Chen-Yu Liu

Research Opportunities with Ultracold Neutrons in the US11/7/2009Santa Fe, NM

New Approaches to Ultra-cold Neutron Production

Chris Lavelle, Greg Manus, Patrick McChesney, Dan Salvat, Yun Chang Shin, Phil Childress, Bill Lozowski, Walt Fox (Indiana University /IUCF)

C. Morris, A. Saunders, M. Makela (LANL)

A.R. Young (NCSU)

Superthermal Principle• Cold neutrons downscatter in the condensed matter, 

giving up almost all their energy, becoming UCN.

• UCN upscattering (the reverse process) is suppressed by cooling the moderator to low temperatures.

2

Canonical Source: superfluid HeliumConcept extended to solids : D2 thin film “dirt” source

3

● Lifetime of UCN in the source material, τ, is the critical parameter to achieve large UCN densities.

● Extract UCN out of the source before it is thermalized ⇒ Spallation N source

Dynamics of UCN Production --Defeat thermal equilibrium

+ Seperation of the source and the storage + a UCN Valve

τ

Volume Scan

• UCN yield saturates above 200 c.c – mean free path = 8 cm– Results from UCN incoherent elastic scattering (random walk).

• No strong evidence of reduced UCN extraction from the finite crystal effects.4

LANL UCN prototype source(2000)

Transparent Crystal(looking down the UCN sourceFrom the top of the guide)

5

• UCN density:(Limited by loss)

• The figure of merit:

Isotop σcoh σinc σa σs/σa purity Debye T

2D

5.59 2.04 0.000519 1.47×10

4 99.82 110

4He

1.13 0 0 ∞ 20

15N

5.23 0.0005 0.000024 2.1×10

5 99.9999 80

16O

4.23 0 0.00010 2.2×10

4 99.95 104

208Pb

11.7 0 0.00049 2.38×10

4 99.93

105

s a

⎟⎟⎠

⎞⎜⎜⎝

⎛∝× ...111 +

σ+

σ+

σστP=ρ

nucl.ab.βupdownucn

Good Candidate Material

Solid Oxygen as a UCN Source

• Electronic spin S=1 in O2 molecules.• Nuclear spin = 0 in 16O• Collinear Anti‐ferromagnetic in 2‐D

– α‐phase, T < 24K.

UCN Production in alpha S‐O2

• Produce UCN through magnon excitations.– Magnetic scattering length ~ 5.4 fm. 

• Null incoherent scattering length.• Small nuclear absorption probability.

6

P.W. Stephens and C.F. Majkrzak, Phys. Rev. B 33, 1 (1986)

⇒ A very large source possible.

σcoh =4.232 barn, σinc = 0 barn,σabs = 0.0001 barn

Neutron-Magnon Scattering in S-O2

• Magnon production energy gap ~ 0.8meV– Magnons partially frozen at T 

< 8K.– Significantly reduce the UCN 

upscattering rate. 

7

0.8meV

Phonon Density of States

Magnon Density of States

Calculated S(α,β) in α‐O2

8

JNN=-2.44 meVJNNN = -1.22 meV

Coupling constant Suggested in P.W. Stephens and C.F. Majkrzak, Phys. Rev. B 33,(1986)

Range of coupling constantsJNN = -3.32 (-2.44) meVJNNN = -0.91 (-1.22) meVJ⊥ < 0(0) meV

UCN production in Solid Oxygen

• Production rate

– P = 2.7 × 10‐8 Φ0 (30K CN in S‐O2)

– P = 3.0 × 10‐8 Φ0 (15K CN in S‐O2)

– P = 1.5 × 10‐8 Φ0 (30K CN in S‐D2)

– Gain ~ 2 relative to S‐D2

• Lifetime

– 375 ms in S‐O2

– 40 ms in S‐D2

– Gain ~ 10

• Volume gain, (l)n, n= 1‐3

– lucn = 380 cm in S‐O2

– lucn = 8 cm in S‐D2

– Gain ~ 50 ‐ 105

9

Compared with S-D2, Gain > 1000 is possible !

10

SC solenoid Cryostat

Pulse-tubeRefrigerator(1.5W @4K)

Cold N

M1BeamMonitor

M2 M3

UCN Guide:polished SS Guide (186 neV)

UCN Detector(ion chamber w/10mbar 3He, 1000mbar CF4)

Target Cell (100 c.c.)

