High power fibre lasers

ORC postgraduate lecture

March 9 2011

Johan Nilsson
Optoelectronics Research Centre
University of Southampton, England

www.orc.soton.ac.uk/hpfl.html

High-power fibre lasers

J. Nilsson, ORC postgraduate lecture, March 9 2011

Why fibre lasers?

• Tight confinement of pump and signal beams

• Long interaction length
– RE-concentrations at quench-free level
– Low heating per unit length
• Control of the propagation properties by the waveguide
– Beam profile / single-mode operation
– Dispersion tailoring
– Distributed waveguide filtering
• High gain
• High power
• High nonlinearities
– Sometimes good, often bad
– High gain & high nonlinearities
“interesting physics”

High-power fibre lasers 2

J. Nilsson, ORC postgraduate lecture, March 9 2011

1
 Fibre laser vs. “bulk” (rod) laser
Optically pumped

Fibre

Tight beam confinement and maximum intensity over arbitrary length.

Pump
Low threshold

High gain efficiency
beam

Even three-level low-concentration systems lase (erbium-doped
fibre amplifiers)

Bulk: Not possible to

maintain tight focusing
over adequate length

High threshold

Pump
beam Laser crystal
High-power fibre lasers 3
J. Nilsson, ORC postgraduate lecture, March 9 2011

Example: Raman amplification (I)

GNp = gR Leff Ip = gR Leff Pp / Aeff

gR Raman gain coefficient (material property)

Leff effective length
Ip pump intensity
Pp pump power
Aeff effective beam area of the beam
GNp gain in nepers: Glin = Pout / Pin = exp (GNp).

Long interaction length and small beam area improves Raman gain.
– Contradiction in bulk amplifier because of DIFFRACTION

Pump
beam Raman amplifier
High-power fibre lasers 4
J. Nilsson, ORC postgraduate lecture, March 9 2011

2
 Example: Raman amplification (II)
• Gain efficiency in Raman bulk amplifier:

GNp / Pp = gR ∫ Aeff-1 (z) dz

= gR ∫ {π ω02 [1 + (λ z)2 / (π ω02 n)2]}-1 dz

= gR π n / λ = 0.43×10-6 W -1 = 1.9 dB/MW

(fused silica at 1060 nm, gR = 1×10-13 m/W, loss neglected)

– Integration extends to infinity but most gain occurs effectively within the
Rayleigh length of the beam

• The Raman gain efficiency in bulk silica is much too low for
significant gain (cw).
-- MW of pump power required for high gain

High-power fibre lasers 5

J. Nilsson, ORC postgraduate lecture, March 9 2011

Example: Raman amplification (III)

• For a fibre, the effective length is determined by the background
loss α [Np/m]

Leff = (1 – exp (–α L)) / α ≈ α-1 for long lengths L

GNp / Pp = gR Leff / Aeff = 14 W -1 = 63 dB/W

at 1060 nm with typical parameters for silica fibre

-- effective area 30 µm2, loss 1 dB/km

• The Raman gain efficiency in fibre is 7.5 orders of magnitude

higher than in bulk silica!

Conclusion: A Raman amplifier requires the tight beam

confinement and large effective length of a fibre
(Or, a very high Raman cross-section)
High-power fibre lasers 6
J. Nilsson, ORC postgraduate lecture, March 9 2011

3
Example: Kilowatt-level Yb-doped laser
• Yb-doped lasers are attractive for their high slope efficiency
-- Optically pumped
• Three-level system so threshold can be high
-- Need high Yb excitation level
• Must be at least 10 cm long for thermal management
-- Ten meter scale more suitable for fibres
• The poor beam quality of kilowatt-level pump diodes increases the pump area
-- Threshold proportional to pump area
-- Replace λ by (M2 λ) in formulas 2
z R M pump λ pump
for Gaussian beam propagation
w pump
=
-- Area proportional to M2
-- For example, M2 = 200
0
π
2
• Average effective pump beam area becomes 7.75 mm over 10 cm
• Threshold in the kW-regime for bulk laser, because of large beam area

Pump waveguiding is required to reduce area & threshold
-- Threshold a few watts for kW fibre laser
• Waveguiding is very useful for lasers
• Also bulk lasers can use waveguiding to reduce the pumped area to, e.g., 1 mm2
• The thin-disk laser is another attractive alternative in the high-power regime, which also
avoids many of the draw-backs of rod-lasers.

High-power fibre lasers 7

J. Nilsson, ORC postgraduate lecture, March 9 2011

Fibres enable
• Very long effective lengths in Raman amplifiers
– 7.5 orders of magnitude improvement of the gain efficiency over bulk
amplifier of same material

• EDFAs with low concentrations but still high gain and gain efficiency
– 3.5 orders of magnitude improvement of the gain efficiency over bulk
amplifier of same material
• EDFAs need to be long

• Low thermal load per unit length in high-power fibre lasers

• Pump beam confinement for low threshold
-- watt-scale threshold in cladding-pumped Yb-doped fibre lasers
-- kW-scale threshold in Yb-doped bulk lasers

Waveguiding or short interaction length essential for Yb lasers
– In practice also Yb-doped bulk lasers use a waveguide for the pump

• Rare-earth doped fibres provide a superb combination of the

excellent properties of rare earths and optical fibres.

High-power fibre lasers 8

J. Nilsson, ORC postgraduate lecture, March 9 2011

4
 Material matters too!
Glass vs. crystal hosts
Property Glass Crystal Comment

Absorption Large Narrow Large BW good for diode pumping

bandwidth
Peak absorption Low High High absorption reduces threshold and device length

Emission BW Large Narrow Broadband amplifiers and tunability

Peak emission Low High Threshold, important for bulk lasers

Thermal Low High Very important for bulk lasers

conductivity
Fibre drawing Easy Difficult

Crystals for bulk, glasses for fibre, fibres for high-gain amplifiers!!
• Table is primarily for rare earth laser ions.
• Transition metals have broad emission spectra even in crystals.
• Transition metals very difficult in glass hosts
High-power fibre lasers 9
J. Nilsson, ORC postgraduate lecture, March 9 2011

Performance advantages
• Unique combination of high gain, high power, high efficiency,
and broad bandwidth
– high gain, high fidelity, and broad bandwidth crucial for telecom
– MOPAs for sophistication and high-power
• High speed control and versatility
• Ytterbium-doped fibres are exceptionally efficient (80%)
• Superior signal amplifiers
– EDFAs have no serious competition

High-power fibre lasers 10

J. Nilsson, ORC postgraduate lecture, March 9 2011

5
 Scalability to high average power
(kW-level) is primary motivation for
current research on fibre lasers

• But, never forget telecom! Erbium-doped fibre

amplifiers are extremely important for
telecommunications at much lower powers, e.g.,
0.1 W.

