PhotonDiag 2020

26 Oct 2020, 09:00 → 28 Oct 2020, 17:10 Europe/Zurich

Virtual only (Paul Scherrer Institut)

Mirjam van Daalen (Paul Scherrer Institut), Pavle Juranic (Paul Scherrer Institut)

The 5th FELs OF EUROPE Conference on FEL Photon Diagnostics, Instrumentation and Beamline Design, PhotonDiag, will take place at the Paul Scherrer Institut (PSI) from 26 to 28 October 2020. 

 

It will focus on the following topics:

• Temporal diagnostics
• Spectral measurements
• Coherence, profile, and position measurements
• Scientific computing, machine learning and large data management
• New developments in photon diagnostics and optics
• New detectors for photon diagnostics
• Photon Diagnostics at synchrotrons
• Photon diagnostics requirements for new experimental methods

There will be Satellite Workshops on 29 October 2020.

                                                 

Book of Abstracts PhotonDiag 2020 Book of Abstracts.pdf

Poster PhotonDiag 2020 poster.pdf

Presentation AAT-presentation_PhotonDiag2020.pdf

Programme Schedule status 21.10.2020.pdf

Monday 26 October

Facility Overview 1

Session 1

Convener: Pavle Juranic (Paul Scherrer Institut)

1 Welcome

Speaker: Gabriel Aeppli (Paul Scherrer Institut)

2 Recent status of laser synchronization system at SACLA

The precise synchronization between the optical laser and the FEL is of importance for the ultrafast pump-probe experiments. For this purpose, two techniques have been developed at SACLA. One is arrival timing monitors for the hard and the soft X-ray FEL beamlines, which have ~4 fs temporal resolution. The other is a new synchronization system, which consists of a mode-locked oscillator combined with a balanced optical-microwave phase detector (BOM-PD). The typical arrival timing jitter of the new system is ~40 fs (RMS). In this presentation, we will report the recent status of the synchronization system at SACLA.

Speaker: Shigeki Owada (Japan Synchrotron Radiation Research Institute)

3 Photon diagnostics status at PAL-XFEL

PAL-XFEL first started supporting regular user experiments in mid 2017, and total 6 instruments are currently operating at two hard X-ray (XSS: X-ray Scattering and Spectroscopy and NCI: Nano Crystallography and Imaging) and one soft X-ray (SSS: Soft X-ray Scattering and Spectroscopy) beamlines. The HX beamline provides more than 1011 photons/pulse in the range from 2.2 to 15 keV with maximum repetition rate of 60 Hz, and about 1012 photons/pulse in the range from 200 to 1200 eV are delivered at the SX beamline.
All the way from XFEL tuning to the end of the user beamtime, photon beam diagnostics is an essential part of the beamline operation. This presentation includes overview of photon beam diagnostic devices at the PAL-XFEL to monitor intensity, position, spectrum, and arrival timing of the XFEL pulses.

Speaker: Intae Eom (Pohang Accelerator Laboratory)

4 Photon diagnostic developments at the Athos and Aramis beamlines of SwissFEL

With the first lasing of Athos (the soft X-ray branch of SwissFEL) in December 2019, SwissFEL is now operating two FELs in parallel. Whereas at the hard X-ray branch (Aramis) we are now mainly focusing on further developing the experimental infrastructure as well as setting up and performing user experiments, at Athos a constant progress of the beamline commissioning and the construction of endstations takes place.

In this talk, we will present the developments of X-ray photon diagnostic tools. We will highlight the facility achievements with specific focus given to spectral and temporal diagnostics of the ultrashort x-ray pulse that are required on the one hand to optimize the accelerator, on the other to provide optimum conditions for the user experiments scheduled at the beamlines. Finally, we will discuss our plans for future developments of these tools for both, Athos and Aramis.

Speaker: Christopher Arrell

10:20 Coffee break

Temporal Diagnostics 1

Session 2

Convener: Christian Erny

5 Nonlinear optics for timing diagnosis at SwissFEL

Timing jitter is a major issue in many pump-probe X-ray FEL experiments. Current implementations of single-shot timing diagnostic tools are based either on X-ray induced changes of the optical properties of solid-state materials, or on the modulation of X-ray generated photoelectron energies via strong field interactions. These methods nearly always rely on nonlinear processes to generate suitable conditions for the measurement.
One application of nonlinear optics is the generation of broadband optical radiation, so-called “white light”, that can be used for broadband spectroscopic measurements or for the creation of ultrashort pulses. White light can be generated in a bulk crystal, a high-pressure gas cell, or a gas-filled hollow core fiber with an intense femtosecond laser pulse. White light generation in a bulk crystal has limited photon flux due to breakdown and multiple-filamentation issues. The filamentation in a high-pressure gas cell and hollow core fiber can provide greater flux in the super-continuum spectral range, but with higher noise. In addition, small changes in the gas pressure can cause significant time drifts which can impact usage in timing diagnostic tools. At SwissFEL, several systems have been developed for the short laser pulse and white light generation in order to achieve higher signal-to-noise ratio and temporal resolution.
Nonlinear optics not only provides the light sources for the timing diagnosis of FELs, but also can be applied for the process used in timing tools. Current realizations of this concept in single-shot timing diagnostics in solid state media are all based on X-ray induced changes in the linear susceptibility. In the optical spectroscopy, however, nonlinear processes often offer an enhanced sensitivity to small perturbations and in some cases can provide nearly background-free signals.
Here we present the potential of using a prototype χ(2) process, second harmonic generation, to measure the X-ray induced changes and to explore its potential applications using for an X-ray timing diagnostic tool.

Figure 1 (a) Measured the correlation between the SHG spectrum encoding and spatial encoding. (b) and (d) present the distributions of the arrival time measured with the SHG spectral timing tool and the spatial timing tool, respectively. (c) The distributions of the difference in the retrieved arrival time between the two timing tools (blue: all signal; red: signals in the timi

Speaker: Yunpei Deng (PSI)

6 Single-shot temporal characterization of SASE XUV pulses at FLASH FEL

We present a Terahertz (THz) field driven streak camera [1] with the capability to deliver the XUV pulse duration and the arrival time information with < 10 fs resolution for each single XUV FEL pulse at FLASH. Pulse durations between ~350 fs and ~10 fs (FWHM) have been measured for different FLASH FEL settings [2]. In particular the arrival time analysis showed the precision with which FLASH can be operated meanwhile. A comparison with the FEL electron bunch arrival time (BAM – beam arrival time monitor) in the FLASH linac section showed a very good correlation (< 15 fs rms). For the simulation and analysis of the streaking process, a standard classical approach was used as well as a quantum mechanical theory, based on strong field approximation. Various factors limiting the temporal resolution of the presented THz streaking setup are investigated and discussed. Special attention is paid to the long and short pulse limit. Additionally SASE (Self-amplified spontaneous emission) pulses are inherently fluctuating in various properties. The pulse resolved characterization of the XUV SASE pulses regarding pulse duration, spectral distribution and pulse energy provides a large set of data that can be used to investigate the dependencies of the different parameters. Using the measurements together with simulations we can disentangle accelerator based fluctuations from pure SASE contributions and provide more insight into the SASE process [3].

