High Gradient Dielectric Wall Accelerators,

1
High Gradient Dielectric Wall Accelerators,

• Yu-Jiuan Chen
• Lawrence Livermore National Laboratory
• Muon Collider Design Workshop
• December 8-12, 2008
• Thomas Jefferson National Laboratory
• This work performed under the auspices of the
  U.S. Department of Energy by Lawrence
• Livermore National Laboratory under Contract
  DE-AC52-07NA27344.

Patents Pending
2
Compact Proton DWA Accelerator Team

• Principal Investigator G. J. Caporaso
• LLNL Team D. T. Blackfield, Y.-J. Chen, S.
  Falabella, G. Guethlein, J. R. Harris, S. A.
  Hawkins, C. L. Holmes, S. D. Nelson, A. C.
  Paul, B. R. Poole, M. A. Rhodes, R.
  Richardson, S. E. Sampayan, D. M. Sanders, J.
  S. Sullivan, L.-F. Wang, J. A. Watson, J.
  Weir1
• TomoTherapy Team D. W. Pearson
• 1. Also Compact Particle Acceleration
  Corporation, USA

3
A new type of induction accelerator promises to
provide a compact proton source for cancer therapy

• Dielectric Wall Accelerator (DWA) for flash x-ray
  radiography
• Important technologies for the DWA
• High gradient insulator technology
• Dielectric materials
• Blumlein development
• Solid-state switch development
• Proton therapy concept
• Source and system tests
• Summary

4
DWA technology originated with a desire for more
compact flash x-ray sources
5
Dielectric Wall Accelerator (DWA) incorporates
pulse forming lines into a high gradient cell
with an insulating wall
Novel Zero Integral Pulse (ZIP) Forming Line with
potential for gt 10 MV/m
Switched load (beam)

• Important elements for the DWA
• High gradient insulators
• PFL architecture
• Switches
• Large size dielectrics with high dielectric
  constant and high bulk breakdown strength

State of the Art Electron Induction Accelerator
0.3 - 0.5 MV/meter Gradient
Patent Pending
6
High gradient insulators (HGIs) perform 2 - 5 x
better than conventional insulators
High Gradient Insulators
Conventional Insulators
U. S. Patent No. 6,331,194
7
A Blumlein pulse generator is formed from two
transmission lines
Output voltage
-
-


-
Closing switch
Matched load
Beam
Output voltage V
0
time
8
The Blumlein systems are more efficient with kAs
beams for high gradient DWAs

• System efficiency increases with beam current up
  to a theoretical maximum of 50 for our
  Blumlein design
• Proton therapy system requires 1 A maximum
  virtually no beam loading
• System efficiency on the order of 0.01
• Radial ZIP line architecture, with varying
  dielectric constant in layers to make Z constant,
  can theoretically deliver 100 of stored energy
  into the beam
• Switch resistive losses and use of less than full
  accelerating pulsewidth will reduce this figure

radial ZIP line - vary dielectric constant in
layers to make Z constant
Switched load (beam)
9
A new castable dielectric is one of the possible
materials for a DWA accelerator
Cast dielectric has high bulk breakdown strength
gt 400 MV/m (small samples) and can have epsilons
from 3 up to 50 for transmission lines
Patent pending
10
Cast dielectric has been successfully used for a
long pulse DWA
?r 10, 1.2 m long, cast ZIP line

• Prototype single arm cell
• 4 cast dielectric, zero integral pulse generating
  lines producing 25 ns pulse
• 4 self-breaking oil switches
• Power coupled to beam through 4 high gradient
  insulators
• 3 MV/m gradient (600 kV) across stack and HGIs
  with 1 kA electron beam load

11
SiC photoconductive switch has demonstrated fast
operation at 27.5 MV/m average gradient
SiC offers the possibility of high voltage, high
current operation at elevated temperature with
long lifetime and low jitter
13 mJ, 1064 nm NdYAG laser
Patent pending
12
Control of the temporal laser profile potentially
can compensate for beam loading in high-current
accelerators

• Photoconductive switch is really a
  light-controlled resistor
• Control of the temporal laser profile can change
  the shape of the accelerating waveform and
  potentially compensate for beam loading
• How well this can work is TBD

Modulator/ Switching Region
High Gain Amplifier Region (IdarknA)
13
DWA can be used in the single pulse traveling
wave" mode for any charge particle
HGI characteristics imply that the highest
gradients will be attained for the shortest pulses
b
Longitudinal Electric Field Plot
Uncharged transmission lines in this region are
not shown
Charged Blumleins