Choppers

Sep~Oct, 2008

Low Statistics Measurements

Cold Neutrons~3×106 n/cm2-s

M1 M2 M3

10 cm ∅

BG : 300 Hz

0.16 Hz

0.03 Hz

÷ 2000

÷ 5.3UCN signal rate ~ 0.34 Hz (D2)

0.15 Hz (O2)S/N ~ 5/1 for O2

Bird’s Eye View

B4CBorated PE

Cold Neutron Monitors• Provide information of incident spectrum and total xs

12

Cold Neutron Cross-section

A

Miller Calc Observed111 3.10 3.09 (L1)200 2.69 2.69 (L2)220 1.90 1.90 (L3)311 1.62 1.61 (L4)222 1.55 ‐331 1.23 ‐420 1 20 1 21 (L5)

316L stainless steel (16-18Cr, 10-14Ni + 2-3Mo), FCC Lattice

UCN Detector TOF Signal(allows for further BG rejection)

O2 Gas Handling System (all VCR)

• Optical cell

14

beta-gamma phase transition(slow cool-down~0.017K/min)

Y. Shin

beta phase (slow vapor deposition)

beta phase (slow cool down)

Comparison between D2 & O2

16

Source: 3.7 cm cell, 6 cm I.D., UCN extracted from downstream

Bears no significance in predicting the performance of a full-scale source.

Cold Neutron Attenuation

• Cold neutron attenuated by Bragg scattering,  φ (E, z)

• UCN: Isotropic, v2dv velocity distribution up to 1 μeV. (upper limit of guide reflection energy)

• UCN Production varies with  the depth of the cell:

• UCN Production Cross‐section:

17

σ = 1.84x10-7 barnsAveraged across the cell and all Neutron energies.

( )∫∫

∞

∞

==

0

0

)0,(

)(),(

dEZE

dEEZEZ

φ

σφσ

A Better Geometry for UCN Collection:

18

UCN collected from the front of the cell, upstream of the beam

Geometry used in 2007

Energy Dependent UCN Production

Energy selection of incident CN using dual “frame overlap”choppers in a de‐phased mode.

A Fermi‐Chopper would be better.

19

20

DCS Measurement

21

DCS Measurement

22

Measurements using Disc Chopper Spectrometer, NIST (2009)

UCN Production predicted using S(q,ω)

23

• Gamma (plastic)• molecular orientation• paramagnetic spin fluctuation

• Beta• Spin diffusion• paramagnetic spin fluctuation

• Alpha• Anti-ferro magnons• Phonons

Cold Neutron

Temperature Dependent UCN 

Production

24

Calculated using S(q,w)

Superthermal behavior

Volume Dependence

25

small

medium

large ??

α-O2

Note that the UCN production in the β-phase is already saturated.

α-O2 β-O2 γ-O2

Large Cell

26

× 0.75

Exp(-3.7/24) = 0.62Exp(-11.5/24) = 0.82

= 0.72

Production is only effective in the front of the target cell.Produced UCN has to transmit through the rest of the cell to be collected.

α-O2

β-O2 γ-O2

Analysis

• Rucn: UCN production rate (determined from S(q,w)

• l1 :CN mfp  ;   l2: UCN mfp 

27

∫ −−−d

UCN lxdlxdxR0

21 )/)(exp()/exp(

UCN Mean Free Path in Solid O2

• Bragg Scattering: 4.232 barn– Lcn = 3.94 cm

• Absorption:  0.44 barn – Labs = 379 cm

• Diffuse scattering from grain boundaries, cracks?– Combined analysis of the 

measurements from 3 cells• UCN mfp : 12 ~ 100 cm

(depending on the details of 

cold neutron attenuation)

28

Magnetic Zeeman Splitting in α‐O2

• The Low energy mode 

splits.

29

R.J. Meier, J H P Colpa, H Sigg, J. Phys. C: Solid State Phys., 17, 4501 (1984).

H//Z

H//Y

H//X

H//ZH//X

H//Y

Magnetic Zeeman Splitting in α‐O2

30Data under analysis:Magnetic field dependent UCN productionField dependent S(q,w) data collected at PHAROS, Lujan Center.

