John Pulsifer, Mark Tillack

1
GIMM experimental setup and tests at prototypical
pulse length
John Pulsifer, Mark Tillack S. S. Harilal, Joel
Hollingsworth
With contributions from Roman Salij (Cabot
Microelectronics Engineered Surface Finishes)
HAPL Project Meeting Princeton, NJ 11-12 December
2006
2
Overview
2 of 16

• GIMM program logic
• Review of front-end/amplifier facility upgrade
• Short-pulse test results
• More evidence of variability in optics
• Efforts toward a thin film fabrication capability

3
3 of 16
4
Review of short-pulse setup using CompEX as
front-end and LPX as amplifier
4 of 16
CompEX pulse is sliced to 4.5 ns
Key challenges are timing and alignment.
5
Nice profiles, but limited to 5 Hz pulse
repetition freq to maintain energy stability
5 of 16
Spatial profile of amplified beam is smooth.
Amplified pulse shape (red) replicates the 4.5 ns
seed pulse.

• System jitter increases with increasing PRF
• 5 Hz PRF limitation due to energy variation
  greater than 10 at higher PRF

6
6 of 16
Based on 1-D thermal diffusion, we previously
applied a large safety factor with long-pulse
testing
Up to 6x104 shots were accumulated for a fluence
curve
Long pulse, M109
Scaled goal
Predicted short-pulse result, M109
IFE goal
7
7 of 16
Short-pulse test results exceeded our expectation
One mirror was used for both 4.5-ns and 25-ns
testing
The damage fluence is higher than expected from
simple scaling

• Damage does not scale like vpulselength (like
  Tmax)
• Effect of cumulative damage? ? s dt

8
Surface temperature effect Time at temperature
may also be a factor
8 of 16
Absorbed heat flux using fixed total energy
Short-pulse induced damage occurs at 23 less
fluence than long-pulse induced damage, not the
expected 50 less fluence.
Surface temperature (2x energy in Compex)
9
9 of 16
Our latest Alumiplated mirror (M109) performed
extremely well (long pulse)
Long pulse, M109
Long pulse, M85 (previous best)
Further evidence of variability in coating and
surface finishing
10
A quality coating with good diamond-turning
provides much better damage resistance
10 of 16
Poor (m80)
Good (m85)
Best (m109)
Alumiplate has not been a highly reproducible
fabrication technique. Best is 3 nm RMS
Roughness, 20 nm P-V
11
CMP may offer a pathway to higher quality and
better quality control
11 of 16

• 1 nm RMS Roughness
• 48 nm P-V

Alumiplate mirror with Chemical-Mechanical Polish
12
CMP mirror damage resistance is comparable to
previous Diamond-Turned mirrors
12 of 16
Damage fluence curves (long-pulse)
Damage morphology of CMP is the same as D-T
grain motion in the coating
13
We are developing in-house mirror fabrication
capability at UCSD
13 of 16

• Thick, thin-film deposition at UCSD Nano3
  facility (also externally at Thin Films, Inc.)
• Alloy development
• Post-processing (CMP, DT) to be done externally

Sputter system at Nano3 facility, UCSD
14
14 of 16
Conclusions

• We have obtained data with 4.5 ns pulses
• Short-pulse damage resistance is better than we
  expected
• Time at temperature probably the reason
• Latest batch of Alumiplate seems to be capable of
  meeting the requirements for an IFE GIMM
• First CMP results were obtained and are
  promising.
• Next Steps
• High cycle, alloys, substrates, large aperture

15
Next-step goals for GIMM RD
15 of 16
?
?
16
16 of 16
Questions?