Fiber Lasers at LLNL J.W. Dawson, M.J. Messerly, J.E. Heebner, P.H. Pax, A.K. Sridharan, A. L. Bullington, R.E. Bonanno, R.J. Beach, C.W. Siders and C.P.J. Barty NIF & PS Directorate Lawrence Livermore National Laboratory This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344
57
Embed
Fiber Lasers at LLNL - Stanford University · PDF fileFiber Lasers at LLNL J.W. Dawson, M.J. Messerly, J.E. Heebner, P.H. Pax, A.K. Sridharan, ... c = 300K — h = 10,000 W/(m2 –
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Fiber Lasers at LLNL
J.W. Dawson, M.J. Messerly, J.E. Heebner, P.H. Pax, A.K. Sridharan,A. L. Bullington, R.E. Bonanno, R.J. Beach, C.W. Siders and C.P.J. Barty
NIF & PS Directorate
Lawrence Livermore National Laboratory
This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344
People like fiber lasers because they are a simple-to-use, low-maintenance, compact source of high-brightness, high-power
laser light with wall plug efficiencies in excess of 30%
Double clad fiber amplifiers are conceptually simple devices
3
•Rare earth doped core absorbs pump light from cladding•Light propagating in core stimulates emissions leading to brightness enhancement•Optical fiber core defines output beam quality•High surface area to volume ratio of core minimizes thermal effects•High intensity over long lengths lead to highly efficiency process I>>Isat
•Yb3+ fibers can achieve 85% optical to optical conversion efficiencies
11
Development of injection seed lasers has been a high priority at LLNL. The NIF laser is seeded by a fully rack-mounted state-of-the-art LLNL-engineered 1053-nm fiber
laser system that operates 24/7 under full computer control.
12LLNL-PRES-426924 April 6, 2010
The Advanced Radiographic Capability project will create a parallel short pulse front end for NIF
Photocathode Drive Laser
Messerly—PS&A Technical Review, May 11-12, 2010 13NIF-0510-18892s2.ppt
LLNL’s Mono-Energetic Gamma-Ray project is employing complementary fiber systems
1.8 mm-mrad rms state-of-the-art emittance measured at 800 pC
T-REX Seed Laser System•World’s first fiber-basedphotocathode driver•Frequency and phase-locked to 3 ppm•Foundation for futureaccelerators and FELs
15LLNL-PRES-426924 April 6, 2010
Both systems are based upon a master oscillator power amplifier architecture
MEGa-Ray Fiber Laser ARC Fiber Laser
Due to time constraints I will focus on the ARC system
16LLNL-PRES-426924 April 6, 2010
Challenges the systems face in order to meet their requirements
• Imperfections in the CFBGs
• Non-linear effects
• Pulse contrast
• Dispersion understanding and control
17LLNL-PRES-426924 April 6, 2010
The mode locked oscillator is a self-similar design based on NPE*
18LLNL-PRES-426924 April 6, 2010
Test data from the oscillator looks good
20NIF-0410-18843s2.ppt
Spare oscillator was tested in NIF MOR for
amplitude and synchronization stability
Phan—Photon Science Technical Review, May 12, 2010
ARC Oscillator stability test in NIF MOR room from 01/07/09 to 04/02/10
We have accumulated over 1 year of data
Data acq down
Air conditioning system down
21NIF-0410-18843s2.ppt Phan—Photon Science Technical Review, May 12, 2010
22LLNL-PRES-426924 April 6, 2010
Amplified spontaneous emission (ASE) is the main challenge to high pre-pulse contrast
23LLNL-PRES-426924 April 6, 2010
The requirement for 1-10nJ of clean stretched light complicates the system design
24LLNL-PRES-426924 April 6, 2010
We investigated several pulse cleaning schemes and chose a resonant saturable absorber
24LLNL-PRES-426924 April 6, 2010
25LLNL-PRES-426924 April 6, 2010
Pulse contrast was measured with a photo-diode and attenuators before and after the cleaner
26LLNL-PRES-426924 April 6, 2010
A high resolution FROG was taken of the output pulse at 49nJ
26LLNL-PRES-426924 April 6, 2010
27LLNL-PRES-426924 April 6, 2010
The clean pulse is stretched to 2.5ns in a chirped fiber Bragg grating (CFBG)
