Citation
Thin film metalization for micro-bimetallic actuators

Material Information

Title:
Thin film metalization for micro-bimetallic actuators
Creator:
Gorrell, Jonathan Frank, 1958-
Publication Date:
Language:
English
Physical Description:
vii, 114 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Alloys ( jstor )
Aluminum ( jstor )
Aluminum alloys ( jstor )
Atoms ( jstor )
Intermetallics ( jstor )
Mechanical relaxation ( jstor )
Oxides ( jstor )
Stress relaxation ( jstor )
Tensile stress ( jstor )
Thin films ( jstor )
Dissertations, Academic -- Materials Science and Engineering -- UF ( lcsh )
Materials Science and Engineering thesis, Ph.D ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1998.
Bibliography:
Includes bibliographical references (leaves 109-113).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Jonathan Frank Gorrell.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
029233568 ( ALEPH )
39556976 ( OCLC )

Downloads

This item has the following downloads:


Full Text










THIN FILM METALLIZATION
FOR MICRO-BIMETALLIC ACTUATORS















By

JONATHAN FRANK GORRELL


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1998















ACKNOWLEDGMENTS


I would like to express my appreciation to advisor Dr. Paul H. Holloway for his

support, guidance and encouragement. I would also like to give special thanks for pro-

viding a new perspective to some of the problems I encountered in this research. Thanks

are also due to my doctoral committee members Dr. DeHoff, Dr. Singh, Dr. Adair, and

Dr. Kumar for their assistance and interest in my work.

This work would not have been possible without the support of EG&G ICSensors

and the DARPA MEMS research program. I am also grateful to Dr. Hal Jerman for his

efforts in obtaining this grant and his support of my work.

I would also like to thank my friends and coworkers for their support and encour-

agement and who made my return to school most enjoyable. I would like to thank Mark

Davidson for providing me with his insight into the workings of a university and scien-

tific instrumentation.

Lastly, I would like to thank my parents James and Billie Gorrell for instilling a

love of learning and the belief in myself that I can achieve the goals I set for myself from

a very early age. Without these values I would have never tried much less have com-

pleted this program.














TABLE OF CONTENTS

ACKNOWLEDMENTS.................... .............. .......... ......................... i

A B ST R A C T.................................................................. ............... ................... vi

CHAPTERS

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

M otivation and Objective.......................... ................ ..................... 1
Scope of Present W ork.................................................. ..................... 6

2 LITERATURE REVIEW............................. .. ....................... 8

Introduction....................... ............. .... ...................... 8
The Bimetallic Strip..................................................... 8
The Simple Bimetallic Cantilever................................. ............ 9
The Bimetallic Disk..................... .. .................... 10
The Effects of Plastic Deformation........................ ................. 12
Atomic Bonding.................................................................... 13
Plastic Deformation In Metals............................. ...................... 17
C reep in M etals.................................................. ................................. 19
Deformation Mechanism Maps (DMMs)............................................ 22
Annealing ................................................................ ...................... 22
Strengthening of Bulk Metals........................................ 24
Ordered Intermetallic Compounds.................... ...................... 27
Thin Film s............................................. ....................... .................... 28
Environmental Stability of Materials....................... ..................... 35

3 EXPERIMENTAL PROCEDURE................................................... 37

Introduction....................... .. ................. ..................... 37
Thin Film Deposition......................... ........ ........................... 38
Copper Gold Heat Treatment........................ .... ................. 39
Stress and Stress Relaxation Measurements...................... ........... ... 39
Powder X-Ray Diffraction (XRD)........................... .................... 40
Atomic Force Microscopy (AFM)................................................... 42
Scanning Electron Microscopy (SEM)...................... ..... ........... 42

iii








Electron-Probe Microanalysis (EPMA)..................... ...... ........... .. 43
Transmision Electron Microscopy (TEM)................................... ............ 43
E tch ing ................................. ................................. ............................. 44
Auger Electron Spectroscopy (AES).......................... ......... ........... .. 44
Curve Fitting........................................................ 45
Numerical M odeling.................................................... 46

4 R E SU LT S............................................................................ ............. 47

Introduction....................... .................................... 47
T201 Aluminum..................................................... 47
5052 Aluminum............................................... 59
2090 Aluminum............................ ............................ 68
C opper..................... ................................. .................................. 70
Titanium, Manganese and Nickel....................... ........ .......... .... 72
Copper-Gold Intermetallics........................................... 74
Aluminum-Titanium Intermetallics................................................ 77
Modeling the Effects of Oxide Thickness on a Bimetallic
Actuator's Curvature................................................ 78

5 D ISCU SSIO N .......................................... ................... ..................... 81

Introduction................ ........................... .... .................... 81
Thin Film Strengthening............................................. ...................... 81
Solid Solution Strengthening..................... ...................... 82
Precipitation and Multiphase Hardening............... ........ ........... .. 84
Ordered Phases and the Order Disorder Transition................................ 85
Stress Relaxation Mechanism in Thin Films........................................... 86
Mode 2 Stress Relaxation......................... .................... 88
Mode 3 Stress Relaxation.................... ..... .................. 89
High Temperature Application.................... ...................... 90
O xidation ................................................................ ...................... 91
Strength at Elevated Temperatures....................... ................. 91
Figure of M erit................................................................................... 92

6 CONCLUSIONS ..................................................................... 94

7 FUTURE WORK............ ........ .................................. 96

APPENDICES

A DERIVATION OF RELATIONSHIP BETWEEN YOUNG'S
MODULUS AND COEFFICIENT OF THERMAL EXPANSION.... 98



iv









B CURVE FITTING OF STRESS RELAXATION DATA........................ 101

C CALCULATION OF CHANGE IN CURVATURE OF A
BIMETALLIC STRIP BASED ON OXIDE THICKNESS................ 106

D MECHANICAL PROPERTIES OF EXAMINED METALS............... 108

BIBLIOGRAPHY ...................................................................................... 109

BIOGRAPHICAL SKETCH......... .................................................. 113








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

THIN FILM METALLIZATION
FOR MICRO-BIMETALLIC ACTUATORS

By

Jonathan Frank Gorrell

May, 1998

Chairperson: Dr. Paul H. Holloway
Major Department: Materials Science and Engineering



In this study, eleven different thin film metallization systems were evaluated for

use in micro-bimetallic actuators for microelectromechanical structures. These films

were evaporated or sputtered onto silicon wafers. The film stress and stress relaxation

were determined by measuring changes in the wafer curvature. The phases and micro-

structure of these films were evaluated with, scanning electron microscopy, transmission

electron microscopy, Auger electron spectroscopy, electron probe micro-analysis, X-ray

diffraction and line shape analysis, and atomic force microscopy.

Bimetallic actuator may be operated to generate either force or displacement. The

displacement mode is dominated by the coefficient of thermal expansion while the force

mode is a function of both Young's modulus and coefficient of thermal expansion of the

active layer material. In both modes the maximum displacement or force is determined

by the material's yield strength. A figure of merit was developed to aid in material

selection.

The 5052 aluminum alloy films showed that solid solution strengthening can

double the yield strength of a thin film. The T201 aluminum alloy films showed that

precipitates can increase yield strength by 2.5 times. The 2090 alloy film oxidized during








the first heating. Based on isothermal stress relaxation data and changes in the micro-

structure of the 5052 and T201 alloy thin films, two mechanisms involving logarithmic

creep have been postulated to cause stress relaxation. One mechanism is movement of

dislocations in slip systems that terminate at the surface while the other is dislocations

moving in slip systems that terminate at grain boundaries.

Copper gold intermetallics films oxidized and plastically deformed before the

order-disorder transformation occurred, but showed that ordered intermetallics have a

lower stress relaxation rate than the solid solution phase. The Al Ti films showed no

stress relaxation at 4500C, plastically deformed only above 500C, and had limited oxida-

tion up to 800C.
Nickel, copper, titanium, and manganese films all oxidized on their first heating to
3500C. The copper film also oxidized at 500C over 48 hours. Calculations also showed
that the passivation oxide on aluminum alloys can significantly reduce preformance a
bimetallic actuator. Thus oxidation resistance is a significant requirement for materials
for thermal actuation.














CHAPTER 1
INTRODUCTION

Motivation and Objective



The miniaturization of electronics has led to what is now called the information

age. More and faster electronics packaged in ever smaller volumes allow massive

information processing and exchange. Information processing systems are, however,

limited by the devices that allow them to perceive and affect the physical world. Sensors

and actuators have been connected to computers and control systems from the early days

of computing, but these sensors and actuators have been, until recently, large electrome-

chanical devices. Researchers have begun to miniaturize electromechanical devices using

the technologies developed for microelectronics [Hog96]. These new micro sensors and

actuators have created a new class of devices called MEMS, microelectromechanical

systems [Mas95, Fra97, Ang83, Bea96, Mad97]. Integration of MEMS sensors and

actuators is enabling the development of mechanical, chemical and biological "smart

systems" that are able to interact with the physical world.

MEMS have already become a significant industry. Yearly sales are in the billions

of dollars and the market continues to grow rapidly [Mas95]. This market is currently

dominated by pressure sensors and accelerometers. The miniaturization of pressure

sensors has reduced sensor cost to the point that in medical applications the sensors can

be discarded after a single use. Accelerometers are predominately used to determine

when to deploy automobile air bags. Micro actuators have yet made as many inroads into

industry. Currently the biggest application of MEMS actuators is for ink jet print heads

























Figure 1.1: SEM image of an EG&G IC Sensors bimetallic
thermally actuated micro-valve.


[ROW95]. The market of MEMS actuators will continue to develop as actuators are

needed to produce a truly "smart system" that can not only monitor processes but also

control them. However, additional research and development are needed to improve the

performance of micro actuators, and this is the general focus of this work.

Micro actuators have been designed based on a number of physical properties

including electrostatic, electromagnetic, piezoelectric, magnetostriction and thermal

activation bimetallicc, shape memory alloys, and thermopneumatic) [Mad97]. Table 1.1

shows a comparison of the different actuation techniques and their differences in perfor-

mance. While all of these techniques have advantages for particular applications, this

study focuses on the thermally activated bimetallic actuator. Bimetallic actuators are of

interest because they can produce a relatively large force or displacement, the force is

relatively constant throughout the travel of the actuator and bimetallic actuators are easily

manufactured using semiconductor processing technology. Micro bimetallic actuators

have already been employed in commercially available products [Mad97, Jer94]. Figure

1.1 shows a bimetallic actuated micro valve produced by EG&G ICSensors produced by









Table 1.1: Comparison of actuation principles [Bel97].


Effect

Par r B c Electro- Piezo- Electro- Thermo-
Parameter Bimetallic
static electric magnetic hydraulic
Fore pr High(O.I 1) Iow(0'. 10') High High High
I rotor, N

Deflection, mm High OOl I) Low (10'. 10') LUw' High High

Work per cycle per L () gh H
l m--2, J High (10') Low (10 ) High High
I mm'. J
Frequency range, Low 01-100 (Mech Resonal (Mcch Resonat L.w
Khz (limiting factorr)(Heat Tanfer frequ y) higheqey) Hea Tranfer

\bltage, V Law (3 12) High (100 300) High Low Low

Current, mA High (0.1 10) Lw (10 10') Low High High

Power Consunption. H,
Power Consumption High(10' 0.1) Low (10 W. 1-) Law High High
w

Efficiency, % 0.01 0.5 000l o.01

Size: Length, mm Semll (1. 1)
Width, mm Sal ( 0.1) Sl Lage Large Lrge

rAddeiional s lih lVtage (Mageti field Media separaton
reqauremtnts Co.nlbhaly)

Cost Low Low High High High





aluminum metallization on silicon. Even though silicon is not a true metal, this structure

is still referred to as a bimetallic actuator.

The objective of this research is to improve the performance of micro-bimetallic

actuators and broaden their range of operating parameters (stress, displacement, tempera-

ture range and stress relaxation) by identifying better or improve the materials of con-

struction. In order to understand how to improve the performance of the bimetallic

actuator, it is important to understand how the device works. A bimetallic actuator is

made by bonding together two different materials that have a large difference in thermal

expansion, such as aluminum and silicon, to create a bimetallic strip or spring that is

heated or cooled. Upon heating, the material with the larger coefficient of thermal expan-








Aluminum Active Layer


Silicon


Passive Layer


Aluminum


Silicon


200C










1000C


Silicon Expansion
Aluminum Expansion


Figure 1.2: Diagram of a bimetallic strip, showing how the differencein
thermal expansion bewteen the active layer, aluminum, and
passive layer, silicon, causes the bimetallic strip to bend.



sion, called the active layer, expands more than the material with the smaller coefficient,
called the passive layer (see Figure 1.2). As these two materials are bonded together, a
shear stress develops between the layers due to the different rates of expansion, producing
a force that causes the structure to bend. Upon cooling the reverse processes occurs and
the bimetallic strip returns to its original position, or deflects in the opposite direction
with continued cooling. The geometry may be a disk, cantilevered beam, or more com-
plex geometry.
While the current generation of bimetallic actuators works well for a number of
applications, their use could be significantly expanded by improving the position and
force stability of these actuators and increasing the operable ambient temperature limit
from 75C to 500C. Force and position stability have been found to be a problem in


... :.: ..L-
* I i;r: ..; .

:' .:''5;H '.


- ;,,,.-.;









Aluminum Valve Closed
Metallizatlon Inlet
Heated
\ i,3: Dkiphragm


Valve Body \

Outlet

Cooling or
Heating Stress Relaxation
of Metal
Aluminum Valve Open
Metallization Inlet

\ _\ 4/, -* _


Valve Body

Outlet


Figure 1.3: Cross Sectional Diagram of bimetallically actuated micro-valve.




some bimetallic actuators such as the one shown in figure 1.1. It is difficult to maintain a

constant flow from these valves over extended periods of time. The opening of the valve

decreases with time. Force and position instability in bimetallic actuators are caused by a

relaxation in the shear stresses between the active and passive layers of the bimetallic

strip which in turn reduces the force bending the structure. Figure 1.3 shows a cross

sectional diagram of a bimetallic actuated micro-valve and the effects of stress relaxation

on the valve operation. The inability to maintain a constant force or position for extend

periods of time complicates the usefulness of the device. These limitations result prima-

rily from creep in the active metal layer of the bimetallic actuators. Thus, improvements

can be achieved through the identification and or development of better metal thin films.








While improvement of materials for bimetallic actuators may at first seem to be a

narrow subject, it requires examining much broader materials questions. How can a thin

metal film be strengthened? What are the atomic mechanisms responsible for the stress

relaxation in thin films? What materials are sufficiently chemically stable at elevated

temperatures to be useful? These questions are of general interest to the entire MEMS

field because micro-machines use thin metal films as structural components, and struc-

tural failure and environmental stability are concerns in any mechanical device. Also, the

continued reduction of interconnect widths in microelectronic integrated circuits lead to

greater stresses in thin metal lines that may fail due to stress induced voiding (SIV)

[Bow74]. Therefore, this research will examine questions of interest to many disciplines

The temperature change of an actuator is normally accomplished by resistive

heating. The passive layer of most micro-bimetallic actuators is silicon, and a resistive

heater is easily made by doping the silicon. A thin silicon oxide layer grown on the

silicon is all that is needed to keep the resistive heater from shorting to the active layer.

Cooling is accomplished by conduction of heat to the valve body and the surrounding

structure. The thermal masses of micro-bimetallic actuators are so small that they can be

cycled rapidly. Other heating and cooling methods could be used, such as a Peltier

element [Cad60]. In the present work, the mechanisms to change the temperature of the

bimetallic actuator is not of direct interest. Instead the present research is concerned with

the mechanical and thermal properties of the materials in the bimetal strip, and how they

affect actuation. Optimization of the performance of bimetallic strips will be the focus,

which directly translates into optimizing the performance of a bimetallic actuator.

Scope of Present Work



The three specific objectives of the present work are as follows. (i) The first

objective is to determine the effectiveness of different strengthening techniques for thin








film metallization. There are a number of techniques used to increase the yield and

ultimate tensile strength of bulk metal alloys, but their effectiveness in thin films has yet

to be determined. Increased yield strengths of the metal thin films would increase the

maximum force and displacement produced by an actuator, and could increase force and

displacement stability. (ii) The second objective is to evaluate use of materials at el-

evated temperatures. The primary requirements for elevated temperature applications are

high strength over their operation temperature range and resistance to creep and oxida-

tion. (iii) The third objective is to identify the mechanism responsible for isothermal

stress relaxation in thin films. It is this relaxation that causes the force and position

instability seen over extended time periods.

To meet these objectives, eleven different materials were deposited and evaluated.

The evaluation focused on determining the material's strength, environmental stability

and rate of stress relaxation. For those materials that showed promise and to understand

the underlying mechanisms responsible for the stress relaxation, the morphology and

microstructure of several samples were examined. The materials that were evaluated fall

into three groups: (i) aluminum alloys, (ii) elemental metals, and (iii) intermetallics. The

aluminum alloys were used to identify the mechanisms responsible for isothermal stress

relaxation, and to determine the effectiveness of the different strengthening techniques.

The elemental metals and intermetallics were evaluated at elevated temperatures. Some

of the intermetallics also exhibited ordering and non-linear thermal properties. The

evaluation of several materials systems allowed a wide variation in properties and there-

fore provided significant insight into the materials requirements for MEMS devices.

The organization of this dissertation is as follows. Chapter 2 is a review of the literature

pertinent to this study. The experimental procedures used to deposit, test, and analyze the

samples are presented in chapter 3. Chapter 4 contains the results of these tests and

analyses. The results are discussed in chapter 5. The conclusions are summarized in

chapter 6 and suggestions for future work are presented in chapter 7.















CHAPTER 2
LITERATURE REVIEW

Introduction



Thin film metallization for bimetallic actuators are the focus of this research,

therefore a number of topics need to be reviewed. The first topic is the theory of the

bimetallic strip. In examining literature, the concern is to determine how the different

materials properties affect the performance of a bimetallic strip. With an understanding

of the effects of materials properties, improved materials selection can be used to opti-

mize performance. After their identification, the origin and interdependence of these

properties can be discussed using the theories of atomic bonding and thermal expansion.

In addition, there are a number of processes that may affect the metallization of a bimetal-

lic strip, such as creep, plastic deformation, and oxidation. Lastly, these bimetallic

actuators are manufactured with thin film technology and the effects of these processes

are discussed. There is no perfect material for use in all bimetallic actuators, since

different applications require different properties. With an understanding of the interde-

pendence of the different materials properties, the tradeoffs that are needed for any given

application can be better selected.

The Bimetallic Strip


The formulation of the bimetallic strip problem was first reported, in part, by G.

Gerald Stoney [Sto09]. Stoney determined the relationship between curvature and stress

in a bimetallic strip. S. Timoshenko [Tim25] then derived the equations for displacement








and force of a bimetallic strip versus change in temperature. Lastly, Townsend [Tow87]

generalized the equations for a multi-layer structure.

The formulation and solution of the bimetallic strip problem are dependent on the

geometry and boundary conditions used, therefore no single solution is generally appli-

cable. Roark's Formulasfor Stress and Strain [You89] provide solutions for many shapes

and boundary conditions. This complexity of solutions and lack of solutions for unusual

shapes has led to the widespread use of finite element analysis to design bimetallic

actuators [Bar94] [Tsa92].

The Simple Bimetallic Cantilever


To determine the effects of materials properties on performance, the simple

bimetallic cantilever, simply supported on only one end, is examined. The simplifying

assumptions made in this case are that the length is much greater than the width or thick-

ness, the materials are perfectly elastic, and the material returns to its original dimension

once the load is removed [Mey84]. Figure 2.1 shows a diagram of this case where E, and

E2 are the Young's moduli, L is the length of the strip, h is the thickness of the strip, a and

b are the thickness of the first and second layer, w is the width of the strip, a, and ca are

the coefficients of thermal expansion, and the curvature of the bimetallic cantilever, k,

for a given change in temperature, AT, is[Tim25]



6(a2 -a )(AT)(l+ )2
k= b
h[3(l+b))2 +( 2 bE2 (1)



For the case of both layers of the strip being the same thickness a = b and equal moduli

the equation for curvature reduces to [Tim25]












Layer1a Loy-w
G F Iaver 1 a
LayrEer 2 h
LoVer 2 b
--- L >


Figure 2.1: Diagram of bimetallic strip



3 (a,- 2)
k = AT. (2)
2 h


The error induced by assuming the moduli are equal is only three percent for a factor of

two change in either material's modulus. Therefore, the primary material's property

affecting the displacement in this case is the difference in the coefficients of thermal

expansion.

The equation for the restraining force, R, that is needed to keep a bimetallic strip

from moving for a given temperature change DT is the maximum force a bimetallic strip

can produce. This restraining force is given by equation 3:



(E + E2) h (3)
R= h 2. ( -a,)AT.
4 h-L


Thus the force produced by a bimetallic strip is a function of both the thermal expansion

and the Young's moduli of the materials.

The Bimetallic Disk



To evaluate the effects of materials properties on performance for a more complex

case, the bimetallic disk is examined. Due to the complexity of the equation describing












; Ni
L 2-
x
\ CA 172
Cu-Be

o 1
--B

CA T201
1 5052
c Contour of
Constant Force Contour of / SMg
Constant Deflection
0
0 10 20 30
Expansion Coefficient (XE-6/OC)

Figure 2.2: Equal potential lines for force and displacement for a bimetallic disk [Jer95].


the bimetallic disk the evaluation was performed by calculating and plotting the
equalpotential lines for maximum displacement (curvature) and force of a bimetallic disk
as a function of Young's moduli and the coefficient of thermal expansion, as shown in
Figure 2.2 [Jer95]. This model is based on a bimetallic strip where the passive layer is
silicon and both the active and passive layers are 5ltm thick. In agreement with equations
(2) and (3) this figure shows that displacement is predominately limited by the thermal
expansion of the active layer material. There is some dependence on the Young's modu-
lus but this occurs only for a modulus lower than that found in most engineering materi-
als. The equal potential line for force shows that thermal expansion is not as significant a
factor as in displacement. Therefore, while the relationships shown in Figure 2.2 for a
bimetallic disk are far more complex than the relationships determined by equations 2








and 3 for a single fixed cantilever, the trends and limiting materials properties are the

same. Displacement in a bimetallic structure is primarily limited by the thermal expan-

sion of the active layer material. In contrast, the force generated by a bimetallic strip is

dependent on both the thermal expansion and Young's modulus of the active layer mate-

rial. These trends and limitations are expected to hold for other configurations as well.

The Effects of Plastic Deformation


The assumption that a material performance is perfectly elastic is only valid for a

range of applied stresses. Beyond this range of stress the material will not return to its

original shape when the stress is removed and the material is said to have undergone

plastic deformation [Mey84](see Figure 2.3). Plastic deformation in either layer of a

bimetallic device invalidates the relationships in equations (1), (2) and (3), and would

reduce the force or displacement produced from the strip. The effects of plastic deforma-



Ultimate Tensile Strength, UTS

Offset Yield Strength,
Yield Strength, YS B
u\/ Breaking Point



CY



Young's Modulus, E, Stress/Strain


Strain, AL/L
Figure 2.3. Typical stress strain curve for a ductile metal.









tion can be modeled as a non-linear reduction of the thermal expansion or Young's modu

lus of a material. While not considered in the above equations that predict the force or

displacement of a bimetallic strip, the stress at which plastic deformation begins, called

the yield strength [Ask89, Mey84] (see Figure 2.3) is also a critical materials parameter.

A material's yield strength determines the maximum force it can exert on another layer

and in turn may determine the maximum force or displacement a bimetallic strip can

achieve.

Therefore, the materials properties of interest depend upon whether the applica-

tion requires force or displacement. In the displacement mode, the material's thermal

expansion and yield strength are the primary limiting properties. In the force mode,

thermal expansion and Young's modulus both contribute to the force generated by a

bimetallic strip, and yield strength is a limiting materials property.

Atomic Bonding



The atomic bonds that hold a material together also determine its thermal expan-

sion and Young's modulus. There are a number of aspects of atomic bonding that affect

these physical properties, and these aspects will be examined in this section. This review

is not intended to cover all aspects of thermal expansion, but only those that affect the

performance of a material in a bimetallic actuator.

The first aspect of atomic bonding that determines a material's rate of thermal

expansion and its Young's moduli is the strength of the atomic bonds. Figure 2.4 shows

the asymmetric potential energy well that an atom is drawn into when it bonds with

another atom [Ask89]. The depth of this potential well determines the strength of the

bond. The deeper the potential well, the stronger the bond. A deep potential well nor-

mally has a larger dE/dx which is proportional to the Young's modulus.

The rate of thermal expansion of a material is due to the asymmetry of the poten-

tial energy well. The addition of energy to a material causes an increase in vibration of














Stime-averaged position of atom

Distance


5 Increase in interatomic distance,








Figure 2.4: Variation in potential energy with distance between atoms.


the atoms. This increase in atomic vibration causes the atoms to move up in the potential

well. As the atoms move higher their time-averaged position increases (see Figure 2.4)

and the material expands [Tou75]. The deeper the potential well, the smaller the change

in the time-averaged position of the atoms for a given increase in temperature, and a
lower thermal expansion is observed. The rate of thermal expansion is frequently ex-

pressed as the coefficient of thermal expansion, CTE. A material's CTE is the derivative

of its thermal expansion curve, AL/L versus temperature. To a first approximation, the
thermal expansion is inversely proportional to the atomic bond strength. Thus, the
Young's modulus and thermal expansion of a material are inversely proportional. Appen-

dix A shows the thermodynamic proof of this relationship.

The second aspect of atomic bonding that affects the rate of thermal expansion
and the Young's Modulus is the directionality of the bonding. The directionality of the

bonding results from the type of bonds that form between atoms, the crystallography of








the material, the number of nearest neighbors, the coordination number, and the bond

length. Graphite is a good example to illustrate the effects of directionality of bonding.

Graphite has a layered structure where the bonds within a layer are strong covalent bonds,

but the bonds between atoms in different layers are much weaker van der Waals bonds

[Eva79]. This difference in bonding causes the c-plane coefficient of thermal expansion

(CTE) to be 2.8xl0-60C- while between the planes the CTE is 4.4x 106OC'.

In many instances, the directionality of bonding does not change the bond type but

still has a major affect on physical properties. Zinc illustrates this point well. The atoms

in zinc are bonded into a hexagonal close packed crystal structure, with metallic bonds

between all atoms. Along the c-axis the CTE for zinc is 5x10-60C, but along the a-axis

it is 65xl0-60C-1 [Tou75]. Thus there is a 13-fold difference in CTE between different

planes due to the crystallography that is determined by the bonding.

There are other materials with small anisotropy but unexpected physical proper-

ties due to unusual crystallographic structure. Manganese is one such metal [Dea52].

While o-manganese has a cubic structure, it does not have a simple face centered cubic

(FCC) or body centered cubic (BCC) crystal structure [Ask89]. The unit cell is based on

the BCC structure but contains 58 atoms in 29 pairs [Dea52]. The CTE for manganese is

22x 106 C-', almost double what would be expected for a material with a melting tem-

perature of 12440C. In comparison, beryllium melts at 12780C and has a CTE of only

12x10-6 'C- [Wea76].

The directionality and crystallographic structure of a material are not always

static. Pure iron undergoes two allotropic transformations between solidification at

15380C and room temperature. Iron first solidifies into a BCC structure at 15380C,

converts to an FCC structure at 13940C, and then converts back to a BCC structure at

9120C [Kra90]. At each allotropic transformation, there is a significant change in thermal

expansion and other physical properties. Discontinuities in a material's CTE could be

useful in a bimetallic actuator. Unfortunately the transitions in iron are far above the










2,0-






0.8--
03
0 1.6
c







0 ,




I-


04L-
-2 0 20 0 40 6W0
TEMPERATURE C

Figure 2.5: Linear thermal expansion for intermetallic compound CuAu and
Cu1Au. The discontinuity in thermal expansion is due to the order-disorder
transion of these compounds [Tou75].



temperature currently being considered for practical devices. However, transformations

are used in shape memory alloy actuators [Mad97].

Another change in bonding that can significantly affect properties is the order-

disorder transformation exhibited by some intermetallic compounds [Tou75]. Intermetal-

lic compounds are made up of two or more elements, A and B. In ordered alloys, the A-B

bond may be stronger or weaker than the A-A or B-B bonds and thus there is a driving

force to form one type of bond over the other [Cah83]. Ordered intermetallic compound

may form directly upon solidification, as in Al Ti and NiAl. However, there are other

intermetallic compounds that first solidify as a solid solution and then order at a lower

temperature. CdMg3, CuAu CuAu, and Cu3Au are all examples of materials that

undergo a solid phase disorder/order transition [Hov64, Fed58, Cul78, War69]. The


CuAu





Cu Au


I /








material may undergo a substantial contraction upon ordering (see Figure 2.5). This

contraction is due to the ability of the atoms to pack tighter in the ordered phases due to

the stronger atomic bonding. This transition may also increase the material's yield

strength, reduce the creep rate, and increase electrical conductivity.

Plastic Deformation In Metals


As mentioned above, plastic deformation in either layer of a bimetallic actuator

will limit the force and displacement of the actuator. Therefore, the yield strength deter-

mines the maximum force and displacement of an actuator. Atomic motion becomes

rapid as the temperature of a metal exceeds half its melting point, therefore the homolo-

gous temperature, TH, is defined to be the ratio of the temperature to the melting point of

the material in degrees Kelvin. At TH < 0.5 dislocation motion is the dominant mecha-

nism of plastic deformation. Inhibition of the motion of dislocations would strengthen

the material. This is in contrast to the thermal expansion and Young's modulus of a

material that can not be changed by inhibiting dislocation motion. In addition, many of

the methods used to inhibit plastic deformation do not significantly alter the CTE or

Young's modulus of the material. Therefore, improved yield strength can be used to

produce a better actuator. In this section, the atomic processes responsible for plastic

deformation will be examined, along with methods to inhibit these operations.

The two main processes that take place in a material during plastic deformation

are: dislocation movement and multiplication. Movement of dislocations produces a

change in shape of the material. However, the dislocation density in an annealed sample

is not great enough to produce the observed plastic deformation. Thus the dislocation

density is increased during plastic deformation by dislocation multiplication [Ree92,

Cah83].

With respect to dislocation motion, they can move through a crystal by either

dislocation glide or climb. Dislocation glide occurs when a dislocation moves on a slip









plane and in a slip direction of the crystal [Ree92]. The combination of a slip direction

lying along a slip plane creates a slip system. There are a limited number of slip systems

and they are determined by the crystallography of the material. Dislocation may not be

able to glide in response to an applied force because a slip system is blocked. On the

other hand, the stress resolved onto that slip plane in the slip direction may be below the

critical value to cause motion of dislocation. This is known as Schmid's law [Ask89].

The second way for a dislocation to move is by climb, in which it moves between parallel

slip planes. Climb allows dislocations to move off blocked slip planes. It occurs by

vacancy movement [Ree92], but both the vacancy concentration and movement are very

dependent on the temperature of the material [Deh93]. Dislocation climb only becomes a

dominant process at high homologous temperature T,. For most metals dislocation climb

becomes significant at T,>0.5 and for most intermetallics at T,>0.7. Above these tem-

peratures dislocations may rapidly climb.

The process of dislocation multiplication can occur in many ways. One of the

more significant is the Frank-Read source [Ree92] by which dislocation glide can in-

crease the dislocation density. In a Frank-Read source, a dislocation line on a slip plane

is pinned or blocked at two points. In response to an applied resolved shear stress the

dislocation bows out between these two points and forms an incomplete dislocation loop.

A critical stress exists, dependent on the dislocation line length, above which complete

dislocation loops form. This mechanism is complicated by the fact that dislocations can

pin one another, thus forming more Frank-Read sources. This is one of the reasons that

metals harden when they are cold worked, which is known as "work hardening" [Bro82].


There generally exists an inverse relationship between temperature and yield

stress independent of the mechanisms responsible for the deformation. For dislocation

glide, random thermal excitation helps dislocation overcome the Peierls-Nabarro barrier

[Mey84] and thus lowers the yield stress. The effects of temperature on dislocation








density and mobility have already been discussed and these effects also lead to reduced

yield stress with increased temperature. While an inverse relationship exists between

yield strength and temperature it is a non-linear function specific to each material.

The stress at which plastic deformation begins is also dependent on whether the

force is applied in tension or compression. There is no general relationship between the

compressive and tensile yield strength and most yield strength data for metals are for

tensile yield strength. As the thin films in a bimetallic actuator can be in tension or

compression the variation in the compressive versus tensile yield stresses could be sig-

nificant.

Lastly, the crystallographic texture (which is the preference for one crystallo-

graphic orientation) of a material also affects its yield strength [Mey84, Bro82]. Baldwin

[Bal46] showed up to 30% change in tensile yield strength based on the direction of the

applied stress relative to the texture of rolled copper sheet.

Creep in Metals


In the preceding section the mechanisms responsible for the rapid plastic deforma-

tion of materials in response to an applied stress were examined. However, materials

continue to plastically deform in response to applied stresses over extended periods of

time (from minutes to years). Time dependent plastic deformation is known as creep. In

this section the different types of creep and the mechanisms responsible for the time

dependent plastic deformation will be discussed.

Figure 2.6 [Cah83] shows the four major types of creep that occur in materials as

a function of the homologous temperature versus the resolved stress, o, normalized by the

shear modulus, i.

At stress levels below which dislocation can move (ao /g < 10-8) anelastic creep

occurs. In anelastic creep, interstitial atoms move to interstitial sites that have been

elongated by the applied stress [Cah83]. This movement of interstitial atoms creates a










T/T
0.5 1.0 Theoretical
10 \ Strength
10 HighTemperature St
Creep (approximate)
10 \ (Andrade Creep)


10 Herring
(/., Nabarro-Coble
Low Temperature Creep
10 Creep
(logarithmic Creep)
10 (,,,/P)
Anelastic Creep
(Recoverable Creep)
10




Figure 2.6: The creep diagram defining the conditions of temperature and stress which
produce the four principle types of creep. The temperature is plotted as the homologous
temperature and the stress is normalized by the shear modulus. a = applied resolved
shear stress, p = shear modulus, and oas. = critical resolved shear stress of a well
annealed crystal [Cah83].



small deformation of the material that is not permanent as the interstitial atoms will

randomly redistribute once the applied force is removed. Since the materials examined in

this study do not contain a significant percentage of interstitial atoms, this creep mecha-

nism is of little concern.

Herring-Nabarro-Coble creep [Her50, Cob63] produces a deformation in the

material by a net mass diffusion. In the Herring-Nabarro model, a net flux of vacancies

move away from the axis of the applied stress produces a net flux of atoms to the axis of

applied tensile stress, thus elongating the sample in the direction of the applied force.

This mechanism of creep requires a high concentration of vacancies with high mobility,

and only occurs at T. > 0.9 [Cah83]. Coble expanded this idea of bulk diffusion to

include the diffusion of atoms along grain boundaries. Grain boundary diffusion has a

lower activation energy than bulk diffusion because a grain boundary is an array of









dislocations. This lower activation energy for diffusion slightly reduces the temperature

at which the creep rate is significant. Coble creep still requires a T > 0.85 [Cah83] to

become active and is therefore not of concern in this study.

The third type of creep is low temperature or logarithmic creep. In this mecha-

nism, deformation occurs through dislocation multiplication, glide and climb. However,

the sources for dislocation multiplication become exhausted. Therefore the creep rate

starts at some initial higher valve and logarithmically approaches zero as the number of

mobile dislocations approaches zero [Cah83]. The rate of logarithmic creep is not a

function of the applied stress, since the rate limiting steps are dislocation climb or multi-

plication which are driven by random thermal excitations. This type of creep is typically

observed at low temperatures (T, < 0.5).

The fourth creep mechanism shown in figure 2.6 is high temperature or Andrade

creep [Cah83]. In this mechanism dislocation glide, climb and multiply in response to

the applied force with the help of thermal excitation to overcome the higher activation

energies. There are three stages to Andrade creep. In stage one, a stress applied to the

sample at a given temperature causes an initial high strain rate that immediately begins to

decline. The reason for the declining strain rate is that the sample begins to work harden.

After a finite strain, equilibrium is established between the rate of work hardening and the

rate of dynamic annealing. Annealing is a heat treatment that eliminates the effects of

cold work, in dynamic annealing this processes is occurring at the deformation tempera-

ture as described below. Second stage creep is a steady state process that produces a

constant strain rate. In the third stage of Andrade creep the sample begins to neck, grain

boundary sliding occurs [Mey84] and voids form inside the sample. This reduces the area

over which the force is carried thus increasing the stress which increases the strain rate,

ultimately resulting in failure. This type of creep is typically observed for T > 0.6.








Deformation Mechanism Maps (DMMs)


In the proceeding sections, several mechanisms have been discussed which lead to

plastic deformation of metals. H.J. Frost [Fro82] has created deformation mechanism

maps (DMM) relating stress, temperature and deformation mechanism. All DMMs are

based on numerical models describing the different deformation mechanisms. This

model is then used to generate a two dimensional contour plots (temperature and stress

axis) that shows which deformation mechanisms are active. These plots are useful in

selecting materials for a given application. If the stress and temperature levels for a given

application are known, then the deformation mechanisms expected for a given application

can be projected. The primary limitation to the use of DMMs is that detailed data re-

quired by the model are not available for all materials and conditions.

DMMs have been used to explain deformation and stress relaxation in thin film

metallization [Kol86, Fro92, Tho93, She96]. These models have been able to accurately

predict plastic deformation in thin films at high strain rates, but none of these DMMs

were able to accurately predict the rate of isothermal stress relaxation. However, only

bulk material properties are currently available to use in DMMs.

Annealin2

In the preceding sections the mechanisms responsible for plastic deformation at

high and low strain rates were discussed. However, since stress relaxation is of concern

in this study, other mechanisms that transform metals must be considered, such as anneal-

ing.

Annealing is the process by which the stored energy in cold worked metal is

released. When a metal is cold worked, that is plastically deformed below the tempera-

ture at which dislocation climb becomes significant, part of the deformation energy is

stored in the metal as defects and lattice distortion [Ree92]. Hundreds of Joules per mole

can be stored in a metal that has been heavily cold worked. The annealing process re-


























(a) (b)
Figure 2.7: Distribution of dislocations in a bent crystal (a) before polygonization, (b)
after polygonization and subgrain coalescence.


leases this stored energy by reducing the defects, lattice distortion and residual stress. In

the case of a thin metal film deposited on a substrate, the shear stresses between the film

and the substrate may also distort the lattice and induce defects. The process of annealing

should result therefore in stress relaxation.

There are four stages in annealing: recovery, recrystallization, grain growth, and

secondary grain growth [Ree92]. For this study the first three stages are of most interest

since they dominate the release of stored energy from cold work.

In recovery there is little change in the mechanical properties of metal, but the

electrical resistance of the metal decreases. This reduced electrical resistance indicates

that random dislocation tangles have begun to polygonize [Cah49] (see Figure 2.7). The

polygonization of the dislocation tangle into subgrain boundaries reduces the free energy

of the system. Ordering also reduces the number of electron scattering centers, resulting

in a lower electrical resistivity. No dislocations are destroyed, therefore the yield stress

does not change.







24
Polygonization and subgrain coalescence also lead to the second stage of anneal-

ing, recrystallization. In recrystallization new unstressed crystals nucleate and grow to

replace the strained, disordered cold worked grains. One theory on the formation/nucle-

ation of these new unstrained crystals is that subgrains coalesce to form high angle grain

boundaries leaving an unstrained crystal in their wake [Cah83]. The difference in ener-

gies provides the driving force for the high angle grain boundary to move outward,

consuming the strained crystals until they are all consumed. This is the growth stage of

the annealing process and it is at this point that the energy of cold work has been removed

and therefore stress has been relaxed.

Grain growth also affects the size and distribution of precipitates within a sample

[Bro82]. Precipitates strain the matrix of the material in which they exist. There is

therefore a driving force for precipitates to coalesce and reduce the strain in the matrix.

In age hardened alloys this is called "over aging" as it reduces the yield strength of the

material. This process would also produce a relaxation of stress.

Strengthening of Bulk Metals


In the preceding sections', mechanisms responsible for stress relaxation in thin

metal films have been reviewed. We will now review the methods used to strengthen

bulk metals to determine which could be used to strengthen thin metal films in bimetallic

actuators. There is limited information on strengthening of thin films.

The key to strengthening any non-brittle metal is to inhibit dislocation glide. This

can be done in several ways. First, the matrix of the metal can be affected to increase the

force needed to move a dislocation in a slip system. This can be done a number of ways,

including straining the matrix with cold work or solid solutions. Second, hard particles or

phases can be placed in the matrix of a material that will physically block the movement

of dislocations. There are five commonly identified strengthening mechanisms that use









one or both of these general approaches. Also, many commercial alloys use two or more

alloying elements and several of the five strengthening mechanisms to ensure that several

means of inhibiting dislocation motion are used. The five strengthening methods used in

bulk alloys and the means by which they inhibit dislocation motion are described below.

The first strengthening method is solid solution strengthening, in which soluble

alloying elements are added to the host metal [Cah83]. Since most alloying elements

have an atomic size different from the host element, the matrix of the metal is strained,

and inhibits dislocation glide. The more the host matrix is strained, the harder it is for

dislocations to move. However, alloying elements generally have limited solubility in

host metals which limits the strengthening that can be achieved. For most of the com-

monly used engineering metals, the solubility and strengthening effects of many alloying

elements have been determined experimentally. The Metals Handbooks Volumes 1 and 2

[Bak 97] contain most of this information. For more unusual materials, the Hume-

Rothery [Ree92] rules for substitutional solid solution strengthening are helpful.

Solid solution strengthening has a number of properties that make it unique

compared with the other strengthening mechanisms. First, the solute atoms are in equi-

librium with the host matrix atoms. Thus there are no driving forces to change the atom

distribution in the metal and solid solution strengthened alloys are expected to be more

stable at elevated temperatures than alloys using some of the other strengthening tech-

niques. This is of particular importance for this study since materials will be used at

elevated temperatures. Secondly, a solid solution strengthened material is expected to be

homogenous. Thus there is limited possibility that electrolytic half cells would cause

corrosion.

The second strengthening mechanism is to reduce grain size. Dislocations gener-

ally pileup at the grain boundaries as they cannot cross them. The strength of a material

has been found to be proportional to the inverse of the square root of the average grain








size [Mey84, Ree92]. This relationship is called the Hall-Petch relationship. A potential

limitation to using this strengthening technique is that grains coarsen at elevated tempera-

tures (T, > 0.6), and this generally reduces the strength of the material.

The third strengthening techniques is work hardening. When a metal is cold

worked (i.e. worked below a temperature at which annealing occurs) the dislocation

density increases greatly. With an increased density, dislocation will be pinned by other

dislocations, which limits dislocation slip. Work hardening can produce seven fold

increase in yield strength for pure aluminum. Fully annealed aluminum has a yield

strength of 15-20 MPa that increases to 50-60 MPa for 40% cold work and 100-120 MPa

for 90% cold work [Bak97]. A concern with work hardening is that the material will

soften as dislocations climb at elevated temperatures during recrystallization.

The fourth way to strengthen some materials are by a martensitic transformation

[Kra90]. This type of transformation occurs only in a few alloys and ceramics. It is

widely known and used in ferrous alloys, but does occur in some copper, aluminum and

titanium alloys. The martensitic transformation is a diffusionless non-equilibrium pro-

cess that occurs by a coordinated shear displacement of atoms. This transformation

stresses the matrix of the material which inhibits dislocation motion. The use of marten-

sitic transformations to strengthen alloys is limited because there are few non-ferrous

alloys which exhibit them [Pet70]. In addition, a martensitic transformation leads to a

non-equilibrium state so there is always a driving force to transform to a lower strength

phase.

The fifth way to strength a metal is by precipitation or multi-phase hardening.

Alloying elements are added above their solubility limit in a host and precipitates form

out of the solid solution to block dislocation motion [Bro82]. In certain cases, a fine

ceramic powder is added to a molten metal in place of an alloying element to provide the

particles in a modification called dispersion hardening. This leads to improved thermal

stability of the alloy. For the greatest precipitation hardening, a large number of small







27
precipitates should be uniformly distributed throughout the host matrix. Thermal stability

is a concern at elevated temperatures since precipitates may coarsen (over age) resulting

in a weakened material.

The presence of a second phase can also strengthen a host matrix by straining it.

A second phase with a coherent interface to the host matrix will increase the strain in the

matrix. A coherent interface does not an array of dislocation between the precipitate and

the matrix [Bro90]. Thus, in addition to blocking dislocation motion, a coherent precipi-

tate strains the matrix, increasing the force needed to move a dislocation. Thermal

stability is a concern, as precipitates may grow by thermally activated diffusion and the

coherent interface may be lost. The loss of the coherent interface substantially reduces

the strength of the alloy.

Ordered Intermetallic Compounds


Ordered intermetallic compounds [Cah85] are of interest in this study because

they are noted for high strength at both room and elevated temperatures with good resis-

tance to creep and oxidation [Wes95]. There are several reports on the use of intermetal-

lics at high temperatures in jet engines [Lim97]. The basic structure of intermetallics was

discussed above in the section on thermal expansion. The mechanical properties are

discussed below.

Intermetallics are a class of materials with structures and properties between pure

metals and ceramics [Pop87]. The general composition on an intermetallic is A B where

A and B are metallic elements and x and y are normally small integers. This more com-

plex composition creates a more complex crystallographic structure which in turn make

dislocation movement more difficult. There are several reasons for this. First, the num-

ber of atomic distances a dislocation has to move to reach an equivalent crystallographic

site is much greater. In a simple AB intermetallic a dislocation has to move two atomic

distances as compared with one atomic distance for a non-ordered metal. In many







28
instances a dislocation will break into pairs of partial dislocations which only move one

atomic distance forming an anti-phase boundary [Bro82]. The creation of an anti-phase

boundary requires energy and increases the critical resolved shear stress for dislocation

glide. In a normal metal a dislocation must break A-A bonds and then reform them to

move. In an intermetallic a dislocation must break the stronger A-B bond to move. Once

the A-B bond is broken the weaker A-A and B-B bonds form (the anti-phase boundary),

which then have to be broken to form the new A-B bond. This process requires more

energy, and thus intermetallics have high yield strengths.

The more complex structure also inhibits creep. The self diffusion rate in inter-

metallics is unusually low [Pop87] due to the large distance between equivalent site, and

a high activation energy for this process. As the creep rate for a material is proportional

to the rate of self diffusion [Cah83], intermetallics have a low creep rate.

Some intermetallics are also resistant to oxidation. Oxidation resistant intermetal-

lics form an adherent passivating oxide such as A12O, on compounds of aluminum,

nickel and titanium.

A concern in using intermetallics in bimetallic actuators is their sensitivity to

composition fluctuations. In A1Ni, as little as 0.5% change in the composition of the

intermetallic can cause dramatic changes in the strength of the metal [Lim96]. Therefore,

the compositional stability of any intermetallic compound should be tested before selec-

tion.

Thin Films


Thin films have a number of unique properties versus their bulk counterparts.

Sinse the materials examined in this study are use as thin films in bimetallic actuators, an

understanding of their properties is needed. There are three primary reasons for the

unique properties of thin films. First, thin films are deposited onto a substrate, resulting








in substrate-thin film interactions. Second, the surface area-to-volume ratio for a thin

film is orders of magnitude higher than for a bulk material, so surface energies and

kinetics play a much greater role. Last, the techniques used to produce thin films are very

much different from the techniques used in the processing of bulk metals, which affects

their microstructure.

There are at least three common categories of thin film deposition processes:

chemical vapor deposition (CVD), physical vapor deposition (PVD), and electroplating

[Ohr92]. In this study we have only used physical vapor deposition methods consisting

of sputtering or thermal evaporation.

In all physical vapor deposition (PVD) processes, a vapor of the source material is

created and condensed on solid surfaces, including the substrate of interest. All PVD is

done in a high vacuum chamber so that the material vapor can travel from the source to

the substrate without reacting with atmospheric gases. In thermal evaporation, the source

material is simply heated until it vaporizes, a very simple and clean process useful to

deposit single element films. The vapor pressure of different material varies so greatly

that the composition of an alloy can not be maintained with thermal evaporation [Ohr92].

In the sputter deposition processes, an inert gas, normally argon, is ionized and ions are

accelerated into the source material or target. When the inert gas ion strikes the target

atoms, its momentum is transferred and by development of a momentum cascade, some

of the target atoms are ejected from the surface. The sputter process is based on momen-

tum transfer, therefore the vapor pressure of the target material does not control the rate

at which atoms are ejected from the target surface. Sputtering can therefore be used to

deposit metal alloys containing a number of elements [Ohr92]. However, differences in

composition may be seen between the deposited film and source material [Zhe97].

After being sputtered or sublimed, the vaporized atoms condense and agglomerate

to form a film [Sor95]. The morphology of the developing film is primarily determined

by the mobility of the condensing atoms [Mac95]. For thermal evaporation, the condens-

















0.3 0.5
SUBSTRATE
TEMPERATURE (TlTm)


Figure 2.8: Diagram of Movchen and Demishin's model of thin film morphology

[Kra90].

ing atoms have little kinetic energy and the surface mobility is primarily controlled by the

substrate temperature. Based on this, Movchen and Demishin [Kra90, Thor74] devel-

oped a model that predicts film morphology based on the homologous temperature of the

substrate (Figure 2.8). In the Movchen and Demishin model, the film morphology is

divided into three zones. In zone 1 (TH < 0.3) the adatom mobility is so low that tapered

crystals form with voids between the crystals. In zone 2 (0.3 < TH < 0.5) the adatom

mobility is great enough that most of the voids fill in and columnar grains form. In zone

3 (TH > 0.5) the mobility is high enough that nearly equiaxed crystals form.

In the sputtering process, the kinetic energy of the condensing atoms can be

controlled by the gas pressure and by biasing the substrate [Win92]. The effects of

substrate temperature and gas pressure have been studied by J.A. Thomton [Tho74], who

developed a model to predict the morphology of sputtered deposited films (see Figure

2.9). Thornton did not divide the sputtered morphology into zones as Movchen and

Demishin did. There is some correlation between these two models as Krauss illustrated

in figure 2.9 [Kra90]. It should also be noted that post deposition annealing of a thin film

may cause the microstructure to develop similar to that which would have developed had

the film been grown at that same temperature [Mac95, Kno91].


















ARGON 10
PRESSURE
(MICRONS)


TEMPERATURE
(TITm)


Figure 2.9: Thornton's model of sputtered thin film morphology with a comparison
to Movchen and Demishin's model of thin film morphology [Kra90].


During the deposition process and post deposition annealing, thin films can

develop a crystallographic texture [Mac95, Kno95]. There are two competing processes

responsible for the development of the thin film texture. In the growth of a columnar

microstructure, even from a melt, the crystallographic planes with the fastest growth

cause the elongated axis of the column [Bro94]. In thin films there is also a large driving

force to reduce surface energy. So the crystallographic planes that have the lowest sur-

face energy have a thermodynamic advantage over faster growing planes with higher

surface energies. The final texture that develops in a thin film is the balance of these two

factors.

There are two types of stresses that develop in thin films: intrinsic and extrinsic.

Intrinsic stresses are inherent to the deposition process and conditions, while extrinsic

stresses are due to the difference in the thermal expansion of the film and substrate.

Extrinsic stresses cause actuation of the bimetallic element.

Intrinsic stresses develop during the deposition process due to limited adatom

mobility [Mac95]. This limited mobility causes the formation of voids and vacancies in

the film. Voids tend to collapse and form grain boundaries. Vacancies migrate to grain







32
boundaries and are destroyed. This causes the film to attempt to contract, but instead it

may develop a state of biaxial tensile stress because it is attached to a rigid substrate. In

sputter deposition it is possible to increase the kinetic energy of the sputtered atoms, by

reducing the sputter gas pressure or biasing the substrate. Under bias, ions may penetrate

the first few layers of the deposited film. This "atomic peening" can reduce the intrinsic

tensile stress and even produce a compressive stress [Mac95]. However, atomic peening

injects a large number of trapped vacancies and implants gas [Mac95]. If the film is

heated to a temperature where vacancies or gasses are mobile, they will migrate to the

grain boundaries, be destroyed or evolved, and change the stress back to tension [Tow87].

In MEMS it is often desirable to have a residual stress near zero at room tempera-

ture. This allows for greater flexibility of design. While it is possible to achieve low

stress by adjusting the sputter parameters, it is sometimes possible to reduce tensile stress

by thermal cycling. If the thin film and substrate are thermally cycled to -196 C (liquid

nitrogen temperature) the metal film will normally contract more than the substrate. If

the film is ductile, it will plastically deform and upon returning to room temperature, the

film will experience a lower tensile stress or even a compressive stress [Bal94]. Unfortu-

nately, elevated temperature cycling of the film will counteract this compressive shift in

stress and re-introduce a tensile stress [Bal94].

A primary concern of this study is stress relaxation in the thin film, defined above

to be a time dependent change in the state of intrinsic plus extrinsic stress. Stress relax-

ation changes the force or displacement of a bimetallic actuator, and may be caused by

plastic deformation, creep, or diffusional processes including precipitation. The process

of plastic deformation and dislocation movement in thin films is restrictive as compared

with bulk metals. There are several reasons; first, a thin film is under a biaxial state of

stress. Thus dislocations can only glide in slip systems that have a component in the z

direction, normal to the film surface, since only these planes have a non-zero resolved

shear stress [Nix89]. If the thin film is textured there will even fewer slip systems with a
















...U ----.-- --------------- --- ---- ---- --.-- --- --- -.-. ..... .....-- -
i} ^ i








-100
0 50 100 150 200 250 300 350
Temperature, C

Figure 2.10: Stress temperature profile for pure sputtered aluminum and
microelectronic alloy Al-Si-Cu [Jer95].


resolved shear stress greater than the critical value. Second, a dislocation moving along a
slip system must nucleate misfit dislocations at the film/substrate interface, which re-
quires additional energy. If a passivating oxide forms on the surface of the film, as in the
case of aluminum, misfit dislocations must be created at both interfaces [Nix89], further

reducing dislocation glide.

The primary method used to test the mechanical properties of thin films is mea-
surement of stress as a function of temperature. As the thin film expands at a different
rate than the substrate, changing the temperature changes the stress in the film. This is an
easy and non-destructive way to test a thin film, but it is limited because the stress and
temperature cannot be varied independently. Instruments that measure stress in this way

include the Tencor Flexus which was used in this study (see description in Chapter 3).







34
The result of this type of testing is a plot of stress as a function of temperature, shown in

Figure 2.10, for sputtered pure Al and Al-1.5%Si-2%Cu. To evaluate stress relaxation

the sample can be heated to a temperature and the stress plotted as a function of time.

For bimetallic actuators, a film that shows no hysteresis in stress versus tempera-

ture over the range used is desired. Hysteresis in this plot indicates that the film is plasti-

cally deforming. Figure 2.10 shows that the metals currently being used to manufacture

bimetallic actuators [Jer96], both exhibit significant hysteresis over the temperature range

from 20C to 300-3500C. The pure aluminum shows little strength over 120C, while the

Al-Si-Cu alloy is only slightly stronger and plastically deforms above 170C under a

compressive stress of 50 MPa. The room temperature residual tensile stress for pure

aluminum is 150 MPa, while the Al-Si-Cu is 200 MPa. Isothermal stress relaxation tests

were also performed on pure aluminum and Al-Si-Cu alloy thin films [Jer96], with both

films showing a 15% reduction in stress over 1000 minutes. At 50'C the pure aluminum

film was initially under a tensile stress of 75 MPa, while the Al-Si-Cu alloy film was

initially under a higher tensile stress of 150 MPa.

There have been only a few studies to examine isothermal, long term stress

relaxation in thin films. The first was performed in 1985 by Hershkovitz, Blech and

Komem. They identified three modes of stress relaxation in thin aluminum films. The

first mode was identified as dislocation glide. The mechanism responsible for the other

two modes were not determined. In 1991, Drapper and Hill studied stress relaxation in

Al-Si-Cu thin films. They concluded that logarithmic creep was the dominant mecha-

nism responsible for stress relaxation. However, Drapper and Hill assumed a single

mechanism was responsible for the stress relaxation and fit there data with a single

exponential function, ignoring the conclusions of Hershkovitz, Blech and Komem. The

single figure of the data and curve fit in their article showed great inaccuracies between

the data and the fit. In 1995 Witvrouw, Proost, Beweerdt, Roussel, and Imec also re-








ported isothermal stress relaxation data from highly tensile stressed Al-Si-Cu thin films.

At low temperature (70C) they attribute stress relaxation to dislocation glide. At higher

temperatures (120C to 140C) they propose that the dislocations are cutting the Al2Cu

precipitates. From these limited studies, no complete model has been developed to

explain isothermal stress relaxation in thin films.

A new mechanism proposed by J. Tersoff for the stress relaxation in epitaxial

films is surface roughing [Ter94]. In this new mechanism the surface of the epitaxial

layer becomes rough, allowing easy nucleation of dislocations. This mechanism has not

yet been tested in polycrystalline thin films.

While strengthening of thin films to reduce stress relaxation and plastic deforma-

tion has not been extensively studied, work to reduce failures caused by thermal cycling

has been reported. Much of this work was done to improve the reliability of Josephson

superconducting devices (SQUID Superconducting Quantum Interference Device)

[Kir80], which were of interest for building ultrahigh-speed computers. These devices

operate at temperatures below 10K (-263C), and need to be able to withstand tempera-

ture cycles form 200C to -263C. Basavaiah and Greiner found the addition of gold to the

lead indium alloy reduced failures, but the mechanisms responsible for the reduction in

failures was not determined [Bas77].

Environmental Stability of Materials


Bimetallic actuators may operate in atmospheric gases at elevated temperatures,

and the materials of construction must be stable. Environmental instabilities of these

films would compromise their physical integrity and cause the device to fail. There are

two environment reactions that need to be considered for this application: oxidation and

corrosion.

Most metals oxidize, and some self-passivate while others require some protective

coatings. This study has focused on metals which self passivate, such as aluminum,








copper, titanium, nickel. For a passivating oxide to be effective it should have the follow-

ing four properties: high thermodynamic and kinetic stability, slow growth rate, adherence

to the metal, and easily form or re-form [Wes95]. The growth rate is of greater impor-

tance in thin films as an oxide could consume the entire film. Most passivating oxides

generally follow a parabolic growth rate equation given by



x = kt (4)


where x is the thickness of the oxide, k is the parabolic-growth-rate constant, and t is

time. Aluminum typically exhibits an inverse logarithmic dependence at low tempera-

tures and parabolic at high temperatures. Thus for thin film applications, materials with a

low parabolic-growth rate constant are needed, such as aluminum and silicon, for el-

evated temperature applications. In addition, the formation of an oxide and or a metal-

oxygen solid solution will introduce some stress into the metal substrate. This change in

stress could affect the force or displacement of a bimetallic strip. This change may be

time dependent as can be seem from equation 4.

Corrosion is also a concern because these devices may be operated in moist air

where the possibility of a galvanic cell exists. One area where a cell could exist is be-

tween the active and passive layers of the bimetallic strip. In this application, there is

normally an oxide layer between these two layers of the bimetallic strip that electrically

isolates them and so a galvanic cell cannot be formed. Disruption of this oxide would

cause this to be a concern. Also, formation of micro-galvanic cells in a two phase alloy,

should be tested [Ask89].
















CHAPTER 3
EXPERIMENTAL PROCEDURE


Introduction


Eleven different materials were deposited as thin films and analyzed in this study

(see Table 3.1). The deposition methods and procedures will be covered in this sections.

The films were first tested to determine their mechanical properties and the best materials

were analyzed further to determine their microstructural and chemical compositions. A

number of analysis techniques were used as discussed below.


Table 3.1: Materials examined in this study.