High-power fibre lasers

J. Nilsson, ORC postgraduate lecture, March 9 2011

17 kW, Yb, MM

Tremendous rate of 10 kW, Yb, MM 3 – 6 kW SM (IPG)

2 kW single-mode (IPG)
2 kW, Yb, MM
progress 1.5 kW (Jena)
1.4 kW single-mode (ORC)
At ORC: MM 1.3 kW, M2 < 3 (Jena)
1.2 1.2 kW single-mode (ORC)
• 2.1 kW YDFL (1100 nm)
1.1
• 0.3 kW 1570 nm Er:Yb co-1 kW, Nd, MM 1 kW, Yb, MM (ORC)
1.0 0.8 kW TDFL
doped fibre laser @ 2 µm (Q-peak)
0.9
• 100 W 2 µm thulium-doped
0.8 kW, Yb, SM
Laser power [kW]

0.8 fibre laser (Dual-end output)

0.7 • 4 W 980 nm YDFL
0.6 610W, Yb SM (ORC) 0.6 kW, SM, polarized (ORC)
• 4 W 920 nm NDFL 0.4 kW SF, SM,
0.5 485W, Nd/Yb, SM
SP Yb MOPA
0.4 352W, Yb, SM (ORC)
366 W JAC Yb (ORC) 0.3 kW EYDFL
0.3 272W, Yb (ORC) 280 W, JAC Yb PCF
135W, Yb, SM 8.4 mJ, 120W, Yb (ORC)
0.2 150W, Nd/Yb 120W, Yb, SF 160 W Er/Yb
0.32 kW, 20 ps,
0.1 110, Yb, SM (SDL) Yb MOPA
30W, Nd 20W, Yb, SF 80 W Tm. 2 µm
10W, Yb, SF 264 W SF, SM,
0.0 87 W, Er/Yb, SFSP Yb MOPA
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 MOPA (ORC)

Year 1.2 mJ, 380 fs, Yb 2.8 mJ, 50/100W, Yb 150 W SF, SM,
7.7 mJ, 10W, Yb (ORC)High-power fibre lasers
Er:Yb MOPA 12
J. Nilsson, ORC postgraduate lecture, March 9 2011

6
 Why high power fibre lasers?
• With more power we can do more!
– Machining, welding, sensing...
• Power is not enough
– 100 W enough for lots of things, but... don’t try this with a light bulb
• Power must be “concentrated” and ideally controllable
– Spatially
– Spectrally
– Temporally (pulsed)
– Brightness
• Fibres are very good for this at high powers
– Long
– Waveguide
• Robust, reliable, flexible, compact, light-weight, manufacturable
– Leverage heavy telecom investments in laser diodes and fibres

High-power fibre lasers 13

J. Nilsson, ORC postgraduate lecture, March 9 2011

Basics
High power fibre sources are:
• Cladding-pumped by multimode laser diodes
• Doped with rare-earth laser ions
• Silica-based
(with few exceptions)

• The fibre converts the relatively low-brightness beam

from a diode source to a much brighter beam with
somewhat lower power

High-power fibre lasers

J. Nilsson, ORC postgraduate lecture, March 9 2011

7
Example: Cladding-pumped fibre amplifier
• Optical fibre doped with rare earth in the core
• The RE ions are optically pumped to an excited state by laser diode
The excited RE ions generate / amplify light via stimulated emission
• Pump coupler combines signal and pump

Signal out
Double-clad
RE-doped fibre
Signal in Pump
diode

Pump coupler fibre

isolator isolator
• All fibre / all fiberized light-path
• Robust
High-power fibre lasers 15
J. Nilsson, ORC postgraduate lecture, March 9 2011

Example: Schematic end-pumped

rare-earth-doped fibre laser
Dichroic mirror for Cavity feedback from
pump / signal Lens perpendicular fibre cleave
combination (4% Fresnel reflection)

Cavity mirror Rare-earth Laser

for external doped fibre output
feedback Pump diode Pump diode

• In this example, fibre is really only the gain medium

• Similar to conventional bulk laser such as Nd:YAG
• Far from the ideal “all-fibre” laser without free-space paths
• Still, the fibre can provide many advantages!

High-power fibre lasers 16

J. Nilsson, ORC postgraduate lecture, March 9 2011

8
 Example: Amplifiers in different
pumping configurations
Dichroic mirror Lens
End-pumped

Signal in Rare-earth Signal out

doped fibre
Pump diode Pump diode

Pump
Side-pumped
couplers Splice • Pump not in same mode
Signal in Signal out as signal
(fibre) (fibre)
• “Space multiplexing”
• “All-fibre”
Pump diode
• Allows for continuous
Pump diode fibre path

High-power fibre lasers 17

J. Nilsson, ORC postgraduate lecture, March 9 2011

Key elements

• Fibre
– Double-clad for cladding-pumping

• Pump source
– Laser diodes

• Pump launch

High-power fibre lasers 18

J. Nilsson, ORC postgraduate lecture, March 9 2011

9
 Cladding-pumped fibre Pump launch:
End- or side-
pumping
Outer cladding:
Low-index polymer-
coating or all-glass
structure

Laser signal
at λs

Inner cladding:
Multimode waveguide to Diode pump
capture pump radiation beam at λ p

Core: Silica glass

RE-doped silica glass Long and thin
May be single-mode 10 m x 0.5 mm
• Cladding-pumping always leads to spatial concentration of light,
from multimode diode to (nearly) diffraction-limited fibre laser output
• Enhancement of (spatial) brightness
High-power fibre lasers 19
J. Nilsson, ORC postgraduate lecture, March 9 2011

