University of Southern California Dissertations and Theses

Integrated photonics assisted electron emission devices

(USC Thesis Other)

Integrated photonics assisted electron emission devices

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content i

Integrated Photonics Assisted Electron Emission Devices

by

Fatemeh Rezaeifar

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
(ELECTRICAL ENGINEERING)

Dissertation Adviser:
Professor Rehan Kapadia

December 2018

Copyright by Fatemeh Rezaeifar 2018

ii

Dedication

To my dearest husband (Alireza), my beloved parents (Soraya and Fereidoon) and my
wonderful brother (Hossein) for their unconditional love, guidance and support on this
journey.

iii

Acknowledgement

First, I would like to express my deepest gratitude to my academic adviser, Prof.Rehan
Kapadia who supported me and provided me the opportunity to do the research I liked to
do in his research group. I appreciate his advice and guidance during my stay in USC
and under his supervision, I’ve become a better researcher. I look forward to many more
years of professional collaboration with him.
I would like to thank Professor Stephen Cronin, Professor Wei Wu, Professor Jayakanth
Ravichandran and Professor Willie W. Ng for serving on my qualifying exam and defense
committee and for their valuable suggestions to improve my research. Besides my
committee, I am grateful to Professor Martin Gundersen for his generous offer on using
the electron emission setup in his lab during the early stage of this project which helped
me significantly toward developing our own 1
st
generation electron emission setup
equipped with guiding optics in to vacuum chamber. I would also like to appreciate Dr.
Ivan Sanchez Esqueda for the collaboration opportunity with him at USC information
science institute on investigating the reliability of advanced CMOS technology and his
support during my stay at ISI.
I would like to thank all the group members in Professor Kapadia’s lab for their help and
support on my study and all the hard and fun time we had together! Dr. Qingfeng Lin,
Debarghya Sarkar, Jubin Hazra, Matthew Yeung, Louis Blankemeier, Ayush Garg,
Xiangyu Chen, Jun Tao and Hyun Uk Chae, Chenhao Ren, Ragib Ahsan and Alex Marely.
I am also thankful to excellent researchers in USC whom I worked with closely, Dr.
iv

Hooman Abediasl, Dr. Sungwon Chung, Dr. Nirakar Poudel, Tom Orvis, Sriram
Subramanian and Sanjana Kerketta.
Lastly, and most importantly, this journey would not have been possible without the
support of my family. I owe a special thanks to my parents and my brother for their
patience on my long absence. Thanks a lot for desiring the bests for me and supporting
me to pursuit my dreams. Also, I am thankful to my family-in-law who have been always
supportive and understanding. Finally, I express deep love to my wonderful husband, my
best friend and my greatest support, Alireza, who has always been with me through this
journey. Thanks sweetheart for your patience on my absence and your constant support
to reach my dream since the first minute we met in 2008!

v

Abstract
Integrated Photonics Assisted Electron Emission Devices
by Fatemeh Rezaeifar
Doctor of Philosophy in Electrical Engineering
University of Southern California,
Professor Rehan Kapadia, Chair

Photon assisted ultrafast electron emission is essential for applications such as
developing free electron lasers (FEL), time resolved electron microscopy and high-power
RF amplifiers using vacuum electron devices (VED). With the advances of Q-switched
and Ti-Sapphire ultrafast laser sources, researchers started working on the
demonstration of sharp electron beam via exposing high-power laser to a sharp metallic
tip and they were successful in achieving electron beam pulse down to 100 fs. However,
one of the most important challenge that hasn’t been addressed is the development of a
technique which enables higher quantum efficiency (QE) of the photon assisted electron
emission devices.
As such this dissertation has two focuses:
(1) Implementing techniques that enable lower threshold dark field emitter. Specifically, I
focused on using a real 2D material such as graphene with excellent electrical and optical
properties and lanthanum hexaboride (LaB6) nanoparticles as a low work function
electron emitter. This hybrid electron emitter in combination with an array of silicon sharp
tips that enhance the local field intensity was used to demonstrate lower threshold field.
The other technique for reducing the threshold field of the dark field emitter is to add an
ultra-thin layer of oxide in between metal-semiconductor (MS) junction to inject a sharp
distribution of hot electron from doped substrate in to the thin top metallic emitter. The
contribution of these hot electrons to field emission is equivalent to electron emission from
vi

an artificial material with lower effective work function. As such, the threshold value is
expected to be smaller compared to the identical MS emitter. The experimental
demonstration on two different stacks of MIS emitter is described in chapter 3.
(2) Demonstrating a novel idea of introducing integrated photonics waveguide and
cavities for electron emission devices to obtain enhanced QE through achieving higher
optical absorption. Less than 1 pA of light electron emission current is being detected
from conventional free space coupled emitters and increasing photon assisted emission
current is necessary for many applications. The theoretical calculation on the emission
device consist of a microfabricated optical cavity, a Fabry-Perot and a heterostructured
thermionic emitter with a small bandgap or metallic thermionic emitter (e.g. LaB6)
deposited on a wider bandgap electrical and thermal conductor (e.g. doped Si) are
discussed in detail in chapter 4 . The significant result is through cavity assisted emitter
the highly-efficient photon to thermal conversion efficiencies > 60% can be achieved
despite small emitter active absorption volumes < 0.01 µm
3
and moderate Q optical
cavities. Through transient simulation, it is found that cavity assisted electron emitter can
be designed with ultra-fast sub-ns thermal response time, and sub 10 ps current response
times with < 10 µW of power required to achieve nA level current emission per tip. After
theoretical evaluations, we experimentally demonstrated the integrated photonics
waveguide assisted electron emitters. First, I explored the fabrication process and then
investigated the effect of the replacing free space coupling with integrated waveguide
assisted evanescent coupling on electron emission. Enhanced electron emission has
been observed using different emitter material on waveguide including (i) graphene, (ii)
lanthanum hexaboride, (iii) hybrid of lanthanum hexaboride on graphene. I measured up
vii

to 10 pA/mW of emission efficiency via evanescent coupling whereas the free space
coupling results in a few fA/mW of emission efficiency. I also observed the emission
process is sensitive to a 2
nd
order power law of the laser intensity at low E-field, which
supports an emission mechanism based on multi-photons absorption followed by over-
the-barrier direct thermionic emission. However, at higher E-field, the emission process
is linearly proportional to laser power, indicating a single photon process followed by
quantum mechanical direct tunneling of hot electron contribution to emission process. We
conclude this technique in combination with on-chip laser source as a solution for future
generation power efficient photon-assisted electron emission device.

viii

Table of Contents

Dedication ii

Acknowledgement iii

Abstract v

1. Introduction 1

1.1 Research Motivation 1
1.2 Problem Definition & Proposed Solution 3
1.3 Related Work 7
1.4 Applications 11
1.5 Dissertation Outline 12
Chapter 1 References 13

2. Background 15

2.1 Electron Emission General Terms 15
2.1.1 Potential Barrier at a Metallic Emitter / Vacuum Interface 15
2.1.2 Space Charge Effect 17
2.1.3 Electron Collision and Scattering 17
2.1.4 Electric and Optical Field Enhancement 18
2.2 Mechanisms of Electron Emission 19
2.2.1 Field Emission 19
2.2.2 Thermionic Emission 20
2.2.3 Photo Emission 22
2.2.4 Intense Photo-Field Emission 23
2.3 Materials for Electron Emission 24
2.3.1 Semiconductor based Emitter 24
2.3.2 Metallic based Emitter 25
2.3.3 Carbon based Emitter (Graphene) 26
Chapter 2 References 28
ix

3. Light & Dark Electron Emission Characterization 29

3.1 Building an Experimental Setup 29
3.1.1 Ultra-high Vacuum Chamber Design and Installation 29
3.1.2 High Voltage Power Supply 29
3.1.3 Electron Emission Detection Equipment 30
3.1.4 CW Tunable Fiber Coupled Laser Source 32
3.1.5 Automated Test Setup 33

3.2 Hybrid Emitter: Fabrication & Characterization 34
3.2.1 Pyramid Silicon Array Fabrication 34
3.2.2 Wet Transfer of Graphene 38
3.2.3 Low Work Function Lanthanum Hexaboride Drop-Casting 40
3.2.4 Work Function Characterization 43
3.2.5 Electron Emission Characterization 44
3.2.6 Summary 50

3.3 Tunneling Assisted Electron Emitter 51
3.3.1 Device Schematic and Fabrication 51
3.3.2 Electron Emission Characterization 54
3.3.3 Summary 57

3.4 Free Space Coupled Photon Assisted Electron Emitter 58
3.4.1 Developing a Setup – LED Source & Graphene Emitter 58
3.4.2 Developing a Setup – Laser Source Characterization 61
3.4.3 LaB6 Free Space Emission Characterization 64
3.4.4 Graphene Free Space Emission Characterization 66
Chapter 3 References 67

4. Modeling of the Integrated Photonic Assisted Electron
Emitter 68

4.1 Theoretical Calculation of Cavity Assisted Thermionic Emission
Device 68
4.1.1 Cavity Assisted Thermionic Device 69
4.1.2 Optical and Thermal Calculations 70
4.1.3 Optical Response Modeling 73
4.1.4 Thermal Response Modeling 77
4.1.5 Summary 82
x

4.2 Theoretical Calculation of Integrated Ultra-Thin Photoemission
Device 83
4.2.1 Device schematic and Optical Calculation 84
4.2.2 Quantum Efficiency Calculation 86
Chapter 4 References 90

5. Demonstration of the Integrated Photonic Waveguide
Assisted Electron Emitter 91

5.1 Waveguide Assisted Electron Emission Device 92
5.1.1 Device Schematic and Electron Excitation 92
5.1.2 Optical V-groove and Waveguide Fabrication 95
5.1.3 Wet Transfer of Graphene and Annealing Process 101
5.1.4 Optical Laser Coupling from Fiber to Waveguide 102

5.2 Characterization of Waveguide Assisted Graphene Emitter 105
5.2.1 Optical Absorption of Graphene on Optical Waveguide 105
5.2.2 Electron Emission Characterization 109

5.3 Characterization of Waveguide Assisted LaB6 Emitter 115
5.4 Characterization of Waveguide Assisted LaB6 on Graphene Emitter
121
5.5 Laser Modulation Effect on Emission 125
5.6 Characterization of Emitter on Oxide/Nitride Substrate 127
Chapter 5 References 128

6. Conclusion and Future Work 129

6.1 Conclusion 129
6.2 Future Work 130
Chapter 6 References 134
1

Chapter 1 – Introduction

1.1 Research Motivation

Electron beam can be generated by optical illumination on metallic, semiconductor
or graphene emitter in which absorbed photon will excite an electron and the excited
electron upon having enough energy will transport to emitter surface and scape to
vacuum. The basis of this phenomena, converting photon to excited electron along with
scattering and electron transport mechanisms have been discovered many years ago and
researcher studied various aspects of this phenomena theoretically and experimentally
since 1960
1-4
utilizing the free space laser illumination on emitter. Figure 1.1 shows the
typical photoemission experimental set up of the free space coupling (direct illumination)
of laser on an electron emitter that requires stringent alignment of laser beam using mirror/
lens on the emitter tip inside vacuum chamber
5
.
With the advances of Q-switched and Ti-Sapphire ultrafast laser sources,
researches on ultrafast electron beam started in the 1980s. The first photoelectron-based
diffraction experiments probing light-driven structural phase transitions with 20 ps time
resolution
6-8
. For this time and the next 30 years, most researches devoted to
demonstrating sharp electron beam via exploring the electron emission from sharp tip
metallic emitter exposed to high power laser radiation
9-15
. In the 1990s, research on
ultrafast electron microscopy accomplished time resolution below 1 ps by incorporating a
single-electron photocathode, driven by high repetition rate lasers, into a high-spatial-
resolution transmission electron microscope
16
. This time resolution has been improved
to about 300 fs in a miniaturized diffraction setup, allowing visualization of ultrafast
2

chemical reaction.

In principle, it would be highly desirable to further enhance this
temporal resolution because of observing a variety of fundamental processes and this
currently triggers most of the experimental activity in ultrafast electron imaging. Photon
assisted ultrafast electron source is also essential for other applications such as free
electron laser source
17
, vacuum electronic high-power THz generation
18
. Up to date,
intensity dependence of the photon assisted electron emission, energy distribution and
emittance of the emitted electrons and effect of the incident light polarization and angle
with respect to emitter were at the top interests of the published papers in this field.
Despite all developments after the initial results on photon assisted electron emission,
this phenomenon suffers from low quantum efficiency as the biggest challenge. The
recent value of the reported QE for Cu used as photocathode in RF photo-gun is around
6e-5
19-20
and it hardly exceeds slightly over 1e-4 only under deep UV (at 193 - 266 nm)
illumination
21
. The theoretical value of quantum efficiency for the current technique has
been studied and researcher expecting to get close to 7e-3 for cupper through higher
optical absorption from structured cupper emitter (sub-wavelength grating structures with
challenging fabrication requirements)
22
. This research intends to address this issue and
propose a novel solution for enhancing QE.

Figure 1.1 Typical photoemission experimental set up with direct illumination of the focused
beam on electron emitter

Ultrafast Laser source
V acuum chamber
Emitter tip
Parabolic mirror / lens
x y z-aligner to
adjust the tip
3

1.2 Problem Definition & Proposed Solution

As discussed in previous section, the major challenge of the photon assisted
ultrafast electron emission devices is their small QE. The small QE in metallic emitter can
be explained using Spicer three steps model
23
. This model is developed in 1960 and
widely being used for describing the photon assisted electron emission process.
According to this model, photon assisted electron emission consists of three steps: (1)
absorption of incident photons and excitation of the electrons, (2) the motion of the
photoexcited electrons to surface and (3) the escape of the electrons over the surface
barrier in to vacuum. The first step in Spicer’s electron emission theory is the electron
excitation via optical absorption. However, most of the metallic cathode demonstrates
poor optical absorption from illuminated photons and reflects the incoming photons. In the
case of graphene, the optical absorption of the monolayer is limited to 2.3% of the
incoming photons. Optical absorption of the semiconductor is higher than metallic and
graphene-based emitter however semiconductor emitter has unstable electron emission
and large energy spread which is not favorable as electron emitter. Therefore, efficient
optical absorption is a key parameter for enhancing the quantum efficiency.
In addition to poor optical absorption, the optical absorption length inside bulk
material (defined by the distance in to a material where the absorption dropped to 1/e of
the incident and is given by 𝑙 𝑜𝑝𝑡 =
𝜆 4𝜋𝑘
) is larger than electron mean free path. For
example, the absorption length at IR wavelength of 800 nm is as large as 40 nm. The
optical absorption length reduces at deep UV laser which is difficult to generate. On the
other hand, the mean free path of electron also depends on material and the thickness of
4

emitter
24
. For bulky metallic emitter a good approximation is around 5 nm. As a result,
great portion of the photo-excited electrons goes through collisions and electron-electron
scattering, losing their energy and cannot get to emitter surface. This is in addition to a
phonon scattering. Hence, only small portion of the photo-excited electrons that have
shorter path to emitter surface experience fewer scattering and they have higher chance
to get to emitter’s surface and emit to vacuum. As such developing an electron emitter
with fewer electron scattering is a second key for enhancing the quantum efficiency.
So far, most of the efforts have been focused to design electron emitters with
proper geometry to localize the incoming field and enhancing the optical absorption from
free space illuminated laser for example sub-wavelength grating. However, the design
criteria for optimizing the absorption via sub-wavelength grating will be challenging for
actual fabrication. It has been demonstrated for increasing the optical absorption of the
free space illuminated laser on Aluminum to 80%, grating with opening of 5 nm or less is
needed
22
. Fabrication of this device will be very challenging. Other efforts have been
focused to investigate promising material as electron emitter for higher optical absorption,
this include carbon-based films and nanofibers
25-26
. However, increasing QE is still a
challenging problem.
In recent years, remarkable advances have been achieved on implementing on
chip III-V laser source as well as integrated photonics devices at the order of tens of µm
size
27-28
. For example, in 2012 Huan Li
29
demonstrated absorption coefficient of 0.2
dB/µm at C-band can be achieved by monolayer of graphene via transferring it directly
on integrated optical waveguide (This is equivalent to 100% absorption within a length
shorter than 1 mm) whereas the graphene free space optical absorption is only 2.3% of
5

the incoming laser power. In 2013, Steven J. Koester
30
described the overview of
waveguide-coupled graphene optoelectronics including its application in optical
modulator, photodetector, etc. In addition, optical Fabry-Perot cavities are a useful
platform for generating enhanced absorption due to storing light and higher interaction
opportunity
31-32-33
. This technique has been used for medical sensor and other areas that
requires enhanced light – matter interaction
34
. Furthermore, recent theoretical work
anticipated that in graphene multiple electron–hole pairs can be created from a single
absorbed photon during energy relaxation of the primary photoexcited electron–hole pair
35,36
and the effect has been investigated experimentally
37
. In the present work, we take
advantages of these development and demonstrate an enhanced photon assisted
electron emission from a graphene layer as an electron emitter directly placed on Silicon
Nitride (Si3N4) optical waveguide. The schematic of the demonstrated device is depicted
in Figure 1.2b. The details of this device will be discussed in chapter 5, however the core
idea of enhancing optical absorption is due to the placement of emitter on waveguide that
causes evanescent optical coupling over longer interaction length (emitter length).
Enhanced optical absorption is equivalent to improving 1
st
step of Spicer’s photoemission
mechanism. The photoexcitation of electrons generates electron distribution at elevated
energy level compared to thermal electron distribution. Then, the photoexcited hot
electrons start their travel to emitter surface. The three steps Spicer’s model that
described previously applies to photon assisted field emission from bulky emitter whereas
in 2D materials, due to their infinitesimal thickness, the photoexcited hot electrons are
capable of escape from all directions and momentum cone angular limitation doesn’t
apply to electron emission from 2D material. The probability of emission from 2D material
6

is given by the integral of the overlap between states in the graphene and vacuum and
the time component is given by the length of the time electrons spend in each state. A
monolayer of graphene provides greater chance for photoexcited hot electron to get to
emitter surface without losing their energy through scattering compared to 3D bulky
emitter (Improving Spicer’s photoemission 2
nd
step). The ability to absorb majority of the
incoming photons and excite electrons within the size at the order of tens of µm increases
the photon assisted emission current density significantly. Last but not the least
advantages of using integrated photonics for electron emission devices is the large-scale
fabrication and reliability of the integrated process. In recent years, giant semiconductor
foundries invested in developing photonics friendly process which is highly beneficial for
reproducing the large-scale photonics assisted electron emission devices with
reasonable cost for customers.

Figure 1.2 (a) Schematics of the free space coupled graphene electron emitter. (b) Schematics
of the integrated Si3N4 waveguide assisted graphene electron emitter.

b
Graphene Emitter
a
Graphene Emitter
Laser
7

1.3 Related Work
Here, I will review the state of the art demonstrations of the photon assisted electron
emission devices and give a summary of their performance. These include fs laser pulse
on Tungsten sharp tip emitter
15, 39
, Carbon based emission devices
25
and Theoretical
design on sub-wavelength grating for electron emission devices
22
. Figure 1.3 illustrate
the summary of major research in this field.
Figure 1.3 (a-c) shows the femtosecond pulse laser assisted electron emission from a
tungsten tip tested with mode locked Ti: sapphire laser that operates at 1 GHz repetition
rate and produces a train of 48 fs pulses at a center wavelength of 810 nm with maximum
average power of 600 mW. Up to 100 fA is measured with average power of 150 mW.
Furthermore, the polarization dependence of the photocurrent indicates photocurrent is
minimum when polarization angle is 90 degrees, in other word if the laser illumination
occurs perpendicular to emitter tip.
Figure 1.3 (d-e) shows a typical set up for photocurrent measurement with the
focus on final electron pulse duration. They show that a field emission tip electron source
that is triggered with a femtosecond laser pulse can generate electron pulses shorter than
the laser pulse duration (~100 fs). In addition, Figure 1.3 (e) shows photocurrent versus
laser power that indicates multi-photon contribution using 810 nm photon (Photon Energy
= 1.53 eV – Tungsten work function = 4.5 eV).
Figure 1.3 (f) shows the effort of trying photoemission performance from electron
rich material such as carbon nano fiber. The field emission threshold observed at 2V/µm.
The photon assisted emission with femtosecond Ti: sapphire laser provided 50 fs pulses
8

at 800 nm with repetition rate of 50 Hz and energy up to 0.8 mJ. Beam diameter set to 4
mm by an aperture. The anode made of glass plate covered with transparent conductive
indium tin oxide (ITO) layer. The irradiated area on cathode was 0.12 cm
2
. Under this
condition, they detected emitted charge of nC level as plotted in Figure 1.3 (f).
Figure 1.3 (g) shows the effort for designing subwavelength grating. This structure
proven to increase the optical absorption. This work
22
demonstrated theoretical design
of the subwavelength grating that enhances absorption up to 80%. Using this design, the
QE can be enhanced as high as 25 times. However, this design requires grating opening
of 5 nm which makes the fabrication of this device challenging.

9

Figure 1.3 Related Work – (a)(b)(c) [15] – (d)(e) [39] – (f) [25] – (g) [22]

Nanocarbon Film
Sub-wavelength grating Emitter
10

Figure 1.4 is a summary plot from several papers
40-44
comparing their reported emission
current versus laser intensity. These works utilizing the conventional free space coupling
of laser on emitter. It can be seen, majority of them using extremely large power intensity.
Among these works, one
40
reports 10-20 pA of current in which they focused on using
high energy photons at 257 nm (4.8eV) and 343 nm (3.6 eV). They were able to reduce
the work function of tungsten from 4.5 eV to 1.6eV using low work function coating as well
as high DC field. At the end of this dissertation, I will add the results we obtained in this
project.

