At the University of Osaka, an international team of researchers has achieved what particle physicists have long sought — a record-breaking proton acceleration of 132 megaelectronvolts using targets made from graphene, a material only a few atomic layers thick. The breakthrough belongs to a broader human effort to miniaturize the instruments of discovery, bringing the power of particle physics out of vast industrial complexes and into hospitals and laboratories. By pairing an ancient element, carbon, with modern laser physics and artificial intelligence, the team has demonstrated that the path
Graphene targets enable record 132 MeV proton acceleration in laser experiments
Protons continuously gaining energy as they ride the moving wave
Mark
So they got protons to 132 million electron volts. What does that actually mean—is that fast?
Mimi
Nearly half the speed of light. For a proton, that's extremely fast. Conventional accelerators can do this, but they're enormous. This team did it with a laser and a piece of graphene you could barely see.
Luke
But is 132 MeV actually a record? The article says it's a record for this specific technique—laser-driven acceleration with graphene targets. That's a narrower claim than "record proton acceleration."
Mimi
Right. There are bigger accelerators that push protons much higher. But the point is doing it compactly, with less infrastructure.
Mark
Why does graphene matter here? Why not just use any thin material?
Mimi
Most ultrathin materials get destroyed by the laser's prepulse—the weak blast that comes before the main pulse. Graphene is thin enough to let the protons accelerate efficiently, but strong enough to survive the prepulse intact.
Luke
How thin are we talking?
Mimi
Four to sixteen atomic layers. A few nanometers.
Mark
And the "surfing" thing—they're not just blasting the protons?
Mimi
No. The laser creates a wave in the plasma, and the protons ride it for several picoseconds, continuously gaining energy. It's more like acceleration over time than a single kick.
Luke
How confident are we in that interpretation? Is that from simulation or observation?
Mimi
Simulation. The computer models describe it that way. The actual observation is the final energy of the protons and the detector signals.
Mark
Which brings us to the neural network. Why was that necessary?
Mimi
The signals from high-energy protons are faint. Scanning millions of microscope images by hand would be impossible. The AI found the real signals with 99.2 percent accuracy.
Luke
That's impressive, but I'd want to know: was that tested on data the network hadn't seen before? How do we know it's not just memorizing the training set?
Mimi
Fair question. The article doesn't specify the validation method. But they're saying this approach will be essential for pushing higher, which suggests they're confident in it.
O Pulso
- Ultrathin targets have always been the Achilles' heel of compact laser-driven accelerators — they shatter under the laser's own preparatory pulse before the real work can begin.
- Graphene's extraordinary combination of atomic thinness and structural resilience allowed it to survive where other materials failed, keeping the target intact long enough for the main laser pulse to arrive.
- Rather than a single violent energy transfer, protons were found to 'surf' an electrostatic wave through the plasma for several picoseconds, continuously gaining energy in a mechanism that surprised even the researchers running the simulations.
- Confirming the record required a convolutional neural network trained to scan millions of microscope images, achieving 99.2% precision in identifying the faint traces left by rare, high-energy proton impacts.
- The 132 MeV result sets a new benchmark for this technique and opens a credible pathway toward compact accelerators capable of targeted cancer therapy, medical imaging, and laboratory astrophysics.
At the University of Osaka, an international team of researchers has achieved what particle physicists have long sought — a record-breaking proton acceleration of 132 megaelectronvolts using targets made from graphene, a material only a few atomic layers thick. The breakthrough belongs to a broader human effort to miniaturize the instruments of discovery, bringing the power of particle physics out of vast industrial complexes and into hospitals and laboratories. By pairing an ancient element, carbon, with modern laser physics and artificial intelligence, the team has demonstrated that the path to higher energies may run not through bigger machines, but through thinner materials and deeper understanding.
A team at the University of Osaka, collaborating with researchers from Japan, Taiwan, the UK, and France, has accelerated protons to 132 megaelectronvolts — nearly half the speed of light — using targets made from graphene just a few nanometers thick. The achievement represents a record for laser-driven ion acceleration, a field that has long promised a more compact alternative to the enormous radio-frequency accelerators that currently dominate particle physics.
