INTRODUCTION TO

RF Module
Introduction to the RF Module
© 1998–2021 COMSOL
Protected by patents listed on www.comsol.com/patents, or see Help>About COMSOL Multiphysics on the File
menu in the COMSOL Desktop for a less detailed lists of U.S. Patents that may apply. Patents pending.
This Documentation and the Programs described herein are furnished under the COMSOL Software License
Agreement (www.comsol.com/comsol-license-agreement) and may be used or copied only under the terms of the
license agreement.
COMSOL, the COMSOL logo, COMSOL Multiphysics, COMSOL Desktop, COMSOL Compiler, COMSOL Server,
and LiveLink are either registered trademarks or trademarks of COMSOL AB. All other trademarks are the property
of their respective owners, and COMSOL AB and its subsidiaries and products are not affiliated with, endorsed by,
sponsored by, or supported by those trademark owners. For a list of such trademark owners, see www.comsol.com/
trademarks.
Version: COMSOL 6.0

Contact Information
Visit the Contact COMSOL page at www.comsol.com/contact to submit general inquiries or
search for an address and phone number. You can also visit the Worldwide Sales Offices page at
www.comsol.com/contact/offices for address and contact information.

If you need to contact Support, an online request form is located at the COMSOL Access page at
www.comsol.com/support/case. Other useful links include:

• Support Center: www.comsol.com/support

• Product Download: www.comsol.com/product-download
• Product Updates: www.comsol.com/support/updates
• COMSOL Blog: www.comsol.com/blogs
• Discussion Forum: www.comsol.com/forum
• Events: www.comsol.com/events
• COMSOL Video Gallery: www.comsol.com/videos
• Support Knowledge Base: www.comsol.com/support/knowledgebase

Part number. CM021004

Contents

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
The Use of the RF Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
The RF Module Physics Interfaces . . . . . . . . . . . . . . . . . . . . . . . . 11
Physics Interface Guide by Space Dimension and Study Type . . . 14
Tutorial Model: Impedance Matching of a Lossy Ferrite 3-Port
Circulator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Impedance Matching. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Model Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
The Lossy Ferrite Material Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

|3
4|
Introduction

The RF Module is used by engineers and scientists to understand, predict, and

design electromagnetic wave propagation and resonance effects in high-frequency
applications. Simulations of this kind result in more powerful and efficient
products and engineering methods. It allows its users to quickly and accurately
predict electromagnetic field distributions, transmission, reflection, and power
dissipation in a proposed design. Compared to traditional prototyping, it offers
the benefits of lower cost and the ability to evaluate and predict entities that are
not directly measurable in experiments. It also allows the exploration of operating
conditions that would destroy a real prototype or be hazardous.
This module covers electromagnetic fields and waves in two-dimensional and
three-dimensional spaces along with traditional circuit-based modeling of passive
and active devices. All modeling formulations are based on Maxwell’s equations or
subsets and special cases of these together with material constitutive relations for
propagation in various media. The modeling capabilities are accessed via
predefined physics interfaces, referred to as Radio Frequency (RF) interfaces,
which allow you to set up and solve electromagnetic models. The RF interfaces
cover the modeling of electromagnetic fields and waves in frequency domain, time
domain, eigenfrequency, and mode analysis.
Under the hood, the RF interfaces formulate and solve the differential form of
Maxwell’s equations together with the initial and boundary conditions. The
equations are solved using the finite element method with numerically stable edge
element discretization in combination with state-of-the-art algorithms for
preconditioning and solution of the resulting sparse equation systems. The results
are presented using predefined plots of electric and magnetic fields, S-parameters,
power flow, and dissipation. You can also display your results as plots of
expressions of the physical quantities that you define freely, or as tabulated derived
values obtained from the simulation.
The workflow is straightforward and can be described by the following steps:
define the geometry, select materials, select a suitable RF interface, define
boundary and initial conditions, define the finite element mesh, select a solver, and
visualize the results. All these steps are accessed from the COMSOL Desktop. The
solver selection step is usually carried out automatically using default settings,
which are tuned for each specific RF interface.
The RF Module’s application library describes the physics interfaces and the
different features through tutorial and benchmark examples for the different
formulations. The library includes antennas, ferrite devices, microwave heating
phenomena, passive devices, scattering and RCS analysis, transmission lines and

|5
waveguides in RF and microwave engineering, tutorial models for education, and
benchmark models for verification and validation of the RF interfaces.
This introduction is intended to give you a jump start in your modeling work. It
has examples of the typical use of the RF Module, a list of the physics interfaces
with a short description, and a tutorial model that introduces the modeling
workflow.

The Use of the RF Module

The RF interfaces are used to model electromagnetic fields and waves in high
frequency applications. The latter means that it covers the modeling of devices that
are above about 0.1 electromagnetic wavelength in size. Thus, it may be used to
model microscale devices or human size devices operating at frequencies above
10 MHz.
RF simulations are frequently used to extract S-parameters characterizing the
transmission and reflection of a device. Figure 1 shows the electric field distribution
in a dielectric loaded H-bend waveguide component. A rectangular TE10
waveguide mode is launched into the inport at the near end of the device and is
absorbed by the outport at the far end. The bend region is filled with silica glass.

Figure 1: Electric field distribution in a dielectric loaded H-bend waveguide. From the application library,
H-Bend Waveguide 3D.

6|
The transmission and reflection of the device are expressed quantitatively by the
S-parameters, shown in Figure 2.

Figure 2: The S-parameters, on a dB scale, as a function of the frequency.

The S-parameters are, by themselves, useful performance measures of a device.

They can also be exported as a Touchstone file for further use in system
simulations.
In Figure 3 and Figure 4, a application library entry for the RF Module shows how
a human head absorbs a radiated wave from an antenna held next to an ear. The
temperature is increased by the absorbed radiation.

|7
Figure 3: The local increase in temperature in a human head due to absorption of electromagnetic energy
from an antenna held next to the ear. From the application library entry Absorbed Radiation (SAR) in
the Human Brain.

