5/23/2012

1
Pioneering 21
st
Century
Electromagnetics and Photonics
The Finite-Difference
Frequency-Domain Method
Raymond C. Rumpf, Ph.D.
I am always doing that which I cannot do,
in order that I may learn how to do it.
Pablo Picasso
Short Course Outline
Background Topics
Numerical methods, electromagnetics, linear algebra,
finite-difference approximations
Break
The Finite-Difference Frequency-Domain Method
Yee grid, Maxwellmatrix, formulation, PML, TF/SF,
solution, post processing
Break
Implementation
3D2D, grids and materials, testing, convergence
Code examples: GMR, photonic crystal, and wire-grid
polarizer
Done!
Short Course on Finite-Difference Frequency-Domain
5/23/2012
2
Background
Background: Numerical Methods
Golden Rule #1
1. All numbers should equal 1
(1.234567) + (0.0123456) = Lost two digits of accuracy!!
Why?
Solution: NORMALIZE EVERYTHING!!!
0
0
E E
c

0
0
H H

c
=

or
0
x k x ' =
0
y k y ' =
0
z k z ' =
0
0
1 m

' =
Short Course on Finite-Difference Frequency-Domain
5/23/2012
3
Background: Numerical Methods
Golden Rule #2
2. Never perform calculations
1. Golden Rule #1.
2. Finite floating point precision introduces round-off errors.
Why?
Solution: MINIMIZE NUMBER OF COMPUTATIONS!!!
1. Take problems as far analytically as possible.
2. Avoid unnecessary computations.
( )
2 2
2
exp
R x y
R
g R
o
= +
| |
=
|
\ .
( )
2 2
2
2
exp
r x y
r
g r
o
= +
| |
=
|
\ .
Short Course on Finite-Difference Frequency-Domain
Background: Numerical Methods
Golden Rule #3
3. Follow rules 1 and 2
1. Improves accuracy by reducing numerical error.
2. Enables codes to model larger and more intensive problems.
Why?
Solution
1. Normalize all parameters.
2. Minimize computations.
Short Course on Finite-Difference Frequency-Domain
5/23/2012
4
Background: Numerical Methods
Benefits and drawbacks
Frequency-Domain Time-Domain
Semi-Analytical Fully Numerical
Fourier-Space Real-Space
+ wideband simulations
+ scales near linearly
+ active & nonlinear devices
+ easily locates resonances
- longitudinal periodicity
- sharp resonances
- memory requirements
- oblique incidence
+ resolves sharp resonances
+ handles oblique incidence
+ longitudinal periodicity
+ can be very fast
- scales at best NlogN
- can miss sharp resonances
- active & nonlinear devices
+ better convergence
+ scales better than SA
+ complex device geometry
- memory requirements
- long uniform sections
+ very fast & efficient
+ layered devices
+ less memory
- convergence issues
- scales poorly
- complex device geometry
+ high index contrast
+ metals
+ resolving fine details
+ field visualization
- slow for low index contrast + moderate index contrast
+ periodic problems
+ very fast and efficient
- field visualization
- formulation difficult
- resolving fine details
Unstructured Grid Structured Grid
+ easy to implement
+ rectangular structures
+ easy for divergence free
- less efficient
- curved surfaces
+ most efficient
+ handles larger structures
+ conforms to curved surfaces
- difficult to implement
- spurious solutions
Short Course on Finite-Difference Frequency-Domain
Background: Numerical Methods
Finite-Difference Frequency-Domain
H j E
E j H
ec
e
V =
V =


Maxwells Equations Matrix Equation
| || | | | A x b =
Numerical Solution
| | | | | |
1
x A b

=
Fields fit to a
discrete grid
Short Course on Finite-Difference Frequency-Domain
5/23/2012
5
Background: Electromagnetics
Gausss law in differential form
Electric fields diverge from positive charges
and converge on negative charges.
- +
v
D V- =

y
x z
D
D D
D
x y z
c
c c
V- = + +
c c c

If there are no charges,

electric fields must form loops.
Short Course on Finite-Difference Frequency-Domain
Background: Electromagnetics
No magnetic charge
Magnetic fields always form loops.
0 B V- =

y
x z
B
B B
B
x y z
c
c c
V- = + +
c c c

Short Course on Finite-Difference Frequency-Domain

5/23/2012
6
Background: Electromagnetics
Consequence of Zero Divergence
The divergence theorems force the electric and magnetic fields to be
perpendicular to the propagation direction of a plane wave.
k D

( )
0
0
jk r
D
de
d
-
V- =
V- =
V-

no charges
0
0
jk d
k d
- =
- =


k

k B

( )
0
0
jk r
B
be
b
-
V- =
V- =
V-

no charges
0
0
jk b
k b
- =
- =


k

Short Course on Finite-Difference Frequency-Domain

Background: Electromagnetics
Amperes law in differential form
D
H J
t
c
V = +
c

y y
x x z z
x y z
H H
H H H H
H a a a
y z z x x y
c c | | | | c c c c | |
V = + +
| | |
c c c c c c
\ .
\ . \ .

Circulating magnetic fields induce currents and/or time varying electric fields.
Currents and/or time varying electric fields induce circulating magnetic fields.
Short Course on Finite-Difference Frequency-Domain
5/23/2012
7
Background: Electromagnetics
Faradays law in differential form
B
E
t
c
V =
c

y y
x x z z
x y z
E E
E E E E
E a a a
y z z x x y
c c | | | | c c c c | |
V = + +
| | |
c c c c c c
\ .
\ . \ .

Circulating electric fields induce time varying magnetic fields.

Time varying magnetic fields induce circulating electric fields.
Short Course on Finite-Difference Frequency-Domain
Background: Electromagnetics
Consequences of curl equations
The curl equations predict electromagnetic waves.
H E k

The curl equations force the electric and magnetic field components of a plane
wave to be perpendicular.
k

Electric Field
Magnetic Field
Electric Field
Magnetic Field
Short Course on Finite-Difference Frequency-Domain
5/23/2012
8
Background: Electromagnetics
Maxwells equations
Divergence Equations
0
v
B
D
V- =
V- =

D
H J
t
B
E
t
c
V = +
c
c
V =
c

Curl Equations
Constitutive Relations
( ) ( ) ( )
( ) ( ) ( )
D t t E t
B t t H t
c

= -
= -


What produces fields
How fields interact
with materials
means convolution -
Short Course on Finite-Difference Frequency-Domain
Background: Electromagnetics
Simplifying Maxwells equations
0
0
B
D
V- =
V- =

H D t
E B t
V = c c
V = c c


( ) ( ) ( )
( ) ( ) ( )
D t t E t
B t t H t
c

= -
= -


1. Assume no charges or current sources: 0
v
=
0 J =

0
0
B
D
V- =
V- =

H j D
E j B
e
e
V =
V =


D E
B H
c

=
=


2. Transform Maxwells equations to frequency-domain:
( )
( )
0
0
H
E

c
V- =
V- =

H j E
E j H
ec
e
V =
V =


3. Substitute constitutive relations into Maxwells equations:
Convolution becomes
multiplication
Note: It is helpful to retain and and not replace
with refractive index n.
Short Course on Finite-Difference Frequency-Domain
5/23/2012
9
Background: Electromagnetics
Physical boundary conditions
Fields tangential to the interface
are continuous across it.
1,T
E
2,T
E
1,T
H
2,T
H
1 1
and c
2 2
and c
Fields normal to the interface
are discontinuous across it.
1 1,N
E c
1 1,N
H
2 2,N
E c
2 2,N
H
Note: normal components of D and B are continuous across an interface.
These are more complicated
boundary condition that have
numerical consequences.
Short Course on Finite-Difference Frequency-Domain
Background: Electromagnetics
Sign Convention for Waves
Forward Propagation Along +x
SIGN CONVENTION #1
ikx
e
+
Refractive Index
N n ik = +
( )

