Low-dimensional asymmetric crystals: fundamental properties and applications

PDF

Download

Open document

Flip pages

Summarize

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Low-dimensional Asymmetric Crystals: Fundamental Properties and Applications
by
Nan Wang
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(MATERIALS SCIENCE)
December 2023
Copyright 2023 Nan Wang

Dedication
To my family for your endless support and love.
ii

Acknowledgements
Embarking on this scholarly journey has been a profound privilege, and it is with deep gratitude that I
acknowledge the multitude of individuals and institutions who have played an integral role in shaping
this dissertation.
First and foremost, I express my deepest gratitude to my esteemed academic advisor, Prof. Han Wang,
for his support and guidance throughout my doctoral study. His unyielding belief in me not only afforded
me an opportunity to embark on a wonderful academic journey but also cultivated my confidence and
passion for scientific research. More importantly, his wisdom and vision provided me with immense support and freedom to explore research in diverse fields, collaborating with many talented and outstanding
researchers. I also offer my sincere appreciation to my dissertation committee members, Prof. Joshua
Yang, Prof. Rajiv Kalia, Prof. Zhenglu Li, and Yu-Tsun Shao, and my prior qualifying committee members,
Prof. Stephen Cronin, Prof. Paulo Branicio, and Prof. Jayakanth Ravichandran. Their collective expertise,
insightful critiques, and rigorous scholarship have been instrumental in refining the depth and breadth of
this dissertation.
The power of collaboration undoubtedly contributed to the completion of my doctoral study. I am
deeply grateful for the invaluable contributions of my collaborators, including Prof. Jayakanth Ravichandran, Prof. Stephen Cronin, Prof. Michelle Povinelli, Prof. Wei Wu, Prof. Hangbo Zhao, and Prof. Joshua
Yang’s group at USC, and researchers from other universities, Prof. Rohan Mishra’s group. My deepest
gratitude goes out to them for not only granting me access to a diverse array of resources but also for
iii

their invaluable expertise, which played an essential role in bringing this work to fruition. Especially, I
would like to thank Prof. Jayakanth Ravichandran for his constructive guidance, and his team members,
Dr. Huandong Chen and Boyang Zhao for their generous dedication to my projects. In addition, I thank
Phillip Sliwoski who taught me how to make and seal quartz tubes for materials growth.
The camaraderie and intellectual exchange with my colleagues and fellow graduate students have been
a wellspring of inspiration. A special acknowledgment goes to Dr. Jiangbin Wu and Dr. Max Lien for their
insightful discussions and invaluable assistance in the collaboration of multiple projects, without which
this work would not have been possible. I also truly express my gratitude to my colleagues, including
Xiaodong Yan, Huan Zhao, Hefei Liu, Zhonghao Du, Jiahui Ma, Silvia Guadagnini, Hung-yu Chen, Hongming Zhang, Ting-Hao Hsu, and Jian Zhao. Through shared challenges and mutual encouragement, we
have forged bonds that I know will endure far beyond this academic journey.
I also would like to extend my heartfelt gratitude to my friends, Jiahao, Tao Jun, Zerui, Deming, TseHsien, Zhiyuan, Mingrui, Huandong, Boyang, Pan Hu, and Bofan, for their unwavering support and companionship throughout my doctoral journey. They stood by my side, offering encouragement during the
challenging moments, and celebrating the milestones with genuine enthusiasm. Their presence provided
a sense of comfort and reassurance, reminding me that I was not alone in this pursuit.
Last but not least, this dissertation is dedicated to my family. I can never express enough gratitude to
my parents. Their unwavering and boundless love, support, and sacrifice have provided me with the opportunity to pursue knowledge and achieve my dreams. I am profoundly grateful for the values you instilled
in me, which have guided me through this challenging yet rewarding path. I also thank my parents-in-law.
Their unselfish assistance allowed me and my wife the time and space to focus on our academic endeavors.
Especially, I would like to dedicate an acknowledgment to someone whose unwavering support and love
have been the foundation upon which this dissertation was built - my wife, Ting Zhao. Throughout this
challenging and arduous journey, you have been my rock, my confidante, and my source of strength. Your
iv

endless patience, understanding, and encouragement have sustained me during the long hours of research
and writing. Your belief in my abilities, even when I doubted myself, has been a constant source of motivation. The arrival of our baby has added a new dimension to our lives, filling each day with immeasurable
joy and inspiration. The resilience and grace with which you manage the responsibilities of motherhood
while continuing to be my greatest supporter are nothing short of extraordinary. The resilience and grace
with which you manage the responsibilities of motherhood while continuing to be my greatest supporter
are nothing short of extraordinary. Our little baby, Melissa, has brought immeasurable joy to our lives and
serves as a constant reminder of the beautiful tapestry we are weaving together. I love you all!
v

Table of Contents
Dedication
Acknowledgements
List of Tables
List of Figures
Abstract
Chapter 1: Introduction
1.1 Background and motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Double helix nanostructures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Quasi one-dimensional perovskite chalcogenide – BaTiS3 . . . . . . . . . . . . . . . . . . . 5
1.4 Black Phosphorus Photodetectors Integrated with Metasurfaces . . . . . . . . . . . . . . . 7
Chapter 2: Polymer-like Mechanical Properties in Inorganic Double Helical Van Der Waals
Semiconductor
2.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3 Crystal synthesis of SnIP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4 Crystal structure characterization of SnIP . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.4.1 Crystal structure and morphology of SnIP . . . . . . . . . . . . . . . . . . . . . . . 15
2.4.2 Raman and photoluminescence measurements on SnIP . . . . . . . . . . . . . . . . 18
2.5 SnIP Young’s modulus measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.5.1 AFM-based nanomechanical bending test for SnIP Young’s modulus . . . . . . . . 19
2.5.2 Brillion scattering spectroscopy for SnIP Young’s modulus . . . . . . . . . . . . . . 23
2.6 Flexibility of the SnIP nanowire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.7 Deformability of the SnIP nanowire . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.8 Summary and outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
Chapter 3: Infrared Polarization Photodetection and Imaging of Quasi One-Dimensional
Perovskite Chalcogenide BaTiS3
3.1 Absract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.3 Anisotropic characterization of BaTiS3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.3.1 Structural anisotropy in BaTiS3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.3.2 Optical anisotropy in BaTiS3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
vi
ii
iii
viii
ix
xi
1
11
33
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3.4 Polarized optoelectronic responses for BaTiS3 photodetectors . . . . . . . . . . . . . . . . 39
3.5 Trap-induced photogating effect in BaTiS3 . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.6 Linearly infrared polarization imaging based on BaTiS3 . . . . . . . . . . . . . . . . . . . . 42
3.7 Summary and outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Chapter 4: Black Phosphorus Molybdenum Disulfide Midwave Infrared Photodiodes with
Broadband Absorption-Increasing Metasurfaces
4.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.3 Fabrication of BP-MoS2 Photodiodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.4 Integrating MIM Metasurface Gratings with BP-MoS2 Photodiodes . . . . . . . . . . . . . 49
4.5 Evaluation of MIM Metasurfaces with BP-MoS2 Photodiodes . . . . . . . . . . . . . . . . . 53
4.6 Summary and outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Chapter 5: An All-Silicon Metalens Integrated with a Mid-Wave Infrared Black Phosphorus
Photodiode
5.1 Abstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
5.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
5.3 Design and fabrication of Silicon metalens . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.4 Fabrication and characterization of BP-MoS2 photodiodes integrated with the metalens . . 67
5.5 Summary and outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Bibliography
vii
45
58
72
. . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

List of Tables
4.1 Structural Parameters of Each Device in Figure 4.3 . . . . . . . . . . . . . . . . . . . . . . . 52
4.2 Responsivity R and Specific Detectivity D* at V DS = -0.5 V for Device 1 and 2 with and
without the Metasurface Grating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.3 Comparison of Nanostructure-Integrated Thin-Film BP Photodetectors and Bulk Material
Photodiodes at Room Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.1 Simulated and measured focal spot locations in z and Enhancement at z = 300 µm. . . . . . 67
viii

List of Figures
1.1 Crystal structure of double helical SnIP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Structure of BaTiS3 single crystal. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 Structure of BaTiS3 single crystal. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.1 Optical images showing the synthetic process . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2 Bulk and nanosized crystals of double helical SnIP . . . . . . . . . . . . . . . . . . . . . . . 16
2.3 SnIP crystal structure and characterization. . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4 SnIP crystal structure and characterization. . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.5 Sample fabrication for the AFM-based nanomechanical bending test. . . . . . . . . . . . . 20
2.6 SnIP Young’s modulus from nanomechanical bending test. . . . . . . . . . . . . . . . . . . 22
2.7 SnIP Young’s modulus from Brillion scattering spectroscopy measurement. . . . . . . . . . 24
2.8 Benchmark of Young’s modulus among different material classes. . . . . . . . . . . . . . . 25
2.9 Flexibility of the SnIP nanowire. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.10 The Raman spectra of SnIP before and after bending test. . . . . . . . . . . . . . . . . . . . 28
2.11 Figure of merit for flexibility of materials. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
2.12 Deformability of the SnIP nanowire. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.13 Benchmark of the deformability in materials of different electronic bandgaps. . . . . . . . 31
3.1 Structural anisotropy in BaTiS3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.2 Optical anisotropy in BaTiS3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
ix

3.3 Polarized optoelectronic responses for BaTiS3 photodetectors . . . . . . . . . . . . . . . . 39
3.4 Polarized optoelectronic responses for BaTiS3 photodetectors . . . . . . . . . . . . . . . . 40
3.5 Trap-induced photogating effect in BaTiS3 . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.6 Schematic diagram of the imaging system . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
3.7 Linearly infrared polarization imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.1 Metasurface-integrated black phosphorus photodiode. . . . . . . . . . . . . . . . . . . . . . 48
4.2 Double- and triple-resonator metasurface gratings. . . . . . . . . . . . . . . . . . . . . . . 50
4.3 Absorption spectra of metasurface gratings with varying geometry. . . . . . . . . . . . . . 51
4.4 Room temperature photocurrent enhancement in two devices. . . . . . . . . . . . . . . . . 54
4.5 Room temperature photocurrent enhancement in two devices. . . . . . . . . . . . . . . . . 55
5.1 Conceptual diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
5.2 Mid-wave infrared silicon metalens design. . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
5.3 Metalens characterization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
5.4 Measured and simulated focal spots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
5.5 Metalens-integrated black phosphorus photodiodes. . . . . . . . . . . . . . . . . . . . . . . 68
5.6 Room temperature metalens-enhanced photodetection. . . . . . . . . . . . . . . . . . . . . 69
x

Abstract
As Moore’s law predicted, conventional silicon-based devices are reaching their physical limit with shrinking devices. Low-dimensional materials bring out a new breakthrough in the semiconductor industry to
extend Moore’s law, because of the atomic-level size characteristics, the surfaces free from defects, and the
sensitivity to electrical and optical control. The reduced dimensionality provides the opportunity to take
advantage of atomically thick layers, leading to unique and extraordinary optical, electronic, mechanical,
thermal, and chemical properties. The potential makes it possible to consider their use in a wide range of
nanotechnologies, including optoelectronics, biological sciences, energy harvesting, and chemical sensing.
As a promising candidate, low-dimensional asymmetric semiconducting materials are a new class of
semiconductors with low symmetry in their crystal structures. These highly asymmetric crystal structures mainly include orthorhombic, monoclinic, and triclinic systems. Due to the crystal asymmetry,
low-dimensional asymmetric nanostructures exhibit unique anisotropic physical characteristics in mechanical, optics, and electronics, which include confined or anisotropic transports, excellent flexibility and
deformability in mechanics, strong anisotropy in optical absorption, and light emission with controlled
directionality. The novel physical characteristics stemming from the asymmetric crystal structure unlock new possibilities for scientists to develop and design innovative semiconductor device applications
in highly integrated photonic, electronic, optoelectronic, and especially flexible and stretchable electronic
xi

systems. This dissertation focuses on the fundamental properties and potential applications of three different low-dimensional asymmetric semiconducting materials, including Tin iodide phosphide (SnIP), Barium
titanium sulfide (BaTiS3), and black phosphorus (BP).
Firstly, We studied the mechanical properties of the newly discovered inorganic double helical semiconductor tin indium phosphate (SnIP). Revealed through both the nanomechanical testing and the Brillouin scattering spectroscopy, this spiral-shaped double helical crystal shows the lowest Young’s modulus
(13.6 GPa) among all known stable inorganic materials. The large elastic (>27%) and plastic (>60%) bending strain are also observed due to the easy slippage between neighboring double helices that are coupled
through van der Waals interactions, leading to the high flexibility and deformability among known semiconducting materials. The results advance the fundamental understanding of the unique polymer-like
mechanical properties in this distinctive class of quasi-one-dimensional crystals and lay the foundation
for their potential applications in a broad range of flexible electronics and nanomechanics disciplines.
Secondly, we explored the potential of BaTiS3 in anisotropic infrared photodetection and polarization
imaging, we fabricated the infrared photodetector based on exfoliated BaTiS3 flakes, and demonstrated that
the BaTiS3 photodetector shows a giant anisotropy in broadband optoelectronic responses from near to
mid-infrared spectrum. Also, this BaTiS3 photodetector shows the different polarization ratios at different
wavelengths. This strong difference in the photodetection anisotropy is clearly observed to demonstrate
that the BTS photodetector can identify the slight differences between different polarized illuminations. In
addition, the frequency dependence of BaTiS3 photodetector shows the dynamic photocurrent as a function of frequency in the range from 200 Hz to 10 kHz. This behavior suggests that the dominant mechanism
at the illumination of infrared light is the trap-induced photogating effect. A broadband-sensitive photoresponse with excellent stability without degradation under ambient atmospheric conditions will help the
polarized infrared imaging for BaTiS3 realized at room temperature. Importantly, a large anisotropic ratio
of BaTiS3 ensures polarized imaging in a scattering environment, opening up possibilities for developing
xii

next-generation polarized infrared imaging technology. Therefore, we systematically investigated linearly
polarized infrared imaging for a designed target using an anisotropic BaTiS3 photodetector.
Thirdly, in order to enable thin-film, high-absorption BP photodetectors, we developed a resonant
metal-insulator-metal (MIM) grating-enhanced thin-film BP-MoS2 photodiode. We first integrated a single resonator MIM structure with a BP–MoS2 photodiode. Our results show that this single-resonator
MIM can increase the absorption around a selected mid-infrared wavelength. After the integration of a
single-resonator MIM structure, the generated photocurrent increased by 6.7 times at 3.39 µm. Then, we
designed and fabricated multiple-resonator MIM structures with broadband absorption spectra that can
be controlled by tuning their geometry. Finally, we integrated broadband MIM structures with BP–MoS2
photodiode and evaluated them by measuring the room temperature responsivity and specific detectivity of BP-MoS2 photodiodes at multiple MWIR wavelengths. We demonstrated that the broadband MIM
structures resulted in enhanced room temperature performance over a broad range of wavelengths. After
the integration of a triple-resonator MIM structure, the photocurrent increased by 7.5 times at 3.39 µm and
by 12.8 times at 3.88 µm. In a word, our results show that broadband metasurface gratings are a scalable
approach for boosting the performance of BP photodiodes over large spectral ranges.
Lastly, to enhance the optical signal collection in BP devices, we directly integrated a designed allsilicon metalens with a BP-MoS2 heterojunction photodiode and experimentally demonstrated its capability to increase photodetector performance in the mid-wave infrared wavelengths. We first designed and
fabricated silicon metalens with cylindrical nanorod unit cells. Then, we measured the silicon metalens by
imaging clear focal spots at various testing wavelengths. The results showed that the silicon metalens can
produce focal spots in the mid-wave infrared wavelengths and increase power density by concentrating
light into a small absorber region. Then, we integrated a BP-MoS2 photodetector onto the same substrate
and demonstrated that the metalens boosts MWIR performance at room temperature. Furthermore, we
integrated a thin-film BP-MoS2 heterojunction photodiode with the silicon metalens. The results showed
xiii

that the metalens boost room temperature performance at multiple testing mid-wave infrared wavelengths.
Photocurrent and responsivity measured at 3.39 µm light focused by the metalens was ≈ 6.65 times higher
than in the case with no metalens. Additionally, photocurrent and responsivity were increased by ≈6.31
times at 3.88 µm in the presence of the metalens. These results demonstrate that integrated metasurface
lenses are an excellent approach for boosting the performance of MWIR BP photodetectors.
xiv

Chapter 1
Introduction
1.1 Background and motivation
With the rapid development of the information society, the efficient distribution, manipulation, and integration of information put forward an ever-increasing urgent need for miniaturization and integration of
logic devices. As Moore’s law predicted, however, conventional silicon-based devices are reaching their
physical limit with shrinking devices. As devices shrink, they start to suffer from surface instabilities,
inadequate carrier mobility, and short-channel effects. To overcome the limitations of further miniaturization, developing different types of emerging materials and advanced electronics is needed to realize the
potential of the next generation of computers.
Low-dimensional materials bring out a new breakthrough in the semiconductor industry to extend
Moore’s law, because of the atomic-level size characteristics, the surfaces free from defects, and the sensitivity to electrical and optical control [1]. Compared to their 3D counterparts, the reduced dimensionality
provides the opportunity to take advantage of atomically thick layers, leading to unique and extraordinary optical, electronic, mechanical, thermal, and chemical properties [2, 3]. The potential of enabling us
to exploit all these properties makes it possible to consider their use in a wide range of nanotechnologies,
including optoelectronics, biological sciences, energy harvesting, and chemical sensing [4, 5]. Although
1

many of common low-dimensional materials, such as graphene, h-BN, and MoS2, show outstanding physical performance in various fields, their relatively symmetric crystal structures result in isotropic material
properties. Low-dimensional asymmetric semiconducting materials are a new class of semiconductors with
low symmetry in their crystal structures. These highly asymmetric crystal structures mainly exist in orthorhombic, monoclinic, and triclinic systems. Due to the crystal asymmetry, low-dimensional asymmetric
nanostructures not only possess excellent properties comparable to those of most studied isotropic materials, but also exhibit unique anisotropic physical characteristics in mechanical, optics, and electronics, [6,
7, 8] which include confined or anisotropic transports, excellent flexibility and deformability in mechanics,
strong anisotropy in optical absorption, and light emission with controlled directionality. The novel physical characteristics stemming from the asymmetric crystal structure unlock new possibilities for scientists
to develop and design innovative semiconductor device applications in highly integrated photonic, electronic, optoelectronic, and especially flexible and stretchable electronic systems. This dissertation focuses
on the fundamental properties and potential applications of three different low-dimensional asymmetric
semiconducting materials, including Tin iodide phosphide (SnIP), Barium titanium sulfide (BaTiS3), and
black phosphorus (BP). [9, 10, 11, 12].
1.2 Double helix nanostructures
The double helical structure has always been of central interest to a broad range of scientific and engineering disciplines from the architectural marvel of Bramante Staircase in the Vatican City to the molecular
structure of deoxyribonucleic acid (DNA) [13, 14, 15]. In particular, this unique geometrical structure has
been of critical importance in genetics and genomics research where the double helix consisting of sugar
phosphates forms the backbone of DNA molecules that are essential for encoding the genetic information
in all living species. [16, 17] In the inorganic materials world, the double helix is one of the most fascinating and landmark quasi-one-dimensional nanostructures. A double helical structure can be formed by
2

Figure 1.1: Crystal structure of double helical SnIP. a. Atomic structure of SnIP along the a and b axis.
Tin-iodide and phosphorus helix are forming a double helical SnIP strand. The bulk b and c the single rod
band structure of SnIP. Band gaps are 1.79 and 2.28 eV, respectively. Figures adopted from reports [23].
twining two single helices with the same axis or a translation along the axis together. This nanostructure
can be either right-handed or left-handed depending on the direction of the rotation. Due to the unique
organization of the two single helices and their synergistic effect, the double-helical structure is expected
to exhibit exceptional and even rather different properties, which have attracted intensive attention to
their amazing morphology-related potential applications [18, 19, 20, 21, 22]. Inorganic solid-state crystals
with analogous double helical atomic lattices, however, have remained largely unexplored by the physical
science community.
3

A few very complicated organic-inorganic hybrid material frameworks were synthesized in the 80s and
90s of the last century, which possess double-helix-like structures at the molecular level but never achieved
crystalline forms. [24, 25] Some purely inorganic double helical crystals were theoretically predicted in
the early 80s[26], but none have been experimentally proven to exhibit double helix crystal lattice[27].
In the crystal, the double helical structure atomic structure would provide a lower Young’s modulus and
higher elasticity. Unlike the double-helical nanostructures formed through the physical winding of carbon
nanotubes, nanowires, and 2D materials[28, 29, 22, 30, 31], such double-helical crystals are expected to
consist of atomic-scale double helices orderly arranged in a highly crystalline manner. Each individual
double helix unit in the crystal lattice would consist of two atomic spiral strands with identical pitches
intertwined about a common axis at the fundamental atomic scale. Tin iodide phosphide (SnIP) recently
surprised the inorganic materials community as the first carbon-free atomic-scale double helix [23]. SnIP
is a quasi-one-dimensional semiconducting van der Waals crystal with atomic-scale double helical nanostructure. The SnIP bulk material can be regarded as a bundle of individual double helices, with each double
helix containing a negatively charged molecular phosphorus helix, surrounded by a positively charged tiniodine helix (Figure 1.1a). The inner and outer strands of an individual double helix have the same chirality,
i.e. either left (M-helix) or right-handed (P-helix). Each two-winding cycle of the intertwined strand pair
contains seven SnIP units, which leads to a 7/2 helix. In the most thermodynamically stable form, alternating M and P helices bind together through van der Waals (vdW) interaction to form the bulk crystal lattice
structure. The unique crystal structure with the covalent bonding within a helix chain and weak van der
Waals interactions between the double helices results in reduced symmetry in its mechanical properties.
According to the DFT calculation[23], the band gap energy for bulk SnIP bundles is 1.79 eV, which differs
from the bandgap of 2.28 eV for a single SnIP double helix by 0.49 eV (Figure 1.1b and 1.1c), suggesting
that the band gap of SnIP double helix bundles can be discretely controlled by the number of individual
double helices in a SnIP bundle. Furthermore, SnIP is estimated to have a high electron mobility [18].
4

