lecture #18 metals and alternative grid schemesemlab.utep.edu/ee5390fdtd/lecture 18 -- metals and...
TRANSCRIPT
10/19/2016
1
Lecture 18 Slide 1
EE 5303
Electromagnetic Analysis Using Finite‐Difference Time‐Domain
Lecture #18
Metals and Alternative Grid Schemes These notes may contain copyrighted material obtained under fair use rules. Distribution of these materials is strictly prohibited
Lecture Outline
Lecture 18 Slide 2
• Review of Selected Topics
– PML Placement
– Two‐dimensional TF/SF
– Computing reflectance and transmittance
• Incorporating metals into FDTD
• Alternative grids
10/19/2016
2
Lecture 18 Slide 3
Review of PML Placement
Lecture 18 Slide 4
PML Placement
PML
PML
PML
PML
It is best practice to place the PML outside any evanescent field. This is best assessed by visualizing the fields during a simulation.
10/19/2016
3
Lecture 18 Slide 5
Review of Two‐Dimensional Total‐Field/Scattered‐Field
Lecture 18 Slide 6
The Total‐Field/Scattered‐Field Framework
Problem Points!
total‐field
scattered‐field
2D FDTD Grid
srcjsrc 1j
10/19/2016
4
We must subtract the source from to make it look like a scattered‐field quantity.
Lecture 18 Slide 7
Correction to Finite‐Difference Equations at the Problem Cells (1 of 2)
On the scattered‐field side of the TF/SF interface, the finite‐difference equation contains a term from the total‐field side. Due to the staggered nature of the Yee grid, this only occurs in the update equation for a magnetic field. In fact, this only occurs in the computation of the curl of E used in the H field update equations.
src
src
src , 1
, 1
,i j i j
i j zE tx t
z tE
CE
y
This is an equation in the scattered‐field, butis a total‐field quantity.
srcjzE
srcjzE
src
src
srcsrc , 1
, 1
s c, ,r i j
z
i j
zi j tEx
t
t
i j
z tEE E
Cy
src src
s crc sr,sr
1
c
, ,
, 1 1i j i j
i j z zE t ti j
z tx t
E EC
y yE
standard curl equation This is a correction term that can be implemented after calculating the curl to inject a source.
Lecture 18 Slide 8
Correction to Finite‐Difference Equations at the Problem Cells (2 of 2)
On the total‐field side of the TF/SF interface, the finite‐difference equation contains a term from the scattered‐field side. Due to the staggered nature of the Yee grid, this only occurs in the update equation for an D field. In fact, this only occurs in the computation of the curl of H used in the D field update equations.
src srcsrc
s
src
2rc 2 2 2
2
, 1, ,
,
, 1t t t t
t
i j i j i jy yi j xt t tH
i j
z
x t
t
H H HC
x
H
y
This is an equation in the scattered‐field, butis a total‐field quantity.
src 1jxH
We must add the source to to make it look like a total‐field quantity.
src 1jxH
src src src
src 22 2
src
2 2
2
src, 1, 1, s,
,
, rc1tt
t
tt t
i j i j i j
xy y t
i ji
i j t tHz t
x t x t
jHH H
x
HC
y
H
src src
src
2
src src
src 2 2 2 2
2
, 1, , ,
, 1src
1
, 1t
t t t t
t
i j
i j i j i j i jy yi j x xt t t t
x t
Hz t
H H H HC
x y yH
standard update equation This is a correction term that can be implemented after calculating the curl to inject a source.
10/19/2016
5
Lecture 18 Slide 9
Calculation of the Source Functions (EzMode)
We calculate the electric field as
src,src i j
z tE g t
We calculate the magnetic field as
src
2
, 1src
02 2t
ri j
rx t
n ttH
yg
c
Half time step difference
Delay through one half of a grid cell
Amplitude due to Maxwell’s equations
The index i indicates the TF/SF correction is incorporated across the entire row of the grid.
