LLNL Researchers Improve Predictive Modeling of Laser Backscatter

Scientists at Lawrence Livermore National Laboratory (LLNL) have improved their ability to predict how laser light interferes with fusion experiments at the National Ignition Facility (NIF). Their research revealed new details about the origins of scattered laser light that can sap NIF’s laser energy, putting experimental results at risk.

Their research, published by Physics of Plasmas and featured on the cover of the journal’s August edition, provides a clearer picture of this stubborn source of energy loss and optics damage from laser light backscatter during fusion experiments.

"We have a much better handle now on the understanding of and ability to predict the magnitude of backscatter in future experiments,” said Andreas Kemp, lead author of the paper, “Modeling stimulated Brillouin backscatter from the inner laser cones during indirect drive inertial confinement fusion experiments at the National Ignition Facility.”

The researchers focused on stimulated Brillouin scatter (SBS), which occurs during fusion ignition experiments at LLNL’s NIF, the world’s most energetic laser. The results also underscore LLNL’s wider strengths in team science and the cumulative efforts of various types of research that Lab scientists constantly conduct while exploring the frontiers of science.

During indirect drive fusion experiments, NIF’s 192 laser beams enter a pencil-eraser sized cylindrical container called a hohlraum, which holds a peppercorn-sized fuel capsule filled with deuterium and tritium (DT).

Beams hitting the hohlraum’s side walls produce a bath of x-rays that cause the fuel capsule to implode, forcing the elements inside to fuse together and generate energy through the same process found in the Sun and the stars.

The first achievement of fusion ignition — when more energy comes out of the process than delivered by the lasers — in a laboratory on Earth occurred at NIF in 2022. The achievement has since been repeated more than ten times.

However, some of that laser light scatters backwards into the laser system, initiating damage to laser optics and reducing the amount of energy that could instead be delivered to the target.

NIF’s laser system delivers energy in two cone-shaped groupings — one inner, one outer— to the upper and lower hemisphere of the hohlraums, precisely adjusting the x-ray fields. That helps control the symmetry of the implosion for a higher energy yield.

Improving the ability to forecast how efficiently energy couples to the fuel capsule is important in a field where even the smallest changes can alter the outcome of an implosion.

The researchers previously studied outer-cone backscatter, which was easier to measure and produced closer agreement between pre-shot modeling predictions and actual experimental results.

The inner cones, however, proved more complicated to simulate within the hydrodynamics physics codes LASNEX and HYDRA used by the experimentalists. Inner-cone models “completely overpredicted the backscatter, not only in magnitude but also how long it was going on,” Kemp said.

To explain overpredictions from 2D simulations, the researchers tried higher-resolution 3D models. Yet the results were the same, “so that wasn’t the reason,” Kemp said.

Then, a separate research effort provided a breakthrough. “Direct Evidence of Multispecies Hydrodynamics in Ignition-Scale Hohlraums,” published in April 2025 in Physical Review Letters, clarified the ways atoms from materials that are typically used in ICF experiments — gold, helium, deuterium, and tritium — mix in their own unique ways.

Previous physics models treated each of those materials as a “single species” that acted in similar fashion, but those simulations also overpredicted neutron yield by about three times greater than what was observed in experiments.

The “multispecies” hydrodynamic research relied on a new wide-field neutron imager to diagnose low-density plasma inside hohlraums to identify where neutrons originated and determine the ion temperatures of those regions. The spindle-shaped imager could capture a wider view of the cauldron of different materials mixing in the extreme conditions inside the hohlraum.

“We had some neutron imagers on NIF, but they're always focused on very small regions,” said LLNL physicist Drew Higginson, that paper’s lead author. “There was no imager that could see the entire hohlraum.”

Lead experimentalist Nobuhiko Izumi developed that imager shape from an idea he conceived in 2002 as a post-doctoral researcher. A deceptively simple question occurred to him while having lunch with his mentor, Richard Lerche, in a Lab cafeteria: Why not turn the aperture “inside out”—replacing the difficult-to-machine void of a high-aspect-ratio aperture with a simple, cigar-shaped solid?

Izumi wrote the spindle shape concept on a paper napkin. About 20 years later, he revisited the idea while reviewing neutron production from images captured in a NIF experiment.

“I saw that as a chance to realize the idea,” Izumi said.

The penumbral neutron imager with a simple tungsten-alloy spindle performed as expected during its first experiment and has helped inform the LASNEX hydrodynamics code. That, in turn, helped Kemp and his research team.

“It allows two plasmas that bounce into each other to not just bounce back, but move through each other,” Kemp said. “Imagine two clouds in the sky and they run into each other. The traditional hydro description would tell you that the clouds will bounce off each other. In this case (with the multispecies model), it’s obvious they will go through each other.”

The revised modeling enabled new measurements of inner-cone backscatter. Repeated experiments “led to a massive improvement in the agreement between the measured values and the simulated values,” Kemp said. “Not only does the magnitude of backscatter shrink, but also the duration of the backscatter.”

Higginson and Izumi also replaced the helium gas used to fill a gold hohlraum with DT gas to better observe nuclear reactions.

That research recorded a strong emission of neutrons midway between the capsule and the hohlraum’s laser entrance hole, Kemp said. “That indicated there’s some kind of density spike developing there late in the implosion,” he said.

Kemp found that the density spike “refracts the laser pulse to a small degree, which then also contributes to the early drop in the backscatter that we were overpredicting earlier.”

Kemp was also part of a different team, led by Sebastian Le Pape, that conducted an experiment at the University of Rochester’s Omega EP Laser System also showing how two different plasmas interpenetrate.

“So now we have three different, completely separate things going on where this plasma interpenetration really matters,” he said. “That’s what makes this paper relevant. Combining these existing numerical techniques in this area of predicting backscatter is a cool, new step in our understanding.”

That understanding has led to techniques such as coating the inside of the hohlraum wall with a low-Z material like tantalum oxide. Low-Z materials contain lighter atoms, while high-Z materials contain heavier atoms.

“The strategy there is to create an environment where not only is the high-Z material like gold blowing off the wall, but you’re mixing it with oxygen,” Kemp said. “And that causes Landau damping to reduce the magnitude of the ion acoustic waves that caused the backscatter in the first place.”

With help from ideas first hatched decades ago, Kemp is looking ahead to when NIF is anticipated to achieve higher laser energies (see “Enhanced Yield Capability Proposal Aims to Boost NIF Yield”) and to a more distant future when a higher fusion yield facility is built.

Higher energy could increase backscatter, so “we have to have confidence in our ability to model this correctly,” Kemp said. “That’s the main value of this, that we’re developing credibility in our modeling abilities step-by-step .”

Joining Kemp and Higginson in the SBS backscatter paper were LLNL colleagues Tom Chapman, Laurent Divol, Eugene Kur, Nuno Lemos, Steve MacLaren, Pierre Michel, David Strozzi, and George Zimmerman.

More information

“LLNL Researchers Employed an AI-Driven Model to Predict Fusion Ignition Shot,” NIF & Photon Science News, August 29, 2025

“Neutron imaging of the deuterium–tritium tamping gas volume in an inertial confinement fusion hohlraums,” Review of Scientific Instruments, October 4, 2024

“Modeling stimulated Brillouin backscatter from outer-cone quads across multiple inertial confinement fusion hohlraum designs,” Physics of Plasmas, April 12, 2024

“Computing Codes, Simulations Helped Make Ignition Possible,” NIF & Photon Science News, April 6, 2023

Follow us on X: @lasers_llnl