Countering space radiation materials damage with machine learning
Jason Rivas is researching materials at the atomic level to improve reliability and resistance of electronics to space radiation.
A PhD student in materials science and engineering at the University of Colorado Boulder, Rivas is tackling a problem that faces any technology that goes beyond the confines of Earth’s atmosphere: damaging bursts of radiation from our sun and other stars.
“We’ve gotten good at shielding electronics,” he said. “But we’re not good at making things radiation hard by design. You can shield anything if you put enough iron, steel and lead around it, but if you add that to a satellite, how many millions of dollars extra is it to launch that weight?”
Working with his PhD advisor, Associate Professor Sanghamitra Neogi, Rivas has earned a major graduate fellowship from the Draper Laboratories in Cambridge, Massachusetts. Through the program, he intends to use computational materials modeling expedited by machine learning to advance the science of space hardened electronics.
The fellowship provides four years of funding for his PhD, as well as access to scientists and engineers at Draper Labs.
“We want to make this research faster and cheaper. Currently the testing requires physically using a neutron beam in a radiation environment. It’s expensive. We think we can change that using with machine learning,” Rivas said.
Developing computational models to map out the effects of ionizing radiation on materials requires exploring the problem at the level of individual atoms.
“We want to determine how much degradation a transistor can stand. It’s called displacement damage. If radiation hits an atom in a material, it displaces that atom, which hits another atom, which hits another atom. How well can that material then return to its original form,” he said.
Tackling the challenges of radiation at an atomic level requires analysis using supercomputers, like CU Boulder’s Alpine system. If the work is successful, it could aid researchers across the spectrum of engineering fields. That prospect is appealing to Rivas, and part of why he chose to pursue a PhD in materials science and engineering: the interdisciplinary nature of the work.
“It’s this intersection of all these different needs. Materials are everywhere. It’s problem solving that means something to the real world,” he said.
Rivas has long been interested in math and science. As a child, he was encouraged by positive teachers and through exploring YouTube channels like 3Blue1Brown, which outlines math problems visually.
“I had a really good calculus teacher in high school. She inspired me. Calculus is just beautiful. Math tells you how the world works,” Rivas said.
Rivas earned his bachelor’s in physics and computer science from Austin Peay State University, which is located near where he grew up in Tennessee. After completing his undergraduate degree, he was drawn to earn a PhD by the prospect of becoming an educator and to break new ground in science.
That eventually led him to Boulder and the materials science and engineering program.
“I want to teach in a college setting. You sort of need a PhD to do that,” Rivas said. “The jobs that come with it are also pretty interesting. Doing research, the problems are self-defined. I get bored doing the same thing everyday.”