ABSTRACT
Title of Dissertation: QUANTUM LIGHT GENERATION
FROM BOUND EXCITONS IN ZNSE
Aziz Karasahin
Doctor of Philosophy, 2022
Dissertation Directed by: Professor Edo Waks
Department of Electrical and Computer Engineering
Quantum light sources and spin based qubits are essential building blocks for on-
chip scalable quantum computation and information processing. To achieve scalability,
information-storing qubits should exhibit long coherence times. These qubits should also
be efficiently interfaced with information-carrying single photons. Semiconductors are
not only able to host such qubits and single photon sources but also they offer a platform
to interface them with the help of photonic structures. Hence, optically active solid-
state qubits such as quantum dots, crystal defects and color centers have been extensively
studied to date in various semiconductors. However, we still lack of a suitable platform
to satisfy all the requirements needed to realize a scalable quantum technology.
Impurities in epitaxially grown ZnSe are particularly promising single photon sources
and qubit candidates due to the direct bandgap of the material and potential for isotopic
purification to achieve nuclear spin-zero background. These impurities possess impurity-
bound electrons that can serve as spin-qubit. They also form impurity-bound excitons
that can generate single photons. Various impurities have been studied in ZnSe, but only
F impurities have been isolated as single emitters to date. Despite the great potential
suggested by previous results, there are many impurities waiting to be explored for their
quantum capabilities.
In this thesis, we study isolated Cl impurities in ZnSe for their photon emission
and spin properties. We utilize a ZnMgSe/ZnSe/ZnMgSe quantum well to increase the
binding energies bound excitons and to better separate donor bound exciton emission
from the free excitons. In the PL spectrum, we observe narrow emission lines around
440 nm, which are originated from the single bound excitons. We calculate the average
binding energy as 15 meV (at least 2 times higher than bulk values) and inhomogeneous
broadening as 6 meV. We confirm the single photon emission by observing clear photon
antibunching in the second order autocorrelation measurements. The time-resolved photo-
luminescence measurements show short radiative lifetimes of 192 ps. Our results demonstrate
first time isolation of donor impurities in an unstructured ZnSe, and provides complete
characterization of radiative properties single Cl bound excitons.
The bound electron of a donor impurity atom can serve as a spin qubit. We verify
that the presence of ground state electron of the Cl donor complex by observing two-
electron satellite emission. We also characterize the Zeeman splitting of the exciton
transitions by performing polarization-resolved magnetic spectroscopy on the single emitters.
We also discover the presence of single biexcitons bound to Cl impurities. We
demonstrate a radiative cascade from the decay of bound biexcitons. The emission exhibits
both single photon statistics and clear temporal correlations revealing the time?ordering
of the cascade.
Finally, we discuss the design of nanophotonic cavities in the ZnSe platform. We
develop a nanofabrication recipe to create suspended photonic crystal cavities. Then, we
optically characterize the fabricated cavities.
The results presented in this thesis provide the first complete study of single Cl
impurities in ZnSe. Based on the results discussed, single Cl impurities in ZnSe manifest
themselves as promising quantum light sources and appealing solid-state qubit candidates.
QUANTUM LIGHT GENERATION FROM
BOUND EXCITONS IN ZNSE
by
Aziz Karasahin
Dissertation submitted to the Faculty of the Graduate School of the
University of Maryland, College Park in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
2022
Advisory Committee:
Professor Edo Waks, Chair & Advisor
Professor Cheng Gong
Professor Avik Dutt
Professor Rajarshi Roy
Professor Oded Rabin, Dean?s Representative
? Copyright by
Aziz Karasahin
2022
Dedication
To my better half, my dear Hatice
ii
Acknowledgments
All praise be to God, the Lord of worlds, for his countless graces and blessings,
endowing me with faith, life, patience and the will to pursue the science in the name of
him.
My most special thanks go to my beloved wife, Hatice, for her continued support at
every moment, her patience in moments of struggle and her the most beautiful heart. She
has always made me smile, even in the most difficult times. Her unconditional love has
enlightened my days and nights, and made our home a lovely nest.
This thesis would not be possible without many supervisors, colleagues and peers.
I would like express my gratitude to my advisor, Edo Waks, for giving me an invaluable
opportunity to work on challenging projects in his group. His scientific approach and
breadth of ideas have made this thesis possible. I am grateful to Dr. Robert M. Pettit
who joined the group at times that I had been struggling with experiments. Thanks to his
experience, endless patience and scientific curiosity, this study gave fruits.
I would like to thank Professor Avik Dutt, Professor Cheng Gong, Professor Rajarshi
Roy and Professor Oded Rabin for agreeing to serve on my thesis committee and for
sparing their invaluable time reviewing the manuscript. I would like to also thank Professor
Mohammad Hafezi for serving in my proposal committee and our exciting collaboration
in the first half of my PhD. He is one the best scientists that I have ever met.
iii
All samples studied in this thesis are supplied by our colleagues, Professor Alexander
Pawlis and his team at Ju?lich, Marvin Marco Jansen, Nils von den Driesch and Yurii
Kotovyi. I thank them for the incredible collaboration.
My friends and peers at the Quantum Nanophotonics Group have enriched my
graduate life in many ways and deserve a special mention. I would like to thank Sabyasachi
Barik for answering my endless questions always with enthusiasm. I learnt many experimental
skills and optical alignment tricks from him. He was not only incredible lab mate but also
a good friend outside of the lab. I did not know that Zhouchen Luo would be the best
office mate when I met him as first person in the campus. His dedication has always made
me amazed. I had many valuable interactions with other group members, to name some,
Subhojit Dutta, Mustafa Buyukkaya, Dr. Changmin Lee, Harjot Singh, Shahriar Aghahei,
Dr. Je-Hyung Kim, Uday Saha, Dr. Dima Farfurnik, Yuqi Zhao, Sam Harper, Tao Cai,
Shuo Sun and many others. They all made the IREAP home far away from my country. I
will always remember my beautiful memories with them.
I am grateful for the support from the staff in the Maryland NanoCenter: Tom
Loughran, Mark Lecates, Jon Hummel and John Abrahams, Jiancun Rao and Sz-Chian
Liou. I especially thank Valery Ray for teaching me how to use SEM and FIB. He
is not only a great expert but also a humble teacher. I would like to also thank all
the staff members in IREAP, including Don Schmadel, Nolan Ballew and Bryan Quinn
for their technical support that enabled complex experiments. I would like to thank
IREAP administrative office staff, including Nancy Boone, Judi Cohn Gorski, Dorothea
Brosius and Leslie Delabar for their great assistance with purchases, payrolls and all other
administrative works.
iv
I would like to acknowledge financial support from the Air Force Office of Scientific
Research (AFOSR), Maryland- ARL Quantum Partnership, National Science Foundation
(NSF), Physics Frontier Center (PFC) at the Joint Quantum Institute (JQI) and German
Research Fund (DFG) for all the projects discussed herein.
Lastly, I would like to thank my mom, dad, sister for being there for these long
years. I have been always fortunate to have parents that supported me since my childhood
and respected my choices. I am grateful for my both grandmothers?s prayers which gave
me the power to finish this journey. I hope this thesis and my success will relieve their
longing eyes.
v
Table of Contents
Dedication ii
Acknowledgements iii
Table of Contents vi
List of Figures viii
List of Abbreviations xi
Chapter 1: Introduction 1
1.1 Quantum Computing and Information Processing . . . . . . . . . . . . . 1
1.2 Some Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Motivation for This Study . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 Thesis Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Chapter 2: Bound Excitons in ZnSe 7
2.1 Crystal Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2 Donor Bound Exciton State . . . . . . . . . . . . . . . . . . . . . . . . . 9
Chapter 3: Material Growth 12
3.1 MBE Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1.1 Delta Doping of Cl Impurities . . . . . . . . . . . . . . . . . . . 14
3.2 The Effect of Quantum Confinement . . . . . . . . . . . . . . . . . . . . 14
Chapter 4: Radiative Properties of Single Bound Excitons 16
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4.2 Photoluminescence Characterization . . . . . . . . . . . . . . . . . . . . 17
4.2.1 Measurement Details . . . . . . . . . . . . . . . . . . . . . . . . 17
4.2.2 Photoluminescence Spectrum . . . . . . . . . . . . . . . . . . . . 18
4.2.3 Binding Energies & Inhomogeneous Broadening . . . . . . . . . 23
4.2.4 Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.3 PLE Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4.4 Lifetime of Emitters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.5 Single Photon Generation . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.6 Temperature Tuning of the Emission . . . . . . . . . . . . . . . . . . . . 37
4.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
vi
Chapter 5: Spin States of Bound Electrons 41
5.1 Two-Electron Satellite Emission . . . . . . . . . . . . . . . . . . . . . . 42
5.2 Donor Electron Energy States under Magnetic Field . . . . . . . . . . . . 45
5.2.1 Energy Splitting . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
5.2.2 Optical Selection Rules . . . . . . . . . . . . . . . . . . . . . . . 48
5.3 Magnetospectroscopy Experiments . . . . . . . . . . . . . . . . . . . . . 50
5.3.1 Measurement Setup . . . . . . . . . . . . . . . . . . . . . . . . . 50
5.3.2 Faraday Magnetic Field . . . . . . . . . . . . . . . . . . . . . . . 51
5.3.3 Voigt Magnetic Field . . . . . . . . . . . . . . . . . . . . . . . . 54
5.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Chapter 6: Cascaded Single Photon Emission from Biexcitons 55
6.1 Radiative Cascades from Biexcitonic Decay . . . . . . . . . . . . . . . . 56
6.2 Experiment Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
6.3 Demonstration of Cascaded Single Photon Emission . . . . . . . . . . . . 59
6.3.1 Characteristics of PL Response . . . . . . . . . . . . . . . . . . . 59
6.3.2 Time Resolved Spectroscopy Results . . . . . . . . . . . . . . . . 61
6.3.3 Photon Correlation Measurements . . . . . . . . . . . . . . . . . 64
6.4 Fine Structure Splitting and Polarization Response . . . . . . . . . . . . . 67
6.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Chapter 7: Design and Fabrication of Nanophotonic Cavities 73
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
7.2 Design of L3 Photonic Crystal Cavities . . . . . . . . . . . . . . . . . . . 75
7.3 Fabrication of Cavities . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
7.4 Optical Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . 81
7.5 Discussions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Chapter 8: Conclusion and Outlook 85
Appendix A: Fabrication Details 89
A.1 Nanofabrication Recipe for ZnSe Photonic Crystals . . . . . . . . . . . . 89
Appendix B: Other Research Contributions 93
B.1 Demonstration of Robust Quantum Photonic Devices with Topological
Edge States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
B.2 Integration of Quantum Emitters with Telecom Photonics . . . . . . . . . 94
Appendix C: List of Publications and Conference Talks 97
Bibliography 99
Bibliography 99
vii
List of Figures
2.1 Crystal structure of ZnSe with a single Cl impurity substituting a Se atom. 9
2.2 Energy band diagram showing the ground state (D0) and excited state
(D0X), along with the bound exciton and two-electron-satellite emission,
shown as blue and red arrow, respectively. . . . . . . . . . . . . . . . . . 10
3.1 Layer structure of the grown material with in-situ doped Cl impurities in
the ZnSe quantum well. . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4.1 Photoluminescence response obtained with the above band excitation. a)
Schematic illustrating the excitation and recombination mechanisms of
the observed FX and D0X lines. b) PL spectrum demonstrating free
exciton (FX), trion (X?) and bound exciton (D0X) emission. . . . . . . . 20
4.2 a) Power dependent intensity of bound exciton (blue) and free exciton
(red), showing saturation of bound exciton lines. b) Power dependent
intensity plotted in log scale for low powers up to 20 ?W , showing linear
dependence for both free exciton and bound exciton emission. . . . . . . 22
4.3 Linewidth of bound exciton line, extracted from Lorentzian curve fitting,
showing a linear dependence on the excitation power. Extrapolation to
zero power shows significant homogeneous broadening, that could be due
to charge instabilities and temperature induced effects. . . . . . . . . . . 24
4.4 Histogram showing the distribution of binding energies of D0X lines, with
an average binding energy of 15 meV and a standard deviation of 6 meV. 26
4.5 PL intensity recorded over 7 hrs showing the stability of the emitter. . . . 27
4.6 Photoluminescence excitation a) Schematic illustration showing how resonant
excitation of FX line creates bound excitons. b) PLE spectrum of bound
exciton line obtained by tuning the pump laser over the FX line. . . . . . 29
4.7 Time resolved lifetime measurement setup a) The measurement configuration
to characterize the timing response of the measurement system. b) The
instrument response function (IRF) showing the limited time resolution
caused by detector timing jitter. . . . . . . . . . . . . . . . . . . . . . . 30
4.8 Time-resolved photoluminescence measurement, showing biexponential
decay with dominant fast decay component of 192 ps and slow decay
component of 2.49 ns. . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
viii
4.9 The schematic of the setup to measure the autocorrelation function from
the D0X line. First, the collected light is filtered with the monochromator,
then photons originated from bound exciton emission are sent to the identical
detectors. We enclose the detectors to avoid the room light contribution. . 34
4.10 a) Coincidence counts of two detectors with variable time delay, obtained
with 60 hours of integration, showing emitted photons are single. b) Noise
corrected second-order auto-correlation function obtained from the single
line, demonstrating almost perfect photon antibunching. . . . . . . . . . 36
4.11 Temperature tuning of the emission. a) PL spectrum obtained at increasing
sample temperatures up to 70K. b) The relative shift of emission energies
of FX and D0X lines from the 3.6K values, showing a similar temperature
dependence for both curves. c) Temperature dependent linewidth showing
that significant broadening starts at 60K . . . . . . . . . . . . . . . . . . 39
5.1 Spectrums showing the TES line with increasing resonant pump. power . 43
5.2 Background corrected and integrated counts from dashed spectral window
showing saturation of TES line with increasing pump power. . . . . . . . 44
5.3 Energy levels and spin configurations of the (ground) bound electron and
(excited) bound exciton states. . . . . . . . . . . . . . . . . . . . . . . . 46
5.4 Zeeman splittings of ground and excited states of bound exciton, along
with optical transitions under Faraday magnetic field. . . . . . . . . . . . 49
5.5 Zeeman splittings of ground and excited states of bound exciton, along
with optical transitions under Voigt magnetic field. . . . . . . . . . . . . 50
5.6 Optical setup to measure PL under the applied magnetic field. a and
b) Schematic demonstrating the configuration of the insert (composed of
XYZ stage, the sample, and the collection lens) for Faraday (Voigt) field
measurements. c) The picture of the Attocube sample insert. . . . . . . . 52
5.7 Polarization resolved photoluminescence spectrum recorded at increasing
magnetic field intensity, showing the splitting of orthogonal circular polarizations. 53
6.1 Schematic of the optical setup used to investigate optical properties of
exciton-biexciton complexes. . . . . . . . . . . . . . . . . . . . . . . . . 58
6.2 Photoluminescence spectrum measured with above band continous wave
laser excitation, showing free exciton (FX), bound exciton (D? X, D?A BX)
and biexciton (D? XX, D?A BXX) states . . . . . . . . . . . . . . . . . . . 60
6.3 a) Power dependent intensity for exciton and biexciton lines b) Energy
levels showing exciton and biexciton states c) Time resolved PL intensity
showing delayed emission from exciton lines in the presence of biexcitonic
emission. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
6.4 PL intensity recorded for 10 minutes showing spectral wandering and
intensity correlations between exciton-biexciton lines, for A (top) and B
lines (bottom). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
ix
6.5 Single photon correlation measurements. (a,b) Autocorrelation measurements
of (a) the localized exciton line, (2)gX (?), and (b) the localized biexciton
line, (2)gXX(?). (c,d) Cross?correlation measurements between the localized
exciton and biexciton lines, (2)gX,XX(?), with (c) pulsed excitation and (d)
continuous wave excitation above the quantum well barrier. . . . . . . . 65
6.6 Polarization resolved spectroscopy. (a) Measured polarized fine structure
of the D? X and D?A AXX lines. (b) Polarization analysis of the D
?
AX and
D?AXX fine structure as a funciton of quarter wave plate angle. . . . . . . 68
7.1 Schematic of L3 cavity. Three missing holes as linear line at the center,
labeled as c, forms the cavity in the triangular lattice. The geometrical
parameters optimized for the design has been labeled as lattice constant
(L), radius (r), displacements (a,b,d). . . . . . . . . . . . . . . . . . . . 75
7.2 FDTD simulation results. a) Electric field intensity distribution of the
fundamental mode at the resonance b) Far field radiation pattern of high-
Q design c) Far field radiation pattern of lower-Q design . . . . . . . . . 77
7.3 Nanofabrication steps to create suspended photonic crystal cavities . . . . 78
7.4 SEM images of fabricated photonic crystal cavities. a) Cross sectional
view of the etched lines, showing the directionality of the dry etch process.
b) Top view showing the triangular lattice with 3 missing holes, forming
L3 cavity. Inset shows the close up image of the cavity region. Rectangular
holes around the structure are etched to facilitate undercutting of GaAs
to make structures suspended. c) Angled view image of the fabricated
device, showing that substrate is not etched completely. . . . . . . . . . . 80
7.5 Layer structure of the fabricated devices. Left image shows the the sample
in-use. Lack of sacrificial layer resulted in non-selective etching of GaAs.
