Robert W. Zwanzig: Formulated nonequilibrium statistical mechanics.

Robert (“Bob”) Zwanzig died on May 15 this year at the age of 86. He had been a member of the National Academy of Sciences since 1972. He is survived by Frances, his wife of 60 years, a daughter, Elizabeth Bennett, and a son, Carl. Frances writes that Bob “was a baker of bread and of made-from-scratch pizza, a maker of dry martinis and hearty soups, an ice dancer, and a walker along the Chesapeake and Ohio Canal.” We remember Bob as the theorist who inspired us, sometimes criticized us, and more than anyone else, explained the foundations of nonequilibrium statistical mechanics. Bob Zwanzig was born and raised in Brooklyn, New York. He graduated from the Polytechnic Institute of Brooklyn in 1948 and then moved West, first to do experimental physical chemistry with Sidney Benson at the University of Southern California, and then to do theory with John Kirkwood at the California Institute of Technology. Zwanzig received his doctorate from the California Institute of Technology in 1952, about a year after he had moved to Yale with Kirkwood in 1951. That was when Zwanzig was introduced to Frances. She was a firstyear graduate student at Yale. Marshall Fixman, coming to Yale to do postdoctoral research with Kirkwood, first met Zwanzig in early 1954. He recalls hearing stories about Zwanzig and other members of the Kirkwood group from the moment he arrived from the Massachusetts Institute of Technology that January: “Berni Alder and Irwin Oppenheim were mentioned as great warriors . . ., [but] Bob Zwanzig was the current top gun, and it was suggested that murky thoughts should not be expressed in his presence.” Zwanzig had already published an impressive work in 1951 with J. H. Irving on quantum hydrodynamics (1). “I was properly intimidated,” Fixman says. Zwanzig returned from winter holiday in time to see Fixman’s initiation in front of Kirkwood and Lars Onsager, among others. “At my seminar” Fixman recalls, “I talked about my PhD work . . . and seemingly replied to Onsager’s objections with sufficient acumen that [Onsager] remained silent for the rest of the talk. Bob was inordinately pleased by this silence and decided on such flimsy evidence that I was worthy of toleration. . . . Bob and Frances, Bob Mazo and his wife Joan, and I had many good times hiking the hills around New Haven and swimming in Long Island Sound. They brightened [my] life . . .. Bob [Zwanzig] had great clarity of mind, as everyone appreciated, and also great warmth of heart.” Fixman’s remembrances hint at what was then the tradition in statistical mechanics: rapid-fire discussion and persistent questioning modulated by a wholesome desire for understanding. Throughout his career, Zwanzig engaged in this type of activity with enthusiasm and vigor. That year, 1954, Zwanzig published a paper introducing the idea that liquid-state properties could be estimated with thermodynamic perturbation theory (2). The importance of this work was not fully appreciated until computer simulation methods were applied more than a decade later to see how accurate a theory based on this insight could be. Zwanzig’s greatest contributions came after he spent a few years on the faculty at Johns Hopkins and then took a position at the National Bureau of Standards. One of us (H.C.A.) was a visiting graduate student in the Statistical Physics Section of the Bureau in 1963. Zwanzig was in the Physical Chemistry Section, but he would frequently drop in to the Statistical Physics offices to talk with the postdocs about their research. It was clear even to this graduate student that Zwanzig understood basic physics, chemistry, statistical mechanics, and mathematics much better than most scientists. Moreover, Zwanzig enjoyed talking with younger researchers about their interests and he shared his insights to aid them in their research. While