All fibres made

at ORC MCVD setup

• Preform & fibre fabrication

– MCVD for silica
– Extrusion, casting, etc., for soft glass
– Fibre drawing
• Post-fabrication
Draw tower processing
– tapers
– splicing
– side-splicing
– gratings

Sophisticated low loss fibre structures with high
power capability

One week lead time

High-power fibre lasers 20

J. Nilsson, ORC postgraduate lecture, March 9 2011

10
 Fibres for high power
Large core / large mode area Multi-core ribbon - Scalable!!
Multi-

Micro-structured
Micro- Helical core fibre
Circular birefringence
“holey” fibre
Filtering
Air core

Birefringent Multi-
Multi-core
Air clad Photonic bandgap (stress)
High pump-NA Delivery, pulse compression,
gas-filled devices J. Nilsson, ORC postgraduateHigh-power fibre lasers
lecture, March 9 2011
21

Diodes
Diode stacks
0.2 – 1 kW, 808, 915, 940, 980 nm
Alternatively, use LOTS of
single-emitter diodes
with combining network!

Power up to ~ 10 W each

JDSU
High-power fibre lasers 22
J. Nilsson, ORC postgraduate lecture, March 9 2011

11
 Diodes (II)
• High power
• Low cost (< £5/W)
• High efficiency
– 80% reported
– 50% – 60% standard pigtailed

• But, multimode = “low” brightness

• Also, don’t perform well pulsed

High-power fibre lasers 23

J. Nilsson, ORC postgraduate lecture, March 9 2011

Brightness: key property of lasers

• Spatial brightness
• “Townes” suggested as unit for brightness
– David Hanna, LASE (Photonics West) plenary presentation, 2005
– After C. H. Townes, co-inventor of the laser
• Related to focusability (beam quality, M2) and intensity

P P P P θ
B= = 2 2 2= 2 = 2 2
ΩA π w θ λ (M ) π w NA 2
2 2 2w

P
= I = BΩ P: Power MM diodes:
A A: Area at focus High power

B=
I Ω: Solid angle Poor beam quality
Ω I: Intensity
Limited brightness

High-power fibre lasers 24

J. Nilsson, ORC postgraduate lecture, March 9 2011

12
 Five orders of magnitude brightness-
enhancement possible with cladding-pumping
• Launch low-brightness high-power diode into inner cladding of fibre!
• The converted output beam emerges from a much smaller core with
much smaller NA than the inner cladding
– Area up to 1000 times smaller in Yb-doped fibre laser
– NA (angles) up to 10 times smaller

Brightness enhancement of over 105 possible in Yb-doped fibre laser
– Ideally M2out = 1; M2pump > 300 possible
2 2
Bout  Pout   w pump NApump  w NA 
= ⋅   = η  pump pump 
B pump  Ppump   rcore NAout   rcore NAout 
Core Inner-cladding
Output

Pump

Outer-cladding Fibre laser

High-power fibre lasers 25

J. Nilsson, ORC postgraduate lecture, March 9 2011

Why fibre for high brightness? Thermal effects:

Enemy #1 for high brightness lasers
Laser crystal Gaussian
(Nd:YAG) laser field
Radial
• Thermal gradients & differences heat flow

stress fracture
– Limits maximum power

Brightness Often
More Important Than
Power

High-power fibre lasers 26

J. Nilsson, ORC postgraduate lecture, March 9 2011

13
 Why fibre for high brightness? Thermal effects:
Enemy #1 for high brightness lasers

Nd:YAG

• Thermal gradients & differences

stress fracture
∆Tcompress Equilibrium,
– Limits maximum power volume-
– Thermal expansion
tensile averaged,
∆Ttensile temperature
and compressive stress

• Parabolic temperature profile in

conventional lasers (Nd:YAG)
– ∆Tmax = ∆Tcompress = ∆Ttensile
∝ power x diameter2 / length

High-power fibre lasers 27

J. Nilsson, ORC postgraduate lecture, March 9 2011

Why fibre for high brightness? Thermal effects:

Enemy #1 for high brightness lasers
Phase front of
aberrated beam

• Aberrated thermal lensing

degraded beam quality
– (degraded brightness)
– dn/dT

“Thermal lensing ruined my life”

– Andy Clarkson

High-power fibre lasers 28

J. Nilsson, ORC postgraduate lecture, March 9 2011

14
 No stress fracture in fibres!
Core Cladding • 10 times thinner
100 • 100 times longer

10,000 times better!!
Temperature increase (K)

80
∆Tcompress
60
• Also temperature
∆Ttensile
40 profile is better!
20 – Stronger in compression
than in tension
0
-400 -300 -200 -100 0 100 200 300 400
Position (microns)

High-power fibre lasers 29

J. Nilsson, ORC postgraduate lecture, March 9 2011

No aberrated thermal lensing in fibres!

• Thermally induced phase shift
across core
∆φ ~ ∆T (dn/dT) L ~ P 100
Temperature increase (K)

– Independent of geometry! 80
– Fibre doesn’t help??
60

• The waveguide “picks up” any 40

transversally dependent phase shift
20
– Beam remains guided mode,
with plane phase front 0
orthogonal to Poynting vector -400 -300 -200 -100 0 100 200 300

– No aberrated phase distortions! Position (microns)

• If the waveguiding is sufficiently

strong!

High-power fibre lasers 30

J. Nilsson, ORC postgraduate lecture, March 9 2011

15
 By and large, fibres do away with the
worst enemy of high power lasers!
• Superb thermal properties make
fibre lasers power scalable to several kilowatts with diffraction-
limited beam quality

• Often, we can simply forget about thermal effects

– Highest power fibre lasers are not limited by thermal effects
(as far as I know)

• Thermal damage to coating most critical thermal limitation?