Right image shows the proposed structure with AlGaAs sacrificial layer. . 82
7.6 The measured spectrum of the sample. Blue line shows PL spectrum from
unpatterned bulk region. The orange curves demonstrates cavity mode
with q-factor of 110. The low q-factor is caused by radiative losses, due
to non-perfect undercutting. . . . . . . . . . . . . . . . . . . . . . . . . 84
x
List of Abbreviations
AlAs Aluminum Arsenide
AlGaAs Aluminum Gallium Arsenide
CCD Charged Coupled Device
Cl Chlorine
CW Continous Wave
EBL Electron Beam Lithography
F Flourine
FDTD Finite Difference Time Domain
FWHM Full Width Half Maximum
FX Free Exciton
GaAs Gallium Arsenide
HH Heavy-hole
ICP-RIE Inductively Coupled Plasma-Reactive Ion Etching
InAs Indium Arsenide
InP Indium Phosphide
LL Light-hole
MBE Molecular Beam Epitaxy
NA Numerical Aperture
PC Photonic Crystal
PECVD Plasma Enhanced Chemical Vapor Deposition
PIC Photonic Integrated Circuit
PL Photoluminescence
PLE Photoluminescence Excitation
QD Quantum Dot
QIP Quantum Information Processing
QW Quantum Well
Se Selenium
TES Two-Electron Satellite
ZnSe Zinc Selenide
ZnMgSe Zinc Magnesium Selenide
SiN Silicon Nitride
SEM Scanning Electron Microscope
xi
Chapter 1: Introduction
1.1 Quantum Computing and Information Processing
In today?s classical digital computers, the information is encoded as bits, 0 or 1,
usually in the form of physical state of a classical system such as stationary charge or
flowing electrical current. The information is processed by manipulation of the state
by using logical gates. As we have witnessed miracles of the electronics technology
in the last century, more challenging computational problems have emerged, sometimes
practically impossible to solve with classical computers. Some of these hard problems
can be addressed by recently developed algorithms based on a new type of computational
approach, called quantum computing [1]. These algorithms are using advantage of the
quantum states of the matter that are not available with classical systems to perform
mathematical operations. In addition to enabling quantum computers, quantum mechanics
can also be used for other applications such as secure communication (quantum key
distribution) [2] and advanced sensing [3]. Hence, due to these great potential benefits,
there is a growing interest in research community to engineer feasible quantum systems
that can realize the proposed operations.
It is hard to predict the form of most suitable quantum hardware that will operate
robustly in a practical setting. The qubits, quantum counterparts of classical bits, can be
1
constructed with various quantum systems such as spin states of electrons [4], superconducting
circuits [5] or polarization of single photons [6]. Although the very basic units of quantum
computation have been demonstrated with all these systems, there are still unsolved
engineering issues to realize truly scalable system to perform complex tasks.
1.2 Some Concepts
Regardless of the form of the quantum device, the requirements to create robust
devices to perform quantum operations can be generalized [7]. The performance of a
quantum computer depends on the capability of satisfying all these requirements. However,
often times, it is hard to find a system that can satisfy all requirements at once. Hence,
the motivation of this research is to study a new platform to examine its capabilities for
the quantum computing and quantum information applications.
Before discussing the content of my research, here, I would like to briefly introduce
some concepts and requirements to the reader.
Stationary qubits are localized quantum states where the information can be written,
read and stored. The quantum state is not often isolated from the environment. Hence, due
to the interactions, fragile quantum information is lost over short a time duration, usually
much less than seconds. For example, spin qubits can be formed by spin of a particle such
as electron. When the qubit is in the superposition state of spin-up and spin-down, the
relative phase and amplitude of superposition state will be preserved for certain amount
of time until the system decays to more favorable energy state. The amount of time that
qubit can store the information is called as decoherence time.
2
Three well known decoherence mechanisms are defined to quantify the total decoherence
time [8]. T1 is called the spin-lattice relaxation time, which represents the time it takes
flipping of spin from one state to another, usually exchanging energy with phonons in
the crystal. T2 is called intrinsic decoherence time, which represents the time it takes
loosing the phase information in the spin superposition state. T ?2 is called spin dephasing
time, which represents the time it takes losing the coherence of spin vector. Usually, T2 is
the shortest among three and it sets the limit time duration a qubit can store information.
Spin dephasing and coherence times are usually limited due to interactions between single
spins and spin background of the host crystal. The broader goal is to find qubits or
engineer materials such that long coherence times can be achieved. The decoherence
times ranging from nanoseconds to seconds have been reported in the literature. The
trapped ions usually are well isolated in the vacuum, hence they exhibit decoherence
times longer than seconds [9]. The solid state systems such as impurities and defects
in crystals usually exhibits much less decoherence times in the order of miliseconds to
nanoseconds [10]. To date, spin coherence times of single donor impurities in ZnSe have
not been reported yet, however, ensemble measurements showed coherence times that are
in the order of nanoseconds [11].
To create interactions between different localized qubits, single quantum particles
that have ability to transport the information (such as photons) should be used. Interfacing
the stationary qubits with photons can allow mapping the qubit state to the photon state
(such as polarization). This interfaces are called spin-photon interfaces. They form the
fundamental pieces to realize quantum gates by allowing reading, writing and manipulating
the spin state [12].
3
To realize efficient interactions between many spin qubits, single photons should
exhibit number of features. First, photons emitted from different emitters should be
indistinguishable in terms of wavelength and polarization. The figure of merit describing
the ?indistinguishability? of photons are called homogeneity. Secondly, we should be
able to create a large number of single photons to have multiple operations within the
decoherence time of qubits. The ability to generate single photons in a given time interval
is called brightness. The brightness of the emitter is limited by the radiative lifetime of
the emitters. Hence, short radiative lifetimes usually result in more number of photons
per second. (higher ?brightness?).
Interfacing stationary qubits and single photons is a challenging task due to the
probabilistic nature of the interaction. To increase the efficiency of the process and to
create more deterministic interactions, nanophotonic devices that enhance light-matter
interactions are employed [13]. Due our capability to process semiconductors to create
such interfaces, solid state qubits based on semiconductors usually offer more practical
and scalable approach to realize scalable quantum computing.
1.3 Motivation for This Study
Impurities in wide-bandgap semiconductors are promising candidates to address all
requirements and to create quantum photonic devices [14?16]. These isolated impurities
generate single photons that exhibit transform-limited linewidths [17,18] and small inhomogeneous
broadening [19,20] through their radiative optical transitions. Impurity atoms also provide
isolated bound electrons [21] or holes [22] that can serve as spin-photon interfaces with
4
long coherence times [23, 24]. Furthermore, semiconductor fabrication and integration
methods allow efficient spin-photon interfaces [25?27]. Hence, these platforms could
satisfy major requirements of scalable quantum technology by combining pristine radiative
properties with long-coherence time spin qubits in a practical setting.
Zinc Selenide (ZnSe) is appealing material as the host crystal for impurities because
it possesses a direct optical bandgap and large natural abundances of nuclear spin-0
isotopes. The direct band gap allows efficient radiative transitions that can generate
bright excitonic emission. Both single fluorine (F) donors [28] and single nitrogen (N)
acceptors [22] have been optically isolated in this host material. Ensemble measurements
have also revealed the potential for optical active spin qubits [11]. Recently, the growth
of isotopically purified (Zn,Mg)Se/ZnSe quantum well structures was also demonstrated
[29]. Spin purification of the host crystal further resulted in extended spin coherence
times [30]. Moreover, donor-bound excitons in nanostructured pillars have been used to
demonstrate indistinguishable single photon emission between two independent emitters
[19], two-photon entanglement [20], and optical pumping of the donor spin [31].
Most studies on impurity-bound exciton emission have focused on the fluorine (F)
impurity measurements, which possesses a smaller atomic radius compared to selenium
(Se) atoms. Due to size difference between F and Se atoms, the crystal quality is worsened
due to the defects generated at high doping concentrations. This effect is called self-
compensation and limits the range of doping. On the other hand, chlorine (Cl) has an
atomic radius closer to that of Se and is therefore more suitable as dopant with a low
probability of self-compensation [32]. This enables Cl doping levels that cover a wide
range of densities between 1016 cm?3 and 1019 cm?3. But all past work on Cl impurity
5
bound exciton emission was in the large ensemble regime, and single Cl donor bound
exciton emission has yet to be studied.
In this thesis, we will focus our attention to one particular type of impurity, Cl
donors in ZnSe. We aim to study the optical characteristics of bound excitonic emission
originated from Cl impurities, and show single photon emission from isolated Cl bound
excitons. While doing that, we will also attempt to understand the spin properties of Cl
impurity bound electrons. Lastly, we will demonstrate the fabrication of photonic devices
in this material platform.
1.4 Thesis Outline
In Chapter 2, we will start with a short introduction of bound excitons in ZnSe
where we will discuss how bound excitons are used to generate single photons. In Chapter
3, we will discuss the growth of the material and incorporation of Cl atoms to the ZnSe
crystal. We will also discuss the motivation for using quantum wells and their effect on
the radiative properties. In Chapter 4, we will detail optical characterization of radiative
transitions of bound excitons. We will demonstrate the single photon emission from
single Cl impurities. Following this discussion, in Chapter 5, we will discuss the energy
levels, Zeeman splittings and allowed optical transitions of impurity bound electron. In
the Chapter 6, we will report the formation of biexcitons from Cl complexes. We will
demonstrate cascaded single photon emission from exciton-biexciton pairs. Finally, in
Chapter 7, we will report our efforts to design and fabricate photonic crystal cavities in
ZnSe platform.
6
Chapter 2: Bound Excitons in ZnSe
Isolated single impurities in crystals may host a bound electron that can serve as a
spin qubit. They can also support bound excitons, excited electron-hole pairs, that can
generate single photon emission. Combining these two capabilities in the same system is
promising to realize spin-photon interfaces, which are fundamental blocks of spin based
quantum computing. To achieve this broader goal, first, a number of fundamental steps
must be realized. These are a) to isolate single impurities b) to realize efficient spin-
photon interfaces c) to create interaction between multiple single qubits on the same chip.
The donor impurities at low doping concentrations can be used to isolate single
photon emission and single spins. The donor bound electron generates a shallow potential
trap for excitons, resulting in a donor-bound exciton emission localized at donor site. The
donor bound exciton emission acts as a single atom-like quantum emitter that can emit
single photons. The spin of the bound donor electron can serve as isolated single spin,
that can be used as a quantum register.
To realize efficient spin-photon interfaces, impurities can be integrated in photonic
cavities. These cavities serve as a medium to increase the interactions between electron
spins and single photons.
To create interactions between multiple single qubits on the same chip, single photons
7
can be used. Photonic pathways such as waveguides enable the efficient transport of
quantum information (encoded on single photons) on the different positions in the same
chip. Hence, broad networks that allow the interaction of multiple localized spins can be
realized.
Zinc Selenide (ZnSe) is appealing as the host crystal to host such impurities because
it possesses a direct optical bandgap and large natural abundances of nuclear spin-0
isotopes [29, 33]. The direct band gap allows efficient radiative transitions that can
generate bright excitonic emission. The existence of nuclear spin-0 isotopes allows to
create spin-free background so that long coherence times can be achieved [30].
In this chapter, we will start with the discussion of the ZnSe crystal structure. We
will discuss how Cl impurity is incorporated in the ZnSe crystal upon doping. Then, we
will discuss the formation of bound excitons and their optical transitions.
2.1 Crystal Structure
Figure 2.1 illustrates the atomic structure of a ZnSe crystal with a single Cl impurity.
In its intrinsic form, zinc and selenium atoms are arranged in a face centered cubic lattice
where each zinc atom is surrounded by 4 selenium atoms. A single Cl impurity replaces
a Se atom [34] when chlorine impurities are introduced into the lattice via in-situ doping
or ion implantation. The Cl impurity serves as a neutral electron donor, consisting of the
positively charged chlorine core and a bound electron, as shown in Figure 2.1.
8
Zn
Se
Cl
Figure 2.1: Crystal structure of ZnSe with a single Cl impurity substituting a Se atom.
2.2 Donor Bound Exciton State
A substitutional Cl donor in the ZnSe crystal provides a single electron. At the room
temperature, this electron is in the conduction band due to the thermal energy. However,
at sufficiently low temperatures, the attractive potential created by the donor atom can
trap the free electron, hence the impurity electron becomes a localized bound electron at
the impurity site. This donor-bound electron may serve as a stable electron spin qubit.
Figure 2.2 shows the energy level structure of the Cl impurity. Optical excitation
above the ZnSe band gap produces free electron-hole pairs in the material. These electron-
hole pairs, also called as free excitons, can become bound to the Cl impurity, and they
form a bound state. This excited state is called the donor-bound exciton state and is
shown as D0X. Bound excitons can radiatively recombine back to the 1s ground state by
9
??0??
?????? ??????
Bound Electron 
States
2??, 2??
??0 1??
Figure 2.2: Energy band diagram showing the ground state (D0) and
excited state (D0X), along with the bound exciton and two-electron-satellite
emission, shown as blue and red arrow, respectively.
emitting a photon around 440 nm.
During the exciton recombination process, the bound electron can be excited temporarily
to its higher energy 2s, 2p states and decay back to its ground state through an Auger
recombination. If this happens, due to energy absorbed by the electron, bound exciton
recombination produces a longer wavelength photon compared to D0X. This emission is
called as two-electron-satellite (TES) emission. TES signal can only be observed in the
presence of the bound electron. Therefore, the observation of TES in the PL spectrum
establishes the presence of the impurity-bound electron.
10
Bound Exciton States
The binding energy of the donor-bound exciton can be approximated by the Hydrogenic
model by using effective mass theory. Previous theoretical calculations show that the
bound state formed by Cl impurity is a shallow state with a small binding energy, which is
less than 7 meV. These bound excitons have Bohr radii of around 5 nm. These calculations
will not be part of this thesis. Hence, we refer the reader to the references for detailed
discussions [35].
11
Chapter 3: Material Growth
All samples studied in this thesis were grown by molecular beam epitaxy (MBE)
process on the GaAs wafers. The quantum well structures are composed of ZnMgSe
barrier layers and Cl doped ZnSe wells.
In this chapter, we will first discuss the layer structure of the samples, growth and
doping process. Then, we will discuss the effects of the quantum wells on the optical
properties of bound excitons.
3.1 MBE Growth
The growth of the material starts with a commercially available GaAs wafer. First,
a thin (a few hundred nm) epi-layer of GaAs is grown on the GaAs wafer to achieve a
clean and atomically flat surface. After the growth of the GaAs epilayer, the surface is
passivated by arsenic cap-layer. Without exposing the surface to the air, the sample is
transferred to the another chamber to continue the growth of ZnSe layers. During the
transfer, ultra high vacuum is maintained to reduce the undesired impurity contribution
from the background and to achieve defect-free growth. The details of the growth process
can be found in the related literature [36?38].
Figure 3.1 shows the epitaxial layer structure of the sample used in this work. The
12
31 nm
4.6 nm
Cl 31 nm
10 nm
Figure 3.1: Layer structure of the grown material with in-situ doped Cl
impurities in the ZnSe quantum well.
sample is composed of a 4.6 nm ZnSe quantum well embedded in a (Zn0.88Mg0.12)Se
barrier on top and bottom each with a thickness of 31 nm. To achieve controlled doping
of impurities while sustaining the high crystalline quality, we grow a thin Cl delta-doped
layer in the middle of the ZnSe quantum well. The delta-doped layer has a Cl sheet
concentration of approximately 109 cm?2. The overall structure is grown on a GaAs
substrate covered with a 10 nm ZnSe buffer layer.
The lattice mismatch between ZnSe and GaAs is between % 0.2 and % 0.4. The
small mismatch allows two dimensional growth and high-quality epitaxial layers [39].
However, due to the small but existent lattice mismatch, the quantum well layer experiences
a compressive strain.
13
3.1.1 Delta Doping of Cl Impurities
The delta doping of chlorine impurities is performed during the MBE growth of
ZnSe/ZnMgSe layers. During the growth of the ZnSe quantum well, ZnCl2 cracker
cell is heated to 300 C? to dissociate Cl atoms. The precise control of doping time
is achieved with the magnetic shutter that can be opened and closed in milliseconds
resolution. Opening the shutter for short amount of time (usually 1 second) allows ?-
doping of impurity atoms within a monolayer at the center of the well. The position
of the Cl impurities within the well directly affects their binding energies due to the
change in the strength of the quantum confinement. Limiting the doping profile to a
single monolayer allows precise location of the Cl atoms (ignoring the diffusion), and it
decreases the binding energy variations.
To calculate the doping concentrations in the ?-doped sample, we first measured
the volume concentration (atoms/cm?3) of impurity atoms in a uniformly doped epilayers
with secondary ion mass spectroscopy (SIMS) and then we calculated the density of
impurity atoms per monolayer.
3.2 The Effect of Quantum Confinement
Using quantum wells serves two purposes. First, due to the quantum confinement
effect, the binding energies of the donor bound excitons are increased [40, 41]. Second,
ZnMgSe barrier prevents the leakage of bound excitons from the the quantum well region,
hence enhanced optical brightness can be achieved [42].
We define the binding energy of the bound exciton as the energy difference between
14
the free exciton emission in ZnSe (EFX) and bound exciton emission of the impurity
(EDX). In the bulk, this binding energy (EFX ? EDX) has been reported as 7 meV
for Cl impurities [43]. Considering the similarity of Bohr radius (about 5 nm) of the
excitons and quantum well thickness (4.6 nm), we expect both free exciton and bound
exciton wavefunctions to be squeezed in the axis of the growth. This confinement causes
an increase both in the binding energy of the free exciton (Eg ? EFX) and the binding
energy of the bound exciton (EFX ?EDX). Using the variational method for hydrogenic
states, the increase in the binding energy can be estimated [44]. Using the quantum
well has increased the binding energies by 2-3 times compared to the bulk values. This
enhancement allowed us better separating the single bound exciton emission from the
free-exciton spectrally.
15
Chapter 4: Radiative Properties of Single Bound Excitons
4.1 Introduction
Isolating impurities in ZnSe allows direct access to the atom-like transitions of
single bound excitons. To understand the radiative properties of these transitions, the
detailed characterization involving spectral and temporal measurements are required. Previous
studies discussed in the Chapter 1 employed the nanopillars to isolate single F impurities.
However, the bound exciton emission is susceptible to surface charge effects and material
strain, both of which are modified due to the pillar geometry. Studying the optical
transitions of donor bound excitons in an unstructured bulk material has greater importance
to understand the both radiative and non-radiative properties. Nevertheless, to my knowledge,
no previous studies have isolated single bound excitons in the bulk material to date.
