the films were completely peeled off in
the process of ultrasonic cleaning in methanol for removal of heat sink compound before
optical characterization was performed. The peeling-off did not occur in fihns with the
nominal deposition condition without helium. From the above facts, the films deposited
with addition of helium gas seem to have different film structure and possible more
internal stress (hence the peeling) compared to the films deposited with the standard

deposition condition.

4.7. The Deposition Rate as a Function of the Acetylene Flow Rate

The flow rate of acetylene gas into the deposition chamber was varied and the
films were deposited with other external deposition conditions fixed. The deposition rate
almost linearly increases with the increasing flow rate as shown in Fig. 4-20. This result
suggests that the flow rate of acetylene source gas is the rate limiting process to form and

grow the films. The neutrals, radicals and ions of acetylene gas are considered to have a

99

 

 

 

2.5

 

 

 

2.0
9
8 9 . E04 .
g- O O
%o 9
g 1 5 ~ I
_ I
.§ - I Etauc I #
E I
O

1.0 ~

0.5 . 1 r 1

0.0 0.5 1.0 1.5 2.0 2.5

Flow Rate of Helium (sccm)

Fig. 4 - 18: Optical bandgap (Eunc and E04) versus flow rate of helium for films
deposited from acetylene and helium gas feed with 200 V of rf
induced substrate bias.

100

 

 

 

2.4

 

 

 

t o
O
O O

2.2 -
a
.2 i'
*5
N
“E
m 2.0 "
M
Q
N
9 ..
'5
El

1.8 '

1.6 I J I I

0.0 0.5 1.0 1.5 2.0 2.5

Flow Rate of Helium (sccm)

Fig. 4 - 19: Index of refraction versus flow rate of helium for films deposited from
acetylene and helium gas feed with 200 V of rf' induced substrate bias.

101

 

 

 

 

350

300
A 250
.5.

E

E, 200

d)

‘5

In

5 150

s

D.

5 100
50
0

 

 

1 I 1 I 1 I 1

 

 

0 10 20 30 40

Flow Rate of Acetylene (sccm)

Fig. 4 - 20: The deposition rate of a-C:H films versus the flowrate of acetylene

gas into the discharge. The pressure of the discharges varied from

0.2 mTorr to 0.45 mTorr as the acetylene flow rates increased from 4
spam to 1‘ «mm

102

 

 

high sticking coefficient in the deposition process from this result. The PPA data also
support that they have high sticking coefficient, which was shown in earlier discussion in
Section 4.2. The deposition rate (~90 nm/min) at 7.0 sccm of acetylene flow rate in the
nominal deposition condition is much higher than the other filtered ion beam and plasma
beam deposition systems used for tetrahedral (hydrogenated) amorphous carbon film

depositions as discussed in Section 4.3.

4.8 Summary

Hydrogenated amorphous carbon films were deposited at pressures in the
submillitorr range and room temperature with the variation of rf induced substrate bias
fiom 80 to -300 V using acetylene source gas. The flow rate of acetylene gas was 7 seem
and the net absorbed input microwave power was about 250 W.

The optical bandgap of the films has a peak value (1.34-1.44 eV for Em, and
1.77-1.83 eV for E04) at -200 V of rf induced substrate bias, which corresponds to 100 eV
of ion energy per carbon atom. The occurrence of a peak value of optical bandgap at -200
V of rf induced substrate bias is in good agreement with the results in the literature and
matches well with the subplantation model of ta-C:H films described in Section 2-3.
Thus, the films deposited at near -200 V of rf induced substrate bias are considered to
have the maximum ratio of carbon sp3 bonding to sp2 bonding. Further, the index of
refraction does not vary much and the hydrogen content is almost uniform over the
experimental range of rf induced substrate bias.

Maintaining a low deposition temperature was critical to obtain the peak value of

Optical bandgap at -200 V of rf induced substrate bias. The optical bandgap was lowered

103

 

 

to 1.19 eV for Emu, at a higher deposition temperature fiom 1.44 eV for Emuc at a lower
deposition temperature. The effect of ion flux to neutral flux ratio on the properties of
films was not present as the films were deposited with the variation of absorbed
microwave input power. But the effect appeared as the films were deposited with a
variation of substrate position and move weakly with variation of pressure. Thus, the
optical bandgap had the peak at ~200 V of rf induced substrate bias for the films
deposited at 3.5 cm substrate position, but did not show the peak for the fihns deposited
at 6.0 cm substrate position. The ion flux to neutral flux ratio is smaller by 40 % at 6.0
cm substrate position as compared at 3.5 cm substrate position. Thus, the effect of ion
flux to neutral flux ratio on the properties of a-C:H films is demonstrated as the
subplantation model predicts. The estimated ion flux to neutral flux ratio at 3.5 cm
deposition position is approximately 10 %, which is much smaller than the usual ta-C:H
film deposition systems of plasma beam systems and filtered ion beam systems in
literature. Hence, the observation of a peak in the optical bandgap can occur at ion flux to
neutral flux ratio as low as 10 % and the threshold ratio of ion flux to neutral flux for the
deposition of ta-C:H films is found to be in the range of 0.06-0.1 at —200 Vrf induced
substrate bias.

The deposition rate increased almost linearly with the increasing acetylene flow
rate, which suggests the carbon species in the discharge have a very high sticking
coefficient in the deposition process. This fact is supported by the data of partial pressure
analysis of the exit gas fiom the deposition chamber, which shows there are few carbon
species in the exit gas. Thus, the flow rate of acetylene gas acts as rate-limiting process of

the film deposition. The deposition rate (~90 run/min) at 7.0 sccm of acetylene flow rate

104

 

 

in the nominal deposition condition is much higher than the other filtered ion beam and
plasma beam deposition systems used for tetrahedral (hydrogenated) amorphous carbon

film depositions.

105

 

 

Chapter 5

5. Films Deposited from Acetylene-Argon and Methane-Argon

Discharges at Pressures in the Millitorr Range

5.1. Introduction

In this chapter, a-C:H films are deposited and characterized at higher pressure
conditions than those of the previous chapter. The objectives in Chapter 5 are to establish
the variation of film properties possible by depositing the fihns using different
hydrocarbon/argon feed gases at different deposition conditions and to understand the
deposition process of the films by investigating the effects of rf induced substrate bias,
pressure and argon flow ratio. The deposition pressure is increased up to the millitorr
range from the submillitorr range, and the typical pressure is 3mTorr. Many plasma-
assisted CVD investigations for a-C:H deposition have been run at pressures in the
millitorr range as discussed in Section 2.4. At this range of pressures, the films are
sometimes deposited with the addition of inert gases or hydrogen gas [20, 25-28, 30, 71-
73]. Work by Mutsukura and coworkers have studied the influence of noble gases He,
Ne, Ar, Kr and Xe on methane plasmas used for the deposition of a-C:H films [74-76].
Their work done using a rf plasma deposition system showed that He, Ne, Ar and Kr can
all work to enhance the hydrocarbon ion flux to the deposition surface. The source gases
used in this study for the films were acetylene gas and methane gas, and argon gas was
used as the inert gas. The usual deposition time was 5 minutes for acetylene-argon

discharge and 10 minutes for methane-argon discharge. The incident microwave input

106

 

 

 

power was about 250 W and the reflected power was very small when compared with the
incident power, thus the reflected power was neglected. The rf induced negative substrate
bias was varied from 0 to -60 V for 0.17 mm thick glass substrate and from 0 to -100 V
for 1 mm thick glass substrate. The substrates used were mounted on the metal substrate
holder without thermally conducting paste (heat sink) and placed at 3.5 cm below the
baseplate. Thus, the substrate temperature was higher than room temperature because of
the heating from the discharge. The substrate temperatures were measured to be around
80-100 C after 5 minutes of deposition time at 3mTorr with the method discussed in
Section 3.4.3.

The deposition pressure and the argon flow rate were also varied for the
deposition of several films to see their effects on the properties of the films. This
variation of deposition condition in this chapter produces a variation of the film
properties. Specific properties studied include deposition rate, density, hydrogen content,

index of refiaction and optical bandgap. The ranges of input variables are summarized in

 

 

Table 5 - 1.

Table 5-1: The input variable space

Input Variables Variable Range

RF induced substrate bias (-V) 0 - 60 for 0.1 mm thick glass substrate

0 - 100 for 1 mm thick glass substrate

 

 

 

Microwave power (W) 200 - 300
Pressure (mTorr) 1 - 5
Hydrocarbon gas flow rate (sccm) 7 or 8 (C2H2 or CH4)

 

 

Argon flow to hydrocarbon gas flow ratio 0 - 1 (Ar/Csz or Ar/CH4)

 

 

107

 

 

 

The discharge properties at pressures in the millitorr range are presented before
the properties of the films to use the discharge properties in the discussion of the
properties of the films. The discharge properties include electron temperature, plasma

density measured by double Langmuir probe and partial pressure analysis of the exit gas.

5.2 Discharge Properties at Pressures in the Millitorr Range

For the investigation of discharge properties of the ECR-CVD deposition system
at pressures in the millitorr range, the electron temperature and plasma density were
measured using a double Langmuir probe for an argon discharge at 200 W microwave
input power and 8 seem argon flow rate versus variation of pressure. The probes were
located at the place where the substrates are normally positioned. The results are shown
in Fig. 5-1 and Fig. 5-2. The electron temperature decreases and plasma density increases
with increasing pressure.

