The Bifrost spectrometer General update Rasmus ToftPetersen Technical

The Bifrost spectrometer General update Rasmus Toft-Petersen Technical University of Denmark

Outline: Changes since last STAP meeting • • • Bifrost instrument outline (yet again) Guide system Tank design Filter and radial collimator Challenges: Background, monitors, detector safety

The team Core team: Liam Whitelegg and Rasmus Toft-Petersen Jonas Birk Rodion Kolevatov Finn Saxild Kim Lefmann Christof Niedermeyer Marko Marton Phillipe Bourges Bjorn Hauback Henrik Ronnow Niels B Christensen Keld Theodor Sylvain Rodrigues

Outline

Front end – To. F diagram Bifrost aims to take advantage of the full ESS pulse.

Backend

Feasibility intrinsically low number Incoming momentum Matrix element, golden rule, physics

Performance Triple-axis spectrometer Two types of experiment on a cold neutron spectrometer, 1) Workhorse Mono Flux 107 -108 n/s/cm 2 Spatial angle 0. 015 steradians Energy transfer Single value Time-of-flight spectrometer Mono Flux 105 -106 n/s/cm 2 Spatial angle 3 steradians Energy transfer Continuous 2) Borderline feasible 1) Inherently small signal/noise – spin liquids Bifrost 2) Small moment – High Tc superconductors 3) Inherently small crystals – ionic conductor Polychromatic Flux magnets Spatial angle 4) Extreme environment needed (> 10 T, 30 m. K, 10 Gpa) Energy transfer 108 -1010 n/s/cm 2 0. 5 steradians Continuous

Scope setting • • • 13. 45 M€ approved Only 4/6 analysers funded No dedicated Bifrost magnet No dilution stick No Polarization analysis Went thourgh TG-2 in March this year Analyser energy Distance from sample [m] Take-off angle [deg] 2 -Theta coverage [deg] Total area of analyser [mm 2] 2. 7 me. V 3. 0 me. V 3. 4 me. V (NIS) 3. 9 me. V 4. 4 me. V (NIS) 5. 0 me. V 55. 11 51. 09 46. 97 43. 04 39. 98 37. 07 4. 6 4. 85 5. 1 5. 35 5. 6 5. 85 6075 7360 8840 10738 13230 15960 1. 05 1. 13 1. 21 1. 32 1. 42 1. 52

Guide validation Changes in the guide: We lengthened the defocusing/focusing sections, and narrowed the straight section width to 6 cm. This was to minimize bandwidth chopper opening/closing times. Validation: We made a Mc. Stas-file from scratch using the CAD-geometries, included gaps and imperfections due to piecewise linerarity of the guide, etc.

Guide validation Flux n/cm 2/s/Å Wavelength [Å]

Guide validation Position [m] Divergence [deg]

Cave design change dilution stick problems • • Weight of dilution stick ~ 50 Kgs Length of dilution stick ~ 1700 mm Hook height required ~ 7. 2 m All of this converges on a 9 m tall cave Structural difficulties Additional racks and supplies If stored within cave, larger footprint needed

New design • • • Predominant sample access now from roof through top loading Moved to dual access via platform access Crane now external to cave Cave height reduced from 9 m to 5. 2 m Sample environment auxiliary equipment positioned on roof • Another picture here

Sample access • • Lead access hatch linked to PSS system Ladder access down to platform Lead lining along base and side to provide shielded access Secondary access through cave labyrinth to sample position • Picture of sample platform

Fitting the detectors

Staggering the analyzers

Tank design We have 9 Q-channels, each covering a 2 theta of 5 degrees or more. We have 5 analysers (not 6), to be installed through hatches on the back. Energies are: 2. 7 me. V, 3. 2 me. V, 3. 8 me. V, 4. 4 me. V and 5. 0 me. V. We went away from symmetric Rowland geometry to make detectors fit to one plane

Point to point focusing Symmetrical Asymmetrical

Be transmission 100 10 -1 10 -2 10 -3 1Å 2Å 3Å 4Å 5ÅÅ

Bifrost Bragg peaks Bragg peak distribution for a standard cubic sample with lattice parameter 4 pi for high and low wavelengths

Shoulder Bragg peaks

Aluminum and graphite scatters at shoulder energies

Transmission versus thickness We decided on an integrated Be-filter and radial collimator. It is the only way to have enough space for aggressive collimation.

Conclusion: Long blades and Be-filter pieces

Integrated filter We decided on an integrated Be-filter and radial collimator. It is the only way to have enough space for aggressive collimation.

Prompt gammas At 5 MW, the prompt gamma intensity is of the order of 5*109 gammas/s/cm 2 The gamma flux at the detector ensembles are of the order of 3*108 gammas per minute. This is on the threshold of discrimination. Lead produces neutrons!

Diagnostics Monitor

Monitors Flux curve: The variation In flux is between 1010 n/s/cm 2/Å and 2*108 n/s/cm 2/Å. a factor of 50 variation. Cold flux at fully open: 1010 n/s/cm 2/Å Cold flux at 0. 1 ms: 5*108 n/s/cm 2/Å Near thermal flux at 0. 1 ms: 107 n/s/cm 2/Å

Monitors Flux 71, 4 ms pulse @ sample Binning Time We use the white beam, and can simulate a nice flux distribution in the frame. But that’s not how it looks in real life.

Monitors Flux 71, 4 ms pulse @ sample Binning Example from TAS Time

Monitors Flux 71, 4 ms pulse @ sample Binning Lets say we need a sampling 3 times better than the pulse duration – 2000 bins Time The flux might be changing, but not much on a time scale smaller than the pulse duration. But you might want Wavelength band - 1. 2 -2. 9 Å Flux: 107 n/s/cm 2/Å Pulse duration: 0. 1 ms. If we want flux determination better than 1, 5 % in each bin we need 5000 cts pr bin: This is 107 counts. We would need 0. 5 % efficiency at 1. 2 Å in 3 mins. That would normally go to 2. 5 % at 6 Å. Gigahertz range… Safety, attenuators, another monitor? Argh.

Day 1 capabilities • Accelerator at 2 MW -> Max flux limited to 1010 n/s/cm 2 • Only 4 working analyzer arrays, to be upgraded to 5 (!) • No polarization analysis on Day 1 (upgrade) • Used Oxford 15 T magnet on Day 1 (upgradable to a newer magnet designet for Bifrost) • Pressure cells to be developed, but Day 1 availability still uncertain.

Thank you for your attention