Simulating Beamlines and User Experiments with ...

Simulating Beamlines and User Experiments with “Synchrotron Radiation Workshop” M. Rakitin and O. Chubar

NSLS-II, Brookhaven National Laboratory, Upton, NY, USA with acknowledgements to: L. Berman, V. Bisogni, Y. Cai, N. Canestrari, Y. Chu, E. DiMasi, A. Fluerasu, M. Fuchs, I. Jarige, K. Kaznatcheev, Q. Shen, D. Schneider, A. Suvorov, R. Sweet, E. Vescovo, L. Wiegart, H. Yan, M. Zhernenkov (BNL), R. Reininger (APS), J.P. Sutter (DLS), L. Samoylova (E-XFEL)

References

Abstract The “Synchrotron Radiation Workshop” (SRW) physical optics computer code has been very successfully used for design and commissioning of X-ray optical beamlines at NSLS-II and other Light Source facilities. Among new functionalities recently implemented in the code one can mention physical optics based “propagators” for grazingincidence focusing mirrors that are extensively used in hard and soft X-ray beamlines, for perfect crystals used in hard X-ray monochromators, and for variable-line-spacing gratings that are the key elements of monochromators for soft X-rays. Detailed partially-coherent emission and wavefront propagation simulations are now routinely performed using the SRW code for the hard and soft X-ray beamlines of NSLS-II at BNL (e.g. HXN, CHX, IXS, SRX, FMX, SMI, ESM, SIX), allowing for better optimization and matching of insertion devices and X-ray optics to scientific goals of these beamlines, and the most efficient use of the high brightness and coherence of the NSLS-II radiation beams. A complement open-source cloud-based computing platform named “Sirepo” is being developed in collaboration with RadiaSoft LLC., allowing scientists to execute SRW code from a web-browser using a convenient user-friendly web-interface. This development may significantly improve user experience and usability of SRW code at NSLS-II and other facilities.

[1]

O. Chubar “Precise computation of electron beam radiation in non-uniform magnetic fields as a tool for the beam diagnostics,” Rev. Sci. Instrum., vol.66 (2), 1872-1874 (1995).

[2]

O. Chubar and P. Elleaume, “Accurate and efficient computation of synchrotron radiation in the near field region”, Proc. of EPAC-98, p. 1177 (1998).

[3]

O. Chubar, L. Berman, Y.S. Chu, A. Fluerasu, S. Hulbert, M. Idir, K. Kaznatcheev, D. Shapiro, Q. Shen, J. Baltser, “Development of partially-coherent wavefront propagation simulation methods for 3rd and 4th generation synchrotron radiation sources”, Proc. of SPIE Vol. 8141, 814107 (2011).

[4]

O. Chubar, Y.S. Chu, K. Kaznatcheev, H. Yan, “Application of partially coherent wavefront propagation calculations for design of coherence-preserving synchrotron radiation beamlines”, Nuclear Instrum. and Methods, vol. A649, Issue 1, pp.118-122 (2011).

[5]

O. Chubar, A. Fluerasu, L. Berman, K. Kaznatcheev, L. Wiegart, “Wavefront propagation simulations for beamlines and experiments with ‘Synchrotron Radiation Workshop’ ”, J. Phys.: Conf. Ser. 425, 162001 (2013).

[6]

N. Canestrari, O. Chubar and R. Reininger, “Partially coherent X-ray wavefront propagation simulations including grazing-incidence focusing optics”, Journal of Synchrotron Radiation, vol. 21, pp. 1110-1121 (2014).

[7]

A. Suvorov, A. Cunsolo, O. Chubar, Y.Q. Cai, “Ultrahigh energy resolution focusing monochromator for inelastic x-ray scattering spectrometer”, Optics Express, vol. 23, Issue 24, pp. 31607-31618 (2015).

[8]

O. Chubar, G. Geloni, V. Kocharyan, A. Madsen, E. Saldin, S. Serkez, Y. Shvyd'ko, J. Sutter, “Ultra-high-resolution inelastic X-ray scattering at high-repetition-rate self-seeded X-ray free-electron lasers”, Journal of Synchrotron Radiation, vol. 23, pp. 410-424 (2016).

[9]

D.L. Bruhwiler, O. Chubar, R. Nagler, J. Krzywinski, A. Boehnlein, “An open software framework for advancement of xray optics simulation and modeling”, Proc. of SPIE Vol. 9209, 92090Z (2014).

The present work was supported by US DOE contract No. DE-AC02-98CH10886, US DOE grant No. DE-SC0011237 and SBIR grant No. DE-SC0011237.

