nEDM @ PSI

Lepton Moments 2014Craigville07/22/2014P. Schmidt-Wellenburg

nEDM @ PSI

Searching for the electric dipole moment

of the neutron

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

2

• 14 Institutions• Eight countries / 14 institutions• 48 Members• 11 PhD students

The collaboration

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

3

PSI in Switzerland

UCN source

& E

DM

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

4

Outline

Introduction• How do we search for the nEDM?

Status at PSI• The UCN source• Statistical sensitivity

Neutron/mercury gyromagnetic ratio

• Magnetometry• Preliminary Results

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

5

The measurement technique

Measure the difference of precession frequencies in parallel/anti-parallel fields:

BBμEEdΔ nn 22

for dn<10-

26

Δω < 60 nHzωL≈30Hz for B0=1μT

Best nedm limit: 2.9×10-

26ecmBaker et al., PRL97(2006)

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

6

st 2

nRF

The Ramsey technique

Free precessionat ω L

Apply /2 spinflip pulse...

Spin “down” neutron...

Second /2 spinflip pulse.

B0↑

B0↑ + Brf

B0↑

B0↑ + Brf

tT

st 2

Ramsey resonance curve

Sensitivity:

a Visibility of resonanceT Time of free precessionN Number of neutronsE Electric field strength

𝜎 ( 𝑓 )= ℏ2𝛼𝑇𝐸 √𝑁

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

7

Ultracold neutrons (UCN)

n

1dσ

T N Storable

neutrons(UCN)

E. Fermi & W.H. Zinn(1946),Y. B. Zeldovich, Sov. Phys. JETP (1959)389

Storage properties arematerial dependent

350 neV ↔ 8 m/s ↔ 500 Å ↔ 3 mK Magnetic∼60 neV/T

Magnetic∼60 neV/T

Gravity102

neV/m

Gravity102

neV/m

StrongVF

StrongVF

V

neV 350 NbVF

& systematic effects

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

8

Overview

Introduction• How do we search for the nEDM?


Neutron/mercury gyromagnetic ratio• Magnetometry• Preliminary Results

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

9

PSI UCN source

ProtonsSpallation targetEn~MeV

D2O moderatorNeutrons thermalized to 25 meV

1m

Main shutter

UCN storage volume

Neutron guide to experiments

UCN convertor (solid D2 @ 5K)

590 MeV 2.2 mA

Golub, R. & Pendlebury, J. MPLA (1975)133Anghel, et. alNIMA (2009) 272

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

10

• 30 UCN/cm3 at beam exit(~550 000 UCN/25 liter)

UCN source performance

Standard operating PulseNorm Pulse

UC

N C

ounts

/s

VAT

CascadeDetector

VAT

1m glass tube

- NiMo coating- Stainless steel flanges- shutter DLC coated

UCN

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

11

The nEDM spectrometer

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

12

• Optimize product

Filling UCN

14000

N (

aft

er

30s

stora

ge)

Pola

riza

tion

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

13

Simultaneous spin detection

~20% increase in sensitivity(for 2014)

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

14

Storage life time

• Chamber made of dPS insulator ring and DLC electrodes

• Two exp fit:ts~30stf ~180s

• Max no. UCN measured after 180s storage:10 500

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

15

Transversal depolarization

Best @ PSIT2(B0↓) : 1158 ± 94sT2(B0↑): 836 ± 63s

2013:T2(B0↓/B0↑)~600s

→ Excellent magnetic field homogeneity

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

16

RAL/Sussex/ILL*

PSI 2013

best avg best avg

E-field 8.8 8.3 12 10.3

Neutrons 14 000

14 000

10 500

6 500

Tfree 130 130 200 180

Tduty 240 240 340 340

Α 0.68 0.51 0.62 0.57

(10-25ecm) 2.0 2.8 1.5 2.8

Neutron EDM sensitivity

2013 data taking: 3266 cycles25 daysStopped by switch break down2013 accumulated sensitivity 6×10-26 e.cm * Best nedm limit:

Baker et al., PRL97(2006)

2σ

TE Nα

~7500

~0.75

<2

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

17

Overview

Introduction• How to search for the nEDM?


