OptiFDTD

Technical Background
Finite-Difference Time-Domain Simulation Design

for Microsoft Windows XP /Vista /7 /8

Table of Contents

1 FDTD Basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.1 2D FDTD Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.1.1 TE waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.1.2 TM waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.2 3D FDTD Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.3 Space and Time Steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.4 OptiFDTD Simulation Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.5 Output data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2 Material Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.1 Constant Dielectrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.2 Lossy Dielectrics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.3 Lorentz Dispersive Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.4 Drude Dispersive Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.5 Lorentz-Drude Dispersive Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.5.1 Lorentz-Drude Model in Frequency Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.5.2 Lorentz-Drude Model in Time Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.5.3 Lorentz-Drude model user interface in OptiFDTD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.5.4 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.6 Nonlinear Materials Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.6.1 Dispersive second-order nonlinear materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.6.2 Dispersive third-order nonlinear material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.6.3 Dispersive Kerr effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.6.4 Dispersive Raman effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.6.5 Nonlinear effects simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.6.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3 OptiFDTD Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.1 PML Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.1.1 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.2 PMC / PEC boundary conditions and plane wave simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.2.1 Perfect Magnetic Conductor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.2.2 Perfect Electric Conductor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3.2.3 Plane waves in symmetric and periodic structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.3 PBC Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4 Light Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.1 Input planes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
4.1.1 CW excitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.1.2 Pulsed excitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
 4.2 A note on Gaussian Beam sizes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
4.3 Point Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
5 Plane Wave Expansion (PWE) method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
5.1 References: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
6 Post-simulation Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
6.1 Discretized Fourier Transform (DFT) and Fast Fourier Transform (FFT) . . . . . . . . . . . . . . . . . . . . 41
6.2 Power calculation and Poynting vector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
6.2.1 x-polarization power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
6.2.2 y-polarization power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
6.2.3 Total power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
6.3 Far Field Transform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
6.3.1 Fraunhofer approximation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
6.3.2 Fresnel-Kirchhoff Diffraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
6.4 Heat Absorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Technical Background

Copyright 2013 Optiwave

All rights reserved.
All OptiFDTD documents, including this one, and the information contained therein, is copyright material.
No part of this document may be reproduced, stored in a retrieval system, or transmitted in any form or by
any means whatsoever, including recording, photocopying, or faxing, without prior written approval of
Optiwave.
Disclaimer
Optiwave makes no representation or warranty with respect to the adequacy of this documentation or the
programs which it describes for any particular purpose or with respect to its adequacy to produce any
particular result. In no event shall Optiwave, its employees, its contractors or the authors of this
documentation, be liable for special, direct, indirect, or consequential damages, losses, costs, charges,
claims, demands, or claim for lost profits, fees, or expenses of any nature or kind.

5
 Technical Background

1 FDTD Basics

OptiFDTD is a powerful, highly integrated, user-friendly software that allows computer aided design and
simulation of advanced passive photonic components.

The OptiFDTD software package is based on the finite-difference time-domain (FDTD) method. The FDTD
method has been established as a powerful engineering tool for integrated and diffractive optics device
simulations. This is due to its unique combination of features, such as the ability to model light propagation,
scattering and diffraction, and reflection and polarization effects. It can also model material anisotropy and
dispersion without any pre-assumption of field behavior such as the slowly varying amplitude approximation.
The method allows for the effective and powerful simulation and analysis of sub-micron devices with very fine
structural details. A sub-micron scale implies a high degree of light confinement and correspondingly, the
large refractive index difference of the materials (mostly semiconductors) to be used in a typical device
design.

1.1 2D FDTD Equations

The FDTD approach is based on a direct numerical solution of the time-dependent Maxwell's curl equations.
When using OptiFDTD in 2D, the photonic device is laid out in the XZ plane. The propagation is along Z. The
Y direction is assumed to be infinite. This assumption removes all the y related derivatives from Maxwell's
equations and splits them into two (TE and TM) independent sets of equations.

The 2D computational domain is shown below. Each mesh point is associated with a specific type of material
and contains information about its properties such as refractive index and dispersion parameters.

