John Pulsifer, Mark Tillack S. S. Harilal, Joel Hollingsworth GIMM experimental setup and tests at prototypical pulse length HAPL Project Meeting Princeton,

John Pulsifer, Mark TillackS. S. Harilal, Joel

Hollingsworth

GIMM experimental setup and tests at prototypical pulse

length

HAPL Project MeetingPrinceton, NJ

11-12 December 2006

With contributions from: Roman Salij (Cabot Microelectronics; Engineered

Surface Finishes)

1. GIMM program logic

2. Review of front-end/amplifier facility upgrade

3. Short-pulse test results

4. More evidence of variability in optics

5. Efforts toward a thin film fabrication capability

Overview

2 of 16

3 of 16

Review of short-pulse setup using CompEX as front-end and LPX as amplifier

Key challenges are timing and alignment.

CompEX pulse is sliced to 4.5 ns

4 of 16

Nice profiles, but limited to 5 Hz pulse repetition freq to maintain energy stability

• System jitter increases with increasing PRF

• 5 Hz PRF limitation due to energy variation greater than 10% at higher PRF

Amplified pulse shape (red) replicates the 4.5 ns seed pulse.

Spatial profile of amplified beam is smooth.

5 of 16

Based on 1-D thermal diffusion, we previously applied a large safety

factor with long-pulse testing

Long pulse, M109

Predicted short-pulse result, M109

IFE goal

Scaled goal

Up to 6x104 shots were accumulated for a fluence curve

6 of 16

Long pulse, M109

Short pulse, M109

Predicted short-pulse result, M109

Short-pulse test results exceeded our expectation

The damage fluence is higher than expected from simple scaling

One mirror was used for both 4.5-ns and 25-ns testing

• Damage does not scale like √pulselength (like Tmax)

• Effect of cumulative damage? ∫ dt

7 of 16

Surface temperature effect; Time at temperature may also be a factor

Absorbed heat flux using fixed total energy

Surface temperature (2x energy in Compex)

Short-pulse induced damage occurs at 30% less

fluence, not 50% less as expected.

8 of 16

Our latest Alumiplated mirror (M109) performed extremely well

(long pulse)

Further evidence of variability in coating and surface finishing

Long pulse, M109

Long pulse, M85 (previous best)

9 of 16

A quality coating with good diamond-turning provides much better damage

resistance

Alumiplate has not been a highly reproducible fabrication technique.

Best is 3 nm RMS Roughness, 20 nm P-V

Poor (m80) Good (m85) Best (m109)

10 of 16

CMP may offer a pathway to higher quality and better quality control

• 1 nm RMS Roughness

• 48 nm P-V

Alumiplate mirror with Chemical-Mechanical Polish

11 of 16

CMP mirror damage resistance is comparable to previous Diamond-

Turned mirrors

Damage morphology of CMP is the same as D-T: grain motion in the coating

Damage fluence curves (long-pulse)

12 of 16

We are developing in-house mirror fabrication capability at

UCSD

• 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

13 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

14 of 16

Next-step goals for GIMM R&D15 of 16

16 of 16

Questions?