• Long length But thermal effects

• Waveguide are starting to reappear
as power levels
• Can’t have long length without waveguide keep on increasing

High-power fibre lasers 31

J. Nilsson, ORC postgraduate lecture, March 9 2011

Multi-kilowatt single-mode
Yb-doped fibre laser
Output power
Fibre: T0120L30087
Output spectrum
2500 Core D/ Cladding D: 50 µm / 850 µm -20
Core NA: 0.06
T0120L30087
L: 20 m
-30 Output @2.1 kW
2000 JTO Contract No.
Absorption: 1 dB/m @976 nm
Pump: 978 nm + 975 nm
FA9451-06-D-0014
Signal: 1095 nm -40
Laser power [W]

1500 Slope efficiency: 74%

Power [dB]

2
M : 1.2 -50
1.2 kW pump
1000
-60

500
June 2007 -70
2.2 kW pump
0 -80
0 500 1000 1500 2000 2500 3000 1000 1020 1040 1060 1080 1100 1120 1140
Launched pump power [W] Wavelength [nm]

• Maximum output (> 2.1 kW) was limited by available pump power
– Not limited by thermal effects!
• Diffraction limited beam quality: M2 = 1.2
– Five orders of magnitude brightness enhancement
• Excellent power handling indicates higher power possible
– 10 kW?
High-power fibre lasers 32
J. Nilsson, ORC postgraduate lecture, March 9 2011

16
 Multi-kW arrangement
Spatial beam combination > 2 kW of output power!
M-HR-P Signal output

Len-1
M-HR-S M-HR-S
M-HR-P Len-2
2×Diode stack M-HR-P
@978 nm, 2×1.1 kW
SPI

M-HR-P M-HR-S
Double-clad M-HR-P
Yb-doped fibre Diode stack
@975 nm, 1.2 kW
ORC

Power supply & Power supply &

Diode controller Diode controller

High-power fibre lasers 33

J. Nilsson, ORC postgraduate lecture, March 9 2011

2 kW YDFL / MOPA
IPG, 2005
“2 kW CW ytterbium fiber laser with record diffraction-limited brightness”, V. Gapontsev, D. Gapontsev,
N. Platonov, O. Shkurikhin, V. Fomin, A. Mashkin, M. Abramov, S. Ferin, CLEO-E 2005, CJ-1-1-THU

YDF: 10 m, 10.6 µm MFD YDF: 6 m, 12 µm MFD YDF: 9 m, 14 µm MFD

Grating Pump 490 W @
coupler Splice
1090 nm 1040 W 1960 W

LD LD LD LD LD LD LD LD LD LD LD LD

20 W x 36 976 nm 20 W x 36 976 nm 20 W x 72 976 nm

LDs in total LDs in total LDs in total

Not really laser, but cascade of low-gain amplifiers (7 dB)

Total pumping 144 x 975 nm LDs with 20 W of power into 100 µm pigtail
-- Total pump power ~ 2880 W for 2 kW output power

SPI Lasers uses similar configuration

High-power fibre lasers
J. Nilsson, ORC postgraduate lecture, March 9 2011

17
 Yb-doped fibre laser power progress

10000
Wavelength ~ 1.1 µm

Tandem-
Telecom
Power [W]

pumping
1000 boom &
bust

100 6.6 dB/year

(Nd) x 10 increase in 1.5 years!

10
1996 1998 2000 2002 2004 2006 2008 2010 2012
Year
• Diode-pumping up to 2 – 2.5 kW of output power
– Limited by available pump power (at least at ORC...)
• Tandem-pumping used for higher powers High-power fibre lasers
J. Nilsson, ORC postgraduate lecture, March 9 2011

Tandem-pumping
• One or several (diode-pumped) fibre lasers pump another
fibre laser or amplifier
– J. D. Minelly, R. I. Laming, J. E. Townsend, W. L. Barnes,
K. P. Jedrzejewski, D. N. Payne, “High-gain fibre power
amplifier tandem-pumped by a 3 W multi-stripe diode”, Proc
OFC, pp. 32-33, 1992
• Er:Yb co-doped fibre amplifier pumped by Nd-doped fibre laser
Diode
Diode
pump
pump – Also possible to cladding-pump a YDFL by short-
wavelength YDFLs cladding-pumped by diodes

High-brightness
YDFLs
pumping
Large core cladding-pumped
Yb final stage amplifier
(brightness convertor)
Diode
pump Diode
pump

High-power fibre lasers

J. Nilsson, ORC postgraduate lecture, March 9 2011

18
 Tandem-pumping:
YDFL pumped by YDFLs
• Pump wavelength 1010 – 1050 nm
– IPG used 1018 nm
• Absorption cross-section is very low so inner-cladding-
to-core area ratio must be quite small
– Increases average dopant concentration experienced by pump
– High-brightness pump sources required
• The quantum defect can be as low as a few percent
which allows for a reduced thermal load

High-power fibre lasers

J. Nilsson, ORC postgraduate lecture, March 9 2011

IPG YLS-5000-SM & YLS-10000-SM

optical scheme
Master oscillator Power amplifier
Diode-pumped YDFL Tandem-pumped YDFA
Output 1 kW @ 1070 nm Output 5 kW @ 1070 nm
YDF: 15 m long with
30 µm diameter, 0.2 NA core
YDF Effective mode area 500 µm2
Grating Pump
coupler

1 kW @ 5 kW
Pump-LDs Pump-LDs 1070 nm Pump
combiner
YDFL

YDFL
YDFL

Pump-YDFLs
300 W @
1018 nm

High-power fibre lasers

J. Nilsson, ORC postgraduate lecture, March 9 2011

19
 IPG YLS-5000-SM characteristics

From IPG

High-power fibre lasers

J. Nilsson, ORC postgraduate lecture, March 9 2011

Ytterbium-doping is
best for power-scaling
• Highest fibre power (multi-kW)
• Highest efficiency (> 80% slope efficiency)
– Reduces heat load
• Simple spectroscopy
• Emission wavelength 1 – 1.1 µm
• Pump wavelength 0.9 – 1 µm 3e-24

3e-24
• Yb can be incorporated in Aluminosilicate
high concentrations
Cross-section [m ]
2

Phosphosilicate
2e-24 Boro-aluminosilicate

2e-24
• Allows for tandem-pumping Absorption Emission
for “ultra-low” thermal load 1e-24

5e-25

0
850 900 950 1000 1050 1100 1150
Wavelength [nm]

High-power fibre lasers 40

J. Nilsson, ORC postgraduate lecture, March 9 2011

20
 Pump launch

End-pumping
Side-pumping
• Primarily for low-brightness pumps
• Not always suitable for high-brightness pumping
(tandem-pumping)