Here, we optically isolate a donor impurity in a bulk material without using any
nanostructure, and characterize the radiative properties of single bound exciton emission.
Optical isolation of single bound excitons requires a sufficiently low density of impurity
atoms and sufficiently large exciton binding energy. The latter is needed to separate bound
exciton emission from the free exciton emission spectrally [40?42]. We achieve both of
these requirements by using a quantum well that is delta-doped with a low concentration
of Cl impurities, as discussed in Chapter 3.
16
In this chapter1, we employ high-resolution optical spectroscopy techniques to study
the radiative properties of single bound excitons. First, we start with photoluminescence
measurements and discuss the spectral properties of the samples investigated. These
measurements give us insight about the binding energies, inhomogeneous broadening
and stability of the emission. Then, in section 4.3, we show results of photoluminescence
excitation, and discuss the origin of the observed emission lines. In section 4.4 , we
discuss the temporal dynamics of the emission by measuring the time-resolved photoluminescence.
Then, in section 4.5, we measure photon correlations and demonstrate the single photon
emission from the Cl bound excitons. Finally, in section 4.6, we discuss methods to tune
the emission wavelength, and illustrate the temperature tuning. Our results demonstrate
the isolation of single Cl impurities in ZnSe and provide complete characterization of
bound exciton emission.
4.2 Photoluminescence Characterization
4.2.1 Measurement Details
To investigate the spectral properties at cryogenic temperatures, we mounted samples
on an XYZ piezo stage and maintained the sample temperature at 3.6K in a closed-loop
helium cryostat. During this study, we used two different cryostats, namely, Montana
Cryostation CR-270 (Montana Instruments), and AttoDry 2100 (Attocube). Although
both systems offer similar cryogenic capabilities, the latter allowed us to apply the magnetic
field with the help of a built-in superconducting magnet.
1The results discussed here will be based on arXiv:2203.05748
17
To perform optical measurements in the wavelength region of the bound exciton
emission, we built a confocal microscope system with free space and fiber optic components,
working in the range of 400 nm - 550 nm. The optical setup was composed of an objective
(NA:0.8), beamsplitters, waveplates and polarizers, filters, attenuators and fiber couplers.
We excited the samples with either 405 nm CW diode laser (Thorlabs) or frequency-
doubled, tunable, 4-ps pulsed laser (Mira Optima 900P) with repetition rate of 76 MHz.
The excitation wavelength of 405 nm corresponds to 3.06 eV, that is higher than both
quantum well and barrier bandgap energies.
To spatially isolate individual impurities in the bulk material, we used confocal
collection to a single-mode fiber. The single mode fiber has numerical aperture of 0.12.
Using a small NA allowed us to filter the unwanted background signal coming from the
neighborhood of the impurity emitter. We analyzed the photoluminescence signal with an
imaging spectrometer (Princeton Instruments, SP-2750). The spectrometer is composed
of an opening slit, a diffraction grating and nitrogen-cooled CCD. Narrowing the opening
slit width to 30 ?m allowed us to achieve a spectral resolution of 15 pm.
4.2.2 Photoluminescence Spectrum
We start by analyzing the photoluminescence spectrum from the sample. We excite
the sample with 405 nm CW laser that creates free electron-hole pairs (called free excitons)
both in the quantum well and the barrier layers. The excitation scheme and creation of
electron-hole pairs are sketched in the Figure 4.1.a. When the electron-hole pairs are
generated in the quantum well, due to the attractive potential created by Cl impurities,
18
a single electron-hole pair may be trapped at a local impurity site and form a localized
bound-exciton. The bound exciton state is not a stable state, hence it decays to its ground
state by electron hole recombination. A single photon emission originates from this
radiative recombination process.
There are three characteristic spectral features of the PL emission that is shown in
Figure 4.1.b. The highest energy emission region of the spectra (labeled as FX, centered
at 436.7 nm) is from the radiative recombination of free excitons in the quantum well.
Close look up of this broad emission shows several sharp peaks within the bandwidth of
FXline. These narrow lines are result of thickness fluctuations in the ZnSe quantum well.
The second broad region denoted by X?corresponds to the negatively charged trion state
in the quantum well [45]. This emission is composed of 2 electrons-single hole pairs. In
addition to these broader features, the distinct narrow feature (labeled as D0X), originates
from excitons bound to the single Cl donors.
Sharp and localized D0X lines are found around 15 meV away from the FX peak.
As we move the position of the sample, we observe different localized lines appear in the
spectrum at different spectral locations (mostly 5-20 meV away from FX), as opposed to
constant energy FX and trion lines. Our main focus here will be on the location dependent
narrow lines that are originated from the bound excitons.
To confirm that bound exciton emission originates from a two-level system with
finite number of optical states, we perform intensity-dependent photoluminescence measurements.
We record the spectrum at increasing excitation powers. Then, we fit the emission lines to
a Lorentzian curve. Using the fitting, we extract linewidth and intensity of the emission
for various excitation powers. We perform the same experiments with both pulsed and
19
a) b)
??1
??0
405 nm
???? ??0?? ??????
2??,2??
1??
Figure 4.1: Photoluminescence response obtained with the above band
excitation. a) Schematic illustrating the excitation and recombination
mechanisms of the observed FX and D0X lines. b) PL spectrum
demonstrating free exciton (FX), trion (X?) and bound exciton (D0X)
emission.
20
CW laser excitation. Both excitation schemes result in similar qualitative response. Here,
we only discuss the latter for the sake of simplicity.
Figure 4.2.a shows the emission intensity of both free exciton and bound exciton
line as a function of the pump intensity. The bound exciton line exhibits an intensity
saturation at pump powers above 5 ?W . In contrast, the free exciton line exhibits no
saturation behavior in the applied excitation power range. The donor bound exciton state
is composed of a single bound exciton, which can be modeled as a two-level system with
a characteristic excited state lifetime. Due to finite number of optical states, the bound
exciton emission saturates at low powers. On the other hand, we can create many free
excitons within the excitation spot which causes FX emission to saturate at much higher
pump powers. The emission intensity at the saturation is dependent on the brightness of
the emitter.
The Figure 4.2.b shows the log scale power dependence of both curves for the low
excitation powers, shown up to 20 ?W . Fitting of both D0X and FX lines up to 3 ?W
shows linear dependence on the excitation power, with P 0.98 and P 1.06, respectively. The
observed power dependence is consistent with the exciton power dynamics that is reported
in the literature [46].
To show the effect of excitation power on the linewidth of the emission, we fit the
bound exciton emission line to Lorentzian curve and extract the linewidth with respect
to the pump power. The linewidth (calculated as FWHM) is shown in the Figure 4.3.
At low powers, the linewidth is limited by the spectrometer resolution of 15 pm. At
increasing powers, we observe a linear power broadening. The extrapolation to the zero
power displays power-independent broadening as 0.01 nm (200 GHz). Although the value
21
a) b)
Figure 4.2: a) Power dependent intensity of bound exciton (blue) and free
exciton (red), showing saturation of bound exciton lines. b) Power dependent
intensity plotted in log scale for low powers up to 20 ? W , showing linear
dependence for both free exciton and bound exciton emission.
22
is close to lifetime limited linewidth of 5 GHz (which will be discussed in later sections),
it is still non-negligible homogeneous broadening. This could be due to combined effect
of charge instabilities and temperature induced effects.
4.2.3 Binding Energies & Inhomogeneous Broadening
Quantum interference between single photons generated by different triggered single
photon sources forms the foundation of many optical quantum information processing
schemes. To realize such quantum interference, one of the fundamental requirement
is indistinguishability of the photons that are emitted by different photon emitters. To
satisfy this requirement, one must ensure that emitted photons have same energies. The
figure of merit that measures the similarity of emitted photons in terms of energy is called
inhomogeneous broadening.
The attractive potential that bounds the exciton to the impurity site can be quantified
as binding energy of the bound exciton. This is the energy needed to remove the bound
exciton from the local impurity site, hence to make it free exciton. The binding energy can
be calculated from the PL spectrum by measuring the relative relative energy difference
between the free exciton (FX) and bound exciton (D0X) emission.
Measuring the distribution of emission energies of emitters from the same sample
will reveal both the inhomogeneous broadening and the average binding energy of bound
excitons. Figure 4.4 plots a histogram of the energies of 29 emission lines acquired at
different locations. X-axis represents the binding energy (emission energy relative to the
free exciton emission). The mean of the distribution is measured as 15 meV, which shows
23
D0X
0.04
0.03
0.02
0.01
0 5 10 15 20 25 30 35
Power (?W)
Figure 4.3: Linewidth of bound exciton line, extracted from Lorentzian curve
fitting, showing a linear dependence on the excitation power. Extrapolation
to zero power shows significant homogeneous broadening, that could be due
to charge instabilities and temperature induced effects.
24
FWHM (nm)
the average binding energy in the quantum well. This value is 2-3 times higher than
reported bulk values (7 meV) [43]. The enhancement of binding energy allows better
separation of FX and D0X emission, and it is direct consequence of using quantum wells,
as discussed in Chapter 3.
We calculate the standard deviation of the emission energies as 6 meV. The emission
energy of each bound exciton depends on various factors such as the strength of quantum
confinement and local strain. The strength of the quantum confinement will depend on
the thickness of the quantum well and the position of the impurity in the quantum well
[47, 48]. Both monolayer variations in the quantum well and diffusion of impurities in
the growth axis may cause variations in emission energies, hence results in distribution of
binding energies.
Although, the observed inhomogeneous broadening is better than epitaxial quantum
dots [49], it is still significant compared to state of art impurity based systems [50].
However, these emitters exhibit very short lifetimes, which results in increased linewidths.
For quantum interference applications, the ratio of the inhomogeneous broadening to the
transform-limited linewidth is a better figure of merit to compare the emitters. Regardless,
precise tuning of optical transition frequencies can be achieved with methods such as
Raman tuning [51], magnetic field tuning [52], or strain tuning [53].
4.2.4 Stability
To investigate the stability of the emission, we monitored changes in the spectrum
for both short time intervals, 10 seconds, and long time intervals, up to 7 hrs. For short
25
10
8
15 mev
6 median binding energy
4
6 mev
2 ?
0
?5 0 5 10 15 20 25 30 35
Binding Energy (meV)
Figure 4.4: Histogram showing the distribution of binding energies of
D0X lines, with an average binding energy of 15 meV and a standard
deviation of 6 meV.
26
# Emitters Found
FX
7
FX D0X 4?104
6 3.5?104
5 3?104
2.5?104
2?104
1.5?104
2 104
1 5000
0
436 436.5 437 440.2 440.5
Wavelength (nm)
Figure 4.5: PL intensity recorded over 7 hrs showing the stability of the emitter.
term stability, we recorded the spectrum with an integration time of 1 ms for a total
duration 10 seconds. The spectrum did not show any significant variation. Figure 4.5
highlights the long-term stability of the D0X. The photoluminescence spectrum over 7
hours under continuous above barrier excitation is recorded. The emission displays no
blinking, though slow spectral wandering is observed over a timescale of hours. We
calculated the standard deviation of spectral wandering as 0.009 nm, which is in the
same order of our spectrometer resolution. Nonexistence of blinking demonstrates of
the stability of the emitter.
27
Time (hr)
4.3 PLE Measurements
To verify that the observed D0X emission originates from a bound state of the
quantum well free exciton, we perform a photoluminescence excitation measurement.
Figure 4.6.a demonstrates the excitation scheme. To avoid excitation of ZnMgSe barrier
layers and selectively excite the ZnSe quantum well free excitons, we resonantly excite
the FX line by tuning the pump laser on FX line. Resonant excitation efficiently creates
free electron-hole pairs in the well. Since the probability of creating bound excitons
depends on the concentration of free excitons, efficient generation of free electron-hole
pairs in the well results in enhanced brightness from the bound exciton.
We use a tunable laser with a fixed pump power of 100 nW. We tune the laser
emission over the free exciton line and monitor the emission from a D0X line. During
this measurement, we also inject a weak above-band laser (about 3 nW at 405 nm).
At this power, the above-band laser produces negligible photoluminescence. However,
it creates a population of electron gas in the barrier that tunnels to the quantum well
and maintain the electron-rich charge configuration in the quantum well [54?56]. In
the absence of above band laser, the quantum well has hole-rich configuration, which
prohibits the observation of bound excitonic emission.
Figure 4.6.b shows the photoluminescence excitation spectrum of a D0X line as
we tune the pump laser over the free exciton line. We observe strong emission from the
D0X line when the laser is resonant with the free exciton. This enhancement suggests that
free excitons in the quantum well non-radiatively relax to form donor-bound excitons at
the impurity.
28
a) b)
???
??0
Excitation ??0??
??
Figure 4.6: Photoluminescence excitation a) Schematic illustration showing
how resonant excitation of FX line creates bound excitons. b) PLE spectrum
of bound exciton line obtained by tuning the pump laser over the FX line.
29
a) b)
Figure 4.7: Time resolved lifetime measurement setup a) The measurement
configuration to characterize the timing response of the measurement
system. b) The instrument response function (IRF) showing the limited time
resolution caused by detector timing jitter.
4.4 Lifetime of Emitters
To characterize the donor-bound exciton state?s lifetime, we perform time-resolved
fluorescence measurements. To excite the emitters, we frequency-double a tunable, mode-
locked, Ti-sapphire laser (Mira 900P, Coherent) by using a nonlinear crystal. This configuration
allows us to get 4 ps pulses at 405 nm with a repetition rate of 76 MHz. We use
single photon avalanche photodiodes (PDM Detector, Micro Photon Devices) and a time
correlated single photon counting module (Picoharp 300, Picoquant) to record the photon
arrival times relative to the laser pulse, shown in Figure 4.7.a.
30
The detector time jitter is the limiting factor to measure the fast dynamics. To
characterize the limits of our measurement capability, we first record the instrument
response function of the overall system at our measurement wavelength. To do that, we
tune the excitation laser to the bound exciton emission wavelength (440 nm), and record
the each detector pulse with photon detectors. Since the laser pulse is very short, it can be
approximated as delta function, hence the histogram of detection times shows the impulse
response, shown in Figure 4.7.b.
To mitigate the slow timing response of detectors, we reconvolved (EasyTau software,
Picoharp) the detector impulse response recorded at the emission wavelength with custom
defined decay function. We extracted the parameters of decay function by fitting the
reconvolved function to the measured data. Extracted data shows a biexponential decay,
with a dominant fast decay time of 192 ps and a weak contribution from a slower component
with a decay time of 2.49 ns. Figure 4.8 shows the histogram of photon arrival times
obtained from a single bound exciton along with the bi-exponential decay fit.
To investigate the origin of biexponential decay, we repeated the same experiment
at higher temperatures. Although only %4 of photons are originated from the slow
component at 4K, the contribution of slow decay to total photons rises to %24 at 60K.
The origin of the slow component may be caused by the repopulation of the D0X state
from longer-lived nearby trapping states and is a matter of further investigation.
The fast decay time of donor bound exciton transitions is an indicator of bright
emission. Combined with small inhomogeneous broadening, these well-defined and fast
emissions are making Cl impurities attractive candidates as efficient single photon source.
31
4K 60K
???
??? ???
Figure 4.8: Time-resolved photoluminescence measurement, showing
biexponential decay with dominant fast decay component of 192 ps and slow
decay component of 2.49 ns.
32
4.5 Single Photon Generation
To optically isolate single bound excitons, we aimed to limit doping concentration
to allow only 1-2 impurity per the optical diffraction spot. The delta doping of Cl impurities
allowed locating the emitters at the center of the quantum well in the growth axis. The
barrier layers serves the purpose of preventing the carrier diffusion into to the lower
bandgap material,i.e. GaAs substrate. The lateral distribution of emitters are random,
but we can search for them by laterally scanning the sample. Hence, by choosing low
doping concentration, we achieve the isolation of individual emitters without need for
further nanostructuring.
In order to demonstrate the single-photon emission, we perform second-order photon
correlation measurements. We use a frequency-doubled Ti:Sapphire laser emitting at 405
nm with a pulse repetition rate of 76 MHz to excite the emitters. We efficiently collect
the emitted photons from the sample by using a high NA objective (NA:0.8). Then, we
couple to collected light to a single mode fiber (NA:0.12) to spatially isolate individual
emitters.
To spectrally filter the narrow D0X emission lines, we use a monochromator and a
narrow slit. The monochromator grating disperses the collected light, and the variable-
width slit allows us to filter wavelength region of interest by tuning the grating angle and
slit-width. This configuration allows us both tuning the center wavelenth and bandwidth
of filtering. After the monochromator, we place a free-space 50:50 beam splitter to split
the incoming photons into two identical paths. Each path has identical single photon
detectors (PDM Detector, Micro Photon Devices) which are connected to time correlated
33
??0??
SPCM
Enclosing Cryostat
Figure 4.9: The schematic of the setup to measure the autocorrelation
function from the D0X line. First, the collected light is filtered with the
monochromator, then photons originated from bound exciton emission are
sent to the identical detectors. We enclose the detectors to avoid the room
light contribution.
34
Laser
single photon counting module (Picoharp 300, Picoquant). During the measurement, we
enclose the photon detectors to avoid any unwanted light. The Figure 4.9 shows the
schematic of the setup.
Photon correlations can be calculated by measuring the photon arrival times between
detectors. In the presence of incoming single photons, we expect to see a dip in the
coincidence counts, because the photon will either be detected in the horizontal or the
vertical detector. This results in photon anti-bunching for zero time delays between
detectors.
To observe the photon anti-bunching, we recorded the histogram of coincidences
for 60 hours. To ensure the stability of the long integration measurements, we constantly
monitored the detector counts and spectrum, and fixed any sample shifts. The Figure
4.10.a shows the raw counts of the second-order auto-correlation measurements performed
on the single line. The photon anti-bunching is clearly observed for the zero time delay.