Material Deposited Thickness Composition or Purity, (wt. %)
prm
T201 Alminum 1.3 A -Al 4.6Cu 0.57Ag 0.36Mn -
T201 Aliminum 1.3 0.2Mg-0.27T.
0.2Mg 0.27Ti
5052 Aluminum 1.3 Al 2.5Mg 0.25Cr

2090 Aluminum 1.3 Al 2.57Cu 2.1Li -0.12Zr

Nickel 0.8 99.98

Titanium 0.55 99.99

Manganese 0.4 99.9

Copper 0.9 99.9

AI,Ti 0.4 Al-99.99. Ti-99.99

CuAu 1.3 Cu-99.99 Au-99.99

CuAu 1.3 Cu-99.99 Au-99.99

CuAu, 1.3 Cu-99.99 Au-99.99








Thin Film Deposition

The thin films examined in this study were deposited by either sputtering or

electron beam evaporation. All the films were deposited on 100mm diameter, [100]

oriented, single crystal silicon wafers. The thickness of each film is shown in table 3.1.

The wafers used for the aluminum alloy and copper gold intermetallics had a 100 nm

thermal oxide grown on them before film deposition. All wafers were cleaned using the

following procedure: 5 minute in an ultrasonic bath for each solvent with a de-ionized

water rinse between solvent for, trichroloethane, acetone, and methanol, then five minutes

in a solution of 75% sulfuric acid and 25% hydrogen peroxide-30% followed by a de-

ionized water rinse and blown dry with dry nitrogen.

The aluminum alloy and copper gold intermetallic films were sputter deposited by

Sputtered Thin Films, Inc., Santa Clara, CA, using an 8-inch DC magnetron sputter gun

running at 5kW of power. The chamber was first pumped down to 2.8x10-7 Torr, then

back filled with argon to a pressure of 8 mTorr. The substrates were heated to 1000C.

The aluminum alloy sputter targets were from bulk sheet stock. Copper gold intermetal-

lics were produced by depositing four layers of pure copper and gold in the proper pro-

portion to produce the required composition. Table 3.2 shows the thickness of the indi-

vidual copper and gold films that were deposited for each of the average compositions.

Table 3.2: Layer thickness deposited for the copper/gold intermetallic films.


C d Copper Layer Gold Layer
Compound
Thickness, gm Thickness,4m

Cu3Au 0.1132 0.4868

CuAu 0.2465 0.3533

CuAu, 0.4059 0.1941








The copper, titanium, nickel, manganese and aluminum-titanium intermetallic

films were deposited by electron beam evaporation. The evaporation system contained

three Telemark Model 211 single pocket electron-guns in the vacuum chamber. These

guns are powered and controlled by a Sloan PAK 12 series 12 kW power supply operated

at 10 kV. The deposition rate and total deposited thickness were measured with quartz

crystal monitors, either an Inficon XTC or Sycon model 100. The deposition rate was

manually controlled and varied between 10 to 15 angstroms per second. The vacuum

chamber was pumped down to 5xl0- Torr before each deposition. Due to outgassing, the

chamber pressure would rise to 5x 10 6Torr during deposition. The substrates were heated

to 1500C +/- 100C with resistive heaters, manually controlled with a variable transformer.

Copper Gold Heat Treatment


The copper gold compounds required a special heat treatment to homogenize the

layers and to develop the long range ordered intermetallic structure [Fed58]. These heat

treatments were performed in a vacuum of 5x 10-6 Torr or better to inhibit oxidation of the

copper. The heat treatments were 12 hours at 500C to homogenize the film, and 12

hours at 2750C followed by 12 hours at 225C to develop the long range order. All

temperatures changes were ramped at I1C per minute using an Omega CN3000 controller.

Stress and Stress Relaxation Measurements


Stress and stress relaxation in the samples were measured with a Tencor Flexus

model 2320. The Flexus 2320 uses an optical lever to measure the curvature of the

sample [Tur96]. Stoney's equation [Sto09], Equation 4, was used to calculate the average

biaxial stress, a, in the film:



E-h2 (4)
(1-uv)-6-R-t









where R is the curvature of the substrate, t is the film thickness, h is the substrate thick-

ness, E is the Young's modulus for the substrate, and v is Poisson's ratio for the substrate.

The silicon substrate's thickness and curvature were measured before the film was

deposited. The curvature of the substrate was then measured after the film was deposited,

and the change in curvature is used in Stoney's equation to calculate the film stress. In

addition to measuring stress at room temperature, the Flexus could heat a sample to

500C or cool it down to -60C under computer control and measure curvature. The

Flexus could also temperature cycle a sample, taking measurements during the cycle to

generate a stress versus temperature plot (see Figure 2.10).

The isothermal stress relaxation in aluminum alloy thin films was also measured

with the Flexus. To obtain a similar starting condition for different tests, each sample

was heated to 3500C at PlC/minute, held at 350C for 30 minutes, and cooled back down

to room temperature at I C/minute. The sample was then taken to the test temperature, at

lC/minute, and held at the test temperature for forty eight hours. Every 15 minutes the

stress in the sample was measured, allowing stress to be plotted versus time (see Figure

3.1). Once the mechanical properties of a film had been measured with the Flexus, the

samples was divided for chemical and microstructural evaluation.

Powder X-Ray Diffraction (XRD)



Powder x-ray diffraction was used to determine the long range ordering in the

copper gold films [War69] and to evaluate crystallite size in the aluminum alloy films

[Cul78]. The system used was a Philips APD 3720 x-ray diffraction systems controlled

with an IBM-PC type computer running Microsoft windows 3.11 with Philips PW1877

Automated Powder Diffraction software, version 3.6g. X-rays from a copper anode

operated at 40 kV and 2 milliamps with a nickel filter were used.






























Time, Minutes
105



















Figure 3.1: Isothermal stress relaxation in 5052 aluminum alloy sample at 125C.



X-ray diffraction is effective for detecting the change in structure that occurs in an

order-disorder transformation. For the copper gold system, the disordered crystallo-

graphic structure is FCC, with atoms sitting randomly on lattice sites. In an ordered

phase, one type of atom, either copper or gold, sits in the corner sites and the other atom

will sit in the face sites. This creates two inter-linked simple cubic or simple tetragonal

crystallographic structure. This change in the crystal structure causes the original diffrac-

tion peaks to change, and the different structure factor causes a number of new peaks to

appear [Cu178]. With a true randomly oriented sample, it is possible to calculate the

degree of ordering in the sample by comparing different peak heights [War69]. However,
thin films have a crystallographic texture that complicates this procedure. As a result it

was difficult to quantify the degree of order in the films.

For aluminum alloys the dislocation structure inside the aluminum crystals, the

grain size and the strain were all evaluated using x-ray diffraction line shape analysis. In
'5






I XI



80 -----------1 ----- 1 ----- 1 ------ 1 ------
0 S00 000 1500 2000 2500 3000
Time, Minutes

Figure 3.1: Isothermal stress relaxation in 5052 aluminum alloy sample at 1251C.



X-ray diffraction is effective for detecting the change in structure that occurs in an

order-disorder transformation. For the copper gold system, the disordered crystallo-

graphic structure is FCC, with atoms sitting randomly on lattice sites. In an ordered

phase, one type of atom, either copper or gold, sits in the comer sites and the other atom

will sit in the face sites. This creates two inter-linked simple cubic or simple tetragonal

crystallographic structure. This change in the crystal structure causes the original diffrac-

tion peaks to change, and the different structure factor causes a number of new peaks to

appear [Cul78]. With a true randomly oriented sample, it is possible to calculate the

degree of ordering in the sample by comparing different peak heights [War69]. However,

thin films have a crystallographic texture that complicates this procedure. As a result it

was difficult to quantify the degree of order in the films.

For aluminum alloys the dislocation structure inside the aluminum crystals, the

grain size and the strain were all evaluated using x-ray diffraction line shape analysis. In








this technique, the (111) and (220) diffraction peaks were selected for evaluation. The

(111) peak was scanned from 28 = 380 to 39.20, and the (220) peak was scanned from 20

= 64.240 to 66, using 0.02 increments and 10 second dwell time. For line shape analysis,

an unstressed quartz standard was scanned in the same diffractometer under the same

conditions. The quartz standard was scanned from 20 = 38.70 to 39.80, and from 20 =

63.50 to 64.7.

Line shape analysis extracts the factors that cause the x-ray diffraction peaks to

shift and/or broaden. There are three main causes of peak broadening in x-ray diffraction.

The first is the diffractometer itself. Instrumental broadening is eliminated by compari-

son to the quartz standard. The other two causes of peak broadening are strain and

crystallite size. Each of these produce a different shape. Strain effects produce a

Gaussian distribution while the crystallite size produce a Cauchy distribution [War69].

The Phillips diffraction software can deconvolute these effects and determine the strain

and crystallite size.

Atomic Force Microscopy (AFM)

A Digital Instruments Nanoscope III atomic force microscope (AFM) was used to

determine the surface topology of the aluminum alloy samples [Hud92]. A silicon tip in

the tapping mode was used to probe the surface of the samples. Digital Instruments

software was used process the scanned images and calculate the RMS surface roughness.

Scanning Electron Microscopy (SEM)

A JEOL 6400 scanning electron microscope was also used to image the surface of

samples. In addition to the secondary and back scattered electron detectors, this system

has an energy dispersive spectrometer (EDS) that allowed qualitative elemental analysis

of surface features [God92]. Samples were first viewed in the secondary electron mode

to evaluate the surface topology. Very bright areas were viewed in the back scattered








electron mode. The true secondary electron yield from an area is more strongly depen-

dent on the topology, whereas the back scattered electron intensity is dominated by the

atomic mass of the excited atoms. Areas that appeared to have different atomic composi-

tions were examined with EDS to determine the concentration of elements.

Electron-Probe Microanalysis (EPMA)

EPMA was used to quantitatively determine the compositions of the aluminum

alloys and the copper gold intermetallic films [Gol92]. A JEOL 733 Superprobe with

four wavelength despersive spectrometers was used. For the aluminum alloy films the

bulk aluminum alloy was used as the analysis standard to determine difference in compo-

sition between the bulk and thin film composition. For the copper gold intermetallic

films, pure copper and gold standards were used with the ZAF (atomic number effects, X-

ray absorption affects, and X-ray fluoresces effects) analysis technique to determine the

composition of these films [Gol92].

Transmision Electron Microscopy (TEM)

A JEOL 200CX analytical transmission electron microscope was used to examine

precipitates and the crystal size of the aluminum alloys. The samples were prepared by

first cutting a 3mm disk from a silicon substrate with a deposited film. These samples

were then thinned from the silicon side to between 150tm and 250.m. Using a dimpling

grinder, the back sides of the samples were hollowed out so that the center of the samples

were approximately 50glm thick. The thinned samples were encapsulated in paraffin and

a small opening (1mm) was made in the paraffin over the dimple. The samples were

etched in a solution of 50% HF and 50% HNO,. The etching process was monitored by

viewing the samples through a low power binocular microscope. Once the acid had

etched through the silicon and reached the aluminum film, etching was stopped. These

samples were ion milled until a small opening appeared in the aluminum, then viewed in

the TEM.








In the TEM, the samples were first viewed in the bright field mode to evaluate the

general structure of the sample [Lor94]. The dark field mode was used to better identify

different phases. Electron diffraction images and patterns were also taken to identify the

phases present by calculating the interplaner spacing. The inter plane spacing, dhkl was

calculated using:



SL=Rdhk (5)


where L is the camera length (for the JEOL 200CX L = 82 cm), X is the electron

wavelength which is based on the accelerating voltage (X = 0.0251nm based on an accel-

erating voltage of 200keV) and R is the radius of the diffraction pattern.

Etching

Grain size is an important parameter when evaluating the strength of a metal.

Etching is the primary means by which the grain size may be made visible in bulk metals,

but this technique was found to be ineffective in this study. Etching thin films to define

grains is problematic. Most etchants attach different crystallographic planes at different

rates. In a material with a randomly oriented crystals, this type of etchant works well.

However, thin films are strongly textured and tend to etch more uniformly, complicating

the determination of crystal size. Some etchants that attack grain boundaries preferen-

tially [Smi67, Gif70] and two were tested for aluminum alloys: 1% hydrofluoric acid and

a 10% sodium hydroxide solution. Satisfactory definition of the grain boundaries was not

achieved.

Auger Electron Spectroscopy (AES)

The Perkin-Elmer PHI 660 Scanning Auger Microprobe was used to evaluate the

oxide thickness on the aluminum alloy films. Auger analysis is very surface sensitive,








containing information primarily from the top three or four atomic layers [Hol80]. The

Auger system contained an argon sputter gun that enables depth profiles to be collected.

To do so, an Auger spectrum was taken, the sample sputtered for 15 seconds, and another

Auger spectrum taken. This process was repeated to the required depth. The aluminum

alloy samples were sputtered until the oxygen signal had decreased to 10% of its original

value. This technique gives a thickness in terms of sputter time. The conditions used

result in an estimated sputter rate of 250 angstroms per minute These data are primarily

used to compare differences in oxide thickness from sample to sample.

Curve Fitting

To curve fit the stress relaxation data, curve fitting techniques were evaluated:

polynomial, single exponential, and double exponential. The double exponential tech-

nique fit the data using the equation:



a(t)=A+Be"' + Ce"2 (5)


where a is stress, t is time, and A, B, C, m, and m2 are constants that are determined from

the data. This technique fit most of the data with the least error.

The process for determining the constant was the linearization of the data. This

was sometimes complex as the two exponential functions needed to be separated to be

linearized. Separation is accomplished by first fitting the stress relaxation data for t > 380

minute. To linearize these data, the value of A in equation 5 was first manually selected

and subtracted from the remaining data. The natural log for the remaining data was then

taken and least squares linear fit to determine the slope, intercept and correlation factor.

The A constant was iterated to obtain the best possible correlation factor. With this step

completed, the first exponential was subtracted from the original data. The natural log of

the difference was then taken and the resulting data again fitted with a linear least square







46
fit to determine the slope and intercept, m and C.. All calculations were done in

MathCad Version 4.0.

Numerical Modeling


The effects of oxide thickness were found to be significant, therefore a numerical

model was developed to quantify this effect. This model was based on the equation

developed by Townsend [Tow87] and implemented in MathCad version 4.0. In this

analysis a three layer structure was modeled. The first layer was a 5tm thick layer of

silicon, and the second layer was 5ltm of aluminum. The third layer of Al203 was varied

from 0 to Inm in thickness. This model was used to calculate the curvature of the struc-

ture, therefore the length was not required. The output was plotted as a percentage

change in curvature for the structure with varying oxide thickness.














CHAPTER 4
RESULTS

Introduction


The results of the testing and analysis will be presented in this chapter which is

organized by material to provide an understanding of their performance.

T201 Aluminum


Figure 4.1 shows the stress versus temperature plot for the T201 alloy film and

may be compared to the literature data shown in figure 2.10. These data were collected

after an initial temperature cycle to 3500C to stabilize the film's microstructure after

deposition. While the intrinsic stress in these films were a function of deposition param-

eters, the first cycle to 350C annealed and consolidated the films and the extrinsic

stresses were so high that they dominated the stresses reported as in Figure 4.1. As can

be seen from these data, the room temperature residual tensile stress is almost 400 MPa

and the film begins to plastically deform at approximately 2000C under a compressive

stress of 50 MPa. This sample was also cooled to -1960C in an attempt to reduce its room

temperature residual stress, but this procedure had no effect on the room temperature

residual stress.

The isothermal stress relaxation seen in these samples at different temperatures

are shown in figures 4.2 through 4.6. These data were curve fit as describe above and the

equation is shown in each plot. Below 1251C the stress relaxation data was fit well by a

double exponential decaying function that suggesting that are two stress relaxation pro-

cesses were occurring. At 1251C and 150C the noise on the data are to large to allow

fitting. Rather than heating at a rate of lbC/minute samples were also heated rapidly

47














300 ------------... ...- ----- -------,----- ------------------------




( 200 ------ --- ---- ---- --------- ------------
a-


(i 100 v -- S .- ^ -




O 0.
CD,

0




-100
0 50 100 150 200 250 300 350

Temperature, C
Figure 4.1: Stress versus temperature plot of T201 aluminum thin film.


390 I -



o(t) = 330.2 + 48.9e -5.45o1 + 9.2e -.096"1
380 -



"370
LO


Time, minutes

Figure 4.2: Isothermal stress relaxation of T201 aluminum thin film at 50C.
































Figure 4.3: Isothermal stress relaxation of T201 aluminum thin film at 75C.


Figure 4.4: Isothermal stress relaxation of T201 aluminum thin film at 100oC.






























Time, minutes
Figure 4.5: Isothermal stress relaxation of T201 aluminum thin film at 125C.

46 1--------1----------------1------- 1 -----


4O O
40 0
0 0 %0 .0

0 0\^^^X
I 00^ 0 *' ^<


36 I I I I
0 500 I000 1500 2000 2500

Time, minutes
Figure 4.6: Isothermal stress relaxation of T201 aluminum thin film at 1500C.









(10C/min) to the testing temperature to evaluate the effect of ramping rate on stress

relaxation. The stress relaxation was little changed, showing that ramping rate had little

affect on the stress relaxation seen in the samples. A compilation of the curve fitting

parameters is shown in table 4.1.

In addition to the mechanical evaluation of the T201 thin films, microstructure

and chemical composition was evaluated. Table 4.2 shows the results of the EPMA

analysis of the thin films and the composition of the bulk alloy, as determined by an

external lab. There are systematic differences in composition between the sputtered thin

film and the bulk alloy presumably due to the sputter deposition process [Zhe97].

Figure 4.7 is a SEM micrograph of a T201 thin film sample that was heat treated

for 60 hours at 100C. Back scatter images of the same regions showed that the bright

areas had a different atomic composition than the areas around them. Figure 4.8 is the

EDS spectrum for the dark areas showing only aluminum. Figure 4.9 shows the EDS

spectrum for the bright areas and both copper and aluminum are detected while silicon is

not. Therefore, the bright areas on the SEM image are due to both topographic features

and differences in chemical composition of the sample, presumably due to precipitates.



Table 4.1: Residual stress after infinite relaxation time, o, change in stress at short (AG)
and long (Aoa) times, and the time constants (t) for the two exponential equations
for T201 aluminum.

Temp. Ao T" AzC
(C) (MPa) (MPa) (min) (MPa) (min)

50 330 9.2 196 49 1820

75 240 13.8 220 26 1940

100 160 608 220 29 1400

125 33 0 0 81 960

150 40 -3 200 0 0










Table 4.2: Chemical composition of sputter deposited T201 aluminum alloy thin
films determined by EPMA, and compared with the bulk alloy composition.


Alloying Bulk Alloy Sputtered Thin Film
Element Weight Percent Weight Percent

Al 93.91 93.19+/-2


Ag 0.57 0.48 +/- 0.04


Cu 4.6 4.59 +/- 0.2


Ti 0.27 0.53 +/- 0.04


Mg 0.29 0.34+/- 0.1


Mn 0.36 0.87+/- 0.1


Figure 4.7: SEM micrograph of T201 aluminum sample after 60 hours at 100"C.









A
L












c s
U I

1 2 3 4 5 6 7 8 9 10
Energy, keV
Figure 4.8: EDS spectra of dark area of SEM image shown in figure 4.7,
showing a lack of copper and silicon.



A
L


C
U




s


1 2 3 4 5 6 7 8 9 10
Energy, keV
Figure 4.9: EDS spectra for bright areas in SEM image, figure 4.7, consistent with
a precipitate rich in copper.


Atomic force microscopy (AFM) was used to evaluate the surface morphology of

the samples, with an image shown in Figure 4.10. Mounds or hillocks about 0.6p1m high

were present on the sample surface. The AFM was also used to evaluate the surface

roughness of these films after different heat treatments. Due to the large mounds that

form on the surface, the RMS roughness varied greatly from position to position on the

same sample. An effort was made to scan areas unaffected by the mounds, but this

required very subjective selection and exclusions of areas of the sample. Thus, the RMS

roughness data were not reliable and are not reported.






























Figure 4.10: AFM amplitude image of T201 aluminum thin film after stress relaxation at
125C for 2 hours. The scanned area is 15 lm by 15 ltm.


The primary strengthening techniques used in the T201 alloy are solid solution

strengthening (due to alloying elements of magnesium and manganese [Dav90, Bro82])

and precipitation hardening (due to alloying element copper). Silver is added to this alloy

(0.57 % weight percent) to stabilize the copper precipitates [Dav90]. The effectiveness of

precipitation hardening is controlled by the size and distribution of the precipitates in the

host crystals [Cah83]. TEM was used to evaluate the formation of precipitates in the

aluminum crystals. Figure 4.11 a shows a plan view TEM image of a T201 sample. The

small (10 nm to 30 nm) dark spots in this image are the CuA12 precipitates in the alumi-

num matrix. Figure 4.1 lb shows the indexed electron diffraction pattern from this sample

with the aluminum diffraction rings and the CuAl, diffraction spots identified.

The bright field TEM images also give some indication as to the size of the

aluminum crystals. Caution must be used in evaluating grain size from this image be-

cause the etching process used in the sample preparation would likely attack small crys-




































""at..


Aluminum Grains
Aluminum Grains


CuAl2 (422)


A (200)
Al (200)


Al (1ll)


Figure 4.11: TEM micrographs of T201 thin film sample. (a) bright field image showing
aluminum crystals and CuAI2 precipitates; (b) indexed electron diffraction pattern show-
ing the diffraction rings for the Al (111) and (200) planes and for the AICu2 (422) plane.


Percipiijiel f C(uAI.


a


i








tals faster than large crystals. Also, a TEM image only examines a small area of the

sample, and a large area must be viewed to determine the true grain size distribution. The

apparent average crystal sizes was between 1 and 2 R.m.

While direct measurement of the aluminum grain size by etching was unsuccess-

ful, formation of the copper rich precipitates gives some indication of the aluminum grain

size. Precipitates tend to nucleate and grow at triple points where three crystals come

together [Ree92]. While the direct calculation of grain size from the precipitation density

is questionable, the comparison of different samples and inferring changes in crystal size

based on changes in precipitation density is reasonable. Figure 4.12 shows a SEM micro-

graph of a T201 sample that was heated to 350C and then cooled to 25C over 48 hours

to test the stability of the microstructure. The number of copper rich particles in figure

4.12 are of the same order as that of figure 4.7, i.e. 150 +/- 10 particles in the 20 by 10

gim area of the micrograph. Therefore, the precipitate size and density of these samples

is apparently stable over several temperature cycles to 350C.


Figure 4.12: SEM image of T201 aluminum sample after temperature
cycling from 3500C to 25oC over 48 hours.









10

0
8

7-

I 6






0 I 0
2S A] Al




0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5,0

SPUTTER TIME, min.

Figure 4.13: Auger depth profile of T201 thin film. The film was sputtered
with 3 keV argon in a 3x3 mm raster.


The effect of oxidation on a bimetallic actuator's performance is a concern. AES

analysis was used to measure the oxide thickness on a T201 thin film sample that had

been cycled to 3500C 5 times. Figure 4.13 shows the Auger depth profile. The rastered

ion gun used in this analysis sputtered A120, at an approximate rate of 25 nm per minute.

Thus the oxide thickness on this sample is approximately 25 to 30 nm.

Powder x-ray diffraction analysis was also performed on the T201 thin film

samples. A broad 20 scan was taken of a sample cycled to 3500C once to determine the

phases present in the sample. Only elemental aluminum and silicon from the substrate

were detected in these scans. XRD scans were then run on samples that were heat treated

at 50C or 1251C for up to 32 hours. The samples were scanned for their (111) and (220)

peaks. Figures 4.14 and 4.15 show the diffraction peaks after different stress relaxation

times at 1250C. Note that there is a substantial shift in the peak position between the

initial scan and those that have undergone stress relaxation for the shortest time of two





















S60K



40K 2 Hours

32 hours
S Hou s
20K



OK
38.0 38.2 38.4 38.6 388 39.0 39.2
20

Figure 4.14: Aluminum ( 11) x-ray diffraction peaks for T201 thin film samples
after various stress relaxation times at 1250C.






4000
0 Hours


3000

C

2 HOURS 16 s
2000




1000 Hou





65.6 65.8 66.0 66.2 66.4 66.6
20



Figure 4.15: Aluminum (220) x-ray diffraction peaks for T201 thin film samples
after various stress relaxation times at 125C.









hours. Line shape analysis was performed on the X-ray data to determine crystallite size

as a function of stress relaxation, but the crystallite sizes varied randomly and no correla-

tion was found.

5052 Aluminum


Figure 4.16 shows the stress versus temperature plot for the 5052 aluminum alloy

thin film. These data were collected after an initial temperature cycle to 3500C to stabi-

lize the film's microstructure. As can be seen from these data, the room temperature

residual tensile stress in the film is 300 MPa versus 400 MPa for T201, 200 MPa for Al-

Si-Cu, and 150 MPa for pure AL. The film begins to plastically deform under a compres-

sive stress of 25 MPa at 2001C. A 5052 sample was also cooled to -1960C in liquid

nitrogen to reduce the room temperature residual stress. Upon warming to room tempera-



400



300



200 --- ------ -----------






Temperature C


Fe 4.16: St------------- ------ress versus temperate pt fr 52

-100
0 50 100 150 200 250 300 350
Temperature, C

Figure 4.16: Stress versus temperature plot for 5052 aluminum.








ture, the tensile stress was reduced from 300 MPa to 125 MPa, a 58% reduction. Starting

at the reduced room temperature stress of 125 MPa the film begins to plastically deform

at a lower temperature of 125oC under a compressive stress of 25 MPa. Upon cooling

from 350C to room temperature, the tensile stress returned to 400 MPa.

The isothermal stress relaxation seen in these films at different temperatures is

shown in figures 4.17 through 4.21, along with the fitted curve and the equation. At 50C

the stress relaxation was best fit with a single exponential curve. For temperatures

between 75C and 125C the stress relaxation data was fit well by a double exponential

curve. At 150C the noise in the stress relaxation data is so large that the fitted curve is

unreliable. Samples were also heated at 10C/minute to the testing temperature to evalu-

ate the effect of ramping rate on stress relaxation. Again, this had little affect on stress

relaxation. The curve fitting parameters are compiled in table 4.3.


Time, minutes

Figure 4.17: Isothermal stress relaxation in 5052 aluminum thin film at 50oC.


































Time, minutes
Figure 4.18: Isothermal stress relaxation in 5052 aluminum thin film at 75C


S1000 1500 2000 2500
Time, minutes
Figure 4.19: Isothermal stress relaxation in 5052 aluminum thin film at 100C
























2 95



90-







80
85 wI I


0 500 1000 1500 2000 2500

Time, minutes
Figure 4.20: Isothermal stress relaxation in 5052 aluminum thin film at 1250C.


u U00 1000 1500 2000 2500

Time, minutes
Figure 4.21: Isothermal stress relaxation in 5052 aluminum thin film at 1500C.







63

Table 4.3: Residual stress after infinite relaxation time, i. change in stress at short (Ao)
and long (A() times, and the time constants (7) for the two exponential equations,
for 5052 aluminum.

Temp. o. Ao ACa, "
("C) (MPa) (MPa) (min) (MPa) (min)

50 240 0 0 30 2800


75 150 12 200 33 1500


100 110 19 170 29 1380


125 80 10 160 17 220


150 0 0 0 58 50000



Table 4.4: Chemical composition of sputter deposited 5052 aluminum alloy determined by
EPMA, compared with the bulk alloy composition.


Alloy Element Bulk Alloy Sputtered Film
Weight Percent Weight Percent

Al 97.25 98.9 +/- 0.5


Mg 2.5 1.73 +/- 0.03


Cr 0.25 0.16 +/- 0.06


The sample that was cooled to -196"C was tested for stress relaxation at 75C. At

this temperature the tensile stress in the film was 50 MPa and there was no stress relax-

ation observed over forty-eight hours of testing.

The EPMA chemical analysis of the 5052 samples is shown in Table 4.4. Again,

there are differences in composition between the sputtered thin film and the bulk alloy,

presumable due to sputter deposition effects [Zhe97].





























Figure 4.22: SEM micrograph of 5052 aluminum alloy thin film sample that was
temperature cycled to 3500C at 10C/minuteand held at 350C for 30 minutes
and then cooled to room temperature over 48 hours.


Figure 4.23: SEM micrograph of 5052 aluminum alloy thin film
that was held at 100C for 60 hours.






























Figure 4.24: AFM height image of 5052 aluminum alloy thin film without heat treat-
ment. Scan area is 5 im by 5 gpm, Z-range 32nm and RMS roughness is 3.77 nm.


Figure 4.25: AFM height image of 5052 aluminum alloy thin film heat treated at 125C
for 32 hours. Scan area is 5 pm by 5 tm, Z-range 60nm and RMS roughness is 7.09 nm.







66
Table 4.5: RMS roughness and Z-range for 5052 aluminum alloy thin films that have
undergone different heat treatments.

50C 125C

Time at RMS
Tempe atur R Z-Range RMS Roughness Z-Range
Temperature Roughness (nm) (nm)
,TTs (m (nm) (nm) (nm)
(Hours) (nm)

0 3.77 32.4 3.77 32.4

2 5.9 47.5 5.9 60.2

4 4.6 35.7 5.6 55.2

8 5.8 51.1 5.6 55.5

16 6.3 49.4 4.6 58.9

32 5.6 51.7 4.7 43.6



Figure 4.22 and 4.23 show SEM micrographs of 5052 alloy thin films that have

undergone different heat treatments. EDS was used to examine the difference between

the light and dark areas seen in figure 4.23. No difference in composition between these

two areas was found.

The changes in AFM surface roughness with stress relaxation in samples that

were heat treated at 50C and 1250C for up to 32 hours is shown in Table 4.5 and figures

4.24 and 4.25. The RMS roughness and Z-range (the height difference between the

highest and lowest point on the sample) are reported in Table 4.5. Figure 4.24 shows the

AFM height image of a 5052 alloy sample with no heat treatment and figure 4.25 shows

the AFM height image of a sample that had been held at 125C for 32 hours. The size of

features in these images are nearly the same, but the RMS roughness and the Z-range

have doubled for the sample held a 125C (Table 4.5). The data in Table 4.5 shows that

the change in surface roughness occurred in the first two hours of the heat treatment, and

that heat treatment temperature has little effect on the rate or the magnitude of the rough-








10

9
0
8

7








2 -Mg Al A] A]
0Mg

0 I I i II
0.0 0,5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4,5 5.0

SPUTTER TIME, min.

Figure 4.26: Auger depth profile of 5052 alloy thin film. Film sputtered with
3 keV argon in a 3 mm x 3 mm raster.



ness change. AES depth profile data in Figure 4.26 shows that the surface oxide was

between 20nm and 30nm thick.

Powder x-ray diffraction analysis of 5052 thin films heat treated at 500C or 125C

for up to 32 hours showed that the (111) peak shifted low 0.1" for heat treated (2 to 32

hours) versus unheated samples. No additional peak shifts were detected for samples

heat treated longer than two hour. The peak shift for the (220) showed the same trend

with a shift of 0.1. Samples were also scanned to detect changes in the A13Mg2 peak

heights or widths that would indicate a change in these precipitates. No measurable

change was seen in these peaks indicating that the precipitates had not coarsened. Line

shape analysis was also performed on the X-ray data to determine crystallite size as a

function of stress relaxation, but no correlation was found with stress relaxation time or

temperature.










200 ----


150 ----- --i- ------ ---- -------- ----
150


0

100 ---- --- ---- ---- ----






-100
0 50 100 150 200 250 300 350
Temperature, C
Figure 4.27: Stress versus temperature plot of 2090 aluminum alloy thin film.


2090 Aluminum

Figure 4.27 shows the stress versus temperature plot for the 2090 aluminum alloy

thin film for the first and second temperature cycle to 350C. The as deposited tensile

stress in this film is 250 MPa before the first cycle. During the first temperature cycle,

after deposition the film stress decreases linearly with increasing temperature up to

140C. Between 140C and 2000C the stress in the film increases with increased tempera-

ture, presumable due vacancies migrating to and being destroyed at the grain boundaries
[Tow87]. This behavior was typical for T201 and 5052 films also during the first thermal

cycle. Once the excess vacancies were eliminated, the slope again becomes linear up to

275C. At 2750C the slope increases and becomes unstable up to 350C, where the film is

under a compressive stress of 70 MPa. Upon cooling, the slope of the stress-temperature

curve is much less than during heating and the film only reaches a tensile stress of 80







69




9 \0


10




4 0



2-

1 C
C
0I I i I
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0

SPUTTER TIME, min.


Figure 4.28: Auger depth profile of 2090 aluminum alloy thin film. Film sputtered with
3 keV argon in a 3x3 mm raster.




MPa at room temperature. On the second temperature cycle the film attains a slightly

higher compressive stress of 80 MPa at 3500C and returns to a tensile stress of 80 MPa

upon returning to room temperature. When the sample was removed from the Flexus it

was noticed that the specular reflection of the film was greatly reduced due to surface

oxide.

Based on the reduced specular reflection and the lower slope of the stress tem-

perature curve it was assumed that the film had oxidized badly. This was confirmed by

the Auger depth profile data from that sample, (see Figure 4.28). As can be seen, the

oxygen concentration falls off very gradually. This indicates that the film had oxidized to

a depth greater than the 5052 or T201 films (see figures 4.26 and 4.13).








Copper

The stress versus temperature plot for the copper thin film is shown in figure 4.29.
While this plot is quite complex the data are consistent with the film being oxidized
during heat treatment. As the stress-temperature plot indicates, the film was stable during
the first thermal cycle to only 1000C. Another film was tested for stress relaxation at
50oC for 22 (see figure 4.30). The copper film oxidized in this test, as indicated by a
greatly reduced the specular reflection, and the film stress increased after 14 hours due to
oxidation.
To inhibit the oxidation of the copper thin film, a 75 nm thick protective alumi-
num layer was deposited on the copper. This sample was then heated to 500C for 6
hours in a tube furnace to test the stability of this protective coating. The film retained its
specular reflection during this test. From an Auger depth profile on this sample the oxide
thickness is approximately 200nm (see figure 4.31).

200

100 --------------------------|-----------


0---"V ^--------------- -----



0 200 ..........

-3 o o .. .. .. .....-- -- '- --- .- -"-: ---. .. ......-- .- -.. --. -... -..... .....


-400 --------.--- -- ----------- -.------ -----,-
-500 ------___---------------------___



-500
0 100 200 300 400 500
Temperature, "C
Figure 4.29: Stress versus temperature plot for a copper thin film.














CO
0- 400 .-------------- --- ------------------- 4-- ---------------------
0.40


30
C,
23 0 -^ .^ .------ .-------J---I- --------.--|---------------.-------- -- ----------- -------


10



-1 0 --------------------- ----------- -------------------- -------- -------------

-10

0 5 10 15 20 2
Tim e, H ours
Figure 4.30: Isothermal stress test of a copper thin film conducted at 500C.


0.0 0.5 10 1.5 2.0 2.5
SPUTTER IME, min.


30 3.5 4.0 4.5 5.0


Figure 4.31: Auger depth profile of a copper film coated with a protective layer of alumi-
num 750 nm thick. The film was sputtered with 3keV argon in a 3x3 mm raster.


CuC
pJ 0









Titanium. Manganese and Nickel


The stress versus temperature plots for titanium, manganese, and nickel are shown

in figures 4.32 through 4.34. All of these films oxidized severely during their first tem-

perature cycle up to 500C, 250C, and 500oC for titanium, manganese and nickel respec-

tively. This was evident from a change in specular reflection, color change of the samples

and by the stress-strain data. For titanium (figure 4.32) the rapid reduction in tensile

stress at 350C to a compressive stress of over 600 MPa at 500C and the film maintain-

ing a high tensile stress (375 MPa) upon cooling to room temperature indicates some

changes in the film from oxidation. Oxidation of the manganese film (figure 4.33) results

in an 800 MPa shift in the stress, becoming highly tensile at just over 2000C. The nickel

film (figure 4.34) shows a change in slope of the stress strain curve at 2750C and the

residual stress at room temperature dropped from 1000 MPa before to 600 MPa after the


1000

800 -------------------------------------

600 ------ ---- ---- ------ ----I------ ----------------- ---- -----^--- ------------

2 0 0 ------ --------- -------- ------------ ---- ---- -,-- .----- --------------

2 200............. ....... ... -..............
) 00

-200
a-




-4 0 0 C o oli. ----- I --... .. .
-600 ---------- ------------------------ --------------


-800 _
0 100 200 300 400 500

Temperature, C

Figure 4.32: Stress versus temperature plot for a titanium thin film.







73
1000

900 Coo

800 ----------------------- W..... .. .

700
uo
S600---------- --- ------- ---------
c,
S500 --......--.....-.-...-----..---.................-------

200----- ------




-100
-100 ------------ ,-'-
0 50 100 150 200 250
Temperature C
Figure 4.33: Stress versus temperature plot for a manganese thin film.

1200


1000 ------------------------------------
100 --------------------- 0 -------------




600 --
600 -------- ----------------.^ ^-----------------S--------------- ---------------------


S400
200 -------------------------- -----------------.------

200


-200 ------------------------------------------------------------------ -----
0


-200
0 100 200 300 400 500
Temperature, C

Figure 4.34: Stress versus temperature plot for a nickel thin film.








thermal cycling due to oxidation. An aluminum protective coating was deposited onto a

manganese sample, but did not inhibit oxidation of the sample.

Copper-Gold Intermetallics


After the copper-gold films had undergone the homogenization and ordering heat

treatment, their chemical composition was analyzed by EPMA (see Table 4.6). The films

were within 2 to 3 percent of the desired composition.

In powder x-ray diffraction all of the films showed a shifts in the diffraction peak

positions form disordered to the ordered state, and additional peaks were observed from

the long range order. As reported above, for textured thin films it was not possible to

calculate the degree of ordering.

The films were thermally cycled in the Flexus up to 450C in an argon atmo-

sphere. Figures 4.35 through 4.37 show the results of these tests. The CuAu films did

not show a change in stress near the order-disorder temperature, 4000C figure 4.35).

The CuAu film showed a marked change in stress at the transition temperature (figure

4.36) as did to a lesser degree the CuAu, film (figure 4.37). However, the order-disorder

transition occurs well above the temperature (300C) at which the films begin to plasti-

cally deform.


Table 4.4: Chemical composition of copper-gold thin films.


Atomic % Atomic %
Cu Al

Cu3Au 77 +/- 0.2 23 +/- 0.2

CuAu 53 +/- 0.4 47 +/- 0.5


CuAu, 22 +/- 0.4 78 +/- 0.4










500 --- ------------------

400
40 0 ----------------- --^ ---- ------- .--- .- ---- ^-- ^---- --- .-- -- .-- ---------------- I --------------.---.---
a
300 --------------------- ------- ------

S200-------------------------

10 0 ^- ^-^--- ^-------- --: --------------- .------- :^ ^----- ------------------ ---------------- -

0


-100
0 100 200 300 400 500
Temperature, C
Figure 4.35: Stress versus temperature plot for a Cu3Au thin film.
600

500-------------- -------

400

S 3 0 0 ------------------ -- -- ---- --------- .------------------ -------. ----------------------^ -----^-----------






0

-100
0 100 200 300 400 500
Temperature, C
Figure 4.36: Stress versus temperature plot for a CuAuthin film.







600
500 --- ---------------------------------------------------------------------------
500 --- -------- ---- -- -

0 O ------------ -- -- ^ -- ^ ^^ -- -- -

3 0 0 ---------. -------- -.----------- .. ---- ---. .--.---.-------- --------------- -- -- ---

^= 200 ---- ----- ------------ .-------- --------------------|-- -------------------4--- ------ .---- ---- .-- --



-100 ----------------------------------------------------------------
0
100

-100
0 100 200 300 400 500
Temperature, C
Figure 4.37: Stress versus temperature plot for a CuAu3 thin film.



325
320 5-------- -----------------------------..................................

320

0 3 15 .. -
Disordered CuAu
310 --... ..
305
^3 305
-4-


0,5 1.0 1.5
Time, days

Figure 4.38: Isothermal stress relaxation of ordered and disordered CuAu,
thin films at 150"C.








A series of isothermal stress relaxation test were performed on these samples.

These tests were conducted in argon since these films oxidize if heated to 150C in air.

The tests were run at 150C to maximize mobility in the samples. Figure 4.38 shows the

isothermal stress relaxation data for an ordered and a disordered CuAu, thin film. The

ordered film has a 2 percent reduction in stress over two days while the disordered film

has a 7 percent reduction in stress over the same period. The isothermal stress relaxation

in the CuAu and Cu3Au films showed no difference between the ordered and disordered

states.

Aluminum-Titanium Intermetallics

The as-deposited tensile stress in the layered AlTi film was 200 MPa. On the

first temperature cycle to 500C the slope of the stress temperature curve changed at

350"C, indicating a change in the structure of the film. At 500C the film was in a state of

no stress. When the film cooled to room temperature it was under a tensile stress of 800

MPa as shown in figure 4.39. The film was cycled up to 500C two additional times and

showed no hysteresis in the stress-temperature plot. An isothermal stress relaxation test

was performed on this sample at 350C and the film showed no stress relaxation over a 48

hour period.

The sample was then tested at Tencor up to 8000C. The results of this test are

shown in figure 4.39. The AlITi film begins to plastically deform above 500"C under a

compressive stress of 200 MPa. At 8000C the film still maintains a compressive stress of

300 MPa. Upon returning to room temperature the film was under a tensile stress of

nearly 1000 MPa. An isothermal stress relaxation test was also performed at 450C and

again the film showed no stress relaxation over a 48 hour period.

The native passivating oxide held up well for all tests at all temperatures. The

specular reflection of the sample was high after all the tests. There was no visible change

in the oxide after cycling to 500C. Figure 4.40 shows the Auger depth profile for an








1000


800 ... .


600 : "

C-


2 0 -" ---...-- -. -- -.





-400
4200 .. ....... ............. ^' .. .. .
*4000-----------i -- ***



0 100 200 300 400 500 600 700 800
Temperature C
Figure 4.39: Stress versus temperature for Al3Ti thin film.



Al Ti film that was heated to a maximum of 500oC. The oxygen concentration drops

sharply after 5 minutes of sputtering. Figure 4.41 shows the Auger depth profile for a

sample heated to 800"C. The oxygen concentration drops to near zero after 25 minutes of

sputtering, indicating that the surface is passivated, but the oxide layer is about three

times thicker for samples heated to 8000C are than samples heated only to 5000C.


Modeling the Effects of Oxide Thickness on a Bimetallic Actuator's Curvature


As has been seen with many of the materials examined in this study oxidation has

a significant impact of the performance of the material. For the aluminum alloys, it was

not possible to test for the effects of the oxide layer since one always forms when alumi-

num is exposed to air. To evaluate the effect of the oxide on the performance of a bime-

tallic strip, the change in curvature of a bimetallic strip as a function of the Al203 layer

thickness was calculated. This calculation is based on a bimetallic strip that consists of 5




























SPUTTER TIME, min.


Figure 4.40: Auger depth profile of AITi sample that was heated to 5000C.


0.0 5 10 15 20 25 30

SPUTTER TIME, min.

Figure 4.41: Auger depth profile of AI3Ti sample that was heated to 8000C.
gm of aluminum on 5 gm of silicon with an oxide whose thickness varied from 0 to 1

















% of Max
Curvature

40-



20



0 ---- I I
0 02 04 0.6 0.8


A1203 Thickness
microns

Figure 4.42: Curvature of a bimetallic strip as a function of oxide thickness as a percent-
age of maximum curvature.



pm. Figure 4.42 shows the results of this calculation. From this calculation it can be

seen that the 30 nm oxide layer found on the aluminum alloy samples reduces the curva-

ture of the bimetallic strip by 0.5 percent, but thicker oxides would have a more signifi-

cant impact. Therefore, controlling the oxide thickness is critical to maximizing the

performance of a bimetallic strip.














CHAPTER 5
DISCUSSION

Introduction



In this chapter the results and conclusions that have been drawn from this research

will be presented and discussed with respect to the three objectives of this research. (i)

effectiveness of strengthening techniques in thin films. (ii) identification of the mecha-

nism responsible for isothermal stress relaxation in thin films. (iii) identification of

materials for high temperature applications. In addition, to assist the development of

bimetallic actuators, a figure of merit has been developed and a number critical param-

eters have been identified to assist in the selection of materials for this application.

Thin Film Strengthening


Of the five strengthening techniques used in bulk metals, this study examines the

effectiveness of two of these, solid solution strengthening and multiphase hardening

(precipitation/age hardening) in thin films. While not normally considered a strengthen-

ing technique, the effectiveness of ordered intermetallics to resist stress relaxation was

also examined. Solid solution and precipitation/age hardening were selected because of

their effectiveness in strengthening bulk metals and because these mechanisms are rela-

tively insensitive to small variation in composition, which improves manufacturability.

The order/disorder transition that occurs in some intermetallics was also of interest due

to the large change in volume that frequently occurs at this transition. The order / disor-

der transition provided a means of evaluating the effectiveness of the ordered phase to

resist stress relaxation.







82
With respect to measuring the strengthening effects of these techniques, there are

a number of parameters that are of interest. With bulk materials, strengthening is usually

evaluated by measuring the yield strength or ultimate tensile strength of the material at a

temperature. In thin film bimetallic structures, this is not possible since stress and tem-

perature are dependent on one another. Thus the temperature and stress at which plastic

deformation begins are interrelated, and is the point on the stress versus temperature

curve where the slope significantly departs from linear. This study is also concerned with

the stress relaxation that occurs in thin films, which determines the stability of a bimetal-

lic actuator. So the strengthening techniques must be evaluated with respect to their

effectiveness at increasing both the temperature and therefore the stress at which plastic

deformation occurs and at reducing isothermal stress relaxation in the film.

While it would be reasonable to assume that the bulk strengthening techniques

would be effective in thin films, the AlSiCu alloy used in microelectronics would suggest

that the strengthen mechanisms are less effective in thin films. In the AlSiCu alloy,

silicon could provide some solid solution strengthening, but the solubility of silicon in

aluminum at room temperature is very low. Copper could provide precipitation strength-

ening, but due to processing is generally in the overaged condition. As shown in Figure

2.9, the strength of the AISiCu alloy is only 30% greater than pure aluminum. In contrast,

age hardened bulk aluminum copper alloys can have yield strengthens of five times those

of pure aluminum. While this suggest that bulk strengthening mechanism may not be as

effective in thin films, it must also be noted that the AISiCu alloy was not developed for

strength but to reduce aluminum spiking into silicon wafer (Si) and to reduce

electromigration (Cu) [Bow74]. On the other hand, the alloys evaluated in this study

were developed for strength.

Solid Solution Strengthening


The aluminum alloy 5052 was selected to test solid solution strengthening. This

alloy is solid solution strengthened by the addition of magnesium. While the majority of








the magnesium stays in solution, some AIMg2 and AlIMg precipitates form. However,

these precipitates do not provide the primary strengthening. Comparing the stress tem-

perature curve for 5052 aluminum (Figure 4.16) with the curves for pure aluminum and

the AISiCu alloy (Figure 2.10) clearly shows that solid solution strengthening is effective

in thin films. The room temperature residual stress for the 5052 alloy is 300MPa as

compared to 150MPa for pure aluminum and 200MPa for the AISiCu alloy. Also plastic

deformation begins in the 5052 alloy sample at 225C under a compressive stress of

40MPa while the pure aluminum and AISiCu alloy begin to deform under the same stress

but at much lower temperatures of 1250C and 175C, respectively. Therefore solid solu-

tion strengthening has significantly increased the performance of the film.

Testing also showed that the 5052 alloy film is more resistant to stress relaxation

than a pure aluminum or AlSiCu films. At 500C, pure aluminum and the AlCuSi alloy

films experience a 17% reduction in stress over 48 hours while the 5052 alloy film only

experience a 6.7% reduction in stress over the same time period. This is even more

significant because the 5052 alloy film was under a tensile stress of 150 MPa while the

pure aluminum was under only a 80 MPa compressive stress.

In addition to the 5052 aluminum alloy the copper / copper gold systems were

examined. The copper gold system was examined primarily because it undergoes an

order disorder transition. However, in the disordered phase it can be considered a solid

solution of the two elements and it can be compared to pure copper. From (Figure 4.29)

the room temperature residual tensile stress for pure copper is 50 MPa and the films

began to plastically deform above 2500C under a compressive stress of 300 MPa. The

stress temperature plot for CuAu (Figure 4.35) shows a residual room temperature tensile

stress in the film of 575 MPa and the film begins to plastically deform at 300C under a

compressive stress of 50 MPa. As with aluminum, solid solutions of copper and gold

nearly doubled the strength of the film. Stress relaxation tests were not done on the pure

copper films due to oxidation.








Given the results of the 5052 aluminum and copper gold, it is clear that solid

solutions are effective in strengthening thin films. It has also been shown that this

strengthening mechanism is effective at reducing stress relaxation in aluminum alloys.

Precipitation and Multiphase Hardening



The T201 aluminum alloy was selected to determine the effectiveness of precipi-

tation hardening to reduce stress relaxation. The T201 aluminum alloy is not strength-

ened only by precipitation hardening. Magnesium and manganese are added for solid

solution strengthening and titanium is added to reduce grain size. Copper is the primary

alloying element and it leads to strengthening by precipitates. Silver is added to stabilize

the copper precipitates. Table 4.2 shows the complete composition of the T201 alloy.

Since this alloy is precipitation and solid solution strengthened, comparisons must be

made to pure aluminum, AISiCu alloy, and the 5052 alloy to evaluate the effectiveness of

the precipitation strengthening.

Figure 4.1 shows the stress-temperature plot for the T201 alloy. The room tem-

perature residual stress is nearly 400 MPa, which is 100 MPa higher than the 5052 alloy.

(Figure 4.16), 2.5 times greater than pure aluminum and 2 times greater than the AlSiCu

alloy (Figure 2.10). The T201 alloy began to plastically deform at 2250C under a com-

pressive stress of 75 MPa, again much higher than pure aluminum or the other aluminum

alloys. At 50oC the T201 alloy thin film undergoes a 12% reduction in stress over 48

hours (Figure 4.2) as compared to 7% for the 5052 samples (Figure 4.17) and a reported

value of 15% for the pure aluminum and AlSiCu alloy. However, at 500C, the T201 film

is under a tensile stress of 390 MPa as compared to 270 MPa for the 5052 alloy films, 75

MPa for the pure aluminum film and 150 MPa for the AlSiCu alloy film. At 125C, the

T201 sample undergoes a 2% stress relaxation over 48 hours under a 114 MPa tensile

stress (Figure 4.5) while the 5052 alloy film undergoes an 18% stress relaxation under a








104 MPa tensile stress (Figure 4.20). Therefore, precipitation/age hardening is effective

at reducing stress relaxation in thin films.

The plane view TEM photomicrograph in Figure 4.11 verifies that precipitates

were present in the aluminum thin film. In this image, precipitates are clearly seen within

aluminum crystals and the diffraction pattern for CuAI2 is seen. The size of these precipi-

tates is on the order of 20nm to 30nm which is appropriate for age hardening [Bro82].

Based on the testing and microstructural examination of the sample it can be

concluded that precipitation/age hardening is effective in thin films. It should also be

noted that the addition of silver to this alloy stabilizes the CuAlI precipitates. Without

this stabilization, the copper would precipitate at the grain boundaries, have a coarse

dispersion rather than a fine dispersion, and be much less effective at strengthening the

film. A coarse dispersion of "overaged" precipitates is the reason for lack of strength in

the AISiCu alloy [Bro82].

Ordered Phases and the Order Disorder Transition


While not normally considered a strengthen technique, ordered phases generally

have a higher strength than their disordered counterparts. As mentioned above, order

disorder transitions were also of interest due to a large change in volume that occurs

during this transition. This volume change could generate large stress or deflection

values in the design of thermally actuated MEMS. Figures 4.37, 4.38 and 4.39 show the

stress temperature curves for the three copper gold compounds tested. While there is a

substantial change in stress seen near the order disorder temperature, this occurs beyond

the point at which the yield strength / temperature conditions lead to plastic deformation

for these films and the order/disorder transition is not of use in this system.

Ordered phases were presumed to have a lower stress relaxation rate than disor-

dered phases since the self diffusion rate, which determines the creep rate and stress

relaxation rate for many materials, were expected to be lower in an ordered phases versus








a disordered phase [Cah83]. Figure 4.38 clearly shows that this presumption was true, for

Cu3Au, the ordered phase showed a 1.6% reduction in stress over two days compared

with a 6.5% reduction in stress for the disordered phase. The disordered film's stress

relaxation rate is almost four time faster than the ordered phase, while the stress is only

8% higher on the disordered film.. The AlITi ordered intermetallic also showed no stress

relaxation at 4500C in two days.

Stress Relaxation Mechanism in Thin Films


While increasing the strength of the aluminum thin films reduced the stress

relaxation rates and magnitudes, it did not eliminate stress relaxation. Thus the mecha-

nisms responsible for stress relaxation in thin films needs to be discussed. In this study

three modes of stress relaxation have been observed. The first mode is plastic deforma-

tion of the thin film [Her85]. If the stress applied by the substrate is greater than the yield

strength of the material, the film will plastically deform and relax the stress. This is seen

as a change in slope of the stress-temperature curves upon heating (Figures 4.1, 4.16,

4.35, 4.36, 4.37, 4.39). The second mode of stress relaxation occurs over several tens of

minutes after a sample has reached temperature. This is seen in the stress relaxation

curves of the T201 and 5052 aluminum alloy films (Figures 4.2-4.4 and 4.18-4.20). The

third mode of stress relaxation occurs over a number of days and can again be seen in the

stress relaxation plots of the T201 and 5052 samples.

The challenge in determining the stress relaxation mechanisms in thin films on a

substrate is that film stress and temperature can not be independently varied. This elimi-

nates the use of an Arrhenius equation to calculate the activation energies of the processes

under investigation. It is possible to manipulate samples at different temperatures to the

same stress, however the microstructure and dislocation array of a material are not state

functions. Therefore, their structures will be quite different at different temperatures even

if the stress is the same. In addition, removing the film from the substrate would, in many








cases, substantially change the film in a number of ways. For self passivating materials,

such as examined in this study, the back side of the film would oxidize upon exposure to

air. It has been shown that dislocations interact differently with oxidized surfaces

[Nix89]. It would also be difficult to create and apply a biaxial stress to a free standing

thin film and to measure strain in the film. Therefore, direct determination of an activa-

tion energy and thus the mechanisms responsible for the different modes of stress relax-

ation is not possible. Other data are needed to determine the mechanisms responsible for

stress relaxation in thin films.

It should be noted that the change in total strain and temperature to reduce the

stress in these films is very small. A change in stress of 10 MPa is roughly equivalent to

a temperature change of only five degrees. For aluminum on silicon, this is a strain of

only 0.00012. Thus very small changes in the dimensions of a film can result in large

changes in the stress in the film.

The inability to calculate activation energies for stress relaxation modes 2 and 3

makes it very difficult to identify their mechanisms. All plausible relaxation mechanism

must be considered including: recovery recrystallization, logarithmic creep, Andrade

creep, bulk diffusion, gain boundary migration, grain boundary sliding, surface to grain

boundary diffusion, polygonization/subgrain coalescence, and precipitate formation or

coarsening. Several of these mechanism can be eliminated as possible causes of both

modes 2 and 3 stress relaxation.

For the aluminum alloys at 150C, the T, is less than 0.5, which is slightly lower

than the temperature at which Andrade creep would be expected. In addition, none of the

stages of Andrade creep would produce an exponential decay as is seen in the stress

relaxation data. Samples heated at higher rates showed no transition from stage 1 creep

to stage 2 creep. In the early stages of relaxation the stress relaxation rate decrease, but

does not go to zero, and the stage I strain rate is not a simple exponential [Cah83] decay








as seen in the data. Stage 2, steady state creep has a constant strain rate that would

produce a constant stress relaxation rate that is not seen in the data. The strain rate in

stage 2 creep has some dependence of the applied stress, but this relationship is not a

simple exponential [Cah83]. In the third stage of Andrade creep the stain rate increases

and so is not consistent that data that has been collected. Therefore, Andrade creep does

not appear to be responsible for the stress relaxation.

Diffusional creep processes are generally significant at temperatures of, TH > 0.6,

which is well above the temperatures at which the current stress relaxation is seen (TH <

0.5). Diffusional process increase exponentially with temperature [Cah83] while the

present stress relaxation shows an inverse relationship to temperature. Diffusional pro-

cesses should result in a zero stress at infinite time instead of the finite residual stress

projected from the furve fits (Tables 4.1 and 4.3).

X-ray line shape analysis did not detected polygonization or subgrain coalescence,

so these mechanism can be eliminated. In addition, subgrain coalescence [Cah83] is a

preliminary step to recovery and recrystallization, and the microstructure of the samples

did not change during stress relaxation (Figures 4.7 and 4.12). Thus recovery, recrystalli-

zation and grain boundary migration do not appear to be responsible for the stress relax-

ation. The formation or coalescence of precipitates does not appear to be a factor. X-ray

diffraction showed no significant change in precipitate concentration over time. In

addition, the same two modes of stress relaxation have been seen in pure aluminum

[Her85]. Having eliminated these potential mechanism of stress relaxation it is necessary

to separately consider the mechanism applicable to each mode.

Mode 2 Stress Relaxation


Mode 2 stress relaxation occurs relatively quickly. It is virtually complete in two

hours. During this time there is a measurable increase in the surface roughness of the

film (Table 4.3). There is also a 0.10 shift in the (111) x-ray diffraction peak. The in-








crease in surface roughness equates to a 0.15% change in the thickness of the 1.2p.m film

which is of the same order as the change in the d spacing calculated from the x-ray peak

shift of 0.1 This indicates that the processes affecting the surface is occurring throughout

the thickness of the film. Surface to grain boundary diffusion is not the likely cause of

this stress relaxation since diffusional processes are temperature sensitive as discussed

above [Ask89], while the roughness and peak shifts are temperature insensitive. Grain

boundary sliding is also not consistent with the results obtained since it normally has a

very high activation energy [Mey84]. A high activation energy is not consistent with a

process that occurs rapidly at low temperature. The film also has a columnar grain

structure [Kra90, Tho74] so the resolved shear stress on most of the grain boundaries will

be very small. If some crystals have a tilted grain boundary, these grains should be

visible with the AFM. Therefore, grain boundary sliding is eliminated as a potential

explanation of stress relaxation.

The only mechanism not yet eliminated to explain stress relaxation is logarithmic

creep which results from dislocation glide on slip systems. It is postulated that mode 2

stress relaxation is caused by the movement of dislocation on slip planes that terminate as

the surface of the film. Dislocation moving in these slip systems should not encounter

dislocation pile ups as would be encountered in slip system that terminate at grain bound-

aries. Thus these dislocations should move and be eliminated quicklt. The degree of

relaxation should only be limited by the number of sources of mobile dislocations. The

increase in roughness could be caused by dislocation bands. Also because of the stress

level and temperature of these films, the DMM's predict that logarithmic creep would be

the dominant process (Figure 2.5) [Cah83].

Mode 3 Stress Relaxation


The mechanisms not yet eliminated for mode 3 stress relaxation are grain bound-

ary sliding, surface to grain boundary diffusion, and logarithmic creep. Surface to grain








boundary diffusion is a diffusional process that should be exponentially dependent upon

temperature. It could increase surface roughness as atoms migrate from the surface to the

grain boundary. However, the data do not show a large temperature dependence and the

surface roughness does not increase over stage 3. The temperature at which the stress

relaxation is occurring is also low for a diffusional process to be active (TH < 0.5). There-

fore, surface to grain boundary diffusion does not appear to be the mechanism responsible

for this mode of stress relaxation. Grain boundary sliding also not likely to be the mecha-

nism responsible for this stress relaxation, since grain boundary sliding should also

increase surface roughness. In addition, the activation energy for grain boundary sliding

is very high and it normally only occurs just before mechanical failure. Furthermore, in a

columnar structure the resolved stress on most grain boundary should be very low. This

only leaves logarithmic creep, which was concluded to be the mechanism for mode 2

stress relaxation. However, how can the same mechanism have a different rate constants?

