University of Southern California Dissertations and Theses

Development of organic-inorganic optical microcavities for studying polymer thin films

(USC Thesis Other)

Development of organic-inorganic optical microcavities for studying polymer thin films

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

DEVELOPMENT OF ORGANIC-INORGANIC OPTICAL
MICROCAVITIES FOR STUDYING POLYMER THIN FILMS

by

Hong Seok Choi

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(MATERIALS SCIENCE)
May 2012

Copyright 2012 Hong Seok Choi
ii

Acknowledgments

I feel very lucky to spend part of my life at USC as I have had a really enjoyable
time and met various people at here both for academic and personal life. I am sure that
my experiences at USC will be a foundation to be part of important roles as one of
researchers in the society.
I would like to thank my PhD advisor, Prof. Andrea M. Armani, for giving me the
opportunity to be part of her research group even with my lack of knowledge about
optical resonators and for her great mentoring both for the research projects and my
personal life during my PhD study at USC. It has been my great pleasure to work with
her. Upon meeting and talking with her first time, I was surprised and also impressed
how well-prepared she was to give a presentation to just one graduate student. I still
remember that she told me “Be positive, and then you will and can learn any necessary
research knowledge and skills.” I was extremely impressed by her diligence and her
desire to help students become great researchers. Whenever I have some questions and
concerns, she was never too busy or too tired to answer or help, but passionate about
helping me to understand my research at every level or to resolve the concerns. And
whenever I have had a meeting with her, I’ve been always inspired by her insightful
comments and suggestions. She is also happy to listen and share my personal concerns
whenever possible.
The Ph.D study is a long journey that could be frustrating and boring. However,
I’ve found that it could also be an enjoyable time full of discovery and learning if you
iii

have a great mentor. The relationship between graduate students and advisors is really
important to be successful both academically and professionally. Prof. Armani really
makes me positive and passionate about my research topic, which ensures that I never get
tired or bored of my research. I am sure these built-up attitudes toward research will lead
to and ensure future success.
I have enjoyed working with all of group members in the lab. I would particularly
like to thank Ce Shi, Simin Mehrabani, Matt Reddick, Shehzad Ismail, Xiaomin Zhang,
Dania Neiroukh, Dr. Rasheeda Hawk and Prof. Heather Hunt for their fruitful
collaborations. Special thanks go to Dr. Jason Gamba, my very first collaborator, for
showing me all the techniques that were needed for optical resonator characterization
while he was still a PhD student at Caltech. I also would like to thank Ashley Maker for
her continuous supply of the energetic cookies for group members and Kelvin Kuo,
Maria Chistiakova, Brian Rose for showing me various aspects of life outside of USC. I
would also like to thank to Ce Shi, Biliang Hu, Simin Mehrabani, Hooman, Kelvin Kuo
and Maria Chistiakova for their extension of the friendship to my family.
I would like to thank my friends who I met at USC for sharing useful discussions
and for spending time with me outside of the lab. Many thanks go to Bojun Kim,
Kyoungmin Ryu, Jinwoo Park, Seounghoon Hong, Hsiusheng Hsu, Pochiang Chen,
Fumiaki Ishikawa, Byoungmin Ahn and many others.
Last but not least, I would like to express my deepest gratitude to my respectable
parents, my lovely wife (Liang-Tzu (Lisa) Lin), my lovely unborn son (Mason) and my
cute dog (Pang-pang). My parents always encourage and support me during my entire life
iv

with invaluable wisdom. My wife, Liang-Tzu (Lisa) Lin, is my soul mate and has shown
me true love, strength, patience, laugh and warmness. My unborn son, Mason, has given
me delight by just his existence and I can’t wait to see him in person . My cute dog, pang-
pang, has become my family member and it was a great decision to have him in our
family. All of them have been kept me to go through any difficulties that I have had
during my PhD study. Therefore, this thesis is devoted to them with my greatest respect
and gratitude as it could not be possible to have this moment without their support and
love.

v

Table of Contents

Acknowledgments............................................................................................................... ii

List of Tables ................................................................................................................... viii

List of Figures .................................................................................................................... ix

Abstract ..........................................................................................................................xx

Chapter 1 Introduction ........................................................................................................1
1.1 Motivation ............................................................................................................ 1
1.2 Chapter overview ................................................................................................. 2
Chapter 1 References ...................................................................................................... 6

Chapter 2 Overview of Optical Microtoroidal Resonators .................................................7
2.1 Introduction .......................................................................................................... 7
2.2 Background ........................................................................................................ 10
2.2.1 Quality (Q) factor ........................................................................................ 10
2.2.2 Mode Volume ............................................................................................. 12
2.3 Microtoroid Fabrication ..................................................................................... 13
2.3.1 Define micropad by patterning oxide layer ................................................ 13
2.3.2 Fabricate microdisk using the xenon difluoride (XeF
2
) etcher ................... 15
2.3.3 Fabricate microtoroid using a carbon dioxide (CO
2
) laser reflow .............. 17
2.4 Device testing procedure .................................................................................... 18
2.4.1 Pulling a tapered optical fiber ..................................................................... 18
2.4.2 Alignment of optical fiber with microtoroid resonator ............................... 20
2.4.3 Getting data ................................................................................................. 21
Chapter 2 References .................................................................................................... 25

Chapter 3 Finite Element Method (FEM).........................................................................27
3.1 Introduction ........................................................................................................ 27
3.2 Microtoroid simulations ..................................................................................... 28
3.3 Microsphere simulations .................................................................................... 33
3.4 Microdisk simulations ........................................................................................ 34
3.5 Conclusion .......................................................................................................... 35
Chapter 3 References .................................................................................................... 36

Chapter 4 Hybrid Silica-Polymer Ultra-High-Q Microresonators ...................................37
4.1 Introduction ........................................................................................................ 37
4.2 Fabrication .......................................................................................................... 37
vi

4.3 Theory ................................................................................................................ 39
4.3.1 Loss mechanisms for hybrid devices .......................................................... 39
4.3.2 Material absorption loss for PMMA and PS ............................................... 40
4.3.3 Determination of β, γ and δ ......................................................................... 41
4.4 Experimental results and discussions ................................................................. 47
4.5 Conclusion .......................................................................................................... 50
Chapter 4 References .................................................................................................... 51

Chapter 5 Thermal nonlinear effects in hybrid optical microresonators ..........................52
5.1 Introduction ........................................................................................................ 52
5.2 Theory ................................................................................................................ 53
5.3 Experimental Results and Discussions ............................................................... 60
5.4 Conclusions ........................................................................................................ 66
Chapter 5 References .................................................................................................... 68

Chapter 6 Studying polymer thin films with hybrid optical microcavities ......................69
6.1 Introduction ........................................................................................................ 69
6.2 Theory ................................................................................................................ 70
6.3 Experimental Results and Discussions ............................................................... 71
6.4 Conclusions ........................................................................................................ 78
Chapter 6 References .................................................................................................... 80

Chapter 7 Thermo-optic coefficient of polyisobutylene ultra thin films measured with
integrated photonic devices ................................................................................................82
7.1 Introduction ........................................................................................................ 82
7.2 Background & Motivation ................................................................................. 82
7.3 Methods and Materials ....................................................................................... 84
7.3.1 Surface-initiated cationic polymerization ................................................... 85
7.3.2 Spin-coating ................................................................................................ 89
7.4 Results and discussions ...................................................................................... 89
7.4.1 Transmission (Material) loss determination ............................................... 89
7.4.2 Thermo-optic coefficient determination ..................................................... 93
7.5 Results and discussions ...................................................................................... 98
Chapter 7 References .................................................................................................. 100

Chapter 8 Measuring interface effects in the glass transition temperature of ultra thin
polymer films ...................................................................................................................103
8.1 Introduction ...................................................................................................... 103
8.2 Background & Motivation ............................................................................... 104
8.3 Thoery .............................................................................................................. 107
8.4 Experimental results and discussions ............................................................... 114
8.5 Conclusions ...................................................................................................... 124
Chapter 8 References .................................................................................................. 126

vii

Bibliography ....................................................................................................................129

Appendix A: Ultimate quality factor of silica microtoroid resonant cavities ..................140
A.1 Introduction ....................................................................................................... 140
A.2 Fabrication ......................................................................................................... 140
A.3 Theory ............................................................................................................... 142
A.4 Experimental results and discussions ................................................................ 144
A.5 Conclusions ....................................................................................................... 146
Appendix A References .............................................................................................. 148

Appendix B: Excitation of Cy5 in self-assembled lipid bilayers using optical
microresonators ................................................................................................................149
B.1 Introduction ....................................................................................................... 149
B.2 Theory ................................................................................................................ 150
B.3 Experimental procedure ..................................................................................... 151
B.3.1 Fabrication of Microspheres ....................................................................... 151
B.3.2 Self Assembly of Lipid Bilayers ................................................................. 152
B.3.3 Confirmation of bilayer structure................................................................ 154
B.4 Experimental results and discussions ................................................................ 158
B.4.1 Characterization of Q factor......................................................................... 158
B.4.2 Measurement of Cy5 Emission .................................................................... 159
B.4.3 Photodynamics of Cy5 ................................................................................. 164
B.5 Conclusions ....................................................................................................... 166
Appendix B References ............................................................................................... 168

Appendix C: Optical Microcavities with a Thiol-Funcionlized Gold Nanoparticle
Polymer Thin Film Coating .............................................................................................171
C.1 Introduction ....................................................................................................... 171
C.2 Motivation ......................................................................................................... 171
C.3 Theory ................................................................................................................ 173
C.4 Materials and Methods ...................................................................................... 174
C.4.1 Synthesis of Gold Nanoparticles .................................................................. 175
C.4.2 Thiol Functionalization of Gold Nanoparticles ........................................... 176
C.4.3 Concentration of PMMA-Gold Nanoparticle Solutions .............................. 176
C.4.4 Characterization of PMMA-Gold Nanoparticle Thin Film ......................... 177
C.4.5 Material Absorption Spectra ........................................................................ 179
C.4.6 Device Fabrication and Lifetime ................................................................. 180
C.5 Results and Discussions .................................................................................... 181
C.5.1 Typical spectra for Q measurement of PMMA-gold coated devices ........... 181
C.5.2 Q measurement results ................................................................................. 182
C.5.3 Photon Lifetime ........................................................................................... 185
C.6 Conclusions ....................................................................................................... 186
Appendix C References ............................................................................................... 187

viii

List of Tables

Table 3-1 The relationship between the frequency and wavelength 30

Table 4-1 Summary of Model and Experimental Fit Parameters. 50

Table 5-1 Experimental Values. 65

Table 7-1 The ellipsometry results of the polyisobutylene ultra thin films 89

Table 7-2 The optical properties of the polyisobutylene ultra thin films 98

Table A-1 Summary for the values of refractive index and quality factor at λ=635,
850 and 980nm and its corresponding boron concentrations (*The error
in a single measurement of an individual resonant linewidth is ± 1×10
7
.
The variation in Q between a significant number of cavities is much
larger, as a result of imperfections in fabrication. Therefore, values
reported in this table are the highest achieved with each type of cavity. 142

Table B-1 FEM Simulation Results 151

Table C-1 Summary of calculations and spectroscopic ellipsometry results. 179

Table C-2 Summary of Model and Experimental Fitting Parameters 185

Table C-3 Conversion of experimental quality factors to photon lifetime 185

ix

List of Figures

Figure 2-1 (a) St. Paul’s Cathedral Whispering Gallery and (b) Microtoroidal
optical resonators with the optical fiber waveguide showing the light
propagation (highlighted as red) 8

Figure 2-2 The geometric optics showing that the resonance occurs when the light
propagates (a) integral fashion and (b) the other cases it will be off
resonance. 8

Figure 2-3 Various types of the WGM optical micro-resonators based on different
geometries 9

Figure 2-4 Overview of the fabrication process. a) Circular oxide pads are
photolithographically defined on a silicon wafer, b) XeF
2
undercuts the
oxide, forming a microdisk, c) a CO
2
laser reflows the oxide, forming a
microtoroid. a) and b) are redenerings and c) is a scanning electron
micrograph. 13

Figure 2-5 (a) The system for top portion of XeF
2
etcher (b) control panel on
computer screen 16

Figure 2-6 A portion of CO
2
laser reflow set up 18

Figure 2-7 (a) Taper pulling set-up and (b) magnified image of its core region 19

Figure 2-8 An example of pulled optical fiber. We can clearly see the rainbow color
like distribution from optical fiber. This is due to optical filtering effect. 20

Figure 2-9 The portion of cavity testing set-up 21

Figure 2-10 (a) Top and (b) side view of aligned resonator with optical fiber 21

Figure 2-11 A schematic diagram of the optical resonator’s testing set -up 22

Figure 2-12 An example of (a) broad and (b) fine scan spectra for silica
microresonators. 24

Figure 3-1 A FEM result of the cross-sectional electric field intensity (

) profile
of the 40(3) μm major(minor) diameter of the silica microtoroidal
optical resonators, with 115
th
TE polarized mode. 29
x

Figure 3-2 A FEM result showing the normalized radial intensity VS radius for the
40(3) μm major(minor) diameter of the silica microtoroidal optical
resonators, with 115
th
TE polarized mode. 30

Figure 3-3 A FEM result showing the various modes existing in the 40(3) μm
major(minor) diameter of the silica microtoroidal optical resonators,
with 115
th
TE polarized mode. 31

Figure 3-4 A FEM result of the cross-sectional electric field intensity (

) profile
for a 100nm thick PMMA coating on the 40(8) μm major(minor)
diameter of the silica microtoroidal optical resonators, operating
wavelength at 980nm. 32

Figure 3-5 FEM results of the cross-sectional electric field intensity (

profile of
the 40μm diameter of the silica microsphere optical resonators, with
115
th
TE polarized mode. 33

Figure 3-6 FEM results of the cross-sectional electric field intensity (

) profile
of the 40μm diameter of the silica microdisk optical resonators, with
115
th
TE polarized mode. 34

Figure 4-1 PMMA film thickness as a function of spin speed 38

Figure 4-2 Optical images shown fabrication procedure. Silica (a) Micropad, (b)
Microdisk, (c) Microtoroid and (d) 100nm thick PMMA coated
microtoroid on Silicon wafer. The bluish color on the toroidal rim is
due to the polymer coating with this coating thickness. 39

Figure 4-3 Material absorption loss measured by Cary 14 UV-Vis
spectrophotometry for (a) PMMA and (b) PS. 41

Figure 4-4 Finite simulation result which shows optical profile (a) without and (b)
with 500nm thickness of PMMA at λ=980nm. It is cross section of one
side toroidal ring. 42

Figure 4-5 Finite simulation results. Percentage of the optical field in the polymer
layer at λ=980nm as a function of the minor diameter as (a) the major
diameter increases from 40 to 100μm. The film thickness is fixed at
500nm. (b) the film thickness change from 100 to 500nm. The major
diameter is fixed at 40μm 43

Figure 4-6 Finite element method results. (a) Percentage of the optical field in the
silica layer as a function of the minor diameter as the PMMA film
thickness increases from 0 to 500nm at (a) λ=850nm and (b) λ=980nm
. The major diameter of the device is fixed at 40μm. 44
xi

Figure 4-7 Finite element method results. (a) Percentage of the optical field in the
PMMA layer as a function of the minor diameter as the PMMA film
thickness increases from 100 to 500nm at (a) λ=850nm and (b)
λ=980nm . The major diameter of the device is fixed at 40μm. 44

Figure 4-8 (a) Scanning electron micrograph of a toroidal microresonator. Finite
element method simulation results for the optical field intensity
distribution: (b) silica microtoroid, (c) hybrid microtoroid with a
100nm thick PS film, and (d) 200nm thick PS film. Note that the
optical field shifts from the silica towards the polymer film as the
thickness of the polymer film increases. The major (minor) diameter is
40(8) μm and λ=850nm. 45

Figure 4-9 Finite element method results. Percentage of the optical field as a
function of the polymer film thickness for PMMA and PS at λ=850,
980nm. 46

Figure 4-10 Summary of FEM simulation result for PMMA (a) λ=850 (b) λ=980
and for PS (c) λ=850 (d) λ=980. 47

Figure 4-11 An example of (a) broad and (b) fine scan spectra for hybrid
microresonators. 48

Figure 4-12 Experimental and theoretical Quality factor (Q) as a function of
polymer thickness. PMMA at (a) 850nm and (b) 980nm. PS at (c)
850nm and (d) at 980nm. The results were fit to an equation of the form
y=ax
b
, which is included as a solid (dashed) red line for the theoretical
(experimental) Q results. The black dotted line indicates the highest Q
demonstrated with a silica toroidal resonant cavity to date, setting an
upper bound on Q. 49

Figure 5-1 Polymer coated (Hybrid) optical microtoroid resonators. a) Pov-Ray
rendering of the polymer coated devices. The yellow line in the cavity
indicates that whispering gallery mode is confined within hybrid
device. Green color indicates polymer film which part of it is cut away
for clarity. b) An optical image of a fabricated hybrid device. 53

Figure 5-2 FEM simulation results. A set of optical field distribution for (a) silica
microtoroid, (b) hybrid microtoroid with a 100nm thick PMMA film
and (c) hybrid microtoroid with a 100nm thick PS film with device size
of 40(8) μm major (minor) diameter at λ=980nm. As can be seen, the
part of optical field is interacting with polymer films both for PMMA
and PS coated hybrid devices due to higher refractive index of coated
polymer compared to silica. 54
xii

Figure 5-3 The normalized radial optical field intensity as a function of radius for
silica (solid black line), 100nm thick PMMA coated (red dashed line)
and 100nm thick PS coated (blue dotted line) devices at λ=980nm. The
zero point indicates that the center of minor diameter. 55

Figure 5-4 Calculated percent of optical field in SiO
2
layer with 40μm major
diameter of the device as a function of minor diameter at =850 and
980nm. The simulations results shows with no film, 100nm and 200nm
of PMMA coatings. As can be seen, the optical field is highly
dependent on polymer film thicknesses and operating wavelengths. 56

Figure 5-5 Calculated resonant wavelength shift as a function of temperature with
silica and hybrid devices at =850nm and 980nm for both PS and
PMMA. a) 100nm thick PS and PMMA films at 850 and b) 980nm. c)
200nm thick PS and PMMA films at 850 and d) 980nm. As can be
seen, depening on device kinds, we can see different optical responses. 58

Figure 5-6 Calculated resonant wavelength shift as a function of temperature with
silica and hybrid devices at =850nm and 980nm for both PS and
PMMA. a) 100nm and 200nm PMMA films at 850 and b) 980nm. c)
100nm and 200nm thick PS films at 850 and d) 980nm. As can be seen,
depening on device kinds, we can see different optical responses. 59

Figure 5-7 The examples of the thermal broadening and compression peaks for (a)
silica and (b) hybrid (PS200nm) devices at λ=850nm. While the
thermal broadening occurs in forward scan for silica device (red shift),
it occurs in backward scan for hybrid device (blue shift). The positive
or negative effective thermal-optic coefficients play a major role to
cause red or blue shift. 61

Figure 5-8 Thermal nonlinear shift determination from under to critical coupled
regime 62

Figure 5-9 The resonance shift for silica, PMMA 130nm and 200nm thick coated
hybrid devices for (a) =850nm and (b) =980nm with different quality
factors as a function of input powers. The resonant shift is larger with
higher Q due to high circulating build up intensity in the optical
cavities. Athermal condition is nearly achieved with 200nm thick
PMMA coated hybrid devices. Lines are a guide to the eye for ease of
comparison. 63

xiii

Figure 5-10 The resonance shift for silica, PS 100nm and 200nm thick coated
hybrid devices for (a) =850nm and (b) =980nm with different quality
factors as a function of input powers. The resonant shift is larger with
higher Q due to high circulating build up intensity in the optical
cavities. Athermal condition is nearly achieved with 100nm thick PS
coated hybrid devices. Lines are a guide to the eye for ease of
comparison. 64

Figure 6-1 Sample stage with integrated heater and temperature control. 72

Figure 6-2 The measured resonance shift as a function of time in air. The resonance
peak shows stable in air, indicating that the current testing set up is
good for thermal sensing purpose. 73

Figure 6-3 The measured resonance shift as a function of time. (a) continuous heat
up to set point and keep constant temp. (b) heat was on and off. 74

Figure 6-4 Comparison of sensor response between silica and hybrid devices. a)
The T is the same for each increment. b) The T is increasing for
each increment. Because of the polymer layer, the hybrid device has a
significantly larger response. Note that the y-axis is the absolute value
of the resonant wavelength shift or the magnitude, for easy comparison
between the two devices. 75

Figure 6-5 The results from experiments like those shown in Figure 6-4 were
compiled to show the shift versus temperature for both the hybrid and
silica devices. Additionally, the theory based on equation 6.1 is
included. The shift for the silica device is significantly smaller than the
shift for the hybrid device, and the theory and experiment are in good
agreement. 76

Figure 6-6 Reproducibility of results a) The measurement was performed
iteratively. b) The data in part a) is re-plotted to emphasize the
hysteretic behavior due to the heating element. 77

Figure 7-1 Reaction scheme for forming the polymer brush on the silica surface
using surface-initiated cationic polymerization. 85

Figure 7-2 Spectra showing the attachment of the polyisobutylene to the silica
surface, and a control spectrum of the silica surface. If the polymer
layer is not thoroughly washed, residual Al from the catalyst is visible
(middle spectra). 88

xiv

Figure 7-3 The integrated photonic sensor. a) Scanning electron micrograph of a
single integrated photonic sensor device. b) Rendering of the sensor,
highlighting the location of the circulating optical field and its
interaction with the polymer film. c) Schematic of the testing set-up
sample stage with the heating element and thermocouple. Multiple
sensors operating at different wavelengths are shown. d) A finite
element method simulation of the optical field distribution in the
optical sensor system, showing the interaction of the optical field with
the polymer film. 91

Figure 7-4 Transmission spectra of a resonant wavelength of the optical cavity. At
resonant wavelengths of the cavity, light is coupled into the cavity,
resulting in a decrease in detected power. 92

Figure 7-5 The dependence of the experimentally measured Q factor on the amount
of light which is coupled into the sensor device. By fitting this data
and extrapolating to the 0 point, the intrinsic Q factor (Q
0
) or material-
limited Q factor (Q
mat
) can be determined. 92

Figure 7-6 The dependence of the resonant wavelength on the temperature. The
slope of this line is d/dT, which is directly proportional to dn/dT, the
desired quantity. 94

Figure 7-7 The dn/dT of silica measured using the optical sensor. The value of the
slope is 1.2E-5 C
-1
, which corresponds to the dn/dT of silica. 95

Figure 7-8 The change in the refractive index as the temperature around the sensor
is incrementally increased with temperature. The thermo-optic
coefficient (dn/dT) is the slope of this graph. As can be clearly
observed, the optical behavior of the spun-coat films and the polymer
brush films was drastically different and the general behavior was
independent of characterization wavelength, as expected. The films
attached using the polymer brush approach exhibited a negative dn/dT
whereas the film deposited using spin-coating demonstrated a positive
dn/dT. 96

Figure 8-1 A scanning electron microscope image of the fabricated silica optical
micro-sensor integrated on a silicon substrate. Inset: A rendering of the
silica sensor in operation. The light is partially confined within the
device, but also interacts with the polymer film coating enabling
detection of the polymer film behavior. 107

xv

Figure 8-2 (a) An example simulation showing the interaction of the optical field
with the substrate, polymer film and air. The sensor size is 40(8)m of
major(minor) diameter and film thickness is 100nm. The operating
wavelength is 765nm. As can be seen, the optical field is clearly able to
easily interrogate the entire polymer film, with the strongest interaction
occurring at the substrate-film interface. (b) FEM simulation results
with a range of film thicknesses at different probing wavelengths.
Sensor size is 40(8)m of major(minor) diameter and film thickness.
The results show that different locations in polymer films can be
probed depending on film thicknesses and probing wavelengths without
embedding additional materials in the film. 110

Figure 8-3 The calculated results showing the relationship between the fraction in
polymer () and polymer film thickness or resonant shifts (a) with 100,
200, 250nm thick PS at =980nm and (b) with 250nm thick PS at
=635, 850, 980 and 1550nm. 111

Figure 8-4 The normailized radial optical field intensity as a function of radius for
0 (dashed black line), 20 (solid black line), 50 (dashed red line), 100
(dotted blue line), 150nm (dash-dotted green line) thick polystyrene
coated optical microresonators at (a) =765 and (b) 1330nm. The zero
point indicates that the surface of the silica (SiO
2
) sensor. 112

Figure 8-5 Effective mode area as a function of polymer film thickness (20, 50,
100, 150nm) at λ=765, 980 and 1330nm with (a) total mode area
(silica, polymer, air) and (b) mode area only in polymer. 113

Figure 8-6 (a) An example of testing results for silica microtoroid and (b) testing
results showing d/dT as a function of temperature at 20, 40, 60 and
80°C. 115

Figure 8-7 Experimental (black square dot) and theoretical (red triangle dot) results
showing d/dT as a function of temperature at 20, 40, 60 and 80°C. 115

Figure 8-8 The feasibility and reproducibility test for continuous thermal history
test. 117

Figure 8-9 Iteration test around Tg to see material degradation by observing optical
response. 118

xvi

Figure 8-10 Optical response as a function of temperature with (a) 20nm and (b)
150nm thick PS coated hybrid device tested at =1330nm. and (c) glass
transition temperature for all film thicknesses tested as a function of
temperature. 119

Figure 8-11 (a) Example testing results with 100nm thick polystyrene film
thickness deposited on silica microtoroid at =980 and 1330nm
showing the change in the effective refractive index of the polymer as a
function of temperature. The T
g
point at 75.7°C at =980 and at
68.8°C =1330 are clearly identifiable. (b) A compilation of all of the
experimentally measured data which shows the glass transition
temperature as a function of temperature with three different
wavelengths (765, 980, 1330nm). As can be seen, the T
g
point is
strongly dependent on the film thickness and the measurement
wavelength. 121

Figure 8-12 Glass transition temperature as a function of effective mode area with
20, 50, 100 and 150 film thicknesses at λ=765, 980 and 1330nm for (a)
total mode area (silica, polymer, air) and (b) mode area only in
polymer. 123

Figure A-1 A scanning electron micrograph of the fabricated silica microtoroid
resonator. 141

Figure A-2 The calculated dependence of the quality factor on the refractive index
and the resonant wavelength. The Q factor decreases as the refractive
index decreases and the wavelength increases. 143

Figure A-3 FEM results which shows different mode volume for (a) λ=635 and (b)
λ=980 144

Figure A-4 A fine scan (the forward scan direction) of the fundamental transverse
mode of the microtoroid fabricated from the film with boron
concentration of 1.63×10
14
cm
-3
in silicon at 848.8nm with a dual-
Lorentz fit. The resonance shows splitting and the quality factor of the
left and right peak is 3.39×10
8
and 2.67×10
8
separately. Inset: Optical
micrograph picture of a microtoroid coupled to a tapered optical fiber
during testing. 145

xvii

Figure A-5 Measured quality factor as a function of refractive index or boron
dopant concentration at three wavelengths: 630, 850 and 980nm. The
Q factor decreases as the refractive index decreases and the wavelength
increases. 146

Figure B-1 A finite element method simulation showing the distribution of the
optical field inside (left) and outside (right) of the microsphere cavity at
980nm. The 5nm-thick lipid bilayer is not visible at this scale. 151

Figure B-2 Images of (a) an optical micrograph of a silica microsphere and (b) a
fluorescent micrograph of a silica microsphere with a self assembled
lipid bilayer, conjugated to Texas Red Dye. 153

Figure B-3 Half of the lipids on the microsphere surface are accessible to a soluble
quencher. Epifluorescent images of a lipid-coated microsphere (a)
before and (b) after the addition of the soluble quencher QSY-7,
reproduced from the main text for the reader’s convenience. (c)
Quantitative measurement of lipid fluorescence before and after the
addition of quencher shows that about half of the lipids are accessible
to quenching, establishing a bilayer structure. 155

Figure B-4 Recovery of fluorescence in a photobleached region indicates that lipids
on the microsphere are mobile. Epifluorescent images of the lipid
bilayer (a) before bleaching, (b) during bleaching, (c) after bleaching
and (d) after 20 minutes of recover. (e) Quantitative measurement of
lipid fluorescence shows that the bilayer recovers 97% of the bleached
fluorescence, indicating that lipids can diffuse in the membrane. 157

Figure B-5 The measurements of the photon lifetime or Q factor, which indicates
the interaction strength between the whispering gallery mode and the
lipid bilayer. The quality factor was determined from the full-width
half maximum (FWHM, ) of the Lorentzian fit (dashed red line) to
the experimental data (solid black line). This measurement was
performed using an undoped lipid bilayer at a) 633 nm and b) 980 nm,
and using a Cy5 conjugated lipid bilayer at c) 633 nm and d) 980 nm. 159

Figure B-6 Schematic of test and measurement set-up. 161

Figure B-7 (a) The emission from the Cy5-conjugated lipid bilayer which is excited
by the optical microresonator. The peak maximum is located at 670.11
nm. The 633 nm excitation laser is blocked using a filter. (b) The
fluorescent decay of the Cy5 in the lipid bilayer which was excited by
the optical microresonator. The decay was fit to a double exponential. 162
xviii

Figure B-8 The spectra of the background was measured, both normalized and non-
normalized 163

Figure B-9 The spectra of the laser, with and without the filter, to verify the
efficacy of the 633nm filter. This pair of spectra was recorded by
imaging the evanescent field of the taper. Note that the y-axis is
plotted on a logarithmic scale. 163

Figure B-10 Comparison of the excitation/emission spectra measured by the
spectrograph, with and without the 633nm filter. This pair of spectra
further verifies that the 633nm filter has minimal impact on the
measurement of the Cy5 emission. The spectrum with the filter is
reproduced from the main text (Figure B-3a) for convenience. Note
that the y-axis is plotted on a logarithmic scale. 164

Figure C-1 (a) Artistic rendering and (b) optical image of the gold coated hybrid
devices. Gold nanoparticles suspended in a PMMA solution are coated
onto the toroid surface. The major diameter for the microtoroid is
approximately 50µm. The gold nanoparticles are too small to visualize
in this optical image. 172

Figure C-2 FEM simulations of the optical field distribution. a) The normalized
radial optical field intensity as a function of radius for silica (solid
black line) and for 10% PMMA-nanoparticle nanocomposite film (red
dashed line). The zero point indicates the center of the minor diameter.
This graph was determined from the optical field distribution for (b)
silica microtoroid and (c) nanoparticle-polymer coated microtoroid
with a 30nm thick film. The device size is 50(5) m major (minor)
diameter and operating wavelength is 635nm. 174

Figure C-3 Schematic of synthesis for thiol-stablized gold nanoparticles. First a
gold hydrosol is synthesized. Then, to replace the –OH on the surface
with –SH, a small amount of 1-dodecanethiol is added, and the gold
was transferred completely to toluene. 175

Figure C-4 Gold concentrations for the solutions with the volume ratio at 0:100,
5:100, 10:100 and 20:100. In other words, gold solutions volume
percentages are 0, 4.76, 9.09, and 16.67, respectively. It shows
concentration changes almost linearly with gold volume percentage.
Results are taken by the inductively coupled plasma-atomic emission
spectrometry. 176

Figure C-5 UV-Vis spectra of gold nanoparticles stabilized by (black) –OH group,
(blue) –SH group without PMMA, and (red) –SH group with PMMA. 180
xix

Figure C-6 Scanning electron micrograph image of an ultra high Q silica
microtoroid resonant cavity. 181

Figure C-7 Transmission spectra of microtoroid coated by spinning PMMA
solution doped with 0.0224Mol/L gold nanoparticles. Although the
data was normalized, some of the data fell above 1.0 due to noise in the
measurement. 182

Figure C-8 Experimental (solid squares) and theoretical (hollow circles) quality
factor of hybrid devices as a function of gold concentrations in PMMA
film at a) 633nm, b) 765nm and c) 980nm. The results were fit
according to the equation of y=ax
b
. 184

xx

Abstract

The optical characterization of polymeric materials to understand their
fundamental behavior is important to study their materials’ properties and to utilize them
in numerous applications. Therefore, this thesis mainly investigates the possibilities for
the development of hybrid organic-inorganic optical resonators for studying polymer thin
films based on whispering-gallery mode (WGM) optical resonators. The long-period light
confinement in WGM optical resonators with low loss materials, such as silica, ensures
high sensitivity or high quality (Q) factor by scanning optical resonators numerous times,
hence extracting the change of information caused in the system during the scan. Such
detection mechanisms can be the change of quality factor, resonant shift or transmission
change.
In this thesis, it is first demonstrated that the loss mechanism for hybrid optical
resonators are material-limited with quality factor over than 10
7
, meaning that other loss
mechanisms such as scattering, radiation, contamination and coupling losses are
minimized with optimized device performance, device fabrication and optical
characterization set-up. The change of optical field characterization is also investigated
with different polymers, film thickness and operating wavelengths by Finite Element
Method (FEM) simulations for the hybrid structure.

