Post on 14-Jan-2016
10/14/2004 SPIN2004 Meeting 1
The MIT-Bates Compton Polarimeterfor the South Hall Ring
W.A. Franklin for the BLAST Collaboration
SPIN2004 Conference Trieste, Italy
October 14, 2004
1. Physics Motivation2. Experimental Apparatus3. Results
10/14/2004 SPIN2004 Meeting 2
BLAST
Collaborating InstitutionsMIT, UNH, Arizona St., Duke, Dartmouth, Vrije University, U.S.
Naval Acad., Ohio, Boston U.
12 Ph.D. students• Neutron Electric Form Factor
V. Ziskin, 10/15 Session 5• e-p Asymmetries on
DeuteriumA. Maschinot, 10/15 Session 5
• Electric-Magnetic Form Factor Ratio of HydrogenC. Crawford, 10/15 Session 5
• Research program in South Hall Ring using Bates Large Acceptance Spectrometer Toroid • Study comprehensively nucleons and light nuclei at low Q2
Experiments underway 2003, running throughout 2004
10/14/2004 SPIN2004 Meeting 3
BLAST Experiment
• South Hall Ring: Intense (175 mA) stored CW polarized electron beams in at 850 MeV
• BLAST Atomic Beam Source: (E. Tsentalovich, 10/15 Session 8)
• BLAST: Symmetric detector with wide momentum transfer bite
• Beam-target polarization product from BLAST asymmetry.• Need rapid nondestructive measurement of beam polarization. • Laser backscattering can provide.
Measure asymmetries using polarized beams and targets
10/14/2004 SPIN2004 Meeting 4
Compton Polarimetry Overview• Compton scattering in highly relativistic frame
Angular distribution compressed into narrow kinematic cone Photon frequencies shifted into gamma regime Detect backscattered photons with compact detector
• Compton scattering cross section • Well known theoretically• Term dependent on electron spin and laser helicity Can extract e- polarization by measuring asymmetries in scattering rates for circularly polarized laser light
Transverse pol. yields asymmetry in azimuthal distribution of scattered photons.Longitudinal pol. yields asymmetry in scattered photon energy spectrum
10/14/2004 SPIN2004 Meeting 5
Compton Polarimetry Below 1 GeV
• Bates seeks precise polarization measurement for each ring fill (15 minutes) for experiments with BLAST.
Bates
JLab
HERA532 nm laser light
• Compton polarimetry is well established at high energy accelerators (Apol~0.5)
• Different challenges exist in applying at energies below 1 GeV. > Analyzing power falling with energy (Apol < 0.05)> Interaction mechanism varies with gamma ray energy> Broader angular distribution for photons> Background from low energy photons> Beam lifetime less than 1 hour
Electron Energy (MeV)
Compton Analyzing Power
Ap
ol
10/14/2004 SPIN2004 Meeting 6
CsI detector
Laser hut
Remotely controlledmirrors
Electron beam
InteractionRegion
Scatteredphotons
SHR Injection Line
RingDipole
BLAST
Laser exit
Laser line
• Design Considerations• Based on NIKHEF Compton Polarimeter• Located upstream of BLAST target to reduce background • Measures longitudinal projection of electron pol.• Scattered gamma trajectory defined by electron momentum
• Polarimeter Layout• Laser in shielded hut with 18 m flight path • Interaction with electron beam in 4 m straight section• Laser mirrors moved remotely • CsI calorimeter 10 m from IR
SHR Compton Polarimeter
10/14/2004 SPIN2004 Meeting 7
Laser System• Laser
• Solid-state continuous-wave, very stable • 5W output at 532 nm
• Optical Transport•Simple, robust lens arrangement for transport to IR and focusing• Mechanical chopper wheel allowing background measurements • Circular polarization by Pockels Cell for rapid helicity reversal• Phase-compensated mirror arrangement
• Interaction Region• 4 degrees of freedom for laser scans. • Laser intercepts stored beam at < 2 mrad• EPICS Control System for slow controls
Y a
ngle
X angle
10/14/2004 SPIN2004 Meeting 8
Scattered Gamma Ray Line
• Align electron beam first to align using bremsstrahlung background. • Movable collimators used to eliminate background from beam halo• Sweep magnet, veto scintillator reject charged particles• Scintillator hodoscope provides position information for beam alignment• Pure CsI calorimeter offers resolution and speed for single photon mode • High rate PMT bases for linear response, stable gain
• Variable thickness stainless steel absorbers for high intensity operation
10/14/2004 SPIN2004 Meeting 9
High Intensity Operation• Measurements made up to 190 mA.
