Discontinuous Galerkin Time-Domain Method in Electromagnetics

Discontinuous Galerkin Time-Domain Method
in Electromagnetics: From Nanostructure

Simulations to Multiphysics Implementations
Item Type Book Chapter
Authors Dong, Ming;Chen, Liang;Li, Ping;Jiang, Lijun;Bagci, Hakan
Citation Dong, M., Chen, L., Li, P., Jiang, L., & Bagci, H. (2022).
Discontinuous Galerkin Time-Domain Method in
Electromagnetics: From Nanostructure Simulations to
Multiphysics Implementations. Advances in Time-Domain
Computational Electromagnetic Methods, 135–198. Portico.
https://doi.org/10.1002/9781119808404.ch4
Eprint version Post-print
DOI 10.1002/9781119808404.ch4
Publisher IEEE
Rights This is an accepted manuscript version of a paper before final

publisher editing and formatting. Archived with thanks to IEEE.
Download date 2024-03-08 23:41:12
Link to Item http://hdl.handle.net/10754/685980

Chapter 1
Discontinuous Galerkin
Time-domain Method in
Electromagnetics: From
Nanostructure Simulations to
Multiphysics Implementations
Ming Dong,1 Liang Chen,1 Ping Li,2 Lijun Jiang,3 and Hakan
Bagci1
1
Electrical and Computer Engineering (ECE) Program, Division of Computer, Electrical,
and Mathematical Science and Engineering (CEMSE), King Abdullah University of
Science and Technology (KAUST), Thuwal, Saudi Arabia
2
Department of Electrical Engineering, Shanghai Jiao Tong University, Shanghai, China
3
Department of Electrical and Electronic Engineering, The University of Hongkong,
Hongkong, China
Abstract: In the last few decades, the discontinuous Galerkin time-domain

(DGTD) method has become widely popular in various fields of engineering
due the fact that it benefits from computational advantages that come
with finite volume and finite element formulations. Similarly, in the field
of computational electromagnetics, the superiority of the DGTD method
has been quickly recognized after first few works on its formulation and
1
implementation to solve Maxwell equations. With further developments in
more recent years, the DGTD method has become one of the preeminent
solutions to tackle a wide variety of challenging large scale electromagnetic
problems including those that require multiphysics modeling.
This chapter starts with a brief introduction to the DGTD method. This
introduction provides the fundamentals of numerical flux, discretization
techniques that rely on vector and nodal basis functions, and incorpora-
tion of absorbing boundary conditions. This is followed by descriptions of
a time-domain boundary integral(TDBI) scheme, which replaces absorb-
ing boundary conditions within the DGTD method, and a multi-step time
integration technique, which uses different time step sizes for the DGTD
and TDBI parts. Numerical results show that both techniques significantly
improve the efficiency, accuracy, and stability of the traditional DGTD
method. Then, the chapter continues with the applications of the DGTD
method to several real-life practical problems. More specifically, it describes
various novel techniques developed to enable the application of the DGTD
method to electromagnetic analysis of nanostructures and graphene-based
devices, and multiphysics simulation of optoelectronic antennas and source
generators. For each application, several numerical examples are provided
to demonstrate the accuracy, efficiency, and robustness of the developed
techniques.
Keywords: discontinuous Galerkin time-domain method, Maxwell’s equa-

tions, electromagnetic simulation, multiphysics simulation, nanostructures,
optoelectronic devices
2
1.1. Introduction to the Discontinuous
Galerkin Time-domain Method
Since the discontinuous Galerkin (DG) method was first applied to solve the
neutron transport equation by Reed and Hill in 1973 [Reed and Hill, 1973], it
has been re-formulated and extended to numerous applications including but
not limited to gas dynamics, semiconductor physics, magneto-hydrodynamics,
fluid dynamics, and electromagnetics [Cockburn et al., 2000]. In the field of
electromagnetics, one of the very first works on a DG method formulated
and implemented to solve the time-domain Maxwell equations is from Bourdel
[Bourdel et al., 1991]. This first DG method uses a zeroth-order (piecewise
constant) discretization scheme but several high-order discontinuous Galerkin
time-domain (DGTD) methods have been developed rather quickly after that.
These methods can account for unstructured meshes and uses various Runge-
Kutta schemes to carry out the time integration [Hesthaven and Warburton,
2002, Kabakian et al., 2004, Fezoui et al., 2005, Shu, 2016a]. Since then, other
methods have been developed increasing the popularity of DGTD in the field
of computational electromagnetics [Hesthaven and Warburton, 2008, Dosopou-
los and Lee, 2010b, Garcı́a De Abajo, 2013, Li et al., 2014a, Lin et al., 2016,
Harmon et al., 2016, Angulo et al., 2015, Yan et al., 2016, Ren et al., 2017,
Chen et al., 2018, Sirenko et al., 2018, Chen and Bagci, 2020b]. Finally, in more
recent years, DGTD has successfully become a powerful numerical technique
to tackle a wide variety of challenging large scale electromagnetic problems
including those that require multiphysics modeling [Gedney et al., 2007, Jin,
2002, Li and Jiang, 2013a,c, Li et al., 2014b, Li and Jiang, 2013b].
DGTD combines advantages of both the finite volume time-domain (FVTD)
3
method and the finite element time-domain (FETD) method [Hesthaven and
Warburton, 2008]. Like FVTD, the information exchange between neighboring
elements is realized through the use of numerical flux. This approach localizes
all spatial operations to individual discretization elements. The resulting local
element-level mass matrix blocks are factorized and stored very efficiently be-
fore time integration/marching starts. Indeed, if an explicit time integration
scheme is used, DGTD becomes an efficient solver with very small memory
footprint (compared to classical time domain FEM). In addition, the block di-
agonal mass matrix makes DGTD very suitable for parallelization. It can easily
be ported on high performance distributed memory systems or multi-threaded
GPUs to solve large scale electromagnetic problems [Klöckner et al., 2009,
Dosopoulos et al., 2011]. Finally, locality of the spatial operations makes it
easier to incorporate h- and/or p-adaptive refinement techniques [Shu, 2016a,
Hesthaven and Warburton, 2008] as well as non-conformal meshing strategies
[Hesthaven and Warburton, 2002].
DGTD also comes with the flexibility of choosing different time-stepping
schemes in different subdomains of the computation domain [Angulo et al.,
2015, Lin et al., 2016]. For example, an efficient explicit scheme can be used
in subdomains with coarser meshes while an unconditionally-stable implicit
solver can be used in subdomains with dense meshes. This approach allows for
the same large time step to be used throughout whole computation domain
potentially reducing the overall computation time.
As a final word, just like other time domain methods, DGTD offers sev-
eral benefits in simulating various electromagnetic problems: It can provide
broadband data with a single execution of the simulation and it allows for
hybridization with circuit solvers that can account for nonlinear components
4
[Li and Jiang, 2013c, Li et al., 2014a, Li and Jiang, 2013a].
In this chapter, we first describe the basic formulation of underlying the
DGTD developed for solving the Maxwell curl equations using vector basis
functions and explain how different boundary conditions can be incorporated
within the solver using numerical flux. This is followed by the description
of a nodal DGTD scheme, where higher-order polynomials are used as basis
functions to expand the three components of the electric and magnetic fields.
Then, the chapter continues with the applications of DGTD to several real-
life practical problems. More specifically, novel techniques developed to enable
the application of DGTD to electromagnetic analysis of nanostructures and
multiphysics simulation of photoconductive devices/antennas are described.
For each application, we present several numerical examples to demonstrate
the accuracy, efficiency and robustness of the developed techniques.
1.1.1. DGTD formulation for Maxwell equations

Analysis of transient electromagnetic field interactions with arbitrarily-shaped
objects calls for numerical solution of Maxwell equations for electric field E(r, t)
and magnetic field H(r, t) [Dosopoulos and Lee, 2010b] or the wave equation
for E(r, t) or H(r, t) [Sun et al., 2017]. Here, we focus on the formulation of
the DGTD schemes for the former and refer the readers to [Sun et al., 2017]
and references therein for the details for DGTD schemes developed for solving
the wave equation. Let J(r, t) represent the volumetric current density located
in the domain of interest, then Maxwell curl equations can be written as [Jin,
5
2002]
∂E(r, t)
ε(r) − ∇ × H(r, t) + J(r, t) = 0
∂t (1.1)
∂H(r, t)
µ(r) + ∇ × E(r, t) = 0
∂t
where ε(r) and µ(r) denote the permittivity and permeability respectively.
Note that for the sake of simplicity in the presentation, in the rest of the
chapter, dependence of the variables on location r and time t is not explicitly
stated (unless it is needed to clarify the formulation/notation).
To develop the DGTD formulation, the Maxwell equations above should
be re-expressed in the conservation form as [Hesthaven and Warburton, 2008]
∂U
+ ∇ · F(U) + S = 0. (1.2)
∂t
This can be achieved by defining field vector U = [E, H]T , excitation vector
S = [J/ε, 0]T and the flux vector
     
 −x̂ × H/ε   −ŷ × H/ε   −ẑ × H/ε 
F(U) =   x̂ +   ŷ +   ẑ. (1.3)
x̂ × E/µ ŷ × E/µ ẑ × E/µ
The flux vector F(U) can be further expressed as a multiplication of the purely
material dependent coefficient matrix A and the field vector U:
F(U) = AU (1.4)
6
where
     
0 − 1ε Cx 0 − 1ε Cy 0 − 1ε Cz
A=  x̂ +   ŷ +   ẑ
     
1 1 1
µ Cx 0 µ Cy 0 µ Cz 0
     
(1.5)
 0 0 0   0 0 1   0 −1 0 
     
Cx = 
 0 0 −1  , Cy =  0 0 0  , Cz =  1 0 0  .
    
     
0 1 0 −1 0 0 0 0 0
The coefficient matrix A determines the characteristics of the Maxwell system.
It is shown later in this section that the eigenvalues of this flux matrix AU
are real (a necessary condition for the hyperbolic system) and they are the
characteristic speed associated with the Maxwell system.
Let Ω and ∂Ω denote the computation domain of interest and its boundary,
respectively. Ω can be divided into a set of non-overlapping tetrahedrons Ωi
with boundary ∂Ωi , i.e., Ω = ∪Ωi . In each mesh element Ωi , i = 1, · · · , N ,
after applying Galerkin testing [Hesthaven and Warburton, 2008] to equation
(1.1) in space with basis function Λi = [Ψei , Ψhi ] (Ψei and Ψhi are used to expand
E and H in element i respectively as explained later) and using the divergence
(Gauss) theorem twice, we obtain
Z
∂U
Λi · + ∇ · F(U) + S dr
Ωi ∂t
Z (1.6)
=− {n̂i · [F ∗ (U) − F(U)]} · Λi dr
∂Ωi
where n̂i is the outward pointing unit normal vector of ∂Ωi and n̂i · F ∗ (U) is
the numerical flux. Since the solutions in adjacent elements are allowed to be
discontinuous, F ∗ must be carefully defined to ensure the uniqueness of the
solution.
7
The definition of the numerical flux starts by solving the corresponding Rie-
mann problem with Rankine-Hugoniot jump condition. To obtain the Rankine-
Hugoniot jump condition [Sankaran, 2007, Jeffrey, 2011], the characteristic plot
of the Maxwell system has to be obtained in advance from the flux matrix de-
fined as
   
 εi 0   0 Ci 
Ă = n̂i · A =  ·  (1.7)
0 µi −C i 0
where εi = diag(εi , εi , εi ), µi = diag(µi , µi , µi ) and
 
 0 ni,z −ni,y 
 
Ci = 
 −ni,z 0 ni,x . (1.8)
 
ni,y −ni,x 0
The flux matrix Ă has six eigenvalues given by [Sankaran, 2007, Jeffrey, 2011]
λ1,4 = 0
λ2,5 = √1 = +ci (1.9)

εi µi
λ3,6 = − √ε1i µi = −ci
and they contain important information about the solution of the Riemann
problem. For hyperbolic system, the solutions remain unchanged while prop-
agating along the characteristic plot xn = λi t. Across each solution, Rankine-
Hugoniot jump condition holds, that is [Hesthaven and Warburton, 2008,
8
Sankaran, 2007, Jeffrey, 2011]
ĂU∗ − ĂUi = −ci (U∗ − Ui ) (1.10)
ĂU∗∗ − ĂU∗ = 0 (1.11)
ĂU∗∗ − ĂUj = cj (U∗∗ − U∗ ) (1.12)
where j ∈ N (i), N (i) is the set of elements that are neighboring element i, and
U∗ and U∗∗ represent two intermediate states. The explicit jump relations for
fields can be simplified as
1. Jump across the characteristic plot xn = −ci t
1
n̂i × (E∗ − Ei ) = −ci (H∗ − Hi ) (1.13)
µi
1
n̂i × (Hi − H∗ ) = −ci (E∗ − Ei ). (1.14)
εi
2. Jump across the characteristic plot xn = 0
n̂i × (E∗∗ − E∗ ) = 0 (1.15)
n̂i × (H∗∗ − H∗ ) = 0. (1.16)
3. Jump across the characteristic plot xn = cj t
1
n̂i × (Ej − E∗∗ ) = cj (Hj − H∗∗ ) (1.17)
µj
1
n̂i × (H∗∗ − Hj ) = cj (Ej − E∗∗ ). (1.18)
εj
With the above jump conditions, we obtain the explicit expression of the
9
numerical flux n̂i · F ∗ (U) = [−n̂i × H∗ , n̂i × E∗ ]T as
(Zi Hi + Zj Hj ) + n̂i × (Ei − Ej )

n̂i × H∗ = n̂i × (1.19)
Zi + Zj
(Yi Ei + Yj Ej ) − n̂i × (Hi − Hj )
n̂i × E∗ = n̂i × (1.20)
Yi + Yj
p
where Zi,j = µi,j /εi,j and Yi,j = 1/Zi,j denote the characteristic impedance
and admittance, respectively. This numerical flux defined by equation (1.19)
and (1.20) is termed “upwind flux”, which is the most commonly used type
of numerical flux the DGTD methods used for electromagnetic simulation. In
the rest of this chapter, it is assumed that upwind numerical flux is unless
otherwise is stated.
Substituting (1.19) and (1.20) into (1.6), we get two separate equations as
Z Z
∂Ei
εi − ∇ × Hi + Ji · Ψei,k dr = Ψei,k · [n̂i × (H∗ − Hi )] dr (1.21)
Ωi ∂t ∂Ω
Z Z i
∂Hi
µi + ∇ × Ei · Ψhi,k dr = Ψhi,k · [n̂i × (Ei − E∗ )] dr. (1.22)
Ωi ∂t ∂Ωi
Next, the (local) electric and magnetic fields in element i, Ei and Hi , are
expanded using vector basis functions Ψei,l and Ψhi,l , respectively as [Jin, 2002]
K
X K
ei,l (t)Ψei,l (r), hi,l (t)Ψhi,l (r).
P
Ei (r, t) = Hi (r, t) = (1.23)
l=1 l=1
where ei,l and hi,l are time-dependent unknown coefficients of Ψei,l and Ψhi,l ,
respectively. Note that the numbers of basis functions used for expanding Ei
and Hi do not have to be the same. Similarly, the number of basis functions
can be different in different elements. But for the sake of simplicity in the
notation, we assume that all elements use the same number of basis functions
10
for both Ei and Hi , and that number is K.
With (1.19) to (1.23), a semi-discrete matrix system can be obtained [Li
et al., 2014b, Li and Jiang, 2015]
∂ei
Mei = Sei hi − ji + Fee ee eh eh
ii ei − Fij ej − Fii hi + Fij hj (1.24)
∂t
∂hi
Mhi = −Shi ei + Fhh hh he he
ii hi − Fij hj + Fii ei − Fij ej (1.25)
∂t
where ei and hi are time-dependent vectors that store unknown basis function
coefficients, Mei and Mhi are the mass matrices, Sei and Shi are the stiffness
ee/eh
matrices, ji is a vector that stores the tested current source, and Fii/ij and
hh/he
Fii/ij are the flux matrices. The entries of these matrices and vectors are given
11
by [Li et al., 2014b, Li and Jiang, 2015]:
[ei ]l = ei,l , [hi ]l = hi,l

Z
e
[Mi ]kl = Ψei,k · εi Ψei,l dr, [Mhi ]kl = Ωi Ψhi,k · µi Ψhi,l dr
R
Ω
Z i
[Sei ]kl = Ψei,k · ∇ × Ψhi,l dr, [Shi ]kl = Ωi Ψhi,k · ∇ × Ψei,l dr
R
Ω
Z i
[ji ](k) = Ψei,k · Ji dr
Ωi
Z
1
[Fee ]
ii kl = Ψe · n̂i × (n̂i × Ψei,l ) dr
Zi + Zj ∂Ωi,j i,k
Z
1
[Fee ]
ij kl = Ψe · n̂i × (n̂i × Ψej,l ) dr
Z
Zj
[Feh ]
ii kl = Ψe · n̂i × Ψhi,l dr (1.26)
Z
eh Zj
[Fij ]kl = Ψe · n̂i × Ψhj,l dr
Z
hh 1
[Fii ]kl = Ψh · n̂i × (n̂i × Ψhi,l ) dr
Yi + Yj ∂Ωi,j i,k
Z
hh 1
[Fij ]kl = Ψh · n̂i × (n̂i × Ψhj,l ) dr
Z
he Yj
[Fii ]kl = Ψh · n̂i × Ψei,l dr
Z
he Yj
[Fij ]kl = Ψh · n̂i × Ψej,l dr.
Note that the semi-discrete equation (1.24) and (1.25) are integrated in time
for all i = 1, · · · , N . Before the time integration/time marching starts, Mei and
Mhi , i = 1, · · · , N , each of which is only of dimension K × K, are inverted
and stored. Consequently, if an explicit scheme is used for time integration
(as described below), the resulting DGTD solver is very efficient with a small
memory footprint (compared to traditional time domain FEM).
12
Next, we describe how the semi-discrete system in equation (1.24) and
(1.25) can be integrated in time to obtain the unknown coefficients ei and hi .
We use two examples: Classical Runge-Kutta method [Hesthaven and War-
burton, 2008] and the leap-frog scheme that relies on central finite difference
approximation [Li et al., 2014b].
The fourth-order Runge-Kutta: The system in (1.24) and (1.25) is in the
form of a first-order ordinary differential equation (ODE) that can be expressed

