Electro-Optically Tunable Metasurfaces for a Comprehensive Control of Properties of Light - CaltechTHESIS

Electro-Optically Tunable Metasurfaces for a Comprehensive Control of Properties of Light - CaltechTHESIS
CaltechTHESIS
A Caltech Library Service
About
Browse
Deposit an Item
Instructions for Students
Electro-Optically Tunable Metasurfaces for a Comprehensive Control of Properties of Light
Citation
Kafaie Shirmanesh, Ghazaleh
(2021)
Electro-Optically Tunable Metasurfaces for a Comprehensive Control of Properties of Light.
Dissertation (Ph.D.), California Institute of Technology.
doi:10.7907/m554-as73.
Abstract
The ability to control electromagnetic wavefront is a central key in optics. Conventional optical components rely on the gradual accumulation of the phase of light as it passes through an optical medium. However, since the accumulated phase is limited by the permittivity of naturally existing materials, such a mechanism often results in bulky devices that are much thicker than the operating wavelength.
During the last several years, metasurfaces (quasi-2D nanophotonic structures) have attracted a great deal of attention owing to their promise to manipulate constitutive properties of electromagnetic waves such as amplitude, phase, and polarization. Metasurfaces are ultrathin arrays of subwavelength resonators, called meta-atoms, where each meta-atom imposes a predefined change on the properties of the scattered light. By precisely designing the optical response of these meta-atoms to an incident wave, metasurfaces can introduce abrupt changes to the properties of the transmitted, reflected, or scattered light, and hence, can flexibly shape the out-going wavefront at a subwavelength scale. This enables metasurfaces to replace conventional bulky optical components such as prisms or lenses by their flat, low-profile analogs. Furthermore, a single metasurface can perform optical functions typically attained by using a combination of multiple bulky optical elements, offering tremendous opportunities for flat optics.
The optical response of a metasurface is typically dictated by the geometrical parameters of the subwavelength scatterers. As a result, most of the reported metasurfaces have been passive, namely have functions that are entirely fixed at the time of fabrication. By making the metasurfaces reconfigurable in their phase, amplitude, and polarization response, one can achieve real-time control of optical functions, and indeed, achieve multi-functional characteristics after fabrication. Dynamical control of the properties of the scattered light is possible by using external stimuli such as electrical biasing, optical pumping, heating, or elastic strain that can give rise to changes in the dielectric function or physical dimensions of the metasurface elements.
In this dissertation, we present the opportunities and challenges towards achieving reconfigurable metasurfaces. We introduce a paradigm of active metasurfaces for real-time control of the wavefront of light at a subwavelength scale by investigating different modulation mechanisms and possible metasurface designs and material platforms that let us effectively employ the desired modulation mechanism. We will present multiple electro-optically tunable metasurface platforms. These electronically-tunable schemes are of great interest owing to their robustness, high energy-efficiency, and reproducibility. We will also show the design and experimental demonstration of active metasurfaces for which the tunable optical response can be tailored in a pixel-by-pixel configuration.
The ability to individually control the optical response of metasurface elements has made active optical metasurfaces to be progressively ubiquitous by enabling a wide range of optical functions such as dynamic holography, light fidelity (Li-Fi), focusing, and beam steering. As a result, reconfigurable metasurfaces can hold an extraordinary promise for optical component miniaturization and on-chip photonic integration. Such compact and high-performance devices with reduced size, weight, and power (SWaP) can be used in future free-space optical communications or light detection and ranging (LiDAR) systems.
Item Type:
Thesis (Dissertation (Ph.D.))
Subject Keywords:
Active metasurfaces, Optics, Nanophotonics, Plasmonics, Indium tin oxide
Degree Grantor:
California Institute of Technology
Division:
Engineering and Applied Science
Major Option:
Applied Physics
Thesis Availability:
Public (worldwide access)
Research Advisor(s):
Atwater, Harry Albert
Group:
Kavli Nanoscience Institute
Thesis Committee:
Vahala, Kerry J. (chair)
Atwater, Harry Albert
Scherer, Axel
Faraon, Andrei
Defense Date:
24 August 2020
Non-Caltech Author Email:
ghazal90kafaie (AT) gmail.com
Record Number:
CaltechTHESIS:09172020-190836007
Persistent URL:
DOI:
10.7907/m554-as73
Related URLs:
URL
URL Type
Description
DOI
Article adapted for Ch. 2
DOI
Article adapted for Ch. 3
DOI
Article adapted for Ch. 5
DOI
Article adapted for Ch. 6
DOI
Article adapted for Ch. 6
ORCID:
Author
ORCID
Kafaie Shirmanesh, Ghazaleh
0000-0003-1666-3215
Default Usage Policy:
No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:
13955
Collection:
CaltechTHESIS
Deposited By:
Ghazaleh Kafaie Shirmanesh
Deposited On:
21 Sep 2020 15:38
Last Modified:
08 Nov 2023 00:12
Thesis Files
Preview
PDF (Thesis)
- Final Version
See Usage Policy.
15MB
Repository Staff Only:
item control page
CaltechTHESIS is powered by
EPrints 3.3
which is developed by the
School of Electronics and Computer Science
at the University of Southampton.
More information and software credits
Electro-optically Tunable Metasurfaces for a
Comprehensive Control of Properties of Light

Thesis by

Ghazaleh Kafaie Shirmanesh

In Partial Fulfillment of the Requirements for
the degree of
Doctor of Philosophy

CALIFORNIA INSTITUTE OF TECHNOLOGY
Pasadena, California

2020
Defended August 24, 2020

ii

 2020
Ghazaleh Kafaie Shirmanesh
ORCID: 0000-0003-1666-3215

iii

ACKNOWLEDGEMENTS
Being a graduate student at Caltech has been an incredible journey that provided me with
everything that I would have asked for. This great experience would not have come
together without the support and input of a phenomenal group of people.
First and foremost, I thank Professor Harry A. Atwater for being a truly unique advisor. I
am incredibly grateful for the freedom he has given me to explore what I liked, and for
teaching me how to develop my self-confidence. I admire him for being keenly intelligent,
insatiably curious, and tirelessly enthusiastic for science. His passion and positive attitude
have been great inspirations to help me push forward when facing hardships and failures
in my experiments. I express my tremendous gratitude for what I have learned from him
both scientifically and personally. I will always be deeply impressed by the nearly
impossible high bar he has set for me, and also for his unwavering support through these
years which encouraged me to become the best version of myself.
I have been very fortunate to be surrounded by an amazing group of individuals in the
Atwater group. I count myself incredibly lucky to have Dr. Ruzan Sokhoyan as a mentor
and a role model since I first joined Caltech. I thank Ruzan for her exceptionally invaluable
guidance. She has been instrumental in all of the work I have done in my doctoral studies,
and this thesis would have not been possible without her guidance and collaboration. Far
beyond being an inspiring scientist and a helpful colleague, I admire the “Boss Lady” for
being a supportive friend and a wonderful source of optimism.
Professor Pin Chieh Wu has been a great colleague in many of my projects, who fascinated
me with his insight and knowledge. I would like to thank him for his assistance throughout
my research, and for his advice and exchange of ideas. I also thank Professor Yu-Jung Lu
(Yuri, the Beautiful) and Wen-Hui (Sophia) Cheng for being exceptionally bright friends
and colleagues with whom I have always had many enjoyable conversations and fun times.
The work in this thesis would not have been possible without my collaborations with the
talented members of the “Metasurface Team” in the Atwater group. I am thankful to all of
them for generating this positive, friendly, and collaborative atmosphere. Especially I
would like to thank Dr. Ragip Pala and Professor Muhammad Alam who have always been

iv
very open, helpful, collaborative, and supportive. I also learned a lot in my discussions
with Dr. Yao-Wei Huang, Dr. Meir Grajower, and Professor Artur Davoyan for which I
express my gratitude. I am thankful to Prachi Thureja and Yonghwi Kim for our
collaborations and invaluable friendship. I am grateful to our collaborators at Samsung
Electronics for supporting my research. Especially, I would like to thank Dr. Duhyun Lee,
Dr. Junghyun Park, and Dr. Seunghoon Han for helpful discussions.
Besides, I thank a large number of teachers and advisors who developed my enthusiasm
for science and I will always be tremendously in debt to all of them for what they have
taught me. I would like to sincerely appreciate my Ph.D. thesis committee members for
their time and helpful pieces of advice. I thank Professor Axel Scherer for his ceaseless
support throughout my Ph.D. I am grateful to Professor Andrei Faraon for what he has
taught me on metasurfaces and for the helpful discussions during our Samsung California
Metaphotonics Cluster review meetings. I express my appreciation to Professor Kerry
Vahala for the insightful guidance and feedback on my Ph.D. projects. I would also like to
thank Professor Oscar Painter for serving on my candidacy exam committee.
During my Ph.D., I have been very fortunate to collaborate with smart people in the field
of photonics. I acknowledge the great help and support I have received from my
collaborators outside Caltech including Professor Hossein Mosallaei, Thej Tumkur
Umanath, Eyal Feigenbaum, Selim Elhadj, and Alejandro Ceballos-Sanchez.
So much appreciation additionally needs to be given to the incredible administrators and
amazing Caltech staff who have been super helpful and friendly, holding everything
together. I thank Jennifer Blankenship, Jonathan Gross, Liz Hormigoso, Lyann Lau, Mabel
Chik, Christy Salinas, Natalie Gilmore, Tess Legaspi, and Connie Rodriguez. I have to
especially thank Christine Jenstad for making it so easy to embrace this place as my home.
I am sincerely grateful to Kam Flower. Not only was she a wonderful admin, but she has
also been an amazing friend, who was always there whenever I needed help with anything.
Moreover, the work I have done during my Ph.D. would have been possible without the
support and training from the Kavli Nanoscience Institute (KNI) at Caltech. So, I would
like to thank the current and past KNI staff: Dr. Guy DeRose, Bert Mendoza, Dr. Matthew

Sullivan Hunt, Alex Wertheim, Nathan Lee, Melissa Melendes, Tiffany Kimoto, and
Jennifer Palmer.
I also thank Rebecca Glaudell for her tireless work for keeping the Atwater labs as safe as
possible.
I am grateful to my officemates Hamid, Pankaj, Arun, and Jeremy, who made my time at
work particularly enjoyable. I especially thank Professor Ognjen Ilic for many empowering
conversations, and for his advice and encouragement. I also thank Georgia for the great
times in the office and for her open-hearted friendship.
Most importantly, I would like to thank my lab-mate, friend, and husband who has always
played an irreplaceable role in all I have achieved so far. He has always been understanding
and supportive of me and my work. Amir, thank you for always being there with words of
love and encouragement, for your patience and accompanying me in all my over-night tool
runs, for brightening my day each day, and for making life so wonderful. Words cannot
express my gratitude for all your help and support, and this would have never happened
without you.
I have been extremely fortunate to be given a family of great kindness and generosity, and
none of what I have done so far would have been possible without them. They patiently
provided me with crucial support during all these years during which I was not able to visit
them. I thank my father, Amir, for his unwavering trust in me, for motivating my curiosity,
and for being my source of guidance and courage in difficult times. I thank my mother,
Fariba, who has been my mental anchor and has been always trying to ensure that we stay
healthy. I also thank my siblings Kimia, Ramin, and Ghazal for their support and love, and
for all the fun times we had while talking over the phone during these years. I have also
been lucky to have some amazing friends who have always been a source of joy and energy
for me. I especially thank Elahe, Hamid, and Sana for their compassion and friendship.
Ghazaleh Kafaie Shirmanesh
July 2020
Pasadena, CA

vi

ABSTRACT

The ability to control electromagnetic wavefront is a central key in optics. Conventional
optical components rely on the gradual accumulation of the phase of light as it passes
through an optical medium. However, since the accumulated phase is limited by the
permittivity of naturally existing materials, such a mechanism often results in bulky devices
that are much thicker than the operating wavelength.
During the last several years, metasurfaces (quasi-2D nanophotonic structures) have
attracted a great deal of attention owing to their promise to manipulate constitutive
properties of electromagnetic waves such as amplitude, phase, and polarization.
Metasurfaces are ultrathin arrays of subwavelength resonators, called meta-atoms, where
each meta-atom imposes a predefined change on the properties of the scattered light. By
precisely designing the optical response of these meta-atoms to an incident wave,
metasurfaces can introduce abrupt changes to the properties of the transmitted, reflected,
or scattered light, and hence, can flexibly shape the out-going wavefront at a
subwavelength scale. This enables metasurfaces to replace conventional bulky optical
components such as prisms or lenses by their flat, low-profile analogs. Furthermore, a
single metasurface can perform optical functions typically attained by using a combination
of multiple bulky optical elements, offering tremendous opportunities for flat optics.
The optical response of a metasurface is typically dictated by the geometrical parameters
of the subwavelength scatterers. As a result, most of the reported metasurfaces have been
passive, namely have functions that are entirely fixed at the time of fabrication. By making
the metasurfaces reconfigurable in their phase, amplitude, and polarization response, one
can achieve real-time control of optical functions, and indeed, achieve multi-functional
characteristics after fabrication. Dynamical control of the properties of the scattered light
is possible by using external stimuli such as electrical biasing, optical pumping, heating,
or elastic strain that can give rise to changes in the dielectric function or physical
dimensions of the metasurface elements.

vii
In this dissertation, we present the opportunities and challenges towards achieving
reconfigurable metasurfaces. We introduce a paradigm of active metasurfaces for real-time
control of the wavefront of light at a subwavelength scale by investigating different
modulation mechanisms and possible metasurface designs and material platforms that let
us effectively employ the desired modulation mechanism. We will present multiple electrooptically tunable metasurface platforms. These electronically-tunable schemes are of great
interest owing to their robustness, high energy-efficiency, and reproducibility. We will also
show the design and experimental demonstration of active metasurfaces for which the
tunable optical response can be tailored in a pixel-by-pixel configuration.
The ability to individually control the optical response of metasurface elements has made
active optical metasurfaces to be progressively ubiquitous by enabling a wide range of
optical functions such as dynamic holography, light fidelity (Li-Fi), focusing, and beam
steering. As a result, reconfigurable metasurfaces can hold an extraordinary promise for
optical component miniaturization and on-chip photonic integration. Such compact and
high-performance devices with reduced size, weight, and power (SWaP) can be used in
future free-space optical communications or light detection and ranging (LiDAR) systems.

viii

PUBLISHED CONTENT AND CONTRIBUTIONS
[1] G. Kafaie Shirmanesh, R. Sokhoyan, R. A. Pala, and H. A. Atwater, “Dual-Gated
Active Metasurface at 1550 nm with Wide (>300°) Phase Tunability,” Nano Lett.,
vol. 18, no. 5, pp. 2957–2963, May 2018, doi: 10.1021/acs.nanolett.8b00351.
G.K.S contributed to conceiving the original idea of dual-gated active metasurfaces,
performed the numerical design, device fabrication, performed the optical
measurements, analyzed numerical and experimental data, built the optical setup
for measurement, and wrote the manuscript.
[2] G. Kafaie Shirmanesh, R. Sokhoyan, P. C. Wu, and H. A. Atwater, “Electrooptically Tunable Multifunctional Metasurfaces,” ACS Nano, vol. 14, no. 6, pp.
6912–6920, Jun. 2020, doi: 10.1021/acsnano.0c01269.
G.K.S contributed to conceiving the original idea, performed the numerical design,
device fabrication, performed the optical measurements, analyzed numerical and
experimental data, designed and built up the PCBs for individual electrical control
of metasurface elements, helped with the build-up of the optical setup for
measurement, and wrote the manuscript.
[3] P. C. Wu, R. Sokhoyan, G. Kafaie Shirmanesh, W.-H. Cheng, and H. A. Atwater,
“Tunable Metasurface for Dynamic Control of Polarization in Near-Infrared,” in
preparation (2020).
G.K.S participated in the conception of the project, contributed to the numerical
design, helped with device fabrication, and participated in the writing of the
manuscript.
[4] G. Kafaie Shirmanesh*, R. Sokhoyan*, and H. A. Atwater, “Modulation of
Spontaneous Emission of Quantum Emitters by Active Metasurfaces,” in
preparation (2020). (*Equal contributors).
G.K.S performed the numerical design and optical simulations, analyzed numerical
data, and wrote the manuscript.
[5] P. C. Wu, R. A. Pala, G. Kafaie Shirmanesh, W.-H. Cheng, R. Sokhoyan, M.
Grajower, M. Z. Alam, D. Lee, H. A. Atwater, “Dynamic Beam Steering with AllDielectric Electro-Optic III–V Multiple-Quantum-Well Metasurfaces,” Nat.
Commun., vol. 10, no. 1, p. 3654, 2019, doi: 10.1038/s41467-019-11598-8.
G.K.S contributed to device fabrication, designed and built up the PCB for
individual electrical control of metasurface elements, and participated in the
writing of the manuscript.

ix
[6] A. Forouzmand, M. M. Salary, G. Kafaie Shirmanesh, R. Sokhoyan, H. A.
Atwater, and H. Mosallaei, “Tunable All-Dielectric Metasurface for Phase
Modulation of the Reflected and Transmitted Light via Permittivity Tuning of
Indium Tin Oxide,” Nanophotonics, vol. 8, no. 3, pp. 415–427, doi:
10.1515/nanoph-2018-0176.
G.K.S participated in the conception of the project, contributed to the numerical
design, and participated in the writing of the manuscript.
[7] Y.-J. Lu, R. Sokhoyan, W.-H. Cheng, G. Kafaie Shirmanesh, A. R. Davoyan, R.
A Pala, K. Thyagarajan, H. A. Atwater, “Dynamically Controlled Purcell
Enhancement of Visible Spontaneous Emission in a Gated Plasmonic
Heterostructure,” Nat. Commun., vol. 8, no. 1, p. 1631, 2017, doi: 10.1038/s41467017-01870-0.
G.K.S. performed material characterization and contributed to writing the
manuscript.

OTHER PUBLICATIONS

[1] G. Kafaie Shirmanesh, R. Sokhoyan, P. C. Wu, and H. A. Atwater, “Electrooptically Tunable Multifunctional Metasurfaces,” arXiv preprint arXiv:1910.02069.
[2] P. Thureja, G. Kafaie Shirmanesh, K. T. Fountaine, R. Sokhoyan, M. Grajower,
and H. A. Atwater, “Array-Level Inverse Design of Beam Steering Active
Metasurfaces,” submitted to ACS Nano (2020).
[3] Y. Kim, P. C. Wu, R. Sokhoyan, K. Mauser, R. Glaudell, G. Kafaie Shirmanesh,
H. A. Atwater, “Phase modulation with electrically tunable vanadium dioxide
phase-change metasurfaces,” Nano Lett. 19 (6), 3961-3968 (2019).
[4] T. U. Tumkur, R. Sokhoyan, M. Su, A. Ceballos-Sanchez, G. Kafaie Shirmanesh,
Y. Kim, H. A. Atwater, E. Feigenbaum, S. Elhadj, “Toward High Laser Power
Beam Manipulation with Nanophotonic Materials: Evaluating Thin Film Damage
Performance,” in preparation (2020).
[5] R. Sokhoyan, G. Kafaie Shirmanesh, Y. -J. Lu, K. Thyagarajan, R. A. Pala, H. A.
Atwater “Tunable Optical Response and Purcell Enhancement of Gated Plasmonic
Structures,” 2017 International Conference on Optical MEMS and Nanophotonics
(OMN), 1-2.
[6] M. Z. Alam, H. W. Lee, Y. W. Huang, R. A. Pala, K. Thyagarajan, G. Kafaie
Shirmanesh, R. Sokhoyan, H. A. Atwater, “Plasmonic Nanophotonic Modulators,”
2017 IEEE Photonics Society Summer Topical Meeting Series (SUM), 193-194.

11

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ......................................................................................... iii
ABSTRACT ..................................................................................................................vi
PUBLISHED CONTENT AND CONTRIBUTIONS ............................................... viii
OTHER PUBLICATIONS ............................................................................................ x
TABLE OF CONTENTS ............................................................................................. 11
LIST OF FIGURES ..................................................................................................... 14
LIST OF TABLES ....................................................................................................... 19
ABBREVIATIONS ..................................................................................................... 20
INTRODUCTION ....................................................................................................... 22
1.1. METASURFACES: MOTIVATIONS AND APPLICATIONS .................................................. 22
1.2. ACTIVE METASURFACES............................................................................................... 26
1.3. THESIS OUTLINE ........................................................................................................... 30

DUAL-GATED ACTIVE METASURFACE FOR WIDE PHASE TUNABILITY .. 32
2.1. ACTIVE METASURFACES FOR MODULATION OF PROPERTIES OF LIGHT ...................... 32
2.2. DESIGN OF DUAL-GATED METASURFACES ................................................................... 34
2.2.1. Electrostatic Simulations to Extract ITO Properties .......................................................... 35
2.2.2. Choice of Plasmonic Metal ................................................................................................. 36
2.2.3. Choice of Gate Dielectric ................................................................................................... 37

2.3. ELECTROMAGNETIC SIMULATIONS OF DUAL-GATED METASURFACES ....................... 37
2.3.1 Modelling the ITO Layer in the Dual-Gated Metasurface Structure ................................... 37
2.3.2. Simulated Amplitude and Phase Response of the Dual-Gated Metasurface ....................... 40
2.3.3. Distribution of Electromagnetic Fields in the Dual-Gated Metasurface ............................ 43

2.4. FABRICATION OF DUAL-GATED TUNABLE METASURFACE .......................................... 45
2.5. EXPERIMENTAL DEMONSTRATION OF THE DUAL-GATED METASURFACE .................... 47
2.5.1. Reflectance measurements .................................................................................................. 47
2.5.2. Phase measurements ........................................................................................................... 49

2.6. THEORETICAL DEMONSTRATION OF BEAM STEERING USING THE DUAL-GATED
METASURFACES ................................................................................................................... 53
2.7. COMPARISON TO SINGLE-GATED TUNABLE METASURFACE ........................................ 57
2.8. CONCLUSIONS AND OUTLOOK ...................................................................................... 58

ELECTRO-OPTICALLY TUNABLE MULTIFUNCTIONAL METASURFACEs . 59
3.1. INTRODUCTION ............................................................................................................. 59
3.2. DESIGN OF ELECTRO-OPTICALLY TUNABLE METASURFACE ELEMENT....................... 63
3.3. OPTICAL MODULATION IN ELECTRO-OPTICALLY TUNABLE METASURFACE ELEMENT
............................................................................................................................................. 66
3.3.1. Calculated Optical Response of the Electro-Optically Tunable Metasurfaces ................... 66
3.3.2. Measured Optical Response of the Electro-Optically Tunable Metasurfaces .................... 70

3.4. MULTIFUNCTIONAL PERFORMANCE OF THE ELECTRO-OPTICALLY UNIVERSAL
METASURFACE .................................................................................................................... 75
3.4.1. Demonstration of Beam Steering Using the Multifunctional Metasurface ......................... 76
3.4.2. Demonstration of Dynamic Focusing Meta-Mirror ............................................................ 84

3.5. CONCLUSIONS AND OUTLOOK ...................................................................................... 89

ELECTRO-OPTICALLY TUNABLE METASURFACES FOR DYNAMIC
POLARIZATION CONTROL .................................................................................... 90

12
4.1. INTRODUCTION ............................................................................................................. 90
4.2. TUNABLE POLARIZATION CONVERSION METASURFACE DESIGN ................................ 92
4.3. OPTICAL RESPONSE OF THE POLARIZATION CONVERSION METASURFACE ................. 93
4.4. DYNAMIC MODULATION OF THE POLARIZATION STATE OF THE REFLECTED BEAM BY
USING TUNABLE POLARIZATION CONVERSION METASURFACE ......................................... 98
4.5. EXPERIMENTAL DEMONSTRATION OF TUNABLE POLARIZATION CONVERSION
METASURFACE .................................................................................................................. 102
4.6. CONCLUSIONS AND OUTLOOK .................................................................................... 106

MODULATION OF SPONTANEOUS EMISSION OF QUANTUM EMITTERS BY
ACTIVE METASURFACES .................................................................................... 108
5.1. PURCELL ENHANCEMENT OF SPONTANEOUS EMISSION FROM QUANTUM EMITTERS 108
5.2. RECONFIGURABLE PURCELL ENHANCEMENT OF SPONTANEOUS EMISSION BY
METASURFACES ................................................................................................................. 110
5.3. COUPLING OF QUANTUM EMITTERS TO THE PLASMONIC ACTIVE METASURFACE .... 116
5.3.1. Effect of ITO and Al2O3 Thickness on the Purcell Enhancement ...................................... 116
5.3.2. Effect of Antenna and Electrode Thickness on the Purcell Enhancement ......................... 118
5.3.3. Effect of Antenna Length, Antenna Width, and Electrode Width on the Purcell Enhancement
.................................................................................................................................................... 119
5.3.4. Effect of Dipole Position on the Purcell Enhancement ..................................................... 120

5.4. BIAS-INDUCED MODULATION OF SPONTANEOUS EMISSION DECAY RATE ................ 121
5.5. CONCLUSIONS AND OUTLOOK .................................................................................... 127

DIELECTRIC TUNABLE METASURFACES ........................................................ 128
6.1. DIELECTRIC TUNABLE METASURFACES ..................................................................... 129
6.2. SI-BASED DIELECTRIC ACTIVE METASURFACES ........................................................ 132
6.2.1. Geometry and Structural Parameters of the Si-based Dielectric Active Metasurfaces .... 132
6.2.2. Optical Response of the Si-based Dielectric Active Metasurfaces .................................... 135

6.3. III-V ALL-DIELECTRIC ACTIVE METASURFACES ....................................................... 139
6.3.1. Characterization of MQW Wafers .................................................................................... 139
6.3.2. Design and Simulation of All-dielectric MQW Metasurface............................................. 143
6.3.3. Fabrication and Measurement of All-dielectric MQW Metasurface ................................ 145
6.3.4. Demonstration of Electrical Beam Switching and Beam Steering with the All-dielectric
MQW Metasurface ...................................................................................................................... 149

6.4. CONCLUSIONS AND OUTLOOK .................................................................................... 154

CONCLUSION AND OUTLOOK ............................................................................ 157
BIBLIOGRAPHY ...................................................................................................... 162
APPENDIX ................................................................................................................ 178
APPENDIX A.1. EFFECT OF THE THICKNESS OF THE ITO LAYER ON THE DEVICE TUNABILITY
........................................................................................................................................... 178
APPENDIX A.2. FABRICATION AND CHARACTERIZATION OF HAOL................................. 179
APPENDIX A.3. ALTERNATIVE METHOD FOR DEPOSITION OF AL BACK REFLECTOR ....... 180
APPENDIX A.4. FABRICATION AND CHARACTERIZATION OF ITO ..................................... 181
APPENDIX B.1. COMPARISON TO THE PREVIOUSLY PROPOSED DESIGN ........................... 184
APPENDIX B.2. CHOOSING THE NUMBER OF METASURFACE PIXELS ................................ 185
APPENDIX B.3. CHANGING METASURFACE REFLECTANCE LEVEL ................................... 185
APPENDIX B.4. PCB LAYOUT FOR DEMONSTRATION OF DYNAMIC BEAM STEERING AND
RECONFIGURABLE FOCUSING USING THE MULTIFUNCTIONAL METASURFACE. .............. 193
APPENDIX B.5. PATTERN LAYOUTS FOR FABRICATION OF THE MULTIFUNCTIONAL
METASURFACE .................................................................................................................. 195
APPENDIX C.1. MEASURING THE BREAKDOWN FIELD OF HAOL FILMS .......................... 200
APPENDIX D.1. CONVERGENCE TEST FOR SIMULATION REGION SIZE .............................. 201

13
APPENDIX D.2. EXPERIMENTAL DEMONSTRATION OF TUNABLE PURCELL ENHANCEMENT
OF SPONTANEOUS EMISSION USING ACTIVE METASURFACES .......................................... 201
APPENDIX E.1. PCB DESIGN FOR DEMONSTRATION OF DYNAMIC BEAM STEERING ....... 207

14

LIST OF FIGURES
Figure 1.1: Schematic illustration of Snell’s law. ........................................................ 23
Figure 1.2: A review of the recent passive metasurfaces. ............................................ 25
Figure 1.3: A review of the recent advances in active metasurfaces. .......................... 29
Figure 2.1: Dual-gated active metasurface. ................................................................. 34
Figure 2.2: Configuration of bias applied to the nanoantennas in the dual-gated active
metasurface. ................................................................................................................. 38
Figure 2.3: Illustration of Case I and Case II for the bias applied to the dual-gated active
metasurface. ................................................................................................................. 39
Figure 2.4: Simulated amplitude response of the dual gated metasurface for Case I and
Case II. ......................................................................................................................... 41
Figure 2.5: Simulated phase response of the dual gated metasurface for Case I and Case
II. .................................................................................................................................. 42
Figure 2.6: Simulated reflectance and phase shift of the dual-gated metasurface as a
function of applied voltage V0 at the wavelength of λ = 1550 nm. .............................. 42
Figure 2.7: Spatial distribution of the electric field intensity within the dual-gated
metasurface for Case I and Case II. ............................................................................. 43
Figure 2.8: Spatial distribution of the z component of the electric field within the dualgated metasurface for Case I and Case II. .................................................................... 44
Figure 2.9: Spatial distribution of the magnetic field intensity within the dual-gated
metasurface for Case I and Case II. ............................................................................. 45
Figure 2.10: Schematic representation of fabrication steps of the dual-gated tunable
metasurface. ................................................................................................................. 46
Figure 2.11: Optical setup for measuring the reflectance spectrum of our dual-gated
metasurfaces. ................................................................................................................ 47
Figure 2.12: Measured amplitude response of the dual gated metasurface for Case I.49
Figure 2.13: Optical setup used for measuring the phase shift of the light reflected from
the tunable dual-gated metasurfaces. ........................................................................... 50
Figure 2.14: The Michelson interferometer-type measurement results used to extract
the phase shift of the light reflected from the dual-gated metasurface when changing
the applied bias. ........................................................................................................... 51
Figure 2.15: Measured phase response of the dual-gated metasurface for Case I. ...... 52
Figure 2.16: Calculated beam steering results using the dual-gated tunable metasurface.
...................................................................................................................................... 55
Figure 2.17: Calculated steering angles using a blazed grating approach. .................. 56
Figure 2.18: Comparison between the dual-gated metasurface and the single-gated
counterpart. .................................................................................................................. 57
Figure 3.1: Schematic demonstration of the multifunctional metasurface with 96
independently addressable metasurface elements. ....................................................... 63
Figure 3.2: Unit cell design of the multifunctional metasurface. ................................ 64
Figure 3.3: Voltage-dependent spatial distribution of the ITO properties. .................. 65
Figure 3.4: Calculated amplitude/phase modulation provided by the multifunctional
metasurface. ................................................................................................................. 67
Figure 3.5: Selecting the operating wavelength of the multifunctional metasurface. . 67
Figure 3.6: Spatial electric field intensity distribution of the multifunctional
metasurface. ................................................................................................................. 68
Figure 3.7: Spatial distribution of the magnetic field intensity in the multifunctional
metasurface. ................................................................................................................. 69

15
Figure 3.8: Spatial distribution of the z-component of the electric field in the
multifunctional metasurface......................................................................................... 69
Figure 3.9: Fabrication steps of the multifunctional metasurface................................ 70
Figure 3.10: Universal measurement setup used for reflectance, phase shift, beam
steering, and focusing measurements. ......................................................................... 71
Figure 3.11: Optical setup used for reflectance measurement. .................................... 72
Figure 3.12: Optical setup used for phase shift measurement. .................................... 73
Figure 3.13: Measured tunable optical response of the multifunctional metasurface. 74
Figure 3.14: Fabrication and measurement of a multifunctional metasurface. ............ 75
Figure 3.15: Analytical and full-wave simulation of the beam steering metasurface. 77
Figure 3.16: Effect of amplitude and phase modulation on the far-field radiation pattern
of the beam steering metasurface. ................................................................................ 78
Figure 3.17: A supercell for a steering angle of 14.6 with different ideal, linear, and
stair-step phase profiles................................................................................................ 80
Figure 3.18: Simulation results of the beam steering metasurface obtained when
unequal reflectance values and linear phase distributions are assumed for all
metasurface pixels. ....................................................................................................... 81
Figure 3.19: Optical setup used for beam steering measurement. ............................... 82
Figure 3.20: Demonstration of dynamic beam steering by the multifunctional
metasurface. ................................................................................................................. 83
Figure 3.21: Theoretical demonstration of a dynamic focusing meta-mirror. ............. 84
Figure 3.22: Optical setup used for focusing performance measurement. .................. 85
Figure 3.23: Experimental demonstration of a dynamic focusing meta-mirror with short
focal length................................................................................................................... 86
Figure 3.24: Experimental demonstration of a dynamic focusing meta-mirror with long
focal length................................................................................................................... 88
Figure 3.25: Possibility of demonstration of focusing meta-mirror with extended focal
lengths using the multifunctional metasurface. ........................................................... 88
Figure 4.1: Schematic illustration of the active polarization conversion metasurface
operating in reflection mode. ....................................................................................... 93
Figure 4.2: Metasurface design principle for demonstration of tunable polarization
conversion. ................................................................................................................... 94
Figure 4.3: Modulation of ITO properties under an applied bias. ............................... 95
Figure 4.4: Amplitude modulation provided by the tunable polarization conversion
metasurface. ................................................................................................................. 96
Figure 4.5: Phase modulation provided by the tunable polarization conversion
metasurface. ................................................................................................................. 97
Figure 4.6: Spatial distribution of the electromagnetic fields in the tunable polarization
conversion metasurface under an applied bias............................................................. 97
Figure 4.7: Calculated polarization conversion performance of the Al-based tunable
metasurface. ................................................................................................................. 99
Figure 4.8: Realization of different polarization states when changing the operating
wavelength. ................................................................................................................ 101
Figure 4.9: Calculated polarization conversion performance of the Au-based tunable
metasurface. ............................................................................................................... 102
Figure 4.10: Fabrication steps of the Au-based tunable polarization conversion
metasurface. ............................................................................................................... 103
Figure 4.11: Measured reflectance spectra of the linearly-polarized reflected beams.
.................................................................................................................................... 104

16
Figure 4.12: Measured reflectance spectra of the circularly-polarized reflected beams.
.................................................................................................................................... 105
Figure 4.13: Measured reflectance spectra of the linearly-polarized reflected beams
along 45 and −45 axes. ........................................................................................... 105
Figure 4.14: Measured Stokes parameters of the tunable polarization conversion
metasurface. ............................................................................................................... 106
Figure 5.1: Schematic illustration of Fermi’s golden rule. ........................................ 111
Figure 5.2: TiN/SiO2/Ag plasmonic heterostructure used for active control of
spontaneous emission of QDs. ................................................................................... 113
Figure 5.3: Field-effect modulation of charge carrier concentration density and
permittivity of ITO. .................................................................................................... 115
Figure 5.4: Schematic of the gate-tunable metasurface used for active control of
spontaneous emission decay rate of quantum emitters via modulation of the local
density of optical states. ............................................................................................. 115
Figure 5.5: Effect of Al2O3 and ITO thickness on the Purcell enhancement. ............ 117
Figure 5.6: Effect of Al2O3 and ITO thickness on the resonance wavelength and peak
intensity of the Purcell enhancement. ........................................................................ 117
Figure 5.7: Effect of antenna thickness on the Purcell enhancement. ....................... 118
Figure 5.8: Effect of antenna width and length, and electrode width on the Purcell
enhancement. ............................................................................................................. 119
Figure 5.9: Effect of dipole position on the Purcell enhancement. ............................ 121
Figure 5.10: Bias-induced modulation of the spontaneous emission decay rate of a
quantum emitter with a dipole moment along the x-direction. .................................. 124
Figure 5.11: Bias-induced modulation of the spontaneous emission decay rate of a
quantum emitter with a dipole moment along the y-direction. .................................. 124
Figure 5.12: Bias-induced modulation of the spontaneous emission decay rate of a
quantum emitter with a dipole moment along the z-direction. .................................. 125
Figure 5.13: Bias-induced modulation of the spontaneous emission decay rate of a
quantum emitter with a randomly-oriented dipole moment. ..................................... 126
Figure 5.14: Active control of the emission from a randomly-oriented dipole embedded
in the tunable metasurface. ........................................................................................ 126
Figure 6.1: Si-based dual-mode tunable metasurface. ............................................... 134
Figure 6.2: Dual-mode Si-based metasurface operating in the reflection mode........ 136
Figure 6.3: Dual-mode Si-based metasurface operating in the transmission mode. .. 138
Figure 6.4: Characterization of MQM wafers............................................................ 141
Figure 6.5: Design and simulation of MQW-based metasurface. .............................. 144
Figure 6.6: Optical setup used for the measurement of the reflection spectrum of the
MQW metasurface. .................................................................................................... 146
Figure 6.7: Measured tunable optical response of the MQW metasurface. ............... 147
Figure 6.8: Demonstration of switchable diffraction grating using MQW metasurface.
.................................................................................................................................... 150
Figure 6.9: Fabricated MQW-based all-dielectric metasurface for demonstration of
dynamic beam steering. ............................................................................................. 151
Figure 6.10: Simulated optical response of the MQW-based metasurface fabricated for
demonstration of dynamic beam steering. ................................................................. 152
Figure 6.11: Measured amplitude and phase modulation provided by the MQW-based
metasurface fabricated for demonstration of dynamic beam steering. ...................... 153
Figure 6.12: Demonstration of dynamic beam steering using all-dielectric MQW
metasurface. ............................................................................................................... 153

17
Figure 6.13: Effect of aperture size on the performance of the MQW beam steering
metasurface. ............................................................................................................... 154
Figure A.1: Effect of the thickness of the ITO layer on the device tunability. .......... 178
Figure A.2: HAOL gate dielectric. ............................................................................ 179
Figure A.3: Surface roughness of the sputtered Al back reflector. ............................ 180
Figure B.1: Comparison between the single-gated Au-based metasurface and the dualgated Al-based metasurface. ...................................................................................... 184
Figure B.2: Beam steering performance of the multifunctional metasurface for different
numbers of individually-controllable metasurface elements (N). .............................. 185
Figure B.3: Effect of antenna thickness on the optical response of the metasurface. 186
Figure B.4: Effect of antenna thickness on the spatial distribution of the electric field
intensity. ..................................................................................................................... 187
Figure B.5: Effect of antenna thickness on the spatial distribution of the x-component
of the electric field. .................................................................................................... 187
Figure B.6: Effect of HAOL thickness on the optical response of the metasurface. . 188
Figure B.7: Effect of HAOL thickness on the spatial distribution of the electric field.
.................................................................................................................................... 188
Figure B.8: Effect of ITO thickness on the optical response of the metasurface. ..... 189
Figure B.9: Effect of ITO collision frequency on the optical response of the metasurface.
.................................................................................................................................... 190
Figure B.10: Effect of the thickness of the lower dielectric layer on the optical response
of the metasurface. ..................................................................................................... 190
Figure B.11: Effect of the refractive index of the lower dielectric layer on the optical
response of the metasurface. ...................................................................................... 191
Figure B.12: Effect of top dielectric coat on reflectance and maximum achievable phase
shift of the metasurface. ............................................................................................. 192
Figure B.13: Near-field distribution of the amplitude of the electric field when the
nanoantennas are covered by a SiO2 top coat layer. .................................................. 192
Figure B.14: Near-field distribution of the x-component of the electric field when the
nanoantennas are covered by an SiO2 top coat layer. ................................................ 193
Figure B.15: Schematic layout of the sample-mounting PCB used to demonstrate
dynamic beam steering and reconfigurable focusing using the multifunctional
metasurface. ............................................................................................................... 194
Figure B.16: Schematic layout of the voltage-deriving PCB used to demonstrate
dynamic beam steering and reconfigurable focusing using the multifunctional
metasurface. ............................................................................................................... 195
Figure B.17: Layout of the photomask used for patterning the contact pads of the
multifunctional metasurface....................................................................................... 196
Figure B.18: Layout of the shadow mask used for patterning the Al2O3 layer of the
multifunctional metasurface....................................................................................... 197
Figure B.19: Layout of the EBL pattern used for patterning the ITO contact pads of the
multifunctional metasurface....................................................................................... 197
Figure B.20: Layout of the shadow mask used for patterning the HAOL layer of the
multifunctional metasurface....................................................................................... 198
Figure B.21: Layout of the EBL pattern used for the antennas and the inner connection
lines of the multifunctional metasurface. ................................................................... 199
Figure B.22: Schematic illustration of the layout of the final multifunctional
metasurface. ............................................................................................................... 199
Figure C.1: Configuration of the MIM structure used to measure the breakdown field
on the HAOL film. ..................................................................................................... 200

18
Figure D.1: Convergence test for the FDTD simulation region size. ........................ 201
Figure D.2: Atomic concentration of different components of Er-doped alumina films
obtained from RBS. ................................................................................................... 203
Figure D.3: Effect of RF power applied to the Er target. .......................................... 204
Figure D.4: Ag-based gate-tunable metasurface designed for demonstration of tunable
spontaneous emission enhancement. ......................................................................... 205
Figure D.5: Optical setup used for measuring the voltage-tunable Purcell enhancement
of quantum emitters embedded in active metasurfaces. ............................................ 205
Figure D.6: Tunable metasurface used for measuring the voltage tunable Purcell
enhancement. ............................................................................................................. 206
Figure E.1: Schematic layout of the sample-mounting PCB used to demonstrate the
MQW beam steering metasurface. ............................................................................. 207
Figure E.2: Schematic layout of the voltage-deriving PCB used to demonstrate the
MQW beam steering metasurface. ............................................................................. 208

19

LIST OF TABLES

Table 2.1: Experimentally measured phase and reflectance values. The reported values
of V0 provide phase shift in 90o steps........................................................................... 54
Table 2.2: Experimentally measured phase and reflectance values. The reported values
of V0 a provide phase shift in 60o steps. ....................................................................... 55
Table A.1: Electrical and Optical parameters obtained from Hall measurements and
spectroscopic ellipsometry for the ITO films deposited using different Ar+O2 flow rates.
.................................................................................................................................... 182
Table C.1: Measured breakdown field of the fabricated HAOL film for demonstration
of active polarization conversion metasurface. .......................................................... 200
Table D.1: Spectroscopic ellipsometry results of the Er-doped alumina films cosputtered on Si substrates in different Ar and O2 gas flow rates at different deposition
temperatures. .............................................................................................................. 202
Table D.2: RBS results of the Er-doped alumina films co-sputtered on Si substrates in
different Ar and O2 gas flow rates at different deposition temperatures. .................. 203

20

ABBREVIATIONS

AFM

atomic forced microscopy

ALD

atomic layer deposition

AOM

acousto-optic modulator

C−V

capacitance−voltage

CCD

charge-coupled device

CMOS

complementary metal–oxide–semiconductor

DAC

digital to analog converter

DBR

distributed Bragg reflector

DC

direct current

DRAM

dynamic random access memories

EBL

electron-beam lithography

EBPG

electron beam pattern generator

EBR

electron-beam resist

ENZ

epsilon-near-zero

F-P

Fabry-Pérot

FDTD

finite difference time domain

FWHM

full width at half maximum

GHz

gigahertz

GM

guided mode

GST

germanium–antimony–tellurium

HAOL

hafnium-aluminum oxide laminate

I−V

current−voltage

IR

infrared

ITO

indium tin oxide

LC

liquid crystal

LCP

left-handed circularly-polarized

LiDAR

light detection and ranging

LDOS

local density of optical states

MA

moving average

MEMS

microelectromechanical systems

Mie-GM

Mie-guided mode

21
MIM

metal-oxide-metal

MIR

mid-infrared

MIS

metal-insulator-semiconductor

MOS

metal-oxide-semiconductor

MQW

multiple-quantum-well

NIR

near-infrared

PB

Pancharatnam-Berry

PCB

printed circuit board

PL

photoluminescence

PML

perfectly matched layer

PMMA

polymethyl methacrylate

PMT

photomultiplier tube

QCSE

quantum-confined Stark effect

QD

quantum dot

Q-factor

quality factor

RBS

Rutherford backscattering spectrometry

RCP

right-handed circularly-polarized

RCWA

rigorous coupled wave analysis

RMS

root mean square

RTA

rapid thermal annealing

SEM

scanning electron microscopy

SIS

semiconductor-insulator-semiconductor

TE

transverse electric

TEM

transmission electron microscopy

THz

terahertz

TM

transverse magnetic

VCSEL

vertical-cavity surface-emitting laser

22
Chapter 1

INTRODUCTION
Recent advances in nanophotonic integration have created burgeoning interest in
controlling the interactions between light and matter at a scale comparable to, or even
smaller than the wavelength of light. Along the same line, the miniaturization of
nanophotonic elements has yielded a new class of low-profile optical components,
known as metasurfaces. Metasurfaces that are artificially designed nanostructured
surfaces have been introduced as compact and planar alternatives to conventional bulky
optical elements. Owing to their well-engineered subwavelength meta-atoms,
metasurfaces provide the possibility to spectrally, temporally, or spatially manipulate
electromagnetic waves with subwavelength spatial resolution. The ability of
metasurfaces to demonstrate compact, high-performance, and low-cost optical devices
has made them a thriving field of nanophotonics over the last decade. In this chapter,
we will provide a brief history of metasurfaces and describe a number of the recently
demonstrated metasurfaces. We then will describe active metasurfaces whose functions
could be reconfigured at a post-fabrication stage. We continue with an overview of the
active metasurfaces that operate based on different active platforms. A short outline of
the thesis contents concludes the chapter.

1.1. Metasurfaces: Motivations and Applications
Metasurfaces are two-dimensional arrays of subwavelength scatterers, referred to as
meta-atoms. Each meta-atom is precisely designed to impose the desired change to the
fundamental attributes of light such as amplitude, phase, polarization, and wavelength.
Accordingly, one will be able to engineer the characteristics of the light interacting with
the metasurface via an appropriate choice of the meta-atoms. This will make
metasurfaces strong candidates to replace bulky optical elements such as gratings,
lenses, polarizers, mirrors, and waveplates.
Moreover, metasurfaces can provide functions unachievable by conventional optical
components. Conventional optical elements, in which phase shift accumulates as the
beam travels over distance in one medium, obey the classical Snell’s law (Fig. 1.1a):

23
sin 𝜃𝑖

= 𝜈𝑖 = 𝜆𝑖 = 𝑛𝑖
sin 𝜃

𝜃𝑟 = 𝜃𝑖

(1.1)

where 𝜃𝑖 , 𝜈𝑖 , 𝜆𝑖 , and 𝑛𝑖 (𝜃𝑡 , 𝜈𝑡 , 𝜆𝑡 , and 𝑛𝑡 ) are the angle, speed, and wavelength of light
and the refractive index in the incidence (transmission) medium, respectively, and 𝜃𝑟
is the angle of reflection.
On the other hand, metasurfaces can introduce abrupt changes to the properties of the
incident beam, and hence, can create phase gradients at the interface of two media (Fig.
1.1b). As a consequence, the generalized Snell’s law was introduced in 2011 to account
for such phase gradients [1]:
𝑛𝑡 sin 𝜃𝑡 − 𝑛𝑖 sin 𝜃𝑖 = 𝑘0−1 ∇𝜙

(1.2)

sin 𝜃𝑟 − sin 𝜃𝑖 = 𝑛𝑖−1 𝑘0−1 ∇𝜙

(1.3)

where 𝜃𝑖 , 𝜃𝑡 , and 𝜃𝑟 are the angles of the incident, transmitted, and reflected beams,
respectively. 𝑘0 denotes the wavenumber in free space, and ∇𝜙 is the phase gradient.

Figure 1.1: Schematic illustration of Snell’s law. (a) Classical Snell’s law for refracted light
passing from one medium to another one, and (b) generalized Snell’s law when there is a phase
gradient at the interface of two media.

It can be inferred from the generalized Snell’s law that a careful choice of the phase
gradient imposed by each meta-atom would enable the generation of the far-field at
will. As a result, metasurfaces can manipulate light in ways not allowed by
conventional optical components [2], [3]. The control offered by metasurfaces over the

24
properties of reflected or transmitted light has given rise to the field of flat optics, which
explores how metasurfaces can be used for the creation of low-profile optical elements
[4].
Thanks to their versatile functionalities, ease of fabrication via planar lithographic
processing, and amenability for integration, metasurfaces have been of growing interest
in optics and photonics. As a result, there has been intensive research to demonstrate
metasurfaces with different functions such as wave-front shaping metasurfaces [5]–[14],
holograms [15]–[21], optical vortex generators [22]–[28], frequency selective [29] and
perfect absorber [30]–[32] devices. Not only have these examples witnessed a plethora
of functions provided by metasurfaces in current devices in a wide range of the
electromagnetic spectrum, but they also pave the way for inspiring many new thrilling
applications such as programmable on-demand optics and photonics in the future.
A two-dimensional array of optical scatterers with subwavelength separation was first
employed to impose phase discontinuities on the propagating light [1]. Each scatterer
consisted of gold (Au) V-shaped antennas, formed from two nanorods with equal
lengths joined together at a certain angle (Fig. 1.2a). With an incident beam being
polarized in a direction neither parallel nor perpendicular to the antenna symmetry axis,
two resonant modes (called symmetric and antisymmetric modes) were excited
simultaneously. Changing the angle between the two nanorods then resulted in a phase
variation of 0 to 2π. By a careful choice of the phase variation along the interface,
optical vortices were generated via planar designer metasurfaces.
In another study, silicon (Si)-based gradient metasurfaces were proposed to
demonstrate optical gratings, lenses, and axicons (Fig. 1.2b) [9]. A 100 nm-thick layer
of poly-silicon deposited on a quartz substrate was patterned to create nanoantennas.
In each case, nanoantennas were patterned to impose a Pancharatnam-Berry (PB) phase
on the incident electromagnetic wave. By precisely designing the space-variant phase
changes over the metasurface, the desired wavefronts were achieved.
Along the same line, high-aspect-ratio metasurfaces were also used to demonstrate high
numerical aperture metalenses at the visible region (Fig. 1.2c) [10]. The metasurface
was composed of nanofins made of amorphous titanium dioxide (TiO2) owing to its
low surface roughness, no absorption at the visible wavelengths, and a sufficiently high
refractive index (~2.4).

