SPIN 2006

1
Transport Mechanisms in Polarized Semiconductor
Photocathodes
International Linear Collider Stanford Linear
Accelerator Center

• K. Ioakeimidi, A. Brachmann, J.E. Clendenin, E.L.
  Garwin, R.E. Kirby, T. Maruyama and C.Y. Prescott
• Stanford Linear Accelerator Center, Menlo Park,
  California 94025, U.S.A.
• R. Prepost
• Department of Physics,
  University of Wisconsin,Madison, Wisconsin 53706,
  U.S.A.
• G.A. Mulhollan, J.C. Bierman
• Saxet Surface Science, Austin, Texas, 78744,
  U.S.A.
• S.A. Gradinaru
• Magma Design, Cupertino, CA, U.S.A

2
Standard photocathode structure
3
Motivation for fast electron transport

• Observed emission spectrum
• Emitted electron polarization

a optical absorption R reflection
coefficient t1 electron lifetime in BBR tem
time of electron emission in vacuum  P0 initial
electron polarization upon excitation with
circularly polarized light S0 surface
recombination velocity tS spin relaxation time
in SL tS1 spin relaxation time in BBR
A. Subashiev et. al. SLAC-PUB 10901
4
Polarization vs time
Diffusion
Drift
Otherwise if we assume that in 100nm films the
scattering rate is
where s is the total number of scattering events
then
5
Scattering rates
BAPe-h EY Ion, POP DPtem
MIN scattering rates
6
Scattering statistics
10,000 electrons photogenerated in the active
region
7
Monte Carlo simulations of emission time
Transit time decreases by an order of magnitude
with field Scattering events decrease-no electron
hole scattering when the active region is
depleted
Drift
Diffusion
If we assume that the depolarization of transport
is 5-15 the expected increase due to the
applied field is 1-10
8
Structure of gridded cathode
MBE grown high surface/low active doping gridded
cathode
Metal grid, Schottky contact
Composition
Thickness
Doping
GaAs surface region 5-10nm 1-
5x1019cm-3 Be doped
GaAs,AlGaAs, GaAsP/GaAs active region
90nm 1014 - 1018 cm-3 Be doped
Al.3Ga.7As buffer
5x1018cm-3 Be doped
p- GaAs substrate, 5x1018cm-3 Zn doped
0.3um W film, Ohmic contact
9
Metal grid Tantalum ring
CB
E1V/100nm
Cs,O layer
GaAs 5 x1016cm-3 Be doping
Vac
GaAs 5 x1019cm-3 Be doping
VB
5nm
95nm
10
Requirements for the cathode

• Schottky contact at the surface ensures voltage
  drop at the active region rather than throughout
  the whole wafer.
• Best bias results are expected when 2 conditions
  are satisfied
• The active region is fully depleted to eliminate
  e-h scattering
• The energy of the accelerated electrons is in the
  region where min scattering rates and emission
  time occur
• For 100nm GaAs samples for 1V applied voltage the
  active region is fully depleted when the doping
  is 1e17cm-3

11
Requirements for the grid

• Deliver the voltage uniformly along the cathode
  surface
• Thick 1micron lines
• Spacing up to 100 microns
• Schottky contact at the metal/GaAs surface to
  drop the applied voltage at the cathode active
  region
• Schottky contact must survive at high temperature
  cleaning!!

12
Potential variation between two wires
Grid spacing for uniformdistribution of voltage
at the cathode surface
13
2 fabrication processes of metallic grids

• Photolithography and lift off
• 30nm thick W grid by e-beam evaporation
• Ebeam and physical mask
• 100nm Au/10nm thick W lines
• Sputtering of W with physical mask
• W lines 1micron thick

14
W/GaAs contacts after high-temperature
heat-cleaning
Before heat-cleaning

• To see the bias effect, the W/GaAs contact
• must be non-Ohmic.
• Non-Ohmic contact must survive
• heat-cleaning.

15
Bias effect observed on QE
I vs. V
QE
Over all resistance is 0.25?
16
Samples
17
IV curves of gridded samples
IV curve sample 19B
IV curve sample 18A
18
QE- gridded samples

• Thin GaAs films with 4mm 2D grid and 48mm pitch

5x1014cm-3
5x1016cm-3
19
Polarization- gridded samples

• Thin GaAs films with 4mm 2D grid and 48mm pitch

5x1014cm-3
5x1016cm-3
20
IQE 100nm GaAs, 5x1016cm-3, 784nm laser
illumination, 4x16mm grid
21
QE of samples with thin lines
39 Thin GaAs 5x1014cm-3 WAu lines
38 Bulk GaAs WAu lines
41 Thin GaAs 5x1014cm-3 W lines
40 Bulk GaAs W lines
22
Polarization of samples with thin lines
38 Bulk GaAs AuW lines
39 Thin GaAs 5x1014cm-3 AuW lines
40 Bulk GaAs W lines
43 Thin GaAs 5x1016cm-3 W lines
23
Sample 51 Thin GaAs with 1 um-thick W lines

• One 570 C heat-cleaning
• W/GaAs contact is Ohmic
• No bias effect on QE
• No Bias effect on polarization

24
Conclusions about the metal grid/lines

• The thickness of the grid needs to be 1um to
  distribute the voltage across the cathode surface
• Electron beam evaporation of W achieves Schottky
  contacts but the grid is thin (lt50nm)
• Sputtered W thick lines are susceptible to heat
  cleaning and dont form Schottky contact at the
  cathode surface
• Will try deposition of Re and plasma etched WN
  deposition

25
QE-Polarization trade-off
100nm GaAs
100nm GaAs
100nm GaAs
Case1 QE decreases because the vacuum level
elevates (No fields applied). The tunneling
probability of the electrons is decreased and the
electrons that scatter (especially in the BBR)
do not escape so the polarization increases.
Case2 High QE after cesiation (no fields
applied). The polarization is decreased because
although the tunneling probability of electrons
is higher than in case 1 (because the level of
vacuum is lower) a lot of electrons that scatter
in the BBR manage to escape in vacuum after
having been depolarized.
Case 3 Forward bias case. The tunneling
probability increases because the electrons are
further from the bottom of the well so the QE
increases. The escape time of the electrons
decreases so the polarization increases. The
well is also wider so the trapping probability is
not as high as in case 1.
26
CONCLUSIONS

• Monte Carlo simulations indicate that the
  QE-Polarization trade off can be broken by
  accelerating the electrons in the active region
• Preliminary experimental results indicate a 1
  increase in polarization
• Further improvement of heat resistant Schottky
  contacts at the surface and the use of strained
  superlattice samples are expected to show higher
  QE and polarization improvement.

27
International Linear Collider Stanford Linear
Accelerator Center

• Thank you!

28
Scattering statistics of electrons photogenerated
above bandgap
5e16cm-3
29
Voltage distribution across the cathode surface