∂u
as ∂t = f (u). Using the fourth order Runge-Kutta method, the solution of this
ODE at time step tn+1 = (n + 1)∆t = tn + ∆t can be expressed as [Hesthaven
and Warburton, 2008]
∆t
un+1 = un + (an + 2bn + 2cn + dn ) (1.27)
6
where an = f (un ), bn = f (un + an ∆t/2), cn = f (un + bn ∆t/2), and dn =
f (un + cn ∆t).
Leap-frog scheme: One can use the central finite difference method to ap-
proximate the two time derivatives in (1.24) and (1.25) at staggered time points
[Li et al., 2014b]:
∂ei en+1 − eni

|n+ 1 ≈ i + O(∆t2 ) (1.28)
∂t 2 ∆t
n+ 32 n+ 12
∂hi h − hi
|n+1 ≈ i + O(∆t2 ). (1.29)
∂t ∆t
Inserting equation (1.28) and (1.29) into equation (1.24) and (1.25) yields
13
the update equations resulting in the so-called leap frog time marching scheme:
Mei en+1
i = Mei eni
(1.30)
n+ 1 n+ 1 n+ 21 n+ 12
+∆t Sei hi 2 − ji 2 + Fee n
ii ei − Fee n
ij ej − Feh h
ii i + Feh h
ij j
n+ 32 n+ 12
Mhi hi = Mhi hi
1 1
(1.31)
h n+1 hh n+ 2 hh n+ 2 he n+1 he n+1
−∆t Si ei − Fii hi + Fij hj − Fii ei + Fij ej .
We should emphasize here that Runge-Kutta method and leap-frog scheme
are used as examples here because of their popularity. Other explicit schemes
can also be used to integrate (1.24) and (1.25) in time [Liu et al., 2012, Angulo
et al., 2015].
1.1.2. Boundary conditions

Boundary conditions are incorporated within DGTD by modifying the numer-
ical flux. The expressions of the upwind flux in equation (1.19) and (1.20) show
that the computation of numerical flux for a given element requires the field
values on that element and its neighbors (this can be thought of as a decom-
position of numerical flux into incoming flux and outgoing flux [Jeffrey, 2011]).
However, if an element is “touching” boundary, the corresponding neighbor
element does not exist and the incoming flux is not immediately available in
the form of field values. In this case, the relevant boundary condition is used
to modify the incoming flux expression and update the numerical flux of the
element touching the boundary. Next, we describe how absorbing boundary
conditions (ABCs) and the boundary conditions for perfect electrically con-
ducting (PEC) and perfect magnetically conducting (PMC) surfaces can be
14
incorporated within DGTD using the numerical flux.
1.1.2.1. Absorbing boundary conditions(ABCs)

In simulations of electromagnetic fields in open region problems, the unbounded
background domain has to be truncated into a finite computation domain [Jin,
2002]. On the truncation surface, a boundary condition that imitates the radia-
tion relation has to be enforced. The first-order Silver-Muller ABC (SM-ABC)
[Mur, 1981, Dosopoulos and Lee, 2010a]
n̂i × Ej = −cj µj n̂i × (n̂i × Hj )

(1.32)
n̂i × Hj = cj εj n̂i × (n̂i × Ej )
is widely used due to its simplicity and ease of implementation even though
its performance is limited. The implementation of SM-ABC in DGTD is done
by simply setting the field values Ej and Hj in the numerical flux expres-
sion to zero since all outgoing waves are assumed to be absorbed without any
reflection. Then, the numerical flux can be expressed as
Zi Hi + n̂i × Ei
n̂i × H∗ = n̂i ×
2Zi
(1.33)
Y E
i i − n̂i × Hi
n̂i × E∗ = n̂i × .
2Yi
Note that in Section 1.1.3, a more accurate approach to truncating computation
domains is described. This approach hybridizes the DGTD scheme with a time
domain boundary integral (TDBI) method, but just like ABCs, the field from
TDBI are coupled to DGTD via the use of the numerical flux. Please read
Section 1.1.3 for details on the implementation of this hybrid method.
15
1.1.2.2. Boundary condition on perfect electrically con-
ducting (PEC) surfaces

On a PEC surface, tangential component of Ej is equal to zero and we set
the characteristic impedance Zj = 0, which simplify the expression of the
numerical flux to
1
n̂i × H∗ = n̂i × (Hi + n̂i × Ei )
Zi (1.34)
∗
n̂i × E = 0.
Alternatively, this expression can be obtained by choosing Zj = Zi , Ej = −Ei ,
and Hj = Hi .
1.1.2.3. Boundary condition on perfect magnetically con-
ducting (PMC) surfaces

On a PMC surface, tangential component of Hj is equal to zero and we set
the characteristic admittance Yj = 0, which simplify the expression of the
numerical flux to
1
n̂i × E∗ = n̂i × (Ei − n̂i × Hi )
Yi (1.35)
∗
n̂i × H = 0.
Alternatively, this expression can be obtained by choosing Yj = Yi , Ej = Ei ,
and Hj = −Hi .
16
1.1.3. Hybridization with Time-domain Bound-
ary Integral (TDBI) method

As briefly mentioned before, when analyzing electromagnetic fields in open
region problems (e.g., computation of fields scattered from an object) using
DGTD, the unbounded background medium has to be truncated into a finite
computation domain [Jin, 2002]. On the truncation surface, a boundary con-
dition that imitates/approximates the radiation relation has to be enforced.
To this end, one can use mathematically exact absorbing boundary conditions
(EACs) [Hagstrom et al., 2010, Sirenko et al., 2013] and their localized but ap-
proximate versions (ABCs) [Mur, 1981, Higdon, 1986]. EACs are constructed
using the outgoing wave modes defined on the truncation surface and there-
fore they can only be enforced on planar and/or spherical boundaries. Their
accuracy depends on the number of these modes, which cannot be easily esti-
mated for a prescribed accuracy of the solution [Hagstrom et al., 2010, Sirenko
et al., 2013]. Localized ABCs are computationally less expensive but their ac-
curacy significantly deteriorates for waves obliquely incident on the truncation
surface [Mur, 1981, Higdon, 1986]. Another approach to imitate the radiation
relation is to “wrap” the computation domain with perfectly matched layers
(PMLs) [Berenger, 1994, Gedney, 1996, Chen et al., 2020b]. The attenuation
in PML does not depend on the incidence angle [Fezoui et al., 2005], but it
decreases at low frequencies [Gedney et al., 2012, Gedney and Zhao, 2010]
making the PML less accurate and/or more computationally expensive. Addi-
tionally, PML-based truncation methods might introduce late time instabilities
in the solution [Abarbanel et al., 2002] and reduce its accuracy due to spurious
reflections at the PML interface [Berenger, 1994, Gedney, 1996].
17
{Einc (r,t), H inc (r,t)}
∂Ω
ε 0 µ0 Jh
∂Γ Je
∂ΩT
ε (r) µ (r)
Figure 1.1: Illustration of the scatterer, computation domain boundary ∂Ω,

Huygens surface ∂Γ, and TF/SF surface ∂ΩT , and the excitation. Reproduced
with permission from [Li et al., 2014b].
In this section, we describe a scheme [Li et al., 2014b, Dong et al., 2020,
2022] that combines DGTD with TDBI to truncate computation domains with-
out suffering from the shortcomings of the other methods briefly described
above. TDBI represents the fields on the truncation surface in the form of a
retarded time boundary integral defined over a Huygens surface enclosing the
scatterer. This results in a mathematically exact replacement for the radiation
condition and ensures that the accuracy of the solution is limited only by the
discretization error. In addition, since the truncation surface is allowed to con-
form to the scatterer’s shape and can be located very close to its surface, the
resulting computation domain can be as small as possible. In the rest of this
section, the formulation underlying this hybrid method, which is abbreviated
as DGTDBI, is described.
The scattering problem under consideration is illustrated in Figure 1.1.
The permittivity and permeability inside the scatterer and in the background
medium are represented by ε(r) and µ(r) and ε0 and µ0 , respectively. The
scatterer is excited by the electromagnetic field Einc (r, t), Hinc (r, t) , and the

18
total and scattered field (TF/SF) technique is used to introduce the excitation
in the computation domain. In Figure 1.1, Ω is the computation domain, ∂Ω
is the boundary of the computation domain (truncation surface), ∂ΩT is the
boundary of the TF region, and ∂Γ is the Huygens surface that is introduced to
facilitate TDBI. The spatial discretization used in Ω is exactly same as the one
described in Section 1.1.1. DGTD solves for the scattered field in the region
between ∂Ω and ∂ΩT , and solves for the total field in the region enclosed by
∂ΩT .
In this set up, let the upwind numerical flux [Fezoui et al., 2005, Dosopoulos
and Lee, 2010b, Hesthaven and Warburton, 2008, Dosopoulos and Lee, 2010a,
Alvarez et al., 2010] associated with element i be expressed as
Zi Hi + n̂i,m × Ei + Z̃ H̃ − n̂i,m × Ẽ
n̂i,m × H∗m = n̂i,m ×
Zi + Z̃
(1.36)
Yi Ei − n̂i,m × Hi + Ỹ Ẽ + n̂i,m × H̃
n̂i,m × E∗m = n̂i,m ×
Yi + Ỹ
where n̂i,m is the outward pointing unit normal vector on ∂Ωi,m , which is
the m-th facet of the boundary ∂Ωi of element i. In (1.36), the outgoing flux
involves Ei and Hi (fields in Ωi ) and the incoming flux involves Ẽ and H̃ (fields
“external” to Ωi ). The incoming flux is used to (i) establish the “connectivity”
between fields in Ωi and other elements “touching” ∂Ωi,m , (ii) enforce the
boundary condition on ∂Ω, and (iii) introduce {Einc , Hinc } on ∂ΩT [Alvarez
19
et al., 2010]. Accordingly, Ẽ and H̃ in (1.36) are selected as


 E∂Ω ∂Ωi,m ∈ ∂Ω
 i,m


Ẽ =
 Ej ± Einc
i,m ∂Ωi,m ∈ ∂ΩT , (1.37)


Ej else




 H∂Ω ∂Ωi,m ∈ ∂Ω
 i,m


H̃ =
 Hj ± Hinc
i,m ∂Ωi,m ∈ ∂ΩT . (1.38)


 Hj

else
Here, j is the index of the element that neighbors element i via ∂Ωi,m , E∂Ω
i,m
and H∂Ω inc inc

i,m are the fields enforced on ∂Ωi,m ∈ ∂Ω, and Ei,m and Hi,m are the
incident fields computed on ∂Ωi,m ∈ ∂ΩT , and finally, ‘+’/‘−’ sign should be
selected if Ωi is inside the TF/SF region. Similarly, Z̃ and Ỹ are selected as

 
 Z0 ∂Ωi,m ∈ ∂Ω
  Y0 ∂Ωi,m ∈ ∂Ω

Z̃ = , Ỹ = (1.39)
 Zj

else  Yj

else
p
where Z0 = µ0 /ε0 and Y0 = 1/Z0 are the wave impedance and admittance
in the background medium.
E∂Ω ∂Ω
i,m and Hi,m [as they appear in (1.37) and (1.38)] on ∂Ωi,m ∈ ∂Ω are
computed using the TDBI method as described next. Let ∂Ωi′ ,m′ represent the
triangular facets that discretize ∂Γ. Here, i′ runs over the indices of elements
that are outside the volume enclosed by ∂Γ and have at least three nodes
residing on ∂Γ, and m′ runs over the indices of each element’s facets that are
described by these nodes. On facet ∂Ωi′ ,m′ , the equivalent electric and magnetic
surface currents, Jhi′ ,m′ and Jei′ ,m′ [Jiao et al., 2001, Shanker et al., 2005] can
20
be expressed as
K
X
Jhi′ ,m′ (r, t) = fih′ ,l′ (t)n̂i′ ,m′ (r) × Ψhi′ ,l′ (r)
l′ =1
(1.40)
XK
Jei′ ,m′ (r, t) = − fie′ ,l′ (t)n̂i′ ,m′ (r) × Ψei′ ,l′ (r).
l′ =1
Then, E∂Ω ∂Ω h
i,m and Hi,m on ∂Ωi,m ∈ ∂Ω are constructed using currents Ji′ ,m′ and
Jei′ ,m′ [Jiao et al., 2001, Shanker et al., 2005]:
XXh i
E∂Ω
i,m (r, t) = µ0 Li′ ,m′ (Jhi′ ,m′ ) − Ki′ ,m′ (Jei′ ,m′ )
i′ m′
XXh i (1.41)
H∂Ω
i,m (r, t) = ε0 Li′ ,m′ (Jei′ ,m′ ) + Ki′ ,m′ (Jhi′ ,m′ )
i′ m′
Here, the operators Li′ ,m′ and Ki′ ,m′ are given by [Jin, 2002, Dong et al., 2020,
2022]
∂t J(r′ , t − R/c0 ) ′
Z
1
Li′ ,m′ (J) = − dr
4π∂Ωi′ ,m′ R
Z t−R/c0 ′
c20 ∇ · J(r′ , t′ ) ′ ′
Z
+ ∇ dt dr (1.42)
4π ∂Ωi′ ,m′ 0 R
J(r′ , t − R/c0 ) ′
Z
1
Ki′ ,m′ (J) = ∇ × dr (1.43)
4π ∂Ωi′ ,m′ R
where R = |r − r′ | is the distance between the integration point r′ ∈ ∂Ωi′ ,m′

√
and the observation point r and c0 = 1/ ε0 µ0 is the speed of light in the
background medium.
To account for TDBI, the semi-discrete system in (1.24) and (1.25) (Section
21
1.1.3) is updated using equation (1.36) - (1.43):
X
εi Mei ∂t ei =[Seh eh ee
i − Fii ]hi + Fii ei + [Feh ee
ij hj − Fij ej ]
j
− tee eh ee eh
i + ti ∓ bi ± bi (1.44)
X
µi Mhi ∂t hi =[−She he hh
i + Fii ]ei + Fii hi − [Fhe hh
ij ej + Fij hj ]
j
− thh he hh he
i − ti ∓ bi ∓ bi (1.45)
where the mass matrices, stiffness matrices and flux matrices are the same
as those in Section 1.1.1. The non-zero entries of the remaining matrices and
vectors are given by

Z
[bpp Ψpi,k · n̂i,m × (n̂i,m × bpi,m )dr
X
i ]k = Ap
m m ∂Ωi,m ∈∂ΩT
Z
[bpp̃ Ψpi,k · n̂i,m × bp̃i,m dr
X
i ]k = Bp
m m ∂Ωi,m ∈∂ΩT
Z (1.46)
[tpp Ψpi,k ipi,m )dr
X
i ]k = Ap · n̂i,m × (n̂i,m ×
m m ∂Ωi,m ∈∂Ω
Z
[tpp̃ Ψpi,k · n̂i,m × ip̃i,m dr
X
i ]k = Bp
m m ∂Ωi,m ∈∂Ω
In the above expressions, p ∈ {e, h}, p̃ = h for p = e and p̃ = e for p = h,
Aem = 1/(Zi + Z̃), Ahm = 1/(Yi + Ỹ ), Bm

e = Z̃/(Z + Z̃), B h = Ỹ /(Y + Ỹ ),
i m i
bei,m = Einc h inc

i,m , bi,m = Hi,m and
(K K
)
XX X X
iei,m = µ0 Li′ ,m′ (fih′ ,l′ n̂i′ ,m′ × Ψhi′ ,l′ ) + Ki′ ,m′ (fie′ ,l′ n̂i′ ,m′ × Ψei′ ,k′ )
i′ m′ l′ =1 l′ =1
(K K
)
XX X X
ihi,m = Ki′ ,m′ (fih′ ,l′ n̂i′ ,m′ × Ψhi′ ,l′ ) − ε0 Li′ ,m′ (fie′ ,l′ n̂i′ ,m′ × Ψei′ ,l′ ) .
i′ m′ l′ =1 l′ =1
22
Figure 1.2: The illustration of temporal interpolation process.
1.1.4. Multi-time stepping scheme of DGTDBI