Due to the unique crystal structure and excellent predicted electronic properties, SnIP is expected to be
an extremely soft and flexible semiconductor, which has a promising potential for applications in flexible
electronics. Therefore, it is significant to advance its mechanical properties, which will lay the foundation
for its potential applications in flexible electronics and nanomechanics disciplines.
1.3 Quasi one-dimensional perovskite chalcogenide – BaTiS3
Optical anisotropy plays a crucial role in photonic and optoelectronic applications, including polarization
control, optical communication, and nonlinear optics. In the past, the mechanical strain was applied in
materials to dynamically control the degree of optical anisotropy, [32, 33] but most materials possess limited strain capability, which will reduce the effectiveness of strain-induced optical anisotropy. Besides,
anisotropic metamaterial or metasurface architectures were designed to satisfy the requirement of large
optical anisotropy, but optical losses and fabrication challenges block their practical use. [34, 35] Among
many crystal structures, perovskite chalcogenides are an emerging class of inorganic semiconductors with
exciting electronic and optical properties, such as ultrahigh light absorption coefficients in the visible to infrared spectrum, good carrier mobility, and small effective masses of charge carriers, which make promising for technological applications in photovoltaics, infrared Optoelectronics, and thermoelectrics [36].
Quasi-one-dimensional crystal structures are made up of one-dimensional substructures with strong covalent bonding which combines with each other together by weak electrostatic interactions [37]. The unique
atomic arrangement induces asymmetry in their crystal structure and their physical properties. Quasione-dimensional perovskite chalcogenides combine the characteristics of perovskite structures, chalcogen
elements, and a one-dimensional structural arrangement. Therefore, quasi-one-dimensional perovskite
chalcogenides are expected to achieve large and accessible in-plane anisotropy.
Barium titanium sulfide (BaTiS3), a hexagonal perovskite chalcogenide with a quasi-one-dimensional
crystal structure, recently become a rising star in the family of quasi-one-dimensional nanostructures
5

Figure 1.2: Structure of BaTiS3 single crystal. a. Illustration of the BaTiS3 structure in perspective view
along the c axis. Blue and orange spheres denote Ba and S atoms, respectively. Octahedra formed by Ti
and six surrounding S atoms are highlighted in green. b. Illustration of the BaTiS3 structure in perspective
view projected down an axis. The high-resolution STEM HAADF images with overlaid atomic models
showing c axis c and an axis d projection of BaTiS3. Figures adopted from reports [38].
due to the giant optical anisotropy [39]. BaTiS3 crystallizes in the polar non-centrosymmetric P63mc
space group, where TiS6 octahedra sharing common faces form the parallel quasi-one-dimensional lattice chains propagating along the c axis and these chains are separated from each other in the ab plane
by Ba2+ cations [39], as shown in Figure 1.2a and 1.2b. Figures 1.2c and 1.2d show the wide field-of-view
high-angle annular dark-field (HAADF) images of a BaTiS3 crystal along with c and axis projections, respectively [38]. Therefore, this quasi-1D structure features easily large and accessible in-plane anisotropy.
In addition, the intra-interaction within the TiS6 cell is the strong covalent bonding while the inter-chain
6

direction is weak ionic bonding, which makes bulk BaTiS3 exfoliated into thin flakes with various thicknesses. Cleaving quasi-one-dimensional bulk crystals reduces the dimensionality of material effectively,
which leads to robust surfaces that retain covalent bonding in atomic columns and thus make the potential of the material for device miniaturization and application in integrated circuits. More interestingly,
this unique quasi-one-dimensional crystal structure BaTiS3 results in the giant optical anisotropy, triggering strong linear dichroism and tremendous birefringence [39, 40, 41]. Moreover, BaTiS3 is a small
direct bandgap semiconductor. The resulting unique optical and electronic properties of BaTiS3 enable the
wavelength-tunable and polarization-sensitive miniaturized photonic devices desired for emerging applications in polarization-sensitive sensing, multispectral imaging, and optical communication. Therefore, it
is of scientific and technological importance to explore the performance of BaTiS3 in infrared polarized
photodetection and imaging.
1.4 Black Phosphorus Photodetectors Integrated with Metasurfaces
Mid-wave infrared (MWIR) photodetectors, which operate in the wavelength range of 3 to 5 µm, have a
wide range of practical applications including optical communication, infrared astronomy, and environmental monitoring. [42, 43] Recently, black phosphorus (BP) has gained significant attention as a promising candidate material for MWIR photodetection due to its unique structure and intriguing properties. [44,
45] BP is a layered material with an orthorhombic crystalline structure. The phosphorus atoms interact
with each other via strong covalent bonding, forming a special puckered honeycomb structure, as shown
in Figure 1.3a. [46] The unusual atomic arrangement of BP reduces the crystalline symmetry, resulting
in the distinct band structure and carrier effective mass between the two perpendicular in-plane crystal
directions, which gives rise to its remarkable in-plane electronic and optical anisotropy. [47, 48, 49] Adjacent layers of BP interact with each other via weak van der Waals (vdW) forces and are arranged in
an AB stacking order with an interlayer distance of 5.3 Å. Hence, BP can be suitably integrated with the
7

Figure 1.3: a. Layered crystal structure of black phosphorus with two adjacent puckered sheets. b. Diagram of the black phosphorus MWIR photodetector operating at 3.39 µm. Figure adopted from report
[45]. c. Resonant MIM metastructure integrating with BP photodetector. Metasurface gratings boost the
performance of BP photodetectors over large spectral ranges. d. Conceptual diagram of a metalens that
collects incident light and focuses it onto the active area.
silicon platform and processed with the standard nanofabrication, bringing about excellent compatibility
with complementary metal-oxide-semiconductor (CMOS) technology. In addition, BP has a moderate direct bandgap (0.3 eV / 4.1 µm in bulk form) and high carrier mobility (1000 cm2 V
−1
s
−1
). [46, 50] The
unique crystal structure and excellent properties enable BP MWIR photodetectors with high responsivities
and specific detectivities at room temperature (Figure 1.3b). [45] Furthermore, the bandgap location of BP
has been tuned by arsenic alloying, the number of layers, and vertical electric fields, which has enabled
spectrally tunable MWIR photodetectors and optics. [51, 52, 53, 54, 55, 56] However, due to the screening effect, bandgap tunability is only possible with small film thicknesses (up to ∼40 nm), which limits a
photodetector’s absorption. [49, 57]
8

To break this limit, researchers have tried a variety of methods to enable high-performance BP devices.
For example, BP photodetectors have achieved a dramatic increase in their photoresponse by utilizing photogating and gate voltage tuning effects. [45, 53] Additionally, integrated nanostructures such as waveguides, plasmonic apertures and gratings, and Fabry-Pérot cavities have experimentally shown to increase
performance in BP photodetectors by increasing the absorption path of incident light. [58, 59, 60, 61, 62,
63, 64, 65, 66] While the nanostructures do not increase the intrinsic absorption of BP (i.e., the imaginary
part of its refractive index), they increase the amount of incident radiation that is collected, which in turn
boosts the device’s signal-to-noise ratio. [63] Furthermore, these nanostructures are often designed to
boost photodetection around specific wavelengths such as telecom bands, but many applications require a
high photoresponse across large ranges of wavelengths. Additionally, devices with small active areas are
desirable for dense arrays because of reduced parasitic effects in the absorber material. However, the small
active area also reduces the optical collection area, resulting in low efficiency in optical absorption.
In a previous study, MIM structures were numerically predicted to increase the amount of light absorbed in thin-film BP (Figure 1.3c). [67] Metal-insulator-metal (MIM) structures consist of a nonmetallic
material that is placed between a metallic back reflector and periodic metallic nanostructure. [68] An MIM
grating creates a lateral resonant mode within the nonmetallic material, and the spectral location of the
resonance depends on the grating’s structural parameters (e.g., dielectric thickness, periodicity). [69, 70]
Incident light then couples to the mode, which recirculates within the structure and increases the amount
of light absorbed by the nonmetallic material. [69] In addition, metalens is a type of lens that uses nanoscale
structures, often made of metamaterials or engineered nanostructures, to focus light (Figure 1.3d). Unlike
traditional lenses, which use curved surfaces to refract light, metalenses use flat surfaces with specifically
designed patterns to manipulate the phase and amplitude of light waves. Metalenses have been established
as planar arrays of subwavelength nanostructures that offer vast degrees of freedom for controlling the
phase, transmission, and polarization of light. In the context of an array, individual metalenses could be
9

fabricated for each pixel, and the concentrated incident light also allows for smaller device pitches.[71, 72,
73] Previous work has demonstrated this concept in a single-pixel III-V detector system. [72] Therefore,
those two nanostructures are expected to boost the performance of BP photodetectors.
10

Chapter 2
Polymer-like Mechanical Properties in Inorganic Double Helical Van
Der Waals Semiconductor
2.1 Abstract
In high-performance flexible and stretchable electronic devices, conventional inorganic semiconductors
made of rigid and brittle materials typically need to be configured into geometrically deformable formats
and integrated with elastomeric substrates, which leads to challenges in scaling down device dimensions
and complexities in device fabrication and integration. Here we report the extraordinary mechanical properties of the newly discovered inorganic double helical semiconductor tin indium phosphate. Revealed
through both the nanomechanical testing and the Brillouin scattering spectroscopy, this spiral-shaped
double helical crystal shows the lowest Young’s modulus (13.6 GPa) among all known stable inorganic
materials. The large elastic (> 27%) and plastic (> 60%) bending strains are also observed and attributed to
the easy slippage between neighboring double helices that are coupled through van der Waals interactions,
leading to the high flexibility and deformability among known semiconducting materials. The results advance the fundamental understanding of the unique polymer-like mechanical properties in this distinctive
class of quasi-one-dimensional crystals and lay the foundation for their potential applications in flexible
electronics and nanomechanics disciplines.
11

2.2 Introduction
Compared with organic semiconductors, inorganic semiconductors have significantly higher field-effect
mobilities and long-term stability, which are desirable for high-performance flexible and stretchable electronics. However, due to their high Young’s modulus, small yield strain, and fracture limit, inorganic semiconductors typically require large deformation forces and have small elastic deformation ranges. Such intrinsically nonbendable and nonstretchable characteristics of inorganic semiconductors significantly limit
their applications in flexible and stretchable electronics where large deformations are involved. To overcome these limitations, several approaches have been presented. For example, the size-dependent mechanical properties of one-dimensional (1D) semiconductors such as silicon nanowires were utilized to
achieve enhanced flexibility as compared to their large-scale counterparts. Moreover, large numbers of
nanowires were used in a single device to make sure the performance would not degrade quickly under repeated bending or stretching. Another approach is leveraging structural designs to form nanoscale
semiconductor thin films into out-of-plane buckled structures or in-plane serpentine structures to accommodate overall structure deformations while avoiding substantial strains in the semiconductor material
itself. However, those approaches yield limited flexibility and stretchability and are challenging to scale
down key device dimensions below tens or hundreds of micrometers, limiting their use in devices with increasing spatial resolution and integration density. Therefore, inorganic semiconductors with polymer-like
mechanical properties, such as low Young’s modulus and high intrinsic stretchability, are highly desirable
for next-generation flexible and stretchable electronic devices.
Double helical structure has always been of central interest to a broad range of scientific and engineering disciplines from the architectural marvel of Bramante Staircase in the Vatican City to the molecular
structure of deoxyribonucleic acid (DNA). In the crystal, the atomic double helical structure would provide
a lower Young’s modulus and higher elasticity. Unlike the double helical nanostructures formed through
the physical winding of carbon nanotubes, nanowires, and 2D materials, such double-helical crystals are
12

expected to consist of atomic-scale double helices orderly arranged in a highly crystalline manner. Each
individual double helix unit in the crystal lattice would consist of two atomic spiral strands with identical
pitches intertwined about a common axis at the fundamental atomic scale. However, inorganic solid-state
crystals with analogous double helical atomic lattices have remained largely unexplored by the physical
science community.
In this work, we studied the mechanical properties (including Young’s modulus, flexibility, and deformability) of the newly discovered semiconducting, carbon-free inorganic double helical van der Waals
crystal tin indium phosphate (SnIP). Comparing Young’s modulus of the SnIP crystal with other air-stable
materials, its Young’s modulus along the double helical strand axis, which has been independently confirmed by both the nanomechanical bending test and the Brillouin scattering measurement, is significantly
smaller than any other known inorganic materials, including all the ceramics, metals, III-V compounds,
as well as low-dimensional materials such as carbon nanotubes, nanowire crystals and all the known 2D
materials. Young’s modulus of SnIP is even comparable to or lower than many organic materials. Here, we
attribute the ultra-low Young’s modulus of SnIP to the unique double helix lattice structure, in which the
strong inter-atomic bonding between the P atoms in the inner strands and between the Sn and I atoms in
the outer strands are along the tangential direction of the helices, leading to the significantly relaxed interactions along the a-axis, analogous to atomic-scale mechanical springs. In addition, a large elastic bending
strain (> 27%) is observed via bending tests, resulting in the high flexibility of SnIP comparable to the polymer. Furthermore, out of the yield strain limit, a high plastic bend strain (> 60%) is also recorded, leading
to a high deformability comparable to metal. Density-functional theory (DFT) calculation reveals that the
low slipping energy between van der Waals coupled neighboring double helices is the reason for high flexibility and deformability. The results advance the fundamental understanding of the unique polymer-like
mechanical properties in this distinctive class of quasi-one-dimensional crystals and lay the foundation
for their potential applications in a broad range of flexible electronics and nanomechanics disciplines.
13

2.3 Crystal synthesis of SnIP
The intriguing double helix structure motivates a significant interest in introducing it into the domain of
inorganic materials. Nilges and co-workers first reported the inorganic atomically precise double helical
structure (SnIP) in 2016.[23] The surprising synthesis of SnIP, featuring an unparalleled DNA-like double
helical configuration, underscores the atomically precise design of nanomaterials. According to the formation mechanisms for SnIP, the SnIP is most probably synthesized via the chemical vapor transport (CVT)
reaction. CVT is an efficient method to synthesize high-quality crystals. The selection of appropriate reactive species and the establishment of a temperature profile play vital roles in transport reactions. Under
equimolar element ratios, SnIP is synthesized in a vapor phase reaction of gaseous white phosphorus (P4)
and SnI2. In the SnIP reaction, two SnI2 molecules can be dimerized spontaneously to form a Sn2I4 dimer.
Further, this Sn2I4 dimer can crucially activate the P4 molecule by attracting the P lone pairs (SnIP), which
leads to successive helix growth[74].
Based on the CVT reaction and the formation mechanism, SnIP can be synthesized efficiently. Figure
2.1a shows the schematic diagram for the synthesis process of SnIP. Precursor materials, Tin (99.995%,
Alfa Aesar), SnI2 (99.999%, Sigma-Aldrich), and red phosphorus (99.99%, Sigma-Aldrich) were mixed in
stoichiometric proportions (Sn:I:P = 1:1:1) with a total weight of 0.4 g and finely ground in an argon-filled
environment to form a homogeneous powder, which was then sealed into an evacuated quartz ampoule (3
mTorr). The ampoule was heated up to 660 °C at a 1.4 °C min−1
ramp rate and held at the reaction temperature for 8 hours. At the reaction temperature of 660 °C, phosphorus and the tin halides, are predominantly
transferred into the gas phase where the phase formation takes place [75]. The ampule was subsequently
cooled down to room temperature with a slow cooling rate of 1 K h−1
to obtain crystalline SnIP. The weak
van der Waals interactions between the single SnIP double helices lead to a pronounced fiber-like behavior
which results in a strong cleavage tendency. Therefore, bulk SnIP can be nanofabricated into the bundles
or even the single double helix mechanically using tape.
14

Figure 2.1: Optical images showing the synthetic process. a. The schematic diagram for the synthesis
process of SnIP. b. Heat treatment in furnaces.
2.4 Crystal structure characterization of SnIP
2.4.1 Crystal structure and morphology of SnIP
The bulk SnIP is a truly one-dimensional (1D) van der Waals crystal, which is formed by assembling the 1D
SnIP double-helixes together via van der Waals interactions, as shown in Figure 2.2a. In a SnIP double-helix,
there are two substructures (strands) of different elemental composition - an inner helical chain consisting
of phosphorus atoms (inner strand) and an outer helical chain consisting of alternating tin and iodide
atoms (outer strand). The inner and outer strands of an individual double helix have the same chirality,
i.e. either left (M-helix) or right-handed (P-helix). Each two-winding cycle of the intertwined strand pair
contains seven SnIP units leading to a 7/2 helix. In the most thermodynamically stable form, alternating
M and P helices bind together through van der Waals interaction to form the bulk crystal lattice structure,
as shown in Figure 2.2a. This SnIP crystal belongs to the C2h space group, with lattice parameters of a
= 0.79 nm, b = 0.98 nm, c = 1.84 nm, and β = 110.1°, which means the center-to-center distance between
neighboring double helices is 0.98 nm. Thus, there are 42 atoms in each unit cell of the bulk SnIP crystal
lattice.
15

Figure 2.2: Bulk and nanosized crystals of double helical SnIP. a. Atomic structure of the SnIP crystal
with the most thermodynamically stable stacking order viewed along the a-axis (left panel) and b-axis
(right panel), respectively. b. Optical image of the as-grown SnIP in a sealed quartz tube. Typical optical
microscopy image c. and scanning electron microscopy image d. of exfoliated SnIP crystals on the SiO2/Si
substrate.
An optical image of the as-grown SnIP bulk crystal is shown in Figure 2.2b. Due to the van der Waals
interactions between the neighboring double helices, the SnIP material can be exfoliated into nanowireshape crystals on the SiO2/Si substrate[40, 76, 77]. The optical micrograph and the scanning electron
microscopy (SEM) image of typical SnIP crystals exfoliated into the nanowire form are shown in Figures
2.2c and 2.2d, respectively. Energy-dispersive X-ray spectroscopy (EDS) mapping of the nanowire crosssection perpendicular to the a-axis shows the expected elemental composition (Figure 2.3a). As shown in
the SEM and EDS images, the typical cross-sections of the exfoliated nanowires are close to a rectangle with
16

Figure 2.3: a. Energy-dispersive X-ray spectroscopy mapping of the cross-section perpendicular to the aaxis of the SnIP crystal showing the uniform presence of tin (red), iodine (green), and phosphorus (yellow)
elements. The scale bar is 200 nm. b. Atomic resolution bright field HR-STEM image of the (100) plane in
the SnIP crystal. c. Dark field HR-STEM image along the [100] direction of the SnIP crystal. The insets in
b and c show the corresponding electron diffraction patterns, respectively.
a width in the range of 0.1-1.5 µm. High-resolution scanning transmission electron microscopy (HR-STEM)
images were obtained using the FEI Titan Themis G2 system with four detectors and spherical aberration.
Chromium and carbon layers were pre-coated on the sample and a thin film was obtained using a focusedion beam (FIB, FEI Helios 450S) with an acceleration voltage of 30 kV. The acceleration voltage was then
increased to 200 kV to obtain the HR-STEM image at higher resolutions. Figure 2.3b shows the bright
field HR-STEM image of the SnIP crystal perpendicular to the [100] direction. As shown in Figure 2.3b,
the bulk SnIP crystal lattice is revealed with both the inner and outer strands of each individual double
helix clearly observed as heptagonal rings in this cross-sectional view. Figure 2.3c presents the dark field
HR-STEM image along the [100] direction of the SnIP crystal, which shows the double helical strands
arranged in parallel along the a-axis. The corresponding selected-area electron diffraction patterns are
also provided in the insets of Figures 2.3b and 2.3c, which confirm the pseudo-hexagonal arrangement
17

Figure 2.4: a. Raman spectra of SnIP measured at room temperature and irradiated with a 532 nm laser.
b. Ambient temperature photoluminescence spectrum of SnIP with a minimum recombination energy of
1.88 eV and a shoulder peak at 1.8 eV.
of the double-helical units perpendicular to the a-axis and their parallel arrangement along the a-axis,
respectively.
2.4.2 Raman and photoluminescence measurements on SnIP
Raman spectroscopy is a robust technique to investigate the crystal structure and bonding characteristics
of solids. We measured the characteristic Raman spectrum of SnIP at room temperature, as shown in
Figure 2.4a. The measured Raman spectrum of SnIP exhibits three characteristic vibration modes (Sn-I,
P-P, and Sn-P bonds), which respectively correspond to the Sn-I breathing mode of the outer SnI-helix
localized at 130 cm-1, P-P stretching mode of the inner P-helix localized at 339 cm-1 and 353 cm-1, and P-P
breathing mode of inner P-helix towards the outer SnI-helix localized at 217 cm-1 and 450 cm-1. As shown
in Figure 2.4a, there were little differences in those three vibration mode positions between bulk SnIP and a
single strand of SnIP, which demonstrates that the properties of bulk SnIP are not significantly affected by
the stacking arrangement of single SnIP strands in the bulk material. Furthermore, those three vibration
modes can be associated with the crystal structure and bonding arrangement in SnIP. The Sn-I breathing
18

vibration is attributed to the strong Sn-I covalent bonding interactions and the weak vdW interactions
between the double helices. The P-P stretching vibration propagating along the P-helix axis stems from
the strong covalent P-P bonding of the inner P-helix. Therefore, the outer sn-I helix does not interfere
with the inner P-helix. On the other hand, the P-P breathing vibration of the inner P-helix perpendicular
to the double helix axis is determined by the dative ionic bonding interactions between the inner P-helix
and outer SnI-helix.
In addition, based on the DFT calculation, SnIP is predicted to be a semiconductor with an indirect
band gap of 1.79 eV and a direct band gap of 1.81 eV. We performed a photoluminescence measurement on
bulk SnIP at Room temperature. As shown in Figure 2.4b, we found that the measured photoluminescence
spectrum of SnIP features a narrow characteristic peak at 1.88 eV with a shoulder peak at 1.8 eV, which is in
great agreement with the calculated band gaps and the previous report by Nilges. In a word, the measured
Raman spectra and photoluminescence spectrum demonstrated that SnIP with a high-quality crystalline
structure was synthesized successfully.
2.5 SnIP Young’s modulus measurements
2.5.1 AFM-based nanomechanical bending test for SnIP Young’s modulus
An atomic force microscope (AFM)-based nanomechanical bending test [78, 79] was employed to measure
Young’s modulus of SnIP crystals along the a-axis. Figure 2.5a shows the schematic diagram of the AFMbased nanomechanical bending test. To perform AFM bending tests, SnIP nanowires were acquired by
mechanical exfoliation and then transferred on a SiO2/Si substrate pre-patterned with well-defined circular
trenches of varying diameters etched to the depth of 300 nm using reactive-ion etching (RIE). The lengths
of the SnIP nanowires are usually around tens micrometers, which is much larger than the diameters of
the circular trenches (4 µm). Therefore, the surface of SiO2/Si substrate that was not etched can provide
19

