The RF system for the CEBAF Polarized Photoinjector (M. Crofford, C. Hovater, G. Lahti, M. Piller, M. Poelker)

Proceedings of the 1999 Particle Accelerator Conference, New York, 1999
W.J. Schneider , P. Adderley, J. Clark, A. Day, B. Dunham+, J. Hansknecht, P. Hartmann, J. Hogan,
R. Kazimi, D. Machie, M. Poelker, J.S. Price+, P.M. Rutt, K. Ryan, C.K. Sinclair and M. Steigerwald
Thomas Jefferson National Accelerator Laboratory, Newport News, VA
Construction is underway at Jefferson Lab on a loadlocked polarized electron source. The design incorporates
all of the essential features of the existing non loadlocked gun and improves on the designs of existing loadlocked guns operating at other labs. When complete, we
expect the new load-locked gun to enhance the versatility
of the JLAB polarized injector.
Historically, load-locked guns have been constructed as a
means of circumventing seemingly insurmountable
obstacles that have prevented labs from delivering
reliable polarized beam to physics end-stations. For
example, at SLAC, prior to the construction of their loadlocked gun, full cathode activation in the main gun
chamber caused high voltage breakdowns. It is believed
that the high voltage breakdowns were associated with
cesium deposition on the cathode electrode during the
initial activation [1]. Once this process (i.e., initial full
activation of the photo cathode) was performed in the
preparation chamber of their load-locked gun, the high
voltage incidents ceased. At MAMI, short cathode
lifetimes (~ hours) necessitated frequent cathode
replacement, a situation that prevented reliable beam
delivery to nuclear physics users during a typical monthslong experiment [2]. The load-locked gun at MAMI now
allows the accelerator staff to change photo cathodes with
minimal delay (few hours) to the nuclear physics
At Jefferson Lab, we have demonstrated that a loadlocked gun is not essential to meet the demanding
requirements of the Jefferson Lab nuclear physics
program. For example, unlike SLAC, we do not have any
high voltage problems associated with doing cathode
activation in the gun proper and, unlike MAMI, our
cathode life at high current is excellent. Over the past two
years, we have identified a number of mechanisms that
contribute to the decay of photo cathode quantum
efficiency. Understanding these decay mechanisms has
allowed us to implement design changes to our non-loadlocked gun that have resulted in exceptional lifetime (1/e
lifetime > than 100 H at 100 µA, > 1000 H at 10 µA) [3].
Work supported by U.S. Department of Energy DE-AC05-84ER40150
Now at GE Medical Systems Milwaukee, WI
0-7803-5573-3/99/$10.00@1999 IEEE.
With our present non-load-locked gun, 10’s of
Coulombs can be delivered to Users before intervention
(i.e., re-cesiation or heat treatment followed by full
activation) is required.
Such polarized source
performance means that intervention can occur during a
scheduled-maintenance day, (every other week at JLAB)
with no impact on the physics program.
Still, the obvious disadvantage of the present non-loadlocked gun is that a bake-out is required when photo
cathode replacement is necessary. We have reduced the
entire replacement process (photo cathode replacement,
bake-out and gun re-commissioning) down to fifty-two
hours beam to beam - albeit small, but still a delay.
Although we do not expect to improve the inherent
performance of the non-load-locked gun, we believe a
load-locked gun will greatly enhance the versatility of the
polarized injector. With a load-locked gun, we could
rapidly change photo cathodes to meet changing demands
of the nuclear physics program, and during biannual
facility development periods; we could change photo
cathodes quickly to conduct photo cathode research using
the superb diagnostics in the injector.
Before making specific design plans, we outlined the
basic features of the Jefferson Lab load-locked gun.
These features are based on our experiences at JLAB and
the experiences of our colleagues at other Labs. They
• Installation of the photo cathode from air to the gun
chamber must take less than six hours.
• The load-lock vacuum chamber must be at ground
potential and there must be no moving parts at high
• The gun and bend magnet must produce beam in the
horizontal plane; the bend magnet must not deflect
the beam more than 15 degrees.
• There will be four chambers; a) main gun chamber,
b) cathode preparation chamber, c) heat cleaning
chamber, and d) atomic hydrogen cleaning chamber
where samples are inserted into the load-lock
mechanism and cleaned with atomic hydrogen
• A fifth chamber may be added for storage of photo
cathodes or cleaned wafers ready for activation.
