Citation
Properties and characteristicsa of LiTaO₃ for integrated-optical device applications

Material Information

Title:
Properties and characteristicsa of LiTaO₃ for integrated-optical device applications
Creator:
Maring, David Blayne, 1969-
Publication Date:
Language:
English
Physical Description:
xi, 179 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Annealing ( jstor )
Apes ( jstor )
Crystals ( jstor )
Electrodes ( jstor )
Lasers ( jstor )
Narrative devices ( jstor )
Phase diagrams ( jstor )
Protons ( jstor )
Waveguides ( jstor )
Wavelengths ( jstor )
Electrooptical devices ( fast )
Integrated optics ( fast )
Lithium tantalate ( fast )
Optical wave guides ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2000.
Bibliography:
Includes bibliographical references (leaves 170-178).
General Note:
Printout.
General Note:
Vita.
Statement of Responsibility:
by David Blayne Maring.

Record Information

Source Institution:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
45077520 ( OCLC )
ocm45077520

Downloads

This item has the following downloads:


Full Text











PROPERTIES AND CHARACTERISTICS OF LiTaO3 FOR
INTEGRATED-OPTICAL DEVICE APPLICATIONS










By

DAVID BLAYNE MARING

















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 2000



































To my friends, family, and loved ones,
for their support.















ACKNOWLEDGMENTS



I would like to express my deepest thanks to Dr. Ramu V. Ramaswamy for his guidance and leadership as my research advisor. His endless hard work to provide funding for an extremely well-equipped lab gives each of his students a unique opportunity to gain valuable experience, obtainable at very few other institutions. Additionally, his high standards for research, along with providing a very professional work environment, develops in each of his students the crucial skills necessary to communicate and compete in the business world.

I must also express much thanks to Dr. Robert Tavlykaev for his detailed guidance and assistance on much of my research. His drive for knowledge and excellence brings out the best in those around him. He is truly an asset to the Photonics Research Lab.

I would additionally like to thank Dr. Martin Uman, Dr. Gijs Bosman, Dr. Arnost Neugroschel, and Dr. David Tanner for their participation on my supervisory committee.

I recognize and appreciate many of my fellow researchers and students, particularly Dr. Sergey Kostritskii, Dr. Yuri Korkishko, Dr. Chris Hussell, Dr. Suning Xie, Dr. Weidong Wang, Dr. Sanjai Sinha, Dr. Scott Samson, Dr. Hyoun-Soo Kim, Dr. Sang-Kook Han, Dr. Hao Feng, Mark Skowronski, Al Ogden, Tetsuya Kishino, and Ryo Chinen. Many of their contributions through individual discussions and group meetings have been very valuable.



ill








I am forever indebted to my parents, Blayne and Nadine Maring, as well as my siblings, Sheri and Steve. They instilled in me from a very young age the desire to achieve excellence in every aspect of life and have supported me, both emotionally and financially, through all of my education.

Finally, I wish to acknowledge some of the many new friends I have made along the way: Juan, Carl, John, Rob, Chip, Jaime, Beth, Heather, Jen, Chad, Jodi, Mike, Kats, and all the others. Thanks for the insane times. I needed them to remain sane myself.






































iv
















TABLE OF CONTENTS




page

ACKNOWLEDGMENTS ........................................... iii

ABSTRACT ............ ....................................... ix

CHAPTERS

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

1.1 Integrated-Optical Devices and Ferroelectric Materials ................ 2
1.2 The Advantages of LiTaO3 ...................................... 4
1.3 Fabrication Processes in LiTaO3 ........... .................... 6
1.4 Motivation and Objectives .................... .................. 9
1.5 Chapter Organization ........................................ 11

2 FABRICATION OF PLANAR AND CHANNEL
WAVEGUIDES IN LiTaO3 ....................................... 15

2.1 Waveguide Formation by the APE Process ......................... 16
2.1.1 The PE/APE Process ............................. 16
2.1.2 Features of the APE Process in LiTaO3 ................... 20
2.1.3 Anomalies of APE Waveguides .......................... 23
2.1.4 APE Waveguide Fabrication Parameters ................ 26
2.2 Waveguide Formation by Metal/Vapor-Indiffusion .................. 30
2.2.1 Ti Metal-Indiffusion ................................. 31
2.2.2 Zn Vapor-Indiffusion ..................................33
2.3 Crystal Repoling ............................................34
2.4 End-Face Polishing of Channel Waveguides ....................... 36
2.5 The Bevel Polish for Planar Waveguides ....................... 37
2.6 Antireflective Coatings .............. ............ .............39
2.6.1 Thin Film Theory for Single-Layer AR Coatings ............ 39
2.6.2 The AR Film Parameters and Deposition .................. 44
2.7 Electroplating .............................................45
2.8 Summary ...........................................47



V









3 CHARACTERISTICS OF Er-DOPED LiTaO3 ........................ 49

3.1 Background .............................................50
3.2 Er-Doping Conditions ................. ..................... 52
3.3 Raman/Fluorescence Measurements ............................ 53
3.3.1 The Raman Effect and Raman Spectroscopy. ............... 53
3.3.2 Fluorescence Spectroscopy .......................... 54
3.3.3 Fluorescence Measurements ........................... 54
3.3.4 Raman Measurements .............................57
3.4 Energy Transfer Upconversion ................. ............... 62
3.5 Li-Treatment ...................................... ..........65
3.6 The Two-Photon Model of the PR Effect: LiTaO3's Advantage
in Data Storage and Lasers ........... ................... 66
3.7 Summary ........... ...................................70

4 WAVEGUIDE CHARACTERIZATION AND
MEASUREMENT TECHNIQUES ................................. 71

4.1 Near-Field Characterization ................................ 72
4.1.1 Description of the Measurement Technique ................ 73
4.1.2 Modal Characterization Results ....................... 73
4.2 Effective Mode Index Measurements .......................... 75
4.2.1 The Prism Coupling Technique .......................... 75
4.2.2 A Modified Prism Coupling Method ................... 78
4.2.3 Measured Mode Index Increments ....................... 79
4.3 Waveguide Loss ................ ........................81
4.3.1 Loss Measurement Setup ............................ 81
4.3.2 Total Insertion Loss ................................... 82
4.3.3 Propagation Loss ................................. 83
4.4 r33 Measurement......................................84
4.4.1 Interferometric Measurement Method .............. . 85
4.4.2 Unity Overlap Method ...............................85
4.5 Power Handling .............................................86
4.5.1 Setup Description .... ................ ................ 87
4.5.2 Results ........................................87
4.6 Direct Index Profiling ............ .......................... 88
4.6.1 The Importance of Direct Profiling ....................... 89
4.6.2 The Reflectivity Setup ................ .............. 91
4.6.3 Accuracy and Repeatability .......................... 95
4.6.4 Measurement Results ................................ 96
4.7 Rocking Curve Analysis ................. ................... 100
4.7.1 The Rocking Curve Measurement ....................... 100
4.7.2 Calculating Lattice Strain ............................. 102
4.7.3 Some Rocking Curve Results .......................... 102
4.8 Summary ............. ...............................103



vi









5 STUCTURAL PHASE DIAGRAM FOR APE:LiTaO3 ................. 105

5.1 Previous Limitations .................................... 107
5.2 Constructing the Structural Phase Diagram ...................... 108
5.3 Analyzing the Structural Phase Diagram ............... ........ . 110
5.3.1 Explaining APE Anomalies ............... ........... . 110
5.3.2 Effect of Crystal Phases on Index Profiles ................ 110
5.4 PE Waveguides ............................................ 113
5.4.1 Raman Analysis of PE Waveguides ...................... 114
5.4.2 Modified Structural Phase Diagram ................... 116
5.5 Summary ............. .............................118

6 STABILITY OF LiTaO3 WAVEGUIDES ............................ 120

6.1 Rocking Curve Analysis of Stability......................... 121
6.2 Using a Directional Coupler to Measure Stability .................. 123
6.2.1 Directional Coupler Theory ......................... 123
6.2.2 Measuring Coupling Length and Index Change ............ 124
6.3 APE and PE Stability Results .................................. 126
6.3.1 APE Results .............. ..................... 126
6.3.2 The Advantage of PE ................................. 129
6.3.3 PE Results .......................... ... ...........131
6.3.4 Limits to Reducing Strain ............................. 135
6.4 Zn/Ti-Indiffused Stability Results ............................ 137
6.4.1 Directional Coupler Stability Results
of Zn:LiTaO3 Waveguides ............................ 138
6.4.2 Comparison of the Waveguide Formation Processes......... 138
6.5 Examining the Crystal Growth Issue ........................... 140
6.5.1 Comparing Different Growers .......................... 141
6.5.2 Accounting for the Li Deficiency ....................... 142
6.6 Summary ................................................... 143

7 EXPLORING THE POTENTIAL OF LiTaO3:
A TRAVELING-WAVE ELECTRO-OPTIC MODULATOR ............. 145

7.1 Fabrication Conditions ................................... 146
7.2 Microwave Characteristics .................................... 147
7.2.1 S-Parameters and Loss............................... 148
7.2.2 Microwave Effective Index ......................... 149
7.2.3 Characteristic Impedance ............................ 149
7.3 DC Response and Power Handling ........................... 150
7.3.1 DC Response ................................... 150
7.3.2 Power Handling Near 1.5ptm........................ 152
7.4 Frequency Response .......................... ............... 154
7.5 Summary ............................................156



vii









8 CONCLUSIONS AND FUTURE WORK ........................... 158

8.1 Conclusions .......................................... 158
8.2 Future Work ................ ........................... 161
8.2.1 Modulator Improvements ............................. 162
8.2.2 Integrated Laser-Modulator Module ..................... 164
8.2.3 Photorefractive Grating ........................... 165

APPENDICES

A DISPERSION OF RUTILE ................................. 166

B DISPERSION OF LITHIUM TANTALATE ............... ...... 168

REFERENCES ..................... ............................... 170

BIOGRAPHICAL SKETCH ......................... .............. 179




































Viii















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

PROPERTIES AND CHARACTERISTICS OF LiTaO3 FOR
INTEGRATED-OPTICAL DEVICE APPLICATIONS By

David Blayne Maring

May 2000


Chairman: Ramu V. Ramaswamy
Major Department: Electrical and Computer Engineering

Continued and rapid growth expounded by the bandwidth explosion in the area of dense-wavelength-division-multiplexed (DWDM) telecommunication and information networks, like the internet, has been and continues to be met by the global deployment of optical fiber. In turn, this has resulted in the unprecedented demand for the integratedoptical components used in these fiber networks. Of particular interest is devices fabricated in ferroelectrics, most notably lithium niobate (LiNbO3) and lithium tantalate (LiTaO3). Although current demands are being met by LiNbO3, its brethren LiTaO3 exhibits a much higher threshold for optical damage, providing the opportunity to fabricate devices, such as lasers and modulators, with much higher output and throughput levels than currently obtainable with LiNbO3. In addition, LiTaO3 has a shorter wavelength ultra-violet (UV) absorption edge (280nm) than LiNbO3 (350nm), permitting nonlinear conversion by frequency doubling to shorter wavelengths with less absorption.


ix








A number of obstacles, however, prevent the large-scale development of optical devices in LiTaO3. The various processes of waveguide formation have not been completely investigated or characterized. For instance, the annealed proton exchange (APE) technique is known to exhibit a number of anomalies which to date have not been understood, the most significant being the temporal instability of the waveguide index increment. Additionally, no work has been performed to characterize the impact of Erindiffusion, necessary for the fabrication of 1.5gm lasers.

In this dissertation, the above issues are addressed. Conditions for the indiffusion of Er in LiTaO3 are identified. Fluorescence and Raman spectroscopy are used to identify energy transfer upconversion (ETU) between neighboring clusters of Er3+ ions as the dominant mechanism of upconversion, leading to increased photorefractivity, a gain-limiting factor for lasers. Upconversion is then decreased through a novel Li-treatment process to reduce clustering, with results showing a significant advantage over LiNbO3. Additionally, the structural phase diagram for APE:LiTaO3 waveguides is constructed and used to explain the previously observed anomalies associated with this process in LiTaO3. The stability of waveguides fabricated by various techniques is also examined, resulting in the determination that instability does not depend on fabrication conditions, as it does in LiNbO3, but rather it is inherent to the crystal, the likely result of an immature growth process. Finally, to demonstrate the potential of LiTaO3 and indicate the desire for higher quality crystals, a high-speed traveling-wave modulator is fabricated and its measured results compared to theory. This work hopefully will inspire crystal growth companies to arrive at a growth process which produces crystals with better crystal stoichiometry, desirable for improved device performance.



X















CHAPTER 1
INTRODUCTION



A discussion leading to the motivation for this dissertation is presented in this chapter. Ever increasing demand for bandwidth in telephone, computer network, and internet applications has resulted in the ever increasing demand worldwide for fiber networks, and explosive growth in the associated integrated-optical devices for these networks such as lasers, filters, and high-speed modulators. The demands imposed by these devices are considered before presenting possible materials to be used in their fabrication. Among the suitable materials, consideration of their properties reveal that lithium tantalate (LiTaO3) possesses a number of distinct advantages, particularly a higher optical damage threshold and a shorter wavelength ultra-violet (UV) absorption edge.

Use of LiTaO3 as a host for the fabrication of integrated-optical devices faces a number obstacles. For instance, the characteristics of Er-indiffusion for laser fabrication have not yet been studied in this material. Additionally, much work needs to be done on the characterization of different waveguide fabrication techniques. The most common technique, annealed proton exchange, is known to exhibit a number of anomalies preventing it from being used in the production of reliable devices, the most significant anomaly being the instability of waveguide index profiles over time. As for the other methods of waveguide formation, no data characterizing relevant issues associated with each process, such as stability, has been reported.






2

In this dissertation, the process of Er-doping in LiTaO3 is characterized in order to examine its impact on the optical and physical properties of the material, and gauge its usefulness for the development of high-power infrared lasers. Additionally, the different processes of waveguide formation are characterized with an aim towards determining the feasibility of identifying fabrication conditions that will produce stable waveguides, and hence reliable devices, in LiTaO3.



1.1 Integrated-Optical Devices and Ferroelectric Materials


Demand for broadband information and communication services for telephone, cable television (CATV), and computer network applications, such as the internet, continues to increase significantly year by year. As illustrated in Fig. 1.1, for the telecommunication industry alone, this trend toward rapidly increasing bandwidth demand is expected



25 Total:
Data and Voic
20




10


5

Voice


1990 1995 2000 2005
year

Fig. 1.1 Relative trend in the demand for telecommunication network bandwidth.






3

to nearly quadruple in the next five years. Because of the large bandwidth it offers, optical fiber has been employed as the transmission medium to meet this demand, operating at wavelengths around 1.5gm because of its decreased loss and low dispersion at this wavelength. In addition, wavelength division multiplexing (WDM) in highly-dense networks is often employed to take further advantage of the fiber's generous bandwidth.

It follows that a large surge in the production of integrated-optical devices for use in broad-band fiber-optic networks, as well as other commercial and military applications, is clearly evident. Some of the most promising devices include the following: travelingwave modulators, high-output 1.5gm lasers, and high-speed optical switches for use in fiber-optic communications systems [Alf84], wavelength-selective tunable filters for use in WDM systems [Hus95], optical parametric oscillators (OPO) for generation of coherent, long wavelength (2.5-4gm) radiation for use in spectroscopy and laser radar, non-linear frequency doublers for blue-light generation needed in optical storage applications [Ahl94a, Ah194c, Nak90], and high-frequency, broad-band surface acoustic wave (SAW) filters for use in optical spectrum analyzers and television applications [Xu91].

In order to fabricate these devices, we must find a suitable material whose characteristics are such that it is able to meet the demands imposed by each device. In particular, it must be transparent in the wavelength regions of interest, for both difference frequency generation of infrared (as in OPOs) and in the visible, for blue-light generation by nonlinear frequency doubling. It should be capable of producing waveguides with low propagation loss and must possess a large electro-optic coefficient, desirable in the fabrication of electro-optic modulators and switches. It must also possess a large nonlinear coefficient and be capable of satisfying the phase matching condition so that difference frequency






4

generation and second-harmonic generation (SHG) of blue light are possible. It should have low acoustic losses to allow the fabrication of SAW devices such as filters. It must be able to serve as a host for the indiffusion of rare-earth materials such as Er, which produces lasing emissions near 1.5[m when optically pumped. Finally, it must be commercially available with high optical quality and at low cost.

The ferroelectric materials lithium niobate (LiNbO3), lithium tantalate (LiTaO3), and potassium titanyl phosphate (KTP) are three of the most promising candidates which satisfy the above conditions and have received widespread use in the fabrication of the aforementioned devices [Ahl94a, Ahl94c, E1H95, Miz92, Nak90, Xu91]. In the next section, we will compare these crystals and demonstrate why LiTaO3 is the favored choice for the fabrication of most devices.



1.2 The Advantages of LiTaO3


Historically, considerable effort has gone into the fabrication, characterization, and modeling of channel waveguides in LiNbO3 [Cao92, Paz94, Yam91 a]. On the other hand, LiTaO3, which is isomorphous to LiNbO3, has received much less attention although it has been shown to possess many advantages over LiNbO3. LiTaO3 is harder and demonstrates a photorefractive damage threshold of 1500W/cm2 at 514.5nm and room temperature, about forty times higher than the 40W/cm2 value of LiNbO3 [Gla72, Mat92b, Miz91, Miz92, Saw91, Spi83]. This improved photorefractive resistance should provide better power handling capabilities of devices fabricated in LiTaO3 [Bur93, McW92, Miz91], making it advantageous for the development of high-throughput modulators and filters, as well as high-output 1.5utm sources with output power levels far exceeding those attainable






5

in LiNbO3. In addition, LiTaO3 has a shorter wavelength UV absorption edge (280nm) than LiNbO3 (350nm), permitting nonlinear conversion by frequency doubling to shorter wavelengths with less absorption.

Though LiTaO3 has a large nonlinear coefficient d33 and an electro-optic coefficient r33 as large as that in LiNbO3 [Miz92, Yuh92], it exhibits a low birefringence, making it unattractive for birefringent phase matching [Ahl91, Mat92b]. However, nonlinear applications, such as second harmonic generation (SHG) of blue light, can be achieved in LiTaO3 with the aid of quasi-phase matching (QPM). Hence, LiTaO3's higher optical damage threshold and low propagation loss make it an ideal material for QPM-SHG of blue light, capable of output power levels far exceeding those of LiNbO3 [Ahl91, E1H95, Yam9lb, Yam92].

The more mature growth technology of LiTaO3 allows the use of larger size crystals at a more reasonable cost than is possible with KTP. In addition, KTP's much higher conductivity may result in large leakage currents, complicating the application of DC bias voltages which are essential in many electro-optic devices.

It is clear, therefore, that there is an established advantage to using LiTaO3 for the fabrication of integrated-optical devices. However, significant research still remains to be done, especially on the characterization and modeling of waveguides formed in LiTaO3. In the next section, some of the techniques used for the fabrication of waveguides in LiTaO3 will be presented and a few of the currently unresolved issues relating to this subject will be discussed, leading to the motivation and focus for this work.






6

1.3 Fabrication Processes in LiTaO3


Several methods exist for the formation of waveguides in ferroelectric materials. For LiTaO3 in particular, there are at least three recognized procedures. They are Ti metalindiffusion, Zn vapor-indiffusion, and annealed proton exchange (APE). Additionally, indiffusion of rare-earth materials, such as Nd and Er, have been examined for the production of laser sources. Though these techniques are all described in greater detail in subsequent chapters, they are introduced here briefly to point out some of the limitations and anomalies associated with each and therefore develop the motivation for this work.

In order to create an active gain region for the facilitation of a laser source or waveguide amplifier in LiTaO3, a rare-earth material such as Nd or Er needs to be indiffused. Nd, when doped in a host material and optically pumped, is known to produce lasing emissions near lgm [Nou95, San92]. For communication applications, however, it is desirable to have emissions near 1.51tm, to take advantage of the decreased loss and low dispersion of optical fiber at this wavelength. Er-doped material when optically pumped is known to produce lasing emissions near 1.5tpm [Gil96]. Though previous work has concentrated on the characterization of Er-indiffusion in LiNbO3 for the fabrication of waveguide amplifiers and lasers [Ami96, Bau96, Hua96], little or no work has been reported on the indiffusion of Er into LiTaO3. As a result, further research is needed to determine the conditions for indiffusion of Er into LiTaO3, identify optimum levels of Er3+ concentration, investigate the effects of Er-doping on the optical properties of the crystal, and explore the various methods for waveguide fabrication after Er-doping to determine the most ideal technique in terms of laser and amplifier performance.






7

Of the three methods for waveguide fabrication outlined above, APE is presently the most common. Compared to the other techniques, APE offers the advantages of simplicity, flexibility, high index increment, and low propagation loss. In addition, APE can be performed at temperatures below the Curie point of the crystal (610'C), assuring that this technique does not disrupt the monodomain structure of the crystal and so no postfabrication repoling process is needed.

However, a number of anomalies are known to exist for APE waveguides in LiTaO3. For instance, it has been shown that the peak extraordinary index increment increases during the initial stages of annealing [Ahl94c, Mat92b] and that the index decreases with H proton concentration as proton concentration exceeds a certain point. These facts suggest a nonlinear dependence of index increment on proton concentration and the possibility to form buried refractive index profiles [Ahl95, Dav95, Mar96a]. Such a buried profile has recently been directly observed [Mar96b] and is exhibited in Fig. 1.2.

Though buried index waveguides

0.025 , can provide the advantages of

0.02 - - improved fiber-to-device coupling

0.015 15 min and reduced scattering losses at the

0.01 _ substrate-air interface, they seri0.005 7.5 min
0.005 ously undermine the adequacy of

0 0.2 0.4 0.6 0.8 1 1.2 1.4 previous theoretical studies, most
0 0.2 0.4 0.6 0.8 1 1.2 1.4
depth (pm) of which relied on the indirect proFig. 1.2 Refractive-index profiles of extraordi- file reconstruction technique of nary polarization for various annealing times measured at X=0.633utm. inverse WKB and the assumption






8

of a monotonically decreasing
0.025
index profile. Most important from
0.02
a practical standpoint, APE
0.015 immediately after
o fabrication waveguides in LiTaO3 have been
0.01 4 weeks after
fabrication shown to exhibit temporal instabili0.005
ties [Ah194c, Mar96b, Mar00],

0 0.5 1 1.5 2 2.5 manifesting themselves in the form depth (pm)
of significant index profile modifiFig. 1.3 Temporal changes in the index
profile (extraordinary polarization) cations for both short and long
measured at X=0.633gm
immediately after fabrication and 4 periods after fabrication. An
weeks later.
example of this is depicted in Fig.

1.3, showing the extraordinary index profile of a waveguide measured immediately after

fabrication, and again four weeks later. These instabilities are believed by some to be due

to the presence of multiple crystal phases within the proton exchanged region [E1H95,

Kor96], though it has not been directly confirmed. This fact prevents LiTaO3 from being

used in the fabrication of practically viable, that is, temporally stable waveguides for

device applications. As a further point, it is known that the introduction of H+, through

proton exchange, into a rare-earth doped region leads to significant quenching of laser

action, or a reduction in excited state lifetime, thereby reducing laser efficiency. As such,

additional methods for waveguide fabrication need be explored.

As stated earlier, the other two recognized techniques of waveguide formation in

LiTaO3 are Ti metal-indiffusion and Zn vapor-indiffusion. While a small amount of work

has been reported in the literature about the success of these techniques, they have not






9

been as seriously investigated as has APE. While it is likely that these techniques may be better suited for the production of waveguides in Er-doped regions because they won't contribute to H+-induced quenching, they have the drawback of needing to be performed at temperatures above the Curie point, thus requiring the additional step of crystal repoling to return the sample to a monodomain state. Additionally, no investigations have been reported as to the temporal stability of waveguides fabricated using these techniques.

From the above discussion, it is clear that additional research still remains to be performed on the characterization of the different waveguide formation processes used in LiTaO3, as well as the characterization of Er-doping and its effect on the crystal's optical properties, before reliable, high-throughput integrated-optic devices can be fabricated in it. This need for continued research formulates the motivation for this work which is outlined in the next section, along with a short outline of the steps to be taken in order to complete this work.



1.4 Motivation and Objectives


As stated earlier, LiTaO3 offers the advantages of high power handling capability and a shorter wavelength UV absorption edge. However, the feasibility of fabricating lowloss, temporally stable waveguides for use in reliable, high-throughput integrated-optical devices such as modulators, filters, and lasers has not yet been demonstrated. To this end, considerable research needs to be performed to characterize both the process of Er-indiffusion and the various methods of waveguide fabrication in LiTaO3. This forms the basis of the motivation for this work, with a specific interest in determining whether fabrication conditions exist which will produce temporally stable waveguides.






10

To begin, an extensive study on the conditions of Er-doping needs to be performed. Fabrication conditions such as Er thickness before indiffusion and temperature and time of indiffusion need to be determined. Er3+ concentration should be at an appropriate level so as to provide significant gain and Er3+ concentration profiles should be at least as deep as waveguide index profiles so that the optical mode has good overlap with the gain medium. The effect of Er-doping on the physical and optical properties of the crystal need also be examined; in particular: how doping effects the photorefractive property of the crystal as well as how much upconverted emission from Er3+ clusters, which reduce gain [Gil96], is present and how to reduce it.

Next, a complete characterization of the APE process of waveguide formation is desired to facilitate the understanding of the associated anomalies and identify fabrication conditions leading to temporally stable waveguides. The aforementioned anomalies have pointed to a nonlinear dependence of refractive index on proton concentration. As such, the structural phase diagram for APE:LiTaO3, relating waveguide surface index increment to proton-induced lattice strain at the surface, needs to be constructed in order to explain these anomalies. Because instability is speculated to be due to the presence of multiple phases within the proton exchanged region, the compilation of this diagram would also aid in the identification of fabrication conditions yielding lower concentration, possibly single-phase waveguides which are believed to be more stable. Once fabricated, the stability of these waveguides needs to be quantified and monitored over time.

Finally, characterization of waveguides formed by Ti metal-indiffusion and Zn vapor-indiffusion needs to be performed. Fabrication conditions leading to the formation of single-mode waveguides at 1.5gm need to be identified. Since these processes cause no








additional lattice strain, as does APE, there is no structural phase diagram to be constructed. However, the stability of waveguides formed by each process needs also to be quantified and monitored over time, so that it may be compared to that of APE waveguides.



1.5 Chapter Organization


In Chapter 2, the various methods of waveguide fabrication used in this research are discussed. A description of the APE process is provided followed by some of the features and anomalies of this process in LiTaO3. Then the actual conditions used for the fabrication of APE waveguides are given. Next, a description of the waveguide formation processes of Ti-indiffusion and Zn vapor-indiffusion is presented. Again, some of the features associated with each process are provided along with the actual fabrication conditions used. The rest of the chapter details some additional processing steps involved in waveguide and device fabrication such as crystal repoling, end-face polishing, the application of anti-reflectivity coatings, and electroplating of thick gold electrodes.

Chapter 3 examines the process of rare-earth Er-doping in LiTaO3, a first step in the fabrication of a 1.55ptm active waveguide laser. Some background information related to the process of Er-doping in LiNbO3 is presented. Then the limitations associated with the use of LiNbO3 are presented, opening the door for the use of LiTaO3 as a host material for Er-doping. Next, the actual conditions used for the doping of LiTaO3 with Er are outlined, after which, characterization is performed by means of Raman spectroscopy and fluorescence measurements to examine the effect of Er-doping on the photorefractive






12

property of the crystal. Lastly, a two-photon model of the photorefractive effect is used to explain how upconverted light caused by Er-doping enhances photorefractivity.

The issue of waveguides in LiTaO3 is returned to in Chapter 4, where waveguide characterization and measurement techniques are discussed and applied. First, a near-field measurement technique is used to examine the region of single-mode operation, in terms of waveguide channel width, for waveguides designed to be single-mode at 1.55gm. Next, a modified prism coupling technique, which exhibits improved accuracy over the conventional technique, is introduced and applied to measure the mode effective indices for waveguides at several different wavelengths. Waveguide loss is examined next, where measurements of total insertion loss and waveguide propagation loss are made, the latter by a Fabry-Perot technique. Then, the electro-optic coefficient r33 is measured directly using a unity-overlap technique which does not require the estimation of an electrical-optical field overlap factor. To demonstrate the superior photorefractive resistance of LiTaO3 waveguides near 1.5gm, plots of output power versus input power are measured next and compared to similar structures in LiNbO3. Finally, the methods of direct index profiling and X-ray diffraction rocking curve analysis are explained and applied for the direct determination of surface index increment An and proton-induced lattice strain _33 at the surface of APE waveguides. These figures are necessary for the construction of the structural phase diagram for APE waveguides performed in Chapter 5.

In order to explain the anomalies associated with the APE process in LiTaO3, the structural phase diagram for this material needs to be constructed. This is done in Chapter 5 where several waveguides with different values of surface concentration are fabricated by APE and measured using the techniques of direct index profiling and rocking curve






13

analysis. Resulting values of surface index increment An and lattice strain 33 are plotted to construct the structural phase diagram. The diagram is then used to explain the previously observed anomalies associated with the APE process in LiTaO3 and the effect of crystal phases on the shape of the waveguide index profile is also examined, with examples of measured profiles from different regions of the structural phase diagram. Lastly, proton exchanged (PE) waveguides fabricated in a dilute source are examined, exhibiting advantages over APE waveguides. These waveguides are characterized by the same direct measurement techniques used on APE waveguides and by Raman spectroscopy to determine the presence of an additional single-phase region of the structural phase diagram obtainable only by PE in a dilute source.

Chapter 6 examines the stability issue of waveguides in LiTaO3. APE waveguides are inspected by rocking curve analysis to track changes in lattice strain F33 over time. For a more quantitative analysis, directional couplers are fabricated from different regions of the phase diagram and changes in measured coupling length are related to corresponding changes in index increment 8(An) through computer simulation of a directional coupler. Results indicated that stability is not caused by the presence of multiple phases, as previously assumed. To determine if stability depends on fabrication conditions, the stability of Ti-indiffused and Zn vapor-indiffused are also measured, showing no improvement in stability, unlike LiNbO3. Suspecting inferior growth technology to be the reason for the instability, crystals from several different growers worldwide, both SAW and optical grade, are obtained and compared in terms of stability only to find that all are similar and unstable.






14

Though critical advancements are needed in the growth technology for LiTaO3, a demand for high-quality optical grade crystals must exist before time and money will be spent on such an effort. To this end, the potential of LiTaO3 for integrated-optical components is demonstrated in Chapter 7 with the design, fabrication, and testing of a symmetric Mach-Zehnder interferometer (MZI) traveling-wave electro-optic modulator. First, the fabrication conditions for the device are provided. Next, some microwave characteristics of the device are examined, including measurements of the microwave loss coefficient, microwave effective index, and characteristic impedance. Then, the DC response of the device is measured in order to obtain the extinction ratio and half-wave voltage. The power handling capability of the modulator is examined next by plotting output power versus input power near 1.5pim. Additionally, the frequency response of the device is measured and compared to theory.

Finally, Chapter 8 presents the conclusions for this dissertation and outlines some future work. A summary of the research performed is stated and the important contributions resulting from this work are emphasized. Some future work to be performed in this area, some of it currently under way, is then outlined.















CHAPTER 2
FABRICATION OF PLANAR AND CHANNEL WAVEGUIDES IN LiTaO3



Planar and channel waveguides are formed by creating an area of higher refractive index in a dielectric material such as LiTaO3. Unlike planar waveguides, channel waveguides confine the optical wave in two dimensions. Most processes result in a roughly semi-elliptical, graded-index profile that decreases monotonically from the substrate surface [Ram88]. Such a guide is schematically represented in Fig. 2.1, where the propagation direction of the optical wave is also shown. These waveguides form the basis for a multitude of devices in the integrated-optic embodiment, in particular, amplitude modulators used in communication systems, high-speed optical switches, non-linear frequency doublers, and optical parametric oscillators.

propagation
direction


high index
channel









substrate

Fig. 2.1 Depiction of a channel waveguide.


15






16

This chapter describes the methods used in the creation of optical waveguides in dielectric substrates. Particular attention is paid to the proton exchange and annealed proton exchange techniques, as well as to the methods of Ti metal-indiffusion and Zn vaporindiffusion, used to fabricate waveguides in LiTaO3. Several distinct features and anomalies attributed to each of the processes are discussed. Specific fabrication parameters are given in order to cover completely the fabrication methods used in this research, and a few additional processing steps are also described.



2.1 Waveguide Formation by the APE Process Waveguides in LiNbO3 and LiTaO3 can be formed by using several methods, the most popular of which are by metal indiffusion at high temperatures, and by the exchange of Li+ ions in the substrate by H+ ions supplied by an acid source -- proton exchange or annealed proton exchange method [Ram88, Sym92]. In this section, the annealed proton exchange method of waveguide fabrication is discussed. The features of waveguides fabricated by this process in LiTaO3 are then described, followed by some of the associated anomalies observed. Finally, the actual fabrication conditions used to produce waveguides by this technique are given.


2.1.1 The PE/APE Process

As discovered by Jackel, Rice, and Veselka in 1982, the proton exchange (PE) process results in a large increase of the extraordinary refractive index of LiNbO3 [Jac82]. The technique was later shown to also be effective in forming waveguides in LiTaO3 by Spillman, Sanford and Soref [Spi83]. Though the PE process initially causes a reduction






17

of the electro-optic effect, the r33 coefficient can be almost completely recovered through post-exchange annealing of the substrate [Ahl94c, Yuh92]. This entire process of ion exchange followed by annealing is referred to as the annealed proton exchange (APE) technique. Annealing causes further diffusion of the H+ ions into the substrate, expanding the area of the waveguide and decreasing the propagation loss [Kan94]. In the case of LiTaO3, its extraordinary index increment Ane after proton exchange has been shown to be lower than that of LiNbO3. However, annealing causes an anomalous index increase in LiTaO3, compared to an anticipated decrease for LiNbO3 [Ahl94c, Mat92b, Yuh92]. This distinct feature of APE in LiTaO3, along with others, will be more thoroughly discussed in the Section 2.1.2.

The APE technique is among the most suitable for not only the fabrication of channel waveguides in LiTaO3, but also for the formation of domain-inverted regions [Ahl94b]. These areas of inverted domain are necessary in the fabrication of integrated devices intended for nonlinear optical applications, like blue light generation by frequency domain inverted


waveguide











Fig. 2.2 Depiction of an APE channel waveguide laid across regions of domain inversion, also formed by APE.








doubling of infrared, where quasi-phase matching (QPM) is required [Ah194c, Nak90]. For this case, the APE waveguides and domain inverted regions are fabricated on the c surface of LiTaO3 (as opposed to the c+ surface for LiNbO3) since domain inversion seems to be initiated on the c- surface [Nak90, Saw91]. When using patterns of domain inversion, such as in QPM, the waveguides are oriented perpendicular to the existing inverted regions, as in Fig. 2.2. Since proton exchange and post-exchange annealing for waveguide fabrication can be done at temperatures well below 4500C, they do not affect the existing inverted regions which have also been created by APE, but annealed at temperatures near the Curie temperature of about 6000C [Fin88, Miz94].

The exchange setup of the APE process is depicted in Fig. 2.3 for channel waveguides. The substrate is first covered by a metal masking layer (usually Al or Ta), in













....... .metal mask










Fig. 2.3 Schematic of the proton exchange process.
. . . ...-.'...-. ,'." . . . ". ... . .
. . . ..'.' "..' ', .'. . . . .. . . ',." . ."
. . . . . ...'. .".�. .'' " .'. . " . .'.' . "







. . . . . . . ..." ,' ' '. . ." . . . . . . . .",
� � � . . . . . . . ..,.'. ... '.' ' . � � � � . "






19

which narrow stripe patterns are delineated by photolithography, exposing regions of the substrate where the exchange is to take place. In the case of planar waveguides, no mask is needed and the entire bulk surface is subjected to proton exchange. The substrate is then immersed in a hot melt of high H+ concentration, usually benzoic or pyrophosphoric acid, where Li+ diffuses out of the substrate and is replaced by H+ from the acid source.

As a result, an exchanged layer of HxLil_xTaO3 (HxLil_-xNbO3 if LiNbO3 is used as the substrate) is formed at the surface with the thickness dependent on the melt temperature T and exchange time t. On the other hand, the refractive index change is practically independent of T and t, being determined largely by the choice of the proton source (acid). Pyrophosphoric acid is normally preferred as the H+ source since it has a larger diffusion coefficient D(t) and results in an increase in the extraordinary index Ane which is 15% higher than what is attainable with benzoic acid [Ahl94c, Kan94]. In addition, pyrophosphoric acid does not boil at typical exchange temperatures (it is liquid up to 3000C) and has a low vapor pressure, allowing work at high temperatures where diffusion is rapid. The higher H+ concentration also allows for more control over the diffusion depth and uniformity of the exchanged layer than is possible with benzoic acid [Miz92]. However, when using pyrophosphoric acid, Ta must be used as the mask since the pyrophosphoric acid will attack and cause pitting in an Al mask.

The APE process in LiTaO3 results in an increase in the extraordinary index only, allowing for the propagation of solely TM modes in a z-cut crystal, as illustrated in Fig. 2.4. For the case of x and y-cut crystals, only TE modes will propagate. Additional features characteristic of the APE process in LiTaO3 are described in the next section.






20

ne TM no no


no








TM


Sz-cut substrate

- TE

Fig. 2.4 Increase of ne only due to proton exchange causes the propagation of solely TM modes in a z-cut crystal.




2.1.2 Features of the APE Process in LiTaO3

As stated earlier, the APE process of waveguide fabrication holds many advantages over other methods, such as simplicity and high index increment. Specifically, in comparison with the method of Ti or Zn- indiffusion, the APE process supports only one polarization and can be performed at much lower temperatures, below the Curie point where no crystal repoling is necessary. Additionally, the waveguide loss is low [Mat92a], and it is less susceptible to photorefractive damage [Mat92a, Mat92b]. We describe here many of the characteristics and peculiarities of the APE process in LiTaO3, making comparisons to the well characterized APE process in LiNbO3 whenever applicable.

The proton exchange process is known to cause an increase in the extraordinary index ne, and a decrease in the ordinary index no [Ahl94c, Ahl95]. This is advantageous






21

in applications where a single polarization is desirable, since only light polarized in the direction of ne will be propagated.

In APE:LiTaO3, there are at least four mechanisms contributing to the total index change An [Ahl95], which, thus, can be expressed as An = AnP + AnM + AnV + Ane (2-1) The AnP term is caused by a change in the spontaneous polarization Ps via the Kerr electro-optic effect. After substitution of H+ for Li+ during proton exchange, H+ is known to occupy positions within the oxygen planes rather than the Li+ sites, reducing the contribution of Li+ to Ps. The AnM term is due to the difference in the molar polarization of Li20 and H20 brought about by the substitution of one for the other in proton exchange. The Anv term is caused by the change in molar volume due to an expansion of the unit cell upon exchanging. Finally, AnE, due to exchange-induced stress, contributes via the elastooptic effect. Previous studies [Fed94] have observed that H:LiTaO3 waveguides on Z and X-cut substrates have only one non-zero component -33 of the deformation tenser. The An" term is proportional to the deformation value 33.

The depth de of the refractive index change after proton exchange is related to the exchange time te by [Ahl94c, Dav95]


de = 2 De(Te)te (2-2) where De(Te) is the temperature dependent diffusion coefficient for the proton exchange process and Te is the exchange temperature. This diffusion coefficient follows the Arrhenius law [Ah194c, Dav95],


De(T) = e0exp (2-3)






22

where Ee is the activation energy for the proton exchange process, Deo is the exchange diffusion constant, and kB is Boltzmann's constant. Similarly, for the annealing process at temperature Ta and duration ta, the following annealing diffusion coefficient Da(Ta) and increase in depth da are found [Dav95]


da = 2 Da(T)ta (2-4)


-E
Da(Ta) = D aex k (2-5)


where Ea is the activation energy for the annealing process. The total depth dT of the waveguide fabricated by the APE process can then be approximated by [Dav95] dT = de + da (2-6) The profile of the index change in the proton exchanged region is known to be step-like [Ahl94c, Kan94]. The APE process in LiTaO3 is known to result in an increase of the extraordinary index and a decrease of the ordinary index. Therefore, only TM modes will be supported in a Z-cut crystal. In the case of X and Y-cut crystals, only TE modes propagate. Upon increasing the relative proton concentration x, the crystal structure of the exchanged layer may transform into another crystal phase as has been indicated by rocking curve measurements of exchanged samples [Fed94]. These transformations are reversible in the sense that the initial phase can be restored upon post-exchange annealing. Post-exchange annealing at elevated temperatures (higher than that of exchange) then causes further diffusion of H+, smoothing out the index profile and expanding the waveguide area while decreasing the propagation loss [Miz92]. Since there is no proton source at the surface, the local concentration within the exchanged layer may decrease to






23

the value corresponding to the phase transition. Annealing also serves to recover the electro-optic r33 and nonlinear d33 coefficients within the exchanged region, known to have been seriously degraded during the exchange process [Ahl93, Ahl94a, Ahl94c, Cao92, Mat92b].

Every phase transition results in a change in the refractive index. The value of proton concentration at which a phase transition occurs is determined by the specific material, as is the dependence of refractive index on proton concentration between the phase transitions. In other words, the dependence of refractive index on proton concentration is specific to the material and is identified as the structural phase diagram of this material. As will be shown below, knowledge of the structural phase diagram is crucial in understanding the properties and anomalies of APE waveguides.


2.1.3 Anomalies of APE Waveguides

Many anomalies exist which are characteristic of the APE process in LiTaO3. Though some research on these peculiarities has been done, they are not yet fully understood. First, the extraordinary index increase Ane that can be achieved by APE in LiTaO3 is as much as an order of magnitude less than that in LiNbO3 [Ahl94c, Mat92b, Yuh92]. It has been suggested that this difference is the result of a larger value of spontaneous polarization in LiNbO3 [Ahl95], which is related to index increment through Eq. (2-1). As for the magnitude of the ordinary index decrease Ano in APE LiTaO3, it has been found to be many times larger than the extraordinary index increase Ane [Ahl94c, Mar96b], as opposed to being of nearly the same value or smaller for the case of APE in LiNbO3 [Ahl95].






24

Second, is the anomalous behavior of waveguide profiles upon annealing. It has been demonstrated that the refractive index increment increases during short annealing times for the APE process in LiTaO3, vastly increasing the area under the index profile [Ahl94c, Mar96b, Mat92b, Yuh92]. In addition, studies performed by varying the H+ concentration in the exchange source indicate that above a certain concentration, index increment actually decreases with increasing H+ concentration [Ah194c]. These peculiarities suggest a nonlinear dependence of the refractive index increment on H+ concentration and the possibility of forming waveguides with buried refractive index profiles [A'h194c, Ahl95, Dav95, Mar96a]. Such profiles have recently been confirmed by direct measurement [Mar96b], using a reflectivity profiling technique to be described in Chapter 4, and are exemplified in Fig. 2.5. This is in direct opposition to the case of APE in LiNbO3, where the index increment is known to decrease upon annealing, maintaining a more or less linear dependence on H+ concentration [McW91].
0.025
The third, and most signifi0.02
cant peculiarity, is that APE
0.015 15 min
=o _waveguides in LiTaO3 have been
0.01
found to exhibit significant tempo0.005 7.5 min
0.005
ral instabilities, appearing in the
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 form of an index decrease in the depth (pm)
refractive index increment for both

Fig. 2.5 Refractive-index profiles (for the short and long periods after fabricaextraordinary polarization) for various annealing times measured at tion [Ah194c, Ah195, Mar00,
X=0.633Ltm.






25

Mat92a]. For certain fabrication
0.025 I
conditions, in the initial hours after
0.02
fabrication, a large decrease in Ane
0.015 immediately after
o - fabrication has been shown to occur, followed
0.01 4 weeks after
fabrication by a much smaller, gradual decrease
0.005
in Ane over the next few months 0 0.5 1 1.5 2 2.5 [Ah195]. The total decrease in depth (pm)
refractive index after several Fig. 2.6 Temporal changes in the index profile (for the extraordinary polarization) months can be of the order of 10-3
measured at X=0.633 pm immediately
after fabrication and 4 weeks later.
[Mar00, Mat92a]. Preliminary

measurements have demonstrated

large temporal changes on the same order, as illustrated in Fig. 2.6. These changes can be a

serious problem for devices which are extremely sensitive to index variations and require

long-term stability, such as directional-couplers and QPM waveguides used for frequency

doubling (to a specified wavelength).

The structural phase diagram for APE LiTaO3 is invaluable for understanding the

properties of APE waveguides fabricated in this material. In particular, knowledge of the

structural phase diagram of LiTaO3 would shed light on the origin of the aforementioned

anomalies, as well as serve to identify fabrication conditions that produce stable, possibly

single-phase waveguides. These waveguides could be utilized in the fabrication of reliable

integrated-optic devices which exploit the previously described advantages of LiTaO3.

In the next section, the actual conditions for the fabrication of APE waveguides in

LiTaO3 are presented.






26

2.1.4 APE Waveguide Fabrication Parameters

Using the APE process, channel waveguides were formed on the -X surface of Xcut LiTaO3 and the c- surface of Z-cut LiTaO3 with the aim of single mode operation at a wavelength of 1.5gtm. In this context, c- and c+ will be used to refer to the negative and positive surfaces of a Z-cut crystal, respectively. We defer the characterization of these waveguides until chapter four, but present here the exact parameters used in their fabrication. Additionally, the fabrication parameters of planar waveguides in X-cut LiTaO3 are presented. These waveguides will be used for reflectivity profiling and rocking curve analysis, to be described later in chapter four.

The c+ and c- surfaces of raw Z-cut LiTaO3 crystal wafers can be identified by taking advantage of the piezoelectric effect of the ferroelectric substrates. Fig. 2.7 demonstrates how a voltmeter was used to quickly press down and release on the surface of the substrate. Because of the internal field set up by the spontaneous polarization of the crystal, free charges exist on the surfaces to counteract this field and maintain electrical neutrality. For example, the c+ and c- surfaces have negative and positive free charges, respectively. By rapidly pressing on the crystal and releasing, the internal field due to polarization is momentarily reduced and the voltmeter can read the potential due to the now unbalanced surface charges, so the individual positive and negative surfaces can be identified. These Z-cut wafers were next diced into rectangular samples of various dimensions, depending on their application, where y-propagating channel waveguides could be formed. These samples were then annealed in air at 2000C for 1 hour to relieve stresses incurred during dicing.






27


Voltmeter





+












substrate

Al foil


pos. meter change = c+ surface up neg. meter change = c- surface up Fig. 2.7 Method used to identify the individual surfaces of a z-cut LiTaO3 wafer.



Samples intended for channel waveguides were subjected to the standard photolithographic process. After cleaning in trichloroethane, acetone, and methanol, the samples were dried at 1200C for 5 minutes. Hoechst Celanese AZ 1370 Photoresist was spun on the c- surfaces at 6000 rpm for 35 seconds, after which the samples were baked at 900C for 35 minutes. This resulted in a photoresist thickness of about 1gm. Each substrate was then covered with a digitally dark photo-mask containing desired patterns, be they interferometers, directional couplers, or just straight stripes ranging in width from 2gm to






28

10lm, in steps of 0.5Rm. Samples were aligned under the photo-mask with the mask aligner so that the propagation direction of the waveguide pattern to be used ran parallel with the y-direction of the sample. Upon a 6 second exposure to ultraviolet (UV) light, the samples were held in Hoechst Celanese AZ 312 developer (diluted 1:1.2) for 45 seconds, whereby the areas of photoresist not covered by the mask guide patterns, and hence exposed to the UV, were removed. A Ta mask was sputtered on at a thickness of about 1000A using sputtering parameters of 280V for about 2 minutes and an Ar flow rate of 10 sccm. Acetone was then used to lift off the remaining photoresist, leaving the sample ready for proton exchange, as depicted in Fig. 2.3. The entire photolithographic process from the spinning of a photoresist film through lift-off is illustrated in Fig. 2.8.

Proton exchange was performed using two different sources, pyrophosphoric acid and glycerin. The exchange was performed in a box furnace at 2600C for various times. A custom made acid wash beaker was used to allow the source and sample to heat up separately within the furnace. After stabilizing at the desired exchange temperature, the sample was immersed into the source without having to open the furnace door and create temperature variations. After exchanging, the sample was removed from the source and allowed to cool to room temperature slowly, in order to reduce thermal strain. Each sample was exchanged independently and new source was used each time.

Once cooled, the samples were washed in deionized (DI) water to remove the remaining source and the Ta mask was etched off using a Ta etchant consisting of hydrofluoric and nitric acids. The PE waveguides exchanged using pyrophosphoric acid were then annealed in air using a tube furnace at various times and temperatures.






29







deposit PR







4

expose to UV and develop









deposit Ta










I .... ... lift-off



Fig. 2.8 Illustration of the photolithographic process up
through lift-off.






30

Planar waveguides were also fabricated in X-cut LiTaO3 using the above exchange condition. Being planar, however, they obviously did not need to go through the photolithographic process and their annealing times varied from 0 (as-exchanged) to 120 minutes. The reason for these varied annealing times will become more obvious in Chapter 4. It is noted here that exchange times longer than 20 minutes resulted in surface cracks in Xcut crystals. This was most likely due to the difference in expansion of the lattice parameters along the X and Z-directions, as noted elsewhere [Mat92b].

In the next section, the other methods of waveguide fabrication will be described, namely Ti metal-indiffusion and Zn vapor-indiffusion. The characteristics of waveguides fabricated by these processes will be described and the conditions for their fabrication will be given.



2.2 Waveguide Formation by Metal/Vapor-Indiffusion


As has been mentioned, APE waveguides in LiTaO3 suffer from temporal instabilities. In addition, the introduction of H+ into the crystal via proton exchange is known to cause quenching within a rare-earth-doped region [Nou92] in LiNbO3, reducing the excited state lifetime and therefore decreasing laser performance. As doping with rareearth Er for laser fabrication is a potential application for LiTaO3 also, a means of waveguide fabrication other than APE needs be explored. Since metal-indiffusion in LiNbO3 is known to be temporally stable, an examination of this process is a good place to start for LiTaO3. As such, waveguide fabrication in LiTaO3 by Ti metal-indiffusion and Zn vapor-indiffusion were examined. The known characteristics and fabrication parameters for each of these processes are presented here.






31

2.2.1 Ti Metal-Indiffusion

Waveguide fabrication by Ti indiffusion is performed by evaporating a Ti layer, possibly patterned by photolithography, onto the surface of a substrate and subsequently indiffusing the metal at high temperatures (>1000oC). Since its discovery in 1974 [Sch75], much work has been done on the characterization and fabrication of Ti indiffused waveguides and devices, especially in LiNbO3. There is a limited amount of data reported in literature about the Ti indiffusion process in LiTaO3. Most of it is from the device perspective [How89, Tan78] and does not characterize the process of waveguide formation in LiTaO3 by Ti indiffusion. As a result, much of the description of this process given here comes from reported data in literature dealing with Ti indiffusion in LiNbO3, which is assumed to be very similar.

The indiffusion of Ti results in a relatively Gaussian index profile [Bur79]. Unlike APE, it produces an increase of both the extraordinary and ordinary indices. In both LiNbO3 and LiTaO3 there is a larger change of the extraordinary refractive index than of the ordinary refractive index [Bur79, Tan77].

The diffusion equations governing diffusion depth and the diffusion coefficient are similar to those given for proton exchange in Eqs. (2-2) and (2-3) [Bur79]. However, the diffusion coefficient for Ti in LiTaO3 is about 1.5 orders of magnitude smaller than that for Ti in LiNbO3 [Bur79].

Upon indiffusion, Ti ions take positions at Nb (Ta) sites in the LiNbO3 (LiTaO3) crystal lattice [INS89]. The mechanism of index increase is thought to be attributed to an increase of the polarisability and to the photoelastic effect caused by the different ionic radii of Ti and Nb (Ta) ions [INS89].






32

For the process of Ti indiffusion, there are no reported anomalies similar to those of APE. Ti waveguides in LiNbO3 appear to be temporally stable, though stability has never been confirmed for the same process in LiTaO3. In addition, this process does not cause a change in the crystal lattice strain, so there is no structural phase diagram with multiple phases similar to that of the APE process. However, the surface of samples have been shown to exhibit swelling in the regions of indiffusion. The thickness of this swelling can be as much as 1.5-2.0 times the thickness of the original Ti metal layer before indiffusion [Hol84]. Additionally, Ti:LiNbO3 waveguides are known to have a greater susceptibility to the photorefractive effect than APE waveguides in LiNbO3 [Fuj93].

Due to the high temperatures involved in this indiffusion process, Li20 outdiffusion occurs [Woo93]. This is a problem because it causes a small increase of the extraordinary index everywhere on the surface of the sample, decreasing the index difference between the waveguide and its surroundings. Also due to the high temperature, when Ti indiffusion is performed in LiTaO3, crystal repoling is necessary as the crystal is heated above its Curie point during the process. The method of crystal repoling is described in Section 2.4.

The fabrication conditions for Ti

waveguides in LiTaO3 are as follows. PhotoliTi thography is done on the c+ surface of a LiTaO3 sample to define a pattern using a digitally clear L'TaO3 Imask, as opposed to a digitally dark mask used Fig. 2.9 Depiction of a Ti pattern for APE. Then, 900A of Ti is e-beam deposited
on LiTaO3 before indiffusion. onto the sample after which lift-off in acetone






33

takes place. The result is a sample with a waveguide pattern of Ti metal on the surface, as depicted in Fig. 2.9 for the case of a straight-channel waveguide. The sample is placed inside a platinum box with powered LiNO3, and the box is loaded inside a tube furnace. Dry 02 is flowed through the tube of the furnace at 20 ccm while the sample is indiffused at 1200oC for 20 hours, with an 8o/min rise and fall ramp rate. The purpose of the powered LiNO3 and flowing 02 is to help reduce Li20 outdiffusion from the sample surface. These conditions result in single-mode waveguides for both polarizations at 1.55pm immediately after fabrication.


2.2.2 Zn Vapor-Indiffusion

Zn vapor-indiffusion is performed by placing a sample inside a vacuum sealed quartz ampule with some amount of Zn wire, granules, or powder and heating the assortment to a temperature above the boiling point for Zn so that it vaporizes and diffuses into the sample. The Zn concentration profile obtained resembles a complementary error function erfc(y/d) distribution, where d is the depth at which the concentration is 14.7% of that at the surface [Ekn87]. As with Ti indiffusion, this process results in an increase of both the extraordinary and ordinary indices [Yoo88]. Likewise, the change in the extraordinary index is larger than that of the ordinary index [Yoo88]. Again, the diffusion equations governing depth and diffusion coefficients are similar to those used for proton exchange and Ti indiffusion, namely Eqs. (2-2) and (2-3) [Yoo88].

Similar to Ti indiffusion, Zn vapor-indiffusion does not appear to show any of the same anomalies as APE, and there is no measured change of crystal lattice strain induced by this process. However, the stability of waveguides fabricated in LiTaO3 by this process have not yet been confirmed. Advantageously, the lower temperature of Zn-indiffusion






34

(-800C) results in negligible Li20 outdiffusion from the sample [Yoo88], though the temperature is still high enough to require sample repoling. Curiously, samples after indiffusion appear slightly grayish-brown in color. This discoloration is likely due to 02 deficiency [Kam73] and is easily removed by a short annealing at about 600oC in air.

To fabricate Zn vapor-indiffused waveguides, photolithography was done on the c+ surface of LiTaO3 samples using a digitally dark mask to define a pattern, similar to APE. However, 5000A of SiO2 was e-beam deposited at a rate of 2k/s to serve as the mask, instead of Ta as in APE. Next the sample was subjected to the lift-off process in acetone and was vacuum sealed in a quartz ampule with powered Zn. The ampule was loaded into the tube furnace and indiffused at 8000C for 5.5 hours with an 8/min rise rate and a natural cool down. Afterwards, the SiO2 mask was etched off with Ta etchant and the sample was annealed on a sapphire plate in the box furnace at 600oC for 30 minutes to remove the discoloration. These conditions resulted in single-mode waveguides for the TM polarization at 1.55gm immediately after fabrication.

In the sections to follow, supplemental processing steps such as crystal repoling, polishing, the application of an antireflectivity (AR) coating, and the electroplating of traveling-wave electrode structures are also described.



2.3 Crystal Repoling


As mentioned earlier, the processes of Er-doping, Ti metal-indiffusion, and Zn vapor-indiffusion are all performed above the Curie temperature for LiTaO3, which varies a bit depending on crystal stoichiometry [Bor95], but is generally about 610oC. As a result, the crystal goes from the ferroelectric state to the paraelectric state, where it is






35

multidomain and there is no net polarization and no electro-optic effect. To return the crystal to its monodomain ferroelectric state, it must be poled. For LiTaO3, this is done by heating the crystal to just above its Curie point and applying an electric field of about 250V/cm [Lev66, Suz93] to re-orient the internal dipoles into a monodomain configuration, as shown in Fig. 2.10. The sample is then cooled to below the Curie point with the field applied, locking the dipoles in place.

Previous efforts to pole individual LiTaO3 samples involved using platinum paint on the c+ and c- surfaces to act as electrodes [Jun90]. Platinum must be used as this is the only metal which does not oxidize at the temperatures needed for poling. However, for the application of waveguide devices, success was only achieved when poling X-cut samples as Z-cut samples were found to suffer from surface damage related to the application of electrodes directly onto the sample surface [Jun90].

For this work, the poling process was modified to be useful for poling Z-cut LiTaO3 samples. First, the sample to be poled had 3000-4000A of SiO2 deposited on the c+ and c- surfaces to protect them from damage done by direct contact of the electrodes. Next, the sample was sandwiched between two sheets of platinum foil with copper wires







electrodes Z V00 0







Fig. 2.10 Illustration of the crystal repoling process.






36

attached and placed inside the box furnace with the wire leads running out the vent at the top of the furnace. The wire leads were connected to a DC power supply. The sample was heated to 700C and a voltage of 30V was applied. The oven was immediately turned off and allowed to cool. Once it had cooled to 4500C, the voltage was removed and the sample allowed to continue cooling to room temperature. As will be seen later in Chapter 4, this poling process was successful and resulted in complete recovery of the electro-optic coefficient r33.

In the next section, the technique of end-face polishing is described. A good endface polish in necessary for accurate optical measurements and low insertion loss.



2.4 End-Face Polishing of Channel Waveguides In order to minimize coupling losses into the channel waveguides, good quality endfaces of the sample, perpendicular to the waveguides, are required. A polishing scheme is outlined here which achieves satisfactory results.

First, about 3000-4000A of SiO2 is e-beam deposited on the waveguide surface of the sample to be polished to protect this surface from scratches during the mounting and polishing process. Then the sample is adhered to a Si wafer, using high temperature QuickStickTM clear wax (melting point of 1350C), with the waveguide surface pressed against the Si. In this manner, a thin (-lm or less) layer of wax can be formed between the sample and the Si wafer which serves to minimize cracking and chipping of the sample endface on the waveguide surface. The Si wafer is then cleaved along all four edges of the sample and polished by hand in silicon carbide (SiC) powder, without using a polishing jig, until the Si is flush with the LiTaO3 on all sides. This is done to reduce cracking






37

and flaking of the Si during polishing with a polishing jig. The structure is then mounted to a polishing jig using low temperature bee's wax. Because of the lower melting temperature of the bee's wax (-70'C), there is no danger of melting the thin layer of clear wax between the Si and the LiTaO3 sample.

The endface of the sample is first lapped by hand in SiC powder. Next, the sample is polished by hand in 5pim aluminum oxide (A1203) powder to remove the deeper scratches caused by the SiC powder. The polishing jig is then allowed to rotate freely on a nylon polishing pad with 1m diamond paste for one hour, followed by 0.25Rm diamond paste for thirty minutes. The LiTaO3-Si structure is then turned around on the polishing jig and the opposite endface is polished in a similar manner. The results of this polishing technique for coupling into channel waveguides have been found to work very well.

Along with end-face polishing, a good bevel polishing of the waveguide surface was also required for reflectivity profiling measurements. The bevel polish method used is outlined in the next section.



2.5 The Bevel Polish for Planar Waveguides


For the reflectivity profiling measurement to be presented in Chapter 4, the resolution of the depth profile is enhanced by bevelling a planar sample at an angle of oa=20. Using this technique, the profiles in the depth direction are magnified by 1/sina. In addition, an extremely fine polish is required so as not to introduce a considerable amount of noise in the reflected signal from the beveled surface. The details of the polishing scheme used to achieve a fine bevel polish are given here.






38

First, a Ta film of thickness around 1200A is deposited on the waveguide surface. This serves to identify the substrate surface during scans of the depth profile by introducing a rapid spike in reflected light intensity. The sample is then mounted to a Si wafer, waveguide surface against the Si, in the same manner as outlined in the previous subsection. The Si here helps to prevent rounding due to polishing at the substrate surface. Now, however, the structure is mounted onto a polishing jig which has a ramp-like grove milled in it at an angle of a=2', as illustrated in Fig. 2.11.

Polishing begins by hand in SiC powder until the surface of the LiTaO3 is reached. Aluminum oxide powder is used next, followed by 1m and 0.25ptm diamond pastes, same as before. As a final step, Ultra-Sol 201A/280 Colloidal Alumina Polishing polished surface


Ta
















polishing jig

sample


Fig. 2.11 Depiction of the polishing jig used to bevel polish planar waveguides at
an angle a = 20. This magnifies the depth profiles by a factor of 1/sina.






39

Compound is used with a generous amount of sodium hydroxide (NaOH) dissolved in DI water on a Buehler Microcloth. After polishing for six minutes, the sample is removed from the Microcloth wheel and quickly submerged in the NaOH solution to prevent a glassy film from forming on the LiTaO3 surface.

In the next section, the theory and fabrication conditions associated with the application of an anti-reflectivity coating are described. In many instances, such a coating was necessary to minimize reflections and reduce measurement noise.



2.6 Antireflective Coatings


For the measurements of Chapter 4 where light is reflected from the beveled surface of a planar waveguide, it was found that reflections from the back surface of the sample substrate were contributing considerable noise into the experimental setup. In addition, for the loss measurements of Chapter 4 and the stability measurements of Chapter 6, reflections from the endfaces of waveguide samples were also introducing some noise and error into the setup. For each of these cases, much of the impact of these reflections was eliminated through the use of a single-layer antireflective (AR) coating on either the back side or the endface of the substrate, as the case may be. Some of the theory used in the development of these AR coatings is presented in this section, as well as the material and thickness needed for the wavelength used and the method of its deposition.


2.6.1 Thin Film Theory for Single-Layer AR Coatings

The development here, though only concerned with a single-layer film, stems from the derivation of the transfer matrix method of multiple thin film layers treated in [Hus95,






40

Ped87]. The film is assumed to be homogenous and isotropic. The film is also assumed to be thin enough so that the path difference between multiply reflected and transmitted beams within the film region is smaller than the coherence length of the light. Therefore, the reflected and transmitted beams can completely interfere.

We begin by referring to Fig. 2.12. The condition shown is that for TE polarization, which is the only case we are concerned with here. Later, the stipulation of normal incidence will be made so that there is no distinction between the TE and TM polarizations. Additionally, it is noted that a convention was chosen whereby as normal incidence is approached, the incident and reflected electric field vectors E are oriented in the same direction, while the magnetic field vectors B switch direction. Though this will introduce a phase shift in the B field at 0 = 00, conventional optics defines the direction of polarization as that of the E-field direction and hence the orientations are acceptable. The quantities E, with subscripts, define the magnitudes of the electric fields at each boundary (a) and (b). For example, Er1 is the sum of all multiply reflected beams at interface (a) which are exiting the film toward the region with index no; Ei2 is the sum of all beams incident at interface (b), directed toward the substrate region with index ns; and so on.

By applying the boundary condition that the tangential components of the E and B fields be continuous across each interface, we get Ea = Eio + Erl = Etl + El (2-7) Eb = Ei2 + Er2 = Et2 (2-8) Ba = BiocosOi- BrlCosi = Btl cosOtl - BilCOSOtl (2-9)


Bb = Bi2cosOtl - Br2COS Ot = Bt2cosOt2 (2-10)






41










E E


B B



no
SOi
Eio Erl
boundary (a) Etd Eil
nf B t film B B

boundary (b) Et2
ns t2 substrate


E

B







Fig. 2.12 Diagram of the reflection from a single-layer thin film used to calculate the total reflectivity R of the film.






42

Now, by using


B = E E = n oE (2-11) V C


0 = n0 Ep0O cosOi (2-12) Yf = nff 00cosOtl1 (2-13) 7s = ns cOsROcsOt2 (2-14) we can write the B-fields as

Ba = 70(Ei0 - Erl) = 7f(Etl - Eil) (2-15) Bb = Yf(Ei2 - Er2) = sEt2 (2-16) The E-fields just below interface (a) and just above interface (b) are related through a phase difference 8 caused by one traversal of the film. Using this fact, we may write Ei2 = Etle-j6 (2-17) Eil = Er2e- (2-18) where


8 = (0 nftcosOtl (2-19) Eqs. (2-17) and (2-18) can be used in the portions of Eqs. (2-8) and (2-16) representing the fields within the film region (below (a) and above (b)) to solve for the fields Etl and Eil in terms of Eb and Bb. These new equations can, in turn, be used in the portions of Eqs. (2-7) and (2-15) representing the fields within the film region to give






43


E = Eb cos6 + Bb(sin (2-20)


Ba = Eb(Jf sin 6) + Bbcos8 (2-21) Eqs. (2-20) and (2-21) may be written in matrix form as


E cos jsi E I a = Yf Bb (2-22) S jYfsin6 cos6 b This 2x2 matrix of elements m11, m12, m21, m22 is known as the transfer matrix of the film.

Now, by substituting the portions of Eqs. (2-7), (2-8), (2-15), and (2-16) which represent the E-fields outside of the film region (above (a) and below (b)) into Eq. (2-22), we can finally relate the fields incident and reflected from the film with those transmitted through the equations

1 + r = m11t + m127st (2-23) 70(1 - r) = m21t + m227st (2-24) where we define the reflection and transmission coefficients by rE - (2-25) Ei(2-26)

Et2
t -E - (2-26)






44

Eqs. (2-23) and (2-24) can be solved to give the reflection coefficient in terms of the transfer matrix elements to yield y0m 11 + yOYsm12 - m21 - Ysm22
r = (2-27) Y0m11 + YoYsm12 + m21 + Ysm22 Finally, the reflectance R is defined by R = Ir 2 (2-28) In the next subsection, the actual conditions for the formation and deposition of an AR coating are given.


2.6.2 The AR Film Parameters and Deposition

Several simplifications were employed in the fabrication of the AR film used for this research. First, the reflection, loss, and stability measurements were all performed at near normal incidence, so that all angles in the above derivation are set to zero. Secondly, quarter-wave thickness films were deposited for each wavelength used in this effort. In this instance,


t = (2-29) 4nf

which forces the phase difference of Eq. (2-19) to become 8 = 7t/2 so that cos6 = 0 and sin8 = 1. Using these conditions, the equation for reflectance R, Eq. (2-28), becomes


R Ons2 - n
R =f 2- (2-30) n ons + nf






45

For the case of LiTaO3 in air, ns = 2.12 and no = 1. In order to have no reflections, R = 0, we must use a film material whose index nf = (ns)1/2 = 1.46. We have chosen magnesium fluoride (MgF2), with an index of 1.35, as a suitable material. For the wavelengths of interest, namely 632.8nm and 1550nm, the resulting film thicknesses are 1172A and 2870A, respectively.

The AR films of MgF2 are easily deposited using an electron-beam (e-beam) deposition system. The details of such a system can be found in [Geo92, Mac86].

In the next section, the conditions used for the electroplating of thick gold electrodes are given. Such thick electrodes are used to minimize rf loss in high-speed traveling-wave modulators, such as the one to be demonstrated in Chapter 7.



2.7 Electroplating


For the demonstration of a high-speed traveling-wave electro-optic modulator to be done later in Chapter 7, thick gold electroplated electrodes were needed. In this section, the conditions used for the fabrication and electroplating of these electrodes are given.

To begin with, a buffer layer of SiO2 2000-3000A thick was e-beam deposited at a rate of 2A/s. For a Z-cut crystal, the buffer layer should be deposited uniformly over the entire waveguide surface. For an X-cut crystal, the buffer layer need only be at each end, outside the electrode interaction length where the electrodes bend and cross the waveguides. Next, 150A of Cr was deposited, followed by 800A of Au. The Au is the seed layer for electroplating and the Cr is to help the gold stick to the sample. Note that






46

the e-beam bell jar must not be opened between the deposition of Cr and Au as the Cr will oxidize and, as a result, the Au will not stick to it.

Photolithography of the electrode pattern was performed next. First, the sample was subjected to 30 minutes of vapor-deposition of photoresist adhesion promoter. Next, Hoechst Celanese AZ4620 photoresist was spun on at 2000rpm for 55 seconds. The resulting photoresist thickness was around 9gm. The sample was then baked at 1000C for 15 minutes. The edge bead was then removed and it was baked again at 1000C for 15 minutes. After cooling, the sample was aligned under the electrode pattern and exposed at 9mW/cm2 for 1.2 minutes. It was then developed in Hoechst Celanese AZ400K developer, diluted 1:4 with DI water, for 4.5 minutes, after which it was subjected to a baking sequence of: 90oC for 10 minutes, 120oC for 30 minutes, 90oC for 10 minutes. Finally, the


DC ammeter power supply

- + 10000 +










sample ,.... . - grate


gold solution

beaker


Fig. 2.13 Setup for the electroplating process.






47

sample was loaded into the reactive ion etcher (RIE) and etched in 02 plasma for 1.5 minutes (30mTorr, 12.4 sccm, 60W) to remove any remaining photoresist from the exposed areas.

The thick gold layer is next electroplated to the sample. The setup for this is depicted in Fig. 2.13. The sample was attached to the negative lead of a DC power supply with the positive lead being attached to a grate. Both were submerged into a beaker containing Orotemp 24 gold solution from Technic, Inc. A 1000I resistor and ammeter were put in the circuit to monitor current flow. The beaker was placed in a temperature-controlled water recirculator and the temperature set to 45C. A current of 0.5mA per cm2 of sample surface was then applied for 30 minutes, after which the temperature and current were both increased. The complete process is as follows:

45oC, 0.5mA/cm2, 30 minutes
50oC, lmA/cm2, 10 minutes
55oC, 1.5mA/cm2, 10 minutes
60oC, 2mA/cm2, 40 minutes

The resulting Au thickness was around 6.5ptm.

After electroplating, lift-off was performed using acetone and a 10:1 solution of H2SO4:H202. Gold etchant, diluted 1:2 with DI water, was used next to remove the gold seed layer. Finally, the sample was loaded into the RIE and the Cr layer was removed using a C12 plasma for 20 minutes (30mTorr, 3.4sccm, 60W).



2.8 Summary


In this chapter, the fabrication of planar and channel waveguides in LiTaO3 was presented. A few methods for fabricating the waveguides were given, with the main focus






48

being on the commonly employed APE technique. Several distinct features of the APE process in LiTaO3 were discussed and a few of the currently unresolved anomalies associated with this process were described. The exact fabrication parameters for the planar and channel waveguides used in this research were next presented, followed by some additional processing steps such as crystal repoling, endface and bevel polishing, the application of an AR coating, and electroplating. In the next chapter, a characterization of Erdoping in LiTaO3 is performed.















CHAPTER 3
CHARACTERISTICS OF Er-DOPED LiTaO3



Because of the demand for integrated optical components for use in communication networks, as outlined in Chapter 1, there is a decisive need for lasers and amplifiers to operate in these systems. While much work has been done on the development of fiber amplifiers, particularly Er-doped fiber amplifiers (EDFA's) [Ara92, De193], efforts continue for the development of rare-earth-doped lasers and amplifiers in LiNbO3 and LiTaO3 because of the advantage they posses for allowing the integration of a laser and external device, such as a modulator, on the same substrate. Because these materials, especially LiNbO3, are currently used for the fabrication of network components like filters and modulators and because selective doping may be used on these substrate to define an active gain region on only a portion of the substrate, the opportunity to bypass coupling between individual devices arises by allowing the integration of a laser and another device on the same module.

Though much work has been done recently on the production of rare-earth-doped waveguide amplifiers and lasers in LiNbO3 [Ami96, Bau96, Hua96], an inherent problem with guided-wave devices in LiNbO3 is their low resistance to photorefractive damage at visible and near-infrared wavelengths. Having a much higher photorefractive damage threshold, LiTaO3 emerges as a promising candidate for the development of waveguide




49






50

lasers and amplifiers. However, much research awaits to be performed on the characterization and effect of rare-earth doping, particularly Er, in LiTaO3.

In this chapter, the effect of Er-doping in LiTaO3 is investigated. Some background material relating to work done in this connection on LiNbO3 is presented, as well as the limitations of LiNbO3, leading to the motivation to examine Er-doped LiTaO3. The conditions used for doping of LiTaO3 are then given, followed by Raman and fluorescence measurements of samples identifying Er3+ clustering as the dominant gain limiting factor. A novel method of reducing clustering is then presented and applied. Finally, a two-photon model is used to exemplify the photorefractive process in these crystals and the inherent advantage of LiTaO3 for optical data storage and laser applications is outlined.



3.1 Background


Doping LiNbO3 and LiTaO3 with rare-earth materials, such as Nd or Er, is known to cause laser emissions when optically pumped [Ami96, Bau96, Nou95]. Much work has been done to demonstrate the lasing effect of Nd-doping and Er-doping in LiNbO3 [Ami96,Bau96]. Doping with Nd is known to produce emissions near ljm. Er, on the other hand, is known to produce emissions near 1.5pm, making it more popular because of its potential application in fiber networks which operate in the 1.5ptm region.

There are currently two pump wavelengths which correspond to an electronic transition of Er3+ ions and are used for the pumping of Er-doped devices. They are 980nm and 1480nm. For the case of waveguide lasers in LiNbO3, however, pumping at 1480nm has some advantages. First, the photorefractive effect is smaller at 1480nm than at 980nm, allowing for the application of higher pump powers. Second, waveguides fabricated to be






51

single-mode near 1.5gm are generally multimode at 980nm, giving rise to the problem of selective mode excitation [Bau96]. Whereas when pumping at 1480nm, the mode sizes of the pump and signal are nearly the same and so overlap is better.

As mentioned above, an inherent problem with guided-wave devices in LiNbO3 is their low resistance to photorefractive (PR) damage at visible and near-infrared wavelengths. PR instability is especially severe for active waveguide devices on account of the high power densities within the waveguides. It has been demonstrated [Ami96, Hua96] that PR damage induces significant gain degradation and temporal instability in Er-doped Ti:LiNbO3 waveguide amplifiers and lasers. Even though the problem can be partially alleviated by an appropriate co-doping technique [Die98, Hua96], the performance of the tested devices is still far from the theoretical optimum [Cac97]. Additional efforts to avoid the PR problem have demonstrated a Z-propagating waveguide laser in Er:LiNbO3 [Ami96]. However, a limitation to this configuration is that it does not allow for the integration of an additional electro-optic component, such a modulator, on the same chip as the additional component would not benefit from the large electro-optic coefficient in the Z-direction.

In light of the described limitations of LiNbO3, LiTaO3 appears an attractive alternative with a PR damage threshold of 1500 W/cm2 at 514.5 nm and room temperature, i.e., about forty times higher than the 40 W/cm2 value of LiNbO3 [Gla72]. Similar to LiNbO3, Er needs to be indiffused into LiTaO3 to serve as the active gain medium for active integrated devices, such as amplifiers and lasers. However, Er-doping is known to lead to upconversion in most host materials[Ara92, De193, Gil96], which is widely recognized as a gain limiting factor in Er-doped amplifiers by inducing PR damage [Ara92, De193,






52

Hua96]. There are two mechanisms of upconversion. The first is the homogeneous upconversion caused by excited-state absorption (ESA) within non-clustered Er3+ ions. This process occurs wherever there are Er3+ ions and therefore is inherent to Er-doped waveguides. The second mechanism is nonradiative energy transfer upconversion (ETU) which occurs in Er3+ cluster sites. This form of upconversion can be suppressed by decreasing the relative concentration of cluster sites through a proper choice of doping technology [Ara92, De193] and/or post-indiffusion processing [Die98, Gil96]. For example, in the case of Er:LiNbO3 the cluster site concentration depends on the Li/Nb ratio [Gil96] and, therefore, upconversion efficiency may be reduced significantly by using a vapor phase equilibration technique to increase the Li content [Die98, Gi196]. The similarity of intrinsic defect structures in LiNbO3 and LiTaO3 [Gop96] would therefore suggest that an increase of the Li/Ta ratio should also reduce upconversion caused by ETU in Er:LiTaO3.

As a result, the effect of Er-doping on the physical and optical properties of LiTaO3 needs to be examined. In particular, how doping affects the PR property of the crystal and how much upconverted emission from Er3+ clusters, which reduce gain [Gil96], is present, as well as methods for its reduction. This study begins in the next section by outlining the conditions used for Er-doping in LiTaO3.



3.2 Er-Doping Conditions


The following factors need be considered when Er-doping with the aim of producing active gain regions for waveguide lasers. First, Er3+ concentration should be at an appropriate level so as to provide significant gain. Next, Er3+ concentration profiles






53

should be at least as deep as waveguide index profiles so that the optical mode has good overlap with the gain medium.

For this study, experimental Z-cut LiTaO3 substrates were doped near the surface by indiffusion of a 10-12 nm thick Er layer at 12000C for 100-120 hours. These fabrication parameters were chosen so as to provide an Er3+ concentration profile appropriate for active guided-wave devices. Some of the Er-doped LiTaO3 samples were processed in molten LiNO3 at 2800C for times ranging from 1 to 44 hours. Such a treatment was reported [Kos97a] to be effective toward increasing the Li content in a near-surface layer (with depth 5-10 jm) of LiNbO3.

In order to study the impact of Er-doping on the photorefractive properties of the crystal and to be able to check for upconversion emissions, Raman and fluorescence measurements were needed. The next section describes these techniques.



3.3 Raman/Fluorescence Measurements


Raman and fluorescence spectroscopy are useful tools for examining the effect of Er-doping in LiTaO3 by allowing for the measurement of upconverted emissions and gauging the amount of photorefractive damage induced. In this section, the techniques of Raman and fluorescence spectroscopy are detailed. Then the results of measurements using these methods are presented.


3.3.1 The Raman Effect and Raman Spectroscopy

When monochromatic laser light is incident upon a sample, it is scattered by the material. Most of the scattered light is of the same frequency as the incident light






54

(Rayleigh Scattering). However, a very small portion of the scattered light, a fraction of a percent, is frequency shifted by an amount equal to the vibrational frequency of the molecules in the material. This is the Raman effect. When the frequency shift is toward smaller frequencies, or smaller energies, it is called Stokes scattering. When the frequency shift is toward higher frequencies, or larger energies, it is called anti-Stokes scattering. A spectrometer can be used to detect the scattered light from the surface of a sample. Detecting and plotting the intensity of this scattered light as a function of frequency shift is referred to as Raman spectroscopy. A Raman spectrum is unique to a given material, or to a certain composition of that material.


3.3.2 Fluorescence Spectroscopy

Fluorescence spectroscopy is similar to Raman spectroscopy except that here the detected light is light re-radiated by ions within the material, in our case Er3+ ions. A plot of detected fluorescence intensity versus frequency, or wavelength, is called a fluorescence spectrum. In this case, wavelength is calculated from measured wavelength shift by: ,(nm) = 107/( e - 1m) (3-1) where k is the wavelength in nm, um is the measured wavenumber in cm-1, and ve is the wavenumber of the excitation source in cm-1 (e.g. 15802cm-1 for a 632.8nm source and 15456cm-1 for a 647nm source).


3.3.3 Fluorescence Measurements

To examine the influence of up-converted light on the PR effect in Er:LiTaO3, the fluorescence spectrum of each Er-doped sample was first measured using a Renishaw Ramascope spectrometer operating at excitation wavelengths of 632.8, 647 and 785nm.






55

In the setup, a linearly polarized laser beam was focused, by a 50X objective, to a 2[tm spot on the optical-grade polished edge of the sample under study. The plane of light polarization was parallel to the optical axis. A Raman microprobe, i.e. a microscope optically coupled to a Raman spectrometer, was used to collect both the out-of-plane scattered light and the fluorescence emission.

There was a significant difference between the fluorescence spectrum produced by pumping at 785nm and that produced by pumping at 647nm or 632.8nm. For 785nm, only intensive Stokes fluorescence (SF) was induced (i.e. fluorescence with photon energies smaller than the excitation energy). For 647nm and 632.8 nm, however, there was anti-Stokes fluorescence (ASF) with both a green emission at 520-560nm (see Fig. 3.1)


2000

647nm, J=0.7x103 W/cm2
......... 632.8nm, J=2x103 W/cm2
1500




1000




500





510 520 530 540 550 560 570
wavelength (nm)

Fig. 3.1 Anti-Stokes fluorescence spectra excited at 632.8nm and 647nm in
the Er:LiTaO3 sample with the highest Er concentration (-0.65
mol%).






56




3 10 4 _ Er:LiTaO3
S- - - - - - ....... virgin LiTaO3


R
S2104 R



E 1 104
0



6 0 -* -iL L L j , 1-li -11- 4- 1
610 620 630 640 650
wavelength (nm)

Fig. 3.2 Fluorescence spectra of the red ASF emission for the Er-doped
sample of Fig. 3.1 (solid) and Raman spectra of virgin LiTaO3 (dashed) excited at X=647nm and J=5x104 W/cm2. R denotes Raman lines and L denotes the excitation laser line after a notch
filter.






1000

100

� 10

1
O -- 632.8nm o 0.1 I _E 647nm
8 0.01I 1 ,I
102 103 104 105 106
input intensity J (W/cm2)


Fig. 3.3 Intensity of the green ASF emission versus input intensity J for excitation at 632.8nm (circles) and 647nm (squares) for the sample shown in Fig.
3.1.






57

and a red emission at 613-628nm (Fig. 3.2), in addition to the SE Within the measurement range of the input intensity J, the SF intensity ISF changed proportionally to J at all excitation wavelengths in accordance with previous studies [Ara92, De193, Gi196]. The dependence of the ASF intensity IASF of the green emission was found to be more involved and had a knee at a certain value Jt of J, with Jt ranging from 104 to 1.1.104 W/cm2 for 632.8nm pumping and about 4*103 W/cm2 for 647nm pumping. As illustrated in Fig. 3.3, IASF(J) is a nonlinear function IASFOCJX for JJt, there is a linear correlation between IASF and J, i.e. x = 1. The dependence of IASF of the red emission had the same character as that of the green emission in Fig. 3.3, but the intensity of the anti-Stokes red emission was smaller than that of the green emission by about 4.5 times. Such an approximately quadratic input-intensity law, observed at a low pump range (< 4-103 W/cm2), suggests that the green and red ASF emissions are attributed to a two-photon absorption process, to be described in Section 3.4. Reduction from a quadratic to a linear dependence at high input intensities J is attributed, in accordance with Ref.[Ara92], to population depletion of the ground state.


3.3.4 Raman Measurements

According to Ref. [Gla72], green light with the observed intensity IASF (see Fig. 3.3) can not alone induce significant "single-color" PR effect in LiTaO3. However, a comparison of the data for IASF (Fig. 3.3) with the results reported for LiNbO3 [Lan98] and Pr:LiNbO3 [Bai97, Gue97] leaves open the possibility that the secondary green emission of Er3+ clusters has enough photon energy and intensity IASF to effect, together with the pumping red light, significant photorefractivity via two-color excitation in Er:LiTaO3.






58

no PR effect The mechanics of two-color excitation is IR detailed later in Section 3.6.
PR effect
To verify the assumption that the

up-converted green emission acts as a Jmin J "self-gating" source, enhancing the PR
(a)
effect via two-photon (or two-color) exciIR
tation without the need for an external gatI . . no PR effect R,min
PR effect ing source, the PR effect was analyzed by Raman spectroscopy measurements. The Jmin
(b) PR effect is known to change the intensity Fig. 3.4 Raman intensity (a) and normalized Raman intensity (b) IR of Raman-shifted lines by an amount
as a function of input intensity, illustrating the effects of which is proportional to the light-induced
photorefractive damage.
extraordinary refractive-index change

An. Since An depends on input power intensity J, the change of IR also depends on J, as illustrated in Fig. 3.4. The dependence of An on J is calculated by the following relation, which was experimentally verified [Kos94]:



'Rmin R I Ans(J) = Ad 1 = A i 1 (3-2)


where Ans(J) is the intensity-dependent steady-state value of An; Jmin is the minimum value of J when photorefractive damage becomes undetectable by Raman spectroscopy;

IR,min and IR are Raman intensities measured at the minimum and an arbitrary value of J, respectively; IR = IRJmin/J is a normalized Raman intensity; and A is a coefficient






59

60
2000 - ------------ .---55 \
S1500
.> - - - 632.8nm w/ Er
------632.8nmw/oEr
1000 50 - - 785nm w/Er ......... 785nm w/o Er

500 45
0 . ,S. "
0 ' ' ' ' ' ' ' ' ' ' ' '''t ' 4 0
280 320 360 400 440 0 2 ld 4 1d 6 1d
Raman shift, wavenumber (cm-1) input intensity J (W/cm2)

Fig. 3.5 Raman spectra of virgin LiTaO3 Fig. 3.6 Normalized Raman intensity of
(solid) and Er:LiTaO3 (dashed) the line at 353cm-1 versus input measured at X=632.8nm and intensity measured at 632.8nm J=2.5x104 W/cm2. and 785nm for samples with and without Er.



dependent upon the setup used for the Raman measurements. The value of A has been determined from Raman measurements of previously studied LiTaO3, Cu:H:LiNbO3 and LiNbO3 samples [Kos94, Kos97a, Kos97b].

Fig. 3.5 shows the measured Raman spectra of virgin LiTaO3 and Er:LiTaO3. The decreased intensity for the Er-doped sample is a result of the increased PR effect. By making this measurement for several different input intensities J, a plot of normalized Raman intensity IR' versus J was constructed and is shown in Fig. 3.6. Note that if the PR effect was absent, IR would be constant versus J (Fig. 3.4). It is then seen from Fig. 3.6 that all samples suffered some degree of photorefractivity as J was increased. From this data, Ans at any value of J can be calculated from Eq. (3-2).

Fig. 3.7 shows the calculated value of Ans for the Er:LiTaO3 sample with the highest Er3+ concentration at the surface (-0.65 mol%) and for a virgin LiTaO3 substrate. There are two important observations to be made. First, the presence of Er3+ ions has






60

14 I I

12 .

10

o 8 - . - 632.8nm w/ Er
- - * 632.8 nm w/o Er
J 6 785nm w/ Er
< ........ 785nm w/o Er
4

2
0 . . . . . . .. . . . . ... . . . . ... .

1 104 3104 5104
input intensity J (W/cm2)


Fig. 3.7 Saturated values of photorefractive index change Ans versus input
intensity J for excitation wavelengths of 632.8nm and 785nm.



increased the PR effect. Second, there is a pronounced dependence of the PR effect on input intensity and this dependence is modified with wavelength. At the 785nm excitation wavelength, the dependence of Ans on J in both the Er-doped sample and the virgin LiTaO3 substrate is nearly linear for J > 103 W/cm2. This was reported [Gla72, Kos97a] to be typical for nominally pure LiTaO3 and LiNbO3 in the studied range of J from 103 to 5*104 W/cm2. The slope a(Ans)/J of this dependence increased proportionally with Er3+ concentration. The proportionality factor has a maximum value of 1.6 in the sample with the highest Er3+ concentration (Fig. 3.7). At 632.8nm, the dependence of Ans on J for the same Er-doped sample has a much more complicated character, as seen in Fig. 3.7. Using a simple mathematical fitting to this data demonstrates the appearance of an additional intensity dependent component Ans(J), superimposed on the linear dependence, given by






61

the following empirical relation:

Ans(J) = D[Er3+] jx (3-3) where D is a constant of proportionality and [Er3+] is the Er3+ concentration. Above a certain threshold value of J, around 104 W/cm2, the value of x is 1. Below this threshold, x is about 2. However, at light intensities lower than 103 W/cm2 the contribution described by Eq. (3-3) becomes insignificant. Finally, at 647nm the dependence of Ans on J was roughly similar to that for 632.8nm. However, accuracy in calculating Ans from spontaneous Raman scattering at 647nm, as well as for J > 5x104 W/cm2 at 632.8nm, was hampered because of interference from SF and resonant Raman scattering. Rough estimation of Ans gives a value of around 2x 10-3 for the highest J used (2x 105 W/cm2).

From a comparison of the data for Ans(J) (Fig. 3.7) with that of Fig. 3.3, it is evident that there is a strong correlation between the up-converted green emission and the increased PR effect at high intensities of J. Such a correlation may be expressed, according to Eq. (3-3), by the following empirical relations: Ans = (E)IASF; IASF = (D/E)[Er3+] jx (3-4) where E is a constant of proportionality. Note, that the experimental dependence of Ans on IASF expressed by Eq. (3-4) is in agreement with the gating dependence of photorefractivity reported for Pr:LiNbO3 using an extrinsic source of gating light [Gue97]. As such, we may conclude that Er:LiTaO3 employs self-gating (i.e. secondary green emissions) to achieve significant two-color photorefractivity from a monochromatic pump.

It is noted that for a quantitative comparison, Er-doped LiNbO3 samples were also prepared with fabrication conditions similar to those reported for active devices [Bau96]. A comparison of the data reported in Fig. 3.7 with the corresponding data for Er-doped






62

LiNbO3 showed that Er-doped LiTaO3 has a much smaller (2.2-4.5 times) PR index change than that of Er-doped LiNbO3 in the intensity range studied. The reason for this advantage will become clearer when we discuss the two-photon model for PR damage later in this chapter. Nonetheless, it indicates that LiTaO3 is the more attractive host for active Er-doped devices where high PR resistance is indispensable.

In the next section, the mechanism responsible for the upconverted emissions, which lead to enhanced photorefractivity, is investigated.



3.4 Energy Transfer Upconversion


In the last section, it was determined that the upconverted emissions act as a selfgating source to, along with the pump source, enhance photorefractivity by the two-photon, or two-color, photorefractive effect. In this section, the mechanism creating the upconverted emissions is investigated. Particular attention is paid to the process of energy transfer upconversion between Er3+ clusters.

The two-photon, multistep processes associated with the red and green ASF emissions are exemplified in Fig. 3.8 which shows a partial energy-level diagram for the Er3+ ion. Electrons in Er3+ ions are first excited to the metastable 4F9/2 state, then excited to the 2(4F9/2) state, after which they decay nonradiatively to the4G11/2, 2H11/2 or 4S3/2 state. The red emissions at 613-628nm correspond to a transition from the 4G11/2 state to the 4j11/2 excited state, while the transitions from the 2H11/2 and 4S3/2 states to the ground state produce the green emissions. The substructure of the green and red emissions reflects the splitting of the ground state and excited states into Stark sublevels. It is important to note that the red emissions at 613-628nm observed here for Er:LiTaO3 were not







63


2(4F9/2)
GI 1/2

(ETU)

S , 2H11/2
3/2

red ASF 4F9/2 613 - 628nm

S- 419/2

SF 4111/2 632.8 or 650 - 660nm green ASF
647nm SF 510 - 570nm
pumpS
785nm 820nm
pump
S415/2


Fig. 3.8 Partial energy-level diagram of Er3+ ions in LiTaO3 depicting the
radiative transitions observed in our measurements. The ETU process, radiative transitions, and nonradiative transitions are represented by thin solid arrows, thick solid arrows, and dashed
arrows, respectively.




observed in previous studies of other Er-doped materials [Ara92, De193, Die98, Red94], including Er:LiNbO3, pumped via the 4F9/2 manifold {Gil96].