Summaries on O2 UCN Source

• β‐O2 has large downscattering (upscattering) cross‐sections– Thin film source (< 2 cm)

• α‐O2 has a smaller UCN cross‐section– UCN Upscattering length > 10 cm (for T < 20K)

– In order to see the superthermal behavior, we need a cell with length > 10 cm.

• Volume scaling is non‐trivial– Bragg scattering of cold neutrons limit the size of the UCN source (especially important for coherent scatters)

– However, one could use a geometry to use multiple‐Bragg scattering and increase the effective length of the source.31

Outlook• Quasi‐elastic scattering 

– Such as, spin diffusion, paramagnetic fluctuation, molecular orientation fluctuation observed in the solid O2 system.

– Possible new type of condensed‐matter based UCN source (NOT super‐thermal)

– VCN source (TCN < 20K)• Ex: β‐O2 at reduced phase transition temperature

• Low temperature phase transitions– O2 in confined geometries (nano‐pore) ⇒ extend beta‐phase down to 10K.  

– Other phase transitions• Structural: He (gas to liquid) at critical point (5.2K, 2.2 atm)

• Magnetic:  Superconducting phase transition, i.e., Pb at 8K; type‐I to type‐II SC transition.

32

Summaries• First observation of UCN production from magnetic excitation.

• Solid Oxygen has comparable UCN production to solid deuterium. If “properly designed”, its low absorption and long UCN lifetime continue to make it an attractive candidate for an intense UCN source.– Large dimension source (1mx1mx1m) is challenging. Cold neutrons don’t penetrate deep inside the source, because of Bragg scatterings.

– The effective large dimension could be achieved with the “flux trap” idea.

33

Our Group at IUCF

Ultra-Cold Neutrons, Neutron EDM, Electron EDM34

Backup Slides

35

Low Energy Neutron Source (LENS) at IUCF

36Proton Linac (13MeV) 1st Target Station: NRE

2nd Target Station

SANS

Sesame

Target/Moderator/Reflector (TMR) Assembly

37

Thin beryllium production target Methane moderator Outer shielding – lead

and poly layersinner reflector – water

9Be(p,n)9B reaction

UCN Module at LENS

38

Concrete Wall

Cryomech4K, 1.5WClosed-cyclerefrigerator

2nd Stage

1st Stage

Water (or Be)Reflector

UCN Guide/UCN Bottle

UCN Detector

B/PE/Pb Shield

~1m

~10cm

Entrance ValveExit Valve

Not to scale

Kernel of the UCN Module

39

Beryllium Reflector

Water cooled production target

(Phase II LENS)

4K Polyethylene

Cold Flux Trap

Solid Oxygen UCN Converter

0.00E+005.00E-051.00E-041.50E-042.00E-042.50E-043.00E-043.50E-044.00E-044.50E-04

-20 -15 -10 -5 0 5 10 15 20

Y Position (cm)

0-5

meV

Flu

x (n

/cm

^2/s

rc) mAscm

nsrccm

n⋅⋅

×→⋅

× −2

102

4 108.1108.3

40

Using a 2.0 cm thick trap, 1000 cc solid oxygen source we can model the volume average cold flux in the oxygen as a function of position. 1 cm left of center is optimal.

The peak cold flux is MCNP by C. Lavelle

18103 −−×≈ cmP

mAcmucn

ucn ⋅≈ 330ρ

41

mAscmcn

cn ⋅⋅×≈Φ 2

10108.1

Solid Oxygen UCN Conversion Rate, P:Calculation based on inelastic magnetic scattering

0-5 meV Neutron Flux, ΦCN(13meV, Be reflector)MCNP simulation validated by theLENS cold flux measurements

UCN Density in source:P×ΦCN × ττ: escape time ~ 30ms

UCN Current output:P ×ΦCN × VV: source volume ~ 1 liter

mAsucnN

mAsucnN

Water

ucn

Be

ucn

⋅×≈

⋅×≈

•

•

5

6

104.6

101.1

Heat Load (MCNP results)

CellPhoton Heat (W)

Neutron Heat (W)

Target Photon Heat (W) Total (W)

sO2 0.20 0.08 0.12 0.39sO2 Al 0.09 0.01 0.04 0.15 0.54Poly 0.39 1.58 0.17 2.13Poly Al 0.20 0.03 0.09 0.32Guide 0.07 0.00 0.01 0.09 2.54

42

With 1mA, 13MeV proton current on Be target

We expect UCN density ~ 30 ucn/cc in the s-O2.UCN fluence: 6×105 ~1×106 ucn/s from a 1 liter s-O2.

