Lepton Polarisation at HERA

1
Lepton Polarisation at HERA

• A brief overview of lepton polarisation - its use
  and measurement at HERA, including testbeam
  analysis of the performance of the upgraded
  Transverse Polarimeter (TPOL).
• Chris Collins-Tooth (ZEUS, IC-London)

2
Outline

• Lepton Beam Polarisation - how, why?
• Measurement of Polarisation - currently, and in
  the future.
• Test Beam setup of the Telescope and Transverse
  Polarimeter
• What is the Telescope and what does it do?
• What data was gathered?
• Analysis of the data
• Multiple Coulomb Scattering, Beam spread, and
  Telescope resolution
• Relative rotations
• between parts of the Telescope
• between the Telescope and the TPOL silicon
• What can be done about these factors?
• What does this tell us about the TPOL silicon -
  resolution,efficiency?
• Conclusions

3
Lepton Beam Polarisation

• Relativistic e/- emit synchrotron radiation in
  curved portions of a storage ring.
• Emission can cause spin flip.
• and flip rates differ.
• e- become polarised antiparallel to the guide
  field, e become polarised parallel to field.
• P(t) Pst 1-exp(-t/Tst) (N
  -N )/(N N )
• Pst was 0.51 at HERA
• Tsttime-constant 20min

60
P()
50
40
30
20
10
0
-10
t(min)
4
Why is polarisation important?

• More accurate knowledge of polarisation state of
  beam will allow new physics to be investigated
  e.g.
• Standard Model no right handed Charged Currents
  (WL / W-R ).
• Plotting sCCobs(P) allows direct measurement of
  right-handed W-R mass. - present limit is 720GeV
  set in 10/2000 by D0.
• Neutral Current (Z0,g ) cross sections split into
  4 at high Q2 if Leptons polarised.
• This allows measurement light quark Neutral
  Current couplings vu vd au ad .
  (complimentary to LEP b,c quarks).

5
How is polarisation currently measured?

• Transversely polarised leptons collide with
  circularly polarised laser light to give
  angular asymmetry.
• Angular asymmetry translates to spatial
  asymmetry.
• Compton-scattered photons enter calorimeter.
• Calorimeter is in two halves to measure up-down
  energy asymm.
• Polarisations measured to 6 (photon position
  measured to within 1000 mm)
• Upgrade will improve accuracy.

6
What does the prototype TPOL look like?

• Only changes to calorimeter section of TPOL.
• New 1cm2 Si strip detector in front of
  calorimeter (80 mm pitch for horizontal strips).
• Due to small beam spot radiation damage may
  occur.
• Movable scintillating fibre to calibrate Si
  response over 5yr lifetime.
• Production TPOL has 6x6cm2 Si strip detector,
  with horizontal and vertical strips (just tested
  at CERN).

7
Test Beam setup
TPOL
Telescope

• 6 GeV e- beam enters from left
• e- beam passes through Telescope then moves into
  the TPOL
• Telescope mounted as close to TPOL as possible on
  movable table
• Telescope has 3 position sensitive detectors
  T1,T2 and T3 (Td was dead material being used
  for a second experiment)

8
The Telescope

• Previously, degree of polarisation estimated
  using energy asymmetry in calorimeter
  (Calorimeter resolution 1000 ?m)
• Now measure polarisation using 80 ?m pitch Si
• Something more accurate needed to probe TPOL
  silicon resolution - the Telescope.
• 3 planes of 50 ?m pitch Si, with horizontal and
  vertical strips.
• T1,T2,T3 detectors roughly 3cm 3cm (TPOL
  silicon 1 cm2)

9
The data

• T1,T2,T3 used to predict Si strip to fire.
• As expected, fitted line has slope 1 ? 0.01
• Offset simply due to T1,2,3 being physically
  larger than TPOL Silicon
• Width of data about fitted line gives indication
  of TPOL resolution

10
The width

• Width 132.15 2.93 ?m
• TPOL Silicon strip pitch 80 ?m
• Intrinsic resolution 80/?12 ?m

0
0
400
-400
800
-800
11
Analysis of the data

• Observed width does not relate directly to the
  TPOL resolution
• Multiple Coulomb Scattering (MCS) of e- beam at
  T1,Td,T2,T3 and TPOL Aluminium Box
• Finite resolution of the Telescope
• e- beam not 100 collimated
• Misalignments of the Telescope detectors T1,T2,T3
• Misalignments of the Telescope (as a whole) and
  the TPOL

12
Telescope internal misalignment

• T1,T2 and T3 could all be misaligned with respect
  to each other.
• Rotations would produce systematic shifts of
  predicted minus actual strip firing from
  left-to-right
• Most important are rotations about beam-axis
  (pictured). A 0.08o rotation would cause a shift
  of 1 strip across breadth of detector

