John Valleau started his scientific life as an experimental electrochemist and he retained an interest in electrochemistry throughout his career. He was born in Toronto  in 1932 and obtained his first degrees from the University of Toronto; his BSc in the Mathematics, Physics and Chemistry Program, a program that no longer exists and that had a spectacular drop-out rate. Students who survived were qualified to carry out graduate studies in  any of these three fields. He remained in Toronto to complete an MSc degree on experimental electrochemistry. When work on deterministic chaos had become one of the most popular scientific topics in the late 1970s and 80s, he showed us his Master’s thesis from  the 1950s where he observed oscillations in an electrochemical system whose origin was not understood at the time. Perhaps this was an opportunity lost but later he made major contributions to other areas.

After this work his interests turned sharply to Theoretical Chemistry and he went to Cambridge to obtain his PhD degree with Christopher Longuet-Higgins. In Cambridge he developed an interest in statistical mechanics and worked on several topics in kinetic theory of gases, especially on the transport properties of rough-sphere systems. After his PhD he returned to Canada and worked at the National Research Council, again doing theory,  before taking up a position in the Chemistry Department at University of Toronto in 1961. He was the only theorist in the department and felt that he should also establish an experimental research program. He started work on several experimental topics including electrochemistry, ultrasonic relaxation and light scattering. In this period he continued work on the kinetic theory of gases, in particular investigations related to molecular beam studies of reactions being carried out in the department.

From the 1960s his main interest was the development and application of Monte Carlo methods to investigate the structures of liquids and electrolyte solutions. In 1968 he spent his sabbatical year in Paris in the group of Loup Verlet where he not only worked on computational aspects of statistical mechanics but also experienced the Paris May riots first-hand and, as a bystander, was hit by a police truncheon. When he returned to Toronto in 1969 he began to work almost exclusively on equilibrium statistical mechanics related to phase transitions and electrolyte solutions using Monte Carlo and integral equation methods, although for a few more years he maintained a small experimental program.

In the 1970s he carried out extensive research on Markov chain Monte Carlo methods. One of his interests was the computation of free energies by such methods. Simple Monte Carlo methods are very inefficient since some of the configurational states of interest have very low probability. His research turned to the development of methods to solve this problem. This work led to the invention of a new technique called umbrella sampling which was presented in several papers published in 1973-77 written in collaboration with Glenn Torrie. This scheme has the double advantage of speeding up convergence of the Markov chain and allowing efficient estimation of the free energy of the system. It is now one of the most widely used methods for free energy computations and rare event sampling in statistical mechanics and is a standard technique in the field. The name umbrella sampling appears in the title of their paper [G.M. Torrie and J.P. Valleau,  Nonphysical Sampling Distributions  in Monte Carlo Free-Energy Estimation: Umbrella Sampling, Journal of Computational Physics 23, 187 (1977)] and Glenn Torrie recalled how it was chosen. According to Glenn Torrie, John suggested umbrella sampling after rejecting several technical-sounding titles because of its versatility, and not as Glenn (and many of us) thought because of the visual appearance of the broad, flat energy distributions.

Towards the end of this period in the 1970s John terminated his small experimental program and he suggested turning his laboratory space into a set of offices to house the theoretical faculty and students then in the department in one area to stimulate interactions. This beginning led in a few years to the formation of the Chemical Physics Theory Group and finally to its expansion to its present form as a distinct research group in the Department. He was an important influence on how the group functioned for many years. John continued his research into the statistical mechanics of liquids and electrolytes until well after his retirement. He will be remembered as a highly original scientist who made important and lasting contributions to his field.

John had high standards as a teacher and expected a lot of his students. He especially enjoyed teaching Thermodynamics and was keen to make the thermodynamic arguments as rigorous as possible in his lectures. He was especially critical of the ways the conditions for spontaneity were taught and developed new ways of deriving these conditions. Debating various formulations of the Second Law of Thermodynamics was a favourite pastime.

He was interested in social issues and was involved in left-wing politics for much of his adult life. In the early 1980s he became involved in the peace movement, especially through the organization Science for Peace, and was always concerned with developing a more just and equal society. He was a stalwart friend with strong opinions and fond of an argument. He had an amazing laugh, loud and infectious, and even if you were at the other end of the building you knew he was having an interesting discussion. He was never dull.

He was 88 years old when he died at Toronto Western Hospital on May 17, 2020.

Ray Kapral

Stu Whittington