Use of Fringe-Resolved Autocorrelation for

the Diagnosis of the Wavelength Stability

of KU-FEL

and single-shot measurement of mid-IR laser spectra

Institute of Advanced Energy

Kyoto University

Takashi Nakajima, Yu Qin, Xiaolong Wang,

Heishun Zen, Toshiteru Kii,

and Hideaki Ohgaki

Outline

• Introduction

• Autocorrelation

intensity autocorrelation (IAC) fringe-resolved autocorrelation (FRAC)

• Idea to estimate the shot-to-shot change of

laser wavelength

spectrally unresolved method using IAC and FRAC

• Single-shot measurement of mid-IR FEL spectra

sum-frequency mixing

• Summary

Introduction

KU-FEL: Oscillator-type FEL for the mid-IR (5-14 μm)

time

macro pulse

micro pulse

~ µs

~ 350 ps

pulse structure of KU-FEL

~ ps

0.2 μm

Shot-to-shot change of laser wavelength

If 1 ps pulse and

1 % change of λ0

scan

12 11.5 13 (μm)

FEL wavelength

1.0

0.5

0

Inte

ns

ity

λ0

different macropulse

different micropulse

scanning-type spectrometer

array-type mid-IR photodetector

For the detection of single-micropulse spectra

Autocorrelation methods

Nonlinear crystal

filter

Photo

detector

τ dttEtE

4

21 )()(

FEL pulse

Movable

stage

FRAC signal

ω

2ω

【Fringe-resolved autocorrelation (FRAC)】

dttItI )()( τ

2ω

Photo

detector

Inter-

ferometer

【Intensity autocorrelation (IAC)】

FEL pulse

Movable

stage

Nonlinear crystal

IAC signal

Delay time τ

τp 2

IAC

sig

nal

Techniques to measure the ps~fs pulse duration

Delay time τ

FRA

C s

ign

al

ω

Idea to estimate the shot-to-shot change of laser wavelength

How does it work? λ(k) λ(k+1)

train of

micropulses time

interference smeared out

for larger delay if λ(k)≠λ(k+1)

c/ period noscillatio

λ(k) λ(k+1) λ(k)+ λ(k+1)

+ = narrowing

replica

4)1()1(

4)()(

)()(

)()()(

tEtE

tEtEI

kk

kk

FRAC

λ(k)

𝒄

λ(k+1)

𝒄

)(exp1

1)1(2ln2exp)( )()(

0

)(

3

2

0

)(

2

)(

0

)(

1

)( ttiaia

tktIatE kkk

k

kkk

M

k

k tEtI1

2)( )()(

IAC and FRAC signals with various fluctuations

autocorrelation

signals

dttItII IAC )()()(

dttEtEIN

k

kk

FRAC

1

4)()( )()()(

Intensity of the macropulse

Field amplitude of the kth micropulse

micropulse k the of

frequency central and duration, intensity, of nfluctuatio :, ,

micropulse kth the offrequency central :

duration micropulse

th

(k)

0

)(

3

)(

2

)(

1

:

kkk

p

aaa

0

21 p

0% 0.5%

1% 1.5%

Delay time (ps) Delay time (ps)

1.53ps 1.36ps

1.09ps 0.89ps

Narrowing of FRAC signals by wavelength fluctuation

After averaging over thousands of shots,

IAC signals : no change

FRAC signals: narrowing by shot-to-shot wavelength change

FRAC signal narrowing

Wavelength fluctuation

For 1 ps pulse,

Qin et al., Jpn.J.Appl.Phys. (in press)

2

2

0

22

0

2

2

22

02

2222

0

2

2

22

4

2ln2exp2

2cos2)1(2ln2

exp

cos2

2lncos

22

)3(2lnexp4

1)()(

p

p

pp

dttEtE

Retrieval of pulse duration from FRAC signals

If we can retrieve the pulse duration from the FRAC signals,

we may only perform the FRAC measurement !