High-power fibre lasers

J. Nilsson, ORC postgraduate lecture, March 9 2011

End-pumping
Focusing lens Rare-earth-doped
core Outer cladding

Pump
beam

Inner cladding
Pros and cons of end-pumping
+ Simplest approach
+ Straightforward fibre fabrication, preparation and setup
+ High efficiency
+ Minimal pump brightness degradation
– At most two pump launch points
– Launch point becomes hot-spot – risk for failure
– Multiple launch points or distributed pump injection preferable
– Pre-empts splicing & “all-fibre” devices (without free-space path)

High-power fibre lasers 42

J. Nilsson, ORC postgraduate lecture, March 9 2011

21
 Side-pumping schemes
V-groove side-pumping
Pump Keopsys
Fused Rare-earth-doped
fibre
Side splice part core Outer cladding
V-groove

Doped C. Renaud, thesis,

Fibre U. Southampton 2001 Focusing lens
Inner cladding
IPG?? Silicone
Coating
rubber Pump beam

GTWave® SPI lasers Tapered fibre bundle

From C. Headley, OFS
Also Sifam, ITF...

High-power fibre lasers 43

J. Nilsson, ORC postgraduate lecture, March 9 2011

Side-pumping schemes (II)

Tapered V-groove side-
End-pumping GTWave Side-splicing
fiber bundle pumping
End user ease of use and
Yes Yes No No No
reconfigurability
Ruggedness Poor Good Medium Good Good
Polymer-coated Polymer-coated
Type of active fiber Any Any Any
(strippable) (strippable)
Poor (free-space Best (continuous Best (continuous fiber Best (continuous
Signal launch efficiency Good
launch) fiber path) path) fiber path)
125 – 500 Arbitrary (> 200 Arbitrary (> 125
Active fiber size Arbitrary 100 – 200 µm?
µm? µm?) µm?)
Size of pump fiber / Arbitrary (same 125 – 200 Arbitrary (same
100 – 200 µm? 100 µm??
emitter as active fiber) µm as active fiber)
Pump launch efficiency Best Good Good Good Good (?)
Pump power handling Best Good Good Good Good
Brightness preservation Best Good Poor Good Medium (?)
Distributed / 4
Pump launch points 2 2 – 18 Arbitrary (typically, multiplexing Arbitrary
possible)
Hot-spots? Yes Yes Yes (?) No No (?)
Notes: These are estimates. Published data is scarce.
Pump power handling of a single launch point is very good with end-pumping, but the
number of launch points is limited to two.
See also C. Headley III, et al., “Tapered fibre bundles for combining laser pumps”, in
Fibre lasers II: technology, systems, and applications, L. N. Durvasula, A. J. W. Brown,
and J. Nilsson, Eds., Proc. SPIE vol. 5709, pp. 263-272 (2005) High-power fibre lasers 44
J. Nilsson, ORC postgraduate lecture, March 9 2011

22
 Spectral coverage of cladding-pumped rare-
earth-doped silica fibres

40
Yb
Output power [dBm]

10 kW
35
single- Er/Yb
mode 0.3 kW Tuning
30 few-moded Tm curves
Nd 1 kW (seriously
1 kW few-moded out of date)
25
multi-
mode
20
1000 1200 1400 1600 1800 2000 2200
Wavelength [nm]

• Nd and Yb also used for lasers in the 900 – 1000 nm range

• Output power up to 10 W
• Also Ho at 2 µm recently High-power fibre lasers 45
J. Nilsson, ORC postgraduate lecture, March 9 2011

Disadvantages
• High nonlinearities
– Sometimes good, often bad
– High gain & high nonlinearities
“interesting physics”

• Low energy storage

– The flip side of high gain efficiency (& tight confinement)
• Already a small stored energy creates a high gain, which leads to
instabilities and large power losses to amplified spontaneous emission

• Low damage threshold

– Especially in pulsed regime

High-power fibre lasers 46

J. Nilsson, ORC postgraduate lecture, March 9 2011

23
 Limitations of fibres
Long with tight beam confinement: Many advantages but some trade-offs
Optical
Optical •• Large
Largecore
core All
Alloperating
operating
damage
damage •• New
Newmaterials
materials regimes
regimes
Nonlinear
Nonlinear •• Large
Largecore
core Primarily
Primarilypulsed
pulsed
degradation
degradation •• Short
Shortfibre
fibre and
andnarrow
narrow
•• Raman
Raman ––Higher
HigherRE-concentration
RE-concentration/ /new
newmaterials
materials linewidth
linewidthregimes
regimes
•• FWM
FWM ––Higher
Higherpumppumpbrightness
brightness
•• SPM
SPM •• Spectral
Spectralfilter
filter
SBS
SBS •• Linewidth
Linewidthbroadening
broadeningofofsignal
signal Very
Verynarrow
narrow
––SBS
SBSnegligible
negligiblefor
for>>10
10GHz
GHzlinewidth
linewidth linewidth
linewidth
––No
NoSBS
SBSfor forpulses
pulses<<55ns
ns Large
Largecore
core&&
•• Linewidth
Linewidthbroadening
broadeningofofgain
gain short
shortfibre
fibrehelps,
helps,
––Temperature,
Temperature,stress,
stress,compositional
compositionalvariations
variations too!
too!
Energy
Energy •• Large
Largecore
core High
Highenergy
energy
storage
storage pulses
pulses
Thermal
Thermal •• Longer
Longerfibres
fibres
damage
damage •• Better
Bettercoatings
coatings
•• Improved
Improvedheatsinking
heatsinking

High-power fibre lasers 47

J. Nilsson, ORC postgraduate lecture, March 9 2011

Size matters!
Silica matters, too!
• Large core to facilitate:
Large core is the most important design –power handling
feature of a high-power fibre laser
–minimization of nonlinear
Typically a few hundred square microns effects
–high energy storage for
pulsed application
–efficient pump absorption &
reduced device length
• Large inner cladding for launch
of high-power pump beams
• Silica-based
9 / 125 µm – Superb power-handling
100 W
40 / 650 µm, > 1 kW
– Excellent control of
parameters

High-power fibre lasers 48

J. Nilsson, ORC postgraduate lecture, March 9 2011

24
 Diffraction-limited large-core fibre lasers
Control of refractive index profile

Conventional process Improved process

In large cores, the
beam follows the
index profile. The
fundamental mode is
NOT diffraction-
limited with ring-
shaped large cores.