The double Langmuir probe was also used to measure a-C:H deposition discharges.
These discharges coated the probe quickly with an insulating film, so before each
measurement the probe was biased negative and sputtered clean of the a-C:H film with an
argon plasma. Once cleaned, the probe could be used to take measurement for about 30
seconds before a new a-C:H insulating coating would disrupt the measurements
significantly, thus requiring the probe be cleaned again. The summary of this data taken
at a pressure of 3mTorr, microwave power of 270 W and argon flow rate of 8 scam is
shown in Table 5-2. The electron temperatures of acetylene-argon discharges are similar,
but are slightly lower than those of the methane-argon discharges. Another observation in
Table 5 -2 is that the ion saturation current shows only a modest change as the gas

composition varies. Specifically, the change in the ion saturation current, at a fixed

108

 

microwave power as the hydrocarbon gas flow rate is adjusted, is less than a 30%
deviation fiom the pure argon plasma value.

Table 5-2: Langmuir probe measurement of argon, methane-argon, and acetylene-argon
discharges. The argon flow rate is constant at 8 seem

 

 

 

 

 

 

CH4 flow rate C2H2 flow rate Saturation ion Electron
(sccm) (sccm) current (mA/cmz) temperature (eV)
0 0 2.33 2.0
8 0 1.67 2.8
4 0 2.11 2.7
0 8 1.96 2.3
0 4 2.96 2.6

 

 

 

 

 

 

An important variable that has a crucial influence on the film properties is the ion
impacting energy onto the surface a growing film as discussed in Section 2-2. The ion
energy is determined by the potential given by |¢,/1+¢t.+ my in the discussion in Section
3.4.2. The potential m, as determined from equation (3-5) and T e values of 1.8-3.0 eV
from Table 5-2 and Fig. 5-1, ranges from 6 to 14 V. ¢dwrmeasured varies from 2-10 V. So
the potential difference between the plasma and the substrate holder is |¢41+(8 to 24) V.
and can be written by |¢,f |+16i8. The plasma sheath thickness on the substrate holder
can be estimated with ¢,f (=100 V, for example), electron density of (0.6-1.8)x10ll cm'3
fiom Fig. 5-2 and T e of the values given above using equation (3-6) and (3-7) in Section
3.4.2. It has ranges of 0.33 mm-0.61 mm with the matrix sheath theory and 0.45 mm-0.99
mm with Child Law sheath theory. When the thickness of the substrate (0.17 mm) is

considered, the ion energy on the surface of a substrate using the above estimation of

109

 

 

plasma sheath thickness on the substrate holder and a given ¢,f =100 V, has ranges of 25-
64 eV with the matrix sheath theory and 54-96 eV with Child Law sheath theory.

Another factor that could influence the ion energy is collisions within the plasma
sheath. The importance of collisions can be assessed by comparing the ion mean fiee path
between collisions with the plasma sheath thickness. The plasma sheath is already
estimated and is less than 1 mm. The mean free path 7).] for an argon ion moving in an
argon background gas of temperature 600 K [77] is approximated as 151/(165p) cm [51]
where p is the pressure in torr. For the pressure range studied in this investigation of 1—5
mTorr the mean fi'ee path ranges from 12-60 mm. Hence the mean fiee path is
significantly longer than the sheath thickness and minimal collision occurs as the ions
transit the plasma sheath. Therefore, the ion energy is well described by knong the
potential across the plasma sheath adjacent to the substrate.

The compositions of the residual gas flow at the pumping port of the chamber for
50 % / 50 % methane-argon mixtures with discharge on and discharge off are shown in
Fig. 5-3 - Fig. 5-4, respectively. Acetylene-argon mixtures with discharge on and
discharge off are shown in Fig. 5-5 and Fig. 5-6, respectively. The mass of various
species is presented in Table 5-3 for convenience in reading the plots of partial pressure
analysis.

Table 5 - 3: Table of mass

 

Mass (amu) l 2 16 26 18 40

 

 

Species H H2 CH4 C2H2 H20 Ar

 

 

 

 

 

 

 

 

110

 

 

3.5

Electron Temperature (eV)
N m
in 'e

5"
O
l

 

 

 

 

4 6

Pressure (mTorr)

10

Fig. 5 - 1: Electron temperature for argon discharges versus pressure in the

ECR-CVD system.

111

 

 

20.0

 

 

 

 

5 15.0 ~
'7:
H
5;
£9 _
E
Q
Q
N
g 10.0 —
O:
5.0
0 2 4 6 8 10

Pressure (mTorr)

Fig. 5 - 2: Plasma density, up, for argon discharges versus pressure in the
ECR-CVD system.

112

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10
Argon-Methane (N 0 Discharge)
CH4

A 1
i:
o
H
3
2

2 0.1
0
I-
9..
E
t
a
a.

0.01

0.001

0 15 20 25 30 35 40 45

Mass (amu)

Fig. 5 - 3: Partial pressure analysis for the methane-argon gas mixture with the
discharge off.

113

 

 

 

 

 

 

 

10
Argon-Methane
l I
E Hz Ar
: CH4
3-
4)
‘5
g 0 1
2
n.
27.2
‘1':
c3
:1.
0.01
I
0.001 ‘ L '

 

 

 

 

0 5 10 15 20 25 30 35 40 45

Mass (amu)

Fig. 5 - 4: Partial pressure analysis for the methane-argon gas mixture with
the discharge on.

114

 

 

10

 

Argon Acetylene (No Discharge)

Csz

 

 

Partial Pressure (“To”)
:5
1—1

 

 

 

0.01 -

 

0.001 J

 

 

 

15 20 25 30 35 40 45

Mass (amu)

Fig. 5 - 5: Partial pressure analysis for acetylene-argon gas mixture with

discharge off.

115

 

 

 

10
Argon—Acetylene

 

Ar

H2

 

 

Partial Pressure (”Torr)
:5
1—1

0.01

 

 

 

 

 

 

0.001 ' '
0 5 10 15 20 25 30 35 40 45

Mass (amu)

Fig. 5 - 6: Partial pressure analysis for acetylene-argon gas mixture with
discharge on.

116

 

 

The discharge conditions for the figures were 3 mTorr pressure, 300 W microwave input
power, 7 seem flow rate of acetylene or methane, 7 seem argon flow rate and zero rf
induced substrate bias voltage. The discharge-off data in Fig. 5-3 and Fig. 5-5 shows
standard cracking patterns for electron impacting ionization. The partial pressure of
carbon-containing species in the exit gas flow with the discharge on is low. The ratio of
methane partial pressure with the discharge on as compared to the discharge off is 0.20
for a methane-argon mixture (7 seem-7 seem) in Fig. 5-4. Similarly, the ratio of acetylene
partial pressure with discharge on to the acetylene partial pressure with discharge off is
0.11 for an acetylene—argon mixture (7 sccm - 7 seem) in Fig. 5-6. Hence, a substantial
portion of the carbon that flows into the plasma discharge is activated by either
excitation, dissociation or ionization and this carbon is deposited on either the substrate
or the chamber walls. The high sticking coefficient of hydrocarbon species in acetylene
discharge was also seen in Section 4-2, already. Also indicated is that the methane-argon
plasma shows a substantial increase in the hydrogen present in the exit gas when the
discharge is on as compared to the discharge off. It should be also noted that the source of
the hydrogen for the discharge-off case in the PPA spectrum of Fig. 5-3 and Fig. 5-5 is

likely dissociation of methane caused by the electron emitter in the PPA unit.

5.3 Film Properties at Pressures in the Millitorr Range

5.3.1 Absorption Coefficients
The optical absorption coefficient versus photon energy of a-CzH films deposited
using a methane-argon plasma are shown in Fig. 5-7 and Fig. 58 Fig. 5-7 shows the

absorption coefficient versus photon energy for four different rf induced substrate biases.

117

 

 

 

 

 

 

 

105 o
ox
00§X
o
X o
o
o
+
.5“ +
'a 10“ +
3
g .
= ,
0
.§ [
a
d)
O
U
s:
.3 RFBias(-V)
EI- 103 : o: 100
8 ' x: 50
.G C O: 25
< t +: 0
19.2“ 1 L 1 I g I L
0.0 1.0 2.0 3.0 4.0

Photon Energy

Fig. 5 - 7: Optical absorption coefficients of films deposited in methane-argon
discharges. Data is plotted versus photon energy at various rf induced
substrate biases.

118

 

Absorption Coefficient (cm '1)

 

10“

 

Ar/CH4 Ratio
0: 0.8

g x: 0.4
" 3 ' O: 0.2

3
10 L '9 4'" + : 0.0

 

 

 

102 1 1 1 1 1 1 1
0.0 1.0 2.0 3.0 4.0

Photon Energy

Fig. 5 - 8: Optical absorption coefficients of films deposited in methane-argon
discharges. Data is plotted versus photon energy at various argon
flow ratios.

119

 

 

 

 

 

 

 

 

105 .
E 10“
ii .
g .
Q .
.§ 1.
a: .
o
c .
U
G ,
.2
H
a.
’5 3
m 10 :
a h
4 I

O
102 l I 1 I 1
0.0 1.0 2.0 3.0

Photon Energy (eV)

Fig. 5 -9: Optical absorption coefficients of films deposited in acetylene-argon
discharges. Data is plotted versus photon energy at various rf induced
substrate biases.

120

 

 

 

The deposition condition for this figure is 200 W input microwave power, 8 seem argon,
8 seem methane, and 3 mTorr pressure. The films were deposited on 1 mm thick glass
substrates. The absorption at a given photon energy increases as the ion energy increases
(i.e., the rf induced substrate bias becomes more negative). Fig. 5-8 shows the absorption
coefficient for four different argon/methane flow rate ratios at the condition of 300 W
microwave power, 8 seem argon, 8 seem methane, 3 mTorr and -50 volts induced
substrate bias. The addition of argon produces films of higher optical absorption. Fig. 5-9
shows the absorption coefficient for films deposited using an acetylene-argon gas
mixture. The deposition conditions include 200 W microwave power, 8 seem acetylene, 8
seem argon, 3 mTorr pressure, and -40, -20 and 0 V rf induced dc substrate bias. The
films were deposited on 1 mm thick glass substrates. The more negative substrate biases

again produced films that are more optically absorbing.