Different resolved stresses on different slip directions could explain this in part, but, the

slip systems are all equivalent within each of the (100),(110) and (ll1) oriented crystals.

For this to be the cause of change in relaxation rate, the different oriented crystals would

have to relax at different rates, and this should be noticeable in the AFM images. Also,

no increase in the roughness of the film was seen in mode 3. It is known that disloca-

tions pile up at grain boundaries [Mey84]. Dislocation climb could move some of these

dislocations into the grain boundary. This would result in a reduction of stress and no

increase in surface roughness. Therefore this mechanism is consistent with the observa-

tion made in this study and is postulated to be the origin of the mode 3 stress relaxation.

High Temperature Application


Micro actuators are needed that can operate in severe environments, such as inside

jet engines. In these severe environments oxidation was assumed to be a major concern.








High temperature application also require materials that maintain their strength at the

operating temperatures. These topics will be examined in this section.

Oxidation


During this study it was found that oxidation is a pervasive problem affecting all

temperatures ranges. As is shown above and in Appendix C, a 30nm thick oxide forming

of the surface of a Sm Al 5pm Si bimetallic strip reduces the curvature of the strip by

0.5%. The 2090 aluminum lithium alloy was found to be unsuitable for bimetallic actua-

tors because the lithium increased the oxidation rate (Figure 4.29). The other elements

tested also behaved poorly. At elevated temperatures, copper, titanium, manganese, and

nickel all oxidized (Figures 4.31, 4.32, 4.34, 4.35). In bulk metals application the thick-

ness of the passivating oxides is of little concern as long it remains intact. In thin films

the passivating oxide can consume a large fraction of the thickness. This is very damag-

ing to a bimetallic actuator because most oxides have a very low CTE which will reduce

the displacement of the device. Thus, a material's oxidation rate is a critical parameter

and those metals with the slowest oxidation rate are best suited for this application.

Aluminum and silicon both have very low oxidation rates due to the formation of passi-

vating oxides and both are well suited for elevated temperature applications [Wes95]. A

thin aluminum layer was also found to be effective at protecting less resistant materials,

such as pure copper.

Strength at Elevated Temperatures


As discussed above, intermetallics have a high resistance to creep due to their low

self diffusion rates. This is of interest for elevated temperature application where a lower

melting point intermetallic with a higher CTE could be used in place of a standard higher

melting point alloy with a low CTE. The Al1Ti intermetallic performed well, showing no







92
stress relaxation at 450C and no oxidation. Unfortunately nickel and copper oxidized, so

comparisons of stress relaxation could not be made.

Figure of Merit


To assist engineers who design bimetallic actuators in selecting the active layer

material, a figure of merit has been developed. The figure of merit is designed to provide

an initial evaluation of the type of material that should be used, i.e. aluminum or nickel.

The figure of merit, FOM, is expressed in equation 5.1 as:



FOM = (MHT -M i Y'S. [(FD CTE 10 'C) + ( E 101 Pa'. (1 FD))] (5.1)

-- W -- --- --P 1---W--
Temperature compensated Displacement Force
yield strength mode mode


where MHT is the maximum homologous temperature (the homologous temperature

above which the yield strength of the material rapidly declines, equal to 0.4 for most

metals and 0.65 for most intermetallics), MOT is the maximum operating temperature of

the actuator, YS is the yield strength of the material at room temperature, CTE is the

coefficient of thermal expansion for the active layer material, E is Young's modulus for

the active layer material and FD, determines whether the priority of the device is force or

displacement (0 to maximize force, and zero to maximize displacement). Table 5.1

shows a comparison of the different figures of merit for different materials at different

operating temperatures and for force or displacement. Note that the values for tin are

negative, indicating it is not suitable at the temperatures. In general the large CTE of

aluminum alloys give the FOMs of ~ 108 for displacement while the higher Young's

modulus and moderate CTE's of nickel and stainless steal give them FOMs of -101"-10"

for force. Note the FOM at 500oC of 1015 for cobalt. This figure of merit is complex,

but could be expanded to also include the environmental stability of the materials.










Table5.1: Figure of merit ratings for different materials at different operating tempera-
tures and optimized for displacement, D, or force, F.

Max. Temp. 100 (-C) 250 ('C) 500 ("C)
Force/
ce D F D F D F
Displacment

Sn -5.8*106 -3.4*107 -2.9*106 -1.7*107 -1.5*106 -9.1*106

5052 Al 1.4*10' 8.8*109 -2.8*10' 1.7*109 -9.3*106 -5.7*10'

T210 Al 1.1*108 1.2*10 -2.4*107 -2.7*109 -8.3*106 -9.0*10'

Mg 8.1*10' 8.0*10' -3.2*107 -3.2*108 -9.6*106 -9.5*10'

Anneal Al 2.8*107 1.3*109 -1.3*107 -6.0*108 -3.7*106 -1.7*10'

75% c.w. A 5.0*107 2.4*109 -2.3*107 -1.1*109 -6.6*106 -3.1*10'

Cu 7.5*106 1.4*1010 2.1*10' 3.8*101' -1.1*106 -2.0*10'1

Ni 1.0*106 6.1*101 1.7*106 9.7*107" 1.1*10' 6.6*1012

Co 4.2*106 5.4*1013 6.5*106 8.3*10" 7.8*107 1.0*101"

410 S.S. 2.4*10' 2.3*10" 3.6*105 3.4*10" 2.6*106 2.5*10"

c.w. cold worked
S.S. stainless steal




Full Text

PAGE 1

7+,1 ),/0 0(7$//,=$7,21 )25 0,&52%,0(7$//,& $&78$7256 %\ -21$7+$1 )5$1. *255(// $ ',66(57$7,21 35(6(17(' 72 7+( *5$'8$7( 6&+22/ 2) 7+( 81,9(56,7< 2) )/25,'$ ,1 3$57,$/ )8/),//0(17 2) 7+( 5(48,5(0(176 )25 7+( '(*5(( 2) '2&725 2) 3+,/2623+< 81,9(56,7< 2) )/25,'$

PAGE 2

$&.12:/('*0(176 ZRXOG OLNH WR H[SUHVV P\ DSSUHFLDWLRQ WR DGYLVRU 'U 3DXO + +ROORZD\ IRU KLV VXSSRUW JXLGDQFH DQG HQFRXUDJHPHQW ZRXOG DOVR OLNH WR JLYH VSHFLDO WKDQNV IRU SURn YLGLQJ D QHZ SHUVSHFWLYH WR VRPH RI WKH SUREOHPV HQFRXQWHUHG LQ WKLV UHVHDUFK 7KDQNV DUH DOVR GXH WR P\ GRFWRUDO FRPPLWWHH PHPEHUV 'U 'H+RII 'U 6LQJK 'U $GDLU DQG 'U .XPDU IRU WKHLU DVVLVWDQFH DQG LQWHUHVW LQ P\ ZRUN 7KLV ZRUN ZRXOG QRW KDYH EHHQ SRVVLEOH ZLWKRXW WKH VXSSRUW RI (*t* ,&6HQVRUV DQG WKH '$53$ 0(06 UHVHDUFK SURJUDP DP DOVR JUDWHIXO WR 'U +DO -HUPDQ IRU KLV HIIRUWV LQ REWDLQLQJ WKLV JUDQW DQG KLV VXSSRUW RI P\ ZRUN ZRXOG DOVR OLNH WR WKDQN P\ IULHQGV DQG FRZRUNHUV IRU WKHLU VXSSRUW DQG HQFRXUn DJHPHQW DQG ZKR PDGH P\ UHWXUQ WR VFKRRO PRVW HQMR\DEOH ZRXOG OLNH WR WKDQN 0DUN 'DYLGVRQ IRU SURYLGLQJ PH ZLWK KLV LQVLJKW LQWR WKH ZRUNLQJV RI D XQLYHUVLW\ DQG VFLHQn WLILF LQVWUXPHQWDWLRQ /DVWO\ ZRXOG OLNH WR WKDQN P\ SDUHQWV -DPHV DQG %LOOLH *RUUHOO IRU LQVWLOOLQJ D ORYH RI OHDUQLQJ DQG WKH EHOLHI LQ P\VHOI WKDW FDQ DFKLHYH WKH JRDOV VHW IRU P\VHOI IURP D YHU\ HDUO\ DJH :LWKRXW WKHVH YDOXHV ZRXOG KDYH QHYHU WULHG PXFK OHVV KDYH FRPn SOHWHG WKLV SURJUDP

PAGE 3

7$%/( 2) &217(176 $&.12:/('0(176 L $%675$&7 YL &+$37(56 ,1752'8&7,21 0RWLYDWLRQ DQG 2EMHFWLYH 6FRSH RI 3UHVHQW :RUN /,7(5$785( 5(9,(: ,QWURGXFWLRQ 7KH %LPHWDOOLF 6WULS 7KH 6LPSOH %LPHWDOOLF &DQWLOHYHU 7KH %LPHWDOOLF 'LVN 7KH (IIHFWV RI 3ODVWLF 'HIRUPDWLRQ $WRPLF %RQGLQJ 3ODVWLF 'HIRUPDWLRQ ,Q 0HWDOV &UHHS LQ 0HWDOV 'HIRUPDWLRQ 0HFKDQLVP 0DSV '00Vf $QQHDOLQJ 6WUHQJWKHQLQJ RI %XON 0HWDOV 2UGHUHG ,QWHUPHWDOOLF &RPSRXQGV 7KLQ )LOPV (QYLURPHQWDO 6WDELOLW\ RI 0DWHULDOV (;3(5,0(17$/ 352&('85( ,QWURGXFWLRQ 7KLQ )LOP 'HSRVLWLRQ &RSSHU *ROG +HDW 7UHDWPHQW 6WUHVV DQG 6WUHVV 5HOD[DWLRQ 0HDVXUHPHQWV 3RZGHU ;5D\ 'LIIUDFWLRQ ;5'f $WRPLF )RUFH 0LFURVFRS\ $)0f 6FDQQLQJ (OHFWURQ 0LFURVFRS\ 6(0f LLL

PAGE 4

(OHFWURQ3UREH 0LFURDQDO\VLV (30$f 7UDQVPLVLQ (OHFWURQ 0LFURVFRS\ 7(0f (WFKLQJ $XJHU (OHFWURQ 6SHFWURVFRS\ $(6f &XUYH )LWWLQJ 1XPHULFDO 0RGHOLQJ 5(68/76 ,QWURGXFWLRQ 7 $OXPLQXP $OXPLQXP $OXPLQXP &RSSHU 7LWDQLXP 0DQJDQHVH DQG 1LFNHO &RSSHU*ROG ,QWHUPHWDOOLFV $OXPLQXP7LWDQLXP ,QWHUPHWDOOLFV 0RGHOLQJ WKH (IIHFWV RI 2[LGH 7KLFNQHVV RQ D %LPHWDOOLF $FWXDWRUfV &XUYDWXUH ',6&866,21 ,QWURGXFWLRQ 7KLQ )LOP 6WUHQJWKHQLQJ 6ROLG 6ROXWLRQ 6WUHQJWKHQLQJ 3UHFLSLWDWLRQ DQG 0XOWLSKDVH +DUGHQLQJ 2UGHUHG 3KDVHV DQG WKH 2UGHU 'LVRUGHU 7UDQVLWLRQ 6WUHVV 5HOD[DWLRQ 0HFKDQLVP LQ 7KLQ )LOPV 0RGH 6WUHVV 5HOD[DWLRQ 0RGH 6WUHVV 5HOD[DWLRQ +LJK 7HPSHUDWXUH $SSOLFDWLRQ 2[LGDWLRQ 6WUHQJWK DW (OHYDWHG 7HPSHUDWXUHV )LJXUH RI 0HULW &21&/86,216 )8785( :25. $33(1',&(6 $ '(5,9$7,21 2) 5(/$7,216+,3 %(7:((1 <281*f6 02'8/86 $1' &2()),&,(17 2) 7+(50$/ (;3$16,21 LY

PAGE 5

% &859( ),77,1* 2) 675(66 5(/$;$7,21 '$7$ & &$/&8/$7,21 2) &+$1*( ,1 &859$785( 2) $ %,0(7$//,& 675,3 %$6(' 21 2;,'( 7+,&.1(66 0(&+$1,&$/ 3523(57,(6 2) (;$0,1(' 0(7$/6 %,%/,2*5$3+< %,2*5$3+,&$/ 6.(7&+ Y

PAGE 6

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n VWUXFWXUH RI WKHVH ILOPV ZHUH HYDOXDWHG ZLWK VFDQQLQJ HOHFWURQ PLFURVFRS\ WUDQVPLVVLRQ HOHFWURQ PLFURVFRS\ $XJHU HOHFWURQ VSHFWURVFRS\ HOHFWURQ SUREH PLFURDQDO\VLV ;UD\ GLIIUDFWLRQ DQG OLQH VKDSH DQDO\VLV DQG DWRPLF IRUFH PLFURVFRS\ %LPHWDOOLF DFWXDWRU PD\ EH RSHUDWHG WR JHQHUDWH HLWKHU IRUFH RU GLVSODFHPHQW 7KH GLVSODFHPHQW PRGH LV GRPLQDWHG E\ WKH FRHIILFLHQW RI WKHUPDO H[SDQVLRQ ZKLOH WKH IRUFH PRGH LV D IXQFWLRQ RI ERWK
PAGE 7

WKH ILUVW KHDWLQJ %DVHG RQ LVRWKHUPDO VWUHVV UHOD[DWLRQ GDWD DQG FKDQJHV LQ WKH PLFURn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r& SODVWLFDOO\ GHIRUPHG RQO\ DERYH r& DQG KDG OLPLWHG R[LGDn WLRQ XS WR r& 1LFNHO FRSSHU WLWDQLXP DQG PDQJDQHVH ILOPV DOO R[LGL]HG RQ WKHLU ILUVW KHDWLQJ WR r& 7KH FRSSHU ILOP DOVR R[LGL]HG DW r& RYHU KRXUV &DOFXODWLRQV DOVR VKRZHG WKDW WKH SDVVLYDWLRQ R[LGH RQ DOXPLQXP DOOR\V FDQ VLJQLILFDQWO\ UHGXFH SUHIRUPDQFH D ELPHWDOOLF DFWXDWRU 7KXV R[LGDWLRQ UHVLVWDQFH LV D VLJQLILFDQW UHTXLUHPHQW IRU PDWHULDOV IRU WKHUPDO DFWXDWLRQ YLL

PAGE 8

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n FKDQLFDO GHYLFHV 5HVHDUFKHUV KDYH EHJXQ WR PLQLDWXUL]H HOHFWURPHFKDQLFDO GHYLFHV XVLQJ WKH WHFKQRORJLHV GHYHORSHG IRU PLFURHOHFWURQLFV >+RJ@ 7KHVH QHZ PLFUR VHQVRUV DQG DFWXDWRUV KDYH FUHDWHG D QHZ FODVV RI GHYLFHV FDOOHG 0(06 PLFURHOHFWURPHFKDQLFDO V\VWHPV >0DV )UD $QJ %HD 0DG@ ,QWHJUDWLRQ RI 0(06 VHQVRUV DQG DFWXDWRUV LV HQDEOLQJ WKH GHYHORSPHQW RI PHFKDQLFDO FKHPLFDO DQG ELRORJLFDO fVPDUW V\VWHPVf WKDW DUH DEOH WR LQWHUDFW ZLWK WKH SK\VLFDO ZRUOG 0(06 KDYH DOUHDG\ EHFRPH D VLJQLILFDQW LQGXVWU\ 0DV@ 7KLV PDUNHW LV FXUUHQWO\ GRPLQDWHG E\ SUHVVXUH VHQVRUV DQG DFFHOHURPHWHUV 7KH PLQLDWXUL]DWLRQ RI SUHVVXUH VHQVRUV KDV UHGXFHG VHQVRU FRVW WR WKH SRLQW WKDW LQ PHGLFDO DSSOLFDWLRQV WKH VHQVRUV FDQ EH GLVFDUGHG DIWHU D VLQJOH XVH $FFHOHURPHWHUV DUH SUHGRPLQDWHO\ XVHG WR GHWHUPLQH ZKHQ WR GHSOR\ DXWRPRELOH DLU EDJV 0LFUR DFWXDWRUV KDYH \HW PDGH DV PDQ\ LQURDGV LQWR LQGXVWU\ &XUUHQWO\ WKH ELJJHVW DSSOLFDWLRQ RI 0(06 DFWXDWRUV LV IRU LQN MHW SULQW KHDGV

PAGE 9

)LJXUH 6(0 LPDJH RI DQ (*t* ,& 6HQVRUV ELPHWDOOLF WKHUPDOO\ DFWXDWHG PLFURYDOYH >52:@ 7KH PDUNHW RI 0(06 DFWXDWRUV ZLOO FRQWLQXH WR GHYHORS DV DFWXDWRUV DUH QHHGHG WR SURGXFH D WUXO\ fVPDUW V\VWHPf WKDW FDQ QRW RQO\ PRQLWRU SURFHVVHV EXW DOVR FRQWURO WKHP +RZHYHU DGGLWLRQDO UHVHDUFK DQG GHYHORSPHQW DUH QHHGHG WR LPSURYH WKH SHUIRUPDQFH RI PLFUR DFWXDWRUV DQG WKLV LV WKH JHQHUDO IRFXV RI WKLV ZRUN 0LFUR DFWXDWRUV KDYH EHHQ GHVLJQHG EDVHG RQ D QXPEHU RI SK\VLFDO SURSHUWLHV LQFOXGLQJ HOHFWURVWDWLF HOHFWURPDJQHWLF SLH]RHOHFWULF PDJQHWRVWULFWLRQ DQG WKHUPDO DFWLYDWLRQ ELPHWDOOLF VKDSH PHPRU\ DOOR\V DQG WKHUPRSQHXPDWLFf >0DG@ 7DEOH VKRZV D FRPSDULVRQ RI WKH GLIIHUHQW DFWXDWLRQ WHFKQLTXHV DQG WKHLU GLIIHUHQFHV LQ SHUIRUn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t* ,&6HQVRUV SURGXFHG E\

PAGE 10

7DEOH &RPSDULVRQ RI DFWXDWLRQ SULQFLSOHV >%HO@ (IIHFW 3DUDPHWHU %LPHWDOOLF (OHFWURn VWDWLF 3LH]Rn HOHFWULF (OHFWURn PDJQHWLF 7KHUPR K\GUDXOLF )RUFH SHU PP 1 +LJK f /RZ OD ff +LJK +LJK +LJK 'HIOHFWLRQ PP +LJK f /RZ r Df /RZ +LJK +LJK :RUN SHU F\FOH SHU PP +LJK f /RZ rf +LJK +LJK )UHTXHQF\ UDQJH .K] OLPLWLQJ IDFWRUf /RZ +HDW 7UDQVIHUf 0HFK 5HVRQDW IUHTXHQF\f +LJK 0HFK 5HVRQDW IUHTXHQF\f /RZ +HDW 7UDQVIHU ?IROWDJH 9 /RZ f +LJK f +LJK /RZ /RZ &XUUHQW P$ +LJK f /RZ r f /RZ +LJK +LJK 3RZHU &RQVXPSWLRQ : +LJK nf /RZ O2/O2f /RZ +LJK +LJK (IILFLHQF\ b 6L]H /HQJWK PP :LGWK PP 6PDOO rOf 6PDOO ?f 6PDOO /DUJH /DUJH /DUJH $GGLWLRQDO UHTXLUHPHQWV +LJK ?EOWDJH +LJK ?EO WDJH 0DWHULDO &RPSDWLELOLW\f 0DJQHWLF ILHOG PDJQHWVf 0HGLD VHSDUDWLRQ &RVW 2: /RZ +LJK +LJK +LJK DOXPLQXP PHWDOOL]DWLRQ RQ VLOLFRQ (YHQ WKRXJK VLOLFRQ LV QRW D WUXH PHWDO WKLV VWUXFWXUH LV VWLOO UHIHUUHG WR DV D ELPHWDOOLF DFWXDWRU 7KH REMHFWLYH RI WKLV UHVHDUFK LV WR LPSURYH WKH SHUIRUPDQFH RI PLFURELPHWDOOLF DFWXDWRUV DQG EURDGHQ WKHLU UDQJH RI RSHUDWLQJ SDUDPHWHUV VWUHVV GLVSODFHPHQW WHPSHUDn WXUH UDQJH DQG VWUHVV UHOD[DWLRQf E\ LGHQWLI\LQJ EHWWHU RU LPSURYH WKH PDWHULDOV RI FRQn VWUXFWLRQ ,Q RUGHU WR XQGHUVWDQG KRZ WR LPSURYH WKH SHUIRUPDQFH RI WKH ELPHWDOOLF DFWXDWRU LW LV LPSRUWDQW WR XQGHUVWDQG KRZ WKH GHYLFH ZRUNV $ ELPHWDOOLF DFWXDWRU LV PDGH E\ ERQGLQJ WRJHWKHU WZR GLIIHUHQW PDWHULDOV WKDW KDYH D ODUJH GLIIHUHQFH LQ WKHUPDO H[SDQVLRQ VXFK DV DOXPLQXP DQG VLOLFRQ WR FUHDWH D ELPHWDOOLF VWULS RU VSULQJ WKDW LV KHDWHG RU FRROHG 8SRQ KHDWLQJ WKH PDWHULDO ZLWK WKH ODUJHU FRHIILFLHQW RI WKHUPDO H[SDQ

PAGE 11

$OXPLQXP $FWLYH /D\HU r& r& )LJXUH 'LDJUDP RI D ELPHWDOOLF VWULS VKRZLQJ KRZ WKH GLIIHUHQFHLQ WKHUPDO H[SDQVLRQ EHZWHHQ WKH DFWLYH OD\HU DOXPLQXP DQG SDVVLYH OD\HU VLOLFRQ FDXVHV WKH ELPHWDOOLF VWULS WR EHQG VLRQ FDOOHG WKH DFWLYH OD\HU H[SDQGV PRUH WKDQ WKH PDWHULDO ZLWK WKH VPDOOHU FRHIILFLHQW FDOOHG WKH SDVVLYH OD\HU VHH )LJXUH f $V WKHVH WZR PDWHULDOV DUH ERQGHG WRJHWKHU D VKHDU VWUHVV GHYHORSV EHWZHHQ WKH OD\HUV GXH WR WKH GLIIHUHQW UDWHV RI H[SDQVLRQ SURGXFLQJ D IRUFH WKDW FDXVHV WKH VWUXFWXUH WR EHQG 8SRQ FRROLQJ WKH UHYHUVH SURFHVVHV RFFXUV DQG WKH ELPHWDOOLF VWULS UHWXUQV WR LWV RULJLQDO SRVLWLRQ RU GHIOHFWV LQ WKH RSSRVLWH GLUHFWLRQ ZLWK FRQWLQXHG FRROLQJ 7KH JHRPHWU\ PD\ EH D GLVN FDQWLOHYHUHG EHDP RU PRUH FRPn SOH[ JHRPHWU\ :KLOH WKH FXUUHQW JHQHUDWLRQ RI ELPHWDOOLF DFWXDWRUV ZRUNV ZHOO IRU D QXPEHU RI DSSOLFDWLRQV WKHLU XVH FRXOG EH VLJQLILFDQWO\ H[SDQGHG E\ LPSURYLQJ WKH SRVLWLRQ DQG IRUFH VWDELOLW\ RI WKHVH DFWXDWRUV DQG LQFUHDVLQJ WKH RSHUDEOH DPELHQW WHPSHUDWXUH OLPLW IURP r& WR r& )RUFH DQG SRVLWLRQ VWDELOLW\ KDYH EHHQ IRXQG WR EH D SUREOHP LQ

PAGE 12

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n ULO\ IURP FUHHS LQ WKH DFWLYH PHWDO OD\HU RI WKH ELPHWDOOLF DFWXDWRUV 7KXV LPSURYHPHQWV FDQ EH DFKLHYHG WKURXJK WKH LGHQWLILFDWLRQ DQG RU GHYHORSPHQW RI EHWWHU PHWDO WKLQ ILOPV

PAGE 13

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n WXUDO IDLOXUH DQG HQYLURQPHQWDO VWDELOLW\ DUH FRQFHUQV LQ DQ\ PHFKDQLFDO GHYLFH $OVR WKH FRQWLQXHG UHGXFWLRQ RI LQWHUFRQQHFW ZLGWKV LQ PLFURHOHFWURQLF LQWHJUDWHG FLUFXLWV OHDG WR JUHDWHU VWUHVVHV LQ WKLQ PHWDO OLQHV WKDW PD\ IDLO GXH WR VWUHVV LQGXFHG YRLGLQJ 6,9f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f 7KH ILUVW REMHFWLYH LV WR GHWHUPLQH WKH HIIHFWLYHQHVV RI GLIIHUHQW VWUHQJWKHQLQJ WHFKQLTXHV IRU WKLQ

PAGE 14

ILOP PHWDOOL]DWLRQ 7KHUH DUH D QXPEHU RI WHFKQLTXHV XVHG WR LQFUHDVH WKH \LHOG DQG XOWLPDWH WHQVLOH VWUHQJWK RI EXON PHWDO DOOR\V EXW WKHLU HIIHFWLYHQHVV LQ WKLQ ILOPV KDV \HW WR EH GHWHUPLQHG ,QFUHDVHG \LHOG VWUHQJWKV RI WKH PHWDO WKLQ ILOPV ZRXOG LQFUHDVH WKH PD[LPXP IRUFH DQG GLVSODFHPHQW SURGXFHG E\ DQ DFWXDWRU DQG FRXOG LQFUHDVH IRUFH DQG GLVSODFHPHQW VWDELOLW\ LLf 7KH VHFRQG REMHFWLYH LV WR HYDOXDWH XVH RI PDWHULDOV DW HOn HYDWHG WHPSHUDWXUHV 7KH SULPDU\ UHTXLUHPHQWV IRU HOHYDWHG WHPSHUDWXUH DSSOLFDWLRQV DUH KLJK VWUHQJWK RYHU WKHU RSHUDWLRQ WHPSHUDWXUH UDQJH DQG UHVLVWDQFH WR FUHHS DQG R[LGDn WLRQ LLLf 7KH WKLUG REMHFWLYH LV WR LGHQWLI\ WKH PHFKDQLVP UHVSRQVLEOH IRU LVRWKHUPDO VWUHVV UHOD[DWLRQ LQ WKLQ ILOPV ,W LV WKLV UHOD[DWLRQ WKDW FDXVHV WKH IRUFH DQG SRVLWLRQ LQVWDELOLW\ VHHQ RYHU H[WHQGHG WLPH SHULRGV 7R PHHW WKHVH REMHFWLYHV HOHYHQ GLIIHUHQW PDWHULDOV ZHUH GHSRVLWHG DQG HYDOXDWHG 7KH HYDOXDWLRQ IRFXVHG RQ GHWHUPLQLQJ WKH PDWHULDOfV VWUHQJWK HQYLURQPHQWDO VWDELOLW\ DQG UDWH RI VWUHVV UHOD[DWLRQ )RU WKRVH PDWHULDOV WKDW VKRZHG SURPLVH DQG WR XQGHUVWDQG WKH XQGHUO\LQJ PHFKDQLVPV UHVSRQVLEOH IRU WKH VWUHVV UHOD[DWLRQ WKH PRUSKRORJ\ DQG PLFURVWUXFWXUH RI VHYHUDO VDPSOHV ZHUH H[DPLQHG 7KH PDWHULDOV WKDW ZHUH HYDOXDWHG IDOO LQWR WKUHH JURXSV Lf DOXPLQXP DOOR\V LLf HOHPHQWDO PHWDOV DQG LLLf LQWHUPHWDOOLFV 7KH DOXPLQXP DOOR\V ZHUH XVHG WR LGHQWLI\ WKH PHFKDQLVPV UHVSRQVLEOH IRU LVRWKHUPDO VWUHVV UHOD[DWLRQ DQG WR GHWHUPLQH WKH HIIHFWLYHQHVV RI WKH GLIIHUHQW VWUHQJWKHQLQJ WHFKQLTXHV 7KH HOHPHQWDO PHWDOV DQG LQWHUPHWDOOLFV ZHUH HYDOXDWHG DW HOHYDWHG WHPSHUDWXUHV 6RPH RI WKH LQWHUPHWDOOLFV DOVR H[KLELWHG RUGHULQJ DQG QRQOLQHDU WKHUPDO SURSHUWLHV 7KH HYDOXDWLRQ RI VHYHUDO PDWHULDOV V\VWHPV DOORZHG D ZLGH YDULDWLRQ LQ SURSHUWLHV DQG WKHUHn IRUH SURYLGHG VLJQLILFDQW LQVLJKW LQWR WKH PDWHULDOV UHTXLUHPHQWV IRU 0(06 GHYLFHV 7KH RUJDQL]DWLRQ RI WKLV GLVVHUWDWLRQ LV DV IROORZV &KDSWHU LV D UHYLHZ RI WKH OLWHUDWXUH SHUWLQHQW WR WKLV VWXG\ 7KH H[SHULPHQWDO SURFHGXUHV XVHG WR GHSRVLW WHVW DQG DQDO\]H WKH VDPSOHV DUH SUHVHQWHG LQ FKDSWHU &KDSWHU FRQWDLQV WKH UHVXOWV RI WKHVH WHVWV DQG DQDO\VHV 7KH UHVXOWV DUH GLVFXVVHG LQ FKDSWHU 7KH FRQFOXVLRQV DUH VXPPDUL]HG LQ FKDSWHU DQG VXJJHVWLRQV IRU IXWXUH ZRUN DUH SUHVHQWHG LQ FKDSWHU

PAGE 15

&+$37(5 /,7(5$785( 5(9,(: ,QWURGXFWLRQ 7KLQ ILOP PHWDOOL]DWLRQ IRU ELPHWDOOLF DFWXDWRUV DUH WKH IRFXV RI WKLV UHVHDUFK WKHUHIRUH D QXPEHU RI WRSLFV QHHG WR EH UHYLHZHG 7KH ILUVW WRSLF LV WKH WKHRU\ RI WKH ELPHWDOOLF VWULS ,Q H[DPLQLQJ OLWHUDWXUH WKH FRQFHUQ LV WR GHWHUPLQH KRZ WKH GLIIHUHQW PDWHULDOV SURSHUWLHV DIIHFW WKH SHUIRUPDQFH RI D ELPHWDOOLF VWULS :LWK DQ XQGHUVWDQGLQJ RI WKH HIIHFWV RI PDWHULDOV SURSHUWLHV LPSURYHG PDWHULDOV VHOHFWLRQ FDQ EH XVHG WR RSWLn PL]H SHUIRUPDQFH $IWHU WKHLU LGHQWLILFDWLRQ WKH RULJLQ DQG LQWHUGHSHQGHQFH RI WKHVH SURSHUWLHV FDQ EH GLVFXVVHG XVLQJ WKH WKHRULHV RI DWRPLF ERQGLQJ DQG WKHUPDO H[SDQVLRQ ,Q DGGLWLRQ WKHUH DUH D QXPEHU RI SURFHVVHV WKDW PD\ DIIHFW WKH PHWDOOL]DWLRQ RI D ELPHWDOn OLF VWULS VXFK DV FUHHS SODVWLF GHIRUPDWLRQ DQG R[LGDWLRQ /DVWO\ WKHVH ELPHWDOOLF DFWXDWRUV DUH PDQXIDFWXUHG ZLWK WKLQ ILOP WHFKQRORJ\ DQG WKH HIIHFWV RI WKHVH SURFHVVHV DUH GLVFXVVHG 7KHUH LV QR SHUIHFW PDWHULDO IRU XVH LQ DOO ELPHWDOOLF DFWXDWRUV VLQFH GLIIHUHQW DSSOLFDWLRQV UHTXLUH GLIIHUHQW SURSHUWLHV :LWK DQ XQGHUVWDQGLQJ RI WKH LQWHUGHn SHQGHQFH RI WKH GLIIHUHQW PDWHULDOV SURSHUWLHV WKH WUDGHRIIV WKDW DUH QHHGHG IRU DQ\ JLYHQ DSSOLFDWLRQ FDQ EH EHWWHU VHOHFWHG 7KH %LPHWDOOLF 6WULS 7KH IRUPXODWLRQ RI WKH ELPHWDOOLF VWULS SUREOHP ZDV ILUVW UHSRUWHG LQ SDUW E\ *HUDOG 6WRQH\ >6WR@ 6WRQH\ GHWHUPLQHG WKH UHODWLRQVKLS EHWZHHQ FXUYDWXUH DQG VWUHVV LQ D ELPHWDOOLF VWULS 6 7LPRVKHQNR >7LP@ WKHQ GHULYHG WKH HTXDWLRQV IRU GLVSODFHPHQW

PAGE 16

DQG IRUFH RI D ELPHWDOOLF VWULS YHUVXV FKDQJH LQ WHPSHUDWXUH /DVWO\ 7RZQVHQG >7RZ@ JHQHUDOL]HG WKH HTXDWLRQV IRU D PXOWLOD\HU VWUXFWXUH 7KH IRUPXODWLRQ DQG VROXWLRQ RI WKH ELPHWDOOLF VWULS SUREOHP DUH GHSHQGHQW RQ WKH JHRPHWU\ DQG ERXQGDU\ FRQGLWLRQV XVHG WKHUHIRUH QR VLQJOH VROXWLRQ LV JHQHUDOO\ DSSOLn FDEOH 5RDUNfV )RUPXODV IRU 6WUHVV DQG 6WUDLQ >%DU@ >7VD@ 7KH 6LPSOH %LPHWDOOLF &DQWLOHYHU 7R GHWHUPLQH WKH HIIHFWV RI PDWHULDOV SURSHUWLHV RQ SHUIRUPDQFH WKH VLPSOH ELPHWDOOLF FDQWLOHYHU VLPSO\ VXSSRUWHG RQ RQO\ RQH HQG LV H[DPLQHG 7KH VLPSOLI\LQJ DVVXPSWLRQV PDGH LQ WKLV FDVH DUH WKDW WKH OHQJWK LV PXFK JUHDWHU WKDQ WKH ZLGWK RU WKLFNn QHVV WKH PDWHULDOV DUH SHUIHFWO\ HODVWLF DQG WKH PDWHULDO UHWXUQV WR LWV RULJLQDO GLPHQVLRQ RQFH WKH ORDG LV UHPRYHG >0H\@ )LJXUH VKRZV D GLDJUDP RI WKLV FDVH ZKHUH ( DQG ( DUH WKH 7LP@ N DDf$7fO Af E LQ )RU WKH FDVH RI ERWK OD\HUV RI WKH VWULS EHLQJ WKH VDPH WKLFNQHVV D E DQG HTXDO PRGXOL WKH HTXDWLRQ IRU FXUYDWXUH UHGXFHV WR >7LP@

PAGE 17

N /D\HU ( /D\HU D ( / /D\HU D A /D\HU E )LJXUH 'LDJUDP RI ELPHWDOOLF VWULS f 7KH HUURU LQGXFHG E\ DVVXPLQJ WKH PRGXOL DUH HTXDO LV RQO\ WKUHH SHUFHQW IRU D IDFWRU RI WZR FKDQJH LQ HLWKHU PDWHULDOfV PRGXOXV 7KHUHIRUH WKH SULPDU\ PDWHULDOfV SURSHUW\ DIIHFWLQJ WKH GLVSODFHPHQW LQ WKLV FDVH LV WKH GLIIHUHQFH LQ WKH FRHIILFLHQWV RI WKHUPDO H[SDQVLRQ 7KH HTXDWLRQ IRU WKH UHVWUDLQLQJ IRUFH 5 WKDW LV QHHGHG WR NHHS D ELPHWDOOLF VWULS IURP PRYLQJ IRU D JLYHQ WHPSHUDWXUH FKDQJH '7 LV WKH PD[LPXP IRUFH D ELPHWDOOLF VWULS FDQ SURGXFH 7KLV UHVWUDLQLQJ IRUFH LV JLYHQ E\ HTXDWLRQ f K/ 7KXV WKH IRUFH SURGXFHG E\ D ELPHWDOOLF VWULS LV D IXQFWLRQ RI ERWK WKH WKHUPDO H[SDQVLRQ DQG WKH
PAGE 18

OL ([SDQVLRQ &RHIILFLHQW ;(r&f )LJXUH (TXDO SRWHQWLDO OLQHV IRU IRUFH DQG GLVSODFHPHQW IRU D ELPHWDOOLF GLVN >-HU@ WKH ELPHWDOOLF GLVN WKH HYDOXDWLRQ ZDV SHUIRUPHG E\ FDOFXODWLQJ DQG SORWWLQJ WKH HTXDOSRWHQWLDO OLQHV IRU PD[LPXP GLVSODFHPHQW FXUYDWXUHf DQG IRUFH RI D ELPHWDOOLF GLVN DV D IXQFWLRQ RI -HU@ 7KLV PRGHO LV EDVHG RQ D ELPHWDOOLF VWULS ZKHUH WKH SDVVLYH OD\HU LV VLOLFRQ DQG ERWK WKH DFWLYH DQG SDVVLYH OD\HUV DUH SP WKLFN ,Q DJUHHPHQW ZLWK HTXDWLRQV f DQG f WKLV ILJXUH VKRZV WKDW GLVSODFHPHQW LV SUHGRPLQDWHO\ OLPLWHG E\ WKH WKHUPDO H[SDQVLRQ RI WKH DFWLYH OD\HU PDWHULDO 7KHUH LV VRPH GHSHQGHQFH RQ WKH
PAGE 19

DQG IRU D VLQJOH IL[HG FDQWLOHYHU WKH WUHQGV DQG OLPLWLQJ PDWHULDOV SURSHUWLHV DUH WKH VDPH 'LVSODFHPHQW LQ D ELPHWDOOLF VWUXFWXUH LV SULPDULO\ OLPLWHG E\ WKH WKHUPDO H[SDQn VLRQ RI WKH DFWLYH OD\HU PDWHULDO ,Q FRQWUDVW WKH IRUFH JHQHUDWHG E\ D ELPHWDOOLF VWULS LV GHSHQGHQW RQ ERWK WKH WKHUPDO H[SDQVLRQ DQG 0H\@VHH )LJXUH f 3ODVWLF GHIRUPDWLRQ LQ HLWKHU OD\HU RI D ELPHWDOOLF GHYLFH LQYDOLGDWHV WKH UHODWLRQVKLSV LQ HTXDWLRQV f f DQG f DQG ZRXOG UHGXFH WKH IRUFH RU GLVSODFHPHQW SURGXFHG IURP WKH VWULS 7KH HIIHFWV RI SODVWLF GHIRUPD )LJXUH 7\SLFDO VWUHVV VWUDLQ FXUYH IRU D GXFWLOH PHWDO

PAGE 20

WLRQ FDQ EH PRGHOHG DV D QRQOLQHDU UHGXFWLRQ RI WKH WKHUPDO H[SDQVLRQ RU $VN 0H\@ VHH )LJXUH f LV DOVR D FULWLFDO PDWHULDOV SDUDPHWHU $ PDWHULDOfV \LHOG VWUHQJWK GHWHUPLQHV WKH PD[LPXP IRUFH LW FDQ H[HUW RQ DQRWKHU OD\HU DQG LQ WXUQ PD\ GHWHUPLQH WKH PD[LPXP IRUFH RU GLVSODFHPHQW D ELPHWDOOLF VWULS FDQ DFKLHYH 7KHUHIRUH WKH PDWHULDOV SURSHUWLHV RI LQWHUHVW GHSHQG XSRQ ZKHWKHU WKH DSSOLFDn WLRQ UHTXLUHV IRUFH RU GLVSODFHPHQW ,Q WKH GLVSODFHPHQW PRGH WKH PDWHULDOfV WKHUPDO H[SDQVLRQ DQG \LHOG VWUHQJWK DUH WKH SULPDU\ OLPLWLQJ SURSHUWLHV ,Q WKH IRUFH PRGH WKHUPDO H[SDQVLRQ DQG $VN@ 7KH GHSWK RI WKLV SRWHQWLDO ZHOO GHWHUPLQHV WKH VWUHQJWK RI WKH ERQG 7KH GHHSHU WKH SRWHQWLDO ZHOO WKH VWURQJHU WKH ERQG $ GHHS SRWHQWLDO ZHOO QRUn PDOO\ KDV D ODUJHU G(G[ ZKLFK LV SURSRUWLRQDO WR WKH
PAGE 21

WKH DWRPV 7KLV LQFUHDVH LQ DWRPLF YLEUDWLRQ FDXVHV WKH DWRPV WR PRYH XS LQ WKH SRWHQWLDO ZHOO $V WKH DWRPV PRYH KLJKHU WKHLU WLPHDYHUDJHG SRVLWLRQ LQFUHDVHV VHH )LJXUH f DQG WKH PDWHULDO H[SDQGV >7RX@ 7KH GHHSHU WKH SRWHQWLDO ZHOO WKH VPDOOHU WKH FKDQJH LQ WKH WLPHDYHUDJHG SRVLWLRQ RI WKH DWRPV IRU D JLYHQ LQFUHDVH LQ WHPSHUDWXUH DQG D ORZHU WKHUPDO H[SDQVLRQ LV REVHUYHG 7KH UDWH RI WKHUPDO H[SDQVLRQ LV IUHTXHQWO\ H[n SUHVVHG DV WKH FRHIILFLHQW RI WKHUPDO H[SDQVLRQ &7( $ PDWHULDOfV &7( LV WKH GHULYDWLYH RI LWV WKHUPDO H[SDQVLRQ FXUYH $// YHUVXV WHPSHUDWXUH 7R D ILUVW DSSUR[LPDWLRQ WKH WKHUPDO H[SDQVLRQ LV LQYHUVHO\ SURSRUWLRQDO WR WKH DWRPLF ERQG VWUHQJWK 7KXV WKH
PAGE 22

WKH PDWHULDO WKH QXPEHU RI QHDUHVW QHLJKERUV WKH FRRUGLQDWLRQ QXPEHU DQG WKH ERQG OHQJWK *UDSKLWH LV D JRRG H[DPSOH WR LOOXVWUDWH WKH HIIHFWV RI GLUHFWLRQDOLW\ RI ERQGLQJ *UDSKLWH KDV D OD\HUHG VWUXFWXUH ZKHUH WKH ERQGV ZLWKLQ D OD\HU DUH VWURQJ FRYDOHQW ERQGV EXW WKH ERQGV EHWZHHQ DWRPV LQ GLIIHUHQW OD\HUV DUH PXFK ZHDNHU YDQ GHU :DDOV ERQGV >(YD@ 7KLV GLIIHUHQFH LQ ERQGLQJ FDXVHV WKH FSODQH FRHIILFLHQW RI WKHUPDO H[SDQVLRQ &7(f WR EH [O2f& ZKLOH EHWZHHQ WKH SODQHV WKH &7( LV [O2nr&n ,Q PDQ\ LQVWDQFHV WKH GLUHFWLRQDOLW\ RI ERQGLQJ GRHV QRW FKDQJH WKH ERQG W\SH EXW VWLOO KDV D PDMRU DIIHFW RQ SK\VLFDO SURSHUWLHV =LQF LOOXVWUDWHV WKLV SRLQW ZHOO 7KH DWRPV LQ ]LQF DUH ERQGHG LQWR D KH[DJRQDO FORVH SDFNHG FU\VWDO VWUXFWXUH ZLWK PHWDOOLF ERQGV EHWZHHQ DOO DWRPV $ORQJ WKH FD[LV WKH &7( IRU ]LQF LV [Of2& EXW DORQJ WKH DD[LV LW LV [On2&n >7RX@ 7KXV WKHUH LV D IROG GLIIHUHQFH LQ &7( EHWZHHQ GLIIHUHQW SODQHV GXH WR WKH FU\VWDOORJUDSK\ WKDW LV GHWHUPLQHG E\ WKH ERQGLQJ 7KHUH DUH RWKHU PDWHULDOV ZLWK VPDOO DQLVRWURS\ EXW XQH[SHFWHG SK\VLFDO SURSHUn WLHV GXH WR XQXVXDO FU\VWDOORJUDSKLF VWUXFWXUH 0DQJDQHVH LV RQH VXFK PHWDO >'HD@ :KLOH DPDQJDQHVH KDV D FXELF VWUXFWXUH LW GRHV QRW KDYH D VLPSOH IDFH FHQWHUHG FXELF )&&f RU ERG\ FHQWHUHG FXELF %&&f FU\VWDO VWUXFWXUH >$VN@ 7KH XQLW FHOO LV EDVHG RQ WKH %&& VWUXFWXUH EXW FRQWDLQV DWRPV LQ SDLUV >'HD@ 7KH &7( IRU PDQJDQHVH LV [On2& DOPRVW GRXEOH ZKDW ZRXOG EH H[SHFWHG IRU D PDWHULDO ZLWK D PHOWLQJ WHPn SHUDWXUH RI r& ,Q FRPSDULVRQ EHU\OOLXP PHOWV DW r& DQG KDV D &7( RI RQO\ [ 2n r& > :HD @ 7KH GLUHFWLRQDOLW\ DQG FU\VWDOORJUDSKLF VWUXFWXUH RI D PDWHULDO DUH QRW DOZD\V VWDWLF 3XUH LURQ XQGHUJRHV WZR DOORWURSLF WUDQVIRUPDWLRQV EHWZHHQ VROLGLILFDWLRQ DW r& DQG URRP WHPSHUDWXUH ,URQ ILUVW VROLGLILHV LQWR D %&& VWUXFWXUH DW r& FRQYHUWV WR DQ )&& VWUXFWXUH DW r& DQG WKHQ FRQYHUWV EDFN WR D %&& VWUXFWXUH DW r& >.UD@ $W HDFK DOORWURSLF WUDQVIRUPDWLRQ WKHUH LV D VLJQLILFDQW FKDQJH LQ WKHUPDO H[SDQVLRQ DQG RWKHU SK\VLFDO SURSHUWLHV 'LVFRQWLQXLWLHV LQ D PDWHULDOfV &7( FRXOG EH XVHIXO LQ D ELPHWDOOLF DFWXDWRU 8QIRUWXQDWHO\ WKH WUDQVLWLRQV LQ LURQ DUH IDU DERYH WKH

PAGE 23

)LJXUH /LQHDU WKHUPDO H[SDQVLRQ IRU LQWHUPHWDOOLF FRPSRXQG &X$X DQG &X$X 7KH GLVFRQWLQXLW\ LQ WKHUPDO H[SDQVLRQ LV GXH WR WKH RUGHUGLVRUGHU WUDQVLRQ RI WKHVH FRPSRXQGV >7RX@ WHPSHUDWXUH FXUUHQWO\ EHLQJ FRQVLGHUHG IRU SUDFWLFDO GHYLFHV +RZHYHU WUDQVIRUPDWLRQV DUH XVHG LQ VKDSH PHPRU\ DOOR\ DFWXDWRUV >0DG@ $QRWKHU FKDQJH LQ ERQGLQJ WKDW FDQ VLJQLILFDQWO\ DIIHFW SURSHUWLHV LV WKH RUGHU GLVRUGHU WUDQVIRUPDWLRQ H[KLELWHG E\ VRPH LQWHUPHWDOOLF FRPSRXQGV >7RX@ ,QWHUPHWDOn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

PAGE 24

PDWHULDO PD\ XQGHUJR D VXEVWDQWLDO FRQWUDFWLRQ XSRQ RUGHULQJ VHH )LJXUH f 7KLV FRQWUDFWLRQ LV GXH WR WKH DELOLW\ RI WKH DWRPV WR SDFN WLJKWHU LQ WKH RUGHUHG SKDVHV GXH WR WKH VWURQJHU DWRPLF ERQGLQJ 7KLV WUDQVLWLRQ PD\ DOVR LQFUHDVH WKH PDWHULDOfV \LHOG VWUHQJWK UHGXFH WKH FUHHS UDWH DQG LQFUHDVH HOHFWULFDO FRQGXFWLYLW\ 3ODVWLF 'HIRUPDWLRQ ,Q 0HWDOV $V PHQWLRQHG DERYH SODVWLF GHIRUPDWLRQ LQ HLWKHU OD\HU RI D ELPHWDOOLF DFWXDWRU ZLOO OLPLW WKH IRUFH DQG GLVSODFHPHQW RI WKH DFWXDWRU 7KHUHIRUH WKH \LHOG VWUHQJWK GHWHUn PLQHV WKH PD[LPXP IRUFH DQG GLVSODFHPHQW RI DQ DFWXDWRU $WRPLF PRWLRQ EHFRPHV UDSLG DV WKH WHPSHUDWXUH RI D PHWDO H[FHHGV KDOI LWV PHOWLQJ SRLQW WKHUHIRUH WKH KRPRORn JRXV WHPSHUDWXUH 7+ LV GHILQHG WR EH WKH UDWLR RI WKH WHPSHUDWXUH WR WKH PHOWLQJ SRLQW RI WKH PDWHULDO LQ GHJUHHV .HOYLQ $W 7+ GLVORFDWLRQ PRWLRQ LV WKH GRPLQDQW PHFKDn QLVP RI SODVWLF GHIRUPDWLRQ ,QKLELWLRQ RI WKH PRWLRQ RI GLVORFDWLRQV ZRXOG VWUHQJWKHQ WKH PDWHULDO 7KLV LV LQ FRQWUDVW WR WKH WKHUPDO H[SDQVLRQ DQG 5HH &DK@ :LWK UHVSHFW WR GLVORFDWLRQ PRWLRQ WKH\ FDQ PRYH WKURXJK D FU\VWDO E\ HLWKHU GLVORFDWLRQ JOLGH RU FOLPE 'LVORFDWLRQ JOLGH RFFXUV ZKHQ D GLVORFDWLRQ PRYHV RQ D VOLS

PAGE 25

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f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n SHUDWXUHV GLVORFDWLRQV PD\ UDSLGO\ FOLPE 7KH SURFHVV RI GLVORFDWLRQ PXOWLSOLFDWLRQ FDQ RFFXU LQ PDQ\ ZD\V 2QH RI WKH PRUH VLJQLILFDQW LV WKH )UDQN5HDG VRXUFH >5HH@ E\ ZKLFK GLVORFDWLRQ JOLGH FDQ LQn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fZRUN KDUGHQLQJf >%UR@ 7KHUH JHQHUDOO\ H[LVWV DQ LQYHUVH UHODWLRQVKLS EHWZHHQ WHPSHUDWXUH DQG \LHOG VWUHVV LQGHSHQGHQW RI WKH PHFKDQLVPV UHVSRQVLEOH IRU WKH GHIRUPDWLRQ )RU GLVORFDWLRQ JOLGH UDQGRP WKHUPDO H[FLWDWLRQ KHOSV GLVORFDWLRQ RYHUFRPH WKH 3HLHUOV1DEDUUR EDUULHU >0H\@ DQG WKXV ORZHUV WKH \LHOG VWUHVV 7KH HIIHFWV RI WHPSHUDWXUH RQ GLVORFDWLRQ

PAGE 26

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n QLILFDQW /DVWO\ WKH FU\VWDOORJUDSKLF WH[WXUH ZKLFK LV WKH SUHIHUHQFH IRU RQH FU\VWDOORn JUDSKLF RULHQWDWLRQf RI D PDWHULDO DOVR DIIHFWV LWV \LHOG VWUHQJWK >0H\ %UR@ %DOGZLQ >%DO@ VKRZHG XS WR b FKDQJH LQ WHQVLOH \LHOG VWUHQJWK EDVHG RQ WKH GLUHFWLRQ RI WKH DSSOLHG VWUHVV UHODWLYH WR WKH WH[WXUH RI UROOHG FRSSHU VKHHW &UHHS LQ 0HWDOV ,Q WKH SUHFHGLQJ VHFWLRQ WKH PHFKDQLVPV UHVSRQVLEOH IRU WKH UDSLG SODVWLF GHIRUPDn WLRQ RI PDWHULDOV LQ UHVSRQVH WR DQ DSSOLHG VWUHVV ZHUH H[DPLQHG +RZHYHU PDWHULDOV FRQWLQXH WR SODVWLFDOO\ GHIRUP LQ UHVSRQVH WR DSSOLHG VWUHVVHV RYHU H[WHQGHG SHULRGV RI WLPH IURP PLQXWHV WR \HDUVf 7LPH GHSHQGHQW SODVWLF GHIRUPDWLRQ LV NQRZQ DV FUHHS ,Q WKLV VHFWLRQ WKH GLIIHUHQW W\SHV RI FUHHS DQG WKH PHFKDQLVPV UHVSRQVLEOH IRU WKH WLPH GHSHQGHQW SODVWLF GHIRUPDWLRQ ZLOO EH GLVFXVVHG )LJXUH >&DK@ VKRZV WKH IRXU PDMRU W\SHV RI FUHHS WKDW RFFXU LQ PDWHULDOV DV D IXQFWLRQ RI WKH KRPRORJRXV WHPSHUDWXUH YHUVXV WKH UHVROYHG VWUHVV D QRUPDOL]HG E\ WKH VKHDU PRGXOXV M[ $W VWUHVV OHYHOV EHORZ ZKLFK GLVORFDWLRQ FDQ PRYH 2FUVVS ff DQHODVWLF FUHHS RFFXUV ,Q DQHODVWLF FUHHS LQWHUVWLWLDO DWRPV PRYH WR LQWHUVWLWLDO VLWHV WKDW KDYH EHHQ HORQJDWHG E\ WKH DSSOLHG VWUHVV >&DK@ 7KLV PRYHPHQW RI LQWHUVWLWLDO DWRPV FUHDWHV D

PAGE 27

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n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

PAGE 28

GLVORFDWLRQV 7KLV ORZHU DFWLYDWLRQ HQHUJ\ IRU GLIIXVLRQ VOLJKWO\ UHGXFHV WKH WHPSHUDWXUH DW ZKLFK WKH FUHHS UDWH LV VLJQLILFDQW &REOH FUHHS VWLOO UHTXLUHV D 7+ >&DK@ WR EHFRPH DFWLYH DQG LV WKHUHIRUH QRW RI FRQFHUQ LQ WKLV VWXG\ 7KH WKLUG W\SH RI FUHHS LV ORZ WHPSHUDWXUH RU ORJDULWKPLF FUHHS ,Q WKLV PHFKDn QLVP GHIRUPDWLRQ RFFXUV WKURXJK GLVORFDWLRQ PXOWLSOLFDWLRQ JOLGH DQG FOLPE +RZHYHU WKH VRXUFHV IRU GLVORFDWLRQ PXOWLSOLFDWLRQ EHFRPH H[KDXVWHG 7KHUHIRUH WKH FUHHS UDWH VWDUWV DW VRPH LQLWLDO KLJKHU YDOYH DQG ORJDULWKPLFDOO\ DSSURDFKHV ]HUR DV WKH QXPEHU RI PRELOH GLVORFDWLRQV DSSURDFKHV ]HUR >&DK@ 7KH UDWH RI ORJDULWKPLF FUHHS LV QRW D IXQFWLRQ RI WKH DSSOLHG VWUHVV VLQFH WKH UDWH OLPLWLQJ VWHSV DUH GLVORFDWLRQ FOLPE RU PXOWLn SOLFDWLRQ ZKLFK DUH GULYHQ E\ UDQGRP WKHUPDO H[FLWDWLRQV 7KLV W\SH RI FUHHS LV W\SLFDOO\ REVHUYHG DW ORZ WHPSHUDWXUHV 7+ f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n WXUH DV GHVFULEHG EHORZ 6HFRQG VWDJH FUHHS LV D VWHDG\ VWDWH SURFHVV WKDW SURGXFHV D FRQVWDQW VWUDLQ UDWH ,Q WKH WKLUG VWDJH RI $QGUDGH FUHHS WKH VDPSOH EHJLQV WR QHFN JUDLQ ERXQGDU\ VOLGLQJ RFFXUV >0H\@ DQG YRLGV IRUP LQVLGH WKH VDPSOH 7KLV UHGXFHV WKH DUHD RYHU ZKLFK WKH IRUFH LV FDUULHG WKXV LQFUHDVLQJ WKH VWUHVV ZKLFK LQFUHDVHV WKH VWUDLQ UDWH XOWLPDWHO\ UHVXOWLQJ LQ IDLOXUH 7KLV W\SH RI FUHHS LV W\SLFDOO\ REVHUYHG IRU 7+

PAGE 29

'HIRUPDWLRQ 0HFKDQLVP 0DSV '00VO ,Q WKH SURFHHGLQJ VHFWLRQV VHYHUDO PHFKDQLVPV KDYH EHHQ GLVFXVVHG ZKLFK OHDG WR SODVWLF GHIRUPDWLRQ RI PHWDOV +)URVW >)UR@ KDV FUHDWHG GHIRUPDWLRQ PHFKDQLVP PDSV '00f UHODWLQJ VWUHVV WHPSHUDWXUH DQG GHIRUPDWLRQ PHFKDQLVP $OO '00V DUH EDVHG RQ QXPHULFDO PRGHOV GHVFULELQJ WKH GLIIHUHQW GHIRUPDWLRQ PHFKDQLVPV 7KLV PRGHO LV WKHQ XVHG WR JHQHUDWH D WZR GLPHQVLRQDO FRQWRXU SORWV WHPSHUDWXUH DQG VWUHVV D[LVf WKDW VKRZV ZKLFK GHIRUPDWLRQ PHFKDQLVPV DUH DFWLYH 7KHVH SORWV DUH XVHIXO LQ VHOHFWLQJ PDWHULDOV IRU D JLYHQ DSSOLFDWLRQ ,I WKH VWUHVV DQG WHPSHUDWXUH OHYHOV IRU D JLYHQ DSSOLFDWLRQ DUH NQRZQ WKHQ WKH GHIRUPDWLRQ PHFKDQLVPV H[SHFWHG IRU D JLYHQ DSSOLFDWLRQ FDQ EH SURMHFWHG 7KH SULPDU\ OLPLWDWLRQ WR WKH XVH RI '00V LV WKDW GHWDLOHG GDWD UHn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n LQJ $QQHDOLQJ LV WKH SURFHVV E\ ZKLFK WKH VWRUHG HQHUJ\ LQ FROG ZRUNHG PHWDO LV UHOHDVHG :KHQ D PHWDO LV FROG ZRUNHG WKDW LV SODVWLFDOO\ GHIRUPHG EHORZ WKH WHPSHUDn WXUH DW ZKLFK GLVORFDWLRQ FOLPE EHFRPHV VLJQLILFDQW SDUW RI WKH GHIRUPDWLRQ HQHUJ\ LV VWRUHG LQ WKH PHWDO DV GHIHFWV DQG ODWWLFH GLVWRUWLRQ >5HH@ +XQGUHGV RI -RXOHV SHU PROH FDQ EH VWRUHG LQ D PHWDO WKDW KDV EHHQ KHDYLO\ FROG ZRUNHG 7KH DQQHDOLQJ SURFHVV UH

PAGE 30

)LJXUH 'LVWULEXWLRQ RI GLVORFDWLRQV LQ D EHQW FU\VWDO Df EHIRUH SRO\JRQL]DWLRQ Ef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f 7KH SRO\JRQL]DWLRQ RI WKH GLVORFDWLRQ WDQJOH LQWR VXEJUDLQ ERXQGDULHV UHGXFHV WKH IUHH HQHUJ\ RI WKH V\VWHP 2UGHULQJ DOVR UHGXFHV WKH QXPEHU RI HOHFWURQ VFDWWHULQJ FHQWHUV UHVXOWLQJ LQ D ORZHU HOHFWULFDO UHVLVWLYLW\ 1R GLVORFDWLRQV DUH GHVWUR\HG WKHUHIRUH WKH \LHOG VWUHVV GRHV QRW FKDQJH

