The influence of different versions of FLUKA and GEANT4 on the calculation of response functions of ionization chambers in clinical proton beams

Objective. To investigate the influence of different versions of the Monte Carlo codes geant4 and fluka on the calculation of overall response functions f(Q) of air-filled ionization chambers in clinical proton beams. Approach. f(Q) factors were calculated for six plane-parallel and four cylindrical ionization chambers with geant4 and fluka. These factors were compared to already published values that were derived using older versions of these codes. Main results. Differences in f(Q) factors calculated with different versions of the same Monte Carlo code can be up to similar to 1%. Especially for geant4, the updated version leads to a more pronounced dependence of f(Q) on proton energy and to smaller f(Q) factors for high energies. Significance. Different versions of the same Monte Carlo code can lead to differences in the calculation of f(Q) factors of up to similar to 1% without changing the simulation setup, transport parameters, ionization chamber geometry modeling, or employed physics lists. These findings support the statement that the dominant contributor to the overall uncertainty of Monte Carlo calculated f(Q) factors are type-B uncertainties.