Electron force field for radiation chemistry? A critical assessment of a low-cost approach

The electron force field (eFF) models electrons as Gaussian wave packets whose centers and widths evolve according to classical equations of motion. This unconventional approach allows the simulation of nonadiabatic electron dynamics at a cost comparable to standard molecular dynamics. Herein, we present the first systematic benchmark of eFF for prototypical problems in water radiation chemistry. We show that the original eFF1 model and the eFF-ECP extension do not reproduce a stable ground-state geometry for a water monomer, whereas the eFF2 model recovers the correct monomer minimum, yet still overestimates hydrogen bond energies in water clusters. For dynamical observables, eFF2 captures key features of water electronic stopping for various projectiles (H⁺, He²⁺, Li³⁺, and electrons), including velocity-dependent projectile charge-state populations. It also reproduces the femtosecond dynamics of valence-ionized water monomers and dimers in good agreement with previous multireference ab initio results. However, cross sections describing proton-water collisions are largely underestimated, both for charge-transfer and fragmentation decay channels. Overall, these results establish eFF as a computationally efficient exploratory framework for ultrafast radiation-induced processes. Its primary value lies in revealing qualitative trends and identifying new mechanistic pathways, rather than providing quantitative reaction probabilities.