• Gun features that have proven to be essential on the
non-load-locked gun (superb vacuum, electrodes
designed specifically for Jefferson Lab beam current
requirements, electron optics that minimize stray
electrons hitting vacuum chamber walls, etc.) must
be incorporated into the load-locked gun design.
Proceedings of the 1999 Particle Accelerator Conference, New York, 1999
A brief description follows of the load-locked gun being
assembled with specific detail on the major points of
preparation chamber is accomplished through a 1.5-inch
ultra high vacuum metal sealed VAT valve [7].
The gun design (figure 1) is a novel one as it makes use
of the better electron optics of a horizontal configuration,
has no moving parts at HV and has all adjacent chambers
at ground potential. Low base pressure (with and without
beam extraction) and wise choice of materials is thought
to be the most important ingredient for long photo
cathode lifetimes. The main gun chamber is manufactured
from a standard six-way stainless steel cross. It has one
220 l/s ion pump and three GP 500 MK 2, SORB-AC
SAES cartridge pumps symmetrically located around the
photo cathode. The non evaporable getter (NEG) pumps
are well suited for pumping CO, CO2 and greatly enhance
the pumping speed (~ 4000 l/s) for hydrogen, the
dominant gas species in the vicinity of the photo cathode.
Pressure in the gun chamber is further reduced, because
all cathode preparation is done in a separate chamber. We
have achieved pressures below 1x10 Torr with a similar
gun [4] pumped by a massive NEG array. That was
measured by an extractor gauge with a measurement limit
claimed to be 1x10 by the manufacturer.
A stainless steel tee supports the cathode electrode
which has a shape designed for high current operation and
is made of titanium alloy (Ti-6Al-4V). The titanium
exhibits better high voltage performance (ie conditioning,
low field emission current at full gun fields, etc.)[5]. The
ceramic, to isolate the 100 kV (120 kV peak) high
potential, is located vertically so that photo cathode
preparation is performed at ground potential. A
molybdenum puck, which carries the cathode wafer, is
similar in concept to the SLAC design. The puck is held
in place inside the tee supporting the cathode with a stem
spring holder and sapphire rollers. The electrode holder,
triple point protectors and internal surfaces of the six-way
cross are electropolished while the electrodes are
metallographically polished with diamond paste. The
anode, also manufactured of titanium alloy, is mounted
on a stainless steel spider with a large open area to
increase vacuum conductance from the gun proper
through a 2.5-inch beam line. A channel cesiator is
provided behind the anode for in situ “touch-up”.
Alignment of the electrodes relative to the beam axis was
accomplished to better than a ½ mm.
A Surface Interface (SI) manipulator [6], on the beam
axis which can both translate and rotate, is used to move
the puck between the preparation chamber and the gun
(figure 2). A silver plated stainless steel adapter, mounted
on the manipulator engages a set of transfer ears inside
the puck to allow attachment and release of the puck.
Movement of the puck into and out of the gun has worked
smoothly although we have not yet baked out the entire
apparatus. Isolation between the gun and the cathode
The cathode preparation chamber contains all of the
components to produce negative electron affinity (NEA)
photocathodes: a stainless steel chamber with eight ports
placed around the circumference. Two ports are used for
the SI manipulators, the others ports are for a 40 l/s
sputter ion pump, a GP 100 SAES NEG, a channel
cesiator, a NF3 oxidizer, an optical window with a mirror
for light and a ring anode. In addition, a SRA [8] residual
gas analyzer (RGA) and an extractor gauge have been
added for vacuum diagnostics. On beam axis of the
chamber is the SI manipulator, previously mentioned, that
transfers the puck into the gun. The puck is transferred
from the on axis to the transverse manipulator via an
aluminum clamp that attaches to the molybdenum puck.
This transfer from gun manipulator to load lock
manipulator has also worked smoothly; again we have not
baked. Pressure in the cathode preparation chamber is
maintained at better than 1E Torr. At some point, we
may add a storage area to the preparation chamber that
will allow us to activate a number of wafers during an
accelerator maintenance day and store them for future
use. Separating the cathode preparation chamber from
the next chamber - heat-cleaning chamber is a 2.5-inch
VAT ultra high vacuum metal sealed valve.
