Calibrating the Electromagnetic Calorimeter

The measurement of the energy response of the electromagnetic calorimeter (EMLAC) was an important part of the E706 physics analysis.  The energy scale calibration dealt with systematic effects due to the physical detector, the data acquisition system, the simulated detector, and the reconstruction package. The cross sections measured by E706 are sensitive to the energy scale calibration.  An incorrect energy scale affects both the normalization and the shape of these cross sections.  The EMLAC energy response was studied using low-p_T pi0's in their two photon decay mode.  Photon energies were adjusted so that the mean pi0 mass matched the PDG value.  Cross-checks included high-p_T pi0's and eta's, the omega (pi0-photon decay mode),  pi0's and eta's where one or both photons converted into e⁺e^- pairs (ZMPs) in the material upstream of the calorimeter (ZMP measurements made with the charged particle spectrometer), and J/Psi's.  The impact of  spectrum on the reconstructed photon energy (~1%) and potential procedural or reconstructor biases were accounted for by calibrating the simulated EMLAC in the same manner as the real detector.  The overall uncertainty in the energy scale calibration was <0.5%.  More information on the calibration may be found in NIM A417, 50 (1998), and the Ph.D. thesis by Michael Begel.

• Systematic Uncertainty
  The systematic uncertainty in the pi0 cross section versus p_T as a function of the uncertainty in the energy scale.
• Pi0 Mass Distribution
  Two photon mass distribution in the pi0 mass region for combinations with p_T>3 GeV/c.
• Pi0, Eta, and Omega Mass Distributions
  Mass distributions in the pi0 and eta (two photon combinations) and omega (pi0-photon combinations) mass regions for combinations with p_T>5 GeV/c.
• Linearity using the Eta
  Eta mass (two photon) compared to PDG value as a function of the eta's energy and p_T.
• Calibration of the Charged Particle Spectrometer
  The tracking system was calibrated to 0.1% with the K_s and J/Psi.
• Pi0 and Eta Mass Distributions
  Mass distributions in the pi0 and eta mass regions for combinations of photons and ZMP electrons at two values of p_T.  The overall masses are low by ~1%.
• Linearity using Converted Photons
  Pi0 and eta mass (photon-ZMP) compared to PDG value versus the energy of the photon and versus the number of radiation lengths traversed in the target.  The energy scale is linear with photon energy, but depends upon amount of target material.  This dependance is consistent with models of bremsstrahlung radiation.
• J/Psi Mass Comparison
  Mass distribution in the J/Psi mass region for e⁺e^- and mu⁺mu^- pairs.  The dielectron mass is ~1% lower than the dimuon mass, consistent with the bremsstrahlung hypothesis.
• Pi0 Mass Distribution (Di-ZMP)
  Mass distribution in the pi0 mass region for ZMP-ZMP combinations with p_T>0.5 GeV/c.  The pi0 mass is ~2% lower than the PDG value, consistent with the bremsstrahlung hypothesis.
• Electron E/P
  Comparison of ZMP electron's energy to momentum.  Both measurements are made following bremsstrahlung losses; E approximately equals P.  

The EMLAC energy response was studied using low-p_T pi0's in their two photon decay mode.  Photon energies were adjusted so that the mean pi0 mass matched the PDG value.  Cross checks included measurements of the pi0 mass at higher p_T, measurements of the eta mass in the two photon decay mode, and measurements of the omega mass in the pi0-photon decay mode.  The eta mass as a function of its energy and p_T was used to examine the linearity of the energy scale calibration.

Photons passing through the material upstream of the EMLAC may convert into an e⁺e^- pair (referred to as a ZMP in this discussion).  The electron's momentum is measured in the charged particle spectrometer (calibrated with the K_s and J/Psi to better than 0.1%).  We reconstruct pi0's and eta's where one photon converted into a ZMP.  The overall masses are low by approximately 1%.  When we compare the reconstructed mass as a function of the photon's energy, we find that the scale is linear.  A comparison of the reconstructed mass as a function of the number of radiation lengths the electrons traversed in the target reveals a strong dependence which is consistent with models of bremsstrahlung radiation.  To check this hypothesis, we compared the J/Psi mass reconstructed in its e⁺e^- and mu⁺mu^- decay modes.  The mass in the dielectron mode is approximately 1% smaller than the mass in the dimuon mode.  Additionally, we can reconstruct the pi0 mass in the charged particle spectrometer for the case where both decay photons converted into e⁺e^- pairs.  The overall mass is approximately 2% low, consistent with the bremsstrahlung hypothesis.  The electron's momentum is measured following the losses due to bremsstrahlung radiation,
and the energy of the electron (measured in the EMLAC) compares well to its momentum (measured in the charged particle spectrometer).  We therefore find that the energy scale measurements using electrons are consistent with those from the photons.

The impact of  spectrum on the reconstructed photon energy (~1%) and potential procedural or reconstructor biases were accounted for by calibrating the simulated EMLAC in the same manner as the real detector.

More information on the calibration may be found in NIM A417, 50 (1998), and the Ph.D. thesis by Michael Begel.