Expectation values of

0

0

1.0

ps p

delay (ps)

Pulse duration is retrieved by the

time-integration of FRAC signals p

1)( FRACI

d

d

d

pFRACId

2ln

21)(

1)( FRACI

Experimental results

step size = 2 μm (= 1/150 ps)

FRAC measurement

exp.envelope

fitted envelope

delay (ps)

exp.envelope

fitted envelope

delay (ps)

step size = 10 μm (= 1/30 ps)

IAC measurement

delay (ps)

exp.data

Gaussian fitting

%4.1

0

66.0

ps p

%0

52.1

66.0

ps p

ps 64.02

9.0 p

0.9 ps

• pulse duration 0.64~0.66 ps by

both IAC and FRAC methods

• wavelength stability is < 1.4 %

Qin et al. (in preparation)

Toward single-shot measurement of Mid-IR FEL pulse spectra

Mid-IR : MCT-based array photodetector (very expensive)

Near–IR: Si-based array photodetetor (cheap!)

Convert the MIR laser pulse into the NIR pulse by sum-frequency mixing !

ω1 (11 μm) + ω2 (1064 nm) → ω3（970 nm）

KU-FEL Nd:YAG or

microchip lasers

SFM signal

(near-IR)

Signal intensity

ISFM ∝ IMIR x I1064nm

< 1 ps 20 ns or

0.8 ns

< 1 ps

Wang et al. (in preparation)

spectrometer

shortpass

filter

nonlinear

crystal

(AgGaS2)

SFM signal

~970 nm

Schematics of sum-frequency mixing

SFM signal detectable?

Experimental setup for sum-frequency mixing

2 mm

5 mJ/pulse for YAG

(20 ns)

100μJ/pulse

for microchip laser

(0.8 ns)

KU-FEL

@11 μm

5mJ/macropulse at 1Hz

~5000 micropulses

YA

G / m

icroch

ip

spectro -meter

SP filter

Spectral resolution

~0.8 nm

KU-FEL

macropulse

control

delay 1064 nm

pulse

λ/2

AgGaS2

Shot-to-shot change of the SFM spectra by 20 ns Nd:YAG laser

Sum of the micropulse spectra in the same macropulse

𝟐𝟎 𝒏𝒔

𝟑𝟓𝟎 𝒑𝒔= 𝟓𝟕

SF

M s

ign

al

966 968

Wavelength (nm)

972 974 970

1064nm pulse

macro pulse

micropulse separation

20 ns

350ps

SFM

Wang et al. (in preparation)

𝛌𝐅𝐄𝐋 = (𝟏

𝛌𝐒𝐅𝐌−

𝟏

𝛌𝟏𝟎𝟔𝟒𝐧𝐦)−𝟏

retrieved

FEL spectrum

Retrieval of the FEL spectra

Wavelength (μm)

raw spectrum

SF

M s

ignal (a

.u.)

Wavelength (nm)

deconvolved

spectrum

FEL spectrum by

monochrometer

FE

L s

ignal (a

.u.)

Wang et al. (in preparation)

Time (μs) wavelength (μm)

Inte

nsi

ty (

a.u

.)

Inte

nsi

ty (

a.u

.) raw

deconvolved

Macropulse shape FEL spectrum

Time (μs) wavelength (μm)

Inte

nsi

ty (

a.u

.)

Inte

nsi

ty (

a.u

.) raw

deconvolved

Macropulse shape FEL spectrum

Time (μs) wavelength (μm)

Inte

nsi

ty (

a.u

.)

Inte

nsi

ty (

a.u

.) raw

deconvolved

Macropulse shape FEL spectrum

Summary

Measurement of the pulse duration and wavelength stability

of KU-FEL using fringe-resolved autocorrelation

Spectrally unresolved detection !

Periodicity of the fringe wavelength information

micropulse duration = 0.64~0.66 ps

wavelength instability <1.4%

Single-shot measurement of the laser spectra of KU-FEL for temporally selected micropulse(s)

Sum-frequency mixing to generate near-IR signals