The Nobel Prize in Chemistry of 2013 to Karplus, Leavitt and Warshel came to me as a shock. A Nobel Prize for “multiscale models of complex chemical systems”, which then overlooked the true founding fathers of the discipline of molecular simulation? I could not bear it! The three scientists – whose qualities are beyond doubt – were prized for having brought computer simulation to a realm of practical, applied problems, as clearly affirmed by the Nobel committee. The typical pedestrian attitude of a chemist, in my view… Indeed, when I think “computer simulation of atomic systems” my mind would never go to the Karpluses and their like. 

It was August 30, 1956 when, at the Brussels Symposium on statistical mechanics of irreversible processes, a young guy from Livermore National Laboratory named Berni Alder, presented a paper co-signed with his colleague Tom Wainwright, titled Molecular dynamics on electronic computers. It was the first report of a substantial amount of computing time spent on the Univac, in use at Livermore since 1953, and on the more powerful IBM704 in Los Alamos. It took quite some courage to present such a work, in front of the likes of Prigogine, Van Hove, Onsager, De Gennes, Kirkwood, Van Kampen, Uhlenbeck, Lebowitz… It was the birth of molecular dynamics, coupling Ludwig Boltzmann’s ingenious intuitions with the power offered by modern electronic computers.

Berni Alder earned his PhD at Caltech under John Kirkwood (picking him over Linus Pauling… what a life!). Part of his thesis revolved around the calculation of the radial distribution function g(r) for hard spheres. Kirkwood did not spare much trust for the growing hype around numerical computing after WWII, but rather pushed the integral equation approach for the theory of the liquid state, together with Percus-Yevick, Yvon-Born-Green and their YBG equation, and the hypernetted-chain (HNC) diagrammatic technique. (Integral equations were still in use in the mid-80s: the first problem I had to scratch when I started my thesis in nuclear structure, was to calculate the g(r) for liquid Helium, using the HNC representation that J. van Leeuwen had developed in 1959 for his PhD). However, both Alder and Bill Wood (which pushed the Monte Carlo method, invented by Fermi and put on the computer by Metropolis & co.) were students of Kirkwood, and he could just witness this new way of doing physics to develop under his own eyes.

Works and papers followed the Bruxelles presentation, on the background of a sharp competition between the two groups in Livermore and Los Alamos. Heated discussions arose between the “livermorean” Alder and the “los-alamitean” Wood, about a subtle problem in statistical mechanics. In fact, one original discovery by Alder and Wainwright was that a system of hard disks under pressure exhibits a liquid-solid phase transition. (“Big deal”, I hear say my friends chemists… Too bad they don’t understand statistical mechanics.) The importance of such an unexpected result can be traced back to the Mermin-Wagner theorem, which states that a 2D system with continuous symmetry cannot display long-range order. Of course, any system with attractive interactions displays a crystalline state below some critical temperature, but how about hard disks with purely repulsive interactions? Now, numerical simulations were for the first time showing some entirely new physics, in contrast with theoretical predictions. A similar result was also obtained a few months later by Wood and Jacobson, by Monte Carlo simulations, and the two short communications appeared together in 1957, in the same issue of the Journal of Chemical Physics; the full length paper by Alder and Wainwright would appear only three years later, apparently because Kirkwood did not seem to fully believe the story. The apparent contradiction would be explained only years later by Kosterlitz and Thouless, with the introduction of the “hexatic” phase lying somewhere between the solid and liquid, which brought them the 2016 Nobel prize in Physics.

Berni continued to obtain interesting and unexpected results with his Livermore group, such as the discovery of the power-law decay of the velocity auto-correlation function (which is at the basis of diffusion phenomena), instead of the exponential decay predicted by diffusion theory, and by the early ’80s he was fully into quantum systems, producing the famous Ceperley-Alder exchange-correlation functional for the electron gas.The photo attached, permanently pinned to the blackboard of Giovanni Ciccotti’s room in Rome University (where I was assistant in Condensed matter physics, under Giovanni’s guidance, from 1998 to 2004) portraits a young Berni, together with Wainwright and the programmer Mary-Ann Mansigh in Livermore, circa 1962.

In the meantime, atomistic simulations had gained ground, with Aneesur Rahman in Argonne extending the method to soft potentials, in a landmark paper on liquid… Argon (pun intended… This was 1964: when I arrived at Argonne for my first term, in 1995, MD codes there still used his old predictor-corrector integration routine). The molecular dynamics method first landed in Europe thanks to Loup Verlet, who had spent two years in New York with J. Lebowitz, where he developed his famous integration algorithm (actually, a rediscovery of a triangulation procedure devised by Newton). Returning to Orsay, where the new IBM650 was just installed, Verlet started the simulation work in the Theoretical physics laboratory (LPTHE) with D. Levesque. However, the actual spreading of the atomistic simulation methods is mostly credited to the creation of the CECAM international center, where a number of meetings and workshops were organised between 1970 and the late ’80s. There, the “first wave” of B. Alder, A. Rahman, I MacDonald, JP Valleau, M. Kalos and others, started breeding the “second wave” of younger colleagues, like G Ciccotti, K Binder, D Frenkel, H Berendsen, S Yip, WG Hoover, and many others. (People like me could be a kind of “third wave”… In all humility, I could learn all the basic atomistic simulation catechism from Ciccotti in Rome, and from Sidney Yip, during my short term at MIT in 1994 and in Argonne 1995-98, where Sidney regularly spent all his summer terms.) 

And so we come to the second story of today, enter William G Hoover. Around the mid ’80s, Bill was very active in the development of non-equilibrium methods, trying to connect irreversibility with molecular simulations, and had collaborated to the Nosé-Hoover thermostat algorithm. Such ideas were born during CECAM meetings and the Varenna School he organised with Ciccotti in 1985. He tried to put these ideas at work, and wrote several papers with other colleagues, seeking publication in Physical Review Letters and Journal of Statistical Physics. These works however didn’t made it through the refereeing barrier. Some time later, during an air flight, Bill overheard two women chatting in Italian about some man, and using several times the expressions “Che stronzo!” (What an assh**e), “Uno stronzo bestiale” (A hell of an assh**e). Not understanding the language, he was however struck by the peculiar sound. Until in a next CECAM meeting, he met again Ciccotti and asked him the meaning. Once he got acquainted with the proper translation (which – knowing well Giovanni – I can only imagine how colourful could have been), he decided that “dr. Stronzo Bestiale” should be the ideal co-author for his rejected papers. So, he resubmitted the papers just by changing the title and adding some “S. Bestiale, University of Palermo” as last author. And the two papers were eventually published. (see: https://doi.org/10.1007/BF01019693 and https://doi.org/10.1063/1.453402 )

Although I never had the chance of meeting him, I reckon Bill must be quite a character. He entered the Department of Applied Science at Livermore in 1970, and retired as Professor Emeritus in 1993, when he was just 57 years old. In his own words, “The gradual politization and bureaucratization of the laboratory, and the concurrent decline of the University of California toward a rock-bottom level of political correctness, made an active retired research career much more attractive [to me] than continuing service to such failing institutions.” He moved with his family to Ruby Valley, a wilderness area in Nevada, where he also met a Mormon community that he seems to enjoy, though being an unrepentant atheist. He still publishes interesting papers about small-system thermodynamics and ergodicity issues, which he signs as “W.G. Hoover, Ruby Valley Research Institute” (so much the worse for the Shanghai university rankings…)

Two stories in one (and some personal notes)

Post navigation


Leave a Reply

Your email address will not be published. Required fields are marked *