Upconversion is known to be strongest in regions of highest Er3+ concentration [Ara92, Gi196], where Er3+ clustering is the greatest. As such, the dominant mechanism of upconversion leading to the red and green ASF emissions observed in Er:LiTaO3 was concluded to be ETU, that is, upconversion by energy transfer between neighboring Er3+ ions within an Er3+ cluster. To confirm this, the effect of the Er-doping level on upconversion efficiency was examined by profiling, in the depth direction, both the ASF and SF emissions within the Er-indiffused near-surface layer. Since ISF was known to be proportional






64


6000 d=lgm

5000
=4gm
4000
Sd=5.51tm 3000
d=8gm
2000 1000

0 0


0 200 400 600 800 1000
Raman shift, wavenumber (cm-1) Fig. 3.9 Stokes fluorescence spectra at various depths d below the surface.




to the Er3+ concentration for pump intensities below 103 W/cm2, it was possible to determine that the Er3+ concentration profile is exponential-like with a 1/e depth of around 56gm (Fig. 3.9). The intensity profile of the ASF emission had, however, a depth of around 2gm. In other words, the ASF emissions emanated from the regions of highest Er3+ concentration (near the surface), where most of the Er3+ clusters resided. This was direct evidence of the dominating role of nonradiative energy transfer between clustered Er3+ ions in the mechanism of upconversion in Er-doped LiTaO3.

Since Er3+ clustering was determined to be the primary factor leading to upconversion, significant reduction of the upconversion efficiency in Er:LiTaO3 was expected to be achieved by a post-indiffusion process to reduce Er3+ clustering. This is pursued in the next section.






65

3.5 Li Treatment


In an effort to decrease the amount of Er3+ clusters, and hence reduce upconversion, a Li-treatment process was employed to increase the Li/Ta ratio. The results of this Li-treatment are discussed in this section.

Some of the fabricated Er:LiTaO3 samples were processed in molten LiNO3. Subsequent fluorescence measurements showed a gradual decrease of IASF with an increase in the process time up to 12-16 hours. At the same time, the intensity of SF remained nearly constant, i.e. the total Er3+ content in the crystal was not affected by this processing. Further increase of the processing time induced only a small change in IASF. The most impressive result was a nearly 54% reduction in the intensity of the green ASF emission as a result of the post-indiffusion processing, as illustrated in Fig. 3.10. This reduction is



3000
20 - - - -Er:LiTaO3, X=647nm, J=0.7x 103 W/cm2 2500 -- Er:LiTaO3, X=632.8nm, J=2x103 W/cm2 t---Li:Er:LiTaO3, X=632.8nm, J=2x103 W/cm2 2000

1500 1000
o


0
525 535 545 555 565
wavelength (nm)

Fig. 3.10 Fluorescence spectra of the green ASF emission from the Er-indiffused sample with the highest Er3+ concentration. The dotted line corresponds to a pump wavelength of X=647nm and input intensity J=0.7x103 W/cm . The solid and dashed lines correspond to X=632.8nm and J=2x103 W/cm2, before and after Li treatment,
respectively.






66

attributed to the decrease of the Er3+ cluster site concentration as a result of the increase in Li content in the near-surface layer. Similar to the effect of vapor phase equilibration in Er:LiNbO3 [Gi196], a redistribution of Er3+ sites occurs in response to a change in the defect populations created upon Er incorporation. It is important to note that the technique of post-indiffusion processing in a Li-containing melt described here has the advantage of simplicity, as well as being performed at temperatures below the Curie point for LiTaO3. This Li-treatment process affects only a shallow near-surface layer. However, it is sufficiently deep to reduce the Er3+ cluster concentration and therefore may be a useful technique for the optimization of active integrated-optical devices utilizing an Er-doped nearsurface layer in LiTaO3 crystals.

To this point, it has been determined that ETU is the dominant mechanism of upconversion in Er:LiTaO3 and that the presence of this upconverted light leads to PR damage via the two-photon, or two-color, PR effect. A two-photon model used to exemplify this process is described in the next section.



3.6 The Two-Photon Model of the PR Effect: LiTaO3's Advantage in Data Storage and Lasers


In this section, a two-photon model is applied to model the PR effect and explain how upconverted light, along with the presence of the longer wavelength incident light, leads to PR damage. During this process, the inherent advantages of LiTaO3 for applications such as optical data storage and laser fabrication will also be described.

To develop a theoretical explanation of the phenomenon of photorefractivity enhancement by Er-doping, the two-photon model of the PR effect has been employed.






67

This model has been developed and e- migration
conduction band
recording light experimentally confirmed for LiNbO3 2-polaron [Jer95, Kos97a] and has been applied gating light
here to LiTaO3 due to the similarity of the intrinsic defect structures between 1- bipolaron 3 - deep trap the two materials [Gop96]. In the valence band
case of undoped lithium niobate Fig. 3.11 Schematic-level diagram of the twophoton PR effect depicting the gat- (LiNbO3) [Kos97a, Lan98] this meching light, longer wavelength recording light and bipolaron level, single anism involves bipolarons as deep
polaron level, and deep trap (outside
the illuminated area). donors and metastable small polarons which provide the high PR sensitivity. Bipolarons and polarons exist due to the intrinsic defect structure of LiNbO3, originating from its lithium deficiency [Sch91]. Near infrared (or red) recording [Bai97, Gue97, Lan98] occurs from the small polaron state consisting of an electron trapped at the antisite defect (NbLi), designated level 2 in Fig. 3.11. This level is normally populated via an extrinsically applied blue-green gating light through the photo-dissociation of bipolarons (level 1 in Fig. 3.11). After the excitation of electrons from the small polaron sites to the conduction band, the electrons drift out of the illuminated area and are retrapped at bipolaron sites and other deep traps (level 3) resulting in PR damage, or PR recording if the application is optical data storage. Such a "two-color" scheme has the advantage of a very high gating ratio, i.e., the ratio of writing sensitivity with and without the second gating color [Gue97, Lan98].






68

The main feature of the two-photon excitation is the nearly linear dependence of Ans on J for high input intensities J, which has been verified experimentally in LiNbO3 [Fuj93, Jer95] and LiTaO3 [Gla72]. Moreover, it has been established [Kos97a], that { (Ans)/J } - N, where N is the concentration of antisite defects. Such a proportionality is in full accordance with the two-photon model.

It is well known that Er [Gil96] and Pr [Gue97] dopants increase N, especially bipolarons, in LiNbO3 through the need for charge compensation of ions with valence state (3+). Hence, data obtained here at the excitation wavelength of 785 nm (Fig. 3.7) is clear evidence of a similar effect for Er3+ in LiTaO3. Moreover, the relative change of N caused by Er-doping can be estimated from a comparison of the slope {8(Ans)/J} observed in the Er-doped sample with the one observed in the virgin congruent substrate. It was found that N is increased 1.6 times by Er-doping in the sample with the highest Er3+ concentration (Fig. 3.7).

Radiation at 785 and 632.8 nm is near the maximum of the absorption band for small polarons, but is not optimal for the photo-induced dissociation of bipolarons, as these wavelengths are very far from the bipolaronic absorption band [Sch91]. Therefore, the generation rate of shallow PR centers is too low to induce significant photorefractivity, even at high light intensities. However, the situation changes when excitation with two wavelengths is used, with photons of the shorter wavelength (green light) exciting bipolarons and increasing the virtual population of single-polaron levels and photons of the other wavelength (red or infrared) sweeping them effectively into the conduction band [Gue97, Lan98], as illustrated in Fig. 3.11. Recent research with additional sources of such "gating" light has shown [Gue97, Lan98] that the two-color scheme has a very high






69

gating ratio (i.e. the ratio of PR sensitivity with and without the second color) and, therefore, can be implemented even with low-power sources (e.g., 0.1 W/cm2 [Lan98] - 1 W/ cm2 [Gue97]). It means that the secondary green emission of Er3+ clusters has enough photon energy and intensity IASF to effect, together with the pumping red light, significant photorefractivity via the two-color excitation.

The difference in excitation of green emission observed between pumping at 785 and 632.8 nm is consistent with previous data [Gil96] on the spectral dependence of upconversion excitation within Er3+ clusters. Such an upconversion is an undesirable effect in Er-doped devices [Ami96, Gil96, Hua96] operating at telecommunication wavelengths. Experimental data reported here demonstrates that this upconversion induces a more pronounced PR effect in LiNbO3 than in LiTaO3. This is because the green emission of Er3+ clusters is within the maximum of the bipolaronic absorption band in LiNbO3 [Sch91] yet is far from the maximum of the absorption band in LiTaO3 [Kos97b]. This fact makes LiTaO3 more attractive for the fabrication of Er-doped waveguide amplifiers. At the same time, data presented here also points to the advantage of LiTaO3 over LiNbO3 among perspective rear-earth-doped PR materials for non-volatile holographic storage with two-photon recording. At an optimal wavelength of extrinsic gating light (ranging from 280 to 370 nm in LiTaO3 [Kos97b]), a saturated value of holographic efficiency should be limited mainly by the intrinsic defects concentration N, having a higher value (1.4 times) in LiTaO3 [Bor95, Kos99]. Because the gating beam generally does not have to be from a coherent light source [Gue97], low-cost light sources such as filtered xenon and mercury lamps can be used for this purpose in Er:LiTaO3.






70

As an additional application, the PR enhancement brought about by Er-doping in LiTaO3 affords the possibility of producing PR gratings for use in WDM filters or waveguide lasers, to name a few devices. Such gratings have been demonstrated in Fedoped LiNbO3, though they are unstable and degrade over time [Huk98].



3.7 Summary


In this chapter, the effects of Er-doping in LiTaO3 were examined. It was found that Er-doping led to upconversion emissions. Energy transfer upconversion between Er3+ ions was determined to be the dominant mechanism for this upconversion and that the presence of this upconverted light led to PR damage. A Li-treatment technique to reduce clustering was described and successfully applied to minimize upconversion. A two-photon model was then applied to model the PR effect and explain how upconverted light, along with the presence of the longer wavelength incident light, caused PR damage through "self-gating". Finally, in addition to its known potential for applications like high-speed modulation and nonlinear frequency doubling, LiTaO3 was also demonstrated to have strong advantages and potential for the applications of optical data storage, lasers, and devices incorporating PR gratings. However, to realize these devices, stable waveguides in LiTaO3 must first be developed. The subject of waveguides is returned to in the next chapter, where waveguides are characterized using a variety of techniques.















CHAPTER 4
WAVEGUIDE CHARACTERIZATION AND MEASUREMENT TECHNIQUES



In the fabrication of integrated optical components incorporating channel waveguides, it is often necessary to know certain characteristics of the waveguide, such as its region of single-mode operation and the depth profile of its refractive index increment, in order to properly predict device performance. Additionally, the proper performance of a device may rely heavily on additional waveguide features, such as it being of low loss and having a large electro-optic coefficient. As a result, it is necessary to be able to accurately measure and determine these characteristics in order to effectively design, model, and fabricate reliable waveguide devices.

This chapter provides the measurement techniques and characterization results of the channel waveguides whose fabrication methods and conditions were given in Chapter 2. All of the waveguides were fabricated in LiTaO3 with the aim of single-mode operation at the wavelength of 1.55gm. First, a technique of near-field characterization is described and used to determine the region of single-mode operation by coupling light from a 1.55gm laser into various width guides and focusing the waveguide outputs onto an IR camera, while observing the mode pattern on a monitor. Next, a modified prism coupling technique is applied to measure the effective mode index increments AN of the waveguides at various wavelengths. Propagation loss measurements are then performed, followed by a setup description and the results on measurements of the r33 electro-optic coefficient. The 71






72

photorefractive characteristic of waveguides in the infrared, near 1.5gm, is measured next and compared to similar structures in LiNbO3. The chapter concludes with some measurements of planar waveguides, including index profile measurements and rocking curve analysis, which will provide data to be used in a subsequent chapter.



4.1 Near-Field Characterization


In this section, a first inspection of the number of modes supported by the channel waveguides of various widths is carried out. This is done using a near-field characterization method whereby the output faces of waveguides excited from the single-mode fiber pigtail of a diode laser source are focused onto a camera so that the number of spots associated with different modes can be counted off of the display monitor.






waveguide
sample
1.55tm (w
diode laser single-mode t IR camera fiber focusing lens








monitor


Fig. 4.1 Schematic of the near-field characterization setup used to initially
determine the number of modes supported by the channel waveguides
of various widths.






73

4.1.1 Description of the Measurement Technique

Light from a 1.55gm diode laser was coupled into each of the channel waveguides via a single-mode fiber pigtail. The complete experimental setup is illustrated in Fig. 4.1. The output face of each waveguide was focused, using a 20x lens, onto the input surface of an infrared (IR) camera, with the output video feed of the camera connected directly to a video display monitor. Using this approach, the number of modes supported by a particular waveguide can easily be determined by counting the number of observed mode spots appearing on the video monitor screen. Some results of this measurement technique are presented next.


4.1.2 Modal Characterization Results By scanning the single-mode fiber from the laser source across the input face of each waveguide, it was possible to create off-axis coupling into any higher order

(a) modes that may have been supported by the waveguide. This was done for each of the channel waveguides in LiTaO3, and the number of modes supported for each was determined. As an example, Fig. 4.2 shows

(b)
the video monitor screen photographs taken Fig. 4.2 Video monitor screen Ssfor a single-mode and double-mode channel snap shots depicting a single-mode (a) and a double-mode (b) channel
waveguide.






74


Fabrication Fabrication Single-Mode
Technique Conditions Region PE: 260oC, 2hrs.
APE A: 300oC, lhr.
(pyrophosphoric PE: 260oC, 1 hr. Z 2.5-6.5gm
acid) A: 300oC, 3 hrs.
PE: 260oC, 30 min.

PE: 260oC, 10 min.
APE A: 340'C, 1 hr. Z 3.5-7.5gm
(pyrophosphoric PE: 260oC, 7 min.
acid) A: 340oC, 10 hrs.

PE PE: 260oC, 24 hrs. X 2-3.5gm
(glycerin)

Ti-indiffusion 900A, 1200oC, 20 hrs. Z 2.5-4.5gm

Zn vapor-indiffusion 800oC, 5.5 hrs Z 2.5-7.5gm


Table 4.1 Near-field characterization results depicting the region of singlemode operation immediately after fabrication for the extraordinary index, in terms of waveguide channel width, for various fabrication
conditions.



There were far too many waveguides produced with varying fabrication conditions, especially using APE, to be presented here. However, Table 4.1 presents some results of modal characterization from each of the fabrication processes used. It is noted that the regions of single-mode operation listed here are for the extraordinary index only. Additionally, waveguides fabricated by APE involved multiple steps of proton exchange followed by annealing in order to achieve the desired waveguide depth, while maintaining a desired proton concentration at the surface. The reason for this will become clear when the structural phase diagram for APE is reconstructed (Chapter 5) and stability issues are addressed (Chapter 6).






75

In the next section, a prism coupling technique is applied to measure the effective mode index increments of channel waveguides at various wavelengths. Such a technique allows for a rough estimation of the magnitude of the index increment within a waveguide, as well as providing some information as to the wavelength dispersion of the index increment.



4.2 Effective Mode Index Measurements


The effective index increments of modes in planar and channel waveguides have been measured successfully for many years by the prism coupling method [Ulr73]. These values not only give some idea on the magnitude of index increase within the guiding region, but can also be used to reconstruct the index profile by the inverse WKB method. This naturally requires that the guide supports at least three modes and exhibit a monotonically decreasing profile. In this section, the underlying principle behind the prism coupling method to obtain effective mode indices N is presented. A modified version of this standard prism coupling technique is then described that provides higher accuracy. This measurement method is applied to determine the effective mode index increments AN for the fundamental modes of APE channel waveguides in LiTaO3.


4.2.1 The Prism Coupling Technique

The standard prism coupling setup is shown in Fig. 4.3. Here, a prism with index np is located some distance h, on the order of a wavelength or less, above the surface of a waveguide. The waveguiding region has an index of ng, while the substrate has an index of ns. The cover index nc is assumed to be unity. The basic principle of operation is based






76




incident beam surface normal n, nc � prism



01 d



h / substrate coupled waveguide


Fig. 4.3 Schematic of the standard prism coupling method used to selectively couple light into individual waveguide modes.



on the coupling of light from the prism to the waveguide by satisfying the phase matching condition for each mode within the waveguide separately.

Light incident on the face (hypotenuse) of the prism at an angle Oi, is refracted as it enters the prism according to Snell's law: n sinOi = n sin1 (4-1) The angle of incidence of the incoming beam is adjusted so as to maintain total internal reflection at the np-nc interface. This condition is satisfied if the angle 02 is greater than the critical angle:


02 > 0c = asin C (4-2)


The coupling of light into modes of the waveguide can occur only if the horizontal components of the k vector in the prism matches the propagation constant of the waveguide. This is referred to as the phase matching condition:






77

N = n sin 02 (4-3) where N is the normalized propagation constant, or effective mode index, of a particular mode of the waveguide. In order to obtain real angles 02, the prism index np, therefore, should be larger than N. For maximum coupling, the gap h between the prism and the waveguide should be as small as possible. Usually it is about a quarter wavelength and can be achieved through the use of clamps, pushing the prism up against the substrate surface. Additionally, the interaction distance d should be comparable to one coupling length. If it is longer, light may be coupled back into the prism. Using simple geometry within the prism and Snell's law, it can be shown that the effective mode index N is related to the incident angle by


N= nsin[A + asin sin0 (4-4)


where A is the prism angle. From this equation, it is obvious that the effective mode index for each mode within the waveguide can be calculated by accurately measuring the angle of incidence of the incoming light which selectively excites each mode, and the angle A of the prism. In addition, the value of no at the wavelength of interest must be known to an accuracy of at least four decimal places, as will be discussed later.

An alternative to this is achieved by considering the path taken by the light for coupling into the waveguide, now in the reverse direction. In this case, light is first launched into the waveguide structure, exciting all of the existing modes. Each of these modes will have a different effective index N value and thus exit the hypotenuse face of the prism at a different angle. The angular spectrum of the emerging light exhibits peaks, called m-lines, corresponding to the modes of different order. The effective index values for each of the






78

modes can then be calculated by measuring the angular deviation of the corresponding mline from the normal to the prism surface. In this case, the fundamental mode will have the largest angular deviation.

While this standard prism coupling technique depends on the measurement of some angular deviation, and this measurement can usually be made very accurately, it does have one significant limitation which needs to be mentioned. In many cases, it is the mode index increment, AN=N-ne, of each of the modes that is of interest, where ne is the bulk extraordinary index value. The accuracy of making such a determination using this prism coupling arrangement is then directly proportional to the prism index np and the bulk index of the substrate ne. Normally, AN needs to be known with an accuracy of 10-4 or better. However, the values of prism and bulk index are usually only known with an accuracy of, at best, 10-3. As such, the standard prism coupling method is not a dependable means to obtain effective mode index increments for APE waveguides. However, a modification to this technique may be employed which requires less accuracy in the values of these indices. This new technique is discussed next.


4.2.2 A Modified Prism Coupling Method

The new prism coupling technique [Tav95] is depicted in Fig. 4.4. Light is coupled into the waveguides via a single-mode fiber, then out-coupled via the prism. As before, the m-lines corresponding to the different excited modes of the waveguide, as well as a region delimiting the continuum of substrate radiation, are seen exiting the prism face. This spectrum is easily observed by placing a screen near to the exiting face of the prism. The angular deviation AB between the edge of substrate radiation and the mode of interest (only the fundamental mode is depicted in the figure) can be precisely measured and






79

surface normal


edge of
substrate 13
radiation ..............prism


fundamental
mode


S............ - fiber
substrate fiber
waveguide


Fig. 4.4 Modified prism coupling technique used to measure mode effective index increments.


substituted into the derivative of the prism coupling equation, Eq. (4-4), with respect to the coupling angle 13:


dN = = os A + as cos (4-5)
dN A + asin l 2 2
d AL P np -(sin)2


Since the measured angle AB is the angle between the substrate and any mode of interest, this new prism coupling equation gives the mode index increment AN=N-ne of this mode above the substrate value. Because AB<

4.2.3 Measured Mode Index Increments

Using the modified prism coupling method, the mode index increments AN of straight-channel APE waveguides in LiTaO3 were measured at four different wavelengths:






80


0.015
..- 632.8 nm


F-/
o .o ....................... ---------------------------------- - ...............................................

0.01
890 nm




1060 nm
-- - - - - - - .. . . . . .... . . . . . . . .. -- - - -- - - -- - - -- - - -- - - -- - - -- -
0 .0 0 5 ..................... IF



V 1310 nm


0

0 2 4 6 8 10 12 channel mask width (gm) Fig. 4.5 Measured fundamental mode index increments at different wavelengths for channel waveguides of various widths obtained from the
modified prism coupler method.



0.6328gm, 0.890gm, 1.060gm, and 1.310gm. A rutile prism with an angle A=45.1290 was used for coupling light out of the waveguide. The refractive index np of the rutile prism at each wavelength was found from a cubic spline interpolation of the dispersion data for rutile presented in Appendix A.

The measured mode increments for the fundamental mode at each wavelength are plotted in Fig. 4.5 as a function of channel mask width. For the experimentally measured






81

index increments at X=l .31tm, these values were found to be in reasonable agreement with those obtained in [Tav95] by using a numerical routine of waveguide modeling [Tav94], under similar waveguide fabrication conditions. This figure also presents some insight into the wavelength dispersion of the index increment.

In addition to having some knowledge of the modal characteristics and magnitude of the index increment, the performance of many devices often depends on the losses incurred within the waveguide. As such, these losses need to be accurately measured. This is the subject of the next section.



4.3 Waveguide Loss


Waveguide loss is an important characteristic affecting device performance. It is usually desirable to have as low loss as possible, in order to minimize the fiber-to-fiber insertion loss of devices, and maximize gain in waveguide lasers. In this section, the techniques used to measure both total insertion loss and propagation loss are described and some measurement results are provided.

4.3.1 Loss Measurement Setup

The experimental setup used to measure both total insertion loss and propagation loss is depicted in Fig. 4.6. Single-mode, polarization-maintaining fiber from an HP


Chopper Lock-In DUT Amplifier HP 8168A IR
Tunable Laser SM Fiber Detector 20x Polarizer
lens

Fig. 4.6 Experimental setup used for measuring waveguide loss.






82

8168A tunable laser source is used to couple light into the waveguide. A 20x lens is used at the waveguide output to collect the light and focus it onto an infrared Ge detector. A polarizer is used in order to insure the detection of only the polarization of interest, from both the sample and the fiber output, and a chopper is used, in conjunction with a lock-in amplifier, to read the detector voltage with as low signal-to-noise ratio as possible.


4.3.2 Total Insertion Loss

To measure total insertion loss, the fiber output is first measured, with no sample in the setup. Then the sample is introduced into the setup and the waveguide output is measured in a similar manner. The difference between the two measurements is the total insertion loss of the waveguide sample.

For most of the waveguide samples used in this research, the total insertion loss was around 2.5-3.0dB. The exact value of insertion loss depended on waveguide width, sample length, fabrication technique and conditions, quality of the endface polish, and quality of the cleaved face of the fiber.

As a best result, the fiber-to-fiber insertion loss of a 5gm wide waveguide fabricated by PE in glycerin at 2600C for 24 hours was measured. To do this, the lens at the output face of the waveguide was replaced with another single-mode fiber and the measurement was made in a similar manner as described above, using a lightwave multimeter to measure optical throughput instead of the Ge detector and lock-in amplifier. The fiberto-fiber insertion loss for this sample was around 3dB.






83


190 ~ I I

180 170

a 160 .C 150 2 140 A 130

120

110 I I , I
1500 1500.02 1500.04 1500.06 1500.08 1500.1 wavelength (nm)

Fig. 4.7 Fabry-Perot transmission response of a 5pim wide PE channel
waveguide.


4.3.3 Propagation Loss

Propagation loss was measured using the waveguide Fabry-Perot resonator technique [Reg85]. The setup is the same as that depicted in Fig. 4.6. Here, however, the sample itself acts as a Fabry-Perot resonator. The measurement is made by changing the wavelength of the laser source and monitoring the measured optical output from the waveguide. The output as a function of wavelength will have a typical Fabry-Perot response, resembling somewhat a sinusoidal response, depending on the amount of propagation loss. By measuring the ratio of the maximum transmission to the minimum transmission, and by knowing the reflectivity of the sample endfaces, the propagation loss is easily calculated [Reg85].

The measured Fabry-Perot transmission response of a waveguide sample prepared by PE in glycerin is shown in Fig. 4.7. The value of propagation loss calculated from this






84

measurement is 0.24dB/cm, comparable to values reported for APE waveguides in LiNbO3 [Hus95]. For the case of Zn vapor-indiffused waveguides, a propagation loss of 0.75dB/cm was measured, comparable to that previously reported for these waveguides in LiTaO3 [Ekn87].

Along with loss, another important characteristic of waveguides is the value of the electro-optic coefficient r33. In the next section, the method of measuring r33 is outlined and performed.



4.4 r33 Measurement


For the fabrication of electro-optic devices, such as modulators, it is desirable to have a large electro-optic coefficient r33. However, the fabrication techniques used here are known to decrease somewhat the value of r33. As stated earlier, PE, especially in pyrophosphoric acid, is known to decrease r33, though it can be recovered to some degree by annealing. The r33 value of waveguides fabricated in lower H+ concentration sources, such as glycerin, where annealing is not used, is not known. Additionally, the techniques of Ti-indiffusion and Zn vapor-indiffusion are performed at temperatures above the Curie point for LiTaO3, so there is no electro-optic effect after waveguide formation. As a result, repoling is needed to restore r33. A direct measurement of r33 is then needed to gauge the effectiveness of the repoling process.

In this section, the technique used to measure the electro-optic coefficient r33 is described and administered. The advantage of the technique used here over other methods is outlined and some results are presented.






85

4.4.1 Interferometric Measurement Method

Most efforts to directly measure r33 in channel waveguides involve the use of an interferometric device, such as a Mach-Zehnder interferometer (MZI), and the application of an electrode pattern to induce a phase shift. For example, the r33 coefficient for an MZI in push-pull configuration is given by: r33 - (4-6) 27cfVLn3
where X is the wavelength, g is the electrode gap, 0 is the induced phase shift, F is the optical-electrical overlap factor, V is the applied voltage, L is the electrode interaction length, and n is the refractive index. The accuracy of this technique depends greatly on the accuracy in determining the value of F, which can only be estimated based on electrode dimensions. However, there is another method which ensures an overlap factor of unity. This method is described next.


4.4.2 Unity Overlap Method

The setup used here to measure r33 is depicted in Fig. 4.8. The polarization-maintaining fiber from a diode laser is connected to a polarization-maintaining fiber coupler. It was a 3dB coupler for the wavelength used, 1.55gm. One output from the coupler is used 20x
lens

diode lasecamera

fiber 3dB DUT beam splitter coupler





video monitor

Fig. 4.8 Experimental setup used to measure the electro-optic coefficient r33.






86

to couple light into a waveguide sample. Then the outputs of both the other coupler arm and the waveguide are imaged, on top of each other, so they interfere on an IR camera. The interference pattern is seen on a video monitor. The waveguide sample had Ta deposited on both the waveguide surface, with a SiO2 buffer layer to reduce loss, and the bottom surface. The Ta served as electrodes and guaranteed an overlap factor of unity. Of course, the limitation of this technique is that it can only be used on Z-cut samples.

To make the measurement, a voltage was applied to the electrodes and the fringe pattern was observed on the monitor. The voltage V required to shift the fringe pattern by one-half period ((=ic) was recorded and the r33 value was calculated from: r33 - g (4-7) VLn
where g is the sample thickness and L is the sample length. The accuracy of this technique was estimated to be �6%.

For a waveguide fabricated by PE in glycerin at 260C for 24 hours, the measured r33 value was 14.5 pm/V. For a Zn vapor-indiffused sample which was repoled, the measured r33 value was 27. lpm/V. This value is close to the r33 value of 30pm/V for bulk, virgin LiTaO3.

In the next section, the issue of power handling of LiTaO3 waveguides is examined. In particular, the photorefractive characteristic of PE waveguides in the infrared, near 1.5jm, is measured for similar structures in both LiTaO3 and LiNbO3.



4.5 Power Handling


LiTaO3 is known to have a threshold for photorefractive damage of more than an order of magnitude larger than that of LiNbO3 at visible wavelengths [Gla72]. In the






87

infrared, however, although photorefractive damage is often considered negligible, damage has been observed at 1.3gm in Ti:LiNbO3 for power levels around 5mW [Har86]. The photorefractive resistance of these waveguides, as well as those fabricated by PE in both LiNbO3 and LiTaO3, at 1.55gm have not been investigated to date.

In this section, the power handling capability of PE waveguides in LiTaO3 operating near 1.5gm is examined and compared to similar structures in LiNbO3. A description of the measurement setup is given first, followed by the measurement results.


4.5.1 Setup Description

Straight-channel waveguides of width 6gm were fabricated by PE in X-cut LiTaO3 using glycerin at 260C for 24 hours and by APE in X-cut LiNbO3 using pyrophosphoric acid at 200C for 50 minutes followed by annealing at 350C for 7 hours with a 2 hour ramp period at each end. To provide high input intensities needed for the measurement, a FiberRaman Laser at 1.48gm was used. Light from the laser was coupled into the waveguides using a single-mode fiber. Waveguide outputs were collected with a 20x lens and focused onto a thermal detector with calibrated power meter. A pinhole was placed in front of the detector and closed around the mode spot at low intensities so as to not inadvertently collect the light from the expanded mode of a waveguide exhibiting photorefractive damage at higher intensities. The output power versus input power was then plotted to check for photorefractive damage.

4.5.2 Results

Fig 4.9 shows the results of photorefractive measurements performed on waveguides of similar channel width in LiTaO3 and LiNbO3. As can be seen, the LiTaO3 sample, shows a reasonable degree of linearity between output and input power over the






88

140 range tested. Therefore, little, if any, degra120
PE:LiTaO3 dation to its performance has been caused 100
80 ,* by photorefractive damage. The LiNbO3 o 60

20 .- .
o 0 . saturation and appears to have already
0 100 200 300 400 500
input power (mW) passed its threshold point for photorefracFig. 4.9. Output power vs. input power tive damage for even the smallest input for 6pm wide straight-channel waveguides in PE:LiTaO3 power tested. Lower levels of input power
and APE:LiNbO3.
were not attainable as they were below the lasing threshold of the laser used in this experiment. The significant attenuation of output power relative to input power illustrates the extremely high photo-induced refractive index change in APE:LiNbO3 waveguides, especially above 100mW.

In the next section, a direct method of index profiling is described. Such a measurement technique is necessary to extract surface index values of waveguide in order to reconstruct the structural phase diagram for APE:LiTaO3 in Chapter 5.



4.6 Direct Index Profiling


In order for devices to take advantage of the high throughput capability offered by LiTaO3, the technology of reliable waveguide fabrication must first be developed. For the APE technique of waveguide fabrication in particular, however, a number of anomalies are known to exist, as was outlined in Chapter 2. In order to explain these anomalies, as well as identify conditions yielding stable waveguides, the structural phase diagram for






89

APE:LiTaO3 is needed. This diagram relates waveguide surface index increment to proton-induced lattice strain at the surface.

In this section, the measurement technique used to extract waveguide index profiles is described. The need for a direct measurement technique is discussed and the experimental setup used to directly measure index profiles is detailed. Then the accuracy of this setup is given and some measured results are presented.


4.6.1 The Importance of Direct Profiling

It is the waveguide index profile that determines the most important characteristics of device performance, including insertion loss, modulation efficiency, nonlinear conversion efficiency, etc. Moreover, from the index profile the value of surface index increment can be obtained, which is needed to reconstruct the structural phase diagram for APE:LiTaO3. In this regard, previous techniques, in most instances, have not attacked the problem directly. Instead, the bulk of previous effort has attempted to approach this problem by analyzing waveguides with indirect techniques. Many of these methods are flawed since the results were only of qualitative nature, due to the assumptions made on the properties of the waveguide index profiles. In particular, a vast majority of the results on temperature-dependent diffusion coefficients have been obtained by using the inverse WKB (IWKB) procedure and, hence, were affected by the fundamental assumption of a monotonically decreasing depth profile with the maximum reached at the substrate surface. On the contrary, recent results have directly demonstrated [Mar96a] the existence of buried profiles where the index profile maximum is reached at a point below the substrate surface. This fact makes the accuracy of the previous results less trustworthy, to say the least. Additionally, even in the case of monotonically decreasing index profiles,






90

0.03 .. there still remains the problem as to
0.025 ..... which profile distribution to apply to
0.02
an IWKB approximation to determine
0.015
0.01 surface index values, i.e., an
0.005
0.005 exponential function, a
0
-0.005 ,, ,,,,, ,,,, complementary error function, a
0 1 2 3 4 5
depth (pm) Gaussian function, or a Fermi Fig. 4.10 Comparison of a directly function. Since many of these
observed index profile (solid)
and an IWKB reconstructed distributions have been reported in the
profile (dotted).
literature for various fabrication

conditions in APE:LiTaO3, application of an unsuitable function to a set of measured mode indices can result in significant error in calculating the surface value of the waveguide index [Mu94]. Such an example is depicted in Fig. 4.10, where the measured effective mode indices and associated WKB reconstruction (dotted line) are compared with the directly measured index profile (solid line). The measured mode indices are within measurement error for the prism coupling technique, however the IWKB reconstruction method could not accurately predict the plateau (almost) region near the substrate surface and hence seriously overestimated the surface index increment.

Finally, the IWKB method requires at least three guided modes and, hence, is not applicable at all to the case of single-mode waveguides. Clearly, extrapolation of the multi-mode waveguide data, whose accuracy is rather questionable, to single-mode structures is even more doubtful. On the other hand, the single-mode regime of operation




Full Text
110
5.3 Analyzing the Structural Phase Diagram
Having reconstructed the structural phase diagram, it can be inspected and used to
explain several of the aforementioned anomalies associated with the APE process in
LiTa03. That is done in this section. In addition, waveguide index profiles were examined
in different region, or phases, of the structural phase diagram in order to gauge the influ
ence of strain on waveguide index profiles.
5.3.1 Explaining APE Anomalies
Analysis of the structural phase diagram lends an explanation for the aforemen
tioned anomaly of reduced index increment with increasing proton concentration
[Mar98a]. As proton concentration is increased via proton exchange, strain and index
increase through the first four phases. However, once the concentration is high enough to
produce a deformation e"3 ~ 4, further increase of protons and strain leads to a decrease in
index. The anomaly of increased index increment upon short annealing can also be
explained [Mar98a]. If a single-step proton exchange process is sufficient enough to sup
ply the required proton concentration and induced strain to form waveguides largely
within the [33 or (34 phase, then subsequent annealing will decrease strain, and hence
increase index.
5.3.2 Effect of Crystal Phases on Index Profiles
A unique perspective of how the concentration induced strain impacts the
waveguide index profile can be obtained by tracking the evolution of the index profile and
corresponding rocking curves as we move through the structural phase diagram in the
direction of decreasing e"3 [Mar98b], Typical examples of index profiles are illustrated in


113
As further annealing is performed and strains are relieved, the presence of the (32
phase becomes more pronounced on rocking curves and index variations below the surface
are less noticeable. At the same time, the surface value of index rises until the profile is no
longer buried. Fig. 5.3(d) shows such a waveguide with a surface e"3 value of 4.06, at the
low-concentration end of the (33 phase. Waveguides in this region have a characteristically
Fermi-like index distribution.
Lastly, Fig. 5.3(e) depicts a waveguide profile with a surface e"3 value of 3.84.
This Gaussian-like index distribution is characteristic of the (31 and (32 phases, as well as
the k and a phases. Rocking curve analysis for waveguides within the (32, (31, and k
phases reveal only the crystal phase at the surface, with a soft shoulder representing the
lower-concentration phases present. On the other hand, waveguides within the a phase are
truly single-phase.
In the next section, PE waveguides are analyzed by Raman spectroscopy and direct
measurement techniques. As a result, a new phase is identified and the structural phase
diagram is updated to include it.
5.4 PE Waveguides
As will be discussed later in Chapter 6, the APE technique had some limitations,
particularly when attempting to fabricate single-mode waveguides at 1,55pm which are of
low concentration, or within the a phase [Mar99]. As a result, PE only using a dilute
source was used to overcome these limitations. The details leading up to such a transition
will be discussed in the next chapter, however, the analysis of PE waveguides is presented
in this section because of their significance in the structural phase diagram for LiTa03.


29
deposit PR
expose to UV
and develop
deposit Ta
lift-off
Fig. 2.8 Illustration of the photolithographic process up
through lift-off.


APE:LiTa03 is needed. This diagram relates waveguide surface index increment to
proton-induced lattice strain at the surface.
In this section, the measurement technique used to extract waveguide index pro
files is described. The need for a direct measurement technique is discussed and the
experimental setup used to directly measure index profiles is detailed. Then the accuracy
of this setup is given and some measured results are presented.
4.6.1 The Importance of Direct Profiling
It is the waveguide index profile that determines the most important characteristics
of device performance, including insertion loss, modulation efficiency, nonlinear
conversion efficiency, etc. Moreover, from the index profile the value of surface index
increment can be obtained, which is needed to reconstruct the structural phase diagram for
APE:LiTa03. In this regard, previous techniques, in most instances, have not attacked the
problem directly. Instead, the bulk of previous effort has attempted to approach this
problem by analyzing waveguides with indirect techniques. Many of these methods are
flawed since the results were only of qualitative nature, due to the assumptions made on
the properties of the waveguide index profiles. In particular, a vast majority of the results
on temperature-dependent diffusion coefficients have been obtained by using the inverse
WKB (IWKB) procedure and, hence, were affected by the fundamental assumption of a
monotonically decreasing depth profile with the maximum reached at the substrate
surface. On the contrary, recent results have directly demonstrated [Mar96a] the existence
of buried profiles where the index profile maximum is reached at a point below the
substrate surface. This fact makes the accuracy of the previous results less trustworthy, to
say the least. Additionally, even in the case of monotonically decreasing index profiles,


64
Fig. 3.9 Stokes fluorescence spectra at various depths d below the surface.
to the Er3+ concentration for pump intensities below 103 W/cm2, it was possible to deter
mine that the Er3+ concentration profile is exponential-like with a 1/e depth of around 5-
6p.m (Fig. 3.9). The intensity profile of the ASF emission had, however, a depth of around
2p.m. In other words, the ASF emissions emanated from the regions of highest Er3+ con
centration (near the surface), where most of the Er3+ clusters resided. This was direct evi
dence of the dominating role of nonradiative energy transfer between clustered Er3+ ions
in the mechanism of reconversion in Er-doped LiTa03.
Since Er3+ clustering was determined to be the primary factor leading to upconver-
sion, significant reduction of the upconversion efficiency in Er:LiTa03 was expected to be
achieved by a post-indiffusion process to reduce Er3+ clustering. This is pursued in the
next section.


131
Fig. 6.7 Measured coupling length of Sam
ple 3. This coupler was comprised
of 7(im channels with an 8(im gap.
The stability of these new waveguides
is examined next.
6.3.3 PE Results
The measured change in cou
pling length for Sample 3 is shown in
Fig. 6.7. The fabrication conditions
and estimated change in index incre
ment 8(An) for this, and all the PE
samples tested, are given in Table 6.3.
This sample was in the a' phase, having a strain value of e"3=4xl0 less than that of
Sample 1 but greater than that of Sample 2. As can be seen from Table 6.3, just 7 days
after fabrication, the index increment for this sample decreased by 8(An)=7xlO"4. Again
compared to the previous samples, this change is greater than that of Sample 2 and
roughly similar to that of Sample 1 for the same time period.
It was evident after just 7 days that Sample 3 would not be stable enough, so in an
effort to improve its stability, it was annealed at 230C for 3 hours. It was then denoted as
Sample 3' (Table 6.3). The measured change in coupling length for this sample is shown
in Fig. 6.8. It was seen from this figure and Table 6.3, that the effect of annealing was to
reduce the coupling length from about 22mm to about 13mm, though the measured value
of strain did not change. Additionally, after 126 days, the change in index increment was
8(An)=1.7xlO~3, larger than that of Sample 2 and less than that of Sample 1 for roughly the
same time period.


50
lasers and amplifiers. However, much research awaits to be performed on the
characterization and effect of rare-earth doping, particularly Er, in LiTa03.
In this chapter, the effect of Er-doping in LiTa03 is investigated. Some back
ground material relating to work done in this connection on LiNb03 is presented, as well
as the limitations of LiNb03, leading to the motivation to examine Er-doped LiTa03. The
conditions used for doping of LiTa03 are then given, followed by Raman and fluorescence
measurements of samples identifying Er3+ clustering as the dominant gain limiting factor.
A novel method of reducing clustering is then presented and applied. Finally, a two-pho
ton model is used to exemplify the photorefractive process in these crystals and the inher
ent advantage of LiTa03 for optical data storage and laser applications is outlined.
3.1 Background
Doping LiNb03 and LiTa03 with rare-earth materials, such as Nd or Er, is known
to cause laser emissions when optically pumped [Ami96, Bau96, Nou95]. Much work has
been done to demonstrate the lasing effect of Nd-doping and Er-doping in LiNb03
[Ami96,Bau96]. Doping with Nd is known to produce emissions near 1p.m. Er, on the
other hand, is known to produce emissions near 1.5pm, making it more popular because of
its potential application in fiber networks which operate in the 1,5pm region.
There are currently two pump wavelengths which correspond to an electronic tran
sition of Er3+ ions and are used for the pumping of Er-doped devices. They are 980nm
and 1480nm. For the case of waveguide lasers in LiNb03, however, pumping at 1480nm
has some advantages. First, the photorefractive effect is smaller at 1480nm than at 980nm,
allowing for the application of higher pump powers. Second, waveguides fabricated to be


19
which narrow stripe patterns are delineated by photolithography, exposing regions of the
substrate where the exchange is to take place. In the case of planar waveguides, no mask
is needed and the entire bulk surface is subjected to proton exchange. The substrate is
then immersed in a hot melt of high H+ concentration, usually benzoic or pyrophosphoric
acid, where Li+ diffuses out of the substrate and is replaced by H+ from the acid source.
As a result, an exchanged layer of HxLii_xTa03 (HxL!_xNb03 if LiNb03 is used
as the substrate) is formed at the surface with the thickness dependent on the melt temper
ature T and exchange time t. On the other hand, the refractive index change is practically
independent of T and t, being determined largely by the choice of the proton source (acid).
Pyrophosphoric acid is normally preferred as the H+ source since it has a larger diffusion
coefficient D(t) and results in an increase in the extraordinary index Ane which is 15%
higher than what is attainable with benzoic acid [Ahl94c, Kan94], In addition, pyrophos
phoric acid does not boil at typical exchange temperatures (it is liquid up to 300C) and
has a low vapor pressure, allowing work at high temperatures where diffusion is rapid.
The higher H+ concentration also allows for more control over the diffusion depth and
uniformity of the exchanged layer than is possible with benzoic acid [Miz92], However,
when using pyrophosphoric acid, Ta must be used as the mask since the pyrophosphoric
acid will attack and cause pitting in an A1 mask.
The APE process in LiTa03 results in an increase in the extraordinary index only,
allowing for the propagation of solely TM modes in a z-cut crystal, as illustrated in Fig.
2.4. For the case of x and y-cut crystals, only TE modes will propagate. Additional fea
tures characteristic of the APE process in LiTa03 are described in the next section.


waveguide index profiles was examined. Fourth, the advantages of PE waveguides over
APE waveguides were detailed, showing better optical mode confinement and the exist
ence of a new single-phase region on the structural phase diagram, obtainable only by PE
in a dilute source. Finally, these PE waveguides were shown to have superior photorefrac-
tive resistance near 1.5|J.m than their counterparts in LiNb03.
On the issue of temporal stability of waveguides in LiTa03, a number of accom
plishments were made. First, rocking curve and directional coupler measurements were
used to show that APE waveguides in high concentration phases are very unstable. Sec
ond, by comparison of low concentration APE and single-phase PE waveguides, it was
found that stability was not improved by fabricating waveguides comprised of only a sin
gle phase, as is the case with LiNb03 and was assumed to be the case with LiTa03.
Rather, stability appeared to be improved by simply reducing strain, or proton concentra
tion. However, there was a limit to which proton concentration can be reduced, especially
at 1.55|im because of the extremely low values of index increment at this wavelength.
Third, a comparison of the stability of PE, Ti-indiffused, and Zn vapor-indiffused
waveguides, where all were found to be unstable, showed that instability does not origi
nate from proton-induced lattice strain or ion mobility and, unlike LiNb03, does not
depend on the fabrication technique. It is likely a function of the crystal. Fourth, crystals
from several growers, of both SAW grade and those claimed to be of optical grade, were
compared to reveal no distinction between grades or growers, showing them all to be
unstable. As a result, it would appear that the only hindrance standing in the way of the
deployment of LiTa03 devices is crystal instability, a result of a less than optimum quality
crystal, probably due to its high growth temperature, where lack of significant demand for


138
6.4.1 Directional Coupler Stability
Results of Zn:LiTaQ3 Waveguides
The measured change in
coupling length for a Zn vapor-indif-
fused directional coupler, whose fabri
cation conditions were given in
Chapter 2, is shown in Fig. 6.12. The
measurements were made at 1.31|im,
starting two days after fabrication.
Immediately after fabrication, the
waveguides comprising the coupler were single-mode at 1.55p.m. However, after two
days they did not guide at all at 1.55p.m. Parameters about the index profile were taken
from Refs. [Ekn87, Y0088], which used identical fabrication conditions, and were used in
the computer simulation to estimate a 8(An) of about 1CT3 from day 2 to day 41. This is
comparable to that seen in PE waveguides, though it does not include the apparently large
change from day 1 to day 2.
6.4.2 Comparison of the Waveguide Formation Processes
As can be seen from the above results, metal-indiffusion appears to also be unsta
ble in LiTa03 [MarOO]. Without getting into a quantitative comparison between 8(An)s of
the different fabrication processes used, a simple comparison was made by tracking the
change in the single-mode region of operation. To do this, straight channel waveguides of
widths 2-10pm, in steps of 0.5pm, were fabricated by Ti-indiffusion, Zn vapor-indiffu
sion, and PE using pure glycerin. The modal analysis technique of Section 4.1 was
1 1
I T I I | 11
-
r-r-t-T
10.5
I n
10
9.5
n
. 1
J0 10
20
I I I I I I
JQ_
days
Fig. 6.12 Measured coupling length of
the Zn vapor-indiffused sample.
This coupler was comprised of
6.5pm channels with an 8pm
gap-


144
Crystals from different growers around the world were then tested and all showed the
same degree of instability. Additionally, there was no distinction between crystals of SAW
grade and those claimed to be of optical grade. These results suggested that the instability
with LiTa03 is the result of an immature growth process, possibly the result of high strain
values or defect sites within the crystal boule related to its Li deficiency due to the
extremely high growth temperature required.
In the next chapter, the potential for LiTa03 in the realm of integrated-optical com
ponents is explored. For this, the design, fabrication, and testing of a high-speed, travel
ing-wave electro-optic modulator is performed. The development and demonstration of
such a device illustrates the potential for LiTa03 as a viable host material, offering the
advantage of higher throughput capability, pending the improvement of the crystal through
further development of the growth process.


76
Fig. 4.3 Schematic of the standard prism coupling method used to selec
tively couple light into individual waveguide modes.
on the coupling of light from the prism to the waveguide by satisfying the phase matching
condition for each mode within the waveguide separately.
Light incident on the face (hypotenuse) of the prism at an angle 0¡, is refracted as it
enters the prism according to Snells law:
csin0. = (4-1)
The angle of incidence of the incoming beam is adjusted so as to maintain total internal
reflection at the np-nc interface. This condition is satisfied if the angle 02 is greater than
the critical angle:
0O > 0 = asin
L c
in ^
n
V PJ
(4-2)
The coupling of light into modes of the waveguide can occur only if the horizontal compo
nents of the k vector in the prism matches the propagation constant of the waveguide. This
is referred to as the phase matching condition:


69
gating ratio (i.e. the ratio of PR sensitivity with and without the second color) and, there
fore, can be implemented even with low-power sources (e.g., 0.1 W/cm2 [Lan98] 1 W/
cm2 [Gue97]). It means that the secondary green emission of Er3+ clusters has enough
photon energy and intensity IASF to effect, together with the pumping red light, significant
photorefractivity via the two-color excitation.
The difference in excitation of green emission observed between pumping at 785
and 632.8 nm is consistent with previous data [G196] on the spectral dependence of up-
conversion excitation within Er3+ clusters. Such an upconversion is an undesirable effect
in Er-doped devices [Ami96, Gil96, Hua96] operating at telecommunication wavelengths.
Experimental data reported here demonstrates that this upconversion induces a more pro
nounced PR effect in LiNb03 than in LiTa03. This is because the green emission of Er3+
clusters is within the maximum of the bipolaronic absorption band in LiNb03 [Sch91] yet
is far from the maximum of the absorption band in LiTa03 [Kos97b], This fact makes
LiTa03 more attractive for the fabrication of Er-doped waveguide amplifiers. At the same
time, data presented here also points to the advantage of LiTa03 over LiNb03 among per
spective rear-earth-doped PR materials for non-volatile holographic storage with two-pho
ton recording. At an optimal wavelength of extrinsic gating light (ranging from 280 to 370
nm in LiTa03 [Kos97b]), a saturated value of holographic efficiency should be limited
mainly by the intrinsic defects concentration N, having a higher value (1.4 times) in
LiTa03 [Bor95, Kos99]. Because the gating beam generally does not have to be from a
coherent light source [Gue97], low-cost light sources such as filtered xenon and mercury
lamps can be used for this purpose in Er:LiTa03.


TABLE OF CONTENTS
page
ACKNOWLEDGMENTS iii
ABSTRACT ix
CHAPTERS
1 INTRODUCTION 1
1.1 Integrated-Optical Devices and Ferroelectric Materials 2
1.2 The Advantages of LiTa03 4
1.3 Fabrication Processes in LiTa03 6
1.4 Motivation and Objectives 9
1.5 Chapter Organization 11
2 FABRICATION OF PLANAR AND CHANNEL
WAVEGUIDES IN LiTa03 15
2.1 Waveguide Formation by the APE Process 16
2.1.1 The PE/APE Process 16
2.1.2 Features of the APE Process in LiTa03 20
2.1.3 Anomalies of APE Waveguides 23
2.1.4 APE Waveguide Fabrication Parameters 26
2.2 Waveguide Formation by Metal/Vapor-Indiffusion 30
2.2.1 Ti Metal-Indiffusion 31
2.2.2 Zn Vapor-Indiffusion 33
2.3 Crystal Repoling 34
2.4 End-Face Polishing of Channel Waveguides 36
2.5 The Bevel Polish for Planar Waveguides 37
2.6 Antireflective Coatings 39
2.6.1 Thin Film Theory for Single-Layer AR Coatings 39
2.6.2 The AR Film Parameters and Deposition 44
2.7 Electroplating 45
2.8 Summary 47


173
[Gue97] H. Guenter, G. Wittmann, R.M. Macfarlane, and R.R. Neurgaonkar, Intensity
Dependence and White-Light Gating of Two-Color Photorefractive Gratings in
LiNb03 Opt. Lett., vol. 22, p. 1305, 1997.
[Har86] G. T. Harvey, G. Astfalk, A. Y. Feldblum, and B. Kassahun, The Photorefrac
tive Effect in Titanium Indiffused Lithium Niobate Optical Directional Cou
plers at 1.3|J.m, IEEE J. Quantum Electron., vol. QE-22, no. 6, p. 939, 1986.
[Hol84] R. J. Holmes and D. M. Smyth, Titanium Diffusion into LiNb03 as a Func
tion of Stoichiometry, J. Appl. Phys., vol. 55, no. 10, p. 3531, 1984.
[How89] P. H. Howerton, W. K. Burns, and R. P. Moeller, Integrated-Optical Mode
Converter/Frequency Shifter in Ti:LiTa03, IGWO89, WCC3-1, p. 250, 1989.
[Hua96] C.-H. Huang, and L. McCaughan, 980-nm-Pumped Er-Doped LiNb03
Waveguide Amplifiers: A comparison with 1484-nm pumping, IEEE J. Sel.
Top. Quantum Electron., vol. 2, p. 367, 1996.
[Huk98] J. Hukriede, I. Nee, D. Kip, and E. Kratzig, Thermally Fixed Reflection Grat
ings for Infrared Light in LiNb03:Ti:Fe Channel Waveguides, Opt. Lett., vol.
23, no. 17, p. 1405, 1998.
[Hus95] C. P. Hussell, A Novel Tunable Filter for Wavelength Division Multiplexed
Communication Systems, Ph.D. Dissertation, University of Florida, Gaines
ville, 1995.
[INS89] Properties of Lithium Niobate, EMIS Datareviews Series No. 5, INSPEC
Pub., London, 1989.
[Jac82] J. L. Jackel, C. E. Rice, and J. J. Vesselka, Proton-Exchange for High Index
Waveguides in LiNb03, Appl. Phys. Lett., vol. 41, p. 607, 1982.
[Jer95] F. Jermann, M. Simon, and E. Kratzig, Photorefractive Properties of Congru
ent and Stoichiometric Lithium Niobate at High Intensities, J. Opt. Soc. Am.
B, vol. 12, p. 2066, 1995.
[Jun90] H.-S. Jung, H. F. Taylor, and O. Eknoyan, Interferometric Polarization-Inde
pendent Modulator in LiTa03, J. Lightwave Technoi, vol. 8, no. 10, p. 1452,
1990.
[Kam73] I. P. Kaminow and J. R. Carruthers, Optical Waveguiding Layers in LiNb03
and LiTa03, Appl. Phys. Lett., vol. 22, no. 7, p. 326, 1973.
[Kan94] D. Kan and G. L. Yip, Annealed Proton-Exchanged Lithium Tantalate
Waveguides Fabricated in Concentrated and Diluted Pyrophosphoric Acid,
GRIN, paper C-25, 1994.


CHAPTER 1
INTRODUCTION
A discussion leading to the motivation for this dissertation is presented in this
chapter. Ever increasing demand for bandwidth in telephone, computer network, and
internet applications has resulted in the ever increasing demand worldwide for fiber net
works, and explosive growth in the associated integrated-optical devices for these net
works such as lasers, filters, and high-speed modulators. The demands imposed by these
devices are considered before presenting possible materials to be used in their fabrication.
Among the suitable materials, consideration of their properties reveal that lithium tantalate
(LiTa03) possesses a number of distinct advantages, particularly a higher optical damage
threshold and a shorter wavelength ultra-violet (UV) absorption edge.
Use of LiTa03 as a host for the fabrication of integrated-optical devices faces a
number obstacles. For instance, the characteristics of Er-indiffusion for laser fabrication
have not yet been studied in this material. Additionally, much work needs to be done on
the characterization of different waveguide fabrication techniques. The most common
technique, annealed proton exchange, is known to exhibit a number of anomalies prevent
ing it from being used in the production of reliable devices, the most significant anomaly
being the instability of waveguide index profiles over time. As for the other methods of
waveguide formation, no data characterizing relevant issues associated with each process,
such as stability, has been reported.


£33=11-4
£33 =6.03
£33=5-16
(c)
£33=4.06
£33=3.84
Fig. 5.3 Measured refractive index profiles for e'33 = 11.4, near the high-concen
tration end of the [34 phase (a), e"3 =6.03, near the low-concentration end
of the [34 phase (b), e'33 =5.16, near the high-concentration end of the [33
phase (c), e'33 =4.06, near the low-concentration end of the [33 phase (d),
and e'33 =3.84, in the [31-[32 phase region (e).


56
Fig. 3.2 Fluorescence spectra of the red ASF emission for the Er-doped
sample of Fig. 3.1 (solid) and Raman spectra of virgin LiTa03
(dashed) excited at X=647nm and J=5xl04 W/cm2. R denotes
Raman lines and L denotes the excitation laser line after a notch
filter.
Fig. 3.3 Intensity of the green ASF emission
versus input intensity J for excitation
at 632.8nm (circles) and 647nm
(squares) for the sample shown in Fig.
3.1.


172
[Die98] V. Dierolf, A.B. Kutsenko, F. Tallian, and W. von der Osten, in EURODIM-98,
8th Europhysical Conference on Defects in Insulating Materials, Abstracts,
Keele University, Keele, Staffs, UK, p.136, July, 1998.
[Dou89] J. C. Doukhan, P. Cordier, and N. Doukhan, Lattice Defects in Lithium Tanta-
late, 43rd Symposium on Frequency Control, p.497, 1989.
[Ekn87] O. Eknoyan, D. W. Yoon, and H. F. Taylor, Low-Loss Optical Waveguides in
Lithium Tantalate by Vapor Diffusion, Appl. Phys. Lett., vol. 51, no. 6, p. 384,
1987.
[E1H95] K. El Hadi, P. Baldi, S. Nouh, M. P. De Micheli, A.Leycuras, V.A.Fedorov,
Yu.N.Korkishko, Control of Proton Exchange for LiTa03 Waveguides and
Crystal Structure of HxTai_xNb03, Optics Letters, vol. 20, no. 16, p. 1698,
1995.
[E1H98] K. El Hadi, V. Rastogi, M. R. Shenoy, K. Thyagarajan, M. De Micheli, and D.
B. Ostrowsky, Spectral Measurement of the Film-Substrate Index Difference
in Proton-Exchanged LiNb03 Waveguides, Appl Opt., vol. 37, no. 27, p.
6463, 1998.
[Fed94] V. A. Federov and Y. N. Korkishko, Crystal Structure and Optical Properties
of Proton-Exchanged LiTa03 Waveguides, Ferroelectrics, vol. 160, p. 185,
1994.
[Fin88] T. Findakly, P. Suchoski, and F. Leonberger, High-Quality LiTa03 Integrated-
Optical Waveguides and Devices Fabricated by the Annealed-Proton-Exchange
Technique, Opt. Lett., vol. 13, no. 9, p. 797, 1988.
[Fuj93] T. Fujiwara, R. S. Srivastava, X. Cao, and R. V. Ramaswamy, Comparison of
Photorefractive Index Change in Proton-Exchanged and Ti-Indiffused LiNb03
Waveguides, Opt. Lett., vol. 18, no. 5, p. 346, 1993.
[Geo92] J. George, Preparation of Thin Films, Marcel Dekker, Inc. New York, 1992.
[Gil96] D. M. Gill, L.McCaughan, and J. C. Wright, Spectroscopic Site Determina
tions in Erbium-Doped Lithium Niobate, Phys. Rev. B, vol. 53, no. 5, p. 2334,
1996.
[Gla72] A. M. Glass, G. E. Peterson, and T. J. Negran, Optical Index Damage in Elec
trooptic Crystals, in Laser Induced Damage in Optical Materials, A. M. Glass
and A. H. Guenter, eds. (Washington, DC, Nat. Bur. Stand., 1972) Spec. Publ.
372, p. 15.
[Gop96] V. Gopalan and M. Gupta, Origin of Internal Field and Vizualization of 180
Domains in Congruent LiTa03 Crystals, J. Appl. Phys., vol. 80, p. 6099, 1996.


177
[Saw91] I. Sawaki and S. Kurimura, Second-Harmonic Generation in Periodically
Domain-Inverted Lithium Tantalate Channel Waveguides, CLEO, CTuV4, p.
166, 1991.
[Sch75] R. V. Schmidt and I. P. Kaminow, Acoustooptic Bragg Deflection in LiNb03
Ti-Diffused Waveguides, IEEE J. Quantum Electron., vol. QE-11, no. 1, p. 57,
1975.
[Sch91] O. F. Schirmer, O Thiemann, and M. Wohlecke, Defects in LiNb03-I. Experi
mental Aspects, J. Phys. Chem. Solids, vol. 52, no. 1, p. 185, 1991.
[Spi83] W. B. Spillman, Jr., N. A. Sanford, and R. A. Soref, Optical Waveguides in
LiTa03 Formed by Proton Exchange, Opt. Lett., vol. 8, no. 9, p. 497, 1983.
[Ste90] J. Steffen, A. Neyer, E. Voges, and N. Hecking, Refractive Index Profile Mea
surement Techniques by Reflectivity Profiling: Vidicon Imaging, Beam Scan
ning, and Sample Scanning, Appl. Opt., vol. 29, no. 30, p. 4468, 1990.
[Suz93] T. Suzuki and O. Eknoyan, Characterization of the Poling Process in Zinc
Diffused Lithium Tantalate Optical Waveguides, Ferroelectrics, vol. 145, p.
119, 1993.
[Sym92] R. Syms and J. Cozens, Optical Guided Waves and Devices, McGraw-Hill,
London, 1992.
[Tan77] G. L. Tangonan, M. K. Barnoski, J. F. Lotspeich, and A. Lee, High Optical
Power Capabilities of Ti-Indiffused LiTa03 Waveguide Modulator Structures,
Appl. Phys. Lett., vol. 30, no. 5, p. 238, 1977.
[Tan78] G. L. Tangonan, D. L. Persechini, J. F. Lotspeich, and M. K. Barnoski, Elec
trooptic Diffraction Modulation in Ti-Indiffused LiTa03, Appl. Opt., vol. 17,
no. 20, p. 3259, 1978.
[Tav94] R. Tavlykaev, K. Kiickelhaus, and E. Voges, Index Profile Reconstruction of
Ti:LiNb03 Structures and Bending Loss Evaluation from Near-Field Measure
ments,/. Opt. Comm., vol. 15, p. 71, 1994.
[Tav95] R. F. Tavlykaev, D. B. Maring, and R. V. Ramaswamy, Refractive-Index Pro
file of Annealed Proton-Exchanged LiTa03 Channel Waveguides from White-
Light Source Measurements, Integrated Photonic Research, vol. 7, 1995 OSA
Technical Digest Series, IThC5-l, p. 46.
[Ueh78] S. Uehara, Calibration of Optical Modulator Frequency Response with Appli
cation to Signal Level Control, Appl. Opt., vol. 17, no. 1, p. 68, 1978.


79
waveguide
Fig. 4.4 Modified prism coupling technique used to measure mode effec
tive index increments.
substituted into the derivative of the prism coupling equation, Eq. (4-4), with respect to the
coupling angle 6:
dN AN
P A(3
cos
A + asin
cosP
(4-5)
Since the measured angle A6 is the angle between the substrate and any mode of interest,
this new prism coupling equation gives the mode index increment AN=N-nc of this mode
above the substrate value. Because ABB and this equation is not sensitive to the value of
B, it can be evaluated at any angle inside AB. The accuracy of this new technique is esti
mated to be 10"4.
4.2.3 Measured Mode Index Increments
Using the modified prism coupling method, the mode index increments AN of
straight-channel APE waveguides in LiTa03 were measured at four different wavelengths:


22
where Ee is the activation energy for the proton exchange process, De0 is the exchange dif
fusion constant, and kB is Boltzmanns constant. Similarly, for the annealing process at
temperature Ta and duration ta, the following annealing diffusion coefficient Da(Ta) and
increase in depth da are found [Dav95]
(2-4)
(2-5)
where Ea is the activation energy for the annealing process. The total depth dT of the
waveguide fabricated by the APE process can then be approximated by [Dav95]
drj, d + d
1 e a
(2-6)
The profile of the index change in the proton exchanged region is known to be
step-like [Ahl94c, Kan94], The APE process in LiTa03 is known to result in an increase
of the extraordinary index and a decrease of the ordinary index. Therefore, only TM
modes will be supported in a Z-cut crystal. In the case of X and Y-cut crystals, only TE
modes propagate. Upon increasing the relative proton concentration x, the crystal structure
of the exchanged layer may transform into another crystal phase as has been indicated by
rocking curve measurements of exchanged samples [Fed94], These transformations are
reversible in the sense that the initial phase can be restored upon post-exchange annealing.
Post-exchange annealing at elevated temperatures (higher than that of exchange) then
causes further diffusion of H+, smoothing out the index profile and expanding the
waveguide area while decreasing the propagation loss [Miz92], Since there is no proton
source at the surface, the local concentration within the exchanged layer may decrease to


N = n sin09
P ^
(4-3)
77
where N is the normalized propagation constant, or effective mode index, of a particular
mode of the waveguide. In order to obtain real angles 02, the prism index np, therefore,
should be larger than N. For maximum coupling, the gap h between the prism and the
waveguide should be as small as possible. Usually it is about a quarter wavelength and
can be achieved through the use of clamps, pushing the prism up against the substrate sur
face. Additionally, the interaction distance d should be comparable to one coupling
length. If it is longer, light may be coupled back into the prism. Using simple geometry
within the prism and Snells law, it can be shown that the effective mode index N is related
to the incident angle by
N n^sin
A + asin
f sinG^
n
V P J
(4-4)
where A is the prism angle. From this equation, it is obvious that the effective mode index
for each mode within the waveguide can be calculated by accurately measuring the angle
of incidence of the incoming light which selectively excites each mode, and the angle A of
the prism. In addition, the value of np at the wavelength of interest must be known to an
accuracy of at least four decimal places, as will be discussed later.
An alternative to this is achieved by considering the path taken by the light for cou
pling into the waveguide, now in the reverse direction. In this case, light is first launched
into the waveguide structure, exciting all of the existing modes. Each of these modes will
have a different effective index N value and thus exit the hypotenuse face of the prism at a
different angle. The angular spectrum of the emerging light exhibits peaks, called m-lines,
corresponding to the modes of different order. The effective index values for each of the


42
Now, by using
B = -V = -CE = nJWoE
(2-11)
y0 = noJso^ocosBi
(2-12)
Vf = 'yV£<#ocose/i
(2-13)
ys = nsJeoHcos(li2
(2-14)
we can write the B-fields as
Ba ~ Yo(EiO Er0 ~ Yf(Et\ Ei0
(2-15)
Bb = ^f^Eil~Erl) = ^sEtl
(2-16)
The E-fields just below interface (a) and just above interface (b) are related through a
phase difference 8 caused by one traversal of the film. Using this fact, we may write
Ei 2 = Ene~1& (2-17)
Eil = Er2e~fi <2-|8>
where
8 = |^j/2yiCOS0rl (2-19)
Eqs. (2-17) and (2-18) can be used in the portions of Eqs. (2-8) and (2-16) representing the
fields within the film region (below (a) and above (b)) to solve for the fields Et] and E¡] in
terms of Eb and Bb. These new equations can, in turn, be used in the portions of Eqs. (2-7)
and (2-15) representing the fields within the film region to give


PROPERTIES AND CHARACTERISTICS OF LiTa03 FOR
INTEGRATED-OPTICAL DEVICE APPLICATIONS
By
DAVID BLAYNE MARING
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
2000


92
Fig. 4.11 Reflectivity profiling setup used for measuring the refractive
index profiles of APE waveguides in LiTa03.


41
Fig. 2.12 Diagram of the reflection from a single-layer thin film used to calcu
late the total reflectivity R of the film.


waveguide devices. The remaining phases showed no change in measured e"3, at least
within the measurement accuracy of this technique.
123
In the next section, a method of more accurately measuring and quantifying the
amount of instability, in terms of index change, is introduced. It involves the use of a
directional coupler to track changes in index increments by measuring corresponding
changes in coupling length.
6.2 Using a Directional Coupler to Measure Stability
In order to determine if fabrication conditions exist which yield temporally stable
waveguides at 1.55|J.m, a method of quantifying the amount of waveguide index change
over time was needed. To do this, a passive directional coupler was employed. Changes
in measured coupling length of the coupler over time were related to corresponding
changes in index increment through computer simulation. In this section, a brief descrip
tion of a directional coupler is provided. Then, the measurement technique used to deter
mine changes in index increment is outlined.
6.2.1 Directional Coupler Theory
A passive directional coupler is shown in Fig. 6.1, where w is the individual
waveguide width, g is the gap of the coupler region, and L is the interaction length of the
coupler section. The device should be designed so that each arm of the coupler is single
mode at the wavelength of interest, in this case 1.55|im, and supports the fundamental
symmetric and anti-symmetric supermodes. These modes are orthogonal to each other


First, the Raman spectra of samples fabricated in both pure and diluted glycerin will be
measured and compared to that of single-phase APE samples to determine phase content.
Next, surface values of index increment and strain are measured to plot the position of
these new samples on the structural phase diagram, indicating the presence of a new sin
gle-phase region obtainable only through PE with dilute sources.
5.4.1 Raman Analysis of PE Waveguides
Waveguides were fabricated in X-cut LiTa03 using PE only in a dilute source of
pure glycerin or glycerin diluted with Li2C03. The waveguides, which exhibited a rela
tively step-like index profile, were analyzed by Raman Spectroscopy for comparison of
phase content against APE waveguides. Fig. 5.4 shows the Raman spectra of samples
Raman shift, wavenumber (cm'1)
Fig. 5.4 Raman spectra of PE samples prepared in pure glycerin
at 260C for various exchange times and of an APE sam
ple with a strain value of e"3 =0.84x10-3, in the a phase.


3
to nearly quadruple in the next five years. Because of the large bandwidth it offers, optical
fiber has been employed as the transmission medium to meet this demand, operating at
wavelengths around 1.5pm because of its decreased loss and low dispersion at this wave
length. In addition, wavelength division multiplexing (WDM) in highly-dense networks is
often employed to take further advantage of the fibers generous bandwidth.
It follows that a large surge in the production of integrated-optical devices for use
in broad-band fiber-optic networks, as well as other commercial and military applications,
is clearly evident. Some of the most promising devices include the following: traveling-
wave modulators, high-output 1.5pm lasers, and high-speed optical switches for use in
fiber-optic communications systems [Alf84], wavelength-selective tunable filters for use
in WDM systems [Hus95], optical parametric oscillators (OPO) for generation of coher
ent, long wavelength (2.5-4pm) radiation for use in spectroscopy and laser radar, non-lin
ear frequency doublers for blue-light generation needed in optical storage applications
[hl94a, hl94c, Nak90], and high-frequency, broad-band surface acoustic wave (SAW)
filters for use in optical spectrum analyzers and television applications [Xu91].
In order to fabricate these devices, we must find a suitable material whose charac
teristics are such that it is able to meet the demands imposed by each device. In particular,
it must be transparent in the wavelength regions of interest, for both difference frequency
generation of infrared (as in OPOs) and in the visible, for blue-light generation by nonlin
ear frequency doubling. It should be capable of producing waveguides with low propaga
tion loss and must possess a large electro-optic coefficient, desirable in the fabrication of
electro-optic modulators and switches. It must also possess a large nonlinear coefficient
and be capable of satisfying the phase matching condition so that difference frequency


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
PROPERTIES AND CHARACTERISTICS OF LiTa03 FOR
INTEGRATED-OPTICAL DEVICE APPLICATIONS
By
David Blayne Maring
May 2000
Chairman: Ramu V. Ramaswamy
Major Department: Electrical and Computer Engineering
Continued and rapid growth expounded by the bandwidth explosion in the area of
dense-wavelength-division-multiplexed (DWDM) telecommunication and information
networks, like the internet, has been and continues to be met by the global deployment of
optical fiber. In turn, this has resulted in the unprecedented demand for the integrated-
optical components used in these fiber networks. Of particular interest is devices fabri
cated in ferroelectrics, most notably lithium niobate (LiNb03) and lithium tantalate
(LiTa03). Although current demands are being met by LiNb03, its brethren LiTa03
exhibits a much higher threshold for optical damage, providing the opportunity to fabri
cate devices, such as lasers and modulators, with much higher output and throughput lev
els than currently obtainable with LiNb03. In addition, LiTa03 has a shorter wavelength
ultra-violet (UV) absorption edge (280nm) than LiNb03 (350nm), permitting nonlinear
conversion by frequency doubling to shorter wavelengths with less absorption.
IX


9
been as seriously investigated as has APE. While it is likely that these techniques may be
better suited for the production of waveguides in Er-doped regions because they wont
contribute to H+-induced quenching, they have the drawback of needing to be performed
at temperatures above the Curie point, thus requiring the additional step of crystal repoling
to return the sample to a monodomain state. Additionally, no investigations have been
reported as to the temporal stability of waveguides fabricated using these techniques.
From the above discussion, it is clear that additional research still remains to be
performed on the characterization of the different waveguide formation processes used in
LiTa03, as well as the characterization of Er-doping and its effect on the crystals optical
properties, before reliable, high-throughput integrated-optic devices can be fabricated in it.
This need for continued research formulates the motivation for this work which is outlined
in the next section, along with a short outline of the steps to be taken in order to complete
this work.
1.4 Motivation and Objectives
As stated earlier, LiTa03 offers the advantages of high power handling capability
and a shorter wavelength UV absorption edge. However, the feasibility of fabricating low-
loss, temporally stable waveguides for use in reliable, high-throughput integrated-optical
devices such as modulators, filters, and lasers has not yet been demonstrated. To this end,
considerable research needs to be performed to characterize both the process of Er-indiffu-
sion and the various methods of waveguide fabrication in LiTa03. This forms the basis of
the motivation for this work, with a specific interest in determining whether fabrication
conditions exist which will produce temporally stable waveguides.


164
its bulk value and thus decrease the value of WK, though this will be done at the expense of
increased propagation loss.
Finally, to achieve bandpass rather than baseband operation, quasi-phase-matching
may be employed by using regions of domain reversal as illustrated in Fig. 8.2. Such a
technique has the advantages over periodically reversed electrodes of reduced microwave
reflections and the possibility of being implemented on an X-cut substrate [Wan97],
8.2.2 Integrated Laser-Modulator Module
Because of its improved photorefractive characteristics after Er-doping, especially
in conjunction with Li-treatment, LiTa03 extends itself to the fabrication of high-power
Fig. 8.3 Schematic of the proposed integrated optical source incorporating an
Er-doped waveguide laser and external high-speed Mach-Zehnder
modulator.


140
cause of instability, as Zn and Ti ions are larger and less mobile, but do not produce
waveguides with improved stability. Therefore, the source of the stability problem may lie
within the crystal itself, the result of an imperfect growth process. This issue is examined
in the next section.
6.5 Examining the Crystal Growth Issue
To this point, all waveguides and devices were fabricated in LiTa03 provided by
the same supplier, namely Deltronic Crystal Industries, Inc. As was eluded to in the last
section, the stability problem is likely not a result of the fabrication conditions, but may lie
within the bulk crystal itself, the result of an improperly grown material. As such, crystals
from different growers need be examined in the same manner for comparison against
those used here. Additionally, because of LiTa03s much higher melting point (1650C),
it is grown at temperatures of greater than 400C higher than that for LiNb03. At such
extreme temperatures, Li20 outdiffusion during growth is a problem, resulting in noncon-
gruent LiTa03 crystals which are somewhat Li deficient, containing only 47.88 mol%
Li20, as opposed to 48.38 mol% for LiNb03.
In this section, the material growth issues outlined above are addressed. LiTa03
samples from several growers, both optical and surface acoustic wave (SAW) grade, are
compared against each other in terms of stability. The issue of Li deficiency is also exam
ined by using Li-treatment (Chapter 3) to introduce more Li into the crystal in an attempt
to improve stability.


sample was loaded into the reactive ion etcher (RIE) and etched in 02 plasma for 1.5 min
utes (30mTorr, 12.4 seem, 60W) to remove any remaining photoresist from the exposed
areas.
The thick gold layer is next electroplated to the sample. The setup for this is
depicted in Fig. 2.13. The sample was attached to the negative lead of a DC power supply
with the positive lead being attached to a grate. Both were submerged into a beaker con
taining Orotemp 24 gold solution from Technic, Inc. A 1000Q resistor and ammeter were
put in the circuit to monitor current flow. The beaker was placed in a temperature-con
trolled water recirculator and the temperature set to 45C. A current of 0.5mA per cm2 of
sample surface was then applied for 30 minutes, after which the temperature and current
were both increased. The complete process is as follows:
45C, 0.5mA/cm2, 30 minutes
50C, ImA/cm2, 10 minutes
55C, 1.5mA/cm2, 10 minutes
60C, 2mA/cm2, 40 minutes
The resulting Au thickness was around 6.5pm.
After electroplating, lift-off was performed using acetone and a 10:1 solution of
H2S04:H202. Gold etchant, diluted 1:2 with DI water, was used next to remove the gold
seed layer. Finally, the sample was loaded into the RIE and the Cr layer was removed
using a Cl2 plasma for 20 minutes (30mTorr, 3.4sccm, 60W).
2.8 Summary
In this chapter, the fabrication of planar and channel waveguides in LiTa03 was
presented. A few methods for fabricating the waveguides were given, with the main focus


149
7.2.2 Microwave Effective Index
The microwave effective index Nm was measured next. Normally, Nm can be esti
mated from the following equation:
(7-1)
where exx and ezz are the relative dielectric constants of the ordinary and extraordinary
axes, respectively. This gives a value of about 4.5 for Nm. However, by using the network
analyzer to measure the group delay along the electrode, Nm can be calculated directly
using
(7-2)
where c is the speed of light, 8 is the measured group delay, and L is the total electrode
length. As before, the group delay through the 1cm Calibration Kit was also measured to
calibrate out the effect of the test fixture. The resulting measured group delay for the fab
ricated modulator was 260ps which, along with L= 1.9cm, gave a microwave effective
index of Nm=4.1.
7.2.3 Characteristic Impedance
The characteristic impedance Z0 of the electrode was estimated from plots of Z0
versus electrode width-to-gap ratio W/G [Chu91, Kub80], For the modulator fabricated
here, W/G=l resulting in a characteristic impedance of close to 402.
In the next section, the DC response of the modulator is measured in order to deter
mine the switching voltage Vn and the extinction ratio. In addition, the power handling
capability of the modulator near 1.5|im is examined.


60
input intensity J (W/cm2)
Fig. 3.7 Saturated values of photorefractive index change Ans versus input
intensity J for excitation wavelengths of 632.8nm and 785nm.
increased the PR effect. Second, there is a pronounced dependence of the PR effect on
input intensity and this dependence is modified with wavelength. At the 785nm excitation
wavelength, the dependence of Ans on J in both the Er-doped sample and the virgin
LiTa03 substrate is nearly linear for J > 103 W/cm2. This was reported [Gla72, Kos97a] to
be typical for nominally pure LiTa03 and LiNb03 in the studied range of J from 103 to
5104 W/cm2. The slope 3(Ans)/3J of this dependence increased proportionally with Er3+
concentration. The proportionality factor has a maximum value of 1.6 in the sample with
the highest Er3+ concentration (Fig. 3.7). At 632.8nm, the dependence of Ans on J for the
same Er-doped sample has a much more complicated character, as seen in Fig. 3.7. Using
a simple mathematical fitting to this data demonstrates the appearance of an additional
intensity dependent component Ans(J), superimposed on the linear dependence, given by


[Ulr73]
[Wan97]
[Woo93]
[Xu91]
[ Yam91 a]
[Yam91b]
[Yam92]
[Y0088]
[Yuh92]
178
R. Ulrich and R. Torge, Measurement of Thin Film Parameters with a Prism
Coupler, Appl. Opt., vol. 12, no. 12, p. 2901, 1973.
W. Wang, R. Tavlykaev, and R. V. Ramaswamy, Bandpass Traveling-Wave
Mach-Zehnder Modulator in LiNb03 with Domain Reversal, IEEE Photon.
Technol. Lett., vol. 9, no. 5, p. 610, 1997.
E. L. Wooten and W. S. C. Chang, Test Stuctures for Characterization of Elec
trooptic Waveguide Modulators in Lithium Niobate, IEEE J. Quantum Elec-
ton., vol. 29, no. 1, p. 161, 1993.
Y. Xu, Ferroelectric Materials and Their Applications, North-Holland,
Amsterdam, 1991.
K. Yamamoto and T. Taniuchi, Characteristics of Pyrophosphoric Acid Pro
ton-Exchanged Waveguides in LiNb03, J. Appl. Phys., vol. 70, no. 11, p.
6663, 1991.
K. Yamamoto, K. Mizuuchi, K. Takeshige, Y. Sasai, and T. Taniuchi, Charac
teristics of Periodically Domain-Inverted LiNb03 and LiTa03 Waveguides for
Second Harmonic Generation, J. Appl. Phys., vol. 70, no. 4, p. 1947, 1991.
K. Yamamoto and K. Mizuuchi, Blue-Light Generation by Frequency Dou
bling of a Laser Diode in a Periodically Domain-Inverted LiTa03 Waveguide,
IEEE Photon. Tech. Lett., vol. 4, no. 5, p. 435, 1992.
D. W. Yoon and O. Eknoyan, Characterization of Vapor Diffused Zn:LiTa03
Optical Waveguides, J. Lightwave Technol., vol. 6, no. 6, p. 877, 1988.
T. Yuhara, K. Tada, and Y. S. Li, Anomalous Refractive Index Change and
Recovery of Electro-Optic Coefficient r33 in Proton-exchanged LiTa03 Optical
Waveguides After Annealing, J. Appl. Phys., vol. 71, no. 8, p. 3966, 1992.
J. M. Zavada, H. C. Casey, Chang-Ho Chen, and A. Loni, Correlation of
Refractive Index Profiles with Substitutional Hydrogen Concentrations in
Annealed Proton-Exchanged LiNb03 Waveguides, Appl. Phys. Lett., vol. 62,
no. 22, p. 2769, 1993.
[Zav93]


96
Fig. 4.13 Scan of focused spot as it trav
els across a step-interface
from bulk LiTa03 to highly
reflective metal.
depth (pm)
Fig. 4.14 Ordinary index profile taken at
five different points across a
sample.
By observing Eq. (4-11), it is seen that the accuracy in measuring index changes is
about the same as the accuracy in measuring AVIV. This method is expected to be capable
of reproducing actual refractive index profiles with an accuracy of An/n 10'4.
In order to test the repeatability of this measurement technique, an x-cut sample of
LiTa03, proton exchanged and then annealed for 15 minutes, was measured at 5 different
points across the width of the sample. The results of this test for the ordinary polarization
are shown in Fig. 4.14. As can be seen from this figure, a high degree of repeatability is
achieved from this method. Results of the same test for the extraordinary polarization
show a similar level of repeatability.
4.6.4 Measurement Results
The extraordinary index profiles for a set of planar samples on X-cut LiTa03 are
shown in Fig. 4.15. The samples were proton exchanged in pyrophosphoric acid at 260C
for 20 minutes, then annealed at 300C, for various times ranging from 0 to 120 minutes.


BIOGRAPHICAL SKETCH
David Blayne Maring was born in Rochester, Minnesota, in 1969. He received a
B.S. degree in electrical engineering from the University of Florida in December 1994.
He has been a research assistant at the Photonics Research Laboratory under the supervi
sion of Professor Ramu V. Ramaswamy since January 1994, after starting his undergradu
ate senior design project with Dr. Ramaswamy during the fall of 1993 in the area of single
mode waveguide fabrication in lithium tantalate. He spent the summer of 1994 perform
ing characterization on waveguides fabricated in lithium tantalate by the annealed proton
exchange technique at Eglin Air Force Base in Fort Walton Beach, Florida.
He received his Master of Engineering degree in electrical engineering from the
University of Florida in August of 1996. The title of the thesis was Characterization and
Direct Index Profiling of Annealed Proton Exchanged Waveguides in LiTa03. Since then
he has been working on his Ph.D. research at the University of Florida in the area of inte
grated-optical devices in lithium tantalate. His current research interests include inte
grated-optical components for optical communication networks.
179


61
the following empirical relation:
Ans(J) = D[Er3+] Jx (3-3)
where D is a constant of proportionality and [Er3+] is the Er3+ concentration. Above a cer
tain threshold value of J, around 104 W/cm2, the value of x is 1. Below this threshold, x is
about 2. However, at light intensities lower than 103 W/cm2 the contribution described by
Eq. (3-3) becomes insignificant. Finally, at 647nm the dependence of Ans on J was
roughly similar to that for 632.8nm. However, accuracy in calculating Ans from spontane
ous Raman scattering at 647nm, as well as for J > 5xl04 W/cm2 at 632.8nm, was ham
pered because of interference from SF and resonant Raman scattering. Rough estimation
of Ans gives a value of around 2x1 O'3 for the highest J used (2xl05 W/cm2).
From a comparison of the data for Ans(J) (Fig. 3.7) with that of Fig. 3.3, it is evi
dent that there is a strong correlation between the up-converted green emission and the
increased PR effect at high intensities of J. Such a correlation may be expressed, according
to Eq. (3-3), by the following empirical relations:
Ans = (E)IASF; Iasf = (D/E)[Er3+] Jx (3-4)
where E is a constant of proportionality. Note, that the experimental dependence of Ans
on IASF expressed by Eq. (3-4) is in agreement with the gating dependence of photorefrac-
tivity reported for Pr:LiNb03 using an extrinsic source of gating light [Gue97], As such,
we may conclude that Er:LiTa03 employs self-gating (i.e. secondary green emissions) to
achieve significant two-color photorefractivity from a monochromatic pump.
It is noted that for a quantitative comparison, Er-doped LiNb03 samples were also
prepared with fabrication conditions similar to those reported for active devices [Bau96].
A comparison of the data reported in Fig. 3.7 with the corresponding data for Er-doped


122
just after fabrication
4-5 days later
e33 x 103
phase
e33 xiO3
phase
10.8
p4
9.83
(34
9.2
P4
8.9
(34
7.7
p4
7.15
(34
6.7
(34
5.02
(34
6.6
(34
5.16
(33
6.52
[34
5.35
(33
6.03
(34
5.35
(34
5.74
(33
5.70
(33
4.14
(33
4.14
(33
3.8
(31,(32
3.8
pi,(32
3.31
P1,P2
3.31
(31,(32
3.07
(31,(32
3.07
P1.P2
2.92
(31,(32
2.92
(3l,|32
2.75
K
2.75
K
1.75
K
1.75
K
0.98
a
0.98
a
0.75
a
0.75
a
Table 6.1 Measured values of strain e"3 taken from APE samples
immediately after fabrication and again 4-5 days later.
Arrows on the left indicate samples which exhibited a
change in e"3. Arrows on the right indicate samples
which underwent a phase transition.


121
directional coupler. Using this technique, the stability of waveguides in different regions of
the structural phase diagram for APE:LiTa03 were analyzed. A comparison, in terms of
stability, between PE/APE waveguides and metal-indiffused waveguides was then made to
determine if stability depends on fabrication conditions or is inherent to the crystal. Finally,
crystals from different vendors were compared to determine the growth maturity for this
material and some of the problems surrounding its growth were examined.
6.1 Rocking Curve Analysis of Stability
In this section, an initial investigation into the long-term stability of APE
waveguides was performed by monitoring changes in rocking curves over time. Samples
fabricated in X-cut LiTa03 using the APE method, and sources outlined in Chapter 5 to
reconstruct the structural phase diagram, were subjected to rocking curve measurements
immediately after fabrication, then again 4-5 days later in order to track changes in the
surface strain value e"3 The results are presented in Table 6.1, which shows the e"3 val
ues, and corresponding phase at the sample surface, just after fabrication and again 4-5
days later. The arrows on the left side of the figure indicate samples which recorded a
change in the e"3 value over the 4-5 day period. The arrows on the right side of the figure
indicate samples which not only showed a change in strain, but also experienced a phase
transition (form the (34 to the (33 phase) in the near-surface layer.
From Table 6.1, it was seen that samples in the high-concentration regions of the
structural phase diagram, particularly the (33 and (34 phases, exhibited significant instabili
ties over the period examined, with a change in e"3 by as much as 20% in some cases.
Hence, these two high-concentration phases are not suitable for the fabrication of reliable


4.1.1 Description of the Measurement Technique
Light from a 1.55fim diode laser was coupled into each of the channel waveguides
via a single-mode fiber pigtail. The complete experimental setup is illustrated in Fig. 4.1.
The output face of each waveguide was focused, using a 20x lens, onto the input surface of
an infrared (IR) camera, with the output video feed of the camera connected directly to a
video display monitor. Using this approach, the number of modes supported by a particu
lar waveguide can easily be determined by counting the number of observed mode spots
appearing on the video monitor screen. Some results of this measurement technique are
presented next.
(b)
Fig. 4.2 Video monitor screen
snap shots depicting a sin
gle-mode (a) and a dou
ble-mode (b) channel
waveguide.
4.1.2 Modal Characterization Results
By scanning the single-mode fiber
from the laser source across the input face of
each waveguide, it was possible to create
off-axis coupling into any higher order
modes that may have been supported by the
waveguide. This was done for each of the
channel waveguides in LiTa03, and the
number of modes supported for each was
determined. As an example, Fig. 4.2 shows
the video monitor screen photographs taken
for a single-mode and double-mode channel
waveguide.


74
Fabrication
Technique
Fabrication
Conditions
Crystal Cut
Single-Mode
Region
APE
(pyrophosphoric
acid)
PE: 260C, 2hrs.
A: 300C, lhr.
PE: 260C, 1 hr.
A: 300C, 3 hrs.
PE: 260C, 30 min.
Z
2.5-6.5pm
APE
(pyrophosphoric
acid)
PE: 260C, 10 min.
A: 340C, 1 hr.
PE: 260C, 7 min.
A: 340C, 10 hrs.
Z
3.5-7.5pm
PE
(glycerin)
PE: 260C, 24 hrs.
X
2-3.5pm
Ti-indiffusion
900, 1200C, 20 hrs.
z
2.5-4.5pm
Zn vapor-indiffusion
800C, 5.5 hrs
z
2.5-7.5pm
Table 4.1 Near-field characterization results depicting the region of single
mode operation immediately after fabrication for the extraordinary
index, in terms of waveguide channel width, for various fabrication
conditions.
There were far too many waveguides produced with varying fabrication condi
tions, especially using APE, to be presented here. However, Table 4.1 presents some
results of modal characterization from each of the fabrication processes used. It is noted
that the regions of single-mode operation listed here are for the extraordinary index only.
Additionally, waveguides fabricated by APE involved multiple steps of proton exchange
followed by annealing in order to achieve the desired waveguide depth, while maintaining
a desired proton concentration at the surface. The reason for this will become clear when
the structural phase diagram for APE is reconstructed (Chapter 5) and stability issues are
addressed (Chapter 6).


The above results seem to indicate that instability is not simply due to the presence
of multiple crystal phases in the waveguiding layer, as was previously assumed. Sample 2,
having two phases was more stable than Samples 3 and 3', which were single-phase.
133
These results do, however, seem to indicate that waveguides with decreased proton con
centration, and strain, do have improved stability, as Sample 2 had the smallest strain
value and also the best stability. As a result, it was decided to try to fabricate waveguides
with decreased strain values. Again, the PE technique afforded the best opportunity as
attempts to fabricate waveguides by APE with strain values smaller than that of Sample 2
required too many successive repetitions of exchange and anneal, especially for X-cut, to
be practical. Additionally, these waveguides suffered from a combination of small An at
1.55|im and severely graded profiles due to annealing, making bending losses so high that
the fabrication of directional couplers, which have bends, was quite impractical. There
fore, it was decided to stick with PE, having the advantages outline in Section 6.3.2, only
with a more dilute proton source, in this case glycerin diluted with Li2C03.
The measured change in cou-
days
pling length of the new sample fabri
cated in diluted glycerin (denoted as
Sample 4 in Table 6.3) is shown in
Fig. 6.9. Because of the more dilute
source, the exchange time increased
from 24 hours to about 80 hours. The
e"3 value for this sample decreased
Fig. 6.9 Measured coupling length of Sam
ple 4. This coupler was comprised only slightly to 3.6. As indicated in
of 7|im channels with an 8p.m gap.


