My years with Gary Leaf at Argonne
When I joined Argonne in 1969, Gary, who had received his PhD at University of Illinois Urbana–Campaign, was already on the scientific staff of the Applied Mathematics Division, the precursor of the Mathematics and Computer Science Division at Argonne. We collaborated regularly over the years, until I left Argonne in 2001. I’ll remember Gary as a great colleague, a generous collaborator, and a solid member of the division.
Gary’s interest and expertise were in applied mathematics and what today we would call computational science. Our first collaboration was in neutron diffusion and neutron transport computations in the context of nuclear reactor research. Nuclear engineers at Argonne used finite difference methods to solve the equations, which led to interesting numerical linear algebra problems. But in the early ‘70s, finite element methods were taking center stage, and we explored both their theory and applications in reactor research. In these efforts we were joined by Arthur Lindeman, who had graduated in nuclear engineering from Purdue. The three of us published several joint papers.
When nuclear energy fell out of favor in the late ‘70s, we shifted our attention to bifurcation and stability problems in theoretical combustion. We studied reaction-diffusion equations, explored dimension-reduction techniques using large-activation energy asymptotics, and complemented our analyses with numerical computations. In this effort, we benefited from the expertise of Professor Bernard Matkowsky of the Applied Mathematics Department at Northwestern University.
In the mid-‘80s, high-temperature superconductors were of intense interest in the Material Science Division (MSD) at Argonne. We collaborated actively with Valerii Vinokur, Alex Koshelev, George Crabtree, and their fellow experimentalists in MSD, who were exploring vortex pinning phenomena in superconducting media. Here, the fundamental equations were variations of the Ginzburg-Landau equations, a system of nonlinear reaction-diffusion equations. Again, our work covered both theory and numerical applications, resulting in many publications, in both the mathematics and physics literature.
Our last collaboration was again with scientists in Argonne’s MSD, this time in micromagnetics, where Marcus Grimsditch was our main contact. We studied the Landau-Lifschitz equations, another instance of a nonlinear reaction-diffusion system. Marcus was interested in spring magnets, which consist of alternating layers of hard and soft magnetic material and display useful hysteresis properties.
Looking back over our joint work, I was always impressed with Gary’s intellect, insight, and skills. He appreciated physics and engineering, as well as applied mathematics and numerical computation. He was a master at Fortran, which he considered the workhorse of scientific computing. And, last but not least, he was a supportive mentor of numerous students and postdocs.
Above all, Gary was good company. He was well versed in history and reminded us often of the facts during conversations at the daily lunch table in the cafeteria or during our walks on the Argonne grounds after lunch. I’ll remember Gary with gratitude and respect.