Inelastic Xray Scattering at NSLS II IXS Program

Inelastic X-ray Scattering at NSLS -II IXS Program, and Current Project Beamline Design Yong Cai With contributions by John Hill, Xianrong Huang, Zhong, Scott Coburn (NSLS-II) IXS@NSLS-II Workshop, February 7 -8, 2008 Alfred Baron (RIKEN), 1 Yuri Shvyd’ko (APS), BROOKHAVEN SCIENCE

IXS: Current Status • A powerful (but weak) probe offering energy and momentum • • resolved information on dynamics and excitations in condensed matter systems Practical applications with highly successful state-of-the-art implementation at ESRF, SPring-8 and APS Broad areas of applications, including but not limited to the followings: Lattice dynamics – Phonons in solids, vibration modes and relaxation processes in disordered systems: routinely operational at ~1 me. V resolution with ~109 photons/sec at ~20 ke. V using symmetric backscattering crystal optics Charge dynamics – Plasmon in simple metals, complex electronic excitations in strongly correlated systems: best resolution at ~50 me. V with ~1011 photons/sec at ~10 ke. V for the most demanding experiments Element specific applications – Chemistry and materials science, materials under extreme conditions of pressure and temperature: x-ray IXS@NSLS-II Workshop, February 7 -8, 2008 Raman scattering, x-ray emission and absorption spectroscopy by partial BROOKHAVEN SCIENCE 2

IXS Opportunities at NSLS-II • Key scientific drivers identified thru community inputs (CDR, user workshops) 0. 1 -me. V scientific case: filling the gap between high and low frequency probes • Visco-elastic crossover behaviors of disordered systems and fluids • New modes in complex fluids and confined systems • Collective modes in lipid membranes and other biomolecular systems 1 -me. V scientific case: • Relaxation processes in disordered systems (glasses, fluids, polymers …) • Phonons in single crystals, surfaces, thin films, high pressure systems, small samples • Exotic excitations in strongly correlated systems 10~50 -me. V (including lower resolution) scientific case: • Non-resonant and resonant inelastic scattering on charge dynamics • X-ray Raman scattering, x-ray emission and absorption spectroscopy by partial fluorescent yield combined with small beam for small samples of micrometer scale, high pressure, high fields, low and high IXS@NSLS-II Workshop, February 7 -8, 2008 temperature … BROOKHAVEN SCIENCE 3

Scientific Mission of the Project Beamline • • Develop advanced instrumentation that enables IXS experiments at extremely high resolution of 0. 1 me. V at ~10 ke. V – a key goal for NSLSII. Why 10 ke. V, as opposed to 20 ke. V or higher? Technical merits: higher brightness and flux for NSLS-II Scientific merits: better momentum resolution for a given solid angle, or the same momentum resolution for larger solid angle Counting efficiency: a complex issue, sample and optics dependent, require careful evaluation Possible scheme based on highly asymmetric backscattering optics proposed by Yuri Shvyd’ko working at 9. 1 ke. V, require considerable R&D (more from Xianrong Huang and Zhong’s talks) What are the alternatives? IXS@NSLS-II Workshop, February 7 -8, 2008 4 BROOKHAVEN SCIENCE

Specifications and Requirements • • Radiation source Provide flux > 1010 phs/s/0. 1 me. V at 9. 1 ke. V, to provide > 109 phs/s/0. 1 me. V at sample, require beamline overall efficiency > 10% Essential source size and divergence stability, vibration and thermal stability for achieving the 0. 1 me. V resolution (more from Xian. Rong Huang’s talk) Beamline and Endstations Two end stations, one with 0. 1 me. V and the other with 1 me. V resolution Primary beam energy at 9. 1 ke. V to match scheme for achieving 0. 1 me. V resolution Beam energy tunable over 7~12 ke. V for added flexibility (e. g. , other refractions) Energy scan range: order of 0. 1 ~ 1 e. V, appropriate for the excitations to be studied Momentum resolution: 0. 1 ~ 0. 4 nm-1 at 9. 1 ke. V Momentum range: up to 80 nm-1 to cover typical BZ sizes, ~120º at 9. 1 ke. V IXS@NSLS-II Workshop, February 7 -8, 2008 BROOKHAVEN SCIENCE 5 Spot size: ≤ 10 μm (H) x 5 μm (V), required for high pressure

