I work with Professor Vincent McKoy, using computers to study electron-molecule collisions in gases This page is an elementary description of that work aimed at nonspecialists.
Most gases are made of molecules. (The exceptions are the “noble gases” such as helium, neon, and argon, which are made of free atoms.) Molecules are, of course, made up of atoms, and atoms in turn contain negatively-charged electrons and positively-charged nuclei. An ordinary gas is electrically neutral: All of the electrons are bound within molecules in such a way that the positive and negative charges balance, so that the gas acts like a collection of neutral particles.
Electrons can be knocked loose when there is a source of sufficient energy: examples are far-ultraviolet light, intense radio-frequency fields, or very high temperatures. Really extreme conditions such as those inside a star reduce everything to a soup of electrons and nuclei--a fully ionized plasma. Under milder (and nonequilibrium) conditions, however, only some molecules are ionized (deprived of an electron) or dissociated (broken apart into atoms and/or smaller molecules). These partially ionized plasmas are extremely interesting from a chemical point of view, because they contain highly reactive fragment molecules and ions that would otherwise be formed only at high temperatures. Indeed, low-temperature, noequilibrium plasmas are used in several major processing steps in the fabrication of semiconductor microelectronics.
When high-energy ionizing radiation encounters living tissue, its energy is dissipated through collisions. Since living things are mostly water, most of those collisions involve water molecules. Ionization of water produces ions and fast electrons that quickly collide with and ionize other water molecules, producing a cascade of more and more, but slower and slower, electrons and ions.
Because this energy deposition process is rapid, and because the slow secondary electrons are so numerous, radiation scientists have known for a long time that most of the damage done to biomolecules by ionizing radiation is caused indirectly by the slow secondaries rather than directly by the primary radiation. It was surprising, though, when Léon Sanche and coworkers found1,2 that slow electrons could cause double-strand breaks in DNA. Indeed, further research showed that quite slow electrons—too slow to cause electronic excitation, much less ionization—could cause single-strand breaks in DNA.3 These observations have stimulated many researchers, including us, to investigate the mechanisms by which slow electrons to “stick to,” and ultimately disrupt, DNA. To date, our research has concentrated on elastic electron collisions with subunits of DNA, including the nucleobases, the backbone sugar deoxyribose, and larger assemblies (nucleosides and the nucleotide dAMP).
In low-temperature plasmas, the upper atmosphere, and other environments where molecules and free electrons mix, electron-molecule collisions drive the dissociation and ionization of molecules, thereby driving the chemistry occuring within the plasma or at surfaces in contact with it. A major goal of plasma scientists and engineers is the development of numerical models that will allow them to predict plasma behavior and properties from first principles. Because electron-molecule collisions play such an important role, any reliable plasma model will require detailed and accurate sets of electron-molecule collision data.
Electron-molecule collisions are governed by the laws of quantum mechanics, and the outcome of any given collision is expressed by a set of probabilities, one for each possible outcome. These probabilities are usually expressed as cross sections, which have units of area, and which can be calculated, at least in principle, from Schrödinger's equation, the basic equation of quantum mechanics. In practice, however, the calculations are enormously demanding, and they grow rapidly more demanding as the size of the molecule is increased.
Indeed, direct solution of Schrödinger's equation is out of the question for all but the simplest physical systems, and we are forced to turn to approximation methods. One powerful class of methods relies on a variational principle for some quantity related to the collision cross section. In our own research, we use the Schwinger Multichannel variational principle for the scattering amplitude as the basis for a computational procedure that is applicable to arbitrary molecules.
Altough variational approximation methods are efficient, the calculations remain enormous, and the main barrier to solving problems of practical interest is lack of adequate computing power. Accordingly, we have adapted our programs to run on massively parallel computers, the most powerful machines available today.