25

Figure 1.2: A review of the recent passive metasurfaces. (a) Scanning electron microscopy
(SEM) image of the metasurface with V-shaped nanoantennas used to generate an optical
vortex (left panel). Each region of the metasurface is composed of one constituent antenna of
the eight-element set shown on top. Measured (top right), and calculated (bottom right) spiral
pattern obtained from the interference of the vortex beam generated via creating a phase shift
azimuthally varying from 0 to 2, and a co-propagating Gaussian beam; adapted from [1]. (b)
SEM image of the Si-based metasurface used as an axicon (top left). The inset shows the
transversal distribution of the Bessel beam generated by the metasurface. SEM image of the
metasurface functioning as a metalens (top right). The inset shows the two-dimensional
intensity profile in the focal plane. SEM image of the fabricated metasurface used as a blazed
grating (bottom left) and its measured diffraction patterns (bottom right) as well as associated
cross-sectional intensity distributions versus the normalized in-plane momentum at different
wavelengths; adapted from [9]. (c) Schematic of the metalens composed of high aspect
ratioTiO2 nanofins (top left), and the top view of the metasurface unit cell (bottom left). By
rotating the nanofins by the angle nf, the required phase is imposed by the meta-atom to the
incident beam, creating the desired wavefront. SEM image of the fabricated metalens (top
right), and intensity distribution in the x-z plane at the wavelength of  = 532 nm (bottom right);
adapted from [10]. (d) SEM image of the metasurface fabricated to serve as a meta-hologram
(top), as well as the calculated (middle) and measured (bottom) holographic images from two
holograms; adapted with permission from [33] © The Optical Society.

26
According to the geometric PB phase, by rotating the nanoantennas by a specific angle,
the phase imposed by each antenna was precisely engineered in order to obtain the
targeted spatial phase profile over the metasurface. As a result, diffraction-limited
focusing was obtained by the metalens with efficiencies of 86%, 73%, and 66% at the
wavelengths of 405 nm, 532 nm, and 660 nm, respectively.
Metasurfaces have also been used to realize meta-holograms [33]. In a related study,
sub-diffraction lattices of Si nanopillars were used to demonstrate transparent metaholograms with superior diffraction and transmission efficiencies (Fig. 1.2d). Each
meta-atom was designed to support multiple electric and magnetic Mie resonances,
resulting in a suppression of the backward scattering at multiple wavelengths. By
changing the size of the nanopillars, and hence, the phase delay imposed by each metaatom, a 2π phase variation across the hologram, accompanied by over 2π phase
variation across the operational spectral bandwidth was achieved. By careful
engineering of the spatial phase profile of the metasurface, a transparent meta-hologram
was obtained that could be used to encode grayscale images.

1.2. Active Metasurfaces
As can be seen, metasurfaces are now demonstrating some of their potential
applications in compact, high-performance, and low-cost optical devices and
components, creating a burgeoning interest in photonic integration. However, despite
tremendous progress in this field, the metasurfaces presented in these prior reports are
passive, which means their properties are fixed at the time of the fabrication and do not
allow for post-fabrication tunability. In other words, metasurfaces have mostly been
designed in an application-specific manner, and the design process resulted in bespoke
architectures tailored to particular applications.
Dynamical control of the properties of the scattered light is possible by using tunable
metasurfaces, for which external stimuli such as electric [34]–[36] and magnetic fields
[37], electrostatic and micromechanical forces [38], chemical reactions [39], [40], and
optical pumping [41] can give rise to changes in the dielectric function or physical
dimensions of the metasurface elements [42], thereby modulating the antenna phase
and amplitude response. The ability to actively and dynamically tune the properties of
metasurfaces would enable dynamic holograms, focusing lenses with reconfigurable

27
focal lengths, and beam steering, a key requirement for future chip-based light detection
and ranging (LiDAR) systems.
Several approaches have been used to actively control the optical response of
metasurfaces in the mid-infrared (MIR) [36], [43]–[48], near-infrared (NIR) [37], [49]–
[55], and visible [56], [57] wavelength ranges. The targeted operating wavelength
usually dictates the appropriate material platform and tuning mechanism to realize
actively tunable metasurfaces.
In the MIR wavelength range, carrier density modulation via the gating of graphene
[43], [44], [58], GaAs [45], or indium tin oxide (ITO) [46], has been employed as a
mechanism to modulate metasurface reflectance. In addition, thermo-optic tuning of
PbTe [47] antennas has yielded actively tunable structures in this wavelength range. An
electronically reconfigurable metasurface based on a graphene-gold resonator
geometry was used to demonstrate phase modulation in the MIR [36]. By altering the
Fermi energy of graphene due to the application of a direct current (DC) bias (Fig. 1.3a),
both the inter- and intra-band contributions to its complex permittivity were tuned,
leading to amplitude and phase modulation. As a result, the phase modulation of 237°
was observed at an operating wavelength of 8.50 μm that was utilized to show beam
steering with an average efficiency of 23%.
In the NIR and visible wavelength ranges, researchers have employed a number of
different physical mechanisms to realize active metasurfaces [37], [49]–[54], [56], [59].
In one study, ITO layers were incorporated into metasurface antennas in order to
demonstrate dynamic electrical control of the phase and amplitude of the plane-wave
reflected from the metasurface in the NIR wavelength range [35]. Due to the field-effect
modulation of the complex refractive index of ITO layers, a phase shift of 180° and
∼30% change in the reflectance were observed when applying 2.5 V gate bias with a
modulation frequency exceeding 10 MHz (Fig. 1.3b). Moreover, by electrical control
over subgroups of metasurface elements, an electrical switching of the ±1st-order
diffracted beams was realized.
In related research, a large-scale two-dimensional nanophotonic phased array was
reported [60]. The metasurface device in which 64×64 optical nanoantennas were
densely integrated on a Si chip, operated based on thermo-optic phase tuning in Si (Fig.

28
1.3c). Employing the mentioned active phase tunability in an 8×8 array, dynamic beam
steering and shaping were observed.
An alternative approach to achieve tunable metasurfaces is to integrate liquid crystal
(LC) active layers into the otherwise passive metasurfaces. As an example,
temperature-dependent refractive index change of a nematic LC has led to a dynamic
tuning of electric and magnetic resonances in all-dielectric Si nanodisk metasurfaces
(Fig. 1.3d) [53]. Consequently, a 5-fold modulation of the transmittance accompanied
by a resonance tuning range of 40 nm was obtained in the telecom spectral range.
In the visible spectral region, other approaches such as using phase-change materials
and ionic transport could lead to the realization of reconfigurable metasurfaces.
Integrating germanium–antimony–tellurium (GST) films into metasurfaces has been
considered in order to achieve nonvolatile and reversible fashion to obtain tunable
metasurfaces [57]. By exploiting a reversible femtosecond-laser-controlled refractive
index transition that occurs in the GST layer, the functionality of the device can be
written, erased, and rewritten (Fig. 1.3e). When writing two-dimensional binary or
greyscale patterns into the device, visible-range reconfigurable bichromatic and multifocus Fresnel zone plates were obtained.
In another study, a novel modulation scheme based on the transport of silver ions
through an alumina (Al2O3) dielectric layer was proposed to obtain tunable
metasurfaces [56]. Such an ionic transport resulted in bias-induced nucleation and
growth of silver nanoparticles in an ITO counter-electrode, and hence, altering the
optical extinction response (Fig. 1.3f). The metasurface that operated at strikingly low
modulation voltages showed up to 30% relative change in reflectance in the visible
spectral range upon application of 5 mV of bias and 78% absolute change in reflectance
upon application of 100 mV of bias.
In addition to the discussed techniques, the reflectance and transmittance of a
metasurface can be mechanically modulated [38] using electrostatic and magnetic
forces [37], [50].
As can be seen, active metasurfaces have gained a great deal of attention because of
their unique ability to provide precise and real-time engineering of the wavefront of
light at the nanoscale.

29

Figure 1.3: A review of the recent advances in active metasurfaces. (a) Schematic of a
graphene-based gate-tunable device for control of reflected phase (left), and phase modulation
at the wavelengths of 8.2 μm, 8.50 μm, and 8.7 μm (right). Circles and lines show the
experiment and simulation results, respectively; adapted from [36]. (b) Measured reflectance
spectrum (top left) as well as measured and simulated phase shift (bottom left) as a function of
applied bias for ITO-based tunable metasurface. The inset shows the interference fringes for
applied biases of 0 and 2.5 V. Schematic of steered diffracted beams via electrical gating of
different numbers of antennas (right); adapted from [35]. (c) Schematic illustration of a 64×64
nanophotonic phased array system (left) in which a laser input from an optical fiber is delivered
equally to each nanoantenna. The inset shows a close-up view of one antenna unit cell.
Simulation (top right) and experimental (bottom right) close-up view of the radiation pattern of
the system designed to generate the MIT logo in the far-field. The inset on the lower right side
of the top right panel shows the targeted MIT logo pattern; adapted from [60]. (d) Schematic
illustration of an LC-based metasurface with Si nanodisk antennas (left). Heating the cell from
the backside of the Si handle wafer will result in a temperature-dependent modulation of the
LC refractive index. Measured transmittance spectra of the metasurface (right). The white
dashed line indicates the phase transition. The red dots and the cyan squares show the resonance
positions of the electric and magnetic resonances, respectively; adapted from [53]. (e)
Schematic of a GST-based reconfigurable photonic device (left) whose optical response could
be written by a femtosecond pulsed laser. Illustration of a lens focusing the lights with two
wavelengths into spatially separated foci (top right) and to the same focal point (bottom right)
on the focal plane; adapted from [57]. (f) Schematic representation of the transport of silver
ions through an alumina layer as a result of applying a DC bias. Increasing the applied bias is
accompanied by an increment in the silver nanoparticles migration; adapted from [56].

30
While these reports indicate options for active control of the scattered light intensity,
phase modulation of the scattered light upon external actuation is of increasing
importance. Different tuning platforms have been proposed to achieve reconfigurable
metasurfaces with wide amplitude/phase modulation. Phase-change materials such as
vanadium dioxide (VO2) and GST enable large-volume index modulation [57], [61],
but their switching speed is limited [51] and they generally require large power
consumption [62] compared to field-effect modulation schemes. Furthermore, with
optical pumping, the area for refractive index modulation is determined by the size of
the focused laser spot, and is thus relatively large, limiting the ability to achieve control
of individual metasurface elements. Prior research has also combined tunable
metasurface optics with microelectromechanical systems (MEMS) technology to
demonstrate varifocal lenses [63].
Moreover, previous work has shown that fabricating metasurfaces on elastomeric
substrates may yield adaptive metalenses [64], strain-multiplexed meta-holograms [65],
and active control of the structural color [66]. However, in MEMS-based and
mechanically stretchable substrate modulation approaches, control of the optical
response is achieved by changing the distance between either adjacent metasurface
elements or entire element arrays and requires a mechanical transducer, which limits
the modulation frequency. While interesting, these approaches are not able to yield
versatile active control over the scattered light wavefront with independent control of
phase and amplitude at a subwavelength scale.
Among the modulation mechanisms, electrical tuning has been proven to be a robust,
energy-efficient, and reversible scheme for tuning active metasurfaces [34]–[36], [45],
[51]–[53], [56], [67]–[71]. In this thesis, we will focus on a number of electrical tuning
mechanisms that have been proposed to achieve active metasurfaces.

1.3. Thesis Outline
Chapter 2 of this dissertation proposes a conceptually new dual-gated electro-optically
tunable metasurface platform that could provide amplitude modulation as well as a
record-high phase shift. In Chapter 3, we present an electro-optically tunable universal
metasurface whose function could be tuned by individually controlling the voltages
applied to each metasurface element, and could serve as a base for future on-chip

31
integrated optical devices. As a proof of principle, dynamic beam steering and
reconfigurable focusing are demonstrated using the universal metasurface. Chapter 4
introduces the use of active metasurfaces to control the polarization of light. It will be
shown that by a careful choice of the bias applied to the metasurfaces, a linearlypolarized incident beam can be scattered as light with linear, elliptical, or circular
polarization.
Chapter 5 focuses on the demonstration of reconfigurable enhancement of the
spontaneous emission decay rate of quantum dots and quantum emitters by integrating
them within tunable metasurfaces. In Chapter 6, alternative approaches to obtain
dielectric tunable metasurfaces will be discussed. It will be shown how dielectric
metasurfaces could be employed for operation in both transmission and reflection
modes and achieving higher efficiency of the device as a result of the small loss
provided by the metasurface. We conclude the dissertation with a discussion of
motivations and challenges towards obtaining tunable metasurfaces in Chapter 7.

32
Chapter 2

DUAL-GATED ACTIVE METASURFACE FOR WIDE PHASE
TUNABILITY
The material in this chapter was in part presented in [34].
Active metasurfaces composed of electrically-reconfigurable nanoscale antenna arrays
can enable real-time control of scattered light amplitude and phase. Achievement of
widely tunable phase and amplitude in chip-based active metasurfaces operating at or
near 1550 nm wavelength has considerable potential for active beam steering, dynamic
hologram rendition, and realization of flat optics with reconfigurable focal lengths.
Previously, electrically-tunable conducting oxide-based reflectarray metasurfaces have
demonstrated dynamic phase control of reflected light with a maximum phase shift of
184° [35]. Here, we introduce a dual-gated reflectarray metasurface architecture that
enables much wider (>300°) phase tunability. We explore light-matter interactions with
dual-gated metasurface elements that incorporate two independent voltage-controlled
metal-oxide-semiconductor (MOS) field-effect channels connected in series to form a
single metasurface element that enables wider phase tunability. Using indium tin oxide
(ITO) as the active metasurface material and a composite hafnia/alumina gate dielectric,
we demonstrate a prototype dual-gated metasurface with a continuous phase shift from
0 to 303° and a relative reflectance modulation of 89% under an applied bias voltage
of 6.5 V.

2.1. Active Metasurfaces for Modulation of Properties of Light
The ability to actively control all the important constitutive properties of light
(wavelength, amplitude, phase, polarization state) via interaction with tunable
nanoscale elements is a grand challenge in nanophotonics. Metasurfaces are twodimensional nanostructured surfaces that enable versatile wavefront control for
scattered light [1]. Metasurfaces can also be viewed as arrays of subwavelength
antennas such that each antenna imposes a predefined phase shift, amplitude change,
and polarization rotation on the scattered light. To date, metasurfaces have been used
to realize focusing mirrors [72], focusing lenses [10], holograms [18], [73], [74], and

33
polarization converters [75], [76]. However, the metasurfaces in these prior reports are
passive, which means their properties are fixed at the time of the fabrication. The ability
to actively and dynamically tune the properties of metasurfaces would enable dynamic
holograms, focusing lenses with reconfigurable focal lengths, and beam steering,
leading to the realization of multiple important devices such as chip-based LiDAR
systems.
Several approaches have been used to actively control the optical response of
metasurfaces in different wavelength ranges. Beam steering has been demonstrated
with chip-based Si photonics phased arrays operating at a wavelength of λ=1550 nm
[78]. In this approach, the phase of each antenna was actively tuned by a waveguidebased thermo-optic phase shifter through an integrated heater on the Si chip. The silicon
photonics approach enabled continuous tuning of the phase of emitted light from 0 to
360° upon application of external bias. However, the large pixel size of the phased array
(9 μm × 9 μm) results in undesired side lobes. Moreover, thermo-optic control limited
the modulation frequency of these phased arrays to less than 50 kHz [78], which is too
slow for versatile beam steering in technologically important LiDAR applications, and
the thermal crosstalk between phase shifters and the photodetectors limited the
detection range to 20° [79].
Metasurfaces offer a different approach to a phased array architecture, in which the
array is intrinsically two-dimensional and the subwavelength antenna dimensions and
antenna spacing can suppress side lobes. Hence, it would be highly desirable to have a
tunable metasurface platform for comprehensive and active control of scattered light in
the NIR spectral range. Our group has previously investigated field-effect modulation
of the carrier density and refractive index of heavily doped semiconductors as an
approach towards NIR actively-tunable metasurfaces [35]. This approach relies on the
field-effect-induced charge accumulation or depletion in the semiconducting electrode
of a nanoscale MOS structure that also serves as a resonant antenna. Using ITO as a
semiconducting layer of the MOS field-effect structure enabled active modulation of
the optical response of plasmonic reflectarray metasurfaces [35], [67], with a
corresponding reflected light phase shift from 0 to 184° for an applied bias between 0
to 2.5 V. While conceptually promising as an approach to active metasurface design, in
order to realize a comprehensively tunable metasurface, a phase shift from 0°
approaching to 360° is desirable.

34
In this chapter, we report the design and fabrication of dual-gated field-effect-tunable
metasurface antenna arrays that enable phase shifts exceeding 300° at a wavelength of
λ=1550 nm.

2.2. Design of Dual-gated Metasurfaces
Our dual-gated metasurface features two charge accumulation/depletion layers within
the dielectric spacer of each active metasurface antenna (Fig. 2.1a). The dual-gated
metasurface structure consists of an aluminum (Al) back reflector, a gatedielectric/ITO/gate-dielectric heterostructure (Fig. 2.1b), and a periodic array of Al
nanoantennas with a ‘fishbone’ pattern. The fishbone nanoantennas are composed of
patch antennas that are connected by Al stripes, which also serve as gate voltage control
electrodes.

Figure 2.1: Dual-gated active metasurface. (a) Schematic of the unit cell of the dual-gated
metasurface, which is composed of an Al back reflector, a bottom gate dielectric, an ITO layer
followed by another gate dielectric, on top of which Al fishbone antennas are located. The
thicknesses of the antenna array, the gate dielectrics, the ITO layer, and the back reflector are
t1 = 40 nm, t2 = 9.5 nm, t3 = 5 nm, and t4 = 80 nm, respectively. The antenna dimensions are l
= 280 nm and w1 = 120 nm, and the electrode width is w2 = 170 nm. The period of the
metasurface is p = 400 nm. A voltage bias Va is applied between the ITO layer and the top
antennas, while another voltage bias Vb is applied between the Al back reflector and the ITO
layer. The two applied voltage biases result in the formation of two accumulation/depletion
regions in the ITO layer at the top and bottom ITO/gate-dielectric interfaces. (b) A magnified
image of the dielectric spacer of the metasurface that consists of the top gate dielectric, the ITO
layer, and the bottom gate dielectric.

Each metasurface element permits the application of two independent DC voltages, i)
between the ITO layer and the fishbone antenna, and ii) between the ITO layer and the
back reflector. As a result, both the top and bottom ITO/gate-dielectric interfaces can
exhibit the charge accumulation or depletion under applied external bias. This design

35
facilitates a large variation of the complex refractive index of the ITO layer via carrier
density modulation at both its top and bottom interfaces (Fig. 2.1b) and is a key reason
for the wide phase tunability of our dual-gated metasurface.

2.2.1. Electrostatic Simulations to Extract ITO Properties
In designing dual-gated metasurfaces, we account for a number of considerations that
can increase the metasurface tunability and efficiency. We choose the ITO carrier
concentration to be N0= 3 × 1020 cm-3 to ensure that the real part of the dielectric
permittivity of the ITO layer is positive at a wavelength of λ=1550 nm when no external
bias is applied. Under bias, a charge accumulation layer is formed in the ITO, and the
real part of the dielectric permittivity of the accumulation layer can change its sign,
undergoing the transition from the optically-dielectric to optically-metallic phase.
When the dielectric permittivity of the accumulation layer is in the epsilon-near-zero
(ENZ) region, which means -1< Re(ε) <1, the optical electric field intensity in the
accumulation layer is strongly enhanced, resulting in the modulation of the intensity
and phase of the scattered light [46], [80]–[82]. The optical electric field enhancement
in the ENZ region of ITO arises from the continuity of the normal component of the
electric displacement as the index approaches zero in this region [80], [81]. This
suggests that increasing the number of accumulation/depletion layers within the active
region of the metasurface antenna can be beneficial for enhancing phase tunability. On
the other hand, since the optical loss of the ITO layer is non-negligible, we design the
ITO layer to be as thin as possible. Based on these considerations, the ITO layer
thickness is chosen to be 5 nm in our dual-gated metasurface (See Appendix A.1).
To accurately calculate the optical response of metasurfaces under applied bias, we
couple the device physics simulations (Device Lumerical) with finite difference time
domain (FDTD) optical simulations (Lumerical). The device physics simulations are
used to determine the charge carrier distribution in the ITO layer under applied bias.
Our electrostatics calculations model the spatial distribution of charge carriers in the
ITO layer embedded in the metasurface. In our device physics calculations, we assume
the work function of Al to be 4.3 eV. We also assume the effective electron mass of
ITO of m* = 0.35 me, with me being the free electron mass and electron mobility of ITO
of 25 cm2V-1s-1. Since our ITO is degenerately doped, we assume no significant

36
contribution of holes to the observed physical processes. In Device Lumerical software,
we insert the holes effective mass of 1×me, and the hole mobility of 1 cm2V-1s-1. In our
simulations, the bandgap of ITO is set to 2.8 eV [83], and the electron affinity of ITO
is chosen as 4.8 eV. The assumed DC permittivity of ITO is 9.3 [84].
Once we identify the spatial distribution of charge under different applied biases, we
then relate the calculated carrier density to the complex dielectric permittivity of ITO
εITO by using the Drude model:
𝜀ITO = 𝜀∞ − 𝜔𝑝2 ⁄(𝜔2 + 𝑖𝜔𝛤)

(2.1)

where the plasma frequency 𝜔𝑝 is given by the following expression:
𝜔𝑝 = √𝑁ITO 𝑒 2 /(𝜀0 𝑚∗ )

(2.2)

Here, NITO is the carrier concentration of ITO, which we extract from the device physics
calculations, e is the electron charge, 𝜀0 is the DC permittivity of vacuum, 𝑚∗ is the
effective electron mass, 𝛤 is the damping constant, 𝜀∞ is a fitting constant, and 𝜔 is the
angular frequency, which is related to the wavelength λ as λ=2πc/ω, where c is the
speed of light in vacuum. When performing optical simulations, we assume that
m*=0.35 me, 𝛾 = 1.8 × 1014 , and 𝜀∞ =3.9.

2.2.2. Choice of Plasmonic Metal
Another parameter that determines the performance of the device is the choice of the
plasmonic metal. The work functions of Al and silver (Ag), which are both near 4.3 eV,
are close to the work function of ITO when the carrier concentration equals N0= 3 ×
1020 cm-3, while the work function of Au (5.1 eV) is higher than that of the ITO. Hence,
using Al or Ag as a metal electrode in the metal/gate-dielectric/ITO capacitor reduces
the zero-bias band bending in the ITO layer compared to an Au electrode. This implies
that in the case of Al or Ag electrodes, one needs to apply lower bias voltages to
overcome the depletion and form an accumulation layer in the ITO at the gatedielectric/ITO interface. Previous research has indicated that Ag can also migrate into
the gate dielectric layers under applied electrical bias [56], [85]. To eliminate this issue,
we use Al, a complementary metal-oxide-semiconductor (CMOS)-compatible material,
as the plasmonic metal in our tunable metasurfaces.

37
2.2.3. Choice of Gate Dielectric
The attainable optical modulation in our tunable metasurface is also determined by the
choice of the gate dielectric material. To enable the largest possible variation of carrier
density in ITO under applied voltage, one would ideally like to have a gate dielectric
with high DC permittivity and high breakdown field. Alumina and hafnia (HfO2) are
among the most commonly used high dielectric constant gate dielectric materials that
are employed in field-effect transistor technology. Al2O3 exhibits good thermal stability
and almost perfect interfacial properties with Si-based substrates, has a large bandgap,
and a high breakdown field of up to 10 MV/cm [86], [87]. However, it suffers from a
relatively low DC permittivity of kAl2O3 = 9. On the other hand, HfO2 is a CMOS
compatible material with a wide bandgap, and a relatively high dielectric constant of
up to kHfO2 = 25. But it exhibits a small breakdown field of 3.1 MV/cm, and high leakage
current induced by its low crystallization temperature.
Previous research has shown that Al2O3/HfO2 nanolaminates, commonly referred to as
hafnium-aluminum oxide laminate (HAOL) materials, can have superior electrostatic
characteristics as compared to both Al2O3 and HfO2 [88]. HAOL structures, which are
grown via consecutive deposition of ultrathin Al2O3 and HfO2 layers, were previously
shown to have the low leakage current and high breakdown field characteristics of
Al2O3, and also the large DC permittivity characteristic of HfO2.
Accordingly, we were able to obtain a custom-developed recipe that provided us HAOL
films with superior electrostatic performance as compared to the Al2O3 and HfO2 films
(see Appendix A.2 for fabrication and characterization of HAOL). Our fabricated
HAOL film showed a DC permittivity of kHAOL = 22, and a breakdown field of EHAOL
= 7.2 MV/cm. As a consequence, by using our custom-made HAOL films as the gate
dielectric in our dual-gated metasurfaces, we have been able to solve one of the most
challenging obstacles of designing voltage-tunable metasurfaces which is the choice of
a gate dielectric with both high DC permittivity and high breakdown field.

2.3. Electromagnetic Simulations of Dual-Gated Metasurfaces
2.3.1 Modelling the ITO Layer in the Dual-Gated Metasurface Structure
As mentioned in the previous section, we model the optical response of our metasurface
under applied bias using FDTD simulations coupled to device physics simulations. In

38
order to apply bias to metasurface elements, fishbone antennas are connected by
electrodes, creating equipotential columns (see Fig. 2.2a). Then by connecting the
obtained columns to external metallic pads, we can control the bias applied to the
nanoantennas. The electrostatic performance of the dual-gated tunable metasurface
element can be viewed as two parallel plate capacitor structures which are connected
in series. Therefore, two independent bias voltages can be applied to each metasurface
element, Va and Vb (Fig. 2.2b). Figure 2.2c shows an SEM image of the fishbone
nanoantennas connected through the electrodes. A focused ion beam (FIB) crosssection image of the nanoantennas is presented in Fig. 2.2d. In order to apply bias to
the nanoantennas, all antennas are connected using connection pads as shown in Fig.
2.2e.

Figure 2.2: Configuration of bias applied to the nanoantennas in the dual-gated active
metasurface. (a) Schematic of the fishbone nanoantennas connected together via electrodes.
(b) Schematic showing the bias application configuration. The nanoantenna arrays are
electrically connected to an external pad to which we apply the voltages. (c) SEM image of the
nanoantennas. The scale bar is 500 nm. (d) FIB cross-section of the nanoantennas. The scale
bar is 500 nm. (e) SEM image of the nanoantennas connected for bias application. The scale
bar is 500 m.

For the sake of simplicity, we assume that |Va|=|Vb| that yields two accessible regimes
of device operation, where sign (Va×Vb) ≥0 (Case I) and where sign (Va×Vb) ≤0 (Case
II). The charge carrier distributions in the 5 nm-thick ITO layer for Case I and Case II
are depicted in Fig. 2.3. Here, the z position varies between 0 and 5 nm, with 0

39
corresponding to the bottom ITO/HAOL interface and 5 nm corresponding to the top
ITO/HAOL interface.

Figure 2.3: Illustration of Case I and Case II for the bias applied to the dual-gated active
metasurface. (a) Schematic of Case I in which there is a simultaneous charge accumulation or
simultaneous charge depletion at both ITO/HAOL interfaces of the ITO layer. In Case I, we
assume Va = V0 and Vb = V0. (b) Schematic of Case II in which the charge accumulation
(depletion) at the top ITO/HAOL interface is always accompanied by the charge depletion
(accumulation) at the bottom ITO/HAOL interface. In Case II, we assume Va = V0 and Vb = −V0.
The charge carrier distribution in the ITO layer as a function of applied voltage V0 for (c) Case
I and (d) Case II. The real part of the dielectric permittivity of the ITO layer as a function of
the applied voltage and position for (e) Case I and (f) Case II at a wavelength of λ = 1550 nm.
The boundaries of the ENZ regions are marked by dashed curves. The imaginary part of the
dielectric permittivity of the ITO layer as a function of the applied voltage and position for (g)
Case I and (h) Case II at a wavelength of λ = 1550 nm.

As can be seen, in Case I, there is a simultaneous charge accumulation or simultaneous
charge depletion at both ITO layer interfaces (see Fig. 2.3a, c, e, g). In Case II, charge
accumulation at the top ITO/HAOL interface is accompanied by charge depletion at the

40
bottom ITO/HAOL interface, or, vice versa, charge depletion at the top ITO/HAOL
interface is accompanied by charge accumulation at the bottom ITO/HAOL interface
(see Fig. 2.3b, d, f, h).
We should note that only the portion of the ITO located directly beneath the Al fishbone
antenna is optically modulated at the top ITO/HAOL interface (see Fig. 2.1b). As seen
in Fig. 2.3, in Case I the ITO dielectric permittivity at the bottom ITO/HAOL interface
is always equal to the dielectric permittivity of the ITO at the top ITO/HAOL interface
beneath the fishbone antenna. This, however, is not true for Case II. In Case II, for
sufficiently large applied voltage magnitude, there is always charge accumulation at
either top or bottom interface of the ITO layer.

2.3.2. Simulated Amplitude and Phase Response of the Dual-Gated Metasurface
After modeling the complex dielectric permittivity of ITO as a function of position and
applied voltage, we calculate the metasurface optical response for different applied
biases under normal-incidence illumination with a transverse magnetic (TM) polarized
plane wave (E-field along x-direction).
Here, the antenna width and length, and the width of the stripe electrode are chosen to
be w1 = 120 nm, l = 280 nm, and w2 = 170 nm (see Fig. 2.1a), respectively in order to
achieve a resonance around our wavelength of interest λ = 1550 nm. It should be noted
that in our simulations, we use 0.025 nm-thick mesh sizes in the active regions of ITO
(≈2 nm-thick) to carefully resolve the inhomogeneous permittivity profiles and capture
the accurate optical response of the unit cell.
Figure 2.4 shows the reflectance response of the dual-gated metasurface in Case I and
Case II as a function of wavelength and applied voltage. In Case I (Figs. 2.4a, c, e), a
large reflectance modulation is observed at positive biases, when the dielectric
permittivity of both the top and bottom ITO interfaces cross into the ENZ region. In
this case, we observe a blue shift of the resonance when the applied bias increases from
0 to 2.5 V. For applied voltages larger than 2.5 V, the resonance red-shifts. This is
consistent with previously reported results [35]. In Case II, we observe a significant
reflectance modulation both at positive and negative biases (Figs. 2.4b, d, f). Moreover,
we observe that in Case II, the reflectance spectrum is invariant to the transformation
V0→−V0. This is due to the fact that in Case II, both at positive and negative biases, the

41
gap plasmon resonance coupled to the ENZ region in ITO that is formed in either the
top or the bottom ITO layer interface.
Figures 2.4e, f show the spectra of relative reflectance change at different applied
voltages for Case I and Case II, respectively. The insets of Figs. 2.4e, f show the relative
reflectance change as a function of applied voltage at a fixed wavelength of λ = 1550
nm. As could be seen, a very large relative reflectance change is achieved in both Case
I and Case II at the operating wavelength.

Figure 2.4: Simulated amplitude response of the dual gated metasurface for Case I and
Case II. Reflectance from the metasurface as a function of wavelength and applied voltage in
(a) Case I, and (b) Case II. Reflectance spectrum for different applied biases for (c) Case I and
(d) Case II. The relative reflectance change spectrum for different applied voltages for (e) Case
I and (f) Case II. The insets show the relative reflectance change as a function of voltage at a
wavelength of 1550 nm.

The phase response of the dual-gated metasurface under applied bias is presented in
Fig. 2.5. Figure 2.5a shows the acquired phase spectrum of Case I for different applied
biases. Here, the acquired phase is defined as a difference between the phases of the
reflected and incoming plane waves calculated at the same spatial point. As can be seen,

42
in this case, a large phase shift is observed at positive biases, when the dielectric
permittivity of both the top and bottom ITO interfaces cross into the ENZ region.
In Case II, similar to the reflectance modulation discussed above, we can observe a
significant phase modulation both at positive and negative biases (Fig. 2.5b).

Figure 2.5: Simulated phase response of the dual gated metasurface for Case I and Case
II. The spectrum of the acquired phase for different applied biases for (a) Case I and (b) Case
II.

As could be seen in Figs. 2.4 and 2.5, the proposed dual-gated metasurface can
significantly alter the amplitude and phase of the reflected light under an applied bias.
In Fig. 2.6, we plot the phase shift and reflectance as a function of applied bias V0 at a
wavelength of λ = 1550 nm. Figure 2.6a, which corresponds to Case I, shows that this
bias configuration can give a continuously tunable phase shift between 70° and −245°
when the applied voltage is varied between V0= −6.5 V and V0= 6.5 V. This amounts to
a total tunable phase shift of 315° derived from Case I.

Figure 2.6: Simulated reflectance and phase shift of the dual-gated metasurface as a
function of applied voltage V0 at the wavelength of λ = 1550 nm. (a) Reflectance and phase
shift of the dual-gated metasurface in Case I, in which there is a simultaneous charge
accumulation or simultaneous charge depletion in the ITO layer at both ITO/HAOL interfaces.
(b) Reflectance and phase shift of the dual-gated metasurface in Case II, in which the charge
accumulation at the top ITO/HAOL interface is always accompanied by the charge depletion
at the bottom ITO/HAOL interface and vice versa. The insets schematically show the charge
distribution in the dielectric spacer of the metasurface.

43
As expected, a phase shift derived from Case II is invariant with respect to the
transformation V0→− V0 (Fig. 2.6b). In Case II, the phase shift smoothly varies between
0 and −275°, when the applied voltage is increased from V0= 0 V to V0= 6.5 V. Thus,
via an appropriate bias application, the proposed dual-gated tunable metasurface can
allow us to attain a tunable phase shift of 345°.

2.3.3. Distribution of Electromagnetic Fields in the Dual-Gated Metasurface
To gain further insight, we study the distribution of the electric and magnetic fields
inside and around the dual-gated active metasurface. In Fig. 2.7, we plot the distribution
of the absolute value of the optical electric field in the metasurface element at the
resonant wavelength of λ =1550 nm. Figure 2.7a shows the spatial distribution of the
optical electric field at zero bias. The bottom part of Fig. 2.7a shows the magnified
region of the dielectric spacer at zero bias. When a DC bias is applied, we observe a
significant variation of the distribution of the optical electric field. Figure 2.7b shows
the optical electric field distribution in Case I at an applied voltage of V0 = 6.5 V. As
seen in Fig. 2.7b, the optical electric field is enhanced at both the top and bottom
ITO/HAOL interfaces due to the ENZ regions that are formed at these interfaces.

Figure 2.7: Spatial distribution of the electric field intensity within the dual-gated
metasurface for Case I and Case II. (a) Distribution of the magnitude of the electric field
inside the metasurface element at no applied bias. The bottom part of (a) shows the magnified
image of the field distribution in the HAOL/ITO/HAOL dielectric spacer of the metasurface.
The close-up of the distribution of the electric field magnitude in the dielectric spacer of the
metasurface when (b) V0 = 6.5 V in Case I, (c) V0 = −6.5 V in Case I, (d) V0 = 6.5 V in Case II,
(e) V0 = −6.5 V in Case II. The operating wavelength is λ =1550 nm.

44
On the other hand, when the applied DC bias in Case I is equal to V0 = −6.5 V, the ITO
layer is depleted at both top and bottom interfaces (Figs. 2.7c), and therefore we do not
observe a significant optical field enhancement in the ITO layer.
In Case II, however, a dramatic optical field enhancement is observed at both positive
and negative applied biases V0= ±6.5 V (Figs. 2.3d, e). In this case, at an applied bias
of V0 = 6.5 V, we observe the optical electric field enhancement in the ITO layer around
the top ITO/HAOL interface due to the ENZ region formed in the ITO layer (Fig. 2.7d).
Similarly, Fig. 2.7e shows that in Case II, the optical electric field is enhanced around
the bottom part of the ITO layer, when the applied bias is equal to V0 = −6.5 V. The
analysis of the optical field profile suggested that strong light confinement in the
dielectric gap of the plasmonic antenna significantly contributes to the observed optical
modulation.
Figure 2.8 shows the spatial distribution of the z-component of the electric field Ez
inside the dielectric spacer of the metasurface, which consists of HAOL/ITO/HAOL
planar layers. The spatial distribution of Ez is calculated at a wavelength of  =1550
nm. Figures 2.8a-c correspond to the bias application configuration that is referred to
as Case I, while Figs. 2.8d-f correspond to the bias application configuration referred
to as Case II. In both Case I and Case II, the assumed values of the applied bias are V0
= −6.5 V, V0 = 0 V, and V0 = 6.5 V. As seen in Figs. 2.8c, d, and f, one can observe a
strong field enhancement at the interfaces of ITO and HAOL.

Figure 2.8: Spatial distribution of the z component of the electric field within the dualgated metasurface for Case I and Case II. Close-up image of the spatial distribution of the z
component of electric field in the HAOL/ITO/HAOL region at the wavelength of λ =1550 nm
for (a) V0 = −6.5 V, (b) V0 = 0, and (c) V0 = +6.5 V in Case I, (d) V0 = −6.5 V, (e) V0 = 0 V,
and (f) V0 = +6.5 V in Case II.

45
Figure 2.8 also shows that the z component of the electric field Ez around the right and
left edges of the antenna are antiparallel to each other.
Figure 2.9 plots the spatial distribution of the absolute value of the magnetic field for
our dual-gated metasurfaces at the wavelength of λ =1550 nm. Figures 2.9a-c
correspond to Case I, while Figs. 2.9d-f correspond to Case II. In both Case I and Case
II, we assume the following values of applied bias voltages: V0 = −6.5 V, V0 = 0 V
and, V0 = 6.5 V. As seen in this figure, the magnetic field is localized in the gap region
between the Al antenna and the back reflector. This proves the existence of a magnetic
dipole resonance. One can also notice that the strength of the magnetic dipole is
strongly altered by changing the applied voltage.

Figure 2.9: Spatial distribution of the magnetic field intensity within the dual-gated
metasurface for Case I and Case II. Spatial distribution of the magnitude of the magnetic
field at the wavelength of λ =1550 nm for (a) V0 = −6.5 V, (b) V0 = 0, and (c) V0 = +6.5 V in
Case I, (d) V0 = −6.5 V, (e) V0 = 0 V, and (f) V0 = +6.5 V in Case II. The horizontal dashed
lines from bottom to the top specify the boundaries between the back reflector and bottom
HAOL, the bottom HAOL and the ITO layer, the ITO layer and the top HAOL, and the top
HAOL and the antennas. The vertical dashed lines outline the patch antenna.

2.4. Fabrication of Dual-gated Tunable Metasurface
Having identified an approach to metasurface design, we fabricate and characterize the
tunable optical response of the dual-gated metasurface. In order to fabricate our gatetunable metasurface, we first perform RCA1 cleaning (H2O: NH4OH: H2O2 = 5: 1: 1)
of 100 Si substrates. We then deposit an 80 nm-thick Al back reflector using e-beam
evaporation (see Appendix A.3). On top of the Al back reflector, we deposit a 9.5 nm-

46
thick HAOL using atomic layer deposition (ALD), as described in Appendix A.2. Next,
we deposit a 5 nm-thick ITO layer on top of the HAOL gate dielectric by using RF
magnetron sputtering in Ar/O2 plasma environment (see Appendix A.4 for fabrication
and characterization of the ITO layer). To characterize our ITO films, we perform Hall
measurements and spectroscopic ellipsometry on 5 nm-thick ITO layers deposited on
quartz and Si substrates, respectively. Once we sputtered the ITO layer, we deposit
another 9.5 nm-thick HAOL layer as the top gate dielectric. Afterward, we spin
electron-beam resist (EBR) on our Si/Al/HAOL/ITO/HAOL planar sample and pattern
fishbone antenna arrays as well as the contact pads via standard electron-beam
lithography (EBL). After developing the e-beam-exposed sample, we deposit 40 nmthick Al by using e-beam evaporation. The dual-gated metasurface is obtained after
performing the lift-off process. Figure 2.10 summarizes the described fabrication steps
of our tunable metasurface.

Figure 2.10: Schematic representation of fabrication steps of the dual-gated tunable
metasurface. Following the direction of the arrows, the fabrication steps can be
summarized as follows: RCA1-cleaning of Si substrates, deposition of Al back reflector by
e-beam evaporation, deposition of bottom HAOL gate dielectric by ALD, sputtering of the
ITO layer, deposition of top HAOL gate dielectric by ALD, spinning bi-layer e-beam resist,
patterning the top antennas with e-beam lithography and development of the exposed resist,
deposition of the Al antennas by e-beam evaporation, and lifting-off the excessive Al and
the resist.

47
It is noteworthy that during fabrication, our samples are patterned to allow for easy
application of a bias between the Al back reflector and the ITO layer (Vb). The Al
fishbone antennas are connected to an external Al pad that allows for facile bias
application between the fishbone antennas and the ITO layer (Va). The electrode pads
are then wire bonded to a compact chip carrier and circuit board for electrical gating.

2.5. Experimental Demonstration of the Dual-gated metasurface
Once we fabricated the dual-gated metasurface as discussed in the previous section, we
do amplitude and phase measurements on the fabricated devices. Optical measurements
are performed by illuminating our metasurfaces with linearly polarized light with
incident electric field aligned with the fishbone antenna (x-direction in Fig. 2.1a). In
our experiments, the bias configuration corresponds to Case I, when Va = V0 and Vb =
V0 (Fig. 2.3a).
2.5.1. Reflectance measurements
Figure 2.11 shows the experimental setup that we use for reflectance measurements.
In our reflectance measurement setup, the metasurface is illuminated by a broadband
laser. The laser beam impinges on the metasurface after passing through an optical
chopper, a polarizer, a 50/50 non-polarizing beam splitter, and a 20x objective lens.
We use a white light source, a flipping mirror, and a charge-coupled device (CCD)
camera to make sure that the laser beam is positioned at the center of the metasurface.
The light reflected from the metasurface is then guided to a germanium (Ge) detector
by the 50/50 beam splitter.

Figure 2.11: Optical setup for measuring the reflectance spectrum of our dual-gated
metasurfaces. A NIR laser beam is shone on the metasurface, and the reflected beam is
collected by a detector.

48
The reflectance is then obtained via
𝑅𝑒𝑓𝑙𝑒𝑐𝑡𝑎𝑛𝑐𝑒 [%] = 100 ×

𝑅𝑚𝑒𝑡𝑎𝑠𝑢𝑟𝑓𝑎𝑐𝑒 −𝑅𝑏𝑎𝑐𝑘𝑔𝑟𝑜𝑢𝑛𝑑
𝑅𝑟𝑒𝑓𝑒𝑟𝑒𝑛𝑐𝑒 −𝑅𝑏𝑎𝑐𝑘𝑔𝑟𝑜𝑢𝑛𝑑

(2.3)

where 𝑅𝑚𝑒𝑡𝑎𝑠𝑢𝑟𝑓𝑎𝑐𝑒 and 𝑅𝑟𝑒𝑓𝑒𝑟𝑒𝑛𝑐𝑒 are the raw reflectance values obtained while
illuminating the metasurface and the Al back reflector, respectively. 𝑅𝑏𝑎𝑐𝑘𝑔𝑟𝑜𝑢𝑛𝑑 is
the background reflectance in the absence of the incident laser beam.
Figure 2.12 displays the measured reflectance spectra at different values of applied
voltage V0. The resonance is observed to blueshift with increasing the voltage from V0
= 0 V to V0 = 2.5 V. When we increase the applied voltage from V0 = 2.5 V to V0 = 6.5
V, the resonance is observed to redshift, and likewise, when we decrease the applied
bias from V0 = 0 V to V0 = −6.5 V, a resonance redshift is observed. These observations
are consistent with our simulation results (Fig. 2.4) which indicate that at an applied
voltage of V0 = 2.5 V, we reach the ENZ region in the ITO accumulation layer,
corresponding to the transition from resonance blueshift to redshift. Figure 2.12b
depicts the spectrum of the relative reflectance change obtained from
𝛥𝑅
𝑅0

𝑅(𝑉)−𝑅(𝑉=0)
𝑅(𝑉=0)

(2.4)

As can be seen, even though the measured reflectance modulation ΔR/R 0 is fairly
broadband, we observe an enhancement in ΔR/R 0 around the metasurface resonant
wavelength. The inset of Fig. 2.12b shows the relative reflectance modulation ΔR/R 0
as a function of applied bias V0 at a wavelength of  = 1550 nm. At a wavelength of 
= 1550 nm, the relative reflectance modulation is much more significant for negative
rather than positive bias voltages.
When V0 decreases from V0 = 0 V to V0 = −6.5 V, we observe a relative reflectance
modulation of 89% at the wavelength of  = 1550 nm. On the other hand, when the
applied voltage V0 increases from V0 = 0 V to V0 = 6.5 V, we observe that the relative
reflectance modulation is only 28%. This implies that at the resonance wavelength, the
formation of multiple ITO charge depletion layers influences the reflectance more
significantly than the formation of multiple charge accumulation layers. This change in
reflectance can be explained by the modulation of the optical electric field in the ITO
layer under applied bias. At V0 = 0 V, the formation of the magnetic dipole leads to
strong absorption in the ITO layer, whereas the applied bias effectively modulates the

49
optical field distribution in the dielectric spacer of the metasurface and leads to a higher
reflectance.
Figure 2.12c displays the reflectance as a function of applied voltage V0 at a wavelength
of  = 1550 nm. As can be seen, a decrease in the reflectance level is observed when
the applied voltage increases from V0 = −6.5 V to V0 = 0.6 V. Once the applied bias
passes V0 = 0.6 V, we observe an increase in the reflectance.

Figure 2.12: Measured amplitude response of the dual gated metasurface for Case I. (a)
Measured reflectance spectra in Case I at different applied voltages V0. (b) The spectra of the
relative reflectance modulation for different applied voltages V0. The inset shows the relative
reflectance change at the wavelength of λ = 1550 nm as a function of applied bias V0. (c)
Measured reflectance in Case I as a function of applied bias V0 and at the fixed wavelength of
λ = 1550 nm.

2.5.2. Phase measurements
After performing reflectance measurements on the gate-tunable metasurface and
identifying the resonance wavelength, we measure the phase shift of the reflected light
under applied bias. For our phase shift measurements, we employ a Michelson
interferometer-type measurement setup shown in Fig. 2.13.
In our interferometer, the beam from a NIR tunable laser is directed towards the sample
via a polarizer, a 50/50 non-polarizing beam splitter, and then a 20x objective lens. By

50
using a white light source and a CCD camera, we ensure that the laser light illuminates
the edge of the metasurface. Therefore, a part of the incoming beam is reflected from
the metasurface, while the other part is reflected from the surrounding planar
Al/HAOL/ITO/HAOL heterostructure, which acts as a built-in phase reference. The
light reflected from the sample is guided to a NIR camera by using a 50/50 beam splitter.
At the same time, the incident laser beam itself is directed to the camera by using
mirrors, where it serves as a reference beam. The images of the formed interference
fringes are recorded by the camera. The camera records two regions of interference
fringes i) the fringes formed via interference of the light reflected from the metasurface
and the reference beam, and ii) the fringes formed via interference of the light reflected
from the HAOL/ITO/HAOL/Al planar heterostructure and the reference beam. Our
interferometry enables accurate phase measurements since, in this configuration, errors
caused by vibrations and other motion instabilities are eliminated.

Figure 2.13: Optical setup used for measuring the phase shift of the light reflected from
the tunable dual-gated metasurfaces. The fringes obtained from the interference of the
incident and the reflected beams are recorded by a NIR camera, and subsequently processed
and fitted. Our fitting procedure enables us to retrieve the relative displacement of the
interference fringes originating from the metasurface and the reference when we apply a bias.

To analyze the phase shift of the light reflected from our tunable metasurface, we
process the images captured by the camera under different applied biases. In these
images, we select one spatial cross-section from the metasurface interference fringes
and another one from the reference fringes area. The intensity values at the crosssections are then interpreted as the curves which are then smoothened by a moving
average (MA) filter.

51
Figure 2.14a shows the interference fringes recorded by the NIR camera for different
applied bias voltages.

Figure 2.14: The Michelson interferometer-type measurement results used to extract
the phase shift of the light reflected from the dual-gated metasurface when changing
the applied bias. (a) The interference fringe patterns captured by the NIR camera. The
dashed lines labeled “R” and “M” show the reference and metasurface fringe cross-sections,
respectively. (b) Extracted intensity data from reference and metasurface fringe crosssections and their MA smoothened curves. (c) Fitted sinusoidal waves for reference and
metasurface fringe cross-sections. Here, 𝛥𝑝 is the distance between the two fixed peaks of
the sinusoidal functions fitted from metasurface and reference fringe cross-sections, and 𝑝0
is the period of the sinusoidal wave, respectively.