The spectral filtering allows us to remove the unwanted photons coming from the
nearby wavelengths. However, there are two more noise sources that contribute to the
coincidence histogram. The first one is the detector dark counts. The second one is due
to the the background counts that are within the bandwidth of the monochromator.
To remove the contributions due to detector dark counts and background emission,
we post-processed the data. The detector dark counts are not correlated with the laser
pulse. Hence, they result in uniform rise in the histogram counts. To remove the dark
count contributions, we calculate the average of the uniform base and remove it uniformly
from coincidence counts.
To remove the background emission contribution, we first quantified the emitter and
35
a) b)
Figure 4.10: a) Coincidence counts of two detectors with variable time delay,
obtained with 60 hours of integration, showing emitted photons are single.
b) Noise corrected second-order auto-correlation function obtained from the
single line, demonstrating almost perfect photon antibunching.
36
background counts on the detectors. The total counts in detector 1 and 2 can be written
as D1(t) = S1(t) + B1(t), D2(t) = S2(t) + B2(t), respectively. S(t) represents the
pure emitter counts, whereas B(t) represents the background. Assuming that background
emission has Poisson distribution, it will result in a uniform pulse train in the histogram.
However, we expect S(t) to show perfect antibunching, hence it will result in a perfect
pulse train except for zero time delay. Here, the counts measured at non-zero time delay
pulses will be determined by D1(t)?D2(t) but only the contribution coming from S1(t)?
S2(t) is actually emitter counts. Here, we take the average of non-zero time delay pulses,
lets say N , and calculate the numeric contribution of background by N ? (D1 ?D2 ?S1 ?
S2)/(D1 ?D2). Finally, we extract this background contribution uniformly from all pulse
trains including the zero time delay pulse. After the noise correction [57], the second-
order correlation exhibits almost perfect anti-bunching with g2 = 0.06 (Figure 4.10.b),
which confirms that the emission line corresponds to a single photon emitter.
4.6 Temperature Tuning of the Emission
We demonstrated, in Section 4.2.3, that Cl bound excitons from the same sample
exhibited small inhomogeneous broadening with standard deviation of 6 meV. However,
to increase the probability of frequency matching of the emission from individual donors,
various tuning mechanisms such as electrical tuning, thermal tuning and strain tuning can
be employed. Here, we will limit ourselves to thermal tuning of emitters.
Figure 4.11.a shows the PL emission of both D0X and FX lines with increased
sample temperatures up to 70K. By increasing the temperature in this range, we tuned
37
the emission by 3.5 meV without much sacrificing from the integrated brightness. At
temperatures higher than 60K, we observed a significant temperature broadening (Figure
4.11.c), which limited our tuning range.
Figure 4.11.b shows the shift of both FX and D0X lines with respect to sample
temperature. Both curves show a similar temperature dependence, proving that wavelength
shift is due to change in the band gap of the material at increased temperatures [58, 59].
We achieved tuning the emission more than half of the standard deviation of the
inhomogeneous broadening. Although our results show that temperature tuning can be
employed to some extent to tune the emission of emission from the different emitters,
other tuning mechanisms such as stark shift or local strain tuning may be investigated for
better tuning capabilities.
38
a) b) c)
Figure 4.11: Temperature tuning of the emission. a) PL spectrum obtained at
increasing sample temperatures up to 70K. b) The relative shift of emission
energies of FX and D0X lines from the 3.6K values, showing a similar
temperature dependence for both curves. c) Temperature dependent linewidth
showing that significant broadening starts at 60K
39
4.7 Summary
In this chapter, we optically isolated single excitons bound to the Cl donor atoms
in ZnSe quantum wells. We observed clear saturation of emission intensity at increased
excitation powers, demonstrating the finite optical states of bound excitons. We confirm
the stable single photon emission by observing a clear photon antibunching. The fast
radiative recombination time below 200 ps demonstrates the potential for high brightness.
Our results establish Cl bound excitons as bright and stable single photon sources.
40
Chapter 5: Spin States of Bound Electrons
The realization of spin-photon interface requires robust and optically accessible
spin qubits. The bound electron of an impurity can serve this purpose because ground
state of the bound electron possesses spins that can be employed as a qubit. In this
chapter1, we will study the both ground and excited states of electron donor bound to
the Cl impurity by using magneto spectroscopy techniques.
In a neutral donor system, at cold temperatures, we expect the donor electron to be
bound to the impurity and become localized at the impurity site. Before discussing the
spin states, we will first start with resonant excitation of bound excitons and observing for
the existence of two-electron-satellite emission. The presence of TES signal allow us to
establish the existence of the bound electron.
Under no external magnetic field, neither ground nor excited states show significant
fine structure. However, the degenerate spin states split in energy by increasing the
magnetic field, further allowing selective optical access to desired qubit states [60]. In the
rest of the chapter, we will discuss how degenerate energy states of single electron system
will split under external magnetic field. Based on this discussion, we will discuss the
selection rules that govern the allowed optical transitions. In the following sections, we
1A manuscript is under preparation discussing the results from this chapter.
41
will introduce our measurement setup to characterize the system. Magneto spectroscopy
results will unleash the detailed level structure under both Faraday and Voigt magnetic
field. Our results will verify the presence of spin qubit states from Cl bound electrons in
ZnSe quantum wells.
5.1 Two-Electron Satellite Emission
Optical excitation above the ZnSe band gap produces free excitons in the quantum
well. These excitons can become bound to the Cl impurity, forming an impurity bound
exciton state D0X. This state can radiatively recombine back its ground state by emitting
a photon around 440 nm. During this process, the bound electron can absorb some energy,
can be excited to higher orbital states such as 2s,2p states. Due to this energy lost, bound
exciton emission is redshifted by around 20 meV. This emission is known as two-electron-
satellite (TES) and shown in Figure 2.2.
The Figure 5.1 shows the PL signal obtained with a weak above-band laser and
varying power of resonant pump laser tuned at the D0X line. Two-electron-satellite (TES)
emission 19.6 meV below the D0X line emerges, and its intensity increases as we increase
the resonant pump power. Satellite emission appears lower in energy by an amount equal
to the spacing of the n=1 and n=2 D0 electron orbitals determined by a Hydrogenic donor
model [43]. The weak nature of this emission is due to the predominant relaxation of D0
state to 1s ground state. However, there is a finite probability of relaxation to any of the
excited (2s, 2p, etc.) states as discussed in the Figure 2.2 . To be able to observe this low-
efficiency emission in the existence of other broad features, we used low-power excitation
42
D0X
19.6 meV
104
TES
1000
0 nW 320 nW
120 nW 420 nW
220 nW
441 442 443 444
Wavelength (nm)
Figure 5.1: Spectrums showing the TES line with increasing resonant pump. power
43
Counts (log)
600
550
500
450
400
350
300
250
0 100 200 300 400
Resonant Pump Power (nW)
Figure 5.2: Background corrected and integrated counts from dashed spectral
window showing saturation of TES line with increasing pump power.
44
Integrated Counts
and long integration times (60 sec). The background corrected integrated intensities with
increasing pump power shows saturation, shown in Figure 5.2. The observation of two-
electron satellite emission supports the presence of an impurity-bound electron [43, 61],
hence potential spin qubit.
5.2 Donor Electron Energy States under Magnetic Field
Energy levels and optical selection rules of electron in an atom can be described
by four fundamental quantum numbers [8]: the principal quantum number (n) (describes
energy eigenstates), angular quantum number (l) (describes the orbital angular momentum),
magnetic quantum number (ml) (describes the orientation of orbital), and the spin quantum
number (ms).
Due to spin-orbit interaction, both the angular momentum (l) and spin momentum
(ms) can contribute to the total angular momentum (J), which can be written as:
J = l + s, (5.1)
It semiconductors, electrons in the conduction band exhibit s-like orbital wavefunction
while holes exhibits p-like orbital wavefunction. Hence, we can assume that l = 0
(ml = 0) for electrons and l = 1 (ml = ?1, 0, 1) for holes while both have s = 1/2
(ms = ?1/2, 1/2). The total angular momentum for s orbitals becomes Js = 1/2 and for
p orbitals, Jp = 3/2 [62].
Considering the spin-orbit coupling, the total angular momentum can have for s-
like states mj = ?1/2,+1/2 and for p-like states mj = ?3/2,?1/2,+1/2,+3/2.
45
??
??0?? (2??) ???? = ?3/2
|1>
??0(1s) ???? = ?1/2
|0>
Figure 5.3: Energy levels and spin configurations of the (ground) bound
electron and (excited) bound exciton states.
When an external magnetic field is applied, these degenerate energy states will become
non-degenerate due to different coupling ratios to external magnetic field. Due to this
perturbation, lifted degeneracy will result in energy differences between the states. In
our MBE grown sample, ZnSe/ZnMgSe on GaAs exhibits internal strain that lifts the
degeneracy between light-hole (LH) and heavy-hole (HH) states in the valence band.
Therefore, heavy-hole states and light-hole states will exhibit mj = ?3/2,+3/2 and
mj = ?1/2,+1/2 respectively.
46
The Figure 5.3 demonstrates how degeneracy is lifted upon the external magnetic
field. In the neutral charge configuration, where bound electron resides in the valence
band, there are two possibilities, as mj = ?1/2 and mj = +1/2. These two states
form the basis for the electron spin qubit. In the excited state, electron-hole pair binds
to the the electron, hence bound exciton is composed of 2 electrons and 1 holes. In this
state, electrons form a singlet, and total angular momentum is determined by the hole of
electron-hole pair.
Due to the internal strain supplied by the quantum well, HH and LH states become
already nondegenerate even without any magnetic field [62]. Also, it is shown that HH
transitions are more profound with higher probability of recombination. In fact, PL
measurements showed only the emissions related to HH states. Therefore, we limit our
attention here only to heavy-hole (HH) transitions.
5.2.1 Energy Splitting
Under the magnetic field, magnetic moment interacts with the magnetic energy, and
the interaction energy can be written as
E = ?M? ? B?z = g?bmsB (5.2)
where g is the Lande? g-factor of a free electron, ?b is the Bohr magneton. Depending
on the electron/hole spin quantum number, the energy difference between states can be
calculated.
In the real crystal, there are some considerations must be taken into account to
47
model the energy splitting. First of all, due to crystal potential, effective g-factors will be
different than free space values. Secondly, electron and hole exhibits different g-factors
which are also dependent on the temperature and crystal conditions such as defect density
and growth conditions [63]. Due to such differences, although the numerical values may
differ from the theoretical calculations, qualitative discussion still will be valid. We will
use the experimental magneto spectroscopy data to extract the effective g-factors for both
holes and electrons.
5.2.2 Optical Selection Rules
As discussed in the previous section, the ground state electron has angular quantum
number l = 0, so it does not have any orbital angular momentum, and orbital shows
perfect spatial symmetry. Hence, the spin of ground state electron will align itself to the
applied magnetic field, independent from the axis. However, excited state hole has p-like
orbital with l = 1. P-orbitals, unlike s-orbitals, do not posses perfect spatial symmetry.
Hence, their reaction to magnetic field will differ depending on the field direction.
In the absence of quantum well confinement, the hole wavefunction can have any
value of ml = ?1, 0, 1. However, since quantum well has thickness around the Bohr
radius of exciton, the quantum confinement will lead the orbital wavefunction to spread
in the in the lateral direction.
When the magnetic field is applied in the growth direction (Faraday configuration),
hole spin will be oriented in the growth direction. However, when magnetic field is
in-plane (Voigt configuration), the magnetic field will compete with the quantum well
48
Faraday + 3?2
+
??0?? ? ~3 ? ????
+ 3?2
?? + 1?2
??0 ~????
?1?2
Figure 5.4: Zeeman splittings of ground and excited states of bound exciton,
along with optical transitions under Faraday magnetic field.
confinement to orient the hole spin. This competition causes a mixture of the hole spin
states in the D0X, allowing additional optical transitions [60].
Conservation of angular momentum only allows radiative emissions if ?mJ =
?1 between the transitions. Figure 5.4 shows the allowed optical transitions in Faraday
geometry. In this configuration, two orthagonal circular polarizations become allowed.
Figure 5.5 shows allowed optical transitions in Voigt geometry. In this configuration
both excited states demonstrate linearly polarized transitions with each of ground state
levels. Optical lambda system formed between transitions allows optical access for manipulating
spin, hence forms the basis for control and readout for spin qubits.
49
Voigt + 3?2 + ?3?2
??0?? ?? ?? 3 ~3 ? ????+ 3?2 ? ? ?2
?? ?? + 1?2
??0 ~????
?1?2
Figure 5.5: Zeeman splittings of ground and excited states of bound exciton,
along with optical transitions under Voigt magnetic field.
5.3 Magnetospectroscopy Experiments
We have discussed the theoretical background for the optical transitions in Faraday
and Voigt geometry. Here, we provide the experimental demonstration by analyzing
magneto-PL emissions from single bound excitons.
5.3.1 Measurement Setup
To characterize our samples under an external magnetic field, we perform polarization
resolved Pl measurements. We keep the samples in the closed-loop helium cryostat
(Attocube) equipped with an an objective lens and XYZ piezo nano-positioners (ANPx51/LT).
To apply static magnetic fields up to 9T, we use cooled superconducting magnets. First,
we collect the PL signal with high-NA objective. Then, using quarter wave plate, half
50
wave plate and polarizer, we set the desired polarization for the collection. Using polarization
optics at different angle configurations allows us to measure the PL at different polarization
basis.
The superconducting magnet has capability to apply magnetic field only in one
axis. To switch between Faraday and Voigt configurations, we change the sample holder
and objective configuration. The rotating the sample holder and replacing the objective
accordingly allow us accessing the sample optically from the top, without changing the
external optics. The Figure 5.6 shows a sketch of both configurations, along with picture
of Attocube insert that holds the objective and sample stage.
5.3.2 Faraday Magnetic Field
The optically allowed transitions in this configuration are presented in Figure 5.4.
The total spin of the bound exciton ground state is given by the single spin-1/2 electron. In
the bound exciton excited state, the donor electron forms a spin-singlet with the electron
of the exciton, giving a total spin determined by the spin-3/2 heavy hole. For increasing
magnetic field, two orthogonally polarized circular emission lines emerge as expected,
shown in Figure 5.7.
Two circular polarized photon emissions; ?+ and ?? circular polarization, can be
distinguished by using a quarter-wave plate with the fast axis at +45? or -45? angle.
Spectra indicated by solid (dashed) lines are recorded for ?+ (??) polarization with
increasing magnetic field intensities as shown in Figure 5.7. The energy difference of the
two allowed transitions is proportional to geff = ge ? 3ghh . From the observed energy
51
(a) (b) (c)
B B
Figure 5.6: Optical setup to measure PL under the applied magnetic field.
a and b) Schematic demonstrating the configuration of the insert (composed
of XYZ stage, the sample, and the collection lens) for Faraday (Voigt) field
measurements. c) The picture of the Attocube sample insert.
52
?+ ?-
9T
8T
7T
6T
5T
4T
3T
2T
1T
0T
440.2 440.4 440.6
Wavelength (nm)
Figure 5.7: Polarization resolved photoluminescence spectrum recorded
at increasing magnetic field intensity, showing the splitting of orthogonal
circular polarizations.
splitting between those two transitions, we infer geff = 0.9 which is in good agreement
with previous results obtained from single fluorine donors in ZnSe. Additionally, we
observe a diamagnetic shift of the D0X emission of 3.25?eV/T 2.
Nonexistence of fine structure splitting and clear splitting into orthogonal circular
polarizations are consistent with the expected Kramer?s degeneracy of the neutral donor
state containing a single electron in its ground state [60].
53
Counts (a.u.)
5.3.3 Voigt Magnetic Field
The optically allowed transitions in this configuration are presented in Figure 5.5.
In Voigt configuration four radiative transitions with linear polarization are allowed, two
pairs with horizontal (H) and vertical (V) polarization. These four transitions generate
two double-connected systems which allow for optical rotation of the electron spin bound
to the neutral chlorine donor. The vertically polarized photons can be distinguished from
the horizontally polarized photons by using a half-wave plate with the fast axis at an angle
of 45?.
5.4 Summary
In this chapter, we first established the presence of the bound donor electron by
observing the signature two-electron-satellite emission. Then, we discussed the interaction
of the magnetic field with spin states of Cl bound exciton. From the obtained magneto-
PL spectra, we experimentally verified the optical selection rules, and inferred the electron
and hole g-factors in Faraday configuration. We confirmed the presence of doubly connected
lambda system, which allows optical access to different spin states. Our results establish
that Cl-bound electron can serve as optically accessible spin qubit with bright excitonic
single photon emission.
54
Chapter 6: Cascaded Single Photon Emission from Biexcitons
In addition to single excitons bound to Chlorine donors, we also identified a different
set of narrow lines (exhibited themselves as paired lines) with distinct spectral properties
in the sample. With the help of detailed measurements such as power dependent saturation
and polarization resolved spetroscopy, we conclude that paired lines can be attributed to
biexciton complexes bound to donors in the quantum well.
In this chapter1, we will discuss the radiative properties of the observed biexcitons.
In Section 6.1, we will first discuss the properties of biexcitons and how they can be
employed to generate cascaded single photon emission. In Section 6.2, we will share the
experimental setup and discuss the techniques that we use to characterize the biexciton
emission. In 6.3, we will start with the discussion of PL characteristics. Then, we will
use the time resolved PL spectroscopy and statistical correlations between observed pair
of lines to prove the existence of cascaded single photon emission. Photon correlation
measurements will prove that observed emission is indeed in the form of cascaded single
photon emission. Finally, in Section 6.4, we will discuss the polarization of the emission
and energy levels of the biexciton state.
1The results discussed here will be based on the arXiv:2203.06280
55
6.1 Radiative Cascades from Biexcitonic Decay
Radiative cascades can be defined as time-ordered emission of photons. If the
emitted pairs of photons exhibit single photon statistics, they can be used as on demand
sources of entangled photons. One way of realization of cascaded emission in semiconductors
is using exciton-biexciton emission pair, where the biexciton consists of two excitons
that form a lower energy complex due to Coulombic interactions [64]. The decay of
the first exciton initiates a radiative cascade in which the photon emitted by the second
exciton may be entangled in polarization [65, 66], time [67], or both [68]. The photon
pairs generated through biexciton decay could be used as sources of entangled light with
applications in quantum communication [69, 70], photonic quantum computing [71], and
quantum metrology [72].
