University of Southern California Dissertations and Theses

Light-matter interactions in engineered microstructures: hybrid opto-thermal devices and infrared thermal emission control

(USC Thesis Other)

Light-matter interactions in engineered microstructures: hybrid opto-thermal devices and infrared thermal emission control

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

ii

Dedication

To my beloved parents

iii

Acknowledgments

Being a graduate student at USC for the past five years has had a far more profound impact
on me than I originally expected, thanks to the guidance and help from many talented and
considerate people in this beautiful campus.
First and foremost, I am deeply thankful to my PhD advisor, Prof. Michelle Povinelli, for her
priceless input and support in my research and all other aspects essential to academics. She is
always available to discuss any issues happily and has put an immense effort on teaching me the
art of writing, oral presentations, and scientific exploration. Her passion, intuition, and insight into
physics and photonics can only be surpassed by her pleasant personality. I will always cherish
her constant support, radiant energy, and constructive advice.
Secondly, I would like to thank my wonderful colleagues for stimulating discussions and
feedback. Aravind Krishnan who is always modest, joyful and a good friend. Our many
conversations over the past years are a wonderful memory. Romil Audhkhasi who has been a
smart and cheerful friend during my last year. I am sure he will have a great career ahead of him.
Next, I am very grateful to Shao-Hua Wu and Roshni Biswas for patiently teaching me their
experiments and design methods. In addition to being great researchers, they were always
cheerful, confident and inspiring examples to look up to. I am also thankful to Mingkun Chen and
Elise Uyehara who were the rising stars in our lab during their internships and I am sure they will
make great future scientists. Moreover, I am thankful to our collaborators Philip Hon, Vladan
Jankovic, Virginia Wheeler, Michael Barako and Luke Sweatlock. I greatly appreciate the EE
department staff, including Kim, Angela and Diane, for smoothly processing everything.
iv

I want to express sincere thanks to Prof. Rehan Kapadia and Prof. Stephen Cronin for being
on my PhD Qualification Exam committee. My deepest regards go to Prof. Wei Wu and Prof.
Jayakanth Ravichandran for being in both my PhD Qualification and Defense committee.
Beyond academics, I am eternally grateful to the support and continuous love of my parents,
brother and sisters. You mean the world to me and I will always love you with all my heart. I am
also grateful to all my friends at USC. Amr Elnakeeb, who is like a brother to me and his company
gave me the joy and the strength to get through my PhD. Amr, Samer, Mostafa and Abo Talib,
you were my bigger family in LA and our memories together are the highlights of my PhD journey.

v

Table of Contents

Dedication ................................................................................................................................. ii
Acknowledgments ................................................................................................................... iii
List of Tables ........................................................................................................................... vii
List of Figures ........................................................................................................................ viii
Abstract .................................................................................................................................. xiii
1. Introduction: Part I ............................................................................................................... 1
2. High Temperature Opto-thermal Memory ........................................................................... 3
2.1 Motivation for Opto-thermal Devices ................................................................................. 3
2.2 Physical Origin of Thermo-optic Bistability ........................................................................ 4
2.3 Experiment ....................................................................................................................... 7
2.3.1 Thermal Memory .......................................................................................................10
2.3.2 Memory Speed ..........................................................................................................12
2.3.3 Memory Reliability .....................................................................................................13
2.3.4 High-temperature operation ......................................................................................14
2.4 Discussion .......................................................................................................................16
3. Heat control on micro-scale ...............................................................................................21
3.1 Background and Motivation .............................................................................................21
3.2 Time-dependent optical and thermal response in water ...................................................22
3.3 Programmable heating ....................................................................................................27
3.3.1 Laser wavelength tuning ...........................................................................................27
3.3.2 Design parameters tuning .........................................................................................28
3.3.3 Spectral shift and bubble generation .........................................................................29
3.3.4 Appendix I .................................................................................................................31

vi

4. Future Work .........................................................................................................................34
4.1 Thermal Logic Gates .......................................................................................................34
4.1.1 Numerical Modeling ..................................................................................................35
4.2 Biological Applications .....................................................................................................38
4.2.1 Microfluidic modelling ................................................................................................39
4.2.2 Experimental demonstration ......................................................................................41
5. Introduction: Part II .............................................................................................................44
6. Coupled resonators for switchable emissivity .................................................................46
6.1 Coupled resonators for switchable emissivity ..................................................................46
6.2 Dark Modes in Metamaterial Absorbers ...........................................................................49
6.3 Dark – Bright Coupled Resonators ..................................................................................50
6.4 Dark- Dark Coupled Resonators ......................................................................................52
7. Broadband Emissivity Switching .......................................................................................56
7.1 Background and motivation .............................................................................................56
7.2 Characterization of infrared optical properties of vanadium dioxide .................................57
7.3 Design optimization .........................................................................................................58
7.4 Measurement of infrared device properties ......................................................................62
7.5 Experimental setup and calibration ..................................................................................63
7.6 Discussion .......................................................................................................................68
7.7 Conclusion .......................................................................................................................70
7.8 Future Work .....................................................................................................................70
7.8.1 Optimized Multilayer Planar Structures .....................................................................70
7.8.2 Thermally- regulating Paint .......................................................................................74
References ..............................................................................................................................77

vii

List of Tables

TABLE I. Comparison of reported works on thermo-optic bistabitity in microresonators. The
temperature rise is defined as the difference in temperature between the high-temperature
bistable state and ambient. ....................................................................................................... 17

viii

List of Figures

Figure 1.1. a) Refractive index (left axis) and absorption length (right axis) of crystalline silicon
from [1] b) Absorption in GRM in a silicon PC slab with lattice constant 380nm and hole size 95nm
.................................................................................................................................................. 2
Figure 2.1. (a,b) Schematic of our two-state thermal memory. The memory device is based on a
photonic crystal slab, consisting of a patterned silicon device layer on a glass substrate. For a
given input laser power, we design the device to support two stable temperatures, referred to as
𝑇𝐿 and 𝑇𝐻 . These are used to store a digital bit with value of 0 or 1, respectively. (c) Schematic
of an absorptive, optical resonance in the slab. The red, dashed line marks the wavelength of
laser illumination, and the arrow marks the direction that the spectrum shifts due to absorptive
heating. (d) Schematic of the absorbed optical power as a function of temperature (input power)
and the power dissipated due to thermal conduction away from the laser spot (output power). The
black dots indicate the two stable temperature states. ............................................................... 6
Figure 2.2. (a) Experimental optical transmission and absorption of the photonic crystal slab. A
SEM image of the fabricated slab is shown in the inset. (b) Measured spectral shift as a function
of input laser power, as determined from measurements at a probe wavelength of 920nm (left
axis). The corresponding device temperature is shown on the right axis. The red and blue curves
show the results for heating (increasing input power) and cooling (decreasing input power),
respectively. ............................................................................................................................... 7
Figure 2.3. (a) Measured steady state output power in response to input powers increasing 0-
70mW (heating) and decreasing 70-0mW (cooling). (b) Input waveform of laser power that
includes two consecutive writing and two consecutive erasing pulses. (c) Transmitted laser power
in response to the input waveform in (b)....................................................................................11
Figure 2.4. (a) Experimental input laser waveform. (b) measured transmitted power in response
to waveform in (a). ...................................................................................................................12
Figure 2.5. (a) 2D COMSOL geometry and simulated heat distribution. (b) Simulated incident
laser waveform, (c) transmitted power, and (d) temperature profile in response to waveform in (b).
.................................................................................................................................................13
Figure 2.6. Output power levels at the standby power input of 40mW after 1-10
6
write/ erase
cycles. .......................................................................................................................................14
Figure 2.7. Simulated temperature of the memory device as a function of the incident laser power
for ambient temperatures 20-60°C. ...........................................................................................15
Figure 2.8. (a) Measured transmission spectrum (dashed line indicates the laser wavelength). (b)
Measured wavelength shift (left axis)/ temperature (right axis) as a function of incident laser power
for ambient temperature of 25 and 100°C. ................................................................................16
ix

Figure 2.9. (a) Peak absorption as a function of intrinsic and extrinsic decay rates, and (b)
corresponding quality factor. Black and maroon circles mark the two fabricated devices, and A-C
mark three different devices with a quality factor of 100 and decreasing peak absorption. ........18
Figure 2.10. (a) Absorptivity spectra of four resonators A-D, with quality factor of 100, and peak
absorptivity of 0.5, 0.1, 0.01 and 0.001, respectively. (b) Corresponding temperature bistablity
curves to devices A-D when illuminated by a laser with 11nm detuning from resonance. (c)
Maximum ambient temperature that the device can passively tolerate before bistability is lost at
the standby power. (d) Minimum power required for memory operation (minimum writing power).
Q is quality factor, d is the device’s diameter and Δλ is the laser detuning. ...............................19
Figure 2.11. Effect of varying quality factor (a), laser detuning (b) and device size (c) of the
memory cell on the maximum operation temperature (left y-axis) and the power requirements
(right y-axis). .............................................................................................................................20
Figure 3.1. schematic of the photonic crystal microfabrication process and transfer to a glass
slide. .........................................................................................................................................23
Figure 3.2. (a,b) Schematic of a plain silicon slab and a photonic crystal nanoheater slab each
illuminated at normal incidence by a plane wave, (c) the corresponding measured transmission
spectrum and FDTD fitted transmission and absorption spectra for the unpatterned slab, (d)
measured and CMT-fitted transmission spectra of the nanoheater slab; dotted blue and red lines
show the laser wavelengths used in measurements. (e, f) Transmitted power from the unpatterned
and photonic-crystal slabs, normalized to the incident power. ...................................................24
Figure 3.3. (a, b, c) Simulated time-dependent transmission, temperature, and absorption. .....26
Figure 3.4. Schematic of a programmable array of microheaters. Different colors are devices of
different resonance wavelengths and/or detuning between laser and resonance wavelengths. 27
Figure 3.5. Experimental transmitted power through the photonic-crystal nanoheater, normalized
to the incident power of 201 mW. Detunings of 4 nm and 6 nm are achieved using laser
wavelengths of 974 nm and 976 nm, respectively. ....................................................................28
Figure 3.6. (a, b) The measured spectra of the two fabricated devices around 805nm and 970nm.
(Inset of the proposed used of these structures to form an array). For this specific measurement,
the two devices were fabricated on different samples and illuminated independently, and (c, d)
Experimental transmitted power through both devices, normalized to the incident power of 140
mW using laser wavelengths of 805.5 nm (green) and 976nm (blue). .......................................29
Figure 3.7. (a) The measured spectra (monitored around 970 nm) of device 2 under CW
illumination by the 805.5 nm laser at various powers. Inset is a microscopic image of the bubble
(black circle to the left) and an unexcited device of 100 µm diameter (dark circle to the right), and
(b) the spectral shift as a function of the laser power before bubble formation. .........................30
Figure 4.1. (a) and (b) Schematics of the thermal inverter gate for insulating and metallic VO 2
states, respectively. (c) Schematic of the thermal response of the PC slab to varying input power.
(d) Truth table showing the thermal inversion. ...........................................................................35
x

Figure 4.2. Measured transmission spectrum of (a) a 500um LSAT substrate, and (b) VO2/ LSAT
stack. ........................................................................................................................................36
Figure 4.3. (a) Schematic of the structure simulated in COMSOL. (b) Calculated thermal
hysteresis curve for the PC slab. (c-e) Simulated response of the thermal inverter design in Fig.
4(a, b); such that (c) is input VO2 temperature, (d) is the VO 2/ LSAT transmissivity response and
(e) is the output PC temperature. ..............................................................................................36
Figure 4.4. (a) Measured optical bistability curve in the PC slab. (b) Projection of the measured
VO2/ LSAT samples in terms of transmission change between insulating to metallic VO 2 states.
.................................................................................................................................................37
Figure 4.5. (a) Schematic of proposed universal “NAND” or “NOR” gate. (b) and (c) Schematics
of power choice and truth table corresponding to each gate. .....................................................38
Figure 4.6. (a) Schematic of unilluminated template of alternating microheater designs with
floating heat-sensitive proteins (ELP’s), and live cells. (b) Schematic of one heater type selectively
activated by laser illumination and attracting the corresponding cells/ proteins. (c) Schematic of
both heaters activated by laser illumination and attracting the corresponding cells/ proteins. ....39
Figure 4.7. COMSOL simulations of thermal and fluidic response to laser illumination of a Si PC
slab microheater positioned (a) horizontally, and (b) vertically. .................................................40
Figure 4.8. Schematic of various forces on a particle placed in the locally heated microfluidic
chamber. ...................................................................................................................................40
Figure 4.9. (a) Calibration of spectral shift as a function of device’s temperature. (b) Response of
a fabricated device to various input laser power mapped to the calibration in (a). .....................41
Figure 4.10. Microscopic snapshots of proteins local phase transition evolution with time. .......42
Figure 6.2. Thermal emissivity switching by tuning resonance wavelength (a) and coupling
coefficient (b). ...........................................................................................................................48
Figure 6.3. Normalized integrated emissivity switching predicted by coupled mode theory for a
two coupled resonator design. Upper panel is for the case when the two resonators are thermally
emitting, and the lower panel is for the case when only one resonator is thermally emitting. .....49
Figure 6.4. FDTD simulations of the cross design in (a) using a dipole cloud and a normal
incidence plane wave. Insets are a schematic of the design, and the field profiles for the two
resonance modes. (b) Dark and bright resonances wavelengths as a function of the cross length
L................................................................................................................................................50
Figure 6.5. FDTD simulation of Dark-Bright mode absorptivity/ emissivity spectrum. Insets are a
schematic of the simulated structure and the magnitude field profile for the two resonance peaks.
.................................................................................................................................................51
xi

Figure 6.6. (a) FDTD normal incidence simulation of a two coupled resonator design with a varying
distance d. (b) Corresponding derived coupled mode theory model for a two coupled resonator
design with a varying coupling coefficient β. ..............................................................................52
Figure 6.7. (a) FDTD normal incidence simulation of a bright mode in a cross metamaterial
resonator (inset shows the schematic and field profiles). Dotted line shows the location of the dark
mode not showing in the normal incidence spectrum. (b) Normal incidence simulation of two
identical resonators coupled at 0.05μm (inset shows the schematic and field profiles). .............53
Figure 6.8. (a) FDTD normal incidence simulation of the dark-dark mode for varying distance (d)
(inset shows the schematic and field profiles). (b) Corresponding peak absorptivity and quality
factor to varying coupling (distance (d)).....................................................................................54
Figure 6.9. (a) FDTD normal incidence simulation of the dark-dark and bright-bright modes for
varying dielectric refractive index (inset shows the schematic). (b) Corresponding dark-dark peak
switching. ..................................................................................................................................54
Figure 7.1. Complex dielectric function of metallic (red) and insulating (blue) states of a 98nm-
thick vanadium oxide film as determined from spectroscopic ellipsometry measurements. Solid
lines give the real part (n) and dashed lines give the imaginary part (k) of the VO 2 refractive index.
.................................................................................................................................................58
Figure 7.2. Schematic and SEM picture of the micro-cones fabrication process. .......................59
Figure 7.3. Calculated 𝛥 𝑃𝑟𝑎𝑑 as a function of height and VO2 thickness for (a) Silicon micro-cones,
(b) parabolic micro-cones, and (c) micro-rods with a fixed period of 1µm. .................................60
Figure 7.4. A comparison of different structures feasibility and switching performances. ...........60
Figure 7.5. Calculated difference in integrated thermal radiation power for an isolated VO 2 thin
film, VO2 thin film with a handle layer of Si, and VO 2/Si/Au multilayer stack. Corresponding
schematics and layer thicknesses are shown in the inset. .........................................................61
Figure 7.6. (a) Schematic of VO 2/ Si, and (b) schematic of VO 2/ Si/ Au layered structures with
respective simulated (c), (d) emissivity and (e), (f) transmissivity corresponding to VO 2 insulating
and metallic states. ...................................................................................................................62
Figure 7.7. Measured and simulated (absorptivity) emissivity spectra of insulating (a) and metallic
(b) state of VO2/Si/Au stack. (c) Calculated integrated radiation power normalized to blackbody
spectrum. ..................................................................................................................................63
Figure 7.8. Overview of experimental setup. (a) Photograph and (b) schematic of vacuum
chamber used to perform thermal measurements. (c) Measured temperature rise as a function of
applied heat load for mirror-gold (low emissivity) and diffuse black (high-emissivity) samples, with
respect to an ambient temperature of 0.5˚C. (d) Measured parasitic heat loss (sum of radiation
losses from perimeter and conduction losses to wire), as a function of temperature, extracted from
mirror gold data in (c). ...............................................................................................................64
xii

Figure 7.9. (a) Measured temperature rise relative to ambient (horizontal axis) as a function of the
applied heat load (vertical axis) for a complete heating and cooling cycle. The inset shows the
derivative applied heat load with respect to temperature for two runs (complete heating/cooling
cycles). (b) Calculated radiative heat flux from the VO 2 surface as a function of temperature.
Constant emissivity curves are plotted in grey. The inset shows the effective emissivity of the
measured sample as function of temperature. ..........................................................................66
Figure 7.10. Thermal Homeostasis. (a) Square wave time-varying input heat power. (b) measured
and calculated response to the input power in for VO 2/ Si/ Au (62nm/ 200µm/ 60nm) structure and
Al2O3/ Si/ Au (480nm/ 200µm/ 60nm). .......................................................................................68
Figure 7.11. Thermal homeostasis in space. (a) Square wave time-varying input heat power. (b)
calculated response to the input power in for VO2/ Si/ Au (62nm/ 200µm/ 60nm) structure and
Al2O3/ Si/ Au (480nm/ 200µm/ 60nm). ......................................................................................69
Figure 7.12. Flow chart of the random walk optimization algorithm. ..........................................71
Figure 7.13. (a) Metallic radiation power as a function of rod height and VO 2 thickness. (b) High
aspect ratio optimal point of metallic P rad. (c) Low aspect ratio optimal point of metallic P rad. .....72
Figure 7.14. (a) Result of applying random walk optimization to a 2x2 unit cell. (b) Metallic state
emissivity spectrum of the initial and final states. ......................................................................73
Figure 7.15. (a) Simulation of the radiation power switching of the four layered structures in (b-d).
.................................................................................................................................................73
Figure 7.16. (a) Schematic of a Si/VO2/ZnSe/Au thermal homeostasis structure. (b) Effect of the
thickness of the ZnSe layer on the normalized thermal radiation power. ...................................74
Figure 7.17. Schematic of the microparticle-based paint. A single layer of close-packed VO2-
coated Si microspheres. The entire microstructure is embedded in polyethylene and resting on
an Al substrate. .........................................................................................................................75
Figure 7.18. Thermal homeostasis and temperature regulation. (a) The microsphere “paint.” (b) A
planar, VO2-coated film. (c) Uncoated film. (d) Boundary conditions used to solve the heat
equation. (e) Radiated thermal power as a function of temperature. The arrows indicate the
direction of heating or cooling processes. The symbols represent the calculated 𝑃𝑟𝑎𝑑𝑇𝑐 for
metallic (squares) or insulating (circles) states. (f) Temperature variation of different structures for
a time-varying heat input flux. ...................................................................................................76

xiii

Abstract

This dissertation studies light interaction with matter in engineered microstructures for two
applications: hybrid opto-thermal devices (Part I), and infrared thermal radiation management
(Part II).
The work on the opto-thermal devices focuses on the use of nano-patterned silicon to enhance
light absorption and control heat on the micro-scale. We utilize critical coupling condition in silicon
photonic crystal resonators to create highly absorptive peaks. This complements the efforts on
plasmonic nanostructures as a primary tool for heating at the micro/ nanoscale. Beside the cost
efficiency of using patterned silicon to control heat at microscale compared to plasmonic, we also
present a scheme where thermo-optic nonlinearity in silicon can be utilized to achieve
programable on-chip heating. Moreover, this scheme is used to create a high temperature thermal
memory that uses laser to read, write and erase digital information, and can potentially operate in
harsh, high temperature environments.
The work on infrared thermal radiation management focuses on developing schemes that use
photonic microstructures to shape thermal emission. We experimentally realize a proof-of-
concept broadband thermal emissivity switching; a property that is used for self-temperature
regulation in space (passive thermal homeostasis). To achieve this, we use a phase change
material, vanadium dioxide (VO2), to provide the emissivity switching, and optical simulations for
design optimization, presenting the first experimental demonstration of a radiative passive thermal
homeostasis device. We next study a coupled-resonators scheme to shape infrared emissivity
spectrum by tuning refractive index and coupling coefficient. We use temporal coupled mode
theory to derive a closed form expression of the emissivity spectrum and FDTD simulations to
verify the effect in a metamaterial perfect absorber design.

