A Joule of Light - University of California, San Diego
Download
Report
Transcript A Joule of Light - University of California, San Diego
UV Laser-Induced Damage to Grazing
Incidence Metal Mirrors
M. S. Tillack, J. E. Pulsifer, K. Sequoia
Mechanical and Aerospace Engineering Department and
Center for Energy Research
3rd International Conference on
Inertial Fusion Science and Applications
Monterey, CA
9 September 2003
Design concept for a power plant GIMM*
85Þ
The reference mirror concept consists
of stiff, light-weight, radiation-resistant
substrates with a thin metallic coating
optimized for high reflectivity
(Al for UV, S-pol, shallow q)
~50 cm
* Sombrero and Prometheus studies, ca. 1992.
Key issues were identified for a GIMM*
• Shallow angle stability
• Damage resistance/lifetime
Goal = 5 J/cm2, 108 shots
• Fabrication & optical quality
• Contamination resistance
• Radiation resistance
S-N curve for Al alloy
* R. L. Bieri and M. W. Guinan, Fusion Technology 19 (1991) 673-678.
We tested several Al fabrication options
• Thin films on superpolished substrates
– CVD SiC, 2-3Å roughness, 2-3 nm flatness over 3 cm
– magnetron sputtering up to 250 nm
– e-beam evaporation up to 2 mm
• Solid polycrystalline metal
– polished
– diamond-turned
• Electroplated and
turned Al
Testing was performed with 25-ns, 248-nm
pulses in a controlled environment
420 mJ, 25 ns, 248 nm
viewing port
beam diagnostics
dump
cube
1/2 waveplate
specimen
mount
cube
dump
In-situ monitoring helped identify
the onset of damage
• Brightfield beam profiling
• Darkfield beam profiling
• Surface imaging dump
test specimen
camera
profiler
microscopy
in-situ imaging
translation
main beam
probe laser
darkfield
Polycrystalline Al is easy to fabricate into a
mirror, but has large grains
• 1-mm 99.999% pure Al, bonded with CA to 3-mm thick Al alloy
• Polished with 5, 1, and 0.04 mm alumina (Al2O3) suspension, or
• Diamond-turned on precision lathe (at GA target fab facility)
~25 nm avg. roughness
Polished Al damages due to plastic
deformation mechanisms
• Exposed for 100 shots in vacuum at 2–5 J/cm2
• Grain boundaries separate
• Slip lines extrude within grains
500 X
500 X
Diamond-turned Al exhibits superior
damage resistance
•
•
•
•
Exposed for 50,000 shots in He at 3–4 J/cm2
No obvious damage
Minimal (if any) grain boundary separation
Polishing appears to introduce impurities and pre-stress the
grain boundaries
Thin film deposition is limited by coating
thickness and surface defects
• Thermal stress, constraints on thickness
• Added complexity of substrate and film requirements
•
•
•
•
Plane stress analysis
10 mJ/cm2 absorbed
Peak stress at interface ~40 MPa
Yield stress is 10-20 MPa
1 mm coating of Al on SiC
Surfa ce
I nte rfa ce
SiC (0.5 um )
SiC (1 um )
SiC (2.5 um )
SiC (5.0 um )
335
330
325
320
315
310
305
300
0.E+00
1.E-08
2.E-08
3.E-08
Time, s
SiC: 10 mm
q”=10 mJ/cm2
300 nm Coating
340
Temperature, K
Al: 20-500 nm
4.E-08
5.E-08
6.E-08
Good coatings were obtained using
superpolished CVD SiC substrates
• Superpolished CVD SiC: 2-3 Å smooth, 2-3 nm flat
• Thin film deposition of Al by magnetron sputtering
and/or e-beam evaporation
• Up to 2 mm Al has been successfully deposited by e-beam
Q uic kTim e™ and a TI FF ( Unco m pr ess ed) deco m pr esso r ar e ne eded t o see t his pict ur e.
Qu ic k Ti me™ and a TI FF (U nc o mpre s s ed ) de c omp res s or a re ne eded to s ee th is p ic tur e.
Thin films are delicate, damage easily and
catastrophically
250 nm e-beam
23,000 shots @4 J/cm2
1.5 mm e-beam
86,000 shots @4 J/cm2
Electroplated Al solves problems with
coating thickness and large grains
• 50-100 mm Al on Al-6061 substrate
• 100,000 shots at 3-4 J/cm2
• No discernable change to the surface
Summary
• Survival above 100,000 shots has yet to be
demonstrated in thin film coatings – damage occurs
due to imperfections and high interfacial stresses.
• Thicker coatings appear to be more robust, but
detrimental effects of grain structures must be avoided.
• Thick (>50 mm) electroplated Al on SiC provided the
best damage response, due to thickness of coating and
small grains. Scale-up and further testing are planned.
Acknowledgements
Thanks to the following for their advice and technical support:
Jim Kaae et. al, General Atomics microfabrication facility
Ed Hsieh et. al, Schafer Corp.
Lee Burns, Rohm & Haas Co. Advanced Materials
Witold Kowbel, MER Corp.
Larry Stelmack, PVD Products, Inc.
John Sethian and the members of the High Average Power Laser Program
This work was funded by US DOE/DP NNSA
Optic scale-up: multiplexed beams enable
smaller, more tolerant final optics
LONG PULSE AMPLIFIER
(~ 100's nsec)
Last Pulse
Demultiplexer
Array
(mirrors)
Multiplexer
Array
(beam
splitters)
Target
FRONT END
( 20 nsec)
Only three pulses shown for clarity