Meeting: Poster Session Summaries
NWChem Meeting on Science Driven Petascale Computing and Capability Development at EMSL
- January 25-26, 2007
- W.R. Wiley Environmental Molecular Sciences Laboratory
- Richland, WA
DFT Study on the Interaction Potentials of Coronene Dimers
- Yan Zhao
- University of Minnesota
State-of-the-art density functional theory has been applied to generate potential energy curves for the sandwich, T-shaped, and parallel-displaced configurations of the large prototype of aromatic pi-pi interactions, the coronene dimer. Results were obtained using a newly developed density functional, M06-2X, with different augmented double zeta and triple zeta basis sets. The high-quality estimates of the DFT potential energy curves for the coronene dimer presented here provide a better understanding of how the strength of pi-pi interactions varies with distance and orientation of the aromatic rings.
Development of TDDFT Gradients
- Huub Van Dam
- CCLLRC Daresbury Laboratory
Overview of the development of TDDFT gradients for the analytical calculation of forces in molecules in the excited state. The technology enables the efficient optimisation of molecular structures of molecules in an excited state. It also opens the door to excited state dynamics as well as the calculation of first order properties. The poster outlines the current developments and presents their current status. It also invites discussions on applications.
Hydrogenation of Dinitrogen Coordinated in Dinuclear Metallocene Complexes
- Petia Bobadova-Parvanova
- formerly Cherry L. Emerson Center for Scientific Computation
- Emory University
The design and synthesis of novel catalysts, capable of hydrogenating molecular nitrogen under mild conditions, is of great fundamental and industrial interest. The almost-century old Haber-Bosch process requires extreme conditions to utilize molecular nitrogen and hydrogen to produce ammonia. Giving the fact that more than 100 million tons of ammonia are produced every year, finding a new catalyst that could reduce the required temperature and pressure would be of great economic advantage. Although a new catalyst has been sought for more than 70 years, a break-through discovery has not been done yet. However, new dinuclear metallocene-N2 complexes have been synthesized and reactions of N2 cleavage and hydrogenation at mild conditions have been observed. These reactions have shown promising possibilities for solving the long-standing problem, but have raised several very intriguing questions.
The results of computational studies of dinitrogen hydrogenation at different dinuclear metallocene-N2 complexes will be presented and the reasons behind the remarkably different experimental reactivity of these complexes will be explained in terms of different properties and intramolecular interactions. The mechanism of the dinitrogen hydrogenation reaction will be compared and the necessary conditions that would lead to successful dinitrogen hydrogenation will be discussed. It will be demonstrated how first-principle electronic structure calculations can answer puzzling questions that emerge from experimental results and help designing new compounds with desired properties.
Modeling of Rotational Behavior of Solid State Ammonia Borane
- Vencislav Parvanov
- Donald Camaioni
- Maciej Gutowski*
- Nancy Hess
- Wendy Shaw
- Abhi Karkamkar
- Tom Autrey
- Pacific Northwest National Laboratory
This work is ongoing investigation of properties of ammonia borane BH3NH3. As material ammonia borane is a solid powder. For all published structures of this material at all temperatures and phases is typical to observe dihydrogen bonds forming network perpendicular to BN bonding. Rotation around BN bonds was often suggested, considering the bond strengths in solid state. Rotational barriers were measured as separate motions of NH3 and BH3 groups using NMR and neutron scattering techniques. Measured activation energies for these motions from different sources show much higher barrier for BH3 of 21-25 kJ/mol than NH3 8-13kJ/mol. We present simple cluster model for calculating these barriers. Our results show that rotation of Borane part has higher barrier than experimentally measured, when considered as independent. We suggest a correction coming from the trend of this molecule to keep the more stable internal conformation. Thus calculated barriers match exactly the experimental value.
Two component DFT methods to study the Stability of Actinide Oxidation States
- Patrick Nichols
- Wibe de Jong
- Eric Bylaska
- Pacific Northwest National Laboratory
The behavior of actinide oxides in solution is largely determined by the stability of the various oxidation states for that particular element. A comprehensive understanding of this aspect of actinide materials will be crucial to future waste processing, separation and storage efforts. This work will present methodologies to facilitate the study of actinide complexes in solution and interfacial interactions through the use of ab initio molecular dynamics. The incorporation of spin-orbit and relativistic scalar pseudopotentials in NWCHEM is a crucial first step. The final goal will allow the use of ZORA and higher order relativistic approximations.