[1] R. Ivanov, J. Liu, G. Brenner, M. Brachmanski, S. Düsterer, FLASH free-electron laser single-shot temporal diagnostic: terahertz-field-driven streaking, J. Synchrotron Rad. 25 26, 2018.
[2] R.Ivanov et al. Single-shot temporal characterization of XUV pulses with duration from ~10 fs to ~350 fs at FLASH, J. Phys B: At. Mol. Opt, accepted, 2020.
[3] I. Bermúdez Macias et al., Study of temporal, spectral and energy fluctuations of SASE FEL pulses, in preparation, 2020.

Speaker: Dr Rosen Ivanov (Deutsches Elektronen-Synchrotron (DESY))

7 New concept for THz based X-ray pulse tomography

In this contribution a new concept for deriving arrivaltime, pulse duration and potentially also the pulse energy of individual XUV/X-ray pulses is presented. The concept is based on the analysis of THz bursts generated by the XUV/X-ray pulses in a novel type of spintronic THz emitter. Note only the arrivaltime jitter but also the duration of the XUV/X-ray pulses can be derived from single-shot electro-optic measurements of the emitted THz waveforms. The potential of the concept is discussed based on first pilot experiments performed recently at the FERMI FEL.

Speaker: Michael Gensch (DLR & TU Berlin)

11:50 Lunch break

Temporal Diagnostics 2

Session 3

Convener: Christian Erny

8 MHz repetition rate Photon Arrival Time Monitor at the European XFEL

Operation of the European X-ray Free Electron Laser (EuXFEL) since 2017 enables novel research of atomic-scale structures and ultrafast phenomena and dynamics. Exploiting the unprecedented high FEL peak brilliance, X-ray induced dynamics can be observed by applying X-ray pump - optical probe schemes. Typically, the timing of pulses from two independent sources can only be controlled down to the level of naturally occurring timing jitter, which limits the temporal resolution of pump and probe experiments. This unsurprisingly calls for a photon arrival time diagnostics tool to precisely determine the timing jitter and to sort and tag the experimental data.
In this contribution, we describe the design of a very compact Photon Arrival time Monitor (PAM) [1] and its commissioning results at the SPB/SFX instrument [2], located at the SASE1 hard X-ray beamline at EuXFEL. This pulse-resolved measurement of timing jitter inside pulse trains with up to MHz repetition rates, together with the determination of inter-train jitter across the 10 Hz pulse trains, represents an important step towards realizing ultrafast experiments[3,4].

References:
[1] J. Liu et al., “Technical Design Report: Photon Arrival Time Monitor (PAM) at the European XFEL,” DOI: 10.22003/XFEL.EU-TR-2017-002 (2017).
[2] A. P. Mancuso, et al., “The Single Particles, Clusters and Biomolecules and Serial Femtosecond Crystallography instrument of the European XFEL: initial installation,” J. Synchrotron Radiat. 26, 660 (2019).
[3] H. J. Kirkwood et al., “Initial observations of the femtosecond timing jitter at the European XFEL,” Opt. Lett. 44, 1650 (2019).
[4] T. sato et al., “Femtosecond timing synchronisation at megahertz repetition rates for an X-ray Free-Electron laser,” Optica 7, 716 (2020).

Speaker: Dr Jia Liu (European XFEL)

9 Post-undulator time-resolved diagnostics of electron and photon beams at SwissFEL with a passive structure.

SwissFEL is a hard X-ray free-electron laser (FEL) facility operating at the Paul Scherrer Institute in Switzerland. We recently installed a passive corrugated structure after the undulator beamline of SwissFEL. The transverse wakefields generated by the electron beam traveling through such a device can be employed to measure the time properties of the electrons. Compared to the standard transverse RF deflector approach, the method is more cost-effective, less sensitive to arrival-time jitter, but the reconstruction of the beam profiles becomes more complicated. A comparison of the longitudinal phase-space of the electron beam with and without lasing conditions also allows reconstructing the time-resolved properties of the produced hard X-ray radiation. We present simulations studies and first experimental results of this new method at SwissFEL.

Speaker: Alexander Malyzhenkov (PSI - Paul Scherrer Institut)

10 Measurement of transient optical phase as a diagnostics technique for XUV pump – optical probe experiments

Measurement of transient optical properties (reflectivity and transmissivity) is routinely performed in extreme-ultraviolet (XUV) pump – optical probe experiments. The optical properties reflect the transient state of the irradiated materials. Here we propose to extend the material diagnostics with an additional measurement of the transient phase change of the optical probe pulse. It can be recorded in parallel to other transient optical properties, enabling access to full information on the complex refractive index and the thickness of the radiation-modified material layer. The latter is essential for investigations of phase transitions progressing in XUV (and X-ray) irradiated materials.
Here we report on the computational study of XUV irradiated silicon and diamond performed in [1]. It shows that the measurement of the optical phase from a probe pulse at correctly tuned pulse parameters can provide a signal strong enough to extract information on transient material properties. The results suggest that in some cases, it is even more preferable to measure the transient phase change than other optical parameters. Such phase measurement, feasible with modern experimental setups, can then be a basis for an improved diagnostics tool for the temporal characteristics of an ultrashort XUV pulse.

[1] V. Tkachenko et al., Optics Letters Vol. 45, 33-36 (2020).

Speaker: Dr Vladimir Lipp (Center for Free-Electron Laser Science CFEL, DESY, Hamburg, Germany)

14:30 Coffee break

Facility Overview 2

Session 4

Convener: Pavle Juranic (Paul Scherrer Institut)

11 FLASH2020+: Photon diagnostics and beam transport

The FLASH2020+ project, a major upgrade program for the high repetition rate XUV and soft X-ray free-electron laser FLASH at DESY, aims at significantly improving the FEL photon beam properties for users. Within the project, both existing FEL lines at FLASH will be equipped with fully tunable undulators capable of delivering photon pulses with variable polarization. One of the two FEL lines will be externally seeded at the full repetition rate that FLASH can provide in burst mode. The other line will exploit novel lasing concepts based on different undulator configurations. Together with an increase in electron beam energy to 1.35 GeV this will extend the wavelength range to the oxygen K-edge in the fundamental harmonic, in order to cover the important elemental resonances for energy research and the entire water window for biological questions. The planned machine upgrade and the resulting new beam properties require a substantial upgrade of the existing photon diagnostics and beam transport.