0 0
0 0
oscillatory Decaying
term in exponential

ik n i x ik Nx ikx
ik nx k x
x
e e e
e e
k
k
+

+
= =
=
Forward Propagation Along +x
SIGN CONVENTION #2
ikx
e

Refractive Index
N n ik =
( )

0 0
0 0
oscillatory Decaying
term exponential

ik n i x ik Nx ikx
ik nx k x
x
e e e
e e
k
k


+
= =
=
U
s
e
d

H
e
r
e
Short Course on Finite-Difference Frequency-Domain
5/23/2012
10
Background: Linear Algebra
Matrices represent sets of equations
11 12 13 14 1
21 22 23 24 2
31 32 33 34 3
41 42 43 44 4
a w a x a y a z b
a w a x a y a z b
a w a x a y a z b
a w a x a y a z b
+ + + =
+ + + =
+ + + =
+ + + =
11 12 13 14 1
21 22 23 24 2
31 32 33 34 3
41 42 43 44 4
a a a a b w
a a a a b x
a a a a b y
a a a a b z
( ( (
( ( (
( ( (
=
( ( (
( ( (

A set of linear algebraic equations can be written in matrix form.
Short Course on Finite-Difference Frequency-Domain
Background: Linear Algebra
Interpretation of matrices
11 12 13 1
a x a y a z b + + =
21 22 23 2
a x a y a z b + + =
31 32 33 3
a x a y a z b + + =
11 12 13 1
21 22 23 2
31 32 33 3
a a a x b
a a a y b
a a a z b
( ( (
( ( (
=
( ( (
( ( (

11 12 13
21 22 23
31 32 33
a a a
a a a
a a a
(
(
(
(

Equation for x
Equation for y
Equation for z
EQUATION FOR RELATION TO
11 12 13
21 22 23
31 32 33
a a a
a a a
a a a
(
(
(
(

Short Course on Finite-Difference Frequency-Domain
5/23/2012
11
Background: Linear Algebra
Matrices have a compact notation
= Ax b
11 12 13 14
21 22 23 24
31 32 33 34
41 42 43 44
a a a a
a a a a
a a a a
a a a a
(
(
(
=
(
(

A
w
x
y
z
(
(
(
=
(
(

x
1
2
3
4
b
b
b
b
(
(
(
=
(
(

b
Matrices and vectors can be represented and treated as single variables.
| || | | | A x b =
or
square matrix
column
vector
column
vector
11 12 13 14 1
21 22 23 24 2
31 32 33 34 3
41 42 43 44 4
a a a a b w
a a a a b x
a a a a b y
a a a a b z
( ( (
( ( (
( ( (
=
( ( (
( ( (

Short Course on Finite-Difference Frequency-Domain
Background: Linear Algebra
Matrices require a special algebra
=
+ = +
AB BA
A B B A
Commutative Laws
( ) ( )
( ) ( )
=
+ + = + +
AB C A BC
A B C A B C
Associative Laws
( )
( ) ( ) ( )
o o o
o o o
+ = +
= =
A B A B
AB A B A B
Multiplication with a Scalar
( )
( )
+ = +
+ = +
A B C AB AC
A B C AC BC
Distributive Laws
11 1
1
?
n
m mn
a a
a a
o
o o
o
o o
+ =
+ + (
(
+ =
(
( + +

A
I A

Addition with a Scalar

= AB BA
Short Course on Finite-Difference Frequency-Domain
5/23/2012
12
Background: Linear Algebra
Special matrices
Zero Matrix
0 0
0 0
(
(
=
(
(

0

Identity Matrix
1
1
0
0
(
(
=
(
(

I
= = 0 A A 0 0
= = I A A I A
+ = + = 0 A A 0 A
= A A 0
Short Course on Finite-Difference Frequency-Domain
Background: Finite-Differences
What is a finite-difference approximation?
1.5 2 1
df f f
dx x

~
A
1
f
2
f
df
dx
x A
second-order accurate
first-order derivative
Short Course on Finite-Difference Frequency-Domain
5/23/2012
13
Background: Finite-Differences
Types of finite-difference approximations
Backward
difference
1.5 2 1
df f f
dx x

~
A
Central
difference
1 2 1
df f f
dx x

~
A
Forward
difference
2 2 1
df f f
dx x

~
A
Short Course on Finite-Difference Frequency-Domain
Background: Finite-Differences
Generalized finite-difference
n
n
i
i i
d f
a
x
f
d
~

Short Course on Finite-Difference Frequency-Domain

5/23/2012
14
Background:
For more information
Electromagnetics
Matthew Sadiku, Elements of Electromagnetics, Saunders, 2000
Constantine Balanis, Advanced Engineering Electromagnetics, Wiley, 1989
Linear Algebra
Howard Anton, Elementary Linear Algebra 8
th
Ed., Wiley, New Jersey, 2000
Gilbert Strang, Linear Algebra and Its Applications 4
th
Ed., Thomson, California,
2006
http://web.mit.edu/18.06/www/Video/video-fall-99.html
Short Course on Finite-Difference Frequency-Domain
Finite-Difference
Frequency-Domain
5/23/2012
15
FDFD Method: Yee Grid
What is a discrete grid?
A grid is
constructed by
dividing space
into discrete cells
Example
physical
(continuous)
field profile
Field is known
only at discrete
points
Representation of
what is actually
stored in memory
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Yee Grid
What is a grid unit cell?
y A
x A
A field component is
assigned to a specific
point within the grid
unit cell.
Whole Grid
A Single Unit Cell
Short Course on Finite-Difference Frequency-Domain
5/23/2012
16
FDFD Method: Yee Grid
Unit cells of Yee grids
Field components are in physically different locations
Field components may reside in different materials even if they are in the
same unit cell
Field components will be out of phase
x
y
z
x
E
y
E
z
E
x
H
y
H
z
H
3D Yee Grid 2D Yee Grids 1D Yee Grid
z
E
x
H
y
H
x
y
x
y
z
H
y
E
x
E
Ez Mode
Hz Mode
z
x
E
y
H
y
E
x
H
E
y
Mode
E
x
Mode
z
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Yee Grid
Another interpretation of the Yee Grid
Short Course on Finite-Difference Frequency-Domain
5/23/2012
17
FDFD Method: Yee Grid
Reasons for using a Yee grid
( )
0 E c V- =

( )
0 H V- =

1. Divergence-free
3. Elegant arrangement
to approximate curl equations
2. Physical boundary
conditions are
naturally satisfied
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Yee Grid
Extended grids
y
x
i
j
222 Grid
44 Grid (Ez Mode)
Short Course on Finite-Difference Frequency-Domain
5/23/2012
18
FDFD Method: Maxwell Matrix
Normalize the magnetic field
E j H e V =

H j E ec V =

Standard Maxwells Curl Equations
Normalized Magnetic Field
377
E
n
H
~

0
0
H j H

c
=

Normalized Maxwells Equations

0 r
E k H V =

0 r
H k E c V =

0 0 0
k e c =
Note:
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Maxwell Matrix
Expand Maxwells equations
0 r
E k H V =

0 r
H k E c V =

( )
( )
( )
0
0
0
y
z
xx x xy y xz z
x z
yx x yy y yz z
y
x
zx x zy y zz z
H
H
k E E E
y z
H H
k E E E
z x
H
H
k E E E
x y
c c c
c c c
c c c
c
c
= + +
c c
c c
= + +
c c
c
c
= + +
c c

( )
( )
( )
0
0
0
y
z
xx x xy y xz z
x z
yx x yy y yz z
y
x
zx x zy y zz z
E
E
k H H H
y z
E E
k H H H
z x
E
E
k H H H
x y



c
c
= + +
c c
c c
= + +
c c
c
c
= + +
c c



Short Course on Finite-Difference Frequency-Domain
5/23/2012
19
FDFD Method: Maxwell Matrix
Normalize the grid
Normalized Maxwells Equations
Normalized Grid
0
x k x ' =
0
y k y ' =
0
z k z ' =
( )
( )
( )
( )
( )
0
0
0
0
0
0
y z
xx x xy y xz z
x z
yx x yy y yz z
y x
zx x zy y zz z
y z
xx x xy y xz z
x z
yx x yy y yz z
y x
zx x zy y zz
E E
k H H H
y z
E E
k H H H
z x
E E
k H H H
x y
H H
k E E E
y z
H H
k E E E
z x
H H
k E E E
x y



c c c
c c c
c c c
c c
= + +
c c
c c
= + +
c c
c c
= + +
c c
c c
= + +
c c
c c
= + +
c c
c c
= + +
c c






