SESAM modelocked solid-state lasers Lecture 2 ... - the Keller Group

SESAM technology – ultrafast lasers for industrial application
SESAM modelocked solid-state lasers
Lecture 2
SESAM
Ursula Keller
U. Keller et al. Opt. Lett. 17, 505, 1992
IEEE JSTQE 2, 435, 1996"
Progress in Optics 46, 1-115, 2004!
Nature 424, 831, 2003"
Gain!
SESAM solved Q-switching problem
for diode-pumped solid-state lasers
Output
coupler!
ETH Zurich, Physics Department, Switzerland
www.ulp.ethz.ch
Ultrafast Laser
Physics
Tutorial, Saturday 2. March 2013, 16:00 – 19:30
Ultrafast Optics 2013
Davos, Switzerland, March 2-8, 2013
ETH Zurich
SESAM
SEmiconductor Saturable Absorber Mirror
"
self-starting, stable, and " reliable modelocking of
diode-pumped ultrafast " solid-state lasers
Semiconductor saturable absorber
SESAM Parameters:
Saturation Fluence, Modulation Depth, Recovery Time ...
100
F sat, A
Saturation fluence
Reflectivity (%)
95
ΔR ns
Non-saturable losses
ΔR
Modulation depth
Pump-probe signal
100-fs excitation pulse
4-ps excitation pulse
90
0 50 100 150 200 250 300
Incident pulse fluence F p ( µJ/cm 2 )
-10
-5
0 5 10 15 20
Pump-probe delay (ps)
25
30
Guidelines how to measure these parameters:"
M. Haiml, R. Grange, U. Keller, Appl. Phys. B 79, 331, 2004"
and with improved accuracy: D. J. H. Maas, et al., Optics Express 16, 7571, 2008"
1

andgap energy E g (eV)
2.4
2.0
1.6
1.2
0.8
GaP
AlGaAs
GaAs-based SESAMs
GaAs
AlAs
InP
InGaAs
0.4 Lattice matched to GaAs"
High refractive index contrast for DBR"
InGaAs strained on GaAs"
0.0
5.4 5.5 5.6 5.7 5.8 5.9
lattice constant a 0 (Å)
AlInGaAs
750 nm
6.0
1 µm
1.3 µm
1.5 µm
InAs
6.1
0.5
0.6
0.7
0.8
1.0
1.3
1.5
3.0
7.0
wavelength λ (µm)
InP-based SESAMs
Lattice matched to InP"
Small refractive index contrast for DBR"
GaInNAs-based SESAMs
Thank you to Stanford, Bell Labs and ETH Zurich
[In] = 36.5 % "
! [N] = 2 %"
GaNAs"
InGaAs!
GaInNAs better lattice matched on
GaAs"
Opt. Lett. 27, 2124, 2002 & Appl. Phys. Lett. 84, 4002, 2004"
Stanford University
Ph.D. student
1985-1989
laser physics
ultrafast measurement
techniques
microwave measurement
tools
Bell Labs, Holmdel
MTS
1989-1993
+ access to state-of-the-art
semiconductor materials (MBE)
ETH Zurich
tenured Professor in Physics
since 1993
+ resources to be fully
empowered
Enabled interdisciplinary approach with the combination of
solid-state lasers, semiconductor physics, and microwave
measurement techniques.
2

20 years of SESAM – looking back
20 years of SESAM
Appl. Phys. B 100, 15-28, 2010
How did it all happen?
This is the paper to read ...
20 years of ultrafast solid-state lasers: invited paper
• Why was it assumed that diode-pumped solid-state lasers
cannot be passively modelocked?
• How was the SESAM invented?
• State-of-the-art performance and future outlook.
First SESAM (first design called A-FPSA)
April 1, 1992 (submitted Nov. 26, 1991)
!
Ultrashort pulse generation with modelocking
A. J. De Maria, D. A. Stetser, H. Heynau
Appl. Phys. Lett. 8, 174, 1966
200 ns/div"
Q-switching problem
in passively modelocked
solid-state lasers:
Active Modelocking
Nd:glass
first passively modelocked laser
Q-switched modelocked!
50 ns/div"
- active modelocking for
solid-state lasers
- dye lasers solved the problem
"
1960 1970 1980 1990 2000
Year"
Flashlamp-pumped
solid-state lasers !
acousto-optic loss modulator
needs RF power and water cooling
3