PAGE 31

3RO\JRQL]DWLRQ DQG VXEJUDLQ FRDOHVFHQFH DOVR OHDG WR WKH VHFRQG VWDJH RI DQQHDOn LQJ UHFU\VWDOOL]DWLRQ ,Q UHFU\VWDOOL]DWLRQ QHZ XQVWUHVVHG FU\VWDOV QXFOHDWH DQG JURZ WR UHSODFH WKH VWUDLQHG GLVRUGHUHG FROG ZRUNHG JUDLQV 2QH WKHRU\ RQ WKH IRUPDWLRQQXFOH DWLRQ RI WKHVH QHZ XQVWUDLQHG FU\VWDOV LV WKDW VXEJUDLQV FRDOHVFH WR IRUP KLJK DQJOH JUDLQ ERXQGDULHV OHDYLQJ DQ XQVWUDLQHG FU\VWDO LQ WKHLU ZDNH >&DK@ 7KH GLIIHUHQFH LQ HQHUn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fRYHU DJLQJf DV LW UHGXFHV WKH \LHOG VWUHQJWK RI WKH PDWHULDO 7KLV SURFHVV ZRXOG DOVR SURGXFH D UHOD[DWLRQ RI VWUHVV 6WUHQJWKHQLQJ RI %XON 0HWDOV ,Q WKH SUHFHGLQJ VHFWLRQVf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

PAGE 32

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n PRQO\ XVHG HQJLQHHULQJ PHWDOV WKH VROXELOLW\ DQG VWUHQJWKHQLQJ HIIHFWV RI PDQ\ DOOR\LQJ HOHPHQWV KDYH EHHQ GHWHUPLQHG H[SHULPHQWDOO\ 7KH 0HWDOV +DQGERRNV 9ROXPHV DQG >%DN @ FRQWDLQ PRVW RI WKLV LQIRUPDWLRQ )RU PRUH XQXVXDO PDWHULDOV WKH +XPH 5RWKHU\ >5HH@ UXOHV IRU VXEVWLWXWLRQDO VROLG VROXWLRQ VWUHQJWKHQLQJ DUH KHOSIXO 6ROLG VROXWLRQ VWUHQJWKHQLQJ KDV D QXPEHU RI SURSHUWLHV WKDW PDNH LW XQLTXH FRPSDUHG ZLWK WKH RWKHU VWUHQJWKHQLQJ PHFKDQLVPV )LUVW WKH VROXWH DWRPV DUH LQ HTXLn OLEULXP ZLWK WKH KRVW PDWUL[ DWRPV 7KXV WKHUH DUH QR GULYLQJ IRUFHV WR FKDQJH WKH DWRP GLVWULEXWLRQ LQ WKH PHWDO DQG VROLG VROXWLRQ VWUHQJWKHQHG DOOR\V DUH H[SHFWHG WR EH PRUH VWDEOH DW HOHYDWHG WHPSHUDWXUHV WKDQ DOOR\V XVLQJ VRPH RI WKH RWKHU VWUHQJWKHQLQJ WHFKn QLTXHV 7KLV LV RI SDUWLFXODU LPSRUWDQFH IRU WKLV VWXG\ VLQFH PDWHULDOV ZLOO EH XVHG DW HOHYDWHG WHPSHUDWXUHV 6HFRQGO\ D VROLG VROXWLRQ VWUHQJWKHQHG PDWHULDO LV H[SHFWHG WR EH KRPRJHQRXV 7KXV WKHUH LV OLPLWHG SRVVLELOLW\ WKDW HOHFWURO\WLF KDOI FHOOV ZRXOG FDXVH FRUURVLRQ 7KH VHFRQG VWUHQJWKHQLQJ PHFKDQLVP LV WR UHGXFH JUDLQ VL]H 'LVORFDWLRQV JHQHUn DOO\ SLOHXS DW WKH JUDLQ ERXQGDULHV DV WKH\ FDQQRW FURVV WKHP 7KH VWUHQJWK RI D PDWHULDO KDV EHHQ IRXQG WR EH SURSRUWLRQDO WR WKH LQYHUVH RI WKH VTXDUH URRW RI WKH DYHUDJH JUDLQ

PAGE 33

VL]H >0H\ 5HH@ 7KLV UHODWLRQVKLS LV FDOOHG WKH +DOO3HWFK UHODWLRQVKLS $ SRWHQWLDO OLPLWDWLRQ WR XVLQJ WKLV VWUHQJWKHQLQJ WHFKQLTXH LV WKDW JUDLQV FRDUVHQ DW HOHYDWHG WHPSHUDn WXUHV 7K f DQG WKLV JHQHUDOO\ UHGXFHV WKH VWUHQJWK RI WKH PDWHULDO 7KH WKLUG VWUHQJWKHQLQJ WHFKQLTXHV LV ZRUN KDUGHQLQJ :KHQ D PHWDO LV FROG ZRUNHG LH ZRUNHG EHORZ D WHPSHUDWXUH DW ZKLFK DQQHDOLQJ RFFXUVf WKH GLVORFDWLRQ GHQVLW\ LQFUHDVHV JUHDWO\ :LWK DQ LQFUHDVHG GHQVLW\ GLVORFDWLRQ ZLOO EH SLQQHG E\ RWKHU GLVORFDWLRQV ZKLFK OLPLWV GLVORFDWLRQ VOLS :RUN KDUGHQLQJ FDQ SURGXFH VHYHQ IROG LQFUHDVH LQ \LHOG VWUHQJWK IRU SXUH DOXPLQXP )XOO\ DQQHDOHG DOXPLQXP KDV D \LHOG VWUHQJWK RI 03D WKDW LQFUHDVHV WR 03D IRU b FROG ZRUN DQG 03D IRU b FROG ZRUN >%DN@ $ FRQFHUQ ZLWK ZRUN KDUGHQLQJ LV WKDW WKH PDWHULDO ZLOO VRIWHQ DV GLVORFDWLRQV FOLPE DW HOHYDWHG WHPSHUDWXUHV GXULQJ UHFU\VWDOOL]DWLRQ 7KH IRXUWK ZD\ WR VWUHQJWKHQ VRPH PDWHULDOV DUH E\ D PDUWHQVLWLF WUDQVIRUPDWLRQ >.UD@ 7KLV W\SH RI WUDQVIRUPDWLRQ RFFXUV RQO\ LQ D IHZ DOOR\V DQG FHUDPLFV ,W LV ZLGHO\ NQRZQ DQG XVHG LQ IHUURXV DOOR\V EXW GRHV RFFXU LQ VRPH FRSSHU DOXPLQXP DQG WLWDQLXP DOOR\V 7KH PDUWHQVLWLF WUDQVIRUPDWLRQ LV D GLIIXVLRQOHVV QRQHTXLOLEULXP SURn FHVV WKDW RFFXUV E\ D FRRUGLQDWHG VKHDU GLVSODFHPHQW RI DWRPV 7KLV WUDQVIRUPDWLRQ VWUHVVHV WKH PDWUL[ RI WKH PDWHULDO ZKLFK LQKLELWV GLVORFDWLRQ PRWLRQ 7KH XVH RI PDUWHQn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

PAGE 34

SUHFLSLWDWHV VKRXOG EH XQLIRUPO\ GLVWULEXWHG WKURXJKRXW WKH KRVW PDWUL[ 7KHUPDO VWDELOLW\ LV D FRQFHUQ DW HOHYDWHG WHPSHUDWXUHV VLQFH SUHFLSLWDWHV PD\ FRDUVHQ RYHU DJHf UHVXOWLQJ LQ D ZHDNHQHG PDWHULDO 7KH SUHVHQFH RI D VHFRQG SKDVH FDQ DOVR VWUHQJWKHQ D KRVW PDWUL[ E\ VWUDLQLQJ LW $ VHFRQG SKDVH ZLWK D FRKHUHQW LQWHUIDFH WR WKH KRVW PDWUL[ ZLOO LQFUHDVH WKH VWUDLQ LQ WKH PDWUL[ $ FRKHUHQW LQWHUIDFH GRHV QRW DQ DUUD\ RI GLVORFDWLRQ EHWZHHQ WKH SUHFLSLWDWH DQG WKH PDWUL[ >%UR@ 7KXV LQ DGGLWLRQ WR EORFNLQJ GLVORFDWLRQ PRWLRQ D FRKHUHQW SUHFLSLn WDWH VWUDLQV WKH PDWUL[ LQFUHDVLQJ WKH IRUFH QHHGHG WR PRYH D GLVORFDWLRQ 7KHUPDO VWDELOLW\ LV D FRQFHUQ DV SUHFLSLWDWHV PD\ JURZ E\ WKHUPDOO\ DFWLYDWHG GLIIXVLRQ DQG WKH FRKHUHQW LQWHUIDFH PD\ EH ORVW 7KH ORVV RI WKH FRKHUHQW LQWHUIDFH VXEVWDQWLDOO\ UHGXFHV WKH VWUHQJWK RI WKH DOOR\ 2UGHUHG ,QWHUPHWDOOLF &RPSRXQGV 2UGHUHG LQWHUPHWDOOLF FRPSRXQGV >&DK@ DUH RI LQWHUHVW LQ WKLV VWXG\ EHFDXVH WKH\ DUH QRWHG IRU KLJK VWUHQJWK DW ERWK URRP DQG HOHYDWHG WHPSHUDWXUHV ZLWK JRRG UHVLVn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n SOH[ FRPSRVLWLRQ FUHDWHV D PRUH FRPSOH[ FU\VWDOORJUDSKLF VWUXFWXUH ZKLFK LQ WXUQ PDNH GLVORFDWLRQ PRYHPHQW PRUH GLIILFXOW 7KHUH DUH VHYHUDO UHDVRQV IRU WKLV )LUVW WKH QXPn EHU RI DWRPLF GLVWDQFHV D GLVORFDWLRQ KDV WR PRYH WR UHDFK DQ HTXLYDOHQW FU\VWDOORJUDSKLF VLWH LV PXFK JUHDWHU ,Q D VLPSOH $% LQWHUPHWDOOLF D GLVORFDWLRQ KDV WR PRYH WZR DWRPLF GLVWDQFHV DV FRPSDUHG ZLWK RQH DWRPLF GLVWDQFH IRU D QRQRUGHUHG PHWDO ,Q PDQ\

PAGE 35

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f ZKLFK WKHQ KDYH WR EH EURNHQ WR IRUP WKH QHZ $% ERQG 7KLV SURFHVV UHTXLUHV PRUH HQHUJ\ DQG WKXV LQWHUPHWDOOLFV KDYH KLJK \LHOG VWUHQJWKV 7KH PRUH FRPSOH[ VWUXFWXUH DOVR LQKLELWV FUHHS 7KH VHOI GLIIXVLRQ UDWH LQ LQWHUn PHWDOOLFV LV XQXVXDOO\ ORZ >3RS@ GXH WR WKH ODUJH GLVWDQFH EHWZHHQ HTXLYDOHQW VLWH DQG D KLJK DFWLYDWLRQ HQHUJ\ IRU WKLV SURFHVV $V WKH FUHHS UDWH IRU D PDWHULDO LV SURSRUWLRQDO WR WKH UDWH RI VHOI GLIIXVLRQ >&DK@ LQWHUPHWDOOLFV KDYH D ORZ FUHHS UDWH 6RPH LQWHUPHWDOOLFV DUH DOVR UHVLVWDQW WR R[LGDWLRQ 2[LGDWLRQ UHVLVWDQW LQWHUPHWDOn OLFV IRUP DQ DGKHUHQW SDVVLYDWLQJ R[LGH VXFK DV $ RQ FRPSRXQGV RI DOXPLQXP QLFNHO DQG WLWDQLXP $ FRQFHUQ LQ XVLQJ LQWHUPHWDOOLFV LQ ELPHWDOOLF DFWXDWRUV LV WKHLU VHQVLWLYLW\ WR FRPSRVLWLRQ IOXFWXDWLRQV ,Q $,1L DV OLWWOH DV b FKDQJH LQ WKH FRPSRVLWLRQ RI WKH LQWHUPHWDOOLF FDQ FDXVH GUDPDWLF FKDQJHV LQ WKH VWUHQJWK RI WKH PHWDO >/LP@ 7KHUHIRUH WKH FRPSRVLWLRQDO VWDELOLW\ RI DQ\ LQWHUPHWDOOLF FRPSRXQG VKRXOG EH WHVWHG EHIRUH VHOHFn WLRQ 7KLQ )LOPV 7KLQ ILOPV KDYH D QXPEHU RI XQLTXH SURSHUWLHV YHUVXV WKHLU EXON FRXQWHUSDUWV 6LQVH WKH PDWHULDOV H[DPLQHG LQ WKLV VWXG\ DUH XVH DV WKLQ ILOPV LQ ELPHWDOOLF DFWXDWRUV DQ XQGHUVWDQGLQJ RI WKHLU SURSHUWLHV LV QHHGHG 7KHUH DUH WKUHH SULPDU\ UHDVRQV IRU WKH XQLTXH SURSHUWLHV RI WKLQ ILOPV )LUVW WKLQ ILOPV DUH GHSRVLWHG RQWR D VXEVWUDWH UHVXOWLQJ

PAGE 36

LQ VXEVWUDWHWKLQ ILOP LQWHUDFWLRQV 6HFRQG WKH VXUIDFH DUHDWRYROXPH UDWLR IRU D WKLQ ILOP LV RUGHUV RI PDJQLWXGH KLJKHU WKDQ IRU D EXON PDWHULDO VR VXUIDFH HQHUJLHV DQG NLQHWLFV SOD\ D PXFK JUHDWHU UROH /DVW WKH WHFKQLTXHV XVHG WR SURGXFH WKLQ ILOPV DUH YHU\ PXFK GLIIHUHQW IURP WKH WHFKQLTXHV XVHG LQ WKH SURFHVVLQJ RI EXON PHWDOV ZKLFK DIIHFWV WKHLU PLFURVWUXFWXUH 7KHUH DUH DW OHDVW WKUHH FRPPRQ FDWHJRULHV RI WKLQ ILOP GHSRVLWLRQ SURFHVVHV FKHPLFDO YDSRU GHSRVLWLRQ &9'f SK\VLFDO YDSRU GHSRVLWLRQ 39'f DQG HOHFWURSODWLQJ >2KU@ ,Q WKLV VWXG\ ZH KDYH RQO\ XVHG SK\VLFDO YDSRU GHSRVLWLRQ PHWKRGV FRQVLVWLQJ RI VSXWWHULQJ RU WKHUPDO HYDSRUDWLRQ ,Q DOO SK\VLFDO YDSRU GHSRVLWLRQ 39'f SURFHVVHV D YDSRU RI WKH VRXUFH PDWHULDO LV FUHDWHG DQG FRQGHQVHG RQ VROLG VXUIDFHV LQFOXGLQJ WKH VXEVWUDWH RI LQWHUHVW $OO 39' LV GRQH LQ D KLJK YDFXXP FKDPEHU VR WKDW WKH PDWHULDO YDSRU FDQ WUDYHO IURP WKH VRXUFH WR WKH VXEVWUDWH ZLWKRXW UHDFWLQJ ZLWK DWPRVSKHULF JDVHV ,Q WKHUPDO HYDSRUDWLRQ WKH VRXUFH PDWHULDO LV VLPSO\ KHDWHG XQWLO LW YDSRUL]HV D YHU\ VLPSOH DQG FOHDQ SURFHVV XVHIXO WR GHSRVLW VLQJOH HOHPHQW ILOPV 7KH YDSRU SUHVVXUH RI GLIIHUHQW PDWHULDO YDULHV VR JUHDWO\ WKDW WKH FRPSRVLWLRQ RI DQ DOOR\ FDQ QRW EH PDLQWDLQHG ZLWK WKHUPDO HYDSRUDWLRQ >2KU@ ,Q WKH VSXWWHU GHSRVLWLRQ SURFHVVHV DQ LQHUW JDV QRUPDOO\ DUJRQ LV LRQL]HG DQG LRQV DUH DFFHOHUDWHG LQWR WKH VRXUFH PDWHULDO RU WDUJHW :KHQ WKH LQHUW JDV LRQ VWULNHV WKH WDUJHW DWRPV LWV PRPHQWXP LV WUDQVIHUUHG DQG E\ GHYHORSPHQW RI D PRPHQWXP FDVFDGH VRPH RI WKH WDUJHW DWRPV DUH HMHFWHG IURP WKH VXUIDFH 7KH VSXWWHU SURFHVV LV EDVHG RQ PRPHQn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

PAGE 37

68%675$7( 7(03(5$785( 77Pf )LJXUH 'LDJUDP RI 0RYFKHQ DQG 'HPLVKLQfV PRGHO RI WKLQ ILOP PRUSKRORJ\ >.UD@ LQJ DWRPV KDYH OLWWOH NLQHWLF HQHUJ\ DQG WKH VXUIDFH PRELOLW\ LV SULPDULO\ FRQWUROOHG E\ WKH VXEVWUDWH WHPSHUDWXUH %DVHG RQ WKLV 0RYFKHQ DQG 'HPLVKLQ >.UD 7KRU@ GHYHOn RSHG D PRGHO WKDW SUHGLFWV ILOP PRUSKRORJ\ EDVHG RQ WKH KRPRORJRXV WHPSHUDWXUH RI WKH VXEVWUDWH )LJXUH f ,Q WKH 0RYFKHQ DQG 'HPLVKLQ PRGHO WKH ILOP PRUSKRORJ\ LV GLYLGHG LQWR WKUHH ]RQHV ,Q ]RQH 7+ f WKH DGDWRP PRELOLW\ LV VR ORZ WKDW WDSHUHG FU\VWDOV IRUP ZLWK YRLGV EHWZHHQ WKH FU\VWDOV ,Q ]RQH 7+ f WKH DGDWRP PRELOLW\ LV JUHDW HQRXJK WKDW PRVW RI WKH YRLGV ILOO LQ DQG FROXPQDU JUDLQV IRUP ,Q ]RQH 7 f WKH PRELOLW\ LV KLJK HQRXJK WKDW QHDUO\ HTXLD[HG FU\VWDOV IRUP ,Q WKH VSXWWHULQJ SURFHVV WKH NLQHWLF HQHUJ\ RI WKH FRQGHQVLQJ DWRPV FDQ EH FRQWUROOHG E\ WKH JDV SUHVVXUH DQG E\ ELDVLQJ WKH VXEVWUDWH >:LQ@ 7KH HIIHFWV RI VXEVWUDWH WHPSHUDWXUH DQG JDV SUHVVXUH KDYH EHHQ VWXGLHG E\ -$ 7KRUQWRQ >7KR@ ZKR GHYHORSHG D PRGHO WR SUHGLFW WKH PRUSKRORJ\ RI VSXWWHUHG GHSRVLWHG ILOPV VHH )LJXUH f 7KRUQWRQ GLG QRW GLYLGH WKH VSXWWHUHG PRUSKRORJ\ LQWR ]RQHV DV 0RYFKHQ DQG 'HPLVKLQ GLG 7KHUH LV VRPH FRUUHODWLRQ EHWZHHQ WKHVH WZR PRGHOV DV .UDXVV LOOXVWUDWHG LQ ILJXUH >.UD@ ,W VKRXOG DOVR EH QRWHG WKDW SRVW GHSRVLWLRQ DQQHDOLQJ RI D WKLQ ILOP PD\ FDXVH WKH PLFURVWUXFWXUH WR GHYHORS VLPLODU WR WKDW ZKLFK ZRXOG KDYH GHYHORSHG KDG WKH ILOP EHHQ JURZQ DW WKDW VDPH WHPSHUDWXUH >0DF .QR@

PAGE 38

$5*21 35(6685( 0,&5216f 68%675$7( 7(03(5$785( 77PO )LJXUH 7KRUQWRQfV PRGHO RI VSXWWHUHG WKLQ ILOP PRUSKRORJ\ ZLWK D FRPSDULVRQ WR 0RYFKHQ DQG 'HPLVKLQf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n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

PAGE 39

ERXQGDULHV DQG DUH GHVWUR\HG 7KLV FDXVHV WKH ILOP WR DWWHPSW WR FRQWUDFW EXW LQVWHDG LW PD\ GHYHORS D VWDWH RI ELD[LDO WHQVLOH VWUHVV EHFDXVH LW LV DWWDFKHG WR D ULJLG VXEVWUDWH ,Q VSXWWHU GHSRVLWLRQ LW LV SRVVLEOH WR LQFUHDVH WKH NLQHWLF HQHUJ\ RI WKH VSXWWHUHG DWRPV E\ UHGXFLQJ WKH VSXWWHU JDV SUHVVXUH RU ELDVLQJ WKH VXEVWUDWH 8QGHU ELDV LRQV PD\ SHQHWUDWH WKH ILUVW IHZ OD\HUV RI WKH GHSRVLWHG ILOP 7KLV fDWRPLF SHHQLQJf FDQ UHGXFH WKH LQWULQVLF WHQVLOH VWUHVV DQG HYHQ SURGXFH D FRPSUHVVLYH VWUHVV >0DF@ +RZHYHU DWRPLF SHHQLQJ LQMHFWV D ODUJH QXPEHU RI WUDSSHG YDFDQFLHV DQG LPSODQWV JDV >0DF@ ,I WKH ILOP LV KHDWHG WR D WHPSHUDWXUH ZKHUH YDFDQFLHV RU JDVVHV DUH PRELOH WKH\ ZLOO PLJUDWH WR WKH JUDLQ ERXQGDULHV EH GHVWUR\HG RU HYROYHG DQG FKDQJH WKH VWUHVV EDFN WR WHQVLRQ >7RZ@ ,Q 0(06 LW LV RIWHQ GHVLUDEOH WR KDYH D UHVLGXDO VWUHVV QHDU ]HUR DW URRP WHPSHUDn WXUH 7KLV DOORZV IRU JUHDWHU IOH[LELOLW\ RI GHVLJQ :KLOH LW LV SRVVLEOH WR DFKLHYH ORZ VWUHVV E\ DGMXVWLQJ WKH VSXWWHU SDUDPHWHUV LW LV VRPHWLPHV SRVVLEOH WR UHGXFH WHQVLOH VWUHVV E\ WKHUPDO F\FOLQJ ,I WKH WKLQ ILOP DQG VXEVWUDWH DUH WKHUPDOO\ F\FOHG WR f& OLTXLG QLWURJHQ WHPSHUDWXUHf WKH PHWDO ILOP ZLOO QRUPDOO\ FRQWUDFW PRUH WKDQ WKH VXEVWUDWH ,I WKH ILOP LV GXFWLOH LW ZLOO SODVWLFDOO\ GHIRUP DQG XSRQ UHWXUQLQJ WR URRP WHPSHUDWXUH WKH ILOP ZLOO H[SHULHQFH D ORZHU WHQVLOH VWUHVV RU HYHQ D FRPSUHVVLYH VWUHVV >%DO@ 8QIRUWXn QDWHO\ HOHYDWHG WHPSHUDWXUH F\FOLQJ RI WKH ILOP ZLOO FRXQWHUDFW WKLV FRPSUHVVLYH VKLIW LQ VWUHVV DQG UHLQWURGXFH D WHQVLOH VWUHVV >%DO@ $ SULPDU\ FRQFHUQ RI WKLV VWXG\ LV VWUHVV UHOD[DWLRQ LQ WKH WKLQ ILOP GHILQHG DERYH WR EH D WLPH GHSHQGHQW FKDQJH LQ WKH VWDWH RI LQWULQVLF SOXV H[WULQVLF VWUHVV 6WUHVV UHOD[n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

PAGE 40

2 7HPSHUDWXUH & )LJXUH 6WUHVV WHPSHUDWXUH SURILOH IRU SXUH VSXWWHUHG DOXPLQXP DQG PLFURHOHFWURQLF DOOR\ $O6L&X >-HU@ UHVROYHG VKHDU VWUHVV JUHDWHU WKDQ WKH FULWLFDO YDOXH 6HFRQG D GLVORFDWLRQ PRYLQJ DORQJ D VOLS V\VWHP PXVW QXFOHDWH PLVILW GLVORFDWLRQV DW WKH ILOPVXEVWUDWH LQWHUIDFH ZKLFK UHn TXLUHV DGGLWLRQDO HQHUJ\ ,I D SDVVLYDWLQJ R[LGH IRUPV RQ WKH VXUIDFH RI WKH ILOP DV LQ WKH FDVH RI DOXPLQXP PLVILW GLVORFDWLRQV PXVW EH FUHDWHG DW ERWK LQWHUIDFHV >1L[@ IXUWKHU UHGXFLQJ GLVORFDWLRQ JOLGH 7KH SULPDU\ PHWKRG XVHG WR WHVW WKH PHFKDQLFDO SURSHUWLHV RI WKLQ ILOPV LV PHDn VXUHPHQW RI VWUHVV DV D IXQFWLRQ RI WHPSHUDWXUH $V WKH WKLQ ILOP H[SDQGV DW D GLIIHUHQW UDWH WKDQ WKH VXEVWUDWH FKDQJLQJ WKH WHPSHUDWXUH FKDQJHV WKH VWUHVV LQ WKH ILOP 7KLV LV DQ HDV\ DQG QRQGHVWUXFWLYH ZD\ WR WHVW D WKLQ ILOP EXW LW LV OLPLWHG EHFDXVH WKH VWUHVV DQG WHPSHUDWXUH FDQQRW EH YDULHG LQGHSHQGHQWO\ ,QVWUXPHQWV WKDW PHDVXUH VWUHVV LQ WKLV ZD\ LQFOXGH WKH 7HQFRU )OH[XV ZKLFK ZDV XVHG LQ WKLV VWXG\ VHH GHVFULSWLRQ LQ &KDSWHU f

PAGE 41

7KH UHVXOW RI WKLV W\SH RI WHVWLQJ LV D SORW RI VWUHVV DV D IXQFWLRQ RI WHPSHUDWXUH VKRZQ LQ )LJXUH IRU VSXWWHUHG SXUH $ DQG $Ob6Lb&X 7R HYDOXDWH VWUHVV UHOD[DWLRQ WKH VDPSOH FDQ EH KHDWHG WR D WHPSHUDWXUH DQG WKH VWUHVV SORWWHG DV D IXQFWLRQ RI WLPH )RU ELPHWDOOLF DFWXDWRUV D ILOP WKDW VKRZV QR K\VWHUHVLV LQ VWUHVV YHUVXV WHPSHUDn WXUH RYHU WKH UDQJH XVHG LV GHVLUHG +\VWHUHVLV LQ WKLV SORW LQGLFDWHV WKDW WKH ILOP LV SODVWLn FDOO\ GHIRUPLQJ )LJXUH VKRZV WKDW WKH PHWDOV FXUUHQWO\ EHLQJ XVHG WR PDQXIDFWXUH ELPHWDOOLF DFWXDWRUV >-HU@ ERWK H[KLELW VLJQLILFDQW K\VWHUHVLV RYHU WKH WHPSHUDWXUH UDQJH IURP & WR r& 7KH SXUH DOXPLQXP VKRZV OLWWOH VWUHQJWK RYHU r& ZKLOH WKH $O6L&X DOOR\ LV RQO\ VOLJKWO\ VWURQJHU DQG SODVWLFDOO\ GHIRUPV DERYH r& XQGHU D FRPSUHVVLYH VWUHVV RI 03D 7KH URRP WHPSHUDWXUH UHVLGXDO WHQVLOH VWUHVV IRU SXUH DOXPLQXP LV 03D ZKLOH WKH $O6L&X LV 03D ,VRWKHUPDO VWUHVV UHOD[DWLRQ WHVWV ZHUH DOVR SHUIRUPHG RQ SXUH DOXPLQXP DQG $O6L&X DOOR\ WKLQ ILOPV >-HU@ ZLWK ERWK ILOPV VKRZLQJ D b UHGXFWLRQ LQ VWUHVV RYHU PLQXWHV $W r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n QLVP UHVSRQVLEOH IRU VWUHVV UHOD[DWLRQ +RZHYHU 'UDSSHU DQG +LOO DVVXPHG D VLQJOH PHFKDQLVP ZDV UHVSRQVLEOH IRU WKH VWUHVV UHOD[DWLRQ DQG ILW WKHUH GDWD ZLWK D VLQJOH H[SRQHQWLDO IXQFWLRQ LJQRULQJ WKH FRQFOXVLRQV RI +HUVKNRYLW] %OHFK DQG .RPHP 7KH VLQJOH ILJXUH RI WKH GDWD DQG FXUYH ILW LQ WKHLU DUWLFOH VKRZHG JUHDW LQDFFXUDFLHV EHWZHHQ WKH GDWD DQG WKH ILW ,Q :LWYURXZ 3URRVW %HZHHUGW 5RXVVHO DQG ,PHF DOVR UH

PAGE 42

SRUWHG LVRWKHUPDO VWUHVV UHOD[DWLRQ GDWD IURP KLJKO\ WHQVLOH VWUHVVHG $O6L&X WKLQ ILOPV $W ORZ WHPSHUDWXUH r&f WKH\ DWWULEXWH VWUHVV UHOD[DWLRQ WR GLVORFDWLRQ JOLGH $W KLJKHU WHPSHUDWXUHV r& WR r&f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n WLRQ KDV QRW EHHQ H[WHQVLYHO\ VWXGLHG ZRUN WR UHGXFH IDLOXUHV FDXVHG E\ WKHUPDO F\FOLQJ KDV EHHQ UHSRUWHG 0XFK RI WKLV ZRUN ZDV GRQH WR LPSURYH WKH UHOLDELOLW\ RI -RVHSKVRQ VXSHUFRQGXFWLQJ GHYLFHV 648,' 6XSHUFRQGXFWLQJ 4XDQWXP ,QWHUIHUHQFH 'HYLFHf >.LU@ ZKLFK ZHUH RI LQWHUHVW IRU EXLOGLQJ XOWUDKLJKVSHHG FRPSXWHUV 7KHVH GHYLFHV RSHUDWH DW WHPSHUDWXUHV EHORZ r. r&f DQG QHHG WR EH DEOH WR ZLWKVWDQG WHPSHUDn WXUH F\FOHV IRUP r& WR r& %DVDYDLDK DQG *UHLQHU IRXQG WKH DGGLWLRQ RI JROG WR WKH OHDG LQGLXP DOOR\ UHGXFHG IDLOXUHV EXW WKH PHFKDQLVPV UHVSRQVLEOH IRU WKH UHGXFWLRQ LQ IDLOXUHV ZDV QRW GHWHUPLQHG >%DV@ (QYLURPHQWDO 6WDELOLW\ RI 0DWHULDOV %LPHWDOOLF DFWXDWRUV PD\ RSHUDWH LQ DWPRVSKHULF JDVHV DW HOHYDWHG WHPSHUDWXUHV DQG WKH PDWHULDOV RI FRQVWUXFWLRQ PXVW EH VWDEOH (QYLURQPHQWDO LQVWDELOLWLHV RI WKHVH ILOPV ZRXOG FRPSURPLVH WKHLU SK\VLFDO LQWHJULW\ DQG FDXVH WKH GHYLFH WR IDLO 7KHUH DUH WZR HQYLURQPHQW UHDFWLRQV WKDW QHHG WR EH FRQVLGHUHG IRU WKLV DSSOLFDWLRQ R[LGDWLRQ DQG FRUURVLRQ 0RVW PHWDOV R[LGL]H DQG VRPH VHOISDVVLYDWH ZKLOH RWKHUV UHTXLUH VRPH SURWHFWLYH FRDWLQJV 7KLV VWXG\ KDV IRFXVHG RQ PHWDOV ZKLFK VHOI SDVVLYDWH VXFK DV DOXPLQXP

PAGE 43

FRSSHU WLWDQLXP QLFNHO )RU D SDVVLYDWLQJ R[LGH WR EH HIIHFWLYH LW VKRXOG KDYH WKH IROORZn LQJ IRXU SURSHUWLHV KLJK WKHUPRG\QDPLF DQG NLQHWLF VWDELOLW\ VORZ JURZWK UDWH DGKHUHQFH WR WKH PHWDO DQG HDVLO\ IRUP RU UHIRUP >:HV@ 7KH JURZWK UDWH LV RI JUHDWHU LPSRUn WDQFH LQ WKLQ ILOPV DV DQ R[LGH FRXOG FRQVXPH WKH HQWLUH ILOP 0RVW SDVVLYDWLQJ R[LGHV JHQHUDOO\ IROORZ D SDUDEROLF JURZWK UDWH HTXDWLRQ JLYHQ E\ [ \ f ZKHUH [ LV WKH WKLFNQHVV RI WKH R[LGH NS LV WKH SDUDEROLFJURZWKUDWH FRQVWDQW DQG W LV WLPH $OXPLQXP W\SLFDOO\ H[KLELWV DQ LQYHUVH ORJDULWKPLF GHSHQGHQFH DW ORZ WHPSHUDn WXUHV DQG SDUDEROLF DW KLJK WHPSHUDWXUHV 7KXV IRU WKLQ ILOP DSSOLFDWLRQV PDWHULDOV ZLWK D ORZ SDUDEROLFJURZWK UDWH FRQVWDQW DUH QHHGHG VXFK DV DOXPLQXP DQG VLOLFRQ IRU HOn HYDWHG WHPSHUDWXUH DSSOLFDWLRQV ,Q DGGLWLRQ WKH IRUPDWLRQ RI DQ R[LGH DQG RU D PHWDO R[\JHQ VROLG VROXWLRQ ZLOO LQWURGXFH VRPH VWUHVV LQWR WKH PHWDO VXEVWUDWH 7KLV FKDQJH LQ VWUHVV FRXOG DIIHFW WKH IRUFH RU GLVSODFHPHQW RI D ELPHWDOOLF VWULS 7KLV FKDQJH PD\ EH WLPH GHSHQGHQW DV FDQ EH VHHP IURP HTXDWLRQ &RUURVLRQ LV DOVR D FRQFHUQ EHFDXVH WKHVH GHYLFHV PD\ EH RSHUDWHG LQ PRLVW DLU ZKHUH WKH SRVVLELOLW\ RI D JDOYDQLF FHOO H[LVWV 2QH DUHD ZKHUH D FHOO FRXOG H[LVW LV EHn WZHHQ WKH DFWLYH DQG SDVVLYH OD\HUV RI WKH ELPHWDOOLF VWULS ,Q WKLV DSSOLFDWLRQ WKHUH LV QRUPDOO\ DQ R[LGH OD\HU EHWZHHQ WKHVH WZR OD\HUV RI WKH ELPHWDOOLF VWULS WKDW HOHFWULFDOO\ LVRODWHV WKHP DQG VR D JDOYDQLF FHOO FDQQRW EH IRUPHG 'LVUXSWLRQ RI WKLV R[LGH ZRXOG FDXVH WKLV WR EH D FRQFHUQ $OVR IRUPDWLRQ RI PLFURJDOYDQLF FHOOV LQ D WZR SKDVH DOOR\ VKRXOG EH WHVWHG >$VN@

PAGE 44

&+$37(5 (;3(5,0(17$/ 352&('85( ,QWURGXFWLRQ (OHYHQ GLIIHUHQW PDWHULDOV ZHUH GHSRVLWHG DV WKLQ ILOPV DQG DQDO\]HG LQ WKLV VWXG\ VHH 7DEOH f 7KH GHSRVLWLRQ PHWKRGV DQG SURFHGXUHV ZLOO EH FRYHUHG LQ WKLV VHFWLRQV 7KH ILOPV ZHUH ILUVW WHVWHG WR GHWHUPLQH WKHLU PHFKDQLFDO SURSHUWLHV DQG WKH EHVW PDWHULDOV ZHUH DQDO\]HG IXUWKHU WR GHWHUPLQH WKHLU PLFURVWUXFWXUDO DQG FKHPLFDO FRPSRVLWLRQV $ QXPEHU RI DQDO\VLV WHFKQLTXHV ZHUH XVHG DV GLVFXVVHG EHORZ 7DEOH 0DWHULDOV H[DPLQHG LQ WKLV VWXG\ 0DWHULDO 'HSRVLWHG 7KLFNQHVV SP &RPSRVLWLRQ RU 3XULW\ ZW bf 7 $OXPLQXP $ &X $J 0Q 0J 7 $OXPLQXP $ 0J &U $OXPLQXP $ &X 8 =U 1LFNHO 7LWDQLXP 0DQJDQHVH &RSSHU $7L $O 7L &X$X &X $X &X$X &X $X &X$X &X $X

PAGE 45

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b VXOIXULF DFLG DQG b K\GURJHQ SHUR[LGHb IROORZHG E\ D GHn LRQL]HG ZDWHU ULQVH DQG EORZQ GU\ ZLWK GU\ QLWURJHQ 7KH DOXPLQXP DOOR\ DQG FRSSHU JROG LQWHUPHWDOOLF ILOPV ZHUH VSXWWHU GHSRVLWHG E\ 6SXWWHUHG 7KLQ )LOPV ,QF 6DQWD &ODUD &$ XVLQJ DQ LQFK '& PDJQHWURQ VSXWWHU JXQ UXQQLQJ DW N: RI SRZHU 7KH FKDPEHU ZDV ILUVW SXPSHG GRZQ WR [On 7RUU WKHQ EDFN ILOOHG ZLWK DUJRQ WR D SUHVVXUH RI P7RUU 7KH VXEVWUDWHV ZHUH KHDWHG WR r& 7KH DOXPLQXP DOOR\ VSXWWHU WDUJHWV ZHUH IURP EXON VKHHW VWRFN &RSSHU JROG LQWHUPHWDOn OLFV ZHUH SURGXFHG E\ GHSRVLWLQJ IRXU OD\HUV RI SXUH FRSSHU DQG JROG LQ WKH SURSHU SURn SRUWLRQ WR SURGXFH WKH UHTXLUHG FRPSRVLWLRQ 7DEOH VKRZV WKH WKLFNQHVV RI WKH LQGLn YLGXDO FRSSHU DQG JROG ILOPV WKDW ZHUH GHSRVLWHG IRU HDFK RI WKH DYHUDJH FRPSRVLWLRQV 7DEOH /D\HU WKLFNQHVV GHSRVLWHG IRU WKH FRSSHUJROG LQWHUPHWDOOLF ILOPV &RPSRXQG &RSSHU /D\HU 7KLFNQHVV _LP *ROG /D\HU 7KLFNQHVV_LP &X$X &X$X &X$X

PAGE 46

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n7RUU EHIRUH HDFK GHSRVLWLRQ 'XH WR RXWJDVVLQJ WKH FKDPEHU SUHVVXUH ZRXOG ULVH WR [ 2n 7RUU GXULQJ GHSRVLWLRQ 7KH VXEVWUDWHV ZHUH KHDWHG WR r& r& ZLWK UHVLVWLYH KHDWHUV PDQXDOO\ FRQWUROOHG ZLWK D YDULDEOH WUDQVIRUPHU &RSSHU *ROG +HDW 7UHDWPHQW 7KH FRSSHU JROG FRPSRXQGV UHTXLUHG D VSHFLDO KHDW WUHDWPHQW WR KRPRJHQL]H WKH OD\HUV DQG WR GHYHORS WKH ORQJ UDQJH RUGHUHG LQWHUPHWDOOLF VWUXFWXUH >)HG@ 7KHVH KHDW WUHDWPHQWV ZHUH SHUIRUPHG LQ D YDFXXP RI ;n 7RUU RU EHWWHU WR LQKLELW R[LGDWLRQ RI WKH FRSSHU 7KH KHDW WUHDWPHQWV ZHUH KRXUV DW r& WR KRPRJHQL]H WKH ILOP DQG KRXUV DW f& IROORZHG E\ KRXUV DW r& WR GHYHORS WKH ORQJ UDQJH RUGHU $OO WHPSHUDWXUHV FKDQJHV ZHUH UDPSHG DW r& SHU PLQXWH XVLQJ DQ 2PHJD &1 FRQWUROOHU 6WUHVV DQG 6WUHVV 5HOD[DWLRQ 0HDVXUHPHQWV 6WUHVV DQG VWUHVV UHOD[DWLRQ LQ WKH VDPSOHV ZHUH PHDVXUHG ZLWK D 7HQFRU )OH[XV PRGHO 7KH )OH[XV XVHV DQ RSWLFDO OHYHU WR PHDVXUH WKH FXUYDWXUH RI WKH VDPSOH >7XU@ 6WRQH\fV HTXDWLRQ >6WR@ (TXDWLRQ ZDV XVHG WR FDOFXODWH WKH DYHUDJH ELD[LDO VWUHVV D LQ WKH ILOP (K f§!f ‘ ‘ 5 D f

PAGE 47

ZKHUH 5 LV WKH FXUYDWXUH RI WKH VXEVWUDWH W LV WKH ILOP WKLFNQHVV K LV WKH VXEVWUDWH WKLFNn QHVV ( LV WKH :DU@ DQG WR HYDOXDWH FU\VWDOOLWH VL]H LQ WKH DOXPLQXP DOOR\ ILOPV >&XO@ 7KH V\VWHP XVHG ZDV D 3KLOLSV $3' [UD\ GLIIUDFWLRQ V\VWHPV FRQWUROOHG ZLWK DQ ,%03& W\SH FRPSXWHU UXQQLQJ 0LFURVRIW ZLQGRZV ZLWK 3KLOLSV 3: $XWRPDWHG 3RZGHU 'LIIUDFWLRQ VRIWZDUH YHUVLRQ J ;UD\V IURP D FRSSHU DQRGH RSHUDWHG DW N9 DQG PLOOLDPSV ZLWK D QLFNHO ILOWHU ZHUH XVHG

PAGE 48

)LJXUH ,VRWKHUPDO VWUHVV UHOD[DWLRQ LQ DOXPLQXP DOOR\ VDPSOH DW r& ;UD\ GLIIUDFWLRQ LV HIIHFWLYH IRU GHWHFWLQJ WKH FKDQJH LQ VWUXFWXUH WKDW RFFXUV LQ DQ RUGHUGLVRUGHU WUDQVIRUPDWLRQ )RU WKH FRSSHU JROG V\VWHP WKH GLVRUGHUHG FU\VWDOORn JUDSKLF VWUXFWXUH LV )&& ZLWK DWRPV VLWWLQJ UDQGRPO\ RQ ODWWLFH VLWHV ,Q DQ RUGHUHG SKDVH RQH W\SH RI DWRP HLWKHU FRSSHU RU JROG VLWV LQ WKH FRUQHU VLWHV DQG WKH RWKHU DWRP ZLOO VLW LQ WKH IDFH VLWHV 7KLV FUHDWHV WZR LQWHUOLQNHG VLPSOH FXELF RU VLPSOH WHWUDJRQDO FU\VWDOORJUDSKLF VWUXFWXUH 7KLV FKDQJH LQ WKH FU\VWDO VWUXFWXUH FDXVHV WKH RULJLQDO GLIIUDFn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

PAGE 49

WKLV WHFKQLTXH WKH f DQG f GLIIUDFWLRQ SHDNV ZHUH VHOHFWHG IRU HYDOXDWLRQ 7KH f SHDN ZDV VFDQQHG IURP r WR r DQG WKH f SHDN ZDV VFDQQHG IURP r WR r XVLQJ r LQFUHPHQWV DQG VHFRQG GZHOO WLPH )RU OLQH VKDSH DQDO\VLV DQ XQVWUHVVHG TXDUW] VWDQGDUG ZDV VFDQQHG LQ WKH VDPH GLIIUDFWRPHWHU XQGHU WKH VDPH FRQGLWLRQV 7KH TXDUW] VWDQGDUG ZDV VFDQQHG IURP r WR r DQG IURP r WR r /LQH VKDSH DQDO\VLV H[WUDFWV WKH IDFWRUV WKDW FDXVH WKH [UD\ GLIIUDFWLRQ SHDNV WR VKLIW DQGRU EURDGHQ 7KHUH DUH WKUHH PDLQ FDXVHV RI SHDN EURDGHQLQJ LQ [UD\ GLIIUDFWLRQ 7KH ILUVW LV WKH GLIIUDFWRPHWHU LWVHOI ,QVWUXPHQWDO EURDGHQLQJ LV HOLPLQDWHG E\ FRPSDULn VRQ WR WKH TXDUW] VWDQGDUG 7KH RWKHU WZR FDXVHV RI SHDN EURDGHQLQJ DUH VWUDLQ DQG FU\VWDOOLWH VL]H (DFK RI WKHVH SURGXFH D GLIIHUHQW VKDSH 6WUDLQ HIIHFWV SURGXFH D *DXVVLDQ GLVWULEXWLRQ ZKLOH WKH FU\VWDOOLWH VL]H SURGXFH D &DXFK\ GLVWULEXWLRQ >:DU@ 7KH 3KLOOLSV GLIIUDFWLRQ VRIWZDUH FDQ GHFRQYROXWH WKHVH HIIHFWV DQG GHWHUPLQH WKH VWUDLQ DQG FU\VWDOOLWH VL]H $WRPLF )RUFH 0LFURVFRS\ $)0 $ 'LJLWDO ,QVWUXPHQWV 1DQRVFRSH ,,, DWRPLF IRUFH PLFURVFRSH $)0f ZDV XVHG WR GHWHUPLQH WKH VXUIDFH WRSRORJ\ RI WKH DOXPLQXP DOOR\ VDPSOHV >+XG@ $ VLOLFRQ WLS LQ WKH WDSSLQJ PRGH ZDV XVHG WR SUREH WKH VXUIDFH RI WKH VDPSOHV 'LJLWDO ,QVWUXPHQWV VRIWZDUH ZDV XVHG SURFHVV WKH VFDQQHG LPDJHV DQG FDOFXODWH WKH 506 VXUIDFH URXJKQHVV 6FDQQLQJ (OHFWURQ 0LFURVFRS\ 6(0 $ -(2/ VFDQQLQJ HOHFWURQ PLFURVFRSH ZDV DOVR XVHG WR LPDJH WKH VXUIDFH RI VDPSOHV ,Q DGGLWLRQ WR WKH VHFRQGDU\ DQG EDFN VFDWWHUHG HOHFWURQ GHWHFWRUV WKLV V\VWHP KDV DQ HQHUJ\ GLVSHUVLYH VSHFWURPHWHU ('6f WKDW DOORZHG TXDOLWDWLYH HOHPHQWDO DQDO\VLV RI VXUIDFH IHDWXUHV >*RG@ 6DPSOHV ZHUH ILUVW YLHZHG LQ WKH VHFRQGDU\ HOHFWURQ PRGH WR HYDOXDWH WKH VXUIDFH WRSRORJ\ 9HU\ EULJKW DUHDV ZHUH YLHZHG LQ WKH EDFN VFDWWHUHG

PAGE 50

HOHFWURQ PRGH 7KH WUDH VHFRQGDU\ HOHFWURQ \LHOG IURP DQ DUHD LV PRUH VWURQJO\ GHSHQn GHQW RQ WKH WRSRORJ\ ZKHUHDV WKH EDFN VFDWWHUHG HOHFWURQ LQWHQVLW\ LV GRPLQDWHG E\ WKH DWRPLF PDVV RI WKH H[FLWHG DWRPV $UHDV WKDW DSSHDUHG WR KDYH GLIIHUHQW DWRPLF FRPSRVLn WLRQV ZHUH H[DPLQHG ZLWK ('6 WR GHWHUPLQH WKH FRQFHQWUDWLRQ RI HOHPHQWV (OHFWURQ3UREH 0LFURDQDOYVLV (30$ (30$ ZDV XVHG WR TXDQWLWDWLYHO\ GHWHUPLQH WKH FRPSRVLWLRQV RI WKH DOXPLQXP DOOR\V DQG WKH FRSSHU JROG LQWHUPHWDOOLF ILOPV >*RO@ $ -(2/ 6XSHUSUREH ZLWK IRXU ZDYHOHQJWK GHVSHUVLYH VSHFWURPHWHUV ZDV XVHG )RU WKH DOXPLQXP DOOR\ ILOPV WKH EXON DOXPLQXP DOOR\ ZDV XVHG DV WKH DQDO\VLV VWDQGDUG WR GHWHUPLQH GLIIHUHQFH LQ FRPSRn VLWLRQ EHWZHHQ WKH EXON DQG WKLQ ILOP FRPSRVLWLRQ )RU WKH FRSSHU JROG LQWHUPHWDOOLF ILOPV SXUH FRSSHU DQG JROG VWDQGDUGV ZHUH XVHG ZLWK WKH =$) DWRPLF QXPEHU HIIHFWV ; UD\ DEVRUSWLRQ DIIHFWV DQG ;UD\ IOXRUHVFHV HIIHFWVf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f ZDV PDGH LQ WKH SDUDIILQ RYHU WKH GLPSOH 7KH VDPSOHV ZHUH HWFKHG LQ D VROXWLRQ RI b +) DQG b +12Y 7KH HWFKLQJ SURFHVV ZDV PRQLWRUHG E\ YLHZLQJ WKH VDPSOHV WKURXJK D ORZ SRZHU ELQRFXODU PLFURVFRSH 2QFH WKH DFLG KDG HWFKHG WKURXJK WKH VLOLFRQ DQG UHDFKHG WKH DOXPLQXP ILOP HWFKLQJ ZDV VWRSSHG 7KHVH VDPSOHV ZHUH LRQ PLOOHG XQWLO D VPDOO RSHQLQJ DSSHDUHG LQ WKH DOXPLQXP WKHQ YLHZHG LQ WKH 7(0

PAGE 51

,Q WKH 7(0 WKH VDPSOHV ZHUH ILUVW YLHZHG LQ WKH EULJKW ILHOG PRGH WR HYDOXDWH WKH JHQHUDO VWUXFWXUH RI WKH VDPSOH >/RU@ 7KH GDUN ILHOG PRGH ZDV XVHG WR EHWWHU LGHQWLI\ GLIIHUHQW SKDVHV (OHFWURQ GLIIUDFWLRQ LPDJHV DQG SDWWHUQV ZHUH DOVR WDNHQ WR LGHQWLI\ WKH SKDVHV SUHVHQW E\ FDOFXODWLQJ WKH LQWHUSODQHU VSDFLQJ 7KH LQWHU SODQH VSDFLQJ GKNO ZDV FDOFXODWHG XVLQJ ;/ 5GKNO f ZKHUH / LV WKH FDPHUD OHQJWK IRU WKH -(2/ &; / FPf ; LV WKH HOHFWURQ ZDYHOHQJWK ZKLFK LV EDVHG RQ WKH DFFHOHUDWLQJ YROWDJH $ OQP EDVHG RQ DQ DFFHOn HUDWLQJ YROWDJH RI NH9f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n WLDOO\ >6PL *LI@ DQG WZR ZHUH WHVWHG IRU DOXPLQXP DOOR\V b K\GURIOXRULF DFLG DQG D b VRGLXP K\GUR[LGH VROXWLRQ 6DWLVIDFWRU\ GHILQLWLRQ RI WKH JUDLQ ERXQGDULHV ZDV QRW DFKLHYHG $XJHU (OHFWURQ 6SHFWURVFRS\ I$(6W 7KH 3HUNLQ(OPHU 3+, 6FDQQLQJ $XJHU 0LFURSUREH ZDV XVHG WR HYDOXDWH WKH R[LGH WKLFNQHVV RQ WKH DOXPLQXP DOOR\ ILOPV $XJHU DQDO\VLV LV YHU\ VXUIDFH VHQVLWLYH

PAGE 52

FRQWDLQLQJ LQIRUPDWLRQ SULPDULO\ IURP WKH WRS WKUHH RU IRXU DWRPLF OD\HUV >+,2@ 7KH $XJHU V\VWHP FRQWDLQHG DQ DUJRQ VSXWWHU JXQ WKDW HQDEOHV GHSWK SURILOHV WR EH FROOHFWHG 7R GR VR DQ $XJHU VSHFWUXP ZDV WDNHQ WKH VDPSOH VSXWWHUHG IRU VHFRQGV DQG DQRWKHU $XJHU VSHFWUXP WDNHQ 7KLV SURFHVV ZDV UHSHDWHG WR WKH UHTXLUHG GHSWK 7KH DOXPLQXP DOOR\ VDPSOHV ZHUH VSXWWHUHG XQWLO WKH R[\JHQ VLJQDO KDG GHFUHDVHG WR b RI LWV RULJLQDO YDOXH 7KLV WHFKQLTXH JLYHV D WKLFNQHVV LQ WHUPV RI VSXWWHU WLPH 7KH FRQGLWLRQV XVHG UHVXOW LQ DQ HVWLPDWHG VSXWWHU UDWH RI DQJVWURPV SHU PLQXWH 7KHVH GDWD DUH SULPDULO\ XVHG WR FRPSDUH GLIIHUHQFHV LQ R[LGH WKLFNQHVV IURP VDPSOH WR VDPSOH &XUYH )LWWLQJ 7R FXUYH ILW WKH VWUHVV UHOD[DWLRQ GDWD FXUYH ILWWLQJ WHFKQLTXHV ZHUH HYDOXDWHG SRO\QRPLDO VLQJOH H[SRQHQWLDO DQG GRXEOH H[SRQHQWLDO 7KH GRXEOH H[SRQHQWLDO WHFKn QLTXH ILW WKH GDWD XVLQJ WKH HTXDWLRQ FUWf $ %H &H f ZKHUH D LV VWUHVV W LV WLPH DQG $ % & P DQG P DUH FRQVWDQWV WKDW DUH GHWHUPLQHG IURP WKH GDWD 7KLV WHFKQLTXH ILW PRVW RI WKH GDWD ZLWK WKH OHDVW HUURU 7KH SURFHVV IRU GHWHUPLQLQJ WKH FRQVWDQW ZDV WKH OLQHDUL]DWLRQ RI WKH GDWD 7KLV ZDV VRPHWLPHV FRPSOH[ DV WKH WZR H[SRQHQWLDO IXQFWLRQV QHHGHG WR EH VHSDUDWHG WR EH OLQHDUL]HG 6HSDUDWLRQ LV DFFRPSOLVKHG E\ ILUVW ILWWLQJ WKH VWUHVV UHOD[DWLRQ GDWD IRU W PLQXWH 7R OLQHDUL]H WKHVH GDWD WKH YDOXH RI $ LQ HTXDWLRQ ZDV ILUVW PDQXDOO\ VHOHFWHG DQG VXEWUDFWHG IURP WKH UHPDLQLQJ GDWD 7KH QDWXUDO ORJ IRU WKH UHPDLQLQJ GDWD ZDV WKHQ WDNHQ DQG OHDVW VTXDUHV OLQHDU ILW WR GHWHUPLQH WKH VORSH LQWHUFHSW DQG FRUUHODWLRQ IDFWRU 7KH $ FRQVWDQW ZDV LWHUDWHG WR REWDLQ WKH EHVW SRVVLEOH FRUUHODWLRQ IDFWRU :LWK WKLV VWHS FRPSOHWHG WKH ILUVW H[SRQHQWLDO ZDV VXEWUDFWHG IURP WKH RULJLQDO GDWD 7KH QDWXUDO ORJ RI WKH GLIIHUHQFH ZDV WKHQ WDNHQ DQG WKH UHVXOWLQJ GDWD DJDLQ ILWWHG ZLWK D OLQHDU OHDVW VTXDUH

PAGE 53

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n WXUH WKHUHIRUH WKH OHQJWK ZDV QRW UHTXLUHG 7KH RXWSXW ZDV SORWWHG DV D SHUFHQWDJH FKDQJH LQ FXUYDWXUH IRU WKH VWUXFWXUH ZLWK YDU\LQJ R[LGH WKLFNQHVV

PAGE 54

&+$37(5 5(68/76 ,QWURGXFWLRQ 7KH UHVXOWV RI WKH WHVWLQJ DQG DQDO\VLV ZLOO EH SUHVHQWHG LQ WKLV FKDSWHU ZKLFK LV RUJDQL]HG E\ PDWHULDO WR SURYLGH DQ XQGHUVWDQGLQJ RI WKHLU SHUIRUPDQFH 7 $OXPLQXP )LJXUH VKRZV WKH VWUHVV YHUVXV WHPSHUDWXUH SORW IRU WKH 7 DOOR\ ILOP DQG PD\ EH FRPSDUHG WR WKH OLWHUDWXUH GDWD VKRZQ LQ ILJXUH 7KHVH GDWD ZHUH FROOHFWHG DIWHU DQ LQLWLDO WHPSHUDWXUH F\FOH WR r& WR VWDELOL]H WKH ILOPfV PLFURVWUXFWXUH DIWHU GHSRVLWLRQ :KLOH WKH LQWULQVLF VWUHVV LQ WKHVH ILOPV ZHUH D IXQFWLRQ RI GHSRVLWLRQ SDUDPn HWHUV WKH ILUVW F\FOH WR r& DQQHDOHG DQG FRQVROLGDWHG WKH ILOPV DQG WKH H[WULQVLF VWUHVVHV ZHUH VR KLJK WKDW WKH\ GRPLQDWHG WKH VWUHVVHV UHSRUWHG DV LQ )LJXUH $V FDQ EH VHHQ IURP WKHVH GDWD WKH URRP WHPSHUDWXUH UHVLGXDO WHQVLOH VWUHVV LV DOPRVW 03D DQG WKH ILOP EHJLQV WR SODVWLFDOO\ GHIRUP DW DSSUR[LPDWHO\ r& XQGHU D FRPSUHVVLYH VWUHVV RI 03D 7KLV VDPSOH ZDV DOVR FRROHG WR '& LQ DQ DWWHPSW WR UHGXFH LWV URRP WHPSHUDWXUH UHVLGXDO VWUHVV EXW WKLV SURFHGXUH KDG QR HIIHFW RQ WKH URRP WHPSHUDWXUH UHVLGXDO VWUHVV 7KH LVRWKHUPDO VWUHVV UHOD[DWLRQ VHHQ LQ WKHVH VDPSOHV DW GLIIHUHQW WHPSHUDWXUHV DUH VKRZQ LQ ILJXUHV WKURXJK 7KHVH GDWD ZHUH FXUYH ILW DV GHVFULEH DERYH DQG WKH HTXDWLRQ LV VKRZQ LQ HDFK SORW %HORZ r& WKH VWUHVV UHOD[DWLRQ GDWD ZDV ILW ZHOO E\ D GRXEOH H[SRQHQWLDO GHFD\LQJ IXQFWLRQ WKDW VXJJHVWLQJ WKDW DUH WZR VWUHVV UHOD[DWLRQ SURn FHVVHV ZHUH RFFXUULQJ $W r& DQG r& WKH QRLVH RQ WKH GDWD DUH WR ODUJH WR DOORZ ILWWLQJ 5DWKHU WKDQ KHDWLQJ DW D UDWH RI Or&PLQXWH VDPSOHV ZHUH DOVR KHDWHG UDSLGO\

PAGE 55

2 7HPSHUDWXUH r& )LJXUH 6WUHVV YHUVXV WHPSHUDWXUH SORW RI 7 DOXPLQXP WKLQ ILOP )LJXUH ,VRWKHUPDO VWUHVV UHOD[DWLRQ RI 7 DOXPLQXP WKLQ ILOP DW r&

PAGE 56

6WHVV 03D 6WHVV 03D )LJXUH ,VRWKHUPDO VWUHVV UHOD[DWLRQ RI 7 DOXPLQXP WKLQ ILOP DW r& )LJXUH ,VRWKHUPDO VWUHVV UHOD[DWLRQ RI 7 DOXPLQXP WKLQ ILOP DW r&

PAGE 57

6WHVV 03D 6WHVV 03D 7LPH PLQXWHV ,VRWKHUPDO VWUHVV UHOD[DWLRQ RI 7 DOXPLQXP WKLQ ILOP DW r& 7LPH PLQXWHV )LJXUH ,VRWKHUPDO VWUHVV UHOD[DWLRQ RI 7 DOXPLQXP WKLQ ILOP DW r&