5.3.2 The Effect of RF Induced Substrate Bias

The effects of the rf induced substrate bias voltage on the film property variations
are shown further in Fig. 5-10 to Fig. 5-14 at pressures in the millitorr range. All of the
films in Fig. 5-10 to Fig. 5-14 were deposited at 3 mTorr, 200 W microwave input power,
7 seem flow rate of acetylene or methane gas and 7 seem flow rate of argon gas with a
variation of rf induced dc substrate bias from 0 to -60 V (i.e., increasing ion energy). The
films were deposited on 0.17 mm thick glass substrates. The films were unable to be
deposited below -60 V of rf induced substrate bias in this deposition process because of
delamination. The film’s deposition rate decreases with increasing ion energy for both

acetylene-argon and methane-argon discharges in Fig. 5-10. The films from acetylene-

121

 

 

 

argon discharges have a higher growth rate than from methane argon discharges at the
same rf induced substrate bias. Fihn mass density increases with the increasing ion
energy for depositions from both acetylene-argon and methane—argon discharges in Fig.
5-11. The mass density of films from acetylene-argon discharges is higher than fiom
methane-argon discharges. The hydrogen content of acetylene-based films changes only
slightly with the variation of rf induced substrate bias, and in contrast, the hydrogen
content of methane-based films decreases greatly with ion energy increases in Fig. 5-12.

The refiactive index increases quickly with the increasing ion energy for
acetylene-based fihns and changes more slowly for methane-based films in Fig. 5-13.
The Tauc optical bandgap Emu, and E04 bandgap (energy at absorption coefficient
a=104cm") decreases with the increasing ion energy in Fig. 5-14. The optical bandgaps
of the CH4 films are consistently higher than those of the acetylene-based films.

The deposition rate of both acetylene-based and methane-based films decreases in
Fig. 5-10 as the rf induced substrate bias is increased. This result agrees with the
reference [73] in which the films were deposited using a rf inductively coupled plasma
reactor with a rf biased substrate. The plasma deposition conditions are expected to be
similar in this system as in ECR systems. In general, for ion-assisted deposition two
competitive processes that determine the deposition rate include an increase in the
sticking coefficient with increases in the bias voltage and an increase in the
sputtering/etching with increases in the bias voltage. An increase in the sticking
coefficient would increase the deposition rate [14, 21, 30], which is not observed

experimentally in this investigation. Rather, the decreasing deposition rate of the films

122

 

 

 

100

80

 

ON
6

Deposition Rate (nm/min)
A
c

N
c

 

 

 

 

0 20 40 60
RF Induced Substrate Bias (-V)

Fig. 5 - 10: Deposition rate versus rf induced substrate bias for methane-based
and acetylene-based films.

123

 

 

 

3.0

 

 

 

 

 

 

 

2.5 -
2.0
E 1.5
39
A?
g 1.0
a 1..
0.5 -
0.0 1 I 1 I 1
0 20 40 60

RF Induced Substrate Bias (-V)

Fig. 5 - 11: Mass density versus rf induced substrate bias for methane-based and
acetylene-based films. ‘

124

 

 

70'

 

ON
6

U!
c

Hydrogen Content (at %)
m
c

40

 

 

 

 

 

20 -
10 -
0 1 1
0 20 40 60

RF Induced Substrate Bias (-V)

Fig. 5 - 12: Hydrogen content (at. %) versus rf induced substrate bias for

methane-based and acetvlene-based films.

125

 

 

 

 

 

Index of Refraction

2.4

1.6

Fig. 5 - 13: Index of refraction versus rf induced substrate bias for
methane-based and acetylene-based films.

 

 

 

 

 

L

 

20
RF Induced Substrate Bias (-V)

40

 

 

 

 

3.0

 

 

 

 

 

 

 

S‘

3.

5‘ 2.0

Of)

'6

I

N

m

“a

'2 E C H

*" tauc9 2 2

8' 1.0 -
.

0.0 ' '

0 20 40 60

RF Induced Substrate Bias (-V)

Fig. 5 - 14: Optical bandgap (Euuc and E04) versus rf induced substrate bias
for methane-based and acetylene-based films.

127

 

 

with increasing bias voltage reflects the influence of a sputtering or etching effect on the
films and/or densification of the films during the deposition process. The decreasing
hydrogen content (Fig. 5-12) in the films and the increasing density of films (Fig. 5-11)
with increasing ion energy primarily support the densification explanation. This
densification with the increasing magnitude of negative induced dc bias on the substrate
can be explained by momentum transfer of the bombarding argon, hydrogen, and/or
hydrocarbon ions to the surface. This bombardment reduces the hydrogen in the film and
also lead to more spZ-bonded carbon [68]. It should also be noted that at higher
argon/methane flow ratios (over 75% argon), argon sputter dominates and no film
deposition occurs in the ECR-CVD system studied.

In Fig. 5-10, the acetylene-based films have higher deposition rates than methane-
based films at a given rf induced substrate bias voltage and flow rate, indicating that the
carbon flux to the deposition substrate in acetylene-argon discharges is higher than in the
methane-argon discharges. The depositing flux of carbon to the surface of the substrate at
a typical experimental deposition rate of 40 nm/min and a film of density 2.0 g/cm3 is
4x10l7 carbon/cmZ/min. For comparison, the peak deposition rate possible can be
estimated assuming all the carbon entering the system is deposited uniformly on the
surfaces defined by the substrate holder, quartz dome, and plasma source region walls.
For this estimate, a flow rate of 7 seem for the hydrocarbon gas is assumed to enter the
plasma source. This flow rate is 1.8x1020 molecules/min or 1.8x1020 carbon-atoms/min
for methane and 3.6x1020 carbon-atoms/min for acetylene. The surface area of the region
defined by the substrate holder, the quartz dome top, and the plasma source/quartz dome

sidewalls is approximately 400 cm2 when we consider the surface to be a surface of

128

 

 

virtual sphere of 10 cm radius. If all the incoming methane carbon atoms are assumed to
be ionized or dissociate and then to flow and stick to the walls defined by this surface
area, the deposition rate for a carbon film of density 2.0 g/cm3 would be 45 nm/min. This
deposition rate is similar in magnitude to that actually deposited on the substrate, which
ranges fi'om 10-40 nm/min as shown in Fig. 5-10. This suggests that most of the methane
injected into the plasma is activated and it is deposited on the substrate or walls. This is
also supported by the partial pressure analyzer data in Fig. 5-3 to Fig. 5-6, in which most
of the carbon in the hydrocarbon gas flows is deposited into the films or on the walls
since hydrocarbon gas is found in the exit gas flow at a low levels of 10% for acetylene
and 20 % for methane when the discharge is on as compared to the level with no
discharge present. For the case of acetylene-based films, the deposition rate is higher
because acetylene has two carbons per molecules instead of the one of methane. Hence
the deposition rate is expected to be approximately two times of the rate of methane. This
approximate doubling of the deposition rate is in fact observed in Fig. 5-10.

Another way to evaluate the fluxes to the deposition surface is to compare the flux
of carbon to the substrate, the flux of ions to the surface, and the flux of neutral species to
the surface. Here, the species consist of argon, hydrogen, and hydrocarbon molecules.
The measured current density to the substrate surface at a pressure of 3 mTorr is typically
in the range of 2-3 mA/cm2 as given in Table 5-2. This corresponds to an ion flux of
approximately ~1 018 ions/min/cmz. The neutral flux can be estimated fiom the neutral gas

temperature and pressure using the ideal gas law and neutral diffusion I‘ given by
[Emu/4 where 11,, is the neutral density and v” is the average neutral speed. This

estimation at a pressure of 3 mTorr and a gas temperature of 600K [77] is ~1020

129

 

species/min/cm2 depending on the species' mass assumed which ranged from 1 amu to 40
amu. This simple flux counting exercise indicates that for each carbon incorporated into
the film, one to a few ions arrive at the deposition surface and 100 or more neutral
species arrive at the surface, some of which may be chemically active neutral radicals
that are deposited.

A comparison of the film properties for methane and acetylene-based films given
in Fig. 5-10 - Fig. 5-14 shows substantial differences between the two discharge types. At
any selected rf induced bias, the acetylene-based films have a higher density, lower
hydrogen content, higher refractive index, and lower optical bandgap as compared to the
methane-based films. To interpret this difference the hydrogen content will first be
examined. The acetylene-based fihns have a substantially lower amount of hydrogen
incorporated during the deposition. First, this is what can be expected to occur because
the hydrogen/carbon ratio for acetylene is l and the hydro gen/carbon ratio for methane is
4. Hence, more hydrogen is available in the methane discharges. An examination of the
partial pressure analyzer (PPA) spectrum of Fig. 5-3 - Fig. 5-6 confirms this since it
shows that the hydrogen concentration in the exit gas increases when the discharge is on
for methane or acetylene as compared to the discharge being off. Further for methane, if
the amplitude of the PPA mass spectrum signals for hydrogen are compared to the argon
(40 amu) signal for the discharge on case, the equivalent number of hydrogen atoms
flowing out of the system for methane is at most twice of the argon flow. Since the input
gas flow of methane is 8 seem (or 32 sccm equivalent hydrogen atom flow), the flow of
argon is 8 seem and the equivalent flow of hydrogen atom out of the system is at most 16

sccm (2 times of argon flow), less than half of the hydrogen that entered the system flows

130

 

 

 

out through the pumping system This leaves more than half of hydrogen (more than 2
hydrogen atoms per carbon atom) being incorporated into the carbon on the substrate and
walls.

