Transcript Document

keV Harmonics from Solid Targets
The Relatvisitic Limit and Attosecond pulses
Matt Zepf
Queens University Belfast
B.Dromey et al. Queen’s University Belfast
K. Krushelnick et al, Imperial College
P. Norreys et al, RAL
Outline
High Harmonic Generation from Solid Targets
Harmonics from solid targets – Background
Experimental results
The relativistic limit – high conversion efficiencies
keV harmonics – coherent fs radiation
Angular distribution- beamed keV radiation
Potential for very bright attosecond pulse generation
Ultra High Harmonic Generation - the principle
• High power pulse tightly focused onto a solid target
• Critical surface oscillates with v approaching c
Relativistically oscillating mirror  = (1+(a0)2/2)1/2
Incident Pulse
Process intrinsically
phased locked for
all harmonics!
Zeptosecond pulses
possible at keV
Reflected Pulse
•
Reflected waveform is modified from sine to ~sawtooth
Harmonic efficiency is FT of reflected waveform
Train of as pulses (analogous to mode-locking)
Typical spectra –
Conversion efficiency follows power law scaling
Conversion efficiency scales
q~n-p
With p=5.5…3.3 for
I=5 1017…1019Wcm-2
(a0=0.6 .. 3)
From Norreys, Zepf et al., PRL, 1832 (1996)
Very high orders become rapidly more efficient at high intensities
e.g. 100th harmonic~I3
PIC predicts q~n-2.5 >1020Wcm-2. (a0>10) and 1000s of orders
Duration of attosecond pulses
Extremely short pulses are possible
by filtering the phase locked HHG
nF
(G. D. Tsakiris et al.,New J. Phys. 8, 19(2006)
Dn=(21/p-1)nF
Harmonic efficiency slope as n-p
Atto pulse efficiency:
~n-p+1~n-1.5
Pulse duration (as)
1000
100
Few as pulses
possible <1keV
10
1
10
100
1000
0.1
0.01
nF Zeptosecond@
>1keV
10000
Realistic experimental configuration
(G. D. Tsakiris et al.,New J. Phys. 8, 19(2006)
Filters (~0.1µm thick) have negligible dispersion
Consequences from the oscillating mirror model
Oscillating Mirror
Flat, sharply defined
critical density surface
• Flatness results in specular
reflection of the harmonics
Surface denting/bowing in
response to laser can change
collimation.
Surface roughness important
for Ångstrom radiation.
• Well defined mirror surface
gives high conversion
efficiency
Phase locked harmonics
– as pulses possible
Harmonic efficiency depends
strongly on plasma scale length, L
L/  0.1-0.2
Short, high contrast pulses appear ideal.
Single cycle pulses to generate atto pulses
Experimental Setup:
Incident laser
pulse: f3 cone
Double plasma
Mirror Setup
Target position
Grating spectrometer or von
Hamos crystal spectrometer
CCD or image
plate detectors
Pulse Energy:
up to 500J
Pulse energy with PM:up to 150 J
Pulse duration: 500-600fs
Contrast (no PM) >107:1
Contrast with PMs:
>1011:1
Peak intensity (with PM) 2.5 1020Wcm-2
Relativistic scaling pREL=2.5
Experimental data from
Vulcan PW shows
p=2.5.2 for a=10
HIGH EFFICIENCY
10-4@60 eV (17nm)
10-6@250eV (4nm)
Extremely high photon numbers
and brightness:
10131 photons
10231ph s-1mrad-2 (0.1%BW)
Published: B. Dromey et al, Nature Physics, 2006
keV harmonics + the efficiency roll-over
Intensity/ /arb. units
Normalised at 1200th order
10
1.5.5x1020 Wcm-2
2.5 .5x1020 Wcm-2
~n-2.55 ±.2
1
Intensity dependent roll-over
10-1
Harmonic efficiency n-2.55
 Relativistic limit
10-2
1200
1414KeV
Order, n
Photon Energy
First coherent, femtosecond,
sub-nm source
3200
3767KeV
I
FWHM 1’ ~ 500fs
t
Roll over scaling confirmed as ~3
Roll-over measurements
83
Roll-over position (order n)
10000
42
1000
100
10
1
1
Vulcan 1996
highest observed
22
10
a0
(6
1020Wcm-2m2)
100
Roll over ~3  10 keV pulse @ a0~30 (1021Wcm-2m2)
Standard contrast (~10-7) – Bright thermal emitters.
Intensity/ arb. units
1
0.8
kT~3keV
0.6
2.5x1020Wcm-2
0.4
kT~1.5keV
0.2
2
7x1019Wcm-2
3
4
5
6
7
Wavelength /Å
8
Planckian Spectrum observed for standard contrast
Signal brightness ~2x HHG signal
Plasma mirrors are essential
Absorption much higher for low contrast pulses.