The SAR (specific absorption rate) value is of specific interest to designers of

mobile telephones and is readily obtained from the simulation.

Figure 4: Log-scale slice plot of the local specific absorption rate (SAR) in a human head.

8|
The RF Module also offers a comprehensive set of features for 2D modeling
including both source driven wave propagation and mode analysis. Figure 5 shows
mode analysis of a step-index profile optical fiber.

Figure 5: The surface plot visualizes the longitudinal component of the electric field in the fiber core. From
the application library entry Step Index Fiber.

Both in 2D and 3D, the analysis of periodic structures is popular. Figure 6 is an

example of a plane wave incidence on a wire grating with a dielectric substrate.

Figure 6: Electric field norm for TE incidence at π/5. From the application library entry Plasmonic Wire
Grating.

|9
It is also possible to perform Body-of-Revolution (BOR) simulations in 2D
axisymmetry. Figure 7 illustrates the modeling of a monoconical antenna with a
coaxial feed.

Figure 7: Axisymmetric model of a monoconical antenna with a coaxial feed. The azimuthal component
of the magnetic field is shown. From the application library entry Conical Antenna.

The RF Module has a vast range of tools to evaluate and export the results, for
example, evaluation of far-field, feed impedance, and scattering matrices
(S-parameters). S-parameters can be exported in the Touchstone file format.
Full-wave electromagnetic field modeling can also be combined with circuit-based
modeling. This is an ideal basis for design, exploration, and optimization. More
complex system models can be created using circuit-based modeling while
maintaining links to full field models for key devices in the circuit allows for design
innovation and optimization on both levels.

10 |
The RF Module Physics Interfaces

The RF Module physics interfaces are based on Maxwell’s equations or subsets

and special cases of these together with material constitutive relations. In the
module, these laws of physics are translated by the RF interfaces to sets of partial
differential equations with corresponding initial and boundary conditions.
The RF interfaces define a number of features. Each feature represents a term or
condition in the underlying equations and may be defined in a geometric entity of
the model, such as a domain, boundary, edge (for 3D components), or point.

Figure 8: The Model Builder (left), and the Settings window for Wave Equation, Electric (right). The
Equation section shows the model equations and the terms added by the Wave Equation, Electric 1 node
to the model equations. The added terms are underlined with a dotted line. The text also explains the
link between the Dielectric node and the values for the relative permittivity.

| 11
Figure 8 uses the Coaxial Waveguide Coupling model from the RF Module
application library to show the Model Builder window and the Settings window
for the selected Wave Equation, Electric 1 feature node. The Wave Equation,
Electric 1 node adds the terms representing Electromagnetic Waves to the model
equations in a selected geometrical domain in the model.
Furthermore, the Wave Equation, Electric 1 feature node may link to the
Materials feature node to obtain physical properties such as relative permittivity —
in this case the relative permittivity of a user-defined dielectric. The properties,
defined by the Dielectric material, can be functions of the modeled physical
quantities, such as temperature. In the same fashion, the Perfect Electric
Conductor 1 feature adds a reflecting boundary condition to truncate the
modeling domain.
Figure 9 shows the Radio Frequency (RF) interfaces as displayed in the Model
Wizard for this module.

Figure 9: The Radio Frequency (RF) physics interfaces as displayed in the Model Wizard.

This module includes RF interfaces ( ) for both frequency-domain and

time-domain modeling. It also includes the Microwave Heating interface that is
found under Heat Transfer>Electromagnetic Heating. Also see Physics Interface
Guide by Space Dimension and Study Type. A brief overview of the RF interfaces
follows.

E LECTROMAGNETIC WAVES , F REQUENCY D OMAIN

The Electromagnetic Waves, Frequency Domain interface ( ) solves a
frequency-domain wave equation for the electric field. The sources can be in the
form of point dipoles, line currents, or incident fields on boundaries or domains.
It is used primarily to model electromagnetic wave propagation in different media
and structures. This physics interface can also be used to solve an eigenvalue
problem for the resonant frequencies and fields of a structure, or to find the
propagating modes of a waveguide or transmission line. Some typical applications
that are simulated with this physics interface are waveguides and transmission
lines, filters and resonators, antennas, and RF connectors and couplers.

12 |
E LECTROMAGNETIC WAVES , T IME E XPLICIT
The Electromagnetic Waves, Time Explicit interface ( ) solves a system of two
first-order partial differential equations (Faraday’s law and Maxwell-Ampère’s law)
for the electric and magnetic fields using the Time Explicit Discontinuous
Galerkin method. The sources can be in the form of volumetric electric or
magnetic currents or electric surface currents or fields on boundaries. It is used
primarily to model electromagnetic wave propagation in linear media. Typical
applications involve the transient propagation of electromagnetic pulses.

E LECTROMAGNETIC WAVES , T RANSIENT

The Electromagnetic Waves, Transient interface ( ) solves a time-domain wave
equation for the electric field. The sources can be in the form of point dipoles, line
currents, or incident fields on boundaries or domains. It is used primarily to model
electromagnetic wave propagation in different media and structures when a
time-domain solution is required — for example, non-sinusoidal waveforms or
nonlinear media. Typical applications involve the propagation of electromagnetic
pulses and the generation of harmonics in nonlinear optical media.

T RANSMISSION L INE
The Transmission Line interface ( ) solves the time-harmonic transmission line
equation for the electric potential. This physics interface is used when solving for
electromagnetic wave propagation along one-dimensional transmission lines and
is available in 1D, 2D and 3D. Eigenfrequency and Frequency Domain study types
are available. The frequency domain study is used for source-driven simulations at
a single frequency or a sequence of frequencies. Typical applications involve the
design of impedance matching elements and networks.

E LECTROMAGNETIC W AVES, A SYMPTOTIC S CATTERING

The Electromagnetic Waves, Asymptotic Scattering interface ( ) is used for
quick studies of the far-field response of a 3D or 2D object to a given background
field. The physics interface sets up a surface electric background field for the
far-field transformation, using the Stratton–Chu formula, performed in the
postprocessing. Use this physics interface in 2D and 3D when approximating the
scattered far-field of an object configured only by a perfect electric conductor
boundary condition.