• More accurate RIP allows

for diffraction-limited No central dip
Central dip fundamental mode M2 value ~ 1.4
M2 value ~ 3.2 • Precise control in large
structures real challenge
J. Sahu, CLEO 2004 High-power fibre lasers 49
J. Nilsson, ORC postgraduate lecture, March 9 2011

Research on core area scaling

Single-mode operation of multimode cores
• Rod fibre
• Straight, rigid, low NA
• Crystal fibre, Jena,
Bordeaux, Aculight...

• Leakage channel fibre

• High leakage loss for
higher order modes
• Chirally coupled core
– “Modal sieve”
• High resonant coupling loss
• U. Bath, IMRA for higher order modes
• U. Michigan

• Index anti-guiding
• Better modal discrimination (?)
• Siegman, CREOL, Clemson
• Higher order mode fibre
• Claim: Higher order mode more
robust than the fundamental mode
• OFS High-power fibre lasers 50
J. Nilsson, ORC postgraduate lecture, March 9 2011

25
 But, core area scaling
doesn’t end all problems
Stimulated
Maximum output power:
Raman
contour lines in kW
scattering

From J. W. Dawson et al.,

"Analysis of the scalability of
diffraction-limited fiber lasers
and amplifiers to high average
power," Opt. Express 16,
13240 (2008)
Thermal beam
Pump launch
distortions

• 36.6 kW maximum output power with assumed parameters

– Diode pumped Yb-doped fibre
– Tandem-pumping would be better!
High-power fibre lasers
J. Nilsson, ORC postgraduate lecture, March 9 2011

Ultimately, need to consider

thermal effects and nonlinearities
• Nonlinearity (SRS) scales as length / area

Thermal beam distortions:

• Waveguiding can no longer ensure an undistorted beam when the
core becomes too large and the guiding too weak
• Allowable thermal load per unit length assumed to scale as 1 / area
– Brown and Hoffman (IEEE JQE 37, p. 207, 2001)
– This can be questioned and needs to be verified experimentally!
• Thermal load per unit length scales as 1 / length (fixed total power)

Thermal beam distortions scale as area / length (?)

• Possible to find optimum value of length / area

– Compromise between SRS and thermal beam distortions
– Maximum power a constant 36.6 kW on ridge

High-power fibre lasers

J. Nilsson, ORC postgraduate lecture, March 9 2011

26
 Why fibre is better than bulk
• Optimum value of length / area ~1000 times larger than the value of
length / area for laser rod
– Fundamentals of beam propagation (diffraction-limited beam &
Gaussian optics) vs. materials constants (thermal conductivity, thermo-
optic constants, nonlinear coefficients...)

Waveguiding helps! The fibre waveguide allows us to find best

trade-off between nonlinearities and thermal distortions, as
governed by fundamentally different materials parameters and
physical aspects,
whereas a rod laser doesn’t! Fibre is the Future
The Future is Now!

High-power fibre lasers

J. Nilsson, ORC postgraduate lecture, March 9 2011

Beam combination

• When core area scaling runs out of steam

• Take the output from several fibre lasers
and combine them
• Maintain beam quality of single fibre!

High-power fibre lasers

J. Nilsson, ORC postgraduate lecture, March 9 2011

27
 Amplifier-based coherent beam combination
(active phasing / phased-array laser / coherent multi-path MOPA)

Need:
• Narrow-linewidth source
– E.g., “single-frequency”
• High-gain amplification to high power
– Preferably efficient
• Single spatial mode, single polarization
• Phase control
Fibres good for coherent beam
combination
• 1 kW x 100 fibres = 100 kW!
T. M. Shay et al., “Self-synchronous and self-referenced coherent
beam combination for large optical arrays”. IEEE. J. Sel. Top.
Quantum Electron. 13, 480 (2007)
E. C. Cheung et al, "Diffractive-optics-based beam combination of
a phase-locked fibre laser array," Opt. Lett. 33, 354-356 (2008)
P. Bourdon et al., Coherent beam combining of fibre amplifier
arrays and application to laser beam propagation through
T. Y. Fan: turbulent atmosphere Proc SPIE 6873, U197-U205 (2008)
CLEO 2004
Tutorial High-power fibre lasers 55
J. Nilsson, ORC postgraduate lecture, March 9 2011

High-power single-frequency
fibre MOPAs
• Seed source operates on single longitudinal mode
– kHz – MHz linewidth, typically
• For coherent beam combination
• Stimulated Brillouin scattering limitation
– Nonlinear scattering effect
– Threshold historically 1 – 10 W

Single-
Amp Amp Amp
frequency
seed

High-power fibre lasers

J. Nilsson, ORC postgraduate lecture, March 9 2011

28
 MOPAs bring out the best of high
power rare earth doped fibres
Cladding-pumped RE-doped fibres offer unique combination of
high power
high efficiency
high gain
broad bandwidth

MOPAs (master oscillator – power amplifiers) allow for
control, sophistication and high power at the same time
Versatile & rapidly reconfigurable

Amp Amp
High power
Seed Amp
output with
characteristics
determined
High control by seed
High power

High-power fibre lasers 57

J. Nilsson, ORC postgraduate lecture, March 9 2011

Stimulated Brillouin scattering key challenge

for single frequency fibre MOPAs
• Weaknesses:
– Stimulated Brillouin scattering (SBS)
– Nonlinear, acousto-optic, effect (electrostriction)
– Because of SBS, fibres are considered poor for single-frequency sources
• Strengths:
– Combination of high gain, high power, high efficiency
– Mode control
– Polarization control
• Conventional strengths of fibre MOPAs
• One weakness, many strengths
• For this to work, we need to fix the SBS!
• If we fix the SBS, we reverse the conventional wisdom that fibres do
not make good high power single frequency amplifiers!