52
The intensity data extracted from the reference fringes (indicated by “R”) and
metasurface fringes (indicated by “M”) is depicted in Fig. 2.14b. Figure 2.14b also
plots MA-filtered curves for both the reference and metasurface fringes. Figure 2.14c
shows the sinusoidal functions fitted to the two mentioned fringe regions.
Considering the offset between these two sinusoidal functions, one can calculate the
phase shift for each applied bias via
𝛥𝑝

𝑃ℎ𝑎𝑠𝑒 𝑠ℎ𝑖𝑓𝑡 = 𝑝

(2.5)

where 𝛥𝑝 is the distance between the two fixed peaks of sinusoidal functions that
correspond to the metasurface fringes and reference fringes, and 𝑝0 is the period of
the sinusoidal wave. Accordingly, the measured interference fringe displacements are
converted into a relative phase shift.
Figure 2.15 shows the measured phase shift values as a function of applied bias
voltage V0 at the laser illumination wavelength of  = 1550 nm. Examples of
interference fringe images recorded at bias voltages of V0 = −6.5 V and V0 = +6.5 V
are shown in the inset of Fig. 2.15. The dashed white lines show the interference
fringes from the metasurface (M) and the reference (R). When we increase the applied
voltage from V0 = 0 V to V0 = +6.5 V, we observe a phase shift of −211.9°, which is
accompanied by a modest relative reflectance modulation of 28% (Fig. 2.12c).

Figure 2.15: Measured phase response of the dual-gated metasurface for Case I. Measured
phase shift of the metasurface as a function of applied voltage V0 at a wavelength of λ = 1550
nm. The insets show the interference fringes at −6.5 V and +6.5 V. The dashed lines labeled as
“R” and “M” indicate the interference fringes from the reference and metasurface, respectively.

Moreover, decreasing the applied voltage from V0 = 0 V to V0 = −6.5 V, we measure a
phase shift of +91° that is consistent with our simulation results shown in Fig. 2.6a.

53
Interestingly, despite the modest phase shift recorded at negative biases V0 < 0, the
reflectance measured at a wavelength  = 1550 nm increases from 13% to 30%.
As can be seen, the overall phase shift of 303° is produced as the applied bias to the
dual-gated metasurface is varied between V0 =−6.5 V and V0 =+6.5 V.

2.6. Theoretical Demonstration of Beam Steering Using the Dual-Gated
Metasurfaces
In the present section, we use the experimentally derived phase shift and reflectance
values to theoretically estimate the beam steering performance of our dual-gated
metasurface. To steer the beam, we use a blazed grating approach. This implies that for
each steering angle, we create a periodic phase pattern along the metasurface. In this
approach, the period of the phase pattern is appropriately chosen to match the desired
steering angle. We use the framework of antenna array theory to analytically calculate
the far-field radiation pattern of our metasurface. In our analytical approach, we
incorporate experimentally measured reflectance and phase shift values into the
calculations of the far-field radiation pattern.
Within the scope of the Fraunhofer approximation [36], [89], the far-field intensity can
be analytically given as
𝐼(𝜃) = |𝐸𝑝𝑎𝑡𝑡 |2 × |𝐴𝐹|2

(2.6)

where 𝐸𝑝𝑎𝑡𝑡 is the far-field radiation pattern of a single metasurface element, while AF
stands for the array factor. θ is the observation angle defined with respect to the z-axis.
Note that the reflected beam is steered in the z-x plane (see Fig. 2.1a). In our case, the
array factor can be written as
𝑖((𝑗−1) 𝑘 𝑑 sin(𝜃)+𝛷𝑗 )
𝐴𝐹 = ∑𝑁
𝑗=1 √𝑅𝑗 𝑒

(2.7)

Here, j numerates the emitter, and N denotes the total number of emitters in the
metasurface. k denotes the free space wavenumber k = 2π/λ, and d gives the distance
between neighboring emitters. Therefore, (𝑗 − 1)𝑘 𝑑 sin(𝜃) gives the phase difference
conditioned by the path length difference to the observation point due to different
emitter positions. 𝑅𝑗 gives the reflectance of the jth emitter, which we extract from our

54
measurement results (Fig. 2.12). 𝛷𝑗 is the actively controlled emitter-imparted phase
that controls the beam deflection (Fig. 2.15).
In this approach, an emitter is defined as a scatterer, which scatters light with a given
phase. For example, an emitter can be composed of a single or multiple metasurface
elements, depending on the voltage application configuration. The array factor captures
the most important features of the far-field radiation pattern, such as the steering angle
and the width of the steered beam. Since our emitter is relatively omnidirectional, we
set |𝐸𝑝𝑎𝑡𝑡 | = 1, and in what follows analyze the antenna factor only. In our calculations
we assume that our metasurface consists of 100 elements, implying that in the case of
our dual-gated metasurface, we can apply 200 independent bias voltages.
As mentioned above, we use a blazed grating approach to steer the beam. In this
approach, we create a constant phase gradient along the metasurface, effectively
creating a periodic phase pattern. Table 2.1 summarizes the experimentally derived
voltages, which yield the relative phase shifts equal to 0°, 90°, 180°, and 270°. The
corresponding measured reflectance values are also shown in the last column of Table
2.1. We define a blazed grating that consists of four metasurface elements with phase
shifts of 0°, 90°, 180°, and 270°. Since we assume that we have N = 100 metasurface
elements, we can effectively create 25 blazes. In this case, the parameter d equals to the
period of our metasurface d = 400 nm. Figure 2.16a shows the calculated far-field
radiation pattern. As can be seen, in this case, the beam is steered to a steering angle of
72°.
In the next step, we again employ four independent phase levels. However, now, we
define a single blaze as: 0°, 0°, 90°, 90°, 180°, 180°, 270°, and 270°. Note, that now in
Eq. (2.7), we set d = 800 nm, and N = 48, implying that in this calculation, we assume
that the metasurface consists of 96 elements. In this case, our blazed grating consists of
12 blazes, and the steering angle is 28° as shown in Fig. 2.16b.
Table 2.1: Experimentally measured phase and reflectance values. The reported values
of V0 provide phase shift in 90o steps.
Phase [o]
-198.6
-108.6
-18.6
71.4

Relative Phase [o]
90
180
270

Voltage V0 [V]
3.5
1.3
0.28
-3.1

Reflectance [%]
18.14
13.51
8.39
21.58

55
We can also use six independent phase levels to demonstrate beam steering. Table 2.2
summarizes the experimentally derived voltages, which yield the phase shifts equal to
0°, 60°, 120°, 180°, 240°, and 300°, which we use to define a unit blaze. In this case, d
= 400 nm, and N = 96. The experimentally measured reflectance values are shown in
the last column of Table 2.2. We can now apply 16 metasurface blazes that yield the
steering angle of 39° as shown in Fig. 2.16c.
Table 2.2: Experimentally measured phase and reflectance values. The reported values of
V0 a provide phase shift in 60o steps.
Phase [o]
-211
-151
-91
-31
29
89

Relative Phase [o]
60
120
180
240
300

Voltage V0 [V]
5.5
2.5
1.35
0.46
-1
-5.35

Reflectance [%]
19.64
16.67
13.42
13.78
17.75
26.85

Figure 2.16: Calculated beam steering results using the dual-gated tunable metasurface.
The far-field radiation pattern for different voltage application configurations with phase
profiles of the metasurface being (a) 0°, 90°, 180°, and 270°, repeated 25 times, (b) 0°, 0°, 90°,
90°, 180°, 180°, 270°, and 270°, repeated 12 times, (c) 0°, 60°, 120°, 180°, 240°, and 300°,
repeated 16 times, and (d) 0°, 0°, 60°, 60°, 120°, 120°, 180°, 180°, 240°, 240°, 300°, and 300°,
repeated 8 times. The inset of each subfigure indicates the angle at which the beam is steered.

56
As a final example, we define a unit blaze as 0°, 0°, 60°, 60°, 120°, 120°, 180°, 180°,
240°, 240°, 300°, and 300°. Thus, a single blaze consists of 12 metasurface elements.
Note that in this case, we set d = 800 nm, and N = 48. When this blaze is periodically
applied, the reflected beam is steered to an angle of 19o. The far-field radiation pattern
of such blazes is shown in Fig. 2.16d.
It should be noted that the steering angle θ obtained from Eqs. (2.6) and (2.7) can also
be calculated by using a simple grating equation

sin(𝜃) = 𝐿

(2.8)

Here, L denotes the length of the blaze.
Figure 2.17 plots the steering angle as a function of the period of the blazed grating
obtained using Eq. (2.8). In Fig. 2.17, the red dots correspond to the blaze periods,
which have been used to produce far-field radiation patterns shown in Fig. 2.16.
Dots a, b, c, d, correspond to Figs. 2.16a, b, c, and d, respectively. The steering angles
predicted by the grating equation (2.8) match well with the steering angles obtained
from the calculations of the far-field radiation pattern. Thus, by using the blazed grating
approach, we can also steer the beam to the other angles shown in Figure 2.17.

Figure 2.17: Calculated steering angles using a blazed grating approach. Steering angle as
a function of the period of the blazed grating. Red dots correspond to the blaze periods, which
were used to produce far-field radiation patterns shown in Fig. 2.16. Dots a, b, c, and d
correspond to Fig. 2.16a, b, c, and d, respectively.

57
2.7. Comparison to Single-gated Tunable Metasurface
To confirm that it is advantageous to use dual-gated metasurfaces as compared to
single-gated ones, we calculate the phase shift of the light reflected from the
metasurface, when only the ITO layer and the fishbone antennas are biased with respect
to each other. Figure 2.18a shows the reflectance as a function of wavelength and
applied bias. Figures 2.18b, c plot the reflectance and relative reflectance change
spectra for different applied voltages.

Figure 2.18: Comparison between the dual-gated metasurface and the single-gated
counterpart. (a) Reflectance from the single-gated metasurface as a function of wavelength
and applied bias, (b) reflectance, and (c) relative reflectance change spectra for different applied
voltages. (d) Reflectance from the single-gated metasurface as a function of applied bias
voltage for three different wavelengths close to the resonance wavelength. (e) Spectra of the
acquired phase for different applied biases. (f) Phase shift as a function of applied voltage at
different wavelengths.

The reflectance as a function of applied bias voltage for three different wavelengths
close to the resonance wavelength is depicted in Fig. 2.18d. Figure 2.18e shows the
spectrum of the acquired phase, and the phase shift as a function of applied voltage at

58
the wavelengths of  =1545 nm,

 =1550 nm, and  =1555 nm are plotted in Fig.

2.18f. As can be seen, when we change the voltage from −6.5 V to +6.5 V, the phase
shift changes from 58° to -212.8° when  =1545 nm, from 75.6° to −185.2° when 
=1550 nm, and from 93° to −156.8° when  =1555 nm.
Therefore, the maximum achievable phase shift for the single-gated metasurface is
~271°, which is 74° smaller than the phase shift obtained from the dual-gated
metasurface. We note that the dual-gated metasurface exhibits an asymmetric response,
enabling a given phase shift to be achieved via multiple different bias configurations,
and hence, greater flexibility in system design for beam steering.

2.8. Conclusions and Outlook
In this chapter, we presented the design and experimental demonstration of a dual-gated
plasmonic reflectarray metasurface that shows wide phase tunability with applied bias
at a wavelength of λ = 1550 nm, and the phase of the reflected light can be continuously
tuned from 0 to 303°. We showed a measured relative reflectance modulation of 89%.
This large optical tunability is achieved both due to the materials employed here and to
the dual-gated metasurface architecture. Each element of our dual-gated metasurfaces
can be viewed as two series-connected MOS field-effect structures to which two
independent bias voltages can be applied, yielding a wider phase tuning range
compared to a single-gated metasurface. Interestingly, in our metasurface, a given
phase shift can be achieved via multiple different bias configurations that yield different
reflectance values, enabling an approach for reflectance modulation at a constant phase.
This feature may be very useful for the design and demonstration of future dynamically
reconfigurable low-profile optical components such as focusing lenses with
reconfigurable focal lengths, dynamic holograms, and beam steering devices.

59
Chapter 3

ELECTRO-OPTICALLY TUNABLE MULTIFUNCTIONAL
METASURFACES
The material in this chapter was in part presented in [95].
Shaping the flow of light at the nanoscale has been a grand challenge for nanophotonics
over decades. It is now widely recognized that metasurfaces represent a chip-scale
nanophotonics array technology capable of comprehensively controlling the wavefront
of light via appropriately configuring subwavelength antenna elements. In this chapter,
we demonstrate a reconfigurable metasurface that is multifunctional, i.e., notionally
capable of providing diverse optical functions in the telecommunication wavelength
regime, using a single compact, lightweight, electronically-controlled array with no
moving parts. By electro-optical control of the phase of the scattered light from each
identical individual metasurface element in an array, we demonstrate a single prototype
multifunctional programmable metasurface that is capable of both dynamic beam
steering

and

reconfigurable

light

focusing.

Reconfigurable

multifunctional

metasurfaces with arrays of tunable optical antennas thus can perform arbitrary optical
functions by programmable array-level control of scattered light phase, amplitude, and
polarization, similar to dynamic and programmable memories in electronics.

3.1. Introduction
Wavefront shaping in traditional optical elements is usually accomplished by gradual
phase changes along the optical paths through spatially varying either the refractive
index profile or the surface topography. Contrarily, the abrupt changes introduced by
metasurfaces to the amplitude, phase, or polarization of the scattered light allows highly
precise manipulation of the wavefront with subwavelength resolution and enables an
unprecedented manipulation of the propagating electromagnetic waves [1]. Rapid
advances in the control of the phase and amplitude of the light scattered from planar
arrays of nanophotonic elements have stimulated the development of metasurfaces that
utilize amplitude/phase-sensitive scattering to enable wavefront engineering [2], [3].
Owing to their well-engineered meta-atoms that act as optically thin scatterers,

60
metasurfaces enable unconventional change of the local state of the interacting light.
This results in several merits such as ultra-compactness, and precise control over lightmatter interaction at the deep subwavelength scale, that are unavailable to conventional
optical elements [96], [97].
A great number of research endeavors accomplished to date, however, have been
mainly devoted to static metasurfaces in which the fixed geometry and compositions of
their constituents results in an inflexible response of the metasurfaces once fabricated.
As a result, developing reconfigurable paradigms is becoming indispensable to elicit a
transition from static to dynamic metasurface devices, and efficaciously harness the
unlimited opportunities that metasurfaces can offer.
Dynamical control of the properties of the scattered light is possible by using tunable
metasurfaces, for which external stimuli can give rise to changes in the dielectric
function of the metasurface elements [42], thereby modulating the antenna phase and
amplitude response. By employing the mentioned external stimuli, one can achieve
tunable metasurfaces by incorporating active materials such as transparent conductive
oxides (TCOs) like ITO [34], [35], [46], [67] or titanium nitride (TiN) [98], liquid
crystals [52], [99], 2-D materials such as graphene [36] and MoS2 [100], [101], phasechange materials [102] and ferroelectrics [103] into metasurfaces. Dynamically
manipulating the incident light can hence initiate extensive innovations for the
application of metasurfaces in optoelectronic devices. Among these active platforms,
TCOs that undergo a reliable and reproducible index change in response to an optical
or electrical stimulus, provide high modulation speed, low energy consumption,
robustness, and wide tuning range, leading them to establish superiority over other
active materials.
The ability of metasurfaces to spectrally, temporally, or spatially manipulate the
wavefront of light with very high spatial resolution is expected to accelerate the
miniaturization of optical devices and integration of optical systems. Despite several
studies conducted on actively reconfigurable metasurface devices to date, developing
an active metasurface platform operating in the NIR wavelength range that would
dynamically tailor the wavefront of scattered light through a pixel-by-pixel
configuration is still remaining an outstanding research challenge. Moreover,
multifunctional reprogrammable metasurface components have not yet been
demonstrated. The realization of a single hardware device that can provide multiple and

61
indeed general functions would further accelerate the impact of metasurfaces and their
applications. Such multifunctionality can be found in electronics technology that has
benefitted from the development of programmable and reprogrammable circuits
composed of identical circuit elements, such as dynamic [104] and static [105] random
access memories and field-programmable gate arrays [106].
In this chapter, we propose a state-of-the-art prototypical platform for a multifunctional
metasurface that could be electronically programmed to achieve multiple optical
functions. As a proof of principle, we utilize our multifunctional metasurface to
demonstrate two of the most essential functions identified to date for metasurfaces,
namely, beam steering and focusing of light.
Optical beam steering is the key element of a broad range of optical systems such as
LiDAR [107], optical interconnects [108], and optical communications [109].
Conventional beam steering devices such as Risley prisms [110], galvanometerscanning mirrors [111], and decentered lenses [112] employ mechanically moving
optical components to steer the incident light. Although mechanical beam steering
systems provide wide steering angular range and a large number of resolvable beam
directions, they suffer from low steering speed due to the inertia of their moving parts
and the weight of their mechanical components [113]. The availability of electronic
beam steering arrays at NIR wavelengths with scanning frequencies above the MHz
range could replace mechanical components with compact and lightweight
optoelectronic alternatives and enable diverse functions unachievable via mechanical
motion.
Reconfigurable metasurfaces have recently been employed to provide dynamic beam
steering in the microwave and NIR regimes by exploiting microfluidic flows [114],
[115], incorporation of phase-change materials [55], and reorientation of liquid crystals
[116]. However, the performance of these devices is limited due to their failure to
provide exquisite control over the phase of the scattered light and accurately generate
a desired phase profile, leaving them unable to demonstrate arbitrary functions.
Alternatively, electro-optic modulation in multiple-quantum-well (MQW) resonant
metasurfaces [117], [118], an intrinsically ultrafast process, has been shown to provide
high-speed modulation and dynamic beam steering, but to date, a limited phase
modulation range has constrained the achievable beam directivity and steering angle
range.

62
Electro-optically controllable beam-switching has also been demonstrated via the
incorporation of TCOs as active material into metasurfaces [35], [62], [119]. However,
individual control over each metasurface element, which is required for more complex
phase distribution patterns, has not been reported. Other researchers have demonstrated
beam steering using waveguide-based thermo-optical phase shifters coupled to
antennas [60], [120]–[123], or by employing frequency-gradient metasurfaces [124].
These chip-based antenna arrays can enable beam steering at visible or infrared (IR)
frequencies, but are application-specific and, hence, have been unable to achieve more
general array functions.
Light focusing is another paramount optical function that plays a fundamental role in
almost every optical system such as imaging, microscopy, optical data storage, and
optical encryption [125]. In order to focus a light beam, its amplitude or phase
distribution has to be spatially varied. In conventional convex lenses, this functionality
is achieved by controlling the optical thickness of materials such as glass to introduce
suitable phase delays. Metasurfaces have given rise to versatile metalenses that can
replace bulky conventional lenses by achieving constructive interference at a focal
point via introducing desired phase differences at certain distances from their center.
This is indeed done by engineering the spatial variation of field amplitude or phase
distribution over arrays of individual metasurface elements at approximately
wavelength-scale or smaller spacing [5], [14], [126]–[128]. Metalenses have
demonstrated the capability to perform high-resolution imaging, wavefront shaping for
aberration correction, and polarization conversion [4].
Reconfigurable metasurfaces have been utilized to realize dynamic focusing by
variation of the overall lens optical thickness or curvature, via liquid crystal
reorientation [129], microfluidic flow [130], [131], or elastic deformation [38].
However, these modes of dynamic focusing do not permit precise tailoring of the lens
focal properties by arbitrary phase control of the lens phase elements.
In this chapter, we design and demonstrate an electro-optically tunable multifunctional
metasurface that can exhibit multiple functions in the NIR wavelength regime using a
single device, via precise tailoring of the phase profile of an optical aperture. In contrast
to the metasurface presented in the previous chapter, where the same bias voltage was
applied to all metasurface elements, the newly-proposed dynamic metasurface platform
allows independent manipulation of addressable subwavelength pixels.

63
Figure 3.1a schematically illustrates our proposed multifunctional metasurface, whose
independently addressable elements enable dynamic control of the wavefront via a
pixel-by-pixel reconfiguration. Using this scheme, we demonstrate a reprogrammable
metasurface whose function can be reconfigured between dynamic beam steering (Fig.
3.1b) and dynamic focusing meta-mirrors, achieving a reconfigurable focal length and
numerical aperture (Fig. 3.1c) by tuning of the gate voltages applied to individual
metasurface elements.

Figure 3.1: Schematic demonstration of the multifunctional metasurface with 96
independently addressable metasurface elements. Schematic of (a) the multifunctional
metasurface whose functionality can be switched between (b) dynamic beam steering and (c)
cylindrical metalens with reconfigurable focal length.

3.2. Design of Electro-Optically Tunable Metasurface Element
Our active metasurface operates by virtue of electrically-tuning the coupling between
a plasmonic metasurface and the ENZ modes via individual control over the
metasurface elements. Plasmonic metasurfaces have recently proven to achieve strong
light-matter interaction through coupling to ENZ modes. In the proposed design, the
coupling strength is modified via control of the charge carrier concentrations in an ITO
layer embedded into the metasurface when operating the device in the ENZ wavelength
regime of ITO. In this fashion, the metasurface can provide possibilities to switch

64
between different regimes of the light-matter interaction. As a result, the amplitude and
phase of the light reflected from the metasurface is expected to be continuously tuned
by applying a DC electric field across the ITO layer [34].
Figures 3.2a, b schematically illustrate the building blocks of our tunable gated fieldeffect metasurface, consisting of an Au back-reflector (tb = 80 nm), on top of which an
Al2O3 layer (tAl = 9.5 nm) is deposited. The Al2O3 layer acts as a dielectric spacer,
adding a degree of freedom for the metasurface optical mode profile design. This layer
is followed by deposition of an ITO layer (ti = 5 nm), a gate dielectric (th = 9.5 nm),
and Au fishbone nanoantennas (ta = 40 nm). The gate dielectric is HAOL, a hybrid
material that simultaneously exhibits high breakdown field and high DC permittivity as
described in Appendix B.2 [34]. The period of the metasurface is picked to be p = 400
nm.

Figure 3.2: Unit cell design of the multifunctional metasurface. Schematic of (a) periodic
array and (b) unit cell of the antenna elements. The metasurface is composed of an Au backreflector, an Al2O3 dielectric layer, an ITO layer, and a HAOL gate dielectric followed by an
Au fishbone antenna. The period of the metasurface is p = 400 nm, and the thickness of the
back-reflector, Al2O3, ITO, and HAOL layers are selected to be tb = 80 nm, tAl = 9.5 nm, ti = 5
nm, and th = 9.5 nm, respectively. The width, length, and thickness of the antenna are wa=130
nm, la = 230 nm, and ta = 40 nm, respectively and the width of the electrode is we = 150 nm.

Applying a DC electric bias between the ITO layer and the nanoantennas enables a
reproducible field-effect-induced modulation of the complex refractive index of ITO in
the NIR wavelength range. By altering the applied electric field, we can modulate the
ITO charge carrier density close to the interface of the ITO and the gate dielectric. By
further increasing the applied bias, the real part of the dielectric permittivity in an
accumulation layer located within ITO takes values between -1 and +1, yielding an

65
ENZ condition. In the ENZ regime, the ITO layer permittivity is varied at NIR
wavelengths by changing the applied DC bias.
In order to calculate the optical response of metasurfaces under applied bias, we first
need to extract the properties of the ITO layer. To this end, we obtain the spatial
distribution of charge carriers in the ITO layer by coupling finite difference time
domain optical simulations (FDTD Lumerical) with device physics simulations (Device
Lumerical). In the device physics calculations, we use the mesh size of 0.025 nm which
had been validated by performing careful convergence tests.
In our device physics calculations, the assumed parameters for ITO are set to be the
same as the ones mentioned in Chapter 2. Once the spatial distributions of charge
carriers under different applied biases are identified, we calculate the complex dielectric
permittivity of ITO 𝜀ITO by using the Drude model provided in Eqs. (2.1) and (2.2).
Figures 3.3a-c show the simulated spatial distribution of charge carrier concentration,
real and imaginary part of the permittivity of ITO for different applied biases. As can
be seen in Figs. 3.3b, by changing the applied bias, an ENZ region is observed within
the ITO layer, at a region close to the interface of the ITO and HAOL film.
The proposed reconfigurable metasurface operates by spectrally overlapping the ENZ
regime of ITO and the geometrical antenna resonance. Subsequently, the width and
length of the antenna, and the width of the electrode are designed to be wa=130 nm, la
= 230 nm, and we = 150 nm, respectively so that a magnetic dipole plasmon resonance
occurs at the wavelengths coinciding with the ENZ regime for ITO, operating in the
telecommunication wavelength regime.

Figure 3.3: Voltage-dependent spatial distribution of the ITO properties. Calculated spatial
distribution of (a) charge carrier density, (b) real, and (c) imaginary part of the permittivity of
ITO for different applied biases.

66
As a result of the spectral overlap of the ENZ regime of ITO and the geometrical
resonance of the metasurface, the metasurface is expected to exhibit a large modulation
of the optical response.
3.3. Optical Modulation in Electro-Optically Tunable Metasurface Element
3.3.1. Calculated Optical Response of the Electro-Optically Tunable Metasurfaces
In order to calculate the optical response of the metasurface, we use full-wave
electromagnetic calculations for our tunable metasurface using FDTD Lumerical. Here,
the metasurface is illuminated by a normal-incidence plane-wave with an electric field
along the direction of the antennas. When performing electromagnetic calculations, the
mesh sizes are adjusted in different layers. To fully take the gradual index change in
the accumulation layer of ITO into account, a mesh size of as thin as 0.025 nm is used
in the active region of the ITO layer.
Figures 3.4a, b show the reflectance and phase shift spectra of the metasurface for
different applied biases. Here, the phase shift is defined as a difference between the
phases of the reflected and incident plane-waves calculated at the same spatial point.
As seen in Fig. 3.4a, at all applied biases, resonant dips are clearly observed at
wavelengths close to  = 1500 nm, which is our wavelength of interest. Figures 3.4c, d
illustrate the simulated reflectance and phase shift as a function of applied bias at
different wavelengths.
As can be seen, when the external bias is changed, we observe a reflectance change that
is accompanied by significant phase modulation. This demonstrates that both the real
and imaginary parts of the refractive index of the active region in the ITO layer are
modulated by the applied bias.
After we confirmed that our designed metasurface can provide both reflectance and
phase modulation, we can then pick the operating wavelength of the beam steering and
focusing devices. To accomplish this, we plan to utilize the metasurface as a phase
modulator, for which the reflectance should ideally remain constant upon the change in
the applied bias. Figure 3.5 shows the maximum reflectance modulation and the
maximum achievable phase shift at different wavelengths. As can be seen, increasing
the maximum achievable phase shift is accompanied by an increase in the amplitude
modulation which is not desirable for utilization of the metasurface as a phase

67
modulator device. As a result, the operating wavelength of  = 1510 nm is chosen in
order to obtain a phase shift of higher than 270 while the maximum reflectance
modulation remains as small as possible.

Figure 3.4: Calculated amplitude/phase modulation provided by the multifunctional
metasurface. Simulated (a) reflectance and (b) phase shift spectra at different bias voltages.
Simulated (c) reflectance and (d) phase of the reflection from the metasurface as a function of
applied voltage for different wavelengths.

Figure 3.5: Selecting the operating wavelength of the multifunctional metasurface.
Simulated maximum achievable reflectance change (square) and phase modulation (diamond)
at different wavelengths.

To gain further insight, we analyze the electromagnetic fields in the metasurface at the
resonance wavelength. The spatial distribution of the electric field intensity, under an
applied bias of −6 V, 0 V, and +6 V are presented in Figs. 3.6a-c. Zoomed-in images

68
of the electric field inside the dielectric spacer of the metasurface, consisting of the
Al2O3/ITO/HAOL heterostructures are shown in the insets. As can be seen, at the
applied bias of +6 V, there is a strong field enhancement in the accumulation region of
the ITO layer which occurs subsequent to the existence of ENZ condition in the
mentioned layer. By leveraging this strong field enhancement in the active region of
the ITO layer, the complex permittivity change can be drastically emphasized.

Figure 3.6: Spatial electric field intensity distribution of the multifunctional metasurface.
The amplitude of the electric field at the operating wavelength of  = 1510 nm for different
voltages of (a) V = −6 V, (b) V = 0, and (c) V = 6 V. The insets show the zoomed-in image of
the Al2O3/ITO/HAOL nano-sandwich.

Our electromagnetic calculations show that the metasurface phase and amplitude
tunabilities are based on an interplay between magnetic plasmon resonance and the
ENZ region of ITO. To show this, the amplitude of the magnetic field inside the
metasurface is shown in Fig. 3.7.
The field distributions are presented at different wavelengths and under different
applied biases. As can be seen in Figs. 3.7d-f, at the resonance wavelength  = 1510
nm, the metasurface supports a large-magnitude magnetic field that is localized in the
gap region between the gold antenna and the gold backplane.
Figure 3.8 shows the spatial distribution of the z-component of the electric field within
the metasurface under different applied biases. As can be seen, the z-components of the
electric field around the right and left edges of the antenna are antiparallel to each other.
This antiparallel field, accompanied by the curl of the current density is consistent with
the large magnetic field shown in Figs. 3.7d-f, indicating that the considered resonance

69
is a magnetic plasmon resonance. We should note that the strength of the magnetic
dipole is strongly modified by altering the applied bias.

Figure 3.7: Spatial distribution of the magnetic field intensity in the multifunctional
metasurface. Spatial distribution of the amplitude of the magnetic field at the operating
wavelength of (a-c)  = 1100 nm, (d-f)  = 1510 nm, and (g-i)  = 1510 nm for different
voltages of (a, d, g) V = −6 V, (b, e, h) V = 0, and (c, f, i) V = 6 V.

Figure 3.8: Spatial distribution of the z-component of the electric field in the
multifunctional metasurface. Spatial distribution of the z-component of the electric field at
the wavelength of  = 1510 nm, under applied bias (a) V = −6 V, (b) V = 0, and (c) V = 6 V.

70
3.3.2. Measured Optical Response of the Electro-Optically Tunable Metasurfaces
After confirming the tunable optical response of the multifunctional metasurface using
simulations, we fabricate the designed structure in order to experimentally obtain the
reflectance and phase shift of the device under applied bias. Figure 3.9 shows the
fabrication steps of the device.

Figure 3.9: Fabrication steps of the multifunctional metasurface. (a) Patterning the
outermost connecting pads, (b) patterning the back reflector, (c) deposing the Al2O3 layer, (d)
patterning the ITO layer, (e) patterning the connecting pads of the ITO layer, (f) depositing the
HAOL gate dielectric, and (g) patterning the antenna array and the inner connecting pads.

To fabricate the multifunctional metasurface device, we first perform a standard
cleaning process on SiO2 (1 m) on Si wafers. Then we spin-coat a layer of S1813
photo-resist on the sample. The outermost parts of the connecting pads as well as some
alignment markers are then patterned using photolithography. After developing the
photoresist, a 10 nm-thick titanium (Ti) layer followed by a 200 nm-thick Au layer is
deposited on the samples using an electron beam evaporator. After lifting-off the excess
Ti-Au parts, we pattern the back reflector by EBL [VISTEC electron beam pattern
generator (EBPG) 5000+] at an acceleration voltage of 100 keV. After developing the

71
electron beam resist, we deposit a 3 nm-thick chromium (Cr) layer followed by an 80
nm-thick Au layer using an electron beam evaporator. After the lift-off process, a 9.5
nm-thick Al2O3 layer is deposited on the samples using ALD through shadow masks.
Then the ITO layer is patterned by EBL, and a 5 nm-thick ITO layer is deposited on
the sample using room-temperature RF magnetron sputtering in Ar/O2 plasma
environment. The deposition pressure is 3 mTorr while the applied RF power is 48W.
Once the excess ITO regions are lifted-off, we pattern the contact pads of the ITO layer
by EBL. After developing the EBR, a 10 nm-thick Ti layer followed by a 200 nm-thick
Au layer is deposited on the samples using an electron beam evaporator. Afterward, a
9.5 nm-thick HAOL layer is deposited on the samples using ALD. The size of the
HAOL film is controlled by using shadow masks during the ALD process. We then
pattern the antennas and the inner contact pad connections by EBL after spinning EBR
on the sample. Once the EBR is developed, we deposit a 2 nm-thick Ge layer followed
by a 40 nm-thick Au layer.
After fabricating the device, we characterize the tunable optical response of the
metasurface by performing amplitude/phase spectrum measurements. In order to
measure all the functions provided by our multifunctional metasurface, we designed
and built a custom optical measurement setup.

Figure 3.10: Universal measurement setup used for reflectance, phase shift, beam steering,
and focusing measurements. NIR-L: NIR laser, M: mirror, I: iris, BS: 50:50 beam splitter, L:
lens, LA: lamp, P: polarizer, ND: neutral density filter, λ/4: quarter-wave plate, SP:
spectrometer, O: objective lens, MTS: metasurface sample, IR-C: IR CCD camera, V-C: visible
CCD camera, MS: 2-axis motorized stage.

72
This measurement setup is capable of measuring the reflectance spectrum, phase shift,
beam steering, and reconfigurable focusing. Figure 3.10 shows our custom-built setup.
Each measurement is performed through a part of our universal setup that will be
discussed in detail in the following sections. For all measurements, we utilize an
uncollimated white light source from a halogen lamp (LA) and a visible CMOS image
sensor camera (V-C) to visualize the sample surface.
Figure 3.11 shows the part of the universal optical setup that is used for reflectance
measurements. For reflectance measurement, the surface of the metasurface sample is
illuminated by an uncollimated white light from a halogen lamp (LA). The incident
light is focused by an objective (O) with a long working distance (Mitutoyo M Plan
Apo 20×, NA = 0.40, WD = 20 mm) after passing through a polarizer (P) and a 50:50
non-polarizing beam splitter (BS). Then the beam reflected from the metasurface is
guided to a spectrometer via a beam splitter (BS) and a mirror (M). The reflectance is
then obtained via

−𝑅

𝑅𝑒𝑓𝑙𝑒𝑐𝑡𝑎𝑛𝑐𝑒 [%] = 100 × 𝑅 𝑠𝑎𝑚𝑝𝑙𝑒 −𝑅𝑑𝑎𝑟𝑘
𝑟𝑒𝑓𝑒𝑟𝑒𝑛𝑐𝑒

𝑑𝑎𝑟𝑘

(3.1)

where 𝑅𝑠𝑎𝑚𝑝𝑙𝑒 is the raw reflectance from the metasurface sample, 𝑅𝑟𝑒𝑓𝑒𝑟𝑒𝑛𝑐𝑒 is the
raw reflectance from a mirror, and 𝑅𝑑𝑎𝑟𝑘 is the background reflectance in the absence
of the incident laser beam.

Figure 3.11: Optical setup used for reflectance measurement. LA: lamp, M: mirror, L: lens,
BS: 50:50 beam splitter, P: polarizer, O: objective lens, MTS: metasurface sample, λ/4: quarterwave plate, SP: spectrometer. The components indicated with red cross marks belong to the
universal setup and are not used in this part of the measurement.

73
In order to experimentally evaluate the phase shift of the beam reflected from the
metasurface, we use a Michelson interferometer system by using a part of the
measurement setup, shown in Fig. 3.12. In this part of the setup, the metasurface sample
is illuminated by a tunable NIR laser. The laser beam illuminates the metasurface after
passing through an iris (I), a 50:50 non-polarizing beam splitter (BS), and an objective
(O) with a long working distance (Mitutoyo M Plan Apo 20×, NA = 0.40, WD = 20
mm). The polarization of the incident beam is controlled by a polarizer (P) to make sure
that the incident electric field is in the direction of the antennas. The reflection from the
sample is then directed towards an IR camera (IR-C) by using the mentioned beam
splitter (BS). In addition to the reflection from the metasurface sample, the laser beam,
which serves as a reference beam, is also guided to the camera by using mirrors (M).
By focusing the incident laser beam on the edge of the metasurface nanoantenna array,
the scattered beam is reflected partly from the metasurface and partly from the Au backplane, resulting in a lateral shift in the interference fringe patterns of the metasurface
and the back-reflector when changing the applied bias. By fitting these two crosssections to sinusoidal functions and obtaining the relative delay between the fitted
sinusoidal curves when changing the applied voltage, we can retrieve the phase shift
acquired due to the applied bias using the method discussed in Chapter 2.

Figure 3.12: Optical setup used for phase shift measurement. NIR-L: NIR laser, M: mirror,
I: iris, BS: 50:50 beam splitter, P: polarizer, ND: neutral density filter, O: objective lens, MTS:
metasurface sample, L: lens, IR-C: IR CCD camera. The components indicated with red cross
marks belong to the universal setup and are not used in this part of the measurement.

74
Figure 3.13 illustrates the measured reflectance (blue curve) and phase shift (red curve)
as a function of applied bias when the same voltage is applied to all metasurface
elements. As seen in Fig. 3.13, at the wavelength of λ=1522 nm, an actively tunable
continuous phase shift of slightly greater than 270o (274), accompanied by a modest
reflectance modulation is obtained. As a result, the operating wavelength of the
fabricated device is picked to be 1522 nm.

Figure 3.13: Measured tunable optical response of the multifunctional metasurface.
Measured reflectance (blue curve) and phase shift (red curve) as a function of applied bias
voltage. The operating wavelength of the fabricated device is chosen to be  = 1522 nm such
that a phase shift greater than 270 accompanied by a moderate amplitude variation could be
obtained.

When analyzing beam steering performance of our multifunctional metasurface in the
following section, we will observe that the reflectance modulation provided by the
metasurface will result in an increased intensity of the undesired side-lobes in the farfield radiation. Moreover, since the complex dielectric permittivity of ITO is
significantly modulated only in a sub-nm-thick layer, a large tunable phase shift is
observed only when the optical field is tightly confined in this sub-nm-thick ITO active
layer. This tight field confinement results in enhanced absorbance, and hence, reduced
reflectance of our active metasurface. By changing the structural parameters of the
metasurface, or using TCOs with higher electron mobilities, such as cadmium oxide
(CdO) [132], one can change the reflectance of the metasurface (see Appendix B.3).

75
3.4. Multifunctional Performance of the Electro-optically Universal Metasurface
Once we validated the modulation performance of the individual metasurface elements,
we investigate the metasurface array beam steering and focusing performance. SEM
images of the fabricated metasurface nanoantennas are shown in Fig. 3.14a. In our
metasurface device, to facilitate an easier and more feasible bias application,
nanoantennas are electrically bus-connected together in one direction, forming
equipotential antenna rows referred to here as a metasurface pixels. Then each pixel is
individually controlled by a separately applied gate voltage. It will be shown in the
following sections how this individual control of the metasurface pixels can lead to the
emersion of a new class of planar photonics components.

Figure 3.14: Fabrication and measurement of a multifunctional metasurface. (a) SEM
image of the nanoantennas of the fabricated gate-tunable metasurface for the demonstration of
dynamic beam steering and a reconfigurable focusing meta-mirror. The scale bars from left to
right are 200 m, 50 m, and 500 nm respectively. (b) Photographic image of the
multifunctional metasurface with 96 independently addressable elements. The scale bar is 5
mm. (c) Sample-mounting PCB to which we wire-bond the multifunctional metasurface pads.
96 metasurface elements’ pads and 4 ITO pads are wire-bonded from the sample to 100
conducting pads on the first PCB. The scale bar is 10 mm. (d) Voltage deriving PCB that
provides 100 voltages controlled by programming microcontrollers. The scale bar is 20 mm.

Figure 3.14b is a photomicrograph of the fabricated array, consisting of 96 individuallycontrollable and identical metasurface pixels. In order to individually bias each of 96

76
different metasurface elements, we design two printed circuit boards (PCBs) shown in
Figs. 3.14c, d (see Appendix B.4 for further details on the PCB design).

3.4.1. Demonstration of Beam Steering Using the Multifunctional Metasurface
After validating the wide phase tunability of our metasurface, and being able to
independently control each metasurface pixel, we employ the metasurface to mold the
scattered wavefront by locally shifting the phase of the beam at will. To this end, we
first, design and demonstrate a dynamic beam steering device. To implement beam
steering, we design the spatial phase profile of the light reflected from the metasurface
by engineering the spatial distribution of the DC bias voltages applied to the 96
metasurface pixels.
To design the spatial phase profile of the metasurface, we employ a multilevel
approximation of a blazed grating approach [133]–[135] that is widely used for
demonstration of beam steering metasurfaces [62], [116]. This approach is
implemented by creating a constant phase gradient along the metasurface, effectively
generating a periodic phase pattern that could provide us specific steering angles. The
phase values are adopted from the unit cell simulations for each applied bias. Here, we
discretize the phase shift acquired by the metasurface pixels into four levels 0, 90,
180, and 270°.
In this configuration, the metasurface acts as a diffraction grating with a reconfigurable
period. Each effective period, hereafter termed a supercell, consists of the metasurface
pixels exhibiting the discretized 4-level phase shift values. When no bias is applied, we
observe only the zeroth-order diffracted beam in the Fourier plane. In other words, the
subwavelength period of the metasurface results in an absence of higher-order
diffracted beams at zero bias. By changing the pixel repetition number (RN) for each
phase shift value within one supercell, we electrically modulate the effective period of
the metasurface array. This results in a shift of the spatial position of the first diffracted
order, enabling manipulation of the far-field radiation.
Figure 3.15a shows the spatial phase profile of the beam steering metasurface for
different RN values. To theoretically study the beam steering performance of the
metasurface, we first perform full-wave electromagnetic simulations using FDTD
Solutions by Lumerical. The far-field radiation patterns of the metasurface obtained

77
from the full-wave simulation can be seen in Fig. 3.15b. Each curve represents the
radiation pattern corresponding to the spatial phase profile illustrated with the same
color depicted in Fig. 3.15a. As can be seen, the metasurface is capable of providing a
maximum steering angle of 70.5 when RN = 1.
After fabricating the metasurface, we noticed a pitch size difference between the
designed metasurface and the fabricated device due to fabrication non-idealities. In
order to take this pitch size difference into account, we use an analytical formulation to
calculate the far-field pattern of the beam steering metasurface by using a blazed grating
approach. In this approach, we use the framework of antenna array theory, using Eqs.
(2.6) and (2.7). By implementing this method using the simulated reflectance and phase
shift values, we obtain the far-field radiation pattern of the beam steering metasurface,
illustrated in Fig. 3.15c.

Figure 3.15: Analytical and full-wave simulation of the beam steering metasurface. (a)
Spatial phase distribution of the beam steering metasurface with different RN values. (b) Fullwave simulation results of the beam steering metasurface for p = 400 nm and  = 1510 nm.
Simulation results of the beam steering metasurface obtained through analytical calculations
for (c) p = 400 nm,  = 1510 nm and (d) p = 504 nm,  = 1522 nm.

78
As can be seen, there is a good agreement between the analytically-obtained radiation
patterns and the full-wave simulation results. Once we confirmed the ability of our
analytical formulation in predicting the far-field radiation pattern, we employ the same
method by using the dimensions obtained from the SEM images to calculate the farfield pattern of our fabricated meta-device (see Fig. 3.15d).
Generally speaking, active metasurfaces rely on “on-resonant” operation. In our
metasurface, the operation wavelength is located around the resonance dip (see Fig.
3.4a). When applying an external bias, we perturb the resonance characteristics,
resulting in a phase shift. However, in addition to the phase shift, we also observe a
reflectance modulation. Hence, when performing beam steering by using the
metasurface, the amplitude of the light scattered by each metasurface element depends
on the applied bias voltage. This undesired amplitude modulation significantly
deteriorates the beam steering performance of our metasurface. Figure 3.16 illustrates
the analytical simulation results of the beam steering performance for the case of equal
amplitudes for all metasurface pixels for different RN values (Figs. 3.16a-f). The phase
shift values are obtained from Fig. 3.4d.

Figure 3.16: Effect of amplitude and phase modulation on the far-field radiation pattern
of the beam steering metasurface. Simulation results of the beam steering metasurface
obtained through analytical calculations when equal reflectance values and step phase
distributions are assumed for all metasurface pixels when (a) RN = 2, (b) RN = 3, (c) RN = 4,
(d) RN = 5, and (e) RN = 6. Analytical simulation results of the beam steering metasurface with
unequal reflectance values and step phase distribution for different metasurface pixels when (f)
RN = 2, (g) RN = 3, (h) RN = 4, (i) RN = 5, and (j) RN = 6 [95].

The beam steering simulation results for the practical metasurface in which the
amplitudes are modulated when changing the applied bias voltage are presented in Figs.

79
3.16g-l. The reflectance and phase shift values are obtained from Figs. 3.4c, d. As seen
in Fig. 3.16, the unequal reflectance values of different metasurface pixels that are
generated as a result of amplitude modulation when changing the applied bias, create
unwanted lobes in the far-field pattern while keeping the steered angle and bandwidth
unchanged. In other words, the ratio of the reflected light intensity at the desired angle
to that of the undesired angles (quantified by the beam directivity [136]) is drastically
decreased due to the variation of the amplitude.
Another important factor that needs to be considered is the detailed choice of the phase
profile across the metasurface that can significantly influence the beam steering
characteristics. Ideally, when steering a beam to a steering angle θ0, the wave reflected
from the metasurface should be a plane-wave such that its wavefront is tilted with
respect to the metasurface plane by an angle θ0. To achieve an ideal tilted plane wave,
one uses the following phase profile across the metasurface: φideal=(j-1) k p sin(θ0),
where j numerates the metasurface element, p is the metasurface period, k is the
wavenumber of the incoming plane wave [137]. To experimentally realize this phase
profile, we need to be able to access phase shifts spanning from 0 to 360. However, in
our metasurface, the maximally accessible phase shift is limited to values around 270.
To implement beam steering by using our multifunctional metasurface, we approximate
the ideal phase profile φideal by stairstep profiles shown in Fig. 3.15a. As seen in Figs.
3.15b, c, the designed step profiles enable beam steering.
One can also realize beam steering by appropriately choosing the period of the phase
profile and linearly changing the phase within the period until reaching the maximal
value of 270. For the ultimate clarity, we plot a period for these three different phase
profiles (ideal, step profile, linear) in the case of a steering angle of 14.6o in Fig. 3.17.
In Fig. 3.17, the dots correspond to the phase values at spatial positions where the
metasurface elements are located, while the solid lines show the phase profiles
effectively created across the metasurface. The stairstep phase profile shown in Fig.
3.17 corresponds to the following spatial distribution of phases: 0, 0, 0, −90, −90,
−90, −180, −180, −180, −270, −270, and −270.
To generate the linear phase profile, the desired steering angles are chosen to be the
same as the ones demonstrated by applying stairstep phase profiles. As a result, for each
steering angle, the number of metasurface pixels (MP) in one supercell is chosen to be

80
𝑀𝑃 = 4 × 𝑅𝑁 to ensure that the lengths of the generated supercells using linear phase
profiles are identical to those used in the stairstep case. Accordingly, the phases within
a supercell are picked to be evenly spaced between 0 and −270. In other words, the
phase shift between metasurface elements within a supercell is chosen as 270/(4 ×
𝑅𝑁 − 1).

Figure 3.17: A supercell for a steering angle of 14.6 with different ideal, linear, and stairstep phase profiles. The blue line corresponds to the ‘ideal’ phase profile, the red line to the
linear phase profile while the black line represents the stairstep phase profile used in the present
work. The horizontal green dashed line indicates the maximal achievable phase shift in our
metasurface [95].

As an example, here, we set a constant phase shift between neighboring metasurface
elements as 270/11 corresponding to the case of RN = 3. In the case of the linear phase
profile, the supercell is given as: 0, −24.5455, −49.0909, −73.6364, −98.1818,
−122.7273, −147.2727, −171.8182, −196.3636, −220.9091, −245.4545, and
−270. For all considered phase profiles (ideal, stairstep, and linear) the supercell
incorporates 12 metasurface pixels, and the length of the supercell (or, equivalently, the
period of the generated blazed grating) is 4.8 μm. The designed stairstep profile
attempts to accurately reproduce both the period and the slope of the ideal phase profile.
On the other hand, the slope of the linear phase profile deviates from the slope of the
ideal phase profile.
As a next step, we calculate the far-field radiation patterns in the cases when linear
phase profiles are employed to create the supercells. The far-field radiation patterns for
linear phase profiles that correspond to different MP values are presented in Figs. 3.18ae. Our simulation results show that employing linear phase profiles leads to slightly

81
better beam steering performance. As seen in Figs. 3.18a-e, the linear phase profile
yields smaller sidelobes at negative angles as compared to the case of the stairstep
profiles. On the other hand, the linear phase profiles generate larger undesired sidelobes
at positive angles. Moreover, for smaller steering angles (larger supercell sizes),
programming the microcontrollers to provide small voltage increments that can lead to
linearly distributed phase profiles would be much more complicated compared to the
cases when 4-level phase distributions are used. As a result, in our experimental
demonstration, we use 4-level stairstep phase profiles with different RN values to
change the size of the supercell and accordingly access different steering angles.

Figure 3.18: Simulation results of the beam steering metasurface obtained when unequal
reflectance values and linear phase distributions are assumed for all metasurface pixels.
(a) MP = 8, (b) MP = 12, (c) MP = 16, (d) MP =20, and (e) MP = 24 [95].

