Performance of the OPAL Si-W Luminometer at LEP I-II

1
Performance of the OPAL Si-W Luminometer at LEP
I-II

• G.Abbiendia, R.G.Kelloggb, D.Stromc

a Dipartimento di Fisica dellUniversita di
Bologna and INFN b Department of Physics,
University of Maryland c University of Oregon,
Department of Physics
10th International Conference on Calorimetry in
High Energy Physics (CALOR02), Caltech, Pasadena,
USA, March 25-29, 2002
2
OPAL Si-W Luminometer
3
Precision Luminosity measurement

• LEP ee- collider (CERN) operated for more than a
  decade
• from 1989 to 1995 at c.m. energies close to the
  Z peak (LEP-I)
• from 1996 to 2000 at higher energies (? 209 GeV,
  LEP-II)

LEP-I achieved high statistics of Z0 events 5
106 ev./exp.
Precision on Integrated Luminosity has to be
better than 1 0/00 to match the statistical
errors

• BHABHA SCATTERING
  
• At small enough angles provides a counting rate
  larger than the
• Z0 event rate
• Angular spectrum 1/?3 ? high sensitivity to
  inner acceptance cut

?? ? 10 ?rad (?R ? 25 ?m) ? 10-3 systematic
error
4
Why ? We want to measure cross sections
In particular to measure the
Invisible Ratio
ee- ? Z0 ? hadrons
Sensitive to the number of neutrinos Nn 2.984
? 0.013
And to contributions of extra new physics (for
ex. cold dark matter)
5
The OPAL Si-W luminometer
OPAL Collaboration, Eur.Phys.J. C14 (2000) 373
2 cylindrical calorimeters encircling the beam
pipe at 2.5 m from the Interaction Point
19 Silicon layers 18 Tungsten layers
Total Depth 22 X0 (14 cm)
Sensitive radius 6.2 14.2 cm
from the beam axis
Total area of Si 1.0 m2/calorimeter
Each detector layer divided into 16 overlapping
wedges
Even and odd layers staggered in ?
Cooling pipes as close as possible to the FE
chips to remove 340 W dissipated in each
calorimeter
6
A detector wedge
Silicon wafer 300 mm thick with R-? pad geometry
(32 x 2 pads)
Pitch R 2.5 mm ? 11.25
Glued to a thick-film ceramic hybrid carrying
the FE electronics
Readout with 4 DC-coupled AMPLEX chips (16
channels in a given ? column each)
Depletion voltage 62 V Bias voltage set to 80
V
The complete luminometer has in total 608 wedges
38,912 channels Total area of Si 2.0 m2
7
Front end electronics
For Ee 45 GeV the charge deposited on one
detector layer at shower maximum is 300-400 mips
(? 1-1.3 pC) spread over a few pads
AMPLEX chip

• Dynamic range sufficient more than 1,000 mips

• Low noise better than 101 S/N ratio for mips

Calibration done in many ways (internal,
external,)

• Overall channel-to-channel uniformity in gain
  1 allows
• ? optimum E resolution for trigger
  thresholds
• ? no need of calibration databases
  for offline reconstruction

• Gain variations inside each AMPLEX even smaller
  ? 0.25

• Cross-talk among channels in each AMPLEX at the
  level of
• 2 /channel subtracted (coherent! 30
  /AMPLEX)

Any residual gain variations depending on the
channel position within each AMPLEX are
cancelled by inverting the channel radial
ordering between the two ? columns of each
wedge.
8
Radiation damage
The calorimeters took a lot of radiation from
occasional catastrophic losses of the electron
beams
Protection system monitoring the bias currents
induced a fast beam dump if the absorbed energy
was greater than 3?108 GeV within 1s
Initial average leakage current (at T22 C)
ltIdark gt 1 nA/cm2 After 8 years of operation
(end 2000) ltIdark gt 12 nA/cm2
(at shower
maximum typical values are 5 times higher)
AMPLEX limit 200 nA/pad
Estimate of the absorbed dose in each calorimeter
from increase of the dark current using meas.
from J.Lauber et al., Nucl. Instrum. Meth. A396
(1997) 165
Total Absorbed Energy ? 5?1012 GeV Effective
absorbed dose ? 4 Krad
0.6 of the detector
At the end of LEP running the SW Luminometer had
only 3 dead ceramics, 2 dead amplexes plus a few
isolated dead pads.
9
Upstream material
Near the inner acceptance cut (R7.7 cm) ? 0.25
X0
Left side

• In the middle of the acceptance
• 2 X0 due to cables and
• beam pipe structures

