Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
EFFECTIVE MODELS FOR OPTICAL PROPERTIES:A STUDY ON 1D, 2D, AND 3D MATERIALS
by
Juan Cuauhtemoc Salazar Gonzalez
A thesis submitted in conformity with the requirementsfor the degree of Doctor of Philosophy
Graduate Department of PhysicsUniversity of Toronto
c© Copyright 2018 by Juan Cuauhtemoc Salazar Gonzalez
Abstract
EFFECTIVE MODELS FOR OPTICAL PROPERTIES:
A STUDY ON 1D, 2D, AND 3D MATERIALS
Juan Cuauhtemoc Salazar Gonzalez
Doctor of Philosophy
Graduate Department of Physics
University of Toronto
2018
In this thesis I employ effective Hamiltonian models to study the electronic structure of
materials. With these models I study charge- and spin-injection induced by optical absorption
processes, and current-injection induced by quantum interference processes between different
orders of absorption. I study the optical response of narrow stripes (“nanoribbons”) of mono-
layer graphene, monolayers of tin atoms (stanene), and alloys of AlGaAs.
First I focus on graphene nanoribbons with zigzag shapes along their lengths, along which
strongly localized (“edge”) states exist at the Fermi level of undoped samples. I present re-
sults for different chemical potentials, showing that edge states are responsible for the main
contribution to the optical response at low photon energies (< 1 eV).
Next I study stanene, a monolayer of tin atoms arranged in a buckled honeycomb lattice
with Dirac-like cones in its bandstructure. The spin-orbit coupling in stanene is significant
and leads to a small band gap opening of about 90 meV. I present a scheme to extract an
effective Hamiltonian model starting from an ab-initio calculation. I keep track of the quality
of the approximations by a careful analysis of the electronic energies and the states obtained
with this effective model. Using this model I study the one- and two-photon absorption, and
spin-injection by circularly polarized light.
Finally, I investigate the optical coherent control of charge currents by two- and three-
photon absorption (“2+3”) processes in alloys of AlαGa1−αAs . An important feature of this
material is that its electronic bandstructure can be tailored to photon energies of interest. Com-
ii
pared to lower orders of interference, such as the 1+2 scheme, in 2+3 coherent control the
laser intensities required for maximal effects are larger, but the optical response is richer, the
number of optical coefficients is larger, the interference processes occur in smaller regions in
the Brillouin zone, and the electronic swarm velocities are higher.
iii
A mis padres, Rosario y Juan;
a mis hermanos, Xochitl, Francisco y Xicotencatl;
a mis tos, Teresita† y Raul†;
y con especial gratitud, a mi esposa Alenita.
To my parents, Rosario and Juan;
to my siblings, Xochitl, Francisco, and Xicotencatl;
to my aunt Teresita† and my uncle Raul†;
and with special gratitude, to my wife Alenita.
iv
Acknowledgements
My most sincere acknowledgements and gratitude go to Prof. John Sipe for his dedicated
and professional PhD supervision. I feel fortunate to have been his PhD student. His high
standards for scientific research and for writing were invaluable for my PhD project. His great
capacity and curiosity for science, his humbleness, and his innate passion to teach and supervise
students and postdocs will always be an example for my career, and a motivation to be a better
professional. I also thank his wife, Maggie Grisdale, for having invited me to their home on
multiple occasions during these years. Their warm hospitality is much appreciated. Special
thanks to my PhD Examination Committee: Professors Sajeev John, Young-June Kim, Daniel
James, and Robin Marjoribanks. Very special thanks go to my external examiner, Prof. Tami
Pereg-Barnea, for her detailed appraisal of my thesis, particularly for having done so during
her sabbatical time.
Next, I would like to thank Jin-Luo Cheng and Rodrigo A. Muniz, two former excellent
postdocs in the group, with whom I had the opportunity to collaborate and from whom I have
learned a lot. I wish you, your wives, and children the best of all in every aspect of life. I also
thank former and current members of the group with whom I had the opportunity to share time.
Special thanks to Julien Rioux for his computational training in my first year; in alphabetic
order, special thanks to Daniel Travo, Federico Gomez-Duque, Zachary Vernon, and Zaheen
Sadeq. My gratitude also goes to Steven Butterworth, Julian Comanean, and Gregory Wu, from
Physics Computing Services for their computational support and for sharing with me tricks of
trade. Thanks also to Krystyna Biel, Teresa Baptista, and Janet Blakely for their kind and
professional help on administrative matters; thanks to Lisa Fannin and Rory McKeown who
were exceptionally kind for processing my final admin work during the last week of December
2017.
I would also like to thank several institutions that supported me in one way or another
during my PhD. The University of Toronto School of Graduate Studies provided a PhD/Masters
scholarship. The Mexico’s Consejo Nacional de Ciencia y Tecnologia (CONACYT) provided
a partial stipend support during the initial years of my PhD. The supercomputing centre Scinet
v
guided me to learn and implement modern practices in scientific software development: special
thanks to the Analysts Ramses van Zon, Marcelo Ponce, and Erik Spence. The Julich Research
Centre hosted me for three weeks during the 45th IFF Spring School and the Fall 2014 workshop
on DFT codes. Special thanks here also to my parents in law, Ivana and Uli, for helping me with
accommodation during these academic stays in Julich. I thank the members of the ABINIT,
Quantum-Espresso, and Wannier90 User Forums, and the entire Free Software Community.
Along this line, I would like to thank Tonatiuh Rangel for his multiple computational advices
on the ABINIT code. Special thanks to the staff and peers at the ashtanga studio Downward
Dog Yoga Centre and to the kind members of the Mindfulness Practice Community sangha
in Roncesvalles: yoga and Zen meditation have become daily practices over the last years
that have reshaped my life for the better. Thanks also to the staff at University of Toronto’s
MacIntosh Sport Medicine Clinic for the world-class physiotherapy that allowed me to go
back to the running track after 15 years of disability.
In the personal arena, I would like to thank my entire family for their support during this
and my previous education. Indeed, I would like to dedicate this work to my parents, Rosario
Gonzalez and Juan Salazar; to my siblings Xochitl, Francisco, and Xicotencatl; to my aunt
and uncle, Teresa† and Raul† Gonzalez, who assisted my parents in raising me up; and to my
wife, Alena Drieschova: I thank all of you for your support and love, for all your material and
emotional sacrifices, and for taking care of each other in my absence.
My most fine gratitude and acknowledgements to my wife; Alenita: you have played a
crucial role in the recent years of my life and during almost my entire PhD degree. Your love,
your warm, playful, and caring company, your acute and frank comments and suggestions,
and your emotional and economic support in the last years were fundamental to complete my
PhD. Thank you also for having introduced me to mindfulness meditation, ashtanga yoga, and
South-Asian cooking. Thank you for helping me to speak English!
Special thanks to former romantic partners who besides sharing part of their lives with me
were also supportive during my studies. Specifically, thanks to Alma G.C. and Cristina Y.T. for
their economic support in the last stages of my undergrad and master’s degrees. Thank you so
vi
much Cristina Y.T. for the flight ticket to Canada to start my PhD and for the economic support
that allowed me to aid my aunt Teresa† during her final illness throughout the course of my
first year in Toronto; my aunt was like a mother for me and your opportune help will always be
present in myself.
Among the people that I had the opportunity to meet outside the Sipe group, I would like
to thank Maria To and Andres Covarrubias for their friendship and support. Andres’s example
prompted me to buy my ever first bike in Kensington market, and within a few weeks I became
a year-round daily cyclist. Having grown up in Mexico made those winter rides in Toronto
a complete adventure and my Humber River rides will certainly be a recollection of my PhD
years. I would also like to express my delight for my continued friendship with former room-
mates, Violet McCrady, Arsalan Ahmad, and Jan van der Tempel: besides your friendship and
company, I also appreciate the number of viewpoints and activities you introduced me to. My
appreciation also goes to Hazem Daoud, for his humble friendship and his help with admin
work while I was out of town. All you enriched my time in Canada and I look forward to a
lifetime friendship.
I would like to thank all the academics who gave me advice, motivation, and guidance
to start and complete my PhD. Special thanks to five Mexican scientists who motivated me
to pursue my PhD abroad, in spite of logistic restrictions and opposite suggestions I faced at
the time: my sincere thanks to Benjamin M. Fregoso (KSU), Tonatiuh Rangel (UC Berke-
ley), Salvador Venegas-Andraca (ITESM-Mexico), Marco Lanzagorta (US NRL), and Luis
Orozco (JQI-UMD); their career examples and their advices during my thought process on
grad schools played an opportune role. Thanks to Chandra Veer Singh (Materials Science,
UofT) and to Paolo Bientinesi and Edoardo DiNapoli (RWTH-Aachen) for their comments
and career advices. My special gratitude to Elizabeth C. for her frank and acute comments and
suggestions. Finally, my sincere apologies to anyone that I have unconsciously omitted.
vii
Contents
Terms and abbreviations xiii
1 Introduction 1
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 The electronic structure problem in condensed matter . . . . . . . . . . . . . . 2
1.3 Effective models: tight-binding and k · p methods . . . . . . . . . . . . . . . . 8
1.4 Effect of the dimensionality of materials . . . . . . . . . . . . . . . . . . . . . 11
1.5 Thesis overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2 Coherent control of current injection in zigzag graphene nanoribbons 15
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.2 Theoretical model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.2.1 Model Hamiltonian . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.2.2 Velocity matrix elements . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.3 Coherent injection and control . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.3.1 Framework . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.3.2 First-order absorption process . . . . . . . . . . . . . . . . . . . . . . 31
2.3.3 Second-order absorption processes . . . . . . . . . . . . . . . . . . . . 33
2.3.4 Current injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
2.4 Doping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.5 Limits of the model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.5.1 Graphene sheet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
viii
2.5.2 Zigzag nanoribbons . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.6 Summary and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
3 An Effective Model for the Electronic and Optical Properties of Stanene 55
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.2 Method for deriving effective models . . . . . . . . . . . . . . . . . . . . . . 59
3.3 Effective model for stanene . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.3.1 First-principles ground state of stanene . . . . . . . . . . . . . . . . . 61
3.3.2 Evaluation of the effective model: accuracy of the approximation . . . 66
3.3.3 Hamiltonian matrices . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.3.4 Accuracy of the energies . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.4 Linear and non-linear optical properties . . . . . . . . . . . . . . . . . . . . . 74
3.5 Tight binding model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
3.6 Summary and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
4 Coherent Control of Two- and Three-photon Absorption in AlGaAs 83
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.2 Optical injection rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.3 Electronic model of AlGaAs . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
4.4 Coherent control with two- and three-photon absorption in AlGaAs . . . . . . . 93
4.4.1 Carrier injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
4.4.2 Current injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
4.5 Summary and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
5 Conclusions 103
A Nonzero injection coefficient components of zincblende lattices 109
Bibliography 112
ix
List of Figures
1.1 Optical responses considered in this thesis . . . . . . . . . . . . . . . . . . . . 13
2.1 Lattice structure of the zigzag nanoribbons (ZZNR) . . . . . . . . . . . . . . . 19
2.2 Bandstructure of ZZNR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.3 Conventional and ERS coherent control schematics . . . . . . . . . . . . . . . 27
2.4 One-photon absorption of ZZNR . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.5 Two-photon absorption of ZZNR . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.6 Stimulated Raman Scattering of ZZNR . . . . . . . . . . . . . . . . . . . . . . 39
2.7 Current injection on ZZNR . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
2.8 Electronic swarm velocity of ZZNR . . . . . . . . . . . . . . . . . . . . . . . 43
2.9 One-photon absorption on doped ZZNR . . . . . . . . . . . . . . . . . . . . . 46
2.10 Two-photon absorption on doped ZZNR . . . . . . . . . . . . . . . . . . . . . 47
2.11 Stimulated Raman scattering on doped ZZNR . . . . . . . . . . . . . . . . . . 48
2.12 Net current injection on ZNR . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.1 Hexagonal lattice of stanene . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.2 Structural relaxation of stanene . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.3 Charge density of stanene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.4 Bandstructure of stanene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.5 Figure of merit of the singular value decomposition . . . . . . . . . . . . . . . 67
3.6 Effective-model and ab-initio bandstructures of stanene . . . . . . . . . . . . . 72
3.7 Effective-model and ab-initio band warping of stanene . . . . . . . . . . . . . 73
x
3.8 Linear absorption and optical conductivity of stanene . . . . . . . . . . . . . . 74
3.9 Spin-density injection of stanene . . . . . . . . . . . . . . . . . . . . . . . . . 76
3.10 Two-photon absorption of stanene . . . . . . . . . . . . . . . . . . . . . . . . 78
4.1 Depiction of the 2+3 QuIC . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4.2 Bandstructure of AlGaAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
4.3 Dielectric function of AlGaAs . . . . . . . . . . . . . . . . . . . . . . . . . . 93
4.4 Two-photon absorption of AlGaAs . . . . . . . . . . . . . . . . . . . . . . . . 94
4.5 Interference of 2- and 3-photon absorption of AlGaAs . . . . . . . . . . . . . . 94
4.6 Interference of 2- and 3-photon absorption of AlGaAs . . . . . . . . . . . . . . 96
4.7 Current injection by 2+3 QuIC on AlGaAs . . . . . . . . . . . . . . . . . . . . 96
4.8 Current injection by 2+3 QuIC on AlGaAs as a function of polarization angle . 97
4.9 Swarm velocity of AlGaAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
xi
List of Tables
1.1 Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.1 Velocity matrix elements of zigzag graphene nanoribbons (ZGNR) . . . . . . . 26
2.2 Onset energies for the lowest energy transitions of ZGNR. . . . . . . . . . . . 33
3.1 Parameter values for an effective model of stanene . . . . . . . . . . . . . . . . 70
xii
Terms and abbreviations
1PA One-Photon Absorption.
2PA Two-Photon Absorption.
3PA Three-Photon Absorption.
BZ Brillouin Zone.
DFT Density-Functional Theory.
DOS Density of States.
ERS Electronic Raman Scattering.
JDOS Joint Density of States.
LCAO Linear Combination of Atomic Orbitals.
LDA Local Density Approximation.
ONCVP Optimized Norm Conserving Vanderbilt Pseudopotentials.
QSHE Quantum Spin Hall Effect.
QuIC Quantum Interference Control.
SVD Singular Value Decomposition.
xiii
TBM Tight-Binding Models.
TMD Transition Metal Dichalcogenide.
VME Velocity Matrix Elements.
ZGNR Zigzag Graphene Nanoribbons.
xiv
Chapter 1
Introduction
1.1 Motivation
One of the most important thrusts of modern technology is miniaturization, as evidenced by the
introduction of the term “nanotechnology” in the late part of the past century to refer to devices
around or below 100 nanometers in size. As this miniaturization continues, technologists and
scientists are faced with new opportunities and challenges. The tinier a device becomes the
more scientists are faced with quantum phenomena such as quantum confinement, quantum
tunnelling, and novel electronic-structure properties, such as electronic dispersion relations
with Dirac cones and topological properties. Such issues represent current areas of research
for the design, control, and operation of devices at nanoscale dimensions.
Traditionally, research in semiconductor physics has focused on silicon and its alloys with
germanium, gallium-arsenide and its alloys with aluminium, and the effect of defects such
as vacancies or dopants, in these materials. The pristine versions of silicon and GaAs have
small bandgaps — 1.11 eV and 1.43 eV, respectively — and exciton binding energies on the
order of a few meV. These semiconductors are usually studied in bulk or in structures where the
electrons and holes are at most weakly confined, such as quantum wells and quantum dots. The
electrical and optical properties of these “conventional” materials are now well characterized,
and in particular, the use of phase properties of one or more short optical pulses to generate and
1
Chapter 1. Introduction 2
control carrier population and carrier currents in semiconductors – referred as coherent control–
have been extensively studied both theoretical and experimentally here at Toronto [1–9].
The subject of this thesis is theoretical and computational studies of the optical properties
and coherent control scenarios in novel semiconductors that depart from this pattern of tra-
ditional materials. Novel materials are always interesting because they offer the possibility
of finding new technical applications. Researchers are also attracted to them because novel
materials can display new phenomena, and therefore can be used to test current theoretical de-
scriptions. If they are unable to explain the observations, then these materials are also useful as
a guide to develop new approaches to understand the behaviour of matter. In the next two sec-
tions I outline the electronic structure problem in condensed matter and two of the most widely
used effective models in this field of study. Then I finalize this chapter with an overview of the
work presented in this thesis.
1.2 The electronic structure problem in condensed matter
A system of M electrons and N nuclei in a solid obey the Schrodinger equation,
i~∂
∂tΨ(X,x, t) = H(X,x, t) Ψ(X,x, t), (1.1)
where H and Ψ are the system’s Hamiltonian and wavefunction, which are functions of the
spatial and spin coordinates of all nuclei and all electrons,
X ≡X1,X2, . . . ,XN
, Xi ≡ (Ri,Σi), (1.2)
x ≡ x1, x2, . . . , xM , xi ≡ (ri, si), (1.3)
where Xi and xi are shorthand notations to refer to the spatial and spin coordinates of nuclei
and electrons. The Hamiltonian includes the kinetic (K) and potential (U) energies from all
Chapter 1. Introduction 3
electrons (elec) and all nuclei (nuc), respectively; that is,
H =[Knuc +Unuc−nuc
]+
[Kelec +Uelec−elec
]+Uelec−nuc, (1.4)
where the last term describes the coupling between electrons and nuclei. The effect of external
electromagnetic fields and relativistic corrections will be outlined below, after I describe some
simplifications to this all-electron Hamiltonian.
Ignoring time dependence and taking Mn as the nuclei masses, m as the electron mass, Zn
as the charge numbers of the nuclei, e = −|e| as the electron charge, ~ as the reduced Plank
constant, and ∇2j as the Laplacian operator indicating derivatives with respect to nucleus or
electron spatial coordinates, the explicit expressions for the kinetic energy terms are
Knuc = −
N∑n=1
~2∇2n
2Mnand Kelec = −
M∑i=1
~2∇2i
2m, (1.5a,b)
while those for the repulsive nucleus-nucleus and electron-electron Coulomb potentials are
Unuc−nuc =12
N,N∑n,ν
e2
4πε0
ZnZν|Rn − Rν|
, and Uelec−elec =12
M,M∑i, j
e2
4πε0
1∣∣∣ri − r j
∣∣∣ , (1.6a,b)
where ε0 is the permittivity of vacuum. Finally, the (attractive) Coulomb coupling term between
the electronic and nuclear parts is
Uelec−nuc = −
N∑n=1
M∑i=1
e2
4πε0
Zn
|Rn − ri|. (1.7)
Since the proton is about 1,800 times heavier than the electron, in most situations in con-
densed matter physics it is considered that the electrons respond instantaneously to changes in
the nuclei coordinates. If one is only interested in the electronic bandstructure properties (as
I am in this thesis) and not in the phonon properties, it is common to assume that the nuclei
are fixed; this assumption is known as the clamped-ion approximation. As such, Knuc can be
assumed to vanish andUnuc−nuc becomes a constant term. In practice, all coordinates of nuclei
Chapter 1. Introduction 4
become just parameters and the Schrodinger equation can be solved for each fixed set X.
Although Eq. (1.4) is substantially simplified with these assumptions, practical solutions
still require a more drastic approximation, known as the single-particle (independent-particle)
approximation [10–13] or simply as the mean-field approximation [14]. In this scheme, the
assumption is that each electron moves in an average potential field Veff(xi) created by all the
other electrons and all nuclei. Then the total Hamiltonian is the sum of one-electron Hamilto-
nians,
H(x, t) =∑
i
Hi(xi, t), Hi(xi, t) = −~2∇2
i
2m+ Veff(xi, t) (1.8a,b)
In summary, the single-particle potential in Eq. (1.8b) Veff contains all the electron-electron
and all the electron-nuclei interactions [14], which are assumed “averaged”; furthermore, Veff
has also all the symmetries of the system. Since choosing an appropriate average potential
Veff is still a hard problem, the solutions usually follow a self-consistent approach that starts
with a reasonable guess for ψi and Veff; if a first-principles approach is sought, then appropriate
exchange-correlation terms in Veff are needed; the later itself is an entire area of active research.
In general terms, externally applied fields in the system are included in the treatment as
follows [15]: externally applied electric and magnetic fields are accounted by introducing
scalar and vector potentials. If time-dependent electromagnetic fields are present (i.e., a laser
field), these can be included in the treatment by applying the minimal coupling prescription,
p → p − eA(r, t), and perturbation theory. Relativistic spin-orbit interactions are described by
the Dirac equation and are considered relevant only for medium to heavy atoms [15]. How-
ever, one usually follows a simpler approach, as I do in this thesis, by adding a Pauli term to
the single-particle Hamiltonian in Eq. (1.8b),
HSO =~
4m2c4
[∇Veff × p
]· σ (1.9)
where p is the momentum operator of a single electron, ~ is the reduced Planck constant, m is
the bare electron mass, c is the speed of light, σ is the dimensionless spin operator σ = 2~−1S,
Chapter 1. Introduction 5
expressed as a vector of Pauli matrices, σ = (σx, σy, σz). With these considerations, the single-
particle Schrodinger equation is
i~∂
∂tψ(x, t) = H(x, t) ψ(x, t), H(x, t) = −
~2∇2
2m+ Veff(x, t) + HSO, (1.10a,b)
where the particle subindex has been removed and we recall that x ≡ r, s, where r and s refer
to the position and the spin-index of the electron, respectively.
Among the fundamental interests in solving Eq. 1.10 is to determine the energy-wavevector
dispersion relation followed by electrons, i.e., the bandstructure. The electronic structure of the
system is determined by minimizing the total energy with the restriction of a normalized wave-
function. In most cases this is an intractable problem, unless the system contains a few atoms
and few electrons, on the order of ten each. However, a number of effective methods have been
developed to solve Eq. 1.10, all of them based on symmetry considerations and/or approxima-
tions that reduce the number of degrees of freedom and allow us to obtain accurate solutions.
In a crystal semiconductor, for example, the solution to Eq. 1.10 can be further simplified by
considering all the translational, rotational, and reflection symmetries of the crystal.
Among the initial steps to simplify Eq. 1.10 is the consideration that electrons in filled
orbitals are mostly localized around the nuclei, and consequently the former can be considered
as united with the later, forming ionic cores; these electrons are referred as core electrons.
The remaining electrons in unfilled orbitals are referred as valence electrons, and are the ones
involved in chemical bonding, and very importantly, are the responsible for the electrical and
optical properties of a solid. As a consequence of this consideration, we can take the sums over
electronic indices to run over the valence electrons only, with the atomic numbers Z modified
accordingly.
The aforementioned classification of the different approaches to solve Eq. 1.10 is not rigid
and modern methodologies include a mix of these. Among the fundamental, first-principles
approaches there is the density-functional theory (DFT), both in time-dependent and time-
independent frameworks. Partially due to the complexity involved in devising time-dependent
Chapter 1. Introduction 6
exchange and correlation potentials, the time-independent (static) DFT has led the progress
in ab-initio DFT methods. In this thesis time-independent DFT methods are used in Ch. 3
as a starting point to develop an effective model. As such, in the following outline of DFT I
restrict to the static case. The reader interested in the time-dependent framework can consult
Refs. [16, 17].
The time-independent DFT is based on the original ideas of Hohenberg, Kohn and Sham
[18, 19]: for every interacting electron system under the influence of an external potential Vext,
there is a local potential, the Kohn-Sham energy potential VKS, that leads to a charge density
equal to that of the interacting system. As such, the (single-particle) Kohn-Sham Equations
(one for each electron) are
HKS[n(x)
]ψi(x) = Ei ψi(x), (1.11)
HKS[n(x)
]= −~2∇2
2m+Vnuc(x) + VH
[n(x)
]+ Vxc
[n(x)
]︸ ︷︷ ︸Kohn-Sham potential
, (1.12)
where
Vnuc(x) = −
N∑n=1
Zne2
|Xn − x|, VH
[n(x)
]=
∫d3x′
e2 n(x′)|x − x′|
, n(x) =
M∑j=1
∣∣∣ψ j(x)∣∣∣2, (1.13)
are the nuclear potential, the Hartree (Coulomb) potential and the charge density due to all
the single-electron wavefunctions; the sums run over all the N nuclei and all the M electrons.
The Vxc[n(r)
]term is the exchange-correlation potential that takes into account all correlation
effects, including the Hartree-exchange terms. Designing good exchange-correlation function-
als is perhaps the most challenging part of implementing transferable1 implementations of the
DFT formalism.
1In the language of pseudopotentials and functionals for DFT, transferability refers to the capability of apseudopotential or a exchange-correlation functional to produce good physical descriptions for a (wide) range ofmaterials and physical conditions. For example, a transferable pseudopotential for carbon should provide physicalresults for an isolated carbon, a carbon nanotube, for graphene, graphite, and diamond.
Chapter 1. Introduction 7
DFT is a vast and developing theory and a further description of it falls outside the scope of
this outline. The reader interested in further details may consult the following: Giustino [20]
for an introductory and modern textbook at the undergrad/graduate level, and the two-volume
series by Martin, Reining, and Ceperly [11, 15], for an advanced research-level description.
I would like to close this Section by mentioning that another class of methods for elec-
tronic structure calculations is based on effective Hamiltonian approaches. Among these are
the tight-binding method and the k · p method (a concise description of both is found in Yu-
Cardona [14], Ch. 2). In general terms, both methods start from different initial assumptions.
Tight-binding methods start from the assumption that electrons are tightly bound to atoms. In
a solid crystal, the atomic separation is comparable to the lattice constant, neighbour electronic
wavefunctions overlap, and some electronic states become delocalized and resemble nearly-
free electron states, hence they are referred as conduction states. The remaining states remain
bounded to the atoms and constitute the core and valence states.
In the other hand, the class of k · p−like methods can be derived from a different initial
assumption, referred as the mean-field approximation, where electrons are assumed to expe-
rience the same average potential V(r). k · p-like methods are widely used in semiconductor
optics because they are based on an extrapolation in terms of (1) energy gaps at a reference
qref point in the BZ and (2) the corresponding oscillator strengths (optical matrix elements) of
the transitions between states at such qref. This bandstructure extrapolation is extended over
a region around qref; the size of which depends on the number of known states at qref. If a
sufficient large number of states at qref is employed, then the entire BZ bandstructure can be in
principle computed.
Modern state-of-the-art methodologies have evolved from both sides of this spectrum.
From the mean-field side we have modern DFT methods based on pseudopotentials and plane-
wave basis sets; from a more atomistic approach, there are modern DFT implementations that
employ basis sets of localized orbital-like functions, like Gaussian or Wannier-like functions.
With the latter one can compute electronic structure properties of materials with unit cells con-
taining hundreds of atoms, in some implementations with an accuracy referred as “planewave
Chapter 1. Introduction 8
precision” [21].
1.3 Effective models: tight-binding and k · p methods
Tight-Binding approximation: an atomistic approach
The tight-binding approximation [10, 14, 22, 23], also known as linear combination of atomic
orbitals (LCAO), starts with the ansatz that crystalline solids are build up from an assembly of
atoms located on a lattice and that electrons are tightly bound to atoms. Ignoring the spin degree
of freedom, using the single-particle approximation, and considering one atom per lattice site,
and one electron per atom, the trial wavefunction at coordinate r is expressed as
Ψk(r) =∑
j
ck, j φ(r − R j), (1.14)
where k is the crystal momentum, R j are lattice-site vectors, and φ is an atomic orbital of
appropriate “orbital character” (i.e., s, px, py, . . .). Since we assume one atom per site and
one electron per atom, then there is only one atomic orbital φ at each lattice point. The trial
wavefunction Ψk(r) satisfies
H Ψk(r) = EkΨk(r). (1.15)
Enforcing the Bloch condition Ψk(r) = ei k·r unk(r), with unk(r + R) = unk(r), the tight-binding
wavefunction transforms to
Ψk(r) =∑
j
eik·R j φ(r − R j). (1.16)
If there are two or more electrons per unit cell, then the trial wavefunction must be chosen
accordingly. Consider for example monolayer graphene, which has two atoms per unit cell, and
each atom resides on two distinct sub-lattices, A and B. Take δ` as the vectors that connect the
Chapter 1. Introduction 9
sites ` =A, B
with the underlying Bravais lattice. Since translations by δ` are not symmetry
operations, then each sublattice must be treated explicitly. The low-energy bandstructure of
graphene is well described by considering only one electron per atom, hence we can write the
tight-binding trial wavefunction as [24]
Ψk(r) = ak ψ(A)k (r) + bk ψ
(B)k (r), (1.17)
which satisfies H Ψk(r) = EkΨk(r) and where
ψ(`)k (r) =
∑j
eik·R j φ(`)(r − R j + δ`), for ` =A, B
. (1.18)
The next step is to compute the total energy, 〈Ψk| Ek |Ψk〉. Two further approximations are
usually made: (1) only “on-site” and few “nearest-neighbours” energy terms have significant
contributions to the total energy, and (2) overlap integrals are small compared to unity.
A point to stress in spite of the latter approximation, is that electrons in this model are
not confined to a given atomic site. Indeed, electrons are mobile throughout the entire crys-
tal, since electrons described by a Bloch state (e.g., in this example enforced in Eq. (1.16))
have electronic velocities given by v(k) = ~−1∂kEk. In Chapter 3 I employ the tight-binding
method to identify structural parameters in an effective model developed from a first-principles
calculation.