As can be implied from Figs. 3.16, to improve the beam directivity, one has to minimize
the amplitude variation when applying an electrical bias to the metasurface element.
This could be achieved by increasing the number of external control knobs, which can
be used to actively control the phase of a metasurface pixel [138]. Increasing the
maximally achievable phase shift could also improve the beam steering performance of
the metasurface since in that case one can use the ‘ideal’ phase profile to steer the beam
(Fig. 3.18).
After confirming the beam steering performance of the metasurface through
simulations, we measure the far-field radiation pattern of the beam steering device in
the Fourier space using the part of the universal setup shown in Fig. 3.19. In this part
of the setup, a coherent beam from a tunable NIR laser illuminates the metasurface
sample. The laser beam is focused on the sample by an objective (O) with a long
working distance (Mitutoyo M Plan Apo 20×, NA = 0.40, WD = 20 mm) after passing
through an iris (I), a polarizer (P), and a 50:50 non-polarizing beam splitter (BS). The
reflected beam is then guided to an IR camera (IR-C) positioned in the Fourier plane,
where the radiation pattern is captured.

82

Figure 3.19: Optical setup used for beam steering measurement. NIR-L: NIR laser, M:
mirror, I: iris, BS: 50:50 beam splitter, P: polarizer, O: objective lens, MTS: metasurface sample,
L: lens, IR-C: IR CCD camera. The components indicated with red cross marks belong to the
universal setup and are not used in this part of the measurement.

Figure 3.20a shows the metasurface spatial phase profiles, for the four-level phase shift
with RN = 2 to RN = 6. In Fig. 3.20a, each gray-shaded region determines one supercell.
In order to provide a better comparison between the simulation and experimental results,
the simulated far-field pattern of the beam steering device using the abovementioned
analytical formula is presented in Fig. 3.20b. It should be noted that the simulations
correspond to the dimensions of our fabricated metasurface that show an average pitch
size of 504 nm. As can be seen, by changing the RN value, the size of the metasurface
supercell is electrically modulated, resulting in reconfigurable beam steering with
quasi-continuous steering angles.
Figure 3.20c shows the measured far-field pattern for our beam steering device. Due to
limitations of our measurement setup, steering angles of higher than 23.5° could not be
captured by the imaging system. As a result, the maximum measured steering angle is
~22°, which corresponds to a repeat number of 2. As expected, by increasing the
effective period of the metasurface, the beam angle became smaller. We also note that
for each RN value, no diffracted order with an intensity equal to that of the desired
steering angle is observed at negative angles, indicating true phase gradient beam
steering rather than switchable diffraction. This confirms that the beam steering is

83
obtained as a result of the asymmetric phase gradient introduced by the subwavelength
metasurface phase elements.

Figure 3.20: Demonstration of dynamic beam steering by the multifunctional
metasurface. (a) The spatial phase distributions of the metasurface elements with different RN
values that are used to create phase gradients resulting in beam steering. (b) Simulation results
of the beam steering metasurface obtained through analytical calculations. Changing the RN
value from 2 to 6, the steering angles of 22.17°, 14.56°, 10.86°, 8.66°, and 7.20° are obtained
through calculations. (c) Experimental results of the beam steering metasurface. Changing the
RN value from 2 to 6, we can obtain the steering angles of 22.19°, 14.43°, 10.91°, 8.51°, and
7.40°. Each steering angle corresponds to the spatial phase distribution of the same color
presented in (a).

84
3.4.2. Demonstration of Dynamic Focusing Meta-Mirror
Using the same concept of controlling the phase imposed by each metasurface pixel,
we are able to demonstrate the use of our multifunctional metasurface as a
reconfigurable lens by developing phase profiles for lenses with different focal lengths.
In order to demonstrate focusing, we design the spatial phase profile of the metasurface
pixels such that the beams reflected from individual pixels are in phase at the desired
focal point. To this end, the phase shift provided by the metasurface is dictated by the
hyperboloidal profile:
2𝜋

𝜙(𝑥) = − 𝜆 (√𝑓 2 + 𝑥 2 − 𝑓)

(3.2)

where λ is the wavelength, f is the focal length and x is the distance of the pixel from
the center of the lens.
Figure 3.21 shows the spatial distribution of the phase shift (diamond) and the
corresponding applied bias voltage (square) required to focus the reflected beam at
focal lengths of 1.5 m, 2 m, and 3 m. These values are extracted from the simulated
phase shift as a function of applied bias (Fig. 3.4d). In order to investigate the focusing
performance, we simulate the multifunctional metasurface under the applied bias
distributions illustrated in Figs. 3.20a-c.

Figure 3.21: Theoretical demonstration of a dynamic focusing meta-mirror. Spatial phase
(diamond) and voltage (square) distribution of a focusing meta-mirror with focal lengths of (a)
f = 1.5 µm, (b) f = 2 µm, and (c) f = 3 µm using the phase shifts obtained from the simulation.
Full-wave simulation of the spatial distribution of the electric field |E|2 for the focusing metamirror with focal lengths of (d) f = 1.5 µm, (e) f = 2 µm, and (f) f = 3 µm.

85
In our full-wave electromagnetic simulations, we modeled a miniaturized lens with a
20 m aperture size since simulating the full metasurface at the small mesh sizes
required for the ITO layer active region is beyond our present numerical simulation
capability.
Figures 3.21d-f illustrate the far-field pattern of the beam reflected from our tunable
metasurface in the x-z plane. As seen in Figs. 3.21d-f, the metasurface can clearly focus
the reflected light at the focal lengths of 1.5 m, 2 m, and 3 m when appropriate bias
voltages are applied to the individual metasurface pixels.
We then experimentally characterize the dynamic focusing meta-mirror once the
focusing performance of our multifunctional metasurface was confirmed by
calculations.
In order to measure the reconfigurable focusing performance of the metasurface, we
use a part of the universal setup depicted in Fig. 3.22.

Figure 3.22: Optical setup used for focusing performance measurement. NIR-L: NIR laser,
M: mirror, I: iris, BS: 50:50 beam splitter, P: polarizer, O: objective lens, MTS: metasurface
sample, L: lens, IR-C: IR CCD camera, MS: 2-axis motorized stage. The components indicated
with red cross marks belong to the universal setup and are not used in this part of the
measurement.

In this setup, the metasurface sample is illuminated by a coherent beam from a tunable
NIR laser. The laser beam is directed to the sample surface passing through an iris (I),
a polarizer (P), an objective (O) with a long working distance (Mitutoyo M Plan Apo
20×, NA = 0.40), and a 50:50 non-polarizing beam splitter (BS). The reflected beam

86
from the metasurface is then captured by an imaging system. The imaging system
consists of an IR camera (IR-C), and an objective lens (O) with long working distance
(Mitutoyo M Plan Apo 20×, NA = 0.40, WD = 20 mm) paired with a tube lens. The
imaging system is then moved from the image plane by a 2-axis motorized stage (MS),
that could move in the x- and y-directions with a resolution of 100 nm, and the intensity
profile of the reflected beam in the x-y plane is captured at different depths.
We then program the voltages applied to each metasurface pixel to experimentally
achieve the desired phase shift values. Figure 3.23 shows the spatial phase profile of
focusing meta-mirrors designed to have focal length of f = 1.5 µm (Fig. 3.23a), f = 2
µm (Fig. 3.23b), and f = 3 µm (Fig. 3.23c). The square points show the ideal required
phase values obtained from Eq. (3.2) and the diamond points represent the phase values
acquired by the metasurface obtained from phase measurements (Fig. 3.13).

Figure 3.23: Experimental demonstration of a dynamic focusing meta-mirror with short
focal length. Spatial phase distribution of focusing meta-mirror with focal lengths of (a) f =
1.5 µm, (b) f = 2 µm, and (c) f = 3 µm. Square points show the ideal phase values and diamond
points present the phase values acquired by the metasurface. Spatial voltage distribution of
focusing meta-mirror with focal lengths of (d) f = 1.5 µm, (e) f = 2 µm, and (f) f = 3 µm.
Measured intensity profile of the beam reflected from the focusing meta-mirror with focal
lengths of (g) f = 1.5 µm, (h) f = 2 µm, and (i) f = 3 µm. The scale bar is 2 μm.

87
After fitting the measured phase values to the ideal curves, we obtain the voltage values
corresponding to the fitted phase profiles. Figures 3.23d-f show the spatial voltage
profile of the focusing meta-mirrors with a focal length of f = 1.5 µm, f = 2 µm, and f =
3 µm.
Using the setup shown in Fig. 3.22, the intensity profile of the reflected beam in the xy plane is recorded. By extracting the cross-sections of the captured intensity profiles
at fixed y values, we reconstruct the intensity profile of the reflected beam in the x-z
plane. Figures 3.23g-i illustrate the metasurface reflected beam intensity profiles in the
x-z plane for the applied bias distributions shown in Figs. 3.23d-f.
As can be seen, the fabricated metasurface focuses the reflected beam at the desired
depths. The scale bars in Figs 3.23g-i are obtained by imaging an object of known size.
When the incident light is polarized perpendicular to the antennas, no focusing is
observed since no phase modulation could be achieved in that polarization. This
observation confirms that the captured focusing originates from the metasurface.
Using the same concept of individually-controlled metasurface pixels, we reprogram
the bias voltages applied to the metasurface in order to experimentally demonstrate a
tunable focusing meta-mirrors with focal length varying from 15 m to 25 m. Figures
3.24a-c show the spatial distribution of the phase shift (diamond) and the corresponding
applied bias voltage (square) required to focus the reflected beam at the focal lengths
of 15 m, 20 m, and 25 m. The voltage distribution is designed using the measured
bias voltage/phase relationship (Fig. 3.13). The measured intensity profiles mapped in
the x-z plane are illustrated in Fig. 3.24d-f.
As can be seen, our universal metasurface is able to provide tunable focusing with
reconfigurable focal lengths only by reprogramming the voltages applied to it. The
focal length could be tuned from small values (1.5 m) and can be extended to large
ones (25 m).
We can further increase the focal length of the metasurface by increasing the number
of metasurface pixels that are individually controlled. Figure 3.25 shows the spatial
phase profile and the corresponding voltage profile of the focusing metasurface with
focal length as large as f = 150 µm (Fig. 3.25a, d), f = 200 µm (Fig. 3.25b, e), and f =
250 µm (Fig. 3.25c, f).

88

Figure 3.24: Experimental demonstration of a dynamic focusing meta-mirror with long
focal length. Spatial phase distribution of focusing meta-mirror with focal lengths of (a) f =
15 µm, (b) f = 20 µm, and (c) f = 25 µm. Square points show the ideal required phase values
and diamond points present the phase values acquired by the metasurface. Spatial voltage
distribution of focusing meta-mirror with focal lengths of (d) f = 15 µm, (e) f = 20 µm, and (f)
f = 25 µm. Measured intensity profile of the beam reflected from the focusing meta-mirror with
focal lengths of (g) f = 15 µm, (h) f = 20 µm, and (i) f = 25 µm. The scale bar is 2 μm.

Figure 3.25: Possibility of demonstration of focusing meta-mirror with extended focal
lengths using the multifunctional metasurface. Spatial phase distribution of focusing metamirror with focal lengths of (a) f = 150 µm, (b) f = 200 µm, and (c) f = 250 µm. Square points
show the ideal required phase values and diamond points present the phase values acquired
by the metasurface. Spatial voltage distribution of focusing meta-mirror with focal lengths of
(d) f = 150 µm, (e) f = 200 µm, and (f) f = 250 µm.

89
The focusing meta-mirrors with the phase/voltage profiles presented in Fig. 3.25 consist
of 288 individually-addressable metasurface pixels. Such tunable focusing meta-mirror
with micro-scale focal length can be potentially applied in many applications, such as
light-field imaging [139] and full-color imaging [140].

3.5. Conclusions and Outlook
In this chapter, we proposed the design and experimental demonstration of an
electrically tunable multifunctional metasurface in the NIR wavelength range. The
multifunctional metasurface is realized via field-effect-induced modulation of TCO
active regions incorporated into the metasurface and is capable of spatiotemporal
modulation of the fundamental attributes of light. As a proof of concept, we designed
phase profiles for our multifunctional metasurface to demonstrate beam steering and
dynamic focusing using the same device via individually controlling each metasurface
pixel. Such a multifunctional metasurface can initiate integrated on-chip electro-optical
devices such as LiDAR systems. Prior research has shown that the reflectance of the
ITO-based active metasurfaces can be considerably enhanced by utilizing ITOintegrated all-dielectric guided-mode resonance mirror designs [141]. The efficiency of
the multifunctional metasurface can possibly be further improved via optimization
algorithms [142]–[145]. It has been previously shown that optimization algorithms may
yield non-trivial structural shapes and metasurface antenna distributions that yield
significantly improved optical performance. In particular, optimization algorithms may
significantly boost the performance of active beam steering metasurfaces [146]. A
worthy direction for future research is to extend the multifunctional metasurface
concept demonstrated here to a two-dimensional phased array architecture. In addition
to enabling beam steering and focusing in two dimensions, such a two-dimensional
array could enable fast and energy-efficient programmable devices such as dynamic
holograms, off-axis lenses, axicons, vortex plates, and polarimeters.

90
Chapter 4

ELECTRO-OPTICALLY TUNABLE METASURFACES FOR
DYNAMIC POLARIZATION CONTROL
Optical polarization is an important characteristic of electromagnetic waves that has a
significant impact on a number of applications, such as information delivery, 3D
imaging, and quantum computation. Optical metasurfaces have attracted immense
attention due to their ability to control constitutive properties of electromagnetic waves
at a subwavelength scale. This makes metasurfaces an appropriate base for the creation
of flat optical devices with novel functionalities. Among the diverse promising
applications that metasurfaces can provide, versatile polarization generation in a
compact device dimension has been of great importance. In this chapter, we will present
an electro-optically tunable metasurface scheme that can generate versatile polarization
states. The proposed ITO-based metasurface can be used for the realization of active
polarization modulation through a field-effect-induced modulation of the ITO
properties. By suitably biasing the metasurface, the linearly-polarized incident light can
be actively converted to a cross-polarized, circularly-polarized, or elliptically-polarized
light.

4.1. Introduction
Controlling light polarization plays a vital role in information delivery, imaging, and
quantum optics [147], [148]. Consequently, there has been a growing demand for
research investigating the modulation of the optical polarization states. Polarization
converters can convert an electromagnetic wave with an undefined polarization state
into a wave with a well-defined polarization such as linear and circular. Conventional
approaches for polarization conversion project the incident electromagnetic field onto
two components with orthogonal linear polarizations through bulky wave-plates made
of birefringence crystals or polymers. Polarization rotation via Faraday and Kerr effects
has also been proposed [149]. However, due to the dispersive nature of natural media,
either complicated structural designs or multilayered films are required to overcome

91
the narrow working bandwidth. This will make such devices to be bulky, and hence
inappropriate for photonic integration.
As mentioned in the previous chapters, metasurfaces can introduce abrupt changes to
the fundamental attributes of scattered light within a subwavelength spatial region [150],
[151]. Owing to their flexibility in structural design, they offer multiple degrees of
freedom for light manipulation, being utilized in flat optics and low-profile
optoelectronic components with diverse functions such as hyperspectral imaging [152],
focusing [11], [153], [154], and wavefront shaping [155], [156]. Metasurfaces have
been employed to demonstrate polarization conversion in a wide range of frequencies
[157], [158] by introducing chirality [159], birefringence [160], or different
electromagnetic amplitudes as well as phase shifts between orthogonal electric field
components [76], [161], [162] using a single compact device. As a result, the
unprecedented capability of metasurfaces to realize polarization manipulation has been
applied in polarization imaging [163], entangled photon generation [164], Stokes
parameters detection [165], [166], and holographic displays [167], [168].
However, the polarization conversion metasurfaces presented so far could only convert
a well-defined-polarized beam to a specific polarization state. A universal polarization
converter, which is capable of realizing multiple polarization states via a single device
is highly desired but still missing. Such universality can be achieved by using tunable
metasurfaces which are obtained by incorporating active media into the otherwise
passive metasurfaces.
Due to the intrinsic limitations of the active materials used in tunable metasurfaces,
dynamic polarization control via active metasurfaces has only been investigated at lowfrequency regions such as MIR [169], [170], terahertz (THz) [171], [172], and gigahertz
(GHz) [173] regimes. Establishing low-profile universal polarization converters that
could provide comprehensive polarization control in visible and NIR spectral ranges
[174] is of great significance in advanced nanophotonics and quantum applications.
Recently, a reconfigurable metasurface consisting of Au meta-aperture and ethyl-red
switching polymer was proposed to experimentally demonstrate active control of the
polarization of transmitted light at a wavelength of 800 nm [175]. However, only
linear-to-elliptical polarization conversion and linear polarization rotation were
realized.

92
In another work, a tunable all-dielectric metasurface capable of dynamic polarization
control at a wavelength of 1535.4 nm was theoretically studied. When applying a bias
voltage to a doped-Si-based metasurface, the electro-optical shift of the Huygens mode
through carrier accumulation resulted in a wide phase difference between two
orthogonal electric components [176]. Nonetheless, the proposed active metasurface
could only access circular and elliptical polarization components with the incident
beam being linearly polarized along an axis with a 45o angle with respect to the
symmetry axis of the unit element. To date, a promising metasurface design for the
feasible realization of arbitrary polarization transformation at the NIR wavelength
regime has not been completely explored yet.
In this chapter, we propose the design and experimental demonstration of a tunable
polarization converter metasurface scheme that can achieve arbitrary polarization states
when illuminated by a beam in the telecommunication wavelength regime with a fixed
polarization. The desired polarization states are obtained via dynamic control of the
amplitude and phase differences between two plasmonic eigenstates of the incident
electromagnetic wave using an ITO-based tunable metasurface.
The metasurface is composed of a metal-insulator-semiconductor heterostructure with
the ITO layer serving as a degenerately doped semiconductor. When applying an
electrical bias between the ITO and the metal gate, the carrier concentration of ITO
undergoes a reproducible change at the gate-dielectric/ITO interface, resulting in a
modulation of the refractive index of the ITO layer. By suitably biasing the metasurface,
the linearly-polarized incident light can be converted to linearly cross-polarized,
circularly-polarized, or elliptically-polarized light.

4.2. Tunable Polarization Conversion Metasurface Design
Figure 4.1 shows the schematic for our proposed dynamic polarization conversion
metasurface, which consists of a 150 nm-thick Al back reflector, a 20 nm-thick
Al2O3/HfO2 nanolaminate (HAOL) gate dielectric, followed by a 5 nm-thick ITO layer,
and an array of Al nanoantenna as the topmost layer. The utilization of HAOL structure
is because of its high breakdown field and large DC permittivity characteristics.

93

Figure 4.1: Schematic illustration of the active polarization conversion metasurface
operating in reflection mode. By suitably biasing the metasurface structure, a linearlypolarized incident beam can be converted to a cross-polarized or circularly-polarized light [177].

4.3. Optical Response of the Polarization Conversion Metasurface
In order to obtain tunable polarization modulation, the metasurface requires to support
two plasmonic eigenstates. As a result, an array of anisotropic Al nanoantennas, rotated
by θ = 45° with respect to the y-axis is designed and utilized as the building block.
Figure 4.2a schematically illustrates the metasurface unit cell. The length and width of
the antennas are chosen to be L = 280 nm, w = 230 nm such that the metasurface
supports resonances at wavelengths close to 1550 nm. The period of the metasurface is
p = 400 nm.
To simulate the optical response of the metasurface, we use the FDTD method
(Lumerical). In our simulations, we use PML boundary condition in the z-direction and
periodic boundary conditions in both the x- and y-directions. Hence, the calculations of
the reflection intensity and phase shift of the Al nanoantennas are performed in an array
configuration. In our simulations, we assume that the incident beam is linearly polarized
with an electric field Ein along the y-axis (with 1 V/m amplitude), and hence, making
an angle of 45° with respect to the long axis of the nanoantennas (see Fig. 4.2b). This
enables simultaneous excitation of two gap plasmon modes associated with the size of

94
the long and short axes of the Al nanoantenna at two distinct wavelengths in the
telecommunication wavelength range, as shown in Fig. 4.2c.
For the first excited mode (blue curve in Fig. 4.2c), the electric field component along
the long-axis of the nanoantenna is dominant, while for the second excited mode
(magenta curve in Fig. 4.2c), the electric field component along the short-axis of the
nanoantenna is influential. By projecting the reflected beam onto the x- and y-axes, one
could see that the x- and y- polarized components of the reflected light can be modulated
by controlling the interaction between the two induced plasmonic modes (see Fig. 4.2b).

Figure 4.2: Metasurface design principle for demonstration of tunable polarization
conversion. (a) Schematic for the tunable metasurface unit structure. The metasurface consists
of a 150 nm-thick Al back reflector, a 20 nm-thick HAOL, a 5 nm-thick ITO, and an array of
Al nanostructures with a thickness of 80 nm. The unit element dimensions are defined as: L =
280 nm, w = 230 nm, θ = 45°, and p = 400 nm. (b) Design principle of tunable polarization
converter. When a y-polarized light interacts with the patch antenna, two gap plasmon modes
are excited. The amplitude and phase of the excited modes can be modulated by biasing the
ITO with respect to the Al mirror. (c) The simulated reflectance spectrum of optimized
metasurface design [177].

Then, we apply a DC bias between the Al back reflector and the ITO layer, leading to
the modulation of the permittivity of ITO. Figure 4.3 shows the spatial distribution of

95
the real part of the ITO permittivity for positive (Fig. 4.3a) and negative (Fig. 4.3b)
applied voltages. Here, the carrier concentration of ITO is assumed to be 2.8×1020 cm3

. As can be seen in Fig. 4.3a, for some applied voltages, the real part of permittivity

approaches zero, creating an ENZ condition in the ITO layer. By coupling this ENZ
region to the resonances provided by the metasurface, one could alter the interaction
between the induced plasmonic modes, leading to modulation of the polarization state
of the reflected light. It is worth mentioning that the ENZ condition shows up only
when the gate voltage is greater than ~3 V.

Figure 4.3: Modulation of ITO properties under an applied bias. Spatial distribution of the
real part of permittivity of a 5 nm-thick ITO layer under (a) positive and (b) negative applied
biases with respect to the back reflector. The operating wavelength is set to 1580 nm [177].

In order to investigate the dynamic behavior of the metasurface, we study the amplitude
and phase of the x- and y-polarized reflected beams under applied bias, when the
metasurface is illuminated by a y-polarized beam.
Figure 4.4 shows the reflectance spectra of the co- and cross-polarized reflected beams
for different applied biases. As can be seen in Fig. 4.4, the intensities of both the x- and
y-polarized reflected beams are modulated when applying bias, confirming the
capability of the proposed metasurface to provide amplitude modulation. It should be
noted that due to the absence of ENZ condition, the metasurface shows much weaker
amplitude modulation under negative applied biases (see Figs. 4.4d, e).

96

Figure 4.4: Amplitude modulation provided by the tunable polarization conversion
metasurface. (a) Simulated reflectance spectra of the co- (top panel) and cross-polarized
(bottom panel) light for four different applied voltages. The inset of the bottom panel shows
the orientation of Al nanoantenna. Simulated reflectance of the (b, c) co- and (d, e) crosspolarized light as a function of wavelength and applied bias for (b, d) positive and (c, e)
negative voltages [177].

In order to obtain a polarization converter, the phase shift provided by the device plays
a vital role. Figure 4.5 shows the phase difference between x- and y-polarized
components of the reflected light as a function of wavelength and applied bias. Here,
the phase difference is defined as ∆φ = φyy – φxy, where φij (i, j = x, y) presents the phase
of the i-polarized reflected beam under a j-polarized illumination. As can be seen in Fig.
4.5, when changing the applied bias, the metasurface can provide a significant phase
modulation. Similar to the amplitude modulation under positive applied biases (Figs.
4.4b, d), a more dominant phase modulation can be achieved at such bias voltages.
In order to further explore the operating principle of the metasurface, we investigate the
near-field coupling conditions at different applied biases by studying the field
distributions of the metasurface. Figure 4.6 shows the z component of the electric field
(Fig. 4.6a) and the intensity of the magnetic field (Fig. 4.6b) at three different applied
biases.
As can be seen, in the absence of applied bias, a strong near-field coupling between the
Al nanoantenna and back reflector is observed, which results in a strong magnetic field
confined within the HAOL gate dielectric layer.

97

Figure 4.5: Phase modulation provided by the tunable polarization conversion
metasurface. The phase difference between x- and y-polarized reflected beams as a function
of wavelength and applied bias. The white solid line marks the wavelength and voltage pairs at
which equal reflectance values are observed for x- and y-polarized components of the reflected
beam [177].

Figure 4.6: Spatial distribution of the electromagnetic fields in the tunable polarization
conversion metasurface under an applied bias. Spatial distribution of (a) z-component of
electric field and (b) intensity of the magnetic field for three different applied biases. Here, the
incident beam is polarized along the y-axis and forms an angle of 45° with the long-axis of the
nanoantennas. The incident plane-wave has an amplitude of 1 V/m and a wavelength of 1580
nm [177].

98
Under a linearly y-polarized illumination, the appearance of the z-component of the
electric field in the x-z plane indicates a linear cross-polarization conversion. It can be
observed that the z-component of the electric field has remarkably higher values in the
y-z plane compared to that of the x-z plane. This will result in the reflected beam to be
mostly aligned along the x-axis due to weaker field confinement.
When the applied bias increases to 4 V, the appearance of the ENZ condition breaks
the field symmetry in the ITO layer. This weakens the magnetic field confinement in
the HAOL layer, as shown in the middle panel of Fig. 4.6b. Moreover, as a result of the
similar values of the z-component of the electric field in the x-z and y-z planes, one can
anticipate observing closer x- and y-polarized reflectance values at an applied bias of 4
V.
Further increasing the applied bias to 14 V, the spatial position of the ENZ condition
shifts towards central regions of the ITO layer (see the right panel of Fig. 4.6a). This
results in a more significant interaction between the Al nanoantenna and the back
reflector and makes the magnetic field to be less confined in the HAOL layer (see right
panel of Fig. 4.6b). At this applied bias, we can expect to see similar reflectance values
for the x- and y-polarized reflected beams because of the close values of the zcomponent of the electric field in the x-z and y-z planes.
The presented results show that by tailoring the ENZ condition of the ITO layer as a
result of changing the applied bias, one can alter the interaction between the two excited
gap plasmon modes, leading to the possibility of modulation of the amplitude, phase,
and polarization state of the reflected light beam.

4.4. Dynamic Modulation of the Polarization State of the Reflected Beam by
Using Tunable Polarization Conversion Metasurface
After confirming the capability of the proposed metasurface design to provide a tunable
optical response, we investigate the tunable polarization conversion function of the
metasurface. Figure 4.7 presents the dynamic polarization conversion at a wavelength
of λ = 1580 nm as a result of the bias applied to the reconfigurable metasurface.
Reflectance values of the x- and y- polarized reflected beams as well as the phase
difference between these two components are shown in Fig. 4.7a. As can be seen, when
the applied bias is varied between −4 V and 2 V, the intensity of the y-polarized

99
reflected beam is negligibly small. This leads to the realization of a y-to-x crosspolarization conversion with about 30% conversion efficiency. Increasing the applied
voltage from 2 V to ~12 V, our metasurface exhibits a linear-to-elliptical polarization
conversion.
The realization of the circularly-polarized reflected beam is the most challenging aspect
of the tunable polarization converter design. To obtain a circularly polarized light, the
reflectance values of the x- and y-polarized beams need to be equal, and a phase
difference of either 90° or −90° needs to be observed between these two components.
As shown in Fig. 4.5, this condition is satisfied at a wavelength of λ = 1580 nm. As a
result, the operating wavelength of the device is chosen to be 1580 nm.

Figure 4.7: Calculated polarization conversion performance of the Al-based tunable
metasurface. (a) Simulated reflectance of the cross-polarized (orange curve) and co-polarized
(olive curve) beams and phase difference between these two components as a function of
applied bias. The metasurface is illuminated by a y-polarized normally-incident beam at an
operating wavelength of 1580 nm. The cyan and magenta shadowed regions indicate the range
of voltages at which y-to-x cross polarization conversion and linear-to-circular polarization
conversion are obtained, respectively. (b) Stokes parameters (normalized to s0) as a function of
applied bias voltages (c) Voltage-dependent path of reflected light’s polarization state on the
Poincaré sphere at a wavelength of 1580 nm [177].

100
It can be seen in Fig. 4.7a that for the applied biases between 12 V and 16 V, the x- and
y-polarized beams have almost equal reflectance values and there is a 90° phase
difference between them. As a result, under an applied bias in this range, the proposed
metasurface can convert a linearly-polarized incident beam into a right-handed
circularly-polarized (RCP) reflected beam. Based on the results shown in Fig. 4.7a,
since all three interesting polarization states are obtained under a positive applied bias
regime, hereafter, we will focus our discussions on the positive voltage range.
In the next step, to further evaluate the polarization conversion performance of the
metasurface, we use Stokes parameters to investigate the generated polarization states
under different applied biases. For an electromagnetic wave propagating along the zaxis, the Stokes parameters can be determined by [165], [166] :

𝑠0 = |𝑟𝑥 |2 + |𝑟𝑦 | = 𝑅𝑥𝑦 + 𝑅𝑦𝑦

(4.1)

𝑠1 = |𝑟𝑥 |2 − |𝑟𝑦 | = 𝑅𝑥𝑦 − 𝑅𝑦𝑦

(4.2)

𝑠2 = 2|𝑟𝑥 ||𝑟𝑦 | cos(△φ) = |𝑟45° | − |𝑟−45° | = 𝑅45° − 𝑅−45°

(4.3)

𝑠3 = 2|𝑟𝑥 ||𝑟𝑦 | sin(△φ) = |𝑟𝑅𝐶𝑃 |2 − |𝑟𝐿𝐶𝑃 |2 = 𝑅𝑅𝐶𝑃 − 𝑅𝐿𝐶𝑃

(4.4)

where rx, ry, r45°, r−45°, are the complex reflection coefficients of the beam that is
linearly-polarized along the x-direction, y-direction, and 45° and −45° with respect to
the y-axis, respectively. rRCP, rLCP are the complex reflection coefficients of RCP and
left-handed circularly-polarized (LCP) beams, respectively. ∆φ = φyy – φxy is the phase
difference between x- and y-components of reflected light under an illumination that is
linearly-polarized along the y-axis (can be obtained from Fig. 4.6).
The calculated Stokes parameters as a function of voltage and wavelength are plotted
in Fig. 4.7b. As can be seen, at an applied bias of ~4 V, the Stokes parameters face a
significant change. This is consistent with the results shown in Fig. 4.7a indicating that
the polarization conversion function of the metasurface alters from a linearly crosspolarization conversion into a linear-to-elliptical polarization conversion at the
mentioned applied bias. This is indeed in accordance with the fact that the ENZ
condition starts to show up at the applied biases larger than ~3 V. Under applied biases
between 12 V and 16 V, s1 and s2 have negligibly small values while s3 exhibits a value
of ~ +1. This verifies that the metasurface can provide a linear-to-circular polarization
conversion under such applied biases.

101
Figure 4.7c plots the voltage-dependent path on the Poincaré sphere of the polarization
states at a wavelength of 1580 nm. It can be seen in Fig. 4.7c that in the absence of the
applied bias, the y-polarized incident beam is converted to an x-polarized light. When
increasing the applied bias, the polarization state moves toward the upper-hemisphere,
corresponding to the right-handed elliptical polarization, and remains at the north-pole
corresponding to the RCP state when the bias voltage keeps increasing. On the other
hand, under a negative applied bias, the polarization state moves toward the lowerhemisphere, corresponding to the right-handed elliptical polarization. These
consequences are consistent with the results presented in Fig. 4.7a.
As a result, it can be seen in Fig. 4.7 that our dynamic metasurface enables conversion
of linearly-polarized incident light into a cross-polarized, circularly-polarized, and
elliptically-polarized reflected waves. By an appropriate choice of the applied bias,
three of the widely used polarization states can be obtained using a single metasurface
device.

Figure 4.8: Realization of different polarization states when changing the operating
wavelength. Voltage-dependent path of the reflected light’s polarization state on the Poincaré
sphere at a wavelength of (a) 1456 nm and (b) 1632 nm [177].

It is worth mentioning that the proposed metasurface is able to access other polarization
states on the Poincaré sphere at other wavelengths (see Fig. 4.8). As can be seen in Fig.
4.8, when increasing the bias voltage from 0 V at an operating wavelength of 1456 nm,
one can alter the polarization state from a left-handed elliptical polarization to a linearly
polarized state along −45° (with respect to the y-axis). The polarization returns to a lefthanded elliptical state when the voltage is further increased. Moreover, applying

102
negative bias voltages at a wavelength of 1456 nm results in the realization of an LCP
beam.
Figure 4.8b also shows that at a wavelength of 1632 nm, electrical switching of the
polarization state between x-directed, right-handed elliptical, and linearly-polarized
along +45° can be attained.
4.5.

Experimental

Demonstration

of

Tunable

Polarization

Conversion

Metasurface
Once we confirmed the successful operation of our reconfigurable polarization
conversion metasurface, we fabricate the metasurface to measure its optical response.
During our fabrication process, we could achieve a higher breakdown voltage of the
HAOL layer when the gate was made of Au rather than Al (see Appendix C.1). As a
result, to experimentally demonstrate the tunable polarization conversion metasurface,
we need to redesign the metasurface based on Au back reflector and Au nanoantennas.
In the new design, the thickness of the back reflector, the HAOL layer, and the
nanoantennas are set to 80 nm, 9.5 nm, and 140 nm, respectively. The period of the
metasurface, the width, and length of the antenna are p = 350 nm, w = 180 nm, and L =
210 nm, respectively.

Figure 4.9: Calculated polarization conversion performance of the Au-based tunable
metasurface. Simulated reflectance of the cross-polarized (orange curve) and co-polarized
(olive curve) beams and phase difference between these two components as a function of
applied bias. The metasurface is illuminated by a y-polarized normally-incident beam at an
operating wavelength of 1520 nm. The cyan and magenta shadowed regions indicate the range
of voltages at which y-to-x cross polarization conversion and linear-to-circular polarization
conversion are obtained, respectively [177].

103
Figure 4.9 shows the reflectance values of the x- and y-polarized reflected beams as
well as the phase difference between them. As can be seen, the new design can provide
linear y-to-x, linear-to-circular, and linear-to-elliptical polarization conversions when
changing the applied bias.
After confirming the tunable polarization conversion performance of the Au-based
metasurface through simulations, we fabricate the metasurface device. Figure 4.10
summarizes the fabrication steps. First, a Si substrate with a 285 nm-thick SiO2 layer
on top is cleaned using standard cleaning processes. Then, the Au back reflector is
patterned using an electron beam pattern generator (EBPG) [VISTEC EBPG 5000+] at
an acceleration voltage of 100 keV after spinning an EBR layer. The exposed EBR
layer is then developed and 3 nm-thick Cr followed by an 80 nm-thick Au layer is
deposited using an e-beam evaporator (Fig. 4.10a).

Figure 4.10: Fabrication steps of the Au-based tunable polarization conversion
metasurface. (a) Patterning the Au back reflector, (b) deposition of the HAOL film, (c)
patterning the ITO layer, and (d) patterning the Au nanoantennas and Au contact pads [177].

After the lift-off process, the HAOL layer is deposited via ALD through shadow masks
(Fig. 4.10b). Then the ITO layer is patterned using EBPG, and then a 5 nm-thick ITO
layer is deposited after developing the exposed resist (Fig. 4.10c). Finally, the
nanoantennas and the connecting pads are patterned via EBPG and 140 nm-thick Au is

104
deposited using an e-beam evaporator. After removing the excess Au and resist, the
final metasurface device is obtained (Fig. 4.10d).
Once we fabricated the tunable polarization conversion metasurface device, we perform
voltage-tunable measurements under a y-polarized normally-incident illumination.
Figures 4.11a, b show the measured spectra of the y- and x-polarized reflected beams
under different applied biases. The spectra of the relative reflectance change for these
two components are plotted in Figs. 4.11c, d. As can be seen, when changing the applied
bias, the reflectance values of both x- and y-polarized reflected beam are modulated.

Figure 4.11: Measured reflectance spectra of the linearly-polarized reflected beams. The
reflectance spectrum of (a) y-polarized and (b) x-polarized reflected beams for different applied
biases. Relative reflectance change spectrum of (c) y-polarized and (d) x-polarized reflected
beams for different applied biases [177].

Figure 4.12 plots the reflectance spectra of the RCP and LCP reflected beams as well
as their relative reflectance changes for different applied biases. As can be seen in Fig.
4.12, when changing the applied bias, both the RCP and LCP reflected beams are
altered under a bias application. Similar to the x-, y-, and circularly-polarized reflected
beams, the reflectance values of the reflected beams linearly-polarized along the 45o
and −45o axes are also tuned when changing the applied bias (see Fig. 4.13).

105

Figure 4.12: Measured reflectance spectra of the circularly-polarized reflected beams. The
reflectance spectrum of (a) LCP and (b) RCP reflected beams for different applied biases.
Relative reflectance change spectrum of (c) LCP and (d) RCP reflected beams for different
applied biases [177].

Figure 4.13: Measured reflectance spectra of the linearly-polarized reflected beams along
45 and −45 axes. The reflectance spectrum of (a) y-polarized and (b) x-polarized reflected
beam for different applied biases. Relative reflectance change spectrum of (c) y-polarized and
(d) x-polarized reflected beam for different applied biases [177].

106
In the next step, we measure the Stokes parameters under an applied bias. Figure 4.14
presents the spectra of s1, s2, and s3 under different applied biases. As can be seen, the
amplitudes of the Stokes parameters are tuned when changing the applied bias. This
verifies a dynamic change in the direction of elliptical polarization. It can also be
observed that at the peak position of s1, both s2 and s3 have large values, resulting in the
absence of a linear cross polarization conversion. Moreover, when s3 locates at its
maximum value, s1 is almost zero. However, s2 possess non-negligible values,
indicating that no linear-to-circular polarization conversion was realized using the
tunable metasurface.
It should be noted that the inability of the fabricated metasurface to realize all expected
polarization states can trace its roots to the non-idealities associated with the device
fabrication. Moreover, the ITO layer incorporated into the fabricated metasurface
device showed different characteristics compared to the ITO layer used for designing
the metasurface elements. This can also result in a different optical response of the
fabricated metasurface than that of the designed structure.

Figure 4.14: Measured Stokes parameters of the tunable polarization conversion
metasurface. Spectrum of (a) s1 (b) s2, and (c) s3 for different applied biases [177].

4.6. Conclusions and Outlook
In this chapter, we proposed an actively tunable metasurface that could provide a
dynamic polarization conversion. The metasurface operates by virtue of field-effectinduced modulation of the properties of an ITO layer integrated within the metasurface
as a result of applied bias. When changing the electrical bias across the active ITO layer,
the carrier concentration at the ITO layer is modulated, resulting in a change of the
refractive index of ITO. By coupling the ENZ region of the ITO layer into the
geometrical resonances of the metasurface, the polarization state of the reflected beam

107
can be tuned at will. By suitably biasing the metasurface structure, the proposed
polarization converter can actively transform the incoming linearly y-polarized light
into a linearly cross-polarized, circularly-polarized, and elliptically-polarized reflected
beams at telecommunication wavelength regime. This dynamic control of the amplitude,
phase, and polarization state of the scattered beam provides prospects for various
applications, such as dynamic wave-plates, low-profile spatial light modulators,
adaptive wavefront control, signal monitoring and detection, and quantum optics.

108
Chapter 5

MODULATION OF SPONTANEOUS EMISSION OF QUANTUM
EMITTERS BY ACTIVE METASURFACES
Emission control of quantum dots and quantum emitters is a cornerstone of modern
high-quality lighting and display technologies. Dynamic emission control of quantum
emitters in an optoelectronic device is usually achieved by changing the optical pump
intensity or injection current density. Here, we propose a mechanism for the temporal
modulation of quantum emitters’ emission intensity at a constant optical pumping rate.
Our mechanism is based on the electronically-controlled modulation of the local
density of optical states (LDOS) at the position of the quantum emitter, resulting in the
modulation of the quantum emitter spontaneous emission rate, radiative decay rate, and
quantum efficiency. We manipulate the LDOS via field-effect-induced optical
permittivity modulation of an ultrathin ITO film, which is incorporated in a gated
Au/ITO/HAOL plasmonic heterostructure. The demonstrated electrical control of the
emission of the quantum emitters provides a new approach for modulating the intensity
of light in displays and other optoelectronic devices.

5.1. Purcell Enhancement of Spontaneous Emission from Quantum Emitters
In recent decades, controlling light emission of quantum emitters has been a grand
challenge of nanophotonics [178]. Different studies have been focusing on enhancing
the spontaneous emission rate [179] of a quantum system via tailoring the LDOS [180],
[181], a phenomenon known as the Purcell effect [182]. In order to enhance the Purcell
factor and improve emission directionality, epitaxial quantum dots (QDs) were first
coupled to dielectric cavities [183]. Integrating emitters into dielectric optical microand nano-cavities with high quality factors (Q-factors) and small mode volumes [184],
[185] can enhance the emission rate of the emitters [186]–[190]. However, due to the
high quality factor of dielectric nanocavities, single QDs need to be positioned at the
maximum field of the cavity so that the emission of the QDs can be spectrally tuned to
match the cavity mode [191].

109
This issue can be resolved by incorporating the quantum emitters into plasmonic
nanostructures owing to their relatively broad spectral responses.
The Purcell factor describing the enhancement of the LDOS in resonant cavities can be
defined by:

𝜆 3𝑄

𝑃𝐹 = 4𝜋2 (𝑛) 𝑉

(5.1)

where 𝑛 is the wavelength of light in the medium, 𝑄 is the quality factor of the
resonance, and 𝑉 is the mode volume.
As can be seen in Eq. (5.1), obtaining micro- and nano-cavities with large values of
Q/V that can result in an increased Purcell factor is a central theme in solid-state
quantum optics.
In plasmonic structures, absorptions and radiative losses lead to small quality factors.
However, excitation of plasmonic modes in such structures enables the confinement of
radiation to sub-wavelength dimensions. This strong field enhancement in the
plasmonic nanostructures provides large values of Q/V. This results in a strongly
modified Purcell factor. As a consequence, coupling quantum emitters to plasmonic
cavities can profoundly enhance the emission decay rate [192]–[194]. Along the same
lines, plasmonic patch antennas consisting of emitters situated in a vertical gap between
a metal disk and a metal plane have been employed to enhance the emission rates of
emitters [195]–[197].
Recent studies have shown large emission rate enhancements, high radiative
efficiencies, and directional emissions by embedding emitters in well-controlled
nanoscale gaps sandwiched between a colloidally synthesized silver nanocube and a
metal film. By controlling the refractive index and thickness of the gap as well as the
dimensions of the nanocube, the resonance wavelength of the nanoscale patch antenna
could be tuned from 500 nm to 900 nm [198]–[200]. In a recent study, an ultrafast and
efficient source of spontaneous emission with a lifetime shorter than 11 ps and Purcell
factor of up to 880 was demonstrated by integrating colloidal and photostable
semiconductor QDs into plasmonic nanopatch antennas. Coupling colloidal QDs to a
plasmonic nanocavity, resulting in a 540-fold increase of the emission decay rate and a
simultaneous 1900-fold increase of total emission intensity was also demonstrated
[201].

110
In another study, Cy5 fluorophores were embedded within the gap region between
colloidally synthesized silver nanocubes and a silver film [202]. The plasmon resonance
of the nanocubes was tuned by varying the nanocube size, leading to 30000-fold
fluorescence enhancements accompanied by a 74-fold enhancement of the spontaneous
emission rate.

5.2. Reconfigurable Purcell Enhancement of Spontaneous Emission by
Metasurfaces
Despite notable progress in this field, most of the micro- and nano-cavities to which
quantum emitters are coupled have fixed properties at the time of fabrication. This fixed
photonic environment leads to a static spontaneous emission decay rate of the emitter,
limiting the functionality of such passive devices. In conventional schemes, the
radiation from quantum emitters can be controlled by altering the optical [203] or
electrical [204] pump intensity within a fixed nanostructured environment.
On the other hand, coupling quantum emitter to nanostructures with tunable local
environments and optical properties could result in dynamic control of the emitter decay
rate at a constant optical pump power.
Based on Fermi’s golden rule, the spontaneous emission decay rate of a dipole is given
by:
2𝜔

𝛾𝑠𝑝 (𝒓) = 3ℏ𝜖 |𝒑|2 𝜌(𝒓, 𝜔) + 𝛾𝑖𝑛𝑡

(5.2)

where 𝒓 is the position, 𝜔 is the emission frequency, 𝜖0 is the permittivity of free space,
𝒑 is the transition dipole moment of the emitter, and 𝛾𝑖𝑛𝑡
is the internal non-radiative

decay rate of the emitter. The local density of optical states, 𝜌(𝒓, 𝜔), is given by:
6𝜔

̂ 𝑝 . 𝐼𝑚{𝑮(𝒓, 𝒓)}. 𝒏
̂𝑝]
𝜌(𝒓, 𝜔) = 𝜋𝑐 2 [𝒏

(5.3)

̂ 𝑝 and 𝐺(𝒓, 𝒓) are the unit vector pointing in the direction of 𝒑 and the dyadic
where 𝒏
Green’s function of the system, respectively.
The dyadic Green’s function of a dipole is determined by the electric field within the
environment in which the dipole is embedded. This electric field is a function of the
properties of the structure. As a result, tailoring the properties of the nanostructure via
application of an external stimulus will enable dynamic control of the dyadic Green’s

111
function, and hence, will result in a reconfigurable spontaneous emission decay rate of
the dipole.
Figure 5.1 is a schematic illustration of the relation between the spontaneous emission
decay rate of a dipole to the properties of the environment in which the dipole is
embedded. In Fig. 5.1, one can see a dipole being embedded within a homogenous
medium (Fig. 5.1a), inhomogeneous medium with fixed properties (Fig. 5.1b), and
inhomogeneous medium with tunable properties (Fig. 5.1c).

Figure 5.1: Schematic illustration of Fermi’s golden rule. Here a dipole is assumed to be
embedded in a (a) homogenous medium, (b) inhomogeneous medium, and (c) voltage-tunable
inhomogeneous medium.

As mentioned in the previous chapters, different modulation mechanisms have been
proposed to modulate the optical response of nanostructures in different wavelength
regimes. Dynamic control of the spontaneous emission rate of epitaxial QDs coupled
to high quality factor photonic crystal cavities has been experimentally demonstrated
[205], [206]. However, these experiments required a cryogenic temperature ambient,
making them less amenable for immediate practical applications.
This limitation can be alleviated by using active plasmonic structures to tune the
emission of broadband room-temperature solid-state emitters. Thanks to their small
optical mode volumes and relatively low quality factors, tunable plasmonic structures
can eliminate the necessity of careful alignment of the quantum emitters and the
structure resonances.
In a recent study, LDOS of visible-emitting colloidal QDs was manipulated via fieldeffect-induced optical permittivity modulation of ultrathin degenerately doped TiN in
a gated TiN/SiO2/Ag plasmonic heterostructure [98]. The heterostructure consisted of
80 nm-thick Ag and 9 nm-thick SiO2 layers in which indium phosphide (InP) QDs were

112
embedded, followed by a 7 nm layer of TiN (see Fig. 5.2a). In this study, degenerately
doped n-type TiN was used because of its ENZ wavelength which is located in the
visible range. Visible-emitting InP/ZnS core-shell colloidal QDs were embedded in the
insulating SiO2 spacer with a filling factor of 9%. The involved QDs were heavy-metalfree, and hence of great application interest, accounting for health and environmental
concerns.
The fabricated TiN films were n-type with carrier densities ranging from 5.9 × 1020 to
4.1 × 1022 cm−3. Depending on the carrier density, the fabricated TiN films could be
optically dielectric (Re(ɛ)>0) or optically plasmonic (Re(ɛ)<0). When a bias was
applied between TiN and Ag, a charge depletion or accumulation layer was formed in
TiN at the interface with SiO2, with tunable real and imaginary parts of the permittivity
(see Fig. 5.2b, c). This resulted in a modulation of the complex refractive index of TiN,
and consequently, tuning of the reflection from the heterostructure via changing the
applied bias (see Fig. 5.2d).
The optical measurements performed on the TiN/SiO2/Ag heterostructures showed a
reflectance increase from 67% to 82% at the QD emission wavelength of λ=630 nm,
when the gate voltage was varied from –1 V to 1 V, with a modulation speed of
exceeding 20 MHz. Due to the modulation of the refractive index of the TiN in its active
region, one could achieve precise control over the LDOS at the position of QDs
embedded in the SiO2 layer (see Fig. 5.2e).
As a result of the Fermi’s golden rule, the LDOS modulation led to a voltage-tunable
lifetime (see Fig. 5.2f) and photoluminescence (PL) (see Fig. 5.2g) of the QDs. The
time-resolved PL intensity measurement of the InP QDs showed a 12% decrease of the
QD lifetime when the gate voltage VG was increased from 0 V to +1 V, and an 18%
increase of the QD lifetime when VG was varied from 0 V to –1 V, leading to a total
amount of 30% modulation of the lifetime of QDs. Moreover, the device provided a
10% relative increase in the PL intensity when the gate voltage VG was varied from 0
V to +1 V, and a 5% relative decrease in the PL intensity when VG was varied from 0
V to –1 V. In addition to the lifetime and PL intensity modulation, a 26% increment of
the radiative decay rate was observed when changing the applied bias. This led to a
56% increase of the quantum yield as the gate voltage VG varied from –1 V to 1 V (see
Fig. 5.2h).