Figure 1.4 Summary of the state of the art reported laser induced emission current as a function
of laser power intensity.
40-43

0.001
0.01
0.1
1
10
100
Current (pA)
10
0

10
3

10
6

10
9

10
12
Laser Intensity (W/cm
2
)
40
41
42
43
11

1.4 Applications
Electron emission is one of the fundamental problem in physics that has many different
applications including, but not limited to electron microscopy
45,46
, electron beam
lithography
47
, space propulsion
48
, high power microwave devices
49,50,51
, free electron
lasers,
52
field effect displays
53
, neutron generation
54-55
,

and ultrafast electron diffraction
56
. Electrically gated field emission devices have been heavily explored in the past,
however the large capacitances due to the proximity of a control gate often limits the
maximum modulation frequency of these devices
57,58
. Optical modulation of electron
emission offers the highest modulation speeds as well as control over a variety of
emission mechanisms such as single and multiphoton photoemission,
59,60
strong field
emission,
61,62
field-emission,
57,58
and thermionic emission
63.
It also opens wider range
of freedom for designing equipment that requires electron source. For example, the
development of ultrafast pulsed electron sources has enabled time-resolved electron
microscopy, which is useful for characterization of atomic-scale motions. Figure 1.5
depicted major applications for which developing high QE ultrafast electron beam is
necessary.

Figure 1.5 Applications that requires high QE ultrafast electron beam. (a) free electron laser
source
64
, (b) Time resolved electron microscopy
65
, (c) DARPA high power amplifier using
vacuum electron devices (VED)
66

Electron Source
X-ray
(a) (c) (b)
12

1.5 Dissertation Outline
The dissertation is organized as follows:
In chapter 2, a background on different mechanisms of the electron emission as
well as typical material used as an electron emitter are reviewed.
In chapter 3, I will describe the development of electron emission set up,
specifically an ultra-high vacuum chamber fully assembled in our lab with the electron
emission generation and detection equipment and optical set up for photo-electron
emission and detection. Then, I will describe the fabrication and characterization process
of three different electron emitter, (1) Hybrid electron emitter consist of array of sharp tip
silicon emitter with graphene and lanthanum hexaboride nano-particles emitter on top. (2)
Tunneling assisted electron emitter consists of thin metal on ultrathin insulator on
semiconductor (MIS) that generates sharp distribution of the hot electrons on metal-side
and finally (3) Photon assisted electron emission via conventional free space coupling
(direct illumination) on graphene and lanthanum hexaboride emitter.
In chapter 4, I explain the theoretical calculation of integrated photonic cavity
coupled electron emission devices. For this part, I focus on optical Fabry-Perot cavity as
a mean to store the light and enhance the light-emitter interaction. The optical, thermal
and electrical response of these emission devices are evaluated. In addition to cavity
assisted electron emitter, I will describe the calculation of QE from ultrathin emitter
evaporated directly above an optical waveguide. The electron scattering plays dominant
role on electron emission and hence the QE. I will describe the effect of electron
scattering considering emitter thickness through numerical calculations.
13

In chapter 5, I will describe the demonstration steps of waveguide assisted electron
emitter including fabrication steps, optical coupling procedure and emitter dark & light
characterization. These tests include investigating the effect of photon energy, laser
power, E-field, emitter material and transient response.
I ll conclude this dissertation in chapter 6 with conclusion on this work and future
directions.
Chapter 1 References
1. Berglund, C. N., and W. E. Spicer. Physical Review 136.4A (1964): A1030.
2. Berglund, C. N., and W. E. Spicer. Physical Review 136.4A (1964): A1044.
3. Kane, Evan O. Physical review 127.1 (1962).
4. Krolikowski, W. F., and W. E. Spicer. Physical Review 185.3 (1969): 882.
5. C. Ropers PRL 98, 043907 (2007).
6. Mourou, G.; Williamson, S. Appl. Phys. Lett. 1982, 41, 44−45.
7. Williamson, S.; Mourou, G.; Li, J. C. M. Phys. Rev. Lett. 1984, 52, 2364−2367.
8. Aeschlimann, M.; Rev. Sci. Instrum. 1995, 66, 1000−1009.
9. J. F. Ready, Phys. Rev. 137, A620 (1965); J. Appl. Phys. 36, 462 (1965).
10. W. L. Knecht, Appl. Phys. Letters 6, 99 (1965).
11. H. Sonnenberg, H. Heffner, and W. Spicer, Appl. Phys. Letters 5, 95 (1964).
12. M. C. Teich et al, Phys. Rev. Letters 13, 611 (1964).
13. E. M. Logothetis and P. L. Hartman, Phys. Rev. Letters Vol 187, 2 (1969).
14. J. H. Bechtel, t W. Lee Smith, ~ and N. Bloembergen, Phys. Rev B VOL 15, 10 (1977).
15. Peter Hommelhoff et al, PRL 96, 077401 (2006).
16. Williamson, J. C. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 5021−5025.
17. Brau, C. A., Free-electron lasers. JSTOR: 1990; Vol. 22.
18. Booske, J. H., et al. IEEE Transactions on Terahertz Science and Technology 1.1 (2011).
19. S. W. Downey et al, Appl. Phys. Lett., vol. 49, pp. 912, (1986).
20. R. Yen et al, Phys. Rev. Lett., vol. 19, pp. 1837, (1984).
21. F. Le Pimpeca et al, Appl. Phys. A 112, 647 (2013).
22. Alexander, Anna et al. JVST A: Vacuum, Surfaces, and Films 34.2 (2016): 021401.
23. W.E. Spicer, Phys. Rev. 112, 114 (1958).
24. K. R. Peschmann et al, J. Applied Phys 44, 2252 (1973).
25. D. A. Lyashenko et al. Journal of the European Optical Society-Rapid publications (2017).
26. MR Bionta et al- Ultramicroscopy, (2015) Elsevier.
27. Il-Sug Chunga et al, APPLIED PHYSICS LETTERS 97, 151113 (2010).
28. Zhiping Zhou, Light: Science & Applications volume 4, page e358 (2015).
29. Huan Li et al, Appl. Phys. Lett. 101, 111110 (2012).
30. S. J. Koester and M. Li, IEEE J. Sel. Top. Quantum Electron. 20(1), 6000211 (2014).
31. F. Rezaeifar and R. Kapadia, J. Vac. Sci. Technol., B 34, 041228 (2016).
32. B Behaghel, Appl. Phys. Lett. 106, 081107 (2015).
33. Arthur Nitkowski, Optics Express, Vol. 16, Issue 16, pp. 11930-11936 (2008).
34. A. Ksendzov and Y. Lin, Optics Letters, Vol 30, Issue 24, pp. 3344-3346 (2005).
14

35. Winzer. T, Nano Lett, 10 (12), pp 4839-4843 (2010).
36. Winzer. T, Phys. Rev. B 85, 24140 (2012).
37. K.J. Tielrooji Nature Physics volume 9, pages 248–252 (2013)
38. J Vogelsang, Nano Lett., 2015, 15 (7), pp 4685–4691
39. B Barwick, New Journal of Physics, Volume 9, (2007).
40. Ding-Shyue Yang et al, PNAS August 24, 2010. 107 (34) 14993-14998;
41. R. Bormann et al, PRL 105, 147601 (2010)
42. Andrej Grubisic et al, Nano Lett, 2012, 12 (9), pp 4823-4829
43. Barwick et al. NEW JOURNAL OF PHYSICS 9 (2007)
44. Anna Mustonen et al, APL 99, 103504 (2011)
45. J. Goldstein, et al. Springer Science (2012).
46. L. Reimer, Springer Science (2013).
47. C. Vieu, et al, Applied Surface Science 164 (1), 111-117 (2000).
48. D. M. Goebel and I. Katz, Fundamentals of Electric Propulsion: Ion and Hall Thrusters.
Hoboken, NJ: Wiley (2008).
49. J. Benford et al. High-Power Microwaves, CRC Press (2007).
50.S. H. Gold and G. S. Nusinovich, American Institute of Physics 68 (11), 3945-3974 (1997).
51.R. J. Barker, N. C. Luhmann, J. H. Booske and G. S. Nusinovich, Modern Microwave and
Millimeter-Wave Power Electronics, by Robert J. Barker (Editor), Neville C. Luhmann (Editor),
John H. Booske (Editor), Gregory S. Nusinovich, pp. 872. Modern Microwave and Millimeter-
Wave Power Electronics John Wiley & Sons. 1 (2005).
52.C. A. Brau, Free Electron Lasers, Science (1990).
53.Q. Wang et al, APL 72 (22), 2912 (1998).
54. A. Persaud et al, JVST B. 29 (2), (2011).
55. A. Persaud et al, Review of Scientific Instrument 83 (2), (2012).
56. A. H. Zewail et al, Annual Review of Physical Chemistry 57, 65-103 (2006).
57. K. Jensen et al, APL 77 (4), 585-587 (2000).
58. K. L. Jensen et al, Solid-state Electronics 45 (6), 831-840 (2001).
59. G. Herink et al, Nature 483 (7388), 190-193 (2012).
60. L. Wimmer et al, Nature Physics 10 (6), 432-436 (2014).
61. R. Bormann et al, PRL 105 (14), 147601 (2010).
62. P. t. Dombi et al, Nano Letters 13 (2), 674-678 (2013).
63. X. Wang et al, Physical Review B 50 (11), 8016 (1994).
64. C Pellegrini, Phys.Scr (2016).
65. Geoffrey H. Campbell, Applied Physics Reviews 1, 041101 (2014);
66. DARPA website- Mr. Dave Tremper.

15

Chapter 2 - Background

2.1 Electron Emission General Terms

2.1.1 Potential Barrier at a Metallic Emitter / Vacuum Interface

Fermi-Dirac distribution describes the energy states of the electrons in a metallic
system and given by:
𝑓 (𝐸 ,𝑇 )=
1
1+𝑒𝑥𝑝 (
𝐸 −𝐸 𝑓 𝑘𝑇
)
(1)
Here 𝑇 is the electron temperature, 𝑘 is the Boltzman constant, and 𝐸 𝑓 is the Fermi
energy. The Fermi energy is the maximum level that one electron can have inside metal
at 𝑇 = 0 K. Fermi-Dirac relation shows electron distribution with energy 𝐸 at temperature
T. The potential barrier seen by electron at metal-vacuum interface at position x=0, with
metal filling the region of space for x < 0 (electron fills up to Fermi level) and the vacuum
for x > 0 is assumed to be U(x):
𝑈 (𝑥 )= 𝐸 𝑓 +𝑊 (2)
Where 𝑊 is the work function of the metallic emitter. The work function is the minimum
amount of energy needed to remove an electron from metallic system (the energy
difference between Fermi and vacuum level). If the electronic states are full up to the
Fermi energy, then an electron requires an additional 𝑊 amount of energy to be removed
from the metal. The work function depends on the material. When we apply an electric
field, the potential barrier becomes:
𝑈 (𝑥 )= 𝐸 𝑓 +𝑊 −𝑒𝐸𝑥 (3)
Where e is the electric charge and E is the applied electric field. The potential seen by
electron modifies with applied field such that electron can quantum mechanically tunnel
16

through it. This potential before and after an applied electric field is shown in Figure 2.1(a).
This potential barrier continuously modifies as electron leaves the metal. The electrons
left the surface of the metal shield the interior of metal from the field of the free electron.
Adding the classic image potential for an electron a distance x from a conducting plane:
𝑈 (𝑥 )= 𝐸 𝑓 +𝑊 −𝑒𝐸𝑥 −
𝑒 2
16𝜋 𝜀 0
𝑥 (4)
This relation indicates the barrier height becomes smaller than the work function of the
material. This reduction of the barrier is called Schottky effect and depicted in Figure
2.1(b).

Figure 2.1 Potential barrier without DC field (dashed line) & with DC field (solid line) (a)
Without space charge effect (b) With space charge effect. φeff is the effective work function.

17

2.1.2 Space Charge Effect

Space charge effects occur when more than one electron is emitted at the same
time and are then repelled from each other by Coulomb repulsion. The number of
electrons emitted and detected depends on many factors including the applied voltage
and the laser intensity as well as the geometrical constraints of the spectrometer and
detector. Space charge broaden the electron pulse duration to longer than the laser pulse.
For the experiment in this dissertation, we used relatively low power CW laser source
compared to other laser assisted emission work and we ignore the space charge effects.

2.1.3 Electron Collision and Scattering

The scattering process occurs inside emitter material before electron emission as
well as after electron emission to vacuum. The first case consists of phonon and impurity
scattering. It is possible to obtain a formula that gives the probability an electron scatter
with a given scatter center (an impurity or a phonon)
1
. The probability is calculated by
means of the Fermi Golden Rule, that is a method to express probability as the result of
a first order perturbation problem. In general, the probability that an electron scatters is
given by the following formula
2
:
𝑊 (𝑘̅
)=
𝛺 (2𝜋 )
3
∫ ∫ ∫ 𝑆 (𝑘̅
,𝑘 ′ ̅
)
∞
0
𝜋 0
𝑑 𝑘 ′
𝑑𝜗𝑑𝜑 2𝜋 0
(5)
where 𝑆 (𝑘̅
,𝑘 ′ ̅
) is the transition rate, the transition probability per unit time that an electron
scatter from the initial state 𝑘̅
to the state 𝑘 ′ ̅
, 𝜑 is the azimuthal angle, θ is the polar angle
of the pseudo-wave vector and Ω is the volume of the crystal.
After electron emission to vacuum, they transport toward anode due to electrostatic
field. However, in this process more than one electron is generated and these electrons
18

experience electrostatic coulomb force from each other and scatter toward different
directions. Parts of them will redirects backward toward the emitter (cathode) and collide
with new-coming electrons. These scattering and collisions are the limiting factor for the
performance of emission device.

2.1.4 Electric and Optical Field Enhancement

(a) Geometrical E-Field Enhancement

To have field emission by applying reasonable voltages, sharp tip cathode is
generally used, which have a half-sphere termination on a nanometer size. For tips, the
field at the end of the apex is given by
𝐹 = 𝛽 𝑉 𝑡𝑖𝑝 (6)
Where 𝑉 𝑡𝑖𝑝 is the applied voltage, and β is related to the tip radius r with 𝛽 ∝
1
𝑟 . Here, β
can be thought of as the field enhancement factor and can reach factors varying from tip
to tip. Thus, the high fields needed to achieve emission can be enhanced geometrically
by the sharpness of the tips. In nanotip systems, since the natural field enhancement is
geometrical, the sharper the tip, the lower the voltage needed to reach observable
emission. Typically, the minimum voltage needed to observe emission is called the
emission threshold. However, in this work we define threshold field as the E-field needed
for 10 µA/cm
2
and turn-on field as the field needed for having 0.1 µA/cm
2
.
(b) Optical Field Enhancement

Optical fields can be enhanced using emitter tip size smaller than the wavelength
of the light. Strong optical field will be localized near sharp tip. For this purpose, tip
diameter should be as small as 100 nm
3
.
19

2.2 Mechanisms of Electron Emission
2.2.1 Field Emission

Electron emission from a metallic surface into vacuum due to a high externally
applied electric field is called field emission
4,5,6
. This phenomenon occurs when a high
enough DC electric field is applied between electrodes such that electrons can quantum
mechanically tunnel through the metal/vacuum interface as shown in Figure 2.2. Electron
field emission was first observed and described by R.W. Wood in 1897
7
. Later W.
Schottky described the mechanisms behind field emission, and a connection between the
applied field and a reduction in the height of the potential barrier
8
. After that, Fowler and
Nordheim described the emission current density
9
. Later, many researchers started using
field emission to study the surface properties of materials experimentally. One of the
outstanding results were used as a method for making the nanotip emitters to obtain an
extremely high field which have an application in systems such as field emission
microscopy (FEM) and scanning electron microscopy (SEM). The relevant calculations
for field emission are made with an assumption that the temperature is 0 K and therefore
this phenomenon sometimes called cold field emission. In cold field emission, a large field
typically beyond 100 V/µm is required to detect the emission current. In practice lower
electric field can be used for electron emission via introducing geometrical field
enhancement to emitter. The Fowler-Nordheim theory is widely being used for correlating
the current density and local electric field to geometrical and material properties of the
emitter and expressed by following equation
9

𝐽 =
𝐴 𝛽 2
𝜑 𝐸 2
.exp(−
𝐵 𝜑 3
2
𝛽𝐸
) (7)
20

Here, J is the current density of the emission, A and B are constant equal to 1.56e-10 A
eV V
-2
and 6.83e3 eV
-3/2
V µm
-1
, respectively. E is the electric field intensity. The electric
field intensity is proportional to the applied tip bias, 𝑉 𝑡𝑖𝑝 as well as tip radius that enhances
the local fields. This parameter is called field enhancement factor, 𝛽 and will be obtained
experimentally once we plot ln (
𝐽 𝑉 𝑡𝑖𝑝 2
) as a function of
1
𝑉 𝑡𝑖𝑝 as I will describe further in chapter
3.
2.2.2 Thermionic Emission

By heating up the system, the Fermi-Dirac distribution changes with temperature
until a portion of the electrons have sufficient energy to escape the work function. This is
called thermionic emission or over the barrier direct emission. This process is governed
by the Richardson-Dushman equation
10,11
.
𝐽 =𝐴 𝜂 𝑇 2
exp (
−𝜑 𝑘𝑇
) (8)
where J is the emitted current density, φ is the material work function, η is a material
dependent pre-factor and A is the Richardson constant defined as

Figure2.2 Potential barrier seen by electron – During field emission process. Large E-field
is applied for bending potential barrier and schottky effect reduced the potential barrier
such that electron can tunnel through it.

21

𝐴 =
4𝜋 𝑚 𝑒 𝑘 2
𝑒 ħ
3
(9)
Where 𝑚 𝑒 is the mass of an electron, e is the charge of an electron and h is the Planck
constant. For thermionic emission, as indicated in Richardson-Dushman relation, there is
no need for an externally applied electric field. Instead, some of the excited electrons
have sufficient energy to overcome the work function and escape from the metal as seen
in Figure 2.3. The emission is governed by electrons whose energies are on the tail end
of the Fermi-Dirac distribution and have energies greater than that of the barrier height.
This means that 𝑘𝑇 must be on the order of 𝑊 . It should be noted, for 2D materials such
as Graphene, the modified Richardson Dushman equation is reported which indicates
thermionic emission has a direct relation to temperature to power three and will be
discussed further in chapter
5
.

Figure 2.3 Potential barrier seen by electron – During thermionic emission process, system
is heated, and electron obtain enough energy to overcome the potential barrier. This
process doesn’t require external applied field.

22

2.2.3 Photo Emission

In photo-field emission the system absorbs photons however the absorbed energy
is smaller than needed value to overcome the barrier. With an externally applied electric
field, a penetrable barrier is formed through which electrons can tunnel as seen in Figure
2.4. The emission rate can be described by the Fowler-Nordheim equation, with the work
function replaced by the effective barrier height
∅
𝑒𝑓𝑓 ′
=∅
𝑠𝑐 ℎ𝑜𝑡𝑡𝑘𝑦 −𝑛 ħ𝜔 (10)
Where ∅
𝑠𝑐 ℎ𝑜𝑡𝑡𝑘𝑦 is the work function described by the Schottky reduction in section 2.1
and n is the number of photons absorbed with nhω < φ. Photo-field emission can be
identified by comparing the emitted current from tips with and without laser illumination
3
.
Assuming the tip radius does not change, the effective work function can be deduced
from the FN equation. Photo-field emission can also be identified in the electron energy
spectra by the asymmetrical shape of the field emission and the sharp edges around the
energies of one-photon and two-photon absorptions
12
.

Figure 2.4 Potential barrier seen by electron – During Photo emission process, electron
absorbs photon and get excited, this process is equivalent to observing smaller effective
work function.

23

2.2.4 Intense Photo-Field Emission

By increasing the laser intensity, the emitters will absorb more photons than
necessary for the electron to overcome the potential barrier defined by the work function.
This result in the emission of electrons with larger energies. The electron gain energy as
large as nhω, where n is the number of photons absorbed and ω is the angular frequency
of the photon. At ultra-high laser intensity, the laser can no longer be considered only
group of coherent photons with energy hω. For a small portion of the optical cycle, the
laser field acts as a strong DC field, creating a penetrable barrier through which electrons
can tunnel. This process is called strong photo-field emission.

24

2.3 Material for Electron Emission

2.3.1 Semiconductor based Emitter

Semiconductor materials have larger optical absorption compared to metallic
emitter specially with the advances on fabrication of the semiconductor nanowires as well
as near to ideal ohmic contact to indium arsenide, they have attracted considerable
scientific interest. However, field-emission based emitter made from semiconductor tips
are often difficult to operate in a stable mode. The Fermi level of semiconductor is below
the conduction band and when an electric field is applied as needed for field electron
emission, insufficient charge carriers are available to maintain a zero potential at the
surface, leading to band bending. The resulting emission process is highly unpredictable
and exhibits a large energy spread of several tens of electron volts of the emitted electron
beam. For these reasons, field emitters are usually made from a sharp metal tip or a tip
of a material with metal-like electronic properties, such as a carbon nanotube or
graphene.

25

2.3.2 Metallic based Emitter
Lanthanum hexaboride (LaB6) has the lower work function (of 2.69 eV) compared
to most of metallic emitter. LaB6 with melting point of 2210 °C has one of the highest
electron emissivities and is stable in vacuum. Hexaboride cathodes are about ten times
"brighter" than tungsten cathodes and have 10–15 times longer lifetime. The physical
difficulties reported with LaB6 is a tendency of Boron to diffuse in to underlying metal
lattice with the formation of interstitial metal-boron alloy. This process let lanthanum
atoms to escape and evaporate. The device continue emission until either all the
interstitial space has been filled or the whole of the lanthanum has evaporated. However,
this material is being used extensively in electron microscopes, microwave tubes, electron
lithography, electron beam welding, X-ray tubes, and free electron lasers.