Neutron/mercury gyromagnetic ratio• Magnetometry • Preliminary Results

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

18

The measurement technique

Measure the difference of precession frequencies in parallel/anti-parallel fields:

BBμEEdΔ nn 22

Statistical accuracy of a magnetometer correcting for a change in B should be better than the neutron sensitivity per cycle:

0 1 Tμn

111 Hz 400fTμ

2Bf Bδ δ

T Nπα

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

19

• Average magnetic field (volume and cycle)

• σB ~ 400 fT• τ > 100 s without HV • s/n ~ 500 without HV

Mercury co-magnetometer

¼ wave platelinear polarizer

Hg lamps

PM

polarization cell

HgO source

B0

≈

1μT

τ = 140s

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

20

Goal: Measure magnetic field drifts via optically detected nuclear magnetic resonance (ODMR) with 199Hg atoms

Mercury co-magnetometer

Hg source

circular polarizer unit

253.7 nm UV light,circularly polarized

UCN precession chamber

PMTpolarization chamber

( M

. Fe

rtl, P

hD

-Th

esi

s 2

01

3 )

• Fourth harmonic generator (IR→VIS→UV)

• Toptica TA FHG (20 mW @ 254 nm)• 50 m away from the nEDM setup

50m

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

21

Hg laser readout

Sensitivity improvement from ~400fT → <100fT

( M

. Fe

rtl, P

hD

-Th

esi

s 2

01

3 )

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

22

𝑅=⟨ 𝑓 UCN ⟩⟨ 𝑓 Hg ⟩

=𝛾 n

𝛾Hg (1∓ 𝜕𝐵𝜕 𝑧

Δ h

|𝐵0|+

⟨𝐵2⊥ ⟩

|𝐵0|2 ∓𝛿Earth+𝛿Hg − lightshift)

• There are two drawbacks using the HgMas field monitor:• Center of mass offset• Non-adiabaticity

Frequency ratio R = fn/fHg

UCN199Hg𝛾Hg

2𝜋≈ 8 Hz / μT

𝛾 n

2𝜋≈ 30 Hz / μT

+ further sys.

CsM - array

m/s vs. m/s

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

23

Cesium gradiometer

Monitoring of vertical magnetic gradients• Six HV CsM

• Ten ground CsM• Stabilized laser• PID phase locked

DAQ

1 2 … 6

7 … 15 16

±120kV

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

24

• Field parameterized by 9 parameters (2nd order)

• The parameters are adjusted to 12 magnetometers readings using least squares

• Residuals are about 30 pT (500 pT using only linear model)

• Jackknife procedure to estimate the error on the gradient G

Gradients from CsM

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

25

Extracting R

Run 6043

Χ2/dof = 1.3

=

𝐾=2𝜋 (𝑇 + 4 𝑡𝜋 ) ⟨ 𝑓 Hg ⟩

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

26


=𝛾 n

𝛾Hg (1∓ 𝜕𝐵𝜕 𝑧

Δ h

|𝐵0|+

⟨𝐵2⊥ ⟩


R vs gradient

B0 upB0 down

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

27


=𝛾 n

𝛾Hg (1∓ 𝜕𝐵𝜕 𝑧

Δ h

|𝐵0|+

⟨𝐵2⊥ ⟩


• There are two drawbacks using the HgMas field monitor:• Center of mass offset• Adiabatic vs Non-adiabatic field sampling

Frequency ratio R = fn/fHg

UCN199Hg𝛾Hg

2𝜋≈ 8 Hz / μT

𝛾 n

2𝜋≈ 30 Hz / μT

+ further sys.

CsM - array

Field maps

m/s vs. m/s

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

28

• Mapping with fluxgate and vector cesium magnetometer

• -20<z<20, -10<r<30, Δφ = 5°

Transversal fields

UCNs: Adiabatic regime

199Hg: Non-adiabatic regime


=𝛾 n

𝛾Hg (1∓ 𝜕𝐵𝜕 𝑧

Δ h

|𝐵0|+

⟨𝐵2⊥ ⟩


Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

29

B0up

λ fEarth

Earth rotation correction

Earth EarthnEarth

Hg n Hg

6

sin

5.3 10

f fγδ λ

f fγ


=𝛾 n

𝛾Hg (1∓ 𝜕𝐵𝜕 𝑧

Δ h

|𝐵0|+

⟨𝐵2⊥ ⟩


Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

30

Vector light shift in 199Hg

Atomic energy levels are influenced by the interaction with light

Spin dependent part can be by modeled as effective magnetic field

(which only alters the Larmor frequency of the Hg atoms, not of the UCN)

Effective magnetic field depends on:• Light intensity• Light polarization• Light frequency• Light beam alignment

All four parameters have to be under control for reliable calculations.