1

Chapter 1 Introduction

1.1 Motivation
As a result of their long photon lifetime which results in high circulating optical
fields, whispering gallery mode optical microcavities have been used in numerous
applications ranging from fundamental physic investigations to telecommunications[1, 2].
One area of research which has emerged over the past decade is the application of
microcavities as biosensors. In this field, they have demonstrated exquisite sensitivity to
small perturbations, such as the attachment of a single molecule or nanoparticle or protein
behavior and kinetics. However, there has been limited research in a similar field:
material properties and behavior.
Optical resonators are an ideal platform for studying material’s properties due to
their high sensitivity and their ability to perform experiments in real-time. While there are
many different geometries of microcavity, in the present thesis, the silica microtoroid
resonator is used due to its planar structure and on-chip fabrication. This geometry
enables the deposition of materials using several different methods. More importantly,
the properties of the silica are well known, which allows a more accurate determination
of the behavior of the material under study.
Organic/inorganic hybrid whispering-gallery-mode (WGM) optical
microresonators offer unique opportunity to investigate and characterize for the coated
organic materials, ranging from simple polymer to nanocomposite materials, on inorganic
silica microresonators. The interaction of the optical field with the coated materials
2

makes it possible to tailor the optical properties of the microresonators. For example, the
fraction of optical field, mode profile and mode area can be modified by coating various
different materials or changing film thicknesses, thereby controlling the resonant
frequencies and increasing/decreasing the optical field interaction with the coated
materials. Combining the results of the Finite Element Method (FEM) simulations and
optical resonator characterization testing set up, the properties of this system can be
determined by accounting for both the silica and the coated material’s properties.
However, several experiments should be performed to evaluate whether this approach
would be possible to characterize the coated materials by using the silica optical
microresonators. Especially, the loss mechanisms for these hybrid structures should be
investigated and understood, so this thesis starts from this experimental and theoretical
investigation with well-known polymers: polymethylmethacrylate (PMMA) and
polystyrene (PS). After understanding the loss mechanisms for this system, more
complicated material properties have been investigated as well as unknown optical
properties of the polyisobutylene (PIB) and nanocomposite materials.

1.2 Chapter overview
The organization of this thesis is as follows:
Chapter 2 gives the background about optical microtoroidal resonators and
whispering gallery mode optical cavities in general. The resonant characteristic for the
optical resonators, such as the loss mechanisms and mode volume are discussed. The
detailed fabrication procedures to make the microtoroidal optical resonators are
3

introduced in detail as well as the importance of each fabrication step. Testing
procedures, such as taper pulling process, alignment and data acquisition, are explained.
Chapter 3 explains how to simulate the optical resonators based on Oxborrow’s
model. It shows how to determine the optical field distribution, radial intensity profile,
resonant frequency, different modes and how modify his model to a more complicated
structure and to different geometries, such as spheres and disks.
Chapter 4 investigates the effect of thin film polymer coatings on silica
microtoroidal optical resonators. The theoretical calculations are initially performed to
determine which parameters among major, minor diameters and film thicknesses have the
most dominate affect on the optical field distribution in these hybrid structure. As a
result, it shows that the change of the film thickness is the most effective way to control
optical profile in these structures. Then, experiments verify theoretical calculations by
coating the resonator with two different polymers: polydimethylmethacrylate (PMMA)
and polystyrene (PS).
Chapter 5 demonstrates a thermally stable optical device by balancing the thermo-
optic coefficient of silica and a coated polymer layer. The studied polymers were PMMA
and PS and the temperature-insensitive devices were successfully fabricated by
optimizing the optical field overlap between silica and PMMA or PS.
Chapter 6 shows the demonstration of a method which is able to detect
temperature-induced changes in the refractive index of polystyrene polymer thin films as
small as 10
-7
. For this research, we made new resonator sample holder to integrate a
heater and thermocouple sensor directly. Also, a custom LabView program is developed
4

to record the resonant shift automatically during data acquisition. Both theoretical
calculations and experiments were performed and showed good agreement.
Chapter 7 shows the possibility of the characterization of the optical properties for
polymeric materials, such as transmission loss and thermo-optic coefficient by using
optical resonators. Due to the limitation of the conventional techniques, the optical
properties of the polyisobutylene (PIB) have been unknown as this polymer is viscous at
room temperature. Two different types of the deposition methods, such as spin-coating or
surface initiated cationic polymerization, were used to investigate and compare the
optical properties of the PIB. The results show significantly different optical properties
depending on the deposition methods.
Chapter 8 shows the study of interface effects by measuring the glass transition
temperature (T
g
) of ultra thin polystyrene films. It is suggested that it may be possible to
selectively interrogate different regions of the film by injecting different operating
wavelengths in the cavity. The preliminary results show that the different T
g
values were
detected when using different operating wavelengths at same polystyrene film thickness,
showing the possibility of the interrogating different regions of the regions in the polymer
film. This research is still under investigation.
Appendix A presents the results of a collaboration with Xiaomin Zhang which
investigated the reason for the performance disparity between microspheres and
microtoroids. It is revealed that the different types of silicon substrate affect the device
performance of the microtoroids. Unlike the microsphere, the microtoroid is fabricated
from thermal oxide not silica fiber, so the dopants diffuse the silicon into the oxide. As a
5

result, it modifies the refractive index of the materials and induces the scattering losses in
the cavity.
Appendix B presents the results for a collaboration with Lindsay Freeman and
members of Prof. Noah Malmstadt’s group using lipid bilayer coated microsphere s. My
contribution focused on simulating the device performance, or the optical field interaction
with a lipid bilayer coated onto the surface of an optical microsphere. In this research, a
method for self assembling the lipid bilayers onto the silica microspheres is developed
and the optical properties of the lipid bilayer were studied by various optical
characterization techniques, such as fluorescence microscopy, resonator testing and
spectrograph.
Appendix C presents the results of a collaboration with Ce Shi which studied
polymer-gold nanoparticle thin films using optical resonators. In this research, we show
that it is possible to tune or tailor the optical properties of this system by changing the
concentration of the thiol-functionalized gold nanoparticle embedded in a PMMA thin
film. The results show that there is strong optical field interaction with the gold
nanoparticles confirmed by the Q decrease with the increase of the nanoparticle
concentration.

6

Chapter 1 References
1. Chremmos, I., Schwelb, O., and Uzunoglu, N., Springer Series in Optical
Sciences, ed. W.T. Rhodes. Vol. 156. 2010, New York: Springer.

2. Vahala, K.J., Optical microcavities. Nature, 2003. 424(6950): p. 839-846.

7

Chapter 2 Overview of Optical Microtoroidal Resonators

2.1 Introduction

Whispering-gallery-mode microresonators are one type of optical resonator that
confines light at the rim of the device by total internal reflection. As a result of their
unique properties and compatibility with conventional silicon processing, these devices
have been intensively studied for various applications such as lasing [1, 2], biosensing [3,
4] and nonlinear optics [5, 6].
The term “whispering gallery mode” was inspired by the Whispering Gallery in
St. Paul’s Cathedral in London (Figure 2 -1a). In this gallery, acoustic waves propagate
around the interior, allowing a whisper at one side of the gallery to be heard at the
opposite side. The same propagation effect is observed in whispering gallery mode
optical resonators, as can be seen in Figure 2-1b. Light is coupled into the cavity when a
waveguide is adjacent to the optical resonator, and appropriate conditions are met. To
first order, the light is confined in the device by total internal reflection and relies on a
refractive index contrast between the resonator and its surrounding environment.
Additionally, unlike in a waveguide which can confine any wavelength, a resonant cavity
can only confine specific wavelengths which are an integral multiple of the device
circumference (Figure 2-2). These wavelengths, also known as the resonant wavelengths
of the cavity, are dependent on the device geometry, material and environment.
8

Figure 2-1 (a) St. Paul’s Cathedral Whispering Gallery and (b) Microtoroidal optical resonators with the optical fiber
waveguide showing the light propagation (highlighted as red)

Figure 2-2 The geometric optics showing that the resonance occurs when the light propagates (a) integral fashion and
(b) the other cases it will be off resonance.

These devices can be made in numerous geometries and from many different
materials. However, because the optical field is most efficiently confined in circular
orbits, the most popular cavities have a circular feature. Figure 2-3 gives a brief
overview of some of these devices. All of the devices in the chart, except for the silica
microsphere, were only invented in the past decade or so, and strongly depend on the
numerous recent advances in micro/nanofabrication technology (material deposition, e-
beam lithography, etching methods, etc). For example, the microtoroid was invented in
9

2002 and the SiN microdisk was invented in 2004. The microsphere was originally
developed in the 1970’s for other applications, and began to be frequently used for
optical applications in the 1980’s; several decades before the other devices in the chart.
Yet, as will be discussed in the following section, it is still one of the highest performing
devices to date.

Figure 2-3 Various types of the WGM optical micro-resonators based on different geometries

The present thesis revolves around the silica microtoroid and microsphere. While
they have similar device performance, the microsphere is not fabricated on a silicon
wafer, making it hard for handling and testing in most applications. From this
perspective, Microdisks and microrings can be also fabricated on-chip, and would seem
to be suitable for much of the work in the present thesis. However, their device
10

performance is significantly worse than the microtoroid, as will be discussed in Section
2.2. Therefore, by considering both device performance and future transition to practical
applications, the choice of the microtoroids seems wise choice.

2.2 Background
2.2.1 Quality (Q) factor
The performance of the optical resonators is quantitatively characterized by the
photon lifetime or quality factor of the device.[7-9] A device with a longer photon life
time has a higher quality (Q) factor. In many applications, this translates to higher device
performance or sensitivity. This metric is governed by a series of loss mechanisms, both
intrinsic and extrinsic to the sensor. These losses can be summarized by the following
expression:

(2.1)
where Q
rad
is the radiation loss, Q
ss
is the surface scattering loss, Q
mat
is the
material loss, Q
cont
is the contamination loss, and Q
coupl
is the coupling loss. The first
four are intrinsic to the cavity (Q
int
), and the fifth is extrinsic to the cavity (Q
ext
).
Unlike in the case of the microsphere, it is not possible to analytically calculate all
terms for the microtoroid, and some must be completed solved via simulations.
However, for most cases, the primary loss mechanisms are either Q
mat
or Q
ss
, and it is
possible to calculate those terms. Q
mat
can be expressed by: [10]



(2.2)
11

where n
eff
is the effective refractive index,  is the wavelength, 
eff
is the effective
material loss. n
eff
=n
silica
+n
air
and 
eff
=
silica
+
air
, where β and δ represent the
percentage of the optical field in silica and air, respectively. Therefore, the sum of β and δ
should be 1.
In 2000, it was shown that Q
ss
is actually comprised of two terms; one represents
surface scattering and one represents internal scattering. Q
ss
and Q
is
can be expressed by:
[10]



(2.3)
where K defines the internal reflection condition, and B are the surface roughness of
the cavity, n
eff
is the effective refractive index,  is the wavelength, and R is the radius of
the cavity.





(2.4)
where T, p,  and  are the temperature of melting, the Pockels coefficient, Boltzman
constant, and isothermic compressibility at room temperature.
Q
is
, Q
mat
and Q
ss
can be calculated by using previously determined values of , B,
n
eff
, p, T,  and , along with the experimental parameters which define K and λ [8, 10].
The thermal oxide loss was taken from previous publication with an assumption that it
would be same as silica [8, 11, 12]. Using these expressions, it is possible to calculate
these terms for the microsphere and the microtoroid.[8, 10, 11].
As may be expected, there are huge differences in the intrinsic quality (Q) factor
depending on the geometries, material’s p roperties and fabrication procedure. The
12

microsphere and toroid have the highest quality factor due to their atomically smooth
surface formed by a laser reflow process and the very low loss of silica. Microdisks
usually have lower quality (Q) factors than microspheres and toroids due to the surface
roughness caused by the fabrication procedure which involves lithography; however, the
control of the device size is somewhat easier. Microrings usually have lowest quality (Q)
factor among the resonant cavities as a result of the high material loss.

2.2.2 Mode Volume
Unlike quality factor describing photon life time in optical resonators, mode
volume defines spatial confinement in the optical resonators. Smaller mode volume
means denser optical confinement in the optical resonators, hence enabling to make lower
threshold microlasers or stronger non-linear effects, such as cavity electrodynamics.
Commonly, the definition of the mode volume is based on the energy density of the
optical modes existing in the optical resonators and it is normally described as the
equation below:

(2.5)
Where V
Mode
, and

represent the mode volume, the permittivity and the
electric field, respectively. More detailed description about the mode volume can be
found elsewhere.[13, 14]

13

2.3 Microtoroid Fabrication
The fabrication process was based on a previously published three step method
summarized here (Figure 2-4):[15] 1) First circular oxide pads are photolithographically
defined and etched using buffered oxide etchant (buffered HF) onto a silicon wafer with
two microns of thermal oxide (silica or SiO
2
). The wafer is rinsed and diced into small
samples (~2mm x ~5mm), 2) Second, the samples are placed into a xenon difluoride
(XeF
2
) etcher. XeF
2
is an isotropic gas phase etchant which is selective for silicon.[16]
Therefore, it does not attack the silica. This second etching process undercuts the silica,
resulting in a silica microdisk on a silicon pillar structure, and 3) Finally, a carbon
dioxide (CO
2
) laser is used to reflow or melt the microdisk, forming a microtoroid. The
wavelength of the CO
2
laser (10.6m) is absorbed by the silica while it is transparent to
silicon.

Figure 2-4 Overview of the fabrication process. a) Circular oxide pads are photolithographically defined on a silicon
wafer, b) XeF
2
undercuts the oxide, forming a microdisk, c) a CO
2
laser reflows the oxide, forming a microtoroid. a)
and b) are redenerings and c) is a scanning electron micrograph.

2.3.1 Define micropad by patterning oxide layer
The silicon wafer with 2μm oxide layer was cleaved to a size of 1”×1” to
minimize the effect of edge-bead with diamond scribe, and then the sample was cleaned
with acetone, methanol and isopropyl-alcohol and dried with N
2
gas. To dry completely,
14

the sample was placed on hot plate at 120 °C for 2 minutes. For the next step, sample was
placed in hexamethyldisilazane (HMDS) exposure set up at least 2 minute to ensure the
adhesion between oxide layer and photoresist. This is widely used process known as
silylation in the semiconductor industry, forming a strong bond to the surface. The other
free bonds readily react with the photoresist.
S1813 photoresist (Shipley) were coated with 3000rpm of spin speed for 1minute
using a spin coater, and then the sample was baked on hot plate for 2 minutes at 95°C to
cure the photoresist and to evaporate the remaining solvents. The sample was exposed
with a circular pattern mask with intensity of 4mJ/cm
2
for 20secs with Karl Suss MJB 3
photomask alinger. With MF-321 (Shipley) developer, the sample were define with clear
circular pattern and wash with DI water followed by N
2
gas drying. After checking the
circular pattern under optical microscopy, the sample was baked on hot plate for 2 minute
at 110°C to reflow the photoresist for protecting the underlying oxide layer.
With improved buffered oxide etchant (BOE), the oxide layer was etched for 19
minutes. To ensure the oxide layer was etched completely, the sample was monitored
carefully and checked with optical microscopy. The sample was washed with acetone,
methanol and isopropyl-alcohol to remove the residue of photoresist and dried with
nitrogen (N
2
) gas. The photoresist should be removed completely during this clean step as
the remaining photoresist will greatly affect the final device performance. Therefore, this
final cleaning process is repeated, and the samples are washed at least twice by holding
different places by a tweezer.

15

2.3.2 Fabricate microdisk using the xenon difluoride (XeF
2
) etcher
The sample was diced to desirable size, and then placed into the chamber of the
xenon difluoride etcher to selectively etch silicon. The xenon difluoride is in a solid
crystalline form at room temperature, but it becomes a gas by sublimation at its vapor
pressure (~3.8 Torr) at room temperature. Due to the possible reaction between XeF
2
and
water, it is important to purge the etching system to remove any moisture and other
contaminants out of the etching chamber before etching process. Otherwise, the xenon
difluoride will react with water, and then will produce HF, which reacts with silicon
dioxide (SiO
2
). This creates defects on the silica microdisk, resulting in defects in the
final microtoroid structure. During this purge process, the surface of the loaded samples
is also dehydrated to avoid forming a thin oxide layer, which significantly affects the
etching rate. Even after the etching process, it is important to do purge process to remove
SiF
4
byproduct and any residual gaseous HF, which forms during the etching. The
reaction mechanism of the xenon difluoride and silicon is described by 2XeF
2
+ Si →
2Xe (g) + SiF
4
(g).
The xenon difluoride is highly selective to silicon and it is isotropic with fast
etching rate. However, it is important to note that the etching rate depends on how many
samples and sizes of the samples loaded in the etcher. The major drawback to use XeF
2

etcher is that there is no well defined etching stops by crystal plane. Therefore, the etched
sample was observed under optical microscopy, then repeated the process until we get
desirable undercut. Figure 2-5 shows the system of XeF
2
that we use for fabricating
16

microdisk resonators. As can be seen, it is consists of the etching chamber, pulsing
chamber, pump, nitrogen gas and XeF
2
chamber.

Figure 2-5 (a) The system for top portion of XeF
2
etcher (b) control panel on computer screen

Chamber
locker
Etching
chamber
17

2.3.3 Fabricate microtoroid using a carbon dioxide (CO
2
) laser reflow
For final step to make silica microtoroid, a carbon dioxide (CO
2
) is used to reflow
or melt the microdisk, forming a microtoroid. The wavelength of the CO
2
laser (10.6μm)
is absorbed by the silica while it is transparent to silicon, which acts as a heat sink and an
important role to define the size of the microtoroid.
For reflow process with CO
2
laser, it is important to position the microdisk in the
center of laser spot, so we use thin glass slide for confirmation of the laser beam which
follows Gaussian distribution. The spot is visible in the camera in the set-up. Then, we
place the sample on the sample holder, and then increase laser power slowly until the
microdisk turns to microtoroidal shape. The optimum power of the CO
2
laser changes
depending on the size of microdisk and silicon pillar. The formation of the microtoroid
from the microdisk can be explained by surface tension with the strain between oxide
laser and silicon wafer during thermal growth of oxide layer. The toroid size is dependent
on the support pillar formed during XeF
2
etching with certain size of microdisk. The laser
exposure last around 1 minute to ensure the surface smoothness of the microtoroid
resonator (possible microsize defects would smoothen during this exposure) and also to
dry out the possible water molecules deposited during storage. The CO
2
testing set-up is
shown in Figure 2-6.

18

Figure 2-6 A portion of CO
2
laser reflow set up

2.4 Device testing procedure
2.4.1 Pulling a tapered optical fiber
While it is important to make good optical microtoroid resonators for optical
device testing, it is also equally important to make low loss tapered optical fiber to
efficiently couple a light in/out of the device. The low loss taper allows the
approximation that Q
coupl
~ (or the coupling losses are zero). Tapered fibers have
demonstrated high efficiency/low-loss coupling to optical resonators [17]. As can be seen
in Figure 2-7, the taper pulling set-up is equipped with a camera to observe the pulling
process in real-time, a high precision pulling control box to control pulling speed and a
flowmeter to control the hydrogen gas.

Camera
ZnSe
lens
CO 2 laser
Beam combiner
Sample holder
19

Figure 2-7 (a) Taper pulling set-up and (b) magnified image of its core region

To pull the optical fiber, we first remove 1” length cladding layer of optical fiber
by a fiber stripper, and then place the optical fiber on the fiber holder. With around
5mL/min of flow rate, we start pulling the optical fiber slowly with real time view on
screen. When the desirable thickness is achieved, we stop the pulling process. To pull a
good tapered optical fiber, it is important to keep the optimum position between hydrogen
torch and optical fiber, proper flow rate and pulling speed. Additional instruments such as
oscilloscope and power meter can be used to monitor the transition between single mode
and multi modes. In Fig 2-8, a successfully pulled optical fiber is shown.
Lamp

Flowmeter

Controller box

Hydrogen
torch

Fiber holder

Camera

20

Figure 2-8 An example of pulled optical fiber. We can clearly see the rainbow color like distribution from optical fiber.
This is due to optical filtering effect.

2.4.2 Alignment of optical fiber with microtoroid resonator
The tapered optical fiber is carefully moved in the fiber holder to the cavity
testing set-up as shown in Figure 2-9. To accurately align the tapered optical fiber and the
resonant cavity, a two axis (top and side) machine vision system in combination with a
nm-resolution motorized stage is used. The top view camera is used to avoid crushing the
microtoroid resonator with the tapered optical fiber, and the side view camera is used to
align the optical fiber parallel to the side of the microtoroid resonator. An example of
aligned resonator with optical fiber is shown in Figure 2-10 which is ready for the testing.
21

Figure 2-9 The portion of cavity testing set-up

Figure 2-10 (a) Top and (b) side view of aligned resonator with optical fiber

2.4.3 Getting data
There are two methods of determining the Q of a resonant cavity. The first based
on measuring the linewidth () of the cavity by recording a resonance spectra. The
second is based on measuring the photon lifetime () of the cavity by performing a ring-
down measurement. A linewidth measurement is accurate as long as the linewidth of the
laser used to perform the measurement is narrower than the linewidth of the cavity. The
22

second method is accurate as long as the ringdown time (photon decay rate) of the cavity
is sufficient long to get a good fit (10ns, given a good oscilloscope, which is a Q~10
million). In our lab, our lasers have linewidths which are sufficiently narrow to be able to
measure Q factors as high as 500 million. Therefore, we use this method, as it is able to
measure Q factors over a larger range (1000 – 500 million).
To measure the linewidth, in our lab we use following main components: Tunable
lasers, NI oscilloscope/digitizer, NI function generator, optical attenuator, photodetectors,
attenuators, sample stages, single mode optical fiber spools and optical table. The
schematic diagram of testing set-up was shown in Figure 2-11.

Figure 2-11 A schematic diagram of the optical resonator’s testing set -up

For resonator characterization, a narrow-linewidth (300kHz) CW tunable lasers
are used. The laser is coupled to the optical devices using a single mode low-loss tapered
optical fiber (Newport). The optical input power is controlled by an attenuator which is
placed in-line between the laser and the optical resonators. A function generator is used
to control the scan speed and range to ensure that these parameters do not distort the
measurements. To determine the loaded quality factor (Q=/, λ=wavelength,
23

λ=linewidth), the resonance linewidth is recorded in the under -coupled regime, where
the coupling between the waveguide and optical resonators is weak, on a high speed NI
oscilloscope/digitizer using LabView software. Testing in the under-coupled regime
ensures minimal optical power transfer to the resonators, hence reducing non-linear
effects, such as thermo-optic effect.
For broad scan, the laser was scanned over the free spectra range (FSR) of the
resonant cavity. An example of broad scan spectra for silica microtoroid is shown in
Figure 2-12 (a). From the measured broad scan spectra, we can clearly see the
fundamental mode separated by the free spectra range of the cavity. The major diameter
of the silica microtoroidal optical resonator was around 64μm.
For fine scan, the scan range, scan speed and laser power was optimized (forward
and backward scans gave similar linewidth measurements). To avoid thermal effects
which mainly caused by thermal optic coefficient of the silica, the input power was
injected as low as possible by pulling out the connector of optical fiber or using an
attenuator. An example of fine scan spectra is shown in Figure 2-12 (b). When fit to a
lorentzian, the full-width-half-max (FWHM) of this spectra corresponds to the linewidth
of the cavity. The quality factor is calculated from the expression: Q
tot
=.
24

Figure 2-12 An example of (a) broad and (b) fine scan spectra for silica microresonators.

25

Chapter 2 References
1. Hsu, H.S., C. Cai, and A.M. Armani, Ultra-low-threshold Er:Yb sol-gel
microlaser on silicon. Optics Express, 2009. 17(25): p. 23265-23271.

2. Tulek, A., D. Akbulut, and M. Bayindir, Ultralow threshold laser action from
toroidal polymer microcavity. Applied Physics Letters, 2009. 94(20): p. 203302.

3. Vollmer, F. and S. Arnold, Whispering-gallery-mode biosensing: label-free
detection down to single molecules. Nature Methods, 2008. 5(7): p. 591-596.

4. Armani, A.M., et al., Label-free, single-molecule detection with optical
microcavities. Science, 2007. 317(5839): p. 783-787.

5. Kippenberg, T.J., et al., Ultralow-threshold microcavity Raman laser on a
microelectronic chip. Optics Letters, 2004. 29(11): p. 1224-1226.

6. Carmon, T. and K.J. Vahala, Visible continuous emission from a silica
microphotonic device by third-harmonic generation. Nature Physics, 2007. 3(6):
p. 430-435.

7. Gorodetsky, M.L., A.D. Pryamikov, and V.S. Ilchenko, Rayleigh scattering in
high-Q microspheres. J. Opt. Soc. Am. B, 2000. 17(6): p. 1051-1057.

8. Gorodetsky, M.L., A.A. Savchenkov, and V.S. Ilchenko, Ultimate Q of optical
microsphere resonators. Optics Letters, 1996. 21(7): p. 453-455.

9. Hunt, H.K. and A.M. Armani, Label-Free Biological and Chemical Sensors.
Nanoscale, 2010. 2(9): p. 1544-1559.

10. Gorodetsky, M.L., A.D. Pryamikov, and V.S. Ilchenko, Rayleigh scattering in
high-Q microspheres. Journal of the Optical Society of America B-Optical
Physics, 2000. 17(6): p. 1051-1057.

11. Vernooy, D.W., et al., High-Q measurements of fused-silica microspheres in the
near infrared. Optics Letters, 1998. 23(4): p. 247-249.

12. Hale, G.M. and M.R. Querry, OPTICAL-CONSTANTS OF WATER IN 200-NM
TO 200-MUM WAVELENGTH REGION. Applied Optics, 1973. 12(3): p. 555-
563.

13. Kippenberg, T.J., Nonlinear Optics in Ultra-high-Q Whispering-Gallery Optical
Microcavities. PhD thesis, California Institute of Technology, 2004.
26

14. Spillane, S.M., Fiber-coupled Ultra-high-Q Microresonators for Nonlinear and
Quantum Optics. 2004, PhD thesis, California Institute of Technology.

15. Armani, D.K., et al., Ultra-high-Q toroid microcavity on a chip. Nature, 2003.
421(6926): p. 925-928.

16. Kovacs, G.T.A., Micromachined Transducers Sourcebook. 1998, New York:
McGraw Hill. 911.

17. Knight, J.C., et al., Phase-matched excitation of whispering-gallery-mode
resonances by a fiber taper. Optics Letters, 1997. 22(15): p. 1129-1131.

27

Chapter 3 Finite Element Method (FEM)

3.1 Introduction
As it is very hard to express the mathematical expression for large domain, FEM
divides the large domain into a number of small and simple elements. Then, the each
element can be expressed by the governing equations and all those equations in each
domain can be solved. Each element is connected to adjacent elements by the equation
that needed to be solved to model physical systems. A wide variety of the physics based
systems from heat flow to structural mechanics, even with extremely complex
geometries, can be solved using FEM.
The models that I have been using to study the optical properties of the optical
resonators during my PhD were based on Mark Oxborrow’s simulations [1]. Unlike in a
spherical cavity where analytical expressions are available, the toroidal cavity must be
solved using a simulation (due to its inseparable wave equation). However, due to its
axisymmetric structure, the complexity can be reduced to a two-dimensional problem
with rotational symmetry.
In this chapter, I will just focus on how to use the Mark Oxborrow’s model to
study optical properties of various microresonators such as microspheres, microtoroids
and microdisks, as the detailed explanation about the models already exists in various
sources [1-3]. Before using Oxborrow’s model, it is important to read these sources to
understand the basics about his model based on the optical resonator theory.
28

3.2 Microtoroid simulations
In Oxborrow’s model, the material of the optical resonator is the silica, which
assumed to be perfect isotropic, with a refractive index of 1.4457 at around λ=852nm and
room temperature. Therefore, it is important to change the refractive index of the silica
under Options  Constants, if operating at different wavelengths, temperatures or
materials. The operating wavelengths can also be changed under this category by
controlling the azimuthal mode order number “M”. Approximate M can be found by
following expression for circular WGM optical microresonators:

(3.1)
; Where n is the refractive index of the material for optical resonator, R is radius
of the optical resonator and  is operating wavelength.
Then, one can change the geometry of the model to desired shape and size
through CAD tool in COMSOL Multiphysics.
After running the model, we can export the resulting image under File  Export
 Image as can be seen in Figure 3-1. The result shows the cross-sectional electric-field
intensity

defined in Postprocessing  Plot parameters. The results show the 115
th

fundamental TE mode controlled by azimuthal mode order “M”. The unit of the axes is
meter, so the simulated frequency should be multiplied by 10
6
as we often deal with m-
size microtoroidal optical resonators. The unit in the geometry can be changed to
microns, but the results should be same. In this result, the resonant frequency is 1.929516
29

 10
14
Hz which is corresponding to λ=1553.718nm with a 40(3) μm major (minor)
diameter silica microtoroidal optical resonators surrounding by free space.

