W. Lorenzon EIC, 8 November 2002 1

The Longitudinal Polarimeterat HERA

W. Lorenzon (Michigan) Collaboration

• Polarization at HERA

• Laser System & Calorimeter

• Statistical and Systematic Precision

• Laser Cavity Project

• Summary and Outlook

W. Lorenzon EIC, 8 November 2002 2

Electron Polarization at HERA

Self polarization of electrons by Synchrotron radiation in the curved sections (“Sokolov-Ternov effect”)

)1()( /tePtP

W. Lorenzon EIC, 8 November 2002 3

Sokolov Ternov Effect

• PST = 0.924 for an ideal, flat machine, independent of any machine parameter.

• ST = 37 min for HERA @ 27.5 GeV (prop. , R3)

• Depolarization effects (D) in a real machine can substantially reduce Pmax:

and effective build-up time constant :

• PST and ST calculable from first principles a simultaneous measurement of Pmax and provides a calibration method:

DST

DSTPP

max

DST

DST

measτP PkPST

ST max

W. Lorenzon EIC, 8 November 2002 4

HERMES Spin Rotators

spin = aorbit

spin orbit

(mrad)

450 12.5 900 25 1800 50

Spin Tune at 27.5 GeVa = E/(0.440 GeV) = 62.5

W. Lorenzon EIC, 8 November 2002 5

Principle of the Pe Measurement withthe Longitudinal Polarimeter

Compton Scattering: e+ e’+

Cross Section: d/dE = d0/dE[1+ PePAz(E)]

d0, Az: known (QED) Pe: longitudinal polarization of e beam P: circular polarization (+-1) of laser beam

e (27.5 GeV) (2.33 eV)

ECalorimeterBack Scattered

Compton photon

W. Lorenzon EIC, 8 November 2002 6

Experimental Setup - Overview

Overview

Calorimeter position

NaBi(WO4)2 crystal calorimeter

W. Lorenzon EIC, 8 November 2002 7

Experimental Setup – Laser System

- M1/2 M3/4 M5/6: phase-compensated mirrors- laser light polarization measured continuously in box #2

W. Lorenzon EIC, 8 November 2002 8

Experimental Setup - Details

W. Lorenzon EIC, 8 November 2002 9

Laser Control - COP

Automatic Control

- Synchronizing laser and electron beam- optimize luminosity- optimize polarization- center beam on calorimeter- control & readjust all parameters: … laser spots on all mirrors using CCD cameras

W. Lorenzon EIC, 8 November 2002 10

Polarimeter Operation I

Single-Photon Mode

Advantages: - large asymmetry: 0.60 (max) - easy comparison to d/dE

Disadvantage: - dP/P = 0.01 in 2.5 h (too long)

As(E) = (–(+= Pc Pe Az(E)

W. Lorenzon EIC, 8 November 2002 11

Polarimeter Operation II

Multi-Photon Mode

Advantages: - eff. independent of brems. Bkg and detector cutoff energy - dP/P = 0.01 in 1 min

Disadvantage: - no easy monitoring of calorimeter performance

Am = (–(+= Pc Pe Ap

Ap = (–(+

= 0.184 (if detector is linear)

W. Lorenzon EIC, 8 November 2002 12

Polarization Determination

Ap = (–(+

= 0.193 (for crystal calorimeter)

dEErEdEdE

Eii )()/(

max

min

Question: Is response function linear over full single to multi-photon range?

0.9% syst. uncertainty 0.8% syst. uncertainty

DESY

CERN

W. Lorenzon EIC, 8 November 2002 13

Polarization Performance

HERA: 220 bunches separated by 96 ns 174 colliding + 15 non-colliding = 189 filled bunches

20 min measurement dP/P = 0.03 in each bunch

time dependence:helpful tool for tuning

W. Lorenzon EIC, 8 November 2002 14

“Non-Intrusive” Method

Possible to ‘empty’ an electron bunch with a powerful laser beam at HERA

W. Lorenzon EIC, 8 November 2002 15

A large Systematic Effect - Solved

In multi-photon mode:1000 ’s 6.8 TeV in detector

Protect PMT’s from saturation insert Ni foil in 3mm air gap

Problem:NBW crystals are 19 X0

small shower leakage with large analyzing power

Ap(w/ foil) = 1.25 x Ap(w/o foil)No light attenuators are used since early 1999

W. Lorenzon EIC, 8 November 2002 16

Systematic Uncertainties

Source Pe/Pe (%)

(2000)

Analyzing Power Ap- response function - single to multi photon transition

Ap long-term stability

Gain mismatchingLaser light polarizationPockels Cell misalignmentElectron beam instability

+- 1.2

(+-0.9)(+-0.8)

+- 0.5+- 0.3

+- 0.2+- 0.4

+- 0.8

Total +-1.6

new sampling calorimeter built and tested at DESY and CERNstatistics limited

W. Lorenzon EIC, 8 November 2002 17

New Sampling Calorimeter

Wave length shifters

Tungsten plates

Scintillator plates

PMT’sTest beam results

DESY

CERN

‘r(E) = 1’

better energy resolution & linearity than NBW calorimeter

W. Lorenzon EIC, 8 November 2002 18

New Sampling Calorimeter - Details

beam view

side view

- sampling & NBW calorimeters sit on movable table (x-y) fast switching possible- one calorimeter is always protected from synchrotron and bremsstrahlung radiation

W. Lorenzon EIC, 8 November 2002 19

Systematic Uncertainties

Source Pe/Pe (%)

(2000)

Pe/Pe (%)

(>2002)

Analyzing Power Ap- response function - single to multi photon transition

Ap long-term stability

Gain mismatchingLaser light polarizationPockels Cell misalignmentElectron beam instability