Beamed keV harmonic radiation - coherent keV radiation
X-ray Signal > 1 keV
1
0.8
0.6
4º FWHM Gaussian fit
to beamed HHG signal
0.4
0.2
-100
50
0
50
100
150
specular
Angle from target normal/deg
(Specular reflection 45º, incident -45º)
X-ray emission above 1keV and 3w is beamed into ~f/3 cone (laser also f/3)
for nm rms roughness targets.
No beaming observed for
-shots with micron rms targets
-shots without plasma mirrors
Surface denting
Laser
Ponderomotive pressure can deform surface.
(under the current conditions some deformation is unavoidable
Denting required to explain our results:~ 0.1m
This would lead to the same divergence for all harmonics in
agreement with results.
 Solution: use shorter pulses to prevent surface deformation
Summary
• Harmonics from solids are efficient way of producing as
pulses up to keV photon energies.
•Ideal for converting ultra high power pulses (100’s of TW)
•HHG in the relativistic limit has been demonstrated.
• Simple geometry for as-pulse production (surface harmonics,
phase locked with flat phase, dispersion free system)
•Two possible schemes:
polarisation switching or single cycle pulses
•Angular divergence limit remains a question mark:
have we reached DL performance?
•Contrast requirements (>1010) are a challenge for fs lasers
Surface roughness
Laser
Surface roughness would impact on the highest orders only
-Unlikely to be a major factor in this experiment
Solution: highly polished targets
Imprinted phase aberration
Phase errors in fundamental beam are passed on to harmonics
Dfn~n DfLaser
Divergence of harmonics can be strongly affected (cf doubling
of high power laser beams)
The cut-off question.
Until recently no firm theoretical basis for a cut-off
Should one expect a cut-off?
Harmonic spectrum is simply FT of reflected waveform
no cut-off infinitely fast risetime components (unphysical)
 Recently: Rollover for n> 42
(Gordienko et al (PRL,93, 115002, 2004)
 Revised theory predicts rollover for n>81/23
(T. Baeva et al, PRE and talk after break)
Very different predictions for reaching 10,000 harmonics:
42: a0=50
81/23: a0=22
What determines the angular distribution?
1) What determines the angular distribution?
Diffraction limited peformance would suggest qharmonic~qLaser/n
 qharmonic~10-4 rad for keV harmonics.
2) Why do keV harmonics beam at all?
Surface roughness should prevent beaming
(Wavelength<< initial surface roughness for keV harmonics)
what reduces the surface roughness
a) smoothing in the expansion phase?
b) Relativistic length contraction
(highest harmonics are only generatedat max. surface )
High Efficiency
Assuming 1J,5fs
(projected ELI front end)
Spectral
range
Number of
photons
Pulse
duration
20-70 eV
(Al filter)
~7 *1015
84 as
80-200 eV
(Zr filter)
~2*1014
38 as
400-1000 eV
(Cu filter)
~2*1012
5 as
Extremely powerful attosecond source
Ultrahigh brightness may be possible with DL performance
Experimental paramters
Pulse Energy (No Plasma Mirror):up to 500J
Pulse energy with PM:
up to 150 J
Pulse duration:
500-600fs
Contrast (no PM)
>107:1
Contrast with PMs:
>1011:1
Spot size:
~7m
Peak intensity (with PM)
2.5 1020Wcm-2
Attosecond pulses by spectral filtering
Removing optical harmonics + fundamental changes wave from
from saw-tooth to individual as-pulses and sub-as pulses
from (G. D. Tsakiris et al.,New J. Phys. 8, 19(2006)
PIC predicts asymptotic limit of pREL~2.5-3
Exact value of p is pulseshape dependent
Gordienko et al. PRL 93, 115001, 2004
Orders > 1000,
keV harmonics!
Conversion efficiency
Conversion efficiency into attosecond pulses
1.00E+00
1.00E-01
1
10
100
1000
10000
1.00E-02
1.00E-03
~n-3/2
1.00E-04
1.00E-05
1.00E-06
1.00E-07
Centre frequency (n)
Conv eff at filter peak: f|~(nf)-p
Bandwidth: Dn~(21/p-1)nF
Pulse efficiency: pulse~(21/p-1)nF-(p-1)~n-3/2
Laser contrast is the key to high efficiency.
1.3x104
1.2x104
1.1x104
Shot 1:
Contrast 1011
(2 plasma mirrors)
Signal (arb)
1x104
9x103
Reference Spectrum (arb.)
Harmonic Spectrum (arb.)
8x103
~
~
1200
~ Strong harmonic signal.
~
b) No plasma mirror
C-line @3.4nm
Contrast ~10-8
1000
800
Shot 2:
Contrast 107
(No plasma mirrors)
C-line @4.01nm
600
Source Broadening increases
linewidth in no PM case
400
200
0
Weak C-line emission
360
380
400
420
440
Pixel number
460
480
500
Harmonics >100x brighter than thermal source in water window