The semi-discrete system in (1.44) and (1.45) is numerically integrated in time
to obtain the samples of ei and hi . It should be emphasized that the time step
size used by DGTD and TDBI does not need to be same. Let ∆tDG represent
the time step size of the integration scheme used. To ensure stability, ∆tDG
has to satisfy a Courant-Friedrichs-Lewy (CFL)-like condition that depends on
the order and type of spatial discretization and time integration schemes as de-
scribed in detail in [Fezoui et al., 2005, Dosopoulos and Lee, 2010b, Hesthaven
and Warburton, 2008, Gedney et al., 2007, Dosopoulos and Lee, 2010a, Alvarez
et al., 2010]. On the other hand, the time step size used by TDBI, denoted by
∆tBI , is only required to resolve the maximum frequency of the excitation and
is independent from the spatial discretization [Yilmaz et al., 2004], [Bagci et al.,
2007]. In a typical scenario ∆tBI ≫ ∆tDG ; therefore, if one uses a single time
step size for the hybrid DGTDBI scheme and sets it to ∆tDG , the computa-
tional requirements of TDBI will increase significantly. Consequently, spatially
discretized retarded-time surface integrals ipi,m are sampled at k∆tBI requiring
the computation of fip′ ,l′ (k∆tBI − R/c0 ). But during marching their samples at
k∆tDG , i.e., fip′ ,l′ (l∆tDG ), are available. Therefore, an interpolation scheme is
needed to compute fip′ ,l′ (k∆tBI − R/c0 ) from fip′ ,l′ (l∆tDG ). From the opposite
side, only samples ipi,m (k∆tBI ) are available during time marching. But time-
23
integration scheme requires samples tpp pp̃
i (l∆tDG ) and ti (l∆tDG ). Therefore,
an interpolation scheme is needed to compute them from ipi,m (k∆tBI ). Both
interpolation operations are carried out using shifted higher-order Lagrange
polynomials [Yilmaz et al., 2004], [Bagci et al., 2007]. Also, time derivative
in operator Li′ ,m′ is moved onto this interpolation function. We should also
emphasize here that, to maintain the explicitness of the time marching, which
is one of the main advantages of DGTD scheme over classical finite element
methods, the minimum distance between any two points on ∂Γ and ∂Ω has to
be larger than c0 ∆tBI . A simple illustration of the linear temporal interpolation
process is shown in Figure 1.2.
1.1.5. Numerical Examples for DGTDBI
( −0.1, 0.173, 0 )
!!
( 0.2, 0, 0 )
(0, 0, 0) !!
0.1 0.015
0.015
!!
( −0.1, − 0.173, 0 )
Figure 1.3: Cluster of three PEC spheres. All dimensions are in meters. Repro-
duced with permission from [Li et al., 2014b].
In this section, we provide two numerical examples to demonstrate the
accuracy and the applicability of DGTDBI. In the first example, the scatterer
is a cluster of three PEC spheres residing in free space (Figure 1.3). Use of
the TDBI approach allows the computation domain to be defined in terms
24
Figure 1.4: Transient scattered electric field computed at the origin of cluster
of PEC spheres. Reproduced with permission from [Li et al., 2014b].
of three equally sized domains, each of which encloses one of the spheres.
Disconnected spherical boundaries of these domains and the Huygens surfaces
are located 0.03 m and 0.015 m away from the sphere surfaces (Figure 1.3). For
the first simulation, the cluster is excited with a plane wave with parameters
fm = 500 MHz, t0 = 13.66 ns, and τm = 2.6 ns. During the simulation, the
transient scattered electric near field computed at the origin is recorded. Figure
1.4 compares this field to that computed using a time-domain integral equation
(TDIE) solver [Yilmaz et al., 2004], [Bagci et al., 2007]. Results agree well. For
the second simulation, τm = 0.65 ns while other excitation parameters are kept
the same. After the time-domain simulation is completed, radar-cross-section
(RCS) of the cluster is computed from the Fourier-transformed currents on
the Huygens surface. Figure 1.5 and 1.6 plot the RCS on the xy- and xz-
planes computed at 2.53 MHz and 1.003 GHz, respectively. Results agree very
well with those obtained from the solution of an in-house method of moments
(MoM) solver. These two figures clearly demonstrate that the “absorption”
enforced by the TDBI approach is accurate even at low frequencies.
25
(a) (b)
Figure 1.5: RCS on (a) xy- and (b) xz- planes computed at 2.53 MHz from
the solutions of the DGTDBI method and MoM. Reproduced with permission
from [Li et al., 2014b].
The second example is a U-shape scatterer residing in free space (Figure
1.7). The boundary of the computation domain and the Huygens surface con-
form to the shape of the scatterer and are located 0.02 m and 0.01 m away
from its surface. The scatterer is excited with a plane wave with parameters
fm = 500 MHz, t0 = 13.66 ns, and τm = 2.6 ns. Two simulations are carried
out: (i) The scatterer surface is PEC. (ii) The scatterer is a dielectric body
with relative permittivity of 4.0. After the time-domain simulations, bistatic
and monostatic RCS of the scatterers are computed. Figure 1.8 (a) and (b)
and Figure 1.9 (a) and (b) plot the bistatic and monostatic RCS of the PEC
and dielectric scatterers computed at a wide range of frequencies, respectively.
Plots clearly demonstrate that the results agree very well with those obtained
from an MoM solver, which further verifies the accuracy and also low-frequency
absorbing capability of the DGTDBI method. Additionally, Figure 1.10 (a) and
(b) compare the RCS of the PEC scatterer on the xz- and yz- planes computed
at 1 GHz to those obtained by the same DGTD method, which uses ABC in-
stead of TDBI. It is clearly shown that ABC is not accurate for this structure.
26
(a) (b)
Figure 1.6: RCS on (a) xy- and (b) xz- planes computed at 1.003 GHz from
the solutions of the DGTDBI method and MoM. Reproduced with permission
from [Li et al., 2014b].
1.1.6. DGTD scheme with nodal basis functions

For “traditional” continuous FEM, vector basis functions often have to be used
instead of nodal basis functions to avoid spurious solutions [Jin, 2002]. On the
other hand, for DGTD, it is found that spurious modes dissipate much faster
than physical modes when upwind or penalized fluxes are used, and therefore
they can be avoided even without using vector basis functions [Hesthaven and
Warburton, 2008]. In addition, for DGTD, even curl-conforming vector basis
functions cannot ensure that the solutions are globally divergence-free. From
this viewpoint, the main advantages of using vector basis functions in FEM do
not hold true for DGTD.
From an implementation perspective, generation high-order nodal basis
functions is much easier and often visualization of simulation results directly
obtained at nodes does not require difficult post-processing operations. Fur-
27
0.05
0.2
0.01
0.01
0.05
0.1
0.3
Figure 1.7: U-shaped scatterer. All dimensions are in meters. Reproduced with
permission from [Li et al., 2014b].
thermore, in many multiphysics problems, all field components are used to
realize the coupling between different sets of equations (representing different
physical phenomena). In nodal basis function approach, all field components
are expanded separately and the continuity between neighboring elements for
each component is recovered through the numerical flux.
All these reasons point out to the fact that nodal basis functions are a
good option for DGTD implementations. In the nodal DG formulation, each
Cartesian component of E and H is expanded using a set of nodal basis func-
tions. A popular choice is the Lagrange polynomials Ψl with nodes located at
Legendre-Gauss-Lobatto (LGL) quadrature points [Hesthaven and Warburton,
2008]
Np Np
X X
Eiν (r, t) ≃ Eiν (rl , t)Ψl (r) = eνi,l (t)Ψl (r) (1.47)
l=1 l=1
Np Np
X X
Hiν (r, t) ≃ Hiν (rl , t)Ψl (r) = hνi,l (t)Ψl (r) (1.48)
l=1 l=1
where i is the element index, l = 1, . . . , Np , Np = (p + 1)(p + 2)(p + 3)/6 is
28
(a) (b)
Figure 1.8: The bistatic (a) and monostatic (b) RCS of the PEC U-shaped
scatterer computed at a range of frequencies changing from 286.4 KHz to 1.0
GHz. Reproduced with permission from [Li et al., 2014b].
the number of interpolation nodes, p is the order of the Lagrange polynomials,
rl denote the locations of the interpolation nodes, ν ∈ {x, y, z} denotes the
components of vectors in the Cartesian coordinate system, and eνi,l (t) and hνi,l (t)
are the time-dependent expanding coefficients to be solved for.
Applying Galerkin testing on element i as is done in the vector DG formu-
lation, one can obtain the component-wise semi-discrete form
∂exi
εi = Dyi hzi − Dzi hyi − Fi fie,x − jxi (1.49)
∂t
∂ey
εi i = Dzi hxi − Dxi hzi − Fi fie,y − jyi (1.50)
∂t
∂ezi
εi = Dxi hyi − Dyi hxi − Fi fie,z − jzi (1.51)
∂t
∂hx
µi i = −Dyi ezi + Dzi eyi + Fi fih,x (1.52)
∂t
∂hyi
µi = −Dzi exi + Dxi ezi + Fi fih,y (1.53)
∂t
∂hz
µi i = −Dxi eyi + Dyi exi + Fi fih,z (1.54)
∂t
where eνi = [eνi,1 , ..., eνi,Np ]T and hνi = [hνi,1 , ..., hνi,Np ]T are unknown vectors,
29
(a) (b)
Figure 1.9: The bistatic (a) and monostatic (b) RCS of the dielectric U-shaped
scatterer computed at a range of frequencies changing from 100 MHz to 1.0
GHz. Reproduced with permission from [Li et al., 2014b].
e,ν
jνi = [ji,1
ν , ..., j ν T ν ν
i,Np ] is the current density vector, ji,l = ji (rl , t), and fi and
fih,ν are coefficients corresponding to the ν components of n̂ × [Hi − H∗i ] and
n̂ × [Ei − E∗i ], respectively. Here, the numerical flux H∗i and E∗i are defined the
same as in equation (1.19) and (1.20).
The operators Dνi = M−1 ν −1

i Si and Fi = Mi Li , where Mi is the mass
matrix defined as
Z
Mi (l, k) = Ψl (r)Ψk (r)dr (1.55)
Ωi
Sνi is the stiffness matrix defined as
Z
Sνi (l, k) = Ψl (r)∂ν Ψk (r)dr (1.56)
Ωi
Li is the surface mass matrix defined as
I
Li (l, k) = Ψl (r)Ψk (r)dr. (1.57)
∂Ωi
30
(a) (b)
Figure 1.10: The calculated RCS in xoz (a) and yoz (b) planes with DGTDBI
method and its comparisons from MoM and DGTD-ABC at 1.00013 GHz.
Reproduced with permission from [Li et al., 2014b].
The semi-discrete system in 1.49-1.54 can be integrated in time using ex-
plicit schemes just like it is done for the semi-discrete system in (1.30)-(1.31),
which is obtained using vector basis functions.
1.2. Application of DGTD to real prob-

lems
1.2.1. Graphene-based Devices

The two-dimensional (2D) materials have received significant attention in the
last decade due to their remarkable electrical, thermal, and mechanical prop-
erties. Graphene, which belongs to the family of 2D materials, is an atomically
thick layer of carbon atoms that are arranged in a honeycomb lattice. In the
last few years, graphene has found various applications in different academic
research fields as well as industry [Sharma et al., 2014, Geim and Novoselov,
31
2007, Schwierz, 2010, Tamagnone et al., 2012, Koppens et al., 2011, Lovat,
2012, Chen et al., 2013]. Simulation of graphene-based electromagnetic de-
vices is one of these research fields. Indeed, several methods, which rely on
method of moments, [Hanson, 2008, Shapoval et al., 2013], the finite differ-
ence time-domain, [Lin et al., 2012, Nayyeri et al., 2013, Feizi et al., 2018],
DGTD [Li et al., 2015, Li and Jiang, 2015, Li et al., 2018], and circuit repre-
sentations [Lovat, 2012], have been developed for numerical characterization of
transmission, reflection, and absorption of electromagnetic fields on graphene
sheets as well as generation of graphene surface plasmon polaritons within
the gigahertz-terahertz frequency band. All these methods model the atomi-
cally thick graphene using an equivalent surface impedance boundary condition
(SIBC) to avoid very fine volumetric discretization of the graphene layer, which
would dramatically increase the number of unknowns and decrease the time-
step size (through CFL condition) in the case of explicit time-domain methods
[Li et al., 2015, Li and Jiang, 2015, Li et al., 2018].
Surface conductivity of SIBC is either a scalar [Lovat, 2012] under electro-
static biasing or a tensor [Hanson, 2008, Lovat, 2012] due to quantum hall effect
under magnetostatic biasing. In either case, surface conductivity is frequency
dependent, which makes its efficient modeling within time-domain methods
challenging. In this section, we introduce two different DGTD-based meth-
ods, which use finite integration and auxiliary differential equation (ADE)
methods, respectively, to take into account the temporal dispersion effects and
anisotropic surface conductivity in time domain [Li et al., 2015, Li and Jiang,
2015].
32
1.2.1.1. A Resistive Boundary Condition to Represent
Graphene within DGTD

SIBC associated with graphene is incorporated within DGTD via the re-
formulation of the numerical flux. First the surface conductivity of graphene
is approximated by a series of rational basis functions generated using fast-
relaxation vector-fitting (FRVF) method in the Laplace-domain [Gustavsen
and Semlyen, 1999], [Gustavsen, 2006]. Then, using inverse Laplace transform,
corresponding time-domain matrix equations in integral form (in time t) are
obtained. Finally, these matrix equations are solved by time-domain finite in-
tegration technique (FIT). For elements not touching the graphene sheet, the
well-known Runge-Kutta method is employed to integrate the Maxwell equa-
tions as described in Section 1.1.1. We emphasize here again that the use of
SIBC significantly reduces the computational requirements of DGTD since the
volumetric mesh that would be defined inside the thin graphene sheet is now
avoided. In the rest of the section, we first present the formulation underly-
ing the DGTD scheme extended to account for graphene’s SIBC model and
provide various numerical examples to show the extended scheme’s accuracy,
efficiency, and applicability.
In the formulation described here, the 2D graphene layer is considered as
an infinitely thin surface. In the absence of an external magnetostatic biasing
field, the surface conductivity of this graphene layer is a scalar denoted by
σg (ω, µc , Γ, T ). Using the Kubo’s formula, σg is expressed as [Hanson, 2008]:
jq 2 R ∞ h ∂fd (ε) ∂fd (−ε) i

σg (ω, µc , Γ, T ) = πℏ(ω−j2Γ) 0 ε ∂ε − ∂ε
2
(1.58)
jq (ω−j2Γ) ∞
R fd (−ε)−fd (ε)
− πℏ2 0 (ω−j2Γ)2 −4(ε/ℏ)2
where ε indicates the energy state, ℏ denotes the reduced Planck constant, −q
33
−1
is the charge of electron, and fd (ε) = e(ε−µc )/kB T + 1

is the Fremi-Dirac
distribution with Boltzmann constant kB . In (1.58), the first term corresponds
to the intraband contribution that can be expressed as [Hanson, 2008]
jq 2 kB T

µc
−µc /kB T
σintra =− 2 + 2ln e +1 (1.59)
πℏ (ω − j2Γ) kB T
while the second term corresponds to the interband contribution, which can
be approximated by
jq 2

2 |µc | − (ω − j2Γ)ℏ
σinter =− ln (1.60)
4πℏ 2 |µc | + (ω − j2Γ)ℏ
when kB T ≪ ℏω and |µc |. On the resistive thin sheet with the finite conduc-
tivity σg , the electromagnetic fields satisfy the following boundary conditions
[Li et al., 2015]:
n̂ × (E2 − E1 ) = 0
(1.61)
2 1
n̂ × (H − H ) = σg Et
where the superscripts 1 and 2 represent the two sides of the graphene sheet,
n̂ is a unit normal vector pointing from side 1 to side 2, and Et = −n̂ ×
n̂ × (E1 + E2 )/2 is the tangential component of the electric field along the

graphene sheet. Note that (1.61) is expressed in frequency domain (or Laplace
domain).
Similar to the fundamental boundary conditions, as discussed in Section
1.1.2, the resistive boundary conditions in (1.61) can be incorporated within
DGTD by reformulating the jump conditions used in the expression of the
numerical flux (see (1.13)-(1.20)) in Section 1.1.1 for the jump conditions).
More specifically, the jump condition for magnetic field for characteristics xn =
34
0 is updated as
n̂i,fg × (H∗∗ ∗
fg − H fg ) = σ g E t (1.62)
for element facets fg that touch the graphene sheet. Here, i represents the
index of these elements. Note that only one facet can touch the graphene sheet
for a given element. Using (1.62) together with the other jump conditions
(1.13)-(1.20) from Section 1.1.1 (or more specifically their frequency-domain
counterparts), the upwind numerical fluxes for facets that touch the graphene
sheet are obtained as

"
Yi Ei + Yj,fg Ej,fg + n̂i,fg × Hj,fg − Hi
n̂i,fg × E∗fg = n̂i,fg ×
Yi + Yj,fg
#
σg Ei + Ej,fg
− (1.63)
2 Yi + Yj,fg
"
Zi Hi + Zj,fg Hj,fg + n̂i,fg × Ei − Ej,fg
n̂i,fg × H∗fg = n̂i,fg ×
Zi + Zj,fg
#
Zj,fg σg n̂i,fg × Ei + Ej,fg
+ . (1.64)
2 Zi + Zj,fg
The numerical flux in (1.63) and (1.64) cannot be directly converted to time
domain since σg is frequency dependent (with a rather complicated function).
To overcome this issue, we first represent σg in terms of a set of rational
basis functions in the Laplace domain (using FRVF technique [Gustavsen and
Semlyen, 1999, Gustavsen, 2006]). Then, we use (1.58) in Laplace-domain DG
equations. Then, via inverse Laplace transform of the resulting equations, a
time marching scheme is obtained for DGTD.
The FRVF technique is applied to a set of frequency samples of σg to
yield the following representation in terms of rational functions [Gustavsen
35
and Semlyen, 1999, Gustavsen, 2006]:
P
up s−p
P
u1 s−1 + u2 s−2 + · · · + uP s−P p=1
σg (s) = −1 −2 −Q
= Q . (1.65)
d0 + d1 s + d2 s + · · · + dQ s
s−q
P
dq
q=0
Here, s is the Laplace state variable, up and dq are the pth and the qth coeffi-
cients, and P and Q are the orders of numerator and denominator, respectively.
The tested DG equations for Maxwell curl equations in Laplace domain are
given as
Z 4
X Z
Ψei,k · εi Ei −s−1 ∇ × Hi dr = s−1 Ψei,k · n̂i,f ×(H∗f − Hi ) dr (1.66)

Ωi f =1 ∂Ωi,f
Z 4
X Z
Ψhi,k · µi Hi +s−1 ∇ × Ei dr = s−1 Ψhi,k · n̂i,f ×(Ei − E∗f ) dr. (1.67)