• full width at half maximum
• u speed of wall excitation
• g Lorentz factor

u?
patent pending
14
Stacks of Blumleins with independent switch
triggers implement the virtual traveling wave
accelerator
HGI
Blumlein
Beam
Patents pending
15
Spark source, capable of providing large proton
current, is a potential compact proton source
Grid assembly
Near Faraday cup
Far Faraday cup
Near Faraday cup Far Faraday cup
Time of flight between 2 Faraday cups indicates
proton extraction
Spark sources
16
We are working with Tomotherapy, Incorporated and
CPAC to develop a compact proton DWA

• 200 MeV protons in 2 meters
• Energy, intensity and spot width variable pulse
  to pulse
• Nanosec pulse lengths
• At least 200 degrees of rotation
• Up to 50 Hz pulse repetition rate
• Less neutron dose (neutrons still produced in the
  patient)
• System will provide CT-guided rotational IMPT
• TomoTherapy has licensed the DWA technology from
  the Lawrence Livermore National Laboratory and
  CPAC has a Cooperative Research and Development
  Agreement (CRADA) with LLNL

artist's rendition of a possible proton therapy
system
17
We have designed a small integrated test of the
accelerator

• F.A.S.T. (First Article System Test) is designed
  to be taken apart and rebuilt many times to
  increase system performance by using improved
  components
• It incorporates the essential elements for the
  accelerator
• HGIs
• Photoconductive switches
• Solid dielectric Blumleins
• Progress is made by finding breakdowns, repairing
  them and increasing performance to find the next
  limiting factor, etc.

18
F.A.S.T. is designed for rapid iteration
7 Blumleins containing photoconductive switches
F.A.S.T. mounted in an oil tank
Sensitive beam current monitor
HGI
Laser line of sight
19
F.A.S.T. was initially operated with sub-standard
switches
1 mm X 1 mm corner of original substrate (5X)

• Breakdown of the low quality dielectric in the
  Blumleins was expected, instead there were
  unexpected switch failures at low charge voltage
• Original SiC switches were defect free
• Replacement material used for FAST had structural
  defects and failed at stresses 20 - 30 x lower
  than original material
• We now have a new (proprietary) source of SiC

1 mm X 1 mm corner of FAST substrate (5X)
20
Near term plans working towards an integrated
test of proton source, HGIs and Photoconductive
switched Blumleins

• A small length of accelerator sufficient to
  verify the accelerator architecture and HGI
  performance with photoconductive switches
• New photoconductive switches over next 5 months
• Optimized dopant levels to lower on resistance
  and improve quantum efficiency
• High voltage packaging
• First Article System Test (F.A.S.T.)
• Integrate new photoconductive switches, solid
  dielectric, HGIs into a proof-of-principle device
• Proton acceleration by 31 December, 2008

21
12/31/08 deliverable requires a 300-keV proton
injector

• F.A.S.T. pulsewidth of 3 ns crossing time of
  injected protons
• gt 300 keV initial energy
• Acceleration provided by induction modules
• 5 induction cells simultaneously triggered
• 300 kV _at_ 30 ns
• SF6 insulation
• Powered by cable Blumleins
• Injector structure similar to that in source test
  stand

22
12/31/08 deliverable requires a 300 keV proton
injector
23
Proton Injector Hardware
5-Induction cells
Vacuum pump
2 100-kV Induction cells
Anode stalk
Thomson spectrometer
Source and Grids
camera
24
Summary

• DWA promises to dramatically increase the
  accelerating gradient of high current
  accelerators
• Good progress is being made on the technologies
  needed for the DWA
• New SiC and GaN material has been obtained for
  switches
• Pulse forming line dielectric materials (gt 400
  MV/M)
• High gradient vacuum insulators can be tested at
  the relevant field stress, sample size and
  pulsewidth
• F.A.S.T. has demonstrated acceleration of
  electrons and will be repeated with new
  photoconductive switches F.A.S.T. will accelerate
  protons by 31 December, 2008
• Spark source is being investigated for the
  potential compact proton source for the proton
  DWA for cancer treatment

25
(No Transcript)
26
High voltage pulser will permit testing of HGIs
at full parameters

• Delivers 3-5 MV into a 160 ? matched load, 12 ns
  pulse
• Will permit testing of hollow HGI insulators at
  full radial scale
• Now fully operational

27
Novel source and electrode system provides great
flexibility for intensity modulated beam
Goal to fit machine in a standard linac
radiation vault

• Only electric focusing fields are used for
  transporting the beam and focusing on the patient
• Variable beam parameters from pulse to pulse
  without using any beam intercepting methods,
  i.e., no range shifting wedges or scattering
  masks
• Variable spot size (2 mm - 2 cm diameter) on each
  pulse
• Variable charge (10 - 80 pC) on each pulse
• Variable beam energy (70 - 250 MeV) on each pulse