E LECTROMAGNETIC W AVES, B OUNDARY E LEMENTS

The Electromagnetic Waves, Boundary Elements interface ( ) solves a
frequency-domain wave equation for the electric field. The formulation is based
on the boundary element method (BEM) and requires the availability of a Green’s

| 13
function. Thus, the physics interface solves the vector Helmholtz equation for
piecewise-constant material properties.
The interface is fully multiphysics enabled and can be coupled seamlessly with the
physics interfaces that are based on the finite element method (FEM). This
approach allows modeling in a FEM-BEM framework, using the strength of each
formulation adequately. The BEM-based interface is especially well suited for
radiation and scattering problems.
The advantage of the boundary element method is that only boundaries need to
be meshed and the degrees of freedom (DOFs) solved for are restricted to the
boundaries.

M ICROWAVE H EATING
The Microwave Heating interface ( ) combines the features of the
Electromagnetic Waves, Frequency Domain interface with those of the Heat
Transfer interface. A predefined interaction automatically sets the electromagnetic
losses as sources for the heat equation. This physics interface is based on the
assumption that the electromagnetic cycle time is short compared to the thermal
time scale (adiabatic assumption).

E LECTRICAL C IRCUIT
The Electrical Circuit interface ( ) can be connected to an RF interface. The
lumped voltage and current variables from the circuits are translated into
boundary conditions applied to the distributed field model. Typical applications
include the modeling of transmission lines and antenna feeding.

Physics Interface Guide by Space Dimension and Study Type

The table below lists the physics interfaces available specifically with this module
in addition to the COMSOL Multiphysics basic license.

PHYSICS INTERFACE ICON TAG SPACE AVAILABLE STUDY TYPE

DIMENSION

AC/DC

Electrical Circuit cir Not space stationary; frequency

dependent domain; time dependent;
small signal analysis,
frequency domain;
eigenfrequency

14 |
PHYSICS INTERFACE ICON TAG SPACE AVAILABLE STUDY TYPE
DIMENSION

Fluid-Structure Interaction

Heat Transfer

Electromagnetic Heating

Microwave Heating1 — 3D, 2D, 2D frequency-stationary;

axisymmetric frequency-transient;
frequency-stationary,
one-way electromagnetic
heating; frequency-transient,
one-way electromagnetic
heating
Radio Frequency

Electromagnetic Waves, ewas 3D, 2D frequency domain

Asymptotic Scattering
Electromagnetic Waves, embe 3D, 2D frequency domain
Boundary Elements
Electromagnetic Waves, emw 3D, 2D, 2D adaptive frequency sweep;
Frequency Domain axisymmetric boundary mode analysis;
eigenfrequency; frequency
domain; frequency domain,
modal; frequency domain,
RF adaptive mesh; frequency
domain source sweep;
mode analysis (2D and 2D
axisymmetric models only);
TEM boundary mode
analysis
Electromagnetic Waves, ewte 3D, 2D, 2D time dependent; time
Time Explicit axisymmetric dependent with FFT
Electromagnetic Waves, temw 3D, 2D, 2D eigenfrequency; time
Transient axisymmetric dependent; time dependent,
modal; time dependent with
FFT

| 15
PHYSICS INTERFACE ICON TAG SPACE AVAILABLE STUDY TYPE
DIMENSION

Transmission Line tl 3D, 2D, 1D eigenfrequency; frequency

domain
1
This physics interface is a predefined multiphysics coupling that automatically adds all the
physics interfaces and coupling features required.

16 |
Tutorial Model: Impedance Matching of a Lossy Ferrite
3-Port Circulator

Introduction
A microwave circulator is a nonreciprocal multiport device. It has the property that
a wave incident on port 1 is routed into port 3 yet a wave incident on port 3 is not
routed back into port 1 but is instead routed into port 2, and so on. This property
of a circulator is used to isolate microwave components from each other, for
example, when connecting a transmitter and a receiver to a common antenna. By
connecting the transmitter, receiver, and antenna to different ports of a circulator,
the transmitted power is routed to the antenna whereas any power received by the
antenna goes into the receiver. Circulators typically rely on the use of ferrites, a
special type of highly permeable and low-loss magnetic material that is anisotropic
for a small RF signal when biased by a much larger static magnetic field. In the
example, a three-port circulator is constructed from three rectangular waveguide
sections joining at 120° and with a ferrite post inserted at the center of the joint.

Port 3

Ferrite post

Port 2
Port 1

The post is magnetized by a static H0 bias field along its axis. The bias field is
supplied by external permanent magnets which are not explicitly modeled in this
tutorial.

| 17
Impedance Matching
An important step in the design of any microwave device is to match its input
impedance for a given operating frequency. Impedance matching is equivalent to
minimizing the reflections back to the inport. The parameters that need to be
determined are the size of the ferrite post and the width of the wider waveguide
section surrounding the ferrite. In this tutorial, these are varied in order to
minimize the reflectance. The scattering parameters (S-parameters) used as
measures of the reflectance and transmittance of the circulator are automatically
computed.
The nominal frequency for the design of the device is chosen as 3 GHz. The
circulator can be expected to perform reasonably well in a narrow frequency band
around 3 GHz, and so a frequency range of 2.8–3.2 GHz is studied. It is desired
that the device operates in single mode. Thus a rectangular waveguide cross
section of 6.67 cm by 3.33 cm is selected to set the cutoff frequency for the
fundamental TE10 mode to 2.25 GHz. The cutoff frequencies for the two nearest
higher modes, the TE20 and TE01 modes, are both at 4.5 GHz, leaving a
reasonable safety margin.

