Laser driven sources of H/D for internal gas targets

Ben Clasie - Massachusetts Institute of Technology 11

Laser driven sources of H/D for internal gas targets

Ben Clasie

MIT Laboratory for Nuclear Science

C. Crawford, D. Dutta, H. Gao, J. Seely, W. Xu


Outline

Introduction and motivation

•Physics motivation for polarized gas targets

•Storage rings and internal gas targets

•Atomic Beam Sources (ABS)

Polarized H/D Laser-Driven Sources/Targets (LDS/LDT)

•Optical pumping

•Spin-temperature equilibrium

•Previous efforts on LDS/LDT

MIT laser-driven target

•Experimental setup (some details on Faraday rotation diagnostics)

•Results and simulations

Summary


Introduction and Motivation

Polarized beams and polarized targets are relatively new technologies

By flipping either the beam or target polarization, small (~10%) changes in the scattering rates are observed

This is an extremely powerful technique as:1) detector efficiencies cancel, and,2) such double-polarization asymmetries are

more sensitive to quantities otherwise difficult to access, for example the nucleon electromagnetic form factors

Nucleon electromagnetic form factors describe the electromagnetic structure of the nucleon


. 3

2 2

( ) ( )

11 ...

6

iq xF q x e d x

q r

Form factors

Kinematics

Form factor

k k'

q

p p'


The Proton Electromagnetic Form Factors

Unpolarized scattering

2 2 2 22 2

2 4

cos 2 2 tan 214 sin 2

p ppE M

M

d E G GG

d EE

Polarization transfer

tan2 2

pt eE

pM l p

PG E E

G P M

Super-ratioPE

PL LL M

PER

PR RM

Ga bA G

GA a bG


Storage ring• Many passes through the target gas

• Large stored current, typically 0.1 to 1A

• , COSY, IUCF, RHIC

• , HERA, Bates, NIKEF, VEPP

Internal gas target• Nuclear polarized H/D for the target can

only be produced in small quantities • Windowless storage cell• Storage cell increases target thickness vs.

jet targets

p

d

e

e

Storage rings and internal gas targets

Stacking

Storage


Storage cells

• 1966 – Idea to use a storage cell to increase the target density (Willy Haeberli)

• 1980 – First test of a storage cell at Wisconsin scattering1000 wall collisionsNo observable depolarization

• The polarized target gas is produced by breaking H/D molecules into atoms, which depolarize quickly on most surfaces

• Recombination produces molecules where little (if any) nuclear polarization is retained

• The storage cell walls are usually coated with teflon or drifilm

( , )H p

Erhard Steffens and Willy Haeberli, Rep. Prog. Phys. 66 (2003) 1887


Atomic Beam Sources (ABS)

Standard technology

Zeeman splitting of the hydrogen hyperfine energy levels

|1|2

|3|4

{Unpol.H

Polarized HSingle state

MFT2-3


Atomic Beam Source (ABS)

• Conventional polarized H/D source• Pure atomic species• High Deuterium tensor polarization

Laser Driven Source (LDS)

• Potentially higher Figure Of Merit • Larger target thickness• Compact design

However …

• Dilution from alkali vapor (potassium or rubidium)

• Drifilm coating deteriorates (~100 hrs) due to the presence of the alkali

Atomic Beam Source (ABS)


Polarized H/D Laser-Driven Sources and Targets (LDS/LDT)

1) A circularly polarized laser is

absorbed by potassium vapor,

which polarizes the potassium

(optical pumping)

2) The vapor is mixed with hydrogen

(H) and spin is transferred to the H

electrons through spin-exchange

collisions

3) The H nuclei are polarized

through the hyperfine interaction

during frequent H-H collisions

hydrogen

potassium

hydrogen

potassium

hydrogen

potassium


Optical pumping

The potassium D1 line is split in a

magnetic field of ~1kG

Photon angular momentum is

transferred to the potassium vapor

polarized potassium

No N2 quench gas is required like 3He targets

Spin-exchange collisions

RadiativeDecays(unpolarized)

+3

1

3

2

2

1jm2

1jm

2/12S4

2/12P4

PumpingRadiativeDecays(unpolarized)

+3

1

3

2

2

1jm2

1jm

2/12S4

2/12P4

Pumping +3

1

3

2

2

1jm2

1jm

2/12S4

2/12P4

Pumping

3Li

11Na

19K

37Rb

55Cs

87Fr

Larger target dilution

Lower spin-exchange cross sectionHigher operating temperature

} Candidates for an LDS

3Li

11Na

19K

37Rb

55Cs

87Fr

Larger target dilution

Lower spin-exchange cross sectionHigher operating temperature

} Candidates for an LDS

HH KK


Spin-Temperature equilibrium (STE)

The H/D nuclear polarization is

given by the spin temperature, β

The H or D nucleus becomes

polarized through H-H or D-D

collisions

STE is reached when:

Deuterium polarization:

Spin exchange rate to H nuclei = spin exchange rate back to H electron

( ) /FmFm e N

H/D hyperfine state population:

Hydrogen polarization: pz = Pe


Radiation trapping

• Fluorescent photons can depolarize the alkali vapor

• T. Walker and L. W. Anderson (1993) suggested using a larger magnetic field in an LDS

• A magnetic field in the kG range shifts the wavelength for + and - absorption

depolarizing fluorescent photons are not absorbed

HOWEVER… The transfer of spin to the H/D nuclei via the hyperfine interaction is reduced at large magnetic fields

Compromise: B ~1.0 kG for hydrogen and less for deuterium.