Ωi f =1 ∂Ωi,f
Note that on the right-hand side of (1.66) and (1.67), when f = fg , the nu-
merical flux expression in (1.63) and (1.64) is used. For other facets, numerical
flux stays same as those in (1.19)-(1.20).
Inserting (1.63) - (1.65) in (1.66) and (1.67), and using the Fourier trans-
form pair
Z Z
−n
s F (s) ↔ ··· | ·{z
f (τ ) dτ · · dτ} (1.68)
| {z } n
n
in the resulting equation yields the following time-domain matrix equations [Li
36
et al., 2015]:
Mei (d0 ei,0 +d1 ei,1 +· · ·+dQ ei,Q )−Sei (d0 hi,1 +d1 hi,2 + · · · + dQ hi,Q+1 )
4
ee,σg σg ee,σ σ
X
= Fee ee eh eh
ii,f ei,f +Fij,f ej,f +Fii,f hi,f +Fij,f hi,f +Fij ej +Fii g ei g (1.69)
f =1 | {z }
Fe,σg
Mhi (d0 hi,0 +d1 hi,1 +· · ·+dQ hi,Q )+Shi (d0 ei,1 +d1 ei,2 +· · · +dQ ei,Q+1 )
4
he,σg σg he,σ σ
X
= Fhh h
ii,f i,f +F hh
h
ij,f i,f +F he
e
ii,f i,f +F he
ij,f i,f +Fij
e ej +Fii g ei g (1.70)
f =1 | {z }
Fh,σg
e/h e/h
where Mi and Si are mass and stiffness matrices and Fee , Feh , Fhh , and
Fhe are the flux matrices. Matrix entries for those matrices can be found in
ee,σg ee,σg he,σg he,σg
Section 1.1.1. The entries of the matrices Fii , Fij Fii , and Fij are
given by:
Zj,fg
Z
ee,σ
[Fii g ]kl = Ψe · n̂i,fg × (n̂i,fg × Ψei,l ) dr
2(Zi + Zj,fg ) ∂Ωi,j i,k
Zj,fg
Z
ee,σg
[Fij ]kl = Ψe · n̂i,fg × (n̂i,fg × Ψej,l ) dr
2(Zi + Zj,fg ) ∂Ωi,j i,k
Z
he,σg 1
[Fii ]kl = Ψh · (n̂i,fg × Ψei,l ) dr
2(Yi + Yj,fg ) ∂Ωi,j i,k
Z
he,σg 1
[Fij ]kl = Ψh · (n̂i,fg × Ψej,l ) dr.
2(Yi + Yj,fg ) ∂Ωi,j i,k
σ ee,σg ee,σg σ he,σg he,σg

In (1.69), (1.70), flux matrices Fe g = Fij + Fii , Fhg = Fij + Fii
are non-zero only for faces (f = fg ) touching the graphene sheet, otherwise
(f ̸= fg ) they are zero. Moreover, terms ei,q and hi,q are column vectors storing
37
the unknown coefficients eki and hli and are defined as
Z Z
eki,q = ··· eki (τ ) dτ
| ·{z
· · dτ}
| {z } q
q
Z Z (1.71)
hli,q = ··· hli (τ ) dτ
| ·{z
· · dτ}
| {z } q
q
The expressions for other column vectors are given as
Q
X Q
X
(e/h)i,f = dq (ei,f /hi,f )q+1 , (e/h)j,f = dq (ej,f /hj,f )q+1
q=0 q=0
(1.72)
P P
σ σ
X X
ei g = up ei,p+1 , ej g = up ej,fg ,p+1
p=1 p=1
Note that all terms in (1.72) involve integrals in time. To enable solution
of (1.69) and (1.70) using a time marching scheme, these integrals in time
are discretized using an FIT scheme. This scheme uses the rectangular rule
[Shapoval et al., 2013], which yields the discretized version of (1.71) with order
p or q at t = (n + 1)δt as:
n
X k3 X
X k2
en+1
i,p/q = (δt)p/q ··· ei,k1 +1
kp/q =0 k2 =0 k1 =0
n k3 X
k2
X X (1.73)
hn+1
i,p/q = (δt)p/q ··· hi,k1 +1
kp/q =0 k2 =0 k1 =0
en+1 n+1
i,0 = ei , hn+1 n+1
i,0 = hi .
To maintain the explicit nature of the time marching scheme, the terms
involving contributions from neighboring elements are discretized using a for-
38
ward rectangular rule. This yields
n
X k3 X
X k2
en+1
j,p/q = (δt)p/q ··· ej,k1
kp/q=0 k2 =0 k1 =0
(1.74)
n
X k3 X
X k2
hn+1
j,p/q = (δt)p/q ··· hj,k1
kp/q=0 k2 =0 k1 =0
After a lengthy mathematical manipulation, (1.69) and (1.70) are expressed
in a more compact form as a matrix system at t = (n + 1)δt
    
M̃ei S̃ei en+1
i F̃ei
= . (1.75)
    
 
S̃hi M̃hi hn+1
i F̃hi
Here, the matrix-blocks are given by
Q Q P
ee,σg
X X X
M̃ei = dq (δt)q Mei − dq (δt)q+1 Fee
ii − up (δt)p+1 Fii
q=0 q=0 p=1
Q
X Q
X
S̃ei = − dq (δt)q+1 Sei + dq (δt)q+1 Feh
ii
q=0 q=0
(1.76)
Q
X Q
X
M̃hi = dq (δt)q Mhi − dq (δt)q+1 Fhh
ii
q=0 q=0
Q Q P
he,σg
X X X
S̃hi = dq (δt)q Shi − dq (δt)q+1 Fhe
ii − up (δt)p+1 Fii
q=0 q=0 p=1
39
and the column vectors are given by
   
XQ  XQ 
F̃ei = − dq (δt)q ẽn+1 e
i,q  Mi +  dq (δt)q+1 h̃n+1 e
i,q+1  Si

q=0 q=0
   
X Q  X Q 
+ dq (δt)q+1 ẽn+1
i,q+1  F ee
ii + d q (δt) q+1 e n+1 ee
j,q+1 Fij
 
q=0 q=0
   
X Q  X Q 
n+1 eh n+1
+ dq (δt)q+1 h̃i,q+1 Fii + dq (δt)q+1 hj,q+1 Feh
    ij
q=0 q=0
   
X P  X P 
ee,σg ee,σ
+ up (δt)p+1 ẽn+1 F
i,p+1  ii + u p (δt) en+1
p+1 j,p+1 F g
   ij
p=1 p=1
   
XQ  XQ 
F̃hi = − dq (δt)q h̃n+1 h
i,q  Mi −  dq (δt)q+1 ẽn+1
i,q+1  Si
h

q=0 q=0
   
X Q  X Q 
+ dq (δt)q+1 h̃n+1 F
i,q+1  ii
hh
+ d q (δt) h n+1
q+1 j,q+1 Fhh
   ij
q=0 q=0
   
X Q  X Q 
+ dq (δt)q+1 ẽn+1 F
i,q+1  ii
he
+ dq (δt) q+1 e n+1
j,q+1  Fij
he
 
q=0 q=0
   
X P  X P 
he,σg he,σ
+ up(δt)p+1 ẽn+1
i,p+1 F ii + u p (δt) p+1 en+1
j,p+1 Fij g
   
p=1 p=1
where
ẽn+1 n+1 n+1

i,p/q = ei,p/q − (δt)p/q ei
h̃n+1 n+1 n+1

i,p/q = hi,p/q − (δt)p/q hi
ẽn+1 n+1
i,0 = 0 , h̃i,0 = 0 .
To efficiently evaluate (e/h)n+1 n+1 n+1

i,p/q , (e/h)j,p/q , and (ẽ/h̃)i,p/q , the following re-
40
cursive schemes can be used [Shapoval et al., 2013]:
(e/h)n+1 n n+1
i/j,p/q = (e/h)i/j,p/q + δt · (e/h)i/j,p−1/q−1
(1.77)
(ẽ/h̃)n+1
i,p/q = (ẽ/h̃)ni,p/q + δt · (ẽ/h̃)n+1
i,p−1/q−1 .
In the remainder of this section, two numerical examples are provided to
Figure 1.11: Fitted magnitude and phase of the surface conductivity σg from
500 MHz to 10 THz with four poles. Reproduced with permission from [Li
et al., 2015].
demonstrate the accuracy and the applicability of the DGTD scheme described
above. In both examples, the excitation is a plane wave with electric field
Einc (r, t) = ŷg(t − k̂ · r/c0 ), and g(t) = exp(−(t − t0 )2 /τm

2 ) cos(2πf (t − t ))
m 0
is a Gaussian pulse with modulation frequency fm , duration τm and delay t0 .
In the first example, the scatterer is an infinitely large graphene sheet that
resides in free space on the xy- plane. The temperature used in the expression of
41
Figure 1.12: The calculated magnitudes of coefficients ΓT , ΓR and ΓA by the
proposed algorithm as well as the theoretical data for µc = 0.3eV. Reproduced
with permission from [Li et al., 2015].
σg is set to T = 300K while µc and Γ are changed for two different simulations.
Similarly, k̂ = ẑ for both of the simulations while τm , fm , and t0 are changed
for the two simulations.
For the first simulation, µc = 0.3eV, the scattering rate Γ = 0.41meV/ℏ.
The FRVF scheme with four poles is applied to the samples of σg in a fre-
quency range between microwave and terahertz. In Figure 1.11, the compar-
ison between the fitted values and the original data is shown for frequencies
between 500 MHz to 10 THz. Excellent agreement is observed, which shows
that FRVF can produce an accurate representation of σg in terms of rational
functions in a wide range of frequencies. The reflection, transmission, and ab-
sorption coefficients (denoted by ΓR , ΓT , and ΓA , respectively) are computed
using the DGTD scheme. The results are shown in Figure 1.12. For compar-
ison, the analytical solutions calculated by the method described in [Hanson,
2008] are also presented in the same figure. Good agreement between two sets
of results is observed.
For the second simulation, the transmission of a plane wave through the
graphene sheet with µc = 0.12eV, Γ = 2 × 1012 Hz is studied. The FRVF with
42
Figure 1.13: Magnitude of calculated transmission coefficient ΓT by the pro-
posed algorithm as well as theoretical data from terra hertz to near-infrared
region. Reproduced with permission from [Li et al., 2015].
seven poles is used to represent σg in the frequency range between 500 GHz
and 100 THz. Note that we use seven poles here since σg has a jump around
60 THz. The transmission coefficient computed using the DGTD scheme and
the analytical expression are compared in Figure 1.13. Again, very good agree-
ment is observed. We should mention here that, for these two simulations, the
computation domain is truncated using the Silver-Muller absorbing boundary
condition [Mur, 1981] (see (1.32) in Section 1.1.2.1), which yields very accurate
results since the wave is normally incident on the truncation surface.
In the second example, the DGTD scheme is validated for a three-dimensional
(3-D) problem. Two simulations are carried out. In the first simulation, the
scatterer is a 5 × 10 µm2 graphene patch that resides in free space on the xy-
plane [Llatser et al., 2012]. Parameters of σg are T = 300K, µc = 0eV, and
Γ= 1
2τ with τ = 10−13 s and the parameters of excitation are fm = 2.5THz,
k̂ = ẑ, τm = 1.274 × 10−13 s, and t0 = 3τm . The FRVF scheme with three
poles and residues is used to represent σg in the frequency range between
43
Figure 1.14: Normalized ECS versus frequency for a freestanding graphene
patch in [Llatser et al., 2012] as its counterpart calculated by integral equation
method. Reproduced with permission from [Li et al., 2015].
500 MHz and 5 THz. For comparison, the normalized extinction cross section
between 0.1 and 4 THz are computed using the DGTD scheme as shown in
Figure 1.14. The reference result in the same figure is obtained by using an
integral equation solver [Llatser et al., 2012]. The good agreement between the
results demonstrates the accuracy and applicability of the DGTD scheme for
3D examples.
In the second simulation, a 50 × 50 µm2 graphene patch that resides in free
space on the xy-plane is characterized. Parameters of σg are T = 300 K , Γ =
2.5meV/ℏ, fm = 2.5THz, τm = 6.37 × 10−14 s. Scattering and absorption cross
sections of the graphene patch are computed for different values of µc . In Figure
1.15, the normalized cross sections between 1 and 10 THz are shown for µc =
0.5, 1.0 and 1.5eV. As expected, plasmon resonances are observed at various
frequencies. The increase in µc results in the up-shift of resonant frequencies. In
addition, the radar cross section between 1 and 10 THz is presented in Figure
1.16. It is found that the peaks in the radar cross section happen at the plasmon
44
Figure 1.15: TSCS and ACS of the 50 × 50 µm2 freestanding graphene patch
corresponding to different chemical potentials µc = 0.5, 1.0, and 1.5 eV. Re-
produced with permission from [Li et al., 2015].
resonant frequencies, which results from the near-field enhancement. In Figure
1.17, the normalized near-field patterns for the y-component of the electric
field obtained in the simulation with µc = 1.5eV at three plasmon frequencies
f1 = 1.76THz, f2 = 4.98THz and f3 = 6.97THz are shown. Noticeable near-
field enhancement is observed at the resonant frequencies. In Figure 1.18, the
normalized far-field scattered patterns in E- and H-planes at the fundamental
plasmon frequency f1 = 1.76THz are also presented. It is interestingly found
that the far-field patterns resemble those of conventional short dipoles, which
45
Figure 1.16: Calculated forward RCS of the 50 × 50 µm2 freestanding graphene
patch corresponding to different chemical potentials µc = 0.5, 1.0, and 1.5eV.
Reproduced with permission from [Li et al., 2015].
is in line with the assertion in [Tamagnone et al., 2012]. The nonsymmetrical
pattern in H-plane is attributed to the boundary condition in (1.61). We should
mention here that, for these two simulations for 3D problems, the computation
domain is truncated using TDBI as explained in Section 1.1.3 to maintain the
accuracy of the solution.
1.2.1.2. A Resistive Boundary Condition and an Auxil-
iary Equation Method to Represent Magnetized
Graphene within DGTD

The section 1.2.1.1 introduces a scheme to analyze the graphene with a scalar
surface conductivity, which is observed in the absence of an external magne-
tostatic biasing. However, when there is a biasing magnetic field, the surface
conductivity becomes a tensor due to quantum Hall effect [Hanson, 2008],
[Lovat, 2012]. Direct incorporation of this anisotropic resistive boundary con-
dition (RBC) (representing the graphene sheet) within DGTD would result a
46
Figure 1.17: Normalized near-field distribution of Ey over the graphene sheet
at (a) f1 = 1.76THz ; (b) f2 = 4.98THz; and (c) f3 = 6.97THz for µc = 1.5eV
case. Reproduced with permission from [Li et al., 2015].
numerical flux in a tensor form. In addition, still, the dispersion of the surface
conductivity has to be accounted for. To address these problems, in this section,
we introduce another DGTD scheme that can account for both the anisotropy
and the dispersion of the surface conductivity. In this scheme, a surface po-
larization current is introduced over the graphene sheet [Li and Jiang, 2015],
[Wang et al., 2013]. This current satisfies an auxiliary differential equation
(ADE) that involves the electric field. With the ADE and the polarization
current, the anisotropic and the dispersive RBC is converted an isotropic and
nondispersive one. Therefore, the time-domain convolution is avoided and the
corresponding numerical flux in the presence of the new RBC is made isotropic.
Suppose a graphene sheet is placed at z = 0 plane and biased by a static
47
Figure 1.18: Normalized far-field pattern in E- and H-plane at f1 = 1.76THz
for µc = 1.5eV. Reproduced with permission from [Li et al., 2015].
magnetic field perpendicular to the graphene plane, i.e., B0 = ẑB0 . Because
of the Lorentz force, the surface conductivity exhibits anisotropy [Sounas and
Caloz, 2012]. To have a fundamental insight into this phenomenon, the motion
of an electron in the presence of an electric field E is studied. First, an x-
polarized electric field E = x̂Ex is considered. In this case, the electron is
accelerated along the x- direction due to the electric force Fe = −eE with −e
denoting the electron charge. Simultaneously, the moving electron with velocity
v will be exposed to a Lorentz force Fm = −ev × B0 along the y- direction.
As a result, two current components along the x- and y- directions will be
generated [Sounas and Caloz, 2012]:
J = σxx Ex x̂ + σyx Ex ŷ (1.78)
where σxx and σyx are conductivities parallel and perpendicular to the electric
field E, respectively. Similarly, for the case with an electric field E = ŷEy , two
48
current components will be generated:
J = −σxy Ey x̂ + σyy Ey ŷ. (1.79)
For graphene with same properties in all directions, the conductivities must
satisfy σxx = σyy and σyx = σxy . Based on this, the generated currents in (1.78)
and (1.79) can be combined together and re-written in a more compact form
as matrix equation [Sounas and Caloz, 2012]:
J = σg · E (1.80)
with E = [Ex , Ey , 0]T and
 
 σxx −σyx 0 
 
σg = 
 σ yx σxx 0 .
 (1.81)
 
0 0 0
The analytical expressions for the surface conductivities σxx and σyx con-
sisting intraband and interband contributions can be obtained based on the
Kubo’s formula [Kubo, 1957] (see (1.59) and (1.60) in Section 1.2.1.1 also).
For frequencies in the terahertz range, the intraband term is dominant over
the interband term [Sounas and Caloz, 2012]. Thus, only the intraband term is
considered here. Based on the fact that the probability of electron transitions
between Landau levels around the Fermi level µc is the strongest over others,
the expressions for σxx and σyx can be approximated by a Drude-like model
49
[Wang et al., 2013, Sounas and Caloz, 2012]:
1 + jωτ
σxx (ω, µc , τ, T, B0 ) = σ0
(ωc τ ) + (1 + jωτ )2
2
(1.82)
ωc τ
σyx (ω, µc , τ, T, B0 ) = σ0
(ωc τ ) + (1 + jωτ )2
2
with
e2 τ kB T

