Determining plasma parameters from X-ray images in Inertial Confinement Fusion

• Caden Schaeffer, Physics, University of Delaware

• Robert Spiers, Physics, University of Delaware
• Arijit Bose, Physics, University of Delaware

Determining plasma parameters from X-ray images in Inertial Confinement Fusion 

Caden Schaeffer, Robert Spiers, Arijit Bose 

Abstract

Inertial Confinement Fusion (ICF) involves irradiating a small capsule (filled with hydrogen isotopes 2H and 3H) with high-power lasers to compress and heat the fusion fuel. These experiments are done at special laboratories around the United States, like the National Ignition Facility (NIF). At peak compression, the compressed core (hotspot) undergoes thermonuclear ignition and burns through the surrounding dense fuel, releasing net energy. During the peak compression of the capsule,  implosion characteristics are determined by analyzing the emitted X-rays. Many variables affect the emitted X-ray flux of the implosion, such as the radius and temperature of the hot spot, shape of the capsule during implosion, and the degree of magnetization. In this work, we present a Python-based machine-learning framework to recover these implosion characteristics. First, this program pre-processes the X-ray image to autonomously remove any meteors (localized bright spots due to capsule imperfections). We employ an inverse Abel transform on our X-ray image to recover the 3D emission from the hotspot before it was projected onto the 2D image plane. Following this, the program fits the implosion variables to the X-ray flux using a Markov-chain Monte Carlo (MCMC) sampling script, which provides estimated values and uncertainty quantification. The method is tested against synthetic x-ray images and well-benchmarked data from a magnetized ICF campaign at the NIF. This presented approach allows physicists at the NIF to determine important implosion characteristics like temperature, magnetization, and implosion shape from an x-ray image.