31
2.2.1 Ti Metal-Indiffusion
Waveguide fabrication by Ti indiffusion is performed by evaporating a Ti layer,
possibly patterned by photolithography, onto the surface of a substrate and subsequently
indiffusing the metal at high temperatures (>1000C). Since its discovery in 1974
[Sch75], much work has been done on the characterization and fabrication of Ti indiffused
waveguides and devices, especially in LiNb03. There is a limited amount of data reported
in literature about the Ti indiffusion process in LiTa03. Most of it is from the device per
spective [How89, Tan78] and does not characterize the process of waveguide formation in
LiTa03 by Ti indiffusion. As a result, much of the description of this process given here
comes from reported data in literature dealing with Ti indiffusion in LiNb03, which is
assumed to be very similar.
The indiffusion of Ti results in a relatively Gaussian index profile [Bur79]. Unlike
APE, it produces an increase of both the extraordinary and ordinary indices. In both
LiNb03 and LiTa03 there is a larger change of the extraordinary refractive index than of
the ordinary refractive index [Bur79, Tan77],
The diffusion equations governing diffusion depth and the diffusion coefficient are
similar to those given for proton exchange in Eqs. (2-2) and (2-3) [Bur79]. However, the
diffusion coefficient for Ti in LiTa03 is about 1.5 orders of magnitude smaller than that
forTi in LiNb03 [Bur79],
Upon indiffusion, Ti ions take positions at Nb (Ta) sites in the LiNb03 (LiTa03)
crystal lattice [INS89]. The mechanism of index increase is thought to be attributed to an
increase of the polarisability and to the photoelastic effect caused by the different ionic
radii of Ti and Nb (Ta) ions [INS89].


127
Sample
Crystal
Cut
APE Conditions
e33 C'3)
Phase
Days
8(An)
1
Z
PE: 260C, 2hrs
A:300C, lhr
PE: 260C, lhr
A: 300C, 2hrs
PE: 260C, 30min
5.43
8
7
1.2x1 O'3
99
2.7xl0'3
239
3.5xl03
2
Z
PE: 260C, lOmin
A:340C, lhr
PE: 260C, 7min
A: 340C, lOhrs
0.95
K
7
1.7x1 O'4
98
7.9x1 O'4
219
1.2x1 O3
Table 6.2 Fabrication conditions, measured e"3 value, associated phase, and esti
mated change in index increment 8(An) at various periods after fabrication
for Samples 1 and 2.
without suffering from surface cracking. Even for the case of Z-cut, repeated exchanges
followed by annealing were necessary to produce single-mode waveguides at 1.55(im with
the desired surface concentration. Had X-cut been used, it would have required many
more steps of exchange followed by annealing to avoid surface damage.
The measured coupling length of Sample 1 over a period of 239 days is shown in
Fig. 6.4. This sample had the highest proton concentration and was in the 8 phase of the
Z-cut diagram. As can be seen from this
figure and Table 6.2, Sample 1 exhibited a
large index change over just the first week
after fabrication. This change continued
over the 239 days it was monitored, reveal
ing a change in index increment 8(An) of
3.5xl0~3. The magnitude of this change
surface £33 (xlO3)
Fig. 6.3 Structural phase diagram for
7-rut APP-I iTaO.,


147
electrodes were fabricated by Au electroplating using the process described in Section 2.8.
The interaction length was 1cm. The center electrode width and the gaps between elec
trodes were all 18|im.
In the next section, some microwave characteristics of the structure are measured.
Parameters such as microwave loss, effective index, and characteristic impedance are
determined.
7.2 Microwave Characteristics
The performance of high-speed modulators is largely dependent upon the micro-
wave transmission and reflection properties of the traveling-wave electrodes. Therefore,
in this section, the microwave characteristics of the modulator structure are measured.
Fig. 7.2 Measured and calibrated S-parameters for the traveling-
wave electrode structure.


In this dissertation, the process of Er-doping in LiTa03 is characterized in order to
examine its impact on the optical and physical properties of the material, and gauge its
usefulness for the development of high-power infrared lasers. Additionally, the different
processes of waveguide formation are characterized with an aim towards determining the
feasibility of identifying fabrication conditions that will produce stable waveguides, and
hence reliable devices, in LiTa03.
1.1 Integrated-Optical Devices and Ferroelectric Materials
Demand for broadband information and communication services for telephone,
cable television (CATV), and computer network applications, such as the internet, contin
ues to increase significantly year by year. As illustrated in Fig. 1.1, for the telecommuni
cation industry alone, this trend toward rapidly increasing bandwidth demand is expected
year
Fig. 1.1 Relative trend in the demand for telecommunication net
work bandwidth.


156
The measured response showed a large deviation from the theoretical response at
low frequencies (l-2GHz). As can be seen from Figs. 7.2 and 7.3, these frequencies corre
spond to an unusually large microwave reflection peak, and hence a dip in microwave
transmission. The electrode structure fabricated here was similar to that used in Ref.
[Wan97], having the exact same value of microwave loss coefficient, but on a LiTa03 sub
strate instead of a LiNb03 substrate. It has been established [Chu91] that the impedance
mismatch of LiTa03 is more serious than that of LiNb03, leading to much higher micro-
wave reflections in LiTa03 for the same W/G ratio of the electrodes. This of course can
be improved by changing the W/G ratio of the electrodes. Further deviations of the mea
sured data from that of the theoretical calculation may be the result of a microwave index
which is frequency dependent.
7.5 Summary
In this chapter, the potential for LiTa03 as a host material for integrated-optical
device applications was demonstrated by the fabrication and testing of a symmetric MZI
modulator with a CPW traveling-wave electrode structure on it. The MZI was fabricated
by PE in glycerin and the electrodes were Au electroplated with a W/G ratio of 1. No fab
rication issues were uncovered which would prevent the development of additional
devices indicating that when crystal quality, and hence stability, is improved, current fabri
cation techniques used in LiNb03 may be directly applied to LiTa03 for the production of
an array of devices.
The microwave characteristics of the device were examined next. The S21 and
parameters were measured and a microwave loss coefficient of 1.0dB/cm VGHz was


REFERENCES
[hl91]
[hl93]
[hl94a]
[hl94b]
[hl94c]
[hl95]
[Alf84]
[Ami96]
[Ara92]
H. hlfeldt, J. Webjorn, and G. Arvidsson, Periodic Domain Inversion and
Generation of Blue Light in Lithium Tantalate Waveguides, IEEE Photon.
Tech. Lett., vol. 3, no. 7, p. 638, 1991.
H. hlfeldt, F. Laurell, and G. Arvidsson, Strongly Reduced Optical Nonlin
earity in Lithium Tantalate Due to Proton Exchange, IEEE Electron. Lett., vol.
29, no. 9, p. 819, 1993.
H. hlfeldt, Nonlinear Optical Properties of Proton-Exchanged Waveguides
in Z-cut LiTa03, J. Appl. Phys., vol. 76, no. 6, p. 3255, 1994.
H. hlfeldt and J. Webjorn, Single-Domain Layers Formed in Multidomain
LiTa03 by Proton Exchange and Heat Treatment, Appl. Phys. Lett., vol. 64,
no. 1, p. 7, 1994.
H. hlfeldt, J. Webjorn, F. Laurell, and G. Arvidsson, Postfabrication
Changes and Dependence on Hydrogen Concentration of the Refractive Index
of Proton-Exchanged Lithium Tantalate Waveguides, J. Appl. Phys., vol. 75,
no. 2, p. 717, 1994.
H. hlfeldt, J. Webjorn, P. A. Thomas, and S. J. Teat, Structural and Optical
Properties of Annealed Proton-Exchanged Waveguides in Z-cut LiTa03, J.
Appl. Phys., vol. 77, no. 9, p. 4467, 1995.
R. C. Alferness, S. K. Korotky, and E. A. J. Marcatili, Velocity-Matching
Techniques for Integrated Optic Traveling Wave Switch/Modulators, IEEE J.
Quantum Electron., vol. QE-20, no. 3, p. 301, 1984.
J. Amin, J.A. Aust, and N.A. Sanford, Z-Propagating Waveguide Lasers in
Rare-Earth-Doped Ti:LiNb03, Appl. Phys. Lett., vol. 69, p. 3785, 1996.
S. Arahira, K. Watanabe, K. Shinozaki, and Y. Ogawa, Successive Excited-
State Absorption Through a Multistep Process in Highly Er3+-Doped Fiber
Pumped by a 1.48-pm Laser Diode, Opt. Lett., vol. 17, no. 23, p. 1679, 1992.
170


8
Fig. 1.3 Temporal changes in the index
profile (extraordinary polarization)
measured at ^=0.633pm
immediately after fabrication and 4
weeks later.
of a monotonically decreasing
index profile. Most important from
a practical standpoint, APE
waveguides in LiTa03 have been
shown to exhibit temporal instabili
ties [hl94c, Mar96b, MarOO],
manifesting themselves in the form
of significant index profile modifi
cations for both short and long
periods after fabrication. An
example of this is depicted in Fig.
1.3, showing the extraordinary index profile of a waveguide measured immediately after
fabrication, and again four weeks later. These instabilities are believed by some to be due
to the presence of multiple crystal phases within the proton exchanged region [E1H95,
Kor96], though it has not been directly confirmed. This fact prevents LiTa03 from being
used in the fabrication of practically viable, that is, temporally stable waveguides for
device applications. As a further point, it is known that the introduction of H+, through
proton exchange, into a rare-earth doped region leads to significant quenching of laser
action, or a reduction in excited state lifetime, thereby reducing laser efficiency. As such,
additional methods for waveguide fabrication need be explored.
As stated earlier, the other two recognized techniques of waveguide formation in
LiTa03 are Ti metal-indiffusion and Zn vapor-indiffusion. While a small amount of work
has been reported in the literature about the success of these techniques, they have not


101
Fig. 4.17 Schematic of the X-ray diffrac
tion measurement process.
Fig. 4.18 Illustration of an
expected rocking curve
for a PE sample.
the angle of incidence 0 is such that Braggs Law is satisfied:
n?i=2dsin0. (4-12)
By slowly rotating, or rocking, the sample during a measurement run, the angle of inci
dence can be changed and the corresponding reflection intensity monitored. Such a pro
cess is referred to as a rocking curve measurement and the resulting data of intensity
versus incident angle is a rocking curve.
For the case of proton exchanged waveguides, the introduction of H+ into the crys
tal increases the lattice spacing d. As a result, the angle 0 at which the Bragg reflection
peak occurs is shifted to a smaller angular value. So a rocking curve measurement of such
a sample will exhibit two peaks as illustrated in Fig. 4.18, one corresponding to the bulk
substrate and one corresponding to the proton exchanged region.


7
Of the three methods for waveguide fabrication outlined above, APE is presently
the most common. Compared to the other techniques, APE offers the advantages of sim
plicity, flexibility, high index increment, and low propagation loss. In addition, APE can
be performed at temperatures below the Curie point of the crystal (610C), assuring that
this technique does not disrupt the monodomain structure of the crystal and so no post
fabrication repoling process is needed.
However, a number of anomalies are known to exist for APE waveguides in
LiTa03. For instance, it has been shown that the peak extraordinary index increment
increases during the initial stages of annealing [hl94c, Mat92b] and that the index
decreases with H proton concentration as proton concentration exceeds a certain point.
These facts suggest a nonlinear dependence of index increment on proton concentration
and the possibility to form buried refractive index profiles [hl95, Dav95, Mar96a], Such
a buried profile has recently been directly observed [Mar96b] and is exhibited in Fig. 1.2.
Though buried index waveguides
can provide the advantages of
improved fiber-to-device coupling
and reduced scattering losses at the
substrate-air interface, they seri
ously undermine the adequacy of
previous theoretical studies, most
of which relied on the indirect pro
file reconstruction technique of
inverse WKB and the assumption
Fig. 1.2 Refractive-index profiles of extraordi
nary polarization for various anneal
ing times measured at 1=0.633p.m.


waveguide lasers operating near the communications wavelength of 1.55p.m. Of particular
interest is the integration of such a laser with an electro-optic modulator on a single, corn-
165
pact module, as illustrated in Fig. 8.3. Such a module has the advantages of low chirp and
eliminates the need for coupling between a separate laser source and modulator. In this
case, the laser cavity reflectors are provided by a multilayer dielectric mirror and a corru
gated waveguide reflector fabricated using a Si overlay [Hus95], The waveguide, at least
in the laser section, would likely be fabricated by Zn vapor-indiffusion because of its high
photorefractive resistance and lack of H+-induced gain quenching.
8.2.3 Photorefractive Grating
As was detailed in Chapter 6, Er:LiTa03 offers some advantage over Er:LiNb03
toward photorefractive (PR) applications such as optical data storage and PR grating fabri
cation. To this end, the fabrication of a PR grating by two-color recording could be real
ized. Such a grating could be used in place of the corrugated reflector of the waveguide
laser or in the development of a tunable WDM filter. The PR grating holds the advantage
over the corrugated reflector [Hus95] of a much simpler fabrication process. However,
some research would need to be performed, particularly into the long-term stability of the
written grating.
The fabrication of such a PR grating involves, first, the doping of the substrate with
Er, using conditions outlined in Chapter 3. Next, the grating can be written using a stan
dard two-beam interference setup [Hus95] and a source of gating light. A nonmonochro-
matic source of gating light in the violet-near UV range (e.g. mercury or xenon lamp) may
be incident on the sample, and an interference pattern formed with aid of a powerful red
laser (HeNe) to record the grating.


5 STUCTURAL PHASE DIAGRAM FOR APE:LiTa03 105
5.1 Previous Limitations 107
5.2 Constructing the Structural Phase Diagram 108
5.3 Analyzing the Structural Phase Diagram 110
5.3.1 Explaining APE Anomalies 110
5.3.2 Effect of Crystal Phases on Index Profiles 110
5.4 PE Waveguides 113
5.4.1 Raman Analysis of PE Waveguides 114
5.4.2 Modified Structural Phase Diagram 116
5.5 Summary 118
6 STABILITY OF LiTa03 WAVEGUIDES 120
6.1 Rocking Curve Analysis of Stability 121
6.2 Using a Directional Coupler to Measure Stability 123
6.2.1 Directional Coupler Theory 123
6.2.2 Measuring Coupling Length and Index Change 124
6.3 APE and PE Stability Results 126
6.3.1 APE Results 126
6.3.2 The Advantage of PE 129
6.3.3 PE Results 131
6.3.4 Limits to Reducing Strain 135
6.4 Zn/Ti-Indiffused Stability Results 137
6.4.1 Directional Coupler Stability Results
of Zn:LiTa03 Waveguides 138
6.4.2 Comparison of the Waveguide Formation Processes 138
6.5 Examining the Crystal Growth Issue 140
6.5.1 Comparing Different Growers 141
6.5.2 Accounting for the Li Deficiency 142
6.6 Summary 143
7 EXPLORING THE POTENTIAL OF LiTa03:
A TRAVELING-WAVE ELECTRO-OPTIC MODULATOR 145
7.1 Fabrication Conditions 146
7.2 Microwave Characteristics 147
7.2.1 S-Parameters and Loss 148
7.2.2 Microwave Effective Index 149
7.2.3 Characteristic Impedance 149
7.3 DC Response and Power Handling 150
7.3.1 DC Response 150
7.3.2 Power Handling Near 1.5pm 152
7.4 Frequency Response 154
7.5 Summary 156
V


Mat92a], For certain fabrication
conditions, in the initial hours after
fabrication, a large decrease in Ane
has been shown to occur, followed
by a much smaller, gradual decrease
in Ane over the next few months
[hl95], The total decrease in
refractive index after several
months can be of the order of 10'3
[MarOO, Mat92a], Preliminary
measurements have demonstrated
large temporal changes on the same order, as illustrated in Fig. 2.6. These changes can be a
serious problem for devices which are extremely sensitive to index variations and require
long-term stability, such as directional-couplers and QPM waveguides used for frequency
doubling (to a specified wavelength).
The structural phase diagram for APE LiTa03 is invaluable for understanding the
properties of APE waveguides fabricated in this material. In particular, knowledge of the
structural phase diagram of LiTa03 would shed light on the origin of the aforementioned
anomalies, as well as serve to identify fabrication conditions that produce stable, possibly
single-phase waveguides. These waveguides could be utilized in the fabrication of reliable
integrated-optic devices which exploit the previously described advantages of LiTa03.
In the next section, the actual conditions for the fabrication of APE waveguides in
Fig. 2.6 Temporal changes in the index profile
(for the extraordinary polarization)
measured at X=0.633(lm immediately
after fabrication and 4 weeks later.
LiTa03 are presented.


single-mode near 1.5pm are generally multimode at 980nm, giving rise to the problem of
selective mode excitation [Bau96], Whereas when pumping at 1480nm, the mode sizes of
the pump and signal are nearly the same and so overlap is better.
As mentioned above, an inherent problem with guided-wave devices in LiNb03 is
their low resistance to photorefractive (PR) damage at visible and near-infrared wave
lengths. PR instability is especially severe for active waveguide devices on account of the
high power densities within the waveguides. It has been demonstrated [Ami96, Hua96]
that PR damage induces significant gain degradation and temporal instability in Er-doped
Ti:LiNb03 waveguide amplifiers and lasers. Even though the problem can be partially
alleviated by an appropriate co-doping technique [Die98, Hua96], the performance of the
tested devices is still far from the theoretical optimum [Cac97], Additional efforts to
avoid the PR problem have demonstrated a Z-propagating waveguide laser in Er:LiNb03
[Ami96], However, a limitation to this configuration is that it does not allow for the inte
gration of an additional electro-optic component, such a modulator, on the same chip as
the additional component would not benefit from the large electro-optic coefficient in the
Z-direction.
In light of the described limitations of LiNb03, LiTa03 appears an attractive alter-
native with a PR damage threshold of 1500 W/cm at 514.5 nm and room temperature, i.e.,
about forty times higher than the 40 W/cm2 value of LiNb03 [Gla72], Similar to LiNb03,
Er needs to be indiffused into LiTa03 to serve as the active gain medium for active inte
grated devices, such as amplifiers and lasers. However, Er-doping is known to lead to
upconversion in most host materials[Ara92, Del93, G196], which is widely recognized as
a gain limiting factor in Er-doped amplifiers by inducing PR damage [Ara92, Del93,


119
within this new phase, obtainable only by PE in dilute sources, were truly of single-phase
content.
In the next chapter, the structural phase diagram constructed here will be used to
study the stability of both APE and PE waveguides in LiTa03. Metal-indiffused
waveguides will also be examined, for stability comparison against APE and PE
waveguides. Additionally, crystals from different growers will be examined to determine
the maturity of the growth process for this crystal.


Second, is the anomalous behavior of waveguide profiles upon annealing. It has
been demonstrated that the refractive index increment increases during short annealing
times for the APE process in LiTa03, vastly increasing the area under the index profile
[hl94c, Mar96b, Mat92b, Yuh92]. In addition, studies performed by varying the H+ con
centration in the exchange source indicate that above a certain concentration, index incre
ment actually decreases with increasing H+ concentration [Ahl94c], These peculiarities
suggest a nonlinear dependence of the refractive index increment on H+ concentration and
the possibility of forming waveguides with buried refractive index profiles [Ahl94c,
hl95, Dav95, Mar96a], Such profiles have recently been confirmed by direct measure
ment [Mar96b], using a reflectivity profiling technique to be described in Chapter 4, and
are exemplified in Fig. 2.5. This is in direct opposition to the case of APE in LiNb03,
where the index increment is known to decrease upon annealing, maintaining a more or
less linear dependence on H+ con
centration [McW91].
The third, and most signifi
cant peculiarity, is that APE
waveguides in LiTa03 have been
found to exhibit significant tempo
ral instabilities, appearing in the
Fig. 2.5 Refractive-index profiles (for the
extraordinary polarization) for vari
ous annealing times measured at
^=0.633|im.
refractive index increment for both
short and long periods after fabrica
tion [hl94c, hl95, MarOO,


54
(Rayleigh Scattering). However, a very small portion of the scattered light, a fraction of a
percent, is frequency shifted by an amount equal to the vibrational frequency of the
molecules in the material. This is the Raman effect. When the frequency shift is toward
smaller frequencies, or smaller energies, it is called Stokes scattering. When the
frequency shift is toward higher frequencies, or larger energies, it is called anti-Stokes
scattering. A spectrometer can be used to detect the scattered light from the surface of a
sample. Detecting and plotting the intensity of this scattered light as a function of
frequency shift is referred to as Raman spectroscopy. A Raman spectrum is unique to a
given material, or to a certain composition of that material.
3.3.2 Fluorescence Spectroscopy
Fluorescence spectroscopy is similar to Raman spectroscopy except that here the
detected light is light re-radiated by ions within the material, in our case Er3+ ions. A plot
of detected fluorescence intensity versus frequency, or wavelength, is called a fluorescence
spectrum. In this case, wavelength is calculated from measured wavelength shift by:
k(nm) = 107/(De-Dm) (3-1)
where X is the wavelength in nm, Dm is the measured wavenumber in cm'1, and "Uc is the
wavenumber of the excitation source in cm'1 (e.g. 15802cm'1 for a 632.8nm source and
15456cm'1 for a 647nm source).
3.3.3 Fluorescence Measurements
To examine the influence of up-converted light on the PR effect in Er:LiTa03, the
fluorescence spectrum of each Er-doped sample was first measured using a Renishaw
Ramascope spectrometer operating at excitation wavelengths of 632.8, 647 and 785nm.


98
For illustration-clarity, all the measured profiles (ta=0 to 120 min), except for the
as-exchanged waveguide, were fitted with splined smooth curves which accentuate profile
evolution upon annealing. Index changes measured on as-exchanged waveguides were
rather low and, in addition, were significantly obscured by the aforementioned ripples,
preventing any fit to the measured curve. The maximum index increment, however,
dramatically increases upon short annealing periods (up to -0.02 for 30-min annealing
time), as does the area under the profiles. The shape of the index profile also undergoes
rapid transformations, revealing buried profiles for samples annealed for 7.5 and 15 min.
Buried profiles, formed in just a single-step annealing process, are most likely to result
from the nonlinear dependence of the refractive index on proton concentration. An
important consequence of this fact is that a rigorous theoretical analysis of APE
waveguides in LiTa03, if possible at all, is hindered to a large extent, until the structural
phase diagram of this material is accurately determined.
Using the reflectivity technique, ordinary profiles were measured in the same
straightforward manner with accuracy and resolution far beyond those of conventional
interferometric methods. The measured scans (X=0.633jim) are summarized in Fig. 4.16.
Note again, curves fitted to the measured distributions with ripples are plotted. The results
indicate, as opposed to the extraordinary index, there are no apparent buried profiles for
the ordinary polarization. As expected, the measured changes in ordinary index are
negative. By comparing Figs. 4.15 and 4.16, it follows that for the same annealing time,
the index distributions for the two polarizations have an approximately identical width.
However, the magnitude of the ordinary index decrease is substantially larger than the
corresponding increase in the extraordinary increase, especially for short annealing times.


117
exactly their position on the structural phase diagram, index profiling and rocking curves
were used to determine surface values of index increment An and strain z" .
The sample prepared in pure glycerin at 260C for 24 hours had measured values
of An=0.02 and e'33=4xl0'3. The sample prepared in glycerin diluted with 0.005mol%
Li2C03 and exchanged at 260C for 80 hours had measured values of An=0.017 and
e33 =3.6x10 As can be seen by inspection of Fig. 5.2, these sample points do not corre
spond to the a phase, or to any phase shown on the structural phase diagram. As such, it
was concluded that there existed a new phase, obtainable only by PE in a dilute source. To
confirm this, an additional data point was constructed. The pure glycerin sample was
annealed at 230C for 4 hours. However, no change in the value of e3 was measured.
Additional annealing was performed at 300C for 4 hours. Again, there was no change in
Fig. 5.7 Modified structural phase diagram for X-cut LiTa03 depicting
the newly discovered a' phase.


159
The accomplishments related to Er-doping of LiTa03 include, first, the definition
of conditions for Er-doping which provide an Er3+ concentration profile appropriate for
active guided-wave devices. Second, the presence of Er3+ was found to lead to upcon-
verted emissions, where energy-transfer-upconversion (ETU) from Er3+ clusters was dis
covered to be the dominant mechanism of this upconversion. Third, through the two-
photon (or two-color) effect, it was shown how upconverted light acted as a self-gating
source to produce, together with the pumping light, significant photorefractivity. Fourth, a
novel technique of Li-treatment was introduced and used to reduce upconversion, and
hence reduce photorefractive damage, by reducing the concentration of Er3+ clusters
which depend on the Li/Ta ratio. Finally, based on the stated results and a comparison
with Er-doped LiNb03, the inherent advantages of Er-doped LiTa03 for applications such
a waveguide laser fabrication, optical data storage, and photorefractive grating fabrication
were detailed.
There were also several accomplishments made on the characterization of
waveguide fabrication techniques in LiTa03. First, the conditions for the fabrication of
single-mode waveguides at 1.55(im by the processes of APE, PE, Ti-indiffusion, and Zn
vapor-indiffusion were outlined. Second, a direct index profiling technique was intro
duced and used to directly determine values of waveguide surface index increment, over
coming previous limitations associated with indirect techniques which had prevented the
construction of the structural phase diagram for APE:LiTa03. Third, using direct mea
surement techniques, the structural phase diagram for APE:LiTa03 was constructed. This
diagram was used to explain several of the previously observed anomalies associated with
the APE process in LiTa03. Additionally, the impact of crystal phases on the shape of


generation and second-harmonic generation (SHG) of blue light are possible. It should
have low acoustic losses to allow the fabrication of SAW devices such as filters. It must be
able to serve as a host for the indiffusion of rare-earth materials such as Er, which pro
duces lasing emissions near 1.5pm when optically pumped. Finally, it must be commer
cially available with high optical quality and at low cost.
The ferroelectric materials lithium niobate (LiNb03), lithium tantalate (LiTa03),
and potassium titanyl phosphate (KTP) are three of the most promising candidates which
satisfy the above conditions and have received widespread use in the fabrication of the
aforementioned devices [hl94a, hl94c, E1H95, Miz92, Nak90, Xu91]. In the next sec
tion, we will compare these crystals and demonstrate why LiTa03 is the favored choice for
the fabrication of most devices.
1.2 The Advantages of LiTaQ3
Historically, considerable effort has gone into the fabrication, characterization, and
modeling of channel waveguides in LiNb03 [Cao92, Paz94, Yam91a], On the other hand,
LiTa03, which is isomorphous to LiNb03, has received much less attention although it has
been shown to possess many advantages over LiNb03. LiTa03 is harder and demonstrates
a photorefractive damage threshold of 1500W/cnr at 514.5nm and room temperature,
about forty times higher than the 40W/cm2 value of LiNb03 [Gla72, Mat92b, Miz91,
Miz92, Saw91, Spi83]. This improved photorefractive resistance should provide better
power handling capabilities of devices fabricated in LiTa03 [Bur93, McW92, Miz91],
making it advantageous for the development of high-throughput modulators and filters, as
well as high-output 1.5pm sources with output power levels far exceeding those attainable


95
scanning distance (jam)
Fig. 4.12 Speaker calibration chart
relating voltage to speaker
excursion. The measured
data (squares) and linear fit
(solid) are shown.
is adjusted to reveal the entire profile, up to
the point of the large intensity, or voltage,
increase marking the substrate surface.
This scan is saved and sent to a computer
where the depth axis is calibrated by know
ing the peak-to-peak voltage input to the
speaker and using the calibration chart to
determine scanning distance in microns.
The vertical axis, or An, is calibrated by
measuring the intensity I, or associated
voltage V, caused by bulk reflection and using Eq. (4-11), noting that AVIV = MU.
4.6.3 Accuracy and Repeatability
The spacial resolution of this measurement system is limited by the spot size.
After bevelling, the effective resolution is improved by the magnification factor. The spot
size of the focused laser beam was estimated experimentally. A sample with a sharp
boundary of Ta metal deposited on the surface was scanned under the spot. The intensity
profile was obtained as the spot traversed the Ta boundary, from bulk region to an area of
higher reflectivity due to the Ta. This reflectivity profile is a convolution of the finite spot
and an ideal step function. The profile is displayed in Fig. 4.13, from which the spot size
was approximated to be about 2|im, given the magnification factor due to bevelling. This
resulted in a measurement resolution of 2/30.76 ~ 0.07|im.


132
Sample
Crystal
Cut
PE Conditions
e33 <10'3)
Phase
Days
8(An)
3
X
glycerin
260C,24hrs
4
a'
7
7x1 O'4
3'
X
Sample 3 annealed
230C, 3hrs
4
a'
7
1.9x1 O'4
50
9.4x1 O'4
126
1.7xl03
4
X
glycerin +
0.005mol% Li2C03
260C, 79.5hrs "
3.6
a'
7
7. lxlO"4
43
2.5xl0'3
74
2.9xl03
5
X
glycerin +
0.005mol% Li2C03
260C, 84hrs
aged: 85C, 4 days
na
na
4
9.9xl0'4
5'
X
more aging of
Sample 5:
95C, 4 days
na
na
7
2x1 O'4
51
2x1 O'3
Table 6.3 Fabrication conditions, measured e"3 value, associated phase, and esti
mated change in index increment 8(An) at various periods after fabrica
tion for Samples 3, 3', 4, 5 and 5".
days
Fig. 6.8 Measured coupling length of Sample
3'. This coupler was comprised of 7pin
channels with an 8p.m gap.


80
channel mask width (|im)
Fig. 4.5 Measured fundamental mode index increments at different wave
lengths for channel waveguides of various widths obtained from the
modified prism coupler method.
0.6328|am, 0.890|am, 1.060|xm, and 1.310|im. A rutile prism with an angle A=45.129 was
used for coupling light out of the waveguide. The refractive index np of the rutile prism at
each wavelength was found from a cubic spline interpolation of the dispersion data for
rutile presented in Appendix A.
The measured mode increments for the fundamental mode at each wavelength are
plotted in Fig. 4.5 as a function of channel mask width. For the experimentally measured


152
Fig. 7.6 Measured input power versus output
power at 1.48pm for the MZI modulator.
fabrication conditions is not known. It may be possible that high concentration
waveguides could be annealed to recover r33 without actually being annealed enough to be
of significantly lower strain or even single-phase.
7.3.2 Power Handling Near 1.5pm
In order to test the power handling capability of this device near 1.5pm, the tech
nique of Section 4.5 was employed. The resulting plot of output power versus input power
is shown in Fig. 7.6. As can be seen, the modulator exhibited a linear response, indicating
no photorefractive damage, for input powers up to nearly 300mW. The fact that the
response of this device deviated from linearity at high input powers when that of straight-
channel waveguides did not is likely due to some slightly asymmetry in the MZI arms or a
non-perfect split at the Y-branch, either one a result of photolithographic imperfections or
limitations.


APPENDIX A
DISPERSION OF RUTILE
The dispersion data for rutile is given in Table A.l. This data is taken from the
manufacturers literature. Index values at an arbitrary wavelength may be found by cubic
spline interpolation.
166


135
24
-i i r i | n i i | i i'~r i ~p i i i 1 p1 i i i -
35
rg i i | i i i i | i ri ¡ pm | i i i i | i i i i.
23
y -5
; ;
22
; ;
21
30
r n -
¡ 20
B
B
: n ;
19
O _


25
I T
18
; j
17
; ;
1fi
?n
*' l l 1 l l ' 1 l l ' 1 l 1 l l Pi ' i~
O 1 2 3 4 5 O 10 20 30 40 50 60
days days
Fig. 6.10 Measured coupling length
of Sample 5. This coupler
was comprised of 7|im
channels with an 8|im gap.
Fig. 6.11 Measured coupling length of
Sample 5". This coupler was
comprised of 7|im channels
with an 8|im gap.
monitored at 1.31|4m and is shown in Fig. 6.11. The stability, however, was not
significantly improved over that of Sample 4.
6.3.4 Limits to Reducing Strain
As the index increment An of PE and APE waveguides is on the order of 10~2 or
smaller for wavelengths near 1.55p.m, the observed instabilities constitute changes in
index increment 8(An) of greater than 10%. Such changes are unacceptable for the fabri
cation of reliable integrated optical devices. As was hinted to here, and is known to be the
case for APE:LiNb03, decreasing the proton concentration, and hence strain, may lead to
improved stability. However, several obstacles limit this achievement.
First, the An of PE and APE waveguides in LiTa03 can be as much as an order of
magnitude smaller than that of LiNb03. Additionally, waveguide devices operating at
1.55|im are required, where, because of dispersion, An values are much smaller than


84
measurement is 0.24dB/cm, comparable to values reported for APE waveguides in
LiNb03 [Hus95]. For the case of Zn vapor-indiffused waveguides, a propagation loss of
0.75dB/cm was measured, comparable to that previously reported for these waveguides in
LiTa03 [Ekn87],
Along with loss, another important characteristic of waveguides is the value of the
electro-optic coefficient r33. In the next section, the method of measuring r33 is outlined
and performed.
4.4 r33 Measurement
For the fabrication of electro-optic devices, such as modulators, it is desirable to
have a large electro-optic coefficient r33. However, the fabrication techniques used here
are known to decrease somewhat the value of r33. As stated earlier, PE, especially in pyro-
phosphoric acid, is known to decrease r33, though it can be recovered to some degree by
annealing. The r33 value of waveguides fabricated in lower H+ concentration sources,
such as glycerin, where annealing is not used, is not known. Additionally, the techniques
of Ti-indiffusion and Zn vapor-indiffusion are performed at temperatures above the Curie
point for LiTa03, so there is no electro-optic effect after waveguide formation. As a result,
repoling is needed to restore r33. A direct measurement of r33 is then needed to gauge the
effectiveness of the repoling process.
In this section, the technique used to measure the electro-optic coefficient r33 is
described and administered. The advantage of the technique used here over other methods
is outlined and some results are presented.


88
140
II1 1 111111111 1 1111 1 11111
O 120
£
PE:LiTa03 *
a ioo


S3 80

o 60


&C

^ 40
Oh
.* APE:LiNb03
P 20
O
0
0
i_llll1llI 1 1lIt l.lijI I 1I I Ii
O 100 200 300 400 500
input power (mW)
Fig. 4.9. Output power vs. input power
for 6pm wide straight-chan
nel waveguides in PE:LiTa03
and APE:LiNb03.
lasing threshold of the laser used in this
range tested. Therefore, little, if any, degra
dation to its performance has been caused
by photorefractive damage. The LiNb03
sample, however, shows some degree of
saturation and appears to have already
passed its threshold point for photorefrac
tive damage for even the smallest input
power tested. Lower levels of input power
were not attainable as they were below the
Iment. The significant attenuation of output
power relative to input power illustrates the extremely high photo-induced refractive index
change in APE:LiNb03 waveguides, especially above lOOmW.
In the next section, a direct method of index profiling is described. Such a mea
surement technique is necessary to extract surface index values of waveguide in order to
reconstruct the structural phase diagram for APE:LiTa03 in Chapter 5.
4.6 Direct Index Profiling
In order for devices to take advantage of the high throughput capability offered by
LiTa03, the technology of reliable waveguide fabrication must first be developed. For the
APE technique of waveguide fabrication in particular, however, a number of anomalies are
known to exist, as was outlined in Chapter 2. In order to explain these anomalies, as well
as identify conditions yielding stable waveguides, the structural phase diagram for


scope of this research, there still remains some work remaining to be performed on the
development of new device technologies in this material, as well as the improvement of
existing devices. This work is described in this section.
8.2.1 Modulator Improvements
The measured frequency response of the modulator fabricated and tested in Chap
ter 7 is reprinted again in Fig. 8.1. As can be seen, there was a large deviation from the
theoretical response at low frequencies (l-2GHz). These frequencies were found to corre
spond to an unusually large microwave reflection peak, and hence a dip in microwave
transmission. The electrode structure fabricated here was the similar to that used in Ref.
[Wan97], having the exact same value of microwave loss coefficient, but on a LiTa03 sub
strate instead of a LiNb03 substrate.
Fig. 8.1 Measured frequency response (squares) and calculated theo
retical response (solid line) for the traveling-wave MZI modu
lator.


44
Eqs. (2-23) and (2-24) can be solved to give the reflection coefficient in terms of the trans
fer matrix elements to yield
Wll+W,t12-m21-V,,22
Wll+WM12 + m21+YSm22
Finally, the reflectance R is defined by
,2
R = \r\ (2-28)
In the next subsection, the actual conditions for the formation and deposition of an AR
coating are given.
2.6.2 The AR Film Parameters and Deposition
Several simplifications were employed in the fabrication of the AR film used for
this research. First, the reflection, loss, and stability measurements were all performed at
near normal incidence, so that all angles in the above derivation are set to zero. Secondly,
quarter-wave thickness films were deposited for each wavelength used in this effort. In
this instance,
t
4nf
(2-29)
which forces the phase difference of Eq. (2-19) to become 8 = 7t/2 so that cos8 = 0 and
sin8 = 1. Using these conditions, the equation for reflectance R, Eq. (2-28), becomes
( 2\2
n0ns~ n f
n0ns + nf2;
R =
(2-30)


102
4.7.2 Calculating Lattice Strain
Previous studies [Fed94] have determined that H:LiTa03 waveguides on X-cut
substrates have only one non-zero component e"3 of the deformation tenser. This value of
strain was determined for each sample by measuring rocking curves with a single and dou
ble-crystal X-ray diffractometer. By measuring the angular deviation A0/J/W between the
peak for the unexchanged substrate and that for the exchanged layer on the rocking curve
of the plane (hkl), e'33 can be found from the expression [Fed94]
-A0
833 tan0
(4-13)
hkl
where 0/?w is the Bragg angle of the strain-free substrate region. For X-cut samples, rock
ing curves were measured for the (220) surface plane. In measurements, the maximum
angular shift on the rocking curve was of primary interest as it corresponds to the maxi
mum deformation e"3 and strain in the APE region. It was confirmed that the maximum
shift corresponds to the strain at the substrate surface. This was done by successive pol
ishing of the sample surface in small steps through the APE region and analyzing the
resulting change in the rocking curve. Likewise, the analysis of the evolution of the rock
ing curve upon annealing has also pointed to a monotonic strain profile with the maximum
reached at the surface. Therefore, the values of both maximum deformation e3 and opti
cal index used to plot the structural phase diagram (Chapter 5) have been measured at the
same point, namely, the substrate surface.
4.7.3 Some Rocking Curve Results
Some measured rocking curves are shown in Fig. 4.19. Fig. 4.19(a) is the rocking
curve for virgin LiTa03. There are two peaks for the substrate because this curve was
taken with a single-crystal diffractometer having two wavelengths. Had a double-crystal


68
The main feature of the two-photon excitation is the nearly linear dependence of
Ans on J for high input intensities J, which has been verified experimentally in LiNb03
[Fuj93, Jer95] and LiTa03 [Gla72], Moreover, it has been established [Kos97a], that
{3(Ans)/3J} ~ N, where N is the concentration of antisite defects. Such a proportionality is
in full accordance with the two-photon model.
It is well known that Er [Gil96] and Pr [Gue97] dopants increase N, especially
bipolarons, in LiNb03 through the need for charge compensation of ions with valence
state (3+). Hence, data obtained here at the excitation wavelength of 785 nm (Fig. 3.7) is
clear evidence of a similar effect for Er3+ in LiTa03. Moreover, the relative change of N
caused by Er-doping can be estimated from a comparison of the slope {3(Ans)/3J}
observed in the Er-doped sample with the one observed in the virgin congruent substrate.
It was found that N is increased 1.6 times by Er-doping in the sample with the highest Er3+
concentration (Fig. 3.7).
Radiation at 785 and 632.8 nm is near the maximum of the absorption band for
small polarons, but is not optimal for the photo-induced dissociation of bipolarons, as
these wavelengths are very far from the bipolaronic absorption band [Sch91]. Therefore,
the generation rate of shallow PR centers is too low to induce significant photorefractivity,
even at high light intensities. However, the situation changes when excitation with two
wavelengths is used, with photons of the shorter wavelength (green light) exciting bipo
larons and increasing the virtual population of single-polaron levels and photons of the
other wavelength (red or infrared) sweeping them effectively into the conduction band
[Gue97, Lan98], as illustrated in Fig. 3.11. Recent research with additional sources of
such "gating" light has shown [Gue97, Lan98] that the two-color scheme has a very high


124
Fig. 6.1 Illustration of a passive directional coupler.
and have different propagation constants P0 and Pi, respectively. The difference in propa
gation constants of these two modes is Ap=P0-P].
From coupled mode theory, the output power ratio of this coupler is expressed as
?2
P.
tm\w+ ep
(6-1)
where 0np represents the amount of residual coupling which occurs at the tapered region
after the interaction length and lc=rc/AP is the coupling length, defined as the length
needed to couple all of the power from guide one into guide two.
6.2.2 Measuring Coupling Length and Index Change
To measure coupling length lc, an array of couplers were fabricated having the
same width w and gap g, but varying interaction lengths L. The output of each arm was
then measured and the normalized output (i.e. P1/(P1+P2)) was plotted as a function of
interaction length L. Such a plot is depicted in Fig. 6.2. Then, the function of Eq. (6-1)
for one arm was fit to the measured data to determine the coupling length lc (Fig. 6.2). It


176
[Miz94] K. Mizuuchi and K. Yamamoto, Domain Inversion in LiTa03 Using Proton
Exchange Followed by Heat Treatment, J. Appl. Phys., vol. 75, no. 3, p. 1311,
1994.
[Mu94] X. Mu, X. Yue, J. Chen, J. Wang, and Z. Shao, Planar Waveguide Refractive
Index Distribution Functions Determined Precisely from Mode Indices,
Applied Optics, vol. 33, no. 15, p. 3227, 1994.
[Nak90] K. Nakamura and H. Shimizu, Ferroelectric Inversion Layers Formed by Heat
Treatment of Proton-Exchanged LiTa03, Appl. Phys. Lett., vol. 56, no. 16, p.
1535, 1990.
[Naz87] M. Nazarathy, D. W. Dolfi, and R. J. Junerman, Spread Spectrum Frequency
Response of Coded Phase Reversal Traveling Wave Modulators, J. Lightwave
Technol., vol. LT-5, no. 10, p. 1433, 1987.
[Nou92] S. Nouh, P. Baldi, M. De Micheli, G. Monnom, D. B. Ostrowsky, E. Lallier,
and M. Papuchon, Laser Action in Nd:LiTa03 Proton Exchange Waveguide
Exhibiting Proton Induced Excited State Lifetime Reduction, Electron Lett.,
vol. 28, p. 2337, 1992.
[Nou95] S. Nouh, P. Baldi, K. El Hadi, M. De Micheli, G. Monnom, D. B. Ostrowsky,
E. Lallier, and M. Papuchon, Fabrication Parameter Optimization of a Low-
Threshold High-Efficiency Proton-Exchanged Waveguide Laser in
Nd:LiTa03, Opt. Lett., vol. 20, no. 13, p. 1468, 1995.
[Paz94] G. R. Paz-Pujalt, D. D. Tuschel, G. Braunstein, T. Blanton, S. T. Lee, and L. M.
Salter, Characterization of Proton Exchange Lithium Niobate Waveguides,
J. Appl. Phys., vol. 76, no. 7, p. 3981, 1994.
[Ped87] F. L. Pedrotti and L. S. Pedrotti, Introduction to Optics, Prentice-Hall, London,
1987.
[Ram88] R. V. Ramaswamy and R. Srivastava, Ion-Exchanged Glass Waveguides: A
Review, J. Lightwave Technol., vol. 6, no. 6, p. 984, 1988.
[Red94] B.R. Reddy, and S.K. Nash-Stevenson, Red-to-Violet and Near Infrared-to-
Green Energy Upconversion in LaF3:Er3+, J. Appl. Phys., vol. 76, no. 6, p.
3896, 1994.
[Reg85] R. Regener and W. Sohler, Loss in Low-Finesse Ti:LiNb03 Optical
Waveguide Resonators, Appl. Phys. B, vol. 36, p. 143, 1985.
[San92] N. A. Sanford, J. A. Aust, K. J. Malone, and D. R. Larson, Nd:LiTa03
Waveguide Laser, Opt. Lett., vol. 17, no. 22, p. 1578, 1992.


143
6.6 Summary
In this chapter, the issue of temporal stability of waveguides fabricated in LiTa03
was examined. As a qualitative investigation, rocking curve analysis was used to track
changes in values of surface strain for APE samples across the structural phase diagram. It
was found that the high-concentration phases exhibited large temporal instability. For a
more quantitative analysis, a directional coupler was used to monitor changes in coupling
length over time, and these changes were related to associated changes in index increment
8(An) through an FDM computer simulation of a directional coupler. For APE
waveguides, it was found that low-concentration waveguides exhibited improved temporal
instability. However, they were still not stable enough. Using the theory that single-phase
waveguides would have superior stability, PE was used to fabricate waveguides in a new
single phase region of the diagram, because of the advantage of PE for better mode con
finement. However, even with aging at low temperatures, results indicated that the stabil
ity was not improved by having waveguides comprised of only a single phase.
Waveguides with simply lower proton concentrations seemed to display the most
improved stability. But a limit to the level to which proton concentration could be reduced
to was reached, based on the extremely low An and poor confinement of these waveguides
at the 1,55pm wavelength.
To determine if stability depended on fabrication conditions, as it does in LiNb03,
metal-indiffused waveguides were examined next. Results showed that all waveguides
were unstable, leading to the conclusion that instability does not originate from ion mobil
ity or induced lattice strain, and does not depend upon fabrication conditions. Rather, it
may be inherent to the crystal, the result of an improper, or immature, growth process.


In the setup, a linearly polarized laser beam was focused, by a 50X objective, to a 2pm
spot on the optical-grade polished edge of the sample under study. The plane of light
polarization was parallel to the optical axis. A Raman microprobe, i.e. a microscope opti
cally coupled to a Raman spectrometer, was used to collect both the out-of-plane scattered
light and the fluorescence emission.
There was a significant difference between the fluorescence spectrum produced by
pumping at 785nm and that produced by pumping at 647nm or 632.8nm. For 785nm,
only intensive Stokes fluorescence (SF) was induced (i.e. fluorescence with photon ener
gies smaller than the excitation energy). For 647nm and 632.8 nm, however, there was
anti-Stokes fluorescence (ASF) with both a green emission at 520-560nm (see Fig. 3.1)
D
od
<
C/3
c

G
O
c
O
C/D
o
G
QG
wavelength (nm)
Fig. 3.1 Anti-Stokes fluorescence spectra excited at 632.8nm and 647nm in
the Er:LiTa03 sample with the highest Er concentration (-0.65
mol%).


3 CHARACTERISTICS OF Er-DOPED LiTa03 49
3.1 Background 50
3.2 Er-Doping Conditions 52
3.3 Raman/Fluorescence Measurements 53
3.3.1 The Raman Effect and Raman Spectroscopy 53
3.3.2 Fluorescence Spectroscopy 54
3.3.3 Fluorescence Measurements 54
3.3.4 Raman Measurements 57
3.4 Energy Transfer Upconversion 62
3.5 Li-Treatment 65
3.6 The Two-Photon Model of the PR Effect: LiTa03s Advantage
in Data Storage and Lasers 66
3.7 Summary 70
4 WAVEGUIDE CHARACTERIZATION AND
MEASUREMENT TECHNIQUES 71
4.1 Near-Field Characterization 72
4.1.1 Description of the Measurement Technique 73
4.1.2 Modal Characterization Results 73
4.2 Effective Mode Index Measurements 75
4.2.1 The Prism Coupling Technique 75
4.2.2 A Modified Prism Coupling Method 78
4.2.3 Measured Mode Index Increments 79
4.3 Waveguide Loss 81
4.3.1 Loss Measurement Setup 81
4.3.2 Total Insertion Loss 82
4.3.3 Propagation Loss 83
4.4 r33 Measurement 84
4.4.1 Interferometric Measurement Method 85
4.4.2 Unity Overlap Method 85
4.5 Power Handling 86
4.5.1 Setup Description 87
4.5.2 Results 87
4.6 Direct Index Profiling 88
4.6.1 The Importance of Direct Profiling 89
4.6.2 The Reflectivity Setup 91
4.6.3 Accuracy and Repeatability 95
4.6.4 Measurement Results 96
4.7 Rocking Curve Analysis 100
4.7.1 The Rocking Curve Measurement 100
4.7.2 Calculating Lattice Strain 102
4.7.3 Some Rocking Curve Results 102
4.8 Summary 103
vi


83
90
80
7 170
7 160
o
* 130
p-
120
110
1500 1500.02 1500.04 1500.06 1500.08 1500.1
wavelength (nm)
Fig. 4.7 Fabry-Perot transmission response of a 5p,m wide PE channel
waveguide.
4.3.3 Propagation Loss
Propagation loss was measured using the waveguide Fabry-Perot resonator tech
nique [Reg85], The setup is the same as that depicted in Fig. 4.6. Here, however, the sam
ple itself acts as a Fabry-Perot resonator. The measurement is made by changing the
wavelength of the laser source and monitoring the measured optical output from the
waveguide. The output as a function of wavelength will have a typical Fabry-Perot
response, resembling somewhat a sinusoidal response, depending on the amount of propa
gation loss. By measuring the ratio of the maximum transmission to the minimum trans
mission, and by knowing the reflectivity of the sample endfaces, the propagation loss is
easily calculated [Reg85].
The measured Fabry-Perot transmission response of a waveguide sample prepared
by PE in glycerin is shown in Fig. 4.7. The value of propagation loss calculated from this


modes can then be calculated by measuring the angular deviation of the corresponding in
line from the normal to the prism surface. In this case, the fundamental mode will have
the largest angular deviation.
While this standard prism coupling technique depends on the measurement of
some angular deviation, and this measurement can usually be made very accurately, it
does have one significant limitation which needs to be mentioned. In many cases, it is the
mode index increment, AN=N-ne, of each of the modes that is of interest, where ne is the
bulk extraordinary index value. The accuracy of making such a determination using this
prism coupling arrangement is then directly proportional to the prism index np and the
bulk index of the substrate ne. Normally, AN needs to be known with an accuracy of 10'4
or better. However, the values of prism and bulk index are usually only known with an
accuracy of, at best, 10 As such, the standard prism coupling method is not a depend
able means to obtain effective mode index increments for APE waveguides. However, a
modification to this technique may be employed which requires less accuracy in the values
of these indices. This new technique is discussed next.
4.2.2 A Modified Prism Coupling Method
The new prism coupling technique [Tav95] is depicted in Fig. 4.4. Light is
coupled into the waveguides via a single-mode fiber, then out-coupled via the prism. As
before, the m-lines corresponding to the different excited modes of the waveguide, as well
as a region delimiting the continuum of substrate radiation, are seen exiting the prism face.
This spectrum is easily observed by placing a screen near to the exiting face of the prism.
The angular deviation A6 between the edge of substrate radiation and the mode of interest
(only the fundamental mode is depicted in the figure) can be precisely measured and


To my friends, family, and loved ones,
for their support.


81
index increments at X= 1.3 lpm, these values were found to be in reasonable agreement
with those obtained in [Tav95] by using a numerical routine of waveguide modeling
[Tav94], under similar waveguide fabrication conditions. This figure also presents some
insight into the wavelength dispersion of the index increment.
In addition to having some knowledge of the modal characteristics and magnitude
of the index increment, the performance of many devices often depends on the losses
incurred within the waveguide. As such, these losses need to be accurately measured.
This is the subject of the next section.
4.3 Waveguide Loss
Waveguide loss is an important characteristic affecting device performance. It is
usually desirable to have as low loss as possible, in order to minimize the fiber-to-fiber
insertion loss of devices, and maximize gain in waveguide lasers. In this section, the tech
niques used to measure both total insertion loss and propagation loss are described and
some measurement results are provided.
4.3.1 Loss Measurement Setup
The experimental setup used to measure both total insertion loss and propagation
loss is depicted in Fig. 4.6. Single-mode, polarization-maintaining fiber from an HP
Fig. 4.6 Experimental setup used for measuring waveguide loss.


86
to couple light into a waveguide sample. Then the outputs of both the other coupler arm
and the waveguide are imaged, on top of each other, so they interfere on an IR camera.
The interference pattern is seen on a video monitor. The waveguide sample had Ta depos
ited on both the waveguide surface, with a Si02 buffer layer to reduce loss, and the bottom
surface. The Ta served as electrodes and guaranteed an overlap factor of unity. Of course,
the limitation of this technique is that it can only be used on Z-cut samples.
To make the measurement, a voltage was applied to the electrodes and the fringe
pattern was observed on the monitor. The voltage V required to shift the fringe pattern by
one-half period (<|)=n) was recorded and the r33 value was calculated from:
r33
_
VLn
(4-7)
where g is the sample thickness and L is the sample length. The accuracy of this technique
was estimated to be +6%.
For a waveguide fabricated by PE in glycerin at 260C for 24 hours, the measured
r33 value was 14.5 pm/V. For a Zn vapor-indiffused sample which was repoled, the mea
sured r33 value was 27.1 pm/V. This value is close to the r33 value of 30pm/V for bulk, vir
gin LiTa03.
In the next section, the issue of power handling of LiTa03 waveguides is exam
ined. In particular, the photorefractive characteristic of PE waveguides in the infrared,
near 1.5|im, is measured for similar structures in both LiTa03 and LiNb03.
4.5 Power Handling
LiTa03 is known to have a threshold for photorefractive damage of more than an
order of magnitude larger than that of LiNb03 at visible wavelengths [Gla72]. In the


62
LiNb03 showed that Er-doped LiTa03 has a much smaller (2.2-4.5 times) PR index
change than that of Er-doped LiNb03 in the intensity range studied. The reason for this
advantage will become clearer when we discuss the two-photon model for PR damage
later in this chapter. Nonetheless, it indicates that LiTa03 is the more attractive host for
active Er-doped devices where high PR resistance is indispensable.
In the next section, the mechanism responsible for the upconverted emissions,
which lead to enhanced photorefractivity, is investigated.
3.4 Energy Transfer Upconversion
In the last section, it was determined that the upconverted emissions act as a self
gating source to, along with the pump source, enhance photorefractivity by the two-pho-
ton, or two-color, photorefractive effect. In this section, the mechanism creating the
upconverted emissions is investigated. Particular attention is paid to the process of energy
transfer upconversion between Er3+ clusters.
The two-photon, multistep processes associated with the red and green ASF emis
sions are exemplified in Fig. 3.8 which shows a partial energy-level diagram for the Er3+
ion. Electrons in Er3+ ions are first excited to the metastable 4F9/2 state, then excited to the
2(4F9/2) state, after which they decay nonradiatively to the4Gjj/2, 2H||;2 or 4S3/2 state.
The red emissions at 613-628nm correspond to a transition from the 4G31/2 state to the
4Ij j/2 excited state, while the transitions from the 2H] 1/2 and 4S3/2 states to the ground
state produce the green emissions. The substructure of the green and red emissions
reflects the splitting of the ground state and excited states into Stark sublevels. It is impor
tant to note that the red emissions at 613-628nm observed here for Er:LiTa03 were not


157
determined. By measuring group delay along the electrode structure a microwave effec
tive index of 4.1 was calculated. A characteristic impedance of 40£2 was found from plots
of impedance versus electrode W/G ratio.
The DC response was then measured and the power handling capability of the
device near 1.5|im was examined. From the measured DC response, an extinction ratio of
-27dB was obtained, along with a switching voltage of 13.6V at DC. As far as power han
dling, the device tracked a linear response of output power versus input power for input
power levels near to 300mW.
Finally, the frequency response of this device was measured from 45MHz to above
10GHz. It was found that the device followed theoretical performance at most frequencies
except for the range of l-2GHz, where a large microwave reflection resulted in diminished
modulation efficiency. The unusually large reflection, as compared to that observed in
LiNb03 for the same electrode design and loss coefficient, was the result of a larger
impedance mismatch for LiTa03. This situation could be improved by changing the W/G
ratio of the electrode to reduce such a mismatch.
In the next chapter, the conclusions to this work are given. The major contribu
tions of this effort are detailed and some future work remaining to be performed is out
lined.


146
waveguide
Mach-Zehnder
Fig. 7.1 Depiction of a symmetric MZI modulator with CPW travel
ing-wave electrodes in X-cut LiTa03.
loss, effective index, and characteristic impedance are determined. Some DC response
measurements are made next to evaluate the switching voltage \n and extinction ratio.
The power handling capability of this structure near 1.5[im is also examined. Lastly, the
frequency response of this modulator is measured and compared to theory.
7.1 Fabrication Conditions
A symmetric MZI modulator was fabricated in X-cut LiTa03 having a coplanar
waveguide (CPW) traveling-wave electrode structure as depicted in Fig. 7.1. The condi
tions for its fabrication are given in this section.
The MZI was fabricated first using PE in glycerin at 260C for 24 hours. The
choice to use PE was merely for simplicity. Any of the other waveguide fabrication pro
cesses could have been used. Next, the sample endfaces were polished using the tech
nique outlined in Section 2.5. Finally, to minimize rf loss in the electrodes, thick


Hua96], There are two mechanisms of reconversion. The first is the homogeneous
upconversion caused by excited-state absorption (ESA) within non-clustered Er3+ ions.
This process occurs wherever there are Er3+ ions and therefore is inherent to Er-doped
waveguides. The second mechanism is nonradiative energy transfer upconversion (ETU)
which occurs in Er3+ cluster sites. This form of upconversion can be suppressed by
decreasing the relative concentration of cluster sites through a proper choice of doping
technology [Ara92, Del93] and/or post-indiffusion processing [Die98, Gil96]. For exam
ple, in the case of Er:LiNb03 the cluster site concentration depends on the Li/Nb ratio
[Gil96] and, therefore, upconversion efficiency may be reduced significantly by using a
vapor phase equilibration technique to increase the Li content [Die98, Gil96]. The simi
larity of intrinsic defect structures in LiNb03 and LiTa03 [Gop96] would therefore sug
gest that an increase of the Li/Ta ratio should also reduce upconversion caused by ETU in
Er:LiTa03.
As a result, the effect of Er-doping on the physical and optical properties of
LiTa03 needs to be examined. In particular, how doping affects the PR property of the
crystal and how much upconverted emission from Er3+ clusters, which reduce gain
[Gil96], is present, as well as methods for its reduction. This study begins in the next sec
tion by outlining the conditions used for Er-doping in LiTa03.
3.2 Er-Doping Conditions
The following factors need be considered when Er-doping with the aim of produc
ing active gain regions for waveguide lasers. First, Er3+ concentration should be at an
appropriate level so as to provide significant gain. Next, Er3+ concentration profiles


130
this research were done so at 632.8nm. Because the index increment is known to exhibit
some dispersion, the actual value of surface index for waveguides operating at 1.55pm is
much lower than that shown in the diagrams, at least two times [E1H98]. The X-cut crys
tal seems to allow for a higher value of surface index increment in the a phase then does
Z-cut, and also does not exhibit such a sharp discontinuity between the a and K phases.
Additionally, the lower-concentration source was expected to have a slower rate of surface
indiffusion, allowing time for diffusion within the sample, thus producing waveguides
whose surface concentration never leaves the a phase during PE and, therefore, would not
require annealing.
The characteristics of waveguides fabricated by PE in X-cut LiTa03 using glycerin
and glycerin diluted with Li2C03 were detailed in Chapter 5. It was found that these con
ditions produced waveguides with a predominately single phase, similar to the known a
phase, but capable of higher An values and higher strain [MarOO]. This new phase was
called the a phase [MarOO] and the revised structural phase diagram is reprinted in Fig.
6.6 for ease of reference. As these waveguides were fabricated by only a one-step, PE
0.04
technique, they were relatively step-like
P1.P2
as compared to their APE counterpart. As
<3 0.02
O
D
on
0.00
a result of this and their larger An values,
mode confinement within the waveguide
was significantly improved, leading, for
0 2 4 6 8 10 12
surface £33 (xlO3)
example, to an increase in coupling length
Fig. 6.6 Structural phase diagram for
X-cut APE:LiTa03.
from about 5 mm to 20mm in couplers
with 6.5|im channels and an 8|im gap.


115
Raman shift, wavenumber (cm1)
Fig. 5.5 Raman spectra of PE samples prepared in pure glycerin,
and glycerin diluted with L2CO3, at 260C for various
exchange times.
exchanged in pure glycerin for times of 24, 48, and 70 hours, and the spectra for an APE
sample with a strain value of e"3 =0.84xl0"3, in the a phase. As can be seen, the sample
exchanged for 24 hours had a spectra somewhat similar to the APE, single-phase sample.
However, samples exchanged for longer periods exhibited a shift in the peak of the Raman
spectra, indicative of samples containing higher concentration, such as K, phases. Fortu
nately, 24 hours of exchange was enough to produce single-mode waveguides at 1.55p.m.
The Raman spectra of samples prepared in diluted glycerin are presented in Fig.
5.5, where they are shown with the spectra for the pure glycerin sample exchanged for 24
hours. As can be seen, all the diluted samples showed a more predominately single-phase
composition then even the pure glycerin sample. Additionally, the spectra did not change


36
attached and placed inside the box furnace with the wire leads running out the vent at the
top of the furnace. The wire leads were connected to a DC power supply. The sample was
heated to 700C and a voltage of 30V was applied. The oven was immediately turned off
and allowed to cool. Once it had cooled to 450C, the voltage was removed and the sam
ple allowed to continue cooling to room temperature. As will be seen later in Chapter 4,
this poling process was successful and resulted in complete recovery of the electro-optic
coefficient r33.
In the next section, the technique of end-face polishing is described. A good end-
face polish in necessary for accurate optical measurements and low insertion loss.
2.4 End-Face Polishing of Channel Waveguides
In order to minimize coupling losses into the channel waveguides, good quality
endfaces of the sample, perpendicular to the waveguides, are required. A polishing
scheme is outlined here which achieves satisfactory results.
o
First, about 3000-4000A of Si02 is e-beam deposited on the waveguide surface of
the sample to be polished to protect this surface from scratches during the mounting and
polishing process. Then the sample is adhered to a Si wafer, using high temperature
QuickStick clear wax (melting point of 135C), with the waveguide surface pressed
against the Si. In this manner, a thin (~lpm or less) layer of wax can be formed between
the sample and the Si wafer which serves to minimize cracking and chipping of the sam
ple endface on the waveguide surface. The Si wafer is then cleaved along all four edges of
the sample and polished by hand in silicon carbide (SiC) powder, without using a polish
ing jig, until the Si is flush with the LiTa03 on all sides. This is done to reduce cracking


takes place. The result is a sample with a waveguide pattern of Ti metal on the surface, as
depicted in Fig. 2.9 for the case of a straight-channel waveguide. The sample is placed
inside a platinum box with powered LiN03, and the box is loaded inside a tube furnace.
Dry 02 is flowed through the tube of the furnace at 20 ccm while the sample is indiffused
at 1200C for 20 hours, with an 8/min rise and fall ramp rate. The purpose of the pow
ered LiN03 and flowing 02 is to help reduce Li20 outdiffusion from the sample surface.
These conditions result in single-mode waveguides for both polarizations at 1.55jim
immediately after fabrication.
2.2.2 Zn Vapor-Indiffusion
Zn vapor-indiffusion is performed by placing a sample inside a vacuum sealed
quartz ampule with some amount of Zn wire, granules, or powder and heating the assort
ment to a temperature above the boiling point for Zn so that it vaporizes and diffuses into
the sample. The Zn concentration profile obtained resembles a complementary error func
tion erfc(y/d) distribution, where d is the depth at which the concentration is 14.7% of that
at the surface [Ekn87], As with Ti indiffusion, this process results in an increase of both
the extraordinary and ordinary indices [Y0088]. Likewise, the change in the extraordinary
index is larger than that of the ordinary index [Y0088]. Again, the diffusion equations
governing depth and diffusion coefficients are similar to those used for proton exchange
and Ti indiffusion, namely Eqs. (2-2) and (2-3) [Y0088].
Similar to Ti indiffusion, Zn vapor-indiffusion does not appear to show any of the
same anomalies as APE, and there is no measured change of crystal lattice strain induced
by this process. However, the stability of waveguides fabricated in LiTa03 by this process
have not yet been confirmed. Advantageously, the lower temperature of Zn-indiffusion


57
and a red emission at 613-628nm (Fig. 3.2), in addition to the SF. Within the measurement
range of the input intensity J, the SF intensity ISF changed proportionally to J at all excita
tion wavelengths in accordance with previous studies [Ara92, Del93, Gil96], The depen
dence of the ASF intensity IASF of the green emission was found to be more involved and
had a knee at a certain value Jt of J, with Jt ranging from 104 to 1.1*104 W/cm2 for
632.8nm pumping and about 4103 W/cm2 for 647nm pumping. As illustrated in Fig. 3.3,
IASf(J) is a nonlinear function I/^sf^J* for J and about 1.7 for 647nm pumping. For J>Jt, there is a linear correlation between IASF and
J, i.e. x = 1. The dependence of IASF of the red emission had the same character as that of
the green emission in Fig. 3.3, but the intensity of the anti-Stokes red emission was
smaller than that of the green emission by about 4.5 times. Such an approximately qua
dratic input-intensity law, observed at a low pump range (< 4*103 W/cm2), suggests that
the green and red ASF emissions are attributed to a two-photon absorption process, to be
described in Section 3.4. Reduction from a quadratic to a linear dependence at high input
intensities J is attributed, in accordance with Ref.[Ara92], to population depletion of the
ground state.
3.3.4 Raman Measurements
According to Ref. [Gla72], green light with the observed intensity IASF (see Fig.
3.3) can not alone induce significant single-color PR effect in LiTa03. However, a
comparison of the data for IASF (Fig. 3.3) with the results reported for LiNb03 [Lan98]
and Pr:LiNb03 [Bai97, Gue97] leaves open the possibility that the secondary green emis
sion of Er3+ clusters has enough photon energy and intensity IASF to effect, together with
the pumping red light, significant photorefractivity via two-color excitation in Er:LiTa03.


Compound is used with a generous amount of sodium hydroxide (NaOH) dissolved in DI
water on a Buehler Microcloth. After polishing for six minutes, the sample is removed
from the Microcloth wheel and quickly submerged in the NaOH solution to prevent a
glassy film from forming on the LiTa03 surface.
In the next section, the theory and fabrication conditions associated with the appli
cation of an anti-reflectivity coating are described. In many instances, such a coating was
necessary to minimize reflections and reduce measurement noise.
2.6 Antireflective Coatings
For the measurements of Chapter 4 where light is reflected from the beveled sur
face of a planar waveguide, it was found that reflections from the back surface of the sam
ple substrate were contributing considerable noise into the experimental setup. In
addition, for the loss measurements of Chapter 4 and the stability measurements of Chap
ter 6, reflections from the endfaces of waveguide samples were also introducing some
noise and error into the setup. For each of these cases, much of the impact of these reflec
tions was eliminated through the use of a single-layer antireflective (AR) coating on either
the back side or the endface of the substrate, as the case may be. Some of the theory used
in the development of these AR coatings is presented in this section, as well as the mate
rial and thickness needed for the wavelength used and the method of its deposition.
2.6.1 Thin Film Theory for Single-Laver AR Coatings
The development here, though only concerned with a single-layer film, stems from
the derivation of the transfer matrix method of multiple thin film layers treated in [Hus95,


125
L (mm)
Fig. 6.2 Example of a plot of normalized output power versus interaction
length L for an array of directional couplers. Measured data is shown
as squares and the theoretical fit to the data, shown as a solid line, indi
cates a coupling length of 2.27mm.
has been shown [Che91] that it is possible to determine lc from as little as two couplers
with close interaction lengths.
To quantify stability, the coupling length was monitored over time in order to track
any change due to index instability. Changes in measured coupling length were then
related to corresponding changes in index increment 8(An) through a Finite Difference
Method (FDM) computer simulation of a directional coupler [MarOO], For each fabrica
tion condition, the index profile of the waveguides comprising the directional coupler were
measured by reflectivity profiling (Chapter 4) so that the appropriate parameters describ
ing the profile could be used in the simulation. The accuracy in measuring coupling length


112
Fig.5.3. Fig. 5.3(a) shows the index profile for an as-exchanged waveguide with a surface
e"33 value of 11.4, at the high-concentration end of the (34 phase. This is the highest strain
attainable without surface cracking. As seen, there is no surface index change. However,
there are large index variations below the surface. Index profile measurements performed
at different wavelengths showed no change in the period of the observed index varia
tions, allowing for the conclusion that the variations are not an interference phenomenon.
Rather, the presence of multiple phases and resulting non-uniform strain is likely to be the
cause of the index variations observed. Indeed, rocking curves for this sample also clearly
indicate the presence of at least two phases within the APE layer, [34 and (33, with the (34
phase being the uppermost, or surface layer. Lower-concentration phases ((32, (31, K, and
a) are also present deeper within the APE layer, but the associated strain profile is too
gradual to allow these phases to be easily discerned on the rocking curves.
Fig. 5.3(b) depicts a waveguide with a step-like index profile and a smaller surface
e"3 value of 6.03, as may be obtained through annealing. Rocking curves for this region
near the low-concentration end of the (34 phase show an increasing content of the addi
tional [33 phase.
Fig. 5.3(c) shows a profile for a waveguide with a surface e"33 value of 5.16, essen
tially in the (33 phase. Waveguides in this region are characterized by buried index profiles
with a peak value Ane = 0.016 reached below the surface, nearly the same value as that of
the low-concentration end of the (34 phase. Rocking curves also reveal the distinct pres
ence of a [32 phase layer, beneath the [33 layer. Note again that the index variations, and
the fact that the waveguide profile is buried, are due to the presence of multiple phases.


of the electro-optic effect, the r33 coefficient can be almost completely recovered through
o
post-exchange annealing of the substrate [Ahl94c, Yuh92], This entire process of ion
exchange followed by annealing is referred to as the annealed proton exchange (APE)
technique. Annealing causes further diffusion of the H+ ions into the substrate, expanding
the area of the waveguide and decreasing the propagation loss [Kan94], In the case of
LiTa03, its extraordinary index increment Ane after proton exchange has been shown to be
lower than that of LiNb03. However, annealing causes an anomalous index increase in
LiTa03, compared to an anticipated decrease for LiNb03 [hl94c, Mat92b, Yuh92], This
distinct feature of APE in LiTa03, along with others, will be more thoroughly discussed in
the Section 2.1.2.
The APE technique is among the most suitable for not only the fabrication of chan
nel waveguides in LiTa03, but also for the formation of domain-inverted regions
[hl94b]. These areas of inverted domain are necessary in the fabrication of integrated
devices intended for nonlinear optical applications, like blue light generation by frequency
domain inverted
Fig. 2.2 Depiction of an APE channel waveguide laid
across regions of domain inversion, also formed by
APE.


6
1.3 Fabrication Processes in LiTaQ3
Several methods exist for the formation of waveguides in ferroelectric materials.
For LiTa03 in particular, there are at least three recognized procedures. They are Ti metal-
indiffusion, Zn vapor-indiffusion, and annealed proton exchange (APE). Additionally,
indiffusion of rare-earth materials, such as Nd and Er, have been examined for the produc
tion of laser sources. Though these techniques are all described in greater detail in subse
quent chapters, they are introduced here briefly to point out some of the limitations and
anomalies associated with each and therefore develop the motivation for this work.
In order to create an active gain region for the facilitation of a laser source or
waveguide amplifier in LiTa03, a rare-earth material such as Nd or Er needs to be indif-
fused. Nd, when doped in a host material and optically pumped, is known to produce las
ing emissions near l|im [Nou95, San92], For communication applications, however, it is
desirable to have emissions near 1,5(im, to take advantage of the decreased loss and low
dispersion of optical fiber at this wavelength. Er-doped material when optically pumped is
known to produce lasing emissions near 1.5|im [Gil96]. Though previous work has con
centrated on the characterization of Er-indiffusion in LiNb03 for the fabrication of
waveguide amplifiers and lasers [Ami96, Bau96, Hua96], little or no work has been
reported on the indiffusion of Er into LiTa03. As a result, further research is needed to
determine the conditions for indiffusion of Er into LiTa03, identify optimum levels of
Er3+ concentration, investigate the effects of Er-doping on the optical properties of the
crystal, and explore the various methods for waveguide fabrication after Er-doping to
determine the most ideal technique in terms of laser and amplifier performance.


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 EVATKABFC_BALQM5 INGEST_TIME 2015-02-05T21:13:42Z PACKAGE AA00024531_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


APPENDIX B
DISPERSION OF LITHIUM TANTALATE
The dispersion of the ordinary and extraordinary refractive indices of LiTa03 is
given in Table B. 1. This data is taken from the manufacturers literature. Index values at
arbitrary wavelengths may be found by cubic spline interpolation.
168


CHAPTER 7
EXPLORING THE POTENTIAL OF LiTa03:
A TRAVELING-WAVE ELECTRO-OPTIC MODULATOR
Integrated-optical devices fabricated in LiTa03 have the advantage of benefiting
from the crystals higher threshold of photorefractive damage. As was deduced in the last
chapter, however, some work on the material growth aspect of this crystal still needs to be
performed before the issue of waveguide stability can be resolved and reliable waveguide
devices can be fabricated. For such work to take place and the crystal quality and growth
process improved, a stronger demand for this crystal in applications other than SAW appli
cations needs to be developed, as was done for LiNb03 [INS89], In this regard, this chap
ter demonstrates the potential that LiTa03 offers for integrated-optical device applications
by fabricating and testing a traveling-wave electro-optic modulator. The successful manu
facturing of this device will illustrate that there are no other unforeseen device fabrication
issues to be conquered in LiTa03 so that when crystal quality is improved, many of the
currently used techniques of device fabrication in LiNb03 may be directly applied to
LiTa03.
This chapter begins by outlining the fabrication conditions used in the develop
ment of a Mach-Zehnder interferometer (MZI) modulator, from waveguide formation to
electrode electroplating. Next, some microwave characteristics of the traveling-wave elec
trode structure are examined. The Sjj and S2] parameters are measured and microwave
145


154
In the next section, the frequency response for this device is measured at frequen
cies from 45MHz to above 10GHz. The measured response is then compared to theory.
7.4 Frequency Response
The frequency response for this device was measured by using the small signal fre
quency-swept method [Ueh78, Wan97]. The setup is shown in Fig. 7.7. Light at 1.55pm
from a tunable laser source was end-fire coupled into the MZI using a single-mode, polar
ization-maintaining fiber. A 1kHz, 2V peak-to-peak (-1V to +1V) signal from a frequency
generator was used to AM modulate the microwave signal which was coupled to the trav
eling-wave electrode via cables from the network analyzer S-parameter test set. This same
signal was used for phase matching at a lock-in amplifier. The modulator was biased at its
minimum output point by applying a DC bias voltage to the bias port of the network ana
lyzer test set. The modulated optical output light was captured with a multi-mode fiber
and focused by a lens onto a low-speed IR detector. The microwave frequency was manu
ally swept from 45MHz to above 10GHz
and the amplitude of the low-frequency
envelope, which was proportional to the
high-speed modulation efficiency, was
measured by using the low-speed photo
detector and the lock-in amplifier.
0 2 4 6 8 10 12 14
frequency (GHz) After measuring the frequency
Fig. 7.8 Powei calibiation plot depict- response of the modulator, the power out
ing sweeper output power
veisus fiequency. put Qf the microwaVe sweeper source
m o
T3
-4
3
Q,
4
o
5-h
& -6
£
cn
p-r-r-,
a
rTr~rn
r* vi
ift....
: In
:
f
PfftT
] n
;
r
T
R...
;
:
i
3 n
1
i
n
I

I
1 :


63
(ETU)
632.8 or
647nm
pump
r\yv
red ASF
613 628nm
y\yv
SF
650 660nm
/xyv
785nm
pump
y\yv
-2(4F9/2)
4g
SF
820nm
/\yv
11/2
Ml/2
' a3/2
4p
" ^9/2
4,
9/2
11/2
green ASF
510 570nm
/a rv
115/2
Fig. 3.8 Partial energy-level diagram of Er3+ ions in LiTa03 depicting the
radiative transitions observed in our measurements. The ETU
process, radiative transitions, and nonradiative transitions are rep
resented by thin solid arrows, thick solid arrows, and dashed
arrows, respectively.
observed in previous studies of other Er-doped materials [Ara92, Del93, Die98, Red94],
including Er:LiNb03, pumped via the 4F9/2 manifold {G196].
Upconversion is known to be strongest in regions of highest Er3+ concentration
[Ara92, Gil96], where Er3+ clustering is the greatest. As such, the dominant mechanism of
upconversion leading to the red and green ASF emissions observed in Er:LiTa03 was con
cluded to be ETU, that is, upconversion by energy transfer between neighboring Er3+ ions
within an Er3+ cluster. To confirm this, the effect of the Er-doping level on upconversion
efficiency was examined by profiling, in the depth direction, both the ASF and SF emis
sions within the Er-indiffused near-surface layer. Since ISF was known to be proportional


126
was better than 0.1mm and the sensitivity of the directional coupler allowed for the track
ing of index changes on the order of 10'5 [Mat92a],
In the next section, the method described in this section of measuring index change
using a directional coupler is applied to APE and PE waveguides. The stability of
waveguides in different regions of the structural phase diagram is examined in effort to
pinpoint fabrication conditions leading to improved temporal stability.
6.3 APE and PE Stability Results
In this section, the results of stability tests for waveguides in different regions of
the structural phase diagram for APE:LiTa03 are presented. First, APE waveguides in
both high-concentration and low-concentration regions are examined. Based on these
results, the move to low-concentration PE waveguides is made for a number of advantages
which will be discussed. Stability results on these PE waveguides are given next and the
section concludes with a discussion on the trade-off associated with decreasing strain to
potentially improve stability and the reduction of An that occurs with decreased strain.
6.3.1 APE Results
The first couplers tested were fabricated in Z-cut LiTa03 using the APE process
and pyrophosphoric acid as a proton source. The conditions for fabrication are given in
Table 6.2. For ease of reference, the structural phase diagram for Z-cut APE:LiTa03 is
reprinted in Fig. 6.3. A Z-cut crystal was chosen because, especially for the high surface
concentration sample (Sample 1), it was extremely difficult to produce a waveguide in X-
cut crystals (using pyrophosphoric acid) deep enough to support a mode at 1.55|im


153
juinn
1kHz, 2V
pp
DC source
frequency
generator
N.A. Test Set
port 1 port 2
N.A. Sweeper
n
microwave cable
lens
MM fiber
detector
test fixture
SM fiber
lock-in
amplifier
Fig. 7.7 Setup used for measuring frequency response by the small
signal frequency-swept method.


It has been established [Chu91] that the impedance mismatch of LiTa03 is more
serious than that of LiNb03, leading to much higher microwave reflections in LiTa03 for
the same W/G ratio of the electrodes. As a result, frequency response could likely be
improved by changing the W/G ratio of the electrodes to minimize microwave reflections.
Further advances to this device could also be made by reducing the half-wave volt
age WK. As the tested modulator was made by PE, it had an electro-optic coefficient r33
which was nearly half that of the bulk value. To improve this, the device could be made by
metal-indiffusion followed by repoling, shown in this research to return the r33 to nearly
waveguide
Mach-Zehnder
microwave source
domain reversal
Fig. 8.2 High-speed traveling-wave Mach-Zehnder interferometer modu
lator in (a) standard configuration with baseband frequency
response and in (b) domain-reversed configuration with bandpass
response.


should be at least as deep as waveguide index profiles so that the optical mode has good
overlap with the gain medium.
For this study, experimental Z-cut LiTa03 substrates were doped near the surface
by indiffusion of a 10-12 nm thick Er layer at 1200C for 100-120 hours. These fabrica
tion parameters were chosen so as to provide an Er3+ concentration profile appropriate for
active guided-wave devices. Some of the Er-doped LiTa03 samples were processed in
molten LiN03 at 280C for times ranging from 1 to 44 hours. Such a treatment was
reported [Kos97a] to be effective toward increasing the Li content in a near-surface layer
(with depth 5-10 |im) of LiNb03.
In order to study the impact of Er-doping on the photorefractive properties of the
crystal and to be able to check for upconversion emissions, Raman and fluorescence mea
surements were needed. The next section describes these techniques.
3.3 Raman/Fluorescence Measurements
Raman and fluorescence spectroscopy are useful tools for examining the effect of
Er-doping in LiTa03 by allowing for the measurement of upconverted emissions and
gauging the amount of photorefractive damage induced. In this section, the techniques of
Raman and fluorescence spectroscopy are detailed. Then the results of measurements
using these methods are presented.
3.3.1 The Raman Effect and Raman Spectroscopy
When monochromatic laser light is incident upon a sample, it is scattered by the
material. Most of the scattered light is of the same frequency as the incident light


43
Ea = Ebcos5 + Bb{jY^)
Ba = Eb(jy fSinS) + BbcosS
Eqs. (2-20) and (2-21) may be written in matrix form as
(2-20)
(2-21)
(2-22)
This 2x2 matrix of elements mll5 m12, m21, m22 is known as the transfer matrix of the
film.
Now, by substituting the portions of Eqs. (2-7), (2-8), (2-15), and (2-16) which
represent the E-fields outside of the film region (above (a) and below (b)) into Eq. (2-22),
we can finally relate the fields incident and reflected from the film with those transmitted
through the equations
1 + r = m^t + st
y0(l-r) = m2lt + m22yst
where we define the reflection and transmission coefficients by
r =
t =
(2-23)
(2-24)
(2-25)
(2-26)


155
Fig. 7.9 Measured frequency response (squares) and calculated theo
retical response (solid line) for the traveling-wave MZI modu
lator.
needed to be measured so the collected data could be properly calibrated. This was done
by attaching the microwave cable lead from the sweeper directly to a calibrated microwave
power meter and measuring power over the frequency range of interest. The resulting plot
of sweeper output power versus frequency in shown in Fig. 7.8. As can be seen, the power
out of the sweeper changes significantly with frequency.
The power calibration data of Fig. 7.8 was then used to calibrate the measured fre
quency response for the modulator. The resulting measured and calibrated response is
shown in Fig. 7.9. Also shown is the theoretical curve for the frequency response, calcu
lated using closed form expressions [Naz87] with an optical refractive index N0=2.12,
microwave effective index Nm=4.1, interaction length L= 1.0cm, microwave loss coefficient
a0=TOdB/cm* VGHz, characteristic impedance Z0=4O£2, and source impedance of 501


40
Ped87], The film is assumed to be homogenous and isotropic. The film is also assumed to
be thin enough so that the path difference between multiply reflected and transmitted
beams within the film region is smaller than the coherence length of the light. Therefore,
the reflected and transmitted beams can completely interfere.
We begin by referring to Fig. 2.12. The condition shown is that for TE polariza
tion, which is the only case we are concerned with here. Later, the stipulation of normal
incidence will be made so that there is no distinction between the TE and TM polariza
tions. Additionally, it is noted that a convention was chosen whereby as normal incidence
is approached, the incident and reflected electric field vectors E are oriented in the same
direction, while the magnetic field vectors B switch direction. Though this will introduce
a phase shift in the B field at 0 = 0, conventional optics defines the direction of polariza
tion as that of the E-field direction and hence the orientations are acceptable. The quanti
ties E, with subscripts, define the magnitudes of the electric fields at each boundary (a)
and (b). For example, Er] is the sum of all multiply reflected beams at interface (a) which
are exiting the film toward the region with index n0; Ei2 is the sum of all beams incident at
interface (b), directed toward the substrate region with index ns; and so on.
By applying the boundary condition that the tangential components of the E and B
fields be continuous across each interface, we get
(2-7)
Eb ~ Ei2 + Er2 ~ Et2
(2-8)
B = 5ncos0-B tCOsG. = B., cos0f1 B-, cosG.,
a w i r1 i t1 tv ii t1
(2-9)
Bb = Bi2COsQn-Br2Cosdtl = Bt2COset2
(2-10)


108
5.2 Constructing the Structural Phase Diagram
In this section, several planar APE waveguides were fabricated and measured via
direct index profiling and rocking curve analysis. From this data, the structural phase dia
gram for X-cut APE:LiTa03 was constructed.
To reconstruct the structural phase diagram for X-cut LiTa03, APE planar
waveguides were fabricated using a variety of proton sources: pyrophosphoric acid, ben
zoic acid, ammonium dihydrophosphate, stearic acid and palmetic acid. Exchange and
anneal times and temperatures were varied as deemed necessary to produce a sufficient
number of data points across the structural phase diagram. It was found that pyrophospho
ric acid and ammonium dihydrophosphate provided the highest proton concentrations pos
sible without causing surface cracks on the sample. This limit of proton concentration is
reached after exchanging in pyrophosphoric acid at 260C for 20 minutes. Further
exchanges lead to surface cracking, which, with the associated concentration induced lat
tice deformation, determined the upper limit of the reconstructed diagram.
The fabricated samples were subjected to direct index profiling and rocking curve
measurements. Values of surface extraordinary index increment Ane were obtained from
observation of the measured index profile. Values of surface lattice strain e"3 were calcu
lated from rocking curve measurements, as detailed in Chapter 4.
In Fig. 5.2, surface values of the extraordinary index change Ane are plotted versus
surface values of the lattice strain e"3 to construct a complete structural phase diagram for
APE planar waveguides in X-cut LiTa03. This figure illustrates that the index change has
an approximately linear dependence on lattice deformation, and therefore on proton con
centration, within each phase. However, the slope of this dependence is different for each


CHAPTER 4
WAVEGUIDE CHARACTERIZATION AND MEASUREMENT TECHNIQUES
In the fabrication of integrated optical components incorporating channel
waveguides, it is often necessary to know certain characteristics of the waveguide, such as
its region of single-mode operation and the depth profile of its refractive index increment,
in order to properly predict device performance. Additionally, the proper performance of a
device may rely heavily on additional waveguide features, such as it being of low loss and
having a large electro-optic coefficient. As a result, it is necessary to be able to accurately
measure and determine these characteristics in order to effectively design, model, and fab
ricate reliable waveguide devices.
This chapter provides the measurement techniques and characterization results of
the channel waveguides whose fabrication methods and conditions were given in Chapter
2. All of the waveguides were fabricated in LiTa03 with the aim of single-mode operation
at the wavelength of 1.55pm. First, a technique of near-field characterization is described
and used to determine the region of single-mode operation by coupling light from a
1.55pm laser into various width guides and focusing the waveguide outputs onto an IR
camera, while observing the mode pattern on a monitor. Next, a modified prism coupling
technique is applied to measure the effective mode index increments AN of the waveguides
at various wavelengths. Propagation loss measurements are then performed, followed by a
setup description and the results on measurements of the r33 electro-optic coefficient. The
71


45
For the case of LiTa03 in air, ns = 2.12 and n0 = 1. In order to have no reflections, R = 0,
we must use a film material whose index nf = (ns)1/2 ~ 1.46. We have chosen magnesium
fluoride (MgF2), with an index of 1.35, as a suitable material. For the wavelengths of
interest, namely 632.8nm and 1550nm, the resulting film thicknesses are 1172 and
o
2870A, respectively.
The AR films of MgF2 are easily deposited using an electron-beam (e-beam) depo
sition system. The details of such a system can be found in [Geo92, Mac86],
In the next section, the conditions used for the electroplating of thick gold elec
trodes are given. Such thick electrodes are used to minimize rf loss in high-speed travel
ing-wave modulators, such as the one to be demonstrated in Chapter 7.
2.7 Electroplating
For the demonstration of a high-speed traveling-wave electro-optic modulator to
be done later in Chapter 7, thick gold electroplated electrodes were needed. In this sec
tion, the conditions used for the fabrication and electroplating of these electrodes are
given.
To begin with, a buffer layer of Si02 2000-3000A thick was e-beam deposited at a
rate of 2A/s. For a Z-cut crystal, the buffer layer should be deposited uniformly over the
entire waveguide surface. For an X-cut crystal, the buffer layer need only be at each end,
outside the electrode interaction length where the electrodes bend and cross the
waveguides. Next, 150 of Cr was deposited, followed by 800 of Au. The Au is the
seed layer for electroplating and the Cr is to help the gold stick to the sample. Note that


85
4.4.1 Interferometric Measurement Method
Most efforts to directly measure r33 in channel waveguides involve the use of an
interferometric device, such as a Mach-Zehnder interferometer (MZI), and the application
of an electrode pattern to induce a phase shift. For example, the r33 coefficient for an MZI
in push-pull configuration is given by:
r33 (4-6)
27trVLn
where A, is the wavelength, g is the electrode gap, <]) is the induced phase shift, V is the opti
cal-electrical overlap factor, V is the applied voltage, L is the electrode interaction length,
and n is the refractive index. The accuracy of this technique depends greatly on the accu
racy in determining the value of T, which can only be estimated based on electrode dimen
sions. However, there is another method which ensures an overlap factor of unity. This
method is described next.
4.4.2 Unity Overlap Method
The setup used here to measure r33 is depicted in Fig. 4.8. The polarization-main
taining fiber from a diode laser is connected to a polarization-maintaining fiber coupler. It
was a 3dB coupler for the wavelength used, 1.55|im. One output from the coupler is used
20x
Fig. 4.8 Experimental setup used to measure the electro-optic coefficient r33.


26
2.1.4 APE Waveguide Fabrication Parameters
Using the APE process, channel waveguides were formed on the -X surface of X-
cut LiTa03 and the c' surface of Z-cut LiTa03 with the aim of single mode operation at a
wavelength of 1.5p.m. In this context, c~ and c+ will be used to refer to the negative and
positive surfaces of a Z-cut crystal, respectively. We defer the characterization of these
waveguides until chapter four, but present here the exact parameters used in their fabrica
tion. Additionally, the fabrication parameters of planar waveguides in X-cut LiTa03 are
presented. These waveguides will be used for reflectivity profiling and rocking curve
analysis, to be described later in chapter four.
The c+ and c~ surfaces of raw Z-cut LiTa03 crystal wafers can be identified by tak
ing advantage of the piezoelectric effect of the ferroelectric substrates. Fig. 2.7 demon
strates how a voltmeter was used to quickly press down and release on the surface of the
substrate. Because of the internal field set up by the spontaneous polarization of the crys
tal, free charges exist on the surfaces to counteract this field and maintain electrical neu
trality. For example, the c+ and c" surfaces have negative and positive free charges,
respectively. By rapidly pressing on the crystal and releasing, the internal field due to
polarization is momentarily reduced and the voltmeter can read the potential due to the
now unbalanced surface charges, so the individual positive and negative surfaces can be
identified. These Z-cut wafers were next diced into rectangular samples of various dimen
sions, depending on their application, where y-propagating channel waveguides could be
formed. These samples were then annealed in air at 200C for 1 hour to relieve stresses
incurred during dicing.