The k · p approximation: a continuous model
The k · p approximation [14, 25] is a continuous semi-empirical model that extrapolates the
band structure of materials from a set of known states. It is based on time-independent degen-
erate perturbation theory, and resembles a Taylor expansion up to second order terms of the
electronic energy as a function of the crystal momentum. The starting assumption is that the
independent-particle approximation holds, and that the electronic states are known at a band
extrema, located at k0. That is, the set En(k0) is known, where n is a band index. Consider a
Chapter 1. Introduction 10
non-degenerate case, and a simple Hamiltonian
H =p2
2m0+ V(r), (1.19)
that satisfies the single-particle Schrodinger equation, Hψnk(r) = Enkψnk(r). Employing the
Bloch theorem for periodic crystals, and assuming that k0 is at the origin of the Brillouin zone,
we replace ψnk(r)→ eik·r unk(r) and the Schrodinger equation takes the form [14],
[p2
2m0+~2k · p
m0+~2k2
2m0+ V(r)
]unk(r) = En(k) unk(r). (1.20)
Here p = −i~∇ is the momentum operator, and V(r) is a periodic potential, V(r) = V(r + R),
unk(r) = unk(r + R) are the periodic parts of the Bloch wavefunctions, and k is the crystal
momentum. Assuming that we know the solutionsunk0(r), En(k0)
at the origin k0 = (0, 0, 0),
Eq. (1.20) reduces to
[p2
2m0+ V(r)
]unk0(r) = En(k0) unk0(r), (1.21)
then we can take the second and third terms in Eq. (1.20) as perturbations to the Hamiltonian
in Eq. (1.21). From standard perturbation theory, we have
En(k) = En,k0 +~2k2
2m0+~2
m20
∑n′,n
∣∣∣〈unk0 |k · p |un′k0〉∣∣∣2
Enk0 − En′k0
. (1.22)
This is the basic equation of the k · p method, from which the electronic energies in a range
of k space are expressed in terms of the known energiesEn(k0)
and the momentum matrix
elements appearing in the third term. Such matrix elements are known as the optical matrix ele-
ments, and are commonly measured in optical experiments from the determination of oscillator
strengths [14]. The k ·p method is routinely used to obtain analytical expressions and effective
masses at high symmetry points [14, 25]. As clearly seen from Eq. 1.22, the precision of the
bandstructure En(k) depends on the number of basis functions unk0 . Hence, for a sufficiently
Chapter 1. Introduction 11
large number of basis functions at the expansion point k0, an accurate bandstructure En(k) can
be obtained over the entire Brillouin Zone. A flavour of the k·p method, known as the envelope
function approach [25], will be employed in Chapter 2 when I study narrow strips of mono-
layer graphene. Then in Chapter 3, starting from a first-principles calculation, I will develop
an effective model that resembles a k · p method; however, the scheme will not be restricted
to a second order dependence on the crystal wavevector k. Importantly, the scheme will use
a minimum basis set to describe states on a relevant region of the Brillouin Zone. Finally, in
Chapter 4 I employ a traditional 30-band k · p method to study the AlαGa1−αAs alloy.
1.4 Effect of the dimensionality of materials
This thesis explores the electronic and optical properties of different materials in one- two, and
three-dimensions. As it is generally known, when the size of a material is reduced such that
electronic quantum confinement takes place in one or more dimensions, then the electronic,
optic and many other material properties change drastically [26–28]. From the uncertainty
principle (∆p ∼ ~/∆x), a simple estimate in one dimension suggests that a spatial restriction
leads to an additional energy term associated with motion, ∆E = (∆p)2/(2m) = ~2/(2m(∆x)2);
we refer to this ∆E as the confinement energy [26]. When this confinement energy becomes
larger than the kinetic energy associated with the thermal motion of the particle, then we expect
electronic behaviour that depends on the length of confinement. In a simple approximation
[26],
∆x .
√~2
m kB T, (1.23)
where m is the mass of the electron, kB is the Boltzmann constant, and T is the temperature.
For an electron in a semiconductor, taking m = 0.1 m0, with m0 as the bare electron mass,
and a cold semiconductor at about 20 K, we obtain an estimate of a confinement length of
∆x . 30 nm, at which quantum confinement behaviour is expected.
Chapter 1. Introduction 12
In general terms, the main consequence of having a material with a dimensionality different
from three is that electrons and holes have a restricted motion along one or more dimensions.
Consequently, the functional form of the density of states (DOS) as a function of the electronic
energy is modified. For instance, charge carriers in a 3D material are free to move in any direc-
tion, hence quantum confinement is absent. Within a parabolic band approximation, the DOS
varies as (E − Egap)1/2. For a 2D material, quantum confinement occurs along one dimension,
and the DOS has a step-wise variation; for a 1D material, confinement occurs in two dimen-
sions, and the DOS varies as (E − Egap)−1/2; finally, for a 0D material (e.g., a quantum dot)
confinement occurs in all three dimensions, and the DOS has the form of a comb.
1.5 Thesis overview
In this thesis I investigate basic electronic and optical properties, as well as optical coherent
control scenarios, of semiconductor materials of different dimensions. All calculations assume
a low saturation regime of excited carriers2 and electronic properties are described within the
single-particle approximation, hence many-body effects are neglected; light-matter interactions
are described with a Fermi golden rule approach.
In Chapter 2 I start by investigating narrow strips of monolayer graphene, commonly known
as graphene nanoribbons. Since the periodic part extends along a single direction, this is con-
sidered a 1D material; hence its density of states displays a rich structure that offers the possi-
bility of having an optical response that varies by orders of magnitude within a small range of
photon energy. Moreover, this material possesses localized states extremely close to the Fermi
level, which are easily controlled by external potentials or adsorbants.
Then in Chapter 3 I investigate the electronic and spin properties of a monolayer of tin
atoms, recently named as “stanene”. In this study I employ an effective Hamiltonian model
extracted from a first-principles calculation. Due to the atomic weight of tin atoms, signifi-
cant spin-orbit coupling is present, which leads to a small band gap opening of about 90 meV.
2Regimes of high density of carriers can be described by the Bloch-Semiconductor Equations. See for instancep. 216 of [59].
Chapter 1. Introduction 13
2hω
1PA 2PA
hω
1+2CI
2+3
hω
32hω
2hω
hωhω−hω
hω
CIERS
ERS: Electron Raman Scattering
CI: Current Injection
PA: Photon Absorption
hω
Figure 1.1: Optical responses considered in this thesis. Electronic properties are describedin the single-particle approximation (many-body effects are neglected) and the light-matterinteraction is described with a Fermi golden rule approach.
Although performing first-principles calculations is becoming routine, calculating the optical
response of materials with such a small fundamental bandgap requires significant computa-
tional expenditure, since the computation of these properties involve integrals over the Bril-
louin Zone (BZ), and the smaller the bandgap the finer the required numerical partitioning of
the BZ that is necessary to resolve the absorption onset. For this reason I propose an effective
model that is free of experimental input and is based solely on an ab-initio calculation of the
electronic dispersion relations of the material, i.e., on its electronic energy vs. crystal momen-
tum relations. Besides the computation of the optical response, I also present figures of merit
to systematically assess the range of validity of the approximations.
As the last case study, in Chapter 4 I employ a well known effective method, the k · p ap-
proximation, to study electronic carrier and current injection rates induced by optically coher-
ent control techniques on the alloy AlαGa1−αAs . It turns out that the electronic bandstructure
of this alloy is easily modified by varying the relative concentration of aluminium and gallium,
i.e., the α parameter, known as the stoichiometric value of the alloy. Hence, this is a prime
example of bandgap engineering, where I aim to study features of the optical response that
arise and display their most significant structure over certain photon energies of interest. The
Chapter 1. Introduction 14
This workSystem
Material Method Relevance 1PA 2PA 3PA CI
1DGraphene Envelope
Edge states X X RNC XRibbons Functions
2D Staneneab initio Small gap↓ SOC X X RNC P
k · p-like QSHE, TI
3D AlGaAsTraditional
Tuneable gap
k · p SOC NA X X XDevices
NA: Not ApplicableP: PlannedRNC: Relevant, but Not Considered (here)SOC: (significant) Spin-Orbit CouplingTI: Topological InsulatorQSHE: Quantum Spin Hall Effect
Table 1.1: Materials and methods studied in this thesis.
novelty on this chapter is not the method I employ, but the optical response of the material I
study: to the best of my knowledge, optical coherent control using the interference between
two- and three-photon absorption processes in AlαGa1−αAs have not yet been reported. On
Figure1.1 I present a sketch of the optical processes I study in this thesis and on Table 1.1 I
outline the materials and methods employed for such purpose. Finally, in Chapter 5 I present
the Conclusions of this thesis and describe suitable lines of future work.
Chapter 2
Coherent control of current injection in
zigzag graphene nanoribbons
Abstract
I present Fermi’s golden rule calculations of the optical carrier injection and the coherent con-
trol of current injection in graphene nanoribbons with zigzag geometry, using an envelope
function approach. This system possesses strongly localized states (flat bands) with a large
joint density of states at low photon energies; for ribbons with widths above a few tens of
nanometers, this system also posses large number of (non-flat) states with maxima and minima
close to the Fermi level. Consequently, even with small dopings the occupation of these local-
ized states can be significantly altered. In this Chapter, I calculate the relevant quantities for
coherent control at different chemical potentials, showing the sensitivity of this system to the
occupation of the edge states. I consider coherent control scenarios arising from the interfer-
ence of one-photon absorption at 2~ωwith two-photon absorption at ~ω, and those arising from
the interference of one-photon absorption at ~ω with stimulated electronic Raman scattering
(virtual absorption at 2~ω followed by emission at ~ω). Although at large photon energies these
processes follow an energy-dependence similar to that of 2D graphene, the zigzag nanoribbons
exhibit a richer structure at low photon energies, arising from divergences of the joint density
15
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 16
of states and from resonant absorption processes, which can be strongly modified by doping.
As a figure of merit for the injected carrier currents, I calculate the resulting swarm velocities.
Finally, I provide estimates for the limits of validity of our model. A modified version of this
chapter was published in Physical Review B 93, 075442 (2016).
2.1 Introduction
The electronic properties of low-dimensional materials depend strongly on the size and ge-
ometry of the system [29, 30]. For instance, the bandstructure of a monolayer and a stripe
of graphene are significantly different. A stripe of graphene is usually referred as a graphene
nanoribbon, where the boundaries impose novel conditions on the wavefunctions; for a zigzag
graphene nanoribbon (ZGNR), the wavefunction vanishes on a single sublattice, A or B, at
each edge. As shown earlier [25,30,31], in ZGNR, there are confined states that extend across
the width of the system, incorporating states from both sublattices. There is also another class
of states strongly localized at each edge, which incorporate states from either one or the other
sublattice; these states are known as edge states. Although confined states are also found
in other types of ribbons, such as armchair, the edge states are present only for zigzag rib-
bons. These edge and confined states provide many of the novel characteristics seen in ZGNR.
Moreover, the energy of these states can be easily tuned by changing the ribbon width, apply-
ing external fields, and functionalizing the system [32, 33]. Since for an undoped ZGNR the
Fermi level coincides with the flat part of the edge states, tuning the doping level allows to
easily control the contribution of the edge states. Given that a 2D graphene sheet lacks these
localized states, a ZGNR offers the advantage of having optical responses that are easily tune-
able. Over the last few years, a number of studies have reported the special properties of these
localized states [25, 30, 31, 34–37] and recent investigations have described more novel prop-
erties and applications [38–43]. At zero energy they have an important role in the electronic
transport properties of both clean and disordered ZGNR, as Luck et al. [39] (and references
therein) have recently shown using a tight-binding formalism with a transfer-matrix approach.
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 17
A detailed review of these localized states in graphene-like systems can be found in Lado et
al. [44]. The optical properties of ZGNR and graphene nano-flakes have been studied from a
number of perspectives [34, 40, 45–50], always showing the strong influence of the edge states
in the dielectric function. First-principles studies of functionalization in graphene ribbons have
shown [32] that the low-energy π electrons at the edges of the ZGNR lead to higher binding
energies as compared with ribbons of different shape edges. Similar studies indicate [33] that
the optical response of functionalized ZGNR depends strongly on the size, shape and location
of the deposited molecule, suggesting functionalization as an effective way of fine-tuning the
electronic and optical properties of ZGNR.
In this Chapter, I investigate the optical injection of carriers and currents in graphene
nanoribbons by means of coherent light fields at ω and 2ω. In general, for arbitrary beams, this
technique is referred as coherent current control. It is based on the fundamental feature that
if the quantum evolution of a system can proceed via several pathways, then the interference
between such pathways can play a determining role in the final state of the system [51, 52]. In
a semiconductor, it is possible to control the injection of carriers [1, 3, 53, 54], spins, electrical
current [9], spin current [6], and even valley current [7], using phase-dependent perturbations,
usually involving coherent beams or pulses of light. In a one-color scheme, the interference is
between transition amplitudes associated with different polarizations [1]. Although carrier in-
jection can be achieved in graphene ribbons with one-color excitation, current injection cannot.
This is due to symmetry considerations, since one-color current injection is characterized by
a third-rank tensor, hence it is only allowed in systems that lack inversion symmetry [1]. Due
to the inversion symmetry in zigzag graphene ribbons, the one-color coherent control process
is forbidden. In a two-color scheme, the interference is between pathways related to photon
absorption processes arising from different phase related beams, one at ω and the other at 2ω.
In this case, current injection is characterized by a fourth-rank tensor; hence it is nonzero for a
ZGNR. In both schemes, the different pathways connect the same initial and final states. Here
our focus is on two-color current injection, and we consider two classes of processes: the first
class arises from the interference of one-photon absorption at 2~ω with two-photon absorption
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 18
at ~ω, and the second class arises from the interference of one-photon absorption at ~ω with
stimulated electronic Raman scattering at ~ω. In general, coherent control injection allows for
the placement of electrons and holes in different bands and portions of the Brillouin Zone as
ω is varied. Thus, as I will show, the current injection is very sensitive to the presence of both
confined and edge states. In line with plausible experiments, we consider nanoribbons with a
width on the order of 20 nanometers, which leads to unit cells containing a few hundreds of
atoms. For this reason, we employ an envelope function strategy to calculate the relevant en-
ergies and velocity matrix elements; the rest of the calculation follows a conventional Fermi’s
golden rule approach to calculate the absorption coefficients.
The Chapter is organized as follows. In Sec. 2.2, we describe the model Hamiltonian
employed to describe the wavefunctions, the resulting bandstructure, and the selection rules
for the velocity matrix elements. In Sec. 2.3, we describe the different carrier injection and
current injection coefficients, including the conventional and Raman contributions. In Sec. 2.4,
we revisit these calculations, but for a p-doped system. This allows us to show the significant
change in the signals that can be accomplished by altering the occupation of the edge states. In
Sec. 2.5, we provide an estimate of the limits of validity of the model employed in this Chapter.
Finally, in Sec. 2.6, we present our final discussions and conclusions.
2.2 Theoretical model
2.2.1 Model Hamiltonian
A zigzag graphene nanoribbon (ZGNR) is a strip of monolayer graphene [55, 56] that has
been cut such that the edges along its length have a zigzag shape, as shown in Fig. 2.1. We
take the ribbon to lie in the (xy) plane, with x as the longitudinal direction along which the
ribbon extends over all space; y then identifies the direction across the ribbon, along which the
electron states are confined.
We assume passivated carbon atoms at the longitudinal boundaries, as if hydrogen atoms
were adsorbed [25,40]; this allows the passivation of any dangling edge states and the neutral-
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 19
ization of the spin moments at the edges [40]. We take W = a√
3 (2N + 2)/6 as the effective
width, where N is the total number of atoms in the unit cell, a = acc√
3 = 0.246 nm is the
graphene lattice constant, and acc is the carbon-carbon distance (see Fig. 2.1). The edge at
y = a/√
3 is formed by A-atoms, while the edge at y = W − a/√
3 is formed by B-atoms. The
lattice vector is a = ax and the atomic sites are set in terms of the graphene lattice vectors,
a1 = (x −√
3y) a/2 and a2 = (−x −√
3y) a/2. The Dirac points of monolayer graphene are
projected [25] into the one-dimensional Brillouin zone of the ZGNR, [−πa ,πa ), as K = 2π
3a and
K′ = −2π3a . We express the total wavefunctions as linear combinations of atomic orbitals ϕ that
x
y
W
acca
aa1a2
, ,= A-site = B-site
. . .. . .
Figure 2.1: (Color online) Illustration of the lattice structure of a zigzag graphene nanoribbon extendedalong x and confined along y. Passivation atoms and carbon atoms are represented by unfilled and filledcircles, respectively; A (B) sites are colored red (cyan) and the unit cell is represented in grey.
are centered at atomic sites A and B,
Ψ(r) =∑RA
ψA(RA)ϕ(r −RA) +∑RB
ψB(RB)ϕ(r −RB). (2.1)
Then, following Marconcini and Macucci [25], we employ the semi-empirical k · p method
to describe Ψ(r) with a smooth envelope function approach. The coefficients ψA and ψB in
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 20
Eq. (2.1) can be written as
ψA(r) = eiK·rFKA (r) + eiK′·rFK′
A (r), (2.2a)
ψB(r) = −eiK·rFKB (r) + eiK′·rFK′
B (r), (2.2b)
where the FK(K′)A(B) (r) are the envelope function components associated with the K(K′) Dirac
point and the orbital at atom A(B)1. In writing Eq. (2.2) we have replaced ψi(Ri) → ψi(r) for
i = A, B, on the basis of two assumptions. First, we assume that atomic orbitals are strongly
localized at their corresponding atom, and second, we assume that the envelope functions are
slow-varying functions of r near the K (K′) Dirac point. These envelope functions satisfy the
Dirac equation,
0 −i∂x − ∂y 0 0
−i∂x + ∂y 0 0 0
0 0 0 −i∂x + ∂y
0 0 −i∂x − ∂y 0
×
FKA (r)
FKB (r)
FK′A (r)
FK′B (r)
=
Eγ
FKA (r)
FKB (r)
FK′A (r)
FK′B (r)
, (2.3)
where γ = (√
3/2) ta, t = 2.70 eV is the nearest-neighbor hopping parameter and vF = γ~−1
is the graphene Fermi velocity. Because of the translational symmetry along x, each envelope
function can be factorized as the product of a propagating plane wave along the length direction
(x), and a function confined along the width direction (y),
FK(r) ≡
FKA (r)
FKB (r)
= eiκx x
ΦKA (y)
ΦKB (y)
, (2.4a)
F K′(r) ≡
FK′A (r)
FK′B (r)
= eiκ′x x
ΦK′A (y)
ΦK′B (y)
, (2.4b)
where κx (κ′x) is the wavevector along the length of the ribbon, measured from the Dirac point
1The hexagonal (“honeycomb”) lattice of graphene is composed by two distinct triangular lattices, A and B.On each sub-lattice all atoms are equivalent.
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 21
K (K′). The dangling π orbitals of the carbon atoms at the edges of the ribbons are passivated
with hydrogen atoms; it is then reasonable to assume that the full wavefunction vanishes at the
lattice sites located at the effective edges. This leads to the following boundary conditions for
the confined part of the envelope functions [25],
ΦKB (y = 0) = 0, ΦK
A (y = W) = 0, (2.5a)
ΦK′
B (y = 0) = 0, ΦK′
A (y = W) = 0. (2.5b)
These boundary conditions and the block diagonal form of the matrix in Eq. (2.3) cause the
envelope functions at K to be uncoupled from their counterparts at K′; therefore, they can be
studied separately2. With the use of Eq. (2.4a), the Dirac equation for the K valley is
γ
0 κx − ∂y
κx + ∂y 0
Φ
KA (y)
ΦKB (y)
= E
ΦKA (y)
ΦKB (y)
. (2.6)
The solutions of Eq. (2.6) are of the form [25],
ΦKA (y) =
γ
E
[(κx − K)AeKy + (κx +K)Be−Ky
], (2.7)
ΦKB (y) = AeKy + Be−Ky, (2.8)
where K =√κ2
x − (E/γ)2. Under the boundary conditions (Eq. (2.5a)), this leads to a relation
between the transverse (K) and the longitudinal (κx) wavenumbers,
e−2KW =κx − K
κx +K, (2.9)
which shows that they are coupled for ZGNR. If K is taken to be real, then Eq. (2.9) reduces
to
κx = K coth (WK) , (2.10)
2This is not necessarily the case of other strip geometries, e.g., arm-chair ribbons, p. 568 of [25].
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 22
and without loss of generality we assume K to be positive. Equation (2.10) supports two
eigensolutions for κx > W−1, which we label as n = 1 for positive energies and n = −1 for
negative energies; both correspond to states strongly confined at the edges, henceforth referred
as edge states [25],
ΦKA (y) =
−2√
LAedge ζ
edgen sinh
[K edge(W − y)
], (2.11)
ΦKB (y) =
2√
LAedge sinh
[K edgey
], (2.12)
ζedgen = n, for n = ±1, (2.13)
where L is a normalization length along the x direction. We have also set K → K edge, and
Aedge is the usual wavefunction normalization coefficient,
Aedge =
√K edge/2
sinh(2K edgeW) − (2K edgeW), (2.14)
and the eigenenergy is
Eedgen = n γ
√κ2
x − (K edge)2. (2.15)
Equations (2.11)–(2.12) indicate that the edge states occupy both sublattices, and that edge
states from one sublattice are localized at one edge (e.g., for y = W, Eq. (2.11) vanishes and
Eq. (2.12) reaches its maximum; for for y = 0 the situation is reversed).
Conversely, if we consider solutions of Eq. (2.9) with K purely imaginary, of the form iKn
with Kn real, then Eq. (2.9) reduces to
κx = Kn cot (WKn) , (2.16)
where, without loss of generality, we take Kn to be positive. These solutions give states that
extend over the full width of the ribbon, and are known simply as confined states; for these
we set Kn → Kconfn and label them by n = ±1,±2,±3, . . ., starting with ±1 for those with
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 23
energies closest to zero. These confined states exist for any real κx, except those with band
index n = ±1, which exist only for κx≤W−1. Hence, the wavevector κx≤W−1 is the point of
the BZ where the edge states couple with the confined states; clearly, the wider the ribbon the
closest this coupling occurs towards the Dirac cones (κx = 0). The dispersion relations of the
confined states with band index n = ±1 connect with that of the edge states; both share the
band index n = ±1 (transition from the red to the blue traces in Fig. 2.3). The confined states
have the form
ΦKA (y) = −i
2√
LAconf
n ζconfn sin
[K conf
n (W − y)], (2.17)
ΦKB (y) = i
2√
LAconf
n sin[K conf
n y], (2.18)
ζconfn = (−1)n+1sgn(n), (2.19)
where
Aconfn =
√K conf
n /2− sin(2K conf
n W) + (2K confn W)
, (2.20)
Econfn = sgn(n) γ
√κ2
x + (K confn )2. (2.21)
We can indicate any of the edge or confined states simply by |nκx〉, where if |n| ≥ 2 the state
is confined, while if |n| = 1 then the state is confined for κx≤W−1, but it is an edge state if
κx > W−1.
Equations (2.15) and (2.21) describe the bandstructure of ZGNR, shown in Figs. 2.2 and
2.3. The edge states are flattened towards the zero energy level for κx > W−1 (Fig. 2.3), whereas
the confined states have a parabolic structure around the Dirac points, with an axis of symmetry
at κx = W−1, except for the two confined states nearest to zero energy, with band index n = ±1
and κx ≤ W−1 (Fig. 2.3). These confined states are associated with the Dirac cones of 2D
graphene. Since the extrema of the confined states occur at κx = W−1 , we can express the band
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 24
energies at such value of κx as
E±1(W−1) = ±γW−1, (2.22a)
E±n(W−1) ≈ ±γW−1√
1 + π2 (n − 1/2)2, (2.22b)
for the edge and confined states, respectively. This indicates that the band gap scales as W−1 and
provides an estimate of the photon energy at which the absorption edge occurs with respect to
the ribbon width W. It turns out that the sign functions (ζedge, ζconf) appearing in the expressions
for the wavefunctions at A-sites, ΦKA (y) [Eq. (2.11) for edge states and Eq. (2.17) for confined
states], alternate for consecutive states, being +1 for the first state above zero energy, −1 for the
next up, and so on; the situation is reversed for negative energies. These alternating sign factors
are attributed to the fact that eigenstates of the ZZGR are eigenstates of parity [25, 30, 31, 45].
This sign factor plays an important role in the selection rules of the quantities we calculate.
Therefore we indicate these sign factors on the bandstructure diagram (Figs. 2.2 and 2.3): a
solid line indicates that the confined part of an A-site component of the envelope function
has ζn = +1, whereas a dashed trace means it has ζn = −1. Fig. 2.3 is an amplification of the
bandstructure of ZZGR around the Dirac point, κx = 0. In this figure I have signaled the critical
point κx = W−1 with a vertical gray line. This is the crystal wavevector at which the edge states
couple with the confined states. In computing the bandstructure in the limit of large widths
W, I find that (1) this coupling tends towards the Dirac point and that (2) the energy difference
between contiguos energy bands decreases fast (not shown). This behaviour of the energy
states is in agreement with a simple model of a particle in a box system. Notice that both the
edge and confined states reside on both sublattices (equations (2.11)-(2.12) and (2.17)-(2.18)).
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 25
-1
0
1
−
π
a−
2π
3a 0 2π
3a
π
a
K′
K
Energy
[eV]
kx
Edge
Confined
Figure 2.2: (Color online) Zigzag nanoribbon bandstructure with 95 zigzag lines (about 20 nm width).Solid and dashed lines distinguish the polarity of the states. The confined states are shown in red andred-dashed lines, while the edge states are shown in blue and blue-dashed lines. The latter are flattenedtowards zero energy. The different polarities of these edge states is more distinguishable in the insetgiven in Fig. 2.3. The horizontal axis corresponds to the total wavevectors kx, measured from theBrillouin zone center, cf. Fig. 2.3.
2.2.2 Velocity matrix elements
We employ the envelope functions given by Eq. (2.4a) in order to calculate the velocity matrix
elements (VME) that describe the coupling between two states |n, κx〉 and |m, κx〉 as,
vnm(κx) =
∫dr
[FK(r)
]†v
[FK(r)
], (2.23)
where κx is a wavenumber and n, m are band indices. The velocity operator is given by
v = [r,H] /(i~), which, together with the Hamiltonian in Eq. (2.6) for the K valley,
H = γ
0 −i∂x − ∂y
−i∂x + ∂y 0
, (2.24)
leads to v = vF(σx, σy), where σx and σy are the Pauli matrices and vF = γ/~ is graphene’s
Fermi velocity.
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 26
Tabl
e2.
1:V
eloc
itym
atri
xel
emen
tsat
the
Dir
acpo
intK
.A
tagi
venκ x
,any
ofth
ese
mat
rix
elem
ents
are
pure
lyre
alor
pure
lyim
agin
ary
(whi
chis
expl
icitl
yin
dica
ted
byth
epr
esen
ce(a
bsen
ce)o
fthe
imag
inar
yun
iti)
.The
corr
espo
ndin
gex
pres
sion
sat
the
othe
rDir
acpo
intK
′ar
eid
entic
al,e
xcep
tth
atth
ex−
com
pone
nts
ofth
em
atri
xel
emen
tsfli
psi
gn;t
hey
-com
pone
nts
ofth
em
atri
xel
emen
tsre
mai
nun
chan
ged.
The
rang
eof
valid
ityfo
rth
isex
pres
sion
sis
give
nin
the
thir
dco
lum
n.Ty
peE
xpre
ssio
nC
ondi
tions
nCon
fvx nm
(κx)
=−
4vF
( ζconf
m+ζco
nfn
) Aco
nfn
Aco
nfm
[ Kconf msi
n(K
conf
nW
)−K
conf
nsi
n(K
conf
mW
)(K
conf
m)2−
(Kco
nfn
)2
]|n|≥
2,|m|≥
2,∀κ x
,or
l|n|≥
2,|m|=
1,κ x<
W−
1 ,or
mC
onf
vy nm(κ
x)=−
i4v F
( ζconf
m−ζco
nfn
) Aco
nfn
Aco
nfm
[ Kconf msi
n(K
conf
nW
)−K
conf
nsi
n(K
conf
mW
)(K
conf
m)2−
(Kco
nfn
)2
]|n|=
1,|m|≥
2,κ x<
W−
1
nEdg
evx nm
(κx)
=−
4vF
( ζedge
m+ζed
gen
) Aed
gen
Aed
gem
[ Kedge
nsi
nh(K
edge
mW
)−K
edge
msi
nh(K
edge
nW
)(K
edge
m)2−
(Ked
gen
)2
] |n|≥
1,|m|≥
1,κ x>
W−
1l
mE
dge
vy nm(κ
x)=−
i4v F
( ζedge
m−ζed
gen
) Aed
gen
Aed
gem
[ Kedge
nsi
nh(K
edge
mW
)−z m
sinh
(Ked
gen
W)
(Ked
gem
)2−
(Ked
gen
)2
]
nCon
fvx nm
(κx)
=i4
v F( ζed
gem
+ζco
nfn
) Aco
nfn
Aed
gem
[ Kconf nsi
nh(K
edge
mW
)−K
edge
msi
n(K
conf
nW
)
(Ked
gem
)2+
(Kco
nfn
)2
] |n|≥
2,|m|=
1,κ x>
W−
1l
mE
dge
vy nm(κ
x)=−
4vF
( ζedge
m−ζco
nfn
) Aco
nfn
Aed
gem
[ Kconf nsi
nh(K
edge
mW
)−K
edge
msi
n(K
conf
nW
)
(Ked
gem
)2+
(Kco
nfn
)2
]
nEdg
el
( Con
f↔
Edg
e) †|n|=
1,|m|≥
2,κ x>
W−
1
mC
onf
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 27
-0.4
-0.2
0
0.2
0.4
-0.2 0 W−1 0.2
µ2
µ1
K
2hω
hω
hω
m
CONV
2hω
−hω
hω
ℓ
ERS
Energy[eV]
κx [nm−1]
+1
−1
+2
−2
−3
−4
n
+3
+4
m
nℓ
Figure 2.3: (Color online) Depiction of the conventional coherent control (CC) scheme (set of arrowson the right) and the ERS CC (left arrows). Confined and edge states are shown in red and blue lines,respectively; solid and dashed lines distinguish the polarity of the states (see also Fig. 2.2). The initial(final) state is m (n) and ` is a virtual state. For m = −3, n = 2, ` = −1, the three purple dots alongκx = 0 pinpoint three states at which both the conventional and the ERS current injection are resonant.The upper boundaries of the grey areas depict Fermi levels of µ1 = −0.10 eV and µ2 = −0.20 eV (p-doped system). The horizontal axis corresponds to wavevectors κx measured from the Dirac point K,cf. Fig. 2.2. The vertices of the parabolic (confined) states occur at κx = W−1.
The resulting expressions are given in Table 2.1, and obey the following selection rules:
vxnm(κx) = 0 if ζn , ζm, (2.25a)
vynm(κx) = 0 if ζn = ζm. (2.25b)
We close this section by mentioning that the solutions corresponding to the Dirac point K′ are
analogous to those presented here for K. As shown by Marconcini et al. [25], the wavefunc-
tions for the A sites, Eqs. (2.11) and (2.17), at the K′ differ by a sign factor from those at K.