1.1.1 TE waves
In the 2D TE case, only Hx, Ey and Hz have nonzero components. In lossless media, Maxwell's equations
take the following form:

where

6
Technical Background
The refractive index of the material is defined by:

Each field is represented by a 2D array: Ey(i,k), Hx(i,k) and Hz(i,k). The indices i and k account for
the number of space steps in the X and Z direction, respectively. In the case of TE, the location of the fields
in the mesh is shown below.

Location of the TE fields components in the computational domain

The TE fields stencil can be explained as follows. The E y field is considered to be the center of the FDTD
space cell. The dashed lines form the FDTD cells. The magnetic fields H x and H z are associated with cell
edges. The locations of the electric fields are associated with integer values of the indices i and k. The Hx
field is associated with integer i and (k + 0.5) indices. The Hz field is associated with (i + 0.5) and
integer k indices. The discrete analog to Maxwell's equations can be defined using the central finite
differences theorem both in space and time:

7
 Technical Background
The superscript n indicates the discrete time step while the indices i and k indicate the discrete space
step along the x and z directions, respectively. The symbols of this equation are defined as follows:

This is the so-called Yee's algorithm applied to the 2D TE case. The central differences used provide
second-order accuracy in both space and time. To conserve accuracy, a space sampling of typically 10 to 20
steps per wavelength is needed. The sampling in time is chosen to ensure numerical stability of the
algorithm. The time step is determined by the Courant limit:

1.1.2 TM waves
In the 2D TM case, only Ex , Hy and Ez have nonzero components. In lossless media, Maxwell's equations
take the following form:

The location of the TM fields in the computational domain follows the same philosophy and is shown below:

8
Technical Background

Location of the TM fields components in the computational domain

Now, the electric field components Ex and Ez are associated with the cell edges, while the magnetic field
Hy is located at the cell center. The TM algorithm can be presented in a way similar to the TE case.

1.2 3D FDTD Equations

In 3D simulations, the space and time steps are defined as follows:

Each field components is represented by a 3D array: Ex(i,j,k), Ey(i,j,k), Ez(i,j,k), Hx(i,j,k),

Hy(i,j,k), Hz(i,j,k). The field components positioning in space is shown below. This arrangement is
called a Yee Cell. E and H field components are interleaved at intervals of 1/2 cell both in space and time.
This way it is possible to solve sequentially all E fields and H fields in the simulation domain using a leapfrog
algorithm.

9
 Technical Background

A 3D Yee cell showing the E and H field components

The time domain vectorial Maxwells equations are given in differential form by:

When the second-order accurate central finite difference theorem is applied to the above Maxwells equation
for time and space derivatives, the 3D-FDTD formulas can be written, using Allen Taflove's notation, as:

10
Technical Background

For H components, and as:

For E components.

11
 Technical Background

1.3 Space and Time Steps

The fundamental constraint of the FDTD method is the step size both for time and space. Space and time
steps relate to the accuracy, numerical dispersion, and the stability of the FDTD method. Many references
and books have discussed these problems. In general, to keep the results as accurate as possible, with a low
numerical dispersion, the mesh size often quoted is "10 cells per wavelength", meaning that the size of
each cell should be 1/10th or less of the shortest wavelength to simulate.

Please note that FDTD is a volumetric computational method, so that if some portion of the computational
space is filled with penetrable material, you must use the wavelength in the material to determine the
maximum cell size.

The following equation can help you determine a suitable mesh size for you simulation:

where nmax is the maximum refractive index value in the computational domain.

Once the cell size is determined, the maximum size for the time step immediately follows the
Courant-Friedrichs-Levy (CFL) condition.

For a 3D FDTD simulation, the CFL condition is:

where v is the speed of the light in the medium.

1.4 OptiFDTD Simulation Procedures

The following is the flow chart for the FDTD simulation in OptiFDTD. It also details the work flow in
OptiFDTD.

12
Technical Background

FDTD Simulation Flow Chart in OptiFDTD

1.5 Output data

The fields propagated by the FDTD algorithm are the time domain fields. At each location of the
computational domain they have a form similar to that given below:

13
 Technical Background

where A is the amplitude of the field at that particular location, G is the wave profile, and is the

corresponding phase. However, the values of A and are not accessible from the time domain field values.

In order to get the full amplitude/phase wave information, we need the stationary complex fields that
correspond to the waveform of the above equation. The complex fields are the source of all useful
information, such as output and reflected powers, overlap integrals with modal fields, etc. Those complex
fields are calculated by a run time Discrete Fourier transform. The final complex fields can be visualized at
specific output Planes located properly in the computational domain.