( ) z
y z
xx x xy y xz z
x z
yx x yy y yz z
y x
zx x zy y zz z
y z
xx x xy y xz z
x z
yx x yy y yz z
y x
zx x zy y zz
E E
H H H
y z
E E
H H H
z x
E E
H H H
x y
H H
E E E
y z
H H
E E E
z x
H H
E E E
x y



c c c
c c c
c c c
c c
= + +
' ' c c
c c
= + +
' ' c c
c c
= + +
' ' c c
c c
= + +
' ' c c
c c
= + +
' ' c c
c c
= + +
' ' c c






z
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Maxwell Matrix
Assume diagonal tensor functions
y
z
xx x
x z
yy y
y
x
zz z
E
E
H
y z
E E
H
z x
E
E
H
x y

c
c
=
' ' c c
c c
=
' ' c c
c
c
=
' ' c c

y
z
xx x
x z
yy y
y
x
zz z
H
H
E
y z
H H
E
z x
H
H
E
x y
c
c
c
c
c
=
' ' c c
c c
=
' ' c c
c
c
=
' ' c c

Short Course on Finite-Difference Frequency-Domain
5/23/2012
20
FDFD Method: Maxwell Matrix
Finite-difference equation for H
x
y
z
xx x
E
E
H
y z

c
c
=
' ' c c

z
x
y x
E
, , i j k
y
E
, , i j k
z
E
, , i j k
x
H
y
H
z
H
, 1, i j k
z
E
+
, , 1 i j k
y
E
+
, , 1 , ,
, 1, , ,
, , , ,
i j k i j k
i j k i j k
y y
z
i j k
x
j k
x
z
i
x
H
E E
y
E
z
E

+
+

=
' ' A A

Short Course on Finite-Difference Frequency-Domain
x z
yy y
E E
H
z x

c c
=
' ' c c

FDFD Method: Maxwell Matrix
Finite-difference equation for H
y
z
x
y
y
E
, , i j k
z
E
x
H
, , i j k
y
H
z
H
, , i j k
x
E
1, , i j k
z
E
+
, , 1 , , 1, , , ,
, , , ,
i j k i j k i j k
i j
i j k
x x z z
k i k
y y
j
y
z x
E E E E
H
+ +

=
' ' A A

, , 1 i j k
x
E
+
Short Course on Finite-Difference Frequency-Domain
5/23/2012
21
y
x
zz z
E
E
H
x y

c
c
=
' ' c c

FDFD Method: Maxwell Matrix
Finite-difference equation for H
z
z
x
y
z
E
x
H
y
H
, , i j k
z
H
, , i j k
x
E
1, , i j k
y
E
+
1, , , ,
, 1, , ,
, , , ,
i j k i j k
i j k
i j
i j k
y y
x x
k i k
z z
j
z
x y
E E
E E
H
+
+

=
' ' A A

, 1, i j k
x
E
+
, , i j k
y
E
Short Course on Finite-Difference Frequency-Domain
y
z
xx x
H
H
E
y z
c
c
c
=
' ' c c

FDFD Method: Maxwell Matrix

Finite-difference equation for E
x
, , , ,
,
1
, 1,
, ,
,
,
,
i j k i j k
i j
i j k i j k
y y
z
k
x
j
xx
z
i k
H H
E
H H
y z
c

=
' ' A A


z
x
y
, , i j k
x
E
y
E
z
E
x
H , , i j k
y
H
, , i j k
z
H
, , 1 i j k
y
H

, 1, i j k
z
H

Short Course on Finite-Difference Frequency-Domain
5/23/2012
22
x z
yy y
H H
E
z x
c
c c
=
' ' c c

FDFD Method: Maxwell Matrix
Finite-difference equation for E
y
z
x
y
x
E
y
E
z
E
y
H
, , i j k
x
H
, , i j k
z
H
, , , , 1 , , 1,
,
,
, , ,
i j k i j k i j k i j k
x x z z
i j i
y
k
y
j k
y
H H H H
z x
E c


=
' ' A A

, , 1 i j k
x
H

1, , i j k
z
H

Short Course on Finite-Difference Frequency-Domain
y
x
zz z
H
H
E
x y
c
c
c
=
' ' c c

FDFD Method: Maxwell Matrix
Finite-difference equation for E
z
z
x
y
x
E
y
E
, , i j k
z
E
, , i j k
y
H
, , i j k
x
H
z
H
, , 1, ,
, , , 1
,
,
, , ,
i j k i j k
i j k i j k
y y i j
x
z
i j
x
zz
k k
E
H H
H H
x y
c

=
' ' A A


, 1, i j k
x
H

1, , i j k
y
H

Short Course on Finite-Difference Frequency-Domain
5/23/2012
23
FDFD Method: Maxwell Matrix
Extended 2D Yee grid (Ez Mode)
1
2
,
, 1,
y
y y
i j
i j i j
x x
H
H H

c

~
c A
1
2
,
, , 1
x
x x
i j
i j i j
H
H
y
H
y

c

~
c A
1
2
,
1, ,
z
z z
i j
i j i j
E
x x
E E
+
+
c

~
c A
1
2
,
, 1 ,
z
z z
i j
i j i j
y
E
y
E
E
+
+
c

~
c A
y
x
i
j
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Maxwell Matrix
Summary of FD approximations
x
y
z
y
z
x z
y
xx
yy
zz
xx
yy
zz
y
z
x z
y
x
x
x
y
z
E
E
E E
E
E
E
E
E
y z
z x
x y
y z
H
H
H
H
H
H H
H
H
z x
x y

c
c
c
c
c
=
' ' c c
c c
=
' ' c c
c
c
=
' ' c c
c
c
=
' ' c c
c c
=
' ' c c
c
c
=
' ' c c

, , 1
, ,
, ,
,
, ,
,
, ,
, 1, , ,
, , 1 , , 1, , , ,
1, , , ,
, 1
,
,
,
, ,
, , ,
, ,
1
,
i j k
x
i j k
y
i j k
z
i j k i
i j k i j k
i j k
i j k
xx
i j k
yy
i j k
y y
z z
i j k i j k i j k i j k
x x z z
i j k i j k
i j k i j k
y y
x x i j k
j
z
k
z z
z
y
E E
y z
E E
E E E E
H
H
H
H
H
z x
x
E E
E
y
H
y
E

+
+
+ +
+

=
' ' A A

=
' ' A A

=
' ' A A

' A

, , , , 1
, , , , 1
, ,
, ,
, ,
, , 1, ,
, , 1, ,
, ,
, ,
, ,
,
, 1
,
,
i j k
x
i j k
y
i
i j k i j k
y
i j k i j k i j k i j k
i j k
xx
i j x x z z
i j k i j k
i j k i j k
k
yy
i j k y
z
k
z
x
z
j y
x
H
H H H H
H H
z
z x
x y
H
E
E
H
E
c
c
c

=
' A

=
' ' A A

=
' ' A A

Short Course on Finite-Difference Frequency-Domain
5/23/2012
24
FDFD Method: Maxwell Matrix
Fields are put into column vectors
2-D Systems
1
E
5
E
9
E
13
E
2
E
6
E
10
E
14
E
3
E
7
E
11
E
15
E
4
E
8
E
12
E
16
E
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
(
(
(
(
(
(
(
(
(
(
(
(
=
(
(
(
(
(
(
(
(
(
(
(
(