PAGE 58

r&PLQf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r& %DFN VFDWWHU LPDJHV RI WKH VDPH UHJLRQV VKRZHG WKDW WKH EULJKW DUHDV KDG D GLIIHUHQW DWRPLF FRPSRVLWLRQ WKDQ WKH DUHDV DURXQG WKHP )LJXUH LV WKH ('6 VSHFWUXP IRU WKH GDUN DUHDV VKRZLQJ RQO\ DOXPLQXP )LJXUH VKRZV WKH ('6 VSHFWUXP IRU WKH EULJKW DUHDV DQG ERWK FRSSHU DQG DOXPLQXP DUH GHWHFWHG ZKLOH VLOLFRQ LV QRW 7KHUHIRUH WKH EULJKW DUHDV RQ WKH 6(0 LPDJH DUH GXH WR ERWK WRSRJUDSKLF IHDWXUHV DQG GLIIHUHQFHV LQ FKHPLFDO FRPSRVLWLRQ RI WKH VDPSOH SUHVXPDEO\ GXH WR SUHFLSLWDWHV 7DEOH 5HVLGXDO VWUHVV DIWHU LQILQLWH UHOD[DWLRQ WLPH D FKDQJH LQ VWUHVV DW VKRUW $FDQG ORQJ $FUWLPHV DQG WKH WLPH FRQVWDQWV [f IRU WKH WZR H[SRQHQWLDO HTXDWLRQV IRU7O DOXPLQXP 7HPS r&f &7c 03Df $D 03Df 7 PLQf $D! 03Df 7 PLQf O22

PAGE 59

7DEOH &KHPLFDO FRPSVLWLRQ RI VSXWWHU GHSRVLWHG 7 DOXPLQXP DOOR\ WKLQ ILOPV GHWHUPLQHG E\ (30$ DQG FRPSDUHG ZLWK WKH EXON DOOR\ FRPSRVLWLRQ $OOR\LQJ (OHPHQW %XON $OOR\ :HLJKW 3HUFHQW 6SXWWHUHG 7KLQ )LOP :HLJKW 3HUFHQW $ $J &X 7L 0J 0Q )LJXUH 6(0 PLFURJUDSK RI 7 DOXPLQXP VDPSOH DIWHU KRXUV DW r&

PAGE 60

(QHUJ\ NH9 )LJXUH ('6 VSHFWUD RI GDUN DUHD RI 6(0 LPDJH VKRZQ LQ ILJXUH VKRZLQJ D ODFN RI FRSSHU DQG VLOLFRQ (QHUJ\ NH9 )LJXUH ('6 VSHFWUD IRU EULJKW DUHDV LQ 6(0 LPDJH ILJXUH FRQVLVWHQW ZLWK D SUHFLSLWDWH ULFK LQ FRSSHU $WRPLF IRUFH PLFURVFRS\ $)0f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

PAGE 61

)LJXUH $)0 DPSOLWXGH LPDJH RI 7 DOXPLQXP WKLQ ILOP DIWHU VWUHVV UHOD[DWLRQ DW r& IRU KRXUV 7KH VFDQQHG DUHD LV SP E\ SP 7KH SULPDU\ VWUHQJWKHQLQJ WHFKQLTXHV XVHG LQ WKH 7 DOOR\ DUH VROLG VROXWLRQ VWUHQJWKHQLQJ GXH WR DOOR\LQJ HOHPHQWV RI PDJQHVLXP DQG PDQJDQHVH >'DY %UR@f DQG SUHFLSLWDWLRQ KDUGHQLQJ GXH WR DOOR\LQJ HOHPHQW FRSSHUf 6LOYHU LV DGGHG WR WKLV DOOR\ b ZHLJKW SHUFHQWf WR VWDELOL]H WKH FRSSHU SUHFLSLWDWHV >'DY@ 7KH HIIHFWLYHQHVV RI SUHFLSLWDWLRQ KDUGHQLQJ LV FRQWUROOHG E\ WKH VL]H DQG GLVWULEXWLRQ RI WKH SUHFLSLWDWHV LQ WKH KRVW FU\VWDOV >&DK@ 7(0 ZDV XVHG WR HYDOXDWH WKH IRUPDWLRQ RI SUHFLSLWDWHV LQ WKH DOXPLQXP FU\VWDOV )LJXUH D VKRZV D SODQ YLHZ 7(0 LPDJH RI D 7 VDPSOH 7KH VPDOO QP WR QPf GDUN VSRWV LQ WKLV LPDJH DUH WKH &X$O SUHFLSLWDWHV LQ WKH DOXPLn QXP PDWUL[ )LJXUH OE VKRZV WKH LQGH[HG HOHFWURQ GLIIUDFWLRQ SDWWHUQ IURP WKLV VDPSOH ZLWK WKH DOXPLQXP GLIIUDFWLRQ ULQJV DQG WKH &X$ GLIIUDFWLRQ VSRWV LGHQWLILHG 7KH EULJKW ILHOG 7(0 LPDJHV DOVR JLYH VRPH LQGLFDWLRQ DV WR WKH VL]H RI WKH DOXPLQXP FU\VWDOV &DXWLRQ PXVW EH XVHG LQ HYDOXDWLQJ JUDLQ VL]H IURP WKLV LPDJH EHn FDXVH WKH HWFKLQJ SURFHVV XVHG LQ WKH VDPSOH SUHSDUDWLRQ ZRXOG OLNHO\ DWWDFN VPDOO FU\V

PAGE 62

$OXPLQXP *UDLQV Ef )LJXUH 7(0 PLFURJUDSKV RI 7 WKLQ ILOP VDPSOH Df EULJKW ILHOG LPDJH VKRZLQJ DOXPLQXP FU\VWDOV DQG &X$ SUHFLSLWDWHV Ef LQGH[HG HOHFWURQ GLIIUDFWLRQ SDWWHUQ VKRZn LQJ WKH GLIIUDFWLRQ ULQJV IRU WKH $ f DQG f SODQHV DQG IRU WKH $&Xf SODQH

PAGE 63

WDLV IDVWHU WKDQ ODUJH FU\VWDOV $OVR D 7(0 LPDJH RQO\ H[DPLQHV D VPDOO DUHD RI WKH VDPSOH DQG D ODUJH DUHD PXVW EH YLHZHG WR GHWHUPLQH WKH WUXH JUDLQ VL]H GLVWULEXWLRQ 7KH DSSDUHQW DYHUDJH FU\VWDO VL]HV ZDV EHWZHHQ DQG SP :KLOH GLUHFW PHDVXUHPHQW RI WKH DOXPLQXP JUDLQ VL]H E\ HWFKLQJ ZDV XQVXFFHVVn IXO IRUPDWLRQ RI WKH FRSSHU ULFK SUHFLSLWDWHV JLYHV VRPH LQGLFDWLRQ RI WKH DOXPLQXP JUDLQ VL]H 3UHFLSLWDWHV WHQG WR QXFOHDWH DQG JURZ DW WULSOH SRLQWV ZKHUH WKUHH FU\VWDOV FRPH WRJHWKHU >5HH@ :KLOH WKH GLUHFW FDOFXODWLRQ RI JUDLQ VL]H IURP WKH SUHFLSLWDWLRQ GHQVLW\ LV TXHVWLRQDEOH WKH FRPSDULVRQ RI GLIIHUHQW VDPSOHV DQG LQIHUULQJ FKDQJHV LQ FU\VWDO VL]H EDVHG RQ FKDQJHV LQ SUHFLSLWDWLRQ GHQVLW\ LV UHDVRQDEOH )LJXUH VKRZV D 6(0 PLFURn JUDSK RI D 7 VDPSOH WKDW ZDV KHDWHG WR r& DQG WKHQ FRROHG WR r& RYHU KRXUV WR WHVW WKH VWDELOLW\ RI WKH PLFURVWUXFWXUH 7KH QXPEHU RI FRSSHU ULFK SDUWLFOHV LQ ILJXUH DUH RI WKH VDPH RUGHU DV WKDW RI ILJXUH LH SDUWLFOHV LQ WKH a E\ SP DUHD RI WKH PLFURJUDSK 7KHUHIRUH WKH SUHFLSLWDWH VL]H DQG GHQVLW\ RI WKHVH VDPSOHV LV DSSDUHQWO\ VWDEOH RYHU VHYHUDO WHPSHUDWXUH F\FOHV WR r& )LJXUH 6(0 LPDJH RI 7 DOXPLQXP VDPSOH DIWHU WHPSHUDWXUH F\FOLQJ IURP r& WR r& RYHU KRXUV

PAGE 64

63877(5 7,0( PLQ )LJXUH $XJHU GHSWK SURILOH RI 7 WKLQ ILOP 7KH ILOP ZDV VSXWWHUHG ZLWK NH9 DUJRQ LQ D [ PP UDVWHU 7KH HIIHFW RI R[LGDWLRQ RQ D ELPHWDOOLF DFWXDWRUfV SHUIRUPDQFH LV D FRQFHUQ $(6 DQDO\VLV ZDV XVHG WR PHDVXUH WKH R[LGH WKLFNQHVV RQ D 7 WKLQ ILOP VDPSOH WKDW KDG EHHQ F\FOHG WR r& WLPHV )LJXUH VKRZV WKH $XJHU GHSWK SURILOH 7KH UDVWHUHG LRQ JXQ XVHG LQ WKLV DQDO\VLV VSXWWHUHG $ DW DQ DSSUR[LPDWH UDWH RI QP SHU PLQXWH 7KXV WKH R[LGH WKLFNQHVV RQ WKLV VDPSOH LV DSSUR[LPDWHO\ WR QP 3RZGHU [UD\ GLIIUDFWLRQ DQDO\VLV ZDV DOVR SHUIRUPHG RQ WKH 7 WKLQ ILOP VDPSOHV $ EURDG VFDQ ZDV WDNHQ RI D VDPSOH F\FOHG WR r& RQFH WR GHWHUPLQH WKH SKDVHV SUHVHQW LQ WKH VDPSOH 2QO\ HOHPHQWDO DOXPLQXP DQG VLOLFRQ IURP WKH VXEVWUDWH ZHUH GHWHFWHG LQ WKHVH VFDQV ;5' VFDQV ZHUH WKHQ UXQ RQ VDPSOHV WKDW ZHUH KHDW WUHDWHG DW r& RU r& IRU XS WR KRXUV 7KH VDPSOHV ZHUH VFDQQHG IRU WKHLU f DQG f SHDNV )LJXUHV DQG VKRZ WKH GLIIUDFWLRQ SHDNV DIWHU GLIIHUHQW VWUHVV UHOD[DWLRQ WLPHV DW r& 1RWH WKDW WKHUH LV D VXEVWDQWLDO VKLIW LQ WKH SHDN SRVLWLRQ EHWZHHQ WKH LQLWLDO VFDQ DQG WKRVH WKDW KDYH XQGHUJRQH VWUHVV UHOD[DWLRQ IRU WKH VKRUWHVW WLPH RI WZR

PAGE 65

)LJXUH $OXPLQXP f [UD\ GLIIUDFWLRQ SHDNV IRU 7 WKLQ ILOP VDPSOHV DIWHU YDULRXV VWUHVV UHOD[DWLRQ WLPHV DW r& )LJXUH $OXPLQXP f [UD\ GLIIUDFWLRQ SHDNV IRU 7 WKLQ ILOP VDPSOHV DIWHU YDULRXV VWUHVV UHOD[DWLRQ WLPHV DW r&

PAGE 66

KRXUV /LQH VKDSH DQDO\VLV ZDV SHUIRUPHG RQ WKH ;UD\ GDWD WR GHWHUPLQH FU\VWDOOLWH VL]H DV D IXQFWLRQ RI VWUHVV UHOD[DWLRQ EXW WKH FU\VWDOOLWH VL]HV YDULHG UDQGRPO\ DQG QR FRUUHODn WLRQ ZDV IRXQG $OXPLQXP )LJXUH VKRZV WKH VWUHVV YHUVXV WHPSHUDWXUH SORW IRU WKH DOXPLQXP DOOR\ WKLQ ILOP 7KHVH GDWD ZHUH FROOHFWHG DIWHU DQ LQLWLDO WHPSHUDWXUH F\FOH WR r& WR VWDELn OL]H WKH ILOPfV PLFURVWUXFWXUH $V FDQ EH VHHQ IURP WKHVH GDWD WKH URRP WHPSHUDWXUH UHVLGXDO WHQVLOH VWUHVV LQ WKH ILOP LV 03D YHUVXV 03D IRU 7 03D IRU $O 6L&X DQG 03D IRU SXUH $/ 7KH ILOP EHJLQV WR SODVWLFDOO\ GHIRUP XQGHU D FRPSUHVn VLYH VWUHVV RI 03D DW r& $ VDPSOH ZDV DOVR FRROHG WR r& LQ OLTXLG QLWURJHQ WR UHGXFH WKH URRP WHPSHUDWXUH UHVLGXDO VWUHVV 8SRQ ZDUPLQJ WR URRP WHPSHUD WRR 7HPSHUDWXUH & )LJXUH 6WUHVV YHUVXV WHPSHUDWXUH SORW IRU DOXPLQXP

PAGE 67

WXUH WKH WHQVLOH VWUHVV ZDV UHGXFHG IURP 03D WR 03D D b UHGXFWLRQ 6WDUWLQJ DW WKH UHGXFHG URRP WHPSHUDWXUH VWUHVV RI 03D WKH ILOP EHJLQV WR SODVWLFDOO\ GHIRUP DW D ORZHU WHPSHUDWXUH RI r& XQGHU D FRPSUHVVLYH VWUHVV RI 03D 8SRQ FRROLQJ IURP r& WR URRP WHPSHUDWXUH WKH WHQVLOH VWUHVV UHWXUQHG WR 03D 7KH LVRWKHUPDO VWUHVV UHOD[DWLRQ VHHQ LQ WKHVH ILOPV DW GLIIHUHQW WHPSHUDWXUHV LV VKRZQ LQ ILJXUHV WKURXJK DORQJ ZLWK WKH ILWWHG FXUYH DQG WKH HTXDWLRQ $W r& WKH VWUHVV UHOD[DWLRQ ZDV EHVW ILW ZLWK D VLQJOH H[SRQHQWLDO FXUYH )RU WHPSHUDWXUHV EHWZHHQ r& DQG r& WKH VWUHVV UHOD[DWLRQ GDWD ZDV ILW ZHOO E\ D GRXEOH H[SRQHQWLDO FXUYH $W r& WKH QRLVH LQ WKH VWUHVV UHOD[DWLRQ GDWD LV VR ODUJH WKDW WKH ILWWHG FXUYH LV XQUHOLDEOH 6DPSOHV ZHUH DOVR KHDWHG DW r&PLQXWH WR WKH WHVWLQJ WHPSHUDWXUH WR HYDOXn DWH WKH HIIHFW RI UDPSLQJ UDWH RQ VWUHVV UHOD[DWLRQ $JDLQ WKLV KDG OLWWOH DIIHFW RQ VWUHVV UHOD[DWLRQ 7KH FXUYH ILWWLQJ SDUDPHWHUV DUH FRPSLOHG LQ WDEOH )LJXUH ,VRWKHUPDO VWUHVV UHOD[DWLRQ LQ DOXPLQXP WKLQ ILOP DW r&

PAGE 68

6WUHVV 03D 6WUHVV 03D 7LPH PLQXWHV )LJXUH ,VRWKHUPDO VWUHVV UHOD[DWLRQ LQ DOXPLQXP WKLQ ILOP DW r& )LJXUH ,VRWKHUPDO VWUHVV UHOD[DWLRQ LQ DOXPLQXP WKLQ ILOP DW r&

PAGE 69

6WUHVV 03D 6WUHVV 03D LJXUH ,VRWKHUPDO VWUHVV UHOD[DWLRQ LQ DOXPLQXP WKLQ ILOP DW r& )LJXUH ,VRWKHUPDO VWUHVV UHOD[DWLRQ LQ DOXPLQXP WKLQ ILOP DW r&

PAGE 70

7DEOH 5HVLGXDO VWUHVV DIWHU LQILQLWH UHOD[DWLRQ WLPH FU FKDQJH LQ VWUHVV DW VKRUW $Ff DQG ORQJ $FVWLPHV DQG WKH WLPH FRQVWDQWV [f IRU WKH WZR H[SRQHQWLDO HTXDWLRQV IRU DOXPLQXP 7HPS r&f &7c 03Df $D 03Df 7 PLQf $D! 03Df PLQf 7DEOH &KHPLFDO FRPSVLWLRQ RI VSXWWHU GHSRVLWHG DOXPLQXP DOOR\ GHWHUPLQHG E\ (30$ FRPSDUHG ZLWK WKH EXON DOOR\ FRPSRVLWLRQ $OOR\ (OHPHQW %XON $OOR\ :HLJKW 3HUFHQW 6SXWWHUHG )LOP :HLJKW 3HUFHQW $, 0J &U 7KH VDPSOH WKDW ZDV FRROHG WR r& ZDV WHVWHG IRU VWUHVV UHOD[DWLRQ DW r& $W WKLV WHPSHUDWXUH WKH WHQVLOH VWUHVV LQ WKH ILOP ZDV 03D DQG WKHUH ZDV QR VWUHVV UHOD[n DWLRQ REVHUYHG RYHU IRUW\HLJKW KRXUV RI WHVWLQJ 7KH (30$ FKHPLFDO DQDO\VLV RI WKH VDPSOHV LV VKRZQ LQ 7DEOH $JDLQ WKHUH DUH GLIIHUHQFHV LQ FRPSRVLWLRQ EHWZHHQ WKH VSXWWHUHG WKLQ ILOP DQG WKH EXON DOOR\ SUHVXPDEOH GXH WR VSXWWHU GHSRVLWLRQ HIIHFWV >=KH@

PAGE 71

)LJXUH 6(0 PLFURJUDSK RI DOXPLQXP DOOR\ WKLQ ILOP VDPSOH WKDW ZDV WHPSHUDWXUH F\FOHG WR r& DW Or&PLQXWHDQG KHOG DW r& IRU PLQXWHV DQG WKHQ FRROHG WR URRP WHPSHUDWXUH RYHU KRXUV )LJXUH 6(0 PLFURJUDSK RI DOXPLQXP DOOR\ WKLQ ILOP WKDW ZDV KHOG DW r& IRU KRXUV

PAGE 72

)LJXUH $)0 KHLJKW LPDJH RI DOXPLQXP DOOR\ WKLQ ILOP ZLWKRXW KHDW WUHDWn PHQW 6FDQ DUHD LV _LP E\ _WP =UDQJH QP DQG 506 URXJKQHVV LV QP )LJXUH $)0 KHLJKW LPDJH RI DOXPLQXP DOOR\ WKLQ ILOP KHDW WUHDWHG DW r& IRU KRXUV 6FDQ DUHD LV WP E\ QP =UDQJH QP DQG 506 URXJKQHVV LV QP

PAGE 73

7DEOH 506 URXJKQHVV DQG =UDQJH IRU DOXPLQXP DOOR\ WKLQ ILOPV WKDW KDYH XQGHUJRQH GLIIHUHQW KHDW WUHDWPHQWV r& r& 7LPH DW 7HPSHUDWXUH +RXUVf 506 5RXJKQHVV QPf =5DQJH QPf 506 5RXJKQHVV QPf =5DQJH QPf )LJXUH DQG VKRZ 6(0 PLFURJUDSKV RI DOOR\ WKLQ ILOPV WKDW KDYH XQGHUJRQH GLIIHUHQW KHDW WUHDWPHQWV ('6 ZDV XVHG WR H[DPLQH WKH GLIIHUHQFH EHWZHHQ WKH OLJKW DQG GDUN DUHDV VHHQ LQ ILJXUH 1R GLIIHUHQFH LQ FRPSRVLWLRQ EHWZHHQ WKHVH WZR DUHDV ZDV IRXQG 7KH FKDQJHV LQ $)0 VXUIDFH URXJKQHVV ZLWK VWUHVV UHOD[DWLRQ LQ VDPSOHV WKDW ZHUH KHDW WUHDWHG DW r& DQG f& IRU XS WR KRXUV LV VKRZQ LQ 7DEOH DQG ILJXUHV DQG 7KH 506 URXJKQHVV DQG =UDQJH WKH KHLJKW GLIIHUHQFH EHWZHHQ WKH KLJKHVW DQG ORZHVW SRLQW RQ WKH VDPSOHf DUH UHSRUWHG LQ 7DEOH )LJXUH VKRZV WKH $)0 KHLJKW LPDJH RI D DOOR\ VDPSOH ZLWK QR KHDW WUHDWPHQW DQG ILJXUH VKRZV WKH $)0 KHLJKW LPDJH RI D VDPSOH WKDW KDG EHHQ KHOG DW r& IRU KRXUV 7KH VL]H RI IHDWXUHV LQ WKHVH LPDJHV DUH QHDUO\ WKH VDPH EXW WKH 506 URXJKQHVV DQG WKH =UDQJH KDYH GRXEOHG IRU WKH VDPSOH KHOG D r& 7DEOH f 7KH GDWD LQ 7DEOH VKRZV WKDW WKH FKDQJH LQ VXUIDFH URXJKQHVV RFFXUUHG LQ WKH ILUVW WZR KRXUV RI WKH KHDW WUHDWPHQW DQG WKDW KHDW WUHDWPHQW WHPSHUDWXUH KDV OLWWOH HIIHFW RQ WKH UDWH RU WKH PDJQLWXGH RI WKH URXJK

PAGE 74

63877(5 7,0( PLQ )LJXUH $XJHU GHSWK SURILOH RI DOOR\ WKLQ ILOP )LOP VSXWWHUHG ZLWK NH9 DUJRQ LQ D PP [ PP UDVWHU QHVV FKDQJH $(6 GHSWK SURILOH GDWD LQ )LJXUH VKRZV WKDW WKH VXUIDFH R[LGH ZDV EHWZHHQ QP DQG QP WKLFN 3RZGHU [UD\ GLIIUDFWLRQ DQDO\VLV RI WKLQ ILOPV KHDW WUHDWHG DW r& RU r& IRU XS WR KRXUV VKRZHG WKDW WKH f SHDN VKLIWHG ORZ ,R IRU KHDW WUHDWHG WR KRXUVf YHUVXV XQKHDWHG VDPSOHV 1R DGGLWLRQDO SHDN VKLIWV ZHUH GHWHFWHG IRU VDPSOHV KHDW WUHDWHG ORQJHU WKDQ WZR KRXU 7KH SHDN VKLIW IRU WKH f VKRZHG WKH VDPH WUHQG ZLWK D VKLIW RI r 6DPSOHV ZHUH DOVR VFDQQHG WR GHWHFW FKDQJHV LQ WKH $O0J SHDN KHLJKWV RU ZLGWKV WKDW ZRXOG LQGLFDWH D FKDQJH LQ WKHVH SUHFLSLWDWHV 1R PHDVXUDEOH FKDQJH ZDV VHHQ LQ WKHVH SHDNV LQGLFDWLQJ WKDW WKH SUHFLSLWDWHV KDG QRW FRDUVHQHG /LQH VKDSH DQDO\VLV ZDV DOVR SHUIRUPHG RQ WKH ;UD\ GDWD WR GHWHUPLQH FU\VWDOOLWH VL]H DV D IXQFWLRQ RI VWUHVV UHOD[DWLRQ EXW QR FRUUHODWLRQ ZDV IRXQG ZLWK VWUHVV UHOD[DWLRQ WLPH RU WHPSHUDWXUH

PAGE 75

2 7HPSHUDWXUH & )LJXUH 6WUHVV YHUVXV WHPSHUDWXUH SORW RI DOXPLQXP DOOR\ WKLQ ILOP $OXPLQXP )LJXUH VKRZV WKH VWUHVV YHUVXV WHPSHUDWXUH SORW IRU WKH DOXPLQXP DOOR\ WKLQ ILOP IRU WKH ILUVW DQG VHFRQG WHPSHUDWXUH F\FOH WR r& 7KH DV GHSRVLWHG WHQVLOH VWUHVV LQ WKLV ILOP LV 03D EHIRUH WKH ILUVW F\FOH 'XULQJ WKH ILUVW WHPSHUDWXUH F\FOH DIWHU GHSRVLWLRQ WKH ILOP VWUHVV GHFUHDVHV OLQHDUO\ ZLWK LQFUHDVLQJ WHPSHUDWXUH XS WR r& %HWZHHQ r& DQG r& WKH VWUHVV LQ WKH ILOP LQFUHDVHV ZLWK LQFUHDVHG WHPSHUDn WXUH SUHVXPDEOH GXH YDFDQFLHV PLJUDWLQJ WR DQG EHLQJ GHVWUR\HG DW WKH JUDLQ ERXQGDULHV >7RZ@ 7KLV EHKDYLRU ZDV W\SLFDO IRU 7 DQG ILOPV DOVR GXULQJ WKH ILUVW WKHUPDO F\FOH 2QFH WKH H[FHVV YDFDQFLHV ZHUH HOLPLQDWHG WKH VORSH DJDLQ EHFRPHV OLQHDU XS WR r& $W r& WKH VORSH LQFUHDVHV DQG EHFRPHV XQVWDEOH XS WR r& ZKHUH WKH ILOP LV XQGHU D FRPSUHVVLYH VWUHVV RI 03D 8SRQ FRROLQJ WKH VORSH RI WKH VWUHVVWHPSHUDWXUH FXUYH LV PXFK OHVV WKDQ GXULQJ KHDWLQJ DQG WKH ILOP RQO\ UHDFKHV D WHQVLOH VWUHVV RI

PAGE 76

63877(5 7,0( PLQ )LJXUH $XJHU GHSWK SURILOH RI DOXPLQXP DOOR\ WKLQ ILOP )LOP VSXWWHUHG ZLWK NH9 DUJRQ LQ D [ PP UDVWHU 03D DW URRP WHPSHUDWXUH 2Q WKH VHFRQG WHPSHUDWXUH F\FOH WKH ILOP DWWDLQV D VOLJKWO\ KLJKHU FRPSUHVVLYH VWUHVV RI 03D DW r& DQG UHWXUQV WR D WHQVLOH VWUHVV RI 03D XSRQ UHWXUQLQJ WR URRP WHPSHUDWXUH :KHQ WKH VDPSOH ZDV UHPRYHG IURP WKH )OH[XV LW ZDV QRWLFHG WKDW WKH VSHFXODU UHIOHFWLRQ RI WKH ILOP ZDV JUHDWO\ UHGXFHG GXH WR VXUIDFH R[LGH %DVHG RQ WKH UHGXFHG VSHFXODU UHIOHFWLRQ DQG WKH ORZHU VORSH RI WKH VWUHVV WHPn SHUDWXUH FXUYH LW ZDV DVVXPHG WKDW WKH ILOP KDG R[LGL]HG EDGO\ 7KLV ZDV FRQILUPHG E\ WKH $XJHU GHSWK SURILOH GDWD IURP WKDW VDPSOH VHH )LJXUH f $V FDQ EH VHHQ WKH R[\JHQ FRQFHQWUDWLRQ IDOOV RII YHU\ JUDGXDOO\ 7KLV LQGLFDWHV WKDW WKH ILOP KDG R[LGL]HG WR D GHSWK JUHDWHU WKDQ WKH RU 7 ILOPV VHH ILJXUHV DQG f

PAGE 77

&RSSHU 7KH VWUHVV YHUVXV WHPSHUDWXUH SORW IRU WKH FRSSHU WKLQ ILOP LV VKRZQ LQ ILJXUH :KLOH WKLV SORW LV TXLWH FRPSOH[ WKH GDWD DUH FRQVLVWHQW ZLWK WKH ILOP EHLQJ R[LGL]HG GXULQJ KHDW WUHDWPHQW $V WKH VWUHVVWHPSHUDWXUH SORW LQGLFDWHV WKH ILOP ZDV VWDEOH GXULQJ WKH ILUVW WKHUPDO F\FOH WR RQO\ r& $QRWKHU ILOP ZDV WHVWHG IRU VWUHVV UHOD[DWLRQ DW r& IRU VHH ILJXUH f 7KH FRSSHU ILOP R[LGL]HG LQ WKLV WHVW DV LQGLFDWHG E\ D JUHDWO\ UHGXFHG WKH VSHFXODU UHIOHFWLRQ DQG WKH ILOP VWUHVV LQFUHDVHG DIWHU KRXUV GXH WR R[LGDWLRQ 7R LQKLELW WKH R[LGDWLRQ RI WKH FRSSHU WKLQ ILOP D QP WKLFN SURWHFWLYH DOXPLn QXP OD\HU ZDV GHSRVLWHG RQ WKH FRSSHU 7KLV VDPSOH ZDV WKHQ KHDWHG WR r& IRU KRXUV LQ D WXEH IXUQDFH WR WHVW WKH VWDELOLW\ RI WKLV SURWHFWLYH FRDWLQJ 7KH ILOP UHWDLQHG LWV VSHFXODU UHIOHFWLRQ GXULQJ WKLV WHVW )URP DQ $XJHU GHSWK SURILOH RQ WKLV VDPSOH WKH R[LGH WKLFNQHVV LV DSSUR[LPDWHO\ QP VHH ILJXUH f )LJXUH 6WUHVV YHUVXV WHPSHUDWXUH SORW IRU D FRSSHU WKLQ ILOP

PAGE 78

2 7LPH +RXUV )LJXUH ,VRWKHUPDO VWUHVV WHVW RI D FRSSHU WKLQ ILOP FRQGXFWHG DW r& )LJXUH $XJHU GHSWK SURILOH RI D FRSSHU ILOP FRDWHG ZLWK D SURWHFWLYH OD\HU RI DOXPLn QXP QP WKLFN 7KH ILOP ZDV VSXWWHUHG ZLWK NH9 DUJRQ LQ D [ PP UDVWHU

PAGE 79

7LWDQLXP 0DQJDQHVH DQG 1LFNHO 7KH VWUHVV YHUVXV WHPSHUDWXUH SORWV IRU WLWDQLXP PDQJDQHVH DQG QLFNHO DUH VKRZQ LQ ILJXUHV WKURXJK $OO RI WKHVH ILOPV R[LGL]HG VHYHUHO\ GXULQJ WKHLU ILUVW WHPn SHUDWXUH F\FOH XS WR r& r& DQG r& IRU WLWDQLXP PDQJDQHVH DQG QLFNHO UHVSHFn WLYHO\ 7KLV ZDV HYLGHQW IURP D FKDQJH LQ VSHFXODU UHIOHFWLRQ FRORU FKDQJH RI WKH VDPSOHV DQG E\ WKH VWUHVVVWUDLQ GDWD )RU WLWDQLXP ILJXUH f WKH UDSLG UHGXFWLRQ LQ WHQVLOH VWUHVV DW r& WR D FRPSUHVVLYH VWUHVV RI RYHU 03D DW r& DQG WKH ILOP PDLQWDLQn LQJ D KLJK WHQVLOH VWUHVV 03Df XSRQ FRROLQJ WR URRP WHPSHUDWXUH LQGLFDWHV VRPH FKDQJHV LQ WKH ILOP IURP R[LGDWLRQ 2[LGDWLRQ RI WKH PDQJDQHVH ILOP ILJXUH f UHVXOWV LQ DQ 03D VKLIW LQ WKH VWUHVV EHFRPLQJ KLJKO\ WHQVLOH DW MXVW RYHU r& 7KH QLFNHO ILOP ILJXUH f VKRZV D FKDQJH LQ VORSH RI WKH VWUHVV VWUDLQ FXUYH DW r& DQG WKH UHVLGXDO VWUHVV DW URRP WHPSHUDWXUH GURSSHG IURP 03D EHIRUH WR 03D DIWHU WKH 7HPSHUDWXUH & )LJXUH 6WUHVV YHUVXV WHPSHUDWXUH SORW IRU D WLWDQLXP WKLQ ILOP

PAGE 80

7HPSHUDWXUH & )LJXUH 6WUHVV YHUVXV WHPSHUDWXUH SORW IRU D PDQJDQHVH WKLQ ILOP 7HPSHUDWXUH & )LJXUH 6WUHVV YHUVXV WHPSHUDWXUH SORW IRU D QLFNHO WKLQ ILOP

PAGE 81

WKHUPDO F\FOLQJ GXH WR R[LGDWLRQ $Q DOXPLQXP SURWHFWLYH FRDWLQJ ZDV GHSRVLWHG RQWR D PDQJDQHVH VDPSOH EXW GLG QRW LQKLELW R[LGDWLRQ RI WKH VDPSOH &RSSHU*ROG ,QWHUPHWDOOLFV $IWHU WKH FRSSHUJROG ILOPV KDG XQGHUJRQH WKH KRPRJHQL]DWLRQ DQG RUGHULQJ KHDW WUHDWPHQW WKHLU FKHPLFDO FRPSRVLWLRQ ZDV DQDO\]HG E\ (30$ VHH 7DEOH f 7KH ILOPV ZHUH ZLWKLQ WR SHUFHQW RI WKH GHVLUHG FRPSRVLWLRQ ,Q SRZGHU [UD\ GLIIUDFWLRQ DOO RI WKH ILOPV VKRZHG D VKLIWV LQ WKH GLIIUDFWLRQ SHDN SRVLWLRQV IRUP GLVRUGHUHG WR WKH RUGHUHG VWDWH DQG DGGLWLRQDO SHDNV ZHUH REVHUYHG IURP WKH ORQJ UDQJH RUGHU $V UHSRUWHG DERYH IRU WH[WXUHG WKLQ ILOPV LW ZDV QRW SRVVLEOH WR FDOFXODWH WKH GHJUHH RI RUGHULQJ 7KH ILOPV ZHUH WKHUPDOO\ F\FOHG LQ WKH )OH[XV XS WR r& LQ DQ DUJRQ DWPRn VSKHUH )LJXUHV WKURXJK VKRZ WKH UHVXOWV RI WKHVH WHVWV 7KH &X$X ILOPV GLG QRW VKRZ D FKDQJH LQ VWUHVV QHDU WKH RUGHUGLVRUGHU WHPSHUDWXUH r& ILJXUH f 7KH &X$X ILOP VKRZHG D PDUNHG FKDQJH LQ VWUHVV DW WKH WUDQVLWLRQ WHPSHUDWXUH ILJXUH f DV GLG WR D OHVVHU GDJUHH WKH &X$X ILOP ILJXUH f +RZHYHU WKH RUGHUGLVRUGHU WUDQVLWLRQ RFFXUV ZHOO DERYH WKH WHPSHUDWXUH r&f DW ZKLFK WKH ILOPV EHJLQ WR SODVWLn FDOO\ GHIRUP 7DEOH &KHPLFDO FRPSRVLWLRQ RI FRSSHUJROG WKLQ ILOPV $WRPLF b &X $WRPLF b $ &X$X &X$X &X$X

PAGE 82

6WUHVV 03D 6WUHVV 03D 2 22 7HPSHUDWXUH & )LJXUH 6WUHVV YHUVXV WHPSHUDWXUH SORW IRU D &X$X WKLQ ILOP )LJXUH 6WUHVV YHUVXV WHPSHUDWXUH SORW IRU D &X$XWKLQ ILOP

PAGE 83

)LJXUH 6WUHVV YHUVXV WHPSHUDWXUH SORW IRU D &X$XWKLQ ILOP )LJXUH ,VRWKHUPDO VWUHVV UHOD[DWLRQ RI RUGHUHG DQG GLVRUGHUHG &X$X WKLQ ILOPV DW r&

PAGE 84

$ VHULHV RI LVRWKHUPDO VWUHVV UHOD[DWLRQ WHVW ZHUH SHUIRUPHG RQ WKHVH VDPSOHV 7KHVH WHVWV ZHUH FRQGXFWHG LQ DUJRQ VLQFH WKHVH ILOPV R[LGL]H LI KHDWHG WR r& LQ DLU 7KH WHVWV ZHUH UXQ DW r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r& WKH VORSH RI WKH VWUHVV WHPSHUDWXUH FXUYH FKDQJHG DW r& LQGLFDWLQJ D FKDQJH LQ WKH VWUXFWXUH RI WKH ILOP $W r& WKH ILOP ZDV LQ D VWDWH RI QR VWUHVV :KHQ WKH ILOP FRROHG WR URRP WHPSHUDWXUH LW ZDV XQGHU D WHQVLOH VWUHVV RI 03D DV VKRZQ LQ ILJXUH 7KH ILOP ZDV F\FOHG XS WR r& WZR DGGLWLRQDO WLPHV DQG VKRZHG QR K\VWHUHVLV LQ WKH VWUHVVWHPSHUDWXUH SORW $Q LVRWKHUPDO VWUHVV UHOD[DWLRQ WHVW ZDV SHUIRUPHG RQ WKLV VDPSOH DW r& DQG WKH ILOP VKRZHG QR VWUHVV UHOD[DWLRQ RYHU D KRXU SHULRG 7KH VDPSOH ZDV WKHQ WHVWHG DW 7HQFRU XS WR r& 7KH UHVXOWV RI WKLV WHVW DUH VKRZQ LQ ILJXUH 7KH $O7L ILOP EHJLQV WR SODVWLFDOO\ GHIRUP DERYH r& XQGHU D FRPSUHVVLYH VWUHVV RI 03D $W r& WKH ILOP VWLOO PDLQWDLQV D FRPSUHVVLYH VWUHVV RI 03D 8SRQ UHWXUQLQJ WR URRP WHPSHUDWXUH WKH ILOP ZDV XQGHU D WHQVLOH VWUHVV RI QHDUO\ 03D $Q LVRWKHUPDO VWUHVV UHOD[DWLRQ WHVW ZDV DOVR SHUIRUPHG DW r& DQG DJDLQ WKH ILOP VKRZHG QR VWUHVV UHOD[DWLRQ RYHU D KRXU SHULRG 7KH QDWLYH SDVVLYDWLQJ R[LGH KHOG XS ZHOO IRU DOO WHVWV DW DOO WHPSHUDWXUHV 7KH VSHFXODU UHIOHFWLRQ RI WKH VDPSOH ZDV KLJK DIWHU DOO WKH WHVWV 7KHUH ZDV QR YLVLEOH FKDQJH LQ WKH R[LGH DIWHU F\FOLQJ WR r& )LJXUH VKRZV WKH $XJHU GHSWK SURILOH IRU DQ

PAGE 85

)LJXUH 6WUHVV YHUVXV WHPSHUDWXUH IRU $O7L WKLQ ILOP $O7L ILOP WKDW ZDV KHDWHG WR D PD[LPXP RI r& 7KH R[\JHQ FRQFHQWUDWLRQ GURSV VKDUSO\ DIWHU PLQXWHV RI VSXWWHULQJ )LJXUH VKRZV WKH $XJHU GHSWK SURILOH IRU D VDPSOH KHDWHG WR r& 7KH R[\JHQ FRQFHQWUDWLRQ GURSV WR QHDU ]HUR DIWHU PLQXWHV RI VSXWWHULQJ LQGLFDWLQJ WKDW WKH VXUIDFH LV SDVVLYDWHG EXW WKH R[LGH OD\HU LV DERXW WKUHH WLPHV WKLFNHU IRU VDPSOHV KHDWHG WR r& DUH WKDQ VDPSOHV KHDWHG RQO\ WR r& 0RGHOLQJ WKH (IIHFWV RI 2[LGH 7KLFNQHVV RQ D %LPHWDOOLF $FWXDWRUfV &XUYDWXUH $V KDV EHHQ VHHQ ZLWK PDQ\ RI WKH PDWHULDOV H[DPLQHG LQ WKLV VWXG\ R[LGDWLRQ KDV D VLJQLILFDQW LPSDFW RI WKH SHUIRUPDQFH RI WKH PDWHULDO )RU WKH DOXPLQXP DOOR\V LW ZDV QRW SRVVLEOH WR WHVW IRU WKH HIIHFWV RI WKH R[LGH OD\HU VLQFH RQH DOZD\V IRUPV ZKHQ DOXPLn QXP LV H[SRVHG WR DLU 7R HYDOXDWH WKH HIIHFW RI WKH R[LGH RQ WKH SHUIRUPDQFH RI D ELPHn WDOOLF VWULS WKH FKDQJH LQ FXUYDWXUH RI D ELPHWDOOLF VWULS DV D IXQFWLRQ RI WKH $, OD\HU WKLFNQHVV ZDV FDOFXODWHG 7KLV FDOFXODWLRQ LV EDVHG RQ D ELPHWDOOLF VWULS WKDW FRQVLVWV RI

PAGE 86

3($. +(,*+7 3($. +(,*+7 63877(5 7,0( PLQ )LJXUH $XJHU GHSWK SURILOH RI $O7L VDPSOH WKDW ZDV KHDWHG WR r& 63877(5 7,0( PLQ )LJXUH $XJHU GHSWK SURILOH RI $O7L VDPSOH WKDW ZDV KHDWHG WR r& SP RI DOXPLQXP RQ SP RI VLOLFRQ ZLWK DQ R[LGH ZKRVH WKLFNQHVV YDULHG IURP WR

PAGE 87

$,2 7KLFNQHVV PLFURQV )LJXUH &XUYDWXUH RI D ELPHWDOOLF VWULS DV D IXQWLRQ RI R[LGH WKLFNQHVV DV D SHUFHQWn DJH RI PD[LPXP FXUYDWXUH SP )LJXUH VKRZV WKH UHVXOWV RI WKLV FDOFXODWLRQ )URP WKLV FDOFXODWLRQ LW FDQ EH VHHQ WKDW WKH QP R[LGH OD\HU IRXQG RQ WKH DOXPLQXP DOOR\ VDPSOHV UHGXFHV WKH FXUYDn WXUH RI WKH ELPHWDOOLF VWULS E\ SHUFHQW EXW WKLFNHU R[LGHV ZRXOG KDYH D PRUH VLJQLILn FDQW LPSDFW 7KHUHIRUH FRQWUROOLQJ WKH R[LGH WKLFNQHVV LV FULWLFDO WR PD[LPL]LQJ WKH SHUIRUPDQFH RI D ELPHWDOOLF VWULS

PAGE 88

&+$37(5 ',6&866,21 ,QWURGXFWLRQ ,Q WKLV FKDSWHU WKH UHVXOWV DQG FRQFOXVLRQV WKDW KDYH EHHQ GUDZQ IURP WKLV UHVHDUFK ZLOO EH SUHVHQWHG DQG GLVFXVVHG ZLWK UHVSHFW WR WKH WKUHH REMHFWLYHV RI WKLV UHVHDUFK Lf HIIHFWLYHQHVV RI VWUHQJWKHQLQJ WHFKQLTXHV LQ WKLQ ILOPV LLf LGHQWLILFDWLRQ RI WKH PHFKDn QLVP UHVSRQVLEOH IRU LVRWKHUPDO VWUHVV UHOD[DWLRQ LQ WKLQ ILOPV LLLf LGHQWLILFDWLRQ RI PDWHULDOV IRU KLJK WHPSHUDWXUH DSSOLFDWLRQV ,Q DGGLWLRQ WR DVVLVW WKH GHYHORSPHQW RI ELPHWDOOLF DFWXDWRUV D ILJXUH RI PHULW KDV EHHQ GHYHORSHG DQG D QXPEHU FULWLFDO SDUDPn HWHUV KDYH EHHQ LGHQWLILHG WR DVVLVW LQ WKH VHOHFWLRQ RI PDWHULDOV IRU WKLV DSSOLFDWLRQ 7KLQ )LOP 6WUHQJWKHQLQJ 2I WKH ILYH VWUHQJWKHQLQJ WHFKQLTXHV XVHG LQ EXON PHWDOV WKLV VWXG\ H[DPLQHV WKH HIIHFWLYHQHVV RI WZR RI WKHVH VROLG VROXWLRQ VWUHQJWKHQLQJ DQG PXOWLSKDVH KDUGHQLQJ SUHFLSLWDWLRQDJH KDUGHQLQJf LQ WKLQ ILOPV :KLOH QRW QRUPDOO\ FRQVLGHUHG D VWUHQJWKHQn LQJ WHFKQLTXH WKH HIIHFWLYHQHVV RI RUGHUHG LQWHUPHWDOOLFV WR UHVLVW VWUHVV UHOD[DWLRQ ZDV DOVR H[DPLQHG 6ROLG VROXWLRQ DQG SUHFLSLWDWLRQDJH KDUGHQLQJ ZHUH VHOHFWHG EHFDXVH RI WKHLU HIIHFWLYHQHVV LQ VWUHQJWKHQLQJ EXON PHWDOV DQG EHFDXVH WKHVH PHFKDQLVPV DUH UHODn WLYHO\ LQVHQVLWLYH WR VPDOO YDULDWLRQ LQ FRPSRVLWLRQ ZKLFK LPSURYHV PDQXIDFWXUDELOLW\ 7KH RUGHUGLVRUGHU WUDQVLWLRQ WKDW RFFXUV LQ VRPH LQWHUPHWDOOLFV ZDV DOVR RI LQWHUHVW GXH WR WKH ODUJH FKDQJH LQ YROXPH WKDW IUHTXHQWO\ RFFXUV DW WKLV WUDQVLWLRQ 7KH RUGHU GLVRUn GHU WUDQVLWLRQ SURYLGHG D PHDQV RI HYDOXDWLQJ WKH HIIHFWLYHQHVV RI WKH RUGHUHG SKDVH WR UHVLVW VWUHVV UHOD[DWLRQ

PAGE 89

:LWK UHVSHFW WR PHDVXULQJ WKH VWUHQJWKHQLQJ HIIHFWV RI WKHVH WHFKQLTXHV WKHUH DUH D QXPEHU RI SDUDPHWHUV WKDW DUH RI LQWHUHVW :LWK EXON PDWHULDOV VWUHQJWKHQLQJ LV XVXDOO\ HYDOXDWHG E\ PHDVXULQJ WKH \LHOG VWUHQJWK RU XOWLPDWH WHQVLOH VWUHQJWK RI WKH PDWHULDO DW D WHPSHUDWXUH ,Q WKLQ ILOP ELPHWDOOLF VWUXFWXUHV WKLV LV QRW SRVVLEOH VLQFH VWUHVV DQG WHPn SHUDWXUH DUH GHSHQGHQW RQ RQH DQRWKHU 7KXV WKH WHPSHUDWXUH DQG VWUHVV DW ZKLFK SODVWLF GHIRUPDWLRQ EHJLQV DUH LQWHUUHODWHG DQG LV WKH SRLQW RQ WKH VWUHVV YHUVXV WHPSHUDWXUH FXUYH ZKHUH WKH VORSH VLJQLILFDQWO\ GHSDUWV IURP OLQHDU 7KLV VWXG\ LV DOVR FRQFHUQHG ZLWK WKH VWUHVV UHOD[DWLRQ WKDW RFFXUV LQ WKLQ ILOPV ZKLFK GHWHUPLQHV WKH VWDELOLW\ RI D ELPHWDOn OLF DFWXDWRU 6R WKH VWUHQJWKHQLQJ WHFKQLTXHV PXVW EH HYDOXDWHG ZLWK UHVSHFW WR WKHLU HIIHFWLYHQHVV DW LQFUHDVLQJ ERWK WKH WHPSHUDWXUH DQG WKHUHIRUH WKH VWUHVV DW ZKLFK SODVWLF GHIRUPDWLRQ RFFXUV DQG DW UHGXFLQJ LVRWKHUPDO VWUHVV UHOD[DWLRQ LQ WKH ILOP :KLOH LW ZRXOG EH UHDVRQDEOH WR DVVXPH WKDW WKH EXON VWUHQJWKHQLQJ WHFKQLTXHV ZRXOG EH HIIHFWLYH LQ WKLQ ILOPV WKH $O6L&X DOOR\ XVHG LQ PLFURHOHFWURQLFV ZRXOG VXJJHVW WKDW WKH VWUHQJWKHQ PHFKDQLVPV DUH OHVV HIIHFWLYH LQ WKLQ ILOPV ,Q WKH $O6L&X DOOR\ VLOLFRQ FRXOG SURYLGH VRPH VROLG VROXWLRQ VWUHQJWKHQLQJ EXW WKH VROXELOLW\ RI VLOLFRQ LQ DOXPLQXP DW URRP WHPSHUDWXUH LV YHU\ ORZ &RSSHU FRXOG SURYLGH SUHFLSLWDWLRQ VWUHQJWKn HQLQJ EXW GXH WR SURFHVVLQJ LV JHQHUDOO\ LQ WKH RYHUDJHG FRQGLWLRQ $V VKRZQ LQ )LJXUH WKH VWUHQJWK RI WKH $O6L&X DOOR\ LV RQO\ b JUHDWHU WKDQ SXUH DOXPLQXP ,Q FRQWUDVW DJH KDUGHQHG EXON DOXPLQXP FRSSHU DOOR\V FDQ KDYH \LHOG VWUHQJWKHQV RI ILYH WLPHV WKRVH RI SXUH DOXPLQXP :KLOH WKLV VXJJHVW WKDW EXON VWUHQJWKHQLQJ PHFKDQLVP PD\ QRW EH DV HIIHFWLYH LQ WKLQ ILOPV LW PXVW DOVR EH QRWHG WKDW WKH $O6L&X DOOR\ ZDV QRW GHYHORSHG IRU VWUHQJWK EXW WR UHGXFH DOXPLQXP VSLNLQJ LQWR VLOLFRQ ZDIHU 6Lf DQG WR UHGXFH HOHFWURPLJUDWLRQ &Xf >%RZ@ 2Q WKH RWKHU KDQG WKH DOOR\V HYDOXDWHG LQ WKLV VWXG\ ZHUH GHYHORSHG IRU VWUHQJWK 6ROLG 6ROXWLRQ 6WUHQJWKHQLQJ 7KH DOXPLQXP DOOR\ ZDV VHOHFWHG WR WHVW VROLG VROXWLRQ VWUHQJWKHQLQJ 7KLV DOOR\ LV VROLG VROXWLRQ VWUHQJWKHQHG E\ WKH DGGLWLRQ RI PDJQHVLXP :KLOH WKH PDMRULW\ RI

PAGE 90

WKH PDJQHVLXP VWD\V LQ VROXWLRQ VRPH $O0J DQG $O0J SUHFLSLWDWHV IRUP +RZHYHU WKHVH SUHFLSLWDWHV GR QRW SURYLGH WKH SULPDU\ VWUHQJWKHQLQJ &RPSDULQJ WKH VWUHVV WHPn SHUDWXUH FXUYH IRU DOXPLQXP )LJXUH f ZLWK WKH FXUYHV IRU SXUH DOXPLQXP DQG WKH $O6L&X DOOR\ )LJXUH f FOHDUO\ VKRZV WKDW VROLG VROXWLRQ VWUHQJWKHQLQJ LV HIIHFWLYH LQ WKLQ ILOPV 7KH URRP WHPSHUDWXUH UHVLGXDO VWUHVV IRU WKH DOOR\ LV 03D DV FRPSDUHG WR 03D IRU SXUH DOXPLQXP DQG 03D IRU WKH $O6L&X DOOR\ $OVR SODVWLF GHIRUPDWLRQ EHJLQV LQ WKH DOOR\ VDPSOH DW r& XQGHU D FRPSUHVVLYH VWUHVV RI 03D ZKLOH WKH SXUH DOXPLQXP DQG $O6L&X DOOR\ EHJLQ WR GHIRUP XQGHU WKH VDPH VWUHVV EXW DW PXFK ORZHU WHPSHUDWXUHV RI r& DQG r& UHVSHFWLYHO\ 7KHUHIRUH VROLG VROXn WLRQ VWUHQJWKHQLQJ KDV VLJQLILFDQWO\ LQFUHDVHG WKH SHUIRUPDQFH RI WKH ILOP 7HVWLQJ DOVR VKRZHG WKDW WKH DOOR\ ILOP LV PRUH UHVLVWDQW WR VWUHVV UHOD[DWLRQ WKDQ D SXUH DOXPLQXP RU $O6L&X ILOPV $W r& SXUH DOXPLQXP DQG WKH $O&X6L DOOR\ ILOPV H[SHULHQFH D b UHGXFWLRQ LQ VWUHVV RYHU KRXUV ZKLOH WKH DOOR\ ILOP RQO\ H[SHULHQFH D b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f WKH URRP WHPSHUDWXUH UHVLGXDO WHQVLOH VWUHVV IRU SXUH FRSSHU LV 03D DQG WKH ILOPV EHJDQ WR SODVWLFDOO\ GHIRUP DERYH r& XQGHU D FRPSUHVVLYH VWUHVV RI 03D 7KH VWUHVV WHPSHUDWXUH SORW IRU &X$X )LJXUH f VKRZV D UHVLGXDO URRP WHPSHUDWXUH WHQVLOH VWUHVV LQ WKH ILOP RI 03D DQG WKH ILOP EHJLQV WR SODVWLFDOO\ GHIRUP DW r& XQGHU D FRPSUHVVLYH VWUHVV RI 03D $V ZLWK DOXPLQXP VROLG VROXWLRQV RI FRSSHU DQG JROG QHDUO\ GRXEOHG WKH VWUHQJWK RI WKH IGP 6WUHVV UHOD[DWLRQ WHVWV ZHUH QRW GRQH RQ WKH SXUH FRSSHU ILOPV GXH WR R[LGDWLRQ

PAGE 91

*LYHQ WKH UHVXOWV RI WKH DOXPLQXP DQG FRSSHU JROG LW LV FOHDU WKDW VROLG VROXWLRQV DUH HIIHFWLYH LQ VWUHQJWKHQLQJ WKLQ ILOPV ,W KDV DOVR EHHQ VKRZQ WKDW WKLV VWUHQJWKHQLQJ PHFKDQLVP LV HIIHFWLYH DW UHGXFLQJ VWUHVV UHOD[DWLRQ LQ DOXPLQXP DOOR\V 3UHFLSLWDWLRQ DQG 0XOWLSKDVH +DUGHQLQJ 7KH 7 DOXPLQXP DOOR\ ZDV VHOHFWHG WR GHWHUPLQH WKH HIIHFWLYHQHVV RI SUHFLSLn WDWLRQ KDUGHQLQJ WR UHGXFH VWUHVV UHOD[DWLRQ 7KH 7 DOXPLQXP DOOR\ LV QRW VWUHQJWKn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n SHUDWXUH UHVLGXDO VWUHVV LV QHDUO\ 03D ZKLFK LV 03D KLJKHU WKDQ WKH DOOR\ )LJXUH f WLPHV JUHDWHU WKDQ SXUH DOXPLQXP DQG WLPHV JUHDWHU WKDQ WKH $O6L&X DOOR\ )LJXUH f 7KH 7 DOOR\ EHJDQ WR SODVWLFDOO\ GHIRUP DW r& XQGHU D FRPn SUHVVLYH VWUHVV RI 03D DJDLQ PXFK KLJKHU WKDQ SXUH DOXPLQXP RU WKH RWKHU DOXPLQXP DOOR\V $W r& WKH 7 DOOR\ WKLQ ILOP XQGHUJRHV D b UHGXFWLRQ LQ VWUHVV RYHU KRXUV )LJXUH f DV FRPSDUHG WR b IRU WKH VDPSOHV )LJXUH f DQG D UHSRUWHG YDOXH RI b IRU WKH SXUH DOXPLQXP DQG $O6L&X DOOR\ +RZHYHU DW r& WKH 7 ILOP LV XQGHU D WHQVLOH VWUHVV RI 03D DV FRPSDUHG WR 03D IRU WKH DOOR\ ILOPV 03D IRU WKH SXUH DOXPLQXP ILOP DQG 03D IRU WKH $O6L&X DOOR\ ILOP $W r& WKH 7 VDPSOH XQGHUJRHV D b VWUHVV UHOD[DWLRQ RYHU KRXUV XQGHU D 03D WHQVLOH VWUHVV )LJXUH f ZKLOH WKH DOOR\ ILOP XQGHUJRHV DQ b VWUHVV UHOD[DWLRQ XQGHU D

PAGE 92

03D WHQVLOH VWUHVV )LJXUH f 7KHUHIRUH SUHFLSLWDWLRQDJH KDUGHQLQJ LV HIIHFWLYH DW UHGXFLQJ VWUHVV UHOD[DWLRQ LQ WKLQ ILOPV 7KH SODQH YLHZ 7(0 SKRWRPLFURJUDSK LQ )LJXUH YHULILHV WKDW SUHFLSLWDWHV ZHUH SUHVHQW LQ WKH DOXPLQXP WKLQ ILOP ,Q WKLV LPDJH SUHFLSLWDWHV DUH FOHDUO\ VHHQ ZLWKLQ DOXPLQXP FU\VWDOV DQG WKH GLIIUDFWLRQ SDWWHUQ IRU &X$ LV VHHQ 7KH VL]H RI WKHVH SUHFLSLn WDWHV LV RQ WKH RUGHU RI QP WR QP ZKLFK LV DSSURSULDWH IRU DJH KDUGHQLQJ >%UR@ %DVHG RQ WKH WHVWLQJ DQG PLFURVWUXFWXUDO H[DPLQDWLRQ RI WKH VDPSOH LW FDQ EH FRQFOXGHG WKDW SUHFLSLWDWLRQDJH KDUGHQLQJ LV HIIHFWLYH LQ WKLQ ILOPV ,W VKRXOG DOVR EH QRWHG WKDW WKH DGGLWLRQ RI VLOYHU WR WKLV DOOR\ VWDELOL]HV WKH &X$O SUHFLSLWDWHV :LWKRXW WKLV VWDELOL]DWLRQ WKH FRSSHU ZRXOG SUHFLSLWDWH DW WKH JUDLQ ERXQGDULHV KDYH D FRDUVH GLVSHUVLRQ UDWKHU WKDQ D ILQH GLVSHUVLRQ DQG EH PXFK OHVV HIIHFWLYH DW VWUHQJWKHQLQJ WKH ILOP $ FRDUVH GLVSHUVLRQ RI fRYHUDJHGf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n GHUHG SKDVHV VLQFH WKH VHOI GLIIXVLRQ UDWH ZKLFK GHWHUPLQHV WKH FUHHS UDWH DQG VWUHVV UHOD[DWLRQ UDWH IRU PDQ\ PDWHULDOV ZHUH H[SHFWHG WR EH ORZHU LQ DQ RUGHUHG SKDVHV YHUVXV