Output Planes

OptiFDTD uses TF/SF (total field/scattering field) technique for the incident plane. Behind the incident plane,
it is the pure reflection field region, when the observation detectors are placed in this region, the reflection
function can be calculated. When the Observation detectors are placed in the field transmission region, the
transmission function can be calculated.

1.6 References
1. Yee, K. S., "Numerical solution of initial boundary value problems involving Maxwell's equations in
isotropic media," IEEE Transactions on Antennas and Propagation, 302-307, (1966).

2. Chu, S. T., Chaudhuri, S.K., "A finite-difference time-domain method for the design and analysis of
guided-wave optical structures," Journal of Lightwave Technology, 2033-2038, (1989).

3.
14
Technical Background
3. Taflove, A., Hagness, S., "Computational Electrodynamics: The Finite-Difference Time-Domain
Method," Second edition, Arthech House, Boston, (2000).

15
 Technical Background

2 Material Models

One of the main advantages of the FDTD approach is the lack of approximations for the propagating
fieldlight is modeled in its full richness and complexity. The other significant advantage is the great variety
of materials that can be consistently modeled within the FDTD context. In this sub-section we make a brief
review of some of the main material properties that can be handled.

2.1 Constant Dielectrics

Constant dielectric materials are expressed by a complex refractive index value ( ) or relative permittivity
value ( ).

Here, is the refractive index indicating the phase velocity information in the medium, while is called the
extinction coefficient, which indicates the amount of absorption loss when the electromagnetic wave
propagates through the material. Note that the sign of the complex part is a matter of convention, which is
important due to possible confusion between loss and gain.

In OptiFDTD, engineering convention is used for the constant dielectric material definition, which
means that a negative imaginary part will express a loss.

Sellmeier equation can be used for calculating the permittivity of the material at the reference wavelength (
):

Where is the strength, is the damping factor or collision wavelength. is the oscillating wavelength

16
Technical Background

2.2 Lossy Dielectrics

Before proceeding with a more detailed description it should be emphasized that in the time domain all the H
and E fields are real quantities. Thus, accounting for loss is possible only through a non-zero conductivity of
the medium:

where

Here we have assumed that

and

corresponds to time-to-frequency domain Fourier transform. The real and imaginary part of the permitivity
can be expressed through the real and imaginary part of the refractive index:

This makes the refractive index approach and the conductivity approach equivalent.

2.3 Lorentz Dispersive Materials

By Lorentz dispersion materials, we mean materials for which the frequency dependence of the dielectric
permitivity can be described by a sum of multiple resonance Lorentzian functions:

where

17
 Technical Background

In the lossless case, the above equation is directly related to the Sellmeier equation which in the three
resonances can be presented as:

In the lossy case, the Sellmeier equation can be written in a generalized form, accounting for a non-zero
damping coefficient as well as for anisotropy in the dispersion properties:

There are different ways to implement the Lorentz model into the FDTD formalism. Here we consider the
so-called polarization equation approach in the single resonance case. It uses the dielectric susceptibility
function:

and the relation between the polarization and the electric field:

. Taking the Fourier transform of the last equation leads to the following differential equation:

If we substitute the polarization term by the current density:

Then, the differential equation becomes:

Finally, the curl of H term in the differential form of the 3D maxwell's equations can take the form:

18
Technical Background

This can be solved using the FDTD method as described in 3D FDTD Equations.

The Lorentz model is only supported in 2D simulations. Lorentz-Drude materials, that are a
generalization of the Lorentz model, are supported both for 2D and 3D simulations.

2.4 Drude Dispersive Materials

Drude dispersive materials in OptiFDTD are characterized by the following dielectric function:

Where is the permittivity for infinity frequency, is the plasma frequency, and is the collision
frequency or damping factor.

The Drude model is only supported in 2D simulations. Lorentz-Drude materials, that are a
generalization of the Drude model, are supported both for 2D and 3D simulations.

2.5 Lorentz-Drude Dispersive Materials

2.5.1 Lorentz-Drude Model in Frequency Domain

It has been shown [1] that a complex dielectric function for some metals can be expressed in the following
form:

This form separates explicitly the intraband effects (usually referred to as free electron effects) from interband

effect (usually referred to as bound-electron effects). The intraband part of the dielectric function is
described by the well known free-electron or Drude model [2,3]:

19
 Technical Background

The interband part of the dielectric function is described by the following simple semi-quantum model
resembling the Lorentz results for dielectrics:

Where:

The plasma frequency associated with intraband transistions can be written as:

Where G0 is the oscillator strength.