µc
σ0 = + 2ln(e−µc /kB T + 1) (1.83)
πℏ2 kB T
where we choose T = 300 K in this chapter, τ is the scattering time, and
ωc ≈ eB0 vF2 /µc is the cyclotron frequency with vF ≈ 106 m/s denoting the
Fermi velocity.
As done in the section 1.2.1.1, the graphene layer is assumed to be an
infinitesimally thin conductive sheet. Therefore, the following boundary con-
ditions over the graphene sheet have to be satisfied
n̂ × (E2 − E1 ) = 0
(1.84)
n̂ × (H2 − H1 ) = σ g · E
Note that (1.81) is expressed in frequency domain and the only difference
with respect to (1.58) in Section 1.2.1.1 is that the surface conductivity is a
tensor.
Since the incorporation of boundary conditions within DGTD is enforced
by redefining the numerical flux, the anisotropic boundary condition in (1.84)
would result in a tensor form flux expression and simultaneously dispersive,
which no doubt complicates the problem significantly (the convolution in the
time-domain is required and the different components of electromagnetic fields
50
are entangled with each other). To address these challenges, an auxiliary polar-
ization surface current J(ω) is introduced by rewriting (1.84) as n̂×(H2 −H1 ) =
J(ω) with J governed by the ADE
Jx (ω) = σxx Ex − σyx Ey

(1.85)
Jy (ω) = σyx Ex + σxx Ey
After some mathematical manipulations, (1.85) can be rewritten as
jωJx (ω) = −2ΓJx (ω) − ωc Jy (ω) + 2Γσ0 Ex (ω)

(1.86)
jωJy (ω) = −2ΓJy (ω) + ωc Jx (ω) + 2Γσ0 Ey (ω)
where Γ = 1/2τ denotes the phenomenological scattering rate. Through the
inverse Fourier transform, the time-domain counterparts of (1.86) are obtained
as
∂J
= C · J + 2Γσ0 E (1.87)
∂t
with
 
 −2Γ −ωc 
C= . (1.88)
ωc −2Γ
With this auxiliary polarization surface current J and the ADE, the anisotropic
and dispersive RBC in (1.81) is transformed to an isotropic and nondispersive
one in (1.88). Then, the incorporation of this scalar RBC within DGTD can be
facilitated in the conventional way by reformulating the numerical flux, which
will not be repeated again in this section.
For the DGTD-formulation, the electric field and magnetic field are ap-
51
proximated in the same way as described in the section 1.2.1.1. Other than
that, the polarization current J is approximated by 2D vector basis functions
φi (r)
K
X
Ji = ci,l (t)φi,l (r) (1.89)
l=1
By applying the DG testing to the Maxwell curl equations and the ADE, we
obtain
Z 4 Z
∂Ei X
Ψei,k Ψei,k · n̂i,f × (H∗f − Hi ) dr

· εi − ∇ × Hi dr =
Ωi ∂t ∂Ωi,f
f =1
Z 4 Z
∂Hi X
Ψhi,k Ψhi,k · n̂i,f × (Ei − E∗f ) dr (1.90)

· µi + ∇ × Ei dr =
Ωi ∂t
f =1 ∂Ωi,f
Z Z Z
∂Ji
φi,k · dr = φi,k · C · Ji dr + 2Γσ0 φi,k · Ei dr.
∂Ωi,fg ∂t ∂Ωi,fg ∂Ωi,fg
The semi-discrete matrix equations for the ADE can then be obtained as
∂ci
J̄i = M̄c ci + 2Γσ0 M̄v ei (1.91)
∂t
where
Z
[Mc ]kl = φi,k · C · φi,l dr
∂Ωi,fg
Z
[Mv ]kl = φi,k · Ψei,l dr (1.92)
∂Ωi,fg
Z
i
J̄ kl = φi,k · φi,l dr.
∂Ωi,fg
We only give the matrix related to the ADE since other matrices are same as
those introduced in the section 1.2.1.1 [Li and Jiang, 2015]. In the remainder of
52
this section, two numerical examples are provided to demonstrate the accuracy
and the applicability of the DGTD scheme described above.
(a) (b)
(c)
Figure 1.19: (a) Total transmission coefficients, (b) Faraday rotation angle,
and (c) cross-polarized transmission coefficients versus frequency for different
magnetostatic biasing. Reproduced with permission from [Li and Jiang, 2015].
As the first example, we use the DGTD scheme to study the Faraday ro-
tation and surface plasmon polarization by varying the chemical potential and
the magnetostatic biasing. An infinitely large graphene sheet placed on the xy-
plane is biased by a z-directed magnetostatic field B0 = ẑB0 . The graphene
sheet is excited a plane wave with electric field Einc (r, t) = x̂g(t−k̂·r/c0 ), where
k̂ = ẑ is the propagation direction, g(t) = exp(−(t − t0 )2 /τm

2 ) cos(2πf (t−t ))
m 0
is a Gaussian pulse, where the pulse’s modulation frequency, duration, and de-
53
lay are set to fm = 5THz, τm = 6.37 × 10−13 s, and t0 = 5tm , respectively.
The first-order SM-ABC [Dosopoulos et al., 2013] (see also ((1.32)) in Section
1.1.2.1) is used to truncate the computation domain along the z- direction,
which is sufficient for this example since all waves are normally incident to
the truncation boundary. In Fig. 19, the total transmission coefficient, Fara-
day rotation angle and cross-polarized transmission coefficient obtained using
the DGTD scheme are presented. For comparison, the analytical results for
total transmission coefficient and Faraday rotation angle are also provided in
Figure 1.19. Very good agreement is observed between the two sets of results.
The results show that the magnetostatic biasing has significant impacts on the
electromagnetic waves propagation through a graphene sheet.
Figure 1.20: Calculated extinction cross-section (ECS) by the proposed DGTD-

RBC algorithm and the reference result by integral equation (IE) method in
[Llatser et al., 2012]. Reproduced with permission from [Li and Jiang, 2015].
The next example validates the accuracy of the DGTD scheme for 3D
examples. First, a 5 × 10 µm unmagnetized graphene patch in [Llatser et al.,
2012] is revisited. The same excitation is employed as in previous example. The
computation domain is truncated using TDBI as explained in Section 1.1.3 to
54
(a)
(b)
Figure 1.21: (a) Total transmission coefficients, (b) Faraday rotation angle,
and (c) cross-polarized transmission coefficients versus frequency for different
magnetostatic biasing. Reproduced with permission from [Li and Jiang, 2015].
maintain the accuracy of the solution. In Figure 1.20, the extinction cross
section computed using the DGTD scheme and an integral equation solver
[Llatser et al., 2012] are compared. The figure shows the results agree well.
Then, a 20 × 100 µm2 graphene patch under biased by a static magnetic field
B0 = 0.25ẑ is analyzed. To show the impacts of chemical potentials on the
graphene resonance, the scattering and the extinction cross sections obtained
under different µc by letting the scattering rate Γ = 2.5meV/ℏ are presented in
55
Figure 1.22: Normalized ECS for substrates Si, SiO2 , and Si3 N4 . Reproduced
with permission from [Li and Jiang, 2015].
Figure 1.21. It is noted that the resonant frequencies shift upwards with higher
chemical potentials, and the resonance becomes much stronger. Next, to study
the effect of the substrate material on the plasmon resonance, the graphene
patch is located on a substrate with thickness h = 2µm. The materials tested
are silicon (Si), silicon-nitrate (Si3 N4 ), and silicon-dioxide (SiO2 ) (all of which
are commonly used by experimental researchers). The extinction cross section
computed under µc = 0.5eV for these three different materials is shown in
Figure 1.22. It is noted that the first two resonant peaks of the extinction cross
section shift to low frequencies with the increasing permittivity due to the
physical dimension of the graphene becoming larger compared with wavelength

√
λ = λ0 / ε r .
56
1.2.2. Multiphysics simulation of optoelectronic
devices
The several advantages the DG methods comes with (as explained in Section
1.1.6) make them very suitable for multiphysics simulations. Indeed, in the
last decade, they have been applied to many multiphysics problems in various
fields, such as coupled simulations of electromagnetic/thermal interactions,
electromagnetic/plasma interactions, and finally electromagnetic/semiconduc-
tor carrier interactions [Homsi et al., 2017, Don, 2019, Dong et al., 2022]. In
general, multiphysics simulations involve solving different partial differential
equations simultaneously, and those partial differential equations are usually
characterized by different scales in space and time. DG permits various types
of mesh elements for discretization and leads to easy implementations of high-
order spatial basis functions and high-order time integration schemes. Further-
more, it allows non-conformal meshes, adaptive h/p-refinement, and local time
stepping. These properties can be utilized to greatly reduce computational re-
quirements when dealing with multiphysics problems [Jacobs and Hesthaven,
2006, Hesthaven and Warburton, 2008, Harmon et al., 2016, Shu, 2016b, Yan
et al., 2016, Chen and Bagci, 2020c,b].
This section presents the application of DG in multiphysics simulations of
photoconductive devices. The operation of these devices relies on the inter-
actions between electromagnetic fields and semiconductor carriers. Numerical
tool has to be account for full-wave electromagnetic interactions, carrier dy-
namics, and their nonlinear coupling. We note here although the DG frame-
work presented in this section is developed for simulation of photoconductive
devices, it can be readily applied to similar optoelectronic devices (such as
57
phototransistors and photovoltaic devices) with only minor modifications.
The operation of photoconductive devices can be analyzed into two stages.
Initially, a bias voltage is applied to the electrodes. The resulting static electric
field changes the carrier distribution. The re-distributed carriers, in turn, affect
the electric potential distribution. The device reaches a non-equilibrium steady-
state mathematically described by a coupled system of Poisson and stationary
drift-diffusion equations [Chen and Bagci, 2020c, Vasileska et al., 2010, Chuang,
2012]
∇ · [ε(r)Es (r)] = q[C(r) + nsh (r) − nse (r)] (1.93)
∇ · Jsc (r) = ±qRs (nse , nsh ) (1.94)
Jsc (r) = qµc (Es )Es (r)nsc (r) ± qdc (Es )∇nsc (r) (1.95)
where Es (r) is the static electric field, ε(r) is the dielectric permittivity, q is the
electron charge, C(r) is the doping concentration, the superscript “s” stands
for static, the subscript c ∈ {e, h} represents the carrier type and hereinafter
the upper and lower signs should be selected for electron (c = e) and hole
(c = h), respectively, nsc (r) is the carrier density, R(ne , nh ) is the recombina-
tion rate, and µc (Es ) and dc (Es ) are the mobility and diffusion coefficient, re-
spectively. From the Einstein relation, dc (Es ) = VT µc (Es ), where VT = kB T /q
is the thermal voltage, kB is the Boltzmann constant, and T is the absolute
temperature.
Two most common recombination processes are the trap assisted recom-
bination described by the Shockley-Read-Hall (SRH) model [Chen and Bagci,
58
2020c, Vasileska et al., 2010, Chuang, 2012]
nse (r)nsh (r) − ni 2

RSRH (nse , nsh ) =
τe (nh1 + nsh (r)) + τh (ne1 + nse (r))
and the three-particle band-to-band transition described by the Auger model
RAuger (nse , nsh ) = [nse (r)nsh (r) − ni 2 ][CeA nse (r) + ChA nsh (r)]
Here, ni is the intrinsic carrier concentration, τe and τh are the carrier lifetimes,
ne1 and nh1 are SRH model parameters related to the trap energy level, and CeA
and ChA are the Auger coefficients. Depending on the semiconductor material,
more recombination models can be included [Vasileska et al., 2010]. The net
recombination rate Rs (nse , nsh ) is given by the summation of all recombination
terms.
To accurately model semiconductor devices, it is important to include
proper mobility models with respect to the semiconductor materials and device
operating conditions [Selberherr, 1984, Vasileska et al., 2010, Sil, 2016, Min,
2017, COM, 2017]. For a photoconductive device working under a high bias
voltage, it is important to consider the high-field mobility model that accounts
for the carrier velocity saturation effect [Vasileska et al., 2010, Sil, 2016, Min,
2017, COM, 2017]. Here, the Caughey-Thomas model [Vasileska et al., 2010,
Moreno et al., 2014a] is used
 !βc βc−1
µ0c E∥ (r)
µc (Es ) = µ0c 1 + 
Vcsat
where E∥ (r) is amplitude of the electric field intensity parallel to the current
flow, µ0c is the low-field mobility, and Vcsat and βc are fitting parameters ob-
59
tained from experimental data.
A transient stage starts when an optical electromagnetic excitation is in-
cident on the device. The photoconductive material absorbs the optical elec-
tromagnetic wave energy induced on the device and generates carriers. The
carriers are driven by both the bias electric field and the optical electromag-
netic fields. The carrier dynamics and electromagnetic wave interactions are
mathematically described by a coupled system of the time-dependent Maxwell
and drift-diffusion equations [Chen and Bagci, 2020b]
∂H(r, t)
µ(r) = −∇ × E(r, t) (1.96)
∂t
∂E(r, t)
ε(r) = ∇ × H(r, t) − [Je (r, t) + Jh (r, t)] (1.97)
∂t
∂nc (r, t)
q = ±∇ · Jc (r, t) − q[R(ne , nh ) − G(E, H)] (1.98)
∂t
Jc (r, t) = qµc (E)E(r, t)nc (r, t) ± qdc (E)∇nc (r, t). (1.99)
Here, E(r, t) and H(r, t) are the total electric and magnetic field intensities,
ne (r, t) and nh (r, t) are the total electron and hole densities, Je (r, t) and Jh (r, t)
are the total current densities due to electron and hole movement, ε(r) and
µ(r) are the dielectric permittivity and permeability, R(ne , nh ) and G(E, H)
are the recombination and generation rates, µc (E) and dc (E) are mobility and
diffusion coefficient, respectively.
Since the bias voltage persists during the transient stage, and the boundary
conditions for Poisson and drift-diffusion equations, e.g., the Dirichlet bound-
ary conditions on the electrodes, do not change, it is assumed that Es (r)
and nsc (r) solved from the Poisson-drift-diffusion system are valid through-
out the transient stage. Therefore, the total field intensities and carrier and
current densities can be separated into stationary and transient components as
60
E(r, t) = Es (r) + Et (r, t), H(r, t) = Hs (r) + Ht (r, t), nc (r, t) = nsc (r) + ntc (r, t),
Jc (r, t) = Jsc (r) + Jtc (r, t), respectively. Here, the superscript “t” stands for
transient components. Similarly, the recombination rate R(ne , nh ) is decom-
posed into stationary and transient components as R(ne , nh ) = Rs (nse , nsh ) +
Rt (nte , nth ). The mobility is assumed to be a function of Es (r) only, i.e., µc (E) ≈
µc (Es ), c ∈ {e, h}. With these definitions, it is straightforward to eliminate the
static components from the left hand of (1.96)–(1.99), yielding a reduced cou-
pled system of time-dependent Maxwell and drift-diffusion equations [Chen
and Bagci, 2020b, Chen et al., 2021]
∂Ht (r, t)
µ(r) = −∇ × Et (r, t) (1.100)
∂t
∂Et (r, t)
ε(r) = ∇ × Ht (r, t) − [Jte (r, t) + Jth (r, t)] (1.101)
∂t
∂nt (r, t)
q c = ±∇ · Jtc (r, t) − q[Rt (nte , nth ) − G(Et , Ht )] (1.102)
∂t
Jtc (r, t) = qµc (Es )([Es (r) + Et (r, t)]ntc (r, t) (1.103)
+ Et (r, t)nsc (r, t)) ± qdc (Es )∇ntc (r, t)
Here, the generation rate describes the generation of carriers upon absorption
of optical electromagnetic wave energy [Saleh and Teich, 2019, Chuang, 2012,
Piprek, 2018]
P abs (r, t)
G(Et , Ht ) = η
E ph
where η is the intrinsic quantum efficiency (number of electron-hole pairs gen-
erated by each absorbed photon), P abs (r, t) is the absorbed power density of
optical waves, E ph = hν is the photon energy, h is the Planck constant, and
ν is the frequency of the optical wave. Here, ν must be high enough such that
61
E ph is large enough to excite electrons, e.g., E ph should be larger than the
bandgap energy E g in direct bandgap semiconductors [Saleh and Teich, 2019,
Chuang, 2012, Piprek, 2018]. Apparently, P abs (r, t) only depends on the op-
tical electromagnetic fields, and hence G(Et , Ht ) is only a function of Et and
Ht . A simple approximation for P abs (r, t) is P abs (r, t) ≈ α Et (r, t) × Ht (r, t) ,
where α is the photon absorption coefficient.
Figure 1.23: Time marching scheme. Source: [Chen and Bagci, 2020b].
A complete multiphysics framework for modeling photoconductive devices
consists of a steady-state solver that solves the Poisson-drift-diffusion system
1.93-1.95 and a transient solver that solves the Maxwell-drift-diffusion sys-
tem 1.96-1.99. The steady-state solutions are used as inputs to the transient
solver. The steady-state solver uses the Gummel iteration method to take the
nonlinearity into account [Chen and Bagci, 2020c]. At every iteration of this
method, one has to solve two partial differential equations, i.e., the nonlin-
ear Poisson equation and the linearized drift-diffusion equation. The transient
solver directly solves the time-dependent Maxwell equations and drift-diffusion
equations using explicit time integration methods and takes into account their
coupling by alternately feeding the updated solutions into each other during
the time marching. Since the carrier response is much slower than the time
62
variation of the electromagnetic waves, the drift-diffusion equations are up-
dated using a larger time-step size (typically 10 times larger). This mixed-step
explicit time integration scheme can greatly reduce the computational cost.
The time marching scheme is illustrated in Fig. 1.23 [Chen and Bagci, 2020b].
For illustration, the time-step size for the DD equations is assumed to be
twice the step size for the Maxwell equations. Let the time steps of the two
systems are synchronized at time T = T ′ . The generation rate GT is first
calculated from Et,T and Ht,T and then used to update the carrier densities
′ ′
nt,T
c (from step T ′ − ∆T ′ to T ′ ). Then, nt,T
c are used to compute the cur-
rent densities Jt,T t,T +∆T

