Cryo EDM at ILL Philip Harris Overview n

Cryo. EDM at ILL Philip Harris

Overview n n n Motivation, history and technique Cryo. EDM: Current status Upgrade Systematic errors Timeline Conclusion P. Harris ORNL, 11 Oct 2012

Cryo. EDM Collaboration C. Baker, S. Balashov, A. Cottle, V. Francis, P. Geltenbort, M. George, K. Green, M. van der Grinten, M. Hardiman, P. Harris, S. Henry, P. Iaydjiev, S. Ingleby, S. Ivanov, K. Katsika, A. Khazov, H. Kraus, A. Lynch, J. M. Pendlebury, M. Pipe, M. Raso-Barnett, D. Shiers, P. Smith, M. Tucker, I. Wardell, H. Yoshiki, D. Wark, D. van der Werf P. Harris ORNL, 11 Oct 2012

Electric Dipole Moments n n n EDMs are P, T odd Complementary study of CPv Constrains models of new physics E + P. Harris ORNL, 11 Oct 2012

History Factor 10 every 8 years on average P. Harris ORNL, 11 Oct 2012

Measurement principle Use NMR on ultracold neutrons in B, E fields. B 0 E <Sz> = + h/2 h (0) h ( ) <Sz> = - h/2 ( ) – ( ) = – 4 E d/ h with appropriate compensation for any changes in B during measurement period. P. Harris ORNL, 11 Oct 2012

Ramsey method of Separated Oscillating Fields 1. “Spin up” neutron. . . 2. Apply /2 spinflip pulse. . . 2 s 3. Free precession. . . 130 s 4. Second /2 spin-flip pulse 2 s P. Harris ORNL, 11 Oct 2012

UCN production in liquid helium Dispersion curve for free neutrons Landau-Feynman dispersion curve for 4 He excitations R. Golub and J. M. Pendlebury Phys. Lett. 53 A (1975), Phys. Lett. 62 A (1977) ln = 8. 9 Å; E = 1. 03 me. V n n n 1. 03 me. V (11 K) neutrons downscatter by emission of phonon in liquid helium at 0. 5 K Upscattering suppressed: Boltzmann factor e-E/k. T means not many 11 K phonons present Observed: C. A. Baker et al. , Phys. Lett. A 308 67 -74 (2002) P. Harris ORNL, 11 Oct 2012

Cryo. EDM overview Neutron beam input Cryogenic Ramsey chamber Transfer section P. Harris ORNL, 11 Oct 2012

Sensitivity • Achieved 60% polarisation in source, but must improve • Successfully applied 10 k. V/cm (same as previous expt); aiming for 20 -30 k. V/cm • Successfully produced, transported, stored UCN, but need to reduce losses • RT-edm: 130 s. So far we have 62 s cell storage time. P. Harris ORNL, 11 Oct 2012

Neutron numbers • • • Anticipated production rate 1. 4 /cc/s Aperture mask x 0. 44 Entrance window scattering x 0. 8 Beam attenuation x 0. 72 Source storage lifetime 91 s Incomplete source filling (200 s): x 0. 89 Gives expected density in source: 30/cc Source volume 10. 5 litres. P. Harris ORNL, 11 Oct 2012

Neutron numbers • Measurement has been somewhat indirect (neutrons taking convoluted paths to detectors) but it appears that we are currently down a factor of ~4: Under investigation • • Alignment/divergence issue? Spectrum affected by upstream instruments? P. Harris ORNL, 11 Oct 2012

Neutron numbers • • • Guides and valves not yet optimal. Ramsey chambers: first attempt yielded storage time 60 s. For next time, improved cleaning; also bakeout. What is limiting storage lifetimes in source and cells. . . ? P. Harris ORNL, 11 Oct 2012

Electric field n n See talk by M. Hardiman Latest feedthru installed designed for 30 k. V (6. 7 k. V/cm); it was run up to 45 k. V (10 k. V/cm). We know how to design feed up to ~80 k. V, and possibly up to ~150 k. V, but. . . will need mild pressurisation of He. P. Harris ORNL, 11 Oct 2012

Detectors n n Solid-state detectors developed for use in LHe Thin surface film of 6 Li. F: n + 6 Li a + 3 H; 82% efficient Fe layer for spin analysis Currently, a peak hidden under g background n n pulse-shape discrimination Now moving to detector with 10 x area, to cover entire guide C. A. Baker et al. , NIM A 487 511520 (2002) P. Harris ORNL, 11 Oct 2012