Overall Beamline Layout R-wall: 28. 3 m Walkway: 60. 8 m Beamline o K B ccupies a h HRM 1. 0 me. V Endstation* KB HRM VFM VC DCM M 0. 1 me. V Endstation - CDDW HRM (0. 7 me. V) - CDDW HRM (0. 1 me. V) - CDW Analyzers (0. 7 me. V) - CDW Analyzers (0. 1 me. V) - larger q range (~80 nm-1) - 0. 1 nm-1 q resolution * (not in the project scope) igh-β strai ght sectio - limited q range FOE - HHL DCM - VFM / VCM IXS@NSLS-II Workshop, February 7 -8, 2008 6 BROOKHAVEN SCIENCE

Optical Scheme for 0. 1 me. V • • CDDW mono to achieve 0. 1 me. V with ~100µrad angular acceptance Multilayer mirror to achieve 5~10 mrad angular acceptance for CDW/CDDW a Δθe = ~ 5μrad E = 9. 1 ke. V for Si(800) Side view ΔE = 0. 1 me. V, φ = Si(800) 89. 6° Si(111) Si(220) ~20μrad divergence < 5: 1 ~100μrad focusin g Si(800) acceptance IXS@NSLS-II Workshop, February 7 -8, 2008 7 BROOKHAVEN SCIENCE

Multilayer Mirror for Analyzer • Requirements: 100: 1 demagnification (from 10 mrad to 0. 1 mrad) 10 mrad angular acceptance Greater than 50% efficiency Possibility for parallelization of data collection E = 9. 1 ke. V m 0 c ~1 CDDW analyzer with ~0. 1 mrad acceptance We hope the beam here is a slightly focused beam with divergence ~0. 1 mrad and beam height ~50 µm We hope the mirror length to be ~10 cm. It can be either a (1) laterally graded multilayer or (2) a parabolically / elliptically bent periodic multilayer IXS@NSLS-II Workshop, February 7 -8, 2008 8 BROOKHAVEN SCIENCE

Conceptual Beamline Layout • Multilayer mirror with 100: 1 demagnification to provide 10 mrad (vertical) angular acceptance for analyzer ~10 m arm length IXS@NSLS-II Workshop, February 7 -8, 2008 9 BROOKHAVEN SCIENCE

Experimental Floor Layout 25 m BM line • • ~53 m Space may be sufficient for only one end station based on current scheme and 10 mrad angular acceptance for analyzer ID line Another ~8 m may be required for a second end station with 5 mrad angular acceptance for analyzer, space for mirrors and phase plates ~61 m + 2 m (possibly) IXS@NSLS-II Workshop, February 7 -8, 2008 10 BROOKHAVEN SCIENCE

Insertion Device • U 20 (IVU), or U 14 (SCU) on a high-β straight section Figure of Merit: Flux / me. V – provide flux > 1010 phs/s/0. 1 me. V at 9. 1 ke. V Name U 20 U 14 Type IVU SCU Period (mm) 20 14 L (m) 3. 0 2. 0 No. Periods 150 143 Bmax (T) 0. 97 1. 68 Kmax 1. 81 2. 20 Tot. Power (k. W) 8. 024 16. 0 8 On-axis 62. 65 103. * U 20: baseline device PD 7 IXS@NSLS-II Workshop, February 7 -8, 2008(k. W/mrad 2) BROOKHAVEN SCIENCE 11