Ab Initio Molecular Dynamics Calculation of Isotope Fractionation Between Borate-Boric Acid in Aqueous Solution
- James Rustad
- University of California, Davis
Ab initio molecular dynamics calculations are used here to calculate vibration frequencies for B(OH)3(aq) and B(OH)4-(aq). We show that previous calculations have either underestimated or omitted altogether a major fractionating vibrational mode. The new results indicate that the 11B partitions into B(OH)4- in water, in contrast to recent experimental measurement of the fractionation factor. The discrepancy appears to result from using finite-temperature vibrational frequencies in the standard harmonic expression for the fractionation factor. While our results connect the measured spectrum to previous harmonic electronic structure calculations, they indicate that harmonic frequencies must be extracted from experimental vibrational spectra before they can be used in the standard expressions.
Coupled-Cluster Linear Response Properties for Very Large Systems Using New Functionality Within NWChem
- Jeff R. Hammonda
- James Francka
- Karol Kowalskib
- Wibe de Jongb
- aUniversity of Chicago
- bPacific Northwest National Laboratory
Coupled-cluster linear response theory, recently added to NWChem using the TCE, is used to study dipole polarizabilties of extended aromatic systems as large as pentacene and with more than 700 basis functions. We evaluate the accuracy of various density-functionals for these systems using CCSD (coupled-cluster singles and doubles). Although the failure of DFT to describe electric properties of extended systems is known, the coupled-cluster results allow us to quantify this breakdown better than with experiment. A systematic comparison of the Sadlej, Dunning and Pople basis sets is also performed.
NWChem Performance Analysis
- Edoardo Aprà
- Oak Ridge National Laboratory
We analyze benchmark results of the NWChem computational chemistry code on parallel computers. Benchmark data was collected applying Quantum Mechanical methods (e.g., Density-Functional Theory, second-order Møller-Plesset theory MP2) on molecules of increasing size (corresponding to increasing computational complexity in terms of aggregate use of memory and disk resources). Analysis of benchmark data will focus both on serial and parallel performance.
Development of the Cyclic Cluster Model for KS-DFT Methods and its Application to Covalent Periodic Systems
- Florian Janetzkoa
- Andreas M. Kösterb
- Dennis R. Salahuba
- aUniversity of Calgary
- bDepartamento de Quimica, CINVESTAV
The Cyclic Cluster Model (CCM) offers an alternative approach for the simulation of extended systems like polymers, surfaces or crystals. It combines the advantages of the Free Cluster Model (FCM) and the Supercell Model (SCM), and the results of CCM calculations can be directly compared with FCM results, since the same methodology (e.g. basis sets, functionals etc.) can be used. In this presentation we compare the three different models and give a detailed description of the CCM. The cyclic cluster formalism for first-principle KS-DFT methods is presented and the implementation of the CCM in the Kohn-Sham density functional theory program (KS-DFT) deMon2k1 is discussed. The deMon2k CCM was applied to covalent carbon-based systems periodic in one, two and three dimensions. Results of deMon2k CCM simulations of transpolyacetylene, graphene and diamond are presented. The optimized structures and electronic properties are compared with results of FCM calculations and available experimental data from the literature.