E
1
2
3
4
5
E
E
E
E
E
(
(
(
( =
(
(
(

E
1-D Systems
1
E
2
E
3
E 4
E 5
E
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Maxwell Matrix
Construction of field column vectors
1
E
5
E
9
E
13
E
2
E
6
E
10
E
14
E
3
E
7
E
11
E
15
E
4
E
8
E
12
E
16
E
1
E
2
E
3
E
4
E
5
E
6
E
7
E
8
E
9
E
10
E
11
E
12
E
13
E
14
E
15
E
16
E
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
E
(
(
(
(
(
(
(
(
(
(
(
(
=
(
(
(
(
(
(
(
(
(
(
(
(

E
MATLAB reshape command
E = E(:);
E = reshape(E,Nx,Ny);
Short Course on Finite-Difference Frequency-Domain
5/23/2012
25
FDFD Method: Maxwell Matrix
Point-by-point multiplication
, r i i
E c
1
E
2
E
3
E 4
E 5
E
1 1 1 1
2 2 2 2
3 3 3 3
4 4 4 4
5 5 5 5
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
r r
r r
r r r
r r
r r
E E
E E
E E
E E
E E
c c
c c
c c
c c
c c
( ( (
( ( (
( ( (
( ( ( = =
( ( (
( ( (
( ( (

E
1 r
c
2 r
c
3 r
c
4 r
c
5 r
c
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Maxwell Matrix
Point-by-point multiplication
, r i i
E c
1
E
2
E
3
E 4
E 5
E
1 1 r
E c
1 1 1 1
2 2 2 2
3 3 3 3
4 4 4 4
5 5 5 5
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
r r
r r
r r r
r r
r r
E E
E E
E E
E E
E E
c c
c c
c c
c c
c c
( ( (
( ( (
( ( (
( ( ( = =
( ( (
( ( (
( ( (

E
1 r
c
2 r
c
3 r
c
4 r
c
5 r
c
Short Course on Finite-Difference Frequency-Domain
5/23/2012
26
FDFD Method: Maxwell Matrix
Point-by-point multiplication
, r i i
E c
1
E
2
E
3
E 4
E 5
E
2 2 r
E c
1 1 r
E c
1 1 1 1
2 2 2 2
3 3 3 3
4 4 4 4
5 5 5 5
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
r r
r r
r r r
r r
r r
E E
E E
E E
E E
E E
c c
c c
c c
c c
c c
( ( (
( ( (
( ( (
( ( ( = =
( ( (
( ( (
( ( (

E
1 r
c
2 r
c
3 r
c
4 r
c
5 r
c
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Maxwell Matrix
Point-by-point multiplication
, r i i
E c
1
E
2
E
3
E 4
E 5
E
2 2 r
E c
1 1 r
E c
3 3 r
E c
1 1 1 1
2 2 2 2
3 3 3 3
4 4 4 4
5 5 5 5
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
r r
r r
r r r
r r
r r
E E
E E
E E
E E
E E
c c
c c
c c
c c
c c
( ( (
( ( (
( ( (
( ( ( = =
( ( (
( ( (
( ( (

E
1 r
c
2 r
c
3 r
c
4 r
c
5 r
c
Short Course on Finite-Difference Frequency-Domain
5/23/2012
27
FDFD Method: Maxwell Matrix
Point-by-point multiplication
, r i i
E c
1
E
2
E
3
E 4
E 5
E
2 2 r
E c
1 1 r
E c
3 3 r
E c
4 4 r
E c
1 1 1 1
2 2 2 2
3 3 3 3
4 4 4 4
5 5 5 5
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
r r
r r
r r r
r r
r r
E E
E E
E E
E E
E E
c c
c c
c c
c c
c c
( ( (
( ( (
( ( (
( ( ( = =
( ( (
( ( (
( ( (

E
1 r
c
2 r
c
3 r
c
4 r
c
5 r
c
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Maxwell Matrix
Point-by-point multiplication
, r i i
E c
1
E
2
E
3
E 4
E 5
E
2 2 r
E c
1 1 r
E c
3 3 r
E c
4 4 r
E c
5 5 r
E c
1 1 1 1
2 2 2 2
3 3 3 3
4 4 4 4
5 5 5 5
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
0 0 0 0
r r
r r
r r r
r r
r r
E E
E E
E E
E E
E E
c c
c c
c c
c c
c c
( ( (
( ( (
( ( (
( ( ( = =
( ( (
( ( (
( ( (

E
1 r
c
2 r
c
3 r
c
4 r
c
5 r
c
Short Course on Finite-Difference Frequency-Domain
5/23/2012
28
FDFD Method: Maxwell Matrix
Derivative operators for Electric fields
1
1
2
i i
i
E E E
x x
+
+
c
~
c A
1
E
2
E
3
E 4
E 5
E
1 1.5
2 2.5
3 3.5
4 4.5
5 5.5
1 1 0 0 0
0 1 1 0 0
1
0 0 1 1 0
0 0 0 1 1
0 0 0 0 1
x
x
E
x x
x
x
E E
E E
E E
x
E E
E E
c
c
c
c
c
c
c
c
c
c
( ( (
( ( (

( ( (
( ( ( = =
A
( ( (

( ( (
( ( (

DE
x
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Maxwell Matrix
Derivative operators for Electric fields
1
1
2
i i
i
E E E
x x
+
+
c
~
c A
1
E
2
E
3
E 4
E 5
E
2 1
E E
x

A
1 1.5
2 2.5
3 3.5
4 4.5
5 5.5
1 1 0 0 0
0 1 1 0 0
1
0 0 1 1 0
0 0 0 1 1
0 0 0 0 1
x
x
E
x x
x
x
E E
E E
E E
x
E E
E E
c
c
c
c
c
c
c
c
c
c
( ( (
( ( (

( ( (
( ( ( = =
A
( ( (

( ( (
( ( (

DE
x
Short Course on Finite-Difference Frequency-Domain
5/23/2012
29
FDFD Method: Maxwell Matrix
Derivative operators for Electric fields
1
1
2
i i
i
E E E
x x
+
+
c
~
c A
1
E
2
E
3
E 4
E 5
E
3 2
E E
x

A
2 1
E E
x

A
1 1.5
2 2.5
3 3.5
4 4.5
5 5.5
1 1 0 0 0
0 1 1 0 0
1
0 0 1 1 0
0 0 0 1 1
0 0 0 0 1
x
x
E
x x
x
x
E E
E E
E E
x
E E
E E
c
c
c
c
c
c
c
c
c
c
( ( (
( ( (

( ( (
( ( ( = =
A
( ( (

( ( (
( ( (

DE
x
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Maxwell Matrix
Derivative operators for Electric fields
1
1
2
i i
i
E E E
x x
+
+
c
~
c A
1
E
2
E
3
E 4
E 5
E
3 2
E E
x

A
2 1
E E
x

A
4 3
E E
x

A
1 1.5
2 2.5
3 3.5
4 4.5
5 5.5
1 1 0 0 0
0 1 1 0 0
1
0 0 1 1 0
0 0 0 1 1
0 0 0 0 1
x
x
E
x x
x
x
E E
E E
E E
x
E E
E E
c
c
c
c
c
c
c
c
c
c
( ( (
( ( (

( ( (
( ( ( = =
A
( ( (

( ( (
( ( (

DE
x
Short Course on Finite-Difference Frequency-Domain
5/23/2012
30
FDFD Method: Maxwell Matrix
Derivative operators for Electric fields
1
1
2
i i
i
E E E
x x
+
+
c
~
c A
1
E
2
E
3
E 4
E 5
E
3 2
E E
x

A
2 1
E E
x

A
4 3
E E
x

A
5 4
E E
x

A
1 1.5
2 2.5
3 3.5
4 4.5
5 5.5
1 1 0 0 0
0 1 1 0 0
1
0 0 1 1 0
0 0 0 1 1
0 0 0 0 1
x
x
E
x x
x
x
E E
E E
E E
x
E E
E E
c
c
c
c
c
c
c
c
c
c
( ( (
( ( (

