Empirical valence bond models for reactive potential energy surfaces: a parallel multilevel genetic program approach

Multiscale Reactive Molecular Dynamics for Absolute pKa Predictions and Amino Acid Deprotonation

Accurately calculating a weak acid's pK_a from simulations remains a challenging task. We report a multiscale theoretical approach to calculate the free energy profile for acid ionization, resulting in accurate absolute pK_a values in addition to insights into the underlying mechanism. Importantly, our approach minimizes empiricism by mapping electronic structure data (QM/MM forces) into a reactive molecular dynamics model capable of extensive sampling. Consequently, the bulk property of interest (the absolute pK_a) is the natural consequence of the model, not a parameter used to fit it. This approach is applied to create reactive models of aspartic and glutamic acids. We show that these models predict the correct pK_a values and provide ample statistics to probe the molecular mechanism of dissociation. This analysis shows changes in the solvation structure and Zundel-dominated transitions between the protonated acid, contact ion pair, and bulk solvated excess proton.

Design of linear ligands for selective separation using a genetic algorithm applied to molecular architecture

Continuous purification of chemical reaction products through adsorption-based operations during workup may present advantages over batch chromatography or crystallization. In pharmaceutical syntheses, however, the desired product is often structurally similar to byproducts or unconverted reactant, so that identifying a suitable adsorption medium is challenging. We developed an in silico screening process to design organic ligands which, when chemically bound to a solid surface, would constitute an effective adsorption for a pharmaceutically relevant mixture of reaction products. This procedure employs automated molecular dynamics simulations to evaluate potential ligands, by measuring the difference in adsorption energy of two solutes which differed by one functional group. Then, a genetic algorithm was used to iteratively improve a population of such ligands through selection and reproduction steps. This procedure identified chemical designs of the surface-bound ligands that were outside the set we considered using chemical intuition. The ligand designs achieved selectivity by exploiting phenyl-phenyl stacking which was sterically hindered in the case of one solution component. The ligand designs had selectivity energies of 0.8-1.6 kcal/mol in single-ligand, solvent-free simulations, if entropic contributions to the relative selectivity are neglected. We believe this molecular evolution technique presents a useful method for the directed exploration of chemical space or for molecular design, when the chemical properties of interest can be efficiently evaluated through simulations.

Multiscale reactive molecular dynamics

Many processes important to chemistry, materials science, and biology cannot be described without considering electronic and nuclear-level dynamics and their coupling to slower, cooperative motions of the system. These inherently multiscale problems require computationally efficient and accurate methods to converge statistical properties. In this paper, a method is presented that uses data directly from condensed phase ab initio simulations to develop reactive molecular dynamics models that do not require predefined empirical functions. Instead, the interactions used in the reactive model are expressed as linear combinations of interpolating functions that are optimized by using a linear least-squares algorithm. One notable benefit of the procedure outlined here is the capability to minimize the number of parameters requiring nonlinear optimization. The method presented can be generally applied to multiscale problems and is demonstrated by generating reactive models for the hydrated excess proton and hydroxide ion based directly on condensed phase ab initio molecular dynamics simulations. The resulting models faithfully reproduce the water-ion structural properties and diffusion constants from the ab initio simulations. Additionally, the free energy profiles for proton transfer, which is sensitive to the structural diffusion of both ions in water, are reproduced. The high fidelity of these models to ab initio simulations will permit accurate modeling of general chemical reactions in condensed phase systems with computational efficiency orders of magnitudes greater than currently possible with ab initio simulation methods, thus facilitating a proper statistical sampling of the coupling to slow, large-scale motions of the system.

A three-state effective Hamiltonian for symmetric cationic diarylmethanes

We analyze the low-energy electronic structure of a series of symmetric cationic diarylmethanes, which are bridge-substituted derivatives of Michler's Hydrol Blue. We use a four-electron, three-orbital complete active space self-consistent field and multi-state multi-reference perturbation theory model to calculate a three-state diabatic effective Hamiltonian for each dye in the series. We exploit an isolobal analogy between the active spaces of the self-consistent field solutions for each dye to represent the electronic structure in a set of analogous diabatic states. The diabatic states can be identified with the bonding structures in classical resonance-theoretic models of cyanine dyes. We identify diabatic states with opposing charge and bond-order localization, analogous to the classical resonance structures, and a third state with charge on the bridge. While the left- and right-charged structures are similar for all dyes, the structure of the bridge-charged diabatic state, and the Hamiltonian matrix elements connected to it, change significantly across the series. The change is correlated with an inversion of the sign of the charge carrier on the bridge, which changes from an electron pair to a hole as the series is traversed.