113

Figure 5.2: TiN/SiO2/Ag plasmonic heterostructure used for active control of spontaneous
emission of QDs. (a) Schematic of the gated plasmonic heterostructure. Cross-sectional TEM
image of the fabricated heterostructure (top left). The scale bar is 10 nm. High-resolution TEM
image of InP QDs with a diameter of 4–5 nm (top right). The scale bar is 5 nm. Measured (b)
real and (c) imaginary parts of the complex dielectric permittivity of TiN films. The gray dotted
line in (b) denotes Re(ɛ) = 0. For comparison, the dielectric permittivity values for gold and
silver [207] are plotted. (d) The measured reflectance spectrum of the gated plasmonic
heterostructure for different applied voltages. The inset shows the heterostructure reflectance
as a function of voltage at a wavelength of λ=630 nm. (e) Calculated LDOS enhancement
spectra at the position of a QD (averaged over QD dipole orientations) for different carrier
densities in a 1 nm-thick modulated TiN layer. The black curve corresponds to a homogeneous
TiN film, which is in the ENZ region. The red curve corresponds to a TiN film with a 1 nmthick modulated TiN layer that is plasmonic but far from the ENZ region. The top panels show
the simulated spatial distribution of the electric field |E| radiated by a QD (λ=630 nm). (f) PL
lifetime of QDs embedded in the gated heterostructure. The inset shows the PL intensity as a
function of time for different gate voltages VG. (g) Modulation of the PL intensity of InP QDs
embedded in the gated heterostructure at the wavelength of λ=630 nm. The inset shows the PL
intensity spectra for different gate voltages. (h) Radiative decay rate (top panel) of InP QDs
(normalized to radiative decay rate at zero bias) embedded in the plasmonic heterostructure as
a function of gate voltage. Dynamically tunable quantum yield (bottom panel) of QDs
(normalized to quantum yield at zero bias).

114
As could be seen, the described platform could provide a promising scheme for
modulating the spontaneous emission decay rate of quantum dots and quantum emitters
in the visible wavelength range.
In the near- and mid-infrared regimes, modulation mechanisms have been adopted to
tune the near-infrared emission of erbium ions coupled to Salisbury-screen type
heterostructure via optically induced phase transition in VO2 [208] and electrically
controlling the Fermi energy of graphene sheets [209], [210]. However, very small
enhancement of the spontaneous emission rate could be achieved in such planar
structures.
To conquer this issue, using reconfigurable plasmonic metasurfaces seems to be a
promising approach to achieve tunable emission control with a notable enhancement of
the spontaneous emission rate of emitters. Tunable metasurfaces can provide unique
electromagnetic environments leading to precise control of the constitutive properties
of radiation. The ability to tune the optical resonances of active metasurfaces is a key
element to engineer enhancement of emission decay rate at desired wavelengths.
In this chapter, we propose an active plasmonic metasurface platform to achieve a
tunable enhancement of the spontaneous emission decay rate of quantum emitters. The
tunable metasurface operates based on a field-effect-induced modulation of the charge
carrier concentrations in an ITO layer when an external DC electric field is applied
across the layer. Figure 5.3 shows the spatial distribution of the number of charge
carrier concentrations (Fig. 5.3a), real (Fig. 5.3b), and imaginary (Fig. 5.3c) parts of the
ITO permittivity when a DC bias is applied between the ITO layer and another gate
with a gate dielectric between them. The bulk charge carrier concentration of the ITO
layer is chosen to be Nb = 11020 cm-3. The metal and the gate dielectric of the metalinsulator-semiconductor (MIS) heterostructure are picked to be gold (Au) and HAOL
[34], respectively.
As can be seen in Fig. 5.3a, when altering the applied bias between the Au gate and the
ITO layer, the number of charge carrier concentrations of the ITO layer is changed
especially in a thin region close to the interface of the ITO and the gate dielectric
(corresponding to the right-border position in Fig. 5.3). This will result in a modulation
of the real and imaginary parts of the electric permittivity of ITO.

115

Figure 5.3: Field-effect modulation of charge carrier concentration density and
permittivity of ITO. Spatial distribution of (a) number of charge carrier concentration, (b) real
and (c) imaginary part of the ITO permittivity for different applied biases. The bulk carrier
concentration of ITO is 1×1020. The right border (z = 7 nm) presents the interface of the ITO
and the gate dielectric.

As a consequence, when the mentioned tunable MIS heterostructure is incorporated
into a metasurface, the field-effect-induced modulation of the electric permittivity of
ITO enables an active tuning of the LDOS. This leads to a reconfigurable enhancement
of the spontaneous decay rate of the quantum emitters embedded within such a photonic
environment.
Figure 5.4 shows a schematic of the proposed active metasurface which is composed
of an Au back reflector on top of which an Al2O3 layer with quantum emitters embedded
within it is located. The Al2O3 layer is followed by an ITO layer on top of which a
HAOL gate dielectric and then Au fishbone antennas are placed.

Figure 5.4: Schematic of the gate-tunable metasurface used for active control of
spontaneous emission decay rate of quantum emitters via modulation of the local density
of optical states. Schematic of the (a) unit cell (the inset shows the top view of the unit) and
(b) periodic array of electro-optically tunable metasurface used for modulation of Purcell
enhancement. The metasurface consists of a gold back reflector, on top of which a host material
(Al2O3) doped by quantum emitters (Er3+ ions) is placed. An ITO layer followed by HAOL
gate-dielectric and gold antennas is then placed on top of the Er-doped Al2O3 layer.

116
The fishbone nanoantennas are composed of patch antennas that are connected by Au
stripes, which also serve as gate voltage control electrodes. As discussed in the previous
chapters, such a structure is able to provide a magnetic dipole plasmon resonance. By
coupling this resonance with the emission of the desired quantum emitters, one can
enhance the spontaneous emission decay rate of the emitters.
In order to achieve a tunable modulation of spontaneous emission decay rate in the NIR
wavelength range, Er3+ ion is chosen as the quantum emitter to be embedded in an
Al2O3 host material. Such a quantum emitter shows emission at a wavelength of 1535
nm.

5.3. Coupling of Quantum Emitters to the Plasmonic Active Metasurface
In order to effectively couple the 1535-nm emission of the Er3+ ions to the metasurface,
we first need to investigate how changing the structural dimensions of the metasurface
affects the resonance wavelength of the metasurface, and accordingly, the enhancement
of the spontaneous emission of the quantum emitter.
The following sections summarize the results of full-wave electromagnetic calculations
for quantum emitters embedded within our tunable metasurface using finite difference
time domain optical simulations (FDTD Lumerical) of one unit cell. In these
simulations, an electric dipole is embedded in the Al2O3 layer of the metasurface. A
perfectly matched layer (PML) is used as the boundary condition in the x-, y-, and zdirections (see Appendix D.1).

5.3.1. Effect of ITO and Al2O3 Thickness on the Purcell Enhancement
Figure 5.5 shows the simulation results describing the effect of the thickness of ITO
and Al2O3 layers on the Purcell enhancement of a quantum emitter. Here, the Purcell
enhancement is calculated using FDTD optical simulations via:

Purcell enhancement = 𝑃𝐸 = 𝑃𝑀𝑇𝑆

ℎ𝑜𝑚

(5.4)

where 𝑃𝑀𝑇𝑆 and 𝑃ℎ𝑜𝑚 are the power radiated by a dipole source in the metasurface and
a homogeneous medium (Al2O3 in this case), respectively.

117
In these calculations, the size of the unit cell is set to be p = 400 nm, and the thickness
of the back reflector, HAOL, antenna, and electrode are chosen to be tBR = 80 nm, tHAOL
= 10 nm, ta = 40 nm, and te = 40 nm, respectively. The length and width of the antenna
and the width of the electrode are la = 250 nm, wa = 100 nm, and we = 100 nm,
respectively in order for the metasurface to show a resonance at a region close to our
wavelength of interest res = 1535 nm. In each case, the quantum emitter (electric dipole)
is placed at the position of (xd, yd, zd) = (100 nm, 0, tAl2O3/2), where (x, y) = (0,0) is the
center of the unit cell and z = 0 is the back-reflector/Al2O3 interface. The dipole moment
is considered to be in the z-direction.

Figure 5.5: Effect of Al2O3 and ITO thickness on the Purcell enhancement. Purcell
enhancement as a function of wavelength and Al2O3 thickness for (a) tITO = 5 nm, (b) tITO = 7
nm, (c) tITO = 9 nm, (d) tITO = 11 nm.

As can be seen in Fig. 5.5, when changing either the ITO or the Al2O3 layer thickness,
the resonance wavelength and the peak value of the Purcell enhancement can be tuned.
In order to delve deeper into this investigation, we present the resonance wavelength
(res) and the value of the Purcell enhancement at resonance (Max {PE}) in Fig. 5.6.

Figure 5.6: Effect of Al2O3 and ITO thickness on the resonance wavelength and peak
intensity of the Purcell enhancement. (a) Resonance wavelength and Maximal Purcell
enhancement, (b) resonance wavelength, and (c) maximal Purcell enhancement for different
Al2O3 and ITO thicknesses.

As can be seen in Fig. 5.6, increasing the thickness of either the Al2O3 or the ITO layer
will redshift the resonance and decrease the maximal Purcell enhancement.

118
5.3.2. Effect of Antenna and Electrode Thickness on the Purcell Enhancement
Another important factor that can affect the Purcell factor of the emitter embedded
within the active metasurface, is the thickness of the antennas and the electrodes. Figure
5.7 summarizes the effect of antenna and electrode thickness (ta = te) on the Purcell
enhancement when other physical dimensions of the metasurface are set to be p = 400
nm, tBR = 80 nm, tAl2O3 = 16 nm, tITO = 7 nm, tHAOL = 10 nm, la = 250 nm, wa = 100 nm,
and we = 100 nm. The position of the emitter is (xd, yd, zd) = (100 nm, 0, 8 nm), where
(x, y) = (0,0) is the center of the unit cell and z = 0 is the back-reflector/Al2O3 interface.
The dipole is assumed to have a dipole moment along the z-direction.
As can be seen in Fig. 5.7a, when the thickness of the antenna is changed, one can see
a shift in the resonance position and also the value of the Purcell enhancement. The
resonance wavelength (res) and the value of the Purcell enhancement at this
wavelength (Max {PE}) for different thicknesses of the antenna are plotted in Fig. 5.7b.
As can be seen in Fig. 5.7b, increasing the antenna thickness from 30 nm to 90 nm, the
resonance blueshifts. Increasing the antenna thickness from 90 nm to 120 nm has almost
no effect on the resonance wavelength. Further increasing the antenna thickness will
first slightly increase and then decrees the resonance wavelength. Besides, a decrease
in the maximal Purcell enhancement value is observed when the antenna thickness is
increased.

Figure 5.7: Effect of antenna thickness on the Purcell enhancement. (a) Purcell
enhancement as a function of wavelength and antenna thickness. (b) Maximal Purcell
enhancement (black curve) and resonance wavelength (red curve) for different antenna
thicknesses.

119
5.3.3. Effect of Antenna Length, Antenna Width, and Electrode Width on the
Purcell Enhancement
Next, we study the effect of antenna length and width, and electrode width on the
emission of the dipole integrated into the metasurface. Figures 5.8a-c show the
spectrum of the Purcell enhancement for different antenna lengths, antenna widths, and
electrode widths, respectively. Here, we set p = 400 nm, tBR = 80 nm, tAl2O3 = 16 nm,
tITO = 7 nm, tHAOL = 10 nm, ta = te = 40 nm. The dipole is again placed at the position of
(xd, yd, zd) = (100 nm, 0, 8 nm), where (x, y) = (0, 0) is the center of the unit cell and z
= 0 is the back-reflector/Al2O3 interface. The dipole is assumed to be oriented along the
z-direction. As can be seen in Figs. 5.8a-c by altering the lateral dimensions of the
antenna and electrode, one can tune the resonance position and the amplitude of the
Purcell enhancement.

Figure 5.8: Effect of antenna width and length, and electrode width on the Purcell
enhancement. Purcell enhancement as a function of wavelength and (a) antenna length, (b)
antenna width, and (c) electrode width. Maximal Purcell enhancement (black curve) and
resonance wavelength (red curve) as a function of (d) antenna length, (e) antenna width, and (f)
electrode width.

The resonance wavelength (res) and the value of the Purcell enhancement at resonance
(Max {PE}) for different antenna lengths, antenna widths, and electrode widths are
presented in Fig. 5.8d, Fig. 5.8e, and Fig. 5.8f, respectively. As can be seen in Fig. 5.8d,
when increasing the antenna lengths, the resonance redshifts. The peak value of the
Purcell enhancement will first increase and then decrease. It is observed in Fig. 5.8e
that an increase in the antenna width will result in a resonance wavelength increment

120
accompanied by a reduction of the Purcell enhancement peak value. Furthermore, Fig.
5.8f shows that increasing the electrode width will lead to a smaller peak value of the
Purcell enhancement. This appends an initial reduction followed by an increment of the
resonance wavelength.
Based on the described calculations, the dimensions of the tunable metasurface are
picked to be p = 400 nm, tBR = 80 nm, tAl2O3 = 16 nm, tITO = 7 nm, tHAOL = 9.5 nm, ta =
te = 40 nm, la = 250 nm, wa = 90 nm, and we = 110 nm. The selected design can provide
resonance in the Purcell enhancement at our wavelength of interest λ = 1535 nm which
is indeed the emission wavelength of Er3+ ions that are being used as our quantum
emitters.

5.3.4. Effect of Dipole Position on the Purcell Enhancement
It can be conveyed from Figs. 5.5-5.8 that a pre-fabrication modulation of the resonance
wavelength and the Purcell enhancement value can be achieved via precisely designing
the dimensions of different layers constituting the active metasurface structure.
However, in addition to the physical properties of the metasurface, the position at which
the dipole is placed is also expected to play a key role in the spontaneous emission
decay rate of the emitter. To investigate this, we studied the effect of the dipole position
on the enhancement of the spontaneous emission decay rate.
Figure 5.9 illustrates the Purcell enhancement spectrum for different dipole positions
(xd, yd, zd). In this set of calculations, the dimensions of the metasurface are assumed to
be equal to the values obtained for the optimized design mentioned in the previous
section. The dipole is also assumed to be oriented along the z-direction.
Each subfigure in Fig. 5.9 depicts the Purcell enhancement spectrum when the dipole
is placed at a fixed (xd, yd) position and different zd values. It should be noted that to
prevent quenching, dipole should not be placed at close proximity of a lossy metallic
layer. As a result, the zd value ranges between z = 2 nm and z = 12 nm that correspond
to 2 nm above the back-reflector/Al2O3 interface, and 4 nm below the Al2O3/ITO
interface, respectively.
As can be seen in Fig. 5.9, at a fixed (xd, yd) value, increasing the z position of the dipole
(zd) will first decrease and then increase the peak value of the Purcell enhancement.
That is due to the fact that when increasing zd, the dipole is first moved away from the

121
lossy back reflector. Further increasing zd, the dipole gets closer to the lossy ITO layer,
which is accompanied by an increase in the field enhancement, and hence, an increment
of the Purcell enhancement. Moreover, Fig. 5.9 shows that a noticeable resonance can
be observed when the dipole is placed within part of the Al2O3 layer that is located
beneath the antenna and the electrode (see Figs. 5.9a,b,e, and f) with a higher Purcell
enhancement when the dipole is moved away from the center of the unit cell (see Fig.
5.9f). It can also be observed in Fig. 5.9 that changing the dipole location will not affect
the resonance position. As a result, the resonance wavelength is only dictated by the
physical dimensions of the metasurface.

Figure 5.9: Effect of dipole position on the Purcell enhancement. Purcell enhancement as
a function of wavelength for the (xd, yd) values of (a) (0, 0), (b) (0, 40 nm), (c) (0, 100 nm), (d)
(0, 200 nm), (e) (50 nm, 0), (f) (100 nm, 0), (g) (150 nm, 0), (h) (200 nm, 0).

5.4. Bias-Induced Modulation of Spontaneous Emission Decay Rate
In the previous section, we figured out how constituent layers of the plasmonic
metasurface could be precisely designed to provide a pre-fabrication modulation of the
metasurface resonance, leading to a prosperous coupling of quantum emitters to the
metasurfaces. After finding the best design that could provide resonance at the emission
wavelength of our desired quantum emitter, we investigate the bias-induced modulation

122
of the Purcell enhancement. To this end, a voltage (V) is applied between the Au
nanoantennas and the ITO layer, leading to a field-effect-induced modulation of the
permittivity of the ITO layer. As a result of the modulation of the LDOS, the
spontaneous emission decay rate could be reconfigured through the alteration of the
applied bias voltage. When modeling the performance of our structure under applied
bias, we use the voltage-dependent complex refractive index of the ITO layer presented
in Fig. 5.3.
Figures 5.10-5.12 illustrate the bias-induced modulation of the spontaneous emission
̂ 𝑝 = 𝑥̂), y- (𝒏
̂𝑝 =
decay rate of the quantum emitter with a dipole moment along the x- (𝒏
̂ 𝑝 = 𝑧̂ ) directions, respectively. The Purcell enhancement spectra for
𝑦̂ ), and z- (𝒏
different applied biases are shown in Figs. 5.10a, 5.11a, and 5.12a.
As can be seen, when changing the applied bias, the Purcell enhancement could be
considerably modulated. It is also notable that when the dipole moment is along x- (Fig.
5.10a) and y- (Fig. 5.11a) directions, no evident resonance can be observed in the
Purcell enhancement. Figures 5.10b, 5.11b, and 5.12b show the relative Purcell
enhancement change obtained from:
𝑃 (𝑉)−𝑃 (𝑉=0)

∆𝑟𝑒𝑙 𝑃𝐸 (𝑉) = 100 × 𝐸 𝑃 (𝑉=0)

(5.5)

When changing the applied bias, notable relative Purcell enhancement changes are
observed at our wavelength of interest  = 1535 nm.
Once a quantum emitter is excited by absorbing a photon, it may undergo a transition
to the ground state either by far-field photon emission (radiative pathway) or via nonradiative processes such as energy transfer via non-radiative dipole-dipole coupling
[211]. As a result, an important factor that needs to be taken into consideration is the
radiative decay rate enhancement provided by the metasurface. Figures 5.10c, 5.11c,
and 5.12c depict the radiative decay rate enhancement obtained through:
𝑃𝑟𝑎𝑑 =

𝑃𝑜𝑢𝑡𝑐𝑜𝑢𝑝𝑙𝑒𝑑
𝑃ℎ𝑜𝑚

(5.6)

where 𝑃𝑜𝑢𝑡𝑐𝑜𝑢𝑝𝑙𝑒𝑑 is the power that is transferred from the whole structure with an
embedded dipole to the surrounding media, and 𝑃ℎ𝑜𝑚 is the power radiated by a dipole
source in a homogeneous Al2O3 layer.

123
It can be seen in Figs. 5.10c, 5.11c, and 5.12c that the radiative decay rate enhancement
is much smaller than the Purcell enhancement. That is because a considerable part of
the power radiated by the quantum emitter is absorbed by the plasmonic lossy structure,
reducing the outcoupled power. The high amount of loss in this plasmonic metasurface
directly relates to the strong field enhancement in the lossy accumulation region of the
ITO layer, and accordingly, is an inevitable component of the operating principle of
our voltage-tunable metasurface. While the metasurface provides a notably smaller
radiative decay rate enhancement over the Purcell enhancement, it can be observed that
changing the applied bias will result in modulation on the power outcoupled from the
quantum emitter embedded within the metasurface.
It can also be seen from Figs. 5.10d, 5.11d, and 5.12d that changing the applied bias
will result in relative radiative decay rate enhancement changes for a dipole, obtained
via:

(𝑉)−𝑃𝑟𝑎𝑑 (𝑉=0)
𝑟𝑎𝑑 (𝑉=0)

∆𝑟𝑒𝑙 𝑃𝑟𝑎𝑑 (𝑉) = 100 × 𝑟𝑎𝑑 𝑃

(5.7)

Another important metric to evaluate the emission properties of a quantum emitter that
quantifies how an excited emitter will relax via the radiative pathway is quantum
efficiency. Figures 5.10e, 5.11e, and 5.12e show the quantum efficiency of the dipole
obtained through:
𝑄𝐸 =

𝑃𝑟𝑎𝑑
𝑃𝐸

(5.8)

As can be seen, when the applied bias is changed, the quantum efficiency of the system
could be efficiently modulated. This can be clearly observed from Figs. 5.10f, 5.11f,
and 5.12f which present the relative quantum efficiency change obtained through:
∆𝑟𝑒𝑙 𝑄𝐸(𝑉) = 100 ×

𝑄𝐸(𝑉)−𝑄𝐸(𝑉=0)
𝑄𝐸(𝑉=0)

(5.9)

It should be noted that in Figs. 5.10-5.12, the dipole is located at a position of (xd, yd,
yd) = (100 nm, 0, 8 nm) where the Al2O3/back reflector interface is located at z = 0.
As can be seen, by changing the bias voltage applied to our tunable metasurface, the
Purcell factor, radiative decay rate, and the quantum efficiency of a quantum emitter
coupled to the metasurface could be modulated.

124

Figure 5.10: Bias-induced modulation of the spontaneous emission decay rate of a
quantum emitter with a dipole moment along the x-direction. (a) Purcell enhancement, (b)
relative Purcell enhancement change, (c) radiative decay rate enhancement, (d) relative
radiative decay rate enhancement change, (e) quantum efficiency, and (f) relative quantum
efficiency change spectra for different applied biases. Vertical dashed lines indicate our
wavelength of interest  = 1535 nm which coincides with the emission wavelength of the
Er3+ions.

Figure 5.11: Bias-induced modulation of the spontaneous emission decay rate of a
quantum emitter with a dipole moment along the y-direction. (a) Purcell enhancement, (b)
relative Purcell enhancement change, (c) radiative decay rate enhancement, (d) relative
radiative decay rate enhancement change, (e) quantum efficiency, and (f) relative quantum
efficiency change spectra for different applied biases. Vertical dashed lines indicate our
wavelength of interest  = 1535 nm which coincides with the emission wavelength of the
Er3+ions.

125

Figure 5.12: Bias-induced modulation of the spontaneous emission decay rate of a
quantum emitter with a dipole moment along the z-direction. (a) Purcell enhancement, (b)
relative Purcell enhancement change, (c) radiative decay rate enhancement, (d) relative
radiative decay rate enhancement change, (e) quantum efficiency, and (f) relative quantum
efficiency change spectra for different applied biases. Vertical dashed lines indicate our
wavelength of interest  = 1535 nm which coincides with the emission wavelength of the
Er3+ions.

The results presented so far, have been obtained for a dipole oriented along one of the
x-, y-, or z-directions. However, in practice, the quantum emitters will have random
dipole moments. As a result, to make the simulations closer to the practical case, one
would average over the cases with the dipole moments along the x-, y-, and z-directions.
Figure 5.13 presents the Purcell enhancement (Figs. 5.13a, b), radiative decay rate
enhancement (Figs. 5.13c, d), and quantum efficiency (Figs. 5.13e, f) spectra for
different applied biases. Here, the spectrum of each parameter is obtained by averaging
over the spectra of the corresponding metric when the dipole moment is oriented along
the x-, y-, and z-directions.
As can be seen, in the case of an averaged dipole moment, the Purcell enhancement,
and the radiative decay rate enhancement values are about 1/3 of those in the case of a
dipole moment along the z-direction. As a consequence, the quantum efficiency values
remain almost the same as that of the z-directed dipole moment. Moreover, it can be
observed that when changing the applied bias, one can easily tune the Purcell
enhancement, the radiative decay rate enhancement, and quantum efficiency at a broad
range close to our wavelength of interest  = 1535 nm.

126

Figure 5.13: Bias-induced modulation of the spontaneous emission decay rate of a
quantum emitter with a randomly-oriented dipole moment. (a) Purcell enhancement, (b)
relative Purcell enhancement change, (c) radiative decay rate enhancement, (d) relative
radiative decay rate enhancement change, (e) quantum efficiency, and (f) relative quantum
efficiency change spectra for different applied biases. Vertical dashed lines indicate our
wavelength of interest  = 1535 nm which coincides with the emission wavelength of the
Er3+ions.

To gain better insight, Fig. 5.14 shows the Purcell enhancement (Fig. 5.14a), radiative
decay rate enhancement (Fig. 5.14b), and quantum efficiency (Fig. 5.14a) of the
averaged system as a function of applied bias at a different operating wavelength
around our wavelength of interest.

Figure 5.14: Active control of the emission from a randomly-oriented dipole embedded in
the tunable metasurface. (a) Purcell enhancement, (b) radiative decay rate enhancement, and
(c) quantum efficiency as a function of applied bias at different wavelengths around the
emission wavelength of the Er3+ ions ( = 1535 nm).

As can be seen, in the presented wavelength range, once could achieve a considerable
amount of modulation in the mentioned metrics. It can also be observed that when

127
increasing the applied bias, the Purcell enhancement first decreases. Further increasing
the applied bias, we will notice an increase in the Purcell enhancement at the voltages
greater than the ENZ voltage of the ITO. At the emission wavelength of the Er3+ ions
( = 1535 nm), modulations of the Purcell enhancement, the radiative decay rate
enhancement, and the quantum efficiency of 24%, 51%, and 36%, are obtained,
respectively.

5.5. Conclusions and Outlook
In this section, we presented the ability to electrostatically control the spontaneous
emission rate of quantum emitters via modulation of LDOS. To this end, the quantum
emitters are coupled to a tunable resonant metasurface whose properties could be
modulated by applying an external bias voltage. The metasurface consists of an ITO
layer that serves as an active medium with tunable properties. By changing the applied
bias, the complex refractive index of the ITO layer is altered, leading to a modulation
of the LDOS. By incorporating quantum emitters into the metasurface, and aligning the
resonant wavelength of the structure with the emission wavelength of the emitter, the
spontaneous emission of the quantum emitters could be drastically enhanced. Then, the
configuration could yield a large modulation of the spontaneous emission when the
applied bias is changed.
The presented optoelectronic device exemplifies how conventional electronic
components, in our case, a metal-oxide-semiconductor capacitor, can be adapted to the
field of nanophotonics. This can be achieved by establishing a bias-controlled
modulation of the emitted intensity without an electrical injection of carriers into the
quantum emitters and can lead to a number of applications. Emission control of
quantum emitters at a constant optical pumping rate is a cornerstone of modern highquality lighting and display technologies, biology, photovoltaic devices, and other
optoelectronics applications.

128
Chapter 6

DIELECTRIC TUNABLE METASURFACES
The material in this chapter was in part presented in [71], [117].
Modulating constitutive properties of light using plasmonic active metasurfaces has
paved the way towards novel functions and applications unachievable using
conventional optical components. While interesting, plasmonic metasurfaces possess a
high amount of optical loss, reducing the efficiency of the device and the system it is
integrated into. Moreover, the plasmonic structures presented so far, require a large
field enhancement at the dielectric spacer sandwiched between a metal back reflector
and metallic antennas. This requirement, make the mentioned metasurfaces work only
in reflection mode. Dielectric metasurfaces are appropriate replacements to the lossy
plasmonic structures by providing higher efficiencies and the ability to operate in both
reflection and transmission modes. In this chapter, we propose and study all-dielectric
electro-optically tunable metasurfaces. Two approaches that could provide dielectric
tunable metasurfaces will be investigated. First, a theoretical approach towards a fieldeffect-based electrically tunable metasurface, which can achieve relatively large phase
modulation in both reflection and transmission modes (dual-mode operation) will be
proposed. The metasurface consists of Si slab followed by an alumina layer, on top of
which Si nanodisks connected via Si nanobars are located. By incorporating an ultrathin
layer of ITO as an electro-optically tunable material between the Si slab and the alumina
layer, we report an approach for active tuning of all-dielectric metasurfaces. Two
separate resonances are excited in the reflection and transmission modes by the Si
nanobars and the Si slab, respectively. This enables highly confined electromagnetic
fields at the ITO-alumina interface. When changing the bias applied between the ITO
layer and the nanobars, the charge carrier concentration, and accordingly, the refractive
index in the ITO accumulation layer will be varied. This leads to 240° phase-agility at
an operating wavelength of 1696 nm for the reflected transverse electric (TE) polarized
beam and 270° phase shift at 1563 nm for the transmitted transverse magnetic (TM)
polarized light. Having independent and isolated control of the reflection and
transmission modes, one would be able to achieve distinctly different functions for each
operation mode.

129
Next, we will show the design and experimental demonstration of high-performance,
low-loss optoelectronic modulators exploiting the electro-optic effect in multiplequantum-well heterostructures [117]. An all-dielectric active metasurface based on
electro-optically tunable III-V MQWs patterned into subwavelength elements will be
presented. Each metasurface element can support a hybrid Mie-guided mode (Mie-GM)
resonance. These volumetric Mie-GM resonances could be actively modulated using
the quantum-confined Stark effect (QCSE). This will lead to a relative reflectance
modulation of 270% and a phase shift from 0 to ~70°. By actively changing the
metasurface period via individually controlling the electrical bias applied to each
metasurface element, we demonstrate beam switching and dynamic beam steering. This
approach could be employed to realize other optical functions such as tunable
metalenses, active polarizers, and flat spatial light modulators.

6.1. Dielectric Tunable Metasurfaces
In the last several years, metasurfaces, artificially-designed arrays of subwavelength
optical scatterers, have been employed to achieve versatile and comprehensive control
of the key constitutive properties of light at the nanoscale [4], [212]–[214]. To date,
metasurfaces have been used to demonstrate a number of low-profile optical
components with important capabilities including wavefront engineering [1], [215],
focusing [7], [8], [72], [140], [153], [216], polarization control and detection [166],
[217]–[220], holographic imaging [73], [221], and quantum light control [164], [222],
[223].
Recently, actively-tunable metasurfaces have emerged as a transformational concept in
the field of photonics owing to their ability to provide post-fabrication modulation, and
thereby, the capability to control the wavefront of the scattered light in real-time. This
leads to the demonstration of a wide range of photonic functions [61], [161], [224], and
hence, the realization of compact and fast low-profile nanophotonic devices capable of
beam steering, active polarization switching, and formation of reconfigurable
metalenses. To date, different approaches have been employed in order to realize
reconfigurable metasurfaces. In these approaches, the reconfigurable metasurfaces are
commonly obtained by incorporating an active material into the otherwise passive

130
metasurface structures. Then, by applying an external stimulus, the dielectric
permittivity of the active material can be dynamically controlled.
Amongst the active meta-devices presented so far, the metasurfaces hybridized with
tunable CMOS compatible materials, such as highly-doped semiconductors like ITO
[46], [80], [84], [90], [93], [119], [225]–[229] have enabled an approach to achieve
strong light-matter interaction with ultrafast control over the optical properties of
individual unit cell elements, which in aggregate enable the realization of tunable
spatially-varying phase/amplitude gradients.
Thanks to its large refractive index change, possessing an ENZ condition, and short
response

time

at

NIR

wavelengths

including

particularly the

1.55

μm

telecommunication wavelength [34], [35], [67], ITO has attracted a great deal of
interest to be used as an electro-optically tunable active medium. Despite remarkable
progress toward ITO-based active metasurfaces in the reflection mode [34], [35], [67],
[230]–[232], there has been a lack of demonstration of reconfigurable devices in the
transmission mode via an ultrathin flat metasurface based on ITO permittivity
modulation.
A theoretical study of a transmissive structure based on ITO-integrated multi-material
nanowires was proposed [233]. However, the non-planar design of the metasurface
would make the fabrication of the device not to be very favorable. In another study,
metallic slits filled with ITO [234] were utilized to provide tunable metasurfaces
working in transmission mode. The design, however, was several wavelengths thick,
and as a result, not appropriate to be employed as a low-profile optical component that
can be used in photonic integration. It is noteworthy that ITO-based tunable
metasurfaces

were

proposed

that

could

provide

modulation

of

the

reflection/transmission amplitude via coupling the incident beam to guided mode (GM)
resonances [235], [236] or employing Huygens modes in dielectric resonators [62].
However, no phase modulation was achievable using these devices. Moreover, tunable
metasurfaces presented so far have been only able to perform phase modulation in
either reflection or transmission mode at a predesigned operating wavelength.
In this chapter, we will present an ITO-based all-dielectric metasurface that could
provide phase modulation in both reflection and transmission modes. The proposed
metasurface is mainly composed of Si which is of particular interest for on-chip

131
photonic integrated devices and can separately control the transformations of both
reflected and transmitted beams in real-time at two different wavelengths (dual-mode
operation). To achieve such a device, one needs to have a careful choice of the
constituent materials and the geometrical shape of nanoresonators. The structural
parameters of the nanostructure should then be optimized in order for the device to
provide dual-mode tunability.
As mentioned in the previous chapters, when applying an external bias across the ITO
layer, one could observe a modulation of the dielectric permittivity in a very small
region. As a result, these hybrid metasurfaces operate by spectrally overlapping the
geometrical antenna resonance and the ENZ permittivity regime, and also spatially
overlapping the metasurface element mode profile with the active ITO layer. This
places a stringent requirement of having resonant excitation of highly confined fields
in the active layer to achieve ITO-based tunable metasurfaces with widely-tunable
optical responses. On the other hand, the strong field confinement in the accumulation
region of ITO will lead to a considerable field absorption in this region. This will result
in a limited efficiency of the ITO-based tunable metasurfaces, necessitating the search
for alternative approaches to achieve reconfigurable metasurfaces.
Along the same line, an all-dielectric GaAs tunable metasurface was previously
proposed that could achieve refractive index modulation via free carrier generation by
using an optical pump [237]. Even though this approach could enable a picosecond
response time, the requirement of an ultrafast pump laser source is not desirable for
many low-power compact nanophotonic applications. Moreover, when using optical
pumping as the modulation mechanism, the area at which the refractive index is tuned
will be determined by the size of the focused laser spot. As a result of a large laser spot
size, individual control of subwavelength metasurface elements would not be possible,
limiting the applications of optically-tunable platforms. Consequently, to achieve
independent control of individual metasurface elements, electrical modulation of the
optical response of the metasurface is preferable.
Prior research has shown that metasurfaces incorporating patterned graphene layers
[58], [238], and ITO-integrated metasurfaces [95] with individually-controllable
elements could actively modulate the properties of the scattered light via application of
a bias voltage. However, the efficiency of the mentioned plasmonic metasurfaces is
low. Thus, developing a dielectric active metasurface platform that can dynamically

132
tailor the wavefront of scattered light by modulation of individual antenna elements
remains an outstanding research challenge.
Electrical tuning of the coupling between metasurface resonances and intersubband
transitions in multiple-quantum-wells has been explored for applications such as
tunable filters and optical modulators at mid-infrared wavelengths [239]–[241].
In this chapter, we will describe an MQW-based electrically tunable metasurface
platform that could provide amplitude and phase modulation in the NIR wavelength
range. This all-dielectric metasurface utilizes III-V compound semiconducting MQW
structures as resonant elements. When applying an electric field across the MQW
metasurface elements, the complex refractive index of the MQW will be electrooptically tuned via the QCSE [242] especially at wavelengths near the MQW bandgap.
This leads to a continuous modulation of the optical response of the metasurface. The
QCSE is widely used in high-performance electro-optical components such as highspeed modulators [243].
In this approach, we combine the well-established MQW technology with
subwavelength antennas to create an active metasurface platform for diverse
nanophotonic applications. Each MQW resonator of our metasurface design supports a
hybrid resonant mode with a relatively high quality factor, enabling optical modulation
under applied bias. This active device concept is then employed to experimentally
demonstrate beam steering by electrically controlling the optical response of individual
metasurface elements.

6.2. Si-based Dielectric Active Metasurfaces
6.2.1. Geometry and Structural Parameters of the Si-based Dielectric Active
Metasurfaces
In order to design the metasurface, first, a parametric study is performed to define a
fixed set of structural parameters that can support strongly confined electromagnetic
modes coupled with the carrier density modulation of ITO in both transmission and
reflection modes at two separate operating wavelengths. The building block of the
proposed Si-based metasurface is a semiconductor-insulator-semiconductor (SIS) unit
cell. The metasurface consisted of a Si backplane followed by an ITO layer on top of
which an alumina layer is located. On top of the alumina layer, Si nanodisk antennas

133
connected via Si nanobars (which serve as bias lines) are placed. The whole structure
is placed on as SiO2 substrate. Figures 6.1a, b show schematics of the metasurface
structure. The thickness of the Si back slab, the ITO layer, and the alumina gatedielectric layer are picked to be hSlab = 200 nm, hITO = 24 nm, and hAl2O3 = 12 nm,
respectively. The radius of the nanodisks is chosen to be R=320 nm, the antennas have
a thickness of hDisk = 310 nm, and the width of the nanobars is w=100 nm. The
metasurface has a period of p = 700 nm.
The ITO layer is degenerately doped with a background carrier concertation of 3×1020
cm-3. Moreover, in order for the top Si nanodisks and nanobars to serve as bias
electrodes for applying voltage, they are selected to be highly n-doped with a
background carrier concentration of 1×1019 cm-3. When applying an external bias
voltage between the Si nanodisks and the ITO layer, an accumulation/depletion layer
will be formed at the ITO-Al2O3 interface depending on the applied bias. The spatial
distribution of charge carrier concentrations within the ITO and Si layers for different
applied voltages are then calculated by self-consistently solving the Poisson and driftdiffusion equations, using the Lumerical Device simulator [244].
The spatial charge carrier distribution inside the ITO active layer as functions of applied
bias is shown in Fig. 6.1c. As can be seen, increasing the external bias voltage from 0
V to 12 V, the carrier concentration will be increased from 3.25×1020 cm-3 to 9.86×1020
cm-3 at the ITO-Al2O3 interface. It should be noted that since the Si back slab is undoped
with a work function quite close to the work function of the degenerately doped ITO, a
negligibly small band bending is observed at the interface of ITO and the Si back slab.
The electron and hole carrier distributions inside the Si active layer as functions of
applied bias voltage are also shown in Figs. 6.1d and f, respectively. It can be noted that
at a threshold voltage of VT = 4V, the thickness of the depletion layer reaches its
maximum. Further increasing the voltage, the hole carriers are generated and
accumulated at the Si-Al2O3 interface. Similar to the electron densities inside the ITO
active layer, the hole densities inside Si exponentially attenuate by moving away from
the Si-Al2O3 interface.
After performing the device physics simulations, the spatial distributions of the charge
carriers within the active regions of the structure are translated to the spatial
distributions of permittivity via the picked dispersion models. Here, to describe the
dielectric permittivity of ITO, we use a Drude model mentioned in Eq. (2.1). In order

134
to obtain the carrier-induced changes of the complex dielectric permittivity of highlydoped Si nanodisks/nanobars, we employ a Plasma-Drude model as
𝑒2

𝜖𝑆𝑖−𝑑𝑜𝑝𝑒𝑑 (𝜔) = 𝜖𝑆𝑖−𝑢𝑛𝑑𝑜𝑝𝑒𝑑 (𝜔) − 𝜖 (𝑚∗ 𝜔+𝑖𝑒/𝜇 + 𝑚∗ 𝜔+𝑖𝑒/𝜇 )

(6.1)

where N (P) is the electron (hole) density, µN (µP) is electron (hole) mobility and
m*N  mP*  is electron (hole) conductivity effective mass. Here, we use µN = 80 cm V S ,

-1 -1

µP = 60 cm2V-1S-1, 𝑚𝑁
= 0.27𝑚0 , and 𝑚𝑃∗ = 0.39𝑚0 [245].

Figure 6.1: Si-based dual-mode tunable metasurface. (a) Schematic presentation of the Sibased tunable metasurface including a periodic array of ITO-integrated Si nanodisks located on
top of a high index Si slab and a SiO2 substrate. The metasurface is capable of providing wide
phase modulation in both reflection and transmission modes at the operating wavelengths of λ1
and λ2. (b) The unit cell of the Si-based active metasurface. Spatial distribution of (c) electron
carriers in ITO, (d) electron, and (e) hole densities in the Si active layer as functions of applied
bias and distance from Si-Al2O3 interface [71].

It should be remarked that while the Si active layer is thicker than the ITO active layer,
the changes in the complex permittivity of Si is much smaller in comparison with that
of ITO. In particular, an almost two-fold decrement in the real part of permittivity of
ITO can be obtained in a region narrower than 3 nm-thick at the ITO-Al2O3 interface,
at the wavelength of  = 1.55 µm. However, one can observe less than 1% increment of

135
the Si permittivity in a region smaller than 20 nm-thick in the case of electron depletion,
and less than 40% decrement in the real part of permittivity in a region narrower than
1.5 nm-thick in the case of hole accumulation at the wavelength of  = 1.55 µm.
Moreover, the formation of the ENZ region in the ITO layer enhances the light-matter
interactions substantially. As a result, the main contributing factor to the tunable optical
response of the metasurface is the charge accumulation inside the ITO layer. It was also
shown previously that the modification of Im{εITO} has an important role in achieving
considerable wide phase modulation through satisfying the critical coupling condition
(so-called impedance matching) [246]–[251].
In the current design, the refractive indices of undoped Si, Al2O3, and SiO2 are derived
from the experimentally-measured data by Palik [252].

6.2.2. Optical Response of the Si-based Dielectric Active Metasurfaces
In order to investigate the optical response of the tunable metasurface, a customdeveloped solver based on rigorous coupled-wave analysis (RCWA) [253], [254]
method is employed.
To study the optical response of the tunable metasurface in reflection mode, the
metasurface is illuminated from the SiO2 substrate by a normal-incident TE polarized
plane-wave with the electric field vector being normal to the plane of incidence as
illustrated in Fig. 6.1a. Figures 6.2a, b show the reflection amplitude and phase of the
metasurface in the absence of gate bias. As can be seen, in this mode, the unit cell
supports a geometrical resonance with a high quality factor, accompanied by ~2π
spectral phase variation at the operating wavelength of λRefl ≈ 1696 nm.
The spatial distribution of the y-component of the magnetic field (|Hy|, color bar) and
the normalized electric displacement currents (Ex and Ez, white arrows) in the x-z plane
are shown in Fig. 6.2c. As can be seen in Fig. 6.2c, there is a rotating electric
displacement current loop around the normal component of the magnetic field which is
a well-known characteristic of magnetic resonances. Moreover, strong confinements of
the magnetic fields could be observed both at the center of the Si nanodisk, far from the
ITO layer and below the Si bias lines at the vicinity of the ITO layer. Having the filed
confinement in a region far from the ITO active layer will not lead to a notable spectral
shift or large phase modulation of the excited geometrical resonance. On the other hand,

136
part of the electromagnetic field confined below the Si bias lines would be able to
enhance the effect of a slight modification of carrier concentration within the ultrathin
ITO accumulation layer due to the vicinity to the ITO layer. As a result, one could
expect a notable modulation of the optical response of the metasurface contributed to
this mentioned field confinement.

Figure 6.2: Dual-mode Si-based metasurface operating in the reflection mode. (a)
amplitude and (b) phase of the reflection from the Si-based dielectric metasurface illuminated
by x-polarized incident beam for the unbiased case in the presence (solid blue lines) and absence
(dashed red lines) of the bias lines. Spatial distribution of the electromagnetic fields in the x-z
plane at the resonant wavelengths of (c) 1696 nm in the presence of the bias lines (d) 1691 nm
in the absence of bias lines. The color bar shows the magnitude of the y-component of the
magnetic field and the white arrows present the normalized electric displacement currents. The
spectrum of the (e) amplitude and (f) phase of the reflection from the metasurface under applied
bias. The dashed lines denote the points for which the real-part of ITO permittivity, Re{εITO},
at the ITO-Al2O3 interface is equal to −1, 0, and 1, illustrating the ENZ region of the ITO
accumulation layer. (g) Amplitude and (h) phase of the reflection from the metasurface as a
function of applied voltage at the operating wavelength of λRefl ≈ 1696 nm. The shadowed
regions present the dielectric breakdown of the gate dielectric [71].

In order to delve deeper into the effect of the bias lines, the amplitude and phase
response of the metasurface without the bias lines are also plotted in Figs. 6.2a, b. As
can be seen, in the absence of the bias lines, a low Q-factor resonance as well as a
spectral phase variation of ~π would be obtained at the operating wavelength of λ =
1691 nm. Figure 6.2d presents the spatial distribution of the magnetic field within the
metasurface with no bias line. As can be seen, no strong field enhancement is observed

137
near the ITO active region in this case. As a consequence, the Si nanobars, which serve
as the electrodes that electrically connect the nanodisk antennas, could give rise to the
required resonant reflection in a critical spectral regime for phase modulation by
breaking the geometrical symmetry of the unit cells.
The spectra of the amplitude and phase of the reflection from the metasurface under an
applied bias voltage are depicted in Figs. 6.2e, f, respectively. As can be seen, when
increasing the applied bias, the resonance will first blueshift and then redshift. Figures
6.2g, h present the reflection amplitude and phase, respectively at the operating
wavelength of λRefl ≈ 1696 nm. One can observe an amplitude modulation accompanied
by large phase modulation of ~240˚ with an applied gate bias lower than the breakdown
threshold of 10.5 V.
After confirming the tunable optical response of the metasurface in reflection mode,
the metasurface is illuminated by a TM polarized incident with the magnetic field
vector being normal to the plane of incidence as shown in Fig. 6.1a, to examine the
optical response of the metasurface in transmission mode. Similar to the reflection
mode, the geometrical resonance of the structure should overlap with the ITO
accumulation layer to achieve relatively large electrical modulation of the optical
response of the metasurface in transmission mode. The amplitude and phase of the
transmission through the metasurface are plotted in Figs. 6.3a, b, respectively.
As can be seen, the transmission response features two resonances: i) a high-Q-factor
resonance and a rapid transmission phase variation at the resonant wavelengths of λ1Tran
≈ 1563 nm, and ii) a low-Q-factor resonance and a slow transmission phase variation
at the resonant wavelengths of λ2Tran ≈ 1641 nm. The spatial distributions of the
electromagnetic fields in the y-z plane at these two resonance wavelengths are depicted
in Figs. 6.3c, d.
As can be seen in Figs. 6.3c, d, at both wavelengths, there is a circulation of the electric
displacement currents around the strengthened magnetic field at the center of the unit
cell, confirming the magnetic nature of the corresponding resonances. It can also be
observed that there is a confinement of the magnetic field between the Si nanodisk and
the Si back slab in the vicinity of the ITO active layer. However, the strength of the
confined magnetic field at λ1Tran is nearly 2.2 times larger than the one at λ2Tran. This
large local field enhancement at λ1Tran could be coupled to the ENZ region of the ITO

138
layer which itself would lead to a significantly enhanced electric field to satisfy the
continuity of the normal displacement field. As a consequence, one can expect a large
phase modulation at the operating wavelength of λ1Tran. It should be noted that unlike
the reflection mode, the presence or absence of the bias lines and choice of their
dimensions had negligible impacts on the transmission response of the unit cell. That
is due to the fact that the bias lines will be effectively transparent in the case of a ypolarized beam which is the illumination used in the transmission mode.

Figure 6.3: Dual-mode Si-based metasurface operating in the transmission mode. (a)
Amplitude and (b) phase of the transmission through the Si-based dielectric metasurface.
Spatial distribution of the electromagnetic fields in the z-y plane at the resonant wavelengths of
(c) λ1Tran ≈ 1563 nm and (d) λ2Tran ≈ 1641 nm. The color bar shows the normal component of
the magnetic field and the white arrows present the normalized electric displacement currents.
Spectrum of the (e) amplitude and (f) phase of the transmission through the metasurface under
applied bias. The dashed lines denote the points for which the real-part of ITO permittivity,
Re{εITO}, at the ITO-Al2O3 interface is equal to −1, 0, and 1, illustrating the ENZ region of the
ITO accumulation layer. (g) Amplitude and (h) phase of the transmission through the
metasurface as a function of applied voltage at the operating wavelength of λ1Tran ≈ 1563 nm.
The shadowed regions present the dielectric breakdown of the gate dielectric [71].

Figures 6.3e, f show the spectra of the amplitude and phase of the transmission through
the metasurface as functions of wavelength and applied bias voltage. As can be seen,
when increasing the bias voltage, the resonance first faces a redshift followed by a
blueshift. It should be noted that the spectral shift and the phase swing of the
transmission resonance as a function of applied bias voltage exhibits an opposite trend
compared to the reflection resonance. The amplitude and phase of the transmission at

139
the fixed operating wavelength of λ1Tran ≈ 1563 nm as functions of the applied bias are
plotted in Figs. 6.3g, h, respectively.
It can be observed that when changing the applied bias, modulation of the transmission
amplitude, accompanied by a phase modulation coverage of ~270° is realized at the
operating wavelength of λ1Tran ≈ 1563 nm.
As seen in this section, using Si-based active metasurfaces, we can achieve dielectric
metasurface elements operating both in transmission and reflection modes (dual-mode
operation) at two distinct operating wavelengths. By altering the bias applied between
Si nanoantennas and an ITO layer, one can tune the amplitude and phase of the
reflected/transmitted light, and accordingly, engineer the wavefront of the scattered
light at will.

6.3. III-V All-Dielectric Active Metasurfaces
As mentioned earlier, despite the large phase modulation achievable by ITO-based
active metasurfaces, the strong field enhancement in the accumulation layer of ITO
results in a high absorption, and accordingly low reflectance/transmittance. This will
lead to a small efficiency of the ITO-based active metasurface, and hence, limit their
applications. To conquer this issue, alternative active platforms like III-V multiple
quantum well heterostructures can be utilized to obtain electro-optically tunable
metasurfaces.