Figure 2.5 Structure of Lanthanum Hexaboride

26

2.3.3 Carbon based Emitter (Graphene)

Graphene is a well-known material for its electric properties. The unit cell of
graphene is plotted in Figure 2.6. The edges parallel to the y-axis are called zigzag edges,
while the edges paralleled to the x-axis are armchair edges. It is reported that an emission
current from the armchair edge is much stronger than that from zigzag edge
13
. The
electron emission efficiency from flat sheet of graphene is low, because electrons emit
only from the abovementioned edges as well as sharp edges due to the decrease of work
function of sheet at edges
14,15
.
On the other hand, silicon pillars or patterned silicon enable E-field enhancement.
Addition of graphene with favorable electronic structure with sharp tip structures could
yield a combined characteristic. For this reason, various efforts have been made to
fabricate silicon-based heterojunction nanostructures as field emitters that exhibit higher
current density at lower electric field. The results enable to reduce the threshold voltage
to below 5V/µm
16,17
.

Figure 2.6 Idealized structure of a single graphene sheet. Copyright
18

27

The graphene on the Si substrate can be investigated through Raman spectroscopy. The
graphene related peaks are ~1384, 1582 and 2693 cm−1, corresponding to the D, G and
2D modes. The D band arises from backscattering of phonons by edges and defects like
corrugation, twisting and edges. The G band is associated with doubly degenerate E2g
phonon modes of the sp
2
hybridized carbon network of graphite. A symmetric 2D peak
originates from a second-order process, indicating the graphene layers. The intensity ratio
of the D-to-G peaks can be applied to indicate the disorder degree of graphene. Also, the
intensity ratio of G to 2D bands indicates number of graphene layer. It has been reported
that I2D/IG ratio decreases with the increase of layer numbers and a value greater than 2.0
indicates the presence of monolayer graphene
19-20
.

Figure 2.7 Idealized Raman spectrum with a 514.5 nm excitation laser
wavelength of pristine single-layer and multi-layer graphene
19
.

28

Chapter 2 References

1. J.M. Sellier, Archimedes, the Free Monte Carlo simulator, arXiv:1207.6575, 2012.
2. JM Zhang, Y. Liu, European Journal of Physics, Vol 37, No 6 (2016)
3. Hommelhoff, Peter, et al. Physical review letters 96.7 (2006): 077401.
4. R. Gomer, Field Emission and Field Ionization, Harvard Univ. Press, Cambridge, MA, (1961).
5. Gomer, Robert. Field emissions and field ionization. American Inst. of Physics, 1992.
6. Good Jr, R. H., and Erwin W. Müller. "Field emission." Electron-Emission Gas Discharges
I/Elektronen-Emission Gasentladungen I. Springer Berlin Heidelberg, 1956. 176-231.
7. R.W. Wood, Phys. Rev. 5, 1, (1897).
8. Schottky, W. "Cold and hot electron discharges." Z. Phys 14 (1923): 63
9. Fowler, Ralph Howard, and L. Nordheim. "Electron emission in intense electric fields."
Proceedings of the Royal Society of London A: Mathematical, Physical and Engineering Sciences.
Vol. 119. No. 781. The Royal Society, 1928.
10. E. L. Murphy and R. H. Good, Jr. Phys. Rev. 102, 1464 (1956)
11. Conyers Herring and M. H. Nichols - Rev. Mod. Phys. 21, 185 (1949).
12. Hirofumi Yanagisawa et al., Phys. Rev. Lett. 107, 087601 (2011)
13. Weiliang Wang et al., Journal of Applied Physics 109, 044304 (2011)
14. Santandrea, S., et al. Applied Physics Letters 98.16 (2011): 163109.
15. Eda, Goki, et al. Applied Physics Letters 93.23 (2008): 233502
16. Lv, Shasha, et al. Scientific reports 5 (2015): 15035.
17. Chang, Tinghsun, et al. "Enhanced electron field emission properties from hybrid
nanostructures of graphene/Si tip array." RSC Advances 5.4 (2015): 2928-2933.
18. Chris Ewels, (http://www.www.ewels.info)
19. Tuinstra, F., and J. Lo Koenig. "Raman spectrum of graphite." The Journal of Chemical
Physics 53.3 (1970): 1126-1130.
20. Zheng Yan and Andrew R Barron, Characterization of Graphene by Raman Spectroscopy
(2015).

29

Chapter 3
Light & Dark Electron Emission Characterization
3.1 Building an Experimental Setup
3.1.1 Ultra-high Vacuum Chamber Design and Installation
The field emission setup is designed and constructed to achieve ultra-high vacuum
within reasonable time. We targeted to have easy loading-unloading system. For this
purpose, a high-speed oerlikon leybold vacuum turbo pump (TURBOVAC 350 iX) is used
with the easy access opening door. In our pumping scheme, there is a gate valve between
main testing chamber and pumping chamber which let us to isolate these two chambers
while loading the new sample in to testing chamber. The bake out process is used for the
initial usage to outgas the moisture. For bake out, the heat tape is used and set the current
setting such that the temperature reaches close to 120 ºC. The base pressure of our
pump goes down to 1e-10 Torr. However, our measurement performed under 1e-8 Torr
that is achievable within reasonable time after loading graphene and metallic emitter
inside chamber. However, the photon assisted electron emission characterizations
require optical fiber inside vacuum chamber. The fiber has plastic jacket and it takes
longer time for achieving high vacuum due to degassing of the fiber’s jacket.

3.1.2 High Voltage Power Supply
This chamber includes electrical feed through as shown in Figure 3.1 from MPF
that can stand up to 30kV. The Spellman high voltage source is used to apply adjustable
30

voltages 0 to 30kV. The source has the current handling is up to 10 mA. The power supply
uses special high-voltage cable provided from Spellman to apply voltage on anode. The
power supply is programmable and operates in remote mode.

3.1.3 Electron Emission Detection Equipment
The electron emission current measurement performed via different equipments
as shown in Figure 3.2, mainly depend on the emission current level. For high current
measurement, we use the Keithley digital multimeter 2100 that can detect from a µA up
to 3A level of current. For smaller current detection, the Keithley picoammeter 6485 is
used that measures from tens of pA to 20 mA level. For current level smaller than pA, we
use the Keysight electrometer B2985 which can measure fA levels of current. These
equipments are sensitive to the input voltage and breakdown at high voltages that occurs
in case of vacuum breakdown from anode to cathode. As such, I added a protection

Figure 3.1 Field emission setup constructed in SSC 508 Lab

31

circuitry before the signal goes in to the detection equipment. The protection circuit shown
in Figure 3.3. The diode leakage under threshold voltage is small and it doesn’t interfere
with equipment measurement of 10 pA or more. As such we used this for picoammeter
protection. The resistor should be high enough in resistance to limit the diode current to
a safe value and should withstand the maximum input voltage without burning up. With
two diodes back-to-back, the circuit will provide protection regardless of the input polarity.

Figure 3.2. Field emission measurement equipment constructed in SSC 508 Lab.
445 nm laser source
Electrometer
HV Source
405 nm laser source
Pulse generator
PC

Figure 3.3 Protection circuitry for equipment.

Picoammeter
Input Signal
32

3.1.4 CW Tunable Fiber Coupled Laser Source
For optical signal, we first used Thorlab driver DC2200 on two fiber coupled LED
source to generate wavelength 385 nm and 660 nm with nominal output power of 23 mW
and 14 mW respectively. The vacuum chamber has optical feedthrough as shown in
Figure 3.1 and we guide the optical power in to vacuum chamber on sample using this
optical feedthrough. The optical feedthrough is from MPF and we selected the multimode
with a diameter of 400 µm with standard NA of 0.22 for transferring optical power. The
fiber coupled LED source has a larger NA, 0.39 and this mismatch result in optical loss
and the power was not enough for generating photo-emission response. Later, we used
the CW laser source with tunable output power from wave-spectrum at 445 nm. The
power is adjustable from 0 ~ 2W along with an optical fiber with a diameter of 200 µm.
We also used tunable and modulated laser source at 405 nm (Figure 3.4). This laser
source has a tunable power 0 ~ 1W along with bundle fiber including 4 optical fibers with
diameter of 105 µm. It accepts up to 1kHz RF signal 0-5V for analog modulation. In
section 3.4, I will describe the characterization of both laser sources which we used for
photon assisted electron emission experiments.
33

3.1.5 Automated Test Setup
The high voltage source and electron emission detection equipments are programmed with
MATLAB for automated controlling. Our laser sources are not equipped with automation and
operated manually.

Figure 3.4 (left) The 445-nm laser system 0 – 2W. (right) The 405-nm laser source 0 – 1W
with modulation.
34

3.2 Hybrid Emitter: Fabrication and Characterization
3.2.1 Pyramid Silicon Array Fabrication
Fabrication procedure of hybrid field emitter on pyramid array is summarized in
Figure 3.5. The silicon tip array structure was fabricated on (1 0 0)-oriented, heavily doped
n-type two-inch silicon wafers. This wafer size was chosen to ensure that there were no
field emission effects from wafer edges and that all the measured emission current occurs
from microfabricated silicon tips. The fabrication procedure starts with removing the native
oxide using BOE 7:1. Then, LOR 5A was spin coated with 500 rpm for 5 s and 3000 rpm
for 60 s and baked at 175 °C for 5 min. Next, AZ5214 photoresist was spin coated under
the same conditions and baked at 110 °C for 50 s followed by a dose of 40 mJ/cm
2
. The
pattern used here was a 1 cm
2
array of circles with 20 µm diameter and 35 µm center to
center distance. After exposure, the sample was baked at 115 °C for 2 min prior to a flood
exposure with a dose of 240 mJ/cm
2
. The second bake and exposure enable image
reversal of the mask pattern. After patterning, silicon pillars were formed by etching silicon
using inductively coupled plasma (ICP) etching, Bosch process. We used an SF6 flow
rate of 100 sccm and an C4F8 flow rate of 1 sccm under a plasma forward power of 700
W. Under this controlled condition, we repeated the process for 125 Bosch cycles and
obtained the pillars with 25 µm height. After this step, we immersed the silicon pillars in a
stirred 30% KOH solution at 75 °C, which resulted in the desired tip shape. The stirring is
crucial step in getting pyramid silicon array and leaving sample inside KOH result in high
aspect ratio silicon rods instead of pyramid structure as shown in Figure 3.6. From
experimental results, the emission properties including the field enhancement factor and
threshold field emission of pyramid silicon array is significantly improved compared to thin
35

rods. The SEM image of the pyramid silicon array is shown in Figure 3.7 that indicates
the uniformity of the pyramid shape over the area of interest. Furthermore, by increasing
KOH etching time (over-etching), sharper tips can be obtained. The SEM images of sharp
tips silicon array is shown in Figure 3.8. For these cases, the height of the pillars reduces
significantly as it is getting etched inside KOH.

Figure 3.5 Pyramid Silicon Array Fabrication Procedure (a) Photo-resist spin coating
(b) Pillar patterning using photo-lithography (c) Silicon pillar fabrication using ICP for
etching silicon (d) Pyramid array formation using KOH etching (e) adding 10 nm gold
contact (f) transferring Graphene followed by drop casting Lanthanum hexaboride nano
particles.

(a) (b)
(c) (d)
(e) (f)
Photoresist Silicon Au
Graphene
LaB6 nanoparticle
36

Figure 3.6 SEM image of silicon pillar that changed to thin rods, the pillars have original
diameter of 10 µm and leaving sample inside hot KOH solution result in thinner silicon
rods. The emission properties of silicon rod are not as significant as pyramid silicon.

Figure 3.7 SEM image of reduced diameter rod vs. pyramid silicon, (a) – (b) the pillars
have original diameter of 10um and I left them in hot KOH solution for reducing the
diameter. (c)-(d) the pillars have original diameter of 20 µm and I obtained pyramid array
with reduced tip after stirring the hot KOH etchant. The emission properties of pyramid
array indicate significant improve in electron emission.

37

In the next step, I evaporated 10 nm Au as a uniform contact above pyramid silicon
tip array. This step is added to fabrication process after observing inconsistency between
measurements over time. Given the Boron element of LaB6 tends to diffuse in to silicon,
electrons no longer experience low work function of LaB6 and I-E curve will have an
inconsistent threshold field over time. We added 10 nm gold to keep the LaB6 nano
particle intact on the surface to maintain the low work function.

Figure 3.8 SEM image of sharp tip array, this tip is fabricated by over-etching pyramid silicon
in hot KOH.

38

3.2.2 Wet Transfer of Graphene
Graphene was transferred on top of pyramid silicon array as graphene emitter. The
graphene used in our work was originally grown on copper foil by the chemical vapor
deposition (CVD) method. The wet transfer technique was used as described in Figure
3.9. First, the copper was completely dissolved in a ferric chloride solution. Thus, the
graphene and PMMA float on the surface of the etchant. After being cleaned multiple
times with deionized (DI) water, Then, PMMA and graphene layer will be washed with
acid for 20 min and later will be washed in DI water. Then this film was manually lifted out
from water by silicon substrate including pillars. Finally, PMMA will be removed using
acetone or rapid thermal annealing (RTA). After transfer, Graphene is characterized using
Raman-spectroscopy (Figure 3.10). The result indicating sharp peak of the 2D band at
2689 cm-1 referring to the second-order vibration caused by the scattering of phonons. I
also observed G band peak at 1580 cm-1 resulting from the E2g vibration mode of sp
2

bonded carbon. There is a correlation between the ratio of the I2D/IG and number of layers
in graphene sheet. It has been reported that I2D/IG ratio decreases with the increase of
layer numbers and a value greater than 2.0 indicates the presence of monolayer
graphene
1
. I measured I2D/IG value of 2.5 which indicates formation of monolayer
graphene on field emission device.

39

Figure 3.10 Pyramid silicon tip array after graphene transfer and Raman spectrum that indicates
monolayer graphene

Figure 3.9 Graphene wet transfer procedure

a. CVD graphene
Cu foil
b. Cu substrate etching
PMMA/Graphene
Copper etchant
c. PMMA/Graphene
in DI water
d. Washing Graphene
in HCl solution
e. Washing Graphene
in DI water
10% HCL
f. Graphene transferred to
sample
40

3.2.3 Low Work Function Lanthanum Hexaboride Drop-Casting
The LaB6 nanoparticles used in this work is in DI water solvent. An enhanced
emission current has been measured from thin LaB6 layer specifically layers thinner than
10 nm
2
. Therefore, the LaB6 nanoparticles with an average size of 3-4 nm were
synthesized using a previously reported method
3
. Briefly, 1.0 g anhydrous LaCl3
(4.1mmol) and 0.95 g NaBH4 (25.1mmol) were mixed under argon for 20 min and then
heated to 360 °C at a rate of 10 °C/sec. The reactants were stirred at 360 °C for 60 min
and then cooled to room temperature. Work-up was performed in air, where MeOH was
used to remove excess NaBH4, HCl to convert residual Na to NaCl, and deionized water
was used to wash out the NaCl. Then we made 1:10 dilution of nanoparticle solution and
isopropyl alcohol and drop casted on graphene sheet on sharp tip array while leaving the
sample under fume hood to allow evaporation of the solvent. The SEM image of the hybrid
emitter after nanoparticle deposition is shown in Figure 3.11. Similarly, I drop casted LaB6
nano particle on graphene that is supported by pyramid silicon and the SEM image is
shown in Figure 3.12. By comparison, we can see the sample with Graphene has a
residue of Graphene transfer.

41

Figure 3.12 (a) SEM image of LaB6 nano particle drop-casted on Graphene
supported by sharp tip silicon pyramid array

Figure 3.11 (a) SEM image of LaB6 nano particle drop-casted on sharp tip silicon
array (b) SEM image of LaB6 nano particle drop-casted on graphene supported by
sharp tip silicon array

42

Following figure shows the summary of the major fabrication steps:

Figure 3.13 SEM images of the fabrication steps for the sharp tip silicon array including (a)
Initial silicon pillars array, (b) Sharp tip silicon array after anisotropic etching inside KOH, and
(c) Cross section tilted view of single sharp tip silicon pillar (d) Cross section view of sharp tip
silicon array after graphene sheet transfer and LaB6 nanoparticles deposition.
(d)
50 µm
5 µm
(e)
50 µm
(a)
50 µm
10 µm
(c)
(b)
(d)
43

3.2.4 Work Function Characterization
This hybrid emitter of LaB6 nanoparticles on graphene sheet has a work function
that is different from pristine LaB6 nanoparticles and graphene sheets. It is well known
that the work function of a material is extremely sensitive to surface orientation, oxides,
organic residues, etc. Thus, to accurately determine the field enhancement factor, the
work function of the samples must be directly measured. To precisely determine the work
function, we used a photoemission spectroscopy (PES) process
4
on the planar versions
of our devices. The measured valence band spectrum with the Fermi level (Ef) at zero
binding energy and the secondary electron cutoff of the photoelectron spectrum is shown
in Figure 3.14.
Based on this experimental data, we used the previously reported method
4
and
measured the value of the effective work function for LaB6 nanoparticle emitter on 10 nm
Au 3.3 eV, which is higher than work function of the pristine bulk LaB6 and it is due to
underlying Au contacts. Furthermore, the work function of the LaB6 nanoparticle on
graphene sheet on 10 nm Au is measured 3.62 eV, slightly higher than LaB6 nanoparticle

Figure 3.14 (a) Valence band spectrum of Au, Graphene, LaB6 and hybrid emitter with the
Fermi level (Ef) at zero binding energy (b) secondary electron cutoff of the photoelectron
spectrum
5000
4000
3000
2000
1000
0
Intensity (Counts)
10 8 6 4 2 0 -2
Binding Energy (eV)
Au
LaB
6
LaB
6
+ Graphene
Graphene
100x10
3
80
60
40
20
0
Intensity (Counts)
1484 1482 1480 1478 1476
Kinetic Energy (eV)
LaB
6
+ Graphene
LaB
6
Graphene
Au

44

on 10 nm Au and significantly lower than pristine graphene sheet. This measurement
verifies graphene emitter work function can be reduced significantly at the surface using
low work function nanoparticles. After fabrication of the devices and materials
characterization, we carried out field emission measurements.
3.2.5 Electron Emission Characterization
5

The field emission characteristics for each device were measured at room
temperature under a vacuum of 10
-8
Torr. The current measurement was carried out using
a Keithley 6485 picoammeter connected directly to our cathode, ensuring that all
measured current was field emission, and no secondary electron emission current was
measured. The field emitter device served as the lower electrode (cathode) and the high
voltage top electrode (anode) made of stainless steel was positioned 1 mm above the
cathode. The patterned area was 1 cm
2
array of sharp tips. The distance was measured
using an MDC linear motion feedthrough. Specifically, the anode was brought into
electrical contact with a non-emitting area of the sample and then retracted by the amount
desired for the anode-cathode separation. The high voltage was applied and swept using
a Spellman high voltage source with a positive voltage. We measured the emission
current from bare silicon, Si/Au, Si/graphene, Si/Au/LaB6, and Si/Au/graphene/LaB6, each
in a planar and tip array configuration. We show the I-E curves in Figure 3.15a and 3.15b
and extracted the threshold field (defined as electric field required for J = 10 µA/cm
2
) as
shown in Table 1. We see that the bare silicon exhibits the highest threshold field of
Vth=12.5 V/µm while the Si/Au/graphene/LaB6 hybrid on an array of silicon tips shows a
much lower threshold field of Vth=2.6 V/µm. We also observed the highest current density
from hybrid Si/Au/graphene/LaB6 emitter at higher field, higher than Si/Au/LaB6 emitter.
45

This is due to contribution of the graphene sheet to emission current which starts at higher
field compared to LaB6 emitter. In addition, the threshold field for emitters on the silicon
tip array is 1.4x-1.7x smaller than the ones on planar (non-patterned) silicon substrates
due to engineered increase of the field enhancement factor. Specifically, the hybrid
emitter on the silicon tip array has the lowest threshold field of Vth=2.6 V/µm, which is 1.7x
lower than the hybrid emitter on planar substrate.
To analyze the results, Fowler-Nordheim theory was used to correlate the current
density and local electric field to the geometrical and material properties of the fabricated
emitter. The Fowler-Nordheim (FN) theory of electron emission is encapsulated by the
following equation
6,7
:
𝐽 =(
𝐴 𝛽 2
𝐸 2
∅
)exp(−
𝐵 ∅
3/2
𝛽𝐸
) (1)
Here, A and B are constants equal to 1.54e-6 A eV V
-2
and 6.83e3 eV
-3/2
V µm
-1
,
respectively. 𝛽 is the field enhancement factor, E is the applied field calculated from the
ratio of the applied voltage to the cathode-anode distance, J is the emission current
density obtained from the total measured current divided by the area of the sharp tip
silicon array, and ∅ is the emitter work function. The field enhancement factor 𝛽 can be
calculated using the slope of the fitted straight line from a curve of ln(
𝐽 𝐸 2
) versus
1
𝐸 if the
work function is known. Typically, the work function of the material and field enhancement
factor are both treated as variables, utilizing the previously measured work function, such
as 4.57 eV for graphene
8
and the work function measured using PES for the emitters
which had the LaB6 nanoparticles. After establishing the work function of our emitters,
we can accurately calculate the field enhancement factor of these samples from the slope
46

of the FN plots as shown in Figure 3.15c-d. The extracted field enhancement factors are
tabulated in Table I.

Figure 3.15. (Color Online) Emission characteristics of different emitter and corresponding FN
plots (a)-(c) on planar substrate, (b)-(d) on sharp tip silicon array, (e) Simulated geometrical field
enhancement, (f) Emission current stability test.

