CryoEDM The search for the electric dipole moment of the neutron

CryoEDM The search for the electric dipole moment of the neutron RAL, STFC Sussex University Oxford University ILL Kure University Swansea University Dr Christine Clarke1 4/19/2011

CryoEDM Motivation (CPV) Method (Ramsey Resonance technique) Overview of CryoEDM apparatus CryoEDM commissioning results from 2010 Experimental challenge: the magnetic field Depolarization Measurement Summary 2

Neutron Physics Where did all this matter come from? Strong CP violation: θ in QCD Lagrangian. This could certainly be enough... but ruled out as it predicts large EDMs. Electroweak CP violation: well-studied in K and B meson sectors... but much too small. Massive neutrinos: Majorana leads to CP violating phases. Beyond the SM: violating phases in two-higgs doublets, supersymmetric models. 4

CPV and nedms If there is an EDM, it is proportional spin (d S). When placed in electric field E... There is a term in Hamiltonian: -d S.E An EDM violates P and T symmetry (below). S E The neutron has a magnetic dipole moment μ along direction of spin S. The EDM is a measure of CP violation within and beyond the SM providing an excellent model independent, background free measure. S μ μ + -- P E S S μ + E Eμ + - T + - 5

EDM predictions SM (weak sector CP violation): 10-31-10-33 e cm. Extensions to SM (additional Higgs, right handed currents, SUSY): 10-25 10-28 e cm. Current limit: dn < 2.9 x 10-26 e cm (QCD sector CP violation and some SUSY constrained therefore need to fine tune to suppress EDMs.). 6

Precession in E and B fields Dipole feels a torque in a field. Torque causes the spin vector to precess around the field direction with Larmor frequency ω. For a particle with both a magnetic dipole and electric dipole moment in parallel B and E fields: Antiparallel B and E fields: ħω0 = 2μB + 2dE = ħ(ωb+ ωe) ħω0 = 2μB - 2dE = ħ(ωb- ωe) Therefore measure change in precession frequency produced by reversing E field (2ωE) to find EDM. 8

Ramsey Resonance To find such a small change in frequency (~10-7 rad s-1), we need to use the Ramsey Resonance method. 1. Fill storage cell with polarised neutrons aligned with holding (B) field and apply E field. 2. Short RF pulse capable of rotating spin vector by π/2 3. Precession period 4. Short RF pulse (π/2 rotation) Then... Empty cell and count the number of neutrons with spin-down and spin-up polarity. Flip relative direction of E and B fields and repeat. 9

Ramsey Resonance cont. If precession frequency = applied RF Resonance. Small differences in frequency add up during free precession as precession and RF source get out of step. Results in fringes and a sensitivity of a small fraction of Larmor period. Working points (marked x) are on steepest part of slope to allow highly sensitive measurements of 2ωE. 10

Improving on current limit with cyrogenic experiment Room Temperature experiment limited by statistics. σd= 2α ET N Better polarisation Higher E field Longer NMR coherence time More neutrons 12

CryoEDM Experiment - Overview <0.9 K superfluid helium stores neutrons for measurement. Decrease uncertainty in measurement of d by: Increasing flux (using superthermal UCN source). Increasing storage time (need ultra-cold neutrons ~10-7 ev). Increasing strength of E field (higher fields in liquid helium). (More efficient detectors.) Overall, approx. 100 times more sensitive 13 Target: 10-28 e cm precision

Apparatus Cold Neutron Beam (input) Polarizer UCN Source Detectors Neutron Guides Resonance Cell HV Probe Magnetic Shielding 14

Location Institut Laue Langevin (Grenoble, France) 15

Cold Neutron Source The Institut LaueLangevin (ILL) in Grenoble operates the world s strongest source for thermal, cold and ultra-cold neutrons Cold neutrons produced at the 58 MW reactor are moderated in liquid deuterium at the Horizontal Cold Source Neutron guide H53 couples to HCS CryoEDM situated at the end of H53 beamline. 16

The Polarizer Polarizing Guide (FeSi coated glass) Permanent magnets provide 300G field along polarizing surface. P = 93 +- 2 % Flux 3 x 107 cm-2 s-1 Å-1 at wavelength 9 Å. 17

Production of Ultra Cold Neutrons Downscattering in ~0.65 K superfluid helium-4 Neutrons downscatter by emission of phonon Upscattering suppressed: Boltzmann factor e-e/kt means not many 11 K phonons present R. Golub and J.M. Pendlebury Phys. Lett. 53A (1975), Phys. Lett. 62A (1977) 18

The guides and storage cells Be-coated Copper guides for neutron transport BeO storage/resonance cells 2-cell design allows simultaneous measurement with E field and without E field (to identify systematic effects). HV Cell neutral cell 20

Apparatus- Horizontal Shields Resonance cells are within low magnetic environment 3 layers mu metal Lead (superconducting) shield Plus superconducting solenoid A lot of helium is required! 21

Detectors Destructive detection Solid state detectors (also in liquid helium) Silicon devices with 6LiF coating n + 6Li3 4He2 + 3H1 Detect charged product. 2 detectors have a Fe layer that reflects one spin polarisation. At end of data cycle, UCN directed to detectors. UCN polarisation is measured. 22

Commissioning (November 2010) Detectors at Tower 1 and in transfer section. Initially operating first valve only- letting neutrons into tower 1. Run Multichannel Analysers and Scalers. 24

Results Commissioning run 1353 First Ultra Cold Neutrons! Counts on Tower 1 detector Counts on Transfer section detector. Energy spectrum shows triton peak. Alpha peak lost in noise of gammas and detector. Counts on Transfer Section detector Energy Spectrum 4 He2 3 H1 25