Innovation: before and after
4-Niveau-System
Ti 3+ :Saphir Laser
How did I invent the SESAM?
acousto-optic modelocker
needs RF power and water cooling
SESAM modelocker
Ti:Saphir Laser
1982
Peter Moulton
MIT Lincoln Lab
Two experiments that should not have worked ...
“Magic Modelocking”
Postdeadline Paper CLEO 1990
Nr. CPDP10
“modelocked Ti:sapphire laser without
an absorber inside the cavity”
... should not work!
Published result later in:
D. E. Spence, P. N. Kean, W. Sibbett,
Optics Lett. 16, 42, 1991
Explained modelocking as a “coupled
cavity modelocking effect” between
two transverse modes
“CPM Ti:sapphire laser”
Postdeadline Paper Ultrafast
Phenomena 1990, Nr. PD11
Y. Ishida, N. Sarukura, H. Nakano
“used a dye saturable absorber inside
the CPM Ti:sapphire laser”
... should not work!
loss
gain
pulse
time
?
Both of them explained later by KLM by Keller et al.
Optics Lett. 16, 1022, 1991 IEEE JQE 28, 2123, 1992
loss
gain
pulse
time
loss
gain
pulse
First Demonstration: D. E. Spence, P. N. Kean, W. Sibbett, Optics Lett. 16, 42, 1991
Explanation: U. Keller et al., Optics Lett. 16, 1022, 1991
•
Kerr Lens Modelocking (KLM)
Advantages of KLM
§ very fast thus shortest pulses
§ very broadband thus broader tunability
Disadvantages of KLM
§ not self-starting
§ critical cavity adjustments
(operated close to the stability limit)
§ saturable absorber coupled to cavity
design (limited application)
time
4

Soliton Laser by Linn Mollenauer
Why did the Sibbett group explain their results as a
“coupled modelocking effect”?
1984
Remember the mind set at that time was:
Passive modelocking of solid-state laser does not work!
Coupled cavity
... but there was the soliton laser from Linn Mollenauer
fiber with negative dispersion (GDD < 0)
soliton pulse compression in coupled cavity with shortens
pulse in main cavity
1989
Coupled Cavity Modelocking by Wilson Sibbett
Coupled Cavity Modelocking by “Blow & Wood”
1988
Coupled cavity
Coupled cavity modelocking (CCM)
fiber with positive dispersion (GDD > 0)
Dispersive pulse broadening in coupled cavity! How should that shorten pulse in main cavity?
Explanation with APM (additive pulse modelocking):
E. P. Ippen, H. A. Haus, L. Y. Liu, J. Opt. Soc. Am. B 6, 1736, 1989
Pulse is broadened yes, ... but SPM makes an intensity dependent phase shift, which
leads to constructive interference at the peak and destructive inteference in the wings with the
pulse from the main cavity. This gives a pulse shortening effect of the coupled cavity.
Coupled cavity
“any nonlinear element in the coupled cavity should shorten the pulse”
theoretical prediction with numerical modeling ... physical insight was missing
5

1990
Coupled Cavity Modelocking by Ursula Keller
1990
Coupled Cavity Modelocking by Ursula Keller
τ p
Resonant passive modelocking (RPM)
L L + δ L δ L
Resonant passive modelocking (RPM)
GaAs 95 Å
75x
GaAl x As 1-x 45 Å, x=0.3
AlAs/GaAlAs 800 nm mirror
semiconductor saturable absorber (“borrowed from my office neighbour Keith Goosen”)
pulse compression in coupled cavity which shortens pulse in main cavity
Surprise:
• no active cavity length stabilization
was necessary
• modelocking was stable
even with large cavity length detuning
U. Keller et al., Optics Lett. 15, 1377, 1990
δ L
δ L
Coupled Cavity Modelocking by Ursula Keller
Invention of the first SESAM device: A-FPSA
Coupled cavity acts like a nonlinear Fabry Perot (FP)
Antiresonant Fabry-Perot Saturable
Absorber (A-FPSA)
Example: Opt. Lett. 18, 217, 1993"
R top
absorber
RPM = amplitude (resonant) nonlinearity
operated at maximum FP reflectivity
requires no cavity length stabilization
APM = phase nonlinearity
operated not at maximum FP reflectivity
requires cavity length stabilization
U. Keller et al., Optics Lett. 17, 505, 1992
R bottom
6