Visualize Fields
Update Ez
Inject TF/SF Source into curl of E
Lecture 18 Slide 10
TF/SF Block Diagram for EzMode
Compute Curl of E
Update H Integrations
Update H Field
Update Dz
Main loop…
Compute Curl of H
Update D Integrations
Inject TF/SF Source into curl of H
Finished!yes
Done?
no
10/19/2016
6
Lecture 18 Slide 11
Review of Computing Reflectance and Transmittance
Lecture 18 Slide 12
Complex Wave Vectors
Purely Real k Purely Imaginary k Complex k
•Uniform amplitude•Oscillations move energy• Considered to be a propagating wave
•Decaying amplitude•Oscillations move energy• Considered to be a propagating wave (not evanescent)
•Decaying amplitude•No oscillations, no flow of energy• Considered to be evanescent
10/19/2016
7
Lecture 18 Slide 13
Evanescent Fields in 2D Simulations
1 2
No critical angle
n n 1 2
1 C
n n
1 2
1 C
n n
Lecture 18 Slide 14
Fields in Periodic Structures
Waves in periodic structures take on the same periodicity as their host.
k
10/19/2016
8
Lecture 18 Slide 15
The Plane Wave Spectrum (1 of 2)
We rearranged terms and saw that a periodic field can also be thought of as an infinite sum of plane waves at different angles. This is the “plane wave spectrum” of a field.
, x yj k m x k m y
m
E x y S m e
Lecture 18 Slide 16
The Plane Wave Spectrum (2 of 2)
, jk m r
m
E x y S m e
,inc
2 20 2
ˆ ˆ
2
x y
x xx
y x
k m k m x k m y
mk m k
k m k n k m
The plane wave spectrum can be calculated as follows
ky is imaginary.
Each wave must be separately phase matched into the medium with refractive index n2.
inck
2n
1n
ky is real. ky is imaginary.
0xk 1xk 2xk 3xk 4xk 5xk 1xk 2xk 3xk 4xk 5xk
10/19/2016
9
Lecture 18 Slide 17
Power Flow From Gratings
1. Power flow is in the direction of the Poynting vector.
2. It is only the z‐component that flows power into and out of the grating.
3. Diffraction efficiency is defined as the fraction of power diffracted into a particular mode.
4. The diffraction efficiencies of the spatial harmonics are
5. Conservation of power requires
2*1 1Re Re
2 2E H k k E
2
1Re
2z
z
Ek
k
,inc
DE z
z
mm
2 2
ref trn,ref ,trn ,refref trn2 2
,inc ,inc ,trninc inc
DE Re DE Rez z r
z z r
S m S mk m k mm m
k kS S
ref trn
1 materials have loss
DE DE 1 materials have no loss
1 materials have gain m m
m m
Lecture 18 Slide 18
Calculating Transmittance and Reflectance
frequency 1frequency 2frequency 3frequency 4
frequency NFREQ
frequency 1frequency 2frequency 3frequency 4
frequency NFREQ
FDTD Simulation
Steady‐State Fields
Spatial Harmonics
Diffraction Efficiencies
Reflectance and
Transmittance
Fourier Transform
t f
FFT
, ,x yi j k m k m
2
ref ,refref 2
,incinc
2
trn ,trn ,reftrn 2
,inc ,trninc
DE Re
DE Re
z
z
z r
z r
S m k mm
kS
S m k mm
kS
ref
trn
DE ,
DE ,x
x
N
N
R f m f
T f m f
10/19/2016
10
Lecture 18 Slide 19
Procedure for FDTD1. Simulate the device using FDTD and calculate the steady‐state field at the reflection and
transmission record planes.
ref trn src, , , and E x f E x f E f
2. Calculate the incident wave vector: ,inc 0 incyk k n
3. Calculate periodic expansion of the transverse wave vector
2 floor 2 , , 1,0,1, floor 2x x x xk m m L m N N 4. Calculate longitudinal wave vector components in the reflected and transmitted regions.