Speaker: Dr Kai Tiedtke (DESY)

12 Present status, achievements, and future developments at FERMI: photon transport and diagnostics from current operations to prospective low-wavelengths upgrade

FERMI, the Italian free electron laser operating at Elettra (Trieste), has been in user operations for eight years.
During this timespan the number of beamlines and endstations has reached the project target (6), and continuous development of the machine has taken place, too. Currently, two calls for proposals are issued per year, offering users the possibility to perform experiments on both FEL undulator lines of FERMI, covering the 100-4nm range. Of course, also the photon transport and diagnostics system (PADReS) has evolved in this years, having to face new requests in terms of several different aspects (e.g., intensity monitors, energy spectrometers, and new optics for extended/lower wavelength operation; multiplexing solutions for the efficient use of the radiation; advanced diagnostics for ultimate focusing; etc.).
A further upgrade of the entire facility is now being discussed and developed, which foresees an increase of the machine electron energy in order to decrease the achievable wavelength to values below 2nm in the fundamental laser harmonic. The goal is to reach such lower wavelengths maintaining the unique characteristics and capabilities of FERMI, which is a seeded FEL. In particular, intensity and wavelength stability, spectral purity, coherence, polarization, and TM-00 photon emission mode should be guaranteed as well as the possibility to produce two-color double pulses and other exotic emission schemes.
Achieving that, a new class of experiments will be possible together with the chance to extend the current experimental capabilities in so-far non-achievable wavelength regions.
Obviously, this ambitious target calls for parallel evolution of the photon transport and diagnostics system in order to be able to efficiently cover the need for online characterizing each single FEL pulse for wavelength extending to 2nm and below. The present setup will need to be upgraded in several ways, from the energy spectrometers to the intensity monitors, from beamline design to the optical mirror coatings, and so on.
The current situation, with the solutions adopted up to now, is presented, and the possible future upgrades and developments are discussed.

Speaker: Marco Zangrando (Elettra Sincrotrone Trieste and IOM-CNR)

13 Status of the EuXFEL photon diagnostics

The European XFEL facility has entered user operation in September 2017. Since that time the baseline photon diagnostics are in routine operation, delivering online monitoring of pulse energies, beam position, and transverse beam profile. More advanced diagnostics such as monitors for spectral and temporal properties were also commissioned in the meantime. This presentation reviews the operational diagnostics devices in the three SASE undulator beamlines, and provides details about new diagnostic capabilities such as MHz-rate spectral and temporal monitoring at hard and soft x-rays, measurements for undulator system optimization, studies of high-repetition rate effects with pulse-resolved intensity monitors, novel diamond diagnostic detectors, absolute pulse energy measurements with calorimeters and gas-based monitors, and application of photon diagnostics for special machine operation modes and studies such as hard x-ray self seeding, two-color lasing and harmonic lasing.

Speaker: Dr Jan Grünert (European XFEL)

14 Status of LCLS and Future Developments

The Linac Coherent Light Source (LCLS) began user operations in 2009 and has performed over 500 full-scale user experiments with ever-expanded operating modes/capabilities. The facility recently emerged from a long down period where new electron transport, undulators, x-ray transport systems and x-ray instruments were installed. First commissioning results of these new systems will be presented as well the status of three ongoing upgrade projects: LCLS-II, LCLS-II-HE and MEC-U.

Speaker: David Fritz (SLAC National Accelerator Lab)

15 FEL Prize presentation

Tuesday 27 October

New Development in Photon Diagnostics and Optics 1

Session 5

Convener: Benedikt Rösner (1 Paul Scherrer Institut, CH-5232 Villigen, Switzerland)

16 Spatiotemporal diagnostics of XFEL pulses via intensity correlation techniques

Characterizing spatiotemporal properties of XFEL pulses is of great importance not only for analyzing experiments, but for giving effective feedbacks to machine operations. As simple and cost effective ways to diagnose XFEL pulses, we have developed X-ray intensity correlation techniques, such as intensity correlation measurements of fluorescence and spontaneous undulator radiation for evaluating XFEL durations [1,2] and spatial profiles of tightly focused XFEL beam [3].
In this presentation, I will talk about the concepts of these techniques and their applications to XFEL pulses from SACLA, as well as the future perspectives.

[1] I. Inoue et al., Phys. Rev. Accel. Beams. 21, 080704 (2018).
[2] I. Inoue et al., J. Synchrotron Rad. 26, 2050 (2019).
[3] N. Nakamura et al., submitted.

Speaker: Dr Ichiro Inoue (RIKEN SPring-8 Center)

17 Newly designed radiometer for synchrotron radiation and free-electron laser in high power range

We designed a new radiometer to measure the absolute power of synchrotron radiation and free-electron laser in the wavelength range from the extreme-ultraviolet (EUV) to x-rays. The target measurable power range is from 1 mW to 1 W, which exceeds that of our former radiometer (a compact radiometer) of 0.01 mW to 150 mW. The new radiometer is directly mounted on a vacuum flange cooled with a fan, and can operate at room temperature or above (around 310 K). The measurement principle of the radiometer is based on a temperature measurement of an absorber, which is a component of the radiometer. The absorber is a cavity type and consists of a tungsten plate and a copper cylinder. The absorptance is higher than 99.5% for photon beams in the wavelength range from EUV to x-ray (20 eV to 60 keV); therefore, a temperature change in the absorber relates to the power of an incident photon beam. The absolute power can be evaluated from a electrical heating power applied to a heater on the absorber, namely a dynamic electrical substitution technique. We have completed the construction of the radiometer and plan to finish checking the performance of the radiometer using an electrical heating by the end of this year.

Speaker: Takahiro Tanaka (National Institute of Advanced Industrial Science and Technology (AIST) )

18 Multi-Focus Off-Axis Zone Plates for Experiments at X-Ray Free Electron Lasers

1. General information
X-ray free electron lasers (XFELs) pave the way towards new and highly exciting experiments in the fields of X-ray imaging and spectroscopy [1,2]. They allow accessing ultrafast and non-linear processes on the nanoscale. However, the stochastic nature of the lasing process causes strong shot-to-shot fluctuations in intensity and spectral composition, which represents a challenge for many experiments in terms of signal normalization. Thus, accurate detection schemes with high correlation between the signal and a reference beam are necessary for normalization purposes to ensure good data quality in experiments that rely on the use of multiple pulses from an XFEL, such as spectroscopy and pump-probe experiments.

2. Methodology
For addressing this challenge, we have designed a new type of X-ray optic that combines the beam splitting functionality of a transmission grating with an off-axis zone plate [3]. This design not only overcomes the disadvantages of two individual optical elements, but introduces the beam splitter as a perfect phase grating, which can be implemented in a defined way to adjust the beam splitting efficiencies for certain diffraction orders precisely.
The grating can be added in two ways either by inverting the zone structure with a certain duty cycle or by shifting the zone structure by a certain factor.

3. Results
We characterized such multi-focus zone plate optics at the SIM beamline at the SLS and they enabled us to conduct the first user experiment at the Spectroscopy and Coherent Scattering beamline at the European XFEL. Here, the overlaid inversion grating generated two intense spots in the focal plane, where one interacted with the sample whereas the other one was used as a reference. In this way, we were able to measure an X-ray absorption spectrum from Ni and NiO thin films at different fluence levels with precise normalization on a shot-to-shot basis. The normalized spectrum showed a clear signal over noise improvement. These results demonstrate the capabilities of such compound optics to enable advanced experiments at XFELs.

