Molecular modeling - Theoretical basis (part 1)

How can molecular modeling be defined, and how can it be considered an engineering technique?

Since the end of the 19th century, the sciences of matter have been based on atomic theory. Engineers concerned with useful transformations of matter rely on atomistic and molecular representations of the processes involved. Based on such microscopic descriptions, the concepts of thermodynamics, chemical kinetics, spectroscopy, crystallography, solid state physics, etc., in principle enable quantitative predictions of macroscopic observables. In practice, however, these concepts have hitherto mainly been used to organize and interpret experimental measurements, and to base predictions on these measurements. This is how, for example, vast databases have been built up listing the thermodynamic properties of pure bodies, which are needed for the initial design of industrial processes.

Today, advances in computer technology have encouraged the deployment of digital simulation techniques as an exploratory tool in all areas of the physical sciences, and as a design tool in engineering. The sciences of matter and process engineering are not to be outdone: for all that, simulation concerns collections of atoms, whose organized aggregation in varying degrees, from small molecules to crystals, underpins the diversity of reality perceived on a human scale, from minerals to living organisms. Probably because chemists were the first to see this as a natural extension of their three-dimensional molecular models based on balls and rods, the term "molecular modeling" has come to be used for this category of simulations or, more appropriately, numerical experiments. The chemical engineer in the broadest sense of the term seeks to produce a material with use value: his first target is a property. The consumer's growing and well-founded sensitivity to the potential harmful effects of artificial substances on human health and the environment means that a complex set of specifications has to be drawn up, with properties to be sought and properties to be avoided. Exploratory research very early on attempted to go beyond the "trial and error" stage and establish "structure/property relationships". These initially proceed by analogy, but reach their greatest maturity when they are grounded in clear physico-chemical determinism.

For engineers, the strongest motivation for using molecular modeling is to link chemical structures and properties via the first principles of physics, thus opening the way to virtual sorting of structures prior to any experimentation, or even preferable to experimentation, as soon as the above-mentioned specifications exceed a certain threshold...