27
28LLNL-PRES-426924 April 6, 2010
View of the CFBG/Pulse Cleaner chassis
29LLNL-PRES-426924 April 6, 2010
After the CFBG, the pulse is amplified, split, sent through 120m of fiber and boosted to 100µJ
30LLNL-PRES-426924 April 6, 2010
Set-up for recompressing pulses and measuring results
32
After passage through the full front end and amplification to 58µJ pulse contrast ~78dB
Unsaturated peak voltage 0.272V
OD used 3.75
ASE voltage 0.01V
Scope impulse response 400ps
Estimated 1ps contrast 61,200,000 or ~78dB
31LLNL-PRES-426924 April 6, 2010
FROG data at 800nJ, B~1.4, 512X512, PC=42.3%
25
FROG data at 97µJ, B~6.2, 512X512, PC=24.0%Out[316]=
Out[359]=
Out[368]=
25
High energy, short pulse fiber lasers employing chirped pulse amplification (Summary)• Impressive results in terms of pulse energy (~mJ) and average power
(>>100W) have been reported in the literature and at conferences
• To date, almost all high energy systems have reported around 1 ps pulse widths, but it is not clear the quality of those pulses was all that good
— Most systems have large temporal pedestals— Frequency converting to the UV, trades pedestal for efficiency
• Large B-integrals (a measure of the amount of self phase modulation) is the key contributor to the temporal pedestal
— Understanding and mitigating B-integral issues needs to be a key R&D thrust for fiber lasers going forward
• A secondary, but still important area of concern is dispersion management of a system that is all glass
— This is needed to generate high quality CPA pulses below 1 ps— R&D in compact stretchers with flexible dispersion is a second R&D
thrust for short pulse fiber lasers going forward
• Fiber lasers are a promising future source of high average power short pulses due to gain media with broad band widths and their natural ability to achieve high average powers
26
33LLNL-PRES-426924 April 6, 2010
Ytterbium doped silica optical fibers are reaching their physically-limited output powers
Pump Coupling
Rupture
Melting
Lensing
SRS
SBS
Damage
BendingBending
Eight physical phenomena govern CW fiber laser
power scaling and all can be expressed as f(d,L)
d = core diameterL = fiber length
35LLNL-PRES-426924 April 6, 2010
Power limit contours
Ridge-like upper bound at 36.9 kW, regardless of fiber’s diameter and length
Contour lines units are kW
SRS
Thermal Lens
Pump Power
37
36
SRS
Thermal Lens
Pump Power
LLNL-PRES-426924 April 6, 2010
Ridge at intersection of SRS and thermal limits
Ridge power is independent of diameter and length
37
J.W. Dawson, M.J. Messerly, R.J. Beach, M.Y. Shverdin, E.A. Stappaerts, A.K. Sridharan, P.H. Pax, J.E. Heebner, C.W. Siders, C.P.J. Barty “Analysis of the scalability of diffraction-limited fiber lasers and amplifiers to high average power,” Optics Express, 2008, 16, 13240 - 13266
€
Plaser = 4πηlaser ⋅ k ⋅ λ
2 ⋅ Γ2 ⋅ ln G( )
2 ⋅ηheat ⋅dndT
⋅ gR
€
L = 4a22 ⋅ηheat ⋅
dndT
⋅ Γ2 ⋅ ln G( )
ηlaser ⋅ k ⋅ λ2 ⋅ gR
37LLNL-PRES-426924 April 6, 2010
SBS limits narrowband lasers to lower powers and shorter lengths (< 4m)
SBS The laser power peak of this ridge is independent of core diameter and fiber length
Thermal Lens
Pump Power
SBS suppression fibers having acoustic anti-guides may improve this limit beyond 1.9kW
1.9
€
Plaser = πηlaser ⋅ k ⋅ λ
2 ⋅ Γ2 ⋅ 21⋅ ln G( )
2 ⋅ηheat ⋅dndT
⋅ gB Δν( )
€
L = a2ηheat ⋅
dndT
⋅ Γ2 ⋅ 42 ⋅ ln G( )
ηlaser ⋅ k ⋅ λ2 ⋅ gB Δν( )
38LLNL-PRES-426924 April 6, 2010
Bending perturbs a mode’s size and shape
r
n
Unbent
Bent
Bend-induced distortion of effective area calculated via aBPM method, assuming step index fiber with NA = 0.06.
r
I
UnbentBent
39LLNL-PRES-426924 April 6, 2010
Mode size limited to ~ 40µm, even for large bends
Geometric effect, NOT material-dependent(Fibers shorter than ~1m do not need to be bent)
Red parameters are material–independent.Our proceedings paper lists references for the following tabulated values.