The above Lorentz-Drude Model can be expressed as the more general equation:

where:

If only the term exists, and , then the general equation describes the Drude model.

If only the terms exist, and then the general model becomes the
Lorentz model.

Reference [1] gives the Lorentz-Drude parameters for 11 noble metals.The OptiFDTD Material Library
manual contains parameters compiled by Optiwave that describe noble metals and other dispersive
materials.

20
Technical Background

2.5.2 Lorentz-Drude Model in Time Domain

The Lorentz-Drude model described above is in the frequency domain. However, FDTD is a time domain
method and therefore would be suitable for broadband simulations. We need to transpose the frequency
domain Lorentz-Drude model to time domain so that FDTD can handle the fullwave analysis for the
Lorentz-Drude material. This transformation to time domain is accomplished by using the Polarization field
within Maxwell's equation. The Lorentz-Drude model in time domain can be expressed as:

We can use the relation between the polarization and the electric field:

Taking the Fourier transform of the last equation leads to the following differential equation:

The FDTD algorithm can be derived from the above equation.

2.5.3 Lorentz-Drude model user interface in OptiFDTD

The Lorentz-Drude model user interface is shown below.

21
 Technical Background

Lorentz-Drude material definition data entry interface

In this dialog, you can enter the coefficients for each resonant term as defined in the previous equations. The
following notation is found:

resonance: number of resonance terms

S: Strength of the corresponding resonance term

P (rad/s): Plasma frequency

R (rad/s): Resonant frequency

D (rad/s): Collision frequency (or damping factor )

2.5.4 References
1. Aleksandar D. Rakic, Aleksandra B. Djurisic, et. al., "Optical Properties of Metallic Films for Vertical -
Cavity Optoelectronic Devices". 1998 Optical Society of America, August, Vol. 37, No. 22, Applied
Optics, pp. 5271-5283.

2. M. I. Markovic and A. D. Rakic, " Determination of reflection coefficients of laser light of wavelength
(0.22 um, 200um) from the surface of aluminum using the Lorentz-Drude model", Appl. Opt. 29,
3479-3483 (1990).

3. M. I. Markovic and A. D. Rakic, " Determination of optical properties of aluminum including electron
reradiation in the Lorentz-Drude Model", Opt. Laser technol. 22, 394-398, (1990).

4.
22
Technical Background
4. Ziolkowski, R. W., "Incorporation of microscopic material models into FDTD approach for ultrafast
optical propagation," IEEE Transactions on Antennas and Propagation, 375-391, (1997).

5. Liang, T., Ziolkowski, R. W., "Dispersion effects on grating-assisted output couplers under ultra-fast
pulse excitations", Microwave and Opt. Tech. Lett., 17, 17-23, (1998).

2.6 Nonlinear Materials Models

In general, the nonlinear behavior is due to the dependence of the polarization P(t) on the applied electric
field, E(t). Assuming an isotropic dispersive material, Maxwell's equations are:

D is defined as:

We can then write the linear polarization term PL as a Lorentz oscillator differential equation:

PL is generally referred to as:

where is the linear relative permittivity.

The nonlinear polarization PNL may come from various models. Currently OptiFDTD can handle four kinds of
nonlinearity:

Dispersive second-order nonlinear materials

Dispersive third-order nonlinear materials

Dispersive Kerr effect

Dispersive Raman effect

2.6.1 Dispersive second-order nonlinear materials

In this model, the nonlinear polarization is:

23
 Technical Background

where is the second order isotropic susceptibility.

In order to simulate second order nonlinear effects, you should input two parameters: the linear relative

permittivity and the second order isotropic susceptibility .

2.6.2 Dispersive third-order nonlinear material

Like the second-order nonlinearity, OptiFDTD takes third-order susceptibility to calculate the nonlinear
polarization:

where is the third order isotropic susceptibility.

In order to simulate third order nonlinear effects, you should input two parameters: the linear relative

permittivity and the third order isotropic susceptibility .

