Design of transient plasma photonic structure mirrors for high-power lasers using deep kernel Bayesian optimisation
Year: 2026
Type: article
Abstract: Ultra-high power lasers are becoming important tools for advancing high-field physics and fusion research. However, their development is constrained by the damage thresholds of conventional optical components; it is challenging to design optical elements capable of withstanding high powers without them becoming impractically large. Here we show that transient plasma photonic structures, formed by the interaction of intercepting laser pulses in gas, can act as compact and robust reflective elements. Because these structures evolve in space and time, and rely on many interdependent parameters, designing optical components using traditional trial-and-error design methods is challenging. We show that machine learning can efficiently explore this complex parameter space to rapidly design robust, high reflectivity plasma mirrors. Moreover, this design process unexpectedly discovers a regime where unchirped laser pulses are compressed. This work demonstrates machine learning as a powerful tool for design, discovery and development of ultra-compact optical components for next-generation lasers. Machine learning is used to design and optimise compact, robust plasma photonic mirrors for ultra-powerful lasers. This approach yields a range of high-reflectivity plasma-based optical components and discovers a new regime for compressing unchirped laser pulses.
Source: Communications Physics
Language: English
Subfield: Nuclear and High Energy Physics
Field: Physics and Astronomy
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $3,990
Awards: G1037004589, The Plasma-CCP Network, G117723289, The Integrated Initiative of European Laser Research Infrastructures, Extension and optimisation of the EPOCH code, G4303546637, G434337789, Multi-scale simulation of intense laser plasma interactions, Lab in a bubble, Multi-scale simulation of intense laser plasma interactions, G6989204204, G8043434492
OpenAlex ID: W7127122907