c and Jc . Jt,T t,T +∆T
c and Jc are then used to update the
Maxwell equations at steps T and T +∆T to produce {Et,T +∆T , Ht,T +∆T } and
{Et,T +2∆T , Ht,T +2∆T }, respectively. The time steps of the two systems match
again at time T + 2∆T = T ′ + ∆T ′ and the process described here is repeated.
Note that G̃T +2∆T is calculated as the average value (GT +∆T + GT +2∆T )/2.
The above multiphysics framework calls for solving four partial differ-
ential equations, namely, the nonlinear Poisson equation and the stationary
drift-diffusion equation [Chen and Bagci, 2020c], the time-dependent Maxwell
equations, and the time-dependent drift-diffusion equations [Chen and Bagci,
2020b]. As discussed before, DG is preferred for multiphysics problems thanks
to its flexibility in discretizing geometries and equations, which helps to cap-
ture the multiscale features involved in the coupled systems of equations. A
special challenge in modeling semiconductor carrier dynamics (by solving drift-
diffusion equation) is that they belong to the group of so-called convection-
dominated convection-diffusion problems. Efficient solution of these problems
has been a challenging research topic in computational mathematics for more
than 40 years [Shu, 2016a]. The difficulty stems from the existence of exponen-
63
tial boundary layers in the solution, which makes numerical methods unstable
unless extremely fine meshes are used. To this end, FEM requires additional
techniques to ensure stability, and FVM (generalized from the Scharfetter-
Gummel scheme) poses restrictions on the meshes [Chen and Bagci, 2020c].
DG seems to be favorable for this type of problem since it is free from these
issues while providing a high-order accuracy. As discussed previously, since all
electric field components described by Poisson and Maxwell equations are used
in the drift-diffusion equations, nodal basis functions are preferred here. With
these discussions, all the four partial differential equations are discretized using
a unified nodal DG framework. DG discretization of Maxwell equations is the
same as that introduced at the beginning of this chapter. In the following, DG
formulations for the remaining three equations are briefly introduced.
The time-dependent drift-diffusion equation is solved using the local DG
(LDG) method. Since the electron and hole drift-diffusion equations only differ
by the sign in front of the drift term, we only discuss the electron drift-diffusion
equation. Also note that the subscript “e” (meaning electron) and the super-
script “t” (meaning transient) are dropped from the variables to simplify the
notation. Since Es (r) is provided as an input to the time-domain simulation,
variables that are functions of it are assumed to change with r. Furthermore,
as the three drift terms in (1.103) are treated in the same way, for brevity,
they are combined into one term and is denoted by v(r, t)n(r, t). Under those
notation simplifications, the electron drift-diffusion equations (1.102)-(1.103)
64
are expressed as the following initial-boundary value problem
∂n(r, t)
= ∇ · [d(r)q(r, t)] + ∇ · [v(r, t)n(r, t)] − R(r, t), r ∈ Ω (1.104)
∂t
q(r, t) = ∇n(r, t), r ∈ Ω (1.105)
n(r, t) = fD (r), r ∈ ∂ΩD (1.106)
n̂(r) · [d(r)q(r, t) + v(r, t)n(r, t)] = fR (r), r ∈ ∂ΩR . (1.107)
Here, q(r, t) is an auxiliary variable introduced to reduce the order of the spa-
tial derivative in the diffusion term, R(r, t) ≡ Rt (nte , nth ) − G(Et , Ht ), ∂ΩD and
∂ΩR represent the surfaces where Dirichlet and Robin boundary conditions are
enforced and fD and fR are the coefficients associated with these boundary con-
ditions, respectively, and n̂(r) denotes the outward normal vector ∂ΩR . For the
problems considered in this work, ∂ΩD represents the electrode/semiconductor
interfaces and fD = 0; and ∂ΩR represents the semiconductor/insulator inter-
faces and fR = 0 indicating no carrier spills out those interfaces [Schroeder,
1994].
DG discretization of the above initial-boundary value problem is similar to
the previous process described for Maxwell equations. Here, the nodal basis
functions (see Section 1.1.6) are used. Testing (1.104) and (1.105) with Ψl (r)
on element i and applying the divergence theorem yield the weak form
Z Z I
ni (r, t)
Ψl (r)dV =− d(r)qi (r, t) · ∇Ψl (r)dV + n̂(r) · (dq)∗ Ψl (r)dS
Ωi ∂t Ωi ∂Ωi
Z
− v(r, t)ni (r, t)·∇Ψl (r)dV (1.108)
Ωi
I Z
∗
+ n̂(r) · (vn) Ψl (r)dS − R(r, t)Ψl (r)dV
∂Ωi Ωi
Z Z I
ν
qi (r, t)Ψl (r)dV =− ni (r, t)∂ν Ψl (r)dV + n̂ν (r)n∗ Ψl (r)dS. (1.109)
Ωi Ωi ∂Ωi
65
Here, n∗ , (dq)∗ , and (vn)∗ are numerical fluxes defined as
n∗ = {n} + 0.5β̂ · n̂ [[n]] (1.110)
(dq)∗ = {dq} − 0.5β̂(n̂ · [[dq]]) (1.111)
(vn)∗ = {vn} + αn̂ [[n]] . (1.112)
where {⊙} = 0.5(⊙− + ⊙+ ) and [[⊙]] = ⊙− − ⊙+ , are “average” and “jump”
operators, ⊙ is a scalar or a vector variable. Superscripts “-” and “+” refer to
variables defined in element i and in its neighboring element, respectively. For
the diffusion term, 1.110 and 1.111 are the so-called LDG alternate flux [Cock-
burn et al., 2000]. The vector β̂ determines the upwind direction of n and (dq).
In LDG, opposite directions are chosen for n and (dq), while the precise direc-
tion of each variable is not important [Cockburn et al., 2000, Hesthaven and
Warburton, 2008, Shu, 2016a]. Here, we choose β̂ = n̂ on each element surface.
1.112 is the local Lax-Friedrichs flux [Hesthaven and Warburton, 2008], with
α = max(|n̂ · v− |, |n̂ · v+ |)/2 [Hesthaven and Warburton, 2008], mimics the
path of the information propagation. On ∂ΩD , n∗ = fD , (dq)∗ = (dq)− and
(vn)∗ = v− fD , and on ∂ΩR , n∗ = n− and (dq)∗ + (vn)∗ = fR .
Expanding ni (r, t) and qkν (r, t), ν ∈ {x, y, z}, using the set of Lagrange
polynomials
Np Np
X X
ni (r, t) ≃ ni (rl , t)Ψl (r) = ni,l (t)Ψl (r) (1.113)
l=1 l=1
Np Np
X X
qiν (r, t) ≃ qiν (rl , t)Ψl (r) = ν
qi,l (t)Ψl (r) (1.114)
l=1 l=1
66
yields the semi-discretized form
ni
Mi = Ci ni + Cii′ nk′ + Di d¯i qi + Dii′ d¯k′ qi′ − Bni (1.115)
∂t
Mqi qi = Gi nk + Gii′ ni′ + Bqi (1.116)
where the global unknown vectors are defined as ni = [ni,1 , ..., ni,Np ]T and
qi = [qxi , qyi , qzi ]T , with qνi = [qi,1

ν , ..., q ν
i,Np ], ν ∈ {x, y, z}.
In 1.115-1.116, Mi and Mqi are mass matrices. Mqi is a 3 × 3 block diagonal
matrix with blocks Mi defined in 1.55. d¯k is a diagonal matrix with entries
d1 , ..., dN , where di = (dxi , dyi , dzi ) and dνi (i) = di (rl ), l = 1, ..., Np , ν ∈ {x, y, z}.
We note that d(r) is assumed isotropic and constant in each element.
Matrices Gi and Gii′ , and Di and Dii′ correspond to the gradient and
the divergence operators, respectively. For the LDG flux, Di = −GTi and
Dii′ = −GTii′ . Gi is a 3Np ×Np matrix and it has contributions from the volume
and surface integral terms on the right hand side of 1.109: Gi = GV S

i + Gi ,
h iT
x y z T
GVi = S̃i S̃i S̃i , GSi = [Lxi Lyi Lzi ] , where S̃νi = −[Sνi ]T and
1 + sign(β̂ · n̂)
I
Lνi (l, k) = θi (k) n̂ν (r)Ψl (r)Ψk (r)dS.
2 ∂Ωii′
Here, ∂Ωii′ denotes the interface connecting element i and i′ and θi (k) selects
the interpolation nodes on the interface


1,
 rk ∈ Ωi , rk ∈ ∂Ωii′
θi (k) = .
0,

otherwise
Matrix Gii′ corresponds to the surface integral term in 1.109, which involves
T
the unknowns from neighboring elements of element i: Gii′ = Lxi′ Lyi′ Lzi′ ,

67
where
1 − sign(β̂ · n̂)
I
Lνi′ (l, k) = θi′ (j) n̂ν (r)Ψl (r)Ψk (r)dS.
2 ∂Ωii′
Similarly, matrix Ci has contributions from the volume integral and the
surface integral related to the drift term on the right hand side of (1.108):
Ci = CV S
i + Ci , where
X Z
CV
i (l, k) =− vν (r, t)∂ν Ψl (r)Ψk (r)dV
ν Ωi
and
I
1X
CSi (l, k) = θi (k) [ n̂ν (r)vν (r, t) + α(r, t)]Ψl (r)Ψk (r)dS.
∂Ωii′ 2 ν
Matrix Cii′ is from the last surface integral term in (1.108), which involves the
unknowns from neighboring elements of the element i:
I
1X
Cii′ (l, k) = θi′ (k) [ n̂ν (r)vν (r, t) − α(r, t)]Ψl (r)Ψk (r)dS.
∂Ωii′ 2 ν
Matrices Bni and Bqi are contributed from the force term and boundary con-
ditions (where element k ′ does not exist) and are expressed as
Z I
Bni (k) = R(r, t)Ψl (r)dV + fR (r)Ψl (r)dS
Ωi ∂Ωi ∩∂ΩR
I
+ n̂(r) · v(r, t)fD (r)Ψl (r)dS
∂Ωi ∩∂ΩD
I
Bq,ν
i (k) = n̂ν (r)fD (r)Ψl (r)dS.
∂Ωi ∩∂ΩD
An explicit third-order total-variation-diminishing Runge-Kutta method
[Shu and Osher, 1988] is used for integrating 1.115-1.116 in time. With initial
68
value n(r, t = 0) = 0, time samples of the unknown vector ni are obtained step
by step in time.
The nonlinear Poisson equation and the stationary drift-diffusion equation
are discretized using similar LDG schemes. Details of their corresponding DG
formulations can be found in [Chen and Bagci, 2020c]. One should note that
these two equations are time-independent and belong to elliptic-type partial
differential equations. Solving them calls for solving global linear systems. For-
tunately, for photoconductive device simulation, one only needs to solve the
steady-state once (for each bias voltage), and the solution can be reused in
different transient studies.
To demonstrate the applicability of the above DG-based multiphysics frame-
work, two practical device simulation examples are presented below. We first
consider a conventional face-to-face dipole photoconductive device [Tani et al.,
1997]. This device has a relatively simple structure and it can be simulated us-
ing the FDTD [Moreno et al., 2014b] or FEM-based [Burford and El-Shenawee,
2016, Bashirpour et al., 2017] approach, where the carrier generation due to the
optical electromagnetic wave excitation is modeled using closed-form analytical
expressions [Moreno et al., 2014b, Burford and El-Shenawee, 2016, Bashirpour
et al., 2017].
Since the focus of the device study is the photocurrent response, we only
consider the central gap region of the device. The cross section of the device is
shown in Fig. 1.24. The width and height of the semiconductor and substrate
layers are 7 µm and 0.5 µm, respectively. Both electrodes are made of gold and
have a width of 1 µm and a height of 0.19 µm. In the transient simulation,
the permittivity of gold is represented using the Drude model [Gedney et al.,
2012]. The Drude model parameters of gold at the optical frequencies [Olmon
69
et al., 2012] as well as other physical parameters are shown in Table 1.3.
1.5
0.5
y (µm)
–0.5
–1.5
–4.5 –3.5 –2.5 –1.5 –0.5 0.5 1.5 2.5 3.5 4.5
x (µm)
–4.5
Figure 1.24: Cross section of the conventional PCD considered in the first
example. Reproduced with permission from [Chen and Bagci, 2020b].
In multiphysics simulations, proper boundary conditions must be set for
each of the above DG solvers, and the device should be meshed with respect
to the characteristic space and time scales corresponding to each equation.
The stationary and time-dependent drift-diffusion equations are solved only
within the photoconductive semiconductor (LT-GaAs) layer, while the Poisson
equation and the time-dependent Maxwell equations are solved in the whole
domain. The boundary conditions are as follows:
1. Stationary and time-dependent drift-diffusion equations: On electrode/semi-
conductor interfaces, metal contacts are assumed ideal Ohmic contacts

√
and Dirichlet boundary condition ne = (C+ C 2 + 4ni 2 )/2, nh = ni 2 /ne ,
which is derived using the local charge neutrality [Schroeder, 1994, Vasileska
et al., 2010], is used. On semiconductor/insulator interfaces, no carrier
spills happen, which corresponds to the homogeneous Robin boundary
condition n̂ · Je(h) = 0 [Schroeder, 1994, Chen et al., 2021].
70
2. Poisson equation: On electrode surfaces, assuming ideal Ohmic contact,
Dirichlet boundary condition φ = Vexternal +VT ln (nse /ni ) is enforced [Vasileska
et al., 2010]. On truncation boundary of the computation domain, homo-
geneous Neumann boundary condition n̂ · ∇φ = 0 is enforced under the
assumption that the static fields reaching the boundary are small and do
not change the solution near the device.
3. Maxwell equations: The computation domain is “wrapped around” by
PMLs [Gedney et al., 2009] truncated by first-order ABCs [Hesthaven
and Warburton, 2008].
The characteristic space and time characteristic scales are as follows:
1. Space scales: Characteristic space scales, more specifically, the electro-
magnetic wavelength at the optical and THz frequencies, the skin depth
of the electromagnetic wave in the metal ld [Olmon et al., 2012], Debye
length lD [Vasileska et al., 2010], and the Peclet number [Trangenstein,
2013] constrain the edge length used in the mesh discretizing the com-
p
putation domain. The Debye length lD = 2εVT /(qnc ) is a measure
of the carrier density variation in space [Vasileska et al., 2010]. Here,
nc is an estimate of maximum achievable carrier density. The Peclet
number ∆d CP , where ∆d is the largest edge length of the local mesh
element and CP = |µc E|/(2Dc ), has to be less than 1 to ensure the sta-
bility of the convection-diffusion component in the drift-diffusion equa-
tions [Trangenstein, 2013]. We should note here that the potential distri-
bution is smooth, therefore the mesh using the edge length determined by
the space scales described above discretizes it accurately. The values of
these space-scale parameters for the materials used in this example (see
Table 1.3) and the corresponding required edge lengths are provided in
71
Table 1.1: Length scales
Quantity Value (nm) Required mesh size (nm)

optical wavelength 800 ∼ 200
THz wavelength 3 × 105 ∼ 105
ld 25 ∼ 10
lD 30 ∼ 10
CP−1 10 ∼ 10
geometrical details 100 ∼ 100
Reproduced with permission from [Chen and Bagci, 2020b].
Table 1.1. We should note that the last number in Table 1.1 is the small-
est geometry dimension and the mesh used for this example represents
the geometry very accurately.
2. Time scales: The Courant-Friedrichs-Lewy (CFL) stability conditions of
Maxwell equations [Hesthaven and Warburton, 2008] and diffusion and
drift components of the drift-diffusion equations [Liu and Shu, 2016,
Wang et al., 2015], constrain the time-step sizes used in the transient
solution. Table 1.2 lists the values of the maximum time-step sizes al-
lowed by these three CFL conditions with the edge length obtained from
the characteristic space scale discussion above. We should note that the
time-step sizes required to resolve the periods of the optical and THz
electromagnetic waves are larger than those required by their CFL con-
ditions. Table 1.2 clearly shows that the drift-diffusion equations can be
integrated using a larger time-step size than the one required for the
integration of Maxwell equations.
Using these scales as a guide, the device is discretized with mesh sizes in
the range [10, 200] nm, where the finest meshes are used near the boundaries of
the semiconductor layer and near the metallic electrodes, the maximum mesh
72
Table 1.2: Time scales
Quantity Required time-step size (ps)

CFL (Maxwell equations) ∼ 10−7
CFL (DD drift term) ∼ 10−6
CFL (DD diffusion term) ∼ 10−6
size in the semiconductor layer and the substrate is 70 nm, and the maximum
mesh size in the rest of the computation domain is 200 nm. The time-step sizes
for the Maxwell and drift-diffusion equations are selected to be 10−7 ps and
5 × 10−7 ps, respectively.
The steady-state solutions are first solved from the Poisson-drift-diffusion
solver and used as inputs in the transient simulation. Here, we focus on the
transient simulation. An aperture source located at y = 0.8 µm generates op-
tical electromagnetic waves propagating from top to bottom. The pulse shape
parameters are given in Table 1.3. The intensity of the electromagnetic field
on the aperture has a Gaussian distribution with beam width 3 µm.
Figs. 1.25 (a) and (b) show Hzt (r, t) and nte (r, t) at several time instants,
respectively, and Figs. 1.26 (a) and (b) show the time signatures of the optical
electromagnetic wave excitation and the generated THz current and their spec-
trum obtained using Fourier transform. Figs. 1.25 (a) and (b) and Fig. 1.26
(a) also serve as a reference for the following discussion.
At t = 0.05 ps, the optical electromagnetic wave generated by the aperture
arrives the semiconductor layer. Because the permittivity of LT-GaAs is high,
a large part of the incident field’s energy is reflected back (see the reflected
wave above the air-semiconductor interface and behind the aperture). At the
same time, in the LT-GaAs layer, the incident field’s energy entering the device
73
Table 1.3: Physical parameters used for the PCD examples
Lasera fc = 375 THz, fw = 25 THz, Power = 0.63 mW

