Click here for a list of group members, past and present.

In our group (which at various points have included chemical, civil, computer, electrical, mechanical and BSE engineering students) we generally try to explain unusual experimental results and suggest new experiments for others to try. We work to relate molecular properties on the atomic level to the interpretation of macroscopic phenomena. To this end we devise and develop computer algorithms to study these complex systems. We also make use of standard commercial codes and freeware programs as appropriate to the project at hand.

In other words, we use theoretical and computational methods to characterize chemical reactions, structures of biomolecules, and properties of nanoclusters and bulk materials. This approach is often called "computational chemistry," or slightly more precisely, "computational molecular modeling."

Some of the methods we use include: Monte Carlo, molecular dynamics, molecular mechanics, simulated annealing, Hartree-Fock theory, density-functional theory, semiempirical MO theory, normal-mode analysis, variational transition-state theory, and dynamical corrections to transition-state theory.

Click here to see the NIH Tutorial on Molecular Mechanics and Quantum Chemistry.

(*) indicates a group alum.

"Quantum-mechanical investigation of the intermolecular distortions associated with the binding of aromatic amines and polyaromatic hydrocarbons to human DNA"

Timothy Isgro, Heather Kulik, Darren Louie, Loni Rodriguez, Daniel To, Christine Boelke(*), Lana H. Chan(*), Kimberly Chung(*), Raquib Hannan, NYU(*), Jo Kang(*), and Tina Kiang(*).

We are applying molecular mechanics, semiempirical, Hartree-Fock and (especially) density functional theory methods to aromatic amines, polyaromatic hydrocarbons, and to covalent adduct complexes of these species, comparing the results to all available experiments.

Garrett Bauer, Eugene Agichtein(*), Miguel Amat(*), Michael Capuzzi(##,*), Gerardo Cardone(*), Dominic Coluccio(#,*), Alexandra Deaconescu(#,*), John de la Parra(#,*), Kunmi Demuren(*), Leonid Grinberg(*), Joshua Kritzer(#,*), Keirnan LaMarche(##,*), Sally Mikhail(*), Kenneth Mui(##) and Guowei Yu(*).

This work is supported by a grant from the Petroleum Research Fund of the American Chemical Society. Students whose names are followed by [#] are 1999 American Chemical Society- Petroleum Research Fund Scholars; [##] indicates 2000 ACS-PRF Scholars.

We are studying the structural properties of the solid and liquid phases of elemental bismuth, as well as bismuth clusters. Pseudopotential calculations have been carried out for elemental bismuth in its solid and liquid forms by other groups. We have turned to the construction of effective pair potentials which mimic the pseudopotential results.

We have developed a a first-generation potential energy model derived from some basic theoretical considerations, and are benchmarking it against the known properties of crystalline and liquid forms of bismuth. Our first communication of this work was published in the Internet Journal of Chemistry (2000). We have also submitted a book chapter describing the Monte Carlo methods we are using to Reviews in Computational Chemistry. We are currently preparing to carry out Monte Carlo simulations of the thermodynamic and structural properties of bismuth clusters and molten bismuth using direct density-functional theory calculations within the pseudopotential approximation.

Check out Bismuth on the Periodic Table of Comic Books!

"Accurate predictions of thermodynamic properties and equilibrium constants of oxides, sulfides and selenides using ab initio and adaptively optimized path-integral Monte Carlo methods"

Joseph Kirtland, Jane Atkinson(*), Denise Bergin(*), Christopher Briscoe(*), Timothy Isgro(*), and Andrew Sweeney(*)

In a series of papers in 1992-1993, Topper and Donald Truhlar and their colleagues developed and tested an adaptively optimized stratified sampling Fourier path-integral (AOSS-FPI) Monte Carlo method for "dynamicaly exact" quantum free energy calculations of small polyatomic molecules. The AOSS-FPI method makes no assumption that vibrations and rotations are separable. Their 1993 calculations showed that water molecules are thermodynamically stable species up to 4000K. This was subsequently confirmed by experimental observation of spectroscopic signatures of hot water molecules within sunspots. These signatures were definitively assigned in 1997 by Tennyson and coworkers. Subsequent calculations which specifically correct for the possibility of dissociation have recently been carried out at these same temperatures by Truhlar and coworkers, and have shown the original calculations to be accurate.