The central obstacle in this field has always been the fragility of ultrathin targets. To accelerate protons efficiently, targets must be as thin as possible — but every laser pulse arrives preceded by a weaker 'prepulse' that destroys delicate materials before the main event. Graphene, a honeycomb lattice of carbon atoms, proved resilient enough to survive this preparatory blast. The Osaka team used targets of just 4, 8, and 16 atomic layers, paired with a relatively long-pulse laser operating at moderate intensity — an unconventional choice that yielded an unexpected result.
Rather than receiving a single sharp kick of energy, the protons were found to ride an electrostatic wave propagating through the laser-generated plasma over several picoseconds. Simulations described this as 'surfing acceleration' — a sustained energy gain that outperformed the brief, violent transfers typical of ultra-short pulse setups.
Confirming the record posed its own challenge. High-energy proton signals are rare and faint, buried among noise across millions of microscope images of detector surfaces. The team trained a convolutional neural network to identify genuine impacts with 99.2% precision — a capability they describe as essential for pushing toward even higher energies in future experiments.
The work sketches a future in which particle accelerators fit inside a single room, serving medical centers with targeted cancer therapies or recreating the interior conditions of distant stars. Graphene has helped close one of the most stubborn gaps on that road.
A team at the University of Osaka, working with researchers across Japan, Taiwan, the UK, and France, has pushed protons to 132 megaelectronvolts—nearly half the speed of light—using a material so thin it exists at the edge of what physics can measure. The breakthrough came not from building a bigger machine, but from solving a problem that has long frustrated scientists trying to shrink particle accelerators down to something a hospital or laboratory could actually use.
Laser-driven ion acceleration has long promised a more compact path to high-energy particles than the massive radio-frequency accelerators that dominate the field today. The applications are real: medical imaging, cancer treatment, and experiments that mimic the physics of distant stars. But there is a catch. To push protons to higher energies, you need thinner targets. The thinner the target, the less material the laser has to work with, and the more efficiently it can accelerate the particles inside it. The problem is that ultrathin targets shatter easily. Every laser pulse is preceded by a weaker "prepulse"—a ghost of the main blast that arrives first and destroys fragile materials before the real work can begin.
The Osaka team solved this by using graphene targets suspended in space, just a few nanometers thick—4, 8, and 16 atomic layers depending on the experiment. Graphene, a single sheet of carbon atoms arranged in a honeycomb, is nearly as thin as matter gets. But it is also remarkably strong. The material could withstand the prepulse and remain intact long enough for the main laser pulse to arrive and do its job. The researchers used a relatively long-pulse laser—1.5 picoseconds, or trillionths of a second—at moderate intensity, a departure from the ultra-short, ultra-intense pulses that dominate this kind of work elsewhere. What they found was unexpected: the protons did not receive a single violent kick from the laser. Instead, they were accelerated over several picoseconds by riding an electrostatic wave that propagated through the plasma created by the laser. The simulations called it "surfing acceleration"—the protons continuously gained energy as they moved with the wave, like a surfer riding a swell that never breaks.
The 132 MeV result is a record for this technique. But the real innovation may lie in how the team confirmed it. High-energy protons are rare, and the signals they leave behind are faint. Detecting them meant scanning millions of microscope images of the detector surface, looking for the tiny marks where individual protons had struck. The researchers trained a convolutional neural network—a form of artificial intelligence—to find these signals automatically. The network reached 99.2 percent precision, distinguishing genuine high-energy proton impacts from background noise with remarkable reliability. This is not a side note. The team says this approach will be essential as they push toward even higher energies, where the signals become fainter still and the noise grows louder.
The work points toward a future where particle accelerators are not cathedrals of steel and magnets, but compact devices that fit in a room. Medical centers could use them for targeted cancer therapy. Laboratories could recreate the conditions of stellar interiors. But that future depends on solving problems like the one graphene just helped crack: how to accelerate particles harder, faster, and with less infrastructure. The Osaka result shows the path is narrowing.
Citações Notáveis
Protons continuously gaining energy as they ride the moving wave rather than receiving a single short kick— Computer simulations of the acceleration mechanism
Quer a matéria completa?
Leia o original em Graphene-Info ↗