An advance in the study of laser / plasma interactions

A new 3D Cellular Particle Simulation (CIP) tool developed by researchers at the Lawrence Berkeley National Laboratory and CEA Saclay allows the simulation of laser / plasma coupling mechanisms previously inaccessible to conventional PIC codes. plasma research. A more detailed understanding of these mechanisms is essential for the development of ultra-compact particle accelerators and light sources that can more effectively and cost-effectively solve the long-standing challenges in medicine, in the industry and in the basic sciences.

In laser-plasma experiments such as those of the Berkeley Lab (BELLA) laser acceleration center and CEA Saclay, a French international research center attached to the French Atomic Energy Commission, very large electric fields in plasmas accelerate particle beams of high energies over much shorter distances compared to existing accelerator technologies. The long-term goal of these laser-plasma accelerators (LPAs) is to build collectors for high energy research one day, but many benefits are already being developed. For example, LPAs can rapidly deposit large amounts of energy into solid materials, creating dense plasmas and subjecting them to extreme temperatures and pressures. They also have the potential to drive free-electron lasers that generate light pulses that only last a few seconds. These extremely short pulses could allow researchers to observe the interactions of molecules, atoms and even subatomic particles on extremely short time scales.

Supercomputer simulations have become increasingly critical for this research, and the Berkeley Lab National Center for Scientific Computing for Energy Research (NERSC) has become an important resource in this effort. By giving researchers access to physical observables such as particle orbits and radiated fields that are difficult to obtain in experiments at extremely short time and length scales, PIC simulations played a major role in understanding, modeling and conducting of high intensity physics experiments. However, the lack of sufficiently computationally accurate PIC codes to model the laser-matter interaction at very high intensities has prevented the development of new sources of particles and light produced by this interaction.

This challenge led the Berkeley Lab / CEA Saclay team to develop their new simulation tool, Warp + PXR, an effort launched during the first cycle of the NERSC Exascale Scientific Applications Program (NESAP). The code combines the widely used 3D PIC code, Warp, with the high performance PICSAR library developed jointly by Berkeley Lab and CEA Saclay. It also exploits a new type of massively parallel pseudo-spectral solver co-developed by Berkeley Lab and CEA Saclay, which significantly improves the accuracy of simulations compared to solvers typically used in plasma research.

In fact, without this new highly scalable solver, "the simulations we are currently performing would not be possible," said Jean-Luc Vay, an experienced physicist at Berkeley Lab, who leads the accelerator modeling program in the lab "Applied Physics and Accelerator Technologies". Division. "As our team has shown in a previous study, this new FFT spectral solver provides much better accuracy than finite difference time domain domain (FDTD) solvers, so we have been able to reach certain parameter spaces that are n & ## ## "This new type of spectral resolver is also at the heart of the new generation of PIC algorithms and adaptive mesh refinement that Vay and his colleagues are developing in the new Warp-code. X as part of the Exascale Computing project of the US Department of Energy.

2D and 3D simulations Both critics

Vay is also co-author of an article published on March 21 in Physical examination X which reports on the first comprehensive study of laser-plasma coupling mechanisms using Warp + PXR. This study combined state-of-the-art experimental measurements performed on CEA Saclay's UHI100 laser setup and state-of-the-art 2D and 3D simulations performed on NERSC's Cori supercomputer and Argonne's Mira and Theta systems. Computer installation of the Argonne National Laboratory. These simulations have allowed the team to better understand the coupling mechanisms between the ultra-intense laser light and the dense plasma that it has created, offering new information on how to optimize the sources ultra-compact particles and light. The benchmark tests with Warp + PXR showed that the code is scalable on 400,000 Cori cores and 800,000 on Mira and that it can speed up the problem resolution time up to three times more problems related to ultra-high intensity physics experiments.

"We can not always repeat or reproduce what happened during the experiment with 2D simulations, we need 3D for that," said co-author Henri Vincenti, group scientist of high intensity physics from CEA Saclay. Vincenti led the theoretical and simulation work for the new study and was a Marie Curie Postdoctoral Fellow in the Berkeley Lab's Vay Group, where he began working on the new code and solver. "3D simulations were also very important to compare the accuracy of the new code with experiments."

For the experiment described in the Physical examination X In this article, researchers at CEA Saclay used a high power femtosecond laser beam (100TW) in CEA's UHI100 facility, focused on a silica target to create a dense plasma. In addition, two diagnoses – a Lanex scintillating screen and an ultra-ultraviolet spectrometer – were applied to study the laser-plasma interaction during the experiment. The diagnostic tools presented additional difficulties in studying time and length scales during the course of the experiment, again making the simulations crucial for the researchers' results.

"Often, in this type of experiment, you can not access the time and length scales involved, especially because in experiments, you have a very intense laser field on your target, so you can not place no diagnosis near the target, "said Fabien Quere. , a researcher who directs the experimental program at CEA and is co-author of the PRX document. "In this type of experiment, we examine the elements emitted by the target that are distant – 10, 20 cm – and taking place in real time, essentially, while the physics is at the micron or submicron scale and at the scale of the sub-multimode in time. we need simulations to decipher what is happening in the experiment. "

"The basic principles simulations we used for this research gave us access to the complex dynamics of laser field interaction, with the solid target at the level of the individual particle orbit details, which allowed us to to better understand what was happening in the experience, "added Vincenti.

These very large simulations with a spectral FFT solver of very high precision were possible thanks to the paradigm shift introduced in 2013 by Vay and his collaborators. In a study published in the Journal of Computational Physics, they observed that, when solving Maxwell's time-dependent equations, the standard method of FFT parallelization (which is global and requires communications between processors across the domain of simulation) could be replaced by a decomposition with local FFTs and limited communications to neighboring processors. In addition to enabling a much stronger and weaker scale up on a large number of computer nodes, the new method is also more energy efficient as it reduces communications.

"With standard FFT algorithms, you need to establish communications across the machine," Vay said. "But the new spectral FFT resolver saves time and energy on the computer, which is a big problem for the new supercomputing architectures that are introduced."