B0

BLS

Beff

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

31

• Change of light intensity 30% / 70%• Change of relative angle B-field / light

beam

Hg lightshift measurement

-1.5

-1.5

1.5

1.5ppm

ppm


=𝛾 n

𝛾Hg (1∓ 𝜕𝐵𝜕 𝑧

Δ h

|𝐵0|+

⟨𝐵2⊥ ⟩


Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

32

Result (submitted to PLB)

γn

γHg

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

33

Error budget

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

34

• The UCN source delivers sufficient statistics for data taking, potential improvements are being identified

• The nEDM experiment is taking data, operational reliability has been improved during shutdown 2014

• As a test of magnetic field control we have measured the most precise gyromagnetic ratio of mercury-199 and neutron.

• We expect with 300 data-days until 2016 a statistical sensitivity of σ ≲10-

26 e⋅cm

Conclusion

Ort, Datum

Philipp Schmidt-Wellenburg

Thank you for your attention.

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

36

detector1

Cascade UCN detector

Standard operating PulseNorm Pulse

UC

N C

ounts

/s

Measurements on beam West-1

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

37

Cs OPM

Servo

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

38

Vector cesium magnetometer

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

39

Cs vector magnetometer

|B|

Bz

By

Bx

75 fT

400 fT

10 – 20 pT

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

40

Analysis of a runThe analysis:Ramsey Fit

Mercury frequency

Cut on cycles with residuals larger than 3 (10 %)

Typical χ2=1.5

aN N

AN N

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

41

UCN depolarization

Measurement of frequency ratio R asfunction of vertical gradient revealed unexpected polarization behavior andnon-linearity for large gradients.

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

42

Hg laser readout

( R

&D

for

n2

ED

M,

test

ed

in

nE

DM

)

Sensitivity improvement from ~400fT → ~100fT

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

43

• Ultracold neutrons with different energies average differently over the storage volume

• UCN with low energies, will depolarize faster:→ contribute less to the

average field measurement,

→ frequency changes non-linear

Gravitationally enhanced depolarization

0 cm

BB E B h E

z

Figures from:P. G. Harris et al, PRD89, 016011(2014)

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

44

Side remark: nEDM sensitivity

RAL/Sx/ILL* PSI 2012 PSI 2013

best avg best avg best avg

E-field 8.8 8.3 8.33 7.9 12 10.3

Neutrons 14 000

14 000

9 000

5 400

10 500

6 500

Tfree 130 130 200 200 200 180

Tduty 240 240 360 360 340 340

α 0.6 0.453 0.65 0.57 0.62 0.57

(10-25ecm) 2.3 3.0 2.3 3.5 1.5 2.8

2013 data taking: 3266 cycles25 daysStopped by switch break down2013 accumulated sensitivity 6×10-26 e.cm

*Best nedm limit:Baker et al., PRL97(2006)

𝜎 (𝑑n )= ℏ2𝛼𝐸𝑇 √𝑁

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

45

Magnetic fields

B0

BT

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

46

• Field parameterized by 9 parameters (next-to-linear)

• The parameters are adjusted to 12 magnetometers readings using least squares

• Residuals are about 30 pT (500 pT using linear model)

• Jackknife procedure to estimate the error on the gradient G

Gradients from CsM

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

47

fn/fHg vs gz

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

48

B0up

λ fEarth

Earth rotation correction

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

49

The pseudo magnetic field

𝑏5=𝐵0

𝑅u −𝑅d

𝑅u+𝑅d

= (0.08 ± 0.56 ) pT

⟨ 𝑓 UCN ⟩⟨ 𝑓 Hg ⟩

=𝛾 n

𝛾Hg (1∓ 𝜕𝐵𝜕 𝑧

Δ h

|𝐵0|+

⟨𝐵2⊥ ⟩

|𝐵0|2 ∓𝛿Earth ±

𝑏5

𝐵0 )

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

50

A new limit

gSgP 2 4:8 1027

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

51

HV performance

+160

-160

HV

(kV

)

• Best HV performance 144 kV → 12.0 kV/cm

• Average E-Field while data taking: 10.3 kV/cm

Philip

p S

chm

idt-W

elle

nburg Le

pto

n M

om

ents 2

01

4, 0

72

1-0

7/2

4/1

4

34

52

Hg co-magnetometer

Extract B field from Larmor frequencyand correct UCN frequency (lamp data)

16 h

1.5 pT

nEDM @ PSI

Documents

Transcript of nEDM @ PSI