8 CONCLUSIONS AND FUTURE WORK
158
8.1 Conclusions 158
8.2 Future Work 161
8.2.1 Modulator Improvements 162
8.2.2 Integrated Laser-Modulator Module 164
8.2.3 Photorefractive Grating 165
APPENDICES
A DISPERSION OF RUTILE 166
B DISPERSION OF LITHIUM TANTALATE 168
REFERENCES 170
BIOGRAPHICAL SKETCH 179
viii


103
(a) (b)
Fig. 4.19 Measured rocking curves for X-cut virgin LiTa03 (a) and PE
LiTa03 (b).
diffractometer been used, only one wavelength would have been present, and only one
peak, the left-most one, would exist on the rocking curve. The measured curve is plotted
as a function of twice the incident angle 0. The angle for the substrate peak of interest is
labeled. Note that these samples were X-cut. For X-cut, the angle for the bulk substrate
0b is around 36.2, while it is around for 41.7 Z-cut.
Fig. 4.19(b) is the rocking curve for a PE sample prepared in glycerin at 260C for
24 hours. For this curve, the substrate peak and the peak corresponding to the PE layer for
the wavelength of interest are labeled. Having the measured values for 0b and A0=0b-0PE,
the value of e"3 can then be calculated from Eq. (4-13) and was found to be 4x1 O'3.
4.8 Summary
This chapter presented the characterization of planar and channel waveguides in
LiTa03 fabricated by APE, Ti-indiffusion, and Zn vapor-indiffusion. The number of


lOfim, in steps of 0.5|im. Samples were aligned under the photo-mask with the mask
aligner so that the propagation direction of the waveguide pattern to be used ran parallel
with the y-direction of the sample. Upon a 6 second exposure to ultraviolet (UV) light,
the samples were held in Hoechst Celanese AZ 312 developer (diluted 1:1.2) for 45 sec
onds, whereby the areas of photoresist not covered by the mask guide patterns, and hence
exposed to the UV, were removed. A Ta mask was sputtered on at a thickness of about
1000 using sputtering parameters of 280V for about 2 minutes and an Ar flow rate of 10
seem. Acetone was then used to lift off the remaining photoresist, leaving the sample
ready for proton exchange, as depicted in Fig. 2.3. The entire photolithographic process
from the spinning of a photoresist film through lift-off is illustrated in Fig. 2.8.
Proton exchange was performed using two different sources, pyrophosphoric acid
and glycerin. The exchange was performed in a box furnace at 260C for various times.
A custom made acid wash beaker was used to allow the source and sample to heat up sep
arately within the furnace. After stabilizing at the desired exchange temperature, the sam
ple was immersed into the source without having to open the furnace door and create
temperature variations. After exchanging, the sample was removed from the source and
allowed to cool to room temperature slowly, in order to reduce thermal strain. Each sam
ple was exchanged independently and new source was used each time.
Once cooled, the samples were washed in deionized (DI) water to remove the
remaining source and the Ta mask was etched off using a Ta etchant consisting of hydrof
luoric and nitric acids. The PE waveguides exchanged using pyrophosphoric acid were
then annealed in air using a tube furnace at various times and temperatures.


ACKNOWLEDGMENTS
I would like to express my deepest thanks to Dr. Ramu V. Ramaswamy for his
guidance and leadership as my research advisor. His endless hard work to provide funding
for an extremely well-equipped lab gives each of his students a unique opportunity to gain
valuable experience, obtainable at very few other institutions. Additionally, his high stan
dards for research, along with providing a very professional work environment, develops
in each of his students the crucial skills necessary to communicate and compete in the
business world.
I must also express much thanks to Dr. Robert Tavlykaev for his detailed guidance
and assistance on much of my research. His drive for knowledge and excellence brings
out the best in those around him. He is truly an asset to the Photonics Research Lab.
I would additionally like to thank Dr. Martin Uman, Dr. Gijs Bosman, Dr. Arnost
Neugroschel, and Dr. David Tanner for their participation on my supervisory committee.
I recognize and appreciate many of my fellow researchers and students, particu
larly Dr. Sergey Kostritskii, Dr. Yuri Korkishko, Dr. Chris Hussell, Dr. Suning Xie, Dr.
Weidong Wang, Dr. Sanjai Sinha, Dr. Scott Samson, Dr. Hyoun-Soo Kim, Dr. Sang-Kook
Han, Dr. Hao Feng, Mark Skowronski, A1 Ogden, Tetsuya Kishino, and Ryo Chinen.
Many of their contributions through individual discussions and group meetings have been
very valuable.


I am forever indebted to my parents, Blayne and Nadine Maring, as well as my sib
lings, Sheri and Steve. They instilled in me from a very young age the desire to achieve
excellence in every aspect of life and have supported me, both emotionally and financially,
through all of my education.
Finally, I wish to acknowledge some of the many new friends I have made along
the way: Juan, Carl, John, Rob, Chip, Jaime, Beth, Heather, Jen, Chad, Jodi, Mike, Kats,
and all the others. Thanks for the insane times. I needed them to remain sane myself.
IV


116
Raman shift, wavenumber (cm-1)
Fig. 5.6 Raman spectra of APE single-phase sample and PE sam
ples prepared in both pure glycerin and diluted glycerin at
260C for the times indicated.
with exchange time meaning that longer exchanges, required to produce waveguides deep
enough to guide, would be possible without increasing surface concentration. For exam
ple, diluting with 0.005 mol% Li2C03 required an exchange time of 80 hours to produce
single-mode waveguides at 1.55p.m. As can be seen in Fig. 5.6, this diluted sample is
comparable, or slightly improved, over the 24 hour pure glycerin sample and the single
phase APE sample.
5.4.2 Modified Structural Phase Diagram
From the above Raman analysis, samples fabricated in pure glycerin for 24 or less
and samples fabricated in diluted glycerin are predominately single phase. To determine


in LiNb03. In addition, LiTa03 has a shorter wavelength UV absorption edge (280nm)
than LiNb03 (350nm), permitting nonlinear conversion by frequency doubling to shorter
wavelengths with less absorption.
Though LiTa03 has a large nonlinear coefficient d33 and an electro-optic coeffi
cient r33 as large as that in LiNb03 [Miz92, Yuh92], it exhibits a low birefringence, mak
ing it unattractive for birefringent phase matching [Ahl91, Mat92b]. However, nonlinear
applications, such as second harmonic generation (SHG) of blue light, can be achieved in
LiTa03 with the aid of quasi-phase matching (QPM). Hence, LiTa03s higher optical
damage threshold and low propagation loss make it an ideal material for QPM-SHG of
blue light, capable of output power levels far exceeding those of LiNb03 [hl91, E1H95,
Yam91b, Yam92],
The more mature growth technology of LiTa03 allows the use of larger size crys
tals at a more reasonable cost than is possible with KTP. In addition, KTPs much higher
conductivity may result in large leakage currents, complicating the application of DC bias
voltages which are essential in many electro-optic devices.
It is clear, therefore, that there is an established advantage to using LiTa03 for the
fabrication of integrated-optical devices. However, significant research still remains to be
done, especially on the characterization and modeling of waveguides formed in LiTa03.
In the next section, some of the techniques used for the fabrication of waveguides in
LiTa03 will be presented and a few of the currently unresolved issues relating to this sub
ject will be discussed, leading to the motivation and focus for this work.


additional lattice strain, as does APE, there is no structural phase diagram to be con
structed. However, the stability of waveguides formed by each process needs also to be
quantified and monitored over time, so that it may be compared to that of APE
waveguides.
1.5 Chapter Organization
In Chapter 2, the various methods of waveguide fabrication used in this research
are discussed. A description of the APE process is provided followed by some of the fea
tures and anomalies of this process in LiTa03. Then the actual conditions used for the fab
rication of APE waveguides are given. Next, a description of the waveguide formation
processes of Ti-indiffusion and Zn vapor-indiffusion is presented. Again, some of the fea
tures associated with each process are provided along with the actual fabrication condi
tions used. The rest of the chapter details some additional processing steps involved in
waveguide and device fabrication such as crystal repoling, end-face polishing, the applica
tion of anti-reflectivity coatings, and electroplating of thick gold electrodes.
Chapter 3 examines the process of rare-earth Er-doping in LiTa03, a first step in
the fabrication of a 1.55(im active waveguide laser. Some background information related
to the process of Er-doping in LiNb03 is presented. Then the limitations associated with
the use of LiNb03 are presented, opening the door for the use of LiTa03 as a host material
for Er-doping. Next, the actual conditions used for the doping of LiTa03 with Er are out
lined, after which, characterization is performed by means of Raman spectroscopy and flu
orescence measurements to examine the effect of Er-doping on the photorefractive


38
First, a Ta film of thickness around 1200 is deposited on the waveguide surface.
This serves to identify the substrate surface during scans of the depth profile by introduc
ing a rapid spike in reflected light intensity. The sample is then mounted to a Si wafer,
waveguide surface against the Si, in the same manner as outlined in the previous subsec
tion. The Si here helps to prevent rounding due to polishing at the substrate surface. Now,
however, the structure is mounted onto a polishing jig which has a ramp-like grove milled
in it at an angle of a=2, as illustrated in Fig. 2.11.
Polishing begins by hand in SiC powder until the surface of the LiTa03 is reached.
Aluminum oxide powder is used next, followed by 1pm and 0.25pm diamond pastes,
same as before. As a final step, Ultra-Sol 201 A/280 Colloidal Alumina Polishing
polished surface
Fig. 2.11 Depiction of the polishing jig used to bevel polish planar waveguides at
an angle a = 2. This magnifies the depth profiles by a factor of 1/sina.


37
and flaking of the Si during polishing with a polishing jig. The structure is then mounted
to a polishing jig using low temperature bees wax. Because of the lower melting temper
ature of the bees wax (~70C), there is no danger of melting the thin layer of clear wax
between the Si and the LiTa03 sample.
The endface of the sample is first lapped by hand in SiC powder. Next, the sample
is polished by hand in 5|im aluminum oxide (A1203) powder to remove the deeper
scratches caused by the SiC powder. The polishing jig is then allowed to rotate freely on a
nylon polishing pad with lpxn diamond paste for one hour, followed by 0.25pm diamond
paste for thirty minutes. The LiTa03-Si structure is then turned around on the polishing
jig and the opposite endface is polished in a similar manner. The results of this polishing
technique for coupling into channel waveguides have been found to work very well.
Along with end-face polishing, a good bevel polishing of the waveguide surface
was also required for reflectivity profiling measurements. The bevel polish method used is
outlined in the next section.
2.5 The Bevel Polish for Planar Waveguides
For the reflectivity profiling measurement to be presented in Chapter 4, the resolu
tion of the depth profile is enhanced by bevelling a planar sample at an angle of a=2.
Using this technique, the profiles in the depth direction are magnified by 1/sina. In addi
tion, an extremely fine polish is required so as not to introduce a considerable amount of
noise in the reflected signal from the beveled surface. The details of the polishing scheme
used to achieve a fine bevel polish are given here.


91
is very common and desirable in practice since it results in devices with reduced crosstalk,
low driving voltages, and broad bandwidths.
To avoid the limitations of the previously used reconstruction techniques, such as
IWKB, it was decided to use a technique of direct index profiling. The bevelled technique
[Mar96a, Ste90], which is a direct technique, does not require waveguiding whatsoever and,
hence, can be used to measure both positive and negative index changes. The latter occur in
APE LiTa03 for the ordinary polarization. The bevelled technique is perfectly suited for
profiling single-mode waveguides after its accuracy has been brought to a level adequate to
resolve small index changes characteristic of single-mode structures. The technique is appli
cable to X-cut LiTa03 to measure changes in both the extraordinary and ordinary index.
Advantageously, electro-optic devices fabricated in this orientation do not require buffer
layers and exhibit much better temporal stability of optical output and improved frequency
response.
4.6.2 The Reflectivity Setup
Direct index profile measurements of APE waveguides in x-cut LiTa03 were made
using a reflectivity technique[Mar96, Ste90], The measurement setup is illustrated in Fig.
4.11. Waveguide profiles in the exchanged region are scanned in the depth direction under
the stationary spot of a focused laser beam. The index change is then measured by moni
toring the change in intensity of light reflected from the sample surface. The relationship
between the variation of reflected light intensity and index change is derived as follows.
The intensity of light reflected from the surface of the substrate /,. is related to the incident
intensity Ia by


Though critical advancements are needed in the growth technology for LiTa03, a
demand for high-quality optical grade crystals must exist before time and money will be
spent on such an effort. To this end, the potential of LiTa03 for integrated-optical compo
nents is demonstrated in Chapter 7 with the design, fabrication, and testing of a symmetric
Mach-Zehnder interferometer (MZI) traveling-wave electro-optic modulator. First, the
fabrication conditions for the device are provided. Next, some microwave characteristics
of the device are examined, including measurements of the microwave loss coefficient,
microwave effective index, and characteristic impedance. Then, the DC response of the
device is measured in order to obtain the extinction ratio and half-wave voltage. The
power handling capability of the modulator is examined next by plotting output power ver
sus input power near 1.5|im. Additionally, the frequency response of the device is mea
sured and compared to theory.
Finally, Chapter 8 presents the conclusions for this dissertation and outlines some
future work. A summary of the research performed is stated and the important contribu
tions resulting from this work are emphasized. Some future work to be performed in this
area, some of it currently under way, is then outlined.


CHAPTER 6
STABILITY OF LiTa03 WAVEGUIDES
The temporal instability of waveguides in LiTa03 is the major hindrance in the
way of their widespread use in practical devices. At least for APE waveguides, the instabil
ities are suspected to be caused by the presence of a mixture of crystal phases that can
transform in time one into another. As a result, the waveguide index, averaged over the
waveguiding area, changes, leading in turn to a variation in the propagation constant of the
guided modes. There is a valid suggestion that a waveguide containing only the low-con
centration a phase will exhibit improved temporal stability [E1H95]. Increased stability of
structures in predominantly one phase (low-concentration or high-concentration) [E1H95]
supports this idea.
In this chapter, the stability of waveguides in different regions of the structural
phase diagram for APE:LiTa03 are examined, along with the stability of waveguides fab
ricated by metal-indiffusion, with the aim of determining the feasibility of defining fabri
cation conditions yielding temporally stable waveguides. For APE waveguides, a
qualitative analysis was performed using rocking curve measurements to monitor any
change in strain values e"3 over time. For a more quantitative analysis, a directional cou
pler was used to track changes in the coupling length of couplers fabricated using
waveguides of varying fabrication conditions. These changes in coupling length were then
related to corresponding changes in index increment through a computer simulation of a
120


107
5.1 Previous Limitations
In this section, the limitations inhibiting previous efforts for complete reconstruc
tion of the structural phase diagram for APE:LiTa03 are described. The need for direct
measurement techniques to overcome these limitations is developed, involving methods
outlined in Chapter 4.
While the structural phase diagram for APE LiNb03 is fairly well known [Kor96,
Kor97, Zav93], the aforementioned specifics of the APE process in LiTa03 have compli
cated a complete and accurate reconstruction of the structural phase diagram for this mate
rial. The main obstacle appears to be the determination of refractive index, especially in
high-concentration, high-strain phases. Indeed, the appearance of a negative slope on the
known portion of the structural phase diagram (An vs. e"3) for Z-cut LiTa03 [E1H95,
Fed94] and rotated cuts of LiTa03 [Fed94] clearly implies the existence of buried index
profiles. Therefore, the accuracy of the IWKB method used for reconstruction of index
profiles is likely to have been compromised. Additionally, the decreasing index in high-
concentration phases results in waveguides supporting very few or no modes. Conse
quently, the IWKB method cannot be applied at all. This circumstance has limited the
range of the known structural phase diagram for X-cut LiTa03 [Fed94],
These limitations can be avoided through the use of direct measurement tech
niques. Specifically, the techniques of direct index profiling and rocking curve analysis,
described in the previous chapter, can be applied to extract surface values of index incre
ment and lattice strain for any waveguide. Through this approach, the structural phase dia
gram can be constructed. This has been done in the next section.


property of the crystal. Lastly, a two-photon model of the photorefractive effect is used to
explain how upconverted light caused by Er-doping enhances photorefractivity.
The issue of waveguides in LiTa03 is returned to in Chapter 4, where waveguide
characterization and measurement techniques are discussed and applied. First, a near-field
measurement technique is used to examine the region of single-mode operation, in terms
of waveguide channel width, for waveguides designed to be single-mode at 1,55|im. Next,
a modified prism coupling technique, which exhibits improved accuracy over the conven
tional technique, is introduced and applied to measure the mode effective indices for
waveguides at several different wavelengths. Waveguide loss is examined next, where
measurements of total insertion loss and waveguide propagation loss are made, the latter
by a Fabry-Perot technique. Then, the electro-optic coefficient r33 is measured directly
using a unity-overlap technique which does not require the estimation of an electrical-opti
cal field overlap factor. To demonstrate the superior photorefractive resistance of LiTa03
waveguides near 1.5|im, plots of output power versus input power are measured next and
compared to similar structures in LiNb03. Finally, the methods of direct index profiling
and X-ray diffraction rocking curve analysis are explained and applied for the direct deter
mination of surface index increment An and proton-induced lattice strain e"3 at the sur
face of APE waveguides. These figures are necessary for the construction of the structural
phase diagram for APE waveguides performed in Chapter 5.
In order to explain the anomalies associated with the APE process in LiTa03, the
structural phase diagram for this material needs to be constructed. This is done in Chapter
5 where several waveguides with different values of surface concentration are fabricated
by APE and measured using the techniques of direct index profiling and rocking curve


Planar waveguides were also fabricated in X-cut LiTa03 using the above exchange
condition. Being planar, however, they obviously did not need to go through the photo
lithographic process and their annealing times varied from 0 (as-exchanged) to 120 min
utes. The reason for these varied annealing times will become more obvious in Chapter 4.
It is noted here that exchange times longer than 20 minutes resulted in surface cracks in X-
cut crystals. This was most likely due to the difference in expansion of the lattice parame
ters along the X and Z-directions, as noted elsewhere [Mat92b],
In the next section, the other methods of waveguide fabrication will be described,
namely Ti metal-indiffusion and Zn vapor-indiffusion. The characteristics of waveguides
fabricated by these processes will be described and the conditions for their fabrication will
be given.
2.2 Waveguide Formation by Metal/Vapor-Indiffusion
As has been mentioned, APE waveguides in LiTa03 suffer from temporal instabil
ities. In addition, the introduction of H+ into the crystal via proton exchange is known to
cause quenching within a rare-earth-doped region [Nou92] in LiNb03, reducing the
excited state lifetime and therefore decreasing laser performance. As doping with rare-
earth Er for laser fabrication is a potential application for LiTa03 also, a means of
waveguide fabrication other than APE needs be explored. Since metal-indiffusion in
LiNb03 is known to be temporally stable, an examination of this process is a good place
to start for LiTa03. As such, waveguide fabrication in LiTa03 by Ti metal-indiffusion and
Zn vapor-indiffusion were examined. The known characteristics and fabrication parame
ters for each of these processes are presented here.


As an additional application, the PR enhancement brought about by Er-doping in
LiTa03 affords the possibility of producing PR gratings for use in WDM filters or
waveguide lasers, to name a few devices. Such gratings have been demonstrated in Fe-
doped LiNb03, though they are unstable and degrade over time [Huk98],
3.7 Summary
In this chapter, the effects of Er-doping in LiTa03 were examined. It was found
that Er-doping led to upconversion emissions. Energy transfer upconversion between Er3+
ions was determined to be the dominant mechanism for this upconversion and that the
presence of this upconverted light led to PR damage. A Li-treatment technique to reduce
clustering was described and successfully applied to minimize upconversion. A two-pho
ton model was then applied to model the PR effect and explain how upconverted light,
along with the presence of the longer wavelength incident light, caused PR damage
through self-gating. Finally, in addition to its known potential for applications like
high-speed modulation and nonlinear frequency doubling, LiTa03 was also demonstrated
to have strong advantages and potential for the applications of optical data storage, lasers,
and devices incorporating PR gratings. However, to realize these devices, stable
waveguides in LiTa03 must first be developed. The subject of waveguides is returned to in
the next chapter, where waveguides are characterized using a variety of techniques.


82
8168A tunable laser source is used to couple light into the waveguide. A 20x lens is used
at the waveguide output to collect the light and focus it onto an infrared Ge detector. A
polarizer is used in order to insure the detection of only the polarization of interest, from
both the sample and the fiber output, and a chopper is used, in conjunction with a lock-in
amplifier, to read the detector voltage with as low signal-to-noise ratio as possible.
4.3.2 Total Insertion Loss
To measure total insertion loss, the fiber output is first measured, with no sample in
the setup. Then the sample is introduced into the setup and the waveguide output is mea
sured in a similar manner. The difference between the two measurements is the total inser
tion loss of the waveguide sample.
For most of the waveguide samples used in this research, the total insertion loss
was around 2.5-3.0dB. The exact value of insertion loss depended on waveguide width,
sample length, fabrication technique and conditions, quality of the endface polish, and
quality of the cleaved face of the fiber.
As a best result, the fiber-to-fiber insertion loss of a 5pm wide waveguide fabri
cated by PE in glycerin at 260C for 24 hours was measured. To do this, the lens at the
output face of the waveguide was replaced with another single-mode fiber and the mea
surement was made in a similar manner as described above, using a lightwave multimeter
to measure optical throughput instead of the Ge detector and lock-in amplifier. The fiber-
to-fiber insertion loss for this sample was around 3dB.


multidomain and there is no net polarization and no electro-optic effect. To return the
crystal to its monodomain ferroelectric state, it must be poled. For LiTa03, this is done by
heating the crystal to just above its Curie point and applying an electric field of about
250V/cm [Lev66, Suz93] to re-orient the internal dipoles into a monodomain
configuration, as shown in Fig. 2.10. The sample is then cooled to below the Curie point
with the field applied, locking the dipoles in place.
Previous efforts to pole individual LiTa03 samples involved using platinum paint
on the c+ and c" surfaces to act as electrodes [Jun90]. Platinum must be used as this is the
only metal which does not oxidize at the temperatures needed for poling. However, for
the application of waveguide devices, success was only achieved when poling X-cut sam
ples as Z-cut samples were found to suffer from surface damage related to the application
of electrodes directly onto the sample surface [Jun90].
For this work, the poling process was modified to be useful for poling Z-cut
LiTa03 samples. First, the sample to be poled had 3000-4000 of Si02 deposited on the
c+ and c" surfaces to protect them from damage done by direct contact of the electrodes.
Next, the sample was sandwiched between two sheets of platinum foil with copper wires
]
+
electrodes'

0 0 0 0 @
§ § § § Q I
Fig. 2.10 Illustration of the crystal repoling process.


in applications where a single polarization is desirable, since only light polarized in the
direction of ne will be propagated.
In APE:LiTa03, there are at least four mechanisms contributing to the total index
change An [Ahl95], which, thus, can be expressed as
An = AnP + AnM + AnV + Ane (2-1)
The AnF term is caused by a change in the spontaneous polarization Ps via the Kerr elec
tro-optic effect. After substitution of H+ for Li+ during proton exchange, H+ is known to
occupy positions within the oxygen planes rather than the Li+ sites, reducing the contribu
tion of Li+ to Ps. The AnM term is due to the difference in the molar polarization of Li20
and H20 brought about by the substitution of one for the other in proton exchange. The
Anv term is caused by the change in molar volume due to an expansion of the unit cell
upon exchanging. Finally, Ane, due to exchange-induced stress, contributes via the elasto-
optic effect. Previous studies [Fed94] have observed that H:FiTa03 waveguides on Z and
X-cut substrates have only one non-zero component e33 of the deformation tenser. The
Anz term is proportional to the deformation value e33.
The depth de of the refractive index change after proton exchange is related to the
exchange time te by [hl94c, Dav95]
de = <2'2>
where De(Te) is the temperature dependent diffusion coefficient for the proton exchange
process and Te is the exchange temperature. This diffusion coefficient follows the Arrhe
nius law [hl94c, Dav95],
VkBTe)
A (2-3)


58
no PR effect
The mechanics of two-color excitation is
detailed later in Section 3.6.
To verify the assumption that the
up-converted green emission acts as a
self-gating source, enhancing the PR
K.min
-no PR effect
PR effect
effect via two-photon (or two-color) exci
tation without the need for an external gat
ing source, the PR effect was analyzed by
Raman spectroscopy measurements. The
PR effect is known to change the intensity
IR of Raman-shifted lines by an amount
which is proportional to the light-induced
extraordinary refractive-index change
An. Since An depends on input power
intensity J, the change of IR also depends on J, as illustrated in Fig. 3.4. The dependence
of An on J is calculated by the following relation, which was experimentally verified
[Kos94]:
(b)
Fig. 3.4 Raman intensity (a) and nor
malized Raman intensity (b)
as a function of input inten
sity, illustrating the effects of
photorefractive damage.
Ans(J) = Atp-ll = a^-iI (3-2)
I IRJmin J [ IR J
where Ans(J) is the intensity-dependent steady-state value of An; Jmin is the minimum
value of J when photorefractive damage becomes undetectable by Raman spectroscopy;
IR m¡n and IR are Raman intensities measured at the minimum and an arbitrary value of J,
respectively; IR = IRJmin/J is a normalized Raman intensity; and A is a coefficient


27
Voltmeter
4
substrate
Al foil
pos. meter change = c+ surface up
neg. meter change = c' surface up
Fig. 2.7 Method used to identify the individual surfaces of
a z-cut LiTa03 wafer.
Samples intended for channel waveguides were subjected to the standard photo
lithographic process. After cleaning in trichloroethane, acetone, and methanol, the sam
ples were dried at 120C for 5 minutes. Hoechst Celanese AZ 1370 Photoresist was spun
on the c' surfaces at 6000 rpm for 35 seconds, after which the samples were baked at 90C
for 35 minutes. This resulted in a photoresist thickness of about 1p.m. Each substrate was
then covered with a digitally dark photo-mask containing desired patterns, be they inter
ferometers, directional couplers, or just straight stripes ranging in width from 2fim to


94
crystal polarization for profiling. The substrate was bevel-polished at an angle a in order
to magnify the depth profiles by a factor of 1/sina. The exact angle of the bevel was a =
1.86, making the magnification factor 30.76. A Ta film deposited on the substrate surface
prior to bevelling served as a reference for marking the position of the substrate surface by
causing a sharp peak in reflected intensity. An AR coating (MgF2, 1172 thick) deposited
on the back of the sample helped to minimize interference and noise caused by backreflec-
tions. Light incident on the substrate surface was chopped and detected via a Si detector
and lock-in amplifier. The lock-in output was fed to a digital oscilloscope. The sample
was scanned, at a low frequency, under the stationary spot of the focused laser beam and
the reflectivity displayed in real-time on the oscilloscope.
The scanning of the sample was achieved by attaching the sample, by means of a
light-weight wooden dowel (diameter 1/8), to the center cone of a 3 audio speaker. The
sample was stuck to the rod using double-sided tape and the rod attached to the speaker
cone using hot glue. A DC-ramp signal drove the speaker, causing it to oscillate linearly
and thus scan the sample back-and-forth under the spot. This ramp signal was also used to
trigger the oscilloscope. The amount of speaker excursion, and thus scanning distance,
was controlled by the peak-to-peak voltage of the input ramp signal to the speaker. A cal
ibration chart of speaker input voltage versus speaker excursion distance was created by
identifying the voltage necessary to cause enough excursion to display on the oscilloscope
the intensity peaks of thin metal stripes, fabricated on a particular sample, with a known
separation between them. This calibration chart is depicted in Fig. 4.12. The entire mea
surement process is then performed as follows. The oscilloscope is used to display the
change in reflected intensity, appearing as a change in voltage AV. The scanning distance


99
depth (pm)
Fig. 4.16 Ordinary index profiles for the set of samples in Fig. 4.15.
Upon annealing, the magnitude of the ordinary index depression, in general, decreases
while the distribution width increases. It is noted that the peak value of the negative
ordinary change for the as-exchanged waveguides is comparable to those reported in
[Ahl94c],
To determine the dispersion of the index change, the measurements were repeated at
7u= 1.32pm and compared with those at A,=0.633pm. The obtained results clearly indicate
that the extraordinary increment exhibits strong dispersion over the spectral range from
0.633pm to 1.32pm. As an example, the extraordinary profile measured on a sample with
ta=30min is shown in Fig. 4.15 by the dashed-dotted line. For this sample, the ratio


174
[Kis95] K. M. Kissa, P. G. Suchoski, and D. K. Lewis, Accelerated Aging of Annealed
Proton-Exchanged Waveguides, J. Lightwave Technol., vol. 13, no. 7, p. 1521,
1995.
[Kor96] Yu.N.Korkishko and V.A.Fedorov, Structural Phase Diagram of HxLij_xNb03
Waveguides: The Correlation Between Structural and Optical Properties,
IEEE J. Sel. Top. Quantum Electron., vol. 2, p. 1, 1996.
[Kor97] Yu. N. Korkishko and V. A. Fedorov, Relationship Between Refractive Indices
and Hydrogen Concentration in Proton-Exchanged LiNb03 Waveguides, J.
Appl. Phys., vol. 82, no .2, p. 171, 1997.
[Kos94] S.M. Kostritskii, and O.M. Kolesnikov, Photoinduced Light Scattering in the
Copper-Doped Li j_x HxNb03 Photorefractive Waveguides, J. Opt. Soc. Am. B,
vol. 11, p. 1674, 1994.
[Kos97a] S.M. Kostritskii and O.G. Sevostyanov, Influence of Intrinsic Defects on
Light-Induced Changes in Refractive Index of Lithium Niobate Crystals,
Appl. Phys. B, vol. 65, p. 527, 1997.
[Kos97b] S.M. Kostritskii, D. Kip, and E. Kratzig, Improvement of Photorefractive
Properties of Proton-Exchanged LiTa03 Waveguides, Appl. Phys. B, vol. 65,
p. 517, 1997.
[Kos99] S. M. Kostritskii, D. B. Maring, R. F. Tavlykaev, and R. V. Ramaswamy,
"Enhancement of the Photorefractive Effect by Er Doping in LiTa03, TOPS
(Trends in Optics and Photonics Series), 1999, v. 27, pp. 72-79. Seventh Topi
cal Meeting on Photorefractive Materials, Effects, and Devices, June 27-30,
1999, Elsinore, Denmark.
[Kub80] K.Kubota, J. Noda, and O. Mikami, Traveling Wave Optical Modulator Using
a Directional Coupler LiNb03 Waveguide, IEEE J. Quantum Electron., vol.
QE-16, no. 7,p. 754, 1980.
[Lan98] D. Lande, S.S. Orlov, and L. Hesselink, Two-Photon Apodization in Lithium
Niobate, Opt. Lett., vol. 23, p. 1399, 1998.
[Lev66] H. J. Levinstein, A. A. Ballman, and C. D. Capio, Domain Structure and Curie
Temperature of Single-Crystal Lithium Tantalate, J. Appl. Phys., vol. 37, p.
4585, 1966.
[Mac86] H. A. Macleod, Thin-Film Optical Filters, Macmillan, New York, 1986.
[Mar96a] D. B. Maring, R. F. Tavlykaev, and R. V. Ramaswamy, Direct Observation of
Buried Refractive-Index Profiles in Annealed Proton-Exchanged LiTa03


171
[Bai97]
[Bau96]
[Bor95]
[Bui-79]
[Bur93]
[Cac97]
[Cao92]
[Che91]
[Chu91]
[Dav95]
[Del93]
Y.S. Bai, R.R. Neurgaonkar, and R. Kachru, High-Efficiency Nonvolatile
Holographic Storage with Two-Step Recording in Praseodymium-Doped Lith
ium Niobate by Use of Continuous-Wave Lasers, Opt. Lett., vol. 22, no. 5, p.
334, 1997.
1. Baumann, S. Bosso, R. Brinkmann, R. Corsini, M. Dinand, A. Greiner, K.
Schafer, J. Sochting, W. Sohler, H. Suche, and R. Wessel, Er-Doped Inte
grated Optical Devices In LiNb03, IEEE J. Sel. Top. Quantum Electron., vol.
2, p. 355, 1996.
P. F. Bordui, R. G. Norwood, C. D. Bird, and J. T. Carella, Stoichiometry
Issues in Single-Crystal Lithium Tantalate, J. Appl. Phys., vol. 78, no. 7, p.
4647, 1995.
W. K. Burns, P. H. Klein, E. J. West, and L. E. Plew, Ti Diffusion in
Ti:LiNb03 Planar and Channel Optical Waveguides, J. Appl. Phys., vol. 50,
no. 10, p. 6175, 1979.
W. K. Burns, M. McWright Howerton, and R. P. Moeller, Performance and
Modeling of Proton Exchanged LiTa03 Branching Modulators, J. Lightwave
Technol., vol. 10, no. 10, p. 1403, 1993.
F. Caccavale, F. Segato, and I. Mansour, A Numerical Study of Erbium-Doped
Active LiNb03 Waveguides by the Beam Propagation Method, J. Lightwave
Technol., vol. 15, no. 12, p. 2294, 1997.
X. F. Cao, R. V. Ramaswamy, and R. S. Srivastava, Characterization of
Annealed Proton Exchanged LiNb03 Waveguides for Nonlinear Frequency
Conversion, J. Lightwave Technol., vol. 10, no. 9, p. 1302, 1992.
H. C. Cheng and R. V. Ramaswamy, Novel Technique for the Measurement of
Coupling Length in Directional Couplers, Technical Digest SPIE OE/Aero
space Sensing, Orlando91, April 1-5, 1991.
W. C. Chuang, W. S. Wang, and J. H. Chu, A Comparison of the Performance
of LiNb03 and LiTa03 Traveling-Wave Phase Modulators, Fiber Int. Opt.,
vol. 10, p. 183, 1991.
G. Davis and N. A. Lindop, Fabrication and Characterization of Pyrophospho-
ric Acid Proton Exchanged Lithium Tantalate Waveguides, J. Appl. Phys.,,
vol. 77, no. 12, p. 6121, 1995.
E. Delevaque, T. Georges, M. Monerie, P. Lamouler, and J.-F. Bayon, Model
ing of Pair-Induced Quenching in Erbium-Doped Silicate Fibers, IEEE Pho
ton. Technol. Lett., vol. 5, no. 1, p. 73, 1993.


18
doubling of infrared, where quasi-phase matching (QPM) is required [hl94c, Nak90],
For this case, the APE waveguides and domain inverted regions are fabricated on the c'
surface of LiTa03 (as opposed to the c+ surface for LiNb03) since domain inversion
seems to be initiated on the c" surface [Nak90, Saw91]. When using patterns of domain
inversion, such as in QPM, the waveguides are oriented perpendicular to the existing
inverted regions, as in Fig. 2.2. Since proton exchange and post-exchange annealing for
waveguide fabrication can be done at temperatures well below 450C, they do not affect
the existing inverted regions which have also been created by APE, but annealed at tem
peratures near the Curie temperature of about 600C [Fin88, Miz94],
The exchange setup of the APE process is depicted in Fig. 2.3 for channel
waveguides. The substrate is first covered by a metal masking layer (usually A1 or Ta), in
Fig. 2.3 Schematic of the proton exchange process.


161
high-quality material dampens motivation for the improvement of its growth technology.
Finally, to demonstrate the potential of LiTa03 and hopefully initiate market interest for
higher quality material, a high-speed traveling-wave electro-optic modulator was fabri
cated and characterized, which compared well with theoretical prediction.
These contributions are significant to the further exploitation of LiTa03 for device
fabrication. They demonstrate that this material has great potential, and advantages, in the
fabrication of a number of high-throughput integrated-optical devices, such as lasers and
modulators, as well as for applications involving photorefractive materials, such as holo
graphic data storage and photorefractive gratings. This is important because it emphasizes
the need for high-quality, optical grade crystals, which in turn encourages crystal growers
to spend time and money on the advancement of growth technology and improved stoichi
ometry for this material beyond that of merely SAW grade.
Some future work related to this research is outlined in the next section. In partic
ular, the application of results obtained here toward the development of integrated-optical
devices is discussed.
8.2 Future Work
Though this research has gone a long way toward the development of integrated-
optical devices in LiTa03, there still remains the obstacle of crystal instability. This limi
tation is likely the result of an immature growth technology for this crystal, possibly
related to defects and strain caused by its high growth temperature and corresponding Li
deficiency. However, the existence of market demand for better crystals should result in
research toward the improvement of growth techniques. Though such work is outside the


134
Table 6.3, this sample was still more unstable than Sample 2. However, to try to fabricate
samples with significantly reduced e"3 values by using an even more dilute source would
have required hundreds of hours of exchange time, due to the decreased proton concentra
tion of the source and the larger depth required to support a mode at the lower An values
associated with decreased e"3 values. Such a fabrication situation is not practical for the
formation of waveguides.
By comparison with Samples 3 and 3', however, it was seen that Sample 4, though
having a smaller strain value, was no more stable than Sample 3 and more unstable then
Sample 3'. Evidently, annealing Sample 3 to produce Sample 3' had some impact on sta
bility. To this end, it was decided to try to artificially age [Kis95] these samples at a low
temperature (<100C). By this means, it might be possible to improve the stability of
these samples, without causing significant alterations to the H+ profile, and hence the
index profile, as annealing at higher temperatures causes.
Sample 5 (Table 6.3) had roughly the same exchange conditions as Sample 4. In
addition, however, it was aged at 85C for 4 days. Though its strain value was not directly
measured, because of its fabrication conditions it must have had an e3 value similar to
that of Sample 4, or slightly smaller. The measured change in coupling for this sample is
shown in Fig. 6.10.
After a period of just 4 days, it was evident that stability had not been improved.
So this sample was subjected to further aging at 95C for 4 more days. It was then denoted
as Sample 5" (Table 6.3). Within days after this aging, the waveguide couplers fell below
cutoff for the fundamental mode at 1.55|im. As a result, the change in coupling length was


104
modes supported by channels of various widths were first determined by a near-field mea
surement technique. Next, the effective mode index increments of modes supported by the
waveguides were measured at four wavelengths and the values for the fundamental mode
at 1.31pm were found to be in good agreement with those numerically calculated [Tav95].
A Fabry-Perot method of determining propagation loss was then discussed and applied to
channel waveguides, yielding a value of 0.24dB/cm for PE waveguides fabricated using a
glycerin source. A method for measuring the electro-optic coefficient which did not
involve estimating a value for the overlap factor of the electrical and optical fields was also
described. This technique gave a value of 27.1pm/V for a repoled Zn vapor-indiffused
sample, indicating that the repoling process was successful. The photorefractive charac
teristic of PE waveguides in LiTa03 and LiNb03 were measured near 1.5pm. The LiNb03
waveguides exhibited significant damage at power levels as low as 40mW, while the
LiTa03 waveguides demonstrated decent power handling to levels as high as 500mW.
Finally, the techniques of direct index profiling and rocking curve analysis were outlined
for their use in determining surface values of index increment An and lattice strain e"3,
used in the next chapter to reconstruct the structural phase diagram for APE:LiTa03.


118
e"3 The sample was annealed one last time at 340C for 6 hours. This time the measured
values of strain and index increment were e"3 =2.3xl0~3 and An=0.011.
The three new data points were added to the structural phase diagram, shown as
filled squares in Fig. 5.7. A straight line was fit through these points back to the origin.
This data, along with the Raman data confirming these samples to be of single-phase, con
firms the presence of a new single-phase region. This new phase, labeled the oT phase in
Fig. 5.7, is obtainable only by PE in a dilute source where the surface concentration never
leaves this phase during exchange.
5.5 Summary
In this chapter, the structural phase diagram for X-cut APE:LiTa03 was con
structed. First, some of the limitations hindering previous efforts to construct this diagram
were outlined, leading to the need for direct index profiling. Next, direct index profiling,
along with rocking curve measurements, were applied to APE samples of various fabrica
tion conditions in order to construct the structural phase diagram for APE:LiTa03, relating
surface index increment to surface values of proton-induced lattice strain. This diagram
was then used to explain several of the previously observed anomalies associated with the
APE process in LiTa03. Additionally, the effect of the presence of multiple crystal phases
on the waveguide index profiles was studied, with profiles from different regions of the
phase diagram being illustrated. Lastly, PE waveguides were analyzed because of their
advantage of higher mode confinement. Measured values of surface index increment and
strain showed these samples to be in a new phase of the diagram. Raman spectra of PE
samples, compared with that of single-phase APE samples, confirmed that waveguides


For this comparison, straight-channel waveguides of widths 2-10|im, in steps of
0.5jim, were fabricated in Z-cut crystals from each supplier using the conditions for Zn
142
vapor-indiffusion given in Chapter 2. The samples from Shin-Etsu were of SAW grade, all
others were claimed to be of optical grade. The method used in Section 6.4.2 of tracking
the change in the region of single-mode operation was used here, with the results pre
sented in Table 6.5. As can be seen, all the samples showed relatively the same degree of
instability, indicating that these materials were all similar after growth, and unstable.
Additionally, there was no distinction between crystals of SAW grade and those claimed to
be of optical grade. From this, it was clear that the growth process for LiTa03 has not yet
reached the maturity level of LiNb03.
6.5.2 Accounting for the Li Deficiency
Because of its high growth temperature, Li20 diffuses out of LiTa03 while it is
growing, leading to a crystal of noncongruent composition [Bor95, Dou89]. In order to
account for this Li deficiency, a noncongruent, Li-rich melt can be used during growth.
Flowever, this results in a crystal boule having a varying amount of Li concentration along
the length of the boule, in addition to requiring some calibration research to determine the
amount of Li to add to the melt to obtain a congruent crystal.
As a preliminary effort to examine if increasing the Li content in virgin LiTa03
would improve stability, a Yamaju sample was subjected to the Li-treatment process
described in Chapter 3 before Zn waveguides were fabricated in it. Here, the sample was
processed in LiN03 at 280C for 21.5 hours. The resulting stability analysis is shown
with the other samples in Table 6.5. As can be seen, this treatment did not significantly
improve the stability of the waveguides.


148
Fig. 7.3 Measured S2i (dotted) and fitted frequency dependent loss
function a = a0Vf (solid) wherea0=l.OdB/cm* VGHz.
7.2.1 S-Parameters and Loss
The microwave transmission characteristics, S-parameters, of the CPW electrode
were measured first using an HP8510C network analyzer and a Universal Test Fixture
from Wiltron. The measured and calibrated microwave reflection (Sjj) and transmission
(S2i) characteristics for the electrode are shown in Fig. 7.2. To construct this plot, the Sj]
and S2i parameters of the electrodes were measured. Then the same parameters were
measured on a 1cm Calibration Kit from Wiltron, designed to simulate the test fixture
holding the sample. The measurements done on the 1cm Calibration Kit were then used to
calibrate out the effect of the test fixture and produce the S] | and S21 curves for the elec
trode only as shown in Fig. 7.2. By fitting the S21 curve with a frequency-dependent loss
function of a = a0Vf, a microwave loss coefficient of a0=1.0dB/cm JGHz was
obtained as depicted in Fig. 7.3.


16
This chapter describes the methods used in the creation of optical waveguides in
dielectric substrates. Particular attention is paid to the proton exchange and annealed pro
ton exchange techniques, as well as to the methods of Ti metal-indiffusion and Zn vapor-
indiffusion, used to fabricate waveguides in LiTa03. Several distinct features and anoma
lies attributed to each of the processes are discussed. Specific fabrication parameters are
given in order to cover completely the fabrication methods used in this research, and a few
additional processing steps are also described.
2.1 Waveguide Formation by the APE Process
Waveguides in LiNb03 and LiTa03 can be formed by using several methods, the
most popular of which are by metal indiffusion at high temperatures, and by the exchange
of Li+ ions in the substrate by H+ ions supplied by an acid source proton exchange or
annealed proton exchange method [Ram88, Sym92], In this section, the annealed proton
exchange method of waveguide fabrication is discussed. The features of waveguides fab
ricated by this process in LiTa03 are then described, followed by some of the associated
anomalies observed. Finally, the actual fabrication conditions used to produce
waveguides by this technique are given.
2.1.1 The PE/APE Process
As discovered by Jackel, Rice, and Veselka in 1982, the proton exchange (PE) pro
cess results in a large increase of the extraordinary refractive index of LiNb03 [Jac82].
The technique was later shown to also be effective in forming waveguides in LiTa03 by
Spillman, Sanford and Soref [Spi83], Though the PE process initially causes a reduction



PAGE 1

3523(57,(6 $1' &+$5$&7(5,67,&6 2) /L7D )25 ,17(*5$7('237,&$/ '(9,&( $33/,&$7,216 %\ '$9,' %/$<1( 0$5,1* $ ',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

7R P\ IULHQGV IDPLO\ DQG ORYHG RQHV IRU WKHLU VXSSRUW

PAGE 3

$&.12:/('*0(176 ZRXOG OLNH WR H[SUHVV P\ GHHSHVW WKDQNV WR 'U 5DPX 9 5DPDVZDP\ IRU KLV JXLGDQFH DQG OHDGHUVKLS DV P\ UHVHDUFK DGYLVRU +LV HQGOHVV KDUG ZRUN WR SURYLGH IXQGLQJ IRU DQ H[WUHPHO\ ZHOOHTXLSSHG ODE JLYHV HDFK RI KLV VWXGHQWV D XQLTXH RSSRUWXQLW\ WR JDLQ YDOXDEOH H[SHULHQFH REWDLQDEOH DW YHU\ IHZ RWKHU LQVWLWXWLRQV $GGLWLRQDOO\ KLV KLJK VWDQn GDUGV IRU UHVHDUFK DORQJ ZLWK SURYLGLQJ D YHU\ SURIHVVLRQDO ZRUN HQYLURQPHQW GHYHORSV LQ HDFK RI KLV VWXGHQWV WKH FUXFLDO VNLOOV QHFHVVDU\ WR FRPPXQLFDWH DQG FRPSHWH LQ WKH EXVLQHVV ZRUOG PXVW DOVR H[SUHVV PXFK WKDQNV WR 'U 5REHUW 7DYO\NDHY IRU KLV GHWDLOHG JXLGDQFH DQG DVVLVWDQFH RQ PXFK RI P\ UHVHDUFK +LV GULYH IRU NQRZOHGJH DQG H[FHOOHQFH EULQJV RXW WKH EHVW LQ WKRVH DURXQG KLP +H LV WUXO\ DQ DVVHW WR WKH 3KRWRQLFV 5HVHDUFK /DE ZRXOG DGGLWLRQDOO\ OLNH WR WKDQN 'U 0DUWLQ 8PDQ 'U *LMV %RVPDQ 'U $UQRVW 1HXJURVFKHO DQG 'U 'DYLG 7DQQHU IRU WKHLU SDUWLFLSDWLRQ RQ P\ VXSHUYLVRU\ FRPPLWWHH UHFRJQL]H DQG DSSUHFLDWH PDQ\ RI P\ IHOORZ UHVHDUFKHUV DQG VWXGHQWV SDUWLFXn ODUO\ 'U 6HUJH\ .RVWULWVNLL 'U
PAGE 4

, DP IRUHYHU LQGHEWHG WR P\ SDUHQWV %OD\QH DQG 1DGLQH 0DULQJ DV ZHOO DV P\ VLEn OLQJV 6KHUL DQG 6WHYH 7KH\ LQVWLOOHG LQ PH IURP D YHU\ \RXQJ DJH WKH GHVLUH WR DFKLHYH H[FHOOHQFH LQ HYHU\ DVSHFW RI OLIH DQG KDYH VXSSRUWHG PH ERWK HPRWLRQDOO\ DQG ILQDQFLDOO\ WKURXJK DOO RI P\ HGXFDWLRQ )LQDOO\ ZLVK WR DFNQRZOHGJH VRPH RI WKH PDQ\ QHZ IULHQGV KDYH PDGH DORQJ WKH ZD\ -XDQ &DUO -RKQ 5RE &KLS -DLPH %HWK +HDWKHU -HQ &KDG -RGL 0LNH .DWV DQG DOO WKH RWKHUV 7KDQNV IRU WKH LQVDQH WLPHV QHHGHG WKHP WR UHPDLQ VDQH P\VHOI ,9

PAGE 5

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

PAGE 6

&+$5$&7(5,67,&6 2) (U'23(' /L7D %DFNJURXQG (U'RSLQJ &RQGLWLRQV 5DPDQ)OXRUHVFHQFH 0HDVXUHPHQWV 7KH 5DPDQ (IIHFW DQG 5DPDQ 6SHFWURVFRS\ )OXRUHVFHQFH 6SHFWURVFRS\ )OXRUHVFHQFH 0HDVXUHPHQWV 5DPDQ 0HDVXUHPHQWV (QHUJ\ 7UDQVIHU 8SFRQYHUVLRQ /L7UHDWPHQW 7KH 7ZR3KRWRQ 0RGHO RI WKH 35 (IIHFW /L7Df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

PAGE 7

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

PAGE 8

&21&/86,216 $1' )8785( :25. &RQFOXVLRQV )XWXUH :RUN 0RGXODWRU ,PSURYHPHQWV ,QWHJUDWHG /DVHU0RGXODWRU 0RGXOH 3KRWRUHIUDFWLYH *UDWLQJ $33(1',&(6 $ ',63(56,21 2) 587,/( % ',63(56,21 2) /,7+,80 7$17$/$7( 5()(5(1&(6 %,2*5$3+,&$/ 6.(7&+ YLLL

PAGE 9

$EVWUDFW RI 'LVVHUWDWLRQ 3UHVHQWHG WR WKH *UDGXDWH 6FKRRO RI WKH 8QLYHUVLW\ RI )ORULGD LQ 3DUWLDO )XOILOOPHQW RI WKH 5HTXLUHPHQWV IRU WKH 'HJUHH RI 'RFWRU RI 3KLORVRSK\ 3523(57,(6 $1' &+$5$&7(5,67,&6 2) /L7D )25 ,17(*5$7('237,&$/ '(9,&( $33/,&$7,216 %\ 'DYLG %OD\QH 0DULQJ 0D\ &KDLUPDQ 5DPX 9 5DPDVZDP\ 0DMRU 'HSDUWPHQW (OHFWULFDO DQG &RPSXWHU (QJLQHHULQJ &RQWLQXHG DQG UDSLG JURZWK H[SRXQGHG E\ WKH EDQGZLGWK H[SORVLRQ LQ WKH DUHD RI GHQVHZDYHOHQJWKGLYLVLRQPXOWLSOH[HG ':'0f WHOHFRPPXQLFDWLRQ DQG LQIRUPDWLRQ QHWZRUNV OLNH WKH LQWHUQHW KDV EHHQ DQG FRQWLQXHV WR EH PHW E\ WKH JOREDO GHSOR\PHQW RI RSWLFDO ILEHU ,Q WXUQ WKLV KDV UHVXOWHG LQ WKH XQSUHFHGHQWHG GHPDQG IRU WKH LQWHJUDWHG RSWLFDO FRPSRQHQWV XVHG LQ WKHVH ILEHU QHWZRUNV 2I SDUWLFXODU LQWHUHVW LV GHYLFHV IDEULn FDWHG LQ IHUURHOHFWULFV PRVW QRWDEO\ OLWKLXP QLREDWH /L1Ef DQG OLWKLXP WDQWDODWH /L7Df $OWKRXJK FXUUHQW GHPDQGV DUH EHLQJ PHW E\ /L1E LWV EUHWKUHQ /L7D H[KLELWV D PXFK KLJKHU WKUHVKROG IRU RSWLFDO GDPDJH SURYLGLQJ WKH RSSRUWXQLW\ WR IDEULn FDWH GHYLFHV VXFK DV ODVHUV DQG PRGXODWRUV ZLWK PXFK KLJKHU RXWSXW DQG WKURXJKSXW OHYn HOV WKDQ FXUUHQWO\ REWDLQDEOH ZLWK /L1E ,Q DGGLWLRQ /L7D KDV D VKRUWHU ZDYHOHQJWK XOWUDYLROHW 89f DEVRUSWLRQ HGJH QPf WKDQ /L1E QPf SHUPLWWLQJ QRQOLQHDU FRQYHUVLRQ E\ IUHTXHQF\ GRXEOLQJ WR VKRUWHU ZDYHOHQJWKV ZLWK OHVV DEVRUSWLRQ ,;

PAGE 10

$ QXPEHU RI REVWDFOHV KRZHYHU SUHYHQW WKH ODUJHVFDOH GHYHORSPHQW RI RSWLFDO GHYLFHV LQ /L7D 7KH YDULRXV SURFHVVHV RI ZDYHJXLGH IRUPDWLRQ KDYH QRW EHHQ FRPn SOHWHO\ LQYHVWLJDWHG RU FKDUDFWHUL]HG )RU LQVWDQFH WKH DQQHDOHG SURWRQ H[FKDQJH $3(f WHFKQLTXH LV NQRZQ WR H[KLELW D QXPEHU RI DQRPDOLHV ZKLFK WR GDWH KDYH QRW EHHQ XQGHUn VWRRG WKH PRVW VLJQLILFDQW EHLQJ WKH WHPSRUDO LQVWDELOLW\ RI WKH ZDYHJXLGH LQGH[ LQFUHn PHQW $GGLWLRQDOO\ QR ZRUN KDV EHHQ SHUIRUPHG WR FKDUDFWHUL]H WKH LPSDFW RI (U LQGLIIXVLRQ QHFHVVDU\ IRU WKH IDEULFDWLRQ RI SP ODVHUV ,Q WKLV GLVVHUWDWLRQ WKH DERYH LVVXHV DUH DGGUHVVHG &RQGLWLRQV IRU WKH LQGLIIXVLRQ RI (U LQ /L7D DUH LGHQWLILHG )OXRUHVFHQFH DQG 5DPDQ VSHFWURVFRS\ DUH XVHG WR LGHQWLI\ HQHUJ\ WUDQVIHU XSFRQYHUVLRQ (78f EHWZHHQ QHLJKERULQJ FOXVWHUV RI (U LRQV DV WKH GRPLQDQW PHFKDQLVP RI XSFRQYHUVLRQ OHDGLQJ WR LQFUHDVHG SKRWRUHIUDFWLYLW\ D JDLQOLPLWn LQJ IDFWRU IRU ODVHUV 8SFRQYHUVLRQ LV WKHQ GHFUHDVHG WKURXJK D QRYHO /LWUHDWPHQW SURFHVV WR UHGXFH FOXVWHULQJ ZLWK UHVXOWV VKRZLQJ D VLJQLILFDQW DGYDQWDJH RYHU /L1E $GGLWLRQn DOO\ WKH VWUXFWXUDO SKDVH GLDJUDP IRU $3(/L7D ZDYHJXLGHV LV FRQVWUXFWHG DQG XVHG WR H[SODLQ WKH SUHYLRXVO\ REVHUYHG DQRPDOLHV DVVRFLDWHG ZLWK WKLV SURFHVV LQ /L7D 7KH VWDELOLW\ RI ZDYHJXLGHV IDEULFDWHG E\ YDULRXV WHFKQLTXHV LV DOVR H[DPLQHG UHVXOWLQJ LQ WKH GHWHUPLQDWLRQ WKDW LQVWDELOLW\ GRHV QRW GHSHQG RQ IDEULFDWLRQ FRQGLWLRQV DV LW GRHV LQ /L1E EXW UDWKHU LW LV LQKHUHQW WR WKH FU\VWDO WKH OLNHO\ UHVXOW RI DQ LPPDWXUH JURZWK SURn FHVV )LQDOO\ WR GHPRQVWUDWH WKH SRWHQWLDO RI /L7D DQG LQGLFDWH WKH GHVLUH IRU KLJKHU TXDOLW\ FU\VWDOV D KLJKVSHHG WUDYHOLQJZDYH PRGXODWRU LV IDEULFDWHG DQG LWV PHDVXUHG UHVXOWV FRPSDUHG WR WKHRU\ 7KLV ZRUN KRSHIXOO\ ZLOO LQVSLUH FU\VWDO JURZWK FRPSDQLHV WR DUULYH DW D JURZWK SURFHVV ZKLFK SURGXFHV FU\VWDOV ZLWK EHWWHU FU\VWDO VWRLFKLRPHWU\ GHVLUn DEOH IRU LPSURYHG GHYLFH SHUIRUPDQFH

PAGE 11

&+$37(5 ,1752'8&7,21 $ GLVFXVVLRQ OHDGLQJ WR WKH PRWLYDWLRQ IRU WKLV GLVVHUWDWLRQ LV SUHVHQWHG LQ WKLV FKDSWHU (YHU LQFUHDVLQJ GHPDQG IRU EDQGZLGWK LQ WHOHSKRQH FRPSXWHU QHWZRUN DQG LQWHUQHW DSSOLFDWLRQV KDV UHVXOWHG LQ WKH HYHU LQFUHDVLQJ GHPDQG ZRUOGZLGH IRU ILEHU QHWn ZRUNV DQG H[SORVLYH JURZWK LQ WKH DVVRFLDWHG LQWHJUDWHGRSWLFDO GHYLFHV IRU WKHVH QHWn ZRUNV VXFK DV ODVHUV ILOWHUV DQG KLJKVSHHG PRGXODWRUV 7KH GHPDQGV LPSRVHG E\ WKHVH GHYLFHV DUH FRQVLGHUHG EHIRUH SUHVHQWLQJ SRVVLEOH PDWHULDOV WR EH XVHG LQ WKHLU IDEULFDWLRQ $PRQJ WKH VXLWDEOH PDWHULDOV FRQVLGHUDWLRQ RI WKHLU SURSHUWLHV UHYHDO WKDW OLWKLXP WDQWDODWH /L7Df SRVVHVVHV D QXPEHU RI GLVWLQFW DGYDQWDJHV SDUWLFXODUO\ D KLJKHU RSWLFDO GDPDJH WKUHVKROG DQG D VKRUWHU ZDYHOHQJWK XOWUDYLROHW 89f DEVRUSWLRQ HGJH 8VH RI /L7D DV D KRVW IRU WKH IDEULFDWLRQ RI LQWHJUDWHGRSWLFDO GHYLFHV IDFHV D QXPEHU REVWDFOHV )RU LQVWDQFH WKH FKDUDFWHULVWLFV RI (ULQGLIIXVLRQ IRU ODVHU IDEULFDWLRQ KDYH QRW \HW EHHQ VWXGLHG LQ WKLV PDWHULDO $GGLWLRQDOO\ PXFK ZRUN QHHGV WR EH GRQH RQ WKH FKDUDFWHUL]DWLRQ RI GLIIHUHQW ZDYHJXLGH IDEULFDWLRQ WHFKQLTXHV 7KH PRVW FRPPRQ WHFKQLTXH DQQHDOHG SURWRQ H[FKDQJH LV NQRZQ WR H[KLELW D QXPEHU RI DQRPDOLHV SUHYHQWn LQJ LW IURP EHLQJ XVHG LQ WKH SURGXFWLRQ RI UHOLDEOH GHYLFHV WKH PRVW VLJQLILFDQW DQRPDO\ EHLQJ WKH LQVWDELOLW\ RI ZDYHJXLGH LQGH[ SURILOHV RYHU WLPH $V IRU WKH RWKHU PHWKRGV RI ZDYHJXLGH IRUPDWLRQ QR GDWD FKDUDFWHUL]LQJ UHOHYDQW LVVXHV DVVRFLDWHG ZLWK HDFK SURFHVV VXFK DV VWDELOLW\ KDV EHHQ UHSRUWHG

PAGE 12

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f DQG FRPSXWHU QHWZRUN DSSOLFDWLRQV VXFK DV WKH LQWHUQHW FRQWLQn XHV WR LQFUHDVH VLJQLILFDQWO\ \HDU E\ \HDU $V LOOXVWUDWHG LQ )LJ IRU WKH WHOHFRPPXQLn FDWLRQ LQGXVWU\ DORQH WKLV WUHQG WRZDUG UDSLGO\ LQFUHDVLQJ EDQGZLGWK GHPDQG LV H[SHFWHG \HDU )LJ 5HODWLYH WUHQG LQ WKH GHPDQG IRU WHOHFRPPXQLFDWLRQ QHWn ZRUN EDQGZLGWK

PAGE 13

WR QHDUO\ TXDGUXSOH LQ WKH QH[W ILYH \HDUV %HFDXVH RI WKH ODUJH EDQGZLGWK LW RIIHUV RSWLFDO ILEHU KDV EHHQ HPSOR\HG DV WKH WUDQVPLVVLRQ PHGLXP WR PHHW WKLV GHPDQG RSHUDWLQJ DW ZDYHOHQJWKV DURXQG SP EHFDXVH RI LWV GHFUHDVHG ORVV DQG ORZ GLVSHUVLRQ DW WKLV ZDYHn OHQJWK ,Q DGGLWLRQ ZDYHOHQJWK GLYLVLRQ PXOWLSOH[LQJ :'0f LQ KLJKO\GHQVH QHWZRUNV LV RIWHQ HPSOR\HG WR WDNH IXUWKHU DGYDQWDJH RI WKH ILEHUfV JHQHURXV EDQGZLGWK ,W IROORZV WKDW D ODUJH VXUJH LQ WKH SURGXFWLRQ RI LQWHJUDWHGRSWLFDO GHYLFHV IRU XVH LQ EURDGEDQG ILEHURSWLF QHWZRUNV DV ZHOO DV RWKHU FRPPHUFLDO DQG PLOLWDU\ DSSOLFDWLRQV LV FOHDUO\ HYLGHQW 6RPH RI WKH PRVW SURPLVLQJ GHYLFHV LQFOXGH WKH IROORZLQJ WUDYHOLQJ ZDYH PRGXODWRUV KLJKRXWSXW SP ODVHUV DQG KLJKVSHHG RSWLFDO VZLWFKHV IRU XVH LQ ILEHURSWLF FRPPXQLFDWLRQV V\VWHPV >$OI@ ZDYHOHQJWKVHOHFWLYH WXQDEOH ILOWHUV IRU XVH LQ :'0 V\VWHPV >+XV@ RSWLFDO SDUDPHWULF RVFLOODWRUV 232f IRU JHQHUDWLRQ RI FRKHUn HQW ORQJ ZDYHOHQJWK SPf UDGLDWLRQ IRU XVH LQ VSHFWURVFRS\ DQG ODVHU UDGDU QRQOLQn HDU IUHTXHQF\ GRXEOHUV IRU EOXHOLJKW JHQHUDWLRQ QHHGHG LQ RSWLFDO VWRUDJH DSSOLFDWLRQV >ƒKOD ƒKOF 1DN@ DQG KLJKIUHTXHQF\ EURDGEDQG VXUIDFH DFRXVWLF ZDYH 6$:f ILOWHUV IRU XVH LQ RSWLFDO VSHFWUXP DQDO\]HUV DQG WHOHYLVLRQ DSSOLFDWLRQV >;X@ ,Q RUGHU WR IDEULFDWH WKHVH GHYLFHV ZH PXVW ILQG D VXLWDEOH PDWHULDO ZKRVH FKDUDFn WHULVWLFV DUH VXFK WKDW LW LV DEOH WR PHHW WKH GHPDQGV LPSRVHG E\ HDFK GHYLFH ,Q SDUWLFXODU LW PXVW EH WUDQVSDUHQW LQ WKH ZDYHOHQJWK UHJLRQV RI LQWHUHVW IRU ERWK GLIIHUHQFH IUHTXHQF\ JHQHUDWLRQ RI LQIUDUHG DV LQ 232Vf DQG LQ WKH YLVLEOH IRU EOXHOLJKW JHQHUDWLRQ E\ QRQOLQn HDU IUHTXHQF\ GRXEOLQJ ,W VKRXOG EH FDSDEOH RI SURGXFLQJ ZDYHJXLGHV ZLWK ORZ SURSDJDn WLRQ ORVV DQG PXVW SRVVHVV D ODUJH HOHFWURRSWLF FRHIILFLHQW GHVLUDEOH LQ WKH IDEULFDWLRQ RI HOHFWURRSWLF PRGXODWRUV DQG VZLWFKHV ,W PXVW DOVR SRVVHVV D ODUJH QRQOLQHDU FRHIILFLHQW DQG EH FDSDEOH RI VDWLVI\LQJ WKH SKDVH PDWFKLQJ FRQGLWLRQ VR WKDW GLIIHUHQFH IUHTXHQF\

PAGE 14

JHQHUDWLRQ DQG VHFRQGKDUPRQLF JHQHUDWLRQ 6+*f RI EOXH OLJKW DUH SRVVLEOH ,W VKRXOG KDYH ORZ DFRXVWLF ORVVHV WR DOORZ WKH IDEULFDWLRQ RI 6$: GHYLFHV VXFK DV ILOWHUV ,W PXVW EH DEOH WR VHUYH DV D KRVW IRU WKH LQGLIIXVLRQ RI UDUHHDUWK PDWHULDOV VXFK DV (U ZKLFK SURn GXFHV ODVLQJ HPLVVLRQV QHDU SP ZKHQ RSWLFDOO\ SXPSHG )LQDOO\ LW PXVW EH FRPPHUn FLDOO\ DYDLODEOH ZLWK KLJK RSWLFDO TXDOLW\ DQG DW ORZ FRVW 7KH IHUURHOHFWULF PDWHULDOV OLWKLXP QLREDWH /L1Ef OLWKLXP WDQWDODWH /L7Df DQG SRWDVVLXP WLWDQ\O SKRVSKDWH .73f DUH WKUHH RI WKH PRVW SURPLVLQJ FDQGLGDWHV ZKLFK VDWLVI\ WKH DERYH FRQGLWLRQV DQG KDYH UHFHLYHG ZLGHVSUHDG XVH LQ WKH IDEULFDWLRQ RI WKH DIRUHPHQWLRQHG GHYLFHV >ƒKOD ƒKOF (+ 0L] 1DN ;X@ ,Q WKH QH[W VHFn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

PAGE 15

LQ /L1E ,Q DGGLWLRQ /L7D KDV D VKRUWHU ZDYHOHQJWK 89 DEVRUSWLRQ HGJH QPf WKDQ /L1E QPf SHUPLWWLQJ QRQOLQHDU FRQYHUVLRQ E\ IUHTXHQF\ GRXEOLQJ WR VKRUWHU ZDYHOHQJWKV ZLWK OHVV DEVRUSWLRQ 7KRXJK /L7D KDV D ODUJH QRQOLQHDU FRHIILFLHQW G DQG DQ HOHFWURRSWLF FRHIILn FLHQW U DV ODUJH DV WKDW LQ /L1E >0L] $KO 0DWE@ +RZHYHU QRQOLQHDU DSSOLFDWLRQV VXFK DV VHFRQG KDUPRQLF JHQHUDWLRQ 6+*f RI EOXH OLJKW FDQ EH DFKLHYHG LQ /L7D ZLWK WKH DLG RI TXDVLSKDVH PDWFKLQJ 430f +HQFH /L7DfV KLJKHU RSWLFDO GDPDJH WKUHVKROG DQG ORZ SURSDJDWLRQ ORVV PDNH LW DQ LGHDO PDWHULDO IRU 4306+* RI EOXH OLJKW FDSDEOH RI RXWSXW SRZHU OHYHOV IDU H[FHHGLQJ WKRVH RI /L1E >ƒKO (+
PAGE 16

)DEULFDWLRQ 3URFHVVHV LQ /L7D4 6HYHUDO PHWKRGV H[LVW IRU WKH IRUPDWLRQ RI ZDYHJXLGHV LQ IHUURHOHFWULF PDWHULDOV )RU /L7D LQ SDUWLFXODU WKHUH DUH DW OHDVW WKUHH UHFRJQL]HG SURFHGXUHV 7KH\ DUH 7L PHWDO LQGLIIXVLRQ =Q YDSRULQGLIIXVLRQ DQG DQQHDOHG SURWRQ H[FKDQJH $3(f $GGLWLRQDOO\ LQGLIIXVLRQ RI UDUHHDUWK PDWHULDOV VXFK DV 1G DQG (U KDYH EHHQ H[DPLQHG IRU WKH SURGXFn WLRQ RI ODVHU VRXUFHV 7KRXJK WKHVH WHFKQLTXHV DUH DOO GHVFULEHG LQ JUHDWHU GHWDLO LQ VXEVHn TXHQW FKDSWHUV WKH\ DUH LQWURGXFHG KHUH EULHIO\ WR SRLQW RXW VRPH RI WKH OLPLWDWLRQV DQG DQRPDOLHV DVVRFLDWHG ZLWK HDFK DQG WKHUHIRUH GHYHORS WKH PRWLYDWLRQ IRU WKLV ZRUN ,Q RUGHU WR FUHDWH DQ DFWLYH JDLQ UHJLRQ IRU WKH IDFLOLWDWLRQ RI D ODVHU VRXUFH RU ZDYHJXLGH DPSOLILHU LQ /L7D D UDUHHDUWK PDWHULDO VXFK DV 1G RU (U QHHGV WR EH LQGLI IXVHG 1G ZKHQ GRSHG LQ D KRVW PDWHULDO DQG RSWLFDOO\ SXPSHG LV NQRZQ WR SURGXFH ODVn LQJ HPLVVLRQV QHDU O_LP >1RX 6DQ@ )RU FRPPXQLFDWLRQ DSSOLFDWLRQV KRZHYHU LW LV GHVLUDEOH WR KDYH HPLVVLRQV QHDU LP WR WDNH DGYDQWDJH RI WKH GHFUHDVHG ORVV DQG ORZ GLVSHUVLRQ RI RSWLFDO ILEHU DW WKLV ZDYHOHQJWK (UGRSHG PDWHULDO ZKHQ RSWLFDOO\ SXPSHG LV NQRZQ WR SURGXFH ODVLQJ HPLVVLRQV QHDU _LP >*LO@ 7KRXJK SUHYLRXV ZRUN KDV FRQn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

PAGE 17

2I WKH WKUHH PHWKRGV IRU ZDYHJXLGH IDEULFDWLRQ RXWOLQHG DERYH $3( LV SUHVHQWO\ WKH PRVW FRPPRQ &RPSDUHG WR WKH RWKHU WHFKQLTXHV $3( RIIHUV WKH DGYDQWDJHV RI VLPn SOLFLW\ IOH[LELOLW\ KLJK LQGH[ LQFUHPHQW DQG ORZ SURSDJDWLRQ ORVV ,Q DGGLWLRQ $3( FDQ EH SHUIRUPHG DW WHPSHUDWXUHV EHORZ WKH &XULH SRLQW RI WKH FU\VWDO r&f DVVXULQJ WKDW WKLV WHFKQLTXH GRHV QRW GLVUXSW WKH PRQRGRPDLQ VWUXFWXUH RI WKH FU\VWDO DQG VR QR SRVWn IDEULFDWLRQ UHSROLQJ SURFHVV LV QHHGHG +RZHYHU D QXPEHU RI DQRPDOLHV DUH NQRZQ WR H[LVW IRU $3( ZDYHJXLGHV LQ /L7D )RU LQVWDQFH LW KDV EHHQ VKRZQ WKDW WKH SHDN H[WUDRUGLQDU\ LQGH[ LQFUHPHQW LQFUHDVHV GXULQJ WKH LQLWLDO VWDJHV RI DQQHDOLQJ >ƒKOF 0DWE@ DQG WKDW WKH LQGH[ GHFUHDVHV ZLWK + SURWRQ FRQFHQWUDWLRQ DV SURWRQ FRQFHQWUDWLRQ H[FHHGV D FHUWDLQ SRLQW 7KHVH IDFWV VXJJHVW D QRQOLQHDU GHSHQGHQFH RI LQGH[ LQFUHPHQW RQ SURWRQ FRQFHQWUDWLRQ DQG WKH SRVVLELOLW\ WR IRUP EXULHG UHIUDFWLYH LQGH[ SURILOHV >ƒKO 'DY 0DUD@ 6XFK D EXULHG SURILOH KDV UHFHQWO\ EHHQ GLUHFWO\ REVHUYHG >0DUE@ DQG LV H[KLELWHG LQ )LJ 7KRXJK EXULHG LQGH[ ZDYHJXLGHV FDQ SURYLGH WKH DGYDQWDJHV RI LPSURYHG ILEHUWRGHYLFH FRXSOLQJ DQG UHGXFHG VFDWWHULQJ ORVVHV DW WKH VXEVWUDWHDLU LQWHUIDFH WKH\ VHULn RXVO\ XQGHUPLQH WKH DGHTXDF\ RI SUHYLRXV WKHRUHWLFDO VWXGLHV PRVW RI ZKLFK UHOLHG RQ WKH LQGLUHFW SURn ILOH UHFRQVWUXFWLRQ WHFKQLTXH RI LQYHUVH :.% DQG WKH DVVXPSWLRQ )LJ 5HIUDFWLYHLQGH[ SURILOHV RI H[WUDRUGLn QDU\ SRODUL]DWLRQ IRU YDULRXV DQQHDOn LQJ WLPHV PHDVXUHG DW SP

PAGE 18

)LJ 7HPSRUDO FKDQJHV LQ WKH LQGH[ SURILOH H[WUDRUGLQDU\ SRODUL]DWLRQf PHDVXUHG DW A SP LPPHGLDWHO\ DIWHU IDEULFDWLRQ DQG ZHHNV ODWHU RI D PRQRWRQLFDOO\ GHFUHDVLQJ LQGH[ SURILOH 0RVW LPSRUWDQW IURP D SUDFWLFDO VWDQGSRLQW $3( ZDYHJXLGHV LQ /L7D KDYH EHHQ VKRZQ WR H[KLELW WHPSRUDO LQVWDELOLn WLHV >ƒKOF 0DUE 0DU22@ PDQLIHVWLQJ WKHPVHOYHV LQ WKH IRUP RI VLJQLILFDQW LQGH[ SURILOH PRGLILn FDWLRQV IRU ERWK VKRUW DQG ORQJ SHULRGV DIWHU IDEULFDWLRQ $Q H[DPSOH RI WKLV LV GHSLFWHG LQ )LJ VKRZLQJ WKH H[WUDRUGLQDU\ LQGH[ SURILOH RI D ZDYHJXLGH PHDVXUHG LPPHGLDWHO\ DIWHU IDEULFDWLRQ DQG DJDLQ IRXU ZHHNV ODWHU 7KHVH LQVWDELOLWLHV DUH EHOLHYHG E\ VRPH WR EH GXH WR WKH SUHVHQFH RI PXOWLSOH FU\VWDO SKDVHV ZLWKLQ WKH SURWRQ H[FKDQJHG UHJLRQ >(+ .RU@ WKRXJK LW KDV QRW EHHQ GLUHFWO\ FRQILUPHG 7KLV IDFW SUHYHQWV /L7D IURP EHLQJ XVHG LQ WKH IDEULFDWLRQ RI SUDFWLFDOO\ YLDEOH WKDW LV WHPSRUDOO\ VWDEOH ZDYHJXLGHV IRU GHYLFH DSSOLFDWLRQV $V D IXUWKHU SRLQW LW LV NQRZQ WKDW WKH LQWURGXFWLRQ RI + WKURXJK SURWRQ H[FKDQJH LQWR D UDUHHDUWK GRSHG UHJLRQ OHDGV WR VLJQLILFDQW TXHQFKLQJ RI ODVHU DFWLRQ RU D UHGXFWLRQ LQ H[FLWHG VWDWH OLIHWLPH WKHUHE\ UHGXFLQJ ODVHU HIILFLHQF\ $V VXFK DGGLWLRQDO PHWKRGV IRU ZDYHJXLGH IDEULFDWLRQ QHHG EH H[SORUHG $V VWDWHG HDUOLHU WKH RWKHU WZR UHFRJQL]HG WHFKQLTXHV RI ZDYHJXLGH IRUPDWLRQ LQ /L7D DUH 7L PHWDOLQGLIIXVLRQ DQG =Q YDSRULQGLIIXVLRQ :KLOH D VPDOO DPRXQW RI ZRUN KDV EHHQ UHSRUWHG LQ WKH OLWHUDWXUH DERXW WKH VXFFHVV RI WKHVH WHFKQLTXHV WKH\ KDYH QRW

PAGE 19

EHHQ DV VHULRXVO\ LQYHVWLJDWHG DV KDV $3( :KLOH LW LV OLNHO\ WKDW WKHVH WHFKQLTXHV PD\ EH EHWWHU VXLWHG IRU WKH SURGXFWLRQ RI ZDYHJXLGHV LQ (UGRSHG UHJLRQV EHFDXVH WKH\ ZRQf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f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

PAGE 20

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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n JOHSKDVH ZDYHJXLGHV ZKLFK DUH EHOLHYHG WR EH PRUH VWDEOH 2QFH IDEULFDWHG WKH VWDELOLW\ RI WKHVH ZDYHJXLGHV QHHGV WR EH TXDQWLILHG DQG PRQLWRUHG RYHU WLPH )LQDOO\ FKDUDFWHUL]DWLRQ RI ZDYHJXLGHV IRUPHG E\ 7L PHWDOLQGLIIXVLRQ DQG =Q YDSRULQGLIIXVLRQ QHHGV WR EH SHUIRUPHG )DEULFDWLRQ FRQGLWLRQV OHDGLQJ WR WKH IRUPDWLRQ RI VLQJOHPRGH ZDYHJXLGHV DW SP QHHG WR EH LGHQWLILHG 6LQFH WKHVH SURFHVVHV FDXVH QR

PAGE 21

DGGLWLRQDO ODWWLFH VWUDLQ DV GRHV $3( WKHUH LV QR VWUXFWXUDO SKDVH GLDJUDP WR EH FRQn VWUXFWHG +RZHYHU WKH VWDELOLW\ RI ZDYHJXLGHV IRUPHG E\ HDFK SURFHVV QHHGV DOVR WR EH TXDQWLILHG DQG PRQLWRUHG RYHU WLPH VR WKDW LW PD\ EH FRPSDUHG WR WKDW RI $3( ZDYHJXLGHV &KDSWHU 2UJDQL]DWLRQ ,Q &KDSWHU WKH YDULRXV PHWKRGV RI ZDYHJXLGH IDEULFDWLRQ XVHG LQ WKLV UHVHDUFK DUH GLVFXVVHG $ GHVFULSWLRQ RI WKH $3( SURFHVV LV SURYLGHG IROORZHG E\ VRPH RI WKH IHDn WXUHV DQG DQRPDOLHV RI WKLV SURFHVV LQ /L7D 7KHQ WKH DFWXDO FRQGLWLRQV XVHG IRU WKH IDEn ULFDWLRQ RI $3( ZDYHJXLGHV DUH JLYHQ 1H[W D GHVFULSWLRQ RI WKH ZDYHJXLGH IRUPDWLRQ SURFHVVHV RI 7LLQGLIIXVLRQ DQG =Q YDSRULQGLIIXVLRQ LV SUHVHQWHG $JDLQ VRPH RI WKH IHDn WXUHV DVVRFLDWHG ZLWK HDFK SURFHVV DUH SURYLGHG DORQJ ZLWK WKH DFWXDO IDEULFDWLRQ FRQGLn WLRQV XVHG 7KH UHVW RI WKH FKDSWHU GHWDLOV VRPH DGGLWLRQDO SURFHVVLQJ VWHSV LQYROYHG LQ ZDYHJXLGH DQG GHYLFH IDEULFDWLRQ VXFK DV FU\VWDO UHSROLQJ HQGIDFH SROLVKLQJ WKH DSSOLFDn WLRQ RI DQWLUHIOHFWLYLW\ FRDWLQJV DQG HOHFWURSODWLQJ RI WKLFN JROG HOHFWURGHV &KDSWHU H[DPLQHV WKH SURFHVV RI UDUHHDUWK (UGRSLQJ LQ /L7D D ILUVW VWHS LQ WKH IDEULFDWLRQ RI D LP DFWLYH ZDYHJXLGH ODVHU 6RPH EDFNJURXQG LQIRUPDWLRQ UHODWHG WR WKH SURFHVV RI (UGRSLQJ LQ /L1E LV SUHVHQWHG 7KHQ WKH OLPLWDWLRQV DVVRFLDWHG ZLWK WKH XVH RI /L1E DUH SUHVHQWHG RSHQLQJ WKH GRRU IRU WKH XVH RI /L7D DV D KRVW PDWHULDO IRU (UGRSLQJ 1H[W WKH DFWXDO FRQGLWLRQV XVHG IRU WKH GRSLQJ RI /L7D ZLWK (U DUH RXWn OLQHG DIWHU ZKLFK FKDUDFWHUL]DWLRQ LV SHUIRUPHG E\ PHDQV RI 5DPDQ VSHFWURVFRS\ DQG IOXn RUHVFHQFH PHDVXUHPHQWV WR H[DPLQH WKH HIIHFW RI (UGRSLQJ RQ WKH SKRWRUHIUDFWLYH

PAGE 22

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n WLRQDO WHFKQLTXH LV LQWURGXFHG DQG DSSOLHG WR PHDVXUH WKH PRGH HIIHFWLYH LQGLFHV IRU ZDYHJXLGHV DW VHYHUDO GLIIHUHQW ZDYHOHQJWKV :DYHJXLGH ORVV LV H[DPLQHG QH[W ZKHUH PHDVXUHPHQWV RI WRWDO LQVHUWLRQ ORVV DQG ZDYHJXLGH SURSDJDWLRQ ORVV DUH PDGH WKH ODWWHU E\ D )DEU\3HURW WHFKQLTXH 7KHQ WKH HOHFWURRSWLF FRHIILFLHQW U LV PHDVXUHG GLUHFWO\ XVLQJ D XQLW\RYHUODS WHFKQLTXH ZKLFK GRHV QRW UHTXLUH WKH HVWLPDWLRQ RI DQ HOHFWULFDORSWLn FDO ILHOG RYHUODS IDFWRU 7R GHPRQVWUDWH WKH VXSHULRU SKRWRUHIUDFWLYH UHVLVWDQFH RI /L7D ZDYHJXLGHV QHDU _LP SORWV RI RXWSXW SRZHU YHUVXV LQSXW SRZHU DUH PHDVXUHG QH[W DQG FRPSDUHG WR VLPLODU VWUXFWXUHV LQ /L1E )LQDOO\ WKH PHWKRGV RI GLUHFW LQGH[ SURILOLQJ DQG ;UD\ GLIIUDFWLRQ URFNLQJ FXUYH DQDO\VLV DUH H[SODLQHG DQG DSSOLHG IRU WKH GLUHFW GHWHUn PLQDWLRQ RI VXUIDFH LQGH[ LQFUHPHQW $Q DQG SURWRQLQGXFHG ODWWLFH VWUDLQ H DW WKH VXUn IDFH RI $3( ZDYHJXLGHV 7KHVH ILJXUHV DUH QHFHVVDU\ IRU WKH FRQVWUXFWLRQ RI WKH VWUXFWXUDO SKDVH GLDJUDP IRU $3( ZDYHJXLGHV SHUIRUPHG LQ &KDSWHU ,Q RUGHU WR H[SODLQ WKH DQRPDOLHV DVVRFLDWHG ZLWK WKH $3( SURFHVV LQ /L7D WKH VWUXFWXUDO SKDVH GLDJUDP IRU WKLV PDWHULDO QHHGV WR EH FRQVWUXFWHG 7KLV LV GRQH LQ &KDSWHU ZKHUH VHYHUDO ZDYHJXLGHV ZLWK GLIIHUHQW YDOXHV RI VXUIDFH FRQFHQWUDWLRQ DUH IDEULFDWHG E\ $3( DQG PHDVXUHG XVLQJ WKH WHFKQLTXHV RI GLUHFW LQGH[ SURILOLQJ DQG URFNLQJ FXUYH

PAGE 23

DQDO\VLV 5HVXOWLQJ YDOXHV RI VXUIDFH LQGH[ LQFUHPHQW $Q DQG ODWWLFH VWUDLQ H DUH SORWWHG WR FRQVWUXFW WKH VWUXFWXUDO SKDVH GLDJUDP 7KH GLDJUDP LV WKHQ XVHG WR H[SODLQ WKH SUHYLn RXVO\ REVHUYHG DQRPDOLHV DVVRFLDWHG ZLWK WKH $3( SURFHVV LQ /L7D DQG WKH HIIHFW RI FU\VWDO SKDVHV RQ WKH VKDSH RI WKH ZDYHJXLGH LQGH[ SURILOH LV DOVR H[DPLQHG ZLWK H[DPn SOHV RI PHDVXUHG SURILOHV IURP GLIIHUHQW UHJLRQV RI WKH VWUXFWXUDO SKDVH GLDJUDP /DVWO\ SURWRQ H[FKDQJHG 3(f ZDYHJXLGHV IDEULFDWHG LQ D GLOXWH VRXUFH DUH H[DPLQHG H[KLELWLQJ DGYDQWDJHV RYHU $3( ZDYHJXLGHV 7KHVH ZDYHJXLGHV DUH FKDUDFWHUL]HG E\ WKH VDPH GLUHFW PHDVXUHPHQW WHFKQLTXHV XVHG RQ $3( ZDYHJXLGHV DQG E\ 5DPDQ VSHFWURVFRS\ WR GHWHUn PLQH WKH SUHVHQFH RI DQ DGGLWLRQDO VLQJOHSKDVH UHJLRQ RI WKH VWUXFWXUDO SKDVH GLDJUDP REWDLQDEOH RQO\ E\ 3( LQ D GLOXWH VRXUFH &KDSWHU H[DPLQHV WKH VWDELOLW\ LVVXH RI ZDYHJXLGHV LQ /L7D $3( ZDYHJXLGHV DUH LQVSHFWHG E\ URFNLQJ FXUYH DQDO\VLV WR WUDFN FKDQJHV LQ ODWWLFH VWUDLQ H RYHU WLPH )RU D PRUH TXDQWLWDWLYH DQDO\VLV GLUHFWLRQDO FRXSOHUV DUH IDEULFDWHG IURP GLIIHUHQW UHJLRQV RI WKH SKDVH GLDJUDP DQG FKDQJHV LQ PHDVXUHG FRXSOLQJ OHQJWK DUH UHODWHG WR FRUUHVSRQGn LQJ FKDQJHV LQ LQGH[ LQFUHPHQW $Qf WKURXJK FRPSXWHU VLPXODWLRQ RI D GLUHFWLRQDO FRXn SOHU 5HVXOWV LQGLFDWHG WKDW VWDELOLW\ LV QRW FDXVHG E\ WKH SUHVHQFH RI PXOWLSOH SKDVHV DV SUHYLRXVO\ DVVXPHG 7R GHWHUPLQH LI VWDELOLW\ GHSHQGV RQ IDEULFDWLRQ FRQGLWLRQV WKH VWDn ELOLW\ RI 7LLQGLIIXVHG DQG =Q YDSRULQGLIIXVHG DUH DOVR PHDVXUHG VKRZLQJ QR LPSURYHn PHQW LQ VWDELOLW\ XQOLNH /L1E 6XVSHFWLQJ LQIHULRU JURZWK WHFKQRORJ\ WR EH WKH UHDVRQ IRU WKH LQVWDELOLW\ FU\VWDOV IURP VHYHUDO GLIIHUHQW JURZHUV ZRUOGZLGH ERWK 6$: DQG RSWLn FDO JUDGH DUH REWDLQHG DQG FRPSDUHG LQ WHUPV RI VWDELOLW\ RQO\ WR ILQG WKDW DOO DUH VLPLODU DQG XQVWDEOH

PAGE 24

7KRXJK FULWLFDO DGYDQFHPHQWV DUH QHHGHG LQ WKH JURZWK WHFKQRORJ\ IRU /L7D D GHPDQG IRU KLJKTXDOLW\ RSWLFDO JUDGH FU\VWDOV PXVW H[LVW EHIRUH WLPH DQG PRQH\ ZLOO EH VSHQW RQ VXFK DQ HIIRUW 7R WKLV HQG WKH SRWHQWLDO RI /L7D IRU LQWHJUDWHGRSWLFDO FRPSRn QHQWV LV GHPRQVWUDWHG LQ &KDSWHU ZLWK WKH GHVLJQ IDEULFDWLRQ DQG WHVWLQJ RI D V\PPHWULF 0DFK=HKQGHU LQWHUIHURPHWHU 0=,f WUDYHOLQJZDYH HOHFWURRSWLF PRGXODWRU )LUVW WKH IDEULFDWLRQ FRQGLWLRQV IRU WKH GHYLFH DUH SURYLGHG 1H[W VRPH PLFURZDYH FKDUDFWHULVWLFV RI WKH GHYLFH DUH H[DPLQHG LQFOXGLQJ PHDVXUHPHQWV RI WKH PLFURZDYH ORVV FRHIILFLHQW PLFURZDYH HIIHFWLYH LQGH[ DQG FKDUDFWHULVWLF LPSHGDQFH 7KHQ WKH '& UHVSRQVH RI WKH GHYLFH LV PHDVXUHG LQ RUGHU WR REWDLQ WKH H[WLQFWLRQ UDWLR DQG KDOIZDYH YROWDJH 7KH SRZHU KDQGOLQJ FDSDELOLW\ RI WKH PRGXODWRU LV H[DPLQHG QH[W E\ SORWWLQJ RXWSXW SRZHU YHUn VXV LQSXW SRZHU QHDU _LP $GGLWLRQDOO\ WKH IUHTXHQF\ UHVSRQVH RI WKH GHYLFH LV PHDn VXUHG DQG FRPSDUHG WR WKHRU\ )LQDOO\ &KDSWHU SUHVHQWV WKH FRQFOXVLRQV IRU WKLV GLVVHUWDWLRQ DQG RXWOLQHV VRPH IXWXUH ZRUN $ VXPPDU\ RI WKH UHVHDUFK SHUIRUPHG LV VWDWHG DQG WKH LPSRUWDQW FRQWULEXn WLRQV UHVXOWLQJ IURP WKLV ZRUN DUH HPSKDVL]HG 6RPH IXWXUH ZRUN WR EH SHUIRUPHG LQ WKLV DUHD VRPH RI LW FXUUHQWO\ XQGHU ZD\ LV WKHQ RXWOLQHG

PAGE 25

&+$37(5 )$%5,&$7,21 2) 3)$1$5 $1' &+$11() :$9(*8,'(6 ,1 /L7D 3ODQDU DQG FKDQQHO ZDYHJXLGHV DUH IRUPHG E\ FUHDWLQJ DQ DUHD RI KLJKHU UHIUDFWLYH LQGH[ LQ D GLHOHFWULF PDWHULDO VXFK DV /L7D 8QOLNH SODQDU ZDYHJXLGHV FKDQQHO ZDYHJXLGHV FRQILQH WKH RSWLFDO ZDYH LQ WZR GLPHQVLRQV 0RVW SURFHVVHV UHVXOW LQ D URXJKO\ VHPLHOOLSWLFDO JUDGHGLQGH[ SURILOH WKDW GHFUHDVHV PRQRWRQLFDOO\ IURP WKH VXEn VWUDWH VXUIDFH >5DP@ 6XFK D JXLGH LV VFKHPDWLFDOO\ UHSUHVHQWHG LQ )LJ ZKHUH WKH SURSDJDWLRQ GLUHFWLRQ RI WKH RSWLFDO ZDYH LV DOVR VKRZQ 7KHVH ZDYHJXLGHV IRUP WKH EDVLV IRU D PXOWLWXGH RI GHYLFHV LQ WKH LQWHJUDWHGRSWLF HPERGLPHQW LQ SDUWLFXODU DPSOLWXGH PRGXODWRUV XVHG LQ FRPPXQLFDWLRQ V\VWHPV KLJKVSHHG RSWLFDO VZLWFKHV QRQOLQHDU IUHn TXHQF\ GRXEOHUV DQG RSWLFDO SDUDPHWULF RVFLOODWRUV VXEVWUDWH )LJ 'HSLFWLRQ RI D FKDQQHO ZDYHJXLGH

PAGE 26

7KLV FKDSWHU GHVFULEHV WKH PHWKRGV XVHG LQ WKH FUHDWLRQ RI RSWLFDO ZDYHJXLGHV LQ GLHOHFWULF VXEVWUDWHV 3DUWLFXODU DWWHQWLRQ LV SDLG WR WKH SURWRQ H[FKDQJH DQG DQQHDOHG SURn WRQ H[FKDQJH WHFKQLTXHV DV ZHOO DV WR WKH PHWKRGV RI 7L PHWDOLQGLIIXVLRQ DQG =Q YDSRU LQGLIIXVLRQ XVHG WR IDEULFDWH ZDYHJXLGHV LQ /L7D 6HYHUDO GLVWLQFW IHDWXUHV DQG DQRPDn OLHV DWWULEXWHG WR HDFK RI WKH SURFHVVHV DUH GLVFXVVHG 6SHFLILF IDEULFDWLRQ SDUDPHWHUV DUH JLYHQ LQ RUGHU WR FRYHU FRPSOHWHO\ WKH IDEULFDWLRQ PHWKRGV XVHG LQ WKLV UHVHDUFK DQG D IHZ DGGLWLRQDO SURFHVVLQJ VWHSV DUH DOVR GHVFULEHG :DYHJXLGH )RUPDWLRQ E\ WKH $3( 3URFHVV :DYHJXLGHV LQ /L1E DQG /L7D FDQ EH IRUPHG E\ XVLQJ VHYHUDO PHWKRGV WKH PRVW SRSXODU RI ZKLFK DUH E\ PHWDO LQGLIIXVLRQ DW KLJK WHPSHUDWXUHV DQG E\ WKH H[FKDQJH RI /L LRQV LQ WKH VXEVWUDWH E\ + LRQV VXSSOLHG E\ DQ DFLG VRXUFH f§ SURWRQ H[FKDQJH RU DQQHDOHG SURWRQ H[FKDQJH PHWKRG >5DP 6\P@ ,Q WKLV VHFWLRQ WKH DQQHDOHG SURWRQ H[FKDQJH PHWKRG RI ZDYHJXLGH IDEULFDWLRQ LV GLVFXVVHG 7KH IHDWXUHV RI ZDYHJXLGHV IDEn ULFDWHG E\ WKLV SURFHVV LQ /L7D DUH WKHQ GHVFULEHG IROORZHG E\ VRPH RI WKH DVVRFLDWHG DQRPDOLHV REVHUYHG )LQDOO\ WKH DFWXDO IDEULFDWLRQ FRQGLWLRQV XVHG WR SURGXFH ZDYHJXLGHV E\ WKLV WHFKQLTXH DUH JLYHQ 7KH 3($3( 3URFHVV $V GLVFRYHUHG E\ -DFNHO 5LFH DQG 9HVHOND LQ WKH SURWRQ H[FKDQJH 3(f SURn FHVV UHVXOWV LQ D ODUJH LQFUHDVH RI WKH H[WUDRUGLQDU\ UHIUDFWLYH LQGH[ RI /L1E >-DF@ 7KH WHFKQLTXH ZDV ODWHU VKRZQ WR DOVR EH HIIHFWLYH LQ IRUPLQJ ZDYHJXLGHV LQ /L7D E\ 6SLOOPDQ 6DQIRUG DQG 6RUHI >6SL@ 7KRXJK WKH 3( SURFHVV LQLWLDOO\ FDXVHV D UHGXFWLRQ