Acknowledgements
We would like to thank everyone involved in the open community proposals No. 2161 and No. 2170 at the European XFEL. This work received funding from the EU-H2020 under grant agreement No. 701647 and No. 654360 NFFA-Europe.

References
[1] P. Emma, et al. Nat. Photon. 4(9), 641–647 (2010).
[2] E. Allaria, et al. Nat. Photon. 6, 699–704 (2012).
[3] F. Döring, et al. Optica 7 (8), 1007-1014 (2020).

Speaker: Florian Döring (1 Paul Scherrer Institut, CH-5232 Villigen, Switzerland)

10:10 Coffee break

Spectral Measurements

Session 6

Convener: Joan Vila-Comamala (Institute for Biomedical Engineering)

19 Resolving spectro-temporal dynamics with all-XUV-optical FEL transient absorption spectroscopy

Extreme ultraviolet (XUV) and x-ray free-electron lasers (FELs), delivering intense ultrashort pulses in the femtosecond regime and even below, create new opportunities for site-specific light-matter interaction, and eventually to control the intrinsic quantum dynamics of a system with atomic scale resolution. Using visible lasers, the technique of ultrafast transient absorption spectroscopy and related multidimensional spectroscopies, which elegantly combine both high spectral and temporal resolution through multi-pulse sequences, are now routinely used both for understanding and controlling photochemical reactions, however typically coupling only to the valence shell of the binding electrons and the intramolecular vibrational degrees of freedom, which is directly linked to the low photon energies being used (from infrared to visible and ultraviolet).

In this talk I will give an overview of our recent efforts to perform all-XUV-optical transient absorption spectroscopy of gas-phase atoms and molecules [1-3]. Hereby we have proven sensitivity to element-specific spectral signatures identifying resonant electronic transitions in the XUV and how they are modified (both Stark shifts and line-shape asymmetry changes) due to the presence of the strong XUV free-electron laser electric field. Furthermore, transient spectro-temporal signatures could be extracted from the data, with a timescale below 3 femtoseconds, and understood through nonlinear interaction of pump and probe beams in the target medium, revealing direct sensitivity to the coherence properties of the FEL pulses. This can be regarded as a precursor to coherent multidimensional spectroscopy in the XUV spectral range.

An important ingredient for such measurements is a profound knowledge of the spectro-temporal properties of the employed FEL pulses. In our experiments we utilize the quasi-instantaneous non-resonant ionization dynamics of neon atoms in response to an FEL pump pulse as an ultrafast optical switch to modulate the XUV absorption properties of our gas target with timing precision below one femtosecond. Thereby, upon spectrally resolving the FEL probe transmission spectrum with a grating-based spectrometer, we are directly sensitive to the average chirp of the FEL probe pulses, revealing a pronounced positive chirp of about 30 fs² for the particular FEL machine settings [4]. Most importantly this is a direct spectro-temporal measurement, tracking the transmitted FEL photon energy spectrum as a function of pump-probe delay. No additional external laser pulses, which are typically challenging to synchronize at the precision level of a femtosecond, are needed for this all-XUV-optical scheme. Besides the importance of in-situ measuring such spectro-temporal pulse properties in transient-absorption and multidimensional-spectroscopy experiments, this new characterization method may also develop into a useful tool in general, extending the diagnostics toolbox of FEL pulses.

I gratefully acknowledge the fruitful collaborations [1-4] during our previous measurement campaigns at the Free-Electron-Laser in Hamburg (FLASH) at DESY.

[1] T. Ding et al., PRL 123, 103001 (2019).
[2] C. Ott et al., PRL 123, 163201 (2019).
[3] M. Rebholz et al., “All-XUV pump-probe transient absorption spectroscopy of the structural molecular dynamics of diiodomethane”, submitted manuscript (2020).
[4] T. Ding et al., “Measuring the frequency chirp of extreme-ultraviolet free-electron laser pulses by transient absorption spectroscopy”, submitted manuscript (2020).

Speaker: Christian Ott

20 High-Resolution Pulse-to-Pulse Spectral Monitoring for SwissFel

The spectrum of SASE XFEL sources exhibits strong variations pulse-to-pulse. As a consequence, XFEL-driven experiments, XFEL optimization or “new-modes” development rely on X-ray spectrometers functioning on a shot-to-shot basis.
We will discuss the potentiality of the setup used at SwissFEL, as well as results obtained during its commissioning.
Several arrangements of transmission gratings, e.g. having a pitch of 100 nm, and bent crystals (e.g. Si(220)) with various bending radii (e.g. between 75-200 mm) allowed measuring the SASE spectrum with high resolution, e.g. about 0.4 eV (FWHM) at 7.1 keV was achieved without gratings. Possible “dependencies” with the FEL profile or bent crystals alignment, as well as “dispersive” XAS measurements will be presented.

Speaker: Yunieski Arbelo Pena

21 Flat field soft X-ray spectrometer based on reflection zone plates

The reflection zone plate (RZP), as a wavelength-dispersive optical element, has been successfully used in several designs of femtosecond (fs) spectrometers and monochromators for soft X-rays [1]. While demonstrating its evident advantages in comparison with conventional VLS gratings, like a high resolving power at the design energy and an excellent signal-to-noise ratio, RZPs on planar substrates suffer from a narrow energy range in parallel spectral registration, limiting the applications of this type of optic. Recent developments in theory, technology, and metrology of RZPs make it possible to fabricate RZPs on spherical substrates with a small radius of curvature down to 2 m.

In this work, we report on first results obtained with newly designed high-resolution soft X-ray fluorescence spectrometer, based on RZPs, which were fabricated on a spherical substrate. High resolution flat field spectra within about ± 50 % around the design energies were measured in the interval from 150 eV to 750 eV, using only two RZPs: The 1st RZP with its design energy of 277 eV covers the band (150 – 550) eV and the 2nd RZP with a design energy of 460 eV covers the band (350 – 750) eV, where the upper bound to this energy range is defined by the Ni coating of the RZPs. The absolute diffraction efficiency reaches 25 % and 20 % at the design energies 277 eV and 460 eV, respectively. The energy resolving power E/E exceeds 2000 in the entire energy range. To compensate the slope error of the substrate, an algorithm for diffractive wavefront correction [2] was used in the calculation of the groove structure.

The development of theory and technology for RZPs on figured substrates opens new possibilities for a considerable improvement of the instrumental performance in soft X-ray spectroscopy, pushing forward the frontiers in time, resolving power and efficiency of dispersive optical elements. Successful results of tests enabled NOB GmbH to start customer-oriented production of RZPs on spherical substrates for spectroscopic applications.

This work was supported by the project REFLEX, Berlin Program ProFiT co-financed by the European Regional Development Fund (ERDF) and project NeuGaR, partly financed by the Federal Ministry for Economic Affairs and Energy.

