GPU Enhancements to the Vorpal Plasma Acceleration Module For Modeling Laser-Solid Interactions

The Berkeley Lab Laser Accelerator (BELLA) is a DOE-funded program being constructed for the purpose of accelerating electrons to 10 GeV in a distance of just 1 meter using laser-plasma acceleration techniques. Tape-based plasma mirrors have been proposed to couple the intense laser pulses into each stage of the multistage accelerator, and these devices are being investigated experimentally. Computational modeling of these plasma mirrors is crucial to their successful implementation, but BELLA scientists currently do not have access to software that is capable of modeling laser interactions with such high-density plasmas. We propose to implement a well-respected implicit particle-in-cell (PIC) algorithm into packages for the Vorpal simulation framework, which is currently in use by BELLA scientists. Implicit particle-in-cell algo- rithms have been shown to make possible such simulations with high-density plasmas over long time scales. The Vorpal simulation framework already has a sophisticated, generic field-solving engine capable of im- plicit solutions to a wide variety of physical problems. However, implicit particles are currently missing from Vorpal, as are the necessary implicit coupling mechanisms between the particle and the field solves. In Phase I, we will implement the missing implicit particle algorithm into Vorpal, along with the implicit moment calculations necessary to couple the particle push with the field solve. These algorithms will be well tested and made easy to use within the Vorpal framework. In Phase II, these algorithms will be ported to the GPU to provide optimized, parallel performance and make detailed simulation of high-density plasma phenomena accessible to individuals without access to supercomputers or large clusters. Improvements will be made to the algorithms based on continuing research into implicit PIC methods, and additional capabilities will be added to the Vorpal framework targeted for high-density plasma modeling, including improved models for ionization. Commercial Applications and other benefits: The implicit PIC algorithms needed to model high-density plasma interactions with lasers are also needed for almost all high-density plasma simulations in industry, such as simulations of plasma processing. Plasma processing is estimated to be a $9 billion market, and one where improved modeling capabilities are highly desired to reduce cost and risk in semiconductor manufacturing. These new capabilities will allow Vorpal to be marketed, for the first time, to other industries such as the plasma antenna industry, which has been of interest to the military for some time. These capabilities will also be of strong interest to other DOE-funded research initiatives where high- density plasmas are used, such as in compact neutron and ion sources and laser-ion and laser-proton accel- erators (for such applications as proton therapy).