Model Definition
One of the rectangular ports is excited by the fundamental TE10 mode. At the
ports, the boundaries are transparent to the TE10 mode. The following equation
applies to the electric field vector E inside the circulator:

∇ × ( μ r ∇ × E ) – k 0  ε r – --------- E = 0
–1 2 jσ
 ωε 0

where μr denotes the relative permeability tensor, ω is the angular frequency, σ is

the conductivity tensor, ε0 is the permittivity of vacuum, εr is the relative
permittivity tensor, and k0 is the free space wave number. In this particular model,
the conductivity is zero everywhere. Losses in the ferrite are introduced as
complex-valued permittivity and permeability tensors. The magnetic permeability
is of key importance as it is the anisotropy of this parameter that is responsible for
the nonreciprocal behavior of the circulator. For simplicity, the rather complicated
material expressions are predefined in a text file that is imported into the model.
The expressions are also included in the next section for reference.

The Lossy Ferrite Material Model

Complete treatises on the theory of magnetic properties of ferrites can be found
in Ref. 1 and Ref. 2. The model assumes that the static magnetic bias field, H0, is
much stronger than the alternating magnetic field of the microwaves, so the
quoted expressions are a linearization for a small-signal analysis around this

18 |
operating point. Under these assumptions, and including losses, the anisotropic
permeability of a ferrite magnetized in the positive z direction is given by:

μ jκ 0
[ μ ] = – jκ μ 0
0 0 μ0

where
κ = – jμ 0 χ xy

μ = μ 0 ( 1 + χ xx )

and the unique elements of the magnetic susceptibility tensor χ are given by:

ω 0 ω m ( ω20 – ω2 ) + ω 0 ω m ω2 α2 αωω m ( ω20 + ω2 ( 1 + α2 ) )

2
- – j -------------------------------------------------------------------------------------
χ xx = ------------------------------------------------------------------------------------- 2
-
( ω20 – ω2 ( 1 + α2 ) ) + 4ω 20 ω2 α2 ( ω 20 – ω2 ( 1 + α2 ) ) + 4ω20 ω2 α2

2ω 0 ω m ω2 α ωω m ( ω20 – ω2 ( 1 + α2 ) )
2
- + j -------------------------------------------------------------------------------------
χ xy = ------------------------------------------------------------------------------------- 2
-
( ω20 – ω2 ( 1 + α2 ) ) + 4ω 20 ω2 α2 ( ω20 – ω2 ( 1 + α2 ) ) + 4ω20 ω2 α2

where
ω 0 = μ 0 γH 0

ω m = μ 0 γM s

μ 0 γ ΔH
α = ------------------
2ω

Here μ0 denotes the permeability of free space; ω is the angular frequency of the
microwave field; ω0 is the precession resonance frequency (Larmor frequency) of
a spinning electron in the applied magnetic bias field, H0; ωm is the electron
Larmor frequency at the saturation magnetization of the ferrite, Ms; and γ is the
gyromagnetic ratio of the electron. For a lossless ferrite (α = 0), the permeability
becomes infinite at ω = ω0. In a lossy ferrite (α ≠ 0), this resonance becomes finite
and is broadened. The loss factor, α, is related to the line width, ΔH, of the
susceptibility curve near the resonance as given by the last expression above. The
material data,
Ms = 5.41·104 A/m, εr = 14.5

| 19
with an effective loss tangent of 2·10-4 and ΔH = 3.18·103 A/m, are taken for
aluminum garnet from Ref. 2. The applied bias field is set to H0 = 7.96·103 A/m.
The electron gyromagnetic ratio taken from Ref. 2 is 1.759·1011 C/kg.

References
1. R.E. Collin, Foundations for Microwave Engineering, 2nd ed., IEEE Press/
Wiley-Interscience, 2000.
2. D.M. Pozar, Microwave Engineering , 3rd ed., John Wiley & Sons Inc, 2004.

Model Wizard

These step-by-step instructions guide you through the design and modeling of the
lossy three-port circulator in 3D. The first part involves the geometric design and
impedance matching at a nominal frequency of 3 GHz. After that, a frequency
sweep is performed to see how well it performs in a frequency band of 400 MHz
centered at 3 GHz. Finally, computation and a Touchstone file export of the
entire S-parameter matrix is performed.
Note: These instructions are for the user interface on Windows but apply, with
minor differences, also to Linux and Mac.
1 To start the software, double-click the COMSOL icon on the
desktop. When the software opens, you can choose to use the
Model Wizard to create a new COMSOL Multiphysics model or
Blank Model to create one manually. For this tutorial, click the
Model Wizard button.
If COMSOL Multiphysics is already open, you can start the
Model Wizard by selecting New from the File menu and then
click Model Wizard .
The Model Wizard guides you through the first steps of setting
up a model. The next window lets you select the dimension of
the modeling space.
2 In the Select Space Dimension window click 3D .
3 In the Select Physics tree, expand Radio Frequency, and
double-click Electromagnetic Waves, Frequency Domain so
that it appears in the Added physics interface list.
Alternatively, select Electromagnetic Waves, Frequency Domain
and click the Add button.

20 |
4 Click Study .
5 In the Studies tree under General Studies, click Frequency Domain .
6 Click Done .

Global Definitions - Parameters and Variables

The geometry is set up using a parameterized approach. This allows you to match
the input impedance to that of the connecting waveguide sections by variation of
two geometric design parameters. These are dimensionless numbers used to scale
selected geometric building blocks.
In this section, two parameters are entered and a set of variables imported from a
file to prepare for drawing the circulator geometry, which is described in the
section Geometry Sequence. Alternatively, a predefined application library file
containing the geometry, parameters, and variables can be imported, as described
in Geometry. If you import the geometry, you only need to review this section for
information.
1 On the Home toolbar click Parameters (or in the Model Builder, right-click
Global Definitions and select Parameters ).
Note: On Linux and Mac, the Home toolbar refers to the specific set of controls
near the top of the Desktop.
2 In the Settings window under Parameters, enter these settings in the table.

