Search for an Electric Dipole Moment of 199 Hg Blayne Heckel

University of Washington

Graduate students: Jennie Chen, Brent Graner Scientific glassblower: Erik Lindahl Support: NSF and DOE Low Energy Nuclear Physics

History of

199

Hg EDM results

1,00E-25 1,00E-26 1,00E-27 1,00E-28 1,00E-29 1,00E-30 Lamoreaux Jacobs Klipstein Fortson 1987 1993 1995 Griffith Swallows Romalis Loftus Fortson 2001 2009 2014 Current sensitivity

B

E E

4-Cell,

199

Hg Magnetometer

EDM sensitive frequency combination w OT w MT w MB w

c

   (  8 3  3

B



z

3 

z

3 )  4

dE

 w

c

  w

m

 1 3  w

o

Cancels up to 2 nd order gradient noise w OB EDM insensitive channels: ω OT ω OB and ( ω OT + ω OB ) – (ω MT monitor for E field correlations odd and even in z, respectively. + ω MB )

Cell Holding Vessel Cell Holding Vessel

Schematic Overview

Pump Probe

Vapor Cells

Improvements to the EDM Experiment  Vapor Cell development: Stable spin lifetimes of 800, 600, 500, and 270 sec Problem of disappearance of Hg within the cells solved  Reduced magnetic field noise – less conducting materials near the cells  Light induced noise and systematic effects eliminated by precession in the dark  Reduced high voltage leakage currents: Measure separately the currents across the cell and to the ground plane  Commercial uv laser provides better stability  We are working on improving the beam pointing stability

Precession in the dark A Optical rotation angle of 24 0 B Advantages : Longer spin coherence times Light induced noise and systematic errors eliminated Insensitive to common mode B field drift

Frequency difference extraction Because we are only interested in cell pair frequency differences during the light-off period, we need only the phase difference between cells at the end of the A period and start of the B period:  w

Dark

   (

B

)

T B

   (

A

) 

T A

We find Δ ω by multiplying the signals from the 2 cells.

From N

2 (

t n

) 

S

(

t n S

2 ) (

t n

 )

Ae



t n



S

(

t n

/   sin(  / w 2

t n

w )   

S

) (

t n and

  / 2 w ) 

A

2

e

 2

t n

/ 

get C

(

t n

)  [

S

  (

t

(

t n n

)   

S

/ (

t

2

n

w ) ) /

N



S

( 

t n

(

t

)

n

  sin(  / w

t

2 w

n

)]  /  ) 2 

and

cos( w

t n

  )

Finally S



MT

(

t n

)  

C MB

(

t n

)  

C MT

(

t n

) 

S



MB

(

t n

)  sin(  w

t n

   )

Filtered Data 10 8 6 4 2 Photo-Diode (V) 2.0

1.5

1.0

0.5

0 50 100 150 200 250 Time (s) Δω MT-MB (mrad)  w

Dark

   (

B

)

T B

   (

A

) 

T A

T A T B

System Performance Average angular frequency relative to first scan Run Number This corresponds to a drift of ~1 µGauss/day

Δω (MT-MB) Middle cell angular frequency difference 0.3 nG/day Δω (OT-OB) Outer cell angular frequency difference

Δω (OT-MT) + Δω (OB-MB) Quadratic field drift channel Δω C EDM sensitive frequency combination 80 pG/day

One day EDM signal ω edm ( n ) = (-1) n [ Δω C ( n-1 ) - 2 Δω C ( n ) + Δω C ( n+1 ) ] /4 ω edm = xx ± 2.0 × 10 -9 rad/s, a factor of 4 improvement over 2009 (3.5 × 10 -29 e-cm/day for 10kV runs)

Data Sequences EDM data is taken in ``sequences’’. Each sequence comprises: • A defined set of cell orientations and ordering in the vessel • Equal number of day long runs at 6kV and 10kV • Equal number of runs with normal and reversed magnetic fields • Equal number of runs with fast and slow high voltage ramp rates • Typically 16 -20 runs total We have completed 7 data sequences and have 5 to go for a complete EDM data set.

For seq. 1-6: at 10kV, d Hg at 6kV, d Hg = -(13.4 ± 5.6) × 10 -30 = -(15.6 ± 9.3) × 10 -30 e-cm e-cm combined, d Hg = -(14.0 ± 4.8) × 10 -30 e-cm (blind offset in place)

Sequence 1-6 χ 2 = 1.05

χ 2 = 1.02

Sequence 1-6

EDM data by Sequence B field up B field down -10 -20 -30 -40 30 20 10 0 10 0 -10 -20 -30 -40 χ 2 30 20 10 kV χ 2 Seq.1

Seq.2

Seq.3

Seq.4

Seq.5 Seq.6

6 kV χ = 1.22

2 = 0.98

Seq.1

Seq.2

Seq.3

Seq.4 Seq.5

Seq.6

Systematic Uncertainties 2009 Error Budget -- Scale with frequency precision X (no sparks observed) X (precession in the dark)

Leakage Currents

We now measure separately the currents flowing down the cell walls and through the dry air.