Figure 3-1 A FEM result of the cross-sectional electric field intensity (

) profile of the 40(3) μm major(minor)
diameter of the silica microtoroidal optical resonators, with 115
th
TE polarized mode.

The corresponding radial field distribution as a function of the radius also can be
determined as can be seen Figure 3-2 under Postprocessing  Line/Extrusion. The graph
was generated same condition as Figure 3-1. The n, l and m represent the radial, angular
and azimuthal mode numbers, respectively.
30

Figure 3-2 A FEM result showing the normalized radial intensity VS radius for the 40(3) μm major(minor) diameter of
the silica microtoroidal optical resonators, with 115
th
TE polarized mode.

The relationship between the frequency and wavelength is shown in Table 3-1.
The major diameter can be controlled by moving whole objects to left or right side and
the minor diameter can be controlled by increasing or decreasing the size. The side
colorbar shows the corresponding optical field intensity as shown in resulting image,
hence making it easier to find the maximum or minimum values of the field intensity.
Frequency (Hz10
14
) Wavelength (nm)
4.736057 633
3.918855 765
3.059106 980
2.254078 1330
1.934144 1550

Table 3- 1 The relationship between the frequency and wavelength

31

The model can also be used to find various modes existing in the optical
resonators as shown in Figure 3-3 by setting desired number of eigenvalues under Solve
 Solver parameters.

Figure 3-3 A FEM result showing the various modes existing in the 40(3) μm major(minor) diameter of the silica
microtoroidal optical resonators, with 115
th
TE polarized mode.
32

Various other quantities such as mode area, mode volume and radiation loss can
also be determined from the model. More detailed descriptions are described elsewhere.
The Oxborrow’s model can be modified to more advanced and complicated
systems such as hybrid optical resonator’s structure. A 100nm thick poly(methyl
methacrylate) (PMMA) coated hybrid structure is shown in Figure 3-4 as an example. For
this model, the additional geometry was created to express PMMA coating and the
domain’s physics was defined under Physics  subdomain settings same as silica
domain, assuming the perfect isotropic material. Also, the optical properties such as
refractive index and azimuthal mode order number changed under Options  Constants.

Figure 3-4 A FEM result of the cross-sectional electric field intensity (

) profile for a 100nm thick PMMA coating
on the 40(8) μm major(minor) diameter of the silica microtoroidal optical resonators, operating wavelength at 980nm.

Various interesting theoretical calculations can be done with this model. For
example, the resonant shift as a function of PMMA coating thickness at different
operating wavelengths could be determined by keeping the M as constant, or the resonant
shift as a function of temperature at constant coating thickness.
33

3.3 Microsphere simulations
The optical mode profile of the microspheres can also be simulated by modifying
the geometry. Every condition such as materials and azimuthal mode order were kept the
same. As a result, the found frequency is 2.034526  10
14
Hz which is corresponding to
λ=1473.525nm with a 40μm diameter silica microsphere optical resonators surrounding
by air environment. As can be seen in Figure 3-5, the microsphere has a reduced axial
confinement compared to the microtoroid shown in Figure 3-1.

Figure 3-5 FEM results of the cross-sectional electric field intensity (

) profile of the 40μm diameter of the silica
microsphere optical resonators, with 115
th
TE polarized mode.

34

3.4 Microdisk simulations
The optical mode profile of the microdisks can also be simulated by modifying
the geometry. Every condition such as materials and azimuthal mode order were kept the
same. As a result, the found frequency is 2.077737  10
14
Hz which is corresponding to
λ=1442.88nm with a 40μm diameter sili ca microdisk optical resonators surrounding by
air environment. The angle of the microdisk’s edge was assumed to be 45°. The different
angle also can be simulated by simply changing the geometry.

Figure 3-6 FEM results of the cross-sectional electric field intensity (

) profile of the 40μm diameter of the silica
microdisk optical resonators, with 115
th
TE polarized mode.

35

3.5 Conclusion
In summary, it has been shown that how to simulate optical microresonators such
as microtoroids, microspheres and microdisks based on the Oxborrow’s model. Starting
from his model, more advanced and complicated also can be simulated nicely as shown
above. The simulation methods shown here are basics for following chapters.

36

Chapter 3 References
1. Oxborrow, M., Traceable 2-D finite-element simulation of the whispering-gallery
modes of axisymmetric electromagnetic resonators. IEEE Transactions on
Microwave Theory and Techniques, 2007. 55(6): p. 1209-1218.

2. Kippenberg, T.J., Nonlinear Optics in Ultra-high-Q Whispering-Gallery Optical
Microcavities. 2004, PhD thesis, California Institute of Technology.

3. Spillane, S.M., Fiber-coupled Ultra-high-Q Microresonators for Nonlinear and
Quantum Optics. 2004, PhD thesis, California Institute of Technology.

37

Chapter 4 Hybrid Silica-Polymer Ultra-High-Q
Microresonators

4.1 Introduction

While silica microcavities have numerous applications, by integrating optically
active elements, such as gain media or nonlinear coating, additional areas of research
become possible. However, it is very difficult to develop this type of optically active
device while maintaining the high performance (high Q) of the passive device. One
approach is to coat the microcavity with a polymer coating, which can act as a host
material. In the present work, we study the effect of thin film polymer coatings on ultra-
high Q (UHQ) silica microtoroid optical microcavities.[1] In the initial work which
developed the theory for all subsequent studies, two different polymers were used:
polydimethylmethacrylate (PMMA) and polystyrene (PS).

4.2 Fabrication
In addition to silica microtroid fabrication procedure described in chapter 2, an
additional spincoating step is necessary to make silica (inorganic) / polymer (organic)
hybrid microtoroid resonators. Two different polymers were used throughout this study,
PMMA (Sigma Aldrich, 15k MW) and PS (Sigma Aldrich, 200k MW). To enable spin-
coating deposition, the polymer was dissolved in ultra-high purity toluene (J.T. Baker,
99.5% min) at 0.5, 2, and 4% for PMMA and 0.5, 1, 2% for PS. After spin-coating, the
film is thermally reflowed for 30 minutes in a gravity oven at 115°C (PMMA) and 105°C
38

(Polystyrene). This thermal reflow allows the hybrid microtoroids to maintain the smooth
surface of the silica devices. Film thickness measurements for PMMA and for
Polystyrene were performed by ellipsometry on bare silicon wafers and compared to the
value in a reference [2]. For the case of PMMA, the film thickness as a function of spin
speed is shown in Figure 4-1. Also, the fabricated devices for each step are shown in
Figure 4-2.

Figure 4-1 PMMA film thickness as a function of spin speed

39

Figure 4-2 Optical images shown fabrication procedure. Silica (a) Micropad, (b) Microdisk, (c) Microtoroid and (d)
100nm thick PMMA coated microtoroid on Silicon wafer. The bluish color on the toroidal rim is due to the polymer
coating with this coating thickness.

4.3 Theory
4.3.1 Loss mechanisms for hybrid devices
In the present hybrid system, it is proposed that material absorption is the
dominant loss mechanism, yielding Q
0
~ Q
mat
. Under this condition, the quality factor of
the device is described by


(4.1)
40













(4.2)
where β, γ and δ represent the percentage of the optical field in silica, polymer, and air,
respectively. Therefore, the sum of β , γ and δ should be 1. The refractive indices of the
PMMA and PS are verified using ellipsometry. The results agree well with commonly
referenced values [3].
4.3.2 Material absorption loss for PMMA and PS
The material loss of the PMMA and PS are measured by Cary 14 UV-Vis
spectrophotometry with different concentrations. Then, by using Beer-Lambert law as
can be seen equation 4.3, we can calculate the absorption coefficient.
(4.3)
where ε is the molar absorptivity, l is path length over which the measurement is taken, c
is the concentration of the species (PMMA or PS), and α is the absorption coefficient of
the material. A is the absorption determined by the spectrophotometer.
The measured material loss for PMMA and PS is in Figure 4-3, and the material
loss of the silica is used from previously published reference [4].
41

Figure 4-3 Material absorption loss measured by Cary 14 UV-Vis spectrophotometry for (a) PMMA and (b) PS.

4.3.3 Determination of β, γ and δ
The final parameter, which should be known for the theoretical Q calculation, is
the fraction of the optical field in each region. We used COMSOL Multiphysics finite
element method and revisited Oxborrow’s simulation result [5]. This model has been
already verified by numerous scientists in the field. The simulation was modified for our
purpose to study field distribution in the layer of silica, polymer and air. The “operating
wavelength” can be set by controlling mode number (M), where M is azimuthal mode
order in the cavity. The two dimensional axial symmetry is chosen for the simulation
which is appropriate for microtoroid resonators. This choice is appropriate to calculate
the percentage of optical field in each region and significantly saves time compared to
three dimensional simulations.
42

The optical field profile of the cross section of a toroid ring is shown as an
example without and with 500nm thickness of PMMA at λ=980nm in Figure 4-4. As
expected the optical field is shifted toward the polymer layer due to higher refractive
index of PMMA than SiO
2
. In addition, the mode shape changed.

(a) (b)
Figure 4-4 Finite simulation result which shows optical profile (a) without and (b) with 500nm thickness of PMMA at
λ=980nm. It is cross section of one side toroidal ring.

The percentage of the optical field in each layer was determined by Power
In

(Silica, Polymer, Air)/Power
Tot
, where Power
In
and Power
Tot
represent the portion of the
optical power in each region and total optical power in the cavity and its surrounding
environment, respectively.
There are three experimentally determined variables: 1) major diameter (D), 2)
minor diameter (d), and 3) polymer film thickness. Before performing experiments, all
three parameters were systematically varied in the COMSOL simulations to determine
which variable had the largest impact on the Q.
43

Specifically, we first simulated a 500nm film thickness of PMMA, and changed
both the major and minor diameters at λ =980nm. As shown in Figure 4-5 (a), there is less
percent change by varying the minor diameter than the major diameter, as expected. For
the next, we simulated with toroid with D=40m and changed film thicknesses and minor
diameter to determine which variable has the most significant impact. As can be seen in
Figure 4-5 (b), we concluded that it is more effective to change the film. From these
simulations, we can easily anticipate that a higher percentage of the field will be in the
polymer film at lower wavelengths, thicker film thicknesses and smaller devices.

Figure 4-5 Finite simulation results. Percentage of the optical field in the polymer layer at λ=980nm as a function of the
minor diameter as (a) the major diameter increases from 40 to 100μm. The film thickness is fixed at 500nm. (b) the
film thickness change from 100 to 500nm. The major diameter is fixed at 40μm

From simulation results above, we decided that the most effective way to study
for this research is the change of film thickness by keeping major and minor diameter
consistently. The simulation results are summarized for silica layer in Figure 4-6 and
PMMA layer in Figure 4-7.
(a) (b)
44

Figure 4-6 Finite element method results. (a) Percentage of the optical field in the silica layer as a function of the
minor diameter as the PMMA film thickness increases from 0 to 500nm at (a) λ=850nm and (b) λ=980nm . The major
diameter of the device is fixed at 40μm.

Figure 4-7 Finite element method results. (a) Percentage of the optical field in the PMMA layer as a function of the
minor diameter as the PMMA film thickness increases from 100 to 500nm at (a) λ=850nm and (b) λ=980nm . The
major diameter of the device is fixed at 40μm.

From the results of PMMA case, the same toroid geometry was used for
polystyrene (PS) coated hybrid devices, and the optical field was modeled with (20, 50,
100, 200) nm thick films for the polystyrene (PS) case.
(a) (b)
(a)
(b)
45

Figure 4-8 shows FEM results for a 40(8)m major(minor) diameter microtoroid
with an SEM image of fabricated silica microtoroid resonatorr. As the Polystyrene (PS)
film thickens, the optical field shifts from the silica into the polymer film (Figure 4-8
b,c,d) as stated previously for PMMA case.

Figure 4-8 (a) Scanning electron micrograph of a toroidal microresonator. Finite element method simulation results for
the optical field intensity distribution: (b) silica microtoroid, (c) hybrid microtoroid with a 100nm thick PS film, and
(d) 200nm thick PS film. Note that the optical field shifts from the silica towards the polymer film as the thickness of
the polymer film increases. The major (minor) diameter is 40(8)m and λ =850nm.

46

In Fig 4-9, we can see summary of simulation results for PMMA and PS which
shows optical field in polymer as a function of film thickness with D(d)=40(8)m at
λ=850nm, 980nm. By using these percentages of the optical field in each region, we can
calculate Q
mat
from equation (1) and the results are summarized in Fig 4-10. It should be
noted that theoretical Q
mat
is highly dependent on the percent of optical field in polymer
layer due to its higher optical absorption loss compared to silica layer.

Figure 4-9 Finite element method results. Percentage of the optical field as a function of the polymer film thickness for
PMMA and PS at λ=850, 980nm.

47

Figure 4-10 Summary of FEM simulation result for PMMA (a) λ=850 (b) λ=980 and for PS (c) λ=850 (d) λ=980.

4.4 Experimental results and discussions
The Q was measured using the linewidth method. An example of broad and fine
scan spectra with a hybrid device is shown in Fig 4-11 with Lorentz-fit for the fine scan.
It should be noted that the Q of every device was taken twice: before and after the
polymer film was applied. Only if Q is over than theoretical value for silica microtoroid,
then the silica microtoroid is coated with polymer layer to measure Q of hybrid device.
48

All Q factors were measured with very low input powers to make sure that no non-linear
effects distorted the Q measurement.

Figure 4-11 An example of (a) broad and (b) fine scan spectra for hybrid microresonators.

The Q of the hybrid devices was measured with four different film thicknesses,
including without a polymer coating film, for PMMA and for PS and at two different
wavelengths (850nm, 980nm). The modeling results and the experimental data are
presented in figure 4-12. It should be noted that Q factors above 10
7
were measured for
hybrid devices fabricated using either PS or PMMA. From our best knowledge, these Q
factors are the highest to date for a hybrid device. The theoretical and experimental data
was fit to an equation of the form y=ax
b
, which is the appropriate form for Q
mat
. As seen
in Table 4-1, there is very good agreement between the values for (a, b) predicted by the
model and the experimental results. Therefore, the Q factor of the hybrid device is
material loss limited.
49

Figure 4-12 Experimental and theoretical Quality factor (Q) as a function of polymer thickness. PMMA at (a) 850nm
and (b) 980nm. PS at (c) 850nm and (d) at 980nm. The results were fit to an equation of the form y=ax
b
, which is
included as a solid (dashed) red line for the theoretical (experimental) Q results. The black dotted line indicates the
highest Q demonstrated with a silica toroidal resonant cavity to date, setting an upper bound on Q [6].

50

Model Experiment
Polymer
Wavelength
(nm) a (x10
9
) b a (x10
9
) b
PS
PS
PMMA
PMMA
850
980
850
980
5.24
4.76
3.11
1.37
-1.49
-1.42
-1.32
-1.19
4.96
4.02
1.52
0.398
-1.53
-1.47
-1.21
-1.06

Table 4- 1 Summary of Model and Experimental Fit Parameters.

4.5 Conclusion
In summary, we have tested silica/PMMA or Polystyrene hybrid microresonators
and the Q was intermediate value between pure silica and pure polymer microresonators
in certain range of coated film thickness. The results showed good agreement for
theoretical and experimental data. As a result, these hybrid microresonators shows
material loss limited which governed by percentage of optical field in each region (silica,
polymer thin film, and air). Higher contrast of refractive indices between silica and
Polystyrene makes more percentage of the optical power go into coating layer compared
to silica and PMMA case at same thickness. With suitable surface functionalized
polymer, these devices would improve detection sensitivity [7, 8] for bio/chem sensing
applications. Also, these types of hybrid UHQ devices will find applications in
telecommunications as optical filters, modulators [9] and real time materials’
characterization such as polymer and nanocomposite thin film dynamics.
51

Chapter 4 References
1. Choi, H.-S., X. Zhang, and A.M. Armani, Hybrid Silica-Polymer Ultra-High-Q
Microresonators. Optics Letters, 2010. 35(4): p. 459-461.

2. Walsh, C.B. and E.I. Franses, Ultrathin PMMA films spin-coated from toluene
solutions. Thin Solid Films, 2003. 429(1-2): p. 71-76.

3. Kasarova, S.N., et al., Analysis of the dispersion of optical plastic materials.
Optical Materials, 2007. 29(11): p. 1481-1490.

4. Pinnow, D.A., et al., Fundamental optical attenuation limits in the liquid and
glassy state with application to fiber optical waveguide materials. Applied Physics
Letters, 1973. 22(10): p. 527-529.

5. Oxborrow, M., Traceable 2-D finite-element simulation of the whispering-gallery
modes of axisymmetric electromagnetic resonators. IEEE Transactions on
Microwave Theory and Techniques, 2007. 55(6): p. 1209-1218.

6. Spillane, S.M., Fiber-coupled Ultra-high-Q Microresonators for Nonlinear and
Quantum Optics. 2004.

7. Dong, C.H., et al., Fabrication of high-Q polydimethylsiloxane optical
microspheres for thermal sensing. Applied Physics Letters, 2009. 94(23): p.
231119.

8. Chao, C.Y. and L.J. Guo, Biochemical sensors based on polymer microrings with
sharp asymmetrical resonance. Applied Physics Letters, 2003. 83(8): p. 1527-
1529.

9. Rabiei, P., et al., Polymer micro-ring filters and modulators. Journal of Lightwave
Technology, 2002. 20(11): p. 1968-1975.

52

Chapter 5 Thermal nonlinear effects in hybrid optical
microresonators

5.1 Introduction

The resonant frequency position in microcavities is very unstable because the
large build-up intensities inside the cavity change the refractive index of the medium
(silica), also known as the thermo-optic effect. This causes linewidth distortion
(broadening and compression) which makes it difficult to accurately interpret
experimental data.
In this research, it is shown that it is possible to design a device which is
thermally stable by balancing the thermo-optic coefficient of silica with that of a second
material. To demonstrate a thermally stable resonant cavity, we coat two distinctly
different polymers, polydimethylmethacrylate (PMMA) and polystyrene (PS) on the
surface of silica microtoroid resonant cavities, forming hybrid optical cavities as shown
in Figure 5-1. It is important to note that the thermo-optic coefficient of silica and the
two polymeric materials have opposite signs. Therefore, by optimizing the optical field
overlap in the two materials, we are able to form a temperature-insensitive device.
53

Figure 5-1 Polymer coated (Hybrid) optical microtoroid resonators. a) Pov-Ray rendering of the polymer coated
devices. The yellow line in the cavity indicates that whispering gallery mode is confined within hybrid device. Green
color indicates polymer film which part of it is cut away for clarity. b) An optical image of a fabricated hybrid device.

5.2 Theory
Depending on the amount of coupled optical power, the refractive index will
either increase or decrease. This change results in a distortion in the resonant linewidth
and a shift in the resonant frequency. An additional mechanism is the mechanical or
physical deformation of the polymer film resulting in a change in the resonant cavity’s
size. This also will change the resonant frequency. This pair is also known as the
thermo-optic effect (dn/dT) and thermal expansion effect (), and it is theoretically
defined as the equation below:
Δ/ΔT=
0
(
eff
+(dn
eff
/dT)/n
eff
) (5.1)
where 
0
, 
eff
, dn
eff
/dT and n
eff
represent cold cavity resonance wavelength, the
effective material expansion, thermal optic coefficients and refractive indices of the
microresonators [1]. In the hybrid system, the definition of 
eff
, dn
eff
/dT and n
eff
should be
54

changed to express optical field distribution in each silica, polymer and air region as
following; 
eff
= β
silica
+ γ
Polymer
+ δ
air,
dn
eff
/dT = β(dn/dT)
silica
+ γ(dn/dT)
Polymer
+δ(dn/dT)
air
and n
eff
= βn
silica
+ γn
Polymer
+ δn
air
. As the β, γ and δ are fraction of optical
field distribution in silica, polymer and air, respectively, the sum of β, γ and δ should be
1.
As mentioned in previous chapters, the β, γ and δ are determined by COMSOL
Multiphysics FEM [2, 3]. In Figure 5-2, a set of FEM simulations results are shown for
silica microtoroid, 100nm thick PMMA or PS coated hybrid microtoroids with device
size of 40(8)m major (minor) diameter at λ=980nm. As can be seen, the optical field
shifts toward polymer layer for both PMMA and PS coated hybrid device compared to
monolithic silica device due to higher refractive index of coated polymers than silica. In
addition, there is more optical field confined in PS coated device compared to PMMA
coated one at same thickness as the refractive index of PS (~1.59) is higher than
refractive index of PMMA (~1.49).

Figure 5-2 FEM simulation results. A set of optical field distribution for (a) silica microtoroid, (b) hybrid microtoroid
with a 100nm thick PMMA film and (c) hybrid microtoroid with a 100nm thick PS film with device size of 40(8)m
major (minor) diameter at λ=980nm. As can be seen, the part of optical field is interacting with polymer films b oth for
PMMA and PS coated hybrid devices due to higher refractive index of coated polymer compared to silica.

55

As can be seen in Figure 5-3, the normalized radial field distribution is shown as a
function of radius where the zero point is the center of the minor diameter to support Fig
5-2. The obvious differences of optical field distribution among the three different
devices are clearly evident.

Figure 5-3 The normalized radial optical field intensity as a function of radius for silica (solid black line), 100nm thick
PMMA coated (red dashed line) and 100nm thick PS coated (blue dotted line) devices at λ=980nm. The zero point
indicates that the center of minor diameter.

The calculated fraction of optical field (β) in silica layer is shown in Figure 5-4
with no film, 100 and 200nm thick PMMA films at both =850 and 980nm as a function
of minor diameters with constant 40μm major diameter. The fraction of the optical
powers in silica layer is determined by Power
In(silica)
/Power
Tot
. For monolithic silica
device, there is more optical field confined in the silica layer with shorter operating
wavelength (850nm) compared to longer operating wavelength (980nm). However, the
field is less in silica with polymer coatings with shorter operating wavelength, indicating
that more field is confined in polymer layer with short operating wavelength compared to
56

longer operating wavelength. This is more apparent when the polymer film is thicker.
This shows that the optical field distribution in hybrid device can be controlled by
coupled light wavelength, polymer kind and film thickness. As results from higher
absorption loss of polymers compared to silica, the quality (Q) factor is decreased as the
polymer film thickens [3].
Even micrometer-scale change of minor diameter play a role to control optical
field in the cavity; it is important to note that nanometer-scale thickness change of a
polymer coating controls the optical field distribution in hybrid devices. Therefore, the
optical field distribution is held constant by film thickness in this research to study the
dependence of resonant shift. More detailed description about the simulation results can
be found elsewhere [3] and previous chapters.

Figure 5-4 Calculated percent of optical field in SiO
2
layer with 40μm major diameter of the device as a function of
minor diameter at =850 and 980nm. The simulations results shows with no film, 100nm and 200nm of PMMA
coatings. As can be seen, the optical field is highly dependent on polymer film thicknesses and operating wavelengths.

57

Based on the equation 5.1, the nature of the red or blue shift can be determined as
a function of temperature. For this theoretical determination, the effective refractive
index (n
eff
) and thermo-optic coefficient (dn
eff
/dT) with different optical devices (silica or
hybrid devices) should be calculated by incorporating the calculated optical field
fractions (β, γ and δ) in each region. This shift is primarily caused by the following
parameters: 1) wavelength, 2) coating thickness, 3) type of polymer and 4) device size.
In Figure 5-5 and 5-6, it is shown that the dependence of resonant wavelength
shifts for silica devices, 100 or 200nm thick PMMA or PS coated hybrid devices as a
function of temperature at =850nm and 980nm for the device size of 40(8) μm major
(minor) diameter. It is important to note that the resonant shift is almost balanced with
100nm thick PS coated or 200nm thick PMMA coated hybrid devices, so an experimental
investigation is performed to verify these theoretical calculations.
In FEM simulations, the thermal optic coefficient used for PMMA is -10510
-6
/K
[4] and for PS is -14010
-6
/K [4]. The thermal optic coefficient for silica is used with
11.910
-6
/K, taken from a literature [5]. Based on the simulations, the calculated resonant
shifts show obviously different behavior due to the differences in effective thermal optic
coefficients and refractive indices. This shift is mainly dependent on the following
parameters: 1) wavelength, 2) coating thickness, 3) type of polymer 4) device size. The
results for this prediction are shown in Figure 5-5 and 5-6.
58

Figure 5-5 Calculated resonant wavelength shift as a function of temperature with silica and hybrid devices at
=850nm and 980nm for both PS and PMMA. a) 100nm thick PS and PMMA films at 850 and b) 980nm. c) 200nm
thick PS and PMMA films at 850 and d) 980nm. As can be seen, depening on device kinds, we can see different optical
responses.

Thermally stable conditions can be determined based on thermal optical
coefficient of polymers and silica and they are 7.83 and 10.18% for PS and PMMA,
respectively. This gives us that the film thickness should be around 110nm for PS and
180nm for PMMA for thermally stable condition at λ=850nm. This is clearly shown by
theoretical calculations in Fig 5-5 or 5-6. It is important to note that while the percentage
of optical field for thermally stable conditions in polymer layer is independent of
wavelength, the film thickness for thermally stable conditions is dependent on
59

wavelength due to the optical field distribution change by polymer film thickness in the
hybrid optical microcavity.

Figure 5-6 Calculated resonant wavelength shift as a function of temperature with silica and hybrid devices at
=850nm and 980nm for both PS and PMMA. a) 100nm and 200nm PMMA films at 850 and b) 980nm. c) 100nm and
200nm thick PS films at 850 and d) 980nm. As can be seen, depening on device kinds, we can see different optical
responses.

The response time for thermal expansion is much slower than the response time
for thermo-optic effect, so only thermal-optic effect is considered in the present system
[6, 7]. In addition, while the Kerr nonlinearity is a dominant nonlinear effect in a
cryogenic environment, it is negligible in an ambient environment [8]. Therefore, this
60

effect is not considered, either. As a result, the calculated resonant shift is based solely on
thermal-optic effect.

5.3 Experimental Results and Discussions
Based on theoretical calculation, a series of devices are chosen to study the
thermal nonlinear effect. One thickness is chosen near a thermally stable condition and
another thickness is chosen far from a thermally stable condition for PMMA or PS coated
devices to compare with theoretical prediction. In addition, the monolithic silica device is
also tested for comparison with known reference data and for comparison with hybrid
devices. The procedure of the detailed device fabrication is described in chapter 2 and 4.
All of quality (Q) factor are recorded with low input power in under-coupled
regime to avoid thermal broadening and compression. The limitation of the quality (Q)
factor is highly dependent on coated polymers due to higher optical absorption than silica
[3].
For the measurement of thermal non linear effect, several parameters are held
constant: 1) device size, 2) scan speed, 3) scan range to minimize any other experimental
variation and inaccuracy. To study and understand thermal nonlinear effect in both silica
and hybrid devices, a series of experiments are performed with variation of following
parameters: 1) quality factor, 2) input power, 3) wavelength and 4) film thickness.
The examples of the thermal broadening and compression resonant peaks are
shown in Figure 5-7 for both silica and 200nm thick PS coated hybrid devices. As silica
has positive thermal-optic coefficient, the resonant shift shows broadening for forward
61

scan while it shows compression for backward scan with monolithic silica microtoroid
(red shift). In contrast to monolithic silica microtoroid, the resonant shift shows
compression for forward scan while it shows broadening for backward scan with 200nm
thick PS coated hybrid microtoroid (blue shift). This is because the effective thermal-
optic coefficient is positive for silica devices, while it is negative for 200nm thick PS
coated hybrid devices. As previously mentioned, silica has positive thermal-optic
coefficeint (11.910
-6
/K), while PS has negative thermal-optic coefficients (-14010
-
6
/K). Therefore, 200nm thick PS coating is enough to change optical response from red to
blue shift. This indicates that the optical field confined in polymer layer play a major role
to change optical response of the optical microcavities.

Figure 5-7 The examples of the thermal broadening and compression peaks for (a) silica and (b) hybrid (PS200nm)
devices at λ=850nm. While the thermal broadening occurs in fo rward scan for silica device (red shift), it occurs in
backward scan for hybrid device (blue shift). The positive or negative effective thermal-optic coefficients play a major
role to cause red or blue shift.

An example of an experimental result, which shows how the shift was determined
from under to critical coupled regime is shown in figure 5-8.
62

Figure 5-8 Thermal nonlinear shift determination from under to critical coupled regime

The first set of experiments are performed with silica devices and 130nm or
200nm thick PMMA coated hybrid devices as a function of input power with different
quality (Q) factors for each device. For 130nm thick coated PMMA hybrid devices, the
film thickness is far from athermal condition while for 200nm thick coated PMMA
hybrid deivces, the film thickness is close to athermal condition. Therefore, resonant
shifts show different optical response for silica and 130nm or 200nm thick PMMA coated
hybrid devices for both at =850nm and 980nm testing wavelength as can be seen in
Figure 5-9. Due to higher build up intensity in the cavities, the resonant shifts are larger
at higher input power and quality factor. This high circulating intensity induces
temperature change, leading to refractive index change, hence changing the optical path
in the cavity.

63

Figure 5-9 The resonance shift for silica, PMMA 130nm and 200nm thick coated hybrid devices for (a) =850nm and
(b) =980nm with different quality factors as a function of input powers. The resonant shift is larger with higher Q due
to high circulating build up intensity in the optical cavities. Athermal condition is nearly achieved with 200nm thick
PMMA coated hybrid devices. Lines are a guide to the eye for ease of comparison.

