Download Status of the polarized source at MAMI

Status of the polarized source at MAMI
Kurt Aulenbacher, Valeri Tioukine, Markus Wiessner, Konrad Winkler
Institutfur Kernphysik der Universitdt Mainz, 55118 Mainz, Germany
Abstract. This talks addresses the operation of the polarized source at the Mainz Microtron MAMI.
The source is operating with selected photocathodes of modulation doped, uniaxially strained layer
photocathodes, which results in an average spin polarization of 80% and a quantum efficiency of
typically 2fj,A/mW. The operative lifetime has been improved by employing a novel activation
technique which reduces transmission losses in the vicinity of the cathode. In addition a considerable
simplification of the laser system has become possible by improving the power output of laser diode
seed lasers so that it is not necessary to employ power amplifier units. It was shown that the potential
for increasing the laser power is limited in our setup because of the thermal resistance between
cathode and the surrounding electrode.
PHOTOCATHODE PROPERTIES AND LIFETIME
IMPROVEMENT
This talk deals with some results and observations concerning the attempt to improve
the operative stability for operation with optimized polarization in our setup operating
at the MAinz Microtron MAMI.
The photocathode that is employed at MAMI is a 150nm thick strained layer cathode [1], with an active region that consists of uniaxially strained GaAsQ 95PQ 05. In recent
years the fabrication at Joffe institute in St. Petersburg/Russia has been optimized, so
that ever higher polarizations and quantum efficiencies are available for production runs
at MAMI. The latest samples achieve typical quantum efficiencies (q.e.'s) of 2 jiA/mW
in the maximum of the polarization spectrum. At this maximum the exciting photon energy is 1.50 eV corresponding to 826 nm wavelength. The available average polarization
is about 80%. An increase of polarization is frequently observed during the quantum
efficiency decay, a drop of a factor two may increase the polarization by 6% (absolute).
This was also observed by Mamaev et al. with the same type of cathodes [2].
The decrease of q.e. is believed to be caused by contamination of the Cesium/Oxygen
(Cs/O) activation layer and is undesirable because of the need for stable operating
conditions. It has been long guessed that very small transmission losses in the vicinity
of the anode of the source lead to this contamination, and that the losses are generated
by photoemission from the edge area of the cathode. It is not possible to suppress edge
emission by mechanical masks because of unwanted focussing effects.
The attempts to create a non mechanical mask have lead to the anodization technique [3] which was successfully applied at JLAB, where the cathode is oxidized outside
a small area around the main laser spot, so that the quantum efficiency is close to zero on
the outer parts of the cathode. Even though we have demonstrated that this method indeed leads to reduced transmission losses and to an increased photocathode lifetime [4]
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/$20.00
1088
we decided to make a further modification to the process. We have applied the Cs/O
evaporation through an aperture of 3 mm ('mask'). This method has the advantage that
no chemical treatment of the sensitive cathodes is necessary, and that it is possible to
do a comparative measurement of lifetimes with the same cathode. The mask activation
suppresses all photoemission outside the 3 mm diameter activation area. Even after several weeks of observation we have not been able to detect photoemission outside this
area, what indicates that the activation layer does not spread laterally on the semiconductor surface. In order to implement this technique we make use of the possibility to
move the cathode holder in vacuum by manipulators, so that the cathode is activated in
the mask and then removed from the mask and transferred to its operational position
in the source. In figure 1 a comparison is presented, it shows the behavior of the same
cathode operated at 200 Mikroamperes, once activated 'nude' and then masked.
16-
yude cathode
^-catnode reactivated
with mask
14-
Nude cathode
12-
Cs:0 Layer
10-
8-
Anodized cathode
Oxide layer
^^^^^^^^^^
B
WW™ ^*W$^~~~----~~^__
0
^d^ <S^H
•
- 2 0
2
4
6
8
10
12
14
16
Mask activated
cathode
18
extracted charge in Coulomb
FIGURE 1. Right: Illustration of the Cs/O coverage of 'nude', 'anodized' and 'mask activated' photocathodes. Left: Comparison q.e.-decay during beam current production.
It is evident that the decrease of q.e. is much smaller in the mask activated case. The
mask activation allows therefore to extract more charge. A measure for the available
charge in practical operation is given by Q = I • i where i (the 'lifetime') is defined as
the time for a 1/e decrease of q.e. from its initial value.
