Transcript Document

Soft x-ray optics and beamlines for
next generation light sources
Mark D Roper
Accelerator Science & Technology Centre
STFC Daresbury Laboratory
Talk Outline
•
•
•
•
•
Photon Properties
Transportation
Diagnostics
Conclusion
Questions
If I could summarise everything that was of concern in
soft x-ray optics for future light sources in 30 minutes,
this lecture would probably not be worth giving.
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Properties of a FLS Light Beam
• Coherent wavefront
– Diffraction limited
• Pulsed
– Shot to shot variation
• Short pulse length
– Transform limited
• High pulse energy
– Damage
• Wavelength dependence
– In ways not familiar from conventional sources
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Transportation
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Where is the source?
• As optic gets closer than ZR, source will look like it
is at infinity.
– Not very likely for an x-ray source
• Still need to ask
–
–
–
–
–
Where is the source?
How big is the source?
What is the M2 propagation factor?
Is it the same horizontally and vertically?
How do these factor vary with wavelength?
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Source characteristics for NLS FEL
Deduced from Genesis simulations using
wavefront propagation (FOCUS code) and
second moment analysis
Roper, Thompson, Dunning. J.Mod.Opt.
(2011)
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Source shape for NLS FEL
Deduced from Genesis simulations
using wavefront propagation
The transport system has to cope
with a source of changing size,
position & quality
Four undulator modules
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Preserving the wavefront
• Reflection imprints defects in the mirror surface
onto the wavefront
QuickTime™ and a
decompressor
are needed to see this picture.
Small defects also
give “speckle”
diffraction patterns
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
M. Zangrando
FERMI@Elettra
[email protected]
Preserving the wavefront
Typical SR mirror
Quic kTime™ and a
dec ompres sor
are needed to s ee this pic ture.
M. Zangrando
FERMI@Elettra
Wavelength
(nm)
40
40
40
10
10
10
5
5
5
1.67
AoI
(°)
6
3
1.5
3
2
1
3
2
1
3
P-V shape error (nm)
=0.25° =0.10°
47
18
95
38
191
76
23
9
35
14
71
28
12
5
18
7.2
36
14
4
2
The demands on optical
manufacturing and
metrology are
unprecedented
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Preserving the wavefront
Don’t forget that a coherent wave will
diffract from the edges of mirrors!!
100 eV. Simulation with FOCUS code
Implications for:
• Diffraction limited focusing
• Wavefront dividing beam-splitters
• Knife-edge position monitors
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Diffraction Limited Focusing
• The fringes will not be (so) visible at a focus
– Size of focus limited by the aperture through diffraction
6: +11%
f = 0.2 m
4: +38%
8: +2.5%
Relative to infinite aperture
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Focus limit from surface errors
Field @ source
FLASH BL3, 98 eV
Field @ focus
Focus source
with imperfect
ellipsoid at 37.5:1
demagnification
PSD of mirrors
Temporal profile
@ source
M.A.Bowler
B.Faatz
F.Siewert
Temporal profile
@ focus
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Beam splitters
• Significant demand for multi-photon experiments
• Wavefront division
– Technologically easier - knife-edged mirror
– Diffraction effects
– Auto-correlator (beam splitter & delay line) at FLASH
• Amplitude division
–
–
–
–
Reflection-reflection or reflection-transmission
Gratings, multi-layers, (crystals)
Pulse length effects
Flatness of thin membranes
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Metrology & Manufacturing
Achieving the highest possible figure accuracy requires collaboration
between the manufacturer and the metrology laboratory
HZB: NOM Metrology Data +
Zeiss: Ion Beam Finishing
Reduction in form error of elliptical
focusing mirrors by factor of 3
H. Thiess, H. Lasser, F. Siewert, NIM A (2009)
F. Siewert, J. Buchheim, T. Zeschke, NIM A (2010)
Before
After
Slope (arcsec)
F. Siewert
HZB
1.6 µrad to 0.5 µrad RMS
Height (nm)
61 nm to 22 nm PV
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Preserving the Pulse Length
• To preserve the pulse length to 1 fs, the optical
path length must be the same to 0.3 µm for all
positions across the wavefront from source to final
image (distance 10’s to 100’s metres).
– Tight control of all aberrations
– Control of penetration depth into multi-layers
– Special attention to dispersing elements like gratings
• The pulse bandwidth must be preserved
– Transform limit
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Gratings and short pulses
• The path length difference with a diffraction grating
will stretch the pulse
• Low line density
gratings and controlled
illumination*
• Conical diffraction
geometry
• Double gratings
* Roper, NIM A (2010)
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Photon Induced Damage
• Damage from the high fluence pulses to the optical surfaces
is a major concern
• The main approach to protection is
– Use the most robust coating (lighter elements)
– Spatially dilute the beam (distance & grazing angle)
– Calculate absorbed dose per atom (geometry, reflectivity,
penetration) and make sure it is below the “damage threshold”
• Amorphous carbon (a-C) most popular XUV coating
(FLASH) but no good >280 eV
• Cr, Ni, even Pt may be needed.
• Damage mechanisms are complicated and not fully
understood
– What is the “damage threshold” (e.g. function of wavelength)
• Effect on structured surfaces
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
a-C Single-shot Damage
Threshold fluence for damage
as a function of grazing angle.