PAGE 27

RI WKH HOHFWURRSWLF HIIHFW WKH U FRHIILFLHQW FDQ EH DOPRVW FRPSOHWHO\ UHFRYHUHG WKURXJK R SRVWH[FKDQJH DQQHDOLQJ RI WKH VXEVWUDWH >$KOF .DQ@ ,Q WKH FDVH RI /L7D LWV H[WUDRUGLQDU\ LQGH[ LQFUHPHQW $QH DIWHU SURWRQ H[FKDQJH KDV EHHQ VKRZQ WR EH ORZHU WKDQ WKDW RI /L1E +RZHYHU DQQHDOLQJ FDXVHV DQ DQRPDORXV LQGH[ LQFUHDVH LQ /L7D FRPSDUHG WR DQ DQWLFLSDWHG GHFUHDVH IRU /L1E >ƒKOF 0DWE ƒKOE@ 7KHVH DUHDV RI LQYHUWHG GRPDLQ DUH QHFHVVDU\ LQ WKH IDEULFDWLRQ RI LQWHJUDWHG GHYLFHV LQWHQGHG IRU QRQOLQHDU RSWLFDO DSSOLFDWLRQV OLNH EOXH OLJKW JHQHUDWLRQ E\ IUHTXHQF\ GRPDLQ LQYHUWHG )LJ 'HSLFWLRQ RI DQ $3( FKDQQHO ZDYHJXLGH ODLG DFURVV UHJLRQV RI GRPDLQ LQYHUVLRQ DOVR IRUPHG E\ $3(

PAGE 28

GRXEOLQJ RI LQIUDUHG ZKHUH TXDVLSKDVH PDWFKLQJ 430f LV UHTXLUHG >ƒKOF 1DN@ )RU WKLV FDVH WKH $3( ZDYHJXLGHV DQG GRPDLQ LQYHUWHG UHJLRQV DUH IDEULFDWHG RQ WKH Fn VXUIDFH RI /L7D DV RSSRVHG WR WKH F VXUIDFH IRU /L1Ef VLQFH GRPDLQ LQYHUVLRQ VHHPV WR EH LQLWLDWHG RQ WKH F VXUIDFH >1DN 6DZ@ :KHQ XVLQJ SDWWHUQV RI GRPDLQ LQYHUVLRQ VXFK DV LQ 430 WKH ZDYHJXLGHV DUH RULHQWHG SHUSHQGLFXODU WR WKH H[LVWLQJ LQYHUWHG UHJLRQV DV LQ )LJ 6LQFH SURWRQ H[FKDQJH DQG SRVWH[FKDQJH DQQHDOLQJ IRU ZDYHJXLGH IDEULFDWLRQ FDQ EH GRQH DW WHPSHUDWXUHV ZHOO EHORZ r& WKH\ GR QRW DIIHFW WKH H[LVWLQJ LQYHUWHG UHJLRQV ZKLFK KDYH DOVR EHHQ FUHDWHG E\ $3( EXW DQQHDOHG DW WHPn SHUDWXUHV QHDU WKH &XULH WHPSHUDWXUH RI DERXW r& >)LQ 0L]@ 7KH H[FKDQJH VHWXS RI WKH $3( SURFHVV LV GHSLFWHG LQ )LJ IRU FKDQQHO ZDYHJXLGHV 7KH VXEVWUDWH LV ILUVW FRYHUHG E\ D PHWDO PDVNLQJ OD\HU XVXDOO\ $ RU 7Df LQ )LJ 6FKHPDWLF RI WKH SURWRQ H[FKDQJH SURFHVV

PAGE 29

ZKLFK QDUURZ VWULSH SDWWHUQV DUH GHOLQHDWHG E\ SKRWROLWKRJUDSK\ H[SRVLQJ UHJLRQV RI WKH VXEVWUDWH ZKHUH WKH H[FKDQJH LV WR WDNH SODFH ,Q WKH FDVH RI SODQDU ZDYHJXLGHV QR PDVN LV QHHGHG DQG WKH HQWLUH EXON VXUIDFH LV VXEMHFWHG WR SURWRQ H[FKDQJH 7KH VXEVWUDWH LV WKHQ LPPHUVHG LQ D KRW PHOW RI KLJK + FRQFHQWUDWLRQ XVXDOO\ EHQ]RLF RU S\URSKRVSKRULF DFLG ZKHUH /L GLIIXVHV RXW RI WKH VXEVWUDWH DQG LV UHSODFHG E\ + IURP WKH DFLG VRXUFH $V D UHVXOW DQ H[FKDQJHG OD\HU RI +[/LLB[7D +[/B[1E LI /L1E LV XVHG DV WKH VXEVWUDWHf LV IRUPHG DW WKH VXUIDFH ZLWK WKH WKLFNQHVV GHSHQGHQW RQ WKH PHOW WHPSHUn DWXUH 7 DQG H[FKDQJH WLPH W 2Q WKH RWKHU KDQG WKH UHIUDFWLYH LQGH[ FKDQJH LV SUDFWLFDOO\ LQGHSHQGHQW RI 7 DQG W EHLQJ GHWHUPLQHG ODUJHO\ E\ WKH FKRLFH RI WKH SURWRQ VRXUFH DFLGf 3\URSKRVSKRULF DFLG LV QRUPDOO\ SUHIHUUHG DV WKH + VRXUFH VLQFH LW KDV D ODUJHU GLIIXVLRQ FRHIILFLHQW 'Wf DQG UHVXOWV LQ DQ LQFUHDVH LQ WKH H[WUDRUGLQDU\ LQGH[ $QH ZKLFK LV b KLJKHU WKDQ ZKDW LV DWWDLQDEOH ZLWK EHQ]RLF DFLG >$KOF .DQ@ ,Q DGGLWLRQ S\URSKRVn SKRULF DFLG GRHV QRW ERLO DW W\SLFDO H[FKDQJH WHPSHUDWXUHV LW LV OLTXLG XS WR r&f DQG KDV D ORZ YDSRU SUHVVXUH DOORZLQJ ZRUN DW KLJK WHPSHUDWXUHV ZKHUH GLIIXVLRQ LV UDSLG 7KH KLJKHU + FRQFHQWUDWLRQ DOVR DOORZV IRU PRUH FRQWURO RYHU WKH GLIIXVLRQ GHSWK DQG XQLIRUPLW\ RI WKH H[FKDQJHG OD\HU WKDQ LV SRVVLEOH ZLWK EHQ]RLF DFLG >0L]@ +RZHYHU ZKHQ XVLQJ S\URSKRVSKRULF DFLG 7D PXVW EH XVHG DV WKH PDVN VLQFH WKH S\URSKRVSKRULF DFLG ZLOO DWWDFN DQG FDXVH SLWWLQJ LQ DQ $ PDVN 7KH $3( SURFHVV LQ /L7D UHVXOWV LQ DQ LQFUHDVH LQ WKH H[WUDRUGLQDU\ LQGH[ RQO\ DOORZLQJ IRU WKH SURSDJDWLRQ RI VROHO\ 70 PRGHV LQ D ]FXW FU\VWDO DV LOOXVWUDWHG LQ )LJ )RU WKH FDVH RI [ DQG \FXW FU\VWDOV RQO\ 7( PRGHV ZLOO SURSDJDWH $GGLWLRQDO IHDn WXUHV FKDUDFWHULVWLF RI WKH $3( SURFHVV LQ /L7D DUH GHVFULEHG LQ WKH QH[W VHFWLRQ

PAGE 30

QF 70 !‘ 7( )LJ ,QFUHDVH RI QH RQO\ GXH WR SURWRQ H[FKDQJH FDXVHV WKH SURSDJDWLRQ RI VROHO\ 70 PRGHV LQ D ]FXW FU\VWDO )HDWXUHV RI WKH $3( 3URFHVV LQ /L7D4 $V VWDWHG HDUOLHU WKH $3( SURFHVV RI ZDYHJXLGH IDEULFDWLRQ KROGV PDQ\ DGYDQWDJHV RYHU RWKHU PHWKRGV VXFK DV VLPSOLFLW\ DQG KLJK LQGH[ LQFUHPHQW 6SHFLILFDOO\ LQ FRPSDUn LVRQ ZLWK WKH PHWKRG RI 7L RU =Q LQGLIIXVLRQ WKH $3( SURFHVV VXSSRUWV RQO\ RQH SRODUL]Dn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ƒKOF ƒKO@ 7KLV LV DGYDQWDJHRXV

PAGE 31

LQ DSSOLFDWLRQV ZKHUH D VLQJOH SRODUL]DWLRQ LV GHVLUDEOH VLQFH RQO\ OLJKW SRODUL]HG LQ WKH GLUHFWLRQ RI QH ZLOO EH SURSDJDWHG ,Q $3(/L7D WKHUH DUH DW OHDVW IRXU PHFKDQLVPV FRQWULEXWLQJ WR WKH WRWDO LQGH[ FKDQJH $Q >$KO@ ZKLFK WKXV FDQ EH H[SUHVVHG DV $Q $Q3 $Q0 $Q9 $QH f 7KH $Q) WHUP LV FDXVHG E\ D FKDQJH LQ WKH VSRQWDQHRXV SRODUL]DWLRQ 3V YLD WKH .HUU HOHFn WURRSWLF HIIHFW $IWHU VXEVWLWXWLRQ RI + IRU /L GXULQJ SURWRQ H[FKDQJH + LV NQRZQ WR RFFXS\ SRVLWLRQV ZLWKLQ WKH R[\JHQ SODQHV UDWKHU WKDQ WKH /L VLWHV UHGXFLQJ WKH FRQWULEXn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ƒKOF 'DY@ GH n! ZKHUH 'H7Hf LV WKH WHPSHUDWXUH GHSHQGHQW GLIIXVLRQ FRHIILFLHQW IRU WKH SURWRQ H[FKDQJH SURFHVV DQG 7H LV WKH H[FKDQJH WHPSHUDWXUH 7KLV GLIIXVLRQ FRHIILFLHQW IROORZV WKH $UUKHn QLXV ODZ >ƒKOF 'DY@ 9N%7Hf $U $RH[3 f

PAGE 32

ZKHUH (H LV WKH DFWLYDWLRQ HQHUJ\ IRU WKH SURWRQ H[FKDQJH SURFHVV 'H LV WKH H[FKDQJH GLIn IXVLRQ FRQVWDQW DQG N% LV %ROW]PDQQfV FRQVWDQW 6LPLODUO\ IRU WKH DQQHDOLQJ SURFHVV DW WHPSHUDWXUH 7D DQG GXUDWLRQ WD WKH IROORZLQJ DQQHDOLQJ GLIIXVLRQ FRHIILFLHQW 'D7Df DQG LQFUHDVH LQ GHSWK GD DUH IRXQG >'DY@ f f ZKHUH (D LV WKH DFWLYDWLRQ HQHUJ\ IRU WKH DQQHDOLQJ SURFHVV 7KH WRWDO GHSWK G7 RI WKH ZDYHJXLGH IDEULFDWHG E\ WKH $3( SURFHVV FDQ WKHQ EH DSSUR[LPDWHG E\ >'DY@ GUM Gf Gf H D f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f WKHQ FDXVHV IXUWKHU GLIIXVLRQ RI + VPRRWKLQJ RXW WKH LQGH[ SURILOH DQG H[SDQGLQJ WKH ZDYHJXLGH DUHD ZKLOH GHFUHDVLQJ WKH SURSDJDWLRQ ORVV >0L]@ 6LQFH WKHUH LV QR SURWRQ VRXUFH DW WKH VXUIDFH WKH ORFDO FRQFHQWUDWLRQ ZLWKLQ WKH H[FKDQJHG OD\HU PD\ GHFUHDVH WR

PAGE 33

WKH YDOXH FRUUHVSRQGLQJ WR WKH SKDVH WUDQVLWLRQ $QQHDOLQJ DOVR VHUYHV WR UHFRYHU WKH HOHFWURRSWLF U DQG QRQOLQHDU G FRHIILFLHQWV ZLWKLQ WKH H[FKDQJHG UHJLRQ NQRZQ WR KDYH EHHQ VHULRXVO\ GHJUDGHG GXULQJ WKH H[FKDQJH SURFHVV >ƒKO ƒKOD $KOF &DR 0DWE@ (YHU\ SKDVH WUDQVLWLRQ UHVXOWV LQ D FKDQJH LQ WKH UHIUDFWLYH LQGH[ 7KH YDOXH RI SURn WRQ FRQFHQWUDWLRQ DW ZKLFK D SKDVH WUDQVLWLRQ RFFXUV LV GHWHUPLQHG E\ WKH VSHFLILF PDWHULDO DV LV WKH GHSHQGHQFH RI UHIUDFWLYH LQGH[ RQ SURWRQ FRQFHQWUDWLRQ EHWZHHQ WKH SKDVH WUDQVLn WLRQV ,Q RWKHU ZRUGV WKH GHSHQGHQFH RI UHIUDFWLYH LQGH[ RQ SURWRQ FRQFHQWUDWLRQ LV VSHn FLILF WR WKH PDWHULDO DQG LV LGHQWLILHG DV WKH VWUXFWXUDO SKDVH GLDJUDP RI WKLV PDWHULDO $V ZLOO EH VKRZQ EHORZ NQRZOHGJH RI WKH VWUXFWXUDO SKDVH GLDJUDP LV FUXFLDO LQ XQGHUVWDQGn LQJ WKH SURSHUWLHV DQG DQRPDOLHV RI $3( ZDYHJXLGHV $QRPDOLHV RI $3( :DYHJXLGHV 0DQ\ DQRPDOLHV H[LVW ZKLFK DUH FKDUDFWHULVWLF RI WKH $3( SURFHVV LQ /L7D 7KRXJK VRPH UHVHDUFK RQ WKHVH SHFXOLDULWLHV KDV EHHQ GRQH WKH\ DUH QRW \HW IXOO\ XQGHUn VWRRG )LUVW WKH H[WUDRUGLQDU\ LQGH[ LQFUHDVH $QH WKDW FDQ EH DFKLHYHG E\ $3( LQ /L7D LV DV PXFK DV DQ RUGHU RI PDJQLWXGH OHVV WKDQ WKDW LQ /L1E >ƒKOF 0DWE $KO@ ZKLFK LV UHODWHG WR LQGH[ LQFUHPHQW WKURXJK (T f $V IRU WKH PDJQLWXGH RI WKH RUGLQDU\ LQGH[ GHFUHDVH $Q LQ $3( /L7D LW KDV EHHQ IRXQG WR EH PDQ\ WLPHV ODUJHU WKDQ WKH H[WUDRUGLQDU\ LQGH[ LQFUHDVH $QH >ƒKOF 0DUE@ DV RSSRVHG WR EHLQJ RI QHDUO\ WKH VDPH YDOXH RU VPDOOHU IRU WKH FDVH RI $3( LQ /L1E >ƒ:@

PAGE 34

6HFRQG LV WKH DQRPDORXV EHKDYLRU RI ZDYHJXLGH SURILOHV XSRQ DQQHDOLQJ ,W KDV EHHQ GHPRQVWUDWHG WKDW WKH UHIUDFWLYH LQGH[ LQFUHPHQW LQFUHDVHV GXULQJ VKRUW DQQHDOLQJ WLPHV IRU WKH $3( SURFHVV LQ /L7D YDVWO\ LQFUHDVLQJ WKH DUHD XQGHU WKH LQGH[ SURILOH >ƒKOF 0DUE 0DWE $KOF@ 7KHVH SHFXOLDULWLHV VXJJHVW D QRQOLQHDU GHSHQGHQFH RI WKH UHIUDFWLYH LQGH[ LQFUHPHQW RQ + FRQFHQWUDWLRQ DQG WKH SRVVLELOLW\ RI IRUPLQJ ZDYHJXLGHV ZLWK EXULHG UHIUDFWLYH LQGH[ SURILOHV >$KOF ƒKO 'DY 0DUD@ 6XFK SURILOHV KDYH UHFHQWO\ EHHQ FRQILUPHG E\ GLUHFW PHDVXUHn PHQW >0DUE@ XVLQJ D UHIOHFWLYLW\ SURILOLQJ WHFKQLTXH WR EH GHVFULEHG LQ &KDSWHU DQG DUH H[HPSOLILHG LQ )LJ 7KLV LV LQ GLUHFW RSSRVLWLRQ WR WKH FDVH RI $3( LQ /L1E ZKHUH WKH LQGH[ LQFUHPHQW LV NQRZQ WR GHFUHDVH XSRQ DQQHDOLQJ PDLQWDLQLQJ D PRUH RU OHVV OLQHDU GHSHQGHQFH RQ + FRQn FHQWUDWLRQ >0F:@ 7KH WKLUG DQG PRVW VLJQLILn FDQW SHFXOLDULW\ LV WKDW $3( ZDYHJXLGHV LQ /L7D KDYH EHHQ IRXQG WR H[KLELW VLJQLILFDQW WHPSRn UDO LQVWDELOLWLHV DSSHDULQJ LQ WKH )LJ 5HIUDFWLYHLQGH[ SURILOHV IRU WKH H[WUDRUGLQDU\ SRODUL]DWLRQf IRU YDULn RXV DQQHDOLQJ WLPHV PHDVXUHG DW A _LP UHIUDFWLYH LQGH[ LQFUHPHQW IRU ERWK VKRUW DQG ORQJ SHULRGV DIWHU IDEULFDn WLRQ >ƒKOF ƒKO 0DU22

PAGE 35

0DWD@ )RU FHUWDLQ IDEULFDWLRQ FRQGLWLRQV LQ WKH LQLWLDO KRXUV DIWHU IDEULFDWLRQ D ODUJH GHFUHDVH LQ $QH KDV EHHQ VKRZQ WR RFFXU IROORZHG E\ D PXFK VPDOOHU JUDGXDO GHFUHDVH LQ $QH RYHU WKH QH[W IHZ PRQWKV >ƒKO@ 7KH WRWDO GHFUHDVH LQ UHIUDFWLYH LQGH[ DIWHU VHYHUDO PRQWKV FDQ EH RI WKH RUGHU RI n >0DU22 0DWD@ 3UHOLPLQDU\ PHDVXUHPHQWV KDYH GHPRQVWUDWHG ODUJH WHPSRUDO FKDQJHV RQ WKH VDPH RUGHU DV LOOXVWUDWHG LQ )LJ 7KHVH FKDQJHV FDQ EH D VHULRXV SUREOHP IRU GHYLFHV ZKLFK DUH H[WUHPHO\ VHQVLWLYH WR LQGH[ YDULDWLRQV DQG UHTXLUH ORQJWHUP VWDELOLW\ VXFK DV GLUHFWLRQDOFRXSOHUV DQG 430 ZDYHJXLGHV XVHG IRU IUHTXHQF\ GRXEOLQJ WR D VSHFLILHG ZDYHOHQJWKf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f PHDVXUHG DW ; OP LPPHGLDWHO\ DIWHU IDEULFDWLRQ DQG ZHHNV ODWHU /L7D DUH SUHVHQWHG

PAGE 36

$3( :DYHJXLGH )DEULFDWLRQ 3DUDPHWHUV 8VLQJ WKH $3( SURFHVV FKDQQHO ZDYHJXLGHV ZHUH IRUPHG RQ WKH ; VXUIDFH RI ; FXW /L7D DQG WKH Fn VXUIDFH RI =FXW /L7D ZLWK WKH DLP RI VLQJOH PRGH RSHUDWLRQ DW D ZDYHOHQJWK RI SP ,Q WKLV FRQWH[W Fa DQG F ZLOO EH XVHG WR UHIHU WR WKH QHJDWLYH DQG SRVLWLYH VXUIDFHV RI D =FXW FU\VWDO UHVSHFWLYHO\ :H GHIHU WKH FKDUDFWHUL]DWLRQ RI WKHVH ZDYHJXLGHV XQWLO FKDSWHU IRXU EXW SUHVHQW KHUH WKH H[DFW SDUDPHWHUV XVHG LQ WKHLU IDEULFDn WLRQ $GGLWLRQDOO\ WKH IDEULFDWLRQ SDUDPHWHUV RI SODQDU ZDYHJXLGHV LQ ;FXW /L7D DUH SUHVHQWHG 7KHVH ZDYHJXLGHV ZLOO EH XVHG IRU UHIOHFWLYLW\ SURILOLQJ DQG URFNLQJ FXUYH DQDO\VLV WR EH GHVFULEHG ODWHU LQ FKDSWHU IRXU 7KH F DQG Fa VXUIDFHV RI UDZ =FXW /L7D FU\VWDO ZDIHUV FDQ EH LGHQWLILHG E\ WDNn LQJ DGYDQWDJH RI WKH SLH]RHOHFWULF HIIHFW RI WKH IHUURHOHFWULF VXEVWUDWHV )LJ GHPRQn VWUDWHV KRZ D YROWPHWHU ZDV XVHG WR TXLFNO\ SUHVV GRZQ DQG UHOHDVH RQ WKH VXUIDFH RI WKH VXEVWUDWH %HFDXVH RI WKH LQWHUQDO ILHOG VHW XS E\ WKH VSRQWDQHRXV SRODUL]DWLRQ RI WKH FU\Vn WDO IUHH FKDUJHV H[LVW RQ WKH VXUIDFHV WR FRXQWHUDFW WKLV ILHOG DQG PDLQWDLQ HOHFWULFDO QHXn WUDOLW\ )RU H[DPSOH WKH F DQG F VXUIDFHV KDYH QHJDWLYH DQG SRVLWLYH IUHH FKDUJHV UHVSHFWLYHO\ %\ UDSLGO\ SUHVVLQJ RQ WKH FU\VWDO DQG UHOHDVLQJ WKH LQWHUQDO ILHOG GXH WR SRODUL]DWLRQ LV PRPHQWDULO\ UHGXFHG DQG WKH YROWPHWHU FDQ UHDG WKH SRWHQWLDO GXH WR WKH QRZ XQEDODQFHG VXUIDFH FKDUJHV VR WKH LQGLYLGXDO SRVLWLYH DQG QHJDWLYH VXUIDFHV FDQ EH LGHQWLILHG 7KHVH =FXW ZDIHUV ZHUH QH[W GLFHG LQWR UHFWDQJXODU VDPSOHV RI YDULRXV GLPHQn VLRQV GHSHQGLQJ RQ WKHLU DSSOLFDWLRQ ZKHUH \SURSDJDWLQJ FKDQQHO ZDYHJXLGHV FRXOG EH IRUPHG 7KHVH VDPSOHV ZHUH WKHQ DQQHDOHG LQ DLU DW r& IRU KRXU WR UHOLHYH VWUHVVHV LQFXUUHG GXULQJ GLFLQJ

PAGE 37

9ROWPHWHU VXEVWUDWH $O IRLO SRV PHWHU FKDQJH F VXUIDFH XS QHJ PHWHU FKDQJH Fn VXUIDFH XS )LJ 0HWKRG XVHG WR LGHQWLI\ WKH LQGLYLGXDO VXUIDFHV RI D ]FXW /L7D ZDIHU 6DPSOHV LQWHQGHG IRU FKDQQHO ZDYHJXLGHV ZHUH VXEMHFWHG WR WKH VWDQGDUG SKRWRn OLWKRJUDSKLF SURFHVV $IWHU FOHDQLQJ LQ WULFKORURHWKDQH DFHWRQH DQG PHWKDQRO WKH VDPn SOHV ZHUH GULHG DW r& IRU PLQXWHV +RHFKVW &HODQHVH $= 3KRWRUHVLVW ZDV VSXQ RQ WKH Fn VXUIDFHV DW USP IRU VHFRQGV DIWHU ZKLFK WKH VDPSOHV ZHUH EDNHG DW r& IRU PLQXWHV 7KLV UHVXOWHG LQ D SKRWRUHVLVW WKLFNQHVV RI DERXW SP (DFK VXEVWUDWH ZDV WKHQ FRYHUHG ZLWK D GLJLWDOO\ GDUN SKRWRPDVN FRQWDLQLQJ GHVLUHG SDWWHUQV EH WKH\ LQWHUn IHURPHWHUV GLUHFWLRQDO FRXSOHUV RU MXVW VWUDLJKW VWULSHV UDQJLQJ LQ ZLGWK IURP ILP WR

PAGE 38

O2ILP LQ VWHSV RI _LP 6DPSOHV ZHUH DOLJQHG XQGHU WKH SKRWRPDVN ZLWK WKH PDVN DOLJQHU VR WKDW WKH SURSDJDWLRQ GLUHFWLRQ RI WKH ZDYHJXLGH SDWWHUQ WR EH XVHG UDQ SDUDOOHO ZLWK WKH \GLUHFWLRQ RI WKH VDPSOH 8SRQ D VHFRQG H[SRVXUH WR XOWUDYLROHW 89f OLJKW WKH VDPSOHV ZHUH KHOG LQ +RHFKVW &HODQHVH $= GHYHORSHU GLOXWHG f IRU VHFn RQGV ZKHUHE\ WKH DUHDV RI SKRWRUHVLVW QRW FRYHUHG E\ WKH PDVN JXLGH SDWWHUQV DQG KHQFH H[SRVHG WR WKH 89 ZHUH UHPRYHG $ 7D PDVN ZDV VSXWWHUHG RQ DW D WKLFNQHVV RI DERXW ƒ XVLQJ VSXWWHULQJ SDUDPHWHUV RI 9 IRU DERXW PLQXWHV DQG DQ $U IORZ UDWH RI VHHP $FHWRQH ZDV WKHQ XVHG WR OLIW RII WKH UHPDLQLQJ SKRWRUHVLVW OHDYLQJ WKH VDPSOH UHDG\ IRU SURWRQ H[FKDQJH DV GHSLFWHG LQ )LJ 7KH HQWLUH SKRWROLWKRJUDSKLF SURFHVV IURP WKH VSLQQLQJ RI D SKRWRUHVLVW ILOP WKURXJK OLIWRII LV LOOXVWUDWHG LQ )LJ 3URWRQ H[FKDQJH ZDV SHUIRUPHG XVLQJ WZR GLIIHUHQW VRXUFHV S\URSKRVSKRULF DFLG DQG JO\FHULQ 7KH H[FKDQJH ZDV SHUIRUPHG LQ D ER[ IXUQDFH DW r& IRU YDULRXV WLPHV $ FXVWRP PDGH DFLG ZDVK EHDNHU ZDV XVHG WR DOORZ WKH VRXUFH DQG VDPSOH WR KHDW XS VHSn DUDWHO\ ZLWKLQ WKH IXUQDFH $IWHU VWDELOL]LQJ DW WKH GHVLUHG H[FKDQJH WHPSHUDWXUH WKH VDPn SOH ZDV LPPHUVHG LQWR WKH VRXUFH ZLWKRXW KDYLQJ WR RSHQ WKH IXUQDFH GRRU DQG FUHDWH WHPSHUDWXUH YDULDWLRQV $IWHU H[FKDQJLQJ WKH VDPSOH ZDV UHPRYHG IURP WKH VRXUFH DQG DOORZHG WR FRRO WR URRP WHPSHUDWXUH VORZO\ LQ RUGHU WR UHGXFH WKHUPDO VWUDLQ (DFK VDPn SOH ZDV H[FKDQJHG LQGHSHQGHQWO\ DQG QHZ VRXUFH ZDV XVHG HDFK WLPH 2QFH FRROHG WKH VDPSOHV ZHUH ZDVKHG LQ GHLRQL]HG ',f ZDWHU WR UHPRYH WKH UHPDLQLQJ VRXUFH DQG WKH 7D PDVN ZDV HWFKHG RII XVLQJ D 7D HWFKDQW FRQVLVWLQJ RI K\GURIn OXRULF DQG QLWULF DFLGV 7KH 3( ZDYHJXLGHV H[FKDQJHG XVLQJ S\URSKRVSKRULF DFLG ZHUH WKHQ DQQHDOHG LQ DLU XVLQJ D WXEH IXUQDFH DW YDULRXV WLPHV DQG WHPSHUDWXUHV

PAGE 39

GHSRVLW 35 H[SRVH WR 89 DQG GHYHORS GHSRVLW 7D OLIWRII )LJ ,OOXVWUDWLRQ RI WKH SKRWROLWKRJUDSKLF SURFHVV XS WKURXJK OLIWRII

PAGE 40

3ODQDU ZDYHJXLGHV ZHUH DOVR IDEULFDWHG LQ ;FXW /L7D XVLQJ WKH DERYH H[FKDQJH FRQGLWLRQ %HLQJ SODQDU KRZHYHU WKH\ REYLRXVO\ GLG QRW QHHG WR JR WKURXJK WKH SKRWRn OLWKRJUDSKLF SURFHVV DQG WKHLU DQQHDOLQJ WLPHV YDULHG IURP DVH[FKDQJHGf WR PLQn XWHV 7KH UHDVRQ IRU WKHVH YDULHG DQQHDOLQJ WLPHV ZLOO EHFRPH PRUH REYLRXV LQ &KDSWHU ,W LV QRWHG KHUH WKDW H[FKDQJH WLPHV ORQJHU WKDQ PLQXWHV UHVXOWHG LQ VXUIDFH FUDFNV LQ ; FXW FU\VWDOV 7KLV ZDV PRVW OLNHO\ GXH WR WKH GLIIHUHQFH LQ H[SDQVLRQ RI WKH ODWWLFH SDUDPHn WHUV DORQJ WKH ; DQG =GLUHFWLRQV DV QRWHG HOVHZKHUH >0DWE@ ,Q WKH QH[W VHFWLRQ WKH RWKHU PHWKRGV RI ZDYHJXLGH IDEULFDWLRQ ZLOO EH GHVFULEHG QDPHO\ 7L PHWDOLQGLIIXVLRQ DQG =Q YDSRULQGLIIXVLRQ 7KH FKDUDFWHULVWLFV RI ZDYHJXLGHV IDEULFDWHG E\ WKHVH SURFHVVHV ZLOO EH GHVFULEHG DQG WKH FRQGLWLRQV IRU WKHLU IDEULFDWLRQ ZLOO EH JLYHQ :DYHJXLGH )RUPDWLRQ E\ 0HWDO9DSRU,QGLIIXVLRQ $V KDV EHHQ PHQWLRQHG $3( ZDYHJXLGHV LQ /L7D VXIIHU IURP WHPSRUDO LQVWDELOn LWLHV ,Q DGGLWLRQ WKH LQWURGXFWLRQ RI + LQWR WKH FU\VWDO YLD SURWRQ H[FKDQJH LV NQRZQ WR FDXVH TXHQFKLQJ ZLWKLQ D UDUHHDUWKGRSHG UHJLRQ >1RX@ LQ /L1E UHGXFLQJ WKH H[FLWHG VWDWH OLIHWLPH DQG WKHUHIRUH GHFUHDVLQJ ODVHU SHUIRUPDQFH $V GRSLQJ ZLWK UDUH HDUWK (U IRU ODVHU IDEULFDWLRQ LV D SRWHQWLDO DSSOLFDWLRQ IRU /L7D DOVR D PHDQV RI ZDYHJXLGH IDEULFDWLRQ RWKHU WKDQ $3( QHHGV EH H[SORUHG 6LQFH PHWDOLQGLIIXVLRQ LQ /L1E LV NQRZQ WR EH WHPSRUDOO\ VWDEOH DQ H[DPLQDWLRQ RI WKLV SURFHVV LV D JRRG SODFH WR VWDUW IRU /L7D $V VXFK ZDYHJXLGH IDEULFDWLRQ LQ /L7D E\ 7L PHWDOLQGLIIXVLRQ DQG =Q YDSRULQGLIIXVLRQ ZHUH H[DPLQHG 7KH NQRZQ FKDUDFWHULVWLFV DQG IDEULFDWLRQ SDUDPHn WHUV IRU HDFK RI WKHVH SURFHVVHV DUH SUHVHQWHG KHUH

PAGE 41

7L 0HWDO,QGLIIXVLRQ :DYHJXLGH IDEULFDWLRQ E\ 7L LQGLIIXVLRQ LV SHUIRUPHG E\ HYDSRUDWLQJ D 7L OD\HU SRVVLEO\ SDWWHUQHG E\ SKRWROLWKRJUDSK\ RQWR WKH VXUIDFH RI D VXEVWUDWH DQG VXEVHTXHQWO\ LQGLIIXVLQJ WKH PHWDO DW KLJK WHPSHUDWXUHV !r&f 6LQFH LWV GLVFRYHU\ LQ >6FK@ PXFK ZRUN KDV EHHQ GRQH RQ WKH FKDUDFWHUL]DWLRQ DQG IDEULFDWLRQ RI 7L LQGLIIXVHG ZDYHJXLGHV DQG GHYLFHV HVSHFLDOO\ LQ /L1E 7KHUH LV D OLPLWHG DPRXQW RI GDWD UHSRUWHG LQ OLWHUDWXUH DERXW WKH 7L LQGLIIXVLRQ SURFHVV LQ /L7D 0RVW RI LW LV IURP WKH GHYLFH SHUn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f DQG f >%XU@ +RZHYHU WKH GLIIXVLRQ FRHIILFLHQW IRU 7L LQ /L7D LV DERXW RUGHUV RI PDJQLWXGH VPDOOHU WKDQ WKDW IRU7L LQ /L1E >%XU@ 8SRQ LQGLIIXVLRQ 7L LRQV WDNH SRVLWLRQV DW 1E 7Df VLWHV LQ WKH /L1E /L7Df FU\VWDO ODWWLFH >,16@ 7KH PHFKDQLVP RI LQGH[ LQFUHDVH LV WKRXJKW WR EH DWWULEXWHG WR DQ LQFUHDVH RI WKH SRODULVDELOLW\ DQG WR WKH SKRWRHODVWLF HIIHFW FDXVHG E\ WKH GLIIHUHQW LRQLF UDGLL RI 7L DQG 1E 7Df LRQV >,16@

PAGE 42

)RU WKH SURFHVV RI 7L LQGLIIXVLRQ WKHUH DUH QR UHSRUWHG DQRPDOLHV VLPLODU WR WKRVH RI $3( 7L ZDYHJXLGHV LQ /L1E DSSHDU WR EH WHPSRUDOO\ VWDEOH WKRXJK VWDELOLW\ KDV QHYHU EHHQ FRQILUPHG IRU WKH VDPH SURFHVV LQ /L7D ,Q DGGLWLRQ WKLV SURFHVV GRHV QRW FDXVH D FKDQJH LQ WKH FU\VWDO ODWWLFH VWUDLQ VR WKHUH LV QR VWUXFWXUDO SKDVH GLDJUDP ZLWK PXOWLSOH SKDVHV VLPLODU WR WKDW RI WKH $3( SURFHVV +RZHYHU WKH VXUIDFH RI VDPSOHV KDYH EHHQ VKRZQ WR H[KLELW VZHOOLQJ LQ WKH UHJLRQV RI LQGLIIXVLRQ 7KH WKLFNQHVV RI WKLV VZHOOn LQJ FDQ EH DV PXFK DV WLPHV WKH WKLFNQHVV RI WKH RULJLQDO 7L PHWDO OD\HU EHIRUH LQGLIIXVLRQ >+RO@ $GGLWLRQDOO\ 7L/L1E ZDYHJXLGHV DUH NQRZQ WR KDYH D JUHDWHU VXVFHSWLELOLW\ WR WKH SKRWRUHIUDFWLYH HIIHFW WKDQ $3( ZDYHJXLGHV LQ /L1E >)XM@ 'XH WR WKH KLJK WHPSHUDWXUHV LQYROYHG LQ WKLV LQGLIIXVLRQ SURFHVV /L RXWGLIIX VLRQ RFFXUV >:RR@ 7KLV LV D SUREOHP EHFDXVH LW FDXVHV D VPDOO LQFUHDVH RI WKH H[WUDRUn GLQDU\ LQGH[ HYHU\ZKHUH RQ WKH VXUIDFH RI WKH VDPSOH GHFUHDVLQJ WKH LQGH[ GLIIHUHQFH EHWZHHQ WKH ZDYHJXLGH DQG LWV VXUURXQGLQJV $OVR GXH WR WKH KLJK WHPSHUDWXUH ZKHQ 7L LQGLIIXVLRQ LV SHUIRUPHG LQ /L7D FU\VWDO UHSROLQJ LV QHFHVVDU\ DV WKH FU\VWDO LV KHDWHG DERYH LWV &XULH SRLQW GXULQJ WKH SURFHVV 7KH PHWKRG RI FU\VWDO UHSROLQJ LV GHVFULEHG LQ 6HFWLRQ 7KH IDEULFDWLRQ FRQGLWLRQV IRU 7L ZDYHJXLGHV LQ /L7D DUH DV IROORZV 3KRWROLn WKRJUDSK\ LV GRQH RQ WKH F VXUIDFH RI D /L7D VDPSOH WR GHILQH D SDWWHUQ XVLQJ D GLJLWDOO\ FOHDU PDVN DV RSSRVHG WR D GLJLWDOO\ GDUN PDVN XVHG )LJ 'HSLFWLRQ RI D 7L SDWWHUQ IRU $3( 7KHQ ƒ RI 7L LV HEHDP GHSRVLWHG RQ /L7D EHIRUH LQGLIn IXVLRQ RQWR WKH VDPSOH DIWHU ZKLFK OLIWRII LQ DFHWRQH

PAGE 43

WDNHV SODFH 7KH UHVXOW LV D VDPSOH ZLWK D ZDYHJXLGH SDWWHUQ RI 7L PHWDO RQ WKH VXUIDFH DV GHSLFWHG LQ )LJ IRU WKH FDVH RI D VWUDLJKWFKDQQHO ZDYHJXLGH 7KH VDPSOH LV SODFHG LQVLGH D SODWLQXP ER[ ZLWK SRZHUHG /L1 DQG WKH ER[ LV ORDGHG LQVLGH D WXEH IXUQDFH 'U\ LV IORZHG WKURXJK WKH WXEH RI WKH IXUQDFH DW FFP ZKLOH WKH VDPSOH LV LQGLIIXVHG DW r& IRU KRXUV ZLWK DQ rPLQ ULVH DQG IDOO UDPS UDWH 7KH SXUSRVH RI WKH SRZn HUHG /L1 DQG IORZLQJ LV WR KHOS UHGXFH /L RXWGLIIXVLRQ IURP WKH VDPSOH VXUIDFH 7KHVH FRQGLWLRQV UHVXOW LQ VLQJOHPRGH ZDYHJXLGHV IRU ERWK SRODUL]DWLRQV DW MLP LPPHGLDWHO\ DIWHU IDEULFDWLRQ =Q 9DSRU,QGLIIXVLRQ =Q YDSRULQGLIIXVLRQ LV SHUIRUPHG E\ SODFLQJ D VDPSOH LQVLGH D YDFXXP VHDOHG TXDUW] DPSXOH ZLWK VRPH DPRXQW RI =Q ZLUH JUDQXOHV RU SRZGHU DQG KHDWLQJ WKH DVVRUWn PHQW WR D WHPSHUDWXUH DERYH WKH ERLOLQJ SRLQW IRU =Q VR WKDW LW YDSRUL]HV DQG GLIIXVHV LQWR WKH VDPSOH 7KH =Q FRQFHQWUDWLRQ SURILOH REWDLQHG UHVHPEOHV D FRPSOHPHQWDU\ HUURU IXQFn WLRQ HUIF\Gf GLVWULEXWLRQ ZKHUH G LV WKH GHSWK DW ZKLFK WKH FRQFHQWUDWLRQ LV b RI WKDW DW WKH VXUIDFH >(NQ@ $V ZLWK 7L LQGLIIXVLRQ WKLV SURFHVV UHVXOWV LQ DQ LQFUHDVH RI ERWK WKH H[WUDRUGLQDU\ DQG RUGLQDU\ LQGLFHV ><@ /LNHZLVH WKH FKDQJH LQ WKH H[WUDRUGLQDU\ LQGH[ LV ODUJHU WKDQ WKDW RI WKH RUGLQDU\ LQGH[ ><@ $JDLQ WKH GLIIXVLRQ HTXDWLRQV JRYHUQLQJ GHSWK DQG GLIIXVLRQ FRHIILFLHQWV DUH VLPLODU WR WKRVH XVHG IRU SURWRQ H[FKDQJH DQG 7L LQGLIIXVLRQ QDPHO\ (TV f DQG f ><@ 6LPLODU WR 7L LQGLIIXVLRQ =Q YDSRULQGLIIXVLRQ GRHV QRW DSSHDU WR VKRZ DQ\ RI WKH VDPH DQRPDOLHV DV $3( DQG WKHUH LV QR PHDVXUHG FKDQJH RI FU\VWDO ODWWLFH VWUDLQ LQGXFHG E\ WKLV SURFHVV +RZHYHU WKH VWDELOLW\ RI ZDYHJXLGHV IDEULFDWHG LQ /L7D E\ WKLV SURFHVV KDYH QRW \HW EHHQ FRQILUPHG $GYDQWDJHRXVO\ WKH ORZHU WHPSHUDWXUH RI =QLQGLIIXVLRQ

PAGE 44

ar&f UHVXOWV LQ QHJOLJLEOH /L RXWGLIIXVLRQ IURP WKH VDPSOH ><@ WKRXJK WKH WHPSHUDWXUH LV VWLOO KLJK HQRXJK WR UHTXLUH VDPSOH UHSROLQJ &XULRXVO\ VDPSOHV DIWHU LQGLIIXVLRQ DSSHDU VOLJKWO\ JUD\LVKEURZQ LQ FRORU 7KLV GLVFRORUDWLRQ LV OLNHO\ GXH WR GHILFLHQF\ >.DP@ DQG LV HDVLO\ UHPRYHG E\ D VKRUW DQQHDOLQJ DW DERXW r& LQ DLU 7R IDEULFDWH =Q YDSRULQGLIIXVHG ZDYHJXLGHV SKRWROLWKRJUDSK\ ZDV GRQH RQ WKH F VXUIDFH RI /L7D VDPSOHV XVLQJ D GLJLWDOO\ GDUN PDVN WR GHILQH D SDWWHUQ VLPLODU WR $3( +RZHYHU $ RI 6L ZDV HEHDP GHSRVLWHG DW D UDWH RI $V WR VHUYH DV WKH PDVN LQVWHDG RI 7D DV LQ $3( 1H[W WKH VDPSOH ZDV VXEMHFWHG WR WKH OLIWRII SURFHVV LQ DFHWRQH DQG ZDV YDFXXP VHDOHG LQ D TXDUW] DPSXOH ZLWK SRZHUHG =Q 7KH DPSXOH ZDV ORDGHG LQWR WKH WXEH IXUQDFH DQG LQGLIIXVHG DW r& IRU KRXUV ZLWK DQ %RU@ EXW LV JHQHUDOO\ DERXW r& $V D UHVXOW WKH FU\VWDO JRHV IURP WKH IHUURHOHFWULF VWDWH WR WKH SDUDHOHFWULF VWDWH ZKHUH LW LV

PAGE 45

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n SOHV DV =FXW VDPSOHV ZHUH IRXQG WR VXIIHU IURP VXUIDFH GDPDJH UHODWHG WR WKH DSSOLFDWLRQ RI HOHFWURGHV GLUHFWO\ RQWR WKH VDPSOH VXUIDFH >-XQ@ )RU WKLV ZRUN WKH SROLQJ SURFHVV ZDV PRGLILHG WR EH XVHIXO IRU SROLQJ =FXW /L7D VDPSOHV )LUVW WKH VDPSOH WR EH SROHG KDG ƒ RI 6L GHSRVLWHG RQ WKH F DQG F VXUIDFHV WR SURWHFW WKHP IURP GDPDJH GRQH E\ GLUHFW FRQWDFW RI WKH HOHFWURGHV 1H[W WKH VDPSOH ZDV VDQGZLFKHG EHWZHHQ WZR VKHHWV RI SODWLQXP IRLO ZLWK FRSSHU ZLUHV @ HOHFWURGHVn  k # i i i i 4 )LJ ,OOXVWUDWLRQ RI WKH FU\VWDO UHSROLQJ SURFHVV

PAGE 46

DWWDFKHG DQG SODFHG LQVLGH WKH ER[ IXUQDFH ZLWK WKH ZLUH OHDGV UXQQLQJ RXW WKH YHQW DW WKH WRS RI WKH IXUQDFH 7KH ZLUH OHDGV ZHUH FRQQHFWHG WR D '& SRZHU VXSSO\ 7KH VDPSOH ZDV KHDWHG WR r& DQG D YROWDJH RI 9 ZDV DSSOLHG 7KH RYHQ ZDV LPPHGLDWHO\ WXUQHG RII DQG DOORZHG WR FRRO 2QFH LW KDG FRROHG WR r& WKH YROWDJH ZDV UHPRYHG DQG WKH VDPn SOH DOORZHG WR FRQWLQXH FRROLQJ WR URRP WHPSHUDWXUH $V ZLOO EH VHHQ ODWHU LQ &KDSWHU WKLV SROLQJ SURFHVV ZDV VXFFHVVIXO DQG UHVXOWHG LQ FRPSOHWH UHFRYHU\ RI WKH HOHFWURRSWLF FRHIILFLHQW U ,Q WKH QH[W VHFWLRQ WKH WHFKQLTXH RI HQGIDFH SROLVKLQJ LV GHVFULEHG $ JRRG HQG IDFH SROLVK LQ QHFHVVDU\ IRU DFFXUDWH RSWLFDO PHDVXUHPHQWV DQG ORZ LQVHUWLRQ ORVV (QG)DFH 3ROLVKLQJ RI &KDQQHO :DYHJXLGHV ,Q RUGHU WR PLQLPL]H FRXSOLQJ ORVVHV LQWR WKH FKDQQHO ZDYHJXLGHV JRRG TXDOLW\ HQGIDFHV RI WKH VDPSOH SHUSHQGLFXODU WR WKH ZDYHJXLGHV DUH UHTXLUHG $ SROLVKLQJ VFKHPH LV RXWOLQHG KHUH ZKLFK DFKLHYHV VDWLVIDFWRU\ UHVXOWV R )LUVW DERXW $ RI 6L LV HEHDP GHSRVLWHG RQ WKH ZDYHJXLGH VXUIDFH RI WKH VDPSOH WR EH SROLVKHG WR SURWHFW WKLV VXUIDFH IURP VFUDWFKHV GXULQJ WKH PRXQWLQJ DQG SROLVKLQJ SURFHVV 7KHQ WKH VDPSOH LV DGKHUHG WR D 6L ZDIHU XVLQJ KLJK WHPSHUDWXUH 4XLFN6WLFNr1 FOHDU ZD[ PHOWLQJ SRLQW RI r&f ZLWK WKH ZDYHJXLGH VXUIDFH SUHVVHG DJDLQVW WKH 6L ,Q WKLV PDQQHU D WKLQ aOSP RU OHVVf OD\HU RI ZD[ FDQ EH IRUPHG EHWZHHQ WKH VDPSOH DQG WKH 6L ZDIHU ZKLFK VHUYHV WR PLQLPL]H FUDFNLQJ DQG FKLSSLQJ RI WKH VDPn SOH HQGIDFH RQ WKH ZDYHJXLGH VXUIDFH 7KH 6L ZDIHU LV WKHQ FOHDYHG DORQJ DOO IRXU HGJHV RI WKH VDPSOH DQG SROLVKHG E\ KDQG LQ VLOLFRQ FDUELGH 6L&f SRZGHU ZLWKRXW XVLQJ D SROLVKn LQJ MLJ XQWLO WKH 6L LV IOXVK ZLWK WKH /L7D RQ DOO VLGHV 7KLV LV GRQH WR UHGXFH FUDFNLQJ

PAGE 47

DQG IODNLQJ RI WKH 6L GXULQJ SROLVKLQJ ZLWK D SROLVKLQJ MLJ 7KH VWUXFWXUH LV WKHQ PRXQWHG WR D SROLVKLQJ MLJ XVLQJ ORZ WHPSHUDWXUH EHHfV ZD[ %HFDXVH RI WKH ORZHU PHOWLQJ WHPSHUn DWXUH RI WKH EHHfV ZD[ ar&f WKHUH LV QR GDQJHU RI PHOWLQJ WKH WKLQ OD\HU RI FOHDU ZD[ EHWZHHQ WKH 6L DQG WKH /L7D VDPSOH 7KH HQGIDFH RI WKH VDPSOH LV ILUVW ODSSHG E\ KDQG LQ 6L& SRZGHU 1H[W WKH VDPSOH LV SROLVKHG E\ KDQG LQ _LP DOXPLQXP R[LGH $f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n WLRQ RI WKH GHSWK SURILOH LV HQKDQFHG E\ EHYHOOLQJ D SODQDU VDPSOH DW DQ DQJOH RI D r 8VLQJ WKLV WHFKQLTXH WKH SURILOHV LQ WKH GHSWK GLUHFWLRQ DUH PDJQLILHG E\ VLQD ,Q DGGLn WLRQ DQ H[WUHPHO\ ILQH SROLVK LV UHTXLUHG VR DV QRW WR LQWURGXFH D FRQVLGHUDEOH DPRXQW RI QRLVH LQ WKH UHIOHFWHG VLJQDO IURP WKH EHYHOHG VXUIDFH 7KH GHWDLOV RI WKH SROLVKLQJ VFKHPH XVHG WR DFKLHYH D ILQH EHYHO SROLVK DUH JLYHQ KHUH

PAGE 48

)LUVW D 7D ILOP RI WKLFNQHVV DURXQG ƒ LV GHSRVLWHG RQ WKH ZDYHJXLGH VXUIDFH 7KLV VHUYHV WR LGHQWLI\ WKH VXEVWUDWH VXUIDFH GXULQJ VFDQV RI WKH GHSWK SURILOH E\ LQWURGXFn LQJ D UDSLG VSLNH LQ UHIOHFWHG OLJKW LQWHQVLW\ 7KH VDPSOH LV WKHQ PRXQWHG WR D 6L ZDIHU ZDYHJXLGH VXUIDFH DJDLQVW WKH 6L LQ WKH VDPH PDQQHU DV RXWOLQHG LQ WKH SUHYLRXV VXEVHFn WLRQ 7KH 6L KHUH KHOSV WR SUHYHQW URXQGLQJ GXH WR SROLVKLQJ DW WKH VXEVWUDWH VXUIDFH 1RZ KRZHYHU WKH VWUXFWXUH LV PRXQWHG RQWR D SROLVKLQJ MLJ ZKLFK KDV D UDPSOLNH JURYH PLOOHG LQ LW DW DQ DQJOH RI D r DV LOOXVWUDWHG LQ )LJ 3ROLVKLQJ EHJLQV E\ KDQG LQ 6L& SRZGHU XQWLO WKH VXUIDFH RI WKH /L7D LV UHDFKHG $OXPLQXP R[LGH SRZGHU LV XVHG QH[W IROORZHG E\ SP DQG SP GLDPRQG SDVWHV VDPH DV EHIRUH $V D ILQDO VWHS 8OWUD6RO $ &ROORLGDO $OXPLQD 3ROLVKLQJ SROLVKHG VXUIDFH )LJ 'HSLFWLRQ RI WKH SROLVKLQJ MLJ XVHG WR EHYHO SROLVK SODQDU ZDYHJXLGHV DW DQ DQJOH D r 7KLV PDJQLILHV WKH GHSWK SURILOHV E\ D IDFWRU RI VLQD

PAGE 49

&RPSRXQG LV XVHG ZLWK D JHQHURXV DPRXQW RI VRGLXP K\GUR[LGH 1D2+f GLVVROYHG LQ ', ZDWHU RQ D %XHKOHU 0LFURFORWK $IWHU SROLVKLQJ IRU VL[ PLQXWHV WKH VDPSOH LV UHPRYHG IURP WKH 0LFURFORWK ZKHHO DQG TXLFNO\ VXEPHUJHG LQ WKH 1D2+ VROXWLRQ WR SUHYHQW D JODVV\ ILOP IURP IRUPLQJ RQ WKH /L7D VXUIDFH ,Q WKH QH[W VHFWLRQ WKH WKHRU\ DQG IDEULFDWLRQ FRQGLWLRQV DVVRFLDWHG ZLWK WKH DSSOLn FDWLRQ RI DQ DQWLUHIOHFWLYLW\ FRDWLQJ DUH GHVFULEHG ,Q PDQ\ LQVWDQFHV VXFK D FRDWLQJ ZDV QHFHVVDU\ WR PLQLPL]H UHIOHFWLRQV DQG UHGXFH PHDVXUHPHQW QRLVH $QWLUHIOHFWLYH &RDWLQJV )RU WKH PHDVXUHPHQWV RI &KDSWHU ZKHUH OLJKW LV UHIOHFWHG IURP WKH EHYHOHG VXUn IDFH RI D SODQDU ZDYHJXLGH LW ZDV IRXQG WKDW UHIOHFWLRQV IURP WKH EDFN VXUIDFH RI WKH VDPn SOH VXEVWUDWH ZHUH FRQWULEXWLQJ FRQVLGHUDEOH QRLVH LQWR WKH H[SHULPHQWDO VHWXS ,Q DGGLWLRQ IRU WKH ORVV PHDVXUHPHQWV RI &KDSWHU DQG WKH VWDELOLW\ PHDVXUHPHQWV RI &KDSn WHU UHIOHFWLRQV IURP WKH HQGIDFHV RI ZDYHJXLGH VDPSOHV ZHUH DOVR LQWURGXFLQJ VRPH QRLVH DQG HUURU LQWR WKH VHWXS )RU HDFK RI WKHVH FDVHV PXFK RI WKH LPSDFW RI WKHVH UHIOHFn WLRQV ZDV HOLPLQDWHG WKURXJK WKH XVH RI D VLQJOHOD\HU DQWLUHIOHFWLYH $5f FRDWLQJ RQ HLWKHU WKH EDFN VLGH RU WKH HQGIDFH RI WKH VXEVWUDWH DV WKH FDVH PD\ EH 6RPH RI WKH WKHRU\ XVHG LQ WKH GHYHORSPHQW RI WKHVH $5 FRDWLQJV LV SUHVHQWHG LQ WKLV VHFWLRQ DV ZHOO DV WKH PDWHn ULDO DQG WKLFNQHVV QHHGHG IRU WKH ZDYHOHQJWK XVHG DQG WKH PHWKRG RI LWV GHSRVLWLRQ 7KLQ )LOP 7KHRU\ IRU 6LQJOH/DYHU $5 &RDWLQJV 7KH GHYHORSPHQW KHUH WKRXJK RQO\ FRQFHUQHG ZLWK D VLQJOHOD\HU ILOP VWHPV IURP WKH GHULYDWLRQ RI WKH WUDQVIHU PDWUL[ PHWKRG RI PXOWLSOH WKLQ ILOP OD\HUV WUHDWHG LQ >+XV

PAGE 50

3HG@ 7KH ILOP LV DVVXPHG WR EH KRPRJHQRXV DQG LVRWURSLF 7KH ILOP LV DOVR DVVXPHG WR EH WKLQ HQRXJK VR WKDW WKH SDWK GLIIHUHQFH EHWZHHQ PXOWLSO\ UHIOHFWHG DQG WUDQVPLWWHG EHDPV ZLWKLQ WKH ILOP UHJLRQ LV VPDOOHU WKDQ WKH FRKHUHQFH OHQJWK RI WKH OLJKW 7KHUHIRUH WKH UHIOHFWHG DQG WUDQVPLWWHG EHDPV FDQ FRPSOHWHO\ LQWHUIHUH :H EHJLQ E\ UHIHUULQJ WR )LJ 7KH FRQGLWLRQ VKRZQ LV WKDW IRU 7( SRODUL]Dn WLRQ ZKLFK LV WKH RQO\ FDVH ZH DUH FRQFHUQHG ZLWK KHUH /DWHU WKH VWLSXODWLRQ RI QRUPDO LQFLGHQFH ZLOO EH PDGH VR WKDW WKHUH LV QR GLVWLQFWLRQ EHWZHHQ WKH 7( DQG 70 SRODUL]Dn WLRQV $GGLWLRQDOO\ LW LV QRWHG WKDW D FRQYHQWLRQ ZDV FKRVHQ ZKHUHE\ DV QRUPDO LQFLGHQFH LV DSSURDFKHG WKH LQFLGHQW DQG UHIOHFWHG HOHFWULF ILHOG YHFWRUV ( DUH RULHQWHG LQ WKH VDPH GLUHFWLRQ ZKLOH WKH PDJQHWLF ILHOG YHFWRUV % VZLWFK GLUHFWLRQ 7KRXJK WKLV ZLOO LQWURGXFH D SKDVH VKLIW LQ WKH % ILHOG DW r FRQYHQWLRQDO RSWLFV GHILQHV WKH GLUHFWLRQ RI SRODUL]Dn WLRQ DV WKDW RI WKH (ILHOG GLUHFWLRQ DQG KHQFH WKH RULHQWDWLRQV DUH DFFHSWDEOH 7KH TXDQWLn WLHV ( ZLWK VXEVFULSWV GHILQH WKH PDJQLWXGHV RI WKH HOHFWULF ILHOGV DW HDFK ERXQGDU\ Df DQG Ef )RU H[DPSOH (U@ LV WKH VXP RI DOO PXOWLSO\ UHIOHFWHG EHDPV DW LQWHUIDFH Df ZKLFK DUH H[LWLQJ WKH ILOP WRZDUG WKH UHJLRQ ZLWK LQGH[ Q (L LV WKH VXP RI DOO EHDPV LQFLGHQW DW LQWHUIDFH Ef GLUHFWHG WRZDUG WKH VXEVWUDWH UHJLRQ ZLWK LQGH[ QV DQG VR RQ %\ DSSO\LQJ WKH ERXQGDU\ FRQGLWLRQ WKDW WKH WDQJHQWLDO FRPSRQHQWV RI WKH ( DQG % ILHOGV EH FRQWLQXRXV DFURVV HDFK LQWHUIDFH ZH JHW f (E a (L (U a (W f % QFRV% W&2V* % FRVI % FRV* D Z L U L W WY LL W f %E %L&2V4Q%U&RVGWO %W&2VHW f

PAGE 51

)LJ 'LDJUDP RI WKH UHIOHFWLRQ IURP D VLQJOHOD\HU WKLQ ILOP XVHG WR FDOFXn ODWH WKH WRWDO UHIOHFWLYLW\ 5 RI WKH ILOP

PAGE 52

1RZ E\ XVLQJ % 9 &( Q-:R( f \ QR-VRARFRV%L f 9I n\9eRFRVHL f \V QV-HR+FRVOL f ZH FDQ ZULWH WKH %ILHOGV DV %D a
PAGE 53

(D (EFRV %E^M
PAGE 54

(TV f DQG f FDQ EH VROYHG WR JLYH WKH UHIOHFWLRQ FRHIILFLHQW LQ WHUPV RI WKH WUDQVn IHU PDWUL[ HOHPHQWV WR \LHOG :fOO:WP9 :fOO:0 P<6P )LQDOO\ WKH UHIOHFWDQFH 5 LV GHILQHG E\ f 5 ?U? f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f ZKLFK IRUFHV WKH SKDVH GLIIHUHQFH RI (T f WR EHFRPH W VR WKDW FRV DQG VLQ 8VLQJ WKHVH FRQGLWLRQV WKH HTXDWLRQ IRU UHIOHFWDQFH 5 (T f EHFRPHV ? QQVa Q I QQV QI 5 f

PAGE 55

)RU WKH FDVH RI /L7D LQ DLU QV DQG Q ,Q RUGHU WR KDYH QR UHIOHFWLRQV 5 ZH PXVW XVH D ILOP PDWHULDO ZKRVH LQGH[ QI QVf a :H KDYH FKRVHQ PDJQHVLXP IOXRULGH 0J)f ZLWK DQ LQGH[ RI DV D VXLWDEOH PDWHULDO )RU WKH ZDYHOHQJWKV RI LQWHUHVW QDPHO\ QP DQG QP WKH UHVXOWLQJ ILOP WKLFNQHVVHV DUH ƒ DQG R $ UHVSHFWLYHO\ 7KH $5 ILOPV RI 0J) DUH HDVLO\ GHSRVLWHG XVLQJ DQ HOHFWURQEHDP HEHDPf GHSRn VLWLRQ V\VWHP 7KH GHWDLOV RI VXFK D V\VWHP FDQ EH IRXQG LQ >*HR 0DF@ ,Q WKH QH[W VHFWLRQ WKH FRQGLWLRQV XVHG IRU WKH HOHFWURSODWLQJ RI WKLFN JROG HOHFn WURGHV DUH JLYHQ 6XFK WKLFN HOHFWURGHV DUH XVHG WR PLQLPL]H UI ORVV LQ KLJKVSHHG WUDYHOn LQJZDYH PRGXODWRUV VXFK DV WKH RQH WR EH GHPRQVWUDWHG LQ &KDSWHU (OHFWURSODWLQJ )RU WKH GHPRQVWUDWLRQ RI D KLJKVSHHG WUDYHOLQJZDYH HOHFWURRSWLF PRGXODWRU WR EH GRQH ODWHU LQ &KDSWHU WKLFN JROG HOHFWURSODWHG HOHFWURGHV ZHUH QHHGHG ,Q WKLV VHFn WLRQ WKH FRQGLWLRQV XVHG IRU WKH IDEULFDWLRQ DQG HOHFWURSODWLQJ RI WKHVH HOHFWURGHV DUH JLYHQ 7R EHJLQ ZLWK D EXIIHU OD\HU RI 6L $ WKLFN ZDV HEHDP GHSRVLWHG DW D UDWH RI $V )RU D =FXW FU\VWDO WKH EXIIHU OD\HU VKRXOG EH GHSRVLWHG XQLIRUPO\ RYHU WKH HQWLUH ZDYHJXLGH VXUIDFH )RU DQ ;FXW FU\VWDO WKH EXIIHU OD\HU QHHG RQO\ EH DW HDFK HQG RXWVLGH WKH HOHFWURGH LQWHUDFWLRQ OHQJWK ZKHUH WKH HOHFWURGHV EHQG DQG FURVV WKH ZDYHJXLGHV 1H[W ƒ RI &U ZDV GHSRVLWHG IROORZHG E\ ƒ RI $X 7KH $X LV WKH VHHG OD\HU IRU HOHFWURSODWLQJ DQG WKH &U LV WR KHOS WKH JROG VWLFN WR WKH VDPSOH 1RWH WKDW

PAGE 56

WKH HEHDP EHOO MDU PXVW QRW EH RSHQHG EHWZHHQ WKH GHSRVLWLRQ RI &U DQG $X DV WKH &U ZLOO R[LGL]H DQG DV D UHVXOW WKH $X ZLOO QRW VWLFN WR LW 3KRWROLWKRJUDSK\ RI WKH HOHFWURGH SDWWHUQ ZDV SHUIRUPHG QH[W )LUVW WKH VDPSOH ZDV VXEMHFWHG WR PLQXWHV RI YDSRUGHSRVLWLRQ RI SKRWRUHVLVW DGKHVLRQ SURPRWHU 1H[W +RHFKVW &HODQHVH $= SKRWRUHVLVW ZDV VSXQ RQ DW USP IRU VHFRQGV 7KH UHVXOWLQJ SKRWRUHVLVW WKLFNQHVV ZDV DURXQG _LP 7KH VDPSOH ZDV WKHQ EDNHG DW r& IRU PLQXWHV 7KH HGJH EHDG ZDV WKHQ UHPRYHG DQG LW ZDV EDNHG DJDLQ DW r& IRU PLQn XWHV $IWHU FRROLQJ WKH VDPSOH ZDV DOLJQHG XQGHU WKH HOHFWURGH SDWWHUQ DQG H[SRVHG DW P:FP IRU PLQXWHV ,W ZDV WKHQ GHYHORSHG LQ +RHFKVW &HODQHVH $=. GHYHOn RSHU GLOXWHG ZLWK ', ZDWHU IRU PLQXWHV DIWHU ZKLFK LW ZDV VXEMHFWHG WR D EDNLQJ VHTXHQFH RI r& IRU PLQXWHV r& IRU PLQXWHV r& IRU PLQXWHV )LQDOO\ WKH )LJ 6HWXS IRU WKH HOHFWURSODWLQJ SURFHVV

PAGE 57

VDPSOH ZDV ORDGHG LQWR WKH UHDFWLYH LRQ HWFKHU 5,(f DQG HWFKHG LQ SODVPD IRU PLQn XWHV P7RUU VHHP :f WR UHPRYH DQ\ UHPDLQLQJ SKRWRUHVLVW IURP WKH H[SRVHG DUHDV 7KH WKLFN JROG OD\HU LV QH[W HOHFWURSODWHG WR WKH VDPSOH 7KH VHWXS IRU WKLV LV GHSLFWHG LQ )LJ 7KH VDPSOH ZDV DWWDFKHG WR WKH QHJDWLYH OHDG RI D '& SRZHU VXSSO\ ZLWK WKH SRVLWLYH OHDG EHLQJ DWWDFKHG WR D JUDWH %RWK ZHUH VXEPHUJHG LQWR D EHDNHU FRQn WDLQLQJ 2URWHPS JROG VROXWLRQ IURP 7HFKQLF ,QF $ 4 UHVLVWRU DQG DPPHWHU ZHUH SXW LQ WKH FLUFXLW WR PRQLWRU FXUUHQW IORZ 7KH EHDNHU ZDV SODFHG LQ D WHPSHUDWXUHFRQn WUROOHG ZDWHU UHFLUFXODWRU DQG WKH WHPSHUDWXUH VHW WR r& $ FXUUHQW RI P$ SHU FP RI VDPSOH VXUIDFH ZDV WKHQ DSSOLHG IRU PLQXWHV DIWHU ZKLFK WKH WHPSHUDWXUH DQG FXUUHQW ZHUH ERWK LQFUHDVHG 7KH FRPSOHWH SURFHVV LV DV IROORZV r& P$FP PLQXWHV r& ,P$FP PLQXWHV r& P$FP PLQXWHV r& P$FP PLQXWHV 7KH UHVXOWLQJ $X WKLFNQHVV ZDV DURXQG SP $IWHU HOHFWURSODWLQJ OLIWRII ZDV SHUIRUPHG XVLQJ DFHWRQH DQG D VROXWLRQ RI +6+ *ROG HWFKDQW GLOXWHG ZLWK ', ZDWHU ZDV XVHG QH[W WR UHPRYH WKH JROG VHHG OD\HU )LQDOO\ WKH VDPSOH ZDV ORDGHG LQWR WKH 5,( DQG WKH &U OD\HU ZDV UHPRYHG XVLQJ D &O SODVPD IRU PLQXWHV P7RUU VFFP :f 6XPPDU\ ,Q WKLV FKDSWHU WKH IDEULFDWLRQ RI SODQDU DQG FKDQQHO ZDYHJXLGHV LQ /L7D ZDV SUHVHQWHG $ IHZ PHWKRGV IRU IDEULFDWLQJ WKH ZDYHJXLGHV ZHUH JLYHQ ZLWK WKH PDLQ IRFXV

PAGE 58

EHLQJ RQ WKH FRPPRQO\ HPSOR\HG $3( WHFKQLTXH 6HYHUDO GLVWLQFW IHDWXUHV RI WKH $3( SURFHVV LQ /L7D ZHUH GLVFXVVHG DQG D IHZ RI WKH FXUUHQWO\ XQUHVROYHG DQRPDOLHV DVVRFLn DWHG ZLWK WKLV SURFHVV ZHUH GHVFULEHG 7KH H[DFW IDEULFDWLRQ SDUDPHWHUV IRU WKH SODQDU DQG FKDQQHO ZDYHJXLGHV XVHG LQ WKLV UHVHDUFK ZHUH QH[W SUHVHQWHG IROORZHG E\ VRPH DGGLn WLRQDO SURFHVVLQJ VWHSV VXFK DV FU\VWDO UHSROLQJ HQGIDFH DQG EHYHO SROLVKLQJ WKH DSSOLFDn WLRQ RI DQ $5 FRDWLQJ DQG HOHFWURSODWLQJ ,Q WKH QH[W FKDSWHU D FKDUDFWHUL]DWLRQ RI (U GRSLQJ LQ /L7D LV SHUIRUPHG

PAGE 59

&+$37(5 &+$5$&7(5,67,&6 2) (U'23(' /L7D %HFDXVH RI WKH GHPDQG IRU LQWHJUDWHG RSWLFDO FRPSRQHQWV IRU XVH LQ FRPPXQLFDn WLRQ QHWZRUNV DV RXWOLQHG LQ &KDSWHU WKHUH LV D GHFLVLYH QHHG IRU ODVHUV DQG DPSOLILHUV WR RSHUDWH LQ WKHVH V\VWHPV :KLOH PXFK ZRUN KDV EHHQ GRQH RQ WKH GHYHORSPHQW RI ILEHU DPSOLILHUV SDUWLFXODUO\ (UGRSHG ILEHU DPSOLILHUV (')$fVf >$UD 'HO@ HIIRUWV FRQn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

PAGE 60

ODVHUV DQG DPSOLILHUV +RZHYHU PXFK UHVHDUFK DZDLWV WR EH SHUIRUPHG RQ WKH FKDUDFWHUL]DWLRQ DQG HIIHFW RI UDUHHDUWK GRSLQJ SDUWLFXODUO\ (U LQ /L7D ,Q WKLV FKDSWHU WKH HIIHFW RI (UGRSLQJ LQ /L7D LV LQYHVWLJDWHG 6RPH EDFNn JURXQG PDWHULDO UHODWLQJ WR ZRUN GRQH LQ WKLV FRQQHFWLRQ RQ /L1E LV SUHVHQWHG DV ZHOO DV WKH OLPLWDWLRQV RI /L1E OHDGLQJ WR WKH PRWLYDWLRQ WR H[DPLQH (UGRSHG /L7D 7KH FRQGLWLRQV XVHG IRU GRSLQJ RI /L7D DUH WKHQ JLYHQ IROORZHG E\ 5DPDQ DQG IOXRUHVFHQFH PHDVXUHPHQWV RI VDPSOHV LGHQWLI\LQJ (U FOXVWHULQJ DV WKH GRPLQDQW JDLQ OLPLWLQJ IDFWRU $ QRYHO PHWKRG RI UHGXFLQJ FOXVWHULQJ LV WKHQ SUHVHQWHG DQG DSSOLHG )LQDOO\ D WZRSKRn WRQ PRGHO LV XVHG WR H[HPSOLI\ WKH SKRWRUHIUDFWLYH SURFHVV LQ WKHVH FU\VWDOV DQG WKH LQKHUn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n VLWLRQ RI (U LRQV DQG DUH XVHG IRU WKH SXPSLQJ RI (UGRSHG GHYLFHV 7KH\ DUH QP DQG QP )RU WKH FDVH RI ZDYHJXLGH ODVHUV LQ /L1E KRZHYHU SXPSLQJ DW QP KDV VRPH DGYDQWDJHV )LUVW WKH SKRWRUHIUDFWLYH HIIHFW LV VPDOOHU DW QP WKDQ DW QP DOORZLQJ IRU WKH DSSOLFDWLRQ RI KLJKHU SXPS SRZHUV 6HFRQG ZDYHJXLGHV IDEULFDWHG WR EH

PAGE 61

VLQJOHPRGH QHDU SP DUH JHQHUDOO\ PXOWLPRGH DW QP JLYLQJ ULVH WR WKH SUREOHP RI VHOHFWLYH PRGH H[FLWDWLRQ >%DX@ :KHUHDV ZKHQ SXPSLQJ DW QP WKH PRGH VL]HV RI WKH SXPS DQG VLJQDO DUH QHDUO\ WKH VDPH DQG VR RYHUODS LV EHWWHU $V PHQWLRQHG DERYH DQ LQKHUHQW SUREOHP ZLWK JXLGHGZDYH GHYLFHV LQ /L1E LV WKHLU ORZ UHVLVWDQFH WR SKRWRUHIUDFWLYH 35f GDPDJH DW YLVLEOH DQG QHDULQIUDUHG ZDYHn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n JUDWLRQ RI DQ DGGLWLRQDO HOHFWURRSWLF FRPSRQHQW VXFK D PRGXODWRU RQ WKH VDPH FKLS DV WKH DGGLWLRQDO FRPSRQHQW ZRXOG QRW EHQHILW IURP WKH ODUJH HOHFWURRSWLF FRHIILFLHQW LQ WKH =GLUHFWLRQ ,Q OLJKW RI WKH GHVFULEHG OLPLWDWLRQV RI /L1E /L7D DSSHDUV DQ DWWUDFWLYH DOWHU QDWLYH ZLWK D 35 GDPDJH WKUHVKROG RI :FP DW QP DQG URRP WHPSHUDWXUH LH DERXW IRUW\ WLPHV KLJKHU WKDQ WKH :FP YDOXH RI /L1E >*OD@ 6LPLODU WR /L1E (U QHHGV WR EH LQGLIIXVHG LQWR /L7D WR VHUYH DV WKH DFWLYH JDLQ PHGLXP IRU DFWLYH LQWHn JUDWHG GHYLFHV VXFK DV DPSOLILHUV DQG ODVHUV +RZHYHU (UGRSLQJ LV NQRZQ WR OHDG WR XSFRQYHUVLRQ LQ PRVW KRVW PDWHULDOV>$UD 'HO *@ ZKLFK LV ZLGHO\ UHFRJQL]HG DV D JDLQ OLPLWLQJ IDFWRU LQ (UGRSHG DPSOLILHUV E\ LQGXFLQJ 35 GDPDJH >$UD 'HO

PAGE 62

+XD@ 7KHUH DUH WZR PHFKDQLVPV RI UHFRQYHUVLRQ 7KH ILUVW LV WKH KRPRJHQHRXV XSFRQYHUVLRQ FDXVHG E\ H[FLWHGVWDWH DEVRUSWLRQ (6$f ZLWKLQ QRQFOXVWHUHG (U LRQV 7KLV SURFHVV RFFXUV ZKHUHYHU WKHUH DUH (U LRQV DQG WKHUHIRUH LV LQKHUHQW WR (UGRSHG ZDYHJXLGHV 7KH VHFRQG PHFKDQLVP LV QRQUDGLDWLYH HQHUJ\ WUDQVIHU XSFRQYHUVLRQ (78f ZKLFK RFFXUV LQ (U FOXVWHU VLWHV 7KLV IRUP RI XSFRQYHUVLRQ FDQ EH VXSSUHVVHG E\ GHFUHDVLQJ WKH UHODWLYH FRQFHQWUDWLRQ RI FOXVWHU VLWHV WKURXJK D SURSHU FKRLFH RI GRSLQJ WHFKQRORJ\ >$UD 'HO@ DQGRU SRVWLQGLIIXVLRQ SURFHVVLQJ >'LH *LO@ )RU H[DPn SOH LQ WKH FDVH RI (U/L1E WKH FOXVWHU VLWH FRQFHQWUDWLRQ GHSHQGV RQ WKH /L1E UDWLR >*LO@ DQG WKHUHIRUH XSFRQYHUVLRQ HIILFLHQF\ PD\ EH UHGXFHG VLJQLILFDQWO\ E\ XVLQJ D YDSRU SKDVH HTXLOLEUDWLRQ WHFKQLTXH WR LQFUHDVH WKH /L FRQWHQW >'LH *LO@ 7KH VLPLn ODULW\ RI LQWULQVLF GHIHFW VWUXFWXUHV LQ /L1E DQG /L7D >*RS@ ZRXOG WKHUHIRUH VXJn JHVW WKDW DQ LQFUHDVH RI WKH /L7D UDWLR VKRXOG DOVR UHGXFH XSFRQYHUVLRQ FDXVHG E\ (78 LQ (U/L7D $V D UHVXOW WKH HIIHFW RI (UGRSLQJ RQ WKH SK\VLFDO DQG RSWLFDO SURSHUWLHV RI /L7D QHHGV WR EH H[DPLQHG ,Q SDUWLFXODU KRZ GRSLQJ DIIHFWV WKH 35 SURSHUW\ RI WKH FU\VWDO DQG KRZ PXFK XSFRQYHUWHG HPLVVLRQ IURP (U FOXVWHUV ZKLFK UHGXFH JDLQ >*LO@ LV SUHVHQW DV ZHOO DV PHWKRGV IRU LWV UHGXFWLRQ 7KLV VWXG\ EHJLQV LQ WKH QH[W VHFn WLRQ E\ RXWOLQLQJ WKH FRQGLWLRQV XVHG IRU (UGRSLQJ LQ /L7D (U'RSLQJ &RQGLWLRQV 7KH IROORZLQJ IDFWRUV QHHG EH FRQVLGHUHG ZKHQ (UGRSLQJ ZLWK WKH DLP RI SURGXFn LQJ DFWLYH JDLQ UHJLRQV IRU ZDYHJXLGH ODVHUV )LUVW (U FRQFHQWUDWLRQ VKRXOG EH DW DQ DSSURSULDWH OHYHO VR DV WR SURYLGH VLJQLILFDQW JDLQ 1H[W (U FRQFHQWUDWLRQ SURILOHV

PAGE 63

VKRXOG EH DW OHDVW DV GHHS DV ZDYHJXLGH LQGH[ SURILOHV VR WKDW WKH RSWLFDO PRGH KDV JRRG RYHUODS ZLWK WKH JDLQ PHGLXP )RU WKLV VWXG\ H[SHULPHQWDO =FXW /L7D VXEVWUDWHV ZHUH GRSHG QHDU WKH VXUIDFH E\ LQGLIIXVLRQ RI D QP WKLFN (U OD\HU DW r& IRU KRXUV 7KHVH IDEULFDn WLRQ SDUDPHWHUV ZHUH FKRVHQ VR DV WR SURYLGH DQ (U FRQFHQWUDWLRQ SURILOH DSSURSULDWH IRU DFWLYH JXLGHGZDYH GHYLFHV 6RPH RI WKH (UGRSHG /L7D VDPSOHV ZHUH SURFHVVHG LQ PROWHQ /L1 DW r& IRU WLPHV UDQJLQJ IURP WR KRXUV 6XFK D WUHDWPHQW ZDV UHSRUWHG >.RVD@ WR EH HIIHFWLYH WRZDUG LQFUHDVLQJ WKH /L FRQWHQW LQ D QHDUVXUIDFH OD\HU ZLWK GHSWK _LPf RI /L1E ,Q RUGHU WR VWXG\ WKH LPSDFW RI (UGRSLQJ RQ WKH SKRWRUHIUDFWLYH SURSHUWLHV RI WKH FU\VWDO DQG WR EH DEOH WR FKHFN IRU XSFRQYHUVLRQ HPLVVLRQV 5DPDQ DQG IOXRUHVFHQFH PHDn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

PAGE 64

5D\OHLJK 6FDWWHULQJf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f 'H'Pf f ZKHUH ; LV WKH ZDYHOHQJWK LQ QP 'P LV WKH PHDVXUHG ZDYHQXPEHU LQ FPn DQG 8F LV WKH ZDYHQXPEHU RI WKH H[FLWDWLRQ VRXUFH LQ FPn HJ FPn IRU D QP VRXUFH DQG FPn IRU D QP VRXUFHf )OXRUHVFHQFH 0HDVXUHPHQWV 7R H[DPLQH WKH LQIOXHQFH RI XSFRQYHUWHG OLJKW RQ WKH 35 HIIHFW LQ (U/L7D WKH IOXRUHVFHQFH VSHFWUXP RI HDFK (UGRSHG VDPSOH ZDV ILUVW PHDVXUHG XVLQJ D 5HQLVKDZ 5DPDVFRSH VSHFWURPHWHU RSHUDWLQJ DW H[FLWDWLRQ ZDYHOHQJWKV RI DQG QP

PAGE 65

,Q WKH VHWXS D OLQHDUO\ SRODUL]HG ODVHU EHDP ZDV IRFXVHG E\ D ; REMHFWLYH WR D SP VSRW RQ WKH RSWLFDOJUDGH SROLVKHG HGJH RI WKH VDPSOH XQGHU VWXG\ 7KH SODQH RI OLJKW SRODUL]DWLRQ ZDV SDUDOOHO WR WKH RSWLFDO D[LV $ 5DPDQ PLFURSUREH LH D PLFURVFRSH RSWLn FDOO\ FRXSOHG WR D 5DPDQ VSHFWURPHWHU ZDV XVHG WR FROOHFW ERWK WKH RXWRISODQH VFDWWHUHG OLJKW DQG WKH IOXRUHVFHQFH HPLVVLRQ 7KHUH ZDV D VLJQLILFDQW GLIIHUHQFH EHWZHHQ WKH IOXRUHVFHQFH VSHFWUXP SURGXFHG E\ SXPSLQJ DW QP DQG WKDW SURGXFHG E\ SXPSLQJ DW QP RU QP )RU QP RQO\ LQWHQVLYH 6WRNHV IOXRUHVFHQFH 6)f ZDV LQGXFHG LH IOXRUHVFHQFH ZLWK SKRWRQ HQHUn JLHV VPDOOHU WKDQ WKH H[FLWDWLRQ HQHUJ\f )RU QP DQG QP KRZHYHU WKHUH ZDV DQWL6WRNHV IOXRUHVFHQFH $6)f ZLWK ERWK D JUHHQ HPLVVLRQ DW QP VHH )LJ f RG & F /! 2 F 2 &' R 4* ZDYHOHQJWK QPf )LJ $QWL6WRNHV IOXRUHVFHQFH VSHFWUD H[FLWHG DW QP DQG QP LQ WKH (U/L7D VDPSOH ZLWK WKH KLJKHVW (U FRQFHQWUDWLRQ PRObf

PAGE 66

)LJ )OXRUHVFHQFH VSHFWUD RI WKH UHG $6) HPLVVLRQ IRU WKH (UGRSHG VDPSOH RI )LJ VROLGf DQG 5DPDQ VSHFWUD RI YLUJLQ /L7D GDVKHGf H[FLWHG DW ; QP DQG [O :FP 5 GHQRWHV 5DPDQ OLQHV DQG / GHQRWHV WKH H[FLWDWLRQ ODVHU OLQH DIWHU D QRWFK ILOWHU )LJ ,QWHQVLW\ RI WKH JUHHQ $6) HPLVVLRQ YHUVXV LQSXW LQWHQVLW\ IRU H[FLWDWLRQ DW QP FLUFOHVf DQG QP VTXDUHVf IRU WKH VDPSOH VKRZQ LQ )LJ

PAGE 67

DQG D UHG HPLVVLRQ DW QP )LJ f LQ DGGLWLRQ WR WKH 6) :LWKLQ WKH PHDVXUHPHQW UDQJH RI WKH LQSXW LQWHQVLW\ WKH 6) LQWHQVLW\ ,6) FKDQJHG SURSRUWLRQDOO\ WR DW DOO H[FLWDn WLRQ ZDYHOHQJWKV LQ DFFRUGDQFH ZLWK SUHYLRXV VWXGLHV >$UD 'HO *LO@ 7KH GHSHQn GHQFH RI WKH $6) LQWHQVLW\ ,$6) RI WKH JUHHQ HPLVVLRQ ZDV IRXQG WR EH PRUH LQYROYHG DQG KDG D NQHH DW D FHUWDLQ YDOXH -W RI ZLWK -W UDQJLQJ IURP WR r :FP IRU QP SXPSLQJ DQG DERXW } :FP IRU QP SXPSLQJ $V LOOXVWUDWHG LQ )LJ ,$6I-f LV D QRQOLQHDU IXQFWLRQ ,AVIA-r IRU --W ZLWK [ RI DERXW IRU QP SXPSLQJ DQG DERXW IRU QP SXPSLQJ )RU -!-W WKHUH LV D OLQHDU FRUUHODWLRQ EHWZHHQ ,$6) DQG LH [ 7KH GHSHQGHQFH RI ,$6) RI WKH UHG HPLVVLRQ KDG WKH VDPH FKDUDFWHU DV WKDW RI WKH JUHHQ HPLVVLRQ LQ )LJ EXW WKH LQWHQVLW\ RI WKH DQWL6WRNHV UHG HPLVVLRQ ZDV VPDOOHU WKDQ WKDW RI WKH JUHHQ HPLVVLRQ E\ DERXW WLPHV 6XFK DQ DSSUR[LPDWHO\ TXDn GUDWLF LQSXWLQWHQVLW\ ODZ REVHUYHG DW D ORZ SXPS UDQJH r :FPf VXJJHVWV WKDW WKH JUHHQ DQG UHG $6) HPLVVLRQV DUH DWWULEXWHG WR D WZRSKRWRQ DEVRUSWLRQ SURFHVV WR EH GHVFULEHG LQ 6HFWLRQ 5HGXFWLRQ IURP D TXDGUDWLF WR D OLQHDU GHSHQGHQFH DW KLJK LQSXW LQWHQVLWLHV LV DWWULEXWHG LQ DFFRUGDQFH ZLWK 5HI>$UD@ WR SRSXODWLRQ GHSOHWLRQ RI WKH JURXQG VWDWH 5DPDQ 0HDVXUHPHQWV $FFRUGLQJ WR 5HI >*OD@ JUHHQ OLJKW ZLWK WKH REVHUYHG LQWHQVLW\ ,$6) VHH )LJ f FDQ QRW DORQH LQGXFH VLJQLILFDQW fVLQJOHFRORUf 35 HIIHFW LQ /L7D +RZHYHU D FRPSDULVRQ RI WKH GDWD IRU ,$6) )LJ f ZLWK WKH UHVXOWV UHSRUWHG IRU /L1E >/DQ@ DQG 3U/L1E >%DL *XH@ OHDYHV RSHQ WKH SRVVLELOLW\ WKDW WKH VHFRQGDU\ JUHHQ HPLVn VLRQ RI (U FOXVWHUV KDV HQRXJK SKRWRQ HQHUJ\ DQG LQWHQVLW\ ,$6) WR HIIHFW WRJHWKHU ZLWK WKH SXPSLQJ UHG OLJKW VLJQLILFDQW SKRWRUHIUDFWLYLW\ YLD WZRFRORU H[FLWDWLRQ LQ (U/L7D

PAGE 68

QR 35 HIIHFW 7KH PHFKDQLFV RI WZRFRORU H[FLWDWLRQ LV GHWDLOHG ODWHU LQ 6HFWLRQ 7R YHULI\ WKH DVVXPSWLRQ WKDW WKH XSFRQYHUWHG JUHHQ HPLVVLRQ DFWV DV D fVHOIJDWLQJf VRXUFH HQKDQFLQJ WKH 35 .PLQ QR 35 HIIHFW ‘ 35 HIIHFW HIIHFW YLD WZRSKRWRQ RU WZRFRORUf H[FLn WDWLRQ ZLWKRXW WKH QHHG IRU DQ H[WHUQDO JDWn LQJ VRXUFH WKH 35 HIIHFW ZDV DQDO\]HG E\ 5DPDQ VSHFWURVFRS\ PHDVXUHPHQWV 7KH 35 HIIHFW LV NQRZQ WR FKDQJH WKH LQWHQVLW\ ,5 RI 5DPDQVKLIWHG OLQHV E\ DQ DPRXQW ZKLFK LV SURSRUWLRQDO WR WKH OLJKWLQGXFHG H[WUDRUGLQDU\ UHIUDFWLYHLQGH[ FKDQJH $Q 6LQFH $Q GHSHQGV RQ LQSXW SRZHU LQWHQVLW\ WKH FKDQJH RI ,5 DOVR GHSHQGV RQ DV LOOXVWUDWHG LQ )LJ 7KH GHSHQGHQFH RI $Q RQ LV FDOFXODWHG E\ WKH IROORZLQJ UHODWLRQ ZKLFK ZDV H[SHULPHQWDOO\ YHULILHG >.RV@ Ef )LJ 5DPDQ LQWHQVLW\ Df DQG QRUn PDOL]HG 5DPDQ LQWHQVLW\ Ef DV D IXQFWLRQ RI LQSXW LQWHQn VLW\ LOOXVWUDWLQJ WKH HIIHFWV RI SKRWRUHIUDFWLYH GDPDJH $QV-f $WSOO DAL, f ,5-PLQ > ,5 ZKHUH $QV-f LV WKH LQWHQVLW\GHSHQGHQW VWHDG\VWDWH YDOXH RI $Q -PLQ LV WKH PLQLPXP YDOXH RI ZKHQ SKRWRUHIUDFWLYH GDPDJH EHFRPHV XQGHWHFWDEOH E\ 5DPDQ VSHFWURVFRS\ ,5 PcQ DQG ,5 DUH 5DPDQ LQWHQVLWLHV PHDVXUHG DW WKH PLQLPXP DQG DQ DUELWUDU\ YDOXH RI UHVSHFWLYHO\ ,5 ,5-PLQLV D QRUPDOL]HG 5DPDQ LQWHQVLW\ DQG $ LV D FRHIILFLHQW

PAGE 69

)LJ 5DPDQ VSHFWUD RI YLUJLQ /L7D VROLGf DQG (U/L7D GDVKHGf PHDVXUHG DW ; QP DQG [O :FP )LJ 1RUPDOL]HG 5DPDQ LQWHQVLW\ RI WKH OLQH DW FPn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n YHUVXV ZDV FRQVWUXFWHG DQG LV VKRZQ LQ )LJ 1RWH WKDW LI WKH 35 HIIHFW ZDV DEVHQW ,5n ZRXOG EH FRQVWDQW YHUVXV )LJ f ,W LV WKHQ VHHQ IURP )LJ WKDW DOO VDPSOHV VXIIHUHG VRPH GHJUHH RI SKRWRUHIUDFWLYLW\ DV ZDV LQFUHDVHG )URP WKLV GDWD $QV DW DQ\ YDOXH RI FDQ EH FDOFXODWHG IURP (T f )LJ VKRZV WKH FDOFXODWHG YDOXH RI $QV IRU WKH (U/L7D VDPSOH ZLWK WKH KLJKn HVW (U FRQFHQWUDWLRQ DW WKH VXUIDFH PRObf DQG IRU D YLUJLQ /L7D VXEVWUDWH 7KHUH DUH WZR LPSRUWDQW REVHUYDWLRQV WR EH PDGH )LUVW WKH SUHVHQFH RI (U LRQV KDV