High-power fibre lasers 58

J. Nilsson, ORC postgraduate lecture, March 9 2011

29
 So, can we fix SBS?
• Yes we can!
• SBS is a high gain, but fragile process
– Scattering off propagating acoustic wave
• Electrostrictively generated
• Heavily damped
• Stokes shift 10 – 20 GHz = frequency of acoustic wave
– Proportional to speed of sound
– Narrow-band interaction due to phase-matching with slow acoustic wave
• Intrinsic linewidth 20 – 50 MHz
Fibre core

Brillouin pump wave

Brillouin Stokes wave
Acoustic wave
• Easy to perturb
– Perturbed, e.g., by intrinsically generated longitudinal thermal variations
• Changes the speed of sound and
thus the Brillouin frequency shift
– Compositional variations also
change the speed of sound High-power fibre lasers 59
J. Nilsson, ORC postgraduate lecture, March 9 2011

1. Longitudinal temperature variations

End-pumped configuration
Diode pump & Yb signal powers • Nearly exponential decay of diode
(counter-propagating) pump & signal in counter-
Change in temperature propagating end-pumped
Change in speed of sound configuration
Change in Brillouin
frequency shift • Yb pump = diode pump
All proportional and
• Yb signal = Brillouin pump!
exponentially
decaying
Brillouin gain

Diode Position
pump
Yb-doped fibre
Yb signal Frequency
Yb signal input
Brillouin gain
Brillouin gain

Brillouin gain

output

Frequency Frequency Frequency

High-power fibre lasers 60
J. Nilsson, ORC postgraduate lecture, March 9 2011

30
 0.4 kW single-frequency
Yb-doped fibre MOPA
500
Measured Output power Backward signal power
Linear fit
400 Expected Brillouin power
Slope efficiency: 72% without suppression
Signal power [W ]

Max. power: >400W 0.30

300 Linewidth

Backward signal pow er [a.u.]

-40
80 mW (seed)

-45
264 W 0.25
-50
0.20
200

Power [dB]
-55

-60
0.15
-65

-70
100 0.10
-75
-120-100 -80 -60 -40 -20 0 20 40 60 80 100 120
Frequency [kHz] 0.05

0 0.00
0 100 200 300 400 500 600 0 1 00 20 0 3 00 4 00
S ig nal pow er [W ]
Launched pum p power [W ]
SBS suppressed by
thermal gradient
> 100 K variation estimated
Y. Jeong, J. Nilsson, J. K. Sahu, D. B. S. Soh, C. Alegria, P. Dupriez, C. A. Codemard, D. N. Payne,
R. Horley, L. M. B. Hickey, L. Wanzcyk, C. E. Chryssou, J. A. Alvarez-Chavez, and P. W. Turner,
“Single-frequency, polarized ytterbium-doped fibre MOPA source with 264 W output power”,
CLEO 2004 CPDD1 High-power fibre lasers 61
J. Nilsson, ORC postgraduate lecture, March 9 2011

Experimental setup
Fibre details: hi-bi Yb-doped fibre with
borosilicate stress rods, core 25 µm,
NA ~ 0.06, OD 380 µm, L = 7 m,
birefringence ~ 2x10-4
Large-core
250 mW 2W 7W λ/2 WP 3W
double-clad
DFB fibre
laser, 80 mW, Isolator
hi-bi YDF 0.4 kW
BS DM
1060 nm
(single- YDFA, YDFA, YDFA, FR
Single-mode
non- non- non-
frequency,
single-
PM PM PM
DM
Single-frequency
Lens Lens Pump Angled Lens
polarization)
Polarizers thru ends (< 60 kHz)
SBS Diode pump Linearly polarized
monitoring source, 975 nm

Output signal
Max output: 0.4 kW (limited by available
Final-stage seed: 3 W @ 1060 nm pump power)
linewidth < 60 kHz Linewidth: < 60 kHz (resolution limited)
Pump: 975 nm Beam quality: M2 = 1.1
Polarization extinction ratio: 16 dB

DARPA Contract No. MDA972-02-C-0049

High-power fibre lasers 62
J. Nilsson, ORC postgraduate lecture, March 9 2011

31
 2. Transversally varying speed of
sound – compositional variations
Brillouin gain Brillouin gain Brillouin gain

Local speed of sound Yb pump = diode pump

Yb signal = Brillouin pump!

Frequency
Brillouin pump

Fibre core
Brillouin Stokes
Acoustic waves
Same wavelength
Frequency
but different
frequencies

Frequency
Brillouin gain

Frequency High-power fibre lasers 63

J. Nilsson, ORC postgraduate lecture, March 9 2011

Fibre approach
• Use dopants that affect speed of sound
and speed of light differently

• Independent control of speed of light

(refractive index) and speed of sound
(“acoustic refractive index”)

• Experimental demonstrations include:

– P. D. Dragic et al., “Optical fibre with an
acoustic guiding layer for stimulated
Brillouin scattering suppression”, CLEO,
CThZ3, 2005.
– M. L. et al., "Al/Ge co-doped large mode
area fibre with high SBS threshold",
Opt. Express 15, 8290-8299 (2007)
– M. D. Mermelstein et al., “11.2 dB SBS
gain suppression in a large mode area
Yb-doped optical fibre”, Proc. SPIE Ming-jun Li, “Managing nonlinearity in optical fibre
6873, U63-U69 (2008) for high-power lasers”, SPIE Newsroom, 2006
– T. Nakanishi et al., “Al2O3-SiO2 core http://newsroom.spie.org/x3863.xml?highlight=x537
highly nonlinear dispersion-shifted fibre
with Brillouin gain suppression
improved by 6.1 dB”, ECOC 2006, PDP
Th4.2.2
High-power fibre lasers 64
J. Nilsson, ORC postgraduate lecture, March 9 2011

32
• Fixing SBS has made single-frequency fibre MOPAs very
attractive for coherent beam combination (active phasing)

• kW-level SF MOPAs possible

• Fibre-based phased-array lasers very interesting topic!

