Download Draft2 - Baartman`s computer at TRIUMF

HIGH POWER ALLISON SCANNER FOR ELECTRONS*
Aurelia Laxdal, Friedhelm Ames, Richard Baartman, Dan Brennan, Shane Rupert Koscielniak,
David Morris, William Rawnsley, Peter Vincent, Graham Waters, Dimo Yosifov
Abstract
TRIUMF’s new Allison emittance scanner is
designed to measure the emittance of a 100-300 keV
electron beam at an average current of 10 mA (1 kW
beam power cw) with a phase space area resolution of
0.032 µm, in a high vacuum environment 10 -9 Torr. The
emittance scanner has been tested with a beam produced
by a 100kV thermionic electron source. The source is
modulated at frequency of 650 MHz and can be operated
with a macro pulse structure to allow duty factors of less
than 0.1 % up to cw operation. The expected transverse
emittance is 4εrms ≤ 20µm. This paper discusses the
engineering challenges of designing an Allison scanner
for a high intensity electron beam, the components and
materials used, as well as the technologies introduced.
Also included are the experimental results of first tests
with the 100 keV source up to 600W @ 99% duty factor.
Emittance contours were measured at high resolution for
a variety of focusing conditions and intensities..
INTRODUCTION
An electron linac (e-Linac) is being designed and
installed at TRIUMF as part of the ARIEL project to
produce radioactive ion beams through photo-fission. The
final specified energy and intensity are 50MeV and
10mA. The electron source for the e-Linac is a thermionic
electron gun (e-gun) that will operate at 300 kV. An
initial electron source test stand operating at 100 kV is
installed for initial beam tests. A variety of beam
diagnostics have been installed in the test stand to
characterize the electron beam both at low duty cycle and
the high duty cycle.
REQUIREMENTS
An Allison type emittance scanner was proposed for
measuring the beam emittance of the 100 keV electron
source, and later for the 300 keV electron source.
TRIUMF’s previous experience with such scanners is for
low power density beams i.e. for pulsed H- beams at 300
keV and 1 µA, and for c.w. heavy ion beams at 50 keV
and a few nA. For the electron beam the scanner had to
be re-engineered to take high intensity beams at higher
beam power densities. The design requirements were: 10
mA full c.w. beam at 100 keV, for a beam size of 1cm.
Also, since the e-Gun operates at 10-9 Torr, the technology
and materials used for the emittance scanner must meet
UHV standards. For this purpose the emittance scanner
assembly was considered to be installed on a 6” CF
flange.
CONCEPTUAL DESIGN
___________________________________________
*Work supported by … THIS INFORMATION MUST BE WITHIN
#[email protected]
THE TEXT & COLUMN MARGINS
Briefly an Allison scanner consists of a mechanical unit
where a front slit is stepped across the beam while
downstream deflecting plates scan the each selected
beamlet across a downstream slit to a Faraday cup to
measure the angular spread (Fig. 1)[ref].
Fig 1: Emittance scanner schematics
For a non-relativistic Allison scanner the necessary scan
voltage for a charged particle of energy E entering with
an angle x’ through the entrance slit and exiting through
the second slit is:
'
V

