Portal de Periódicos da CAPES

Editorial: A Tribute to Michele Parrinello: From Physics via Chemistry to Biology

2005; Wiley; Volume: 6; Issue: 9 Linguagem: Inglês

10.1002/cphc.200500427

1439-7641

Wanda Andreoni, Dominik Marx, Michiel Sprik,

Phase Equilibria and Thermodynamics

It is an honor and a pleasure to dedicate this Festschrift to Professor Michele Parrinello on the occasion of his 60th birthday. His various original contributions to modern computational methods and algorithms in the realm of molecular sciences are invaluable and we are confident that they will enjoy an ever growing impact on our understanding thereof. Indeed, his work has already opened many new avenues for the study of processes and properties in diverse fields ranging from liquids to chemical reactions and biochemistry, bridging the gap from fundamental methodology to groundbreaking applications. It is instructive to retrace Parrinello's path in science, from the analytic approach of the statistical theory of liquids to the large-scale computing of electronic systems, from the phenomenological models of ionic compounds to ab initio water, from liquid metals to amorphous silicon to DNA. No matter how complex and branched this path may appear at a first glance into his curriculum vitae, after a more thoughtful analysis one can recognize in it the logic and the drive of the pioneer, moved by an insatiable eagerness for new ways into the complexity of the structure of matter. The "modern" theory of liquids was newly born when, in the early 1970s, Parrinello joined Mario Tosi and later Norman March in their enthusiastic search for an understanding of the growing experimental data on molten salts and liquid metals. The disordered states of matter, from liquids to amorphous solids, would never stop fascinating Parrinello's curiosity and would be the topic of many studies performed with a variety of ever more sophisticated methods up to current times. Parrinello's introduction to computational physics/chemistry happened only later, in 1980, when he went to visit Argonne National Laboratory with the wish to collaborate with Anees Rahman, the pioneer of molecular dynamics simulations of "realistic liquids" (1964). This cooperation marked a striking turning point in Parrinello's career. Although a novice in the field, very soon he contributed the crucial idea which would become the "Parrinello–Rahman method" for molecular simulations in a cell of variable shape. Being rooted in the extended Lagrangian framework just introduced by Hans Andersen (1980), the Parrinello–Rahman method allows structural phase transitions of solids "to happen" during a simulation. The Physical Review Letters article from 19801 marked a milestone in the field of atomistic computer simulations and, together with a couple of subsequent papers,2–4 accumulated well over 1500 citations to date. In the same fruitful period, Parrinello and Rahman also worked out what they called a "convenient tool for the numerical investigation of the properties of an electron solvated in molten KCl"5 by combining Feynman path integrals with molecular dynamics exploiting again the potential power of extended Lagrangians. This piece of work also became a highly cited paper in view of its broad application range for many-body quantum simulations, extending much beyond F centers in molten salts, up to present times. Parrinello's experience at Argonne National Laboratories would soon reveal to be of vast importance. The "computational physicist" part of his soul was born there. Moreover, Parrinello's interest in uncovering the complexity of real materials had been stimulated once for all with molecular dynamics as his method of choice. However, due to the limitations inherent in the simple (semi-) empirical potentials used for the description of interatomic interactions at the time it was difficult to approach realistic systems. Overcoming this obstacle would become Parrinello's goal from that time on. In those days, materials such as semiconductors and metals were the almost exclusive subject of computational solid-state physicists especially busy in interpreting the overwhelming amount of data coming out of experiments on electronic and optical properties of bulk solids and surfaces. In spite of common theoretical grounds, the computational solid-state physicists were to a large extent a separate community from that of computational quantum chemists who were focused on ab initio calculations of atoms and isolated molecules. The activities in both fields were rooted in a rather static view of the world. There was little interaction with the colleagues studying the chemical physics of liquids who were necessarily more interested in atomic motion and dynamics. In the early 1980s solid-state physicists were starting to discover the power of density functional theory (introduced by Walter Kohn and co-workers in 1964 and 1965) in predicting structural characteristics of a number of materials and were slowly moving to applications in the realm of clusters and even molecules. Parrinello's encounter with Roberto Car in Trieste, who had a strong background in DFT and electronic structure calculations, was providential. Their first joint paper "Unified Approach for Molecular Dynamics and Density-Functional Theory",6 submitted in August 1985 to Physical Review Letters, marked the beginning of a new era in atomistic simulation of real matter. More than 3600 citations are reported to date, which interestingly come from a large variety of sources. Car and Parrinello's seminal paper is one of the Physical Review Letters articles with the highest interdisciplinarity of citations ever. Indeed an ideal platform emerged from it for exchange of views and collaboration between computational physicists and computational chemists towards