Experimental and simulated H2O2 in pure water under conventional and ultra‐high dose rates for single pulse proton irradiation

Abstract Background Despite the remarkable sparing of normal tissue by FLASH radiotherapy, the fundamental mechanisms that link physics to biological outcomes remain unclear. Among water radiolysis species produced after irradiation, hydrogen peroxide () is a final product resulting from hydroxyl radical ( and reactive oxygen species /) reactions, sources of biological damage. Many experiments have shown that increasing the dose rate, reduces yields, supporting hypotheses related to a transient hypoxia during irradiation. Reproducing experimental data using Monte Carlo simulations can be challenging due to incomplete or ambiguous information about the actual experimental conditions, such as the thorough measurement of oxygen or pH levels. Purpose Through water radiolysis experiments and simulations, we propose to understand the fundamental mechanisms responsible for the production of under varying oxygen and dose rate conditions. A new Geant4‐DNA chemistry module, managing pulse duration, is specifically tested. Methods The purified water samples were irradiated with a 67.5 MeV proton beam delivered by the ARRONAX isochronous cyclotron (IBA Cyclone 70XP) at dose rates ranging from a conventional dose rate (CDR, 0.2 Gy/s) to ultra‐high dose rates (UHDR, ∼6 kGy/s). The concentration of , used as a final endpoint providing insight into earlier processes, was quantified using the Ghormley triiodide protocol. We reproduced irradiation conditions using the GATE and Geant4‐DNA Monte Carlo libraries. Beam alignment and dose homogeneity were verified using a gamma‐index method. A Geant4‐DNA chemistry module (Geant4 11.4‐beta version) was used to calculate the time‐dependent evolution of radiolytic yields for , , , species until 15 min post irradiation, taking into account the oxygen concentration, pH, absorbed dose, and pulse duration. Results Under aerated conditions, for CDR, the simulated (2.18) and experimental (2.13) G() are in close agreement. At higher dose rates, the decrease of G() is very similar between experiments and simulations. Under deaerated conditions, simulated G() decreased from 1.56 at 0.26 Gy/s to 1.22 at 42 Gy/s, with relative differences of 1.3% and 0.8 % compared to the experiment. The impact of content is evaluated through simulation studies and discussed. Conclusions The Geant4‐DNA chemistry module reproduces with a good agreement experimental yields measured under different oxygen levels and dose rates. Both experiments and simulations show an oxygen dependent decrease in G() under UHDR conditions. Simulations indicate an impact of the content at physiological level. As perspectives, we aim at studying the role of hydrated electron associated to the understanding of and effects through time resolved radicals’ measurement.