A High Power, Spallation-Driven Ultracold Neutron Source

1. A High Power, Spallation-Driven Ultracold Neutron Source K. K. Leung, M. Makela, G. Muhrer, C. Morris, R .W. Pattie, A. Saunders, A. R. Young

2. Outline • Motivation • A Brief Review of Status in 2015 • The Optimized Geometry • Cryogenics and UCN Extraction

3. Motivation and Goals • UCN experiments: EDM, beta-decay (lifetime and angular correlations), limits on short range forces and gravitational state studies, limits on the electric charge of the neutron, mirror-neutron and dark matter searches, N-Nbar oscillation searches • Goal for UCN Source Figure of merit for loading a particular experiment to the highest density involves some combination of total UCN production (“current”) and limiting UCN density in the source →ideally highest density with shortest required storage time All of these efforts currently limited in part by available UCN density! Goal of this talk to answer question: could a UCN source be competitive with CN for an NNbar experiment?

4. NNbar with UCN hadron tracking and calorimeter outer detector and muon veto vacuum vessel magnetic shielding n amplitude sampled when UCN hits surface Box filled with UCN gas…each bounce is a “sample” then many samples/neutron, longer average flight times (~1/3 sec), but very large UCNcurrent required

5. Pros and Cons Advantages: • No long, shielded beamline required: more compact and less $ • Some sources soon available: possibly much less expensive • Same ability to turn “on” and “off” effect w/magnetic field Disadvantages: • Limits less stringent than those obtained with CN beam geometry

6. An example geometry: Based on SD2 source now almost constructed at PULSTAR (1 MW) Transport codes vetted during development of UCNA experiment

7. Preliminary results for base case(annhilation det eff = 1, 1 year running): NCState geometry, 4 cm thick SD2, 18 cm guides, 0.050s SD2 lifetime, and we only model storable UCN Primary flux: 1.2 x 107 (below 305 neV) -- 3.5 MW Box loading efficiency: 30% 325s avg. residency in experiment Best case: diffuse walls, specular floor discovery potential = 2.3x109 Ns τnn > 1.1x108 s

8. Ultimate Reach with PULSTAR • “straightforward gains” • 4 years of running τnn> 2.2x108 s • “speculative gains” • Multiple reflections (x1-4) Serebrov and Fomin; coherent n amplitude enhancement (x2) Golub and Yoshiki • Compound parabolic concentrators in floor • Optimized, higher “m” wall coatings

9. Comparison (4y expt) PULSTAR (3.5 MW) SD2, optimized: τnn> 2.2x108 s (original predicted primary current 1.2x107) SuperK: τnn> 2.7x108 s PSI SD2 (106 UCN/s max): τnn> 7.6x107 s ~ same as ILL turbine (measured current < ~1x106) FRM II: τnn> 5.3x108 s (perhaps more) (predicted current = 7x107 with extraction from SD2) SerebrovLHe: τnn> 7.4x108 s (predicted extracted current = 7x107, 60% loading efficiency) 1.6 K spallation-driven LHe: τnn> 2.0x109 s (predicted extracted current > 5x108 , 60% loading efficiency) (possibly 2 or more times greater if full flux from ILL source used) These are very Interesting! Geometry under development Factor of 2 gain in efficiency under investigation…

10. A strategy for the development of an optimized, high integrated flux UCN source Ideally: (1) For selected source materials (SD2, LHe, SO2) identify UCN source operating temperature and cryogenic strategy (2) Optimize UCN production and beam-current one can handle for moderator geometry, consistent with (1) Given what we’ve learned in the past 10 years and will learn in the next few years, this program could define an optimal geometry for available beam power at next generation sources… What we actually did was pick a specific idea and explore…

11. Conceptual Neutronics Modeling of a High Current, UCN Production Geometry G. Muhrer and A. R. Young Guiding concepts: Build from existing 200 kW, Lujan center target Use existing, vetted cryogenic kernels Use strategies benchmarked in development of Mark III target Use large volume (40 l) LHe as converter Inspired by a suggestion from Masuda and the source design of Serebrov –use source as coolant! Assume 100W cooling power available for LHe(e.g. CERN subcooled He system) – adjust beam power and shielding to accommodate heat! J. Carpenter, R. Kleb, T. Postol, R. Steuk, and D. Mildner, Nuclear Instruments and Methods in Physics Research 189, 485 (1981).

12. UCN production function Scattering kernel near 1 K in equilibrium w/saturated vapor pressure (low P) Korobkina et al., Phys. Lett. A 301, 462 (2002)

13. The Starting Geometry 40 liters Philosophy: Use moderating shielding material to isolate low-temperature components from gamma and particle heating Use a large target and raster beam to keep energy density down (~35 W/cm3, comparable to Lujan Mark III), so edge cooling possible

14. 2015: A first stage of optimization ~Same as optimized Mark III target! 1.0*108 UCN/s/100mA Heat load @ 100mA ≡ 80KW Total heat: 13.9 W Neutron heat: 10.8 W Photon heat: 2.4 W Proton heat: 0.7 W 7.14*108 UCN/s/100W (heat in the He) 800 MeV p+ 800 MeV p+ 53cm 40L-He Be(20K) W W D2 (70% ortho, 19K) D2O (5cm) Bi(300K) Dominated by neutron-produced heat Good Start – with 710 kW of proton beam, can produce × 108 UCN/s! A. Young, T. Huegle, M. Makela, C. Morris, G. Muhrer, and A. Saunders, Physics Procedia 51, 93 (2014).

