Simulation conditions: Reflectvity: 0.99 Bulk Attenuation Length: 2.5 m Sample length: 11 cm Measurement location: 3 cm
I ran these “photon scattering” simulations under the same conditions as the one above when I was running attenuation measurement simulations. The scattering algorithm works as follows (pseudocode):
x=StepLength;
prob=1-e^(-x/ScatterLeng);
if(randUniform(0,1)<prob){ setMomentumDirectionRandom();}
I'm pretty sure this procedure works because when I try to find the modes (optical photons propagating circularly around the edges, see the figures from the previous meeting), I can't see anything well-defined. This seems to indicate that the scattering algorithm is flinging the optical photons out of their modes.
The PE detection results of the sim:
Photon scattering simulations
| Scattering length | Muon Peak (PE) | Error | Sigma | SigErr |
|---|---|---|---|---|
| None (normal) | 1988 | 31 | 300 | 53 |
| 1000 nm | 1896 | 15 | 235 | 22 |
| 200 nm | 1899 | 16 | 253 | 24 |
| 50 nm | 1864 | 19 | 326 | 29 |
I chose to set 50 nm as my lower limit because a quick calculation yielded that for 0.1% Gd-TMHD by mass, the average spacing between Gd atoms (assuming uniform distribution) would be ~20 nm. We know these scatters won't happen every time, so 50 nm is a generous lower limit.
If anything, these particle scatters reduce the total number of PEs detected. We don't really know if this approach accurately represents what we think might be happening in the scintillator. However, this simple test indicates that there doesn't seem to be any benefit to adding random isotropic optical photon scattering inside the scintillator.