11 GeV PV M

1
11 GeV PV Møller Detector Considerations BRAINST
ORMING JLab Workshop August 2008 Michael
Gericke and Dave Mack
2
This is a current mode experiment From the
point of view of detector development Every
undesirable effect that we dont design away to
begin with will increase our RMS width in the
signal. Some of them will introduce the
potential for false asymmetries. There are no
cuts (short of beam properties) once the data is
taken. A custom tailored detector set is
paramount ! Simple is better !!
3

• What do(nt) we (I) know ?
• several proposed spectrometer designs (but no
  collimators)
• we (I) have some idea of the focal plane profile
  (FPP) shape
• but no information about rate or q2 variation
  (do we care ?)
• we dont yet know where the background is
  hitting the FP
• These are important factors in determining the
  detector geometry, material and type!
• Nonetheless, what one CAN say about possible
  detectors
• the experiment should be statistics limited we
  want to suppress excess noise (electronic and
  detector geometry), i.e. as close to counting
  statistics as possible
• ideally, we want to be insensitive to anything
  but electrons
• we want something that works (realistically) and
  can be
• funded
• These already constrain to a large extend what
  technology we should use

4
Detector Cause and Effect - Driving Issues
Given by exp. Detector physical Signal
properties design choices FPP Type
(Technology) Yield (light, etc) Rate Geometry
(Shape) Yield uniformity Q2 Active Material Q2
uniformity Rad. Dose Shower Material E, Q2,
spatial resolution Background Readout Backgroun
d rejection Linearity Noise
5
Basic Detector Technology

• We can get some things out of the way
  immediately
• Cerenkov (bare quartz) Rad hard, largely
  insensitive to soft photon background, hard
  to shape, can have low signal (light) yield,
  good noise performance, expensive,
• Cerenkov Shower Calorimeter Rad hard,
  insensitive to photon background, can
  accommodate quartz fibers/rods for odd
  shapes (as in E158), larger excess noise, can
  have much larger light yields, expensive,
• PSICs Rad hard, must have radiator shields to
  remove background sensitivity and increase
  signal yield, inexpensive, handles weird FFP
  shapes, larger excess noise,
• Scintillator Not rad hard enough

Am I missing something ?
6
Focal Plane Profile Shape

• We have 4 (?) spectrometer designs with slightly
  different FPPs.
• The profile shape dictates the minimum detector
  geometry constraints which in turn affects all
  other detector properties
• yield weird detector geometries produce less
  light at the photo-cathode (Cerenkov) (PSICs
  presumably less sensitive to this unless you
  have to do really weird things )
• Y unif. complicated detector geometries produce
  light yield non uniformities across the
  detector
• q2 unif. if the focus is not uniform and the
  rate or light yield is not flat over the FPP
  then extended detector sizes give rise to q2
  bias
• (need better spatial resolution)
• backgr. larger geometries invite more background
  
• All of the above then in turn influence excess
  noise and the yield and q2 non-uniformities
  produce systematic false asymmetries with
  helicity correlated beam effects.

7
Willie Falk 3 toroid design
Put a thin rectangular (?) quartz bar there (a
la Qweak) (20 cm in x) Maybe encase in
tungsten ?
x m
z m
Calculation and plot by Kent Paschke
8
How important is this region? Same rate ? Same
q2 ? There are obvious problems with
interference between neighboring
sections. Keeping these away is a collimation
problem but at what cost in statistics?
Calculation and plot by Willie Falk
9
But is there an e-p radiative tail in here ?
Calculation and plot by Kent Paschke
10
Annulus sections of a PSIC. Or a quartz shower
calorimeter a la E158. This would allow
binning in q2 if the focus is not so good.
2 Toroid Calculation and plot by Willie Falk
11
Kent Paschke Nested Toroids
Candidate for a ring shaped detector again.
12
Krishna Kumar and Luis Mercado quads
Focal plane profile is a ring. Use a set of ring
detectors. (out of what ?)
13
The End
14
Q2 Bias
Average momentum transfer calculated
from collimator apertures and detector geometry.
The photoelectron yield varies with hit
location along the detector ! The Q2
distribution is not uniform across the bar ! How
big is mean Q2 bias introduced by PE weighing ?
15
No NPE Weighting
A detector asymmetry will be calculated by
averaging left and right PMT asymmetries. Q2
bias is troubling in combination with radiation
damage and PMT aging ! Non uniform Q2 bias
across the detector is troubling in combination
with helicity correlated beam motion !
Left PMT NPE Weighting
Right PMT NPE Weighting
Sum PMT NPE Weighting
Back
16
Detector Thickness and Excess Noise
Optimal quartz thickness based on excess noise
simulations at 0 degree tilt-angle. QWeak
Statistical Error Excess Noise Modeled as a
contribution from photoelectron noise and
shower noise Shower activity inside the
detector increases with detector thickness. The
number of PEs will decrease as the detector is
made thinner to suppress shower activity. The two
competing processes lead to an optimal detector
thickness which minimizes the total excess noise.
17
Bialkali Cathode S20 Cathode

Detector thickness was selected at 1.25 cm
Back
18
The 10 keV to 1 MeV photon rate is as high as the
elastic electron rate ! Photons with E lt 10 keV
mostly stopped in detector housing or
wrapping. Photons with 10 keV E lt 1MeV
potentially stopped in the detector. Photons with
E 1 MeV deposit 10. Photons with E 10 MeV
produce 30 of electron Cherenkov light (photon
rate is down by 2 orders of magnitude for E 10
MeV).
electrons
Back
19
Lead Pre-Radiator Study
Can we cut soft photon background using a
pre-radiator? Questions How thick does this
radiator have to be? Can we live with the excess
noise ?
20
Simulate various radiator thicknesses and
establish an ideal thickness that minimizes the
excess noise while attenuating the soft photons
Excess noise a function of photoelectron yield
and shower size
Overall asymmetry error with excess detector
noise
21
Simulations were run for 8 different setups with
the lead radiator thickness varied between 1 and
4 cm. Lead radiation length 0.5 cm Shower max
is reached at 4 radiation lengths --- on
these grounds it is expected that the minimum
in excess noise is reached at about 2 cm
A 2 cm lead radiator would produce about 12
excess noise requiring about 370 hours of
additional running time but keep it in our back
pocket if we end up seeing too much background
with beam.
Back
22
Position Sensitive Ion Chambers (PSICs)

• Fused silica-based Cerenkov detectors are
  expensive/difficult to sculpt to match the shape
  of a crude hardware focus.
• An ion chamber with an optimized pre-radiator is
  very promising
• a clever E158 implementation had
• good time response, good linearity,
• low susceptibility to dielectric
• breakdown.
• Ion chambers are intrinsically rad-hard with the
  signal size determined by geometry and pressure.
• By partitioning the anode into strips, it is
  possible to make detectors with radial
  resolutions of lt 1 cm.
• Cost will be dominated by the electronics.

23
PSICs Minimum Position Resolution

• Simulation
• Ee 4.5 GeV
• 1.9 cm W (5.4 X0)
• (shower max!)
• 10 cm, 1 atm He gas
• Minimum position resolution is a few mm (
  rMoliere)
• Need to control point to point variations in the
  gas column

M. Gericke (U. Manitoba)
24
Simulations of the energy resolution and the
corresponding excess noise for a PSIC detector
with various pre-radiator strengths. The
interplay between the number of shower particles
and the corresponding energy deposition yields an
optimal radiator tickness.
Back