The behavior in the acetylene-argon plasmas is different. The hydrogen content in
the exit gas flow does not increase as much as the methane-argon case when the
discharge is on indicating that most of the hydrogen is incorporated into the carbon film
deposited on the walls and the substrate. This hydrogen incorporation rate is therefore
less than 1 hydrogen per carbon because the input gas acetylene has 1 hydrogen atom for
each carbon atom. Hence, the methane discharges have an abundance of hydrogen which
is both incorporated into the carbon on the walls and the substrate and pumped out of the
system, and the acetylene discharge has some hydrogen that is incorporated into the
carbon deposited on the walls and the substrate. It should also be noted that as the bias on
the substrate is made more negative the hydrogen is driven out of the films as shown in
Fig. 5-12.

The density of the films increased with the increasing magnitude of the rf induced
substrate bias and the acetylene-based films are denser than the methane-based films in
Fig. 5 -1 1. The less hydrogen content at higher magnitude of the rf induced substrate bias
and of acetylene-based films is considered to make the films more dense. The index of
refraction increases with the increasing magnitude of the rf induced substrate bias and is
higher for the acetylene-based films as compared to the methane-based films in Fig. 5-13.
The index of refi'action of materials is dependent on the density of materials and their
compositions. The variation of the index of refiaction can be explained with the hydrogen

content of the films such as that the less hydrogen content at higher magnitude of the rf

131

 

 

 

induced substrate bias and for acetylene-based films made denser films giving a higher
refractive index.

The optical bandgap decreases with the increasing magnitude of the rf induced
substrate bias for both films from acetylene and methane, and the optical bandgap of
acetylene-based films has lower values than those of methane-based films for films
grown with the same rf induced bias as indicated in the bandgap energy, E04, data of
Fig. 5-14. The optical bandgap is the material property determined by the composition of
the material. In this case, the composition of the films is determined by the hydrogen
content, sp2 fraction and sp3 fraction. The hydrogen content is considered to induce the
variation of the optical bandgap variation in Fig. 5-14. This can be seen more clearly in
the data that shows the E04 optical bandgap versus hydrogen content as shown in Fig. 5-
15. A decrease in the hydrogen content produces a lower optical bandgap energy. In the
material's structure this is attributed to the decreasing hydrogen content reducing the sp3
C-H bonding leaving more graphite-like (sp2 carbon) bonds in the films [68]. These sp2
bonds are primarily responsible for the light absorption in the visible-near infrared
spectral range [67] as it is seen that the graphite is dark and black with naked eyes.

More insight into the bonding within the films can be obtained by plotting the
optical bandgap versus film mass density as shown in Fig. 5-16. In general, a-C:H films
are predominantly composed of three bond types including C-C sp3 bonds, C-C sp2 bonds
and OH bonds. The density is a good indicator at the higher range of densities (>2.4
g/cm3) of the sp3/sp2 ratio as discussed in Section 2.1 with high density films having a

high sp3/sp2 ratio. At the lower density range (less than 2.0 g/cm3), the density is most

132

 

 

 

 

Optical Bandgap, E04 (eV)

3.5

 

 

 

 

3.0 r
I
2.5 '
O
2.0 - o '
I
O§C2H2

1.5 L- IzCH4
1.0 I 1 I 1 I 1 I

20 30 40 50 60

Hydrogen Content (at. % )

70

Fig. 5 - 15: Variation of optical bandgap versus hydrogen content for

133

acetylene-based films and methane-based films.

 

 

 

 

 

3.5 E
O: C2H2
.3 CH4
I + 3 Csz [3]
3.0 r
E I +
as? 2.5 -
:2 O
s.
7% ++
+
E 2.0 - I o
a
.2
2% - +
CI.
1.5 -
a-C:H ta-C:H
1.0 1 1 1 3
0.5 1.0 1.5 2.0 2.5 3.0
Density (g/cm3)

Fig. 5 - 16: Variation of optical bandgap versus mass density for
acetylene—based films and methane-based films deposited in
this study and in the study of Ref. [16].

134

dependent on the amount of hydrogen incorporated as shown in Fig. 5-11 and Fig. 5—12.
Fig. 5-16, which shows both our data and data from [3, 78], can be interpreted as (1) the
low density films (less than 1.5 g/cm3) being polymer-like with high hydrogen content
and many C-H bonds, (2) the intermediate density a-C:H films (2.0 g/cm3) having a
reduced number of OH bonds and having more C-C bonds with some being of the sp2
type, and (3) the high density ta-C:H films (greater than 2.4 g/cm3) having few C-H
bonds and an increase in the sp3 C-C bonds and reduction in sp2 C—C bonds. This reduced
sp2 concentration in the densest films produces the increase in the optical bandgap shown
in Fig. 5-16. Thus, the optical bandgap is mainly determined by the hydrogen content for
low density films and by the carbon-carbon sp3 fraction for high density films. Thus the
films deposited at pressures in the millitorr range have decreasing optical bandgap with
the increasing magnitude of the rf induced substrate bias because the sp2 bondings
increases with the variation of the bias. On the other hand, the films deposited at
submillitorr pressure (0.2 mTorr) have a peak value of optical bandgap at -200 V of rf
induced substrate bias and the occurrence of the peak is considered to be due to the

variation of sp3 fraction as discussed in Chapter 4.

5.3.3 The Effect of Pressure

The effect of pressure on the film properties including refractive index, deposition
rate, and optical bandgap is shown in Fig. 5-17 - Fig. 5-19. The deposition condition in
the figures is 200 W microwave input power, 8 sccm flow rate of acetylene or methane
gas, 8 seem flow rate of argon gas, -20 V rf bias for acetylene-based films or -50V for
methane-based films. The two different rf induced substrate bias conditions were selected

so that both the methane-based and the acetylene-based films had similar densities and

135

hydrogen content at a 3 mTorr deposition condition. The deposition rate of the acetylene-
based films is larger than the methane-based films for all pressures considered and the
rate for both discharges have the increasing trend as the pressure increases. Both the
refiactive index and the optical bandgap change minimally with the variation of pressure
for methane-based films. When the pressure is increased, the refractive index of
acetylene-based films has a decreasing trend and the optical band gap has an increasing
trend. The pressure dependence given in Fig. 5-17 -Fig. 5-19 showed that the acetylene
film properties including optical bandgap and refi'active index are changed as a function
of pressure.

The above results for the pressure variation suggest the ion bombardment effect
decreases with the increasing discharge pressure and changes significantly for argon-
acetylene discharge cases, and changes minimally for argon-methane discharge cases.
The decreasing trend of ion bombardment effect with the increasing pressure is expected
because the ion density does not increase as fast as neutral flux as the discharge pressure
increases (as seen in Fig. 5-1) resulting in the decrease of the ratio of ion flux to neutral
flux onto the substrates. To estimate this change in flux ratio, the argon discharge data
shown in Fig. 5-1 will be used. This data first indicates a drOp in the ion density at l
mTorr to 40 % of the ion density at 8 mTorr. Next, for this same pressure variation the
neutral density and hence neutral flux would change by a factor of 8 giving a value at 1
mTorr that is less than 15% of the value at 8 mTorr. Combining these two observations
gives that the ion- flux to neutral flux ratio would then increase by a factor of about 3

when the pressure drop from 8 to l mTorr. The changed film properties at lower

136

pressures are therefore produced by the higher percentage of ions compared to neutral
species reaching the substrate.

The acetylene-based films responded more sensitively to the variation of
discharge pressure. This is explained with a significant fraction of the ions being heavier
ions in argon-acetylene discharges. The momentum of ions is proportional to the square
root of ion's mass when the ions have the same energy. And the ions onto a substrate are
supposed to have the same energy because they have traveled through the same sheath
potential regardless of the ion types. Thus, heavy ions like argon and acetylene have
bigger momentum and transfer bigger momentum onto the growing films (i.e. bigger ion
bombardment effect) than light ions like hydrogen molecules. Table 5-4 shows the
momentum of several ion types normalized by the momentum of an atomic hydrogen ion
under the condition of the same ion energy.

Table 5-4: The dependence of momentum on mass. The momentum of several ion types

is normalized by that of an atomic hydrogen ion under the condition of the same ion
energy and it is designated by Mx/MH

 

 

 

 

IOIl type 14+ H2+ CI‘I4+ C2H2+ {311'+
Mass (amu) '1 2 16 26 40
Mx/MH 1 1.4 4.0 5.1 6.3

 

 

 

 

 

 

 

The argon-acetylene discharges have a higher fraction of the heavier argon ions than
hydrogen ions as compared to argon-methane discharges. This higher concentration of
argon was shown by the PPA measurements. Hence the ion bombardment effect is much
more significant for the acetylene-based films. Further, as the pressure is reduced, the ion

flux to neutral flux ratio is increased showing significant film property changes for

137

Deposition Rate (nm/min)

120

 

100 -

 

co
G
i

O\
G
l

CH4

uh
C
I

20-

 

 

0 I l l I I
0 l 2 3 4 5 6

Pressure (mTorr)

 

Fig. 5 - 17 : Deposition rate of methane and acetylene-based films versus
deposition pressure.

138

 

Index of Refraction

 

 

 

 

2.4 ~

2.2 ~

2.0 -

1.8 -

1.6 ‘ ‘ ' ' '
0 l 2 3 4 5

Pressure (mTorr)

Fig. 5 - 18: Index of refraction of methane and acetylene deposited films
versus deposition pressure.