Stainless steel absorbers act as neutral density filter to control rate.
• Signal-to-background tractable at high currents (beam size increases)
• Energy calibration stable on short time scale for high rates in CsI
• Small systematic correction to asymmetry for absorber thickness
• Polarization reduction sometimes observed in raising beam current (tune spreading). • Rapid feedback for retuning Ring.
10/14/2004 SPIN2004 Meeting 10
Data Acquisition
• VME-based system• Rapid digitization
– 100 MHz, 12-bit buffered ADC– External triggering – Single event mode
• High readout and sorting speed
– DMA for high CW event rates (> 100 kHz)
– Pulse shape discrimination and pile-up rejection
– Rapid spin sorting capability
• Integration
– Linked with BLAST analysis
– Synch with BLAST event stream
– Include EPICS information
Fast and reliable data acquisition system is very important
Developed locally (T. Akdogan)
10/14/2004 SPIN2004 Meeting 11
• Analysis begins with raw ADC spectra
– L1 - Laser on, right-handed Pcirc
– L2 - Laser on, left-handed Pcirc
– B1 - Laser off, right-handed Pcirc
– B2 - Laser off, left-handed Pcirc
• Establish energy calibration based on Compton edge and ADC pedestal• Normalization for background subtraction from bremsstrahlung tail• Asymmetry between laser helicities formed as function of gamma energy • Fit asymmetry data with function representing polarimeter analyzing power(GEANT simulation)
Data Analysis
10/14/2004 SPIN2004 Meeting 12
Fill-by-Fill Polarization Results
• Polarization reversed in electron source on fill-by-fill basis• Polarization monitored continuously• Typical precision of 4-5% for ~15 minute fill• Gaussian profile to results
Time (hours)
Pola
riza
tion
10/14/2004 SPIN2004 Meeting 13
Cumulative Results
• Database of results for BLAST experiment in blocks of ~4 hrs• Polarization stable within few percent as a function of time. • Changes usually correlated with beam properties.• Mean polarization (2004): 0.663, (July-Sep, 2004): 0.654• Long term errors determined by systematics
10/14/2004 SPIN2004 Meeting 14
Consistency Checks
• Expected sinusoidal dependence on injection angle• Avg. Injection polarization 0.71 +/- 0.04 (20 MeV transmission pol.).• Measure zero asymmetries with unpolarized beam• Consistency of results for two helicity states
Wien Filter Voltage
Sto
red
Pola
riza
tion
10/14/2004 SPIN2004 Meeting 15
Spin Flipping in South Hall Ring
• Spin flipper
• Reverse direction of beam pol. while stored
• Separate instrumental asymmetries from polarization differences from source
• Adiabatic spin flip
• Rf magnetic field
• Ramp through resonant frequency.
• Performance in South Hall Ring
• Achieved efficiency > 98% (Michigan Spin Physics Group)
• < 1 sec to induce spin flip.
• Spin flip period of ~5 min.
Resonance Scan
frf (kHz)
Pfin
al
10/14/2004 SPIN2004 Meeting 16
Automated Spin Flip
• Spin flipper controlled through Compton DAQ for time synchronization. • Execute flip based on time or beam current • Polarization data sorted from beginning of fill to flip (dark) and post-flip (open)• Eliminates false asymmetry due to geometric effects• Verifies equality of beam polarization for two helicities at injection to < .01
Pola
riza
tion
Time (hrs)
10/14/2004 SPIN2004 Meeting 17
Systematic Error Estimation
•Total systematic error in avg. beam pol. estimated at 0.04. •Working to reduce, sufficient for BLAST experiments.
Small analyzing power makes systematic error reduction crucial
Error Contribution DP
Modeling of Apol and energy calibration
0.03
Pile-up 0.01
Beam misalignment and PITA
~0.05 single pol. state, < 0.01 average pol
Laser circular polarization 0.005
Spin precession uncertainty 0.003
10/14/2004 SPIN2004 Meeting 18
Summary
• MIT-Bates operates a laser backscattering Compton polarimeter during internal target experiments with BLAST at 850 MeV.
• High intensity operation has been successful with stored currents exceeding 175 mA.
• Polarization results are generally stable with time. Certain changes in beam conditions can produce depolarization.
• Systematic error control and calibration uncertainty meet level needed for BLAST experiments. Use of consistency checks provides reduction of systematic errors.