[1] A. Erko, C. Braig, and H. Löchel, “Spectrometers and monochromators for femtosecond soft x-ray sources,” Proc. SPIE 11108, 111080J (2019).
[2] J. Probst, C. Braig, E. Langlotz, I. Rahneberg, M. Kühnel, T. Zeschke, F. Siewert, T. Krist, and A. Erko, “Conception of diffractive wavefront correction for XUV and soft x-ray spectroscopy,” Appl. Opt. 59, 2580 – 2590 (2020).

Speaker: Alexei Erko (Institute for Applied Photonics)

11:40 Lunch break

Coherence, Profile and Position Measurements

Session 7

Convener: Rasmus Ischebeck (Paul Scherrer Institut)

22 FEL beam diagnostics and source metrology

After the first decade of operability, the strong quest for results around the Free Electron Laser (FEL) facilities has been positively filled, giving the opportunity to make one step back from the sample-level and to focus on subtle, still open topics concerning the FEL source metrology and coherence characterization. Because of the complexity of the emission process, important parameters such as the effective source position and dimension may be a-priori not known and depend on the required machine optimization, or can be adjusted to meet the experiment needs. The idea of using wavefront sensing techniques, usually employed for optics tuning, is captivating because of their shot-to-shot operability and accuracy, which make them robust and suitable as a feedback for FEL machine-tuning operations. Here we review the results of source metrology measurements at the FERMI seeded-FEL performed at distinct machine configurations, by means of Hartmann wavefront sensing. The effects of the transport optics, which may introduce curvature alterations, are discussed as well, with particular attention to the case in which shot-to-shot properties of the source are investigated (e.g. source position fluctuations) and local curvature effects of the optics come into play.

Speaker: Dr Michele Manfredda (Elettra Sincrotrone Trieste S.C.p.A.)

23 Autocorrelation measurement of the FEL coherence and pulse length using a magnetic delaying chicane in the undulator beamline

A delay of the electron beam with respect to the FEL radiation field can be used for auto-correlation measurements since the phase is imprinted both in the radiation field as well as the microbunching. As long as the delay is within the coherence length of the SASE spikes there is interference, resulting in a modulation of the output power, similar to a phase shifter. For longer delays the bunching is overlapped with a different spike, which shows in average no dependence on the delay due to the arbitrary phase relation between the bunching and the radiation of the other spike. Even further delays will separate the radiation field from the electron bunch reducing further the output power of the FEL signal, since the FEL starts from the spontaneous radiation again. Therefore a scan of the delay can give information on both the coherence length and the FEL pulse length with a resolution on the scale of the radiation wavelength. We present simulation examples for SwissFEL that show the validity and feasibility of the method.

Speaker: Sven Reiche (Paul Scherrer Institut)

24 Towards Pulse-by-pulse XFEL Beam Measurement

Characterization of XFEL beams can be challenging, requiring the use of several different types of detectors to measure beam properties. Sensors fabricated using electronic grade single crystal diamond have been shown to have rapid response and enable measurement of signals over a wide dynamic range for synchrotron beams. To study the utility for measurement of XFEL beams, flux linearity and temporal response measurements were performed at the SwissFEL Bernina station with two such devices through a collaboration formed during PhotonDiag2018. Comparisons were made with synchrotron beam measurements (flux linearity, spatial uniformity and temporal response) performed at NSLS-II using the same detectors prior to and subsequent to FEL beam testing. The results indicate a relationship between flux linearity and temporal response, which may provide guidance for improved detector design. Development of these monitors holds promise for enabling measurement of the intensity (and potentially also position) of single XFEL pulses over a significant dynamic range.

Speaker: Jen Bohon (Los Alamos National Laboratory)

14:40 Coffee break

New Detectors for Photon Diagnostics

Session 8

Convener: Aldo Mozzanica (Paul Scherrer Institut)

25 New Detectors for Photon Diagnostics

Diagnosing ionizing radiation has been subject of research as long as applications in this field exist. Many techniques to reveal temporal and spatial information about a beam of ionizing photons are intrusive. To minimize this intrusiveness, plasma could be a suitable intermediate state for potential diagnostics to be developed.
The in this contribution elaborated novel diagnostic is based on using a very small fraction of the radiation to partially photo-ionize a low pressure background gas. As a result, this background gas is turned into a plasma which can be – based on its permittivity - further characterized by traditional plasma diagnostics.
For this characterization, the beam of photons is sent through a cylindrical metal pillbox cavity filled with a sensing gas at low pressure. The free electrons in the plasma induced in this cavity are probed by having them interacting with a standing electromagnetic wave in the microwave frequency range (a few GHz). Tracking temporally resolved the resonance frequency of this standing wave reveals information about the radiation beam in terms of power, alignment and stability.
This contribution elaborates on the physical principles of the method and highlights the newest developments towards in-line beam monitoring.

Speaker: Dr Job Beckers (Eindhoven University of Technology)

26 High-sensitivity Femtosecond X-ray Optical Cross-Correlator for Next Generation Free-Electron Lasers

The large-scale nature of free-electron lasers (FELs) often precludes control over various noise sources affecting the arrival time of the X-ray pulses. To achieve high temporal resolution (<100 fs) in laser pump/X-ray probe experiments, an arrival time monitor (ATM), such as an X-ray/optical cross-correlator, is implemented. Typically, an X-ray-induced change of the optical properties of a target is used to derive the arrival time of the X-rays relative to the optical pump laser. These ATMs often require soft X-ray fluence levels of >1 mJ/cm$^2$. The next generation of high-repetition FELs, such as the European XFEL and the Linac Coherent Light Source II (LCLS-II), take advantage of superconducting accelerators to enable repetition rates of up to 1 MHz. As a result of the increase in repetition rate, the energy per X-ray pulse will decrease. The X-ray fluences required by current ATMs cannot be achieved with high repetition rate FELs or when using few and sub-fs X-ray pulses. The novel ATM scheme described here solves this problem by achieving unprecedented sensitivity to soft X-rays.

Current ATM designs at FELs use a portion of the pump laser beam at a wavelength of 800 nm or white light in the visible range to probe an X-ray induced material change. However, state-of-the-art optical timing distribution systems operate at a wavelength of 1550 nm and are an enabling technology for sub-10-fs timing jitter between X-rays and pump lasers. Therefore, the ATM presented here cross-correlates X-rays and 1550 nm optical pulses to pave the way to directly derive the X-ray timing jitter measured with respect to an optical master oscillator.

The X-ray/optical cross-correlator is based on a time-to-space mapping geometry and takes advantage of 1) a specially designed multi-layer target, and 2) an interferometric detection scheme. The sample used here is made of a thin film of germanium, sputtered on a 2 μm thick diamond layer grown by Chemical Vapor Deposition (CVD) onto a silicon dioxide layer on top of a silicon substrate. This multi-layer structure shows optical interference effects (etalon) in reflectivity, which are designed to exhibit a node in the reflection spectrum around 1550 nm. The X-ray induced changes to the sample shift the center wavelength of this node leading to dramatic amplitude and phase changes of the reflected 1550 nm light, allowing to detect extremely weak X-rays when combined with an interferometric detection scheme.

