John Pulsifer and Mark Tillack

1
GIMM testing to millions of shots and beyond
John Pulsifer and Mark Tillack
HAPL Project Meeting 8-9 April 2008 Santa Fe, NM
2
Our current goal is to test alloy mirrors at very
high shot count to complete our proof of
principle demonstration of laser damage

• Challenges of high-cycle testing
• Need high quality mirrors
• Goal of 5 J/cm2, 3x108 shots
• Need a reliable, high quality laser
• Need control over the environment
• Need time a lot of time (months)

• Status
• Al-1Cu is our current standard due to better
  performance than alumiplate
• Fabricated in the nano3 lab at UCSD by sputter
  coating on Si
• Typical test runs of 107 shots, fluence limited
  by homogenizer
• Contamination has become an issue for high cycle
  testing

3
Previous future plans andprogress since Oct.
2007
Plans
Progress

• Achieve higher fluence with a different
  homogenizer (larger aperture, AR-coated)
• Identify and eliminate chamber contaminant source
  
• Continue shooting to achieve 108 shots

Peak of 4 J/cm2 now possible with homogenized
beam. Homogenization is essential for high
quality data.
Source term better understood. Near-term
solution Ne in chamber. Has implications for a
power plant.
Data acquired in the range of 107 New damage
morphology discovered. New issue with absorption.
4
Contamination has become a major issue with
longer-term exposures

• The problem
• Coatings on windows and on test sample
  (offsetting)
• Not seen previously due to differences in
  exposure time and vacuum
• Appears under hard vacuum (107 Torr), even with
  a cryopump
• Not unique problems exist in space telescopes
  (e.g. JWST)

• From top to bottom
• 1 million shots _at_ 1.4 x 10-7 Torr
• 16.1 million shots _at_ 1.1 x 10-6 Torr
• 1.4 million shots _at_ 1.2 x 10-5 Torr

Entrance window
Test mirror
5
Spectroscopic analysis was performed on a W
witness plate to determine impurities
Electron spectroscopy for chemical analysis

• Source term was identified
• Pump oil was found in cryo head
• ESCA identified C and O
• Near-term solution
• Chamber baking
• Ne background gas in chamber
• A scroll pump might help

• Implications for a power plant, and possible
  future RD
• Our chamber is more controlled than a power plant
• A power plant will have a continuous source term
  from targets
• This issue needs further study!

6
Facility improvements enable us to obtain higher
quality data

• Homogenizer
• Addition of AR coatings helped (increased from 3
  to 4 J/cm2)
• Combined w/ contamination control, we can
  maintain 4 J/cm2
• We can now perform uniform exposures over 10 mm2
  area
• Chamber baking
• 30 C for 13.5 hrs
• Base pressure improved
• to lt107 from lt104 Torr
• Leak rate improved
• to 7 x 107 from 1 x 105 Pa-m3/s
• LPX
• no problems since
• power supply repairs

7
Latest results of high-cycle testing6.7 million
shots at 4 J/cm2
unhomogenized
homogenized
8
Damage morphology in Al-1Cu at 6x106 shots
consists of triangular features
m119, 3 J/cm2 failure _at_ 5,770,860 shots
Laser direction
m120, 4 J/cm2 failure _at_ 6,734,665 shots
9
Triangular damage features at 6x106 shots
Laser direction

• This is the 1st known observation of this
  phenomenon.
• Is it debonding? Can it be prevented?

10
Latest results of high-cycle testingDamage
resistance is lower than expected

• The alloys better low-cycle damage resistance is
  not maintained at high shot count
• Is it fundamental to metal mirror damage?
• Similar degradation observed with DT vs. CMP
• Finish seems more important at low cycle,
  inherent microstructure at high cycle
• Al-1Cu is stronger why doesnt it exhibit
  better high-cycle behavior?
• Contamination may have compromised earlier data.
• Poor polarization purity may be affecting our
  results
• Could be the result of homogenization
• Every point receives peak fluence every shot
• If homogenization is the reason, we need to
  diagnose and eliminate weak spots in the mirrors

11
Mirror reflectivity is lower than expected

• 97 specular reflectivity in Al-1Cu (and
  Alumiplate) at 85 using the excimer laser
  (in-situ) with current optics
• Part of this may be non-specular scattering
• Imperfect polarization may be the cause
• Need to revisit this and demonstrate that our
  mirrors have acceptable absorptivity

12
Summary and Future Plans

• Data have been obtained up to 6.7x106 shots
  withhomogenized fluence levels up to 4 J/cm2
  over 10 mm2
• Contamination appears to be a larger concern than
  previously understood. More study is
  recommended.
• High cycle data is discouraging. We need to
  resolve this.
• We need to resolve the polarization/reflectivity
  issue first
• We need to understand the source of triangular
  damage.
• We would like to try making/testing a 0.5Cu
  alloy
• Limitations imposed by the homogenizer limit our
  database
• Take more unhomogenized data to fill in the
  curves?
• Simulate higher fluence data with lower angles
  (e.g. 80)?