PAGE 70

LQSXW LQWHQVLW\ :FPf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} :FP 7KH VORSH $QVfRI WKLV GHSHQGHQFH LQFUHDVHG SURSRUWLRQDOO\ ZLWK (U FRQFHQWUDWLRQ 7KH SURSRUWLRQDOLW\ IDFWRU KDV D PD[LPXP YDOXH RI LQ WKH VDPSOH ZLWK WKH KLJKHVW (U FRQFHQWUDWLRQ )LJ f $W QP WKH GHSHQGHQFH RI $QV RQ IRU WKH VDPH (UGRSHG VDPSOH KDV D PXFK PRUH FRPSOLFDWHG FKDUDFWHU DV VHHQ LQ )LJ 8VLQJ D VLPSOH PDWKHPDWLFDO ILWWLQJ WR WKLV GDWD GHPRQVWUDWHV WKH DSSHDUDQFH RI DQ DGGLWLRQDO LQWHQVLW\ GHSHQGHQW FRPSRQHQW $QV-f VXSHULPSRVHG RQ WKH OLQHDU GHSHQGHQFH JLYHQ E\

PAGE 71

WKH IROORZLQJ HPSLULFDO UHODWLRQ $QV-f '>(U@ -[ f ZKHUH LV D FRQVWDQW RI SURSRUWLRQDOLW\ DQG >(U@ LV WKH (U FRQFHQWUDWLRQ $ERYH D FHUn WDLQ WKUHVKROG YDOXH RI DURXQG :FP WKH YDOXH RI [ LV %HORZ WKLV WKUHVKROG [ LV DERXW +RZHYHU DW OLJKW LQWHQVLWLHV ORZHU WKDQ :FP WKH FRQWULEXWLRQ GHVFULEHG E\ (T f EHFRPHV LQVLJQLILFDQW )LQDOO\ DW QP WKH GHSHQGHQFH RI $QV RQ ZDV URXJKO\ VLPLODU WR WKDW IRU QP +RZHYHU DFFXUDF\ LQ FDOFXODWLQJ $QV IURP VSRQWDQHn RXV 5DPDQ VFDWWHULQJ DW QP DV ZHOO DV IRU [O :FP DW QP ZDV KDPn SHUHG EHFDXVH RI LQWHUIHUHQFH IURP 6) DQG UHVRQDQW 5DPDQ VFDWWHULQJ 5RXJK HVWLPDWLRQ RI $QV JLYHV D YDOXH RI DURXQG [ 2n IRU WKH KLJKHVW XVHG [O :FPf )URP D FRPSDULVRQ RI WKH GDWD IRU $QV-f )LJ f ZLWK WKDW RI )LJ LW LV HYLn GHQW WKDW WKHUH LV D VWURQJ FRUUHODWLRQ EHWZHHQ WKH XSFRQYHUWHG JUHHQ HPLVVLRQ DQG WKH LQFUHDVHG 35 HIIHFW DW KLJK LQWHQVLWLHV RI 6XFK D FRUUHODWLRQ PD\ EH H[SUHVVHG DFFRUGLQJ WR (T f E\ WKH IROORZLQJ HPSLULFDO UHODWLRQV $QV (f,$6) ,DVI '(f>(U@ -[ f ZKHUH ( LV D FRQVWDQW RI SURSRUWLRQDOLW\ 1RWH WKDW WKH H[SHULPHQWDO GHSHQGHQFH RI $QV RQ ,$6) H[SUHVVHG E\ (T f LV LQ DJUHHPHQW ZLWK WKH JDWLQJ GHSHQGHQFH RI SKRWRUHIUDF WLYLW\ UHSRUWHG IRU 3U/L1E XVLQJ DQ H[WULQVLF VRXUFH RI JDWLQJ OLJKW >*XH@ $V VXFK ZH PD\ FRQFOXGH WKDW (U/L7D HPSOR\V VHOIJDWLQJ LH VHFRQGDU\ JUHHQ HPLVVLRQVf WR DFKLHYH VLJQLILFDQW WZRFRORU SKRWRUHIUDFWLYLW\ IURP D PRQRFKURPDWLF SXPS ,W LV QRWHG WKDW IRU D TXDQWLWDWLYH FRPSDULVRQ (UGRSHG /L1E VDPSOHV ZHUH DOVR SUHSDUHG ZLWK IDEULFDWLRQ FRQGLWLRQV VLPLODU WR WKRVH UHSRUWHG IRU DFWLYH GHYLFHV >%DX@ $ FRPSDULVRQ RI WKH GDWD UHSRUWHG LQ )LJ ZLWK WKH FRUUHVSRQGLQJ GDWD IRU (UGRSHG

PAGE 72

/L1E VKRZHG WKDW (UGRSHG /L7D KDV D PXFK VPDOOHU WLPHVf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n JDWLQJ VRXUFH WR DORQJ ZLWK WKH SXPS VRXUFH HQKDQFH SKRWRUHIUDFWLYLW\ E\ WKH WZRSKR WRQ RU WZRFRORU SKRWRUHIUDFWLYH HIIHFW ,Q WKLV VHFWLRQ WKH PHFKDQLVP FUHDWLQJ WKH XSFRQYHUWHG HPLVVLRQV LV LQYHVWLJDWHG 3DUWLFXODU DWWHQWLRQ LV SDLG WR WKH SURFHVV RI HQHUJ\ WUDQVIHU XSFRQYHUVLRQ EHWZHHQ (U FOXVWHUV 7KH WZRSKRWRQ PXOWLVWHS SURFHVVHV DVVRFLDWHG ZLWK WKH UHG DQG JUHHQ $6) HPLVn VLRQV DUH H[HPSOLILHG LQ )LJ ZKLFK VKRZV D SDUWLDO HQHUJ\OHYHO GLDJUDP IRU WKH (U LRQ (OHFWURQV LQ (U LRQV DUH ILUVW H[FLWHG WR WKH PHWDVWDEOH ) VWDWH WKHQ H[FLWHG WR WKH )f VWDWH DIWHU ZKLFK WKH\ GHFD\ QRQUDGLDWLYHO\ WR WKH*MM +__ RU 6 VWDWH 7KH UHG HPLVVLRQV DW QP FRUUHVSRQG WR D WUDQVLWLRQ IURP WKH VWDWH WR WKH ,M M H[FLWHG VWDWH ZKLOH WKH WUDQVLWLRQV IURP WKH +@ DQG 6 VWDWHV WR WKH JURXQG VWDWH SURGXFH WKH JUHHQ HPLVVLRQV 7KH VXEVWUXFWXUH RI WKH JUHHQ DQG UHG HPLVVLRQV UHIOHFWV WKH VSOLWWLQJ RI WKH JURXQG VWDWH DQG H[FLWHG VWDWHV LQWR 6WDUN VXEOHYHOV ,W LV LPSRUn WDQW WR QRWH WKDW WKH UHG HPLVVLRQV DW QP REVHUYHG KHUH IRU (U/L7D ZHUH QRW

PAGE 73

(78f RU QP SXPS U?\Y UHG $6) QP \?\Y 6) QP [\Y QP SXPS \?\Y )f f§J 6) QP ?\Y 0O n D S A JUHHQ $6) QP D UY )LJ 3DUWLDO HQHUJ\OHYHO GLDJUDP RI (U LRQV LQ /L7D GHSLFWLQJ WKH UDGLDWLYH WUDQVLWLRQV REVHUYHG LQ RXU PHDVXUHPHQWV 7KH (78 SURFHVV UDGLDWLYH WUDQVLWLRQV DQG QRQUDGLDWLYH WUDQVLWLRQV DUH UHSn UHVHQWHG E\ WKLQ VROLG DUURZV WKLFN VROLG DUURZV DQG GDVKHG DUURZV UHVSHFWLYHO\ REVHUYHG LQ SUHYLRXV VWXGLHV RI RWKHU (UGRSHG PDWHULDOV >$UD 'HO 'LH 5HG@ LQFOXGLQJ (U/L1E SXPSHG YLD WKH ) PDQLIROG ^*@ 8SFRQYHUVLRQ LV NQRZQ WR EH VWURQJHVW LQ UHJLRQV RI KLJKHVW (U FRQFHQWUDWLRQ >$UD *LO@ ZKHUH (U FOXVWHULQJ LV WKH JUHDWHVW $V VXFK WKH GRPLQDQW PHFKDQLVP RI XSFRQYHUVLRQ OHDGLQJ WR WKH UHG DQG JUHHQ $6) HPLVVLRQV REVHUYHG LQ (U/L7D ZDV FRQn FOXGHG WR EH (78 WKDW LV XSFRQYHUVLRQ E\ HQHUJ\ WUDQVIHU EHWZHHQ QHLJKERULQJ (U LRQV ZLWKLQ DQ (U FOXVWHU 7R FRQILUP WKLV WKH HIIHFW RI WKH (UGRSLQJ OHYHO RQ XSFRQYHUVLRQ HIILFLHQF\ ZDV H[DPLQHG E\ SURILOLQJ LQ WKH GHSWK GLUHFWLRQ ERWK WKH $6) DQG 6) HPLVn VLRQV ZLWKLQ WKH (ULQGLIIXVHG QHDUVXUIDFH OD\HU 6LQFH ,6) ZDV NQRZQ WR EH SURSRUWLRQDO

PAGE 74

)LJ 6WRNHV IOXRUHVFHQFH VSHFWUD DW YDULRXV GHSWKV G EHORZ WKH VXUIDFH WR WKH (U FRQFHQWUDWLRQ IRU SXPS LQWHQVLWLHV EHORZ :FP LW ZDV SRVVLEOH WR GHWHUn PLQH WKDW WKH (U FRQFHQWUDWLRQ SURILOH LV H[SRQHQWLDOOLNH ZLWK D H GHSWK RI DURXQG SP )LJ f 7KH LQWHQVLW\ SURILOH RI WKH $6) HPLVVLRQ KDG KRZHYHU D GHSWK RI DURXQG SP ,Q RWKHU ZRUGV WKH $6) HPLVVLRQV HPDQDWHG IURP WKH UHJLRQV RI KLJKHVW (U FRQn FHQWUDWLRQ QHDU WKH VXUIDFHf ZKHUH PRVW RI WKH (U FOXVWHUV UHVLGHG 7KLV ZDV GLUHFW HYLn GHQFH RI WKH GRPLQDWLQJ UROH RI QRQUDGLDWLYH HQHUJ\ WUDQVIHU EHWZHHQ FOXVWHUHG (U LRQV LQ WKH PHFKDQLVP RI UHFRQYHUVLRQ LQ (UGRSHG /L7D 6LQFH (U FOXVWHULQJ ZDV GHWHUPLQHG WR EH WKH SULPDU\ IDFWRU OHDGLQJ WR XSFRQYHU VLRQ VLJQLILFDQW UHGXFWLRQ RI WKH XSFRQYHUVLRQ HIILFLHQF\ LQ (U/L7D ZDV H[SHFWHG WR EH DFKLHYHG E\ D SRVWLQGLIIXVLRQ SURFHVV WR UHGXFH (U FOXVWHULQJ 7KLV LV SXUVXHG LQ WKH QH[W VHFWLRQ

PAGE 75

/L 7UHDWPHQW ,Q DQ HIIRUW WR GHFUHDVH WKH DPRXQW RI (U FOXVWHUV DQG KHQFH UHGXFH XSFRQYHU VLRQ D /LWUHDWPHQW SURFHVV ZDV HPSOR\HG WR LQFUHDVH WKH /L7D UDWLR 7KH UHVXOWV RI WKLV /LWUHDWPHQW DUH GLVFXVVHG LQ WKLV VHFWLRQ 6RPH RI WKH IDEULFDWHG (U/L7D VDPSOHV ZHUH SURFHVVHG LQ PROWHQ /L1 6XEn VHTXHQW IOXRUHVFHQFH PHDVXUHPHQWV VKRZHG D JUDGXDO GHFUHDVH RI ,$6) ZLWK DQ LQFUHDVH LQ WKH SURFHVV WLPH XS WR KRXUV $W WKH VDPH WLPH WKH LQWHQVLW\ RI 6) UHPDLQHG QHDUO\ FRQVWDQW LH WKH WRWDO (U FRQWHQW LQ WKH FU\VWDO ZDV QRW DIIHFWHG E\ WKLV SURFHVVLQJ )XUn WKHU LQFUHDVH RI WKH SURFHVVLQJ WLPH LQGXFHG RQO\ D VPDOO FKDQJH LQ ,$6) 7KH PRVW LPSUHVVLYH UHVXOW ZDV D QHDUO\ b UHGXFWLRQ LQ WKH LQWHQVLW\ RI WKH JUHHQ $6) HPLVVLRQ DV D UHVXOW RI WKH SRVWLQGLIIXVLRQ SURFHVVLQJ DV LOOXVWUDWHG LQ )LJ 7KLV UHGXFWLRQ LV : WLQ RR D ' 2 8 2 & D R O O O O O O O O O O O O O O O O O O O O O L L L L L O O O O O O (U/L7D "L QP [O :FP (Q/L7D2M "L QP [O :FP f§ /L(U/L7D2L ; QP [O :FP ZDYHOHQJWK QPf )LJ )OXRUHVFHQFH VSHFWUD RI WKH JUHHQ $6) HPLVVLRQ IURP WKH (ULQGLI IXVHG VDPSOH ZLWK WKH KLJKHVW (U FRQFHQWUDWLRQ 7KH GRWWHG OLQH FRUUHVSRQGV WR D SXPS ZDYHOHQJWK RI ; QP DQG LQSXW LQWHQVLW\ [O :FP 7KH VROLG DQG GDVKHG OLQHV FRUUHVSRQG WR f QP DQG [O :FP EHIRUH DQG DIWHU /L WUHDWPHQW UHVSHFWLYHO\

PAGE 76

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n FLHQWO\ GHHS WR UHGXFH WKH (U FOXVWHU FRQFHQWUDWLRQ DQG WKHUHIRUH PD\ EH D XVHIXO WHFKn QLTXH IRU WKH RSWLPL]DWLRQ RI DFWLYH LQWHJUDWHGRSWLFDO GHYLFHV XWLOL]LQJ DQ (UGRSHG QHDUn VXUIDFH OD\HU LQ /L7D FU\VWDOV 7R WKLV SRLQW LW KDV EHHQ GHWHUPLQHG WKDW (78 LV WKH GRPLQDQW PHFKDQLVP RI XSFRQYHUVLRQ LQ (U/L7D DQG WKDW WKH SUHVHQFH RI WKLV XSFRQYHUWHG OLJKW OHDGV WR 35 GDPDJH YLD WKH WZRSKRWRQ RU WZRFRORU 35 HIIHFW $ WZRSKRWRQ PRGHO XVHG WR H[HPn SOLI\ WKLV SURFHVV LV GHVFULEHG LQ WKH QH[W VHFWLRQ 7KH 7ZR3KRWRQ 0RGHO RI WKH 35 (IIHFW /L7D4fV $GYDQWDJH LQ 'DWD 6WRUDJH DQG /DVHUV ,Q WKLV VHFWLRQ D WZRSKRWRQ PRGHO LV DSSOLHG WR PRGHO WKH 35 HIIHFW DQG H[SODLQ KRZ XSFRQYHUWHG OLJKW DORQJ ZLWK WKH SUHVHQFH RI WKH ORQJHU ZDYHOHQJWK LQFLGHQW OLJKW OHDGV WR 35 GDPDJH 'XULQJ WKLV SURFHVV WKH LQKHUHQW DGYDQWDJHV RI /L7D IRU DSSOLFDn WLRQV VXFK DV RSWLFDO GDWD VWRUDJH DQG ODVHU IDEULFDWLRQ ZLOO DOVR EH GHVFULEHG 7R GHYHORS D WKHRUHWLFDO H[SODQDWLRQ RI WKH SKHQRPHQRQ RI SKRWRUHIUDFWLYLW\ HQKDQFHPHQW E\ (UGRSLQJ WKH WZRSKRWRQ PRGHO RI WKH 35 HIIHFW KDV EHHQ HPSOR\HG

PAGE 77

Ha PLJUDWLRQ FRQGXFWLRQ EDQG L Q f r‘ UHFRUGLQJ OLJKW r-;:? n JDWLQJ OLJKW $9: SRODURQ ELSRODURQ GHHS WUDS YDOHQFH EDQG 7KLV PRGHO KDV EHHQ GHYHORSHG DQG H[SHULPHQWDOO\ FRQILUPHG IRU /L1E >-HU .RVD@ DQG KDV EHHQ DSSOLHG KHUH WR /L7D GXH WR WKH VLPLODULW\ RI WKH LQWULQVLF GHIHFW VWUXFWXUHV EHWZHHQ WKH WZR PDWHULDOV >*RS@ ,Q WKH FDVH RI XQGRSHG OLWKLXP QLREDWH )LJ 6FKHPDWLFOHYHO GLDJUDP RI WKH WZR SKRWRQ 35 HIIHFW GHSLFWLQJ WKH JDW /L1Ef >.RVD /DQ@ WKLV PHFK LQJ OLJKW ORQJHU ZDYHOHQJWK UHFRUGn LQJ OLJKW DQG ELSRODURQ OHYHO VLQJOH DQLVP LQYROYHV ELSRODURQV DV GHHS SRODURQ OHYHO DQG GHHS WUDS RXWVLGH WKH LOOXPLQDWHG DUHDf GRQRUV DQG PHWDVWDEOH VPDOO SRODURQV ZKLFK SURYLGH WKH KLJK 35 VHQVLWLYLW\ %LSRODURQV DQG SRODURQV H[LVW GXH WR WKH LQWULQVLF GHIHFW VWUXFWXUH RI /L1E RULJLQDWLQJ IURP LWV OLWKLXP GHILFLHQF\ >6FK @ 1HDU LQIUDUHG RU UHGf UHFRUGLQJ >%DL *XH /DQ@ RFFXUV IURP WKH VPDOO SRODURQ VWDWH FRQVLVWLQJ RI DQ HOHFWURQ WUDSSHG DW WKH DQWLVLWH GHIHFW 1E/Lf GHVLJQDWHG OHYHO LQ )LJ 7KLV OHYHO LV QRUPDOO\ SRSXODWHG YLD DQ H[WULQVLFDOO\ DSSOLHG EOXHJUHHQ JDWLQJ OLJKW WKURXJK WKH SKRWRGLVVRFLDWLRQ RI ELSRODURQV OHYHO LQ )LJ f $IWHU WKH H[FLWDWLRQ RI HOHFWURQV IURP WKH VPDOO SRODURQ VLWHV WR WKH FRQGXFWLRQ EDQG WKH HOHFWURQV GULIW RXW RI WKH LOOXPLQDWHG DUHD DQG DUH UHWUDSSHG DW ELSRn ODURQ VLWHV DQG RWKHU GHHS WUDSV OHYHO f UHVXOWLQJ LQ 35 GDPDJH RU 35 UHFRUGLQJ LI WKH DSSOLFDWLRQ LV RSWLFDO GDWD VWRUDJH 6XFK D WZRFRORU VFKHPH KDV WKH DGYDQWDJH RI D YHU\ KLJK JDWLQJ UDWLR LH WKH UDWLR RI ZULWLQJ VHQVLWLYLW\ ZLWK DQG ZLWKRXW WKH VHFRQG JDWLQJ FRORU >*XH /DQ@

PAGE 78

7KH PDLQ IHDWXUH RI WKH WZRSKRWRQ H[FLWDWLRQ LV WKH QHDUO\ OLQHDU GHSHQGHQFH RI $QV RQ IRU KLJK LQSXW LQWHQVLWLHV ZKLFK KDV EHHQ YHULILHG H[SHULPHQWDOO\ LQ /L1E >)XM -HU@ DQG /L7D >*OD@ 0RUHRYHU LW KDV EHHQ HVWDEOLVKHG >.RVD@ WKDW ^$QVf-` a 1 ZKHUH 1 LV WKH FRQFHQWUDWLRQ RI DQWLVLWH GHIHFWV 6XFK D SURSRUWLRQDOLW\ LV LQ IXOO DFFRUGDQFH ZLWK WKH WZRSKRWRQ PRGHO ,W LV ZHOO NQRZQ WKDW (U >*LO@ DQG 3U >*XH@ GRSDQWV LQFUHDVH 1 HVSHFLDOO\ ELSRODURQV LQ /L1E WKURXJK WKH QHHG IRU FKDUJH FRPSHQVDWLRQ RI LRQV ZLWK YDOHQFH VWDWH f +HQFH GDWD REWDLQHG KHUH DW WKH H[FLWDWLRQ ZDYHOHQJWK RI QP )LJ f LV FOHDU HYLGHQFH RI D VLPLODU HIIHFW IRU (U LQ /L7D 0RUHRYHU WKH UHODWLYH FKDQJH RI 1 FDXVHG E\ (UGRSLQJ FDQ EH HVWLPDWHG IURP D FRPSDULVRQ RI WKH VORSH ^$QVf-` REVHUYHG LQ WKH (UGRSHG VDPSOH ZLWK WKH RQH REVHUYHG LQ WKH YLUJLQ FRQJUXHQW VXEVWUDWH ,W ZDV IRXQG WKDW 1 LV LQFUHDVHG WLPHV E\ (UGRSLQJ LQ WKH VDPSOH ZLWK WKH KLJKHVW (U FRQFHQWUDWLRQ )LJ f 5DGLDWLRQ DW DQG QP LV QHDU WKH PD[LPXP RI WKH DEVRUSWLRQ EDQG IRU VPDOO SRODURQV EXW LV QRW RSWLPDO IRU WKH SKRWRLQGXFHG GLVVRFLDWLRQ RI ELSRODURQV DV WKHVH ZDYHOHQJWKV DUH YHU\ IDU IURP WKH ELSRODURQLF DEVRUSWLRQ EDQG >6FK@ 7KHUHIRUH WKH JHQHUDWLRQ UDWH RI VKDOORZ 35 FHQWHUV LV WRR ORZ WR LQGXFH VLJQLILFDQW SKRWRUHIUDFWLYLW\ HYHQ DW KLJK OLJKW LQWHQVLWLHV +RZHYHU WKH VLWXDWLRQ FKDQJHV ZKHQ H[FLWDWLRQ ZLWK WZR ZDYHOHQJWKV LV XVHG ZLWK SKRWRQV RI WKH VKRUWHU ZDYHOHQJWK JUHHQ OLJKWf H[FLWLQJ ELSRn ODURQV DQG LQFUHDVLQJ WKH YLUWXDO SRSXODWLRQ RI VLQJOHSRODURQ OHYHOV DQG SKRWRQV RI WKH RWKHU ZDYHOHQJWK UHG RU LQIUDUHGf VZHHSLQJ WKHP HIIHFWLYHO\ LQWR WKH FRQGXFWLRQ EDQG >*XH /DQ@ DV LOOXVWUDWHG LQ )LJ 5HFHQW UHVHDUFK ZLWK DGGLWLRQDO VRXUFHV RI VXFK JDWLQJ OLJKW KDV VKRZQ >*XH /DQ@ WKDW WKH WZRFRORU VFKHPH KDV D YHU\ KLJK

PAGE 79

JDWLQJ UDWLR LH WKH UDWLR RI 35 VHQVLWLYLW\ ZLWK DQG ZLWKRXW WKH VHFRQG FRORUf DQG WKHUHn IRUH FDQ EH LPSOHPHQWHG HYHQ ZLWK ORZSRZHU VRXUFHV HJ :FP >/DQ@ : FP >*XH@f ,W PHDQV WKDW WKH VHFRQGDU\ JUHHQ HPLVVLRQ RI (U FOXVWHUV KDV HQRXJK SKRWRQ HQHUJ\ DQG LQWHQVLW\ ,$6) WR HIIHFW WRJHWKHU ZLWK WKH SXPSLQJ UHG OLJKW VLJQLILFDQW SKRWRUHIUDFWLYLW\ YLD WKH WZRFRORU H[FLWDWLRQ 7KH GLIIHUHQFH LQ H[FLWDWLRQ RI JUHHQ HPLVVLRQ REVHUYHG EHWZHHQ SXPSLQJ DW DQG QP LV FRQVLVWHQW ZLWK SUHYLRXV GDWD >*@ RQ WKH VSHFWUDO GHSHQGHQFH RI XS FRQYHUVLRQ H[FLWDWLRQ ZLWKLQ (U FOXVWHUV 6XFK DQ XSFRQYHUVLRQ LV DQ XQGHVLUDEOH HIIHFW LQ (UGRSHG GHYLFHV >$PL *LO +XD@ RSHUDWLQJ DW WHOHFRPPXQLFDWLRQ ZDYHOHQJWKV ([SHULPHQWDO GDWD UHSRUWHG KHUH GHPRQVWUDWHV WKDW WKLV XSFRQYHUVLRQ LQGXFHV D PRUH SURn QRXQFHG 35 HIIHFW LQ /L1E WKDQ LQ /L7D 7KLV LV EHFDXVH WKH JUHHQ HPLVVLRQ RI (U FOXVWHUV LV ZLWKLQ WKH PD[LPXP RI WKH ELSRODURQLF DEVRUSWLRQ EDQG LQ /L1E >6FK@ \HW LV IDU IURP WKH PD[LPXP RI WKH DEVRUSWLRQ EDQG LQ /L7D >.RVE@ 7KLV IDFW PDNHV /L7D PRUH DWWUDFWLYH IRU WKH IDEULFDWLRQ RI (UGRSHG ZDYHJXLGH DPSOLILHUV $W WKH VDPH WLPH GDWD SUHVHQWHG KHUH DOVR SRLQWV WR WKH DGYDQWDJH RI /L7D RYHU /L1E DPRQJ SHUn VSHFWLYH UHDUHDUWKGRSHG 35 PDWHULDOV IRU QRQYRODWLOH KRORJUDSKLF VWRUDJH ZLWK WZRSKRn WRQ UHFRUGLQJ $W DQ RSWLPDO ZDYHOHQJWK RI H[WULQVLF JDWLQJ OLJKW UDQJLQJ IURP WR QP LQ /L7D >.RVE@f D VDWXUDWHG YDOXH RI KRORJUDSKLF HIILFLHQF\ VKRXOG EH OLPLWHG PDLQO\ E\ WKH LQWULQVLF GHIHFWV FRQFHQWUDWLRQ 1 KDYLQJ D KLJKHU YDOXH WLPHVf LQ /L7D >%RU .RV@ %HFDXVH WKH JDWLQJ EHDP JHQHUDOO\ GRHV QRW KDYH WR EH IURP D FRKHUHQW OLJKW VRXUFH >*XH@ ORZFRVW OLJKW VRXUFHV VXFK DV ILOWHUHG [HQRQ DQG PHUFXU\ ODPSV FDQ EH XVHG IRU WKLV SXUSRVH LQ (U/L7D

PAGE 80

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n WRQ PRGHO ZDV WKHQ DSSOLHG WR PRGHO WKH 35 HIIHFW DQG H[SODLQ KRZ XSFRQYHUWHG OLJKW DORQJ ZLWK WKH SUHVHQFH RI WKH ORQJHU ZDYHOHQJWK LQFLGHQW OLJKW FDXVHG 35 GDPDJH WKURXJK fVHOIJDWLQJf )LQDOO\ LQ DGGLWLRQ WR LWV NQRZQ SRWHQWLDO IRU DSSOLFDWLRQV OLNH KLJKVSHHG PRGXODWLRQ DQG QRQOLQHDU IUHTXHQF\ GRXEOLQJ /L7D ZDV DOVR GHPRQVWUDWHG WR KDYH VWURQJ DGYDQWDJHV DQG SRWHQWLDO IRU WKH DSSOLFDWLRQV RI RSWLFDO GDWD VWRUDJH ODVHUV DQG GHYLFHV LQFRUSRUDWLQJ 35 JUDWLQJV +RZHYHU WR UHDOL]H WKHVH GHYLFHV VWDEOH ZDYHJXLGHV LQ /L7D PXVW ILUVW EH GHYHORSHG 7KH VXEMHFW RI ZDYHJXLGHV LV UHWXUQHG WR LQ WKH QH[W FKDSWHU ZKHUH ZDYHJXLGHV DUH FKDUDFWHUL]HG XVLQJ D YDULHW\ RI WHFKQLTXHV

PAGE 81

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n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

PAGE 82

SKRWRUHIUDFWLYH FKDUDFWHULVWLF RI ZDYHJXLGHV LQ WKH LQIUDUHG QHDU _LP LV PHDVXUHG QH[W DQG FRPSDUHG WR VLPLODU VWUXFWXUHV LQ /L1E 7KH FKDSWHU FRQFOXGHV ZLWK VRPH PHDn VXUHPHQWV RI SODQDU ZDYHJXLGHV LQFOXGLQJ LQGH[ SURILOH PHDVXUHPHQWV DQG URFNLQJ FXUYH DQDO\VLV ZKLFK ZLOO SURYLGH GDWD WR EH XVHG LQ D VXEVHTXHQW FKDSWHU 1HDU)LHOG &KDUDFWHUL]DWLRQ ,Q WKLV VHFWLRQ D ILUVW LQVSHFWLRQ RI WKH QXPEHU RI PRGHV VXSSRUWHG E\ WKH FKDQQHO ZDYHJXLGHV RI YDULRXV ZLGWKV LV FDUULHG RXW 7KLV LV GRQH XVLQJ D QHDUILHOG FKDUDFWHUL]Dn WLRQ PHWKRG ZKHUHE\ WKH RXWSXW IDFHV RI ZDYHJXLGHV H[FLWHG IURP WKH VLQJOHPRGH ILEHU SLJWDLO RI D GLRGH ODVHU VRXUFH DUH IRFXVHG RQWR D FDPHUD VR WKDW WKH QXPEHU RI VSRWV DVVRn FLDWHG ZLWK GLIIHUHQW PRGHV FDQ EH FRXQWHG RII RI WKH GLVSOD\ PRQLWRU ZDYHJXLGH )LJ 6FKHPDWLF RI WKH QHDUILHOG FKDUDFWHUL]DWLRQ VHWXS XVHG WR LQLWLDOO\ GHWHUPLQH WKH QXPEHU RI PRGHV VXSSRUWHG E\ WKH FKDQQHO ZDYHJXLGHV RI YDULRXV ZLGWKV

PAGE 83

'HVFULSWLRQ RI WKH 0HDVXUHPHQW 7HFKQLTXH /LJKW IURP D ILP GLRGH ODVHU ZDV FRXSOHG LQWR HDFK RI WKH FKDQQHO ZDYHJXLGHV YLD D VLQJOHPRGH ILEHU SLJWDLO 7KH FRPSOHWH H[SHULPHQWDO VHWXS LV LOOXVWUDWHG LQ )LJ 7KH RXWSXW IDFH RI HDFK ZDYHJXLGH ZDV IRFXVHG XVLQJ D [ OHQV RQWR WKH LQSXW VXUIDFH RI DQ LQIUDUHG ,5f FDPHUD ZLWK WKH RXWSXW YLGHR IHHG RI WKH FDPHUD FRQQHFWHG GLUHFWO\ WR D YLGHR GLVSOD\ PRQLWRU 8VLQJ WKLV DSSURDFK WKH QXPEHU RI PRGHV VXSSRUWHG E\ D SDUWLFXn ODU ZDYHJXLGH FDQ HDVLO\ EH GHWHUPLQHG E\ FRXQWLQJ WKH QXPEHU RI REVHUYHG PRGH VSRWV DSSHDULQJ RQ WKH YLGHR PRQLWRU VFUHHQ 6RPH UHVXOWV RI WKLV PHDVXUHPHQW WHFKQLTXH DUH SUHVHQWHG QH[W Ef )LJ 9LGHR PRQLWRU VFUHHQ VQDS VKRWV GHSLFWLQJ D VLQn JOHPRGH Df DQG D GRXn EOHPRGH Ef FKDQQHO ZDYHJXLGH 0RGDO &KDUDFWHUL]DWLRQ 5HVXOWV %\ VFDQQLQJ WKH VLQJOHPRGH ILEHU IURP WKH ODVHU VRXUFH DFURVV WKH LQSXW IDFH RI HDFK ZDYHJXLGH LW ZDV SRVVLEOH WR FUHDWH RIID[LV FRXSOLQJ LQWR DQ\ KLJKHU RUGHU PRGHV WKDW PD\ KDYH EHHQ VXSSRUWHG E\ WKH ZDYHJXLGH 7KLV ZDV GRQH IRU HDFK RI WKH FKDQQHO ZDYHJXLGHV LQ /L7D DQG WKH QXPEHU RI PRGHV VXSSRUWHG IRU HDFK ZDV GHWHUPLQHG $V DQ H[DPSOH )LJ VKRZV WKH YLGHR PRQLWRU VFUHHQ SKRWRJUDSKV WDNHQ IRU D VLQJOHPRGH DQG GRXEOHPRGH FKDQQHO ZDYHJXLGH

PAGE 84

)DEULFDWLRQ 7HFKQLTXH )DEULFDWLRQ &RQGLWLRQV &U\VWDO &XW 6LQJOH0RGH 5HJLRQ $3( S\URSKRVSKRULF DFLGf 3( r& KUV $ r& OKU 3( r& KU $ r& KUV 3( r& PLQ = SP $3( S\URSKRVSKRULF DFLGf 3( r& PLQ $ r& KU 3( r& PLQ $ r& KUV = SP 3( JO\FHULQf 3( r& KUV ; SP 7LLQGLIIXVLRQ ƒ r& KUV ] SP =Q YDSRULQGLIIXVLRQ r& KUV ] SP 7DEOH 1HDUILHOG FKDUDFWHUL]DWLRQ UHVXOWV GHSLFWLQJ WKH UHJLRQ RI VLQJOHn PRGH RSHUDWLRQ LPPHGLDWHO\ DIWHU IDEULFDWLRQ IRU WKH H[WUDRUGLQDU\ LQGH[ LQ WHUPV RI ZDYHJXLGH FKDQQHO ZLGWK IRU YDULRXV IDEULFDWLRQ FRQGLWLRQV 7KHUH ZHUH IDU WRR PDQ\ ZDYHJXLGHV SURGXFHG ZLWK YDU\LQJ IDEULFDWLRQ FRQGLn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f DQG VWDELOLW\ LVVXHV DUH DGGUHVVHG &KDSWHU f

PAGE 85

,Q WKH QH[W VHFWLRQ D SULVP FRXSOLQJ WHFKQLTXH LV DSSOLHG WR PHDVXUH WKH HIIHFWLYH PRGH LQGH[ LQFUHPHQWV RI FKDQQHO ZDYHJXLGHV DW YDULRXV ZDYHOHQJWKV 6XFK D WHFKQLTXH DOORZV IRU D URXJK HVWLPDWLRQ RI WKH PDJQLWXGH RI WKH LQGH[ LQFUHPHQW ZLWKLQ D ZDYHJXLGH DV ZHOO DV SURYLGLQJ VRPH LQIRUPDWLRQ DV WR WKH ZDYHOHQJWK GLVSHUVLRQ RI WKH LQGH[ LQFUHn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n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

PAGE 86

)LJ 6FKHPDWLF RI WKH VWDQGDUG SULVP FRXSOLQJ PHWKRG XVHG WR VHOHFn WLYHO\ FRXSOH OLJKW LQWR LQGLYLGXDO ZDYHJXLGH PRGHV RQ WKH FRXSOLQJ RI OLJKW IURP WKH SULVP WR WKH ZDYHJXLGH E\ VDWLVI\LQJ WKH SKDVH PDWFKLQJ FRQGLWLRQ IRU HDFK PRGH ZLWKLQ WKH ZDYHJXLGH VHSDUDWHO\ /LJKW LQFLGHQW RQ WKH IDFH K\SRWHQXVHf RI WKH SULVP DW DQ DQJOH c LV UHIUDFWHG DV LW HQWHUV WKH SULVP DFFRUGLQJ WR 6QHOOfV ODZ mFVLQ f 7KH DQJOH RI LQFLGHQFH RI WKH LQFRPLQJ EHDP LV DGMXVWHG VR DV WR PDLQWDLQ WRWDO LQWHUQDO UHIOHFWLRQ DW WKH QSQF LQWHUIDFH 7KLV FRQGLWLRQ LV VDWLVILHG LI WKH DQJOH LV JUHDWHU WKDQ WKH FULWLFDO DQJOH 2 DVLQ / F LQ A Q 9 3f 7KH FRXSOLQJ RI OLJKW LQWR PRGHV RI WKH ZDYHJXLGH FDQ RFFXU RQO\ LI WKH KRUL]RQWDO FRPSRn QHQWV RI WKH N YHFWRU LQ WKH SULVP PDWFKHV WKH SURSDJDWLRQ FRQVWDQW RI WKH ZDYHJXLGH 7KLV LV UHIHUUHG WR DV WKH SKDVH PDWFKLQJ FRQGLWLRQ

PAGE 87

1 Q VLQ 3 A f ZKHUH 1 LV WKH QRUPDOL]HG SURSDJDWLRQ FRQVWDQW RU HIIHFWLYH PRGH LQGH[ RI D SDUWLFXODU PRGH RI WKH ZDYHJXLGH ,Q RUGHU WR REWDLQ UHDO DQJOHV WKH SULVP LQGH[ QS WKHUHIRUH VKRXOG EH ODUJHU WKDQ 1 )RU PD[LPXP FRXSOLQJ WKH JDS K EHWZHHQ WKH SULVP DQG WKH ZDYHJXLGH VKRXOG EH DV VPDOO DV SRVVLEOH 8VXDOO\ LW LV DERXW D TXDUWHU ZDYHOHQJWK DQG FDQ EH DFKLHYHG WKURXJK WKH XVH RI FODPSV SXVKLQJ WKH SULVP XS DJDLQVW WKH VXEVWUDWH VXUn IDFH $GGLWLRQDOO\ WKH LQWHUDFWLRQ GLVWDQFH G VKRXOG EH FRPSDUDEOH WR RQH FRXSOLQJ OHQJWK ,I LW LV ORQJHU OLJKW PD\ EH FRXSOHG EDFN LQWR WKH SULVP 8VLQJ VLPSOH JHRPHWU\ ZLWKLQ WKH SULVP DQG 6QHOOfV ODZ LW FDQ EH VKRZQ WKDW WKH HIIHFWLYH PRGH LQGH[ 1 LV UHODWHG WR WKH LQFLGHQW DQJOH E\ 1 QAVLQ $ DVLQ I VLQ*A Q 9 3 f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n SOLQJ LQWR WKH ZDYHJXLGH QRZ LQ WKH UHYHUVH GLUHFWLRQ ,Q WKLV FDVH OLJKW LV ILUVW ODXQFKHG LQWR WKH ZDYHJXLGH VWUXFWXUH H[FLWLQJ DOO RI WKH H[LVWLQJ PRGHV (DFK RI WKHVH PRGHV ZLOO KDYH D GLIIHUHQW HIIHFWLYH LQGH[ 1 YDOXH DQG WKXV H[LW WKH K\SRWHQXVH IDFH RI WKH SULVP DW D GLIIHUHQW DQJOH 7KH DQJXODU VSHFWUXP RI WKH HPHUJLQJ OLJKW H[KLELWV SHDNV FDOOHG POLQHV FRUUHVSRQGLQJ WR WKH PRGHV RI GLIIHUHQW RUGHU 7KH HIIHFWLYH LQGH[ YDOXHV IRU HDFK RI WKH

PAGE 88

PRGHV FDQ WKHQ EH FDOFXODWHG E\ PHDVXULQJ WKH DQJXODU GHYLDWLRQ RI WKH FRUUHVSRQGLQJ LQn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n RU EHWWHU +RZHYHU WKH YDOXHV RI SULVP DQG EXON LQGH[ DUH XVXDOO\ RQO\ NQRZQ ZLWK DQ DFFXUDF\ RI DW EHVW $V VXFK WKH VWDQGDUG SULVP FRXSOLQJ PHWKRG LV QRW D GHSHQGn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f FDQ EH SUHFLVHO\ PHDVXUHG DQG

PAGE 89

ZDYHJXLGH )LJ 0RGLILHG SULVP FRXSOLQJ WHFKQLTXH XVHG WR PHDVXUH PRGH HIIHFn WLYH LQGH[ LQFUHPHQWV VXEVWLWXWHG LQWR WKH GHULYDWLYH RI WKH SULVP FRXSOLQJ HTXDWLRQ (T f ZLWK UHVSHFW WR WKH FRXSOLQJ DQJOH G1 B $1 3 f $ FRV $ DVLQ FRV3 f 6LQFH WKH PHDVXUHG DQJOH $ LV WKH DQJOH EHWZHHQ WKH VXEVWUDWH DQG DQ\ PRGH RI LQWHUHVW WKLV QHZ SULVP FRXSOLQJ HTXDWLRQ JLYHV WKH PRGH LQGH[ LQFUHPHQW $1 1QF RI WKLV PRGH DERYH WKH VXEVWUDWH YDOXH %HFDXVH $%m% DQG WKLV HTXDWLRQ LV QRW VHQVLWLYH WR WKH YDOXH RI % LW FDQ EH HYDOXDWHG DW DQ\ DQJOH LQVLGH $% 7KH DFFXUDF\ RI WKLV QHZ WHFKQLTXH LV HVWLn PDWHG WR EH 0HDVXUHG 0RGH ,QGH[ ,QFUHPHQWV 8VLQJ WKH PRGLILHG SULVP FRXSOLQJ PHWKRG WKH PRGH LQGH[ LQFUHPHQWV $1 RI VWUDLJKWFKDQQHO $3( ZDYHJXLGHV LQ /L7D ZHUH PHDVXUHG DW IRXU GLIIHUHQW ZDYHOHQJWKV

PAGE 90

FKDQQHO PDVN ZLGWK _LPf )LJ 0HDVXUHG IXQGDPHQWDO PRGH LQGH[ LQFUHPHQWV DW GLIIHUHQW ZDYHn OHQJWKV IRU FKDQQHO ZDYHJXLGHV RI YDULRXV ZLGWKV REWDLQHG IURP WKH PRGLILHG SULVP FRXSOHU PHWKRG _DP _DP _[P DQG _LP $ UXWLOH SULVP ZLWK DQ DQJOH $ r ZDV XVHG IRU FRXSOLQJ OLJKW RXW RI WKH ZDYHJXLGH 7KH UHIUDFWLYH LQGH[ QS RI WKH UXWLOH SULVP DW HDFK ZDYHOHQJWK ZDV IRXQG IURP D FXELF VSOLQH LQWHUSRODWLRQ RI WKH GLVSHUVLRQ GDWD IRU UXWLOH SUHVHQWHG LQ $SSHQGL[ $ 7KH PHDVXUHG PRGH LQFUHPHQWV IRU WKH IXQGDPHQWDO PRGH DW HDFK ZDYHOHQJWK DUH SORWWHG LQ )LJ DV D IXQFWLRQ RI FKDQQHO PDVN ZLGWK )RU WKH H[SHULPHQWDOO\ PHDVXUHG

PAGE 91

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n QLTXHV XVHG WR PHDVXUH ERWK WRWDO LQVHUWLRQ ORVV DQG SURSDJDWLRQ ORVV DUH GHVFULEHG DQG VRPH PHDVXUHPHQW UHVXOWV DUH SURYLGHG /RVV 0HDVXUHPHQW 6HWXS 7KH H[SHULPHQWDO VHWXS XVHG WR PHDVXUH ERWK WRWDO LQVHUWLRQ ORVV DQG SURSDJDWLRQ ORVV LV GHSLFWHG LQ )LJ 6LQJOHPRGH SRODUL]DWLRQPDLQWDLQLQJ ILEHU IURP DQ +3 )LJ ([SHULPHQWDO VHWXS XVHG IRU PHDVXULQJ ZDYHJXLGH ORVV

PAGE 92

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n VXUHG LQ D VLPLODU PDQQHU 7KH GLIIHUHQFH EHWZHHQ WKH WZR PHDVXUHPHQWV LV WKH WRWDO LQVHUn WLRQ ORVV RI WKH ZDYHJXLGH VDPSOH )RU PRVW RI WKH ZDYHJXLGH VDPSOHV XVHG LQ WKLV UHVHDUFK WKH WRWDO LQVHUWLRQ ORVV ZDV DURXQG G% 7KH H[DFW YDOXH RI LQVHUWLRQ ORVV GHSHQGHG RQ ZDYHJXLGH ZLGWK VDPSOH OHQJWK IDEULFDWLRQ WHFKQLTXH DQG FRQGLWLRQV TXDOLW\ RI WKH HQGIDFH SROLVK DQG TXDOLW\ RI WKH FOHDYHG IDFH RI WKH ILEHU $V D EHVW UHVXOW WKH ILEHUWRILEHU LQVHUWLRQ ORVV RI D SP ZLGH ZDYHJXLGH IDEULn FDWHG E\ 3( LQ JO\FHULQ DW r& IRU KRXUV ZDV PHDVXUHG 7R GR WKLV WKH OHQV DW WKH RXWSXW IDFH RI WKH ZDYHJXLGH ZDV UHSODFHG ZLWK DQRWKHU VLQJOHPRGH ILEHU DQG WKH PHDn VXUHPHQW ZDV PDGH LQ D VLPLODU PDQQHU DV GHVFULEHG DERYH XVLQJ D OLJKWZDYH PXOWLPHWHU WR PHDVXUH RSWLFDO WKURXJKSXW LQVWHDG RI WKH *H GHWHFWRU DQG ORFNLQ DPSOLILHU 7KH ILEHU WRILEHU LQVHUWLRQ ORVV IRU WKLV VDPSOH ZDV DURXQG G%

PAGE 93

R r S ZDYHOHQJWK QPf )LJ )DEU\3HURW WUDQVPLVVLRQ UHVSRQVH RI D SP ZLGH 3( FKDQQHO ZDYHJXLGH 3URSDJDWLRQ /RVV 3URSDJDWLRQ ORVV ZDV PHDVXUHG XVLQJ WKH ZDYHJXLGH )DEU\3HURW UHVRQDWRU WHFKn QLTXH >5HJ@ 7KH VHWXS LV WKH VDPH DV WKDW GHSLFWHG LQ )LJ +HUH KRZHYHU WKH VDPn SOH LWVHOI DFWV DV D )DEU\3HURW UHVRQDWRU 7KH PHDVXUHPHQW LV PDGH E\ FKDQJLQJ WKH ZDYHOHQJWK RI WKH ODVHU VRXUFH DQG PRQLWRULQJ WKH PHDVXUHG RSWLFDO RXWSXW IURP WKH ZDYHJXLGH 7KH RXWSXW DV D IXQFWLRQ RI ZDYHOHQJWK ZLOO KDYH D W\SLFDO )DEU\3HURW UHVSRQVH UHVHPEOLQJ VRPHZKDW D VLQXVRLGDO UHVSRQVH GHSHQGLQJ RQ WKH DPRXQW RI SURSDn JDWLRQ ORVV %\ PHDVXULQJ WKH UDWLR RI WKH PD[LPXP WUDQVPLVVLRQ WR WKH PLQLPXP WUDQVn PLVVLRQ DQG E\ NQRZLQJ WKH UHIOHFWLYLW\ RI WKH VDPSOH HQGIDFHV WKH SURSDJDWLRQ ORVV LV HDVLO\ FDOFXODWHG >5HJ@ 7KH PHDVXUHG )DEU\3HURW WUDQVPLVVLRQ UHVSRQVH RI D ZDYHJXLGH VDPSOH SUHSDUHG E\ 3( LQ JO\FHULQ LV VKRZQ LQ )LJ 7KH YDOXH RI SURSDJDWLRQ ORVV FDOFXODWHG IURP WKLV

PAGE 94

PHDVXUHPHQW LV G%FP FRPSDUDEOH WR YDOXHV UHSRUWHG IRU $3( ZDYHJXLGHV LQ /L1E >+XV@ )RU WKH FDVH RI =Q YDSRULQGLIIXVHG ZDYHJXLGHV D SURSDJDWLRQ ORVV RI G%FP ZDV PHDVXUHG FRPSDUDEOH WR WKDW SUHYLRXVO\ UHSRUWHG IRU WKHVH ZDYHJXLGHV LQ /L7D >(NQ@ $ORQJ ZLWK ORVV DQRWKHU LPSRUWDQW FKDUDFWHULVWLF RI ZDYHJXLGHV LV WKH YDOXH RI WKH HOHFWURRSWLF FRHIILFLHQW U ,Q WKH QH[W VHFWLRQ WKH PHWKRG RI PHDVXULQJ U LV RXWOLQHG DQG SHUIRUPHG U 0HDVXUHPHQW )RU WKH IDEULFDWLRQ RI HOHFWURRSWLF GHYLFHV VXFK DV PRGXODWRUV LW LV GHVLUDEOH WR KDYH D ODUJH HOHFWURRSWLF FRHIILFLHQW U +RZHYHU WKH IDEULFDWLRQ WHFKQLTXHV XVHG KHUH DUH NQRZQ WR GHFUHDVH VRPHZKDW WKH YDOXH RI U $V VWDWHG HDUOLHU 3( HVSHFLDOO\ LQ S\UR SKRVSKRULF DFLG LV NQRZQ WR GHFUHDVH U WKRXJK LW FDQ EH UHFRYHUHG WR VRPH GHJUHH E\ DQQHDOLQJ 7KH U YDOXH RI ZDYHJXLGHV IDEULFDWHG LQ ORZHU + FRQFHQWUDWLRQ VRXUFHV VXFK DV JO\FHULQ ZKHUH DQQHDOLQJ LV QRW XVHG LV QRW NQRZQ $GGLWLRQDOO\ WKH WHFKQLTXHV RI 7LLQGLIIXVLRQ DQG =Q YDSRULQGLIIXVLRQ DUH SHUIRUPHG DW WHPSHUDWXUHV DERYH WKH &XULH SRLQW IRU /L7D VR WKHUH LV QR HOHFWURRSWLF HIIHFW DIWHU ZDYHJXLGH IRUPDWLRQ $V D UHVXOW UHSROLQJ LV QHHGHG WR UHVWRUH U $ GLUHFW PHDVXUHPHQW RI U LV WKHQ QHHGHG WR JDXJH WKH HIIHFWLYHQHVV RI WKH UHSROLQJ SURFHVV ,Q WKLV VHFWLRQ WKH WHFKQLTXH XVHG WR PHDVXUH WKH HOHFWURRSWLF FRHIILFLHQW U LV GHVFULEHG DQG DGPLQLVWHUHG 7KH DGYDQWDJH RI WKH WHFKQLTXH XVHG KHUH RYHU RWKHU PHWKRGV LV RXWOLQHG DQG VRPH UHVXOWV DUH SUHVHQWHG

PAGE 95

,QWHUIHURPHWULF 0HDVXUHPHQW 0HWKRG 0RVW HIIRUWV WR GLUHFWO\ PHDVXUH U LQ FKDQQHO ZDYHJXLGHV LQYROYH WKH XVH RI DQ LQWHUIHURPHWULF GHYLFH VXFK DV D 0DFK=HKQGHU LQWHUIHURPHWHU 0=,f DQG WKH DSSOLFDWLRQ RI DQ HOHFWURGH SDWWHUQ WR LQGXFH D SKDVH VKLIW )RU H[DPSOH WKH U FRHIILFLHQW IRU DQ 0=, LQ SXVKSXOO FRQILJXUDWLRQ LV JLYHQ E\ U f WU9/Q ZKHUH $ LV WKH ZDYHOHQJWK J LV WKH HOHFWURGH JDS @f LV WKH LQGXFHG SKDVH VKLIW 9 LV WKH RSWLn FDOHOHFWULFDO RYHUODS IDFWRU 9 LV WKH DSSOLHG YROWDJH / LV WKH HOHFWURGH LQWHUDFWLRQ OHQJWK DQG Q LV WKH UHIUDFWLYH LQGH[ 7KH DFFXUDF\ RI WKLV WHFKQLTXH GHSHQGV JUHDWO\ RQ WKH DFFXn UDF\ LQ GHWHUPLQLQJ WKH YDOXH RI 7 ZKLFK FDQ RQO\ EH HVWLPDWHG EDVHG RQ HOHFWURGH GLPHQn VLRQV +RZHYHU WKHUH LV DQRWKHU PHWKRG ZKLFK HQVXUHV DQ RYHUODS IDFWRU RI XQLW\ 7KLV PHWKRG LV GHVFULEHG QH[W 8QLW\ 2YHUODS 0HWKRG 7KH VHWXS XVHG KHUH WR PHDVXUH U LV GHSLFWHG LQ )LJ 7KH SRODUL]DWLRQPDLQn WDLQLQJ ILEHU IURP D GLRGH ODVHU LV FRQQHFWHG WR D SRODUL]DWLRQPDLQWDLQLQJ ILEHU FRXSOHU ,W ZDV D G% FRXSOHU IRU WKH ZDYHOHQJWK XVHG _LP 2QH RXWSXW IURP WKH FRXSOHU LV XVHG [ )LJ ([SHULPHQWDO VHWXS XVHG WR PHDVXUH WKH HOHFWURRSWLF FRHIILFLHQW U

PAGE 96

WR FRXSOH OLJKW LQWR D ZDYHJXLGH VDPSOH 7KHQ WKH RXWSXWV RI ERWK WKH RWKHU FRXSOHU DUP DQG WKH ZDYHJXLGH DUH LPDJHG RQ WRS RI HDFK RWKHU VR WKH\ LQWHUIHUH RQ DQ ,5 FDPHUD 7KH LQWHUIHUHQFH SDWWHUQ LV VHHQ RQ D YLGHR PRQLWRU 7KH ZDYHJXLGH VDPSOH KDG 7D GHSRVn LWHG RQ ERWK WKH ZDYHJXLGH VXUIDFH ZLWK D 6L EXIIHU OD\HU WR UHGXFH ORVV DQG WKH ERWWRP VXUIDFH 7KH 7D VHUYHG DV HOHFWURGHV DQG JXDUDQWHHG DQ RYHUODS IDFWRU RI XQLW\ 2I FRXUVH WKH OLPLWDWLRQ RI WKLV WHFKQLTXH LV WKDW LW FDQ RQO\ EH XVHG RQ =FXW VDPSOHV 7R PDNH WKH PHDVXUHPHQW D YROWDJH ZDV DSSOLHG WR WKH HOHFWURGHV DQG WKH IULQJH SDWWHUQ ZDV REVHUYHG RQ WKH PRQLWRU 7KH YROWDJH 9 UHTXLUHG WR VKLIW WKH IULQJH SDWWHUQ E\ RQHKDOI SHULRG _f Qf ZDV UHFRUGHG DQG WKH U YDOXH ZDV FDOFXODWHG IURP U f B 9/Q f ZKHUH J LV WKH VDPSOH WKLFNQHVV DQG / LV WKH VDPSOH OHQJWK 7KH DFFXUDF\ RI WKLV WHFKQLTXH ZDV HVWLPDWHG WR EH b )RU D ZDYHJXLGH IDEULFDWHG E\ 3( LQ JO\FHULQ DW & IRU KRXUV WKH PHDVXUHG U YDOXH ZDV SP9 )RU D =Q YDSRULQGLIIXVHG VDPSOH ZKLFK ZDV UHSROHG WKH PHDn VXUHG U YDOXH ZDV SP9 7KLV YDOXH LV FORVH WR WKH U YDOXH RI SP9 IRU EXON YLUn JLQ /L7D ,Q WKH QH[W VHFWLRQ WKH LVVXH RI SRZHU KDQGOLQJ RI /L7D ZDYHJXLGHV LV H[DPn LQHG ,Q SDUWLFXODU WKH SKRWRUHIUDFWLYH FKDUDFWHULVWLF RI 3( ZDYHJXLGHV LQ WKH LQIUDUHG QHDU _LP LV PHDVXUHG IRU VLPLODU VWUXFWXUHV LQ ERWK /L7D DQG /L1E 3RZHU +DQGOLQJ /L7D LV NQRZQ WR KDYH D WKUHVKROG IRU SKRWRUHIUDFWLYH GDPDJH RI PRUH WKDQ DQ RUGHU RI PDJQLWXGH ODUJHU WKDQ WKDW RI /L1E DW YLVLEOH ZDYHOHQJWKV >*OD@ ,Q WKH

PAGE 97

LQIUDUHG KRZHYHU DOWKRXJK SKRWRUHIUDFWLYH GDPDJH LV RIWHQ FRQVLGHUHG QHJOLJLEOH GDPn DJH KDV EHHQ REVHUYHG DW _LP LQ 7L/L1E IRU SRZHU OHYHOV DURXQG P: >+DU@ 7KH SKRWRUHIUDFWLYH UHVLVWDQFH RI WKHVH ZDYHJXLGHV DV ZHOO DV WKRVH IDEULFDWHG E\ 3( LQ ERWK /L1E DQG /L7D DW _LP KDYH QRW EHHQ LQYHVWLJDWHG WR GDWH ,Q WKLV VHFWLRQ WKH SRZHU KDQGOLQJ FDSDELOLW\ RI 3( ZDYHJXLGHV LQ /L7D RSHUDWn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n OHFW WKH OLJKW IURP WKH H[SDQGHG PRGH RI D ZDYHJXLGH H[KLELWLQJ SKRWRUHIUDFWLYH GDPDJH DW KLJKHU LQWHQVLWLHV 7KH RXWSXW SRZHU YHUVXV LQSXW SRZHU ZDV WKHQ SORWWHG WR FKHFN IRU SKRWRUHIUDFWLYH GDPDJH 5HVXOWV )LJ VKRZV WKH UHVXOWV RI SKRWRUHIUDFWLYH PHDVXUHPHQWV SHUIRUPHG RQ ZDYHJXLGHV RI VLPLODU FKDQQHO ZLGWK LQ /L7D DQG /L1E $V FDQ EH VHHQ WKH /L7D VDPSOH VKRZV D UHDVRQDEOH GHJUHH RI OLQHDULW\ EHWZHHQ RXWSXW DQG LQSXW SRZHU RYHU WKH

PAGE 98

f§,f§,f§ f§f§f§f§f§f§f§f§ f§f§f§ f§f§f§f§f 2 e 3(/L7D r D LRR f f 6 f f R f f t& f A 2K r $3(/L1E  3 2 ‘ LBOf§Of§Of§Of§f§Of§Of§, f§Of§,f§W OOf§Lf§Mf§, f§, ,f§Lf§ 2 LQSXW SRZHU P:f )LJ 2XWSXW SRZHU YV LQSXW SRZHU IRU SP ZLGH VWUDLJKWFKDQn QHO ZDYHJXLGHV LQ 3(/L7D DQG $3(/L1E ODVLQJ WKUHVKROG RI WKH ODVHU XVHG LQ WKLV UDQJH WHVWHG 7KHUHIRUH OLWWOH LI DQ\ GHJUDn GDWLRQ WR LWV SHUIRUPDQFH KDV EHHQ FDXVHG E\ SKRWRUHIUDFWLYH GDPDJH 7KH /L1E VDPSOH KRZHYHU VKRZV VRPH GHJUHH RI VDWXUDWLRQ DQG DSSHDUV WR KDYH DOUHDG\ SDVVHG LWV WKUHVKROG SRLQW IRU SKRWRUHIUDFn WLYH GDPDJH IRU HYHQ WKH VPDOOHVW LQSXW SRZHU WHVWHG /RZHU OHYHOV RI LQSXW SRZHU ZHUH QRW DWWDLQDEOH DV WKH\ ZHUH EHORZ WKH ,PHQW 7KH VLJQLILFDQW DWWHQXDWLRQ RI RXWSXW SRZHU UHODWLYH WR LQSXW SRZHU LOOXVWUDWHV WKH H[WUHPHO\ KLJK SKRWRLQGXFHG UHIUDFWLYH LQGH[ FKDQJH LQ $3(/L1E ZDYHJXLGHV HVSHFLDOO\ DERYH O22P: ,Q WKH QH[W VHFWLRQ D GLUHFW PHWKRG RI LQGH[ SURILOLQJ LV GHVFULEHG 6XFK D PHDn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

PAGE 99

$3(/L7D LV QHHGHG 7KLV GLDJUDP UHODWHV ZDYHJXLGH VXUIDFH LQGH[ LQFUHPHQW WR SURWRQLQGXFHG ODWWLFH VWUDLQ DW WKH VXUIDFH ,Q WKLV VHFWLRQ WKH PHDVXUHPHQW WHFKQLTXH XVHG WR H[WUDFW ZDYHJXLGH LQGH[ SURn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f SURFHGXUH DQG KHQFH ZHUH DIIHFWHG E\ WKH IXQGDPHQWDO DVVXPSWLRQ RI D PRQRWRQLFDOO\ GHFUHDVLQJ GHSWK SURILOH ZLWK WKH PD[LPXP UHDFKHG DW WKH VXEVWUDWH VXUIDFH 2Q WKH FRQWUDU\ UHFHQW UHVXOWV KDYH GLUHFWO\ GHPRQVWUDWHG >0DUD@ WKH H[LVWHQFH RI EXULHG SURILOHV ZKHUH WKH LQGH[ SURILOH PD[LPXP LV UHDFKHG DW D SRLQW EHORZ WKH VXEVWUDWH VXUIDFH 7KLV IDFW PDNHV WKH DFFXUDF\ RI WKH SUHYLRXV UHVXOWV OHVV WUXVWZRUWK\ WR VD\ WKH OHDVW $GGLWLRQDOO\ HYHQ LQ WKH FDVH RI PRQRWRQLFDOO\ GHFUHDVLQJ LQGH[ SURILOHV

PAGE 100

)LJ &RPSDULVRQ RI D GLUHFWO\ REVHUYHG LQGH[ SURILOH VROLGf DQG DQ ,:.% UHFRQVWUXFWHG SURILOH GRWWHGf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f DUH FRPSDUHG ZLWK WKH GLUHFWO\ PHDVXUHG LQGH[ SURILOH VROLG OLQHf 7KH PHDVXUHG PRGH LQGLFHV DUH ZLWKLQ PHDVXUHPHQW HUURU IRU WKH SULVP FRXSOLQJ WHFKQLTXH KRZHYHU WKH ,:.% UHFRQVWUXFWLRQ PHWKRG FRXOG QRW DFFXUDWHO\ SUHGLFW WKH SODWHDX DOPRVWf UHJLRQ QHDU WKH VXEVWUDWH VXUIDFH DQG KHQFH VHULRXVO\ RYHUHVWLPDWHG WKH VXUIDFH LQGH[ LQFUHPHQW )LQDOO\ WKH ,:.% PHWKRG UHTXLUHV DW OHDVW WKUHH JXLGHG PRGHV DQG KHQFH LV QRW DSSOLFDEOH DW DOO WR WKH FDVH RI VLQJOHPRGH ZDYHJXLGHV &OHDUO\ H[WUDSRODWLRQ RI WKH PXOWLPRGH ZDYHJXLGH GDWD ZKRVH DFFXUDF\ LV UDWKHU TXHVWLRQDEOH WR VLQJOHPRGH VWUXFWXUHV LV HYHQ PRUH GRXEWIXO 2Q WKH RWKHU KDQG WKH VLQJOHPRGH UHJLPH RI RSHUDWLRQ

PAGE 101

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n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n WRULQJ WKH FKDQJH LQ LQWHQVLW\ RI OLJKW UHIOHFWHG IURP WKH VDPSOH VXUIDFH 7KH UHODWLRQVKLS EHWZHHQ WKH YDULDWLRQ RI UHIOHFWHG OLJKW LQWHQVLW\ DQG LQGH[ FKDQJH LV GHULYHG DV IROORZV 7KH LQWHQVLW\ RI OLJKW UHIOHFWHG IURP WKH VXUIDFH RI WKH VXEVWUDWH LV UHODWHG WR WKH LQFLGHQW LQWHQVLW\ ,D E\

PAGE 102

)LJ 5HIOHFWLYLW\ SURILOLQJ VHWXS XVHG IRU PHDVXULQJ WKH UHIUDFWLYH LQGH[ SURILOHV RI $3( ZDYHJXLGHV LQ /L7D

PAGE 103

U ?Q? r f ZKHUH Q LV WKH VXEVWUDWH ORFDO UHIUDFWLYH LQGH[ HLWKHU RUGLQDU\ RU H[WUDRUGLQDU\ %\ WDNLQJ WKH GHULYDWLYH RI WKLV HTXDWLRQ ZLWK UHVSHFW WR Q WKH FKDQJH LQ UHIOHFWHG LQWHQVLW\ FDQ EH UHODWHG WR WKH LQFLGHQW LQWHQVLW\ DFFRUGLQJ WR G,U GQ BU Qf Q f R f 1RZ VLQFH ZKDW ZDV DFWXDOO\ PHDVXUHG ZDV WKH UHIOHFWHG LQWHQVLW\ ,U DQG WKH FKDQJH LQ UHIOHFWHG LQWHQVLW\ GOU E\ VXEVWLWXWLQJ (T f LQWR (T f RQH REWDLQV GO GQ Q ? Q f§ f Q 5HDUUDQJLQJ WKLV HTXDWLRQ DQG GURSSLQJ VXEVFULSWV D UHODWLRQVKLS EHWZHHQ WKH FKDQJH LQ LQGH[ DQG WKH FRUUHVSRQGLQJ FKDQJH ƒ, LQ UHIOHFWHG LQWHQVLW\ LV GHULYHG $Q ^QE QM Q 9`$> En f ZKHUH $Q LV WKH LQGH[ FKDQJH DQG LV WKH EXON LQGH[ YDOXH HLWKHU RUGLQDU\ RU H[WUDRUGLn QDU\ ([DFW EXON LQGH[ YDOXHV RI Q DQG QH DW WKH ZDYHOHQJWK RI LQWHUHVW QP ZHUH IRXQG E\ FXELF VSOLQH LQWHUSRODWLRQ RI WKH GLVSHUVLRQ GDWD IRU /L7D SUHVHQWHG LQ $SSHQn GL[ % 7KHVH YDOXHV ZHUH Q DQG QH $Q ;FXW FU\VWDO KDG WR EH XVHG LQ RUGHU WR EH DEOH WR VHH WKH H[WUDRUGLQDU\ LQGH[ QH FKDUDFWHUL]HG E\ WKH LQGH[ DORQJ WKH = D[LV $ SRODUL]HU DQG KDOIZDYH SODWH ZHUH XVHG DW WKH ODVHU RXWSXW WR VHOHFW WKH GHVLUHG

PAGE 104

FU\VWDO SRODUL]DWLRQ IRU SURILOLQJ 7KH VXEVWUDWH ZDV EHYHOSROLVKHG DW DQ DQJOH D LQ RUGHU WR PDJQLI\ WKH GHSWK SURILOHV E\ D IDFWRU RI VLQD 7KH H[DFW DQJOH RI WKH EHYHO ZDV D r PDNLQJ WKH PDJQLILFDWLRQ IDFWRU $ 7D ILOP GHSRVLWHG RQ WKH VXEVWUDWH VXUIDFH SULRU WR EHYHOOLQJ VHUYHG DV D UHIHUHQFH IRU PDUNLQJ WKH SRVLWLRQ RI WKH VXEVWUDWH VXUIDFH E\ FDXVLQJ D VKDUS SHDN LQ UHIOHFWHG LQWHQVLW\ $Q $5 FRDWLQJ 0J) ƒ WKLFNf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ff WR WKH FHQWHU FRQH RI D f DXGLR VSHDNHU 7KH VDPSOH ZDV VWXFN WR WKH URG XVLQJ GRXEOHVLGHG WDSH DQG WKH URG DWWDFKHG WR WKH VSHDNHU FRQH XVLQJ KRW JOXH $ '&UDPS VLJQDO GURYH WKH VSHDNHU FDXVLQJ LW WR RVFLOODWH OLQHDUO\ DQG WKXV VFDQ WKH VDPSOH EDFNDQGIRUWK XQGHU WKH VSRW 7KLV UDPS VLJQDO ZDV DOVR XVHG WR WULJJHU WKH RVFLOORVFRSH 7KH DPRXQW RI VSHDNHU H[FXUVLRQ DQG WKXV VFDQQLQJ GLVWDQFH ZDV FRQWUROOHG E\ WKH SHDNWRSHDN YROWDJH RI WKH LQSXW UDPS VLJQDO WR WKH VSHDNHU $ FDOn LEUDWLRQ FKDUW RI VSHDNHU LQSXW YROWDJH YHUVXV VSHDNHU H[FXUVLRQ GLVWDQFH ZDV FUHDWHG E\ LGHQWLI\LQJ WKH YROWDJH QHFHVVDU\ WR FDXVH HQRXJK H[FXUVLRQ WR GLVSOD\ RQ WKH RVFLOORVFRSH WKH LQWHQVLW\ SHDNV RI WKLQ PHWDO VWULSHV IDEULFDWHG RQ D SDUWLFXODU VDPSOH ZLWK D NQRZQ VHSDUDWLRQ EHWZHHQ WKHP 7KLV FDOLEUDWLRQ FKDUW LV GHSLFWHG LQ )LJ 7KH HQWLUH PHDn VXUHPHQW SURFHVV LV WKHQ SHUIRUPHG DV IROORZV 7KH RVFLOORVFRSH LV XVHG WR GLVSOD\ WKH FKDQJH LQ UHIOHFWHG LQWHQVLW\ DSSHDULQJ DV D FKDQJH LQ YROWDJH $9 7KH VFDQQLQJ GLVWDQFH

PAGE 105

VFDQQLQJ GLVWDQFH MDPf )LJ 6SHDNHU FDOLEUDWLRQ FKDUW UHODWLQJ YROWDJH WR VSHDNHU H[FXUVLRQ 7KH PHDVXUHG GDWD VTXDUHVf DQG OLQHDU ILW VROLGf DUH VKRZQ LV DGMXVWHG WR UHYHDO WKH HQWLUH SURILOH XS WR WKH SRLQW RI WKH ODUJH LQWHQVLW\ RU YROWDJH LQFUHDVH PDUNLQJ WKH VXEVWUDWH VXUIDFH 7KLV VFDQ LV VDYHG DQG VHQW WR D FRPSXWHU ZKHUH WKH GHSWK D[LV LV FDOLEUDWHG E\ NQRZn LQJ WKH SHDNWRSHDN YROWDJH LQSXW WR WKH VSHDNHU DQG XVLQJ WKH FDOLEUDWLRQ FKDUW WR GHWHUPLQH VFDQQLQJ GLVWDQFH LQ PLFURQV 7KH YHUWLFDO D[LV RU $Q LV FDOLEUDWHG E\ PHDVXULQJ WKH LQWHQVLW\ RU DVVRFLDWHG YROWDJH 9 FDXVHG E\ EXON UHIOHFWLRQ DQG XVLQJ (T f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a _LP

PAGE 106

)LJ 6FDQ RI IRFXVHG VSRW DV LW WUDYn HOV DFURVV D VWHSLQWHUIDFH IURP EXON /L7D WR KLJKO\ UHIOHFWLYH PHWDO GHSWK SPf )LJ 2UGLQDU\ LQGH[ SURILOH WDNHQ DW ILYH GLIIHUHQW SRLQWV DFURVV D VDPSOH %\ REVHUYLQJ (T f LW LV VHHQ WKDW WKH DFFXUDF\ LQ PHDVXULQJ LQGH[ FKDQJHV LV DERXW WKH VDPH DV WKH DFFXUDF\ LQ PHDVXULQJ $9,9 7KLV PHWKRG LV H[SHFWHG WR EH FDSDEOH RI UHSURGXFLQJ DFWXDO UHIUDFWLYH LQGH[ SURILOHV ZLWK DQ DFFXUDF\ RI $QQ n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r& IRU PLQXWHV WKHQ DQQHDOHG DW r& IRU YDULRXV WLPHV UDQJLQJ IURP WR PLQXWHV

PAGE 107

GHSWK _LPf )LJ ([WUDRUGLQDU\ LQGH[ SURILOHV IRU YDULRXV DQQHDOLQJ WLPHV $OO SURn ILOHV ZHUH PHDVXUHG DW "L SP H[FHSW WKDW VKRZQ E\ WKH GDVKHGGRWWHG OLQH ZKLFK ZDV PHDVXUHG DW K SP PLQ DQQHDOLQJ WLPHf 7KH DFWXDO PHDVXUHG H[WUDRUGLQDU\ LQGH[ SURILOH ZLWK ULSSOHV IRU PLQ DQQHDOLQJ PHDVXUHG DW fL _LPf LV VKRZQ E\ WKH GDVKHG OLQH ,W VKRXOG EH QRWHG WKDW WKH PHDVXUHG FXUYHV H[KLELWHG ULSSOHV VKRZQ E\ WKH GDVKHG OLQH RQO\ IRU D VDPSOH DQQHDOHG IRU WD PLQf ZKRVH RULJLQ QHHGV IXUWKHU FODULILFDWLRQ 7DNHQ DORQH SRVVLEOH LQWHUIHUHQFH EHWZHHQ WKH EHDPV UHIOHFWHG IURP WKH EHYHOOHG DQG ERWWRP VXUIDFHV RU D VXUIDFH SURILOH GXH WR SROLVKLQJ FDQQRW H[SODLQ WKLV SKHQRPHQRQ VLQFH YLUWXDOO\ QR ULSSOHV ZHUH REVHUYHG EH\RQG WKH ZDYHJXLGH FURVVVHFWLRQ DUHD )XUWKHUPRUH WKH PDJQLWXGH RI WKHVH ULSSOHV YDULHV GHFUHDVHVf ZLWK DQQHDOLQJ WLPH

PAGE 108

)RU LOOXVWUDWLRQFODULW\ DOO WKH PHDVXUHG SURILOHV WD WR PLQf H[FHSW IRU WKH DVH[FKDQJHG ZDYHJXLGH ZHUH ILWWHG ZLWK VSOLQHG VPRRWK FXUYHV ZKLFK DFFHQWXDWH SURILOH HYROXWLRQ XSRQ DQQHDOLQJ ,QGH[ FKDQJHV PHDVXUHG RQ DVH[FKDQJHG ZDYHJXLGHV ZHUH UDWKHU ORZ DQG LQ DGGLWLRQ ZHUH VLJQLILFDQWO\ REVFXUHG E\ WKH DIRUHPHQWLRQHG ULSSOHV SUHYHQWLQJ DQ\ ILW WR WKH PHDVXUHG FXUYH 7KH PD[LPXP LQGH[ LQFUHPHQW KRZHYHU GUDPDWLFDOO\ LQFUHDVHV XSRQ VKRUW DQQHDOLQJ SHULRGV XS WR IRU PLQ DQQHDOLQJ WLPHf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f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

PAGE 109

GHSWK SPf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

PAGE 110

$QSPf$Q_OPf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f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aN IURP DQ ;UD\ VRXUFH LV LQFLGHQW XSRQ WKH VXUIDFH RI WKH VDPSOH $ GHWHFWRU LV XVHG WR PHDVXUH WKH LQWHQVLW\ RI WKH UHIOHFWHG UDGLDWLRQ 7KLV UHIOHFWHG LQWHQVLW\ LV PD[LPXP ZKHQ

PAGE 111

)LJ 6FKHPDWLF RI WKH ;UD\ GLIIUDFn WLRQ PHDVXUHPHQW SURFHVV )LJ ,OOXVWUDWLRQ RI DQ H[SHFWHG URFNLQJ FXUYH IRU D 3( VDPSOH WKH DQJOH RI LQFLGHQFH LV VXFK WKDW %UDJJfV /DZ LV VDWLVILHG Q"L GVLQ f %\ VORZO\ URWDWLQJ RU URFNLQJ WKH VDPSOH GXULQJ D PHDVXUHPHQW UXQ WKH DQJOH RI LQFLn GHQFH FDQ EH FKDQJHG DQG WKH FRUUHVSRQGLQJ UHIOHFWLRQ LQWHQVLW\ PRQLWRUHG 6XFK D SURn FHVV LV UHIHUUHG WR DV D URFNLQJ FXUYH PHDVXUHPHQW DQG WKH UHVXOWLQJ GDWD RI LQWHQVLW\ YHUVXV LQFLGHQW DQJOH LV D URFNLQJ FXUYH )RU WKH FDVH RI SURWRQ H[FKDQJHG ZDYHJXLGHV WKH LQWURGXFWLRQ RI + LQWR WKH FU\Vn WDO LQFUHDVHV WKH ODWWLFH VSDFLQJ G $V D UHVXOW WKH DQJOH DW ZKLFK WKH %UDJJ UHIOHFWLRQ SHDN RFFXUV LV VKLIWHG WR D VPDOOHU DQJXODU YDOXH 6R D URFNLQJ FXUYH PHDVXUHPHQW RI VXFK D VDPSOH ZLOO H[KLELW WZR SHDNV DV LOOXVWUDWHG LQ )LJ RQH FRUUHVSRQGLQJ WR WKH EXON VXEVWUDWH DQG RQH FRUUHVSRQGLQJ WR WKH SURWRQ H[FKDQJHG UHJLRQ

PAGE 112

&DOFXODWLQJ /DWWLFH 6WUDLQ 3UHYLRXV VWXGLHV >)HG@ KDYH GHWHUPLQHG WKDW +/L7D ZDYHJXLGHV RQ ;FXW VXEVWUDWHV KDYH RQO\ RQH QRQ]HUR FRPSRQHQW H RI WKH GHIRUPDWLRQ WHQVHU 7KLV YDOXH RI VWUDLQ ZDV GHWHUPLQHG IRU HDFK VDPSOH E\ PHDVXULQJ URFNLQJ FXUYHV ZLWK D VLQJOH DQG GRXn EOHFU\VWDO ;UD\ GLIIUDFWRPHWHU %\ PHDVXULQJ WKH DQJXODU GHYLDWLRQ $-: EHWZHHQ WKH SHDN IRU WKH XQH[FKDQJHG VXEVWUDWH DQG WKDW IRU WKH H[FKDQJHG OD\HU RQ WKH URFNLQJ FXUYH RI WKH SODQH KNOf Hn FDQ EH IRXQG IURP WKH H[SUHVVLRQ >)HG@ $ WDQ f KNO ZKHUH "Z LV WKH %UDJJ DQJOH RI WKH VWUDLQIUHH VXEVWUDWH UHJLRQ )RU ;FXW VDPSOHV URFNn LQJ FXUYHV ZHUH PHDVXUHG IRU WKH f VXUIDFH SODQH ,Q PHDVXUHPHQWV WKH PD[LPXP DQJXODU VKLIW RQ WKH URFNLQJ FXUYH ZDV RI SULPDU\ LQWHUHVW DV LW FRUUHVSRQGV WR WKH PD[Ln PXP GHIRUPDWLRQ H DQG VWUDLQ LQ WKH $3( UHJLRQ ,W ZDV FRQILUPHG WKDW WKH PD[LPXP VKLIW FRUUHVSRQGV WR WKH VWUDLQ DW WKH VXEVWUDWH VXUIDFH 7KLV ZDV GRQH E\ VXFFHVVLYH SROn LVKLQJ RI WKH VDPSOH VXUIDFH LQ VPDOO VWHSV WKURXJK WKH $3( UHJLRQ DQG DQDO\]LQJ WKH UHVXOWLQJ FKDQJH LQ WKH URFNLQJ FXUYH /LNHZLVH WKH DQDO\VLV RI WKH HYROXWLRQ RI WKH URFNn LQJ FXUYH XSRQ DQQHDOLQJ KDV DOVR SRLQWHG WR D PRQRWRQLF VWUDLQ SURILOH ZLWK WKH PD[LPXP UHDFKHG DW WKH VXUIDFH 7KHUHIRUH WKH YDOXHV RI ERWK PD[LPXP GHIRUPDWLRQ Hf DQG RSWLn FDO LQGH[ XVHG WR SORW WKH VWUXFWXUDO SKDVH GLDJUDP &KDSWHU f KDYH EHHQ PHDVXUHG DW WKH VDPH SRLQW QDPHO\ WKH VXEVWUDWH VXUIDFH 6RPH 5RFNLQJ &XUYH 5HVXOWV 6RPH PHDVXUHG URFNLQJ FXUYHV DUH VKRZQ LQ )LJ )LJ Df LV WKH URFNLQJ FXUYH IRU YLUJLQ /L7D 7KHUH DUH WZR SHDNV IRU WKH VXEVWUDWH EHFDXVH WKLV FXUYH ZDV WDNHQ ZLWK D VLQJOHFU\VWDO GLIIUDFWRPHWHU KDYLQJ WZR ZDYHOHQJWKV +DG D GRXEOHFU\VWDO

PAGE 113

Df Ef )LJ 0HDVXUHG URFNLQJ FXUYHV IRU ;FXW YLUJLQ /L7D Df DQG 3( /L7D Ef GLIIUDFWRPHWHU EHHQ XVHG RQO\ RQH ZDYHOHQJWK ZRXOG KDYH EHHQ SUHVHQW DQG RQO\ RQH SHDN WKH OHIWPRVW RQH ZRXOG H[LVW RQ WKH URFNLQJ FXUYH 7KH PHDVXUHG FXUYH LV SORWWHG DV D IXQFWLRQ RI WZLFH WKH LQFLGHQW DQJOH 7KH DQJOH IRU WKH VXEVWUDWH SHDN RI LQWHUHVW LV ODEHOHG 1RWH WKDW WKHVH VDPSOHV ZHUH ;FXW )RU ;FXW WKH DQJOH IRU WKH EXON VXEVWUDWH E LV DURXQG r ZKLOH LW LV DURXQG IRU r =FXW )LJ Ef LV WKH URFNLQJ FXUYH IRU D 3( VDPSOH SUHSDUHG LQ JO\FHULQ DW r& IRU KRXUV )RU WKLV FXUYH WKH VXEVWUDWH SHDN DQG WKH SHDN FRUUHVSRQGLQJ WR WKH 3( OD\HU IRU WKH ZDYHOHQJWK RI LQWHUHVW DUH ODEHOHG +DYLQJ WKH PHDVXUHG YDOXHV IRU E DQG $ E3( WKH YDOXH RI H FDQ WKHQ EH FDOFXODWHG IURP (T f DQG ZDV IRXQG WR EH [ 2n 6XPPDU\ 7KLV FKDSWHU SUHVHQWHG WKH FKDUDFWHUL]DWLRQ RI SODQDU DQG FKDQQHO ZDYHJXLGHV LQ /L7D IDEULFDWHG E\ $3( 7LLQGLIIXVLRQ DQG =Q YDSRULQGLIIXVLRQ 7KH QXPEHU RI

PAGE 114

PRGHV VXSSRUWHG E\ FKDQQHOV RI YDULRXV ZLGWKV ZHUH ILUVW GHWHUPLQHG E\ D QHDUILHOG PHDn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n WHULVWLF RI 3( ZDYHJXLGHV LQ /L7D DQG /L1E ZHUH PHDVXUHG QHDU SP 7KH /L1E ZDYHJXLGHV H[KLELWHG VLJQLILFDQW GDPDJH DW SRZHU OHYHOV DV ORZ DV P: ZKLOH WKH /L7D ZDYHJXLGHV GHPRQVWUDWHG GHFHQW SRZHU KDQGOLQJ WR OHYHOV DV KLJK DV P: )LQDOO\ WKH WHFKQLTXHV RI GLUHFW LQGH[ SURILOLQJ DQG URFNLQJ FXUYH DQDO\VLV ZHUH RXWOLQHG IRU WKHLU XVH LQ GHWHUPLQLQJ VXUIDFH YDOXHV RI LQGH[ LQFUHPHQW $Q DQG ODWWLFH VWUDLQ H XVHG LQ WKH QH[W FKDSWHU WR UHFRQVWUXFW WKH VWUXFWXUDO SKDVH GLDJUDP IRU $3(/L7D

PAGE 115

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ƒKOF@ 7KHVH SHFXOLDULWLHV VXJJHVW D QRQn OLQHDU GHSHQGHQFH RI SURWRQ FRQFHQWUDWLRQ RQ UHIUDFWLYH LQGH[ ZKLFK PD\ UHVXOW LQ EXULHG LQGH[ SURILOHV IRU FHUWDLQ IDEULFDWLRQ FRQGLWLRQV DV ZDV FRQILUPHG LQ WKH SUHYLRXV FKDSWHU 7R H[SODLQ WKH REVHUYHG DEQRUPDO EHKDYLRU LW LV WKHUHIRUH LPSRUWDQW WR XQGHUVWDQG WKH UHODWLRQVKLS EHWZHHQ SURWRQ FRQFHQWUDWLRQ DQG LQGH[ FKDQJH 6LQFH WKH LQWURGXFWLRQ RI SURWRQV FDXVHV FU\VWDO GHIRUPDWLRQ DQG VWUDLQ WKH UHODWLRQVKLS EHWZHHQ SURWRQLQGXFHG ODWn WLFH VWUDLQ DQG LQGH[ FKDQJH LV HTXDOO\ DV LPSRUWDQW 6XFK D UHODWLRQ KDV EHFRPH LGHQWLILHG DV WKH VWUXFWXUDO SKDVH GLDJUDP >.RU@ QRW WR EH FRQIXVHG ZLWK WKH SKDVH GLDJUDPV XVHG LQ SK\VLFDO FKHPLVWU\ WR UHODWH WHPSHUDWXUH WR SURWRQ FRQFHQWUDWLRQ DQG LV LOOXVWUDWHG LQ )LJ ,Q JHQHUDO WKLV GLDJUDP LV XVXDOO\ SORWWHG XVLQJ ODWWLFH VWUDLQ Hf UDWKHU WKDQ + SURWRQ FRQFHQWUDWLRQ DV WKH H PHDVXUHPHQW LV QRQGHVWUXFWLYH DQG VLPSOHU WR PDNH .QRZOHGJH

PAGE 116

RI WKH VWUXFWXUDO SKDVH GLDJUDP IRU $3(/L7D LV LQGLVSHQVDEOH LQ GHYHORSLQJ DGYDQFHG KLJKWKURXJKSXW HOHFWURRSWLF DQG QRQOLQHDU JXLGHGZDYH GHYLFHV ZLWK ORZ LQVHUWLRQ ORVV DQG WHPSRUDOO\ VWDEOH SHUIRUn PDQFH ,Q WKLV FKDSWHU WKH VWUXFWXUDO SKDVH GLDJUDP IRU $3(/L7D LV UHFRQVWUXFWHG 6RPH OLPLWDWLRQV WKDW KDYH KDPSHUHG SUHYLn RXV HIIRUWV WR FRQVWUXFW WKH GLDJUDP DUH ILUVW GLVFXVVHG 7KHQ GLUHFW PHDVXUHPHQW WHFKn QLTXHV RXWOLQHG LQ WKH SUHYLRXV FKDSWHU DUH HPSOR\HG IRU WKH VXFFHVVIXO UHFRQVWUXFWLRQ RI WKH VWUXFWXUDO SKDVH GLDJUDP 7KH GLDJUDP LV WKHQ XVHG WR H[SODLQ WKH SUHYLRXVO\ PHQWLRQHG DQRPDOLHV DVVRFLDWHG ZLWK WKH $3( SURFHVV LQ /L7D )LQDOO\ 3( ZDYHJXLGHV DUH DQDn O\]HG IRU LQFOXVLRQ LQ WKH VWUXFWXUDO SKDVH GLDJUDP 'LUHFW PHDVXUHPHQW WHFKQLTXHV DQG 5DPDQ VSHFWUD RI WKHVH VDPSOHV DUH XVHG WR LGHQWLI\ D QHZ VLQJOHSKDVH UHJLRQ RI WKH VWUXFn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

PAGE 117

3UHYLRXV /LPLWDWLRQV ,Q WKLV VHFWLRQ WKH OLPLWDWLRQV LQKLELWLQJ SUHYLRXV HIIRUWV IRU FRPSOHWH UHFRQVWUXFn WLRQ RI WKH VWUXFWXUDO SKDVH GLDJUDP IRU $3(/L7D DUH GHVFULEHG 7KH QHHG IRU GLUHFW PHDVXUHPHQW WHFKQLTXHV WR RYHUFRPH WKHVH OLPLWDWLRQV LV GHYHORSHG LQYROYLQJ PHWKRGV RXWOLQHG LQ &KDSWHU :KLOH WKH VWUXFWXUDO SKDVH GLDJUDP IRU $3( /L1E LV IDLUO\ ZHOO NQRZQ >.RU .RU =DY@ WKH DIRUHPHQWLRQHG VSHFLILFV RI WKH $3( SURFHVV LQ /L7D KDYH FRPSOLn FDWHG D FRPSOHWH DQG DFFXUDWH UHFRQVWUXFWLRQ RI WKH VWUXFWXUDO SKDVH GLDJUDP IRU WKLV PDWHn ULDO 7KH PDLQ REVWDFOH DSSHDUV WR EH WKH GHWHUPLQDWLRQ RI UHIUDFWLYH LQGH[ HVSHFLDOO\ LQ KLJKFRQFHQWUDWLRQ KLJKVWUDLQ SKDVHV ,QGHHG WKH DSSHDUDQFH RI D QHJDWLYH VORSH RQ WKH NQRZQ SRUWLRQ RI WKH VWUXFWXUDO SKDVH GLDJUDP $Q YV Hf IRU =FXW /L7D >(+ )HG@ DQG fURWDWHGf FXWV RI /L7D >)HG@ FOHDUO\ LPSOLHV WKH H[LVWHQFH RI EXULHG LQGH[ SURILOHV 7KHUHIRUH WKH DFFXUDF\ RI WKH ,:.% PHWKRG XVHG IRU UHFRQVWUXFWLRQ RI LQGH[ SURILOHV LV OLNHO\ WR KDYH EHHQ FRPSURPLVHG $GGLWLRQDOO\ WKH GHFUHDVLQJ LQGH[ LQ KLJK FRQFHQWUDWLRQ SKDVHV UHVXOWV LQ ZDYHJXLGHV VXSSRUWLQJ YHU\ IHZ RU QR PRGHV &RQVHn TXHQWO\ WKH ,:.% PHWKRG FDQQRW EH DSSOLHG DW DOO 7KLV FLUFXPVWDQFH KDV OLPLWHG WKH UDQJH RI WKH NQRZQ VWUXFWXUDO SKDVH GLDJUDP IRU ;FXW /L7D >)HG@ 7KHVH OLPLWDWLRQV FDQ EH DYRLGHG WKURXJK WKH XVH RI GLUHFW PHDVXUHPHQW WHFKn QLTXHV 6SHFLILFDOO\ WKH WHFKQLTXHV RI GLUHFW LQGH[ SURILOLQJ DQG URFNLQJ FXUYH DQDO\VLV GHVFULEHG LQ WKH SUHYLRXV FKDSWHU FDQ EH DSSOLHG WR H[WUDFW VXUIDFH YDOXHV RI LQGH[ LQFUHn PHQW DQG ODWWLFH VWUDLQ IRU DQ\ ZDYHJXLGH 7KURXJK WKLV DSSURDFK WKH VWUXFWXUDO SKDVH GLDn JUDP FDQ EH FRQVWUXFWHG 7KLV KDV EHHQ GRQH LQ WKH QH[W VHFWLRQ