Results Commissioning Run 1383 First 50s: Fill source. Valve in transfer section is vertical closing off the source, allowing build-up. Next 10s: Move valve to horizontal position allowing UCN to fill all volumes apart from detector volume. Then valve back to vertical, emptying volume between valve and cells through detectors. Counts on Transfer Section detector Source Cells Valve in intermediate position Leaking Detector Source Cells Detector 26

Polarisation Some detectors have Fe layer- measure one polarisation of neutrons. Can measure polarisation. Correct for detector efficiencies and backgrounds. Neutron count rate Polarisation of neutrons with time Open detector (no Fe) Detector with Fe layer 27

Conclusions from 2010 run Base temperature 530-620 mk Storage lifetime of tsource ~ 85 s, tneutral cell ~ 70 s, tht cell ~ 75 s. The polarisation of the UCN is rather low: 50% down to 30%. A Rabi resonance has not been found: the visibility of fringes is reduced by the low polarisation and overall low UCN numbers. Total shielding factor of ~5500 The temporal stability of the magnetic field is of the order of a nt. A static electric field with +40 kv on the HT electrode has been kept over the storage cell. 29

UCN Numbers and Storage Time Many factors contribute to the low UCN numbers but most can be fixed. Smaller aperture for radiation safety Scattering with Be window into our cryostat Further attenuation in source volume Detectors not 100% efficient and do not cover much area. Storage time reduced due to cracks where neutrons get lost. Contamination leads to lost neutrons Neutrons lost due to fall from source to transfer section (energetic neutrons can escape). 31

Polarisation Polarisation α = (Nup Ndown) / (Nup + Ndown) α = 2(Nup Ntotal) / Ntotal Usually α ~0.4 (often 0). Spin relaxation in two locations: region of polariser. entrance to Horizontal Shield. Hard to maintain polarisation during precession- need field uniform to 0.3nT/m. 32

Improvements Increased shielding allows larger aperture. Be coated Al window has less scattering. Align parts better (there are 1 mm gaps currently). Clean and bake-out parts. Need to increase HV (applied 6.7kV/cm... 30kV/cm looks feasible with pressurised helium). Better detectors- cover more area, greater efficiency Boron-doped scintillator? Remove all superconductors. 33

Importance of Storage Time Need to increase storage time: Increases number of neutrons for counting Increases Ramsey precession time 34

Magnetic Environment We look for a change in the precession frequency of polarised neutrons with a reverse of E field. But variations in B field that correlate to the direction of E will produce a false EDM. Shielding in place (superconducting lead shield and 3 layers mu metal). Superconducting solenoid plus trim coils for homogeneous B field. Also, neutrons will lose polarisation if not handled carefully. 37

Depolarisation of Neutrons Neutrons must travel from the source to the Ramsey Cells We have Guide Fields to give high fields adiabatic transport of the neutrons. Fractional rate at which the precession frequency changes [1/B x (db/dt)] has to be much less than the precession frequency 38

Superconductors Meissner effect- currents on surface of superconductor mean zero field within bulk... But the currents also produce a field beyond the surface. Cannot use indium seals. Restriction on superconducting loops for SQUIDs. Also, cannot have any components made from superconducting materials... In particular, the Superfluid Containment Vessel. 39

Susceptometer The titanium alloy used in the experiment went superconducting ~6K 40

SCV Conclusion This result contributed to the decision to discard the SCV (which leaked) and prompted the investigation of other materials. New material investigation: CuBe? G10? PVC? Gold wire seals? PTFE? Must be non-magnetic. Induction effects In the interim, the collaboration are using a Stainless Steel SCV. Mildly magnetic and not obviously possible... 41

Field within the Resonance Cell Within the Resonance Cell, neutrons make many passes during the long storage time. Changing magnetic fields within Resonance Cell change spin alignment. Length of free precession period is set by how many passes neutrons can make before losing polarisation. With a stainless steel SCV, can get a Rabi Resonance Possibly get a short Ramsey Resonance. For the free precession time, the neutrons were depolarized in 10s. Far off the 130s we aim for. 42

SQUIDs Goal: need to know the B field to within 0.1 pt over neutron storage time. We use DC SQUIDs situated ~1.5m from Resonance cell, coupled to superconducting pick up loops in the superfluid containment vessel. 44

SQUID operation SQUIDs measure relative changes in B (not absolute) find absolute B by measuring Larmor frequency and then SQUIDs track changes in field strength. (Has not been tested yet). Untracked resets and flux jumps will require us to go back to the Larmor frequency to re-ascertain the absolute field Need continuous digitisation and excellent software to deal with this so our working point is not lost. (Completed). Multiple loops to extrapolate field in Ramsey cell. (Studies ongoing) EMI needs to be controlled for stable operation of SQUIDs. 49

SQUID signals Tracks magnetic fields (agreement with fluxgate magnetometers but with better resolution) 50

SQUID Challenges The desire for high resolution has left us very susceptible to EMI issues. Cell valve motors caused additional noise in SQUIDs. Oscillating magnetic field makes SQUIDs lose lock. Many periods of SQUIDs not tracking magnetic fields 51 1 cycle 1 cycle

CryoEDM - Conclusion Current status: Commissioning apparatus at H53 beam position ILL. Currently working on an improved valve design, improved SCV, improved shielding, improved detectors, four-cell storage cells... This allows for an EDM measurement with 10-27 e cm sensitivity (order of magnitude improvement) with a couple of years of date. Likely stop data taking in 2013 and in 2014 move apparatus to new location with 20x higher intensity 9Å beam. Achieve sensitivity of 10-27 e cm quickly and then 10-28 e cm. We will therefore... Find an EDM > 10-28 e cm or... Give a new upper limit to the neutron EDM (ruling out SUSY/ indicating the existence of a mechanism for EDM suppression). 53