Figure 2.5: Sample fabrication for the AFM-based nanomechanical bending test. a. Schematic diagram of
the AFM-based nanomechanical bending test. b. AFM height mapping of the suspended SnIP nanowire.
The inset shows the height profile of the nanowire along the blue dashed line. The scale bar is 1 µm.
adequate support to make SnIP nanowires suspend well across the circular trenches after the transfer
process. The nanowire-shape SnIP crystals were then suspended across these openings. In addition, to
improve the measurement accuracy, the metal (Cr/Au) pads with a thickness (400 nm) larger than the
heights of SnIP nanowires (< 300 nm) were deposited at the trench edges to anchor both ends of SnIP
nanowires as the anchoring points, as shown in the AFM height image in Figure 2.5b.
The bending test on the anchored nanowires was performed using AFM probing. The effective suspended length is determined from the distance-deformation curve. During the nanomechanical bending
test, the AFM probe is scanned along the SnIP nanowire. At each measurement position, the force applied
by the probe (F) is kept constant and the resulting deformation (δ) at the point of contact is recorded. Based
on elastic beam-bending theory, the deformation (δ) as a function of the distance (d) for beams under point
load with both ends fixed should follow [80]
δ(d) = 64d
3
(L − d)
3D/L6
20

where L is the effective suspended length of the SnIP nanowire, D is the deformation at the center of
the SnIP nanowire when the load is applied at the same location, and d is the distance from the effective
anchoring point reference position to the position of the probe where the force is applied.
As shown in Figure 2.6b, the experimental results are in excellent agreement with the theoretical model
prediction. Moreover, Young’s modulus E along the a-axis can be extracted based on E = FL3
/(192ID) [80].
The area moment of inertia (I) of the SnIP nanowire under vertical loading is calculated based on its crosssectional profile experimentally obtained using the AFM as shown in the inset of Figure 2.5b. To further
validate the experiment results, finite element analysis is used to simulate the displacement as a function of
the distance δ(d) based on the extracted Young’s modulus E. Simulations for the bending of SnIP nanowire
crystals during the AFM measurements were performed using the commercial software ABAQUS. SnIP
nanowire geometries with measured cross-sectional profiles were imported. Ten-node tetrahedral (C3D10)
elements were adopted with refined mesh (> 370,000 elements) to ensure accuracy. Loading was applied
along the centreline on the top surface of the nanowires. The bending is in the linear elastic regime
modeled using the fitted Young’s modulus and a Poisson’s ratio of 0.3 (values of 0.2-0.4 were simulated
which showed negligible differences). The simulated probe-response scenario bears closer similarity to
the actual experimental configuration and hence is expected to provide better accuracy as compared to
the analytical model. The simulation matches well with the experimental results and also agrees with
the analytical results as shown in Figure 2.6b. A representative nanowire deformation calculated by finite
element analysis with applied load at the center is shown in Figure 2.6a. An indentation force F = 187 nN
was applied at the center of the nanowire calculated by finite element analysis.
To further improve the accuracy of Young’s modulus measurement, a total of nine samples with various
measurement configurations, especially the effective suspended length L (varying from 2 µm to 10 µm),
are measured. The average value of Young’s modulus E obtained across all the measured samples is 13.55
± 0.78 GPa (Figure 2.6c). The small residual error of the obtained value also validates the reliability of the
21

Figure 2.6: SnIP Young’s modulus from nanomechanical bending test. a. Vertical displacement distribution of a representative SnIP nanowire under an indentation force F=187 nN applied at the center of the
nanowire calculated by FEA (deformation scaled by 10x for visualization). b. The deformation-distance
curves obtained from the experimental measurement, analytical theory, and finite element analysis, respectively, for the sample shown in Figure 2.5b. c. Relationship between the FL3
/D and moment of inertia
(I) of nine different samples. d. Theoretical tensile stress-strain relationship of the SnIP crystal along the
a-axis calculated using DFT.
measurement. Moreover, the indentation force F is chosen to be relatively small (about hundreds of nN)
during all the measurements to ensure that the strain in the SnIP nanowire (E) is well below the elastic
limit. Using the sample in Figure 2.6b as an example, when the force is applied at the center, the bending
strain can be estimated using the analytical expression
Emax ≈
M t
2EI
22

where M =
F L
8
is the bending moment at the point of load, and t is the thickness of the nanowire
[80]. The calculated Emax = 0.23 % agrees with the simulation from finite element analysis (0.23 %) and is
also well within the elastic regime (< 9 %) as predicted by DFT calculations [81, 82, 83] (to be discussed
later). Structural relaxation and stress calculations were performed using the Vienna ab initio simulation
package (VASP) with the exchange-correlation potential of generalized gradient approximation (GGA). To
have an accurate stress value, each structure was fully relaxed until the force was below the 0.0001 eV Å−1
with the k-space mesh of 6 × 6 × 3 and a cutoff energy of 500 eV. The dispersion relations of the acoustic
phonon modes were calculated using the DFPT combined with the phonopy code. The theoretical tensile
stress (σ)-strain relationship of SnIP along the a-axis as calculated by DFT is shown in Figure 2.6d. The
DFT calculation predicts Young’s modulus of E ∼ 13.06 GPa for the SnIP crystal as obtained by fitting the
initial linear regime (up to 9 % here) of the calculated σ-ε curve, which is in close agreement with the
nanomechanical bending test results. Beyond the elastic regime, SnIP crystal can sustain a much larger
strain without brittle rupture, unlike most brittle inorganic semiconductors (such as silicon and SiC) [84,
85]. The DFT calculation indicates that the SnIP crystal can be drastically strained before reaching the
fracture point (32 %) and there is a wide region between the yield point (10 %) and the fracture point,
resembling organic ductile materials [86].
2.5.2 Brillion scattering spectroscopy for SnIP Young’s modulus
To further obtain Young’s modulus of SnIP at a more fundamental level, the Brillouin scattering spectroscopy [87] measurement is carried out to probe the photon-phonon interactions in the SnIP crystals.
The Brillouin scattering spectra of SnIP were measured in the backscattering geometry by a confocal microscope system, which is composed of a confocal microscope with a 20× objective lens (numerical apertur
e= 0.42) and the (3+3)-pass tandem Fabry Pérot interferometers with high contrast (1015). A photon counting head (detector, Hamamatsu H10682-110) is also equipped and the wavelength of the incident laser is
23

Figure 2.7: SnIP Young’s modulus from Brillion scattering spectroscopy measurement. a. Brillouin scattering spectra of SnIP along the a-axis. b. Acoustic phonon dispersion relations along the a-axis calculated
using DFT. One longitudinal acoustic (LA) and two transverse acoustic (TA1 and TA2) phonon branches
are shown.
532 nm. The linear polarization direction of the incident and scattered beam was tuned by the half-wave
plate.
As shown in Figure 2.7a, under the VV polarization configuration (the polarizations of both the incident
and scattered light are fixed along the a-axis of the SnIP crystal), a pair of peaks are observed around ±19.6
GHz with 532 nm wavelength incident laser, corresponding to the scattering of light by the longitudinal
acoustic (LA) phonon along the a-axis of the SnIP crystal. Thus, the velocity of sound along the a-axis
([100] direction) can be extracted [88] as:
Vs[100] =
fλ
2nsin(θ/2)
where f is the Brillouin scattering frequency, λ is the wavelength of the incident laser light, n is the refractive index along the a-axis of the SnIP crystal and θ is the angle between the incident and scattered
beams. Here, we adopt the refractive index n = 3.1 based on the DFT calculation, and the angle θ = 180°
24

Figure 2.8: Benchmark of Young’s modulus for SnIP among different material classes.
due to the back-scattering geometry, which results in Vs[100] = 1.68 km/s. The Young’s modulus can hence
be obtained as [89]:
E = ρV 2
s[100] = 13.8 GP a
where ρ = 4.77 g/cm3
is the mass density of SnIP. This result is consistent with Young’s modulus value
obtained using the nanomechanical bending test. In solid-state crystals, the velocity of sound is equivalent
to the slope of the dispersion curve for the corresponding acoustic phonons. Figure 2.7b shows the acoustic
phonon dispersion relations along the Γ to X direction ([100]) as calculated by DFT. For small values of the
wave vector (k), the acoustic phonon dispersions are approximately linear and the experimental results
from the Brillouin scattering match well onto the calculated dispersion curve for the longitudinal acoustic
phonon along the [100] direction as indicated in Figure 2.7b. Here, we would also like to note that previous
25

work reported the experimentally measured Young’s modulus in SnIP to be well over 190 GPa due to
possible inaccuracy in the experiment or sample degradation, even though the same work also mentioned
that the theoretical value of SnIP bulk modulus should be around 14.9 GPa [90].
Comparing Young’s modulus of the SnIP crystal with other air-stable materials, its E along the aaxis, which has been independently confirmed by both the nanomechanical bending test and the Brillouin
scattering measurement, is significantly smaller than any other known inorganic materials, including all
the ceramics, metals, III-V compounds, as well as low-dimensional materials such as carbon nanotube,
nanowire crystals, and all the known 2D materials. The Young’s modulus of SnIP is even comparable to or
lower than many organic materials such as bamboo, wood, and polyoxymethylene (POM) [91], as shown
in Figure 2.8. Here, we attribute the ultra-low Young’s modulus of SnIP to the unique double helix lattice
structure, in which the strong inter-atomic bonding between the P atoms in the inner strands and between
the Sn and I atoms in the outer strands are along the tangential direction of the helices, leading to the
significantly relaxed interactions along the a-axis, analogous to atomic-scale mechanical springs.
2.6 Flexibility of the SnIP nanowire
To further characterize the flexibility of the material, SnIP nanowire crystals were bent into a U-shape
by fixing one end of the nanowires and moving the other end using a micro-manipulated probe tip [92].
In this test, one end of the SnIP nanowire was anchored in place and a micro-manipulated probe tip was
used to bend the nanowire. Figure 2.9a-2.9e shows the sequential stages of a representative bending and
recovery process as recorded using an optical microscope. The bending strain is extracted based on the
observed radius of curvature:
Eb ≈
W
2R
26

Figure 2.9: Bending test for flexibility of the SnIP nanowire. a-e. Optical images recorded at the different
stages of the bending test showing full recovery to the initial configuration after release. The bending was
performed using a micro-manipulated probe tip. One end of the nanowire was anchored with a metal pad.
f. Top panel: zoomed-in optical image of the bent SnIP nanowire marked with the dashed line in (a-e);
Bottom panel: AFM height map of the same SnIP nanowire as in the top panel.
where W is the width of the nanowire and R is the bending radius along the centreline of the nanowire
[93], as shown in the upper panel of Figure 2.9f. The width of the nanowire w is confirmed to be 600 nm
using AFM measurement as shown in the lower panel of Figure 2.9f, which leads to Eb ≈ 27.3 % with R ≈
1.1 µm. The nanowire fully recovers to its initial configuration very quickly (well within one second) after
being released from this giant strain (> 27 %).
Raman spectra of the SnIP nanowire were measured both before and after this bending strain test at the
same bent point. No significant change in the Raman spectra before and after the bending test is observed
27

Figure 2.10: The Raman spectra of SnIP before bending, after 1st bending, and after 10th bending test. Both
spectra were obtained at the same bent position. It indicates that the materials property remains the same
(corresponding to the good repeatability) after 10 times bending
as shown in Figure 2.10, indicating that there is no noticeable change in the local strain after the recovery.
This bending strain is clearly higher than that observed in other semiconducting nanowires [78, 94]. The
flexibility figure of merit of SnIP is calculated according to
fF OM =
λy
E
where λy is the yield stress [95], and compared with other materials as shown in Figure 2.11. Due to
the low Young’s modulus E and relatively large λy, the fF OM of SnIP is higher than most of not only
low-dimensional materials but also polymers. The flexibility (f ) of the materials is further related to its
thickness (h) [95], f = (2/h)fF OM . Thus, once the thickness is narrowed down to the single double helices
(∼1 nm), the flexibility of SnIP would increase drastically, compared with the current width w (∼600 nm).
28

Figure 2.11: Figure of merit for flexibility of materials
2.7 Deformability of the SnIP nanowire
Moreover, more than 20 samples were elastically tested showing a large elastic bending range (7.8 % -
27.3 %), and plastic bending was observed when the strain exceeded 29.2 %, as shown in Figure 2.12a.
Furthermore, the plastic bending strain can be up to 60 % without fracture. We attribute this observation
in SnIP to the relatively weak van der Waals interaction between neighboring double helices in the SnIP
crystal, leading to relative slipping between the helices during the bending, hence resulting in the large
elastic and plastic bending strain.
To confirm the van der Waals interactions between the neighboring double helices in the SnIP crystal,
we further calculated the mapping of the charge density and electron localization function (ELF) using DFT
as shown in Figure 2.12b, which indicates that the charge density is localized within the helical strands
and is much weaker between neighboring double helices units. The slippage energy (Es) along the [100]
direction is calculated to be 0.041 eV/atom.
29

Figure 2.12: Deformability of the SnIP nanowire. a. Bending strains measured in 35 SnIP nanowire samples.b. Calculated charge density and electron localization function (ELF) of the SnIP crystal. c. Free
energy during the slipping process. d. Energy variation as a function of the interdouble-helix distance
during the slipping along the a-axis.
The mirror-symmetrical slipping period is divided into 14 steps to calculate the energy differences
(Figure 2.12c) and 8 of them are shown in Figure 2.12d. The Es of SnIP is lower than that of some 2D
materials, such as MoS2 and InSe [96], indicating a slippery interface along the a-axis. Meanwhile, the
cleavage energy (Ec) is calculated to be 0.064 eV/atom, which is higher than that of MoS2 and graphene
[96] and higher than the Es of SnIP. It shows that the SnIP double helices units in the crystal favor slipping
(smaller Es) to release the local strain caused by the bending. Moreover, during the slipping, the double
helices are expected to bind tightly together (larger Ec), leading to a larger observed elastic strain than
30

Figure 2.13: Benchmark of the deformability in materials of different electronic bandgaps.
the predicted elastic limit without considering the slipping effect. The slipping between neighboring basal
planes has also been experimentally observed during the bending of 2D van der Waals materials such as
hexagonal boron nitride [97], in which the interlayer coupling is also based on van der Waals interaction
similar to the inter-helix coupling in SnIP. Furthermore, we adopt the deformability factor proposed in Ref.
[96]
Ξ = (Ec/Es)(1/E)
(in units of GPa−1
) to evaluate the bendability of the SnIP nanowire. In Figure 2.13, the deformability of
SnIP crystal is benchmarked with respect to other materials of different bandgaps. Due to the much lower
Young’s modulus, the deformability of SnIP is the highest among all the known inorganic semiconductors,
higher than both InSe [96] and Ag2S [98], and is comparable to ductile metals such as titanium.
31

2.8 Summary and outlook
This study revealed the ultra-low Young’s modulus of inorganic double helical crystal SnIP. Such extraordinary mechanical property originates from the unique crystal lattice of this material in which the strong
inter-atomic bonding is tangential to the a-axis of the crystal, resulting in an atomic-scale spring-like double helical structure. The large elastic and plastic bending strain is also observed due to the pliability of
the double helices and the presence of easy slippage pathways along the a-axis, giving rise to a record
high flexibility and deformability factor among inorganic semiconductor materials. The ultra-low Young’s
modulus and high elastic bending strain in this unique inorganic semiconductor crystal, combined with
its high predicted electron mobility [18], indicates promising potential for applications in a broad range of
nanomechanical and nanoelectronics disciplines, especially where giant stretchability and bendability are
desired.
32

Chapter 3
Infrared Polarization Photodetection and Imaging of Quasi
One-Dimensional Perovskite Chalcogenide BaTiS3
3.1 Absract
High optical anisotropy usually gives rise to broadband birefringence and large dichroism, which play
an important role in various potential applications including polarizing optics, optical communication,
and sensing technology. As a quasi-one-dimensional perovskite chalcogenide, Barium titanium sulfide
(BaTiS3), was demonstrated to possess giant in-plane optical anisotropy and narrow direct bandgap, which
enables the high-performance infrared polarization sensing technology. Therefore, it is of scientific significance to study the optoelectronic properties of BaTiS3 and explore its applications in infrared polarization
photodetection and imaging. Here we systematically investigated the performance of BaTiS3 in anisotropic
photoresponse at different infrared wavelengths using an anisotropic BaTiS3 photodetector. Polarized infrared imaging was successfully achieved at room temperature with good stability. Overall, this work paves
the way for the advancement of next-generation infrared polarization sensing and imaging technology.
33

3.2 Introduction
Optical anisotropy lies in its ability to tremendously influence the behavior of light in different directions
within a material. This property has far-reaching implications in various scientific and technological applications, including polarization control [99, 100], optical modulation [101, 102], and enhanced sensing
and imaging [103, 104, 44]. Optical anisotropy in materials can be achieved through various methods,
often involving the manipulation of the material’s structure or properties. For instance, applying mechanical strain to a material can dynamically control the degree of optical anisotropy, [32, 33] but the
effectiveness of strain-induced optical anisotropy is limited by the compatibility and ultimate strain of materials. Anisotropic metamaterial or metasurface architectures were tailored to the needs of large optical
anisotropy, but fabrication challenges and optical losses block their practical use. [34, 35] Recently, some
2D materials with reduced symmetry crystal structure, such as black phosphorus (BP), tellurene (Te), and
germanium diselenide (GeSe2), attracted considerable attention owing to their intrinsic in-plane optical
anisotropy. [45, 104, 103, 105] However, their optical anisotropy remains limited, and material degradation
and fabrication block their use in optical and optoelectronic systems. Perovskite chalcogenides represent
a burgeoning class of inorganic semiconductors distinguished by their compelling electronic and optical
characteristics. Notably, these materials exhibit ultrahigh light absorption coefficients across the visible to
infrared spectrum, coupled with excellent carrier mobility and small effective masses of charge carriers.
These remarkable properties position perovskite chalcogenides as highly promising candidates for a range
of technological applications, including photovoltaics, infrared optoelectronics, and thermoelectrics. [36]
Quasi-one-dimensional crystal structures are characterized by the presence of one-dimensional substructures held together by strong covalent bonding. These substructures interact with each other through
weak electrostatic forces [37]. The distinctive atomic arrangement gives rise to asymmetry in their crystal
structure, consequently influencing their optical and electronic properties. Quasi-one-dimensional perovskite chalcogenides integrate the distinctive features of perovskite structures and chalcogen elements
34

with an asymmetric one-dimensional structural configuration. As a result, these materials are anticipated
to exhibit substantial and accessible in-plane anisotropy.
Barium titanium sulfide (BaTiS3), a hexagonal perovskite chalcogenide with a quasi-one-dimensional
crystal structure, recently become a rising star in the family of quasi-one-dimensional materials [39]. The
quasi-one-dimensional crystal structure contributes to its exceptionally large optical anisotropy. BaTiS3
crystals exhibit an unprecedentedly high, broadband infrared birefringence (∼0.76) covering from midinfrared to long-infrared, as well as a large window of linear dichroism in the infrared region (1.6 µm to
4.5 µm). Moreover, linear dichroism conversion was discovered in BaTiS3 ultrathin flakes.[106] More interestingly, BaTiS3 was proved to be a semiconductor with a narrow direct bandgap. Those unique and
interesting properties promise next-generation photonic detection devices for modern infrared polarization sensing and imaging applications. Here, in this work, we exfoliated the bulk BaTiS3 into thin flakes and
fabricated the thin flake BaTiS3 photodetector for infrared polarization photodetection and imaging. First,
we evaluated the performance of infrared photodetection for BaTiS3 devices at different wavelengths. They
show a usually large anisotropy in photoresponse. In addition, we measured the frequency dependence of
BaTiS3 photodetector, which indicates that the dominant mechanism at the illumination of infrared light
is the trap-induced photogating effect. Finally, based on its anisotropic photoresponse, we successfully
demonstrated the polarized infrared imaging capability of the fabricated BaTiS3 photodetector. In a work,
this work reveals that BaTiS3 holds promise for high-performance and polarization-sensitive broadband
photodetectors and imaging facilities.
35

3.3 Anisotropic characterization of BaTiS3
3.3.1 Structural anisotropy in BaTiS3
BaTiS3 crystallizes in the polar non-centrosymmetric P63mc space group, where TiS6 octahedra sharing
common faces form the parallel quasi-1D lattice chains propagating infinitely along the c axis and these
chains are separated from each other in the ab plane by Ba2+ cations [39], as illustrated in Figure 3.1a.
The quasi-one-dimensional chains of BaTiS3 and their hexagonal configuration are revealed by the highresolution transmission electron microscopy (HR-TEM) measurements showing the c-axis and a-axis of
the crystal (Figure 3.1b). Corresponding electron diffraction images are also presented in the insets of
Figure 3.1b, which show the hexagonal and square reciprocal space along the c-axis and a-axis, respectively [40]. Therefore, this quasi-one-dimensional structure features easily large and accessible in-plane
anisotropy. In addition, the intra-interaction within the TiS6 cell is the strong covalent bonding while
the inter-chain direction is weak ionic bonding, which makes bulk BaTiS3 exfoliated into thin flakes with
below two hundred nanometers, as shown in Figure 3.1c and 3.1d. Cleaving quasi-one-dimensional bulk
crystals reduces the dimensionality of material effectively, which leads to robust surfaces that retain covalent bonding in atomic columns and thus make the material’s potential for device miniaturization and
application in integrated circuit [37].
3.3.2 Optical anisotropy in BaTiS3
This unique quasi-one-dimensional crystal structure BaTiS3 results in a giant optical anisotropy. Figures
3.2a and 3.2b show the transmission and reflection spectra with incident light polarized parallel and perpendicular to the c axis, which was measured by the polarization-resolved infrared spectroscopy [39]. The
absorption edge was observed at 4.5 µm when the polarized light was parallel to the c-axis. Whereas, the
absorption edge was blueshifted to 1.6 µm (0.76 eV) when the polarization was perpendicular to the c axis.
36