Detection of polarised UCN n n Observed ~60% polarised downscattered neutrons Should be able to improve on this - Upcoming measurement of source polarisation P. Harris ORNL, 11 Oct 2012

T 1 n n n Longitudinal polarisation T 1 is fairly straightforward to hold: field mustn’t change too fast for precession to follow Issues last time with superconducting material around source/guide region. . . Need to watch also sensitive area at entrance to shields, where field is low P. Harris ORNL, 11 Oct 2012

T 2 n n Transverse polarisation T 2 is more delicate. Depends largely on variation of Bz within trap volume – causes dephasing: goes as We are aiming for ~1 n. T across the bottle Currently in commissioning phase. SS plates at end of superfluid containment vessel (SCV) distort field. Modelling suggests that with correction coils we can reach T 2 ~ 30 s with current SCV. Working on non-magnetic SCV. P. Harris ORNL, 11 Oct 2012

Sensitivity summary: Current n Room-temperature expt final sensitivity ~2 E-25 ecm/day n n n We can come within factor 4 -5 of this in 2013 by n n n Took 12 years of incremental developments from known technology Systematics limited (geometric phase effect) increasing detector area x 10: technology now proved refurbishing damaged detector-valve: in hand applying ~70 k. V (previously ~40 k. V): should be straightforward opening beam aperture from 43 to 50 mm: depends on radiation levels retaining polarisation: superconducting material has been removed There may be additional improvements beyond this n n n a peak above background (detector improvement) Polarisation to 60% or more (improved guide field) Increasing cell storage lifetime (insulator bakeout) (we will achieve these by 2014) P. Harris ORNL, 11 Oct 2012

Shutdown, move to new beamline Mid-2013: Have to vacate current location. ILL to shut down for a year; we move to new dedicated beamline. n New beam 4 x more intense; and dedicated n Due to become operational mid-2014 n Beam must then be characterised (9 A flux, divergence, stability, polarisation) n We will then have access to the area (late 2014) to move our apparatus into it. P. Harris ORNL, 11 Oct 2012

Upgrade 2013 -15 n n Not yet fully costed Major upgrade to experiment: n Cryogenics design changes: n n Upgrade to back-to-back cells (or possibly 4 cells) n n n B-field stability improves x 500, for systematics Construction of non-magnetic SCV n n n 2 x neutrons Cancellation of some systematic effects Installation of inner superconducting magnetic shield n n Pressurise the liquid helium: increase E field x 2 -3 Improves depolarisation: better T 2 Overcome geometric-phase systematic error Net result: n n Order of magnitude improvement in sensitivity P. Harris Commensurate improvement in systematics. ORNL, 11 Oct 2012

Systematics: General n Systematics minimised by highly symmetric data taking: n n B and E field reversals Alternating either side and above/below middle of central Ramsey fringe Upgrade: opposite E in adjacent cells Possibly also neutron magnetometers in adjacent (outer) cells, for 4 -cell system P. Harris ORNL, 11 Oct 2012

Systematics: B field fluctuations n n At present, Pb shield too short: flux lines clip coil end, inducing current in whole coil Introduces common-mode noise, limiting sensitivity to 1 E-27 e. cm Figure: JMP P. Harris ORNL, 11 Oct 2012

Systematics: B field fluctuations n We plan to add an inner superconducting shield. Scale model work in lab (MH) suggests that this can bring increased shielding factor ~500. ISS Figure: JMP P. Harris ORNL, 11 Oct 2012

Systematics: B field fluctuations n Can also (or instead) add Pb end caps, calculated to give factor ~250 improvement Figure: JMP P. Harris ORNL, 11 Oct 2012

Systematics: Geometric phase Bnet Br Bnet Bv Bottle (top view) Bv . . . so particle sees additional rotating field Frequency shift E Looks like an EDM, but scales with d. B/dz Bnet Bv Br Bnet P. Harris ORNL, 11 Oct 2012

Systematics: Geometric phase n n n For neutrons, Scales as 1/B 2; increase B 5 x to obtain factor 25 protection <1 n. T/m 3 E-29 e. cm P. Harris ORNL, 11 Oct 2012

Systematics: E x v n Translational: n n n Vibrations may warm UCN, cause CM to rise ~1 mm in 300 s 3 E-6 m/s If E, B misaligned 0. 05 rad. , gives 2 E-29 e. cm Rotational: n n Net rotation damped quickly (~1 s): matt walls Delay before NMR pulses allows rotation to die away Neutrons enter E-field cells centred horizontally; no preferred rotation Below 1 E-29 e. cm P. Harris ORNL, 11 Oct 2012