Compared to silica devices, 130nm thick coated PMMA hybrid devices show
reduced red shift while 200nm thick coated devices show blue shift due to the change of
effective thermal-optic cofficient of the system at both tested wavelengths. Thermally
stable condition is almost achieved with 200nm thick PMMA coated hybrid devices at
both wavelengths as can be seen in Figure 5-9. This thickness is close enough values
based on theoretical calcuation, which is 180nm at λ=850nm and 190nm at λ=98 0nm, for
thermally stable conditions with a device size of a 40(8) μm major (minor) diameter.
However, rather larger shifts are observed with 980nm compared to 850nm. This might
be the difference of material absorption for tested wavelengths.
The second set of experiments is performed with 100nm or 200nm thick PS
coated hybrid devices at λ=850 and 980nm as a function of input power with different
quality (Q) factors. For 100nm thick coated PS hybrid devices, the film thickness is close
to the athermal condition while for 200nm thick coated PS hybrid deivces, the film
thickness is far from athermal condition. As a result, the 100nm coated one shows almost
64

balanced resonant shift, while the 200nm coated one shows obvious blue shift for both at
=850nm and 980nm testing wavelength as can be seen in Figure 5-10. Same as PMMA
coated hybrid devices, due to higher build up intensity in the cavities, the resonant shifts
are larger at higher input power and quality factor. However, near thermally stable
conditions, the resonant shift is rather independent on the quality (Q) factor compared to
non-athermal condition. For easy comparison the same resonant shift data for silica
devices are also included in Figure 5-10.

Figure 5-10 The resonance shift for silica, PS 100nm and 200nm thick coated hybrid devices for (a) =850nm and (b)
=980nm with different quality factors as a function of input powers. The resonant shift is larger with higher Q due to
high circulating build up intensity in the optical cavities. Athermal condition is nearly achieved with 100nm thick PS
coated hybrid devices. Lines are a guide to the eye for ease of comparison.

The experimental results which are plotted in Figures 5-9 and 5-10 in the
experimental section are detailed in Table 5-1.

65

Polymer
Coating

(nm)
Film Thickness
(nm)
Q Δλ (pm)
(at 1000µW)
none 633 NA 4.68×10
7
8.19
5.74×10
7
10.56
none 850 NA 2.58×10
7
11.38
3.24×10
7
30.48
none 980 NA 3.38×10
6
21.47
5.71×10
7
90.72
PS 850 100 9.62×10
5
1.62
3.27×10
6
2.35
PS 850 200 2.73×10
5
-9.78
3.67×10
5
-25.95
5.11×10
5
-53.23
PS 980 100 1.13×10
6
4.60
2.65×10
6
7.64
PS 980 200 3.30×10
5
-61.89
PMMA 850 130 1.20×10
6
13.53
5.38×10
6
19.82
PMMA 850 200 3.62×10
5
-0.80
6.26×10
5
-2.00
PMMA 980 130 6.60×10
5
4.80
2.20×10
6
16.32
PMMA 980 200 1.79×10
5
-7.12
2.82×10
5
-16.89

Table 5- 1 Experimental Values.

As can be seen from Figure 5-9 and 5-10, the optical response is quite different
between PMMA and PS coated device. This is mainly because they have different
66

refractive indices and thermal-optic coefficients. Due to the lower refractive index of
PMMA than PS, the optical field has less interaction with a PMMA layer than a PS layer
of the same thickness. Also, in addition to refractive index, due to the lower thermal-optic
coefficient of PMMA than PS, the thermally stable condition is achieved with a much
thicker film for PMMA coated devices than for PS coated devices.
The experimental testing results match well with theoretical calculation for
thermally stable conditions for both PMMA and PS coated devices. Therefore, the
proposed method could be applied to the other polymers to control optical response in the
cavity for optical field enhancement in polymer layers for various applications such as
nanocomposite’s study and microlasers.

5.4 Conclusions
In summary, the resonant shifts induced by thermal nonlinear effect are
investigated for PS or PMMA coated hybrid devices as well as silica microresonators
both for experiments and theoretical calculations. The results show that the resonant
shifts for the WGM optical microcavities are highly dependent on the 1) quality factor 2)
input power, 3) wavelength, and 4) film thickness. While monolithic silica devices
always shows red shift, PS or PMMA coated hybrid devices show a redshift, a blueshift
or no shift by controlling the film thicknesses of polymers, hence controlling the optical
field in each region.
The thermally stable conditions are shown with 100nm PS coated devices and
200nm PMMA coated devices at both λ=850nm and 980nm. The optical field distribution
67

is highly dependent on refractive indices of medium used for optical microcavities. There
are good agreement between experimental results and theoretical calculations. Therefore,
the theoretical analysis could be extended to additional hybrid systems with other
polymers. With proper choice of film thickness, the hybrid devices would be used as
environmentally stable bio/chem. sensing purpose. In addition, using functional polymers
and elastomeric polymers, various other hybrid devices could be developed with similar
approach [9, 10].

68

Chapter 5 References
1. Carmon, T., L. Yang, and K.J. Vahala, Dynamical thermal behavior and thermal
self-stability of microcavities. Optics Express, 2004. 12(20): p. 4742-4750.

2. Oxborrow, M., Traceable 2-D finite-element simulation of the whispering-gallery
modes of axisymmetric electromagnetic resonators. IEEE Transactions on
Microwave Theory and Techniques, 2007. 55(6): p. 1209-1218.

3. Choi, H.S., X.M. Zhang, and A.M. Armani, Hybrid silica-polymer ultra-high-Q
microresonators. Optics Letters. 35(4): p. 459-461.

4. Weber, M.J., Handbook of optical materials. 2003: CRC Press.

5. Han, M. and A. Wang, Temperature compensation of optical microresonators
using a surface layer with negative thermo-optic coefficient. Optics Letters, 2007.
32(13): p. 1800-1802.

6. Choi, H.-S. and A.M. Armani, Thermal non-linear effects in hybrid optical
microresonators. Applied Physics Letters, 2010. 97(22): p. 223306.

7. He, L., et al., Compensation of thermal refraction effect in high-Q toroidal
microresonator by polydimethylsiloxane coating. Applied Physics Letters, 2008.
93(20).

8. Treussart, F., et al., Evidence for intrinsic Kerr bistability of high-Q microsphere
resonators in superfluid helium. European Physical Journal D, 1998. 1(3): p. 235-
238.

9. Armani, A.M., A. Srinivasan, and K.J. Vahala, Soft lithographic fabrication of
high Q polymer microcavity arrays. Nano Letters, 2007. 7(6): p. 1823-1826.

10. Dalton, L.R., P.A. Sullivan, and D.H. Bale, Electric Field Poled Organic Electro-
optic Materials: State of the Art and Future Prospects. Chemical Reviews. 110(1):
p. 25-55.

69

Chapter 6 Studying polymer thin films with hybrid optical
microcavities

6.1 Introduction
As the applications for polymer and nanocomposite coatings are increasing, there
is additional pressure to more fully understand how changes in temperature affect the
mechanical and optical properties of polymer thin films, at the nano-scale level.[1, 2]
Currently, researchers employ a suite of methods, including ellipsometry, nano-
indentation, calorimetry, and thin-film spectrophotometry.[2-4] However, no single
method is able to resolve both the mechanical and optical changes in the thin films in
real-time (sub-microsecond accuracy) with high resolution. For example, indentation
methods, like nano-indentation or atomic force microscopy, which have the highest
spatial resolution, have limited abilities to determine a sample’s optical properties and
they typically degrade a sample which limits the ability to perform a measurement
iteratively.[5]
In addition, there are limited non-destructive characterization methods which are
able to perform these high resolution measurements. In this chapter, we demonstrate a
method which is able to detect temperature-induced changes in the refractive index of
polystyrene polymer thin films as small as 10
-7
. This approach is based on optical
microcavity resonators. The experimental results agree well with the theoretical
simulations.
70

6.2 Theory
Optical sensing devices based on evanescent field whispering gallery mode
resonant cavities have demonstrated the ability to detect subtle changes in the optical
properties of the device. Specifically, the circulating optical field interacts with the
microcavity, and any changes in the material are detected as a change in the resonant
wavelength of the device, which is determined by the refractive index and the geometry
of the resonator, among other properties. Variations in temperature induce changes in the
refractive index of the material (thermo-optic behavior) and changes in the material size
(expansion). As previously mentioned in chapter 5, the governing equation for these
effects is:[6]
Δ/ΔT=
0
(
eff
+(dn
eff
/dT)/n
eff
) (6.1)
For example, using silicon resonant cavities, researchers have demonstrated the
sensitivity of 83 pm/°C, which corresponds to a change in the refractive index of the
device 1.3710
-4
.[7] Alternatively, using surface plasmon devices and temperature
stabilization, this sensitivity can be improved to over 10
-7
refractive index units.[8] This
sensitivity improvement is the result of several advancements, one of which is the longer
photon lifetime or higher quality factor (Q) of the device. If an ultra-high Q microcavity
(Q>10
8
) is used, the sensitivity will be improved further.
If the microcavity is coated with a polymer film, creating a hybrid cavity, it can
be used to measure the optical properties and the mechanical behavior of the film.
Because the optical and mechanical effects in polymers have very different response
times, it is possible to de-convolve the two behaviors.
71

Hybrid systems are slightly more complex to analytically describe, because the

eff
, dn
eff
/dT and n
eff
must incorporate the polymer and the device material properties
according to the degree of interaction with the circulating optical field, which is
determined by the device geometry and material properties.[9] Therefore, by optimizing
device design, it is possible to maximize the overlap of the optical field with the polymer,
enhancing the device’s ability to study the polymer film.
In the present work, a hybrid device is developed with the express design of
exploring the temperature-dependent refractive index of a polymer film. All
experiments are performed at temperatures which are significantly away from the glass
transition temperature of the polymer to ensure that the polymer does not mechanically
deform. The hybrid system consists of a silica resonant cavity conformally coated with
polystyrene (PS). PS is chosen as the coating polymer because it has higher refractive
index than silica allowing a large overlap of the optical field with the polymer film, it
does not readily absorb water ensuring minimal artifacts, and it has very low optical loss.
The hybrid microtoroid silica resonators are fabricated using a previously
described method [10, 11] and described well in previous chapter 2 and 4. In the present
work, 200k molecular weight polystyrene (Sigma Aldrich) was used.

6.3 Experimental Results and Discussions
The detailed procedure for optical characterization is described in chapter 2. For
these measurements, a 980nm laser was used and an optical attenuator was placed in-line
to control the input power.
72

To determine the resonant shift data as a function of temperature, a heater and
thermocouple sensor are integrated directly into the resonator sample holder as can be
seen Figure 6-1. The heater (Omega CSH-102100/120V) is mounted directly under the
optical devices. Silver conductive epoxy (MG chemicals) is used between the heater and
the stage to ensure effective heat transfer from the heater to the sample holder, to
minimize heat lost to the environment, and to hold the heater in place. A surface-mount
fast-response thermocouple sensor (Omega SA1XL) is attached immediately adjacent to
the optical device to accurately read the temperature in real-time. The response time for
this thermocouple sensor is less than 0.15 seconds. Both the heater and the sensor are
connected to a benchtop controller (Omega CSC32 series).

Figure 6-1 Sample stage with integrated heater and temperature control.

The resonant shift results are automatically recorded by a custom LabView
program. Initial control experiments are performed to ensure the reliability and stability
of this stage and the testing software. From these experiments, it is determined that there
73

is negligible variation in resonance wavelength over the timeframe of a typical
experiment as can be seen Figure 6-2 (less than 0.3pm over four minutes in air). During
the record of the resonant shift upon heating and cooling, the author try to keep the
constant coupling efficiency as shown in Figure 6-3a. As can be seen in Figure 6-3, the
resonant shift shows clear optical response (thermal history) upon heating, cooling and
equilibrium.

Figure 6-2 The measured resonance shift as a function of time in air. The resonance peak shows stable in air, indicating
that the current testing set up is good for thermal sensing purpose.

74

Figure 6-3 The measured resonance shift as a function of time. (a) continuous heat up to set point and keep constant
temp. (b) heat was on and off.

Initial experiments studied the response of the hybrid device with that of a single
material (silica) device. Specifically, two different experiments were performed: 1)
increasing the temperature by the same amount (N>5 times) and 2) increasing the
temperature by a different amount. These results are shown in Figures 6-4a and b. By
performing both experiments, we are able to determine the linearity of the sensor
response and the stability of the sensor after the temperature change. The slight decline
after each step for both the silica and the hybrid device is the result of the heater over-
shooting the set temperature and then correcting.
75

Figure 6-4 Comparison of sensor response between silica and hybrid devices. a) The T is the same for each increment.
b) The T is increasing for each increment. Because of the polymer layer, the hybrid device has a significantly larger
response. Note that the y-axis is the absolute value of the resonant wavelength shift or the magnitude, for easy
comparison between the two devices.

Figure 6-5 shows all of these results compiled into a single graph. As can be
easily observed, the monolithic silica microtoroids exhibit a redshift, however, the hybrid
microtoroids show blueshift. A simple linear fit to the data provides a significant amount
of insight into the dominant detection mechanism. The slope of the silica data is usually
around 11pm/°C, which agrees well with theoretical value as can be seen 6-5. This
indicates that the primary cause for the resonant frequency change in the silica device
over this temperature range is the change in refractive index of the device.

76

Figure 6-5 The results from experiments like those shown in Figure 6-4 were compiled to show the shift versus
temperature for both the hybrid and silica devices. Additionally, the theory based on equation 6.1 is included. The
shift for the silica device is significantly smaller than the shift for the hybrid device, and the theory and experiment are
in good agreement.

In the hybrid device, the slope is -36pm/°C. If the device does not experience an
expansion, then this value represents the effective sensitivity of the device which relates
to effective dn/dT. The change from a positive to a negative value is the result of the
negative thermo-optic coefficient of PS. Using the geometry of the device, the dn/dT
values of PS and silica from literature[12, 13], and the testing wavelength, the effective
sensitivity was calculated to be between -28pm/°C (PS=250nm) to -40pm/°C
(PS=300nm). This value is extremely similar to the experimental results, indicating that
over this temperature range, the polymer film does not deform. At higher temperatures,
based on the expansion coefficient of PS and the completion between the two effects
(thermo-optic inducing a blue shift and expansion inducing a red-shift), we would expect
to see shift of approximately -4pm/
o
C, which is easily resolvable.
77

Based on the previous results, we can determine the threshold sensitivity of the
device to changes in the refractive index of polystyrene. Specifically, based on the
measured noise threshold with the present set-up (0.3pm), the theoretical threshold
refractive index sensitivity is 4.5x10
-7
.
Finally, experiments are performed to determine if the hybrid sensor is able to
monitor the behavior of PS over several heat and cool cycles, or if it experiences
degradation. To perform this experiment, we first heat the sensor, and then let it cool to
the initial temperature. As can be seen in Figure 6-6a/b, the signal produced by the hybrid
device in response to the temperature increase and decrease is fairly consistent over all
eight cycles. This result verifies that the signal produced by the hybrid device is stable,
and the basic device structure does not rapidly degrade. Both of these are key features of
a thin film measurement device.

Figure 6-6 Reproducibility of results a) The measurement was performed iteratively. b) The data in part a) is re-plotted
to emphasize the hysteretic behavior due to the heating element.

78

It is important to comment on the hysteretic behavior which is clearly evident in
Figure 6-6b. For clarity, the heating-cooling cycle is counter-clockwise. The time-scale
is far too slow for the hysteresis to be related to the hybrid device, which has a
millisecond response time, including the limitations in the current detection system
(function generator, laser, etc). Therefore, the rate of heating is determined by the heating
element, while the rate of cooling is determined by heat transfer from the sample to the
environment and to the substrate.

6.4 Conclusions
In conclusion, we have experimentally and theoretically verified the ability of
optical microcavities to non-destructively detect the temperature dependent optical
properties of polymer thin film coatings. Specifically, we measured the dn/dT of silica
and hybrid devices using an optical cavity and concluded that the sensing signal was
solely generated by the change in refractive index, with no contribution from material
expansion. By performing finite element method simulations, we calculated the
interaction of the optical field with the polymer film and developed an accurate model of
the system which had good agreement with the experimental results. As a result of
minimal device degradation over the course of the iterative experiments, the response is
extremely reproducible and, in the temperature range studied, the sensor response is very
linear. Because this sensor system accurately monitors the effect of temperature on the
optical properties of the polymer material with significantly improved sensitivity over
currently available methods, it will find numerous uses in studying how temperature
79

changes effect and degrade the optical properties of polymer thin films, which have
applications ranging from solar cells to nanocomposite coatings.[14, 15]

80

Chapter 6 References
1. Kropka, J.M., V. Pryamitsyn, and V. Ganesan, Relation between glass transition
temperatures in polymer nanocomposites and polymer thin films. Physical
Review Letters, 2008. 101(7).

2. Huang, Y. and D.R. Paul, Physical aging of thin glassy polymer films monitored
by optical properties. Macromolecules, 2006. 39(4): p. 1554-1559.

3. Mundra, M.K., et al., Confinement, composition, and spin-coating effects on the
glass transition and stress relaxation of thin films of polystyrene and styrene-
containing random copolymers: Sensing by intrinsic fluorescence. Polymer, 2006.
47(22): p. 7747-7759.

4. Priestley, R.D., et al., Structural relaxation of polymer glasses at surfaces,
interfaces and in between. Science, 2005. 309(5733): p. 456-459.

5. Burghard, Z., et al., Toughening through Nature-Adapted Nanoscale Design.
Nano Letters, 2009. 9(12): p. 4103-4108.

6. Carmon, T., L. Yang, and K.J. Vahala, Dynamical thermal behavior and thermal
self-stability of microcavities. Optics Express, 2004. 12(20): p. 4742-4750.

7. Kim, G.D., et al., Silicon photonic temperature sensor employing a ring resonator
manufactured using a standard CMOS process. Optics Express. 18(21): p. 22215-
22221.

8. Davis, L.J. and M. Deutsch, Surface plasmon based thermo-optic and temperature
sensor for microfluidic thermometry. Review of Scientific Instruments, 2010.
81(11): p. -.

9. Choi, H.-S. and A.M. Armani, Thermal non-linear effects in hybrid optical
microresonators. Applied Physics Letters, 2010. 97(22): p. 223306.

10. Armani, D.K., et al., Ultra-high-Q toroid microcavity on a chip. Nature, 2003.
421(6926): p. 925-928.

11. Choi, H.-S., X. Zhang, and A.M. Armani, Hybrid Silica-Polymer Ultra-High-Q
Microresonators. Optics Letters, 2010. 35(4): p. 459-461.

12. He, L., et al., Compensation of thermal refraction effect in high-Q toroidal
microresonator by polydimethylsiloxane coating. Applied Physics Letters, 2008.
93(20).

81

13. Weber, M.J., Handbook of optical materials. 2003: CRC Press.

14. Vendamme, R., et al., Robust free-standing nanomembranes of organic/inorganic
interpenetrating networks. Nature Materials, 2006. 5(6): p. 494-501.

15. Kietzke, T., et al., Novel approaches to polymer blends based on polymer
nanoparticles. Nature Materials, 2003. 2(6): p. 408-U7.

82

Chapter 7 Thermo-optic coefficient of polyisobutylene ultra
thin films measured with integrated photonic devices

7.1 Introduction
The optical properties of polymeric materials, such as transmission loss and
thermo-optic coefficient, determine their utility in numerous applications, ranging from
nanotechnology to the automotive and aerospace industries. However, due to the wide
variation in the physical properties of polymers, many are unsuited for characterization
using conventional techniques; consequently, their optical properties are unknown. One
such polymer is polyisobutylene (PIB), which is viscous at room temperature, and
therefore, is not compatible with conventional transmission loss and the thermo-optic
coefficient characterization techniques, as they rely on contact measurements. To
overcome this, we have developed an integrated, microscale optical sensor that relies on
an evanescent wave to study the material’s optical behavior. Using this device, we
successfully determined the refractive index, the transmission loss and the thermo-optic
coefficient of ultra-thin-films of polyisobutylene. The films are deposited on the sensor’s
silica surface using either spin-coating or surface initiated cationic polymerization,
demonstrating the flexibility of this approach.

7.2 Background & Motivation
Ultra-thin polymer thin films have found a wide array of applications, ranging
from the electronics industry to the photovoltaics industry.[1-4] Ultimately, a polymer’s
83

utility in a specific application is determined by its relevant fundamental physical
properties, such as its optical characteristics. For example, low refractive index and low
transmission loss polymers have found application in designing hybrid optical
components, such as polymer lenses and optical fiber, whereas high absorption polymers
with good photo-electron conversion efficiencies are used in the design and construction
of polymer solar cells.[5]
Performing the requisite measurements of these fundamental properties has
proven challenging because of the diversity of polymeric materials available to be
studied, particularly when thin films, rather than bulk materials, are the object of
investigation. As a result, numerous techniques have been developed over the past
century to study polymer thin film properties. Using these methods, researchers recently
discovered that the material properties of polymer thin films vary significantly depending
on the processing techniques used to deposit the polymer in thin film form.[6, 7] For
example, the optical and electrical properties of pure and iodine-doped polyaniline thin
films change when using radio frequency versus alternating current plasma
polymerization techniques, due to the polymer adopting different structures or
conformations during deposition.[8] Therefore, it is critical to not only characterize the
optical material characteristics of polymer thin films, but to also determine the impact of
the film preparation method on the aforementioned properties.
PIB was chosen as the model system for this study due to its unique bulk material
properties, such as its gas impermeability, that give rise to its utility in a variety of non-
film applications.[9-11] For example, it polymerizes at -50 ˚C, in contrast to other
84

common polymers, such as polystyrene (PS) and polymethylmethacrylate (PMMA),
which polymerize at much higher temperatures (between 22-24˚C).[12, 13] Additionally,
PIB has one of the weakest temperature dependencies on structural relaxation and
viscosity, and demonstrates strong inelastic scattering even at temperatures well above its
glass transition.[14] However, for low molecular weight formulations it is not solid at
room temperature and is typically permanently “tacky,” while for higher molecular
weights, it is a soft solid, unless it is crosslinked with isoprene to form butyl rubber.[15,
16] Therefore, the optical properties of this material have been extremely difficult to
characterize both in the bulk and thin film regimes, limiting its potentially high use in
optical applications that require thin film configurations.

7.3 Methods and Materials
In order to evaluate the optical properties of PIB ultra-thin films, the films were
coated onto the surface of the integrated silica sensor via two different methods: spin-
coating and surface-initiated cationic polymerization. These two approaches for film
deposition are significantly different, and we hypothesize that the deposition method
should impact their optical properties. For example, surface-initiated cationic
polymerization creates an ordered structure (i.e., a polymer brush) consisting of oriented
polymer chains that are tethered by one end to a substrate’s surface while the other end is
free. As tethering density and chain length increases, the chains repel each other to avoid
overlapping, resulting in a brush-like structure.[17] While there are many methods of
forming polymer brushes, by using surface-initiated cationic polymerization, it is possible
85

to achieve a high grafting density and a uniform film thickness directly from the
isobutylene monomer.[18] In contrast, if the films are created via spin-coating, the
resulting polymer film will be completely disordered.
7.3.1 Surface-initiated cationic polymerization
A surface-initiated cationic polymerization technique was selected because of its
ability to tolerate different functional monomers and its flexibility in experimental
conditions; the conditions shown in Figure 7-1 were used and were based on prior
literature.[19, 20]

Figure 7-1 Reaction scheme for forming the polymer brush on the silica surface using surface-initiated cationic
polymerization.

These conditions were chosen due to their simplicity and their ability to create
ultra-thin films of PIB on silica surfaces, rather than high molecular weight bulk
polymers, which is more typical of other polymerization conditions. The polymerization
is a three-step process, using an alkyl halide, attached to an organosilane as the surface
initiator for the “polymerization from” reaction. Here, the organosilane is used to bridge
86

the inorganic surface and the initiator via reaction of the surface hydroxyls with the
chlorine leaving group attached to silicon. Initially, the sensor was hydroxylated using an
O
2
plasma etcher, and the organosilane, with the surface initiator attached, was deposited
using vapor deposition. Following the vapor deposition step, the devices were placed in a
small reaction flask, which was then sealed with a rubber septum. Dichloromethane (15
mL, 99.5%, VWR) and diethylaluminum chloride (0.053 mL, catalyst, 97.00%, Sigma
Aldrich) were added to the flask using syringe transfer techniques, and cooled to below -
50 ˚C using a dry ice / acetone . Finally, condensed isobutylene (2.5 g, Sigma Aldrich,
IB) was added to the flask. The flask, still in the dry ice / acetone bath, was placed on an
incubating rocker, allowing the solution to be gently mixed for 30 minutes. The reaction
was then quenched by adding chilled methanol (10 mL, 99.8%, VWR), to the sealed
flask. The devices were then carefully removed and washed, on the incubating rocker,
with n-pentane (98.0%, Alfa Aesar) to remove absorbed PIB, IB monomer, and entrained
catalyst. The same parameters were used to form PIB brushes on the control wafers for
the ellipsometry measurements to determine the refractive index and film thickness
(Table 7-1). As can be seen Figure 7-2, in addition to the ellipsometric measurements,
the PIB polymer brush ultra-thin films were subjected to X-ray photoelectron
spectroscopy (XPS, M-Probe ESCA) using a 1487 eV Al Kα source, and a survey scan
from 0 to 1000 binding eV, to identify all chemical species on the surface, and to verify
that the surface-initiated cationic polymerization was performed.
Figure 7-2 shows a series of XPS spectra which were taken at three different
stages in the surface-initiated cationic polymerization process: 1) the background (100) Si
87

wafer with SiO
2
on the surface, 2) after the addition of the coupling agent and
polymerization of the polyisobutylene on the surface, and 3) after the polymerization and
additional washing (to remove surface adsorbed polymer). As shown in Figure 7-2, the
peaks correspond to the growth of the PIB polymer brush on the surface. In this case, we
see a growth of the C 1s peak at 285 eV in comparison to the underlying silicon wafer
peaks, showing that the polymer has been attached to the surface. The small C 1s peak
on the background spectrum is from adventitious carbon in the sample chamber that is
never completely removed; therefore, comparison of the ratio of the C 1s to Si peak areas
between samples is more representative of the addition of polymeric groups than solely
the presence of the C 1s peak. Additionally, in the initial PIB spectrum, we see the
presence of Al and Cl atoms, from the catalyst. By fully washing the substrate, these
species can be fully removed from the ultra-thin films; the presence of Al and Cl would
negatively impact the ability of the integrated silica sensor to obtain the optical
characteristics described previously. Thorough washing removes any catalyst that is
adsorbed onto the surface.
88

Figure 7-2 Spectra showing the attachment of the polyisobutylene to the silica surface, and a control spectrum of the
silica surface. If the polymer layer is not thoroughly washed, residual Al from the catalyst is visible (middle spectra).

To determine the initiator density on the surface of the samples,
thermogravimetric analysis (TGA) measurements (TA instruments, Q5000 IR) were
performed with a temperature ramp rate of 5
o
C/min to 600
o
C, with a 120 minute hold at
600
o
C. This approach is commonly used to determine initiator and polymer film
densities.[21] Based on complementary TGA measurements, the initiator density was
approximately 1x10
15
molecules/mm
2
on the silica surface. Previous researchers have
studied the concentration of hydroxyl groups on silicon surfaces immediately after
piranha etching, a process similar to O
2
plasma, and obtained densities of 5x10
14

molecules/mm
2
.[22] It is expected that a SiO
2
surface would have a higher density of
89

hydroxyl groups. Therefore, given the measured value, it is clear that the previously
detailed process is achieving uniform and dense surface coverage, enabling the polymer
brush formation.
7.3.2 Spin-coating
To form PIB thin films via spin-coating for ellipsometric measurements, 1 wt%
solutions of PIB (1.7E6 gmol
-1
, Acros Organics) in toluene (99.7%, VWR) were prepared
and spun onto (100) silicon wafers with 2 m of thermal oxide at 4000 rpm for 60
seconds. The same parameters were used to spin coat the sensor devices for the dn/dT
and transmission loss measurements. To enable a straightforward comparison of the two
materials, similar film thicknesses are used with the sensor (Table 7-1).
Attachment
Method
Refractive
index
Thickness
[nm]
Polymer
Brush
1.482 76

Spin-Coat 1.40233 48

Table 7- 1 The ellipsometry results of the polyisobutylene ultra thin films

7.4 Results and discussions
7.4.1 Transmission (Material) loss determination
The primary metric used to describe the performance of a resonant cavity is its
quality factor (Q), which describes the photon lifetime in the device. As demonstrated in
Chapter 4, in polymer-coated devices, this is directly related to the transmission or
90

material loss of the resonant cavity and the equation 4.1 in Chapter 4 can be changed to
following expression to determine transmission loss of the coated polymers:

(7.1)
To determine the transmission or material loss of the coated PIB, the fraction of
optical field in silica, polymer and air was determined using Finite Element Method
simulations (COMSOL Multiphysics) as detailed in previous chapters.[23, 24] The
microtoroid resonant cavity’s major (minor) diameter was held constant at 45(6) m. A
SEM image and rendering are shown in Figure 7-3 (a) and (b). The values of the
polyisobutylene (PIB) film thicknesses and refractive indices for both spin-coating and
surface-initiated cationic polymerization were used as shown in Table 7-1. The refractive
index for silica was taken from literature. One example simulation result is shown in
Figure 7-3 (d). Therefore, by solving the above equation, it is possible to determine the
transmission loss with a high degree of accuracy by simply measuring the Q of the cavity.
While Q
mat
or the intrinsic Q of the resonant cavity is constant, Q
coupl
is dependent
on the amount of power coupled into the cavity. Therefore, to accurately determine the
intrinsic Q of the resonant cavity, it is necessary to measure the quality factor at several
different coupling conditions, and then fit the data. This measurement is performed by
coupling light from a pair of narrow linewidth, continuous wave (CW) tunable lasers,
centered at either =765 or 980nm, into the optical cavity using a tapered optical fiber
waveguide.[25]
91

Figure 7-3 The integrated photonic sensor. a) Scanning electron micrograph of a single integrated photonic sensor
device. b) Rendering of the sensor, highlighting the location of the circulating optical field and its interaction with the
polymer film. c) Schematic of the testing set-up sample stage with the heating element and thermocouple. Multiple
sensors operating at different wavelengths are shown. d) A finite element method simulation of the optical field
distribution in the optical sensor system, showing the interaction of the optical field with the polymer film.

A sample resonance spectrum of a PIB coated microcavity recorded using a
980nm laser is shown in Figure 7-4. To calculate the Q factor for this individual
measurement, the spectra is fit to a Lorentzian (shown in red).
92

Figure 7-4 Transmission spectra of a resonant wavelength of the optical cavity. At resonant wavelengths of the cavity,
light is coupled into the cavity, resulting in a decrease in detected power.