Figure 3.1: Structural anisotropy in BaTiS3. a. Schematic lattice structure of the BaTiS3 lattice viewed
along the c-axis and the a-axis of the crystal. The c-axis view of the crystal has hexagonal symmetry
and the a-axis view of the crystal illustrates the TiS6 chains along the c-axis. b. Atomic resolution TEM
images showing c axis (left) and an axis (right) projection of the BaTiS3 crystal. The corresponding electron
diffraction patterns are shown as the insets. c. SEM image of a BaTiS3 crystal needle. d. Optical micrograph
of BaTiS3 sample mechanically cleaved along a–c plane. Figures adopted from reports [40].
More interestingly, the giant optical anisotropy triggers strong linear dichroism and tremendous birefringence. Figure 3.2d indicates the extracted wavelength-dependent birefringence, linear dichroism, and
normalized dichroism. BaTiS3 displays a broad window of dichroism in the infrared with absorption edges
at 1.5 µm and 4.5 µm for light polarized parallel and perpendicular to the crystal c axis, respectively. It also
displays an unprecedentedly giant broadband birefringence of up to 0.76 in the mid-infrared between 8
µm and in the long-infrared 16.7 µm. This performance exceeds that of other reported highly anisotropic
materials, including the best liquid crystals. For example, the birefringence of BaTiS3 is more than twice
37

Figure 3.2: Optical anisotropy in BaTiS3. a. Infrared transmission spectra for incident light polarized
perpendicular (orange) and parallel (dark green) to the c-axis. b. Infrared reflection spectra for incident
light polarized perpendicular (orange) and parallel (dark green) to the c axis. c. Real (ϵ1) and imaginary
(ϵ2) parts of the dielectric function for polarization perpendicular and parallel to the c axis, extracted
from a combination of ellipsometry and polarization-resolved transmission/reflectance measurements. d.
Birefringence, linear dichroism and normalized dichroism for wavelengths from 210 nm to 16 µm. Figures
adopted from reports [39].
that of rutile (0.29) and an order of magnitude larger than commercially used long-wave infrared birefringent materials [39, 40]. The remarkable optical characteristics, including the substantial birefringence
and a broadband linear dichroism window observed in the infrared wavelengths, can be attributed to the
pronounced anisotropy inherent in the crystal structure of BaTiS3. This anisotropy is further influenced
by distinct chemical composition variances along both inter-chain and intra-chain directions within the
crystal lattice. The synergistic effect of these factors imparts unique and desirable optical properties of
BaTiS3, making it an intriguing material for applications involving tailored manipulation of light in specific wavelength ranges.
38

Figure 3.3: Polarized optoelectronic responses for BaTiS3 photodetectors. a. The schematic structure of
the BaTiS3 photodetector. b. Optical micrograph of the BaTiS3 photodetector. Inset: Polar plots of the
Raman spectra intensity of the A2
1g mode versus the polarization angle at the 1.71 eV excitations. c. Polar
diagram of the polarized photocurrent under 2.0 V drain bias and 0 V gate bias for the incident wavelengths
of 3.39 µm. d. Polar diagram of the polarized photocurrent under 2.0 V drain bias and 0 V gate bias for the
incident wavelengths of 1.55 µm.
3.4 Polarized optoelectronic responses for BaTiS3 photodetectors
To explore the potential of BaTiS3 in infrared photoelectronics, based on exfoliated BaTiS3 flakes, we fabricated the BaTiS3 photodetector with the contact direction parallel to the c-direction, as shown in Figure
3.3a. Based on anisotropic Raman spectra in the region of A2
1g mode, we can identify the c-direction of
BaTiS3 (The inset of Figure 3.3b) [40]. The channel length of this device is 4 µm and the channel width of 3
µm, crossing the BaTiS3 flake (Figure 3.3b). The gate bias is fixed at 0 V, the drain bias is varied from -2.0 to
2.0 V, and the incident lights are nonpolarized. The polarized photoresponses of the device were also investigated under linear polarized illuminations. Figures 3.3c and 3.3d display the photocurrent of BaTiS3-based
39

Figure 3.4: Polarized optoelectronic responses for BaTiS3 photodetectors with contact direction perpendicular to the c-direction. a. Optical micrograph of the BaTiS3 photodetector with a contact direction
perpendicular to the c-direction. b. Polar diagram of the polarized photocurrent under 2.0 V drain bias
and 0 V gate bias for the incident wavelengths of 1.55 µm. c. Transistor transfer characteristic for Vg = 0
V. d. Dependence of the photocurrent with polarization angles (Vg = 0 V and Vds = 2 V).
photodetector illuminated under 3.39 and 1.55 µm at 300 K temperature, respectively, showing its feasibility
for photodetecting at near-infrared and mid-infrared wavelength range. More importantly, the photoresponse performances demonstrate that the BaTiS3 photodetector shows a giant anisotropy in broadband
optoelectronic responses from near to mid-infrared spectrum. Under mid-infrared incident light (3.39 µm),
the BaTiS3 photodetector shows a large photoresponse anisotropic ratio (11). The photoresponsivity is 58
mA/W. Under the near-infrared incident light (1.55 µm), the photoresponse anisotropic ratio of the BaTiS3
photodetector is slightly smaller than that under the mid-infrared light. The photoresponsivity is 80 mA/W.
However, this photoresponse anisotropic ratio is still larger than the widely researched 2D materials such
40

Figure 3.5: Trap-induced photogating effect in BaTiS3. Photocurrent versus modulation frequency under
incident light of 1.55 µm (Vg = 0 V and Vds = 2 V).
as black phosphorous [45], and tellurene [104, 44]. More importantly, this observation of the slight difference in the photodetection anisotropy under different incident light demonstrates clearly that the BaTiS3
photodetector can identify the slight differences between different polarized illuminations. In addition,
we also fabricated the BaTiS3 photodetector with a contact direction perpendicular to the c-direction of
BaTiS3 (Figure 3.4a). This photodetector shows a smaller photoresponse anisotropic ratio than the BaTiS3
photodetector with contact direction parallel to the c direction, as shown in Figures 3.4b and 3.4d. That is
probably attributed to the higher carrier mobility in the c direction, which leads to shorter carrier transit
time along the c-direction, contributing to large polarization photosensitivity.
3.5 Trap-induced photogating effect in BaTiS3
In addition, we measured the frequency dependence of BaTiS3 photodetector. It shows the dynamic photocurrent as a function of frequency in the range from 200 Hz to 10 kHz (Figure 3.5). As shown in Figure
3.5, at the low-frequency range, the photocurrent drops quickly with increasing frequency, due to the
photogating effect. When the modulation frequency of the incident light was larger than 1000 Hz, the
41

photogating effect was ruled out, and the pure photoconductive effect dominated. This behavior suggests
that the dominant mechanism at the illumination of infrared light is the trap-induced photogating effect at
low-frequency [104, 45]. In the case of the photogating effect, the potential inhomogeneity can result from
a random distribution of trapped charges at the interfaces and vacancies, dislocations, or grain boundaries
in materials. Trap states with energies above the Fermi level are empty and are able to capture electrons.
Trap states with energies below the Fermi level are negatively charged because they are filled with electrons. As a result, they are able to capture free holes. The localized states trap one type of photocarriers
and therefore prolong the recombination lifetime of the other type. If the lifetime of a carrier is longer
than its transit time, it will make several transits through the material between the contacts, provided that
the contacts are able to replenish carriers drawn off at the opposite side by the injection of an equivalent
carrier, which is required by the charge neutrality condition. The free carrier continues to circulate until
it is annihilated by recombination, leading to a large photoconductive gain. This gain will contribute to
the generation of photocurrent. In addition, we can induce that the photoresponse anisotropy at low frequency is due to not only the optical anisotropy but also the transport anisotropy and the anisotropy of
the defect distribution.
3.6 Linearly infrared polarization imaging based on BaTiS3
High anisotropy ratio, high photoresponsivity, and excellent stability are basic requirements for realistic
imaging applications. A high photoresponsivity anisotropy of the stable quasi-one-dimensional BaTiS3
with in-plane anisotropy crystal structure opens up possibilities for developing polarized infrared imaging
technology. Therefore, we investigated infrared linearly-polarization imaging for a designed target using
fabricated BaTiS3 photodetectors. [103]. Firstly, we designed and built up the imaging system based on the
schematics of the imaging system shown in Figure 3.6. In this imaging system, the polarization direction
of the laser can be changed by rotating the half-wave plate. A broadband spectrum range from visible
42

Figure 3.6: Schematic diagram of the imaging system. The polarization direction of the laser can be changed
by rotating a half-wave plate.
to infrared can be covered. The key point of this system is using a programmable motor to control the
position of the laser spot on the target, then the BaTiS3 photodetector will get the optical signal of the
reflected light from each pixel of the target. Finally, we can get the photocurrents on all pixels of the target
and thus achieve infrared polarization imaging. The resolution of the imaging is determined by the size
of the laser spot and the pixel size we set. After building up and validating the imaging system, we used
this setup to systematically explore the linear-polarized infrared imaging ability of an anisotropic BaTiS3
photodetector. With the movement of the target, the position-dependent current of the BaTiS3 device
could be recorded in real-time by a lock-in amplifier, which was further transformed into a photocurrent
mapping image. Figure 3.7 show the polarization-dependent infrared images measured under 1.55 µm
illumination with four different polarization angles of 0 ◦
, 45 ◦
, 90 ◦
and 135 ◦
, respectively. A highresolution image of a “USC” pattern with a high photocurrent contrast ratio is captured at a polarization
angle of 0 ◦
, while an image with a smaller photocurrent contrast ratio is obtained at a polarization angle
of 90 ◦
. Therefore, a polarization contrast ratio over 5 is achieved, revealing the superior polarized light
imaging capability of the BaTiS3 device. These results unambiguously demonstrate the great potential of
BaTiS3 for polarization-sensitive photodetection and imaging.
43

Figure 3.7: Linearly infrared polarization imaging. a. Schematic diagrams of the BaTiS3 photodetector. b.
Perspective view of a BaTiS3 crystal plate with blue and orange spheres representing barium and sulfur
atoms, respectively. TiS6 octahedra are highlighted in green. c. Polar diagram of the polarized photocurrent under 2.0 V drain bias and 0 V gate bias for the incident wavelengths of 1.55 µm. d. Schematic of the
imaging target and the expected image.
3.7 Summary and outlook
In summary, we systematically investigated the photodetection performance of BaTiS3 photodetectors.
Using the fabricated BaTiS3 photodetectors, we also successfully achieved linear-polarized infrared imaging. BaTiS3 photodetectors show high photoresponsivity and large photoresponse anisotropy under the
illumination of near-infrared and mid-infrared. Based on the anisotropic photoresponses, we further
demonstrated the linear-polarized imaging capability of BaTiS3 photodetectors. As an emerging quasione-dimensional material, BaTiS3 photodetector will undoubtedly promote the development of modern
infrared polarization photodetection and imaging technology.
44

Chapter 4
Black Phosphorus Molybdenum Disulfide Midwave Infrared
Photodiodes with Broadband Absorption-Increasing Metasurfaces
4.1 Abstract
Black phosphorus (BP) has been established as a promising material for room-temperature mid-wave infrared (MWIR) photodetectors. However, many of its attractive optoelectronic properties are often observable only at smaller film thicknesses, which inhibits photodetector absorption and performance. In this
work, we show that metasurface gratings increase the absorption of BP-MoS2 heterojunction photodiodes
over a broad range of wavelengths in the MWIR. We designed, fabricated, and characterized metasurface gratings that increase absorption at selected wavelengths or broad spectral ranges. We evaluated the
broadband metasurfaces by measuring the room temperature responsivity and specific detectivity of BPMoS2 photodiodes at multiple MWIR wavelengths. Our results show that broadband metasurface gratings
are a scalable approach for boosting the performance of BP photodiodes over large spectral ranges.
4.2 Introduction
Recently, black phosphorus (BP) has been identified as a promising material for mid-wave infrared (MWIR,
3-5 µm) photodetection.[44] Due to their remarkably high performance at room temperature (D* values
45

exceeding 1010 cm Hz1/2 W−1
) and ease of integration with many substrates, [107, 108, 109, 110, 45,
111, 112] MWIR BP photodetectors have significant potential to benefit civilian and defense applications.
BP is an anisotropic 2D layered material and consists of sheets of phosphorus atoms held together by
van der Waals forces.[113, 46] It exhibits a direct bandgap of 4.1 µm or 300 eV in its bulk form and a
high carrier mobility of 1000 cm−1 V
−1
s
−1
. [46, 50] Additionally, its bandgap is statically tunable by
arsenic alloying and absorber thickness and dynamically tunable with electric displacement fields. [51, 52,
53, 54, 55, 56] However, these attractive material properties are often observable only with smaller film
thicknesses, which reduces photodetector absorption and performance. [49, 57]
To address this challenge, a variety of methods have been successfully tested with BP photodetectors
to enable high-performance devices. Device architectures that enable photogating and gate voltage tuning
effects can dramatically increase the photoresponse of BP. [45, 53] Additionally, integrated nanostructures such as waveguides, plasmonic apertures and gratings, and Fabry-Pérot cavities have experimentally
shown to increase performance in BP devices. [58, 59, 60, 61, 62, 63, 64, 65, 66] While the nanostructures
do not increase the intrinsic absorption of BP (i.e., the imaginary part of its refractive index), they increase
the amount of incident radiation that is collected, which in turn boosts the device’s signal-to-noise ratio. These nanostructures are often designed to boost photodetection around specific wavelengths such as
telecom bands, but many applications require a high photoresponse across large ranges of wavelengths.
In a previous study, metal-insulator-metal (MIM) structures were numerically predicted to increase the
amount of light absorbed in thin-film BP. MIM structures consist of a nonmetallic material that is placed
between a metallic back reflector and a periodic metallic nanostructure. [68] An MIM grating creates a lateral resonant mode within the nonmetallic material, and the spectral location of the resonance depends on
the grating’s structural parameters (e.g., dielectric thickness, and periodicity). [69, 70] Incident light then
couples to the mode, which recirculates within the structure and increases the amount of light absorbed
by the nonmetallic material. [69]
46

We integrated BP-MoS2 photodiodes with MIM metasurface gratings that have absorption-increasing
capabilities that are expandable to large wavelength ranges. Previously, we showed that an MIM grating
enhanced the performance of a BP photoconductor. [63] But the MIM increased responsivity only at a
select wavelength, and the BP photoconductor’s architecture resulted in moderate performance. Incorporating an MoS2 layer with a BP photodetector results in a more robust heterostructure photodetection
mechanism. A BP-MoS2 interface, in which the BP is a p-type MWIR absorber and the MoS2 is an n-type
hole barrier, creates a pn heterojunction that is capable of room temperature photodetection in the MWIR.
[111, 114, 115, 116, 117, 118, 119, 120] The MoS2 also serves as a transparent optical window and a passivation layer that protects the BP from oxidation. [121, 122] In this work, we first show that integrating a MIM
with a BP-MoS2 photodiode increases absorption in the MWIR around a selected wavelength. Then, we
adapt the MIM’s structure to support multiple resonances that expand the range of wavelengths in which
absorption is increased. Finally, we evaluate broadband MIMs by measuring BP-MoS2 photodiode responsivity at multiple testing wavelengths. To our knowledge, we present the first demonstration of a BP-MoS2
photodiode with an integrated nanostructure that boosts absorption across a large spectral range.
4.3 Fabrication of BP-MoS2 Photodiodes
Our study begins with fabricating BP-MoS2 photodiodes. A schematic of a BP-MoS2 photodiode is shown
in Figure 4.1a. The photodiode consists of a 50 nm thick Au back reflector, BP layer with thickness T BP ,
MoS2 layer with thickness TM o, and 50 nm thick Ti/Au contacts. The Au back reflector and Ti/Au contacts
support a vertical bias voltage V DS. Figure 4.1b is a scanning electron microscope (SEM) image of a
BP-MoS2 photodiode. The photodiode was fabricated onto a silicon substrate with 300 nm of thermal
oxide, which was cleaned with piranha solution and oxygen ashing. The Au back reflector was deposited
with e-beam evaporation (CHA Industries) before transferring the BP and MoS2 onto the substrate. The
heterostructure layers were created by mechanically exfoliating BP and MoS2 flakes from bulk crystals
47

Figure 4.1: Metasurface-integrated black phosphorus photodiode. a. Device schematic of a black phosphorus (BP) molybdenum disulfide photodiode structure, in which T BP and TM o are the thicknesses of the
BP and MoS2 layers respectively. The thickness of the Au back reflector and Ti/Au contacts is 50 nm. b.
Scanning electron microscope (SEM) image of a BP-MoS2 photodiode. The angle between the armchair
(AC) and the zigzag (ZZ) is denoted with θ. c. Absorption spectra (A = 1 - R - T, T = 0) of the device in a
measured with a Fourier-transform infrared spectrometer (FTIR) and linearly polarized light along the AC
(θ = 0°) and ZZ (θ = 90°) directions. The inset of c is the corresponding Raman spectra with linearly polarized 532 nm light along the AC and ZZ directions. The Ag2 peak is maximized when light is polarized in
the AC direction (θ = 0°), which also maximizes absorption. d. Device schematic of a BP-MoS2 photodiode
after fabricating a periodic metasurface grating with period A and width W on top of the MoS2 layer. e.
SEM image of the same device in b after integration of the metasurface. The inset is a zoomed-in view of
the periodic metasurface grating. f. Simulated and measured absorption spectra of the device in b at linear
polarization angles θ = 0° and θ = 90° after integrating the metasurface. Absorption is maximized with
light that is polarized along the AC axis (θ = 0°), which is perpendicular to the grating lines. The inset is
the simulated intensity profile of |H|2
at 3 µm. The simulated absorption spectra and field intensity profile
were calculated with finite-difference time-domain (FDTD) simulations.
(procured from HQ graphene) and transferring them onto the substrate with polydimethylsiloxane (PDMS)
stamps. We performed the exfoliation and transfer processes in an argon-purged glovebox to minimize
oxidation and exposure to water vapor. [123] For each device in this study, we selected BP samples with
thicknesses around 80 nm and MoS2 samples with thicknesses around 11 nm. The thicknesses of the
flakes were estimated with optical images due to their thickness-dependent color and later confirmed with
atomic force microscopy (Bruker). [122] Once the BP-MoS2 interface was transferred onto the substrate,
48

we identified the orthogonal axes of BP and applied the Ti/Au contacts with electron beam lithography
(Raith) and e-beam evaporation (CHA Industries).
We determined the direction of BP’s orthogonal axes with a Fourier-transform infrared spectrometer
(FTIR, Bruker) and Raman spectrometer (Renishaw). BP is an anisotropic material with high and low
absorption axes that are respectively labeled armchair (AC) and zigzag (ZZ) (the labels AC and ZZ are
based on its orthorhombic crystal structure). [46] Figure 4.1c shows the absorption spectra of the device
in Figure 4.1b along its AC and ZZ axes. The device has layer thicknesses T BP = 83.1 and TM o = 11.7 nm,
and the two absorption spectra in Figure 4.1c were measured with an FTIR and linearly polarized light
along the AC or ZZ axes. The angle of linear polarization θ is shown in Figure 4.1b, and absorption is
estimated with A = 1 - R - T (T = 0 because no light transmits through the Au back reflector). From these
spectra, it is apparent that absorption is higher along the AC axis than the ZZ axis. In addition to the FTIR
measurements, we include the Raman spectra of the device as the inset of Figure 4.1c. The Raman spectra
were measured with 532 nm excitation linearly polarized along the AC or ZZ axis. We found that for our
samples, the Ag2 peak is maximized along the AC axis, as indicated by Figure 4.1c. [122]
4.4 Integrating MIM Metasurface Gratings with BP-MoS2 Photodiodes
We integrated MIM metasurface gratings with BP-MoS2 photodiodes. Figure 4.1d shows the device structure of a BP-MoS2 photodiode with an integrated MIM grating that is perpendicular to the AC axis. The
MIM supports a lateral resonant mode by enclosing the BP and MoS2 layers between a metallic back reflector and periodic grating with period A and width W. [69, 68] When incident light impinges on the
MIM, it excites the mode which thereby increases its round-trip absorption. [67] We fabricated MIMs onto
photodiodes by patterning gratings with e-beam lithography (Raith) and then successively depositing 5
nm of Ti and 50 nm of Au with e-beam evaporation (CHA Industries). Figure 4.1e is an SEM image of the
same device in Figure 4.1b after integrating the MIM. The inset of Figure 4.1e shows clearly distinguishable
49

Figure 4.2: Double- and triple-resonator metasurface gratings. a. Metasurface-integrated BP-MoS2 photodiode with two resonators. A single unit cell with period 2A incorporates two gratings both with period A:
one grating has a width W1 and the other has a width W2. b. BP-MoS2 photodiode with a triple-resonator
grating. Similar to the double resonator system, one unit cell with period 3A includes three gratings with
period A and distinct widths W1, W2, and W3. c. Measured and simulated absorption spectra (A = 1 - R -
T, T = 0) of a double-resonator metasurface grating. Two resonant peaks are present, and T BP = 75.4 and
TM o = 11.7 nm. d. Measured and simulated absorption of a triple-resonator metasurface grating (T BP =
83.2 and TM o = 12.4 nm). Compared to the spectra in c, a third peak is present due to the inclusion of a
third resonator with width W3. Measured and simulated absorption spectra were respectively obtained
with a Fourier-transform infrared spectrometer (FTIR) and finite-difference time-domain (FDTD) simulations.
grating lines. Figure 4.1f shows the measured absorption spectra (FTIR) of the MIM-integrated BP-MoS2
photodiode at linear polarization angles θ = 0° and θ = 90°. After the addition of the MIM (A = 450 and W
= 235 nm), there is a clear resonant peak that appears when θ = 0° (light is perpendicular to the grating).
Absorption is minimized when θ = 90° because the MIM acts as a reflector when linearly polarized light is
parallel to the grating. [63, 124] In addition to the measured spectra, the simulated absorption spectrum of
the MIM-integrated photodiode is included in Figure 4.1f. The simulated spectrum is in good agreement
with the measured spectrum, and it was calculated with finite-difference time-domain (FDTD) simulations.
The optical constants of BP and MoS2 were borrowed from References [50] and [125], respectively. The
50

