Neural network sampling of Bethe-Heitler process in particle-in-cell codes

TL;DR

This paper introduces a neural network approach to efficiently simulate the Bethe-Heitler process in particle-in-cell codes, achieving comparable accuracy to traditional tables but with significantly reduced memory requirements.

Contribution

The authors develop and validate a neural network model that predicts Bethe-Heitler pair production cross-sections, enabling more memory-efficient simulations in PIC codes.

Findings

Neural network matches accuracy of pre-calculated tables.

Memory usage is reduced by a factor of 100.

Performance is comparable in laser-plasma interaction scenarios.

Abstract

This study uses neural networks to improve Monte Carlo (MC) implementations of the Bethe-Heitler process in Particle-In-Cell (PIC) codes. We provide a neural network that is as accurate as pre-calculated tables, and requires a hundred times less memory to store. It is trained to predict Bethe-Heitler pair production cross-sections for atomic numbers 1-50 and photon energies between 1 MeV and 10 GeV in the PIC code OSIRIS. We first validate our approach against a theoretical estimate in a simplified context. We later prove that both approaches have similar performance in a typical relativistic laser-plasma interaction scenario. The large memory decrease accessible with neural networks will enable introducing more advanced cross-section models for Bethe-Heitler pair production and other QED mechanisms in the MC modules of PIC codes.