2.6.3 Dispersive Kerr effect

If the time scale over which the medium changed is greater than the pulse width, we should take into account
the effects of the finite response time of the medium. Followed by Prof. Richard W. Ziolkowski 's work [1-4],
OptiFDTD treats the nonlinear effect with a finite response time as well as an instantaneous manner by
solving the phenomenological susceptibility equation simultaneously with Maxwell's equation:

where

Therefore, in order to simulate the Kerr effect, you should setup three parameters, , , and .

24
Technical Background

2.6.4 Dispersive Raman effect

Raman model allows another way to simulate the nonlinear phenomenon where the nonlinear susceptibility
was modeled by a second-order derivative equation which is related to the resonant wavelength and the
response time

where

In order to simulate the Raman effect, setup four parameters: , , , and .

2.6.5 Nonlinear effects simulation

To observe nonlinear effects in common used materials, a high-intensity light source is required. You should
pay special attention to the input wave amplitude and/or the power level; each model with different
parameters may need different input power or amplitude. If the input power is too low, you may not observe
the nonlinear phenomenon. If the power is too strong, the nonlinear effect may be over-saturable.

2.6.6 References
1. Ziolkowski, Richard W., Judkins, Justin B., "Full-wave vector Maxwell equation modeling of the
self-focusing of ultrashort optical pulses in a nonlinear Kerr medium exhibiting a finite response time",
J. Opt. Soc. Am. B, 2, 186-198 (1993).

2. Ziolkowski, Richard W., Judkins, Justin B., "Nonlinear finite-difference time-domain modeling of linear
and nonlinear corrugate waveguides", J. Opt. Soc. Am. B, 9, 1565-1575, (1994).

3. Ziolkowski, Richard W., Judkins, Justin B., "Applications of the nonlinear fainted difference
time-domain(NL-FDTD) method to pulse propagation in nonlinear media: self-focusing and
linear-nonlinear interfaces", Radio Science, 901-911, (1993).

4. Ziolkowski, Richard W., "The incorporation of microscopic material models into the FDTD approach for
ultrafast optical pulse simulations", IEEE Trans. On Antenna and Propagation, 3, 375-391, (1997).

5. Joseph, Rose M., Taflove, Allen, "FDTD Maxwell's equations models for nonlinear electrodynamics
and optic", IEEE Trans. On Antenna and Propagation, 3, 364-374, (1997).

6.
25
 Technical Background
6. Goorjian, Peter M., Taflove, Allen, Joseph, Rose M., "Computational modeling of Femtosecond optical
soliton from Maxwell's equation", IEEE Journal of Quantum electronics, 10, 2416-2422, (1992).

7. Joseph, Rose M., Taflove, Allen, " Spatial soliton deflection mechanism indicated by FD-TD Maxwell's
equations modeling", IEEE Photonics Technology Letters, 10, 1251-1254, (1994).

26
Technical Background

3 OptiFDTD Boundary Conditions

Boundary conditions describe the behavior of light crossing the borders of the simulation domain. There can
be 4 types of boundaries:

Perfectly Matched Layers (PML)

Perfect Electric Conductor (PEC)

Perfect Magnetic Conductor (PMC)

Periodic Boundary Condition (PBC)

3.1 PML Boundary Conditions

The basic FDTD algorithm must be modified at the boundaries of the computational window where suitable
numerical absorbing boundary conditions (ABC) are applied. This is one of the most challenging parts of
FDTD simulations. There are several choices for the type of boundary conditions. The Perfectly Matched
Layer (PML) boundary conditions have the best performance. Our FDTD simulator uses the Anisotropic PML,
or so-called Un-split PML (UPML) version. The theory of the UPML is very well explained in some of the
references given here. The UPML boundary conditions are physical rather than numerical because their
implementation is based on a Maxwellian formulation rather than on a mathematical model. Their absorbing
properties are physically equivalent to the properties of an absorbing uni-axial anisotropic medium with the
following permittivity and permeability tensors:

Where:

A plane wave incident on a half space composed of the above uni-axial medium with an interface in the x =
const plane is purely transmitted into it. The reflectionless property is completely independent of the angle of
incidence, polarization and frequency of the incident wave.