Insertion Device Optimization • Optimize to maximize flux at 9. 1 ke. V: magnet period vs length vs minimum gap Until superconducting devices become an option, the U 20 -5 m seems to be the Basic Undulator Parameters #1 #2 #3 #4 #5 Baseline Proposed best option Name Type Period (mm) Length (m) No. of Periods Minimum gap (mm) Bpeak (T) Keff Minimum 5 th harmonic energy, linear mode (e. V) Total power (k. W) Power density (k. W/mrad 2) U 20 -3 m PMU U 22 -6 m PMU 20 3 150 5 0. 969 1. 81 22 6 272 7. 5 0. 725 1. 4893 8099. 45 8. 024 62. 65 9210. 20 8. 955 84. 32 U 21 -4. 5 m U 20 -4. 5 m PMU 21 20 4. 5 214 225 7. 5 7 0. 85 0. 88 1. 66671 1. 64336 8518. 10 9. 244 78. 15 9091. 00 9. 922 85. 02 U 20 -5 m PMU 20 5 250 7 0. 88 1. 64336 9091. 00 11. 02 94. 47 Performance at 9. 1 ke. V, 5 th harmonic (for 9. 3 m high-β straight, beta_x=20. 8, beta_y=2. 94) Brightness (phs/sec/mrad 2/mm 2/0. 1%BW) Total Flux (phs/sec/0. 1%BW) Brightness ratio with baseline: Flux ratio with baseline: 4. 68 E+20 4. 60 E+20 4. 25 E+20 8. 00 E+14 1. 06 E+15 1. 01 E+15 IXS@NSLS-II Workshop, February 7 -8, 2008 1. 00 0. 98 0. 91 1. 00 1. 33 1. 26 12 4. 90 E+20 5. 42 E+20 1. 20 E+15 1. 35 E+15 1. 05 1. 16 1. 50 1. 69 BROOKHAVEN SCIENCE

Extended Long Straight • • • It is possible to obtained one 15 m extended long straight by extending one 9. 3 m high-β straight and shortening two adjacent 6. 6 m low-β straights of the current lattice Preliminary solution indicates that it would be possible to put in two U 20 -5 m devices with refocusing magnet. Potentially more than triple (3. 38 times) the flux compared to one U 20 -3 m baseline device There accelerator issues (tune shift compensation, radiation shielding) and cost impact. βx = 18. 5, βy = 3. 5 for extended LS βx = 0. 8, βy = 0. 8 for short straights IXS@NSLS-II Workshop, February 7 -8, 2008 13 BROOKHAVEN SCIENCE

Straight Section • • High-β straight section (9. 3 m) is required for a long device Additional benefits associated with a high-β compared to a low-β straight section : • Bigger horizontal source size but smaller horizontal divergence, leading to smaller horizontal photon beam size shorter horizontal mirrors • Lower power load within the central cone of the photon beam delivered to the first optics heatload is more manageable for first optics, even for two U 19 undulators. • Important: Flux within the central cone remains the same! Distanc K Low-β High-β Photon beam size and power (one U 19, at 9. 1 ke. V) thru angular aperture e, m of 4 s(H)x 4 s(V). Aperture Beam Size 2 Power, (Hx. V), μrad 2 30 1. 71 3 30 0. 98 1 (Hx. V), mm 79. 2 x 22. 8 2. 4 x 0. 7 W 118 65 (Hx. V), μrad 2 29. 2 x 20. 0 (Hx. V), mm 1. 0 x 0. 6 W 38 21 IXS@NSLS-II Workshop, February 7 -8, 2008 14 BROOKHAVEN SCIENCE

First Optics Enclosure (FOE) C sc ool r e W een d f lu hi. te b ea Bl m ad s e BP lit B M co rem llim ss at tra or hl un D ia g m C o m ool nd as ed w k fi ind xe o d w PM W B hi st rem te b op s e st am ra hl & un D g C M B an t ua dr Q • DCM: cryogenic cooled Si(111). FEA analysis on first crystal of DCM shows < ± 5 µrad slope error at 115 W heat load For high-β straights, thermal effect would be less a problem (lower heat load). VF M • • IXS@NSLS-II Workshop, February 7 -8, 2008 15 BROOKHAVEN SCIENCE

Beamline Collimating and Focusing Optics • Candidate: Bimorph mirrors • Supplied by SESO • Multiples of 150 mm in length • Each segment can be bent independently (2 electrodes for cylindrical and 4 electrodes for elliptical) • Bimorph can in principle correct its own tangential polishing slope errors and also correct for tangential errors introduced by other optics • KB bimorph pair can in principle correct for sagittal and tangential errors • Small beam size allows short mirrors with small incident angle BROOKHAVEN SCIENCE 16 (2. 5 mrad) and Si substrate for IXS@NSLS-II Workshop, February 7 -8, 2008