( ( (
( ( ( = =
A
( ( (

( ( (
( ( (

DE
x
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Maxwell Matrix
Derivative operators for Electric fields
1
1
2
i i
i
E E E
x x
+
+
c
~
c A
1
E
2
E
3
E 4
E 5
E
3 2
E E
x

A
2 1
E E
x

A
4 3
E E
x

A
5 4
E E
x

A
5 6
E E
x

A
1 1.5
2 2.5
3 3.5
4 4.5
5 5.5
1 1 0 0 0
0 1 1 0 0
1
0 0 1 1 0
0 0 0 1 1
0 0 0 0 1
x
x
E
x x
x
x
E E
E E
E E
x
E E
E E
c
c
c
c
c
c
c
c
c
c
( ( (
( ( (

( ( (
( ( ( = =
A
( ( (

( ( (
( ( (

DE
x
Short Course on Finite-Difference Frequency-Domain
5/23/2012
31
FDFD Method: Maxwell Matrix
Derivative operators for Magnetic fields
0.5 1
1.5 2
2.5 3
3.5 4
4.5 5
1 0 0 0 0
1 1 0 0 0
1
0 1 1 0 0
0 0 1 1 0
0 0 0 1 1
x
x
H
x
x
x
x
H H
H H
H H
x
H H
H H
c
c
c
c
c
c
c
c
c
c
( (
(
( (
(

( (
(
( (
( = =
A ( (
(

( (
(
( (
(

D H





1
H

2
H

3
H

4
H

5
H

x
1
1
2
i i
i
H H H
x x

c
~
c A

Short Course on Finite-Difference Frequency-Domain
FDFD Method: Maxwell Matrix
Derivative operators for Magnetic fields
0.5 1
1.5 2
2.5 3
3.5 4
4.5 5
1 0 0 0 0
1 1 0 0 0
1
0 1 1 0 0
0 0 1 1 0
0 0 0 1 1
x
x
H
x
x
x
x
H H
H H
H H
x
H H
H H
c
c
c
c
c
c
c
c
c
c
( (
(
( (
(

( (
(
( (
( = =
A ( (
(

( (
(
( (
(

D H





1
H

2
H

3
H

4
H

5
H

x
1
1
2
i i
i
H H H
x x

c
~
c A

0 1
H H
x

A

Short Course on Finite-Difference Frequency-Domain
5/23/2012
32
FDFD Method: Maxwell Matrix
Derivative operators for Magnetic fields
0.5 1
1.5 2
2.5 3
3.5 4
4.5 5
1 0 0 0 0
1 1 0 0 0
1
0 1 1 0 0
0 0 1 1 0
0 0 0 1 1
x
x
H
x
x
x
x
H H
H H
H H
x
H H
H H
c
c
c
c
c
c
c
c
c
c
( (
(
( (
(

( (
(
( (
( = =
A ( (
(

( (
(
( (
(

D H





1
H

2
H

3
H

4
H

5
H

x
1
1
2
i i
i
H H H
x x

c
~
c A

2 1
H H
x

A

0 1
H H
x

A

Short Course on Finite-Difference Frequency-Domain
FDFD Method: Maxwell Matrix
Derivative operators for Magnetic fields
0.5 1
1.5 2
2.5 3
3.5 4
4.5 5
1 0 0 0 0
1 1 0 0 0
1
0 1 1 0 0
0 0 1 1 0
0 0 0 1 1
x
x
H
x
x
x
x
H H
H H
H H
x
H H
H H
c
c
c
c
c
c
c
c
c
c
( (
(
( (
(

( (
(
( (
( = =
A ( (
(

( (
(
( (
(

D H





1
H

2
H

3
H

4
H

5
H

x
1
1
2
i i
i
H H H
x x

c
~
c A

2 1
H H
x

A

0 1
H H
x

A

3 2
H H
x

A

Short Course on Finite-Difference Frequency-Domain
5/23/2012
33
FDFD Method: Maxwell Matrix
Derivative operators for Magnetic fields
0.5 1
1.5 2
2.5 3
3.5 4
4.5 5
1 0 0 0 0
1 1 0 0 0
1
0 1 1 0 0
0 0 1 1 0
0 0 0 1 1
x
x
H
x
x
x
x
H H
H H
H H
x
H H
H H
c
c
c
c
c
c
c
c
c
c
( (
(
( (
(

( (
(
( (
( = =
A ( (
(

( (
(
( (
(

D H





1
H

2
H

3
H

4
H

5
H

x
1
1
2
i i
i
H H H
x x

c
~
c A

2 1
H H
x

A

0 1
H H
x

A

3 2
H H
x

A

4 3
H H
x

A

Short Course on Finite-Difference Frequency-Domain
FDFD Method: Maxwell Matrix
Derivative operators for Magnetic fields
1
1
2
i i
i
H H H
x x

c
~
c A

2 1
H H
x

A

0 1
H H
x

A

3 2
H H
x

A

4 3
H H
x

A

5 4
H H
x

A

0.5 1
1.5 2
2.5 3
3.5 4
4.5 5
1 0 0 0 0
1 1 0 0 0
1
0 1 1 0 0
0 0 1 1 0
0 0 0 1 1
x
x
H
x
x
x
x
H H
H H
H H
x
H H
H H
c
c
c
c
c
c
c
c
c
c
( (
(
( (
(

( (
(
( (
( = =
A ( (
(

( (
(
( (
(

D H





1
H

2
H

3
H

4
H

5
H

x
Short Course on Finite-Difference Frequency-Domain
5/23/2012
34
FDFD Method: Numerical BCs
Simplest boundary conditions
1 1
1
1 1
2
1 1
3
1 1
4
1
5
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0 0
x x
x x
E
x x x
x x
x
E
E
E
E
E
A A
A A
A A
A A
A
+ ( (
( (
+
( (
( ( = +
( (
+
( (
( (

DE
Dirichlet Boundary Conditions
6
Assume 0 E =
Periodic Boundary Conditions
6 1
Assume E E =
1
E
2
E
3
E 4
E 5
E
6
E
1
E
2
E
3
E 4
E 5
E 1
E
1 1
1
1 1
2
1 1
3
1 1
4
1
5
1
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
x x
x x
E
x x x
x x
x x
E
E
E
E
E
A
A A
A A
A A
A A
A
+ ( (
( (
+
( (
( ( = +
( (
+
( (
( (

+

DE
x
x
5 1
E E
x

A
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Numerical BCs
Pseudo-periodic boundary conditions
6 1
Assume
x x
jk
E E e
A
=
1
E
2
E
3
E 4
E 5
E 6
E
1 1
1
1 1
2
1 1
3
1 1
4
1
5
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
jk
x x
x x
x x
E
x
e
x
x x
x x
x
E
E
E
E
E
A
A
A A
A A
A A
A A
A
+ ( (
( (
+
( (
( ( = +
( (
+
( (
( (

+

DE
x
Short Course on Finite-Difference Frequency-Domain
5/23/2012
35
FDFD Method: Maxwell Matrix
Derivative operators on a 4x4 grid
1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 1 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 1 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0
0 0 0 0 0 0 0 0
0
0
0 0 1 1 0
1 E
x
x
+
+
+

+
+
+

+
+
+
=
A
D
0 0 0
0 0 0 0 0 0 0 0 0 0 0 1 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1
0 0 0 0 0 0 0 0 0 0 0 1
0
0 0 0 0
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
+
(
(
+
(
+ (
(

1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0
0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0
0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0
0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0
0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0
0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0
0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0
0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0
0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0
0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0
0 0 0 0 0 0 0 0 0 0 1 0
1 E
y
y
+
+
+
+
+
+
+
+
+
+

=
A
D
0 0 1 0
0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1
0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
+
(
( +
(

