Kim's 'K'omputational 'K'emistry Project Area

San Diego Supercomputer Center, PO Box 85608, San Diego CA 92186-9784
UPS, Fedex 10100 Johns Hopkins Drive, La Jolla CA 92093
(619) 534-5000

Software | Research | Collabortors | Publications | Kims Home Page

Applications Projects
Designing the molecular world has been coined as the scientific frontier of the 21st century and our research efforts direct a combined program of computational prediction and experimental verification which strives to understand the design principles for the molecular constructions. Part architects and part sculptors of chemical structures, we use principles of symmetry and form to augment our research. An example includes the questioning of long held beliefs about the nature of a well known chemical structure like benzene, which lead us to provocative speculations about the fundamentals of aromatic molecules. With the aid of computational quantum chemical packages such as GAMESS and QMView , the team is able to give these speculations specific structural focus, and finally, the synthetic wing at UCSD creates these molecules, "in substance," and tests the predictions using modern NMR spectroscopy and X-ray crystallography.
Not every study leads to structural designs. Sometimes it is necessary to correct, or at least question, notions that were created by experimental or computational errors. A clear case of this is in the classic report of extremely long bonds in molecules like 9,10-dimethyleneanthracene dimer photoisomer or lepidopterene . Ab initio calculations that we performed in this area did not agree with the original experiment and when we rechecked the X-ray structure of the former compound, we found the original structure was in error. This now opens up questions concerning the mechanisms through which orbital interactions lead to bond lengthening, questions we are now investigating.
The size of these chemical structures and the need for high accuracy makes the computational effort of these projects extremely intensive. Some of our most exciting problems require extensive basis sets and/or dynamic correlation, utilizing density functional methods on systems of greater than 50 atoms, all of which makes machine size a severe limitation. Our team is certainly not platform limited, accessing the C90, Paragon, T3D, SP2 and alpha-farm cluster, to be able to maintain our usual assault on stereochmeical problems. We benefit from the work we put into making our packages portable to almost any platform.
The importance of the kinds of research done in our efforts can be found if one realizes that aromatic subunits are found in a large number of everyday materials. Petrochemicals from aircraft fuel to common charcoal, depend on aromatic chemistry. The agrochemical and pharmaceutical industry produces vast varieties of aromatic and heteroaromatic molecules as the bases to their most active drugs and agricultural aids. Modern textiles and dyes would be unthinkable without aromatics. On the negative side, aromatic residues have a heavy environmental impact as well. Clearly, progress in amplifying the benefits and reducing the drawbacks will come from understanding and design.