Moreover, the velocity operator at the K′ has the form v = vF(σx,−σy). This, together with
the properties of the envelope functions at both valleys, causes the x component of the VME at
K′ to have opposite sign of those at K; the y components of the VME are the same near K as
near K′.
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 28
2.3 Coherent injection and control
2.3.1 Framework
In this section, we describe the general framework of the two-color coherent control scheme.
As mentioned in the Introduction, the quantum interference is between pathways associated
with photon absorption processes arising from different phase related beams. These pathways
connect the same initial and final states, as shown for the processes in Fig. 2.3, where we
consider the two-color scheme with beams at ω and at 2ω. This figure depicts the two classes
of processes I study in this chapter.
The first, conventional processes, are those where current injection arises due to the in-
terference of one-photon absorption (1PA) at 2~ω and two-photon absorption (2PA) of (two)
photons with energy ~ω [1]; this is depicted with the set of arrows on the right of Fig. 2.3, un-
der the label “CONV”. In the remaining of the discussion, we label variables associated with
conventional processes with a subindex ‘C’.
The second class of processes arise in experiments on narrow band gap or gapless materials,
with ~ω > Eg, where Eg is the energy band gap. Under this condition, current injection can
arise due to the interference of 1PA at ~ω and stimulated electronic Raman scattering (ERS) at
~ω [5]. This ERS is indicated by the set of arrows at 2~ω and ~ω in the left of Fig. 2.3, under
the label “ERS”. We refer to variables associated with this Raman processes with a subindex
‘R’. We mention that in coherent control experiments on typical semiconductors, the beam
frequencies employed are such that ~ω < Eg < 2~ω, and, consequently, the ERS current is
absent because 1PA at ~ω is impossible.
Following van Driel and Sipe [1, 2], we calculate the one- and two-photon carrier injection
and current injection rates due to the interaction with a classical electromagnetic field
E(t) = E(ω)e−iωt +E(2ω)e−2iωt + c.c., (2.26)
in the long wavelength limit, where ω is the fundamental frequency. The interaction between
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 29
the electric field and the electron system is accounted by the minimal coupling prescription in
the Hamiltonian of Eq. (2.24); we do the usual replacement p j → p j − eA j(t), for j = (x, y),
with p j = −i~∂ j, and obtain the interaction Hamiltonian that acts as the perturbation,
Hint(t) = −ev ·A(t), (2.27)
where e = −|e| is the electron charge andA(t) is the vector potential associated with the electric
field, E(t) = −∂A(t)/∂t. We treat this problem using standard time-dependent perturbation
theory and Fermi’s golden rule. Since we are interested in 1PA, 2PA and ERS processes, the
unitary evolution operator U(t) is expanded perturbatively up to second order,
U(t) = e−iH0t/~Uint(t) (2.28)
where
Uint(t) =1 + (i~)−1∫ t
−∞
Vint(t1) dt1 + (i~)−2∫ t
−∞
Vint(t1) dt1
∫ t1
−∞
Vint(t2) dt2 + . . . (2.29)
and
Vint(t) = eiH0t/~ Hint(t) e−iH0t/~. (2.30)
Under the perturbation of Eq. (2.27), the evolution of the system’s state∣∣∣Υ〉 is not just the
ground state∣∣∣0〉, but it also contains an amplitude of the excited state |nmκx〉 (this ket corre-
sponds to a state with an electron-hole pair),
∣∣∣Υ(t)〉 = c0(t)|0〉 + cnmκx(t)|nmκx〉 + . . . , (2.31)
where∣∣∣cnmκx(t)
∣∣∣2 is the probability that the system is at∣∣∣nmκx〉; the missing terms in Eq. (2.31)
correspond to higher order excitations, which we neglect in this Chapter. The carrier injection
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 30
and the current injection rates are given by
n =1L
∑nmκx
ddt
∣∣∣cnmκx(t)∣∣∣2, (2.32)
Ja =1L
∑nmκx
e[va
nn(κx) − vamm(κx)
] ddt
∣∣∣cnmκx(t)∣∣∣2, (2.33)
respectively, where L is the normalization length introduced below Eq. (2.13). To describe
the optical processes we are interested, we compute cnmκx(t) up to second order (a tutorial
derivation is given by van Driel and Sipe in Ref. [1]; a more recent review is given by Rioux
and Sipe in Ref. [4]). Then, the expressions for these injection rates get the form,
n(1) = ξab(ω)Ea(−ω)Eb(ω), (2.34)
n(2)C = ξabcd
C (ω)Ea(−ω)Eb(−ω)Ec(ω)Ed(ω), (2.35)
n(2)R = ξabcd
R (ω)Ea(−2ω)Eb(−ω)Ec(2ω)Ed(ω), (2.36)
Ja = ηabcd(ω) Eb(−ω)Ec(−ω)Ed(2ω) + c.c., (2.37)
where repeated indexes indicate summation, ω is the fundamental frequency, n(1) and n(2)C(R)
account for the first- and second-order absorption processes, respectively; overall n refers to the
rate of injected carriers per unit length along the ribbon (carriers per unit length per unit time).
The 1PA coefficient is described by a second-order tensor, ξab, while the 2PA and the ERS
absorption coefficients are described by fourth-order tensors, ξabcdC and ξabcd
R , respectively. Here,
Ja includes the electron and hole contributions to the current (charge per unit time), injected
per unit time along the ribbon. The current injection coefficient η(ω) in Eq. (2.37) includes the
conventional and the ERS contributions, i.e., η(ω) = ηC(ω) + ηR(ω). In the following sections,
we give the full expressions for these coefficients. Note that the coefficients can be chosen such
that ξabcdC = ξbacd
C = ξbadcC and ηabcd = ηacbd.
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 31
2.3.2 First-order absorption process
We calculate the expressions for the coefficients ξ and η appearing in Eq. (2.34)–(2.37) using
Fermi’s golden rule. For the one-photon absorption coefficient, we obtain
ξab(ω) =4πe2
~2
∑nm
∫fmn(κx)
dκx
2πva
nm(κx) vb∗nm(κx)
ω2nm(κx)
δ(ωnm(κx) − ω), (2.38)
where we have gone from a sum over states to an integral over reciprocal space by L−1 ∑κx→
(2π)−1∫
dκx. In this expression the sum∑
nm runs over all bands, filled and empty (simi-
larly for the other response functions considered here); ωnm(κx) = ~−1Enm(κx) and Enm(κx) =
En(κx) − Em(κx) is the energy difference between two states at a given κx. A factor of two
has been included to account for spin degeneracy, which I do throughout this chapter. The
x−components of the VME at the K and K′ valleys differ just by a sign while the y−components
of the VME are the same. Consequently, since all integrals over reciprocal space include pairs
of VME, the integration over κx can be restricted to a single valley, K, and another factor of
two included to account for the contribution of the K′ valley.
The occupation of the states is described by the Fermi-Dirac distribution. In all of our
integrals over reciprocal space fmn(κx) = fm(κx) − fn(κx), with fn(κx) = [1 + e(En(κx)−µ)/(kBT )]−1 at
temperature T and chemical potential µ. Until the end of Sec. IV, we confine ourselves to zero
temperature, hence fn(κx) = θ(En(κx) − µ), where θ is the Heaviside step function. Because of
the selection rules for the VME, Eq. (2.25), the only nonzero components of the one-photon
coefficient are ξxx and ξyy, which we plot in Fig. 2.4 for a system at zero chemical potential. As
a comparison 3, we include plots of Wξxx2D, where W is the effective width of the ribbon,
ξxx2D(ω) = 2σ0(~ω)−1, (2.39)
and ξxx2D = ξ
yy2D is the 1PA coefficient for a 2D monolayer of graphene [3]; hereσ0 = gsgve2/(16~)
is the universal optical conductivity of graphene, and gs = 2, gv = 2 are the spin and valley
3At large photon energies, the two-photon absorption coefficients for zigzag nanoribbons drop off with thefifth power of the photon energy, as they do for a monolayer of graphene.
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 32
degeneracies, respectively. For ZGNR, the main difference between the two 1PA coefficients is
that ξyy diverges at zero photon energy, due to a divergence in the joint density of states (JDOS)
between bands n = +1 and n = −1. In contrast, for such a pair of bands ξxx is identically
zero, due to the VME selection rules. For an undoped ZGNR, ξxx displays its first divergence
at about 0.15 eV, which is the value of the band gap at zero Fermi level, and corresponds to the
onset of the transitions (2,−1) and (1,−2) at κx = W−1; these four states give the initiation en-
ergy for ξxx. In the following we indicate a transition from band m to band n by (n,m); hence,
for zero chemical potential and zero temperature, the possible transitions have m ≤ −1 and
n ≥ 1. In general, the ξxx and ξyy 1PA coefficients possess an infinite number of divergences
that arise due to the infinite number of parabolic bands in the bandstructure. Indeed, the JDOS
between states with band index n and m,
JDOSnm(E) = gsgv
∫dκx δ(E − Enm(κx)), (2.40)
can be shown to diverge as (E − Egapnm )−1/2 for the confined states, and as E−1 for edge states,
where E is the photon energy and Egapnm is the energy band gap between bands n and m. In fre-
quency space, these divergences occur at photon energies E such that E = Egapnm ; in reciprocal
space, they occur at κx points where argument of the delta function has a zero derivative. The
absorption coefficients inherit these JDOS divergences if the associated velocity matrix ele-
ments are nonzero at the κx at which dEnm/dκx = 0. The sensitivity of an experiment to these
divergences would depend on the resolution of the photon energy and on the magnitude of the
velocity matrix elements, as well as on the presence of scattering effects that are not included in
this simple treatment. In every pertaining Figure, we signal the location of these JDOS diver-
gences by small green ticks. An interesting characteristic of ξxx and ξyy is that the divergence at
the initiation energy always involves an edge state (see Table 2.2); this is reasonable, as these
states are involved in the minimum band gap for an undoped system.
As mentioned above, the sum over states runs over all bands, filled and empty, but for a
given photon energy range (e.g. 0 − 0.5 eV, as in Fig. 2.4) the sum requires a finite number of
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 33
bands. We refer to this as the “full” response. In order to highlight the contribution of the edge
states, we also compute the response coefficients with a restricted sum over states∑
nm, such
that n or m are ±1, e.g., (n,m) = (1,−1), (1,−2), (2,−1), . . .; we refer to this as the “edge”
contribution and in the appropriate figures we plot it with black-dashed lines. This allows us
to easily identify the contribution to 1PA from states at bands ±1. At low photon energies
such contribution is dominant: for ξxx, all transitions at photon energies ~ω < 0.350 eV are
from or to edge states; for ξyy, all transitions at photon energies ~ω < 0.439 eV are from
or to edge states. Consequently, at low-photon energies the “full” and “edge” contributions
are indistinguishable. This is shown in Fig. 2.4 (see also Table 2.2), where for comparison
we also plot Wξxx2D, where ξxx
2D is the 1PA coefficient of graphene calculated [3] at the same
level of approximation adopted here; it is clear how the presence of the edge states in ZGNR
significantly modifies the 1PA. Finally, we mention that the Dirac delta functions appearing in
all our expressions are treated with an interpolation scheme [57].
Table 2.2: Onset energies for the lowest energy transitions for an undoped cold ZGNR. Tuples (n,m)indicate a transition from band m to band n and every onset energy indicates the position of a JDOSdivergence. The peak number is as indicated in Fig. 2.4.
Peak ξxx ξyy
number E (eV) Transition E (eV) Transition
1 0.149 (2,−1), (1,−2) 0.000 (1,−1)2 0.323 (4,−1), (1,−4) 0.236 (3,−1), (1,−3)3 0.350 (3,−2), (2,−3) 0.410 (5,−1), (1,−5)4 0.498 (6,−1), (1,−6) 0.439 (4,−2), (2,−4)...
......
......
2.3.3 Second-order absorption processes
Conventional process
In this section, we start by considering the second order process related to the absorption of two
photons of energy ~ω, indicated by the rightmost arrows in Fig. 2.3. Carrying the perturbation
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 34
0
1
0 0.1 0.2 0.3 0.4 0.5
1 2 3 4
(b)
ξyy(ω
)[
ms−
1V
−2]
Photon energy hω [eV]
0
1
×109
1 2 3 4
(a)
ξxx(ω
)[
ms−
1V
−2]
FullEdgeWξxx
2D
Figure 2.4: (Color online) One photon absorption spectrum for a ZGNR of 95 zigzag lines (about20 nm width). The definitions of the full and edge contributions are given in the last paragraph of Sec.II B. For comparison, we include Wξxx
2D (red dot-dashed curves), where ξxx2D = ξ
yy2D is the 1PA coefficient
for graphene [3], given by Eq. (2.39). The green ticks along the horizontal axis indicate photon energiesat which JDOS divergences occur, which are numbered in concordance with Table 2.2.
calculation up the second order, we obtain the two-photon absorption (2PA) coefficient,
ξabcdC (ω) =
64πe4
~4
∑nm
∫fmn(κx)
dκx
2π
Vab∗C;nm V
cdC;nm
ω4nm(κx)
δ(ωnm(κx) − 2ω), (2.41)
where
Vi j
C;nm ≡ ~∑`
v in` v j
`m + v jn` v i
`m
2E` − En − Em + i βC, (2.42)
which we regard as the effective velocity matrix element (effective VME) for the second order
conventional process (C) process and the sum over ` corresponds to the virtual electron and
virtual hole contributions [1]. Although this sum runs over all bands (filled and empty), a
converged value is obtained for ` = 20 bands for a photon energy range of 0–1 eV.
In Eq. (2.42) we introduce a small constant βC in order to broaden resonant processes
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 35
(discussed below). This is done in a phenomenological basis, as a simple way to account for
the finite lifetime of the excited states. Throughout this thesis I set the value of this broadening
constant to the energy equivalent of room temperature, about 26 meV. By varying the value
of βC, I find that values closer to 26 meV provide a good compromise between preserving
the structure of the responce functions and broadening the resonant processes. At a more
fundamental level, deriving the interacting-particle Green function of an electron in a periodic
potential, the imaginary broadening constant βC can be shown to be related to the inverse of
the finite lifetime of the excitation (see, for example, page 615 of [58]).
From the selection rules for the regular VME, Eq. (2.25), we obtain the selection rules for
V,
VxxC;nm = 0 if ζn , ζm, (2.43a)
VyyC;nm = 0 if ζn , ζm, (2.43b)
VxyC;nm = 0 if ζn = ζm, (2.43c)
and from this we identify eight nonzero ξabcdC components, four of them independent, namely
ξxxxxC , ξ
xxyyC =
(ξ
yyxxC
)∗, ξ
xyxyC = ξ
xyyxC = ξ
yxxyC = ξ
yxyxC , and ξ
yyyyC , which we show in Fig. 2.5.
A feature of these coefficients is that the onset of the two-photon absorption signal is at the
minimum band gap between bands (2,−1), except for ξxyxyC , which has its onset at 0 eV; this
follows from the selection rules for the effective VME, which are inherited from the usual
VME, and indicate that the transition (1,−1) is allowed.
As we found for the 1PA coefficients ξab, the 2PA coefficients ξabcdC suffer from divergences,
but for the 2PA coefficients they are of two types: JDOS divergences and effective-VME-
divergences. The latter results when the nominal virtual state lies at the average of the energies
between two transition states, |nκx〉 and |mκx〉, i.e., when (see Eq. (2.42))
E` = (En + Em)/2. (2.44)
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 36
Such condition corresponds to a resonant 2PA and an instance where this occurs is indicated
on Fig. 2.3 by the three dots along the vertical line at κx = 0. In Fig. 2.5 we distinguish these
two types of divergences by small vertical lines of different color; a green tick indicates the
presence of a JDOS-divergence, while a red tick indicates the presence of an effective-VME-
divergence. In order to broaden the latter resonances, a small damping constant βC of 20 meV
was introduced in the denominator of Eq. (2.42). This value, which is close to the thermal
energy kBT associated with room temperature, was chosen arbitrarily. A more detailed theory
would be necessary to indicate how these resonances are really broadened; the choice we make
here simply allows us to identify easily where these resonances occur in our calculations. We
mention that the onset of ξxxxxC is due to the transitions (2,−1) and (1,−2), which are free from
resonances because the matrix elements to the intermediate states (one of the edge bands ±1
that would lead to a divergent condition) are forbidden by the selection rules. Therefore, in the
photon energy range 0 to 0.15 eV, the coefficient ξxxxxC is free of resonances.
We present the ξabcdC coefficients in Fig. 2.5, and identify the edge contributions to them
(black-dashed lines). As we found for ξab, for ξabcdC the edge states make a dominant contri-
bution at low photon energies, and are involved at the onset of 2PA. As a comparison 4, in
Fig. 2.5, we include plots of Wξabcd2D , where W is the effective width of the ribbon,
ξxxxx2D (ω) = 8gsgv~e4v2
F(2~ω)−5, (2.45)
and ξxxxx2D = ξ
yyyy2D = ξ
xyxy2D = ξ
xyyx2D = −ξ
xxyy2D are the 2PA coefficients for a 2D monolayer of
graphene [3]; as before, gs = 2 and gv = 2 are the spin and valley degeneracies, respectively.
ERS process
Now we consider another second order process involving light at 2ω and light at ω, stimulated
electronic Raman scattering, which can be characterized as virtual absorption at 2~ω followed
by emission at ~ω; see the left diagram in Fig. 2.3. This process exists in semiconductors
4At large photon energies, the two-photon absorption coefficients for zigzag nanoribbons drop off with thefifth power of the photon energy, as they do for a monolayer of graphene.
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 37
0
2
4
6
8
0 0.05 0.1 0.15 0.2 0.25
Wξyyyy
2Dր
(d)
ξyyyy
C(ω
)[
m3s−
1V
−4]
Fundamental photon energy hω [eV]
×10−5
0
1
2
3
4
ւWξxyxy
2D
(c)
ξxyxy
C(ω
)[
m3s−
1V
−4]
×10−4
−6
−4
−2
0
2
4
6
տWξxxyy
2D
(b)
ξxxyy
C(ω
)[
m3s−
1V
−4]
×10−6
0
1
2
3
4
5
6
ւWξxxxx2D
(a)
ξxxxx
C(ω
)[
m3s−
1V
−4]
Full
Edge
×10−6
Figure 2.5: (Color online) Nonzero two-photon absorption coefficients ξabcd
C for aZGNR with 95 zigzag-lines (about 20 nmwidth). The definitions of the full and edgecontributions are given in the last paragraphof Sec. II B. On each panel, we includeWξabcd
2D (red dot-dashed curves), where ξabcd2D
(Eq. (2.45) and text below) is for a graphenesheet [3]. The green (red) ticks along thehorizontal axis indicate the photon energiesat which JDOS divergences (resonances) oc-cur.
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 38
when the fundamental photon energy is larger than the band gap, which is always the case for
an undoped ZGNR, because the edge states provide a zero-gap system. Following an earlier
treatment of graphene [5], we find the ERS carrier injection to be
ξabcdR (ω) =
2πe4
~4
∑nm
∫fmn
dκx
2π
V d a ∗R;nmV
b cR;nm
ω4nm
δ(ωnm(κx) − ω), (2.46)
where the effective VME for the ERS process are
Vi j
R;nm ≡ ~∑`
v in`v
j`m
E`n − Enm + iβR+
v jn` v i
`m
E`m + Enm + iβR
. (2.47)
As in Eq. (2.42), βR is a small constant introduced to broaden resonant processes and the sum
over ` runs over all bands (filled and empty), but a converged value is obtained for ` = 30 bands
for a photon energy range of 0–1 eV. The first term in the sum of Eq. (2.47) corresponds to
photo-emission by an electron, and the second to photo-emission by a hole [5]. Note that due
to the different frequencies involved in Eq. (2.36), symmetrization of VijR is unnecessary. The
selection rules forVijR are the same as those forVij
C (Eq. (2.43)); note, however, thatVijR , V
jiR,
although VijR and Vji
R satisfy the same selection rule. From this we identify six nonzero terms
for the ERS carrier injection coefficient, ξxxxxR , ξxyyx
R =(ξ
yxxyR
)∗, ξxxyy
R =(ξ
yyxxR
)∗, ξxyxy
R , ξyxyxR , and
ξyyyyR . As do the conventional coefficients, the ERS coefficients suffer from JDOS and effective-
VME divergences, the later arising whenever
E` = 2En − Em or (2.48a)
E` = 2Em − En (2.48b)
is satisfied. These conditions correspond to resonant processes, when a state is located at an
energy |Enm(κx)| above (below) the final (initial) state n (m). As in Eq. (2.42), a small damping
constant βR of 20 meV was introduced in the denominators of Eq. (2.47). All of these ERS
coefficients present a large number of these resonances, causing ξabcdR to be highly sensitive to
the value of the βR parameter. However, these resonances are of small magnitude for the energy
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 39
range chosen for Fig. 2.6, hence they are not apparent. As shown, three of these components
have their onset at zero photon energy, because the symmetry properties of the involved matrix
elements allow for transitions between the two edge states.
0
1
2
0 0.05 0.1 0.15 0.2 0.25
(c)
ξyxyx
R(ω
)[
m3s−
1V
−4]
Fundamental photon energy hω [eV]
×10−4
0
1
2
3
4 (b)
ξxyxy
R(ω
)[
m3s−
1V
−4]
×10−5
0
0.5
1
1.5
2(a)
ξxxxx
R(ω
)[
m3s−
1V
−4]
Full
Edge
Wξ′abcd2D
×10−7
0
1
0 0.05 0.1 0.15 0.2 0.25
(f)
ξyyyy
R(ω
)[
m3s−
1V
−4]
Fundamental photon energy hω [eV]
×10−6
0
1
2
3 (e)
|ξxyyx
R(ω
)|[
m3s−
1V
−4]
×10−7
0
1
2(d)
ξxxyy
R(ω
)[
m3s−
1V
−4] ×10−5
Figure 2.6: (Color online) ERS carrier injection tensor, as given by Eq. (2.46). The definitions of thefull and edge contributions are given in the last paragraph of Section II.B. Notice that the edge statesplay a dominant contribution to the ERS absorption process, due to the large amount of resonant states.The green (red) ticks along the horizontal axis indicate the photon energies at which JDOS divergences(resonances) occur. The red dot-dashed lines indicate the ERS processes for 2D graphene [5].
2.3.4 Current injection
Injection coefficients
We begin with the expression for ηC, the current injection coefficient characterizing the con-
ventional process. Here the interference between the 2PA at ~ω with 1PA at 2~ω (see the
right diagram in Fig. 2.3) leads to net current injection coefficients (including electron and hole
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 40
0
1
2
0 0.05 0.1 0.15 0.2 0.25
(c)
|ηxyyx(ω
)|[
Cm
3s−
2V
−3]
Fundamental photon energy hω [eV]
×10−11
0
1
2(b)
|ηxxyy(ω
)|[
Cm
3s−
2V
−3]
×10−11
0
1
2(a)
|ηxxxx(ω
)|[
Cm
3s−
2V
−3]
Full
Edges
Wηabcd
2D
×10−12
Figure 2.7: (Color online) Nonzero net current in-jection coefficients, including the conventional andERS contributions, i.e. η(ω) = ηC(ω) + ηR(ω). Thedefinitions of the full and edge contributions aregiven in the last paragraph of Section II.B. On eachpanel, we include Wηabcd
2D (dot-dashed red curves),where ηabcd
2D (Eq. (2.51) and text below) is for agraphene sheet [5]. The red ticks along the horizon-tal axis indicate the energies at which resonancesoccur; a damping constant of 20 meV is introducedto broaden such resonances. The dips observed inthese coefficients arise due to negative contributionsto the conventional and ERS currents, in turn due tothe shape of the involved matrix elements.
contributions) given by [1]
ηabcdC (ω) =
16iπe4
~3
∑nm
∫fmn
dκx
2π
(va
nn − vamm
)Vbc∗
C;nmvdnm
ω3nm
δ(ωnm(κx) − 2ω). (2.49)
From the selection rules for the regular and the effective VME, Eq. (2.25) and Eq. (2.43), we
identify three nonzero current injection coefficients, ηxxxxC , η
xyyxC , and ηxxyy
C = ηxyxyC . Notice that
for all these tensors the first Cartesian component is x: Due to the confinement of the ribbons
along the y direction (see Fig. 2.1), the current injection can only flow along the x direction,
and all tensor components ηyabcC are zero.
Turning to the expression for ηR, the current injection coefficient characterizing the inter-
ference between the ERS discussed above and the 1PA at ω (see the left diagram in Fig. 2.3),
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 41
including both electron and hole contributions we find
ηabcdR (ω) =
2iπe4
~3
∑nm
∫fmn
dκx
2π
(va
nn − vamm
)ω3
nm
[vb∗
nmVcdR;nm + vc∗
nmVbdR;nm
]δ(ωnm(κx) − ω), (2.50)
where VR is given by Eq. (2.47). On the basis of the matrix elements selection rules, we
identify three nonzero ERS current injection coefficients, ηxxxxR , ηxyyx
R , and ηxxyyR = η
xyxyR .
Over the frequency range shown in Fig. 2.7, the conventional and the ERS current injection
coefficients are of the same order, dropping off as the inverse of the third power of the photon
energy, as do the coefficients for graphene [5]. Thus we only plot the total injection coefficients
ηabcd = ηC(ω) + ηR(ω). For comparison, we include plots of Wηabcd2D (with the respective values
of the Cartesian indices), where
ηxxxx2D (ω) = i
34
gsgve4v2F(2~ω)−3, (2.51)
and ηxxxx2D = 3ηxxyy
2D = 3ηxyyx2D are the net current injection coefficients for a 2D monolayer of
graphene [5]; as before, gs = 2 and gv = 2 are the spin and valley degeneracies, respectively.
As we saw for carrier injection, the edge states provide the strongest contribution at the onset
of current injection. Another characteristic of these coefficients is that ηxxxx has its onset at the
band gap between bands (2,−1), while ηxxyy and ηxyyx have their onset at 0 eV. This is due to
the selection rules that the matrix elements involved in both the conventional and ERS process
satisfy, allowing transitions between bands (1,−1). An important characteristic of the current
injection coefficients is that they are free of JDOS divergences, because the diagonal matrix
elements in their respective expressions, Eqs. (2.49) and (2.50), are identically zero at the κx at
which the minimum gap occurs. However, a number of effective VME resonances do exist at
photon energies indicated by the small red ticks in Fig.2.7, such that Eq. (2.44) is satisfied. As
explained before, the magnitude of these resonances is broadened by a small damping constant.
These coefficients are shown in Fig. 2.7, where we present the net current injection arising from
the addition of the conventional and ERS contributions, i.e., η(ω) = ηC(ω) + ηR(ω).
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 42
Swarm velocities
The numerical values of the coefficients ξab, ξabcdC(R), and ηabcd
C(R) do not immediately give a sense of
the average velocities with which the electrons and holes are injected. Sometimes an average,
or swarm velocity is introduced to indicate this [1]. In the system considered here, we could
introduce a swarm velocity for both the conventional and ERS processes, according to
V C(R) =1e
JC(R)(ω)
n(1)(Ω) + n(2)C(R)(ω)
, (2.52)
where Ω = 2ω for V C because JC arises from the interference of 1PA at 2ω with 2PA at
ω, while Ω = ω for V R because JR arises from the interference of 1PA at ω with the ERS
described above. Besides describing an average speed that characterizes the injected carriers,
one can consider maximizing Eq. (2.52) by using appropriate phases in the optical beams, and
adjusting the relative amplitudes of the light at ω and 2ω. Considering just the swarm velocity
of the conventional process, such optimization leads to equal 1PA and 2PA, and it follows that
the intensity of the fundamental beam at ω should be about half an order of magnitude larger
that of the beam at 2ω, for a fundamental photon energy of about 0.4 eV. In contrast, the swarm
velocity of the ERS process depends only on the intensity of the beam at 2ω. Further, in trying
to optimize the net swarm velocity, determined by the total current injected divided by the total
carrier density injected, one finds that the beam at 2ω should have an intensity about an order of
magnitude larger than the beam at ω. Since in typical experiments the beam at 2ω is obtained
by second harmonic generation of part of the beam at ω, this would be impractical. Thus we
calculate the conventional and Raman swarm velocities for typical [53] beam intensities of
the fundamental and second harmonic fields, shown in Fig. 2.8. We complement these carrier
velocities with the total average velocity of the injected carriers
V tot =1e
JR(ω) + JC(ω)
n(1)(ω) + n(1)(2ω) + n(2)C (ω) + n(2)
R (ω), (2.53)
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 43
0
1
0 0.05 0.1 0.15 0.2 0.25
×1026 (b)
n[(m
·s)
−1]
Photon energy hω [eV]
Conv
ERS
Total
-1012345
×105 (a)
ImV
[ms−
1] Conv swarm
ERS swarmWeighted Average
Figure 2.8: (Color online) Swarm and (weighted) average velocities (top), accompanied by the carrierdensity rates (bottom) along x due to ηxxxx for typical [53] beam intensities of the fundamental andsecond harmonic fields. The average carrier velocities (black-dashed line) for ηxxyy and ηxyxy are of thesame order, but their net components have a smooth onset at zero photon energy.
also evaluated at typical [53] beam intensities. These carrier velocities are shown in Fig. 2.8. As
a reference, at the photon energy of 0.25 eV, the maximum swarm velocity of the conventional
process for a monolayer of graphene is 2.9 × 105 ms−1. Hence the carrier velocities in ZGNR
are comparable to those on a monolayer of graphene, as might be expected.
2.4 Doping
In the previous sections, we investigated the carrier and current injection at zero chemical
potential. Since the dispersion relations of the edge states in ZGNR have a zero band gap and
are flattened for κx > W−1 (Fig. 2.3), those states are always involved at the onset energy of
all of the optical response coefficients studied here. This suggests that doping is an effective
method to alter the population of these two bands and the current that can be injected by
the optical transitions between them. In this section, we revisit the calculations of ξab, ξabcdC(R)
and ηabcd for a negative chemical potential, corresponding to a p-doped system. Besides the
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 44
modified contribution from the edge states, we will also see significant modification in the
contributions from other bands, particularly in the region near the K and K′ points, where
doping leads to either a “valley” of filled states (n-doped), or a “hill” of unfilled states (p-
doped); see Fig. 2.3.
