D.-A. Luh, A. Brachmann, J. E. Clendenin, T. Desikan, E. L. Garwin,

1
Recent Polarized Photocathode RD at SLAC
D.-A. Luh, A. Brachmann, J. E. Clendenin, T.
Desikan, E. L. Garwin, S. Harvey, R. E. Kirby, T.
Maruyama, and C. Y. Prescott Stanford Linear
Accelerator Center, Stanford, CA 94025 R.
Prepost Department of Physics, University of
Wisconsin, Madison, WI 53706
2
Highlights

• Current cathode in use (high-gradient-doped
  strained GaAsP)
• Growth and preparation techniques for
  photocathodes and their weakness
• Possible solutions/improvements and current
  progress

3
High-Gradient-Doped Strained GaAsP

• Currently used in the accelerator
• Peak polarization 82 _at_805nm
• QE 0.4 _at_ 805nm
• No charge limit effect with available laser
  energy

4
High-Gradient-Doped Strained GaAsP

• Cathode Growth
• Grown by Bandwidth Semiconductor
• Metal-Organic-Chemical-Vapor-Deposition (MOCVD)
• Zn-doping
• Cathode preparation
• Anodized at 2.5V to form a 3 nm oxide layer
• Waxed to a glass for cutting
• Degreased in boiling Trichloroethane.
• Stripped surface oxide layer by NH4OH
• Transferred into loadlock immediately.
• Heat-cleaned at 600C for one hour
• Activated by Cs/NF3 co-deposition
• Heat-cleaned and activated twice

5
Weakness of Current Cathode Growth and
Preparation Techniques

• MOCVD
• The base pressure of MOCVD growth chamber is in
  high-vacuum range, compared with ultra
  high-vacuum in other techniques.
• MOCVD requires higher growth temperature.
• MOCVD growth mechanism is complicated.
• Zn-doping
• The diffusion coefficient of Zn in GaAs is high
  at the heat-cleaning temperature we use.
• The heat-cleaning capability of Zn-doped cathodes
  is limited.
• Single strained layer
• Strain relaxation in thick strained layers causes
  lower polarization.

6
Dopant Loss during Heat-Cleaning

• High-gradient-doped cathode shows charge limit
  effect after three activations at 600?C.

7
SIMS Analysis

• SIMS (Secondary Ion Mass Spectroscopy) analysis
  confirms Zn dopant loss after repeated
  heat-cleaning at 600C.

8
Strain Relaxation in Thick Strained Layers

• Strained layers start relaxing beyond critical
  thickness (10nm).
• Strained layers relax partially until reaching
  practical limit (100nm).
• Strain relaxation ? Lower polarization

Active Layer Thickness (nm) Polarization ()
MO5-5868 90 82
MO5-6007 170 70
9
Possible Improvements on Cathode Growth and
Preparation

• MBE (Molecular Beam Epitaxy) growth High
  quality films
• Ultra-high-vacuum environment
• Lower growth temperature and simpler growth
  mechanism
• More choices on dopants
• Be/C doping better heat-cleaning capability
• Lower impurity diffusion coefficients in GaAs at
  high temperature
• As-capped cathodes -- Lower heat-cleaning
  temperature
• Atomic-hydrogen cleaning Lower heat-cleaning
  temperature
• Superlattice structure Preserve strain in
  active layers ? higher polarization

10
MBE vs. MOCVD

• Both SVT-3982 and MO5-5868 are high-gradient-doped
  strained GaAsP.
• SVT-3982 is MBE-grown Be-doped (SVT Associates).
• MO5-5868 is MOCVD-grown Zn-doped (Bandwidth
  Semiconductor).
• Preliminary result shows that MBE-grown cathode
  has better performance.
• Heat-cleaning capability of Be-doped cathodes
  need to be determined.

11
Atomic-Hydrogen Cleaning

• The goal to achieve good QE with lower
  heat-cleaning temperature
• Thanks to Matt Poelker of Jefferson Lab for many
  discussions and helps.
• Cathodes are atomic-hydrogen cleaned, and then
  transferred into activation chamber through
  loadlock.

12
Preliminary Results from Atomic-Hydrogen Cleaning
System

• GaAs Reference Cathode stripped its surface
  oxide by NH4OH, heat-cleaned, and activated
• GaAs Test Cathode No NH4OH stripping. Cleaning
  procedures are indicated in the figure.
• Atomic-hydrogen cleaning shows promising results.
  Cleaning condition needs to be optimized.

13
Superlattice Photocathodes

• Critical thickness (10nm) limits the size of
  strained active region.
• Multiple quantum wells to preserve strain
• Strained layers sandwiched between unstrained
  layers
• The thickness of single strained layer is less
  than critical thickness.
• Band structure calculation to determine cathode
  structure parameters (well width, barrier width,
  and phosphorus fraction, etc.)
• X-ray diffraction to characterize cathode
  structure (layer thickness, composition, and
  strain, etc.)
• Photoluminescence to check cathode band structure

14
Superlattice Band Structure Calculations

• kp transfer matrix method (S. L. Chuang, Phys.
  Rev. B 43 9649 (1991))
• Dm transmission and reflection at interfaces,
• Pm propagation and decay in layers

• Set AN2 1, BN2 0 Change incident electron
  energy, and look at 1/A1 for transmittivity.
• Transmittivity maximum ? Resonant tunneling ?
  Energy level

15
Multiple Quantum Well Simulation
16
Multiple Quantum Well Simulation
widthBarrier 50nm

• QE Band Gap
• Polarization HH-LH Splitting

Effective Band Gap
HH-LH Splitting
17
X-Ray Diffraction -- Theory

• Braggs Law n ? 2 d sin?
• All lattice planes contribute to Bragg
  diffraction
• Every layer contributes a Bragg peak
• Repeating series of thin layers causes additional
  peaks

18
X-Ray Diffraction Rocking Curves

• Test cathode strained GaAs
• (004) scan distance between layers

GaAs Bulk
Graded GaAs1-xPx
GaAs0.64P0.36
Strained GaAs
19
Strained Superlattice GaAsP SVT-3682 and SVT-3984
T. Nishitani et al, SPIN2000 Proceedings p.1021
20
Strained superlattice GaAsP SVT-3682 and SVT-3984
CB1
1.65 eV
HH1
0.86 eV
LH1
GaAsP
GaAs
GaAsP
GaAs
GaAsP

• Photoluminescence confirms the simulation
  prediction

21
Rocking Curve (004) scan from SVT-3682

• Both SVT-3682 and SVT-3984 are superlattice
  cathodes
• MBE grown Be-doped (SVT Associates).
• Barrier width 30Å
• Well width 30Å
• Phosphorus fraction in GaAsP 0.36
• Layer number 16
• Highly-doped surface layer thickness 50Å
• XRD analysis on SVT-3682
• Well Width Barrier Width 32Å
• Phosphorus fraction in GaAsP 0.36

22
Superlattice Cathode Performance

• Peak polarization gt 85
• Good QE
• SVT-3984 was tested in Gun Test Lab at SLAC, and
  there was no charge limit effect with available
  laser energy.

23
Conclusion

• MBE-grown Be-doped cathodes show equal or better
  performance than MOCVD-grown Zn-doped cathodes.
• Preliminary test of atomic-hydrogen cleaning
  shows promising result.
• First strained superlattice cathodes show very
  good performance.

To do

• Study the heat-cleaning capability of Be-doped
  and C-doped cathodes.
• Optimize the process of atomic-hydrogen cleaning.
• Study As-capped cathodes.
• Test superlattice cathodes with different
  structure parameters