Basic R&D Studies for Lower Emittance
Polarized Electron Guns
C. Suzuki, T. Nakanishi, S. Okumi, F. Furuta, K. Wada, T. Nishitani,
M. Yamamoto, T. Hirose, M. Kuwahara, R. Mizuno, N. Yamamoto,
H. Matsumotoa), M. Yoshiokaa), H. Horinakab), K. Wadab),
T. Matsuyama b) and H. Kobayakawac)
Nagoya University, Department of Physics, Nagoya 464-8602, Japan
a) KEK High Energy Accelerator Research Organization, Tsukuba 305-0801, Japan
b) Osaka Prefecture University, Department of Physics and Electronics, Osaka 599-8531, Japan
c) Nagoya University, Department of Materials Processing Engineering, Nagoya 464-8603, Japan
Abstract. In order to produce the lower emittance electron beam, the higher field gradient must
be used for the gun. From this point of view, we try to develop both of a polarized DC-gun and
RF-gun. In case of a 200 ke V polarized DC-gun at Nagoya University, the accelerating gradient
at the photocathode surface was designed to be 3 MV/m, and the dark current emitted from the
SUS electrodes was suppressed below 1 nA. To increase the gradient, we tested the property of
pure Titanium as a new electrode material. The tested electrode showed small dark current (~1
nA) even at field gradient of 88 MV/m, which is as twice as higher than that of SUS electrode.
Concerning the feasibility of a polarized RF-gun, it seems difficult for the NEA surface to
survive in high field gradient of 100 MV/m. Therefore, we proposed a new type of polarized
electron source using two-photon excitation method, for which it is not necessary to use the
NEA surface to extract electrons into vacuum. For this method, the polarization higher than
90 % was already demonstrated by the photoluminescence measurement using the bulk GaAs
crystal.
1. INTRODUCTION
Recently, an electron gun using an NEA (Negative Electron Affinity) GaAs
photocathode has been noticed not only as a polarized electron source but also as a
good candidate of an electron source for the FEL (Free Electron Laser) machine. It has
the possibility to produce a beam with emittance less than 1.071 mm-mrad/mm (emitter
radius) [1]. In order to keep such a low emittance against the space charge effect at the
gun, the accelerating gradient higher than 10 MV/m should be required, which
corresponds to the gun bias-voltage of 300~500 kV.
As well known, the NEA performance is extremely sensitive to the surface
contamination and is easily degraded by back-bombardment of ionized residual gas
produced by the collisions with electron beam. In addition of UHV level of 10"11 torr,
the field emission dark currents between HV electrodes must be kept below 10 nA,
since the dark currents stimulate the desorption of gases from the anode [2]. There are
two ways for approaching the lower emittance by the higher field-gradient DC-gun or
by RF-gun. However, the technical difficulties are much greater for these guns, as the
CP675, Spin 2002:15th Int'l. Spin Physics Symposium and Workshop on Polarized Electron
Sources and Polarimeters, edited by Y. L Makdisi, A. U. Luccio, and W. W. MacKay
© 2003 American Institute of Physics 0-7354-0136-5/03/$20.00
1068
field gradient at the cathode surface is assumed to be one or two orders of magnitude
higher (10~100 MV/m) than those of existing DC-guns. At moment, both approaches
have the common NEA lifetime problem and many R/D studies are required to
overcome this problem. At first, the effort to reduce the dark current for high gradient
DC-gun is described, and next the idea of a new PES using two-photon excitation
method is described and discussed.
2. DARK CURRENT STUDY OF TITANIUM ELECTRODE
The future project to build the FEL machine, which is based on ERL (Energy
Recovery Linac) as the fourth generation of synchrotron light source, has been
discussed at Cornell and KEK intensively. The key technology is considered to
develop a 500 keV DC-gun that can produce the lowest level of emittance (<1.0rc
mm-mrad/mm) for the highest level of average current (~100 mA) from the NEAGaAs photocathode. In this gun, the accelerating gradient at NEA surface is required
to be more than 10 MV/m, and the dark current problem becomes more serious. In a
series of our dark current studies, the property of Titanium as an electrode material has
been just recently tested.
2.1 Experimental Apparatus
A study to reduce the field emission dark current has been continued by our group
for several years. The same test apparatus was used to supply a high DC field gradient
(<100 MV/m) and to measure the dark current from a pair of sample electrodes under
UHV (~10'n ton) condition.