LT-GaAs ϵr = 13.26, µr = 1.0, α = 1 µm−1 , η = 1.0
SI-GaAs ϵr = 13.26, µr = 1.0
Metalb ϵ∞ = 1, ωp = 9.03/ℏ, γ = 0.053q/ℏ
Temperature 300 K
Vbias 10 V
C 1.3 × 1016 cm−3
ni 9 × 106 cm−3
µ0e = 8000cm2 /V/s, µ0h = 400cm2 /V/s
Mobility Vesat = 1.725×107 cm/s, Vhsat = 0.9×107 cm/s
βe = 1.82, βh = 1.75
τe = 0.3ps, τh = 0.4ps
Recombination ne1 = nh1 = 4.5 × 106 cm−3
CeA = ChA = 7 × 10−30 cm6 /s
a
fc is the center frequency and fw is the bandwidth.
b
Drude model parameters [Olmon et al., 2012].
is partially absorbed and carriers are generated near the air-semiconductor in-
terface. At t = 0.25 ps, the incident field reaches its pulse peak and the electron
density increases to ∼ 1011 cm−3 . The short incident field pulse passes quickly
and, after t = 0.5 ps, only some scattered fields reside in the high permittiv-
ity region due to internal reflections. During that time, the electron density
keeps increasing until the excitation pulse decays to of 20% its peak value at
t ≈ 0.4 ps, after t ≈ 0.4 ps, nte (r, t) decays slowly due to the recombination.
Comparing the electron density distributions at different time instants, it
can be clearly observed that electrons move toward the anode on the left side.
The picture of holes (not shown) is similar but with holes moving toward the
cathode on the right side with a lower speed. The resulting current shown in
74
t = 0.05 ps t = 0.25 ps
t = 0.5 ps t = 1 ps
(a)
t=0.05 ps
t=0.25 ps
t=0.5 ps
t=0.75 ps
t=1 ps
t=1.25 ps
t=1.5 ps
t=1.75 ps
(b)
Figure 1.25: (a) |Hz (r, t)| computed using the DG scheme at different time
instants. The dotted line shows the aperture of optical EM wave excitation.
The gray box indicates the LT-GaAs region. (b) ne (r, t) computed using the
DG scheme at different time instants. Reproduced with permission from [Chen
and Bagci, 2020b].
Figs. 1.26 (a) and (b) match well with the result presented in [Moreno et al.,
2014b], where the latter has been shown to agree with experimental results very
well [Moreno et al., 2014b]. Because of the simplicity of the single interface
75
(a)
(b)
Figure 1.26: (a) The time signature and (b) the spectrum of the optical EM
wave excitation and the generated THz current. Reproduced with permission
from [Chen and Bagci, 2020b].
scattering, the recorded electromagnetic field intensity in the semiconductor
layer has almost the same pulse shape as the optical excitation signal. This
explains why the analytical generation rate [Moreno et al., 2014b] works very
well.
Next, a plasmonic photoconductive device is simulated for comparison. The
cross section of the device is shown in Fig. 1.27. For this example, gold nanos-
tructures are added between the two electrodes. The periodicity, the duty cycle,
and the thickness of the nanograting are 180 nm, 5/9, and 190 nm, respec-
tively. The surfaces of the nanostructures are modeled as floating potential
surfaces for the stationary state simulation (i.e., for solution of the Poisson
equation) [Chen and Bagci, 2020c, Chen et al., 2020a]. Here, finer meshes are
76
1.69
0.5
y (µm)
–0.5
–1.5
–4.5
–4.5 –3.5
–3.5 –2.5
–2.5 –1.5
–1.5 –0.5
–0.5 0.5
0.5 1.5
1.5 2.5
2.5 3.5
3.5 4.5
4.5
x x(µm)
(µm)
Figure 1.27: Cross section of the plasmonic PCD. Reproduced with permission
from [Chen and Bagci, 2020b].
used near and inside the nanostructures to correctly resolve the geometry and
the exponential decay of the plasmonic fields. In this region, the mesh has
a minimum edge length of 3 nm and the maximum edge length of the mesh
is allowed to reach 70 nm in the GaAs layers and 200 nm in the rest of the
computation domain. The corresponding time-step size used for the time in-
tegration of Maxwell equations is 0.3 × 10−7 ps. Other simulation parameters
and boundary conditions are same as those in the previous example.
Figs. 1.28 (a) and (b) show Hzt (r, t) and nte (r, t) at several time instants,
respectively. For comparison, Fig. 1.28 uses the same color scale as Fig. 1.25
for the previous example.
Fig. 1.28 shows that the transient electromagnetic fields on the plasmonic
photoconductive device are much stronger compared with the conventional
photoconductive device. In addition, plasmonic mode patterns are observed
at t = 0.05 ps and t = 0.25 ps. Accordingly, the level of electron density in
Fig. 1.28 is much higher and shows an inhomogeneous pattern. As time march-
ing goes on, electrons drift toward the anode. Figs. 1.29 (a) and (b) compare
77
t = 0.05 ps t = 0.25 ps
t = 0.5 ps t = 1 ps
(a)
t=0.05 ps
t=0.25 ps
t=0.5 ps
t=0.75 ps
t=1 ps
t=1.25 ps
t=1.5 ps
t=1.75 ps
(b)
Figure 1.28: (a) |Hz (r, t)| computed using the DG scheme at different time
instants. The dotted line shows the aperture of optical EM wave excitation.
The gray box indicates the LT-GaAs region. (b) ne (r, t) computed using the
DG scheme at different time instants. Reproduced with permission from [Chen
and Bagci, 2020b].
the time signatures of the current induced on the plasmonic photoconductive
device and the conventional one (previous example) and their spectrum ob-
tained using Fourier transform. The figures clearly show that the inclusion of
78
500
Conventional PCA
Plasmonic PCA
400 Plasmonic PCA (/7.315)
300
200
100
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2
Time (ps)
(a)
500
Amplitude (a.u.)
1
400
0.5
300
0
0 1 2 3 4 5
200 Conventional PCA Frequency (THz)
Plasmonic PCA
100 Plasmonic PCA (/7.315)
0
0 0.5 1 1.5 2 2.5
Time (ps)
(b)
Figure 1.29: (a) The time signature and (b) the spectrum of the THz current
generated on the plasmonic PCD and on the conventional PCD. Reproduced
with permission from [Chen and Bagci, 2020b].
the nanostructures in the photoconductive device design enhances the current
by almost 7 times.
The above examples show the applicability of the multiphysics DG frame-
work. However, due to the high computational cost, only the central gap region
of the device is considered. Two extensions toward more efficient simulation of
photoconductive devices have been reported. In [Chen et al., 2021], a unit-cell
scheme is presented for fast characterization of nanostructure-enhanced pho-
toconductive devices. By using a “potential-drop” boundary condition, this
scheme models a biased photoconductive device within a unit cell of the nanos-
79
tructure and thus greatly reduces the computational cost. Another challenge
in modeling full-structure photoconductive devices is the inclusion of both op-
tical and terahertz (THz) frequency electromagnetic waves. The wavelength of
optical and THz waves differs by almost 400 times. Correspondingly, the THz
antenna is typically 400 times larger than the size of the nanostructure. This
scale difference makes the direct simulation of full-structure photoconductive
devices almost impossible. This challenge has been addressed by a DG-based
dual-mesh scheme [Chen and Bagci, 2020a], where the optical and the THz
waves are treated separately in two DG solvers and their two-way couplings
are taken into account using high-order interpolations. With this dual-mesh
scheme, the space-time-dependent conductivity in photoconductive materials
can be strictly modeled in the THz antenna simulation. Thus, the impedance
matching between different photoconductive structures and THz antennas can
be strictly analyzed.
Bibliography
ATLAS user’s manual: device simulation software. Silvaco, Int. St. Clara, CA,
2016.
Semiconductor Module User’s Guide, version 5.3a. COMSOL Multiphysics,
2017.
Minimos-NT User Manual. Inst. Microelectron. (TU Vienna) Glob. TCAD
Solut. GmbH, St. Clara, CA, 2017.
Transient Electromagnetic-Thermal Simulation of Dispersive Media Using
DGTD Method. IEEE Trans. Electromagn. Compat., 61(4):1305–1313, aug
80
2019.
S. Abarbanel, D. Gottlieb, and J. S. Hesthaven. Long Time Behavior of the
Perfectly Matched Layer Equations in Computational Electromagnetics. J.
Sci. Comput., 17(1-4):405–422, 2002.
J. Alvarez, Luis D. Angulo, M. Fernández Pantoja, A. Rubio Bretones, and
Salvador G. Garcia. Source and boundary implementation in vector and
scalar DGTD. IEEE Trans. Antennas Propag., 58(6):1997–2003, jun 2010.
Luis Angulo, Jesus Alvarez, Mario Pantoja, Salvador Garcia, and A Bretones.
Discontinuous Galerkin Time Domain Methods in Computational Electrody-
namics: State of the Art. Forum Electromagn. Res. Methods Appl. Technol.,
10, 2015.
Hakan Bagci, Ali E. Yilmaz, Jian Ming Jin, and Eric Michielssen. Fast and
rigorous analysis of EMC/EMI phenomena on electrically large and complex
cable-loaded structures. IEEE Trans. Electromagn. Compat., 49(2):361–381,
may 2007.
M Bashirpour, S Ghorbani, M Kolahdouz, M Neshat, M Masnadi-Shirazi, and
H Aghababa. Significant performance improvement of a terahertz photo-
conductive antenna using a hybrid structure. RSC Adv., 7(83):53010–53017,
2017.
Jean Pierre Berenger. A perfectly matched layer for the absorption of electro-
magnetic waves. J. Comput. Phys., 114(2):185–200, oct 1994.
Françoise Bourdel, Pierre-Alain Mazet, and Philippe Helluy. Resolution of
the non-stationary or harmonic Maxwell equations by a discontinuous finite
element method: Application to an E.M.I. (electromagnetic impulse) case.
81
In Proc. 10th Int. Conf. Comput. methods Appl. Sci. Eng. {Computing}
methods Appl. Sci. Eng., pages 405–422, USA, mar 1991.
Nathan Burford and Magda El-Shenawee. Computational modeling of plas-
monic thin-film terahertz photoconductive antennas. J. Opt. Soc. Am. B,
33(4):748–759, 2016.
Liang Chen and Hakan Bagci. A Dual-mesh Framework for Multiphysics Sim-
ulation of Photoconductive Terahertz Devices. In 2020 33rd Gen. Assem.
Sci. Symp. Int. Union Radio Sci. URSI GASS 2020, aug 2020a.
Liang Chen and Hakan Bagci. Multiphysics simulation of plasmonic photocon-
ductive devices using discontinuous Galerkin methods. IEEE J. Multiscale
Multiphys. Comput. Tech., 5:188–200, 2020b.
Liang Chen and Hakan Bagci. Steady-state simulation of semiconductor de-
vices using discontinuous Galerkin methods. IEEE Access, 8:16203–16215,
2020c.
Liang Chen, Kostyantyn Sirenko, and Hakan Bagci. Multiphysics Analysis of
Plasmonic Photomixers under Periodic Boundary Conditions using Discon-
tinuous Galerkin Time Domain Method. In Prog. Electromagn. Res. Symp.
August 1-4, 2018 in Toyama, Japan, 2018.
Liang Chen, Ming Dong, and Hakan Bagci. Modeling floating potential con-
ductors using discontinuous Galerkin method. IEEE Access, 8:7531–7538,
2020a.
Liang Chen, Mehmet Burak Ozakin, Shehab Ahmed, and Hakan Bagci.
A Memory-efficient Implementation of Perfectly Matched Layer with
82
Smoothly-varying Coefficients in Discontinuous Galerkin Time-Domain
Method. IEEE Trans. Antennas Propag., 2020b.
Liang Chen, Kostyantyn Sirenko, Ping Li, and Hakan Bagci. Efficient dis-
continuous Galerkin scheme for analyzing nanostructured photoconductive
devices. Opt. Express, 29(9):12903–12917, 2021.
Pai Yen Chen, Christos Argyropoulos, and Andrea Alu. Terahertz antenna
phase shifters using integrally-gated graphene transmission-lines. IEEE
Trans. Antennas Propag., 61(4):1528–1537, 2013.
Shun Lien Chuang. Physics of photonic devices, volume 80. John Wiley and
Sons, 2012.
Bernardo Cockburn, George E. Karniadakis, and Chi-Wang Shu. The Devel-
opment of Discontinuous Galerkin Methods. In Dev. Discontinuous Galerkin
Methods, pages 3–50. Springer, Berlin, Heidelberg, 2000.
Ming Dong, Ping Li, and Hakan Bagci. An Explicit Time Domain Finite
Element Boundary Integral Method with Element Level Domain Decompo-
sition for Electromagnetic Scattering Analysis. In 14th Eur. Conf. Antennas
Propagation, EuCAP 2020, mar 2020.
Ming Dong, Liang Chen, Lijun Jiang, Ping Li, and Hakan Bagci. An explicit
time domain finite element boundary integral method for analysis of electro-
magnetic scattering. IEEE Trans. Antennas Propag., pages 1–1, 2022. doi:
10.1109/TAP.2022.3142319.
Stylianos Dosopoulos and Jin Fa Lee. Interconnect and lumped elements mod-
eling in interior penalty discontinuous Galerkin time-domain methods. J.
Comput. Phys., 229(22):8521–8536, nov 2010a.
83
Stylianos Dosopoulos and Jin Fa Lee. Interior penalty discontinuous galerkin
finite element method for the time-dependent first order maxwell’s equations.
IEEE Trans. Antennas Propag., 58(12):4085–4090, dec 2010b.
Stylianos Dosopoulos, Judith D. Gardiner, and Jin-Fa Lee. An MPI/GPU
parallelization of an interior penalty discontinuous Galerkin time domain
method for Maxwell’s equations. Radio Sci., 46(3), jun 2011.
Stylianos Dosopoulos, Bo Zhao, and Jin Fa Lee. Non-conformal and parallel
discontinuous Galerkin time domain method for Maxwell’s equations: EM
analysis of IC packages. J. Comput. Phys., 238:48–70, apr 2013.
Mina Feizi, Vahid Nayyeri, and Omar M. Ramahi. Modeling Magne-
tized Graphene in the Finite-Difference Time-Domain Method Using an
Anisotropic Surface Boundary Condition. IEEE Trans. Antennas Propag.,
66(1):233–241, jan 2018.
Loula Fezoui, Stéphane Lanteri, Stéphanie Lohrengel, and Serge Piperno. Con-
vergence and stability of a discontinuous Galerkin time-domain method for
the 3D heterogeneous Maxwell equations on unstructured meshes. ESAIM
Math. Model. Numer. Anal., 39(6):1149–1176, nov 2005.
F. Javier Garcı́a De Abajo. Graphene nanophotonics. Science (80-. )., 339
(6122):917–918, feb 2013.
Stephen D. Gedney. An anisotropie perfectly matched layer-absorbing medium
for the truncation of FDTD lattices. IEEE Trans. Antennas Propag., 44(12):
1630–1639, 1996.
Stephen D. Gedney and Bo Zhao. An auxiliary differential equation formula-
84
tion for the complex-frequency shifted PML. IEEE Trans. Antennas Propag.,
58(3):838–847, mar 2010.
Stephen D. Gedney, Chong Luo, Bryan Guernsey, J. Alan Roden, Robert Craw-
ford, and Jeffrey A. Miller. The Discontinuous Galerkin Finite Element Time
Domain method (DGFETD): A high order, globally-explicit method for par-
allel computation. In IEEE Int. Symp. Electromagn. Compat., 2007.
Stephen D Gedney, Chong Luo, J Alan Roden, Robert D Crawford, Bryan
Guernsey, Jeffrey A Miller, Tyler Kramer, and Eric W Lucas. The discon-
tinuous Galerkin finite-element time-domain method solution of Maxwell’s
equations. Appl. Comput. Electromagn. Soc. J., 24(2):129, 2009.
Stephen D. Gedney, John C. Young, Tyler C. Kramer, and J. Alan Roden. A
discontinuous Galerkin finite element time-domain method modeling of dis-
persive media. IEEE Trans. Antennas Propag., 60(4):1969–1977, apr 2012.
A. K. Geim and K. S. Novoselov. The rise of graphene. Nat. Mater., 6(3):
183–191, mar 2007.
Bjørn Gustavsen. Improving the pole relocating properties of vector fitting.
IEEE Trans. Power Deliv., 21(3):1587–1592, jul 2006.
Bjørn Gustavsen and Adam Semlyen. Rational approximation of frequency
domain responses by vector fitting. IEEE Trans. Power Deliv., 14(3):1052–
1059, 1999.
Thomas Hagstrom, Timothy Warburton, and Dan Givoli. Radiation boundary
conditions for time-dependent waves based on complete plane wave expan-
sions. J. Comput. Appl. Math., 234(6):1988–1995, aug 2010.
85
George W. Hanson. Dyadic Green’s functions and guided surface waves for a
surface conductivity model of graphene. J. Appl. Phys., 103(6):064302, mar
2008.
Michael Harmon, Irene M Gamba, and Kui Ren. Numerical algorithms based
on Galerkin methods for the modeling of reactive interfaces in photoelectro-
chemical (PEC) solar cells. J. Comput. Phys., 327:140–167, 2016.
J. S. Hesthaven and T. Warburton. Nodal high-order methods on unstructured
grids: I. Time-domain solution of Maxwell’s equations. J. Comput. Phys.,
181(1):186–221, sep 2002.
Jan S. Hesthaven and Tim Warburton. Nodal Discontinuous Galerkin Methods,
volume 54 of Texts in Applied Mathematics. Springer New York, New York,
NY, 2008.
Robert L. Higdon. Absorbing Boundary Conditions for Difference Approxima-
tions to the Multi-Dimensional Wave Equation. Math. Comput., 47(176):
437, oct 1986.
L Homsi, C Geuzaine, and L Noels. A coupled electro-thermal Discontinu-
ous Galerkin method. J. Comput. Phys., 348:231–258, 2017. ISSN 0021-
9991. doi: https://doi.org/10.1016/j.jcp.2017.07.028. URL https://www.
sciencedirect.com/science/article/pii/S0021999117305363.
G B Jacobs and J S Hesthaven. High-order nodal discontinuous Galerkin
particle-in-cell method on unstructured grids. J. Comput. Phys., 214(1):
96–121, 2006.
Ian Jeffrey. Finite-Volume Simulations of Maxwell’s Equations on Unstructured
Grids. PhD thesis, University of Manitoba, 2011.
86
D. Jiao, A. A. Ergin, B. Shanker, E. Michielssen, and J. M. Jin. A fast higher-
order time-domain finite element-boundary integral method for 3-D electro-
magnetic scattering analysis. In IEEE Antennas Propag. Soc. AP-S Int.
Symp., volume 4, pages 334–337, 2001.
J.M. Jin. The Finite Element Method in Electromagnetics. John Wiley and
Sons. Inc., New York, second edition, 2002.
Adour V. Kabakian, Vijaya Shankar, and William F. Hall. Unstructured grid-
based discontinuous Galerkin method for broadband electromagnetic simu-
lations. J. Sci. Comput., 20(3):405–431, jun 2004.
A. Klöckner, T. Warburton, J. Bridge, and J. S. Hesthaven. Nodal discontin-
uous Galerkin methods on graphics processors. J. Comput. Phys., 228(21):
7863–7882, nov 2009.
Frank H.L. Koppens, Darrick E. Chang, and F. Javier Garcı́a De Abajo.
Graphene plasmonics: A platform for strong light-matter interactions. Nano
Lett., 11(8):3370–3377, aug 2011.
Ryogo Kubo. Statistical’Mechanical Theory of Irreversible Processes. I. Gen-
eral Theory and Simple Applications to Magnetic and Conduction Problems.
J. Phys. Soc. Japan, 12(6):570–586, dec 1957.
Ping Li and Li Jun Jiang. A hybrid electromagnetics-circuit simulation method
exploiting discontinuous Galerkin finite element time domain method. IEEE
Microw. Wirel. Components Lett., 23(3):113–115, 2013a.
Ping Li and Li Jun Jiang. Simulation of electromagnetic waves in the magne-
tized cold plasma by a DGFETD method. IEEE Antennas Wirel. Propag.
Lett., 12:1244–1247, 2013b.
87
Ping Li and Li Jun Jiang. Integration of arbitrary lumped multiport circuit
networks into the discontinuous galerkin time-domain analysis. IEEE Trans.
Microw. Theory Tech., 61(7):2525–2534, 2013c.
Ping Li and Li Jun Jiang. Modeling of magnetized graphene from microwave to
thz range by DGTD with a scalar RBC and an ADE. IEEE Trans. Antennas
Propag., 63(10):4458–4467, oct 2015.
Ping Li, Li Jun Jiang, and Hakan Bagci. Cosimulation of electromagnetics-
circuit systems exploiting DGTD and MNA. IEEE Trans. Components,
Packag. Manuf. Technol., 4(6):1052–1061, 2014a.
Ping Li, Yifei Shi, Li Jun Jiang, and Hakan Bagci. A hybrid time-domain
discontinuous galerkin-boundary integral method for electromagnetic scat-
tering analysis. IEEE Trans. Antennas Propag., 62(5):2841–2846, 2014b.
Ping Li, Li Jun Jiang, and Hakan Bagci. A Resistive Boundary Condition
Enhanced DGTD Scheme for the Transient Analysis of Graphene. IEEE
Trans. Antennas Propag., 63(7):3065–3076, jul 2015.
Ping Li, Li Jun Jiang, and Hakan Bagci. Discontinuous Galerkin Time-Domain
Modeling of Graphene Nanoribbon Incorporating the Spatial Dispersion Ef-
fects. IEEE Trans. Antennas Propag., 66(7):3590–3598, jul 2018.
Chao Ping Lin, Su Yan, Robert R. Arslanbekov, Vladimir Kolobov, and
Jian Ming Jin. A DGTD algorithm with dynamic h-adaptation and lo-
cal time-stepping for solving Maxwell’s equations. In 2016 IEEE Antennas
Propag. Soc. Int. Symp. APSURSI 2016 - Proc., pages 2079–2080, oct 2016.
Hai Lin, Mario F. Pantoja, Luis D. Angulo, Jesus Alvarez, Rafael G. Martin,
and Salvador G. Garcia. FDTD modeling of graphene devices using complex
88
conjugate dispersion material model. IEEE Microw. Wirel. Components
Lett., 22(12):612–614, 2012.
Meilin Liu, Kostyantyn Sirenko, and Hakan Bagci. An efficient discontinu-
ous Galerkin finite element method for highly accurate solution of maxwell
equations. IEEE Trans. Antennas Propag., 60(8):3992–3998, 2012.
YunXian Liu and Chi-Wang Shu. Analysis of the local discontinuous Galerkin
method for the drift-diffusion model of semiconductor devices. Sci. China
Math., 59(1):115–140, 2016.
Ignacio Llatser, Christian Kremers, Albert Cabellos-Aparicio, Josep Miquel
Jornet, Eduard Alarcón, and Dmitry N. Chigrin. Graphene-based nano-
patch antenna for terahertz radiation. Photonics Nanostructures - Fundam.
Appl., 10(4):353–358, oct 2012.
Giampiero Lovat. Equivalent circuit for electromagnetic interaction and trans-
mission through graphene sheets. IEEE Trans. Electromagn. Compat., 54
(1):101–109, feb 2012.
Enrique Moreno, M F Pantoja, F G Ruiz, J B Roldan, and S G Garcia. On
the numerical modeling of terahertz photoconductive antennas. J. Infrared
Millim. Terahertz Waves, 35(5):432–444, 2014a.
Enrique Moreno, Mario Fernandez Pantoja, Salvador G Garcia, Amelia Rubio
Bretones, and Rafael Gómez Martin. Time-domain numerical modeling of
{TH}z photoconductive antennas. IEEE Trans. THz Sci. Technol., 4(4):
490–500, 2014b.
Gerrit Mur. Absorbing Boundary Conditions for the Finite-Difference Approx-
89
imation of the Time-Domain Electromagnetic-Field Equations. IEEE Trans.
Electromagn. Compat., EMC-23(4):377–382, 1981.
Vahid Nayyeri, Mohammad Soleimani, and Omar M. Ramahi. Wideband mod-
eling of graphene using the finite-difference time-domain method. IEEE
Trans. Antennas Propag., 61(12):6107–6114, 2013.
Robert L Olmon, Brian Slovick, Timothy W Johnson, David Shelton, Sang-
Hyun Oh, Glenn D Boreman, and Markus B Raschke. Optical dielectric
function of gold. Phys. Rev. B, 86(23):235147, 2012.
J Piprek, editor. Handbook of Optoelectronic Device Modeling and Simulation.
CRC Press, Boca Raton, 2018.
W.H. Reed and T.R. Hill. Triangular mesh methods for the neutron transport
equation. In Proc. Am. Nucl. Soc., pages 1–23, 1973.
Q Ren, Y Bian, L Kang, P L Werner, and D H Werner. Leap-Frog Contin-
uous–Discontinuous Galerkin Time Domain Method for Nanoarchitectures
With the Drude Model. J. Light. Technol., 35(22):4888–4896, 2017. doi:
10.1109/JLT.2017.2760913.
Bahaa E A Saleh and Malvin Carl Teich. Fundamentals of photonics. Hoboken,
NJ, USA: John Wiley and Sons, 2019.
Krishnaswamy Sankaran. Accurate domain truncation techniques for time-
domain conformal methods. PhD thesis, ETH ZUTICH, 2007.
Dietmar Schroeder. Modelling of Interface Carrier Transport for Device Sim-
ulation. Springer-Verlag Wien, 1994.
90
Frank Schwierz. Graphene transistors. Nat. Nanotechnol., 5(7):487–496, may
2010.
Siegfried Selberherr. Analysis and simulation of semiconductor devices.
Springer-Verlag Wien New York, 1984.
Balasubramaniam Shanker, Mingyu Lu, A. Arif Ergin, and Eric Michielssen.
Plane-wave time-domain accelerated radiation boundary kernels for FDTD
analysis of 3-D electromagnetic phenomena. IEEE Trans. Antennas Propag.,
53(11):3704–3716, nov 2005.
Olga V. Shapoval, Juan Sebastian Gomez-Diaz, Julien Perruisseau-Carrier,
Juan R. Mosig, and Alexander I. Nosich. Integral equation analysis of plane
wave scattering by coplanar graphene-strip gratings in the thz range. IEEE
Trans. Terahertz Sci. Technol., 3(5):666–674, 2013.
Pankaj Sharma, Julien Perruisseau-Carrier, Clara Moldovan, and Adrian Mihai
Ionescu. Electromagnetic performance of RF NEMS graphene capacitive
switches. IEEE Trans. Nanotechnol., 13(1):70–79, jan 2014.
Chi Wang Shu. Discontinuous galerkin methods for time-dependent convection
dominated problems: Basics, recent developments and comparison with other
methods. In Lect. Notes Comput. Sci. Eng., volume 114, pages 369–397.
Springer Verlag, 2016a.
Chi Wang Shu. Discontinuous Galerkin methods for time-dependent convection
dominated problems: Basics, recent developments and comparison with other
methods. In Gabriel R Barrenechea, Franco Brezzi, Andrea Cangiani, and
Emmanuil H Georgoulis, editors, Build. Bridg. Connect. Challenges Mod.
91
Approaches to Numer. Partial Differ. Equations, pages 371–399. Springer
International Publishing, Cham, 2016b.
Chi Wang Shu and Stanley Osher. Efficient implementation of essentially non-
oscillatory shock-capturing schemes. J. Comput. Phys., 77(2):439–471, 1988.
K Sirenko, Y Sirenko, and H Bagci. Exact Absorbing Boundary Conditions
for Periodic Three-Dimensional Structures: Derivation and Implementation
in Discontinuous Galerkin Time-Domain Method. IEEE J. Multiscale Mul-
tiphys. Comput. Tech., 3:108–120, 2018.
Kostyantyn Sirenko, Meilin Liu, and Hakan Bagci. Incorporation of exact
boundary conditions into a discontinuous galerkin finite element method
for accurately solving 2d time-dependent maxwell equations. IEEE Trans.
Antennas Propag., 61(1):472–477, 2013.
Dimitrios L. Sounas and Christophe Caloz. Gyrotropy and nonreciprocity of
graphene for microwave applications. IEEE Trans. Microw. Theory Tech.,
60(4):901–914, apr 2012.
Qingtao Sun, Qiwei Zhan, Qiang Ren, and Qing Huo Liu. Wave Equation-
Based Implicit Subdomain DGTD Method for Modeling of Electrically Small
Problems. IEEE Trans. Microw. Theory Tech., 65(4):1111–1119, apr 2017.
M. Tamagnone, J. S. Gómez-Dı́az, J. R. Mosig, and J. Perruisseau-Carrier.
Analysis and design of terahertz antennas based on plasmonic resonant
graphene sheets. J. Appl. Phys., 112(11):114915, dec 2012.
Masahiko Tani, Shuji Matsuura, Kiyomi Sakai, and Shin-ichi Nakashima. Emis-
sion characteristics of photoconductive antennas based on low-temperature-
grown GaAs and semi-insulating GaAs. Appl. Opt., 36(30):7853–7859, 1997.
92
John A Trangenstein. Numerical solution of elliptic and parabolic partial dif-
ferential equations with CD-ROM. Cambridge, UK: Cambridge University
Press, 2013.
Dragica Vasileska, Stephen M Goodnick, and Gerhard Klimeck. Computa-
tional Electronics: Semiclassical and quantum device modeling and simula-
tion. Boca Raton, FL, USA: CRC press, 2010.
Haijin Wang, Chi-Wang Shu, and Qiang Zhang. Stability and error estimates of
local discontinuous Galerkin methods with implicit-explicit time-marching
for advection-diffusion problems. SIAM J. Numer. Anal., 53(1):206–227,
2015.
Xiang Hua Wang, Wen Yan Yin, and Zhizhang Chen. Matrix exponential
FDTD modeling of magnetized graphene sheet. IEEE Antennas Wirel.
Propag. Lett., 12:1129–1132, 2013.
Su Yan, Andrew D Greenwood, and Jian-Ming Jin. Modeling of plasma for-
mation during high-power microwave breakdown in air using the discontinu-
ous Galerkin time-domain method. IEEE J. Multiscale Multiphys. Comput.
Tech., 1:2–13, 2016.
Ali E. Yilmaz, Jian Ming Jin, and Eric Michielssen. Time domain adaptive in-
tegral method for surface integral equations. IEEE Trans. Antennas Propag.,
52(10):2692–2708, oct 2004.
93