Reaction Paths and Excited States in H2O2+OH→HO2+H2O
- Bojana Ginovskaa
- Donald Camaionib
- Michel Dupiusb
- aWashington State University Tri-Cities
- bPacific Northwest National Laboratory
The mechanism of the hydrogen abstraction reaction H2O2+OH→HO2+H2O in gas phase was studied, using DFT (MPW1K) level of theory. We located 2 pathways for the reaction, both going through the same intermediate complex OH-H2O2, but via two distinct transition state structures that differ by the orientation of the hydroxyl hydrogen relative to the incipient hydroperoxy hydrogen. In one case, these hydrogens are on same side of the plane made by the 3 oxygen atoms (Transition state A) and in the other these hydrogens are on opposite sides of the plane (Transition state B). The precursor complex is 6.3 kcal mol-1 more stable than the reactants. Transition states A and B are respectively 0.9 and 1.5 kcal mol-1, above the reactants. The complex formed on the product side of the reaction is 41.1 kcal mol-1 below the energy of the reactants. The first two excited states were calculated for selected points of both pathways using time-dependent DFT, multiconfigurational quasi-degenerate-perturbation theory (MCQDPT2/ CASSCF) and equation of motion coupled cluster singles, doubles model (EOM-CCSD) EOMCCSD energies and completely renormalized EOM-CCSD(T)(IA) correction. An avoided crossing between the two excited states was found on both reaction pathways, on the product side of the barrier to H-transfer on the ground state surface, near the transition states. On the first excited state surfaces, there is a barrier for H abstraction of 19.6 kcal mol-1 for transition states A and 22.2 kcal mol-1 for transition states B, with respect to the precursor complex. The hydrogen transfer on the second excited state surface is barrierless. Our TDDFT calculations show that in the precursor complex, the first excited state is 0.31 eV, and the second excited state 3.33 eV above the ground state. The respective excited states for the successor complex are 1.18 eV and 6.5 eV above the ground state. The vertical excitation energies at the transition state A for the first and the second excited state are 0.74 eV and 2.43 eV respectively. For transition state B, these energies are 0.55 eV and 2.18 eV. The values for the energies given above are not zero-point corrected.
Direct Dynamics Trajectory Study of F- + CH3OOH Reactive Collisions
- Jose G. Lopeza
- Grigoriy Vaynera
- U.Lourderaja
- Srirangam V. Addepallia
- William L. Hasea
- Shuji Katob
- Wibe A. de Jongc
- Theresa L. Windusc
- aTexas Tech University
- bUniversity of Colorado
- cPacific Northwest National Laboratory
In this porter we present results of a theoretical investigation of the reaction dynamics of methyl hydroperoxide with F-. In order to study this reaction, we performed direct dynamics trajectory simulations at the B3LYP/6-311+G(d,p) level of theory using the general chemical dynamics program VENUS1 interfaced with the electronic structure theory software package NWChem.2 The trajectories were initiated in the reactant and transition state regions of the F- + CH3OOH potential energy surface and were propagated for a maximum of 4 ps. The simulation results are in excellent agreement with a previous experimental study.3 We observe two product channels, HF + CH2O + HO- and HF + CH3OO-. The reaction path followed by the trajectories that form the main reaction products, HF + CH2O + HO-, occurs via an ECO2 mechanism and differs from that defined by the intrinsic reaction coordinate. This suggests that the trajectories follow a path controlled by the dynamics instead of the path of steepest descent.
Kinetics of Triscarbonato Uranyl Reduction to Insoluble U+4 by Aqueous Ferrous Iron: Theoretical Study
- Matthew C. F. Wander1,2
- Sebastien Kerisit3
- Kevin M. Rosso3
- Martin A. A. Schoonen1,2
- 1Department of Geosciences, Stony Brook University, Stony Brook, NY 11794-2100
- 2Center for Environmental and Molecular Science (CEMS), Stony Brook University
- 3Pacific Northwest National Laboratory (PNNL), Richland, WA
Uranium is a pollutant whose mobility is tied to its oxidation state. While U(VI) in the form of the uranyl cation is readily reduced by a range of natural reductants, but carbonate greatly reduces its reduction potential. Very little is known about the elementary processes involved in uranium reduction from U(VI) to U(V) to U(IV) in general. In this study, we examine the theoretical kinetics of ET from ferrous iron to triscarbonato uranyl in aqueous solution. A combination of molecular dynamics (MD) simulations and density functional theory (DFT) electronic structure calculations using NWchem are employed to compute the ET parameters that enter into Marcus' model, including the thermodynamic driving force, reorganization energies, and electronic coupling matrix elements. MD simulations predict that two ferrous iron atoms will bind in an inner-sphere fashion to the three-membered carbonate ring of tricarbonato uranyl, forming the charge-neutral Fe2UO2(CO3)3(H2O)8 complex. The first ET step converting U(VI) to U(V) is predicted by DFT to occur at a rate on the order of 1 s-1. The second ET step converting U(V) to U(IV) is predicted to be significantly endergonic involving the catalytic transfer of two protons prior to the ET and would occur at a much slower rate of 1x10-18M/s-1. Therefore, U(V) is a kinetically stabilized end-product in this ET system.