4
gx
E
(
D

2

)
The maximum required scan voltage V m is:
2
2
2
8
Eg
V

(
D

4

)
m
for a maximum analyzable angle:
x'm2
gD
2

For 100 keV, the relativistic parameters are γ = 1.2, β =
0.55 and for 300 keV γ = 1.59, β = 0.79 respectively. The
maximum analyzable angle and the maximum voltage
must be scaled by a factor of:
k2(11)
For the 100 keV electron beam k = 1.09 and k =1.23 for
the 300 keV electron beam.
The (relativistic) maximum voltage and maximum angle
are:
V
m
_Re
l
V
mk
'
xm_Rel kx'm
DESIGN
In order to satisfy the design requirements without
jeopardizing the quality of the scans, the length of the
deflecting plates was chosen to be as long as possible
within the boundary of the CF flange. This defines a
length of 45mm, with a gap of 3.5 mm . Since the top
plate is biased at high voltage, the distance δ between the
electrostatic
plates
and
the
pair
of
slits
upstream/downstream them is chosen at 2 mm. The slit
gap for both entrance and exit slits is 0.038 mm. The
design parameters are also optimized to allow an
achievable maximum deflecting voltage and a reasonable
maximum analyzable angle. See Fig.1 and Table 1 for a
summary of the chosen design parameters.
Beam Energy
100 keV
300 keV
on the Tungsten plate and 300 deg C on the Tungsten
slits.
Table 1:
ENGINEERING & MATERIALS
The emittance scanner materials used inside vacuum
chamber are: Oxygen free Copper, selected for its high
thermal conductivity and for machining purposes,
Tungsten, for its high meting point and low vapour
pressure, SS and Aluminium Nitride (AlN) insulators, for
their low porosity and low out-gassing rate at high
temperatures in UHV. The machining technologies used
are: CNC, electron beam welding of Copper, ceramic
brazing to SS, electrical discharge machining (EDM) and
wire EDM, explosive bonding of Copper to SS and
explosive bonding of Copper to Tungsten.
The emittance scanner body is made of Oxygen free
Copper explosively bonded to SS, so it contains also the
vacuum flange: see Fig 2. The seal of this flange is a very
light ESI spring energized Metal C-Ring made of Silver
plated Inconel. Less than 5000 pounds clamping force is
required to compress the seal, so very light non-magnetic
SS hardware is used.
The water cooling of the emittance head is achieved
through 2 parallel water line of 0.25 inch diameter. The
parallel water lines have different lengths, so the flow
percentage is slightly higher on the upper branch which is
shorter compared with the lower branch. The water enters
and leaves the emittance head from its top part, which is
also the vacuum flange, through 2 VCRs welded directly
on the emittance head body: see Fig.2. To preserve the
UHV, the main supply and return cooling lines are placed
into an inner tube, at atmospheric pressure. Here the
assembly has a vacuum arrangement that is formed of:
atmosphere-vacuum-atmosphere sandwich.
The front slits, made of Tungsten, are covered by a
plate, made of Tungsten explosively bonded to Copper, to
stop the high power beam and efficiently remove the heat
through thermal conduction. This plate has a 1.25mm slit
width and collimates most of the beam.. ANSYS steady
tate thermal simulations were conducted for a simplified
3D model of a 1cm2 beam spot and with a beam power of
1 kW. See Fig.3. The simulation show 526 deg C on the
Tungsten plate, 270 deg C on the Copper bonded material
Fig 2: Emittance scanner head and assembly
FABRICATION
The explosive bonded materials are made by High Energy
Metals, Inc. in USA. The AlN insulators are manufactured
by Omley Inc. in USA. The EDM and wire EDM of the
Tungsten material is done by Innovative Tool & Die Inc
in Canada. The rest of the machining and the assembly
were performed at TRIUMF. The special ESI vacuum
seal is manufactured by Parker Inc. in USA.
INSTALLATION AND CONTROLS
Prior to installation in the beam line the emittance scanner
was cleaned to UHV standards with degreasing in an
ultrasound bath. Vacuum and electrical tests were
performed. Baking was done directly in situ flowing hot
air at 200 deg C through the cooling lines.
Emittance scans are controlled from EPICS running under
Linux, on a VME based CPU. The following types of
VME modules are employed: stepping motor controller
for positioning the slit; DAC for controlling the voltage
ramp; and variable gain current amplifier/digitizer for
scanner current.
The number and size of mechanical steps and the range
and size of the voltage ramp are selected via a GUI
interface. These settings can be saved and recalled. At
each step the DAC sweeps the deflector plate voltage in
Fig.3 ANSYS thermal simulations
and the scanner current is read and stored. When the scan is complete, the scan parameters: position, voltage and current
data are written to a file and made available to emittance analysis software. The slit position and voltage may be
controlled independently of the emittance software and the current display. This allows the emittance scanner to quickly
find the beam during the initial setup. The present readback device has a 10 Hz update rate so the emittance scanner has
a variable delay with a minimum value of 100 ms, so a coarse emittance scan of 21 positions by 21 angles takes 58
seconds and a detailed scan of 41 positions by 41 angles takes 190 seconds. [TREK PS, current amplifier QSX?,
accuracy?, pulsed beam?]
PERFORMANCE AND RESULTS
Scans were taken at 60keV for different average beam intensities and duty cycles ranging from 3 µA, 0.1% duty factor
to 11 mA, 99.9%. The scan taken at the highest beam power density is for a beam rms size of 2.69 mm at 11 mA c.w.
and 60 keV.
Fig.4 is one such beam, roughly 700 Watts/cm^2. In a 0.03 mm-mrad pixel of phase space, this is 4 microA at the peak
of the emittance figure. The noise is around 1 nA on this gain range. This allows detail down to the 98% contour. In
principle, more orders of magnitude sensitivity are available: a similar scanner installed in our radioactive beam facility
can measure currents to the fraction of a pA level. The data file, consisting of 6561 current readings, 81 positions and
81 voltages or angles in this particular case, is processed and contour-plotted using a MATLAB script. At high beam
power levels, the processing includes background subtraction: the current in the pixels along the lower edge of the
emittance plot, where the beam is entering the first slit but deflected too far to make it through the second slit, is used to
characterize this background. It arises from a heating effect causing a few nA positive current reading. At the very
highest current densities, the entry slit closes from thermal expansion, and this gives an anomalous dip in the centre of
the emittance figure. It is readily verified by lowering the duty factor.
The rms emittance for the data of Fig. 4 is found to be 10.1 µm, while the 39% emittance is 7.1 µm. For a perfectly
gaussian beam, the rms emittance and the 39% emittance are equal. The enlarged rms emittance is due to the “bowtie”shaped distortion evident in the figure. The origin of this kind of distortion is thought to be space charge combined with
a non-optimal Pierce geometry of the electron gun; it is under investigation.
Fig. 4
.
(1)
REFERENCES
[1] C. Petit-Jean-Genaz and J. Poole, “JACoW, A Service to the Accelerator Community,” EPAC’04, Lucerne, July 2004,
THZCH03, p. 249 (2004); http://www.JACoW.org.
[2] A. Name et al., Phys. Rev. Lett. 25 (1997) 56.
[3] A.N. Other, “A Very Interesting Paper,” EPAC’96, Sitges, June 1996, MOPCH31, p. 7984 (1996); http://www.JACoW.org {no
period after URL}
[4] F.E. Black et al., This is a Very Interesting Book, (New York: Knopf, 2007), 52.
[5] G.B. Smith et al., “Title of Paper,” MOXAP07, these proceedings.