6.3.1. Characterization of MQW Wafers
In order to design a tunable metasurface with a reconfigurable optical response, we first
need to characterize the available MQW wafers.
The epitaxial III-V heterostructure building block of our MQW metasurface consists of
an n-doped GaAs substrate, a distributed Bragg reflector (DBR), and an undoped MQW
layer. The DBR layer is composed of 20 pairs of alternating layers of n-doped
Al0.9Ga0.1As (76.5 nm-thick) and n-doped GaAs (65 nm-thick) with the n-doped
Al0.9Ga0.1As as the topmost layer. A p-doped GaAs contact layer with a carrier density
of 1019 cm-3 is then grown on top of the MQWs. The thickness of the undoped MQW
layer and the p-doped GaAs contact layer are 1.23 μm and 50 nm, respectively.

140
In order to characterize the DBR layer, we measure the reflectance spectrum of the
planar MQW/DBR/GaAs structure as shown in Fig. 6.4a. As can be seen in Fig. 6.4a,
a reflectance of ~ 100% is obtained at the wavelengths ranging from 915 nm to 990 nm,
indicating that the DBR can act as a high-quality mirror in this wavelength range.
Moreover, a sharp reflectance dip can be observed at the wavelength of ~915 nm which
originates from near-bandgap absorption in the MQW layer.
As a next step, we try to investigate the tunable optical response of the MQW under an
applied bias. When applying a DC electric field across the quantum wells, one can
expect enabling electrical modulation of the MQW complex refractive index [242] due
to the shift of the interband transition energy by the QCSE. For our quantum well
heterostructures, the expected modulation of the real part of the refractive index is on
the order of Δn = 0.01 [255]. In order to be able to experimentally observe this small
variation of the real part of the refractive index, we integrate a Fabry-Pérot (F-P)
resonant cavity around the MQWs. The structure supporting the F-P resonance is
obtained by depositing a 35 nm-thick semitransparent Au film as the top mirror on top
of the MQWs. A 2 nm-thick Ti film is used to improve the adhesion of the Au to the
top p-doped GaAs layer. Figure 6.4b plots the measured reflectance spectrum of the
fabricated DBR/MQW/Au F-P cavity. As can be seen, the structure can exhibit a
narrow resonance at a wavelength of 932.7 nm. This narrow resonance will be later
used to enhance the optical modulation caused by the variation of the complex
refractive index of the MQWs under applied bias.
In order to design the metasurface using the MQW heterostructure, we first need to
identify the tunable optical response of the MQWs at different wavelengths by shifting
the position of the F-P resonance to the desired spectral position. To this end, a number
of planar DBR/MQW/polymethyl methacrylate (PMMA)/Au heterostructures (see the
inset of Fig. 6.4c) that could support high-Q-factor F-P resonances are fabricated. By
changing the thickness of the PMMA layer, and as a result, changing the cavity length,
we could alter the spectral position of the high-Q resonances supported by these planar
heterostructures. This would provide us a database of the change in the real (∆n) and
imaginary (∆k) parts of the refractive index of the MQW at different wavelengths.
At a fixed PMMA thickness, and hence, a fixed unbiased resonant wavelength, the ∆n
and ∆k values are evaluated by examining the shift of the resonant wavelength, and the
change of the full width at half maximum (FWHM) of the resonance, respectively under

141
applied bias. Then, utilizing the same resonant mode (i.e. the first F-P cavity resonant
mode), the ∆n and ∆k values are obtained at different wavelengths by changing the
thickness of the PMMA layer in the DBR/MQW/PMMA/Au heterostructure. Figure
6.4c shows the measured reflectance spectra for different thicknesses of the PMMA
layer.

Figure 6.4: Characterization of MQM wafers. Measured reflectance spectra of (a) a bare
DBR/MQW and (b) a DBR/MQW/Ti/Au F-P cavity, and (c) DBR/MQW/PMMA/Au
heterostructures for different thicknesses of the PMMA layer (t) under no applied bias. The
insets show the schematics of corresponding structures. The shadowed region in (a) indicates
the wavelength range shown in (b). (d) The measured reflectance spectrum of the
DBR/MQW/Ti/Au F-P resonant cavity as a function of applied bias. (e) Measured wavelength
shifts (black dots) and FWHM difference (blue dots) at the first F-P resonant mode of the
DBR/MQW/PMMA/Au planar layers with different PMMA thicknesses under applied bias
[117].

142
As can be seen, increasing the thickness of the PMMA layer would redshift the first FP resonant mode. For PMMA thicknesses greater than 210 nm, the second and third FP resonant modes start showing up.
Once we measured the reflectance spectra of the DBR/MQW/Au and the
DBR/MQW/PMMA/Au planar heterostructures in the absence of the applied bias, we
then measure their reflectance modulations under applied bias. To facilitate bias
application, Ohmic contacts made of Ti (20 nm)/Pt (30 nm)/Au (300 nm) and Ge (43
nm)/Ni (30 nm)/Au (87 nm) are deposited on the topmost p-doped GaAs layer and at
the back of the n-doped GaAs substrate, respectively. Then by applying a bias between
the GaAs substrate (low potential) and the top Ohmic contact (high potential), the
reflectance from our F-P resonant MQW samples is measured.
Figure 6.4d plots the measured reflectance spectrum of the DBR/MQW/Au planar layer
as a function of applied bias. As can be seen, when applying an external bias, a shift of
the resonant wavelength accompanied by a significant reflectance modulation is
observed. It confirms the modulation of both the real and imaginary parts of the MQW
refractive index by the applied bias. It should also be noted that a larger optical
modulation is observed at shorter wavelengths, near the semiconductor band edge.
These measurements also show that the optimal wavelength for large reflectance
modulation is expected to be between 915 nm and 920 nm.
The same measurements are performed to obtain the reflectance spectra of the planar
DBR/MQW/PMMA/Au heterostructures with different PMMA thicknesses, for
different applied biases. These measurements also show stronger amplitude modulation
and larger wavelength shifts at shorter wavelengths. Figure 6.4e presents the measured
bias-induced wavelength shift and variation of the FWHM of the first F-P resonance
supported by the planar DBR/MQW/PMMA/Au heterostructure. Here, the wavelength
shift ∆λ = λ (V = –10 V) – λ (V = 0), and the FWHM difference FWHM (V = –10 V) –
FWHM (V = 0) are defined as an amount of the spectral shift in the resonance position
and the change in FWHM when the bias is changed from 0 V to –10 V.
It can also be seen in Fig. 6.4e that larger FWHM changes are obtained at shorter
wavelengths. These results are consistent with the analysis described in prior work
[255], which reported the III-V compound MQW design used in our work. Based on
the trend shown in Fig. 6.4e, we can conclude that the strongest refractive index

143
modulation can be obtained at wavelengths very close to the absorption band edge of
our quantum wells, which is expected to be at the wavelength of ~915 nm.

6.3.2. Design and Simulation of All-dielectric MQW Metasurface
Once we identified the optimal wavelength range for observing a tunable optical
response of the MQWs, we design an all-dielectric metasurface based on the MQW
samples. The first approach to achieve a tunable MQW metasurface is to effectively
use the relatively modest bias-induced refractive index change of the MQWs to obtain
a significant modulation of the optical response of the metasurface.
As a result, the first step is to design a structure that can support a high-Q resonant
mode near the semiconductor band edge. The fundamental electric or magnetic dipole
modes of typical dielectric resonators do not possess sufficiently high quality factors.
Consequently, our metasurface is designed such that it could support a guided-mode
resonance hybridized with a higher-order Mie resonance, at a wavelength slightly
beyond the MQW band edge absorption. Figure 6.5a shows a schematic demonstration
of the proposed electrically tunable all-dielectric III-V MQW resonator-based
metasurface. The resonator design features a double-slit structure, where the double
slits are partially etched into the MQW layer (see the inset of Fig. 6.5b). Owing to the
inherently large real part of the refractive index (n ≈ 3.62) of our MQWs, we can design
subwavelength resonator elements with a width of 700 nm.
The designed metasurface has a period of p = 910 nm. The widths of the slits are w1 =
110 nm, w2 = 210 nm, w = 180 nm, and there is a g = 100 nm-wide gap between the
slits. The height of the slits and the thickness of the MQW layer are h = 40 nm, and t =
1230 nm, respectively.
Figure 6.5b shows the simulated reflectance spectrum of our metasurface element. As
can be seen, two distinct resonant modes are evident at the wavelengths of 915.9 nm
and 936.3 nm. In order to simulate the optical response of the metasurface, we use the
FDTD method (Lumerical). The calculations of the reflectance spectrum of the MQW
resonators are performed in an array configuration by using PML and periodic
boundary conditions in z- and x-directions, respectively. The resonators are assumed to
be infinite in the y-direction, and the incoming wave is considered to normally impinge
on the metasurface along the z-direction. For the sake of simplicity, the refractive

144
indices of the n-Al0.9Ga0.1As, n-Al0.31Ga0.69As, GaAs0.6P0.4, and InGaP are set as a
constant of 3, 3.39+0.004i, 3.3+0.004i, 3.2+0.004i, respectively.

Figure 6.5: Design and simulation of MQW-based metasurface. (a) Schematic illustration
of the all-dielectric MQW metasurface. The designed metasurface consists of an n-doped GaAs
substrate, a DBR layer, and a 1.23 m-thick MQW layer, followed by a 50 nm-thick p-doped
GaAs layer with a doping level of 1019 cm-3 grown on top of the MQWs as the top contact. (b)
The simulated reflectance spectrum of the designed MQW-based metasurface under xpolarized normal-incidence illumination. A schematic of the unit cell of the proposed
metasurface is presented in the inset. The structural parameters of the unit element are designed
to be p = 910 nm, w1 = 110 nm, w2 = 210 nm, w = 180 nm, g = 100 nm, h = 40 nm, and t = 1230
nm. Spatial distributions of (c) the amplitude of the z-component of the electric field and (d)
intensity of the magnetic field at the wavelengths corresponding to the resonant dips shown in
(b) [117].

In order to study the nature of these two resonances, we investigate the amplitude of
the z-component of the electric field and the magnetic field intensity at these resonant
wavelengths, as shown in Figs. 6.5c, and d, respectively. It can be observed in the
spatial field distributions that the metasurface element supports a high-order Mie
resonance at the wavelength of 915.9 nm (denoted by the pink circle). The high-order
Mie resonant mode is identified to be dominated by the magnetic octupolar mode based
on the multipole decomposition analysis [256]–[258].
Further investigation of the electromagnetic field profiles, demonstrates that the
designed metasurface can support a GM resonance propagating along the x-direction at

145
the wavelength of 915.9 nm that arise mainly from the partially etched double-slit
structures. This leads to an electric field leaking into the air gaps separating the
metasurface elements. As a result, the resonant mode observed at the wavelength of
915.9 nm is interpreted to be a coupling of a Mie resonance and a GM resonance,
referred to here as a hybrid Mie-GM resonance. It is noteworthy that when extending
the simulation range to the shorter wavelengths, a mode splitting is detected due to the
coupling of two resonant modes.
Analyzing the spatial profiles of the electromagnetic field at the wavelength of 936.3
nm (denoted by golden star-shaped polygon), we can describe the resonance appearing
at this wavelength to be an F-P resonance propagating along the z-direction of the 1.23μm tall resonators coupled to a GM resonance propagating along the x-direction of the
700 nm-wide resonators.

6.3.3. Fabrication and Measurement of All-dielectric MQW Metasurface
After finding the optimal design, we fabricate the MQW-based metasurface to
experimentally investigate its optical response. To fabricate the metasurface samples,
first, a bottom Ohmic contact consisting of Ge/Ni/Au with the thicknesses of 43 nm/30
nm/87 nm is deposited on the n-doped GaAs substrate of the MQW wafer by using
electron-beam evaporation. Next, a 950 PMMA A9 electron-beam resist with the
thickness of 1.5 μm is spin-coated on the front side of the prepared MQW wafer at the
speed of 4000 rpm for 60 s. The spin-coated sample is then baked on a hot plate at a
temperature of 180 °C for 3 minutes. Subsequently, the top Ohmic contacts as well as
some alignment markers are patterned on the EBR using an electron-beam direct-write
lithography system [VISTEC electron beam pattern generator (EBPG) 5000+] at an
acceleration voltage of 100 keV with a current of 5 nA. Following the exposure, the
sample is developed, and the metals are deposited using electron-beam evaporation.
After the lift-off processes, ZEP 520A EBR is spin-coated at the speed of 4000 rpm for
60 s, and the sample is then baked on a hot plate for 3 minutes at 180 °C. Afterward,
the double-slit structures are patterned by the mentioned EBPG system with a current
of 0.3 nA.
After baking the sample for 2 minutes at 110 °C and then developing it at about 10°C
for 90 s, the patterned ZEP 520A EBR is used as a mask for the dry etch process

146
employed for the fabrication of double slits. In order to etch the slits, a III-V compound
semiconductor etcher (ICP-RIE, Oxford Instruments System) is utilized with gas flow
rates of Cl2: Ar = 5 sccm: 30 sccm under 5 mTorr chamber pressure for 80 seconds.
After removing the ZEP 520A EBR using remover PG, the double slits are obtained.
Finally, a third EBPG process patterns the resonators, followed by developing the
exposed EBR. Then, the same dry etching technique is used with a chamber pressure
of 3 mTorr for 8 minutes and the MQW-based metasurface sample is obtained after
removal of the ZEP 520A EBR using remover PG.
Once we fabricated our sample, we use the optical setup presented in Fig. 6.6 in order
to optically characterize the reflectance of the MQW metasurface. In this setup, a
coherent NIR laser beam (Toptica Photonics CTL 950) is utilized as a light source. The
laser beam is focused on the sample using a long working distance objective with 10×
magnification and 0.28 numerical aperture. Moreover, an uncollimated white light
source from a halogen lamp is used to visualize the sample surface. Then, the reflected
beam from the metasurface is captured by a power meter as a detector (Thorlabs
PM100D).

Figure 6.6: Optical setup used for the measurement of the reflection spectrum of the
MQW metasurface. Schematic of the custom-built setup. M: mirror, ND: neutral density filter
(Thorlabs NDC-50C-4M), I: iris, L: lens, P: linear polarizer (Thorlabs LPNIR100-MP), BS:
beam splitter (Thorlabs CCM1-BS014), O: objective (Mitutoyo 10× magnification with 0.28
numerical aperture), PM: power meter [117].

Figure 6.7a shows an SEM image of the fabricated MQW metasurface sample. The
measured reflectance spectra of the metasurface under different applied biases with the
incoming light being polarized perpendicular to the MQW stripes are shown in Fig.
6.7b. As can be seen, in the absence of the applied bias, the two resonant dips consistent
with the simulation results (Fig. 6.5b) are observed. When changing the applied bias
from 0 to –10 V, a significant red-shift accompanied by an intensity decrement of the

147
shorter-wavelength resonance, corresponding to the hybrid Mie-GM resonance, is
achieved. The red-shift of the resonance is obtained as a result of an increase of the real
part of the refractive index, enabling the voltage-induced phase modulation provided
by the metasurface. The alteration of the reflectance intensities at the resonant
wavelengths indicates the change in the imaginary part of the refractive index, enabling
amplitude modulation provided by the tunable MQW-based metasurface.

Figure 6.7: Measured tunable optical response of the MQW metasurface. (a) SEM image
of MQW-based hybrid Mie-GM resonators with double slits. (b) Measured reflectance spectra
of the metasurface for different applied bias voltages. (c) The measured relative reflectance
spectrum of the hybrid Mie-GM resonant metasurface as a function of applied bias. (d)
Measured phase modulation at the wavelengths of 917 nm (red dots) and 924 nm (blue dots).
Each data point corresponds to an average phase shift measured at four different positions on
the sample while each error bar indicates the standard deviation of the obtained four data points
[117].

At the wavelength of 938 nm, a negligible shift of the resonance wavelength and a small
decrease of the on-resonance reflectance are observed. Due to the small refractive index
change and the absence of high-Q resonances at the wavelengths longer than 940 nm,
no significant reflectance modulation is achieved at these wavelengths. As a result, we

148
will focus on the tunable resonance at shorter wavelengths which corresponded to the
hybrid Mie-GM resonance.
To gain further insight, Fig. 6.7c presents the relative reflectance spectra at different
applied biases obtained via
Δ𝑅
𝑅0

𝑅(𝑉)−𝑅(𝑉=0)
𝑅(𝑉=0)

(6.2)

As can be seen in Fig. 6.7c, decreasing the applied bias from 0 to –10 V results in a
strong relative reflectance modulation. In particular, a relative reflectance modulation
as high as 270% can be obtained at the wavelength of 917 nm. Increasing the
wavelength would lead to a decrease in the relative reflectance modulation (a relative
reflectance modulation of −45% is observed at the wavelength of 925 nm). Thus, the
proposed III-V MQW resonator-based metasurface seems to be a promising candidate
for tunable amplitude modulation with the operating wavelength around 917 nm.
Once we confirmed that our tunable MQW metasurface could provide large amplitude
modulation, we experimentally evaluate the phase shift of the reflected beam under
applied bias at wavelengths of 917 nm and 924 nm using a Michelson interferometer
system [34], [35], [70]. To measure the phase shift, an incident laser spot is illuminated
on the edge of the resonator-based metasurface. As a result, the reflected beam is partly
coming from the metasurface, and partly from the unpatterned MQW heterostructure.
The interference fringes captured by the camera are then processed and fitted. By
considering the unpatterned MQW heterostructure as a built-in phase reference, the
phase shift is determined by calculating the relative displacement of interference fringes
between the hybrid Mie-GM resonator region and the unpatterned region. The
measured phase shifts as functions of applied bias at wavelengths of 917 nm (red dots)
and 924 nm (blue dots) are plotted in Fig. 6.7d.
At the wavelength of 917 nm, decreasing the applied bias from 0 V to–7 V results in a
continuous increase in the phase shift by 70°. Further decreasing the applied bias from
–7 V to –10 V leads to a decrease of the phase shift to 50°. Moving the operating
wavelengths away from the hybrid Mie-GM resonance results in weaker phase
modulations, with the obtained phase shift being only 12° at the wavelength of 924 nm.
It should be noted that the large (small) phase shift, i.e., large (small) change of the real
part of the refractive index accompanied by the significant (negligible) reflectance
modulation, i.e., large (small) modulation of the imaginary part of the refractive index

149
obtained at the wavelength of 917 nm (924 nm), is consistent with the Kramers–Kronig
relation.

6.3.4. Demonstration of Electrical Beam Switching and Beam Steering with the
All-dielectric MQW Metasurface
As mentioned in the previous section, the proposed electro-optically tunable MQW
metasurface can achieve an extremely strong relative reflectance modulation of ~270%
accompanied by a phase shift of ~70°. As a next step, we employ the metasurface as an
amplitude modulator to obtain an electrically switchable grating that could demonstrate
dynamic beam diffraction, detected as a far-field radiation pattern. This requires active
control of reflectance of the independent groups of metasurface elements, which is
achieved via patterning the edges of the metasurface to selectively apply a bias to
independent element groups. The switchable diffraction grating is created by
fabricating a metasurface with similar structural dimensions as mentioned before.
However, the resonant stripes are electrically connected in parallel in groups of three,
with the adjacent group of three resonant stripes being isolated (see Figs. 6.8a, b).
In the absence of an applied bias, due to the subwavelength period of the metasurface
p = Λ = 910 nm, only the zeroth-order diffracted beam is observed as a single output
beam in the Fourier plane. When applying a negative bias to the groups of connected
MQW resonators, the effective period of the metasurface array is electronically
increased to 6 × p = Λ’ = 5460 nm. As a result of the effective period being greater than
the wavelength of light, first-order diffracted beams appear at an angle defined by the
grating equation
𝑚𝜆

𝜃 = sin−1 ( 𝑝 )

(6.3)

where pg is the period of the grating and m is the diffraction order.
Figure 6.8c presents the schematic of the optical setup used to measure the far-field
radiation pattern of the metasurface in the Fourier space. The metasurface sample is
illuminated by a coherent NIR laser beam (Toptica Photonics CTL 950) at a wavelength
of 917 nm, corresponding to the hybrid Mie-GM resonant mode supported by the unit
elements of the metasurface. The laser beam is focused on the sample surface by a longworking-distance objective with 10× magnification and 0.28 numerical aperture. An

150
uncollimated white-light source from a halogen lamp is used to visualize the sample
surface. The radiation pattern of the reflected beam is captured directly by a CCD
camera positioned in the Fourier plane. Experimentally measured diffraction patterns
for different applied bias voltages are shown in Figs. 6.8d, e.

Figure 6.8: Demonstration of switchable diffraction grating using MQW metasurface. (a)
Schematic of the dynamic diffraction grating which was realized by electronically changing the
effective grating period. (b) SEM image of the fabricated sample used for demonstration of the
dynamic beam switching. (c) Schematic of the optical setup used for far-field radiation pattern
measurements. M: mirror, ND: neutral density filter, I: iris, BS: beam splitter, O: objective, L:
lens. (d) Experimentally measured far-field radiation pattern at a wavelength of 917 nm as a
function of diffraction angle and applied voltage. The intensities were normalized to the light
intensity at 0°. (e) Intensity of the scattered light as a function of the diffraction angle at an
applied bias of 0 V (top panel) and –10 V (bottom panel) [117].

As can be seen, the first-order diffracted beam is only observed when there is a
significantly large reflectance contrast between the resonator groups. For the applied
bias voltages between –3 V and 0 (–3 V  V  0), specular reflection from the
metasurface is observed. Further decreasing the applied bias (V ≤ –3 V), the first-order
diffracted beam appears at an angle of ~9.66° in the Fourier plane. It should be noted
that at an incident wavelength of 924 nm, due to the absence of significant reflectance
and phase modulations, no first-order diffracted beam could be observed.
In addition to switchable beam diffraction, we also experimentally demonstrated beam
steering using the amplitude modulation provided by the tunable MQW metasurface.

151
In order to change the spatial position of the first diffraction order in the far-field
radiation pattern, the amplitude modulation imposed by individual metasurface
elements needs to be independently controlled. As a result, we designed and fabricated
another metasurface in which each unit element is electrically isolated by fully etching
the air gap between the quantum well slabs (see Figs. 6.9a-d). To realize dynamic beam
steering, we design two PCBs to individually apply bias to each metasurface element
across the antenna array.

Figure 6.9: Fabricated MQW-based all-dielectric metasurface for demonstration of
dynamic beam steering. (a) Photographic image of the fabricated gate-tunable metasurface
for the demonstration of dynamic beam steering. The metasurface sample is mounted on a PCB,
and 63 metasurface elements and the ground bottom contact are wire-bonded to 64 contact pads
on the PCB. (b-d) SEM images of the gate-tunable metasurface [117].

The sample is mounted on the first PCB, and 63 individual metasurface elements as
well as the bottom contact are wire-bonded to 64 conducting pads on the PCB (see
Appendix E.1). Each conducting pad of the first PCB is then connected to an external
pin on the second PCB. This PCB is capable of providing 64 independent voltages that
could be individually controlled through the reference voltages derived by an external
power supply (Keithley 2400). The second PCB has different configurations of voltage
paths. By switching between different configurations, we could electrically change the
grating period of the metasurface.

152
It should be emphasized that the high refractive index and large thickness of the MQW
slabs make the resonant mode sensitive to minor variations of metasurface structural
parameters. This would lead to different electromagnetic field profiles between the first
and the second hybrid Mie-GM metasurfaces. As a result, to examine the functionality
of the fabricated metasurface as a beam steering device, we first need to study its optical
response.
Figure 6.10 shows the simulated reflectance spectrum (Fig. 6.10a) and the distribution
of the electromagnetic fields (Figs. 6.10b, c) when the incoming light is polarized
perpendicularly to the MQW stripes. As can be seen, our beam steering metasurface
supports a hybrid Mie-GM resonant mode at a wavelength near the band edge
absorption of the MQWs, which leads to the mentioned reflectance and phase
modulation. Here the structural parameters of the metasurface are chosen to be w1 =
100 nm, w2 = 120 nm, w = 110 nm, gt = 110 nm, gu = 90 nm, t = 1230 nm, h = 80 nm,
and p = 780 nm. Then, we measure the spectral response as well as the active optical
modulation of the metasurface by applying an identical electrical bias to all metasurface
elements.

Figure 6.10: Simulated optical response of the MQW-based metasurface fabricated for
demonstration of dynamic beam steering. (a) Simulated reflectance spectrum of the second
hybrid Mie-GM resonator array. Simulated (b) amplitude of x-component of the electric field,
and (c) magnetic field intensity at the wavelength of 917 nm.

Figure 6.11 shows the measured reflectance spectrum (Fig. 6.11a) and the reflectance
change spectrum (Fig. 6.11b) for different applied biases with the incoming light being
polarized perpendicular to the MQW stripes. The phase shift provided by the
metasurface at two wavelengths of 917 nm and 924 nm as a function of applied bias is
plotted in Fig. 6.11c. As can be seen, under an applied bias, the fabricated metasurface
sample yields about 80% relative reflectance modulation with a phase shift of ~42°.

153

Figure 6.11: Measured amplitude and phase modulation provided by the MQW-based
metasurface fabricated for demonstration of dynamic beam steering. Measured (a)
reflectance and (b) reflectance change spectrum of the fabricated metasurface array for different
applied bias voltages. Here, the reflectance of resonator under a –10 V bias is utilized as the
reference. (c) The measured phase modulation at the wavelengths of 917 nm (red dots) and 924
nm (blue dots). Each data point corresponds to an average phase shift measured at four different
positions on the sample while each error bar indicates the standard deviation of the obtained
four data points.

After confirming the amplitude modulation achievable by the metasurface, we steer the
reflected beam by individually addressing each metasurface element. Figure 6.12
shows the measured intensity of the far-field radiation pattern in which the first-order
diffraction peaks are indicated by black arrows.

Figure 6.12: Demonstration of dynamic beam steering using all-dielectric MQW
metasurface. The measured intensity of dynamic beam steering at the wavelength of 917 nm
by electrically changing the period of the metasurface. Black arrows indicate the position of the
first diffraction order in each case. The right column illustrates how the spatial arrangement of
electrical bias changes the periodicity of the metasurface. The scale bars are 3 μm [117].

154
As can be seen, by changing the effective period of the metasurface, we can realize
dynamic beam steering with the first-order diffraction angle becoming smaller as the
metasurface period is increased via electrical control. It should be noted that the
sidelobes around the zeroth-order diffraction beam appear because of the finite aperture
effect. Figure 6.13 shows the simulated far-field radiation pattern of the metasurface
with 30 (Fig. 6.13a) and 120 (Fig. 6.13b) unit elements at the wavelength of 917 nm.
In order to take the finite aperture effect into account, top hat is utilized as the
illumination condition when processing the far-field projection in Lumerical FDTD
simulation. As can be seen, when increasing the number of metasurface elements, both
the zeroth- and first-order diffracted beams become narrower and they would be more
dominant over the undesired sidelobes.

Figure 6.13: Effect of aperture size on the performance of the MQW beam steering
metasurface. Simulated results of a far-field beam pattern for different effective periods of
metasurface Λ. The total number of metasurface unit element was assumed to be (a) 30 and (b)
120. Black arrows indicate the position of the first diffraction order [117].

6.4. Conclusions and Outlook
In this chapter, we proposed two platforms for achieving dielectric tunable
metasurfaces. First, a theoretical study of an active all-dielectric Si-based metasurface
with independent operations in both reflection and transmission modes (dual-mode
operation) was presented. The unit cell of the proposed design is composed of an Al2O3ITO stand-off layer sandwiched between the Si backplane and Si nanodisks connected

155
by Si nanobars. The desired geometrical resonances are excited via the high-index Si
nanobars for TE polarized reflected beam, and by Si nanodisks for TM polarized
transmitted light. By overlapping the confined resonant optical mode with the ENZ
transition of the ITO active layer considerable phase tunability is obtained both in
reflection and transmission modes.
By utilizing a coupled electrical and optical modeling through linking Lumerical
Device and RCWA solvers, the inhomogeneity and voltage-dependency of the ITO and
Si layers are obtained, leading to a comprehensive study of the optical response of the
metasurface under applied bias. When applying an external voltage, the ITO-integrated
SIS unit cell could separately achieve phase modulations as large as 240° and 270° in
reflection mode at the wavelength of 1696 nm and transmission mode at the wavelength
of 1563 nm, respectively. By individually controlling the metasurface elements, one
will achieve large arrays with electrically phase-tunable elements, capable of dynamic
beam manipulation and wavefront engineering in reflection and transmission modes.
Next, we proposed the design and experimental demonstration of an all-dielectric
electro-optically tunable metasurface platform based on monolithically grown III-V
compound semiconductor MQW heterostructures. The metasurface consists of an array
of MQW elements supporting hybrid Mie-GM resonances. When applying a DC
electric field across the resonators, the QCSE results in an actively tunable optical
response in the NIR wavelength range. This leads to a 270% relative reflectance
modulation accompanied by a 70° phase shift at the hybrid Mie-GM-resonant
wavelength of 917 nm. Employing the metasurface as an amplitude modulator, we were
able to demonstrate a dynamic diffraction grating by electrically connecting
metasurface elements in groups of three. When applying a bias to every other group of
connected resonators, the effective period of the metasurface is actively changed,
leading to the first-order diffracted beam being switched on and off. As a next step, we
demonstrated a dynamic beam steering by individually controlling the voltage applied
to and hence, the reflectance modulation by each metasurface element.
A low leakage current density in our samples confirms that the tunable optical response
of the metasurface is achieved via an electro-optic effect rather than charge carrier
injection. As a result, the proposed metasurface platform offers a high modulation speed
of MHz frequencies or higher [240], [242], [243]. The tunable optical response
achievable by the presented active metasurface platform and the ability to individually

156
control each metasurface element could be utilized for the realization of dynamicallytunable, ultrathin optical components, such as tunable metalenses with reconfigurable
focal lengths and numerical apertures, on-chip beam steering devices, active polarizers,
and flat spatial-light modulators.
These monolithically grown MQW-based active metasurfaces show the potential to be
integrated with existing light-emitting devices, such as vertical-cavity surface-emitting
lasers (VCSELs). This integrated device can serve as a base for future on-chip light
detection and ranging systems. It should be noted that the performance of the proposed
all-dielectric metasurface can be further improved by utilizing alternative quantum well
systems with larger modulation of the real part of the refractive index and lower optical
loss [259], [260].

157
Chapter 7

CONCLUSION AND OUTLOOK
Controlling the flow of light at the nanoscale has always been an awe-inspiring
challenge of optics and photonics. Optical metasurfaces that can control fundamental
properties of light with high spatial resolution have been the subject of intensive
research over the last few years. Owing to their capability to mimic the functionality of
conventional bulky optical elements with much higher resolutions, the study of
metasurfaces has become a thriving field of research. Not only can metasurfaces
outperform their bulky optical element counterparts, but they can also provide unique
functionalities not achievable via conventional diffractive optics.
Unlike passive metasurfaces whose functions are fixed at the time of fabrication, active
metasurfaces comprising arrays of reconfigurable nanoantennas that can provide
dynamic control over fundamental attributes of light have garnered widespread interest.
Notably, active metasurfaces consisting of arrays of programmable nanoantennas offer
a powerful approach to reconfigurable array architectures for optical wavefront shaping.
Among several active platforms provided for obtaining tunable metasurfaces, electrical
modulation was shown to be a robust, reproducible, and high-speed scheme for tuning
of the function of metasurfaces.
In this dissertation, we described electro-optically tunable metasurface platforms to
control the properties of radiation. First, a novel dual-gated design based on ITOintegrated metasurfaces was proposed that could provide amplitude modulation as well
as a record-high phase shift in the NIR wavelength regime. An ITO-based tunable
metasurface was then employed to demonstrate universal metasurfaces whose function
could be tuned by changing the bias voltages applied to the metasurface elements. By
individually controlling the voltage applied to each metasurface pixel, the phase
imposed by a single metasurface pixel was controlled at will. As a result, engineering
the spatial phase profile of the metasurface led to the realization of an arbitrary
wavefront. As a proof of concept, the metasurface was utilized to demonstrate dynamic
beam steering and tunable focusing with reconfigurable focal length and numerical
aperture in the telecommunication wavelength range.

158
The ITO-based tunable metasurface was further presented to be able to provide
dynamic control of the polarization state. By a careful choice of the bias voltage applied
to the metasurface, linear-to-linear, linear-to-circular, and linear-to-elliptical
polarization conversions were realized.
Using a similar approach, we used an ITO-based plasmonic metasurface to actively
enhance the radiative decay rate of the quantum emitters embedded within the
metasurface. By carefully coupling the quantum emitters to the geometrical resonance
of the plasmonic metasurface, a considerable Purcell enhancement of the spontaneous
emission decay rate of the quantum emitters was obtained. By altering the bias voltage
applied to the metasurface, the Purcell enhancement, the radiative decay rate, and the
quantum efficiency of the quantum emitters could be tuned.
Then, alternative approaches to achieve dielectric tunable metasurfaces were proposed.
Unlike the aforementioned reflect-array plasmonic metasurfaces, a Si-based
metasurface was introduced that could provide amplitude and phase modulation both
in reflection, and transmission modes via integrating ITO within the metasurface.
Furthermore, to conquer the low-efficiency issue of the ITO-based metasurfaces, an
alternative platform based on III-V MQW heterostructures was presented that could
provide high-efficiency amplitude and phase modulation using the QCSE. By
employing the modulation achieved when changing the applied bias, dynamic beam
switching and beam steering were realized using the metasurface.
As can be seen, the ability of active metasurface platforms to provide dynamic and
comprehensive control over the elemental properties of light can realize novel
functionalities in wavefront manipulation.
The planar, subwavelength-thick nature of active metasurfaces, their compatibility with
conventional micro/nano-fabrication techniques, and their dexterity to dynamically
control different degrees of freedom of light make them promising for the realization
of the next generation of compact, high-performance integrated photonic systems.
Integrated photonics is finding use in a wide range of areas including
telecommunications, laser-based radars, data communications, sensing, and the
emerging world of autonomous vehicles.
In such integrated photonic systems, miniature tunable metasurface elements can
achieve the functions provided by a group of multiple bulky conventional elements.

159
The small size of the integrated photonic systems at a cost that is friendly for largescale production and use makes them a paradigm-changing platform for optical
elements and system design for various applications.
However, despite all the advancements made in the field of active metasurfaces during
the past few years, there are still several unsolved challenges that need to be addressed.
One area that requires significant advancements is the improvement of the efficiency
of the plasmonic metasurfaces. One limiting factor is the small reflectance value of the
ITO-based plasmonic metasurfaces. This results in a decreased efficiency of the device.
In order to increase the reflectance from these metasurfaces, one can utilize antennalevel shape or topology optimization algorithms [37-43] to obtain non-trivial antenna
designs that provide larger reflectance levels while maintaining a phase shift that is
large enough for the realization of wavefront shaping devices. Moreover, using
transparent conducting oxides with higher electron mobilities, like CdO is expected to
increase the reflectance values.
Furthermore, the non-idealities of the plasmonic active metasurfaces result in the low
efficiency of such devices. Radio frequency (RF) electronic amplifiers consist of
phased-arrays with an ‘ideal’ antenna response, i.e., with a constant amplitude and a
smooth variation of phase from 0 to 2π. In contrast, the presented active metasurfaces
do not allow direct phase modulation in the time domain. In other words, metasurface
phased-arrays rely on phase modulation via permittivity tuning near optical resonances
in nanoantennas, leading to an inherently non-ideal antenna response. As shown in
Chapter 3, these non-idealities include non-unity reflectance of the metasurface, the
achievable phase shift of smaller than 2π, and co-variation of amplitude and phase shift
as the applied bias changes.
Ideal phased-array performance at optical frequencies can be obtained despite the use
of highly non-ideal nanoantennas as components by applying an inverse design
approach at the array-level that capitalizes on the tunability of nanoantennas in active
metasurfaces to optimize the desired function, thus overcoming the limitations posed
by non-ideal metasurface phase and amplitude tuning. By employing array-level
optimization techniques, we can attain highly non-intuitive array phase and amplitude
profiles that compensate for the device non-idealities, and hence significantly enhance
the metasurface performance.

160
Moreover, one can utilize material-level inverse design protocols [49, 50] in order to
further enhance the efficiency of active metasurface devices. As an outlook, we expect
that by combining array-, antenna-, and material-level inverse design approaches, a new
era for co-design of materials, devices and systems can be accessed for nanophotonics
via obtaining highly efficient multifunctional metasurfaces capable of providing many
functions that enable unprecedented space-time control of the scattered light wavefront.
Another issue worth addressing is the number of available degrees of freedom that can
be controlled using a single metasurface device. Increasing the number of controllable
properties that can be encoded with a metasurface will result in an increase in the
number of functions provided by the mentioned device, and hence, the realization of
multifunctional metasurface devices. Even though metasurfaces with independent
control over phase and polarization [14, 133], and metasurfaces operating at multiple
wavelengths [135, 137, 228] and angles [15, 354] have been proposed, comprehensive
independent control over multiple degrees of freedom is still missing. As an example,
by increasing the control knobs via employing multi-gated configurations, one could
be able to achieve phase modulation at a constant amplitude and vice versa.
Moreover, the metasurfaces presented in this dissertation could provide onedimensional wavefront control. In these structures, electrical addressing of individual
stripe-like metasurface elements (called as metasurface pixels) was done from the edge
of the metasurface chip. In order to obtain a two-dimensional control of individual
metasurface elements rather than rows of metasurface pixels, one needs to align a twodimensional electrical circuitry with the active metasurface. To this end, an electrical
addressing approach, that is typically used in CMOS image sensors or dynamic
random-access memories (DRAMs) can be utilized. In the case of reflective
metasurfaces, electrical interconnects can be built underneath the metasurface to
minimize potential photonic artifacts caused by the interconnect structure. In this
approach, each array element can be connected via two transistors, and the array
element is electrically activated only when both transistors are open. Then, rather than
addressing all array elements at the same time, the array elements can be addressed
sequentially (one element at a time) via column and row driving lines.
Additionally, the fabrication of active subwavelength metasurfaces using the existing
low-cost large-scale foundry technologies is challenging especially in the UV, visible,
and NIR wavelength regimes. As a consequence, the manufacturing processes of active

161
metasurfaces should be modified to be compatible with large-scale fabrication
techniques such as roll-to-roll nano-imprint, soft, and DUV lithography to obtain large,
high-NA devices that have a significant industrial impact.

162

BIBLIOGRAPHY
[1]

[2]

[3]

[4]
[5]
[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]
[16]

[17]

N. Yu et al., “Light Propagation with Phase Discontinuities: Generalized Laws
of Reflection and Refraction,” Science (80-. )., vol. 334, no. 6054, pp. 333–337,
Oct. 2011, doi: 10.1126/science.1210713.
H.-H. Hsiao, C. H. Chu, and D. P. Tsai, “Fundamentals and Applications of
Metasurfaces,” Small Methods, vol. 1, no. 4, p. 1600064, Apr. 2017, doi:
10.1002/smtd.201600064.
A. V Kildishev, A. Boltasseva, and V. M. Shalaev, “Planar photonics with
metasurfaces.,” Science, vol. 339, no. 6125, p. 1232009, Mar. 2013, doi:
10.1126/science.1232009.
N. Yu and F. Capasso, “Flat optics with designer metasurfaces.,” Nat. Mater.,
vol. 13, no. 2, pp. 139–150, Feb. 2014, doi: 10.1038/nmat3839.
X. Chen et al., “Dual-polarity plasmonic metalens for visible light.,” Nat.
Commun., vol. 3, p. 1198, 2012, doi: 10.1038/ncomms2207.
M. Kang, T. Feng, H.-T. Wang, and J. Li, “Wave front engineering from an array
of thin aperture antennas,” Opt. Express, vol. 20, no. 14, pp. 15882–15890, 2012,
doi: 10.1364/OE.20.015882.
F. Aieta et al., “Aberration-Free Ultrathin Flat Lenses and Axicons at Telecom
Wavelengths Based on Plasmonic Metasurfaces,” Nano Lett., vol. 12, no. 9, pp.
4932–4936, Sep. 2012, doi: 10.1021/nl302516v.
X. Ni, S. Ishii, A. V Kildishev, and V. M. Shalaev, “Ultra-thin, planar, Babinetinverted plasmonic metalenses,” Light Sci. Appl., vol. 2, no. 4, pp. e72–e72, 2013,
doi: 10.1038/lsa.2013.28.
D. Lin, P. Fan, E. Hasman, and M. L. Brongersma, “Dielectric gradient
metasurface optical elements,” Science (80-. )., vol. 345, no. 6194, pp. 298 LP –
302, Jul. 2014, doi: 10.1126/science.1253213.
M. Khorasaninejad, W. T. Chen, R. C. Devlin, J. Oh, A. Y. Zhu, and F. Capasso,
“Metalenses at visible wavelengths: Diffraction-limited focusing and
subwavelength resolution imaging,” Science (80-. )., vol. 352, no. 6290, pp. 1190
LP – 1194, Jun. 2016, doi: 10.1126/science.aaf6644.
M. Khorasaninejad et al., “Multispectral Chiral Imaging with a Metalens,” Nano
Lett.,
vol.
16,
no.
7,
pp.
4595–4600,
Jul.
2016,
doi:
10.1021/acs.nanolett.6b01897.
M. Khorasaninejad et al., “Achromatic Metasurface Lens at Telecommunication
Wavelengths,” Nano Lett., vol. 15, no. 8, pp. 5358–5362, Aug. 2015, doi:
10.1021/acs.nanolett.5b01727.
F. Aieta, M. A. Kats, P. Genevet, and F. Capasso, “Multiwavelength achromatic
metasurfaces by dispersive phase compensation,” Science (80-. )., vol. 347, no.
6228, pp. 1342 LP – 1345, Mar. 2015, doi: 10.1126/science.aaa2494.
M. Khorasaninejad et al., “Achromatic Metalens over 60 nm Bandwidth in the
Visible and Metalens with Reverse Chromatic Dispersion,” Nano Lett., vol. 17,
no. 3, pp. 1819–1824, Mar. 2017, doi: 10.1021/acs.nanolett.6b05137.
X. Ni, A. V Kildishev, and V. M. Shalaev, “Metasurface holograms for visible
light,” Nat. Commun., vol. 4, no. 1, p. 2807, 2013, doi: 10.1038/ncomms3807.
L. Huang et al., “Three-dimensional optical holography using a plasmonic
metasurface,” Nat. Commun., vol. 4, no. 1, p. 2808, 2013, doi:
10.1038/ncomms3808.
X. Zhang et al., “Metasurface-based broadband hologram with high tolerance to

163

[18]

[19]
[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

fabrication errors,” Sci. Rep., vol. 6, no. 1, p. 19856, 2016, doi:
10.1038/srep19856.
G. Zheng, H. Mühlenbernd, M. Kenney, G. Li, T. Zentgraf, and S. Zhang,
“Metasurface holograms reaching 80% efficiency.,” Nat. Nanotechnol., vol. 10,
no. 4, pp. 308–312, Apr. 2015, doi: 10.1038/nnano.2015.2.
D. Wen et al., “Helicity multiplexed broadband metasurface holograms,” Nat.
Commun., vol. 6, no. 1, p. 8241, 2015, doi: 10.1038/ncomms9241.
X. Li et al., “Athermally photoreduced graphene oxides for three-dimensional
holographic images,” Nat. Commun., vol. 6, no. 1, p. 6984, 2015, doi:
10.1038/ncomms7984.
M. Khorasaninejad, A. Ambrosio, P. Kanhaiya, and F. Capasso, “Broadband and
chiral binary dielectric meta-holograms,” Sci. Adv., vol. 2, no. 5, p. e1501258,
May 2016, doi: 10.1126/sciadv.1501258.
Y. Yang, W. Wang, P. Moitra, I. I. Kravchenko, D. P. Briggs, and J. Valentine,
“Dielectric Meta-Reflectarray for Broadband Linear Polarization Conversion
and Optical Vortex Generation,” Nano Lett., vol. 14, no. 3, pp. 1394–1399, Mar.
2014, doi: 10.1021/nl4044482.
E. Karimi, S. A. Schulz, I. De Leon, H. Qassim, J. Upham, and R. W. Boyd,
“Generating optical orbital angular momentum at visible wavelengths using a
plasmonic metasurface,” Light Sci. Appl., vol. 3, no. 5, pp. e167–e167, 2014, doi:
10.1038/lsa.2014.48.
M. I. Shalaev, J. Sun, A. Tsukernik, A. Pandey, K. Nikolskiy, and N. M.
Litchinitser, “High-Efficiency All-Dielectric Metasurfaces for Ultracompact
Beam Manipulation in Transmission Mode,” Nano Lett., vol. 15, no. 9, pp. 6261–
6266, Sep. 2015, doi: 10.1021/acs.nanolett.5b02926.
F. Bouchard, I. De Leon, S. A. Schulz, J. Upham, E. Karimi, and R. W. Boyd,
“Optical spin-to-orbital angular momentum conversion in ultra-thin
metasurfaces with arbitrary topological charges,” Appl. Phys. Lett., vol. 105, no.
10, p. 101905, Sep. 2014, doi: 10.1063/1.4895620.
E. Maguid, I. Yulevich, D. Veksler, V. Kleiner, M. L. Brongersma, and E.
Hasman, “Photonic spin-controlled multifunctional shared-aperture antenna
array,” Science (80-. )., vol. 352, no. 6290, pp. 1202 LP – 1206, Jun. 2016, doi:
10.1126/science.aaf3417.
K. E. Chong et al., “Polarization-Independent Silicon Metadevices for Efficient
Optical Wavefront Control,” Nano Lett., vol. 15, no. 8, pp. 5369–5374, Aug.
2015, doi: 10.1021/acs.nanolett.5b01752.
G. Li et al., “Spin-Enabled Plasmonic Metasurfaces for Manipulating Orbital
Angular Momentum of Light,” Nano Lett., vol. 13, no. 9, pp. 4148–4151, Sep.
2013, doi: 10.1021/nl401734r.
K. Sarabandi and N. Behdad, “A Frequency Selective Surface With Miniaturized
Elements,” IEEE Trans. Antennas Propag., vol. 55, no. 5, pp. 1239–1245, 2007,
doi: 10.1109/TAP.2007.895567.
F. Cheng, J. Gao, T. S. Luk, and X. Yang, “Structural color printing based on
plasmonic metasurfaces of perfect light absorption.,” Sci. Rep., vol. 5, p. 11045,
Jun. 2015, doi: 10.1038/srep11045.
W. Zhu, F. Xiao, M. Kang, and M. Premaratne, “Coherent perfect absorption in
an all-dielectric metasurface,” Appl. Phys. Lett., vol. 108, no. 12, p. 121901, Mar.
2016, doi: 10.1063/1.4944635.
Y. Li and B. M. Assouar, “Acoustic metasurface-based perfect absorber with
deep subwavelength thickness,” Appl. Phys. Lett., vol. 108, no. 6, p. 63502, Feb.