The numerical implementation of the UPML in a 2D (X-Z) computational window requires the introduction of
such perfectly matched absorbing layers on all the sides. The corner regions need special attention. In these
regions the tensor must be modifed to:

27
 Technical Background

The minimization of the numerical reflectance of the Anisotropic PML layers requires spatial scaling of the
conductivity profile from zero (at the interface of the PML) to a maximum value at the end of the
computational window:

where L is the thickness of the Anisotropic PML. Typical values for the parameter m are between 2 and 4.

3.1.1 References
1. Brenger, J. P., "A perfectly matched layer for the absorption of electromagnetic waves," Journal of
Computational Physics, 114, 185-200, (1994).

2. Gedney, S. D., "An anisotropic perfectly matched layer absorbing media for the truncation of FDTD
lattices," IEEE Transactions on Antennas and Propagation, 1630-1639, (1996).

3. Taflove, A., "Advances in Computational ElectrodynamicsThe Finite-Difference Time-Domain

Method", Artech House, Boston, Ch. 5, (1998).

3.2 PMC / PEC boundary conditions and plane wave simulations

Perfect Magnetic Conductor (PMC) and Perfect Electric Conductor (PEC) boundary conditions can be used
as "mirrors" for specific field components and use symmetry to reduce the computational domain size. They
are mainly used in the following types of simulations:

Plane wave simulation

Domain reduced simulation for symmetric, periodic, or photonic band gap structures

3.2.1 Perfect Magnetic Conductor

The following graph shows the field set up across a PMC wall (zero thickness) and its image value:

28
Technical Background

Field across a PMC and image area

is the perpendicular electric component at PMC wall, and is the parallel magnetic component. At a
PMC surface, the magnetic field components tangential to the surface vanish.

A PMC boundary acts as a symmetry axis for a structure such as depicted below:

PMC wall in a symmetric waveguide excited by symmetric TE waveguide mode

3.2.2 Perfect Electric Conductor

The following graph shows the field set up across a PEC wall (zero thickness) and its image value:

29
 Technical Background

Field across a PEC and image area

is the perpendicular electric component at PEC wall, and is the parallel magnetic component. At a
PEC surface, the electric field components tangential to the surface vanish.

3.2.3 Plane waves in symmetric and periodic structures

The graph below shows how a plane wave simulation can be realized for a symmetric structure. For a 2D TE
simulation, the boundaries along the propagation should be set to PMC boundary condition to obtain a plane
wave.

Plane wave in TE simulation

For a 2D TM simulation, the the boundaries along the propagation should be set to PEC boundary condition
to obtain a plane wave.

30
Technical Background

Plane wave in TM simulation

For a 3D simulation, the plane wave realization depends on the wave polarization and the boundary condition
setup at different edges of the propagation plane.

If the wave goes in z-direction, and the input wave is in y-direction polarization, then the y plane (x-z)
boundaries should be set to PEC and the x-plane (y-z plane) boundaries set to PMC.

Y-polarization plane wave (z-direction propagation) with boundary conditions

If the wave goes in z-direction, and the input wave is in x-direction polarization, then the y plane (x-z)
boundaries should be set to PMC and x-plane (yz plane) boundaries set to PEC.

31
 Technical Background

X-polarization plane wave (z-direction propagation) with boundary conditions

The plane wave shape can be set by setting the rectangular wave parameters in the Input Wave Properties
dialog box.

3.3 PBC Boundary Conditions

OptiFDTD provides the option to use simplified Periodic Boundary Conditions (PBC). PBC can work with
other boundary conditions such as Anisotropic PML, PMC, and PEC. With PBC, you can generate a simple
plane wave simulation or periodic layout simulations.

The simplified PBC is based on Bloch's Theorem:

For a periodic layout, the wave function is written as the combination of a cell periodic component and a
wave-like component:

The equation above needs to have a predefined value for the k-vector; which is not easily feasible for FDTD
simulations, especially for Pulsed input. Therefore, we simplified this problem by setting the k-vector in one
specific direction as zero, which then leads to the input wave being the axis-propagated wave, and then the
equation becomes:

In OptiFDTD, is the period, is the field component at one edge of the simulation domain,
is the corresponding field value at the boundary of the opposite edge.