The lossy ferrite material model is set up by referring to global variables. For
convenience the definitions are stored in an external text file that is imported
into the model. The external text file also contains comments.

| 21
Note: The location of the text files vary based on the installation. For example, if
the installation is on your hard drive, the file path might be similar to
C:\Program Files\COMSOL\COMSOL60\Multiphysics\applications\.
1 On the Home toolbar click Variables and select Global Variables .
2 Go to the Settings window for Variables. Under Variables click Load from
File .
3 Browse to the application library folder \RF_Module\Ferrimagnetic_Devices
and double-click the file lossy_circulator_3d_parameters.txt. The
variables are imported into the table.

Geometry

In the Global Definitions section, you entered parameters and imported variables
in preparation for drawing the geometry. To learn how to draw the circulator, go
to Geometry Sequence.

22 |
To save time, a predefined model containing the parameters, variables, and
geometry can instead be opened from the Application Libraries window.
1 On the Home toolbar click Windows and select Application Libraries .
2 In the Application Libraries window, under RF Module>Ferrimagnetic Devices
double-click lossy circulator 3d geom to open it.
Discard Untitled.mph, a blank model with parameters and variables created in
the previous section.
Once the geometry is either drawn or imported, you can then experiment with
different dimensions by changing the values of sc_chamfer and sc_ferrite and
re-running the geometry sequence.

Mate rials

The next step is to add material settings to the model. The air that fills most of the
volume is available as a built-in material. The lossy ferrite has material assigned to
it later, and illustrates how external material data can also be entered directly into
the electromagnetic waves model. The walls of the waveguide sections are
modeled as perfect conductors and do not require a material.
1 On the Home toolbar click Add Material .
2 In the Materials tree under Built-In, right-click Air and choose Add to
Component 1 .

3 Click Add Material again to close the Add Material window.

| 23
Ele c tr omag netic Wa ve s, F r eq u en c y Do m a i n

The ferrite enters in the physics interface as a separate, user-defined equation

model, referring to the global variables defined in the section Global Definitions -
Parameters and Variables.

Wave Equation, Electric 2

1 On the Physics toolbar click Domains and choose Wave Equation,
Electric .
A new node called Wave Equation, Electric 2 is added to the Model Builder.
The nodes with a ‘D’ in the upper left corner indicate a default node.
2 To get a view of the interior part of the circulator, click the Wireframe
Rendering button on the Graphics toolbar.
3 Select Domain 2 only.

Domain 2

Note: There are many ways to select geometric entities. When you know the
domain to add, such as in this exercise, you can click the Paste Selection
button located beside the Selection list and enter the information in the
Selection text field. In this example enter 2 in the Paste Selection window. For
more information about selecting geometric entities in the Graphics window, see
the COMSOL Multiphysics Reference Manual.

24 |
4 Go to the Settings window for Wave
Equation, Electric 2. Under Electric
Displacement Field:
- From the Electric displacement
field model list, select Dielectric
loss.
- From the ε′ list, select User
defined. In the associated text
field, enter eps_r_p.
- From the ε′′ list, select User
defined. In the associated text
field, enter eps_r_b.
5 Under Magnetic Field, from the μr
list, select User defined and Full.
6 In the μr table, enter the settings as
in the figure to the right.
7 Under Conduction Current from
the σ list, select User defined and
leave the default value at 0.

Now add ports for excitation and transmission.

Port 1, Port 2, and Port 3

1 On the Physics toolbar click Boundaries and choose Port . Port 1 is
added to the Model Builder.
2 Select Boundary 1 only for Port 1.
3 Go to the Settings window for Port. Under Port Properties from the Type of
port list, select Rectangular.

| 25
4 From the Wave excitation at this port Boundary 19
list, select On.
5 On the Physics toolbar click
Boundaries and choose Port
to add another Port node. For
Port 2:
- Select Boundary 18
- Select Rectangular from the Type Boundary 18
Boundary 1
of port list
6 Add another Port node. For Port
3:
- Select Boundary 19
- Select Rectangular from the Type of port list

The node sequence in the Model Builder should match this figure.

Me sh

The mesh automatically aligns to the geometry. Physics related considerations can
also be incorporated into the mesh. In particular, the mesh needs to resolve the
local wavelength and, for lossy domains, the skin depth. The skin depth in the
ferrite is large compared to the size of the domain so the main concern is to resolve
the local wavelength. This is done by providing maximum mesh sizes per domain.
The rule of thumb is to use a maximum element size that is one fifth of the local
wavelength (at the maximum frequency) or less.

26 |
Free Tetrahedral 1
1 On the Mesh toolbar click Free
Tetrahedral .
2 Right-click Free Tetrahedral 1
and choose Size .
3 Go to the Settings window for Size.
Under Geometric Entity Selection
from the Geometric entity level list,
select Domain.
4 Select Domain 1 only. This is the
domain filled with air inside the
waveguide.
5 Under Element Size click the Custom
button.
6 Under Element Size Parameters
select the Maximum element size
check box. Enter 1.5e-2 in the text
field.

Size 2
1 Right-click Free Tetrahedral 1
and choose Size . A second Size
node is added to the sequence.
2 Go to the Settings window for Size.
Under Geometric Entity Selection
from the Geometric entity level list, select Domain.
3 Select Domain 2 only. This is the ferrite post domain.
4 Under Element Size click the Custom button.
5 Under Element Size Parameters select the Maximum element size check box.
Enter 4.5e-3 in the text field.
6 In the Settings window for Size click Build All .

| 27
The node sequence in the Model Builder and the mesh should match the figure
below.

Study 1

The final step is to solve for the nominal frequency and inspect the results for
possible modeling errors.
1 In the Model Builder expand the Study 1 node, then click Step 1: Frequency
Domain .
2 Go to the Settings window for Frequency Domain. Under Study Settings, in the
Frequencies text field, enter 3[GHz].
3 In the Model Builder, right-click Study 1 and choose Compute .

Results

Electric Field
The default multislice plot shows the electric field norm. It is best viewed from
above, so click the Go to XY View button on the Graphics toolbar.

28 |
The electric field norm gives a good indication of where the main power is flowing
and where there are standing waves due to reflections from the impedance
mismatch at the center.

