Large-scale molecular dynamic simulation of a lipopolysaccharide membrane solvated in a 4.2 nm water box.

Andrew R. Felmy, Eric J. Bylaska, James R. Rustad, and T. P. Straatsma, Pacific Northwest National Laboratory

Research Objectives
Our effort consists of molecular-level simulations in two key areas of geochemistry and biogeochemistry: (1) microbial surface mediated processes: the effects of lipopolysaccharides present on gram-negative bacteria; and (2) mineral surface interactions: providing a molecular-scale understanding of surface complexation reactions at oxide, oxyhydroxide, and silicate minerals.

Computational Approach
We use a variety of computational chemistry methods, including density functional theory, molecular mechanics/dynamics, Car-Parrinello, and kinetic theories. Besides NWChem, we also use parameterized classical potential models for the interaction of water and hydroxide with Fe/Al surfaces. These models are being used to calculate bulk and surface properties. These models are based upon parameterizations from ab initio calculations, and they have been particularly successful in predicting structures, surface charging, and water chemistry of iron-oxide surfaces.

Accomplishments
Plane-wave pseudopotential methods were used to investigate the structures and total energies of AlOOH and FeOOH in the five canonical oxyhydroxide structures: diaspore (goethite), boehmite (lepidocrocite), akaganeite, guyanaite, and grimaldiite. The local density approximation was used in conjunction with ultrasoft pseudopotentials in full optimizations of both AlOOH and FeOOH in each of these structures. Structures are in reasonably good agreement with experiment, with lattice parameters and bond lengths within 3% of the experimental ones.

An important new code development has benefited from our NERSC computer time. A parallel projector augmented-wave code has recently been completed and is currently in an extensive testing phase. This code will allow us to simulate many new types of materials at a first-principles level, including iron oxides.

An isodesmic procedure based upon density functional theory was developed to predict accurate reaction thermodynamics for important redox half-reactions in the solution phase. This work is an extension of our previous work in which we developed a scheme for predicting the thermodynamics of SN2 reactions in the solution phase.

Significance
Subsurface microbial processes can control the rates of oxidation/ reduction reactions, modify and enhance mineral dissolution and precipitation reactions, and adsorb metals and other ions at the microbial surface. Current theoretical understanding of these processes, which are believed to occur either directly at the microbial surface or at the microbe interface, is very limited.

The ubiquitous occurrence, high specific surface area, and strong binding to a large number of cations, anions, metal ions, and organic chelates makes Fe/Al oxides and oxyhydroxides important adsorbing surfaces. Much of what is known about these adsoption processes on Fe/Al oxides is based upon macroscopic measurements, and relatively little is known at the microscopic level about what types of binding sites exist at oxide surfaces. Difficulties in characterizing the structure and energetics of these sites obstruct the development of improved thermodynamic models for adsorption.

Publications
James R. Rustad, David A. Dixon, Kevin M. Rosso, and Andrew R. Felmy, “Trivalent ion hydrolysis reactions: A linear free-energy relationship based on density functional electronic structure calculations,” J. Am. Chem. Soc. 121, 3234 (1999).

Eric J. Bylaska, David A. Dixon, and Andrew R. Felmy, “The free energies of reactions of chlorinated methanes with aqueous monovalent anions: Application of ab initio electronic structure theory,” J. Phys. Chem. A 104, 610 (2000).

James R. Rustad and Kevin M. Rosso, “The structures and energies of AlOOH and FeOOH polymorphs form plane wave pseudopotential calculations,” American Mineralogist (submitted, 2000).