A series of measurements was performed to determine the intrinsic Q, and sample
data is shown in Figure 7-5. From this data, the intrinsic Q or the Q
mat
of the resonant
cavity is determined.

Figure 7-5 The dependence of the experimentally measured Q factor on the amount of light which is coupled into the
sensor device. By fitting this data and extrapolating to the 0 point, the intrinsic Q factor (Q
0
) or material-limited Q
factor (Q
mat
) can be determined.

93

Table 7-2 contains the transmission loss measurements of the two films at =765
and 980nm. As expected, the material loss slightly increases as the wavelength increases
from the visible towards the near-IR as a result of the overtone absorption of the C-H
vibrations.[2] In addition, there is also a dependence on the deposition method. The
optical loss is nearly 40x lower at both wavelengths if the film is deposited using the
spin-coating method as compared to the surface-initiated cationic polymerization method.
The transmission loss values for the spin-coated films are similar to those of other
commonly used optical polymers, such as polystyrene and polymethylmethacrylate,
measured using the same approach and deposited using spin-coating.[26, 27]

Quite
possibly, the higher transmission loss is the result of residual catalyst on the surface.
7.4.2 Thermo-optic coefficient determination
As the refractive index of the device changes, the resonant wavelength changes.
By monitoring changes in the resonant wavelength of the device, changes in the
refractive index can be also determined with 1E-7 resolution demonstrated in Chapter
6.[26]
The thermo-optic coefficient (dn/dT) reflects the change in the refractive index
per unit change in temperature of a material. To determine the thermo-optic coefficient
(dn/dT) of the polymer using the integrated photonic sensor, same sample stage described
in Chapter 6 was used as shown in Figure 7-3 (c) and the position of the resonant
wavelength was monitored and recorded while the temperature of the device was
incrementally increased. A sample measurement is shown in Figure 7-6.
94

Figure 7-6 The dependence of the resonant wavelength on the temperature. The slope of this line is d/dT, which is
directly proportional to dn/dT, the desired quantity.

The measured  value can be converted into a n value using the simple
expression:

(7.2)
where /T is an experimentally measured value and dn
eff
/dT is the effective
thermo-optic coefficient. To determine the dn/dT of the polymer (dn
polymer
/dT), dn
eff
/dT
is expanded [27]:

(7.3)
Then, inserting equation 7.3 into equation 7.2, using the values for () from
the FEM results, and rearranging, it is possible to solve for dn
polymer
/dT with a high
degree of accuracy. To verify that this approach was accurate, the dn/dT of silica was
measured. The results are shown below in Figure 7-7, and the calculated value was 1.2E-
5 C
-1
. The conventionally used value for the dn/dT of silica is 1.19E-5 C
-1
.
95

Figure 7-7 The dn/dT of silica measured using the optical sensor. The value of the slope is 1.2E-5 C
-1
, which
corresponds to the dn/dT of silica.

Figure 7-8 contains the results for the dn/dT measurements. These results,
particularly the results for the polymer which was deposited via spin-coating, are very
surprising, given that the majority of polymers have negative dn/dT values of
approximately -1E-4.[28] Therefore, these values merit some discussion. As described
by Prod’homme, the dn/dT of a material arises from a temperature dependent change in
density, which is determined by two factors, the volumetric thermal expansion coefficient
() and the electronic polarizability temperature coefficient () of the material[29, 30]:

δ
δ

β
δ
δ

β (7.4)

where  is the density and n is refractive index. When the polarizability () is the
greater of the two terms and is positive, the dn/dT of the material is positive. This
behavior is observed in both silica and silicon.[29, 30] However, in polymeric materials,
96

there is a competition between the terms, and the thermal expansion () typically
dominates, resulting in a negative dn/dT.

Figure 7-8 The change in the refractive index as the temperature around the sensor is incrementally increased with
temperature. The thermo-optic coefficient (dn/dT) is the slope of this graph. As can be clearly observed, the optical
behavior of the spun-coat films and the polymer brush films was drastically different and the general behavior was
independent of characterization wavelength, as expected. The films attached using the polymer brush approach
exhibited a negative dn/dT whereas the film deposited using spin-coating demonstrated a positive dn/dT.

In the present case, the deposition method is clearly impacting the relative
magnitudes of the electronic polarizability and the thermal expansion coefficients of the
polyisobutylene thin films. Namely, when the polymer is deposited using spin-coating,
the polarizability term is dominating the dn/dT, whereas when the polymer is grafted to
the sensor, the thermal expansion term dominates the dn/dT of the material. Therefore,
the grafted polyisobutylene displays the classic polymer behavior where the thermal
expansion term is dominant; however, in the spin-coated film the polarizability term is
dominant.
97

We hypothesize that the cause for this transition could be due to differences in
density caused by the two deposition methods. From a theoretical standpoint, this could
be explained by looking at the relative magnitudes of the refractive indices of both films.
Specifically, the refractive index values obtained were 1.40233 and 1.482 for the films
deposited using spin-coating and grafting, respectively. The dependence of the density
() on the refractive index (n) is conventionally described by the Lorentz-Lorenz
equation[31]:

(7.5)

where  is the density of the film, n is the refractive index of the film, R is the
molar refraction and M is the molar mass of the monomer unit. From this equation, it is
clear that as the refractive index increases, the density of the film increases. Therefore,
from these ellipsometry results, one can infer that the grafted layer is significantly denser
than the spun coat layer. From equation (4), it becomes apparent that the magnitude of
the –(n

term will decrease as the density decreases, resulting in an increase in the
overall value of dn/dT.
There are several possible experimental reasons for this difference in densities
between the two films. For example, previous work has shown that ultra-thin, spin-
coated PIB polymer films undergo an effect called “pinning”, where the polymer chains
are irreversibly bound to the substrate. Consequently, the chains are unable to relax, and
even when the polymer film is annealed above the glass transition temperature, it does
not behave like a liquid. This behavior was not observed in polymer brush structures of
the same material.[32] Additionally, entanglement effects in PIB have been extensively
98

studied. Previous research has shown that PIB thin films with high molecular weights
deposited via spin-coating, such as those used in the present work, would be highly
entangled. However, polymer brushes of the same molecular weight have shown a lower
degree of entanglement.[33, 34] Given the combination of the high initiator densities
measured and the previously mentioned effects, it is not surprising that the polymer brush
films could be denser than the spin-coated films. However, it is quite possible that
additional mechanisms are contributing to the observed density difference.
Attachment
Method
Wavelength
[nm]
Transmission
Loss [m
-1
]
dn/dT [C
-1
]
Polymer
Brush
765 1.004E3 -1.13E-2
980 1.551E3 -8.8E-3
Spin-Coat 765 24.24 1.8E-4
980 48.14 1.1E-4

Table 7- 2 The optical properties of the polyisobutylene ultra thin films

7.5 Results and discussions
In this chapter, we used an integrated photonic sensor to characterize the
transmission loss and thermo-optic coefficient of ultra thin films of polyisobutylene,
deposited on silica surfaces using either spin-coating or surface initiated cationic
polymerization. All characterization measurements were performed at room temperature
under ambient conditions. The optical loss of both types of polymer films increased
slightly as the testing wavelength transitioned from the visible to the near-IR due to the
increased overtone absorption of the C-H. While the thermo-optic coefficients of the
grafted films behaved in the classic manner, the thermo-optic coefficient of the spin-
99

coated film was positive, in part, we hypothesize, due to a lower density. Additionally,
the new characterization platform that was developed to perform these complex
measurements will enable future generations of measurements on similar complex
polymeric materials. This work represents an initial effort to characterize many of the
basic material properties of this very unusual polymeric material, enabling its integration
into a wide range of applications.[2, 14, 17]

100

Chapter 7 References
1. Park, J.Y., N.R. Hendricks, and K.R. Carter, Solvent-Assisted Soft Nanoimprint
Lithography for Structured Bilayer Heterojunction Organic Solar Cells.
Langmuir, 2011. 27(17): p. 11251-11258.

2. Ma, H., A.K.Y. Jen, and L.R. Dalton, Polymer-based optical waveguides:
Materials, processing and devices. Advanced Materials, 2002. 14(19): p. 1339-
1365.

3. Yuan, Y.B., et al., Efficiency enhancement in organic solar cells with ferroelectric
polymers. Nature Materials, 2011. 10(4): p. 296-302.

4. Tang, C.B., et al., Evolution of block copolymer lithography to highly ordered
square arrays. Science, 2008. 322(5900): p. 429-432.

5. Huang, P.-T., Y.-S. Chang, and C.-W. Chou, Preparation of porous poly(3-
hexylthiophene) by freeze-dry method and its application to organic
photovoltaics. Journal of Applied Polymer Science, 2011. 122(1): p. 233-240.

6. Hitrik, M., et al., Preparation and Characterization of Mono- and Multilayer Films
of Polymerizable 1,2-Polybutadiene Using the Langmuir–Blodgett Technique.
Langmuir, 2011.

7. Bernardini, C., et al., Polymers at the Water/Air Interface, Surface Pressure
Isotherms, and Molecularly Detailed Modeling. Langmuir, 2010. 26(14): p.
11850-11861.

8. Sajeev, U., et al., On the optical and electrical properties of rf and a.c. plasma
polymerized aniline thin films. Bulletin of Materials Science, 2006. 29(2): p. 159-
163.

9. Kunal, K., et al., Polyisobutylene: A most unusual polymer. Journal of Polymer
Science Part B-Polymer Physics, 2008. 46(13): p. 1390-1399.

10. Puskas, J.E. and G. Kaszas, Polyisobutylene-based thermoplastic elastomers: A
review. Rubber Chemistry and Technology, 1996. 69(3): p. 462-475.

11. Puskas, J.E. and Y. Chen, Biomedical Application of Commercial Polymers and
Novel Polyisobutylene-Based Thermoplastic Elastomers for Soft Tissue
Replacement†. Biomacromolecules, 2004. 5(4): p. 1141-1154.

101

12. Chan, N., M.F. Cunningham, and R.A. Hutchinson, Continuous Controlled
Radical Polymerization of Methyl Acrylate in a Copper Tubular Reactor.
Macromolecular Rapid Communications, 2011. 32(7): p. 604-609.

13. Mueller, L., et al., Synthesis of high molecular weight polystyrene using AGET
ATRP under high pressure. European Polymer Journal, 2011. 47(4): p. 730-734.

14. Kunal, K., et al., Polyisobutylene: A most unusual polymer. Journal of Polymer
Science Part B: Polymer Physics, 2008. 46(13): p. 1390-1399.

15. Ferry, J., Viscous Properties of Polyisobutylene. Physics, 1935. 6(11): p. 356.

16. Billmeyer, F.W., Textbook of Polymer Science. 3rd ed. 1984: John Wiley &
Sons.

17. Ayres, N., Polymer brushes: Applications in biomaterials and nanotechnology.
Polymer Chemistry, 2010. 1(6): p. 769-777.

18. Zhao, B. and W.J. Brittain, Synthesis of Polystyrene Brushes on Silicate
Substrates via Carbocationic Polymerization from Self-Assembled Monolayers.
Macromolecules, 1999. 33(2): p. 342-348.

19. Binder, W.H., et al., Grafting Polyisobutylene from Nanoparticle Surfaces:
Concentration and Surface Effects on Livingness. Macromolecules, 2009. 42(19):
p. 7379-7387.

20. Vidal, A., A. Guyot, and J.P. Kennedy, Silica-grafted polyisobutylene and butyl
rubber. Polymer Bulletin, 1980. 2(5): p. 315-320.

21. Parvole, J., et al., Initiator-Grafted Silica Particles for Controlled Free Radical
Polymerization: Influence of the Initiator Structure on the Grafting Density.
Macromolecular Rapid Communications, 2003. 24(18): p. 1074-1078.

22. Lobau, J., et al., Adsorption of alkyl-trichlorosilanes on glass and silicon: a
comparative study using sum-.frequency spectroscopy and XPS. Thin Solid
Films, 1996. 289: p. 272-281.

23. Choi, H.-S., X. Zhang, and A.M. Armani, Hybrid Silica-Polymer Ultra-High-Q
Microresonators. Optics Letters, 2010. 35(4): p. 459-461.

24. Oxborrow, M., How to simulate the whispering-gallery-modes of dielectric
microresonators in FEMLAB/COMSOL - art. no. 64520J. Laser Resonators and
Beam Control IX, 2007. 6452: p. J4520-J4520.
102

25. Little, B.E., J.P. Laine, and H.A. Haus, Analytic theory of coupling from tapered
fibers and half-blocks into microsphere resonators. Journal of Lightwave
Technology, 1999. 17(4): p. 704-715.

26. Choi, H.S., S. Ismail, and A.M. Armani, Studying polymer thin films with hybrid
optical microcavities. Optics Letters, 2011. 36(12): p. 2152-2154.

27. Choi, H.-S. and A.M. Armani, Thermal non-linear effects in hybrid optical
microresonators. Applied Physics Letters, 2010. 97(22): p. 223306.

28. Zhang, Z., et al., Thermo-optic coefficients of polymers for optical waveguide
applications. Polymer, 2006. 47(14): p. 4893-4896.

29. Diemeer, M.B.J., Polymeric thermo-optic space switches for optical
communications. Optical Materials, 1998. 9: p. 192-200.

30. Pokrass, M., Z. Burshtein, and R. Gvishi, Thermo-optic coefficient in some
hybrid organic/inorganic fast sol-gel glasses. Optical Materials, 2010. 32: p. 975-
981.

31. Reichelt, S., et al., Functionalization of solid surfaces with hyperbranched
polyesters to control protein adsorption Colloids and Surfaces B: Biointerfaces,
2009. 69(2): p. 169-177.

32. Fujii, Y., et al., Shear Modulus of a Polymer Brush. Macromolecules, 2010.
43(9): p. 4310-4313.

33. Charlesby, A. and B.J. Bridges, NMR Measurement of Entanglement Density in
Irradiated Polyisobutylene. Radiation Physics and Chemistry, 1982. 19(2): p. 155-
165.

34. Plazek, D.J., X.D. Zheng, and K.L. Ngai, Viscoelastic Properties of Amorphous
Polymers. 1. Different Temperature Dependences of Segmental Relaxation and
Terminal Dispersion. Macromolecules, 1992. 25(19): p. 4920-4924.

103

Chapter 8 Measuring interface effects in the glass transition
temperature of ultra thin polymer films

8.1 Introduction
The fundamental behavior of polymeric materials at the nanoscale and at
interfaces is markedly different from the bulk behavior. One simple example is the glass
transition temperature, which, depending on the interface, can either increase or decrease
with the thickness of the polymer thin film. However, despite the recent advances in
computational methods which have enabled simulations of many of these effects, it has
not been possible for experimentalists to provide complementary data in all cases, in
large part due to instrumentation limitations. One of the challenges is fundamental to the
region of the polymer film of interest: the substrate-polymer interface. Currently, the
majority of methods study polymer films from the air interface inward. As a result, the
recorded signal has decreased significantly by the time it has reached the polymer-
substrate boundary. To address this challenge, we have developed a new method based
on an integrated photonic sensor and used it to characterize the glass transition
temperature of polystyrene ultra thin films (<150nm) at different regions within the film.
Unlike previous methods, the detection signal is generated at the polymer-substrate
interface, enabling high resolution characterization of polymer behavior in this region.
Additionally, by changing the operational wavelength of the device, we are able to
selectively interrogate different regions of the film, fully characterizing the polymer thin
film from the polymer-substrate interface to the air-polymer interface. At these different
104

locations within the film, we have detected markedly different glass transition
temperatures, spanning nearly a 30
°
C range at some film thicknesses, indicating that
layers with distinctly different behavior would exist within the film. However, there is
some discrepancy between the absolute values and results obtained using alternative
methods; therefore, it is possible that there is some trapped solvent in the films. This is
currently under investigation.

8.2 Background & Motivation
Ultra-thin polymer films are widely used throughout the electronics, energy, and
defense industries.[1-4] For example, new approaches to lithographic patterning based on
self-assembly rely on understanding their thermodynamic behavior and interface
interactions, and polymer-based solar cells are an emerging technology which shows
significant promise.[2, 5] While the chemical and physical properties of the polymers
used in these applications vary greatly, many of the fundamental characterization
experiments are similar. For example, researchers typically determine the glass transition
temperature (T
g
), which describes the temperature at which a n amorphous polymer
undergoes a transition from a solid to a rubbery or viscous liquid.[6, 7] Clearly, this
temperature determines the suitability of a given polymer for many applications.
However, the T
g
of supported ultra-thin polymer films is markedly different from
the bulk behavior.[6, 8, 9] Specifically, the glass transition temperature (T
g
) is dependent
on both the thickness of the film and its interaction with the supporting substrate.[10]
Additionally, because different measurement methods probe different regions within a
105

given film, the measured T
g
values for the same film-substrate system vary greatly. Since
the first observation of these effects, numerous experimental methods have been applied
to both free and supported thin films in an effort to understand this behavior, and
complimentary theoretical models have been developed [10-19]. For example,
researchers have concluded that if there is a strong interaction with the substrate, the T
g

increases as the film thickness decreases, whereas the T
g
decreases in free standing films
[17-19].
Recent studies have begun to investigate the concept of multiple glass transition
temperatures inside a film, suggesting the change of mobilities within the films
depending on the locations. For example, by embedding an intermediate silica slide to
improve the sensitivity of Raman microspectrometry, researchers have shown that there
are two glass transition temperatures in supported PMMA thin films [20]. It is proposed
that the multiple temperatures arise from different interactions within the film.
Specifically, the lower T
g
is the result of inter/intra-chain interactions with enhanced
mobility which are similar to the bulk materials, and the higher T
g
is the result of
increased polymer-surface interactions. More recently, using multi-wavelength
ellipsometry, researchers have observed a similar phenomenon with PMMA films
supported on activated silicon surfaces by observing the change in the of polarization of
the incident light[21].
Theoretical models of this behavior have been developed which range from a 2-
layer system to an N-layer film.[10, 19, 21-23] The 2-layer model includes a dead or
rigid layer which has an expansion coefficient of approximately zero. This layer arises
106

from the pinning of the polymer to the substrate, and typically is approximated as 2-3nm
thick for amorphous polymers.[22] The behavior of the second layer, which includes the
polymer-air interface, is dominated by the increased chain mobility as the temperature is
increased. For ultra thin films (<~30nm), previous research has shown that the 2-layer
model accurately captures the polymer film behavior.[22] In contrast, for thicker films, a
multi-layer model is necessary to further distinguish the intermediate regions of the film
from the air-film interface.[10, 21]
Therefore, it is important to be able to study the entire film. However, one
limitation with the currently available both of these methods, such as Raman
microspectrometry and dielectric relaxation spectroscopy, is that they interrogate the
polymer film from the air-film interface. Because the detection signal decreases with
film depth, the understanding of this multi-layer behavior the polymer-substrate
interaction is limited, unless high intensity energy sources are used or intermediate
surfaces, such as fluorescent dye molecules[24, 25] or silica slides[24, 25], are embedded
within the film to enhance the signal. By interrogating the film at the substrate-film
interface, the sensitivity to changes in this location will be increased, hence enabling the
possible approach for further understanding of polymer’s thin film dynamics in nm -
resolution.
One approach to solving this challenge is to combine the refractive index
sensitivity of ellipsometry with integrated resonant cavity-based sensors (Figure 8-1). In
the present work, using this ultra-sensitive technique, we first develop a theoretical model
for this detection platform, and then we perform finite element method simulations to
107

better understand the interactions between the optical field and the ultra-thin polymer
film. Finally, we experimentally determine glass transition temperature at different
locations within a polystyrene film by changing the interrogation wavelength. The
different wavelengths change the optical field distribution in the sensor, thereby changing
the incident angle into the polymer thin films. As a result, the sensor is able to
interrogate from the substrate-polymer interface to the air-polymer interface.
Additionally, we vary the film thickness and explore how this impacts the T
g
.

Figure 8-1 A scanning electron microscope image of the fabricated silica optical micro-sensor integrated on a silicon
substrate. Inset: A rendering of the silica sensor in operation. The light is partially confined within the device, but also
interacts with the polymer film coating enabling detection of the polymer film behavior.

8.3 Thoery
As mentioned, in solid-supported ultra-thin polymer films, there is a strong
dependence of the glass transition temperature on the polymer film thickness due to
strong interface effects. One recently developed predictive model for the dependence of
the glass transition temperature on the film thickness incorporates both the mobility of
108

the polymer within the film and the interface effects through the following expression
[26]:

(8.1)
where T
g,
is the glass transition temperature of the bulk material, t is the film
thickness, k is a measure of the interaction strength between the polymer film and the
substrate, and  is a material-dependent fitted parameter which is related to the segmental
polymer chain length. For free standing films, k is set equal to zero. Therefore,
depending on the nature of the substrate-film interaction and the polymer, dramatically
different behavior can be expected.
One method of determining the T
g
of a film is detecting changes in refractive
index as a function of temperature. The change is conventionally described by the
Lorentz-Lorenz equation[27] as introduced in Chapter 7:

(8.2)

Therefore, as a film is heated, the density decreases, thereby changing the
refractive index. This will enable us to understand how the density of the polymer thin
film changes with temperature variation by observing the change in refractive index. At
T
g
, there is a discontinuity in this otherwise gradual process due to an abrupt change of
the density of the polymer. This change causes a shift in the slope of the refractive index.
Therefore, by accurately measuring the change in refractive index (dn/dT) and
determining the second derivative (d
2
n/dT
2
), the T
g
can be determined very precisely.
The precision of this measurement relies on the accuracy of measuring the change
in the refractive index. In the present platform, the primary detection signal is a change
109

in the resonant wavelength (), which is directly related to the change in the effective
refractive index though the following expression as described in previous chapters [28]:
Δ=
o
(R/R+ Δn/n)=(
eff
+(dn
eff
/dT)/n
eff
) 
o
T (8.3)

Using Finite Element Method (FEM) simulations, we determine the optical field
decay length, the % of the field in the polymer film and the mode area. All three
parameters provide insight into the interaction of the optical field with the polymer film
and its ability to detect various aspects of the substrate-polymer behavior.
An example simulation showing the interaction of the optical field with the
substrate, polymer film and air is shown in Figure 8-2a with 100nm polystyrene coating
on SiO
2
substrate at λ=765nm. The optical field is clearly able to interrogate the entire
polymer film, with the strongest interaction occurring at the substrate-film interface. A
compilation of several simulations in which film thickness and wavelength were
simultaneously varied is shown in Figure 8-2b. As can be seen, the optical field in the
polymer thin film is highly dependent on interrogation wavelength and the film thickness.
Therefore, it would be possible to control the interrogation region by controlling these
parameters to study different areas within the thin films.
110

Figure 8-2 (a) An example simulation showing the interaction of the optical field with the substrate, polymer film and
air. The sensor size is 40(8)m of major(minor) diameter and film thickness is 100nm. The operating wavelength is
765nm. As can be seen, the optical field is clearly able to easily interrogate the entire polymer film, with the strongest
interaction occurring at the substrate-film interface. (b) FEM simulation results with a range of film thicknesses at
different probing wavelengths. Sensor size is 40(8)m of major(minor) diameter and film thickness. The results show
that different locations in polymer films can be probed depending on film thicknesses and probing wavelengths without
embedding additional materials in the film.

From the Figure 8-2b, we can get the relationship between the fraction in polymer
() and polymer film thickness and an example expression at =980nm is below:

(8.4)

By considering only thermo-optic effect for simplicity, equation 8.3 can be
expressed as following:

P    
6
10 0306 . 2 
111

(8.5)

In Figure 8-3, a set of the calculated results based on equation 8.5 is shown. The
left one shows the results at =980nm with 100, 200 and 250nm thick PS and the right
one shows the results at =635, 850, 980 and 1550nm with 250nm thick PS.

Figure 8-3 The calculated results showing the relationship between the fraction in polymer () and polymer film
thickness or resonant shifts (a) with 100, 200, 250nm thick PS at =980nm and (b) with 250nm thick PS at =635, 850,
980 and 1550nm.

In Figure 8-4, a set of FEM simulations results is shown for 0, 20, 50, 100, and
150nm thick polystyrene coated optical microresonators at =765 and 1330nm with a
device size of 40(8)m major(minor) diameter. As can be seen, the optical field shifts
towards the polymer layer as the film thickness increases. By comparing both graphs,
there is a stronger field confinement in the polymer layer at 765nm than at 1330nm, for
the same film thickness. However, the evanescent tail is much longer at 1330nm.
Therefore, through judicious selection of the testing wavelength, researchers could be
able to study different regions of the polymer film.






















  
 
Silica polymer Coated
o
s eff
dT
dn
dT
dn
n

 

) (
112

Figure 8-4 The normailized radial optical field intensity as a function of radius for 0 (dashed black line), 20 (solid black
line), 50 (dashed red line), 100 (dotted blue line), 150nm (dash-dotted green line) thick polystyrene coated optical
microresonators at (a) =765 and (b) 1330nm. The zero point indicates that the surface of the silica (SiO
2
) sensor.

However, to fully understand this device’s sensing behavior, it is necessary to
compare the total interaction of the optical field with the polymer film. While the film
penetration depth is one critical parameter, it does not fully capture this system.
Therefore, to develop a quantitative model, it is necessary to investigate not just a single
axial direction (shown in Figure 8-4), but a cross section of the optical field overlap with
the polymer film. The cross section can be determined by calculating the optical mode
area of the device. From this simulation, the portion of the optical mode in the polymer
film can be determined.
To understand how the effective mode area (A
eff
) is related to the sensor geometry
and interrogation wavelengths, following expression is used to calculate the effective
mode area [29]:

(8.6)

113

Figure 8-5a shows the total effective mode area as a function of polystyrene
thickness at λ=765, 980 and 1330nm. As can be seen, there is a significant difference in
the total effective mode area among the different interrogation wavelengths. As expected,
the mode area is larger with longer wavelengths than with shorter wavelengths. In
addition, as the film thickness changes, the mode area also changes. This effect is
especially apparent at short wavelengths where the mode area is significantly reduced as
the film thickness increases.

Figure 8-5 Effective mode area as a function of polymer film thickness (20, 50, 100, 150nm) at λ=765, 980 and
1330nm with (a) total mode area (silica, polymer, air) and (b) mode area only in polymer.

A more direct comparison can be made by considering only the effective mode
area in the polymer film. Figure 8-5b shows the effective modes area in the polymer
region as a function of polystyrene thickness at all interrogation wavelengths that were
studied. As can be seen, the mode area is larger at longer wavelength than at shorter
wavelength in the polymer region and the variation of the mode area is also larger at
longer wavelength as the polymer thickens. Additionally, the mode area changes
depending on the interrogation wavelength at same thickness. This confirms that it is
114

possible to control probe area by changing the wavelength, hence enabling the detection
of different T
g
within the polymer film.

8.4 Experimental results and discussions
It is important to note that silica is inherently hydrophilic while silicon is
inherently hydrophobic. However, silicon rapidly grows a thin native oxide upon
exposure to air. Therefore, the entire sensor device is hydrophilic, to different degrees.
As a result, no subsequent processing steps, such as an O
2
plasma treatments or piranha
etches, were necessary in order to deposit uniform thin films of polystyrene.
Light was coupled into the sensor from a series of continuous wavelength narrow
linewidth lasers (765, 980, 1300nm) using single-mode tapered optical fiber waveguides.
The input power into the sensor was controlled by using an attenuator to avoid any
unwanted non-linear effects, ensuring minimal distortion of the resonant lineshape. The
scan speed and scan range of the laser were held fixed at 100Hz and 2V for all
experiments.
Initial testing was carried out with silica microtoroid without a polymer coating at
discrete temperature range. An example of testing results is shown in Figure 8-6a and the
testing results for silica microtoroid is shown in Figure 8-6b.

115

Figure 8-6 (a) An example of testing results for silica microtoroid and (b) testing results showing d/dT as a function of
temperature at 20, 40, 60 and 80°C.

As can expected d/dT shows pretty constant over wide range of temperature
variation for silica microtororid and this is a reference data set that can be used to explain
the difference between monolithic silica microtoroid and hybrid microtoroid, hence
enabling interpret polymer behavior based on experimental results.
A 250nm thick PS was shown for initial test of hybrid devices as this thickness
shows quite difference optical response compared to silica microtoroid and the result is
shown in Figure 8-7.

Figure 8-7 Experimental (black square dot) and theoretical (red triangle dot) results showing d/dT as a function of
temperature at 20, 40, 60 and 80°C.
116

As can be seen from the results in Figure 8-7, there seems to be a discrete
behavior between 60 and 80°C, but it can not be sure as it is not continuous scanning of
thermal history of the coated polymer and it would be also possible we look at different
optical modes. In addition for this measurement method, we should always find a
fundamental mode to keep track of d/dT, which means need to test any possible
candidate of resonant peaks due to complex mode behaviors of microtroidal optical
resonators. Therefore, we decide to find stable peaks (ideally within several hundred pm
variation up to around 100°C) and keep track of the peaks for continuous thermal history
of the coated polymer.
With this new approach, initial thickness is chosen to be 100nm thick PS as the
resonant peak is stable with this thickness at =980nm. To check the feasibility and
reproducible, the testing was performed several times for same temperature range and the
results are shown in Figure 8-8.

117

Figure 8-8 The feasibility and reproducibility test for continuous thermal history test.

As can be seen, the forward and backward scans show similar trend both for
resonant shift and quality factor, indicating that the data is reliable. Additional data
around T
g
are also taken to see whether there is material degradation or not and the
results are shown in Figure 8-9.
118

Figure 8-9 Iteration test around Tg to see material degradation by observing optical response.

Using the equation 8.3 and the results from the simulations and experiments, the
experimentally measured values for  were converted to values for n. Figure 8-10
shows the experimental results which were used to determine the T
g
of the films at
1330nm with 20 and 150nm thick PS coated hybrid devices. Similar sets of
measurements were performed for all other combinations of wavelengths (765, 980 and
1330nm) and film thicknesses (20, 50, 100 and 150nm).
As can be seen in Figure 8-10a, there is an abrupt change in the response at
91.25°C, indicating that there is a discontinuity of the optical and thermal properties of
the polystyrene thin film. Therefore, T
g
is defined at this abrupt change from the data
analysis.[15, 17, 30]

119

Figure 8-10 Optical response as a function of temperature with (a) 20nm and (b) 150nm thick PS coated hybrid device
tested at =1330nm. and (c) glass transition temperature for all film thicknesses tested as a function of temperature.