We consider two negative Fermi levels, µ1 = −0.1 eV and µ2 = −0.2 eV, which in Fig. 2.3
we indicate by the upper boundaries of the grey areas. The value of −0.1 eV is interesting
because, at this chemical potential, the flat part of band −1 (i.e., the region where κx > W−1, cf.
Fig. 2.3) contains empty states; this condition allows transitions from lower energy bands with
final states in band −1, but also disables transitions from band −1 to upper bands. The second
value, µ = −0.2 eV, is interesting because at this potential a “hill” of unfilled states arises in
the first parabolic band (band −2 in Fig. 2.3) at energies below our nominal value of zero.
We present the results of the calculations of 1PA coefficients for those values of the chem-
ical potential in Fig. 2.9. In an undoped sample, the JDOS divergences in ξxx at low photon
energies are due to the onset of the transitions (2,−1), (1,−2), (4,−1), and (1,−4) (see Table 2.2
and Fig. 2.4). Since all of these transitions involve bands ±1, any nonzero chemical potential
has the capacity to significantly alter the 1PA at these photon energies. For instance, if the
Fermi level is at −0.1 eV, then the flat part of band −1 contains empty states, and the low pho-
ton energy divergences are removed. In addition, at this chemical potential transitions of the
type (−1, n), for n odd and < −1 are permitted. However, the contributions to the 1PA from
these new transitions are of smaller magnitude than the contribution from the (1,−2) transition,
which is unaffected by the −0.1 eV doping. For this reason, the (1,−2) transition remains as
the main contribution to the ξxx coefficient at low photon energies at this chemical potential
(see Fig. 2.9).
At the Fermi level −0.2 eV, the edge states are completely empty, as are the states at the
higher points of band −2 near the K and K′ points. This condition allows transitions of the
type (−2, n), for n even and < −2, and also forbids transitions of the type (n,−2), for n odd and
≥ 1, and κx near the K and K′ points. It is this latter restriction which significantly changes the
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 45
ξxx coefficient near its onset. A further decrease in the Fermi level would consistently remove
the divergences in ξxx at low photon energies. All these observations were confirmed with a
band-by-band calculation of ξxx.
The effect of doping the system has a larger influence on the onset energy of ξyy that on
that of ξxx. This is because the JDOS divergences at low photon energies relevant for ξyy are
due to the transitions (1,−1), (3,−1), and (1,−3) (cf. Table 2.2). Therefore, even for small
doping, the large contribution coming from the transitions between the two edge states (bands
±1, κx > W−1) is significantly decreased, and leads to a greater change of the magnitude of
ξyy than of the magnitude of ξxx. A special signature of ξyy for µ = −0.2 eV (dark-violet
signal, Fig. 2.9 b)) is the presence of two narrow peaks at 0.045 and 0.075 eV; the first of these
peaks is due to the (−1,−2) transition, while the second is from the (−2,−3) transition. These
two transitions are active only for those κx states at which the “hill” of band −2 is empty (see
Fig. 2.3). Notably, the transition (−2,−3) brings a new JDOS divergence because it is active
over a range of reciprocal space that includes κx = W−1, where both bands have their maximum
and their energy difference Enm(κx) has a zero derivative (see the discussion below Eq. (2.40)).
In general, all these new transitions involve more JDOS divergences if the range of κx over
which they are active includes the κx at which the band pairs have their maxima or minima. For
instance, the divergences 1–4 in Fig. 2.9 are the same as those in Fig. 2.4 and Table 2.2, but
the divergences 5–6 arise due to the new transitions allowed at nonzero chemical potentials: in
Fig. 2.9 a), at the chemical potential −0.20 eV, the divergence 5 at 0.179 eV is due to the tran-
sition (−2,−4), which is active over a range of κx that includes the κx at which bands (−2,−4)
have their maxima, hence a new JDOS divergence appears. Likewise for ξyy in Fig. 2.9 b)
at µ = −0.20 eV: divergences 5 at 0.089 eV and 6 at 0.268 eV exist because the transitions
(−2,−3) and (−2,−5) are active over regions of reciprocal space that include the κx at which
such bands have their maxima.
In Fig. 2.10, 2.11, and 2.12 we present the nonzero ξabcdC , ξabcd
R , and ηabcd coefficients for
selected nonzero Fermi levels. As was seen for ξab, doping the ZGNR has the effect of modi-
fying the responses around their onset energy, either due to the removal of some transitions, or
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 46
due to the appearance of new ones, which in the undoped system were forbidden because the
initial and final states were filled [e.g. (−1,−2) or (−1,−3)]. This shows that doping is an ef-
fective way of modifying the carrier and current injection in ZGNR, where the most significant
changes are due to the removal of density of states at the edge bands.
We close this section by mentioning that we performed finite temperature calculations at
room temperature; this was achieved by implementing a temperature dependence of the Fermi
factors through the Fermi-Dirac distribution. We found that the only significant change is in
that the onset energy of the coefficients ξab, ξabcdC(R), and ηabcd are smaller. However, the mag-
nitudes of the coefficients at energies near the lower onsets are several orders of magnitude
smaller that the magnitudes of the corresponding coefficients at zero temperature near their
energy onsets.
0
2
4
6
0 0.1 0.2 0.3 0.4 0.5
1 2 3 45 6
(b)
ξyy(ω
)[
ms−
1V
−2]
Photon energy hω [eV]
0
1
2
×109
1 2 3 45
(a)
ξxx(ω
)[
ms−
1V
−2]
µ [eV]0
−0.10−0.20
Figure 2.9: (Color online) One photon absorption coefficients as a function of the photon energy forselected Fermi levels corresponding to p-doped samples. The ZGNR has 95 zigzag lines (about 20 nmwidth). For nonzero chemical potentials, some transitions become impossible and some new transitionsarise, possibly leading to new JDOS divergences (e.g. divergences 5 and 6). Divergences 1–4 are thesame as in Fig. 2.4.
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 47
0
2
4
6
8
0 0.05 0.1 0.15 0.2 0.25
0
1
2×10−5←−
×10−3−→
(d)
ξyyyy
C(ω
)[
m3s−
1V−4]
Fundamental photon energy hω [eV]
0
1
2
3
4
0
2
4
6
8×10−4
(c)
ξxyxy
C(ω
)[
m3s−
1V−4]
0
1
2
3
4
5
0
1
2
3
4
5×10−6
(b)
|ξxxyy
C(ω
)|[
m3s−
1V−4]
0
1
2
3
4
5
6
0
1
2
3
4
5
6×10−6
(a)
ξxxxx
C(ω
)[
m3s−
1V−4] µ [eV]
0
−0.10
−0.20
Figure 2.10: (Color online) Two-photon ab-sorption coefficients for selected Fermi lev-els corresponding to p-doped samples. TheZGNR has 95 zigzag lines (about 20 nmwidth). For panels where two different verti-cal scales are present, i.e. panel (d), the scaleon the left (right) is for undoped (doped)cases (arrows below the factors indicate theordinate for which they apply). A dampingconstant βC = 20 meV was introduced.
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 48
0
1
2
0 0.05 0.1 0.15 0.2 0.25
0
1
2
(c)
ξyxyx
R(ω
)[
m3s−
1V−4]
Fundamental photon energy hω [eV]
×10−4←−
×10−5−→
0
1
2
3
4
0
1
2×10−5
(b)
ξxyxy
R(ω
)[
m3s−
1V−4]
0
1
2
0
2
4
6
8
10×10−7
(a)ξxxxx
R(ω
)[
m3s−
1V−4]
µ [eV]0
−0.10
−0.20
0
0.5
1
0 0.05 0.1 0.15 0.2 0.25
0
0.5
1
×10−6
(f)
ξyyyy
R(ω
)[
m3s−
1V−4]
Fundamental photon energy hω [eV]
0
1
2
3
0
1
2
3×10−7
(e)
|ξxyyx
R(ω
)|[
m3s−
1V−4]
0
1
2
0
1
2
3
4
(d)
ξxxyy
R(ω
)[
m3s−
1V−4]
×10−5←−
×10−6−→
Figure 2.11: (Color online) ERS carrier injection coefficients for selected Fermi levels corresponding top-doped samples. The ZGNR has 95 zigzag lines (about 20 nm width). For panels where two differentvertical scales are present, i.e. (c) and (d), the scale on the left (right) is for undoped (doped) cases(arrows below the factors indicate the ordinate for which they apply). Notice that at µ = −0.20 meVsome resonances are absent, e.g., at 0.15 eV in (a) and (f); this is because, at this Fermi level, the statesat which these resonances are present for the undoped system, now contain empty states. A dampingconstant βC = 20 meV was introduced.
2.5 Limits of the model
The model employed in this Chapter inherits the limits of applicability of time-dependent per-
turbation theory, which is restricted to situations of low electron-hole pair densities [59] (for
high injection densities a density matrix formalism could be employed to study the dynamics).
The regime of validity of the perturbation treatment used here can be estimated: we require the
populated fraction of excited states accessible to a typical Gaussian pulse to be small.
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 49
01234567
0 0.05 0.1 0.15 0.2 0.25
01234567
(c)
|ηxyyx(ω
)|[
Cm
3s−
2V−3]
Fundamental photon energy hω [eV]
×10−11←−
×10−12−→
0
1
2
0
1
2
(b)
|ηxxyy(ω
)|[
Cm
3s−
2V−3]
×10−10←−
×10−11−→
0
1
2
0
1
2
×10−12 (a)
|ηxxxx(ω
)|[
Cm
3s−
2V−3]
µ [eV]0
−0.10
−0.20
Figure 2.12: (Color online) Net current in-jection tensors (conventional plus ERS con-tributions) for selected Fermi levels corre-sponding to p-doped samples. The ZGNRhas 95 zigzag lines (about 20 nm width).For panels where two different verticalscales are present, i.e. (b) and (c), the scaleon the left (right) is for undoped (doped)cases (arrows below the factors indicate theordinate for which they apply). A dampingconstant of 20 meV was introduced.
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 50
2.5.1 Graphene sheet
As a reference, we first consider monolayer graphene. When the electric fields of the optical
beams are all aligned along x, the one- and two-photon injection coefficients for a 2D graphene
sheet are [3] given by Eqs. (2.39) and (2.45). For each of ξxx2D and ξxxxx
2D , we set the number of
carriers injected per unit area to be less than the number of states per unit area accessible to the
optical beam. Then taking the beam intensity as I(ω) = 2ε0c|E(ω)|2, we arrive to
I(2ω) <ε0cαω
2πv2F(∆t)2ξxx
2D(2ω), (2.54)
I2(ω) <(2ε0c)2αω
2πv2F(∆t)2ξxxxx
2D (ω), (2.55)
where α is the time-bandwidth product for the optical beam (which we take as 0.44, typical for
a Gaussian beam), ∆t is the pulse-duration, and vF ≈ 106 m/s is graphene’s Fermi velocity.
2.5.2 Zigzag nanoribbons
The estimate for the nanoribbon case is similar to the graphene sheet, aside from the fact that
the areal ratios become length ratios, i.e. for each one of 1PA and 2PA coefficients we set the
number of carriers injected per unit length to be less than the number of states per unit length
accessible to the optical beam, giving us
I(2ω) <ε0cα
π(∆t)2ξxx(2ω)(|ve| + |vh|), (2.56)
I2C(R)(ω) <
(2ε0c)2 α
π(∆t)2ξxxxxC(R)(ω)(|ve| + |vh|)
, (2.57)
where α and ∆t where defined previously, ve is the velocity of the injected electrons in the
conduction band, given by the matrix element vnn, and vh is the velocity of the holes injected in
the valence band, given by vmm. Equation (2.57) provides the expression for the conventional
(C) and ERS processes (R).
In order to compare the limiting intensities of our model for a graphene sheet and for
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 51
ZGNR, we assume a typical pulse duration of 220 fs and beam wavelengths of 3.2µm and
1.6µm for the ω and 2ω beams [53]. Then we identify the states that contribute at these two
wavelengths, and find that, on average, |ve| + |vh| ≈ vF . From Eqs. (2.54) and (2.56), at λ =
1.6 µm,
IGraphene(2ω)IRibbons(2ω)
=ωξxx(2ω)
2vF ξxx2D(2ω)
≈ 2.6, (2.58)
and from Eqs. (2.55) and (2.57), at λ = 3.2 µm,
IGraphene(ω)IRibbons(ω)
=
√ωξxxxx
C (ω)2vF ξ
xxxx2D (ω)
≈ 1.6. (2.59)
Equations (2.58) and (2.59) indicate that the limiting intensities of our model are similar for a
graphene sheet and for a ZGNR, within an order of magnitude.
We find that, under the assumptions made in this section, the estimated limit for the beam
intensities at ω in the ZGNR and the 2D graphene are about two orders of magnitude below
the intensities used in some experiments [53] on 2D graphene, where coherent current injection
was observed. Due to relaxation processes, of course, the number of allowed carrier excitations
below saturation is expected to be higher than our estimates, leading to larger values of the
beam intensities for which a perturbation approach would be valid. Based on the estimates in
Eqs. (2.58) and (2.59), if relaxation processes affect the ribbon samples as effectively as they
do for 2D samples, we can expect coherent control in ZGNR to be observable at the higher
intensities used in 2D graphene experiments.
2.6 Summary and discussion
We have calculated the response coefficients for one- and two-photon charge injection and the
two-color current injection in a graphene zigzag nanoribbon; we use the semi-empirical k · p
method to describe the electron wavefunctions by smooth envelope functions.
The only nonzero one-photon injection coefficients correspond to the case of all-x or all-y
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 52
aligned fields, i.e., ξxx and ξyy. These two coefficients possess a rich structure of divergences,
caused by divergences of the joint-density-of-states originating from the infinite set of parabolic
bands present in the zigzag nanoribbon. These two coefficients have distinct selection rules for
the allowed transitions.
The two-photon carrier injection coefficients drop off as the fifth power of the photon en-
ergy at large photon energies, as they do for monolayer graphene. Moreover, these coeffi-
cients possess two classes of divergencies. One corresponds to the joint-density-of-states di-
vergences associated with the parabolic bands. The second class corresponds to divergences
arising from resonant conditions, when the two-photon absorption processes arise from sequen-
tial one-photon absorption processes between real states. In our calculation here we broadened
these resonances phenomenologically, but a more sophisticated treatment of these resonantly
enhanced transitions is an outstanding problem on which we hope this work will encourage
further study. The onset of the signals is determined by the minimum energy band gap and the
selection rules for these coefficients.
We calculated the electron and hole contributions to the conventional and the stimulated
electronic Raman scattering (ERS) current injection processes, finding that the only nonzero
components are associated with current injected along the length of the nanoribbon, as ex-
pected. The behavior of these coefficients as a function of the photon energy follows the be-
havior of 2D graphene [∼ (~ω)−3] at large photon energies, aside of the resonances present in
the ribbons. We have also calculated the so-called swarm velocity of the injected electrons,
which inherits a rich structure as a function of the photon energy due to the details of the struc-
ture of the injection coefficients. All these calculations were presented for a system at zero
Fermi level and zero temperature. However, we also carried finite temperature calculations
and found that, within this model, finite temperatures only account for changes at the onset
of the signals, which are several orders of magnitude smaller than the nominal values at zero
temperature.
Lower bound estimates on the permissible incident intensities for which the calculations
here can be valid were presented. They are similar to those of monolayer graphene, where
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 53
coherent current injection has been observed at much higher intensities than these simple es-
timates, which do not take into account the relaxation effects in the excited populations. Thus
experiments to demonstrate coherent current injection in ZGNR seem to us to be in order.
For experiments contemplated for ribbons of different width than those studied here, it is
important to note that simple scaling arguments show that the wider the ribbon, the stronger
the confinement of the energy bands. As shown in the calculations presented in this Chapter,
at low photon energies, the band gap follows a linear relation with respect to the inverse of
the ribbon width. Consequently, increasing the width of the ribbon decreases the energy band
gap between any pair of bands. This in turn shifts the onset energy of the response coefficients
towards zero energy and increases the number of JDOS divergences per photon energy. For
instance, the onset of the response coefficients when light is polarized along the length of the
ribbon is determined by the bangap between bands (1,−2) (see Fig. 2.3). For such pair of
bands, a linear fit shows that the band gap depends on the ribbon width W as Egap1,−2 ≈ aW−1
with a = 2.98 eV · nm. Besides altering the onset energy of the responses, a larger width also
leads to a larger magnitude of the injection coefficients, larger than would be expected simply
on the basis of the increase in material; e.g., a width increase of about 15% doubles the size of
ηxxxx.
As the outstanding signature of the zigzag nanoribbons are the strongly localized edge
states, we have identified their contribution to the carrier- and current-injection processes. In
all cases the edge states always participate in the onset of the signals. This lead us to consider a
second scenario to study these localized states: given that the dispersion relations of these states
are flattened towards zero energy for certain regions in k-space, we re-visited our calculations
considering doped scenarios. We found that that even small doping levels allow for signifi-
cant changes around the onset energy of the signals. This is because the large joint-density
of states present between the edge states is diminished with nonzero chemical potentials. Due
to the relative ease of doping graphene systems, the present work shows that zigzag nanorib-
bons offer an excellent opportunity to investigate scenarios in which electrical currents can be
generated and controlled optically. While more sophisticated treatments of the electron states
Chapter 2. Coherent control of current injection in zigzag graphene nanoribbons 54
and the inclusion of electron-electron interaction [34, 44] will undoubtedly add to the richness
of the injection processes, we hope that the description given here will motivate all-optical
current injection experiments. Although coherent control has been studied and observed on
graphene sheets, zigzag graphene nanoribbons have the advantage of having optical responses
that depend strongly on the geometry and width of the ribbon. Moreover, as shown in the liter-
ature, the localized states present in these ribbons are highly sensitive to external fields, doping
and functionalization. All these characteristics endow graphene zigzag ribbons with a richness
absent in simpler graphene sheets.
Chapter 3
An Effective Model for the Electronic and
Optical Properties of Stanene
Abstract
The existence of several 2D materials with heavy atoms has recently been demonstrated. The
electronic and optical properties of these materials can be accurately computed with numer-
ically intensive density functional theory methods. However, it is desirable to have simple
effective models that can accurately describe these properties at low energies. Here I present
an effective model for stanene that is reliable for electronic and optical properties for photon
energies up to 1.1 eV. For this material, I find that a quadratic model with respect to the lattice
momentum is the best suited for calculations based on the bandstructure, even with respect
to band warping. I also find that splitting the two spin-z subsectors is a good approximation,
which indicates that the lattice buckling can be neglected in calculations based on the band-
structure. I illustrate the applicability of the model by computing the linear optical injection
rates of carrier and spin densities in stanene. Our calculations indicate that an incident circu-
larly polarized optical field only excites electrons with spin that matches its helicity. A modified
version of this chapter was published in Physical Review Materials 1, 054006 (2017).
55
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 56
3.1 Introduction
The experimental isolation of single layers of graphene nearly a decade ago has inspired a
search for new 2D materials [60, 61]. Among those recently studied are silicene [62, 63], zinc-
oxide [64], and the transition metal dichalcogenides [65,66]. There is also substantial research
on other elemental 2D materials, including the remaining elemental crystallogens [67–69],
elemental pnictogens, such as nitrogene [70], phosphorene [71], arsenene [72], antimonene
[73,74], and bismuthene [75], as well as members from other families [76,77]. One of the most
interesting materials in this group is stanene, a monolayer of Sn atoms arranged in a buckled
honeycomb lattice. Due to the heavy Sn atoms, the spin-orbit coupling (SOC) is expected to
be strong and to lead to nontrivial topological properties of the bands that make stanene a 2D
topological insulator [78]. The strong SOC is predicted to open band gaps of 88 meV at the
K and K′ points of the Brillouin zone [78, 79], and thus the quantum spin Hall effect, with its
characteristic spin polarized edge modes free of backscattering from non-magnetic impurities,
could in principle be observed at room temperature. Recently, monolayers of stanene have
been epitaxially grown [80], and phase-change laser ablation techniques [81] have been used
to produce few-layer stanene. Experiments probing high photon energy absorption properties
of few-layered stanene have also been reported [82].
While the electronic and optical properties of crystalline materials can be studied with
modern ab initio methods, the numerical task can be challenging. It is thus desirable to have
simple effective models that reliably reproduce the basic properties of materials, at least over
energy ranges of interest. In order to compute electronic and optical properties from an ef-
fective model, it is necessary to know the Hamiltonian and the Lax connection1, which gives
important geometric information about the basis of the quantum states [84] in the model. Two
of the most common types of effective models for crystals are tight-binding and k · p models.
In tight-binding models, the basis of states is defined in terms of a set of Wannier func-
tions that are exponentially localized in space; it is always possible to obtain such a set of
1Notice that in this description I refer to the connections in the Brillouin zone introduced by Melvin Lax [83].They should not be confused with the connections related to Lax pairs introduced by Peter Lax.
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 57
functions for a block of electronic bands with vanishing total Chern number that do not cross
others [85, 86]. The Hamiltonian and the Lax connection are respectively expressed in terms
of hopping parameters and dipole matrix elements. The hopping parameters can be inferred
from bandstructure properties, obtained either from experiments or from first-principle calcu-
lations. In contrast, the Lax connection parameters are harder to deduce since they are usually
obtained from electronic and optical properties. When the Wannier functions are well local-
ized, the overlap between them – and consequently the matrix elements for any operator –
can be restricted to only nearest neighbor atomic sites; the model is then usually simple and
has relatively few parameters that need to be inferred. However, if the Wannier functions at
sites further apart have a considerable overlap, the number of free parameters increases signif-
icantly. While this is not a major problem for determining hopping parameters, it leads to a
large number of dipole parameters that are hard to fit.
In k · p models, the basis of states consists of the periodic parts u`q (r) of Bloch wavefunc-
tions ψ`q (r) = eiq·ru`q (r) /√
(2π)D for a set of bands ` at a reference point q in the Brillouin
zone (BZ) of dimension D. Since the basis is independent of the lattice momentum k, the Lax
connection is null for a k · p model, which simplifies the calculation of electronic and optical
properties. However, k · p models also have drawbacks. For instance, the Hamiltonian has
a fixed form that is quadratic in the lattice momentum k, but its free parameters are only as-
sociated with the linear terms in k, as the quadratic term is related to the electron bare mass.
Because of that, the only way to introduce more parameters in the Hamiltonian is to increase
the number of bands in the model, even if the additional bands are irrelevant except for aiding
in the fitting of the band energies of interest. Also, since the periodic functions depend on k,
the basis needs to include the states of several bands at the reference q point in order to span the
state of a single band at other k points in the BZ. Thus k ·p models for the whole Brillouin zone
usually include several bands, but describe only a few of them accurately, a fact that increases
the number of parameters to be inferred. Moreover, the accuracy of the states [84, 87, 88] at
a point k in the BZ decreases with the distance from the reference point q, and since results
are usually reported without a standard measure of the error, it is not possible to know exactly
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 58
where the approximation becomes unacceptable.
In this chapter I develop an effective model for stanene that is similar to a k·p model but that
is free of the drawbacks pointed out in the previous paragraph. We keep track of the accuracy
of the eigenstates, and the free parameters of the Hamiltonian are not restricted to the linear
terms in the lattice momentum k. Starting from an ab initio set of wavefunctions, we expand
the eigenstates at a region of the BZ in terms of the states at a reference point q in that region.
For a finite set of bands, this expansion is not unitary, as the basis set is incomplete. In order
to preserve unitarity, we approximate this expansion by a unitary transformation [89] using
a singular value decomposition (SVD), the singular values of which provide a measure of the
accuracy of the eigenstates. This transformation allows the same basis to be used for a region of
the BZ, so the Lax connection is null as desired. A Taylor expansion of the Hamiltonian matrix
written in this basis with respect to the lattice momentum k then gives the free parameters
of our model. For stanene we use three regions in the BZ, around the points K, K′, and
Γ. We obtain an effective model that is accurate for transition energies up to 1.1 eV, with a
quadratic expansion for each reference point. We find that the band warping is well accounted
for by a quadratic model, and that a cubic model does not improve upon it significantly. We
also find that neglecting some small parameters leads to the separation of the spin sectors
in our model; such approximation is accurate within a tolerance corresponding to the room
temperature energy.
To illustrate the applicability of our model, we compute the one-photon injection rate coef-
ficients for carrier and spin densities in stanene. We predict that an incident circularly polarized
optical field with photon energy close to the gap only excites electrons with spins that match
the helicity of the optical field. This result suggests the possibility of employing stanene in
optically-controlled spin pump applications.
The outline of this chapter is as follows: In Sec. 3.2 I present the procedure to obtain the
effective model; in Sec. 3.3.2 I apply it to stanene and analyze the accuracy of the eigenstates
and the eigenenergies, including the band warping. In Sec. 3.4 I use our model to compute
linear optical absorption rates of stanene. I end with a discussion of our results in Sec. 3.6.
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 59
3.2 Method for deriving effective models
Bloch’s theorem asserts that the eigenstates ψ`k (r) of a periodic Hamiltonian functionH (r,−i~∇) =
H (r + R,−i~∇), where R is a lattice vector, can be written as
ψ`k (r) =1√
(2π)Deik·ru`k (r) , (3.1)
where u`k (r) = u`k (r + R) are periodic functions. In typical ab initio calculations, a very large
number of basis functions uak (r), which usually consist of plane waves or atomic orbitals, are
used to specify the Bloch HamiltonianH (r,−i~∇ + ~k) by the matrix elements
Habk = 〈uak|Hk|ubk〉
≡ Ω−1uc
∫uc
dr u∗ak (r)H (r,−i~∇ + ~k) ubk (r) , (3.2)
where Ωuc is the volume of the unit cell. The Hamiltonian matrix Hk consisting of these
elements is then diagonalized, and provides the eigenstates and eigenenergies corresponding
to each electronic band ` at the lattice momentum k. We denote the diagonalized matrix by
Hk and its corresponding eigenstates by |u`k〉. If the large set of basis functions in the ab initio
calculation are taken to be the same for different lattice momenta, say q and k, we can compute
the overlap matrix between states,Wk;q, with matrix elements
Wm`k;q = 〈umq|u`k〉 = Ω−1uc
∫uc
dr u∗mq (r) u`k (r) . (3.3)
The overlap matrix allows us to decompose the states u`k (r) at k in terms of those at the
reference point q in the BZ and to use the states umq (r) as a basis for any k point in the region
of the BZ around q. In order to have a simple effective model, it is desirable to include only
a small number of bands in the basis set. However, if only a few functions umq (r) = 〈r | umq〉
are included in the basis, even the states umk (r) = 〈r | umk〉 corresponding to the same block
of bands at other k point in the BZ neighborhood might not be completely spanned by them.
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 60
This means that the overlap matrixWk;q might not be unitary when restricted to a small set of
bands. Here we ensure the unitarity of the model by replacingWk;q with a unitary matrix based
on its “singular value” decomposition (SVD). In the remaining of this discussion we drop the
subindex indicating the reference q point in the BZ where it does not lead to confusion. In its
singular form, the overlap matrixWk is written as
Wk = UkΣkV†k , (3.4)
where Uk and Vk are orthogonal matrices, and Σk is a diagonal matrix with its elements as the
singular values. For a complex matrix W, such decomposition always exists, U and V are
unitary (i.e., UU† = I = UU†), and the “singular values” are guaranteed to be positive (for a
proof, see for example, Sec 2.5.2 and 2.5.6 of [90]). IfWk were a unitary matrix, Σk would
be the identity matrix I, thus a simple “unitary approximation” toWk is to replace Σk with the
identity matrix as
Wk → Wk ≡ UkV†k . (3.5)
An obvious measure for the accuracy of this approximation is the difference I− Σk. For each k
in a region around the reference q point in the BZ, the approximate unitary overlap matrix Wk
allows the expansion of the states |u`k〉 in terms of the basis |umq〉 as
|u`k〉 =∑
m
Wm`k∣∣∣umq
⟩. (3.6)
The next step is to use the above equation to write the Hamiltonian matrix Hk for each k in
terms of the states∣∣∣umq
⟩at the reference q point in the BZ. Since |u`k〉 are the eigenstates of the
(diagonalized) Hamiltonian matrix Hk, then
Hm`k = 〈umk| Hk |u`k〉 = δm`E`k. (3.7)
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 61
We write the elements of the Hamiltonian matrix for lattice momentum k expressed in the∣∣∣u`q⟩
basis as
Hm`k =⟨umq
∣∣∣Hk∣∣∣u`q⟩ , (3.8)
and using Eq. (3.6), the matrix Hk is related to Hk through the unitary matrix Wk that performs
the change of basis
Hk = WkEkW†
k , (3.9)
where Ek is a diagonal matrix with diagonal elements E`k. Since the basis of states ∣∣∣umq
⟩ is
independent of k, its Lax connection vanishes, ξm`k ≡ i⟨umq
∣∣∣∇k∣∣∣u`q⟩ = 0. Consequently, such
a basis is suitable for expanding the Hamiltonian matrix Hk around q simply as
Hk = Hq + κ · ∇kHk∣∣∣k=q + O
(κ2
)+ O
(κ3
). . . , (3.10)
where κ = k−q. If the basis were dependent on the lattice momentum k, the expansion would
include a correction given by the Lax connection.
In summary, the overlap matrixWk from an ab initio calculation is replaced by its unitary
approximation Wk, the diagonalized Hamiltonian is written in a basis that is independent of the
lattice momentum k, and a Taylor expansion of its matrix elements gives the free parameters
in our model. We now turn to discuss the application of this procedure to stanene.
3.3 Effective model for stanene
3.3.1 First-principles ground state of stanene
We start by obtaining the electronic wavefunctions from a first-principles calculation, in the
framework of Density Functional Theory (DFT) and the Local Density Approximation (LDA),
using the freely available ABINIT code [91, 92]. The wavefunctions are expanded in a ba-
sis of planewaves; the size of the basis is determined by a kinetic-energy cutoff of 653 eV
(≈ 24 Ha), corresponding to 6166 planewaves. The crystal (ionic) potential is modeled us-
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 62
z
x b
y
x a
a2
a1
δ1
δ2
δ3
= A-site = B-site
Figure 3.1: Hexagonal lattice of stanene with Sn atoms at A and B sites. The lattice vectorsare denoted by a1 and a2, and we show the unit cell with a gray rhombus. The interatomicdistance projected on the plane is a, and along the vertical direction it is b, so the interatomicdistance is
√a2 + b2.
ing the Optimized Norm-Conserving Vanderbilt Pseudopotentials (ONCVP) [93], which is a
modern alternative for pseudopotential representations that provides a compromise between
computational efficiency and accuracy for the current demands in automated, high-throughput,
computational materials science research2. We take 14 out of the 50 Sn electrons as valence
electrons, and the others are assumed clamped. We converge the ground-state total energy up
to 2.7 meV, leading to a 12 × 12 k-point mesh. Since we simulate the Sn monolayer with a
supercell model, we introduce an interlayer vacuum space of 11.42 Å, such that spurious inter-
layer interactions are negligible; with this amount of vacuum space, the total energy remains
unchanged within 2.7 meV if the vacuum space is incremented.
