RF Devices – Why and How

Jeremiah Holzbauer, Ph.D.

USPAS – Applied Electromagnetism Lecture 1
January 2019 - Knoxville
First, Sources and Recommendations

• Much thanks to Tom Wangler whom I’ve taken many, very

useful images from.
• Recommended Literature:
– RF Linear Accelerators – T. Wangler
• RF and Beam Optics for Linear Machines
• Extensive history with NC structures at Los Alamos
– RF Superconductivity for Accelerators – H. Padamsee
• Formative figure in Superconducting RF
• Lots of SRF theory and application, including cavity processing
and limitations
– Microwave Electronics – J.C. Slater
• Head of the MIT Rad Lab and Bell Laboratories (!)
• Difficult to find, but outstandingly written and understandable

2 J. Holzbauer | Introduction to RF - Lecture 1 2/27/2019

Course Outlook

• My goal is to give you the tools you need to work with

(around/in connection with) RF systems
• The class will be roughly an equal mix of:
– EM Theory (Analytical calculations, scaling, derivations)
– Practical Concerns (Implementation, design, and limitations)
• Start with RF Basics
– Maxwell Equations, Helmholtz Equations
– Standard Transmission Line/Cavity Geometries
– Figures of Merit and Optimization
• Move to Real Design
– Practical Concerns and Implications (fabrication, cleaning, etc.)
– Many Examples of RF structures in use today

3 J. Holzbauer | Introduction to RF - Lecture 1 2/27/2019

Quick Introduction

4 J. Holzbauer | Introduction to RF - Lecture 1 2/27/2019

Rutherford’s Desire
• Ernest Rutherford’s discovery of the structure of
the nucleus in 1909 opened new fields of
experimentation
– Nuclear Physics
– Eventually, High Energy Physics, Light
Sources, etc.
• These fields (and eventually others) require
particle “accelerators”
– Rutherford expressed a long-standing
“ambition to have available for study a copious
supply of atoms and electrons which have
individual energy far transcending that of the
alpha and beta particles” available from natural
sources.
– This desire has lead to a century of
accelerator physics research

5 J. Holzbauer | Introduction to RF - Lecture 1

High Energy Particles

• What does the science want? • What do they REALLY want?

– More Energy!
• Different science available Wall power
– Controllable/Tunable Energy!
• Dynamic behavior studies
• Fine structure investigations like
resonances
– More Intensity!
• Take data faster
• Study rare processes
– Rare isotopes
– Neutrinos
– Rare particle decays
– Variety of Particle Beams!
Desired Beam
• The ability to create and use
beams of any element/isotopes

6 J. Holzbauer | Introduction to RF - Lecture 1

What force to use?

• Strong • Radioactive decay

– Limited Natural Sources (some
• Weak artificial sources, not much better)
• Electromagnetic – Limited Intensity
– Very specific energies
• Gravity – Limited set of available beams
• Extremely weak
– Using the Sun’s gravity well, you could
get a proton up to ~22 MeV
– No comments on the practicality
• Electric Fields!
– Magnetic fields can’t do work
Δ𝐸 = 𝐹 ⋅ Δ𝑥 = 𝑞𝑣Ԧ × 𝐵 ⋅ 𝑣Δ𝑡
Ԧ =0

7 J. Holzbauer | Introduction to RF - Lecture 1

Most Simple Accelerator

• We don’t want Magnetic

Fields
– Steering should be left to
the magnets (most of the
time), we’re just talking
about energy gain
– If you assume B = 0, then
we only get static electric
fields
– We want only longitudinal
electric fields, strong in the
desired direction

8 J. Holzbauer | Introduction to RF - Lecture 1

Static Electric Fields

• Electrostatic Accelerators
– Limited energy gain (60 MeV/q)
– Can accelerate DC beams (used
often for particle sources)
• Tandem Accelerators
– By changing the particle charge from
negative to positive, twice the energy
can be achieved (limited current)

9 J. Holzbauer | Introduction to RF - Lecture 1

The Benefits of RF Fields

10 J. Holzbauer | Introduction to RF - Lecture 1

Wideroe Drift Tube Linac

• First RF accelerator conceived

and demonstrated by Wideroe
in 1927 in Aachen, Germany
• RF voltage of 25 kV from 1 MHz
oscillator was applied to single
electrode between two ground
planes
• Accelerated potassium ions to
50 keV, two gaps for twice the
voltage
• Sloan and Lawrence built one of
these style linacs with 30
electrodes, applying 42 kV to
get mercury ions up to 1.36 Wideroe’s Device (from his thesis)
MeV

11 J. Holzbauer | Introduction to RF - Lecture 1

How do we create V(t)?