800
600
400
200
0
Emission Current (µA)
8 6 4 2 0
E-Field (V/µm)
Sharp tip substrate
Graphene + LaB
6
Graphene
LaB
6
Au
Silicon

-32
-28
-24
-20
-16
ln(J/E
2
) (µA / V
2
)
0.4 0.3 0.2 0.1 0.0
1/E (µm/V)
Planar substrate
Silicon Au
Graphene LaB
6
Graphene + LaB
6
-32
-28
-24
-20
-16
ln(J/E
2
) (µA/V
2
)
0.5 0.4 0.3 0.2 0.1 0.0
1/E (µm/V)
Sharp tip substrate
Silicon Au
Graphene LaB
6
Graphene + LaB
6
0
2.17
4.17
6.17
7.75
|E| V/µm
400
300
200
100
0
Current Density (µA/cm
2
)
300 200 100 0
Time (min)
Graphene (Planar)
Graphene + LaB
6 (Planar)
Graphene + LaB
6 (Sharp tip)

(a) (b)
(c)
(d)
(e)
(f)
300
250
200
150
100
50
0
Emission Current (µA)
12 8 4 0
E-Field (V/µm)
Planar substrate
Graphene + LaB
6
Graphene
LaB
6
Au
Silicon
47

The field enhancement factors from emitter on planar substrates are due to local
protrusions within emitters that enhance the electric field as well as the nature of the
constants in the FN equation. Specifically, those constants do not include material specific
information, and are derived from certain assumptions. However, since we have
measured all samples with both planar and tip array geometries, we can look at the
relative change of the field enhancement factor with no loss of accuracy. It should be
noted that the total electric field enhancement factor of hybrid emitters can be represented
by
9

𝛽 overall
= 𝛽 geometry
.𝛽 emitter
(2)
where, 𝛽 geometry
, 𝛽 emitter
are the geometrical field enhancement factor from the silicon
tip array, and field enhancement factor due to emitter material and surface roughness,
respectively. The value of 𝛽 emitter
was obtained from emitter on planar silicon substrate
and 𝛽 overall
was also obtained from hybrid emitter experimentally. Therefore, the value of
the geometrical field enhancement, 𝛽 geometry
can be calculated. The obtained value of
𝛽 geometry
for graphene is 2.51, for LaB6 is 1.57 and for the hybrid emitter is 2.16. To
validate the observed results, we used Finite Element Method Magnetics (FEMM)
software to calculate the expected geometrical field enhancement. The geometry of the
silicon tip array was obtained from SEM images, including the geometry of the tip itself.
The microfabricated tip has a spherical end with a diameter of 300 nm. In our simulation,
1000 V was applied to the HV electrode, which was separated by 1 mm from the silicon
tip array. This configuration results in 1 V/µm uniform field in the case of a planar cathode;
however as shown in figure 3.15e, the electric field intensity is enhanced by 6x at very
48

small area of the spherical part of the silicon tip and 2.5x at larger area of the lower level
of silicon tip. This result is in good agreement with the experimental observation.
As the graphene emitter should have the least surface roughness in the planar form,
we expect it would give results most closely matching simulation, which is what is
observed. However, the measured field enhancement factor value difference between
planar and tip array LaB6 emitters is quite a bit lower. We believe this discrepancy occurs
due to the drop casting method of the LaB6 particles on the planar silicon, where it is
possible for aggregation to occur, which could cause deviations from the planar geometry
assumed here. This drives up the field enhancement factor of the planar electrode and
reduces the difference with the silicon tip array.
TABLE I. Threshold Field Vth for J = 10 µA/cm
2
and Field Enhancement Factor

V th (V/µm) for J= 10
µA/cm
2

Field Enhancement
Factor 𝛽
Planar silicon 12.5 450
Silicon Tip Array 5.8 1217
Au on Planar Silicon 8.2 1275
Au on Silicon Tip Array 4.9 2488
Graphene on Planar Silicon 6.4 904
Graphene on Silicon Tip Array 4.5 2270
LaB6 on Planar Silicon 4.8 1094
LaB6 on Silicon Tip Array 3.2 1716
LaB6 on graphene on Planar Silicon 4.5 1285
LaB6 on graphene on Silicon Tip Array 2.6 2775

49

Stability of the field emission current is another important parameter. The result of
stability test for 6 hours is shown in figure 3.15f and indicate that there is no degradation
throughout the stability test. We observed stable electron emission from the graphene
emitter during the testing period. In addition, emission current from LaB6 nanoparticle
emitters shows small variation at the beginning before stabilizing. This longer time was
required for LaB6 nanoparticle to get rid of moisture and to desorb residual gas molecules
after drop casting them on the emitter surface. This trend of initial variation in emission
current observed and reported by others as well
10
.
BandDiagram of the Hybrid Emitter

Adding LaB6 nanoparticle on Graphene will change the band diagram of the hybrid
emitter as shown in Figure 3.16. As such, electron starts travelling from heavily doped
silicon toward the surface of the hybrid emitter. Electrons at the surface of hybrid emitter
requires energy at the order of LaB6 work function (2.7 eV which is smaller than graphene)
to emit to vacuum.

Figure 3.16 Band Diagram of the hybrid emitter from TCAD Sentaurus simulation
-100
-50
0
50
100
Energy (meV)
30 20 10 0
Position (nm)
E
c
E
F
6
4
2
0
-2
Energy (eV)
30 25 20 15 10 5 0
Position (nm)
E
c
E
v
E
F
Vacuum Level
1 2 3
4
4. Silicon – n Type
1e19
1. LaB6 ( 4 nm)
2. Graphene (1 nm)
3. Gold (10 nm)
𝑒 −
𝑒 −
𝑒 −
50

3.2.6 Summary
In conclusion, this section reported a technique for independent engineering the field
enhancement and work function of a field emission device. Specifically, the silicon tip
array enhances the local field intensity in the graphene emitter and the LaB6 nanoparticles
deposited on the surface of the graphene reduces the effective work function of the
emitter. We have showed experimentally that these two techniques drastically improve
the electron emission performance, reducing the threshold field by about 5x. We also
performed simulations of the geometrical field enhancement factor to evaluate the
contribution of the geometrical field enhancement and find excellent agreement with our
experimental results. In the future, higher aspect ratio pillars can easily be made to
improve the field enhancement factor from the present value of 2-4 as compared to planar
to upwards of 25, which would result in ultra-low turn on voltages of Vth~0.25 V for these
devices.

51

3.3 Tunneling Assisted Electron Emitter
3.3.1 Device Schematic and Fabrication
The tunneling effect contribution in electron emission has been investigated since
1960 after C.A Mead introduced the new class of emission device that employs the
principle of tunnel emission
11
. The initial investigation focused on MIM diode structures
made of alumina sandwiched between two aluminum layers in which the top metal should
be very thin. The thickness of the dielectric layers in the early investigation was up to 500
nm and this ends up requiring an extremely high field up to hundreds of V/µm to initiate
the tunneling electron injection and observation of emission current limited to hundreds
of nA. After this initial report, other researchers started reproducing their preliminary
results as well as investigating the structure using other materials and thickness
12-16
. In
1999, H-J Fitting et al investigated the tunnel electron emission from MIS structure
consists of Au (8-12 nm) - SiO2 (10-500 nm) - n type Si sandwiched structure in to vacuum
via electron counting by a chaneltron. They measured emitted electron energies up to
200 eV for thick oxides layers under extremely high field strength of 500V/µm
16
. To initiate
the injection of electron through thick dielectric layer, they grounded the top Au layer and
applied negative driving bias of 20 V and higher to the substrate semiconductor. Because
of driving bias, the n-type silicon substrate Fermi level shifted to higher value compared
to Au Fermi level and result in triangular barrier enabling tunnel injection of electron from
the conduction band of Si in to SiO2 at very high field. Parts of the injected electron scatter
inside SiO2 and others with enough energy go through the oxide layer, arrive to top metal
and will be emitted through the thin top Au electrode in to vacuum.
52

Later, researchers focused on making needle shape array of emitters or using
carbon nanotube array to enhance the electric field intensity for field electron emission
17,18,19
. However, complex fabrication and non-uniformity was the big challenge of the
sharp tip arrays emitter whereas the planar emitters are simple and easy to fabricate. In
2000, Binh et al.
20
proposed a planar structure for cathode by modifying the electronic
properties of the underneath surface layer, called solid-state field-controlled emission
(SSE). They added an ultrathin wide band-gap semiconductor layer on a metal that result
in lowering the surface barrier up to 0.1 eV. In 2004, V. Semet et al
21
. presented the
experimental measurements of the field electron emission from a multilayer planar
nanostructured emitter. They described the introduction of the multilayer concept gave
them more parameters for the control of SSE process, such as the bulk interfacial barrier
in addition to the surface barrier and explained their results based on a dual barrier
mechanism for electron emission through nanostructured layers. In 2012, Changzhi Gu
22
reported the field electron emission properties of a planar cathode consists of
diamond/CoSi2/Si quantum well nanostructure and demonstrated the main emission
properties were modified by varying the CoSi2 thickness and they reported low-field, high
emission current and controlled electron emitter.
In this section, I intend to report the emission properties of planar metal-insulator-
semiconductor (MIS) emitter with ultrathin dielectric layer to generate a narrowband and
sharp distribution of hot electron injection from heavily doped substrate to metal layer
using least fabrication process and no additional bias across dielectric layer. Specifically,
I fabricated two different stacks of MIS emitters and measured the electron emission
characteristics from these emitters and their counterpart without dielectric layer under
53

high vacuum and compared the results. The experimental results are interpreted using
the band diagram analysis.
Figure 3.17 shows the schematic of MIS and MS emitters used in this work for
investigating the direct tunneling assisted electron emission. The first stack of MIS emitter
consists of 10 nm thin layer of Au as top emitter metal evaporated on three different
thickness (5 - 10 - 15 Aº) of ultrathin aluminum oxide (Al2O3) as insulator layer which is
deposited by atomic layer deposition (ALD) on the heavily doped N
++
silicon (resistivity =
0.001-0.005 Ω.cm) semiconductor substrate. The native oxides on substrates are etched
via HF right before ALD process. The second stack of MIS emitter consists of 10 nm thin
layer of conductive Strontium Ruthenate (SrRuO3) on 12 Aº ultrathin epitaxially grown film
of Lanthanum Aluminate (LaAlO2) on Nb doped Strontium Titanate (Nb- SrTiO3). For this
sample, the thin films were synthesized by pulsed laser deposition (PLD) with in situ
monitoring of reflection high-energy electron diffraction (RHEED) on TiO2-terminated
(001) Nb: doped (wt% = 0.7) SrTiO3 substrates. The substrates were prepared by etching
with 50:1 buffered HF solution and annealing at 1050 ºC in flowing O2. The films were

Figure 3.17 (a) schematic of the MIS emission device, (b) schematic of the MS emission device
HV
A

54

grown using a single-crystal LaAlO3 target and a polycrystalline SrRuO3 target with a
target-substrate distance of 7.5 cm, using a 248 nm KrF laser and an energy fluence of
~1.5 J/cm
2
. Both films were grown with the substrate heated to 750 ºC. The LaAlO3 film
was grown in 10
-3
mbar O2 with a laser rep rate of 1 Hz. The SrRuO3 film was grown
inside 2×10
-2
mbar O2 with a laser repetition rate of 5 Hz. After growth the chamber was
flushed with oxygen to 400 mbar before cooling to sample to room temperature. The MS
emitter is prepared with same thickness of top metal and without dielectric layer on a
same substrate used for the above mentioned two cases.
3.3.2 Electron Emission Characterization
I investigated the field emission characteristics for both cases at room temperature
under a vacuum of 10
-7
Torr. The current measurement was carried out using a Keithley
6485 picoammeter connected directly to our cathode. The emitter device served as the
lower electrode (cathode) and the high voltage top electrode (anode) made of stainless
steel was positioned 1 mm above the cathode. The distance was measured using an
MDC linear motion feedthrough. Specifically, the anode was brought into electrical
contact with a non-emitting area of the sample and then retracted by the amount desired
for the anode-cathode separation. The high voltage was applied and swept using a
Spellman high voltage source with a positive voltage. The I-E curve measured from first
sets of emitters is shown in figure 3.18(a) and the corresponding Fowler Nordheim (FN)
plot is shown in figure 3.18(b). The emission curve for Au on 5 Aº and 10 Aº of dielectric
layer shifted to lower threshold value compared to Au on silicon substrate. I also noticed
the emitter with dielectric thickness of 15 Aº demonstrate similar emission curve as the
emitter with no dielectric layer. From these results, one can interpret that direct tunneling
55

cannot occurs through the dielectric thickness of 15 Aº or higher. The FN plot indicate
steeper slope for emitter with 5 Aº and 10 Aº of dielectric layer compared to emitter without
or with 15 Aº dielectric layer. The field enhancement factor from Au with no dielectric layer
is back calculated as 650 assuming the work function of the Au is 5.1eV. This value
indicates the roughness in top Au layer from our metal evaporation. Assuming this is
constant throughout different samples and using this value of field enhancement factor
(650), I calculated the work function for emitters with dielectric layer using FN plot. I found
the work function of 3 eV for the emitter with 5 Aº dielectric layer, 3.5 eV for emitter with
10 Aº and 5.1 for the emitter with 15 Aº dielectric layer. The work function values for
emitter with 5 Aº and 10 Aº are lower than the work function of the Au. This smaller work
function is due to the distribution of the injected hot electron from heavily doped silicon in
to top Au emitter which lies above Au Fermi level. At the absence of dielectric layer, the
electron transport from heavily doped silicon to metal layer result in a wide band
distribution of lower energy electrons that are not able to contribute in emission process.
Comparable results are observed for second set of emitters. The I-E curves for emitter
with and without dielectric layer of LaAlO3 is shown in figure 3.18(c) and the
corresponding FN plot is shown in figure 3.18(d). I observed the threshold field shift to
lower value. The metallic SrRuO3 and Nb-SrTiO3 substrate form a Schottky barrier
whereas adding LaAlO3 dielectric layer avoids the band bending and let the sharp
distribution of hot electron injection in SrRuO3 contributes in electron emission.
Furthermore, knowing the work function of the SrRuO3 is 5.2eV, I calculated the field
enhancement factor of the emitter without dielectric layer as 370 related to the roughness
through epitaxial growth. Assuming same enhancement factor, the effective work function
56

of the emitter with LaAlO3 becomes 3.6 eV. This is the energy level that hot electron
injected from heavily doped substrate located.
The band diagram of the Au emitter, Al2O3 dielectric layer and silicon substrate is
shown in figure 3.19(a). The ALD-Al2O3 has an energy-band gap of 6.65 eV and electron
affinity of 2.58 eV
23
. The work function of the Au set to 5.1 eV and we used heavily doped
silicon substrate (doping = 1e20 cm
-3
). For these heavily doped silicon, the band bending

Figure 3.18 (a) J-E characteristics of the MIS and MS emitter using Alumina as insulator and
Au as Metal, (b) the corresponding FN plot, (c) J-E characteristics of the MIS and MS emitter
using LAO as insulator and SRO as Metal, (d) the corresponding FN plot.
(a) (b)
(c) (d)
500
400
300
200
100
0
J (µA/cm
2
)
20 15 10 5 0
E (V/µm)
SrRuO
3
+ 12 Aº LAO
SrRuO
3
+ no LAO
-24
-22
-20
-18
-16
Ln(J/E
2
) (µA/V
2
)
0.20 0.15 0.10 0.05 0.00
1/E (µm/V)
SrRuO
3
+ 12 Aº LAO
SrRuO
3
+ no LAO

500
400
300
200
100
0
J (µA/cm
2
)
16 12 8 4 0
E (V/µm)
Au + no Alumina
Au + 5 Aº Alumina
Au + 10 Aº Alumina
Au + 15 Aº Alumina

-24
-22
-20
-18
-16
-14
-12
Ln(J/E
2
) (µA/V
2
)
0.20 0.15 0.10 0.05 0.00
1/E (µm/V)
Au + no Alumina
Au + 5 Aº Alumina
Au + 10 Aº Alumina
Au + 15 Aº Alumina
57

of the conduction band is extremely small and only high energy electrons get to metal
layer after direct tunneling through dielectric layer. As such it generates the narrow and
sharp distribution of the hot electrons that have sufficient energy to contribute to emission
process.

3.3.3 Summary
In conclusion, I have investigated the I-E properties of MIS emitter with their metal-
semiconductor (MS) counterparts and noticed the threshold field shifts to lower value.
This trend was observed from the stack of 10 nm Au / 1 nm Al2O3 deposited on heavily
doped n type silicon substrate and from the stack of 10 nm SrRuO3 acting as metal on
1.2 nm LaAlO3 deposited on Nb-doped SrTiO3 substrate. For these cases, the ultrathin
dielectric layer allows sharp distribution of hot electron injection to top thin metal layer
and their contribution to electron emission shift the threshold field to lower value. The
results are highly sensitive to dielectric layer thickness given hot electron injection (direct
tunneling) from heavily doped semiconductor to metal exponentially decreases as the
dielectric thickness increases.

Figure 3.19 (a) Band diagram of the MIS emitter

Semiconductor Metal
Ef Ef
e
58

3.4 Free Space Coupled Photon Assisted Electron
Emitter
3.4.1 Developing a Setup - LED Source & Graphene Emitter
After testing the dark field emission characteristics from hybrid emitters as well as
tunneling assisted MIS emitter, I have started a set up for photon assisted electron
emission characterization. The LED source at 660 nm with nominal power of 14 mW using
400 µm diameter fiber was used. For initial free space illumination, I made a sample
consists of graphene on planar heavily doped N++ silicon and coupled the LED source in
to fiber inside vacuum chamber to illuminate on graphene. This experiment scheme is
shown in Figure 3.20. For the setup, I matched the fiber diameter (400 µm) with optical
flange as well as the numerical aperture (NA) of the flange and fibers all matched to be
0.22, however the NA of the LED source is wider, 0.39 and this mismatch result in optical
loss. The Fiber is set tilted above sample such that it shines graphene directly, however
given free space loss and small optical absorption of graphene mono-layer (2.3%), I
couldn’t observe significant changes on emission current at this optical power level using
Keithley picoammeter.
I continued with measuring the photo-current from graphene layer. I added two
electrical contacts on graphene and I used the laser pen with wavelength of 650 nm (2
eV) and nominal optical power of 50 mW working with batteries. The optical power at the
fiber end was measured 30 mW. This fiber is shining from the top as shown in Figure
3.21(c). At this level of optical power, I verified 100 nA of photo-current. Also, I verified
the photo-current by switching laser pen ON-OFF every 5 sec. The results are shown in
Figure 3.21 (b)-(d).
59

Figure 3.20 Schematic of the setup with optical source, Thorlab LED source of 385 nm and
nominal power of 23 mW. The red lines are optics and black ones are electrical line. The fiber
optics diameter is 400 µm and optical flange has a same opening. The NA of the fibers and optical
flange are 0.22 while the NA of the LED source is wider, 0.39. Thus, less optical power is coupled
and being transmitted through fiber.
Sample
Pico-ammeter
LED source
HV source
60

Given optical losses from flange and other mismatches between fibers and source
NA reduces the actual power illuminated on sample, we decided to modify the setup via
purchasing a higher power laser source as well as an electrometer with lower noise floor
for detecting photon assisted electron emission current. In the following sections, I will
use two laser sources, a fiber coupled 445 nm CW laser system (corresponding to 2.8
eV) with adjustable power up to 2W and fiber coupled 405 nm CW laser system
(corresponding to 3.06 eV) with adjustable power up to 1W. The characterization of these
laser source as well as free-space photon assisted electron emission results.

Figure 3.21 (a) Graphene with electrical contact for measuring the photo-current, (b) 100 nA
of photo-current was observed, (c) Sample and fiber optic on probe station (d) Verifying the
photo-current by switching the laser pen On-Off every 5 sec at 200mV applied bias.

-186.0
-185.5
-185.0
-184.5
-184.0
Current (µA)
2.0 1.8 1.6 1.4 1.2 1.0
Voltage (mV)
Light
Dark
367.7
367.6
367.5
367.4
367.3
367.2
Current (µA)
40 30 20 10 0
Time(s)
(a)
(b)
(c)
(d)
61

3.4.2 Developing a Setup – Laser Source Characterization
The CW laser source at 445 nm and 405 nm with tunable output power had been
used for this project. These laser sources are fiber coupled and optical power transferred
in to vacuum chamber through an optical feedthrough. The optical feedthrough is from
MPF, the multimode with a diameter of 400 µm with standard NA of 0.22 for transferring
optical power. The optical fiber for 445 nm laser source has the standard NA of 0.22 and
a diameter of 180 µm core with multiple cladding and jacket layer as described previously.
Identical optical fiber has been used inside the vacuum chamber to guide the laser
illuminating on sample. As such, the mismatch between the core size of the fiber diameter
and flange along with two FC/PC connectors results in optical loss. Figure 3.22 shows
the optical loss characterization. The loss occurs due optical flange and connectors. The
x axis indicates the current goes to laser diode. From these two characterizations, we

Figure 3.22 The optical flange loss characterization (a) 445nm cleaved fiber output power after
laser source (b) 445nm fiber output power after optical flange

40
30
20
10
0
Laser Power (mW)
0.4 0.3 0.2 0.1 0.0
Laser Display Current
Output Before Optical Flange
445 nm Laser
5
4
3
2
1
0
Laser Power (mW)
0.4 0.3 0.2 0.1 0.0
Laser Display Current
Output After Optical Flange
445 nm Laser
(a) (b)
62

found 15% to 20% of the source optical power will be transferred through optical
feedthrough to the sample.
The actual optical power illuminating on emitter sample is the power after optical
feedthrough. This power is characterized using natural density (ND) filter to protect the
optical power meter at higher output power. For this purpose, the output power at lower
power range has been measured with and without ND filter to characterize the ND filter
specification at 445 nm. The ND filter was NE10B-B, AR coated absorptive with nominal
optical density of 1 (nominal attenuation of 10X). The measurement result is shown in
figure 3.23a. We measured 13X attenuation at 445 nm wavelength using this ND filter. I
extended this measurement with ND filter to higher range and then used the attenuation
factor to calculate the actual optical power at the end of cleaved fiber after optical flange.
Figure 3.23b shows the optical power as a function of laser display current. From these
characterizations, we found the laser power corresponding to driver current. Furthermore,
we found this source can illuminate up to 250 mW on an emitter.