xiv

Part I: Hybrid Opto-Thermal devices
1

Chapter 1
1. Introduction: Part I
Photonic crystal slabs (PCS) featuring two dimensional (2D) arrays of periodic holes, exhibit
Fano shaped resonances on a smoothly varying background due to the interference between the
Fabry-Perot oscillations of the slab and the guided resonance modes (GRM) excited within the
slab [1, 2]. Due to the resonantly enhanced optical near-field, these devices are extensively used
as filters [3, 4] and sensors [5, 6]. They have also been used for enhancement of fluorescence in
dyes [7, 8], single molecules [9], and colloidal quantum dots [10]. Properly designed modes in
these structures leads to enhanced coupling of normally incident light. These modes have shown
to increase the light extraction from lasers [11, 12] and LEDs [13, 14]. However, most of the
silicon-based PC slabs focus on GRMs in optical communication band where silicon is non-
absorptive resulting in high quality (Q) factor modes [15].
Recently, however, due to the proliferation of the solar research, there has been interest in
exploiting GRMs in the absorptive regime of silicon. Too high of absorption in the substrate
destroys the resonance due to the loss associated with increased absorption. However, silicon is
moderately absorptive between the wavelength range of 365nm to 1127nm; this range
corresponds to the separation between silicon's direct and indirect bandgap (Figure 1.1a) energy.
This region of low absorptivity leads to the existence of resonance modes [16-18] with moderately
high quality factors (a few hundred). Due to the resonantly enhanced electric field within the
substrate, the GRMs are associated with enhanced absorption (Figure 1.1b) which is exploited in
photovoltaic applications. This suggests that we should be able to achieve resonantly enhanced
thermo-optic effects in this regime. This is of particular interest for achieving nonlinear absorption
and heat control as in this proposed work.
2

Figure 1.1. a) Refractive index (left axis) and absorption length (right axis) of crystalline silicon from [1] b)
Absorption in GRM in a silicon PC slab with lattice constant 380nm and hole size 95nm

3

Chapter 2

2. High Temperature Opto-thermal Memory
A version of the results in this chapter was published as Ref. [19]
2.1 Motivation for Opto-thermal Devices
The powerful computational capabilities provided by modern electronics ultimately fail in
extreme environments, driving interest in the development of alternative computation
architectures [20-23] . Thermal analogues to electronic devices have been proposed as
alternative route to storing and processing information, including thermal diodes [24-27],
transistors [28, 29], logic [30, 31], and memories. The basic element in a thermal memory is the
ability of a device to support bistable states at two different temperatures; the two states are used
to encode a single bit of information. Theoretical proposals for thermal memories have used a
variety of different physical mechanisms to provide bistability [32-36]. Experiments have focused
on the use of the metal-insulator phase transition in vanadium dioxide to demonstrate thermal
memory operation [37, 38]. To date, switching times are on the order of 10’s of milliseconds or
greater in theory [32-36], and larger than 1 second in experiment [37, 38]. In this paper, we
propose and experimentally demonstrate an alternative approach to thermal memories that uses
an all-silicon system to reduce the experimental switching times below 500µs.
Our approach is based on the extensive literature on optical bistability in integrated photonic
systems. Optical bistability in microresonators has been studied both theoretically and
experimentally at wavelengths near 1550nm [39-48]. If the refractive index of a material changes
with increasing light intensity, the resonance wavelength of a microresonator shifts with input
optical power. The nonlinear feedback can give rise to optically bistable states. A subset of work
on optical bistability has focused on thermo-optic bistability, where the intensity-dependent
4

wavelength shift arises from absorptive heating, via the thermo-optic effect. In this case, the
bistable states correspond to slightly different temperatures [40, 48-51]. As a result, this class of
devices can, in fact, be viewed as thermal memories. In this case, a laser is used to read and
write a bit of information, while different temperatures inside the device are used to encode the 0
and 1 states.
In this work, we consider thermal memory operation within high- and fluctuating-temperature
environments. For this purpose, we deliberately increase the temperature difference between 0
and 1 states by an order of magnitude or greater relative to previous experiments [46, 52]. To
achieve this goal, we shift our operating wavelength to the 800-1000nm range and exploit direct
absorption in silicon, while using a resonant mode to achieve strong thermo-optic nonlinearities.
We thus achieve thermo-optic bistability with a temperature difference between states of more
than 200°C. Below, we demonstrate reading, writing, and erasing of a single bit. Using a photonic
device provides extremely high endurance, in excess of 1 million cycles. Moreover, our approach
allows operation up to high temperatures at which electronic memories fail, as well as operation
in environments where the temperature is highly fluctuating. Our proof-of-concept devices have
measured temperature ranges as large as 25-100°C. In addition, our use of the all-silicon platform
offers the advantages of a standard, CMOS-compatible fabrication process with no need for
vacuum operation.
2.2 Physical Origin of Thermo-optic Bistability
The geometry of our thermal memory device is shown in Fig. 2.1a. This type of structure is
commonly known as a photonic crystal slab[1]. It consists of a thin silicon device layer, patterned
by a periodic array of holes, resting on a silicon dioxide substrate. Our goal will be to encode a bit
of information in the internal temperature of the slab. This is indicated schematically in Figs. 2.1a
and 2.1b, where 0 and 1 values of the bit are chosen to correspond to a lower (𝑇 𝐿 ) and higher
5

(𝑇 𝐻 ), temperature, respectively. Below, we explain how to design the photonic crystal to support
bistable temperature states.
The key feature of our approach is an absorptive, optical resonance [53]. We have previously
studied such resonances in Refs. [54-57]. The physical meaning of such a resonance can be
understood from Fig. 2.1c. For a low-power laser incident perpendicular to the photonic crystal
slab, the light absorption depends on the laser wavelength (blue curve). The absorption peaks at
the resonant wavelength. In the photonic-crystal literature, this mode is known as a Γ-point mode
of the band structure, lies above the light line, and can be excited with normally-incident light [1].
The optical absorption is nonlinear with laser power. If the laser wavelength is chosen slightly
above the low-power absorption curve (wavelength indicated by red, dashed line), the initial
absorption is low. But as light is absorbed, the slab starts to heat up. Due to the thermo-optic
effect in silicon, its refractive index changes with power. The change in refractive index shifts the
resonance peak to the right (black arrow). The steady-state shift depends on laser power.
Using coupled mode theory (CMT) [53], we can relate the absorbed laser power to the internal
temperature of the slab [57]. The result is a peaked function of temperature, shown schematically
in Fig. 2.1d (red line). Mathematically,
𝑃 abs
= 𝑃 in
2𝛾 𝑖 𝛾 𝑟 (𝜔 𝑜𝑝
− 𝜔 0
+
𝜔 0
𝑛 0

𝑑𝑛
𝑑𝑇
ΔT)
2
+ ( 𝛾 𝑖 + 𝛾 𝑟 )
2

,
Eq. 2.1
where 𝜔 𝑜𝑝
is the operating frequency of the laser, equal to 2𝜋 𝑐 /𝜆 𝑜𝑝
, 𝑐 is the speed of light in free
space, 𝛾 𝑟 is the decay rate of the resonance due to radiation loss, 𝛾 𝑖 is the decay rate due to
material absorption loss (assumed to be independent of pump power), 𝑑𝑛 /𝑑𝑇 is the thermo-optic
coefficient of silicon, and 𝑛 0
is the refractive index of silicon at room temperature. The absorbed
power heats the device.
6

Figure 2.1. (a,b) Schematic of our two-state thermal memory. The memory device is based on a photonic
crystal slab, consisting of a patterned silicon device layer on a glass substrate. For a given input laser
power, we design the device to support two stable temperatures, referred to as 𝑇 𝐿 and 𝑇 𝐻 . These are used
to store a digital bit with value of 0 or 1, respectively. (c) Schematic of an absorptive, optical resonance in
the slab. The red, dashed line marks the wavelength of laser illumination, and the arrow marks the direction
that the spectrum shifts due to absorptive heating. (d) Schematic of the absorbed optical power as a function
of temperature (input power) and the power dissipated due to thermal conduction away from the laser spot
(output power). The black dots indicate the two stable temperature states.
Conductive heat loss cools the device. Since the incident laser beam is finite in size and
covers the center of the total device area, power is conducted away from the beam spot to the
surroundings. Due to the high thermal conductivity of both silicon and glass, conductive loss
dominates over radiative and convective cooling, for sufficient temperature rise at the beam spot.
We can assume that the conductive loss increases as the temperature gradient increases, with a
functional form of 𝑘 ∇T.
We can find the steady-state temperature values by setting the absorbed power equal to the
conductive loss, following Ref. [52]:
7

𝑘 ∇T = 𝑃 in
2𝛾 𝑖 𝛾 𝑟 (𝜔 𝑜𝑝
− 𝜔 0
+
𝜔 0
𝑛 0

𝑑𝑛
𝑑𝑇
ΔT)
2
+ ( 𝛾 𝑖 + 𝛾 𝑟 )
2

,
Eq. 2.2
2.3 Experiment
The photonic crystal is made by patterning a square array of circular holes in a dielectric slab,
using standard fabrication methods [58]. The inset to Fig. 2.2a shows a scanning electron
microscope image of the fabricated device. In brief, we used electron-beam lithography and
reactive-ion etching to pattern a silicon on insulator (SOI) wafer, with a silicon slab thickness of
340nm, a lattice constant (center-to-center distance between holes) of 380nm, and a hole
diameter of 129nm. The total patterned area is circular, with a diameter of 100μm. After patterning,
the silica was removed via wet etching, and the 340nm silicon device layer was transferred to a
1.6mm-thick, ground and polished glass disc [58]. Glass is transparent at the laser wavelength
used in the experiment (805.5nm).

Figure 2.2. (a) Experimental optical transmission and absorption of the photonic crystal slab. A SEM image
of the fabricated slab is shown in the inset. (b) Measured spectral shift as a function of input laser power,
as determined from measurements at a probe wavelength of 920nm (left axis). The corresponding device
temperature is shown on the right axis. The red and blue curves show the results for heating (increasing
input power) and cooling (decreasing input power), respectively.
8

The normal-incidence, low-power transmission spectrum of the photonic crystal was
characterized using an unpolarized, broadband white-light source and a spectrometer (Ocean
Optics USB 4000). The measured transmission spectrum is depicted by the solid black line in Fig.
2.2(a). It shows a prominent dip at 797.5nm, corresponding to the resonance wavelength. The
resonance wavelength and quality factor can be tuned by changing the photonic crystal hole size
or spacing [59].
We used the CMT model to fit the experimental transmission spectrum and calculate the
corresponding absorption. We obtained the fitting parameters of 𝛾 𝑟 =6.2 THz, 𝛾 𝑖 =2.9 THz, 𝑡 𝑠 =0.9,
and 𝑟 𝑠 =0.1, where 𝑡 𝑠 and 𝑟 𝑠 are the direct transmission and reflection coefficients, respectively.
Figure 2.2a shows that fitted and experimental results for transmission are in good agreement.
Using the given fitting parameters, we obtain the absorption spectrum shown by the blue line in
Fig. 2.2(a). At resonance, the absorption is 50%, which is approximately 100x higher than the
absorption of the unpatterned slab [56].
To demonstrate thermal bistability, we first measured the spectral shift at a probe wavelength
due to laser heating, and then used a calibration experiment to map the spectral shift to
temperature.
To measure spectral shift, we illuminated the fabricated device with a continuous wave (CW)
laser with a wavelength of 805.5nm. This value is indicated by the dashed red line in Fig. 2.2(a).
The diameter of the laser spot focused on the device was 14µm. The laser power was manually
increased from 0-70mW in discrete steps. For each value of laser power, we measured the
spectral shift by monitoring the peak position of a second resonance mode, close to 920nm, with
the spectrometer. The shift in the peak position for the second mode is denoted as ∆𝜆 . A long
pass filter with a cut-on wavelength of 850nm was used at the input to the spectrometer to avoid
saturation from the CW laser. The same procedure was then repeated for decreasing power from
70-0mW.
9

The left axis of Figure 2.2b shows the spectral shift at the probe wavelength as a function of
laser power for increasing and decreasing power. For laser powers between 30 and 50mW, two
different values of spectral shift are possible. One is obtained by increasing the power from zero,
while the other is obtained by decreasing the power from a value above 50mW.
To translate the spectral shift into a temperature, we conducted a calibration experiment, as
in Ref. [57]. Using a Thorlabs heating lens tube to heat the device externally, we monitored the
device temperature using a SeeK™ Thermal camera. Simultaneously, we measured the spectral
shift corresponding to this known temperature. We used the calibration data to map the measured
spectral shift induced by laser heating, shown in Fig. 2.2b, to a temperature. The temperature is
shown on the right axis.
The results show that the device has two stable temperature states for laser powers between
approximately 30 and 50mW. For a given laser power, the internal temperature of the device
depends on the illumination history. Referring to the schematics in Figs. 2.1c and 2.1d, we may
interpret the experimental results as follows.
Suppose we start from zero laser power and increase to a fixed value within the bistable
region; we will choose 40mW for concreteness. The absorptive peak will shift to higher
wavelength (to the right on Fig. 2.1c). Meanwhile, the temperature will increase from room
temperature until the absorptive and conductive power balance. This occurs at the lower of the
two bistable states in Fig. 2d, 𝑇 𝐿 . From Fig. 2b, an input power of 40mW corresponds to 𝑇 𝐿 = 52°C
for the lower-temperature state.
Now suppose instead that we start at a very high laser power, corresponding to much higher
temperature. As the laser power is decreased, the temperature falls until it settles at the higher-
temperature state, 𝑇 𝐻 . From Fig. 2.1b, the temperature in the higher temperature bistable state is
𝑇 𝐻 = 84°C for an input of 40mW.
10

The ability of the device to take on two different temperatures at the same laser power forms
the basis of our 1-bit thermal memory. We will use the laser to control the device temperature,
demonstrating that we can toggle between 𝑇 𝐿 (“0”) and 𝑇 𝐻 (“1”) states. As shown below, the laser
transmission also provides a convenient method for reading out the bit’s value.
2.3.1 Thermal Memory

To operate our thermal memory, we must choose values for the laser operating power. It is
convenient to observe the operation of the memory by measuring the transmitted laser power
through the device. Fig. 2.3a shows the measured output power as a function of input laser power,
for both heating (increasing power) and cooling (decreasing power) curves. As in Fig. 2.2b, we
clearly observe a hysteresis loop. We choose a “standby” power of 40mW, in the middle of the
hysteresis loop. The standby power is the default value of the laser, in between write and erase
pulses. At the standby power, the output power can take on one of two values, 34mW (state “0”,
the lower temperature state) or 19mW (state “1”, the higher temperature state). We choose an
input write power of 56mW, to the right of the hysteresis loop (right-most yellow dot in Fig. 2.3a)
and an erase power of 22mW, to the left of the hysteresis loop (left-most yellow dot).
The measured input power to the thermal memory as a function of time is shown in Figure
2.3b. We include two consecutive write pulses, followed by two consecutive erase pulses. After
each write or erase pulse, the input power returns to the standby value. The laser pulses were
generated using a Labview control program and have a duration of 1.2ms. The input power was
measured using a transimpedance amplifier (Thorlabs) with a response time on the order of a few
picoseconds.
Figure 2.3c shows the measured output power as a function of time. The measurement was
again made with a transimpedance amplifier. Initially, the output power is approximately 34mW.
At the conclusion of the write pulse, the output power falls to approximately 19mW. These two
output power levels correspond to states “0” and “1”, respectively. The write pulse thus flips the
11

thermal memory state from “0” to “1”. As expected, applying a second write pulse has no effect
on the state. At the conclusion of the write pulse, the device remains in state “1”. Applying an
erase pulse flips the state from “1” back to “0”. Again, a second erase pulse has no effect on the
state, leaving the device in state “0”. This experiment directly illustrates the operation of the
memory device.
During the write and erase pulses, the device exhibits transient behavior. The time-dependent,
output power shown within the grey-shaded regions of Fig. 2.3c arises from the dynamics of the
resonance shift. This gives rise to brief spikes and dips at the beginning of the first write pulse
and the first erase pulse (boundary of white and grey-shaded regions). However, the output power
quickly stabilizes at the end of each write or erase pulse, easily allowing determination of the
output power in the standby state (white regions) and corresponding readout of the “0”/ ”1” value
of the bit. We note that our memory is volatile; to retain a bit of information, the laser power must
be maintained at its stand-by value. Turning the laser off completely will erase the bit.

Figure 2.3. (a) Measured steady state output power in response to input powers increasing 0-70mW
(heating) and decreasing 70-0mW (cooling). (b) Input waveform of laser power that includes two
consecutive writing and two consecutive erasing pulses. (c) Transmitted laser power in response to the
input waveform in (b).
12

2.3.2 Memory Speed
The speed of our thermal memory device will fundamentally be determined by the response
times for heating and cooling of the device. To provide an upper bound, we illuminated the device
with the input laser waveform shown in Fig. 2.4a. In this experiment, the power changes directly
from the minimum power (erase) to the maximum power (write) and vice versa, corresponding to
the highest and lowest temperatures of the device during normal operation. Fig. 2.4b shows that
the measured transient response is on a time scale less than 500µs for both heating and cooling.
This time is at least one order of magnitude better than the response times recently reported in
Refs. [32-36].

Figure 2.4. (a) Experimental input laser waveform. (b) measured transmitted power in response to
waveform in (a).
We can reproduce similar results for the thermal response time via numerical simulations.
Details of our approach are given in Refs. [56, 57]. We used the CMT fitted parameters from Fig.
2.2a to calculate the time dependent absorption, which was fed to COMSOL to solve the time-
dependent 2D heat equation. The geometry used in the simulation is shown in Fig. 2.5a. Side and
bottom boundaries were set to room temperature, while top boundary was thermally insulated. A
heat source with a Gaussian spatial profile (full width at half maximum of 14μm) is positioned at
the disk center. To approximately model the effect of the holes, we used the same modified silicon
parameters as in Ref. [56]. Figure 2.5c shows that the time-dependent output power in response
13

to the input power in Fig. 2.5b is in good agreement with the measurement in Fig 2.4b. The
temperature response with time is shown in Fig. 2.5d and illustrates toggling between high and
low temperature states.