Entangled on demand photon generation from biexcitons are actively being studied
in various material platforms including III-V quantum dots [73] and in 2D materials
[74]. Impurity atoms in semiconductors are another material platform that shows promise
for realizing both sources of quantum light and spin-photon interfaces [15]. Although
evidence for biexciton formation has been observed in these materials from isoelectronic
impurity dyads [75] and other impurity centers [76, 77], cascaded emission has yet to be
demonstrated.
56
6.2 Experiment Setup
The observation of biexcitons depends on the position on the sample. The most
clear feature of them was the pair of lines (mostly separated by 3 nm), which showed equal
fine structure splitting. We used the same experimental setup for initial photoluminescence
measurements, that we have discussed in the bound excitons Chapter 4. To perform
polarization-resolved spectroscopy on the narrow emission lines, we fix the excitation
polarization with a linear polarizer with an extinction ratio of 1:106. Photoluminescence
signal is filtered from the excitation laser with a notch filter and is passed through a
polarization analyzer consisting of a quarter?wave plate, half?wave plate,and linear polarizer.
Changing the waveplate angles allowed us to investigate the complex polarization behavior.
One important modification was required to measure the photon correlations between
pairs of observed emission lines. Our initial setup was designed to measure auto-correlations
of each line, hence a single monochromator (SP-2750, Princeton Instruments) was sufficient
to filter the signal. In the new configuration, we introduced a 50:50 beamsplitter before
the monochromator, and employed an additional identical monochromator. Using two
monochromators allowed us performing narrow filtering at two different wavelengths
of interest. We placed identical single photon detectors (PDM Detector, Micro Photon
Devices) after each monochromator?s exit slit, and measured the time correlations between
two detector detections by using time correlated single photon counting module (Picoharp
300). The Figure 6.1 shows the sketch of the experimental setup.
57
405 nm
excitation
SMF
pol SMF
filter HWP 50:50 BS SPAD
50:50 BS
PL QWP pol
monochromator
3.6 K CCD
OL
SPAD
50:50 BS
TCSPC
SPAD
positioner
Figure 6.1: Schematic of the optical setup used to investigate optical
properties of exciton-biexciton complexes.
58
monochromator
6.3 Demonstration of Cascaded Single Photon Emission
In this section, we will present a set of measurements that demonstrates cascaded
single photon emission from exciton-biexciton pairs. The first indicator of biexcitonic
behavior is coming from photoluminescence measurements. We observe clear emission
line-pairs that exhibit a power dependence consistent with an exciton-biexciton cascade.
To further reveal the cascading, we use time-resolved decay of the emission performed at
low powers, where we only observe single excitons, and high powers where we observe
both emission lines. Comparison of two reveals that due to decay of biexcitons, main
exciton line is delayed by the lifetime of biexciton decay. Along with lifetime data, photon
correlation measurements conclusively reveal the time-ordered nature of the cascade.
6.3.1 Characteristics of PL Response
Using the continuous wave optical excitation above the barrier bandgap at 3.06 eV
(405 nm), we efficiently excited the excitons in the quantum well. We show the photoluminescence
spectrum from the sample at 3.6 K in Figure 6.2. The highest energy line labeled as (FX)
is due to heavy hole free exciton as discussed in the previous chapter. In addition, we
observe a discrete pair of lines D? X and D?A AXX at lower energies. We also observe
a second pair of lines visible in the spectrum, D?BX and D
?
BXX, which we attribute to
exciton-biexciton emission from a separate impurity center.
Figure 6.3.a shows the excitation power dependence of the emission intensity of
the D?AX and D
?
AXX lines. Although both lines show similar linewidths, we observe
markedly different behaviors for each line under increasing excitation power. In the low
59
FX
D*AX
D*BX
D* XX D
*
AXX
B
436 438 440 442 444
wavelength (nm)
Figure 6.2: Photoluminescence spectrum measured with above band
continous wave laser excitation, showing free exciton (FX), bound exciton
(D? ?AX, DBX) and biexciton (D
?
AXX, D
?
BXX) states
excitation power regime we fit the intensity of each line to I(p) ? pk, where k is a
fit parameter. For the D?AX line we observe the exponent k = 0.95, indicating nearly
linear dependence of the emission intensity on the excitation power. On the other hand,
for the D?AXX line we observe k = 1.40, indicating a super?linear dependence on the
excitation power. The nearly linear power dependence of the D?AX line and the superlinear
power dependence of the D?AXX line, followed by saturation of the emission intensity at
higher excitation powers, are characteristic of localized exciton and biexciton emission
pairs [78].
We also investigate power law behavior of the emission intensity for the D?BX and
D?BXX lines, such that I(p) ? pk. For the D?BX line we observed the exponent k = 1.09,
while for the D?BXX line we observed k = 1.36 in the low excitation power regime.
60
counts (arb.)
a b 1 c 10
?2
D*XX
105 10?3
10?4
104 D*X 10?5
0.5
0 2 4
103
D*AX
D*AXX
102 D* 0
0.1 1 10 0 1 2
power (?W) delay (ns)
Figure 6.3: a) Power dependent intensity for exciton and biexciton lines
b) Energy levels showing exciton and biexciton states c) Time resolved PL
intensity showing delayed emission from exciton lines in the presence of
biexcitonic emission.
Observing the similar power law trend for both D?AX, D
?
AXX and D
?
BX, D
?
BXX lines
support the assignment of each line as a localized exciton and biexciton, respectively.
6.3.2 Time Resolved Spectroscopy Results
Figure 6.3.b shows the expected energy level diagram of the biexciton, exciton and
ground states. As decay paths reveal, the cascaded emission is originated in the case of
biexciton emission. To demonstrate that emission from D? X and D?A AXX are cascaded, we
perform time?resolved lifetime measurements of each line in the emission pair. We excite
the sample with 3 ps optical pulses with a photon energy of 3.06 eV from a frequency
doubled Ti:Sapphire laser and filter the emission with a grating monochromator. Figure
61
counts /s
PL counts (x10-2)
6.3.c shows the time?resolved measurements of the optical emission from the D?AX and
D?AXX lines. In all panels the recorded transients are normalized by the total number of
recorded counts to better show the relative shape of each decay. The solid lines are fits to
the data using a multiexponential function convolved with the detector response.
Emission from the D?AXX line shows a biexponential decay. By fitting a biexponential
decay, we observe a fast time constant of ?XX1 = 106 ps first and a small contribution from
a slower time constant of ?XX2 = 4.6 ns second. Emission from the D
?
AX line rises more
slowly and reaches its peak intensity only after the onset of D?AXX decay. The delay
between peak emission of the D?AXX and D
?
AX lines is ? 110 ps, comparable to ?XX1 ,
suggesting that the decay of D? ?AXX feeds into the DAX state. The decay of the D
?
AX line
also exhibits a biexponenial decay with a fast time consant of ?X1 = 142 ps and a slower
time constant of ?X2 = 3.5 ns. The slow decay components, ?2, present in both D
?
AX and
D?AXX emission may indicate the influence of deeper trapping states, dark excitons, or
possible carrier recapture from the quantum well.
To further reveal the time correlations between lines, we monitored each doublet
for signs of spectral wandering over a period of 10 minutes. In Figure 6.4, we show
the slow drift of the mean emission energy of the D? ?AX and DAXX doublets, and the
D?BX and D
?
BXX doublets respectively. The D
?
A family of lines show nearly identical
behavior, consistent with the picture proposed in the main text of cascaded emission
between a localized biexciton and exciton. The D?B family of lines also show nearly
identical behavior, though we note that the spectral wandering of the D?B lines are qualitatively
distinct from the D?A lines. This difference suggests that D
?
BX and D
?
BXX are localized at
a separate impurity complex from D?A.
62
5
0
D*AX
*
?5 DAXX
5 D*BX
D*BXX
0
?5
0 100 200 300 400 500 600
time (s)
Figure 6.4: PL intensity recorded for 10 minutes showing spectral wandering
and intensity correlations between exciton-biexciton lines, for A (top) and B
lines (bottom).
63
?? (pm)
6.3.3 Photon Correlation Measurements
More definitive proof of cascaded single emission comes from the study of photon
statistics. To demonstrate that each line of emission is due to single photon emission, we
measured second order auto-correlation from each emission line in the Hanbury Brown
and Twiss configuration. We excited the emitters with 4 picosecond pulse laser at 405
nm with a repetition rate of 76 MHz. Then, after the spectral filtering, we recorded the
auto-correlations of both exciton and biexciton line for over 40 hrs. Top and bottom
panel of Figure 6.5.a show the second?order autocorrelations of the D?AX and D
?
AXX
lines respectively. Each correlation is showing photon anti-bunching, given by g(2)(0) <
0.5, which verifies the single photon nature of the emitted light. For D?AX we measured
(2)
gX (0) = 0.38, while for D
?
AXX we measured
(2)
gXX(0) = 0.43.
In order to verify the correlations between the cascaded single photons, we perform
cross-correlation measurements between the D? ?AXX and DAX emission lines. We used
the modified setup, shown in Figure 6.1, with two identical monochromators as spectral
filtering devices. We used the pulse excitation (same as auto-correlation measurements),
and recorded the counts from each emission line in a separate detector for a time period
of 48 hrs. During the long integration, we kept the samples under dark, and ensured the
stability of the measurement apparatus by carefully monitoring the counts in detectors.
Figure 6.5.b shows the cross?correlation with pulsed excitation which displays significant
photon bunching of (2)gX,XX(0) = 1.78. The observation of photon bunching, given by
(2)
gX,XX(0) > 1, establishes the enhanced probability of detecting a single photon from
the D?AX line after first detecting a single photon from the D
?
AXX line. Photon bunching
64
a 2 c 3 d
1 1.78
0.5 0.38 1.5 2
0
1
b
1 1
0.5
0.5 0.43
0 0 0
?30 ?20 ?10 0 10 20 30 ?40 ?20 0 20 40 ?2 ?1 0 1 2
delay (ns) delay (ns) delay (ns)
Figure 6.5: Single photon correlation measurements. (a,b) Autocorrelation
measurements of (a) the localized exciton line, (2)gX (?), and (b) the localized
biexciton line, (2)gXX(?). (c,d) Cross?correlation measurements between the
localized exciton and biexciton lines, (2)gX,XX(?), with (c) pulsed excitation
and (d) continuous wave excitation above the quantum well barrier.
65
g(2)(?) g(2)XX X (?)
g(2)X,XX (?)
g(2)X,XX (?)
provides conclusive evidence for the cascaded relationship between the two emission lines
and is therefore direct evidence for the presence of the localized biexciton initiating a
radiative cascade.
Measuring the cross-correlation with continuous wave excitation reveals more information
about the time dynamics of the cascade. Therefore, we also measured the cross-correlations
by using a CW excitation. We used 405 nm CW diode laser (Thorlabs) and the same
experimental setup. We recorded the coincidence counts with a time-correlated single
photon counting module (Picoharp 300, Picoquant) for a time period of 42 hrs. Figure
6.5.d, reveals characteristic temporal asymmetry of (2)gX,XX(?) expected near ? = 0 for
cascaded emission [79]. The asymmetry is a signature of (2)gX,XX(?) sharply transitioning
between antibunching for ? < 0 and bunching for ? > 0. Antibunching arises in the
continuous wave measurement because detection of a photon from the D?AX state projects
the system into the impurity ground state, out of which a biexciton photon cannot be
emitted. The temporal response of our correlation experiment (? 145 ps) prevented the
direct observation of antibunching for ? < 0 at the required time scale, but still allowed
the clear observation of the resulting asymmetry of the bunching profile. The fit function
is a solution to a three?level rate equation model taking into account the experimentally
measured lifetimes of the D?AX and D
?
AXX states convolved with the measurement system
response [79].
66
6.4 Fine Structure Splitting and Polarization Response
We consistently observe a fine structure splitting from both exciton and biexciton
lines. In this section, we will study the polarization characteristics of the fine structure to
reveal more insight about the energy level structure of the exciton-biexciton pairs. First
observation is that each doublet composed of orthogonal elliptical polarizations. Figure
6.6.a shows polarization resolved spectra of each line. The spectra reveal a fine structure
splitting between the orthogonal elliptical components of ? ?290 ?eV for the D?AX and
D?AXX lines, respectively. The equal magnitude but opposite sign of the fine structure
splitting suggests that the splitting originates in the localized exciton state D?AX [80].
A more detailed polarization analysis is performed to quantify the ellipticity of the
emission from of each fine structure component. We perform the polarization analysis
using a combination of a quarter-wave plate, half-wave plate, and linear polarizer. Changing
the angle of waveplates and recording the modulated PL intensity allowed us to access
the relative polarization state of each line. Figure 6.6.b shows the modulated intensities
for various angle configurations. We characterize the ellipticity of each line based on the
fits in Figure 6.6.b
We model the emitted photoluminescence electric field with an arbitrary polarization
in the plane as
?? ??? E0,x ?E = ??PL , (6.1)
E ei??0,y
where E0,x and E0,y are the normalized amplitudes of field along the horizontal and
67
a b
? 1 *+ D* * DAX? AX DAXX-
0.5
0
1 D*AXX
0.5
0
440.0 440.5 443.0 443.5 ?50 0 50 100 150
wavelength (nm) QWP angle (deg)
Figure 6.6: Polarization resolved spectroscopy. (a) Measured polarized fine
structure of the D?AX and D
?
AXX lines. (b) Polarization analysis of the D
?
AX
and D?AXX fine structure as a funciton of quarter wave plate angle.
68
counts (arb.)
normalized intensity
vertical directions, and ?? = ?y ? ?x is the relative phase difference between each
component. Each optical element can be modeled by a matrix that indicates it?s effect on
the polarization components of incident light that are either parallel or perpendicular to
the element?s birefringent axis. For the quarter?wave plate, half?wave plate, and linear
polarizer we have respectively,
?? ?? ?? ? ? ??1 0 1 0M i?/4 ?QWP = e ? ? ,M i?/2?HWP = e ? ??? ,M = ???0 0??POL ? ,
0 ?i 0 ?1 0 1
where the fast axis of each wave plate and the transmission axis of the polarizer were each
taken to be vertical. A rotation of each element by an angle ? can be considered with the
rotation matrix
?? ?cos ? ? sin ??
R(?) = ?? ?? , (6.2)
sin ? cos ?
such that M(?) = R(?)?1MR(?). The transmission intensity of the photoluminescence
signal through the polarization analyzer can then be described by
|E |2t = |MPOL(?p)MHWP(?h)MQWP(?q)EPL|2, (6.3)
where ?p, ?h, and ?q are the rotation angles of the polarizer, half?wave plate, and quarter?
wave plate respectively. In the experiment described in the main text, the transmission
axis of the polarizer is taken to be vertical and the fast axis of the half?wave plate is
69
aligned to the polarizer, such that ?p = ?h = 0. Therefore, the transmission intensity of
the polarization analyzer is given as
| |2 1E = |ei??t E0,y(1? i cos 2?q) + iE0,x sin 2?q|2. (6.4)
2
In the experiment, the transmission intensity is monitored as the quarter?wave plate angle
is rotated. The data is then fit to Equation 6.4, taking E0,x and ?? as fit parameters and
restricting E0,y by the normalization condition to determine the polarization of the emitted
photoluminescence.
The Stokes parameters of the signal can be directly determined using the extracted
values of the amplitudes and relative phase for the components of EPL by then calculating
(x) (x)? (y) (y)?
s0 = EPLEPL + EPLEPL
(x) (x)? (y) (y)?
s1 = EPLEPL ? EPLEPL
(x) (y)? (y) (x)?
s2 = E(PLEPL + EPLEPL )
(x) (y)? (y) (x)?
s3 = i EPLEPL ? EPLEPL ,
where s0 gives the total intensity, s1 determines the prevelance of vertical and horizontal
components of the total intensity, s2 determines the prevelance of diagonal and anti?
diagonal components, and s3 determines the prevelance of right and left handed circular
components. From each fit we determined the ellipticity angle ?, defined in terms of the
polarization ellipse as tan? = ?, where ? is the ratio between the ellipse?s minor and
70
major axes. In terms of the calculated Stokes parameters, the angle ? can be expressed as
sin 2? = s3/s0.
The angle ? therefore assumes the value ? = 0 (??/4) for linear (?? circular)
polarizations. For the D?AX doublet we found ?p1,p2 = ?0.13?, while for the D?AXX
doublet we found ?p1,p2 = ?0.14? in close similarity.
The ellipticity of the fine structure components observed here contrasts with typical
observations of neutral biexciton cascades in self?assembled quantum dots where the
fine structure components are well described by orthogonal linear polarizations [65].
The magnitude of the fine structure splitting and optical polarization of each line are
closely related to the confinement symmetry and charge configuration of the localized
excitons, and typically emerges from a confinement symmetry below D2d [80]. Possible
candidates for the observed impurity center are a neutral chlorine donor (D0), an ionized
Chlorine donor (D+), which is stable in ZnSe and known to have a higher excitonic
binding energy than the neutral donor [43], or potential multi?donor complexes which
cannot be ruled out. In particular, the observed elliptical polarization of the fine structure
components may be explained by anisotropic electron?hole exchange in the presence of a
highly charged biexciton cascade [81,82]. Further understanding of the impurity center?s
charge configuration will enable strategies for minimizing the observed fine structure
splitting, which is a requirement for realizing a source of entangled photons. Strategies
for minimizing the fine structure splitting may include tuning with electric or magnetic
fields or strain [83?85].
