Dusty plasmas, swirling with charged particles and found everywhere from Saturn’s rings to lab experiments on Earth, are more mysterious than physicists once thought.
In a new study published July 31 in PNAS, researchers used a physics-tailored machine learning model to decode the forces acting between particles in these dynamic systems. Their approach not only captured the complex, nonreciprocal forces that traditional models struggle to explain, but also revealed surprising deviations from long-accepted plasma theories.
Learning Physics From the Inside Out
Dusty plasmas are electrified soups of ions, electrons, and dust-sized particles. These systems are famously difficult to model, especially in real-life lab settings where particles can vary in size, move unpredictably, and are constantly influenced by a chaotic plasma environment.
To make sense of it all, a team led by Wentao Yu, Eslam Abdelaleem, Ilya Nemenman, and Justin C. Burton at Emory University built a machine learning model that does more than fit curves. It incorporates the known physical symmetries and constraints of dusty plasmas into its neural network architecture.
“The model’s accuracy enables precise measurements of particle charge and screening length, identifying large deviations from common theoretical assumptions,” the authors wrote.
How the Model Works
The team recorded 3D trajectories of up to 18 charged dust particles using a scanning laser sheet and high-speed camera. These trajectories served as input for a neural network that learned how the particles interacted, both with each other and with the surrounding environment.
Unlike off-the-shelf machine learning, this model was built to respect the rules of physics. It used:
- A network for interparticle forces, trained on symmetry-aware inputs
- A second network for external environmental forces
- A third for velocity-dependent drag forces from surrounding gas
By comparing these computed forces to actual particle accelerations, the model inferred key physical properties like mass, charge, and interaction strength.
Big Surprises in the Fine Details
One of the study’s biggest revelations: the interaction range, known as the screening length, depends on the particle size. That challenges the longstanding assumption that screening depends solely on plasma conditions.
Even more striking, particle charge did not scale with mass the way classical orbital-motion-limited (OML) theory predicts. In theory, charge should grow with the cube root of mass. Instead, the team found power law exponents ranging from 0.30 to 0.80 depending on gas pressure.
“Even when the particle charge is inferred at the same z-position, where plasma properties should be the same for all particles, the power p can vary substantially from the expected value of 1/3,” the authors noted.
Using Dust to Probe the Universe
The dusty plasma used in this study isn’t just a lab curiosity. Similar particle systems exist in planetary rings, comet tails, and star-forming regions. And by analyzing how charged dust moves in a plasma, researchers can indirectly study electric fields, gas flows, and energy dissipation in those environments.
In this experiment, the model’s accuracy was so strong that it predicted acceleration with an R² value greater than 0.99. And it could infer each particle’s mass using two entirely independent approaches—force drag and interparticle fitting—with close agreement.
Beyond Plasma: A Tool for Complexity
This research isn’t just about dusty plasma. It sets a precedent for how physics-informed AI can reveal hidden laws in any many-body system, whether you’re studying colloidal fluids, migrating birds, or cell behavior.
“Our ability to identify unknown physics from experimental data demonstrates how ML-powered approaches can guide new routes of scientific discovery in many-body systems,” the researchers concluded.
What’s Next?
The findings open several paths forward. Why does screening length scale with particle size? How do ion-neutral collisions in plasma affect charge accumulation? The deviations from OML theory suggest there’s still much to learn about how particles behave under non-ideal conditions.
And as the authors note, “the ability to surpass intuition and avoid biased assumptions is an essential first step in discovering new scientific laws from experiments.”
Journal: Proceedings of the National Academy of Sciences (PNAS)
DOI: 10.1073/pnas.2505725122
Published: July 31, 2025
ScienceBlog.com has no paywalls, no sponsored content, and no agenda beyond getting the science right. Every story here is written to inform, not to impress an advertiser or push a point of view.
Good science journalism takes time — reading the papers, checking the claims, finding researchers who can put findings in context. We do that work because we think it matters.
If you find this site useful, consider supporting it with a donation. Even a few dollars a month helps keep the coverage independent and free for everyone.