Figure 2.5. (a) 2D COMSOL geometry and simulated heat distribution. (b) Simulated incident laser
waveform, (c) transmitted power, and (d) temperature profile in response to waveform in (b).
2.3.3 Memory Reliability

The reliability of memory devices is typically described by the write/ erase endurance. This
can be quantified by the number of the write/ erase cycles the memory device can undergo before
its two states become indistinguishable. The oxide traps in NAND flash memories alter the
threshold voltage, causing a large degradation after 1K write/ erase cycles [60, 61]. The
endurance of our thermo-optic memory was tested by running periodic write/ erase pulses
separated by standby pulses. Each pulse for this experiment was 3ms, which makes a complete
write/ erase cycle of 12ms (including a standby pulse after each write/erase). We ran the
experiment for 10
6
cycles and the output power levels for both states were monitored as shown
in Fig. 2.6. Our device shows less than 3% change in the output levels after 10
6
cycles, yielding
an extremely high endurance compared to conventional electronic flash memories.
14

Figure 2.6. Output power levels at the standby power input of 40mW after 1-10
6
write/ erase cycles.
2.3.4 High-temperature operation

The thermal hysteresis loop shown in Fig. 2.2b is the key element that allows the memory
function, i.e. storage of binary information. The hysteresis loop depends on the initial detuning
between the laser and the resonance wavelengths. Changing the ambient temperature shifts the
resonance, decreasing detuning and changing the hysteresis loop. Using the model described in
section 2.3.2 and the parameters of the device shown in Fig. 2.2a, the temperature dependence
of the hysteresis loop was simulated in Fig 2.7. As the ambient temperature increases, the loop
moves to a lower power range and becomes narrower. For this particular device, the hysteresis
loop persists up to a temperature of 60°C. In general, the maximum allowable temperature the
device can support depends on the quality factor and the vicinity to critical coupling (see
Supplemental Information). For operation at fixed wavelength, the power requirements are
minimized by operating near critical coupling and using the lowest quality factor consistent with
the desired maximum environmental temperature.
15

Figure 2.7. Simulated temperature of the memory device as a function of the incident laser power for
ambient temperatures 20-60°C.
For a larger detuning between the resonance and the laser wavelength, a higher temperature
difference between the two memory states can be achieved. Figure 2.8a shows the measured
transmission spectra of a fabricated device with a guided mode resonance near 965nm. The
wavelength shift of the device was monitored under illumination with a laser wavelength of 976nm,
as shown in Figure 2.8b (11nm detuning). The hysteresis width is approximately 120mW, and the
temperature difference between the two states is approximately 200°C, compared to 20mW and
50°C for the device in Fig. 2.2b.
To demonstrate high ambient temperature operation, the device was heated using a resistive
heater, and the temperature was monitored using a SeeK™ thermal camera. The red curve in
Fig. 2.8b shows that the device shows a hysteresis loop at an ambient temperature as high as
100°C. This memory device can thus operate over a temperature range of at least 25-100°C.
Due to the overlap between the loops, the memory can operate in a highly-fluctuating
temperature environment. We can, for example, choose the standby, write and erase powers to
be 70, 180 and 20mW, and the boundary between “0” and “1” states to be at 𝛥𝜆 =10nm. Then the
two states could clearly be distinguished, even for ambient temperature fluctuations over the
entire 25-100°C range.
16

Figure 2.8. (a) Measured transmission spectrum (dashed line indicates the laser wavelength). (b) Measured
wavelength shift (left axis)/ temperature (right axis) as a function of incident laser power for ambient
temperature of 25 and 100°C.
2.4 Discussion
The switching times demonstrated here are much shorter than for previously-reported thermal
memories. To achieve further reduction in switching time, alternative resonator designs may be
useful. Examining the literature on thermo-optic bistability in microresonators, a number of
devices with shorter switching times have been demonstrated (Table I) [40, 49, 62-64]. It is
important to note that these works did not consider memory performance within a fluctuating-
temperature environment; the temperature of the surrounding environment was kept fixed. As a
result, lower temperature rises, corresponding to lower operating powers, could be used.
However, the resonator geometries considered in Table I, which have more tightly-confined
modes than the guided-resonance devices we study here, may help reduce the switching time in
our experiments. While these past works did not aim to achieve critical coupling, a favorable
condition for our current application, modified designs that incorporate both smaller mode volume
and critical coupling are an interesting direction for future research.

17

TABLE I. Comparison of reported works on thermo-optic bistabitity in microresonators. The temperature
rise is defined as the difference in temperature between the high-temperature bistable state and ambient.
Device Time (μs)
Wavelength
(nm)
Power
(mW)
Temperature rise
ΔT (K)
Wavelength
shift (nm)
Ref.
Si high-Q PhC cavity 0.1 1535 0.04 0.1 [40]
InP PhC nano-cavity 4 800 0.165 3.2 0.3 [62]
Si MIM disk resonator 2.6 1555 0.5 1.6 0.15 [63]
Graphene-on-Si3N4
ring resonator
0.235 1555 40 0.3 [64]
Si PhC guided mode
resonance
500 800-1000 50-170 50-250 5-17
This
work

From coupled-mode theory [65], the absorptivity of our resonator is given by
𝐴 =
2𝛾 𝑖 𝛾 𝑟 ( 𝜔 − 𝜔 0
)
2
+ ( 𝛾 𝑖 + 𝛾 𝑟 )
2

, Eq. 2.3
where 𝛾 𝑖 and 𝛾 𝑟 are the intrinsic (absorption) and extrinsic (radiative) decay rates, respectively.
Figure 2.9(a) shows the peak absorption as a function of both decay rates. The red region marks
the critical coupling condition, 𝛾 𝑖 = 𝛾 𝑟 , at which peak absorptivity of 0.5 is achieved. Figure 2.9(b)
shows the quality factor as a function of the decay rates. Our devices have quality factors of 130
and 150.

18

Figure 2.9. (a) Peak absorption as a function of intrinsic and extrinsic decay rates, and (b) corresponding
quality factor. Black and maroon circles mark the two fabricated devices, and A-C mark three different
devices with a quality factor of 100 and decreasing peak absorption.
To understand the effect of critical coupling and quality factor on our thermal memory, we
carried out thermal simulations solving the time dependent heat diffusion equation in COMSOL
for four different devices. Devices A-D have a quality factors of 100 and peak absorptivity of 0.5,
0.1, 0.01 and 0.001, respectively as shown in Fig. 2.9(a) and illustrated in Fig. 2.10(a). Figure
2.10(b) shows the thermal bistability curves for the four resonators upon laser illumination at 11nm
above the resonance wavelength. The two stable states are at similar temperatures for all four
resonators and the required input power for the existence of bistability increases as peak
absorptivity decreases. Namely, devices A, B, C and D require powers of at least 0.06, 0.3, 1.7
and 17W, respectively.

19

Figure 2.10. (a) Absorptivity spectra of four resonators A-D, with quality factor of 100, and peak absorptivity
of 0.5, 0.1, 0.01 and 0.001, respectively. (b) Corresponding temperature bistablity curves to devices A-D
when illuminated by a laser with 11nm detuning from resonance. (c) Maximum ambient temperature that
the device can passively tolerate before bistability is lost at the standby power. (d) Minimum power required
for memory operation (minimum writing power). Q is quality factor, d is the device’s diameter and Δλ is the
laser detuning.
The maximum ambient temperature allowable by each of the four devices is shown in Fig.
2.10(c). The standby power for each device is chosen to be immediately at the lower jump of the
hysteresis loop. Figure 2.10(c) shows that the maximum ambient temperature is slightly improved
for lower peak absorptivity, at the cost of logarithmically increasing power requirements (Fig.
2.10(d)). It is thus desirable to be near critical coupling. Figure 2.10(c) also shows that as the
quality factor increases, the maximum tolerable ambient temperature increases. However, the
required power increases, as shown in Fig. 2.10(d). It is thus desirable to use the lowest Q that
can accommodate the desired environmental temperature swing.
We further study the effect of varying each of the resonance quality factor, the laser detuning,
and the device diameter on the maximum operation temperature and required power. Figure
2.11(a) and (b) shows that there is a power/ maximum temperature tradeoff for increasing quality
factor and laser detuning. For a given power budget, and device size, quality factor and laser
detuning must be mutually optimized. For the device diameter, however, the required power
rapidly decreases as the device size goes down at the expense of a small change in the maximum
operation temperature (Fig. 2.11(c)). The small device size also improves the device speed which
suggests that smaller device size is always desirable.
20

Figure 2.11. Effect of varying quality factor (a), laser detuning (b) and device size (c) of the memory cell on
the maximum operation temperature (left y-axis) and the power requirements (right y-axis).
21

Chapter 3
3. Heat control on micro-scale
A version of the results in this chapter was published as Ref. [57]
3.1 Background and Motivation
The use of laser illumination to generate heat has led to valuable applications in several fields,
such as nanochemistry [66], optofluidics [67] and biosensing [68, 69]. Noble metal nanoparticles
have widely been used as nanoheaters, due to enhanced light absorption at their plasmonic
resonances[70-72]. Lithographically-defined structures offer further advantages, including
increased absorption strength, the ability to tune the resonance wavelength, and control over the
heat pattern [70, 73]
Recent work by the authors [56] has introduced a dielectric alternative for laser-induced
nanoheaters. In contrast to previous work on metallic particles and films, we use silicon photonic-
crystal slabs as efficient heaters. We have shown that a highly absorptive optical mode can be
obtained by patterning a 2D periodic array of holes inside a silicon slab and operating in the 700-
1000 nm wavelength range, falling within the biological transparency window. Our silicon
nanoheater features a much narrower resonance linewidth than plasmonic structures. As a result,
heating occurs suddenly and rapidly.
In this work, we demonstrate that the thermal response of the device can be “programmed in”
to the nanostructure. The resonant wavelength of the device depends on the hole size and
spacing in the photonic crystal. By adjusting the detuning of the laser from the resonant
wavelength, the response time (delay from laser turn-on to temperature jump) and maximum
temperature rise can be controlled. We study the operation of photonic-crystal nanoheaters in
water for potential lab-on-chip applications. We further report the use of our nanoheaters to
22

generate controllable, stable microbubbles. For CW illumination at 976 nm wavelength, the
observation of microbubbles indicates heating from room temperature to at least 100 °C for a
laser power density of 1.18 mW/μm
2
. The programmable heating capabilities we demonstrate
open up the door for fine spatial and temporal control of temperature on the microscale, with
potential use in lab-on-chip and nano-chemistry applications.
3.2 Time-dependent optical and thermal response in water

First, we experimentally study the time-dependent response of the nanoheater in water. We
compare the response of the photonic-crystal slab nanoheater to an unpatterned slab of the same
thickness, for reference. We measure the optical transmission changes through the slabs as a
function of time. We then develop a model of the relationship between transmission and
temperature, to show how changes in transmission serve as a signature of optical heating.
The photonic crystal nanoheater is formed by patterning a square array of circular holes in a
340 nm thick silicon slab. Laser light is normally incident on the device. Schematic diagrams of
unpatterned and photonic-crystal slabs are shown in Figures 1(a) and 1(b). We will study the
characteristics of both structures, in order to show how nanopatterning affects optical
transmission and enhances heating.
For the photonic crystal slab, the lattice constant is 450 nm, and the hole diameter is 135 nm.
The total patterned area is circular, with a diameter of 100 μm. We followed the same fabrication
procedure in Ref. [74], except that we transferred the silicon membrane to a 1.6 mm-thick, ground
and polished glass disc. A schematic of the photonic crystal microfabrication process and transfer
to a glass slide is shown in Fig. 3.1. A 1.5 mm-thick fluidic chamber filled with water was created
on top of the device, using a lens-holder with a rubber O-ring as a sidewall pressed against an
identical glass disc on the top of the chamber. The normal-incidence transmission spectra of the
photonic crystal, as well as that of an unpatterned silicon slab of the same thickness in the same
23

chamber, were characterized using a broadband white-light source and a spectrometer (Ocean
Optics USB 4000).

Figure 3.1. schematic of the photonic crystal microfabrication process and transfer to a glass slide.
The measured transmission spectra are depicted by the black lines in Figures 3.2(c) and (d).
The unpatterned slab (Figure 3.2(c)) has a nearly featureless transmission spectrum. The small
slope in transmission results from Fabry-Perot effects: the Fabry-Perot fringe spacing is
significantly larger spacing than the wavelength range shown here. The photonic crystal slab
(Figure 3.2(d)) has a Fano resonance near 970 nm. The origin of the characteristic, asymmetric
lineshape of Fano resonances in photonic-crystal slabs is described in detail in Ref. [1].
To calculate the absorption of the unpatterned slab, we used the Lumerical FDTD solver to
simulate transmission and absorption. Figure 1c shows that the unpatterned silicon slab has only
about 0.5% light absorption for the given thickness and wavelength range. For the photonic
crystal slab, we used the methods in Appendix I to fit the experimental spectrum to the CMT
theory model in equation (1) and obtained the fitting parameters of
r
 =3.0177 THz,
i
 =1.916
THz,
s
t =0.65, and
s
r =0.6. Figure 1d shows that fitted and experimental results are in good
24

agreement. Using the given fitting parameters, we obtain the absorption spectrum shown in Figure
3.2(d). At resonance, the absorption is 45%. By nanopatterning the slab, the absorption is thus
enhanced by a factor of about 90.

Figure 3.2. (a,b) Schematic of a plain silicon slab and a photonic crystal nanoheater slab each illuminated
at normal incidence by a plane wave, (c) the corresponding measured transmission spectrum and FDTD
fitted transmission and absorption spectra for the unpatterned slab, (d) measured and CMT-fitted
transmission spectra of the nanoheater slab; dotted blue and red lines show the laser wavelengths used in
measurements. (e, f) Transmitted power from the unpatterned and photonic-crystal slabs, normalized to
the incident power.
To measure the time-dependent optical transmission, we illuminated the slabs with a near-
infrared, 3S Photonics, CHP1999 laser operating at 976 nm. The laser is passed through a 19
mm lens and focused through an objective onto the sample. This results in a Gaussian beam with
a spot size of 30 μm on the sample. The laser intensity is modulated using a laser diode controller
(Thorlabs) using a square pulse with modulation frequency of 5 Hz and 0.3% duty cycle. The rise
time of the laser pulse is approximately 1 μs. The transmitted beam collected by the objective is
fed to a fiber-coupled trans-impedance amplifier (Thorlabs) with a response time on the order of
a few picoseconds.
25

Figures 1(e) and (f) show the transmission as a function of time for the unpatterned and
photonic-crystal slabs, respectively. In Figure 3.2(e), the unpatterned slab shows a nearly flat
transmission response as a function of time. However, for the photonic crystal slab, the
transmission exhibits a downward peak (Figure 3.2(f)). From the absorption curve in Figure 3.2(d),
for a laser detuning of 6 nm from resonance (laser wavelength of 976 nm), nearly 10% of the
incident power is absorbed at t=0. Consequently, the slab starts to gradually heat up. Due to the
positive thermo-optic coefficient of silicon, the transmission spectrum incurs a redshift, i.e. shifts
to the right. As the spectrum shifts across the laser wavelength, the transmission decreases with
time until it hits a minimum and then rises again. Figure 3.2(f) shows that the response time,
defined as the time elapsed from turning the laser on to the upward jump in the transmission, can
be controlled by varying the incident power. As the power is increased, the response time is
shortened.
In previous work on similar nanoheaters operating in air, it was shown that the sudden rise in
transmission corresponds to a sudden temperature rise [56]. Here we show that a similar
conclusion applies in water. To model the temporal heating and absorption profiles, we use
COMSOL Multiphysics to solve for the temporal heat distribution using coupled optical, thermal
and fluidic modules, following the methods in Refs. [56] and [67]. A detailed description of the
simulation is given in Appendix I. It incorporates a fluidic module to explicitly account for
convective flow and convective cooling.
We simulated an axisymmetric structure consisting of a silicon disk of thickness 340 nm with
a 5 μm water layer on top. The simulation cell had a radius of 0.5 mm. We took the bottom surface
of the disk to be thermally insulated, and the top boundary of water to be open fluidic and thermal
boundary, to mimic the much taller fluidic channel of 1.6mm thickness used in experiment. The
sidewalls of the disk and water are fixed at room temperature. A heat source with a Gaussian
spatial profile (full width at half maximum of 30 μm) is positioned at the center of the disk. To
26

approximately model the effect of the holes, we used the same modified silicon parameters as in
Ref. [56].

Figure 3.3. (a, b, c) Simulated time-dependent transmission, temperature, and absorption.
First, we consider the blue curves in Figures 3.3 that correspond to the illumination of the
photonic crystal with the fitted spectrum shown in Figure 3.2(b) by a 976 nm wavelength laser.
Figure 3.3(a) shows the time-dependent transmission for 200mW laser power. The shape of the
curve is in good agreement with the measurements in Figure 3.2(f). The corresponding
temperature profile in Figure 3.3b shows that the initial smooth decrease in transmission with time
corresponds to gradual heating of the slab. After a response time of around 320 μs, sudden jumps
27

in both transmission and temperature occur simultaneously. Figure 3.3(c) further shows that light
absorption by the slab reaches its peak at the jump. This indicates that the sudden rise in
transmission and temperature occur when the resonance wavelength shifts through the laser
wavelength.
3.3 Programmable heating
An important feature of our nanoheater is the programmability of the time-dependent
temperature response: both the response time and the temperature after the jump can be
controlled by varying the detuning of the laser from the resonance. Figure 3.4 shows a schematic
of the proposed use of this feature for fine temporal and spatial control of heat on microscale.

Figure 3.4. Schematic of a programmable array of microheaters. Different colors are devices of different
resonance wavelengths and/or detuning between laser and resonance wavelengths.
3.3.1 Laser wavelength tuning

We first carried out simulations to see the effect of varying the detuning on the temporal
response of the nanoheater. The red curves in Figures 3.3(a), (b), and (c) show the effect of
reducing the detuning on the temporal transmission, temperature and absorption. The
transmission dip (Figure 3.3(a)) and temperature jump (Figure 3.3(b)) happen faster for the
28

smaller detuning. This can be attributed to the fact that the absorption is maximum at the
resonance; the closer the laser wavelength is to resonance, the higher the absorption is at t=0,
as shown in Figure 3.3(c).
We verified the effects of detuning experimentally. In Figure 3.5, the effects of 4 nm and 6 nm
detunings between the laser and the resonance wavelength were achieved by using 974 nm and
976 nm lasers. As in the simulation, the response time can be controlled by changing the laser
detuning, with a lower detuning corresponding to a faster response.

Figure 3.5. Experimental transmitted power through the photonic-crystal nanoheater, normalized to the
incident power of 201 mW. Detunings of 4 nm and 6 nm are achieved using laser wavelengths of 974 nm
and 976 nm, respectively.
3.3.2 Design parameters tuning

In the previous section, we controlled the detuning by changing the laser wavelength used for
a fixed device. The temperature response can also be programmed into the device design itself.
We fabricated another photonic crystal slab with a lattice constant of 430 nm, and a hole diameter
of 129 nm. The measured spectra of the new and the old designs are shown in Figures 5(a) and
(b) for two different spectral regions. It can be noticed that device 1 has a strong resonance near
976 nm and a nearly flat response near 805nm and vice versa for device 2. We then characterized
the temporal transmission response for each device under both laser wavelengths of 805.5 nm
and 976 nm. The laser power used is 140 mW and spot size of 10 µm in diameter for both lasers.
29

Figure 3.6. (a, b) The measured spectra of the two fabricated devices around 805nm and 970nm. (Inset of
the proposed used of these structures to form an array). For this specific measurement, the two devices
were fabricated on different samples and illuminated independently, and (c, d) Experimental transmitted
power through both devices, normalized to the incident power of 140 mW using laser wavelengths of 805.5
nm (green) and 976nm (blue).
It can be noticed from Figs. 3.6(c), (d) that each device shows the redshift in the transmission
when illuminated near its resonance and almost flat response otherwise. This demonstrates that
by varying the hole size and/or spacing, different time-dependent heating profiles at different spots
on the sample can be achieved. The temperature profile across the sample could thus be written
in to the nanopatterned design.
3.3.3 Spectral shift and bubble generation
Finally, spectral shift under CW 805.5 nm laser illumination was directly measured using the
spectrometer (Ocean Optics USB 4000). Thorlabs - NF808-34 notch filter was used to attenuate
the laser power and the spectrum was monitored around 970 nm. As shown in Figure 3.7(a) and
(b) the device spectrum incurs an increasing redshift with power. When the power is increased
30

beyond a certain threshold (in this case between 120 mW and 145 mW), the water reaches the
boiling point and stable microbubbles are formed as shown in the inset of Figure 6(a). Due to light
scattering by the bubble, the transmission spectrum magnitude goes down when the power is
increased beyond 145 mW.