Field Dependent UCN Production

43

Physics in Simulation Guide Wall and Foil Losses

• In D2 source– Inelastic scattering in D2

– 2.7% Para, 0.1% HD

– Absorption

– Cracks: Modeled as 20 barns of inhomogenity scattering  from atchison total XS measurement

• Guides – Surface Roughness

– Diffusivity, Specular/non‐specular

– too much and it kills all loss, below that no diff

– welds

• Aluminum Windows 

– 50% transmitting by measurement.

– Assigned reasonable diffusivity #’s

– 80% transmitting.  Remainder assigned to inhomogeneity scattering.

44

D2 data to validate the simulation

45

Temperature (K)0 5 10 15 20

UC

N S

igna

lS

(uni

tless

)

0

10-10

2x10-10

3x10-10

4x10-10

5x10-10

6x10-10

7x10-10

8x10-10

Non-linear FitSimple SubtractionSimulationSimulation w/ Cracks

UCN Production Rate

46

• Cold Neutron Flux8×106 n/cm2-s (S. Penttila)

• Detector efficiency– 80% under the full energy peak

– 50% transmission

• Guide transport efficiency (Geant4 simulation)– 6% with Al cell (0.010” thick) + 2 meters of SS guide

• UCN production rate:

0.15 /s = (8×106 n/cm2‐s) × P × 100 cm3 × 0.8 × 0.5 × 0.06P ≈ 0.8 × 10‐8 Φ ucn/cm3‐s      : agrees with the calculation     

47

• No hysteresis

• A prominent feature around alpha-beta phase transition.

• Absence of the superthermal temperature dependence in alpha phase???

UCN Production in Solid O2

(10 hours)

Magnetic Critical Scattering

48

Frequency spectrum of the spontaneous fluctuations in magnetization??? W. Marshall, R.D. Lowde

“magnetic correlations and neutron scattering”

UCN productionIn our geometry

New tool: UCN

θ θ

• E < 335 neV (Ni58) • T < 4 mK• Velocity < 8m/s• λ > 500 Å

( ) ρbmπ=rrδb

mπ=V coh

niii

n

22 22 hh ∑ −

UCN Production vs. CN Transmission(PSI run at 2005)

5010−4

10−2

0

0.2

0.4

0.6

0.8

1

Energy (eV)

Proba

bility

Scattered CN (Vapor Deposition)

10−4

10−2

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Energy (eV)

Scattered CN (Freeze from Liquid)

Warm Up

Cool Down

UCN yield is strongly correlated with CN scattering.

Material: solid O2

Probe the Magnon Mechanism using a B field

• An external magnetic field to perturb the magnon dispersion curve– Change the density of states.– Optimize UCN production.

• Definitive demonstration of the magnetic origin. 51

C. Uyeda at. al., J. Phys. Soc. Jpn. 54, 1107 (1985)

An unique feature of oxygen!

Spin flop transition around 7 Tesla.

Experimental Program to Investigate Novel UCN Source • Previous Experience

– LANL– PSI

• 2007‐2008 (July – October)– Lujan Center (ER2) Flight Path 12

• UCN production under B field• CN TOF transmission • E dependent Production• UCN gravity spectrometer / Velocity filter/He‐3 filter

– PHAROS: one week beam time to measure S(alpha, beta) in solid oxygen under high field.

– SANS: Quantify the size of crystal grains, voids & cracks.• UCN Source coupled to LENS at IUCF.

52

UCN Production under B Fields

53

No suppressed scattering.Transition TemperatureIncreases by ~ 2K.

Density of States

54A. De Bernabe, G. J. Cuello, F.J. Bermejo, F. R. Trouw, A.P.J. Jansen, PRB, 58, 14442 (1998).

10 meV or 12.5 meV ???

Question:Is the UCN production originatedfrom “magnetic” scattering?