Right side
Effects of degraded energy resolution are
important and corrected for
10
Transverse shower profiles
Dense absorber (W) and compact longitudinal size
Sharp shower core (FWHM lt1 pad 2.5 mm) Broad
tails to almost 10 pads
Peak finding based on 2nd derivative of the pad
signals sensitive to overlapping showers
Inefficiency of cluster finding lt 10-5
11
Two-cluster resolution efficiency
Determined from radiative Bhabha events with 2nd
cluster with E2 gt 5 GeV well separated from
primary one
2nd cluster overlayed pad signals rotated until
?1?2
? ? 50 at ?R ? 1.0 cm (i.e. 4 pad widths)
12
Radial reconstruction
Radial coordinate is first determined in each
layer
Resolution varies strongly across a pad 300
/750 mm at pad boundary / center
Pad maximum on a given layer taken as basic tool,
due to the sharp core of e.m. showers
Then good layer coordinates from 2X0 to 10X0 are
projected onto the reference layer at 7X0 and
averaged there (reference layer is chosen near
the average shower maximum to minimize systematic
effects)
Average R distribution
Periodical structure following the radial pitch
Average R coordinate is influenced by the
non-uniform resolution ? Smoothing algorithm
applied
13
Radial metrology
Key issue for the Luminosity meas absolute
radial dimensions of the detector
Geometry of Si-wafers very precisely defined from
industry (lt 1 mm) Sharp visible pad boundaries
created with photo-lithographic
technique (aluminization)
Very accurate positioning and monitoring of
detector wedges on layers, using microscopes
(with 2 mm precision) and micro-manipulators
Measured radii in each detector half-layer have
RMS 1.3 mm
Final precision on the Absolute average radius
4.4 mm taking into account mechanical
deformations, temperature effects and meas.errors
14
Test beam measurements
Alignment of the calorimeter w.r.t a
high-resolution Si-strip beam telescope
determined using a beam of 100 GeV muons (FE
electronics sensitive to mips) Alternate m beam
with 45 GeV electrons
Position Resolution of the Average Smoothed R
130 mm at pad boundaries 170 mm at pad centers
Pad boundary bias
15
Pad boundary image
Fit with an Error Function
Layer 7X0
parameters w pad boundary
transition width Roff radial offset between
50 probability and nominal pad
boundary
Difference in Roff between e and m ?
pad boundary bias Due to R-? pad geometry
16
Pad boundary bias
Depends on the shower width Parameterized as a
function of w
LEP-I vs LEP-II (at R9.45 cm)
E 96-101 GeV
Shrinkage of the shower core with increasing
energy position resolution should improve
accordingly
E 45.5 GeV
Width (microns)
Depth (radiation lengths)
17
Anchoring R
The reconstructed radial coordinate is sensitive
to the distribution and type of material in
front of the detector as well as to the incidence
angle of the particles Test beam configuration
could not reproduce the exact features of the
OPAL running, so an indirect approach is
followed anchoring
pad boundary image in OPAL data
In the end the inner acceptance cut gets a
correction of ? 5-10 mm with an uncertainty of
? 3.5 mm The outer acceptance cut gets ? 10-20 mm
with an uncertainty of ? 6 mm
18
Energy spectrum of selected events
Summed energies measured in Right and Left
calorimeters
In radiative events with 1 photon from ISR pT
balance implies that

• By cutting on the Acollinearity DR RR - RL
  one can
• reduce the background
• study the energy response and reduce the impact
  of the low E response

19
Energy response
Energy resolution almost constant At LEP-I
DE/E 3.8 / 4.5 (Right / Left cal.) (E ? 45
GeV) At LEP-II DE/E 5.0 (Right and Left
cal.) (E ? 104 GeV)
Differences between Right and Left calorimeter as
well as from LEP-I and LEP-II are due to
different preshowering material
without DR cut
Energy resolution was not critical for the
Luminosity meas.
with DR cut 10 mrad
Systematic error due to E meas is reduced by
almost a factor 3 with the DR cut
20
Final Error on Luminosity
After all the effort on Radial reconstruction the
dominant systematic error is related to Energy
(mostly tail in the E response and nonlinearity)
Quantitatively (OPAL
Collaboration, Eur.Phys.J. C14 (2000) 373)
Systematic Error (?10-4)
Energy 1.8
Inner Anchor 1.4
Radial Metrology 1.4
Total Experimental Systematic Error 3.4 ?
10-4
Theoretical Error on Bhabha cross section 5.4 ?
10-4
21
Summary

• OPAL Si-W luminometer measured small angle
  Bhabha e at LEP
• for 8 years (1993-2000)
• Very reliable and efficient operation in a
  non-trivial background
• environment. Negligible loss of Si-diodes and
  FE electronics.
• Energy resolution 4,
• almost constant from E 45 GeV to E 100
  GeV
• Good efficiency to resolve close lying clusters
  
• ? ? 50? for ?R ? 1.0 cm (4 pads)
• Good S/N ratio for MIPs (10)
• Fine radial and longitudinal segmentation (2.5
  mm ?1X0) allowed a
• radial resolution of 130-170 mm with a
  residual bias ? 7 mm
• This was crucial to obtain the experimental
  systematic error on the
• Integrated Luminosity of only 3.4 ? 10-4