PAGE 118

&RQVWUXFWLQJ WKH 6WUXFWXUDO 3KDVH 'LDJUDP ,Q WKLV VHFWLRQ VHYHUDO SODQDU $3( ZDYHJXLGHV ZHUH IDEULFDWHG DQG PHDVXUHG YLD GLUHFW LQGH[ SURILOLQJ DQG URFNLQJ FXUYH DQDO\VLV )URP WKLV GDWD WKH VWUXFWXUDO SKDVH GLDn JUDP IRU ;FXW $3(/L7D ZDV FRQVWUXFWHG 7R UHFRQVWUXFW WKH VWUXFWXUDO SKDVH GLDJUDP IRU ;FXW /L7D $3( SODQDU ZDYHJXLGHV ZHUH IDEULFDWHG XVLQJ D YDULHW\ RI SURWRQ VRXUFHV S\URSKRVSKRULF DFLG EHQn ]RLF DFLG DPPRQLXP GLK\GURSKRVSKDWH VWHDULF DFLG DQG SDOPHWLF DFLG ([FKDQJH DQG DQQHDO WLPHV DQG WHPSHUDWXUHV ZHUH YDULHG DV GHHPHG QHFHVVDU\ WR SURGXFH D VXIILFLHQW QXPEHU RI GDWD SRLQWV DFURVV WKH VWUXFWXUDO SKDVH GLDJUDP ,W ZDV IRXQG WKDW S\URSKRVSKRn ULF DFLG DQG DPPRQLXP GLK\GURSKRVSKDWH SURYLGHG WKH KLJKHVW SURWRQ FRQFHQWUDWLRQV SRVn VLEOH ZLWKRXW FDXVLQJ VXUIDFH FUDFNV RQ WKH VDPSOH 7KLV OLPLW RI SURWRQ FRQFHQWUDWLRQ LV UHDFKHG DIWHU H[FKDQJLQJ LQ S\URSKRVSKRULF DFLG DW r& IRU PLQXWHV )XUWKHU H[FKDQJHV OHDG WR VXUIDFH FUDFNLQJ ZKLFK ZLWK WKH DVVRFLDWHG FRQFHQWUDWLRQ LQGXFHG ODWn WLFH GHIRUPDWLRQ GHWHUPLQHG WKH XSSHU OLPLW RI WKH UHFRQVWUXFWHG GLDJUDP 7KH IDEULFDWHG VDPSOHV ZHUH VXEMHFWHG WR GLUHFW LQGH[ SURILOLQJ DQG URFNLQJ FXUYH PHDVXUHPHQWV 9DOXHV RI VXUIDFH H[WUDRUGLQDU\ LQGH[ LQFUHPHQW $QH ZHUH REWDLQHG IURP REVHUYDWLRQ RI WKH PHDVXUHG LQGH[ SURILOH 9DOXHV RI VXUIDFH ODWWLFH VWUDLQ H ZHUH FDOFXn ODWHG IURP URFNLQJ FXUYH PHDVXUHPHQWV DV GHWDLOHG LQ &KDSWHU ,Q )LJ VXUIDFH YDOXHV RI WKH H[WUDRUGLQDU\ LQGH[ FKDQJH $QH DUH SORWWHG YHUVXV VXUIDFH YDOXHV RI WKH ODWWLFH VWUDLQ H WR FRQVWUXFW D FRPSOHWH VWUXFWXUDO SKDVH GLDJUDP IRU $3( SODQDU ZDYHJXLGHV LQ ;FXW /L7D 7KLV ILJXUH LOOXVWUDWHV WKDW WKH LQGH[ FKDQJH KDV DQ DSSUR[LPDWHO\ OLQHDU GHSHQGHQFH RQ ODWWLFH GHIRUPDWLRQ DQG WKHUHIRUH RQ SURWRQ FRQn FHQWUDWLRQ ZLWKLQ HDFK SKDVH +RZHYHU WKH VORSH RI WKLV GHSHQGHQFH LV GLIIHUHQW IRU HDFK

PAGE 119

)LJ 7KH FRPSOHWH VWUXFWXUDO SKDVH GLDJUDP IRU $3( ;FXW /L7D UHODWLQJ VXUIDFH LQGH[ FKDQJH $QH WR SURWRQLQGXFHG ODWWLFH GHIRUPDWLRQ H DW WKH VXUIDFH 6WUDLJKWOLQH VHJPHQWV VROLGf ZHUH ILW WR PHDVXUHG GDWD SRLQWV VTXDUHVf /RZHUFDVH OHWWHUV Df WKURXJK Hf PDUN WKH ORFDWLRQV RI ZDYHJXLGHV ZLWK FRUUHn VSRQGLQJ LQGH[ SURILOHV VKRZQ LQ )LJ SKDVH DQG WKH WUDQVLWLRQV EHWZHHQ SKDVHV DUH PDUNHG E\ VKDUS GLVFRQWLQXLWLHV 1RWH WKDW WKH VORSH GLIIHUHQFH EHWZHHQ WKH > DQG > SKDVHV LV VPDOO DQG FDQ RQO\ EH GLVWLQJXLVKHG RQ fURWDWHGf FXWV >)HG@ $V VXFK D VLQJOH OLQHDU ILW WKURXJK ERWK RI WKHVH SKDVHV ZDV XVHG ,Q WKH QH[W VHFWLRQ WKH VWUXFWXUDO SKDVH GLDJUDP LV DQDO\]HG DQG XVHG WR H[SODLQ WKH REVHUYHG DQRPDOLHV RI WKH $3( SURFHVV LQ /L7D ,Q DGGLWLRQ WKH FKDUDFWHULVWLFV RI ZDYHJXLGH SURILOHV LQ GLIIHUHQW UHJLRQV RI WKH VWUXFWXUDO SKDVH GLDJUDP DUH H[DPLQHG

PAGE 120

$QDO\]LQJ WKH 6WUXFWXUDO 3KDVH 'LDJUDP +DYLQJ UHFRQVWUXFWHG WKH VWUXFWXUDO SKDVH GLDJUDP LW FDQ EH LQVSHFWHG DQG XVHG WR H[SODLQ VHYHUDO RI WKH DIRUHPHQWLRQHG DQRPDOLHV DVVRFLDWHG ZLWK WKH $3( SURFHVV LQ /L7D 7KDW LV GRQH LQ WKLV VHFWLRQ ,Q DGGLWLRQ ZDYHJXLGH LQGH[ SURILOHV ZHUH H[DPLQHG LQ GLIIHUHQW UHJLRQ RU SKDVHV RI WKH VWUXFWXUDO SKDVH GLDJUDP LQ RUGHU WR JDXJH WKH LQIOXn HQFH RI VWUDLQ RQ ZDYHJXLGH LQGH[ SURILOHV ([SODLQLQJ $3( $QRPDOLHV $QDO\VLV RI WKH VWUXFWXUDO SKDVH GLDJUDP OHQGV DQ H[SODQDWLRQ IRU WKH DIRUHPHQn WLRQHG DQRPDO\ RI UHGXFHG LQGH[ LQFUHPHQW ZLWK LQFUHDVLQJ SURWRQ FRQFHQWUDWLRQ >0DUD@ $V SURWRQ FRQFHQWUDWLRQ LV LQFUHDVHG YLD SURWRQ H[FKDQJH VWUDLQ DQG LQGH[ LQFUHDVH WKURXJK WKH ILUVW IRXU SKDVHV +RZHYHU RQFH WKH FRQFHQWUDWLRQ LV KLJK HQRXJK WR SURGXFH D GHIRUPDWLRQ H a IXUWKHU LQFUHDVH RI SURWRQV DQG VWUDLQ OHDGV WR D GHFUHDVH LQ LQGH[ 7KH DQRPDO\ RI LQFUHDVHG LQGH[ LQFUHPHQW XSRQ VKRUW DQQHDOLQJ FDQ DOVR EH H[SODLQHG >0DUD@ ,I D VLQJOHVWHS SURWRQ H[FKDQJH SURFHVV LV VXIILFLHQW HQRXJK WR VXSn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

PAGE 121

e e e Ff e e )LJ 0HDVXUHG UHIUDFWLYH LQGH[ SURILOHV IRU Hn QHDU WKH KLJKFRQFHQn WUDWLRQ HQG RI WKH > SKDVH Df H QHDU WKH ORZFRQFHQWUDWLRQ HQG RI WKH > SKDVH Ef Hn QHDU WKH KLJKFRQFHQWUDWLRQ HQG RI WKH > SKDVH Ff Hn QHDU WKH ORZFRQFHQWUDWLRQ HQG RI WKH > SKDVH Gf DQG Hn LQ WKH >> SKDVH UHJLRQ Hf

PAGE 122

)LJ )LJ Df VKRZV WKH LQGH[ SURILOH IRU DQ DVH[FKDQJHG ZDYHJXLGH ZLWK D VXUIDFH H YDOXH RI DW WKH KLJKFRQFHQWUDWLRQ HQG RI WKH SKDVH 7KLV LV WKH KLJKHVW VWUDLQ DWWDLQDEOH ZLWKRXW VXUIDFH FUDFNLQJ $V VHHQ WKHUH LV QR VXUIDFH LQGH[ FKDQJH +RZHYHU WKHUH DUH ODUJH LQGH[ YDULDWLRQV EHORZ WKH VXUIDFH ,QGH[ SURILOH PHDVXUHPHQWV SHUIRUPHG DW GLIIHUHQW ZDYHOHQJWKV VKRZHG QR FKDQJH LQ WKH fSHULRGf RI WKH REVHUYHG LQGH[ YDULDn WLRQV DOORZLQJ IRU WKH FRQFOXVLRQ WKDW WKH YDULDWLRQV DUH QRW DQ LQWHUIHUHQFH SKHQRPHQRQ 5DWKHU WKH SUHVHQFH RI PXOWLSOH SKDVHV DQG UHVXOWLQJ QRQXQLIRUP VWUDLQ LV OLNHO\ WR EH WKH FDXVH RI WKH LQGH[ YDULDWLRQV REVHUYHG ,QGHHG URFNLQJ FXUYHV IRU WKLV VDPSOH DOVR FOHDUO\ LQGLFDWH WKH SUHVHQFH RI DW OHDVW WZR SKDVHV ZLWKLQ WKH $3( OD\HU > DQG ZLWK WKH SKDVH EHLQJ WKH XSSHUPRVW RU VXUIDFH OD\HU /RZHUFRQFHQWUDWLRQ SKDVHV DQG Df DUH DOVR SUHVHQW GHHSHU ZLWKLQ WKH $3( OD\HU EXW WKH DVVRFLDWHG VWUDLQ SURILOH LV WRR JUDGXDO WR DOORZ WKHVH SKDVHV WR EH HDVLO\ GLVFHUQHG RQ WKH URFNLQJ FXUYHV )LJ Ef GHSLFWV D ZDYHJXLGH ZLWK D VWHSOLNH LQGH[ SURILOH DQG D VPDOOHU VXUIDFH H YDOXH RI DV PD\ EH REWDLQHG WKURXJK DQQHDOLQJ 5RFNLQJ FXUYHV IRU WKLV UHJLRQ QHDU WKH ORZFRQFHQWUDWLRQ HQG RI WKH SKDVH VKRZ DQ LQFUHDVLQJ FRQWHQW RI WKH DGGLn WLRQDO > SKDVH )LJ Ff VKRZV D SURILOH IRU D ZDYHJXLGH ZLWK D VXUIDFH H YDOXH RI HVVHQn WLDOO\ LQ WKH SKDVH :DYHJXLGHV LQ WKLV UHJLRQ DUH FKDUDFWHUL]HG E\ EXULHG LQGH[ SURILOHV ZLWK D SHDN YDOXH $QH UHDFKHG EHORZ WKH VXUIDFH QHDUO\ WKH VDPH YDOXH DV WKDW RI WKH ORZFRQFHQWUDWLRQ HQG RI WKH SKDVH 5RFNLQJ FXUYHV DOVR UHYHDO WKH GLVWLQFW SUHVn HQFH RI D > SKDVH OD\HU EHQHDWK WKH > OD\HU 1RWH DJDLQ WKDW WKH LQGH[ YDULDWLRQV DQG WKH IDFW WKDW WKH ZDYHJXLGH SURILOH LV EXULHG DUH GXH WR WKH SUHVHQFH RI PXOWLSOH SKDVHV

PAGE 123

$V IXUWKHU DQQHDOLQJ LV SHUIRUPHG DQG VWUDLQV DUH UHOLHYHG WKH SUHVHQFH RI WKH SKDVH EHFRPHV PRUH SURQRXQFHG RQ URFNLQJ FXUYHV DQG LQGH[ YDULDWLRQV EHORZ WKH VXUIDFH DUH OHVV QRWLFHDEOH $W WKH VDPH WLPH WKH VXUIDFH YDOXH RI LQGH[ ULVHV XQWLO WKH SURILOH LV QR ORQJHU EXULHG )LJ Gf VKRZV VXFK D ZDYHJXLGH ZLWK D VXUIDFH H YDOXH RI DW WKH ORZFRQFHQWUDWLRQ HQG RI WKH SKDVH :DYHJXLGHV LQ WKLV UHJLRQ KDYH D FKDUDFWHULVWLFDOO\ )HUPLOLNH LQGH[ GLVWULEXWLRQ /DVWO\ )LJ Hf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

PAGE 124

)LUVW WKH 5DPDQ VSHFWUD RI VDPSOHV IDEULFDWHG LQ ERWK SXUH DQG GLOXWHG JO\FHULQ ZLOO EH PHDVXUHG DQG FRPSDUHG WR WKDW RI VLQJOHSKDVH $3( VDPSOHV WR GHWHUPLQH SKDVH FRQWHQW 1H[W VXUIDFH YDOXHV RI LQGH[ LQFUHPHQW DQG VWUDLQ DUH PHDVXUHG WR SORW WKH SRVLWLRQ RI WKHVH QHZ VDPSOHV RQ WKH VWUXFWXUDO SKDVH GLDJUDP LQGLFDWLQJ WKH SUHVHQFH RI D QHZ VLQn JOHSKDVH UHJLRQ REWDLQDEOH RQO\ WKURXJK 3( ZLWK GLOXWH VRXUFHV 5DPDQ $QDO\VLV RI 3( :DYHJXLGHV :DYHJXLGHV ZHUH IDEULFDWHG LQ ;FXW /L7D XVLQJ 3( RQO\ LQ D GLOXWH VRXUFH RI SXUH JO\FHULQ RU JO\FHULQ GLOXWHG ZLWK /L& 7KH ZDYHJXLGHV ZKLFK H[KLELWHG D UHODn WLYHO\ VWHSOLNH LQGH[ SURILOH ZHUH DQDO\]HG E\ 5DPDQ 6SHFWURVFRS\ IRU FRPSDULVRQ RI SKDVH FRQWHQW DJDLQVW $3( ZDYHJXLGHV )LJ VKRZV WKH 5DPDQ VSHFWUD RI VDPSOHV 5DPDQ VKLIW ZDYHQXPEHU FPnf )LJ 5DPDQ VSHFWUD RI 3( VDPSOHV SUHSDUHG LQ SXUH JO\FHULQ DW r& IRU YDULRXV H[FKDQJH WLPHV DQG RI DQ $3( VDPn SOH ZLWK D VWUDLQ YDOXH RI H [ LQ WKH D SKDVH

PAGE 125

5DPDQ VKLIW ZDYHQXPEHU FPf )LJ 5DPDQ VSHFWUD RI 3( VDPSOHV SUHSDUHG LQ SXUH JO\FHULQ DQG JO\FHULQ GLOXWHG ZLWK /&2 DW r& IRU YDULRXV H[FKDQJH WLPHV H[FKDQJHG LQ SXUH JO\FHULQ IRU WLPHV RI DQG KRXUV DQG WKH VSHFWUD IRU DQ $3( VDPSOH ZLWK D VWUDLQ YDOXH RI H [O LQ WKH D SKDVH $V FDQ EH VHHQ WKH VDPSOH H[FKDQJHG IRU KRXUV KDG D VSHFWUD VRPHZKDW VLPLODU WR WKH $3( VLQJOHSKDVH VDPSOH +RZHYHU VDPSOHV H[FKDQJHG IRU ORQJHU SHULRGV H[KLELWHG D VKLIW LQ WKH SHDN RI WKH 5DPDQ VSHFWUD LQGLFDWLYH RI VDPSOHV FRQWDLQLQJ KLJKHU FRQFHQWUDWLRQ VXFK DV SKDVHV )RUWXn QDWHO\ KRXUV RI H[FKDQJH ZDV HQRXJK WR SURGXFH VLQJOHPRGH ZDYHJXLGHV DW SP 7KH 5DPDQ VSHFWUD RI VDPSOHV SUHSDUHG LQ GLOXWHG JO\FHULQ DUH SUHVHQWHG LQ )LJ ZKHUH WKH\ DUH VKRZQ ZLWK WKH VSHFWUD IRU WKH SXUH JO\FHULQ VDPSOH H[FKDQJHG IRU KRXUV $V FDQ EH VHHQ DOO WKH GLOXWHG VDPSOHV VKRZHG D PRUH SUHGRPLQDWHO\ VLQJOHSKDVH FRPSRVLWLRQ WKHQ HYHQ WKH SXUH JO\FHULQ VDPSOH $GGLWLRQDOO\ WKH VSHFWUD GLG QRW FKDQJH

PAGE 126

5DPDQ VKLIW ZDYHQXPEHU FPf )LJ 5DPDQ VSHFWUD RI $3( VLQJOHSKDVH VDPSOH DQG 3( VDPn SOHV SUHSDUHG LQ ERWK SXUH JO\FHULQ DQG GLOXWHG JO\FHULQ DW r& IRU WKH WLPHV LQGLFDWHG ZLWK H[FKDQJH WLPH PHDQLQJ WKDW ORQJHU H[FKDQJHV UHTXLUHG WR SURGXFH ZDYHJXLGHV GHHS HQRXJK WR JXLGH ZRXOG EH SRVVLEOH ZLWKRXW LQFUHDVLQJ VXUIDFH FRQFHQWUDWLRQ )RU H[DPn SOH GLOXWLQJ ZLWK PROb /L& UHTXLUHG DQ H[FKDQJH WLPH RI KRXUV WR SURGXFH VLQJOHPRGH ZDYHJXLGHV DW SP $V FDQ EH VHHQ LQ )LJ WKLV GLOXWHG VDPSOH LV FRPSDUDEOH RU VOLJKWO\ LPSURYHG RYHU WKH KRXU SXUH JO\FHULQ VDPSOH DQG WKH VLQJOHn SKDVH $3( VDPSOH 0RGLILHG 6WUXFWXUDO 3KDVH 'LDJUDP )URP WKH DERYH 5DPDQ DQDO\VLV VDPSOHV IDEULFDWHG LQ SXUH JO\FHULQ IRU RU OHVV DQG VDPSOHV IDEULFDWHG LQ GLOXWHG JO\FHULQ DUH SUHGRPLQDWHO\ VLQJOH SKDVH 7R GHWHUPLQH

PAGE 127

H[DFWO\ WKHLU SRVLWLRQ RQ WKH VWUXFWXUDO SKDVH GLDJUDP LQGH[ SURILOLQJ DQG URFNLQJ FXUYHV ZHUH XVHG WR GHWHUPLQH VXUIDFH YDOXHV RI LQGH[ LQFUHPHQW $Q DQG VWUDLQ ] 7KH VDPSOH SUHSDUHG LQ SXUH JO\FHULQ DW r& IRU KRXUV KDG PHDVXUHG YDOXHV RI $Q DQG Hn [On 7KH VDPSOH SUHSDUHG LQ JO\FHULQ GLOXWHG ZLWK PROb /L& DQG H[FKDQJHG DW r& IRU KRXUV KDG PHDVXUHG YDOXHV RI $Q DQG H [ $V FDQ EH VHHQ E\ LQVSHFWLRQ RI )LJ WKHVH VDPSOH SRLQWV GR QRW FRUUHn VSRQG WR WKH D SKDVH RU WR DQ\ SKDVH VKRZQ RQ WKH VWUXFWXUDO SKDVH GLDJUDP $V VXFK LW ZDV FRQFOXGHG WKDW WKHUH H[LVWHG D QHZ SKDVH REWDLQDEOH RQO\ E\ 3( LQ D GLOXWH VRXUFH 7R FRQILUP WKLV DQ DGGLWLRQDO GDWD SRLQW ZDV FRQVWUXFWHG 7KH SXUH JO\FHULQ VDPSOH ZDV DQQHDOHG DW r& IRU KRXUV +RZHYHU QR FKDQJH LQ WKH YDOXH RI Hf ZDV PHDVXUHG $GGLWLRQDO DQQHDOLQJ ZDV SHUIRUPHG DW r& IRU KRXUV $JDLQ WKHUH ZDV QR FKDQJH LQ )LJ 0RGLILHG VWUXFWXUDO SKDVH GLDJUDP IRU ;FXW /L7D GHSLFWLQJ WKH QHZO\ GLVFRYHUHG Dn SKDVH

PAGE 128

H 7KH VDPSOH ZDV DQQHDOHG RQH ODVW WLPH DW r& IRU KRXUV 7KLV WLPH WKH PHDVXUHG YDOXHV RI VWUDLQ DQG LQGH[ LQFUHPHQW ZHUH H [Oa DQG $Q 7KH WKUHH QHZ GDWD SRLQWV ZHUH DGGHG WR WKH VWUXFWXUDO SKDVH GLDJUDP VKRZQ DV ILOOHG VTXDUHV LQ )LJ $ VWUDLJKW OLQH ZDV ILW WKURXJK WKHVH SRLQWV EDFN WR WKH RULJLQ 7KLV GDWD DORQJ ZLWK WKH 5DPDQ GDWD FRQILUPLQJ WKHVH VDPSOHV WR EH RI VLQJOHSKDVH FRQn ILUPV WKH SUHVHQFH RI D QHZ VLQJOHSKDVH UHJLRQ 7KLV QHZ SKDVH ODEHOHG WKH R7 SKDVH LQ )LJ LV REWDLQDEOH RQO\ E\ 3( LQ D GLOXWH VRXUFH ZKHUH WKH VXUIDFH FRQFHQWUDWLRQ QHYHU OHDYHV WKLV SKDVH GXULQJ H[FKDQJH 6XPPDU\ ,Q WKLV FKDSWHU WKH VWUXFWXUDO SKDVH GLDJUDP IRU ;FXW $3(/L7D ZDV FRQn VWUXFWHG )LUVW VRPH RI WKH OLPLWDWLRQV KLQGHULQJ SUHYLRXV HIIRUWV WR FRQVWUXFW WKLV GLDJUDP ZHUH RXWOLQHG OHDGLQJ WR WKH QHHG IRU GLUHFW LQGH[ SURILOLQJ 1H[W GLUHFW LQGH[ SURILOLQJ DORQJ ZLWK URFNLQJ FXUYH PHDVXUHPHQWV ZHUH DSSOLHG WR $3( VDPSOHV RI YDULRXV IDEULFDn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

PAGE 129

ZLWKLQ WKLV QHZ SKDVH REWDLQDEOH RQO\ E\ 3( LQ GLOXWH VRXUFHV ZHUH WUXO\ RI VLQJOHSKDVH FRQWHQW ,Q WKH QH[W FKDSWHU WKH VWUXFWXUDO SKDVH GLDJUDP FRQVWUXFWHG KHUH ZLOO EH XVHG WR VWXG\ WKH VWDELOLW\ RI ERWK $3( DQG 3( ZDYHJXLGHV LQ /L7D 0HWDOLQGLIIXVHG ZDYHJXLGHV ZLOO DOVR EH H[DPLQHG IRU VWDELOLW\ FRPSDULVRQ DJDLQVW $3( DQG 3( ZDYHJXLGHV $GGLWLRQDOO\ FU\VWDOV IURP GLIIHUHQW JURZHUV ZLOO EH H[DPLQHG WR GHWHUPLQH WKH PDWXULW\ RI WKH JURZWK SURFHVV IRU WKLV FU\VWDO

PAGE 130

&+$37(5 67$%,/,7< 2) /L7D :$9(*8,'(6 7KH WHPSRUDO LQVWDELOLW\ RI ZDYHJXLGHV LQ /L7D LV WKH PDMRU KLQGUDQFH LQ WKH ZD\ RI WKHLU ZLGHVSUHDG XVH LQ SUDFWLFDO GHYLFHV $W OHDVW IRU $3( ZDYHJXLGHV WKH LQVWDELOn LWLHV DUH VXVSHFWHG WR EH FDXVHG E\ WKH SUHVHQFH RI D PL[WXUH RI FU\VWDO SKDVHV WKDW FDQ WUDQVIRUP LQ WLPH RQH LQWR DQRWKHU $V D UHVXOW WKH ZDYHJXLGH LQGH[ DYHUDJHG RYHU WKH ZDYHJXLGLQJ DUHD FKDQJHV OHDGLQJ LQ WXUQ WR D YDULDWLRQ LQ WKH SURSDJDWLRQ FRQVWDQW RI WKH JXLGHG PRGHV 7KHUH LV D YDOLG VXJJHVWLRQ WKDW D ZDYHJXLGH FRQWDLQLQJ RQO\ WKH ORZFRQn FHQWUDWLRQ D SKDVH ZLOO H[KLELW LPSURYHG WHPSRUDO VWDELOLW\ >(+@ ,QFUHDVHG VWDELOLW\ RI VWUXFWXUHV LQ SUHGRPLQDQWO\ RQH SKDVH ORZFRQFHQWUDWLRQ RU KLJKFRQFHQWUDWLRQf >(+@ VXSSRUWV WKLV LGHD ,Q WKLV FKDSWHU WKH VWDELOLW\ RI ZDYHJXLGHV LQ GLIIHUHQW UHJLRQV RI WKH VWUXFWXUDO SKDVH GLDJUDP IRU $3(/L7D DUH H[DPLQHG DORQJ ZLWK WKH VWDELOLW\ RI ZDYHJXLGHV IDEn ULFDWHG E\ PHWDOLQGLIIXVLRQ ZLWK WKH DLP RI GHWHUPLQLQJ WKH IHDVLELOLW\ RI GHILQLQJ IDEULn FDWLRQ FRQGLWLRQV \LHOGLQJ WHPSRUDOO\ VWDEOH ZDYHJXLGHV )RU $3( ZDYHJXLGHV D TXDOLWDWLYH DQDO\VLV ZDV SHUIRUPHG XVLQJ URFNLQJ FXUYH PHDVXUHPHQWV WR PRQLWRU DQ\ FKDQJH LQ VWUDLQ YDOXHV H RYHU WLPH )RU D PRUH TXDQWLWDWLYH DQDO\VLV D GLUHFWLRQDO FRXn SOHU ZDV XVHG WR WUDFN FKDQJHV LQ WKH FRXSOLQJ OHQJWK RI FRXSOHUV IDEULFDWHG XVLQJ ZDYHJXLGHV RI YDU\LQJ IDEULFDWLRQ FRQGLWLRQV 7KHVH FKDQJHV LQ FRXSOLQJ OHQJWK ZHUH WKHQ UHODWHG WR FRUUHVSRQGLQJ FKDQJHV LQ LQGH[ LQFUHPHQW WKURXJK D FRPSXWHU VLPXODWLRQ RI D

PAGE 131

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n XHV DQG FRUUHVSRQGLQJ SKDVH DW WKH VDPSOH VXUIDFH MXVW DIWHU IDEULFDWLRQ DQG DJDLQ GD\V ODWHU 7KH DUURZV RQ WKH OHIW VLGH RI WKH ILJXUH LQGLFDWH VDPSOHV ZKLFK UHFRUGHG D FKDQJH LQ WKH H YDOXH RYHU WKH GD\ SHULRG 7KH DUURZV RQ WKH ULJKW VLGH RI WKH ILJXUH LQGLFDWH VDPSOHV ZKLFK QRW RQO\ VKRZHG D FKDQJH LQ VWUDLQ EXW DOVR H[SHULHQFHG D SKDVH WUDQVLWLRQ IRUP WKH WR WKH SKDVHf LQ WKH QHDUVXUIDFH OD\HU )URP 7DEOH LW ZDV VHHQ WKDW VDPSOHV LQ WKH KLJKFRQFHQWUDWLRQ UHJLRQV RI WKH VWUXFWXUDO SKDVH GLDJUDP SDUWLFXODUO\ WKH DQG SKDVHV H[KLELWHG VLJQLILFDQW LQVWDELOLn WLHV RYHU WKH SHULRG H[DPLQHG ZLWK D FKDQJH LQ H E\ DV PXFK DV b LQ VRPH FDVHV +HQFH WKHVH WZR KLJKFRQFHQWUDWLRQ SKDVHV DUH QRW VXLWDEOH IRU WKH IDEULFDWLRQ RI UHOLDEOH

PAGE 132

MXVW DIWHU IDEULFDWLRQ GD\V ODWHU H [ SKDVH H [L2 SKDVH S 3 S > SL 33 33 O_ . . D D D D 7DEOH 0HDVXUHG YDOXHV RI VWUDLQ H WDNHQ IURP $3( VDPSOHV LPPHGLDWHO\ DIWHU IDEULFDWLRQ DQG DJDLQ GD\V ODWHU $UURZV RQ WKH OHIW LQGLFDWH VDPSOHV ZKLFK H[KLELWHG D FKDQJH LQ H $UURZV RQ WKH ULJKW LQGLFDWH VDPSOHV ZKLFK XQGHUZHQW D SKDVH WUDQVLWLRQ

PAGE 133

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n WLRQ RI D GLUHFWLRQDO FRXSOHU LV SURYLGHG 7KHQ WKH PHDVXUHPHQW WHFKQLTXH XVHG WR GHWHUn PLQH FKDQJHV LQ LQGH[ LQFUHPHQW LV RXWOLQHG 'LUHFWLRQDO &RXSOHU 7KHRU\ $ SDVVLYH GLUHFWLRQDO FRXSOHU LV VKRZQ LQ )LJ ZKHUH Z LV WKH LQGLYLGXDO ZDYHJXLGH ZLGWK J LV WKH JDS RI WKH FRXSOHU UHJLRQ DQG / LV WKH LQWHUDFWLRQ OHQJWK RI WKH FRXSOHU VHFWLRQ 7KH GHYLFH VKRXOG EH GHVLJQHG VR WKDW HDFK DUP RI WKH FRXSOHU LV VLQJOHn PRGH DW WKH ZDYHOHQJWK RI LQWHUHVW LQ WKLV FDVH _LP DQG VXSSRUWV WKH IXQGDPHQWDO V\PPHWULF DQG DQWLV\PPHWULF VXSHUPRGHV 7KHVH PRGHV DUH RUWKRJRQDO WR HDFK RWKHU

PAGE 134

)LJ ,OOXVWUDWLRQ RI D SDVVLYH GLUHFWLRQDO FRXSOHU DQG KDYH GLIIHUHQW SURSDJDWLRQ FRQVWDQWV 3 DQG 3L UHVSHFWLYHO\ 7KH GLIIHUHQFH LQ SURSDn JDWLRQ FRQVWDQWV RI WKHVH WZR PRGHV LV $S 33@ )URP FRXSOHG PRGH WKHRU\ WKH RXWSXW SRZHU UDWLR RI WKLV FRXSOHU LV H[SUHVVHG DV 3 WP?Z HfS f ZKHUH QS UHSUHVHQWV WKH DPRXQW RI UHVLGXDO FRXSOLQJ ZKLFK RFFXUV DW WKH WDSHUHG UHJLRQ DIWHU WKH LQWHUDFWLRQ OHQJWK DQG OF UF$3 LV WKH FRXSOLQJ OHQJWK GHILQHG DV WKH OHQJWK QHHGHG WR FRXSOH DOO RI WKH SRZHU IURP JXLGH RQH LQWR JXLGH WZR 0HDVXULQJ &RXSOLQJ /HQJWK DQG ,QGH[ &KDQJH 7R PHDVXUH FRXSOLQJ OHQJWK OF DQ DUUD\ RI FRXSOHUV ZHUH IDEULFDWHG KDYLQJ WKH VDPH ZLGWK Z DQG JDS J EXW YDU\LQJ LQWHUDFWLRQ OHQJWKV / 7KH RXWSXW RI HDFK DUP ZDV WKHQ PHDVXUHG DQG WKH QRUPDOL]HG RXWSXW LH 333ff ZDV SORWWHG DV D IXQFWLRQ RI LQWHUDFWLRQ OHQJWK / 6XFK D SORW LV GHSLFWHG LQ )LJ 7KHQ WKH IXQFWLRQ RI (T f IRU RQH DUP ZDV ILW WR WKH PHDVXUHG GDWD WR GHWHUPLQH WKH FRXSOLQJ OHQJWK OF )LJ f ,W

PAGE 135

/ PPf )LJ ([DPSOH RI D SORW RI QRUPDOL]HG RXWSXW SRZHU YHUVXV LQWHUDFWLRQ OHQJWK / IRU DQ DUUD\ RI GLUHFWLRQDO FRXSOHUV 0HDVXUHG GDWD LV VKRZQ DV VTXDUHV DQG WKH WKHRUHWLFDO ILW WR WKH GDWD VKRZQ DV D VROLG OLQH LQGLn FDWHV D FRXSOLQJ OHQJWK RI PP KDV EHHQ VKRZQ >&KH@ WKDW LW LV SRVVLEOH WR GHWHUPLQH OF IURP DV OLWWOH DV WZR FRXSOHUV ZLWK FORVH LQWHUDFWLRQ OHQJWKV 7R TXDQWLI\ VWDELOLW\ WKH FRXSOLQJ OHQJWK ZDV PRQLWRUHG RYHU WLPH LQ RUGHU WR WUDFN DQ\ FKDQJH GXH WR LQGH[ LQVWDELOLW\ &KDQJHV LQ PHDVXUHG FRXSOLQJ OHQJWK ZHUH WKHQ UHODWHG WR FRUUHVSRQGLQJ FKDQJHV LQ LQGH[ LQFUHPHQW $Qf WKURXJK D )LQLWH 'LIIHUHQFH 0HWKRG )'0f FRPSXWHU VLPXODWLRQ RI D GLUHFWLRQDO FRXSOHU >0DU22@ )RU HDFK IDEULFDn WLRQ FRQGLWLRQ WKH LQGH[ SURILOH RI WKH ZDYHJXLGHV FRPSULVLQJ WKH GLUHFWLRQDO FRXSOHU ZHUH PHDVXUHG E\ UHIOHFWLYLW\ SURILOLQJ &KDSWHU f VR WKDW WKH DSSURSULDWH SDUDPHWHUV GHVFULEn LQJ WKH SURILOH FRXOG EH XVHG LQ WKH VLPXODWLRQ 7KH DFFXUDF\ LQ PHDVXULQJ FRXSOLQJ OHQJWK

PAGE 136

ZDV EHWWHU WKDQ PP DQG WKH VHQVLWLYLW\ RI WKH GLUHFWLRQDO FRXSOHU DOORZHG IRU WKH WUDFNn LQJ RI LQGH[ FKDQJHV RQ WKH RUGHU RI n >0DWD@ ,Q WKH QH[W VHFWLRQ WKH PHWKRG GHVFULEHG LQ WKLV VHFWLRQ RI PHDVXULQJ LQGH[ FKDQJH XVLQJ D GLUHFWLRQDO FRXSOHU LV DSSOLHG WR $3( DQG 3( ZDYHJXLGHV 7KH VWDELOLW\ RI ZDYHJXLGHV LQ GLIIHUHQW UHJLRQV RI WKH VWUXFWXUDO SKDVH GLDJUDP LV H[DPLQHG LQ HIIRUW WR SLQSRLQW IDEULFDWLRQ FRQGLWLRQV OHDGLQJ WR LPSURYHG WHPSRUDO VWDELOLW\ $3( DQG 3( 6WDELOLW\ 5HVXOWV ,Q WKLV VHFWLRQ WKH UHVXOWV RI VWDELOLW\ WHVWV IRU ZDYHJXLGHV LQ GLIIHUHQW UHJLRQV RI WKH VWUXFWXUDO SKDVH GLDJUDP IRU $3(/L7D DUH SUHVHQWHG )LUVW $3( ZDYHJXLGHV LQ ERWK KLJKFRQFHQWUDWLRQ DQG ORZFRQFHQWUDWLRQ UHJLRQV DUH H[DPLQHG %DVHG RQ WKHVH UHVXOWV WKH PRYH WR ORZFRQFHQWUDWLRQ 3( ZDYHJXLGHV LV PDGH IRU D QXPEHU RI DGYDQWDJHV ZKLFK ZLOO EH GLVFXVVHG 6WDELOLW\ UHVXOWV RQ WKHVH 3( ZDYHJXLGHV DUH JLYHQ QH[W DQG WKH VHFWLRQ FRQFOXGHV ZLWK D GLVFXVVLRQ RQ WKH WUDGHRII DVVRFLDWHG ZLWK GHFUHDVLQJ VWUDLQ WR SRWHQWLDOO\ LPSURYH VWDELOLW\ DQG WKH UHGXFWLRQ RI $Q WKDW RFFXUV ZLWK GHFUHDVHG VWUDLQ $3( 5HVXOWV 7KH ILUVW FRXSOHUV WHVWHG ZHUH IDEULFDWHG LQ =FXW /L7D XVLQJ WKH $3( SURFHVV DQG S\URSKRVSKRULF DFLG DV D SURWRQ VRXUFH 7KH FRQGLWLRQV IRU IDEULFDWLRQ DUH JLYHQ LQ 7DEOH )RU HDVH RI UHIHUHQFH WKH VWUXFWXUDO SKDVH GLDJUDP IRU =FXW $3(/L7D LV UHSULQWHG LQ )LJ $ =FXW FU\VWDO ZDV FKRVHQ EHFDXVH HVSHFLDOO\ IRU WKH KLJK VXUIDFH FRQFHQWUDWLRQ VDPSOH 6DPSOH f LW ZDV H[WUHPHO\ GLIILFXOW WR SURGXFH D ZDYHJXLGH LQ ; FXW FU\VWDOV XVLQJ S\URSKRVSKRULF DFLGf GHHS HQRXJK WR VXSSRUW D PRGH DW _LP

PAGE 137

6DPSOH &U\VWDO &XW $3( &RQGLWLRQV H &rnf 3KDVH 'D\V $Qf = 3( r& KUV $r& OKU 3( r& OKU $ r& KUV 3( r& PLQ [ 2n [On [Of = 3( r& O2PLQ $r& OKU 3( r& PLQ $ r& O2KUV [ 2n [ 2n [ 2f 7DEOH )DEULFDWLRQ FRQGLWLRQV PHDVXUHG H YDOXH DVVRFLDWHG SKDVH DQG HVWLn PDWHG FKDQJH LQ LQGH[ LQFUHPHQW $Qf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n LQJ D FKDQJH LQ LQGH[ LQFUHPHQW $Qf RI [Oa 7KH PDJQLWXGH RI WKLV FKDQJH VXUIDFH e [O2f )LJ 6WUXFWXUDO SKDVH GLDJUDP IRU UXW $33, L7D2

PAGE 138

DJUHHV ZLWK SUHYLRXVO\ UHSRUWHG GDWD >0DWD@ $V ZDV GLVFRYHUHG HDUOLHU E\ PRQLWRULQJ FKDQJHV LQ URFNLQJ FXUYHV 6HFWLRQ f ZDYHJXLGHV ZLWK KLJK SURWRQ FRQFHQWUDWLRQ QHDU WKH XSSHU OLPLWV RI WKH SKDVH GLDJUDP DUH KLJKO\ XQVWDEOH DQG WKHUHIRUH XQVXLWDEOH IRU )LJ 0HDVXUHG FRXSOLQJ OHQJWK RI 6DPSOH 7KLV FRXSOHL FRQ UHOLDEOH GHYLFH IDEULFDWLRQ VLVWHG RI SP FKDQQHOV ZLWK D SP JDS MW MV ZRUWK QRWLQJ WKDW DW OHDVW IRU WKH FDVH RI KLJK FRQFHQWUDWLRQ VDPn SOHV ZKLFK DUH LQ SKDVHV RI GHFUHDVLQJ VORSH WKDW RYHU WLPH ERWK WKH VXUIDFH VWUDLQ DQG WKH LQGH[ LQFUHPHQW ZHUH IRXQG WR GHFUHDVH 7KLV LV LQ FRQWUDGLFWLRQ WR WKH HYROXWLRQ RI WKH LQGH[ SURILOH XSRQ DQQHDOLQJ ZKLFK ZDV IRXQG WR IROORZ WKH VWUXFWXUDO SKDVH GLDJUDP DQG VKRZ DQ LQFUHDVH LQ LQGH[ DV WKH VDPSOH ZDV DQQHDOHG GHFUHDVLQJ VWUDLQ WKURXJK WKH KLJK FRQFHQWUDWLRQ SKDVHV 7KH UHDVRQ IRU WKLV SHFXOLDULW\ LV QRW FOHDU EXW LW ZRXOG VHHP WR VXJn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

PAGE 139

KLJKVWUDLQ KLJKFRQFHQWUDWLRQ SKDVHV LQ WKH ZDYHJXLGLQJ OD\HU +RZHYHU HYHQ JUHDWHU VWDELOLW\ LV QHFHVVDU\ IRU UHOLDEOH GHYLFH RSHUDWLRQV 7KHUHIRUH LQYHVWLJDn WLRQ RI ZDYHJXLGHV LQ WKH D SKDVH ZDV QHFHVVDU\ 7KH $GYDQWDJH RI 3( 7HPSRUDO LQVWDELOLWLHV ZHUH VXVn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n WLRQ VRXUFH JO\FHULQ DQG WR XVH ;FXW FU\VWDOV 7KH VWUXFWXUDO SKDVH GLDJUDPV SORWWHG LQ )LJ 0HDVXUHG FRXSOLQJ OHQJWK RI 6DPSOH 7KLV FRXSOHU FRQn VLVWHG RI SP FKDQQHOV ZLWK DQ SP JDS

PAGE 140

WKLV UHVHDUFK ZHUH GRQH VR DW QP %HFDXVH WKH LQGH[ LQFUHPHQW LV NQRZQ WR H[KLELW VRPH GLVSHUVLRQ WKH DFWXDO YDOXH RI VXUIDFH LQGH[ IRU ZDYHJXLGHV RSHUDWLQJ DW SP LV PXFK ORZHU WKDQ WKDW VKRZQ LQ WKH GLDJUDPV DW OHDVW WZR WLPHV >(+@ 7KH ;FXW FU\Vn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n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e [O2f H[DPSOH WR DQ LQFUHDVH LQ FRXSOLQJ OHQJWK )LJ 6WUXFWXUDO SKDVH GLDJUDP IRU ;FXW $3(/L7D IURP DERXW PP WR PP LQ FRXSOHUV ZLWK _LP FKDQQHOV DQG DQ _LP JDS

PAGE 141

)LJ 0HDVXUHG FRXSOLQJ OHQJWK RI 6DPn SOH 7KLV FRXSOHU ZDV FRPSULVHG RI LP FKDQQHOV ZLWK DQ LP JDS 7KH VWDELOLW\ RI WKHVH QHZ ZDYHJXLGHV LV H[DPLQHG QH[W 3( 5HVXOWV 7KH PHDVXUHG FKDQJH LQ FRXn SOLQJ OHQJWK IRU 6DPSOH LV VKRZQ LQ )LJ 7KH IDEULFDWLRQ FRQGLWLRQV DQG HVWLPDWHG FKDQJH LQ LQGH[ LQFUHn PHQW $Qf IRU WKLV DQG DOO WKH 3( VDPSOHV WHVWHG DUH JLYHQ LQ 7DEOH 7KLV VDPSOH ZDV LQ WKH Dn SKDVH KDYLQJ D VWUDLQ YDOXH RI H [O OHVV WKDQ WKDW RI 6DPSOH EXW JUHDWHU WKDQ WKDW RI 6DPSOH $V FDQ EH VHHQ IURP 7DEOH MXVW GD\V DIWHU IDEULFDWLRQ WKH LQGH[ LQFUHPHQW IRU WKLV VDPSOH GHFUHDVHG E\ $Qf [O2 $JDLQ FRPSDUHG WR WKH SUHYLRXV VDPSOHV WKLV FKDQJH LV JUHDWHU WKDQ WKDW RI 6DPSOH DQG URXJKO\ VLPLODU WR WKDW RI 6DPSOH IRU WKH VDPH WLPH SHULRG ,W ZDV HYLGHQW DIWHU MXVW GD\V WKDW 6DPSOH ZRXOG QRW EH VWDEOH HQRXJK VR LQ DQ HIIRUW WR LPSURYH LWV VWDELOLW\ LW ZDV DQQHDOHG DW r& IRU KRXUV ,W ZDV WKHQ GHQRWHG DV 6DPSOH n 7DEOH f 7KH PHDVXUHG FKDQJH LQ FRXSOLQJ OHQJWK IRU WKLV VDPSOH LV VKRZQ LQ )LJ ,W ZDV VHHQ IURP WKLV ILJXUH DQG 7DEOH WKDW WKH HIIHFW RI DQQHDOLQJ ZDV WR UHGXFH WKH FRXSOLQJ OHQJWK IURP DERXW PP WR DERXW PP WKRXJK WKH PHDVXUHG YDOXH RI VWUDLQ GLG QRW FKDQJH $GGLWLRQDOO\ DIWHU GD\V WKH FKDQJH LQ LQGH[ LQFUHPHQW ZDV $Qf [O2a ODUJHU WKDQ WKDW RI 6DPSOH DQG OHVV WKDQ WKDW RI 6DPSOH IRU URXJKO\ WKH VDPH WLPH SHULRG

PAGE 142

6DPSOH &U\VWDO &XW 3( &RQGLWLRQV H nf 3KDVH 'D\V $Qf ; JO\FHULQ r&KUV Dn [ 2n n ; 6DPSOH DQQHDOHG r& KUV Dn [ 2n [ 2n [Of ; JO\FHULQ PROb /L& r& KUV Dn O[O2 [On [Of ; JO\FHULQ PROb /L& r& KUV DJHG r& GD\V QD QD [On n ; PRUH DJLQJ RI 6DPSOH r& GD\V QD QD [ 2n [ 2n 7DEOH )DEULFDWLRQ FRQGLWLRQV PHDVXUHG H YDOXH DVVRFLDWHG SKDVH DQG HVWLn PDWHG FKDQJH LQ LQGH[ LQFUHPHQW $Qf DW YDULRXV SHULRGV DIWHU IDEULFDn WLRQ IRU 6DPSOHV n DQG GD\V )LJ 0HDVXUHG FRXSOLQJ OHQJWK RI 6DPSOH n 7KLV FRXSOHU ZDV FRPSULVHG RI SLQ FKDQQHOV ZLWK DQ SP JDS

PAGE 143

7KH DERYH UHVXOWV VHHP WR LQGLFDWH WKDW LQVWDELOLW\ LV QRW VLPSO\ GXH WR WKH SUHVHQFH RI PXOWLSOH FU\VWDO SKDVHV LQ WKH ZDYHJXLGLQJ OD\HU DV ZDV SUHYLRXVO\ DVVXPHG 6DPSOH KDYLQJ WZR SKDVHV ZDV PRUH VWDEOH WKDQ 6DPSOHV DQG n ZKLFK ZHUH VLQJOHSKDVH 7KHVH UHVXOWV GR KRZHYHU VHHP WR LQGLFDWH WKDW ZDYHJXLGHV ZLWK GHFUHDVHG SURWRQ FRQn FHQWUDWLRQ DQG VWUDLQ GR KDYH LPSURYHG VWDELOLW\ DV 6DPSOH KDG WKH VPDOOHVW VWUDLQ YDOXH DQG DOVR WKH EHVW VWDELOLW\ $V D UHVXOW LW ZDV GHFLGHG WR WU\ WR IDEULFDWH ZDYHJXLGHV ZLWK GHFUHDVHG VWUDLQ YDOXHV $JDLQ WKH 3( WHFKQLTXH DIIRUGHG WKH EHVW RSSRUWXQLW\ DV DWWHPSWV WR IDEULFDWH ZDYHJXLGHV E\ $3( ZLWK VWUDLQ YDOXHV VPDOOHU WKDQ WKDW RI 6DPSOH UHTXLUHG WRR PDQ\ VXFFHVVLYH UHSHWLWLRQV RI H[FKDQJH DQG DQQHDO HVSHFLDOO\ IRU ;FXW WR EH SUDFWLFDO $GGLWLRQDOO\ WKHVH ZDYHJXLGHV VXIIHUHG IURP D FRPELQDWLRQ RI VPDOO $Q DW _LP DQG VHYHUHO\ JUDGHG SURILOHV GXH WR DQQHDOLQJ PDNLQJ EHQGLQJ ORVVHV VR KLJK WKDW WKH IDEULFDWLRQ RI GLUHFWLRQDO FRXSOHUV ZKLFK KDYH EHQGV ZDV TXLWH LPSUDFWLFDO 7KHUHn IRUH LW ZDV GHFLGHG WR VWLFN ZLWK 3( KDYLQJ WKH DGYDQWDJHV RXWOLQH LQ 6HFWLRQ RQO\ ZLWK D PRUH GLOXWH SURWRQ VRXUFH LQ WKLV FDVH JO\FHULQ GLOXWHG ZLWK /L& 7KH PHDVXUHG FKDQJH LQ FRX GD\V SOLQJ OHQJWK RI WKH QHZ VDPSOH IDEULn FDWHG LQ GLOXWHG JO\FHULQ GHQRWHG DV 6DPSOH LQ 7DEOH f LV VKRZQ LQ )LJ %HFDXVH RI WKH PRUH GLOXWH VRXUFH WKH H[FKDQJH WLPH LQFUHDVHG IURP KRXUV WR DERXW KRXUV 7KH H YDOXH IRU WKLV VDPSOH GHFUHDVHG )LJ 0HDVXUHG FRXSOLQJ OHQJWK RI 6DPn SOH 7KLV FRXSOHU ZDV FRPSULVHG RQO\ VOLJKWO\ WR $V LQGLFDWHG LQ RI _LP FKDQQHOV ZLWK DQ SP JDS

PAGE 144

7DEOH WKLV VDPSOH ZDV VWLOO PRUH XQVWDEOH WKDQ 6DPSOH +RZHYHU WR WU\ WR IDEULFDWH VDPSOHV ZLWK VLJQLILFDQWO\ UHGXFHG H YDOXHV E\ XVLQJ DQ HYHQ PRUH GLOXWH VRXUFH ZRXOG KDYH UHTXLUHG KXQGUHGV RI KRXUV RI H[FKDQJH WLPH GXH WR WKH GHFUHDVHG SURWRQ FRQFHQWUDn WLRQ RI WKH VRXUFH DQG WKH ODUJHU GHSWK UHTXLUHG WR VXSSRUW D PRGH DW WKH ORZHU $Q YDOXHV DVVRFLDWHG ZLWK GHFUHDVHG H YDOXHV 6XFK D IDEULFDWLRQ VLWXDWLRQ LV QRW SUDFWLFDO IRU WKH IRUPDWLRQ RI ZDYHJXLGHV %\ FRPSDULVRQ ZLWK 6DPSOHV DQG n KRZHYHU LW ZDV VHHQ WKDW 6DPSOH WKRXJK KDYLQJ D VPDOOHU VWUDLQ YDOXH ZDV QR PRUH VWDEOH WKDQ 6DPSOH DQG PRUH XQVWDEOH WKHQ 6DPSOH n (YLGHQWO\ DQQHDOLQJ 6DPSOH WR SURGXFH 6DPSOH n KDG VRPH LPSDFW RQ VWDn ELOLW\ 7R WKLV HQG LW ZDV GHFLGHG WR WU\ WR DUWLILFLDOO\ DJH >.LV@ WKHVH VDPSOHV DW D ORZ WHPSHUDWXUH r&f %\ WKLV PHDQV LW PLJKW EH SRVVLEOH WR LPSURYH WKH VWDELOLW\ RI WKHVH VDPSOHV ZLWKRXW FDXVLQJ VLJQLILFDQW DOWHUDWLRQV WR WKH + SURILOH DQG KHQFH WKH LQGH[ SURILOH DV DQQHDOLQJ DW KLJKHU WHPSHUDWXUHV FDXVHV 6DPSOH 7DEOH f KDG URXJKO\ WKH VDPH H[FKDQJH FRQGLWLRQV DV 6DPSOH ,Q DGGLWLRQ KRZHYHU LW ZDV DJHG DW r& IRU GD\V 7KRXJK LWV VWUDLQ YDOXH ZDV QRW GLUHFWO\ PHDVXUHG EHFDXVH RI LWV IDEULFDWLRQ FRQGLWLRQV LW PXVW KDYH KDG DQ Hf YDOXH VLPLODU WR WKDW RI 6DPSOH RU VOLJKWO\ VPDOOHU 7KH PHDVXUHG FKDQJH LQ FRXSOLQJ IRU WKLV VDPSOH LV VKRZQ LQ )LJ $IWHU D SHULRG RI MXVW GD\V LW ZDV HYLGHQW WKDW VWDELOLW\ KDG QRW EHHQ LPSURYHG 6R WKLV VDPSOH ZDV VXEMHFWHG WR IXUWKHU DJLQJ DW r& IRU PRUH GD\V ,W ZDV WKHQ GHQRWHG DV 6DPSOH 7DEOH f :LWKLQ GD\V DIWHU WKLV DJLQJ WKH ZDYHJXLGH FRXSOHUV IHOO EHORZ FXWRII IRU WKH IXQGDPHQWDO PRGH DW _LP $V D UHVXOW WKH FKDQJH LQ FRXSOLQJ OHQJWK ZDV

PAGE 145

L L U L Q L L L LnaU L aS L L L S L L L UJ L L L L L L L UL c SP L L L L L L L L \ ’ U Q c % % Q 2 B B ’ ’ f§ ,f 7 M IL "Q rn n O O O O n n O O n n n O n ‘ O n O n 3L n n La 2 2 GD\V GD\V )LJ 0HDVXUHG FRXSOLQJ OHQJWK RI 6DPSOH 7KLV FRXSOHU ZDV FRPSULVHG RI _LP FKDQQHOV ZLWK DQ _LP JDS )LJ 0HDVXUHG FRXSOLQJ OHQJWK RI 6DPSOH 7KLV FRXSOHU ZDV FRPSULVHG RI _LP FKDQQHOV ZLWK DQ _LP JDS PRQLWRUHG DW _P DQG LV VKRZQ LQ )LJ 7KH VWDELOLW\ KRZHYHU ZDV QRW VLJQLILFDQWO\ LPSURYHG RYHU WKDW RI 6DPSOH /LPLWV WR 5HGXFLQJ 6WUDLQ $V WKH LQGH[ LQFUHPHQW $Q RI 3( DQG $3( ZDYHJXLGHV LV RQ WKH RUGHU RI a RU VPDOOHU IRU ZDYHOHQJWKV QHDU SP WKH REVHUYHG LQVWDELOLWLHV FRQVWLWXWH FKDQJHV LQ LQGH[ LQFUHPHQW $Qf RI JUHDWHU WKDQ b 6XFK FKDQJHV DUH XQDFFHSWDEOH IRU WKH IDEULn FDWLRQ RI UHOLDEOH LQWHJUDWHG RSWLFDO GHYLFHV $V ZDV KLQWHG WR KHUH DQG LV NQRZQ WR EH WKH FDVH IRU $3(/L1E GHFUHDVLQJ WKH SURWRQ FRQFHQWUDWLRQ DQG KHQFH VWUDLQ PD\ OHDG WR LPSURYHG VWDELOLW\ +RZHYHU VHYHUDO REVWDFOHV OLPLW WKLV DFKLHYHPHQW )LUVW WKH $Q RI 3( DQG $3( ZDYHJXLGHV LQ /L7D FDQ EH DV PXFK DV DQ RUGHU RI PDJQLWXGH VPDOOHU WKDQ WKDW RI /L1E $GGLWLRQDOO\ ZDYHJXLGH GHYLFHV RSHUDWLQJ DW _LP DUH UHTXLUHG ZKHUH EHFDXVH RI GLVSHUVLRQ $Q YDOXHV DUH PXFK VPDOOHU WKDQ

PAGE 146

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n QHQWLDO LQGH[ SURILOHV >(+@ ZKLFK DJDLQ ZRXOG KDYH SRRU FRQILQHPHQW DQG ODUJH EHQGn LQJ ORVVHV 7KLUG EHFDXVH RI WKH ORZ $Q YDOXHV ZDYHJXLGHV ZRXOG EH UHTXLUHG WR EH TXLWH GHHS a_OPf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n HOV DUH ORZ HQRXJK LW LV SRVVLEOH WR IDEULFDWH ZDYHJXLGHV LQ /L1E E\ $3( ZKLFK H[KLELW QR PHDVXUDEOH WHPSRUDO LQVWDELOLW\ >0DWD@ /LNHZLVH QR GDWD KDV EHHQ UHSRUWHG RQ WKH LQVWDELOLW\ RI 7LLQGLIIXVHG ZDYHJXLGHV LQ /L1E )RU WKH FDVH RI /L7D LI VWDELOLW\ GHSHQGV RQ IDEULFDWLRQ FRQGLWLRQV DV LW GRHV LQ /L1E WKHQ WKH SURGXFWLRQ RI VWDEOH 3( DQG $3( ZDYHJXLGHV LV KLQGHUHG HVSHFLDOO\ DW

PAGE 147

_LP EHFDXVH RI WKH LQKHUHQWO\ ORZ YDOXHV RI LQGH[ LQFUHPHQW $Q +RZHYHU PHWDO LQGLIIXVHG ZDYHJXLGHV VKRXOG WKHQ EH VWDEOH DV WKH\ DUH LQ /L1E 7KLV VLWXDWLRQ LV H[DPLQHG LQ WKH QH[W VHFWLRQ ZKHUH WKH VWDELOLW\ RI =Q YDSRULQGLIIXVHG DQG 7LLQGLIIXVHG ZDYHJXLGHV LV PHDVXUHG =Q7L,QGLIIXVHG 6WDELOLW\ 5HVXOWV ,I VWDELOLW\ GHSHQGV RQ ZDYHJXLGH IDEULFDWLRQ FRQGLWLRQV WKHQ PHWDOLQGLIIXVHG ZDYHJXLGHV LQ /L7D VKRXOG EH VWDEOH DV WKH\ DUH LQ /L1E 7KHVH ZDYHJXLGHV SURn GXFH QR VWUDLQ DV GRHV 3( $GGLWLRQDOO\ WKHLU LRQV DUH ODUJHU LQ VL]H WKDQ WKH + LRQV LQWURGXFHG E\ SURWRQ H[FKDQJH VR DQ\ LQVWDELOLWLHV GXH WR LRQ PRELOLW\ ZLWKLQ WKH FU\VWDO VKRXOG DOVR EH LPSURYHG 0HWDOLQGLIIXVHG ZDYHJXLGHV DOVR KDYH WKH DGGHG DGYDQWDJHV WKDW WKHLU HOHFWURRSWLF FRHIILFLHQW U LV FRPSOHWHO\ UHFRYHUDEOH E\ UHSROLQJf XQOLNH PDQ\ 3($3( ZDYHJXLGHV DQG WKH\ GRQfW FRQWULEXWH WR TXHQFKLQJ RI DFWLYH (UGRSHG JDLQ UHJLRQV ,Q WKLV VHFWLRQ WKH VWDELOLW\ RI =Q YDSRULQGLIIXVHG DQG 7LLQGLIIXVHG ZDYHJXLGHV LV H[DPLQHG WR GHWHUPLQH LI VWDELOLW\ LQ /L7D GHSHQGV RQ IDEULFDWLRQ FRQGLWLRQV VLPLODU WR /L1E 7KH GLUHFWLRQDO FRXSOHU LV XVHG DJDLQ WR PHDVXUH WKH DPRXQW RI FKDQJH LQ LQGH[ LQFUHPHQW IRU =Q YDSRULQGLIIXVHG ZDYHJXLGHV 7KHQ WKH PRGDO FKDUDFWHUL]DWLRQ WHFKn QLTXH RI 6HFWLRQ LV XVHG WR WUDFN DQG FRPSDUH FKDQJHV LQ WKH UHJLRQ RI VLQJOHPRGH RSHUDWLRQ IRU ZDYHJXLGHV IDEULFDWHG E\ 3( 7LLQGLIIXVLRQ DQG =Q YDSRULQGLIIXVLRQ

PAGE 148

'LUHFWLRQDO &RXSOHU 6WDELOLW\ 5HVXOWV RI =Q/L7D4 :DYHJXLGHV 7KH PHDVXUHG FKDQJH LQ FRXSOLQJ OHQJWK IRU D =Q YDSRULQGLI IXVHG GLUHFWLRQDO FRXSOHU ZKRVH IDEULn FDWLRQ FRQGLWLRQV ZHUH JLYHQ LQ &KDSWHU LV VKRZQ LQ )LJ 7KH PHDVXUHPHQWV ZHUH PDGH DW _LP VWDUWLQJ WZR GD\V DIWHU IDEULFDWLRQ ,PPHGLDWHO\ DIWHU IDEULFDWLRQ WKH ZDYHJXLGHV FRPSULVLQJ WKH FRXSOHU ZHUH VLQJOHPRGH DW SP +RZHYHU DIWHU WZR GD\V WKH\ GLG QRW JXLGH DW DOO DW SP 3DUDPHWHUV DERXW WKH LQGH[ SURILOH ZHUH WDNHQ IURP 5HIV >(NQ <@ ZKLFK XVHG LGHQWLFDO IDEULFDWLRQ FRQGLWLRQV DQG ZHUH XVHG LQ WKH FRPSXWHU VLPXODWLRQ WR HVWLPDWH D $Qf RI DERXW &7 IURP GD\ WR GD\ 7KLV LV FRPSDUDEOH WR WKDW VHHQ LQ 3( ZDYHJXLGHV WKRXJK LW GRHV QRW LQFOXGH WKH DSSDUHQWO\ ODUJH FKDQJH IURP GD\ WR GD\ &RPSDULVRQ RI WKH :DYHJXLGH )RUPDWLRQ 3URFHVVHV $V FDQ EH VHHQ IURP WKH DERYH UHVXOWV PHWDOLQGLIIXVLRQ DSSHDUV WR DOVR EH XQVWDn EOH LQ /L7D >0DU22@ :LWKRXW JHWWLQJ LQWR D TXDQWLWDWLYH FRPSDULVRQ EHWZHHQ $QffV RI WKH GLIIHUHQW IDEULFDWLRQ SURFHVVHV XVHG D VLPSOH FRPSDULVRQ ZDV PDGH E\ WUDFNLQJ WKH FKDQJH LQ WKH VLQJOHPRGH UHJLRQ RI RSHUDWLRQ 7R GR WKLV VWUDLJKW FKDQQHO ZDYHJXLGHV RI ZLGWKV SP LQ VWHSV RI SP ZHUH IDEULFDWHG E\ 7LLQGLIIXVLRQ =Q YDSRULQGLIIXn VLRQ DQG 3( XVLQJ SXUH JO\FHULQ 7KH PRGDO DQDO\VLV WHFKQLTXH RI 6HFWLRQ ZDV 7 , f§ ’ UUW7 Q Q , , , -4B GD\V )LJ 0HDVXUHG FRXSOLQJ OHQJWK RI WKH =Q YDSRULQGLIIXVHG VDPSOH 7KLV FRXSOHU ZDV FRPSULVHG RI SP FKDQQHOV ZLWK DQ SP JDS

PAGE 149

'D\V 3( ;FXWf 3( =FXWf =Q =FXWf 7L =FXWf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

PAGE 150

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fV PXFK KLJKHU PHOWLQJ SRLQW r&f LW LV JURZQ DW WHPSHUDWXUHV RI JUHDWHU WKDQ r& KLJKHU WKDQ WKDW IRU /L1E $W VXFK H[WUHPH WHPSHUDWXUHV /L RXWGLIIXVLRQ GXULQJ JURZWK LV D SUREOHP UHVXOWLQJ LQ QRQFRQ JUXHQW /L7D FU\VWDOV ZKLFK DUH VRPHZKDW /L GHILFLHQW FRQWDLQLQJ RQO\ PROb /L DV RSSRVHG WR PROb IRU /L1E ,Q WKLV VHFWLRQ WKH PDWHULDO JURZWK LVVXHV RXWOLQHG DERYH DUH DGGUHVVHG /L7D VDPSOHV IURP VHYHUDO JURZHUV ERWK RSWLFDO DQG VXUIDFH DFRXVWLF ZDYH 6$:f JUDGH DUH FRPSDUHG DJDLQVW HDFK RWKHU LQ WHUPV RI VWDELOLW\ 7KH LVVXH RI /L GHILFLHQF\ LV DOVR H[DPn LQHG E\ XVLQJ /LWUHDWPHQW &KDSWHU f WR LQWURGXFH PRUH /L LQWR WKH FU\VWDO LQ DQ DWWHPSW WR LPSURYH VWDELOLW\

PAGE 151

'D\V
PAGE 152

)RU WKLV FRPSDULVRQ VWUDLJKWFKDQQHO ZDYHJXLGHV RI ZLGWKV _LP LQ VWHSV RI MLP ZHUH IDEULFDWHG LQ =FXW FU\VWDOV IURP HDFK VXSSOLHU XVLQJ WKH FRQGLWLRQV IRU =Q YDSRULQGLIIXVLRQ JLYHQ LQ &KDSWHU 7KH VDPSOHV IURP 6KLQ(WVX ZHUH RI 6$: JUDGH DOO RWKHUV ZHUH FODLPHG WR EH RI RSWLFDO JUDGH 7KH PHWKRG XVHG LQ 6HFWLRQ RI WUDFNLQJ WKH FKDQJH LQ WKH UHJLRQ RI VLQJOHPRGH RSHUDWLRQ ZDV XVHG KHUH ZLWK WKH UHVXOWV SUHn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
PAGE 153

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f WKURXJK DQ )'0 FRPSXWHU VLPXODWLRQ RI D GLUHFWLRQDO FRXSOHU )RU $3( ZDYHJXLGHV LW ZDV IRXQG WKDW ORZFRQFHQWUDWLRQ ZDYHJXLGHV H[KLELWHG LPSURYHG WHPSRUDO LQVWDELOLW\ +RZHYHU WKH\ ZHUH VWLOO QRW VWDEOH HQRXJK 8VLQJ WKH WKHRU\ WKDW VLQJOHSKDVH ZDYHJXLGHV ZRXOG KDYH VXSHULRU VWDELOLW\ 3( ZDV XVHG WR IDEULFDWH ZDYHJXLGHV LQ D QHZ VLQJOH SKDVH UHJLRQ RI WKH GLDJUDP EHFDXVH RI WKH DGYDQWDJH RI 3( IRU EHWWHU PRGH FRQn ILQHPHQW +RZHYHU HYHQ ZLWK DJLQJ DW ORZ WHPSHUDWXUHV UHVXOWV LQGLFDWHG WKDW WKH VWDELOn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n LW\ RU LQGXFHG ODWWLFH VWUDLQ DQG GRHV QRW GHSHQG XSRQ IDEULFDWLRQ FRQGLWLRQV 5DWKHU LW PD\ EH LQKHUHQW WR WKH FU\VWDO WKH UHVXOW RI DQ LPSURSHU RU LPPDWXUH JURZWK SURFHVV

PAGE 154

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n SRQHQWV LV H[SORUHG )RU WKLV WKH GHVLJQ IDEULFDWLRQ DQG WHVWLQJ RI D KLJKVSHHG WUDYHOn LQJZDYH HOHFWURRSWLF PRGXODWRU LV SHUIRUPHG 7KH GHYHORSPHQW DQG GHPRQVWUDWLRQ RI VXFK D GHYLFH LOOXVWUDWHV WKH SRWHQWLDO IRU /L7D DV D YLDEOH KRVW PDWHULDO RIIHULQJ WKH DGYDQWDJH RI KLJKHU WKURXJKSXW FDSDELOLW\ SHQGLQJ WKH LPSURYHPHQW RI WKH FU\VWDO WKURXJK IXUWKHU GHYHORSPHQW RI WKH JURZWK SURFHVV

PAGE 155

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n FDWLRQV QHHGV WR EH GHYHORSHG DV ZDV GRQH IRU /L1E >,16@ ,Q WKLV UHJDUG WKLV FKDSn WHU GHPRQVWUDWHV WKH SRWHQWLDO WKDW /L7D RIIHUV IRU LQWHJUDWHGRSWLFDO GHYLFH DSSOLFDWLRQV E\ IDEULFDWLQJ DQG WHVWLQJ D WUDYHOLQJZDYH HOHFWURRSWLF PRGXODWRU 7KH VXFFHVVIXO PDQXn IDFWXULQJ RI WKLV GHYLFH ZLOO LOOXVWUDWH WKDW WKHUH DUH QR RWKHU XQIRUHVHHQ GHYLFH IDEULFDWLRQ LVVXHV WR EH FRQTXHUHG LQ /L7D VR WKDW ZKHQ FU\VWDO TXDOLW\ LV LPSURYHG PDQ\ RI WKH FXUUHQWO\ XVHG WHFKQLTXHV RI GHYLFH IDEULFDWLRQ LQ /L1E PD\ EH GLUHFWO\ DSSOLHG WR /L7D 7KLV FKDSWHU EHJLQV E\ RXWOLQLQJ WKH IDEULFDWLRQ FRQGLWLRQV XVHG LQ WKH GHYHORSn PHQW RI D 0DFK=HKQGHU LQWHUIHURPHWHU 0=,f PRGXODWRU IURP ZDYHJXLGH IRUPDWLRQ WR HOHFWURGH HOHFWURSODWLQJ 1H[W VRPH PLFURZDYH FKDUDFWHULVWLFV RI WKH WUDYHOLQJZDYH HOHFn WURGH VWUXFWXUH DUH H[DPLQHG 7KH 6MM DQG 6@ SDUDPHWHUV DUH PHDVXUHG DQG PLFURZDYH

PAGE 156

ZDYHJXLGH 0DFK=HKQGHU )LJ 'HSLFWLRQ RI D V\PPHWULF 0=, PRGXODWRU ZLWK &3: WUDYHOn LQJZDYH HOHFWURGHV LQ ;FXW /L7D ORVV HIIHFWLYH LQGH[ DQG FKDUDFWHULVWLF LPSHGDQFH DUH GHWHUPLQHG 6RPH '& UHVSRQVH PHDVXUHPHQWV DUH PDGH QH[W WR HYDOXDWH WKH VZLWFKLQJ YROWDJH ?Q DQG H[WLQFWLRQ UDWLR 7KH SRZHU KDQGOLQJ FDSDELOLW\ RI WKLV VWUXFWXUH QHDU >LP LV DOVR H[DPLQHG /DVWO\ WKH IUHTXHQF\ UHVSRQVH RI WKLV PRGXODWRU LV PHDVXUHG DQG FRPSDUHG WR WKHRU\ )DEULFDWLRQ &RQGLWLRQV $ V\PPHWULF 0=, PRGXODWRU ZDV IDEULFDWHG LQ ;FXW /L7D KDYLQJ D FRSODQDU ZDYHJXLGH &3:f WUDYHOLQJZDYH HOHFWURGH VWUXFWXUH DV GHSLFWHG LQ )LJ 7KH FRQGLn WLRQV IRU LWV IDEULFDWLRQ DUH JLYHQ LQ WKLV VHFWLRQ 7KH 0=, ZDV IDEULFDWHG ILUVW XVLQJ 3( LQ JO\FHULQ DW r& IRU KRXUV 7KH FKRLFH WR XVH 3( ZDV PHUHO\ IRU VLPSOLFLW\ $Q\ RI WKH RWKHU ZDYHJXLGH IDEULFDWLRQ SURn FHVVHV FRXOG KDYH EHHQ XVHG 1H[W WKH VDPSOH HQGIDFHV ZHUH SROLVKHG XVLQJ WKH WHFKn QLTXH RXWOLQHG LQ 6HFWLRQ )LQDOO\ WR PLQLPL]H UI ORVV LQ WKH HOHFWURGHV WKLFN

PAGE 157

HOHFWURGHV ZHUH IDEULFDWHG E\ $X HOHFWURSODWLQJ XVLQJ WKH SURFHVV GHVFULEHG LQ 6HFWLRQ 7KH LQWHUDFWLRQ OHQJWK ZDV FP 7KH FHQWHU HOHFWURGH ZLGWK DQG WKH JDSV EHWZHHQ HOHFn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

PAGE 158

)LJ 0HDVXUHG 6L GRWWHGf DQG ILWWHG IUHTXHQF\ GHSHQGHQW ORVV IXQFWLRQ D D9I VROLGf ZKHUHD O2G%FPr 9*+] 63DUDPHWHUV DQG /RVV 7KH PLFURZDYH WUDQVPLVVLRQ FKDUDFWHULVWLFV 6SDUDPHWHUV RI WKH &3: HOHFWURGH ZHUH PHDVXUHG ILUVW XVLQJ DQ +3& QHWZRUN DQDO\]HU DQG D 8QLYHUVDO 7HVW )L[WXUH IURP :LOWURQ 7KH PHDVXUHG DQG FDOLEUDWHG PLFURZDYH UHIOHFWLRQ 6MMf DQG WUDQVPLVVLRQ 6Lf FKDUDFWHULVWLFV IRU WKH HOHFWURGH DUH VKRZQ LQ )LJ 7R FRQVWUXFW WKLV SORW WKH 6M@ DQG 6L SDUDPHWHUV RI WKH HOHFWURGHV ZHUH PHDVXUHG 7KHQ WKH VDPH SDUDPHWHUV ZHUH PHDVXUHG RQ D FP &DOLEUDWLRQ .LW IURP :LOWURQ GHVLJQHG WR VLPXODWH WKH WHVW IL[WXUH KROGLQJ WKH VDPSOH 7KH PHDVXUHPHQWV GRQH RQ WKH FP &DOLEUDWLRQ .LW ZHUH WKHQ XVHG WR FDOLEUDWH RXW WKH HIIHFW RI WKH WHVW IL[WXUH DQG SURGXFH WKH 6@ DQG 6 FXUYHV IRU WKH HOHFn WURGH RQO\ DV VKRZQ LQ )LJ %\ ILWWLQJ WKH 6 FXUYH ZLWK D IUHTXHQF\GHSHQGHQW ORVV IXQFWLRQ RI D D9I D PLFURZDYH ORVV FRHIILFLHQW RI D G%FP} -*+] ZDV REWDLQHG DV GHSLFWHG LQ )LJ

PAGE 159

0LFURZDYH (IIHFWLYH ,QGH[ 7KH PLFURZDYH HIIHFWLYH LQGH[ 1P ZDV PHDVXUHG QH[W 1RUPDOO\ 1P FDQ EH HVWLn PDWHG IURP WKH IROORZLQJ HTXDWLRQ f ZKHUH H[[ DQG H]] DUH WKH UHODWLYH GLHOHFWULF FRQVWDQWV RI WKH RUGLQDU\ DQG H[WUDRUGLQDU\ D[HV UHVSHFWLYHO\ 7KLV JLYHV D YDOXH RI DERXW IRU 1P +RZHYHU E\ XVLQJ WKH QHWZRUN DQDO\]HU WR PHDVXUH WKH JURXS GHOD\ DORQJ WKH HOHFWURGH 1P FDQ EH FDOFXODWHG GLUHFWO\ XVLQJ f ZKHUH F LV WKH VSHHG RI OLJKW LV WKH PHDVXUHG JURXS GHOD\ DQG / LV WKH WRWDO HOHFWURGH OHQJWK $V EHIRUH WKH JURXS GHOD\ WKURXJK WKH FP &DOLEUDWLRQ .LW ZDV DOVR PHDVXUHG WR FDOLEUDWH RXW WKH HIIHFW RI WKH WHVW IL[WXUH 7KH UHVXOWLQJ PHDVXUHG JURXS GHOD\ IRU WKH IDEn ULFDWHG PRGXODWRU ZDV SV ZKLFK DORQJ ZLWK / FP JDYH D PLFURZDYH HIIHFWLYH LQGH[ RI 1P &KDUDFWHULVWLF ,PSHGDQFH 7KH FKDUDFWHULVWLF LPSHGDQFH = RI WKH HOHFWURGH ZDV HVWLPDWHG IURP SORWV RI = YHUVXV HOHFWURGH ZLGWKWRJDS UDWLR :* >&KX .XE@ )RU WKH PRGXODWRU IDEULFDWHG KHUH :* O UHVXOWLQJ LQ D FKDUDFWHULVWLF LPSHGDQFH RI FORVH WR  ,Q WKH QH[W VHFWLRQ WKH '& UHVSRQVH RI WKH PRGXODWRU LV PHDVXUHG LQ RUGHU WR GHWHUn PLQH WKH VZLWFKLQJ YROWDJH 9Q DQG WKH H[WLQFWLRQ UDWLR ,Q DGGLWLRQ WKH SRZHU KDQGOLQJ FDSDELOLW\ RI WKH PRGXODWRU QHDU _LP LV H[DPLQHG

PAGE 160

'& 5HVSRQVH DQG 3RZHU +DQGOLQJ 7KH '& UHVSRQVH RI D PRGXODWRU LV D SORW RI WKH PRGXODWRU RXWSXW YHUVXV DSSOLHG '& YROWDJH )URP WKLV SORW SDUDPHWHUV VXFK DV WKH KDOIZDYH YROWDJH RU VZLWFKLQJ YROWn DJH 9MM DQG H[WLQFWLRQ UDWLR FDQ EH GHWHUPLQHG 7KH '& UHVSRQVH IRU WKH PRGXODWRU IDEULn FDWHG KHUH LV PHDVXUHG LQ WKLV VHFWLRQ ,Q DGGLWLRQ WKH SRZHU KDQGOLQJ FDSDELOLW\ RI WKLV GHYLFH LV DOVR FKDUDFWHUL]HG IRU RSHUDWLRQ QHDU SP '& 5HVSRQVH 7KH VHWXS IRU PHDVXULQJ '& UHVSRQVH LV VKRZQ LQ )LJ /LJKW DW SP IURP D WXQDEOH ODVHU VRXUFH ZDV FRXSOHG LQWR WKH PRGXODWRU YLD D VLQJOHPRGH SRODUL]DWLRQPDLQn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f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

PAGE 161

)LJ 6FKHPDWLF RI WKH VHWXS IRU PHDVXULQJ WKH '& UHVSRQVH RI DQ HOHFn WURRSWLF PRGXODWRU DSSOLHG YROWDJH 9f )LJ 0HDVXUHG '& UHVSRQVH VTXDUHVf DQG WKHRUHWLFDO ILW VROLG OLQHf RI WKH 0=, PRGXODWRU

PAGE 162

)LJ 0HDVXUHG LQSXW SRZHU YHUVXV RXWSXW SRZHU DW SP IRU WKH 0=, PRGXODWRU IDEULFDWLRQ FRQGLWLRQV LV QRW NQRZQ ,W PD\ EH SRVVLEOH WKDW KLJK FRQFHQWUDWLRQ ZDYHJXLGHV FRXOG EH DQQHDOHG WR UHFRYHU U ZLWKRXW DFWXDOO\ EHLQJ DQQHDOHG HQRXJK WR EH RI VLJQLILFDQWO\ ORZHU VWUDLQ RU HYHQ VLQJOHSKDVH 3RZHU +DQGOLQJ 1HDU SP ,Q RUGHU WR WHVW WKH SRZHU KDQGOLQJ FDSDELOLW\ RI WKLV GHYLFH QHDU SP WKH WHFKn QLTXH RI 6HFWLRQ ZDV HPSOR\HG 7KH UHVXOWLQJ SORW RI RXWSXW SRZHU YHUVXV LQSXW SRZHU LV VKRZQ LQ )LJ $V FDQ EH VHHQ WKH PRGXODWRU H[KLELWHG D OLQHDU UHVSRQVH LQGLFDWLQJ QR SKRWRUHIUDFWLYH GDPDJH IRU LQSXW SRZHUV XS WR QHDUO\ P: 7KH IDFW WKDW WKH UHVSRQVH RI WKLV GHYLFH GHYLDWHG IURP OLQHDULW\ DW KLJK LQSXW SRZHUV ZKHQ WKDW RI VWUDLJKW FKDQQHO ZDYHJXLGHV GLG QRW LV OLNHO\ GXH WR VRPH VOLJKWO\ DV\PPHWU\ LQ WKH 0=, DUPV RU D QRQSHUIHFW VSOLW DW WKH
PAGE 163

MXLQQ N+] 9 SS '& VRXUFH IUHTXHQF\ JHQHUDWRU 1$ 7HVW 6HW SRUW SRUW 1$ 6ZHHSHU Q PLFURZDYH FDEOH OHQV 00 ILEHU GHWHFWRU WHVW IL[WXUH 60 ILEHU ORFNLQ DPSOLILHU )LJ 6HWXS XVHG IRU PHDVXULQJ IUHTXHQF\ UHVSRQVH E\ WKH VPDOO VLJQDO IUHTXHQF\VZHSW PHWKRG

PAGE 164

,Q WKH QH[W VHFWLRQ WKH IUHTXHQF\ UHVSRQVH IRU WKLV GHYLFH LV PHDVXUHG DW IUHTXHQn FLHV IURP 0+] WR DERYH *+] 7KH PHDVXUHG UHVSRQVH LV WKHQ FRPSDUHG WR WKHRU\ )UHTXHQF\ 5HVSRQVH 7KH IUHTXHQF\ UHVSRQVH IRU WKLV GHYLFH ZDV PHDVXUHG E\ XVLQJ WKH VPDOO VLJQDO IUHn TXHQF\VZHSW PHWKRG >8HK :DQ@ 7KH VHWXS LV VKRZQ LQ )LJ /LJKW DW SP IURP D WXQDEOH ODVHU VRXUFH ZDV HQGILUH FRXSOHG LQWR WKH 0=, XVLQJ D VLQJOHPRGH SRODUn L]DWLRQPDLQWDLQLQJ ILEHU $ N+] 9 SHDNWRSHDN 9 WR 9f VLJQDO IURP D IUHTXHQF\ JHQHUDWRU ZDV XVHG WR $0 PRGXODWH WKH PLFURZDYH VLJQDO ZKLFK ZDV FRXSOHG WR WKH WUDYn HOLQJZDYH HOHFWURGH YLD FDEOHV IURP WKH QHWZRUN DQDO\]HU 6SDUDPHWHU WHVW VHW 7KLV VDPH VLJQDO ZDV XVHG IRU SKDVH PDWFKLQJ DW D ORFNLQ DPSOLILHU 7KH PRGXODWRU ZDV ELDVHG DW LWV PLQLPXP RXWSXW SRLQW E\ DSSO\LQJ D '& ELDV YROWDJH WR WKH ELDV SRUW RI WKH QHWZRUN DQDn O\]HU WHVW VHW 7KH PRGXODWHG RSWLFDO RXWSXW OLJKW ZDV FDSWXUHG ZLWK D PXOWLPRGH ILEHU DQG IRFXVHG E\ D OHQV RQWR D ORZVSHHG ,5 GHWHFWRU 7KH PLFURZDYH IUHTXHQF\ ZDV PDQXn DOO\ VZHSW IURP 0+] WR DERYH *+] DQG WKH DPSOLWXGH RI WKH ORZIUHTXHQF\ HQYHORSH ZKLFK ZDV SURSRUWLRQDO WR WKH KLJKVSHHG PRGXODWLRQ HIILFLHQF\ ZDV PHDVXUHG E\ XVLQJ WKH ORZVSHHG SKRWRn GHWHFWRU DQG WKH ORFNLQ DPSOLILHU IUHTXHQF\ *+]f $IWHU PHDVXULQJ WKH IUHTXHQF\ )LJ 3RZHL FDOLELDWLRQ SORW GHSLFW UHVSRQVH RI WKH PRGXODWRU WKH SRZHU RXWn LQJ VZHHSHU RXWSXW SRZHU YHLVXV ILHTXHQF\ SXW 4I WKH PLFURZD9H VZHHSHU VRXUFH P R 7  4 ff§} R K t e FQ SUU D U7f§UaUQ Ur YL LIW ,Q I 3IIW7f @ Q U 7 5 L Q L Q ‘ ‘ ‘

PAGE 165

)LJ 0HDVXUHG IUHTXHQF\ UHVSRQVH VTXDUHVf DQG FDOFXODWHG WKHRn UHWLFDO UHVSRQVH VROLG OLQHf IRU WKH WUDYHOLQJZDYH 0=, PRGXn ODWRU QHHGHG WR EH PHDVXUHG VR WKH FROOHFWHG GDWD FRXOG EH SURSHUO\ FDOLEUDWHG 7KLV ZDV GRQH E\ DWWDFKLQJ WKH PLFURZDYH FDEOH OHDG IURP WKH VZHHSHU GLUHFWO\ WR D FDOLEUDWHG PLFURZDYH SRZHU PHWHU DQG PHDVXULQJ SRZHU RYHU WKH IUHTXHQF\ UDQJH RI LQWHUHVW 7KH UHVXOWLQJ SORW RI VZHHSHU RXWSXW SRZHU YHUVXV IUHTXHQF\ LQ VKRZQ LQ )LJ $V FDQ EH VHHQ WKH SRZHU RXW RI WKH VZHHSHU FKDQJHV VLJQLILFDQWO\ ZLWK IUHTXHQF\ 7KH SRZHU FDOLEUDWLRQ GDWD RI )LJ ZDV WKHQ XVHG WR FDOLEUDWH WKH PHDVXUHG IUHn TXHQF\ UHVSRQVH IRU WKH PRGXODWRU 7KH UHVXOWLQJ PHDVXUHG DQG FDOLEUDWHG UHVSRQVH LV VKRZQ LQ )LJ $OVR VKRZQ LV WKH WKHRUHWLFDO FXUYH IRU WKH IUHTXHQF\ UHVSRQVH FDOFXn ODWHG XVLQJ FORVHG IRUP H[SUHVVLRQV >1D]@ ZLWK DQ RSWLFDO UHIUDFWLYH LQGH[ 1 PLFURZDYH HIIHFWLYH LQGH[ 1P LQWHUDFWLRQ OHQJWK / FP PLFURZDYH ORVV FRHIILFLHQW D 72G%FPr 9*+] FKDUDFWHULVWLF LPSHGDQFH = 2e DQG VRXUFH LPSHGDQFH RI 

PAGE 166

7KH PHDVXUHG UHVSRQVH VKRZHG D ODUJH GHYLDWLRQ IURP WKH WKHRUHWLFDO UHVSRQVH DW ORZ IUHTXHQFLHV O*+]f $V FDQ EH VHHQ IURP )LJV DQG WKHVH IUHTXHQFLHV FRUUHn VSRQG WR DQ XQXVXDOO\ ODUJH PLFURZDYH UHIOHFWLRQ SHDN DQG KHQFH D GLS LQ PLFURZDYH WUDQVPLVVLRQ 7KH HOHFWURGH VWUXFWXUH IDEULFDWHG KHUH ZDV VLPLODU WR WKDW XVHG LQ 5HI >:DQ@ KDYLQJ WKH H[DFW VDPH YDOXH RI PLFURZDYH ORVV FRHIILFLHQW EXW RQ D /L7D VXEn VWUDWH LQVWHDG RI D /L1E VXEVWUDWH ,W KDV EHHQ HVWDEOLVKHG >&KX@ WKDW WKH LPSHGDQFH PLVPDWFK RI /L7D LV PRUH VHULRXV WKDQ WKDW RI /L1E OHDGLQJ WR PXFK KLJKHU PLFUR ZDYH UHIOHFWLRQV LQ /L7D IRU WKH VDPH :* UDWLR RI WKH HOHFWURGHV 7KLV RI FRXUVH FDQ EH LPSURYHG E\ FKDQJLQJ WKH :* UDWLR RI WKH HOHFWURGHV )XUWKHU GHYLDWLRQV RI WKH PHDn VXUHG GDWD IURP WKDW RI WKH WKHRUHWLFDO FDOFXODWLRQ PD\ EH WKH UHVXOW RI D PLFURZDYH LQGH[ ZKLFK LV IUHTXHQF\ GHSHQGHQW 6XPPDU\ ,Q WKLV FKDSWHU WKH SRWHQWLDO IRU /L7D DV D KRVW PDWHULDO IRU LQWHJUDWHGRSWLFDO GHYLFH DSSOLFDWLRQV ZDV GHPRQVWUDWHG E\ WKH IDEULFDWLRQ DQG WHVWLQJ RI D V\PPHWULF 0=, PRGXODWRU ZLWK D &3: WUDYHOLQJZDYH HOHFWURGH VWUXFWXUH RQ LW 7KH 0=, ZDV IDEULFDWHG E\ 3( LQ JO\FHULQ DQG WKH HOHFWURGHV ZHUH $X HOHFWURSODWHG ZLWK D :* UDWLR RI 1R IDEn ULFDWLRQ LVVXHV ZHUH XQFRYHUHG ZKLFK ZRXOG SUHYHQW WKH GHYHORSPHQW RI DGGLWLRQDO GHYLFHV LQGLFDWLQJ WKDW ZKHQ FU\VWDO TXDOLW\ DQG KHQFH VWDELOLW\ LV LPSURYHG FXUUHQW IDEULn FDWLRQ WHFKQLTXHV XVHG LQ /L1E PD\ EH GLUHFWO\ DSSOLHG WR /L7D IRU WKH SURGXFWLRQ RI DQ DUUD\ RI GHYLFHV 7KH PLFURZDYH FKDUDFWHULVWLFV RI WKH GHYLFH ZHUH H[DPLQHG QH[W 7KH 6 DQG SDUDPHWHUV ZHUH PHDVXUHG DQG D PLFURZDYH ORVV FRHIILFLHQW RI G%FP} 9*+] ZDV

PAGE 167

GHWHUPLQHG %\ PHDVXULQJ JURXS GHOD\ DORQJ WKH HOHFWURGH VWUXFWXUH D PLFURZDYH HIIHFn WLYH LQGH[ RI ZDV FDOFXODWHG $ FKDUDFWHULVWLF LPSHGDQFH RI e ZDV IRXQG IURP SORWV RI LPSHGDQFH YHUVXV HOHFWURGH :* UDWLR 7KH '& UHVSRQVH ZDV WKHQ PHDVXUHG DQG WKH SRZHU KDQGOLQJ FDSDELOLW\ RI WKH GHYLFH QHDU _LP ZDV H[DPLQHG )URP WKH PHDVXUHG '& UHVSRQVH DQ H[WLQFWLRQ UDWLR RI G% ZDV REWDLQHG DORQJ ZLWK D VZLWFKLQJ YROWDJH RI 9 DW '& $V IDU DV SRZHU KDQn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n WLRQV RI WKLV HIIRUW DUH GHWDLOHG DQG VRPH IXWXUH ZRUN UHPDLQLQJ WR EH SHUIRUPHG LV RXWn OLQHG

PAGE 168

&+$37(5 &21&/86,216 $1' )8785( :25. &RQFOXVLRQV 6HYHUDO FRQWULEXWLRQV WRZDUG WKH LPSOHPHQWDWLRQ RI LQWHJUDWHGRSWLFDO GHYLFHV EHQn HILWLQJ IURP WKH KLJK SKRWRUHIUDFWLYH WKUHVKROG DQG VKRUWHU ZDYHOHQJWK 89 DEVRUSWLRQ HGJH RI /L7D KDYH UHVXOWHG EDVHG RQ WKLV UHVHDUFK ,Q SDUWLFXODU D FRPSOHWH FKDUDFWHUn L]DWLRQ RQ WKH HIIHFW RI (UGRSLQJ IRU DFWLYH ZDYHJXLGH ODVHUV DQG RSWLFDO GDWD VWRUDJH DSSOLFDWLRQV KDV EHHQ SHUIRUPHG RSHQLQJ WKH GRRU IRU WKH GHYHORSPHQW RI VXFK GHYLFHV ZKLFK ZHUH VKRZQ WR KDYH VHYHUDO DGYDQWDJHV ZKHQ IDEULFDWHG LQ /L7D $GGLWLRQDOO\ YDULRXV PHWKRGV RI ZDYHJXLGH IDEULFDWLRQ QDPHO\ $3( 3( 7LLQGLIIXVLRQ DQG =Q YDSRU LQGLIIXVLRQ ZHUH UHVHDUFKHG LQ GHWDLO )URP WKLV WKH VWUXFWXUDO SKDVH GLDJUDP IRU $3(/L7D ZDV FRQVWUXFWHG DQG XVHG WR H[SODLQ WKH SUHYLRXVO\ REVHUYHG DQRPDOLHV DVVRn FLDWHG ZLWK WKLV SURFHVV LQ /L7D DQG DOVR WR LGHQWLI\ $3(3( IDEULFDWLRQ FRQGLWLRQV \LHOGLQJ ZDYHJXLGHV H[KLELWLQJ LPSURYHG WHPSRUDO VWDELOLW\ )LQDOO\ WKH RQO\ IDFWRU KLQn GHULQJ WKH ZLGHVSUHDG XVH RI /L7D QDPHO\ WKH LQVWDELOLW\ RI ZDYHJXLGHV ZDV IRXQG WR QRW EH GHSHQGHQW XSRQ IDEULFDWLRQ PHWKRGV EXW UDWKHU WR EH LQKHUHQW WR WKH FU\VWDO LWVHOI WKH OLNHO\ UHVXOW RI D VOLJKWO\ XQGHUGHYHORSHG JURZWK WHFKQRORJ\ IRU WKLV FU\VWDO 7KH PDMRU DFFRPSOLVKPHQWV IRU WKLV ZRUN DUH GHWDLOHG EHORZ