PAGE 93

D GLVRUGHUHG SKDVH >&DK@ )LJXUH FOHDUO\ VKRZV WKDW WKLV SUHVXPSWLRQ ZDV WUXH IRU &X$X WKH RUGHUHG SKDVH VKRZHG D b UHGXFWLRQ LQ VWUHVV RYHU WZR GD\V FRPSDUHG ZLWK D b UHGXFWLRQ LQ VWUHVV IRU WKH GLVRUGHUHG SKDVH 7KH GLVRUGHUHG ILOPfV VWUHVV UHOD[DWLRQ UDWH LV DOPRVW IRXU WLPH IDVWHU WKDQ WKH RUGHUHG SKDVH ZKLOH WKH VWUHVV LV RQO\ b KLJKHU RQ WKH GLVRUGHUHG ILOP 7KH $O7L RUGHUHG LQWHUPHWDOOLF DOVR VKRZHG QR VWUHVV UHOD[DWLRQ DW r& LQ WZR GD\V 6WUHVV 5HOD[DWLRQ 0HFKDQLVP LQ 7KLQ )LOPV :KLOH LQFUHDVLQJ WKH VWUHQJWK RI WKH DOXPLQXP WKLQ ILOPV UHGXFHG WKH VWUHVV UHOD[DWLRQ UDWHV DQG PDJQLWXGHV LW GLG QRW HOLPLQDWH VWUHVV UHOD[DWLRQ 7KXV WKH PHFKDn QLVPV UHVSRQVLEOH IRU VWUHVV UHOD[DWLRQ LQ WKLQ ILOPV QHHGV WR EH GLVFXVVHG ,Q WKLV VWXG\ WKUHH PRGHV RI VWUHVV UHOD[DWLRQ KDYH EHHQ REVHUYHG 7KH ILUVW PRGH LV SODVWLF GHIRUPDn WLRQ RI WKH WKLQ ILOP >+HU@ ,I WKH VWUHVV DSSOLHG E\ WKH VXEVWUDWH LV JUHDWHU WKDQ WKH \LHOG VWUHQJWK RI WKH PDWHULDO WKH IGP ZLOO SODVWLFDOO\ GHIRUP DQG UHOD[ WKH VWUHVV 7KLV LV VHHQ DV D FKDQJH LQ VORSH RI WKH VWUHVVWHPSHUDWXUH FXUYHV XSRQ KHDWLQJ )LJXUHV f 7KH VHFRQG PRGH RI VWUHVV UHOD[DWLRQ RFFXUV RYHU VHYHUDO WHQV RI PLQXWHV DIWHU D VDPSOH KDV UHDFKHG WHPSHUDWXUH 7KLV LV VHHQ LQ WKH VWUHVV UHOD[DWLRQ FXUYHV RI WKH 7 DQG DOXPLQXP DOOR\ ILOPV )LJXUHV DQG f 7KH WKLUG PRGH RI VWUHVV UHOD[DWLRQ RFFXUV RYHU D QXPEHU RI GD\V DQG FDQ DJDLQ EH VHHQ LQ WKH VWUHVV UHOD[DWLRQ SORWV RI WKH 7 DQG VDPSOHV 7KH FKDOOHQJH LQ GHWHUPLQLQJ WKH VWUHVV UHOD[DWLRQ PHFKDQLVPV LQ WKLQ ILOPV RQ D VXEVWUDWH LV WKDW ILOP VWUHVV DQG WHPSHUDWXUH FDQ QRW EH LQGHSHQGHQWO\ YDULHG 7KLV HOLPLn QDWHV WKH XVH RI DQ $UUKHQLXV HTXDWLRQ WR FDOFXODWH WKH DFWLYDWLRQ HQHUJLHV RI WKH SURFHVVHV XQGHU LQYHVWLJDWLRQ ,W LV SRVVLEOH WR PDQLSXODWH VDPSOHV DW GLIIHUHQW WHPSHUDWXUHV WR WKH VDPH VWUHVV KRZHYHU WKH PLFURVWUXFWXUH DQG GLVORFDWLRQ DUUD\ RI D PDWHULDO DUH QRW VWDWH IXQFWLRQV 7KHUHIRUH WKHLU VWUXFWXUHV ZLOO EH TXLWH GLIIHUHQW DW GLIIHUHQW WHPSHUDWXUHV HYHQ LI WKH VWUHVV LV WKH VDPH ,Q DGGLWLRQ UHPRYLQJ WKH ILOP IURP WKH VXEVWUDWH ZRXOG LQ PDQ\

PAGE 94

FDVHV VXEVWDQWLDOO\ FKDQJH WKH ILOP LQ D QXPEHU RI ZD\V )RU VHOI SDVVLYDWLQJ PDWHULDOV VXFK DV H[DPLQHG LQ WKLV VWXG\ WKH EDFN VLGH RI WKH ILOP ZRXOG R[LGL]H XSRQ H[SRVXUH WR DLU ,W KDV EHHQ VKRZQ WKDW GLVORFDWLRQV LQWHUDFW GLIIHUHQWO\ ZLWK R[LGL]HG VXUIDFHV >1L[@ ,W ZRXOG DOVR EH GLIILFXOW WR FUHDWH DQG DSSO\ D ELD[LDO VWUHVV WR D IUHH VWDQGLQJ WKLQ ILOP DQG WR PHDVXUH VWUDLQ LQ WKH ILOP 7KHUHIRUH GLUHFW GHWHUPLQDWLRQ RI DQ DFWLYDn WLRQ HQHUJ\ DQG WKXV WKH PHFKDQLVPV UHVSRQVLEOH IRU WKH GLIIHUHQW PRGHV RI VWUHVV UHOD[n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r& WKH 7+ LV OHVV WKDQ ZKLFK LV VOLJKWO\ ORZHU WKDQ WKH WHPSHUDWXUH DW ZKLFK $QGUDGH FUHHS ZRXOG EH H[SHFWHG ,Q DGGLWLRQ QRQH RI WKH VWDJHV RI $QGUDGH FUHHS ZRXOG SURGXFH DQ H[SRQHQWLDO GHFD\ DV LV VHHQ LQ WKH VWUHVV UHOD[DWLRQ GDWD 6DPSOHV KHDWHG DW KLJKHU UDWHV VKRZHG QR WUDQVLWLRQ IURP VWDJH FUHHS WR VWDJH FUHHS ,Q WKH HDUO\ VWDJHV RI UHOD[DWLRQ WKH VWUHVV UHOD[DWLRQ UDWH GHFUHDVH EXW GRHV QRW JR WR ]HUR DQG WKH VWDJH VWUDLQ UDWH LV QRW D VLPSOH H[SRQHQWLDO >&DK@ GHFD\

PAGE 95

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f 'LIIXVLRQDO SURFHVV LQFUHDVH H[SRQHQWLDOO\ ZLWK WHPSHUDWXUH >&DK@ ZKLOH WKH SUHVHQW VWUHVV UHOD[DWLRQ VKRZV DQ LQYHUVH UHODWLRQVKLS WR WHPSHUDWXUH 'LIIXVLRQDO SURn FHVVHV VKRXOG UHVXOW LQ D ]HUR VWUHVV DW LQILQLWH WLPH LQVWHDG RI WKH ILQLWH UHVLGXDO VWUHVV SURMHFWHG IURP WKH IXUYH ILWV 7DEOHV DQG f ;UD\ OLQH VKDSH DQDO\VLV GLG QRW GHWHFWHG SRO\JRQL]DWLRQ RU VXEJUDLQ FRDOHVFHQFH VR WKHVH PHFKDQLVP FDQ EH HOLPLQDWHG ,Q DGGLWLRQ VXEJUDLQ FRDOHVFHQFH >&DK@ LV D SUHOLPLQDU\ VWHS WR UHFRYHU\ DQG UHFU\VWDOOL]DWLRQ DQG WKH PLFURVWUXFWXUH RI WKH VDPSOHV GLG QRW FKDQJH GXULQJ VWUHVV UHOD[DWLRQ )LJXUHV DQG f 7KXV UHFRYHU\ UHFU\VWDOOLn ]DWLRQ DQG JUDLQ ERXQGDU\ PLJUDWLRQ GR QRW DSSHDU WR EH UHVSRQVLEOH IRU WKH VWUHVV UHOD[n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f 7KHUH LV DOVR D ,R VKLIW LQ WKH f [UD\ GLIIUDFWLRQ SHDN 7KH LQ

PAGE 96

FUHDVH LQ VXUIDFH URXJKQHVV HTXDWHV WR D b FKDQJH LQ WKH WKLFNQHVV RI WKH _UP ILOP ZKLFK LV RI WKH VDPH RUGHU DV WKH FKDQJH LQ WKH G VSDFLQJ FDOFXODWHG IURP WKH [UD\ SHDN VKLIW RI r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n DULHV 7KXV WKHVH GLVORFDWLRQV VKRXOG PRYH DQG EH HOLPLQDWHG TXLFNOW 7KH GHJUHH RI UHOD[DWLRQ VKRXOG RQO\ EH OLPLWHG E\ WKH QXPEHU RI VRXUFHV RI PRELOH GLVORFDWLRQV 7KH LQFUHDVH LQ URXJKQHVV FRXOG EH FDXVHG E\ GLVORFDWLRQ EDQGV $OVR EHFDXVH RI WKH VWUHVV OHYHO DQG WHPSHUDWXUH RI WKHVH ILOPV WKH '00fV SUHGLFW WKDW ORJDULWKPLF FUHHS ZRXOG EH WKH GRPLQDQW SURFHVV )LJXUH f >&DK@ 0RGH 6WUHVV 5HOD[DWLRQ 7KH PHFKDQLVPV QRW \HW HOLPLQDWHG IRU PRGH VWUHVV UHOD[DWLRQ DUH JUDLQ ERXQGn DU\ VOLGLQJ VXUIDFH WR JUDLQ ERXQGDU\ GLIIXVLRQ DQG ORJDULWKPLF FUHHS 6XUIDFH WR JUDLQ

PAGE 97

ERXQGDU\ GLIIXVLRQ LV D GLIIXVLRQDO SURFHVV WKDW VKRXOG EH H[SRQHQWLDOO\ GHSHQGHQW XSRQ WHPSHUDWXUH ,W FRXOG LQFUHDVH VXUIDFH URXJKQHVV DV DWRPV PLJUDWH IURP WKH VXUIDFH WR WKH JUDLQ ERXQGDU\ +RZHYHU WKH GDWD GR QRW VKRZ D ODUJH WHPSHUDWXUH GHSHQGHQFH DQG WKH VXUIDFH URXJKQHVV GRHV QRW LQFUHDVH RYHU VWDJH 7KH WHPSHUDWXUH DW ZKLFK WKH VWUHVV UHOD[DWLRQ LV RFFXUULQJ LV DOVR ORZ IRU D GLIIXVLRQDO SURFHVV WR EH DFWLYH 7+ f 7KHUHn IRUH VXUIDFH WR JUDLQ ERXQGDU\ GLIIXVLRQ GRHV QRW DSSHDU WR EH WKH PHFKDQLVP UHVSRQVLEOH IRU WKLV PRGH RI VWUHVV UHOD[DWLRQ *UDLQ ERXQGDU\ VOLGLQJ DOVR QRW OLNHO\ WR EH WKH PHFKDn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ff DQG f RULHQWHG FU\VWDOV )RU WKLV WR EH WKH FDXVH RI FKDQJH LQ UHOD[DWLRQ UDWH WKH GLIIHUHQW RULHQWHG FU\VWDOV ZRXOG KDYH WR UHOD[ DW GLIIHUHQW UDWHV DQG WKLV VKRXOG EH QRWLFHDEOH LQ WKH $)0 LPDJHV $OVR QR LQFUHDVH LQ WKH URXJKQHVV RI WKH ILOP ZDV VHHQ LQ PRGH ,W LV NQRZQ WKDW GLVORFDn WLRQV SLOH XS DW JUDLQ ERXQGDULHV >0H\@ 'LVORFDWLRQ FOLPE FRXOG PRYH VRPH RI WKHVH GLVORFDWLRQV LQWR WKH JUDLQ ERXQGDU\ 7KLV ZRXOG UHVXOW LQ D UHGXFWLRQ RI VWUHVV DQG QR LQFUHDVH LQ VXUIDFH URXJKQHVV 7KHUHIRUH WKLV PHFKDQLVP LV FRQVLVWHQW ZLWK WKH REVHUYDn WLRQ PDGH LQ WKLV VWXG\ DQG LV SRVWXODWHG WR EH WKH RULJLQ RI WKH PRGH VWUHVV UHOD[DWLRQ +LJK 7HPSHUDWXUH $SSOLFDWLRQ 0LFUR DFWXDWRUV DUH QHHGHG WKDW FDQ RSHUDWH LQ VHYHUH HQYLURQPHQWV VXFK DV LQVLGH MHW HQJLQHV ,Q WKHVH VHYHUH HQYLURQPHQWV R[LGDWLRQ ZDV DVVXPHG WR EH D PDMRU FRQFHUQ

PAGE 98

+LJK WHPSHUDWXUH DSSOLFDWLRQ DOVR UHTXLUH PDWHULDOV WKDW PDLQWDLQ WKHLU VWUHQJWK DW WKH RSHUDWLQJ WHPSHUDWXUHV 7KHVH WRSLFV ZLOO EH H[DPLQHG LQ WKLV VHFWLRQ 2[LGDWLRQ 'XULQJ WKLV VWXG\ LW ZDV IRXQG WKDW R[LGDWLRQ LV D SHUYDVLYH SUREOHP DIIHFWLQJ DOO WHPSHUDWXUHV UDQJHV $V LV VKRZQ DERYH DQG LQ $SSHQGL[ & D QP WKLFN R[LGH IRUPLQJ RI WKH VXUIDFH RI D SP $ SP 6L ELPHWDOOLF VWULS UHGXFHV WKH FXUYDWXUH RI WKH VWULS E\ b 7KH DOXPLQXP OLWKLXP DOOR\ ZDV IRXQG WR EH XQVXLWDEOH IRU ELPHWDOOLF DFWXDn WRUV EHFDXVH WKH OLWKLXP LQFUHDVHG WKH R[LGDWLRQ UDWH )LJXUH f 7KH RWKHU HOHPHQWV WHVWHG DOVR EHKDYHG SRRUO\ $W HOHYDWHG WHPSHUDWXUHV FRSSHU WLWDQLXP PDQJDQHVH DQG QLFNHO DOO R[LGL]HG )LJXUHV f ,Q EXON PHWDOV DSSOLFDWLRQ WKH WKLFNn QHVV RI WKH SDVVLYDWLQJ R[LGHV LV RI OLWWOH FRQFHUQ DV ORQJ LW UHPDLQV LQWDFW ,Q WKLQ ILOPV WKH SDVVLYDWLQJ R[LGH FDQ FRQVXPH D ODUJH IUDFWLRQ RI WKH WKLFNQHVV 7KLV LV YHU\ GDPDJn LQJ WR D ELPHWDOOLF DFWXDWRU EHFDXVH PRVW R[LGHV KDYH D YHU\ ORZ &7( ZKLFK ZLOO UHGXFH WKH GLVSODFHPHQW RI WKH GHYLFH 7KXV D PDWHULDOfV R[LGDWLRQ UDWH LV D FULWLFDO SDUDPHWHU DQG WKRVH PHWDOV ZLWK WKH VORZHVW R[LGDWLRQ UDWH DUH EHVW VXLWHG IRU WKLV DSSOLFDWLRQ $OXPLQXP DQG VLOLFRQ ERWK KDYH YHU\ ORZ R[LGDWLRQ UDWHV GXH WR WKH IRUPDWLRQ RI SDVVLn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

PAGE 99

VWUHVV UHOD[DWLRQ DW r& DQG QR R[LGDWLRQ 8QIRUWXQDWHO\ QLFNHO DQG FRSSHU R[LGL]HG VR FRPSDULVRQV RI VWUHVV UHOD[DWLRQ FRXOG QRW EH PDGH )LJXUH RI 0HULW 7R DVVLVW HQJLQHHUV ZKR GHVLJQ ELPHWDOOLF DFWXDWRUV LQ VHOHFWLQJ WKH DFWLYH OD\HU PDWHULDO D ILJXUH RI PHULW KDV EHHQ GHYHORSHG 7KH ILJXUH RI PHULW LV GHVLJQHG WR SURYLGH DQ LQLWLDO HYDOXDWLRQ RI WKH W\SH RI PDWHULDO WKDW VKRXOG EH XVHG LH DOXPLQXP RU QLFNHO 7KH ILJXUH RI PHULW )20 LV H[SUHVVHG LQ HTXDWLRQ DV )20 0+70SWMn96 > )' ‘ &7( f Wff ( ‘ nnr 3Df ‘ )' ffI f n Y n n Y n 7HPSHUDWXUH FRPSHQVDWHG 'LVSODFHPHQW \LHOG VWUHQJWK PRGH : RQ PRGH ZKHUH 0+7 LV WKH PD[LPXP KRPRORJRXV WHPSHUDWXUH WKH KRPRORJRXV WHPSHUDWXUH DERYH ZKLFK WKH \LHOG VWUHQJWK RI WKH PDWHULDO UDSLGO\ GHFOLQHV HTXDO WR IRU PRVW PHWDOV DQG IRU PRVW LQWHUPHWDOOLFVf 027 LV WKH PD[LPXP RSHUDWLQJ WHPSHUDWXUH RI WKH DFWXDWRU <6 LV WKH \LHOG VWUHQJWK RI WKH PDWHULDO DW URRP WHPSHUDWXUH &7( LV WKH FRHIILFLHQW RI WKHUPDO H[SDQVLRQ IRU WKH DFWLYH OD\HU PDWHULDO ( LV
PAGE 100

7DEOH )LJXUH RI PHULW UDWLQJV IRU GLIIHUHQW PDWHULDOV DW GLIIHUHQW RSHUDWLQJ WHPSHUDn WXUHV DQG RSWLPL]HG IRU GLVSODFHPHQW RU IRUFH ) 0D[ 7HPS r&f r&f f&f )RUFH 'LVSODFPHQW ) ) ) 6Q r r rr r rr rr $ rn r r r rr rr 7$ rr rnr r r rr rr 0J r rr r rr rr r $QQHDO $ r r} r rr rr rr b FZ $, r r r r rr rr &X r r r rnr rr rnr 1L r rnr rr rnr rr r &R rr rf rr rn r rn 66 rV r rr r rr rn FZ FROG ZRUNHG 66 VWDLQOHVV VWHDO

PAGE 101

&+$37(5 &21&/86,216 ,Q WKLV VWXG\ WKH WZR PRGHV RI RSHUDWLRQ IRU D ELPHWDOOLF DFWXDWRU IRUFH DQG GLVSODFHPHQW ZHUH VWXGLHG DQG HDFK PRGH KDV GLIIHUHQW PDWHULDOV UHTXLUHPHQWV WR PD[Ln PL]H SHUIRUPDQFH 7KH GLVSODFHPHQW PRGH LV GRPLQDWHG E\ WKH GLIIHUHQFH LQ WKH &7(fV RI WKH PDWHULDOV RI ELPHWDOOLF VWULS :KLOH SHUIRUPDQFH LQ WKH IRUFH PRGH LV D IXQFWLRQ RI ERWK
PAGE 102

7KH RUGHUGLVRUGHU WUDQVIRUPDWLRQ IRXQG LQ WKH FRSSHU JROG V\VWHP ZKLOH SURGXFn LQJ D ODUJH FKDQJH LQ VWUHVV IRU D VPDOO WHPSHUDWXUH FKDQJH RFFXUUHG DW WHPSHUDWXUHV ZKLFK JHQHUDWHG VWUHVVHV ZKLFK ZHUH DERYH WKHLU \LHOG VWUHQJWKV ,Q DGGLWLRQ WKHVH ILOPV R[LGL]HG DW WKH WHPSHUDWXUHV QHFHVVDU\ IRU WKH RUGHUGLVRUGHU WUDQVLWLRQ 7KHUHIRUH RUGHU GLVRUGHU WUDQVIRUPDWLRQV LQ WKH &X$X V\VWHP LV QRW XVHIXO IRU ELPHWDOOLF DFWXDWLRQ 7KHVH ILOPV SURYLGHG D PHDQV RI FRPSDULQJ WKH VWUHVV UHOD[DWLRQ UDWHV IRU RUGHUHG DQG GLVRUGHUVROLG VROXWLRQ SKDVHV 7KH RUGHUHG SKDVH VKRZHG RQH TXDUWHU WKH VWUHVV UHOD[n DWLRQ DV FRPSDUHG WR WKH GLVRUGHUHG SKDVH 7KHUHIRUH RUGHUHG LQWHUPHWDOOLF WKLQ ILOPV KDYH D KLJK UHVLVWDQFH WR VWUHVV UHOD[DWLRQ 7KLV ZDV DOVR VHHQ LQ WKH $O7L ILOPV WKDW VKRZHG QR VWUHVV UHOD[DWLRQ RYHU KRXUV DW r& 7KH $O7L ILOPV DOVR VKRZHG KLJK VWUHQJWK RQO\ SODVWLFDOO\ GHIRUPLQJ DW r& RU KLJKHU DQG R[LGDWLRQ ZDV RQO\ VLJQLILFDQW DERYH r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n YHUHO\ XSRQ WKHLU ILUVW KHDWLQJ WR r& 7KH FRSSHU ILOP HYHQ R[LGL]HG DW r& RYHU KRXUV 7KHUHIRUH PDWHULDOV IRU XVH LQ ELPHWDOOLF DFWXDWRUV PXVW KDYH D YHU\ ORZ R[LGDWLRQ UDWH

PAGE 103

&+$37(5 )8785( :25. ,Q WKH DUHD RI VWUHQJWKHQLQJ PHFKDQLVPV DGGLWLRQDO ZRUN FRXOG EH GRQH WR HYDOXn DWH GLIIHUHQW VROXWHV DQG SUHFLSLWDWHV WKDW FRXOG EH XVHG $V IRU WKH VWUHVV UHOD[DWLRQ PHFKDQLVPV DGGLWLRQDO WKHRUHWLFDO DQG DQDO\WLFDO ZRUN VKRXOG EH GRQH WR VXSSRUW RU GLVSURYH WKH FRQFOXVLRQV UHDFKHG LQ WKLV VWXG\ )URP WKH DQDO\WLFDO SHUVSHFWLYH WKH XVH RI QLFNHO DV D EDVH PDWHULDO KDV WKH DGYDQWDJHV WKDW 7(0 VDPSOHV FDQ EH PRUH HDVLO\ SURn GXFHG ZLWK OHVV FRQFHUQ WKDW WKH VDPSOH SUHSDUDWLRQ LV DIIHFWLQJ WKH GDWD 2[LGDWLRQ SUHVHQWV DQ LQWHUHVWLQJ SUREOHP IRU ELPHWDOOLF DFWXDWRUV EHFDXVH LW UHGXFHV GLVSODFHPHQW EXW DOVR LQFUHDVHV WKH VWUHQJWK RI WKH WKLQ ILOP >1L[@ 3XWWLQJ D SURWHFWLYH SRO\PHU FRDWLQJ RQ DQ DOXPLQXP WKLQ ILOP WR LQKLELW R[LGDWLRQ ZRXOG VLJQLILn FDQWO\ UHGXFH WKH \LHOG VWUHQJWK RI WKH DOXPLQXP EXW WKH R[LGH ZRXOG QRW UHGXFH GLVn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

PAGE 104

PD\ EH H[SRVHG WR KLJKO\ FXUYHG VWUXFWXUHV ZLWK ODUJH VWUHVV JUDGLHQWV DFURVV WKH ILOP WKLFNQHVV 7KLV FRXOG KDG GUDPDWLF HIIHFWV RQ WKH VWUHVV UHOD[DWLRQ ZLWKLQ WKH ILOPV DQG VKRXOG EH H[DPLQHG

PAGE 105

$33(1',; $ '(5,9$7,21 2) 5(/$7,216+,3 %(7:((1 <281*f6 02'8/86 $1' &2()),&,(17 2) 7+(50$/ (;3$16,21 7KH IROORZLQJ UHODWLRQVKLS EHWZHHQ D PDWHULDOfV .XO@ 6WDUWLQJ ZLWK WKH WKHUPRG\QDPLF UHODWLRQVKLS JLYHQ LQ HTXDWLRQ $ >'HK@ G9 9DL7 )3G3 $Of ZKHUH 9 LV YROXPH 3 LV SUHVVXUH 7 LV WHPSHUDWXUH D LV WKH FRHIILFLHQW RI WKHUPDO H[SDQn VLRQ DQG S LV WKH FRHIILFLHQW RI FRPSUHVVLELOLW\ %\ DOJHEUDLF PDQLSXODWLRQ DQG GLIIHUHQn WLDWLRQ ZLWK UHVSHFW WR SUHVVXUH ZKLOH NHHSLQJ WKH YROXPH WKH VDPH HTXDWLRQ $ LV GHn ULYHG G9 G7 G3 9D 9DDU G7? B 93 ?G3M\ 9D 3 DGI9 $f 7KH SDUWLDO GHULYDWLYH RI WHPSHUDWXUH ZLWK UHVSHFW WR SUHVVXUH DW FRQVWDQW YROXPH LV D SRVLWLYH YDOXHG IXQFWLRQ EHFDXVH WKH WHPSHUDWXUH RI D V\VWHP PXVW LQFUHDVH LQ UHVSRQVH WR LQFUHDVHG SUHVVXUH IRU WKH V\VWHP WR PDLQWDLQ D FRQVWDQW YROXPH

PAGE 106

7XUQLQJ WR WKH HTXDWLRQV WKDW GHILQH DQG UHODWH WKH GLIIHUHQW PDWHULDOV SURSHUWLHV >'LH@ HTXDWLRQ $ GHILQH WKH EXON PRGXOXV RI D PDWHULDO 3 A ‘ $f 7KH EXON PRGXOXV LV WKHQ UHODWHG WR WKH VKHDU PRGXOXV DQG
PAGE 107

%\ FRPELQLQJ HTXDWLRQV $ DQG $ WKH UHODWLRQVKLS EHWZHHQ WKH
PAGE 108

$33(1',; % &859( ),77,1* 2) 675(66 5(/$;$7,21 '$7$ 7KH VWUHVV UHOD[DWLRQ GDWD ZHUH FROOHFWHG RYHU D KRXU SHULRG 6WUHVV PHDVXUHn PHQWV ZHUH WDNHQ HYHU\ PLQXWHV SURGXFLQJ GDWD SRLQWV SHU VWUHVV UHOD[DWLRQ WHVW 7KHVH GDWD ZHUH H[SRUWHG IURP WKH 7HQFRU )OH[XV PRGHO LQWR DQ $6&,, IRUPDWWHG ILOH 7KLV $6&,, ILOH ZDV WKHQ LPSRUWHG LQWR 0LFURVRIW ([FHO ZKHUH DOO EXW WKH VWUHVV GDWD SRLQWV ZHUH HOLPLQDWHG 7KH GDWD ZHUH H[SRUWHG DV DQ $6&,, GDWD ILOH FRQWDLQLQJ RQO\ WKH VWUHVV GDWD SRLQWV GHOLPLWHG E\ FDUULDJH UHWXUQV DQG UHDG LQWR 0DWKFDG YHUn VLRQ ZKHUH WKH IROORZLQJ SURFHGXUH ZDV XVHG WR ILW WKH GDWD 7 $O # 44& 7KH GDWD DUH EHLQJ ILW WR DQ HTXDWLRQ RI WKH IRUP 6WUHVVWf $ % H[SPLnWf &H[SAWf L R LL 8VHG WR LQGH[ GDWD SRLQWV M Rf 8VHG WR LQGH[ GDWD SRLQWV ZKHQ QRW XVLQJ ILUVW GDWD SRLQWV U R 8VHG WR LQGH[ FROXPQV RI GDWD G 5($'351DLRRf 5HDGLQJ LQ GDWD 'LU U R =HURLQJ GDWD DUUD\ RQRW 7KLV LV WKH $ FRQVWDQW DQG LV PDQQXDOO\ VHOHFWHG 'OULR 'R 'OUL OQ'c L f RQRWf 6HWWLQJ XS DQ DUUD\ ZLWK WKH ORJ RI WKH GDWD 'LU GX 'LUM &RQYHUWLQJ GDWD SRVLWLRQ WR WLPH LQ PLQXWHV [ 'LU \ 'LU &RQYHUWLQJ GDWD LQWR VLQJOH FROXP DUUD\V rFM [M \F 'LUM 5HPRYLQJ WKH ILUVW GDWD SRLQWV IRU ILUVW ILW

PAGE 109

3HUIRUPLQJ OHDVW VTXDUHV OLQHDU ILW WR WKH QDWXUDO ORJDULWKP RI WKH GDWD P VORSH[F\Ff P Ra &DOFXODWLQJ WKH VORSH RI WKH OLQH 7KLV ,V WKH U_ FRQVWDQW E LQWHUFHSW[F\Ff &DOFXODWLQJ WKH \ ,QWHUFHSW RI WKH OLQH E E H[SEf E 6KRZLQJ WKH % FRQVWDQW F FRUU[F\Ff F A &DOFXODWLQJ WKH FRUUHODWLRQ IDFWRU IRU WKH ILW FDO P[ E &DOFXODWLQJ WKH FXUYH IURP GDWD ILW FRQVWDQWV 3ORW RI ILWWLQJ FXUYH ZLWK ORJ RI GDWD ,Q RI 6WUHVV WLPH PLQXWHV

PAGE 110

HFDLc RQRW H[SFDLMf &DOFXODWLRQ RI WKH ILW WR WKH GDWD 3ORW RI ILUVW FXUYH WR GDWD MR [M ;M 6HWWLQJ XS DUUD\V GLII 'M HFDLM 5HPRYLQJ WKH ILUVW FXQH ILW FRPSRQHQW IURP WKH GDWD LQGLLMI LQGLIA 7DNLQJ WKH QDWXUDO ORJDULWKP RI WKH GDWD 3UHIRUPLQJ OHDVW VTXDUHV OLQHDU ILW WR WKH QDWXUDO ORJDULWKP RI WKH VHFRQG VHW RI GDWD P IORRHGO &DOFXODWLQJ WKH VORSH RI WKH OLQH 7KLV ,V WKH U FRQVWDQW E LQWHUFHSW[LQGLII &DOFXODWLQJ WKH \ ,QWHUFHSW RI WKH OLQH E F H[SEf F 6KRZLQJ WKH & FRQVWDQW F FRUU[ ,QGLII F &DOFXODWLQJ WKH FRUUHODWLRQ IDFWRU IRU WKH ILW

PAGE 111

FDLM P[M E &DOFXODWLQJ FXUYH IURP GDWD ILW FRQVWDQWV HFDOM H[SAFDMM 3ORW RI ILWWLQJ FXUYH ZLWK ORJ RI GDWD 6WUHVV 03D

PAGE 112

FDOW P[ E HFDOW" H[SAFDOWA HFDOIM HFDOM ‘ FFDOW &DOFXODWLQJ ILWWLQJ &XUYH 3ORW RI ILWWHG FXUYH DQG GDWD 6WUHVV 03D HFDO&M HFDOM RQRW 3ORW RI WKH WZR FXUYH FRPSRQHQWV 6WUHVV 03D

PAGE 113

$33(1',; & &$/&8/$7,21 2) &)/$1*( ,1 &859$785( 2) $ %,0(7$//,& 675,3 %$6(' 21 2;,'( 7+,&.1(66 7KH UHGXFWLRQ LQ FXUYDWXUH RI D ELPHWDOOLF VWULS LV FDOFXODWHG DV D IXQFWLRQ RI WKH R[LGH WKLFNQHVV WKDW IRUPV RQ WKH VXUIDFH RI WKH DOXPLQXP ,Q WKLV PRGHO WKH EDVH ELPHn WDOOLF VWULS LV RQVLVWLQJ RI SP RI DOXPLQXP RQ SP RI VLOLFRQ 7KH R[LGH WKLFNQHVV YDULHV IURP WR SP LQ WKLFNQHVV ZLWK QR FKDQJH LQ WKH DOXPLQXP WKLFNQHVV 7KLV PRGHO ZDV EDVHG RQ WKH HTXDWLRQV GHYHORSHG E\ 3+ 7RZQVHQG >7RZ@ DQG ZDV LPSOHPHQWHG LQ 0DWKFDG YHUVLRQ OD\HUV OD\HU OD\HUV GHOWD R ORRRR 6HWWLQJ XS DUUD\V L M 'HLWD7 &KDQJH LQ WHPSHUDWXUH GR LR 6HW VWHS VL]H LQFUHDVH LQ R[LGH WKLFNQHVV 0DWHULDO SURSHUWLHV 6L $W $O P W 02 rrGHOWDU GHOWD &7( n &7(M &7( ( O 8 3D ( rr3D ( rn3D &DOFXODWLRQ RI FXUYDWXUH MR L L M L LLR G 'HLWD7&7( G2G2 fD\HU OD\HU O727GGf r r 9GOW

PAGE 114

9 D W OD\HUL LGm 9& 9r L ALGHKM GHOOD a 1SODLQ 9 ( W ‘ L LGF 1SWRUBW727B IW727 9H ‘sGFOX =M L UL GHOOD L =(GGX M \ 5 W L L LGHKD N68EGGXf r 727GH1 AnG8X U fr r 9f§f 727mf@ 'NGFMX ‘ \ ORR &DOFXODWLQJ FXUYDWXUH DV D SHUFHQWDJH RI PD[LPXP FXUYDWXUH WR NR }nA GHOWD GHOWD PLFURQV

PAGE 115

$33(1',; 0(&+$1,&$/ 3523(57,(6 2) (;$0,1(' 0(7$/6 $OOR\ 0HOWLQJ 3RLQW r&f
PAGE 116

%,%/,2*5$3+< $QJ $VN %DN %DO %DO %DU %DV %HD %H, %RZ %UR &DG &DK $QJHO -% 7HUU\ 6& DQG %DUWK 3: 6FLHQWLILF $PHULFDQ f f $VNHODQG '5 7KH 6FLHQFH DQG (QJLQHHULQJ RI 0DWHULDOV QG (GLWLRQ 3:6.HQW 3XEOLVKLQJ &RPSDQ\ %RVWRQ f %DNHU +f 0HWDOV +DQGERRNV $60 ,QWHUQDWLRQDO 0HWDOV 3DUN 2KLR f %DOGZLQ :0 7UDQV $,0( f %DOGZLQ )f +ROORZD\ 3+ %RUGHORQ 0 :DWNLQV 75 3DSHU SUHVHQWHG DW WKH 0DWHULDOV 5HVHDUFK 6RFLHW\ &RQIHUHQFH 6DQ )UDQFLVFR %DUWK 3: %HDWW\ && )LHOG /$ %DNHU -: DQG *RUGRQ *% 6ROLG6WDWH 6HQVRU DQG $FWXDWRU :RUNVKRS +LOWRQ +HDG 6RXWK &DUROLQD -XQH S f %DVDYDLDK 6 DQG *UHLQHU + $SSO 3K\V f f %HDUGVH\ 7 6FLHQWLILF $PHULFDQ f f %HORY 9 '$53$ 6HPLDQQXDO 3URJUHVV 5HSRUW RQ %LPHWDOOLF $FWXDWRUV IRU 0(06 FRQWUDFW QXPEHU '$%7& 'HFHPEHU %RZHU $) DQG )UHXQG /% $SSO 3K\V f f %URRNV &5 +HDW 7UHDWPHQW 6WUXFWXUH DQG 3URSHUWLHV RI 1RQIHUURXV $OOR\V $PHULFDQ 6RFLHW\ IRU 0HWDOV 0HWDOV 3DUN 2KLR f &DGRII DQG 0LOOHU ( 7KHUPRHOHFWULF 0DWHULDOV DQG 'HYLFHV 5HLQKROG 3XEOLVKLQJ &RUSRUDWLRQ 1HZ
PAGE 117

&DK &DKQ 5: DQG +DDVHQ 5 3K\VLFDO 0HWDOOXUJ\ UG HGLWLRQ 1RUWK +ROODQG 3K\VLFV 3XEOLVKLQJ 1HZ
PAGE 118

,OO +HU 0 +HUVKNRYLW] %OHFK ,$ DQG .RPHQ < 7KLQ 6ROLG )LOPV f +RJ +RJDQ +f 1HZ 6FLHQWLVW f f +,2 +ROORZD\ 3+ $GYDQFHV ,Q (OHWURQLFV DQG (OHFWURQ 3K\VLFV f +RY +RYL 9 DQG 3DDODVVDOR 3 $FWD 0HWDOOXUJLFD f +XG +XGVRQ -% 6XUIDFH 6FLHQFH DQ ,QWURGXFWLRQ %XWWHUZRUWK+HLQHPDQQ %RVWDQ f +XP +XPPHO 5( (OHFWURQLF 3URSHUWLHV RI 0DWHULDOV QG (GLWLRQ 6SULQJHU 9HUODJ 1HZ
PAGE 119

0DF 0DG 0DV 0H\ 1L[ 2KU 3RS 3RW 5HH 5LH 5RZ 6KH 6PL 6UR 6WR 6XK 6XK 7D\ 0DFKOLQ (6 0DWHULDOV 6FLHQFH LQ 0LFURHOHFWURQLFV 7KH 5HODWLRQVKLS %HWZHHQ 7KLQ )LOP 3URFHVVLQJ DQG 6WUXFWXUH *LUR 3UHVV &URWRQRQ +XGVRQ 1HZ
PAGE 120

7HU 7KR 7LP 7RX 7RZ 7VD 7XU :DU :HD :LQ :LW
PAGE 121

%,2*5$3+,&$/ 6.(7&+ -RQDWKDQ ) *RUUHOO ZDV ERUQ RQ )HEUXDU\ LQ 3DOR $OWR &DOLIRUQLD WR -DPHV DQG %LOOLH *RUUHOO +H ILQLVKHG KLV VHFRQGDU\ HGXFDWLRQ DW 0RQWD 9LVWD +LJK 6FKRRO LQ &XSHUWLQR &DOLIRUQLD LQ +H ZHQW WR &DOLIRUQLD 3RO\WHFKQLF 6WDWH 8QLYHUn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fV JURXS DQG LQLWLDOO\ VWXGLHG QRYHO WHFKQLTXHV IRU GLDPRQG JURZWK 2QH \HDU ODWHU KH EHJDQ H[DPLQLQJ WKLQ ILOP PHWDOOL]DWLRQ IRU XVH LQ PLFURPDFKLQHV ,Q 0D\ RI -RQDWKDQ UHFHLYHG KLV 3K' LQ 0DWHULDOV 6FLHQFH DQG (QJLQHHULQJ 8SRQ JUDGXDWLRQ -RQDWKDQ UHWXUQHG WR LQGXVWU\ DFFHSWLQJ D SRVLWLRQ DW 0RWRUROD 6HPLFRQGXFWRU LQ 3KRHQL[ $UL]RQD

PAGE 122

, FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW LV FRQIRUPV WR DFFHSWn DEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ L!FAI\nXU\n 3DXO + +ROORZD\ &KDLUPDQ 3URIHVVRU RI 0DWHULDOV 6FLHQFH DQG (QJLQHHULQJ FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW LV FRQIRUPV WR DFFHSWn DEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ $VKRN .XPDU $VVLVWDQW 3URIHVVRU RI 0HFKDQLFDO (QJLQHHULQJ FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW LV FRQIRUPV WR DFFHSWn DEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ 3URIHVVRU RI 0DWHULDOV 6FLHQFH DQG (QJLQHHULQJ FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW LV FRQIRUPV WR DFFHSWn DEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ `$ 5DMLY 5 6LQJK 3URIHVVRU RI 0DWHULDOV 6FLHQFH DQG (QJLQHHULQJ FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW LV FRQIRUPV WR DFFHSWn DEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ nDWH 3URIHVVRU RI 0DWHULDOV 6FLHQFH DQG (QJLQHHULQJ

PAGE 123

7KLV GLVVHUWDWLRQ ZDV VXEPLWWHG WR WKH *UDGXDWH )DFXOW\ RI WKH &ROOHJH RI (QJLn QHHULQJ DQG WR WKH *UDGXDWH 6FKRRO DQG ZDV DFFHSWHG DV SDUWLDO IXOILOOPHQW RI WKH UHTXLUHn PHQWV IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ 0D\ 6 :LQIUHG 0 3KLOOLSV 'HDQ &ROOHJH RI (QJLQHHULQJ .DUHQ $ +ROEURRN 'HDQ *UDGXDWH 6FKRRO

PAGE 124

Y /' 656_7< 2) )/25,'$ m9 PL +LOO ,,


THIN FILM METALLIZATION
FOR MICRO-BIMETALLIC ACTUATORS
By
JONATHAN FRANK GORRELL
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1998

ACKNOWLEDGMENTS
I would like to express my appreciation to advisor Dr. Paul H. Holloway for his
support, guidance and encouragement. I would also like to give special thanks for pro¬
viding a new perspective to some of the problems I encountered in this research. Thanks
are also due to my doctoral committee members Dr. DeHoff, Dr. Singh, Dr. Adair, and
Dr. Kumar for their assistance and interest in my work.
This work would not have been possible without the support of EG&G ICSensors
and the DARPA MEMS research program. I am also grateful to Dr. Hal Jerman for his
efforts in obtaining this grant and his support of my work.
I would also like to thank my friends and coworkers for their support and encour¬
agement and who made my return to school most enjoyable. I would like to thank Mark
Davidson for providing me with his insight into the workings of a university and scien¬
tific instrumentation.
Lastly, I would like to thank my parents James and Billie Gorrell for instilling a
love of learning and the belief in myself that I can achieve the goals I set for myself from
a very early age. Without these values I would have never tried much less have com¬
pleted this program.

TABLE OF CONTENTS
ACKNOWLEDMENTS i
ABSTRACT vi
CHAPTERS
1 INTRODUCTION 1
Motivation and Objective 1
Scope of Present Work 6
2 LITERATURE REVIEW. 8
Introduction 8
The Bimetallic Strip 8
The Simple Bimetallic Cantilever. 9
The Bimetallic Disk 10
The Effects of Plastic Deformation 12
Atomic Bonding 13
Plastic Deformation In Metals 17
Creep in Metals 19
Deformation Mechanism Maps (DMMs) 22
Annealing 22
Strengthening of Bulk Metals 24
Ordered Intermetallic Compounds 27
Thin Films 28
Enviromental Stability of Materials 35
3 EXPERIMENTAL PROCEDURE 37
Introduction 37
Thin Film Deposition 38
Copper Gold Heat Treatment 39
Stress and Stress Relaxation Measurements 39
Powder X-Ray Diffraction (XRD) 40
Atomic Force Microscopy (AFM) 42
Scanning Electron Microscopy (SEM) 42
iii

Electron-Probe Microanalysis (EPMA) 43
Transmisión Electron Microscopy (TEM) 43
Etching 44
Auger Electron Spectroscopy (AES) 44
Curve Fitting 45
Numerical Modeling 46
4 RESULTS 47
Introduction 47
T201 Aluminum 47
5052 Aluminum 59
2090 Aluminum 68
Copper. 70
Titanium, Manganese and Nickel 72
Copper-Gold Intermetallics 74
Aluminum-Titanium Intermetallics 77
Modeling the Effects of Oxide Thickness on a Bimetallic
Actuator’s Curvature 78
5 DISCUSSION 81
Introduction 81
Thin Film Strengthening 81
Solid Solution Strengthening 82
Precipitation and Multiphase Hardening 84
Ordered Phases and the Order Disorder Transition 85
Stress Relaxation Mechanism in Thin Films 86
Mode 2 Stress Relaxation 88
Mode 3 Stress Relaxation 89
High Temperature Application 90
Oxidation 91
Strength at Elevated Temperatures 91
Figure of Merit 92
6 CONCLUSIONS 94
7 FUTURE WORK 96
APPENDICES
A DERIVATION OF RELATIONSHIP BETWEEN YOUNG’S
MODULUS AND COEFFICIENT OF THERMAL EXPANSION.... 98
iv

B CURVE FITTING OF STRESS RELAXATION DATA 101
C CALCULATION OF CHANGE IN CURVATURE OF A
BIMETALLIC STRIP BASED ON OXIDE THICKNESS 106
D MECHANICAL PROPERTIES OF EXAMINED METALS 108
BIBLIOGRAPHY. 109
BIOGRAPHICAL SKETCH 113
v

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
THIN FILM METALLIZATION
FOR MICRO-BIMETALLIC ACTUATORS
By
Jonathan Frank Gorrell
May, 1998
Chairperson: Dr. Paul H. Holloway
Major Department: Materials Science and Engineering
In this study, eleven different thin film metallization systems were evaluated for
use in micro-bimetallic actuators for microelectromechanical structures. These films
were evaporated or sputtered onto silicon wafers. The film stress and stress relaxation
were determined by measuring changes in the wafer curvature. The phases and micro¬
structure of these films were evaluated with, scanning electron microscopy, transmission
electron microscopy, Auger electron spectroscopy, electron probe micro-analysis, X-ray
diffraction and line shape analysis, and atomic force microscopy.
Bimetallic actuator may be operated to generate either force or displacement. The
displacement mode is dominated by the coefficient of thermal expansion while the force
mode is a function of both Young’s modulus and coefficient of thermal expansion of the
active layer material. In both modes the maximum displacement or force is determined
by the material’s yield strength. A figure of merit was developed to aid in material
selection.
The 5052 aluminum alloy films showed that solid solution strengthening can
double the yield strength of a thin film. The T201 aluminum alloy films showed that
precipitates can increase yield strength by 2.5 times. The 2090 alloy film oxidized during
vi

the first heating. Based on isothermal stress relaxation data and changes in the micro¬
structure of the 5052 and T201 alloy thin films, two mechanisms involving logarithmic
creep have been postulated to cause stress relaxation. One mechanism is movement of
dislocations in slip systems that terminate at the surface while the other is dislocations
moving in slip systems that terminate at grain boundaries.
Copper gold intermetallics films oxidized and plastically deformed before the
order-disorder transformation occurred, but showed that ordered intermetallics have a
lower stress relaxation rate than the solid solution phase. The Al,Ti films showed no
stress relaxation at 450°C, plastically deformed only above 500°C, and had limited oxida¬
tion up to 800°C.
Nickel, copper, titanium, and manganese films all oxidized on their first heating to
350°C. The copper film also oxidized at 50°C over 48 hours. Calculations also showed
that the passivation oxide on aluminum alloys can significantly reduce preformance a
bimetallic actuator. Thus oxidation resistance is a significant requirement for materials
for thermal actuation.
vii

CHAPTER 1
INTRODUCTION
Motivation and Objective
The miniaturization of electronics has led to what is now called the information
age. More and faster electronics packaged in ever smaller volumes allow massive
information processing and exchange. Information processing systems are, however,
limited by the devices that allow them to perceive and affect the physical world. Sensors
and actuators have been connected to computers and control systems from the early days
of computing, but these sensors and actuators have been, until recently, large electrome¬
chanical devices. Researchers have begun to miniaturize electromechanical devices using
the technologies developed for microelectronics [Hog96]. These new micro sensors and
actuators have created a new class of devices called MEMS, microelectromechanical
systems [Mas95, Fra97, Ang83, Bea96, Mad97]. Integration of MEMS sensors and
actuators is enabling the development of mechanical, chemical and biological “smart
systems” that are able to interact with the physical world.
MEMS have already become a significant industry. Yearly sales are in the billions
of dollars and the market continues to grow rapidly [Mas95]. This market is currently
dominated by pressure sensors and accelerometers. The miniaturization of pressure
sensors has reduced sensor cost to the point that in medical applications the sensors can
be discarded after a single use. Accelerometers are predominately used to determine
when to deploy automobile air bags. Micro actuators have yet made as many inroads into
industry. Currently the biggest application of MEMS actuators is for ink jet print heads
1

2
Figure 1.1: SEM image of an EG&G IC Sensors bimetallic
thermally actuated micro-valve.
[ROW95]. The market of MEMS actuators will continue to develop as actuators are
needed to produce a truly “smart system” that can not only monitor processes but also
control them. However, additional research and development are needed to improve the
performance of micro actuators, and this is the general focus of this work.
Micro actuators have been designed based on a number of physical properties
including electrostatic, electromagnetic, piezoelectric, magnetostriction and thermal
activation (bimetallic, shape memory alloys, and thermopneumatic) [Mad97]. Table 1.1
shows a comparison of the different actuation techniques and their differences in perfor¬
mance. While all of these techniques have advantages for particular applications, this
study focuses on the thermally activated bimetallic actuator. Bimetallic actuators are of
interest because they can produce a relatively large force or displacement, the force is
relatively constant throughout the travel of the actuator and bimetallic actuators are easily
manufactured using semiconductor processing technology. Micro bimetallic actuators
have already been employed in commercially available products [Mad97, Jer94]. Figure
1.1 shows a bimetallic actuated micro valve produced by EG&G ICSensors produced by

Table 1.1: Comparison of actuation principles [Bel97].
3
Effect
Parameter
Bimetallic
Electro¬
static
Piezo¬
electric
Electro¬
magnetic
Thermo-
hydraulic
Force per
1 mm2, N
High (0.1.. 1)
Low (la7 .. 10-’)
High
High
High
Deflection, mm
High (0.01.. 1)
Low (io\.iaJ)
Low
High
High
Work per cycle per
1 mm2, J
High (10-7)
Low (10*1*)
High
High
Frequency range,
Khz (limiting factor)
Low 0.1-100
(Heal Transfer)
(Mech. Resonat
frequency)
High
(Mech. Resonat
frequency)
Low
Heat Transfer
\foltage, V
Low (3 ..12)
High (100 .. 300)
High
Low
Low
Current, mA
High (0.1 ..10)
Low (10-5 .. 10-4)
Low
High
High
Power Consumption,
W
High (10'4...0.1)
Low (lO-L.-lO-4)
Low
High
High
Efficiency, %
0.01 ... 0.5
0.001...0.01
Size: Length, mm
Width, mm
Small (10-2...l)
Small (10-\..0.1)
Small
Large
Large
Large
Additional
requirements
High \foltage
High \bl tage
(Material
Compatibility)
Magnetic field
(magnets)
Media separation
Cost
O.W
Low
High
High
High
aluminum metallization on silicon. Even though silicon is not a true metal, this structure
is still referred to as a bimetallic actuator.
The objective of this research is to improve the performance of micro-bimetallic
actuators and broaden their range of operating parameters (stress, displacement, tempera¬
ture range and stress relaxation) by identifying better or improve the materials of con¬
struction. In order to understand how to improve the performance of the bimetallic
actuator, it is important to understand how the device works. A bimetallic actuator is
made by bonding together two different materials that have a large difference in thermal
expansion, such as aluminum and silicon, to create a bimetallic strip or spring that is
heated or cooled. Upon heating, the material with the larger coefficient of thermal expan-

4
Aluminum Active Layer
20°C
100°C
Figure 1.2: Diagram of a bimetallic strip, showing how the differencein
thermal expansion bewteen the active layer, aluminum, and
passive layer, silicon, causes the bimetallic strip to bend.
sion, called the active layer, expands more than the material with the smaller coefficient,
called the passive layer (see Figure 1.2). As these two materials are bonded together, a
shear stress develops between the layers due to the different rates of expansion, producing
a force that causes the structure to bend. Upon cooling the reverse processes occurs and
the bimetallic strip returns to its original position, or deflects in the opposite direction
with continued cooling. The geometry may be a disk, cantilevered beam, or more com¬
plex geometry.
While the current generation of bimetallic actuators works well for a number of
applications, their use could be significantly expanded by improving the position and
force stability of these actuators and increasing the operable ambient temperature limit
from 75“C to 500°C. Force and position stability have been found to be a problem in

5
Heating
Aluminum
Valve Open
Cooling or
Stress Relaxation
of Metal
Figure 1.3: Cross Sectional Diagram of bimetallically actuated micro-valve.
some bimetallic actuators such as the one shown in figure 1.1. It is difficult to maintain a
constant flow from these valves over extended periods of time. The opening of the valve
decreases with time. Force and position instability in bimetallic actuators are caused by a
relaxation in the shear stresses between the active and passive layers of the bimetallic
strip which in turn reduces the force bending the structure. Figure 1.3 shows a cross
sectional diagram of a bimetallic actuated micro-valve and the effects of stress relaxation
on the valve operation. The inability to maintain a constant force or position for extend
periods of time complicates the usefulness of the device. These limitations result prima¬
rily from creep in the active metal layer of the bimetallic actuators. Thus, improvements
can be achieved through the identification and or development of better metal thin films.

6
While improvement of materials for bimetallic actuators may at first seem to be a
narrow subject, it requires examining much broader materials questions. How can a thin
metal film be strengthened? What are the atomic mechanisms responsible for the stress
relaxation in thin films? What materials are sufficiently chemically stable at elevated
temperatures to be useful? These questions are of general interest to the entire MEMS
field because micro-machines use thin metal films as structural components, and struc¬
tural failure and environmental stability are concerns in any mechanical device. Also, the
continued reduction of interconnect widths in microelectronic integrated circuits lead to
greater stresses in thin metal lines that may fail due to stress induced voiding (SIV)
[Bow74], Therefore, this research will examine questions of interest to many disciplines
The temperature change of an actuator is normally accomplished by resistive
heating. The passive layer of most micro-bimetallic actuators is silicon, and a resistive
heater is easily made by doping the silicon. A thin silicon oxide layer grown on the
silicon is all that is needed to keep the resistive heater from shorting to the active layer.
Cooling is accomplished by conduction of heat to the valve body and the surrounding
structure. The thermal masses of micro-bimetallic actuators are so small that they can be
cycled rapidly. Other heating and cooling methods could be used, such as a Peltier
element [Cad60]. In the present work, the mechanisms to change the temperature of the
bimetallic actuator is not of direct interest. Instead the present research is concerned with
the mechanical and thermal properties of the materials in the bimetal strip, and how they
affect actuation. Optimization of the performance of bimetallic strips will be the focus,
which directly translates into optimizing the performance of a bimetallic actuator.
Scope of Present Work
The three specific objectives of the present work are as follows, (i) The first
objective is to determine the effectiveness of different strengthening techniques for thin

7
film metallization. There are a number of techniques used to increase the yield and
ultimate tensile strength of bulk metal alloys, but their effectiveness in thin films has yet
to be determined. Increased yield strengths of the metal thin films would increase the
maximum force and displacement produced by an actuator, and could increase force and
displacement stability, (ii) The second objective is to evaluate use of materials at el¬
evated temperatures. The primary requirements for elevated temperature applications are
high strength over ther operation temperature range and resistance to creep and oxida¬
tion. (iii) The third objective is to identify the mechanism responsible for isothermal
stress relaxation in thin films. It is this relaxation that causes the force and position
instability seen over extended time periods.
To meet these objectives, eleven different materials were deposited and evaluated.
The evaluation focused on determining the material’s strength, environmental stability
and rate of stress relaxation. For those materials that showed promise and to understand
the underlying mechanisms responsible for the stress relaxation, the morphology and
microstructure of several samples were examined. The materials that were evaluated fall
into three groups: (i) aluminum alloys, (ii) elemental metals, and (iii) intermetallics. The
aluminum alloys were used to identify the mechanisms responsible for isothermal stress
relaxation, and to determine the effectiveness of the different strengthening techniques.
The elemental metals and intermetallics were evaluated at elevated temperatures. Some
of the intermetallics also exhibited ordering and non-linear thermal properties. The
evaluation of several materials systems allowed a wide variation in properties and there¬
fore provided significant insight into the materials requirements for MEMS devices.
The organization of this dissertation is as follows. Chapter 2 is a review of the literature
pertinent to this study. The experimental procedures used to deposit, test, and analyze the
samples are presented in chapter 3. Chapter 4 contains the results of these tests and
analyses. The results are discussed in chapter 5. The conclusions are summarized in
chapter 6 and suggestions for future work are presented in chapter 7.

CHAPTER 2
LITERATURE REVIEW
Introduction
Thin film metallization for bimetallic actuators are the focus of this research,
therefore a number of topics need to be reviewed. The first topic is the theory of the
bimetallic strip. In examining literature, the concern is to determine how the different
materials properties affect the performance of a bimetallic strip. With an understanding
of the effects of materials properties, improved materials selection can be used to opti¬
mize performance. After their identification, the origin and interdependence of these
properties can be discussed using the theories of atomic bonding and thermal expansion.
In addition, there are a number of processes that may affect the metallization of a bimetal¬
lic strip, such as creep, plastic deformation, and oxidation. Lastly, these bimetallic
actuators are manufactured with thin fdm technology and the effects of these processes
are discussed. There is no perfect material for use in all bimetallic actuators, since
different applications require different properties. With an understanding of the interde¬
pendence of the different materials properties, the tradeoffs that are needed for any given
application can be better selected.
The Bimetallic Strip
The formulation of the bimetallic strip problem was first reported, in part, by G.
Gerald Stoney [Sto09], Stoney determined the relationship between curvature and stress
in a bimetallic strip. S. Timoshenko [Tim25] then derived the equations for displacement
8

9
and force of a bimetallic strip versus change in temperature. Lastly, Townsend [Tow87]
generalized the equations for a multi-layer structure.
The formulation and solution of the bimetallic strip problem are dependent on the
geometry and boundary conditions used, therefore no single solution is generally appli¬
cable. Roark’s Formulas for Stress and Strain [You89] provide solutions for many shapes
and boundary conditions. This complexity of solutions and lack of solutions for unusual
shapes has led to the widespread use of finite element analysis to design bimetallic
actuators [Bar94] [Tsa92],
The Simple Bimetallic Cantilever
To determine the effects of materials properties on performance, the simple
bimetallic cantilever, simply supported on only one end, is examined. The simplifying
assumptions made in this case are that the length is much greater than the width or thick¬
ness, the materials are perfectly elastic, and the material returns to its original dimension
once the load is removed [Mey84], Figure 2.1 shows a diagram of this case where E, and
E2 are the Young’s moduli, L is the length of the strip, h is the thickness of the strip, a and
b are the thickness of the first and second layer, w is the width of the strip, a! and a2 are
the coefficients of thermal expansion, and the curvature of the bimetallic cantilever, k,
for a given change in temperature, AT, is[Tim25]
k =
6(a2-a,)(AT)(l + ^)2
b
en
For the case of both layers of the strip being the same thickness a = b and equal moduli
the equation for curvature reduces to [Tim25]

10
k-
Layer 1 E
Layer 2 a E
L
Layer 1 a
^ Layer 2 b
Figure 2.1: Diagram of bimetallic strip
(2)
The error induced by assuming the moduli are equal is only three percent for a factor of
two change in either material’s modulus. Therefore, the primary material’s property
affecting the displacement in this case is the difference in the coefficients of thermal
expansion.
The equation for the restraining force, R, that is needed to keep a bimetallic strip
from moving for a given temperature change DT is the maximum force a bimetallic strip
can produce. This restraining force is given by equation 3:
(3)
4
h-L
Thus the force produced by a bimetallic strip is a function of both the thermal expansion
and the Young’s moduli of the materials.
The Bimetallic Disk
To evaluate the effects of materials properties on performance for a more complex
case, the bimetallic disk is examined. Due to the complexity of the equation describing

3
li
0-1 1 1
0 10 20 30
Expansion Coefficient (XE-6/°C)
Figure 2.2: Equal potential lines for force and displacement for a bimetallic disk [Jer95],
the bimetallic disk the evaluation was performed by calculating and plotting the
equalpotential lines for maximum displacement (curvature) and force of a bimetallic disk
as a function of Young’s moduli and the coefficient of thermal expansion, as shown in
Figure 2.2 [Jer95], This model is based on a bimetallic strip where the passive layer is
silicon and both the active and passive layers are 5pm thick. In agreement with equations
(2) and (3) this figure shows that displacement is predominately limited by the thermal
expansion of the active layer material. There is some dependence on the Young’s modu¬
lus but this occurs only for a modulus lower than that found in most engineering materi¬
als. The equal potential line for force shows that thermal expansion is not as significant a
factor as in displacement. Therefore, while the relationships shown in Figure 2.2 for a
bimetallic disk are far more complex than the relationships determined by equations 2

12
and 3 for a single fixed cantilever, the trends and limiting materials properties are the
same. Displacement in a bimetallic structure is primarily limited by the thermal expan¬
sion of the active layer material. In contrast, the force generated by a bimetallic strip is
dependent on both the thermal expansion and Young’s modulus of the active layer mate¬
rial. These trends and limitations are expected to hold for other configurations as well.
The Effects of Plastic Deformation
The assumption that a material performance is perfectly elastic is only valid for a
range of applied stresses. Beyond this range of stress the material will not return to its
original shape when the stress is removed and the material is said to have undergone
plastic deformation [Mey84](see Figure 2.3). Plastic deformation in either layer of a
bimetallic device invalidates the relationships in equations (1), (2) and (3), and would
reduce the force or displacement produced from the strip. The effects of plastic deforma-
Figure 2.3. Typical stress strain curve for a ductile metal.

