Calorimetry: simulation and reconstruction

Status report to GlueX colllaboration meeting
Newport News, Virginia, May 12-13, 2005
Richard Jones, University of Connecticut

General Strategy Forward calorimeter: lead glass array Central calorimeter: lead-SciFi barrel Backward calorimeter: lead-scintillator sandwich

I. General Strategy

• Three types of calorimeters, three different approaches to simulation, reconstruction
• Common strategy in three cases
  1. detailed simulation
     • module-based, not necessarily the complete detector
     • structural geometry described in microscopic detail
     • detailed simulation of detector response
     • standalone simulation code, not integrated with entire detector
     • generally expensive in terms of compute and storage resources
     • only way to understand detector response from first principles
     • useful for optimizing detector parameters and readout
      
     
  2. full-detector simulation
     • geometry more coarse-grained than detailed simulation
     • detail roughly matches the level of readout segmentation
     • full physics simulation, but detector response is modelled
     • optimized for efficient use of compute,storage resources
     • only way to understand integrated detector performance
     • essential for developing reconstruction code
      
     
  3. reconstruction package
     • existing prototypes based on code developed by previous experiments
     • integration into general reconstruction framework yet to come

II. Forward Calorimeter: lead glass array

1. detailed simulation
   • mature simulator exists, developed at UConn
   • incorporates detailed data for lead glass provided by I.U. group
     1. refractive index as a function of wavelength 350-800 nm
     2. absorption coefficient as a function of wavelength 350-800 nm
     3. photocathode efficiency as a function of wavelength 350-800 nm
   • accounts for polarization and wavelength-dependence of:
     1. Cerenkov process
     2. refraction, reflection
     3. absorption in bulk and at surfaces
     4. quantum efficiency
   • predicts photoelectron yield as a function of shower particle:
     1. velocity
     2. direction w.r.t. block axis
     3. position in block
     4. path length
   • main detector issues:
     1. wrapping (absorptive, reflective, air gap)
     2. coupling to phototube
     3. blind spot
    
   
2. full-detector simulation
   • incorporates parameterized response from detailed sim
   • detailed showering (avoids shower library problems)
   • still needs work to match detailed sim at Eγ > 3 GeV (non-linearities) -- not yet in production
   • cvs version simply returns deposited energy sum per block as the hit E value
   • multiple hits treated with a simple two-hit separation cut
    
   
3. reconstruction package
   • existing prototypes based on code developed by E852 and Radphi
   • integration into general reconstruction framework underway at Jlab

III. Central Calorimeter: lead-SciFi barrel

1. detailed simulation
   • program exists
   • work underway by Regina group (Papandreou et.al.)
    
   
2. full-detector simulation
   • present version is rudimentary, but incorporates
     1. detailed showering
     2. exponential attenuation
     3. internal propagation delays at ceffective
     4. energy sums at the module-end level
     5. energy-weighted time values
     6. simple Δt cut on multiple hits
   • anticipated developments include
     1. increased readout segmentation (in cvs, certification pending)
     2. photoelectron statistics in σ(E)
     3. photoelectron statistics propagated into σ(t)
     4. matching to test run data, especially for low-E range 20-70 MeV
    
   
3. reconstruction package
   • existing reconstruction code from KLOE is being dusted off at Regina
   • integration into general reconstruction framework yet to come

IV. Backward Calorimeter: lead-scintillator sandwich

1. detailed simulation
   • program exists
   • work underway by the Florida State group (Eugenio et.al.)
    
   
2. full-detector simulation
   • detector not present in the initial release of HDGEANT
   • detector geometry recently added (in cvs, certification pending)
   • hits generation code in development
   • more news from Florida State group
    
   
3. reconstruction package
   • more news from Florida State group