Progress Report
Cerenkov detector simulation
Michael Gericke started with the simulation of the
cerenkov light progagation/detection of the detector using a stable
strip down version (v32) of my Qweak_Geant4 code.
This version does not include the magnetic field of the Mini Torus and
the three collimators. The shielding wall is included but with very
generous octant openings.
Geant4 setup details for the cerenkov detector:
- PMT only covers the area of the bar (so it has a retangular shape
instead of a disk/cylinder). This PMT mockup shape was choosen in order
to avoid the detection
of leakage photons escaping the bar and hitting the
photocathode.
- The PMT has a 0.5mm thick entrance window (may be too thin) made from
lime glass. I assigned a refraction index of n=1.52 for lime
glass
independent of the wavelength.
I could not find any document/table about
the refraction index for different wavelength.
- The PMT photocathode is defined as an optical border surface between
the PMT window (lime glass) and the vacuum of the PMT tube. The
electromagnetic properties are
set to "dielectric->metal", because the photocathode is
an alkali metal (Geant4 takes care of the Fresnel Reflection from a
metal).
- There is quasi no light absorption in the PMT entrance window
- PMT quantum efficiency given by Neven
(http://dilbert.physics.wm.edu/elog/Software/42
, see for
QE_XP4572_bialkali_limeglass.dat )
- Fused Silica (SiO2): refraction index , absorption length and
reflectivity values taken from Neven's reports
(optical properties measured by the Babar collaboration for the DIRC
bar)
.
I have added some SiO2 refraction index values for
some wavelengths in order to get a smoother overall description.
Geant4 applies an linear interpolation between two given values.
The SiO2 refraction index can be found on http://dilbert.physics.wm.edu/elog/Software/43
, the online calculator reproduces Neven's/DIRC SiO2 measured
refraction index values.
- The shape of the cerenkov bar ends where I attach the PMTs is
different than the cuts defined by Juliette: http://dilbert.physics.wm.edu/elog/Construction/77
See http://dilbert.physics.wm.edu/elog/Software/79
for some visualization of the cerenkov light propagation
with the nominal tilt angle of +20.8 deg; this means that the detector
is
perpendicular to the track envelope.
The simulation was first done for a straight bar with a tilting angle
of 20.8deg and later on for the V shape under the same condition but
for different tilt angles
The tilting angle is defined to the vertical axis; a positive angle
tilts the detector towards the target (0deg = vertical, 20.8 = perp. to
mean track).
The results of the simulation including the
new handlebar geometry can be seen on http://dilbert.physics.wm.edu/elog/Software/93
.
A picture of the handlebar design can be seen on http://dilbert.physics.wm.edu/elog/Software/100
.
All simulations
are done for a QTOR scaling factor BFIELD
= 1.04 and a detector position at Z=570cm assuming an average
track angle of 20.8deg for the radial
location
Conclusion: The
cerenkov light detection depends strongly on the tilt angle and
detector shape !
Important for Qweak (simulation) is that the average Q2 is not the
average Q2 defined by cerenkov detector hits (e.g. applying 2D cut area
on a plane), because in reality we will measure
an average cerenkov light weighted Q2 "seen" by the PMTs . Therefore we
also weighted in the simulation the Q2 of an electron hitting the 3D
detector with the number of associated photoelectrons.
It seems that the weighting does not change the Q2 distribution much,
which are good news. It means that we don't have to simulate the
cerenkov light for collimator/detector optimization, so we save
a lot of simulation time (2000 primary events including Cerenkov
light : ~3h, w/o Cerenkov light: ~20min).
However in preliminary Geant4
simulation runs, where the collimators are included, the
<Q2> seen by the cerenkov detector/PMTs is higher than the
nominal Q2=0.03,
I see more like Q2=0.04-0.05 with a RMS of ~0.015 (Z=570cm),
*preliminary* result on http://dilbert.physics.wm.edu/elog/Software/101
Collimator 3 issues
In the QTOR drawings review meeting (12-03-2004) Stan Sobczynski
requested a 5cm minimum gap between the last collimator and the
beginning of the QTOR structure.
In the Geant4 simulation I stumbled on tracking ambiguities caused by
overlapping volumes of collimator 3 and the main magnet, see
http://dilbert.physics.wm.edu/elog/Software/95
and http://dilbert.physics.wm.edu/elog/Software/99
.
Question: Is the collimator 3 now finalized ? Maybe collimator 3 can be
embedded in the magnet support structure ?
SolidWorks drawings
I put my SolidWorks drawings on http://dilbert.physics.wm.edu/qweak/Drawings/Qweak
. The subdirectiry Tony contains Target, MiniTorus, Collimators +
support ; the subdirectory Pitt contains
Target, MiniTorus, Collimators + support, and Stan's MainMagnet from
December 2004. There are also STEP files (AP203 and AP214) included, so
Paulo Medeiros from JLab can import
my drawings into IDEAS (to be verified).
That's all folks !