71
6.5 Summary
In this chapter, we reported the observation of biexcitons in a chlorine doped ZnSe
quantum well. The power dependent intensity measurements showed that excitons and
biexcitons exhibited linear and superlinear power dependence, respectively. We confirmed
the single photon nature of both exciton and biexciton line by measuring the second order
auto correlation measurements in Hanbury Brown and Twiss configuration. The resulting
radiative cascade was clearly observed in both time?resolved photoluminescence and
single?photon correlation measurements. The observed fast decay times of 106 ps and
142 ps for the biexciton and exciton further highlight the promise of this system to
produce fast and bright emission of single photons and photon pairs. These results provide
a path for the development of on?demand sources of entangled photons with impurity
atoms in semiconductors.
72
Chapter 7: Design and Fabrication of Nanophotonic Cavities
7.1 Introduction
Photonic crystals (PCs) are devices that have periodically modulated refractive
index. The local refractive index changes can be achieved with various techniques such as
removing the material (making holes), doping and placing a different materials periodically
[86].
The propagation of electromagnetic waves in a medium obeys certain phase relations.
The modulation of refractive index by introducing periodic elements in the length scales
of optical wavelengths can lead to suppression of certain modes while supporting the
others. Changing the local refractive indexes and periodicity allows engineering photonic
crystals to open a photonic band gap, which prohibits the propagation of photons for
certain directions within some frequency range [87].
Introducing local defects in a photonic crystal can lead to lateral confinement of
electromagnetic mode in the defect site in the photonic band gap. The out-of-plane
confinement is achieved by using thin slab of material suspended in the air. The total
internal reflection provides the out-of-plane confinement, hence 3D confinement of the
mode is achieved. Such devices are called photonic crystal cavities due to their highly
localized, cavity-like, mode distributions. The optical properties of defect modes can be
73
engineered by tailoring the geometrical parameters of defect site such as size , shape and
type of defects. Hence, designing photonic crystal with defects allows controlling the
mode profile, resonant frequency, mode volume and quality factors [88].
Integration of photonic crystal cavities with quantum emitters serves multiple purposes.
First, the coupling of quantum emitter radiation to cavity modes can lead to increase of
the brightness of emitter via the Purcell effect [89]. Secondly, it allows engineering the
radiation pattern of the emission to increase the directionality of emission [90]. Due
to coupling of emitter to the cavity mode, the far field radiation pattern is governed
by the field distribution of the cavity which can be engineered by tailoring the defect
parameters. One greater application is to exploit the cavity quantum electrodynamics.
When the emitter is strongly coupled to the cavity mode, the reflectivity of the cavity can
be modulated by the existence of excited state coupled to the cavity resonance [91]. This
can lead to realization of quantum gates with single qubits [92].
Photonic crystal cavities integrated with quantum emitters have been demonstrated
in various semiconductor platforms. Here, we will focus our attention to design,fabrication
and testing of a spesific photonic crystal cavity in ZnSe designed to operate at wavelengths
of bound exciton emission. Here, we limit our interest only to showing a proof of concept
device that can work in ZnSe. Hence, our results are far from being comprehensive, but
establishes that photonic structures can be designed and manufactured in ZnSe for future
integration of quantum emitters.
74
L
d b a c c c a b d
2r
Figure 7.1: Schematic of L3 cavity. Three missing holes as linear line at the
center, labeled as c, forms the cavity in the triangular lattice. The geometrical
parameters optimized for the design has been labeled as lattice constant (L),
radius (r), displacements (a,b,d).
7.2 Design of L3 Photonic Crystal Cavities
The starting photonics crystal consists of a slab of material with etched holes in the
form of triangular lattice. Removing the three holes at the center allows the confinement
of the mode, and forms the basis for the L3 cavity [93]. The Figure 7.1 shows the
schematic of the cavity design, where missing holes are represented by dashed circles,
labeled c. The quality factor of the resonance can be increased by decreasing the radiation
losses. This can be achieved varying the size of the etched holes nearby the cavity and by
displacing the end holes in the outward direction.
75
The cavity supports multiple modes. The fundamental mode (called as M1) has
longest wavelength and highest quality factor. The wavelengths of resonances depend
on the lattice constant (labeled as L). Hence, changing the periodicity of holes allows
matching the resonance wavelength to bound exciton emission wavelength, 440 nm.
We performed FDTD simulations by using a commercial simulation package (ANSYS
Lumerical) to find the physical parameters of the design. The thickness of suspended
ZnSe slab is set to 100 nm. Changing the parameters namely lattice constant (L), hole
radius (r) and displacement of nearby holes (a,b,d), we optimized the quality factor at the
wavelength of interest.
In the first design, Q factor of 50, 000 has been achieved by setting the L = 134 nm,
r = 27 nm, a = 33 nm. The Figure 7.2.a shows the electric field intensity of the
M1 mode at resonance. The cavity mode is confined in the region where the missing
holes are located. The field intensity decreases rapidly in the lateral direction since no
modes are supported to the photonic band gap. Although this design achieves very high
quality factors, there are two issues. First, the far field radiation pattern (shown in Figure
7.2.b) exhibits two strong lobes at high angles. Due to limited numerical aperture of
the optical setup, this mode is hard to collect with the objective lens, hence it is not
feasible in terms of measurement perspective. The another issue is the physical sizes
of holes and displacements. The numbers mentioned above are already in the limit of
our fabrication resolution, hence this design is hard to fabricate robustly even by using
state-of-art nanofabrication techniques.
To address both issues, we re-designed the cavity by sacrificing from quality factors.
In the new design, we achieved Q-factor of 35, 000 by setting L = 134 nm, r = 38 nm,
76
a) b) c)
|E|
Figure 7.2: FDTD simulation results. a) Electric field intensity distribution
of the fundamental mode at the resonance b) Far field radiation pattern of
high-Q design c) Far field radiation pattern of lower-Q design
a = 27 nm and d = 27 nm. The far field pattern of the new design shows that increased
ratio of the emitted radiation is within the acceptance cone of our objective. Therefore,
we have chosen to proceed with these parameters.
7.3 Fabrication of Cavities
The sample is composed of 100 nm thick ZnSe/ZnMgSe quantum well layer
epitaxially grown on GaAs substrate. The details of the quantum well and sample growth
have been discussed in the Chapter 3. The Figure 7.3 depicts the steps of fabrication
recipe.
To fabricate the photonic crystal cavities, we used high resolution e-beam lithography
with directional etching. We discuss the detailed steps below.
77
100pA / 100keV 
ZEP520A (300nm)
SiN (120nm) SiN (120nm)
ZnSe QW (100nm) ZnSe QW (100nm)
GaAs GaAs
1) Nitride Dep. (PECVD) 2) Resist Coating 3) E-Beam Lithography 4) Developing
5) Etching of Mask (ICP) 6) Etching of ZnSe (ICP) 7) Undercutting (Wet Etch)
Figure 7.3: Nanofabrication steps to create suspended photonic crystal cavities
78
In the first step, we deposited 120 nm SiN by using Plasma Enhanced Chemical
Vapor Deposition (PECVD). The thin SiN layer serves as hard mask while etching the
ZnSe. The thickness of SiN has been determined by measuring the etch rates ZnSe and
selectivity of recipe between ZnSe and SiN.
Secondly, we spin-coated a positive e-beam resist (ZEP520A) diluted with a solvent
(Anisole). The diluting the resist and using high spin coating speeds (4000 rpm) allowed
us achieving very thin layer of resist (160 nm) to achieve the better resolution needed for
small holes.
Then, we performed electron beam lithography (Elionix ELS-G100) by using 100
pA current with 100keV acceleration energy. The small radius of holes requires very
high resolution, therefore very careful alignment of the beam is necessary to achieve the
increased resolution. The exposed regions of resist is then developed by using developers
MIBK and ZED.
The directional etching is composed of two steps. In the first step, we transferred the
resist pattern to the SiN mask layer by using F based Inductively Coupled Plasma Etching
(Oxford ICP-RIE). Then, we used another recipe optimized for directional etching of
ZnSe to transfer the mask pattern to the ZnSe layer.
To make ZnSe slab suspended, we used a wet chemical etch step to selectively etch
GaAs underneath. To allow the etchant to reach underneath uniformly, we placed large
rectangular holes around the PC. Considering the small size of etched holes, it is critical
to have very selective wet etch, etching GaAs while not etching the ZnSe layer.
The Figure 7.4 shows SEM images of the fabricated device after all steps. Using
a wet etch recipe that is not perfectly selective results in enlarged holes. Our wet etch
79
a) b) c)
500 nm
Figure 7.4: SEM images of fabricated photonic crystal cavities. a) Cross
sectional view of the etched lines, showing the directionality of the dry etch
process. b) Top view showing the triangular lattice with 3 missing holes,
forming L3 cavity. Inset shows the close up image of the cavity region.
Rectangular holes around the structure are etched to facilitate undercutting of
GaAs to make structures suspended. c) Angled view image of the fabricated
device, showing that substrate is not etched completely.
80
recipe did not have perfect selectivity, therefore we had to limit etching time not to distort
holes more than tolerable limits. That resulted in partial etching of the substrate material,
as seen in the Figure 7.4.c.
To address the undercut issue, we recommend using a modified layer structure as
shown in Figure 7.5. Using a sacrificial layer such as AlGaAs will allow us using more
selective chemical such as HF, and hence will result in better undercut while maintaining
the size and shape of the etched holes. Nonetheless, we will discuss the measurement
results of only the original structure in the next section. We leave the rest as an open
question to the future fellow researchers.
7.4 Optical Characterization
To measure cavity properties, we mount the sample in a closed-cycle cryostat and
we kept the sample at 3.6K. We excite the sample with 405 nm CW laser, that excites both
free excitons and bound excitons. We collect this emission with a confocal microscope
with an objective lens, NA of 0.7. Even in the absence of narrow lines originated from
bound excitons, sample PL shows a faint background in the broader range of wavelengths
between (437 nm to 442 nm). We use this background emission to hunt for cavity
resonances.
Due to fabrication imperfections and non-perfect modeling of refractive index of
ZnSe/ZnMgSe layer, the fabricated structures may deviate from the designed optical
optical response. To mitigate those effects, we fabricated various designs with varying
geometrical parameters. Then, we scanned the sample and look for signs for cavity
81
ZnMgSe (14%Mg) [32 nm] ZnMgSe (14%Mg) [32 nm]
ZnSe QW delta-doped with Cl [3.6 m] ZnSe QW delta-doped with Cl [3.6 m]
ZnMgSe (14%Mg) [32 nm] ZnMgSe (14%Mg) [32 nm]
ZnSe buffer [10 nm] ZnSe buffer [10 nm]
AlGaAs Sacrificial layer
001-GaAs substrate 001-GaAs substrate
Figure 7.5: Layer structure of the fabricated devices. Left image shows the
the sample in-use. Lack of sacrificial layer resulted in non-selective etching
of GaAs. Right image shows the proposed structure with AlGaAs sacrificial
layer.
82
resonance.
The Figure 7.6 shows the spectrum of the device with a cavity mode. The blue
curve shows the reference spectrum measured at the unstructured portion of the sample.
On the photonic crystal structure, the center wavelength of the FX line is slightly red-
shifted due to the strain relief, and it is consistent when measured on other devices in the
same chip. The enhanced peak at 445, on the other hand, is device dependent and only
brightens at the center of the photonic crystal, where the cavity is located. We measure the
quality factor of this peak as 110, which is two orders of magnitude smaller than FDTD
results. The difference can be explained by non-perfect etching of the substrate. As seen
in the Figure 7.4, non-etched substrate creates lossy path towards the high-refractive index
substrate, significantly decreases the quality factor of the cavity.
7.5 Discussions
In this chapter, we discussed our attempts to realize photonic crystal cavities on
ZnSe platform. Our efforts to achieve high-q cavities coupled to donor bound excitons
are interrupted mainly due to non-availability of the material with the sacrificial layer.
However, the photonic designs and fabrication results showing patterning of ZnSe establishes
that ZnSe photonic cavities can be integrated with donor bound excitons. We worked
towards this goal by working with our collaborators who are specializing in the MBE
growth. However, those studies will be the topic of another thesis.
83
Figure 7.6: The measured spectrum of the sample. Blue line shows PL
spectrum from unpatterned bulk region. The orange curves demonstrates
cavity mode with q-factor of 110. The low q-factor is caused by radiative
losses, due to non-perfect undercutting.
84
Chapter 8: Conclusion and Outlook
Using solid state qubits to realize scalable quantum computing has a great advantage
due to our capabilities to process solid state materials with modern nanofabrication methods.
Integrating photonic devices with solid state qubits in the same platform enables realization
of efficient spin-photon interfaces to read/write quantum information by using photons.
Furthermore, lossless photonic pathways allows the routing of quantum information encoded
photons to any place on the chip.
Great benefits of scalability and photonic integration have led to numerous efforts
to implement qubits in solid-state systems. Quantum light sources and solid state qubits
have been extensively explored for various platform ranging from quantum dots to color
centers in diamond. However, we are still lacking a qubit system that provides the full set
of capabilities required to achieve scalable quantum technology.
Donor impurities in ZnSe were proposed as the basis for single photon sources and
qubits. To this extent, only flourine impurities have been isolated as single impurities to
date.
In this thesis we aimed to investigate single Cl impurities in ZnSe for their radiative
properties and quantum applications. We achieved the optical isolation of single impurities
in the unstructured ZnSe by using low doping concentration samples and high-resolution
85
confocal spectroscopy.
Donor bound excitons in ZnSe exhibit very small binding energies that leads to
short radiative lifetimes and bright emission. However, due to background emission
originated from free excitons in the material, high single photon purities are hard to
achieve. To increase the binding energies of bound excitons, we employed the quantum
confinement effect. We delta doped Cl impurities in the center of ZnSe/ZnMgSe quantum
wells. Using the quantum well increased the confinement of bound excitons, resulting in
high brightness and high binding energies.
We performed complete characterization of radiative properties of Cl bound excitons
in Chapter 4. The photoluminescence measurements showed that Cl bound excitons have
narrow linewidths with small inhomogeneous broadening. Time resolved photoluminescence
data show that they also exhibit short lifetimes in the order of 200 ps. We confirmed
the single photon emission by measuring the second order auto correlations in Hanbury
Brown and Twiss configuration. Our study shows that Cl bound excitons are promising
single photon sources with high brightness and small inhomogeneity. Future studies
may focus on the improvement of single photon emission by integrating emitters with
nanophotonic devices such as photonic crystals and bulls-eye cavities. Our first results
establish that temperature tuning can be employed at some extent to tune the emission
wavelength. The other tuning mechanisms such as strain tuning and Stark shift may also
be investigated to increase the tuning capability.
We studied the energy levels of the Cl bound electron in Chapter 5 to investigate the
potential for spin-qubit applications. We started with the discussion of energy levels of
single electron system and recombination pathways for the excited state. The observation
86
of two-electron-satellite emission upon resonant excitation established that Cl impurity
hosts a single electron in its ground state. After confirming the presence of the ground
state electron, we discussed how energy states of neutral donor evolves under the applied
magnetic field. Polarization-resolved photoluminescence measurements revealed allowed
optical transitions and effective g-factors. Our results establishes that Cl impurity hosts a
single ground state electron that can serve as a potential spin qubit. To realize the optical
control of the electron spin, in-plane magnetic field may be used in the future studies. In
Voigt configuration, optical pumping of desired spin states can be realized and electron
spin dephasing times can be characterized. To improve the spin coherence times, spin
purification of the host crystal may be performed and dilution fridge temperatures may be
exploited.
In Chapter 6, we reported the observations of exciton-biexciton pairs from the
same Cl complex. We confirmed the single photon nature of both exciton and biexciton
emission lines by measuring the photon statistics. The time resolved photoluminescence
data showed short lifetimes from both exciton and biexciton line. Finally, we studied
photon emission correlations between exciton and biexciton pairs, proving the cascaded
emission. Our results are first time demonstration of cascaded single photon emission
from an impurity based system. Demonstration of fast, cascaded single photon emission
from Cl complexes provides a path for the development of on?demand sources of entangled
photons with impurity atoms in semiconductors.
In Chapter 7, we discussed the design, fabrication and testing of photonic crystal
cavities on ZnSe platform. We, first started with FDTD simulation results, discussing
the cavity mode and far-field emission pattern. Then, we developed a nanofabrication
87
recipe to create suspended photonic crystal cavities, and showed images of the devices
that we fabricated. The optical characterization showed that cavities exhibit high losses,
which limit their quality factors. We identified the source of the loss as etched parts of
the substrate, which is caused by non-perfect undercut recipe. To increase the selectivity,
we proposed a modification in the MBE growth process. Future studies may involve
using sacrificial layers to allow easier undercut of the samples. Moreover, other photonic
devices such as bulls-eye cavities or nanobeam cavities may be investigated.
The results presented in this thesis provide a first complete study of single Cl
impurities in ZnSe. Promising radiative properties, suitable spin states and potential to
integrate emitters with photonic devices demonstrate significant potential of Cl bound
excitons. Based on the works presented in this thesis and further potentials to be realized,
single Cl impurities in ZnSe manifest themselves as attractive quantum light sources and
appealing solid-state based qubit candidate.
88
Appendix A: Fabrication Details
A.1 Nanofabrication Recipe for ZnSe Photonic Crystals
Step 0: Cleaning
Before starting the fabrication, the sample is cleaned thoroughly. The cleaning steps
are as follows:
? Acetone [90 seconds]
? Methanol [90 seconds]
? IPA [90 seconds]
? DI Water [90 seconds]
? Hot Plate [90 seconds]
Step 1: Silicon Nitride Deposition
120 nm Silicon Nitride (SiN) is deposited with PECVD.