Discontinuous Galerkin Time-Domain Method in Electromagnetics

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Discontinuous Galerkin Time-Domain Method in Electromagnetics

Uploaded by

Copyright:

Available Formats

Discontinuous Galerkin Time-Domain Method

in Electromagnetics: From Nanostructure

Item Type Book Chapter

Authors Dong, Ming;Chen, Liang;Li, Ping;Jiang, Lijun;Bagci, Hakan

Eprint version Post-print

Rights This is an accepted manuscript version of a paper before final

Download date 2024-03-08 23:41:12

Link to Item http://hdl.handle.net/10754/685980

Abstract: In the last few decades, the discontinuous Galerkin time-domain

Keywords: discontinuous Galerkin time-domain method, Maxwell’s equa-

has been re-formulated and extended to numerous applications including but

not limited to gas dynamics, semiconductor physics, magneto-hydrodynamics,

fluid dynamics, and electromagnetics [Cockburn et al., 2000]. In the field of

electromagnetics, one of the very first works on a DG method formulated

and implemented to solve the time-domain Maxwell equations is from Bourdel

[Bourdel et al., 1991]. This first DG method uses a zeroth-order (piecewise

constant) discretization scheme but several high-order discontinuous Galerkin

of computational electromagnetics [Hesthaven and Warburton, 2008, Dosopou-

recent years, DGTD has successfully become a powerful numerical technique

to tackle a wide variety of challenging large scale electromagnetic problems

2002, Li and Jiang, 2013a,c, Li et al., 2014b, Li and Jiang, 2013b].

DGTD combines advantages of both the finite volume time-domain (FVTD)

Warburton, 2008]. Like FVTD, the information exchange between neighboring

all spatial operations to individual discretization elements. The resulting local

fore time integration/marching starts. Indeed, if an explicit time integration

be ported on high performance distributed memory systems or multi-threaded

GPUs to solve large scale electromagnetic problems [Klöckner et al., 2009,

Dosopoulos et al., 2011]. Finally, locality of the spatial operations makes it

easier to incorporate h- and/or p-adaptive refinement techniques [Shu, 2016a,

Hesthaven and Warburton, 2008] as well as non-conformal meshing strategies

[Hesthaven and Warburton, 2002].

DGTD also comes with the flexibility of choosing different time-stepping

schemes in different subdomains of the computation domain [Angulo et al.,

in subdomains with coarser meshes while an unconditionally-stable implicit

potentially reducing the overall computation time.

eral benefits in simulating various electromagnetic problems: It can provide

In this chapter, we first describe the basic formulation of underlying the

functions and explain how different boundary conditions can be incorporated

of a nodal DGTD scheme, where higher-order polynomials are used as basis

life practical problems. More specifically, novel techniques developed to enable

the application of DGTD to electromagnetic analysis of nanostructures and

multiphysics simulation of photoconductive devices/antennas are described.

For each application, we present several numerical examples to demonstrate

the accuracy, efficiency and robustness of the developed techniques.

1.1.1. DGTD formulation for Maxwell equations

chapter, dependence of the variables on location r and time t is not explicitly

stated (unless it is needed to clarify the formulation/notation).

To develop the DGTD formulation, the Maxwell equations above should

be re-expressed in the conservation form as [Hesthaven and Warburton, 2008]

S = [J/ε, 0]T and the flux vector

material dependent coefficient matrix A and the field vector U:

The coefficient matrix A determines the characteristics of the Maxwell system.

characteristic speed associated with the Maxwell system.

respectively. Ω can be divided into a set of non-overlapping tetrahedrons Ωi

with boundary ∂Ωi , i.e., Ω = ∪Ωi . In each mesh element Ωi , i = 1, · · · , N ,

after applying Galerkin testing [Hesthaven and Warburton, 2008] to equation

E and H in element i respectively as explained later) and using the divergence

(Gauss) theorem twice, we obtain

discontinuous, F ∗ must be carefully defined to ensure the uniqueness of the

mann problem with Rankine-Hugoniot jump condition. To obtain the Rankine-

where εi = diag(εi , εi , εi ), µi = diag(µi , µi , µi ) and

λ2,5 = √1 = +ci (1.9)

λ3,6 = − √ε1i µi = −ci