The X-ray/optical cross-correlator described here has been tested during an experiment on the LCLS soft X-ray beamline at an X-ray photon energy of 530 eV and a pulse repetition rate of 120 Hz. We achieved a ~100-fold increase in the relative signal change as compared to previously demonstrated techniques. The resolution of the timing measurement is estimated to 2.8 fs (rms), making this cross-correlator well suited for sub-10-fs timing jitter operations at LCLS and other FELs.

Speaker: Stefan Droste (SLAC National Accelerator Laboratory)

27 The new ePix10k megapixel hard x-ray area detector at the LCLS

The ePix10k2M is a new large area detector specifically developed for X-ray Free Electron Laser applications. The hybrid pixel detector was developed at SLAC to provide a hard X-ray area detector with a high dynamic range, running at the 120 Hz repetition rate of the Linac Coherent Light Source (LCLS). The detector has a dynamic range from single photon counting up to 10.000 photons/pixel/pulse at 8keV. The high dynamic range is achieved with 3 distinct gain settings (Low, Medium, High) as well as two autoranging modes (high-to-low and medium-to-low).

Here we evaluate the detector performance in comparison with the previously deployed CSPAD.
The external dimensions of the two detectors are similar, making the upgrade from CSPAD to ePix10k straightforward for most setups at LCLS, with the sPix10k improving on experimental performance. The main detector during an experiment, such as the large area ePix10k, is used for primary signal detection. However, this detector is also often used for normalization and for diagnostics of the experimental setup, it is therefore crucial that these multi-purpose detectors are well understood and calibrated to facilitate the best scientific output of the limited XFEL beamtime.

Here we present the first measurements on this new ePix10k detector and evaluate the performance under typical XFEL conditions during an LCLS x-ray diffuse scattering experiment measuring the 9.5 keV x-ray photons scattered from a thin liquid jet.
The SLAC developed ePix cameras all utilize a similar platform and are designed to provide an upgrade path for future high repetition rate XFELs such as LCLS-II and LCLS-II-HE.

Speaker: Tim van Driel (SLAC National Accelerator Laboratory)

16:10 Coffee break

New Development in Photon Diagnostics and Optics 2

Session 9

Convener: Benedikt Rösner (1 Paul Scherrer Institut, CH-5232 Villigen, Switzerland)

28 Diffractive wavefront correction for nearly diffraction-limited soft X-ray spectroscopy

We report on the development of an accurate and flexible concept to compensate for two-dimensional slope errors of mirrors or reflective gratings as used in wavelength-dispersive soft X-ray spectroscopy. Our approach enables an ultra-high spectral resolving power with low-cost optical components [1].

Modern soft X-ray spectrometers often use a toroidal mirror in combination with a variable line space grating. However, the curved mirror / grating substrate typically suffers from figure errors, leading to an aberrated wavefront. In close analogy to computer-generated holograms, that phase distortions are converted into a customized groove structure of the grating, the so-called “diffractive wavefront corrector”. Either the probed height profile (ex situ) of the substrate or deterministic phase retrieval from the intensity distribution along the propagating X-ray beam (in situ) may be used to derive the phase [2].

We consider popular instrumental configurations, like the “Hettrick-Underwood” setup or the compact “all-in-one” device, where the grating is inscribed in the curved mirror, and apply an appropriate variety of experimental procedures and the associated computational framework to evaluate the optical path difference (OPD). This function yields the grating vector field of the DWC as its gradient.

Characteristic, slightly aperiodic and – in general – asymmetric groove structures are obtained, which differ from the regular line density distribution of, e.g., analogous reflection zone plates on a scale of ~ µm. Lamellar gratings of that kind may be fabricated conveniently by direct laser writing, especially in their three-dimensional (3D) implementation on a curved substrate.

Ray tracing simulations of our DWC spectrometers predict an energy resolution which is pushed close to the diffraction limit around the design energy, while the photon flux of the initial, uncorrected instrument can be maintained.

This work is funded by the Bundesministerium für Wirtschaft und Energie within the project “NeuGaR” (ZF4302303SY8).

[1] C. Braig, J. Probst, E. Langlotz, I. Rahneberg, M. Kühnel, A. Erko, T. Krist, and C. Seifert, “Diffraction compensation of slope errors on strongly curved grating substrates,” Proc. SPIE 11109, 111090U (2019).

[2] J. Probst, C. Braig, E. Langlotz, I. Rahneberg, M. Kühnel, T. Zeschke, F. Siewert, T. Krist, and A. Erko, “Conception of diffractive wavefront correction for XUV and soft X-ray spectroscopy,” Appl. Opt. 59, 2580 – 2590 (2020).

Speaker: Dr Christoph Braig (Institut für angewandte Photonik e.V.)

PhotonDiag-CB.pdf

29 Design of an amplitude-splitting hard x-ray delay line with sub-nanoradian stability

Measurements of the dynamics of a material on the fs-to-ns temporal scale and atomic length scale provide essential information to guide our understanding of disordered materials. Split-pulse x-ray photon correlation spectroscopy (spXPCS) was recognized early in the development of the Linac Coherent Light Source as a technique that could possibly provide critical measurements of these dynamics. To conduct spXPCS measurements at fs-to-ns scale, two identical x-ray pulses are delivered to the same spot on the sample, along the same direction, with a tunable time separation over the time scale of interest. This time regime is currently beyond the capability of the two-pulse accelerator based machine operation modes at free electron laser (FEL) facilities, and was expected to be reached with the x-ray analog of a visible optical split-delay instrument.

Such x-ray optical systems, usually referred to as x-ray split-delay optics, have seen many iterations since the first x-ray FEL facility became operational. Thin Bragg crystals were initially favored and explored as the beam splitter. However, residual strain and beam-induced thermal-mechanical issues in the thin crystals have prevented their effective use for XPCS experiments so far. Later, wave-front splitting designs adopted edge-polished silicon crystals for dividing the wave-front into two halves, and demonstrated stable and routine delivery of two microfocused x-ray spots to the sample with promising stability for pump probe experiments. However, a non-negligible crossing angle between the two beams, as a result of the wave-front splitting mechanism, poses a serious challenge for realizing visibility spectroscopy measurements. Wave-front splitting also makes the output beam properties sensitive to beam pointing jitters, and this degrades the effective spatial overlap.

Sun et al. proposed an all channel-cut split-delay design that demonstrated superb pointing stability which is a critical requirement for two-pulse x-ray experiments. This design, however, suffered from the pulse front tilt and photon energy dispersion due to the use of asymmetric reflections.

