TPOL Test Beam Telescope

1
TPOL Test Beam Telescope

• Chris Collins-Tooth (ZEUS, IC-London)

2
Outline

• Test Beam setup of the Telescope and TPOL
• 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?
• TPOL silicon resolution
• TPOL silicon efficiency
• Summary

3
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)

4
The Telescope
60
P()
50
40
30

• 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)

20
10
0
-10
t(min)
5
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

6
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
7
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

8
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

9
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

10
Correction of misalignment

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

11
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

12
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
• Attempting to use GEANT to improve estimate

13
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

14
TPOL Resolution

• From Monte-Carlo Simulation we obtain RTPOL? 83
  ?m
• From Error Propagation we obtain RTPOL? 55 ?m

15
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)
16
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.

17
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

18
Summary

• 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 uniform at 97.8