2 22 2,ref 0 ref ,trn 0 trn y x y xk m k n k m k m k n k m
5. Normalize the steady‐state fields to the source
ref ref src trn trn srcˆ ˆ, , , ,E x f E x f E f E x f E x f E f
6. Calculate the complex amplitudes of the spatial harmonics
ref ref trn trnˆ ˆ, FFT , , FFT ,S m f E x f S m f E x f
7. Calculate the diffraction efficiencies of the spatial harmonics
8. Calculate reflectance, transmittance, and conservation of energy.
ref trnDE , DE , m m
R f m f T f m f C f R f T f
frequency
2 2,ref ,trn refref ref trn trn
,inc ,inc trn
DE , , Re DE , , Rey y
y y
k m k mm f S m f m f S m f
k k
Lecture 18 Slide 20
Incorporating Metals into FDTD
10/19/2016
11
Lecture 18 Slide 21
What are Metals?
Metals are materials with very high conductivity and usually very negative dielectric constant. The electric field approaches zero inside a metal.
Material Conductivity (S/m)
Glass 10‐12
Carbon 3×104
Mercury 106
Lead 3×106
Tin 9×106
Iron 1.03×107
Nickel 1.45×107
Aluminum 3.96×107
Gold 4.1×107
Copper 5.76×107
Silver 6.1×107
Increa
sing Conductivity
Methods for Incorporating Metals
• Extreme Dielectric Constant– Easiest because no modification to the code is necessary, but it does not account for loss.
• Perfect Electric Conductor– Requires minimal modification to the code, but does not account for loss.
• – Requires greater modification to the formulation of the update equations. It can account for loss, but cannot account for frequency dependence.
• Lorentz‐Drude Model– Requires a much more complicated formulation and implementation, but it can account for loss and frequency dependence.
Lecture 18 Slide 22
Easier Im
plementation
More Accu
rate Sim
ulatio
n
10/19/2016
12
Lecture 18 Slide 23
Method #1: Extreme Dielectric Constant
Recall the update equations for the electric field:
, , , , , ,
, , ,
1 1 1
i j i j i j i j i j i j
x x y y z zi j i j i jt t t t t t t t t t t txx zzyy
E D E D E D
In metals, the electric field approaches zero. We can force this to happen by choosing dielectric constants that are very large (i.e. 103 or higher).
,,
,,
,,
As , 0
As , 0
As , 0
i ji j
xx x t t
i ji j
yy y t t
i ji j
zz z t t
E
E
E
Lecture 18 Slide 24
Method #2: Perfect Electric Conductor
Recall the update equations for the electric fields:
, , , , , ,,, ,
1 1 1 i j i j i j i j i j i ji ji j i j
x Ex x y Ey y z Ez zt t t t t t t t t t t tE m D E m D E m D
We can force the fields to zero by setting the update coefficients to 0 everywhere there is a metal and 1 everywhere that there is not.
Define the placement of metals using three arrays:
, ,, , , ,PEC PEC PEC
i j ki j k i j k
x y z
We modify the update coefficients as follows:
, , , , , ,
1 1
, , , , , ,
1 1
, , , , , ,
1 1
PEC
PEC
PEC
i j k i j k i j k
Ex x Ex
i j k i j k i j k
Ey y Ey
i j k i j k i j k
Ez z Ez
m m
m m
m m
10/19/2016
13
Lecture 18 Slide 25
Method #3: Conductivity Recall from Lecture 10 that we can incorporate material conductivity into Maxwell’s equations as follows:
H E
E H Et t
We incorporate metals by retaining this conductivity term, deriving new update equations, and revising the FDTD algorithm.