Figure 4.3: Absorption spectra of metasurface gratings with varying geometry. a. Measured and b. simulated absorption spectra (A = 1 - R - T, T = 0) of three single-resonator devices (MIM 1-3). With a fixed
A and an increasing W, the resonant peak shifts to higher wavelengths. The structural parameters for all
devices in this figure are included in the text as Table 4.1. c. Measured and d. simulated absorption spectra of three double-resonator devices (MIM 4-6). As W1 decreases, the lower-wavelength resonant peak
blueshifts. Conversely, as W2 increases, the higher-wavelength resonant peak redshifts. e. Measured and
f. simulated absorption spectra of two triple-resonator devices (MIM 7-8). As W1, W2, and W3 all increase, the absorption spectrum is shifted to higher wavelengths. Measured spectra were obtained with
a Fourier-transform infrared spectrometer (FTIR), and simulated absorption spectra were calculated with
finite-difference time-domain (FDTD) simulations.
inset of Figure 4.1f is the simulated intensity profile of |H|2
at λ = 3 µm, in which the field is strongly
confined to the BP-MoS2 layers. Next, we adapted the MIM grating’s structure to broaden its absorption
spectrum.
The absorption of an MIM structure can be expanded by adding additional resonators within the grating’s unit cell. Figure 4.2a shows a MIM structure with two distinct grating widths within one unit cell.
The unit cell’s period is then increased 2A while containing two gratings with distinct widths W1 and W2.
Similarly, Figure 4.2b is a MIM with three gratings W1, W2, and W3 inside one unit cell with period 3A.
51

The additional gratings W2 and W3 support additional resonances that can be used to broaden an MIM’s
absorption spectrum. [126, 127, 128, 129, 130] Figure 4.2c shows the measured (FTIR) and simulated (FDTD)
absorption spectrum of the double-resonator system in Figure 4.2a. Two resonant peaks are observable,
which arise from the two distinct grating widths W1 and W2. The device’s structural parameters are T BP
= 75.4 nm, TM o = 11.7 nm, A = 450 nm, W1 = 200 nm, and W2 = 300 nm. The spectra were obtained at θ
= 0° in order to maximize absorption. And to accommodate for the larger unit cell periods of the doubleand triple-resonator MIMs, we obtained each measured absorption spectrum in this study with an FTIR
spot size of 50 × 50 µm2
. The spot size was controlled with a square knife-edge aperture in our instrument,
and the side length of the square spot was much larger than each MIM’s unit cell period. Moreover, Figure
4.2d is the measured and simulated absorption spectra for the triple-resonator system in Figure 4.2b with
structural parameters being T BP = 83.2 nm, TM o = 12.4 nm, A = 600 nm, W1 = 230 nm, W2 = 320 nm,
and W3 = 410 nm. Because of the addition of the third grating W3, a third resonant peak is clearly visible.
Compared to the single-resonator system in Figure 4.1, the multi-resonator MIMs in Figure 4.2 increase
absorption across a greater spectral range.
Table 4.1: Structural Parameters of Each Device in Figure 4.3
MIM T BP (nm) TM o (nm) A (nm) W1 (nm) W2 (nm) W3 (nm)
1 83.1 11.7 450 235
2 78.3 12.6 450 280
3 77.9 11.3 450 310
4 75.4 11.7 450 200 300
5 79.6 12.7 450 230 320
6 76.3 12.2 450 245 330
7 83.2 12.4 600 230 320 410
8 82.0 10.3 600 280 370 460
The spectral position of the single, double, and triple MIM’s resonant peaks can be controlled with
geometry. Figures 4.3a and 4.3b respectively show the measured (FTIR) and simulated (FDTD) absorption
spectra of three single-resonator MIM devices. Each device in Figures 4.3a and 4.3b has a fixed period A =
450 nm and varying grating width W1. As W1 increases, the resonant mode shifts to higher wavelengths.
52

A list of the structural parameters for each device used for Figure 4.3 can be found in Table 4.1. Figures 4.3c
and 4.3d show the measured and simulated absorption spectra of three double-resonator MIM devices. The
spectral location of the distinct resonant peaks can be controlled by independently varying W1 and W2.
The lower-wavelength peak is controlled by varying W1, and the higher-wavelength peak is controlled
by varying W2. Similar to the single-resonator system, an increase in either W1 or W2 for a fixed period
A will result in the corresponding resonance shifting to higher wavelengths. Figures 4.3e and 4.3f are
the measured and simulated absorption spectra of two triple-resonator MIM samples. By simultaneously
increasing W1, W2, and W3, the absorption spectrum, which exhibits three distinct peaks, shifts to higher
wavelengths while retaining the relative spectral distance between each resonant peak.
4.5 Evaluation of MIM Metasurfaces with BP-MoS2 Photodiodes
To evaluate our broadband MIM structures, we compared the room temperature photodetection performance of BP-MoS2 photodiodes before and after integrating single-resonator and triple-resonator gratings.
We denote the single-resonator photodiode as Device 1 and the triple-resonator photodiode as Device 2,
which have the structural parameters of MIM 1 and MIM 7 in Table 4.1 respectively. Figure 4.4 and Figure
4.5 compare the absorption spectra, dark current, and photocurrent at testing laser wavelengths λ1 = 3.39
µm and λ2 = 3.88 µm for the two devices. Each photodiode has an active area of approximately 50 × 50
µm2
and was measured at room temperature. Absorption spectra were obtained with an FTIR spectrometer, and the dark currents were obtained with a source measure unit (Keithley). To measure photocurrent,
the outputs of both lasers (Thorlabs) were modulated with an optical chopper (Thorlabs) at 200 Hz, and
the generated photocurrents were extracted with a lock-in amplifier (Stanford Research Systems). Both
lasers were set to emit at 1.2 W cm−2
, which resulted in approximately 30 µW of power impinging on each
photodiode at both λ1 and λ2. The polarization of the lasers’ output was manipulated with a polarizer and
half-wave plate (Thorlabs).
53

Figure 4.4: Room temperature photocurrent enhancement in Device 1 with a single-resonator metasurface
grating. a. Measured absorption spectra (A = 1 - R - T, T = 0) of Device 1. Device 1 has same the structural
parameters as MIM 1 in Table 4.1. The two testing wavelengths λ1 = 3.39 and λ2 = 3.88 µm are depicted with
vertical lines. b. Dark current in Device 1 as a function of applied voltage bias V DS. c. Measured current
in Device 1 with 30 µW of linearly polarized λ1 = 3.39 µm illumination (I On) before and after integrating
the metasurface, and current with no illumination (I Of f , dark current). The inset is the linearly polarized
responsivity at λ1 = 3.39 µm and bias V DS = -0.5 V after integrating the metasurface. The angle of linear
polarization θ is defined in Figure 4.1. d. Measured current in Device 1 with 30 µW of linearly polarized λ2
= 3.88 µm illumination (I On) before and after integrating the metasurface, and current with no illumination
(I Of f , dark current). The inset is the linearly polarized responsivity at λ2 = 3.88 µm and bias of V DS = -0.5
V after integrating the metasurface.
Device 1 is a single-resonator MIM BP-MoS2 photodiode. Figure 4.4a shows the absorption spectra of
the single resonant peak, which has a center wavelength near λ = 3 µm, and vertical lines that represent λ1
and λ2. The dark current in Device 1 is shown in Figure 4.4b. Figure 4.4c is the measured photoresponse
of Device 1 at λ1 = 3.39 µm before and after integrating the single-resonator MIM grating. I On is the
current when the device is illuminated with approximately 30 µW of linearly polarized light, and I Of f is
the current under no illumination (dark current). The linearly polarized incident radiation is aligned to the
AC axis of the device. After the integration of the MIM structure, the generated photocurrent increased
by 6.7 times at 3.39 µm and a reverse bias of V DS = -0.5 V. Photocurrent generation under reverse bias was
significantly greater than photocurrent generation under forward bias, which is consistent with refs [111],
[65] and [121]. This behavior is likely due to the relative thicknesses of the BP p-type absorber and MoS2
n-type hole barrier, which affect the energy band and carrier dynamics. The inset of Figure 4.4c tracks the
54

Figure 4.5: Room temperature photocurrent enhancement in Device 2 with a triple-resonator metasurface
grating. a. Measured absorption spectra of Device 2. Device 2 has the same structural parameters as MIM
7 in Table 4.1. b. Dark current in Device 2 as a function of applied voltage bias V DS. c. I Of f and I On in
Device 2 with 30 µW of linearly polarized λ1 = 3.39 µm light before and after integrating the metasurface. d.
I Of f and I On in Device 2 with 30 µW of linearly polarized λ2 = 3.88 µm light before and after integrating the
metasurface. The insets of c and d are the linearly polarized responsivity after integrating the metasurface
at λ1 = 3.39 and λ2 = 3.88 µm, respectively, and bias V DS = -0.5 V.
polarization-resolved responsivity R at a bias of V DS = -0.5 V as a function of the linear polarization angle
θ (see Figure 4.1b). R is maximized at 737 mA W−1 when θ = 0° and minimized when θ = 90°. We define
R as extracted photocurrent divided by incident power. Figure 4.4d shows I On and I Of f at λ2 = 3.88 µm.
Similar to Figure 4.4c, I On was measured with linearly polarized incident light at θ = 0° before and after
the inclusion of the MIM grating. After the inclusion of the MIM, the photoresponse increased by 3.9 times
at 3.88 µm and a bias of V DS = -0.5 V. The inset of Figure 4.4d is the polarization-resolved responsivity
at a bias of V DS = -0.5 V as a function of θ, which exhibits a maximum responsivity of 191 mA W−1
and
the same anisotropic behavior as the inset of Figure 4.4c. The increase in photocurrent and responsivity
is greater at λ1 than at λ2, which we attribute to the absorption peak in Figure 4.4a being further from
wavelength than λ1 than λ2.
Device 2 is a triple-resonator MIM BP-MoS2 photodiode. Its absorption spectrum is shown in Figure
4.5a, and its dark current is shown in Figure 4.5b. Figure 4.5c shows I Of f and I On at λ1 = 3.39 µm measured
before and after the integration of the triple-resonator MIM, and the photocurrent increased by 7.5 times
55

Table 4.2: Responsivity R and Specific Detectivity D* at V DS = -0.5 V for Device 1 and 2 with and without
the Metasurface Grating
λ1 λ1
R (mA W−1
) D* (cm Hz1/2 W−1
) R (mA W−1
)
Device 1 no grating 110 6.37 x 108
49
with grating 737 4.26 x 109
191
Device 2 no grating 109 8.83 x 108
49
with grating 818 6.62 x 109
614
at a bias of V DS = -0.5 V. The inset of Figure 4.5c, which is the polarization-resolved responsivity at a
bias of V DS = -0.5 V, has a maximum responsivity of 818 mA W−1
. Likewise, at λ2 = 3.88 µm, Figure 4.5d
shows I Of f and I On measured with and without the integrated MIM, and the inset of Figure 4.5d is the
polarization-resolved responsivity at a bias of V DS = -0.5 V. At λ2, the MIM increased photocurrent by
12.8 times and maximized responsivity at 614 mA W−1
. In contrast to Device 1, the triple-resonator MIM
of Device 2 increased photocurrent by more than 7 times at both λ1 and λ2, which is due to the broad
absorption spectra that arise from the multiresonator MIM system. Table 4.2 summarizes the performance
enhancement from the integrated MIM by listing R and specific detectivity D* for both devices and testing
wavelengths. D* is calculated with D* = RAdect
1/2
(2eI Of f )
−1/2
, in which R is the responsivity, Adect is the
active area, e is the elementary charge constant, and I Of f is the dark current. At V DS = -0.5 V, I Of f = 2.33
µA for Device 1 and I Of f = 1.19 µA for Device 2, which corresponds to room temperature dark current
densities of 9.3 × 10−2
and 4.8 × 10−2 A cm−2
for Device 1 and 2, respectively. In future applications,
dark current could be suppressed with gate-tuning mechanisms that decrease the conductance of the BP
channel. [51]
Finally, we compared Device 2 to the room temperature performance of other nanostructure-integrated
thin-film BP photodetectors. Table 4.3 lists the peak wavelength, nanostructure type, room temperature
D* before integrating the nanostructure, D* after integrating the nanostructure (D*’), and the ratio D*’/D*.
Additionally, we included one BP-MoS2 photodiode with a thick absorber layer (T BP > 100 nm) and two
56

Table 4.3: Comparison of Nanostructure-Integrated Thin-Film BP Photodetectors and Bulk Material Photodiodes at Room Temperature
ref material λ (µm) D* (cm Hz1/2 W−1
) nanostructure enhanced D* (D*’, cm Hz1/2 W−1
) D*’/D*
this work BP/MoS2 3.39 8.83 x 108 Device 2 6.62 × 109
7.5
this work BP/MoS2 3.88 3.89 x 108 Device 2 4.98 × 109
12.8
[63] BP 3.37 2.25 x 107 metallic grating 1.44 × 109
6.4
[62] BP 1.55 9.34 x 109 plasmonic antenna 1.43 × 1010 1.43
[65] BP/MoS2 3.0 N/A Fabry–Pérot cavity 1.7 × 109 N/A
[121] BP/MoS2 3.8 1.1 × 1010 N/A N/A N/A
[131] HgCdTe 3.4 5.5 × 1010 N/A N/A N/A
[131] PbSe 3.0 1.0 × 109 N/A N/A N/A
commercially available bulk material photodiodes for room temperature MWIR sensing. [131] Table 4.3
highlights that broadband MIM structures are an effective approach for increasing BP detector performance
at multiple peak wavelengths.
4.6 Summary and outlook
In conclusion, we demonstrated a broadband and scalable approach to increasing the performance of BPMoS2 photodiodes. We fabricated BP-MoS2 photodiodes and showed that single-resonator MIM structures
increase the absorption around a selected wavelength. Then, we characterized double- and triple-resonator
MIM structures with broadband absorption spectra that can be controlled by tuning their geometry. The
broadband absorption spectra resulted in enhanced room temperature performance at multiple wavelengths. The process of integrating MIM structures in one lithographic and deposition step is scalable
to large arrays of BP-MoS2 photodiodes that could be fabricated by large-area growth methods. [132, 133,
134, 135, 136] Each pixel in such an array could have different integrated MIMs that are tailored to specific wavelengths or wavelength ranges. Additionally, MIM metasurfaces and gold gratings can be applied
to other 2D and conventional MWIR sensing materials, such as graphene and HgCdTe, to improve overall performance. [137, 138] Overall, our results show that integrated MIM structures enhance BP-MoS2
photodiodes over a broad range of wavelengths.
57

Chapter 5
An All-Silicon Metalens Integrated with a Mid-Wave Infrared Black
Phosphorus Photodiode
5.1 Abstract
Black phosphorus (BP) is a promising material for mid-wave infrared (MWIR) photodetection because of
its direct tunable bandgap and compatibility with a wide range of substrates, particularly silicon. However,
optical signal collection in BP devices can be limited by its absorber thickness or active area. Dielectric
metasurface lenses, which can be directly integrated into a device’s substrate, are a scalable approach to
increase the signal collected by BP photodetectors. This work presents the design, fabrication, and characterization of an all-silicon metasurface lens for focusing MWIR light through its substrate. Then, a BP-MoS2
heterojunction photodetector is integrated with the metalens, and MWIR photodetection increases at room
temperature. These results demonstrate that integrated metasurface lenses are an excellent approach for
boosting the performance of MWIR BP photodetectors.
5.2 Introduction
Mid-wave infrared (3-5 µm, MWIR) photodetectors have been established as a crucial technology in fields
such as astronomy, remote sensing, and defense. [42, 43, 139] Recently, the 2D layered material black
58

phosphorus (BP) has been identified as a promising material for MWIR sensing due to its high performance
at room temperature and ease of integration with silicon. [44, 45, 140, 141, 142, 112] BP consists of sheets
of phosphorus atoms held together by van der Waals forces and exhibits a direct bandgap (300 meV or 4.1
µm in bulk form), high carrier mobility (≈1000 cm2 V
−1
s
−1
), and bandgap tunable by arsenic alloying,
number of layers, and electric displacement fields. [46, 53, 51, 52] However, optical signal collection in BP
photodetectors can often be limited.
Methods for efficiently collecting impinging light are crucial for BP photodetectors. For instance, BP
devices with tunable bandgaps must have small absorber thicknesses, which reduces photodetector absorption. [53, 49] To address this challenge, nanostructures such as silicon waveguides, plasmonic devices,
and Fabry-Pérot cavities have been shown to enhance the performance of BP photodetectors by increasing
the absorption path of incident light. [60, 59, 62, 65, 63, 143, 58] Additionally, devices with small active
areas are desirable for dense arrays because of reduced parasitic effects in the absorber material. However,
the small active area also reduces the optical collection area, thereby diminishing the signal-to-noise ratio
(SNR). We propose to use integrated metasurface lenses to boost the SNR of BP photodetectors.
Metalenses have been established as planar arrays of subwavelength nanostructures that offer vast
degrees of freedom for controlling the phase, transmission, and polarization of light. [144, 145] When
integrated with a photodetector’s substrate, a metalens that is on the opposite side of the substrate to the
device will focus light onto the absorber layer. [71, 72] Figure 5.1 depicts this concept: A metalens with the
area Alens concentrates light, through the substrate, into the photodetector’s absorbing area Adect. The
collection of light from the larger Alens to the smaller Adect increases the incident power density on the
detector and thereby boosts the SNR. Additionally, the approach in Figure 5.1 is scalable to focal plane
arrays, which will be enabled by large-scale BP growth methods. [107, 132, 135] In the context of an array,
individual metalenses could be fabricated for each pixel, and the concentrated incident light also allows
for smaller device pitches. [71, 72, 73] Previous work has demonstrated this concept in a single-pixel III-V
59

Figure 5.1: Conceptual diagram of a metalens-integrated black phosphorus photodiode. Normally incident
mid-wave infrared light is focused by the metalens with area Alens onto a photodetector with a smaller
area Adect, which increases the collected signal.
detector system. [72] In this work, we present the first experimental demonstration of a metalens directly
integrated with the substrate of a 2D layered material MWIR photodetector.
We directly integrated a silicon metalens with a BP-MoS2 heterojunction photodiode’s substrate and
experimentally demonstrated its capability to increase photodetector performance in the MWIR. We designed and measured silicon metalenses that produce focal spots in the MWIR and concentrate light into a
small absorber region. Then, we integrated a BP-MoS2 photodetector onto the same substrate and demonstrated that the metalens boosts MWIR performance at room temperature. A schematic of a BP-MoS2
photodetector is shown in Figure 5.1, in which the metallic contacts are insulated from the substrate with
Al2O3. BP-MoS2 creates a pn heterojunction in which BP is a p-type MWIR absorber and MoS2 is an n-type
hole barrier. [117, 115, 121, 118, 116, 111, 114, 120] In addition to reducing majority carrier diffusion from
60

BP to suppress dark current, the MoS2 functions as a transparent optical MWIR window and a passivation
layer that protects the BP from water vapor and oxygen. [121, 119]
5.3 Design and fabrication of Silicon metalens
We designed a silicon metalens with cylindrical nanorod unit cells. Silicon was selected because of its
transparency in the MWIR and high refractive index, and cylindrical nanorods were selected because of
their polarization insensitivity that arises from rotational symmetry. [146, 147] The metalens is designed
to focus light by reconstructing the target phase equation given by [147]
φ(x, y) = 2π −
2πn
λ
(
p
x
2 + y
2 + z
2 − f ) (5.1)
where x and y denote the position on the metalens area, n is the refractive index of the medium, and f is
the focal length at λ. [147] In this work, we use f = 300 µm, which is the thickness of the substrate, and
n = 3.43, which is the refractive index of the substrate (silicon is relatively dispersion-less in the MWIR).
The required phase at a point x, y on the metalens’ area is calculated with Equation 5.1, and a nanorod
that imparts the required phase is placed at that position x, y. To provide phases ranging from 0 to 2π, a
collection of nanorods was created by varying the unit cell’s geometry.
The geometry of a nanorod unit cell is depicted in Figure 5.2a, and the structural parameters are its
height h, diameter d, and gap spacing g. h and g are respectively fixed at 2.5 and 0.3 µm, and d is varied
to impart relative phase shifts from 0 to 2π. Figure 5.2b shows the simulated relative phase shift and
transmittance of the unit cell at λ = 3 and 4 µm calculated over d = 0.8 to 1.7 µm. Our range of diameters
results in 0 to 2π phase coverage at both λ = 3 and 4 µm, which allows the collection of unit cells to be
used for metalens designs at either wavelength.
61