42LLNL-PRES-426924 April 6, 2010
Tm SilicaProperties Units Tm Silica Er Silica Yb
PhosphateYAG Yb Silica
Rupture Modulus W/m 2460 2460
Thermal Conductivity W/(m-K) 1.38 1.38
Melt Temperature K 1983 1983
dn/dT 1/K 1.18X10-5 1.18X10-5
Raman Gain m/W 10-13 10-13
Brillouin Gain m/W 5X10-11 5X10-11
Damage Fluence W/µm2 35 35
Pump Brightness W/(µm2-Sr) 0.018 0.1
Pump Absorption dB/m 450 250
Laser Efficiency -- 0.7 0.85
Heat Fraction -- 0.3 0.1
Pump Clad NA -- 0.45 0.45
Laser Wavelength nm 2040 1078
€
Plaser = 4πηlaser ⋅ k ⋅ λ
2 ⋅ Γ2 ⋅ ln G( )
2 ⋅ηheat ⋅dndT
⋅ gR
43LLNL-PRES-426924 April 6, 2010
Tm Silica is virtually identical to Yb Silica!
Silica-based, so results not changed much. Biggest impact is pump laser brightness(not yet mature at 795 nm).Experiments suggest that heating in Tm is worse than expected (QE < 2:1), so plots may be optimistic.
SRS Limited SBS Limited
Longer wavelength balances higher heat load,leading to no change
36.2 1.9
44LLNL-PRES-426924 April 6, 2010
Er SilicaProperties Units Tm Silica Er Silica Yb
PhosphateYAG Yb Silica
Rupture Modulus W/m 2460 2460 2460
Thermal Conductivity W/(m-K) 1.38 1.38 1.38
Melt Temperature K 1983 1983 1983
dn/dT 1/K 1.18X10-5 1.18X10-5 1.18X10-5
Raman Gain m/W 10-13 10-13 10-13
Brillouin Gain m/W 5X10-11 5X10-11 5X10-11
Damage Fluence W/µm2 35 35 35
Pump Brightness W/(µm2-Sr) 0.018 0.015 0.1
Pump Absorption dB/m 450 20 250
Laser Efficiency -- 0.7 0.85 0.85
Heat Fraction -- 0.3 0.1 0.1
Pump Clad NA -- 0.45 0.45 0.45
Laser Wavelength nm 2040 1590 1078
45LLNL-PRES-426924 April 6, 2010
Er Silica, SRS limited case
SRS Limited: Pump Absorption Issue
Er:SiO2 might reach 50% more power than Tm or Yb due to longer wavelength.
Er:SiO2 combines highest potential power and eye safety. However, work is needed to increase pump brightness (1530-1545nm) AND raise doping concentrations.
SRS Limited: 10X higher brightness
54
54
46LLNL-PRES-426924 April 6, 2010
Er Silica, SBS limited case
SBS Limited: Pump Absorption Issue SBS Limited: 10X higher brightness
2.8
2.8
Er:SiO2 might reach 50% more power than Tm or Yb due to longer wavelength.
Er:SiO2 combines highest potential power and eye safety. However, work is needed to increase pump brightness (1530-1545nm) AND raise doping concentrations.
47LLNL-PRES-426924 April 6, 2010
Yb PhosphateProperties Units Tm Silica Er Silica Yb
PhosphateYAG Yb Silica
Rupture Modulus W/m 2460 2460 70 2460
Thermal Conductivity W/(m-K) 1.38 1.38 0.49 1.38
Melt Temperature K 1983 1983 723 1983
dn/dT 1/K 1.18X10-5 1.18X10-5 -5.1X10-6 1.18X10-5
Raman Gain m/W 10-13 10-13 0.5-20X10-13 10-13
Brillouin Gain m/W 5X10-11 5X10-11 2.5X10-11 5X10-11
Damage Fluence W/µm2 35 35 6.5 35
Pump Brightness W/(µm2-Sr) 0.018 0.015 0.1 0.1
Pump Absorption dB/m 450 20 5200 250
Laser Efficiency -- 0.7 0.85 0.6 0.85
Heat Fraction -- 0.3 0.1 0.4 0.1
Pump Clad NA -- 0.45 0.45 0.64 0.45
Laser Wavelength nm 2040 1590 1053.7 1078
48LLNL-PRES-426924 April 6, 2010
Yb PhosphateProperties Units Tm Silica Er Silica Yb
PhosphateYAG Yb Silica
Rupture Modulus W/m 2460 2460 70 2460
Thermal Conductivity W/(m-K) 1.38 1.38 0.49 1.38
Melt Temperature K 1983 1983 723 1983
dn/dT 1/K 1.18X10-5 1.18X10-5 -5.1X10-6 1.18X10-5
Raman Gain m/W 10-13 10-13 0.5-20X10-13 10-13
Brillouin Gain m/W 5X10-11 5X10-11 2.5X10-11 5X10-11
Damage Fluence W/µm2 35 35 6.5 35
Pump Brightness W/(µm2-Sr) 0.018 0.015 0.1 0.1
Pump Absorption dB/m 450 20 5200 250
Laser Efficiency -- 0.7 0.85 0.6 0.85
Heat Fraction -- 0.3 0.1 0.4 0.1
Pump Clad NA -- 0.45 0.45 0.64 0.45
Laser Wavelength nm 2040 1590 1053.7 1078
49LLNL-PRES-426924 April 6, 2010
Yb PhosphateSRS Limited (gR=20X10-13 m/W)
Phosphates don’t dissipate heat well, so melting and damage are issues.