10 kV I Gas = 0.4 pA (2009) 0 Volts I Cell = 0.08 pA (2009) 2014: With tin oxide coated ground planes (rather than gold), we see 10 times less leakage current -- no photo-electrons: I gas = 0.04 pA (2014), I cell = ?

Issues to address

• Better understanding of the long time constants associated with the leakage currents • Identification of the dominant source of excess noise (simulated data with Gaussian photon shot noise added results in 40% smaller EDM values) • Identification of the source for occasional runs with 2-3 sigma correlations between the outer cell frequency difference and high voltage

Summary • We anticipate a 199 Hg EDM result with a factor of roughly 4 improvement in statistical precision by the end of 2014.

• The data, so far, looks reasonable but there remains work to do concerning systematic errors.

• We are constructing shorter vapor cells to allow us to increase the applied electric field in our current EDM apparatus.

Laser System

SDL MOPA: 500 mW at 1015 nm 1 st Doubler: 130 mW at 507 nm 2 nd Doubler: 6 mW at 254 nm

Transverse Pumping / Optical Rotation

Pump Phase 254 nm s + Probe Phase 254 nm Linear B B w L Linear Polarizer Detector

Transverse Pumping / Optical Rotation

Optical Rotation Angle Probe Absorption Probe Pump

Typical 24 Hour Run

Final Dataset and Statistical Error

d( 199 Hg) = (0.49 ± 1.29

stat )x10 -29 e cm 0.1 nHz (~ 7.5 ppt)

Systematic Errors and Tests for Systematic Effects No Statistically Significant Dependence on: • • • The Vapor Cells or Electrodes (or their orientation) The DAQ Channel Ordering The Vessels

Systematic Error Budget 99% of Total Error Statistical Error 12.90

Bounds on CP Violating Parameters d( 199 Hg) = (0.49 ± 1.29

stat ± 0.76

sys ) x 10 -29 e cm | d( 199 Hg) | < 3.1 x 10 -29 e cm (95% CL) Quark Chromo EDMs Proton EDM Semi-Leptonic Interactions: QCD Phase Neutron EDM Electron EDM

Confidence Levels:

199 Hg (95%), 205 TI (90%), TIF (95%)

Current Status

Six-fold Reduction of Noise • • • More uniform and stable magnetic field Detection of both polarization states of transmitted light ``In the Dark’’ elimination of uv light induced noise and systematic error Old Analysis ``In the Dark’’

Current Status

Improved Hg Vapor Cells: so far, natural Hg test cells • • Hydroxy bonded rather than glued – should last forever 600 sec spin lifetime (distilled wax + smaller magnetic field gradients) Remaining Tasks: • • Construct enriched Hg vapor cells Acquire a new uv laser source

Summary

Our 2009 Result led to a New Limit on the EDM of 199 Hg

| d( 199 Hg) | < 3.1 x 10 -29 e cm (95% CL) • • Factor of 7 Reduction in Previous Upper Limit Improved Bounds on CP Violating Parameters

Upgrading the Current Apparatus

• • Expect Factor of 5 improvement in Experimental Sensitivity Expect to begin data collection later this year

Among his many accomplishments, Norman Ramsey founded the research field of EDM measurements and developed many of the techniques needed to do such precise measurements. He will always be an inspiration to us.

Reasons to Expect more T (CP) Violation

• • •

The observed matter-antimatter asymmetry:

CP violation in the SM is too small to account for Baryogenesis.

The Strong CP problem:

Why is  QCD so small?

The Standard Model is incomplete:

Extensions to the SM, such as SUSY, introduce new phases that lead to new sources of CP violation often 10 6 times larger than the SM for EDMs.

From the 199 Hg EDM to Models for CP Violation 199 Hg Atomic EDM

Atomic Physics

199 Hg Schiff Moment

Nuclear Physics

CP-Violating Pion-Nucleon Coupling

QCD

CP-Violating QCD Term, Quark Chromo-EDMs

SUSY, etc …

Model-Dependent CP-Violating Parameters Semileptonic Interactions:

C S C P C T

Hyperfine Coupling:

d(e)

Contributions to S from p, n EDMs

d(p) d(n) Naturalness Parameters

Constrained MSSM

Measuring an EDM via Larmor Precession w

L B E



d