Approach to high-accuracy partially-coherent emission and wavefront propagation simulations Averaging (over phase-space volume occupied by e-beam) of the intensity (or mutual intensity, cross-spectral density, or Wigner distribution, …) obtained from electric field emitted by an electron and propagated through an optical system: I  ( x, y )   I 1 ( x, y; xe , ye , ze , xe , ye ,  e ) f ( xe , ye , ze , xe , ye ,  e ) dxe dye dze dxe dye d e

I 1 ( x, y; xe , ye , ze , xe , ye ,  e )  E1 ( x, y; xe , ye , ze , xe , ye ,  e ) M 1 ( x, y, ~x , ~y ; xe , ye , ze , xe , ye ,  e )  E1 ( x, y; xe , ye , ze , xe , ye ,  e )E*1 ( ~x , ~y ; xe , ye , ze , xe , ye ,  e )   B1 ( x, y,  x , y ; xe , ye , ze , xe , ye ,  e ) ~ E1 ( x, y; xe , ye , ze , xe , ye ,  e )  E*1 ( ~x , ~y ; xe , ye , ze , xe , ye ,  e ) exp i ( x ~x   y ~y ) d~x d~y  c 

Evolution and current status of SRW code

Sirepo – open source cloud computing framework for X-ray source and optics simulations

 First official version of SRW was developed at ESRF in 1997-98 (written in C++, interfaced to IGOR Pro); compiled versions are distributed from: http://www.esrf.eu/Accelerators/Groups/I nsertionDevices/Software/SRW

SR calculator:

Wavefront propagation through beamline:

2

This method is general and accurate. For the most part, it is already implemented in SRW code. However, it can be CPU-intensive, requiring parallel calculations on a multi-core server or a small cluster. Several approaches are considered for increasing the efficiency, including use of low-discrepancy sequences (collaboration with R. Lindberg, K.-J. Kim, X. Shi, ANL), “improved Monte-Carlo” type techniques, as well as “coherent mode decomposition”. NOTE: the smaller the e-beam emittance (the higher the radiation coherence) – the faster is the convergence of simulations with this general method. NOTE: convolution can be valid in some cases, such as pure projection geometry, focusing by a thin lens, diffraction at one slit, etc. ~ ~ I  ( x, y )   I 1 ( x  ~xe , y  ~ye ) f ( ~xe , ~ye ) d~xe d~ye

If convolution is valid, the calculations can be accelerated dramatically. The validity of the convolution relation can be easily verified numerically.

Partially-coherent wavefront propagation simulations for CHX @ NSLS-II

 SRW was released to Open Source in 2012 under BSD type license. To make the release possible, permissions were obtained from all previously contributed Institutes: ESRF, European XFEL, SOLEIL, DIAMOND, BNL, and from US DOE.  The main Open Source repository, containing all C/C++ sources, C API, all interfaces and project development files, is on GitHub: https://github.com/ochubar/SRW

http://beta.sirepo.com/light https://expdev.nsls2.bnl.gov/light

 SRW for Python (2.7.x and 3.x, 32- and 64bit) versions were released in 2012.  Phase 2 SBIR Grant was obtained in 2015 from DOE by RadiaSoft LLC and BNL for the development of Sirepo – a webfriendly software framework around SRW.

GitHub: https://github.com/radiasoft Docker: https://hub.docker.com/u/radiasoft Vagrant: https://vagrantcloud.com/radiasoft

Tracking intensity and degree of transverse coherence at a sample (CHX @ NSLS-II) Intensity Distribution

Degree of Transverse Coherence In Horizontal Mid-Plane In Vertical Mid-Plane

Angular Intensity (far field) after Two Slits separated by 10 µm

Approach towards simulation of complete experiments (CHX @ NSLS-II) Known beamline performance and index of refraction maps for samples allow to simulate realistic scattering pattern for future experiments radial average cut

In Horizontal Plane (after vertical slits)

Intensity Distributions for E = 10 keV S1x= 44 μm S1y= 1 mm

Before CRL (@35.8 m)

Before SS1 (@33.5 m)

Before KL (@44 m)

 (r1 , r2 ,  )  | W (r1 , r2 ,  ) | [W (r1 , r1 ,  )W (r2 , r2 ,  )]1/2 W (r1 , r2 ,  ) ~  E (r1 ,  ) E * (r2 ,  ) 

At Sample (@48.5 m)

In Vertical Plane (after horizontal slits)

Flux: 1013 ph/s/.1%bw

Speckle Pattern propagated to the Detector Position

hor. coherence length: ~9.4 µm vert. coherence length: ~13.4 µm Good agreement with 2-slit interference simulation results

IVU21

0m

Partially-coherent wavefront propagation simulations for FMX, IXS, ESM beamlines Grazing-incidence focusing mirrors VLS grating (ESM @ NSLS-II) Advanced high-resolution crystal optics with imperfections (FMX @ NSLS-II) (IXS @ NSLS-II) Beamline Design: HFM

Horizontal Plane

20 m Vertical Plane

SSA

40 m

Horizontal SSA Size: 30 μm Photon Energy: 12.7 keV Flux at Sample: ~5.4x1013 ph/s/.1%bw

KB

60 m

Sample Plane

KB simulated using Grazing-Incidence “Thick Optical Element” Propagator based on “Local Ray-Tracing”. KB Surface Height Error simulated by corresponding Phase Shifts (“Masks”) in Transverse Plane at Mirror Locations.

