Theoretical Chemistry: Theory or Practice

by Sergei Manzhos

"Discovery of new materials or cleaning up the environment, theoretical chemists
find themselves working to benefit the society in most practical ways."

Whether discovering new materials or cleaning up the environment, theoretical chemists find themselves working to benefit society in practical ways.

With this article, we introduce a series of publications that showcase the application of scientific knowledge and tools in different professions. People working in different fields - from biotech to the financial industry - will be sharing their perspective on the use of scientific knowledge. While there is a common understanding of the importance of science for research and development and, of course, for academia, we will show that most modern professions rely on scientific knowledge and tools that students on school benches might consider purely theoretical and inapplicable in their future careers, and therefore uninteresting.
We will also try to provide a brief overview of tools and software used in particular areas. We start the series with an account of a seemingly purely academic discipline: theoretical chemistry.

There is no pure science

The conceptual division of science into pure and applied has been deeply entrenched in the minds of the general public, if it was not in the minds of scientists themselves. Fortunately, recent technological advances have blurred the distinction. The distinction between science and engineering is also being questioned. When I was a university student, I could see the frustration of my peers at having to take the courses in differential equations, statistics, higher algebra or quantum mechanics. Students who are not determined to pursue an academic career tend to consider those disciplines as a necessary evil, a suffering they have to go through to pass the exams, obtain the degree, and pursue the profession of their choice without ever having to come back to those scary concepts. Those who dream about building their career in one of the many professions that have faithfully served the society for centuries are probably right to think so. If, however, you want to work in any of the - let's call them "modern" - industries born or transformed after the world developed information technologies, neglecting those seemingly purely theoretical disciplines in your student years might cost you dear in both the figurative sense of limiting your career choices and consequently also literally by limiting your future income.

Here we would like to describe how theoretical chemists contribute to solve some of the most practical problems modern societies face. Theoretical chemists are people who use theories and computational tools to compute properties of substances and materials. Theoretical chemistry is really a mix of disciplines, mainly physics, chemistry, mathematics, informatics. Usually knowledge of all these disciplines is necessary to do theoretical chemistry.

Where solutions to the world's problems come from

Let us consider two of the hottest issues that the world faces today: environmental pollution and clean energy. We have seen political debates, we have seen people taking it to the streets, and we have seen governments losing majorities or being overthrown when these issues make it to the top of society's priority list. But who is he who can practically resolve these issues?

Pollution means that chemical elements or compounds which harm human health are present in the atmosphere, water, food chain. To reduce that harm, one needs (i) identify the harmful substances, sometimes at very low concentrations (several ppm or even ppb - parts per million or parts per billion) (ii) capture or destroy them to prevent their release into the environment or to neutralize those already released.

Identification of chemical substances is often done by spectroscopic means: one shines light (visible, IR or UV, but X-rays and even gamma rays are also used) on the compound which may adsorb certain radiation frequencies, emit its own characteristic radiation or even disintegrate into parts (dissociate). One measures the extent of adsorption/emission/dissociation depending on the frequency of the radiation used - spectroscopic signature - and compares it to the spectroscopic signatures of known species for identification. The most precise measurements are done with laser radiation. The knowledge of quantum physics is necessary both to design the measurement and to understand its result. The spectroscopic signatures are often complex signals with many overlapping peaks. To help disentangle them, theoretical chemists compute them and compare to the experimental results. This computation involves the solution of the Schrodinger equation, which is a differential equation that governs the behavior of all atoms in chemical compounds. That problem is then recast into the problem of solving a huge system of linear equations (which can contain millions of equations). Specially designed methods for matrix manipulations are needed to solve it.

We started with the detection of hazardous chemicals and we have come to differential equations and linear algebra - pure math, isn't it.