Note, you will end up creating three new materials arrays:
, , xx yy zz
Lecture 18 Slide 26
Method #4: Lorentz‐Drude Model
Recall from Lecture 10 that we can write the constitutive relation as
0D E P
22 2
1 0,
Mm
pm m m
fP E
j
We can implement this in the time‐domain as follows
2 20,m m m m m m p
m m
J t J t P t f E tt
J t P tt
10/19/2016
14
Lecture 18 Slide 27
Placing Metals on a 2D Grid
Ez Mode
xy
zH yExE
Hz Mode
Bad placement of metals Good placement of metals
For the Ezmode, the electric field is always tangential to metal interfaces and few problems arise when modeling metallic structures.
Hz ModeFor the Hzmode, the electric field can be polarized perpendicular to metal interfaces. This is problematic and it is best to place metals with the outermost fields being tangential to the interfaces.
Lecture 18 Slide 28
Alternative Grids
10/19/2016
15
Lecture 18 Slide 29
Drawbacks of Uniform Grids
Uniform grids are the easiest to implement, but do not conform well to arbitrary structures and exhibit high anisotropic dispersion.
Anisotropic Dispersion (see Lecture 10) Staircase Approximation (see Lecture 18)
Lecture 18 Slide 30
Hexagonal Grids
Hexagonal grids are good for minimizing anisotropic dispersion suffered on Cartesian grids. This is very useful when extracting phase information.
See Text, pp. 101‐103.
Yee‐FDTD
Hex‐FDTD
10x
Phase Velocity as a Function of Propagation Angle
57°0° 115° 172° 229° 286° 344°
10/19/2016
16
Lecture 18 Slide 31
Nonuniform Orthogonal Grids (1 of 2)
Nonuniform orthogonal grids are still relatively simple to implement and provide some ability to refine the grid at localized regions.
See Text, pp. 464‐471.
Lecture 18 Slide 32
Nonuniform Orthogonal Grids (2 of 2)
Uniform Grid Simulation• 80×110×16 cells• 140,800 cells
Nonuniform Grid Simulation• 64×76×16 cells• 77,824 cells
Conclusion: Roughly 50% memory and time savings.
10/19/2016
17
Lecture 18 Slide 33
Curvilinear Coordinates
Maxwell’s equations can be transformed from curvilinear coordinates to Cartesian coordinates to conform to curved boundaries of a device.
M. Fusco, “FDTD Algorithm in Curvilinear Coordinates,” IEEE Trans. Ant. and Prop., vol. 38, no. 1, pp. 76‐89, 1990.See Text, pp. 484‐492.
Lecture 18 Slide 34
Structured Nonorthogonal Grids
This is a particularly powerful approach for simulating periodic structures with oblique symmetries.
M. Fusco, “FDTD Algorithm in Curvilinear Coordinates,” IEEE Trans. Ant. and Prop., vol. 38, no. 1, pp. 76‐89, 1990.
10/19/2016
18
Lecture 18 Slide 35
Irregular Nonorthogonal Unstructured Grids
Unstructured grids are more tedious to implement, but can conform to highly complex shapes while maintaining good cell aspect ratios and global uniformity.
ln x
P. Harms, J. Lee, R. Mittra, “A Study of the Nonorthogonal FDTD Method Versus the Conventional FDTD Technique for Computing Resonant Frequencies of Cylindrical Cavities,” IEEE Trans. Microwave Theory and Techniq., vol. 40, no. 4, pp. 741‐746 , 1992.
Comparison of convergence rates
Lecture 18 Slide 36
Bodies of Revolution (Cylindrical Symmetry)
Three‐dimensional devices with cylindrical symmetry can be very efficiently modeled using cylindrical coordinates.
even odd
0
even odd
0
, , , cos , sin
, , , cos , sin
m
m
E e m e m
H h m h m
Devices with cylindrical symmetry have fields that are periodic around their axis. Therefore, the fields can be expanded into a Fourier series in .
Due to a singularity at r=0, update equations for fields on the z axis are derived differently.
See Text, Chapter 12
10/19/2016
19
Lecture 18 Slide 37
Some Devices with Cylindrical Symmetry
Bent Waveguides Dipole AntennasCylindrical Waveguides
Conical Horn Antenna
Focusing Antennas
Diffractive Lenses