13
tion can be modeled as a non-linear reduction of the thermal expansion or Young’s modu¬
lus of a material. While not considered in the above equations that predict the force or
displacement of a bimetallic strip, the stress at which plastic deformation begins, called
the yield strength [Ask89, Mey84] (see Figure 2.3) is also a critical materials parameter.
A material’s yield strength determines the maximum force it can exert on another layer
and in turn may determine the maximum force or displacement a bimetallic strip can
achieve.
Therefore, the materials properties of interest depend upon whether the applica¬
tion requires force or displacement. In the displacement mode, the material’s thermal
expansion and yield strength are the primary limiting properties. In the force mode,
thermal expansion and Young’s modulus both contribute to the force generated by a
bimetallic strip, and yield strength is a limiting materials property.
Atomic Bonding
The atomic bonds that hold a material together also determine its thermal expan¬
sion and Young’s modulus. There are a number of aspects of atomic bonding that affect
these physical properties, and these aspects will be examined in this section. This review
is not intended to cover all aspects of thermal expansion, but only those that affect the
performance of a material in a bimetallic actuator.
The first aspect of atomic bonding that determines a material’s rate of thermal
expansion and its Young’s moduli is the strength of the atomic bonds. Figure 2.4 shows
the asymmetric potential energy well that an atom is drawn into when it bonds with
another atom [Ask89]. The depth of this potential well determines the strength of the
bond. The deeper the potential well, the stronger the bond. A deep potential well nor¬
mally has a larger dE/dx which is proportional to the Young’s modulus.
The rate of thermal expansion of a material is due to the asymmetry of the poten¬
tial energy well. The addition of energy to a material causes an increase in vibration of

14
the atoms. This increase in atomic vibration causes the atoms to move up in the potential
well. As the atoms move higher their time-averaged position increases (see Figure 2.4)
and the material expands [Tou75]. The deeper the potential well, the smaller the change
in the time-averaged position of the atoms for a given increase in temperature, and a
lower thermal expansion is observed. The rate of thermal expansion is frequently ex¬
pressed as the coefficient of thermal expansion, CTE. A material’s CTE is the derivative
of its thermal expansion curve, AL/L versus temperature. To a first approximation, the
thermal expansion is inversely proportional to the atomic bond strength. Thus, the
Young’s modulus and thermal expansion of a material are inversely proportional. Appen¬
dix A shows the thermodynamic proof of this relationship.
The second aspect of atomic bonding that affects the rate of thermal expansion
and the Young’s Modulus is the directionality of the bonding. The directionality of the
bonding results from the type of bonds that form between atoms, the crystallography of

15
the material, the number of nearest neighbors, the coordination number, and the bond
length. Graphite is a good example to illustrate the effects of directionality of bonding.
Graphite has a layered structure where the bonds within a layer are strong covalent bonds,
but the bonds between atoms in different layers are much weaker van der Waals bonds
[Eva79]. This difference in bonding causes the c-plane coefficient of thermal expansion
(CTE) to be 2.8xlO’60C1, while between the planes the CTE is 4.4xlO'6°C'.
In many instances, the directionality of bonding does not change the bond type but
still has a major affect on physical properties. Zinc illustrates this point well. The atoms
in zinc are bonded into a hexagonal close packed crystal structure, with metallic bonds
between all atoms. Along the c-axis the CTE for zinc is 5xl0‘6OC, but along the a-axis
it is 65xl0'6OC'1 [Tou75]. Thus there is a 13-fold difference in CTE between different
planes due to the crystallography that is determined by the bonding.
There are other materials with small anisotropy but unexpected physical proper¬
ties due to unusual crystallographic structure. Manganese is one such metal [Dea52],
While a-manganese has a cubic structure, it does not have a simple face centered cubic
(FCC) or body centered cubic (BCC) crystal structure [Ask89], The unit cell is based on
the BCC structure but contains 58 atoms in 29 pairs [Dea52]. The CTE for manganese is
22xl0'6OC, almost double what would be expected for a material with a melting tem¬
perature of 1244°C. In comparison, beryllium melts at 1278°C and has a CTE of only
12x 1 O'6 °C1 [ Wea7 6].
The directionality and crystallographic structure of a material are not always
static. Pure iron undergoes two allotropic transformations between solidification at
1538°C and room temperature. Iron first solidifies into a BCC structure at 1538°C,
converts to an FCC structure at 1394°C, and then converts back to a BCC structure at
912°C [Kra90], At each allotropic transformation, there is a significant change in thermal
expansion and other physical properties. Discontinuities in a material’s CTE could be
useful in a bimetallic actuator. Unfortunately the transitions in iron are far above the

16
Figure 2.5: Linear thermal expansion for intermetallic compound CuAu and
Cu,Au. The discontinuity in thermal expansion is due to the order-disorder
transion of these compounds [Tou75].
temperature currently being considered for practical devices. However, transformations
are used in shape memory alloy actuators [Mad97].
Another change in bonding that can significantly affect properties is the order-
disorder transformation exhibited by some intermetallic compounds [Tou75], Intermetal¬
lic compounds are made up of two or more elements, A and B. In ordered alloys, the A-B
bond may be stronger or weaker than the A-A or B-B bonds and thus there is a driving
force to form one type of bond over the other [Cah83]. Ordered intermetallic compound
may form directly upon solidification, as in AljTi and NiAl. However, there are other
intermetallic compounds that first solidify as a solid solution and then order at a lower
temperature. CdMg3, CuAu3, CuAu, and CUjAu are all examples of materials that
undergo a solid phase disorder/order transition [Hov64, Fed58, CuI78, War69], The

17
material may undergo a substantial contraction upon ordering (see Figure 2.5). This
contraction is due to the ability of the atoms to pack tighter in the ordered phases due to
the stronger atomic bonding. This transition may also increase the material’s yield
strength, reduce the creep rate, and increase electrical conductivity.
Plastic Deformation In Metals
As mentioned above, plastic deformation in either layer of a bimetallic actuator
will limit the force and displacement of the actuator. Therefore, the yield strength deter¬
mines the maximum force and displacement of an actuator. Atomic motion becomes
rapid as the temperature of a metal exceeds half its melting point, therefore the homolo¬
gous temperature, TH, is defined to be the ratio of the temperature to the melting point of
the material in degrees Kelvin. At TH < 0.5 dislocation motion is the dominant mecha¬
nism of plastic deformation. Inhibition of the motion of dislocations would strengthen
the material. This is in contrast to the thermal expansion and Young’s modulus of a
material that can not be changed by inhibiting dislocation motion. In addition, many of
the methods used to inhibit plastic deformation do not significantly alter the CTE or
Young’s modulus of the material. Therefore, improved yield strength can be used to
produce a better actuator. In this section, the atomic processes responsible for plastic
deformation will be examined, along with methods to inhibit these operations.
The two main processes that take place in a material during plastic deformation
are: dislocation movement and multiplication. Movement of dislocations produces a
change in shape of the material. However, the dislocation density in an annealed sample
is not great enough to produce the observed plastic deformation. Thus the dislocation
density is increased during plastic deformation by dislocation multiplication [Ree92,
Cah83],
With respect to dislocation motion, they can move through a crystal by either
dislocation glide or climb. Dislocation glide occurs when a dislocation moves on a slip

18
plane and in a slip direction of the crystal [Ree92]. The combination of a slip direction
lying along a slip plane creates a slip system. There are a limited number of slip systems
and they are determined by the crystallography of the material. Dislocation may not be
able to glide in response to an applied force because a slip system is blocked. On the
other hand, the stress resolved onto that slip plane in the slip direction may be below the
critical value to cause motion of dislocation. This is known as Schmid’s law [Ask89],
The second way for a dislocation to move is by climb, in which it moves between parallel
slip planes. Climb allows dislocations to move off blocked slip planes. It occurs by
vacancy movement [Ree92], but both the vacancy concentration and movement are very
dependent on the temperature of the material [Deh93], Dislocation climb only becomes a
dominant process at high homologous temperature TH. For most metals dislocation climb
becomes significant at TH>0.5 and for most intermetallics at TH>0.7. Above these tem¬
peratures dislocations may rapidly climb.
The process of dislocation multiplication can occur in many ways. One of the
more significant is the Frank-Read source [Ree92] by which dislocation glide can in¬
crease the dislocation density. In a Frank-Read source, a dislocation line on a slip plane
is pinned or blocked at two points. In response to an applied resolved shear stress the
dislocation bows out between these two points and forms an incomplete dislocation loop.
A critical stress exists, dependent on the dislocation line length, above which complete
dislocation loops form. This mechanism is complicated by the fact that dislocations can
pin one another, thus forming more Frank-Read sources. This is one of the reasons that
metals harden when they are cold worked, which is known as “work hardening” [Bro82],
There generally exists an inverse relationship between temperature and yield
stress independent of the mechanisms responsible for the deformation. For dislocation
glide, random thermal excitation helps dislocation overcome the Peierls-Nabarro barrier
[Mey84] and thus lowers the yield stress. The effects of temperature on dislocation

19
density and mobility have already been discussed and these effects also lead to reduced
yield stress with increased temperature. While an inverse relationship exists between
yield strength and temperature it is a non-linear function specific to each material.
The stress at which plastic deformation begins is also dependent on whether the
force is applied in tension or compression. There is no general relationship between the
compressive and tensile yield strength and most yield strength data for metals are for
tensile yield strength. As the thin films in a bimetallic actuator can be in tension or
compression the variation in the compressive versus tensile yield stresses could be sig¬
nificant.
Lastly, the crystallographic texture (which is the preference for one crystallo¬
graphic orientation) of a material also affects its yield strength [Mey84, Bro82], Baldwin
[Bal46] showed up to 30% change in tensile yield strength based on the direction of the
applied stress relative to the texture of rolled copper sheet.
Creep in Metals
In the preceding section the mechanisms responsible for the rapid plastic deforma¬
tion of materials in response to an applied stress were examined. However, materials
continue to plastically deform in response to applied stresses over extended periods of
time (from minutes to years). Time dependent plastic deformation is known as creep. In
this section the different types of creep and the mechanisms responsible for the time
dependent plastic deformation will be discussed.
Figure 2.6 [Cah83] shows the four major types of creep that occur in materials as
a function of the homologous temperature versus the resolved stress, a, normalized by the
shear modulus, jx.
At stress levels below which dislocation can move (crcrss/(X < 10“8) anelastic creep
occurs. In anelastic creep, interstitial atoms move to interstitial sites that have been
elongated by the applied stress [Cah83]. This movement of interstitial atoms creates a

20
Figure 2.6: The creep diagram defining the conditions of temperature and stress which
produce the four principle types of creep. The temperature is plotted as the homologous
temperature and the stress is normalized by the shear modulus, a = applied resolved
shear stress, it = shear modulus, and c = critical resolved shear stress of a well
annealed crystal [Cah83],
small deformation of the material that is not permanent as the interstitial atoms will
randomly redistribute once the applied force is removed. Since the materials examined in
this study do not contain a significant percentage of interstitial atoms, this creep mecha¬
nism is of little concern.
Herring-Nabarro-Coble creep [Her50, Cob63] produces a deformation in the
material by a net mass diffusion. In the Herring-Nabarro model, a net flux of vacancies
move away from the axis of the applied stress produces a net flux of atoms to the axis of
applied tensile stress, thus elongating the sample in the direction of the applied force.
This mechanism of creep requires a high concentration of vacancies with high mobility,
and only occurs at TH > 0.9 [Cah83]. Coble expanded this idea of bulk diffusion to
include the diffusion of atoms along grain boundaries. Grain boundary diffusion has a
lower activation energy than bulk diffusion because a grain boundary is an array of

21
dislocations. This lower activation energy for diffusion slightly reduces the temperature
at which the creep rate is significant. Coble creep still requires a TH > 0.85 [Cah83] to
become active and is therefore not of concern in this study.
The third type of creep is low temperature or logarithmic creep. In this mecha¬
nism, deformation occurs through dislocation multiplication, glide and climb. However,
the sources for dislocation multiplication become exhausted. Therefore the creep rate
starts at some initial higher valve and logarithmically approaches zero as the number of
mobile dislocations approaches zero [Cah83], The rate of logarithmic creep is not a
function of the applied stress, since the rate limiting steps are dislocation climb or multi¬
plication which are driven by random thermal excitations. This type of creep is typically
observed at low temperatures (TH < 0.5).
The fourth creep mechanism shown in figure 2.6 is high temperature or Andrade
creep [Cah83], In this mechanism dislocation glide, climb and multiply in response to
the applied force with the help of thermal excitation to overcome the higher activation
energies. There are three stages to Andrade creep. In stage one, a stress applied to the
sample at a given temperature causes an initial high strain rate that immediately begins to
decline. The reason for the declining strain rate is that the sample begins to work harden.
After a finite strain, equilibrium is established between the rate of work hardening and the
rate of dynamic annealing. Annealing is a heat treatment that eliminates the effects of
cold work, in dynamic annealing this processes is occurring at the deformation tempera¬
ture as described below. Second stage creep is a steady state process that produces a
constant strain rate. In the third stage of Andrade creep the sample begins to neck, grain
boundary sliding occurs [Mey84] and voids form inside the sample. This reduces the area
over which the force is carried thus increasing the stress which increases the strain rate,
ultimately resulting in failure. This type of creep is typically observed for TH > 0.6.

Deformation Mechanism Maps íDMMsl
22
In the proceeding sections, several mechanisms have been discussed which lead to
plastic deformation of metals. H.J. Frost [Fro82] has created deformation mechanism
maps (DMM) relating stress, temperature and deformation mechanism. All DMMs are
based on numerical models describing the different deformation mechanisms. This
model is then used to generate a two dimensional contour plots (temperature and stress
axis) that shows which deformation mechanisms are active. These plots are useful in
selecting materials for a given application. If the stress and temperature levels for a given
application are known, then the deformation mechanisms expected for a given application
can be projected. The primary limitation to the use of DMMs is that detailed data re¬
quired by the model are not available for all materials and conditions.
DMMs have been used to explain deformation and stress relaxation in thin film
metallization [K0I86, Fro92, Tho93, She96], These models have been able to accurately
predict plastic deformation in thin films at high strain rates, but none of these DMMs
were able to accurately predict the rate of isothermal stress relaxation. However, only
bulk material properties are currently available to use in DMMs.
Annealing
In the preceding sections the mechanisms responsible for plastic deformation at
high and low strain rates were discussed. However, since stress relaxation is of concern
in this study, other mechanisms that transform metals must be considered, such as anneal¬
ing.
Annealing is the process by which the stored energy in cold worked metal is
released. When a metal is cold worked, that is plastically deformed below the tempera¬
ture at which dislocation climb becomes significant, part of the deformation energy is
stored in the metal as defects and lattice distortion [Ree92]. Hundreds of Joules per mole
can be stored in a metal that has been heavily cold worked. The annealing process re-

23
Figure 2.7: Distribution of dislocations in a bent crystal (a) before polygonization, (b)
after polygonization and subgrain coalescence.
leases this stored energy by reducing the defects, lattice distortion and residual stress. In
the case of a thin metal film deposited on a substrate, the shear stresses between the film
and the substrate may also distort the lattice and induce defects. The process of annealing
should result therefore in stress relaxation.
There are four stages in annealing: recovery, recrystallization, grain growth, and
secondary grain growth [Ree92]. For this study the first three stages are of most interest
since they dominate the release of stored energy from cold work.
In recovery there is little change in the mechanical properties of metal, but the
electrical resistance of the metal decreases. This reduced electrical resistance indicates
that random dislocation tangles have begun to polygonize [Cah49] (see Figure 2.7). The
polygonization of the dislocation tangle into subgrain boundaries reduces the free energy
of the system. Ordering also reduces the number of electron scattering centers, resulting
in a lower electrical resistivity. No dislocations are destroyed, therefore the yield stress
does not change.

24
Polygonization and subgrain coalescence also lead to the second stage of anneal¬
ing, recrystallization. In recrystallization new unstressed crystals nucleate and grow to
replace the strained, disordered cold worked grains. One theory on the formation/nucle-
ation of these new unstrained crystals is that subgrains coalesce to form high angle grain
boundaries leaving an unstrained crystal in their wake [Cah83]. The difference in ener¬
gies provides the driving force for the high angle grain boundary to move outward,
consuming the strained crystals until they are all consumed. This is the growth stage of
the annealing process and it is at this point that the energy of cold work has been removed
and therefore stress has been relaxed.
Grain growth also affects the size and distribution of precipitates within a sample
[Bro82], Precipitates strain the matrix of the material in which they exist. There is
therefore a driving force for precipitates to coalesce and reduce the strain in the matrix.
In age hardened alloys this is called “over aging” as it reduces the yield strength of the
material. This process would also produce a relaxation of stress.
Strengthening of Bulk Metals
In the preceding sections’, mechanisms responsible for stress relaxation in thin
metal films have been reviewed. We will now review the methods used to strengthen
bulk metals to determine which could be used to strengthen thin metal films in bimetallic
actuators. There is limited information on strengthening of thin films.
The key to strengthening any non-brittle metal is to inhibit dislocation glide. This
can be done in several ways. First, the matrix of the metal can be affected to increase the
force needed to move a dislocation in a slip system. This can be done a number of ways,
including straining the matrix with cold work or solid solutions. Second, hard particles or
phases can be placed in the matrix of a material that will physically block the movement
of dislocations. There are five commonly identified strengthening mechanisms that use

25
one or both of these general approaches. Also, many commercial alloys use two or more
alloying elements and several of the five strengthening mechanisms to ensure that several
means of inhibiting dislocation motion are used. The five strengthening methods used in
bulk alloys and the means by which they inhibit dislocation motion are described below.
The first strengthening method is solid solution strengthening, in which soluble
alloying elements are added to the host metal [Cah83], Since most alloying elements
have an atomic size different from the host element, the matrix of the metal is strained,
and inhibits dislocation glide. The more the host matrix is strained, the harder it is for
dislocations to move. However, alloying elements generally have limited solubility in
host metals which limits the strengthening that can be achieved. For most of the com¬
monly used engineering metals, the solubility and strengthening effects of many alloying
elements have been determined experimentally. The Metals Handbooks Volumes 1 and 2
[Bak 97] contain most of this information. For more unusual materials, the Hume-
Rothery [Ree92] rules for substitutional solid solution strengthening are helpful.
Solid solution strengthening has a number of properties that make it unique
compared with the other strengthening mechanisms. First, the solute atoms are in equi¬
librium with the host matrix atoms. Thus there are no driving forces to change the atom
distribution in the metal and solid solution strengthened alloys are expected to be more
stable at elevated temperatures than alloys using some of the other strengthening tech¬
niques. This is of particular importance for this study since materials will be used at
elevated temperatures. Secondly, a solid solution strengthened material is expected to be
homogenous. Thus there is limited possibility that electrolytic half cells would cause
corrosion.
The second strengthening mechanism is to reduce grain size. Dislocations gener¬
ally pileup at the grain boundaries as they cannot cross them. The strength of a material
has been found to be proportional to the inverse of the square root of the average grain

26
size [Mey84, Ree92]. This relationship is called the Hall-Petch relationship. A potential
limitation to using this strengthening technique is that grains coarsen at elevated tempera¬
tures (Th > 0.6), and this generally reduces the strength of the material.
The third strengthening techniques is work hardening. When a metal is cold
worked (i.e. worked below a temperature at which annealing occurs) the dislocation
density increases greatly. With an increased density, dislocation will be pinned by other
dislocations, which limits dislocation slip. Work hardening can produce seven fold
increase in yield strength for pure aluminum. Fully annealed aluminum has a yield
strength of 15-20 MPa that increases to 50-60 MPa for 40% cold work and 100-120 MPa
for 90% cold work [Bak97], A concern with work hardening is that the material will
soften as dislocations climb at elevated temperatures during recrystallization.
The fourth way to strengthen some materials are by a martensitic transformation
[Kra90], This type of transformation occurs only in a few alloys and ceramics. It is
widely known and used in ferrous alloys, but does occur in some copper, aluminum and
titanium alloys. The martensitic transformation is a diffusionless non-equilibrium pro¬
cess that occurs by a coordinated shear displacement of atoms. This transformation
stresses the matrix of the material which inhibits dislocation motion. The use of marten¬
sitic transformations to strengthen alloys is limited because there are few non-ferrous
alloys which exhibit them [Pet70], In addition, a martensitic transformation leads to a
non-equilibrium state so there is always a driving force to transform to a lower strength
phase.
The fifth way to strength a metal is by precipitation or multi-phase hardening.
Alloying elements are added above their solubility limit in a host and precipitates form
out of the solid solution to block dislocation motion [Bro82]. In certain cases, a fine
ceramic powder is added to a molten metal in place of an alloying element to provide the
particles in a modification called dispersion hardening. This leads to improved thermal
stability of the alloy. For the greatest precipitation hardening, a large number of small

27
precipitates should be uniformly distributed throughout the host matrix. Thermal stability
is a concern at elevated temperatures since precipitates may coarsen (over age) resulting
in a weakened material.
The presence of a second phase can also strengthen a host matrix by straining it.
A second phase with a coherent interface to the host matrix will increase the strain in the
matrix. A coherent interface does not an array of dislocation between the precipitate and
the matrix [Bro90]. Thus, in addition to blocking dislocation motion, a coherent precipi¬
tate strains the matrix, increasing the force needed to move a dislocation. Thermal
stability is a concern, as precipitates may grow by thermally activated diffusion and the
coherent interface may be lost. The loss of the coherent interface substantially reduces
the strength of the alloy.
Ordered Intermetallic Compounds
Ordered intermetallic compounds [Cah85] are of interest in this study because
they are noted for high strength at both room and elevated temperatures with good resis¬
tance to creep and oxidation [Wes95]. There are several reports on the use of intermetal-
lics at high temperatures in jet engines [Lim97]. The basic structure of intermetallics was
discussed above in the section on thermal expansion. The mechanical properties are
discussed below.
Intermetallics are a class of materials with structures and properties between pure
metals and ceramics [Pop87]. The general composition on an intermetallic is A Bt where
A and B are metallic elements and x and y are normally small integers. This more com¬
plex composition creates a more complex crystallographic structure which in turn make
dislocation movement more difficult. There are several reasons for this. First, the num¬
ber of atomic distances a dislocation has to move to reach an equivalent crystallographic
site is much greater. In a simple AB intermetallic a dislocation has to move two atomic
distances as compared with one atomic distance for a non-ordered metal. In many

28
instances a dislocation will break into pairs of partial dislocations which only move one
atomic distance forming an anti-phase boundary [Bro82]. The creation of an anti-phase
boundary requires energy and increases the critical resolved shear stress for dislocation
glide. In a normal metal a dislocation must break A-A bonds and then reform them to
move. In an intermetallic a dislocation must break the stronger A-B bond to move. Once
the A-B bond is broken the weaker A-A and B-B bonds form (the anti-phase boundary),
which then have to be broken to form the new A-B bond. This process requires more
energy, and thus intermetallics have high yield strengths.
The more complex structure also inhibits creep. The self diffusion rate in inter¬
metallics is unusually low [Pop87] due to the large distance between equivalent site, and
a high activation energy for this process. As the creep rate for a material is proportional
to the rate of self diffusion [Cah83], intermetallics have a low creep rate.
Some intermetallics are also resistant to oxidation. Oxidation resistant intermetal¬
lics form an adherent passivating oxide such as A1203 on compounds of aluminum,
nickel and titanium.
A concern in using intermetallics in bimetallic actuators is their sensitivity to
composition fluctuations. In AINi, as little as 0.5% change in the composition of the
intermetallic can cause dramatic changes in the strength of the metal [Lim96], Therefore,
the compositional stability of any intermetallic compound should be tested before selec¬
tion.
Thin Films
Thin films have a number of unique properties versus their bulk counterparts.
Sinse the materials examined in this study are use as thin films in bimetallic actuators, an
understanding of their properties is needed. There are three primary reasons for the
unique properties of thin films. First, thin films are deposited onto a substrate, resulting

29
in substrate-thin film interactions. Second, the surface area-to-volume ratio for a thin
film is orders of magnitude higher than for a bulk material, so surface energies and
kinetics play a much greater role. Last, the techniques used to produce thin films are very
much different from the techniques used in the processing of bulk metals, which affects
their microstructure.
There are at least three common categories of thin film deposition processes:
chemical vapor deposition (CVD), physical vapor deposition (PVD), and electroplating
[Ohr92], In this study we have only used physical vapor deposition methods consisting
of sputtering or thermal evaporation.
In all physical vapor deposition (PVD) processes, a vapor of the source material is
created and condensed on solid surfaces, including the substrate of interest. All PVD is
done in a high vacuum chamber so that the material vapor can travel from the source to
the substrate without reacting with atmospheric gases. In thermal evaporation, the source
material is simply heated until it vaporizes, a very simple and clean process useful to
deposit single element films. The vapor pressure of different material varies so greatly
that the composition of an alloy can not be maintained with thermal evaporation [Ohr92].
In the sputter deposition processes, an inert gas, normally argon, is ionized and ions are
accelerated into the source material or target. When the inert gas ion strikes the target
atoms, its momentum is transferred and by development of a momentum cascade, some
of the target atoms are ejected from the surface. The sputter process is based on momen¬
tum transfer, therefore the vapor pressure of the target material does not control the rate
at which atoms are ejected from the target surface. Sputtering can therefore be used to
deposit metal alloys containing a number of elements [Ohr92], However, differences in
composition may be seen between the deposited film and source material [Zhe97],
After being sputtered or sublimed, the vaporized atoms condense and agglomerate
to form a film [Sor95]. The morphology of the developing film is primarily determined
by the mobility of the condensing atoms [Mac95]. For thermal evaporation, the condens-

30
SUBSTRATE
TEMPERATURE (T/Tm)
Figure 2.8: Diagram of Movchen and Demishin’s model of thin film morphology
[Kra90],
ing atoms have little kinetic energy and the surface mobility is primarily controlled by the
substrate temperature. Based on this, Movchen and Demishin [Kra90, Thor74] devel¬
oped a model that predicts film morphology based on the homologous temperature of the
substrate (Figure 2.8). In the Movchen and Demishin model, the film morphology is
divided into three zones. In zone 1 (TH < 0.3) the adatom mobility is so low that tapered
crystals form with voids between the crystals. In zone 2 (0.3 < TH < 0.5) the adatom
mobility is great enough that most of the voids fill in and columnar grains form. In zone
3 (T > 0.5) the mobility is high enough that nearly equiaxed crystals form.
In the sputtering process, the kinetic energy of the condensing atoms can be
controlled by the gas pressure and by biasing the substrate [Win92], The effects of
substrate temperature and gas pressure have been studied by J.A. Thornton [Tho74], who
developed a model to predict the morphology of sputtered deposited films (see Figure
2.9). Thornton did not divide the sputtered morphology into zones as Movchen and
Demishin did. There is some correlation between these two models as Krauss illustrated
in figure 2.9 [Kra90]. It should also be noted that post deposition annealing of a thin film
may cause the microstructure to develop similar to that which would have developed had
the film been grown at that same temperature [Mac95, Kno91].

31
ARGON
PRESSURE
(MICRONS)
SUBSTRATE
TEMPERATURE
(TITml
Figure 2.9: Thornton’s model of sputtered thin film morphology with a comparison
to Movchen and Demishin’s model of thin film morphology [Kra90].
During the deposition process and post deposition annealing, thin films can
develop a crystallographic texture [Mac95, Kno95]. There are two competing processes
responsible for the development of the thin film texture. In the growth of a columnar
microstructure, even from a melt, the crystallographic planes with the fastest growth
cause the elongated axis of the column [Bro94], In thin films there is also a large driving
force to reduce surface energy. So the crystallographic planes that have the lowest sur¬
face energy have a thermodynamic advantage over faster growing planes with higher
surface energies. The final texture that develops in a thin film is the balance of these two
factors.
There are two types of stresses that develop in thin films: intrinsic and extrinsic.
Intrinsic stresses are inherent to the deposition process and conditions, while extrinsic
stresses are due to the difference in the thermal expansion of the film and substrate.
Extrinsic stresses cause actuation of the bimetallic element.
Intrinsic stresses develop during the deposition process due to limited adatom
mobility [Mac95]. This limited mobility causes the formation of voids and vacancies in
the film. Voids tend to collapse and form grain boundaries. Vacancies migrate to grain

32
boundaries and are destroyed. This causes the film to attempt to contract, but instead it
may develop a state of biaxial tensile stress because it is attached to a rigid substrate. In
sputter deposition it is possible to increase the kinetic energy of the sputtered atoms, by
reducing the sputter gas pressure or biasing the substrate. Under bias, ions may penetrate
the first few layers of the deposited film. This “atomic peening” can reduce the intrinsic
tensile stress and even produce a compressive stress [Mac95]. However, atomic peening
injects a large number of trapped vacancies and implants gas [Mac95]. If the film is
heated to a temperature where vacancies or gasses are mobile, they will migrate to the
grain boundaries, be destroyed or evolved, and change the stress back to tension [Tow87].
In MEMS it is often desirable to have a residual stress near zero at room tempera¬
ture. This allows for greater flexibility of design. While it is possible to achieve low
stress by adjusting the sputter parameters, it is sometimes possible to reduce tensile stress
by thermal cycling. If the thin film and substrate are thermally cycled to -196 “C (liquid
nitrogen temperature) the metal film will normally contract more than the substrate. If
the film is ductile, it will plastically deform and upon returning to room temperature, the
film will experience a lower tensile stress or even a compressive stress [Bal94], Unfortu¬
nately, elevated temperature cycling of the film will counteract this compressive shift in
stress and re-introduce a tensile stress [Bal94],
A primary concern of this study is stress relaxation in the thin film, defined above
to be a time dependent change in the state of intrinsic plus extrinsic stress. Stress relax¬
ation changes the force or displacement of a bimetallic actuator, and may be caused by
plastic deformation, creep, or diffusional processes including precipitation. The process
of plastic deformation and dislocation movement in thin films is restrictive as compared
with bulk metals. There are several reasons; first, a thin film is under a biaxial state of
stress. Thus dislocations can only glide in slip systems that have a component in the z
direction, normal to the film surface, since only these planes have a non-zero resolved
shear stress [Nix89], If the thin film is textured there will even fewer slip systems with a

33
O 50 100 150 200 250 300 350
Temperature, C
Figure 2.10: Stress temperature profile for pure sputtered aluminum and
microelectronic alloy Al-Si-Cu [Jer95].
resolved shear stress greater than the critical value. Second, a dislocation moving along a
slip system must nucleate misfit dislocations at the film/substrate interface, which re¬
quires additional energy. If a passivating oxide forms on the surface of the film, as in the
case of aluminum, misfit dislocations must be created at both interfaces [Nix89], further
reducing dislocation glide.
The primary method used to test the mechanical properties of thin films is mea¬
surement of stress as a function of temperature. As the thin film expands at a different
rate than the substrate, changing the temperature changes the stress in the film. This is an
easy and non-destructive way to test a thin film, but it is limited because the stress and
temperature cannot be varied independently. Instruments that measure stress in this way
include the Tencor Flexus which was used in this study (see description in Chapter 3).

34
The result of this type of testing is a plot of stress as a function of temperature, shown in
Figure 2,10, for sputtered pure A1 and Al-1.5%Si-2%Cu. To evaluate stress relaxation
the sample can be heated to a temperature and the stress plotted as a function of time.
For bimetallic actuators, a film that shows no hysteresis in stress versus tempera¬
ture over the range used is desired. Hysteresis in this plot indicates that the film is plasti¬
cally deforming. Figure 2.10 shows that the metals currently being used to manufacture
bimetallic actuators [Jer96], both exhibit significant hysteresis over the temperature range
from 20"C to 300-350°C. The pure aluminum shows little strength over 120°C, while the
Al-Si-Cu alloy is only slightly stronger and plastically deforms above 170°C under a
compressive stress of 50 MPa. The room temperature residual tensile stress for pure
aluminum is 150 MPa, while the Al-Si-Cu is 200 MPa. Isothermal stress relaxation tests
were also performed on pure aluminum and Al-Si-Cu alloy thin films [Jer96], with both
films showing a 15% reduction in stress over 1000 minutes. At 50°C the pure aluminum
film was initially under a tensile stress of 75 MPa, while the Al-Si-Cu alloy film was
initially under a higher tensile stress of 150 MPa.
There have been only a few studies to examine isothermal, long term stress
relaxation in thin films. The first was performed in 1985 by Hershkovitz, Blech and
Komem. They identified three modes of stress relaxation in thin aluminum films. The
first mode was identified as dislocation glide. The mechanism responsible for the other
two modes were not determined. In 1991, Drapper and Hill studied stress relaxation in
Al-Si-Cu thin films. They concluded that logarithmic creep was the dominant mecha¬
nism responsible for stress relaxation. However, Drapper and Hill assumed a single
mechanism was responsible for the stress relaxation and fit there data with a single
exponential function, ignoring the conclusions of Hershkovitz, Blech and Komem. The
single figure of the data and curve fit in their article showed great inaccuracies between
the data and the fit. In 1995 Witvrouw, Proost, Beweerdt, Roussel, and Imec also re-

35
ported isothermal stress relaxation data from highly tensile stressed Al-Si-Cu thin films.
At low temperature (70°C) they attribute stress relaxation to dislocation glide. At higher
temperatures (120°C to 140°C) they propose that the dislocations are cutting the Al2Cu
precipitates. From these limited studies, no complete model has been developed to
explain isothermal stress relaxation in thin films.
A new mechanism proposed by J. Tersoff for the stress relaxation in epitaxial
films is surface roughing [Ter94], In this new mechanism the surface of the epitaxial
layer becomes rough, allowing easy nucleation of dislocations. This mechanism has not
yet been tested in polycrystalline thin films.
While strengthening of thin films to reduce stress relaxation and plastic deforma¬
tion has not been extensively studied, work to reduce failures caused by thermal cycling
has been reported. Much of this work was done to improve the reliability of Josephson
superconducting devices (SQUID - Superconducting Quantum Interference Device)
[Kir80], which were of interest for building ultrahigh-speed computers. These devices
operate at temperatures below 10°K (-263°C), and need to be able to withstand tempera¬
ture cycles form 20°C to -263°C. Basavaiah and Greiner found the addition of gold to the
lead indium alloy reduced failures, but the mechanisms responsible for the reduction in
failures was not determined [Bas77].
Enviromental Stability of Materials
Bimetallic actuators may operate in atmospheric gases at elevated temperatures,
and the materials of construction must be stable. Environmental instabilities of these
films would compromise their physical integrity and cause the device to fail. There are
two environment reactions that need to be considered for this application: oxidation and
corrosion.
Most metals oxidize, and some self-passivate while others require some protective
coatings. This study has focused on metals which self passivate, such as aluminum,

36
copper, titanium, nickel. For a passivating oxide to be effective it should have the follow¬
ing four properties: high thermodynamic and kinetic stability, slow growth rate, adherence
to the metal, and easily form or re-form [Wes95]. The growth rate is of greater impor¬
tance in thin films as an oxide could consume the entire film. Most passivating oxides
generally follow a parabolic growth rate equation given by
x2 = kt (4)
where x is the thickness of the oxide, kp is the parabolic-growth-rate constant, and t is
time. Aluminum typically exhibits an inverse logarithmic dependence at low tempera¬
tures and parabolic at high temperatures. Thus for thin film applications, materials with a
low parabolic-growth rate constant are needed, such as aluminum and silicon, for el¬
evated temperature applications. In addition, the formation of an oxide and or a metal-
oxygen solid solution will introduce some stress into the metal substrate. This change in
stress could affect the force or displacement of a bimetallic strip. This change may be
time dependent as can be seem from equation 4.
Corrosion is also a concern because these devices may be operated in moist air
where the possibility of a galvanic cell exists. One area where a cell could exist is be¬
tween the active and passive layers of the bimetallic strip. In this application, there is
normally an oxide layer between these two layers of the bimetallic strip that electrically
isolates them and so a galvanic cell cannot be formed. Disruption of this oxide would
cause this to be a concern. Also, formation of micro-galvanic cells in a two phase alloy,
should be tested [Ask89].

CHAPTER 3
EXPERIMENTAL PROCEDURE
Introduction
Eleven different materials were deposited as thin films and analyzed in this study
(see Table 3.1). The deposition methods and procedures will be covered in this sections.
The films were first tested to determine their mechanical properties and the best materials
were analyzed further to determine their microstructural and chemical compositions. A
number of analysis techniques were used as discussed below.
Table 3.1: Materials examined in this study.
Material
Deposited Thickness
pm
Composition or Purity, (wt. %)
T201 Aluminum
1.3
A1 - 4.6Cu - 0.57Ag - 0.36Mn -
0.2Mg - 0.27TÍ
5052 Aluminum
1.3
A1 - 2.5Mg - 0.25Cr
2090 Aluminum
1.3
A1 - 2.57Cu- 2.1U - 0.12Zr
Nickel
0.8
99.98
Titanium
0.55
99.99
Manganese
0.4
99.9
Copper
0.9
99.9
A13Ti
0.4
Al-99.99, Ti-99.99
Cu,Au
1.3
Cu-99.99 Au-99.99
CuAu
1.3
Cu-99.99 Au-99.99
CuAu,
1.3
Cu-99.99 Au-99.99
37

Thin Film Deposition
38
The thin films examined in this study were deposited by either sputtering or
electron beam evaporation. All the films were deposited on 100mm diameter, [100]
oriented, single crystal silicon wafers. The thickness of each film is shown in table 3.1.
The wafers used for the aluminum alloy and copper gold intermetallics had a 100 nm
thermal oxide grown on them before film deposition. All wafers were cleaned using the
following procedure: 5 minute in an ultrasonic bath for each solvent with a de-ionized
water rinse between solvent for, trichroloethane, acetone, and methanol, then five minutes
in a solution of 75% sulfuric acid and 25% hydrogen peroxide-30% followed by a de¬
ionized water rinse and blown dry with dry nitrogen.
The aluminum alloy and copper gold intermetallic films were sputter deposited by
Sputtered Thin Films, Inc., Santa Clara, CA, using an 8-inch DC magnetron sputter gun
running at 5kW of power. The chamber was first pumped down to 2.8xl0'7 Torr, then
back filled with argon to a pressure of 8 mTorr. The substrates were heated to 100°C.
The aluminum alloy sputter targets were from bulk sheet stock. Copper gold intermetal¬
lics were produced by depositing four layers of pure copper and gold in the proper pro¬
portion to produce the required composition. Table 3.2 shows the thickness of the indi¬
vidual copper and gold films that were deposited for each of the average compositions.
Table 3.2: Layer thickness deposited for the copper/gold intermetallic films.
Compound
Copper Layer
Thickness, |im
Gold Layer
Thickness,|im
Cu,Au
0.1132
0.4868
CuAu
0.2465
0.3533
CuAu,
0.4059
0.1941

39
The copper, titanium, nickel, manganese and aluminum-titanium intermetallic
films were deposited by electron beam evaporation. The evaporation system contained
three Telemark Model 211 single pocket electron-guns in the vacuum chamber. These
guns are powered and controlled by a Sloan PAK 12 series 12 kW power supply operated
at 10 kV. The deposition rate and total deposited thickness were measured with quartz
crystal monitors, either an Inficon XTC or Sycon model 100. The deposition rate was
manually controlled and varied between 10 to 15 angstroms per second. The vacuum
chamber was pumped down to 5xl0'7Torr before each deposition. Due to outgassing, the
chamber pressure would rise to 5x 1 O'6 Torr during deposition. The substrates were heated
to 150°C +/- 10°C with resistive heaters, manually controlled with a variable transformer.
Copper Gold Heat Treatment
The copper gold compounds required a special heat treatment to homogenize the
layers and to develop the long range ordered intermetallic structure [Fed58]. These heat
treatments were performed in a vacuum of 5X10'6 Torr or better to inhibit oxidation of the
copper. The heat treatments were 12 hours at 500°C to homogenize the film, and 12
hours at 275“C followed by 12 hours at 225°C to develop the long range order. All
temperatures changes were ramped at 1°C per minute using an Omega CN3000 controller.
Stress and Stress Relaxation Measurements
Stress and stress relaxation in the samples were measured with a Tencor Flexus
model 2320. The Flexus 2320 uses an optical lever to measure the curvature of the
sample [Tur96]. Stoney’s equation [Sto09], Equation 4, was used to calculate the average
biaxial stress, a, in the film:
Eh2
(1 —1>) ■ 6 ■ R -1
a =
(4)

40
where R is the curvature of the substrate, t is the film thickness, h is the substrate thick¬
ness, E is the Young’s modulus for the substrate, and v is Poisson’s ratio for the substrate.
The silicon substrate’s thickness and curvature were measured before the film was
deposited. The curvature of the substrate was then measured after the film was deposited,
and the change in curvature is used in Stoney’s equation to calculate the film stress. In
addition to measuring stress at room temperature, the Flexus could heat a sample to
500°C or cool it down to -60°C under computer control and measure curvature. The
Flexus could also temperature cycle a sample, taking measurements during the cycle to
generate a stress versus temperature plot (see Figure 2.10).
The isothermal stress relaxation in aluminum alloy thin films was also measured
with the Flexus. To obtain a similar starting condition for different tests, each sample
was heated to 350°C at 1 “C/minute, held at 350°C for 30 minutes, and cooled back down
to room temperature at l°C/minute. The sample was then taken to the test temperature, at
l°C/minute, and held at the test temperature for forty eight hours. Every 15 minutes the
stress in the sample was measured, allowing stress to be plotted versus time (see Figure
3.1). Once the mechanical properties of a film had been measured with the Flexus, the
samples was divided for chemical and microstructural evaluation.
Powder X-Rav Diffraction ( XRDl
Powder x-ray diffraction was used to determine the long range ordering in the
copper gold films [War69] and to evaluate crystallite size in the aluminum alloy films
[Cul78]. The system used was a Philips APD 3720 x-ray diffraction systems controlled
with an IBM-PC type computer running Microsoft windows 3.11 with Philips PW1877
Automated Powder Diffraction software, version 3.6g. X-rays from a copper anode
operated at 40 kV and 2 milliamps with a nickel filter were used.

41
Figure 3.1: Isothermal stress relaxation in 5052 aluminum alloy sample at 125°C.
X-ray diffraction is effective for detecting the change in structure that occurs in an
order-disorder transformation. For the copper gold system, the disordered crystallo¬
graphic structure is FCC, with atoms sitting randomly on lattice sites. In an ordered
phase, one type of atom, either copper or gold, sits in the corner sites and the other atom
will sit in the face sites. This creates two inter-linked simple cubic or simple tetragonal
crystallographic structure. This change in the crystal structure causes the original diffrac¬
tion peaks to change, and the different structure factor causes a number of new peaks to
appear [Cul78], With a true randomly oriented sample, it is possible to calculate the
degree of ordering in the sample by comparing different peak heights [War69], However,
thin films have a crystallographic texture that complicates this procedure. As a result it
was difficult to quantify the degree of order in the films.
For aluminum alloys the dislocation structure inside the aluminum crystals, the
grain size and the strain were all evaluated using x-ray diffraction line shape analysis. In

42
this technique, the (111) and (220) diffraction peaks were selected for evaluation. The
(111) peak was scanned from 20 = 38° to 39.2°, and the (220) peak was scanned from 26
= 64.24° to 66° using 0.02° increments and 10 second dwell time. For line shape analysis,
an unstressed quartz standard was scanned in the same diffractometer under the same
conditions. The quartz standard was scanned from 20 = 38.7° to 39.8°, and from 20 =
63.5° to 64.7°.
Line shape analysis extracts the factors that cause the x-ray diffraction peaks to
shift and/or broaden. There are three main causes of peak broadening in x-ray diffraction.
The first is the diffractometer itself. Instrumental broadening is eliminated by compari¬
son to the quartz standard. The other two causes of peak broadening are strain and
crystallite size. Each of these produce a different shape. Strain effects produce a
Gaussian distribution while the crystallite size produce a Cauchy distribution [War69].
The Phillips diffraction software can deconvolute these effects and determine the strain
and crystallite size.
Atomic Force Microscopy (AFM1
A Digital Instruments Nanoscope III atomic force microscope (AFM) was used to
determine the surface topology of the aluminum alloy samples [Hud92], A silicon tip in
the tapping mode was used to probe the surface of the samples. Digital Instruments
software was used process the scanned images and calculate the RMS surface roughness.
Scanning Electron Microscopy ÍSEM1
A JEOL 6400 scanning electron microscope was also used to image the surface of
samples. In addition to the secondary and back scattered electron detectors, this system
has an energy dispersive spectrometer (EDS) that allowed qualitative elemental analysis
of surface features [God92]. Samples were first viewed in the secondary electron mode
to evaluate the surface topology. Very bright areas were viewed in the back scattered

43
electron mode. The trae secondary electron yield from an area is more strongly depen¬
dent on the topology, whereas the back scattered electron intensity is dominated by the
atomic mass of the excited atoms. Areas that appeared to have different atomic composi¬
tions were examined with EDS to determine the concentration of elements.
Electron-Probe Microanalvsis ÍHPMA1
EPMA was used to quantitatively determine the compositions of the aluminum
alloys and the copper gold intermetallic films [Gol92]. A JEOL 733 Superprobe with
four wavelength despersive spectrometers was used. For the aluminum alloy films the
bulk aluminum alloy was used as the analysis standard to determine difference in compo¬
sition between the bulk and thin film composition. For the copper gold intermetallic
films, pure copper and gold standards were used with the ZAF (atomic number effects, X-
ray absorption affects, and X-ray fluoresces effects) analysis technique to determine the
composition of these films [Gol92],
Transmisión Electron Microscopy (TEMI
A JEOL 200CX analytical transmission electron microscope was used to examine
precipitates and the crystal size of the aluminum alloys. The samples were prepared by
first cutting a 3mm disk from a silicon substrate with a deposited film. These samples
were then thinned from the silicon side to between 150pm and 250pm. Using a dimpling
grinder, the back sides of the samples were hollowed out so that the center of the samples
were approximately 50pm thick. The thinned samples were encapsulated in paraffin and
a small opening (1 mm) was made in the paraffin over the dimple. The samples were
etched in a solution of 50% HF and 50% HNOv The etching process was monitored by
viewing the samples through a low power binocular microscope. Once the acid had
etched through the silicon and reached the aluminum film, etching was stopped. These
samples were ion milled until a small opening appeared in the aluminum, then viewed in
the TEM.

44
In the TEM, the samples were first viewed in the bright field mode to evaluate the
general structure of the sample [Lor94]. The dark field mode was used to better identify
different phases. Electron diffraction images and patterns were also taken to identify the
phases present by calculating the interplaner spacing. The inter plane spacing, dhkl was
calculated using:
XL = Rdhkl (5)
where L is the camera length (for the JEOL 200CX L = 82 cm), X is the electron
wavelength which is based on the accelerating voltage (A, = 0.025 lnm based on an accel¬
erating voltage of 200keV) and R is the radius of the diffraction pattern.
Etching
Grain size is an important parameter when evaluating the strength of a metal.
Etching is the primary means by which the grain size may be made visible in bulk metals,
but this technique was found to be ineffective in this study. Etching thin films to define
grains is problematic. Most etchants attach different crystallographic planes at different
rates. In a material with a randomly oriented crystals, this type of etchant works well.
However, thin films are strongly textured and tend to etch more uniformly, complicating
the determination of crystal size. Some etchants that attack grain boundaries preferen¬
tially [Smi67, Gif70] and two were tested for aluminum alloys: 1% hydrofluoric acid and
a 10% sodium hydroxide solution. Satisfactory definition of the grain boundaries was not
achieved.
Auger Electron Spectroscopy fAESt
The Perkin-Elmer PHI 660 Scanning Auger Microprobe was used to evaluate the
oxide thickness on the aluminum alloy films. Auger analysis is very surface sensitive,

45
containing information primarily from the top three or four atomic layers [H0I8O]. The
Auger system contained an argon sputter gun that enables depth profiles to be collected.
To do so, an Auger spectrum was taken, the sample sputtered for 15 seconds, and another
Auger spectrum taken. This process was repeated to the required depth. The aluminum
alloy samples were sputtered until the oxygen signal had decreased to 10% of its original
value. This technique gives a thickness in terms of sputter time. The conditions used
result in an estimated sputter rate of 250 angstroms per minute These data are primarily
used to compare differences in oxide thickness from sample to sample.
Curve Fitting
To curve fit the stress relaxation data, curve fitting techniques were evaluated:
polynomial, single exponential, and double exponential. The double exponential tech¬
nique fit the data using the equation:
cr(t)=A + Be"'1+Ce"2 (5)
where a is stress, t is time, and A, B, C, m, and m2 are constants that are determined from
the data. This technique fit most of the data with the least error.
The process for determining the constant was the linearization of the data. This
was sometimes complex as the two exponential functions needed to be separated to be
linearized. Separation is accomplished by first fitting the stress relaxation data for t > 380
minute. To linearize these data, the value of A in equation 5 was first manually selected
and subtracted from the remaining data. The natural log for the remaining data was then
taken and least squares linear fit to determine the slope, intercept and correlation factor.
The A constant was iterated to obtain the best possible correlation factor. With this step
completed, the first exponential was subtracted from the original data. The natural log of
the difference was then taken and the resulting data again fitted with a linear least square

46
fit to determine the slope and intercept, m2 and C.. All calculations were done in
MathCad Version 4.0.
Numerical Modeling
The effects of oxide thickness were found to be significant, therefore a numerical
model was developed to quantify this effect. This model was based on the equation
developed by Townsend [Tow87] and implemented in MathCad version 4.0. In this
analysis a three layer structure was modeled. The first layer was a 5|im thick layer of
silicon, and the second layer was 5|im of aluminum. The third layer of A1203 was varied
from 0 to lnm in thickness. This model was used to calculate the curvature of the struc¬
ture, therefore the length was not required. The output was plotted as a percentage
change in curvature for the structure with varying oxide thickness.

CHAPTER4
RESULTS
Introduction
The results of the testing and analysis will be presented in this chapter which is
organized by material to provide an understanding of their performance.
T201 Aluminum
Figure 4.1 shows the stress versus temperature plot for the T201 alloy film and
may be compared to the literature data shown in figure 2.10. These data were collected
after an initial temperature cycle to 350°C to stabilize the film’s microstructure after
deposition. While the intrinsic stress in these films were a function of deposition param¬
eters, the first cycle to 350°C annealed and consolidated the films and the extrinsic
stresses were so high that they dominated the stresses reported as in Figure 4.1. As can
be seen from these data, the room temperature residual tensile stress is almost 400 MPa
and the film begins to plastically deform at approximately 200°C under a compressive
stress of 50 MPa. This sample was also cooled to -196DC in an attempt to reduce its room
temperature residual stress, but this procedure had no effect on the room temperature
residual stress.
The isothermal stress relaxation seen in these samples at different temperatures
are shown in figures 4.2 through 4.6. These data were curve fit as describe above and the
equation is shown in each plot. Below 125°C the stress relaxation data was fit well by a
double exponential decaying function that suggesting that are two stress relaxation pro¬
cesses were occurring. At 125°C and 150°C the noise on the data are to large to allow
fitting. Rather than heating at a rate of l°C/minute samples were also heated rapidly
47

O 50 100 150 200 250 300 350
Temperature, C
Figure 4.1: Stress versus temperature plot of T201 aluminum thin film.
Figure 4.2: Isothermal stress relaxation of T201 aluminum thin film at 50°C.

Stess, MPa Stess, MPa
49
Figure 4.3: Isothermal stress relaxation of T201 aluminum thin film at 75°C.
Figure 4.4: Isothermal stress relaxation of T201 aluminum thin film at 100°C.

Stess, MPa Stess, MPa
50
Time, minutes
4.5: Isothermal stress relaxation of T201 aluminum thin film at 125°C.
Time, minutes
Figure 4.6: Isothermal stress relaxation of T201 aluminum thin film at 150°C.

51
(10°C/min) to the testing temperature to evaluate the effect of ramping rate on stress
relaxation. The stress relaxation was little changed, showing that ramping rate had little
affect on the stress relaxation seen in the samples. A compilation of the curve fitting
parameters is shown in table 4.1.
In addition to the mechanical evaluation of the T201 thin films, microstructure
and chemical composition was evaluated. Table 4.2 shows the results of the EPMA
analysis of the thin films and the composition of the bulk alloy, as determined by an
external lab. There are systematic differences in composition between the sputtered thin
film and the bulk alloy presumably due to the sputter deposition process [Zhe97].
Figure 4.7 is a SEM micrograph of a T201 thin film sample that was heat treated
for 60 hours at 100°C. Back scatter images of the same regions showed that the bright
areas had a different atomic composition than the areas around them. Figure 4.8 is the
EDS spectrum for the dark areas showing only aluminum. Figure 4.9 shows the EDS
spectrum for the bright areas and both copper and aluminum are detected while silicon is
not. Therefore, the bright areas on the SEM image are due to both topographic features
and differences in chemical composition of the sample, presumably due to precipitates.
Table 4.1: Residual stress after infinite relaxation time, o, change in stress at short (AcJ
and long (AcrJ times, and the time constants (x) for the two exponential equations
forT20l aluminum.
Temp.
(°C)
CT¡
(MPa)
Aa<
(MPa)
T<
(min)
Aa>
(MPa)
T>
(min)
50
330
9.2
196
49
1820
75
240
13.8
220
26
1940
lOO
160
608
220
29
1400
125
33
0
0
81
960
150
40
-3
200
0
0

Table 4.2: Chemical compsition of sputter deposited T201 aluminum alloy thin
films determined by EPMA, and compared with the bulk alloy composition.
52
Alloying
Element
Bulk Alloy
Weight Percent
Sputtered Thin Film
Weight Percent
A1
93.91
93.19+/-2
Ag
0.57
0.48 +/- 0.04
Cu
4.6
4.59 +/- 0.2
Ti
0.27
0.53 +/- 0.04
Mg
0.29
0.34+/-0.1
Mn
0.36
0.87+/-0.1
Figure4.7: SEM micrograph of T201 aluminum sample after 60 hours at 100°C.

53
Energy, keV
Figure 4.8: EDS spectra of dark area of SEM image shown in figure 4.7,
showing a lack of copper and silicon.
Energy, keV
Figure 4.9: EDS spectra for bright areas in SEM image, figure 4.7, consistent with
a precipitate rich in copper.
Atomic force microscopy (AFM) was used to evaluate the surface morphology of
the samples, with an image shown in Figure 4.10. Mounds or hillocks about 0.6pm high
were present on the sample surface. The AFM was also used to evaluate the surface
roughness of these films after different heat treatments. Due to the large mounds that
form on the surface, the RMS roughness varied greatly from position to position on the
same sample. An effort was made to scan areas unaffected by the mounds, but this
required very subjective selection and exclusions of areas of the sample. Thus, the RMS
roughness data were not reliable and are not reported.

54
Figure 4.10: AFM amplitude image of T201 aluminum thin film after stress relaxation at
125°C for 2 hours. The scanned area is 15 pm by 15 pm.
The primary strengthening techniques used in the T201 alloy are solid solution
strengthening (due to alloying elements of magnesium and manganese [Dav90, Bro82])
and precipitation hardening (due to alloying element copper). Silver is added to this alloy
(0.57 % weight percent) to stabilize the copper precipitates [Dav90], The effectiveness of
precipitation hardening is controlled by the size and distribution of the precipitates in the
host crystals [Cah83], TEM was used to evaluate the formation of precipitates in the
aluminum crystals. Figure 4.11 a shows a plan view TEM image of a T201 sample. The
small (10 nm to 30 nm) dark spots in this image are the CuAl, precipitates in the alumi¬
num matrix. Figure 4.1 lb shows the indexed electron diffraction pattern from this sample
with the aluminum diffraction rings and the CuA12 diffraction spots identified.
The bright field TEM images also give some indication as to the size of the
aluminum crystals. Caution must be used in evaluating grain size from this image be¬
cause the etching process used in the sample preparation would likely attack small crys-

55
Figure 4.11: TEM micrographs of T201 thin film sample, (a) bright field image showing
aluminum crystals and CuA12 precipitates; (b) indexed electron diffraction pattern show¬
ing the diffraction rings for the A1 (111) and (200) planes and for the A1Cu2(422) plane.

56
tais faster than large crystals. Also, a TEM image only examines a small area of the
sample, and a large area must be viewed to determine the true grain size distribution. The
apparent average crystal sizes was between 1 and 2 pm.
While direct measurement of the aluminum grain size by etching was unsuccess¬
ful, formation of the copper rich precipitates gives some indication of the aluminum grain
size. Precipitates tend to nucleate and grow at triple points where three crystals come
together [Ree92], While the direct calculation of grain size from the precipitation density
is questionable, the comparison of different samples and inferring changes in crystal size
based on changes in precipitation density is reasonable. Figure 4.12 shows a SEM micro¬
graph of a T201 sample that was heated to 350°C and then cooled to 25°C over 48 hours
to test the stability of the microstructure. The number of copper rich particles in figure
4.12 are of the same order as that of figure 4.7, i.e. 150 +/- 10 particles in the ~ 20 by 10
pm area of the micrograph. Therefore, the precipitate size and density of these samples
is apparently stable over several temperature cycles to 350°C.
Figure 4.12: SEM image of T201 aluminum sample after temperature
cycling from 350°C to 25°C over 48 hours.

57
0.0 0,5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
SPUTTER TIME, min.
Figure 4.13: Auger depth profile of T201 thin film. The film was sputtered
with 3 keV argon in a 3x3 mm raster.
The effect of oxidation on a bimetallic actuator’s performance is a concern. AES
analysis was used to measure the oxide thickness on a T201 thin film sample that had
been cycled to 350°C 5 times. Figure 4.13 shows the Auger depth profile. The rastered
ion gun used in this analysis sputtered A120, at an approximate rate of 25 nm per minute.
Thus the oxide thickness on this sample is approximately 25 to 30 nm.
Powder x-ray diffraction analysis was also performed on the T201 thin film
samples. A broad 20 scan was taken of a sample cycled to 350°C once to determine the
phases present in the sample. Only elemental aluminum and silicon from the substrate
were detected in these scans. XRD scans were then run on samples that were heat treated
at 50°C or 125°C for up to 32 hours. The samples were scanned for their (111) and (220)
peaks. Figures 4.14 and 4.15 show the diffraction peaks after different stress relaxation
times at 125°C. Note that there is a substantial shift in the peak position between the
initial scan and those that have undergone stress relaxation for the shortest time of two

58
Figure 4.14: Aluminum (111) x-ray diffraction peaks for T201 thin film samples
after various stress relaxation times at 125°C.
Figure 4.15: Aluminum (220) x-ray diffraction peaks for T201 thin film samples
after various stress relaxation times at 125°C.

59
hours. Line shape analysis was performed on the X-ray data to determine crystallite size
as a function of stress relaxation, but the crystallite sizes varied randomly and no correla¬
tion was found.
5052 Aluminum
Figure 4.16 shows the stress versus temperature plot for the 5052 aluminum alloy
thin film. These data were collected after an initial temperature cycle to 350°C to stabi¬
lize the film’s microstructure. As can be seen from these data, the room temperature
residual tensile stress in the film is 300 MPa versus 400 MPa for T201,200 MPa for Al-
Si-Cu, and 150 MPa for pure AL. The film begins to plastically deform under a compres¬
sive stress of 25 MPa at 200°C. A 5052 sample was also cooled to -196°C in liquid
nitrogen to reduce the room temperature residual stress. Upon warming to room tempera-
0 50 100 150 200 250 300 350
Temperature, C
Figure 4.16: Stress versus temperature plot for 5052 aluminum.

60
ture, the tensile stress was reduced from 300 MPa to 125 MPa, a 58% reduction. Starting
at the reduced room temperature stress of 125 MPa the film begins to plastically deform
at a lower temperature of 125°C under a compressive stress of 25 MPa. Upon cooling
from 350°C to room temperature, the tensile stress returned to 400 MPa.
The isothermal stress relaxation seen in these films at different temperatures is
shown in figures 4.17 through 4.21, along with the fitted curve and the equation. At 50°C
the stress relaxation was best fit with a single exponential curve. For temperatures
between 75°C and 125°C the stress relaxation data was fit well by a double exponential
curve. At 150°C the noise in the stress relaxation data is so large that the fitted curve is
unreliable. Samples were also heated at 10°C/minute to the testing temperature to evalu¬
ate the effect of ramping rate on stress relaxation. Again, this had little affect on stress
relaxation. The curve fitting parameters are compiled in table 4.3.
Figure 4.17: Isothermal stress relaxation in 5052 aluminum thin film at 50°C.

Stress, MPa Stress, MPa
61
Time, minutes
Figure 4.18: Isothermal stress relaxation in 5052 aluminum thin film at 75°C
Figure 4.19: Isothermal stress relaxation in 5052 aluminum thin film at 100°C
3000

Stress, MPa Stress, MPa
62
;igure 4.20: Isothermal stress relaxation in 5052 aluminum thin film at 125°C.
Figure 4.21: Isothermal stress relaxation in 5052 aluminum thin film at 150°C.