(
(

(
(
(

1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 1 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 1 0 0 0 0 0 0 0
0 0 0 0 0 0 0
0
0 1 1 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 1 1 0 0
0
0
1 H
x
x
+
+
+
+
+
+
+
+
+
+
+
=
A
D
0 0
0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 1 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1
0
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
( +
(
+
(
(
+
(
+ (
(
+
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0
1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0
0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0
0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0
0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0
0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0
0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0
0 0 0 0 0 0 1 0 0 0 1 0 0 0 0
1 H
y
y
+
+
+
+
+
+
+
+
+
+
+
=
A
D
0
0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0
0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0
0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0
0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0
0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
( +
(
+
(
(
+
(
+ (
(
+
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Maxwell Matrix
Finally in matrix form!!!
, , 1 , , , 1, , ,
, , 1 , , 1, , , ,
1, , ,
, ,
, ,
,
,
, , 1,
,
, ,
, ,
, ,
,
i j k i j k i j k i j k
y y z z
i j k i j k i j k
i j
i j
i j k
x
i j k
y
i j k
k
x x z z
i j k i j k
k
xx
i j k
i j k i j k
y y x
yy
i j k
zz z
x
y z
z x
H
E E E E
E
x y
H
H
E E E
E E
E E

+ +
+ +
+ +

=
' ' A A

=
' ' A A

=
' ' A A

z y
x z
y x
E E
y z xx
E E
z x yy
E E
x y
y
zz
x
z
=
=
=
E E
E E
E E
D D
D D
D D
H
H
H

, , , , 1 , , , 1,
, , , , 1 , , 1, ,
,
, ,
, ,
, 1, , , , , 1,
, ,
,
, , ,
,
,
i j k i j k i j k i j k
y y z z
i j k i j k i j k i j k
x x z z
i j k
i j k
xx
i j k
yy
i j k i j k i j
i j k
zz
k
i j k
x
i j k
y y x
y
i j k
z
x
H H H H
H
y z
z x
H H H
H H
y
H
E
H
x
E
E
c
c
c




=
' ' A A

=
' ' A A

=
' ' A A




H H
y z xx
H H
z x yy
H H
x y
z y
x
x
z
y z x zz
y
=
=
=
H H
H H
E D D
D D E
E D D H H

xx
yy
x
y
z y x z
z
z
y
x z
y z
z x
x y
H
H
E E
E E
E E H

c c
=
' ' c c
c c
=
' ' c c
c c
=
' ' c c

xx
yy
zz
z y
x z
y x
x
y
z
y z
z
E
E
E
H H
H
x
y
H H
x
H
c
c
c
c c
=
' ' c c
c c
=
' ' c c
c c
=
' ' c c



Short Course on Finite-Difference Frequency-Domain
5/23/2012
36
FDFD Method: Maxwell Matrix
Summary
E E
y z z y xx x
E E
z x x z yy y
E E
x y y x zz z
H H
y z z y xx x
H H
z x x z yy y
H H
x y y x zz z
=
=
=
=
=
=
D E D E H
D E D E H
D E D E H
D H D H E
D H D H E
D H D H E

E j H
H j E
e
ec
V =
V =


0
0
r
r
E k H
H k E

c
V =
V =

No charges
y z
xx x
x z
yy y
y x
zz z
y z
xx x
x z
yy y
y x
zz z
E E
H
y z
E E
H
z x
E E
H
x y
H H
E
y z
H H
E
z x
H H
E
x y

c
c
c
c c
=
' ' c c
c c
=
' ' c c
c c
=
' ' c c
c c
=
' ' c c
c c
=
' ' c c
c c
=
' ' c c

Normalized H
Normalized Grid Finite-Difference Approximation
, , 1 , , , 1, , ,
, , , ,
, , 1 , , 1, , , ,
, , , ,
1, , , , , 1, , ,
, , , ,
, , , 1,
i j k i j k i j k i j k
y y i j k i j k z z
xx x
i j k i j k i j k i j k
i j k i j k x x z z
yy y
i j k i j k i j k i j k
y y i j k i j k x x
zz z
i j k i j k
y z z
E E E E
H
y z
E E E E
H
z x
E E E E
H
x y
H H H
y

+ +
+ +
+ +

=
' ' A A

=
' ' A A

=
' ' A A

' A

, , , , 1
, , , ,
, , , , 1 , , 1, ,
, , , ,
, , 1, , , , , 1,
, , , ,
i j k i j k
y i j k i j k
xx x
i j k i j k i j k i j k
i j k i j k x x z z
yy y
i j k i j k i j k i j k
y y i j k i j k x x
zz z
H
E
z
H H H H
E
z x
H H H H
E
x y
c
c
c

=
' A

=
' ' A A

=
' ' A A

Matrix Form
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Maxwell Matrix
Hints for other formulations
E E
y z z y xx x
E E
z x x z yy y
E E
x y y x zz z
H H
y z z y xx x
H H
z x x z yy y
H H
x y y x zz z
=
=
=
=
=
=
D E D E H
D E D E H
D E D E H
D H D H E
D H D H E
D H D H E

Fully Numerical Methods BPM, Waveguides, & Semi-Analytical Methods
E
y z y xx x
E
x x z yy y
E E
x y y x zz z
H
y z y xx x
H
x x z yy y
H H
x y y x zz z
d
dz
d
dz
d
dz
d
dz
=
=
=
=
=
=
D E E H
E D E H
D E D E H
D H H E
H D H E
D H D H E

Short Course on Finite-Difference Frequency-Domain
5/23/2012
37
FDFD Method: Maxwell Matrix
Block Matrix Form
E E
y z z y xx x
E E
z x x z yy y
E E
x y y x zz z
H H
y z z y xx x
H H
z x x z yy y
H H
x y y x zz z
=
=
=
=
=
=
D E D E H
D E D E H
D E D E H
D H D H E
D H D H E
D H D H E

Fully Numerical Methods
E E
z y x xx x
E E
z x y yy y
E E
y x z zz z
( ( ( (
( (
( (
=
( (
( (
( (
( (


0 D D E 0 0 H
D 0 D E 0 0 H
D D 0 E 0 0 H

H H
z y x xx x
H H
z x y yy y
H H
y x z zz z
( ( ( (
( (
( (
=
( (
( (
( (
( (


0 D D H 0 0 E
D 0 D H 0 0 E
D D 0 H 0 0 E

E
= C E H

H
= C H E

E H ' V =

H E c ' V =

Short Course on Finite-Difference Frequency-Domain

FDFD Method: Formulation
3D FDFD method
Block Matrix Form
E
H
=
=
C E H
C H E

x x
y y
z z
xx xx
yy yy
zz zz
H H
z y
H H H E
z x
H H
y x
( (
( (
= =
( (
( (

( (
( (
= =
( (
( (

(
(
= =
(
(

H E
H H E E
H E
0 0 0 0
0 0 0 0
0 0 0 0
0 D D 0 D
C D 0 D C
D D 0

E E
z y
E E
z x
E E
y x
(
(

(
(

D
D 0 D
D D 0
Matrix Wave Equations
( )
( )
1
1
E H
H E

=
=
C C H 0
C C E 0

= AE 0

For 3D analysis, A is
usually too big to solve
by simple means.
For information on 3D analysis, see
R. C. Rumpf, A. Tal, S. M. Kuebler, Rigorous electromagnetic analysis of volumetrically complex
media using the slice absorption method, J. Opt. Soc. Am. A 24, 3123-3134 (2007).
Short Course on Finite-Difference Frequency-Domain
5/23/2012
38
FDFD Method: Formulation
Reduction to two dimensions
H H
x y y x zz z
E
y z xx x
E
x z yy y
=
=
=
D H D H E
D E H
D E H

For problems uniform along

the z-direction,
E H
z z
= = D D 0
Maxwells equation decouple into two distinct modes:
Ez Mode Hz Mode
E E
x y y x zz z
H
y z xx x
H
x z yy y
=
=
=
D E D E H
D H E
D H E

z
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Formulation
2D matrix wave equations
H H
x y y x zz z
E
y z xx x
E
x z yy y
=
=
=
D H D H E
D E H
D E H