In addition, in the beginning of the T
g
measurement testing, we performed this
experiment iteratively (an example is shown in Figure 8-9, by thermally cycling the film
across the T
g
, to determine if there was any degradation in the film. The location of the
T
g
does not change over several iterations, indicating that there is no degradation of the
film and that the measurement method is not destructive.[28] By compiling a series of
measurements performed with different film thicknesses at  = 1330nm, we are able to
120

form a composite graph (Figure 8-10c) which shows that T
g
decreases as the film
thickness increases. This change is because the chain mobility is suppressed with thinner
films while chain mobility is enhanced with thicker films, hence decreasing the glass
transition temperature.[31-34]
Another example result, which clearly shows the discontinuity at the T
g
due to the
change in the refractive index of the polymer film, are shown in Figure 8-11(a). In this
experiment results, we see that slightly different T
g
are detected for the 980 and 1330nm
interrogation wavelengths with the same 100nm polystyrene thin film. The refractive
index change exhibits a linear response before and after the T
g
with an abrupt change in
the slope at the T
g
for both wavelengths due to the discontinuity of refractive index.
Again, this change is related to the change in the mobility of the polymer films at T
g
.
While the T
g
is 75.7°C at λ=980nm, it is 68.8°C at λ=1330nm, indicating that glass
transition temperature is strongly dependent on interrogation wavelengths due to different
incident angles and probe depths. This variation suggests that there are different layers
within the film that have different mobilities. Therefore, this method could be used to
investigate how the polymer’s properties differ in the different regions within a polymer
thin film.
All of the experimentally measured glass transition temperature values, including
those in Figure 8-11 (a), are compiled in Figure 8-11 (b). This data was fit to equation
(1). T
g
,

and  were set according to previously established experimental values for this
system [26, 35], and k was determined from the fit to the data. The k values were [5, 7,
121

9] for [765, 980, 1330] nm. Therefore, this single graph provides a wealth of
information, both about the substrate-polymer behavior and the measurement method.

Figure 8-11 (a) Example testing results with 100nm thick polystyrene film thickness deposited on silica microtoroid at
=980 and 1330nm showing the change in the effective refractive index of the polymer as a function of temperature.
The T
g
point at 75.7°C at =980 and at 68.8°C =1330 are clearly identifiable. (b) A compilation of all of the
experimentally measured data which shows the glass transition temperature as a function of temperature with three
different wavelengths (765, 980, 1330nm). As can be seen, the T
g
point is strongly dependent on the film thickness and
the measurement wavelength.

First, the T
g
is strongly dependent on the measuring wavelength and the polymer
film thickness, verifying that the sensor is able to isolate specific regions of the film by
using different wavelengths. For example, when using the same wavelength, the T
g
decreases as the polystyrene film thickness increases. Similarly, if different wavelengths
are used to study a single film thickness, distinctly different T
g
values are determined.
This variation suggests that there are strong interface effects in polystyrene films on silica
surfaces, which is further confirmed by the magnitude of the k values, which indicate that
silica is a favorable substrate for polystyrene.[21, 26] The increase in the k value which
trends with the wavelength is attributed to the increase in the sampling region.
122

In studying the 765nm measurements in Figure 8-11 (b) more closely, it becomes
apparent that the T
g
values are nearly constant across the film thickness range. The
simulations suggest that measurements at this wavelength only interrogate the region of
the film which is closest to the polymer-substrate interface, and therefore, one would
expect this small degree of variation. In other words, when using this wavelength
(765nm), the overall magnitude of the T
g
with different film thicknesses is smaller than
with the other wavelengths (980 and 1330nm), indicating that the interrogation area and
depth is closer to substrate-polymer interface.
In contrast, when 1330nm wavelength is used, there is a significant variation.
From the simulations, this wavelength should interrogate both the substrate-polymer
interface and intermediate layers of the film; therefore, there should be a strong
dependence on the film thickness. When the film thickness is 20nm, the 1330nm
wavelength is measuring the behavior near polymer-substrate interface, and the value is
similar to that measured at 765nm. However, as the film thickness increases, the T
g

decreases significantly. This behavior is clearly different than that measured with the
765nm laser. Therefore, by using different wavelengths and polymer thin film
thicknesses, it is possible to selectively interrogate different regions of ultra-thin polymer
films, enabling the study of the substrate-polymer interface.
One exception is the 20nm thin film. There is minimal variation in the T
g
across
the different interrogation wavelengths. This indicates that the 20nm film has minimal
mobility within the film or is strongly constrained to the silica substrate. As discussed in
the introduction, this combination of film thickness and substrate falls within the range
123

which can be approximated by a 2-layer model, and as such has fundamentally different
behavior from that observed in thicker films. Specifically, there is reduced mobility
throughout the film, increasing the glass transition temperature.[22]
In Figure 8-12a, the relationship between the glass transition temperature and the
total effective mode area is shown with all the wavelengths and thicknesses. As can be
seen, the T
g
decreases as the mode area becomes larger at same thickness. This is because
the mode area controls the probe areas within the film, hence increasing and decreasing
optical field intensity of the sensor depending on interrogation wavelengths.
This relationship becomes more apparent when considering only the portion of
the effective mode area which resides in the polymer film. In Figure 8-12b, the measured
T
g
values as a function of the mode area in polymer region is shown with all the
wavelengths and film thicknesses. As the amount of the polymer film which is sampled
increases, the measured T
g
decreases, indicating that the T
g
decreases as one moves from
the substrate-polymer interface towards the middle of the film.

Figure 8-12 Glass transition temperature as a function of effective mode area with 20, 50, 100 and 150 film thicknesses
at λ=765, 980 and 1330nm for (a) total mode area (silic a, polymer, air) and (b) mode area only in polymer.

124

However, the glass transition temperature of PS is typically above 100C. Previous
work has shown that trapped solvent can reduce the temperature. In order to completely
remove the solvent, researchers often anneal samples more than 24hours. Therefore,
while the results are definitely promising and indicative that the proposed approach
works, it appears that there might be trapped solvent within the film layer. This potential
issue is currently under investigation.

8.5 Conclusions
In this study, we have developed an ultra-sensitive multi-wavelength, evanescent
field sensor to study the dependence of the glass transition temperature on polymer film
thickness. By measuring the dependence of the T
g
profile within the polymer film, we
have proposed and validated one explanation for the observed variation in measurements.
Specifically, as the thickness in the polymer film increases, the mobility of the polymer
increases as a result of a decrease in interface effects due to substrate interactions. This
increase in mobility results in a decrease in the glass transition temperature.
Additionally, the glass transition temperature is highest at the substrate-polymer interface
and lowest at the air-polymer interface.
We experimentally validated this proposed model by measuring the glass
transition temperature at different locations within a series of polystyrene (PS) ultra-thin
films supported on a silica (SiO
2
) substrate. By varying the operational wavelength of
the sensor, we interrogated different regions of the polymer film, allowing the T
g
within
the polymer film to be measured at different locations, including at the substrate-polymer
125

interface. However, given the large difference between the measured values and
previously reported values, it is likely that there is trapped solvent in the polymer film.
An improved understanding of the thermodynamic properties of polymeric materials at
interfaces will enable advances numerous fields of science and engineering, particularly
in the electronics, energy, and defense industries. [1-4, 36]

126

Chapter 8 References
1. Park, J.Y., N.R. Hendricks, and K.R. Carter, Solvent-Assisted Soft Nanoimprint
Lithography for Structured Bilayer Heterojunction Organic Solar Cells.
Langmuir, 2011. 27(17): p. 11251-11258.

2. Yuan, Y.B., et al., Efficiency enhancement in organic solar cells with ferroelectric
polymers. Nature Materials, 2011. 10(4): p. 296-302.

3. Yoo, J.E., et al., Directly patternable, highly conducting polymers for broad
applications in organic electronics. Proceedings of the National Academy of
Sciences, 2010. 107(13): p. 5712-5717.

4. Laiho, A., et al., Controlling the dimensionality of charge transport in organic
thin-film transistors. Proceedings of the National Academy of Sciences, 2011.

5. Tang, C.B., et al., Evolution of block copolymer lithography to highly ordered
square arrays. Science, 2008. 322(5900): p. 429-432.

6. Frank, C.W., et al., Structure in thin and ultrathin spin-cast polymer films.
Science, 1996. 273(5277): p. 912-915.

7. Becker, J.S., et al., Comparative surface dynamics of amorphous and
semicrystalline polymer films. Proceedings of the National Academy of Sciences,
2011. 108(3): p. 977-982.

8. Grohens, Y., et al., Glass transition of stereoregular poly(methyl methacrylate) at
interfaces. Langmuir, 1998. 14(11): p. 2929-2932.

9. Napolitano, S. and M. Wübbenhorst, The lifetime of the deviations from bulk
behaviour in polymers confined at the nanoscale. Nature Communications, 2010.
2: p. 260.

10. Clough, A., et al., Glass Transition Temperature of Polymer Films That Slip.
Macromolecules, 2011. 44(6): p. 1649-1653.

11. Mansfield, K.F. and D.N. Theodorou, Molecular dynamics simulation of a glassy
polymer surface. Macromolecules, 1991. 24(23): p. 6283-6294.

12. Wallace, W.E., J.H. van Zanten, and W.L. Wu, Influence of an impenetrable
interface on a polymer glass-transition temperature. Physical Review E, 1995.
52(4): p. R3329-R3332.

127

13. Xie, L., et al., Positronium Formation as a Probe of Polymer Surfaces and Thin
Films. Physical Review Letters, 1995. 74: p. 4947-4950.

14. Baschnagel, J. and K. Binder, On the influence of hard walls on structural
properties in polymer glass simulation. Macromolecules, 1995. 28(20): p. 6808-
6818.

15. Keddie, J.L., R.A.L. Jones, and R.A. Cory, Size-dependent depression of the
glass-transition temperature in polymer-films. Europhysics Letters, 1994. 27(1):
p. 59-64.

16. Kritikos, G. and A.F. Terzis, Theoretical investigation of polymers near surface of
various molecular weights, architecture and external parameters by mean-field
variable-density model. Polymer, 2009. 50(22): p. 5314-5325.

17. Forrest, J.A., et al., Effect of free surfaces on the glass transition temperature of
thin polymer films. Physical Review Letters, 1996. 77(10): p. 2002-2005.

18. Torres, J.A., P.F. Nealey, and J.J. de Pablo, Molecular simulation of ultrathin
polymeric films near the glass transition. Physical Review Letters, 2000. 85(15):
p. 3221-3224.

19. Keddie, J.L., R.A.L. Jones, and R.A. Cory, Interface and surface effects on the
glass-transition temperature in thin polymer films. Faraday Discussions, 1994. 98:
p. 219-230.

20. Vignaud, G., et al., Multiple Glass-Transition Temperatures in Thin Supported
Films of Isotactic PMMA as Revealed by Enhanced Raman Scattering. Langmuir,
2005. 21(19): p. 8601-8604.

21. El Ouakili, A., et al., Multiple glass transition temperatures of polymer thin films
as probed by multi-wavelength ellipsometry. Thin Solid Films, 2011. 519(6): p.
2031-2036.

22. DeMaggio, G.B., et al., Interface and surface effects on the glass transition in thin
polystyrene films. Physical Review Letters, 1997. 78(8): p. 1524-1527.

23. Erber, M., et al., Polystyrene with different topologies: Study of the glass
transition temperature in confined geometry of thin films. European Polymer
Journal, 2010. 46(12): p. 2240-2246.

24. Ellison, C.J. and J.M. Torkelson, The distribution of glass-transition temperatures
in nanoscopically confined glass formers. Nature Materials, 2003. 2: p. 695-700.
128

25. Rotella, C., et al., Distribution of Segmental Mobility in Ultrathin Polymer Films.
Macromolecules, 2010. 43(20): p. 8686-8691.

26. Ahn, S.I., et al., Glass transition temperature of polymer nanocomposites:
Prediction from the continuous-multilayer model. Journal of Polymer Science B,
2009. 47: p. 2281-2287.

27. Reichelt, S., et al., Functionalization of solid surfaces with hyperbranched
polyesters to control protein adsorption Colloids and Surfaces B: Biointerfaces,
2009. 69(2): p. 169-177.

28. Choi, H.S., S. Ismail, and A.M. Armani, Studying polymer thin films with hybrid
optical microcavities. Optics Letters, 2011. 36(12): p. 2152-2154.

29. Kippenberg, T.J., S.M. Spillane, and K.J. Vahala, Kerr-nonlinearity optical
parametric oscillation in an ultrahigh-Q toroid microcavity. Physical Review
Letters, 2004. 93(8).

30. Wallace, W.E., J.H. Vanzanten, and W.L. Wu, INFLUENCE OF AN
IMPENETRABLE INTERFACE ON A POLYMER GLASS-TRANSITION
TEMPERATURE. Physical Review E, 1995. 52(4): p. R3329-R3332.

31. Ao, Z.M. and Q. Jiang, Size effects on miscibility and glass transition temperature
of binary polymer blend films. Langmuir, 2006. 22(3): p. 1241-1246.

32. Priestley, R.D., et al., Structural relaxation of polymer glasses at surfaces,
interfaces and in between. Science, 2005. 309(5733): p. 456-459.

33. El Ouakili, A., et al., Multiple glass transition temperatures of polymer thin films
as probed by multi-wavelength ellipsometry. Thin Solid Films. 519(6): p. 2031-
2036.

34. Tanaka, K., et al., Interfacial Mobility of Polymers on Inorganic Solids. Journal of
Physical Chemistry B, 2009. 113(14): p. 4571-4577.

35. Weber, M.J., Handbook of Optical Materials. 2003, Boca Raton, FL: CRC Press.

36. Chen, W., H.B. Lu, and S.R. Nutt, The influence of functionalized MWCNT
reinforcement on the thermomechanical properties and morphology of epoxy
nanocomposites. Composites Science and Technology, 2008. 68(12): p. 2535-
2542.

129

Bibliography

Ahn, S. I., C. W. Ohk, et al. (2009). "Glass transition temperature of polymer
nanocomposites: Prediction from the continuous-multilayer model." Journal of Polymer
Science B 47: 2281-2287.

Ao, Z. M. and Q. Jiang (2006). "Size effects on miscibility and glass transition
temperature of binary polymer blend films." Langmuir 22(3): 1241-1246.

Armani, A. M., R. P. Kulkarni, et al. (2007). "Label-free, single-molecule detection with
optical microcavities." Science 317(5839): 783-787.

Armani, A. M., A. Srinivasan, et al. (2007). "Soft lithographic fabrication of high Q
polymer microcavity arrays." Nano Letters 7(6): 1823-1826.

Armani, D. K., T. J. Kippenberg, et al. (2003). "Ultra-high-Q toroid microcavity on a
chip." Nature 421(6926): 925-928.

Ayres, N. (2010). "Polymer brushes: Applications in biomaterials and nanotechnology."
Polymer Chemistry 1(6): 769-777.

Baschnagel, J. and K. Binder (1995). "On the influence of hard walls on structural
properties in polymer glass simulation." Macromolecules 28(20): 6808-6818.

Becker, J. S., R. D. Brown, et al. (2011). "Comparative surface dynamics of amorphous
and semicrystalline polymer films." Proceedings of the National Academy of Sciences
108(3): 977-982.

Bernardini, C., S. D. Stoyanov, et al. (2010). "Polymers at the Water/Air Interface,
Surface Pressure Isotherms, and Molecularly Detailed Modeling." Langmuir 26(14):
11850-11861.

Billmeyer, F. W. (1984). Textbook of Polymer Science, John Wiley & Sons.

Binder, W. H., R. Zirbs, et al. (2009). "Grafting Polyisobutylene from Nanoparticle
Surfaces: Concentration and Surface Effects on Livingness." Macromolecules 42(19):
7379-7387.

Brian, A. A. and H. M. McConnell (1984). "ALLOGENEIC STIMULATION OF
CYTO-TOXIC T-CELLS BY SUPPORTED PLANAR MEMBRANES." Proceedings of
the National Academy of Sciences of the United States of America-Biological Sciences
81(19): 6159-6163.
130

Burghard, Z., L. Zini, et al. (2009). "Toughening through Nature-Adapted Nanoscale
Design." Nano Letters 9(12): 4103-4108.

Carmon, T. and K. J. Vahala (2007). "Visible continuous emission from a silica
microphotonic device by third-harmonic generation." Nature Physics 3(6): 430-435.

Carmon, T., L. Yang, et al. (2004). "Dynamical thermal behavior and thermal self-
stability of microcavities." Optics Express 12(20): 4742-4750.

Castellana, E. T. and P. S. Cremer (2006). "Solid supported lipid bilayers: From
biophysical studies to sensor design." Surface Science Reports 61(10): 429-444.

Chan, N., M. F. Cunningham, et al. (2011). "Continuous Controlled Radical
Polymerization of Methyl Acrylate in a Copper Tubular Reactor." Macromolecular Rapid
Communications 32(7): 604-609.

Chao, C. Y. and L. J. Guo (2003). "Biochemical sensors based on polymer microrings
with sharp asymmetrical resonance." Applied Physics Letters 83(8): 1527-1529.

Charlesby, A. and B. J. Bridges (1982). "NMR Measurement of Entanglement Density in
Irradiated Polyisobutylene." Radiation Physics and Chemistry 19(2): 155-165.

Chen, W., H. B. Lu, et al. (2008). "The influence of functionalized MWCNT
reinforcement on the thermomechanical properties and morphology of epoxy
nanocomposites." Composites Science and Technology 68(12): 2535-2542.

Choi, H.-S. and A. M. Armani (2010). "Thermal non-linear effects in hybrid optical
microresonators." Applied Physics Letters 97(22): 223306.

Choi, H.-S., X. Zhang, et al. (2010). "Hybrid Silica-Polymer Ultra-High-Q
Microresonators." Optics Letters 35(4): 459-461.

Choi, H. S., S. Ismail, et al. (2011). "Studying polymer thin films with hybrid optical
microcavities." Optics Letters 36(12): 2152-2154.

Choi, H. S., X. M. Zhang, et al. "Hybrid silica-polymer ultra-high-Q microresonators."
Optics Letters 35(4): 459-461.

Chremmos, I., Schwelb, O., and Uzunoglu, N., Springer Series in Optical Sciences, ed.
W.T. Rhodes. Vol. 156. 2010, New York: Springer.

Clough, A., D. Peng, et al. (2011). "Glass Transition Temperature of Polymer Films That
Slip." Macromolecules 44(6): 1649-1653.

131

Dalton, L. R., P. A. Sullivan, et al. "Electric Field Poled Organic Electro-optic Materials:
State of the Art and Future Prospects." Chemical Reviews 110(1): 25-55.

Davis, L. J. and M. Deutsch (2010). "Surface plasmon based thermo-optic and
temperature sensor for microfluidic thermometry." Review of Scientific Instruments
81(11): -.

DeMaggio, G. B., W. E. Frieze, et al. (1997). "Interface and surface effects on the glass
transition in thin polystyrene films." Physical Review Letters 78(8): 1524-1527.

Diaz, A., S. Kubo, et al. (2007). Tunable refractive index materials with metallic nano-
spheres dispersed in organic liquids. Procs. SPIE.

Diemeer, M. B. J. (1998). "Polymeric thermo-optic space switches for optical
communications." Optical Materials 9: 192-200.

Dong, C. H., L. He, et al. (2009). "Fabrication of high-Q polydimethylsiloxane optical
microspheres for thermal sensing." Applied Physics Letters 94(23): 231119.

El Ouakili, A., G. Vignaud, et al. (2011). "Multiple glass transition temperatures of
polymer thin films as probed by multi-wavelength ellipsometry." Thin Solid Films
519(6): 2031-2036.

Ellison, C. J. and J. M. Torkelson (2003). "The distribution of glass-transition
temperatures in nanoscopically confined glass formers." Nature Materials 2: 695-700.

Erber, M., U. Georgi, et al. (2010). "Polystyrene with different topologies: Study of the
glass transition temperature in confined geometry of thin films." European Polymer
Journal 46(12): 2240-2246.

Fair, R. B. (1997). "Physical Models of Boron Diffusion in Ultrathin Gate Oxides."
Journal of The Electrochemical Society 144(2): 708-717.

Ferry, J. (1935). "Viscous Properties of Polyisobutylene." Physics 6(11): 356.

Forrest, J. A., K. DalnokiVeress, et al. (1996). "Effect of free surfaces on the glass
transition temperature of thin polymer films." Physical Review Letters 77(10): 2002-
2005.

Frank, C. W., V. Rao, et al. (1996). "Structure in thin and ultrathin spin-cast polymer
films." Science 273(5277): 912-915.

Freeman, L. M., S. Li, et al. "Excitation of Cy5 in self-assembled lipid bilayers using
optical microresonators." Applied Physics Letters 98(14).
132

Fujii, Y., Z. Yang, et al. (2010). "Shear Modulus of a Polymer Brush." Macromolecules
43(9): 4310-4313.

Gafni, A., R. L. Modlin, et al. (1975). "Analysis of fluorescence decay curves by means
of the Laplace transformation." Biophysical Journal 15(3): 263-280.

Goedhart, J., L. van Weeren, et al. "Bright cyan fluorescent protein variants identified by
fluorescence lifetime screening." Nature Methods 7(2): 137-U74.

Gorodetsky, M. L., A. D. Pryamikov, et al. (2000). "Rayleigh scattering in high-Q
microspheres." Journal of the Optical Society of America B-Optical Physics 17(6): 1051-
1057.

Gorodetsky, M. L., A. D. Pryamikov, et al. (2000). "Rayleigh scattering in high-Q
microspheres." J. Opt. Soc. Am. B 17(6): 1051-1057.

Gorodetsky, M. L., A. A. Savchenkov, et al. (1996). "Ultimate Q of optical microsphere
resonators." Optics Letters 21(7): 453-455.

Gotzinger, S., L. D. Menezes, et al. (2006). "Controlled photon transfer between two
individual nanoemitters via shared high-Q modes of a microsphere resonator." Nano
Letters 6(6): 1151-1154.

Grohens, Y., M. Brogly, et al. (1998). "Glass transition of stereoregular poly(methyl
methacrylate) at interfaces." Langmuir 14(11): 2929-2932.

Haiss, W., N. T. Thanh, et al. (2007). "Determination of size and concentration of gold
nanoparticles from UV-vis spectra." Anal Chem 79(11): 4215-21.

Hale, G. M. and M. R. Querry (1973). "OPTICAL-CONSTANTS OF WATER IN 200-
NM TO 200-MUM WAVELENGTH REGION." Applied Optics 12(3): 555-563.

Han, M. and A. Wang (2007). "Temperature compensation of optical microresonators
using a surface layer with negative thermo-optic coefficient." Optics Letters 32(13):
1800-1802.

He, L., Y. F. Xiao, et al. (2008). "Compensation of thermal refraction effect in high-Q
toroidal microresonator by polydimethylsiloxane coating." Applied Physics Letters
93(20).

Hitrik, M., V. Gutkin, et al. (2011). "Preparation and Characterization of Mono- and
Multilayer Films of Polymerizable 1,2-Polybutadiene Using the Langmuir–Blodgett
Technique." Langmuir.

133

Hsu, H. S., C. Cai, et al. (2009). "Ultra-low-threshold Er:Yb sol-gel microlaser on
silicon." Optics Express 17(25): 23265-23271.

Huang, P.-T., Y.-S. Chang, et al. (2011). "Preparation of porous poly(3-hexylthiophene)
by freeze-dry method and its application to organic photovoltaics." Journal of Applied
Polymer Science 122(1): 233-240.

Huang, Y. and D. R. Paul (2006). "Physical aging of thin glassy polymer films monitored
by optical properties." Macromolecules 39(4): 1554-1559.

Hunt, H. K. and A. M. Armani (2010). "Label-Free Biological and Chemical Sensors."
Nanoscale 2(9): 1544-1559.

Kasarova, S. N., N. G. Sultanova, et al. (2007). "Analysis of the dispersion of optical
plastic materials." Optical Materials 29(11): 1481-1490.

Keddie, J. L., R. A. L. Jones, et al. (1994). "Interface and surface effects on the glass-
transition temperature in thin polymer films." Faraday Discussions 98: 219-230.

Keddie, J. L., R. A. L. Jones, et al. (1994). "SIZE-DEPENDENT DEPRESSION OF THE
GLASS-TRANSITION TEMPERATURE IN POLYMER-FILMS." Europhysics Letters
27(1): 59-64.

Kietzke, T., D. Neher, et al. (2003). "Novel approaches to polymer blends based on
polymer nanoparticles." Nature Materials 2(6): 408-U7.

Kim, G. D., H. S. Lee, et al. "Silicon photonic temperature sensor employing a ring
resonator manufactured using a standard CMOS process." Optics Express 18(21): 22215-
22221.

Kim, H. K., S. J. Kang, et al. (1999). "Highly efficient organic/inorganic hybrid nonlinear
optic materials via sol-gel process: synthesis, optical properties, and photobleaching for
channel waveguides." Chemistry of materials 11(3): 779-788.

Kippenberg, T. J. (2004). "Nonlinear Optics in Ultra-high-Q Whispering-Gallery Optical
Microcavities." PhD thesis, California Institute of Technology.

Kippenberg, T. J., S. M. Spillane, et al. (2003). "Fabrication and coupling to planar high-
Q silica disk microcavities." Applied Physics Letters 83(4): 797-799.

Kippenberg, T. J., S. M. Spillane, et al. (2004). "Ultralow-threshold microcavity Raman
laser on a microelectronic chip." Optics Letters 29(11): 1224-1226.

134

Kippenberg, T. J., S. M. Spillane, et al. (2004). "Kerr-nonlinearity optical parametric
oscillation in an ultrahigh-Q toroid microcavity." Physical Review Letters 93(8).

Klotz, A., C. Barzen, et al. (1999). "Sensitivity enhancement of transducers for total
internal reflection fluorescence." SPIE International Symposium on Integrated
Optoelectronic Devices, San Jose: 3620-3653.

Knight, J. C., G. Cheung, et al. (1997). "Phase-matched excitation of whispering-gallery-
mode resonances by a fiber taper." Optics Letters 22(15): 1129-1131.

Kovacs, G. T. A. (1998). Micromachined Transducers Sourcebook. New York, McGraw
Hill.

Kritikos, G. and A. F. Terzis (2009). "Theoretical investigation of polymers near surface
of various molecular weights, architecture and external parameters by mean-field
variable-density model." Polymer 50(22): 5314-5325.

Kropka, J. M., V. Pryamitsyn, et al. (2008). "Relation between glass transition
temperatures in polymer nanocomposites and polymer thin films." Physical Review
Letters 101(7).

Kunal, K., M. Paluch, et al. (2008). "Polyisobutylene: A most unusual polymer." Journal
of Polymer Science Part B-Polymer Physics 46(13): 1390-1399.

Kunal, K., M. Paluch, et al. (2008). "Polyisobutylene: A most unusual polymer." Journal
of Polymer Science Part B: Polymer Physics 46(13): 1390-1399.

Laiho, A., L. Herlogsson, et al. (2011). "Controlling the dimensionality of charge
transport in organic thin-film transistors." Proceedings of the National Academy of
Sciences.

Larsson, C., M. Rodahl, et al. (2003). "Characterization of DNA immobilization and
subsequent hybridization on a 2D arrangement of streptavidin on a biotin-modified lipid
bilayer supported on SiO2." Analytical Chemistry 75(19): 5080-5087.

Li, S., P. C. Hu, et al. "Confocal Imaging to Quantify Passive Transport across
Biomimetic Lipid Membranes." Analytical Chemistry 82(18): 7766-7771.

Little, B. E., J. P. Laine, et al. (1999). "Analytic theory of coupling from tapered fibers
and half-blocks into microsphere resonators." Journal of Lightwave Technology 17(4):
704-715.

135

Lobau, J., A. Rumphorst, et al. (1996). "Adsorption of alkyl-trichlorosilanes on glass and
silicon: a comparative study using sum-.frequency spectroscopy and XPS." Thin Solid
Films 289: 272-281.

Ma, H., A. K. Y. Jen, et al. (2002). "Polymer-based optical waveguides: Materials,
processing and devices." Advanced Materials 14(19): 1339-1365.

Malitson, I. H. (1965). "INTERSPECIMEN COMPARISON OF REFRACTIVE INDEX
OF FUSED SILICA." Journal of the Optical Society of America 55(10P1): 1205-&.

Mansfield, K. F. and D. N. Theodorou (1991). "Molecular dynamics simulation of a
glassy polymer surface." Macromolecules 24(23): 6283-6294.

Mueller, L., W. Jakubowski, et al. (2011). "Synthesis of high molecular weight
polystyrene using AGET ATRP under high pressure." European Polymer Journal 47(4):
730-734.

Mundra, M. K., C. J. Ellison, et al. (2006). "Confinement, composition, and spin-coating
effects on the glass transition and stress relaxation of thin films of polystyrene and
styrene-containing random copolymers: Sensing by intrinsic fluorescence." Polymer
47(22): 7747-7759.

Napolitano, S. and M. Wübbenhorst (2010). "The lifetime of the deviations from bulk
behaviour in polymers confined at the nanoscale." Nature Communications 2: 260.

Oxborrow, M. (2007). "How to simulate the whispering-gallery-modes of dielectric
microresonators in FEMLAB/COMSOL - art. no. 64520J." Laser Resonators and Beam
Control IX 6452: J4520-J4520.

Oxborrow, M. (2007). "Traceable 2-D finite-element simulation of the whispering-
gallery modes of axisymmetric electromagnetic resonators." IEEE Transactions on
Microwave Theory and Techniques 55(6): 1209-1218.

Park, J. Y., N. R. Hendricks, et al. (2011). "Solvent-Assisted Soft Nanoimprint
Lithography for Structured Bilayer Heterojunction Organic Solar Cells." Langmuir
27(17): 11251-11258.

Parvole, J., G. Laruelle, et al. (2003). "Initiator-Grafted Silica Particles for Controlled
Free Radical Polymerization: Influence of the Initiator Structure on the Grafting
Density." Macromolecular Rapid Communications 24(18): 1074-1078.

Pei, Y., F. Yao, et al. (2010). "Refractive index of silver nanoparticles dispersed in
polyvinyl pyrrolidone nanocomposite." Journal of Modern Optics 57(10): 872-875.
136

Pinnow, D. A., T. C. Rich, et al. (1973). "Fundamental optical attenuation limits in the
liquid and glassy state with application to fiber optical waveguide materials." Applied
Physics Letters 22(10): 527-529.