Chalupský et al., Appl Phys Lett 95 031111 (2009)
Two regions of damage
 Central ablation
 Peripheral expansion (graphitization)
Electron transport in the a-C is key in
determining the absorbed dose per atom
below the critical angle
Damage occurs well below
the melt threshold
Nomarski microscope images of
damage by 13.5 nm radiation.
Beam at normal incidence (left)
and 18.5° grazing angle (right).
FLASH Measurements
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Multi-shot damage in a-C
• Multi-shot damage observed in a-C
• Each shot is below threshold for
single shot damage
– 0.5 J/cm2, 46.9 nm, 1.7 ns, CDL
– 5 shots  no observable damage
– 10 - 40 shots  progressive erosion
Juha et al., J. Appl. Phys 105, 093117
(2009)
University of
L’Aquila
a-C complex behaviour
10 shots
 Low fluence multi-shot =>
photo-induced erosion
without chemical change
 High fluence => expansion
due to graphitization
40 shots
AFM Image
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Active Optics
• Significant usage of active optics is certain
• Achieving better quality foci
– Use plane surfaces (easier to make) and benders
– Correct residual errors in the manufactured surface
– Correct wavefront distortion caused by errors in the
surface of other optics
• Tailored focusing
– Different spot sizes (without sitting off-focus)
– Tailored spot shapes (e.g. top hat, Lorentzian)
• Compensating for the moving source position
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
FERMI@Elettra K-B System
Both DiProI and Low Density
Matter will use a KB active
optics system to give a small
spot taking into account the
source variation between
FEL1 and FEL2 and the
necessary optical quality of
the surfaces, achievable only
on plane surfaces.
M. Zangrando
FERMI@Elettra
QuickTime™ and a
decompressor
are needed to see this picture.
QuickTime™ and a
decompressor
are needed to see this picture.
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Modelling
• Geometric vs physical optics
– Ray-tracing will still play a big part in designing a
beamline
• Checking aberrations, alignment tolerances etc
• Because it’s fast!!
– Previous slides show physical optics simulations are
essential
• Modelling with Genesis source simulations
• Determining the actual source properties
• Coherence effects from apertures and mirror imperfections
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Diagnostics
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Diagnostics
• The challenge of the ideal diagnostic
– Measure every pulse in real time (@ Hz to MHz)
– Non-invasive
• Transparent to the beam
• Require no special optics
– In situ
• Always “on-line”
• To measure
–
–
–
–
–
–
Pulse energy
Pulse length
Longitudinal and transverse intensity profiles
Timing jitter (relative to something useful) at fs or as level
Spectral content (and phases)
Polarisation
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Pulse Length Measurement
• Cross-correlation with IR laser
– Side-band generation in the presence of an intense IR
field during the photo-ionisation of a noble gas by the
FEL beam
FEL beam needs to be focused impacts on beamline layout
Multi-shot (scan laser delay)
Photo-electrons must be
spectrally analysed
Meyer, M., et al.,
Two-colour photoionization in xuv freeelectron and visible laser fields.
Phys Rev, 2006. 74, 011401
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Pulse Length Measurement
• Single shot cross-correlation by looking at the
intensity and number of the sidebands
• Also gives jitter information (relative to IR laser)
Radcliffe, P., et al.,
Single-shot characterization of
independent femtosecond extreme
ultraviolet free electron and infrared
laser pulses.
Appl Phys Lett, 2007. 90, 131108
Other approaches include
“time to space” mapping
Cunovic, S., et al., Time-to-space mapping in a
gas medium for the temporal characterization
of vacuum-ultraviolet pulses. Appl Phys Lett,
2007. 90, 121112
FEL at 89.9 eV
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Pulse Length “Holy Grail”
• Intensity autocorrelation gives only limited pulse profile
information
• A soft x-ray analog of FROG or SPIDER is needed for
complete pulse characterisation
– Requires a non-linear process to give a signal that is proportional to
the autocorrelation function
• Beam mixing (FROG) or spectral shear (SPIDER)
– (Almost) certainly will involve measuring photo-electrons
• Two-photon ionisation (one or two colours)
• Single-photon multiple-ionisation
• Optical phase and spectral information encoded onto photoelectrons, requires electron spectrometers
• Challenging experiments, limited by spectrometer performance
and wavelength coverage may be limited by gases available
– Autocorrelation, so no timing jitter info
Remetter, T., et al.,
Attosecond electron wave packet
interferometry. Nat Phys, 2006. 2, 323
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Other Diagnostics
• Wavefront
– Hartmann sensor
• Pulse energy
– Gas cell
– Can be expanded to measure wavelength & harmonics
• Spectrum
– VLS grating spectrometer (zeroth order to experiment)
• Position & angle
– Blade monitors (diffraction & damage)
– Ionisation chambers (sensitivity and accuracy)
• Polarisation
– Wideband ML Polarimetry (F. Schäfers)
– Full Stokes vector in a single shot??
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Conclusions
• Ultra-short and transversely coherent SXR pulses present a
new challenge to the beamline designer
–
–
–
–
–
Spectral dependence of even the most basic source properties
Diffractive disruption to the wavefront
Stretching the pulse
The risk of damaging the optical surfaces
Requirement for physical optics modelling
• We also have to account for the shot to shot variation in the
source
– Diagnostics need to be an integrated part of the beamline
• Many areas are at least partly addressed
– There is more that needs to be done
– Progress will follow as sources come on stream
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]
Thank you for your attention
5 March 2012
FLS2012, Thomas Jefferson National Laboratory
[email protected]