Intercomparison of Low-Energy Electron Transport Calculations By Different Monte Carlo Track-Structure Simulation Codes

Objective. To evaluate the uncertainty of physical features of low-energy electron transport in liquid water due to the use of different Monte Carlo track-structure (MCTS) codes.
Approach. Several MCTS codes developed specifically for liquid water, namely, Geant4-DNA, PHITS-TS, RITRACKS, NASIC, and PARTRAC are used to calculate the electronic stopping power, the pathlength and absorption ranges, the dose-point-kernel, and the frequency-mean (y ̅_F) and dose-mean (y ̅_D) lineal energy for electron energies from 20 eV to 100 keV. Results from MCNP6’s single-event mode, which correspond to unit-density scaled water vapor, are also included for comparison. All the examined codes adopt similar electron transport methodologies but different cross section models and datasets. Data from ICRU Report 90 were used to benchmark the medium- to high-energy performance of the codes. The uncertainty of each calculated quantity was evaluated by the relative standard deviation (RSD) and the maximum relative difference (MRD).
Main Results. For energies above ~1 keV, the RSD and MRD among the MCTS codes for liquid water are, on average, between 5‒25% and 20‒100%, respectively. At lower energies, and especially below 100 eV, the differences are significantly higher with the RSD reaching up to ~70% and the MRD up to ~700%. The MCNP6 results are comparable to those of the liquid water codes, except for energies below 0.1‒1 keV and/or nanometer targets where differences may reach (or exceed) ~100%. Fairly good agreement with the ICRU data is found with the average deviation between 1‒16% and the maximum deviation between 3‒20%, with the exact values depending on the particular code.
Significance: Due to the lack of experimental data, our understanding of electron transport in liquid water is practically limited by the dispersion of the results of MCTS codes. The present results reveal significant differences among MCTS codes at low energies, especially below ~100 eV, potentially compromising the accuracy of DNA damage simulations where such electrons play a key role. The present work highlights the need for further development of the physics models used in MCTS codes to reduce the uncertainties associated with low-energy electron transport calculations in liquid water.