Figure 3.23 The 445 nm fiber coupled laser source characterization (a) ND filter
characterization at 445 nm (b) Laser source characterization for complete range.
250
200
150
100
50
0
Laser Power (mW)
2.5 2.0 1.5 1.0 0.5 0.0
Laser Display Current
With ND filter
Calculated
445 nm Laser
60
50
40
30
20
10
0
Laser Power (mW)
0.8 0.6 0.4 0.2 0.0
Laser Display Current
With ND filter
Without ND filter

445 nm Laser
(a) (b)
63

Similar procedure was performed for characterizing the 405-nm laser source. The 405-
nm fiber coupled laser source uses a bundle fiber (4 fibers) with larger diameter (total 400
µm) before the optical flange and due to this size matching between fiber and optical
feedthrough, the optical loss is smaller in comparison with 445 nm laser setup. The ND
filter demonstrated 14X attenuation at 405 nm as shown in Figure 3.24a. Then, I
characterized the laser source at higher power. The result is shown in Figure 3.24b. From
these measurements, I found 405 nm laser setup can illuminate maximum power of 390
mW on emitter sample.

Figure 3.24 (a) ND filter characterization at 405 nm (b) The optical 405 nm laser source
characterization using ND filter.
20
15
10
5
0
Laser Power (mW)
80x10
-3
60 40 20 0
Laser Display Current
Without ND filter
With ND filter
405 nm Laser
400
300
200
100
0
Laser Power (mW)
0.4 0.3 0.2 0.1 0.0
Laser Display Current
Measured (ND filter)
Calculated
405 nm Laser
(a)
(b)
64

3.4.3 LaB6 Free Space Emission Characterization
LaB6 nanoparticle has a work function as small as 2.7eV. As such, I illuminated
laser on LaB6 nanoparticle dropcasted on a heavily doped N++ silicon substrate. The
cleaved fiber is illuminating on this emitter from side. The light versus dark emission curve
is shown in Figure 3.25b. It can be seen up to 0.5 pA of emission current was achieved
using direct illumination on this sample. The current-time signal at different electric field
is shown in Figure 3.26. It can be seen the emission current follows laser pulse modulation
(ON-OFF time) and the value of the current increases slightly as the E-field between
sample and anode increases. The small current is due to small optical absorption in small
volume of LaB6 nanoparticles as well as reflection from gold contact.

Figure 3.25 Electron emission characterization form LaB6 Emitter – Dark versus light via direct
illumination of laser.
Free Space Coupling of Laser
(a) (b)
1.0
0.8
0.6
0.4
0.2
0.0
Current (pA)
3.0 2.0 1.0 0.0
E (V/µm)
Dark
Light

65

Figure 3.26 Time dependent emission current at different electric field using 405 nm laser
source.
0.4
0.3
0.2
0.1
0.0
Current (pA)
200 150 100 50 0
Time (sec)
E = 0.5 V/µm
E = 1.8 V/µm
66

3.4.4 Graphene Free Space Emission Characterization
Similarly, I characterized an emission from a free space laser coupled graphene emitter
that is transferred on a heavily doped N++ silicon substrate. The light versus dark
emission curve is shown in Figure 3.27a. It can be seen up to 1 pA of emission current
was achieved using direct illumination on this sample. The current signal at different
electric field is shown in Figure 3.27b. The emission current less than 2pA is observed
and emission current follows laser pulse as it turns ON and OFF.

Figure 3.27 Dark and light electron emission characterization form Graphene Emitter – Light is
a direct illumination of laser on emitter.
Free Space Coupling of Laser
(a) (b)
2.0
1.5
1.0
0.5
0.0
Current (pA)
3.0 2.0 1.0 0.0
E(V/µm)
Dark
Light

Figure 3.28 Time dependent emission current at different electric field.
3.0
2.5
2.0
1.5
1.0
0.5
0.0
Current (pA)
0.8 0.6 0.4 0.2 0.0
Time (min)
E = 0.8 V/µm
E = 1.3 V/µm
67

Chapter 3 References
1. Tuinstra, F., and J. Lo Koenig. The Journal of Chemical Physics 53.3 (1970): 1126-1130.
2. MP Kirley et al, J. Appl. Phys. 111 (6), 063717 (2012).
3. Tracy M Mattox, Ankit Agrawal, and Delia J Milliron, Chem. Mater. 27 (19), 6620 (2015).
4. Rudy Schlaf, (University of South Florida, 2014).
5. Fatemeh Rezaeifar et al, Journal of Vacuum Science & Technology B 35, 062202 (2017)
6. TE Stern, BS Gossling, and RH Fowler, presented at the Proc. Royal. Soc. A : Mathematical,
Physical
and Engineering Sciences, 124(795), 699-723 (1929).
7. Ralph Howard Fowler and L Nordheim, presented at the Proc. Royal. Soc. A: Mathematical,
Physical and Engineering Sciences, 1928.
8. Young-Jun Yu et al, Nano. lett. 9 (10), 3430 (2009).
9. Ryan Miller, YY Lau, and John H Booske, J. Appl. Phys. 106 (10), 104903 (2009).
10. Dattatray J Late et al, Appl. Phys. Lett. 89 (12), 123510 (2006).
11. C. A Mead, Journal of Applied Physics 32, 646 (1961).
12. J. Cohen, Journal of Applied Physics 33, 1999 (1962).
13. D. J. DiMaria et al, Phys. Rev.Lett. 57, 3213 (1986).
14. A Govyadinov, IVNC 2004.
15. Lasse B. Thomsen et al, JVST B, Nanotechnology and Microelectronics: Materials,
Processing, Measurement, and Phenomena 27, 562 (2009).
16. H-J Fitting et al, J. Phys. D: Appl. Phys. 32 (1999).
17. Walt A. de Heer et al; Progress in Surface Science, Vol 42, Issue 1-4, (1993)
18. Science 17, Vol 270, issue 5239, pp 1179-1180 (1995)
19. F. Rezaeifar et al, JVST B, Volume 35, Issue 6, (2017)
20. Vu Thien Binh and Ch. Adessi, Phys. Rev. Lett. 85, 864 (2000).
21. V. Semet et al, Appl. Phys. Lett. 84, 1937 (2004);
22. Changzhi Gu Scientific Reports volume 2, Article number: 746 (2012)
23. ML Huang, APL 89, (2006)

68

Chapter 4
Modeling of Integrated Photonic Assisted Emission
Emitter

4.1 Theoretical Calculation of Cavity Assisted
Thermionic Emission Device

Realization of on-chip, low-power, high-speed, spatially addressable electron
emission arrays would be potentially transformative for a variety of civilian and military
applications. Laser induced field emission is extensively being explored for this purpose.
However, in general, optical approaches rely on free-space coupling of an optical beam
onto electron emitters, a process that is highly inefficient, particularly when utilizing
nanostructured tips. Furthermore, free-space coupling to nanostructures places stringent
requirements on incident laser alignment and is challenging to implement if nanoscale
alignment between incident photons and arrays of millions of emission tips is required.
To overcome these challenges, in this research, we explore utilizing optical cavities to
enable nanoscale control over the spatial interaction between the photon electric field and
nanostructured electron emission tips. Integrated photonic cavities can be implemented
in various form to trap the light. After light is trapped inside the cavity, it reflects multiple
times producing standing wave for certain resonance frequencies. This standing wave
provides greater chance for light to interact with species placed close to optical cavities.
Two well-known forms of optical cavities are Fabry-Perot and ring resonator and in this
section, I utilize the properties of Fabry-perot optical cavity to design a new class of
efficient electron emitter. The enhanced optical absorption in emitter result in increasing
69

the efficiency of the electron emission. Furthermore, integration of photo-emission
devices provides easier path to design larger and complicated emission system such as
locally addressable electron emission source.

4.1.1 Cavity Assisted Thermionic Device
1

The first device I investigated is microscale optical cavities evanescently coupled to
adjacent emitters. This method enables an efficient and ultrafast optically modulated, on-
chip electron emitters. The device schematic is shown in Figure 4.1.
It consists of a micro fabricated Fabry–Perot resonator as an optical cavity, and a
heterostructure thermionic emitter with a small bandgap or metallic thermionic emitter
(e.g., LaB6) deposited on a wider bandgap electrical and thermal conductor (e.g., doped
Si). The photon with resonant wavelength will be absorbed efficiently by the small
bandgap/metallic emitter. The doped silicon serves as electrical and thermal conduction
for LaB6 emitter.

Figure 4.1 (a) Top view of cavity coupled thermionic emitter, (b) cross-sectional view of cavity
coupled thermionic emitter, and (c) Three-dimensional schematic view of cavity coupled
thermionic emitter.

70

4.1.2 Optical and Thermal Calculations

First, I consider the theoretical calculation assuming the reflectivity of the FP resonator
set to be 𝑅 𝑏 , single pass absorption in emitter, 𝐴 𝑠 . For this calculation, I assume the
intrinsic and scattering loss of the optical waveguide are negligible. I also define total
fraction of the photon injected in to cavity and absorbed by emitter, 𝐴 𝑇 and photon lifetime
inside cavity 𝜏 𝑃 . Under this condition, if we assume cavity has two identical mirrors (both
𝑅 𝑏 ), we can obtain an expression for 𝐴 𝑇 as follows:
𝐴 𝑇 =
𝐴 𝑠 1−(1−𝐴 𝑠 )𝑅 𝑏 (1)

As shown in Eq. (1), the total absorbed power is the ratio of the single pass photon
absorption in the emitter to the total photon loss per single trip. Thus, total absorption can
be optimized by either maximizing mirror reflectivity (𝑅 𝑏 −−>1) , or maximizing the single
pass absorption of the emitter (𝐴 𝑠 −−>1) .

Figure 4.2 (a) Total photon absorption in emitter, AT, as a function of single pass emitter
absorption, AS, for multiple cavity mirror reflectivity values, Rb. (b) Cavity photon lifetime as
a function of AS for varied Rb.

71

Figure 4.2(a) shows the relationship between total absorption, 𝐴 𝑇 , and single pass
absorption 𝐴 𝑠 , for increasing 𝑅 𝑏 ,. Critically, even emitters with small single pass
absorption, 𝐴 𝑠 , can exhibit near 100% total absorption by increasing the cavity mirror
reflectivity. However, increasing 𝑅 𝑏 ,comes at the cost of the cavity photon lifetime,
potentially limiting the ultimate modulation frequency of the emitter. Using optical cavity
lasing condition and considering rate of change in photon number inside cavity, Eq. (2)
can be derived for our structure that shows the relationship between the cavity photon
lifetime and the device parameters
2

1
𝜏 𝑃 =
𝑣 𝑔 𝐿 (log
1
𝑅 𝑏 +log
1
1−𝐴 𝑠 ) (2)

where vg is the group velocity of the mode and L is the cavity length. From this relation,
photon lifetime is plotted as a function of 𝐴 𝑠 for different 𝑅 𝑏 values, as shown in Figure
4.2 (b). Note that the photon lifetime in the cavity will be limited by the mirror loss to values
less than 10 ps, indicating that the factor controlling the overall time response of the
emitter will be the thermal response. To determine the thermal and current response, I
use the Richardson–Dushman equation in conjunction with a lumped thermal circuit
model for both steady-state and cooling transient responses. To convert the absorbed
photon flux to a thermal flux, I assume electrons are in equilibrium with the lattice, which
is reasonable for steady state behavior, as well as the transient behavior explored here
due to the fast carrier relaxation time,
3
which are typically on the sub picosecond
timescale, as compared to the thermal relaxation time here, which are greater than 10 ps.
The thermal and current responses are determined by the thermal mass, thermal
conductivity, and work function of the emitters. The steady state model of the system can
72

be written by assuming the dominant source of heat loss from the LaB6 emitter is
conduction through the Si fin. Then the steady-state ΔT of the LaB6 emitter as
𝛥𝑇 =
𝑃 𝐴𝑏𝑠 𝐿 𝑘𝐴
(3)

where 𝑃 𝐴𝑏𝑠 is the optical power absorbed by the LaB6, 𝜅 is the thermal conductivity of the
Si fin, A is the area of the fin in the plane of the substrate, and L is the length of the fin
from the LaB6 to the substrate. The cooling transient response of the emitter can be
written as 𝑇 (𝑡 )=(𝑇 𝑚 −𝑇 𝑏 )𝑒 −
𝑡 𝜏 𝑡 ℎ
+𝑇 𝑏 (4)

where 𝑇 𝑚 is the initial maximum temperature of the emitter, 𝑇 𝑏 is the bulk temperature,
and 𝜏 𝑡 ℎ
=
𝑚 𝐶 𝑝 𝑘 𝑒𝑚
𝐴 𝑒𝑚
, where 𝑚 is the mass, 𝐶 𝑝 is the heat capacity, and 𝑘 𝑒𝑚
𝐴 𝑒𝑚
is the thermal
conductance of the emitter. This can be rewritten as 𝜏 𝑡 ℎ
=
𝜌 𝑒𝑚
𝐶 𝑝 𝑑 𝑒𝑚
𝑘 𝑒𝑚
𝐴 𝑒𝑚
, where 𝜌 𝑒𝑚
is density
of the emitter material, and 𝑑 𝑒𝑚
is the emitter thickness, illustrating the critical role of
emitter thickness in modulation speed of the device. The current density can then be
estimated using the Richardson–Dushman equation
𝐽 =𝐴𝑏 𝑇 2
exp(−
∅
0
𝑘 𝐵 𝑇 ) (5)

with Ab = 29 A/cm
2
K
2
, A is Richardson constant, b is material factor for LaB6,
4
and ∅
0
=
2.7 eV for LaB6
5
. Later I use the current density on all the emitting surface to calculate
the total emitted current from the emitter. At this stage, three general design rules are:
(1) For practical electron emitter designs, the cavity properties may be tuned to
enable near unity photon absorption efficiency.
(2) Efficient steady state devices require using an emitter with the smallest
thermal conductance possible and are not limited by the LaB6 thermal mass.
73

(3) For ultrafast devices, both small absorber thicknesses and high emitter
thermal conductance are required.

4.1.3 Optical Response Modeling

Here, I used a 3D FDTD Maxwell equation solver to find the optical absorption
spectrum in the emitter as a function of both position and time. The optical absorption
results from the FDTD solver are then used as inputs for a 3D thermal transport
simulation, enabling to ascertain (1) the steady state relationship between optical power
injected into the cavity and emitter temperature, and (2) the transient thermal (cooling)
response of the emitter.
The simulation structure, as shown in Figure 4.1(c), has two Bragg mirrors placed
at the two ends of a 3 µm silicon waveguide to form an optical cavity. The details and
characteristics of the mirrors are shown in Figure 4.4. Silicon/LaB6 emitter of 1 µm length,
70 nm width, and 140 nm thickness is placed with edge-to-edge separation of 50 nm from
the cavity. The cavity width and height are set to 500 and 220 nm, respectively. Figures
4.3(c)-(d) show cross section and top view of the E field profiles of the cavity and emitter.
Also, figure 4.3(e) is the optical absorption profile on the emitter obtained from the
simulation. Two key features are observed: first, the presence of the evanescent coupling
to the emitter modifies the optical mode, as shown by the nonuniformity of the mode in
figure 4.3(d), illustrating the need for the full simulations to obtain accurate solutions.
Second, we can see that coherent light causes non-uniform absorption in the emitter
itself, which must be considered when determining the heating profile of the emitter.
Figure 4.3(a) shows the emitter absorption spectrum. For this specific device, the peak
absorption is around 25%, normalized to power injected into the cavity. As a reference, I
74

also plotted the absorption of a focused beam of light (free space) on the same emitter.
Here, the enhancement due to cavity is an order of magnitude and can be increased
further through device optimization. I also show the time dependence of the optical field
in the cavity [figure 4.3(b)], obtained by monitoring the field versus time at the middle of
the optical cavity. Importantly, it can be seen the photon lifetime in this configuration is
less than 500 fs.

Figure 4.3 (a) Absorption spectrum of the emitter. (b) Electric field vs time inside
the cavity illustrating photon lifetime inside cavity. (c) Cross sectional view of the
photon electric field profile in the optical cavity-emitter system. (d) Top view of the
photon electric field profile in the optical cavity emitter system. (e) Top view of
absorbed power in the emitter due to photons injected into the optical cavity.

0.30
0.25
0.20
0.15
0.10
0.05
0.00
Absorption
1.22 1.20 1.18 1.16
Wavelength (µm)
Cavity
Free Space

0.30
0.25
0.20
0.15
0.10
0.05
0.00
E-Field (V/m)
2000 1500 1000 500 0
Time (fs)
75

Figure 4.4 (a) Structure of the optical cavity and Bragg mirror. (b) Spectral properties of Bragg
mirror

1.0
0.8
0.6
0.4
0.2
0.0
1.24 1.22 1.20 1.18 1.16
Wavelength[µm]
Reflection
Transmission
Substrate Leakage
Cladding Leakage

Figure 4.5 (a) Absorption vs LaB6 emitter thickness, illustrating effect of changing the single
pass absorption, 𝐴 𝑠 , while keeping cavity properties constant. (b) Absorption vs cavity-emitter
distance, illustrating the decay of the cavity evanescent mode. (c) Absorption vs emitter length
and (d) cavity mirror reflectivity.

0.5
0.4
0.3
0.2
0.1
0.0
Absorption
3 2 1 0
Emitter Length (µm)
Emitter Thickness = 140 nm
Emitter Distance = 50nm
0.8
0.6
0.4
0.2
0.0
Absorption
300 200 100 0
Distance from Cavity (nm)
Emitter Thickness = 140 nm
Emitter Length = 1 µm
0.4
0.3
0.2
0.1
0.0
Absorption
250 200 150 100 50 0
Emitter Thickness (nm)
Emitter Length = 1 µm
Emitter Distance = 50nm
0.4
0.3
0.2
0.1
0.0
Absorption
1.0 0.8 0.6 0.4 0.2 0.0
Reflectivity
Emitter Thickness = 140 nm
Emitter Length = 1 µm
Emitter Distance = 50nm
76

To determine the sensitivity of the performance to physical device parameters, I
simulate the device while varying emitter thickness, T, emitter-cavity distance, d, emitter
length, L, and Bragg mirror reflectivity, 𝑅 𝑏 . Each of the emitter parameters explored
essentially changes the single pass absorption of the emitter, 𝐴 𝑠 . First, I explore the effect
of increasing emitter thickness as illustrated in figure 4.5(a). Quantitatively, the absorption
increases until T=150 nm and then saturates. This occurs due to the increasing overlap
between the emitter and the cavity mode with increasing thickness, corresponding to an
increase in 𝐴 𝑠 as emitter thickness increases. The shape of the mode is visible in figure.
4.3(c), essentially showing that the mode does not exist evenly across the height of the
cavity, with the bulk of the mode concentrated in the center, matching well with the
observed thickness versus absorption trend. The dependence on cavity-emitter distance
[figure. 4.5(b)] demonstrates a reduction in absorption roughly following the expected
exponential dependence of the evanescent coupling. Importantly, the absorption peak
reaches above 60% at zero emitter-cavity distance, corresponding to the emitter touching
the cavity. The absorption versus emitter length curve is shown in figure. 4.5(c). Finally,
the effect of tuning the cavity properties by adjusting the mirror reflectivity is explored,
with the results shown in figure. 4.5(d). This is achieved by tuning the Bragg reflector
design. In the regime of reflectivity’s explored here, the relation between total absorption
and reflectivity is roughly linear. Importantly, through the simulation, I observed the simple
equations are in good overall agreement with the trends obtained through detailed
modeling, suggesting that they may be used to develop general design rules for photonic
assisted electron emission devices, with computationally expensive detailed simulations
used only to fine-tune the performance.
77

4.1.4 Thermal Response Modeling

After obtaining the optical behavior, we simulate the effect of the emitter design on
the thermal response via 3D thermal simulations using COMSOL. The heat input to the
emitter was extracted from the optical simulation and imported into the thermal simulation,
details of the absorption profiles are shown in Figure. 4.6. To accurately model the thermal
properties of the device, I utilized experimentally determined temperature dependent
thermal conductivity and heat capacity values for silicon
6
as well as the interfacial thermal
resistance between LaB6 and silicon
7
The power reported here takes into consideration
conductive thermal losses, as well as blackbody radiation, and cooling due to the
Nottingham effect
8
. The power loss for each mechanism has been separately plotted for
T < 2500K and shown in Figure 4.7. The steady state thermal results are plotted for three
cases: (1) fin length L = 1 µm with bulk Si temperature dependent thermal conductivity,
(2) fin length L = 1 µm with nanostructured Si reduced thermal conductivity
9
of 1.5 W/m
K, and (3) fin length L = 100 µm with reduced thermal conductivity of kSi = 1.5 W/m.K. I
used Richardson–Dushman equation for calculating the steady state current considering
all emitting surfaces after simulating the temperature distribution of the emitter and silicon
fin device. In figure. 4.8(a), the emitter current is plotted as a function of optical power
injected into the cavity. The reduction in powers needed for thicker LaB6 is due to the
improved absorption efficiency for the thicker devices explored here. Critically, with proper
design a significant ΔT = 1700K can be achieved with 10 µW levels of injected optical
power, or 5.9 nW/K, enabled by the targeted optical absorption. Figure 4.8(b) plots emitter
current versus injected power for bulk silicon which has greater thermal conductivity and
is suitable for fast modulation applications. In Figure 4.8(c), I plot the required steady state
78

power for different emitter thickness, and as expected, thicker emitters are highly efficient
due to the larger optical absorption.
Next, I evaluated the transient thermal response of these devices. In the proposed device,
3 µm cavity and 1 µm LaB6 emitter 50 nm from the cavity, the transient absorption results
show that optical absorption from a single pulse occurs in about 1 ps, allowing to assume
that the electrons and phonons are at the same temperature. After heating, the emitter
cools due to the Si substrate, which is assumed to be a heat sink at 300 K. Figure 4.8(d)
shows both the temperature and current profiles of a single emitter with a LaB6 thickness
of 80 nm. From this calculation, the current response is significantly faster than the
thermal response, as expected from the R-D equation. Here, I plot the results for heating
to 1200K (the peak temperature is 1214 K), but the general behavior is similar for higher
temperatures as well. Figure 4.8(e) plots the required input optical pulse energy to heat
the LaB6 emitters to a temperature of around 1200K as a function of thickness. The
reduction in pulse energy with thickness is due to the improved optical absorption. We
note that for accurate charge emission calculation when pulse absorption occurs
significantly in the sub-picosecond regime, multiphoton emission will need to be
considered using the generalized Fowler–Dubridge theory. Finally, figure. 4.8(f) plots the
full width half maximum of both the temperature–time and the current–time as a function
of LaB6 thickness. Importantly, we see that for the smallest LaB6 thicknesses explored
here, temperature-time FWHMs of 10 ps are possible, and current time-FWHMs of <1 ps
are possible. Potentially enabling both ultrafast thermionic guns as well as on-chip single
electron sources.