Ez Mode Hz Mode
E E
x y y x zz z
H
y z xx x
H
x z yy y
=
=
=
D E D E H
D H E
D H E

1
1
E
x xx y z
E
y yy x z

=
=
H D E
H D E

1
1
H
x xx y z
H
y yy x z

=
=
E D H
E D H

( ) ( )
1 1 H E H E
x yy x z y xx y z zz z

= D D E D D E E
( ) ( )
1 1 E H E H
x yy x z y xx y z zz z

= D D H D D H H

1 1
E z
H E H E
E x yy x y xx y zz

=
= + +
A E 0
A D D D D
1 1
H z
E H E H
H x yy x y xx y zz

=
= + +
A H 0
A D D D D

AE = DHX/URyy*DEX + DHY/URxx*DEY + ERzz; AH = DEX/ERyy*DHX + DEY/ERxx*DHY + URzz;

Short Course on Finite-Difference Frequency-Domain
5/23/2012
39
FDFD Method: Formulation
These cannot yet be solved
1

= = Ax b x A b
General solution procedure
Solution of matrix wave equation
1

E z z E

= = = A E 0 E A 0 0 trivial solution!!
A source must be incorporated.
1

E z z E

= = A E f E A f
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Boundary Conds
Perfectly matched layer (1 of 2)
PML
PML
P
M
L
P
M
L
Problem
Space
1
y
s =
1
x
s =
1
x y z
s s s = = =
1
y
s =
1
x
s =
Perfectly Matched Layer (PML)
Short Course on Finite-Difference Frequency-Domain
5/23/2012
40
FDFD Method: Boundary Conds
Perfectly matched layer (2 of 2)
( )
( )
0
0
r
r
E k s H
H k s E

c
V =
V =

0 0
0 0
0 0
y z
x
x z
y
x y
z
s s
s
s s
s
s
s s
s
(
(
(
(
( =
(
(
(
(

( ) ( ) ( )
( ) ( ) ( )
( ) ( ) ( )
0
0
0
0
0
0
1
1
1
x x x
y y y
z z z
s x a x x
jk
s y a y y
jk
s z a z z
jk
q
o
q
o
q
o
(
= +
(

(
= +
(

(
= +
(

( ) ( )
( ) ( )
( ) ( )
max
max
max
1
1
1
p
x x
p
y y
p
z z
a x a x L
a y a y L
a z a z L
= +
= +
= +
( )
( )
( )
2
max
2
max
2
max
sin
2
sin
2
sin
2
x
x
y
y
z
z
x
x
L
y
y
L
z
z
L
t
o o
t
o o
t
o o
| |
=
|
\ .
| |
= |
|
\ .
| |
=
|
\ .
Maxwells eqs. with PML
Computing PML Parameters
max
max
0 5
3 5
1
a
p
o
s s
s s
~
Short Course on Finite-Difference Frequency-Domain
FDFD Method: TF/SF Source
Total-field / scattered-field framework
Problem
rows
t
o
t
a
l
-
f
i
e
l
d
s
c
a
t
t
e
r
e
d
-
f
i
e
l
d
Short Course on Finite-Difference Frequency-Domain
5/23/2012
41
FDFD Method: TF/SF Source
Compute source field
( )
1,1
src
,
exp
x y
x y
N N
f
j k k
f
(
(
(
= + =
(

(

f x y
Compute source as it would exist in a completely homogeneous grid.
( )
inc
src
jk r
E r e
-
=

This is NOT the f term in AE=f

Unit amplitude plane wave
Short Course on Finite-Difference Frequency-Domain
FDFD Method: TF/SF Source
Compute SF masking matrix, Q
t
o
t
a
l
-
f
i
e
l
d
s
c
a
t
t
e
r
e
d
-
f
i
e
l
d
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
1 1 1 1 1
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
0 0 0 0 0
1
1
0
0
0
(
(
(
(
=
(
(
(
(

Q

Short Course on Finite-Difference Frequency-Domain
5/23/2012
42
FDFD Method: TF/SF Source
Compute TF/SF source vector
Source isolated to the scattered-field:
Source isolated to the total-field:
Quantity that must be subtracted from TF terms:
but only from SF equations:
Quantity that must be added to SF terms:
but only from TF equations:
Corrected matrix problem:
or:
scat src
= f Qf
( )
tot src
= f I Q f
tot
Af
tot
QAf
scat
Af
( )
scat
I Q Af
( )
tot scat
+ = AE QAf I Q Af 0
( )
src
= AE QA AQ f
( )
src
= AE = f f QA AQ f
Corrects SF Equations
Corrects TF Equations
Short Course on Finite-Difference Frequency-Domain
FDFD Method: TF/SF Source
Example simulation (1 of 2)
Materials Field Q
Materials Field Q
PML
PML
PML
PML
PML
PML
PML
PML
PML
PML
PML
PML
Scattered-Field Scattered-Field
Scattered-Field Scattered-Field
Total-Field Total-Field
Total-Field Total-Field
1
1.0 n =
1
1.0 n =
2
3.0 n =
Short Course on Finite-Difference Frequency-Domain
5/23/2012
43
FDFD Method: TF/SF Source
Example simulation (2 of 2)
Materials Field Q
Materials Field Q
PML
PML
PML
PML
PML
PML
PML
PML
PML
PML
PML
PML
Scattered-Field Scattered-Field
Scattered-Field Scattered-Field
Total-Field Total-Field
Total-Field Total-Field
1
1.0 n =
1
1.0 n =
2
3.0 n =
Short Course on Finite-Difference Frequency-Domain
FDFD Method:
Formulation Summary
1 1 H E H E
E x yy x y xx y zz

= A D D D D
1 1 E H E H
H x yy x y xx y zz

= A D D D D
Matrix Wave Equations
( )
src
= f QA AQ f
Source Vector
FDFD Matrix Problem
E E
A E = f
H H
A H = f

Short Course on Finite-Difference Frequency-Domain

5/23/2012
44
FDFD Method: Solution
Matrix division in MATLAB
1
E = A f
Matrix Division
WARNING: Do not compute inverse of A
MATLAB Implementation
E = A\f;
WARNING: Do not perform
E = inv(A)*f;
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Solution
Iterative algorithms
Good for very large and sparse matrices
Many algorithms exist
Arguably more accurate for large matrices
Usually do not have to explicitly compute matrix A
Popular iterative algorithms for FDFD
Generalized minimum residue (GMRES)
Biconjugate gradient (BCG)
Short Course on Finite-Difference Frequency-Domain
5/23/2012
45
FDFD Method: Solution
Slice absorption method
1 0 1 1 1 2 1 1 0 1 1 1 3 1
2 1 2 2 2 3 2
3 2 3 3 3 4 3 3 1 3 3 3 4 3

' ' ' ' + + = + + =
+ + =
' ' ' ' + + = + + =
a E b E c E f a E b E c E f
a E b E c E f
a E b E c E f a E b E c E f
R. C. Rumpf, A. Tal, S. M. Kuebler, Rigorous electromagnetic analysis of volumetrically complex media using the slice absorption method, J. Opt. Soc.
Am. A 24, 3123-3134 (2007).
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Post Processing
Diffraction by periodic structures
Reflected Power
Transmitted Power
Diffraction Efficiency
( )
inc
DE
m
P
m
P
=
Short Course on Finite-Difference Frequency-Domain
5/23/2012
46
FDFD Method: Post Processing
Compute spatial harmonics
ref
E
trn
E
Step 1: Extract E
ref
and E
trn
Step 2: Remove phase tilt
( ) ( )
ref ref
x
jk x
E x E x e
+
' =
( ) ( )
trn trn
x
jk x
E x E x e
+
' =
Step 3: Compute FFT
( ) ( )
{ } ref ref
FFT S m E x ' =
( ) ( )
{ } trn trn
FFT S m E x ' =
Note: Some FFT algorithms require
that you divide by the number of points
and shift after calculation.
Eref = fftshift(fft(Fref))/Nx;
Etrn = fftshift(fft(Ftrn))/Nx;
SF
TF
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Post Processing
Compute diffraction efficiencies
Step 2: Compute DE
( ) ( )
ref
2
,
ref inc
Re
z m
z
k
R m S m
k
(
=
(