All simulations presented in this study were carried out with the Ansys Lumerical FDTD simulation
module. Silicon nanorod unit cell data was obtained with periodic boundary conditions in the x- and ydirections and a plane wave source in the z-direction with normal incidence. The resulting transmission
and phase shift from a given unit cell were extracted with frequency-domain monitors placed in front of and
behind the structure. Metalens focal spots were simulated with an FDTD region of 120 × 120 µm2 within
a full metalens design. The FDTD region was defined with perfectly matched layers at the boundaries.
To reduce simulation times, near-field data was recorded at single wavelengths and anti-symmetric and
symmetric conditions were respectively imposed in the x- and y-directions. Focal spots were calculated
with the built-in near-to-far field projection tool.
In contrast to periodic designs in which the centers between each nanostructure are fixed, we fixed the
distance between the edges of each nanostructure. [148] By imposing a constant spacing g between the
edges of the nanorod and unit cell, a greater number of unit cells can occupy the same area than if d + 2g
summed to a fixed period (i.e., g is not fixed). [71, 148] Additionally, a fixed g ensures that two nanorods
are never in direct contact with each other, which aids fabrication processes and avoids unwanted coupling
effects. [71] We designed our metalens with the fixed gap spacing constraint.
We designed a metalens to have a diameter of 275 µm and focus λ = 3 µm light to a focal length of f =
300 µm inside a silicon substrate. The center part of the metalens design is shown in Figure 5.2c, in which
each circular marker represents one nanorod. To begin the design, a nanorod of d = 1.17 µm is placed at
x, y = 0 µm to impart the required phase shift calculated with Equation 5.1 (the phase at x, y = 0 µm is
2π). The second nanorod is placed by calculating the required phase at the next position, which is found
by moving in the +x-direction according to the fixed gap spacing constraint. Nanorods, with the same
diameter as the second nanorod, are then evenly distributed along a ring that is defined by the position
of the second nanorod. [148] This process is repeated until the metalens design reaches the desired final
size. In Figure 5.2c, as the nanorods’ diameters decrease as their positions move further from the origin.
62

Figure 5.2: Mid-wave infrared silicon metalens design. a. Silicon nanorod unit cell with structural parameters height h, diameter d, and gap spacing g. b. Simulated transmittance and relative phase shift at 3
and 4 µm for a nanorod unit cell with variable d, h = 2.5 µm, and g = 0.3 µm. The relative phase shift is
normalized to 2π. c. Constructed metalens design map that focuses mid-wave infrared light through a 300
µm substrate. Only the center part of the full metalens design is shown. Each marker depicts one nanorod,
and the size and color of the markers represent the nanorods’ diameter d.
However, there is an abrupt increase in diameter, which is due to ensuring a continuous phase profile
while selecting nanorods within the range of diameters in Figure 5.2b. We fabricated and characterized
the design in Figure 5.2.
Each device studied in this work was fabricated on a double-side polished, 300 µm thick, and undoped
Si substrate (float-zone). First, the Si substrates were cleaned with oxygen ashing for 2 min and then a
Nano-strip (VWR) bath for 30 min at 60 °C. Alignment markers for the front side of the substrates were
patterned onto a photoresist layer with a direct laser writer at 375 nm (Heidelberg). Then, 5 nm of Ti
and 50 nm of Au were successively deposited with e-beam deposition (CHA Industries) and lifted off in
63

acetone for 1 h. Identical alignment markers (identical in position and size) were then created on the back
side of the substrates using the same patterning and deposition processes. The direct laser writer was
capable of aligning patterns to markers on either side of a substrate, which enables the precise positioning
of alignment markers on the back side of the substrates to identical markers on the front side. After both
sides of the Si substrate had matching sets of alignment markers, metalenses were fabricated onto the front
side and photodetectors onto the back side.
Metalenses were etched into the front side of Si substrates. The metalens design was pattered into
a poly(methyl methacrylate) (PMMA, Kayaku Advanced Materials) resist layer using e-beam lithography
(EBL, Raith). Then, metalenses were etched into the substrate using SF6 and O2 cryoetching at -120 °C
(Oxford). The patterned PMMA acted as an etch mask, which was subsequently dissolved in acetone.
Following the cryogenic etch step, the metalens’ focal spot was measured before fabricating BP-MoS2
photodetectors onto the same substrates.
Figure 5.3a is a scanning electron microscope (SEM) image of a fabricated metalens on a silicon substrate that is undoped, double-side polished, and 300 µm thick. The inset of Figure 5.3a shows clearly
distinguishable nanorods with vertical sidewall profiles (90 ± 1 ° measured from the plane of the substrate,
extracted with SEM), which were achieved with a SF6-O2 cryoetching process. [149] All fabrication details
are included in the Methods section. We measured the focal spot of the metalens with the setup in Figure
5.3b. First, a filtered 1200 K blackbody source with an aperture of 1.5 cm is positioned to illuminate the
metalens sample. We assumed that the distance between the source and sample (≈30 cm) was sufficiently
far enough to approximate normally impinging collimated plane waves. Then, a MWIR camera (640 × 512
pixels, Jet Propulsion Laboratory) with a 36× objective records xy-plane images inside the silicon substrate
at various positions along the z-direction (z = 0 µm is the metalens, and z = 300 µm is the opposite edge of
the substrate). It is important to note that the camera’s translation in z is multiplied by the refractive index
64

Figure 5.3: Metalens characterization. a. Scanning electron microscope image of a silicon metalens with a
diameter of 275 µm. The inset shows clearly distinguishable nanorods. b. Experimental setup for characterizing the focal spots of the metalens in a. To approximate normally impinging plane waves, a filtered
1200 K blackbody source is placed sufficiently far from the metalens sample. Then, a mid-wave infrared
camera creates an xyz image cube by reconstructing xy-plane images taken at discrete positions in z. c.
Measured xz-plane image of the metalens’ focal spot at λ = 3050 nm. The focal spot is concentrated on the
silicon-air interface, which is denoted by the dashed line. d. Measured and simulated normalized intensity
profile of the metalens at z = 300 µm and λ = 3050 nm.
of silicon, and proof of this subtlety can be found in ref. [71]. Finally, an xyz image cube is reconstructed
from the xy-plane images, and the xz-plane images reveal the location of the focal spot.
We placed a λ = 3050 nm narrowband filter in front of the blackbody source to image the focal spot
that is shown in Figure 5.3c. There is a clear focal spot at the silicon-air interface at z = 300 µm, and the
metalens concentrates light in the area of the lens Alens into a nominal photodetector area Adect. Figure
5.3d shows the measured and simulated normalized intensity profile of focal spot along the z = 300 µm
line. The simulated intensity profile was calculated with an FDTD simulation of the metalens structure.
Details regarding metalens simulations can be found in the Methods section.
We measured and simulated the focal spot of our metalens at multiple wavelengths. Figure 5.4 shows
the focal spot of the metalens at λ = 3050, 3390, 3900, 4220, and 4665 nm. The measured and simulated
spatial location of the focal spots are generally in good agreement. Additionally, the location of the focal
spot decreases in z as λ increases, which is consistent in both the measured and simulated xz-plane images.
In this work, we did not attempt to address the chromatic aberrations in our metalens. However, recent
65

Figure 5.4: Measured and simulated focal spots at five testing wavelengths (3050, 3390, 3900, 4220, and 4665
nm). The measured and simulated position and chromatic trend of the focal spots are in good agreement.
The dashed lines indicate the silicon-air interface.
work has shown that optimization routines can be utilized to create MWIR focal spots with a reduced
chromatic response. [145, 71] At each of our testing wavelengths, the metalens was able to concentrate
MWIR light into a clear spot.
By using the simulated and measured images of the focal spot, we predicted how the metalens’ ability
to focus light affects photocurrent generation. We defined Enhancement as the ratio of the intensity, within
a detector area that is in the presence of the metalens, to the transmitted intensity within that same area in
the absence of the metalens. [73] Qualitatively, Enhancement is the increased signal at the detector due to
the metalens concentrating light. We assumed a detector area of 30 × 30 µm2
. Table 5.1 lists the simulated
and measured focal spot locations and Enhancement at our five testing wavelengths. The simulated values
are generally higher than the measured values, which could be from roughness introduced in our etching
process, loss from our silicon substrate, or the fabricated nanorods being less transmissive than their design
in Figure 5.2b.
66

Table 5.1: Simulated and measured focal spot locations in z and Enhancement at z = 300 µm.
λ [nm] Simulated z location[µm] Measured z location[µm] Simulated Enhancement Measured Enhancement
3050 300 305 5.02 4.39
3390 265 280 12.34 6.83
3900 230 235 6.74 6.39
4220 210 205 3.48 1.47
4665 185 190 2.65 1.12
After characterizing the metalens’ chromatic focal spot, we measured its capability to enhance room
temperature MWIR photodetection. To quantify the enhancement in photodetection, we characterized
and compared two photodiodes: one photodiode’s substrate has an integrated metalens, and the other
photodiode’s substrate has no integrated metalens. Figure 5.5a shows a device with an integrated metalens
and BP-MoS2 photodiode fabricated onto the opposite sides of its substrate. Likewise, Figure 5.5b shows
a device with only a BP-MoS2 photodiode fabricated onto one side of the substrate. By comparing the
generated photocurrent in the metalens-enhanced photodiode to that of the device with no metalens, we
were able to quantify the photodetection enhancement due to the metalens.
5.4 Fabrication and characterization of BP-MoS2 photodiodes integrated
with the metalens
We fabricated the two BP-MoS2 photodiodes for comparison with the same processes, layer thicknesses,
and active areas. Each BP-MoS2 heterojunction was exfoliated from bulk crystals (procured from HQ
Graphene) with BP/MoS2 layer thicknesses of 36/9 nm, which were measured with an atomic force microscope. The approximate overlap/active area of the BP-MoS2 is 30 × 30 µm2
for both detectors. One
photodiode was aligned to the center of the metalens (i.e., the location of the focal spot), and the other
photodiode was fabricated onto a blank Si substrate.
67

Figure 5.5: Metalens-integrated black phosphorus photodiodes. a. Photodiode with an integrated metalens:
impinging light is focused onto the active region, which increases the incident power on the detector’s
active region. b. Photodiode with no integrated metalens: impinging light is not focused onto the active
region. c. Measured absorption spectrum (A = 1 – R – T) of a black phosphorus molybdenum disulfide
photodetector along the high absorption (armchair, AC) and low absorption (zigzag, ZZ) axes of the device.
Inset: Scanning electron microscopy image of the photodetector.
BP-MoS2 Photodetectors were fabricated onto the back side of the Si substrates. BP and MoS2 flakes
were mechanically exfoliated from bulk crystals (procured from HQ Graphene) onto a polydimethylsiloxane (PDMS) sheet in an O2- and H2O-free glovebox. The thickness of the layered material samples with
optical images were estimated; the pigments of BP/MoS2 flakes were thickness dependent. BP/MoS2 flakes
with thicknesses ≈35/10 nm were identified for the heterojunction and transferred onto the substrates.
Then, 50 nm Al2O3 dielectric insulators were patterned into a PMMA resist layer with EBL (Raith) and
deposited with e-beam deposition (Angstrom). Finally, metallic contacts were deposited on top of the
dielectric insulators by first patterning a PMMA resist layer with EBL (Raith) and then successively depositing 5 nm of Ti and 60 nm of Au with e-beam deposition (CHA Industries). Photodetectors that were
not integrated with the metalens were fabricated onto blank substrates, i.e., substrates with no metalens
etched into its front side.
After fabricating the photodetectors, we measured and identified the high- and low-absorption axes
(armchair/AC and zigzag/ZZ, respectively) of BP with a linearly polarized Fourier-transform infrared spectrometer. [63, 122] Figure 5.5c shows the absorption, estimated by A = 1 – T – R, of one BP-MoS2 photodiode along its AC and ZZ axes. The absorption is notably higher along the AC axis than the ZZ axis. The
68

Figure 5.6: Room temperature metalens-enhanced photodetection. a. Measured current as a function of
voltage bias for a photodetector with no integrated metalens: I Of f is the dark current and I On is the current
under λ = 3.39 or 3.88 µm illumination.a. I Of f and I On for a photodetector with an integrated metalens:
I On exceeds that of the photodetector with no integrated metalens. c. Generated photocurrent IP H = I On
– I Of f and responsivity at λ = 3.39 µm for both the metalens-enhanced and unenhanced photodetectors.
d. Generated photocurrent IPH and responsivity at λ = 3.88 µm for both the metalens-integrated and
unintegrated photodetectors. At λ = 3.39 and 3.88 µm, the metalens-integrated photodetector exhibits
higher generated photocurrent and responsivity than the photodetector with no integrated metalens. All
photocurrent measurements were performed at room temperature and with linearly polarized incident
light aligned to the high absorption (AC axis) of each photodetector.
inset of Figure 5.5c is an SEM image of the photodetector after depositing Ti/Au contacts. Even though the
BP-MoS2 stack rests on undoped silicon (float-zone), we suspended the metallic contacts with Al2O3 to
ensure electrical isolation between the metal and substrate. After fabricating both the metalens-enhanced
and unenhanced BP-MoS2 photodiodes, we measured photocurrent generation in both detectors at room
temperature.
We tested the two photodiodes shown in Figure 5.5a and 5.5b at two laser wavelengths λ = 3.39 µm
and 3.88 µm. Linearly polarized radiation from both lasers (Thorlabs) was fixed to ≈0.84 W cm−2
(≈ 7.5
69

µW of incident power on the active area) and modulated at 200 Hz with an optical chopper (Thorlabs).
Both photodetectors were illuminated with incident light aligned to their high absorption AC axes. A voltage bias was applied across the heterojunction with a source measure unit (Keithley), and the generated
photocurrent was extracted with a lock-in amplifier (Stanford Research Systems) at room temperature.
Over the duration of our measurements, we did not observe any significant degradation in device performance, which we attribute to passivation provided by the MoS2 layer. [150] Figure 5.6a shows the dark
current I Of f and current with λ = 3.39 and 3.88 µm illumination I On of the photodiode with no integrated
metalens. Similarly, Figure 5.6b plots I Of f and I On at λ = 3.39 and 3.88 µm for the photodiode with the
integrated metalens. I On is notably higher in the photodiode with the integrated metalens, and I Of f is
similar in both devices. The IV curves in Figure 5.6a,b reveal that both devices exhibit partial photoconductor behavior rather than exclusive diode behavior. [117] In previous studies, BP-MoS2 heterojuction
devices have been shown to exhibit simultaneous diode and photoconductor behavior. Figures 5.6c and
5.6d show the generated photocurrent and extracted responsivity of both photodiodes at λ = 3.39 and 3.88
µm, respectively. Photocurrent is defined as the total current with subtracted dark current (IP H = I On -
I Of f ), and responsivity is defined as the ratio of generated photocurrent to incident power. Photocurrent
and responsivity measured with λ = 3.39 µm light focused by the metalens was ≈6.65 times higher than in
the case with no metalens. Additionally, photocurrent and responsivity were increased by ≈6.31 times at λ
= 3.88 µm in the presence of the metalens. To evaluate how the integrated metalens increased SNR at both
testing wavelengths, we calculated specific detectivity D* for both photodiodes with the responsivity and
dark current data in Figure 5.6. D* is calculated with D* = R A1/2
(2eI Of f )
−1/2
, where R is responsivity, A
is the detector area, e is the elementary charge constant, and I Of f is the dark current. The detector with
no metalens exhibited D* = 3.03 × 108
and 1.56 × 108
cm Hz1/2 W−1
at λ = 3.39 and 3.88 µm, respectively.
And the detector with the integrated metalens exhibited D* = 2.02 × 109
and 9.81 × 108
cm Hz1/2 W−1
at λ
70

and 3.88 µm, respectively. At both λ = 3.39 and 3.88 µm, the integrated metalens enhanced photodetector
SNR.
5.5 Summary and outlook
We integrated an all-silicon metalens with a BP photodetector to enhance the optical signal collection in
the MWIR. We characterized the metalens by imaging clear focal spots at various testing wavelengths and
showed that it increases power density by focusing light through the substrate. Then, we integrated a thinfilm BP-MoS2 heterojunction photodetector with the metalens. We showed that the metalens increases
room temperature performance at multiple testing wavelengths in the MWIR. In conclusion, we demonstrated the first successful integration of a dielectric metalens with a 2D layered material photodetector
in the MWIR. Overall, our results establish integrated metalenses as an excellent and scalable approach to
enhancing BP photodetectors with limited signal collection.
71

Bibliography
[1] Konstantin S. Novoselov, Artem Mishchenko, Alexandra Carvalho, and Antonio H. Castro Neto.
“2D materials and van der Waals heterostructures”. In: Science 353.6298 (2016), aac9439.
[2] Andre K. Geim. “Graphene: Status and Prospects”. In: Science 324.5934 (2009), pp. 1530–1534.
[3] Andre K. Geim and Konstantin S. Novoselov. “The rise of graphene”. In: Nature Materials 324
(2007), pp. 183–191.
[4] Na Li, Qinqin Wang, Cheng Shen, Zheng Wei, Hua Yu, Jing Zhao, Xiaobo Lu, Guole Wang,
Congli He, and Li Xie et al. “Large-scale flexible and transparent electronics based on monolayer
molybdenum disulfide field-effect transistors”. In: Nature Electronics 3.11 (2020), pp. 711–717.
[5] Fengnian Xia, Han Wang, Di Xiao, Madan Dubey, and Ashwin Ramasubramaniam.
“Two-dimensional material nanophotonics”. In: Nature Photonics 8 (2014), pp. 899–907.
[6] Jean-Christophe Blancon, Matthieu Paillet, Huy N. Tran, Xuan T. Than, Samuel A. Guebrou,
Anthony Ayari, Alfonso S. Miguel, Ngoc-Minh Phan, Ahmed-Azmi Zahab, and
Jean-Louis Sauvajol et al. “Direct measurement of the absolute absorption spectrum of individual
semiconducting single-wall carbon nanotubes”. In: Nature Communications 4.2542 (2013).
[7] Sheng Wang, Sihan Zhao, Zhiwen Shi, Fanqi Wu, Zhiyuan Zhao, Lili Jiang, Kenji Watanabe,
Takashi Taniguchi, Alex Zettl, and Chongwu Zhou et al. “Nonlinear Luttinger liquid plasmons in
semiconducting single-walled carbon nanotubes”. In: Nature Materials 19 (2020), pp. 986–991.
[8] A. Kis, G. Csányi, J.-P. Salvetat, Thien-Nga Lee, E. Couteau, A. J. Kulik, W. Benoit, J. Brugger, and
L. Forró. “Reinforcement of single-walled carbon nanotube bundles by intertube bridging”. In:
Nature Materials 3 (2004), pp. 153–157.
[9] Jie Xiang, Wei Lu, Yongjie Hu, Yue Wu, Hao Yan, and Charles M. Lieber. “Ge/Si nanowire
heterostructures as high-performance field-effect transistors”. In: Nature 441 (2006), pp. 489–493.
[10] Luojun Du, Yanchong Zhao, Linlu Wu, Xuerong Hu, Lide Yao, Yadong Wang, Xueyin Bai,
Yunyun Dai, Jingsi Qiao, and Md Gius Uddin et al. “Giant anisotropic photonics in the 1D van der
Waals semiconductor fibrous red phosphorus”. In: Nature Communications 12.4822 (2021).
72

[11] Jing-Kai Qin, Pai-Ying Liao, Mengwei Si, Shiyuan Gao, Gang Qiu, Jie Jian, Qingxiao Wang,
Si-Qi Zhang, Shouyuan Huang, and Adam Charnas et al. “Raman response and transport
properties of tellurium atomic chains encapsulated in nanotubes”. In: Nature Electronics 3 (2020),
pp. 141–147.
[12] Shaohua Liu, Lu Han, Yingying Duan, Shunsuke Asahina, Osamu Terasaki, Yuanyuan Cao,
Ben Liu, Liguo Ma, Jialiang Zhang, and Shunai Che. “Synthesis of chiral TiO2 nanofibre with
electron transition-based optical activity”. In: Nature Communications 3.1215 (2012).
[13] James D. Watson and Francis H. C. Crick. “Molecular Structure of Nucleic Acids: A Structure for
Deoxyribose Nucleic Acid”. In: Nature 171 (1953), pp. 737–738.
[14] Dang Sheng Su. “Inorganic Materials with Double-Helix Structures”. In: Angewandte Chemie
International Edition 50.21 (2011), pp. 4747–4750.
[15] Volker Berl, Ivan Huc, Richard G Khoury, Michael J Krische, and Jean-Marie Lehn.
“Interconversion of single and double helices formed from synthetic molecular strands”. In:
Nature 407.6805 (2000), pp. 720–723.
[16] Julius Marmur. “A procedure for the isolation of deoxyribonucleic acid from micro-organisms”.
In: Journal of molecular biology 3.2 (1961), 208–IN1.
[17] Watson Jd and Crick Fh. “Genetical implications of the structure of deoxyribonucleic acid.” In:
Nature 171.4361 (1953), pp. 964–967.
[18] Xinru Li, Ying Dai, Yandong Ma, Mengmeng Li, Lin Yu, and Baibiao Huang. “Landscape of
DNA-like inorganic metal free double helical semiconductors and potential applications in
photocatalytic water splitting”. In: Journal of Materials Chemistry A 5 (2017), pp. 8484–8492.
[19] Haruhiko Morito and Hisanori Yamane. “Double-Helical Silicon Microtube”. In: Angewandte
Chemie International Edition 49.21 (2010), pp. 3638–3641.
[20] Yuanyuan Shang, Yibin Li, Xiaodong He, Shanyi Du, Luhui Zhang, Enzheng Shi, Shiting Wu,
Zhen Li, Peixu Li, and Jinquan Wei et al. “Highly Twisted Double-Helix Carbon Nanotube Yarns”.
In: ACS Nano 7.2 (2013), pp. 1446–1453.
[21] Sanghyun Ju, Antonio Facchetti, Yi Xuan, Jun Liu, Fumiaki Ishikawa, Peide Ye, Chongwu Zhou,
Tobin J. Marks, and David B. Janes. “Fabrication of fully transparent nanowire transistors for
transparent and flexible electronics”. In: Nature Nanotechnology 2 (2007), pp. 378–384.
[22] Yin Liu, Jie Wang, Sujung Kim, Haoye Sun, Fuyi Yang, Zixuan Fang, Nobumichi Tamura,
Ruopeng Zhang, Xiaohui Song, and Jianguo Wen et al. “Helical van der Waals crystals with
discretized Eshelby twist”. In: Nature 570 (2019), pp. 358–362.
[23] Daniela Pfister, Konrad Schäfer, Claudia Ott, Birgit Gerke, Rainer Pöttgen, Oliver Janka,
Maximilian Baumgartner, Anastasia Efimova, Andrea Hohmann, and Peer Schmidt et al.
“Inorganic Double Helices in Semiconducting SnIP”. In: Advanced Materials 28 (2016),
pp. 9783–9791.
73

