My research falls in the field of computational chemistry. What is computational chemistry? Well, there are volumes and volumes on that subject, but I will cut it short here. Basically computational chemistry uses computers to simulate the behavior of molecules in the real world. Right away you can see the advantage of this; no more messy and dangerous chemical spills, no more exploding reactions, no more costly clean-up… You can determine the behavior of toxic chemicals without risk to the environment, you can get an idea of the most feasible reaction pathway to a new molecule before going into the lab to make it. The list goes on. One other thing to point out before going into details, the field is HUGE! Computational chemistry has come a long way since people were using boxes and boxes of punch cards to get an idea (and not always such a good one) about triatomic molecules. From highly accurate gas phase small molecule calculations to molecular dynamics simulations of biologically active molecules in solution and everything in between, computers have become a permanent tool of chemical research.

So where do I fit into the grand scheme of computational chemistry? In the end, all chemistry comes down to chemical bonds and therefore, my research focuses on the nature of bonding. The computer solves the mathematical equations that describe the forces holding atoms together to make molecules. I then go in and pick apart that solution to find out what are the factors contributing to a particular bond and how much does each contribute.

The method of choice for this type of work is valence bond (VB) theory. If you have taken a general chemistry course and drawn Lewis structures then you are familiar with VB. Unfortunately, undergraduate textbooks tend to trash VB as an incorrect but efficient and easy to use model. VB is presented as a primer for the more advanced "correct" molecular orbital (MO) theory. When it comes right down to it, neither MO nor VB are "correct", they are both just models that help us to explain the way atoms come together. When taken to the limit, both models are exactly the same. They just approach the problem from different angles.

The power of VB lies in its clear conceptual picture of the two electron bond based ideas of Lewis(1) and Langmuir.(2) Let’s take the Heitler-London(3) (HL) treatment of the simplest bond, H2, as an example. Here, the bond is made by pairing electrons on each H atom, fab=Ha._.Hb. The HL treatment can be extended by allowing for both electrons to be on Ha, faa=Ha:- Hb+ or on Hb, fbb=Ha+ :Hb-. These three VB configurations can then be combined to give the overall wavefunction, y=c1fab+c2faa+c3fbb, where c1-3 are coefficients that describe how much each configuration contributes to the whole. Much of the early work on the VB formalism was done by Pauling(4) and extended to treat molecules much more complicated than H2.

Thanks in large part to the efforts of Shaik,(5) Pross,(6) and Hiberty, (7) this elegant approach has lead to great advances in the understanding of reactivity and bonding in organic molecules. However, similar advances have been lacking for transition metal (TM) complexes due to the computational difficulty of such molecules. The normal ideas of bonding break down in transition metals complexes which are comprised of mostly electron pair donor/acceptor coordinate bonds rather than the traditional covalent bonds that share single electrons on each atom. My research seeks to apply the conceptual benefits of valence bond theory to chemical bonding in transition metals and thus, develop a unified bonding theory that applies to the entire periodic table.

Initial work in this area is focusing on the accurate in situ prediction of s bond energies in Lewis acid adducts. Once the validity of the method is established, this procedure will be used to elucidate s and p bonding in the TM-O/O2 bond. Work on the TM-O/O2 bond will start with the simple equatorial ligand-free 2+ ion and gradually build up to the biologically and catalytically important prophyrin (see picture on home page)

With the recent increases in computational speed, power and storage capacity a modest sized personal computer is now adequate for this type of research. Marist college presently has an IBM Pentium III PC with 256 M memory dedicated to computational research. In addition, I have access to a 30 node Linux PVM cluster consisting of 256 MB Pentium 4 IBM PCs. Aside from two hours a day for two days a week for computer science classes, this cluster is available for research. All of these machines are currently running the computational chemistry software GAUSSIAN, GAMESS, GAMESOL, XIAMEN99, MOLEKEL, and MOLDEN. Additional software available for teaching and possibly research purposes include HYPERCHEM, TBLMTO, DYNASOLVER, and VB2000.

Current Group Members:

Tait Takatani ('06)

Megan Licata (graduated ’02)

Christin Palombo (graduated'04)

Alyson Fiorillo (graduated'04)

Glen Ferguson (graduated'04)

Refrences:

1) Lewis, G. N. J. Am. Chem. Soc. 1916, 38, 762.

2) Langmuir, I., J. Am. Chem. Soc., 1919, 41, 868 and 1543

3) Heitler, W.; London, H. F. Z. Physik 1927, 44, 455.