Study 1

The remaining work is to vary the two design parameters to minimize reflections
at the nominal frequency. To do this, perform parametric sweeps over the design
parameters (scale factors).
Modify the study to vary the scale factor determining the size of the ferrite post.
The study type is still Frequency Domain.

Parametric Sweep
The parametric sweep over the scale factor is added as an extension to the
frequency domain study.
1 On the Study toolbar click Parametric Sweep .
2 Go to the Settings window for Parametric Sweep. Under Study Settings click
Add under the Parameter names table.
3 In the Parameter names list, select sc_ferrite (Geometry scale factor).

| 29
4 Under Study Settings click the Range button under the Parameter names
table. In the Range dialog box:
- In the Start text field, enter 0.5.
- In the Step text field, enter 0.003.
- In the Stop text field, enter 0.53.
Note: Alternatively, enter range(0.5,0.003,0.53) in the Parameter value list.

5 Click Replace.
6 On the Study toolbar click Compute .

30 |
Results

S-parameter (emw)
The reflection coefficient versus the scale parameter is plotted by default. The plot
of the S-parameter indicates a minimum for a scale factor of 0.518. Freeze the
parameter at this value and add a new study to vary the next scale factor.

Global Definitions - Parameters

1 On the Home toolbar click Parameters and select Parameter 1.

| 31
2 In the Settings window under Parameters, in the Expression column enter
0.518 in the sc_ferrite row.

Study 1

Parametric Sweep
1 In the Model Builder under Study 1, click Parametric Sweep .
2 Go to the Settings window for Parametric Sweep. Under Study Settings in the
Parameter names list, select sc_chamfer (Geometry scale factor).
3 Under Study Settings click the Range button under the Parameter names
table. In the Range dialog box:
- In the Start text field, enter 2.8.
- In the Step text field, enter 0.04.
- In the Stop text field, enter 3.2.
Note: Alternatively, enter range(2.8,0.04,3.2) in the Parameter value list.
4 Click Replace.
5 Go to the Settings window for Global 1 under S-parameter (emw). Enter
sc_chamfer in the Expression under x-Axis Data.
6 On the Study toolbar click Compute . Or in the Model Builder, click
Study 1 and press F8.

32 |
Results

S-parameter (emw)
The plot of the S-parameter indicates a minimum for a scale factor of about 3.0.
Leave the parameter at this value and add a study for the frequency response.

Study 2

1 On the Home toolbar, click Add Study .

| 33
2 Under General Studies select Frequency Domain .

3 Click Add Study .

4 Click Add Study again to close the Add Study window.
5 In the Model Builder under Study 2, click Step 1: Frequency Domain .

6 Go to the Settings window for Frequency Domain. Under Study Settings click
the Range button.
7 From the Entry method list, select Number of values.
- In the Start text field, enter 2.8[GHz].
- In the Stop text field, enter 3.2[GHz].
- In the Number of values text field, enter 21.
8 Click Replace.
9 On the Study toolbar click Compute . Or click Study 2 and press F8.

Results

The probe with the reflection coefficient versus the frequency is automatically
logged to a table and plotted while solving. At the last frequency, there are
pronounced standing waves. Look at the center frequency instead.

34 |
1 In the Model Builder under Results, click Electric Field (emw) 2 .

2 Go to the Settings window for 3D Plot Group. Under Data from the Parameter
value (freq (GHz)) list, select 3.
3 Click the Plot button. Click the Go to XY View button .

At the center frequency most of the standing waves are gone.

S-parameter (emw) 1

| 35
Finally look at all the S-parameters plotted versus the frequency. Go to the Global
1 plot in the S-Parameter (emw) 1 plot group. emw.S11dB, emw.S21dB, and
emw.S31dB are plotted by default.

This is the frequency response of the final design.

Ele c tr om ag net i c Wa ve s (e mw )

At this stage, it is convenient to partially invoke some higher-level control of what

port is excited. This is done by letting the value of a parameter decide. The
parameter can in turn be controlled by the solver when computing the full
S-parameter matrix by exciting one port at a time. Another advantage is that when
this control is activated, a plot group for the S-parameters is automatically created.
1 On the Home toolbar click Electromagnetic Waves, Frequency Domain .

36 |
2 Go to the Settings window for Electromagnetic Waves, Frequency Domain.
Under Port Sweep Settings select the Use manual port sweep check box.
Note: The default variable name
PortName is automatically added to the
Port parameter name text field but must
be declared as a global parameter to be
available for the parametric sweep.
So far, only the first port has been excited
so the full S-parameter matrix remains to be calculated by exciting one port at a
time. This is also needed in order to confirm that the circulator behaves as desired.
In the RF Module, this procedure is referred to as performing a port sweep.
During this stage, the S-parameters can optionally be exported to a Touchstone
file for documentation purposes and for use in external system simulation tools.
Add the name of the Touchstone file.
3 In the Settings window, select Export Touchstone file. Enter
lossy_circulator_3d.s3p in the Touchstone file export text field.

Note: The Touchstone file is saved in the start directory of COMSOL

Multiphysics unless a different file path is specified with the Browse button.
When there is the solution of a model, exporting a Touchstone file can also be
performed by right-clicking the Export node under Results and selecting
Touchstone or by selecting Touchstone under Data in the Results ribbon toolbar
(Windows users) or the Results context menu (Mac and Linux users).
The following instructions show how to set manually the necessary parameters and
parametric sweep running the port sweep for reference purposes. Once you have

| 37
a new Frequency Domain study, all other steps can be added automatically by
clicking the Configure Sweep Settings button.

Global Definitions - Parameters

1 On the Home toolbar click Parameters and select Parameter 1.

2 In the Settings window under Parameters:
- In the Name column, enter PortName
In the Expression column, enter 1

Study 3

1 On the Home toolbar, click Add Study .

2 Under General Studies select Frequency Domain .

3 Click Add Study .

4 Click Add Study again to close the Add Study window.
Use the this third study for the port sweep. The study is solved for a single
frequency to keep simulation time to a minimum although it is possible to solve
for a range of frequencies. All solutions are needed in order to display the
S-parameter matrix in a table so this setting has to be changed.
5 Go to the Settings window for Frequency Domain. Under Study Settings, in the
Frequencies text field, enter 3[GHz].