139

 

2.5

5"
G
I

 

Optical Bandgap (eV)
p—i
u:

L"
G
I

 

 

 

 

0.5 I I I I I
0 1 2 3 4 5

Pressure (mTorr)

Fig. 5 - 19: Optical bandgap (Em, and E04) of methane and acetylene
deposited films versus deposition pressure.

140

acetylene based fihns. A possible additional reason is the change of the fiaction of ion
types coming from the variation of the partial pressure of gas types in the discharge as the
discharge pressure increases. The ratio of the partial pressure of argon to the partial
pressure of hydrogen is expected to decrease for argon-acetylene discharges as the
discharge pressure increases. This occurs because the hydrogen partial pressure will be
increased by the increased dissociation of acetylene gas at higher pressure. So the ratio of
the partial pressure of argon gas to the partial pressure of hydrogen gas can be higher at
low pressure than at high pressure in argon- acetylene discharges. That will result in a
lower ratio of the argon ions to hydrogen ions, thus a lower ion bombardment effect in
the deposition process at higher discharge pressures.

The methane-based films show minimal change for the pressure variation. In the
argon-methane discharge the partial pressure of hydrogen gas is almost at the same level
as the partial pressure of argon, as seen in Fig. 54, while in the argon-acetylene
discharge the level of partial pressure of hydrogen is about 1/3 of the argon level, as seen
in Fig. 5-6. Thus, the hydrogen will compete more with argon in ionization and dilute
more the argon ion bombardment effect in the deposition process in the argon-methane
discharge than in the argon-acetylene discharge. Hence, the argon ion bombardment
effect is expected smaller in argon-methane discharges than in argon-acetylene
discharges. Therefore, the decreased ion flux to neutral flux ratio produced by the
increased discharge pressure only shows a slight effect on the film properties in the
figures. And the probable change of ion type fractions by the variation of discharge
pressure does not seem to change enough to show its large effect on the film properties in

the argon-methane discharges

l4l

5.3.4 The Effect of Argon Flow Rate

The film properties and deposition rate versus argon flow ratio (i.e. the ratio of
argon flow rate to acetylene flow rate) for both the methane-argon and acetylene-argon
discharges are shown in Fig. 5-20 - Fig. 5-22. These films were deposited at 3 mTorr
pressure and 200 W microwave input power with -50 V rf induced substrate bias in the
methane-argon discharge and -20 V rf induced substrate bias in the acetylene-argon
discharge. The two different rf induced substrate bias conditions were selected, so that
both the methane-based and the acetylene—based films had similar densities and hydrogen
content at a 3 mTorr deposition condition. The figures show that the deposition rate and
optical bandgap have decreasing trends for argon-methane discharge, and the index of
refraction has increasing trend for both methane-based films and acetylene-based films as
the argon flow ratio increases. The trends are the same with those of the rf induced
substrate bias effect case. Thus, the increased argon flow rate results in the increased ion
bombardment effect in the deposition process as like the case of the increased magnitude
of the rf induced substrate bias. This fact is very natural and expected because the
increased argon flow will increase the argon partial pressure, thus argon ions resulting in
the enhanced ion bombardment effect.

The data indicates that the addition of argon has no or only a small effect on the
films deposited fiom the acetylene gas and a significant influence on the fihns deposited
from methane. Specifically, both the optical bandgap and the deposition rate for the
acetylene-argon films show no variation versus argon addition of O to 0.8 argon flow

ratio. Increasing the argon flow ratio lowers the fihn's optical bandgap for methane-based

142

 

120

110

 

 

 

Deposition Rate (nm/min)

\O
G
r

 

 

80 1 I 1 I 1 I 1
0.0 0.2 0.4 0.6 0.8

Argon Flow Ratio

 

Fig. 5 - 20: Deposition rate of methane and acetylene deposited films versus
argon flow ratio.

143

 

2.2

2.1 -

 

 

 

Index of Refraction

 

 

 

0.0 0.2 0.4 0.6 0.8
Argon Flow Ratio

Fig. 5 - 21: Index of refraction of methane and acetylene deposited films
versus argon flow ratio.

144

Optical Bandgap (eV)

 

2.5

E"
c
I

 

 

L"
u-

1"
c

 

 

 

  

 

 

 

 

13...... €sz 4
L ‘T
0 5 1 I 1 I 1 J 1 7
0.0 0.2 0.4 0.6 0.8
Argon Flow Ratio

Fig. 5 - 22: Optical bandgap (Eta.ch and E...) of methane and acetylene
deposited films versus argon flow ratio.

145

films but it does not effect the acetylene-based films as shown in Fig. 5-21. Similarly, the
deposition rate and the refractive index change more versus argon flow ratio in methane
based fihns as compared to acetylene based films as shown in Fig. 5-20 and Fig. 5-22.
Thus, the methane-based films are more sensitive to the argon addition than the
acetylene-based films. The partial pressure of hydrogen is higher in the argon-methane
discharge than in the argon-acetylene discharge as seen in Fig. 5-4 and Fig. 5-6. So the
addition of argon can more effectively replace the hydrogen gas by the added argon gas
to some extent in the argon-methane discharges as compared to the argon-acetylene
discharge. Thus, the argon ion bombardment effect can be more effectively enhanced by
the addition of a certain amount of argon in the argon-methane discharge than in argon-
acetylene discharge. For an another reason the momentum of argon ions is 4.5 times and
1.6 times bigger than hydrogen ions and methane ions, respectively (see Table 5-4 for the
momentum). On the other hand, the partial pressure of hydrogen is lower in argon-
acetylene discharges, thus the added argon will mainly compete in ionization with
acetylene gas. But the momentum of argon ion is only 1.2 times bigger than acetylene
ions. Thus, the ion bombardment effect of the increased argon ions is not very different
fiom the acetylene ions in the argon—acetylene discharge, but is much bigger than the
hydrogen ions and methane ions in the argon-methane discharges. Therefore, the
methane-based films were more sensitive to the addition of argon than the acetylene-

based films.

146

5.4 Summary

Hydrogenated amorphous carbon films were deposited at pressures in the millitorr
range using discharges of argon-acetylene and argon-methane mixtures. The input
microwave power was 250 W. The rf induced substrate bias, deposition pressure and
argon flow rate were varied to investigate their effect on the film properties.

The film properties varied from higher deposition rate, lower refractive index,
higher optical bandgap, lower density, higher hydrogen content films to lower deposition
rate, higher refractive index, lower optical bandgap, higher density, lower hydrogen
content films as the magnitude of rf induced substrate bias increased. The rf induced
substrate bias provides the ion bombardment energy onto the surface of the growing film.
The higher ion bombardment energy is expected to break the bonds of adsorbed
hydrocarbon species and to expel hydrogen from the film structure resulting in the
densification of the film structure more easily and more effectively. The films of less
hydrogen content will have higher density and lower optical bandgap. The densification
of films will reduce the deposition rate and increase the refractive index.

The two discharge types of argon-acetylene and argon-methane mixtures
produced significantly different film properties. Fihns deposited from argon-acetylene
discharges have a higher deposition rate, higher film density, lower hydrogen content,
higher refractive index and lower optical bandgap than the film deposited from argon-
methane discharges. Acetylene molecules have two times more carbon atoms and two

times less hydrogen atoms in them than methane molecules, which explains the higher

147

deposition rate and lower hydrogen content of the acetylene-based films than the
methane-based films. The higher density, higher refi'active index and lower optical gap of
the acetylene-based films are also explained by the lower hydrogen content of the films.
The film properties are summarized in Table 5-5 from Fig. 5-10 to Fig. 5-14 for easy
comparison of the effects of rf induced substrate bias and the two different discharges.

Table 5-5: Comparison of film properties fi'om argon (SO %)—methane (50 %) and argon

(50 %)-acetylene (50 %) discharges, and from two different rf induced substrate biases of
O and -60 V.

 

 

 

 

 

 

 

 

 

Discharges Argon-methane Argon-acetylene
RF induced 0 (V) -60 (V) 0 (V) -60 (V)
substrate bias
Deposition rate 40 10 80 6O
(Hm/min)
Density 0.9 2.2 1.6 2.4
(g/cm3)
Hydrogen content 65 3O 33 25
(%)
Refractive index 1.8 1.9 1.9 2.2
E04 3.1 1.8 2.4 1.5
feV)
Em“, 2.0 1.2 1.3 0.8
16V)

 

 

 

 

 

As the deposition pressure increased, the deposition rate and the optical bandgap
increased and the refractive index decreased for both methane-based films and acetylene-
based films. These facts suggest the films can be densificated more effectively at lower
pressure and also suggest the ion bombardment effect decreases with the increasing
discharge pressure and changes significantly for argon-acetylene discharge cases, and
changes minimally for argon-methane discharge cases. The decreasing trend of ion
bombardment effect with the increasing pressure is expected because the ion density does

not increase as fast as neutral flux as the discharge pressure increases resulting in the

148

 

 

 

 

 

 

 

decrease of the ratio of ion flux to neutral flux onto the substrates. Thus, the ion
bombardment effect decreased as the discharge pressure increases and this explains the
pressure effect on the film properties. The acetylene-based films responded more
sensitively to the variation of discharge pressure. This is explained with a significant
fiaction of the ions being heavier ions in argon-acetylene discharges.

As the argon flow ratio increased into the discharge chamber, the optical bandgap
decreased and the refractive index increased. This fact suggests the films were
densiflcated as the argon flow ratio increased and is an expected result because the argon
ion bombardment will be enhanced with the increased argon flow ratio.