We observe a local q.e.-decrease which is restricted to the vicinity of the emission
spot: The local extractable charge from one beamspot of the laser (of the area At)
is restricted to Q/At = 3 • 104C/cm2 or about 20 Coulomb per beamspot. This has
negligible effect on availability, since the beam spot position can be changed remotely
during operation with a minimum impact on the injection parameters. For the total
charge that can be extracted from the cathode (for one activation) we observe differences
if operated on long time scale (moderate average current of 10-30 juA, typical for
operation with strained layer cathodes) and on short time scale (test experiments with
bulk GaAs cathodes of high q.e. and high average currents of 200 juA). In the first case
we found Q=40 and in the second case Q>100 Coulomb.
Even in the first regime it was possible to operate all experiments in the last year
continuously with high polarization over their scheduled runtime.
1089
LASER STUDIES FOR 2.5 GHZ PULSED OPERATION
The Master oscillator power amplifier (MOPA) system was first described with 0.5 GHz
operation for accelerator injection by Poelker [5] and transferred to 2.5 GHz for our
purposes later on [6]. It was considered as a mature system which allows to produce
sufficient optical power at a high stability. However an unforeseen problem occurred
when the vendor of the power amplifier elements stopped production in 1999, whereas
the master oscillator laser diodes remain available.
Our last amplifier burnt out in Spring 2002 after about 4000 hours of operation. Even
though power amplifiers are available from a new manufacturer1, we found favorable circumstances to omit the power amplifier and hence to obtain a considerable simplification
of the system. Presently our laser system consists out of the master oscillator laser alone.
The reason for this possibility is found in the ever higher quantum efficiencies (and good
lifetimes) that can be achieved with strained layer cathodes. The master oscillator is a
high power single mode laser diode which is capable of producing 200 mW of average
power in c.w. operation. With the typical initial efficiency of 2/j.A/mW the diode can in
principle deliver enough power to drive the accelerator to the limit of its capabilities (100
jUA) and still have a considerable reserve for optical transmission losses and decreasing
q.e..
Plaser>150mW BF=2,3W
t/ps
t/ps
Double-stub
tuner
Master
Diode
7
=o
2.5 Ghz
optical
power out
60
80
100
120
140
160
180 200
Average laser power / mW
FIGURE 2. Upper part: Pulse distortion at high average powers. Lower left: schematic of diode r.f.drive. Lower right: transmission as a function of average output power for different r.f. power levels.
The remaining problem for high power output is the requirement for 2.5 GHz pulse
operation which is essential to achieve good transmission through the accelerator. The
Toptica AG, Munich, Germany, model Toptica TA100
1090
typical operation consists of driving the diode with a d.c.-current and superimposing
r.f.-power by a bias-t (see figure 2). Improved r.f.-coupling to the diode is achieved with
the help of a double stub tuner.
We found that increasing the average power of the diode resulted in severe pulse distortion and pulse lengthening (see upper part of figure 2). Consequently the accelerator
transmission decreases with increasing power. However we found that this can be compensated to some extent by increasing the r.f.-power at the diode. The lower right part of
figure 2 shows the dependence on the r.f. power level. We find that more than lOOmW
of output power can be realized at > 90% transmission. Since the total charge from the
source is limited a high transmission leads to an optimized continuous operation of the
cathode.
Presently all running experiments at MAMI can be supplied by this simple system.
However, in order to be prepared for future eventualities (e.g. cathodes with smaller
q.e.'s) we will install a MOPA system in parallel to the existing system. An additional
problem that must be solved before the high power capabilities of the MOPA can be
employed, is discussed in the next section.
TEMPERATURE EFFECTS IN CATHODES
We investigated the temperature increase that is generated by the operation at average
laser powers which would be typical for operation of MAMI at the limits of its capabilities (lOOjuA). In order to do this under realistic conditions we have employed a non
invasive technique. The luminescence radiation generated by the laser absorption in the
semiconductor was observed and spectrally resolved (see figure 3). For a bulk-GaAs
cathode - which is assumed to behave thermally identical to a strained layer cathode the position of the luminescence spectrum is related to the temperature by the Varshni
equation [7].
photocathode
Thermal contact
between cathode
and holder
Pump laser
luminescence radiation
to spectrometer
780
800
820
840
860
880
900
920
940
Wavelength [nm]
FIGURE 3. Left part: schematic situation for the laser-heating experiment, Right side: Observed spectral shift with average power
The thermal conditions in this experiment where chosen in a way that they are similar
to the operation in the polarized source: We used the same photocathode holder, and the
1091
holder was coupled to room temperature by a good thermal conductance. The limiting
thermal resistance should therefore be the thermal contact between cathode and holder.
