Simulation of the propagation of scintillation light through argon

This stage of the simulation carries scintillation photons from their birth to their end, either absorbed by some passive material or hitting a sensitive detector and successfully initiating a signal.

Two steps and one half are simulated here:

• where the photons end
• how long they take to get to where they end
• whether they initiate a signal in the optical sensors (which would not belong here but it is in fact for convenience the first step simulated)

Because of the type of the parametrizations used in this stages, the first two steps are simulated independently, so that it is possible to find a scintillation photon born two centimetres away from a PMT, heading straight to it, and still taking microseconds to get there.

Where scintillation photons end

Scintillation photon transportation through the cryostat is parametrized via a lookup table called “photon visibility library”, to which its own whole wiki page is devoted. The total of the scintillation photons in a step is split between all the PMT according to the values of that table, and a large part of them is lost, i.e. not assigned to any channel: those are the ones absorbed in passive material.

When scintillation photons end

big fat to-do

How scintillation photons end

Assuming that a scintillation photon reaches the sensitive surface of a photodetector (in ICARUS, the wavelength-shifter-coated surface of one of the 180 Hamamatsu R5912 PMT within one the cryostat), it has a certain probability to convert into an electric signal (photoelectron) that can be detected. This process includes:

• probability of the TPB layer to absorb the photon and re-emit it as a visible photon (λ ≈ 430 nm); if this does not happen, the PMT photocathode will not convert it
• probability that this emission happens toward the photocathode: the emission is isotripic, so this is a 50%
• probability of this visible photon to cross the rest of the TPB without being reabsorbed and reach the PMT photocathode; our coating is quite thin, so this one is likely high;
• probability of the photocathode to convert that visible photon into an electron (“photoelectron”) inside the PMT; this is quoted by the manifacturer to be ≈25% at λ ≈ 430 nm wavelength.

All these probabilities have been measured together, resulting in what we call “quantum efficiency” (and LArSoft code calls ScintPrescale) of about (12.1 ± 1.0)%.

Since we are simulating million of scintillation photons one by one and at the end only ⅛ does something we can see, the evaluation of the quantum efficiency is actually made as the first step: would this photon convert into a photoelectron if it managed to reach a PMT? This assumes the quantum efficiency to be constant in time and between PMT, which is a reasonable approximation. So for each step only 12.1% of the expected scintillation photons is given the chance of propagating.