Plazek, D. J., X. D. Zheng, et al. (1992). "Viscoelastic Properties of Amorphous
Polymers. 1. Different Temperature Dependences of Segmental Relaxation and Terminal
Dispersion." Macromolecules 25(19): 4920-4924.

Pokrass, M., Z. Burshtein, et al. (2010). "Thermo-optic coefficient in some hybrid
organic/inorganic fast sol-gel glasses." Optical Materials 32: 975-981.

Priestley, R. D., C. J. Ellison, et al. (2005). "Structural relaxation of polymer glasses at
surfaces, interfaces and in between." Science 309(5733): 456-459.

Puskas, J. E. and Y. Chen (2004). "Biomedical Application of Commercial Polymers and
Novel Polyisobutylene-Based Thermoplastic Elastomers for Soft Tissue Replacement†."
Biomacromolecules 5(4): 1141-1154.

Puskas, J. E. and G. Kaszas (1996). "Polyisobutylene-based thermoplastic elastomers: A
review." Rubber Chemistry and Technology 69(3): 462-475.

Rabiei, P., W. H. Steier, et al. (2002). "Polymer micro-ring filters and modulators."
Journal of Lightwave Technology 20(11): 1968-1975.

Reichelt, S., K.-J. Eichhorn, et al. (2009). "Functionalization of solid surfaces with
hyperbranched polyesters to control protein adsorption " Colloids and Surfaces B:
Biointerfaces 69(2): 169-177.

Rokhsari, H. and K. J. Vahala (2004). "Ultralow loss, high Q, four port resonant couplers
for quantum optics and photonics." Physical Review Letters 92(25).

Rotella, C., S. Napolitano, et al. (2010). "Distribution of Segmental Mobility in Ultrathin
Polymer Films." Macromolecules 43(20): 8686-8691.

Sajeev, U., C. Mathai, et al. (2006). "On the optical and electrical properties of rf and a.c.
plasma polymerized aniline thin films." Bulletin of Materials Science 29(2): 159-163.

Schafer, L. V., D. H. de Jong, et al. "Lipid packing drives the segregation of
transmembrane helices into disordered lipid domains in model membranes." Proceedings
of the National Academy of Sciences of the United States of America 108(4): 1343-1348.

Schouten, S., P. Stroeve, et al. (1999). "DNA adsorption and cationic bilayer deposition
on self-assembled monolayers." Langmuir 15(23): 8133-8139.

137

Sepúlveda, B., P. C. Angelomé, et al. (2009). "LSPR-based nanobiosensors." Nano
Today 4(3): 244-251.

Snyder, P. G., M. C. Rost, et al. (1986). "Variable angle of incidence spectroscopic
ellipsometry: Application to GaAs-Al[sub x]Ga[sub 1 - x]As multiple heterostructures."
Journal of Applied Physics 60(9): 3293-3302.

Spillane, S. M. (2004). Fiber-coupled Ultra-high-Q Microresonators for Nonlinear and
Quantum Optics.

Spillane, S. M., T. J. Kippenberg, et al. (2003). "Ideality in a fiber-taper-coupled
microresonator system for application to cavity quantum electrodynamics." Physical
Review Letters 91(4): 043902.

Streetman, B. G. and S. Banerjee (1999). Solid State Electronic Devices, Prentice Hall.

Tanaka, K., Y. Tateishi, et al. (2009). "Interfacial Mobility of Polymers on Inorganic
Solids." Journal of Physical Chemistry B 113(14): 4571-4577.

Tang, C. B., E. M. Lennon, et al. (2008). "Evolution of block copolymer lithography to
highly ordered square arrays." Science 322(5900): 429-432.

Torres, J. A., P. F. Nealey, et al. (2000). "Molecular simulation of ultrathin polymeric
films near the glass transition." Physical Review Letters 85(15): 3221-3224.

Treussart, F., V. S. Ilchenko, et al. (1998). "Evidence for intrinsic Kerr bistability of
high-Q microsphere resonators in superfluid helium." European Physical Journal D 1(3):
235-238.

Tulek, A., D. Akbulut, et al. (2009). "Ultralow threshold laser action from toroidal
polymer microcavity." Applied Physics Letters 94(20): 203302.

Vahala, K. J. (2003). "Optical microcavities." Nature 424(6950): 839-846.

Vendamme, R., S. Y. Onoue, et al. (2006). "Robust free-standing nanomembranes of
organic/inorganic interpenetrating networks." Nature Materials 5(6): 494-501.

Vernooy, D. W., V. S. Ilchenko, et al. (1998). "High-Q measurements of fused-silica
microspheres in the near infrared." Optics Letters 23(4): 247-249.

Vidal, A., A. Guyot, et al. (1980). "Silica-grafted polyisobutylene and butyl rubber."
Polymer Bulletin 2(5): 315-320.

138

Vignaud, G., J.-F. Bardeau, et al. (2005). "Multiple Glass-Transition Temperatures in
Thin Supported Films of Isotactic PMMA as Revealed by Enhanced Raman Scattering."
Langmuir 21(19): 8601-8604.

Vollmer, F. and S. Arnold (2008). "Whispering-gallery-mode biosensing: label-free
detection down to single molecules." Nature Methods 5(7): 591-596.

Wallace, W. E., J. H. van Zanten, et al. (1995). "Influence of an impenetrable interface on
a polymer glass-transition temperature." Physical Review E 52(4): R3329-R3332.

Walsh, C. B. and E. I. Franses (2003). "Ultrathin PMMA films spin-coated from toluene
solutions." Thin Solid Films 429(1-2): 71-76.

Wang, L. L., A. K. Gaigalas, et al. (2005). "Optical properties of Alexa (TM) 488 and Cy
(TM) 5 immobilized on a glass surface." Biotechniques 38(1): 127-132.

Weber, M. J. (2003). Handbook of optical materials, CRC Press.

Weiss, D. S., V. Sandoghdar, et al. (1995). "SPLITTING OF HIGH-Q MIE MODES
INDUCED BY LIGHT BACKSCATTERING IN SILICA MICROSPHERES." Optics
Letters 20(18): 1835-1837.

Xie, L., G. B. DeMaggio, et al. (1995). "Positronium Formation as a Probe of Polymer
Surfaces and Thin Films." Physical Review Letters 74: 4947-4950.

Yang, J., J. Y. Lee, et al. (2003). Langmuir 19(24): 10361.

Yoo, J. E., K. S. Lee, et al. (2010). "Directly patternable, highly conducting polymers for
broad applications in organic electronics." Proceedings of the National Academy of
Sciences 107(13): 5712-5717.

Yuan, Y. B., T. J. Reece, et al. (2011). "Efficiency enhancement in organic solar cells
with ferroelectric polymers." Nature Materials 10(4): 296-302.

Zhang, X., H.-S. Choi, et al. (2010). "Ultimate quality factor of silica microtoroid
resonant cavities." Applied Physics Letters 96(15): 153304.

Zhang, Z., P. Zhao, et al. (2006). "Thermo-optic coefficients of polymers for optical
waveguide applications." Polymer 47(14): 4893-4896.

139

Zhao, B. and W. J. Brittain (1999). "Synthesis of Polystyrene Brushes on Silicate
Substrates via Carbocationic Polymerization from Self-Assembled Monolayers."
Macromolecules 33(2): 342-348.

Zhu, J. G., S. K. Ozdemir, et al. "On-chip single nanoparticle detection and sizing by
mode splitting in an ultrahigh-Q microresonator." Nature Photonics 4(1): 46-49.

140

Appendix A: Ultimate quality factor of silica microtoroid
resonant cavities

A.1 Introduction
Silica optical microcavities with quality (Q) factors above 100 million have
applications throughout science and engineering. While both the microtoroid and
microsphere resonant cavity have demonstrated Q>100 million, only the microsphere has
surpassed 1 billion. Surprisingly, the reason for this performance disparity is directly
related to type of silicon substrate used in the fabrication process. In the present work,
the theoretical Q of planar toroidal silica resonant cavities is calculated and compared to
experimental results from a series of devices fabricated from oxide on doped silicon
wafers. As predicted, the Q depends on the substrate dopant concentration.

A.2 Fabrication
For the purpose of this study, several doping (Boron) concentrations of silicon
wafer with 2μm thick oxide layer were purchased from Montco Silicon. Figure A-1
shows one of fabricated silica microtoroid. Please refer to Chapter 2 for more detailed
fabrication procedure.
141

Figure A- 1 A scanning electron micrograph of the fabricated silica microtoroid resonator.

To measure the refractive indices of the thermal oxides, variable-angle
spectroscopic ellipsometry using a J.A. Woollam Co. VASE® instrument and was used
and summarized in Table A-1. Data is measured from 500 to 1100nm at three angles of
incidence (64, 69 and 74) near 75 to enhance sensitivity in the data analysis [1]. From
acquired  and , the refractive indices are then determined by fitting with previously
determined Si-SiO
2
material system. The parameters for Si and SiO
2
are taken from the
database provided by J.A. Woollam.

142

Boron
concentration
(cm
-3
)
Refractive index Quality factor (Q)*×10
8

630nm 850nm 980nm 630nm 850nm 980nm
1.63E14
1.32E15
1.47E16
7.98E18
7.68E19
1.4657
1.4694
1.467
1.4695
1.3834
1.4608
1.4678
1.463
1.4648
1.3799
1.4582
1.449
1.4609
1.4626
1.3659
3.0
4.53
2.3
1.94
1.31
3.39
1.85
1.27
1.13
0.35
2.0
1.4
1.5
1.16
0.43

Table A- 1 Summary for the values of refractive index and quality factor at λ=635, 850 and 980nm and its
corresponding boron concentrations (*The error in a single measurement of an individual resonant linewidth is ±
1×10
7
. The variation in Q between a significant number of cavities is much larger, as a result of imperfections in
fabrication. Therefore, values reported in this table are the highest achieved with each type of cavity.

A.3 Theory
Unlike the microsphere, the microtoroid is fabricated from thermal oxide not
silica fiber.[2] Thermal oxide is grown by oxidizing a silicon wafer.[3] Previous
research has shown that dopants diffuse from the silicon into the oxide, at a rate which is
dependent on the growth conditions.[3, 4] These dopants can modify the refractive
index, which in turn will affect the Q of the cavity. In the case of boron, the refractive
index is decreased.
Using the previously detailed expressions for Q
mat
, Q
ss
and Q
is
, and assuming the
Q
coupl
~0, we can develop an equation for Q: Q
-1
= Q
is
-1
+Q
ss
-1
+Q
mat
-1
. This expression was
calculated, and the dependence of the Q on the refractive index and the resonant
wavelength is shown in Figure A-2. It is obvious from the graph that the quality factor
(Q) is highly dependent on the change of refractive index caused by different boron
143

concentrations in the cavity. Also, the quality factor depends on the wavelength because
of the change in the mode volume as can be seen in Figure A-3. This change modifies
the effective absorption coefficient and the internal scattering at each wavelength. In
addition, it is anticipated that the larger mode volume in longer wavelength increases the
interaction between the optical field and boron dopants, hence leading to more scattering
and material losses in the cavity.

Figure A- 2 The calculated dependence of the quality factor on the refractive index and the resonant wavelength. The Q
factor decreases as the refractive index decreases and the wavelength increases.

144

Figure A- 3 FEM results which shows different mode volume for (a) λ=635 and (b) λ=980

A.4 Experimental results and discussions
To ensure accurate comparison between theoretical calculation and experimental
testing data, we measured quality factor at three different wavelengths (635, 850 and
980nm) for characterizing the dependence of different doping concentrations. A single
mode tapered optical fiber is used to couple light from a single mode, tunable narrow
linewidth CW laser into the microtoroid as detailed in Chapter 2.
One example of measured quality factor from forward scan can be seen in Fig A-4
at 850nm for a silica microtoroid fabricated on a silicon wafer with a boron concentration
of 1.63 ×10
14
cm
-3
. The splitting of the resonant linewidth results from coupling into the
clockwise (CW) and counter-clockwise modes [5]. To fit the spectra, a dual Lorentizian
fitting was used to determine exact linewidth from each resonance peak. As can be seen
from the spectra, both resonance peaks shows ultra high quality factor which is in excess
145

of 100 million, indicating the low coupling losses present in the system and the high
quality of the optical cavity.

Figure A- 4 A fine scan (the forward scan direction) of the fundamental transverse mode of the microtoroid fabricated
from the film with boron concentration of 1.63 ×10
14
cm
-3
in silicon at 848.8nm with a dual-Lorentz fit. The resonance
shows splitting and the quality factor of the left and right peak is 3.39 ×10
8
and 2.67 ×10
8
separately. Inset: Optical
micrograph picture of a microtoroid coupled to a tapered optical fiber during testing.

The highest quality factor was plotted in Fig A-5 as a function of refractive index
or boron dopant concentration at the three wavelengths. To refer the specific values of
measured quality factor and refractive indices, it was summarized in Table
A-1. As can be seen clearly from Fig A-5, the quality factor is highly dependent on
refractive index and the wavelength. This dependence of the Q on refractive index and
wavelength shows same trend as theoretical calculation as previously shown in Figure A-
2. This trend confirms that the diffusion of the dopant from the silicon to the oxide is
significantly affecting the Q in these devices.
146

Figure A- 5 Measured quality factor as a function of refractive index or boron dopant concentration at three
wavelengths: 630, 850 and 980nm. The Q factor decreases as the refractive index decreases and the wavelength
increases.

A.5 Conclusions
In this research, the dependence of doping concentration was studied. The silicon
wafer was purchased from Montco Silicon to ensure accurate experimental measurement.
Initially, only materials loss was considered for theoretical calculation. However, several
experimental results show that additional terms are necessary to explain the dependence
of doping concentration in the cavity. For final model, we include Q
is
, Q
ss
and Q
mat
for
predicting the resonator device behavior with combinations of wavelengths and refractive
indices. With this consideration, the value of quality factor both for theoretical and
experimental data shows same trend. This supports our assumption. It was revealed from
our study that the dopants diffusing from the silicon into the oxide layer has negative
impact for quality factor which will eventually reduce device performance, especially for
147

sensitivity. Therefore, it is suggested that microtoroid resonant cavities should be
fabricated on an undoped or intrinsic silicon wafer to ensure high quality factor. Such a
performance improvement could impact numerous areas of research and technology,
including bio/chem-detection [6, 7], nonlinear optics [8, 9], fundamental physics [9, 10],
and telecommunications [10].

148

Appendix A References
1. Snyder, P.G., et al., Variable angle of incidence spectroscopic ellipsometry:
Application to GaAs-Al[sub x]Ga[sub 1 - x]As multiple heterostructures. Journal
of Applied Physics, 1986. 60(9): p. 3293-3302.

2. Armani, D.K., et al., Ultra-high-Q toroid microcavity on a chip. Nature, 2003.
421(6926): p. 925-928.

3. Streetman, B.G. and S. Banerjee, Solid State Electronic Devices. 5 ed. 1999:
Prentice Hall. 558.

4. Fair, R.B., Physical Models of Boron Diffusion in Ultrathin Gate Oxides. Journal
of The Electrochemical Society, 1997. 144(2): p. 708-717.

5. Weiss, D.S., et al., SPLITTING OF HIGH-Q MIE MODES INDUCED BY
LIGHT BACKSCATTERING IN SILICA MICROSPHERES. Optics Letters,
1995. 20(18): p. 1835-1837.

6. Zhu, J.G., et al., On-chip single nanoparticle detection and sizing by mode
splitting in an ultrahigh-Q microresonator. Nature Photonics. 4(1): p. 46-49.

7. Armani, A.M., et al., Label-free, single-molecule detection with optical
microcavities. Science, 2007. 317(5839): p. 783-787.

8. Kippenberg, T.J., et al., Ultralow-threshold microcavity Raman laser on a
microelectronic chip. Optics Letters, 2004. 29(11): p. 1224-1226.

9. Hsu, H.S., C. Cai, and A.M. Armani, Ultra-low-threshold Er:Yb sol-gel
microlaser on silicon. Optics Express, 2009. 17(25): p. 23265-23271.

10. Rokhsari, H. and K.J. Vahala, Ultralow loss, high Q, four port resonant couplers
for quantum optics and photonics. Physical Review Letters, 2004. 92(25).

149

Appendix B: Excitation of Cy5 in self-assembled lipid bilayers
using optical microresonators

B.1 Introduction
Recently, microtoroids and microspheres have played an increasingly important
role in biological and chemical detection[1]. The evanescent field of the microcavity is
approximately 150nm long and leaks into the surrounding environment. The evanescent
field interacting with molecules within that range enables detection and excitation of
fluorescent probes. The ultra-high-Q of the microcavities has demonstrated single
molecule detection capabilties[2]. The advantage of this method is label-free detection, as
there are no fluorescent probes required for single molecule detection. While previous
research has focused on antibody/antigen detection for diagnostics, opportunities for
other detection schemes are possible. Specifically, a biomimetic coating functionalized
onto the surface of the resonator could enable 3-dimensional tracking of a molecule
through a cell membrane[3]. This expands upon previous research of 2-dimensional
tracking. In addition, the photodynamics of fluorescent probes could be studied within the
biomimetic coatings[4], improving on previous research, as microcavities have enhanced
temporal resolution. While optical microcavities have been used to excite fluorescent
molecules in the past[5], they have been limited to cavity quantum electrodynamics and
interaction of large fluorescent nanoparticles.
In this research, a lipid bilayer is functionalized onto the surface of an optical
microsphere. The lipid bilayers are 5nm thick, making them within range of the
150

evanescent field generated by the microcavity. Lipid bilayers are commonly used as
models for cell membranes, and have primarily been used in conjunction with fluorescent
microscopy in order to study transport of molecules through the cell membrane[6]. In the
present work, a method for self assembling the lipid bilayers onto the silica microspheres
is developed and the corresponding bilayer and optical properties are characterized.

B.2 Theory
The optical field is only partially confined inside the silica microsphere cavity. A
portion of it evanesces or leaks into the environment, thereby interacting with the lipid
bilayer. COMSOL Multiphysics simulations were performed to determine the exact
percentage of the optical field in the microcavity, the lipid bilayer and in air.[7-9] These
percentages are represented by β , γ and δ, respectively. To perform the modeling, the
refractive index of both the bilayer and the silica was needed. The refractive index of the
microsphere (silica) was taken from the COMSOL library and the refractive index of the
lipid bilayer was taken from reference [10].
The field distribution was modeled assuming a 200m diameter sphere with either
a 5nm or 10nm thick bilayer. However, it is commonly known that lipid bilayers are
~5nm thick. The resonant wavelength was set to 633nm or 980nm in each simulation by
controlling mode number (M), where M is azimuthal mode order in the cavity.
All of the percentages are shown in Table B-1; however, it is commonly known
that a lipid bilayer is 5nm thick. Figure B-1 shows FEM simulation results for a 5nm
thick lipid bilayer on a microsphere at 980nm.
151

Optical Field distribution (%)
Wavelength Thickness of
Bilayer (nm)
SiO
2

Lipid Bilayer

Air
δ
635nm 5 99.742 0.0173832 0.241
10 99.72 0.0370975 0.243
980nm 5 99.663 0.0190103 0.318
10 99.64 0.0395782 0.32

Table B- 1 FEM Simulation Results

Figure B- 1 A finite element method simulation showing the distribution of the optical field inside (left) and outside
(right) of the microsphere cavity at 980nm. The 5nm-thick lipid bilayer is not visible at this scale.

B.3 Experimental procedure
B.3.1 Fabrication of Microspheres
The silica microsphere resonators were fabricated by melting the tip of silica
optical fiber (Newport) using a CO
2
laser.[11, 12] First, the cladding was removed from
the optical fiber using cladding strippers. Then, the optical fiber was partially tapered
152

using an optical fiber tapering set-up in which the fiber is simultaneously pulled on either
end while it is heated in the middle with an oxyhydric flame.[13-15] Then the taper fiber
is cleaved. By tapering the fiber before forming the microsphere, it is possible to
fabricate smaller microspheres. Lastly, the tapered end is melted using a 50W CO
2
laser
(Synrad) at low intensity (Figure B-2 (a)).
B.3.2 Self Assembly of Lipid Bilayers
Solid-supported lipid bilayers are fabricated by releasing either a giant unilamellar
vesicle (GUV) solution or a small unilamellar vesicle (SUV) solution onto a hydrophilic
surface. When the vesicles make contact with the surface, they break apart and self
assemble to form a uniform bilayer. SUVs were chosen as the preferred method for the
functionalization of microspheres, as their smaller vesicle size (50nm) led to a lower
chance of seeing clustering.
The SUV solution is fabricated by mixing a 2:2:1 ratio of DPPC (1,2-dipalmitoyl-
sn-glycero-3-phosphocholine), DOPC (1,2-dioleoyl-sn-glycero-3-phosphocholine), and
cholesterol, dissolved in chloroform and vortexed. Then, the chloroform is evaporated by
applying argon gas for 10 minutes, followed by placing the test tube into a vacuum
chamber for 1 hour. Then, phosphate buffered saline is added to the tube and vortexed,
followed by sonication for 45 minutes. The sonication bath breaks up the multilamellar
vesicles into small unilamellar vesicles. Finally, the optical microspheres are immersed
into the SUV solution for 36 hours, with the lipid bilayer self assembling via vesicle
adsorption and fusion[16]. Over this time period, the vesicles form a smooth, 5nm thick
bilayer (Figure B-2 (b)).
153

The uniformity and bilayer structure have been confirmed by FRAP and FRET
measurements[17]. FRAP measurements showed 97% recovery of a focused light spot
after photobleaching, proving the fluidity of the bilayer. FRET measurements used a
QSY-7 amine quencher (Invitrogen), which when added to the microspheres, reduced the
intensity down to 51%. This shows that the surrounding layer is a bilayer, as
approximately only 50% of the bilayer was affected by the quencher.
During fabrication of the multilamellar vesicles, TR-DPPE (Texas Red dye
conjugated to 1,2-dipalmitoyl-sn-glycero-3-phosphoethanoamine) and Cy5-DMPE (Cy5
dye conjugated to 1,2-dimyristoyl-sn-glycero-3-phosphoethanolamine) are used for
fluorescent microscopy and emission measurement purposes, respectively. Cy5-DMPE is
synthesized via a reaction with DMPE (Avanti) and N-hydroxysuccinimidyl ester (GE
Healthcare).

Figure B- 2 Images of (a) an optical micrograph of a silica microsphere and (b) a fluorescent micrograph of a silica
microsphere with a self assembled lipid bilayer, conjugated to Texas Red Dye.

154

B.3.3 Confirmation of bilayer structure
1) Monolayer quenching

Lipid-coated microspheres were prepared by incubating silica microspheres in a
liposome suspension, as described previously. The lipid mixture included TR-DPPE as a
labeled species. Following liposome deposition, the microsphere was transferred to a
microscopy chamber containing fresh buffer (150 mM NaCl, 100 mM phosphate, pH
7.55) and imaged by epifluorescent microscopy. After an initial image was taken, an
aliquot of solution containing the resonance transfer quencher QSY-7 amine was added
and the microsphere was imaged again. As shown in Figure B-3a and B-3b, the
fluorescence intensity from the microsphere dropped significantly after the addition of
quencher. Figure B-3c shows that this change in intensity was approximately 50%. The
addition of more QSY-7 resulted in no further changes in fluorescence, indicating that
approximately half of the lipid was inaccessible to the buffer. This inaccessible lipid
fraction corresponds to the inner leaflet of the membrane. The fact that half of the lipid is
accessible to quenching by a soluble species indicates that the lipids take on a molecular
bilayer structure.
155

Figure B- 3 Half of the lipids on the microsphere surface are accessible to a soluble quencher. Epifluorescent images of
a lipid-coated microsphere (a) before and (b) after the addition of the soluble quencher QSY-7, reproduced from the
main text for the reader’s convenience. (c) Quantitative measurement of lipid fluorescence before and after the addition
of quencher shows that about half of the lipids are accessible to quenching, establishing a bilayer structure.

156

2) Fluorescence recovery after photobleaching (FRAP)

FRAP experiments were performed on a lipid coated microsphere as described
above. After an initial image was taken (Figure B-4a), the microscope excitation aperture
was closed down to its smallest setting and the intensity of the excitation lamp was
maximized. The illumination was focused on a spot on the surface of the microsphere for
20 minutes (Figure B-4b), after which the fluorescent intensity in this spot could be seen
to significantly decrease (Figure B-4c). The intensity in the bleached spot was measured
again after 20 minutes, at which time it could be seen to have recovered nearly all of its
original fluorescence (Figure B-4d). This indicates that the labeled lipids are mobile, as
expected for a molecular bilayer membrane. Figure B-4e shows a quantitative
measurement of the intensity from the bleached spot before bleaching, after bleaching,
and after 20 minutes’ recovery. 97% of the initial intensity is recovered.
157

Figure B- 4 Recovery of fluorescence in a photobleached region indicates that lipids on the microsphere are mobile.
Epifluorescent images of the lipid bilayer (a) before bleaching, (b) during bleaching, (c) after bleaching and (d) after 20
minutes of recover. (e) Quantitative measurement of lipid fluorescence shows that the bilayer recovers 97% of the
bleached fluorescence, indicating that lipids can diffuse in the membrane.
158

B.4 Experimental results and discussions
B.4.1 Characterization of Q factor
To verify the interaction of the evanescent field with the lipid bilayer, the quality
factor of a lipid bilayer-coated resonant cavity is determined using bilayers containing or
not containing Cy5-DMPE. This measurement is performed in both the visible (633 nm)
and the near-IR (980 nm) using a pair of single-mode, tunable external cavity continuous
wavelength lasers centered at those wavelengths. The intrinsic modal linewidth and
hence the intrinsic Q is then computed using a resonator-waveguide coupling model.[18]
The absorption maximum of Cy5 is approximately 650 nm. In the present case, the
quality factor is inversely proportional to the material absorption.[12] The precise nature
of this dependence can be calculated from the previously described finite element method
simulations by determining the intensity of the optical field which interacts with the Cy5-
DMPE. At 980 nm, the material absorption of the Cy5-DMPE is similar to that of the
lipid bilayer; consequently, the simulations predict the photon lifetime of the device
should be the same with or without dye present. In contrast, at 633 nm, the high optical
absorption of the Cy5 should reduce the photon lifetime and decrease the measured
quality factor. As shown in Figure B-5, the quality factor of the cavity at 633 nm
decreases by nearly two orders of magnitude when the Cy5-DMPE is present (1.594x10
7

to 5.308x10
5
) while the quality factor at 980 nm remains constant (6.601x10
6
to
3.033x10
6
). Therefore, the evanescent tail is strongly overlapping with the lipid bilayer
and interacting with the Cy5-DMPE.
159

Figure B- 5 The measurements of the photon lifetime or Q factor, which indicates the interaction strength between the
whispering gallery mode and the lipid bilayer. The quality factor was determined from the full-width half maximum
(FWHM, ) of the Lorentzian fit (dashed red line) to the experimental data (solid black line). This measurement was
performed using an undoped lipid bilayer at a) 633 nm and b) 980 nm, and using a Cy5 conjugated lipid bilayer at c)
633 nm and d) 980 nm.

B.4.2 Measurement of Cy5 Emission
The Cy5 was excited via the evanescent tail of the whispering gallery mode of the
optical cavity using a tunable laser which was centered at 633nm. Depending on the
experiment being performed, the laser continuously scanned over a 0.01nm to 0.04nm
wavelength range at a frequency of approximately 100Hz. Therefore, given this scan
range and the linewidth of the cavity (10
6
=Q=, =0.0006 for 10
6
or 0.00006nm for
160

10
7
), the Cy5 dye was excited approximately 0.1 to 1% of the time (0.0006/0.04=1.5%
and 0.00006/0.04=0.15%).
The position of the waveguide with respect to the optical cavity was monitored
throughout the experiment, using a custom built Navitar machine vision system
comprised of a top and side view imaging system.
To measure the emission spectra of Cy5, a port was added into the side view
camera optical column with an integrated beam splitter. As shown in the schematic in
Figure B-6, a custom adapter was machined to connect the fiber optic cable from the
spectrograph (Andor Shamrock spectrograph with a Newton CCD detector) to the optical
column (Navitar). Through a judicious choice of port location, the beam splitter allowed
50% of the light to pass and go to the camera while reflecting and focusing 50% of the
light onto the face of the optical fiber, thus maximizing photon collection. Finally, a
custom 633nm filter (Chroma) was inserted between the optical column and the
microsphere to block the excitation laser.
161

Figure B- 6 Schematic of test and measurement set-up.

As shown in Figure B-7(a), the Cy5 dye in the lipid bilayer is efficiently excited
using the evanescent field of the microcavity with minimal input power, and the emission
spectra is accurately recorded with excellent signal-to-noise. Previous research has
shown that Cy5 has an emission wavelength of 670 nm.[19] The peak emission
wavelength of Cy5 excited using the optical cavity is 670.11 nm, which is in excellent
agreement with previous results. Therefore, the experiments demonstrate that the optical
cavity method is able to accurately excite the fluorescent molecules conjugated to lipid
bilayers.
162

Figure B- 7 (a) The emission from the Cy5-conjugated lipid bilayer which is excited by the optical microresonator.
The peak maximum is located at 670.11 nm. The 633 nm excitation laser is blocked using a filter. (b) The fluorescent
decay of the Cy5 in the lipid bilayer which was excited by the optical microresonator. The decay was fit to a double
exponential.

Using Andor’s SOLIS software, data was automatically acquired with an
exposure time of 3 seconds. A series of control experiments were performed to ensure
that the measured emission spectra shown in Figure B-7(a) was not an artifact or
background signal. Specifically, the following set of experiments was performed:
1) The spectra of the background was measured, both normalized and non-
normalized (Figure B-8)
2) The spectra of the laser, with and without the filter, to verify the efficacy of
the filter (Figure B-9)
3) The emission spectra of Cy5, with and without the 633nm filter, to verify that
the 633nm filter has minimal impact on the measurement of the Cy5 emission.
(Figure B-10)
163

Figure B- 8 The spectra of the background was measured, both normalized and non-normalized

Figure B- 9 The spectra of the laser, with and without the filter, to verify the efficacy of the 633nm filter. This pair of
spectra was recorded by imaging the evanescent field of the taper. Note that the y-axis is plotted on a logarithmic
scale.

164

Figure B- 10 Comparison of the excitation/emission spectra measured by the spectrograph, with and without the 633nm
filter. This pair of spectra further verifies that the 633nm filter has minimal impact on the measurement of the Cy5
emission. The spectrum with the filter is reproduced from the main text (Figure B-3a) for convenience. Note that the
y-axis is plotted on a logarithmic scale.