164

[33]
[34]

[35]
[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

2016, doi: 10.1063/1.4941338.
L. Wang et al., “Grayscale transparent metasurface holograms,” Optica, vol. 3,
no. 12, pp. 1504–1505, 2016, doi: 10.1364/OPTICA.3.001504.
G. Kafaie Shirmanesh, R. Sokhoyan, R. A. Pala, and H. A. Atwater, “Dual-Gated
Active Metasurface at 1550 nm with Wide (>300°) Phase Tunability,” Nano Lett.,
vol. 18, no. 5, pp. 2957–2963, May 2018, doi: 10.1021/acs.nanolett.8b00351.
Y.-W. Huang et al., “Gate-Tunable Conducting Oxide Metasurfaces,” Nano Lett.,
vol. 16, no. 9, pp. 5319–5325, Sep. 2016, doi: 10.1021/acs.nanolett.6b00555.
M. C. Sherrott et al., “Experimental Demonstration of >230° Phase Modulation
in Gate-Tunable Graphene–Gold Reconfigurable Mid-Infrared Metasurfaces,”
Nano Lett., vol. 17, no. 5, pp. 3027–3034, May 2017, doi:
10.1021/acs.nanolett.7b00359.
J. Valente, J.-Y. Ou, E. Plum, I. J. Youngs, and N. I. Zheludev, “A magnetoelectro-optical effect in a plasmonic nanowire material,” Nat. Commun., vol. 6,
no. 1, p. 7021, 2015, doi: 10.1038/ncomms8021.
H.-S. Ee and R. Agarwal, “Tunable Metasurface and Flat Optical Zoom Lens on
a Stretchable Substrate,” Nano Lett., vol. 16, no. 4, pp. 2818–2823, Apr. 2016,
doi: 10.1021/acs.nanolett.6b00618.
X. Yang et al., “Reconfigurable all-dielectric metasurface based on tunable
chemical systems in aqueous solution.,” Sci. Rep., vol. 7, no. 1, p. 3190, Jun.
2017, doi: 10.1038/s41598-017-03439-9.
J. Li, S. Kamin, G. Zheng, F. Neubrech, S. Zhang, and N. Liu, “Addressable
metasurfaces for dynamic holography and optical information encryption,” Sci.
Adv., vol. 4, no. 6, p. eaar6768, Jun. 2018, doi: 10.1126/sciadv.aar6768.
L. Deng et al., “Direct Optical Tuning of the Terahertz Plasmonic Response of
InSb Subwavelength Gratings,” Adv. Opt. Mater., vol. 1, no. 2, pp. 128–132, Feb.
2013, doi: 10.1002/adom.201200032.
A. M. Shaltout, V. M. Shalaev, and M. L. Brongersma, “Spatiotemporal light
control with active metasurfaces.,” Science, vol. 364, no. 6441, May 2019, doi:
10.1126/science.aat3100.
N. Dabidian et al., “Electrical Switching of Infrared Light Using Graphene
Integration with Plasmonic Fano Resonant Metasurfaces,” ACS Photonics, vol.
2, no. 2, pp. 216–227, Feb. 2015, doi: 10.1021/ph5003279.
Y. Yao et al., “Broad Electrical Tuning of Graphene-Loaded Plasmonic
Antennas,” Nano Lett., vol. 13, no. 3, pp. 1257–1264, Mar. 2013, doi:
10.1021/nl3047943.
Y. C. Jun et al., “Epsilon-near-zero strong coupling in metamaterialsemiconductor hybrid structures.,” Nano Lett., vol. 13, no. 11, pp. 5391–5396,
2013, doi: 10.1021/nl402939t.
J. Park, J.-H. Kang, X. Liu, and M. L. Brongersma, “Electrically Tunable
Epsilon-Near-Zero (ENZ) Metafilm Absorbers,” Sci. Rep., vol. 5, no. 1, p. 15754,
2015, doi: 10.1038/srep15754.
T. Lewi, H. A. Evans, N. A. Butakov, and J. A. Schuller, “Ultrawide Thermooptic Tuning of PbTe Meta-Atoms,” Nano Lett., vol. 17, no. 6, pp. 3940–3945,
Jun. 2017, doi: 10.1021/acs.nanolett.7b01529.
P. P. Iyer, M. Pendharkar, C. J. Palmstrøm, and J. A. Schuller, “Ultrawide
thermal free-carrier tuning of dielectric antennas coupled to epsilon-near-zero
substrates,” Nat. Commun., vol. 8, no. 1, p. 472, 2017, doi: 10.1038/s41467-01700615-3.
P. Cencillo-Abad, J.-Y. Ou, E. Plum, J. Valente, and N. I. Zheludev, “Random

165

[50]

[51]

[52]

[53]
[54]

[55]

[56]

[57]

[58]

[59]

[60]

[61]

[62]

[63]

[64]

[65]

access actuation of nanowire grid metamaterial,” Nanotechnology, vol. 27, no.
48, p. 485206, 2016, doi: 10.1088/0957-4484/27/48/485206.
J.-Y. Ou, E. Plum, J. Zhang, and N. I. Zheludev, “An electromechanically
reconfigurable plasmonic metamaterial operating in the near-infrared.,” Nat.
Nanotechnol., vol. 8, no. 4, pp. 252–255, Apr. 2013, doi: 10.1038/nnano.2013.25.
Z. Zhu, P. G. Evans, R. F. J. Haglund, and J. G. Valentine, “Dynamically
Reconfigurable Metadevice Employing Nanostructured Phase-Change
Materials.,” Nano Lett., vol. 17, no. 8, pp. 4881–4885, Aug. 2017, doi:
10.1021/acs.nanolett.7b01767.
M. Decker et al., “Electro-optical switching by liquid-crystal controlled
metasurfaces,” Opt. Express, vol. 21, no. 7, pp. 8879–8885, 2013, doi:
10.1364/OE.21.008879.
J. Sautter et al., “Active Tuning of All-Dielectric Metasurfaces,” ACS Nano, vol.
9, no. 4, pp. 4308–4315, Apr. 2015, doi: 10.1021/acsnano.5b00723.
A. Olivieri, C. Chen, S. Hassan, E. Lisicka-Skrzek, R. N. Tait, and P. Berini,
“Plasmonic
Nanostructured
Metal–Oxide–Semiconductor
Reflection
Modulators,” Nano Lett., vol. 15, no. 4, pp. 2304–2311, Apr. 2015, doi:
10.1021/nl504389f.
C. R. de Galarreta et al., “Nonvolatile Reconfigurable Phase-Change
Metadevices for Beam Steering in the Near Infrared,” Adv. Funct. Mater., vol.
28, no. 10, p. 1704993, Mar. 2018, doi: 10.1002/adfm.201704993.
K. Thyagarajan, R. Sokhoyan, L. Zornberg, and H. A. Atwater, “Millivolt
Modulation of Plasmonic Metasurface Optical Response via Ionic Conductance,”
Adv. Mater., vol. 29, no. 31, p. 1701044, Aug. 2017, doi:
10.1002/adma.201701044.
Q. Wang et al., “Optically reconfigurable metasurfaces and photonic devices
based on phase change materials,” Nat. Photonics, vol. 10, no. 1, pp. 60–65,
2016, doi: 10.1038/nphoton.2015.247.
P. C. Wu, N. Papasimakis, and D. P. Tsai, “Self-Affine Graphene Metasurfaces
for Tunable Broadband Absorption,” Phys. Rev. Appl., vol. 6, no. 4, p. 44019,
Oct. 2016, doi: 10.1103/PhysRevApplied.6.044019.
C. H. Chu et al., “Active dielectric metasurface based on phase-change medium
(Laser Photonics Rev. 10(6)/2016),” Laser Photon. Rev., vol. 10, no. 6, p. 1063,
Nov. 2016, doi: 10.1002/lpor.201670068.
J. Sun, E. Timurdogan, A. Yaacobi, E. S. Hosseini, and M. R. Watts, “Largescale nanophotonic phased array.,” Nature, vol. 493, no. 7431, pp. 195–199, Jan.
2013, doi: 10.1038/nature11727.
K. Dong et al., “A Lithography-Free and Field-Programmable Photonic
Metacanvas,” Adv. Mater., vol. 30, no. 5, p. 1703878, Feb. 2018, doi:
10.1002/adma.201703878.
A. Howes, W. Wang, I. Kravchenko, and J. Valentine, “Dynamic transmission
control based on all-dielectric Huygens metasurfaces,” Optica, vol. 5, no. 7, pp.
787–792, 2018, doi: 10.1364/OPTICA.5.000787.
S. He, H. Yang, Y. Jiang, W. Deng, and W. Zhu, “Recent Advances in MEMS
Metasurfaces and Their Applications on Tunable Lens,” Micromachines, vol. 10,
no. 8, p. 505, Jul. 2019, doi: 10.3390/mi10080505.
A. She, S. Zhang, S. Shian, D. R. Clarke, and F. Capasso, “Adaptive metalenses
with simultaneous electrical control of focal length, astigmatism, and shift,” Sci.
Adv., vol. 4, no. 2, p. eaap9957, Feb. 2018, doi: 10.1126/sciadv.aap9957.
S. C. Malek, H.-S. Ee, and R. Agarwal, “Strain Multiplexed Metasurface

166

[66]

[67]

[68]

[69]

[70]

[71]

[72]

[73]

[74]

[75]

[76]

[77]

[78]

[79]

[80]

Holograms on a Stretchable Substrate,” Nano Lett., vol. 17, no. 6, pp. 3641–3645,
Jun. 2017, doi: 10.1021/acs.nanolett.7b00807.
M. L. Tseng, J. Yang, M. Semmlinger, C. Zhang, P. Nordlander, and N. J. Halas,
“Two-Dimensional Active Tuning of an Aluminum Plasmonic Array for FullSpectrum Response,” Nano Lett., vol. 17, no. 10, pp. 6034–6039, Oct. 2017, doi:
10.1021/acs.nanolett.7b02350.
J. Park, J.-H. Kang, S. J. Kim, X. Liu, and M. L. Brongersma, “Dynamic
Reflection Phase and Polarization Control in Metasurfaces,” Nano Lett., vol. 17,
no. 1, pp. 407–413, Jan. 2017, doi: 10.1021/acs.nanolett.6b04378.
P. P. Iyer, M. Pendharkar, and J. A. Schuller, “Electrically Reconfigurable
Metasurfaces Using Heterojunction Resonators,” Adv. Opt. Mater., vol. 4, no.
10, pp. 1582–1588, Oct. 2016, doi: 10.1002/adom.201600297.
A. Komar et al., “Electrically tunable all-dielectric optical metasurfaces based
on liquid crystals,” Appl. Phys. Lett., vol. 110, no. 7, p. 71109, Feb. 2017, doi:
10.1063/1.4976504.
Y. Kim et al., “Phase Modulation with Electrically Tunable Vanadium Dioxide
Phase-Change Metasurfaces.,” Nano Lett., vol. 19, no. 6, pp. 3961–3968, Jun.
2019, doi: 10.1021/acs.nanolett.9b01246.
A. Forouzmand, M. M. Salary, G. Kafaie Shirmanesh, R. Sokhoyan, H. A.
Atwater, and H. Mosallaei, “Tunable All-Dielectric Metasurface for Phase
Modulation of the Reflected and Transmitted Light via Permittivity Tuning of
Indium Tin Oxide,” Nanophotonics, vol. 8, no. 3, pp. 415–427, doi:
10.1515/nanoph-2018-0176.
A. Pors, M. G. Nielsen, R. L. Eriksen, and S. I. Bozhevolnyi, “Broadband
Focusing Flat Mirrors Based on Plasmonic Gradient Metasurfaces,” Nano Lett.,
vol. 13, no. 2, pp. 829–834, Feb. 2013, doi: 10.1021/nl304761m.
Y.-W. Huang et al., “Aluminum Plasmonic Multicolor Meta-Hologram,” Nano
Lett., vol. 15, no. 5, pp. 3122–3127, May 2015, doi:
10.1021/acs.nanolett.5b00184.
W. T. Chen et al., “High-Efficiency Broadband Meta-Hologram with
Polarization-Controlled Dual Images,” Nano Lett., vol. 14, no. 1, pp. 225–230,
Jan. 2014, doi: 10.1021/nl403811d.
L.-J. Black, Y. Wang, C. H. de Groot, A. Arbouet, and O. L. Muskens, “Optimal
Polarization Conversion in Coupled Dimer Plasmonic Nanoantennas for
Metasurfaces,” ACS Nano, vol. 8, no. 6, pp. 6390–6399, Jun. 2014, doi:
10.1021/nn501889s.
P. C. Wu et al., “Versatile Polarization Generation with an Aluminum Plasmonic
Metasurface,” Nano Lett., vol. 17, no. 1, pp. 445–452, Jan. 2017, doi:
10.1021/acs.nanolett.6b04446.
M. J. R. Heck, “Highly integrated optical phased arrays: photonic integrated
circuits for optical beam shaping and beam steering,” Nanophotonics, vol. 6, no.
1, pp. 93–107, 2017, doi: https://doi.org/10.1515/nanoph-2015-0152.
A. Yaacobi, J. Sun, M. Moresco, G. Leake, D. Coolbaugh, and M. R. Watts,
“Integrated phased array for wide-angle beam steering,” Opt. Lett., vol. 39, no.
15, pp. 4575–4578, 2014, doi: 10.1364/OL.39.004575.
C. V Poulton et al., “Coherent solid-state LIDAR with silicon photonic optical
phased arrays,” Opt. Lett., vol. 42, no. 20, pp. 4091–4094, 2017, doi:
10.1364/OL.42.004091.
H. W. Lee et al., “Nanoscale Conducting Oxide PlasMOStor,” Nano Lett., vol.
14, no. 11, pp. 6463–6468, Nov. 2014, doi: 10.1021/nl502998z.

167
[81] A. P. Vasudev, J.-H. Kang, J. Park, X. Liu, and M. L. Brongersma, “Electrooptical modulation of a silicon waveguide with an ‘epsilon-near-zero’ material,”
Opt. Express, vol. 21, no. 22, pp. 26387–26397, 2013, doi:
10.1364/OE.21.026387.
[82] H. Zhao, Y. Wang, A. Capretti, L. D. Negro, and J. Klamkin, “Broadband
Electroabsorption Modulators Design Based on Epsilon-Near-Zero Indium Tin
Oxide,” IEEE J. Sel. Top. Quantum Electron., vol. 21, no. 4, pp. 192–198, 2015,
doi: 10.1109/JSTQE.2014.2375153.
[83] A. Klein et al., “Transparent Conducting Oxides for Photovoltaics: Manipulation
of Fermi Level, Work Function and Energy Band Alignment,” Mater. (Basel,
Switzerland), vol. 3, no. 11, pp. 4892–4914, Nov. 2010, doi:
10.3390/ma3114892.
[84] F. Yi, E. Shim, A. Y. Zhu, H. Zhu, J. C. Reed, and E. Cubukcu, “Voltage tuning
of plasmonic absorbers by indium tin oxide,” Appl. Phys. Lett., vol. 102, no. 22,
p. 221102, Jun. 2013, doi: 10.1063/1.4809516.
[85] J. Lee and W. D. Lu, “On-Demand Reconfiguration of Nanomaterials: When
Electronics Meets Ionics,” Adv. Mater., vol. 30, no. 1, p. 1702770, Jan. 2018,
doi: 10.1002/adma.201702770.
[86] P. D. Ye et al., “GaN metal-oxide-semiconductor high-electron-mobilitytransistor with atomic layer deposited Al2O3 as gate dielectric,” Appl. Phys. Lett.,
vol. 86, no. 6, p. 63501, Jan. 2005, doi: 10.1063/1.1861122.
[87] C. L. Hinkle et al., “GaAs interfacial self-cleaning by atomic layer deposition,”
Appl. Phys. Lett., vol. 92, no. 7, p. 71901, Feb. 2008, doi: 10.1063/1.2883956.
[88] D. Cao et al., “Low-temperature plasma-enhanced atomic layer deposition of
HfO2/Al2O3 nanolaminate structure on Si,” J. Vac. Sci. Technol. B, vol. 33, no.
1, p. 01A101, Sep. 2014, doi: 10.1116/1.4895010.
[89] Joseph W. Goodman, “Introduction to fourier optics,” Introduction to Fourier
optics, 2nd ed., Publisher: New York, NY: McGraw-Hill, 1996. Series: McGrawHill series in electrical and computer engineering ; Electromagnetics. ISBN:
0070242542,
vol.
1.
p.
441,
2006,
[Online].
Available:
[90] E. Feigenbaum, K. Diest, and H. A. Atwater, “Unity-Order Index Change in
Transparent Conducting Oxides at Visible Frequencies,” Nano Lett., vol. 10, no.
6, pp. 2111–2116, Jun. 2010, doi: 10.1021/nl1006307.
[91] J. . Kim, F. Cacialli, A. Cola, G. Gigli, and R. Cingolani, “Hall measurements of
treated indium tin oxide surfaces,” Synth. Met, vol. 111, pp. 363–367, 2000.
[92] A. Melikyan et al., “Surface plasmon polariton absorption modulator,” Opt.
Express, vol. 19, no. 9, pp. 8855–8869, 2011, doi: 10.1364/OE.19.008855.
[93] F. Michelotti, L. Dominici, E. Descrovi, N. Danz, and F. Menchini, “Thickness
dependence of surface plasmon polariton dispersion in transparent conducting
oxide films at 1.55 μm,” Opt. Lett., vol. 34, no. 6, pp. 839–841, 2009, doi:
10.1364/OL.34.000839.
[94] T. J. M. Bayer, A. Wachau, A. Fuchs, J. Deuermeier, and A. Klein, “Atomic
Layer Deposition of Al2O3 onto Sn-Doped In2O3: Absence of Self-Limited
Adsorption during Initial Growth by Oxygen Diffusion from the Substrate and
Band Offset Modification by Fermi Level Pinning in Al2O3,” Chem. Mater., vol.
24, no. 23, pp. 4503–4510, Dec. 2012, doi: 10.1021/cm301732t.
[95] G. K. Shirmanesh, R. Sokhoyan, P. C. Wu, and H. A. Atwater, “Electro-optically
Tunable Multifunctional Metasurfaces,” ACS Nano, vol. 14, no. 6, pp. 6912–
6920, Jun. 2020, doi: 10.1021/acsnano.0c01269.

168
[96] F. Aieta, P. Genevet, N. Yu, M. A. Kats, Z. Gaburro, and F. Capasso, “Out-ofPlane Reflection and Refraction of Light by Anisotropic Optical Antenna
Metasurfaces with Phase Discontinuities,” Nano Lett., vol. 12, no. 3, pp. 1702–
1706, Mar. 2012, doi: 10.1021/nl300204s.
[97] L. Huang et al., “Dispersionless Phase Discontinuities for Controlling Light
Propagation,” Nano Lett., vol. 12, no. 11, pp. 5750–5755, Nov. 2012, doi:
10.1021/nl303031j.
[98] Y.-J. Lu et al., “Dynamically Controlled Purcell Enhancement of Visible
Spontaneous Emission in a Gated Plasmonic Heterostructure,” Nat. Commun.,
vol. 8, no. 1, p. 1631, 2017, doi: 10.1038/s41467-017-01870-0.
[99] G. Si, Y. Zhao, E. S. P. Leong, and Y. J. Liu, “Liquid-Crystal-Enabled Active
Plasmonics: A Review,” Mater. (Basel, Switzerland), vol. 7, no. 2, pp. 1296–
1317, Feb. 2014, doi: 10.3390/ma7021296.
[100] B. Lee et al., “Fano Resonance and Spectrally Modified Photoluminescence
Enhancement in Monolayer MoS2 Integrated with Plasmonic Nanoantenna
Array,” Nano Lett., vol. 15, no. 5, pp. 3646–3653, May 2015, doi:
10.1021/acs.nanolett.5b01563.
[101] W. Zheng, F. Fan, M. Chen, S. Chen, and S.-J. Chang, “Optically pumped
terahertz wave modulation in MoS2-Si heterostructure metasurface,” AIP Adv.,
vol. 6, no. 7, p. 75105, Jul. 2016, doi: 10.1063/1.4958878.
[102] Y. Chen, X. Li, X. Luo, S. A. Maier, and M. Hong, “Tunable near-infrared
plasmonic perfect absorber based on phase-change materials,” Photonics Res.,
vol. 3, no. 3, pp. 54–57, 2015, doi: 10.1364/PRJ.3.000054.
[103] S. Yu, J. Zhang, L. Li, H. Zheng, and W. Zhang, “Low loss, high tunable
BaZr0.2Ti0.8O3/BaSn0.85Ti0.15O3 heterostructure thin films,” Appl. Phys.
Lett., vol. 109, no. 14, p. 142905, Oct. 2016, doi: 10.1063/1.4964462.
[104] R. D. Adams, “Dynamic Random Access Memories,” in High Performance
Memory Testing: Design Principles, Fault Modeling and Self-Test, Boston, MA:
Springer, 2003, pp. 77–87.
[105] R. D. Adams, “Static Random Access Memories,” in High Performance Memory
Testing: Design Principles, Fault Modeling and Self-Test, Boston, MA: Springer,
2003, pp. 17–46.
[106] S. D. Brown, J. R. Francis, J. Rose, and Z. G. Vranesic, “Introduction to FPGAs,”
in Field-Programmable Gate Arrays, Boston, MA: Springer, 1992, pp. 1–11.
[107] B. Schwarz, “Mapping the world in 3D,” Nat. Photonics, vol. 4, no. 7, pp. 429–
430, 2010, doi: 10.1038/nphoton.2010.148.
[108] C. Chaintoutis et al., “Free Space Intra-Datacenter Interconnects Based on 2D
Optical Beam Steering Enabled by Photonic Integrated Circuits,” Photonics , vol.
5, no. 3. 2018, doi: 10.3390/photonics5030021.
[109] F. Feng, I. H. White, and T. D. Wilkinson, “Free Space Communications With
Beam Steering a Two-Electrode Tapered Laser Diode Using Liquid-Crystal
SLM,” J. Light. Technol., vol. 31, no. 12, pp. 2001–2007, 2013, [Online].
Available: http://jlt.osa.org/abstract.cfm?URI=jlt-31-12-2001.
[110] B. D. Duncan, P. J. Bos, and V. Sergan, “Wide angle achromatic prism beam
steering for infrared countermeasure applications,” Opt. Eng., vol. 42, no. 4, pp.
1038–1047, Apr. 2003, doi: 10.1117/1.1556393.
[111] M. Jofre et al., “Fast beam steering with full polarization control using a
galvanometric optical scanner and polarization controller,” Opt. Express, vol. 20,
no. 11, pp. 12247–12260, 2012, doi: 10.1364/OE.20.012247.
[112] J. L. Gibson, B. D. Duncan, E. A. Watson, and J. S. Loomis, “Wide-angle

169
decentered lens beam steering for infrared countermeasures applications,” Opt.
Eng., vol. 43, no. 10, pp. 2312–2321, Oct. 2004, doi: 10.1117/1.1789137.
[113] C. W. Oh, Z. Cao, E. Tangdiongga, and T. Koonen, “Free-space transmission
with passive 2D beam steering for multi-gigabit-per-second per-beam indoor
optical wireless networks,” Opt. Express, vol. 24, no. 17, pp. 19211–19227, 2016,
doi: 10.1364/OE.24.019211.
[114] A. H. Naqvi and S. Lim, “A Beam-Steering Antenna With a Fluidically
Programmable Metasurface,” IEEE Trans. Antennas Propag., vol. 67, no. 6, pp.
3704–3711, 2019, doi: 10.1109/TAP.2019.2905690.
[115] L. Chen, Y. Ruan, and H. Y. Cui, “Liquid metal metasurface for flexible beamsteering,” Opt. Express, vol. 27, no. 16, pp. 23282–23292, 2019, doi:
10.1364/OE.27.023282.
[116] S.-Q. Li, X. Xu, R. Maruthiyodan Veetil, V. Valuckas, R. Paniagua-Domínguez,
and A. I. Kuznetsov, “Phase-only transmissive spatial light modulator based on
tunable dielectric metasurface,” Science (80-. )., vol. 364, no. 6445, pp. 1087 LP
– 1090, Jun. 2019, doi: 10.1126/science.aaw6747.
[117] P. C. Wu et al., “Dynamic Beam Steering with All-Dielectric Electro-Optic III–
V Multiple-Quantum-Well Metasurfaces,” Nat. Commun., vol. 10, no. 1, p. 3654,
2019, doi: 10.1038/s41467-019-11598-8.
[118] P. P. Iyer, M. Pendharkar, C. J. Palmstrøm, and J. A. Schuller, “III–V
Heterojunction Platform for Electrically Reconfigurable Dielectric
Metasurfaces,” ACS Photonics, vol. 6, no. 6, pp. 1345–1350, Jun. 2019, doi:
10.1021/acsphotonics.9b00178.
[119] V. J. Sorger, N. D. Lanzillotti-Kimura, R.-M. Ma, and X. Zhang, “Ultra-compact
silicon nanophotonic modulator with broadband response,” Nanophotonics, vol.
1, no. 1, pp. 17–22, 2012, doi: https://doi.org/10.1515/nanoph-2012-0009.
[120] C. T. DeRose et al., “Electronically controlled optical beam-steering by an active
phased array of metallic nanoantennas,” Opt. Express, vol. 21, no. 4, pp. 5198–
5208, 2013, doi: 10.1364/OE.21.005198.
[121] J. K. Doylend, M. J. R. Heck, J. T. Bovington, J. D. Peters, L. A. Coldren, and
J. E. Bowers, “Two-dimensional free-space beam steering with an optical phased
array on silicon-on-insulator,” Opt. Express, vol. 19, no. 22, pp. 21595–21604,
2011, doi: 10.1364/OE.19.021595.
[122] J. J. López et al., “Planar-lens Enabled Beam Steering for Chip-scale LIDAR,”
in Conference on Lasers and Electro-Optics, 2018, p. SM3I.1, doi:
10.1364/CLEO_SI.2018.SM3I.1.
[123] R. Fatemi, B. Abiri, A. Khachaturian, and A. Hajimiri, “High sensitivity active
flat optics optical phased array receiver with a two-dimensional aperture,” Opt.
Express, vol. 26, no. 23, pp. 29983–29999, 2018, doi: 10.1364/OE.26.029983.
[124] A. M. Shaltout et al., “Spatiotemporal light control with frequency-gradient
metasurfaces.,” Science, vol. 365, no. 6451, pp. 374–377, Jul. 2019, doi:
10.1126/science.aax2357.
[125] M. L. Tseng et al., “Metalenses: Advances and Applications,” Adv. Opt. Mater.,
vol. 6, no. 18, p. 1800554, Sep. 2018, doi: 10.1002/adom.201800554.
[126] A. Arbabi, Y. Horie, M. Bagheri, and A. Faraon, “Dielectric metasurfaces for
complete control of phase and polarization with subwavelength spatial
resolution and high transmission.,” Nat. Nanotechnol., vol. 10, no. 11, pp. 937–
943, Nov. 2015, doi: 10.1038/nnano.2015.186.
[127] C.-H. Liu et al., “Ultrathin van der Waals Metalenses.,” Nano Lett., vol. 18, no.
11, pp. 6961–6966, Nov. 2018, doi: 10.1021/acs.nanolett.8b02875.

170
[128] S. Colburn, A. Zhan, and A. Majumdar, “Varifocal zoom imaging with large
area focal length adjustable metalenses,” Optica, vol. 5, no. 7, pp. 825–831, 2018,
doi: 10.1364/OPTICA.5.000825.
[129] H.-C. Lin, M.-S. Chen, and Y.-H. Lin, “A review of electrically tunable focusing
liquid crystal lenses,” vol. 12, Dec. 2011, doi: 10.4313/TEEM.2011.12.6.234.
[130] N. Chronis, G. L. Liu, K.-H. Jeong, and L. P. Lee, “Tunable liquid-filled
microlens array integrated with microfluidic network,” Opt. Express, vol. 11, no.
19, pp. 2370–2378, 2003, doi: 10.1364/OE.11.002370.
[131] W. Zhu et al., “A flat lens with tunable phase gradient by using random access
reconfigurable metamaterial.,” Adv. Mater., vol. 27, no. 32, pp. 4739–4743, Aug.
2015, doi: 10.1002/adma.201501943.
[132] Y. Yang et al., “Femtosecond optical polarization switching using a cadmium
oxide-based perfect absorber,” Nat. Photonics, vol. 11, no. 6, pp. 390–395, 2017,
doi: 10.1038/nphoton.2017.64.
[133] Z. Gongjian, Z. Man, and Z. Yang, “Wave front control with SLM and
simulation of light wave diffraction,” Opt. Express, vol. 26, no. 26, pp. 33543–
33564, 2018, doi: 10.1364/OE.26.033543.
[134] G. J. Swanson, “Binary Optics Technology: The Theory and Design of MultiLevel Diffractive Optical Elements,” Cambridge, MA, 1989.
[135] G. J. Swanson, “Binary Optics Technology: Theoretical Limits on the
Diffraction Efficiency of Multilevel Diffractive Optical Elements,” Cambridge,
MA, 1991.
[136] C. A. Balanis, “Fundamental Parameters and Figures-of-Merit of Antennas,” in
Antenna Theory: Analysis and Design, 3rd ed., New York: John Wiley & Sons,
1997, pp. 44–57.
[137] D. Engström, J. Bengtsson, E. Eriksson, and M. Goksör, “Improved beam
steering accuracy of a single beam with a 1D phase-only spatial light modulator,”
Opt. Express, vol. 16, no. 22, pp. 18275–18287, 2008, doi:
10.1364/OE.16.018275.
[138] S. Han, S. Kim, S. Kim, T. Low, V. W. Brar, and M. S. Jang, “Complete
Complex Amplitude Modulation with Electronically Tunable Graphene
Plasmonic Metamolecules,” ACS Nano, vol. 14, no. 1, pp. 1166–1175, Jan. 2020,
doi: 10.1021/acsnano.9b09277.
[139] R. J. Lin et al., “Achromatic metalens array for full-colour light-field imaging,”
Nat. Nanotechnol., vol. 14, no. 3, pp. 227–231, 2019, doi: 10.1038/s41565-0180347-0.
[140] S. Wang et al., “A broadband achromatic metalens in the visible,” Nat.
Nanotechnol., vol. 13, no. 3, pp. 227–232, 2018, doi: 10.1038/s41565-017-00524.
[141] A. Forouzmand and H. Mosallaei, “Electro-optical Amplitude and Phase
Modulators Based on Tunable Guided-Mode Resonance Effect,” ACS Photonics,
vol. 6, no. 11, pp. 2860–2869, Nov. 2019, doi: 10.1021/acsphotonics.9b00950.
[142] T. Phan et al., “High-efficiency, large-area, topology-optimized metasurfaces,”
Light Sci. Appl., vol. 8, no. 1, p. 48, 2019, doi: 10.1038/s41377-019-0159-5.
[143] S. D. Campbell, D. Sell, R. P. Jenkins, E. B. Whiting, J. A. Fan, and D. H. Werner,
“Review of numerical optimization techniques for meta-device design [Invited],”
Opt. Mater. Express, vol. 9, no. 4, pp. 1842–1863, 2019, doi:
10.1364/OME.9.001842.
[144] D. Liu, Y. Tan, E. Khoram, and Z. Yu, “Training Deep Neural Networks for the
Inverse Design of Nanophotonic Structures,” ACS Photonics, vol. 5, no. 4, pp.

171
1365–1369, Apr. 2018, doi: 10.1021/acsphotonics.7b01377.
[145] S. Jafar-Zanjani, S. Inampudi, and H. Mosallaei, “Adaptive Genetic Algorithm
for Optical Metasurfaces Design,” Sci. Rep., vol. 8, no. 1, p. 11040, 2018, doi:
10.1038/s41598-018-29275-z.
[146] H. Chung and O. D. Miller, “Tunable metasurface inverse design for 80%
switching efficiencies and 144$^\circ$ angular steering,” Oct. 2019, doi:
[147] “Fundamentals of Optics.” pp. 31–104, 2005, doi: 10.1201/9781420039528.ch2.
[148] R. Shankar, “Fundamentals of physics II : electromagnetism, optics, and
quantum mechanics.” p. 591, 2016.
[149] S. Vandendriessche et al., “Giant Faraday Rotation in Mesogenic Organic
Molecules,” Chem. Mater., vol. 25, no. 7, pp. 1139–1143, Apr. 2013, doi:
10.1021/cm4004118.
[150] S. Sun, Q. He, J. Hao, S. Xiao, and L. Zhou, “Electromagnetic metasurfaces:
physics and applications,” Adv. Opt. Photonics, vol. 11, no. 2, pp. 380–479, 2019,
doi: 10.1364/AOP.11.000380.
[151] X. Luo, “Engineering Optics 2.0: A Revolution in Optical Materials, Devices,
and Systems,” ACS Photonics, vol. 5, no. 12, pp. 4724–4738, Dec. 2018, doi:
10.1021/acsphotonics.8b01036.
[152] F. Yesilkoy et al., “Ultrasensitive hyperspectral imaging and biodetection
enabled by dielectric metasurfaces,” Nat. Photonics, vol. 13, no. 6, pp. 390–396,
2019, doi: 10.1038/s41566-019-0394-6.
[153] B. H. Chen et al., “GaN Metalens for Pixel-Level Full-Color Routing at Visible
Light,” Nano Lett., vol. 17, no. 10, pp. 6345–6352, Oct. 2017, doi:
10.1021/acs.nanolett.7b03135.
[154] Q. Guo et al., “Compact single-shot metalens depth sensors inspired by eyes of
jumping spiders,” Proc. Natl. Acad. Sci., vol. 116, no. 46, pp. 22959 LP – 22965,
Nov. 2019, doi: 10.1073/pnas.1912154116.
[155] Y.-Y. Xie et al., “Metasurface-integrated vertical cavity surface-emitting lasers
for programmable directional lasing emissions,” Nat. Nanotechnol., vol. 15, no.
2, pp. 125–130, 2020, doi: 10.1038/s41565-019-0611-y.
[156] H. Sroor et al., “High-purity orbital angular momentum states from a visible
metasurface laser,” Nat. Photonics, 2020, doi: 10.1038/s41566-020-0623-z.
[157] Y. Hu et al., “All-dielectric metasurfaces for polarization manipulation:
principles and emerging applications,” Nanophotonics, no. 0, p. 20200220, 2020,
doi: https://doi.org/10.1515/nanoph-2020-0220.
[158] H.-T. Chen, A. J. Taylor, and N. Yu, “A Review of Metasurfaces: Physics and
Applications,” Reports Prog. Phys., vol. 79, no. 7, p. 76401, 2016, doi:
10.1088/0034-4885/79/7/076401.
[159] P. Yu et al., “Controllable optical activity with non-chiral plasmonic
metasurfaces,” Light Sci. Appl., vol. 5, no. 7, pp. e16096–e16096, 2016, doi:
10.1038/lsa.2016.96.
[160] Z. Shi et al., “Continuous angle-tunable birefringence with freeform
metasurfaces for arbitrary polarization conversion,” Sci. Adv., vol. 6, no. 23, p.
eaba3367, Jun. 2020, doi: 10.1126/sciadv.aba3367.
[161] P. C. Wu et al., “Broadband Wide-Angle Multifunctional Polarization Converter
via Liquid-Metal-Based Metasurface,” Adv. Opt. Mater., vol. 5, no. 7, p.
1600938, Apr. 2017, doi: 10.1002/adom.201600938.
[162] Z. Wu, Y. Ra’di, and A. Grbic, “Tunable Metasurfaces: A Polarization Rotator
Design,” Phys. Rev. X, vol. 9, no. 1, p. 11036, Feb. 2019, doi:

172
10.1103/PhysRevX.9.011036.
[163] I. S. Osborne, “A metasurface polarization camera,” Science (80-. )., vol. 365,
no. 6448, pp. 40 LP – 42, Jul. 2019, doi: 10.1126/science.365.6448.40-j.
[164] T. Stav et al., “Quantum entanglement of the spin and orbital angular momentum
of photons using metamaterials,” Science (80-. )., vol. 361, no. 6407, pp. 1101
LP – 1104, Sep. 2018, doi: 10.1126/science.aat9042.
[165] A. Pors, M. G. Nielsen, and S. I. Bozhevolnyi, “Plasmonic metagratings for
simultaneous determination of Stokes parameters,” Optica, vol. 2, no. 8, pp.
716–723, 2015, doi: 10.1364/OPTICA.2.000716.
[166] P. C. Wu et al., “Visible Metasurfaces for On-Chip Polarimetry,” ACS Photonics,
vol. 5, no. 7, pp. 2568–2573, Jul. 2018, doi: 10.1021/acsphotonics.7b01527.
[167] L. Jin et al., “Noninterleaved Metasurface for (2(6)-1) Spin- and WavelengthEncoded Holograms.,” Nano Lett., vol. 18, no. 12, pp. 8016–8024, Dec. 2018,
doi: 10.1021/acs.nanolett.8b04246.
[168] R. Zhao et al., “Multichannel vectorial holographic display and encryption,”
Light Sci. Appl., vol. 7, no. 1, p. 95, 2018, doi: 10.1038/s41377-018-0091-0.
[169] J. Li et al., “Optical Polarization Encoding Using Graphene-Loaded Plasmonic
Metasurfaces,” Adv. Opt. Mater., vol. 4, no. 1, pp. 91–98, Jan. 2016, doi:
10.1002/adom.201500398.
[170] H. Cheng et al., “Dynamically tunable broadband mid-infrared cross
polarization converter based on graphene metamaterial,” Appl. Phys. Lett., vol.
103, no. 22, p. 223102, Nov. 2013, doi: 10.1063/1.4833757.
[171] L. Cong, N. Xu, J. Gu, R. Singh, J. Han, and W. Zhang, “Highly flexible
broadband terahertz metamaterial quarter-wave plate,” Laser Photon. Rev., vol.
8, no. 4, pp. 626–632, Jul. 2014, doi: 10.1002/lpor.201300205.
[172] R. H. Fan et al., “Freely tunable broadband polarization rotator for terahertz
waves,” Advanced Materials, vol. 27, no. 7. pp. 1201–1206, 2015, doi:
10.1002/adma.201404981.
[173] W. Li et al., “A Reconfigurable Polarization Converter Using Active
Metasurface and Its Application in Horn Antenna,” IEEE Trans. Antennas
Propag., vol. 64, no. 12, pp. 5281–5290, 2016, doi: 10.1109/TAP.2016.2620484.
[174] Z.-K. Zhou et al., “Quantum plasmonics get applied,” Prog. Quantum Electron.,
vol. 65, pp. 1–20, 2019, doi: https://doi.org/10.1016/j.pquantelec.2019.04.002.
[175] M.-X. Ren, W. Wu, W. Cai, B. Pi, X.-Z. Zhang, and J.-J. Xu, “Reconfigurable
metasurfaces that enable light polarization control by light.,” Light. Sci. Appl.,
vol. 6, no. 6, p. e16254, Jun. 2017, doi: 10.1038/lsa.2016.254.
[176] M. M. Salary and H. Mosallaei, “Tunable All-Dielectric Metasurfaces for PhaseOnly Modulation of Transmitted Light Based on Quasi-bound States in the
Continuum,” ACS Photonics, Jun. 2020, doi: 10.1021/acsphotonics.0c00554.
[177] P. C. Wu, R. Sokhoyan, G. K. Shirmanesh, W.-H. Cheng, and H. A. Atwater,
“Tunable Metasurface for Dynamic Control of Polarization in Near-Infrared,”
Prep., 2020.
[178] Y. Shirasaki, G. J. Supran, M. G. Bawendi, and V. Bulović, “Emergence of
colloidal quantum-dot light-emitting technologies,” Nat. Photonics, vol. 7, no. 1,
pp. 13–23, 2013, doi: 10.1038/nphoton.2012.328.
[179] E. Yablonovitch, “Inhibited Spontaneous Emission in Solid-State Physics and
Electronics,” Phys. Rev. Lett., vol. 58, no. 20, pp. 2059–2062, May 1987, doi:
10.1103/PhysRevLett.58.2059.
[180] L. Novotny and B. Hecht, Principles of nano-optics, 2nd edn. Cambridge
University Press, 2012.

173
[181] M. Pelton, “Modified spontaneous emission in nanophotonic structures,” Nat.
Photonics, vol. 9, no. 7, pp. 427–435, 2015, doi: 10.1038/nphoton.2015.103.
[182] E. M. Purcell, H. C. Torrey, and R. V Pound, “Resonance Absorption by Nuclear
Magnetic Moments in a Solid,” Phys. Rev., vol. 69, no. 1–2, pp. 37–38, Jan. 1946,
doi: 10.1103/PhysRev.69.37.
[183] V. S. C. M. Rao and S. Hughes, “Single Quantum Dot Spontaneous Emission in
a Finite-Size Photonic Crystal Waveguide: Proposal for an Efficient ``On Chip’’
Single Photon Gun,” Phys. Rev. Lett., vol. 99, no. 19, p. 193901, Nov. 2007, doi:
10.1103/PhysRevLett.99.193901.
[184] J. T. Robinson, C. Manolatou, L. Chen, and M. Lipson, “Ultrasmall Mode
Volumes in Dielectric Optical Microcavities,” Phys. Rev. Lett., vol. 95, no. 14,
p. 143901, Sep. 2005, doi: 10.1103/PhysRevLett.95.143901.
[185] S. Nakayama, S. Ishida, S. Iwamotoa, and Y. Arakawa, “Effect of cavity mode
volume on photoluminescence from silicon photonic crystal nanocavities,” Appl.
Phys. Lett., vol. 98, no. 171102, 2011, doi: 10.1063/1.3582035.
[186] J. M. Gérard, B. Sermage, B. Gayral, B. Legrand, E. Costard, and V. ThierryMieg, “Enhanced Spontaneous Emission by Quantum Boxes in a Monolithic
Optical Microcavity,” Phys. Rev. Lett., vol. 81, no. 5, pp. 1110–1113, Aug. 1998,
doi: 10.1103/PhysRevLett.81.1110.
[187] D. Englund et al., “Controlling the Spontaneous Emission Rate of Single
Quantum Dots in a Two-Dimensional Photonic Crystal,” Phys. Rev. Lett., vol.
95, no. 1, p. 13904, Jul. 2005, doi: 10.1103/PhysRevLett.95.013904.
[188] J. Bleuse et al., “Inhibition, Enhancement, and Control of Spontaneous Emission
in Photonic Nanowires,” Phys. Rev. Lett., vol. 106, no. 10, p. 103601, Mar. 2011,
doi: 10.1103/PhysRevLett.106.103601.
[189] A. David, H. Benisty, and C. Weisbuch, “Photonic crystal light-emitting
sources.,” Rep. Prog. Phys., vol. 75, no. 12, p. 126501, Dec. 2012, doi:
10.1088/0034-4885/75/12/126501.
[190] H. Altug, D. Englund, and J. Vučković, “Ultrafast photonic crystal nanocavity
laser,” Nat. Phys., vol. 2, no. 7, pp. 484–488, 2006, doi: 10.1038/nphys343.
[191] K. Hennessy et al., “Quantum nature of a strongly coupled single quantum dotcavity system.,” Nature, vol. 445, no. 7130, pp. 896–899, Feb. 2007, doi:
10.1038/nature05586.
[192] K. J. Russell, T.-L. Liu, S. Cui, and E. L. Hu, “Large spontaneous emission
enhancement in plasmonic nanocavities,” Nat. Photonics, vol. 6, no. 7, pp. 459–
462, 2012, doi: 10.1038/nphoton.2012.112.
[193] A. Kinkhabwala, Z. Yu, S. Fan, Y. Avlasevich, K. Müllen, and W. E. Moerner,
“Large single-molecule fluorescence enhancements produced by a bowtie
nanoantenna,” Nat. Photonics, vol. 3, no. 11, pp. 654–657, 2009, doi:
10.1038/nphoton.2009.187.
[194] E. J. R. Vesseur, F. J. G. de Abajo, and A. Polman, “Broadband Purcell
enhancement in plasmonic ring cavities,” Phys. Rev. B, vol. 82, no. 16, p. 165419,
Oct. 2010, doi: 10.1103/PhysRevB.82.165419.
[195] R. Esteban, T. V Teperik, and J. J. Greffet, “Optical Patch Antennas for Single
Photon Emission Using Surface Plasmon Resonances,” Phys. Rev. Lett., vol. 104,
no. 2, p. 26802, Jan. 2010, doi: 10.1103/PhysRevLett.104.026802.
[196] C. Belacel et al., “Controlling Spontaneous Emission with Plasmonic Optical
Patch Antennas,” Nano Lett., vol. 13, no. 4, pp. 1516–1521, Apr. 2013, doi:
10.1021/nl3046602.
[197] M. Yi et al., “Plasmonic Interaction Between Silver Nano-Cubes and a Silver

174
Ground Plane Studied by Surface-Enhanced Raman Scattering,” Plasmonics, vol.
6, no. 3, pp. 515–519, 2011, doi: 10.1007/s11468-011-9230-7.
[198] G. M. Akselrod et al., “Probing the mechanisms of large Purcell enhancement in
plasmonic nanoantennas,” Nat. Photonics, vol. 8, no. 11, pp. 835–840, 2014, doi:
10.1038/nphoton.2014.228.
[199] J. B. Lassiter et al., “Plasmonic Waveguide Modes of Film-Coupled Metallic
Nanocubes,” Nano Lett., vol. 13, no. 12, pp. 5866–5872, Dec. 2013, doi:
10.1021/nl402660s.
[200] C. T. Yuan et al., “Modification of Fluorescence Properties in Single Colloidal
Quantum Dots by Coupling to Plasmonic Gap Modes,” J. Phys. Chem. C, vol.
117, no. 24, pp. 12762–12768, Jun. 2013, doi: 10.1021/jp401993r.
[201] T. B. Hoang, G. M. Akselrod, and M. H. Mikkelsen, “Ultrafast RoomTemperature Single Photon Emission from Quantum Dots Coupled to Plasmonic
Nanocavities,” Nano Lett., vol. 16, no. 1, pp. 270–275, Jan. 2016, doi:
10.1021/acs.nanolett.5b03724.
[202] A. Rose et al., “Control of Radiative Processes Using Tunable Plasmonic
Nanopatch Antennas,” Nano Lett., vol. 14, no. 8, pp. 4797–4802, Aug. 2014, doi:
10.1021/nl501976f.
[203] J. Grandidier et al., “Gain-Assisted Propagation in a Plasmonic Waveguide at
Telecom Wavelength,” Nano Lett., vol. 9, no. 8, pp. 2935–2939, Aug. 2009, doi:
10.1021/nl901314u.
[204] V. L. Colvin, M. C. Schlamp, and A. P. Alivisatos, “Light-emitting diodes made
from cadmium selenide nanocrystals and a semiconducting polymer,” Nature,
vol. 370, no. 6488, pp. 354–357, 1994, doi: 10.1038/370354a0.
[205] A. Laucht et al., “Electrical control of spontaneous emission and strong coupling
for a single quantum dot,” New J. Phys., vol. 11, no. 2, p. 23034, 2009, doi:
10.1088/1367-2630/11/2/023034.
[206] C.-Y. Jin et al., “Ultrafast non-local control of spontaneous emission,” Nat.
Nanotechnol., vol. 9, no. 11, pp. 886–890, 2014, doi: 10.1038/nnano.2014.190.
[207] P. B. Johnson and R. W. Christy, “Optical Constants of the Noble Metals,” Phys.
Rev. B, vol. 6, no. 12, pp. 4370–4379, Dec. 1972, doi: 10.1103/PhysRevB.6.4370.
[208] S. Cueff et al., “Dynamic control of light emission faster than the lifetime limit
using VO2 phase-change.,” Nat. Commun., vol. 6, p. 8636, Oct. 2015, doi:
10.1038/ncomms9636.
[209] K. J. Tielrooij et al., “Electrical control of optical emitter relaxation pathways
enabled by graphene,” Nat. Phys., vol. 11, no. 3, pp. 281–287, 2015, doi:
10.1038/nphys3204.
[210] Z. Ma, W. Cai, Y. Xiang, M. Ren, X. Zhang, and J. Xu, “Dynamic spontaneous
emission control of an optical emitter coupled to plasmons in strained graphene,”
Opt. Express, vol. 25, no. 19, pp. 23070–23081, 2017, doi:
10.1364/OE.25.023070.
[211] A. Reserbat-Plantey et al., “Electromechanical control of nitrogen-vacancy
defect emission using graphene NEMS,” Nat. Commun., vol. 7, no. 1, p. 10218,
2016, doi: 10.1038/ncomms10218.
[212] F. Ding, A. Pors, and S. I. Bozhevolnyi, “Gradient metasurfaces: a review of
fundamentals and applications,” Reports Prog. Phys., vol. 81, no. 2, p. 26401,
2017, doi: 10.1088/1361-6633/aa8732.
[213] M. Chen, M. Kim, A. M. H. Wong, and G. V Eleftheriades, “Huygens’
metasurfaces from microwaves to optics: a review,” Nanophotonics, vol. 7, no.
6, pp. 1207–1231, 2018, doi: https://doi.org/10.1515/nanoph-2017-0117.