Scaling A-FPSA towards a single QW SESAM
20 years of SESAM
1995
Adjustable parameter: top reflector
1996
Single 15 nm GaAs quantum well embedded inside a Bragg mirror
L. R. Brovelli et al., Electron. Lett., vol. 31, 287, 1995
Introduction of the ackronym: SESAM
Invited review article
September 1996, IEEE JSTQE
Moving from Ti:sapphire to diode-pumped ss-laser
Moving from Ti:sapphire to diode-pumped ss-laser
was not trivial .... why?
Ti:sapphire laser
@ 800 nm
Opt. Lett. 15, 1377, 1990
1991
RPM modelocked Nd:YLF
bandgap energy E g (eV)
2.4
2.0
1.6
1.2
0.8
0.4
0.0
5.4
GaP
AlGaAs
5.5
GaAs
5.6
AlAs
5.7
InGaAs
5.8
InP
5.9
AlInGaAs
750 nm
6.0
1 µm
1.0
1.3 µm
1.5 µm
InAs
6.1
0.5
0.6
0.7
0.8
1.3
1.5
3.0
7.0
wavelength λ (µm)
GaAs 95 Å
75x
GaAl x As 1-x 45 Å, x=0.3
AlAs/GaAlAs 800 nm mirror
Nd:YLF laser
@ 1.064 µm
IEEE JQE 28, 1710, 1992
1992
A-FPSA modelocked Nd:YLF
Nd:YLF laser pumped with
Ti:sapphire laser
1993
“mailed A-FPSA to California”
lattice constant a 0 (Å)
LT = low temperature
MBE growth
A-PFSA modelocked
diode-pumped Nd:YLF and
Nd:YAG laser
7

It also works to modelock fibers ...
LT grown SESAMs to reduce absorber recovery time
1993
“carried an A-FPSA to TU Vienna”
Adjustable parameter:
absorber recovery time
More on LT MBE growth and ion implantation to
reduce recovery time of SESAMs:
Appl. Phys. Lett., vol. 75, pp. 1437-1439, 1999
Appl. Phys. Lett., vol. 74, pp. 1993-1995, 1999
Appl. Phys. Lett., vol. 74, 3134-3136, 1999
Appl. Phys. Lett., vol. 74, pp. 1269-1271, 1999
Physica B: Condensed Matter, vol. 273-274, pp. 733-736, 1999
Stable pulses even with a “long” net gain window
Stable pulses even with a “long” net gain window
loss
gain
loss
Why is the pulse stable
with a long net gain window?
gain
R. Paschotta, U. Keller, Appl. Phys. B 73, 653, 2001
absorber delays pulse!
loss
Fully saturated absorber:"
leading edge of pulse"
negligible loss for"
has significant loss from "
trailing edge of pulse"
the saturable absorber"
pulse
time
Kerr lens modelocking (KLM)
Fast saturable absorber
D. E. Spence, P. N. Kean, W. Sibbett
Opt. Lett. 16, 42, 1991
pulse
time
Soliton modelocking
“not so fast” saturable absorber
F. X. Kärtner, U. Keller,
Opt. Lett. 20, 16, 1995
time
Dominant stabilization process with a relatively long net-gain window:!
Picosecond domain: absorber delays pulse"
The pulse is constantly moving backward and can swallow any noise
growing behind itself."
Femtosecond domain: dispersion in soliton modelocking"
8