The relaxation of the atomic positions leaves the atoms at the (x, y) coordinates of a hon-
eycomb lattice, i.e., one Sn atom at (0, 0) and another at (a1 + a2) /3, see Fig. 3.1. The lattice
2Other pseudopotential representations like the “ultra-soft pseudopotentials” [94] or the “projected-augmentedwaves” [95, 96] method are standard alternatives, but they attain efficiency and accuracy at the expense of mathe-matical complexity of the inherent quantities in the pseudopotential method [93].
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 63
0
0.2
0.4
0.6
0.8
1
1.2
0 2 4 6 8 10 12
zcoordinate(A
)
BFGS Iteration
Initial coordinate of site-Bno d-electrons z00
z01
z02
z03
4.5
4.6
4.7
0 2 4 6 8 10 12
Lattice
constan
t(A
)
BFGS Iteration
Initial lattice constantno d-electrons c01
c02
c03
c04
c05
Figure 3.2: Relaxation of the atomic coordinates. The xy coordinates remain at the usual equilibriumpositions (not shown here) for hexagonal 2D lattices with two atoms in the unit cell: in the basis of thereciprocal vectors, site A is at (0, 0) and site-B is at (1/3, 1/3). In contrast, the z coordinates convergeto out-of-plane positions. In this Figure we show the evolution of the z coordinates of the atom atsite-B during the structural relaxation (atoms at sites-A follow the same). Each curve represents theevolution for different starting positions. All of these curves converge to the same final value of 0.418 Å,except when we neglect the d-electrons. Indeed, as shown in both panels, when the d-orbitals of Sn areassumed as core states (black curves), the out-of-plane atomic coordinates and the lattice constant areunderestimated. That is, for stanene, the d-electrons must be considered as semicore states due to thespatial extension of these orbitals (not shown). The structural relaxation of atoms at sites-A is mirrored,with negative values of z, hence the buckling distance is 0.836 Å. The relaxation is done using theABINIT code, employing its Broyden-Fletcher-Goldfarb-Shanno (BFGS) built-in routine.
vectors are a1 = a(3x +
√3y
)/2 and a2 = a
(3x −
√3y
)/2, where a = 2.66 Å is the inter-
atomic distance projected on the plane. In contrast, the relaxation of the z-coordinates leads to
out-of-plane coordinates ±0.418 Å, giving rise to a “buckling distance” of b = 0.836 Å, such
that the interatomic distance is√
a2 + b2. This small buckling has been shown to enhance the
overlap between π and σ orbitals, leading to an equilibrium configuration in materials where
the π-π bonding is relatively weak [78, 97]. The evolution of the relaxation process of the
atomic z-coordinates and the lattice constant a is shown on Fig. 3.2, where we clearly see that,
for stanene, the d-orbitals must be considered as semi-core states, due to their relatively long
spatial extension (not shown). The first nearest neighbor vectors are δ1 = a2 (x +
√3y) + bz,
δ2 = a2 (x −
√3y) + bz and δ3 = −ax + bz. In Fig. 3.1, we show the crystal lattice of stanene.
With these structural parameters we proceed to obtain the key ingredient in density-functional
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 64
DFT Charge Density n(x, y, z) (e−/Bohr3)(at z = 0)
-6 -4 -2 0 2 4 6
x (Bohr)
-6
-4
-2
0
2
4
6
y(B
ohr)
0
0.05
0.1
0.15
0.2
0.25
-6-4
-20
24
6 -6-4
-20
24
6
0.120.140.160.180.20.220.240.26
DFT Charge Density n(x, y, z) (e−/Bohr3)(at z = 0)
x (Bohr)
y (Bohr)
0.120.140.160.180.20.220.24
Figure 3.3: ab-initio DFT charge density of stanene. The highest values of charge density occur at thelocations of the atoms, hence these maximal values map the hexagonal lattice of stanene (cf. Fig. 3.1).The tin atoms are modeled with an ONCVPSP pseudopotential [93] (see main text), wich accounts forthe s, p and d electronic orbitals. The inclusion of the d-electrons do have an influence on the finalstructural parameters (lattice constant and atomic coordinates), cf. Fig. 3.2.
theory, the self-consistent electronic charge density n(x, y, z). On Fig. 3.3 I show two represen-
tations of the contour levels of n(x, y, z = 0), where we clearly identify the honeycomb lattice
of stanene. The maxima of n(x, y, z = 0) corresponds to the locations of the tin atoms. Once
the self-consistent charge density is obtained, then the Kohn-Sham Hamiltonian is diagonal-
ized one more time, maintaining the charge density and the Kohn-Sham potential unchanged.
This step is usually referred as the “non-self-consistent” calculation of the bandstructure (elec-
tronic dispersion relations). Such process give us the eigenstates and eigenenergies. The plot
of the eigenenergies along a path that connects crystal wavevectors of high symmetry in the
“irreducible” part of the BZ (IBZ) is referred as the bandstructure. On Fig. 3.4 I plot the band-
structure of stanene along the typical M → Γ → K → M path in the IBZ for a hexagonal
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 65
-3
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
a)
c1
v1
v2
Γ
MK
Energy(eV)
0.7
0.8
0.9
1
b)c1
v1 v2
Σk
q
six-band model
0.7
0.8
0.9
1
M Γ K M
c)c1
v1
Σk
q
four-band model
Figure 3.4: a) The ab initio bandstruc-ture of stanene with the bands includedin our effective model highlighted. Theband gaps at K and K′ have a value of88 meV. At Γ, the minimal transition isat 0.472 eV (transition v1 → c1) and thesecond is at 0.808 eV (transition v2 →
c1). All bands are doubly (spin) degen-erate. The dashed (gray) bands are notdescribed by our model. b) The singularvalues (the elements of the diagonal ma-trix Σk, Eq. (3.4)) with q = Γ as the ref-erence point. The shaded area indicatesthe region where the singular values areall greater than 0.9, and the unitary ap-proximation Σk → I is acceptable. c)Same as b), but for q = K as the refer-ence point.
lattice. All the bands are spin degenerate since stanene has space inversion symmetries. How-
ever, SOC in stanene leads to a gap opening due to a removal of orbital degeneracy at the Dirac
point [78]. As shown in Fig. 3.4, the bandstructure of stanene has gaped Dirac cones at the K
and K′ points with a gap of 0.088 eV. At Γ, the first transition occurs at 0.472 eV and the next
one at 0.808 eV. At M the first transition is at 1.55 eV; hence I ignore that region of the BZ in
the effective model, as the focus of this work is on energies ranges up to 1.1 eV. Our effective
model contains only states with lattice momentum around the K, K′, and Γ points, it includes
six bands around the Γ point and only four bands around K and K′ points.
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 66
3.3.2 Evaluation of the effective model: accuracy of the approximation
Once the ab initio wavefunctions are computed, we proceed to obtain the overlap matrix
(Eq. (3.3)) between the periodic functions at the reference point q and the other points in its
neighborhood in the BZ; this is done for each of the regions of interest in the BZ, namely the
regions around the reference points K, K′, and Γ. The overlap matrices of the ab initio wave-
functions can be approximated by unitary matrices based on singular value decompositions
according to Eq. (3.5). In order to determine the region of the BZ where this approximation
is accurate, in Fig. 3.4 we plot the elements of the diagonal matrix Σk (the singular values)
for the reference points K and Γ; the results for the K′ point are similar to those of K. In
Fig. 3.4, we also highlight the regions where each element of Σk is greater than 0.9, which is
taken as our tolerance for the approximation in Eq. (3.5). Notice that the highlighted regions
encompass every point on the BZ where optical transitions with photon energies below 1.1 eV
are possible.
In order to have a measure of the accuracy of the states that is easier to be visualized, we
define a figure of merit
δΣ (k) = n−1√
Tr (Σk − I)2, (3.11)
where n is the number of bands included in the model. In Fig. 3.5 we present the figure of merit
δΣ (k) for the three regions of interest in the BZ. We notice that the error indicated by δΣ (k) is
lower than 5% for large neighborhoods around the reference points.
3.3.3 Hamiltonian matrices
Having established the regions where the approximation of the states is valid, we now turn to
the approximation of the Hamiltonian matrix. We expand the matrix elements of the Hamilto-
nian Hk directly as in Eq. (3.10), and report the results below. Since we use a basis independent
of the lattice momentum for the neighborhood of the BZ around each reference point, the Lax
connection is null for each of these neighborhoods, ξabk = 0.
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 67
M
q = K′
q = K
q = Γ
M
q = K′
q = K
q = Γ
M
q = K′
q = K
q = Γ
SVD error isolines1 %2 %3 %4 % Figure 3.5: (Color online) Figure of
merit of the unitary approximation Σk →
I , as defined by Eq. (3.11), for the threeregions of the BZ centered at the ref-erence points q = Γ,K,K′; each qis marked with black dots. The yel-low lines connecting all contiguous pairsof K and K′ points indicate boundariesof the first Brillouin zone (cf. inset ofFig. 3.4, a)).
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 68
K and K′ points
The valleys around the K and K′ points are similar in our model, so we present the matrices
associated with each of them together, and use the valley parameter τ = 1 to refer to K and
τ = −1 to refer to K′. At the K and K′ points, the wavefunctions have a predominant character
of pz orbitals located at an atom in the unit cell. So we can use si and σi to respectively
denote the Pauli matrices in the spin and sublattice sectors; here i = 0, x, y, z, as we adopt the
convention of denoting the identity as the zeroth Pauli matrix. In this notation, the Hamiltonian
is written in terms of the matrices si ⊗ σ j.
Up to linear order in the lattice momentum κ = k − q, where q = K,K′, we find
H(1)τκ =∆K (−τsz ⊗ σz + s0 ⊗ σ0) + ζ(1)
K as0 ⊗(κxσx + τκyσy
)− λ(1)
K a(κysx − κxsy
)⊗ σz, (3.12)
where in the second part of the first term (i.e., ∆K s0 ⊗ σ0) we add an energy shift ∆K such
that the top of the valence band is at zero energy. Equation (3.12) is the linear-order term
in the crystal wavevector κ of our effective Hamiltonian for stanene. As we will mention
shortly, the last term in equation (3.12) can be neglected due to the relatively small value of
the parameter λ(1)K . This simplification is in agreement with a simple estimate of the spin-
orbit interaction of a monolayer material, which indicates that the spin polarization of the
material is along the direction perpendicular to the plane; this is expected for mono-layered
materials with mirror symmetry about the plane [98]. The simplified version of equation (3.12)
is in agreement with the Kane-Mele model (equations (2)-(3) in [98]3), which predicts that,
at sufficiently low energies, the spin-orbit effects on mono-layered material with atoms in a
honeycomb lattice lead to the opening of a gap accompanied by the quantum spin Hall effect
(QSHE). More specifically, it is the term (∆Kτ)sz ⊗ σz that leads to a QSHE. Following the
procedure outlined by Kane-Mele [98], the QSHE in stanene could be confirmed by describing
the electronic states with a Hamiltonian model for the entire Brillouin Zone; the QSHE states
3Notice that equation (4) in [98] does not apply in the context of this chapter. Such term refers to a “Rashbacontribution” that arises when the mirror symmetry about the plane is broken, e.g., due to a substrate or a perpen-dicular electric field.
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 69
are commonly visualized solving such model for a strip (“ribbon”) geometry. More recently ,
Chamon et al [99] provided a general classification of all gap-opening relativistic masses (such
as the Haldane mass term) of two-dimensional Dirac Hamiltonians that involve spin, valley,
and sub-lattice contributions. The subset of such masses that lead (support) the SQH effect is
in agreement with the Kane-Mele model [98]4
The quadratic terms in κ are
H(2)τκ = − ζ(2)
K a2s0 ⊗
[τκxκyσx +
12
(κ2
x − κ2y
)σy
]− v(2)
K a2 |κ|2 s0 ⊗ σ0 + ϑ(2)K a2τ |κ|2 sz ⊗ σz (3.13)
+ η(2)K a2τ
[(κ2
x − κ2y
)sx − 2κxκysy
]⊗ σz,
where the values of the parameters are shown in Table 3.1. Neglecting the relatively small
parameters λ(1)K and η(2)
K leads to a separation of the spin subsectors, since without them H(1)τk
and H(2)τk do not have terms with sx and sy, the only matrices with cross-spin elements. The spin
separation is expected for lattices without buckling, and it indicates that the lattice buckling
can be neglected in calculations involving k close to the expansion point q.
The parameters v(2)K and ϑ(2)
K can also be neglected, and the three parameters ∆K , ζ(1)K and ζ(2)
K
are the only ones needed for our model to give band energies that match those from DFT within
a tolerance of room temperature energy. We nevertheless report the negligible parameters λ(1)K ,
η(2)K , v(2)
K and ϑ(2)K , because their physical significance can be identified with the help of a pz-
orbital tight-binding model, as we discuss in Section 3.5. Finally, we provide an analytical
expression for the band energies around the K and K′ points obtained from our effective model.
4Notice that Chamon et al [99] denote the Pauli matrices that act on different sublattice and valley states byτ and σ, respectively, whereas in this thesis and in the Kane-Mele model [98] we use we σ and τ for sublatticesand valleys, respectively. Pauli matrices acting on spin states are denoted by s in these three works.
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 70
All values in eV
∆K = 0.044 ζ(1)K = 0.67 ζ(2)
K = 0.33
K λ(1)K = 0.03 v(2)
K = 0.03
K′ ϑ(2)K = 0.03
η(2)K = 0.02
EcΓ
= 0.37 ζ(1)Γ1 = 1.23 v(2)
Γc = 0.34
Γ Ev1Γ
= −0.10 ζ(1)Γ2 = 1.16 v(2)
Γ1 = 0.45
Ev2Γ
= −0.44 v(2)Γ2 = 0.34
ζ(2)Γv = 0.35
Table 3.1: Parameter values for the effectivemodel. The upper half of the Table lists the param-eters for the K and K′ valleys (Eqs. (3.12)–(3.13)).The parameters λ(1)
K , η(2)K , v(2)
K and ϑ(2)K can be ne-
glected without significant changes in the band en-ergies. Neglecting the parameters λ(1)
K and η(2)K alone
already leads to a separation of the spin subsectors.The parameter values for the Γ valley (Eqs. (3.15)–(3.16)) are listed in the lower half of this Table; neg-ligible parameters were also omitted.
Neglecting the small parameters mentioned in the previous paragraph, we have
E±τκ = ∆K ±
√∆2
K + X2κ +Y2
κ, (3.14a)
Xκ = aκx
(ζ(1)
K − τζ(2)K aκy
), (3.14b)
Yκ = ζ(1)K aκy −
12τζ(2)
K a2(κ2
x − κ2y
), (3.14c)
where the positive and negative signs of the square root correspond to the conduction and
valence bands respectively.
Γ point
At the Γ point, the wavefunctions cannot be easily associated with a sublattice, but they can
still be identified according to spin, so we continue using si to denote the Pauli matrices acting
on the spin sector of the Hilbert space. Up to linear order in the lattice momentum, here
κ = k − q = k since q = Γ, we find
H(1)Γκ = s0 ⊗
E(c)
Γ0 0
0 E(v1)Γ
0
0 0 E(v2)Γ
+ aκxs0 ⊗
0 ζ(1)
Γ1 ζ(1)Γ2
ζ(1)Γ1 0 0
ζ(1)Γ2 0 0
+ aκysz ⊗
0 −iζ(1)
Γ1 iζ(1)Γ2
iζ(1)Γ1 0 0
−iζ(1)Γ2 0 0
,(3.15)
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 71
while the quadratic terms in κ are
H(2)Γκ = 1
2a2 |κ|2 s0 ⊗
v(2)
Γc 0 0
0 −v(2)Γ1 0
0 0 −v(2)Γ2
+ 12a2
(κ2
x − κ2y
)s0 ⊗
0 0 0
0 0 ζ(2)Γv
0 ζ(2)Γv 0
+a2κxκysz ⊗
0 0 0
0 0 iζ(2)Γv
0 −iζ(2)Γv 0
.(3.16)
The values of the parameters are presented in Table 3.1. Here we have omitted negligible
parameters. The parameters reported constitute the minimum set necessary to describe the
energies with an accuracy equivalent to room temperature when compared to the bands from
DFT. Notice that the model for the valley at the Γ point can also be separated in two spin
sectors.
3.3.4 Accuracy of the energies
The accuracy of the Taylor expansion of the Hamiltonian matrices in the previous subsection
can be determined by comparing the band energies obtained from our model with those from
the ab initio calculation. In Fig. 3.6 we present the band energies obtained from models in-
cluding first-, second-, and third-order expansions of the Hamiltonian on the lattice momentum
difference κ; third-order expansions are not discussed further in this work. We also show the
ab initio bands for comparison, and focus on the regions where the approximation for the states
is accurate as discussed in Sec. 3.3.2. From Fig. 3.6, we see that keeping the cubic terms in
the Hamiltonian expansion is unnecessary to reproduce the ab initio band energies around the
Γ point, while for the region around the K point (and equivalently the K′ point) it is actually
detrimental to go beyond the second-order expansion.
A plot of band energies along a simple path through a region of the BZ is not enough to
establish the accuracy of the bands from our model in that entire region. Analyzing the band
warping is a way to ensure that the good agreement displayed in Fig. 3.6 is not coincidental to
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 72
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
M Γ K M
c1
v1
v2
E1
Energy
(eV)
REF1st2nd3rd
E2
E3
E4
E5
Figure 3.6: Comparison of the band energies obtained from the Taylor expansions of theHamiltonian matrices (dashed lines) in Sec. 3.3.3 with those from the ab initio calculation(continuous gray line). The band energies from the effective model are plotted only in theshaded regions, where approximation for the states is accurate as discussed in Sec. 3.3.2.
the directions associated with that plot. In Fig. 3.7 we show isoenergy lines for each relevant
band obtained from our model and those from the ab initio computation. The latter are shown
as pairs of lines that enclose an energy range equivalent to room temperature, which is taken as
our tolerance for energy accuracy. We compare the band warping corresponding to expansions
of the Hamiltonian that are quadratic and cubic on the lattice momentum difference κ; on
Fig. 3.7 we show that the cubic expansion does not improve upon the quadratic one. Thus we
confirm that the quadratic expansion provides the best model for the bandstructure of stanene
for excitation energies up to 1.1 eV.
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 73
|KM|4
c1 at E3
Γ
|KM|4
c1 at E5
K
|KM|4
v1 at E2
Γ|KM|
4
v1 at E4
K
|KM|4
v2 at E1
Γ
Figure 3.7: Band warping of the relevant bands around the reference points in the BZ inour model. Red-dashed and green-dotted lines indicate respectively second and third orderexpansions of the Hamiltonian matrices. The thin gray lines are ab initio energy isolines thatenclose a range of energy equivalent to room temperature. To give a sense of proportion weinclude a line segment of length one fourth of the distance KM. The isolevels Ei and bandlabels vi and c1 are as indicated in Fig. 3.6.
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 74
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1
hω (eV)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
ξxx(ω
)(1/V
2s)×
1016
a)
BZ Region
FullK + K′
Γ
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1
hω (eV)
0.00
0.05
0.10
0.15
Reσxx(ω
)(m
A/V
)
b)
BZ Region
FullK + K′
Γ
Figure 3.8: Linear optical absorption properties computed with the effective model ofSec. 3.3.2. a) One-photon absorption coefficient ξxx (ω) and b) real part of the optical con-ductivity σxx (ω) of stanene. The contributions from the regions around the K and K′ points(dot-dashed red line) and the Γ point (dashed black line) in the BZ are shown separately, alongwith that from the full BZ (solid blue line); we stress that the “full” signal indeed contains con-tributions from all crystal momenta k around K(K′) and Γ for which one-photon transitionsless than 1.1 eV are possible; consequently, it is equivalent to a full BZ calculation, within thelimits of validity of our model.
3.4 Linear and non-linear optical properties
The optical properties of a crystalline system depend only on the Hamiltonian matrix and the
Lax connection [100]. Since the Lax connection is null in the basis of our model ξabk = 0,
the velocity matrix elements are simply given by v (k) = ~−1∇kH (k). We consider the optical
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 75
injection rates of carrier and spin densities, given by
ddt
n = ξab (ω) Ea (ω) Eb (−ω) , (3.17)
ddt
S z = ζzab (ω) Ea (ω) Eb (−ω) , (3.18)
where we use the convention of summing repeated indices, E (t) = E (ω) e−iωt + c.c. is an inci-
dent optical field, and the tensors ξab (ω) and ζzab (ω) are the carrier and spin density injection
coefficients
ξab (ω) =2πe2
~2ω2
∑cv
∫d2k
(2π)2 vacvkvb
vckδ (ω − ωcvk) , (3.19)
ζzab (ω) =2πe2
~2ω2
∑cv
∫d2k
(2π)2
(S z
cc − S zvv)
vacvkvb
vckδ (ω − ωcvk) , (3.20)
where v and c are respectively valence and conduction band indices, e = −|e| is the electron
charge, vacvk are the velocity matrix elements, S z
cc = ±~/2 and S zvv = ∓~/2 are the spin-z matrix
elements of respectively the conduction and valence bands, and ~ωcvk = ~ωck − ~ωvk are band
energy differences. In numerical calculations, we approximate the Dirac delta function in the
above equations by a Lorentzian function with a broadening width of 6 meV.
In Fig. 3.8, we present plots of the linear optical absorption coefficient ξxx (ω) and the
real part of the optical conductivity Reσxx (ω), which are related to each other by ξxx (ω) =
2Reσxx (ω) / (~ω). As the frequency increases, the absorption begins at the band gap energy
0.088 eV due to electronic transitions at the K and K′ valleys in the BZ. The contribution
from Γ has an absorption onset at 0.472 eV, and a second absorption onset at 0.808 eV, when
electronic transitions from the second valence band are allowed.
For photon energies close to the band gap, stanene has the interesting property that circu-
larly polarized light excites mostly electrons with the spin that matches its helicity. Similar
characteristics have been identified and studied in other monolayers, such as silicene [101].
This feature can be seen from our linear model for the K and K′ points in Eq. 3.12, which
can be separated in spin sectors, and the expressions of ξ (ω) and ζ (ω) for a Dirac cone [6, 7].
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 76
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1
hω (eV)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
h−
1Im
ζzyx(ω
)(1/V
2s)×
1016
Im ζzyx = −Im ζzxy a)
BZ Region
FullK + K′
Γ
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1
hω (eV)
−1.0
−0.5
0.0
0.5
1.0
Sz/(hn
)
−Im ζzyx
hRe ξxx
b)
BZ Region
FullK + K′
Γ
Figure 3.9: Spin-density injectioncomputed with the effective modelof Sec. 3.3.2. a) Spin densityinjection coefficient ζzyx (ω) and b)Spin polarization of injected carri-ers with circularly polarized light−Im ζzyx (ω) /
[~ξxx (ω)
]for stanene,
with h = 1. The contributions fromthe regions around the K and K′points (dot-dashed red line) and theΓ point (dashed black line) in the BZare shown separately, along with thatfrom the full BZ (solid blue line), inthe sense described in the caption ofFig. 3.8.
For circular polarizations, the light field propagating along the z direction can be written as
E (ω) = Eωph, where h = ±1 is the helicity, and ph = (x + ihy) /√
2. Then the expression for
each spin s = ±1 is
ξhτs (ω) =Θ (ω − 2∆K) e2
8~2ω
(1 + hs
2∆K
ω
)2
, (3.21)
which is independent of valley, and where Θ (x) is the step function, valued as zero or unity
if x < 0 or x > 0, respectively. From Eq. (3.21) we see that the spin polarization is maximal
for photon energies corresponding to the gap, and it decreases for larger photon energies. The
injection coefficient of an arbitrary quantity for circularly polarized light, ηh (ω), is given in
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 77
terms of its Cartesian components as
ηh (ω) =12
[ηxx (ω) + ηyy (ω)
]+ ih
12
[ηyx (ω) − ηxy (ω)
]= ηxx (ω) + i h ηyx (ω) , (3.22)
where the relations ηxx (ω) = ηyy (ω) and ηyx (ω) = −ηxy (ω) due to the symmetries of a buck-
led honeycomb lattice were used. For carrier and spin densities in stanene, we also have
ξxx (ω) = Re ξxx (ω) and ξyx (ω) = 0, as well as ζzxx (ω) = 0 and ζzyx (ω) = i Im ζzyx (ω). So the
coefficients for circular polarizations are simply ξh (ω) = Re ξxx (ω) and ζh (ω) = −h Im ζzyx (ω).
We present plots of the spin density injection coefficient ζzxy (ω) computed with our effective
model in Fig. 3.9 a), which shows the same frequency regimes discussed for ξxx (ω). In Fig. 3.9
b) we show the spin polarization of injected carriers for circularly polarized light. Even for ex-
citations at the Γ valley there is still a helicity-spin coupling, although the net spin polarization
is partially canceled by the excitations at the K and K′ valleys.
We note that helicity-spin coupling is due to the sign of the mass term ∆K in each Dirac
cone [102–104], which also explains why stanene shows the spin Hall effect. We also point
out that the helicity-spin coupling in stanene is analogous to the helicity-valley coupling in
TMDs [105].
As an example of application of our model to study non-linear optics, now we present
results for the two-photon absorption (2PA) rate, defined by
ddt
n2 (2ω) =ξabcd (2ω) Ea (ω) Eb (ω) Ec (−ω) Ed (−ω) , (3.23)
where ξabcd (2ω) is the two-photon absorption coefficient, expressed as
ξabdg (2ω) =2πe4
~4ω4
∑cv
∫d2k
(2π)2
∑c′
vacc′v
bc′v
ω − ωc′v−
∑v′
vbcv′v
av′v
ω − ωcv′
∗ × ∑c′′
vdcc′′v
gc′′v
ω − ωc′′v−
∑v′′
vgcv′′v
dv′′v
ω − ωcv′′
× δ (2ω − ωcv) , (3.24)
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 78
0.0 0.2 0.4 0.6 0.8 1.0 1.2
2hω (eV)
−50
0
50
100
150
200
250
300
350
ξabcd
2(2ω
)(m
2/V
4s)
a)
K + K′ + Γregions
xxxx
xxyy
xyxy
0.4 0.6 0.8 1.0 1.2−0.10
−0.05
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.0 0.2 0.4 0.6 0.8 1.0 1.2
2hω (eV)
−0.02
0.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
ξabcd
2(2ω
)(m
2/V
4s)
only Γregion
b)
xxxx
xxyy
xyxy
Figure 3.10: Two-photon absorptionof stanene. On panel a) we showthe contribution from k-points in theneighbourhood of the three specialpoints Γ, K and K′. The inset is anamplification around the two-photonenergy range where the Γ point con-tributes to this responce. On panelb) we present solely the contributionsfrom k-points around Γ.
where all quantities are as defined below Eq. (3.19). From symmetry considerations, we find
that the only nonzero components of this coefficient are ξxxxx, ξxxyy and ξxyxy. On Fig. 3.10
we present the evaluation of ξabcd (2ω). As expected, the onset of these coefficients start when
the two-photon energy 2~ω equals the fundamental bandgap Egap, which occurs at K. The
contributions from k-points in the neighbourhood of K dominate the 2PA signal, whereas the
contributions arising form the Γ point are negligible as compared to those from K. On the
inset of Fig. 3.10-a) we present a zoom-in over the photon energy range over which the two
band edges at Γ occur (cf. Fig. 3.4). To clarify further the contributions from the Γ point, on
Fig. 3.10-b) we evaluate the 2PA coefficients considering solely k-points near Γ. From both
panels a) and b) we clearly identify two onsets at 0.472 eV and at 0.808 eV, arising from the
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 79
two band-edges at Γ (i.e., transitions from the two top valence bands to the first conduction
band). Although the relative magnitude of the 2PA from Γ is much smaller than that from
K(K′), the relevance of the former in an experimental or applied setting would depend on the
photon energy of the applied light field.
3.5 Tight binding model
Tight-binding models (TBM) have successfully been used to describe electronic states in the
full Brillouin zone (BZ) in different monolayer materials, such as silicene, germanene and
stanene [106, 107]; the description of the full BZ usually requires the inclusion of basis sets
with s, px, py and pz orbitals. In this section we discuss a TBM that includes only pz orbitals
on a buckled honeycomb lattice, which is enough to describe the states around the K and K′
points in the BZ, and in that region it agrees with the models of Liu [106] and Ezawa [107].
This TBM is unable to describe the states at the Γ point because their orbital character are not
purely pz. For instance, from a DFT calculation, we find that at the Γ point the orbital character
of the first conduction band is 73% s, 24% pz and 3% d, while that of the top valence band is
96% a mix of px and py, and about 4% d character.
This basis is not in Bloch’s form5, and it allows us to write all the hopping parameters in
terms of the nearest neighbor vectors δn instead of the lattice vectors an. Using the notation
employed in the main text, the Hamiltonian is written in terms of the matrices si ⊗ σ j. With
these conventions and employing a usual tight-binding framework [55], the nearest-neighbour
(NN) hopping term in the Hamiltonian is
HNNk = −t
3∑n=1
s0 ⊗
0 e−ik·δn
eik·δn 0
= −t
3∑n=1
s0 ⊗[cos (k · δn)σx + sin (k · δn)σy
],
(3.25)
5When the periodic functions of a basis satisfy the condition u`k+G (r) = e−iG·ru`k (r) , where G is a reciprocallattice vector and ` is a band index, the Bloch wavefunctions are periodic over the Brillouin zone, φ`k+G (r) =
φ`k (r) , and the basis is said to be in Bloch’s form.