79

Figure 4.6 Absorption profile through LaB6 emitter at different height, h = 0 nm
indicates the interface between laB6 and silicon. The absorption is stronger at interface
level as well as top surface of LaB6 interface with air.

70 nm
1 µm
1 µm
70 nm
Emitter
80

Figure 4.7 (a) Comparison between various power loss mechanisms. (b) Temperature
distribution along emitter and silicon fin depth. The bottom side of the silicon fin set to
constant room temperature and all other walls are thermally isolated.

(b)
T = 293 K
Silicon
LaB6
81

Figure 4.8 (a) Single emitter current vs optical power injected into the cavity for
nanostructured, low-thermal conductivity Si (LaB6-substrate distance, LFin¼100 lm).
Microwatts of optical power enable heating of tip by >1500 K. (b) Single emitter
current vs optical power injected into the cavity for bulk-Si, LFin¼1 lm and
temperature dependent thermal conductivity. (c) Injected optical power required to
achieve LaB6 temperature of 2000K as a function of LaB6 thickness. Both Si thermal
conductivity and LaB6-substrate distance (LFin) are varied here. (d) Transient thermal
and current response for an 80 nm thick LaB6 emitter. (e) Optical pulse energy required
to heat emitter by 1000K as a function of LaB6 emitter thickness. (f) Full-width half-
maximum of the thermal and current responses as a function of emitter thickness.
Illustrates the tradeoff between efficiency and speed for fixed cavity properties

1
10
100
1000
T
FWHM
(ps)
120 80 40 0
Emitter Thickness (nm)
0.1
1
10
100
J
FWHM
(ps)
0.0001
0.001
0.01
0.1
1
10
100
Power (mW)
160 120 80 40 0
Emitter Thickness (nm)
K(T)
Reduced K, t = 1 µm
Reduced K, t = 100 µm
T
LaB6
= 2000K
1400
1200
1000
800
600
400
Temperature (K)
0.0001 0.01 1
Time (ns)
0.8
0.6
0.4
0.2
0.0
Current (pA)
LaB
6
Thickness = 80 nm
400
350
300
250
200
Pulse Energy (pJ)
160 120 80 40 0
Emitter Thickness (nm)
3.0
2.5
2.0
1.5
1.0
0.5
0.0
Emitter Current (nA)
12 10 8 6 4 2 0
Power (µW)
Nano Si
LaB
6
Thickness
t = 80 nm
t = 100 nm
t = 120 nm
t = 140nm

12
10
8
6
4
2
0
Emitter Current (nA)
30 25 20 15 10 5 0
Power (mW)
LaB
6
Thickness
I = 80 nm
I = 100nm
I = 120nm
I = 140 nm
Bulk Si
82

4.1.5 Summary

In conclusion, in this section I explored a device platform that enables efficient and
ultrafast optical modulation of thermionic emitters by coupling an on-chip heterostructured
thermionic emitter with an optical cavity. First, I identified the critical device parameters,
using them to develop simple equations that elucidated the steady-state and transient
properties of these devices. Next, device performance was carefully evaluated through
full 3-D optical and thermal simulations using accurate geometries and materials
parameters. The full simulation results also enabled validation of the simple analytical
relations describing device performance. Importantly, it was found that through proper
device design, steady-state tip heating of T >1700 K could be achieved with less than
~10 µW of injected optical power or ~5.9 nW/K, which could potentially be reduced
through further optimization. By changing the properties of the emitter, ultra-fast
thermionic current responses <10 ps and thermal transient responses <1 ns are shown
to be possible with this platform. While only thermionic emission was explored here, the
general approach of coupling optical cavities to electron emission micro-/nano-structures
is a potentially rich field, enabling engineering of photoemission, multiphoton and multi-
color emission processes, as well as mixed mode electron emission processes.
Additionally, by designing the emission structures to support surface plasmon or surface
plasmon polariton modes, these approaches could enable efficient generation and
emission of nonequilibrium electrons. Finally, by enabling simple on-chip integration, the
approach outlined here opens the space for ultra-fast, optically-modulated electron
beams in microscale devices.
83

4.2 Theoretical Calculation of Integrated Ultra-Thin
Photoemission Device

Photoelectron emission from metallic surface has been investigated theoretically
and experimentally for many years and numerous photo-emission report has been
published with the advances in ultrashort high intensity laser pulse source. The
photoemission process is widely described by three-step Spicer’s model to obtain
expression for the quantum efficiency of the photocathode. In this model, (1) electron
absorbs the incident photon, (2) high energy excited electron travel to the surface of the
photocathode, and (3) last step is electron escape (emission) from metal-vacuum
interface in to vacuum. The conventional metallic photocathode suffers from low optical
absorption and even if electrons absorb photon, due to small e-e free flight length and
hence scattering mechanisms, their chance for getting to the surface of photocathode is
low. This mean free path for most metals are less than 5 nm and becomes even smaller
as the thickness of the metallic layer reduces. The mean free path of LaB6 film at
thickness below 50 nm is as small as 2 nm.
10
. On the other hand, the optical absorption
length of bulk metal (defined by the distance in to a material where the absorption dropped
to 1/e of the incident) is given by 𝑙 𝑜𝑝𝑡 =
𝜆 4𝜋𝑘
which depends on photon energy and material
properties of the emitter. For photon in visible range (𝜆 =445nm), the optical length can be
as large as 60 nm for LaB6 which indicates substantial portion of the excited electron will
scatter several times and will not be able to get to surface of the photocathode.
Here, I utilize the recent advances of integrated photonic specially the implementation of
integrated light source and introduce ultra-thin and nano-particle integrated photocathode
design along with their quantum efficiency calculation. Throughout FDTD optical
84

simulation and three steps Spicer model, I evaluate the quantum efficiency of this class
of emission device and describe the advantage of using integrated photocathode versus
conventional free space illuminated one.
4.2.1 Device Schematic and Optical Calculation

The proposed device schematic is shown in Figure 4.9 (b-c). In these schemes, instead
of directly illuminating photon above bulky emitter (Figure 4.9 (a)), we propose of using
integrated III-V light source coupling in to optical waveguide underneath the photocathode
and evanescently couple the photon in to emitter. We consider two cases for
photocathode, (1) ultra-thin flat LaB6 emitter, (2) LaB6 nanoparticle emitter above the
optical waveguide. For nanoparticle emitter, a thin layer of optically transparent electrode
such as ITO should be used for electrical contact to collect the current. Nanoparticle
enables larger electron escape probability because the scape cone limitation doesn’t exist
in nanoparticles and electron can get to surface from any direction. The advantages of
using nanoparticle is being described by other researchers
11
. We used LaB6 due to its
small work function (2.69 eV). This choice let us focus on wavelength range 445 nm (this
wavelength is corresponding to photon energy of 2.8 eV, larger than LaB6 work function
and same as our new laser source) which can be obtained using InGaN on chip laser
source. For nanoparticle cathode, I used thin layer of ITO that has a large transmission
(> 60%) at 445 nm. Also, Si3N4 is used to serve as optical waveguide due to its small
optical loss in the abovementioned wavelength. The flat LaB6 emitter with two thickness,
2 nm (at the order of electron mean free path) and 20 nm (much larger than mean free
path of electron) is placed directly above Si3N4 optical waveguide. Figure 4.10 shows the
absorption profile on the top surface of these flat LaB6 emitter above the optical
85

waveguide. From these plots, it can be seen most of the photon is absorbed within short
length of flat LaB6 cathode due to its large absorption coefficient at visible range. In
addition, the top view images are at the interface of emitter and vacuum as expected
thicker photocathode has smaller absorption on its top interface simply because it is far
from the waveguide – emitter interface where photons start to be absorbed evanescently.
As expected, the absorption per length for ultra-thin 2 nm emitter is smaller than the
thicker 20 nm one. However, I considered long enough length (L = 30 µm) for both cases
such that they absorb more than 95% of the input light. Later, I intend to evaluate the
effect of scattering for these two cases.

Figure 4.9 (a) Direct illumination on metallic emitter which result in large reflection
and only small fraction can be absorbed, (b) Proposed approach via utilizing integrated
optics to evanescently couple photon to thin metallic emitter (c) Proposed approach on
nano-particle emitter that enable larger scape probability.

(a) (b) (c)
86

4.2.2 Quantum Efficiency Calculation

Three steps Spicer model is widely used for calculating quantum efficiency of the
electron emission devices. The first step is absorption of a photon and excitation of an
electron. The second step is electron transport to the surface of the device, and the third
step is electron emission. A probability associated with the step (1) is calculated via:
P
1
=
𝛼 ∙∫ f(E−ℏω)𝐷 (E−ℏω)[1−f(E)𝐷 (E)]dE
∞
E
wf
∫ f(E−ℏω)𝐷 (E−ℏω)[1−f(E)𝐷 (E)]dE
∞
−∞
(6)

Figure 4.10 Top and cross-sectional view of the absorption profile in ultra-thin emitter
via integrated optics approach (photon energy = 2.7 eV) (a-b) cathode thickness = 2 nm
(c-d) cathode thickness = 20 nm

30 um
Top View
Z = 2nm
Side View
2 nm
Y
X
Y
Z
0
1
0.2
0.4
0.6
0.8
X

Top View
Z = 20nm
30 um
Side View
20 nm
0
1
0.2
0.4
0.6
0.8
X

Y
X
Y
Z
(a)
(b)
(c)
(d)
87

The density of state (DOS) for lanthanum hexaboride is shown in Figure 4.11. We
approximate the exact DOS with the function plotted in Figure 4.11 (b). Near Fermi level
it is reasonable to approximate this DOS with a rectangular pulse function. In addition,
the matrix element coupling the initial and final states are assumed to be constant over
the relevant energies. α is the probability of a photon exciting an electron. Thus, P
1

describes the probability of an electron being excited to an energy above the work function
energy. A probability associated with step (2) is approximated classically as:
P
2
= ∫
1
2
sin (θ)e
−
1−z
cos(θ)L
θ
0
(E
wf
)
0
dθ (7)
where L is the mean free length. θ
0
is the maximum angle relative to the surface normal
vector beyond which an electron of a given energy cannot escape, and z is the distance
of the excited electron from the surface of the device. Based on the form of P
1
and P
2
, any

Figure 4.11 Density of State (DOS) of LaB6 obtained from the self-consistent band
structure calculation and approximation of the DOS used for calculation

88

electron that makes it to the surface escapes. Thus, the probability associated with the
third step, P
3
, is unity. The total probability of escape is then:
P=P
1
P
2
P
3
(8)
However, this formulation assumes that the excited electron occupies a final state near
the work function energy. In calculating quantum efficiency, a more accurate approach
considers the distribution of excited electrons. An approximation for this excited electron
distribution is included in the quantum efficiency formulation below. Additionally, the
above equations assume that all electrons are excited at the same distance from the
surface. The excited electron distribution is considered with the calculation outlined
below. Furthermore, given the absorption uniformity changes with the thickness of the
emitter film, I consider the spatial distribution of the absorption acquired from FDTD
Lumerical simulation (Absorption (x, y, z)). Thus, the quantum efficiency calculation is
stated as follows:
QE=P
above_wf
∗∫ Distribution (E)
∫ ∫ ∫ ∫ Absorption (x,y,z)
θ
0
(E)
0
x
max
x
min
y
max
y
min
z
max
z
min
∙
∞
E
wf
1
2
sin (θ)e
−
1−z
cos(θ)L
dθdxdydzdE
where P
above_wf
is given by:
P
above_wf
=
∫ f(E−ℏω)[1−f(E)]dE
∞
E
wf
∫ f(E−ℏω)[1−f(E)]dE
∞
−∞
(9)

This gives the probability that an excited electron has energy above the work function
energy. Distribution(E) is given by:
Distribution (E)=
f(E−ℏω)[1−f(E)]
∫ f(E−ℏω)[1−f(E)]dE
∞
E
wf
(10)

89

Distribution (E) is a normalized distribution of excited electrons above E
wf
. This
distribution assumes constant density of states and slow varying matrix element. The
remaining terms in the quantum efficiency equation above are equivalent to those given
in P
2
which describe the probability of an excited electron reaching the surface with
sufficient energy to escape. Figure 4.12 compares the QE of ultra-thin photocathode with
thicker one. Here the length of the photocathodes is same as described before (L = 30
µm) and given high optical absorption of LaB6, more than 95% of the light can be
absorbed evanescently through this length of photocathode. The difference of these two
cases will be on number of the electron that can reach to surface of the photocathode
given the short mean free path of electron in LaB6.
From results in Figure 4.12, we understand it is possible to achieve quantum efficiency
as high as 0.5% by guiding photon energy of 3.5eV to integrated ultra-thin LaB6
photocathode. The saturating trend in flat integrated photocathode coming from the fact

Figure 4.12 Calculated Quantum efficiency of the flat LaB6 photocathode for two
different thickness.
0.8
0.6
0.4
0.2
0.0
Quantum Efficiency (%)
5.0 4.5 4.0 3.5 3.0 2.5
Photon Energy (eV)
Thin Photocathode , 2 nm
Thick Photocathode, 20 nm
Emitter Length = 30 µm
Absorption = 95%
90

that at higher energy DOS reduced significantly whereas the probability of scape and
approach to surface is increasing after absorption of high energy photon. As such, these
two-opposite factors create the saturating trend. Furthermore, as expected ultrathin
cathode provides higher quantum efficiency.

Chapter 4 Reference

1. Fatemeh Rezaeifar, JVST B, Vol 34, issue 4, 041228, (2016).
2. Shun Lien Chuang, Physics of Photonic Devices, 2
nd
Edition, Wiley (2009).
3.X. Wang, D. M. Riffe, Y.-S. Lee and M. Downer, Physical Review B 50 (11), 8016 (1994).
4. Bernhard Wolf. Handbook of ion source, CRC Press 27-28 (1995).

5. D. Goebel et al, Review of scientific instruments 56 (9), 1717-1722 (1985).
6. C. Glassbrenner et al, Physical Review 134 (4A), A1058 (1964).
7. H. Wang et al, Material Transaction. 48(9), 2349-2352 (2007).
8. F.M. Charbonnier et al, PRL.13, 397 (1964).
9. A. I. Hochbaum et al, Nature 451 (7175), 163-167 (2008).
10. Peschmann, K. R. et al, Journal of Applied Physics 44.5 (1973): 2252-2256.
11. AV Uskov, arXiv:1312.1508v1 [physics. optics] 2013.

91

Chapter 5
Demonstration of the Integrated Photonic Waveguide
Assisted Electron Emitter

State of the art fast electron emission devices rely on a free space coupling of photons
on the emitter. Despite half a century investigation, this technology suffers from low
quantum efficiency (QE) and enhancing QE is the major challenge in this field. Replacing
free space coupling technique with integrated photonics evanescent coupling enhances
the optical absorption and lead to have a power efficient electron emitter beyond the
available emitters. In this chapter, for the first time, I report the experimental
demonstration of enhanced photon assisted electron emission from a monolayer
graphene sheet placed over an integrated optical waveguide to transport the photons and
evanescently couple them to graphene electron emitter layer. This evanescent coupling
occurs through longer interaction length and photons are absorbed efficiently compared
to the limited absorption of graphene (2.29% for monolayer graphene) from free space
laser coupling. I describe and compare the measurement result of evanescent coupling
assisted electron emission with the conventional free space coupling assisted electron
emission. The effect is investigated through different material while sweeping the E-field
intensity and laser power. We conclude this technique in combination with on-chip laser
source as a solution for future generation power efficient photon-assisted electron
emission device.

92

5.1 Waveguide Assisted Electron Emission Device

5.1.1 Device Schematic and Electron Excitation
Schematics of the free space laser coupling to electron emitter is shown in Figure 5.1a.
The free space laser coupling requires stringent alignment of the focused beam on emitter
under the anode inside vacuum chamber. Another challenge is the reflection of the
illuminated beam occurs from emitter surface. To reduce these reflections, fabrication
techniques such as sub-wavelength grating shape was utilized on the metallic emitter;
however, the technique is challenging given the extreme dimension and precision it
requires. In addition, the spatial control of the electron generation is not feasible via free
space laser coupling. The illuminated area defines the electron generation profile.

Figure 5.1 (a) Free space coupled graphene emitter, (b) Evanescent coupled graphene emitter.
(b)
Graphene Emitter
(a)
Graphene Emitter
Laser
93

In contrast to free space coupling, the waveguide assisted evanescent coupling using on-
chip laser source controls the exact spatial generation of electrons without the need of
stringent optical alignment inside the vacuum chamber. The lithography process enables
fabrication of the desired waveguide shape for the precise control of electron generation
profile. Schematics of the waveguide assisted emitter is shown in Figure 5.1b. This
emitter consists of optical Si3N4 waveguide for transferring photons underneath the
electron emitter layer. A monolayer of graphene was transferred directly above the optical
Si3N4 waveguide as an electron emitter. Two Ti-Au contacts were evaporated on both
sides of the waveguide for conducting emission current to external measurement
equipment. In this prototype, the optical fiber was aligned to waveguide inside the V-
groove (U-groove) for coupling laser from the external source in to the waveguide. In the
next section, I describe the entire fabrication process of the waveguide assisted electron
emission device.
The optical absorption of graphene at ultraviolet-visible (UV-VIS) spectrum range arises
primarily from interband transitions between the valence and conduction bands as
depicted in Figure 5.2a. The interaction with light creates a distribution of optically excited
hot carriers (multiple electron-hole pairs) at elevated energy level compared to thermal
carriers at equilibrium (Figure 5.2b). These hot carriers have following pathways; first is
going through emission very fast. This step strongly depends on the strength of the E-
field. Under stronger E-field where the potential barrier width is narrower, the hot carriers
go through direct tunneling emission, whereas under lower E-field intensity, only the high
energy hot carriers go through over the barrier thermionic emission and low energy hot
carriers are unable of emission due to wide potential barrier width. The pathway for hot
94

carriers that don’t go through emission is relaxing to thermal equilibrium (Figure 5.2b).
Thermalization occurs through two competing processes: carrier-carrier scattering and
inelastic phonon scattering. The photoexcited carrier-carrier scattering ends up energy
transfer among the carriers involving the collision and in phonon emission process, the
energy of the hot carriers will be converted heat. The second scattering pathway is
responsible for the contribution of the hot carrier to thermionic emission during
thermalization. These scattering processes inside graphene are characterized by
scattering rate as fast as 10
12
- 10
14
per second and consequently the lifetime of the
generated hot electrons is ultrashort. This emphasis the importance of fast extraction of
hot electrons before thermalization. The expected current from hot electron and thermal
electron emission versus collecting E-field is shown in Figure 5.2b.