Pressure : 1000 mTorr
Table Heater : 300 ?C
Chiller Control: 70 ?C
Time: 5 minutes
89
Conditioning:
Purge with 500 sccm N2 [30 seconds]
Gases:
? %5 SiH4 %95 N2 [400 sccm]
? NH3 [20 sccm]
Powers:
? LF : 30 Watts [7 seconds]
? RF : 20 Watts [13 seconds]
Step 2: Resist Coating
ZEP 520A (5) : Anisole (3) [mass ratio]
4000 rpm, 50 seconds
Thickness: 150 nm
Step 3: E-Beam Lithography
Beam Current: 100 pA
Acceleration: 100 keV
Write Field: 100 ?m
Pitch: 80 pm
Dose: 300 ?C/cm2
Step 4: Etching of Silicon Nitride Mask
Pressure : 10 mTorr
Table Temperature : 5 ?C
90
He Backing : 7.5 sccm
Time : 43 seconds [full etch]
Gases:
? CHF3 [45 sccm]
? SF6 [5 sccm]
Powers:
? RF : 50 Watts
? ICP : 500 Watts
Step 5: Removal of Resist Residue
Hot PG remover [? 2 hrs]
Step 6: Etching of ZnSe
Pressure : 4 mTorr
Table Temperature : 60 ?C
He Backing : No
Time : 17 seconds
Gases:
? CHF3 [45 sccm]
? SF6 [5 sccm]
Powers:
? RF : 50 Watts
91
? ICP : 500 Watts
Step 7: Removing Hard Mask
22 seconds of Step 4
Step 8: Undercut
Time: 2 minutes
Mixture H2O2 [7mL] + H2O [190mL]+ NaOH [1.72 gram]
92
Appendix B: Other Research Contributions
B.1 Demonstration of Robust Quantum Photonic Devices with Topological
Edge States
In this work (Appendix C: Publication J.3), we use the topological insulator concepts
from the condensed matter physics to implement topologically robust edge states at the
interface between two topologically distinct photonic crystals. We use finite-difference-
time-domain simulation (FDTD) techniques to design two perturbed honeycomb photonic
crystals, for which equilateral triangular air holes are shifted concentrically inward and
outward direction, respectively. We show that confined edge modes exist at the interface
with chiral propagation properties. To show the presence of the guided mode coupled to
the embedded quantum emitter, we excite the Gallium Arsenide/Indium Arsenide (GaAs/InAs)
quantum dots and measure the transmission at both ends by outcoupling the light via
diffraction gratings. To demonstrate the helical light matter coupling, we apply a magnetic
field to induce Zeeman splitting, and separate two orthogonal circular polarizations of
quantum dot emission. Then, we spectrally resolve the opposite polarizations in opposite
ends of the waveguide, which proves the chiral coupling of the emitter. We verify the
robustness of the waveguide by implementing various bends on straight waveguide and
93
measuring the transmission.
By using a similar topological framework, in another work (Appendix C: Publication
J.4), we demonstrate a topological resonator that exhibits strong light-matter interactions.
In our earlier work, we demonstrate that sharp bends do not cause any significant scattering
due to the topological protection. We use this advantage to make sharp bends that form
closed loop circuit-track resonator. We use FDTD simulations to determine the length of
the resonator to create measurable free spectral range. Then, we fabricate our structures
with nanofabrication techniques. We demonstrate a cavity induced Purcell enhancement
of 3.4 by measuring the intensity enhancement of the quantum dot. Our novel experimental
results open an avenue to design complex nanophotonic circuits for quantum information
processing and to study novel quantum many-body dynamics.
In this project, I developed the fabrication process for GaAs and partially contributed
the characterization of devices. I employed electron beam lithography and dry etching
steps to achieve high-resolution fabrication of photonic crystals. I optimized the e-beam
layout to allow realization of sharp triangular features. For the characterization of the
samples, I used a home-built optical measurement setup, and performed polarization-
resolved photo-luminescence measurements.
B.2 Integration of Quantum Emitters with Telecom Photonics
Quantum networks have great potential applications such as distributed quantum
information processing and quantum key distribution. One of the many requirements
to realize such networks is to efficiently extract the quantum light from the solid state
94
systems and to efficiently couple it to the existing telecom photonics infrastructure. In
this project, we integrate Indium phosphide/Indium Arsenide (InP/InAs) quantum dots
emitting at telecom wavelengths to telecom fiber optics and lithium niobate photonics.
We use finite-difference-time-domain methods to design InP nanobeam waveguides with
tapering tip to efficiently couple quantum dot emission into much larger waveguide modes
such as integrated lithium niobate waveguides and optical fibers. The one end of the
waveguide utilizes distributed Bragg gratings to reflect light into the tapered side. By
increasing the taper length, we adiabatically change the mode size of the nanobeam and
match it to mode size of the corresponding waveguide mode. We fabricate our designs
with e-beam lithography and directional plasma etching.
In the first part of the project (Appendix C: Publication J.6), we transfer InP nanobeam
waveguides to the another photonic chip by using a technique called as ?pick-and-place?.
In this method, we utilize Tescan FIB/SEM system integrated with a motorized tungsten
probe, and remove the suspended nanobeam waveguides from the sample and transfer
them to the on top of pre-fabricated lithium niobate waveguides. Due to the evanescent
coupling between two waveguide modes, we are able to efficiently extract the quantum
dot emission into the lithium niobate waveguides. Then, we outcouple the light from the
lithium niobate photonic integrated circuit to characterize the coupling efficiency and to
measure photon statistics. In the second part of the project (Appendix C: Publication J.5),
we use a similar nanobeam waveguide design to couple light directly to tapered single
mode fiber optic cables. Using tapered nanobeam waveguide and tapered fiber tip, we
show efficient coupling between the single quantum emitters and the single mode fiber.
Our results establish that quantum dots emitting at telecom wavelengths can be integrated
95
with mature photonic integrated circuits and existing fiber optic networks.
In this project, I designed the nanobeam waveguides with Bragg gratings and developed
the fabrication process to create suspended InP devices. I used the FDTD simulation
techniques to optimize the waveguide geometry in coordination with other designers. I
employed electron beam lithography and dry etching steps to achieve high-resolution
fabrication of the suspended nanobeam waveguides. In coordination with the transfer
team, I optimized the nanofabrication recipe to increase the yields of the pick-and-place
process.
96
Appendix C: List of Publications and Conference Talks
J.1. A. Karasahin, R.M. Pettit, N. von den Driesch, M.M. Jansen, A. Pawlis, E. Waks,
?Single quantum emitters with spin ground states based on Cl bound excitons in
ZnSe?, arXiv:2203.05748, under review.
J.2. R.M. Pettit, A. Karasahin, N. von den Driesch, M.M. Jansen, A. Pawlis, E. Waks,
?Correlations between cascaded photons from spatially localized biexcitons in ZnSe?,
arXiv:2203.06280, under review.
J.3. S. Barik, A. Karasahin, C. Flower, T. Cai, H. Miyake, W. DeGottardi, M. Hafezi,
E. Waks, ?A topological quantum optics interface?, Science, 359, 6376 (2018).
J.4. S. Barik, A. Karasahin, S. Mittal, E. Waks, M. Hafezi, ?Chiral quantum optics
using a topological resonator?, Physical Review B, 101, 20 (2020).
J.5. C.M. Lee, M.A. Buyukkaya, S. Aghaeimeibodi, A. Karasahin, C.J.K. Richardson,
E. Waks, ?A fiber-integrated nanobeam single photon source emitting at telecom
wavelengths?, Applied Physics Letters, 114, 17 (2019).
J.6. S. Aghaeimeibodi, B. Desiatov, J.H. Kim, C.M. Lee, M.A. Buyukkaya, A. Karasahin,
C.J.K. Richardson, R.P. Leavitt, M. Lonc?ar, E. Waks, ?Integration of quantum dots
with lithium niobate photonics?, Applied Physics Letters 113, 22 (2018).
97
J.7. Z. Luo, S. Sun, A. Karasahin, A.S. Bracker, S.G. Carter, M.K. Yakes, D. Gammon,
E. Waks, ?A Spin?Photon Interface Using Charge-Tunable Quantum Dots Strongly
Coupled to a Cavity?, Nano Letters 19, 17 (2019).
C.1. A. Karasahin, R. Pettit, N. von den Driesch, M. Jansen, A. Pawlis, E. Waks,
?Single photon emission from donor bound excitons in ZnSe quantum wells?, Bulletin
of the American Physical Society, G67.00002 (2022).
C.2. A. Karasahin, R. M. Pettit, M. M. Jansen, A. Pawlis, and E. Waks, ?Quantum
emission from Cl doped ZnSe quantum wells?, in Frontiers in Optics/ Laser Science,
OSA Technical Digest (Optical Society of America) (2020).
C.3. A. Karasahin, M. Jansen, A. Pawlis, E. Waks, ?Donor-bound excitons in Cl doped
ZnSe quantum wells?, Bulletin of the American Physical Society, 65, 1 (2020).
98
Bibliography
[1] Ashley Montanaro. Quantum algorithms: An overview. npj Quantum Information,
2(1):1?8, jan 2016.
[2] Peter W. Shor and John Preskill. Simple proof of security of the BB84 quantum key
distribution protocol. Physical Review Letters, 85(2):441?444, jul 2000.
[3] Lior Cohen, Elisha S. Matekole, Yoni Sher, Daniel Istrati, Hagai S. Eisenberg, and
Jonathan P. Dowling. Thresholded Quantum LIDAR: Exploiting Photon-Number-
Resolving Detection. Physical Review Letters, 123(20):203601, nov 2019.
[4] Christoph Kloeffel and Daniel Loss. Prospects for spin-based quantum computing
in quantum dots. Annual Review of Condensed Matter Physics, 4(1):51?81, mar
2013.
[5] John Clarke and Frank K. Wilhelm. Superconducting quantum bits. Nature,
453(7198):1031?1042, jun 2008.
[6] Pieter Kok, W. J. Munro, Kae Nemoto, T. C. Ralph, Jonathan P. Dowling, and
G. J. Milburn. Linear optical quantum computing with photonic qubits. Reviews
of Modern Physics, 79(1):135?174, jan 2007.
[7] David P. Divincenzo. The Physical Implementation of Quantum Computation.
Scalable Quantum Computers, pages 1?13, jan 2005.
[8] Marlan O. Scully and M. Suhail Zubairy. Quantum Optics. Quantum Optics, sep
1997.
[9] T. D. Ladd, K. Sanaka, Y. Yamamoto, A. Pawlis, and K. Lischka. Fluorine-doped
ZnSe for applications basic solid state physics in quantum information processing.
Physica Status Solidi (B) Basic Research, 247(6):1543?1546, jun 2010.
[10] Shun Kanai, F. Joseph Heremans, Hosung Seo, Gary Wolfowicz, Christopher P.
Anderson, Sean E. Sullivan, Giulia Galli, David D. Awschalom, and Hideo Ohno.
Generalized scaling of spin qubit coherence in over 12,000 host materials. feb 2021.
99
[11] A. Greilich, A. Pawlis, F. Liu, O. A. Yugov, D. R. Yakovlev, K. Lischka,
Y. Yamamoto, and M. Bayer. Spin dephasing of fluorine-bound electrons in ZnSe.
Physical Review B - Condensed Matter and Materials Physics, 85(12), mar 2012.
[12] Johannes Borregaard, Anders S?ndberg S?rensen, and Peter Lodahl. Quantum
Networks with Deterministic Spin?Photon Interfaces. Advanced Quantum
Technologies, 2(5-6):1800091, jun 2019.
[13] Ali W. Elshaari, Wolfram Pernice, Kartik Srinivasan, Oliver Benson, and Val
Zwiller. Hybrid integrated quantum photonic circuits. Nature Photonics, 14(5):285?
298, apr 2020.
[14] Gary Wolfowicz, F. Joseph Heremans, Christopher P. Anderson, Shun Kanai,
Hosung Seo, Adam Gali, Giulia Galli, and David D. Awschalom. Quantum
guidelines for solid-state spin defects. Nature Reviews Materials, 6(10):906?925,
apr 2021.
[15] David D. Awschalom, Ronald Hanson, Jo?rg Wrachtrup, and Brian B. Zhou.
Quantum technologies with optically interfaced solid-state spins. Nature Photonics,
12(9):516?527, aug 2018.
[16] J. R. Weber, W. F. Koehl, J. B. Varley, A. Janotti, B. B. Buckley, C. G. Van De
Walle, and D. D. Awschalom. Quantum computing with defects. Proceedings of the
National Academy of Sciences of the United States of America, 107(19):8513?8518,
may 2010.
[17] Matthew E. Trusheim, Benjamin Pingault, Noel H. Wan, Mustafa Gu?ndog?an,
Lorenzo De Santis, Romain Debroux, Dorian Gangloff, Carola Purser, Kevin C.
Chen, Michael Walsh, Joshua J. Rose, Jonas N. Becker, Benjamin Lienhard,
Eric Bersin, Ioannis Paradeisanos, Gang Wang, Dominika Lyzwa, Alejandro R.P.
Montblanch, Girish Malladi, Hassaram Bakhru, Andrea C. Ferrari, Ian A. Walmsley,
Mete Atatu?re, and Dirk Englund. Transform-Limited Photons from a Coherent Tin-
Vacancy Spin in Diamond. Physical Review Letters, 124(2):023602, jan 2020.
[18] Carlo Bradac, Weibo Gao, Jacopo Forneris, Matthew E. Trusheim, and Igor
Aharonovich. Quantum nanophotonics with group IV defects in diamond. Nature
Communications, 10(1):1?13, dec 2019.
[19] Kaoru Sanaka, Alexander Pawlis, Thaddeus D. Ladd, Klaus Lischka, and
Yoshihisa Yamamoto. Indistinguishable photons from independent semiconductor
nanostructures. Physical Review Letters, 103(5), aug 2009.
[20] Kaoru Sanaka, Alexander Pawlis, Thaddeus D. Ladd, Darin J. Sleiter, Klaus
Lischka, and Yoshihisa Yamamoto. Entangling single photons from independently
tuned semiconductor nanoemitters. Nano Letters, 12(9):4611?4616, sep 2012.
[21] Xiayu Linpeng, Maria L.K. Viitaniemi, Aswin Vishnuradhan, Y. Kozuka, Cameron
Johnson, M. Kawasaki, and Kai Mei C. Fu. Coherence Properties of Shallow Donor
Qubits in Zn O. Physical Review Applied, 10(6):064061, dec 2018.
100
[22] S. Strauf, P. Michler, M. Klude, D. Hommel, G. Bacher, and A. Forchel. Quantum
optical studies on individual acceptor bound excitons in a semiconductor. Physical
Review Letters, 89(17):177403/1?177403/4, 2002.
[23] D. D. Sukachev, A. Sipahigil, C. T. Nguyen, M. K. Bhaskar, R. E. Evans, F. Jelezko,
and M. D. Lukin. Silicon-Vacancy Spin Qubit in Diamond: A Quantum Memory
Exceeding 10 ms with Single-Shot State Readout. Physical Review Letters,
119(22):223602, nov 2017.
[24] Xiayu Linpeng, Todd Karin, Mikhail V. Durnev, Mikhail M. Glazov, Ru?diger Schott,
Andreas D. Wieck, Arne Ludwig, and Kai Mei C. Fu. Optical spin control and
coherence properties of acceptor bound holes in strained GaAs. Physical Review B,
103(11):115412, mar 2021.
[25] Mete Atatu?re, Dirk Englund, Nick Vamivakas, Sang Yun Lee, and Joerg Wrachtrup.
Material platforms for spin-based photonic quantum technologies. Nature Reviews
Materials, 3(5):38?51, apr 2018.
[26] Zhouchen Luo, Shuo Sun, Aziz Karasahin, Allan S. Bracker, Samuel G. Carter,
Michael K. Yakes, Daniel Gammon, and Edo Waks. A Spin-Photon Interface
Using Charge-Tunable Quantum Dots Strongly Coupled to a Cavity. Nano Letters,
19(10):7072?7077, oct 2019.
[27] Alisa Javadi, Dapeng Ding, Martin Hayhurst Appel, Sahand Mahmoodian,
Matthias Christian Lo?bl, Immo So?llner, Ru?diger Schott, Camille Papon, Tommaso
Pregnolato, S?ren Stobbe, Leonardo Midolo, Tim Schro?der, Andreas Dirk
Wieck, Arne Ludwig, Richard John Warburton, and Peter Lodahl. Spin-photon
interface and spin-controlled photon switching in a nanobeam waveguide. Nature
Nanotechnology, 13(5):398?403, mar 2018.
[28] K. De Greve, S. M. Clark, D. Sleiter, K. Sanaka, T. D. Ladd, M. Panfilova, A. Pawlis,
K. Lischka, and Y. Yamamoto. Photon antibunching and magnetospectroscopy of a
single fluorine donor in ZnSe. Applied Physics Letters, 97(24):241913, dec 2010.
[29] Alexander Pawlis, Gregor Mussler, Christoph Krause, Benjamin Bennemann,
Uwe Breuer, and Detlev Grutzmacher. MBE Growth and Optical Properties of
Isotopically Purified ZnSe Heterostructures. ACS Applied Electronic Materials,
1(1):44?50, jan 2019.
[30] Erik Kirstein, Evgeny A. Zhukov, Dmitry S. Smirnov, Vitalie Nedelea, Phillip
Greve, Ina V. Kalitukha, Viktor F. Sapega, Alexander Pawlis, Dmitri R. Yakovlev,
Manfred Bayer, and Alex Greilich. Extended spin coherence of the zinc-vacancy
centers in ZnSe with fast optical access. Communications Materials, 2(1):1?8, sep
2021.
[31] Darin J. Sleiter, Kaoru Sanaka, Y. M. Kim, Klaus Lischka, Alexander Pawlis, and
Yoshihisa Yamamoto. Optical pumping of a single electron spin bound to a fluorine
donor in a ZnSe nanostructure. Nano Letters, 13(1):116?120, jan 2013.
101
[32] K. Ohkawa, T. Mitsuyu, and O. Yamazaki. Characteristics of Cl-doped ZnSe layers
grown by molecular-beam epitaxy. Journal of Applied Physics, 62(8):3216?3221,
aug 1987.
[33] H. Morkoc?, S. Strite, G. B. Gao, M. E. Lin, B. Sverdlov, and M. Burns. Large-band-
gap SiC, III-V nitride, and II-VI ZnSe-based semiconductor device technologies.