Systematics: 2 nd order E x v n n n Perpendicular component, adds in quadrature to B. Prop. to E 2; gives signal if E reversal is asymmetric Cancellations (back-to-back cells; B reversals) reduce effect to < 3 E-29 e. cm P. Harris ORNL, 11 Oct 2012

Systematics: m metal hysterisis n n n Room-temp expt: Pickup in B coil from E field reversals; return flux causes hysterisis in m metal Coil here is SC, not power-supply driven Inner shield is SC also Small effect from trim coils, enhanced by any misalignments Net estimate < 1 E-30 e. cm P. Harris ORNL, 11 Oct 2012

Systematics: E induced cell movement n n n n Electrostatic forces of order 1 N; E 2 Asymmetry perhaps ~1% of this Radial gradients of order 3 n. T/m Must keep radial displacement on E reversal symmetric to ~ 0. 01 mm Cancellation with double cell Symmetric voltages to ~2% Net effect < 1 E-28 e. cm P. Harris ORNL, 11 Oct 2012

Systematics: Leakage currents n n n Azimuthal current components generate axial contributions to B Cancellation in adjacent cells Conservative estimate: 1 n. A 5 E-29 e. cm In reality LHe should keep currents much below this? New source of current: ionisation from UCN decay electrons (10 -100 p. A? , but preferentially axial) P. Harris ORNL, 11 Oct 2012

Systematics: HV supply contamination n n HV circuit isolated as far as possible to minimise earth contamination. Feedback line far from cells. Separate computer control. 10 k. Hz ripple on HV line can “pull” resonant freq. Estimate 1 E-30 e. cm Likewise 50 Hz ripple: estimate ~1 E-29 e. cm Directly generated AC B fields negligible P. Harris ORNL, 11 Oct 2012

Systematics: Summary Effect B fluctuations Geometric phase Exv translational Exv rotational Exv 2 nd order m metal hysterisis E-induced cell movement Leakage currents HV line contamination Size (e. cm) 1 x 10 -30 3 x 10 -29 2 x 10 -29 1 x 10 -29 3 x 10 -29 1 x 10 -30 1 x 10 -28 5 x 10 -29 1 x 10 -29 P. Harris ORNL, 11 Oct 2012

Sensitivity timeline Date Item factor 2002 RT-edm 2010 Cryo. EDM commission ecm/year Comment 1. 7 E-26 Baseline 1. 7 E-24 2012 Large-area detector 3. 5 2012 HV to 70 k. V 1. 6 2012 Repair detector valve 1. 3 2012 Polarisation 60% 1. 5 2012 Aperture to 50 mm 1. 2 2013 Ramsey time to 60 s 1. 8 2013 See alpha peak 1. 4 2014 New beam 2. 0 5. 5 E-26 Quite likely by 2012, but we do not count on it by then 2. 7 E-26 ILL produced this estimate 2. 2 1. 2 E-26 1. 5 8. 3 E-27 1. 3 6. 4 E-27 1. 9 3. 3 E-27 Requires pressurisation. Lab tests show this is realistic. 2. 3 E-27 Guaranteed part of upgrade 2014 Recover missing input flux? 2014 Improve cell storage lifetime to 100 s 2014 Match aperture to beam 2015 HV to 135 k. V 4. 9 E-25 Proven 3. 1 E-25 OK to 50 k. V, lab tests suggest should work at 70 k. V 2. 5 E-25 Repair – should be fine 1. 7 E-25 Seen in source. Should transfer ok to cells. 1. 4 E-25 Will increase radiation levels slightly, but should be ok 7. 7 E-26 Requires change of SCV baseplates Depends on geometry match to new beam. Not guaranteed, but haven't yet tried most obvious solutions (e. g. bakeout), so improvement likely Likely 2015 Four-cell system 1. 4 2015 Polarisation to 90% 1. 5 2013 -15 Inner supercond. shield 2013 -15 Cryogenics Included in upgrade 2013 -15 Non-magnetic SCV Included in upgrade 1. 6 E-27 No known reason why not Lab tests on scale model shows factor 500 P. Harris ORNL, 11 Oct 2012

New collaborators n n Swansea is interested in joining us soon. There is still plenty of room for new collaborators! Grants panel would like to see us recruiting from overseas. P. Harris ORNL, 11 Oct 2012

Conclusions • • • Cryo. EDM is now commissioning New beamline 2014 Aim to start running ~2015 No “showstoppers” evident Goal (for now) ~3 E-27 e. cm There’s still room on board! P. Harris ORNL, 11 Oct 2012