Figure 5.2 (a) Electron excitation that results in electron emission or thermalization. (b) Hot
electron and thermal electron excitations, corresponding currents.
Vacuum
ħ𝜔 E
f
- -
++
Graphene
-
- -
(a) (b)
95

5.1.2 Optical V-groove and Waveguide Fabrication

Optical waveguides fabrication procedure is summarized in Figure 5.3. First step
is adding hard mask on the back and front side of lightly p-doped (1-10 Ω.cm) silicon
wafer (100). These hard masks are added using PECVD and will protect the silicon wafer
during KOH etching, a process used for making V-groove. The hard mask on the backside
consist of 4 µm SiO2 and 0.72 µm Si3N4. For front side, I used 2 µm of SiO2 and 12 nm of
Si3N4. The front side mask should be removed later for making V-groove and therefore I
used thinner hard mask on front side. This process of hard mask fabrication is shown in
Figure 5.3(a).
Second, photoresist AZ5214 (PR) should be spin coated on both sides of the
silicon wafer. Front side PR is used for V-groove patterning while the backside PR will
protect the silicon wafer against BOE 7:1 that will be used in later steps. The spin coat
specification is 500 rpm for 5 sec and 3000 rpm for 1 min with baking at 100 C for 1 min.
The exposure energy of 80 mJ/cm
2
(4.5 mW/cm
2
for 18 sec) is used for PR AZ5214 and
result in developing time of 1 min. After developing, wafer hard baking performs at 150 C
for 30 min before wet etching of oxide using BOE 7:1. This step is necessary for protecting
V-groove walls. BOE 7:1 removed the Si3N4 and SiO2 from front side and opened a
window for V-groove fabrication. This step takes up to 11 min for the abovementioned
thickness and the result is depicted in Figure 5.3(c). In the next step, hot KOH solution is
used for making V-groove as shown in Figure 5.3(d). The SEM image of V-groove is
shown in Figure 5.4(a).
Alternatively, this step can be done using silicon deep etching (Bosch process) via
Oxford ICP to fabricate U-groove. For etching silicon deep down to 100 µm, we should
96

protect the PR during etching process. As such, we changed the flow rate of C4F8 gas to
zero during etching process and use only SF6 to etch the silicon. During deposition of
passivation, we use very small flow of C4F8, 25 sccm to keep the PR intact for longer
time. The RF power during deposition time while C4F8 flowing in to chamber reduced to
half to protect the PR even more. Under these circumstances, we can fabricate U-groove
which will be used for coupling light from fiber to waveguide. The SEM image of U-groove
is shown in Figure 5.4(b). The advantage of the U-groove is due to dry etching of silicon,
we don’t need a thick hard mask and photoresist on the back side of the wafer which is
necessary in case of wet etching of silicon using KOH solution.
After making V-groove (U-groove), a thickness of 5 µm Si3N4 deposited all over
the chip using PECVD and thick photoresist AZ4620 will be spin coated on it all over the
chip. The specification of this spin coating process is 500 rpm 15 sec and 2000 rpm for
25 sec with baking at 100 degree for 2 min. This photoresist need at least 2 hours of
resting before exposure. After resting, the waveguide pattern aligned on top of V-groove
with exposure of 450 mJ.cm2 (4.5 mW/cm
2
for 100 sec) and an additional 45 min of
resting is needed after exposure before developing. Developing process in MF-319 takes
about 4-5 min. After developing the waveguide pattern, Reactive Ion Etching process is
used to etch the Si3N4 and made waveguide as shown in Figure 5.3 (e). The SEM Images
of final waveguide and its cross-section images are shown in Figure 5.4(c)-(d)-(e)-(f).
In the next step, Ti-Au contact is added to both side of the waveguide to serve as
contact for graphene that will be transferred on top of optical waveguide later. Graphene
transfer is same as discussed in chapter 3, a wet transfer technique following with rapid
thermal annealing (RTA) at 450 ºC for 3 min (Figure 5.3(g)).
97

Figure 5.3 Optical waveguide fabrication procedure and fiber alignment.
(b)
(d)
(e)
(f)
(g) (h)
(a)
(c)
SiN
Si
SiN
Si
Photoresist
Photoresist
Si
SiO2
SiO2
Si
Waveguide
Ti-Au
Graphene Emitter
Optical
Fiber
98

Figure 5.4 (a) SEM Image of V-groove (b) U-groove, (c)-(d) Waveguide top view, (e)
Straight and tapered waveguide tilted view, (f) Waveguide cross section
Waveguide
100 µm
(b)
(c)
50 µm
10 µm
50 µm
(d)
(b)
100 µm
(a)
100 µm
(e) (f)
99

Then fiber alignment is implemented using coupling set up shown in Figure 5.5. Optical
bare fiber with diameter of 180 µm will be positioned on V-groove (Figure 5.3(h)) and
aligned properly to waveguide using microscope as it is shown in Figure 5.6. The pattern
of the output beam is then observed with IR Camera after being focused throughout the
lens placed closely to the output of the waveguide. Then epoxy is used to fix the fiber on
V-groove and required several hours IR heating lamp treatment. As it can be seen in
Figure 5.7, the waveguide is made long enough such that it can be placed under high
voltage electrode within a few mm spacing while epoxy is far enough and not exposed to
high voltage.

Figure 5.5 Optical Coupling Set up used for fiber alignment.
100

Figure 5.6 Optical Fiber Coupling Alignment under microscope and Image of the output beam

Figure 5.7 The optical image of the integrated waveguide assisted emitter before fiber
alignment, (b) A set of fabricated waveguide sample, (c) Optical fiber inside V-groove after
alignment and epoxy treatment
U-groove
Straight Waveguide
Tapered Waveguide
50 µm
Straight Waveguide
Contacts
U-grooves
V-groove
Waveguide
Optical Fiber
(a) (b)
(c)
101

5.1.3 Wet Transfer of Graphene and Annealing Process

In the next step, graphene will be transferred on top of optical waveguide with the
same method as described previously and RTA at 450C for 3 min used to remove the
PMMA residue. The optical image and Raman spectrum after graphene transfer is shown
in Figure 5.8 (on Si3N4 waveguide) and 5.9 (on SiO2 cladding layer).

Figure 5.8 Graphene transferred on Optical waveguide (a) Optical Image (b) Raman Spectrum
above SiN Waveguide indicates multiple layer of Graphene

100x10
3
80
60
40
20
0
Counts
3000 2500 2000 1500
Raman Shift (cm
-1
)
(a) (b)

Figure 5.9 Graphene transferred on optical waveguide (a) Optical Image (b) Raman Spectrum
above SiO2 cladding layer indicates single layer of Graphene
8000
6000
4000
2000
0
Counts
3000 2500 2000 1500
Raman Shift (cm
-1
)
(b) (a)
102

5.1.4 Optical Laser Coupling from Fiber to Waveguide

I described the laser power characterization at the end of the cleaved fiber after
optical feedthrough in section 3.4.2. Here, I describe the characterization of the optical
power coupled from a cleaved fiber after optical feedthrough in to integrated waveguide.
For this, I cut the waveguide after length of 5 mm and after covering the initial tapered
waveguide section as shown in figure 5.10 (a), free space optical power meter from
Thorlab was used as a power detector. Similar characterization was performed on a same
waveguide with shorter length of 1 mm and the results are plotted in Figure 5.10 (b). The
output power after 1 mm gives an estimation of the coupled power that is propagating
inside the optical waveguide beneath an electron emitter. It can be seen for 100 mW of
optical power inside fiber, 1 mW is detected at the output of waveguide that is 1% of
coupling efficiency. The difference between the optical power at the end of cleaved fiber
and the coupled power after 1 mm is due to the size mismatch of optical fiber and the
waveguide. This size mismatch result in mode mismatch and small coupling efficiency.
Assuming a linear relation for laser output after threshold current, Figure 5.10 (c) shows
the full range of optical power being coupled to waveguide as a function of power at fiber
end.
In addition, from these measurements, we have an estimation of the optical loss inside
the waveguide. This optical loss is mainly due to sidewall roughness and scattering. The
waveguide optical loss is characterized as 4 dB/mm.

103

Figure 5.10 The optical flange loss characterization (a) schematic of the optical path for 405nm
laser considering an optical flange (b) Optical power at waveguide end as a function of power
at fiber as an estimation for coupling
Waveguide
Detector
Laser Source
405 nm
Fiber
Flange
Pin
Pout
(b)
(a)
1200
1000
800
600
400
200
0
P
out
(µW)
100 80 60 40 20 0
Fiber Output (mW)
Length = 1 mm
Length = 5 mm
(c)
104

The optical power at the end of 5 mm waveguide as a function of power at fiber end was
characterized experimentally for a full range of laser source power. The result is shown
in Figure 5.11. This measurement will be useful for measuring graphene absorption after
transferring it on waveguide. Basically, the comparison of the output power before and
after graphene transfer indicates the optical absorption in graphene layer. The absorption
in graphene will be discussed in the next section.

Figure 5.11 The optical flange loss characterization (a) schematic of the optical path for 405nm
laser considering an optical flange (b) Optical power at waveguide end as a function of power
at fiber as an estimation for coupling
Waveguide
Detector
Laser Source
405 nm
Fiber
Flange
Pin
Pout
120
100
80
60
40
20
0
Waveguide End - P
out
(µW)
400 300 200 100 0
Fiber End - P
in
(mW)
Considering Flange Loss
(b)
(a)
105

5.2 Characterization of Waveguide Assisted Graphene
Emitter

5.2.1 Optical Absorption of Graphene on Optical Waveguide

Graphene’s optical properties have been intensively studied with a free-space
incident of a light, giving the single-pass absorption of 2.3%
1
. Over the past decades,
variety of structural design has been suggested to enhance the optical absorption. For
example, researchers tried to place graphene near plasmonic structure or place the
graphene inside a resonant Fabry-Perot cavity
2,3
. These structures exhibit an improved
absorption; however, they require very challenging fabrication due to the need to build
devices such as subwavelength grating with less than 5 nm precision or a multilayer
dielectric mirror on top of the graphene.
In recent years, the integrated optoelectronics device using two-dimensional
graphene sheet-based has caught attentions
4,5
and these devices require a
characterization of optical absorption of graphene above optical waveguide. As it is
mentioned above, absorption of normal incident on to a single layer graphene is 2.3%.
This method is depicted in Figure 5.12(a) from reference paper
6
. Given single layer
graphene provides an interaction length equal to the thickness of the layer (as small as 1
nm), it results in major portion of optical power transmitted through the layer. Transferring
graphene above an optical waveguide increases the interaction length as light have an
opportunity to be absorbed through longer length. In this technique, the limitation will be
a device size (waveguide length). Graphene above waveguide is exposed to evanescent
optical mode of the waveguide and absorption occurs through the waveguide length as
106

depicted in Figure 5.12(b) from reference. The overlap between the graphene and
waveguide evanescent mode is shown in Figure 5.12(c). This reference and other works
7
characterized the optical absorption of graphene at C-band above the optical waveguide
with relatively smaller geometry. For this work, I used larger waveguide and we focus on
higher energy photons at 405 nm laser source. The waveguide width is 50 µm and height
is 5 µm. I used a waveguide cut as shown in Figure 5.13(a)-(b) and measured the output
power before and after transferring 4 mm layer of graphene on it. As it can be seen from
the results (Figure 5.13(c)), after transferring 4 mm graphene above waveguide, the
output power reduced significantly indicating the absorbed power by graphene. The result
indicates 90% of optical absorption in graphene layer. This is 10 dB optical absorption
over 4 mm length of graphene sheet; in other words, 25 dB / cm or 0.025 dB / µm in visible
range. This absorption coefficient can be enhanced further via reducing the waveguide
height (5 µm) to thinner optical waveguide to generate strong evanescent coupling from
waveguide to graphene. The choice of thick waveguide result in multimode waveguide
where the distance between the peak of the mode and graphene is larger and as such
the evanescent coupling is weaker and absorption coefficient becomes smaller.

Figure 5.12 (a) The optical absorption in Graphene sheet through direct illumination, (b) The
optical absorption in Graphene sheet above optical waveguide, (c) Optical mode overlap with
graphene layer [6]

107

Figure 5.13 The (a) sample configuration – 4 mm graphene on optical waveguide, (b) The
optical output power from waveguide before and after transferring 4mm graphene sheet, (c)
Optical absorption in Graphene sheet in dB
16
12
8
4
0
Absorption (dB)
80 60 40 20 0
P
Fiber End
(mW)
4 mm Graphene sheet
1.0
0.8
0.6
0.4
0.2
0.0
Normalized Power
WG End
80 60 40 20 0
Power
Fiber End
(mW)
No Graphene
4 mm Graphene
Graphene Emitter
(a) (b)
(c) (d)
108

Using FDTD solver
8
, I investigated the optical absorption of the graphene from excited
optical modes inside the Si3N4 waveguide. Figure 5.14 shows the absorption of
fundamental and higher order mode through graphene sheet placed over Si3N4
waveguide. Here the mode number indicates number of vertical peaks of the mode for
one horizontal peak. It can be seen the absorption of graphene layer increases
significantly for higher order mode. This is because of the proximity of the peak of optical
mode to top of the waveguide where graphene layer is transferred. By comparing the
experimental results with this simulation, it can be understood the excitation of higher
order mode results in 25 dB / cm optical absorption in graphene above the thick Si3N4
waveguide. It should be noted that using on-chip laser source enable easier optical

Figure 5.14 Optical mode inside waveguide and the corresponding optical absorption to
graphene layer. Higher order mode results in larger optical absorption.

50 µm
5 µm
0.08
𝑑𝐵
𝑐𝑚
1.3
𝑑𝐵
𝑐𝑚
0.74
𝑑𝐵
𝑐𝑚
0.32
𝑑𝐵
𝑐𝑚
Absorption increases at higher mode
30
25
20
15
10
5
0
Absorption (dB/cm)
20 15 10 5 0
Mode Number
#1
#6
#11
#15
109

coupling to thinner waveguide in which lower order mode with stronger peak close to
waveguide top can be excited for efficient evanescent coupling to graphene layer.
5.2.2 Electron Emission Characterization

The field emission characteristics for emission device were measured at room
temperature under a vacuum of 5e10
-8
Torr. Photo-current detection was carried out
using a Keysight B2985A electrometer connected via triaxial cable directly to cathode for
low noise measurement. The result of the free space coupled photon assisted electron
emission from graphene emitter on a heavily doped silicon substrate was discussed in
chapter 3. As shown in Figure 3.28, up to 1.5 pA of emission current was detected under
the illumination of 390 mW. This is equivalent to 4 fA/mW efficiency.
After characterizing the free space laser coupled emission, we measured the emission
from graphene sheet on Si3N4 waveguide or a “waveguide assisted emission device”. The
result of dark and light currents is shown in Figure 5.15. For this device, we measured as
large as 40 pA of emission current with optical power of 5 mW evanescently coupled from
waveguide to graphene emitter. As such, the efficiency of electron emission from
graphene through evanescent laser coupling calculated to be 8 pA/mW. This is 2000X
improvement in the efficiency of electron emission process compared to free space laser
coupled emission device. It should be noted the photon assisted electron emission starts
at relatively small E-field from integrated emission device. The emission current up to 5
pA was detected at E-field of smaller than 1 V/µm.
The transient response (I-time signals) under different optical absorbed power has been
measured. The free space coupled photon assisted I-time signals was shown in chapter
110

3 for different E-fields. Even at higher E-field photon assisted current doesn’t exceed 1.5
pA. In contrast, the emission current increases significantly as the laser power increases
for the integrated waveguide assisted device. The plot is shown in Figure 5.15(b).

Figure 5.15 Emission Characterization of Graphene on waveguide (a) I-E characteristics
indicating the light current increases with optical power, (b) I-time signal showing the
photoemitted current pulse at different laser power (these are results from 405 nm laser source)

0.1
1
10
100
Current (pA)
4 3 2 1 0
E (V/µm)
Dark
Light

80
60
40
20
0
Current (pA)
3.0 2.5 2.0 1.5 1.0 0.5 0.0
Time (min)
0.45 mW
1.3 mW
2.25 mW
3.2 mW
3.9 mW
4.4 mW
4.8 mW
E = 3.2 V/µm
(a)
(b)
111

Figure 5.16 summarizes the emission current versus coupled laser power to
graphene at different E-field. For lower E-field, the photocurrent curve is nonlinear with
respect to optical power (fitted with 2
nd
order polynomial) indicating multi-photon emission
process of generating hot carriers for over the barrier thermionic emission. It is known
9,10

that the temperature of the hot carriers in graphene scales with laser power approximately
as T ∼ 𝑃 1/3
. The more accurate steady state relations between the input power and hot
carrier temperature is formulated by considering the emission of both optical

and acoustic
phonons. Here, emission through acoustic phonons includes both supercollision
9,12
(SC)
and normal collision (NC) cooling mechanisms. The steady state relations between the
input power and hot carrier temperature is expressed as
11
:
𝑃 =9.62
𝑔 𝑎𝑐
2
𝐷𝑂𝑆 2
(µ)𝐾 𝐵 3
ħ𝑘 𝐹 𝑙 (𝑇 3
−𝑇 0
3
)+𝜋 𝑔 𝑎𝑐
2
𝐷𝑂𝑆 2
(µ) ħ𝑘 𝐹 2
𝑠 2
𝑘 𝐵 (𝑇 −𝑇 0
)+
𝜋 𝐸 𝑜𝑝
𝑔 𝑜𝑝
2
ħ

∫ 𝑓 𝐹𝐷
(𝐸 ,µ,𝑇 )𝐷𝑂𝑆 (𝐸 )(1−𝑓 𝐹𝐷
(𝐸 −𝐸 𝑜𝑝
,µ,𝑇 ))
∞
−∞
𝐷𝑂𝑆 (𝐸 −𝐸 𝑜𝑝
) (1)

Figure 5.16 Optical mode inside waveguide and the corresponding optical absorption to
graphene layer. Higher order mode results in larger optical absorption.

112

After the electron temperature is known, we can find the thermionic current from a
monolayer of graphene using
13
:
(2)
Where β = e(kB)
3
/πħ
3
(νf)
2
=115.8 A m
2
K
-3
and ΔΦ = (e
3
F / 4πε0)
1/2
represent the reduction
in potential barrier where F is the electric field. These two relations (1) & (2) indicate the
nonlinear relation between thermionic photo-emission current and optical power at lower
E-field where potential barrier width is wide and tunneling emission is not playing a role
in electron emission. Assuming the optical absorption in graphene follows the power in
waveguide, I used the optical mode inside waveguide with one horizontal peak (as
assumed for simulation) and calculate the normalized absorption profile over 4 mm of the
graphene emitter length as shown in Figure 5.17(a). In addition, using relation (1) and
absorption profile, I calculated the electron temperature profile for accumulated absorbed
power of 4.8 mW in graphene and plotted the results in Figure 5.17(b). Due to the higher
optical absorption in graphene at the initial part of the waveguide, electron temperature
rises to 1100K inside the first 1mm length. This region has a significant role in the
thermionic emission current. Figure 5.18 shows the total expected thermionic current.
𝐽 =𝛽 𝑇 3
exp[
−(∅−𝐸 𝐹 −Δ∅)
𝐾 𝐵 𝑇 ]

Figure 5.17 (a) Optical absorption profile, (b) Temperature profile over waveguide

0 1 2 3 4
P = 5 mW
K
Length (mm)
Width (µm)
P = 5 mW
Width (µm)
0 1 2 3 4
Length (mm)
𝑃 𝑎𝑏𝑠 P = 5 mW
(a) (b)
113

Other findings from figure 5.16 is the linear relation of photo-current with respect to optical
power at higher E-field indicating single photon process generating hot electrons that
tunnel through the narrow potential barrier width. Tunneling probability as a function of E-
field (F) and potential barrier (ΦB) is calculated using WKB approximation for triangular
barrier:
𝑇 (𝐹 ,∅
𝐵 )=exp[
−2
ℎ
∫
{2𝑚 (𝑒 ∅
𝐵 −𝑒𝐹𝑥 )}
1
2
∅
𝐵 𝑒𝐹 ⁄
0
𝑑𝑥 ] =exp [
−4(2𝑚𝑒 )
1
2∅
𝐵 3/2

3ħ𝐹 ] (3)
The result is shown in figure 5.19. The electric field on x axis is the applied E-field
considering the dark field enhancement factor of 510 for our graphene emitter on
waveguide. The field enhancement factor of 510 is calculated from the slope of the
Fowler-Nordheim plot assuming the work function of the graphene is 4.57 eV. I also
considered the tunneling from
ℎ𝜔 2
above the Fermi level as well as barrier lowering due to
image charge. In Figure 5.19, it can be seen for lower E-field, the tunneling probability is

Figure 5.18 Calculated thermionic emission current in graphene

250
200
150
100
50
0
Current (pA)
10 8 6 4 2 0
Power (mW)
E = 1V/µm
E = 2V/µm
114

very small and negligible. Therefore, electrons must absorb multiple photons to be able
to overcome the barrier height and emit to vacuum (nonlinear relation at lower E-field).

For the linear region of emission current versus laser power, the generation profile can
be obtained from absorption profile assuming a single photon excite a single electron.
Given these electrons at energy level that is not sufficient for thermionic emission, the
dominant mechanism of emission is direct tunneling, and total current can be calculated
using
𝐼 =∫𝐺 (𝑥 ,𝑦 )×𝑒 ×𝑇𝑟 𝑑𝑆 (4)
Where G(x, y) is the generation rate (#/m
2
.sec) in each area element, e is the electron
charge and Tr is tunneling probability obtained previously. The linear relation between the
final photocurrent and optical power can be seen from this relation. This calculation is for
higher E-field where tunneling is the dominant mechanism of emission.

Figure 5.19 Tunneling probability is extremely small at lower E-field.

0.5
0.4
0.3
0.2
0.1
0.0
Tunneling Probability
4 3 2 1 0
E (V/µm)
115

5.3 Characterization of Waveguide Assisted LaB
6

Emitter
As described in chapter 3, I observed the improved emission characteristics such as
higher emission current and lower threshold field from lanthanum hexaboride (LaB6)
nanoparticles as an electron emitter
14,15
. This is because LaB6 nanoparticles have a low
work function as characterized in chapter 3. Here, I investigated the effect of evanescent
laser coupling in to the hybrid of low workfunction LaB6 nanoparticles on thin gold contact
above optical waveguide. For this emitter, I used 20 nm thin layer of gold as uniform
electrical contact. The schematics of the device is shown in Figure 5.20 (b). As reference,
I also characterized the emission properties of the 20 nm gold on waveguide (Figure 5.20
(a)). The dark emission properties of these two emitter is shown in Figure 5.20 (c). It can
be seen the dark current threshold field of the gold is larger ( > 7 V/µm) compared to the
dark current threshold field of the LaB6 nanoparticles on gold ( > 3.5 V/µm) . As such, the
gold dark current is not interfering (contributing in to) the emission characterization of
LaB6 nanoparticles on waveguide and they only act as an electrical contact.

116

Figure 5.20 Schematic of waveguide assisted emission device – (a) reference sample, gold
emitter. (b) LaB6 on gold emitter (c) Dark current of the reference gold emitter versus LaB6 on
gold emitter
Gold Emitter
LaB6 on Gold Emitter
(a)
(b)
(c)
0.01
0.1
1
10
100
1000
Current (pA)
8 6 4 2 0
E (V/µm)
Gold on WG
LaB
6
on Gold on WG
Dark Current
117

The light electron emission properties (I-E & I-time) from gold on waveguide is shown in
Figure 5.21. The photon assisted current doesn’t exceed 2pA for the E-field intensity
below 4 V/µm. This is due to large work function of the gold. As such, light emission from
gold contacts doesn’t interfere with photon assisted emission characterization of LaB6
nanoparticles that occurs below 4 V/µm.