[24] Jean M. Lehn, Annie Rigault, Jay S. Siegel, Jack Harrowfield, Bernard Chevrier, and Dino Moras.
“Spontaneous assembly of double-stranded helicates from oligobipyridine ligands and copper(I)
cations: structure of an inorganic double helix”. In: Proceedings of the National Academy of
Sciences 84.9 (1987), pp. 2565–2569.
[25] Victoria Soghomonian, Qin Chen, Robert C. Haushalter, Jon Zubieta, and Charles J. O’Connor.
“An Inorganic Double Helix: Hydrothermal Synthesis, Structure, and Magnetism of Chiral
[(CH3)2NH2]K4[V10O10(H2O)2(OH)4(PO4)7]·4H2O”. In: Science 259.5101 (1993), pp. 1596–1599.
[26] Alexander S Ivanov, Andrew J Morris, Konstantin V Bozhenko, Chris J Pickard, and
Alexander I Boldyrev. “Inorganic Double-Helix Structures of Unusually Simple
Lithium–Phosphorus Species”. In: Angewandte Chemie-International Edition 51.33 (2012), p. 8330.
[27] W Hönle and HG von Schnering. “Zur struktur von lip und ksb”. In: Zeitschrift für
Kristallographie-Crystalline Materials 155.1-4 (1981), pp. 307–314.
[28] Meng-Qiang Zhao, Qiang Zhang, Gui-Li Tian, and Fei Wei. “Emerging double helical
nanostructures”. In: Nanoscale 6 (2014), pp. 9339–9354.
[29] Seiji Motojima, Masayuki Kawaguchi, Koji Nozaki, and Hiroshi Iwanaga. “Growth of regularly
coiled carbon filaments by Ni catalyzed pyrolysis of acetylene, and their morphology and
extension characteristics”. In: Applied Physics Letters 56 (1990), pp. 321–323.
[30] Yuzhou Zhao, Chenyu Zhang, Daniel D. Kohler, Jason M. Scheeler, John C. Wright,
Paul M. Voyles, and Song Jin. “Supertwisted spirals of layered materials enabled by growth on
non-Euclidean surfaces”. In: Science 370.6515 (2020), pp. 442–445.
[31] Deep Jariwala, Vinod K. Sangwan, Lincoln J. Lauhon, Tobin J. Marksab, and Mark C. Hersam.
“Carbon nanomaterials for electronics, optoelectronics, photovoltaics, and sensing”. In: Chemical
Society Reviews 42 (2013), pp. 2824–2860.
[32] Yibin Zhao, Zhengwei Du, Licheng Wang, Mingyan Liu, Bing Yao, Xudong Hu, Libo Gao,
Fang Wu, Cong Liu, Xiaoming Li, et al. “Revealing the Optical Anisotropy of ReS2 Nanostructures
by Strain Engineering”. In: ACS Applied Nano Materials (2023).
[33] J Ortega-Gallegos, A Lastras-Martınez, LF Lastras-Martınez, and RE Balderas-Navarro. “Optical
anisotropy induced by mechanical strain around the fundamental gap of GaAs”. In: physica status
solidi c 5.8 (2008), pp. 2561–2564.
[34] Mikhail A Kats, Patrice Genevet, Guillaume Aoust, Nanfang Yu, Romain Blanchard,
Francesco Aieta, Zeno Gaburro, and Federico Capasso. “Giant birefringence in optical antenna
arrays with widely tailorable optical anisotropy”. In: Proceedings of the National Academy of
Sciences 109.31 (2012), pp. 12364–12368.
[35] Wenfu Zhang, Jihong Liu, Wei-Ping Huang, and Wei Zhao. “Giant birefringence of periodic
dielectric waveguides”. In: IEEE Photonics Journal 3.3 (2011), pp. 512–520.
[36] Abhishek Swarnkar, Wasim J. Mir, Rayan Chakraborty, Metikoti Jagadeeswararao, Tariq Sheikh,
and Angshuman Nag. “Are Chalcogenide Perovskites an Emerging Class of Semiconductors for
Optoelectronic Properties and Solar Cell?” In: Chemistry of Materials 31.3 (2019), pp. 565–575.
74

[37] Jian Wang and Kirill Kovnir. “Giant anisotropy detected”. In: Nature Photonics 12 (2018),
pp. 378–385.
[38] Bo Sun, Shanyuan Niu, Raphael P. Hermann, Jaeyun Moon, Nina Shulumba, Katharine Page,
Boyang Zhao, Arashdeep S. Thind, Krishnamurthy Mahalingam, and
JoAnna Milam-Guerrero et al. “High frequency atomic tunneling yields ultralow and glass-like
thermal conductivity in chalcogenide single crystals”. In: Nature Communications 11.6039 (2020).
[39] Shanyuan Niu, Graham Joe, Huan Zhao, Yucheng Zhou, Thomas Orvis, Huaixun Huyan,
Jad Salman, Krishnamurthy Mahalingam, Brittany Urwin, and Jiangbin Wu et al. “Giant optical
anisotropy in a quasi-one-dimensional crystal”. In: Nature Photonics 12 (2018), pp. 392–396.
[40] Jiangbin Wu, Xin Cong, Shanyuan Niu, Fanxin Liu, Huan Zhao, Zhonghao Du,
Jayakanth Ravichandran, Ping-Heng Tan, and Han Wang. “Linear Dichroism Conversion in
Quasi-1D Perovskite Chalcogenide”. In: Advanced Materials 31.33 (2019), p. 1902118.
[41] Boyang Zhao, Md S. B. Hoque, Gwan Y. Jung, Hongyan Mei, Shantanu Singh, Guodong Ren,
Milena Milich, Qinai Zhao, Nan Wang, and Huandong Chen. “Orientation-Controlled Anisotropy
in Single Crystals of Quasi-1D BaTiS3”. In: Chemistry of Materials (2022).
[42] Marcia J Rieke, Douglas M Kelly, and Scott D Horner. “Overview of James Webb Space Telescope
and NIRCam’s Role”. In: Cryogenic Optical Systems and Instruments XI. Vol. 5904. SPIE. 2005,
p. 590401.
[43] Thomas S Pagano. CubeSats and NanoSats for Remote Sensing. Vol. 9978. 2016.
[44] Jiangbin Wu, Nan Wang, Xiaodong Yan, and Han Wang. “Emerging low-dimensional materials
for mid-infrared detection”. In: Nano Research 14.6 (2021), pp. 1863–1877.
[45] Qiushi Guo, Andreas Pospischil, Maruf Bhuiyan, Hao Jiang, He Tian, Damon Farmer,
Bingchen Deng, Cheng Li, Shu-Jen Han, Han Wang, et al. “Black phosphorus mid-infrared
photodetectors with high gain”. In: Nano letters 16.7 (2016), pp. 4648–4655.
[46] Fengnian Xia, Han Wang, and Yichen Jia. “Rediscovering black phosphorus as an anisotropic
layered material for optoelectronics and electronics”. In: Nature communications 5.1 (2014),
p. 4458.
[47] Xiaomu Wang, Aaron M Jones, Kyle L Seyler, Vy Tran, Yichen Jia, Huan Zhao, Han Wang,
Li Yang, Xiaodong Xu, and Fengnian Xia. “Highly anisotropic and robust excitons in monolayer
black phosphorus”. In: Nature nanotechnology 10.6 (2015), pp. 517–521.
[48] Likai Li, Jonghwan Kim, Chenhao Jin, Guo Jun Ye, Diana Y Qiu, Felipe H Da Jornada, Zhiwen Shi,
Long Chen, Zuocheng Zhang, Fangyuan Yang, et al. “Direct observation of the layer-dependent
electronic structure in phosphorene”. In: Nature nanotechnology 12.1 (2017), pp. 21–25.
[49] Tony Low, Rafael Roldán, Han Wang, Fengnian Xia, Phaedon Avouris, Luis Martın Moreno, and
Francisco Guinea. “Plasmons and screening in monolayer and multilayer black phosphorus”. In:
Physical review letters 113.10 (2014), p. 106802.
[50] A Morita. “Semiconducting black phosphorus”. In: Applied Physics A 39 (1986), pp. 227–242.
75

[51] Matin Amani, Emma Regan, James Bullock, Geun Ho Ahn, and Ali Javey. “Mid-wave infrared
photoconductors based on black phosphorus-arsenic alloys”. In: ACS nano 11.11 (2017),
pp. 11724–11731.
[52] Bilu Liu, Marianne Köpf, Ahmad N Abbas, Xiaomu Wang, Qiushi Guo, Yichen Jia, Fengnian Xia,
Richard Weihrich, Frederik Bachhuber, Florian Pielnhofer, et al. “Black arsenic–phosphorus:
layered anisotropic infrared semiconductors with highly tunable compositions and properties”.
In: Advanced Materials 27.30 (2015), pp. 4423–4429.
[53] Xiaolong Chen, Xiaobo Lu, Bingchen Deng, Ofer Sinai, Yuchuan Shao, Cheng Li, Shaofan Yuan,
Vy Tran, Kenji Watanabe, Takashi Taniguchi, et al. “Widely tunable black phosphorus
mid-infrared photodetector”. In: Nature communications 8.1 (2017), p. 1672.
[54] Shaofan Yuan, Doron Naveh, Kenji Watanabe, Takashi Taniguchi, and Fengnian Xia. “A
wavelength-scale black phosphorus spectrometer”. In: Nature Photonics 15.8 (2021), pp. 601–607.
[55] Binjie Zheng, Junzhuan Wang, Tianye Huang, Xin Su, Yi Shi, and Xiaomu Wang. “Single-detector
black phosphorus monolithic spectrometer with high spectral and temporal resolution”. In:
Applied Physics Letters 120.25 (2022).
[56] Souvik Biswas, Meir Y Grajower, Kenji Watanabe, Takashi Taniguchi, and Harry A Atwater.
“Broadband electro-optic polarization conversion with atomically thin black phosphorus”. In:
Science 374.6566 (2021), pp. 448–453.
[57] Bingchen Deng, Vy Tran, Yujun Xie, Hao Jiang, Cheng Li, Qiushi Guo, Xiaomu Wang, He Tian,
Steven J Koester, Han Wang, et al. “Efficient electrical control of thin-film black phosphorus
bandgap”. In: Nature communications 8.1 (2017), p. 14474.
[58] Shifan Wang, Robert James Chapman, Brett C Johnson, Inna Krasnokutska, Jean-Luc J Tambasco,
Kibret Messalea, Alberto Peruzzo, and James Bullock. “Integration of black phosphorus
photoconductors with lithium niobate on insulator photonics”. In: Advanced Optical Materials
11.2 (2023), p. 2201688.
[59] Li Huang, Bowei Dong, Xin Guo, Yuhua Chang, Nan Chen, Xin Huang, Wugang Liao,
Chunxiang Zhu, Hong Wang, Chengkuo Lee, et al. “Waveguide-integrated black phosphorus
photodetector for mid-infrared applications”. In: ACS nano 13.1 (2018), pp. 913–921.
[60] Nathan Youngblood, Che Chen, Steven J Koester, and Mo Li. “Waveguide-integrated black
phosphorus photodetector with high responsivity and low dark current”. In: Nature photonics 9.4
(2015), pp. 247–252.
[61] Che Chen, Nathan Youngblood, Ruoming Peng, Daehan Yoo, Daniel A Mohr, Timothy W Johnson,
Sang-Hyun Oh, and Mo Li. “Three-dimensional integration of black phosphorus photodetector
with silicon photonics and nanoplasmonics”. In: Nano letters 17.2 (2017), pp. 985–991.
[62] Prabhu K Venuthurumilli, Peide D Ye, and Xianfan Xu. “Plasmonic resonance enhanced
polarization-sensitive photodetection by black phosphorus in near infrared”. In: ACS nano 12.5
(2018), pp. 4861–4867.
76

[63] Max R Lien, Nan Wang, Jiangbin Wu, Alexander Soibel, Sarath D Gunapala, Han Wang, and
Michelle L Povinelli. “Resonant Grating-Enhanced Black Phosphorus Mid-Wave Infrared
Photodetector”. In: Nano Letters 22.21 (2022), pp. 8704–8710.
[64] Ruijuan Tian, Linpeng Gu, Yingke Ji, Chen Li, Yuxin Chen, Siqi Hu, Zhiwen Li, Xuetao Gan, and
Jianlin Zhao. “Black phosphorus photodetector enhanced by a planar photonic crystal cavity”. In:
ACS Photonics 8.10 (2021), pp. 3104–3110.
[65] Wei Yan, Vivek Raj Shresha, Quentin Jeangros, Nima Sefidmooye Azar,
Sivacarendran Balendhran, Christophe Ballif, Kenneth Crozier, and James Bullock. “Spectrally
selective mid-wave infrared detection using Fabry-Pérot cavity enhanced black phosphorus 2D
photodiodes”. In: ACS nano 14.10 (2020), pp. 13645–13651.
[66] Shyam Narayan Singh Yadav, Po-Liang Chen, Chang Hua Liu, and Ta-Jen Yen. “Plasmonic
Metasurface Integrated Black Phosphorus-Based Mid-Infrared Photodetector with High
Responsivity and Speed”. In: Advanced Materials Interfaces 10.10 (2023), p. 2202403.
[67] Romil Audhkhasi and Michelle L Povinelli. “Gold-black phosphorus nanostructured absorbers for
efficient light trapping in the mid-infrared”. In: Optics Express 28.13 (2020), pp. 19562–19570.
[68] Xianliang Liu, Tatiana Starr, Anthony F Starr, and Willie J Padilla. “Infrared spatial and frequency
selective metamaterial with near-unity absorbance”. In: Physical review letters 104.20 (2010),
p. 207403.
[69] Yanko Todorov, Lorenzo Tosetto, Jean Teissier, Aaron Maxwell Andrews, Pavel Klang,
Raffaele Colombelli, Isabelle Sagnes, Gottfried Strasser, and Carlo Sirtori. “Optical properties of
metal-dielectric-metal microcavities in the THz frequency range”. In: Optics express 18.13 (2010),
pp. 13886–13907.
[70] Jun Wu, Changhe Zhou, Junjie Yu, Hongchao Cao, Shubin Li, and Wei Jia. “TE polarization
selective absorber based on metal-dielectric grating structure for infrared frequencies”. In: Optics
Communications 329 (2014), pp. 38–43.
[71] Shuyan Zhang, Alexander Soibel, Sam A Keo, Daniel Wilson, Sir Rafol, David Z Ting, Alan She,
Sarath D Gunapala, Federico Capasso, et al. “Solid-immersion metalenses for infrared focal plane
arrays”. In: Applied Physics Letters 113.11 (2018).
[72] Tobias Wenger, Richard Muller, Cory J Hill, Anita Fisher, David Z Ting, Daniel Wilson,
Sarath D Gunapala, and Alexander Soibel. “Infrared nBn detectors monolithically integrated with
metasurface-based optical concentrators”. In: Applied Physics Letters 121.18 (2022).
[73] Tobias Wenger, Richard Muller, Daniel Wilson, Sarath D Gunapala, and Alexander Soibel. “Large
metasurface-based optical concentrators for infrared photodetectors”. In: AIP Advances 11.8
(2021).
[74] Markus RP Pielmeier and Tom Nilges. “Formation mechanisms for phosphorene and SnIP”. In:
Angewandte Chemie International Edition 60.12 (2021), pp. 6816–6823.
[75] André Utrap, Ng Yan Xiang, and Tom Nilges. “A yield-optimized access to double-helical SnIP via
a Sn/SnI2 approach”. In: Journal of Crystal Growth 475 (2017), pp. 341–345.
77

[76] Jiangbin Wu, Hung-Yu Chen, Ning Yang, Jun Cao, Xiaodong Yan, Fanxin Liu, Qibin Sun, Xi Ling,
Jing Guo, and Han Wang. “High tunnelling electroresistance in a ferroelectric van der Waals
heterojunction via giant barrier height modulation”. In: Nature Electronics 3 (2020), pp. 466–472.
[77] Minhao He, Yuhao Li, Jiaqi Cai, Yang Liu, Kenji Watanabe, Takashi Taniguchi, Xiaodong Xu, and
Matthew Yankowitz. “Symmetry breaking in twisted double bilayer graphene”. In: Nature Physics
17 (2020), pp. 26–30.
[78] Shiliang Wang, Zhiwei Shan, and Han Huang. “The Mechanical Properties of Nanowires”. In:
Advanced Science 4.4 (2017), p. 1600332.
[79] Bin Wu, Andreas Heidelberg, and John J. Boland. “Mechanical properties of ultrahigh-strength
gold nanowires”. In: Nature Materials 4 (2005), pp. 525–529.
[80] James M. Gere and Stephen P. Timoshenko. Mechanics of materials. PWS Publishing Company,
1997.
[81] Georg Kresse and Daniel Joubert. “From ultrasoft pseudopotentials to the projector
augmented-wave method”. In: PHYSICAL REVIEW B 59.3 (1999).
[82] Georg Kresse and Jürgen Furthmüller. “Efficient iterative schemes for ab initio total-energy
calculations using a plane-wave basis set”. In: PHYSICAL REVIEW B 54.16 (1996).
[83] Peter E. Blöchl. “Projector augmented-wave method”. In: PHYSICAL REVIEW B 50.24 (1994).
[84] VI Ivashchenko, PEA Turchi, and VI Shevchenko. “Simulations of the mechanical properties of
crystalline, nanocrystalline, and amorphous SiC and Si”. In: Physical Review B 75.8 (2007),
p. 085209.
[85] David Roundy and Marvin L Cohen. “Ideal strength of diamond, Si, and Ge”. In: Physical Review B
64.21 (2001), p. 212103.
[86] Yu-Fei Wu, Ying-Wu Zhou, Biao Hu, Xiaoxu Huang, and Scott Smith. “Fused structures for safer
and more economical constructions”. In: Frontiers of Structural and Civil Engineering 14.1 (2020),
pp. 1–9.
[87] Stephen C. Cusack and Andrew Miller. “Determination of the elastic constants of collagen by
Brillouin light scattering”. In: Journal of Molecular Biology 135.1 (1979), pp. 39–51.
[88] Jia-Min Lai, Ya-Ru Xie, and Jun Zhang. “Detection of electron-phonon coupling in
two-dimensional materials by light scattering”. In: Nano Research 14 (2021), pp. 1711–1733.
[89] Louis R. Testardi and Joachim J. Hauser. “Sound velocity in amorphous Ge and Si”. In: Solid State
Communications 21.11 (1977), pp. 1039–1041.
[90] Claudia Ott, Felix Reiter, Maximilian Baumgartner, Markus Pielmeier, Anna Vogel, Patrick Walke,
Stefan Burger, Michael Ehrenreich, Gregor Kieslich, and Dominik Daisenberger et. al. “Flexible
and Ultrasoft Inorganic 1D Semiconductor and Heterostructure Systems Based on SnIP”. In: Solid
State Communications 29.18 (2019), p. 1900233.
[91] Earl R. Parker. MATERIALS DATA BOOK. For Engineers and Scientists. McGraw-Hill, 1967.
78

[92] Dai-Ming Tang, Cui-Lan Ren, Ming-Sheng Wang, Xianlong Wei, Naoyuki Kawamoto, Chang Liu,
Yoshio Bando, Masanori Mitome, Naoki Fukata, and Dmitri Golberg. “Mechanical properties of Si
nanowires as revealed by in situ transmission electron microscopy and molecular dynamics
simulations”. In: Nano letters 12.4 (2012), pp. 1898–1904.
[93] Lihua Wang, Kun Zheng, Ze Zhang, and Xiaodong Han. “Direct atomic-scale imaging about the
mechanisms of ultralarge bent straining in Si nanowires”. In: Nano letters 11.6 (2011),
pp. 2382–2385.
[94] Hongti Zhang, Jerry Tersoff, Shang Xu, Huixin Chen, Qiaobao Zhang, Kaili Zhang, Yong Yang,
Chun-Sing Lee, King-Ning Tu, Ju Li, et al. “Approaching the ideal elastic strain limit in silicon
nanowires”. In: Science advances 2.8 (2016), e1501382.
[95] Jun Peng and G Jeffrey Snyder. “A figure of merit for flexibility”. In: Science 366.6466 (2019),
pp. 690–691.
[96] Tian-Ran Wei, Min Jin, Yuecun Wang, Hongyi Chen, Zhiqiang Gao, Kunpeng Zhao, Pengfei Qiu,
Zhiwei Shan, Jun Jiang, Rongbin Li, et al. “Exceptional plasticity in the bulk single-crystalline van
der Waals semiconductor InSe”. In: Science 369.6503 (2020), pp. 542–545.
[97] AP Rooney, Zheling Li, Wen Zhao, Ali Gholinia, Aleksey Kozikov, G Auton, Feng Ding,
RV Gorbachev, RJ Young, and SJ Haigh. “Anomalous twin boundaries in two dimensional
materials”. In: Nature communications 9.1 (2018), pp. 1–7.
[98] Xun Shi, Hongyi Chen, Feng Hao, Ruiheng Liu, Tuo Wang, Pengfei Qiu, Ulrich Burkhardt,
Yuri Grin, and Lidong Chen. “Room-temperature ductile inorganic semiconductor”. In: Nature
Materials 17.5 (2018), pp. 421–426.
[99] Michael F Weber, Carl A Stover, Larry R Gilbert, Timothy J Nevitt, and Andrew J Ouderkirk.
“Giant birefringent optics in multilayer polymer mirrors”. In: Science 287.5462 (2000),
pp. 2451–2456.
[100] Sina Abedini Dereshgi, Thomas G Folland, Akshay A Murthy, Xianglian Song, Ibrahim Tanriover,
Vinayak P Dravid, Joshua D Caldwell, and Koray Aydin. “Lithography-free IR polarization
converters via orthogonal in-plane phonons in α-MoO3 flakes”. In: Nature communications 11.1
(2020), p. 5771.
[101] Weiliang Ma, Pablo Alonso-González, Shaojuan Li, Alexey Y Nikitin, Jian Yuan,
Javier Martın-Sánchez, Javier Taboada-Gutiérrez, Iban Amenabar, Peining Li, Saül Vélez, et al.
“In-plane anisotropic and ultra-low-loss polaritons in a natural van der Waals crystal”. In: Nature
562.7728 (2018), pp. 557–562.
[102] Weiliang Ma, Guangwei Hu, Debo Hu, Runkun Chen, Tian Sun, Xinliang Zhang, Qing Dai,
Ying Zeng, Andrea Alù, Cheng-Wei Qiu, et al. “Ghost hyperbolic surface polaritons in bulk
anisotropic crystals”. In: Nature 596.7872 (2021), pp. 362–366.
[103] Lei Tong, Xinyu Huang, Peng Wang, Lei Ye, Meng Peng, Licong An, Qiaodong Sun, Yong Zhang,
Guoming Yang, Zheng Li, et al. “Stable mid-infrared polarization imaging based on quasi-2D
tellurium at room temperature”. In: Nature communications 11.1 (2020), pp. 1–10.
79

