polarimetry in hall ahallaweb.jlab.org/equipment/moller/talks/polar_halla_12gev.pdfmøller...

Outline

Polarimetry in Hall A

E.Chudakov1

1Hall A, JLab

Moller-12 Workshop, Aug 2008

E.Chudakov Moller-12 Workshop, Aug 2008 Polarimetry in Hall A 1

Outline

Outline

1 IntroductionCompton and Møller Polarimetry

2 Polarimeter in Hall ACompton PolarimeterMøller PolarimeterHigh field upgradeAtomic hydrogen trap

3 SummarySummary


Introduction Polarimeters in Hall A Summary

Compton versus Møller Polarimetry

Compton Polarimetry Møller Polarimetry

A = − 79

lab ∼ 180 mbster

Kinematics/asymmetry

• Rad. corr. to Born < 0.1%

• Detect γ at 0, e− < Ebeam

• Strong dAdk ′ - need σEγ/Eγ 1

• A ∝ kE at E < 20 GeV• k ′ ∼ 4γ2k• T ∝ 1/(σ · A2) ∝ 1/k2 × 1/E2

• Rad. corr. to Born < 0.3%

• Detect the e− at θCM ∼ 90

• dAdθCM

|90 ∼ 0 - good systematics

• A(E) ∼ const , dσdΩ ∼ s−1

• Coincidence - no background





A = − 79

lab ∼ 180 mbster

Polarized target

• Plaser ∼ 100%• Non-invasive measurement

• Ferromagnetic target PT ∼ 8% > 1 µm: invasive Beam IB < 2− 4 µA (heating) Levchuk effect Low PT ⇒ dead time Syst. error σ(PT )∼ 2% for B < 2 T∼ 0.3% for B > 3 T





A = − 79

lab ∼ 180 mbster

AccuracySyst. error 3→50 GeV:∼ 1. → 0.5%Hard at < 1 GeV: (JLab project)∼0.8%

Syst. error ∼ 3% typically,0.5%(1%?) at high magn. fields



Compton Polarimeter at low enegy: CW cavity

λP=1kW

=1064 nm, k=1.65 eV

Electron Beam Electrons detector

Photons detectorMagnetic Chicane

k’

E’

E

• Beam: 1.5-6 GeV• Beam: 5− 100 µA at 500 MHz• Laser: 1064 nm, 0.24 W• Fabry-Pérot cavity ×4000 ⇒

1 kW• Crossing angle 23 mrad• e− detector - Silicon µ-strip• γ detector - calorimeter

Stat: 1.0% 30 min, 4.5 GeV, 40 µA

Syst: 1.2% at 4.5 GeV

Upgrade Plans - 1% at 0.85 GeV• Laser: 532 nm, 0.1 W• Cavity ×15000 ⇒ 1.5 kW• Detector upgrade



Electron and Photon Detection

Electron Detector

• 4 planes × 48 strips × 650 µm• New: 192 strips × 250 µm• Calibration for photon detector

Photon Detector

• PbW, 2× 2× 23 cm3

• 1 central crystal used• New: one large GSO crystal

function

RðADC; kÞ ¼ A eðADCADC0Þ2=2s2

R ; ADCXADC0

RðADC; kÞ ¼

A ð1 dÞ eðADCADC0Þ2=2s2

L þ Zþ ðd ZÞADC4

ADC40

" #,

ADCpADC0 ð9Þ

where A, ADC0 and sR=L are Gaussian para-meters, and Z, d denote proportional amplitudesP4ð0Þ=A and P4ðx0Þ=A, as described in Fig. 4. A isfixed by normalizing the integral of the responsefunction to 1 in the denominator of Eq. (7). Theremaining five parameters are functions of thescattered photon energy k, fitted to data from allelectron detector strips which fired. The Gaussianwidths sR=L are corrected for smearing due to thewidth of the electron strips (sE 5MeV).

The electron detector cannot be put closer thana few mm to the beam axis and thus restricts therange over which the response function can bedetermined. For instance, only photon energiesbetween 150 and 340MeV (Compton edge) couldbe explored with a 4.6GeV beam. The determina-

tion of the calorimeter response function is wellcontrolled inside this energy range but the extra-polation to lower energy induces larger systematicerrors (see Section 6).

5.2. Calibration and analyzing power

The response function measured during aspecific reference run has to be corrected for meangain variations when used to analyze a later run.To this end a calibration coefficient l is introducedwhich accounts for gain corrections

RðADC; kÞ ¼ RðADC

l; kÞ (10)

l is fitted to the experimental spectrum of each run(Fig. 5) using the convolution of the unpolarizedCompton cross-section ds0ðkÞ=dk with the re-sponse function

dNðADCÞ

dADC¼

Z kmax

0

ds0ðkÞdk

RðADC; kÞdk. (11)

The probability of photon detection is deducedfrom Eq. (7), where the lower integration bound-ary ADCs is replaced by ADCs=l. The analyzing

ARTICLE IN PRESS

Fig. 5. Fit of the experimental photon spectrum using the smeared cross-section. The fit range is restricted to the validity energy range

of the modelling.