Heating the photocathode samples is accomplished in a
separate chamber in a manner that differs from techniques
used at other labs. The heat-cleaning chamber is
fabricated from two six-way stainless steel crosses and
one water-jacketed spool piece where the heating takes
place. A SI manipulator allows transfer of the puck
between the load-locked or the cathode preparation
chambers into the heat-cleaning chamber. A Research
Inc. model 4085 infrared spot heater, powered by a
Chromolox Port –12221 control system is used to heat the
wafer to ~ 600 C at a ramp rate up and down of 1 degree
C per second. The chamber is also equipped with active
cooling. The heater is capable of 750 W although we have
found that 375 W appears to be sufficient and have
limited the power supply. A thermocouple in the IR
beam is presently used for control of the heater. We are in
the process of developing the parameters (rates of
heating, thermocouple location, pressure rise, etc.) for the
heat-cleaning chamber.
Separating the cathode
preparation chamber from the heat-cleaning chamber is a
2.5-inch VAT ultra high vacuum metal sealed valve. We
hope to maintain ultra high vacuum using a combination
of a 40 l/s ion pump and a GP100 SAES NEG to
minimize the pressure rise during the heating cycle and to
remove the hydrogen which is desorbed from the wafer
due to the hydrogen cleaning process. A similar 2.5-inch
Proceedings of the 1999 Particle Accelerator Conference, New York, 1999
VAT ultra high vacuum metal sealed valve separates the
heat cleaning chamber from the atomic hydrogen cleaning
The primary function of this chamber is to introduce a
number of different wafers on pucks and store them for
eventual processing. Processing that takes place in the
load lock is atomic hydrogen cleaning which is now the
only cleaning method used to prepare photo cathodes at
JLAB [9]. To maximize quantum efficiency we have
incorporated an in situ atomic hydrogen cleaner. The
highest quantum efficiencies achieved at JLAB have been
wafers that were cleaned in situ in a low voltage test
chamber. Similar efficiencies have been achieved when
samples obtained in a portable cleaning chamber were
transferred through air and installed in a gun. A Balzer
model TMU-071 turbo-pump is used during atomic
hydrogen cleaning where typical values for temperature
and pressure near the sample are 350 C and 1x10 Torr
respectively. The load lock makes use of the identical
infrared heater presently used in the heat-cleaning
chamber to raise sample temperature for atomic hydrogen
cleaning. It is anticipated that in situ atomic hydrogen
cleaning will provide a means to clean exotic photo
cathode materials for which wet chemistry techniques are
incapable; for example, chalcopyrites that contain silicon.
Fig. 1
In summary, we have discussed the design of the
polarized electron gun and its associated processing
chambers. We have developed a 100 kV photoemission
electron source, which currently supports the delivery of
highly, polarized, high average current CW electron
beams with long cathode operational lifetime. The new
load-locked source incorporates the capability of
exchanging, cleaning (both atomic hydrogen and heat)
and activating (co-deposition of cesium and NF3) the
photo cathodes without breaking the necessary ultra high
vacuum. The gun chamber proper is separated from the
preparation, heating and cleaning load lock chambers so
that it is capable of reaching the extreme high vacuum
regime to obtain long cathode lifetimes.
1. J.E. Clendenin, “Polarized Electron Sources”, IEEE 0-7803-3053,
2. E. Reichert, Nucl. Inst. Meth. A, Vol. 391, Pg 498-506, (1997).
3. M. Poelker et al, LE 98 Workshop, St. Petersburg, Russia, (1998).
4. C.K. Sinclair, invited paper this conference.
5. R. Latham, High Voltage Vacuum Insulation, Academic Press, (`
Fig. 2
6. Surface Interface Inc., 110 Pioneer Way, Suite D, Mountain View,
CA 94041, (415) 965-8205.
7. VAT Inc., 500 W. Cummings Pk, Woburn, MA 01801, (617) 9351446.
8. Stanford Research Systems, SRS, 1290 D Reamwood Ave.,
Sunnyvale, CA 94089, (408) 744-9040
9. C.K. Sinclair, et al, Atomic Hydrogen Cleaning of Semiconductor
Cathodes PAC ’97 Proceedings, Vancouver, B.C., (1997).