On estimating the photoelectron yield and the resultant inefficiency of a photomultiplier-based detector

We present three methods for estimating the photoelectron yield from a photomultiplier illuminated by an LED. We show that the three estimates are consistent. These estimates disagree by ∼ 25% from the popular but incorrect assumption that the pulse height distribution is Poisson-like in the number of photoelectrons.

Further, computations involving Gaussian spreading and/or Landau fluctuations are then performed. These computations should model the behavior of actual scintillator or Cherenkov counters. These computations enable us to estimate how often a counter will miss a count.