PAGE 169

7KH DFFRPSOLVKPHQWV UHODWHG WR (UGRSLQJ RI /L7D LQFOXGH ILUVW WKH GHILQLWLRQ RI FRQGLWLRQV IRU (UGRSLQJ ZKLFK SURYLGH DQ (U FRQFHQWUDWLRQ SURILOH DSSURSULDWH IRU DFWLYH JXLGHGZDYH GHYLFHV 6HFRQG WKH SUHVHQFH RI (U ZDV IRXQG WR OHDG WR XSFRQ YHUWHG HPLVVLRQV ZKHUH HQHUJ\WUDQVIHUXSFRQYHUVLRQ (78f IURP (U FOXVWHUV ZDV GLVn FRYHUHG WR EH WKH GRPLQDQW PHFKDQLVP RI WKLV XSFRQYHUVLRQ 7KLUG WKURXJK WKH WZR SKRWRQ RU WZRFRORUf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n GXFHG DQG XVHG WR GLUHFWO\ GHWHUPLQH YDOXHV RI ZDYHJXLGH VXUIDFH LQGH[ LQFUHPHQW RYHUn FRPLQJ SUHYLRXV OLPLWDWLRQV DVVRFLDWHG ZLWK LQGLUHFW WHFKQLTXHV ZKLFK KDG SUHYHQWHG WKH FRQVWUXFWLRQ RI WKH VWUXFWXUDO SKDVH GLDJUDP IRU $3(/L7D 7KLUG XVLQJ GLUHFW PHDn VXUHPHQW WHFKQLTXHV WKH VWUXFWXUDO SKDVH GLDJUDP IRU $3(/L7D ZDV FRQVWUXFWHG 7KLV GLDJUDP ZDV XVHG WR H[SODLQ VHYHUDO RI WKH SUHYLRXVO\ REVHUYHG DQRPDOLHV DVVRFLDWHG ZLWK WKH $3( SURFHVV LQ /L7D $GGLWLRQDOO\ WKH LPSDFW RI FU\VWDO SKDVHV RQ WKH VKDSH RI

PAGE 170

ZDYHJXLGH LQGH[ SURILOHV ZDV H[DPLQHG )RXUWK WKH DGYDQWDJHV RI 3( ZDYHJXLGHV RYHU $3( ZDYHJXLGHV ZHUH GHWDLOHG VKRZLQJ EHWWHU RSWLFDO PRGH FRQILQHPHQW DQG WKH H[LVWn HQFH RI D QHZ VLQJOHSKDVH UHJLRQ RQ WKH VWUXFWXUDO SKDVH GLDJUDP REWDLQDEOH RQO\ E\ 3( LQ D GLOXWH VRXUFH )LQDOO\ WKHVH 3( ZDYHJXLGHV ZHUH VKRZQ WR KDYH VXSHULRU SKRWRUHIUDF WLYH UHVLVWDQFH QHDU _-P WKDQ WKHLU FRXQWHUSDUWV LQ /L1E 2Q WKH LVVXH RI WHPSRUDO VWDELOLW\ RI ZDYHJXLGHV LQ /L7D D QXPEHU RI DFFRPn SOLVKPHQWV ZHUH PDGH )LUVW URFNLQJ FXUYH DQG GLUHFWLRQDO FRXSOHU PHDVXUHPHQWV ZHUH XVHG WR VKRZ WKDW $3( ZDYHJXLGHV LQ KLJK FRQFHQWUDWLRQ SKDVHV DUH YHU\ XQVWDEOH 6HFn RQG E\ FRPSDULVRQ RI ORZ FRQFHQWUDWLRQ $3( DQG VLQJOHSKDVH 3( ZDYHJXLGHV LW ZDV IRXQG WKDW VWDELOLW\ ZDV QRW LPSURYHG E\ IDEULFDWLQJ ZDYHJXLGHV FRPSULVHG RI RQO\ D VLQn JOH SKDVH DV LV WKH FDVH ZLWK /L1E DQG ZDV DVVXPHG WR EH WKH FDVH ZLWK /L7D 5DWKHU VWDELOLW\ DSSHDUHG WR EH LPSURYHG E\ VLPSO\ UHGXFLQJ VWUDLQ RU SURWRQ FRQFHQWUDn WLRQ +RZHYHU WKHUH ZDV D OLPLW WR ZKLFK SURWRQ FRQFHQWUDWLRQ FDQ EH UHGXFHG HVSHFLDOO\ DW _LP EHFDXVH RI WKH H[WUHPHO\ ORZ YDOXHV RI LQGH[ LQFUHPHQW DW WKLV ZDYHOHQJWK 7KLUG D FRPSDULVRQ RI WKH VWDELOLW\ RI 3( 7LLQGLIIXVHG DQG =Q YDSRULQGLIIXVHG ZDYHJXLGHV ZKHUH DOO ZHUH IRXQG WR EH XQVWDEOH VKRZHG WKDW LQVWDELOLW\ GRHV QRW RULJLn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

PAGE 171

KLJKTXDOLW\ PDWHULDO GDPSHQV PRWLYDWLRQ IRU WKH LPSURYHPHQW RI LWV JURZWK WHFKQRORJ\ )LQDOO\ WR GHPRQVWUDWH WKH SRWHQWLDO RI /L7D DQG KRSHIXOO\ LQLWLDWH PDUNHW LQWHUHVW IRU KLJKHU TXDOLW\ PDWHULDO D KLJKVSHHG WUDYHOLQJZDYH HOHFWURRSWLF PRGXODWRU ZDV IDEULn FDWHG DQG FKDUDFWHUL]HG ZKLFK FRPSDUHG ZHOO ZLWK WKHRUHWLFDO SUHGLFWLRQ 7KHVH FRQWULEXWLRQV DUH VLJQLILFDQW WR WKH IXUWKHU H[SORLWDWLRQ RI /L7D IRU GHYLFH IDEULFDWLRQ 7KH\ GHPRQVWUDWH WKDW WKLV PDWHULDO KDV JUHDW SRWHQWLDO DQG DGYDQWDJHV LQ WKH IDEULFDWLRQ RI D QXPEHU RI KLJKWKURXJKSXW LQWHJUDWHGRSWLFDO GHYLFHV VXFK DV ODVHUV DQG PRGXODWRUV DV ZHOO DV IRU DSSOLFDWLRQV LQYROYLQJ SKRWRUHIUDFWLYH PDWHULDOV VXFK DV KRORn JUDSKLF GDWD VWRUDJH DQG SKRWRUHIUDFWLYH JUDWLQJV 7KLV LV LPSRUWDQW EHFDXVH LW HPSKDVL]HV WKH QHHG IRU KLJKTXDOLW\ RSWLFDO JUDGH FU\VWDOV ZKLFK LQ WXUQ HQFRXUDJHV FU\VWDO JURZHUV WR VSHQG WLPH DQG PRQH\ RQ WKH DGYDQFHPHQW RI JURZWK WHFKQRORJ\ DQG LPSURYHG VWRLFKLn RPHWU\ IRU WKLV PDWHULDO EH\RQG WKDW RI PHUHO\ 6$: JUDGH 6RPH IXWXUH ZRUN UHODWHG WR WKLV UHVHDUFK LV RXWOLQHG LQ WKH QH[W VHFWLRQ ,Q SDUWLFn XODU WKH DSSOLFDWLRQ RI UHVXOWV REWDLQHG KHUH WRZDUG WKH GHYHORSPHQW RI LQWHJUDWHGRSWLFDO GHYLFHV LV GLVFXVVHG )XWXUH :RUN 7KRXJK WKLV UHVHDUFK KDV JRQH D ORQJ ZD\ WRZDUG WKH GHYHORSPHQW RI LQWHJUDWHG RSWLFDO GHYLFHV LQ /L7D WKHUH VWLOO UHPDLQV WKH REVWDFOH RI FU\VWDO LQVWDELOLW\ 7KLV OLPLn WDWLRQ LV OLNHO\ WKH UHVXOW RI DQ LPPDWXUH JURZWK WHFKQRORJ\ IRU WKLV FU\VWDO SRVVLEO\ UHODWHG WR GHIHFWV DQG VWUDLQ FDXVHG E\ LWV KLJK JURZWK WHPSHUDWXUH DQG FRUUHVSRQGLQJ /L GHILFLHQF\ +RZHYHU WKH H[LVWHQFH RI PDUNHW GHPDQG IRU EHWWHU FU\VWDOV VKRXOG UHVXOW LQ UHVHDUFK WRZDUG WKH LPSURYHPHQW RI JURZWK WHFKQLTXHV 7KRXJK VXFK ZRUN LV RXWVLGH WKH

PAGE 172

VFRSH RI WKLV UHVHDUFK WKHUH VWLOO UHPDLQV VRPH ZRUN UHPDLQLQJ WR EH SHUIRUPHG RQ WKH GHYHORSPHQW RI QHZ GHYLFH WHFKQRORJLHV LQ WKLV PDWHULDO DV ZHOO DV WKH LPSURYHPHQW RI H[LVWLQJ GHYLFHV 7KLV ZRUN LV GHVFULEHG LQ WKLV VHFWLRQ 0RGXODWRU ,PSURYHPHQWV 7KH PHDVXUHG IUHTXHQF\ UHVSRQVH RI WKH PRGXODWRU IDEULFDWHG DQG WHVWHG LQ &KDSn WHU LV UHSULQWHG DJDLQ LQ )LJ $V FDQ EH VHHQ WKHUH ZDV D ODUJH GHYLDWLRQ IURP WKH WKHRUHWLFDO UHVSRQVH DW ORZ IUHTXHQFLHV O*+]f 7KHVH IUHTXHQFLHV ZHUH IRXQG WR FRUUHn VSRQG WR DQ XQXVXDOO\ ODUJH PLFURZDYH UHIOHFWLRQ SHDN DQG KHQFH D GLS LQ PLFURZDYH WUDQVPLVVLRQ 7KH HOHFWURGH VWUXFWXUH IDEULFDWHG KHUH ZDV WKH VLPLODU WR WKDW XVHG LQ 5HI >:DQ@ KDYLQJ WKH H[DFW VDPH YDOXH RI PLFURZDYH ORVV FRHIILFLHQW EXW RQ D /L7D VXEn VWUDWH LQVWHDG RI D /L1E VXEVWUDWH )LJ 0HDVXUHG IUHTXHQF\ UHVSRQVH VTXDUHVf DQG FDOFXODWHG WKHRn UHWLFDO UHVSRQVH VROLG OLQHf IRU WKH WUDYHOLQJZDYH 0=, PRGXn ODWRU

PAGE 173

,W KDV EHHQ HVWDEOLVKHG >&KX@ WKDW WKH LPSHGDQFH PLVPDWFK RI /L7D LV PRUH VHULRXV WKDQ WKDW RI /L1E OHDGLQJ WR PXFK KLJKHU PLFURZDYH UHIOHFWLRQV LQ /L7D IRU WKH VDPH :* UDWLR RI WKH HOHFWURGHV $V D UHVXOW IUHTXHQF\ UHVSRQVH FRXOG OLNHO\ EH LPSURYHG E\ FKDQJLQJ WKH :* UDWLR RI WKH HOHFWURGHV WR PLQLPL]H PLFURZDYH UHIOHFWLRQV )XUWKHU DGYDQFHV WR WKLV GHYLFH FRXOG DOVR EH PDGH E\ UHGXFLQJ WKH KDOIZDYH YROWn DJH :. $V WKH WHVWHG PRGXODWRU ZDV PDGH E\ 3( LW KDG DQ HOHFWURRSWLF FRHIILFLHQW U ZKLFK ZDV QHDUO\ KDOI WKDW RI WKH EXON YDOXH 7R LPSURYH WKLV WKH GHYLFH FRXOG EH PDGH E\ PHWDOLQGLIIXVLRQ IROORZHG E\ UHSROLQJ VKRZQ LQ WKLV UHVHDUFK WR UHWXUQ WKH U WR QHDUO\ ZDYHJXLGH 0DFK=HKQGHU PLFURZDYH VRXUFH GRPDLQ UHYHUVDO )LJ +LJKVSHHG WUDYHOLQJZDYH 0DFK=HKQGHU LQWHUIHURPHWHU PRGXn ODWRU LQ Df VWDQGDUG FRQILJXUDWLRQ ZLWK EDVHEDQG IUHTXHQF\ UHVSRQVH DQG LQ Ef GRPDLQUHYHUVHG FRQILJXUDWLRQ ZLWK EDQGSDVV UHVSRQVH

PAGE 174

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

PAGE 175

ZDYHJXLGH ODVHUV RSHUDWLQJ QHDU WKH FRPPXQLFDWLRQV ZDYHOHQJWK RI SP 2I SDUWLFXODU LQWHUHVW LV WKH LQWHJUDWLRQ RI VXFK D ODVHU ZLWK DQ HOHFWURRSWLF PRGXODWRU RQ D VLQJOH FRUQ SDFW PRGXOH DV LOOXVWUDWHG LQ )LJ 6XFK D PRGXOH KDV WKH DGYDQWDJHV RI ORZ FKLUS DQG HOLPLQDWHV WKH QHHG IRU FRXSOLQJ EHWZHHQ D VHSDUDWH ODVHU VRXUFH DQG PRGXODWRU ,Q WKLV FDVH WKH ODVHU FDYLW\ UHIOHFWRUV DUH SURYLGHG E\ D PXOWLOD\HU GLHOHFWULF PLUURU DQG D FRUUXn JDWHG ZDYHJXLGH UHIOHFWRU IDEULFDWHG XVLQJ D 6L RYHUOD\ >+XV@ 7KH ZDYHJXLGH DW OHDVW LQ WKH ODVHU VHFWLRQ ZRXOG OLNHO\ EH IDEULFDWHG E\ =Q YDSRULQGLIIXVLRQ EHFDXVH RI LWV KLJK SKRWRUHIUDFWLYH UHVLVWDQFH DQG ODFN RI +LQGXFHG JDLQ TXHQFKLQJ 3KRWRUHIUDFWLYH *UDWLQJ $V ZDV GHWDLOHG LQ &KDSWHU (U/L7D RIIHUV VRPH DGYDQWDJH RYHU (U/L1E WRZDUG SKRWRUHIUDFWLYH 35f DSSOLFDWLRQV VXFK DV RSWLFDO GDWD VWRUDJH DQG 35 JUDWLQJ IDEULn FDWLRQ 7R WKLV HQG WKH IDEULFDWLRQ RI D 35 JUDWLQJ E\ WZRFRORU UHFRUGLQJ FRXOG EH UHDOn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n GDUG WZREHDP LQWHUIHUHQFH VHWXS >+XV@ DQG D VRXUFH RI JDWLQJ OLJKW $ QRQPRQRFKUR PDWLF VRXUFH RI JDWLQJ OLJKW LQ WKH YLROHWQHDU 89 UDQJH HJ PHUFXU\ RU [HQRQ ODPSf PD\ EH LQFLGHQW RQ WKH VDPSOH DQG DQ LQWHUIHUHQFH SDWWHUQ IRUPHG ZLWK DLG RI D SRZHUIXO UHG ODVHU +H1Hf WR UHFRUG WKH JUDWLQJ

PAGE 176

$33(1',; $ ',63(56,21 2) 587,/( 7KH GLVSHUVLRQ GDWD IRU UXWLOH LV JLYHQ LQ 7DEOH $O 7KLV GDWD LV WDNHQ IURP WKH PDQXIDFWXUHUfV OLWHUDWXUH ,QGH[ YDOXHV DW DQ DUELWUDU\ ZDYHOHQJWK PD\ EH IRXQG E\ FXELF VSOLQH LQWHUSRODWLRQ

PAGE 177

:DYHOHQJWK 0Pf Q 7DEOH $ 'LVSHUVLRQ GDWD IRU UXWLOH

PAGE 178

$33(1',; % ',63(56,21 2) /,7+,80 7$17$/$7( 7KH GLVSHUVLRQ RI WKH RUGLQDU\ DQG H[WUDRUGLQDU\ UHIUDFWLYH LQGLFHV RI /L7D LV JLYHQ LQ 7DEOH % 7KLV GDWD LV WDNHQ IURP WKH PDQXIDFWXUHUfV OLWHUDWXUH ,QGH[ YDOXHV DW DUELWUDU\ ZDYHOHQJWKV PD\ EH IRXQG E\ FXELF VSOLQH LQWHUSRODWLRQ

PAGE 179

:DYHOHQJWK Q3 Qf IWPf 7DEOH %O 'LVSHUVLRQ GDWD IRU /L7D

PAGE 180

5()(5(1&(6 >ƒKO@ >ƒKO@ >ƒKOD@ >ƒKOE@ >ƒKOF@ >ƒKO@ >$OI@ >$PL@ >$UD@ + ƒKOIHOGW :HEMRUQ DQG $UYLGVVRQ f3HULRGLF 'RPDLQ ,QYHUVLRQ DQG *HQHUDWLRQ RI %OXH /LJKW LQ /LWKLXP 7DQWDODWH :DYHJXLGHVf ,((( 3KRWRQ 7HFK /HWW YRO QR S + ƒKOIHOGW ) /DXUHOO DQG $UYLGVVRQ f6WURQJO\ 5HGXFHG 2SWLFDO 1RQOLQn HDULW\ LQ /LWKLXP 7DQWDODWH 'XH WR 3URWRQ ([FKDQJHf ,((( (OHFWURQ /HWW YRO QR S + ƒKOIHOGW f1RQOLQHDU 2SWLFDO 3URSHUWLHV RI 3URWRQ([FKDQJHG :DYHJXLGHV LQ =FXW /L7Df $SSO 3K\V YRO QR S + ƒKOIHOGW DQG :HEMRUQ f6LQJOH'RPDLQ /D\HUV )RUPHG LQ 0XOWLGRPDLQ /L7D E\ 3URWRQ ([FKDQJH DQG +HDW 7UHDWPHQWf $SSO 3K\V /HWW YRO QR S + ƒKOIHOGW :HEMRUQ ) /DXUHOO DQG $UYLGVVRQ f3RVWIDEULFDWLRQ &KDQJHV DQG 'HSHQGHQFH RQ +\GURJHQ &RQFHQWUDWLRQ RI WKH 5HIUDFWLYH ,QGH[ RI 3URWRQ([FKDQJHG /LWKLXP 7DQWDODWH :DYHJXLGHVf $SSO 3K\V YRO QR S + ƒKOIHOGW :HEMRUQ 3 $ 7KRPDV DQG 6 7HDW f6WUXFWXUDO DQG 2SWLFDO 3URSHUWLHV RI $QQHDOHG 3URWRQ([FKDQJHG :DYHJXLGHV LQ =FXW /L7Df $SSO 3K\V YRO QR S 5 & $OIHUQHVV 6 .RURWN\ DQG ( $ 0DUFDWLOL f9HORFLW\0DWFKLQJ 7HFKQLTXHV IRU ,QWHJUDWHG 2SWLF 7UDYHOLQJ :DYH 6ZLWFK0RGXODWRUVf ,((( 4XDQWXP (OHFWURQ YRO 4( QR S $PLQ -$ $XVW DQG 1$ 6DQIRUG f=3URSDJDWLQJ :DYHJXLGH /DVHUV LQ 5DUH(DUWK'RSHG 7L/L1Ef $SSO 3K\V /HWW YRO S 6 $UDKLUD :DWDQDEH 6KLQR]DNL DQG < 2JDZD f6XFFHVVLYH ([FLWHG 6WDWH $EVRUSWLRQ 7KURXJK D 0XOWLVWHS 3URFHVV LQ +LJKO\ (U'RSHG )LEHU 3XPSHG E\ D SP /DVHU 'LRGHf 2SW /HWW YRO QR S

PAGE 181

>%DL@ >%DX@ >%RU@ >%XL@ >%XU@ >&DF@ >&DR@ >&KH@ >&KX@ >'DY@ >'HO@ <6 %DL 55 1HXUJDRQNDU DQG 5 .DFKUX f+LJK(IILFLHQF\ 1RQYRODWLOH +RORJUDSKLF 6WRUDJH ZLWK 7ZR6WHS 5HFRUGLQJ LQ 3UDVHRG\PLXP'RSHG /LWKn LXP 1LREDWH E\ 8VH RI &RQWLQXRXV:DYH /DVHUVf 2SW /HWW YRO QR S %DXPDQQ 6 %RVVR 5 %ULQNPDQQ 5 &RUVLQL 0 'LQDQG $ *UHLQHU 6FKDIHU 6RFKWLQJ : 6RKOHU + 6XFKH DQG 5 :HVVHO f(U'RSHG ,QWHn JUDWHG 2SWLFDO 'HYLFHV ,Q /L1Ef ,((( 6HO 7RS 4XDQWXP (OHFWURQ YRO S 3 ) %RUGXL 5 1RUZRRG & %LUG DQG 7 &DUHOOD f6WRLFKLRPHWU\ ,VVXHV LQ 6LQJOH&U\VWDO /LWKLXP 7DQWDODWHf $SSO 3K\V YRO QR S : %XUQV 3 + .OHLQ ( :HVW DQG / ( 3OHZ f7L 'LIIXVLRQ LQ 7L/L1E 3ODQDU DQG &KDQQHO 2SWLFDO :DYHJXLGHVf $SSO 3K\V YRO QR S : %XUQV 0 0F:ULJKW +RZHUWRQ DQG 5 3 0RHOOHU f3HUIRUPDQFH DQG 0RGHOLQJ RI 3URWRQ ([FKDQJHG /L7D %UDQFKLQJ 0RGXODWRUVf /LJKWZDYH 7HFKQRO YRO QR S ) &DFFDYDOH ) 6HJDWR DQG 0DQVRXU f$ 1XPHULFDO 6WXG\ RI (UELXP'RSHG $FWLYH /L1E :DYHJXLGHV E\ WKH %HDP 3URSDJDWLRQ 0HWKRGf /LJKWZDYH 7HFKQRO YRO QR S ; ) &DR 5 9 5DPDVZDP\ DQG 5 6 6ULYDVWDYD f&KDUDFWHUL]DWLRQ RI $QQHDOHG 3URWRQ ([FKDQJHG /L1E :DYHJXLGHV IRU 1RQOLQHDU )UHTXHQF\ &RQYHUVLRQf /LJKWZDYH 7HFKQRO YRO QR S + & &KHQJ DQG 5 9 5DPDVZDP\ f1RYHO 7HFKQLTXH IRU WKH 0HDVXUHPHQW RI &RXSOLQJ /HQJWK LQ 'LUHFWLRQDO &RXSOHUVf 7HFKQLFDO 'LJHVW 63,( 2($HURn VSDFH 6HQVLQJ 2UODQGRf $SULO : & &KXDQJ : 6 :DQJ DQG + &KX f$ &RPSDULVRQ RI WKH 3HUIRUPDQFH RI /L1E DQG /L7D 7UDYHOLQJ:DYH 3KDVH 0RGXODWRUVf )LEHU ,QW 2SW YRO S 'DYLV DQG 1 $ /LQGRS f)DEULFDWLRQ DQG &KDUDFWHUL]DWLRQ RI 3\URSKRVSKR ULF $FLG 3URWRQ ([FKDQJHG /LWKLXP 7DQWDODWH :DYHJXLGHVf $SSO 3K\V YRO QR S ( 'HOHYDTXH 7 *HRUJHV 0 0RQHULH 3 /DPRXOHU DQG -) %D\RQ f0RGHOn LQJ RI 3DLU,QGXFHG 4XHQFKLQJ LQ (UELXP'RSHG 6LOLFDWH )LEHUVf ,((( 3KRn WRQ 7HFKQRO /HWW YRO QR S

PAGE 182

>'LH@ 9 'LHUROI $% .XWVHQNR ) 7DOOLDQ DQG : YRQ GHU 2VWHQ LQ (852',0 WK (XURSK\VLFDO &RQIHUHQFH RQ 'HIHFWV LQ ,QVXODWLQJ 0DWHULDOV $EVWUDFWV .HHOH 8QLYHUVLW\ .HHOH 6WDIIV 8. S -XO\ >'RX@ & 'RXNKDQ 3 &RUGLHU DQG 1 'RXNKDQ f/DWWLFH 'HIHFWV LQ /LWKLXP 7DQWD ODWHf UG 6\PSRVLXP RQ )UHTXHQF\ &RQWURO S >(NQ@ 2 (NQR\DQ : (+@ (O +DGL 3 %DOGL 6 1RXK 0 3 'H 0LFKHOL $/H\FXUDV 9$)HGRURY (+@ (O +DGL 9 5DVWRJL 0 5 6KHQR\ 7K\DJDUDMDQ 0 'H 0LFKHOL DQG % 2VWURZVN\ f6SHFWUDO 0HDVXUHPHQW RI WKH )LOP6XEVWUDWH ,QGH[ 'LIIHUHQFH LQ 3URWRQ([FKDQJHG /L1E :DYHJXLGHVf $SSO 2SW YRO QR S >)HG@ 9 $ )HGHURY DQG < 1 .RUNLVKNR f&U\VWDO 6WUXFWXUH DQG 2SWLFDO 3URSHUWLHV RI 3URWRQ([FKDQJHG /L7D :DYHJXLGHVf )HUURHOHFWULFV YRO S >)LQ@ 7 )LQGDNO\ 3 6XFKRVNL DQG ) /HRQEHUJHU f+LJK4XDOLW\ /L7D ,QWHJUDWHG 2SWLFDO :DYHJXLGHV DQG 'HYLFHV )DEULFDWHG E\ WKH $QQHDOHG3URWRQ([FKDQJH 7HFKQLTXHf 2SW /HWW YRO QR S >)XM@ 7 )XMLZDUD 5 6 6ULYDVWDYD ; &DR DQG 5 9 5DPDVZDP\ f&RPSDULVRQ RI 3KRWRUHIUDFWLYH ,QGH[ &KDQJH LQ 3URWRQ([FKDQJHG DQG 7L,QGLIIXVHG /L1E :DYHJXLGHVf 2SW /HWW YRO QR S >*HR@ *HRUJH 3UHSDUDWLRQ RI 7KLQ )LOPV 0DUFHO 'HNNHU ,QF 1HZ *LO@ 0 *LOO /0F&DXJKDQ DQG & :ULJKW f6SHFWURVFRSLF 6LWH 'HWHUPLQDn WLRQV LQ (UELXP'RSHG /LWKLXP 1LREDWHf 3K\V 5HY % YRO QR S >*OD@ $ 0 *ODVV ( 3HWHUVRQ DQG 7 1HJUDQ f2SWLFDO ,QGH[ 'DPDJH LQ (OHFn WURRSWLF &U\VWDOVf LQ /DVHU ,QGXFHG 'DPDJH LQ 2SWLFDO 0DWHULDOV $ 0 *ODVV DQG $ + *XHQWHU HGV :DVKLQJWRQ '& 1DW %XU 6WDQG f 6SHF 3XEO S >*RS@ 9 *RSDODQ DQG 0 *XSWD f2ULJLQ RI ,QWHUQDO )LHOG DQG 9L]XDOL]DWLRQ RI r 'RPDLQV LQ &RQJUXHQW /L7D &U\VWDOVf $SSO 3K\V YRO S

PAGE 183

>*XH@ + *XHQWHU :LWWPDQQ 50 0DFIDUODQH DQG 55 1HXUJDRQNDU f,QWHQVLW\ 'HSHQGHQFH DQG :KLWH/LJKW *DWLQJ RI 7ZR&RORU 3KRWRUHIUDFWLYH *UDWLQJV LQ /L1E 2SW /HWW YRO S >+DU@ 7 +DUYH\ $VWIDON $ < )HOGEOXP DQG % .DVVDKXQ f7KH 3KRWRUHIUDFn WLYH (IIHFW LQ 7LWDQLXP ,QGLIIXVHG /LWKLXP 1LREDWH 2SWLFDO 'LUHFWLRQDO &RXn SOHUV DW _-Pf ,((( 4XDQWXP (OHFWURQ YRO 4( QR S >+RO@ 5 +ROPHV DQG 0 6P\WK f7LWDQLXP 'LIIXVLRQ LQWR /L1E DV D )XQFn WLRQ RI 6WRLFKLRPHWU\f $SSO 3K\V YRO QR S >+RZ@ 3 + +RZHUWRQ : %XUQV DQG 5 3 0RHOOHU f,QWHJUDWHG2SWLFDO 0RGH &RQYHUWHU)UHTXHQF\ 6KLIWHU LQ 7L/L7Df ,*:2f :&& S >+XD@ &+ +XDQJ DQG / 0F&DXJKDQ fQP3XPSHG (U'RSHG /L1E :DYHJXLGH $PSOLILHUV $ FRPSDULVRQ ZLWK QP SXPSLQJf ,((( 6HO 7RS 4XDQWXP (OHFWURQ YRO S >+XN@ +XNULHGH 1HH .LS DQG ( .UDW]LJ f7KHUPDOO\ )L[HG 5HIOHFWLRQ *UDWn LQJV IRU ,QIUDUHG /LJKW LQ /L1E7L)H &KDQQHO :DYHJXLGHVf 2SW /HWW YRO QR S >+XV@ & 3 +XVVHOO $ 1RYHO 7XQDEOH )LOWHU IRU :DYHOHQJWK 'LYLVLRQ 0XOWLSOH[HG &RPPXQLFDWLRQ 6\VWHPV 3K' 'LVVHUWDWLRQ 8QLYHUVLW\ RI )ORULGD *DLQHVn YLOOH >,16@ f3URSHUWLHV RI /LWKLXP 1LREDWHf (0,6 'DWDUHYLHZV 6HULHV 1R ,163(& 3XE /RQGRQ >-DF@ / -DFNHO & ( 5LFH DQG 9HVVHOND f3URWRQ([FKDQJH IRU +LJK ,QGH[ :DYHJXLGHV LQ /L1Ef $SSO 3K\V /HWW YRO S >-HU@ ) -HUPDQQ 0 6LPRQ DQG ( .UDW]LJ f3KRWRUHIUDFWLYH 3URSHUWLHV RI &RQJUXn HQW DQG 6WRLFKLRPHWULF /LWKLXP 1LREDWH DW +LJK ,QWHQVLWLHVf 2SW 6RF $P % YRO S >-XQ@ +6 -XQJ + ) 7D\ORU DQG 2 (NQR\DQ f,QWHUIHURPHWULF 3RODUL]DWLRQ,QGHn SHQGHQW 0RGXODWRU LQ /L7Df /LJKWZDYH 7HFKQRL YRO QR S >.DP@ 3 .DPLQRZ DQG 5 &DUUXWKHUV f2SWLFDO :DYHJXLGLQJ /D\HUV LQ /L1E DQG /L7Df $SSO 3K\V /HWW YRO QR S >.DQ@ .DQ DQG /
PAGE 184

>.LV@ 0 .LVVD 3 6XFKRVNL DQG /HZLV f$FFHOHUDWHG $JLQJ RI $QQHDOHG 3URWRQ([FKDQJHG :DYHJXLGHVf /LJKWZDYH 7HFKQRO YRO QR S >.RU@ .RU@ .RV@ 60 .RVWULWVNLL DQG 20 .ROHVQLNRY f3KRWRLQGXFHG /LJKW 6FDWWHULQJ LQ WKH &RSSHU'RSHG /L MB[ +[1E 3KRWRUHIUDFWLYH :DYHJXLGHVf 2SW 6RF $P % YRO S >.RVD@ 60 .RVWULWVNLL DQG 2* 6HYRVW\DQRY f,QIOXHQFH RI ,QWULQVLF 'HIHFWV RQ /LJKW,QGXFHG &KDQJHV LQ 5HIUDFWLYH ,QGH[ RI /LWKLXP 1LREDWH &U\VWDOVf $SSO 3K\V % YRO S >.RVE@ 60 .RVWULWVNLL .LS DQG ( .UDW]LJ f,PSURYHPHQW RI 3KRWRUHIUDFWLYH 3URSHUWLHV RI 3URWRQ([FKDQJHG /L7D :DYHJXLGHVf $SSO 3K\V % YRO S >.RV@ 6 0 .RVWULWVNLL % 0DULQJ 5 ) 7DYO\NDHY DQG 5 9 5DPDVZDP\ (QKDQFHPHQW RI WKH 3KRWRUHIUDFWLYH (IIHFW E\ (U 'RSLQJ LQ /L7Df 7236 7UHQGV LQ 2SWLFV DQG 3KRWRQLFV 6HULHVf Y SS 6HYHQWK 7RSLn FDO 0HHWLQJ RQ 3KRWRUHIUDFWLYH 0DWHULDOV (IIHFWV DQG 'HYLFHV -XQH (OVLQRUH 'HQPDUN >.XE@ ..XERWD 1RGD DQG 2 0LNDPL f7UDYHOLQJ :DYH 2SWLFDO 0RGXODWRU 8VLQJ D 'LUHFWLRQDO &RXSOHU /L1E :DYHJXLGHf ,((( 4XDQWXP (OHFWURQ YRO 4( QR S >/DQ@ /DQGH 66 2UORY DQG / +HVVHOLQN f7ZR3KRWRQ $SRGL]DWLRQ LQ /LWKLXP 1LREDWHf 2SW /HWW YRO S >/HY@ + /HYLQVWHLQ $ $ %DOOPDQ DQG & &DSLR f'RPDLQ 6WUXFWXUH DQG &XULH 7HPSHUDWXUH RI 6LQJOH&U\VWDO /LWKLXP 7DQWDODWHf $SSO 3K\V YRO S >0DF@ + $ 0DFOHRG 7KLQ)LOP 2SWLFDO )LOWHUV 0DFPLOODQ 1HZ 0DUD@ % 0DULQJ 5 ) 7DYO\NDHY DQG 5 9 5DPDVZDP\ f'LUHFW 2EVHUYDWLRQ RI %XULHG 5HIUDFWLYH,QGH[ 3URILOHV LQ $QQHDOHG 3URWRQ([FKDQJHG /L7D

PAGE 185

:DYHJXLGHVf ,QWHJUDWHG 3KRWRQLF 5HVHDUFK YRO 26$ 7HFKQLFDO 'LJHVW 6HULHV ,:% S >0DUE@ % 0DULQJ 5 ) 7DYO\NDHY DQG 5 9 5DPDVZDP\ f$QRPDOLHV RI 5HIUDFn WLYH,QGH[ 3URILOHV 2EVHUYHG 'LUHFWO\ LQ $QQHDOHG 3URWRQ([FKDQJHG ;FXW /L7D :DYHJXLGHVf (OHFWURQLFV /HWWHUV YRO QR S >0DUD@ % 0DULQJ 0DUE@ % 0DULQJ 5 ) 7DYO\NDHY 5 9 5DPDVZDP\ 0DU@ % 0DULQJ 6 0 .RVWULWVNLL 5 ) 7DYO\NDHY DQG 5 9 5DPDVZDP\ +LJK3RZHU 6LQJOH0RGH /L7D :DYHJXLGHV ZLWK ,PSURYHG 7HPSRUDO 6WDn ELOLW\f ,QWHJUDWHG 3KRWRQLFV 5HVHDUFK 26$ 7HFKQLFDO 'LJHVW 57X.O S >0DU22@ % 0DULQJ 5 ) 7DYO\NDHY 5 9 5DPDVZDP\ DQG 6 0 .RVWULWVNLL f:DYHJXLGH 6WDELOLW\ LQ /L7Df WR EH VXEPLWWHG WR $SSO 3K\V >0DWD@ 3 0DWWKHZV DQG $ 5 0LFNHOVRQ f,QVWDELOLWLHV LQ $QQHDOHG 3URWRQ ([FKDQJH :DYHJXLGHV LQ /LWKLXP 7DQWDODWHf $SSO 3K\V YRO QR S >0DWE@ 3 0DWWKHZV DQG $ 5 0LFNHOVRQ f3URSHUWLHV RI 3URWRQ ([FKDQJH :DYHJXLGHV LQ /LWKLXP 7DQWDODWHf $SSO 3K\V YRO QR S >0F:@ 0 0F:ULJKW +RZHUWRQ : %XUQV 3 5 6NHDWK DQG $ 6 *UHHQEODWW f'HSHQGHQFH RI 5HIUDFWLYH ,QGH[ RQ +\GURJHQ &RQFHQWUDWLRQ LQ 3URWRQ ([FKDQJHG /L1Ef ,((( 4XDQWXP (OHFWURQ YRO QR S >0F:@ 0 0F:ULJKW +RZHUWRQ DQG : %XUQV f3KRWRUHIUDFWLYH (IIHFWV LQ 3URWRQ ([FKDQJHG /L7D 2SWLFDO :DYHJXLGHVf /LJKWZDYH 7HFKQRO YRO QR S >0L]@ 0L]XXFKL 0L]@ 0L]XXFKL DQG
PAGE 186

>0L]@ 0L]XXFKL DQG 0X@ ; 0X ; 1DN@ 1DNDPXUD DQG + 6KLPL]X f)HUURHOHFWULF ,QYHUVLRQ /D\HUV )RUPHG E\ +HDW 7UHDWPHQW RI 3URWRQ([FKDQJHG /L7Df $SSO 3K\V /HWW YRO QR S >1D]@ 0 1D]DUDWK\ : 'ROIL DQG 5 -XQHUPDQ f6SUHDG 6SHFWUXP )UHTXHQF\ 5HVSRQVH RI &RGHG 3KDVH 5HYHUVDO 7UDYHOLQJ :DYH 0RGXODWRUVf /LJKWZDYH 7HFKQRO YRO /7 QR S >1RX@ 6 1RXK 3 %DOGL 0 'H 0LFKHOL 0RQQRP % 2VWURZVN\ ( /DOOLHU DQG 0 3DSXFKRQ f/DVHU $FWLRQ LQ 1G/L7D 3URWRQ ([FKDQJH :DYHJXLGH ([KLELWLQJ 3URWRQ ,QGXFHG ([FLWHG 6WDWH /LIHWLPH 5HGXFWLRQf (OHFWURQ /HWW YRO S >1RX@ 6 1RXK 3 %DOGL (O +DGL 0 'H 0LFKHOL 0RQQRP % 2VWURZVN\ ( /DOOLHU DQG 0 3DSXFKRQ f)DEULFDWLRQ 3DUDPHWHU 2SWLPL]DWLRQ RI D /RZ 7KUHVKROG +LJK(IILFLHQF\ 3URWRQ([FKDQJHG :DYHJXLGH /DVHU LQ 1G/L7Df 2SW /HWW YRO QR S >3D]@ 5 3D]3XMDOW ' 7XVFKHO %UDXQVWHLQ 7 %ODQWRQ 6 7 /HH DQG / 0 6DOWHU f&KDUDFWHUL]DWLRQ RI 3URWRQ ([FKDQJH /LWKLXP 1LREDWH :DYHJXLGHVf $SSO 3K\V YRO QR S >3HG@ ) / 3HGURWWL DQG / 6 3HGURWWL ,QWURGXFWLRQ WR 2SWLFV 3UHQWLFH+DOO /RQGRQ >5DP@ 5 9 5DPDVZDP\ DQG 5 6ULYDVWDYD f,RQ([FKDQJHG *ODVV :DYHJXLGHV $ 5HYLHZf /LJKWZDYH 7HFKQRO YRO QR S >5HG@ %5 5HGG\ DQG 6. 1DVK6WHYHQVRQ f5HGWR9LROHW DQG 1HDU ,QIUDUHGWR *UHHQ (QHUJ\ 8SFRQYHUVLRQ LQ /D)(Uf $SSO 3K\V YRO QR S >5HJ@ 5 5HJHQHU DQG : 6RKOHU f/RVV LQ /RZ)LQHVVH 7L/L1E 2SWLFDO :DYHJXLGH 5HVRQDWRUVf $SSO 3K\V % YRO S >6DQ@ 1 $ 6DQIRUG $ $XVW 0DORQH DQG 5 /DUVRQ f1G/L7D :DYHJXLGH /DVHUf 2SW /HWW YRO QR S

PAGE 187

>6DZ@ 6DZDNL DQG 6 .XULPXUD f6HFRQG+DUPRQLF *HQHUDWLRQ LQ 3HULRGLFDOO\ 'RPDLQ,QYHUWHG /LWKLXP 7DQWDODWH &KDQQHO :DYHJXLGHVf &/(2 &7X9 S >6FK@ 5 9 6FKPLGW DQG 3 .DPLQRZ f$FRXVWRRSWLF %UDJJ 'HIOHFWLRQ LQ /L1E 7L'LIIXVHG :DYHJXLGHVf ,((( 4XDQWXP (OHFWURQ YRO 4( QR S >6FK@ 2 ) 6FKLUPHU 2 7KLHPDQQ DQG 0 :RKOHFNH f'HIHFWV LQ /L1E, ([SHULn PHQWDO $VSHFWVf 3K\V &KHP 6ROLGV YRO QR S >6SL@ : % 6SLOOPDQ -U 1 $ 6DQIRUG DQG 5 $ 6RUHI f2SWLFDO :DYHJXLGHV LQ /L7D )RUPHG E\ 3URWRQ ([FKDQJHf 2SW /HWW YRO QR S >6WH@ 6WHIIHQ $ 1H\HU ( 9RJHV DQG 1 +HFNLQJ f5HIUDFWLYH ,QGH[ 3URILOH 0HDn VXUHPHQW 7HFKQLTXHV E\ 5HIOHFWLYLW\ 3URILOLQJ 9LGLFRQ ,PDJLQJ %HDP 6FDQn QLQJ DQG 6DPSOH 6FDQQLQJf $SSO 2SW YRO QR S >6X]@ 7 6X]XNL DQG 2 (NQR\DQ f&KDUDFWHUL]DWLRQ RI WKH 3ROLQJ 3URFHVV LQ =LQF 'LIIXVHG /LWKLXP 7DQWDODWH 2SWLFDO :DYHJXLGHVf )HUURHOHFWULFV YRO S >6\P@ 5 6\PV DQG &R]HQV 2SWLFDO *XLGHG :DYHV DQG 'HYLFHV 0F*UDZ+LOO /RQGRQ >7DQ@ / 7DQJRQDQ 0 %DUQRVNL ) /RWVSHLFK DQG $ /HH f+LJK 2SWLFDO 3RZHU &DSDELOLWLHV RI 7L,QGLIIXVHG /L7D :DYHJXLGH 0RGXODWRU 6WUXFWXUHVf $SSO 3K\V /HWW YRO QR S >7DQ@ / 7DQJRQDQ / 3HUVHFKLQL ) /RWVSHLFK DQG 0 %DUQRVNL f(OHFn WURRSWLF 'LIIUDFWLRQ 0RGXODWLRQ LQ 7L,QGLIIXVHG /L7Df $SSO 2SW YRO QR S >7DY@ 5 7DYO\NDHY .LLFNHOKDXV DQG ( 9RJHV f,QGH[ 3URILOH 5HFRQVWUXFWLRQ RI 7L/L1E 6WUXFWXUHV DQG %HQGLQJ /RVV (YDOXDWLRQ IURP 1HDU)LHOG 0HDVXUHn PHQWVf 2SW &RPP YRO S >7DY@ 5 ) 7DYO\NDHY % 0DULQJ DQG 5 9 5DPDVZDP\ f5HIUDFWLYH,QGH[ 3URn ILOH RI $QQHDOHG 3URWRQ([FKDQJHG /L7D &KDQQHO :DYHJXLGHV IURP :KLWH /LJKW 6RXUFH 0HDVXUHPHQWVf ,QWHJUDWHG 3KRWRQLF 5HVHDUFK YRO 26$ 7HFKQLFDO 'LJHVW 6HULHV ,7K&O S >8HK@ 6 8HKDUD f&DOLEUDWLRQ RI 2SWLFDO 0RGXODWRU )UHTXHQF\ 5HVSRQVH ZLWK $SSOLn FDWLRQ WR 6LJQDO /HYHO &RQWUROf $SSO 2SW YRO QR S

PAGE 188

>8OU@ >:DQ@ >:RR@ >;X@ > <@ >=DY@

PAGE 189

%,2*5$3+,&$/ 6.(7&+ 'DYLG %OD\QH 0DULQJ ZDV ERUQ LQ 5RFKHVWHU 0LQQHVRWD LQ +H UHFHLYHG D %6 GHJUHH LQ HOHFWULFDO HQJLQHHULQJ IURP WKH 8QLYHUVLW\ RI )ORULGD LQ 'HFHPEHU +H KDV EHHQ D UHVHDUFK DVVLVWDQW DW WKH 3KRWRQLFV 5HVHDUFK /DERUDWRU\ XQGHU WKH VXSHUYLn VLRQ RI 3URIHVVRU 5DPX 9 5DPDVZDP\ VLQFH -DQXDU\ DIWHU VWDUWLQJ KLV XQGHUJUDGXn DWH VHQLRU GHVLJQ SURMHFW ZLWK 'U 5DPDVZDP\ GXULQJ WKH IDOO RI LQ WKH DUHD RI VLQJOHn PRGH ZDYHJXLGH IDEULFDWLRQ LQ OLWKLXP WDQWDODWH +H VSHQW WKH VXPPHU RI SHUIRUPn LQJ FKDUDFWHUL]DWLRQ RQ ZDYHJXLGHV IDEULFDWHG LQ OLWKLXP WDQWDODWH E\ WKH DQQHDOHG SURWRQ H[FKDQJH WHFKQLTXH DW (JOLQ $LU )RUFH %DVH LQ )RUW :DOWRQ %HDFK )ORULGD +H UHFHLYHG KLV 0DVWHU RI (QJLQHHULQJ GHJUHH LQ HOHFWULFDO HQJLQHHULQJ IURP WKH 8QLYHUVLW\ RI )ORULGD LQ $XJXVW RI 7KH WLWOH RI WKH WKHVLV ZDV f&KDUDFWHUL]DWLRQ DQG 'LUHFW ,QGH[ 3URILOLQJ RI $QQHDOHG 3URWRQ ([FKDQJHG :DYHJXLGHV LQ /L7Df 6LQFH WKHQ KH KDV EHHQ ZRUNLQJ RQ KLV 3K' UHVHDUFK DW WKH 8QLYHUVLW\ RI )ORULGD LQ WKH DUHD RI LQWHn JUDWHGRSWLFDO GHYLFHV LQ OLWKLXP WDQWDODWH +LV FXUUHQW UHVHDUFK LQWHUHVWV LQFOXGH LQWHn JUDWHGRSWLFDO FRPSRQHQWV IRU RSWLFDO FRPPXQLFDWLRQ QHWZRUNV


90
Fig. 4.10 Comparison of a directly
observed index profile (solid)
and an IWKB reconstructed
profile (dotted).
there still remains the problem as to
which profile distribution to apply to
an IWKB approximation to determine
surface index values, i.e., an
exponential function, a
complementary error function, a
Gaussian function, or a Fermi
function. Since many of these
distributions have been reported in the
literature for various fabrication
conditions in APE:LiTa03, application of an unsuitable function to a set of measured
mode indices can result in significant error in calculating the surface value of the
waveguide index [Mu94], Such an example is depicted in Fig. 4.10, where the measured
effective mode indices and associated WKB reconstruction (dotted line) are compared
with the directly measured index profile (solid line). The measured mode indices are
within measurement error for the prism coupling technique, however the IWKB
reconstruction method could not accurately predict the plateau (almost) region near the
substrate surface and hence seriously overestimated the surface index increment.
Finally, the IWKB method requires at least three guided modes and, hence, is not
applicable at all to the case of single-mode waveguides. Clearly, extrapolation of the
multi-mode waveguide data, whose accuracy is rather questionable, to single-mode
structures is even more doubtful. On the other hand, the single-mode regime of operation


151
Fig. 7.4 Schematic of the setup for measuring the DC response of an elec
tro-optic modulator.
applied voltage (V)
Fig. 7.5 Measured DC response (squares) and theoretical fit
(solid line) of the MZI modulator.


CHAPTER 3
CHARACTERISTICS OF Er-DOPED LiTa03
Because of the demand for integrated optical components for use in communica
tion networks, as outlined in Chapter 1, there is a decisive need for lasers and amplifiers to
operate in these systems. While much work has been done on the development of fiber-
amplifiers, particularly Er-doped fiber amplifiers (EDFAs) [Ara92, Del93], efforts con
tinue for the development of rare-earth-doped lasers and amplifiers in LiNb03 and LiTa03
because of the advantage they posses for allowing the integration of a laser and external
device, such as a modulator, on the same substrate. Because these materials, especially
LiNb03, are currently used for the fabrication of network components like filters and
modulators and because selective doping may be used on these substrate to define an
active gain region on only a portion of the substrate, the opportunity to bypass coupling
between individual devices arises by allowing the integration of a laser and another device
on the same module.
Though much work has been done recently on the production of rare-earth-doped
waveguide amplifiers and lasers in LiNb03 [Ami96, Bau96, Hua96], an inherent problem
with guided-wave devices in LiNb03 is their low resistance to photorefractive damage at
visible and near-infrared wavelengths. Having a much higher photorefractive damage
threshold, LiTa03 emerges as a promising candidate for the development of waveguide
49


175
Waveguides, Integrated Photonic Research, vol. 6, 1996 OSA Technical
Digest Series, IWB3-1, p. 435.
[Mar96b] D. B. Maring, R. F. Tavlykaev, and R. V. Ramaswamy, Anomalies of Refrac
tive-Index Profiles Observed Directly in Annealed Proton-Exchanged, X-cut
LiTa03 Waveguides, Electronics Letters, vol. 32, no. 16, p. 1473, 1996.
[Mar98a] D. B. Maring, Yu. N. Korkishko, R. F. Tavlykaev, R. V. Ramaswamy, and J.
M. Zavada, "Evolution of Crystal Phases and Refractive Index Profiles in X-
cut Annealed Proton-Exchanged LiTa03, Integrated Photonics Research,
OSA Technical Digest, ITuI2-l, p.301, 1998.
[Mar98b] D. B. Maring, R. F. Tavlykaev, R. V. Ramaswamy, Yu. N. Korkishko, V. A.
Fedorov, and J. M. Zavada, "Effect of Crystal Phases on Refractive Index Pro
files of Annealed Proton-Exchanged Waveguides in X-cut LiTa03, Appl.
Phys. Lett., vol. 73, p.423, 1998.
[Mar99] D. B. Maring, S. M. Kostritskii, R. F. Tavlykaev, and R. V. Ramaswamy,
"High-Power Single-Mode LiTa03 Waveguides with Improved Temporal Sta
bility, Integrated Photonics Research, OSA Technical Digest, RTuK4-l,
p.281, 1999.
[MarOO] D. B. Maring, R. F. Tavlykaev, R. V. Ramaswamy, and S. M. Kostritskii,
Waveguide Stability in LiTa03, to be submitted to J. Appl. Phys.
[Mat92a] P. J. Matthews and A. R. Mickelson, Instabilities in Annealed Proton
Exchange Waveguides in Lithium Tantalate, J. Appl. Phys., vol. 71, no. 11, p.
5310, 1992.
[Mat92b] P. J. Matthews and A. R. Mickelson, Properties of Proton Exchange
Waveguides in Lithium Tantalate, J. Appl. Phys., vol. 72, no. 7, p. 2562, 1992.
[McW91] M. McWright Howerton, W. K. Burns, P. R. Skeath, and A. S. Greenblatt,
Dependence of Refractive Index on Hydrogen Concentration in Proton
Exchanged LiNb03, IEEE J. Quantum Electron., vol. 27, no. 3, p. 593, 1991.
[McW92] M. McWright Howerton and W. K. Burns, Photorefractive Effects in Proton
Exchanged LiTa03 Optical Waveguides, J. Lightwave Technol., vol. 10, no. 2,
p. 142, 1992.
[Miz91] K. Mizuuchi, K. Yamamoto, and T. Taniuchi, Blue-Light Generation by
Quasi-Phase-Matched Second-Harmonic Generation in LiTa03, CLEO,
CTuV3, p. 164, 1991.
[Miz92] K. Mizuuchi and K. Yamamoto, Characteristics of Periodically Domain-
Inverted LiTa03, J. Appl. Phys., vol. 72, no. 11, p. 5061, 1992.


128
agrees with previously reported data
[Mat92a], As was discovered earlier by
monitoring changes in rocking curves
(Section 6.1), waveguides with high
proton concentration, near the upper
limits of the phase diagram, are highly
unstable and therefore unsuitable for
Fig. 6.4 Measured coupling length of
Sample 1. This couplei con- reliable device fabrication.
sisted of 4pm channels with a
5pm gap. jt js worth noting that, at least
for the case of high concentration sam
ples which are in phases of decreasing slope, that over time, both the surface strain and the
index increment were found to decrease. This is in contradiction to the evolution of the
index profile upon annealing, which was found to follow the structural phase diagram and
show an increase in index as the sample was annealed, decreasing strain, through the high
concentration phases. The reason for this peculiarity is not clear, but it would seem to sug
gest that the processes of aging and annealing do not have the same effect on the
waveguide or the crystal.
The measured coupling length of Sample 2 over a period of 219 days is shown in
Fig. 6.5. This sample was in the k phase of the Z-cut diagram. As can be seen from Table
6.2, the change in index increment for this sample is much smaller than that of Sample 1.
Note that because of its longer coupling length, the measured change in coupling length of
Sample 2 is larger, but its corresponding change in index is still smaller. This confirms the
previously held notions that stability can be improved by avoiding the presence of


20
nc TM
> TE
Fig. 2.4 Increase of ne only due to proton exchange causes the
propagation of solely TM modes in a z-cut crystal.
2.1.2 Features of the APE Process in LiTaQ3
As stated earlier, the APE process of waveguide fabrication holds many advantages
over other methods, such as simplicity and high index increment. Specifically, in compar
ison with the method of Ti or Zn- indiffusion, the APE process supports only one polariza
tion and can be performed at much lower temperatures, below the Curie point where no
crystal repoling is necessary. Additionally, the waveguide loss is low [Mat92a], and it is
less susceptible to photorefractive damage [Mat92a, Mat92b]. We describe here many of
the characteristics and peculiarities of the APE process in LiTa03, making comparisons to
the well characterized APE process in LiNb03 whenever applicable.
The proton exchange process is known to cause an increase in the extraordinary
index ne, and a decrease in the ordinary index n0 [hl94c, hl95]. This is advantageous


167
Wavelength
(M-m)
n
0.4358
2.853
0.4916
2.723
0.4960
2.715
0.5461
2.652
0.5770
2.623
0.5791
2.621
0.6907
2.5555
0.7082
2.548
1.0140
2.483
1.5296
2.451
2.0000
2.399
2.5000
2.387
3.0000
2.380
3.5000
2.367
4.0000
2.350
4.5000
2.322
5.0000
2.290
5.5000
2.200
Table A. 1 Dispersion data for rutile.


3.5 Li Treatment
In an effort to decrease the amount of Er3+ clusters, and hence reduce upconver-
sion, a Li-treatment process was employed to increase the Li/Ta ratio. The results of this
Li-treatment are discussed in this section.
Some of the fabricated Er:LiTa03 samples were processed in molten LiN03. Sub
sequent fluorescence measurements showed a gradual decrease of IASF with an increase in
the process time up to 12-16 hours. At the same time, the intensity of SF remained nearly
constant, i.e. the total Er3+ content in the crystal was not affected by this processing. Fur
ther increase of the processing time induced only a small change in IASF. The most
impressive result was a nearly 54% reduction in the intensity of the green ASF emission as
a result of the post-indiffusion processing, as illustrated in Fig. 3.10. This reduction is
3
W
tin
oo
<
a
O
G
O
C/3
a
o
3
03
| l l l l | l l l l | l l l l | l l l l | l l l l | l i i i | i i l l | l l l l
- - -Er:LiTa03, ?i=647nm, J=0.7xl03 W/cm2
EnLiTaOj, ?i=632.8nm, J=2xl03 W/cm2
I Li:Er:LiTaOi, X=632.8nm, J=2xl03 W/cm2
525 535 545 555
wavelength (nm)
565
Fig. 3.10 Fluorescence spectra of the green ASF emission from the Er-indif-
fused sample with the highest Er3+ concentration. The dotted line
corresponds to a pump wavelength of X=647nm and input intensity
J=0.7xl03 W/cm2. The solid and dashed lines correspond to
7=632.8nm and J=2xl03 W/cm2, before and after Li treatment,
respectively.


Wavelength
nP
n
(ftm)
3.0
2.080
2.076
2.8
2.087
2.083
2.6
2.094
2.090
2.4
2.099
2.095
2.2
2.105
2.101
2.0
2.112
2.107
1.8
2.117
2.112
1.6
2.121
2.117
1.4
2.127
2.124
1.2
2.134
2.131
1.0
2.143
2.139
0.9
2.149
2.145
0.8
2.158
2.154
0.7
2.170
2.165
0.6
2.188
2.183
0.5
2.221
2.216
Table B.l Dispersion data for LiTa03.


87
infrared, however, although photorefractive damage is often considered negligible, dam
age has been observed at 1.3|im in Ti:LiNb03 for power levels around 5mW [Har86], The
photorefractive resistance of these waveguides, as well as those fabricated by PE in both
LiNb03 and LiTa03, at 1.55|im have not been investigated to date.
In this section, the power handling capability of PE waveguides in LiTa03 operat
ing near 1.5fim is examined and compared to similar structures in LiNb03. A description
of the measurement setup is given first, followed by the measurement results.
4.5.1 Setup Description
Straight-channel waveguides of width 6|im were fabricated by PE in X-cut LiTa03
using glycerin at 260C for 24 hours and by APE in X-cut LiNb03 using pyrophosphoric
acid at 200C for 50 minutes followed by annealing at 350C for 7 hours with a 2 hour ramp
period at each end. To provide high input intensities needed for the measurement, a Fiber-
Raman Laser at 1.48|im was used. Light from the laser was coupled into the waveguides
using a single-mode fiber. Waveguide outputs were collected with a 20x lens and focused
onto a thermal detector with calibrated power meter. A pinhole was placed in front of the
detector and closed around the mode spot at low intensities so as to not inadvertently col
lect the light from the expanded mode of a waveguide exhibiting photorefractive damage
at higher intensities. The output power versus input power was then plotted to check for
photorefractive damage.
4.5.2 Results
Fig 4.9 shows the results of photorefractive measurements performed on
waveguides of similar channel width in LiTa03 and LiNb03. As can be seen, the LiTa03
sample, shows a reasonable degree of linearity between output and input power over the


106
of the structural phase diagram for
APE:LiTa03 is indispensable in developing
advanced high-throughput electro-optic and
nonlinear guided-wave devices with low
insertion loss and temporally stable perfor
mance.
In this chapter, the structural phase
diagram for APE:LiTa03 is reconstructed.
Some limitations that have hampered previ
ous efforts to construct the diagram are first discussed. Then, direct measurement tech
niques, outlined in the previous chapter, are employed for the successful reconstruction of
the structural phase diagram. The diagram is then used to explain the previously mentioned
anomalies associated with the APE process in LiTa03. Finally, PE waveguides are ana
lyzed for inclusion in the structural phase diagram. Direct measurement techniques and
Raman spectra of these samples are used to identify a new, single-phase region of the struc
tural phase diagram, not seen before with APE structures. Note that this chapter is only
concerned with APE and PE waveguides as the other two methods of waveguide fabrication
used for this work, namely, Ti-indiffusion and Zn vapor-indiffusion, are not known to
exhibit such anomalies and do not induce any lattice strain, therefore have no structural
Fig. 5.1 Example of a structural phase
diagram, relating surface
index increment An to either
H+ proton concentration or
proton-induced lattice strain
e" at the surface.
phase diagram.


93
/ =
r \n+\
(4-8)
where n is the substrate local refractive index, either ordinary or extraordinary. By taking
the derivative of this equation with respect to n, the change in reflected intensity can be
related to the incident intensity according to
dIr
dn
_r = 4 (n-1) J
(n + 1)
3 o
(4-9)
Now, since what was actually measured was the reflected intensity Ir and the change in
reflected intensity dlr by substituting Eq. (4-8) into Eq. (4-9), one obtains
dl
dn
n
( \
4
1
n
(4-10)
n
Rearranging this equation and dropping subscripts, a relationship between the change in
index and the corresponding change I in reflected intensity I is derived
An 1
" 4{nb nj I
n
V}A[
b'
(4-11)
where An is the index change and is the bulk index value, either ordinary or extraordi
nary.
Exact bulk index values of n0 and ne at the wavelength of interest, 632.8 nm, were
found by cubic spline interpolation of the dispersion data for LiTa03 presented in Appen
dix B. These values were n0 = 2.176 and ne = 2.181. An X-cut crystal had to be used in
order to be able to see the extraordinary index ne, characterized by the index along the Z-
axis. A polarizer and half-wave plate were used at the laser output to select the desired


100
An(0.633pm)/An(1.32|lm) is about 1.9, i.e. the index increment almost doubles when the
wavelength is halved.
As stated earlier, the direct profiling technique was used to extract surface index
increment values of planar waveguides in order to reconstruct the structural phase diagram
for APE:LiTa03. The other information needed for this reconstruction is surface values of
proton-induced lattice strain. These values were obtained through rocking curve analysis,
which is the subject of the next section.
4.7 Rocking Curve Analysis
In order to obtain surface values of proton-induced lattice strain e"3 needed to
reconstruct the structural phase diagram for APE:LiTa03, rocking measurements must be
performed. In this section, a description of the rocking curve measurement is given based
on the theory of Braggs Law and utilizing an X-ray diffractometer. Given a rocking curve,
the calculation required to determine the surface value of e"3 is given. Finally, an actual
rocking curve is presented. Note that this procedure is only performed on proton-
exchanged waveguides as both Ti-indiffusion and Zn vapor-indiffusion are not known to
produce any measurable lattice strain.
4.7.1 The Rocking Curve Measurement
The lattice spacing d of a crystal is measured using an X-ray diffractometer. The
schematic for such a measurement is illustrated in Fig. 4.17. Radiation of wavelength ~k
from an X-ray source is incident upon the surface of the sample. A detector is used to
measure the intensity of the reflected radiation. This reflected intensity is maximum when


137
1.55|im, because of the inherently low values of index increment An. However, metal-
indiffused waveguides should then be stable, as they are in LiNb03. This situation is
examined in the next section, where the stability of Zn vapor-indiffused and Ti-indiffused
waveguides is measured.
6.4 Zn/Ti-Indiffused Stability Results
If stability depends on waveguide fabrication conditions, then metal-indiffused
waveguides in LiTa03 should be stable, as they are in LiNb03. These waveguides pro
duce no strain, as does PE. Additionally, their ions are larger in size than the H+ ions
introduced by proton exchange, so any instabilities due to ion mobility within the crystal
should also be improved. Metal-indiffused waveguides also have the added advantages
that their electro-optic coefficient r33 is completely recoverable (by repoling), unlike many
PE/APE waveguides, and they dont contribute to quenching of active Er-doped gain
regions.
In this section, the stability of Zn vapor-indiffused and Ti-indiffused waveguides is
examined to determine if stability in LiTa03 depends on fabrication conditions, similar to
LiNb03. The directional coupler is used again to measure the amount of change in index
increment for Zn vapor-indiffused waveguides. Then, the modal characterization tech
nique of Section 4.1 is used to track and compare changes in the region of single-mode
operation for waveguides fabricated by PE, Ti-indiffusion, and Zn vapor-indiffusion.


136
shown in the structural phase diagrams. As a result, PE must be used, because of the
resulting profile shape, to maximize confinement at such low An values.
Second, to significantly decrease strain values using PE, very dilute sources would
be required. However, this means that exchange times would run hundreds of hours in
order to fabricate waveguides deep enough to support a mode at 1.55|im, because of the
low An and diluted source. Such a situation is not practically viable. Additionally,
increasing the exchange temperature in order to reduce exchange time would lead to expo
nential index profiles [E1H98], which again would have poor confinement and large bend
ing losses.
Third, because of the low An values, waveguides would be required to be quite
deep (~10|lm). At this depth, there is a decreased overlap factor between the optical and
electrical fields for electro-optic devices. In addition, the large mode would exhibit high
bending losses.
Finally, while there appears to be some trade-off between improving stability and
adjusting fabrication parameters to reduce strain at the expense of decreasing An and
device performance, there still lies the possibility that acceptable levels of stability cannot
be achieved, at least by the PE process. For the case of LiNb03, stability is known to be
dependent upon fabrication parameters. For example, as long as proton concentration lev
els are low enough, it is possible to fabricate waveguides in LiNb03 by APE which exhibit
no measurable temporal instability [Mat92a], Likewise, no data has been reported on the
instability of Ti-indiffused waveguides in LiNb03.
For the case of LiTa03, if stability depends on fabrication conditions, as it does in
LiNb03, then the production of stable PE and APE waveguides is hindered, especially at


10
To begin, an extensive study on the conditions of Er-doping needs to be performed.
Fabrication conditions such as Er thickness before indiffusion and temperature and time
of indiffusion need to be determined. Er3+ concentration should be at an appropriate level
so as to provide significant gain and Er3+ concentration profiles should be at least as deep
as waveguide index profiles so that the optical mode has good overlap with the gain
medium. The effect of Er-doping on the physical and optical properties of the crystal need
also be examined; in particular: how doping effects the photorefractive property of the
crystal as well as how much upconverted emission from Er3+ clusters, which reduce gain
[G196], is present and how to reduce it.
Next, a complete characterization of the APE process of waveguide formation is
desired to facilitate the understanding of the associated anomalies and identify fabrication
conditions leading to temporally stable waveguides. The aforementioned anomalies have
pointed to a nonlinear dependence of refractive index on proton concentration. As such,
the structural phase diagram for APE:LiTa03, relating waveguide surface index increment
to proton-induced lattice strain at the surface, needs to be constructed in order to explain
these anomalies. Because instability is speculated to be due to the presence of multiple
phases within the proton exchanged region, the compilation of this diagram would also aid
in the identification of fabrication conditions yielding lower concentration, possibly sin
gle-phase waveguides which are believed to be more stable. Once fabricated, the stability
of these waveguides needs to be quantified and monitored over time.
Finally, characterization of waveguides formed by Ti metal-indiffusion and Zn
vapor-indiffusion needs to be performed. Fabrication conditions leading to the formation
of single-mode waveguides at 1.5pm need to be identified. Since these processes cause no


97
depth (|im)
Fig. 4.15 Extraordinary index profiles for various annealing times. All pro
files were measured at ?i=0.633pm, except that shown by the
dashed-dotted line, which was measured at h= 1.32pm (30-min
annealing time). The actual measured extraordinary index profile
with ripples for 15-min annealing (measured at )i=0.633|im) is
shown by the dashed line.
It should be noted that the measured curves exhibited ripples (shown by the dashed line
only for a sample annealed for ta=15min) whose origin needs further clarification. Taken
alone, possible interference between the beams reflected from the bevelled and bottom
surfaces or a surface profile due to polishing cannot explain this phenomenon, since
virtually no ripples were observed beyond the waveguide cross-section area. Furthermore,
the magnitude of these ripples varies (decreases) with annealing time.


photorefractive characteristic of waveguides in the infrared, near 1,5|im, is measured next
and compared to similar structures in LiNb03. The chapter concludes with some mea
surements of planar waveguides, including index profile measurements and rocking curve
analysis, which will provide data to be used in a subsequent chapter.
4.1 Near-Field Characterization
In this section, a first inspection of the number of modes supported by the channel
waveguides of various widths is carried out. This is done using a near-field characteriza
tion method whereby the output faces of waveguides excited from the single-mode fiber
pigtail of a diode laser source are focused onto a camera so that the number of spots asso
ciated with different modes can be counted off of the display monitor.
waveguide
Fig. 4.1 Schematic of the near-field characterization setup used to initially
determine the number of modes supported by the channel waveguides
of various widths.


109
Fig. 5.2 The complete structural phase diagram for APE X-cut LiTa03,
relating surface index change Ane to proton-induced lattice
deformation e"3 at the surface. Straight-line segments (solid)
were fit to measured data points (squares). Lower-case letters
(a) through (e) mark the locations of waveguides with corre
sponding index profiles shown in Fig. 5.3.
phase and the transitions between phases are marked by sharp discontinuities. Note that
the slope difference between the [31 and [32 phases is small and can only be distinguished
on rotated cuts [Fed94], As such, a single linear fit through both of these phases was
used.
In the next section, the structural phase diagram is analyzed and used to explain the
observed anomalies of the APE process in LiTa03. In addition, the characteristics of
waveguide profiles in different regions of the structural phase diagram are examined.


34
(~800C) results in negligible Li20 outdiffusion from the sample [Y0088], though the
temperature is still high enough to require sample repoling. Curiously, samples after
indiffusion appear slightly grayish-brown in color. This discoloration is likely due to 02
deficiency [Kam73] and is easily removed by a short annealing at about 600C in air.
To fabricate Zn vapor-indiffused waveguides, photolithography was done on the c+
surface of LiTa03 samples using a digitally dark mask to define a pattern, similar to APE.
However, 5000A of Si02 was e-beam deposited at a rate of 2A/s to serve as the mask,
instead of Ta as in APE. Next the sample was subjected to the lift-off process in acetone
and was vacuum sealed in a quartz ampule with powered Zn. The ampule was loaded into
the tube furnace and indiffused at 800C for 5.5 hours with an 8Ymin rise rate and a natu
ral cool down. Afterwards, the Si02 mask was etched off with Ta etchant and the sample
was annealed on a sapphire plate in the box furnace at 600C for 30 minutes to remove the
discoloration. These conditions resulted in single-mode waveguides for the TM polariza
tion at 1.55pm immediately after fabrication.
In the sections to follow, supplemental processing steps such as crystal repoling,
polishing, the application of an antireflectivity (AR) coating, and the electroplating of
traveling-wave electrode structures are also described.
2.3 Crystal Repoling
As mentioned earlier, the processes of Er-doping, Ti metal-indiffusion, and Zn
vapor-indiffusion are all performed above the Curie temperature for LiTa03, which varies
a bit depending on crystal stoichiometry [Bor95], but is generally about 610C. As a
result, the crystal goes from the ferroelectric state to the paraelectric state, where it is


59
Fig. 3.5 Raman spectra of virgin LiTa03
(solid) and Er:LiTa03 (dashed)
measured at X=632.8nm and
J=2.5xl04 W/cm2.
Fig. 3.6 Normalized Raman intensity of
the line at 353cm'1 versus input
intensity measured at 632.8nm
and 785nm for samples with
and without Er.
dependent upon the setup used for the Raman measurements. The value of A has been
determined from Raman measurements of previously studied LiTa03, Cu:H:LiNb03 and
LiNb03 samples [Kos94, Kos97a, Kos97b],
Fig. 3.5 shows the measured Raman spectra of virgin LiTa03 and Er:LiTa03. The
decreased intensity for the Er-doped sample is a result of the increased PR effect. By
making this measurement for several different input intensities J, a plot of normalized
Raman intensity IR' versus J was constructed and is shown in Fig. 3.6. Note that if the PR
effect was absent, IR' would be constant versus J (Fig. 3.4). It is then seen from Fig. 3.6
that all samples suffered some degree of photorefractivity as J was increased. From this
data, Ans at any value of J can be calculated from Eq. (3-2).
Fig. 3.7 shows the calculated value of Ans for the Er:LiTa03 sample with the high
est Er3+ concentration at the surface (-0.65 mol%) and for a virgin LiTa03 substrate.
There are two important observations to be made. First, the presence of Er3+ ions has


CHAPTER 2
FABRICATION OF PFANAR AND CHANNEF WAVEGUIDES IN LiTa03
Planar and channel waveguides are formed by creating an area of higher refractive
index in a dielectric material such as LiTa03. Unlike planar waveguides, channel
waveguides confine the optical wave in two dimensions. Most processes result in a
roughly semi-elliptical, graded-index profile that decreases monotonically from the sub
strate surface [Ram88]. Such a guide is schematically represented in Fig. 2.1, where the
propagation direction of the optical wave is also shown. These waveguides form the basis
for a multitude of devices in the integrated-optic embodiment, in particular, amplitude
modulators used in communication systems, high-speed optical switches, non-linear fre
quency doublers, and optical parametric oscillators.
substrate
Fig. 2.1 Depiction of a channel waveguide.
15


67
e~ migration
conduction band
i
n *
1
recording light
, *JXW\
'
gating light
A/VW
2 polaron
1 bipolaron 3 deep trap
valence band
This model has been developed and
experimentally confirmed for LiNb03
[Jer95, Kos97a] and has been applied
here to LiTa03 due to the similarity of
the intrinsic defect structures between
the two materials [Gop96]. In the
case of undoped lithium niobate
Fig. 3.11 Schematic-level diagram of the two-
photon PR effect depicting the gat- (LiNb03) [Kos97a, Lan98] this mech-
ing light, longer wavelength record
ing light and bipolaron level, single anism involves bipolarons as deep
polaron level, and deep trap (outside
the illuminated area). donors and metastable small polarons
which provide the high PR sensitivity.
Bipolarons and polarons exist due to the intrinsic defect structure of LiNb03, originating
from its lithium deficiency [Sch91 ]. Near infrared (or red) recording [Bai97, Gue97,
Lan98] occurs from the small polaron state consisting of an electron trapped at the antisite
defect (NbLi), designated level 2 in Fig. 3.11. This level is normally populated via an
extrinsically applied blue-green gating light through the photo-dissociation of bipolarons
(level 1 in Fig. 3.11). After the excitation of electrons from the small polaron sites to the
conduction band, the electrons drift out of the illuminated area and are retrapped at bipo
laron sites and other deep traps (level 3) resulting in PR damage, or PR recording if the
application is optical data storage. Such a "two-color" scheme has the advantage of a very
high gating ratio, i.e., the ratio of writing sensitivity with and without the second gating
color [Gue97, Lan98],


150
7.3 DC Response and Power Handling
The DC response of a modulator is a plot of the modulator output versus applied
DC voltage. From this plot, parameters such as the half-wave voltage, or switching volt
age, Vjj and extinction ratio can be determined. The DC response for the modulator fabri
cated here is measured in this section. In addition, the power handling capability of this
device is also characterized for operation near 1.5p.m.
7.3.1 DC Response
The setup for measuring DC response is shown in Fig. 7.4. Light at 1,55pm from a
tunable laser source was coupled into the modulator via a single-mode, polarization-main
taining fiber. The modulator output was collected with a 20x lens and focused onto an IR
detector. A chopper and lock-in amplifier were used to measure detector voltage with low
signal-to-noise ratio as a DC voltage was applied to the device electrodes.
The measured DC response is shown in Fig. 7.5. An extinction ratio, the ratio of
minimum output to maximum output, of -27dB was measured. By fitting a theoretical
cosine response for this modulator to the measured data, a of 13.6V was determined.
Lower values of Wn (about half) have been reported for metal-indiffused, repoled
modulators in LiTa03 [Chu91]. As was seen in Section 4.4, PE waveguides in glycerin
had an electro-optic coefficient r33 of half the bulk value while metal-indiffused, repoled
waveguides had an r33 of close to the bulk value. This explains the discrepancy. Lower
values of have also been reported for APE modulators in LiTa03 [Fin88], The
fabrication conditions in this reference likely resulted in waveguides in a different region
of the structural phase diagram. The exact dependence of r33 on proton concentration or


75
In the next section, a prism coupling technique is applied to measure the effective
mode index increments of channel waveguides at various wavelengths. Such a technique
allows for a rough estimation of the magnitude of the index increment within a waveguide,
as well as providing some information as to the wavelength dispersion of the index incre
ment.
4.2 Effective Mode Index Measurements
The effective index increments of modes in planar and channel waveguides have
been measured successfully for many years by the prism coupling method [Ulr73], These
values not only give some idea on the magnitude of index increase within the guiding
region, but can also be used to reconstruct the index profile by the inverse WKB method.
This naturally requires that the guide supports at least three modes and exhibit a monoton-
ically decreasing profile. In this section, the underlying principle behind the prism cou
pling method to obtain effective mode indices N is presented. A modified version of this
standard prism coupling technique is then described that provides higher accuracy. This
measurement method is applied to determine the effective mode index increments AN for
the fundamental modes of APE channel waveguides in LiTa03.
4.2.1 The Prism Coupling Technique
The standard prism coupling setup is shown in Fig. 4.3. Here, a prism with index
np is located some distance h, on the order of a wavelength or less, above the surface of a
waveguide. The waveguiding region has an index of ng, while the substrate has an index
of ns. The cover index nc is assumed to be unity. The basic principle of operation is based


48
being on the commonly employed APE technique. Several distinct features of the APE
process in LiTa03 were discussed and a few of the currently unresolved anomalies associ
ated with this process were described. The exact fabrication parameters for the planar and
channel waveguides used in this research were next presented, followed by some addi
tional processing steps such as crystal repoling, endface and bevel polishing, the applica
tion of an AR coating, and electroplating. In the next chapter, a characterization of Er-
doping in LiTa03 is performed.


32
For the process of Ti indiffusion, there are no reported anomalies similar to those
of APE. Ti waveguides in LiNb03 appear to be temporally stable, though stability has
never been confirmed for the same process in LiTa03. In addition, this process does not
cause a change in the crystal lattice strain, so there is no structural phase diagram with
multiple phases similar to that of the APE process. However, the surface of samples have
been shown to exhibit swelling in the regions of indiffusion. The thickness of this swell
ing can be as much as 1.5-2.0 times the thickness of the original Ti metal layer before
indiffusion [Hol84]. Additionally, Ti:LiNb03 waveguides are known to have a greater
susceptibility to the photorefractive effect than APE waveguides in LiNb03 [Fuj93],
Due to the high temperatures involved in this indiffusion process, Li20 outdiffu-
sion occurs [Woo93], This is a problem because it causes a small increase of the extraor
dinary index everywhere on the surface of the sample, decreasing the index difference
between the waveguide and its surroundings. Also due to the high temperature, when Ti
indiffusion is performed in LiTa03, crystal repoling is necessary as the crystal is heated
above its Curie point during the process. The method of crystal repoling is described in
Section 2.4.
The fabrication conditions for Ti
waveguides in LiTa03 are as follows. Photoli
thography is done on the c+ surface of a LiTa03
sample to define a pattern using a digitally clear
mask, as opposed to a digitally dark mask used
Fig. 2.9 Depiction of a Ti pattern for APE. Then, 900 of Ti is e-beam deposited
on LiTa03 before indif
fusion. onto the sample after which lift-off in acetone


129
high-strain, high-concentration phases in
the waveguiding layer. However, even
greater stability is necessary for reliable
device operations. Therefore investiga
tion of waveguides in the a phase was
necessary.
6.3.2 The Advantage of PE
Temporal instabilities were sus
pected to be related to the presence of
multiple crystal phases in the waveguide layer, and the possible transformation of these
from one to another over time, as was observed in Section 6.1. As such, there was a valid
suggestion that a waveguide containing only the low-concentration a phase will exhibit
superior temporal stability [E1H95]. To fabricate a waveguides, a significant drawback to
the APE technique using pyrophosphoric acid was discovered. Because of the high proton
concentration of this source, PE led to very shallow, highly concentrated waveguides
which required extensive annealing to reach the a, or even the K, phase. As a result, their
profiles tended to be very graded and mode confinement became poor, or extinct [Mar99],
This problem was especially acute in the region of discontinuity between the a and K
phases, particularly for the Z-cut. APE waveguides supporting a single mode at 1.55pm in
the K phase lost this mode by further annealing them to the a phase because of the large
drop in index increment.
In effort to avoid these limitations, it was decided to switch to a lower-concentra
tion source, glycerin, and to use X-cut crystals. The structural phase diagrams plotted in
Fig. 6.5 Measured coupling length of
Sample 2. This coupler con
sisted of 6.5pm channels with an
8pm gap.


23
the value corresponding to the phase transition. Annealing also serves to recover the
electro-optic r33 and nonlinear d33 coefficients within the exchanged region, known to
have been seriously degraded during the exchange process [hl93, hl94a, Ahl94c,
Cao92, Mat92b],
Every phase transition results in a change in the refractive index. The value of pro
ton concentration at which a phase transition occurs is determined by the specific material,
as is the dependence of refractive index on proton concentration between the phase transi
tions. In other words, the dependence of refractive index on proton concentration is spe
cific to the material and is identified as the structural phase diagram of this material. As
will be shown below, knowledge of the structural phase diagram is crucial in understand
ing the properties and anomalies of APE waveguides.
2.1.3 Anomalies of APE Waveguides
Many anomalies exist which are characteristic of the APE process in LiTa03.
Though some research on these peculiarities has been done, they are not yet fully under
stood. First, the extraordinary index increase Ane that can be achieved by APE in LiTa03
is as much as an order of magnitude less than that in LiNb03 [hl94c, Mat92b, Yuh92], It
has been suggested that this difference is the result of a larger value of spontaneous polar
ization in LiNb03 [Ahl95], which is related to index increment through Eq. (2-1). As for
the magnitude of the ordinary index decrease An0 in APE LiTa03, it has been found to be
many times larger than the extraordinary index increase Ane [hl94c, Mar96b], as
opposed to being of nearly the same value or smaller for the case of APE in LiNb03
[W95].


46
the e-beam bell jar must not be opened between the deposition of Cr and Au as the Cr will
oxidize and, as a result, the Au will not stick to it.
Photolithography of the electrode pattern was performed next. First, the sample
was subjected to 30 minutes of vapor-deposition of photoresist adhesion promoter. Next,
Hoechst Celanese AZ4620 photoresist was spun on at 2000rpm for 55 seconds. The
resulting photoresist thickness was around 9|im. The sample was then baked at 100C for
15 minutes. The edge bead was then removed and it was baked again at 100C for 15 min
utes. After cooling, the sample was aligned under the electrode pattern and exposed at
9mW/cm2 for 1.2 minutes. It was then developed in Hoechst Celanese AZ400K devel
oper, diluted 1:4 with DI water, for 4.5 minutes, after which it was subjected to a baking
sequence of: 90C for 10 minutes, 120C for 30 minutes, 90C for 10 minutes. Finally, the
Fig. 2.13 Setup for the electroplating process.


CHAPTER 8
CONCLUSIONS AND FUTURE WORK
8.1 Conclusions
Several contributions toward the implementation of integrated-optical devices ben
efiting from the high photorefractive threshold and shorter wavelength UV absorption
edge of LiTa03 have resulted based on this research. In particular, a complete character
ization on the effect of Er-doping for active waveguide lasers and optical data storage
applications has been performed, opening the door for the development of such devices
which were shown to have several advantages when fabricated in LiTa03. Additionally,
various methods of waveguide fabrication, namely APE, PE, Ti-indiffusion, and Zn vapor-
indiffusion, were researched in detail. From this, the structural phase diagram for
APE:LiTa03 was constructed and used to explain the previously observed anomalies asso
ciated with this process in LiTa03, and also to identify APE/PE fabrication conditions
yielding waveguides exhibiting improved temporal stability. Finally, the only factor hin
dering the widespread use of LiTa03, namely the instability of waveguides, was found to
not be dependent upon fabrication methods, but rather to be inherent to the crystal itself,
the likely result of a slightly underdeveloped growth technology for this crystal. The
major accomplishments for this work are detailed below.
158


139
Days
PE (X-cut)
PE (Z-cut)
Zn (Z-cut)
Ti (Z-cut)
0
2-3|im
2,2.5-6|im
2-5|im
2.5-4.5|im
1
2-4,4.5|im
2.5,3-7.5|im
4-7.5,8|itm
2
3.5,4-6p.m
3
2-4.5|im
2.5-7.5,8pm
5.5,6-10pm
5
4-6.5 pm
8
2-4.5|im
2.5,3-8(im
11
7-10|im
20
4.5-7p,m
22
2-5|im
3.5-9fim
26
7,7.5-lOpm
29
4.5-7.5|im
Table 6.4 Regions of single-mode operation, in terms of channel width, for the
extraordinary mode measured at 1.55p.m. Samples shown include those
fabricated by PE, Zn vapor-indiffusion and Ti-indiffusion.
employed to determine the region of single-mode operation, in terms of waveguide width,
at 1.55|im for each of the samples. The single-mode region was monitored over time.
The results of this analysis are depicted in Table 6.4. From this, it is seen that all
fabrication processes suffered a shift in the region of single-mode operation of at least 2-
3|_im over the period they were examined. As a result, it is concluded that the stability
issue in LiTa03 does not have the same dependence on fabrication conditions as does
LiNb03. Additionally, instability does not appear to be caused by induced lattice strain, as
the metal-indiffused waveguides have no measurable e"3, but are still as, or more,
unstable as PE and APE waveguides. Likewise, H+ mobility also does not appear to be the


A number of obstacles, however, prevent the large-scale development of optical
devices in LiTa03. The various processes of waveguide formation have not been com
pletely investigated or characterized. For instance, the annealed proton exchange (APE)
technique is known to exhibit a number of anomalies which to date have not been under
stood, the most significant being the temporal instability of the waveguide index incre
ment. Additionally, no work has been performed to characterize the impact of Er-
indiffusion, necessary for the fabrication of 1.5p,m lasers.
In this dissertation, the above issues are addressed. Conditions for the indiffusion
of Er in LiTa03 are identified. Fluorescence and Raman spectroscopy are used to identify
energy transfer upconversion (ETU) between neighboring clusters of Er3+ ions as the
dominant mechanism of upconversion, leading to increased photorefractivity, a gain-limit
ing factor for lasers. Upconversion is then decreased through a novel Li-treatment process
to reduce clustering, with results showing a significant advantage over LiNb03. Addition
ally, the structural phase diagram for APE:LiTa03 waveguides is constructed and used to
explain the previously observed anomalies associated with this process in LiTa03. The
stability of waveguides fabricated by various techniques is also examined, resulting in the
determination that instability does not depend on fabrication conditions, as it does in
LiNb03, but rather it is inherent to the crystal, the likely result of an immature growth pro
cess. Finally, to demonstrate the potential of LiTa03 and indicate the desire for higher
quality crystals, a high-speed traveling-wave modulator is fabricated and its measured
results compared to theory. This work hopefully will inspire crystal growth companies to
arrive at a growth process which produces crystals with better crystal stoichiometry, desir
able for improved device performance.


13
analysis. Resulting values of surface index increment An and lattice strain e"3 are plotted
to construct the structural phase diagram. The diagram is then used to explain the previ
ously observed anomalies associated with the APE process in LiTa03 and the effect of
crystal phases on the shape of the waveguide index profile is also examined, with exam
ples of measured profiles from different regions of the structural phase diagram. Lastly,
proton exchanged (PE) waveguides fabricated in a dilute source are examined, exhibiting
advantages over APE waveguides. These waveguides are characterized by the same direct
measurement techniques used on APE waveguides and by Raman spectroscopy to deter
mine the presence of an additional single-phase region of the structural phase diagram
obtainable only by PE in a dilute source.
Chapter 6 examines the stability issue of waveguides in LiTa03. APE waveguides
are inspected by rocking curve analysis to track changes in lattice strain e"3 over time.
For a more quantitative analysis, directional couplers are fabricated from different regions
of the phase diagram and changes in measured coupling length are related to correspond
ing changes in index increment 8(An) through computer simulation of a directional cou
pler. Results indicated that stability is not caused by the presence of multiple phases, as
previously assumed. To determine if stability depends on fabrication conditions, the sta
bility of Ti-indiffused and Zn vapor-indiffused are also measured, showing no improve
ment in stability, unlike LiNb03. Suspecting inferior growth technology to be the reason
for the instability, crystals from several different growers worldwide, both SAW and opti
cal grade, are obtained and compared in terms of stability only to find that all are similar
and unstable.


141
Days
Yamaju
Sichuan
Shin-Etsu
Deltronic
Yamaju/Li
0
2,2.5-8pm
2.5-7.5,8pm
2-6.5,7pm
2-5pm
2-4.5pm
1
5,5.5-10pm
5-9pm
4.5-9.5pm
4-7.5,8pm
2.5,3-6.5pm
3
5.5,6-10pm
5-9.5pm
4.5-9.5,10pm
5.5,6-10pm
6
3.5-7.5pm
9
3.5-8pm
11
7-10pm
12
6,6.5-10pm
5- 10pm
5.5-10pm
14
3.5-8pm
20
6.5,7-10|im
5-10pm
5.5-10pm
26
7,7.5-lOpm
Table 6.5 Regions of single-mode operation, in terms of channel width, for the
extraordinary mode measured at 1.55pm. Samples shown are from
four different vendors, and one that was Li-treated before waveguide
fabrication.
6.5.1 Comparing Different Growers
LiTa03 samples of both optical and SAW grade were obtained from four different
suppliers world-wide. The suppliers were: Yamaju Ceramics Co., Ltd., Sichuan Institute
of Piezoelectric and Acousto-Optic Technology, Shin-Etsu Chemical Co. Ltd., and Del-
tronic Crystal Industries, Inc. Using these samples, a comparison could be made between
different growers to determine if the samples previously used, grown by Deltronic, were
inferior, or if all LiTa03 grown world-wide was similar, and unstable. Additionally, hav
ing both optical and SAW grade material would give some information about the maturity
of the growth process for this crystal. Optical grade should have lower impurity levels and
smaller strain and therefore, should be distinguishable from SAW grade.


66
attributed to the decrease of the Er3+ cluster site concentration as a result of the increase in
Li content in the near-surface layer. Similar to the effect of vapor phase equilibration in
Er:LiNb03 [Gil96], a redistribution of Er3+ sites occurs in response to a change in the
defect populations created upon Er incorporation. It is important to note that the technique
of post-indiffusion processing in a Li-containing melt described here has the advantage of
simplicity, as well as being performed at temperatures below the Curie point for LiTa03.
This Li-treatment process affects only a shallow near-surface layer. However, it is suffi
ciently deep to reduce the Er3+ cluster concentration and therefore may be a useful tech
nique for the optimization of active integrated-optical devices utilizing an Er-doped near
surface layer in LiTa03 crystals.
To this point, it has been determined that ETU is the dominant mechanism of
upconversion in Er:LiTa03 and that the presence of this upconverted light leads to PR
damage via the two-photon, or two-color, PR effect. A two-photon model used to exem
plify this process is described in the next section.
3.6 The Two-Photon Model of the PR Effect:
LiTaQ3s Advantage in Data Storage and Lasers
In this section, a two-photon model is applied to model the PR effect and explain
how upconverted light, along with the presence of the longer wavelength incident light,
leads to PR damage. During this process, the inherent advantages of LiTa03 for applica
tions such as optical data storage and laser fabrication will also be described.
To develop a theoretical explanation of the phenomenon of photorefractivity
enhancement by Er-doping, the two-photon model of the PR effect has been employed.


CHAPTER 5
STRUCTURAL PHASE DIAGRAM FOR APE:LiTa03
Waveguide devices fabricated in LiTa03 by APE have the advantage of benefiting
from the high-power throughput capability of the crystal. However, the APE process in
LiTa03 is known to exhibit a number of anomalies. These anomalies were discussed at
length in Chapter 2 and so are only briefly reviewed here. APE waveguides in LiTa03 have
been found to exhibit significant short-term and long-term instabilities of the refractive
index increment [Mat92a]. It has been shown that the total hydrogen content and refractive
index increase upon annealing [Ahl94c, Mat92b] and that the index decreases as proton
concentration increases beyond a certain level [hl94c], These peculiarities suggest a non
linear dependence of proton concentration on refractive index which may result in buried
index profiles for certain fabrication conditions, as was confirmed in the previous chapter.
To explain the observed abnormal behavior, it is therefore important to understand
the relationship between proton concentration and index change. Since the introduction of
protons causes crystal deformation and strain, the relationship between proton-induced lat
tice strain and index change is equally as important. Such a relation has become identified
as the structural phase diagram [Kor96], not to be confused with the phase diagrams used
in physical chemistry to relate temperature to proton concentration, and is illustrated in Fig.
5.1. In general, this diagram is usually plotted using lattice strain e3 rather than H+ proton
concentration, as the e" measurement is nondestructive and simpler to make. Knowledge
105