Figure 5.21 Electron emission characterization from gold on waveguide (a) Light versus dark
current at 405 nm laser source, (b) Emission current-time pulse at different E-field using 405
nm laser source.
2.0
1.5
1.0
0.5
0.0
Current (pA)
40 30 20 10 0
Time (sec)
E = 1.5V/µm
E = 3.6V/µm
2
4
6
0.1
2
4
6
1
2
Current (pA)
4 3 2 1 0
E (V/µm)
Dark
Light
(a)
(b)
118

The dark and light I-E emission characteristics of LaB6 nanoparticles on gold is shown
in Figure 5.22. From these results, the photon assisted emission current is detected at E-
field as small as 0.5 V/µm. Then, the emission current increases until at sufficiently high
E-field intensity, field emission process overcome the photon assisted emission current.
This region is after 4V/ µm where the dark current starts increasing exponentially.

Figure 5.22 Dark & Light I-E characteristics of LaB6 nanoparticle emitter on waveguide.
0.01
0.1
1
10
100
Current (pA)
5 4 3 2 1 0
E (V/µm)
Dark
Light
119

The emission current map as a function of E-field and laser power is summarized in
Figure 5.23. The photon assisted emission current above 50 pA is detected from LaB6
nano particles emitter on waveguide.
The I-time signals from LaB6 emitter on waveguide at different laser power is shown in
Figure 5.24. The signals are at low E-field (1.1 V/µm) and high E-field (3.8 V/µm). The
current versus laser power shows linear relation at low and high field indicating single
photon process is dominant for both cases. The work function of the LaB6 nano particles
on thin gold has been characterized in chapter 3 and it turns out to be 3.3eV. One photon
with energy of 3.06 eV is sufficient for generating hot electrons that tunnel through the
barrier bended via small E-field.

Figure 5.23 Emission characterization current map as a function of E-field and laser power.
Current (pA)
120

Figure 5.24 I-time signal from LaB6 nanoparticle on waveguide, (a) at low E-field, (b) at high
E-field. (c) current Vs. power indicating single photon process from LaB6.
120
100
80
60
40
20
0
Current (pA)
8 6 4 2 0
Time (min)
3 mW
3.5 mW
4 mW
4.2 mW
4.4 mW
4.8 mW
E = 3.8 V/µm - High Field
40
30
20
10
0
Current (pA)
8 6 4 2 0
Time (min)
3 mW
3.5 mW
4 mW
4.2 mW
4.8 mW
E = 1.1 V/µm - Low Field
6
1
2
4
6
10
2
4
6
100
Current (pA)
1
2 3 4 5 6 7 8 9
10
Power (mW)
LaB
6
on Gold on WG - 405 nm
E = 3.8 V/µm
E = 1.1 V/µm
(a) (b)
(c)
121

5.4 Characterization of waveguide assisted LaB
6

on Graphene Emitter

Next, I investigated the electron emission characterisics of LaB6 nanoparticles drop
casted above graphene on optical waveguide. This hybrid combination of emitters verified
to have an enhanced emission properties previously (as described in chapter 3). For the
hybird emitter, I observed higher emission current and lower threshold field. Here, I used
the same hybrid combination of graphene and LaB6 nanoparticles above optical
waveguide. The schematics of the device is shown in Figure 5.25.

Figure 5.25 Waveguide assisted emission device – the emitter includes LaB6 nanaoparticles
above graphene emitter on optical waveguide.
LaB6 on Graphene Emitter
122

The light emission characteristic of LaB6 nanoparticles on graphene above
waveguide is shown in Figure 5.26. The light currents are obtained using 405 nm laser
source. I observed higher current up to 80 pA from this hybrid emitter prior to field
emission phenomenon with optical power smaller than 5 mW. This is equivalent to 17
pA/mW of electron emission efficiency. Furthermore, the photon assisted electron
emission current starts at E-field as small as 0.5 V/µm. I detected 10 pA of emission
current from this hybrid emitter at E-field smaller than 1 V/µm .

Figure 5.26 Emission characteristics of the waveguide assisted LaB6 on graphene emitter .
0.01
0.1
1
10
100
1000
Current (pA)
4 3 2 1 0
E (V/µm)
Dark
LaB
6
on Graphene on WG - 405 nm
Light - 2.7 mW Light - 3.3 mW
Light - 4 mW Light - 4.5 mW
123

The emission current map as a function of E-field and laser power is summarized
in Figure 5.27. It can be seen the photon assisted emission current above 80 pA is
detected from LaB6 nano particles emitter on waveguide.

Figure 5.27 Current map of LaB6 nanoparticle above graphene on waveguide as a function of laser
power and E-field.

Current (pA)
124

The emission current-time signals are shown in figure 5.28. These are the pulses
at low and high E-field. The current values are increasing as the laser power increases,
however for this emitter, due to fast modulation of laser light, the current doesn’t reach to
steady state as fast as laser modulation frequency. The separate test on the effect of
laser modulation on emission current shows15-20 sec is needed for transient signals to
reach to steady state as shown in Figure 5.29.

Figure 5.28 (a) Waveguide assisted emission device – the emitter includes LaB6 on gold emitter
(b) reference gold emitter without LaB6
40
30
20
10
0
Current (pA)
160 120 80 40 0
Time (sec)
E = 1.2 V/µm - Low Field
2 mW
2.6 mW
3.3 mW
4 mW
4.3 mW
4.5 mW
4.8 mW
(a) (b)
160
120
80
40
0
Current (pA)
150 100 50 0
Time (sec)
1.2 mW
2 mW
2.6 mW
3.3 mW
4 mW
4.3 mW
4.5 mW
4.8 mW
E = 3.4 V/µm - High Field

Figure 5.29 The current-time signals indicating at least 15-20 sec is needed for reaching to
steady state.
120
80
40
0
Current (pA)
120 80 40 0
Time (sec)
120
80
40
0
Current (pA)
120 80 40 0
Time (sec)
> 1 min
> 1 min
100 pA
65 pA
Ts
Ts
(a) (b)
125

5.5 Laser Modulation Effect on Emission

The 405nm laser source has the capability of being electrically modulated. The
frequency and duty cycle of the driver current pulse are modulated via signal generator
to modulate the laser pulse and I investigated the effect of these modulations on emission
current. Duty cycles of 10%-50% and 80% are tested and I observed the emission current
follows the duty cycle of the incoming laser pulse. Figure 5.30 shows the results. At pulse
frequency of 0.1Hz and duty cycle of 10% , the emission current pulse can’t reach to its
maximum value whereas duty cycle of 50% is enough to get to maximum value.

Figure 5.30 Laser pulse duty cycle modulation effect on emission current.
0.8
0.6
0.4
0.2
0.0
Current (pA)
30 20 10 0
Time (sec)
800V-100mHz -
Duty 10% - Max laser
0.8
0.6
0.4
0.2
0.0
Current (pA)
30 20 10 0
Time (sec)
800V-100mHz -
Duty 50% - Max laser
0.8
0.6
0.4
0.2
0.0
Current (pA)
30 20 10 0
Time (sec)
800V-100mHz -
Duty 80% - Max laser
(a)
(b) (c)
126

Then, I set the duty cycle to 50% and I tried various modulation frequency, 0.1Hz to 1Hz
and observed the emission currents follows the laser modulation frequency. Furthermore,
the emission current doesn’t reach to maximum and steady state value of the current at
higher laser modulation frequency. Hence the characterizations in this dissertation are
with lower modulation frequency (0.1Hz or smaller) .

Figure 5.31 The current-time signal at different modulation frequency for laser source.
Emission current doesn’t reach to maximum value at higher modulation frequency. Thus, all the
characterization in this thesis are at lower modulation frequency.
2.0
1.5
1.0
0.5
0.0
Current (pA)
30 20 10 0
Time (sec)
f = 100 mHz
2.0
1.5
1.0
0.5
0.0
Current (pA)
7 6 5 4 3 2 1 0
Time (sec)
f = 500 mHz
2.0
1.5
1.0
0.5
0.0
Current (pA)
16 12 8 4 0
Time (sec)
f = 200 mHz
2.0
1.5
1.0
0.5
0.0
Current (pA)
4 3 2 1 0
Time (sec)
f = 1000 mHz
(a) (b)
(c) (d)
127

5.6 Characterization of Emitter on Oxide/Nitride
Substrate

In the integrated photonics waveguide assisted electron emission devices, a monolayer
of graphene is placed above an optical waveguide made of silicon nitride that is above
cladding silicon dioxide layer. Given the thermal conductivity of the nitride/oxide are small,
there is a chance of heat trap in oxide/nitride layer that may heat up graphene and result
in thermionic emission current. Here, I investigated this possibility to evaluate the
significance of graphene heating effect via free space coupled laser. The schematic of
the device is shown in figure 5.32 a. The gold contacts are used to direct the emission
current to external circuitry. The emission result is shown in figure 5.32 b for different E-
field. It can be seen this current doesn’t exceed 1 pA indicates the free space coupling to
graphene creates negligible heating and doesn’t generate noticeable thermionic emission
current.

Figure 5.32 The current-time signal at different modulation frequency for laser source. Given
at least 10-20 sec is required for steady state current, we decided to run the experiment at lower
modulation frequency.
Graphene Emitter
1.5
1.0
0.5
0.0
Intensity(arb)
200 150 100 50 0
Time(sec)
E = 1 V/µm
E = 2 V/µm
a b
128

Chapter 5 Reference
1. Nair, R. R. et al. Science 2008, 320, 1308
2. Furchi, M. et al. Microcavity-integrated graphene photodetector. Nano Lett. 2012, 12,
2773−2777.
3.Ferreira, A. et al. Graphene-based photodetector with two cavities. Phys. Rev. B 2012, 85,
115438.
4. Berardi Sensale-Rodriguez, Journal of Lightwave Technology, Vol 33, Issue 5, 2015.
5. Martin Mittendorff et al, ACS Photonics, 2017, 4 (2), pp 316–321
6. Li, H., Anugrah, Y., Koester, S. J. & Li, M. Optical absorption in graphene integrated on silicon
waveguides. Appl. Phys. Lett. 101 (2012)
7. Kovacevic, G.; Yamashita, S. Waveguide design parameters impact on absorption in graphene
coated silicon photonic integrated circuits. Opt. Express 2016, 24, 3584−3591.
8. https://www.lumerical.com/tcad-products/fdtd/
9. Graham et al, Nature Phys.9, 103-108 (2013).
10. Sun D et al, Nature Nanotech. 7, 114-118 (2012).
11. Q Ma et al, Nature Physics volume 12, pages 455–459 (2016)
12. Song, J. C., Reizer, M. Y. & Levitov, L. S. Disorder-assisted electron-phonon scattering and
cooling pathways in graphene. Phys Rev Lett 109, 106602 (2012).
13. Shi-Jun Liang and L.K. Ang, Physical Review Applied 3, 014002 (2015)
14. F. Rezaeifar et al, J. Vac. Sci. Technol. B, Vol 35, Issue 6, 062202 (2017).
15. F. Rezaeifar and R. Kapadia, IEEE (IVEC), 193-194 (2018).

129

Chapter 6 Conclusion and Future Work
6.1 Conclusion
In this dissertation, I described the demonstration of our proposal on the replacing free
space laser coupling with evanescent laser coupling approach as an efficient method of
laser coupling to electron emitters. The idea of using integrated photonics waveguide for
laser coupling not only provides efficient electron emitters but also allow the exact spatial
control over the generated electron beam.
Figure 6.1 compares the emission current versus laser intensity reported in this
dissertation with the previously reported value
1-4
. We observed above 50 pA of emission
current for laser intensity smaller than 10 W/cm
2
. This is significantly higher amount of
photon assisted emission current with much smaller laser intensity compared to the
previously reported value. In conclusion, the evanescent laser coupling to 2D material via
integrated photonics waveguide enable significant enhancement of electron emission
current.

Figure 6.1 Comparison of this work with other state of the art reported laser induced emission
current as a function of laser power intensity.
1-4

0.001
0.01
0.1
1
10
100
Current (pA)
10
0

10
3

10
6

10
9

10
12
Laser Intensity (W/cm
2
)
1
2
3
4
This work
130

6.2 Future Work
There are a few directions for the future of this project:
1. Implementing optical cavities:
Micron size optical cavities such as ring resonator that enable enhanced absorption
within a very small area can be fabricated. The schematics of the ring resonator with
graphene emitter and LaB6 nanoparticles is shown in Figure 6.3 (a). At resonance, the
photons will be trapped and have greater chance to interact with emitter within the small
size. In this dissertation, I used relatively optical waveguide (4mm length), however
cavities are useful in reducing device size and achieving higher current density. In
addition, these devices become very useful and beneficial for making large scale
emission device. The schematic of the array of emitters via optical cavity is shown in
Figure 6.2(b).
In this scheme, a grating coupler (1) is used for coupling photons in to waveguide. Then
a beam splitter or multi-mode interferometer device (2) can be used to split the optical

Figure 6.2 (a) Optical ring resonator with graphene and LaB6 emitter transferred above it, (b)
the array of ring resonators for achieving higher current
(a) (b)
1
2
3
131

power in to several branches that each include the optical ring resonator cavity (3) for
evanescent coupling in to electron emitter which is transferred above them (Figure 6.2a).
Figure 6.3 shows the design of the optical ring resonator with quality factor of 7000 at
photon energy close to 3 eV. The exact resonance wavelength can be affected by
process variation and the characterization of this device requires tunable laser source.

Figure 6.3 Optical ring resonator design at visible wavelength, (a) dimension of the single mode
optical waveguide used for the ring design, (b) The design parameter for ring resonator, (c)
Power inside ring resonator at resonance, (d) Ring resonator spectral response
-8
-6
-4
-2
0
2
Transmission (dB)
420.0 419.0 418.0 417.0
Wavelength (nm)
250 nm
200 nm
Lc = 4 μm
R = 6 μm
gap = 100 nm
Ring on resonance
(a) (b)
(c) (d)
132

Figure 6.4 is the design of grating couplers for the wavelength range close to 3eV. These
grating along with ring resonator can be used for coupling photons and ring resonator
characterization.

Figure 6.4 Grating coupler design at visible wavelength, (a) The dimension of the grating
couplers, (b) Power coupling from fiber to waveguide via grating coupler, (c) Grating coupler
spectral response
Fiber (22 degree)
To Waveguide
Grating Coupler
200 nm
387 nm
595 nm
SiN Waveguide
Duty cycle = 0.41
40
30
20
10
0
Coupling (%)
430 420 410 400
Wavelength (nm)
(a)
(b)
(c)
133

2. Photon energy dependency and emission speed
In this dissertation, the investigation focused on specific photon energy (close to 3.1 eV)
using a CW laser source. It is known that the photoexcited carrier density is proportional
to energy of the incoming photons
5
. In future, the tunable laser source generating from
deep UV (226 nm) to NIR enable the investigation of higher photon’s energy on electron
emission mechanism. In addition to tunable CW laser source, a new high power ultrafast
laser source will be available in the lab that enables the investigation of the transient
response of these emission devices as well.

Figure 6.5 (a) The photoexcited carrier density increases for higher photon energy, (b) The new
tunable and high-power laser source capable of generating deep UV (5.5eV high photon energy)
via harmonic generators.
a
b
134

3. Optical coupling improvement:
Integration of III-V laser along with SiN integrated photonics devices (waveguide &
cavities) to couple laser in to device efficiently (reduce the coupling loss). This will enable
us to use optical waveguide with thinner height to excite the optical mode close to
fundamental mode. The advantages will be the concentration of the optical mode peak at
the vicinity of electron emitter on top of the waveguide. Hence stronger evanescent
optical coupling enables achieving larger optical absorption coefficient. The direct result
of larger optical absorption coefficient is smaller device size in other words, enhanced
current density. Figure 6.6 shows the demonstration of hybrid chip
6
.

Chapter 6 Reference
1. Ding-Shyue Yang et al, PNAS August 24, 2010. 107 (34) 14993-14998;
2. R. Bormann et al, PRL 105, 147601 (2010)
3. Andrej Grubisic et al, Nano Lett, 2012, 12 (9), pp 4823-4829
4. Barwick et al. NEW JOURNAL OF PHYSICS 9 (2007)
5. K.J. Tielrooij et al, Nature Physics vol 9, 248–252 (2013)
6. Jocelyn Durei et al, "Realization of back-side heterogeneous hybrid III-V/Si DBR lasers for
silicon photonics," Proc. SPIE 9750, Integrated Optics: Devices, Materials, and Technologies XX,
97500O (15 March 2016)

Figure 6.6 Optical coupling in to waveguide via hybrid III-V / silicon laser approach
6

Abstract (if available)

Abstract Photon assisted ultrafast electron emission is essential for applications such as developing free electron lasers (FEL), time resolved electron microscopy and high-power RF amplifiers using vacuum electron devices (VED). With the advances of Q-switched and Ti-Sapphire ultrafast laser sources, researchers started working on the demonstration of sharp electron beam via exposing high-power laser to a sharp metallic tip and they were successful in achieving electron beam pulse down to 100 fs. However, one of the most important challenge that hasn’t been addressed is the development of a technique which enables higher quantum efficiency (QE) of the photon assisted electron emission devices. ❧ As such this dissertation has two focuses: ❧ (1) Implementing techniques that enable lower threshold dark field emitter. Specifically, I focused on using a real 2D material such as graphene with excellent electrical and optical properties and lanthanum hexaboride (LaB₆) nanoparticles as a low work function electron emitter. This hybrid electron emitter in combination with an array of silicon sharp tips that enhance the local field intensity was used to demonstrate lower threshold field. The other technique for reducing the threshold field of the dark field emitter is to add an ultra-thin layer of oxide in between metal-semiconductor (MS) junction to inject a sharp distribution of hot electron from doped substrate in to the thin top metallic emitter. The contribution of these hot electrons to field emission is equivalent to electron emission from an artificial material with lower effective work function. As such, the threshold value is expected to be smaller compared to the identical MS emitter. The experimental demonstration on two different stacks of MIS emitter is described in chapter 3. ❧ (2) Demonstrating a novel idea of introducing integrated photonics waveguide and cavities for electron emission devices to obtain enhanced QE through achieving higher optical absorption. Less than 1 pA of light electron emission current is being detected from conventional free space coupled emitters and increasing photon assisted emission current is necessary for many applications. The theoretical calculation on the emission device consist of a microfabricated optical cavity, a Fabry-Perot and a heterostructured thermionic emitter with a small bandgap or metallic thermionic emitter (e.g. LaB₆) deposited on a wider bandgap electrical and thermal conductor (e.g. doped Si) are discussed in detail in chapter 4. The significant result is through cavity assisted emitter the highly-efficient photon to thermal conversion efficiencies > 60% can be achieved despite small emitter active absorption volumes < 0.01 µm³ and moderate Q optical cavities. Through transient simulation, it is found that cavity assisted electron emitter can be designed with ultra-fast sub-ns thermal response time, and sub 10 ps current response times with < 10 µW of power required to achieve nA level current emission per tip. After theoretical evaluations, we experimentally demonstrated the integrated photonics waveguide assisted electron emitters. First, I explored the fabrication process and then investigated the effect of the replacing free space coupling with integrated waveguide assisted evanescent coupling on electron emission. Enhanced electron emission has been observed using different emitter material on waveguide including (i) graphene, (ii) lanthanum hexaboride, (iii) hybrid of lanthanum hexaboride on graphene. I measured up to 10 pA/mW of emission efficiency via evanescent coupling whereas the free space coupling results in a few fA/mW of emission efficiency. I also observed the emission process is sensitive to a 2nd order power law of the laser intensity at low E-field, which supports an emission mechanism based on multi-photons absorption followed by over-the-barrier direct thermionic emission. However, at higher E-field, the emission process is linearly proportional to laser power, indicating a single photon process followed by quantum mechanical direct tunneling of hot electron contribution to emission process. We conclude this technique in combination with on-chip laser source as a solution for future generation power efficient photon-assisted electron emission device.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Semiconductor devices for vacuum electronics, electrochemical reactions, and ultra-low power in-sensor computing

PDF

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

PDF

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

PDF

Nonlinear optical nanomaterials in integrated photonic devices

PDF

Optimized nanophotonic designs for thermal emission control

PDF

Silicon photonics integrated circuits for analog and digital optical signal processing

PDF

Silicon integrated devices for optical system applications

PDF

Application of optical forces in microphotonic systems

PDF

Integrating material growth and device physics: building blocks for cost effective emerging electronics and photonics devices

PDF

Active integrated photonic devices in single crystal LiNbO₃ micro-platelets and a hybrid Si-LiNbO₃ platform

PDF

Efficiency droop in indium gallium nitride light emitters: an introduction to photon quenching processes

PDF

III-V semiconductor heterogeneous integration platform and devices for neuromorphic computing

PDF

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

PDF

Analysis of photonic crystal double heterostructure resonant cavities

PDF

Integrated large-scale monolithic electro-optical systems in standard SOI CMOS process

PDF

Light‐assisted, templated self asssembly using photonic crystal slabs

PDF

Developing improved silica materials and devices for integrated optics applications

PDF

Development of integrated waveguide biosensors and portable optical biomaterial analysis systems

PDF

Optical studies in photonics: terahertz detection and propagation in slot waveguide

PDF

Optimizing nanoemitters using quasi-aperiodicity and photoelectrochemical etching

Asset Metadata

Creator Rezaeifar, Fatemeh (author)

Core Title Integrated photonics assisted electron emission devices

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Electrical Engineering

Publication Date 11/28/2018

Defense Date 10/05/2018

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

Tag electron emission,graphene emitter,hot electron emitter,integrated photonics,low threshold emitter,OAI-PMH Harvest,optical silicon nitride waveguide,quantum efficiency

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Kapadia, Rehan (committee chair), Cronin, Stephen (committee member), Ravichandran, Jayakanth (committee member), Wu, Wei (committee member)

Creator Email rezaeifa@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-107386

Unique identifier UC11675426

Identifier etd-RezaeifarF-6983.pdf (filename),usctheses-c89-107386 (legacy record id)

Legacy Identifier etd-RezaeifarF-6983.pdf

Dmrecord 107386

Document Type Dissertation

Format application/pdf (imt)

Rights Rezaeifar, Fatemeh

Type texts

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

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

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