175
[214] P. Genevet, F. Capasso, F. Aieta, M. Khorasaninejad, and R. Devlin, “Recent
advances in planar optics: from plasmonic to dielectric metasurfaces,” Optica,
vol. 4, no. 1, pp. 139–152, 2017, doi: 10.1364/OPTICA.4.000139.
[215] X. Ni, N. K. Emani, A. V Kildishev, A. Boltasseva, and V. M. Shalaev,
“Broadband Light Bending with Plasmonic Nanoantennas,” Science (80-. )., vol.
335, no. 6067, pp. 427 LP – 427, Jan. 2012, doi: 10.1126/science.1214686.
[216] W. T. Chen et al., “A broadband achromatic metalens for focusing and imaging
in the visible,” Nat. Nanotechnol., vol. 13, no. 3, pp. 220–226, 2018, doi:
10.1038/s41565-017-0034-6.
[217] X. Ding et al., “Ultrathin Pancharatnam–Berry Metasurface with Maximal
Cross-Polarization Efficiency,” Adv. Mater., vol. 27, no. 7, pp. 1195–1200, Feb.
2015, doi: 10.1002/adma.201405047.
[218] Y. Zhao and A. Alù, “Manipulating light polarization with ultrathin plasmonic
metasurfaces,” Phys. Rev. B, vol. 84, no. 20, p. 205428, Nov. 2011, doi:
10.1103/PhysRevB.84.205428.
[219] N. A. Rubin et al., “Polarization state generation and measurement with a single
metasurface,” Opt. Express, vol. 26, no. 17, pp. 21455–21478, 2018, doi:
10.1364/OE.26.021455.
[220] W. T. Chen et al., “Integrated plasmonic metasurfaces for spectropolarimetry,”
Nanotechnology, vol. 27, no. 22, p. 224002, 2016, doi: 10.1088/09574484/27/22/224002.
[221] J. Lin, P. Genevet, M. A. Kats, N. Antoniou, and F. Capasso, “Nanostructured
Holograms for Broadband Manipulation of Vector Beams,” Nano Lett., vol. 13,
no. 9, pp. 4269–4274, Sep. 2013, doi: 10.1021/nl402039y.
[222] P. K. Jha, X. Ni, C. Wu, Y. Wang, and X. Zhang, “Metasurface-Enabled Remote
Quantum Interference,” Phys. Rev. Lett., vol. 115, no. 2, p. 25501, Jul. 2015, doi:
10.1103/PhysRevLett.115.025501.
[223] K. Wang et al., “Quantum metasurface for multiphoton interference and state
reconstruction,” Science (80-. )., vol. 361, no. 6407, pp. 1104 LP – 1108, Sep.
2018, doi: 10.1126/science.aat8196.
[224] A. Nemati, Q. Wang, M. Hong, and J. Teng, “Tunable and reconfigurable
metasurfaces and metadevices,” Opto-Electronic Adv., vol. 1, no. 5, p. 180009,
2018, doi: 10.29026/oea.2018.180009.
[225] V. E. Babicheva, A. Boltasseva, and A. V Lavrinenko, “Transparent conducting
oxides for electro-optical plasmonic modulators,” Nanophotonics, vol. 4, no. 2,
pp. 165–185, 2015, doi: https://doi.org/10.1515/nanoph-2015-0004.
[226] J. A. Dionne, K. Diest, L. A. Sweatlock, and H. A. Atwater, “PlasMOStor: A
Metal−Oxide−Si Field Effect Plasmonic Modulator,” Nano Lett., vol. 9, no. 2,
pp. 897–902, Feb. 2009, doi: 10.1021/nl803868k.
[227] V. E. Babicheva et al., “Towards CMOS-compatible nanophotonics: Ultracompact modulators using alternative plasmonic materials,” Opt. Express, vol.
21, no. 22, pp. 27326–27337, 2013, doi: 10.1364/OE.21.027326.
[228] L. Jin, Q. Chen, W. Liu, and S. Song, “Electro-absorption Modulator with Dual
Carrier Accumulation Layers Based on Epsilon-Near-Zero ITO,” Plasmonics,
vol. 11, no. 4, pp. 1087–1092, 2016, doi: 10.1007/s11468-015-0146-5.
[229] J. Baek, J.-B. You, and K. Yu, “Free-carrier electro-refraction modulation based
on a silicon slot waveguide with ITO,” Opt. Express, vol. 23, no. 12, pp. 15863–
15876, 2015, doi: 10.1364/oe.23.015863.
[230] A. Forouzmand and H. Mosallaei, “Tunable two dimensional optical beam
steering with reconfigurable indium tin oxide plasmonic reflectarray

176
metasurface,” J. Opt., vol. 18, no. 12, p. 125003, 2016, doi: 10.1088/20408978/18/12/125003.
[231] A. Forouzmand, M. M. Salary, S. Inampudi, and H. Mosallaei, “A Tunable
Multigate Indium-Tin-Oxide-Assisted All-Dielectric Metasurface,” Adv. Opt.
Mater., vol. 6, no. 7, p. 1701275, Apr. 2018, doi: 10.1002/adom.201701275.
[232] A. Forouzmand and H. Mosallaei, “Real-Time Controllable and Multifunctional
Metasurfaces Utilizing Indium Tin Oxide Materials: A Phased Array
Perspective,” IEEE Trans. Nanotechnol., vol. 16, no. 2, pp. 296–306, 2017, doi:
10.1109/TNANO.2017.2662638.
[233] M. M. Salary and H. Mosallaei, “Electrically Tunable Metamaterials Based on
Multimaterial Nanowires Incorporating Transparent Conductive Oxides,” Sci.
Rep., vol. 7, no. 1, p. 10055, 2017, doi: 10.1038/s41598-017-09523-4.
[234] J. Park and S. J. Kim, “Subwavelength-spaced transmissive metallic slits for
360-degree phase control by using transparent conducting oxides,” Appl. Opt.,
vol. 57, no. 21, pp. 6027–6031, 2018, doi: 10.1364/AO.57.006027.
[235] S. J. Kim and M. L. Brongersma, “Active flat optics using a guided mode
resonance,” Opt. Lett., vol. 42, no. 1, pp. 5–8, 2017, doi: 10.1364/OL.42.000005.
[236] X. Hu, Q. Chen, L. Wen, L. Jin, H. Wang, and W. Liu, “Modulating Spatial Light
by Grating Slot Waveguides With Transparent Conducting Oxides,” IEEE
Photonics Technol. Lett., vol. 28, no. 15, pp. 1665–1668, 2016, doi:
10.1109/LPT.2016.2565507.
[237] M. R. Shcherbakov et al., “Ultrafast all-optical tuning of direct-gap
semiconductor metasurfaces,” Nat. Commun., vol. 8, no. 1, p. 17, 2017, doi:
10.1038/s41467-017-00019-3.
[238] V. W. Brar, M. S. Jang, M. Sherrott, J. J. Lopez, and H. A. Atwater, “Highly
Confined Tunable Mid-Infrared Plasmonics in Graphene Nanoresonators,” Nano
Lett., vol. 13, no. 6, pp. 2541–2547, Jun. 2013, doi: 10.1021/nl400601c.
[239] A. Benz, I. Montaño, J. F. Klem, and I. Brener, “Tunable metamaterials based
on voltage controlled strong coupling,” Appl. Phys. Lett., vol. 103, no. 26, p.
263116, Dec. 2013, doi: 10.1063/1.4859636.
[240] J. Lee et al., “Ultrafast Electrically Tunable Polaritonic Metasurfaces,” Adv. Opt.
Mater., vol. 2, no. 11, pp. 1057–1063, Nov. 2014, doi: 10.1002/adom.201400185.
[241] R. Sarma et al., “A metasurface optical modulator using voltage-controlled
population of quantum well states,” Appl. Phys. Lett., vol. 113, no. 20, p. 201101,
Nov. 2018, doi: 10.1063/1.5055013.
[242] Y.-H. Kuo et al., “Strong quantum-confined Stark effect in germanium quantumwell structures on silicon,” Nature, vol. 437, no. 7063, pp. 1334–1336, 2005, doi:
10.1038/nature04204.
[243] P. Chaisakul et al., “Recent progress in GeSi electro-absorption modulators,” Sci.
Technol. Adv. Mater., vol. 15, no. 1, p. 14601, 2014, doi: 10.1088/14686996/15/1/014601.
[244] “Lumerical Inc.” http://www.lumerical.com/tcad-products/device/.
[245] G. Backenstoss, “Conductivity Mobilities of Electrons and Holes in Heavily
Doped Silicon,” Phys. Rev., vol. 108, no. 6, pp. 1416–1419, Dec. 1957, doi:
10.1103/PhysRev.108.1416.
[246] H. A. Haus, Waves and fields in optoelectronics. Englewood Cliffs, NJ :
Prentice-Hall, [1984] ©1984.
[247] S. Fan, W. Suh, and J. D. Joannopoulos, “Temporal coupled-mode theory for the
Fano resonance in optical resonators,” J. Opt. Soc. Am. A, vol. 20, no. 3, pp. 569–
572, 2003, doi: 10.1364/JOSAA.20.000569.

177
[248] K. X. Wang, Z. Yu, S. Sandhu, and S. Fan, “Fundamental bounds on decay rates
in asymmetric single-mode optical resonators,” Opt. Lett., vol. 38, no. 2, pp.
100–102, 2013, doi: 10.1364/OL.38.000100.
[249] Y. Yu, Y. Chen, H. Hu, W. Xue, K. Yvind, and J. Mork, “Nonreciprocal
transmission in a nonlinear photonic-crystal Fano structure with broken
symmetry,” Laser Photon. Rev., vol. 9, no. 2, pp. 241–247, Mar. 2015, doi:
10.1002/lpor.201400207.
[250] K. X. Wang, Z. Yu, S. Sandhu, V. Liu, and S. Fan, “Condition for perfect
antireflection by optical resonance at material interface,” Optica, vol. 1, no. 6,
pp. 388–395, 2014, doi: 10.1364/OPTICA.1.000388.
[251] J. W. Yoon and R. Magnusson, “Fano resonance formula for lossy two-port
systems,” Opt. Express, vol. 21, no. 15, pp. 17751–17759, 2013, doi:
10.1364/OE.21.017751.
[252] E. D. Palik, Handbook of Optical Constants of Solids. New York: Academic
Press, 1985.
[253] J. Cheng, S. Inampudi, and H. Mosallaei, “Optimization-based Dielectric
Metasurfaces for Angle-Selective Multifunctional Beam Deflection,” Sci. Rep.,
vol. 7, no. 1, p. 12228, 2017, doi: 10.1038/s41598-017-12541-x.
[254] M. G. Moharam and T. K. Gaylord, “Rigorous coupled-wave analysis of planargrating diffraction,” J. Opt. Soc. Am., vol. 71, no. 7, pp. 811–818, 1981, doi:
10.1364/JOSA.71.000811.
[255] Y. Cho et al., “Optical device including three coupled quantum well structure,”
US20150286078A1, 2015.
[256] V. Savinov, V. A. Fedotov, and N. I. Zheludev, “Toroidal dipolar excitation and
macroscopic electromagnetic properties of metamaterials,” Phys. Rev. B, vol. 89,
no. 20, p. 205112, May 2014, doi: 10.1103/PhysRevB.89.205112.
[257] P. C. Wu et al., “Optical Anapole Metamaterial,” ACS Nano, vol. 12, no. 2, pp.
1920–1927, Feb. 2018, doi: 10.1021/acsnano.7b08828.
[258] A. Y. Zhu et al., “Giant intrinsic chiro-optical activity in planar dielectric
nanostructures,” Light Sci. Appl., vol. 7, no. 2, p. 17158, 2018, doi:
10.1038/lsa.2017.158.
[259] Z. Xu, C. Wang, W. Qi, and Z. Yuan, “Electro-optical effects in straincompensated InGaAs/InAlAs coupled quantum wells with modified potential,”
Opt. Lett., vol. 35, no. 5, pp. 736–738, 2010, doi: 10.1364/OL.35.000736.
[260] H. Feng, J. P. Pang, M. Sugiyama, K. Tada, and Y. Nakano, “Field-induced
optical effect in a five-step asymmetric coupled quantum well with modified
potential,” IEEE J. Quantum Electron., vol. 34, no. 7, pp. 1197–1208, 1998, doi:
10.1109/3.687863.

178

APPENDIX

Appendix A.1. Effect of the Thickness of the ITO Layer on the Device Tunability
In designing our dual-gated metasurface, the choice of the thickness of the ITO layer
plays a crucial role in the amount of tunability our metasurface can provide. Figure A.1
shows the maximum achievable relative reflectance change and maximum phase
modulation obtained when changing the thickness of the ITO layer. As can be seen,
increasing the thickness of the ITO layer will result in a smaller amount of achievable
modulation. This means that the thinner the ITO layer, the more modulation can be
achieved. On the other hand, we were not able to perform Hall measurement and
ellipsometry fitting on ultra-thin ITO layers. As a result, we chose the thickness of the
ITO layer to be 5 nm.
The presented results provide evidence of the fact that further increasing the number of
layers will not further increase the phase shift coverage. This is due to the fact that both
the HfO2/Al2O3 nanolaminate and the ITO layer have to be thicker than a specific lower
limit to provide the desirable properties.

Figure A.1: Effect of the thickness of the ITO layer on the device tunability. (a) Relative
reflectance change spectrum for different thicknesses of the ITO layer, and (b) maximum
achievable relative reflectance change as a function of ITO layer thickness. (c) Phase shift
spectrum for different thicknesses of the ITO layer, and (d) maximum achievable phase
modulation as a function of ITO thickness.

179
Appendix A.2. Fabrication and Characterization of HAOL
During our fabrication process, we grow thin HAOL films by using ALD and compare
their properties with separately grown Al2O3 and HfO2 films.
The deposition is performed at 150oC by using a thermal recipe in the ALD tool (Fiji
G2 Plasma Enhanced Atomic Layer Deposition System). During our ALD process, we
use tetrakis (ethylmethylamino) hafnium, [(CH3)(C2H5)N]4Hf as a precursor for
hafnium (Hf), trimethyl aluminum, and Al(CH3)3 as a precursor for Al, while using
water as an oxidant. In order to fabricate HAOL, we adopt two growth periods, with
each period consisting of 10 cycles of Al2O3 and 30 cycles of HfO2. Immediately after
the deposition, we perform rapid thermal annealing (RTA) in a nitrogen atmosphere.
The RTA is performed for 30 seconds at a temperature of 600°C.
In order to determine the growth per cycle rates of Al2O3 and HfO2 films, we fabricate
the Al2O3 and HfO2 control samples on Si substrates. We use 2×10 cycles to grow Al2O3
and 2×30 cycles to grow HfO2. We use transmission electron microscopy (TEM), as
well as capacitance−voltage (C−V) and current−voltage (I−V) measurements to
characterize the deposited films. The thicknesses of the fabricated Al2O3, HfO2, and
HAOL films are obtained to be 1.54 nm, 7.67 nm, and 9.46 nm, respectively (see Fig.
A.2).

Figure A.2: HAOL gate dielectric. TEM images of (a) the Al2O3 control sample deposited
via 20 ALD cycles, (b) the HfO2 control sample deposited via 60 ALD cycles, and (c)
Al2O3/HfO2 nanolaminate, which served as a gate dielectric in our metasurface. The
Al2O3/HfO2 nanolaminate is grown via ALD. Our ALD process consists of two growth periods,
each including 10 cycles of Al2O3 and 30 cycles of HfO2, followed by a 30-second-long rapid
temperature annealing treatment at 600oC. The inset shows the deposition sequence of the
nanolaminate. The scale bar is 2 nm.

180
As shown in Fig. A.2, while Al2O3 and HfO2 layers are amorphous, the HAOL layer is
partially crystallized after the RTA treatment. The TEM images indicate that, as
expected, there is a thin native oxide layer formed on Si substrates. To enable electrical
characterization of the dielectric films, we deposit HAOL layers on continuous Al
bottom electrodes that were deposited using an e-beam evaporator. We then sputter Al
top electrodes while using shadow masks.
To identify the DC permittivities of the films, we use the C–V measurements of the
fabricated MOS capacitors at 100 kHz. The DC permittivities of the fabricated Al2O3,
HfO2, and HAOL films are obtained to be kAl2O3 = 10.5, kHfO2 = 17.8, and kHAOL = 22,
respectively. By using I–V measurements performed on metal-oxide-metal (MIM)
structures, we identify that the breakdown fields of the fabricated Al2O3, HfO2, and
HAOL films are EAl2O3 = 7.4 MV/cm, EHfO2 = 3.1 MV/cm, and EHAOL = 7.2 MV/cm,
respectively.

Appendix A.3. Alternative Method for Deposition of Al Back Reflector
In order to deposit the Al back reflector, one can also use sputtering. It should be noted
that when depositing the back reflector, the surface roughness of the layer is a key
element that has to be considered. Figure A.3 shows the root mean square (RMS)
roughness of the sputtered Al with different applied power and process pressures
obtained through performing atomic forced microscopy (AFM).

Figure A.3: Surface roughness of the sputtered Al back reflector. RMS roughness of the
sputtered Al as a function of sputtering power and process pressure.

181
As can be seen, when increasing either the sputtering power or the process pressure,
the RMS surface roughness will be increased. As a result, in order to achieve a smooth
Al film via sputtering, small power and process pressure should be used.

Appendix A.4. Fabrication and Characterization of ITO
In order to deposit our ITO films, we use room-temperature RF sputtering. The
deposition pressure and the applied RF power are 3 mTorr and 48 W, respectively. The
plasma is struck by using argon (Ar) gas with a flow rate of 20 sccm. To control the
amount of oxygen deficiency, and hence, the charge carrier concentration of the ITO
layer, we inject argon/oxygen gas (Ar/O2:90/10) with a tunable flow rate [80], [90].
We sputter ITO films on quartz and Si substrates by changing the Ar+O2 flow rate while
keeping the other parameters constant. Using the mentioned deposition parameters, the
deposition rate of ITO is obtained to be about 1.11 nm/minute. Thus, we sputter ITO
for 4.5 minutes to obtain 5 nm-thick ITO films. Then we perform Hall measurements
and spectroscopic ellipsometry on the films deposited on the quartz and silicon
substrates, respectively [91].
After extracting the charge carrier concentration 𝑁𝐼𝑇𝑂 and electron mobility µ of the
ITO films from Hall measurements, we obtain the complex permittivity of the ITO films
via an ellipsometry fit to a single Drude function using Eqs. (2.1) and (2.2). The highfrequency permittivity ε∞, damping rate 𝛤 , and electron effective mass 𝑚∗ are
determined via fitting the Drude model to the measured ellipsometry data. The
electrical and optical constants obtained from Hall measurements and spectroscopic
ellipsometry are listed in Table A.1.
Another important consideration is that when fabricating our dual-gated metasurface,
we deposit the HAOL film on top of the ITO layer. Since the HAOL layer needs to be
RTA-treated at 600 °C for 30 seconds, we need to take the effect of the RTA treatment
on properties of ITO into account. To investigate this effect, we fabricate two identical
ITO samples and perform RTA treatment at 600°C for 30 seconds on one of the samples.
We do Hall measurements and ellipsometry on both samples and compare the results.
As seen in Table A.1, the fitted parameters are in good agreement with the expected
final thicknesses of the films and literature values for the constants (𝛤 = 0.1185 eV,
𝑚∗ = 0.35 𝑚𝑒 and 𝜖∞ = 3.9), which we use to define the dielectric permittivity of ITO

182
in our simulations [80], [92], [93]. We consider the bulk charge carrier concentration
of ITO to be 𝑁𝐼𝑇𝑂 = 3×1020 cm−3 which draws parallels to the plasma frequency of 𝜔𝑝
= 1.0874 eV.
Table A.1: Electrical and Optical parameters obtained from Hall measurements and
spectroscopic ellipsometry for the ITO films deposited using different Ar+O2 flow rates.
Ar+O2
flow
rate
[sccm]

Fitted
thickness
[nm] as
deposited

Fitted
thickness
[nm]
after
RTA

𝜖∞ as
deposited

𝜖∞ after
RTA

𝜔𝑝
[eV] as
deposit
ed

𝜔𝑝
[eV]
after
RTA

𝛾 [eV] as
deposited

𝛾 [eV]
after
RTA

4.3637

4.3137

6.0853

5.8447

1.8516

1.924

0.16245

0.14188

0.4

5.3566

5.1242

6.4603

5.402

1.9679

1.2989

0.14092

0.12521

0.5

5.2988

5.3237

5.1834

4.8832

1.4075

0.16379

0.12981

0.6

4.0852

6.4846

5.338

5.0306

1.4496

0.9440
1.0185

0.15081

0.11095

0.7

5.5826

5.3170

5.9536

4.689±0

1.7932

0.13828

0.1543

0.8

5.4923

5.8453

5.6552

5.1296

1.5872

0.8660
1.1351

0.14384

0.14262

0.9

5.6060

5.5593

5.1672

5.6363

1.285

1.4352

0.14187

0.13105

6.2157

6.0063

5.5049

5.4699

1.4416

1.2529

0.13843

0.12189

When fabricating our metasurface, we deposit ITO at Ar+O2 flow rates of 0.6 sccm. In
this case, the plasma frequency and the charge carrier concentration of ITO after RTA
treatment are 𝜔𝑝 = 1.0185 eV and 𝑁 = 2.6319×1020 cm−3 , respectively. Note that
after depositing top gate dielectric on ITO, the carrier concentration of ITO is expected
to increase due to the leakage of oxygen from the ITO layer into the dielectric that
occurs during the ALD process [94]. As a result, we expect the carrier concentration of
ITO in our final device to be slightly higher than the values we obtain via Hall
measurements.
It should be noted that the smoothness of the ITO films incorporated into our
metasurface is of great importance. To this end, we deposited different test samples
using different sputtering powers and temperatures. To investigate the roughness of
the deposited ITO films, we performed AFM on our ITO samples and figured out that
sputtering at room temperature with applied RF power of 48 W would result in the

183
ITO films with an average RMS roughness of ~ 0.3 nm. This was expected since
having rough and porous ITO films would not let us perform Hall measurements
and/or fit the ellipsometry results with the Drude model.

184
Appendix B.1. Comparison to the Previously Proposed Design
In Chapter 2, a dual-gated Al-based metasurface was used, while, in the design
proposed in Chapter 3, a single-gated Au-based metasurface is presented. In the design
presented in Chapter 2, all metasurface elements were connected and the same bias
voltage was applied to all of them. However, the metasurface presented in the current
chapter operates by individually controlling each metasurface pixel. This requires the
metasurface pixels to be electronically isolated. The requirement of individuallyaddressable metasurface elements makes the single-gated metasurface easier to be
fabricated compared to its dual-gated counterpart.
In the proposed single-gated metasurface, the lower dielectric layer acts as an optical
dielectric, removing the limitation of using HAOL as the lower dielectric layer.
Replacing the HAOL layer by a lower index dielectric (Al2O3) results in a higher
reflectance minimum. In addition, the proposed Au-based metasurface can provide a
smaller amplitude modulation accompanied by a phase shift comparable to the singlegated Al-based metasurface. This supports the operating principle of the
multifunctional metasurface that mainly relies on the phase modulation. Figure B.1
illustrates the amplitude and phase shift of the Al-based and Au-based metasurfaces as
a function of applied biases. As can be seen, the proposed metasurface can provide a
more modest amplitude modulation with an elevated reflectance minimum level
compared to the previous metasurface design.

Figure B.1: Comparison between the single-gated Au-based metasurface and the dualgated Al-based metasurface. Reflectance and phase shift as a function of applied bias for the
new design compared with the previous dual-gated metasurface design.

185
Appendix B.2. Choosing the Number of Metasurface Pixels
In order to implement beam steering, we need to find the number of metasurface
elements that are individually controlled. Figure B.2 shows the efficiency of the beam
steering for different numbers of metasurface elements. In this case, the 4-level phase
profile and a repetition number of RN = 3 are used. As one can see, increasing the
number of metasurface elements will decrease the width of the first diffraction order.
However, increasing the number of metasurface elements will increase the surface
area of the metasurface, leading to a higher probability of the gate dielectric
breakdown due to the existence of pinholes. Besides, increasing the number of
individually-controllable metasurface elements will increase the number of pads to
be wire-bonded, and hence, elevates the complexity of the device fabrication. Here,
we chose to use 96 different metasurface elements to be independently controlled. As
can be seen in Fig. B.2, this number of metasurface pixels would provide a satisfying
beam steering performance.

Figure B.2: Beam steering performance of the multifunctional metasurface for
different numbers of individually-controllable metasurface elements (N). (a) N = 24,
(b) N = 48, (c) N = 72, and (d) N = 96.

Appendix B.3. Changing Metasurface Reflectance Level
Due to the strong field confinement in the ITO layer of the metasurface at ENZ
condition, which is indeed the operation principle of the modulation provided by the
metasurface, these plasmonic active metasurfaces show small reflectance values. It

186
should be noted that in the proposed tunable metasurface, one can always see a tradeoff
between the amplitude of the reflection from the metasurface and the maximum
achievable phase shift.
In this section, we will show how changing the structural parameters of the metasurface
can enhance the reflectance level and the maximum achievable phase shift defined by
Max. phase shift = Acquired phase (V = 6V) − Acquired phase (V = −6V) (A3.1)
Figure B.3 shows the reflectance and phase shift of the reflection from the metasurface
for different thicknesses of the antenna (ranging from 20 nm to 220 nm). As can be
seen, when increasing the thickness of the antenna, the resonance will first blueshift
and then redshift. Moreover, for an increased antenna thickness, the reflectance peak
intensity at zero bias increases while the phase shift decreases. As a result, in order to
obtain a metasurface with a high phase shift, one needs to use shorter antennas.

Figure B.3: Effect of antenna thickness on the optical response of the metasurface. (a)
Unbiased reflectance spectrum, (b) a false-color illustration of the reflectance spectrum, and (c)
phase shift as a function of applied bias for different thicknesses of the nanoantennas.

Spatial

distributions

of

the

intensity

of

the

electric

field

in

the

Al2O3/ITO/heterostructure nanosandwich for three different antenna thicknesses at
applied biases of 0 V and 6 V in the x-z plane are shown in Fig. B.4. As can be seen in
Fig. B.4, increasing the thickness of the antenna will not affect the field concentration
in the active region of the ITO layer. However, investigating the spatial distribution of
the x-component of the electric field (see Fig. B.5) shows that when increasing the
thickness of the nanoantennas (Fig. B.5 e, f), a huge amount of field will be confined
between adjacent antenna elements. That is the reason for which changing the antenna
thickness will influence the reflectance and the phase shift provided by the metasurface.

187

Figure B.4: Effect of antenna thickness on the spatial distribution of the electric field
intensity. Spatial distribution of the amplitude of the electric field of the metasurface with
antenna thickness of (a, b) ta = 20 nm, (c, d) ta = 60 nm, and (e, f) ta = 220 nm, under an applied
bias of (a, c, e) 0 V, and (b, d, f) 6 V in the x-z plane.

Figure B.5: Effect of antenna thickness on the spatial distribution of the x-component of
the electric field. Spatial distribution of the amplitude of the electric field of the metasurface
with antenna thickness of (a, b) ta = 20 nm, (c, d) ta = 60 nm, and (e, f) ta = 220 nm, under an
applied bias of (a, c, e) 0 V, and (b, d, f) 6 V in the x-z plane.

Another important factor that is expected to influence the optical response of the
metasurface is the HAOL layer thickness. Figure B.6 presents the unbiased reflectance
and phase shift spectrum for a metasurface with HAOL thicknesses of 9.5 nm, 19 nm,
and 28.5 nm. As can be seen, when increasing the HAOL thickness, the reflection

188
resonance blueshifts. Besides, an unbiased reflectance increase accompanied by a phase
shift reduction is observed when the thickness of the HAOL increases. As can be seen
in Figs. B.6d, e, at both HAOL thicknesses of 9.5 nm and 28.5 nm, a curved magnetic
field can be observed between the antenna and the back reflector, confirming the
magnetic nature of the resonances.

Figure B.6: Effect of HAOL thickness on the optical response of the metasurface. (a)
Unbiased reflectance spectrum, (b) a false-color illustration of the reflectance spectrum, and (c)
phase shift as a function of applied bias for different thicknesses of the HAOL layer. Spatial
distribution of the y-component of the magnetic field in the x-z plane for the HAOL thickness
of (d) 9.5 nm and (e) 28.5 nm in the absence of the applied bias.

Figure B.7: Effect of HAOL thickness on the spatial distribution of the electric field.
Spatial distribution of the (a-d) amplitude and (e-h) x-component of the electric field of the
metasurface with HAOL thickness of (a, b, e, f) th = 9.5 nm and (c, d, g, h) th = 28.5 nm, under
an applied bias of (a, c, e, g) 0 V, and (b, d, f, h) 6 V in the x-z plane.

189
In order to further investigate the effect of the HAOL thickness, the spatial distributions
of the amplitude, and the x-component of the electric field are illustrated in Fig. B.7. It
can be seen in Fig. B.7 that while the metasurface supports a high field concentration
in the active region of the ITO layer when the HAOL layer is changed, one can observe
a huge amount of confined x-component of the electric field for thinner HAOL layers.
This coupling between the adjacent antenna elements plays an important role in the
tunable optical response of the active metasurface.
Another important factor that plays a crucial role in the tunable optical response of the
metasurface is the intrinsic and structural properties of the ITO layer that serves as the
active medium. Figure B.8 shows the unbiased reflectance spectrum as well as the
phase shift for the metasurface with different ITO thicknesses. As can be seen in Fig.
B.8, increasing the thickness of the ITO layer will result in a blueshift of the reflection
resonance. Furthermore, the reflectance peak value increases and the phase shift
decreases when we increase the ITO thickness.

Figure B.8: Effect of ITO thickness on the optical response of the metasurface. (a)
Unbiased reflectance spectrum, (b) a false-color illustration of the reflectance spectrum, and (c)
phase shift as a function of applied bias for different thicknesses of the ITO layer.

As mentioned before, due to the strong field enhancement in the accumulation layer of
the ITO layer, one expects a large amount of field absorption in this region. As a result,
the imaginary part of the permittivity of ITO is a key element in the obtained optical
response of the metasurface. In Fig. B.9, we investigate imaginary cases for which the
collision frequency of the ITO layer is altered.
As can be seen, increasing the collision frequency of ITO will decrease the unbiased
reflectance value at the resonance wavelength while keeping the obtained phase shift
unchanged. As a result, to increase the reflectance level of the metasurface and hence,

190
the efficiency of the device, one could replace ITO with transparent conducting oxides
with smaller collision frequencies (larger mobilities) such as CdO.
When designing the metasurface unit cell, we used the lower Al2O3 layer as a dielectric
spacer, adding to the degrees of freedom in obtaining the optimal design. As a result, it
is expected that changing the thickness and refractive index of this dielectric spacer
layer can affect the optical response of the metasurface.

Figure B.9: Effect of ITO collision frequency on the optical response of the metasurface.
(a) Unbiased reflectance spectrum, (b) a false-color illustration of the reflectance spectrum, and
(c) phase shift as a function of applied bias for different collision frequencies of an imaginary
ITO layer. The inset in (a) shows a zoomed-in image of the reflectance spectrum around the
resonance wavelength.

Figure B.10 presents the unbiased reflectance spectrum as well as the phase shift for
the metasurface with different thicknesses of the Al2O3 layer.

Figure B.10: Effect of the thickness of the lower dielectric layer on the optical response of
the metasurface. (a) Unbiased reflectance spectrum, (b) a false-color illustration of the
reflectance spectrum, and (c) phase shift as a function of applied bias for different thicknesses
of the lower Al2O3 layer.

191
As can be seen in Fig. B.10, when increasing the thickness of the lower dielectric layer,
the reflection resonance will blueshift, and the reflectance value at the resonance
wavelength will first decrease and then increase. It can also be observed that increasing
the thickness of the lower dielectric layer will cause the phase shift to be first increased
and then decreased.
The effect of the refractive index of the lower dielectric layer is summarized in Fig.
B.11 which shows the reflectance and phase shift of the reflection from the metasurface
for different refractive indices of the lower dielectric layer ranging from 1 to 3. As can
be seen, when increasing the refractive index of the lower dielectric layer, the onresonance unbiased reflectance value first decreases and then increases. Moreover, the
reflection resonance redshifts, accompanied by a phase shift decrement.

Figure B.11: Effect of the refractive index of the lower dielectric layer on the optical
response of the metasurface. (a) Unbiased reflectance spectrum, (b) a false-color illustration
of the reflectance spectrum, and (c) phase shift as a function of applied bias for different
refractive indices of the lower dielectric layer.

As can be seen in Figs. B.10 and B.11, in order to change the phase shift or the
reflectance dip provided by the metasurface, one needs to alter the thickness and /or the
refractive index of the lower dielectric layer.
Another method to alter the reflectance from the metasurface is to cover the antennas
by a dielectric layer. Figure B.12 show the unbiased reflectance and phase shift spectra
from the metasurface for different thicknesses of a SiO2 layer covering the metasurface
area. As can be seen, increasing the thickness of the top SiO2 layer up to 80 nm would
increase the reflectance from the metasurface. However, a 120 nm-thick SiO2 layer will
drastically decrease the maximum achievable phase shift. It should be noted that the
dimensions of the antenna and the electrode are adjusted to la = 210 nm, wa = 115 nm,
and we = 130 nm such that the resonance wavelength of the metasurface, when covered
by an 80 nm-thick SiO2 layer, is the same as that of the original metasurface.

192

Figure B.12: Effect of top dielectric coat on reflectance and maximum achievable phase
shift of the metasurface. (a) Reflectance spectrum, (b) zoomed-in image of the reflectance
spectrum, and (c) maximum achievable phase shift for different top SiO2 thickness values.

Spatial distributions of the amplitude and x-component of the electric field within the
metasurface are presented in Figs. B.13 and B.14, respectively. The local field
distributions are shown for the top SiO2 layer’s thicknesses of 0 and 110 nm, and at the
bias voltages of 0 and 6 V.

Figure B.13: Near-field distribution of the amplitude of the electric field when the
nanoantennas are covered by a SiO2 top coat layer. (a) No top coat and V = 0, (b) 110 nmthick top coat and V = 0, (c) zoomed-in image of the no top coat and V = 0 case in the
Al2O3/ITO/HAOL nanosandwich, (d) zoomed-in image of the 110 nm-thick top coat and V =
0 case in the Al2O3/ITO/HAOL nanosandwich, (e) No top coat and V = 6 V, (f) 110 nm-thick
top coat and V = 6 V, (g) zoomed-in image of the no top coat and V = 6 V case in the
Al2O3/ITO/HAOL nanosandwich, (h) zoomed-in image of the 110 nm-thick top coat and V =
6 V case in the Al2O3/ITO/HAOL nanosandwich.

As can be seen in Figs. B.13 and B.14, covering the metasurface with a SiO2 top coat
does not affect the field concentration in the active region of ITO. However, when
adding the top coat layer, a higher refractive index of the top coat compared to the air
causes a high amount of field to be confined in the regions between adjacent antennas.

193
The constructive coupling between the adjacent antennas can lead to an increased
unbiased on-resonance reflectance for some SiO2 thicknesses.

Figure B.14: Near-field distribution of the x-component of the electric field when the
nanoantennas are covered by an SiO2 top coat layer. (a) No top coat and V = 0, (b) 110 nmthick top coat and V = 0, (c) No top coat and V = 6 V, (d) 110 nm-thick top coat and V = 6 V.

Appendix B.4. PCB Layout for Demonstration of Dynamic Beam Steering and
Reconfigurable Focusing Using the Multifunctional Metasurface.
In order to individually bias each of 96 different metasurface elements, we designed
two PCBs as shown in Figs. 3.14c, d. Figure B.15 shows the schematic layout of the
sample-mounting PCB. The metasurface sample is mounted on this PCB (outlined by
the rectangle denoted by star-shaped polygon), and 96 metasurface elements’ pads, as
well as the ITO pads (to be used as the ground), are wire-bonded from the sample to
100 pads located around the sample on the PCB. Each of these pads is connected to a
pin of 8-pin header connectors (denoted by diamond).

194

Figure B.15: Schematic layout of the sample-mounting PCB used to demonstrate dynamic
beam steering and reconfigurable focusing using the multifunctional metasurface.
Rectangles denoted by star-shaped polygon and diamond outline the metasurface sample and
8-pin connector headers, respectively.

Each 8-pin header on the sample-mounting PCB is then connected to one corresponding
8-pin header connector on the voltage-deriving PCB shown in Fig. B.16. The
independent bias voltage provided at each pin is produced by programming 8-bit digital
to analog converters (DACs) denoted by white triangles in Fig. B.16. Every set of three
DACs is programmed by an Arduino Nano microcontroller board based on the
ATmega328P (Arduino Nano 3.x) denoted by the white hexagon in Fig. B.16. In order
to provide the desired voltages at the output ports of the DACs, the input ports of the
DACs are connected to the digital outputs of the Arduino microcontrollers and are then
programmed via computer by using the Arduino Software (IDE). Each analog output
of the DACs is then connected to one pin of the 8-pin header connectors, denoted by
white circles in Fig. B.16.

195

Figure B.16: Schematic layout of the voltage-deriving PCB used to demonstrate dynamic
beam steering and reconfigurable focusing using the multifunctional metasurface. The
white hexagon, triangle, and circle indicate the Arduino Nano, the DAC, and the 8-pin header,
respectively.

Appendix B.5. Pattern Layouts for Fabrication of the Multifunctional
Metasurface

196

Fabrication steps of the TCO-based multifunctional metasurface are presented in Fig.
3.9. As mentioned in Chapter 3, in order to fabricate the universal metasurface, first the
outermost parts of the connection pads are patterned using photolithography. To
expedite the fabrication process, a 4” SiO2 (1 m) on Si wafer is patterned in order to
provide 7 samples. Figure B.17 shows the layout of the photomask used for this process.
As can be seen in Fig. B.17a, 96 connecting pads as well as some alignment markers
are patterned for 7 samples with sizes of 26 mm  26 mm. A zoomed-in view of a single
sample’s contact pads with alignment markers at the corners is depicted in Fig. B.17b.

Figure B.17: Layout of the photomask used for patterning the contact pads of the
multifunctional metasurface. (a) Photomask used for simultaneously patterning the contact
pads and the alignment markers of 7 samples. (b) Zoomed-in view of the pattern used for each
sample.

After patterning the outermost pads and the alignment markers on the 4” wafer, the
resist is developed and a 10 nm-thick Ti layer followed by a 200 nm-thick Au layer is
deposited on the samples using an electron beam evaporator. After the lift-off process,
the back reflectors of the 7 samples are patterned using EBL and employing the
alignment markers obtained through the photolithography step. The size of each back
reflector is 2.5 mm  2.5 mm. After developing the EBR, a 3 nm-thick Cr layer followed
by an 80 nm-thick Au layer is deposited on the samples using an electron beam
evaporator. Then a 9.5 nm-thick Al2O3 layer is deposited on the samples using a 4”
shadow mask presented in Fig. B18.

197

Figure B.18: Layout of the shadow mask used for patterning the Al2O3 layer of the
multifunctional metasurface. 7 square holes with dimensions of 3.5 mm  3.5 mm are used
to deposit Al2O3 on the 7 samples located 26 mm apart on a 4” wafer.

Once the Al2O3 layer is deposited on the samples, we pattern 7 squares with sizes of
500 m  500 m to be used as the ITO layer of our samples using EBL and employing
the alignment markers. After developing the exposed EBR, the wafer is diced to the 7
separated pieces. We then sputter ITO layers with different Ar/O2 ratios on different
samples using the recipe described in Appendix A.4. Afterward, the contact pads of the
ITO layer are patterned on each sample via EBL using the layout illustrated in Fig. B19.

Figure B.19: Layout of the EBL pattern used for patterning the ITO contact pads of the
multifunctional metasurface. 4 pads are patterned in order to apply bias to the ITO layer of
the multifunctional metasurface sample.

After developing the EBR, a 10 nm-thick Ti layer followed by a 200 nm-thick Au layer

198
is deposited on the samples using an electron beam evaporator. Once the excess resist
and films are lifted off, we deposit the HAOL layer on each sample using the shadow
mask presented in Fig. B20, and following the recipe described in Appendix A.2. It
should be noted that the outer dimensions of the shadow masks are picked to be slightly
larger than the sample size (i.e. 28 mm  28 mm) to prevent deposition of the dielectric
on the contact pads of the chips from the edges.

Figure B.20: Layout of the shadow mask used for patterning the HAOL layer of the
multifunctional metasurface. A 3 mm  3 mm hole is milled through a 28 mm  28 mm
stainless steel mask.

In the next step, the antennas as well as the inner contact lines are patterned on each
sample using EBL. In order to minimize the write time, the connection lines are
patterned in three steps to enable the beam current adjustments. Figure B.21 presents
the patterns used for the e-beam lithography of the antennas and the inner connection
lines.

199

Figure B.21: Layout of the EBL pattern used for the antennas and the inner connection
lines of the multifunctional metasurface. Layout of (a) antennas and inner connection lines,
(b) zoomed-in view of the antennas and the innermost connection lines, and (c) zoomed-in view
of the antenna patterns.

After developing the exposed EBR, a 2 nm-thick Ge layer followed by a 40 nm-thick
Au layer is deposited on the samples using an electron beam evaporator. Once the liftoff process is done, the final sample is obtained as depicted in Fig. B.22.

Figure B.22: Schematic illustration of the layout of the final multifunctional metasurface.

200
Appendix C.1. Measuring the Breakdown Field of HAOL Films
In order to measure the breakdown field of the HAOL films, we use the configuration
presented in Fig. C.1. To fabricate this structure, first, a metallic bottom-contact with
the length, width, and thickness of 9.5 mm, 500 m, and 80 nm, respectively is
deposited on a SiO2 (1 m) on Si substrates using EBL an e-beam evaporator. On top
of the bottom-contact, a 6 mm  6 mm HAOL layer is deposited through shadow masks
via ALD. Then, two top-contacts with the same dimensions as the bottom-contact, but
in a direction perpendicular to it, are deposited.

Figure C.1: Configuration of the MIM structure used to measure the breakdown field on
the HAOL film. Au and Al were used as the contact pads in two distinct samples. The inset
shows a photograph of the fabricated MIM structure.

Performing I-V measurements on the MIM structures fabricated when using Au and Al
as the contact pads, the breakdown fields of the HAOL films were obtained as
summarized in Table C.1. As can be seen, when Au is used as the contact pad, a higher
breakdown field of the HAOL film is obtained. That is why we use Au-based
metasurfaces in order to experimentally demonstrate active polarization conversion.
Table C.1: Measured breakdown field of the fabricated HAOL film for demonstration of
active polarization conversion metasurface.
Metal contacts
material

V+ (I = 40
μA/cm2)

Positive
breakdown field
(MV/cm)

V- (I = 40
μA/cm2)

Au
Al

5.8 V
5.0 V

6.4
5.5

−4.8 V
−5.2 V

Negative
breakdown
field
(MV/cm)
−5.3
−5.7

201
Appendix D.1. Convergence Test for Simulation Region Size
In order to make sure that the size of the FDTD simulation region is large enough, we
performed a convergence test using FDTD Lumerical. Figure D.1 shows the Purcell
enhancement spectrum for different simulation region sizes along the x-, y-, and zdirections. These simulations confirmed that for the simulation regions larger than 1000
nm, the Purcell enhancement results converged. As a result, the size of our simulation
region was picked to be 1600 nm, 1600 nm, and 2800 nm along the x-, y-, and zdirections, respectively.

Figure D.1: Convergence test for the FDTD simulation region size. Purcell enhancement
spectra for different simulation region sizes along the (a) x- and y-, and (b) z-direction.

Appendix D.2. Experimental Demonstration of Tunable Purcell Enhancement of
Spontaneous Emission Using Active Metasurfaces
In order to experimentally investigate the possibility of modulation of spontaneous
emission using reconfigurable metasurfaces, we fabricated a TCO-based active
metasurface with quantum emitters embedded in. To this end, we first needed to find
the best deposition recipe by which we could obtain Er-doped alumina with the desired
concentration of the Er ions. Different co-sputtering recipes have been tried to obtain
Er-doped alumina layers. In these co-sputtering processes, two RF powers were
simultaneously applied to the Er and Al sputtering targets in an Ar and O2 gas
environment. In order to characterize the co-sputtered films, we first performed
spectroscopy ellipsometry to find the thickness of the Er-doped alumina films cosputtered on Si substrates. Table D.1 summarizes the fitted thickness of the films cosputtered in different Ar and O2 gas ratios and at different deposition temperatures. It
should be noted that we co-sputter Er-doped Al2O3 at high temperature (550 oC) to have

202
active Er3+ ions. As can be seen in Table D.1, changing the Ar and O2 gasses flow rates
and the deposition temperature does not affect the deposition rate of the Er-doped
alumina layer.

Table D.1: Spectroscopic ellipsometry results of the Er-doped alumina films co-sputtered
on Si substrates in different Ar and O2 gas flow rates at different deposition temperatures.
Sample

Power
applied
to Al
target
(W)

Power
applied
to Er
target
(W)

Ar gas
flow
rate
(sccm)

O2 gas
flow
rate
(sccm)

200
200
200

30
30
30

20
20
20

0.5

Ar + O2
(Ar/O2 :
90/10)
gas
flow
rate
(sccm)
10

Deposition Deposition
temperature rate
(oC)
(nm/min)

RT
550
RT

3.75
3.75
3.75

We also noticed that the deposition rate and the refractive index of the Er-doped
alumina film slightly decrease at higher processing pressures. Moreover, a slight
decrease of the refractive index was found at higher temperatures
Once we obtained the deposition rate and refractive index of the co-sputtered Er-doped
alumina films, we used Rutherford Backscattering Spectrometry (RBS) to find the
atomic concentration of different components in our films. In an RBS measurement,
high-energy (MeV) He2+ions (alpha particles) are directed onto the sample, and the
energy distribution and yield of the backscattered He2+ ions at a given angle
are measured.
Table D.2 presents the RBS results for different samples co-sputtered in different Ar
and O2 gas ratios. As can be seen in Table D.2, changing the deposition temperature
can have a negligible effect on the atomic concentration of different components. On
the other hand, changing the Ar and O2 gasses flow rate ratio can change the
decomposition of the co-sputtered Er-doped alumina films. This can also be seen in the
concentration profile of different components as a function of depth as plotted in Fig.
D.2.

203
Table D.2: RBS results of the Er-doped alumina films co-sputtered on Si substrates in
different Ar and O2 gas flow rates at different deposition temperatures.
Sample

RBS
thickness
(Å)
Film
Bulk
Film
Bulk
Film
Bulk

2250
2200
1980

Atomic concentration (at %)

Si
100
100
100

Al
39.0
39.7
40.1

60.2
59.5
59.2

Er
0.5
0.5
0.4

Assumed
density
(at/cc)
Ar
0.3
0.3
0.3

6.66 e22
5.00 e22
6.65 e22
5.00 e22
6.65 e22
5.00 e22

Figure D.2: Atomic concentration of different components of Er-doped alumina films
obtained from RBS. Concentration of different components as a function of depth for (a)
Sample #1, (b) Sample #2, and (c) Sample #3 as described in Table D.1.

In order to further investigate the effect of different sputtering parameters, we studied
the effect of the power applied to the Er and Al targets. Our measurements showed that
the deposition rate is mainly influenced by the sputtering power on the Al target, and
using RF power supply leads to a 2–4 times lower deposition rate compared to the DC

204
power supply. Moreover, we noticed that the erbium concentration is mainly influenced
by the sputtering power applied to the Er target as shown in Fig. D.3. As can be seen in
Fig. D.3, by changing the power applied to the Er target, the desired Er concentration
of 0.1-0.5 at % can be achieved.

Figure D.3: Effect of RF power applied to the Er target. Er concentration in Er-doped
alumina films as a function of applied RF power, obtained from RBS.

Once we obtained the optimum recipe leading to the desired Er-doped alumina layer,
we fabricated the whole metasurface structure to investigate gate-tunable spontaneous
emission enhancement of the quantum emitters. Figure D.4a shows the schematic of
the metasurface designed to be fabricated. The designed metasurface consists of a 75
nm-thick Ag back reflector. As discussed above, in order to obtain active Er3+ ions, the
Er-doped Al2O3 layer is supposed to be co-sputtered at high temperature (550 oC). As a
result, to prevent heating the back reflector, we first deposited a 6 nm-thick undoped
Al2O3 layer at room temperature, and then we deposited Er-doped alumina at 550 oC on
top of it. This layer was followed by a 7 nm-thick ITO layer with a bulk carrier
concentration of 11020. Then, a 5 nm-thick HfO2 gate dielectric was deposited using
ALD, followed by the Ag fishbone antennas. In order for the metasurface to show a
resonance at the emission wavelength of the Er3+ ions (res = 1535 nm), the antenna
length, antenna width, and the electrode width were chosen to be 260 nm, 160 nm, and
200 nm, respectively. Figure D.4b shows the Purcell enhancement provided by the
metasurface for different applied biases.

205

Figure D.4: Ag-based gate-tunable metasurface designed for demonstration of tunable
spontaneous emission enhancement. (a) Schematic of the designed metasurface, and (b)
Purcell enhancement spectrum for different applied biases.

Once we fabricated the metasurface, the photoluminescence intensity from the emitters
embedded within the metasurface was measured using the optical setup presented in
Fig. D.5.

Figure D.5: Optical setup used for measuring the voltage-tunable Purcell enhancement of
quantum emitters embedded in active metasurfaces.

In this setup, an Ar laser at 488nm was used for excitation. The laser light was focused
on an Acousto-optic modulator (AOM) which was controlled by a pulse generator. The
light was then directed to a vacuum sample holder on which our metasurface was

206
mounted. The luminescence from the sample was collimated and then focused on an IR
photomultiplier tube (PMT). The data read by the PMT was interpreted by voltage in a
lock-in amplifier and then shown on the computer.
Figure D.6a shows an SEM image of the fabricated metasurface with a zoomed-in view
of the nanoantennas shown in Fig. D.6b. The measured PL intensity spectra for different
applied biases are presented in Fig. D.6c. Figure D.6d illustrates the integrated
measured intensity as a function of the applied voltage. As can be seen, when changing
the applied bias, the PL intensity and the integrated intensity could be tuned.

Figure D.6: Tunable metasurface used for measuring the voltage tunable Purcell
enhancement. SEM image of the (a) nanoantennas connected to external pads for bias
application and (b) zoomed-in image of the nanoantennas. The scale bar is (a) 50 m and (b) 1
m. (c) The measured PL intensity spectrum for different applied biases. (d) Integrated
measured PL intensity as a function of applied bias.

When repeating the measurements, we noticed that the voltage-tunable PL
measurement was not reproducible over the course of a few days while the gate
dielectric and the ITO layer had kept their original properties. This can trace its roots
in the migration of Ag ions through the Al2O3 layer and the fast degradation of Ag
nanoantennas. As a result, the voltage-tunable PL intensity is more likely interpreted to
be obtained as a result of ion migration instead of field-effect modulation of the ITO
layer.

207
Appendix E.1. PCB Design for Demonstration of Dynamic Beam Steering
Figure E.1 shows the schematic layout of the PCB on which the MQW metasurface
sample is mounted. The sample is mounted on the vertical rectangle space (denoted by
star-shaped polygon). Then 63 pads on the sample as well as the bottom contact (to be
used as the ground) are wire-bonded to 64 pads placed on the PCB around the sample.
Each of the mentioned pads is connected to one pin of an 8-pin box header drawn by
vertical white rectangles in Fig. E.1 (denoted by diamond). Then, each connector header
on the sample-mounting PCB is connected to an 8-pin header on the voltage-deriving
board shown in Fig. E.2. Depending on the RN value used to change the effective period
of the metasurface, one configuration of voltages (marked by RN# on the board) is
selected.

Figure E.1: Schematic layout of the sample-mounting PCB used to demonstrate the MQW
beam steering metasurface. Rectangles denoted by star-shaped polygon and diamond outline
the metasurface sample and 8-pin connector headers, respectively.

208

Figure E.2: Schematic layout of the voltage-deriving PCB used to demonstrate the MQW
beam steering metasurface. Depending on the required RN value, the 8-pin headers on the
sample-mounting PCB were connected to one column of the 8-pin headers located on the
voltage-deriving PCB.