Figure 3 shows the spectral shift that is observed when the optical power is increased.
From the spectral shift we conclude that the temperature increase is about 40 degree per
100 mW.
Under this circumstances it seems not appropriate to use higher average powers than
100 mW because it is believed that the lifetime of the Cs/O surface layer will be reduced
drastically at temperatures higher than 50°C. From this we conclude that the operation
of strained GaAs at average currents of 100 juA will possibly suffer from the increased
temperature. We will therefore start work to decrease the thermal resistance of the
cathode/holder interface in the near future.
ACKNOWLEDGMENTS
This work was supported by the Deutsche Forschungsgemeinschaft (DFG) through the
SFB 443 and by the European Union through INTAS grant 9900125.
Further we want to thank Dima Orlov and Uli Waigel from MPI Heidelberg for
productive discussions concerning photoluminescence spectroscopy.
REFERENCES
1. Drescher, P., Andresen, H. G., Aulenbacher, K., Bermuth, J., Dombo, T., Euteneuer, H., Faleev, N. N.,
Fischer, H., Galaktionov, M. S., v. Harrach, D., Hartmann, P., Hoffmann, J., Jennewein, P., Kaiser, K.H., Kobis, S., Kovalenkov, O. V., Kreidel, H. J., Langbein, J., Mamaev, Y. A., Nachtigall, C, Petri, M.,
Pliitzer, S., Reichert, E., Schemies, M., Steffens, K.-H., Steigerwald, M., Subashiev, A. V., Trautner,
H., Vinokurov, D. A., Yashin, Y. P., and Yavich, B. S., Appl, Phys, A, 63, 203-206 (1996).
2. Mamaev, Y. A., Yashin, Y. P., Subashev, A. V., Ambrajei, A. N., and Rochansky, A. V., "Temperature
dependence of electron spin dynamics," in Spin 2000, 14th international spin physics symposium,
AIP proceedings, Vol. 570, edited by K. Hatanaka, T. Nakano, K. Imai, and H. Ejiri, AlP-publishing,
Melville, New York, 2001, pp. 920-925.
3. Sinclair., C., "Performance of the Jefferson Laboratory polarized electron source at high average
current," in International workshop on polarized sources and targets, (PST99), edited by A. Gute,
S. Lorenz, and E. Steffens, University of Erlangen, D-91058 Erlangen, 1999, pp. 222-230.
4. Aulenbacher, K., Euteneuer, H., Jennewein, P., Kaiser, K.-H., Kreidel, H. J., v. Harrach, D., Reichert,
E., Schuler, J., Tioukine, V., Wiessner, M., and Winkler, K., "New results from the Mainz polarized
electron facilities," in Spin 2000,14th international spin physics symposium, AIP proceedings, Vol.
570, edited by K. Hatanaka, T. Nakano, K. Imai, and H. Ejiri, AlP-publishing, Melville, New York,
2001, pp. 949-954.
5. Poelker, M., Appl. Phys. Lett., 67, 2762-2764 (1995).
6. Aulenbacher, K., Euteneuer, H., v. Harrach, D., Hartmann, P., Hoffmann, J., Jennewein, P., Kaiser,
K.-H., Kreidel, H. J., Leberig, M., Nachtigall, C., Reichert, E., Schemies, M., Schuler, J., Steigerwald,
M., and Zalto, C., "High capture efficiency for the polarized beam at MAMI by r.f. synchronized
photoemission," in Proceedings sixth European accelerator conference (EPAC98), edited by S. Myers,
L. Lilijeby, C. Petit-Jean-Genaz, J. Poole, and K.-G. Rensfeldt, Institute of physics publishing, Bristol
andPhiladalphia, 1998, pp. 1388-1390.
7. Varshni, Y. P., Physica (Utrecht), 34, 149-154 (1967).
1092