Figure 3.7. (a) The measured spectra (monitored around 970 nm) of device 2 under CW illumination by the
805.5 nm laser at various powers. Inset is a microscopic image of the bubble (black circle to the left) and
an unexcited device of 100 µm diameter (dark circle to the right), and (b) the spectral shift as a function of
the laser power before bubble formation.
Photothermal microbubble generation has several applications in photoacoustic imaging [75-
77], microparticle trapping [78-81] and lab-on-chip devices [82]. The programmability of the
proposed microheaters offers the advantage of controlling the bubble’s power threshold by
changing the laser detuning and/or device parameters. Namely, devices 1 and 2 shown in figure
5 have 93 mW and 140 mW bubble thresholds, under 976 nm and 805.5 nm laser excitation,
respectively with all other conditions kept the same. Both the resonance quality factor and
detuning of the laser contribute to this shift in the bubble threshold.
In previous work, the design of absorptive resonances in PC slabs has also been used to control
heat on micro-scale [57]. The proposed photonic crystal microheaters exhibit unique temperature
behavior as a function of time. This is a consequence of using a relatively narrow absorptive
resonance for heating [56]. The elapsed time between laser turn-on and jump is called the
response time.
31

3.3.4 Appendix I
We use Comsol Multiphysics to solve for the temporal heat distribution using coupled optical,
thermal and fluidic modules, following the methods in Refs. [56, 67].
The transmission spectrum is given by coupled mode theory (CMT) [83] as
𝑇 𝑡𝑟𝑎𝑛𝑠 ( 𝑡 ) =
( 𝑡 𝑠 𝛾 𝑖 )
2
+ (𝑡 𝑠 (𝜔 𝑜𝑝
− 𝜔 0
′
( 𝑡 ) ) + 𝑟 𝑠 𝛾 𝑟 )
2
(𝜔 𝑜𝑝
− 𝜔 0
′
( 𝑡 ) )
2
+ ( 𝛾 𝑟 + 𝛾 𝑖 )
2
, Eq. 3.1
where 𝜔 𝑜𝑝
is the operating frequency of the laser, 𝛾 𝑟 is the decay rate of the resonance due to
radiation loss, 𝛾 𝑖 is the decay rate due to material absorption loss, and 𝑡 𝑠 and 𝑟 𝑠 are the direct
transmission and reflection coefficients, respectively. The frequency of the guided resonance
mode, 𝜔 0
′
( 𝑡 ) depends on the time-dependent temperature of the slab as
𝜔 0
′
( 𝑡 ) = 𝜔 0
−
𝜔 0
𝑛 0
𝑑𝑛
𝑑𝑇
( 𝑇 ( 𝜌 , 𝑧 , 𝑡 )− 𝑇 0
) , Eq. 3.2
where 𝜔 0
is the initial frequency of the guided resonance mode (e.g., the frequency at room
temperature, in the absence of laser heating), 𝑑𝑛 /𝑑𝑇 is the thermo-optic coefficient of silicon, and
𝑛 0
is the refractive index at room temperature (𝑇 0
). Since silicon has a positive thermo-optic
coefficient, the frequency in equation (2) goes down as the temperature increases in time,
redshifting the spectrum.
The temporal temperature distribution 𝑇 ( 𝑡 ) inside the photonic crystal slab is governed by the
3D heat diffusion equation:
𝜌 𝑠 𝐶 𝑠 𝜕𝑇 ( 𝜌 , 𝑧 , 𝑡 )
𝜕𝑡
= ∇( 𝑘 ∇𝑇 ( 𝜌 , 𝑧 , 𝑡 ) )+ 𝑃 𝑎𝑏𝑠 ( 𝑡 ) ,
Eq. 3.3
32

where 𝑘 is the thermal conductivity of silicon, 𝜌 𝑠 its density, and 𝐶 𝑠 its specific heat capacity at
constant pressure. The heat source is determined by the optical power absorbed at resonance,
which from coupled mode theory is [53]
𝑃 𝑎𝑏𝑠 ( 𝑡 ) = 𝑃 𝑖𝑛
2𝛾 𝑖 𝛾 𝑟 (𝜔 𝑜𝑝
− 𝜔 0
′
( 𝑡 ) )
2
+ ( 𝛾 𝑟 + 𝛾 𝑖 )
2
,
Eq. 3.4
We neglect the absorption in water, since the absorption coefficient is 3 orders of magnitude
lower than that of silicon at 976 nm; thus 𝑞 ( 𝑟 ) = 0. Hence, the heat distribution in the surrounding
water is governed by the following equation [84]:
𝜌 𝑤 𝐶 𝑤 [
𝜕𝑇 ( 𝜌 , 𝑧 , 𝑡 )
𝜕𝑡
+ ∇. ( 𝑇 ( 𝜌 , 𝑧 , 𝑡 ) v( 𝜌 , 𝑧 , 𝑡 ) ) ] − 𝑘 ∇
2
𝑇 ( 𝜌 , 𝑧 , 𝑡 ) = 0, Eq. 3.5
where v( 𝜌 , 𝑧 , 𝑡 ) is the fluid velocity, 𝑘 is the thermal conductivity of water, 𝜌 𝑤 its mass density, and
𝐶 𝑤 its specific heat capacity at constant pressure. Due to the temperature increase that takes
place within the fluid around the structure, the fluid experiences some reduction of its mass
density, which yields upward convection. The general equation governing the fluid velocity profile
is the Navier Stokes equation [84]:
𝜕 v
𝜕𝑡
+ ( v. ∇) v = 𝜈 ∇
2
v + 𝑓 𝑡 ℎ
( 𝑇 ) , Eq. 3.6
where 𝜈 is the viscosity of water and 𝑓 𝑡 ℎ
( 𝑇 ) is the force per unit mass due to temperature non-
uniformity. This thermal force can be estimated using the Boussinesq approximation. This
approximation accounts for the temperature dependence of the mass density by adding an
external buoyancy force term that is dependent on the temperature distribution;
𝑓 𝑡 ℎ
( 𝑇 ) = 𝛽𝑔𝛿𝑇 𝑢 𝑧 , Eq. 3.7
where g is the gravitational acceleration, β the dilatation coefficient of water, 𝛿𝑇 = 𝑇 − 𝑇 0
is the
temperature increase, and 𝑢 𝑧 is the upward unit vector in the 𝑧 direction.
33

At each time step, we use a COMSOL model including coupled heat transfer in solids/ fluids
and laminar flow to solve equations 2 through 5 self-consistently for 𝑇 ( 𝑡 ) , 𝜔 0
′
( 𝑡 ) , 𝑣 ( 𝑡 ) , and 𝑃 𝑎𝑏𝑠 ( 𝑡 ) ,
given the parameters 𝜔 0
, 𝛾 𝑖 , and 𝛾 𝑟 , using methods similar to Refs. [67] and [56]. Note that
throughout the COMSOL simulation, we assume implicitly that 𝑇 , 𝜔 0
′
, v and 𝑃 𝑎𝑏𝑠 are functions of
position ( 𝜌 , 𝑧 ) ; to obtain 𝑇 𝑡𝑟𝑎𝑛𝑠 ( 𝑡 ) , we use the values at the center of the nanoheater ( 𝜌 = 𝑧 = 0) .

34

Chapter 4
4. Future Work
For future work, we propose a modified design that utilizes enhanced absorption in PC slabs
to make high temperature thermal logic gates. Here, we summarize a design of a thermal inverter
and an extension to a universal “NAND” and “NOR” gates. Moreover, we propose and show
preliminary experimental results of a biological application to localized micro-heating in selective
cell trapping.
4.1 Thermal Logic Gates
Figures 4.1(a) and 4.1(b) show schematics of the proposed thermal inverter design. A basic
component is a temperature-controlled transmissivity switch with high transmission at low
temperatures. This switch uses a vanadium dioxide phase change material on a relatively
transparent substrate (LSAT). The switch is then placed between a constant optical power input
and a PC slab with an absorptive GRM as shown in Figs 4.1(a) and 4.1(b). In the insulating state
of VO2, the transmitted power (P1) is designed to switch the PC to the hot state (TH). As the input
temperature to the VO 2 film increases, it switches to a more lossy metallic state with lower
transmitted power (P2). This power is designed to keep the PC slab in the cold state (TL) as shown
in Fig. 4.1(c). The resulting truth table of this design is shown in Fig. 4.1(d) with the output
temperature state being an inverted version of the input.
35

Figure 4.1. (a) and (b) Schematics of the thermal inverter gate for insulating and metallic VO2 states,
respectively. (c) Schematic of the thermal response of the PC slab to varying input power. (d) Truth table
showing the thermal inversion.

4.1.1 Numerical Modeling
To numerically model our device, we use a semi-empirical model. We first characterize the
transmissivity of a bare LSAT double sided polished substrate. Figure 4.2(a) shows a measured
constant transmissivity of 80% across the wavelength range of 500-1000nm. We then deposit a
70nm VO2 film using pulsed laser deposition (PLD) and measure transmissivity switching of the
VO2/ LSAT stack shown in Figure 4.2(b). For this test sample, switching on the order of 10-20%
is measured with the peak around the wavelength of 940nm.

36

Figure 4.2. Measured transmission spectrum of (a) a 500um LSAT substrate, and (b) VO2/ LSAT stack.
To further enhance the transmission switching, an optimized design of the sample will control the
VO2 thin film thickness, substrate material and micro-patterning.
We use FDTD to design a PC slab with a near critically coupled GRM at 800nm. We then use
coupled mode theory to describe the resonance and input the fitted parameters to a thermal
COMSOL model shown schematically in Fig. 4.3(a). Figure 4.3(b) shows the simulated hysteresis
curve of the PC slab in response to varying optical power input. Figures 4.3(c-d) show the
simulated response of the design in Figs. 4.3(a, b). For a low heat input, VO 2 transmissivity of
27% corresponds to an incident power of 54mW on the PC, which heats the slab to 438 K. As the
input temperature increases above the critical temperature of VO 2 (350 K), the transmissivity
switches down to 14%, corresponding to a transmitted power of 28mW, which keeps the output
PC slab temperature at 311 K. Figures 4.3(c, d) show the temperature inverting scheme.

Figure 4.3. (a) Schematic of the structure simulated in COMSOL. (b) Calculated thermal hysteresis curve
for the PC slab. (c-e) Simulated response of the thermal inverter design in Fig. 4(a, b); such that (c) is input
VO2 temperature, (d) is the VO2/ LSAT transmissivity response and (e) is the output PC temperature.

37

4.1.2 Experimental demonstration
For a preliminary experimental demonstration, we use ebeam lithography to fabricate a PC
slab with a near critically coupled GRM at 800nm (~50% absorptive at resonance). The measured
power hysteresis curve is shown in Fig. 4.4(a). This curve shows that a power switching of at least
45-27mW is needed to switch the PC slab temperature between its two stable states. This working
constraint is shown in Fig. 4.4(b). Samples 1 and 2 with VO2 thicknesses of 100 and 70nm,
respectively are projected on Fig. 4.4(b), showing that sample 2 is able to satisfy this working
area constraint.

Figure 4.4. (a) Measured optical bistability curve in the PC slab. (b) Projection of the measured VO2/ LSAT
samples in terms of transmission change between insulating to metallic VO2 states.
The same design can be extended to a universal “NAND” or “NOR” gate as shown in Figure
4.5. The design includes two constant laser powers (P0) that go through two independent
temperature-controlled transmissivity switches and collected by a beam splitter to heat a PC slab
as shown in Fig. 4.5(a). Figures 4.5(b) and 4.5(c) show that by controlling the power value (P),
the same design can be used to make either a “NAND” (Fig. 4.5(a)) or a “NOR” gate (Fig. 4.5(c)).
We will carry numerical simulations to optimize the system utilizing a narrower hysteresis
width of the PC slab and enhanced VO2 transmissivity switching. Further device optimization will
38

also consider maximizing speed and minimizing operating power. The speed of the device is
limited by the phase transition switching in VO 2 and the heating/ cooling time constant in the PC
slab. Previous work has shown that these time constants can be designed to work with switching
times on the order of picoseconds [85]. Key elements for speed are the size of the laser spot, the
design of the switching temperatures and the heating mechanism in VO2 [86, 87]. We will then
rum a proof of concept experiment, similar to Ref. [19], measuring time-dependent output
temperature as a function of input temperature, characterizing the actual speed of the device.

Figure 4.5. (a) Schematic of proposed universal “NAND” or “NOR” gate. (b) and (c) Schematics of power
choice and truth table corresponding to each gate.

4.2 Biological Applications
Since our designed absorptive GRM also provide significant localized heating on the
microscale, it can be used for biological applications in a microfluidic environment. Figure 9 shows
a schematic of a proposed application using a checkerboard template of alternating patterned
heaters operating at different wavelengths and heated to different temperatures (T1 and T2). ELP’s
are heat sensitive proteins that have been shown previously to change phase above a designed
transition temperature [88]. The phase transition occurs when the nano-scaled particle proteins
39

aggregate to form larger micro-particles upon heating. These proteins can be attached to different
types of living cells and hence selectively delivering them to the heated spots on chip as shown
in Fig. 4.6. Figures 4.6(b) and (c) show that with a checkerboard template selective cell trapping
at specific location can be controlled using laser illumination.

Figure 4.6. (a) Schematic of unilluminated template of alternating microheater designs with floating heat-
sensitive proteins (ELP’s), and live cells. (b) Schematic of one heater type selectively activated by laser
illumination and attracting the corresponding cells/ proteins. (c) Schematic of both heaters activated by
laser illumination and attracting the corresponding cells/ proteins.

4.2.1 Microfluidic modelling
We use a similar model to section 4.2.1 with the addition of a water column and considering
coupled heat transfer and laminar flow physics in COMSOL. Figure 4.7 shows the simulation of a
horizontally (a) and vertically (b) positioned chambers. With the local heating of 460K a maximum
symmetric convection flow of 18um/s is obtained in the horizontal case compared to a maximum
of 250μm/s upwards flow in the vertical case.
40

Figure 4.7. COMSOL simulations of thermal and fluidic response to laser illumination of a Si PC slab
microheater positioned (a) horizontally, and (b) vertically.
This heat stimulated flow is necessary to bring the floating particles in the chamber to the
locally heated spot using convection drag forces [80]. When the particles get near the heated
spot, thermophoresis and optical forces become significant as shown schematically in Figure 4.8
for the vertically placed chamber. From the force balance, particles near the surface of the PC
heater with form a semi-ring around the bottom half of the heated spot. This can be seen from the
balance of the thermophoresis and convection forces in the bottom half compared to them acting
in the same direction (pushing particles away) in the upper half. The radius of the ring can be
calculated from the force balance which will be a function of the maximum temperature and the
fluid properties.

Figure 4.8. Schematic of various forces on a particle placed in the locally heated microfluidic chamber.
41

4.2.2 Experimental demonstration
For preliminary experimental demonstration, we fabricate a microheater with an absorptive
resonance near 800 nm. The thermo-optic effect in silicon cause the spectrum to red-shift in
response to heating. This shift can also be used as a probe to optically measure the temperature
of the locally heated spot. For temperature measurements, we first make a calibration experiment,
in which the entire sample is heated to a known temperature and the spectral shift of the device
is monitored as shown in Fig. 4.9(a). We compare the measured points to the simulations using
thermo-optic coefficient in Si, which show a good match. We then use the calibration curve to
measure the temperature of the locally (laser) heated devices. Figure 4.9(b) shows the
transmission spectra in response to increasing laser power with temperature mapping for each
power using the calibration experiment.

Figure 4.9. (a) Calibration of spectral shift as a function of device’s temperature. (b) Response of a
fabricated device to various input laser power mapped to the calibration in (a).
Figure 4.10, shows an actual preliminary experiment of the local heating of the heat sensitive
proteins in a vertically positioned chamber. The patterned PC crystal region is 100um and the
laser spot size is around 50um. The time evolution of the proteins phase transition shows that the
proteins get collimated selectively around the heated spot forming a semi-ring at the bottom side
as predicted by the model. While this type of proteins is reversible upon cooling (laser turned off),
other types can be engineered for permanent trapping. Using this preliminary experiment, we will
42

carry out a comprehensive study of a solution with a mixture of proteins and living cells. The
convection flow helps bringing the proteins to the heated site, and a careful design of force
balance can be used to selectively trap cells. The transition temperature of the proteins can be
controlled to either keep the cells alive or killed after trapping.

Figure 4.10. Microscopic snapshots of proteins local phase transition evolution with time.
43

Part II: Infrared Thermal Emission
Control

44

Chapter 5.
5. Introduction: Part II
Emissivity control in the infrared wavelength range has been subject to extensive research in
the recent years [89]. Several designs have shown that using nanophotonics structures, custom
broadband and narrowband IR absorptivity/ emissivity spectra can be achieved [90, 91].
Advances to metamaterial design have led to new venues for emissivity spectral shaping. Since
its first experimental demonstration by Landy et al. [92], the metamaterial absorber/ emitter has
undergone a rapid and prosperous development, finding a wide variety of practical applications
including selective thermal emitters [93], wavelength-tunable microbolometers [94] and refractive
index sensing [95]. The typical structure of the metamaterial absorber is a three-layered
architecture, consisting of arrays of patterned metallic sub-wavelength structure on top of a
dielectric spacer, backed by a thick metallic ground plane. The top periodic structure is
responsible for the electric response of the metamaterial absorber. The metallic ground plane is
thicker than the penetration depth of the incident wave to eliminate any transmission. The coupling
of two metallic layers and the dielectric spacer determines its magnetic response [96]. Therefore,
by altering the geometry of the ERR and changing the thickness of the dielectric layer, the
effective permittivity ε and permeability μ can be tuned independently, resulting in an impedance
match to the free space, and thus perfect absorption of the incident wave at certain frequencies
[97].
By manipulating the electric and magnetic properties, the operation of metamaterial absorbers
has been verified from microwave range [92], through terahertz [98], infrared [96] and into the
visible [99]. The previous work on metamaterial absorbers focused on the design of bright mode
resonances. Here we study the dark absorptive modes in these structures and propose a scheme
45

for narrowband emissivity tuning using micro resonators coupling. Single wavelength as well as
integrated emissivity tuning can be achieved in this scheme by tuning relative resonance
wavelengths and/ or coupling coefficient between the resonators.

