Minutes of the 9 January 2013 Liquid Argon Photon Detector Meeting
Present: Tyler Alion, Mike Kirby, Herb Greenlee,
Tom Junk, Matt Szydagis, Eric Church,
Victor Gehman, Jonathan Insler, Zepeng Li,
Denver Whittington, Brian Baptista,
Stan Seibert, Jim Stewart, Jim Strait, David Muller
Apologies to those omitted.
Ben Jones has been the point person to develop the photon detector simulation and reconstruction for MicroBooNE -- we'd like to follow the steps he took for LBNE. His technical manual is available at
https://indico.fnal.gov/getFile.py/access?resId=0&materialId=0&confId=6188
There are some important differences however -- MicroBooNE has PMT's on racks, while LBNE has WLS acrylic bars and SIPM's. LBNE is also a much larger detector. Most of the strategy should be similar. Ben has taken care to optimize the computing usage of the optical simulation.
Eric described the procedure for making gdml files and Ben's approach for adding optical detectors. It is a straightforward adaptation to make bars instead of ellipsoids, where the locations of the bars are given by the centers in the perl script. Tyler volunteered to add the bars to the geometry, after he is done with the wire-wrapping for the TPC geometry. Time estimate -- 1.5 weeks.
We brought up the issue of the coordinate axes. LArSoft has x along the drift direction, y vertical (against gravity),
z horizontal and perpendicular to the collection wires. Jim Stewart proposed a similar coordinate system, but with z pointing up and y along what LArSoft calls z. Signs adjusted to make everything right-handed. Mike Kirby is in favor of keeping the LArSoft convention as there is software inertia from ArgoNeut and MicroBooNE for the LArSoft coordinate system. Tom suggests translating coordinates, though this could cause more confusion than any convenience advantage. There is an advantage for physicists and engineers to speak in the same coordinates. Mike will write a short note describing the LArSoft convention.
Jim Stewart reminded us of the need to study the effects of shadowing and reflections from the wire planes. There are five -- a grid, U, V, collection, and a ground plane, between the drift volume and the acrylic bars. Approximate wire size is 100 microns, with a wire spacing of 5 mm. Bruce Baller did a study of shadowing. It is
anisotropic as the fraction of the area masked by the wires depends on the angle of incidence. We will probably first just parameterize it as an overall effect on the efficiency of the photon detectors, but will need the number associated with this. Future studies can include more detail on reflections from the wires and anisotropy. Stan can do these studies in CHROMA, but a comparison study with G4 should be done.
The absorption length of UV light in argon is somewhat uncertain. We can start with the parameterization used by MicroBooNE, and there is a measurement from ICARUS -- the absorption length is of the order of 2 meters. It depends on the nitrogen content of the argon. Purification systems do not remove nitrogen and we are
stuck with what comes from the supplier.
The response of the acrylic bars is approximately a falling exponential, with an attenuation length of also about 2 to 2.5 meters. David Muller says after the meeting that it isn't exactly an exponential, and we can easily put in a better function. The mechanical geometry for the LBNE FD's bars is given in David Warner's talk at Houston -- LBNE DocDB 6605. We are interested in studying the impact on physics measurements of varying the geometrical parameters, but the studies are computationally intensive (the MicroBooNE photon library takes hundreds of Mbytes of space and 200,000 hours of CPU to generate). Many studies, like removing acrylic bars or deadening them on one face so they only measure light from one TPC and not two can be done without regenerating the photon library. But adding bars or moving them around requires a regeneration. So we suggest building the first library with twice the acrylic bars and SIPM sensors. Tyler will put these in the geometry.
Zepeng Li volunteers to be the point person for building the photon library. Optimization of the volume definitions of the photon origin needs to be done -- care must be taken to follow the geometrical effects near the APA's and also faithfully reproduce the attenuation. The library only models isotropically emitted scintillation light. Cherenkov light requires two extra dimensions of the library. Modeling photon arrival locations along the bars increases the dimension of the library by one as well. If we discretize the arrival locations into ten bins along the bar length, we get a factor of ten increase in the library size. But if we find that the attenuation length in the acrylic is different from expected, we will have to rebuild the photon library.
All of these tasks need to be done for the 35T prototype. Its geometry is similar enough to the FD geometry that the same perl scripts can be used and a subset geometry put in. An asymmetric drift volume may require a trick or two to incorporate.
Jim Stewart mentioned that there is a plan to put a photon detector into LAPD, and that simulation of this system would help understand the results of this test. It is possible to predict the light yield for cosmics that satisfy an external paddle trigger, and compare this with observations. It would require an extra person to work on this simulation however. The geometry of LAPD is different enough from the LBNE FD and the 35T prototype that the perl scripts cannot be used. It may be simpler to go with a standalone G4 simulation. Discussion with the Indiana group on what we should target to learn from this test will help us plan what we need to simulate. If the signal yield differs from predictions, it could be due to attenuation in the argon, the acrylic, efficiency of the SIPM, or mismodeling of reflections, and it would be nice to be able to extract as much information about the unknowns as possible from the measurements.
There are minutes attached to this event.
Show them.