[104] Chenfei Shen, Yihang Liu, Jiangbin Wu, Chi Xu, Dingzhou Cui, Zhen Li, Qingzhou Liu,
Yuanrui Li, Yixiu Wang, Xuan Cao, et al. “Tellurene photodetector with high gain and wide
bandwidth”. In: ACS nano 14.1 (2019), pp. 303–310.
[105] Shengxue Yang, Yanhan Yang, Minghui Wu, Chunguang Hu, Wanfu Shen, Yongji Gong, Li Huang,
Chengbao Jiang, Yongzhe Zhang, and Pulickel M Ajayan. “Highly in-plane optical and electrical
anisotropy of 2D germanium arsenide”. In: Advanced Functional Materials 28.16 (2018), p. 1707379.
[106] Jiangbin Wu, Xin Cong, Shanyuan Niu, Fanxin Liu, Huan Zhao, Zhonghao Du,
Jayakanth Ravichandran, Ping-Heng Tan, and Han Wang. “Linear Dichroism Conversion in
Quasi-1D Perovskite Chalcogenide”. In: Advanced Materials 31.33 (2019), p. 1902118.
[107] Parikshit Sahatiya, S Solomon Jones, Venkat Mattela, and Sushmee Badhulika. “Direct growth of
black phosphorus (p-type) on a flexible substrate with dual role of two-dimensional ZnO (n-type)
as effective passivation and enabling highly stable broadband photodetection”. In: ACS Applied
Electronic Materials 1.7 (2019), pp. 1076–1083.
[108] Likai Li, Yijun Yu, Guo Jun Ye, Qingqin Ge, Xuedong Ou, Hua Wu, Donglai Feng, Xian Hui Chen,
and Yuanbo Zhang. “Black phosphorus field-effect transistors”. In: Nature nanotechnology 9.5
(2014), pp. 372–377.
[109] Xuesong Li, Bingchen Deng, Xiaomu Wang, Sizhe Chen, Michelle Vaisman, Shun-ichiro Karato,
Grace Pan, Minjoo Larry Lee, Judy Cha, Han Wang, et al. “Synthesis of thin-film black
phosphorus on a flexible substrate”. In: 2D Materials 2.3 (2015), p. 031002.
[110] A Mogulkoc, YEŞİM Mogulkoc, AN Rudenko, and MI Katsnelson. “Polaronic effects in monolayer
black phosphorus on polar substrates”. In: Physical Review B 93.8 (2016), p. 085417.
[111] Mingsheng Long, Anyuan Gao, Peng Wang, Hui Xia, Claudia Ott, Chen Pan, Yajun Fu, Erfu Liu,
Xiaoshuang Chen, Wei Lu, et al. “Room temperature high-detectivity mid-infrared photodetectors
based on black arsenic phosphorus”. In: Science advances 3.6 (2017), e1700589.
[112] Mingqiang Huang, Mingliang Wang, Cheng Chen, Zongwei Ma, Xuefei Li, Junbo Han, and
Yanqing Wu. “Broadband black-phosphorus photodetectors with high responsivity”. In: Advanced
Materials 28.18 (2016), pp. 3481–3485.
[113] Dinh Loc Duong, Seok Joon Yun, and Young Hee Lee. “van der Waals layered materials:
opportunities and challenges”. In: ACS nano 11.12 (2017), pp. 11803–11830.
[114] Tu Hong, Bhim Chamlagain, Tianjiao Wang, Hsun-Jen Chuang, Zhixian Zhou, and Ya-Qiong Xu.
“Anisotropic photocurrent response at black phosphorus–MoS 2 p–n heterojunctions”. In:
Nanoscale 7.44 (2015), pp. 18537–18541.
[115] Yexin Deng, Zhe Luo, Nathan J Conrad, Han Liu, Yongji Gong, Sina Najmaei, Pulickel M Ajayan,
Jun Lou, Xianfan Xu, and Peide D Ye. “Black phosphorus–monolayer MoS2 van der Waals
heterojunction p–n diode”. In: ACS nano 8.8 (2014), pp. 8292–8299.
[116] Lei Ye, Hao Li, Zefeng Chen, and Jianbin Xu. “Near-infrared photodetector based on MoS2/black
phosphorus heterojunction”. In: Acs Photonics 3.4 (2016), pp. 692–699.
80

[117] Feng Wu, Qing Li, Peng Wang, Hui Xia, Zhen Wang, Yang Wang, Man Luo, Long Chen,
Fansheng Chen, Jinshui Miao, et al. “High efficiency and fast van der Waals hetero-photodiodes
with a unilateral depletion region”. In: Nature communications 10.1 (2019), p. 4663.
[118] Liming Wang, Xuming Zou, Jun Lin, Jiayang Jiang, Yuan Liu, Xingqiang Liu, Xu Zhao,
Yu Fang Liu, Johnny C Ho, and Lei Liao. “Perovskite/black phosphorus/MoS2 photogate reversed
photodiodes with ultrahigh light on/off ratio and fast response”. In: ACS nano 13.4 (2019),
pp. 4804–4813.
[119] S Maimon and GW Wicks. “nBn detector, an infrared detector with reduced dark current and
higher operating temperature”. In: Applied Physics Letters 89.15 (2006).
[120] Hongtao Yuan, Xiaoge Liu, Farzaneh Afshinmanesh, Wei Li, Gang Xu, Jie Sun, Biao Lian,
Alberto G Curto, Guojun Ye, Yasuyuki Hikita, et al. “Polarization-sensitive broadband
photodetector using a black phosphorus vertical p–n junction”. In: Nature nanotechnology 10.8
(2015), pp. 707–713.
[121] James Bullock, Matin Amani, Joy Cho, Yu-Ze Chen, Geun Ho Ahn, Valerio Adinolfi,
Vivek Raj Shrestha, Yang Gao, Kenneth B Crozier, Yu-Lun Chueh, et al. “Polarization-resolved
black phosphorus/molybdenum disulfide mid-wave infrared photodiodes with high detectivity at
room temperature”. In: Nature Photonics 12.10 (2018), pp. 601–607.
[122] Xi Ling, Shengxi Huang, Eddwi H Hasdeo, Liangbo Liang, William M Parkin, Yuki Tatsumi,
Ahmad RT Nugraha, Alexander A Puretzky, Paul Masih Das, Bobby G Sumpter, et al.
“Anisotropic electron-photon and electron-phonon interactions in black phosphorus”. In: Nano
letters 16.4 (2016), pp. 2260–2267.
[123] Yohannes Abate, Deji Akinwande, Sampath Gamage, Han Wang, Michael Snure, Nirakar Poudel,
and Stephen B Cronin. “Recent progress on stability and passivation of black phosphorus”. In:
Advanced Materials 30.29 (2018), p. 1704749.
[124] Itsunari Yamada, Kenji Kintaka, Junji Nishii, Satoshi Akioka, Yutaka Yamagishi, and
Mitsunori Saito. “Mid-infrared wire-grid polarizer with silicides”. In: Optics letters 33.3 (2008),
pp. 258–260.
[125] Kazi M Islam, Ron Synowicki, Timothy Ismael, Isaac Oguntoye, Nathan Grinalds, and
Matthew D Escarra. “In-Plane and Out-of-Plane Optical Properties of Monolayer, Few-Layer, and
Thin-Film MoS2 from 190 to 1700 nm and Their Application in Photonic Device Design”. In:
Advanced Photonics Research 2.5 (2021), p. 2000180.
[126] Romil Audhkhasi, Max R Lien, and Michelle L Povinelli. “Experimental Implementation of
Metasurfaces for Secure Multi-Channel Image Encryption in the Infrared”. In: Advanced Optical
Materials 11.12 (2023), p. 2203155.
[127] Shinpei Ogawa and Masafumi Kimata. “Metal-insulator-metal-based plasmonic metamaterial
absorbers at visible and infrared wavelengths: a review”. In: Materials 11.3 (2018), p. 458.
[128] Cheng-Wen Cheng, Mohammed Nadhim Abbas, Chao-Wei Chiu, Kun-Ting Lai, Min-Hsiung Shih,
and Yia-Chung Chang. “Wide-angle polarization independent infrared broadband absorbers
based on metallic multi-sized disk arrays”. In: Optics express 20.9 (2012), pp. 10376–10381.
81

[129] Zhi Hao Jiang, Seokho Yun, Fatima Toor, Douglas H Werner, and Theresa S Mayer. “Conformal
dual-band near-perfectly absorbing mid-infrared metamaterial coating”. In: ACS nano 5.6 (2011),
pp. 4641–4647.
[130] Serkan Butun and Koray Aydin. “Structurally tunable resonant absorption bands in ultrathin
broadband plasmonic absorbers”. In: Optics express 22.16 (2014), pp. 19457–19468.
[131] Peng Wang, Hui Xia, Qing Li, Fang Wang, Lili Zhang, Tianxin Li, Piotr Martyniuk,
Antoni Rogalski, and Weida Hu. “Sensing infrared photons at room temperature: from bulk
materials to atomic layers”. In: Small 15.46 (2019), p. 1904396.
[132] Joshua B Smith, Daniel Hagaman, and Hai-Feng Ji. “Growth of 2D black phosphorus film from
chemical vapor deposition”. In: Nanotechnology 27.21 (2016), p. 215602.
[133] Jian Mao, Orlando Ortiz, Junjia Wang, Alexandre Malinge, Antonella Badia, and
Stéphane Kéna-Cohen. “Langmuir–Blodgett fabrication of large-area black phosphorus-C 60 thin
films and heterojunction photodetectors”. In: Nanoscale 12.38 (2020), pp. 19814–19823.
[134] Ming Zhao, Haolei Qian, Xinyue Niu, Wei Wang, Liao Guan, Jian Sha, and Yewu Wang. “Growth
mechanism and enhanced yield of black phosphorus microribbons”. In: Crystal Growth & Design
16.2 (2016), pp. 1096–1103.
[135] Yijun Xu, Xinyao Shi, Yushuang Zhang, Hongtao Zhang, Qinglin Zhang, Zengli Huang,
Xiangfan Xu, Jie Guo, Han Zhang, Litao Sun, et al. “Epitaxial nucleation and lateral growth of
high-crystalline black phosphorus films on silicon”. In: Nature communications 11.1 (2020),
p. 1330.
[136] Andrew J Mannix, Andrew Ye, Suk Hyun Sung, Ariana Ray, Fauzia Mujid, Chibeom Park,
Myungjae Lee, Jong-Hoon Kang, Robert Shreiner, Alexander A High, et al. “Robotic
four-dimensional pixel assembly of van der Waals solids”. In: Nature nanotechnology 17.4 (2022),
pp. 361–366.
[137] In-Ho Lee, Daehan Yoo, Phaedon Avouris, Tony Low, and Sang-Hyun Oh. “Graphene acoustic
plasmon resonator for ultrasensitive infrared spectroscopy”. In: Nature nanotechnology 14.4
(2019), pp. 313–319.
[138] Leland Nordin, Priyanka Petluru, Abhilasha Kamboj, Aaron J Muhowski, and Daniel Wasserman.
“Ultra-thin plasmonic detectors”. In: Optica 8.12 (2021), pp. 1545–1551.
[139] Chuan Li, George D Skidmore, and CJ Han. “Uncooled VOx infrared sensor development and
application”. In: Infrared Technology and Applications Xxxvii. Vol. 8012. SPIE. 2011, pp. 541–548.
[140] Chaojian Hou, Lijun Yang, Bo Li, Qihan Zhang, Yuefeng Li, Qiuyang Yue, Yang Wang, Zhan Yang,
and Lixin Dong. “Multilayer black phosphorus near-infrared photodetectors”. In: Sensors 18.6
(2018), p. 1668.
[141] Michael Engel, Mathias Steiner, and Phaedon Avouris. “Black phosphorus photodetector for
multispectral, high-resolution imaging”. In: Nano letters 14.11 (2014), pp. 6414–6417.
82

[142] Michele Buscema, Dirk J Groenendijk, Sofya I Blanter, Gary A Steele, Herre SJ Van Der Zant, and
Andres Castellanos-Gomez. “Fast and broadband photoresponse of few-layer black phosphorus
field-effect transistors”. In: Nano letters 14.6 (2014), pp. 3347–3352.
[143] Skylar Deckoff-Jones, Hongtao Lin, Derek Kita, Hanyu Zheng, Duanhui Li, Wei Zhang, and
Juejun Hu. “Chalcogenide glass waveguide-integrated black phosphorus mid-infrared
photodetectors”. In: Journal of Optics 20.4 (2018), p. 044004.
[144] Wei Ting Chen, Alexander Y Zhu, and Federico Capasso. “Flat optics with dispersion-engineered
metasurfaces”. In: Nature Reviews Materials 5.8 (2020), pp. 604–620.
[145] Shuming Wang, Pin Chieh Wu, Vin-Cent Su, Yi-Chieh Lai, Cheng Hung Chu, Jia-Wern Chen,
Shen-Hung Lu, Ji Chen, Beibei Xu, Chieh-Hsiung Kuan, et al. “Broadband achromatic optical
metasurface devices”. In: Nature communications 8.1 (2017), p. 187.
[146] Jean-Marc Fedeli and Sergio Nicoletti. “Mid-infrared (Mid-IR) silicon-based photonics”. In:
Proceedings of the IEEE 106.12 (2018), pp. 2302–2312.
[147] Mohammadreza Khorasaninejad, Alexander Yutong Zhu, Charles Roques-Carmes,
Wei Ting Chen, Jaewon Oh, Ishan Mishra, Robert C Devlin, and Federico Capasso.
“Polarization-insensitive metalenses at visible wavelengths”. In: Nano letters 16.11 (2016),
pp. 7229–7234.
[148] Yanyan Zhou, Ting Hu, Yu Li, Nanxi Li, Yuan Dong, Dongdong Li, Yuan Hsing Fu, Qize Zhong,
Zhengji Xu, Shiyang Zhu, et al. “A performance study of dielectric metalens with
process-induced defects”. In: IEEE Photonics Journal 12.3 (2020), pp. 1–14.
[149] Y Wu, DL Olynick, A Goodyear, C Peroz, S Dhuey, X Liang, and S Cabrini. “Cryogenic etching of
nano-scale silicon trenches with resist masks”. In: Microelectronic engineering 88.8 (2011),
pp. 2785–2789.
[150] Youngwoo Son, Daichi Kozawa, Albert Tianxiang Liu, Volodymyr B Koman, Qing Hua Wang, and
Michael S Strano. “A study of bilayer phosphorene stability under MoS2-passivation”. In: 2D
Materials 4.2 (2017), p. 025091.
83

Abstract (if available)

Abstract As Moore’s law predicted, conventional silicon-based devices are reaching their physical limit with shrinking devices. Low-dimensional materials bring out a new breakthrough in the semiconductor industry to extend Moore’s law, because of the atomic-level size characteristics, the surfaces free from defects, and the sensitivity to electrical and optical control. The reduced dimensionality provides the opportunity to take advantage of atomically thick layers, leading to unique and extraordinary optical, electronic, mechanical, thermal, and chemical properties. The potential makes it possible to consider their use in a wide range of nanotechnologies, including optoelectronics, biological sciences, energy harvesting, and chemical sensing.

As a promising candidate, low-dimensional asymmetric semiconducting materials are a new class of semiconductors with low symmetry in their crystal structures. These highly asymmetric crystal structures mainly include orthorhombic, monoclinic, and triclinic systems. Due to the crystal asymmetry, low-dimensional asymmetric nanostructures exhibit unique anisotropic physical characteristics in mechanical, optics, and electronics, which include confined or anisotropic transports, excellent flexibility and deformability in mechanics, strong anisotropy in optical absorption, and light emission with controlled directionality. The novel physical characteristics stemming from the asymmetric crystal structure unlock new possibilities for scientists to develop and design innovative semiconductor device applications in highly integrated photonic, electronic, optoelectronic, and especially flexible and stretchable electronic systems. This dissertation focuses on the fundamental properties and potential applications of three different low-dimensional asymmetric semiconducting materials, including Tin iodide phosphide (SnIP), Barium titanium sulfide (BaTiS3), and black phosphorus (BP).

Firstly, We studied the mechanical properties of the newly discovered inorganic double helical semiconductor tin indium phosphate (SnIP). Revealed through both the nanomechanical testing and the Brillouin scattering spectroscopy, this spiral-shaped double helical crystal shows the lowest Young’s modulus (13.6 GPa) among all known stable inorganic materials. The large elastic (>27%) and plastic (>60%) bending strain are also observed due to the easy slippage between neighboring double helices that are coupled through van der Waals interactions, leading to the high flexibility and deformability among known semiconducting materials. The results advance the fundamental understanding of the unique polymer-like mechanical properties in this distinctive class of quasi-one-dimensional crystals and lay the foundation for their potential applications in a broad range of flexible electronics and nanomechanics disciplines.

Secondly, we explored the potential of BaTiS3 in anisotropic infrared photodetection and polarization imaging, we fabricated the infrared photodetector based on exfoliated BaTiS3 flakes and demonstrated that the BaTiS3 photodetector shows a giant anisotropy in broadband optoelectronic responses from near to mid-infrared spectrum. Also, this BaTiS3 photodetector shows the different polarization ratios at different wavelengths. This strong difference in the photodetection anisotropy is clearly observed to demonstrate that the BTS photodetector can identify the slight differences between different polarized illuminations. In addition, the frequency dependence of the BaTiS3 photodetector shows the dynamic photocurrent as a function of frequency in the range from 200 Hz to 10 kHz. This behavior suggests that the dominant mechanism at the illumination of infrared light is the trap-induced photogating effect. A broadband-sensitive photoresponse with excellent stability without degradation under ambient atmospheric conditions will help the polarized infrared imaging for BaTiS3 realized at room temperature. Importantly, a large anisotropic ratio of BaTiS3 ensures polarized imaging in a scattering environment, opening up possibilities for developing next-generation polarized infrared imaging technology. Therefore, we systematically investigated linearly polarized infrared imaging for a designed target using an anisotropic BaTiS3 photodetector.

Thirdly, in order to enable thin-film, high-absorption BP photodetectors, we developed a resonant metal-insulator-metal (MIM) grating-enhanced thin-film BP-MoS2 photodiode. we first integrated a single-resonator MIM structure with a BP–MoS2 photodiode. Our results show that this single-resonator MIM can increase the absorption around a selected mid-infrared wavelength. After the integration of a single-resonator MIM structure, the generated photocurrent increased by 6.7 times at 3.39 µm. Then, we designed and fabricated multiple resonator MIM structures with broadband absorption spectra that can be controlled by tuning their geometry. Finally, we integrated broadband MIM structures with BP-MoS2 photodiode and evaluated them by measuring the room temperature responsivity and specific detectivity of BP-MoS2 photodiodes at multiple MWIR wavelengths. We demonstrated that the broadband MIM structures resulted in enhanced room temperature performance over a broad range of wavelengths. After the integration of a triple-resonator MIM structure, the photocurrent increased by 7.5 times at 3.39 µm and by 12.8 times at 3.88 µm. In a word, our results show that broadband metasurface gratings are a scalable approach for boosting the performance of BP photodiodes over large spectral ranges.

Lastly, to enhance the optical signal collection in BP devices, we directly integrated a designed all-silicon metalens with a BP-MoS2 heterojunction photodiode and experimentally demonstrated its capability to increase photodetector performance in the mid-wave infrared wavelengths. We first designed and fabricated silicon metalens with cylindrical nanorod unit cells. Then, we measured the silicon metalens by imaging clear focal spots at various testing wavelengths. The results showed that the silicon metalens can produce focal spots in the mid-wave infrared wavelengths and increase power density by concentrating light into a small absorber region. Then, we integrated a BP-MoS2 photodetector onto the same substrate and demonstrated that the metalens boosts MWIR performance at room temperature. Furthermore, we integrated a thin-film BP-MoS2 heterojunction photodiode with the silicon metalens. The results showed that the metalens boost room temperature performance at multiple testing mid-wave infrared wavelengths. Photocurrent and responsivity measured at 3.39 µm light focused by the metalens was ≈ 6.65 times higher than in the case with no metalens. Additionally, photocurrent and responsivity were increased by ≈6.31 times at 3.88 µm in the presence of the metalens. These results demonstrate that integrated metasurface lenses are an excellent approach for boosting the performance of MWIR BP photodetectors.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Optoelectronic, thermoelectric, and photocatalytic properties of low dimensional materials

Asset Metadata

Creator Wang, Nan (author)

Core Title Low-dimensional asymmetric crystals: fundamental properties and applications

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Materials Science

Degree Conferral Date 2023-12

Publication Date 11/18/2024

Defense Date 09/05/2023

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

Tag absorption-enhanced metasurfaces and metalens,applications,asymmetric,BP-MoS2 photodetector,fundamental properties,infrared polarization photodetection and imaging,low-dimension semiconductors,mechanical properties,OAI-PMH Harvest,optical anisotropy

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Wang, Han (committee chair), Kalia, Rajiv (committee member), Li, Zhenglu (committee member), Shao, Yu-Tsun (committee member), Yang, Joshua (committee member)

Creator Email nanwang366@gmail.com,wang366@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC113777639

Unique identifier UC113777639

Identifier etd-WangNan-12473.pdf (filename)

Legacy Identifier etd-WangNan-12473

Document Type Dissertation

Format theses (aat)

Rights Wang, Nan

Internet Media Type application/pdf

Type texts

Source 20231127-usctheses-batch-1108 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

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

Repository Name University of Southern California Digital Library

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

Repository Email cisadmin@lib.usc.edu