How does one make those offending substances go away? It is usually done my making them react to produce innocuous or immobilized molecules or by making them disintegrate. First of all, one can guess that harmful compounds do not often react on their own with ambient molecules and materials to produce less harmful substances - then we would not have many of them floating around. But they can disintegrate or react into non-harmful substances in the presence of other molecules or materials - catalysts. An everyday example is catalytic converters found in many cars but also in industrial installations. The role of automobile converters is usually the oxidation of CO to CO2, of unburnt hydrocarbons to CO2 and water, and the reduction of nitrous oxides. These reactions happen at the surface of a catalyst the best of which are platinum, palladium and rhodium. In fact, most solid state catalysts are very expensive materials. Catalysis consumes today about 9/10 of the world supply of rhodium (which is more expensive than Au - another good catalyst) and 2/3 of platinum. This is a problem. Some reactions we know how to catalyze, but the catalyst is too expensive, and some reactions we do not even know how to.

Therefore, the discovery of new, more effective and cheaper catalysts is a huge environmental and economic priority. Theoretical chemists contribute to this process by computing catalytic activity of different substances for different reactions. This is done by again solving the Schr?dinger equation, and sometimes by computing the time evolution of atoms involved in the reaction. This boils down to solving differential equations. The knowledge of quantum mechanics, chemistry, mathematics, and classical mechanics all together is required here.

What about clean energy? We will consider only one telling example. Alerted students must have heard about fuel cells - devices that convert chemical bond energy into electricity. They also operate by catalysis. They can use a variety of fuels. Most known are hydrogen fuel cells - it is this type of fuel cells that is supposed to power Vancouver's fleet of buses servicing the Olympic village. But hydrogen is not easy to handle and is not cheap - a fill up might cost of the order of $1000. Alcohols are considered a promising alternative fuel, and the so-called direct methanol fuel cells have been developed for use in mobile electrical products such as batteries, but never commercialized. The reason is that there is no practical catalyst for the oxidation of methanol, which is the key reaction in the operation of the methanol fuel cell. Theoretical chemists are now modeling this system with different candidate catalysts to make the commercialization of this clean and safe energy source possible.

The above are just two of multiple applications of theoretical chemistry to our wellbeing. But I think they already give one an idea of how much modern technology and our ability to respond to challenges the world faces depend on basic science in more direct ways than most people recognize.

Tools

Mostly theoretical chemists use computational tools and software. Those programs often perform day- or month-long calculations of solutions of differential equations needed to compute molecular structures and chemical reactions; they are therefore run mostly on computational servers under the Unix operational system. The most widely known and used software for this purpose is called Gaussian. Gaussian is a commercial product developed by scientists from all over the world, and Canadian scientists contributed to it as well. It is used in companies involved in drug and catalysis design as well as in universities. Other software optimized for particular applications is also widely used (VASP, SIESTA, ADF are some of the more common).

To comprehend better and to represent the results of the computations, visualization software is used. The whole set of commonly known (PowerPoint, Adobe Photoshop, Adobe Illustrator) and less known specialized software (such as GaussView) finds its use here.

Data analysis tools are widely used - for statistical analysis, regression, function fitting... Besides the ubiquitous Excel, SPSS, Origin, MatLab are widely used as well as software developed for specific applications

Usually, programming skills are also required. Object oriented programming languages are often used, such Visual Basic or Visual C with external libraries depending on the application, but the oldest programming language FORTRAN is still widely employed. Researchers at pharmaceutical companies who develop drugs of the future often use Gaussian for computing molecular structures and reactions - and it is written in the ancient programming language FORTRAN.

There is a growing use of higher level and symbolic programming environments such as Maple (developed in Canada at the University of Waterloo) and MatLab. Maple is a symbolic computation software, i.e. it performs formulae manipulation and derivations which are often complex to the point of being intractable when done by hand. MatLab is a high-level programming environment that dispenses researchers with minutia programming and also allows for symbolic calculations. It comes with a set of libraries that makes it easy to do computations typically done for particular applications: signal processing and analysis, statistical analysis, image processing, valuation of financial instruments etc. These are some reasons why it is often used in industry.

Freeware. The programs mentioned above can be quite expensive (for example, MatLab costs about $1000 plus about $500 for each library). However, there is freeware that doubles most of the functions of expensive software. I would in particular recommend R, Scilab, and Octave as substitutions for MatLab; for data analysis, OpenOffice performs most of the functions of Excel.