The effect of argon flow ratio was higher for the methane-based films than the
acetylene-based films. Because the hydrogen species in the argon-methane discharge are
at a higher level than in the argon-acetylene discharge at 50 % argon and 50 %
hydrocarbon flow rates, the addition of more argon will replace hydrogen species with
the argon more effectively in argon-methane discharges than in argon-acetylene
discharges. Therefore, as the argon flow ratio increases, the argon to hydrogen species
ratio in the argon-methane discharges will increases more quickly than in the argon-
acetylene discharges enhancing the argon ion bombardment effect. Thus, the methane-
based films have higher fihn property variation as the argon flow ratio changes.

Therefore, the properties of the fihns deposited at pressures in the millitorr range
with the argon-hydrocarbon gas mixtures are strongly influenced by the argon ion
bombardment energy and argon ion flux on the surface of grong films. The argon ion
energy is determined by the rf induced substrate bias and the argon ion flux on the

surface of growing films is dependent on the argon to hydrogen species ratio in the

149

discharge chamber. The species ratio is varied by the deposition pressure and the ratio of
argon flow to hydrocarbon gas flow into the deposition chamber. The ion bombardment
effect determined the hydrogen content and the variation of film properties in the film
deposition at pressures in the millitorr range can be mainly explained with the
consideration of the hydrogen content of the films.

The variation of the properties in the films deposited at pressures in the millitorr
range can be mainly explained with the consideration of the hydrogen content of the
films. In contrast, the variation of film properties in the film deposition at pressures in the
submillitorr range is mainly attributed to the film composition of sp3/sp2 ratio as

discussed in Chapter 4.

150

 

Chapter 6

6. Conclusions

Hydrogenated carbon films were deposited in an ECR-CVD system with a rf
biased substrate using acetylene, acetylene-argon and methane-argon gases mixtures. The
films were deposited at pressures in the submillitorr range (0.2-0.6 mTorr) for acetylene
discharges and at pressures in the millitorr range (1-5 mTorr) for acetylene-argon and
methane-argon discharges. The former clearly showed the effects of ion energy and ion
flux to neutral flux ratio on the film properties in the process of deposition and the latter
revealed that the two discharge types of argon-methane and argon-acetylene mixture
produced significantly different film properties and the optical properties of the films can
be controlled by the variation of deposition conditions. The study of discharge properties
provides some information on the discharge deposition conditions such as ion energy,
rate-limiting process of the deposition, ionization levels of the discharges and a rough
estimation of ion types.

The films deposited at pressures in the submillitorr range have a peak (1.3 eV for
Em, and 1.8 eV for E04) in their optical bandgap at -200 V of rf induced substrate bias,
when operated at 0.2 mTorr with acetylene feed gas. The result is consistent with other
researcher's investigations and matches well with the subplantation model showing that
an sp3 peak occurs at ion energies of 90-100 eV per carbon atom in the deposition
process. The variation of optical bandgap in films deposited at conditions near -200 V of

rf induced substrate bias is rendered primarily by the carbon sp3 to sp2 ratio. The high

151

  

values of optical bandgaps of films deposited at lower magnitude of rf induced substrate
biases (< 100 V) are considered due to high hydrogen content of the films.

The low deposition temperature was critical for the films to have the peak value
of optical bandgap at -200 V of rf induced substrate bias. In particular the optical
bandgap of films deposited at higher temperature gave a lower value (1.19 eV for Emuc)
than the peak value (1.44 eV for Emuc) at -200 V of rf induced substrate bias. The effects
of pressure and microwave power on the film's properties were not clearly shown in this
investigation within the deposition variable space.

The effect of substrate position was clear so that the peak of the optical bandgaps
appeared at substrate position located closer to the plasma generation region where the
ion flux to neutral flux ratio is larger. The result indicates that the ion flux to neutral flux
ratio is also a critical factor to obtain films of a high optical bandgap with a high sp3 ratio.
Thus the deposition of the films at pressures in the submillitorr range clearly showed the
effect of ion energy and ion flux to neutral flux ratio on the film properties. The flux ratio
was roughly estimated to be about 10 % and the threshold ratio of ion flux to neutral flux
for deposition of ta—C:H films is found to be in the range of 0.06-0.1. The occurrence of
peak of the optical bandgap at -200 V of rf induced substrate bias conforms to the
explanation of the subplantation model of deposition for ECR deposition and
demonstrated agreement with the other ta-C:H film depositions in the literature.
However, the ion flux to neutral flux ratio of this investigation (10 %) is estimated lower
than the other values (more than 90 %) in literature.

The deposition at pressures in the submillitorr range, therefore, showed that ta—

C:H films can be deposited with the microwave ECR deposition system of this

152

investigation and the deposition rate (90 nm/min) is higher than those of the plasma beam
deposition (15 nm/min) and the filtered ion beam deposition.

For the films deposited at pressures in the the millitorr range the film properties
varied from higher deposition rate, lower refiactive index, higher optical bandgap, lower
density, higher hydrogen content films deposited at low rf induced biases to lower
deposition rate, higher refractive index, lower optical bandgap denser, lower hydrogen
content films deposited at high ion energies. The two discharge types of argon-methane
and argon-acetylene mixture produced significantly different film properties including
higher density, lower hydrogen content, higher optical absorption, higher refractive
indices and higher deposition rates for the acetylene-based films as compared to the
methane—based fihns. Insight into the deposition mechanism for each of these discharge
types was gained by studying the variation in film properties produced by variations in
the deposition conditions. Two specific results are, first, that as the pressure is changed
the methane/argon films do not change in properties, whereas, the acetylene-based films
do change properties, and second when the argon flow ratio in acetylene and methane
discharges is changed, the methane-based films show a large change in properties and the
acetylene-based films do not change properties.

The properties of the films deposited at pressures in the millitorr range with the
argon-hydrocarbon gas mixtures are strongly influenced by the argon ion bombardment
energy and argon ion flux on the surface of growing films. The argon ion energy is
determined by the rf induced substrate bias and the argon ion flux on the surface of
growing films is dependent on the argon to hydrogen species ratio in the discharge

chamber. The species ratio is varied by the deposition pressure and the ratio of argon

153

flow to hydrocarbon gas flow into the deposition chamber. The ion bombardment effect
determined the hydrogen content and the variation of film properties in the film
deposition at pressures in the millitorr range can be mainly explained with the
consideration of the hydrogen content of the films.

The film property and deposition rate variations with discharge conditions
together with data collected on the discharge itself using a partial pressure analyzer and a
Langmuir probe provides a picture of the discharge deposition conditions. The
conclusions reached for both the methane and acetylene discharges is that most of the
carbon entering the discharge in the hydrocarbon gas flow is activated to be either an ion
or a neutral radical which deposits on either the substrate or the chamber walls. The rate-
limiting process for the deposition is the flow of carbon species into the plasma source.
The deposition in the methane-argon discharge proceeds with the dominant species in the
processing chamber being hydrogen and argon with the relative percentages of each
changing based on the input flow rate of each gas. For fihns deposited from acetylene-
argon discharges the acetylene is activated and it deposits on the walls, leaving argon as
dominant species in the processing chamber.

Comparison of the results of depositions at submillitorr and millitorr range of
deposition pressure yields that low pressure reduces the neutral flux to the surface.
Further, the removal of argon and the application of a -200 V rf induced substrate bias in
the submillitorr range of pressure provide the proper ion bombardment ion energy onto
the layer of growing fihn for deposition of ta-C:H giving the peak value of optical
bandgap at the —200 V rf induced substrate bias. The study of deposition at pressures in

the millitorr range also showed the film properties could be varied to some extent by the

154

selection of deposition source gases and varying the other deposition conditions such as
rf induced substrate bias, deposition pressure and argon flow ratio. The variation of film
properties in the film deposition at pressures in the millitorr range can be mame
explained with the consideration of the hydrogen content of the films. But in contrast the
variation of fihn properties in the film deposition at pressures in the submillitorr range is
mainly attributed to the film composition of sp3/sp2 ratio.

In this investigation the direct measurement of fihn composition, i.e. percent sp3,
percent spz, percent argon and percent helium was not done. The film composition
governs the film's mechanical, optical properties. The measurement techniques of ion
types in the plasmas, their ratios in the plasma and the exact carbon ion flux to neutral
flux ratio onto the substrate were not completed in this study. Therefore the unmeasured
or undetermined film composition, the internal variables of plasma and film properties
would be useful future investigations to fully understand the film deposition and film
properties.

The film's electrical properties like dielectric constant, bandgap and resistance,
mechanical properties like hardness, friction coefficient, stress, heat conduction
coefficient and coefficient of thermal expansion, and etch properties would be interesting
investigation areas for the optical, mechanical and electrical applications of the films. The
applications could be optical filters, protective coatings, active electrical devices, and

insulating layers in electronic packaging.

155

 

List of References

[1]

[2]

[3]

[4]

[5]

[6]

P. Koidal, C. Wild, R. Locher, and R. E. Sah, Amorphous Hydrogenated Carbon
Films and Related Materials: Plasma Deposition and Film Properties. in
Diamond and Diamond-like Films and Coatings, R. E. Clausing, L. L. Horton, J.

C. Angus, and P. Koidl, Eds. New YorK: Plenum Press, 1991.

J. Robertson, “Deposition of Diamond -Like Carbon,” Phil. Trans. R. Soc. Land.

A, vol. 342, pp. 277-286, 1993.

M. Weiler, S. Sattel, T. Giessen, K. Jung, H. Ehrhardt, V. S. Veerasamy, and J.
Robertson, “Preparation and Properties of Highly Tetrahedral Hydrogenated

Amorphous Carbon,” Phys. Rev. B, vol. 53, pp. 1594, 1996.

G. F. Zhang, L. J. Guo, Z. T. Liu, X. K. Xiu, and X. Zheng, “Studies on
Diamondlike Carbon Films for Antireflection Coatings of Infrared Optical

Materials,” J. Appl. Phys, vol. 76, pp. 705, 1994.