Laser pulse with pulse envelope
Soliton modelocking: GDD negative, n 2 > 0
E( t) = A( t)e iω 0t
Pulse envelope A(t)
Group
Delay Velocity
Dispersion
Self-
Phase-
Modulation
Gain
Slow
Absorber
Outputcoupler
~
E(ω)
Δω 0
Electric field E(t)
E( t) = 1
2π
∫
E ( ω )e iωt dω
Master equation:
Opt. Lett 20, 16, 1995 and IEEE JSTQE 2, 540, 1996
A out ( T ,t) = e −q(t ) A in ( T ,t)
A out ( T ,t) = ⎡⎣ 1 − q( t)
⎤⎦ A in ( T ,t)
⇒ ΔA( T ,t) = −q( t) A( T ,t)
0
0
A(Δω)
~
Δω
ω 0
E( t) = 1 E ( ω 0
+ Δω )e i ( ω 0 + Δω )t dΔω
2π
∫ = 1
ω
A ( Δω ) = E ( ω 0
+ Δω )
2π eiω 0t
∫
A ( Δω )e iΔωt dΔω
T R
∂
∂T A T ,t
⎛ ⎝
⎜
( ) = iD ∂ 2
( ) 2
− iδ A T ,t
2
∂t
⎞
⎠
⎟ A T ,t
⎛
⎝
⎜
( ) + g − l + D g
A ( T ,t ) = ⎛
A sech ⎛ t ⎞ ⎞
⎝
⎜ 0 ⎝
⎜
τ ⎠
⎟
⎠
⎟ exp ⎡
iφ T ⎤
⎢ 0
⎣ T
⎥ + continuum τ = 4 D
R ⎦
δ ⋅ F p,L
∂ 2
∂t − q( t)
⎞
2 ⎠
⎟ A T ,t
( ) = 0
!
Ultrashort pulse generation with modelocking
A. J. De Maria, D. A. Stetser, H. Heynau
Appl. Phys. Lett. 8, 174, 1966
Nd:glass
first passively modelocked laser
Q-switched modelocked!
200 ns/div"
50 ns/div"
Q-switching problem
in passively modelocked
solid-state lasers:
- active modelocking for
solid-state lasers
- dye lasers solved the problem
"
1960 1970 1980 1990 2000
Year"
Flashlamp-pumped
solid-state lasers !
9

SESAM Parameters:
Saturation Fluence, Modulation Depth, Loss ...
100
F sat, A
Saturation fluence
Reflectivity (%)
95
ΔR ns
Non-saturable losses
ΔR
Modulation depth
Q-switched mode locking is avoided if...
C. Hönninger, R. Paschotta, F. Morier-Genoud, M. Moser, and U. Keller,
JOSA B 16, 46 (1999)
E P 2 > E sat,L
E sat,A
ΔR
Q-switched mode locking
QML
= A eff, A
F sat,A
ΔR
QML noise
QML
90
0 50 100 150 200 250 300
Incident pulse fluence F p ( µJ/cm 2 )
Guidelines how to measure these parameters:"
M. Haiml, R. Grange, U. Keller, Appl. Phys. B 79, 331, 2004"
0
10 20 30 40
Time (multiples of round trip time)
cw mode locking
cw ML
cw ML
with improved accuracy: D. J. H. Maas, et al., Optics Express 16, 7571, 2008"
0 10 20 30 40
Time (multiples of round trip time)
Appl. Phys. B 58, 347, 1994
SESAMs to reduce Q-switching instabilities
SESAM design
2005
1) SESAMs with low saturation fluence:
antiresonant
resonant
E-SESAM
SiO 2
4
4
0.4
400
3
2
1
3
2
1
200
0
-200
0.3
0.2
0.1
Air
-400
0
0
0.0
5600 6000 6400 6800
1280 1320 1360
4
4
4
1000
3
3
500
3
2
2
0
2
1
1 -500
1
Air
2
2
0
2
-1000
0
0
5600 6000 6400 6800
1280 1320 1360
0
4
4
4
400
3
3
200
3
Air
1
1 -200
1
0
0
-400
5600 6000 6400 6800
1280 1320 1360
0
Distance from substrate (nm)
Wavelength (nm)
Refractive index
Refractive index
Refractive index
Standing wave pattern
GDD and field enhanc. factor
GaAs
AlAs
GaAs
AlAs
GaAs
AlAs
GaInNAs absorber
Field enhancement |E n | 2
Field enhancement |E n | 2
Field enhancement |E n | 2
Group delay dispersion (fs 2 )
GDD (fs 2 )
GDD (fs 2 )
Field enhancement factor ξ
Field enhancement factor ξ Field enhancement factor ξ
10

Anti-resonant high-finesse SESAM (first SESAM design)
Optics Lett., vol. 17, 505, 1992
IEEE JSTQE, vol. 2, 435, 1996
Anti-resonant low-finesse SESAM
Anti-resonant Fabry-Perot
ϕ rt
= ϕ b
+ ϕ t
+ 2knd = ( 2m + 1)π
Anti-resonant Fabry-Perot
ϕ rt
= ϕ b
+ ϕ t
+ 2knd = ( 2m + 1)π
ϕ b
= π
ϕ t
= 0
d = m λ
2n
ϕ b
= π
ϕ t
= 0
d = m λ
2n , m = 1
IEEE JSTQE, vol. 2, 435, 1996
Anti-resonant low-finesse SESAM
High- versus low-finesse anti-resonant SESAM
final high-finesse
anti-resonant SESAM
R top = 96%
bottom Bragg mirror
R bottom ≈ 100%
Anti-resonant Fabry-Perot
ϕ rt
= ϕ b
+ ϕ t
+ 2knd = ( 2m + 1)π
optical pulse
spectrum
SESAM without R top
AR-coated
high-finesse SESAM design:
AlGaAs/AlAs Al x Ga 1-x As SiO 2 /TiO 2
x = 0.09
saturable
absorber
ϕ b
= 0
ϕ t
= 0
d = λ
4n
Absorber embedded into a Bragg reflector
(quarter-wave stack of alternating high and
low index material)
Interferometric
autocorrelation (IAC)
SESAM modelocked Ti:sapphire laser
I. D. Jung et al, Opt. Lett. 20, 1559, 1995
11

High- versus low-finesse anti-resonant SESAM
High- versus low-finesse anti-resonant SESAM
final low-finesse
anti-resonant SESAM
bottom Bragg mirror
R bottom ≈ 100%
low-finesse SESAM design:
AlGaAs/AlAs GaAs SA AlO 2
d
nd = λ 2
SESAM modelocked Ti:sapphire laser
I. D. Jung et al, Opt. Lett. 20, 1559, 1995
SESAM modelocked Ti:sapphire laser: I. D. Jung et al, Opt. Lett. 20, 1559, 1995
Ultrabroadband SESAMs
Semiconductor AlGaAs/AlAs Bragg mirror replaced by silver mirror:
Ultrabroadband SESAM
25 pairs
Al 0.17 Ga 0.83 As/AlAs
4 pairs
ZnTe/BaF 2
4 pairs
Al 0.8 Ga 0.2 As/CaF 2
R. Fluck et al., Optics Lett. 21, 743, 1996
S. Schön et al., Appl. Phys. Lett. 77, 782, 2000
12

Ultrabroadband SESAM
SESAMs to reduce Q-switching instabilities
n
n o ●
n e ▲
2005
1) SESAMs with low saturation fluence:
2) SESAMs with inverse saturable absorption:
S. Schön et al., Appl. Phys. Lett. 77, 782, 2000
Inverse saturable absorption (ISA)
Consequences for the QML Threshold
SESAM reflectivity for a pulse fluence F p
the reflectivity decreases
at higher pulse energies
the roll-over =
inverse saturable absorption
R ISA
(F p
) = R P
(F p
)
• F 2 is the inverse slope
of the roll over
− F p
F 2
• The smaller F 2 , the
stronger is the roll-over
No inverse sat. absorption
E P
> E sat,L
ΔR
S
C. Hönninger, R. Paschotta, F. Morier-Genoud,
M. Moser, and U. Keller
J. Opt. Soc. Am. B 16, 46 (1999)
E P Intracavity pulse energy
E sat,L Gain saturation energy
S = E P /E sat,A Saturation Parameter
ΔR ⎛
E P
> E sat,L
S 1 − S 2 ⎞
⎜ 2 ⎟
⎝ S 0
⎠
R. Grange, M. Haiml, R. Paschotta, G. J. Spühler,
L. Krainer, O. Ostinelli, M. Golling, U. Keller
Appl. Phys. B 80, 151 (2005)
S 0
=
F 2 ΔR
F sat
• Less demands on the mode size in the gain medium
• More flexibility for cavity design
• Easier to suppress Q-switched mode locking
With inverse sat. absorption
Saturation parameter at
maximum reflectivity
13