Shadow Ray Tracing with FEA Results Case Model , LN 2 cooled DCM (εx = 0. 55 nm-rad) Spot size , μm (No distortion) Flux, ph/sec (No distortion) Spot size, μm (W/ distortion) Flux, ph/s/e. V (W/ distortion) 1 Lo-β, K=0. 981, 1 U 19 -3 m, 22 W 2. 0 x 1. 8 1. 50 x 1013 2. 1 x 2. 1 1. 49 x 1013 2 Hi-β, K=0. 981, 1 U 19 -3 m, 7 W 7. 0 x 1. 8 1. 49 x 1013 7. 0 x 2. 0 1. 49 x 1013 2 a Same as 2, but water cooled 7. 0 x 1. 8 1. 49 x 1013 16 x 7. 4 1. 51 x 1013 3 Lo-β, K=1. 714, 1 U 19 -3 m, 71 W 2. 0 x 1. 9 2. 88 x 1013 2. 3 x 4. 2 2. 81 x 1013 3 a Same as 3, but hot spot at 2. 0 x 1. 9 2. 88 x 1013 2. 1 x 0. 5 125 K IXS@NSLS-II Workshop, February 7 -8, 2008 2. 92 x 1013 3 b Same as 3, but 2 U 19 -3 m, 2. 0 x 2. 017 5. 80 x 1013 2. 9 x 5. 9 BROOKHAVEN SCIENCE 5. 14 x 1013

0. 1 me. V Endstation Layout Features • Pre-focusing (up to 5: 1) to reduce Instrument NSLS-II, 3 m U 19 length of collimator (C) crystal for 0. 1 me. V HRM • KB mirrors after HRM for micro focusing to ≤ 10(H) x 5(V) μm 2 • Temperature scan of energy transfer • Analyzer(s) based on the CDW or CDDW scheme with multi-layer focusing mirror(s) to achieve 10(H)x 5(V) mrad 2 angular acceptance • High (0. 1 nm-1) momentum resolution in forward scattering, limited scan range Intensity at Sample Extremely clean energy resolution > 1 x 109 • ph/sec/0. 1 me. V (estimate)function APS, 5 m U 30 1 x 109 ph/sec/1 me. V SPring-8, 4. 5 m U 32 5 x 109 ph/sec/1 me. V IXS@NSLS-II Workshop, February 7 -8, 2008 18 BROOKHAVEN SCIENCE

1 me. V Endstation Layout Features • Pre-focusing (up to 2: 1) to reduce length of collimator (C) crystal for 0. 7 me. V HRM • KB mirrors after HRM for micro focusing to ≤ 10(H) x 5(V) μm 2 • Temperature scan of energy transfer • Analyzer(s) based on the CDW or CDDW scheme with multi-layer collimating mirror to achieve 10(H)x 5(V) mrad 2 angular acceptance A natural step in achieving 0. 1 me. V resolution • Large momentum scan range (up to 80 nm-1 at ) for single crystal studies, Instrument Intensity Sample require phase plates NSLS-II, 3 m U 19 > 1 x 1010 ph/sec/1 me. V • Extremely clean energy resolution (estimate) function 9 APS, 5 m U 30 1 x 10 ph/sec/1 me. V SPring-8, 4. 5 m U 32 5 x 109 ph/sec/1 me. V IXS@NSLS-II Workshop, February 7 -8, 2008 19 BROOKHAVEN SCIENCE

Schedule Summary • 2008~2011: Design phase, interaction with BAT, actively pursue 0. 1 me. V R&D • 2011~2013: Construction phase, long lead-time procurements to begin in final • 2014: Integrated testing phase, to be ready to take beam IXS@NSLS-II Workshop, February 7 -8, 2008 20 BROOKHAVEN SCIENCE

Outstanding Issues • • • Multiple undulators on extended straight for more flux (depends on AS design) tuning compensation for beam dynamics, heat load on second undulator, cost impact CDDW scheme to achieve 0. 1 me. V HRM: Substantial R&D effort required (proof of scheme by end 2010, final design phase) Multilayer mirror to achieve 10 mrad angular acceptance for CDW / CDDW analyzers: (proof of scheme by early 2010, preliminary design phase) Risks with current approach and mitigation strategies: 1 me. V is promising, 0. 1 me. V remains a major challenge, aggressively pursue the R&D and seek alternatives Possibility to combine two end stations into one: energy resolution tuning Parallel data collection: multiple analyzers, area detector IXS@NSLS-II Workshop, February 7 -8, 2008 21 Your inputs are important! BROOKHAVEN SCIENCE