Here we present a modified version of the all channel-cut split-delay system design. It realizes amplitude-splitting and the earlier technical difficulties are expected to be overcome. In this new design, a pair of $\pi$-phase-shift transmission gratings are introduced as the amplitude splitter and recombiner. An additional pair of asymmetric channel-cuts are introduced in the delayed branch to eliminate the dispersion and pulse front tilt. A performance analysis of wave propagation through a prototype model using dynamical diffraction theory revealed that sub-nanoradian relative pointing stability during a delay scan can be achieved.

We believe that this new design, once realized, will represent a significant step towards realizing split-pulse x-ray photon correlation spectroscopy at FEL facilities investigating ultrafast equilibrium dynamics of disordered matter.

Speaker: Haoyuan Li (Stanford Linear Accelerator Center)

Wednesday 28 October

Photon Diagnostics at Synchrotrons

Session 10

Convener: Volker Schlott (Paul Scherrer Institut)

30 The use of online photon beam position measurements to improve synchrotron performance

For users to take advantage of the low emittance and high coherence photon beams produced at synchrotron light sources, sub-micron source-point stability is desired over timescales from milliseconds to hours. However, it is insufficient to only monitor the source electron beam; the X-ray beam itself must be monitored and, where required, included in feedback loops to meet the stringent requirements of beamlines and users. Variations in the photon flux intensity or profile can severely reduce the quality of the X-ray data collected, thus it is important to make accurate, online, and preferably non-destructive measurements of the photon beam itself. The use of photon beam diagnostics at synchrotrons is well-established, and is an essential tool to commission, optimise, and improve the light delivered to users. This paper reviews various photon diagnostics techniques in use at different synchrotrons, and discusses how their measurements are applied to improve both machine and beamline performance.

Speaker: Chris Bloomer (Diamond Light Source Ltd.)

31 Modelling the effects of optical vibrations on photon beam parameters using ray-tracing software

A method to simulate beam parameters observed at a beamline sample point in the presence of motion of optical components has been developed at Diamond Light Source. Stationary ray-tracing simulations are used to model the impact on the beam stability caused by dynamic motion of optical elements. Ray-tracing simulations using SHADOW3 in OASYS, completed over multiple iterations and stitched together, permit the modelling of a pseudo-dynamic beamline. As beamline detectors operating at higher frequencies become more common, the beam stability becomes crucial. Synchrotron ring upgrades to low emittance lattices require increased stability of beamlines in order to conserve beam brightness. By simulating the change in beam size and position an estimate of the impact certain motions have on stability is possible. The results presented in this paper focus on modelling the physical vibration of optical elements. However, the basic principle can be applied to any parameter which dynamically changes. Multiple situations can be analysed in succession without manual inputs. In this paper we describe the simulation code and present the results obtained. This method can be applied during beamline design and operation for the identification of optical elements that may introduce large errors in the beam properties at sample point.

Speaker: Ms Claire Houghton (Diamond Light Source)

11:40 Lunch Break

Photon Diagnostics Requirements for new Experimental Methods

Session 11

Convener: Gregor Knopp (Paul Scherrer Institut)

32 Probing THz driven solids by second harmonic generation

THz sources at FLASH utilize spent electron beam from an soft X-ray FEL to generate very intense (up to 150µJ), tunable frequency (1-300THz) and ultrafast narrow-band (~10%) THz pulses, which are naturally synchronized to soft X-ray pulses. This unique combination allows for wide range of element specific pump-probe experiments in physics, material science and biology.
THz pulses are completely temporally characterized by a jitter corrected Electro-Optic-Sampling(EOS) THz pulse characterization over a broad spectral range (0.1 - 8 THz).
In this work we report on the installation of new probing technique that enables study changes of material symmetry, by second harmonic generation of the probing fs-laser. Instrument allows detection of 2nd harmonic both in transmission and reflection, allowing probing of bulk and surface effects and matching THz excitation depth, e.g. at and off of absorption resonances. Instrument employs arrival time jitter correction and has temporal resolution better than 10fs.

Speaker: Dr Nikola Stojanovic (DLR)

33 Characterizing isolated attosecond pulses with angular streaking

The development of isolated attosecond pulses (IAP) with the free-electron laser at LCLS called for a high-resolution, single-shot diagnostic method. In this talk I will discuss the use of angular streaking to characterize IAPs. An IAP with sufficiently high photon energy will ionize an atomic system to produce photoelectrons. In the presence of the IR field, the energy of the ionized photoelectron will be modulated, or streaked, depending on the phase of the IR laser field at the time of ionization. Through the streaking interaction, information about the temporal profile of the IAP will be encoded in the energy (and momentum) distribution of the emitted photoelectrons. Using a circularly polarized streaking laser, the temporal profile of the electron wavepacket is encoded in the angular distribution of streaked photoelectrons.

Speaker: Siqi Li

34 Multi Resolution Cookiebox Optimized for Future Free Electron Laser Experiments - MRCOFFEE

A major upgrade of the LCLS facility, the LCLS-II project, is now underway. LCLS-II is being developed as a high-repetition rate X-ray laser with two simultaneously operating, independently tunable FELs. It features a 4 GeV continuous wave superconducting linac that is capable of producing uniformly spaced (or programmable) ultrafast X-ray laser pulses at a repetition rate up to 1 MHz spanning the energy range from 0.25 to 5 keV. Furthermore, the XLEAP sub-femtosecond soft X-ray pulse generation program is scalable to LCLS-II repetition rate [1].

We have designed, based on the Viefhaus et al. Cookiebook [2], an angle-resolving array of 16 electron time-of-flight spectrometers that allow wide and adjustable energy acceptance windows. By interleaving detector retardations, we enable simultaneous angle-resolved photo-electron and Auger electron spectroscopy as is required by cutting edge molecular frame spectroscopies and diffraction. The spectrometer array will be available for spectral-polarimetry measurements as well as polarization sensitive attosecond resolving temporal characterization of LCLS-II pulses. This multi-polarization and multi-color spectral diagnostic/experiment endstation will have an energy resolution better than the expected seeding spectrum and the SASE spectral features.

We would like to present the next generation of high resolution electron spectroscopy endstation and FEL generated X-ray pulse diagnostic – MRCOFFEE. This talk will present the important science opportunities, new diagnostic capabilities of this newly designed TMO instrument/diagnostic.