63
Table 4.3: Residual stress after infinite relaxation time, cr change in stress at short (Ac.)
and long (Acrj times, and the time constants (x) for the two exponential equations,
for 5052 aluminum.
Temp.
(°C)
CTi
(MPa)
Aa<
(MPa)
T,
(min)
Act>
(MPa)
(min)
50
240
0
0
30
2800
75
150
12
200
33
1500
100
110
19
170
29
1380
125
80
10
160
17
220
150
0
0
0
58
50000
Table 4.4: Chemical compsition of sputter deposited 5052 aluminum alloy determined by
EPMA, compared with the bulk alloy composition.
Alloy Element
Bulk Alloy
Weight Percent
Sputtered Film
Weight Percent
AI
97.25
98.9 +/- 0.5
Mg
2.5
1.73+/- 0.03
Cr
0.25
0.16+/- 0.06
The sample that was cooled to -196°C was tested for stress relaxation at 75°C. At
this temperature the tensile stress in the film was 50 MPa and there was no stress relax¬
ation observed over forty-eight hours of testing.
The EPMA chemical analysis of the 5052 samples is shown in Table 4.4. Again,
there are differences in composition between the sputtered thin film and the bulk alloy,
presumable due to sputter deposition effects [Zhe97],

64
Figure 4.22: SEM micrograph of 5052 aluminum alloy thin film sample that was
temperature cycled to 350°C at l°C/minuteand held at 350°C for 30 minutes
and then cooled to room temperature over 48 hours.
Figure 4.23: SEM micrograph of 5052 aluminum alloy thin film
that was held at 100°C for 60 hours.

65
Figure 4.24: AFM height image of 5052 aluminum alloy thin film without heat treat¬
ment. Scan area is 5 |im by 5 |im, Z-range 32nm and RMS roughness is 3.77 nm.
Figure 4.25: AFM height image of 5052 aluminum alloy thin film heat treated at 125°C
for 32 hours. Scan area is 5 (tm by 5 nm, Z-range 60nm and RMS roughness is 7.09 nm.

66
Table 4.5: RMS roughness and Z-range for 5052 aluminum alloy thin films that have
undergone different heat treatments.
50°C
125°C
Time at
Temperature
(Hours)
RMS
Roughness
(nm)
Z-Range
(nm)
RMS Roughness
(nm)
Z-Range
(nm)
0
3.77
32.4
3.77
32.4
2
5.9
47.5
5.9
60.2
4
4.6
35.7
5.6
55.2
8
5.8
51.1
5.6
55.5
16
6.3
49.4
4.6
58.9
32
5.6
51.7
4.7
43.6
Figure 4.22 and 4.23 show SEM micrographs of 5052 alloy thin films that have
undergone different heat treatments. EDS was used to examine the difference between
the light and dark areas seen in figure 4.23. No difference in composition between these
two areas was found.
The changes in AFM surface roughness with stress relaxation in samples that
were heat treated at 50°C and 125“C for up to 32 hours is shown in Table 4.5 and figures
4.24 and 4.25. The RMS roughness and Z-range (the height difference between the
highest and lowest point on the sample) are reported in Table 4.5. Figure 4.24 shows the
AFM height image of a 5052 alloy sample with no heat treatment and figure 4.25 shows
the AFM height image of a sample that had been held at 125°C for 32 hours. The size of
features in these images are nearly the same, but the RMS roughness and the Z-range
have doubled for the sample held a 125°C (Table 4.5). The data in Table 4.5 shows that
the change in surface roughness occurred in the first two hours of the heat treatment, and
that heat treatment temperature has little effect on the rate or the magnitude of the rough-

67
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
SPUTTER TIME, min.
Figure 4.26: Auger depth profile of 5052 alloy thin film. Film sputtered with
3 keV argon in a 3 mm x 3 mm raster.
ness change. AES depth profile data in Figure 4.26 shows that the surface oxide was
between 20nm and 30nm thick.
Powder x-ray diffraction analysis of 5052 thin films heat treated at 50°C or 125°C
for up to 32 hours showed that the (111) peak shifted low 0.1° for heat treated (2 to 32
hours) versus unheated samples. No additional peak shifts were detected for samples
heat treated longer than two hour. The peak shift for the (220) showed the same trend
with a shift of 0.1°. Samples were also scanned to detect changes in the Al3Mg2 peak
heights or widths that would indicate a change in these precipitates. No measurable
change was seen in these peaks indicating that the precipitates had not coarsened. Line
shape analysis was also performed on the X-ray data to determine crystallite size as a
function of stress relaxation, but no correlation was found with stress relaxation time or
temperature.

68
O 50 100 150 200 250 300 350
Temperature, "C
Figure 4.27: Stress versus temperature plot of 2090 aluminum alloy thin film.
2090 Aluminum
Figure 4.27 shows the stress versus temperature plot for the 2090 aluminum alloy
thin film for the first and second temperature cycle to 350°C. The as deposited tensile
stress in this film is 250 MPa before the first cycle. During the first temperature cycle,
after deposition the film stress decreases linearly with increasing temperature up to
140°C. Between 140°C and 200°C the stress in the film increases with increased tempera¬
ture, presumable due vacancies migrating to and being destroyed at the grain boundaries
[Tow87]. This behavior was typical for T201 and 5052 films also during the first thermal
cycle. Once the excess vacancies were eliminated, the slope again becomes linear up to
275°C. At 275°C the slope increases and becomes unstable up to 350°C, where the film is
under a compressive stress of 70 MPa. Upon cooling, the slope of the stress-temperature
curve is much less than during heating and the film only reaches a tensile stress of 80

69
SPUTTER TIME, min,
Figure 4.28: Auger depth profile of 2090 aluminum alloy thin film. Film sputtered with
3 keV argon in a 3x3 mm raster.
MPa at room temperature. On the second temperature cycle the film attains a slightly
higher compressive stress of 80 MPa at 350°C and returns to a tensile stress of 80 MPa
upon returning to room temperature. When the sample was removed from the Flexus it
was noticed that the specular reflection of the film was greatly reduced due to surface
oxide.
Based on the reduced specular reflection and the lower slope of the stress tem¬
perature curve it was assumed that the film had oxidized badly. This was confirmed by
the Auger depth profile data from that sample, (see Figure 4.28). As can be seen, the
oxygen concentration falls off very gradually. This indicates that the film had oxidized to
a depth greater than the 5052 or T201 films (see figures 4.26 and 4.13).

Copper
70
The stress versus temperature plot for the copper thin film is shown in figure 4.29.
While this plot is quite complex the data are consistent with the film being oxidized
during heat treatment. As the stress-temperature plot indicates, the film was stable during
the first thermal cycle to only 100°C. Another film was tested for stress relaxation at
50°C for 22 (see figure 4.30). The copper film oxidized in this test, as indicated by a
greatly reduced the specular reflection, and the film stress increased after 14 hours due to
oxidation.
To inhibit the oxidation of the copper thin film, a 75 nm thick protective alumi¬
num layer was deposited on the copper. This sample was then heated to 500°C for 6
hours in a tube furnace to test the stability of this protective coating. The film retained its
specular reflection during this test. From an Auger depth profile on this sample the oxide
thickness is approximately 200nm (see figure 4.31).
Figure 4.29: Stress versus temperature plot for a copper thin film.

O 5 10 15 20 25
Time, Hours
Figure 4.30: Isothermal stress test of a copper thin film conducted at 50°C.
Figure 4.31: Auger depth profile of a copper film coated with a protective layer of alumi¬
num 750 nm thick. The film was sputtered with 3keV argon in a 3x3 mm raster.

72
Titanium. Manganese and Nickel
The stress versus temperature plots for titanium, manganese, and nickel are shown
in figures 4.32 through 4.34. All of these films oxidized severely during their first tem¬
perature cycle up to 500°C, 250°C, and 500°C for titanium, manganese and nickel respec¬
tively. This was evident from a change in specular reflection, color change of the samples
and by the stress-strain data. For titanium (figure 4.32) the rapid reduction in tensile
stress at 350°C to a compressive stress of over 600 MPa at 500°C and the film maintain¬
ing a high tensile stress (375 MPa) upon cooling to room temperature indicates some
changes in the film from oxidation. Oxidation of the manganese film (figure 4.33) results
in an 800 MPa shift in the stress, becoming highly tensile at just over 200°C. The nickel
film (figure 4.34) shows a change in slope of the stress strain curve at 275°C and the
residual stress at room temperature dropped from 1000 MPa before to 600 MPa after the
0 100 200 300 400 500
Temperature, C
Figure 4.32: Stress versus temperature plot for a titanium thin film.

73
Temperature C
Figure 4.33: Stress versus temperature plot for a manganese thin film.
0 100 200 300 400 500
Temperature, C
Figure 4.34: Stress versus temperature plot for a nickel thin film.

74
thermal cycling due to oxidation. An aluminum protective coating was deposited onto a
manganese sample, but did not inhibit oxidation of the sample.
Copper-Gold Intermetallics
After the copper-gold films had undergone the homogenization and ordering heat
treatment, their chemical composition was analyzed by EPMA (see Table 4.6). The films
were within 2 to 3 percent of the desired composition.
In powder x-ray diffraction all of the films showed a shifts in the diffraction peak
positions form disordered to the ordered state, and additional peaks were observed from
the long range order. As reported above, for textured thin films it was not possible to
calculate the degree of ordering.
The films were thermally cycled in the Flexus up to 450°C in an argon atmo¬
sphere. Figures 4.35 through 4.37 show the results of these tests. The Cu3Au films did
not show a change in stress near the order-disorder temperature, 400°C 9figure 4.35).
The CuAu film showed a marked change in stress at the transition temperature (figure
4.36) as did to a lesser dagree the CuAu3 film (figure 4.37). However, the order-disorder
transition occurs well above the temperature (300°C) at which the films begin to plasti¬
cally deform.
Table 4.4: Chemical composition of copper-gold thin films.
Atomic %
Cu
Atomic %
A1
Cu,Au
77 +/- 0.2
23 +/- 0.2
CuAu
53 +/- 0.4
47 +/- 0.5
CuAu,
22 +/- 0.4
78 +/- 0.4

Stress, MPa Stress, MPa
O 1OO 200 300 400 500
Temperature, C
Figure 4.35: Stress versus temperature plot for a Cu,Au thin film.
Figure 4.36: Stress versus temperature plot for a CuAuthin film.

600
76
Figure 4.37: Stress versus temperature plot for a CuAu,thin film.
Figure 4.38: Isothermal stress relaxation of ordered and disordered CuAu,
thin films at 150°C.

77
A series of isothermal stress relaxation test were performed on these samples.
These tests were conducted in argon since these films oxidize if heated to 150°C in air.
The tests were run at 150°C to maximize mobility in the samples. Figure 4.38 shows the
isothermal stress relaxation data for an ordered and a disordered CuAu3 thin film. The
ordered film has a 2 percent reduction in stress over two days while the disordered film
has a 7 percent reduction in stress over the same period. The isothermal stress relaxation
in the CuAu and Cu,Au films showed no difference between the ordered and disordered
states.
Aluminum-Titanium Intermetallics
The as-deposited tensile stress in the layered Al3Ti film was 200 MPa. On the
first temperature cycle to 500°C the slope of the stress temperature curve changed at
350°C, indicating a change in the structure of the film. At 500°C the film was in a state of
no stress. When the film cooled to room temperature it was under a tensile stress of 800
MPa as shown in figure 4.39. The film was cycled up to 500°C two additional times and
showed no hysteresis in the stress-temperature plot. An isothermal stress relaxation test
was performed on this sample at 350°C and the film showed no stress relaxation over a 48
hour period.
The sample was then tested at Tencor up to 800°C. The results of this test are
shown in figure 4.39. The Al,Ti film begins to plastically deform above 500°C under a
compressive stress of 200 MPa. At 800°C the film still maintains a compressive stress of
300 MPa. Upon returning to room temperature the film was under a tensile stress of
nearly 1000 MPa. An isothermal stress relaxation test was also performed at 450°C and
again the film showed no stress relaxation over a 48 hour period.
The native passivating oxide held up well for all tests at all temperatures. The
specular reflection of the sample was high after all the tests. There was no visible change
in the oxide after cycling to 500°C. Figure 4.40 shows the Auger depth profile for an

1000
78
Figure 4.39: Stress versus temperature for Al3Ti thin film.
Al,Ti film that was heated to a maximum of 500°C. The oxygen concentration drops
sharply after 5 minutes of sputtering. Figure 4.41 shows the Auger depth profile for a
sample heated to 800°C. The oxygen concentration drops to near zero after 25 minutes of
sputtering, indicating that the surface is passivated, but the oxide layer is about three
times thicker for samples heated to 800°C are than samples heated only to 500°C.
Modeling the Effects of Oxide Thickness on a Bimetallic Actuator’s Curvature
As has been seen with many of the materials examined in this study oxidation has
a significant impact of the performance of the material. For the aluminum alloys, it was
not possible to test for the effects of the oxide layer since one always forms when alumi¬
num is exposed to air. To evaluate the effect of the oxide on the performance of a bime¬
tallic strip, the change in curvature of a bimetallic strip as a function of the AI203 layer
thickness was calculated. This calculation is based on a bimetallic strip that consists of 5

PEAK HEIGHT PEAK HEIGHT
79
SPUTTER TIME, min.
Figure 4.40: Auger depth profile of Al,Ti sample that was heated to 500°C.
SPUTTER TIME, min.
Figure 4.41: Auger depth profile of Al3Ti sample that was heated to 800°C.
pm of aluminum on 5 pm of silicon with an oxide whose thickness varied from 0 to 1

100
80
AI2O3 Thickness
microns
Figure 4.42: Curvature of a bimetallic strip as a funtion of oxide thickness as a percent¬
age of maximum curvature.
pm. Figure 4.42 shows the results of this calculation. From this calculation it can be
seen that the 30 nm oxide layer found on the aluminum alloy samples reduces the curva¬
ture of the bimetallic strip by 0.5 percent, but thicker oxides would have a more signifi¬
cant impact. Therefore, controlling the oxide thickness is critical to maximizing the
performance of a bimetallic strip.

CHAPTER 5
DISCUSSION
Introduction
In this chapter the results and conclusions that have been drawn from this research
will be presented and discussed with respect to the three objectives of this research, (i)
effectiveness of strengthening techniques in thin films, (ii) identification of the mecha¬
nism responsible for isothermal stress relaxation in thin films, (iii) identification of
materials for high temperature applications. In addition, to assist the development of
bimetallic actuators, a figure of merit has been developed and a number critical param¬
eters have been identified to assist in the selection of materials for this application.
Thin Film Strengthening
Of the five strengthening techniques used in bulk metals, this study examines the
effectiveness of two of these, solid solution strengthening and multiphase hardening
(precipitation/age hardening) in thin films. While not normally considered a strengthen¬
ing technique, the effectiveness of ordered intermetallics to resist stress relaxation was
also examined. Solid solution and precipitation/age hardening were selected because of
their effectiveness in strengthening bulk metals and because these mechanisms are rela¬
tively insensitive to small variation in composition, which improves manufacturability.
The order/disorder transition that occurs in some intermetallics was also of interest due
to the large change in volume that frequently occurs at this transition. The order / disor¬
der transition provided a means of evaluating the effectiveness of the ordered phase to
resist stress relaxation.
81

82
With respect to measuring the strengthening effects of these techniques, there are
a number of parameters that are of interest. With bulk materials, strengthening is usually
evaluated by measuring the yield strength or ultimate tensile strength of the material at a
temperature. In thin film bimetallic structures, this is not possible since stress and tem¬
perature are dependent on one another. Thus the temperature and stress at which plastic
deformation begins are interrelated, and is the point on the stress versus temperature
curve where the slope significantly departs from linear. This study is also concerned with
the stress relaxation that occurs in thin films, which determines the stability of a bimetal¬
lic actuator. So the strengthening techniques must be evaluated with respect to their
effectiveness at increasing both the temperature and therefore the stress at which plastic
deformation occurs and at reducing isothermal stress relaxation in the film.
While it would be reasonable to assume that the bulk strengthening techniques
would be effective in thin films, the AlSiCu alloy used in microelectronics would suggest
that the strengthen mechanisms are less effective in thin films. In the AlSiCu alloy,
silicon could provide some solid solution strengthening, but the solubility of silicon in
aluminum at room temperature is very low. Copper could provide precipitation strength¬
ening, but due to processing is generally in the overaged condition. As shown in Figure
2.9, the strength of the AlSiCu alloy is only 30% greater than pure aluminum. In contrast,
age hardened bulk aluminum copper alloys can have yield strengthens of five times those
of pure aluminum. While this suggest that bulk strengthening mechanism may not be as
effective in thin films, it must also be noted that the AlSiCu alloy was not developed for
strength but to reduce aluminum spiking into silicon wafer (Si) and to reduce
electromigration (Cu) [Bow74]. On the other hand, the alloys evaluated in this study
were developed for strength.
Solid Solution Strengthening
The aluminum alloy 5052 was selected to test solid solution strengthening. This
alloy is solid solution strengthened by the addition of magnesium. While the majority of

83
the magnesium stays in solution, some Al3Mg2 and Al2Mg precipitates form. However,
these precipitates do not provide the primary strengthening. Comparing the stress tem¬
perature curve for 5052 aluminum (Figure 4.16) with the curves for pure aluminum and
the AlSiCu alloy (Figure 2.10) clearly shows that solid solution strengthening is effective
in thin films. The room temperature residual stress for the 5052 alloy is 300MPa as
compared to 150MPa for pure aluminum and 200MPa for the AlSiCu alloy. Also plastic
deformation begins in the 5052 alloy sample at 225°C under a compressive stress of
40MPa while the pure aluminum and AlSiCu alloy begin to deform under the same stress
but at much lower temperatures of 125°C and 175°C, respectively. Therefore solid solu¬
tion strengthening has significantly increased the performance of the film.
Testing also showed that the 5052 alloy film is more resistant to stress relaxation
than a pure aluminum or AlSiCu films. At 50°C, pure aluminum and the AlCuSi alloy
films experience a 17% reduction in stress over 48 hours while the 5052 alloy film only
experience a 6.7% reduction in stress over the same time period. This is even more
significant because the 5052 alloy film was under a tensile stress of 150 MPa while the
pure aluminum was under only a 80 MPa compressive stress.
In addition to the 5052 aluminum alloy the copper / copper gold systems were
examined. The copper gold system was examined primarily because it undergoes an
order disorder transition. However, in the disordered phase it can be considered a solid
solution of the two elements and it can be compared to pure copper. From (Figure 4.29)
the room temperature residual tensile stress for pure copper is 50 MPa and the films
began to plastically deform above 250°C under a compressive stress of 300 MPa. The
stress temperature plot for Cu3Au (Figure 4.35) shows a residual room temperature tensile
stress in the film of 575 MPa and the film begins to plastically deform at 300°C under a
compressive stress of 50 MPa. As with aluminum, solid solutions of copper and gold
nearly doubled the strength of the fdm. Stress relaxation tests were not done on the pure
copper films due to oxidation.

84
Given the results of the 5052 aluminum and copper gold, it is clear that solid
solutions are effective in strengthening thin films. It has also been shown that this
strengthening mechanism is effective at reducing stress relaxation in aluminum alloys.
Precipitation and Multiphase Hardening
The T201 aluminum alloy was selected to determine the effectiveness of precipi¬
tation hardening to reduce stress relaxation. The T201 aluminum alloy is not strength¬
ened only by precipitation hardening. Magnesium and manganese are added for solid
solution strengthening and titanium is added to reduce grain size. Copper is the primary
alloying element and it leads to strengthening by precipitates. Silver is added to stabilize
the copper precipitates. Table 4.2 shows the complete composition of the T201 alloy.
Since this alloy is precipitation and solid solution strengthened, comparisons must be
made to pure aluminum, AlSiCu alloy, and the 5052 alloy to evaluate the effectiveness of
the precipitation strengthening.
Figure 4.1 shows the stress-temperature plot for the T201 alloy. The room tem¬
perature residual stress is nearly 400 MPa, which is 100 MPa higher than the 5052 alloy.
(Figure 4.16), 2.5 times greater than pure aluminum and 2 times greater than the AlSiCu
alloy (Figure 2.10). The T201 alloy began to plastically deform at 225°C under a com¬
pressive stress of 75 MPa, again much higher than pure aluminum or the other aluminum
alloys. At 50°C the T201 alloy thin film undergoes a 12% reduction in stress over 48
hours (Figure 4.2) as compared to 7% for the 5052 samples (Figure 4.17) and a reported
value of 15% for the pure aluminum and AlSiCu alloy. However, at 50°C, the T201 film
is under a tensile stress of 390 MPa as compared to 270 MPa for the 5052 alloy films, 75
MPa for the pure aluminum film and 150 MPa for the AlSiCu alloy film. At 125°C, the
T201 sample undergoes a 2% stress relaxation over 48 hours under a 114 MPa tensile
stress (Figure 4.5) while the 5052 alloy film undergoes an 18% stress relaxation under a

85
104 MPa tensile stress (Figure 4.20). Therefore, precipitation/age hardening is effective
at reducing stress relaxation in thin films.
The plane view TEM photomicrograph in Figure 4.11 verifies that precipitates
were present in the aluminum thin film. In this image, precipitates are clearly seen within
aluminum crystals and the diffraction pattern for CuA12 is seen. The size of these precipi¬
tates is on the order of 20nm to 30nm which is appropriate for age hardening [Bro82],
Based on the testing and microstructural examination of the sample it can be
concluded that precipitation/age hardening is effective in thin films. It should also be
noted that the addition of silver to this alloy stabilizes the CuA12 precipitates. Without
this stabilization, the copper would precipitate at the grain boundaries, have a coarse
dispersion rather than a fine dispersion, and be much less effective at strengthening the
film. A coarse dispersion of “overaged” precipitates is the reason for lack of strength in
the AlSiCu alloy [Bro82].
Ordered Phases and the Order Disorder Transition
While not normally considered a strengthen technique, ordered phases generally
have a higher strength than their disordered counterparts. As mentioned above, order
disorder transitions were also of interest due to a large change in volume that occurs
during this transition. This volume change could generate large stress or deflection
values in the design of thermally actuated MEMS. Figures 4.37, 4.38 and 4.39 show the
stress temperature curves for the three copper gold compounds tested. While there is a
substantial change in stress seen near the order disorder temperature, this occurs beyond
the point at which the yield strength / temperature conditions lead to plastic deformation
for these films and the order/disorder transition is not of use in this system.
Ordered phases were presumed to have a lower stress relaxation rate than disor¬
dered phases since the self diffusion rate, which determines the creep rate and stress
relaxation rate for many materials, were expected to be lower in an ordered phases versus

86
a disordered phase [Cah83], Figure 4.38 clearly shows that this presumption was true, for
Cu,Au, the ordered phase showed a 1.6% reduction in stress over two days compared
with a 6.5% reduction in stress for the disordered phase. The disordered film’s stress
relaxation rate is almost four time faster than the ordered phase, while the stress is only
8% higher on the disordered film.. The Al3Ti ordered intermetallic also showed no stress
relaxation at 450°C in two days.
Stress Relaxation Mechanism in Thin Films
While increasing the strength of the aluminum thin films reduced the stress
relaxation rates and magnitudes, it did not eliminate stress relaxation. Thus the mecha¬
nisms responsible for stress relaxation in thin films needs to be discussed. In this study
three modes of stress relaxation have been observed. The first mode is plastic deforma¬
tion of the thin film [Her85], If the stress applied by the substrate is greater than the yield
strength of the material, the fdm will plastically deform and relax the stress. This is seen
as a change in slope of the stress-temperature curves upon heating (Figures 4.1,4.16,
4.35, 4.36, 4.37,4.39). The second mode of stress relaxation occurs over several tens of
minutes after a sample has reached temperature. This is seen in the stress relaxation
curves of the T201 and 5052 aluminum alloy films (Figures 4.2-4.4 and 4.18-4.20). The
third mode of stress relaxation occurs over a number of days and can again be seen in the
stress relaxation plots of the T201 and 5052 samples.
The challenge in determining the stress relaxation mechanisms in thin films on a
substrate is that film stress and temperature can not be independently varied. This elimi¬
nates the use of an Arrhenius equation to calculate the activation energies of the processes
under investigation. It is possible to manipulate samples at different temperatures to the
same stress, however the microstructure and dislocation array of a material are not state
functions. Therefore, their structures will be quite different at different temperatures even
if the stress is the same. In addition, removing the film from the substrate would, in many

87
cases, substantially change the film in a number of ways. For self passivating materials,
such as examined in this study, the back side of the film would oxidize upon exposure to
air. It has been shown that dislocations interact differently with oxidized surfaces
[Nix89], It would also be difficult to create and apply a biaxial stress to a free standing
thin film and to measure strain in the film. Therefore, direct determination of an activa¬
tion energy and thus the mechanisms responsible for the different modes of stress relax¬
ation is not possible. Other data are needed to determine the mechanisms responsible for
stress relaxation in thin films.
It should be noted that the change in total strain and temperature to reduce the
stress in these films is very small. A change in stress of 10 MPa is roughly equivalent to
a temperature change of only five degrees. For aluminum on silicon, this is a strain of
only 0.00012. Thus very small changes in the dimensions of a film can result in large
changes in the stress in the film.
The inability to calculate activation energies for stress relaxation modes 2 and 3
makes it very difficult to identify their mechanisms. All plausible relaxation mechanism
must be considered including: recovery , recrystallization, logarithmic creep, Andrade
creep, bulk diffusion, gain boundary migration, grain boundary sliding, surface to grain
boundary diffusion, polygonization/subgrain coalescence, and precipitate formation or
coarsening. Several of these mechanism can be eliminated as possible causes of both
modes 2 and 3 stress relaxation.
For the aluminum alloys at 150°C, the TH is less than 0.5, which is slightly lower
than the temperature at which Andrade creep would be expected. In addition, none of the
stages of Andrade creep would produce an exponential decay as is seen in the stress
relaxation data. Samples heated at higher rates showed no transition from stage 1 creep
to stage 2 creep. In the early stages of relaxation the stress relaxation rate decrease, but
does not go to zero, and the stage 1 strain rate is not a simple exponential [Cah83] decay

88
as seen in the data. Stage 2, steady state creep has a constant strain rate that would
produce a constant stress relaxation rate that is not seen in the data. The strain rate in
stage 2 creep has some dependence of the applied stress, but this relationship is not a
simple exponential [Cah83], In the third stage of Andrade creep the stain rate increases
and so is not consistent that data that has been collected. Therefore, Andrade creep does
not appear to be responsible for the stress relaxation.
Diffusional creep processes are generally significant at temperatures of, TH> 0.6,
which is well above the temperatures at which the current stress relaxation is seen (TH <
0.5). Diffusional process increase exponentially with temperature [Cah83] while the
present stress relaxation shows an inverse relationship to temperature. Diffusional pro¬
cesses should result in a zero stress at infinite time instead of the finite residual stress
projected from the furve fits (Tables 4.1 and 4.3).
X-ray line shape analysis did not detected polygonization or subgrain coalescence,
so these mechanism can be eliminated. In addition, subgrain coalescence [Cah83] is a
preliminary step to recovery and recrystallization, and the microstructure of the samples
did not change during stress relaxation (Figures 4.7 and 4.12). Thus recovery, recrystalli¬
zation and grain boundary migration do not appear to be responsible for the stress relax¬
ation. The formation or coalescence of precipitates does not appear to be a factor. X-ray
diffraction showed no significant change in precipitate concentration over time. In
addition, the same two modes of stress relaxation have been seen in pure aluminum
[Her85]. Having eliminated these potential mechanism of stress relaxation it is necessary
to separately consider the mechanism applicable to each mode.
Mode 2 Stress Relaxation
Mode 2 stress relaxation occurs relatively quickly. It is virtually complete in two
hours. During this time there is a measurable increase in the surface roughness of the
film (Table 4.3). There is also a 0. Io shift in the (111) x-ray diffraction peak. The in-

89
crease in surface roughness equates to a 0.15% change in the thickness of the 1.2fxm film
which is of the same order as the change in the d spacing calculated from the x-ray peak
shift of 0.1°. This indicates that the processes affecting the surface is occurring thoughout
the thickness of the film. Surface to grain boundary diffusion is not the likely cause of
this stress relaxation since diffusional processes are temperature sensitive as discussed
above [Ask89], while the roughness and peak shifts are temperature insensitive. Grain
boundary sliding is also not consistent with the results obtained since it normally has a
very high activation energy [Mey84], A high activation energy is not consistent with a
process that occurs rapidly at low temperature. The film also has a columnar grain
structure [Kra90, Tho74] so the resolved shear stress on most of the grain boundaries will
be very small. If some crystals have a tilted grain boundary, these grains should be
visible with the AFM. Therefore, grain boundary sliding is eliminated as a potential
explanation of stress relaxation.
The only mechanism not yet eliminated to explain stress relaxation is logarithmic
creep which results from dislocation glide on slip systems. It is postulated that mode 2
stress relaxation is caused by the movement of dislocation on slip planes that terminate as
the surface of the film. Dislocation moving in these slip systems should not encounter
dislocation pile ups as would be encountered in slip system that terminate at grain bound¬
aries. Thus these dislocations should move and be eliminated quicklt. The degree of
relaxation should only be limited by the number of sources of mobile dislocations. The
increase in roughness could be caused by dislocation bands. Also because of the stress
level and temperature of these films, the DMM’s predict that logarithmic creep would be
the dominant process (Figure 2.5) [Cah83],
Mode 3 Stress Relaxation
The mechanisms not yet eliminated for mode 3 stress relaxation are grain bound¬
ary sliding, surface to grain boundary diffusion, and logarithmic creep. Surface to grain

90
boundary diffusion is a diffusional process that should be exponentially dependent upon
temperature. It could increase surface roughness as atoms migrate from the surface to the
grain boundary. However, the data do not show a large temperature dependence and the
surface roughness does not increase over stage 3. The temperature at which the stress
relaxation is occurring is also low for a diffusional process to be active (TH < 0.5). There¬
fore, surface to grain boundary diffusion does not appear to be the mechanism responsible
for this mode of stress relaxation. Grain boundary sliding also not likely to be the mecha¬
nism responsible for this stress relaxation, since grain boundary sliding should also
increase surface roughness. In addition, the activation energy for grain boundary sliding
is very high and it normally only occurs just before mechanical failure. Furthermore, in a
columnar structure the resolved stress on most grain boundary should be very low. This
only leaves logarithmic creep, which was concluded to be the mechanism for mode 2
stress relaxation. However, how can the same mechanism have a different rate constants?
Different resolved stresses on different slip directions could explain this in part, but, the
slip systems are all equivalent within each of the (100),(110) and (111) oriented crystals.
For this to be the cause of change in relaxation rate, the different oriented crystals would
have to relax at different rates, and this should be noticeable in the AFM images. Also,
no increase in the roughness of the film was seen in mode 3. It is known that disloca¬
tions pile up at grain boundaries [Mey84], Dislocation climb could move some of these
dislocations into the grain boundary. This would result in a reduction of stress and no
increase in surface roughness. Therefore this mechanism is consistent with the observa¬
tion made in this study and is postulated to be the origin of the mode 3 stress relaxation.
High Temperature Application
Micro actuators are needed that can operate in severe environments, such as inside
jet engines. In these severe environments oxidation was assumed to be a major concern.

91
High temperature application also require materials that maintain their strength at the
operating temperatures. These topics will be examined in this section.
Oxidation
During this study it was found that oxidation is a pervasive problem affecting all
temperatures ranges. As is shown above and in Appendix C, a 30nm thick oxide forming
of the surface of a 5pm A1 - 5pm Si bimetallic strip reduces the curvature of the strip by
0.5%. The 2090 aluminum lithium alloy was found to be unsuitable for bimetallic actua¬
tors because the lithium increased the oxidation rate (Figure 4.29). The other elements
tested also behaved poorly. At elevated temperatures, copper, titanium, manganese, and
nickel all oxidized (Figures 4.31,4.32, 4.34, 4.35). In bulk metals application the thick¬
ness of the passivating oxides is of little concern as long it remains intact. In thin films
the passivating oxide can consume a large fraction of the thickness. This is very damag¬
ing to a bimetallic actuator because most oxides have a very low CTE which will reduce
the displacement of the device. Thus, a material’s oxidation rate is a critical parameter
and those metals with the slowest oxidation rate are best suited for this application.
Aluminum and silicon both have very low oxidation rates due to the formation of passi¬
vating oxides and both are well suited for elevated temperature applications [Wes95]. A
thin aluminum layer was also found to be effective at protecting less resistant materials,
such as pure copper.
Strength at Elevated Temperatures
As discussed above, intermetallics have a high resistance to creep due to their low
self diffusion rates. This is of interest for elevated temperature application where a lower
melting point intermetallic with a higher CTE could be used in place of a standard higher
melting point alloy with a low CTE. The Al3Ti intermetallic performed well, showing no

92
stress relaxation at 450°C and no oxidation. Unfortunately nickel and copper oxidized, so
comparisons of stress relaxation could not be made.
Figure of Merit
To assist engineers who design bimetallic actuators in selecting the active layer
material, a figure of merit has been developed. The figure of merit is designed to provide
an initial evaluation of the type of material that should be used, i.e. aluminum or nickel.
The figure of merit, FOM, is expressed in equation 5.1 as:
FOM = ( YS'5. [( FD ■ CTE ■ 106<€)+ (E ■ 10'° Pa1 ■ ( 1 - FD ))f (5.1)
' V ' ' V ' ' V '
Temperature compensated Displacement Force
yield strength mode mode
where MHT is the maximum homologous temperature (the homologous temperature
above which the yield strength of the material rapidly declines, equal to 0.4 for most
metals and 0.65 for most intermetallics), MOT is the maximum operating temperature of
the actuator, YS is the yield strength of the material at room temperature, CTE is the
coefficient of thermal expansion for the active layer material, E is Young’s modulus for
the active layer material and FD, determines whether the priority of the device is force or
displacement (0 to maximize force, and zero to maximize displacement). Table 5.1
shows a comparison of the different figures of merit for different materials at different
operating temperatures and for force or displacement. Note that the values for tin are
negative, indicating it is not suitable at the temperatures. In general the large CTE of
aluminum alloys give the FOMs of ~ 10s for displacement while the higher Young’s
modulus and moderate CTE’s of nickel and stainless steal give them FOMs of ~10n-1012
for force. Note the FOM at 500°C of 1015 for cobalt. This figure of merit is complex,
but could be expanded to also include the environmental stability of the materials.

93
Table5.1: Figure of merit ratings for different materials at different operating tempera¬
tures and optimized for displacement, D, or force, F.
Max. Temp.
100 (°C)
250 (”C)
500 (°C)
Force/
Displacment
D
F
D
F
D
F
Sn
-5.8*106
-3.4*107
-2.9*10*
-1.7*107
-1.5*10*
-9.1*10*
5052 A1
1.4*10*
8.8*10’
-2.8*107
1.7*10’
-9.3*10*
-5.7*10*
T210A1
1.1*10*
1.2*10'°
-2.4*107
-2.7*10’
-8.3*10*
-9.0*10*
Mg
8.1*107
8.0*10*
-3.2*107
-3.2*10*
-9.6*10*
-9.5*10’
Anneal A1
2.8*107
1.3*10’
-1.3*107
-6.0*10*
-3.7*10*
-1.7*10*
75% c.w, AI
5.0*107
2.4*10’
-2.3*107
-1.1*10’
-6.6*10*
-3.1*10*
Cu
7.5 *106
1.4*1010
2.1*107
3.8*10'°
-1.1*10*
-2.0*10'°
Ni
1.0*106
6.1*1010
1.7*10*
9.7*10'°
1.1*10*
6.6*1012
Co
4.2*10*
5.4*10‘3
6.5*10*
8.3*10'3
7.8*107
1.0*10'5
410 S.S.
2.4*10*
2.3*10"
3.6*10*
3.4*10"
2.6*10*
2.5*10'2
c.w. cold worked
S.S. stainless steal

CHAPTER 6
CONCLUSIONS
In this study, the two modes of operation for a bimetallic actuator, force and
displacement, were studied and each mode has different materials requirements to maxi¬
mize performance. The displacement mode is dominated by the difference in the CTE’s
of the materials of bimetallic strip, While performance in the force mode is a function of
both Young’s modulus and the CTE difference. In both modes the yield strength limits
the maximum displacement or force produced by an actuator. Based on these results this
research examined mechanisms to strengthen high CTE thin film metallization.
Two strengthen mechanisms were evaluated, solid solution strengthening and
multi-phase/age hardening. The 5052 aluminum alloy thin film, which is a solid solution
strengthened alloy, was twice as strong as pure aluminum, having a room temperature
residual stress of 300 MPa as compared to 150 MPa for pure aluminum. The T201
aluminum alloy thin film, which is a solid solution and multi-phase/age hardened alloy,
was 2.5 times stronger than the pure aluminum with a room temperature residual stress of
400 MPa. Therefore, solid solution and multi-phase/age hardening are effective at
strengthening thin films. However, as with bulk alloys, multi-phase/age hardening is
susceptible to over aging as was seen in the Al-Si-Cu alloy films, which only showed a
room temperature residual stress of 200MPa.
Two modes of isothermal stress relaxation were observed in the 5052 and T201
alloy thin films at temperatures between 50°C and 250°C. In the first mode, stress relax¬
ation is caused by the movement of dislocation in slip systems that terminate at the thin
film surface. The second mode is caused by the movement of dislocation in slip systems
that terminate at grain boundaries.
94

95
The order-disorder transformation found in the copper gold system, while produc¬
ing a large change in stress for a small temperature change, occurred at temperatures
which generated stresses which were above their yield strengths. In addition, these films
oxidized at the temperatures necessary for the order/disorder transition. Therefore, order/
disorder transformations in the Cu/Au system is not useful for bimetallic actuation.
These films provided a means of comparing the stress relaxation rates for ordered and
disorder/solid solution phases. The ordered phase showed one quarter the stress relax¬
ation as compared to the disordered phase. Therefore, ordered intermetallic thin films
have a high resistance to stress relaxation. This was also seen in the Al3Ti films that
showed no stress relaxation over 48 hours at 450°C. The Al3Ti films also showed high
strength, only plastically deforming at 500°C or higher, and oxidation was only significant
above 800°C. Al,Ti is therefore a promising candidate for high temperature actuators.
Oxidation was found to have a pervasive effect on all materials evaluated in this
study. The oxide formation on aluminum thin films has been shown to increase the
strength of the film [Nix89], but calculation performed in this study showed that the
oxide also reduces the displacement in a bimetallic actuator. Oxidation was found to
have an even greater detrimental effect on some of the other materials tested. Copper,
nickel, titanium, manganese, and the 2090 aluminum alloy thin films all oxidized se¬
verely upon their first heating to 350°C. The copper film even oxidized at 50°C over 48
hours. Therefore, materials for use in bimetallic actuators must have a very low oxidation
rate.

CHAPTER 7
FUTURE WORK
In the area of strengthening mechanisms, additional work could be done to evalu¬
ate different solutes and precipitates that could be used. As for the stress relaxation
mechanisms, additional theoretical and analytical work should be done to support or
disprove the conclusions reached in this study. From the analytical perspective, the use of
nickel as a base material has the advantages that TEM samples can be more easily pro¬
duced with less concern that the sample preparation is affecting the data.
Oxidation presents an interesting problem for bimetallic actuators because it
reduces displacement but also increases the strength of the thin film [Nix89], Putting a
protective polymer coating on an aluminum thin film to inhibit oxidation would signifi¬
cantly reduce the yield strength of the aluminum, but the oxide would not reduce dis¬
placement. How these two factor interact needs to be better understood. Also, as an
oxides can strengthen a thin film what other surface treatments would be effective at
strengthening thin films?
Protective coatings would require an additional patterning and deposition step in
the manufacturing process and so are of limited interest. This put the focus of future
work on self passivating or non-oxidizing materials.
Moving to topics not covered in this study, micro bimetallic actuators may be
cycled millions of time. There is therefore concern about fatigue in the active layer
material. This is an area that will become more critical as more actuators are more widely
implemented.
Lastly, this study only examined thin films on wafers where the film had very little
curvature or stress gradient through the thickness of the film. In real actuators, thin films
96

97
may be exposed to highly curved structures with large stress gradients across the film
thickness. This could had dramatic effects on the stress relaxation within the films and
should be examined.

APPENDIX A
DERIVATION OF RELATIONSHIP BETWEEN YOUNG’S
MODULUS AND COEFFICIENT OF THERMAL EXPANSION
The following relationship between a material’s Young’s modulus and its coeffi¬
cient of thermal expansion was derived by N. Kulkarni and R. DeHoff [Kul97], Starting
with the thermodynamic relationship given in equation A1 [Deh93],
dV= FadT- vpdP
(Al)
where V is volume, P is pressure, T is temperature, a is the coefficient of thermal expan¬
sion and p is the coefficient of compressibility. By algebraic manipulation and differen¬
tiation with respect to pressure while keeping the volume the same, equation A2 is de¬
rived.
(A2)
The partial derivative of temperature with respect to pressure at constant volume is a
positive valued function because the temperature of a system must increase in response to
increased pressure for the system to maintain a constant volume.
98

99
Turning to the equations that define and relate the different materials properties
[Die84], equation A3 define the bulk modulus, K, of a material:
P = . (A3)
The bulk modulus is then related to the shear modulus, G, and Young’s Modulus, E, by
equation A4:
K =
(A4)
The shear modulus is related to the Young’s modulus by equation A5 where v is Poisson
ratio:
G =
E
2( 1 + u ) '
(A5)
Combining equations A4 and A5, the relationship between the bulk compressibility and
the Young’s modulus is obtained as equation A6:
K =
E
3 + 6t)
(A6)
By combining equations A2 and A3 the relationship between the bulk compressibility and
the coefficient of thermal expansion is obtained, equation A7:
(A7)

100
By combining equations A6 and A7, the relationship between the Young’s modulus and
the coefficient of thermal expansions is derived in equation A8:
3 + 6t)
(A8)
Equation A8 clearly shows that there is an inverse relationship between the Young’s
modulus and a material’s coefficient of thermal expansion.

APPENDIX B
CURVE FITTING OF STRESS RELAXATION DATA
The stress relaxation data were collected over a 48 hour period. Stress measure¬
ments were taken every 15 minutes producing 192 data points per stress relaxation test.
These data were exported from the Tencor Flexus model 2320 into an ASCII formatted
file. This ASCII file was then imported into Microsoft Excel where all but the 192 stress
data points were eliminated. The data were exported as an ASCII data file containing
only the 196 stress data points delimited by carriage returns and read into Mathcad ver¬
sion 4.0 where the following procedure was used to fit the data.
T201 Al @ 10QQC
The data are being fit to an equation of the form
Stress(t) = A + B -exp(mi't) + C-exp^-t)
i o.. i9i Used to index data points
j - o„ 166 Used to index data points when not using first 25 data points
r:= o.. 3 Used to index columns of data
d := READPRN(a2ioo) Reading in data
Dir r := o Zeroing data array
cnot = 160.6 This is the A constant, and is mannually selected
Dlri.o = D;,o Dlri,2:= ln(D¡. i “ onot) Setting up an array with the log of the data
Dir., = d, ¡ Dirj 3 = 15 D. 0 Converting data position to time, in minutes
x, = Dir, 3 y, = Dir, 2 Converting data into single colum arrays
*cj =xj+25 ye, Dirj+25 2 Removing the first 25 data points for first fit.
101

102
Performing least squares linear fit to the natural logarithm of the data
m := slope(xc,yc)
m=-7 121-1 o-5 Calculating the slope of the line. This Is the r-| constant
b = intercep(xc.yc) Calculating the y Intercept of the line
b =3.366 b = exp(b) b =28.963 Showing the B constant
c = corr(xc.yc)
c =-o 833625^3 Calculating the correlation factor for the fit
cal := rnxj + b Calculating the curve from data fit constants
Plot of fitting curve with log of data
In of Stress
0 500
1500 2000
1000
time, minutes
2500
3000

ecai¡ .= onot+ exp(caij) Calculation of the fit to the data
103
Plot of first curve to data
j-o..25 x2j Xj Setting up arrays
diff = Dj, - ecaij Removing the first cun/e fit component from the data
indiijf in(dif^ Taking the natural logarithm of the data
Preforming least squares linear fit to the natural logarithm of the second set of
data.
m2=-fl0ooe444494dl11 Calculating the slope of the line. This Is the constant
b2 = interceptx2,indiff Calculating the y Intercept of the line
b2 = 1.919
c =exp(b2) c =6.815 Showing the C constant
c2 - corr(x2, Indiff
c2 =-0.917
Calculating the correlation factor for the fit

cai2j = m2-x2. t b2 Calculating curve from data fit constants
104
ecal2j := exp^ca!2jj
Plot of fitting curve with log of data
Stress
MPa

calt2 = m2-Xj + b2
ecaiq = exp(caiq) Calculating fitting Curve
ecalt = ecalj +- ecalt^
Plot of fitted curve and data
105
Stress
MPa
ecalOj = ecalj - onot
Plot of the two curve components
Stress
MPa
time, minutes

APPENDIX C
CALCULATION OF CHANGE IN CURVATURE OF A BIMETALLIC STRIP BASED
ON OXIDE THICKNESS
The reduction in curvature of a bimetallic strip is calculated as a function of the
oxide thickness that forms on the surface of the aluminum. In this model the base bime¬
tallic strip is onsisting of 5pm of aluminum on 5pm of silicon. The oxide thickness
varies from 0 to 1pm in thickness with no change in the aluminum thickness. This model
was based on the equations developed by PH. Townsend [Tow87] and was implemented
in Mathcad version 4.0
layers - 3 layer = 1.. layers
delta = o.. loooo Setting up arrays
i = 1.3
j = 1..3
DeitaT = 200 Change in temperature
do 200 io 6 Set step size increase in oxide thickness
Material properties
Si Af Al203
= 5 10 '
h, delta
= 5 10
CTE, =2.6-10'6
CTEj = 25 10
CTE, = 5 5 10 7
E, = 1.910UPa
E2=71010Pa E3=5.8 10llPa
Calculation of curvature
d. - DeitaT CTE, dO-dO
“ayer layer
tTOTdella ^ lI.l + l2,l f *3,delta
¡01 I l Í
10 l I j
110 1:
[-1 1 1 oj
106

107
- V a t.
/ i layer,i i,d«
v, V*
/ .V%-della’
della
Y
Nplau-
V E t
/ i i i.dc
: Nptar^tTOr., f.TOr
Ve-, .i***.
Zj i *i,dclu 2
i
EEjWb'^
^* y E-t
Z_J J J-deha
j
y F- L
/ i i l.deha
(kS,,hddul * (lTOTdeN. ^Hi.ddjM'i.de».)2- ('T°W]
Dk*Hu ' -y >»° Calculating curvature as a percentage of maximum curvature
to6
delta 3. delta
microns

APPENDIX D
MECHANICAL PROPERTIES OF EXAMINED METALS
Alloy
Melting Point
(°C)
Young’s Modulus
(GPa)
CTE
(ppm/°C)
Pure A1
650
69
25
Al-Si-Cu
620
70
23
T210
620
70
22.7
5052
610
70
24.8
2090
610
70
23
Cu
1085
128
16.5
Ni
1453
207
13.3
Ti
1668
120
8.4
Mn
1244
120*
22
Al3Ti
1340
100
14
CiijAu
950*
CuAu
910*
15
CuAu3
925*
*-estimated
108

BIBLIOGRAPHY
Ang83
Ask89
Bak97
Bal46
Bal94
Bar94
Bas77
Bea96
BeI97
Bow74
Bro82
Cad60
Cah49
Angelí, J.B., Terry, S.C. and Barth, P.W., Scientific American 248 (4),
44(1984)
Askeland, D.R., The Science and Engineering of Materials 2nd Edition
(PWS-Kent Publishing Company, Boston, 1989)
Baker H„ Metals Handbooks (ASM International, Metals Park, Ohio,
1997)
Baldwin, W.M., Trans. AIME, 166, 599 (1946)
Baldwin, E, Holloway, P.H., Bordelon, M., Watkins, T.R., Paper
presented at the Materials Research Society Conference, San Francisco
1994
Barth, P.W., Beatty, C.C., Field, L.A., Baker, J.W. and Gordon, G.B.,
Solid-State Sensor and Actuator Workshop, Hilton Head, South Carolina,
June 13-16, p.248 (1994)
Basavaiah, S. and Greiner, J. H., J. Appl. Phys. 48 (11), 4630 (1977)
Beardsey, T., Scientific American 275 (1), 32 (1996)
Belov, V., DARPA Semiannual Progress Report on Bimetallic Actuators
for MEMS, contract number: DABT63-95-C-0038, December 1997
Bower, A.F. and Freund, L.B., J. Appl. Phys. 74 (6), 3855 (1993)
Brooks, C.R., Heat Treatment, Structure and Properties of Nonferrous
Alloys (American Society for Metals, Metals Park Ohio, 1982)
Cadoff, I. and Miller, E., Thermoelectric Materials and Devices (Reinhold
Publishing Corporation, New York, 1960)
Cahn, R.W. J. Inst. Met. LXXVI, 121 (1949)
109

Cah83
110
Cahn, R.W. and Haasen, R, Physical Metallurgy 3rd edition (North-
Holland Physics Publishing, New York, 1983)
Che96
Chen, L.Q. and Wang, Y, J. Mat. 48 (12), 13 (1996)
Cob63
Coble, R.L., J. Appl. Phys. 34 (6), 1679 (1963)
Cul78
Cullity, B.D., Elements of X-Ray Diffraction 2nd Edition (Addison-Westly
Publishing Company, Inc., Reading Massachusetts, 1978)
Dav90
Davis, J.R., Metals Handbook Volume 2, Properties and Selection:
Nonferrous Alloys and Special-Purpose Materials Iff1' Edition (American
Society for Metals International, Metals Park Ohio, 1990)
Dea52
Dean, R.S., Electrolytic Manganese and Its Alloys (The Ronald Press
Company, New York, 1952)
Deh93
Dehoff, R.T., Thermodynamics in Materials Science (McGraw Hill, Inc.,
New York, 1993)
Die84
Dieter, G.E., Mechanical Metallurgy Second Edition (McGraw-Hill
International Book Company, New York, 1984)
Eva79
Evans, R.C., An Introduction to Crystal Chemistry, Second Edition
(Cambridge University Press, New York, 1979)
Dra91
Draper, B.L. and Hill, T.A., J. Vac. Sci. Tech. B 9 (4), 1956 (1991)
Fed58
Feder, R., Mooney, M. and Nowick, A.S., Acta Met. 6, 266 (1958)
Fra97
Frank, R„ IC&S, April, 41 (1997)
Fro92
Frost, H.J., Mat. Res. Soc. Symp. Proc. 256, 3 (1992)
Fro82
Frost, HJ. and Ashby J., Deformation-Mechanism Maps (Pergamon Press,
London, 1982)
Gif70
Gifkins, R.C., Optical Microscopy of Metals (American Elsevier Publish¬
ing Company, Inc., New York, 1970)
Gol92
Goldstein, J.I., Newbury, D.E., Echlin, P, Joy, D.C., Roming, A.D. Jr.,
Lyman, C.E., Fiori, C. and Lifshin, E., Scanning Electron Microscopy and
X-Ray Microanalysis (Plenum Press, New York, 1992)
Her50
Herring, C., J. Appl. Phys. 21, 437 (1950)

Ill
Her85
M. Hershkovitz, Blech, I.A. and Komen, Y., Thin Solid Films, 130, 87
(1985)
Hog96
Hogan, H„ New Scientist, 150 (2036), 28 (1996)
H0I8O
Holloway, P.H., Advances In Eletronics and Electron Physics, 54, 241
(1980)
Hov64
Hovi, V. and Paalassalo, R, Acta Metallurgica, 12, 723 (1964)
Hud92
Hudson, J.B. Surface Science an Introduction (Butterworth-Heinemann,
Bostan, 1992)
Hum93
Hummel, R.E., Electronic Properties of Materials 2nd Edition (Springer-
Verlag, New York, 1993)
Jer94
Jerman, H., J. Micromach. Microeng. 4, 210 (1994)
Jer95
Jerman, H., DARPA Semiannual Progress Report on Bimetallic Actuators
for MEMS, contract number: DABT63-95-C-0038, December 1995
Jer96
Jerman, H„ DARPA Semiannual Progress Report on Bimetallic Actuators
for MEMS, contract number: DABT63-95-C-0038, June 1996
Kir80
Kircher, C. J. and Murakami, M., Science, +, 944 (1980)
Klu74
Klug, H.P., X-Ray Diffraction Procedures for Polycrystalline and
Amorphous Materials (Wiley International, New York, 1974)
Kno91
Knorr, D. B., Tracy, D. P. and Lu, T. M., Mat. Res. Soc. Symp., 202, 199
(1991)
K0I86
Koleshko, V.M., Relitsky, V.F. and Kiryushin, Thin Solid Films 142, 199
(1986)
Kra90
Krauss, G., Steels: Heat Treatment and Processing Principles (ASM
International, Materials Park, Ohio, 1990)
Lim97
Lim, Y, Microstruture, Phase Stability and Mechanical Properties of
NiAL Alloys Containing Zr and Si (Dissertation, University of Florida,
1997)
Lor94
Loretto, M.H., Electron Beam Analysis of Mater, Second Edition
(Chapman and Hall, New York, 1994)

Mac95
Mad97
Mas95
Mey84
Nix89
Ohr92
Pop87
Pot97
Ree92
Rie88
Row95
She96
Smi69
Sro95
Sto09
Suh86
Suh89
Tay61
112
Machlin, E.S., Materials Science in Microelectronics: The Relationship
Between Thin Film Processing and Structure (Giro Press, Croton-on-
Hudson, New York, 1995)
Madou, M, Fundamentals of Microfabrication (CRC Press LLC, New
York, 1997)
Maseeh, F., Solid State Technology, October, 50 (1995)
Meyers, M.A, and Chawla, K.K., Mechanical Metallurgy Principles and
Applications (Prentice-Hall, Inc., Englewood Cliffs, NJ, 1984)
Nix, W.D., Metal. Trans A, 20A, 2217 (1989)
Ohring, M., The Materials Science of Thin Films (Academic Press, Inc.,
Harcourt Brace Jovanovich, Publishers, New York, 1992)
Pope, D.P., Mat. Res. Soc. Symp. Proc. 81, 4 (1987)
Pottenger, M., Eyre, B., Kruglick, E. and Lin, G., Solid State Technology,
40 (9), 89(1997)
Reed-Hill, R.E. and Abbaschien, R., Physical Metallurgy Principles
(PWS-Kent Publishing Company, Boston, 1992)
Riethmüller, W. and Benecke, W., IEEE Transactions on Electron Devices
35(6), 758(1988)
Rowan, C., Laser Focus World, August 1995, 81
Shen, Y.L. and Suresh, S„ Acta Met. 44 (4), 337 (1996)
Smith, J.C., Metals Reference Book Volume I (Plenum Press, New York,
1969)
Srolovitz, D.J. and Goldiner, M.G., J. of Met., 47 (3), 32 (1995)
Stoney, G. G., Proc. Roy. Soc. London, 82, 172 (1909)
Suhir, E., J. App. Mech., 53, 657 (1986)
Suhir, E., J. App. Mech., 56, 595 (1989)
Taylor, A., X-Ray Metallography ( John Wiley and Sons, Inc., New York,
1961)

113
Ter94
Tho93
Tim25
Tou75
Tow87
Tsa92
Tur96
War69
Wea76
Win92
Wit95
You89
Tersoff, J. and LeGoues, F.K., Phy. Rev. Let. 72 (22), 3570 (1992)
Thouless, M.D., Gupta, J. and Harper, J.M.E., J. Mat. Res. 8, 1845 (1993)
Timoshenko, S., J. Opt. Soc. Am., 11, 23 (1925)
Touloukian, Y.S., Kirby, R.K., Taylor, R.E. and Desai, P.D.,
Thermophysical Properties of Matter Volume 12 (IFI / Plenum, New
York, 1975)
Townsend, P.H., Inelastic Strain in Thin Films (Dissertation, Stanford
University, Palo Alto CA, 1987)
Tsai, M.Y. and Morton, J., Mechanics of Materials 13, 117 (1992)
Turlo, J., R and D Magazine, June, 37 (1996)
Warren, B.E., X-Ray Diffraction (Addison-Westly Publishing Company,
Inc. Reading Massachusetts, 1969)
Weast, R., Handbook of Chemistry and Physics, 57th Edition (CRC Press,
Cleveland, Ohio, 1976)
Windishmann, H„ Crt. Rev. Sol. St. Mat. Sci., 17 (6), 547 (1992)
Witvrouw, A., Proost, J., DeWeerdt, B., Roussel, P.H. and Imec, K.M.,
Mat. Res.Soc. Symp.Proc. 356, 441 (1995)
Young, W.C., Roark's Formulas for Stress and Strain Sixth edition
(McGraw-Hill Book Company, 1989)
Zheng, L. P„ Li, S. Y. and Li, R. S„ J. Vac. Sci. Tech. A, 15 (4), 2431
(1997)
Zhe91

BIOGRAPHICAL SKETCH
Jonathan F. Gorrell was born on February 26, 1958, in Palo Alto California to
James and Billie Gorrell. He finished his secondary education at Monta Vista High
School in Cupertino California in 1976. He went to California Polytechnic State Univer¬
sity at San Luis Obsipo and in 1980 he received his Bachelor of Science degree in applied
mathematics. Jonathan then started his professional career as a programmer writing
assemblers for micro-processor. Over the next twelve years he worked in engineering,
customer support, marketing and sales for a number of computer related companies.
During this time he took a four year assignment in Europe. In 1991, Jonathan decided to
change careers and move into materials engineering. He was accepted into the Materials
Science and Engineering program at the University of Florida and started classes in the
fall of 1992. Two years later, Jonathan joined Dr. Paul Holloway’s group and initially
studied novel techniques for diamond growth. One year later he began examining thin
film metallization for use in micro-machines. In May of 1998 Jonathan received his
Ph.D. in Materials Science and Engineering. Upon graduation, Jonathan returned to
industry accepting a position at Motorola Semiconductor in Phoenix Arizona.
114

I certify that I have read this study and that in my opinion it is conforms to accept¬
able standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
Paul H. Holloway, Chairman
Professor of Materials
Science and Engineering
I certify that I have read this study and that in my opinion it is conforms to accept¬
able standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
Ashok Kumar
Assistant Professor of
Mechanical Engineering
I certify that I have read this study and that in my opinion it is conforms to accept¬
able standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
Professor of Materials
Science and Engineering
I certify that I have read this study and that in my opinion it is conforms to accept¬
able standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Philosophy. .
K }A
Rajiv R. Singh
Professor of Materials
Science and Engineering
I certify that I have read this study and that in my opinion it is conforms to accept¬
able standards of scholarly presentation and is fully adequate, in scope and quality, as a
dissertation for the degree of Doctor of Philosophy.
" 'J,
ate Professor of Materials
Science and Engineering

This dissertation was submitted to the Graduate Faculty of the College of Engi¬
neering and to the Graduate School and was accepted as partial fulfillment of the require¬
ments for the degree of Doctor of Philosophy
May, 199S
Winfred M. Phillips
Dean, College of Engineering
Karen A. Holbrook
Dean, Graduate School

J
.780
199 v
LD
1780
199?
.(3(973
UNIVERS^Y OF FLORIDA
ó ¿«Vi'S" ■i1""1mi '"i in mu 1111
3 1262 08554 9482



xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E41GQ6C9L_Q80D2K INGEST_TIME 2013-01-23T13:52:47Z PACKAGE AA00012979_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES