femtosecond laser Daniele Brida 17.02.2016 Universitt Konstanz - - PowerPoint PPT Presentation

femtosecond laser
SMART_READER_LITE
LIVE PREVIEW

femtosecond laser Daniele Brida 17.02.2016 Universitt Konstanz - - PowerPoint PPT Presentation

How to build an Er:fiber femtosecond laser Daniele Brida 17.02.2016 Universitt Konstanz Konstanz Universitt Konstanz Ultrafast laser Time domain : pulse train Frequency domain: comb 3 26.03.2016 Universitt Konstanz Frequency comb


slide-1
SLIDE 1

Universität Konstanz

How to build an Er:fiber femtosecond laser

Daniele Brida 17.02.2016

slide-2
SLIDE 2

Universität Konstanz

Konstanz

slide-3
SLIDE 3

Universität Konstanz

Ultrafast laser

26.03.2016 3

Time domain : pulse train Frequency domain: comb

slide-4
SLIDE 4

Universität Konstanz

Frequency comb laser

26.03.2016 4

Time domain : pulse train Frequency domain: comb

slide-5
SLIDE 5

Universität Konstanz

Mode locking

26.03.2016 5

Establish a precise phase relation between the modes of the cavity with a well defined phase -> pulses

slide-6
SLIDE 6

Universität Konstanz

Mode locking: How to

26.03.2016 6

Solution: Nonlinearity Kerr lens mode locking

slide-7
SLIDE 7

Universität Konstanz

Ti:sapphire laser

26.03.2016 7

Time domain : pulse train Frequency domain: comb

slide-8
SLIDE 8

Universität Konstanz

Fiber lasers

Guided operations: the mode is confined in an optical fiber PRO

  • Virtually alignment free
  • Robustness
  • Weakly affected by the environment
  • Stability

CONS

  • Careful design (you cannot optimize it)
  • (Low power)
  • (dispersion managment)
slide-9
SLIDE 9

Universität Konstanz

Possible Gain Media

Yb: 1030 nm Er: 1550 nm Tm/Ho: ~2000 nm … In general: rare earth ions in silica matrix

slide-10
SLIDE 10

Universität Konstanz

CW vs femtosecond

CW laser diode Mirror Femtosecond laser

  • > short pulses
  • > frequency comb

PROBLEM: dispersion

slide-11
SLIDE 11

Universität Konstanz

Linear propagation of short pulses Examples

slide-12
SLIDE 12

Universität Konstanz

Er:fiber laser

slide-13
SLIDE 13

Universität Konstanz

Er3+ ions as gain medium

1550 high transparency window for fused silica True 3-level system Lasing at 1550 requires significant population inversion!!

slide-14
SLIDE 14

Universität Konstanz

Er3+ ions more details

3 level system Lifetime of the lasing level is fairly long: 10 ms Green fluorescence

slide-15
SLIDE 15

Universität Konstanz

Mode locking operations in a fiber laser

Three approaches:

  • Active modulation
  • Instantaneous Nonlinearity
  • Ultrafast saturable absorber
slide-16
SLIDE 16

Universität Konstanz

Femtosecond fiber laser 1: figure of 8

Asymmetry in the path between clockwise and counterclockwise propagation The isolator is the lossy component

slide-17
SLIDE 17

Universität Konstanz

Femtosecond fiber laser 2: Polarization Rotation

Nonlinearity: XPS Typically it requires outcoupling to free space within the oscillator

slide-18
SLIDE 18

Universität Konstanz

Fiber GVD1.55 μm (ps2/km) Length (mm) F1

  • 19.7

528 F2

  • 4.76

2340 F3 0.9 393 EDF 19 680

Femtosecond fiber laser 2: Polarization Rotation

slide-19
SLIDE 19

Universität Konstanz

Femtosecond fiber laser 3: Saturable Absorber

Saturable Absorber Mirror SAM works as a mirror only if the optical power in the cavity is sufficiently high It has to show a dynamical behavior and recover the “lossy” condition really quickly

slide-20
SLIDE 20

Universität Konstanz

Femtosecond fiber laser 3: Saturable absorber

slide-21
SLIDE 21

Universität Konstanz

Germanium Saturable Absorber Mirror

slide-22
SLIDE 22

Universität Konstanz

InGaAs Saturable Absorber Mirror

Direct gap semiconductor

GaAs at the center of the Brillouin zone

slide-23
SLIDE 23

Universität Konstanz

InGaAs Saturable Absorber Mirror

slide-24
SLIDE 24

Universität Konstanz

Solitonic Oscillator

Solitonic propagation condition Where The pulse temporal profile is:

slide-25
SLIDE 25

Universität Konstanz

Solitonic Oscillator

Transform Limit pulse duration of approximately 300 fs Output power 2/3 mW

slide-26
SLIDE 26

Universität Konstanz

Femtosecond fiber laser 3: Saturable absorber

slide-27
SLIDE 27

Universität Konstanz

Femtosecond fiber laser 3: Saturable absorber

slide-28
SLIDE 28

Universität Konstanz

Femtosecond fiber laser: polarization

slide-29
SLIDE 29

Universität Konstanz

Discussion

Noise performances (Shot noise) Environmental robustness Optimization Pulse energy VS

slide-30
SLIDE 30

Universität Konstanz

Femtosecond Er:Fiber-Amplifier

Single pass amplifier 2.5 m long gain medium (Er:PM-Fiber) with normal dispersion 980 nm pump light injected from both sides (each with 700 mW) Amplification up to 330 mW, → Pin/Pout ≈ 500 Spectral broadening due to SPM (Self Phase Modulation) and other nonlinear effects in EDF and collimator fiber Recompression of the pulse in a silicon prism compressor

slide-31
SLIDE 31

Universität Konstanz

Nonlinear amplifier: dispersion managment

26.03.2016 31

Optimization of the nonlinearity during amplification by a pre-stretching fiber Also the pump diode coupling is a degree of freedom 1 co-propagating, 1 counterpropagating to optimize the inversion profile in the EDF

slide-32
SLIDE 32

Universität Konstanz

Amplifier

Bandwidth Dl = 70 nm Pulse duration TFWHM = 130 fs Degree of Polarisation > 98% 330 mW before compressor and 305 mW after compressor Pulse energy: 8 nJ Almost perfect synchronisation possible (43 as)

slide-33
SLIDE 33

Universität Konstanz

General Setup

1500 1550 1600 0.0 0.2 0.4 0.6 0.8 1.0 1540 1555 1570 0.0 0.2 0.4 0.6 0.8 1.0

  • 400 -200

200 400 0.0 0.2 0.4 0.6 0.8 1.0

Ep = 8 nJ P = 320 mW

Phase (rad)



Time (fs)

tFWHM = 128 fs

Wavelength (nm) Wavelength (nm)

P = 2.5 mW

Normalized intensity

Dl = 5.4 nm

Oscillator Spectrum Amplifier Spectrum Reconstructed FROG attosecond timing jitter:

  • F. Adler, et al.,
  • Opt. Lett. 32, 3504 (2007)

tailored spectra:

  • A. Sell, G. Krauss et al.,
  • Opt. Express 17, 1070 (2009)
slide-34
SLIDE 34

Universität Konstanz

Variable Pulse Compression

Compression in silicon prism sequence  variable prechirp Pumping of highly nonlinear fiber  tunability of dispersive wave and soliton Collimation with off-axis parabolic mirror  no chromatic aberration

slide-35
SLIDE 35

Universität Konstanz

Nonlinear Pulse Propagation

Quantitative modeling without free parameters: Gain/absorption Dispersion up to 6th order (measured via white-light interferometry) Instantaneous Kerr nonlinearity Retarded Raman effect Amplitude and phase spectra

  • f pump (measured via FROG)

 Central design tool with predictive power

                               

   1 1 2 3 3 2 2

) ( ) , ( ) , ( ) , ( 6 2 2 ) , (           

  

d R z A z A i z A i i t z A

z

slide-36
SLIDE 36

Universität Konstanz

Tailored Spectra in Highly Nonlinear Fibers I

Two-stage process 1st step: soliton compression in standard telecom fiber (l ≈ 10 cm, ØCore = 10.5 µm) Spectrum broadens and pulse is compressed to 14 fs

slide-37
SLIDE 37

Universität Konstanz

Tailored Spectra in Highly Nonlinear Fibers II

2nd step: four-photon interactions in HNF (ØCore = 4 µm) Spectrum splits into two components: Soliton Dispersive wave

slide-38
SLIDE 38

Universität Konstanz

Tuning via Prechirp

Control of nonlinear frequency shift: prechirp of pump (determines minimum pulse duration before HNF) Pout > 30 mW (dispersive wave) and > 50 mW (soliton) Spectral range covered: 800 nm to 2400 nm

time evolution in precompression fiber

0.8 1.0 1.2 1.4 1.6 1.8 2 2.2 2.4 1 2 3 4 5 6 7 Wavelength (m) Spectral power (arb.unit.)

spectral evolution in HNF

slide-39
SLIDE 39

Universität Konstanz

Ultrabroad Spectra I

Optimized dispersion profiles for ultrabroadband and unstructured spectra Quantitative agreement between simulation and experiment Maximum spectral width in dispersive wave: Dl = 580 nm Pout = 23 mW Compression in glass prism compressor

slide-40
SLIDE 40

Universität Konstanz

7.8 fs Dispersive Wave

Retrieved pulse duration: tp = 7.8 fs

 two optical cycles Bandwidth limit: 7.0 fs Good agreement between measured and retrieved spectrum Perfect match between measured and calculated autocorrelation

  • A. Sell, et al. Opt. Express 17, 1070 (2009)
slide-41
SLIDE 41

Universität Konstanz

Few-Cycle Soliton from HNF 2

Retrieved pulse duration: tp = 31 fs 5 optical cycles Fourier limit: 23 fs Average output power: 55 mW

slide-42
SLIDE 42

Universität Konstanz

Single-Cycle Setup

I I I l l l

slide-43
SLIDE 43

Universität Konstanz

Single-Cycle Pulse Synthesis

Large delay times Dt: second-order auto- and cross-correlations Decreasing Dt:

Cross- correlation shifts towards center Amplitude of central fringe increases strongly

Maximum amplitude for Dt = 0

slide-44
SLIDE 44

Universität Konstanz

Single-Cycle Pulse Characterization

Separate FROG analysis of spectral amplitude and phase of soliton and dispersive wave Amplitude ratio: linear spectrum Two missing parameters left for total characterization: Linear slope (time delay Dt) Relative phase Dj between dispersive wave and soliton

slide-45
SLIDE 45

Universität Konstanz

Single-Cycle Pulses: Results

Determination of phase spectrum from FROG traces and least-square fit of Dj and Dt to second-order autocorrelation Temporal amplitude and phase via Fourier transform Retrieved pulse duration: tp = 4.3 fs Pulse energy: Ep = 1 nJ  Single cycle of light in the telecom wavelength regime

slide-46
SLIDE 46

Universität Konstanz

Carrier-Envelope Phase Control

rep

1/ T f 

D 2

CEO

f

rep

f

rep CEO n

nf f f  

  • frequency spectrum

consists of equidistant lines with CEO-frequency

  • ffset
  • slippage of carrier

envelope phase due to group and phase velocity mismatch

  • control of CEO-frequency

essential for:

  • nonlinear physics
  • metrology
slide-47
SLIDE 47

Universität Konstanz

Passive CEP Stabilization: Input Spectra

Idea: generation of phase-stable pulses at 1550 nm via DFG, from ultrabroadband HNF spectrum goal: seed source with carrier-envelope offset frequency set to zero and subsequent amplification

CEO

f

CEO 

f

rep

f

 passive phase locking of fs-Er:fiber technology at full repetition rate of 40 MHz

slide-48
SLIDE 48

Universität Konstanz

Second Harmonic

0, j 20, 2 j + /2

Self Phase Modulation

0, j 0, j + /2

OPA Pump Signal  s, j s Signal  s, j s

Amplification does not affect CEP

Difference Frequency

 1, j 1  2, j 2  1 -  2, j 1 - j 2 - /2 white light generation in a sapphire plate supercontinuum by a hollow fiber

CEP and nonlinear processes

slide-49
SLIDE 49

Universität Konstanz

Difference Frequency Generation

DF  1  2 jDF  j 1  j 2   1 2 DF = 1- 2

(2)

if fields are phase-locked:

j 1 = j 2 + Δj j DF = Δj   (const.)

Difference-frequency generation (DFG) allows:

 manipulation of the CEP  generation of MIR light

DFG between two pulses carrying the same CEP leads to automatic phase-stabilization of the DF pulse

slide-50
SLIDE 50

Universität Konstanz

  • separation of dispersive wave and soliton for compression
  • difference frequency generation in PPLN

General Setup

  • generation of ultrabroad

spectrum in HNF

  • modulation of spectrum

via chirp of the seed pulse

≙ 1550 nm

slide-51
SLIDE 51

Universität Konstanz

  • DFG tunable from

1400 nm – 1600 nm

  • broadband DFG output
  • complete background

suppression with two 1550 nm Bragg-mirrors

Phase-locked Pulses at 1550 nm

slide-52
SLIDE 52

Universität Konstanz

Reamplification of Phaselocked Seed

6 synchronized output ports after preamp high power fiber amplifiers for extreme nonlinear optics frequency comb applications

slide-53
SLIDE 53

Universität Konstanz

Output Performance of Amplifiers

average power P = 2.1 mW @ each port after preamplifier average power P = 330 mW after main amplifier pulse duration tp = 115 fs after prism compressor inherently phase-locked 8 nJ pulses at full 40 MHz repetition rate

slide-54
SLIDE 54

Universität Konstanz

Characterization of Absolute Phase Stability

 

2

) 2 / 2 ) ( sin( ) sin(  j   j       t t

spectrum modulated by:

CEP stable

stationary inteference fringes

) sin( 1 j   

  • ctave spanning

spectrum

slide-55
SLIDE 55

Universität Konstanz

Long-term Stability of Passive Phase Lock

  • acquisition of 1000 spectra over 8 s
  • RMS of phase amounts to 0.219 rad
  • excellent long-term stability

for time-domain applications

  • integration time of 4 ms implies

average over 160,000 pulses

  • good fringe visibility indicates

extremely good short-term stability

slide-56
SLIDE 56

Universität Konstanz

Seeding Yb and Tm amplifiers

Seed high power fiber laser starting with a compact Er:fiber system. Problem: supercontinuum coherence at the output of standard PCFs Yb 1064 nm Power scalable up to a multiW regime Mature technology Tm 1950 nm Broad gain bandwidth Particularly promising for future application Dispersive wave Soliton

slide-57
SLIDE 57

Universität Konstanz

Supercontinuum coherence

Interference between the SCs generated by two distinct branches

  • f the system
slide-58
SLIDE 58

Universität Konstanz

First proof of Tm:amplifier

10 MHz Tm:amplifer 9 W pump power Amplification at 1950 nm with 2.46 W

  • utput average power
slide-59
SLIDE 59

Universität Konstanz

High repetition rate for maximum sensitivity

Er:fiber femtosecond laser seeding a high power Yb:fiber amplifier Multibranch design for advanced ultrafast applications

60 W total output power at 10 MHz repetition rate

slide-60
SLIDE 60

Universität Konstanz

Noise Performance and Long-Term Stability

peak-to-peak fluctuation: < ± 0.3% during 72 h of operation at full power

slide-61
SLIDE 61

Universität Konstanz

White Light Generation

  • 2.5 W from Yb:fiber amplifier

( less than 5% of the available power at 10 MHz! )

  • Focused into 3 mm YAG

 2 octave spanning spectrum

0.4 0.6 0.8 1.2 1.4 1.6 0.0 0.5 1.0 Intensity (norm.) Wavelength (µm)