[1] Duris, J, et. al. (2019). Tunable Isolated Attosecond X-ray Pulses with Gigawatt Peak Power from a Free-Electron Laser. Nature Photonics, 14(January). https://doi.org/10.1038/s41566-019-0549-5
[2] Hartmann, N., et al. (2018). Attosecond time-energy structure of X-ray free-electron laser pulses. Nature Photonics, 12(4), 215–220. https://doi.org/10.1038/s41566-018-0107-6

Speaker: Peter Walter (SLAC)

15:40 Coffee break

Scientific Computing, Machine Learning and large Data Management

Session 12

Convener: Derek Heinrich Feichtinger (Paul Scherrer Institut)

35 Adaptive Feedback and Machine Learning for Tuning Complex and Time Varying Systems

Machine learning (ML) approaches such as neural networks (NN) and Gaussian processes (GP) are powerful tools that can learn input-output models of complex systems directly from data. Because of their strengths ML methods have been growing in popularity in all areas of science and engineering. One limitation of such methods is that if the system which generated the training data changes with time, the accuracy of the models that were learned based on that data begins to drift. Model independent feedback methods that are robustness to noise and can adapt to time varying systems exist, but their limitations include a chance of getting stuck in a local minimum if searching over a very large parameter space. By combining the ability of ML data-based methods to learn the global features of large parameter spaces with model-independent adaptive feedback, we can build an overall system that is both robust and globally optimal. At Los Alamos National Laboratory (LANL) we have begun studies of adaptive ML methods for time varying systems. Our initial results have focused on particle accelerators and their beams which are large complex systems in which both the components (hundreds of coupled magnets and RF cavities) and the beams (complex objects living in 6D (x,y,z,x',y',E) phase space) vary unpredictably with time. The time variation and complexity of accelerators results in days - weeks of hands on tuning after outages or when switching between significantly different beam setups. In this talk I will start with an overview of adaptive model-independent methods that we have developed and implemented at particle accelerator facilities around the world for automated beam tuning and accelerator optimization including the LANSCE linear accelerator at LANL, the EuXFEL at DESY, the AWAKE experiment at CERN, the FACET-II plasma wakefield accelerator facility at SLAC, and at the LCLS FEL at SLAC. I will then present some of the first adaptive ML results that we demonstrated at the LCLS FEL for adaptive tuning of the longitudinal phase space (time vs energy) of the LCLS electron beam and will present an overview of adaptive ML approaches. Finally I will discuss how these same principles can be applied to the experimental beam lines at accelerators which are also complex systems that require frequent and lengthy tuning.

Speaker: Alexander Scheinker

36 Application of automatic differentiation based ptychography for single-shot SASE FEL beams characterization

Wavefront sensing and characterization of the spatial and coherent properties of the Free-Electron-Lasers (FELs) radiation is vital for experiment planning, beamline optics alignment and the photon diagnostics. At the Free-electron LASer at Hamburg (FLASH), the Hartmann wavefront sensing is a typical technique used for the single-shot beam characterization. However, it is working in the assumption of the fully spatially coherent radiation and has resolution limited by mechanical components. Ptychography is another promising technique, which can provide us with high resolution complex wavefields of the FEL radiation.

Ptychography is a scanning coherent diffraction imaging technique, which utilizes mutual overlap between neighboring scan positions for enforcing of the additional constraints during the reconstruction. These constraints make it robust and free of the reconstruction artifacts possible in the other CDI techniques. Ptychography allows to reconstruct complex sample function together with the complex wavefield of the illumination with diffraction limited resolution, and thus, it is frequently used for the beam characterization purposes.

Typically, ptychographical reconstruction assumes constant illumination, and can treat partially coherent illumination by usage of multimodal beam assumption. However, in this case, modal composition also should be constant for each shot, and spatial shot-to-shot fluctuations should be minimized. This conditions cannot be easily fulfilled at Self-Amplified Spontaneous Emission (SASE) FELs with their fluctuating and partially spatially coherent beams.

In this work, we suggest to use a novel automatic differentiation (AD) based ptychographical engine which is able to perform reconstructions under these conditions. AD allows to split ptychographical algorithm into independent forward model, which can be flexibly changed, and optimization routine. This flexibility allows to adapt the forward model to the specifics of ptychographical experiments at SASE FEL and to perform the reconstructions in the assumption of multi-modal spatially fluctuating illumination, with the unique modal composition for each individual shot.

We present an AD-based ptychographical engine applicable for ptychography at FELs. With its help we were able to reconstruct complex wavefields of the FLASH2 radiation during the ptychographical measurements. Applications of AD-based ptychography to beam characterization and possibility of reconstruction of individual shots characteristics will be discussed.

Speaker: Konstantin Kharitonov (Deutsches Elektronen-Synchrotron DESY, Hamburg, 22607, Germany)

37 Developing Real-time Services for Large Volume Experiment-Data Analysis utilizing Supercomputing and Cloud technologies at CSCS (SELVEDAS)

Introduction
The ongoing developments in accelerators, detectors and experiment automation is leading to a rapid growth of data generated during experiments. A viable solution is utilizing suitable infrastructures that allow additional remote high performance capacity for processing and analysis of data from the experimental facilities with larger data volumes and higher processing needs. The SELVEDAS project proposes a hybrid cloud infrastructure, offering scalable and extensible services for data management and analysis to Swiss academic users by leveraging high performance computing (HPC), storage, networking as well as cloud technologies and orchestration. The on-demand services perform as a highly efficient remote data processing system providing fast feedback and analysis with the long time storage and archival of petabytes of data.

Approach
Hybrid cloud
The SELVEDAS project develops a hybrid cloud extension to the PSI infrastructure by giving access to the supercomputing resources at CSCS for experiment data analysis. The solutions rely on the easy transferability of workloads between systems comprising of HPC and cloud technologies. The architecture is data-driven (figure 1). The proposed services at CSCS are accessible through a RESTful API, FirecREST.

On demand service
Online analysis at PSI refers to processing of data while the scientist is using an instrument. Hence an on-demand service for advance resource reservation is implemented to realize the requirement of availability and to provide a tight feedback loop for the experiments 1. To end users, HPC resources can then be of no differences from the facility’s on-site IT infrastructure resources.

Data catalog extension
Since 04/2018, CSCS and PSI jointly operate the PetaByte Archive located at CSCS. The PetaByte Archive provides user services for long term data storage and retrieval of experimental data from PSI large scale facilities. Archiving and retrieval of data is facilitated by the Data Catalogue (SciCat). The SELVEDAS project extends SciCat by integrating it with the FirecREST API and provides services for the analysis of experiment data stored in the PetaByte Archive at CSCS (figure 2).

Cross-site authentication
Authentication and authorization are based on the PSI user account and access rights. PSI client uses service accounts from CSCS for accessing FirecREST API (figure 3). This separation of responsibilities frees PSI users from needing CSCS's personal accounts, improves user scalability, and allows decentralized authentication since it's done directly by PSI clients to an OIDC server (Keycloak) in CSCS.

Performance Result
The performance evaluation is tested on the typical experiment dataset. Figure 4&5 reports results of workflows and the job submission, one compared between two workflows with the large and small dataset, and another one compared between different GPUs for the job submission.

Conclusion
The SELVEDAS project demonstrates the feasibility of a hybrid cloud infrastructure supporting on-demand services with fast feedback analysis and analysis for archived data on the petabyte archive. The approach has been developed to provide the flexibility and extension to allow other institutions or domains to adopt similar approaches.

References
1. https://cug.org/proceedings/cug2019_proceedings/includes/files/pap138s2-file1.pdf.
2. Related Projects: EOSC, DAAS and PaNOSC

Speaker: Dr Mei-Chih Chang (Paul Scherrer Institute (PSI))