Step 1: Compute wave vector
components
, ,inc
2
x m x
m
k k
t
=
A
( )
*
2
ref 2
, 0 ref , z m x m
k k n k
(
=
(

( )
*
2
trn 2
, 0 trn , z m x m
k k n k
(
=
(

( ) ( )
trn
2
, ref
trn inc
trn
Re
z m
E
z
k
T m S m
k

(
=
(

( ) ( )
trn
2
,
ref
trn inc
trn
Re
z m
H
z
k
T m S m
k
c
c
(
=
(

Note: These equations assume the
source has unit amplitude
,..., 2, 1, 0,1, 2,...,
2 2
x x
N N
m =
Short Course on Finite-Difference Frequency-Domain
5/23/2012
47
FDFD Method: Post Processing
Compute total power
( )
tot
m
R R m =

Step 1: Compute total reflected power

( )
tot
m
T T m =

Step 2: Compute total transmitted power

Step 3: Verify conservation of energy
tot tot
1 R T + =
Note: This conservation condition is only
obeyed when no materials have loss or gain.
Short Course on Finite-Difference Frequency-Domain
FDFD Method: Algorithm
1. Construct FDFD Problem
a. Define your problem
b. Choose a grid
c. Assign materials to the grid
2. Implement PML
a. Compute s
x
, s
y
, and s
z
b. Incorporate into c
r
and
r
3. Construct Matrix Problem
a. Compute derivative matrices
b. Construct diagonal materials
matrices
c. Compute A
d. Compute source
i. compute source field
ii. compute Q
iii. compute source vector f
4. Solve Matrix Problem: E = A
-1
f;
5. Post Process Data
a. Extract E
ref
and E
trn
b. Remove phase tilt
c. FFT the fields
e. Compute wave vector terms
d. Compute diffraction efficiencies
f. Verify conservation of energy
Short Course on Finite-Difference Frequency-Domain
5/23/2012
48
FDFD Method:
For more information
Literature
W. Sun, K. Liu, C. A. Balanis, Analysis of Singly and Doubly Periodic
Absorbers by Frequency-Domain Finite-Difference Method, IEEE
Trans. Ant. and Prop. 44, 798-805 (1996)
S. Wu, E. N. Glytsis, Finite-number-of-periods holographic gratings
with finite-width incident beams: analysis using the finite-difference
frequency-domain method, J. Opt. Soc. Am. A 19, 2018-2029 (2002)
Raymond Rumpfs Ph.D. Dissertation
R. C. Rumpf, Design and optimization of nano-optical elements by
coupling fabrication to optical behavior, Ph.D. dissertation, University
of Central Florida, 2006.
See chapter 3, pp. 6081
http://purl.fcla.edu/fcla/etd/CFE0001159
Short Course on Finite-Difference Frequency-Domain
Implementation
5/23/2012
49
Implementation: 3D2D
Uniform devices
p
e
r
i
o
d
i
c

b
o
u
n
d
a
r
y
p
e
r
i
o
d
i
c

b
o
u
n
d
a
r
y
Absorbing boundary
Absorbing boundary
Short Course on Finite-Difference Frequency-Domain
Implementation: 3D2D
Effective index method
1,eff
n
2,eff
n
1,eff
n
2,eff
n
Effective indices are best computed
by modeling the vertical cross section
as a slab waveguide.
A simple average index can also
produce good results.
Short Course on Finite-Difference Frequency-Domain
5/23/2012
50
Implementation: Grid & Materials
(1) Choose initial grid resolution
Must resolve the minimum wavelength
( ) ( )
0
min max ,
10
n x y
N
N

(

A s >
x
A
Must resolve the minimum structural dimension
( ) min
1
d d
d
d
N
N
A s >
Initial grid resolution is the smallest number
computed above
| | min ,
x y d
A = A = A A
y
A
Short Course on Finite-Difference Frequency-Domain
Implementation: Grid & Materials
(2) Snap grid to critical dimensions
Decide what dimensions along each axis are critical
Compute how many grid cells comprise d
c
, and round UP
( )
( )
ceil
ceil
x x x
y y y
M d
M d
= A
= A
Adjust grid resolution to fit this dimension in grid EXACTLY
Typically this is a lattice constant or grating period along x
Typically this is a film thickness along y
and
x y
d d
x x x
y y y
d M
d M
A =
A =
i
n
i
t
i
a
l

g
r
i
d
c
r
i
t
i
c
a
l

d
i
m
e
n
s
i
o
n
a
d
j
u
s
t
e
d

g
r
i
d
c
r
i
t
i
c
a
l

d
i
m
e
n
s
i
o
n
Short Course on Finite-Difference Frequency-Domain
5/23/2012
51
Implementation: Grid & Materials
(3) Compute total grid size
x
A
y
A
y
N
x
N
Dont forget to add cells for PML!
Must often add space between
PML and device.
PML space
2 2
x
x
x
y
y
y
N
N N N
A
=
A
A
= + +
A
Problem Space
Buffer Space
Buffer Space
PML
PML
Note: This is particularly important
when modeling devices with large
evanescent fields.
Short Course on Finite-Difference Frequency-Domain
Implementation: Grid & Materials
(4) Compute 2X grid
1X Grid 2X Grid
Short Course on Finite-Difference Frequency-Domain
5/23/2012
52
Implementation: Grid & Materials
(5) Assign materials
Direct Averaged
Short Course on Finite-Difference Frequency-Domain
Implementation: Grid & Materials
(6) Extract materials onto 1X grid
2X Grid
x
y
z
E
z
H
,
x y
H E
,
y x
H E
zz
c
zz

,
xx yy
c
,
yy xx
c
Field and materials assignments
z
E
x
H
y
H
x
y
Ez Mode
x
y
z
H
y
E
x
E
Hz Mode
I II
III IV
1X Grids
I II
III IV
Short Course on Finite-Difference Frequency-Domain
5/23/2012
53
Implementation: Grid & Materials
(7) Oh yeah, metals!
Perfect Electric Conductors
10000
r
c =
1 1
0 0 0 1 0 0
m
M M
E f
E
E f
( ( (
( ( (
( ( (
( ( ( =
( ( (
( ( (
( ( (



or
Include Tangential Fields at Boundary (TM modes!)
0
m
E =
x
y
z
H
y
E
x
E
Hz Mode
Bad placement of metals Good placement of metals
Short Course on Finite-Difference Frequency-Domain
Implementation: Testing
Guided-mode resonance filters
Both a diffraction grating
and a waveguide
Incident
Reflected
Transmitted
1
2
134 nm
314 nm
1.0
1.52
2.0
2.1
0.5
L
L
T
n
n
n
n
f
=
A =
=
=
=
=
=
2
n
L
n
H
n
1
n
Guided-mode resonance filters are excellent devices for benchmarking and testing numerical
codes because they are so sensitive to their geometry, refractive indices, and dispersion.
Hz Ez
S. Tibuleac R. Magnusson, Reflection and transmission guided-mode resonance filters, J. Opt. Soc. Am. A 14, 1617-1626 (1997)
Short Course on Finite-Difference Frequency-Domain
5/23/2012
54
Implementation: Testing
Convergence
Grid Resolution
A
n
s
w
e
r
Probably good enough
Short Course on Finite-Difference Frequency-Domain
Simulation
Examples
5/23/2012
55
Simulation Examples:
Guided-Mode Resonance Filter
Short Course on Finite-Difference Frequency-Domain
Simulation Examples:
Photonic Crystal
Short Course on Finite-Difference Frequency-Domain
5/23/2012
56
Simulation Examples:
Broadband Polarizer
Short Course on Finite-Difference Frequency-Domain