a common perspective. At the very heart of this novel method was again the extended Lagrangian concept, this time electronic orbitals playing the role of the added auxiliary fields. Thus, Car and Parrinello not only set the ions but also the electrons in motion by classical molecular dynamics. The scientific success came immediately, their ingeniously elegant method became very soon known as the (often misspelled!) "Car–Parrinello method" (CP method), and the official awards arrived only a few years later. Roberto Car and Michele Parrinello were awarded the "Hewlett-Packard Europhysics Prize" of the European Physical Society (1990) and the "Rahman Prize" of the American Physical Society (1995). Although born in condensed matter physics—silicon and its close relatives being "The System"—the Car–Parrinello method has been proven to be a most powerful "tool" for the study of chemical processes. Thus, starting in the early 1990s, during his time at the IBM Zurich Laboratory, encouraged by Heini Rohrer, Parrinello shifted the focus of his interests towards the world of chemistry. The time was ripe for it. Major progress achieved in the implementation of DFT—like the generalized gradient approximation of the exchange-correlation functional—had started to provide an accurate and efficient methodology to deal also with molecules and had thus recently convinced also many quantum chemists to embrace it. With this improvement in the functional, the Car–Parrinello method was becoming ready to treat complex and condensed molecular systems opening up the field for applications to realistic chemical systems. Chemistry brought with it a fundamental challenge, the prediction of reaction rates, and as such it caught Parrinello's attention. This meant the precise elucidation of the underlying reaction mechanisms, and calculation of activation free energies. Using the CP method combined with constrained molecular dynamics and thermodynamic integration, a step forward was made by Parrinello and his group in the study of a number of well-known chemical reactions in the mid 1990s. Moreover, this enabled Parrinello and his associates at IBM to engage in collaborations with chemical industry, which he would continue also after leaving the company and moving to the Max-Planck-Institut für Festkörperforschung in Stuttgart. In particular, such collaborations led to the pioneering dynamical studies of polymerization reactions7, 8 and Ziegler–Natta heterogeneous catalysis.9, 10 The time at the Max Planck Institute, where Parrinello enjoyed freedom, support and ample resources, was extremely fertile for his work in chemistry. Intrigued by the power of David Chandler's "transition path sampling",11, 12 Parrinello went back to the very roots of classical mechanics,13 a subject that he is proud of having studied in all its glorious detail as a student. A major leap forward, however, was made more recently by the introduction of the metadynamics method14 by Parrinello's group in Lugano. By using only minimal a priori knowledge of the ongoing processes, this method enables one to investigate reaction mechanisms and to compute free energy landscapes. As Parrinello put it in reply to a somewhat nasty question at a conference: "I now moved at least from Newton to Gibbs - from the 17th to the 19th century …︁". A number of extensions, in particular the elegant and easy-to-use extended Lagrangian formulation,15 and several, again, pioneering applications have followed that original article. To mention just one example, a recent combination of metadynamics with the Parrinello–Rahman method has proved capable of exploring and even predicting the crystal structures of truly complex molecular solids. Surely, this is only the beginning of another success story with major implications much beyond chemistry. For his contributions to the field of chemistry Parrinello has been awarded several prizes, the "Boys–Rahman Prize" of the Royal Society of Chemistry (1994), the "Award in Theoretical Chemistry" of the American Chemical Society (2001), the "Schrödinger Medal" of the World Association of Theoretically Oriented Chemists (2005) to name a few. In his scientific oevre—currently about 400 original research papers that are cited 20 000 times with an ever growing impact per year—Parrinello has studied a large number of different systems, but there is one to which his name is especially linked: water, the compound that is most essential to life on Earth. A little over 20 years after the seminal "Molecular Dynamics Study of Liquid Water" by Aneesur Rahman and Frank Stillinger appeared in the literature (1971), the pioneering "Ab initio" liquid water simulation16 was published by Parrinello and co-workers. That landmark study is the beginning of a new era for the research on water, solvation, liquid state chemistry, and hydrogen-bonded systems in general in which computer simulations are taking a leading role not only for the deepening of the microscopic understanding but also in the prediction of unforeseen properties and phenomena. Again, the response of the community in terms of citations, currently over 1500, of this hallmark and related publications on charge defects in hydrogen-bonded system is unambiguous.17–22 The extension of Car–Parrinello method to realistic aqueous systems has had another far-reaching implication, not only for the progress of Parrinello's own research, but also for scientific computing in general. Indeed, with this achievement Parrinello's work has opened the way to merge density functional electronic structure calculation and molecular dynamics in a powerful computational tool for the study of the highly complex systems of biochemistry and biophysics. His interests in this field started already in the mid 1990s and included an early implementation of a