Probing the electron edm with cold molecules

E.A. Hinds

Columbus Ohio, 23 June, 2010

Centre for Cold MatterImperial College London

+ +

++

polarisable vacuum with increasinglyrich structure at shorter distances:

(anti)leptons, (anti)quarks, Higgs (standard model)beyond that: supersymmetric particles ………?

How an electron gets structure

-

point electron

electronspin

+-edm

Electric dipole moment (EDM)

Left -Right

MSSM f ~ /a p

Multi Higgs

MSSM

f ~ 1

10-24

10-22

10-26

10-28

10-30

10-32

10-34

10-36

eEDM (e.cm)

Our experiment on YbF molecules

explores this region

Standard Model

de < 1.6 x 10-27 e.cm

Commins (2002)

Excluded region (Tl atomic beam)

10-18 Deb.

The magnetic moment problem

Suppose de = 5 x 10-28 e.cm(the region we explore)

In a field of 10kV/cm de.E _ 1 nHz ~

When does mB.B equal this ?

B _ 1 fG~

It seems impossible to control B at this levelespecially when applying a large E field

A clever solution

E

electric field

de

amplification

atom or molecule containing electron

(Sandars)

For more details, see E. A. H. Physica Scripta T70, 34 (1997)

Interaction energy

-de E•

F PPolarization factor

Structure-dependent

relativistic factor ~ 10 (Z/80)3 GV/cm

Our experiment uses a molecule – YbF

EDM interaction energy is a million times larger (mHz)

mHz energy now “only” requires nG stray field control

0

5

10

15

20

0 10 20 30

Applied field E (kV/cm)

Eff

ecti

ve f

ield

E

(G

V/c

m)

Amplification in YbF

16 GV/cm

| -1 > | +1 >

| 0 >

The lowest two levels of YbF

Goal: measure the splitting 2dehE to ~1mHz

F=1

F=0

E

-dehE

+dehE

+-

+-

X2S+ (N = 0,v = 0)

170 MHz

Flu

ores

cenc

e

Time of flight (ms)

Time-of-flight profile

2.1 2.3 2.5 2.7 2.9 3.1

How it is done

Pulsed YbF beam

PumpA-X Q(0) F=1

ProbeA-X Q(0) F=1

PMT

3K beam

F=1

F=0

rf pulseB HV+

HV-

Scanning the B-field

-200 -100 100 200 B (nT)

Ch 15 Cold Molecules, eds. Krems, Stwalley and Friedrich, (CRC Press 2009)

Measuring the edm

Applied magnetic field

Det

ecto

r co

un

t ra

te

B0

-EE

Interferometer phase f = 2(mB + dehE)t/hInterferometer phase f = 2(mB)t/h-

de

-B0

Modulate everything

• Generalisation of phase-sensitive

detection

• Switch periodically on short timescale

but randomly on long timescale.

• Sequence chosen to minimize noise.

• Measure all 512 correlations.

±E±B

±B

±rf2f±rf1f

±rf2a±rf1a

±laser f±rf

spin interferometer signal

9 switches:

512 possible correlations

7 Non-zero correlations

• 504 correlations should be zero and they are!

correlation mean s mean/s fringe slope calibration beam intensity f-switch changes rf amplitude E drift

E asymmetry E asymmetry inexact π pulse

• We don’t look at the last one (EDM).

** Don’t look at the mean edm **

• We don’t know what result to expect.

• Still, to avoid inadvert bias we hide the mean edm.

• A random blind offset is added that only the computer knows.

• More important than you might think.– e.g. Jeng, Am. J. Phys. 74 (7), 2006.

+20

-20

mean

ED

M (

10

-26 e

.cm

)7381 measurements (~6 min each) at 11.6

kV/cm.

