Transcript 20 TILLACK 23Oct0830mtillack

Progress on GIMM mirror
development and damage testing
Mark Tillack, John Pulsifer,
Sam Yuspeh, Matt Aralis
HAPL Project Meeting
22-23 October 2008
Madison, WI
1 of 13
Our current goal is to test alloy mirrors at
very high shot count to complete our proof
of principle demonstration of laser damage
2 of 13
Progress since April on fabrication and testing:
• Fabrication
•
The cause of triangular damage features seen last time was isolated
and corrected.
•
12-µm thick Al-0.5%Cu alloy mirrors were fabricated by thin film
deposition at UCSD and then CMP finished at Cabot.
• Testing
•
•
•
107 shots have been achieved now @3 J/cm2 with no sign of damage.
A new homogenizer was designed enable higher fluence (8-9 J/cm2).
We remain optimistic that alloy mirrors will exceed damage
requirements.
In April we were suffering in the
“Tillack zone”: the closer we approach the
end, the more difficult the road becomes
unhomogenized
homogenized
3 of 13
The cause of early failure was
determined to be delamination
resulting from improper fabrication
4 of 13
New wafers sent for CMP were damaged
during polishing (thrown from the wheel):
•
Initially suggested cause was high coating
roughness before polishing (P-V of 5 µm).
•
Analysis revealed that the coating had poor
adhesion.
Wrinkle
m119, 3 J/cm2, 5770860 shots
Improved cleaning and handling
procedures result in good coating adhesion
5 of 13
• An expert on metallic coatings at UCSD was consulted
• New wafer cleaning and coating procedures were established
• SiO2 growth is now avoided and contamination reduced
Poor adhesion vs. Good adhesion
A razor can’t peel the coating when adhesion is strong
6 of 13
An additional benefit of the new fabrication
procedure is a higher quality surface
• 12-µm coating of Al-0.5%Cu on each; properly cleaned wafer
results in near specular quality prior to polishing.
Coating on poorly cleaned wafer
Coating on properly cleaned wafer
P-V: 5 microns
P-V: 200 nm
RMS: 0.28 microns
RMS: 16 nm
Coating thickness 12 µm
Coating thickness 12 µm
7 of 13
Grain size remains small in thicker coatings
with 0.5% Cu
Al-1%Cu SEM image
Al-0.5%Cu SEM image (poorly cleaned wafer)
Al-0.5%Cu SEM image (properly-cleaned wafer)
The thicker coating (with poor adhesion)
already shows signs of improvement
(no damage at 107 shots)
8 of 13
In fact, the 12-µm poorly-adhered mirror
surface improved after exposure to
10 million shots @3 J/cm2, without failure
unexposed area,
3.66 nm RMS
500 µm
laser exposed area,
2.4 nm RMS
9 of 13
Our homogenizer is essential to provide 10 of 13
quality data, but has limited the achievable
fluence to inconveniently low levels
• Homogenization provides predictable,
reproducible results
• Depth of field limits achievable fluence
• The current setup is limited to 3-4 J/cm2
11 of 13
Alternative design options for the
optical path have been explored
• Large standoff desired for reasonable
depth of field at the target.
• Existing 2” optics limit fluence to
4 J/cm2 since some light is not
collected (80% energy recovery)
• 3-inch optics (fl=2m and 0.72m) with
same focal geometry will provide
5 J/cm2 with 2.78:1 demagnification
(0.9 mm image), 95% energy recovery
• 3-inch optics (fl=2m and 0.6m) with
3.38:1 demagnification will provide 89 J/cm2 (0.74 mm image)
2.5 x 2.5 mm
12 of 13
Beam path redesign should allow >8 J/cm2
with a reasonable footprint size and profile
At image plane
5 mm away
10 mm away
10% lower
peak fluence
20% lower
peak fluence
• With 2.78:1 demagnification, the footprint is ~0.9 mm x 10 mm.
• Placing the image plane at the center of the footprint, then 5 mm
away is the worst case profile at 2.78:1
• Imaging results suggest that some amount of out-of-focus is
tolerable, such that 8 J/cm2 (or more) is reasonable at 3.38:1
• The footprint will reduce to 0.74 mm x 8 mm
13 of 13
Summary and Future Plans
• Data have been obtained up to 107 shots with
homogenized fluence levels of 3 J/cm2 over 10 mm2.
• Problems with coatings have been resolved and better
mirrors are now being fabricated.
• Testing at 8 J/cm2 or higher is possible with our improved
imaging homogenizer design.
• Testing beyond 107 shots at 8 J/cm2 is planned after CMP
polishing 12-µm thick, well-adhered Al-0.5%Cu mirrors.
• Further study is needed on chamber contamination and
mirror reflectivity issues, reported earlier.