dynamical plane wave-based quantum mechanics/molecular mechanics (QM/MM) scheme coupling DFT to biomolecular force fields. We believe that this extension of the Car–Parrinello methodology to biological systems, that has had a number of proselytes since, has the potential for a huge impact on their understanding on the atomic scale. Tracing back his scientific trajectory, a characteristic feature certainly is Parrinello's quest for fresh and nontraditional computer simulation methodologies and algorithms. The source of his success is undoubtedly his broad overview and synthetic approach to computational science which is manifested in a sometimes uncanny intuition of what is feasible and what is not. Being the "maestro of extended Lagrangians", having introduced quite a few successful ones,1, 5, 6, 23–27 the Car–Parrinello Lagrangian without doubt can be considered the master piece. Many of his technical advances opened up new avenues to address old and outstanding questions in the molecular sciences. As an additional illustration we mention an igniting paper on O(N) linear scaling methods in electronic structure calculations.28 Indeed, the vision of "…︁problems where multiple length scales are important"—now so much on fashion in electronic structure calculations—was already the driving force behind a publication from 1987 in which Car and Parrinello proposed a real-space CP approach23 using nonuniform meshes. Also the more application-oriented work of Parrinello often contained a significant innovation in the field, for example, his calculations of spectroscopic (IR, Raman, NMR,…︁) properties of molecular solids and liquids, which combined the variational perturbation methods, developed earlier and applied in solid-state physics, and the "modern theory of polarization". These outstanding achievements in the physical sciences are honored by being elected member/fellow of the American Physical Society (1991), the International Academy of Quantum Molecular Sciences (1995), the Berlin-Brandenburgische Akademie der Wissenschaften (2000), the Royal Society (2004), the European Academy of Sciences (2004), and the World Association of Theoretically Oriented Chemists (2005) among others.1 So far we have only retraced a few milestones in Michele Parrinello's career, leaving much of his other achievements that materialized in more than 50 "Highly Cited" papers to the attached publication list. However the reason why we dedicate this set of research articles to him is not only our wish to respectfully acknowledge his many scientific achievements, but also the pleasure to take an opportunity to write about our friend Michele, our collaborator, colleague and mentor. That complex and still well-structured path we have recognized in the development of his career has been accompanied by an analogous path in his personal life. It is often said that people born on the ocean feel a strong desire to leave, travel and explore new shores. This is certainly true for Michele, who still very young decided to leave that warm sunshine of his Sicily to look for new opportunities in the northern city of Bologna, where he received his doctorate in Physics, the Laurea, with a thesis on particle physics using relativistic quantum field theory (1968). After his return to his hometown Messina (1970), Michele was there happily reborn as a condensed matter theorist thanks to the charismatic influence of Mario Tosi. Still his stay was not bound to last long. Although a source of new interests and inspiration as well the beginning of friendly and long-lasting collaborations, the experience at the University of Messina was fertile for new ambitions and dreams for Michele. Only a few years later, in 1975, he was ready to leave again, towards the gray skies of London to work with Norman March at Imperial College (1975–1976). His friends in Messina—Tina Abramo, Carlo Caccamo and Paolo Giaquinta—had hoped to keep him there when he came back from England in 1976 but this was only an illusion. As Paolo would say many years later "Michele will never sit and stop moving if a new challenge comes up". Indeed at that time Michele accepted to start a new life experience in Trieste (1977) where Mario Tosi and Erio Tosatti would soon join him, moving from Rome to build up the condensed matter division of the International Centre for Theoretical Physics, ICTP. But also Trieste would not be his home for long. When in 1980 Michele decided to visit Aneesur Rahman in Argonne, he did not know that this event was going to be the turning point of his life and career. He had the intention to stay for three months only but ended up to remain for two years, thanks to their successful collaboration. Today, an inconspicuous photograph of Anees Rahman working at his terminal is a fixed-point in Michele's various offices. An ever growing fame accompanied him on his return to Trieste, where a few years later he would invent the ab initio molecular dynamics method with Roberto Car. In spite of the pleasant and stimulating atmosphere of the scientific community in Trieste and the genuine attachment he had to his friends there—above all to Erio Tosatti—soon was he ready to pack and leave Italy again for many years. First at IBM Zurich Research Laboratory (1989-1994), where he had already been visiting twice (in the summers 1987 and 1988), then at the Max-Planck-Institut für Festkörperforschung in Stuttgart (1994–2001), then again in Switzerland as Professor at the ETH Zurich. However this time he would settle more south—very close to Italy. Initially at the Swiss Center for Scientific Computing at Manno (close to Lugano) that he led up to 2003, and more recently sharing his time between the computational science group he has founded in Lugano (at the campus of the Università della Svizzera Italiana) and the