We are carrying out AOSS-FPI calculations on a variety of poorly characterized small-molecule compounds, and developing reliable protocols for computations of the molecular entropy, heat capacity and enthalpy as a function of temperature. We are investigating the calculation of equilibrium constants for a number of reactions which we find interesting. Finally, we are investigating and implementing new theoretical methods which have been developed by other groups.

In 1998 we were asked to write a review article of our path-integral Monte Carlo methods. The review was published in a special issue of Advances in Chemical Physics, which focused on molecular Monte Carlo methods, in 1999. You can order this volume at most online bookstores; it is entitled Monte Carlo Methods in Chemical Physics.

"A Parallel Tempering and J-walking Monte Carlo interface to SPARTAN"

Denise Bergin(*)

Jump-walking (or J-walking) Monte Carlo is a powerful method for conformational sampling which helps to overcome the problem of quasiergodicity, which is inherent in ordinary Monte Carlo and molecular dynamics calculations. Parallel tempering overcomes some of J-walking's limitations and ensures the maintenance of detailed balance during the simulation.

We have developed a J-walking Monte Carlo interface to SPARTAN 5.1, called SPARTAN-JMC, which allows all of the methods within SPARTAN to be used to carry out jump-walking Monte Carlo calculations on molecular and cluster systems. The method can be used for conformation searching as well as for classical thermodynamic computations. Future work will focus on optimizing and testing the code and expanding it to parallel tempering simulations and to SPARTAN 02.

Jodie Mueller, Denise Bergin(*), Paul Sweeney(*), and Francis Torres(*)

Snapshot of a simulation carried out using MAGWALKER, our in-house Monte Carlo program (mostly developed by Topper at The University of Rhode Island). Graphic created using XMol (courtesy of Minnesota Supercomputer Center, Inc., now Network Computing Services ). This project is a collaboration with Prof. David Freeman, Department of Chemistry, University of Rhode Island. An article on neutral clusters by Alex Matro(URI), Freeman (URI) and Topper appeared in the Journal of Chemical Physics, and Topper/Freeman have also carried out extensive calculations of the solid phase, looking at finite-size effects and comparing atomistic simulations to compressible Ising lattices (a preprint is available on request). We are currently working on the cation clusters and on the solid state.

Eugene Agichtein(*), Jean Barmash(*), Francis Torres(*), and Guowei Yu(*).

In addition to our studies of ammonium chloride and bismuth clusters, we have also studied the structural properties of clusters of water molecules. Pursuant to thus interest, we have tested a number of global optimization techniques, including simulated annealing via the "Boltzmann simplex" described in Numerical Recipes (2nd ed.), Monte Carlo methods and genetic algorithms. We are always interested in trying out new techniques.

A WWW poster on our work on water clusters was presented in the Third Annual Electronic Computational Chemistry Conference, ECCC-3; a paper describing our results on water clusters was published in 1998 in the Journal of Molecular Structure (THEOCHEM).

"Applications of the Reactive Islands theory of microscopic reaction rates to unimolecular rearrangement reactions"

This project could be loosely described as "using chaos theory to describe and characterize molecular reaction dynamics," and is closely related to the area of quantum chaos. This not an active area of research for our group at present but Cooper students have expressed interest in this work in recent years. The genesis of this project was a research collaboration with Nelson De Leon, C. Clay Marston, Manish A. Mehta, and Alfredo Ozorio de Almeida.

Here's an illustration of numerically reconstructed cylindrical phase-space separatrix manifolds in a molecular model of isomerization between two equivalent conformers (the "De Leon-Berne model"). The graphic was originally created by Mehta and subsequently modified by Francis Torres(*) to show the direction of the phase-space flow within the cylinders. A review article by RQT on this work appeared in Reviews in Computational Chemistry (January 1997, Volume 10).

The cylindrical manifold concept provides a useful, general framework for understanding and visualizing nonstatistical effects in reaction dynamics. All trajectories which cross the transition state must flow through its interior. Its cross-section rigorously constitutes the phase-space variational transition state surface described by Keck. Intersections of the stable and unstable manifolds determine the qualitative nature of the recrossing dynamics as well as the timescales for "activation" (transition from trapped motion to reactive motion) to occur.

Back to Cooper Union Chemical Engineering and Chemistry

Cooper Union Home Page