+40

-40

mean

ED

M (

10

-26 e

.cm

)

1440 measurements (~6 min each) at 3.3 kV/cm

Initial edm results and std. error

-147± 5 ×10-28 e.cm

mean-10 10 (10-26 e.cm)

edm distribution

11.6 kV/cmBlind results

includes blind offset

mean-20 20 (10-26 e.cm)

edm distribution

3.3 kV/cm-144±22 ×10-28 e.cm

with same offset

Two systematic error corrections

• Magnetic field noise B fluctuations synchronous with E reversal

produce false EDM.

We measure and correct: (-2.2 ± 1.6) ×10-28 e.cm.

• Electric field asymmetry “Reversal” changes magnitude of E (slightly)

This induces a false EDM We measure and correct: (+0.4 ± 0.4) ×10-28

e.cm.

Negligible systematics

• HV charging current: can magnetize shields

< 0.3 ×10-28 e.cm B that reverses with E

•B. E. Sauer, Dhiren Kara, J. J. Hudson, M. R. Tarbutt, E. A. Hinds “A robust floating nanoammeter”. Rev. Sci. Instr. 79, 126102 (2008).

• HV leakage current: makes B that reverses with E i < 2 nA < 0.8 ×10-28 e.cm

“A robust floating nanoammeter” Rev. Sci. Instr. 79, 126102 (2008).Faraday Discussions, 142, 37 (2009)

• Changing E direction along beamline

< 0.03 ×10-28 e.cm

geometric phase

F=0

F=1

sensitive to E (Stark shift)

Mapping the E field

3K beamrf pulse

B

170.9170.8170.7

Frequency (MHz)

020

4060

80100

0

100

200

300

400

500

600

Frequency

(M

Hz)

Molecule arrival time

PRA 76 033410 (2007)

Electric field variation along plates

Position along plates (m)

-0.75%

0

+0.75%

0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4

gap is constantto ± 90 μm !

One last check...Identical run, but with E reversal

disabled

9000 edm measurements (3 months)

false edm < 4 ×10-28 e.cm

No systematics observed.

80

60

40

200

-20ED

M (

10

-28 e

.cm

) NO! EDM changes across the

pulse

One last, last check!

PMT

Do front and back of pulsemeasure the same EDM?

What’s going on ???The edm should not care where it is measured

(front or back of pulse)

The edm should not care how fast molecules are (570 or 590 m/s)

We find that this “chirp” is proportional to

(i) detuning of 1st rf pulse (ii) asymmetry of E reversal

Can be corrected, but we’d like a detailed model

Current status

• Close to a measurement at the level of 4×10-28

e.cm

• Systematics seem under control at this level, but we’d like to understand the chirp.

• Expect to unveil the result later this summer

New cryogenic buffer gas source of YbF

YbF beam

YAGablationlaser

3K He gas cell

Yb+AlF3

target

Spectroscopy inside: New J.Phys. 11, 123026 (2009)

the real thing

upper LIFdetector

lower LIFdetector

YbF beam

170MHz

Q(0) hyperfine splitting Lower PMTUpper PMT6

5

4

3

2

1

0

mol

ecul

es /s

r/pu

lse

(10

10) 170

Beam intensity

174Q(0)

170 MHz hfs

fluorescence spectrum

5 1010 molecules/sr/pulse in N=0 F=1 25 higher intensity than best previous!

relative frequency (MHz)

Beam velocity

LIF at lower PMT z = 23mm

LIF at upper PMT z = 132mm

0 1 2 3 4 5 6time (ms)

v = 150 m/stuneable byHe pressure100-200 m/s

700 ms

700 ms

EDM prospects with new source

25 more molecules

=> 20 better signal:noise ratio

Also, ideal source fordecelerating or trapping many species

and for laser cooling CaF, BaF, SrF to K.

4 longer interaction time

=> access to low 10-29 e.cm range!

Acknowledgements

Royal Society EPSRC Euroquam CHIMONO

Ben SauerJony Hudson

Mike Tarbutt

Dhiren Kara (Ph..)

Joe Smallman (P...) EDM

Rich Hendricks

Sarah Skoff (Ph..)

Nick Bulleid (P...)

buffer gassource