All-optical 100-GHz pulse generation at 1.5 µm
Autocorrelation
Optical spectrum
SESAM modelocked Er-Yb:glass (ERGO) laser
enabled world-record optical communication results
Time-Bandwidth Products, Inc. makes such lasers commercially available
ERGO laser, http://www.time-bandwidth.com
Photo of actual setup
SESAM modelocked Er-Yb:glass laser
Pulse repetition rate: 101 GHz
Max. output-power: 35 mW
Optical bandwidth: 2.6 nm
Pulse width: 1.6 ps (1.7x TBP)
A. E. H. Oehler et al., Opt. Express 16, 21930, 2008
Resonant SESAM: both SAM and negative GDD
Resonant SESAM: both SAM and negative GDD
Position of saturable absorber (SA):
(a) close to a node (lower loss)
(b) at peak of standing wave: strong
absorption dip at resonance
D. Kopf et al, Opt. Lett. 21, 486, 1996
D. Kopf et al, Opt. Lett. 21, 486, 1996
14

Resonant SESAM: both SAM and negative GDD
SESAM without damage (even at >100 kW)
2012
Position of saturable absorber (SA):
(a) close to a node (lower loss)
(b) at peak of standing wave: strong
absorption dip at resonance
D. Kopf et al, Opt. Lett. 21, 486, 1996
High average power lasers: SESAM ML TDL
TDL
A. Giesen, et al., Appl. Phys. B 58, 365 (1994)
Moving from ss-lasers to semiconductor lasers
Vertical external cavity surface emitting laser (VECSEL)
Modelocked integrated external-cavity surface emitting laser (MIXSEL)
Appl. Phys. B, vol. 88, Nr. 4, pp. 493-497, 2007
Modelocked VECSEL
C. J. Saraceno et al., >270 W (!)
Opt. Express 20, 23535, 2012
First time >10 µJ pulse energy from a SESAM modelocked Yb:YAG thin disk laser (TDL)
Opt. Express 16, 6397, 2008 and CLEO Europe June 2007
26 µJ with a multipass gain cavity and larger output coupling of 70% (Trumpf/Konstanz)
Opt. Express 16, 20530, 2008
Important Milestones:
!
MIXSEL
MIXSEL with 6.4 W average output power: Opt. Express 18, 27582, 2010
femtosecond VECSEL with > 1 W average output power: Opt. Express 19, 8108, 2011
15

More info on the web ...
• Web page of Prof. Ursula Keller at ETH Zurich: http://www.ulp.ethz.ch
• All papers are available to download as PDFs:
http://www.ulp.ethz.ch/publications/paper
• SESAM milestones: http://www.ulp.ethz.ch/research/Sesam
• Ultrafast solid-state laser: get started with the book chapter ...
http://www.ulp.ethz.ch/research/UltrafastSolidStateLasers
• VECSEL and MIXSEL: http://www.ulp.ethz.ch/research/VecselMixsel
• Frequency combs: http://www.ulp.ethz.ch/research/FrequencyComb
• Attoscience, Attoclock, Attoline:
• http://www.ulp.ethz.ch/research/Attosecondscience
• http://www.ulp.ethz.ch/research/Attoline
• http://www.ulp.ethz.ch/research/Attoclock
• Viewgraphs to graduate lecture course: Ultrafast Laser Physics
http://www.ulp.ethz.ch/education/ultrafastlaserphysics/viewgraphs
16