Laser-driven ion acceleration has long been touted as a potential alternative to conventional accelerators, which rely on kilometers of radio-frequency cavities and enormous infrastructure to push particles to high energies. In principle, an intense laser striking a thin foil can generate electric fields billions of times stronger than those in standard accelerator technology, accelerating ions over microscopic distances. The catch has always been the target. The most established mechanism, target normal sheath acceleration, works best with foils tens of nanometers to micrometers thick, but thinner targets generally produce higher energies because the laser can push electrons through them more efficiently, setting up the strong charge-separation fields that drag ions along.

Herein lies the vulnerability that has frustrated researchers for years. High-power laser systems rarely deliver a perfectly clean pulse. Before the main high-intensity burst arrives, a weak but damaging precursor known as the prepulse leaks through from the amplifier chain. For ultrathin foils, even a modest prepulse can heat the material, expand it, or shatter it entirely before the main pulse ever arrives, destroying the very conditions needed for efficient acceleration. This is why many record-setting experiments have relied on thicker targets or pulse-cleaning techniques, accepting compromises in performance to keep the target intact long enough to be useful.

The Osaka team, led by first author Takumi Minami and senior author Yasuhiro Kuramitsu, attacked the problem from an unusual angle: they chose a material that is simultaneously one of the thinnest and one of the toughest known to science. Graphene, a single-atom-thick sheet of carbon arranged in a honeycomb lattice, combines extreme thinness with remarkable mechanical strength and thermal conductivity. The researchers used multilayer large-area suspended graphene targets, with stacks of four, eight, and sixteen layers, each layer contributing roughly a nanometer or less of thickness. These targets proved robust enough to survive the laser’s prepulse and remain intact until the main pulse, delivered by the LFEX laser system at an intensity of 1×10^19 watts per square centimeter with normal incidence, arrived to do the real work.

What happened next distinguishes this experiment from conventional sheath acceleration. According to simulations performed by the team, the long laser pulse did not merely deliver a single impulsive kick to the electrons and ions. Instead, it generated a propagating electrostatic wave that moved through the plasma created at the target surface. Protons, light enough to respond quickly to electric fields, were able to catch this moving wave and ride it, gaining energy continuously as the wave swept forward. This surfing mechanism extended the acceleration period to several picoseconds, an eternity compared with the femtosecond timescales of typical ultrashort-pulse experiments, and allowed the protons to accumulate far more energy than a single brief push could provide.

By using ultrathin graphene layers and a relatively long laser pulse, we are able to accelerate protons for an extended period and reach a record energy of 132 MeV, Minami explained. Our results show that long-duration acceleration can push proton energies beyond those typically achieved with shorter laser pulses. The statement captures the central insight of the work: duration matters. In laser-driven acceleration, the final ion energy depends on how long the particle remains in phase with an accelerating field. A moving electrostatic wave acts like a traveling booster, and if the wave propagates through the plasma in step with the protons, those protons can surf it for an extended ride, ratcheting up their velocity the whole way.

Confirming that protons had genuinely reached 132 MeV was itself a formidable challenge, and it is here that the experiment took on a distinctly modern character. High-energy protons are vanishingly rare in the particle soup produced when a laser strikes a target, and the signals they leave behind in detectors are faint and easily confused with background noise or with heavier ion species. The team deployed a battery of diagnostic instruments, including CR-39 stack detectors, a Thomson parabola spectrometer, and an electron and ion spectrometer, to characterize the accelerated particles. The CR-39 detector, a solid-state nuclear track detector, records ion impacts as microscopic pits that grow into visible craters after chemical etching, with pit sizes that vary depending on the charge-to-mass ratio of the impacting particle.

The challenge is not only to produce these rare high-energy protons, but also to reliably identify them, Kuramitsu noted. We need to search millions of detector images for signals left by individual ions and distinguish the highest-energy protons from background noise. Manual inspection of millions of microscope images is simply not feasible, so the researchers turned to artificial intelligence. They trained a convolutional neural network, a class of machine-learning model originally developed for image recognition, to classify the etch pits left on the CR-39 sheets after two hours of chemical etching. The pit-size distribution showed two distinct peaks, one corresponding to proton tracks with a charge-to-mass ratio of one and another to heavier ions with charge-to-mass ratios of one-half or less, and the network learned to separate them with remarkable fidelity.

The neural network achieved 99.2 percent precision in one high-energy measurement, giving the team the statistical confidence to confirm the presence of proton signals at the record 132 MeV energy. This marriage of laser physics and machine learning points toward a broader trend in the field. The researchers note that further developments are under way in real-time online ion detectors, and that combining AI-based analysis with such detectors could eventually allow laser experiments to analyze their own results and optimize their own parameters on the fly. The vision is a self-tuning, autonomous laser system in which the loop from shot to diagnosis to adjustment closes without human intervention, dramatically accelerating the pace of parameter scans that currently take weeks of laboratory time.

The implications extend well beyond a single record number. Compact, high-energy proton sources are coveted for applications ranging from proton radiography and cancer therapy research to fundamental studies of warm dense matter and astrophysical plasmas, and conventional accelerator technology is often too large, too expensive, or too inflexible to serve all of these needs. If long-pulse lasers driving durable graphene targets can routinely deliver tens to hundreds of MeV protons, the dream of laboratory-scale ion sources with accelerator-grade performance moves closer to reality. The Osaka result demonstrates that the field’s obsession with ever-shorter pulses may not be the only path forward, and that sometimes the winning strategy is to let the protons catch a wave and ride it, sustained by a target tough enough to hold its ground until the wave arrives.

Subject of Research: Laser-driven proton acceleration to record energies using long-pulse lasers and ultrathin graphene targets

Article Title: Protons ride moving waves to reach record energy with long-pulse lasers

Article References: Protons ride moving waves to reach record energy with long-pulse lasers. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: laser-driven ion acceleration, graphene targets, proton acceleration, long-pulse laser, electrostatic waves, LFEX laser, convolutional neural network, CR-39 detector, Thomson parabola spectrometer, plasma physics, particle accelerators, University of Osaka