Rough Pump
A schematic view of this test stand and the
System
geometrical shapes of the cathode and anode
are shown in Fig. 1. The sample electrode can
be replaced to compare the performance of
various kinds of electrodes. The field gradients
can be changed by control both of the gap
separation of the electrodes (0.5~20 mm) and
of the bias voltage (0-150 kV). The dark
current emitted from the cathode is collected at
the anode that is electrically isolated from
ground level and measured by a pico-ampere
meter. The dark current study for SUS and
copper electrodes were already done by this
test stand [3].
FIGURE 1. A schematic view of test stand.
2.2 Experiment
A new titanium electrode was made of JIS grade-2, which contains 99.4 % of Ti.
After the machining, the electrode was finished to mirror like surface by buff polishing.
1069
The surface was cleaned by the 5 minutes high pressure rinsing (80 kg/cm2) with
final ultra-pure-water rinsing. The electrode was installed to the chamber in class-100
clean room immediately after this rinsing. Then, the apparatus was pumped down to
10"11 torr after the baking at 250°C for 1 week. The dark current was measured after
careful current conditioning for 2 weeks, which means that no break down was
occurred during the high voltage conditioning up to 100 kV. Then, the gap separation
was set to 1.0 mm for dark current measurement. Fig.2 shows the results for Ti
electrode together with SUS and Cu data for comparison. The preparation procedures
for each electrode are summarized in Table 1.
gap=1.0mm
Obviously, Ti electrode achieved the
:
!
'
i
!
_......._......... -r-------|------i--f---|----—
highest field gradient of 88 MV/m for the
i t piireTi
(NK-clbanZ) H i
small dark current level of 1 nA. This ^<t
36MV7m i l l
i 1 8iMV/m
value is as twice as higher than that of ^ 1000 --.---L--^
carefully fabricated SUS electrode. As ^
i ^Jcu | | !
; i 47MV/ni 7
[
well known, the attainable field gradient e
i 1
i
i 1
i
depends on the electrode material. Ti has £
t-i--t---tt-t--higher melting point, lower secondary u 400
electron emission rate, and lower ion- <
!
1
!/ 1
200
Tf----;-----?f ----;-•---sputtering rate than SUS and Cu. However,
J
I
J
\
it is not yet understood completely which
20
40
60
80
100
120
property of metal contributes to achieve
FIELD GRADIENT (MV/m)
high field gradient. Further studies should
be needed from this point of view.
FIGURE 2. Results of dark current measurement.
4-+---W-4-—
i;.....L....iL..j........_
i 1
TABLE 1. Preparation procedures for Ti, SUS, and Cu electrodes.
Electrode
Material
Surface finish
Ti
JIS grade-2
Buff polishing
SUS
Clean-Z1
Electro-chemical buffing
Cu
Class-10FHC with HIP2 Diamond turning with ethanol
Rinsing
High pressure rinsing
Hot ultra pure water
Ultra pure water___
1
Clean-Z is a new type of SUS material, which contains much fewer impurities than normal SUS 316L.
Hot Iso-static Pressing.
The details of preparation procedures for SUS and Cu electrodes are described in Reference [3].
2
2.3 Future Plan of Dark Current Study
Now we are planning to fabricate Ti electrodes for our 200 keV polarized electron
gun in order to obtain higher accelerating gradient to minimize the emittance growth.
Basic studies for the dark current have been also continued with the aims of revealing
its origin and mechanism. New electrodes made of Mo and Nb will be tested following
the study of pure and alloy of Ti material.
3. A NEW PES USING TWO PHOTON EXCITATION
RF-guns are now used at several laboratories as the injector system of accelerator,
because they can produce an electron beam with high peak current and short bunch
length of ps-range. However, the feasibility of a polarized RF-gun is not yet
1070
established. It was reported by Novosibirsk group that the lifetime of GaAs with NEA
surface is really too short in the RF-cavity [4]. From this result, we expected that a
mechanism with something new for production of polarized electron beam is needed
to solve the lifetime problem under 100 MV/m gradient, and proposed a new
possibility to use the two-photon excitation mechanism.
3.1 Principle
The principle of two-photon excitation method is rather simple. Fig. 3 (b) shows
two-photon adsorption mechanism. If we use right-circular polarized photon, only one
type of transition is allowed, that is, from the valence band with mj=-3/2 to the
conduction band with mj=+l/2. The expected advantages of this two-photon excitation
are summarized as follows, 1) even a bulk-GaAs can give the highest polarization of
100 % in principle so that the super-lattice or strained layer is not necessary. 2) The
electrons excited by two-photons have enough energy to escape into vacuum through
the small PEA surface. For example, if the laser wavelength of 1500 nm is used, the
obtained energy by two-photon adsorption is 1.65 eV that is larger by 0.23 eV than the
band gap energy of GaAs
(-1.42 eV).