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 80
and the next-nearest-neighbor (NNN) term is
HNNNk = −t′
∑n,m,n
[eik·(δm−δn) + e−ik·(δm−δn)
]s0 ⊗ σ0
= −2t′∑
n,m,ncos (k · (δm − δn)) s0 ⊗ σ0,
(3.26)
without the spin-orbit coupling. The spin-orbit coupling changes the next-nearest-neighbor
hopping matrices according to
t′s0 ⊗ σ0 → t′s0 ⊗ σ0 + iλS O
(δm × δn
)· s ⊗ σz, (3.27)
where λS O is the spin-orbit coupling parameter. The last term in the above equation can be
further separated in two parts by decomposing the δm × δn vector as
λS Oδm × δn = λzz + λbz × (δm − δn) , (3.28)
where
λz =a
√a2 + 4b2
λS O, λb =2b
√a2 + 4b2
λS O, (3.29)
according to the lattice buckling; the lattice parameters a and b are depicted in Fig. 3.1.
In order to compare the tight-binding model with the one described in Sec. 3.3, we now
perform an expansion in powers of κ around the K and K′ points in the BZ, to which we re-
spectively associate τ = +1 and τ = −1. Applying a further change of basis to the B sublattice,
uBk (r)→ ieiK·δ3uBk (r), the linear term is
H(1)τκ = −τ 9
2λza2sz ⊗ σz + 32 tas0 ⊗
(κxσx + τκyσy
)+3t′s0 ⊗ σ0 −
92λba3
(κysx − κxsy
)⊗ σz,
(3.30)
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 81
where the term 3t′s0 ⊗ σ0 is simply an energy shift and can be removed. The quadratic term is
H(2)τκ = −3
4 ta2s0 ⊗[τκxκyσx + 1
2
(κ2
x − κ2y
)σy
]−9
4 t′a2κ2s0 ⊗ σ0 +(
32
)3λza4τκ2sz ⊗ σz
+(
32
)3λba4τ
[(κ2
x − κ2y
)sx − 2κxκysy
]⊗ σz.
(3.31)
Now we compare this tight-binding model described by Eqs. (3.30)-(3.31) to our effective
model around the K and K′ points in the BZ described by Eqs. (3.12)-(3.13). The relations
between the respective first order parameters are
∆K =92λza2, ζ(1)
K =32 t, λ(1)
K =92λba2, (3.32)
and for the second order ones, we have
ζ(2)K =3
4 t, v(2)K =9
4 t′, ϑ(2)K =
(32
)3λza2, η(2)
K =(
32
)3λba2, (3.33)
Since λb = 2bλz/a, we can take t, t′ and λz to be the only independent parameters of the
tight-binding model; numerical values for them can be obtained from Table 3.1. Consequently,
λ(1)K =
2ba
∆K , ζ(2)K =
12ζ(1)
K , ϑ(2)K =
34
∆K , η(2)K =
3b2a
∆K . (3.34)
This tells us that ∆K , ζ(1)K and v(2)
K can be taken as the only independent parameters in Table 3.1,
just as the 3 independent parameters for the tight-binding. The parameters t′ and v(2)K can be
neglected, though, so the relevant parameters are only two: t and λz for tight-binding, and ∆K
and ζ(1)K in our effective model. The relations above are satisfied by the parameters shown in
Table 3.1.
Chapter 3. An EffectiveModel for the Electronic and Optical Properties of Stanene 82
3.6 Summary and discussion
We have presented an effective model that accurately describes the electronic and optical prop-
erties of stanene for low photon energies. We started from an ab initio calculation of the
bandstructure of stanene, which allowed us to identify the parameters in the model. Our model
includes a minimum set of energy states: 6 bands around the Γ point in the BZ, and 4 bands
around the K and K′ points. We provided measures for the accuracy of the approximations for
states and for energies, so we can identify the range of validity of the model.
We found that a quadratic model with respect to the lattice momentum is the best suited for
calculations based on the bandstructure. Even the band warping from DFT calculations is better
reproduced by the quadratic rather than a cubic model. We also found that the lattice buckling
can be neglected. This is confirmed by verifying that a separation of the states according to
spin-z subsectors is a good approximation for the band energies. In Section 3.5, we discuss the
physical significance of some parameters in our model by comparing it to a pz-orbital tight-
binding model expanded around the K and K′ regions of the BZ. Finally, we illustrated the
applicability of the model by computing linear optical absorption rates of stanene. We high-
lighted the coupling of circularly polarized light with the electronic spin, which underscores
the potential of stanene for optical-spintronic applications.
The model proposed here can accurately describe optical properties of stanene up to pho-
ton energies of 1.1 eV, which is suitable for a wide range of optical experiments. Compared
with a usual k · p method, our model requires fewer parameters to describe the bandstructure;
we also provide a figure of merit to determine the portion of the Brillouin Zone where the ap-
proximation is sensible. We expect that this simple model will be useful in understanding and
suggesting experiments on this promising material, and that the procedure described here will
be used to extract effective models from ab initio calculations for other 2D materials.
Chapter 4
Coherent Control of Two- and
Three-photon Absorption in AlGaAs
Abstract
As described in Chapter 2, optical absorption in a semiconductor crystal can lead to carrier
and spin injection into specific regions of the bandstructure of a crystal, and the interference
of different pathways of absorption can lead to current injection of charge and spin carriers.
The magnitude and direction of such injected currents can be manipulated by controlling the
phases and polarizations of the optical fields that lead to their respective carrier injections. This
technique is to referred as Quantum Interference Control (QuIC), and can also been exploited
to measure the parameters of the incident fields.
In this Chapter I study QuIC technique involving two- and three-photon (“2+3”) absorp-
tion processes. As compared to QuIC of one- and two-photon (“1+2”) absorption processes
(Chapter 2), the 2+3 QuIC has different requirements for the laser frequencies and the opti-
cal processes occur in smaller regions of the Brillouin zone. Experiments demonstrating 2+3
QuIC current injection in AlGaAs are being finalized at by Prof. Steven Cundiff’s experimental
group at the University of Michigan. A modified version of this chapter is being prepared for
submission to the Physical Review B.
83
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 84
4.1 Introduction
Quantum interference between different optical processes arises when two optical beams of
different frequencies can lead to the same transition. In a crystal, amplitudes for different opti-
cal processes leading to electron-hole excitations can interfere constructively in some regions
of the Brillouin zone (BZ), and destructively in others. By controlling the polarizations and
phases of the incident fields, it is possible to excite carriers in selected localized regions of the
BZ. Such Quantum Interference Control (QuIC), using 1- and 2-photon absorption processes,
has been used for current injection in semiconductors [4, 108, 109], graphene [3, 9, 110], topo-
logical insulators [6, 111, 112], and transition metal dichalcogenides [7, 113], as well as spin
current injection in semiconductors [114–118]. It has also been theoretically investigated for
current injection in graphene nanoribbons [119], spin currents in topological insulators [6], and
spin and valley currents in transition metal dichalcogenides [7]. The currents injected via QuIC
have also been exploited to determine parameters of the optical fields responsible for their in-
jection [120–122]. This method has found application in the measurement [123, 124] and
stabilization of the carrier-envelope phase of a train of octave-spanning laser pulses [125–127].
In crystalline materials, every instance of QuIC studied to date has involved 1- and 2-photon
absorption processes.
In this chapter I present a theoretical study of QuIC with 2- and 3-photon processes in Al-
GaAs. The derivation of the expressions for the optical injection coefficients is presented, and
then I evaluate them for different stoichiometries of AlGaAs using a 30-band k·pmodel. I com-
pute all the symmetry-allowed injection coefficients corresponding to different polarizations of
the incident fields, and analyze their frequency dependence over a range where the injection of
carriers that do not contribute to the current is avoided. That is, considering 3-photon absorp-
tion of photons at energy ~ω and 2-photon absorption of photons at energy 3~ω/2, we require
2~ω to be less than the band gap. The alloy AlGaAs is an ideal material for 2+3 QuIC, as
its stoichiometry can be chosen to yield a band gap appropriate for the available laser wave-
lengths. The 2+3 QuIC process is of interest both because the use of all-nonlinear optical
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 85
processes leads to excitations even more localized in the BZ than does 1+2 QuIC, and because
2+3 QuIC could be used to stabilize the carrier-envelope phase of a train of laser pulses that do
not span an octave. Experiments demonstrating 2+3 QuIC of injection current in AlGaAs are
in progress [128].
The outline of this chapter is the following: In Sec. 4.2 I present the equations necessary
to compute the optical injection rates for a generic material. In Sec. 4.3 I describe the model
used for AlGaAs. In Sec. 4.4 I present our results for carrier and current injection from 2- and
3-photon absorption (2PA and 3PA) processes in AlGaAs. I also discuss the efficiency of the
current injection by analyzing the swarm velocity, and computing the optimal laser intensities.
In Sec. 4.5 I discuss the implications of our results and present our conclusions.
4.2 Optical injection rates
Assuming the independent particle approximation, I consider a system in the presence of the
external perturbationVext (t) described by a Hamiltonian
H (t) = H0 +Vext (t) , (4.1)
whereH0 is the Hamiltonian in the absence of any external perturbation. In the basis of eigen-
states ofH0,
H0 =∑nk
~ωnk a†nk ank, (4.2)
Vext (t) =∑mnk
a†mk (t) Vmnk (t) ank (t) . (4.3)
In the interaction picture, the creation and annihilation fermion operators are
a†nk (t) = a†nkeiωnt and ank (t) = anke−iωnt, (4.4a)
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 86
and the time-evolution operator can be expanded as
U (t) = 1 +
∞∑N=1
∫ t
t0
dtN
i~Vext (tN)· · ·
∫ t2
t0
dt1
i~Vext (t1) . (4.5)
The terms of each order inVext can be obtained from the previous one by
UN (t) =
∫ t
t0
dtN
i~
∑mnk
a†mkVmnk (tN) ankeiωmntNUN−1 (tN) , (4.6)
whereU0 (t) = 1, ωmnk = ωmk − ωnk, and Emk = ~ωmk is the bandstructure. We are interested
in the excitation of an electron from a valence band v to a conduction band c due to the external
field. This excited state is |cvk〉 = a†ckavk|gs〉, where |gs〉 is the ground state of H0 with filled
valence bands. The state of the system is described by
|ψ〉 = U (t) |gs〉 = γ0|gs〉 +∑cvk
γcvk (t) |cvk〉 + . . . , (4.7)
where the coefficients
γcvk (t) = 〈cvk|U (t) |gs〉 (4.8)
contain the information we want. The expectation value of the density 〈M〉 of a quantity
associated with an operator
M(t) =∑mnk
a†mk(t) Mmnk ank(t), (4.9)
due only to the excitation of states |cvk〉 is
〈M〉 =1
LD
∑cvc′v′k
γ∗c′v′kγcvk 〈c′v′k|M |cvk〉
=1
LD
∑cvc′v′k
(Mc′ckδv′v − Mv′vkδc′c) γ∗c′v′kγcvke−iωc′v′kteiωcvkt, (4.10)
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 87
where L is a normalization length, and D is the spatial dimension of the system. For a full
HamiltonianH (t) that follows from a Hamiltonian for a single particle of the form
H (x,p; t) =1
2m
[p − eA (t)
]2+ HSO
(x,p − eA (t)
)+ Vlat (x) , (4.11)
where x and p are position and momentum operators, HSO is the spin-orbit term, and Vlat (x)
is the lattice potential energy. Here we neglect a contribution to the interaction that is solely
a function of time (∼ [A (t)]2), for it will not lead to any transitions, and we work in a gauge
where the electric fieldE (t), assumed independent of position, is fully described by the vector
potentialA (t). The interaction term in the Hamiltonian and the velocity operator take the form
Vext (t) = −ev ·A (t) and v = −1e∂H
∂A, (4.12)
respectively, where e = − |e| is the charge of the electron. Indeed, the interaction is of the form
we consider for any unperturbed Hamiltonian for a single particle that is at most quadratic in
the momentum. Taking the vector potential as
A (t) =∑α
Aαe−i(ωα+iε)t = −∑α
iωα
Eαe−i(ωα+iε)t, (4.13)
with ωα = ±ω, ±3ω/2; here ε → 0+ describes turning on the field from t = −∞. The γcvk (t)
coefficients can be expanded as γ(N)cvk (t) = 〈cvk| UN (t) |gs〉 following the expansion ofU (t) as
given by Eq. (4.6) for an incident optical field, so we can write the coefficients γ(N)cvk (t) as
γ(N)cvk (t) = R(N)
cvk
e−i(ΩN−ωcv+iε)t
ΩN − ωcvk + iε, (4.14)
where ΩN = ω1 + . . . + ωN . The coefficients R(N)cvk involve the electric field amplitudes Eα
R(N)cvk = R(N)a...b
cvk
(ωα, . . . , ωβ
)Eaα . . . E
bβ (4.15)
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 88
where repeated indices are summed; here superscripts refer to Cartesian indices and subscripts
to incident frequency components. For the lower orders we have
R(1)acvk (ωα) =
∑α
ie~ωα
vacvk, (4.16)
R(2)abcvk
(ωα, ωβ
)=
∑αβ
−e2
~2ωαωβ
∑c′
vacc′kvb
c′vk
ωβ − ωc′vk−
∑v′
vbcv′kva
v′vk
ωβ − ωcv′k
, (4.17)
R(3)abdcvk
(ωα, ωβ, ωδ
)=
∑αβγ
ie3
~3ωαωβωδ
[X −Y −Z
], (4.18)
where
X =∑
c′
vacc′k
ωα − ωcc′k
∑c′′
vbc′c′′kvd
c′′vk
ωδ − ωc′′vk−
∑v′
vdc′v′kvb
v′vk
ωδ − ωc′v′k
(4.19)
Y =∑
v′
vav′vk
ωα − ωv′vk
∑c′
vbcc′kvd
c′v′k
ωδ − ωc′v′k−
∑v′′
vdcv′′kvb
v′′v′k
ωδ − ωcv′′k
(4.20)
Z =∑c′v′
vbcv′kva
v′c′kvdc′vk
(ωα − ωv′c′k) (ωδ − ωc′vk)+
vdcv′kva
v′c′kvbc′vk
(ωδ − ωcv′k) (ωα − ωv′c′k)
(4.21)
We are interested in the non-oscillatory response of the system, so we focus on the ΩN = ΩN′
contributions to Eq. (4.10). In order to compute the injection rate d 〈M〉 /dt associated with
that equation, it is important to realize that
ddt
(γ∗c′v′kγcvke−iωc′v′kteiωcvkt
)t=0,ε→0
=∑
ΩN=ΩN′
2π δ (ΩN − ωcv) R(N′)∗c′v′k R
(N)cvk
∣∣∣∣∣∣ωcv=ωc′v′
. (4.22)
The fact that the R(N)cvk coefficients are always accompanied by δ (ΩN − ωcv) in the expression
for the response allows for substitutions 3~ω − ωcv = 0; this is used to simplify R(3)abdcvk in
Eq. (4.18). The expression for the injection rate of 〈M〉 due to the interference of an N′ photon
process with an N photon process is
ddt〈M〉 = µabd...,pq... (Ω) Ea
−αEb−βE
d−δ . . . E
pρEq
σ . . . + c.c., (4.23)
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 89
3ω/2
ω
ω
ω
3ω/2
3ω/2
ω
ω
ω
3ω/2Figure 4.1: Depiction of 2+3 QuICshowing the destructive (left) andconstructive (right) interference indifferent regions of the Brillouinzone.
where there are N′ frequency labels (α, β, δ, ...) and N frequency labels (ρ, σ, ...). The injection
rate coefficient µabd...,pq... (Ω) is assembled from extracting the terms multiplying the electric
field amplitudes in R(N′)∗c′v′kR(N)
cvk of Eq. (4.22) together with the appropriate matrix elements
appearing in Eq. (4.10) for 〈M〉. I give examples below.
Quantum interference of 2- and 3-photon processes
The processes of 3PA with frequency ω and 2PA with frequency 3ω/2 can interfere since the
total frequency for each of them is Ω = 3ω. For such processes the frequencies are all equal
in the equations (4.17) and (4.18) for the coefficients R(2)cvk and R(3)
cvk, and symmetrizing their
components leads to some simplifications. Using ωβ = ωcvk − ωα and ωα = ωβ = 3ω/2 in
Eq. (4.17), the second order coefficient can be written as
R(2)abcvk
(3ω2,
3ω2
)=−4e2
9~2ω2
∑m
vacmkvb
mvk(3ω2 − ωmvk
) , (4.24)
and using ωβ +ωγ = ωcvk −ωα and ωα = ωβ = ωγ = ω in Eq. (4.18), the third order coefficient
can be written as
R(3)abdcvk (ω,ω, ω) =
ie3
~3ω3
∑mn
vacmkvb
mnkvdnvk
(ω − ωcmk) (ω − ωnvk). (4.25)
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 90
Notice that the denominators in Eqs. (4.24) and (4.25) are minimal for m, n = c, v, so the domi-
nant contributions to R(2)cvk always involve intraband velocity matrix elements, but R(3)
cvk also has
contributions from interband velocity matrix elements 1. Intraband velocity matrix elements
are associated with the corresponding band dispersion, vannk = ∂a
kωnk, which vanishes at the k
point corresponding to the bandgap. Thus R(2)cvk is zero for total photon energies corresponding
to the band gap, and increases for larger excess photon energies. The dependence of R(3)cvk on
the total photon energy is different, as it depends on both interband and intraband velocity ma-
trix elements. For total photon energies just above the gap, R(3)cvk is determined mainly by the
interband matrix elements, but as the photon excess energy increases R(3)cvk becomes dominated
by the intraband matrix elements, since the electronic transitions occur at k points with larger
band dispersion.
The injection rate coefficients corresponding to the interference of 2- and 3-photon pro-
cesses can then be computed as
µabd, f g2+3 (Ω) = 2π
∫dk
(2π)D
∑cvc′v′
(Mc′ckδv′v − Mv′vkδc′c
)δωcv=ωc′v′ R(3)abd∗
c′v′k R(2) f gcvk δ (Ω − ωcv) ,
(4.26)
following Eqs. (4.10) and (4.23), where we have taken the continuous momentum limit. For
the plots in the next sections we use a frequency broadening ∆ corresponding to ~∆ = 13meV.
The factor R(3)abd∗c′v′k R(2) f g
cvk changes sign under a transformation k → −k, and this is the mech-
anism for constructive versus destructive interference in opposite points of the Brillouin zone.
In Fig. 4.1 we illustrate constructive versus destructive interference of 2- and 3-photon pro-
cesses at opposite points in the Brillouin zone.
1The denominators in Eqs. (4.24) and (4.25) do not lead to any divergences because of the assumption that2~ω is below the gap.
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 91
4.3 Electronic model of AlGaAs
I use a 30-band k ·pmodel for computing the electronic bands. The model has free parameters
associated with energies and momentum matrix elements at the Γ point, and the parameters
are adjusted to match the experimental results for band energies from −5 eV to 4 eV, such that
computations of optical absorption coefficient are expected to be reliable for photon energies
up to 6 eV.
L Γ X U,K3
2
1
0
1
2
3
4
5
ω (eV
)
Al0. 2Ga0. 8As
L Γ X U,K3
2
1
0
1
2
3
4
5
ω (eV
)
Al0. 35Ga0. 65As
Figure 4.2: Electronic bandstructure of AlαGa1−αAs for two different stoichiometries.
Using the Γ point as the expansion point for a k · p model, the effective Hamiltonian that
acts only on the periodic part of an energy eigenfunction of crystal momentum k is
Heff = H +~
mk · p +
~2k2
2m, (4.27)
where H is the Hamiltonian (4.11) with the vector potential set equal to zero; in this model [88]
we neglect the k dependence of the effective spin-orbit term. The second term on the right-
hand-side is the usual k·p contribution, and the last term is the contribution to the kinetic energy
only due to the lattice momentum. The basis of states has 8 sets [88], 4 of them corresponding
to the Γ1 representation of the point group Td (or 43m), 3 corresponding to the Γ4 representation,
and 1 to the Γ3 representation. The Γ1 representation has only 1 state, Γ4 has 3 states, and Γ3
has 2 states, so in total we have 4 × 1 + 3 × 3 + 1 × 2 = 15 states before considering spin; we
denote these states as |A〉, |B〉, etc. Tensor products of these are taken with spin states to get 30
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 92
states in all. Terms 〈A|Heff |B〉 are then 2 × 2 matrices, and take the form
〈A|Heff |B〉 = EAδABσ0 +i3∆AB · σ + iPAB · kσ0 +
~2k2
2mδABσ0, (4.28)
where σ0 is the unit 2 × 2 matrix and the components of σ are the usual Pauli matrices. The
free parameters of the model are the energies EA, the matrix elements of the spin-orbit term
∆AB, and the matrix elements of the momentum operator PAB. Since the basis for the states is
the same at every k point [129], the corresponding 2×2 matrices corresponding to the velocity
operator 〈A|v|B〉 are diagonal in the spin sector,
〈A| va |B〉 =1~
∂
∂ka 〈A|Heff |B〉 =
(i~
PaAB +
~ka
mδAB
)σ0, (4.29)
from which the matrix elements of the velocity operator between the energy eigenstates can be
determined.
For GaAs [88] and AlAs [130] I use reported parameters adjusted for room temperature,
while the parameters for AlαGa1−αAs are obtained from a linear interpolation according to the
stoichiometry. This approximation is accurate within an energy tolerance corresponding to
room temperature 2. The chosen parameters lead to effective masses and g-factors that are in
good agreement with experimental data. More important for the problems I consider, the band
structures and linear optical absorption spectra are also in good agreement with experimental
data. In Fig. 4.2 I show the relevant electronic bands for two different stoichiometries. As a
validation of the bandstructures obtained with the 30-band model, I computed bandstructures
for α = 0 (GaAs) and for α = 1 (AlAs) and found good agreement with first-principles (DFT)
calculations (not shown here).
In Fig. 4.3 I show the imaginary parts of the corresponding dielectric functions, which are
related to the 1-photon absorption rates (or carrier injection) by Im ε (Ω) = ~ξxx (Ω) /2ε0.
2The corrections that are quadratic on the stoichiometry parameter α are small, and do not lead to significantchanges in the band energies within a tolerance given by room temperature.
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 93
1.8 2.0 2.2 2.4 2.6 2.8
Ω(eV)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Im ε
(Ω)/ε 0
broadening = 13 meV
α= 0. 20
α= 0. 35
.
Figure 4.3: Imaginary part of the dielec-tric function for different stoichiometries α.The vertical lines indicate the correspondingbandgaps.
4.4 Quantum interference control using two- and three-photon
absorption in AlGaAs
I consider two incident fields of different frequencies with amplitudes
Eω = Eω eiφω eω and E3ω/2 = E3ω/2 eiφ3ω/2 e3ω/2, (4.30)
where Eω > 0 and E3ω/2 > 0 are the field magnitudes, the unit vectors eω and e3ω/2 indicate
their polarizations, and φω and φ3ω/2 indicate their phases. We also define the phase parameter
∆φ = 2φ3ω/2 − 3φω, which will be useful later. We assume that the field at 3ω/2 has a weaker
intensity than the field at ω, and we demand that the frequencies satisfy 2ω < ∆g < 3ω, where
Eg = ~∆g is the electronic gap. Therefore only 3PA processes are important for the lower
frequency fieldEω, while only 2PA processes are relevant for the higher frequency fieldE3ω/2;
the 3PA associated with E3ω/2 is weaker due to the lower intensity of the field, and we neglect
it.
I focus on Al concentrations α such that 0.18 . α . 0.38, since AlαGa1−αAs with α too
small has a band gap smaller than 2~ω for telecommunication wavelengths (~ω ≈ 0.8 eV), and
AlαGa1−αAs with α too large is too reactive.
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 94
1.8 2.0 2.2 2.4
hΩ (eV)
0.0
0.2
0.4
0.6
0.8
1.0ξ 2
(Ω)
(m/V
4s)
×105 Al0.2Ga0.8As
xxxxxyxyxxyy
2.0 2.2 2.4 2.6 2.8
hΩ (eV)
0
1
2
3
4
5
6
7
ξ 2(Ω
)(m/V
4s)
×104 Al0.35Ga0.65As
xxxxxyxyxxyy
Figure 4.4: Two-photon carrier injection coefficients for two different stoichiometries.
1.8 2.0 2.2 2.4
hΩ (eV)
0
1
2
3
4
5
6
7
ξ 3(Ω
)(m
3/V
6s)
×10−13 Al0.2Ga0.8As
xxxxxxxxxxyyxxyxxyxyzxyzxxyyzz
2.0 2.2 2.4 2.6 2.8
hΩ (eV)
0.0
0.5
1.0
1.5
2.0
2.5
ξ 3(Ω
)(m
3/V
6s)
×10−13 Al0.35Ga0.65As
xxxxxxxxxxyyxxyxxyxyzxyzxxyyzz
Figure 4.5: Interference of 2- and 3-photon carrier injection coefficients for two different stoi-chiometries.
4.4.1 Carrier injection
I track the number of injected carriers by calculating the number of electrons in the conduction
bands, which corresponds to the operator
N =∑ck
a†ckack, (4.31)
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 95
so we use ncc′ = δcc′ and nvv′ = 0 for the carrier density matrix elements in Eq. (4.26). The
optical injection of carriers due to 2PA and 3PA processes, as well as their interference, is
characterized by the tensors ξ, according to
ddt〈n〉2 = ξabcd
2 (3ω) Ea−3ω/2Eb
−3ω/2Ec3ω/2Ed
3ω/2, (4.32)
ddt〈n〉3 = ξ
abcde f3 (3ω) Ea
−ωEb−ωEc
−ωEdωEe
ωE fω, (4.33)
ddt〈n〉2+3 = ξabcde
2+3 (3ω) Ea−ωEb
−ωEc−ωEd
3ω/2Ee3ω/2 + c.c., (4.34)
where ~Ω = 3~ω is the total transition energy. If the band gap of the material is smaller than
5ω/2 there would be an additional contribution to n2 proportional to E−ωE−3ω/2EωE3ω/2, but
these carriers do not contribute to the interference between 2PA and 3PA, which is our main
interest. The coefficients are calculated as
ξabde2 (3ω) = 2π
∫dk
(2π)D
∑cv
Rab∗cvkRde
cvkδ (3ω − ωcv) , [three IC] (4.35)
ξabde f g3 (3ω) = 2π
∫dk
(2π)D
∑cv
Rabd∗cvk Re f g
cvkδ (3ω − ωcv) , [three IC] (4.36)
ξabde f2+3 (3ω) = 2π
∫dk
(2π)D
∑cv
Rabd∗cvk Re f
cvkδ (3ω − ωcv) [five IC]. (4.37)
The number of independent non-zero components of the tensors ξ2, ξ2+3, and ξ3 is deter-
mined by the symmetries of the zincblende lattice. The text within square brackets in Equa-
tions (4.35)–(4.37) indicates the number of nonzero, independent components (“IC”). I present
the complete list in the Appendix A.
In Figs. 4.4, 4.6 and 4.5, we show the frequency dependence of the independent compo-
nents of the coefficients ξabcd2 (3ω), ξabcde f
3 (3ω), and ξabcde2+3 (3ω), respectively. Notice that the
3PA coefficient is large for frequencies right above the band gap, while the coefficient for 2PA
nearly vanishes for similar frequencies. As discussed following Eqs. (4.24) and (4.25), the
dominant contribution to 2PA always involves intraband velocity matrix elements, which cor-
respond to the band dispersion, so they vanish at the Γ point of the Brillouin zone. The 3PA
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 96
1.8 2.0 2.2 2.4
hΩ (eV)
0.0
0.2
0.4
0.6
0.8
1.0
−ξ 2
+3(Ω
)(m
2/V
5s)
×10−5 Al0.2Ga0.8As
xxxyzxyyyzxyzxx
2.0 2.2 2.4 2.6 2.8
hΩ (eV)
0
1
2
3
4
5
6
7
−ξ 2
+3(Ω
)(m
2/V
5s)
×10−6 Al0.35Ga0.65As
xxxyzxyyyzxyzxx
Figure 4.6: Interference of two- and three-photon carrier injection coefficients for differentstoichiometries.
has contributions from interband velocity matrix elements, which in general do not vanish at
Γ.
1.8 2.0 2.2 2.4
hΩ (eV)
−0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
η 2+
3(Ω
)(A
m3/V
5s)
×10−17 Al0.2Ga0.8As
xxxxxxxxxxyyyxxxxyyxxyxxyxxyyyxxxyxyyyzzxxzxxyyzxxyzyz
2.0 2.2 2.4 2.6 2.8
hΩ (eV)
0.0
0.5
1.0
1.5
η 2+
3(Ω
)(A
m3/V
5s)
×10−17 Al0.35Ga0.65As
xxxxxxxxxxyyyxxxxyyxxyxxyxxyyyxxxyxyyyzzxxzxxyyzxxyzyz
Figure 4.7: Two- and three-photon interference current injection coefficients for differentstoichiometries.
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 97
4.4.2 Current injection
The injected current density 〈J〉 due to the quantum interference between two- and three-
photon absorption processes is given by
ddt〈Ja〉2+3 = η
abcde f2+3 (3ω) Eb
−ωEc−ωEd
−ωEe3ω/2E f
3ω/2 + c.c., (4.38)
where ~Ω = 3~ω is the total photon energy. I list the independent components of the injection
tensor coefficient ηabcde f2+3 (3ω) in Appendix A. In Fig. 4.7 I show the frequency dependence of
the independent components of the coefficient ηabcde f2+3 (3ω) for different stoichiometries. The
plots show that some components change sign as the frequency increases. This is due to the
competing contributions due to intraband and interband velocity matrix elements to the R(3)cvk
coefficients. For low excess photon energies, the excited carriers are close to the Γ point in the
BZ, and the interband contribution is the most important, as the band dispersion is small. For
larger photon excess energies, the excited carriers are located further from the Γ point in the
BZ, so the band dispersion is large and the intraband contributions are more important.
0.0 0.5 1.0 1.5 2.0
θ/π
−3
−2
−1
0
1
2
3
η 2+
3(Ω
)(A
m3/V
5s)
×10−17 Al0.2Ga0.8As
e3ω/2 = θ
eω = θ
0.0 0.5 1.0 1.5 2.0
θ/π
−3
−2
−1
0
1
2
3
η 2+
3(Ω
)(A
m3/V
5s)
×10−17 Al0.2Ga0.8As
e3ω/2 = eω
e3ω/2 = z × eω
Figure 4.8: Coefficient for injection current along the x direction as the polarizations of theincident fields are rotated in the x-y plane. (Top) Either eω or e3ω/2 is rotated while the otheris fixed along the x direction. (Bottom) In both cases eω = θ is rotated, and e3ω/2 is eitherparallel or perpendicular to it. The total photon energy is ~Ω = 2.4 eV (≈ 520 nm) in bothcases.