15. Inverse geometry (5): Neutron spectrum in He-4

16. Work since 2015 Results Preliminary… • Further optimized neutron transport properties • Updated the UCN production cross-sections to account for (P,T) of our source (also cross-checked MCNP-X results with MCNP-6 using finer binning & lower threshold) • Established some critical parameters for cryogenics (diameter of heat exchanger conduit) • Crude analysis of UCN extraction

17. Inverse Geometry: Optimization

18. After Optimization: Parameterized optimized were: Thickness of D2O pre-moderator Thickness of LD2 moderator Position of target

19. Target Position Moving the target back relative to the source produces gains in production (at least 20% more over LD2 gain) × 109 UCN/s with 1 MW of beam power ~x2 in prod rate/100uA relative to Lujan ~0.1 heat deposition/100uA relative to Lujan!

20. UCN Production – optimum geometry • Updated cross sections (results consistent with 2015 results at 20% level) • Investigated expected T and P dependence of production (based on recent measurements) – about 5% variation from initial, low T, P production K. K. H. Leung, S. Ivanov, F. M. Piegsa, M. Simson, and O. Zimmer, Phys. Rev. C 93, 025501 (2016). K. H. Andersen, et al., Journal of Physics: Condensed Matter 6, 821 (1994). K. Andersen, “Private communication,” (2015).

21. Cryogenics

22. A Cryogenics System Conductive Cooling through large diameter pipe to heat exchanger • Use sub-cooled He (pressurized to 1 atm) • Strongly suppresses thermal “shocks” and bubbling • Practical advantages for long term maintenance (suppress air-in leaks) • Use flowing 2-phase He to carry heat from exchanger • 3-stage cold compressor systems developed at CERN

23. Experimental data available for the temperature and pressure range of interest • Analyze for our system – start with large diameter heat exchanger (~ 14 cm) to remove 100 W w/ ΔT < 0.05 K with L = 70 cm • Base temperature ~ 1.6 K for 3 cold compressors, 1.4 K with 4 compressors P. Lebrun and L. Tavian,Cooling with Superfluid Helium, contribution to the CAS-CERN Accelerator School: Superconductivity for Accelerators, Erice, Italy, 24 April - 4 May 2013, edited by R. Bailey.

24. UCN Extraction (in progress)

25. Model assuming worst case: UCN entering heat exchanger are lost

26. Competition between loss in heat exchanger and upscatter UCN resident in source for less than a few seconds • UCN resident in source for less than a few seconds, so 3 sources of loss considered • Upscatter • Loss into Heat Exchanger (assume 100% absorbing) • Walls (negligible)

27. Note: upscatter losses reduce by more than a factor of 2 at 1.4 K… but limited by losses into heat exchanger aperture

28. UCN reflecting membrane with small aperture and very high heat conductivity

29. Competition between loss in heat exchanger and upscatter UCN resident in source for less than a few seconds Identified UCN reflecting membrane with smaller aperture in heat exchanger (motivates dropping operating T) – factor of 2.5 smaller losses

30. Moderator R&D G. Muhrer, T. Heugle • G. Muhrer: Moderator materials in confined spaces ( H2O in silica microspheres): • reduction of radiation damage • manipulation of excitation spectrum (more of the “right” phonons) • change of solidification temperature • more predictable crystallization process • (2) T. Huegle:Triphenlymethane a) M. Hartl, L. Daemen and G. Muhrer, Microporous and Mesoporous Materials, 161 (2012) 7-13b) G. Muhrer, M. Hartl, M. Mocko, F. Toveson and L. Daemen, NIMA, 681 (2012) 91-93

31. Triphenylmethane Inelastic Neutron Scattering Spectrum • Melting point 93°C, boiling point 359°C • Three rotational modes of phenyl rings provide low energy excitations which prevent upscattering • Comparably stable in radiation field

32. Conclusions • We present a source geometry capable of producing larger currents of UCN than any proposed or constructed source, and densities comparable to the most powerful sources being proposed • Our approach was to develop a shielded moderating design to ensure the dominant heat-load came from useful neutrons, and then to optimize with the constraint of available cooling. UCN extraction is still under development • The cryogenics approach is based on operating and industrially available solutions – and shoud be very stable against thermal “shocks” • Ongoing moderator research at ESS, J-PARC, SNS, LENS etc... could play an important role in defining the optimized geometry!

33. Preliminary Production Study Summary [21] A. Young, T. Huegle, M. Makela, C. Morris, G. Muhrer, and A. Saunders, Physics Procedia 51, 93 (2014). Beryllium canister increases the UCN flux by 50%. Beryllium canister decreases the heat load in the He by about 15%. Heavy water pre-moderators decrease the heat load in the He by 20-25% (depending on the thickness). No significant UCN flux increase has been observed from introducing a heavy water pre-moderator. Liquid D2 moderator lowers the gamma heating in the He significantly.

34. Lujan Geometry

35. Lujan like geometry 9.42*107 UCN/s/100mA Heat load @ 100mA ≡ 80KW Total heat: 66.7 W Neutron heat: 44.7 W Photon heat: 18.7 W Proton heat: 3.3 W 1.41*108 UCN/s/100W( in the He) 30cm

37. D2O Pre-Moderator Thickness Varying the D2O thickness: initial choice of 5 cm very good – no change in UCN prod

38. LD2 Moderator Thickness Varying the LD2 thickness: initial choice thin, increase to 19 cm provides increase in allowed beam power on target (reduced heat)