13
Predicted minus Actual shifts for T3,T2 and T1
from L?R
T3
T2
T1

• Using T1,T2 to predict T3 (left) we observe a
  shift from left to right of approximately 50
  microns

14
Correction of misalignment

• Iterative process invoked
• Rotating T3 by 0.27o flattened off all the plots
  (to within errors)

15
TPOL misalignment

• The TPOL silicon had no vertical strips
• Track through T1,T2,T3 used to predict vertical
  strip to fire to give indication of horizontal
  position of impact
• No discernable shift observed before or after T3
  rotation applied
• Correction for rotations caused no discernable
  reduction in observed width

16
MCS, Telescope resolution and beam collimation

• Simple Monte-Carlo simulation using PDG formula
  for MCS with gaussian width
  ?s(13.6MeV/?cp) ?(x/Xo) (10.038 ?n x/ Xo)
• Telescope resolution and beam collimation are
  small factors in comparison to MCS
• Together, all these factors contribute 102 ?m to
  the width
• Subtracting in quadrature, the TPOL resolution
  obtained is ?(1322-1022) ?83 ?m
• But - MCS is not actually gaussian.
• Attempted to use GEANT to improve estimate

17
GEANT v3.13

• Geant v3.13 used to simulate MCS through 5
  surfaces
• Gives width of 138 ?m 1.5
• cf Observed width 132 ?m 3
• There are other errors on this simulation, 1.5 ?m
  only statistical error.
• Consistent with good performance of TPOL Silicon.

18
Error propagation
T1
T2
Td
T3
Al
T4 (TPOL)
m3
m3
m3?3
z

• Alternative approach to Monte-Carlo
• Use errors introduced by MCS etc., and propagate
  them to the TPOL silicon
• T4T3Z4m3Z4?3(Z4-ZAl)?Al
• ?Var(T4)Var(T3)Z42Var(m3)Z42Var(?3)(Z4-ZAl)2Va
  r(?Al)Z4Cov(T3,m3)
• T3T1-Z1m3-Zd?d-Z2?2
• ?Var(m3)(1/Z12)(T1-T3-Zd?d-Z2?2)
• Variances are calculated, (e.g.
  Var(T1,T2,T3)?2T1,T2,T3 208 ?m)
• ?T4?Var(T4)120 ?m Error on TPOL Si due to
  uncertainties in Telescope
• Resolution of TPOL Si ?1322-120255 ?m

19
TPOL Resolution

• Obviously other effects (eg MCS) mask the
  resolution, so we can only say that TPOL Si
  results are consistent with theory.
• From Monte-Carlo Simulation with PDG formula we
  obtain RTPOL? 83 ?m
• From Error Propagation we obtain RTPOL? 55 ?m

20
TPOL Silicon Efficiency
Ratio of hits NOT registering in TPOL

• Telescope used to predict TPOL events
• Efficiency is the ratio of
• events with TPOL Silicon signal
• events predicted by Telescope
• TPOL Silicon edges found by looking at ratio of
  hits not registering in the TPOL as a function of
  position
• Log plot reveals edges where Telescope predicts
  TPOL hits but TPOL does not register
• Horizontal edges at 13000 20000 ?m
• Vertical edges at 11000 18000 ?m
• Consistent with active area of Silicon

Horizontal Position (?m)
Vertical Position (?m)
21
Efficiency cuts

• Using the edges from previous slide, we must
  remove predicted events from efficiency
  calculations where they miss the boundaries of
  the TPOL Si
• Figures show
• events predicted by the Telescope and registered
  by the TPOL (black)
• For effect, events in red are added. They are
  predicted events which had no TPOL response, but
  the event was inside the opposite direction
  boundary, and so should have registered.
• Clearly, the boundaries look correct.

22
Final Efficiency

• With the positional cuts made, the efficiency of
  the TPOL silicon can be calculated
• This is the ratio of
• events with a TPOL silicon response
• all predicted events inside the boundary
• The efficiency is uniform across the detector, at
  97.8

23
Conclusions

• Upgrade should improve accuracy of asymmetry
  measurement.
• Currently aiming for better than 1 error on P(t)
• Misalignments, Coulomb Scattering and other
  factors contribute significantly to observed
  width of 132 ?m.
• TPOL Silicon resolution TBA but preliminary
  studies suggest 55 and 83 ?m
• TPOL Silicon efficiency spatially uniform at
  97.8
• More accurate knowledge of polarisation state of
  beam will allow new physics to be investigated
  e.g.
• Standard Model no right handed Charged Currents
  (WL / W-R ).
• Plotting sCCobs(P) allows direct measurement of
  right-handed W-R mass. - present limit is 720GeV
  set in 10/2000 by D0.
• Neutral Current (Z0,g ) cross sections split into
  4 at high Q2 if Leptons polarised.
• This allows measurement light quark Neutral
  Current couplings vu vd au ad . (complimentary to
  LEP b,c quarks).

24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
(No Transcript)