T. Walker and L. W. Anderson, Nucl. Instr. And Meth. A334, 313 (1993)


A. Kastler (1950) first proposed using light to produce

atoms with nuclear polarization

Previous efforts on LDS/LDT

After the development of lasers with high power and narrow linewidths, development of an early LDS began at Argonne National Laboratory in the late 1980’s

A. Kastler, J. Phys Radium 11, 225 (1950)

In 1998, an LDT was used for the first time in a physics experiment at IUCF

In the mid to late 1990’s, LDS and LDT projects were begun at the University of Erlangen and at MIT


Argonne LDS

H flow = 1.7 1018 atoms/s, f = 0.75, Pe = 0.51D flow = 0.86 1018 atoms/s, f = 0.75, Pe = 0.47

M. Poelker et al., Phys. Rev. A. 50 2450 (1994)M. Poelker et al., Nucl. Instr. and Meth. A 364 58 (1995)

Originally tested in a source configuration (LDS)

More wall collisions from a storage cell will reduce the polarization and degree of dissociation

Extremely good results were obtained in this source configuration


Results from the pzz polarimeter (Argonne, 1998)

J. A. Fedchak et al., Nucl. Instr. and Meth. A 417 182 (1998)

pzz polarimeter based on work by Price and Haeberli

D+ ions accelerated from the target region

In the reaction:

D + 3H n + 4He

Neutron angular distribution is anisotropic if D is tensor polarized


Verification of STE at Argonne

B = 600GSTE conditions

B = 3600GNon-STETransfer of polarization to

the nucleus is suppressed at large magnetic fields

Solid and dashed line in the first graph are from theory that assumes STE

Non-STE theory was used in the second graph

Correction for wall depol.

pzz under operating conditions agree with STE

J. A. Fedchak et al., Nucl. Instr. and Meth. A 417 182 (1998)


IUCF Laser-Driven Target

Doct. Thesis R. V. Cadman, University of Illinois at Urbana-ChampaignR. V. Cadman et al., Phys. Rev. Lett. 86, 967 (2001)C. E. Jones et al., PST99, p 204M. A. Miller et al., PST97, p148R. V. Cadman et al., PST97, p 437H. Gao et al, PST95, p67

The Illinois target was moved to IUCF in 1996

Modifications:

No transport tube

Low B field region

Storage cell was 40cm 3.2cm 1.3cm with rectangular cross section

Nuclear polarization from proton scatteringHydrogen:

Deuterium:Average pz = 14.5%

Average pz= 10.2%


IUCF 1998 H and D run (CE66 and CE68)

Measurements with the electron polarimeter should agree with the nuclear polarization

However: from the graphs and for both H and D,

f 0.45, Pe 0.41

From STE, we should get

H vector pol: 13.7%D vector pol: 17.4%

Conclusion: H is in STE, D is not in STE


Erlangen Laser-Driven Source

Doct. Thesis J. Wilbert, Uni. Erlangen.http://eomer.physik.uni-erlangen.de/forschung/forschung.htmlJ. Stenger et al., Nucl. Instr. and Meth. A 384 333 (1997)

Developed many diagnostic tools for the LDS

All important operating parameters can be monitored and/or optimized

Dissociator optical monitorFaraday rotation monitorBreit-Rabi polarimeter


Verification of STE at Erlangen

J. Stenger et al., Phys. Rev. Lett. 78, 4177 (1997)

Hydrogen flow 41017 atoms/sB = 1500 GPe = 0.51 0.02

A Breit-Rabi polarimeter is an inverted ABS

Transitions between the hyperfine states are possible

All results are consistent with STE

Measurements from a Breit-Rabi polarimeter


MIT Laser-Driven Target


MIT Laser-Driven Target

Gas panel

Magnetic field

Pump laser system

Probe laser system

Glassware/coating

Dissociator

Storage cell

Heaters

Polarimeter

Vacuum pumps

Control software

Polarimeter


Faraday rotation diagnostics

• The Faraday effect is the rotation of linear polarized light by a medium in a magnetic field ( )

• Provides information on the alkali vapor: density, polarization, and, polarization time constants

ˆ //k B


Faraday rotation diagnostics

Linearly polarized light can be decomposed into two circular counter-rotating components σ+ and σ-

Faraday effect occurs when a B-field is applied

Ln n

c

n+, n- refractive index for σ+, σ-

where, V and α are Verdet Coefficients,

J. Stenger et al., Nucl. Instr. and Meth. A 384 333 (1997)

Population differences in the Ms = +1/2 and -1/2 ground states result from optical pumping

( , ) 1 ( , )

n n

V B NL P B

,P n n

Adapted from D. Budker, et. al., Rev. Mod. Phys. 74, 1153 (2002)