46

Chapter 6
6. Coupled resonators for switchable emissivity

6.1 Coupled resonators for switchable emissivity
In this chapter we study a system of coupled optical resonators in the mid infrared range. We
first use temporal coupled mode theory (CMT) as a computationally efficient way of design to
calculate emission spectral profile. We show that by tuning coupling coefficient as well as relative
resonance wavelength we can tune spectral profile. We then use a metamaterial based thermal
emitter design to verify the developed theory using FDTD simulations. The good match between
the developed theory and the numerical simulations of a sample deign suggests potential use of
the theory to explore the emissivity switching limits using this method.
Figure 6.1 shows a schematic of three coupled resonators with a coupling coefficient β. The
two side resonators act as thermal emitters with thermal noise sources n1 and n2. They are also
designed to operate in the dark mode, which means they can only couple to the middle resonator
but not to an out of plane port. This dark condition is dictated by the mode field symmetry as will
be illustrated later in this chapter. The middle resonator in this scheme is the only one that couples
to a plane wave port or operates in a bright mode. We start by writing the temporal coupled mode
theory equations describing the field intensities as follows

Figure 6.1. Schematic of a three coupled resonating structure
47

𝑑 𝑎 1
𝑑𝑡
= 𝑗 𝜔 1
𝑎 1
+ 𝑗𝛽 𝑎 0
−
1
𝜏 0
𝑎 1
+ √
2
𝜏 0
𝑛 1
,
Eq. 6.1

𝑑 𝑎 0
𝑑𝑡
= 𝑗 𝜔 0
𝑎 0
+ 𝑗𝛽 𝑎 1
+ 𝑗𝛽 𝑎 1
+ √
2
𝜏 𝑒 𝑆 0
+
−
1
𝜏 𝑒 𝑎 0
,
Eq. 6.2

𝑑 𝑎 2
𝑑𝑡
= 𝑗 𝜔 2
𝑎 2
+ 𝑗𝛽 𝑎 0
−
1
𝜏 0
𝑎 2
+ √
2
𝜏 0
𝑛 2
,
Eq. 6.3

𝑆 0
−
= −𝑆 0
+
+ √
2
𝜏 𝑒 𝑎 0
,
Eq. 6.4
where, a0, a1 and a2 are the field amplitudes of the three resonators, β is the coupling coefficient,
𝜏 0
is the intrinsic and 𝜏 𝑒 is the extrinsic decay rates. By solving Eqs. [6.1-6.4], the power spectral
density at the output port 𝑆 0
−
can be written as
𝑃 ( 𝜔 )
=
4Θ( 𝜔 , 𝑇 )
2𝜋 𝜏 0
𝜏 𝑒 𝛽 2
[( 𝜔 − 𝜔 1
)
2
+ (
1
𝜏 0
)
2
+ ( 𝜔 − 𝜔 2
)
2
+ (
1
𝜏 0
)
2
]
|(𝑗 ( 𝜔 − 𝜔 1
)+ (
1
𝜏 0
)) (𝑗 ( 𝜔 − 𝜔 0
)+ (
1
𝜏 𝑒 )) (𝑗 ( 𝜔 − 𝜔 2
)+ (
1
𝜏 0
)) + 𝛽 2
(𝑗 ( 2𝜔 − 𝜔 1
− 𝜔 2
)+
2
𝜏 0
)|
2

Eq. 6.5
where the noise source correlation is [100]
⟨n
∗
( 𝜔 ) 𝑛 ( 𝜔 ′
) ⟩ =
Θ( 𝜔 , 𝑇 )
2𝜋 𝛿 ( 𝜔 − 𝜔 ′
) ,
Eq. 6.6
and,

48

Θ( 𝜔 , 𝑇 ) =
ℏ𝜔 𝑒 ℏ𝜔 𝑘𝑇
− 1
,
Eq. 6.7
Equation (6.5) suggests that the emission spectrum splits into three resonance peaks
depending in location and magnitude on the detuning of the resonance wavelengths of the side
resonators ( Δλ) relative to the middle one as well as the coupling coefficient. Using the derived
form in Eq. (6.5), Fig. 6.2 (a) shows that by tuning the relative resonance wavelength of the micro
resonators, the emission power spectrum can significantly be tuned. A similar effect can be
achieved by tuning the coupling coefficient between the resonators as shown in Fig. 6.2(b).

Figure 6.2. Thermal emissivity switching by tuning resonance wavelength (a) and coupling coefficient (b).
We calculate the integrated emissivity for a simplified model of only two resonators as shown
in Fig. 6.3. The upper panel of Fig. 6.3 shows that normalized integrated emissivity can be
switched between 1.33 and 1 for the case when the two resonators are thermally emitting. If only
the dark mode resonator is a thermal emitter, the integrated emissivity can be completely switched
between 1 and 0 as shown in the bottom panel of Fig. 6.3.
49

Figure 6.3. Normalized integrated emissivity switching predicted by coupled mode theory for a two coupled
resonator design. Upper panel is for the case when the two resonators are thermally emitting, and the lower
panel is for the case when only one resonator is thermally emitting.
6.2 Dark Modes in Metamaterial Absorbers
Metamaterials, through changes of the size and shape of subwavelength metallic elements,
permit the amplitude and frequency tuning of both the electric and magnetic response of
electromagnetic radiation. Although this electromagnetic response is resonant and narrow band,
there have been demonstrations of exotic metamaterials operating in all relevant bands for
frequencies below visible [101-106]. The resonant nature of metamaterials results in a strong
focusing of the electric field within gaps of the structure, thus providing a means to dynamically
control the resonance frequency, phase, and amplitude [107-109].
Figure 6.4 shows an example design of a metamaterial thermal emitter that can be designed
to operate in the IR range. It consists of a 100 nm thick gold cross on top of Al2O3 layer and a gold
back reflector (inset of Fig. 6.4(a)). To simulate the dark modes in this structure, we use a dipole
cloud excitation. A cloud of randomly positioned dipoles, with random phases is positioned within
the micro resonator. Another randomly positioned cloud of time domain monitors is placed to
measure the field amplitude. We then perform a fast Fourier transform (FFT) on the monitors
readings to obtain the spectrum. Figure 6.4(a) shows the simulated spectra resulting from normal
50

incidence simulations (red curve), and a dipole cloud simulation (blue curve). The normal
incidence simulation shows a single resonance peak around 6.2 μm, which corresponds to the
bright mode in the wavelength range plotted. The dipole cloud simulations show one more mode
at a lower wavelength that doesn’t couple to a plane wave (dark mode). The insets of Fig.6.4(a)
show the electric field profiles of both modes. The symmetries in these profiles confirm that only
the higher wavelength mode can couple to an out of plane wave, while the lower peak is a dark
mode. Both the dark and the bright modes can be readily tuned by tuning the cross length (L), as
shown in Fig. 6.4(b).

Figure 6.4. FDTD simulations of the cross design in (a) using a dipole cloud and a normal incidence plane
wave. Insets are a schematic of the design, and the field profiles for the two resonance modes. (b) Dark
and bright resonances wavelengths as a function of the cross length L.
6.3 Dark – Bright Coupled Resonators
To achieve thermal emission switching studied in section 6.1, we consider a reduced model of
two absorptive resonators operating at closely spaced frequencies 𝜔 0
and 𝜔 1
featuring a bright
and a dark mode, respectively. Using CMT as in section 6.1, a closed form power spectral density
for the thermal emission can be written as

51

𝑃 ( 𝜔 ) |
𝑆 0
− =
4Θ( 𝜔 , 𝑇 )
2𝜋 𝜏 0
𝜏 𝑒 [𝛽 2
+ ( 𝜔 − 𝜔 1
)
2
+ (
1
𝜏 0
)
2
]
|(𝑗 ( 𝜔 − 𝜔 1
)+ (
1
𝜏 0
)) (𝑗 ( 𝜔 − 𝜔 0
)+ (
1
𝜏 𝑒 +
1
𝜏 0
)) + 𝛽 2
|
2
,
Eq. 6.8
Figure 6.5 shows FDTD simulation of a dark-bright structure in the metamaterial design shown
in the index. For this design we use cross lengths (L) of 1.7 and 1.27µm which result in a dark
and a bright mode at the wavelength of ~4.7μm, respectively. The emissivity spectrum show that
the interaction of the two modes results in two peaks to either side of the original resonance
location. The lower wavelength mode is a result of the constructive interference between the two
modes and the higher wavelength peak corresponds to a bright mode field in one of the resonators
shifted by the dark resonator. The asymmetry in the peaks results from the fact that the two modes
are not identical in line shape and each resonator’s individual resonance wavelength slightly shifts
because of the larger period in the coupled structure.

Figure 6.5. FDTD simulation of Dark-Bright mode absorptivity/ emissivity spectrum. Insets are a schematic
of the simulated structure and the magnitude field profile for the two resonance peaks.
In Fig. 6.6 we compare the developed CMT to the FDTD simulations of the example design.
In simulations, we use two resonators with a dark and a bright mode and tune the coupling by
tuning the distance between them, shown in Fig. 6.6(a). For coupled mode theory, we use the
derived Eq. (6.5) and tune β in multiples of the intrinsic decay rate. At very large distance, the
resulting spectrum corresponds to zero coupling in the CMT model. As the coupling increases
52

(distance decreases in FDTD), the two resonators start to couple splitting the peak into two
smaller peaks. The distance between the split resonance peaks increases as the coupling
increase resulting in dynamic tunability of the spectrum. An alternative way of changing the
coupling, is by tuning the refractive index of the dielectric layer.

Figure 6.6. (a) FDTD normal incidence simulation of a two coupled resonator design with a varying distance
d. (b) Corresponding derived coupled mode theory model for a two coupled resonator design with a varying
coupling coefficient β.
6.4 Dark- Dark Coupled Resonators
The interaction of the bright-dark modes in the previous section has shown the capability of
tuning the resultant spectrum by tuning the coupling coefficient between the two resonators. In
this section, we consider coupling Dark-Dark modes by bringing two identical crossed at a small
distance d. In this scheme, in the limit of no coupling, the two resonators are dark and thus result
in no emission. When the coupling is high enough, we will show that it results in a narrow bright
emissivity peak with an amplitude of 1.
53

Figure 6.7 shows that coupled at a distance of 0.05 µm, both the bright and the dark modes split
into higher quality factor dark and bright peaks. The dark-dark (bright) peak has a linewidth of 112
nm at the wavelength of ~5.0 µm.

Figure 6.7. (a) FDTD normal incidence simulation of a bright mode in a cross metamaterial resonator (inset
shows the schematic and field profiles). Dotted line shows the location of the dark mode not showing in the
normal incidence spectrum. (b) Normal incidence simulation of two identical resonators coupled at 0.05μm
(inset shows the schematic and field profiles).
In the limit of the weak coupling between the two resonators this dark-dark (bright) peak
disappears as it turns into a dark mode. This effect is simulated by tuning the distance between
the two resonators as shown in Fig. 6.8 (a). The corresponding quality factor and peak absorptivity
are shown on Fig. 6.8 (b). The peak value can be tuned between 1.0 and <0.001 by tuning the
distance between the two resonators from 0.05 to 0.50μm. This novel approach shows a novel
use of coupled dark modes to generate a high quality factor completely switchable emissivity
peak.
54

Figure 6.8. (a) FDTD normal incidence simulation of the dark-dark mode for varying distance (d) (inset
shows the schematic and field profiles). (b) Corresponding peak absorptivity and quality factor to varying
coupling (distance (d)).
Another approach for switching the dark-dark mode is by tuning the refractive index of the
dielectric layer (Al2O3). To maintain the status of the dark- dark (dark) mode, we use the symmetry
shown in the inset of Fig. 6.9 (a). The result of the refractive index tuning is shown in Fig. 6.9 (a).
Zooming on the narrowband effect, Fig. 6.9 (b) shows that the peak absorptivity at ~5.0µm can
be switched down by at least five orders of magnitude with a tuning factor η=1.5.

Figure 6.9. (a) FDTD normal incidence simulation of the dark-dark and bright-bright modes for varying
dielectric refractive index (inset shows the schematic). (b) Corresponding dark-dark peak switching.
55

In conclusion, we presented the use of coupled optical resonators to tune infrared thermal
emissivity spectrum. We used coupled mode theory to predict the spectrum tunability and used
FDTD to simulate a sample design showing the effect predicted by the theory. The concept of
dark- bright and dark-dark mode coupled resonators provides an additional degree of freedom as
a useful tool for infrared spectral shaping.
56

Chapter 7
7. Broadband Emissivity Switching
7.1 Background and motivation
We present an experimental demonstration of passive, dynamic thermal regulation in a solid-
state system with temperature-dependent thermal emissivity switching. We achieve this effect
using a multilayered device, comprised of a vanadium dioxide (VO 2) thin film on a silicon substrate
with a gold back reflector. We experimentally characterize the optical properties of the VO 2 film
and use the results to optimize device design. Using a calibrated, transient calorimetry experiment
we directly measure the temperature fluctuations arising from a time-varying heat load. Under
laboratory conditions, we find that the device regulates temperature better than a constant
emissivity sample. We use the experimental results to validate our thermal model, which can be
used to predict device performance under the conditions of outer space. In this limit, thermal
fluctuations are halved with reference to a constant-emissivity sample.
The use of material design techniques to control the thermal emissive properties of matter has
emerged as topic of great interest in current research of intelligent, radiative thermal control. A
variety of microstructures have been used for this purpose, including multilayer films [110, 111],
microparticles [112], photonic crystals [113], and metamaterials [114, 115]. One particularly
interesting application of emissive control is the design of materials that self-regulate their
temperature [116, 117], a property we term thermal homeostasis [118, 119]. Such a capability is
likely to be useful for a variety of applications including satellite thermal control, for which
traditional solutions require either electrical power or moving parts [120-123].
57

The key physical principle required for passive thermal regulation is strong temperature-
dependent integrated emissivity. The phase change material vanadium dioxide (VO2), in
particular, exhibits a dramatic change to its optical properties across a thermally narrow phase
transition [124, 125]. With proper design, VO 2-based microstructures can achieve a sharp
increase in thermal emissivity across the phase transition temperature near 68
o
C [125, 126].
Intuitively, when the material temperature is below the transition temperature, the emissivity is
low, and the object retains heat. When the material temperature exceeds the transition
temperature, emissivity increases, and the object loses heat. This negative feedback regulates
the material near the temperature of the phase transition [118]. Recent works have demonstrated
experimentally broadband emissivity switching for both planar [127] and meta-reflector designs
[128]. However, no direct measurement of thermal regulation has been performed. In this paper,
we present an experimental method for studying dynamic thermal regulation due to infrared
emissive switching. We therefore demonstrate direct evidence of reduction in thermal fluctuations
due to emissive switching at the VO 2 phase transition.
7.2 Characterization of infrared optical properties of vanadium dioxide
Vanadium dioxide has a phase transition at a critical temperature (Tc) of approximately 68 °C
[129]. The infrared optical properties of VO 2 switch between a low- loss, semi-transparent material
(referred to in this article as the insulating state), and a lossy, more reflective material (referred to
in this article as the metallic state). Various works in the literature have measured the optical
constants of VO2 in the visible and near IR [130-133]. More recent work measured the infrared
optical constants of VO2 thin films grown using pulsed layer deposition (PLD) ) [127, 134],
sputtering, and sol-gel [135]. It was found that the growth technique influences the optical
properties due to the quality of the thin crystalline films [135].
We used atomic layer deposition (ALD) to deposit a VO 2 thin film on a Si substrate. Compared
to traditional growth methods, ALD allows deposition of highly conformal VO 2 films over large
58

areas [136]. The optical constants at temperatures above and below the VO 2 phase transition
were measured using spectroscopic ellipsometry. The deposition process and measurement
method are described in detail in the Methods section, which also lists the ellipsometric fitting
parameters. Figure 7.1 shows the real (n, solid lines) and imaginary (k, dashed lines) parts of the
complex refractive index of the insulating (blue line) and metallic (red line) VO 2 states. We observe
that both n and k change significantly between the two states. The higher value of k in the metallic
state indicates an increase in loss over the entire 2 to 30 μm range.

Figure 7.1. Complex dielectric function of metallic (red) and insulating (blue) states of a 98nm-thick
vanadium oxide film as determined from spectroscopic ellipsometry measurements. Solid lines give the real
part (n) and dashed lines give the imaginary part (k) of the VO2 refractive index.
7.3 Design optimization
Previous work by our group [118] presented a silicon micro-cone design that can be optimized
to achieve >81% switching in radiated power between hot and cold states. To fabricate this
structure, we use the method shown in Fig. 7.2. We start with a 300 µm-thick silicon substrate,
coated with a 1μm hard etching mask of SiO 2. We then use optical lithography to define a disc
array pattern. Isotropic dry etching using standard Bosch process is then performed to etch the
hard mask, followed by anisotropic SF6 and O2 to etch the curved sides. The SEM image of the
final parabolic cone structure is shown in Fig. 7.2.
59

Figure 7.2. Schematic and SEM picture of the micro-cones fabrication process.
Using the measured optical properties of our ALD-grown VO2 shown in Fig. 7.1, we
recalculated the switching in radiation power for three structures: perfect micro-cones, parabolic
micro-cones, and micro-rods. Thermal emissivity spectrum 𝘀 ( 𝜆 , 𝑇 ) is calculated using the ISU-
TMM package [137, 138], an implementation of the plane-wave-based transfer matrix method.
The simulation calculates absorptivity at normal incidence, where absorptivity is equal to
emissivity by Kirchoff’s law. The wavelength range shown is chosen to be 2.5–30 μm; outside this
range, the blackbody radiance at room temperature is negligible. The normalized thermal
radiation power was calculated as
𝑃 𝑟𝑎𝑑 ( 𝑇 ) =
∫ 𝑑𝜆 ∙ 𝐼 𝐵𝐵
( 𝜆 , 𝑇 )∙ 𝘀 ( 𝜆 , 𝑇 )
30 𝜇𝑚
2.5 𝜇𝑚
∫ 𝑑𝜆 ∙ 𝐼 𝐵𝐵
( 𝜆 , 𝑇 )
30 𝜇𝑚
2.5 𝜇𝑚
, Eq. 7.1
where 𝐼 𝐵𝐵
( 𝜆 , 𝑇 ) is the blackbody radiance, and 𝘀 ( 𝜆 , 𝑇 ) is the emissivity spectrum.
The figure of merit is defined as the difference between radiated power in the hot state and the
cold state 𝛥 𝑃 𝑟𝑎𝑑 . Figure 7.3 shows that perfect cones and parabolic cones give figure of merit
values of 0.5542 and 0.5059, respectively while micro-rods have a values of 0.4901.
60

Figure 7.3. Calculated 𝛥 𝑃 𝑟𝑎𝑑 as a function of height and VO2 thickness for (a) Silicon micro-cones, (b)
parabolic micro-cones, and (c) micro-rods with a fixed period of 1µm.
Figure 7.4 shows a comparison of fabrication feasibility versus performance. The two leftmost
columns show that high aspect ratio (h/a~20) is needed to achieve the best value of 𝛥 𝑃 𝑟𝑎𝑑 in
either perfect or parabolic micro-cones. Micro-pillars, however, can give a value of ~0.40 with an
aspect ratio value of ~1.0. The parabolic micro-cones structure achieved with our recipe (h/a=0.5)
gives a value of ~0.33.

Figure 7.4. A comparison of different structures feasibility and switching performances.
We next compare this to the performance of planar VO2 film. Figure 7.5 shows that for an
isolated VO2 thin film, 𝛥 𝑃 𝑟𝑎𝑑 is positive for thicknesses less than ~6μm. For experimental
61

convenience, we add a silicon handle layer with a thickness of 200μm (red line in Fig. 2). 𝛥 𝑃 𝑟𝑎𝑑
is again positive for thickness below ~3μm, with a peak value of 0.22 at a thickness of 800nm.
Adding a gold back reflector to the VO2/Si stack enhances the peak value to 0.3 at a smaller VO2
thickness of 75nm. Smaller thicknesses are highly desirable for ALD fabrication.