Varying Cold Neutron Spectrum

55

Preliminary

30% increase using a higher energy neutrons

TOF

Faster Neutrons

Slower Neutrons

Scaling to Neutron Current

• Seppo Pentilla, Chris Crawford, Yun Shin Conducted Absolute Neutron Spectrum Measurement immediately following our run.

April 3, 2009 Christopher M. Lavelle, IUCF 56

]//[)(1056.5 24000

20001

5 scmniMi

cn ∑=

•×=φ ]//[1065.8 26 scmncn ×=φ

Volume Dependence

57UCN Production in α-O2 scales with volume, but that in β-O2 does not.

BGsmall

large

Venti ??

α-O2

“Diffusivity” – a heavy handle

58

UCN Detector

April 3, 2009 Christopher M. Lavelle, IUCF 59

Nucl. Instr. and Meth. A (2009), doi:10.1016/j.nima.2008.11.099

15% of the signalLost to low end noise

Grow Optically Transparent O2 Solid

• Recipe (A. Jezowski, et. Al, PRL, 71, 97, 1993)• Annealed slightly below triple point (55K) for 6 h

• Ramp down to 43.9K with 1K/h (through gamma phase)

• Ramp through the gamma‐beta transition (43.9K to 42.5K) for 100 h

• Ramp through beta phase with 1.5K/h

• Ramp through the beta‐alpha transition for 5 hours

(The whole process takes about 6 days)

• Monitor: – Cold neutron scattering cross‐section

– small angle scattering 

– ultrasound transducer60

61

• Incoherent scattering (σinc = 2.04 barn)• The difference of singlet and triplet scattering

• Coherent contribution ( σ coh= 5.59 barn)– In a cold neutron flux with a continuous spectrum, more

neutrons could participate in the UCN production.

UCN production in Solid D2

(1,1.73,0)

(1,1,0)

(1,0,0)

• Nuclear Absorption ( σ abs= 0.0005 barn for 2200m/s neutrons)

Sources of UCN Loss in Solid O2

• Elastic scattering due to cracks & voids– Small angle neutron scattering

• Incoherent elastic scattering (diffuse scattering)– Nuclear: zero– Magnetic: non‐zero for paramagnetic phase (liquid, γ, β); zero for 

coherent phase (α).

• Magnon/Phonon upscattering– Temperature dependent 

• 20K: ~8 barns• 10K: ~1.5 barns

– However, we did not see a strong temperature dependence on UCN yield and CN scattering ???

• Upscattering due to rotational states transitions at high temperatures– Only odd rotational states are present.   (ground state: J=1)– Libron mode

62

For the rest 2 weeks’ of beam• Finish the long procedure of crystal growing (by 10/1, TODAY)• Turn on the B field (10/2~10/7)

– (2 days) UCN production vs. |B|, T=5 K, 22K, 26K, 32K – (1 day) Zeeman splitting, T=5K, B=5.5T, scan over Ei– (1 day) Fix B=5.5T, UCN vs Temperature (warm up) – (1 day) BG with empty cell (for both 10Hz & 20Hz operation)– (3 days) Calibration using D2 source at 5K + 1.5 atm D2 gas at 20K

• Change into a small cell (10/8~10/11 or 10/14)– (1 day) Installation + UCN shutter– (2 days for fast cool‐down, 6 days for slow cool‐down)

• Get a handle of the UCN transport inside the target• Single scattering vs multiple scattering 

• If enough time (10/12~10/14), try– 1.5 atm He gas at 5K– O2 confined in carbon napo‐pore– UCN spectrum measurement through a He‐3 filter 63

Challenges

• 1.8 K cryogenic.– Larger gamma heating and smaller thermal conductivity than S‐D2 ⇒ challenges on cryogenic engineering

– Requires a fast thermal break (50K to 2K) over a few cm. 

• Shortened UCN mean free path – Polycrystalline sample formation – Ozone formation ⇒ additional incoherent scattering, resulting in a 

reduced mean free path. (Low radiation level at LENS helps.) 64

1st experimental attempt with S-O2

(2000, LANSCE) LANL UCN team

• Use the S‐D2 prototype source to test S‐O2

• 20 c.c. S‐O2 at 4.5K• Preliminary results:

– UCN are produced and                                             bottled.– 1/5 UCN produced than                                            from 

S‐D2 with the same                                                   condition– No distinct temperature                                         

dependence

• Limitation: – Diffusive solid oxygen film along the UCN guide all the way up to 

below the L‐N2 radiation shield (~ 1m).