• What do we need?
– Correct Frequency
– Accelerating fields that are
easy to access
– “Clean” accelerating field
distribution
– Reasonable mechanical
properties
– Efficiency energy storage

• Coaxial Waveguide Modes!

• Cylindrical Waveguide Modes!
– Everything else is just
topological adjustments of
these

12 J. Holzbauer | Introduction to RF - Lecture 1

How are these cavities judged?

Maximum
Transit Time Longitudinal
Charge Accelerating
Factor Focusing Term
Voltage

• R/Q • Geometry Factor

– Measure efficiency of transferring – Measure of efficiency of energy
energy to the beam storage in the cavity

13 J. Holzbauer | Introduction to RF - Lecture 1

Other Design Considerations
• Mechanical Issues • Higher Order Modes
– In operation, a cavity is exposed to
– As the beam passes through the
many different pressures that deform
cavity, it can excite ALL cavity modes
the cavity
– Strong beam asymmetry or offset
– How this deformation changes the
increases the strength of this coupling
cavity frequency and performance
must be well understood – These modes must be strongly
damped or they can cause emittance
growth or beam breakup

14 J. Holzbauer | Introduction to RF - Lecture 1

Cavity Design for Different Accelerators
• Synchrotrons (Ring Machines) • Linacs (Linear Accelerators)
– The beam sees the cavity MANY times,
– Single (or low #) pass machine
low gradient is typical
– High Gradient is KEY (reduces # of
– Field must be very clean and stable
cavities needed, therefore $$$)
– Very heavy higher order mode damping
– Reliability and ease of fabrication is
– Very large aperture very important (many cavities)
– Acceleration and bunching – Efficiency of operation also important

15 J. Holzbauer | Introduction to RF - Lecture 1

Summary by Tom Wangler (RF Linear Accelerators)
• In DC accelerators the energy gain is limited by the maximum applied
voltage, which is limited by electric breakdown.

• In RF accelerators (linacs, synchrotrons, cyclotrons) the final energy can

exceed the maximum applied voltage, which is applied repeatedly to the
beam. The final energy is limited only by economics.

• Synchrotrons are limited to low beam currents by beam instabilities

associated with the repetitive cycling of the beam from turn to turn through
unavoidable focusing lattice errors.

• Cyclotrons are not pulsed but are limited to low beam currents by weak
focusing and same inherent circular machine instabilities.

• Linacs can deliver high beam currents because they can provide strong
focusing to confine the beam and are not subject to circular-machine
repetitive instabilities.

16 J. Holzbauer | Introduction to RF - Lecture 1 2/27/2019

Maxwell to Helmholtz

17 J. Holzbauer | Introduction to RF - Lecture 1 2/27/2019

Basic Equations

• Maxwell’s Equations
• 𝛻⋅𝐷 =𝜌
are very general,
govern all classical 𝜕𝐵
• 𝛻×𝐸 = −
electromagnetic 𝜕𝑡
interactions • 𝛻⋅𝐵 =0
𝜕𝐷
Ԧ
• 𝛻×𝐻 =𝐽+
𝜕𝑡

• Lorentz Force
Where D = 𝜖0 𝐸 and B = 𝜇0 𝐻
• Stored Energy density • 𝐹Ԧ = 𝑞 𝐸 + 𝑣Ԧ × 𝐵
in EM fields 1 1 2
• Poynting Vector is • 𝑢= 𝜖0 𝐸2 + 𝐵 in vacuum
2 𝜇0
useful conceptual tool 1
(direction of energy • 𝑆Ԧ = 𝐸 × 𝐵 in vacuum
𝜇0
flow in EM fields)

18 J. Holzbauer | Introduction to RF - Lecture 1 2/27/2019

Helmholtz Equation

Any solution to the wave 𝜕2

• Starting from 𝑐2𝛻2 − 𝐸(𝑟,
Ԧ 𝑡) = 0
equation with a time- 𝜕𝑡 2
independent form can be
• Assume 𝐸 𝑟,
Ԧ 𝑡 = 𝐸𝑠 𝑟Ԧ 𝑇 𝑡
represented by the
Helmholtz Equation 𝛻 2 𝐸𝑠 1 𝑑2 𝑇
• Gives =
𝐸𝑠 𝑐 2 𝑇 𝑑𝑡 2
The solutions we want for • Assume both sides equal to a
RF applications are constant −𝑘 2
sinusoidal in time with
frequency ω. • Gives:
• 𝛻 2 + 𝑘 2 𝐸𝑠 = 0
Particles will generally be 𝑑2
accelerated in vacuum, so • + 𝜔2 𝑇 = 0
𝑑𝑡 2
we can assume 𝜌 = 𝐽Ԧ =
𝜖𝑟 = 𝜇𝑟 = 0, 𝑐 2 𝜇0 𝜖0 = 1
• We want: 𝑇 = 𝑒 𝑖𝜔𝑡 Good!

