Strained Superlattice GaAs photocathodes at JLab M. Baylac

1
Strained Superlattice GaAs photocathodes at
JLabM. Baylac

• Qweak collaboration meeting
• August 17, 2004

2
Polarized Electron Guns at JLab
HV insulator
Photoemission from GaAs semiconductor
NEG pumps
NEG-coated Beamline
Strained GaAs in Gun2 (old material) Strained-s
uperlattice GaAs in Gun3 (new material)
3
Strained layer GaAs photocathodes

• From 1998 through 2003, we have used strained
  layer GaAs photocathodes at JLab (Bandwidth
  Semiconductor, Inc.).
• Reliable, well understood material.
• Stained-layer GaAs provides
• Good polarization P 75 at 840 nm
• Moderate quantum efficiency QE 0.2 at 840 nm
  
• Limitations that keep polarization lt 80
• limited band splitting
• relaxation of the strain for thickness gt critical
  thickness (10 nm)

e
4
Strained GaAs/GaAsP superlattice

• Very thin quantum well layers alternating with
  lattice-mismatched barrier layers
• Each superlattice layer is lt critical thickness
• Natural splitting of valence band adds to the
  strain-splitting
• Developed by SLAC with SVT Associates, Inc.
• SLAC-PUB-10331 (2004), submitted to
  Appl.Phys.Lett
• First samples received at JLab October 2003,
  characterized at the injector test cave

5
Superlattice structure
SVT associates, per SLAC specs.
6
Quantum Efficiency

QE ()
QE 1 versus 0.2 from strained layer material
we operate here
Wavelength (nm)
7
Beam polarization
Polarization ()
Highest polarization ever measured at the Test
Cave
Wavelength for Good QE and Polarization
Wavelength (nm)
8
Analyzing power (aka QE anisotropy)
Analyzing power smaller by factor of 3 compared
with strained-layer GaAs 4 versus 12 This
means smaller inherent intensity position
asymmetries on beam.
Analyzing power ()
Wavelength for good QE and polarization
Wavelength (nm)
9
QE vs hydrogen cleaning
Typical H-dose to clean anodized samples
Drawback Delicate material Cant clean with
atomic hydrogen Makes it tough to anodize edge
of cathode
QE ()
Hydrogen exposure time (min)
10
Superlattice Photocathodes at CEBAF

• Several failed attempts to load superlattice
  photocathodes inside tunnel guns
• Successful installation of un-anodized
  superlattice photocathode in Gun 3 (March, 2004)
• Activation gave QE 0.4 at 780 nm (vs 1 in
  test cave)
• Used during HAPPEx-He and portion of HAPPEx-H
  (June, 2004)

11
Poor lifetime

• Frequent spot moves were required to maintain 40
  ?A beam current at Hall A
• every week at start of run, every day as we
  approached July 4 shutdown!
• HAPPEx-He OK. HAPPEx-H not so good. Injector
  conditions changing too often. HC asymmetries
  were not stable.
• Poor gun lifetime atypical of CEBAF
  photoinjector.

QE profile after 3 weeks of running
12
Polarimetry in hall A

• Compton (D. Lhuillier)
• 5 MeV Mott (J. Grames)

P 86 ? 3
e
13
Parity quality beam?

• Short run numerous spot moves
• gt Jury is still out. Poor gun lifetime made it
  difficult to assess performance of superlattice
  photocathode from a parity violation experiment
  perspective.
• HAPPEX reports
• Charge asymmetry OK
• for both photocathodes
• Position asymmetries
• were smaller using gun2
• strained layer photocathode
• (no active position feedback)

14
Surface Charge Limit

• QE drops as laser power increases photoelectrons
  build up in band bending region create opposing E
  field that reduces NEA G.A. Mulhollan et al,
  Phys. Lett. A 282, 309 (2001)
• Reduces maximum available beam current. Lose
  laser headroom. Makes for shorter operating
  lifetime of gun.

QE is not constant
15
Lasers

• Our new commercial Ti-Sapphire lasers provide
  more laser power ( 300 mW) compared to our old
  diode lasers ( 50 mW).
• They are wavelength tunable. Now we can tune to
  peak polarization.
• Successful and reliable running since G0.
• Ti-Sapp laser superlattice photocathode a good
  match for high current Qweak experiment. 300 mW
  laser power QE of 1 can provide 1800 uA beam
  current.
• Max current only 360 uA with strained layer
  cathode. Not as much headroom.

http//www.tbwp.com
16
Summary

• Highest polarization ever measured at JLab P
  86
• Measurements of many samples at test stand
  indicates this is no fluke.
• 5 times higher QE than strained layer material.
• Smaller analyzing power should provide smaller
  inherent charge and position asymmetry. (Recent
  HAPPEx results do not support this claim.)
• Delicate material, more difficult to handle.
  Cannot be H-cleaned.
• Cant recover QE from a dirty superlattice,
  unlike strained layer
• We suffered surface charge limit. QE drops with
  increasing laser power. A concern for high
  current experiments like Qweak.

e
17
Outlook

• Poor lifetime due to supperlattice? Doubt it
• Gun 3 has a bad lifetime in 2003 using strained
  layer
• Un-anodized wafer increases damage on the wafer
• Reworked Gun 3 over the shutdown, hoping to boost
  lifetime
• QE lower in the tunnel than in test cave
• Hopefully due to the gun itself, not the wafer
• Received arsenic capped samples easier to handle
  and anodize (to be tested in lab)
• Smaller inherent HC asymmetries? Surface charge
  limit? Need more operating experience.