The following graph shows the relationship between these values:

32
Technical Background

PBC applied to a 2D periodic mesh

33
 Technical Background

4 Light Sources

Light sources are necessary to inject energy into the FDTD simulation domain. In OptiFDTD you can have 3
main light sources types

Input planes

Point sources

Total-Field / Scattered Field (TF/SF) simulation regions

Input planes are used when light can be injected though a plane in the simulation domain. It propagates in a
particular direction and has an intensity distribution across the incidence plane. Point sources can be used to
simulate point emitters, for example quantum wells or quantum dots. TF/SF regions are useful for simulating
nanoparticle scattering for example. More details about this type of source are given in the User's Reference
Manual.

4.1 Input planes

The FDTD numerical scheme yields the solution of an initial value problem. At the first time step of the
simulation all the fields are set to zero. The algorithm needs the initial field excitation that will be propagated
through the computational domain. The FDTD Total/Reflected Field formulation gives the methodology in
defining a +Z propagating incident field. The computational domain is separated into two sub-regions the
total field region and the reflected field region . The plane separating these regions is called the incident field
(see below).

34
Technical Background

Total/Reflected field formulation

In the Total Field Region, the structures of interest are designed. The interactions between the incident field
and the structures will take place in this region. That is why the Total Field Region contains information for
both the incident and scattered (reflected) waves. In the Reflected Field region the geometry is uniform and
the propagating waves are presented by the fields reflected from the Total Field Region. There are no objects
in this region and the signal will not be reflected back to the total field region.

CW or pulsed excitation can be used. One can consider the incident field as being generated by a flashlight
located on the incidence plane facing the +Z direction. Before starting the simulation the flashlight is turned
off and the field values in the whole computational domain are equal to zero. The flashlight is switched on at t
= 0 and illuminates only the Total Field Region. If the excitation scheme is perfect, there should not be any
light detected by an observer located in the Reflected Field Region, unless there are some obstacles which
would generate the reflections. The incident wave can be generated by specifying the exact field distribution
on the incident plane at each time interval.

4.1.1 CW excitation
In CW excitation, the time dependence of the incident field is a single frequency sinusoidal function. For
example the incident Ey field has the following form:

35
 Technical Background

where A is the field amplitude, is the transverse field distribution at the incident plane location
. The initial phase offset is the phase difference between points in the incidence plane. This offset
can be adjusted to define the direction of the incident field. is the frequency of the input
wave. In the CW case the optical wave propagates until it reaches the stationary state everywhere in the
computational window.

4.1.2 Pulsed excitation

For pulsed excitations the incident field has the form:

where

is the pulse envelope function, is the time offset and is the pulse width parameter.

For pulsed excitation, the time stepping continues until the desired late-time pulse response is observed at
the field points of interest.

4.2 A note on Gaussian Beam sizes

In OptiFDTD, a transverse Gaussian Beam is expressed as

where is the center position and T is the half width.

In general, the Gaussian beam radius is the radius at which the field amplitude and intensity drop to 1/e and
, respectively. The Gaussian beam size or Gaussian beam spot size is the beam diameter which is
two times the beam radius.

Suppose the beam size is noted a. We have

This means that:

36
Technical Background

Please use the above formula to set the half width in OptiFDTD for your desired Gaussian Beam size.

4.3 Point Sources

Point source can also be referred to as a radiation source. The wave radiates away from a single point in
space in a concentric spherical pattern (see below).This input wave is only functional with one field
component at one single point. You can specify which field component is enabled.

A Point source

In the case of CW excitation, the point source is expressed in an update equation form as:

where (i,j,k) is the user-specified position and F is the user specified component, A is the point source
amplitude, T is the time delay and w is the angle frequency

For pulsed excitation, the point source is expressed as:

where is time offset and T is the half width.

37
 Technical Background

Point sources can work with other kinds of sources (such as input planes).

When a Point source is used as the key input. The normalized power calculation will be
disabled in analyzer

38
Technical Background

5 Plane Wave Expansion (PWE) method

The Maxwell equation in a transparent, time-invariant, source free, and non-magnetic medium can be written
in the following form:

where

This equation is sometimes called the Master Equation, and represents an Hermitian eigen-problem, which
would not be applicable if the wave equation were derived in terms of the electric field.

The Bloch theorem states that, due to infinite periodicity, the magnetic field will take the form:

Where:

for all combinations of lattice vectors R . Thus, we end up with the master equation in operator form:

This is the fundamental equation, which needs to be solved. The equation is transformed into a finite problem
by expanding the magnetic field in a finite basis of simple plane waves. Different approaches can be explored
to solve the final discretized problem. The result of solving the discretized problem is the dispersion
relationship between the frequencies of the modes and wave vector k, usually plotted in the form of a band
diagram.

