Autobiography of Fumio Hirata

Hirata
2007 Condensed Matter Physics  
Born in Nagasaki prefecture in Japan, right after the world war two, I grew up in a country side of Fukuoka prefecture, where I had my education until high school. After graduating from the high school, I entered Hokkaido University to study the natural science. The largest turning point as a student came to me when I knocked on the office door of Prof. Arakawa in 1969, who became my thesis advisor, in order to consult my study in the graduate course. He told me enthusiastically how "water" and
more » ... "aqueous solutions" are important in science, how challenging it is to understand the problems experimentally as well as theoretically, and how interesting it is in the biological connections. In fact, at that time, the science related to water was about to flourish both in experiments and in theories accelerated in particular by the bio-scientific motivation such as the hydrophobic interactions. Both experimentalists and theorists were interested in the anomalous behavior which water exhibits in thermodynamic properties, such as the existence of the maximum density at 4 degree, and were trying to explain the behavior in terms of the structure of the liquid. Actually, the X-ray diffraction study of water, triggered by the landmark work by Bernal and Fowler in 1933, was producing the structural data which unequivocally indicate the existance of some tetrahedral coordination similar to that in ice. Concerning the theoretical studies, a couple of statistical mechanics theories, both based on a structural model of water, had been prevailing among the scientists in the field: one by Nemethy and Scheraga and the other by Pople. The model by Nemethy and Scheraga was a mixture model in which water is a mixture of nonbonded monomers and hydrogen-bonded clusters having one to four hydrogen bonds. Nemethy and Scheraga applied the free volume theory, a mean field theory based on the solid-like structure of liquids, to their model to reproduce the thermodynamic quantities and the cluster-size distribution. Pople proposed the bent-hydrogen-bond model of water structure, in which the thermal fluctuation of the ice-like structure in the tetrahedral coordination is represented by the "bending" of hydrogenbond from the linear configuration. Arakawa, who was basically an experimental chemist in the field of ultrasonic measurements of aqueous solutions, was also the first chemist in Japan to contribute to the statistical mechanics theory of liquids, especially of water. Arakawa had already proposed a couple of theories for water, both based on the two state model, hydrogen bonded or non-hydrogenbonded: one using the free volume theory, and the other employing the Brag-Williams-type theory of phase-transition. I believe that Arakawa's theories are the first treatment based on the two state model of water, although it has not been well appreciated in the recent booming of the two state model of water motivated by the finding of two phases in the amorphous ice, low density and high density. In any case, those theories which employ the structural model of water have only historical significance, since the essential theoretical problem to predict the structure of water from the first principle was untouched at that time. Advised by Arakawa, I started my life as a baby scientist with an experimental study of the adiabatic compressibility concerning an aqueous solution of a series of the tetra-alkyl ammonium salts, and published my first paper. (1972) In the results, the temperature dependence of the compressibility reverses in between tetraethyl-and tetrapropyl-ammonium ions. The results were indicating some qualitative change in the hydration structure between those ammonium ions. The finding itself was not so big, but the study determined the direction of my future scientific activity. By struggling with the interpretation of the results of the experiment in order to extract microscopic view concerning what is going on in the hydration structure, I realized that there is no theory applicable to the process, and decided to change my research field to the statistical mechanics of liquids. During my Ph. D. thesis, I studied and developed three different theoretical treatments of water and aqueous solutions. One of those is the scaled particle theory, originated by Reiss, Frisch, and Lebowitz, and applied successfully to aqueous solutions by Pierotti. Applying the theory to the partial molar volume of ions in aqueous solutions, I proposed a new method to decompose the volume into the "intrinsic" and "electrostriction" terms, which had conventionally been done in empirical manner. (1973) The other theories that had a great impact on my thesis work were the analytical solutions of the Ornstein-Zernike (OZ) equations with the PY and MSA closures. After several months of struggling for understanding the theories, I picked up the analyt-299 ical solutions for the primitive model of electrolyte solutions obtained by Weisman and Lebowitz, and applied it to Friedman's model of the solutions. At that time, Friedman was developing a new theory of ionic solutions, in which the solvent induced interaction between two ions in water is modeled by the so-called "Gurney terms". It was the first serious attempt to put forward the electrolyte solution theory beyond the "primitive model". I combined the Freidman model and the Weisman-Lebowitz solution into the Zwanzig-type perturbation theory of liquids to describe the realistic aqueous solutions of electrolytes: the primitive model and the Gurney term as the reference and the perturbation, respectively. (1975) The theory was successful in the sense that it could produce the Gurney terms which are consistent with the empirical models of ion hydration proposed by Frank and Wen as well as Samoilov in 1957. However, the Gurney terms are essentially phenomenological, and this did not give a complete satisfaction to me. I then started to study formalisms which might be capable of being applied to water, and found the seminar papers written by Chandler, Andersen, and Weeks. The papers were proposing two models of molecular liquids, which are closely related to each other, namely the reference interaction-site model (RISM), and the embedded site model (ESM). I naturally picked up ESM, because it employed the analytical solution of the PY equation for the hard-sphere system as the reference for the diagrammatic perturbation expansion of the site-site pair correlation functions of molecular liquids. Needless to say, I was already familiar with the analytical solutions, while at that time I did not have numerical techniques sufficient to solve the RISM integral equation. In order to apply the ESM theory to the structure of water, I made up a model of water, which is similar to that proposed by Ben-Naim. The hydrogen-bond is represented by a square-well-type interaction between the interaction sites which are located symmetrically at the vertices of a tetrahedron embedded in a hard-sphere. The electrostatic interaction was taken into account by placing an ideal dipole at the center of spheres. Regarding the hydrogen-bond as well as the dipole-dipole interaction as a perturbation to the hard-sphere interaction, I applied the ESM theory to the model water system. (1977) The theory and the model qualitatively reproduced the two characteristics of water structure, reflected in the oxygen-oxygen radial distribution functions (RDF). It has been well regarded since the paper by Bernal and Fowler that the first peak of RDF is considerably narrower than that in simple liquids, and that the second peak appears at the position similar to that in ice, both being manifestations of the ice-like tetrahedral coordination of the hydrogen-bond network in water. Although the theory did not attract too much attention, it probably was the first attempt to "qualitatively" reproduce the "signature" of water structure from the first principle.
doi:10.5488/cmp.10.3.299 fatcat:3j7efohbizdw5jutundo256ytq