"^I^ Conduction band
It means the
complete NEA surface is not
required to extract electrons.
This method seems applicable
for a polarized RF-gun for
which the ultra-short (0.1 ps)
laser pulse with highest peak
power can be used.
__
__
__ __ __ __
m
^m mj=3/2
J=-3/2 mj=-i/2 mj =i/2 mj=3/2
4=^/2^=172 Valence band -^-z^
(a)
(b)
m = 3/2
J - m^!72
FIGURE 3. Optical transitions in (a) superlattice or strained GaAs layer by circular polarized singlephoton excitation, and in (b) bulk-GaAs by circular polarized two-photon excitation.
3.2 Measurement of Photoluminescence Polarization
Photoluminescence spectra were measured to estimate the spin polarization of
conduction band electrons excited by two-photon absorption. Fig. 4 shows the
experimental set up, where a bulk-GaAs wafer was mounted in a cryostat and cooled
down to 90 K. An optical parametric oscillator was used as a light source that has
wavelength of 1500 nm, pulse width of 120 fs, repetition rate of 82 MHz, and average
power of 80 mW. The luminescence from the sample through a monochrometer was
detected by photomultiplier tube. The results are summarized in Fig. 5, where f (I")
indicates the right (left) circularly polarized component. From this result, the average
photoluminescence polarization (Pj) was evaluated to be 29 %, which corresponds to
58 % of the electron polarization. The spin polarization of electrons in conduction
band is relaxed during drifting toward the surface. The initial spin polarization (Pi) at
the instant of excitation is estimated from the equation as
P,=0.5Pt
1071
(1)
where Ts is the spin relaxation time and T is the lifetime of conduction band electron.
The measurement of time-resolved-photoluminescence was carried out at the same
time by using a Ti: Sapphire laser and a streak scope. Ts of 78 ps and Tof 50 ps were
obtained by analyzing of the decay curves of the photoluminescence polarization and
the photoluminescence intensity [5]. Substituting PI, Ts and T values into the equation
(1), the initial spin polarization was estimated to be as high as 92 %. Therefore, it was
really confirmed that the highest polarization could be obtained by this new method.
Optical parametric oscillator
Polarizer
Wave plate
±X/4
£>
|
s
Ref.
„
_____ P.M.T.
.
Sample Wave plate Monochrometer
1.44
1.46
1.48
1.50
1.52
o.o
Photon Energy (eV)
FIGURE 4. Experimental set up to measure the
photoluminescence polarization.
FIGURE 5. Spectra of right- and left-circular
polarized lights of photoluminescence.
3.3 Further plan of Two-photon excitation study
Following the photoluminescence measurement, the first attempt to extract
polarized electrons excited by two-photon absorption was tested by using a 70 kV DCgun and a Mott polarimeter. In this experiment, a bulk GaAs was illuminated by
Nd:YAG laser which has the shorter wavelength of 1064 nm and the much longer
pulse width of 20 ns than those of the photoluminescence measurement. As a result,
the extracted beam polarization was rather small, although the polarization was really
increased from 0 % to 13 % when the laser power was increased up to 55 jtlJ/pulse. It
is obvious that more shorter pulse width (< 1.0 ps) laser must be used to extract high
polarization and high QE beam. Otherwise, the highest polarization by two-photon
absorption is diluted by the large background contributions induced by single photon
absorption process. Concerning the RF-gun, we have started collaboration with KEK
to put the Cs2Te photocathode for practical use of ATF damping ring. There, we are
planning to continue the two-photon experiment by using this laser system which has
7.2 ps pulse width, -10 jllJ/pulse energy, and 357 MHz repetition rate.
REFERENCES
1. T. Nakanishi, "Polarized Electron Source using NEA-GaAs Photocathode" in Proceedings of Linac 2002,
Kyongju in Korea (http://linac2002.postech.ac.kr/db/proceeding/FR103.PDF)
2. K. Togawa et al, Nucl Instr. andMeth. A 414 431-455 (1998).
3. C. Suzuki et al, Nucl. Instr. andMeth. A 462 337-348 (2001) and AIP Conf. Proc. 570, 1009-1011 (2000).
4. A. V. Aleksandrov et al. Proceedings of EPAC98 1450- (1998),
http :/Vaccelconf.web. cern. ch/AccelConf/e9 8/PAPERS/TUP02 J.PDF
5. T. Matsuyama et al, Jpn. J. Appl Phys. 40 L555 (2001)
1072