5.1 References:
1. J.D. Joannopoulos, R.D. Meade, and J.N. Winn, "Photonic crystals, Molding the flow of light,
Princeton University Press, 1995.

2. S.G. Johnson, J.D. Joannopoulos, Block-iterative frequency-domain methods for Maxwell's equations
in a planewave basis", Optics Express 8, no.3, p.173-190, 2000.

3.
39
 Technical Background
3. S.Guo, S.Albin, "Simple plane wave implementation for photonic crystal calculations", Optics Express
11, no.2, p.167-175, 2003.

40
Technical Background

6 Post-simulation Data Analysis

OptiFDTD provides comprehensive post-simulation data analysis tools in the Simulator and Analyzer.

6.1 Discretized Fourier Transform (DFT) and Fast Fourier Transform

(FFT)
As a time domain simulation method; FDTD can get all the relevant spectral response with a single
simulation. To get the spectral response, you need to use the frequency DFT or FFT functions in the
observation point analyzer. The spectral response given with DFT is as follows:

where s(n) describes time domain response, N is the number of time steps, and is the angular
frequency.

FFT analysis uses the traditional fast Fourier transform scheme to get a spectral response from the zero
frequency to the cutoff frequency . The frequency domain sampling step is . In
general, the sampling frequency step for FFT is comparable to the wavelength due to the fact that the FDTD
required time step is very small. Therefore, the FFT results may have a larger error than that of the DFT
results, but the FFT method is much faster than DFT.

6.2 Power calculation and Poynting vector

For the z-direction propagation wave. The total power in x-y plane can be divided into two power values:
x-direction polarized z-direction propagation power (Pz-x) and y-direction polarized z-direction propagation
power (Pz-y). The corresponding formulas are:

6.2.1 x-polarization power

6.2.2 y-polarization power

41
 Technical Background

6.2.3 Total power

The z-direction Poynting for a point (i,j,) in x-y plane is:

The Poynting vector is a complex value. In OptiFDTD, only the amplitudes are displayed.

6.3 Far Field Transform

6.3.1 Fraunhofer approximation

Narrow angle far-field transform being used in OptiFDTD is based on the Fraunhofer approximation:

At a large distance d, the far field position can be expressed by the far field angle approximation:

Where the x-directional angle is the angle between the original yz-plane and the shortest straight line

connecting the point and the Y axis, and y-directional angle is the angle between the orginal xz-plane
and the shortest straight line connecting the point and the x axis. The far field angle is also shown below.

42
Technical Background

Far-field angle

Using the angle approximation described above, the far-field formula can be simplified as:

Please note that the above formula assumes that the far-field plane is far away from the near field. OptiFDTD
uses the above equation to calculate the narrow angle far field transform.

6.3.2 Fresnel-Kirchhoff Diffraction

Wide angle far field transform is based on the Fresnel-Kirchhoff diffraction formula.

Where R is the vector from near-field to far-field.

The far-field position can be expressed using the far field angle and the far-field distance d. So in the wide
angle far field transform, the user needs to specify the far-field distance.

43
 Technical Background

6.4 Heat Absorption

It is known that the real electromagnetic power transmitted through a closed surface S into the volume V is
equal to the power loss produced by conduction current resulting in Joule heating plus the power loss
resulting from polarization damping forces.

From the complex Poynting vector theorem, the time-average heat absorption intensity is expressed as:

where is the wavelength-dependant conductivity.

In OptiFDTD, for each cell in a given observation area, the heat absorption intensity for each polarization
then can be expressed as:

Where the subscript letter, x, y and z means the polarization direction, i and j is the position index.

Then the total heat absorption for one observation slice will be the integral of the heat absorption intensity in
the volume

Where du and dv are the mesh size in the observation plane, is the third dimensional (vertical to the
observation area) space step. The total heat absorption in the whole observation slice will be

Heat absorption calculation need the conductivity information from the materials. Permittivity and conductivity
can be expressed in terms of refractive indexes and extinction and vice-versa:

44
Technical Background

So the permittivity and conductivity are:

For Lorentz-Drude dispersive materials,

The conductivity can be expressed as:

45