65

Measurement of the UCN production efficiency of solid D2 and comparison with the

solid cryogenic materials CD4 and O2

Collaboration 

F. Atchison, B. Blau,  B. van den Brandt, T. Brys, M. Daum, P. Hautle, R. Henneck, S. Heule, M. Kasprzak, K. Kirch, A. Knecht, J. A. Konter, A. 

Pichlmaier, M. Wohlmuther, G. Zsigmond  (PSI)

K. Bodek, M. Kuzniak, J. Zejma  (Jagiellonian University)P. Geltenbort, C. Plonka  (ILL)A. Holley, A. R. Young (NCSU)

C.‐Y. Liu, Y. Shin (IU)C. L. Morris, A. Saunders, (LANL)Y. Pokotilovski (JINR, Dubna)

D. Tortorella (TUM)66

Experimental Setup at PSI (Oct. 2005)

FunSpin beamline in SINQ– ΦCN = (4.5±1.0)×107/cm2‐s‐mA, with 1.2mA on SINQ target– UCN count rate in the detector of 0.4/s, with a S/N ratio of 40 to 1.– Similar count rate is expected using solid oxygen, assuming the same 

UCN extraction efficiency.67

F. Atchison, et. Al., Phys. Rev. C. 71, 054601 (2005) F. Atchison, et. Al., Phys. Rev. Lett, 99, 262502 (2007)

2.83m

Results of UCN Production in Solid O2 at PSI

• No superthermal temperature dependence.

– unknown source of UCN loss.

• UCN yield is correlated with how the crystal is prepared.

• The UCN yield (best number) is ~ 3 times less than s‐D2.

• A peak in the α‐β phase transition. (critical scattering?)

68

Cold Neutron Transmission (TOF)

• Less scattering compared with D2.

• Bragg edges

• Additional Bragg peak in α‐phase.    ⇒ the presence of a magnetic ordering.

69

Correlation of UCN production vs CN scattering

70UCN production was not effected by temperature or phase.Something (other than downscattering) is dominating the yield of UCN.

Calculated S(α,β) in α‐O2

71

JNN=-2.44 meVJNNN = -1.22 meV

Compared with Data, the low energyMode is too low!

Basic technique: measure precession in external fields

h

BBB

μω 2−=

h

EdEE

2=ω

h

EdEBantiEBE

4=Δ≡− − ωωω

• Apply static B, E||B

• Look for Δω on reversal of E

• Larmor frequency:

(~ 500 Hz for B ~ 1G)

μB

dE

• dE: additional precession:

Figure: E. N. Fortson, et al., Physics Today 56 6 (2003) 33

Complementary sensitivity to LHC

Parameter space defined by the masses of the Higgs (μ) and gauge boson (M1) superpartners. The light blue bands are required for baryogenesis via an electroweak mechanism. The red band is excluded by LEP 2. The grey region is excluded by the requirement that the lightest supersymmetric particle be a neutralino, and thus a candidate for cold dark matter. The prospective reach of the LHC is indicated by the green lines and arrows, while the possible reach of a future linear collider is given by the blue curves. The region below the dashed red curve is excluded by an electron EDM experiment (assuming no signal) with the sensitivity indicated.

Experiment Overview

~ 5 m

Total Cross‐section Measurement

75

PSI

Lujan

At FP12, we are not able to measureneutrons <1meV, because of the frame overlap.

Unable to reconfigure the chopper to take beam at 10Hz.

Free Gas

76

77

Superthermal UCN Production

Landau-Feynman's dispersiocurve for superfluid 4He

Kinetic energy of a free neutron

11 K

0.7/Å

( ) ( )( ) ( )Tω~ωnωn+=

σσ

vnσσΦ=τP=ρ

up

down

updownucn /exp11

0 ∝⎟⎟⎠

⎞⎜⎜⎝

⎛×

Superthermal gain

( ) ( ) 1/exp1

−Tω=ωn

Two level system:

9 A CN°

UCN

• Superfluid 4He– Energy excitation is isotropic.– Neutron scattering is isotropic.

• UCN can accumulate until the production rate = loss rate

R. Golub and J. M. Pendlebury, Phys. Lett, A53, 133 (1975)