19 J. Holzbauer | Introduction to RF - Lecture 1 2/27/2019

Bring it all together - Free Space and Beyond

• Plugging this back into Maxwell’s Equations, we get: 𝜔 = 𝑘𝑐

Ԧ
• Solving for 𝐸 = 𝐸0 𝑒 −𝑖𝑘⋅𝑟−𝑖𝜔𝑡
𝑖
• Note that the magnetic field is specified: 𝐵 = 𝑘 ×𝐸
𝜔
• While we’ve solved for one frequency, we don’t have to only
work in one frequency
• If required we can solve for as many frequencies as required
and add using Fourier Superposition to give the full, final field
distribution
• Now, we have a powerful tool, just solve the Helmholtz
Equation with whatever boundary conditions are appropriate
to give us the RF fields in a structure of our choosing!

20 J. Holzbauer | Introduction to RF - Lecture 1 2/27/2019

Boundary Conditions

21 J. Holzbauer | Introduction to RF - Lecture 1 2/27/2019

Boundary Conditions

• For most of this class, we’ll only be considering surface

boundary conditions
• For magnet design, it’s often important to simulate the device
plus its surroundings because those fields (can) penetrate
through materials, significantly complicating simulation
• For RF, metallic walls serves as an effectively perfect shield
with very small (negligible) penetration depth
• Additionally, RF transmission will be bounded and well
behaved, not free travel of plane waves
• Free propagation and interaction with EM fields and charges,
like in a Free-Electron Laser are significantly more complex
and will not be covered here (See G. Stupakov’s class at
USPAS)

22 J. Holzbauer | Introduction to RF - Lecture 1 2/27/2019

Perfect Conductor Boundary Conditions

• Gauss’/Faraday’s Laws • Gauss’/Ampere’s Laws

applied to a pillbox/loop on applied to a pillbox/loop on
the metallic surface gives the metallic surface gives
two conditions: two conditions:
• 𝐷𝑛1 − 𝐷𝑛2 = 𝜌𝑠 • 𝐵𝑛1 − 𝐵𝑛2 = 0
• 𝐸𝑡1 − 𝐸𝑡2 = 0 • 𝐻𝑡1 − 𝐻𝑡2 = 𝐽𝑠
• Any field in the metal (2) will • Again, surface currents are
drive currents, so the only fine, but not inside the
stable configuration is with material, so 𝐵𝑛2 = 𝐻𝑡2 = 0.
𝐷𝑛2 = 𝐸𝑡2 = 0.
Electric Metallic Boundary Condition: Magnetic Metallic Boundary Condition:
𝐸𝑡𝑟𝑎𝑛𝑠𝑣𝑒𝑟𝑠𝑒 = 0 𝐻𝑛𝑜𝑟𝑚𝑎𝑙 = 0

23 J. Holzbauer | Introduction to RF - Lecture 1 2/27/2019

Normal Conducting Materials – Skin Depth

• The charge/current density

at the surface isn’t infinite.
• Fields do penetrate into the
material some amount.
• Fortunately, the generalized
Helmholtz equation still
applies, and gives
• 𝛿 −1 = 𝜋𝑓𝜇0 𝜎
• We will get to losses later
For copper at 1 GHz, this gives a Magnetic fields drive surface currents
2um skin depth. This perturbation on in a thin layer near the surface to
the very small for most applications. oppose the fields penetrating the
(Also, very simulation intensive) metal

24 J. Holzbauer | Introduction to RF - Lecture 1 2/27/2019

Additional Boundaries - In ComputerLandTM

The boundary conditions • Symmetry Boundaries

given are for real materials
– In 2D sims - rotational symmetry
and will always represent
the physical boundaries of – In 3D sims - used to reduce size of
your device. the simulation and generally takes
one of the following forms:
Additional types of • Electric – just like metallic wall
boundaries are used to • Magnetic - opposite conditions of
reduce the size of the electric B.C.
simulation or represent an • Periodic – Fields must be related at
interface with the outside two planes
world.
– Ports to the outside – Places for
power to flow into or out of your
I’ll talk about losses later. structure
• More complex B.C. that are defined in
software.
25 J. Holzbauer | Introduction to RF - Lecture 1 2/27/2019
Basic RF Design Process – So Far

1. Define closed boundary that has all the features you want to
simulate.
2. Choose appropriate boundary conditions, including fixed
frequency or eigenmode solver?
3. Use software to solve the Helmholtz Equation for the system
4. Repeat. Many times.
• So, now the question becomes what geometries should we
use?
• In the next section, we’ll cover the most common geometries,
play with them a while, then start to torture them into the most
unreasonable geometries we can think of and see what’s
useful.

26 J. Holzbauer | Introduction to RF - Lecture 1 2/27/2019