+- 1.2

(0.9)(0.8)

+- 0.5+- 0.3

+- 0.2+- 0.4

+- 0.8

+- 0.8(+-0.2)(+-0.8)

+- 0.5+-0.2+-0.2+-0.2+-0.3

Total +-1.6 +-1.0

new sampling calorimeter built and tested at DESY and CERNstatistics limited expected precision (multi-photon mode)

W. Lorenzon EIC, 8 November 2002 20

Polarization-2000

HERMES, H1, ZEUS and Machine Group

Goal: Fast and precise polarization measurements of each electron bunch

Task: major upgrade to Transverse Polarimeter (done) upgrade laser system for Longitudinal Polarimeter

Update on Laser Cavity System(courtesy F. Zomer)

W. Lorenzon EIC, 8 November 2002 21

Precision Measurements (QCD):

Neutral Current (NC) and Charged Current (CC) cross-sections very sensitive to Pe at high Q2

Parton densities at large x

Electroweak Physics:

NC qZ couplings: vq, aq

CC W boson (propagator) mass

Physics Beyond Standard Model:

Right-handed CC interaction Enhanced sensitivity for searching for Leptoquarks, Rp violating SUSY

Necessities of having Fast & Precise Pe MeasurementThe Physics Consideration

Example: NC cross-section

Up to Q2 = s = 105

W. Lorenzon EIC, 8 November 2002 22

A Comparison of Different Polarimeters

Expected precision0.7W

Fabry-Perot Cavity

PL e- rate rate (n) (Pe)stat (Pe)syst

LPOL: 33MW 0.1KHz 1000 /pulse 1%/min ~2% pulse laser multi- mode (all bunches) i.e. 0.01 /bc 1%/(>30min) (bc=bunch crossing) (single bunch)

TPOL: 10W 10MHz 0.01 /bc 1-2%/min ~4% cw laser single- mode (all bunches) <2% (cw=continuous wave) (upgrade) New LPOL: 5kW 10MHz 1 /bc 0.1%/6s per mill cw laser few- mode (all bunches) 1%/min (single bunch)

W. Lorenzon EIC, 8 November 2002 23

A High Gain Fabry-Perot Cavity

Use Fabry-Perot cavity to amplify laser power:

All optical components are fixed rigidly on an optical table

S

HERA e beam

(A proven technology at Jefferson Lab)

Linearly polarizedphotons

Circularlypolarizedphotons

W. Lorenzon EIC, 8 November 2002 24

Final

cavity Mount for travel

Fabry-Perot Cavity

W. Lorenzon EIC, 8 November 2002 25

beam pipe

‘laser axis’

bellows

Fabry-Perot Cavity - Details

W. Lorenzon EIC, 8 November 2002 26

beam scanmeasurement

x

y

Intensity

Beam intensity after cavity (Gaussian in principle)

W. Lorenzon EIC, 8 November 2002 27

Laser Cavity Timelines

• Mechanics: – cavity arrived beginning July at Orsay

• August-Sept: vacuum tests– optical mounts and cavity mirror mounts

• Being done in LAL workshop (finished in Sept.)• Feedback and cavity gain

– still low, investigations being done & wait the final system for more more tests (mirror alignment syst. variations)

• Laser polarisation– per mill level almost reached after 1 year of work…

• Test of the final setup: – started in September

• Installation in HERA tunnel:– during March shutdown (16 weeks)

W. Lorenzon EIC, 8 November 2002 28

New LPOL (few-photon mode): Existing LPOL (multi-photon mode):

Compton and Bremsstrahlung edges Up to 1000 produced per pulseclearly visible Signal/background ratio improvedBackground determination and > 5TeV measured in the detector!Calibration easy Calibration difficult(Pe)syst: per mill level expected Non-linearity main syst. error

Photon Detection and Systematic Uncertainty of Pe

S=+1

S=-1

S=-1 S=+1

W. Lorenzon EIC, 8 November 2002 29

Summary on Laser Cavity

New polarimeter with a Fabry-Perot cavity:

High statistics precision: Large gain in cw laser power (~104)

0.1% every 6s (all bunches) 1% every minute (single bunch)

Small systematic uncertainties: Few-photon mode

Per mill level

Fast and precise measurement of Pe:

Valuable feedback for HERA machine to achieve maximum polarization Necessary to fully exploit the physics potential at HERA after upgrade

Cavity prototype being tested at LAL, Orsay Final set-up will be installed during the shutdown in spring 2002

W. Lorenzon EIC, 8 November 2002 30

Polarization after Lumi Upgrade

• “old” configuration: spin rotators for ZEUS & H1 off• First dedicated spin tuning: encouraging progress• What is polarization w/ all spin rotators turned on?

W. Lorenzon EIC, 8 November 2002 31

TPOL/LPOL Ratio

LPOL: operate at beam current > 5 mA to prevent signal to be too close to pedestal

Still under investigation

TPOL:various analysis codes give different answers!

Strategy:investigate ratio for good and bad fills and look for any dependencies

good fill

W. Lorenzon EIC, 8 November 2002 32

Conclusions

• Longitudinal Polarization is currently measured with – 1% statistical uncertainty per minute (for all bunches) – 1.6% systematic uncertainty (with NBW calorimeter)– ~1% systematic uncertainty (with new Sampling calorimeter)

• After upgrade to optical cavity is finished, – systematic uncertainty: 1% or less.– statistical uncertainty: 1-2% for each individual bunch

• Longitudinal Polarimeter measures rate (energy) differences

• Compton Scattering is ideal for Ebeam > 1 GeV

– large asymmetries: A ~ 2*E

– polarization measured/monitored continuously