Use of the Common Component Architecture Approach
- Curtis L. Janssen
- Sandia National Laboratories
We review the use of the Common Component Architecture approach within the quantum chemistry domain to tackle the software engineering challenges which arise as advanced algorithms are adopted and growing numbers of software packages are integrated to study complex, coupled physical phenomena. The development of common interfaces has allowed the adoption of advanced optimization solvers and high-level interchangeability of quantum chemistry packages. Components have been created which manage multiple levels of parallelism, providing much more efficient usage of parallel machines. Early efforts towards low-level integration of chemistry packages are examined. The ability to share intermediate data expands the capabilities available to any one software package, thereby enabling the rapid development of advanced methods. New methods for the study of reactions involving heavy elements, which depend on our component environment, are highlighted.
Electron/Hole Transport in TiO2 from First Principles
- N. Aaron Deskins
- Michel Dupuis
- Pacific Northwest National Laboratory
This work focuses on the intrinsic electron/hole transport in stoichiometric TiO2. Charge hopping is described by a polaron model, whereby a negative/positive polaron is localized at a Ti3+/O- site and hops to an adjacent Ti4+/O2- site. Polaron hopping is described via Marcus theory formulated for polaronic systems and quasi-equivalent to Emin/Holstein/Austin/Mott theory. We obtain the relevant parameters in the theory (namely the activation energy δG*, the reorganization energy ?, and the electronic coupling matrix elements Vab) for selected crystallographic directions in rutile and anatase, using periodic DFT+U and Hartree-Fock cluster calculations. The DFT+U method was required to correct the well-known electron self-interaction error in DFT in the calculation of polaronic wavefunctions. Our results give non-adiabatic activation energies of similar magnitude in rutile and anatase, all near ~ 0.3 eV for electron hopping and ~ 0.5 eV for hole hopping. The electronic coupling matrix element, Vab, was determined to be largest for electron polaron hopping parallel to the c direction in rutile and indicative of adiabatic transfer (thermal hopping mechanism) with a value of 0.20 eV, while the other directions investigated in both rutile and anatase gave Vab values about one order of magnitude smaller and indicative of diabatic transfer (tunneling mechanism) in anatase. Adiabatic transfer (large Vab) was predominantly seen for bulk hole transport. Results also show a larger activation energy for hole transport on the rutile (110) surface compared to bulk.
Characterization of nanostructure of solvated polymer electrolyte membrane using atomistic simulations
- Arun Venkatnathan
- Ram Devanathan
- Michel Dupuis
- Chemical and Materials Science Division, Fundamental Science Directorate, Pacific Northwest National Laboratory
Polymer Electrolyte Membrane fuel cells (PEMFC) play a key role in a hydrogen economy due to their high efficiency and minimal pollution and are promising candidates for various domestic and industrial applications. However, limitations in their performance under desirable fuel cell operating conditions of temperature points to the need of developing new membrane materials for use in PEMFC. The design and development of new membrane materials requires an understanding of the nanostructure of these membranes, their chemical structure and reactivity, and how these influence proton conductivity under low hydration conditions. As a first step we carried out classical molecular dynamics simulations to examine the structure of NafionTM (DuPont) membranes under varying conditions of membrane hydration and temperature. The nanostructure of hydrated Nafion membrane, vehicular transport of the solvated protons and water molecules along with structural and dynamical properties of the hydrated Nafion under different conditions will be presented.
This work is supported by the Division of Chemical Sciences, Office of Basic Energy Sciences, U.S. Department of Energy (DOE). Battelle operates the Pacific Northwest National Laboratory for the DOE.