Figure 7.5. Calculated difference in integrated thermal radiation power for an isolated VO2 thin film, VO2
thin film with a handle layer of Si, and VO2/Si/Au multilayer stack. Corresponding schematics and layer
thicknesses are shown in the inset.
The use of a gold back reflector also prevents any background thermal radiation from being
transmitted through the device, a useful property for thermal homeostasis. Figure 7.6 shows a
comparison between the VO2/Si (Fig. 7.6(a)) and VO2/Si/Au (Fig. 7.6(b)) systems. Smoothed lines
are superimposed as a guide to the eye. Both structures have higher broadband emissivity in the
metallic state of VO2 than in the insulating state (Figs. 7.6(c) and 7.6(d)). A key difference between
the two structures, however, is their transmissivity. The VO 2/Si/Au structure has zero
transmissivity in both the metallic and insulating states. When it is used to cover an external body
(e.g. experimental sample holder, or object whose temperature we wish to regulate) the total
thermal emission depends only on the emissivity of the VO 2/Si/Au stack, not that of the external
body. We thus use a gold back reflector in experiments.
62

Figure 7.6. (a) Schematic of VO2/ Si, and (b) schematic of VO2/ Si/ Au layered structures with respective
simulated (c), (d) emissivity and (e), (f) transmissivity corresponding to VO2 insulating and metallic states.
7.4 Measurement of infrared device properties
We fabricated a VO2/Si/Au device with a VO 2 thickness of 62nm, close to the optimal value
calculated in Figure 7.5. We measured the infrared absorptivity as a function of wavelength for
both the insulating and metallic states using FTIR. Figure 7.7 compares the experimental
measurement to smoothed simulation results (see Methods for details). Simulations and FTIR
measurements in Figs. 7.7(a, b) show similar broadband emissivity switching: the emissivity is
higher in the metallic state than the insulating state. The integrated difference in radiation power
𝛥 𝑃 𝑟𝑎𝑑 calculated from the spectra is equal to 0.29 (simulated spectra) and 0.22 (experimental
spectra) (Fig. 7.7(c)). The results suggest that the fabricated sample should emit significantly
more heat in the hot (metallic) state, a necessary feature for thermal homeostasis.
Simulations capture most of the measured FTIR spectral features. An offset between the
simulated and measured spectra in the insulating state of Fig 7.7(a) is observed at wavelengths
below 15um. This is likely due to a difference between the optical constants of silicon in
experiment and the values used in simulation (taken from Ref. [139]).
63

Figure 7.7. Measured and simulated (absorptivity) emissivity spectra of insulating (a) and metallic (b) state
of VO2/Si/Au stack. (c) Calculated integrated radiation power normalized to blackbody spectrum.
7.5 Experimental setup and calibration
We designed an experiment to directly test the temperature regulation capabilities of our
device. A photograph and schematic of the experiment are shown in Figures 7.8(a) and 7.8(b),
respectively. Device samples are mounted on either side of a ceramic heater, containing an
embedded thermocouple. The entire structure is suspended in a vacuum chamber, which has an
interior black surface to prevent infrared reflection. The chamber is submerged in an ice bath at
an ambient temperature of To = 0.5°C. The heat load on the sample is varied by changing the
input power to the heater, and the resulting temperature is recorded by the thermocouple.
As shown in the bottom portion of Figure 7.8(b), the system loses heat through two
mechanisms: radiation from the sample, and parasitic loss (which includes both radiation from the
perimeter of the sample and conduction to the wire). To calibrate the parasitic loss, we use gold
mirrors with low, constant emissivity and measure the temperature rise as a function of applied
heat load (yellow circles in Figure 7.8(c)). The experiment is conducted by using a complete
heating and cooling cycle. The temperature is first increased in discrete steps, and then
decreased again. At each temperature value, we allow 45 minutes for the system to reach steady
state. The parasitic heat loss (Qloss(T)) is determined from these results. For each value of applied
64

heat load Q, we subtracted the calculated, net radiative loss of gold to obtain Qloss(T), plotted in
Fig. 7.8(d).

Figure 7.8. Overview of experimental setup. (a) Photograph and (b) schematic of vacuum chamber used to
perform thermal measurements. (c) Measured temperature rise as a function of applied heat load for mirror-
gold (low emissivity) and diffuse black (high-emissivity) samples, with respect to an ambient temperature
of 0.5˚C. (d) Measured parasitic heat loss (sum of radiation losses from perimeter and conduction losses
to wire), as a function of temperature, extracted from mirror gold data in (c).
To probe the dynamic range of our measurement system, we also measure a diffuse-black
sample with a high total emissivity. Results are shown in Figure 7.8(c). The data curve for the
diffuse-black sample is well separated from the curve for the mirror-gold sample. These two
measurements, at the extremes of high and low emissivity, define an operational window for our
subsequent, variable-emissivity measurements.
7.6 Measurement of device emissivity
We next measure the temperature rise as a function of applied heat load for our VO2 devices.
The results are shown in Figure 7.9(a). The heating and cooling curves trace a hysteresis window
around the VO2 phase transition. Inside this window, for constant applied heat load, there is a
65

temperature difference as large as ~5 °C between the heating and cooling curves. The overlap of
the heating and cooling curves above and below the hysteresis window suggests that there is
negligible system drift in the experimental setup. The location of the phase transition can be more
readily observed by plotting the derivative of the heat load-temperature curve (Figure 7.9(a),
inset). During heating, the response dQ/dT peaks in the red, shaded region, indicating the
transition from insulator to metal at ~80 °C. Upon cooling, dQ/dT peaks at a lower temperature
~60 °C, indicating transition back to insulator. We ran this measurement over two complete
heating/cooling cycles (circles and diamonds) to ensure that there was minimal run-to-run
variation in thermal response.
We can use the data from Fig. 7.9(a) along with the calibration curve in Fig. 7.8(d) to determine
the radiative heat flux emitted by the VO 2 sample, 𝑄 𝑟𝑎𝑑 ′′
( 𝑇 ) . In steady state, the net heat input to
the system is equal to the output:
𝘀𝑄 ( 𝑇 )+ 𝘀 ( 𝑇 ) 𝜎 𝐴𝑇
0
4
= 𝑄 𝑟𝑎𝑑 ′′
( 𝑇 ) 𝐴 + 𝑄 𝑙𝑜𝑠𝑠 ( 𝑇 ) , Eq. 7.1
where the second term on the left-hand side represents the absorptive heat flux at the sample
surface due to ambient radiation. This equation can be solved to yield
𝘀 ( 𝑇 ) =
Q( 𝑇 )− Q
loss
( 𝑇 )
𝜎𝐴 ( 𝑇 4
− 𝑇 0
4
)
, Eq. 7.2
assuming that
𝑄 𝑟𝑎𝑑 ′′
( 𝑇 ) = 𝘀 ( 𝑇 ) 𝜎 𝑇 4
, Eq. 7.3
The radiative heat flux is plotted in Figure 7.9(b). Superimposed on the graph are contour
lines indicating constant values of emissivity. The graph shows that along the heating curve, the
radiative heat flux increases sharply near the upper edge of the hysteresis loop. This corresponds
to an increase in emissivity. Along the cooling curve, the radiative flux drops at the lower edge of
66

the loop, corresponding to a decrease in emissivity. The inset of Figure 7.9(b) replots the data to
show emissivity as a function of temperature. It can be seen that εins = 0.22 in the insulator phase,
and εmet = 0.46 in the metallic phase. These values are consistent with those measured using
FTIR microscopy ( εins = 0.22 and εmet = 0.44).

Figure 7.9. (a) Measured temperature rise relative to ambient (horizontal axis) as a function of the applied
heat load (vertical axis) for a complete heating and cooling cycle. The inset shows the derivative applied
heat load with respect to temperature for two runs (complete heating/cooling cycles). (b) Calculated
radiative heat flux from the VO2 surface as a function of temperature. Constant emissivity curves are
plotted in grey. The inset shows the effective emissivity of the measured sample as function of
temperature.

67

7. Dynamic Thermal Regulation
To demonstrate dynamic thermal regulation, we apply a time-varying heat load and measure
the resulting temperature. The input power is plotted in Fig. 7.10(a), and has the form of a square
wave with power levels of 0.22 and 0.59W.
For reference, we first measure a near constant-emissivity structure with an alumina top layer.
(Al2O3/ Si/ Au with the corresponding thicknesses of 480nm/ 200µm/ 60nm). The experimental,
time-dependent temperature data is shown by the red, dotted line in Fig. 7.10(b). In response to
an increase in input power, the measured temperature rises and then plateaus. When the input
power is decreased, the temperature drops again and stabilizes at a lower value. The total range
of temperature fluctuation measured is 56°C (red arrows). The measured results can be
accurately reproduced using a numerical heat transfer model given by
𝜌𝐶 𝐿 𝑐 𝑑𝑇 ( 𝑡 )
𝑑𝑡
=
𝐼𝑉 − 𝑄 𝑙𝑜𝑠𝑠 ( 𝑇 )
𝐴 − 𝜎𝘀 ( 𝑇 ) ( 𝑇 4
− 𝑇 0
4
) ,
Eq. 7.4
where ρ is the effective material density (kg ∕m
3
), C is the effective heat capacitance (J ∕K- kg),
LC is the characteristic length scale of the system (m), and 𝑇 0
=0.5°C is the ambient temperature.
The numerical solution to Eq. 7.4 is shown by the red, solid line in Fig. 7.10(b). Physically, the
response time of the device is determined by the effective heat capacity, material density,
diffusion length and the emissivity of the system. The simulation shows an excellent match to
experiment for a fitted value of ρ C LC =5500 J/(m
2
-K).
We next measure the performance of our variable-emissivity VO2 device. The experimental
data is shown by the blue, dotted line in Figure 7.10(b). In comparison to the constant-emissivity
Al2O 3 device, the total temperature fluctuations are reduced to a value of 50°C. The data can
again be well modeled by Equation 7.4, as shown in Figure 7.10. Physically, the strong change
68

in emissivity at the phase transition decreases the total temperature fluctuation resulting from a
given heat load. This result illustrates the principle of thermal homeostasis.

Figure 7.10. Thermal Homeostasis. (a) Square wave time-varying input heat power. (b) measured and
calculated response to the input power in for VO2/ Si/ Au (62nm/ 200µm/ 60nm) structure and Al2O3/ Si/ Au
(480nm/ 200µm/ 60nm).
7.6 Discussion
In space applications, under ideal conditions, radiative loss is the only heat dissipation
mechanism; parasitic losses vanish. We can use our thermal model to predict the performance of
our VO2 device under these conditions. In the absence of parasitic losses, thermal self-regulation
of the device is far more effective than under laboratory conditions. We choose input powers of
0.037 W and 0.146 W to ensure that the radiative heat loss from the sample is the same. In this
case, the temperature fluctuations in the VO 2 device are again around 50°C, as in the experiment
of Fig. 7.10. However, the fluctuations for the constant-emissivity Al2O3 sample are now 108°C.
This increase is due to the absence of the parasitic loss pathway. The VO 2 sample can therefore
self-regulate its own temperature far better than the constant-emissivity sample.
In fact, the magnitude of fluctuations in the VO 2 device can be predicted directly from Fig.
7.9(b). In the absence of parasitic loss, the steady state radiative heat flux is equal to the input
69

power per unit area. From Fig. 7.9(b), a value of 112 W/m
2
corresponds to a temperature of ~39
°C, while a value of 442 W/m
2
corresponds to a temperature of ~89 °C. These values correspond
well with those obtained in the simulation of Fig. 7.11(b). For the constant emissivity sample, the
temperature fluctuations are much higher. Approximating the Al2O3 sample with a constant
emissivity of 0.35, the lower power level corresponds to a temperature of 6 °C, whereas the upper
power level corresponds to a temperature of 114 °C, lying well outside the edges of Fig. 7.9(b).
This corresponds to the larger fluctuation of 108 °C seen in Fig. 7.11(b). A lower bound on the
fluctuations is given by the width of the hysteresis curve. For our experimental device, this is close
to 20°C. Further improvement in material quality can bring this number down substantially, as
observed in literature [140-142]. Another route to performance improvement is to incorporate
microstructured designs [115, 118, 119] to increase the total difference in radiated power between
metal and insulator states. In this case, for fixed value of temperature fluctuation, the device is
expected to accommodate a larger variation in input heat load. The experimental and thermal
modeling methods form a general platform for further investigation of dynamic thermal regulation
in variable-emissivity systems.

Figure 7.11. Thermal homeostasis in space. (a) Square wave time-varying input heat power. (b) calculated
response to the input power in for VO2/ Si/ Au (62nm/ 200µm/ 60nm) structure and Al2O3/ Si/ Au (480nm/
200µm/ 60nm).
70

7.7 Conclusion
We have directly demonstrated dynamic, passive thermal regulation via experiments on a VO 2
phase-change device. Our device is designed to optimize the increase in radiated power at the
phase transition. This trend allows the sample to “self-regulate” its temperature in response to a
time-varying, input heat load. Under laboratory conditions, the VO 2 device shows a reduction in
thermal fluctuations relative to a constant-emissivity device. Using a thermal model, we can
extrapolate the device performance to conditions typical of outer space, where radiation is the
only heat loss pathway and parasitic losses vanish. Our results demonstrate that emissivity
switching can reduce the thermal fluctuations by up to a factor of 2.
Recent investigations [143, 144] have shown flexibility in tuning the phase-transition
temperature of VO2 from 28 to 63°C through doping, addition of dopant atoms, or alloying films.
This suggests that various devices could be designed to regulate temperature around fixed values
in this range. Moreover, transfer techniques [33] may allow additional flexibility in choice of
substrate. In terms of ultimate applications, the work presented here provides a key step toward
understanding a larger trade space, one that incorporates not only material selection, but also
system-level concerns such as payload target temperature and solar heat load.
7.8 Future Work
7.8.1 Optimized Multilayer Planar Structures
In the previous section, we have experimentally shown that an optimized VO2 planar structure
is able to self-regulate its temperature. To enhance the thermal regulation (be resistant to higher
fluctuations in input heat), further optimization can aim at increasing the switching in the radiation
power.
Using a quasi-periodic structure adds one more degree of freedom that can be used to
maximize the figure of merit (Δ𝑃 𝑟𝑎𝑑 ). Figure 7.12 shows a flow chart for a random walk algorithm
that utilizes a 2x2 cells of the micro-pillars structure. At each iteration, a small update of the pillars’
71

location is randomly picked. The figure of merit (FOM) is then calculated using TMM simulations.
If the FOM is increased, the updated location is kept, otherwise a new update is chosen from the
original state.

Figure 7.12. Flow chart of the random walk optimization algorithm.
The starting point is a low aspect ratio (h/a=1), chosen form Fig. 7.13 (a). From the spectrum
in Fig. 7.13(b), a quasi-periodic structure aims at flattening the emissivity peak around 8μm, and
thus increasing the integrated radiation power.
72

Figure 7.13. (a) Metallic radiation power as a function of rod height and VO2 thickness. (b) High aspect ratio
optimal point of metallic Prad. (c) Low aspect ratio optimal point of metallic Prad.
Figure 7.14 (a) shows a result of 30 iterations using the random walk algorithm to optimize
the FOM. We can see that the slight improvement in FOM from 0.7748 to 0.7800 corresponds to
flattening the spectral peak in Fig. 7.14 (b). Further optimization using the same algorithm can
use a larger super cell (3x3), different pillar diameters in each cell, relax the geometrical
constraints by allowing overlap and/ or random walk with random restarts to avoid local maxima.
73

Figure 7.14. (a) Result of applying random walk optimization to a 2x2 unit cell. (b) Metallic state emissivity
spectrum of the initial and final states.
Another option is to further optimize the planar structure by introducing a small cavity between
the phase change material and the back reflector. With integrated emissivity switching around
20% achieved in the preliminary experiment in the previous section, various combinations of
multilayer structures can be used to enhance the switching to nearly 50% as shown in Figure
7.15.

Figure 7.15. (a) Simulation of the radiation power switching of the four layered structures in (b-d).
74

By creating an additional cavity below the VO 2 layer and above the Au mirror, field
enhancement causes the switching to increase to about 46% as highlighted in Fig. 7.16(b). We
will use numerical simulations to optimize a multilayered structure for maximum integrated
emissivity switching. The preliminary data suggests that an optimized structure has the potential
to perform as good as the optimum micro-cone structure for normal incidence emissivity
switching.

Figure 7.16. (a) Schematic of a Si/VO2/ZnSe/Au thermal homeostasis structure. (b) Effect of the thickness
of the ZnSe layer on the normalized thermal radiation power.
7.8.2 Thermally- regulating Paint
A version of the results in this chapter was published as Ref. [119]
For large area (paint) applications, we propose here a design for VO 2-coated microparticles that
can achieve similar temperature regulation properties to the previous micro-cone design.
Incorporation of such microparticles into paint could greatly widen the potential applications of
thermal homeostasis.
Our microparticle system is shown schematically in Figure 7.17. The layer of particles is made
up of VO2-coated silicon spheres resting on aluminum. In the eventual application, we assume
that the microparticles would be mixed into an infrared-transparent polymer such as polyethylene
for coating onto the substrate. For computational simplicity, we assume that the microparticle
75

layer has a hexagonal, close-packed configuration. VO2 has phase transition from insulating state
to metallic state at 𝑇 𝑐 ~330 K [143, 145-147]. The phase transition is accompanied by a dramatic
change in optical properties. For the most effective temperature regulation, the structure should
be designed to have as low emissivity as possible in the insulating state, and as high emissivity
as possible in the metallic state [118].

Figure 7.17. Schematic of the microparticle-based paint. A single layer of close-packed VO2-coated Si
microspheres. The entire microstructure is embedded in polyethylene and resting on an Al substrate.
Figure 7.18 presents the temperature regulation effect achieved by our thermal homeostatic
“paint” (Fig. 7.18(a)). For reference, we also consider the planar structure (Fig. 7.18(b)) and an
uncoated Al substrate (Fig. 7.18(c)). Note that for each case, we use the corresponding set of
structural parameters that gives the largest ∆𝑃 𝑟𝑎𝑑 . The thickness of the substrate for all cases is
300 μm. Given the calculated values of 𝑃 𝑟𝑎𝑑 ( 𝑇 𝑐 ) , which are shown as symbols in Fig 7.18(e), we
assume a model for 𝑃 𝑟𝑎𝑑 ( 𝑇 ) that includes the hysteresis of the phase transition [118]. The full
model of 𝑃 𝑟𝑎𝑑 ( 𝑇 ) is shown by the curves in Fig 7.18(e). The microsphere “paint” has a larger
difference in 𝑃 𝑟𝑎𝑑 between the insulating (filled, green circle) and metallic (filled, green square)
states, than the planar case (hollow, magenta symbols).
Given a fluctuating heat input, the microparticle paint significantly reduces temperature
variation. We apply the periodic, time-varying input power shown in the top panel of Fig. 7.18(f),
which imitates, for example, time-varying solar illumination or internal heat load. The value of 𝑃 𝑖𝑛

oscillates between 50 and 350 W/m
2
on a time scale of 1000s. To demonstrate the thermal
dynamics of the system, we solve the time-dependent heat equation to obtain the temperature of
the system [19]. In response to the fluctuating input, the temperature of all three structures
76

oscillates (Fig. 7.18(f), bottom). The amplitude of oscillation is largest for uncoated Al. The VO 2-
coating on a flat Al substrate reduces the oscillations from 175 K (dotted, black curve) to 100 K
(dotted, magenta curve). In comparison, the microparticles design has a nearly constant
temperature response (solid, green curve): the fluctuation amplitude is reduced to 12.9 K, a factor
of ~13x smaller than for uncoated Al.
We note that the modulation time scale of 1000s shown in Fig. 7.18 (f) (top) was sufficient for
the microsphere structure to reach an equilibrium temperature at each value of 𝑃 𝑖𝑛
(green curve
in Fig. 7.18 (f) (bottom)). For uncoated Al, the radiated power (𝑃 𝑟𝑎𝑑 , shown in Fig. 7.18 (e)) is
much lower, and the time scale for equilibration is longer. It is thus the case that if 𝑃 𝑖𝑛
is modulated
more slowly, the temperature fluctuation for uncoated Al will be larger, and the overall reduction
in temperature fluctuation provided by the microsphere design will be greater.