A. H. Lettington, Applications of Diamond-Like (Hard Carbon) Films. in
Diamond and Diamond-like Films and Coatings, R. E. Clausing, L. L. Horton, J.

C. Angus, and P. Koidl, Eds. New York: Plenum Press, 1991.

M. Allon-Alaluf, L. Klibanov, A. Seidman, and N. Croitoru, “Metal Contacts and
Electrical Processes in Amorphous Diamond-Like Carbon Films,” Diamond Relt.

Mater., vol. 5, pp. 1275, 1996.

156

 

[7]

[8]

[9]

[10]

[11]

[12]

M. Allon-Alaluf, L. Klibanov, and N. Croitoru, “Iodine Doping of Amorphous

Diamond-like Carbon Films,” Diamond Relat. Mater., vol. 5, pp. 1497, 1996.

L. Klibanov, M. Allon-Alaluf, N. Croitoru, and A. Seidman, “Study of
Photoconductivity in Thin Amorphous Diamond-like Carbon (a: DLC) Films
Prepared by Glow Discharge Technique,” Diamond Relat. Mater., vol. 5, pp.

1414, 1996.

N. Konofaos and C. B. Thomas, “Characterization of Heterojunction Devices
Constructed by Amorphous Diamondlike Films on Silicon,” J. Appl. Phys, vol.

81,pp.6238,1997.

S. Egret, J. Robertson, W. I. Milne, and F. J. Clough, “Diamond-like Carbon
Metal-Semiconductor-Metal Switches for Active Matrix Displays,” Diamond

Relat. Mater., vol. 6, pp. 879, 1997.

C. H. Lee and K. S. Lirn, “Boron-Doped Amorphous Diamondlike Carbon as a
New p-type Window Material in Amorphous Silicon p-i-n Solar Cells,” Appl.

Phys. Lett, vol. 72, pp. 106, 1998.

Y. Catherine, Preparation Techniques for Diamond—Like Carbon. in Diamond
and Diamond-like Films and Coatings, R. E. Clausing, L. L. Horton, J. C. Angus,

and P. Koidl, Eds. New York: Plenum Press, 1991.

157

[13]

[14]

[15]

[16]

[17]

[13]

A. Grill and V. Patel, “Effects of Bias and Inert Gas on Properties of Diamond-
Like Carbon Deposited by dc. PA CVD,” Diamond and Related Materials, vol.

4,pp.62,1994.

J. W. Zou, K.Reinchelt, K. Schimidt, and B. Dischler, “The Deposition and Study

of Hard Carbon Films,” J.Appl. Phys, vol. 65, pp. 3914, 1989.

M. Weiler, S. Sattel, K.Jung, H.Ehrhardt, V. S. Veerasamy, and J. Robertson,
“Highly Tetrahedral, Diamond-like Amorphous Hydrogenated Carbon Prepared

fiom a Plasma Beam Source,” Appl. Phys. Lett, vol. 64, pp. 2797, 1994.

P. J. Fallon, V. S. Veerasamy, C. A. Davis, J. Robertson, G. Amaratunga, W. I.
Milne, and J. Koskinen, “Properties of Filtered-Ion Beam Deposited Diamondlike

Carbon as a Function of Ion Energy,” Phys. Rev. B, vol. 48, pp. 4777, 1993.

Y. Lifshitz, G. D. Lempert, S. Lotter, l. Avigal, C. U. Saguy, and R. Kalish, “The
Influence of Substrate-Temperature during Ion-Beam Deposition on the
Diamond-er or Graphitic Nature of Carbon-Films,” Diamond and Relat. Mater.,

vol. 2, pp. 285, 1993.

Y. Lifshitz, G. D. Lempert, S. Lotter, I. Avigal, C. U. Saguy, R. Kalish, J.Kulik,
D. Marton, and J. W. Rabalais, “The Effect of Ion Energy on the Diamond-like
Graphitic (sp3/sp2) Nature of Carbon Films Deposited by Ion-Beam,” Diamond

Relat. Mater., vol. 3, pp. 542, 1994.

158

 

 

[19]

[20]

[21]

[22]

[23]

[24]

C. Ronning, E. Dreher, J .-U. Thiele, P. Oelhafen, and H. Hofsass, “Electronic and
Atomic Structure of Undoped and Doped ta-C Fihns,” Diamond Relat. Mater.,

vol. 6, pp. 830, 1997.

S. F. Yoon, H. Yang, A. Rusli, J. Ahn, and Q. Zhang, “The Effects of Self-
Generated DC Bias on the Characteristics of Diamond-like Carbon Fihns

Prepared using ECR-CVD,” Diamond Relat. Mater, vol. 7, pp. 70, 1998.

M. Zarrabian, N. Fourches-Coulon, G. Turban, M. Lancin, and C. Marhic, “Effect
of Negative Bias Voltage on a-C:H Films Deposited in Electron. Cyclotron

Resonance Plasma,” Diamond Relat. Mater. , vol. 6, pp. 542, 1997.

S. C. Kuo, E. E. Kunhardt, and A. R. Srivatsa, “Deposition of Diamond-like
Carbon Fihn Using Electron Cyclotron Resonance Plasma,” Appl. Phys. Lett, vol.

59,pp.2532,199l.

T. Fujita and O. Matsumoto, “Deposition of Carbonaceous Films using ECR
Plasma Apparatus Deposition of Colorless, Tranparent and Semiconducting Film

from Methane Plasma,” J. Electrochem. Soc., vol. 136, pp. 2624, 1989.

A. Zeinert, B. Racine, K. Zellama, M. Zarrabian, G. Turban, and A. Sadki,
“Comparative Study of the Optical and Structural Properties of a-C:H Films
Deposited by Conventional Diode and Microwave ECR Assisted RF Plasma,”

Diamond Relat. Mater., vol. 7, pp. 486, 1998.

159

 

[25]

[26]

[27]

[28]

[29]

[30]

P. W. Pastel and W. J. Varhue, “The Effect of Radio Frequency Substrate Biasing
in the Deposition of Diamond-like Carbon Films in an Electron Cyclotron

Resonance Discharge,” J. Vac.Sci. T echnol. A, vol. 9, pp. 1129, 1991.

S. F. Yoon, H. Yang, Rusli, J. Ahn, Q. Zhang, and T. L. Poo, “Influence of
Substrate Temperature and Microwave Power on the Properties of a-C:H Fihns
Prepared using the ECR-CVD Method,” Diamond Relat. Mater. , vol. 6, pp. 1683,

1997.

S. F. Yoon, Rusli, J. Ahn, Q. Zhang, Y. S. Wu, and H. Yang, “DC Bias Effects in
the Deposition of a—C:H Fihns using the Screen-grid Method in Electron

Resonance Chemical Vapor Deposition,” Diamond and Relat. Mater., vol. 7, pp.

1213,1998.

F. S. Pool and Y. H. Shing, “Deposition of Diamondlike Films by Electron

Cyclotron Resonance Microwave Plasma,” J. Appl. Phys, vol. 68, pp. 62, 1990.

P. S. Andry, P. W. Pastel, and W. J. Varhue, “Comparison of Diamond-like
Carbon Fihn Deposition by Electron Cyclotron Resonance with Benzene and

Methane,” J. Mater. Res, vol. 11, pp. 221, 1996.

K. Kuramoto, Y. Domoto, H. Hirano, S. Kiyama, and S. Tsuda, “High Quality
Diamondlike Carbon Thin Film Fabricated by ECR Plasma CVD,” Appl. Surf.

Sci, vol. 113/114, pp. 227, 1997.

160

 

[31]

[32]

[33]

[34]

[35]

[36]

[3 7]

[38]

J. C. Angus and Y. Wang, Diamond-like Hydrocarbon and Carbon Films. in
Diamond and Diamond-like Films and Coatings, R. E. Clausing, L. L. Horton, J.

C. Angus, and P. Koidl, Eds. New York: Plenum Press, 1991.

W. Jacob and W. Moller, “On the Structure of Thin Hydrocarbon Films,” Appl.

Phys. Lett., vol. 63, pp. 1771, 1993.

P. Reinke, W. Jacob, and W. Moller, “Influence of the Ion energy on the Growth
and Structure of Thin Hydrocarbon Fihns,” J. Appl. Phys, vol. 74, pp. 1354,

1993.

S. f. J. Robertson, “Electronic Structure of Diamond-like carbon,” Diamond Relat.

Mater., vol. 6, pp. 212, 1997.

J. Ristein, J. Shafer, and L. Ley, “Effective Correlation Energies for Defects in a-
C:H fiom a Comparison of Photoelectron Yield and Electron Spin Resonance

Experiments,” Diamond Relat. Mater., vol. 4, pp. 508, 1995.

J. F. 6. al., “Structure and Bonding of Hydrogen Plasma Generated Carbon Films:

An Electron Loss Study,” Solid State Commun, vol. 47, pp. 687, 1983.

J. Robertson, “Structural Models of a-C and a-C:H,” Diamond Relat. Mater., vol.

4,pp.297,l995.

C. H. Lee, W. R. L. Lambrecht, B. Segall, P. C. Kelires, T. Frauenheirn, and U.
Stephan, “Electron Structure of Dense Amorphous Carbon,” Phys. Rev. B, vol. 49,

pp.1l448,l994.

161

[39]

[40]

[41]

[42]

[43]

[44]

[45]

[46]

[47]

D. A. Drabold, P. A. Fedders, and P.Strumm, “Theory of Diamondlike

Amorphous Carbon,” Phys. Rev. B, vol. 49, pp. 16415, 1994.