Scuola Normale Superiore in Pisa. Looking back, his affiliations appear to be, in a certain sense though, a secondary issue. Michele has always been able to create his "own local universe" by attracting excellent students, post-docs and visitors wherever he had decided to settle. Nobody can predict whether other shores will ever attract him again, but now one can say that he has, at least part-time, returned to Italy and his presence is strong again in the Italian scientific community. Indeed, in his heart, Michele has never left his country. However fascinated he might have been by one aspect or the other of the different countries he has lived in, his roots have always prevailed. Few people know how closely Michele has always liked to follow what was going on in Italy, from the major events of its cumbersome political scene to minor stories, from the latest movies to the most popular songs. In fact one of his main concerns in any foreign place was where to buy his beloved Corriere della Sera. A number of settings have characterized Michele's life. In each one he has certainly left an indelible trace and each one must in turn have left an indelible memory in him. Probably the most unforgettable office is the one he shared with his close friend Erio Tosatti at the ICTP. The two of them were often engaged in the discussion of fundamental questions of physics, however mixed with jokes with a distinct Italian twist and laughs that were resonating in the corridor at the second floor of the ICTP building, attracting crowds of visitors, post-docs and students. Whoever has visited Erio and Michele during those times cannot forget that high, metastable barricade, wholly made up of paper, in the form of articles, books and scrabbled sheets, which seemed to be there to discourage anybody from trying to disturb. The drive to enter and join them was strong though and Erio was kindly shifting piles of papers from one spot to the other, as sign of hospitality, but ended up generating paper avalanches and "contributing to increase the entropy of the universe" (as Erio would say). Some stories become legend with the years. There is a famous one about Michele and Roberto from that summer 1985 when they were polishing their first CP code and writing the final version of their renowned Physical Review Letters article—the 5th most cited PRL paper ever. They were staying at the IBM Research Laboratory in Yorktown Heights, visiting Socrates Pantelides. They used to work late at night and when they finally returned home, in the darkness of Westchester County, tired and excited, talking and laughing, they were involuntarily transforming from the kind scientists of the day into the frightening night-killers for all the animals of the region who were previously used to cross safely those same roads. Apparently they could not avoid it and still tell about it with horror mixed to a sensation of adventure. "To Michele's discharge—says Roberto—I was at the steering wheel during these trips…︁but it is only a partial discharge because Michele was clearly my accomplice." Michele's sense of humor, his jokes and witty remarks is something we appreciate and like very much in his personality. We have enjoyed working with him on many projects, sometimes originating from a "fuzzy equation", that is, an idea, written on the blackboard in the huge letters of his very characteristic hand writing. We like to remember with a smile that, during intense phases of coding, the worn-out ESSL manual was just called "The Book" by Michele, and how closely his face approached the screen when helping to search for programming errors or when looking carefully at plots. Typical is also his intensely focused interest in the progress of only one or two projects at a time, and the way he transfers his own sense of urgency and enthusiasm to his collaborators. A visionary leader and passionate scientist, Michele has attracted in his orbit a whole community of people who directly or indirectly interact among each other. One remarkable example in which this community recognizes itself is the CPMD package, Version 1.0 being born in summer 1993 when Jürg Hutter joined the IBM Zurich Lab. In this continuously evolving code Jürg's devoted work and the efforts of many among Michele's students, collaborators and colleagues have converged incrementally. The product of all this work is since January 2001 available on the net29 at www.cpmd.org for the general academic user. The CPMD experience is very special as is the ever growing role and impact this code has had and continues to have in the education of new generations in all the advanced simulation methods introduced by Michele. Certainly it has a lot to do with the character of the inventor, who is absolutely not interested in commercializing software and making money with it but instead can be proud of having succeeded in what he always wanted to be above all: a free scientist. All of us like to recall Michele in his relaxed moments, when he enjoys company, new and old friends, and shares his experience, his memories, and his philosophical view of life. Michele has his close friends among his collaborators of a life and—as he says—appreciates their friendship as one of the best gifts that his profession has given to him. He repeated it also on that February 12, 2003, when he received the Honorary Doctorate in Chemistry from the University of his home-town Messina, in that magic moment when family and many friends were reunited in a magnificent scenario where science, art and natural beauty were intermixed. Dear Michele, we hope that you like this set of original research articles that some of your many friends, students, collaborators, and colleagues have written in your honor. We wish you a very Happy Birthday and many more fruitful years in science!