Figure 7.18. Thermal homeostasis and temperature regulation. (a) The microsphere “paint.” (b) A planar,
VO2-coated film. (c) Uncoated film. (d) Boundary conditions used to solve the heat equation. (e) Radiated
thermal power as a function of temperature. The arrows indicate the direction of heating or cooling
processes. The symbols represent the calculated 𝑃 𝑟𝑎𝑑 ( 𝑇 𝑐 ) for metallic (squares) or insulating (circles)
states. (f) Temperature variation of different structures for a time-varying heat input flux.

77

References
1. S. Fan and J. D. Joannopoulos, "Analysis of guided resonances in photonic crystal slabs,"
Physical Review B 65, 235112 (2002).
2. K. B. Crozier, V. Lousse, O. Kilic, S. Kim, S. Fan, and O. Solgaard, "Air-bridged photonic
crystal slabs at visible and near-infrared wavelengths," Physical Review B 73, 115126 (2006).
3. Z. Qiang, H. Yang, L. Chen, H. Pang, Z. Ma, and W. Zhou, "Fano filters based on transferred
silicon nanomembranes on plastic substrates," Applied Physics Letters 93, 061106 (2008).
4. W. Zhou, Z. Ma, H. Yang, Z. Qiang, G. Qin, H. Pang, L. Chen, W. Yang, S. Chuwongin, and
D. Zhao, "Flexible photonic-crystal Fano filters based on transferred semiconductor
nanomembranes," Journal of Physics D: Applied Physics 42, 234007 (2009).
5. M. Huang, A. A. Yanik, T.-Y. Chang, and H. Altug, "Sub-wavelength nanofluidics in photonic
crystal sensors," Opt. Express 17, 24224-24233 (2009).
6. M. El Beheiry, V. Liu, S. Fan, and O. Levi, "Sensitivity enhancement in photonic crystal slab
biosensors," Opt. Express 18, 22702-22714 (2010).
7. L. C. Estrada, O. E. Martinez, M. Brunstein, S. Bouchoule, L. Le-Gratiet, A. Talneau, I.
Sagnes, P. Monnier, J. A. Levenson, and A. M. Yacomotti, "Small volume excitation and
enhancement of dye fluorescence on a 2D photonic crystal surface," Opt. Express 18, 3693-
3699 (2010).
8. B. Zhen, S.-L. Chua, J. Lee, A. W. Rodriguez, X. Liang, S. G. Johnson, J. D. Joannopoulos,
M. Soljačić, and O. Shapira, "Enabling enhanced emission and low-threshold lasing of organic
molecules using special Fano resonances of macroscopic photonic crystals," Proceedings of
the National Academy of Sciences 110, 13711 (2013).
9. T. Kaji, T. Yamada, R. Ueda, and A. Otomo, "Enhanced Fluorescence Emission from Single
Molecules on a Two-Dimensional Photonic Crystal Slab with Low Background Emission," The
Journal of Physical Chemistry Letters 2, 1651-1656 (2011).
10. N. Ganesh, W. Zhang, P. C. Mathias, E. Chow, J. A. N. T. Soares, V. Malyarchuk, A. D. Smith,
and B. T. Cunningham, "Enhanced fluorescence emission from quantum dots on a photonic
crystal surface," Nature Nanotechnology 2, 515 (2007).
11. J. Mouette, C. Seassal, X. Letartre, P. Rojo-Romeo, J.-L. Leclercq, P. Regreny, P.
Viktorovitch, E. Jalaguier, P. Perreau, and H. Moriceau, "Very low threshold vertical emitting
laser operation in InP graphite photonic crystal slab on silicon," in Electronics Letters,
(Institution of Engineering and Technology, 2003), pp. 526-528.
12. F. Raineri, G. Vecchi, A. M. Yacomotti, C. Seassal, P. Viktorovitch, R. Raj, and A. Levenson,
"Doubly resonant photonic crystal for efficient laser operation: Pumping and lasing at low
group velocity photonic modes," Applied Physics Letters 86, 011116 (2004).
13. A. A. Erchak, D. J. Ripin, S. Fan, P. Rakich, J. D. Joannopoulos, E. P. Ippen, G. S. Petrich,
and L. A. Kolodziejski, "Enhanced coupling to vertical radiation using a two-dimensional
photonic crystal in a semiconductor light-emitting diode," Applied Physics Letters 78, 563-565
(2001).
14. S. Noda, M. Fujita, and T. Asano, "Spontaneous-emission control by photonic crystals and
nanocavities," Nature Photonics 1, 449 (2007).
78

15. L. J. Martínez, N. Huang, J. Ma, C. Lin, E. Jaquay, and M. L. Povinelli, "Design and optical
characterization of high-Q guided-resonance modes in the slot-graphite photonic crystal
lattice," Opt. Express 21, 30975-30983 (2013).
16. C. Lin and M. L. Povinelli, "Optical absorption enhancement in silicon nanowire arrays with a
large lattice constant for photovoltaic applications," Opt. Express 17, 19371-19381 (2009).
17. E. Garnett and P. Yang, "Light Trapping in Silicon Nanowire Solar Cells," Nano Letters 10,
1082-1087 (2010).
18. S. E. Han and G. Chen, "Optical Absorption Enhancement in Silicon Nanohole Arrays for Solar
Photovoltaics," Nano Letters 10, 1012-1015 (2010).
19. A. M. Morsy, R. Biswas, and M. L. Povinelli, "High temperature, experimental thermal memory
based on optical resonances in photonic crystal slabs," APL Photonics 4, 010804 (2019).
20. J. B. Casady and R. W. Johnson, "Status of silicon carbide (SiC) as a wide-bandgap
semiconductor for high-temperature applications: A review," Solid-State Electronics 39, 1409-
1422 (1996).
21. P. G. Neudeck, R. S. Okojie, and C. Liang-Yu, "High-temperature electronics - a role for wide
bandgap semiconductors?," Proceedings of the IEEE 90, 1065-1076 (2002).
22. R. W. Johnson, J. L. Evans, P. Jacobsen, J. R. Thompson, and M. Christopher, "The changing
automotive environment: high-temperature electronics," IEEE Transactions on Electronics
Packaging Manufacturing 27, 164-176 (2004).
23. O. Y. Loh and H. D. Espinosa, "Nanoelectromechanical contact switches," Nature
Nanotechnology 7, 283 (2012).
24. B. Li, L. Wang, and G. Casati, "Thermal Diode: Rectification of Heat Flux," Physical Review
Letters 93, 184301 (2004).
25. C. R. Otey, W. T. Lau, and S. Fan, "Thermal Rectification through Vacuum," Physical Review
Letters 104, 154301 (2010).
26. N. Li, J. Ren, L. Wang, G. Zhang, P. HÃ¤nggi, and B. Li, "Colloquium: Phononics:
Manipulating heat flow with electronic analogs and beyond," Reviews of Modern Physics 84,
1045-1066 (2012).
27. P. Ben-Abdallah and S.-A. Biehs, "Phase-change radiative thermal diode," Applied Physics
Letters 103, 191907 (2013).
28. P. Ben-Abdallah and S.-A. Biehs, "Near-Field Thermal Transistor," Physical Review Letters
112, 044301 (2014).
29. J. Ordonez-Miranda, Y. Ezzahri, J. Drevillon, and K. Joulain, "Transistorlike Device for Heating
and Cooling Based on the Thermal Hysteresis of VO 2," Physical Review Applied 6, 054003
(2016).
30. L. Wang and B. Li, "Thermal Logic Gates: Computation with Phonons," Physical Review
Letters 99, 177208 (2007).
31. P. Ben-Abdallah and S.-A. Biehs, "Towards Boolean operations with thermal photons,"
Physical Review B 94, 241401 (2016).
79

32. L. Wang and B. Li, "Thermal Memory: A Storage of Phononic Information," Physical Review
Letters 101, 267203 (2008).
33. M. Elzouka and S. Ndao, "Near-field NanoThermoMechanical memory," Applied Physics
Letters 105, 243510 (2014).
34. V. Kubytskyi, S.-A. Biehs, and P. Ben-Abdallah, "Radiative Bistability and Thermal Memory,"
Physical Review Letters 113, 074301 (2014).
35. C. Khandekar and A. W. Rodriguez, "Thermal bistability through coupled photonic
resonances," Applied Physics Letters 111, 083104 (2017).
36. S. A. Dyakov, J. Dai, M. Yan, and M. Qiu, "Near field thermal memory based on radiative
phase bistability of VO 2," Journal of Physics D: Applied Physics 48, 305104 (2015).
37. X. Rongguo, B. C. Tinh, V. Binni, Z. Qingxin, S. C. Haur, L. Baowen, and T. J. T. L., "An
Electrically Tuned Solid‐State Thermal Memory Based on Metal–Insulator Transition of
Single‐Crystalline VO2 Nanobeams," Advanced Functional Materials 21, 1602-1607 (2011).
38. K. Ito, K. Nishikawa, and H. Iizuka, "Multilevel radiative thermal memory realized by the
hysteretic metal-insulator transition of vanadium dioxide," Applied Physics Letters 108,
053507 (2016).
39. P. E. Barclay, K. Srinivasan, and O. Painter, "Nonlinear response of silicon photonic crystal
microresonators excited via an integrated waveguide and fiber taper," Opt. Express 13, 801-
820 (2005).
40. M. Notomi, A. Shinya, S. Mitsugi, G. Kira, E. Kuramochi, and T. Tanabe, "Optical bistable
switching action of Si high-Q photonic-crystal nanocavities," Opt. Express 13, 2678-2687
(2005).
41. T. Uesugi, B.-S. Song, T. Asano, and S. Noda, "Investigation of optical nonlinearities in an
ultra-high-Q Si nanocavity in a two-dimensional photonic crystal slab," Opt. Express 14, 377-
386 (2006).
42. A. M. Yacomotti, F. Raineri, G. Vecchi, P. Monnier, R. Raj, A. Levenson, B. Ben Bakir, C.
Seassal, X. Letartre, P. Viktorovitch, L. Di Cioccio, and J. M. Fedeli, "All-optical bistable band-
edge Bloch modes in a two-dimensional photonic crystal," Applied Physics Letters 88, 231107
(2006).
43. M.-K. Kim, I.-K. Hwang, S.-H. Kim, H.-J. Chang, and Y.-H. Lee, "All-optical bistable switching
in curved microfiber-coupled photonic crystal resonators," Applied Physics Letters 90, 161118
(2007).
44. M. Notomi, T. Tanabe, A. Shinya, E. Kuramochi, and H. Taniyama, "On-Chip All-Optical
Switching and Memory by Silicon Photonic Crystal Nanocavities," Advances in Optical
Technologies 2008(2008).
45. L.-D. Haret, T. Tanabe, E. Kuramochi, and M. Notomi, "Extremely low power optical bistability
in silicon demonstrated using 1D photonic crystal nanocavity," Opt. Express 17, 21108-21117
(2009).
46. R. Shankar, I. Bulu, R. Leijssen, and M. Lončar, "Study of thermally-induced optical bistability
and the role of surface treatments in Si-based mid-infrared photonic crystal cavities," Opt.
Express 19, 24828-24837 (2011).
80

47. Q. M. Ngo, K. Q. Le, and V. D. Lam, "Optical bistability based on guided-mode resonances in
photonic crystal slabs," J. Opt. Soc. Am. B 29, 1291-1295 (2012).
48. P.-C. Hui, D. Woolf, E. Iwase, Y.-I. Sohn, D. Ramos, M. Khan, A. W. Rodriguez, S. G.
Johnson, F. Capasso, and M. Loncar, "Optical bistability with a repulsive optical force in
coupled silicon photonic crystal membranes," Applied Physics Letters 103, 021102 (2013).
49. V. R. Almeida and M. Lipson, "Optical bistability on a silicon chip," Opt. Lett. 29, 2387-2389
(2004).
50. N. Vukovic, N. Healy, P. Mehta, T. D. Day, P. J. A. Sazio, J. V. Badding, and A. C. Peacock,
"Thermal nonlinearity in silicon microcylindrical resonators," Applied Physics Letters 100,
181101 (2012).
51. F. Ramiro-Manzano, N. Prtljaga, L. Pavesi, G. Pucker, and M. Ghulinyan, "Thermo-optical
bistability with Si nanocrystals in a whispering gallery mode resonator," Opt. Lett. 38, 3562-
3565 (2013).
52. T. Carmon, L. Yang, and K. J. Vahala, "Dynamical thermal behavior and thermal self-stability
of microcavities," Opt. Express 12, 4742-4750 (2004).
53. D. L. C. Chan, I. Celanovic, J. D. Joannopoulos, and M. Soljačić, "Emulating one-dimensional
resonant Q-matching behavior in a two-dimensional system via Fano resonances," Physical
Review A 74, 064901 (2006).
54. C. Lin, L. J. Martínez, and M. L. Povinelli, "Experimental broadband absorption enhancement
in silicon nanohole structures with optimized complex unit cells," Opt. Express 21, A872-A882
(2013).
55. R. Biswas and M. L. Povinelli, "Photonic surfaces for designable nonlinear power shaping,"
Applied Physics Letters 106, 061110 (2015).
56. R. Biswas and M. L. Povinelli, "Sudden, Laser-Induced Heating through Silicon
Nanopatterning," ACS Photonics 2, 1681-1685 (2015).
57. A. M. Morsy, R. Biswas, and M. L. Povinelli, "Sequential and selective localized optical heating
in water via on-chip dielectric nanopatterning," Opt. Express 25, 17820-17831 (2017).
58. C. Lin, L. J. Martínez, and M. L. Povinelli, "Fabrication of transferrable, fully suspended silicon
photonic crystal nanomembranes exhibiting vivid structural color and high-Q guided
resonance," Journal of Vacuum Science & Technology B, Nanotechnology and
Microelectronics: Materials, Processing, Measurement, and Phenomena 31, 050606 (2013).
59. P. Pottier, L. Shi, and Y.-A. Peter, "Determination of guided-mode resonances in photonic
crystal slabs," J. Opt. Soc. Am. B 29, 109-117 (2012).
60. S. Aritome, R. Shirota, G. Hemink, T. Endoh, and F. Masuoka, "Reliability issues of flash
memory cells," Proceedings of the IEEE 81, 776-788 (1993).
61. R. Shirota, B. J. Yang, Y. Y. Chiu, H. T. Chen, S. F. Ng, P. Y. Wang, J. H. Chang, and I.
Kurachi, "New Method to Analyze the Shift of Floating Gate Charge and Generated Tunnel
Oxide Trapped Charge Profile in NAND Flash Memory by Program/Erase Endurance," IEEE
Transactions on Electron Devices 62, 114-120 (2015).
62. M. Brunstein, R. Braive, R. Hostein, A. Beveratos, I. Robert-Philip, I. Sagnes, T. J. Karle, A.
M. Yacomotti, J. A. Levenson, V. Moreau, G. Tessier, and Y. De Wilde, "Thermo-optical
81

dynamics in an optically pumped Photonic Crystal nano-cavity," Opt. Express 17, 17118-
17129 (2009).
63. Y. Shi, X. Chen, F. Lou, Y. Chen, M. Yan, L. Wosinski, and M. Qiu, "All-optical switching of
silicon disk resonator based on photothermal effect in metal–insulator–metal
absorber," Opt. Lett. 39, 4431-4434 (2014).
64. C. Qiu, Y. Yang, C. Li, Y. Wang, K. Wu, and J. Chen, "All-optical control of light on a graphene-
on-silicon nitride chip using thermo-optic effect," Scientific Reports 7, 17046 (2017).
65. D. L. C. Chan, I. Celanovic, J. D. Joannopoulos, and M. Soljačić, "Emulating one-dimensional
resonant $Q$-matching behavior in a two-dimensional system via Fano resonances," Physical
Review A 74, 064901 (2006).
66. L. Cao, D. N. Barsic, A. R. Guichard, and M. L. Brongersma, "Plasmon-Assisted Local
Temperature Control to Pattern Individual Semiconductor Nanowires and Carbon
Nanotubes," Nano letters 7, 3523-3527 (2007).
67. J. S. Donner, G. Baffou, D. McCloskey, and R. Quidant, "Plasmon-Assisted Optofluidics," ACS
Nano 5, 5457-5462 (2011).
68. S. S. Aćimović, M. P. Kreuzer, M. U. González, and R. Quidant, "Plasmon Near-Field Coupling
in Metal Dimers as a Step toward Single-Molecule Sensing," ACS Nano 3, 1231-1237 (2009).
69. J. N. Anker, W. P. Hall, O. Lyandres, N. C. Shah, J. Zhao, and R. P. Van Duyne, "Biosensing
with plasmonic nanosensors," Nat Mater 7, 442-453 (2008).
70. G. Baffou and R. Quidant, "Thermo-plasmonics: using metallic nanostructures as nano-
sources of heat," Laser & Photonics Reviews 7, 171-187 (2013).
71. B. Paolo, H. Jer-Shing, and H. Bert, "Nanoantennas for visible and infrared radiation," Reports
on Progress in Physics 75, 024402 (2012).
72. Z. Qin and J. C. Bischof, "Thermophysical and biological responses of gold nanoparticle laser
heating," Chemical Society Reviews 41, 1191-1217 (2012).
73. G. Baffou, E. B. Urena, P. Berto, S. Monneret, R. Quidant, and H. Rigneault, "Deterministic
temperature shaping using plasmonic nanoparticle assemblies," Nanoscale 6, 8984-8989
(2014).
74. C. Lin, L. J. Martínez, and M. L. Povinelli, "Fabrication of transferrable, fully suspended silicon
photonic crystal nanomembranes exhibiting vivid structural color and high-Q guided
resonance," Journal of Vacuum Science & Technology B 31, 050606 (2013).
75. E.-A. Brujan and A. Vogel, "Stress wave emission and cavitation bubble dynamics by
nanosecond optical breakdown in a tissue phantom," Journal of Fluid Mechanics 558, 281-
308 (2006).
76. M. Eghtedari, A. Oraevsky, J. A. Copland, N. A. Kotov, A. Conjusteau, and M. Motamedi,
"High Sensitivity of In Vivo Detection of Gold Nanorods Using a Laser Optoacoustic Imaging
System," Nano letters 7, 1914-1918 (2007).
77. C. Kim, E. C. Cho, J. Chen, K. H. Song, L. Au, C. Favazza, Q. Zhang, C. M. Cobley, F. Gao,
Y. Xia, and L. V. Wang, "In vivo molecular photoacoustic tomography of melanomas targeted
by bio-conjugated gold nanocages," ACS Nano 4, 4559-4564 (2010).
82