J. Robertson, “The Deposition Mechanism of Diamond-like a-C and a-CzH,”

Diamond Relat. Mater., vol. 3, pp. 361, 1994.

J. Robertson, “Deposition Mechanism of a-C and a-CzH,” Non-Crystalline Solids,

vol. 164-166, pp. 1115,1993.

J. Robertson, “Deposition Mechanism for Promoting sp3 Bonding in Diamond-

 

like Carbon,” Diamond Relat. Mater., vol. 2, pp. 984, 1993.

J. C. Angus, P. Koidl, and S. Domitz, “Carbon Thin F ilms,” in Plasma Deposited

Thin Films, J. Mort and J. Jansen, Eds. Boca Raton: CRC Press, 1986.

W. Moller, “Hydrogen Trapping and Transport in Carbon,” J. Nucl. Mater., vol.

162,pp.138,1989.

J. F. Ziegler, J. P. Biersack, and U. Littmark, Stopping and Range of Ions in

Solids. New York: Pergamon, 1985.

F. Seitz and J. S. Koehier, Solid State Physics, vol. Vol. 2. New York: Academic

Press, 1956.

I. Koponen, M. Hakovirta, and R. Lappalainen, “Modeling the Ion Energy

Dependence of the sp3/sp2 Bonding Ratio in amorphous Diamondlike Films

162

[48]

[49]

[50]

[51]

[52]

[53]

Produced with a Mass-Separated Ion beams,” J. App. Phys, vol. 78, pp. 5837,

1995.

J. Asmussen, T. A. Grotjohn, and D. K. Reinhard, Lecture note on EE989,

Lecture 13, p20, 1999.

R. Kleber, M. Weiler, A. Kruger, S. Sattel, G. Kunz, K. Jung, and H. Ehrhardt,
“Influence of Ion Energy and Flux Composition on the Properties of Plasma-
Deposited Amorphous Carbon and Amorphous Hydrogenated Carbon Fihns,”

Diamond Relat. Mater., vol. 2, pp. 246, 1993.

J. Asmussen, “Electron Cyclotron Resonance Microwave Discharges for Etching
and Thin Film Deposition,” in Handbook of Plasma Processing Technology, S.
M. Rossnagel, J. J. Cuomo, and W. D. Westwood, Eds. New Jersey: Noyes

Publications, 1990.

M. A. Lieberman and A. J. Lichtenberg, Principles of Plasma Discharges and

Materials Processing. New York: John Wiley & Sons, Inc., 1994.

B. K. Kim and T. A. Grotjohn, “Hydrogenated Amorphous Carbon Films
Deposited in an ECR-CVD Discharge Reactor using Acetylene,” Diamond Relat.

Mater. , pp. (accepted for publication).

B. K. Kim and T. A. Grotjohn, “Investigation of an ECR Plasma Source used for
the Deposition of Diamond-like Carbon films,” presented at The 23rd IEEE

International Conference on Plasma Science, Boston, MA, 1996.

163

 

[54]

[55]

[56]

[57]

[58]

[59]

B. K Kim and T. A. Grotjohn, “Investigation of the Optical Properties of

Diamond-like Carbon Films,” MRS Proceedings, vol. 498, pp. 109-114, 1997.

B. K. Kim and T. A. Grotjohn, “Characterization of the Optical and Electrical
Properties of Amorphous Carbon Films,” presented at Material Research Society

Fall Meeting, Boston, MA, 1997.

B. K. Kim, “Plasma Discharge Properties of Methane-Argon and Acetylene-
Argon ECR-CVD Discharges,” presented at International Conference of Plasma

Science, Raleigh, NC, 1998.

B. K. Kim and T. A. Grotjohn, “Amorphous Carbon Films Grown from Methane-
Argon and Acetylene-Argon ECR-CVD Discharges,” presented at The 10th
European Conference on Diamond-like Materials, Carbon Nonotubes, Nitrides

and silicon Carbide, Prague, Czech Republic, 1999.

B. K. Kim and T. A. Grotjohn, “Investigation of the Plasma Properties and Fluxes
in a Hydrogenated Amorphous Carbon Deposition Process,” presented at 46th
International Symposium of the American Vacuum Society, Seattle, Washington,

1 999.

B. K. Kim and T. A. Grotjohn, “Comparison of a-CzH Fihns Deposited fi‘om
Methane—Argon and Acetylene-Argon Mixtures by Electron Cyclotron
Resonance-Chemical Vapor Deposition Discharges,” Diamond Relat. Mater.,

2000.

164

 

[60]

[61]

[62]

[63]

[64]

[65]

J. Hopwood, D. K. Reinhard, and J. Asmussen, “Experimental Conditions for
Uniform Anisotropic Etching of Silicon with a Microwave Electron Cyclotron

Resonance Plasma System,” J. Vac. Sci. Technol. B., vol. 6, pp. 1896, 1988.

J. Hopwood, D. K. Reinhard, and J. Asmussen, “Charged Particle Densities and
Energy Distributions in a Multipolar Electron Cyclotron Resonant Plasma Etching

Sources,” J. Vac. Sci. Techno]. A, vol. 8, pp. 3103, 1990.

L. Ley, “Photoemission and Optical Properties,” in The Physics of Hydrogenated
Amorphous Silicon II, J. D. Joannopoulos and G. Lucovsky, Eds. Berlin:

Springer-Verlag, 1984.

N. Mutsukura, S.-i. Inoue, and Y. Machi, “Deposition Mechanism of
Hydrogenated Hard-carbon Films in a CH4 rf Discharge Plasma,” J.Appl. Phys,

vol. 72, pp. 43, 1992.

S. Liu, S. Gangopadhyay, G. Sreenivas, S. S. Ang, and H. A. Naseem, “Infrared
Studies of Hydrogenated Amorphous Carbon (a—CzH) and Its Alloys (a-

C:H,N,F),” Phys. Rev. B, vol. 55, pp. 13020, 1997.

K. Mui, D. K. Basa, F. W. Smith, and R. Corderman, “Optical Constants of Series
of Amorphous Hydrogenated Silicon-Carbon Alloy films: Dependence of Optical
Response on Film Microstructure and Evidence for Homogeneous Chemical

Ordering,” Phys. Rev. B, vol. 35, pp. 8089, 1987.

165

 

 

 

[66]

[67]

[68]

[69]

[70]

[71]

[72]

B. Dischler, A. Bubenzer, and P. Koidl, “Hard Coatings with Low Optical

Absorption,” Appl. Phys. Lett., vol. 42, pp. 636, 1983.

A. Grill and B. S. Meyerson, “Development and Status of Diamondlike Carbon,”
in Synthetic Diamond: Emerging C VD Science and Technology, K. Spear and J.

P. Dismkes, Eds. new York: John Wiley & Sons Inc., 1994.

S. Kaplan, F. Jansen, and M. Machonkin, “Characterization of Amorphous
Carbon-Hydrogen Films by Solid-State Neuclear Magnetic-Resonance,” Appl.

Phys. Lett., vol. 47, pp. 750, 1985.

J. Schwan, S. Ulrich, K. Jung, H. Ehrhardt, R. Samlenski, and R. Brenn,
“Deposition of ta-C:H Films by r.f. Plasma Discharges,” Diamond Relat. Mater.,

vol. 4, pp. 304, 1995.

J. A. Hopwood, “Macroscopic Properties of a Multipolar Electron Cyclotron
resonance Microwave-Cavity Plasma Source for Anisotropic Silicon Etching,” in

ph. D. Dissertation: Michigan State University, 1990.

L. H. Chou, “Hydrogenated Amorphous Carbon Films Prepared by Plasma-

Enhanced Chemical-Vapor Deposition,” J. Appl. Phys, vol. 72, pp. 2027, 1992.

A. Gehan, J. Amaratunga, S. Ravi, and P. Silva, “Influence of dc Bias Voltage on
the Refractive Index and Stress of Carbon-Diamond Films Deposited from a

CH4/Ar rf Plasma,” J. Appl. Phys, vol. 70, pp. 5374, 1991.

166

 

[73]

[741

[75]

[76]

[771

[78]

D. L. Pappas and J. Hopewood, “Deposition of Diamond-like Carbon using a
Planar Radio-Frequency Induction Plasma,” J. Vac. Sci. Technol, vol. A, 12, pp.

1576, 1994.

N. Mutsukura and K Miyatani, “Deposition of Diamond-like Carbon Film in

CH4-He r.f. Plasma,” Diamond Relat. Mater., vol. 4, pp. 342, 1995.

N. Mutsukura and K. Yoshida, “Deposition of DLC Films in CH4/Ar and

CH4/Xe r.f. Plasma,” Diamond Relat. Mater, vol. 5, pp. 919, 1996.

N. Mutsukura and K. Yoshida, “Deposition of DLC Films in CH4/Ne and

CH4/Kr r.f. Plasmas,” Diamond Relat. Mater., vol. 6, pp. 547, 1997.

J. Hopwood and J. Asmussen, “Neutral Gas Temperatures in a Multipolar
Electron-Cyclotron Resonance Plasma,” Appl. Phys. Lett., vol. 58, pp. 2473,

1991.

P. Koidl, C. Wild, R. Rocher, and R. E. Sah, “Amorphous, Hydrogenated Carbon
Fihns and Related Materials: Plasma Deposition and Fihn Properties,” in
Diamond and Diamond-like Films and Coatings, R. E. Clausing, L. L. Horton, J.

C. Angus, and P. Koidl, Eds. New York and London: Plenum Press, 1991.

167

32.5...

 

. Ly 2:...
1. e .

.4521.
P )1...»
P. .. , .

... :5»:

 

 

 

 

.5 .7.
r: . . is

 

91! 2!

fem