38 |
Parametric Sweep
The parametric sweep is used to control which port is excited. It overrides the
settings on individual port features and drives one port at a time using 1 W of
input power.
1 In the Model Builder under Study 3click
Parametric Sweep .
2 Go to the Settings window for Parametric Sweep.
Under Study Settings, in the Parameter names list, select PortName.
3 In the Parameter value list, enter (space separated) 1 2 3.

4 Alternatively, the above steps setting up the parametric Sweep can be automated
by clicking the Configure Sweep Settings button.
5 On the Study toolbar click Compute .

Results

After the computation finishes, the Touchstone file can be inspected in a text
editor. The S-parameter matrix can also be displayed in a table.
1 In the Model Builder under Results, right-click Derived Values and choose
Global Matrix Evaluation .

| 39
2 Go to the Settings window for Global Matrix Evaluation. Under Data from the
Dataset list, choose Study 3/Parametric Solutions 2.

3 In the upper-right corner of the Expression

section, click Replace Expression .
From the menu, choose Electromagnetic
Waves, Frequency
Domain>Ports>S-parameter,
dB>S-parameter, dB (emw.SdB).
4 Click the Evaluate button and the
S-parameter matrix displays in the Table
window located in the lower-right side of the COMSOL Desktop. Alternatively
you can open the Table window from the Home toolbar by clicking Windows
and choosing Table .

Note: The matrix is nonsymmetric, which is typical for a device based on a

gyrotropic material. There are three groups of matrix elements that differ in the
fourth digit. These should in theory be equal within each group so this gives
some indication of the discretization errors.

40 |
Definitions

As a final step, create a plot and also use it as a model thumbnail. From this plot,
it should be possible to identify the model at first glance so it has to display the
geometry and some characteristic simulation results. First change to the default
3D view and switch off the grid.
1 Click the Go to Default 3D View button on the Graphics toolbar.
2 On the Definition toolbar click View .
3 Go to the Settings window for View.
4 Under View click to clear the Show grid check box.

| 41
Results

1 In the Model Builder under Results, click

Electric Field (emw) 3 .
2 Go to the Settings window for 3D Plot Group. Under
Plot Settings from the View list, choose View 3.
3 Click the Plot button.
Next, delete the multislice and create a single slice.
1 In the Model Builder window, expand Electric Field
(emw) 3 node, right-click Multislice 1 and choose
Delete (or press Delete on the keyboard).
2 Click Yes.
3 Click Electric Field (emw) 3 . On the Electric Field
toolbar click Slice .
4 In the Settings window for Slice under Plane data
from the Plane list, choose xy-planes. In the Planes
text field, enter 1

5 Under Coloring and Style, choose HeatCameraLight from the Color table
Add a deformation proportional to the electric field to the slice.
1 Right-click Slice 1 and choose Deformation .
2 Go to the Settings window for Deformation. In the upper-right corner of the
Expression section, click Replace Expression . From the menu, choose
Electromagnetic Waves, Frequency Domain>Electric>Electric field
(emw.Ex,emw.Ey,emw.Ez).

42 |
3 Under Expression select the Description check box.

Display the magnetic field as arrows. Use logarithmic length scaling to make sure
that the arrows are clearly visible everywhere. Place the arrows well above the slice.
1 In the Model Builder, right-click Electric Field (emw) 3 and choose Arrow
Volume .

2 Go to the Settings window for Arrow Volume. Click Replace Expression .

3 From the menu, choose Electromagnetic Waves, Frequency
Domain>Magnetic>Magnetic field (emw.Hx,emw.Hy,emw.Hz).
4 Under Expression select the Description check box.

| 43
5 Under Arrow Positioning:
- In the Points text field for x grid
points, enter 45.
- In the Points text field for y grid
points, enter 45.
- For z grid points from the Entry
method list, select Coordinates.
- For z grid points in the Coordinates
text field, enter 0.1/3.
6 Under Coloring and Style from the
Arrow length list, choose Logarithmic.
From the Color list, choose Green.
The port excitation can now be selected for the plot group. For the model
thumbnail, select the second port.
1 In the Model Builder click Electric Field (emw) 3 .
2 In the Settings window for 3D Plot Group under Data, choose 2 from the
Parameter value (PortName) list.
3 Click the Plot button.

Select this plot to use as a model thumbnail.

1 In the Model Builder under Results, click Electric Field (emw) 3 .

44 |
2 Click the Root node (the first node in the model tree). On the Settings window
for Root, under Presentation, click Set from Graphics Window.
This concludes the modeling session unless you want to practice drawing the
“Geometry Sequence”, which continues below.

G e o me t r y S e q u e n c e

In Global Definitions - Parameters and Variables parameters were entered to prepare

for drawing the circulator geometry. Once the geometry is created, you can then
experiment with different dimensions by changing the values of sc_chamfer and
sc_ferrite and re-running the geometry sequence. These step-by-step
instructions build the same geometry that is contained in the application library
file lossy_circulator_3d_geom (imported in the section Geometry).
The geometry is built by first defining a 2D cross section of the 3D geometry in a
work plane. The 2D geometry is then extruded into 3D.
Note: Repeat the modeling steps in page 19-20. You need to complete the first
two sections Model Wizard and Global Definitions - Parameters and Variables before
defining the geometry.
Start by defining one arm of the circulator, then copy and rotate it twice to build
all three arms.

Rectangle 1
1 On the Geometry toolbar click Work Plane .
2 Under Work Plane 1 right-click Plane Geometry and choose Rectangle .
3 Go to the Settings window for Rectangle. Under Size:
- In the Width text field, enter 0.2-0.1/(3*sqrt(3)).
- In the Height text field, enter 0.2/3.
4 Under Position:
- In the xw text field, enter -0.2.
- In the yw text field, enter -0.1/3.
5 Click the Build Selected button.