A.Suvorov, Y.Q.Cai, J.P.Sutter, O.Chubar, Proc. of SPIE Vol. 9209, 92090H (2014)

R.Reininger, S.L.Hulbert, P.D.Johnson, J.T.Sadowski, D.E.Starr, O.Chubar, T.Valla, E.Vescovo, Rev. Sci. Instrum. 83, 023102 (2012)

Extended testing of new Physical Optics Propagator for Crystals IXS Monochromators contain: DCM: 2 Crystals HRM: 4 Crystals of HRM

Part.-Coherent Wavefront Propagation Simulations: N.Canestrari, V.Bisogni, A.Walter, Y.Zhu, J.Dvorak, E.Vescovo, O.Chubar, Proc. of SPIE Vol. 9209, 92090I (2014)

In these simulations, the horizontal secondary source slit size was set to be equal to the vertical size (Δx = Δy); mirrors’ height / slope errors were not taken into account (to be included in next series of simulations).

Intensity Distributions at Sample Without Mirror Errors

Mirror Slope Error

With Mirror Errors

Energy Resolution

Spatial Resolution

Flux (finite-bandwidth) at Sample

as functions of the Secondary Source (Monochromator Exit) Slits

Spectral Flux at Sample

Mirror Height Profile Error

hor. cuts (y = 0)

Mirror Surface Error is not taken into account

vert. cuts (x = 0)

E0 ≈ 9131.7 eV

∆𝐸 Τ𝐸 > (𝑚𝑁)−1

Characterizing undulator source and performing e-beam diagnostics (HXN @ NSLS-II) Spectral Flux & Intensity Distributions at 5th Harmonic at Different Undulator Gaps at 30.4 m

On-Axis Spectral Flux per Unit Surface Good Agreement with Accelerator Physics data: εx= 2.1 nm for Bare Lattice, εx= 0.9 nm with 3x7 m DW closed

Eph ≈ 8.0 keV 5th UR Harm. Low-Beta Straight Section of NSLS-II βx= 1.84 m βy= 1.17 m

~Poor Agreement with Accelerator Physics data: σE/E= 0.5x10-3 for Bare Lattice, σE/E= 0.9x10-3 with 3x7 m DW closed UR based e-beam diagnostics was used at ESRF (P. Elleaume et al.) and at APS (A. Lumpkin, E. Gluskin et al.)

Harmonic width is sensitive to ebeam Energy Spread (and other factors, e.g. undulator field quality) Intensity in “Side Lobe” is sensitive to e-beam Horizontal Angular Divergence

On-Axis Gap Spectrum at ~8.0 keV Photon Energy (5th Harmonic)

Two different VLS Gratings (160 mm long) were used: 𝑎0 = 800 lines/mm for E = 20 eV; 𝑎0 = 600 lines/mm for E = 60, 100 eV

SRW enables advanced commissioning of SRX Beamline @ NSLS-II

Improvement of Undulator Performance

X-Ray Diagnostics

5th Harmonic ~6.4 mm Gap

Simulations

Undulator: λu= 20 mm, Lu = 3 m Low-Beta Straight Section of NSLS-II: βx= 1.84 m (σx’= 22 µrad at εx= 0.9 nm) βy= 1.17 m (σy’= 2.6 µrad at εy= 8 pm)

Measurement with Ion Chamber Detector

Undulator simulation successes:

Measured and Calculated Intensity Distributions on YAG Screen at ~200 μm Vertical Size of S0 Slit at fully open SSA , E = 8 keV

• Identify case of an under-performing InVacuum Undulator • Find reason for the reduction of spectral performance (small misalignment of magnet arrays)

Measurements after Monochromator (Scintillator Screen + Lens +CCD)

• Design and implement most efficient correction.

Calculations (SRW)

5th Harmonic ~6.8 mm Gap

Measurements

Fitting of Measured Intensity Distributions in Diffraction Patterns by Calculated ones allows to estimate Effective Source Sizes and X-Ray Coherence Characteristics.

Measurements using Ion Chamber Detector, spring 2015.

Measurements in 2015-2016. Thanks to SRX Team! Measurements in spring 2015. Thanks to HXN Team!

Work supported by SBIR grant No. DE-SC0011237