In order to illustrate some aspects of the different tensor components, in Fig. 4.8 I plot
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 98
the injection current for different polarizations of the incident fields in a typical experimental
scenario. We assume that the sample has electrodes mounted such that they always measure
the current along the [100] crystal direction, which we denote by x. In the first case we keep
either eω or e3ω/2 fixed along the x direction, while the other field is rotated in the x-y plane
and points along the direction θ = x cos θ + y sin θ, where y corresponds to the [010] crystal
direction. In the second scenario, the polarizations of both incident fields are rotated in the
x-y plane and they are kept either parallel or perpendicular to each other. In Fig. 4.8, we show
that the current is largely along the eω direction regardless of the e3ω/2 direction. However,
the magnitude of the current depends significantly on the e3ω/2 direction, and it is maximal for
e3ω/2 = eω.
1.8 2.0 2.2 2.4
hΩ (eV)
−0.2
0.0
0.2
0.4
0.6
0.8
1.0
Vsw
arm
(m/s
)
×106 Al0.2Ga0.8As
e3ω/2 = eω
e3ω/2 = z × eω
2.0 2.2 2.4 2.6 2.8
hΩ (eV)
−0.2
0.0
0.2
0.4
0.6
0.8
1.0
Vsw
arm
(m/s
)
×106 Al0.35Ga0.65As
e3ω/2 = eω
e3ω/2 = z × eω
Figure 4.9: Swarm velocity assuming optimum interference between 2PA and 3PA. Thecurrent is measured along the x direction, and so is the polarization of the lower frequencyfield eω = x, while we consider two cases for the polarization of the higher frequency field:e3ω/2 = x and e3ω/2 = y.
Swarm velocity
Since the excited carriers respond to the induced voltage due to the injected current, and usually
screen it at least partially, a good measure of the efficiency of the current injection is the swarm
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 99
velocity, defined as
vswarm =
ddt 〈J〉
e ddt 〈n〉
, (4.39)
which represents the average contribution to the injection current due to one excited electron.
Since 〈n〉2+3 〈n〉2 + 〈n〉3, the total density of carriers is 〈n〉 ' 〈n〉2 + 〈n〉3, and for both light
beams polarized along the x direction we have a swarm velocity of magnitude
vswarm =2∣∣∣ηxxxxxx
2+3 (3ω)∣∣∣ E3
ωE23ω/2
|e|(ξxxxxxx
3 (3ω) E6ω + ξxxxx
2 (3ω) E43ω/2
) , (4.40)
where we have chosen ∆φ = π/2 to optimize the magnitude of the numerator. The whole ex-
pression is optimized by choosing the intensities of the two beams appropriately; the condition
to be satisfied is ξxxxx2 (3ω) E4
3ω/2 = ξxxxxxx3 (3ω) E6
ω , which corresponds to an equal number of
carriers injected by 2-photon absorption and 3-photon absorption. If this holds,
vswarm =
∣∣∣ηxxxxxx2+3 (3ω)
∣∣∣|e|
√ξxxxxxx
3 (3ω) ξxxxx2 (3ω)
. (4.41)
In Fig. 4.9 I plot this expression, together with the expression that would result if the beam
of frequency ω were polarized in the x direction while the one of frequency 3ω/2 in the y
direction, which is the same as Eq. (4.41) but with ηxxxxxx2+3 (3ω) replaced by η
xxxxyy2+3 (3ω); as
well, ξxxxx2 (3ω) should also be replaced by ξyyyy
2 (3ω), but they are equal. We see that different
stoichiometries give similar values for the swarm velocity if the frequency is adjusted according
to the electronic band gap of the system. The fact that higher Ga concentrations lead to larger
injected currents (see Fig. 4.7) is only due to a higher carrier injection. Yet with appropriate
laser intensities it is possible to reach the same levels of injected current densities with any Al
concentration, although the laser frequencies and intensities at which the maximum is achieved
depend on the Al concentration.
Notice that the 2+3 QuIC swarm velocity is about twice its equivalent for 1+2 QuIC. This
is an indication that the distribution of carriers injected in the BZ is sharper for 2+3 QuIC com-
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 100
pared to 1+2 QuIC. We further confirm that by computing the variance of the lattice momentum
k of the electrons injected in the conduction band for both 1+2 QuIC and 2+3 QuIC,
σa1+2 =
⟨(ka)2
⟩1+2− 〈ka〉
21+2 , σa
2+3 =⟨(ka)2
⟩2+3− 〈ka〉
22+3 (4.42)
respectively. For the incident fields polarized along the x direction, we find
〈k〉1+2 = (4.9, 0, 0) × 10−2Å−1, 〈k〉2+3 = (5.7, 0, 0) × 10−2Å
−1, (4.43a,b)
σ1+2 = (3.4, 4.3, 4.3) × 10−3Å−2, σ2+3 = (2.8, 2.2, 2.2) × 10−3Å
−2, (4.43c,d)
which indicates that the distribution of injected electrons in the BZ is sharper for 2+3 QuIC,
especially in the directions transverse to the polarization of the field (y and z).
Laser intensities
Our calculations are performed in the perturbative regime, the validity of which requires that
the fraction of the injected carrier population density relative to the total density of states nmax
in the range of energies covered by the laser pulse be small. We thus consider these calculations
to be valid when
〈n〉2 + 〈n〉3 < 0.1 nmax, (4.44)
where the fraction 0.1 is chosen arbitrarily, as an estimate of unsaturated regime. The carrier
injection due to the 2- and 3-photon interference 〈n〉2+3 mostly has the effect of concentrating
the carrier injection in some region of the BZ, but it does not contribute significantly to the total
number of injected carriers compared to 〈n〉2 and 〈n〉3. For the estimates of laser intensities we
consider the incident fields to be both polarized along the x direction, so for a laser pulse of
duration T we require
[ddt〈n〉2 +
ddt〈n〉3
]T < 0.1nmax, (4.45)
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 101
[ξxxxx
2 (3ω) E43ω/2 + ξxxxxxx
3 (3ω) E6ω
]T < 0.1nmax. (4.46)
The maximum density of states nmax that can be injected is determined by analyzing the vol-
ume V corresponding to the excited states in the BZ. We denote by kΩ the momentum cor-
responding to the energy difference ~ωcvk = ~Ω between the conduction and valence bands,
so V = 4πk2Ω∆k, where ∆k = dk
dωcvk∆ωcvk is related to the frequency broadening ∆ω = 2π/T
associated with the time duration of the pulse. The derivative of the band energy corresponds
to the velocities of electrons in the conduction and valence bands, vΩ = dωcvdk = dωc
dk −dωvdk , so
V = 8π2k2Ω/(vΩT ). The volume in the BZ associated with one quantum state is V1 = (2π/L)3,
where L is the normalization length of the sample. The number of states that can be excited is
then V/V1, and their spatial density is
nmax =V
V1L3 =k2
Ω
πvΩT. (4.47)
For optimal interference, there should be equal densities of carriers injected by 2- and 3-photon
absorption, 〈n〉2 = 〈n〉3, which according to Eq. (4.46) gives
ξxxxx2 (3ω) E4
3ω/2 = ξxxxxxx3 (3ω) E6
ω < 0.05k2
Ω
πvΩT2 . (4.48)
The maximal amplitudes Eω and E3ω/2 of the incident fields can then be estimated from the
extreme of the inequality in the above equation. For the stoichiometry of α = 0.2, pulses with
duration T = 150 fs, and total photon energy ~Ω = 2.4 eV, we have
Eω = 1.24 × 108 Vm, and E3ω/2 = 6.05 × 107 V
m. (4.49)
The intensities in the material medium with these field amplitudes are given by Iϕ = (2ε0c) nϕ E2ϕ,
with ϕ =ω, 3
2ω,
Iω = 26.5GWcm2 , I3ω/2 = 6.54
GWcm2 . (4.50)
Chapter 4. Coherent Control of Two- and Three-photon Absorption in AlGaAs 102
For these values, the injected current density is
〈Jx〉 = 2ηxxxxxx2+3 (3ω) E3
ωE23ω/2T = 6.25
MAcm2 . (4.51)
It should be noted that these are just estimates, as the limit of carrier density is set arbitrarily
in Eq. (4.44). Note that we are ignoring scattering of the injected carriers. This means that
the true maximal intensities would be larger than our estimates here, since there is room for
more photon absorption as scattering depletes some of the excited states. Also note that in
this treatment the electron-electron interaction has been neglected; were it included, the phase
parameter would be shifted. However this shift is usually very small for zincblende semicon-
ductors, except for frequencies very close to the band gap [131].
4.5 Summary and discussion
QuIC using 2- and 3-photon processes leads to carrier excitation even more localized in the
BZ than 1+2 QuIC. One consequence of this is the fact that the swarm velocity is higher in
2+3 QuIC than in 1+2 QuIC. Another interesting difference is that in 2+3 QuIC the current
injection coefficients tend to change sign as the total photon energy is increased, while in 1+2
QuIC they typically do not. This happens because interband velocity matrix elements are
responsible for the largest contribution to the 3PA coefficient at low photon energies, but at
higher photon energies the intraband velocity matrix elements dominate. As well, for a given
material 2+3 QuIC can result from a larger set of laser frequencies than 1+2 QuIC. Since
only nonlinear optical processes are involved in 2+3 QuIC, the laser intensities required for
maximal effect are higher than for 1+2 QuIC, but still moderate. Also, the fields have a power
law attenuation as they propagate through the absorbing material, instead of the exponential
attenuation of linear absorption. Thus a waveguide geometry is desirable. And while QuIC in
waveguides presents some challenges, as it raises issues of phase and mode matching, it also
presents opportunities as it would allow for easy integration with devices on-chip.
Chapter 5
Conclusions
Effective models to compute electronic and optical properties of semiconductor materials are
of fundamental and practical interest; these simple models enable researchers to study physical
properties at low computational cost, and also provide the simplicity that allows for a focus on
particular phenomena. The study of reduced dimensionality materials is crucial not just from
a fundamental point if view, but also from a practical perspective; as progress on the minia-
turization of technology continues, and as more complex on-chip devices become feasible, the
study of quantum confinement effects becomes more relevant. An important area of research
in this field is the study of optical processes in semiconductor materials. The relevance of this
topic is only expected to grow as opto-electronic devices continue to become ubiquitous; as
these gadgets tend towards nanometric dimensions, the study of the optoelectronic properties
becomes crucial, since quantum confinement strongly modifies the (joint) density of electronic
states accessible to a given photon energy. In this thesis I have presented a study of reduced-
dimensionality semiconductor materials employing effective Hamiltonian models; I presented
a study of narrow strips of monolayer graphene (1D), of monolayer stanene (2D) and of the
alloy AlαGa1−αAs (3D). I presented calculations of linear and nonlinear optical response, in-
cluding charge-carrier injection by one- and two-photon absorption processes, spin-injection,
and the coherent control of injected charge current. Some of these results were contrasted
against first-principles calculations based on density-functional theory. In the following lines
103
Chapter 5. Conclusions 104
I present the relevant findings of these studies and conclude this chapter with an outline of
possible routes for future research.
In Chapter 2 I presented a study of narrow strips of graphene, commonly known as graphene
nanorribons. Electrons are confined along two dimensions, i.e., periodicity is unidimensional.
I confirmed that the electronic dispersions strongly depend on the boundaries of the graphene
strip. I focused on zigzag boundaries, because they led to strongly confined edge states along
the zigzag borders; these states are found to be precisely at the Fermi level for an undoped
material, and as such, the large number of occupied states per unit length allocated in the edge
states is easily controlled by doping. Another important characteristic of this system is that
—within the k · p framework— the bandstructure is entirely composed of parabolic-like en-
ergy bands, aside from the edge-states. These parabolic bands lead to a rich structure due to
singularities in the joint-density-of-states (JDOS). This was followed by a detailed study of
the velocity matrix elements, in which their selection rules were identified. Then I presented
the one- and two-photon carrier injection coefficients, together with their respective selection
rules. Similarly to graphene, the zigzag ribbons support stimulated electronic Raman scatter-
ing (ERS), which in general is present when the fundamental photon energy is larger than the
optical bandgap; this scattering process is characterized by virtual absorption at 2~ω followed
by emission at ~ω. In general, due to the boundary conditions, and the uni-dimensional peri-
odicity of the ribbons, different nonzero optical coefficients have different selection rules, in
contrast with a monolayer of graphene.
One of the key findings in this chapter is that, aside from the JDOS divergencies present
in all the optical coefficients studied here, the second order absorption processes (2PA and
ERS) also possess a class of divergences associated with resonant conditions, when the optical
process in consideration connects real states. Once these absorption processes were analyzed,
I proceeded to study interference processes among them. This interference process leads to
current injection of charge carriers. First I studied current injection arising from the interfer-
ence of sequential absorption of two single photons with two-photon absorption, and then that
from the interference of ERS absorption with 1PA. Although all these absorption and current
Chapter 5. Conclusions 105
injection coefficients inherit the rich structure of the JDOS, I found they follow a general trend
similar to that of 2D graphene.
Since the outstanding feature of zigzag nanoribbons is the existence of localized edge states
along their length and confined states across their width, in this work I identified the contri-
bution to the optical responses that involves edge states (edge ↔ edge and edge ↔ confined
transitions). This is referred here as the “only-edge” contribution, and it is thoroughly com-
pared with the “full” contribution that originates from any combination of optical transitions
(“only-edge” plus confined↔ confined). From this comparison I found that the edge states are
always involved in the onset of the absorption and the current injection processes, and that at
low-photon energies the edge states have the largest contributions. Since the energy of these
edge states is near zero, this suggested a study of the effects of positive doping. I found that, at
low photon energies, all the optical coefficients of ZGNR are extremely sensitive to changes in
the chemical potential. Due to the relative ease with which the chemical potential of graphene
can be altered, zigzag nanoribbons offer an excellent opportunity to generate and control charge
currents by optical fields and by doping techniques. Moreover, due to the strong dependence of
the electronic bandstructure with respect to the width and boundary conditions (edge shape),
the study of graphene ribbons proves relevant from both fundamental and practical perspec-
tives.
In Chapter 3 I studied another graphene-like semiconductor material, stanene, a monolayer
of tin atoms arranged in a buckled honeycomb lattice. Stanene is a promising material for
fundamental studies and practical applications, given the topological properties and the signif-
icant spin-orbit interaction present in its electronic structure; moreover, monolayer stanene on
a substrate and free-standing few-layered stanene have recently been obtained experimentally.
My contribution in this chapter was the development of an effective Hamiltonian model that
accurately describes the bandstructure of stanene at low photon energies employing a minimal
basis set. This simple model is based on a singular value decomposition of overlap matrices
obtained in a preliminary first-principles calculation. This effective model resembles a tradi-
tional k · p method but it differs from the latter in that the electronic energies described by
Chapter 5. Conclusions 106
the simple model presented here uses a minimal basis set and it is not restricted to a second
order dependence on the crystal momentum. I also presented a detailed analysis of the approx-
imations made, and I provided figures of merit for the eigenenergies and eigenstates obtained
with this simple model. During this analysis I found that a second order expansion of the
bandstructure energies is more accurate than higher order expansions on k for stanene; indeed,
the two-dimensional band warping of the bandstructure is well described by the second order
model, as compared with the band warping predicted by a first-principles calculation. Another
significant finding in this work was that the lattice buckling in stanene can be neglected for the
calculation of the optical responses presented in this third chapter; this simplification led to a
separation of the effective Hamiltonian matrix into spin subsectors. I also found that the model
predicts the coupling of circularly polarized light with the spin of the electron; this points to
a potential application of stanene for spintronic applications. Since the procedure described
in this chapter depends solely on a preliminary first-principles calculation, that is, since this
effective model is free of phenomenological input, I expect this method can be applied to study
other bidimensional materials.
After having studied mono- and bi-dimensional novel materials in Chapter 2 and Chapter
3, respectively, in Chapter 4 I focused my attention on a “traditional” tridimensional alloy
material, AlαGa1−αAs . Here I presented results for charge injection based on two- and three-
photon (2+3) absorption processes, and results of optical coherent control of charge current
based on the interference of these two color optical phenomena. Although a vast amount
of literature exist on this material, this class of properties has not yet been reported in the
literature. As in previous chapters, I performed this study using a k · p model, but here I used
a 30-band basis set and I also included the description of spin-orbit interaction, as I also did
on Chapter 3. In general terms, this chapter presents an example of bandstructure engineering
in which the objective was to study optical coherent control of charge currents for suitable
values of the stoichometric parameter α in AlαGa1−αAs . As compared with optical coherent
control based on one- and two-photon (1+2) absorption processes (reported elsewhere), the
study presented here showed that coherent control based on 2+3 processess occurs on a smaller
Chapter 5. Conclusions 107
Brillouin Zone volume, and consequently the swarm velocities reported here are higher in this
latter case. The laser intensities required to maximize the injection of current based on 2+3
processes are higher as compared with 1+2 processes, since in the former case only non-linear
processes are involved.
Future Work
In this thesis I have presented studies of electronic structure and optical properties of some
semiconductors in the framework of simple model Hamiltonians. Although some of these
properties were contrasted with first-principles results, an obvious extension of this work could
include the calculation of all the optical coefficients from first principles. This extension
could still be restricted to a single-particle (non-interacting) picture (as was done here with
simple Hamiltonians) in the framework of density-functional theory and the local-density ap-
proximation (DFT/LDA). Yet a further level of sophistication could include the description of
many-body effects to describe the electron-electron interaction beyond the scope of exchange-
correlation pseudopotentials, as it is commonly done in the DFT/LDA. The best description
of these many-body effects to date is given by the GW approximation, which is commonly
performed with a preliminary DFT calculation of the ground-state wavefunctions. These DFT
wavefunctions are then used as a basis set to express the Green functions in the GW method.
Further sophistications can also include the description of electron-hole (neutral) excitations by
solving the Bethe-Salpeter equations. Another way to extend the work presented here is to em-
ploy time-dependent first-principles methods to describe time- and frequency-dependent quan-
tities; to this day time-dependent DFT is one of the most widely used schemes for this kind of
description. As time-dependent-DFT is a single-particle framework, it has a much lower com-
putational cost than the (many-body) GW method, however it faces important challenges, such
as the accurate description of time-dependent correlation effects and time-dependent structural
forces.
Extensions of the work done in this thesis for specific materials or specific material geome-
tries can include the study of nanoribbons made of monolayered materials other than graphene,
Chapter 5. Conclusions 108
e.g., silicene, germananene, stanene, h-BN, and SiC, among others. Several of these monolayer
materials have recently been produced. The effective model presented in Chapter 3 could be
extended to study the optical coherent control of charge currents and the effects of external
DC fields. Finally, the study presented in Chapter 4 can be extended to study alloys other
than AlGaAs and by studying waveguide heterostructures to for potential applications of the
optically-injected charge currents.
Appendix A
Nonzero injection coefficient components
of zincblende lattices
AlGaAs in the virtual crystal approximation forms a zincblende lattice, which has the sym-
metry of point group Td (or 43m). The optical responses we consider in this work involve
tensors of rank 4 up to 6. With Td symmetries [132], generic rank-4 tensors have 21 non-zero
components of which 4 are independent, rank-5 tensors have 60 non-zero and 10 independent
components, and rank-6 tensors have 183 non-zero and 31 independent components.
However, the tensors representing the optical processes have a few more specific restric-
tions due to their relation to the optical fields, as the indices associated with the same incident
field are symmetrized. With these considerations the number of nonzero independent compo-
nents are
Number of
Tensor independent components
ξ2 three
ξ3 five
ξ2+3 three
η2+3 nine
109
Appendix A. Nonzero injection coefficient components of zincblende lattices 110
The complete list is as follows. ξ2 has three independent components
ξxxxx2 = P (x, y, z) , ξ
xyxy2 = ξ
xyyx2 = P (x, y, z) , ξ
xxyy2 = P (x, y, z) , (A.1a,b,c)
where P (x, y, z) indicates all the possible permutations of (x, y, z) in the indices. The tensor
ξ2+3 has three independent components
ξxxxyz2+3 = P (x, y, z) , (A.2a)
ξxxyxz2+3 = ξ
xxyzx2+3 = ξ
xyxxz2+3 ξ
xyxzx2+3 = ξ
yxxxz2+3 = ξ
yxxzx2+3 = P (x, y, z) , (A.2b)
ξxyzxx2+3 = ξ
yxzxx2+3 = ξ
yzxxx2+3 = P (x, y, z) , (A.2c)
while ξ3 has five independent components
ξxxxxxx3 = P (x, y, z) , (A.3a)
ξxxxxyy3 = ξ
xxxyxy3 = ξ
xxxyyx3 ξ
yyxxxx3 = ξ
yxyxxx3 = ξ
xyyxxx3 = P (x, y, z) , (A.3b)
ξxxyxxy3 = ξ
xxyxyx3 = ξ
xxyyxx3 = ξ
xyxxxy3 = ξ
xyxxyx3 = ξ
xyxyxx3
= ξyxxxxy3 = ξ
yxxxyx3 = ξ
yxxyxx3 = P (x, y, z) , (A.3c)
ξxxyyzz3 = ξ
xxyzyz3 = ξ
xxyzzy3 = ξ
xyxyzz3 = ξ
xyxyzyz3 = ξ
xyxzzy3
= ξyxxyzz3 = ξ
yxxyzyz3 = ξ
yxxzzy3 = P (x, y, z) , (A.3d)
ξxyzxyz3 = ξ
xyzzxy3 = ξ
xyzyzx3 = ξ
xyzzyx3 ξ
xyzxzy3 = ξ
xyzyxz3 = P (x, y, z) . (A.3e)
Appendix A. Nonzero injection coefficient components of zincblende lattices 111
Finally, the tensor η2+3 has nine independent components
ηxxxxxx2+3 = P (x, y, z) , (A.4a)
ηxxxxyy2+3 = P (x, y, z) , (A.4b)
ηxxyyxx2+3 = η
xyxyxx2+3 = η
xyyxxx2+3 = P (x, y, z) , (A.4c)
ηxxxyxy2+3 = η
xxxyyx2+3 = η
xxyxxy2+3 = η
xxyxyx2+3 = η
xyxxxy2+3 = η
xyxxyx2+3 = P (x, y, z) , (A.4d)
ηyxxxxy2+3 = η
yxxxyx2+3 = P (x, y, z) , (A.4e)
ηyxxyxx2+3 = η
yxyxxx2+3 = η
yyxxxx2+3 = P (x, y, z) , (A.4f)
ηxxyyzz2+3 = η
xyxyzz2+3 = η
xyyxzz2+3 = P (x, y, z) , (A.4g)
ηxxyzyz2+3 = η
xxzyyz2+3 = η
xyxzyz2+3 = η
xyzxyz2+3 = η
xzxyyz2+3 = η
xzyxyz2+3 = P (x, y, z) , (A.4h)
ηxyyzxz2+3 = η
xyzyxz2+3 = η
xzyyxz2+3 = η
xyyzzx2+3 = η
xyzyzx2+3 = η
xzyyzx2+3 = P (x, y, z) . (A.4i)
Bibliography
[1] H.M. van Driel and J.E. Sipe. Ultrafast Phenomena in Semiconductors, chapter 5: Co-
herent Control of Photocurrents in Semiconductors, pages 261–306. Springer, 2001.
[2] H.M. van Driel and J.E. Sipe. Coherent control: Applications in semiconductors. In
Robert D. Guenther, editor, Encyclopedia of Modern Optics, pages 137–143. Elsevier,
Oxford, 2005.
[3] J. Rioux, Guido Burkard, and J. E. Sipe. Current injection by coherent one- and two-
photon excitation in graphene and its bilayer. Phys. Rev. B, 83:195406, May 2011.
[4] J. Rioux and J.E. Sipe. Optical injection processes in semiconductors. Physica E: Low-
dimensional Systems and Nanostructures, 45(0):1 – 15, 2012.
[5] J. Rioux, J. E. Sipe, and Guido Burkard. Interference of stimulated electronic raman
scattering and linear absorption in coherent control. Phys. Rev. B, 90:115424, Sep 2014.
[6] Rodrigo A. Muniz and J. E. Sipe. Coherent control of optical injection of spin and
currents in topological insulators. Phys. Rev. B, 89:205113, May 2014.
[7] Rodrigo A. Muniz and J. E. Sipe. All-optical injection of charge, spin, and valley cur-
rents in monolayer transition-metal dichalcogenides. Phys. Rev. B, 91:085404, Feb 2015.
[8] Kiran M. Rao and J. E. Sipe. Coherent photocurrent control in a magnetic field through
quantum interference. Phys. Rev. B, 84:205313, Nov 2011.
112
Bibliography 113
[9] Kiran M. Rao and J. E. Sipe. Coherent photocurrent control in graphene in a magnetic
field. Phys. Rev. B, 86:115427, Sep 2012.
[10] N. Mermin B. Ashcroft. Solid State Physics. Brooks Cole, 1972.
[11] Richard M. Martin. Electronic Structure: Basic Theory and Practical Methods. Cam-
bridge University Press, 2008.
[12] Marvin L. Cohen and James R. Chelikowsky. Electronic Structure and Optical Proper-
ties of Semiconductors. Springer, 2 edition, 1988.
[13] Yousef Saad, James R. Chelikowsky, and Suzanne M. Shontz. Numerical methods for
electronic structure calculations of materials. SIAM Review, 52(1):3–54, 2010.
[14] P. Yu and M. Cardona. Fundamentals of Semiconductors: Physics and Materials Prop-
erties. Springer, 2014.
[15] Richard M. Martin, Lucia Reining, and David M. Ceperley. Interacting Electrons: The-
ory and Computational Approaches. Cambridge University Press, 2016.
[16] Carsten Ullrich. Time-Dependent Density-Functional Theory. Oxford, 2012.
[17] Miguel Marques, Neepa Maitra, and Angel Rubio. Fundamentals of Time-Dependent
Density-Functional Theory. Springer, 2012.
[18] P. Hohenberg and W. Kohn. Inhomogeneous electron gas. Phys. Rev., 136:B864–B871,
1964.
[19] W. Kohn and L. J. Sham. Self-consistent equations including exchange and correlation
effects. Phys. Rev., 140:A1133–A1138, 1965.
[20] Feliciano Giustino. Materials Modeling using Density-Functional Theory. Oxford Uni-
versity Press, 2017. Undergrad/Graduate level.
Bibliography 114
[21] Chris-Kriton Skylaris, Peter D. Haynes, Arash A. Mostofi, and Mike C. Payne. Intro-
ducing ONETEP: Linear-scaling density functional simulations on parallel computers.
The Journal of Chemical Physics, 122(8):084119, 2005.
[22] John Singleton. Band theory and electronic properties of solids. Oxford Master Series
in Condensed Matter Physics. Oxford University Press, 2001.
[23] Robert A. Evarestov. Quantum Chemistry of Solids: The LCAO First Principles Treat-
ment of Crystals. Springer, 2006.
[24] Jean-Noel Fuchs and Mark Oliver Goerbig. Introduction to the physical properties of
graphene: Lecture notes. Laboratoire de Physique des Solides, 2008. Available online.
[25] P. Marconcini and M. Macucci. The k · p method and its application to graphene,
carbon nanotubes and graphene nanoribbons: the Dirac equation. La Rivista del Nuovo
Cimento, pages 489–584, 2011.
[26] Mark Fox. Optical Properties of Solids. Oxford Master Series in Physics (Book 3).
Oxford University Press, 2010.
[27] Emil Roduner. Nanoscopic Materials. The Royal Society of Chemistry, 2 edition, 2014.
[28] Marvin L. Cohen and Steven G. Louie. Fundamentals of Condensed Matter Physics.
Cambridge University Press, 2016.
[29] T.Ogawa and Y.Kanemitsu. Optical Properties of Low–Dimensional Materials. World
Scientific, 1996.
[30] Kyoko Nakada, Mitsutaka Fujita, Gene Dresselhaus, and Mildred S. Dresselhaus. Edge
state in graphene ribbons: Nanometer size effect and edge shape dependence. Phys. Rev.
B, 54:17954–17961, Dec 1996.
[31] L. Brey and H. A. Fertig. Electronic states of graphene nanoribbons studied with the
Dirac equation. Phys. Rev. B, 73:235411, Jun 2006.
Bibliography 115
[32] Hoonkyung Lee, Marvin L. Cohen, and Steven G. Louie. Selective functionalization of
halogens on zigzag graphene nanoribbons: A route to the separation of zigzag graphene
nanoribbons. Applied Physics Letters, 97(23):233101–1, 2010.
[33] Gomes da Rocha, Clayborne, Koskinen, and Hakkinen. Optical and electronic properties
of graphene nanoribbons upon adsorption of ligand-protected aluminum clusters. Phys.
Chem. Chem. Phys., 16:3558–3565, 2014.
[34] Li Yang, Marvin L. Cohen, and Steven G. Louie. Magnetic edge-state excitons in zigzag
graphene nanoribbons. Phys. Rev. Lett., 101:186401, Oct 2008.
[35] Jens Kunstmann, Cem Ozdogan, Alexander Quandt, and Holger Fehske. Stability of
edge states and edge magnetism in graphene nanoribbons. Phys. Rev. B, 83:045414, Jan
2011.
[36] M. Bellec, U. Kuhl, G. Montambaux, and F. Mortessagne. Manipulation of edge states
in microwave artificial graphene. New J. Phys., 16:113023, 2014.
[37] P. Delplace, D. Ullmo, and G. Montambaux. Zak phase and the existence of edge states
in graphene. Phys. Rev. B, 84:195452, Nov 2011.
[38] Oleg V. Yazyev. A guide to the design of electronic properties of graphene nanoribbons.
Accounts of Chemical Research, 46(10):2319–2328, 2013. PMID: 23282074.
[39] J M Luck and Y Avishai. Unusual electronic properties of clean and disordered zigzag
graphene nanoribbons. Journal of Physics: Condensed Matter, 27(2):025301, 2015.
[40] N Fujita, P.J. Hasnip, M.I.J. Probert, and Yuan.J. Theoretical study of core-loss electron
energy-loss spectroscopy at graphene nanoribbon edges. Journal of Physics: Condensed
Matter, 27(30):305301, 2015.