Copy 1
1 Right-click Plane Geometry and from the Transforms menu select
Copy .

| 45
2 To select the object r1, move the
mouse pointer to the Graphics window
and hover over the rectangle r1 so that
it turns red. Then click on the
rectangle when it is red so that it turns
blue. The object r1 is added to the
Input objects list on the Settings
window for Copy. Click the Build
Selected button.

Note: To turn on the geometry labels in

the Graphics window, in the Model
Builder under
Geometry 1>WorkPlane 1>Plane
Geometry, click the View 2 node. Go to
the Settings window for View and select
the Show geometry labels check box.

Rotate 1
1 Right-click Plane Geometry and from the Transforms menu select
Rotate .
2 To select only the object copy1, which lies underneath the object r1, move the
mouse pointer to the Graphics window and hover over the rectangle r1 so that

46 |
 it turns red. While the rectangle r1 is red, use the mouse wheel and then click
when the rectangle underneath (copy1) is selected.
3 Go to the Settings window for Rotate. Under Rotation Angle in the Rotation
text field, enter 120.
4 Click the Build Selected button and then click the Zoom Extents
button on the Graphics toolbar. The geometry should match the figure so
far.

Copy 2
1 Right-click Plane Geometry and choose Transforms>Copy .
2 Select the object r1 only and add it to the Input objects list in the Settings
window for Copy.
3 Click the Build Selected button.

Rotate 2
1 Right-click Plane Geometry and choose Transforms>Rotate .
2 Select the object copy2 only.

| 47
Note: If you cannot locate copy2 in the Graphics window, click the Home
toolbar, then click Windows and choose Selection List. From the Selection
List, right-click copy2 and choose Add to Selection to add it to the Input objects
list.

3 Go to the Settings window for Rotate. Under Rotation Angle in the Rotation
text field, enter -120.
4 Click the Build Selected button and then click the Zoom Extents button
on the Graphics toolbar.
The geometry should match this figure.

48 |
Next, unite the three arms to one object.

Union 1
1 Under Work Plane 1 right-click Plane Geometry and choose Booleans and
Partitions>Union .
2 Select the objects r1, rot1, and rot2 only and add these to the Input objects list
in the Settings window for Union.
3 Click the Build Selected button. There is one object created called uni1.
Now, build the central connecting region and add the ferrite domain. During
these stages, the geometric design parameters are used. Start by creating a triangle
connecting the arms and then by subtracting a copy of what has already been
drawn from a circle of proper radius.

Circle 1
1 Right-click Plane Geometry and choose Circle .
2 In the Settings window for Circle under Size and Shape, enter 0.2/
(3*sqrt(3)) in the Radius text field.
3 Click the Build Selected button.

Copy 3
1 Right-click Plane Geometry and choose Transforms>Copy . A Copy 3
node is added to the sequence.
2 Select the object uni1 only.

| 49
Difference 1
1 Right-click Plane Geometry and
choose Booleans and
Partitions>Difference .
2 Select the object c1 only to add it to
the Objects to add list in the Settings
window for Difference.
3 Go to the Settings window. To the left
of the Objects to subtract section, click
the Activate Selection button.
4 On the Home toolbar click
Windows and choose Selection
List.
5 From the Selection List, right-click
copy3 and choose Add to Selection to
add it to the Objects to subtract list.
6 Click the Build Selected button.
The geometry should match this figure so
far.

Now, rotate the newly created triangle 180 degrees and use one scaled copy of it
to create linear fillets for impedance matching. Use another scaled copy to define
the ferrite.

50 |
Rotate 3
1 In the Model Builder under Work Plane 1, right-click Plane Geometry and
choose Transforms>Rotate .
2 Select the object dif1 only.
3 Go to the Settings window for Rotate. Under Rotation Angle in the Rotation
text field, enter 180.
4 Click the Build Selected button.

Copy 4
1 Right-click Plane Geometry and choose Transforms>Copy .
2 Select the object rot3 only.
It is now time to apply the first scaling for the impedance matching.

Scale 1
1 Right-click Plane Geometry and choose Transforms>Scale .
2 Go to the Settings window for Scale. Under Scale Factor in the Factor text field,
enter sc_chamfer (one of the parameters entered in the step Global Definitions
- Parameters and Variables).
3 Select the object copy4 only.
4 Click the Build Selected button.

Union 2
1 Right-click Plane Geometry and choose Boolean Operations>Union .
2 Select the objects uni1 and sca1 only. Use the Selection List to select the objects
if required.
3 Go to the Settings window for Union. Under Union click to clear the Keep
interior boundaries check box.
4 Click the Build Selected button.

| 51
The geometry should match this figure.

Next, apply the scaling for the ferrite region.

Scale 2
1 In the Model Builder right-click
Plane Geometry and choose
Transforms>Scale .
2 Select the object rot3 only.
3 Go to the Settings window for Scale.
Under Scale Factor in the Factor text
field, enter sc_ferrite.
4 Click the Build Selected button.

Extruding the 2D cross-section into a

3D solid geometry finalizes the
geometry.

Extrude 1
1 Right-click Work Plane 1 and
choose Extrude .
2 Go to the Settings window for
Extrude. Under Distances from Plane in the associated table, enter 0.1/3 in the
Distances (m) column.

52 |
3 Click the Build Selected button and then click the Zoom Extents button
on the Graphics toolbar.

Form Union
1 In the Model Builder click Form Union (fin) .
2 In the Settings window for Form Union click Build All .
The final sequence of Geometry nodes in the Model Builder should match the
figure.

The last step finalizes the geometry and turns it into a form suitable for the
simulation by removing duplicate faces, for example. It is performed automatically
when material is added or when physics features are defined, but it is good practice

| 53
to perform it manually as any error messages from this step may be confusing when
appearing at a later stage. The geometry should match this figure.

Note: If you skipped to this section to learn how to create the geometry, you can
now return to the next tutorial step: Materials.

54 |