1n n


Probe laser system

Ti:Sapph laser tunable from 700 to 850nm

0.001nm linewidth

Low power required <1mW


Measurement of Faraday rotation

Pp = incident probe laser power

Ph = (horizontally polarized transmitted power)/Pp

Pv = (vertically polarized transmitted power)/Pp

,

,

arctanv v backgroundh

v h h background

P P

P P

Analyzing power is greatest when the initial is 45º

rotate the Faraday polarimeter (or a half waveplate)

Faraday rotation from the glassware must be subtracted

Technique is very useful when the incident power, Pp , is not constant

B=0

B=0

B = 155 mT

J. Stenger et al., Nucl. Instr. and Meth. A 384 333 (1997)

pumpblocked

Pump open


Faraday rotation results

Theory curves

Pump beam choppedProbe beam chopped

Characteristic time for the potassium polarization to decay

Make best fit using Verdet Coeffs nK = 1.6 1011 atoms/cm3

PK = -41% (EOM off), -56% (EOM on)


Monte-Carlo simulation

H/D atoms move in straight lines between wall collisions (molecular flow)

Depolarization and recombination coefficients, depol = 0.00146, recomb = 0.0006


Monte-Carlo results

Spin-exchange Transport TargetCell Tube Cell Total

(1) 1180 200 135 1515(2) 1145 195 155 1495(3) 1140 195 295 1630

(1) Atoms that leave the center sampling hole(2) Atoms that leave the off-center hole(3) Atoms that leave the ends of the target cell

Spin-exchange Transport TargetCell Tube Cell Total

(1) 1180 200 135 1515(2) 1145 195 155 1495(3) 1140 195 295 1630

(1) Atoms that leave the center sampling hole(2) Atoms that leave the off-center hole(3) Atoms that leave the ends of the target cell

Average number of wall collisions

Wall collision results Polarization results


MIT LDT preliminary results

fα = degree of dissociation

Pe = H electron polariz.

H nuclear polariz. (pz)

FOM = Figure Of Merit

= flowpz2, or, thickness pz2

Results for hydrogen only (first priority)

Measurements were made without an Electro-Optic Modulator (EOM)

Future tests with a diamond coating0

0.1

0.2

0.3

0.4

0.75 1 1.25 1.5 1.75 2

H2 Flow rate (1018 atoms/s)

FO

M (

1017 a

tom

s/s

)

0

1

2

3

4

5

6

FO

M (

1013 a

tom

s/c

m2 )

Preliminary results

0

20

40

60(%)

f

Pe

Improves Pe


Figure Of Merit (FOM)

• (atoms/cm2)

• FOM is a measure of the target performance, it is inversely proportional to the running time of an experiment

• (atoms/cm2)

•

, dLThickness t

2FOM t p

p = average polarization as seen by the beam

3flow Lt

d T


Figure Of Merit (FOM) (cont.)

• is the usual definition of FOM, however, there are other considerations

• How do we compare the performance of two different types of polarized targets? - smallest error bars

• may be more useful

• How do we compare the performance of polarized sources at different facilities? - storage cell geometry is usually restricted by beam halo

• This comparison is difficult as there are spin-exchange collisions and wall collisions in the storage cell

•

2t p

2

expbeam erimentt p f f

2 2

collisions geometry temperature recombination dilusourc ne tioFlt p fow f f fp f

1temperaturef

T

sourc ae tomsf pp 1

. 2 1recombf f f


FOM resultsHermes (ABS) `96 - `01 BLAST (ABS) (units)

Gas H D H

Flow (F) 6.5 4.6 2.5 (1016 atoms/s)

thicknesss (t) 7.5 14 3.0 (1013 atoms/cm2)

pz 0.88 0.85 0.45

F pz2 0.50 0.33 0.051 (1017 atoms/s)

t pz2 5.8 10.1 0.61 (1013 atoms/cm2)

IUCF (LDT) 1998 MIT (LDT) Prelim. (units)Gas H D HFlow (F) 1.0 1.0 1.1 (1018 atoms/s)thicknesss (t) 0.3 0.4 1.5 (1015 atoms/cm2)f 0.48 0.48 0.56

Pe,atomic 0.45 0.45 0.37

pz 0.145 0.102

F pz2 0.21 0.10 0.34 (1017 atoms/s)

t pz2 (f ) 0.63 (2.3) 0.42 (1.5) 4.7 (2.7) (1013 atoms/cm2)

E.C. Aschenauer ,International Workshop on QCD: Theory and Experiment, Martina Franca, Italy, Jun 16 - 20, 2001 HERMES target cell has elliptical cross section 29 x 9.8 mm

IUCF target cell had a rectangular cross section 32 x 13 mm

FOM

FOM


Summary

Laser driven sources and targets can provide H/D with high polarization at flow rates in excess of 1018 atoms/s

These offer a more compact design than conventional atomic beam sources and may provide a higher overall FOM

Faraday rotation diagnostics provide important information on the alkali number density, polarization and time constants

Documents

Laser driven sources of H/D for internal gas targets