78. P. Rogers and A. Neild, "Selective particle trapping using an oscillating microbubble," Lab on
a Chip 11, 3710-3715 (2011).
79. Y. Xie, C. Zhao, Y. Zhao, S. Li, J. Rufo, S. Yang, F. Guo, and T. J. Huang, "Optoacoustic
tweezers: a programmable, localized cell concentrator based on opto-thermally generated,
acoustically activated, surface bubbles," Lab on a Chip 13, 1772-1779 (2013).
80. C. Zhao, Y. Xie, Z. Mao, Y. Zhao, J. Rufo, S. Yang, F. Guo, J. D. Mai, and T. J. Huang, "Theory
and experiment on particle trapping and manipulation via optothermally generated bubbles,"
Lab on a Chip 14, 384-391 (2014).
81. Y. Zheng, H. Liu, Y. Wang, C. Zhu, S. Wang, J. Cao, and S. Zhu, "Accumulating microparticles
and direct-writing micropatterns using a continuous-wave laser-induced vapor bubble," Lab
on a Chip 11, 3816-3820 (2011).
82. A. Hashmi, G. Yu, M. Reilly-Collette, G. Heiman, and J. Xu, "Oscillating bubbles: a versatile
tool for lab on a chip applications," Lab on a Chip 12, 4216-4227 (2012).
83. S. Fan, W. Suh, and J. D. Joannopoulos, "Temporal coupled-mode theory for the Fano
resonance in optical resonators," J. Opt. Soc. Am. A 20, 569-572 (2003).
84. E. Guyon, H. J. P., L. Petit, and C. D. Mitescu, Physical Hydrodynamics (Oxford University
Press, USA, 2001).
85. J. B. Khurgin, G. Sun, W. T. Chen, W.-Y. Tsai, and D. P. Tsai, "Ultrafast Thermal Nonlinearity,"
Scientific Reports 5, 17899 (2015).
86. A. Crunteanu, J. Givernaud, J. Leroy, D. Mardivirin, C. Champeaux, J.-C. Orlianges, A.
Catherinot, and P. Blondy, "Voltage- and current-activated metal–insulator transition in VO2-
based electrical switches: a lifetime operation analysis," Science and Technology of Advanced
Materials 11, 065002 (2010).
87. X. Chen and Q. Lv, "Theoretical analysis on the switching speed and driving power dissipation
for optical switch based on VO2 thin film," Optik 124, 579-582 (2013).
88. P. Shi, Y.-A. Lin, M. Pastuszka, H. Cui, and J. A. MacKay, "Triggered Sorting and Co-
Assembly of Genetically Engineered Protein Microdomains in the Cytoplasm," Advanced
Materials 26, 449-454 (2014).
89. T. Inoue, M. De Zoysa, T. Asano, and S. Noda, "Realization of narrowband thermal emission
with optical nanostructures," Optica 2, 27-35 (2015).
90. J. G. Fleming, S. Y. Lin, I. El-Kady, R. Biswas, and K. M. Ho, "All-metallic three-dimensional
photonic crystals with a large infrared bandgap," Nature 417, 52-55 (2002).
91. J.-J. Greffet, R. Carminati, K. Joulain, J.-P. Mulet, S. Mainguy, and Y. Chen, "Coherent
emission of light by thermal sources," Nature 416, 61-64 (2002).
92. N. I. Landy, S. Sajuyigbe, J. J. Mock, D. R. Smith, and W. J. Padilla, "Perfect Metamaterial
Absorber," Physical Review Letters 100, 207402 (2008).
93. X. Liu, T. Tyler, T. Starr, A. F. Starr, N. M. Jokerst, and W. J. Padilla, "Taming the Blackbody
with Infrared Metamaterials as Selective Thermal Emitters," Physical Review Letters 107,
045901 (2011).
94. T. Maier and H. Brückl, "Wavelength-tunable microbolometers with metamaterial absorbers,"
Opt. Lett. 34, 3012-3014 (2009).
83

95. N. Liu, M. Mesch, T. Weiss, M. Hentschel, and H. Giessen, "Infrared Perfect Absorber and Its
Application As Plasmonic Sensor," Nano Letters 10, 2342-2348 (2010).
96. X. Liu, T. Starr, A. F. Starr, and W. J. Padilla, "Infrared Spatial and Frequency Selective
Metamaterial with Near-Unity Absorbance," Physical Review Letters 104, 207403 (2010).
97. C. M. Watts, X. Liu, and W. J. Padilla, "Metamaterial Electromagnetic Wave Absorbers,"
Advanced Materials 24, OP98-OP120 (2012).
98. H. Tao, C. M. Bingham, A. C. Strikwerda, D. Pilon, D. Shrekenhamer, N. I. Landy, K. Fan, X.
Zhang, W. J. Padilla, and R. D. Averitt, "Highly flexible wide angle of incidence terahertz
metamaterial absorber: Design, fabrication, and characterization," Physical Review B 78,
241103 (2008).
99. K. Aydin, V. E. Ferry, R. M. Briggs, and H. A. Atwater, "Broadband polarization-independent
resonant light absorption using ultrathin plasmonic super absorbers," Nature Communications
2, 517 (2011).
100. L. Zhu, S. Sandhu, C. Otey, S. Fan, M. B. Sinclair, and T. Shan Luk, "Temporal coupled
mode theory for thermal emission from a single thermal emitter supporting either a single
mode or an orthogonal set of modes," Applied Physics Letters 102, 103104 (2013).
101. M. C. K. Wiltshire, J. B. Pendry, I. R. Young, D. J. Larkman, D. J. Gilderdale, and J. V.
Hajnal, "Microstructured Magnetic Materials for RF Flux Guides in Magnetic Resonance
Imaging," Science 291, 849 (2001).
102. M. Bayindir, K. Aydin, E. Ozbay, P. Markoš, and C. M. Soukoulis, "Transmission properties
of composite metamaterials in free space," Applied Physics Letters 81, 120-122 (2002).
103. S. Linden, C. Enkrich, M. Wegener, J. Zhou, T. Koschny, and C. M. Soukoulis, "Magnetic
Response of Metamaterials at 100 Terahertz," Science 306, 1351 (2004).
104. T. J. Yen, W. J. Padilla, N. Fang, D. C. Vier, D. R. Smith, J. B. Pendry, D. N. Basov, and
X. Zhang, "Terahertz Magnetic Response from Artificial Materials," Science 303, 1494 (2004).
105. V. M. Shalaev, W. Cai, U. K. Chettiar, H.-K. Yuan, A. K. Sarychev, V. P. Drachev, and A.
V. Kildishev, "Negative index of refraction in optical metamaterials," Opt. Lett. 30, 3356-3358
(2005).
106. S. Zhang, W. Fan, B. K. Minhas, A. Frauenglass, K. J. Malloy, and S. R. J. Brueck,
"Midinfrared Resonant Magnetic Nanostructures Exhibiting a Negative Permeability," Physical
Review Letters 94, 037402 (2005).
107. H.-T. Chen, W. J. Padilla, J. M. O. Zide, A. C. Gossard, A. J. Taylor, and R. D. Averitt,
"Active terahertz metamaterial devices," Nature 444, 597-600 (2006).
108. H.-T. Chen, J. F. O'Hara, A. K. Azad, A. J. Taylor, R. D. Averitt, D. B. Shrekenhamer, and
W. J. Padilla, "Experimental demonstration of frequency-agile terahertz metamaterials,"
Nature Photonics 2, 295-298 (2008).
109. H.-T. Chen, W. J. Padilla, M. J. Cich, A. K. Azad, R. D. Averitt, and A. J. Taylor, "A
metamaterial solid-state terahertz phase modulator," Nature Photonics 3, 148-151 (2009).
110. A. P. Raman, M. A. Anoma, L. Zhu, E. Rephaeli, and S. Fan, "Passive radiative cooling
below ambient air temperature under direct sunlight," Nature 515, 540 (2014).
84

111. O. Ilic, C. M. Went, and H. A. Atwater, "Nanophotonic Heterostructures for Efficient
Propulsion and Radiative Cooling of Relativistic Light Sails," Nano Letters 18, 5583-5589
(2018).
112. A. R. Gentle and G. B. Smith, "Radiative Heat Pumping from the Earth Using Surface
Phonon Resonant Nanoparticles," Nano Letters 10, 373-379 (2010).
113. L. Zhu, A. P. Raman, and S. Fan, "Radiative cooling of solar absorbers using a visibly
transparent photonic crystal thermal blackbody," Proceedings of the National Academy of
Sciences 112, 12282 (2015).
114. M. M. Hossain, B. Jia, and M. Gu, "A Metamaterial Emitter for Highly Efficient Radiative
Cooling," Advanced Optical Materials 3, 1047-1051 (2015).
115. K. Sun, C. A. Riedel, Y. Wang, A. Urbani, M. Simeoni, S. Mengali, M. Zalkovskij, B.
Bilenberg, C. H. de Groot, and O. L. Muskens, "Metasurface Optical Solar Reflectors Using
AZO Transparent Conducting Oxides for Radiative Cooling of Spacecraft," ACS Photonics 5,
495-501 (2018).
116. W. Li and S. Fan, "Nanophotonic control of thermal radiation for energy applications
[Invited]," Opt. Express 26, 15995-16021 (2018).
117. M. Ono, K. Chen, W. Li, and S. Fan, "Self-adaptive radiative cooling based on phase
change materials," Opt. Express 26, A777-A787 (2018).
118. S.-H. Wu, M. Chen, M. T. Barako, V. Jankovic, P. W. C. Hon, L. A. Sweatlock, and M. L.
Povinelli, "Thermal homeostasis using microstructured phase-change materials," Optica 4,
1390-1396 (2017).
119. M. Chen, A. M. Morsy, and M. L. Povinelli, "Design of VO2-coated silicon microspheres
for thermally-regulating paint," Opt. Express 27, 21787-21793 (2019).
120. R. Osiander, J. Champion, A. Darrin, J. Sniegowski, S. Rodgers, D. Douglas, and T.
Swanson, "Micromachined louver arrays for spacecraft thermal control radiators," in 39th
Aerospace Sciences Meeting and Exhibit (American Institute of Aeronautics and Astronautics,
2001).
121. T. D. Swanson and G. C. Birur, "NASA thermal control technologies for robotic spacecraft,"
Applied Thermal Engineering 23, 1055-1065 (2003).
122. R. Osiander, S. L. Firebaugh, J. L. Champion, D. Farrar, and M. A. G. Darrin,
"Microelectromechanical devices for satellite thermal control," IEEE Sensors Journal 4, 525-
531 (2004).
123. H. Nagano, Y. Nagasaka, and A. Ohnishi, "Simple Deployable Radiator with Autonomous
Thermal Control Function," Journal of Thermophysics and Heat Transfer 20, 856-864 (2006).
124. G. Xu, C.-M. Huang, M. Tazawa, P. Jin, and D.-M. Chen, "Nano-Ag on vanadium dioxide.
II. Thermal tuning of surface plasmon resonance," Journal of Applied Physics 104, 053102
(2008).
125. M. A. Kats, R. Blanchard, S. Zhang, P. Genevet, C. Ko, S. Ramanathan, and F. Capasso,
"Vanadium Dioxide as a Natural Disordered Metamaterial: Perfect Thermal Emission and
Large Broadband Negative Differential Thermal Emittance," Physical Review X 3, 041004
(2013).
85

126. H. Wang, Y. Yang, and L. Wang, "Wavelength-tunable infrared metamaterial by tailoring
magnetic resonance condition with VO2 phase transition," Journal of Applied Physics 116,
123503 (2014).
127. H. Kim, K. Cheung, R. C. Y. Auyeung, D. E. Wilson, K. M. Charipar, A. Piqué, and N. A.
Charipar, "VO2-based switchable radiator for spacecraft thermal control," Scientific Reports
9, 11329 (2019).
128. K. Sun, C. A. Riedel, A. Urbani, M. Simeoni, S. Mengali, M. Zalkovskij, B. Bilenberg, C. H.
de Groot, and O. L. Muskens, "VO2 Thermochromic Metamaterial-Based Smart Optical Solar
Reflector," ACS Photonics 5, 2280-2286 (2018).
129. H. W. Verleur, A. S. Barker, and C. N. Berglund, "Optical Properties of VO2 between 0.25
and 5 eV," Physical Review 172, 788-798 (1968).
130. H. Kakiuchida, P. Jin, S. Nakao, and M. Tazawa, "Optical Properties of Vanadium Dioxide
Film during Semiconductive–Metallic Phase Transition," Japanese Journal of Applied Physics
46, L113-L116 (2007).
131. J. B. Kana Kana, J. M. Ndjaka, G. Vignaud, A. Gibaud, and M. Maaza, "Thermally tunable
optical constants of vanadium dioxide thin films measured by spectroscopic ellipsometry,"
Optics Communications 284, 807-812 (2011).
132. M. Currie, M. A. Mastro, and V. D. Wheeler, "Characterizing the tunable refractive index
of vanadium dioxide," Opt. Mater. Express 7, 1697-1707 (2017).
133. J. Sun and G. K. Pribil, "Analyzing optical properties of thin vanadium oxide films through
semiconductor-to-metal phase transition using spectroscopic ellipsometry," Applied Surface
Science 421, 819-823 (2017).
134. J. A. Ramirez-Rincon, C. L. Gomez-Heredia, A. Corvisier, J. Ordonez-Miranda, T.
Girardeau, F. Paumier, C. Champeaux, F. Dumas-Bouchiat, Y. Ezzahri, K. Joulain, O. Ares,
and J. J. Alvarado-Gil, "Thermal hysteresis measurement of the VO2 dielectric function for its
metal-insulator transition by visible-IR ellipsometry," Journal of Applied Physics 124, 195102
(2018).
135. Z. Z. Chenghao Wan, David Woolf, Colin M. Hessel, Jura Rensberg, Joel M. Hensley,
Yuzhe Xiao, Alireza Shahsafi, Jad Salman, Steffen Richter, Yifei Sun, M. Mumtaz Qazilbash,
Rüdiger Schmidt-Grund, Carsten Ronning, Shriram Ramanathan, Mikhail A. Kats, "Optical
properties of thin-film vanadium dioxide from the visible to the far infrared," arXiv:1901.02517
(2019).
136. M. Currie, M. A. Mastro, and V. D. Wheeler, "Atomic Layer Deposition of Vanadium
Dioxide and a Temperature-dependent Optical Model," J Vis Exp, 57103 (2018).
137. M. Li, X. Hu, Z. Ye, K.-M. Ho, J. Cao, and M. Miyawaki, "Higher-order incidence transfer
matrix method used in three-dimensional photonic crystal coupled-resonator array
simulation," Opt. Lett. 31, 3498-3500 (2006).
138. M. Li, Z. Y. Li, K.-M. Ho, J. R. Cao, and M. Miyawaki, "High-efficiency calculations for
three-dimensional photonic crystal cavities," Opt. Lett. 31, 262-264 (2006).
139. E. D. Palik, & Ghosh, G. , Handbook of optical constants of solids (Academic Press, San
Diego, 1998).
86

140. I. Takahashi, M. Hibino, and T. Kudo, "Thermochromic Properties of Double-Doped
VO2Thin Films Prepared by a Wet Coating Method Using Polyvanadate-Based Sols
Containing W and Mo or W and Ti," Japanese Journal of Applied Physics 40, 1391-1395
(2001).
141. R. Zhang, H.-B. Jin, D. Guo, J. Zhang, Z. Zhao, Y. Zhao, and J.-B. Li, "The role of Fe
dopants in phase stability and electric switching properties of Fe-doped VO2," Ceramics
International 42, 18764-18770 (2016).
142. X. Li, S. Zhang, L. Yang, X. Li, J. Chen, and C. Huang, "A convenient way to reduce the
hysteresis width of VO2(M) nanomaterials," New Journal of Chemistry 41, 15260-15267
(2017).
143. C. Batista, R. M. Ribeiro, and V. Teixeira, "Synthesis and characterization of VO2-based
thermochromic thin films for energy-efficient windows," Nanoscale Research Letters 6, 301
(2011).
144. K. Nishikawa, K. Yatsugi, Y. Kishida, and K. Ito, "Temperature-selective emitter," Applied
Physics Letters 114, 211104 (2019).
145. Z. Yang, C. Ko, and S. Ramanathan, "Metal-insulator transition characteristics of VO2 thin
films grown on Ge(100) single crystals," Journal of Applied Physics 108, 073708 (2010).
146. D. Fan, Q. Li, and P. Dai, "Temperature-dependent emissivity property in
La0.7Sr0.3MnO3 films," Acta Astronautica 121, 144-152 (2016).
147. J. B. Kana Kana, G. Vignaud, A. Gibaud, and M. Maaza, "Thermally driven sign switch of
static dielectric constant of VO2 thin film," Optical Materials 54, 165-169 (2016).

Abstract (if available)

Abstract This dissertation studies light interaction with matter in engineered microstructures for two applications: hybrid opto-thermal devices (Part I), and infrared thermal radiation management (Part II). ❧ The work on the opto-thermal devices focuses on the use of nano-patterned silicon to enhance light absorption and control heat on the micro-scale. We utilize critical coupling condition in silicon photonic crystal resonators to create highly absorptive peaks. This complements the efforts on plasmonic nanostructures as a primary tool for heating at the micro/ nanoscale. Beside the cost efficiency of using patterned silicon to control heat at microscale compared to plasmonic, we also present a scheme where thermo-optic nonlinearity in silicon can be utilized to achieve programmable on-chip heating. Moreover, this scheme is used to create a high temperature thermal memory that uses laser to read, write and erase digital information, and can potentially operate in harsh, high temperature environments. ❧ The work on infrared thermal radiation management focuses on developing schemes that use photonic microstructures to shape thermal emission. We experimentally realize a proof-of-concept broadband thermal emissivity switching

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Resonant light-matter interactions in nanophotonic structures: for manipulating optical forces and thermal emission

PDF

Optimized nanophotonic designs for thermal emission control

PDF

Light matter interactions in engineered structures: radiative thermal management & light-assisted assembly of reconfigurable optical matter

PDF

Designable nonlinear power shaping photonic surfaces

PDF

Integrated photonics assisted electron emission devices

PDF

Nonlinear optical nanomaterials in integrated photonic devices

PDF

Light management in nanostructures: nanowire solar cells and optical epitaxial growth

PDF

Microphotonics for tailored infrared emission and secure data encryption

PDF

Application of optical forces in microphotonic systems

PDF

In-situ characterization of nanoscale opto-electronic devices through optical spectroscopy and electron microscopy

PDF

Perovskite chalcogenides: emerging semiconductors for visible to infrared optoelectronics

PDF

Tailoring asymmetry in light-matter interaction with atomic-scale lattice modulations

PDF

Phase change heterostructures for electronic and photonic applications

PDF

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

Asset Metadata

Creator Morsy, Ahmed Moustafa (author)

Core Title Light-matter interactions in engineered microstructures: hybrid opto-thermal devices and infrared thermal emission control

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Electrical Engineering

Publication Date 05/03/2020

Defense Date 05/01/2020

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

Tag cell trapping,coupled mode theory,coupled resonators,infrared emissivity,metamaterial,nanophotonics,OAI-PMH Harvest,optical heating,photonic crystals,thermal homeostasis,thermal memory,thermo-optic nonlinearity

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Povinelli, Michelle Lynn (committee chair), Ravichandran, Jayakanth (committee member), Wu, Wei (committee member)

Creator Email morsy@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-295070

Unique identifier UC11665969

Identifier etd-MorsyAhmed-8398.pdf (filename),usctheses-c89-295070 (legacy record id)

Legacy Identifier etd-MorsyAhmed-8398.pdf

Dmrecord 295070

Document Type Dissertation

Rights Morsy, Ahmed Moustafa

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 a...

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