Journal of Applied Physics, 76(3):1363?1398, aug 1994.
[34] S. Po?ykko?, M. Puska, and R. Nieminen. Chlorine-impurity-related defects in ZnSe.
Physical Review B - Condensed Matter and Materials Physics, 57(19):12164?
12168, may 1998.
[35] N. V. Bondar?, V. V. Tishchenko, and M. S. Brodin. Exciton energy states
and photoluminescence spectra of the strained-layer ZnS-ZnSe superlattices.
Semiconductors, 34(5):568?574, 2000.
[36] Torsten Rieger, Thomas Riedl, Elmar Neumann, Detlev Gru?tzmacher, Jo?rg K.N.
Lindner, and Alexander Pawlis. Strain Compensation in Single ZnSe/CdSe
Quantum Wells: Analytical Model and Experimental Evidence. ACS Applied
Materials and Interfaces, 9(9):8371?8377, mar 2017.
[37] Alexander Pawlis, Gregor Mussler, Christoph Krause, Benjamin Bennemann,
Uwe Breuer, and Detlev Grutzmacher. MBE Growth and Optical Properties of
Isotopically Purified ZnSe Heterostructures. ACS Applied Electronic Materials,
1(1):44?50, jan 2019.
[38] Johanna Jan?en, Felix Hartz, Till Huckemann, Christian Kamphausen, Malte Neul,
Lars R. Schreiber, and Alexander Pawlis. Low-Temperature Ohmic Contacts to
n-ZnSe for all-Electrical Quantum Devices. ACS Applied Electronic Materials,
2(4):898?905, apr 2020.
[39] M. C. Tamargo. Structural characterization of GaAs/ZnSe interfaces. Journal
of Vacuum Science & Technology B: Microelectronics and Nanometer Structures,
6(2):784, jun 1988.
[40] H. Mariette, F. Dalbo, N. Magnea, G. Lentz, and H. Tuffigo. Optical investigation of
confinement and strain effects in CdTe/Cd1-xZnxTe single quantum wells. Physical
Review B, 38(17):12443?12448, dec 1988.
[41] A. V. Kavokin, V. P. Kochereshko, G. R. Posina, I. N. Uraltsev, D. R. Yakovlev,
G. Landwehr, R. N. Bicknell-Tassius, and A. Waag. Effect of the electron
Coulomb potential on hole confinement in II-VI quantum wells. Physical Review B,
46(15):9788?9791, oct 1992.
[42] A. Pawlis, T. Berstermann, C. Bru?ggemann, M. Bombeck, D. Dunker, D. R.
Yakovlev, N. A. Gippius, K. Lischka, and M. Bayer. Exciton states in shallow
ZnSe/(Zn,Mg)Se quantum wells: Interaction of confined and continuum electron
and hole states. Physical Review B - Condensed Matter and Materials Physics,
83(11):115302, mar 2011.
102
[43] P. J. Dean, D. C. Herbert, C. J. Werkhoven, B. J. Fitzpatrick, and R. N.
Bhargava. Donor bound-exciton excited states in zinc selenide. Physical Review
B, 23(10):4888?4901, may 1981.
[44] G. Bastard. Hydrogenic impurity states in a quantum well: A simple model.
Physical Review B, 24(8):4714?4722, oct 1981.
[45] O. Homburg, K. Sebald, P. Michler, J. Gutowski, H. Wenisch, and D. Hommel.
Negatively charged trion in ZnSe single quantum wells with very low electron
densities. Physical Review B - Condensed Matter and Materials Physics,
62(11):7413?7419, sep 2000.
[46] Conrad Spindler, Thomas Galvani, Ludger Wirtz, Germain Rey, and Susanne
Siebentritt. Excitation-intensity dependence of shallow and deep-level
photoluminescence transitions in semiconductors. Journal of Applied Physics,
126(17):175703, nov 2019.
[47] Francisco A.P. Osario, Marcos H. Degani, and Oscar Hipalito. Bound impurity in
GaAs-Ga1-Al As quantum-well wires. Physical Review B, 37(3):1402?1405, jan
1988.
[48] R. T. Senger and K. K. Bajaj. Binding energies of excitons in II-VI compound-
semiconductor based quantum well structures. Physica Status Solidi (B) Basic
Research, 241(8):1896?1900, jul 2004.
[49] Ravitej Uppu, Leonardo Midolo, Xiaoyan Zhou, Jacques Carolan, and Peter Lodahl.
Quantum-dot-based deterministic photon?emitter interfaces for scalable photonic
quantum technology. Nature Nanotechnology, 16(12):1308?1317, oct 2021.
[50] L. J. Rogers, K. D. Jahnke, T. Teraji, L. Marseglia, C. Mu?ller, B. Naydenov,
H. Schauffert, C. Kranz, J. Isoya, L. P. McGuinness, and F. Jelezko.
Multiple intrinsically identical single-photon emitters in the solid state. Nature
Communications, 5:4739, jan 2014.
[51] A. Sipahigil, R. E. Evans, D. D. Sukachev, M. J. Burek, J. Borregaard, M. K.
Bhaskar, C. T. Nguyen, J. L. Pacheco, H. A. Atikian, C. Meuwly, R. M. Camacho,
F. Jelezko, E. Bielejec, H. Park, M. Lonc?ar, and M. D. Lukin. An integrated diamond
nanophotonics platform for quantum-optical networks. Science, 354(6314):847?
850, nov 2016.
[52] R. E. Evans, M. K. Bhaskar, D. D. Sukachev, C. T. Nguyen, A. Sipahigil, M. J.
Burek, B. Machielse, G. H. Zhang, A. S. Zibrov, E. Bielejec, H. Park, M. Lonc?ar,
and M. D. Lukin. Photon-mediated interactions between quantum emitters in a
diamond nanocavity. Science, 362(6415):662?665, nov 2018.
[53] Liang Zhai, Matthias C. Lo?bl, Jan Philipp Jahn, Yongheng Huo, Philipp Treutlein,
Oliver G. Schmidt, Armando Rastelli, and Richard J. Warburton. Large-range
frequency tuning of a narrow-linewidth quantum emitter. Applied Physics Letters,
117(8):083106, aug 2020.
103
[54] J. Puls, G. V. Mikhailov, S. Schwertfeger, D. R. Yakovlev, F. Henneberger, and
W. Faschinger. High-excitation effects in the optical properties of ?-doped ZnSe
quantum wells. Physica Status Solidi (B) Basic Research, 227(2):331?337, 2001.
[55] M. Hayne, A. Usher, A. S. Plaut, and K. Ploog. Optically induced density depletion
of the two-dimensional electron system in GaAs/AlxGa1-xAs heterojunctions.
Physical Review B, 50(23):17208?17216, dec 1994.
[56] E. A. Zhukov, A. Greilich, D. R. Yakovlev, K. V. Kavokin, I. A. Yugova, O. A.
Yugov, D. Suter, G. Karczewski, T. Wojtowicz, J. Kossut, V. V. Petrov, Yu K.
Dolgikh, A. Pawlis, and M. Bayer. All-optical NMR in semiconductors provided
by resonant cooling of nuclear spins interacting with electrons in the resonant spin
amplification regime. Physical Review B - Condensed Matter and Materials Physics,
90(8):085311, aug 2014.
[57] Je Hyung Kim, Christopher J.K. Richardson, Richard P. Leavitt, and Edo Waks.
Two-Photon Interference from the Far-Field Emission of Chip-Integrated Cavity-
Coupled Emitters. Nano Letters, 16(11):7061?7066, nov 2016.
[58] Yi hong Wu, Kenta Arai, and Takafumi Yao. Temperature dependence of the
photoluminescence of ZnSe/ZnS quantum-dot structures. Physical Review B -
Condensed Matter and Materials Physics, 53(16):R10485?R10488, apr 1996.
[59] L. Malikova, Wojciech Krystek, Fred H. Pollak, N. Dai, A. Cavus, and M. Tamargo.
Temperature dependence of the direct gaps of ZnSe and ZnCdSe. Physical Review
B - Condensed Matter and Materials Physics, 54(3):1819?1824, jul 1996.
[60] M. Bayer, G. Ortner, O. Stern, A. Kuther, A. A. Gorbunov, A. Forchel, P. Hawrylak,
S. Fafard, K. Hinzer, T. L. Reinecke, S. N. Walck, J. P. Reithmaier, F. Klopf,
and F. Scha?fer. Fine structure of neutral and charged excitons in self-assembled
In(Ga)As/(Al) GaAs quantum dots. Physical Review B - Condensed Matter and
Materials Physics, 65(19):1953151?19531523, may 2002.
[61] P. J. Dean, W. Stutius, G. F. Neumark, B. J. Fitzpatrick, and R. N. Bhargava.
Ionization energy of the shallow nitrogen acceptor in zinc selenide. Physical Review
B, 27(4):2419?2428, feb 1983.
[62] Youngmin Martin Kim. Fluorine Donor Bound Electron Spins As Qubits. PhD
Dissertation Thesis, pages 1?104, 2013.
[63] M. Oestreich, S. Hallstein, and A. Heberle. Temperature and density dependence
of the electron Lande? g factor in semiconductors. Physical Review B - Condensed
Matter and Materials Physics, 53(12):7911?7916, mar 1996.
[64] Yoshihisa Yamamoto, Charles Santori, and Matthew Pelton. Regulated and
entangled photons from a single quantum dot. Physical Review Letters,
84(11):2513?2516, mar 2000.
104
[65] Jin Liu, Rongbin Su, Yuming Wei, Beimeng Yao, Saimon Filipe Covre da Silva,
Ying Yu, Jake Iles-Smith, Kartik Srinivasan, Armando Rastelli, Juntao Li, and
Xuehua Wang. A solid-state source of strongly entangled photon pairs with high
brightness and indistinguishability. Nature Nanotechnology, 14(6):586?593, apr
2019.
[66] Hui Wang, Hai Hu, T. H. Chung, Jian Qin, Xiaoxia Yang, J. P. Li, R. Z. Liu,
H. S. Zhong, Y. M. He, Xing Ding, Y. H. Deng, Qing Dai, Y. H. Huo, Sven
Ho?fling, Chao Yang Lu, and Jian Wei Pan. On-Demand Semiconductor Source
of Entangled Photons Which Simultaneously Has High Fidelity, Efficiency, and
Indistinguishability. Physical Review Letters, 122(11):113602, mar 2019.
[67] Harishankar Jayakumar, Ana Predojevic?, Thomas Kauten, Tobias Huber, Glenn S.
Solomon, and Gregor Weihs. Time-bin entangled photons from a quantum dot.
Nature Communications, 5(1):1?5, jun 2014.
[68] Maximilian Prilmu?ller, Tobias Huber, Markus Mu?ller, Peter Michler, Gregor Weihs,
and Ana Predojevic?. Hyperentanglement of Photons Emitted by a Quantum Dot.
Physical Review Letters, 121(11):110503, sep 2018.
[69] Christoph Simon, Hugues De Riedmatten, Mikael Afzelius, Nicolas Sangouard,
Hugo Zbinden, and Nicolas Gisin. Quantum repeaters with photon pair sources
and multimode memories. Physical Review Letters, 98(19):190503, may 2007.
[70] Francesco Basso Basset, Mauro Valeri, Emanuele Roccia, Valerio Muredda, Davide
Poderini, Julia Neuwirth, Nicolo? Spagnolo, Michele B. Rota, Gonzalo Carvacho,
Fabio Sciarrino, and Rinaldo Trotta. Quantum key distribution with entangled
photons generated on demand by a quantum dot. Science Advances, 7(12):6379,
mar 2021.
[71] Sergei Slussarenko and Geoff J. Pryde. Photonic quantum information processing:
A concise review. Applied Physics Reviews, 6(4):041303, oct 2019.
[72] Vittorio Giovannetti, Seth Lloyd, and Lorenzo MacCone. Advances in quantum
metrology. Nature Photonics, 5(4):222?229, mar 2011.
[73] Daniel Huber, Marcus Reindl, Johannes Aberl, Armando Rastelli, and Rinaldo
Trotta. Semiconductor quantum dots as an ideal source of polarization-entangled
photon pairs on-demand: A review. Journal of Optics (United Kingdom),
20(7):073002, jun 2018.
[74] Yu Ming He, Oliver Iff, Nils Lundt, Vasilij Baumann, Marcelo Davanco, Kartik
Srinivasan, Sven Ho?fling, and Christian Schneider. Cascaded emission of single
photons from the biexciton in monolayered WSe 2. Nature Communications,
7(1):1?6, nov 2016.
[75] S. Marcet, R. Andre?, and S. Francoeur. Excitons bound to Te isoelectronic
dyads in ZnSe. Physical Review B - Condensed Matter and Materials Physics,
82(23):235309, dec 2010.
105
[76] F. Sarti, G. Mun?oz Matutano, D. Bauer, N. Dotti, S. Bietti, G. Isella, A. Vinattieri,
S. Sanguinetti, and M. Gurioli. Multiexciton complex from extrinsic centers
in AlGaAs epilayers on Ge and Si substrates. Journal of Applied Physics,
114(22):224314, dec 2013.
[77] Nicola Dotti, Francesco Sarti, Sergio Bietti, Alexander Azarov, Andrej Kuznetsov,
Francesco Biccari, Anna Vinattieri, Stefano Sanguinetti, Marco Abbarchi, and
Massimo Gurioli. Germanium-based quantum emitters towards a time-reordering
entanglement scheme with degenerate exciton and biexciton states. Physical Review
B - Condensed Matter and Materials Physics, 91(20):205316, may 2015.
[78] S. M. Ulrich, S. Strauf, P. Michler, G. Bacher, and A. Forchel. Triggered
polarization-correlated photon pairs from a single CdSe quantum dot. Applied
Physics Letters, 83(9):1848?1850, aug 2003.
[79] E. Moreau, I. Robert, L. Manin, V. Thierry-Mieg, J. M. Ge?rard, and I. Abram.
Quantum cascade of photons in semiconductor quantum dots. Physical Review
Letters, 87(18):183601?1?183601?4, oct 2001.
[80] V. D. Kulakovskii, G. Bacher, R. Weigand, T. Ku?mmell, A. Forchel,
E. Borovitskaya, K. Leonardi, and D. Hommel. Fine structure of biexciton emission
in symmetric and asymmetric CdSe/ZnSe single quantum dots. Physical Review
Letters, 82(8):1780?1783, feb 1999.
[81] I. A. Akimov, A. Hundt, T. Flissikowski, and F. Henneberger. Fine structure of the
trion triplet state in a single self-assembled semiconductor quantum dot. Applied
Physics Letters, 81(25):4730?4732, dec 2003.
[82] X. Y. Chang, X. M. Dou, B. Q. Sun, Y. H. Xiong, Z. C. Niu, H. Q. Ni, and D. S.
Jiang. Optical transitions of positively charged excitons and biexcitons in single
InAs quantum dots. Journal of Applied Physics, 106(10):103716, nov 2009.
[83] A. J. Bennett, M. A. Pooley, R. M. Stevenson, M. B. Ward, R. B. Patel, A. Boyer
De La Giroday, N. Sko?d, I. Farrer, C. A. Nicoll, D. A. Ritchie, and A. J. Shields.
Electric-field-induced coherent coupling of the exciton states in a single quantum
dot. Nature Physics, 6(12):947?950, oct 2010.
[84] R. M. Stevenson, R. J. Young, P. See, D. G. Gevaux, K. Cooper, P. Atkinson,
I. Farrer, D. A. Ritchie, and A. J. Shields. Magnetic-field-induced reduction of the
exciton polarization splitting in InAs quantum dots. Physical Review B - Condensed
Matter and Materials Physics, 73(3):033306, jan 2006.
[85] Stefan Seidl, Martin Kroner, Alexander Ho?gele, Khaled Karrai, Richard J.
Warburton, Antonio Badolato, and Pierre M. Petroff. Effect of uniaxial stress on
excitons in a self-assembled quantum dot. Applied Physics Letters, 88(20):203113,
may 2006.
106
[86] John D. Joannopoulos, Steven G. Johnson, Joshua N. Winn, and Robert D. Meade.
Photonic crystals: Molding the flow of light. Photonic Crystals: Molding the Flow
of Light (Second Edition), page 446, 2011.
[87] Juntao Li, Thomas P. White, Liam O?Faolain, Alvaro Gomez-Iglesias, and
Thomas F. Krauss. Systematic design of flat band slow light in photonic crystal
waveguides. Optics Express, 16(9):6227, apr 2008.
[88] Dirk Englund, Ilya Fushman, and Jelena Vuc?kovic?. General recipe for designing
photonic crystal cavities. Optics Express, 13(16):5961, aug 2005.
[89] V. S.C. Manga Rao and S. Hughes. Single quantum-dot Purcell factor and ? factor in
a photonic crystal waveguide. Physical Review B - Condensed Matter and Materials
Physics, 75(20):205437, may 2007.
[90] Sylvain Combrie?, Nguyen Vi Quynh Tran, and Alfredo De Rossi. Directive emission
from high-Q photonic crystal cavities through band folding. Optics InfoBase
Conference Papers, 79(4):041101, jan 2009.
[91] Edo Waks and Jelena Vuckovic. Dipole induced transparency in drop-filter cavity-
waveguide systems. Physical Review Letters, 96(15):153601, 2006.
[92] Shuo Sun, Hyochul Kim, Glenn S. Solomon, and Edo Waks. A quantum phase
switch between a single solid-state spin and a photon. Nature Nanotechnology,
11(6):539?544, feb 2016.
[93] A. R.A. Chalcraft, S. Lam, D. O?Brien, T. F. Krauss, M. Sahin, D. Szymanski,
D. Sanvitto, R. Oulton, M. S. Skolnick, A. M. Fox, D. M. Whittaker, H. Y. Liu, and
M. Hopkinson. Mode structure of the L3 photonic crystal cavity. Applied Physics
Letters, 90(24):241117, jun 2007.
107