[41] Marta Pelc, Eric Suarez Morell, Luis Brey, and Leonor Chico. Electronic conduc-
tance of twisted bilayer nanoribbon flakes. The Journal of Physical Chemistry C,
119(18):10076–10084, 2015.
Bibliography 116
[42] D. Gosalbez-Martınez, D. Soriano, J.J. Palacios, and J. Fernandez-Rossier. Spin-filtered
edge states in graphene. Solid State Communications, 152(15):1469 – 1476, 2012. Ex-
ploring Graphene, Recent Research Advances.
[43] Thomas Stegmann and Axel Lorke. Edge magnetotransport in graphene: A combined
analytical and numerical study. Annalen der Physik, 527(9-10):723–736, 2015.
[44] J.L. Lado, N. Garcıa-Martınez, and J. Fernandez-Rossier. Edge states in graphene-like
systems. Synthetic Metals, 210, Part A:56 – 67, 2015. Reviews of Current Advances in
Graphene Science and Technology.
[45] Han Hsu and L. E. Reichl. Selection rule for the optical absorption of graphene nanorib-
bons. Phys. Rev. B, 76:045418, Jul 2007.
[46] Ken-ichi Sasaki, Keiko Kato, Yasuhiro Tokura, Katsuya Oguri, and Tetsuomi Sogawa.
Theory of optical transitions in graphene nanoribbons. Phys. Rev. B, 84:085458, Aug
2011.
[47] Takahiro Yamamoto, Tomoyuki Noguchi, and Kazuyuki Watanabe. Edge-state signa-
ture in optical absorption of nanographenes: Tight-binding method and time-dependent
density functional theory calculations. Phys. Rev. B, 74:121409, Sep 2006.
[48] M. Berahman, M. Asad, M. Sanaee, and M.H. Sheikhi. Optical properties of chi-
ral graphene nanoribbons: a first principle study. Optical and Quantum Electronics,
47(10):3289–3300, 2015.
[49] Wen-Huan Zhu, Guo-Hui Ding, and Bing Dong. The enhanced optical conductivity for
zigzag-edge graphene nanoribbons with applied gate voltage. Applied Physics Letters,
100(10):103101–1, 2012.
[50] Deborah Prezzi, Daniele Varsano, Alice Ruini, Andrea Marini, and Elisa Molinari. Op-
tical properties of graphene nanoribbons: The role of many-body effects. Phys. Rev. B,
77:041404, Jan 2008.
Bibliography 117
[51] E.A. Manykin and A.M. Afanas’ev. On one possibility of making a medium transparent
by multiquantum resonance. J.Exptl. Theor. Phys., 25(2):828–830, November 1967.
Original in Russian: ZhETF 52, No. 5, p. 1246-1249 (1967).
[52] E.A. Manykin. Quantum interference and coherent control. Laser Phys., 11:60, 2001.
[53] Dong Sun, Charles Divin, Julien Rioux, John E. Sipe, Claire Berger, Walt A. de Heer,
Phillip N. First, and Theodore B. Norris. Coherent control of ballistic photocurrents in
multilayer epitaxial graphene using quantum interference. Nano Letters, 10(4):1293–
1296, 2010.
[54] B. M. Fregoso and Sinisa Coh. Intrinsic surface dipole in topological insulators. Journal
of Physics: Condensed Matter, 27(42):422001, 2015.
[55] A. H. Castro-Neto, F. Guinea, N. M. R. Peres, K. S. Novoselov, and A. K. Geim. The
electronic properties of graphene. Rev. Mod. Phys., 81:109–162, Jan 2009.
[56] S. Das Sarma, Shaffique Adam, E. H. Hwang, and Enrico Rossi. Electronic transport in
two-dimensional graphene. Rev. Mod. Phys., 83:407–470, May 2011.
[57] We handle Dirac delta integrals of the form I(ω) =∫
dkFnmkδ(ω − ωnmk) by doing an
interpolation of the integrand, such that I(ω) =∑N−1
i=0
[ki+1−Qiki+1−ki
Fnmki|∆nmki |
+Qi−ki
ki+1−ki
Fnmki+1|∆nmki+1 |
]×θ(Qi − ki)θ(ki+1 −Qi), where Qi is the interpolated k-point that satisfies the Dirac delta,
Qi =ki+1(ω−ωnmki )+ki(ωnmki+1−ω)
ωnmki+1−ωnmki, and θ(k) is the unit step function. This interpolation scheme
requires convergence on a single parameter, the number of k points. More simple nu-
merical treatments of the Dirac delta integrals with broadening functions (Lorentzian or
Gaussian functions) require a larger number of k-points to reach convergence.
[58] J. Solyom. Fundamentals of the Physics of Solids, Vol 3. Springer, 2010.
[59] Hartmut Haug and Stephan W. Koch. Quantum Theory Of The Optical And Electronic
Properties Of Semiconductors. World Scientific Publishing, 5th edition, 2009.
Bibliography 118
[60] Pere Miro, Martha Audiffred, and Thomas Heine. An atlas of two-dimensional materi-
als. Chem. Soc. Rev., 43:6537–6554, 2014.
[61] Ankur Gupta, Tamilselvan Sakthivel, and Sudipta Seal. Recent development in 2D ma-
terials beyond graphene. Progress in Materials Science, 73:44–126, Aug 2015.
[62] Motohiko Ezawa and Guy Le Lay. Focus on silicene and other 2D materials. New
Journal of Physics, 17(9):090201, 2015.
[63] Jijun Zhao, Hongsheng Liu, Zhiming Yu, Ruge Quhe, Si Zhou, Yangyang Wang,
Cheng Cheng Liu, Hongxia Zhong, Nannan Han, Jing Lu, Yugui Yao, and Kehui Wu.
Rise of silicene: A competitive 2D material. Progress in Materials Science, 83:24–151,
Oct 2016.
[64] Trilochan Sahoo, Sanjeev K. Nayak, Pandian Chelliah, Manasa K. Rath, and Bhaskar
Parida. Observations of two-dimensional monolayer zinc oxide. Materials Research
Bulletin, 75:134–138, 3 2016.
[65] Qing Hua Wang, Kourosh Kalantar-Zadeh, Andras Kis, Jonathan N. Coleman, and
Michael S. Strano. Electronics and optoelectronics of two-dimensional transition metal
dichalcogenides. Nat Nano, 7(11):699–712, 11 2012.
[66] Xiaodong Xu, Wang Yao, Di Xiao, and Tony F. Heinz. Spin and pseudospins in layered
transition metal dichalcogenides. Nat Phys, 10(5):343–350, 05 2014.
[67] M. Houssa, B. van den Broek, E. Scalise, B. Ealet, G. Pourtois, D. Chiappe, E. Cin-
quanta, C. Grazianetti, M. Fanciulli, A. Molle, V. V. Afanas’ev, and A. Stesmans. The-
oretical aspects of graphene-like group IV semiconductors. Applied Surface Science,
291:98–103, 2 2014.
[68] Elisabeth Bianco, Sheneve Butler, Shishi Jiang, Oscar D. Restrepo, Wolfgang Windl,
and Joshua Goldberger. Stability and exfoliation of germanane: A germanium graphane
analogue. ACS Nano, 7(5):4414–4421, 2013.
Bibliography 119
[69] A Acun, L Zhang, P Bampoulis, M Farmanbar, A van Houselt, A N Rudenko, M Lin-
genfelder, G Brocks, B Poelsema, M I Katsnelson, and H J W Zandvliet. Ger-
manene: the germanium analogue of graphene. Journal of Physics: Condensed Matter,
27(44):443002, 2015.
[70] V. Ongun Ozcelik, O. Uzengi Akturk, E. Durgun, and S. Ciraci. Prediction of a two-
dimensional crystalline structure of nitrogen atoms. Phys. Rev. B, 92:125420, 2015.
[71] J. M. Pereira and M. I. Katsnelson. Landau levels of single-layer and bilayer phospho-
rene. Phys. Rev. B, 92(7):075437, Aug 2015.
[72] C. Kamal and Motohiko Ezawa. Arsenene: Two-dimensional buckled and puckered
honeycomb arsenic systems. Phys. Rev. B, 91:085423, Feb 2015.
[73] O. Uzengi Akturk, V. Ongun Ozcelik, and S. Ciraci. Single-layer crystalline phases of
antimony: Antimonenes. Phys. Rev. B, 91:235446, 2015.
[74] Giovanni Pizzi, Marco Gibertini, Elias Dib, Nicola Marzari, Giuseppe Iannaccone, and
Gianluca Fiori. Performance of arsenene and antimonene double-gate MOSFETs from
first principles. Nature Communications, 7:12585, Aug 2016.
[75] E. Akturk, O. Uzengi Akturk, and S. Ciraci. Single and bilayer bismuthene: Stability
at high temperature and mechanical and electronic properties. Phys. Rev. B, 94:014115,
2016.
[76] A. J. Mannix, X.-F. Zhou, B. Kiraly, J. D. Wood, D. Alducin, B. D. Myers, X. Liu,
B. L. Fisher, U. Santiago, J. R. Guest, M. J. Yacaman, A. Ponce, A. R. Oganov, M. C.
Hersam, and N. P. Guisinger. Synthesis of borophenes: Anisotropic, two-dimensional
boron polymorphs. Science, 350(6267):1513–1516, Dec 2015.
[77] B. M. Fregoso, Takahiro Morimoto, and Joel E. Moore. Quantitative relationship be-
tween polarization differences and the zone-averaged shift photocurrent. Phys. Rev. B,
96:075421, Aug 2017.
Bibliography 120
[78] Yong Xu, Binghai Yan, Hai-Jun Zhang, Jing Wang, Gang Xu, Peizhe Tang, Wenhui
Duan, and Shou-Cheng Zhang. Large-gap quantum spin hall insulators in tin films.
Phys. Rev. Lett., 111:136804, Sep 2013.
[79] B van den Broek, M Houssa, E Scalise, G Pourtois, V V Afanas’ev, and A Stesmans.
Two-dimensional hexagonal tin: ab initio geometry, stability, electronic structure and
functionalization. 2D Materials, 1(2):021004, 2014.
[80] Feng-feng Zhu, Wei-jiong Chen, Yong Xu, Chun-lei Gao, Dan-dan Guan, Can-hua Liu,
Dong Qian, Shou-Cheng Zhang, and Jin-feng Jia. Epitaxial growth of two-dimensional
stanene. Nat Mater, 14(10):1020–1025, 10 2015.
[81] Sumit Saxena, Raghvendra Pratap Chaudhary, and Shobha Shukla. Stanene: Atomically
thick free-standing layer of 2D hexagonal tin. Scientific Reports, 6:31073, 08 2016.
[82] Raghvendra Pratap Chaudhary, Sumit Saxena, and Shobha Shukla. Optical properties
of stanene. Nanotechnology, 27(49):495701, 2016.
[83] Melvin Lax. Symmetry Principles in Solid State and Molecular Physics. Dover Publi-
cations, 2012. Reprint of the John Wiley & Sons, 1974 edition.
[84] Morten Willatzen and Lok Yan Voon. The k·p Method. Springer, 2009.
[85] Nicola Marzari and David Vanderbilt. Maximally localized generalized wannier func-
tions for composite energy bands. Phys. Rev. B, 56:12847–12865, Nov 1997.
[86] Nicola Marzari, Arash A. Mostofi, Jonathan R. Yates, Ivo Souza, and David Vanderbilt.
Maximally localized wannier functions: Theory and applications. Rev. Mod. Phys.,
84:1419–1475, Oct 2012.
[87] Manuel Cardona and Fred H. Pollak. Energy-band structure of germanium and silicon:
The k · p method. Phys. Rev., 142:530–543, Feb 1966.
Bibliography 121
[88] Soline Richard, Frederic Aniel, and Guy Fishman. Energy-band structure of Ge, Si, and
GaAs: A thirty-band k · p method. Phys. Rev. B, 70:235204, Dec 2004.
[89] J. L. Cheng, C. Salazar, and J. E. Sipe. Optical properties of functionalized graphene.
Phys. Rev. B, 88:045438, Jul 2013.
[90] Golub and van Loan. Matrix Computations. John Hopkins University Press, 3rd edition,
1996.
[91] X. Gonze et al. Abinit: First-principles approach to material and nanosystem properties.
Comput. Phys. Commun., 180(12):2582 – 2615, 2009.
[92] X. Gonze et al. Recent developments in the abinit software package. Comput. Phys.
Commun., 205:106 – 131, 2016.
[93] D. R. Hamann. Optimized norm-conserving vanderbilt pseudopotentials. Phys. Rev. B,
88:085117, Aug 2013.
[94] David Vanderbilt. Soft self-consistent pseudopotentials in a generalized eigenvalue for-
malism. Phys. Rev. B, 41:7892–7895, 1990.
[95] P. E. Blochl. Projector augmented-wave method. Phys. Rev. B, 50:17953–17979, Dec
1994.
[96] G. Kresse and D. Joubert. From ultrasoft pseudopotentials to the projector augmented-
wave method. Phys. Rev. B, 59:1758–1775, 1999.
[97] S. Cahangirov, M. Topsakal, E. Akturk, H. Sahin, and S. Ciraci. Two- and one-
dimensional honeycomb structures of silicon and germanium. Phys. Rev. Lett.,
102:236804, Jun 2009.
[98] C. L. Kane and E. J. Mele. Quantum spin hall effect in graphene. Phys. Rev. Lett.,
95:226801, 2005.
Bibliography 122
[99] Claudio Chamon, Chang-Yu Hou, Christopher Mudry, Shinsei Ryu, and Luiz Santos.
Masses and majorana fermions in graphene. Physica Scripta, 2012(T146):014013,
2012.
[100] Rodrigo A. Muniz and J. E. Sipe. Optical response of effective models for crystals.
arXiv, 2017.
[101] Motohiko Ezawa. Spin-valley optical selection rule and strong circular dichroism in
silicene. Phys. Rev. B, 86:161407, 2012.
[102] Shinsei Ryu, Christopher Mudry, Chang-Yu Hou, and Claudio Chamon. Masses in
graphenelike two-dimensional electronic systems: Topological defects in order parame-
ters and their fractional exchange statistics. Phys. Rev. B, 80:205319, Nov 2009.
[103] Luiz Santos, Shinsei Ryu, Claudio Chamon, and Christopher Mudry. Topological qubits
in graphenelike systems. Phys. Rev. B, 82:165101, Oct 2010.
[104] Claudio Chamon, Chang-Yu Hou, Christopher Mudry, Shinsei Ryu, and Luiz Santos.
Masses and majorana fermions in graphene. Physica Scripta, 2012(T146):014013,
2012.
[105] Di Xiao, Gui-Bin Liu, Wanxiang Feng, Xiaodong Xu, and Wang Yao. Coupled spin and
valley physics in monolayers of MoS2 and other group-VI dichalcogenides. Phys. Rev.
Lett., 108:196802, May 2012.
[106] Cheng-Cheng Liu, Hua Jiang, and Yugui Yao. Low-energy effective hamiltonian involv-
ing spin-orbit coupling in silicene and two-dimensional germanium and tin. Phys. Rev.
B, 84:195430, Nov 2011.
[107] Motohiko Ezawa. A topological insulator and helical zero mode in silicene under an
inhomogeneous electric field. New J. Phys, 14(033003), 2012.
Bibliography 123
[108] R. Atanasov, A. Hache, J. L. P. Hughes, H. M. van Driel, and J. E. Sipe. Coherent control
of photocurrent generation in bulk semiconductors. Physical Review Letters, 76:1703,
1996.
[109] A. Hache, Y. Kostoulas, R. Atanasov, J. L. P. Hughes, J. E. Sipe, and H. M. van Driel.
Observation of coherently controlled photocurrent in unbiased, bulk gaas. Phys. Rev.
Lett., 78:306–309, Jan 1997.
[110] D. Sun, C. Divin, J. Rioux, J. E. Sipe, C. Berger, W. A. de Heer, P. N. First, and T. B.
Norris. Coherent control of ballistic photocurrents in multilayer epitaxial graphene using
quantum interference. Nano Lett., 10:1293, 2010.
[111] D. A. Bas, K. Vargas-Velez, S. Babakiray, T. A. Johnson, P. Borisov, T. D. Stanescu,
D. Lederman, and A. D. Bristow. Coherent control of injection currents in high-quality
films of Bi2Se3. Appl. Phys. Lett., 106:041109, 2015.
[112] Derek A. Bas, Rodrigo A. Muniz, Babakiray Sercan, David Lederman, J. E. Sipe, and
Alan D. Bristow. Identification of photocurrents in topological insulators. Opt. Express,
24:23583, 2016.
[113] Qiannan Cui and Hui Zhao. Coherent control of nanoscale ballistic currents in transition
metal dichalcogenide ReS2. ACS Nano, 9:3935, 2015.
[114] R. D. R. Bhat and J. E. Sipe. Optically injected spin currents in semiconductors. Physical
Review Letters, 85:5432–5435, Dec 2000.
[115] M. J. Stevens, A. L. Smirl, R. D. R. Bhat, J. E. Sipe, and H. M. van Driel. Coherent
control of an optically injected ballistic spin-polarized current in bulk GaAs. J. Appl.
Phys., 91:4382, 2002.
[116] Martin J. Stevens, Arthur L. Smirl, R. D. R. Bhat, Ali Najmaie, J. E. Sipe, and H. M. van
Driel. Quantum interference control of ballistic pure spin currents in semiconductors.
Phys. Rev. Lett., 90:136603, 2003.
Bibliography 124
[117] J. Hubner, W. W. Ruhle, M. Klude, D. Hommel, R. D. R. Bhat, J. E. Sipe, and H. M. van
Driel. Direct observation of optically injected spin-polarized currents in semiconductors.
Phys. Rev. Lett., 90:216601, 2003.
[118] Hui Zhao, Eric J. Loren, H. M. van Driel, and Arthur L. Smirl. Coherence control of
hall charge and spin currents. Phys. Rev. Lett., 96:246601, Jun 2006.
[119] Cuauhtemoc Salazar, J. L. Cheng, and J. E. Sipe. Coherent control of current injection
in zigzag graphene nanoribbons. Phys. Rev. B, 93:075442, Feb 2016.
[120] P. A. Roos, Xiaoqin Li, R. P. Smith, Jessica A. Pipis, T. M. Fortier, and S. T. Cun-
diff. Solid-state carrier-envelope phase stabilization via quantum interference control of
injected photocurrents. Opt. Lett., 30:735, 2005.
[121] P. A. Roos and S. T. Cundiff. Using quantum interference in semiconductors to control
the phase of ultrashort laser pulses. Laser Physics, 15:769, 2005.
[122] R.P. Smith, P. A. Roos, J. K. Wahlstrand, J. A. Pipis, M. B. Rivas, and S. T. Cundiff. Op-
tical frequency metrology of an iodine-stabilized he-ne laser using the frequency comb
of a quantum-interference-stabilized mode-locked laser. J. Res. Nat. Inst. Sci. Tech.,
112:289, 2007.
[123] P. A. Roos, Xiaoqin Li, J. R. Pipis, and S. T. Cundiff. Solid-state carrier-envelope-phase
noise measurements with intrinsically balanced detection. Opt. Exp., 12:4255, 2004.
[124] T. M. Fortier, P. A. Roos, D. J. Jones, S. T. Cundiff, R. D. R. Bhat, and J. E. Sipe. Carrier-
envelope phase-controlled quantum interference of injected photocurrents in semicon-
ductors. Phys. Rev. Lett., 92:147403, 2004.
[125] Steven T. Cundiff and Jun Ye. Colloquium. Rev. Mod. Phys., 75:325–342, Mar 2003.
[126] J. Ye and S.T. Cundiff. Femtosecond Optical Frequency Comb: Principle, Operation,
and Applications. Springer, Norwell, MA, 2005.
Bibliography 125
[127] P. A. Roos, Xiaoqin Li, J. A. Pipis, Tara M. Fortier, Steven T. Cundiff, R. D. R. Bhat, and
J. E. Sipe. Characterization of carrier-envelope phase-sensitive photocurrent injection in
a semiconductor. J. Opt. Soc. Am. B, 22:362, 2005.
[128] Kai Wang, Rodrigo A. Muniz, J.E. Sipe, and S.T. Cundiff. In preparation, 2017.
[129] Cuauhtemoc Salazar, Rodrigo A. Muniz, and J. E. Sipe. Effective model for the elec-
tronic and optical properties of stanene. Phys. Rev. Materials, 1:054006, 2017.
[130] N. Fraj, I. Saidi, S. Ben Radhia, and K. Boujdaria. Band structures of AlAs, GaP, and
SiGe alloys: A 30 k · p model. J. Appl. Phys., 102:053703, 2007.
[131] R. D. R. Bhat and J. E. Sipe. Excitonic effects on the two-color coherent control of
interband transitions in bulk semiconductors. Phys. Rev. B, 72:075205, 2005.
[132] E. G. Sauter. Nonlinear Optics. Wiley-Interscience, New York, 1996.
[133] Rodrigo A. Muniz, Cuauhtemoc Salazar, Kai Wang, J. E. Sipe, and S. T. Cundiff. Quan-
tum interference control of two- and three-photon processes in alαga1−αas. In peparation
for Phys. Rev. B, 2017.
[134] Yoichi Ando. Topological insulator materials. Journal of the Physical Society of Japan,
82(10):102001, 2013.
[135] A. Bansil, Hsin Lin, and Tanmoy Das. Colloquium: Topological band theory. Rev. Mod.
Phys., 88(2), Jun 2016.
[136] N.B. Baranova, A.N. Chudinov, and B.Ya. Zel’dovich. Polar asymmetry of photoion-
ization by a field with 〈E3〉 , 0. theory and experiment. Optics Communications, 79(1-
2):116–120, 1990.
[137] N.B. Baranova, A.N. Chudinov, and B.Ya. Zel’dovich. Reply to comments on “polar
asymmetry of photoionization by a field with 〈E3〉 , 0. theory and experiment”. Optics
Communications, 85(4):371, 1991.
Bibliography 126
[138] Ganesh R. et al. Bhimanapati. Recent advances in two-dimensional materials beyond
graphene. ACS Nano, 9(12):11509–11539, 2015.
[139] M. Born and R. Oppenheimer. Zur quantentheorie der molekeln. Annalen der Physik,
389(20):457–484, 1927.
[140] P. Brumer and M. Shapiro. Control of unimolecular reactions using coherent light.
Chemical Physics Letters, 126:541, 1986.
[141] Bo Cai, Shengli Zhang, Zhong Yan, and Haibo Zeng. Noncovalent molecular doping of
two-dimensional materials. ChemNanoMat, 1(8):542–557, 2015.
[142] Ce Chen, Yi-Yian Yin, and D. S. Elliott. Interference between optical transitions. Phys.
Rev. Lett., 64:507–510, 1990.
[143] Xianping Chen, Ruishen Meng, Junke Jiang, Qiuhua Liang, Qun Yang, Chunjian Tan,
Xiang Sun, Shengli Zhang, and Tianling Ren. Electronic structure and optical properties
of graphene/stanene heterobilayer. Phys. Chem. Chem. Phys., 18(24):16302–16309,
2016.
[144] Ching-Kai Chiu, Jeffrey C.Y. Teo, Andreas P. Schnyder, and Shinsei Ryu. Classification
of topological quantum matter with symmetries. Rev. Mod. Phys., 88(3), Aug 2016.
[145] Bo-Hung Chou, Zhi-Quan Huang, Chia-Hsiu Hsu, Feng-Chuan Chuang, Yu-Tzu Liu,
Hsin Lin, and Arun Bansil. Hydrogenated ultra-thin tin films predicted as two-
dimensional topological insulators. New Journal of Physics, 16(11):115008, 2014.
[146] Domenico Di Sante, Alessandro Stroppa, Paolo Barone, Myung-Hwan Whangbo,
and Silvia Picozzi. Emergence of ferroelectricity and spin-valley properties in two-
dimensional honeycomb binary compounds. Phys. Rev. B, 91:161401, Apr 2015.
[147] E. Dupont, P. B. Corkum, H. C. Liu, M. Buchanan, and Z. R. Wasilewski. Phase-
controlled currents in semiconductors. Phys. Rev. Lett., 74:3596–3599, 1995.
Bibliography 127
[148] Motohiko Ezawa. Monolayer topological insulators: Silicene, germanene, and stanene.
Journal of the Physical Society of Japan, 84(12):121003, 2015.
[149] C. Hartwigsen, S. Goedecker, and J. Hutter. Relativistic separable dual-space gaussian
pseudopotentials from H to Rn. Phys. Rev. B, 58:3641–3662, Aug 1998.
[150] D. R. Hartree. The wave mechanics of an atom with a non-coulomb central field. part i.
theory and methods. Mathematical Proceedings of the Cambridge Philosophical Soci-
ety, 24(1):89–110, 1928.
[151] Michel Houssa, Bas van den Broek, Konstantina Iordanidou, Anh Khoa Augustin Lu,
Geoffrey Pourtois, Jean-Pierre Locquet, Valery Afanas’ev, and Andre Stesmans. Topo-
logical to trivial insulating phase transition in stanene. Nano Research, 9(3):774–778,
2015.
[152] Indranil Paul and Gabriel Kotliar. Thermal transport for many-body tight-binding mod-
els. Phys. Rev. B, 67:115131, 2003.
[153] C. Kamal, Aparna Chakrabarti, and Motohiko Ezawa. Direct band gaps in group IV-VI
monolayer materials: Binary counterparts of phosphorene. Phys. Rev. B, 93(12), Mar
2016.
[154] Koke S. et al. Direct frequency comb synthesis with arbitrary offset and shot-noise-
limited phase noise. Nat. Photonics, 4:462, 2010.
[155] Gershon Kurizki, Moshe Shapiro, and Paul Brumer. Phase-coherent control of photocur-
rent directionality in semiconductors. Phys. Rev. B, 39:3435–3437, 1989.
[156] N. M. Lawandy. Comments on “polar asymmetry of photoionization by a field with
〈E3〉 , 0. theory and experiment” ”. Optics Communications, 85:369, 1991.
[157] E. A. Manykin. Quantum interference and coherent control. Laser Physics, 11:60, 2001.
Bibliography 128
[158] Lars Matthes, Olivia Pulci, and Friedhelm Bechstedt. Massive dirac quasiparticles in
the optical absorbance of graphene, silicene, germanene, and tinene. Journal of Physics:
Condensed Matter, 25(39):395305, 2013.
[159] L Matthes, O Pulci, and F Bechstedt. Optical properties of two-dimensional honeycomb
crystals graphene, silicene, germanene, and tinene from first principles. New Journal of
Physics, 16(10):105007, 2014.
[160] V. Meunier, A.G. Souza Filho, E.B. Barros, and M.S. Dresselhaus. Physical properties of
low-dimensional sp2-based carbon nanostructures. Rev. Mod. Phys., 88(2), May 2016.
[161] Hendrik J. Monkhorst and James D. Pack. Special points for brillouin-zone integrations.
Phys. Rev. B, 13:5188–5192, Jun 1976.
[162] Ana Suarez Negreira, William G. Vandenberghe, and Massimo V. Fischetti. Ab initio
study of the electronic properties and thermodynamic stability of supported and func-
tionalized two-dimensional sn films. Phys. Rev. B, 91:245103, Jun 2015.
[163] K. S. Novoselov. Electric field effect in atomically thin carbon films. Science,
306(5696):666–669, Oct 2004.
[164] R. Reiter, U. Derra, S. Birner, B. Terres, F. Libisch, J. Burgdorfer, and C. Stampfer.
Negative quantum capacitance in graphene nanoribbons with lateral gates. Phys. Rev. B,
89:115406, Mar 2014.
[165] F. Schwierz, J. Pezoldt, and R. Granzner. Two-dimensional materials and their prospects
in transistor electronics. Nanoscale, 7(18):8261–8283, 2015.
[166] M. Shapiro and P. Brumer. Principles of the Quantum Control of Molecular Processes.
Wiley-Interscience, New York, 2003.
[167] M. Shapiro and P. Brumer. Laser control of product quantum state populations in uni-
molecular reactions. Journal of Chemical Physics, 84:4103, 1986.
Bibliography 129
[168] M. Shapiro, J. W. Hepburn, and P. Brumer. Simplified laser control of unimolecular
reactions: Simultaneous (ω1, ω3) excitation. Chemical Physics Letters, 149:451, 1988.
[169] J. C. Slater and G. F. Koster. Simplified LCAO method for the periodic potential prob-
lem. Phys. Rev., 94:1498–1524, Jun 1954.
[170] Grazyna Staszewska, Przemysław Staszewski, Nathan E. Schultz, and Donald G. Truh-
lar. Many-body tight-binding model for aluminum nanoparticles. Phys. Rev. B,
71:045423, 2005.
[171] Weiwei Sun, Yunguo Li, Baotian Wang, Xue Jiang, Mikhail I. Katsnelson, Pavel Ko-
rzhavyi, Olle Eriksson, and Igor Di Marco. A new 2D monolayer BiXene, m 2 C (M =
Mo, Tc, Os). Nanoscale, 8(34):15753–15762, 2016.
[172] Sukosin Thongrattanasiri and F. Javier Garcıa de Abajo. Optical field enhancement by
strong plasmon interaction in graphene nanostructures. Phys. Rev. Lett., 110:187401,
Apr 2013.
[173] M. Torrent, F. Jollet, F. Bottin, G. Zerah, and X. Gonze. Implementation of the projector
augmented-wave method in the ABINIT code: Application to the study of iron under
pressure. Computational Materials Science, 42(2):337 – 351, 2008.
[174] Kuljit S. Virk and J. E. Sipe. Optical injection and terahertz detection of the macroscopic
berry curvature. Phys. Rev. Lett., 107:120403, 2011.
[175] J. K. Wahlstrand et al. Electric field-induced coherent control in gaas: polarization
dependence and electrical measurement. Opt. Express, 19:22563, 2011.
[176] Assa Auerbach and Daniel P Arovas. Introduction to Frustrated Magnetism, chapter 14,
pages 365–377. Springer, 2011.
[177] Andor Kormanyos, Viktor Zolyomi, Neil D. Drummond, Peter Rakyta, Guido Burkard,
and Vladimir I. Fal’ko. Monolayer MoS2: Trigonal warping, the Γ valley, and spin-orbit
coupling effects. Phys. Rev. B, 88:045416, Jul 2013.