SRS Limited (gR=0.5X10-13 m/W)
Damage
Melt
19
3
50LLNL-PRES-426924 April 6, 2010
Yb PhosphateSRS Limited (gR=20X10-13 m/W)
Phosphates don’t dissipate heat well, so melting and damage are issues.High loss (~3dB/m*) is also a concern
SRS Limited (gR=0.5X10-13 m/W)
SBS Limited
Damage
Melt
Melt
1
19
3
* Lee, Digonnet, Sinha, Urbanek, Byer, IEEE J. Selected Topics in QE 15, 93-102 (2009)
51LLNL-PRES-426924 April 6, 2010
Losses lead to over-riding efficiency term, ηfiber
Allowed length vs. fiber loss
Phosphates limit lengths to ~ 1 m
For absorption losses, deposited heat must also be considered.
silica
phosphates
(Assume gain = 10 dB) η’s
52LLNL-PRES-426924 April 6, 2010
Yb Phosphate length ~ 1 m, due to losses(not SBS or SRS)
At 3 dB/m, loss limits power to <500 WAt 1 dB/m, loss limits power to <1 kWLimits are Damage, Melting and Thermal Lens
1m SRS Limited 0.5m SRS Limited
SRS Melt
Thermal
53LLNL-PRES-426924 April 6, 2010
Yb YAG (single crystal & ceramic)Properties Units Tm Silica Er Silica Yb
Yb YAG (1 of 2)SRS Limited (Broadened to 10 nm)SRS Limited (1 nm)
SRS is more pronounced in YAG than SiO2, though bandwidth broadening may overcome this.Loss and fiber flexibility are issues, leading to length restrictions which may ultimately limit power to 10’s of kW.
Damage Damage
105
33
56LLNL-PRES-426924 April 6, 2010
Yb YAG (2 of 2)
SBS limit in YAG may be high, but if it exceeds the SRS limit, the latter dominates. On the right, we only consider SBS values that match the SRS limit (~ 10-13 m/W)Loss and fiber flexibility are issues, leading to length restrictions which may ultimately limit power to 10’s of kW.
SBS Limited (Upper limit SBS gain [5*10-12 m/W])
SBS Limited (At SRS Bandwidth broadened gain [10-13 m/W])
Damage
120
17
57LLNL-PRES-426924 April 6, 2010
Lineouts of YAG based contour plots at 1m
If losses are < 1 dB/m, then SRS-limited and SBS-limited lasers are both limited to 10’s of kW. For SBS-limited (narrowband) lasers, this is a significant improvement.
1m SRS Limited 1m SBS Limited
SBS
Dam
age
Thermal Lens
58
Pulse energy limits in fibers appear to have been reached
SRS
Extractable EnergyDamageBending
Kerr Lens Self Focusing
4MW hard limit
We believe that any further scaling of fiber laser average power and pulse energy will require fundamental changes in the fiber
itself. This will require the ability to make
new fibers.
59LLNL-PRES-426924 April 6, 2010
We have the technical expertise to draw fibers and the capability can be obtained for a modest cost
• Tower features
—Tower height: 8.2 m
—Preforms up to 1 m long and 50 mm dia.
—Fiber from 80 to 500 µm
—UV single acrylate coating
—Tractor for pulling rods and canes from 0.5 to 2.0 mm with automated cane cutter
—Pressure control system with 2 control points capable of vacuum to +100 kPa
—2nd pyrometer enabling tower to draw soft glass fibers
• Tower Cost: $600K
• Facility Cost: ~$600K
• Will provide the ability to make new waveguide designs in a 1-2 week timeframe depending upon the complexity of the design
Draw tower
Quartz rods and tubes and even Yb3+-doped starting glass is easily obtainable commercially
There are a number sources of raw materials including Momentive, Schott and Kigre
Photonic crystal fibers will be made by the “stack and draw” technique
We intend to employ this capability to investigate waveguide designs with the potential to scale power and pulse energies
beyond the limits we have computed for simple cases
Future directions
•Short pulse lasers—Plan to deploy ARC system on NIF—Developing improved systems for MEGa-Ray and LBNL
•Power and energy scaling—Have internally funded R&D program to investigate new
waveguide designs—Have externally funded program to look at beam combining of
pulsed sources
•Long term goal: Increase fiber laser power and pulse energy while shortening output pulse width in order to enable sources capable of addressing new science applications such as X-ray, EUV seed sources and laser based particle acceleration