S. Escoffier et al. / Nuclear Instruments and Methods in Physics Research A 551 (2005) 563–574568



Existing Compton Polarimeter at 1064 nm

source A errorAsymmetry

Statistical 0.80%Position and angle 0.30%Background 0.05%Dead time 0.10%Cuts 0.10%

LightPolarization 0.50%

Analyzing powerResponse function 0.45%Calibration 0.60%Pile up 0.45%Rad. correction 0.26%Total systematic 1.15%Total 1.40%

E (GeV)

δPe/

Pe

(%)

Syst.

Syst. + 1% Stat.

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

0 1 2 3 4 5 6 7 8



Hall A Møller Polarimeter

1998 - commissioned2005 - target upgrade

2008 - target upgradeplanned (high field)

-80

-60

-40

-20

0

20

40

0 100 200 300 400 500 600 700 800Z cm

Y cm

(a)

Targ

et

Col

limat

or

Coils Quad 1 Quad 2 Quad 3 Dipole

Detector

non-scatteredbeam

-20

-15

-10

-5

0

5

10

15

20

0 100 200 300 400 500 600 700 800Z cm

X cm

(b)B→

0.8–6. GeVMinimal Levchukσstat = 1% in∼2–3 minBZ ∼ 25 mT fieldFoil at 20 to fieldFoils 5–30µmthickBeam <2µA



Systematic Errors

The goal for the systematic error

Variable ErrorOLD Present PREX goal

Target polarization 3.5% 2.0% 0.5%Target angle 0.5% 0.5% 0.0%Analyzing power 0.3% 0.3% 0.3%Levchuk effect 0.2% 0.2% 0.2%Dead time 0.3% 0.3% 0.3%Others - - 0.3%Total 3.6% 2.1% ∼1.0%



How to measure the “real” beam?

PREX will run at 50 µATarget heating < 50K to avoid errors on depolarizationRun the injector as close to the regular running as possible

Average current 〈Ibeam〉 < 2 µA ⇒∆T<30 KLaser cycle: 1 ms pulse at 30 Hz∆Tpulse ≈ 12 KBeat frequency laser–chopper:500 MHz ⇒ 500/4 MHzLower average rate ⇒ 1% statistically in10–20 min



Appendix:Bunch suppression

Options (from the draft of a paper by M.Poelker et al)

G0: laser running at 499/16MHz - too long to installFor regular bunch charges: laser at Flaser < FRFbunch suppresssion on the chopper.Beat frequency condition (FRF = 499MHz):Flaser · (n + 1) = FRF · n,n = 3, 4, 7, 15, 31, ... - “magic” numbers

regular Flaser = FRF n = 15 continuous



Appendix:Beat frequency mode - leak through

Pulses overlapτpulse ∼200 ps @50 µAτpulse grows with Ibeam(electro-repulsion)Fully open slit 110 psNo leak: ∆τ > 160 ps

Appendix:Optimization

n=15 same slit ∆τ = 133 ps, contamination ∼ 5% - badn=7 same slit ∆τ = 285 ps, no contamination; other slit∆τ = 95 ps leak ∼ 30% - invasive for other hallsn=4 other slit ∆τ = 166 ps non-invasive?



Macro-pulsing - tune beam

Pulses ∆t > 4 µs at repetition rate k × 30 HzLimitation: at Iinst = 50 µA accelerator stabilization time∼ 100 µsNo micro-suppresssion: ∆t = 1 ms at 30 HzMicro-suppresssion n=4: ∆t = 1 ms at 120 Hz



Polarized atomic hydrogen in a cold magnetic trapE. Chudakov and V. Luppov, IEEE Trans. Nucl. Sci. 51, 1533 (2004).

Ultra-cold traps

30K

0.3K

H

Solenoid 8T

beamStorage Cell

4 cm

40 cm

Atom H1: ~µ ≈ ~µe, E = −~µ~BPopulation ∝ exp(−E/kT )At 300 mK Pe ∼ 1− 10−5

Density ∼ 3 · 1015cm−3

Lifetime > 1 hStat. 1% in 10 min at 100 µA

Contamination and Depolarization at 100µA CEBAF

• Hydrogen molecules < 2 · 10−5

• Upper states |c〉 and |d〉 < 10−5

• Excited states < 10−5

• Helium, residual gas <0.1% - measurable• Depolarization by beam RF < 5 · 10−5

• Ion, electron contamination < 10−5

• Ionization heating < 10−10

Expected depolarization < 10−4

Limitations• Problems ∝ I2

b/F ⇒ “continuous” beam• Complexity of the target

Advantages

• Expected accuracy < 0.5%

• Non-invasive, continuous, the same beamE.Chudakov Moller-12 Workshop, Aug 2008 Polarimetry in Hall A 13


Summary

Møller polarimetry in Hall A

Old: ∼ 2.1%, using 8% Fe at ∼ 25 mT , invasiveNew: ∼ 1%, using 8% Fe at ∼ 4 T, invasiveMay provide < 0.5% accuracy with 100% polarizedhydrogen, non-invasive, continuous

Compton polarimetry in Hall A

Accuracy ∼ 1%(4GeV ) ⇒ 0.5%(11GeV ), non-invasive


polarimetry in hall ahallaweb.jlab.org/equipment/moller/talks/polar_halla_12gev.pdfmøller...

Documents