• A fibre is an engineerable gain medium

High-power fibre lasers 65

J. Nilsson, ORC postgraduate lecture, March 9 2011

Example: Control of Kerr nonlinearity

in micro-structured fibres
γ ≈ 0.01 /(W.km) PBG fibre High power
beam delivery
and processing
γ ≈ 0.1 /(W.km) Large mode area HF

γ ≈ 1 /(W.km) Standard fibre

γ ≈ 70 /(W.km) Small-core silica HF

High
nonlinear
γ ≈ 1100 /(W.km) Extruded Bi-HF gain

• More than 5 orders of magnitude From David

Richardson
engineering control possible!
High-power fibre lasers 66
J. Nilsson, ORC postgraduate lecture, March 9 2011

33
Example: Dispersion control in holey fibres

Anomalous dispersion
possible down to the visible
(λ0 ~550 nm)
[Wadsworth et al. Elect. Lett. 36,
2000]

Dispersion flattening
[Monro et al, WDM Tops Vol. 29,
[d=hole diameter 1999]
Λ = hole spacing]
High dispersion when holes
large (dispersion
compensation)
[Birks et al IEEE Phot. Tech.
Lett. 11, 1999]
From David Richardson
High-power fibre lasers 67
J. Nilsson, ORC postgraduate lecture, March 9 2011

A fibre is an
engineerable
gain medium!

Make the most of this!

High-power fibre lasers 68

J. Nilsson, ORC postgraduate lecture, March 9 2011

34
 Applications

J. Nilsson, High-power
High power fibre
fibre sources
lasers
Short course
J. Nilsson, ORC SC748, Photonics
postgraduate West,
lecture, Jan 9252011
March 2009

Unique characteristics enable unique devices

• High-gain broadband amplifiers for telecom
– Erbium-doped fibre amplifier in particular, also Raman amplifiers
– High gain efficiency and low threshold
– Single-mode operation and low polarization-dependence crucial fibre attributes
• High-power broadband ASE-sources
– Require high gain
• Amplifiers for ultrashort pulses
– Require broad bandwidth
– The controllable dispersion provides additional advantages
• Efficient high-power Nd-doped fibre amplifiers at 900 nm (Ti:Al2O3 replacement?)
– Requires suppression of dominating 1100 nm transition
– Possible with distributed fibre filter
– Also S-band EDFAs
• Distributed feedback fibre lasers -- with FBGs written into the fibre
• High-power MOPAs with superb control of optical parameters
– Impossible to obtain directly in high-power lasers
– Fibre amplifiers work very well in high-power MOPA systems
• And many more, including high-power tunable fibre lasers
And, of course... High efficiency and geometry of
Yb-doped fibres ideal for high powers
High-power fibre lasers 70
J. Nilsson, ORC postgraduate lecture, March 9 2011

35
 Application areas
 Industrial
 Materials processing, welding, printing ….
 Aerospace & defence
 Lidar, range-finding, directed energy, remote sensing...
 Pump sources
 Lasers
 Nonlinear converters
 UV (lithography, etc.), X-rays, visible, supercontinuum…
 Medical
 Imaging, laser surgery, cosmetics…
 Scientific
 Telecoms (?)
 Raman lasers, high-power multi-port EYDFA, free-space
communications, ...
 Displays (?)
 Requires visible sources
High-power fibre lasers 71
J. Nilsson, ORC postgraduate lecture, March 9 2011

Summary
• Thermal properties make fibres excellent for high-power,
high-brightness operation
– Low threshold helps, too
• Ytterbium-doped fibre lasers available at 10 kW
• Fibres are engineerable
• Limited core area limits further power-scaling
• Multi-path fibre-MOPA-based phased-array lasers open
up for multi-element power-scaling
• MOPAs enable high control at kW level
– Temporal
– Spatial
– Spectral

High-power fibre lasers 72

J. Nilsson, ORC postgraduate lecture, March 9 2011

36
 MOPAs bring out the best of high
power rare earth doped fibres
Cladding-pumped RE-doped fibres offer unique combination of
high power
high efficiency
high gain
broad bandwidth

MOPAs (master oscillator – power amplifiers) allow for
control, sophistication and high power at the same time
Versatile & rapidly reconfigurable

Amp Amp
High power
Seed Amp
output with
characteristics
determined
High control by seed
High power

High-power fibre lasers 73

J. Nilsson, ORC postgraduate lecture, March 9 2011

Fibre MOPAs the perfect light source

High power
Light-driven
fibre MOPA,
process
flexible with
(e.g., materials
any output
processing)
characteristics

Feedback
with
intelligence Monitoring

High-power fibre lasers 74

J. Nilsson, ORC postgraduate lecture, March 9 2011

37
 Books on Rare Earth Doped
Fibre Amplifiers and Lasers
France, P. W., ed., Optical fibre lasers and amplifiers (Blackie 1991) -- Old and thin but quite
good.
Bjarklev, Anders, Optical fiber amplifiers: design and system applications (Artech House, 1993) -
- Provides good insight. Mostly EDFAs. Good introduction.
Digonnet, Michel J. F., ed., Rare earth doped fiber lasers and amplifiers, 2nd ed. (Marcel Dekker
2001) -- Best coverage of fibre lasers and physics.
Desurvire, Emmanuel, Erbium-doped fiber amplifiers: principles and applications (Wiley, 1994) --
Lots on noise, lots of everything except WDM. In-depth and difficult at times. Exclusively
EDFAs.
Desurvire, E., D. Bayart, B. Desthieux, S. Bigo, Erbium-doped fiber amplifiers – device and
system developments (Wiley 2002) – Covers aspects such as WDM and other
developments since Desurvire’s first book.
Sudo, Shoichi, Ed., Optical fiber amplifiers (Artech House, 1997) -- Lots on fabrication.
Becker, P. C., N. A. Olsson, and J. R. Simpson, Erbium-doped fiber amplifiers: fundamentals
and technology (Academic Press 1999) – Less mathematical than Desurvire but still quite
comprehensive. Lots on WDM. Exclusively EDFAs.
R. W. Berdine and R. A. Motes, Introduction to High Power Fiber Lasers (Directed Energy
Professional Society; ISBN-10: 0979368731, 2009)
D. J. Richardson, J. Nilsson, and W. A. Clarkson, “High power fiber lasers: current status and
future perspectives”, J. Opt. Soc. Am. B 27, B63-B92 (2010)
Not book but 30 page review article

High-power fibre lasers

J. Nilsson, ORC postgraduate lecture, March 9 2011

38