LET estimation in simple geometry for fast neutrons is not consistent

Hello everyone,

I’m trying to simulate dose-weighted LET (Linear Energy Transfer) function for combined G4Tubs+G4Cons solids under various irradiations. I’ve managed to fit PhysicsList for e- (10 MeV) and Co60 sources showing reasonable values within 10% of well published literature data (i.e. 0.20 keV/um for electrons, 0.35 keV/um for Co-60 for G4EmLivermorePhysics), while simulation of 14.8 MeV neutrons gives me 65(2) keV/um in pretty number of PhysicsList combinations which is not consistent with expected 90-100 (e.g. D.K. Bewley 1968 Calculated LET Distributions of Fast Neutrons on JSTOR estimated as 92). Depending on the scoring method for neutrons I was able to reach 82 with QGSP_BIC_AllHP which might be understood as within 10% difference, but this method doesn’t reproduce expected LET for electrons. Are there any more recent data of fast neutrons LET which Geant4 based on?

My code is available at GitHub - mr-xak/g4epindorph: Geant4 simulation of LET inside epindorph which requires header-only Boost library variant for building, so I’ve put some executables in the case you want to reproduce the case or you need info of exact scoring methods.