Bayesian optimization of laser wakefield acceleration in the self-modulated regime (SM-LWFA) aiming to produce molybdenum-99 via photonuclear reactions

Abstract

While laser wakefield acceleration (LWFA) in the bubble regime demands ultra-short, high-peak-power laser pulses, operation in the self-modulated regime (SM-LWFA) works with more relaxed pulse conditions, albeit at the cost of lower beam quality. Modern laser systems can deliver few-terawatt pulses with tens of femtoseconds at kilohertz repetition rates. These systems are well-suited for developing SM-LWFA applications where high average energy and charge are prioritized over beam quality. Such beams could be used to generate high-energy bremsstrahlung photons, capable of triggering photonuclear reactions to produce radioisotopes like molybdenum-99. This isotope decays into technetium-99m, the most widely used medical radioisotope, with over 30 million applications worldwide per year. This work explores the use of Bayesian optimization to maximize the energy and charge of electron beams accelerated via SM-LWFA. Particle-in-cell (PIC) simulations model a 5 TW, ~60 fs-long (FWHM) Gaussian laser pulse, propagating through tailored hydrogen gas-density profiles. In these simulations, over multiple iterations, the algorithm optimizes a set of input parameters characterizing the gas-density profile and the laser focal position. Three distinct profiles, with lengths ranging from 200 to 400 μm and combining ramps and plateaus, were investigated. Optimal configurations produced electron beams with median energies ranging from 14 to 17 MeV and charges of 600 to 1300 pC, considering electrons with energies above 8 MeV. Preliminary estimates of molybdenum-99 yields for the optimal beams were obtained using their phase space data from PIC simulations as radiation sources in Monte Carlo simulations irradiating a tantalum-molybdenum target.