B.4.3 Photodynamics of Cy5
Additional experiments are performed to probe the photodynamics of the
fundamental dye molecule in the lipid bilayer environment using optical microresonators.
Specifically, the rate of fluorescence decay is monitored and recorded by the Andor
spectrograph. These results are then fit to a standard double exponential model (Figure
B-7b).[20] This form is chosen because it most accurately models the complex
photobleaching behavior, which involves both radiative and non-radiative decay rates.
As shown in Figure B-7b, the fluorescence decay is detected using this method with high
fidelity, and the results fit the conventional, double exponential decay model extremely
well. The half-life of Cy5 in the lipid bilayer is calculated as detailed the below, and is
124 s. This value is comparable to the half-life of antibody-labeled Cy5 which was
excited using a similar manner.[21]
165

To demonstrate that we can use this method to detect the photodynamics or
bleaching behavior of Cy5 embedded in a self-assembly lipid bilayer, we use the Andor
SOLIS software to monitor the emission or intensity of fluorescence every 10 seconds for
1500 seconds. We then fit the intensity decay to the multi-parameter exponential
function below, representing the multiple decay pathways possible in this system:
α

α

The intensity of the optical field interacting with the fluorophore when the cavity
is on resonance is 1.1 MW/cm
2
, based on the microsphere diameter and the input power.
However, it is important to remember that these pulses have a repetition rate of 100Hz
and they are on resonance for a fraction of a second. Therefore, this is not a continuous
excitation. From this fit, we determine the following set of values for our system:

1
0.24930

2
0.65314

1
548.47

2
40.898

To determine the half life of Cy5 in the phospholipid bilayer, it is necessary to
combine 
1
and 
2
into a single  value. This calculation, which is based on
reference[20], is shown below:
α α

α

α

α

α

α

This half-life is comparable to the half-life of antibodies with Cy5 attached to
them which were excited using a waveguide.[21]
166

B.5 Conclusions
In conclusion, we have demonstrated a new self-assembly method for
functionalizing lipid bilayers onto the surface of microspheres. We then characterized the
device performance by calculating the Q factors, verifying that the functionalization
method could be used for future single molecule studies. Finally, we confirmed the
existence of the lipid bilayer on the microsphere via two methods. First, we showed that
the Q factor significantly decreased using a 633nm laser compared to a 980nm, caused by
the absorption of the Cy5 dye that was conjugated to the lipid bilayer. Secondly, we
collected the emission spectrum of Cy5-DMPE dye conjugated to a lipid bilayer that was
excited on the microcavity using a 633nm laser.
The system proposed in this paper can find various applications in biological and
chemical science. Primarily, we expect this system to be used for the application of
studying the transport of molecules through the cell membrane
7
. The system could also
be useful for studying fluorescent probes conjugated within thin films
6
. The advantage to
this system is it is compatible with thin films beyond lipid bilayers. Only a small portion
of the film is excited at once, allowing for the rest of the film to be used multiple times.
This enables future studies on how high optical intensities affect thin biofilms.
Additional advantages to this system include the ability to conjugate other
molecules within the lipid bilayer. It is possible to insert molecules such as DNA, RNA,
and proteins inside the bilayer[22, 23]. This can be useful in studying how fluorescent
probes interact with these molecules[24]. The proposed system is flexible and has
numerous potential applications in the field of sensing, as the bilayers are highly
167

customizable. The surface functionalization technique of self-assembling a thin bilayer
film onto the surface of the microcavity will enhance previous sensing studies, and
provide a platform for future single molecule detection studies.

168

Appendix B References
1. Hunt, H.K. and A.M. Armani, Label-free biological and chemical sensors.
Nanoscale. 2(9): p. 1544-1559.

2. Armani, A.M., et al., Label-free, single-molecule detection with optical
microcavities. Science, 2007. 317(5839): p. 783-787.

3. Schafer, L.V., et al., Lipid packing drives the segregation of transmembrane
helices into disordered lipid domains in model membranes. Proceedings of the
National Academy of Sciences of the United States of America. 108(4): p. 1343-
1348.

4. Goedhart, J., et al., Bright cyan fluorescent protein variants identified by
fluorescence lifetime screening. Nature Methods. 7(2): p. 137-U74.

5. Gotzinger, S., et al., Controlled photon transfer between two individual
nanoemitters via shared high-Q modes of a microsphere resonator. Nano Letters,
2006. 6(6): p. 1151-1154.

6. Li, S., P.C. Hu, and N. Malmstadt, Confocal Imaging to Quantify Passive
Transport across Biomimetic Lipid Membranes. Analytical Chemistry. 82(18): p.
7766-7771.

7. Oxborrow, M., Traceable 2-D finite-element simulation of the whispering-gallery
modes of axisymmetric electromagnetic resonators. Ieee Transactions on
Microwave Theory and Techniques, 2007. 55(6): p. 1209-1218.

8. Oxborrow, M., How to simulate the whispering-gallery-modes of dielectric
microresonators in FEMLAB/COMSOL - art. no. 64520J. Laser Resonators and
Beam Control IX, 2007. 6452: p. J4520-J4520

9. Choi, H.-S., X. Zhang, and A.M. Armani, Hybrid Silica-Polymer Ultra-High-Q
Microresonators. Optics Letters, 2010. 35(4): p. 459-461.

10. Schouten, S., P. Stroeve, and M.L. Longo, DNA Adsorption and Cationic Bilayer
Deposition on Self- Assembled Monolayers. Langmuir, 1999. 15: p. 8133.

11. Gorodetsky, M.L., A.A. Savchenkov, and V.S. Ilchenko, Ultimate Q of optical
microsphere resonators. Optics Letters, 1996. 21(7): p. 453-455.

12. Vernooy, D.W., et al., High-Q measurements of fused-silica microspheres in the
near infrared. Optics Letters, 1998. 23(4): p. 247-249.
169

13. Kippenberg, T.J., et al., Fabrication and coupling to planar high-Q silica disk
microcavities. Applied Physics Letters, 2003. 83(4): p. 797-799.

14. Little, B.E., J.P. Laine, and H.A. Haus, Analytic theory of coupling from tapered
fibers and half-blocks into microsphere resonators. Journal of Lightwave
Technology, 1999. 17(4): p. 704-715.

15. Spillane, S.M., et al., Ideality in a fiber-taper-coupled microresonator system for
application to cavity quantum electrodynamics. Physical Review Letters, 2003.
91(4): p. -.

16. Brian, A.A. and H.M. McConnell, ALLOGENEIC STIMULATION OF CYTO-
TOXIC T-CELLS BY SUPPORTED PLANAR MEMBRANES. Proceedings of
the National Academy of Sciences of the United States of America-Biological
Sciences, 1984. 81(19): p. 6159-6163.

17. Freeman, L.M., et al., Excitation of Cy5 in self-assembled lipid bilayers using
optical microresonators. Applied Physics Letters. 98(14).

18. Spillane, S.M., et al., Ideality in a fiber-taper-coupled microresonator system for
application to cavity quantum electrodynamics. Physical Review Letters, 2003.
91(4): p. 043902.

19. Wang, L.L., A.K. Gaigalas, and V. Reipa, Optical properties of Alexa (TM) 488
and Cy (TM) 5 immobilized on a glass surface. Biotechniques, 2005. 38(1): p.
127-132.

20. Gafni, A., R.L. Modlin, and L. Brand, Analysis of fluorescence decay curves by
means of the Laplace transformation. Biophysical Journal, 1975. 15(3): p. 263-
280.

21. Klotz, A., et al., Sensitivity enhancement of transducers for total internal
reflection fluorescence. SPIE International Symposium on Integrated
Optoelectronic Devices, San Jose, 1999: p. 3620-3653.

22. Castellana, E.T. and P.S. Cremer, Solid supported lipid bilayers: From
biophysical studies to sensor design. Surface Science Reports, 2006. 61(10): p.
429-444.

170

23. Schouten, S., P. Stroeve, and M.L. Longo, DNA adsorption and cationic bilayer
deposition on self-assembled monolayers. Langmuir, 1999. 15(23): p. 8133-8139.

24. Larsson, C., M. Rodahl, and F. Hook, Characterization of DNA immobilization
and subsequent hybridization on a 2D arrangement of streptavidin on a biotin-
modified lipid bilayer supported on SiO2. Analytical Chemistry, 2003. 75(19): p.
5080-5087.

171

Appendix C: Optical Microcavities with a Thiol-Funcionlized
Gold Nanoparticle Polymer Thin Film Coating

C.1 Introduction
Polymer coatings endow ultra-high-Q dielectric resonators with nonlinear
properties, impacting numerous applications. However, minimal research combining
microcavities with polymer-nanoparticle coatings to tune or tailor the optical properties
of the system has been performed. One challenge is maintaining the high performance of
the optical device while in the presence of nanoparticles. In the present work, a toroidal
microcavity is coated with a polymethylmethacrylate thin film containing thiol-
functionalized gold nanoparticles. The thiol-functionalization ensures that the
nanoparticles are uniformly distributed throughout the film. The quality factors of these
devices are above 5 million and are in good agreement with the theoretical predictions.

C.2 Motivation
By coating a silica device with a polymer film, it is possible to form a hybrid
organic-inorganic device, which exhibits many advantages over pure silica devices. For
example, they can be used to stabilize and remove many of the non-linear effects which
typically plague pure silica devices as shown in chapter 5. [1] They can also be used to
study the fundamental behavior of the polymer film. However, all previous work using
hybrid devices has focused on combining silica or silicon devices with simple
homopolymers. If the polymer film is used as a matrix for nanoparticles, it will enable
172

numerous investigations into the interactions between microcavities and nanoparticles.
Additionally, unlike nanoparticles which are simply physio-adsorbed to the surface, the
system will be stable for significant periods of time, allowing for storage of the devices.
One of the key challenges in fabricating such a platform device is it is critical to
maintaining a high quality factor and a strong optical field overlap with the embedded
nanoparticles.
In the present study, different concentrations of gold nanoparticles in
polymethylmethacrylate (PMMA) films are coated on a toroidal whispering gallery mode
resonant cavity device (Figure 1), and the impact of the hybrid film on the device
performance (cavity Q factor) is experimentally determined and compared to simulation
results.

Figure C- 1 (a) Artistic rendering and (b) optical image of the gold coated hybrid devices. Gold nanoparticles
suspended in a PMMA solution are coated onto the toroid surface. The major diameter for the microtoroid is
approximately 50µm. The gold nanoparticles are too small to visualize in this optical image.

173

C.3 Theory
In previous work as shown in Chapter 4, it was shown that in hybrid devices the
material loss is the dominant factor, yielding Q
o
~Q
mat
.[2] Therefore, the quality factor
can be described as:
Q
mat
=2πn
eff
/λα
eff
(C.1)
In this present hybrid system, these are expressed as n
eff
= βn
silica
+γn
gold-polymer
+
δn
air
, and α
eff
= βα
silica
+γα
gold-polymer
+ δα
air
, where β, γ, and δ represent the percentage of the
optical field in silica, gold-PMMA nanocomposite, and air, respectively. According to
Beer-Lambert law, 
gold
=c
gold
, where c
gold
is the concentration of the gold nanoparticles.
Therefore, through a simple substitution, we can easily draw the conclusion that
Q
mat
=A*c
gold
k
, where k should be equal to -1 and A is a constant which is proportional to
2 n/. It is important to note that in the present work, the polymer-nanoparticle film is
treated as an effective media.[3] This assumption has been verified by comparing
ellipsometric experimental data with theoretical calculations based on an effective media
model of the refractive index at 633nm.
To calculate a theoretical Q factor, it is first necessary to determine , , and  by
modeling the optical field distribution using COMSOL Multiphysics.[4] In the present set
of simulations, the device geometry is held constant at 50(5) m major (minor) diameter.
The film thicknesses held fixed at 35nm for 0 and 5% PMMA-nanoparticle solutions,
30nm for 10% PMMA-nanoparticle solutions and 25nm for 30% PMMA-nanoparticle
solutions, based on the values measured by ellipsometry. The refractive index of silica is
taken from literature while the refractive indices for the PMMA-nanoparticle thin films
174

are experimentally determined by ellipsometry.[5] The percentage of the optical field in
each region is determined by Power
In (Silica/Nanocomposite/Air)
/ Power
Tot
. As can be seen in
Figure C-2, the presence of the PMMA-nanoparticle film modifies the location of the
optical field. Combining the values for the optical field intensity distributions (β, γ, δ)
with UV-Vis absorption measurements and refractive index measurements, a theoretical
Q can be determined. These are plotted in Figure 4 to enable direct comparison with the
experimental values.

Figure C- 2 FEM simulations of the optical field distribution. a) The normalized radial optical field intensity as a
function of radius for silica (solid black line) and for 10% PMMA-nanoparticle nanocomposite film (red dashed line).
The zero point indicates the center of the minor diameter. This graph was determined from the optical field distribution
for (b) silica microtoroid and (c) nanoparticle-polymer coated microtoroid with a 30nm thick film. The device size is
50(5) m major (minor) diameter and operating wavelength is 635nm.

C.4 Materials and Methods
Based on previous research results, polymethylmethacrylate (PMMA) is chosen
as the host for the gold nanoparticles.[2] It has been shown that Q factors as high as
175

1x10
7
can be achieved with PMMA thin film coatings. Although PMMA can be easily
dissolved in toluene, the standard synthesis for gold nanoparticles occurs in water.
Therefore, after synthesizing the –OH group stabilized gold nanoparticles, by adding
toluene and a small amount of 1-dodecanethiol, the gold nanoparticles were transferred
completely from water into toluene and stabilized by a thiol group (Figure C-3). Using a
1% weight PMMA solution, a series of solutions containing both gold nanoparticles and
PMMA at different volume ratios (0:100, 5:100, 10:100, and 30:100) were made. The
concentrations are confirmed using an inductively coupled plasma (0, 0.0224, 0.0433,
0.0999) Mol/L.

Figure C- 3 Schematic of synthesis for thiol-stablized gold nanoparticles. First a gold hydrosol is synthesized. Then, to
replace the –OH on the surface with –SH, a small amount of 1-dodecanethiol is added, and the gold was transferred
completely to toluene.

C.4.1 Synthesis of Gold Nanoparticles
The gold nanospheres are synthesized by combining HAuCl
4
with freshly
prepared Sodium borohydride (NaBH
4
) aqueous solution. The gold nanoparticles form as
the NaBH
4
reduces the Au(III).[6] The resulting nanoparticle solution consists of gold
nanospheres suspended in water.
176

C.4.2 Thiol Functionalization of Gold Nanoparticles
The gold nanoparticles are functionalized with a thiol group by adding 10ml
toluene mixed with 10 mg of 1-dodecanethiol into the prepared solution in last step. 2mL
of concentrated hydrochloric acid was added immediately.[6] After 5min stirring, two
miscible layers formed and gold spheres are completely transferred into toluene and
stabilized by thiol group.
C.4.3 Concentration of PMMA-Gold Nanoparticle Solutions
To convert the volume percentage into concentrations, inductively coupled
plasma (ICP) measurements were taken. As shown in Figure C-4, the relationship
between the volume ratio and the ICP measured concentration is extremely linear,
indicating that the solutions were well-mixed and the nanoparticles were evenly
distributed throughout the solution.

Figure C- 4 Gold concentrations for the solutions with the volume ratio at 0:100, 5:100, 10:100 and 20:100. In other
words, gold solutions volume percentages are 0, 4.76, 9.09, and 16.67, respectively. It shows concentration changes
almost linearly with gold volume percentage. Results are taken by the inductively coupled plasma-atomic emission
spectrometry.

177

C.4.4 Characterization of PMMA-Gold Nanoparticle Thin Film
As mentioned previously, we used an effective media model for the refractive
index of the nanoparticle-polymer system. This assumption was verified by comparing a
theoretical calculation of the refractive index based on an effective media model with
ellipsometry measurements. The following paragraphs explain the calculations with a
comparison to the ellipsometry measurements at the end.
The refractive index of the 0.0990 mol/L gold nanoparticles in a PMMA film was
calculated at 633nm. The effective medium approximation allows the effective dielectric
function to be written as:[7]
ε

ε

ε ε

ε ε

ε ε

ε ε

; (C.2)
where

is the effective dielectric constant of the medium, is the dielectric
constant of the gold nanoparticles,

is the dielectric constant of the matrix, and f is
the volume fraction of the gold nanoparticles in the film. By using the Maxwell-Garnett
mixing rule, the electromagnetic material properties of the nanoparticle-polymer film can
be calculated by breaking it down into its components, namely the dielectric function of
the PMMA and the dielectric function of the gold nanoparticles. Specifically, the gold
nanosphere’s dielectric function can be written as: [7]
ε ε

γ

γ

(C.3)

where

, r represent the complex permittivity of gold bulk
material, the plasmonic resonance frequency, the frequency of optical wave, the damping
constant in the free-electron Drude model, the proportionality factor of the order of unit,
the Fermi velocity of gold, and the radius of the gold nanoparticles.
178

To determine the refractive index of the film, we only need the real parts in the
equation. Therefore, equation (2) becomes:

γ

γ

; (C.4)

Based on previously published results of gold[9],

is 0.193 at the wavelength
of 635nm,

is 9.03eV, A is 1.4 of the order of unity,

and

is
0.052eV. At 635nm, we calculated the refractive index for 15nm and 20nm radius
nanoparticles to be 1.9527 and 1.2730 respectively.
To further study the effective refractive index of the nanoparticle-polymer film
and compare with the experimental data, we only need to address the question of gold
volume ratio. One example calculation is shown; however, all calculations at 633nm
were performed using the same procedure based on 17nm radius diameter, and the values
are included in Table C-1.
For this example, we will use the 0.0990mol/L gold nanoparticles in 1.0% weight
PMMA solution at 633nm. The refractive index is calculated for 15nm and 20nm radius
nanoparticles. Through simple unit conversion, the f for 0.0990 mol/L solution is 11.6%.
Plugging the calculated f and into equation 1, the refractive indices for the two different
polymer-nanoparticle films are 1.5336 and 1.4873 for the 15nm and 20nm radius
nanoparticle-polymer films, respectively.
To experimentally measure the refractive index, a series of PMMA-nanoparticle
films were spun on silicon wafers at 4000 rpm for 60 seconds. The refractive index and
thickness of the PMMA-Gold nanoparticle thin film was determined using spectroscopic
ellipsometry (Table C-1). As can be seen, the thickness of the film decreases slightly
179

with gold nanoparticle concentration and the refractive index increases with gold
nanoparticle concentration, as expected.
Using the same parameters (wavelength, particle concentration, etc) as in the
effective media model calculations previously detailed, we experimentally measure a
value of 1.5054 which is in extremely good agreement with the values of 1.5401 and
1.4650 for the 15nm and 20nm nanoparticles. Therefore, this system can be modeled as
an effective media.
Gold
Concentration
(mol/L)
Thickness
(nm)
Refractive Index
(calculated)
Refractive Index
(measured)
633nm
(15nm)
633nm
(20nm)
633nm 765nm 980nm
0 34.36 1.4945 1.4945 1.4945 1.4907 1.4881
0.0224 33.41 1.5066 1.4866 1.4967 1.4931 1.4906
0.0433 31.89 1.5170 1.4798 1.5007 1.4975 1.4953
0.0990 26.84 1.5401 1.4650 1.5054 1.5022 1.5000

Table C- 1 Summary of calculations and spectroscopic ellipsometry results.

C.4.5 Material Absorption Spectra
All solutions of gold nanoparticles functionalized by –OH group and –SH group
were characterized using UV-Vis. As can be seen in the spectra in Figure C-5, there is a
clearly identifiable peak at 520nm for -SH stabilized gold spheres before mixing with
PMMA, which corresponds to a well-known absorption wavelength of gold particles at
radius of 17nm~20nm [8]. After mixing with PMMA solution at ratio 2:10, the peak
180

shows red shift to around 524nm, which is the result of the particles being transferred to a
higher refractive index media. It is important to note that the solution still has only a
single peak, indicating that the spheres did not aggregate. All spectra are normalized by
their intensity at 450nm.

Figure C- 5 UV-Vis spectra of gold nanoparticles stabilized by (black) –OH group, (blue) –SH group without PMMA,
and (red) –SH group with PMMA.

C.4.6 Device Fabrication and Lifetime
After fabricating silica microtoroid, the polymer-nanoparticle solution is spin
coated onto the toroidal surface and the film is thermally reflowed in a gravity oven at
115 °C for 30 min (Figure C-6). This thermal reflow reduces imperfections in the
polymer film, thereby decreasing scattering losses. The PMMA-nanoparticle thin film is
subsequently applied to the surface of the SiO
2
cavity at 4000 rpm for 60 seconds. A
SEM image of the fabricated device is shown in Figure C-6.

181

Figure C- 6 Scanning electron micrograph image of an ultra high Q silica microtoroid resonant cavity.

These devices have been stored in our lab under ambient conditions (room
temperature, pressure, air) for several months without degradation. Using a 765nm laser,
we tested a pure PMMA coated device and a 0.0224mol/L nanoparticle-PMMA coated
device after storage for three months. The Q for both devices did not decrease within
experimental error.

C.5 Results and Discussions
C.5.1 Typical spectra for Q measurement of PMMA-gold coated devices
It should be noted that the resonant wavelength of the cavity is dependent on the
the specific properties of that cavity, such as the refractive index of the cavity (silica), the
refractive index of the polymer-nanoparticle coating, and the geometry. As a result, each
cavity has a unique resonant wavelength. Therefore, figure C-7 is a representative spectra
182

of a microtoroid coated by a PMMA solution doped with 0.0224Mol/L gold nanoparticles
at 983nm.

Figure C- 7 Transmission spectra of microtoroid coated by spinning PMMA solution doped with 0.0224Mol/L gold
nanoparticles. Although the data was normalized, some of the data fell above 1.0 due to noise in the measurement.

C.5.2 Q measurement results
To experimentally measure the quality factor of the devices, the narrow-linewidth
(300kHz) tunable CW lasers operating at 633nm, 765nm, and 980nm to the optical cavity
were used.[9, 10] At these wavelengths, the spectral overlap between the plasmon
resonance of the nanoparticles and the polymer-coated microtoroid is negligible, as the
plasmon resonance is approximately 530nm. All spectra are taken in under-coupled
regime with less than 5W of input power to minimize thermal broadening.[2]
Figure C-8 shows the experimentally measured quality factors for the PMMA-
nanoparticle coated toroidal resonant cavities across the range of nanoparticle
concentrations. As can be observed, the quality factor decreases as the concentration of
183

gold nanoparticle increases, indicating that the optical field is strongly interacting with
the nanoparticles. However, at these gold nanoparticle concentrations, the Q factors are
above 10
6
at all wavelengths studied. The theoretical and the experimental Q values
shown in Figure C-8 are fit to an equation of the form y=ax
b
. The parameters (a, b)
determined from this fit are in table C-2.
184

Figure C- 8 Experimental (solid squares) and theoretical (hollow circles) quality factor of hybrid devices as a function
of gold concentrations in PMMA film at a) 633nm, b) 765nm and c) 980nm. The results were fit according to the
equation of y=ax
b
.

185

Table C- 2 Summary of Model and Experimental Fitting Parameters

As can be seen from both the graphs and the table, there is clearly excellent
agreement between the theoretical prediction and the experimental values. Therefore, the
Q factor is material, and not surface scattering, limited. This strong agreement justifies
our primary series of hypothesis: 1) that the effective media model is correct for this
system and 2) that the Q is limited by material loss (Q
tot
~Q
mat
), therefore scattering from
the nanoparticles plays a minor role.
C.5.3 Photon Lifetime
The conversion of the quality factors shown in figure C-8 in the main paper to the
photon lifetime () in the cavity is:

. (C.5)
The results are listed in Table C-3.
Gold
Concentration
(mol/L)
Photon Lifetime (ns)
633nm 765nm 980nm
0 6.41 6.53 9.51
0.0224 1.91 2.65 2.94
0.0433 0.89 1.03 1.42
0.0990 0.60 0.68 0.95

Table C- 3 Conversion of experimental quality factors to photon lifetime

Wavelength
(nm)
Model Experiment

a (x10
5
) b a (x 10
5
) b
633 3.3538 -0.7564 1.7234 -0.9145
765 2.7995 -0.7902 1.3222 -0.9847
980 2.7490 -0.8033 1.9110 -0.8867
186

C.6 Conclusions
In conclusions, we have experimentally and theoretically investigated the impact
of PMMA-nanoparticle coatings on the quality factor of ultra-high-Q devices. We have
demonstrated Q factors above 10
6
for several different concentrations of gold
nanoparticles and over a wide range of wavelengths. The measured Q values are in good
agreement with the theoretically predicted ones based on a material-limited Q model
using an effective media model for the nanoparticle-polymer coatings. The ability to
combine gold nanoparticle-PMMA thin films with ultra-high-Q devices will impact a
wide range of disciplines; for example, developing improved LSPR sensors and
understanding nonlinear optics in hybrid organic-inorganic systems.[11, 12]

187

Appendix C References
1. Choi, H.-S. and A.M. Armani, Thermal non-linear effects in hybrid optical
microresonators. Applied Physics Letters, 2010. 97(22): p. 223306.

2. Choi, H.-S., X. Zhang, and A.M. Armani, Hybrid Silica-Polymer Ultra-High-Q
Microresonators. Optics Letters, 2010. 35(4): p. 459-461.

3. Pei, Y., et al., Refractive index of silver nanoparticles dispersed in polyvinyl
pyrrolidone nanocomposite. Journal of Modern Optics, 2010. 57(10): p. 872-875.

4. Oxborrow, M., Traceable 2-D finite-element simulation of the whispering-gallery
modes of axisymmetric electromagnetic resonators. IEEE Transactions on
Microwave Theory and Techniques, 2007. 55(6): p. 1209-1218.

5. Malitson, I.H., INTERSPECIMEN COMPARISON OF REFRACTIVE INDEX
OF FUSED SILICA. Journal of the Optical Society of America, 1965. 55(10P1):
p. 1205-&.

6. Yang, J., et al., Langmuir, 2003. 19(24): p. 10361.

7. Pei, Y., et al., Refractive index of silver nanoparticles dispersed in polyvinyl
pyrrolidone nanocomposite. J Mod Optic, 2010. 57(10): p. 872.

8. Haiss, W., et al., Determination of size and concentration of gold nanoparticles
from UV-vis spectra. Anal Chem, 2007. 79(11): p. 4215-21.

9. Kippenberg, T.J., et al., Fabrication and coupling to planar high-Q silica disk
microcavities. Applied Physics Letters, 2003. 83(4): p. 797-799.

10. Zhang, X., H.-S. Choi, and A.M. Armani, Ultimate quality factor of silica
microtoroid resonant cavities. Applied Physics Letters, 2010. 96(15): p. 153304.

11. Sepúlveda, B., et al., LSPR-based nanobiosensors. Nano Today, 2009. 4(3): p.
244-251.

12. Kim, H.K., et al., Highly efficient organic/inorganic hybrid nonlinear optic
materials via sol-gel process: synthesis, optical properties, and photobleaching for
channel waveguides. Chemistry of materials, 1999. 11(3): p. 779-788.

Abstract (if available)

Abstract The optical characterization of polymeric materials to understand their fundamental behavior is important to study their materials' properties and to utilize them in numerous applications. Therefore, this thesis mainly investigates the possibilities for the development of hybrid organic-inorganic optical resonators for studying polymer thin films based on whispering-gallery mode (WGM) optical resonators. The long-period light confinement in WGM optical resonators with low loss materials, such as silica, ensures high sensitivity or high quality (Q) factor by scanning optical resonators numerous times, hence extracting the change of information caused in the system during the scan. Such detection mechanisms can be the change of quality factor, resonant shift or transmission change. ❧ In this thesis, it is first demonstrated that the loss mechanism for hybrid optical resonators are material-limited with quality factor over than 10^7, meaning that other loss mechanisms such as scattering, radiation, contamination and coupling losses are minimized with optimized device performance, device fabrication and optical characterization set-up. The change of optical field characterization is also investigated with different polymers, film thickness and operating wavelengths by Finite Element Method (FEM) simulations for the hybrid structure.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Development of hybrid optical microcavities for Plasmonic laser and improving biodetection

PDF

Silica sol-gel thin film coatings for integrated photonic devices

PDF

Organic materials for linear and nonlinear optical microdevices

PDF

Optical simulation and development of novel whispering gallery mode microresonators

PDF

Developing improved silica materials and devices for integrated optics applications

PDF

Fabrication and characterization of yoroidal resonators for optical process improvement

PDF

Development of novel optical materials for whispering gallery mode resonator for nonlinear optics

PDF

Surface-enhanced on-chip whispering gallery mode resonators for improved nonlinear optics

PDF

Integration of practical high sensitivity whispering gallery mode resonator sensors

PDF

Titania and hybrid titania - silica sol-gel thin films and their applications in integrated optical devices

PDF

Optically triggered smart polymers for environmental monitoring

PDF

Microresonator-based Kerr frequency comb for high-speed optical communications and signal processing

PDF

Real time surface analysis of complex oxide thin films during pulsed laser deposition

PDF

Development of optical devices for applications in photonic integrated circuit and sensing

PDF

Engineering solutions for biomaterials: self-assembly and surface-modification of polymers for clinical applications

PDF

Development of hybrid microsensors for environmental monitoring and biodetection

PDF

Phase change heterostructures for electronic and photonic applications

PDF

Development and characterization of transparent metal/ceramic and ceramic/ceramic nanomultilayers

PDF

Nonlinear optical nanomaterials in integrated photonic devices

PDF

A comparative study of plasma conditions, microstructure and residual stress in sputtered thin films

Asset Metadata

Creator Choi, Hong Seok (author)

Core Title Development of organic-inorganic optical microcavities for studying polymer thin films

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Materials Science

Degree Conferral Date 2012-05

Publication Date 03/20/2012

Defense Date 02/15/2012

Publisher Los Angeles, California (original), University of Southern California (original), University of Southern California. Libraries (digital)

Tag material characterization,non linear,OAI-PMH Harvest,optical microresonators,organic-inorganic,polymer,thin films

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Armani, Andrea M. (committee chair), Brutchey, Richard L. (committee member), Nutt, Steven R. (committee member)

Creator Email cded1004@hotmail.com,hongchoi@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC11292862

Unique identifier UC11292862

Identifier etd-ChoiHongSe-529.pdf (filename)

Legacy Identifier etd-ChoiHongSe-529

Dmrecord 210532

Document Type Dissertation

Format theses (aat)

Rights Choi, Hong Seok

Internet Media Type application/pdf

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu