Active Computing Research Projects

CIR Projects

EMSL's CIR projects focus on environmental molecular science basic and applied research areas that address the environmental problems and research needs facing the U.S. Department of Energy (DOE) and the nation.

• A Systematic Molecular Modeling Study of the Effect of Lipid Bilayer Composition on Resistance to Alcohol-Induced Changes
  Project Lead: Roland Faller
  Project Lead Institution: University of California, Davis
  Research Area: Biological Interaction Dynamics
  FY10 Allocation: 500,000 node-hours
  Abstract

  Production of ethanol and longer alcohols like butanol from cellulosic raw materials in a biorefinery setting is a promising sustainable alternative to petroleum-based energy and chemicals. While yeasts can be genetically engineered to degrade a variety of sugars or even cellulose directly, alcohol tolerance is a key issue in increasing the productivity and of a biorefinery-based production. A better understanding of alcohol tolerance in microorganisms and the effects of lipid bilayer composition on resistance to alcohols would allow us to take a more rational approach to improving biorefinery production strains. This is the main driving force behind this proposal.

  We have evidence that the changes in sugar-to-alcohol-metabolism in yeast cells are in part due to changes in the cell membrane in the presence of alcohol. These changes in the lipid bilayer include a structural transition called interdigitation which most likely affects the conformation of any transmembrane protein in the membrane and eventually disables them. Also mechanical properties of the membrane are crucial for protein function and they are altered in an alcohol containing environment as well.

  This proposal aims at a detailed understanding of this issues and aims at a multiscale simulation approach to the interaction of phospholipid bilayers with alcohols. The first part of the proposal focuses on models of yeast membranes with alcohols which are very important for improving the efficiency of bioreactors. This is based on a USDA funded project to study alcohol tolerance of different biomembranes in detail. The PI of this proposal is co-PI on the USDA project. In the second part we address large scale effects on mechanical properties by changing the environment.
• Advanced Biomolecular Simulations - Development and Applications
  Project Lead: T. P. Straatsma
  Project Lead Institution: Pacific Northwest National Laboratory
  Research Area: Biological Interaction Dynamics
  Proposal 30994 - FY10 Allocation: 1,000,000 node-hours
  Abstract

  This project focuses on two important components of biomolecular modeling and simulation: i) the development of novel molecular modeling capabilities and their efficient parallel implementations, and ii) the application of these capabilities for the study of environmentally important complex biological systems. Our research will focus on massively parallel implementations of methodologies that enable peptide and protein simulations at significantly larger scales compared to current applications, in the time and spatial domains, as well as in the complexity of interaction models for large, heterogeneous biomolecular assemblies. These include the implementation of multi-ensemble technologies, such as replica-exchange, in the NWChem framework, with the promise of reaching high scalability in molecular dynamics simulations for highly complex systems. The second methodology targeted in this project is further development of Brownian dynamics methodology to allow millisecond time-scale simulations of multi-protein complexation. This will be implemented in the SDAMM software, based on the concept of the Simulation of Diffusional Association (SDA) program for systems consisting of rigid protein domains with or without flexible linkers between them with modeling of the forces between the rigid domains based on their atomic detail structures.

  These methodologies will be applied to molecular simulations of a number of environmentally important biomolecular systems. We will continue our work on characterizing the effect of the Gram negative lipopolysaccharide outer membranes on the stability and dynamics of trans-membrane proteins. The primary focus is on a characterization of the free energy profile for material transport across the membrane through porins and protein channels by calculation of the potentials of mean force of ions and small molecules moving through trans-membrane ion channel proteins and specific and non-specific porins. We will investigate the effect of environmental, ionic compounds on the structural integrity of the Gram-negative microbial outer membrane of the environmentally ubiquitous microbe Pseudomonas aeruginosa. The second system to be studied is the MHC class I proteins which are of great importance for the fundamental understanding of the interaction of the immune system with the environment and changes of the environment. These are membrane-bound proteins of the immune system that present antigenic peptides at the cell surface for which peptide loading in the endoplasmic reticulum is a crucial step involving interactions with several proteins to form a nano-molecular loading complex. We will investigate the coupling of conformational flexibility and loading characteristics in long time replica exchange MD simulations. The Brownian dynamics capabilities will be demonstrated in studies of multi-domain signaling proteins and how post-translational modification of flexible as well as rigid parts affects their structure, dynamics, and interactions and therefore function. This model is also applicable to membrane attached proteins often involving flexible regions and domain reorganizations. The simulation of protein solutions is important to understand phenomena such as cellular crowding or protein mixtures. The application components will be reported in the peer-reviewed scientific literature.
• Computational Chemical Dynamics of Complex Systems
  Project Lead: Donald G. Truhlar
  Project Lead Institution: University of Minnesota
  Research Area: Interfacial Phenomena
  FY10 Allocation: 640,000 node-hours
  Abstract

  The objective of this project is to develop and apply innovative high-performance computing techniques and simulation methods in order to address computationally challenging problems in chemical dynamics, with special emphasis on the critical problems in environmental science and chemical engineering facing the DOE and the nation. The proposal is concerned with several fundamental areas of research including thermochemical kinetics and rate constants, photochemistry and spectroscopy, chemical and phase equilibria, and heterogeneous catalysis. These areas are important for solar energy, fuel-cell technology, environmental remediation, weather modeling, pollution modeling, and atmospheric chemistry. These computationally intensive studies will be carried out with new high-throughput integrated software that we have been developing. The development of compatible, portable, scalable, and user-friendly computational tools that combine electronic structure packages with dynamics codes and efficient sampling algorithms will be continued as part of this project. The proposal features four EMSL themes: atmospheric aerosol chemistry, biological interactions and dynamics, geochemistry and subsurface science, and science of interfacial phenomena. In the field of atmospheric aerosol chemistry, we propose a study of nucleation phenomena which play a pivotal role in many atmospheric and technological processes. We propose to develop paradigm-shifting, scalable computational approaches for modeling the nucleation, structure, and properties of nano-droplets.

  Another aspect of our research is the development of efficient and robust methods for analytical representations of multidimensional potential energy surfaces for photochemical reactions including those of environmental and energetic importance. In the field of biological interactions and dynamics, we will study explicit polarization effects in various molecules and biochemical systems, such as protein residues and hydrogen bond complexes using re-parameterized semiempirical models and molecular mechanics force fields. We will also explore Feynman path integral methods in order to incorporate quantum effects such as tunneling and zero-point energy into the treatment of large molecules. In the field of geochemistry and subsurface science, we propose large-scale Monte Carlo simulations of silica melts. Silica plays a significant role in the chemistry and mineralogy of the Earth's crust and mantle. In the field of interfacial phenomena, we propose to provide molecular-level insights on retention mechanisms in reversed-phase liquid chromatography. These mechanisms are not well understood and there remain many open questions on bonded-phase conformation, solvent penetration, solvophobic versus lipophilic interactions, and partition versus adsorption. We are interested in studying interfacial phenomena related to heterogeneous catalysis, especially, involving transition-metal compounds and zeolite frameworks. In particular, we plan to study the electrochemistry of water oxidation by binuclear copper and ruthenium catalysts. The water oxidation process is a difficult component of the challenge to efficiently convert solar radiation into a chemical fuel. Another area of interest is a study of nano-gold clusters in order to relate their optical and structural properties to their enhanced catalytic activity in CO oxidation. We will also develop new potential energy functions for the study of adsorption isotherms of hydrocarbons in zeolites and we will interface the new potentials with a Monte Carlo Gibbs ensemble algorithm to calculate the adsorption isotherms. In addition, substantial efforts will be put in the modeling of ion solvation and ion transport through biological membranes, geological minerals, or in electrolytes.
• Computational Design of Catalysts: The Control of Chemical Transformation to Minimize the Environmental Impact of Chemical Processes
  Project Lead: David A. Dixon
  Project Lead Institution: University of Alabama
  Research Area: Interfacial Phenomena
  Proposal 34000 - FY10 Allocation: 1,800,000 node-hours
  Abstract
  Catalysis is governed by a delicate balance between a myriad of competing bond making and breaking processes including adsorption, reaction, desorption, and surface diffusion that occur at active catalytic centers. These processes are explicitly controlled by the intrinsic bonds that form between the reactants or intermediates and the active catalytic site as well as by the local nanoscale environment about the site. For example, the specific atomic structural configuration and chemical makeup of the ligands that surround an active organometallic center in homogeneous catalysis; and the pore size, local acidity, and specific atomic structure in zeolite control their activity and selectivity in ways that are similar to that found in enzymes. Similarly, supported metal particles and metal oxides are complicated by their complex and ill-defined structure. The catalytic behavior is governed by their size and shape, interaction with the support, composition and atomic configuration for metal alloys and mixed metal oxides, the influence of solution as well as the presence of electric fields or applied potentials. We propose to use advanced computational chemistry approaches implemented on massively parallel computers to develop a quantitative description of catalysts so as to develop new design criteria and to develop new understanding of the physical phenomena that occur at different spatial and temporal scales and that underlie catalytic behavior. Catalysis is about improving kinetics and catalyst design will require quantitative information about transition states for critical reaction processes. Currently information about transition states, especially geometric and spectral information, is only readily accessible by computational methods. Yet such information is critical if we are to develop a language that can describe the events at the atomic level that describe homogeneous, heterogeneous and bio-catalysis. Computational chemistry is an enabling tool for addressing challenges in the optimal design of processes for controlling and enabling chemical transformations leading to processes that have high selectivity, have minimal environmental impact, and are optimal in their use of energy. We will apply the techniques of computational chemistry to address a variety of problems in catalysis science including: oxidative dehydrogenation; organic oxidation chemistry and selectivity; hydrogenation of alkenes and isomerization of alkenes and alkanes; mixed homogeneous/heterogeneous catalysts; transition metal oxides; ligand design for C-C coupling reactions; water-gas-shift (WGS) and preferential oxidation (PROX) of CO in the presence of H2 with metals and metal oxides; catalytic partial oxidation of methane over metal nanoparticles; structure function relationships for zeolite acid catalysis; C-H bond activation over promoted metal sulfides; metal particle sintering on alumina; effect of the metal-support interaction on catalytic conversion of CO2; O2 dissociation at a (111) FCC metal surface; acid properties of polyoxometallate clusters; formation of C-C bonds from C1 species, specifically during the formation of hydrocarbons and alcohols from mixtures of H2 and CO on Fe, Co, and Rh; reactivity of small molecules gamma-Al2O3 supported catalysts; and quantitative prediction of reaction intermediates for homogeneous electrocatalysis.
• Computer Simulations of Protein-Nanomaterial Interactions
  Project Lead: Roberto D. Lins
  Project Lead Institution: Pacific Northwest National Laboratory
  Research Area: Biological Interaction Dynamics
  FY10 Allocation: 700,000 node-hours
  Abstract

  The development of nanotechnology has led to an increasing interest in interfacing biological molecules with nanomaterials. Proteins have been used to functionalize nanomaterials controlling their synthesis, assembly and influencing their properties for applications from biosensing to delivery. In turn, the surface chemistry of a nanoparticle also influences the structure and function of the adsorbed proteins. Some specific interactions between nanomaterials and proteins in living cells have recently been shown to be cause of a series of adverse health effects. Therefore, the understanding of the molecular basis of how nanomaterial properties affect protein structure and function becomes imperative for the design of not only smart bionanocomposites with biosensing or delivery properties, but also to ensure the safety of the designed nanomaterials. This proposal seeks high performance computing allocation for the development and application of computational methodologies aiming at the characterization of the effects of nanomaterials on protein structure, stability and dynamics. The rationale of such interactions will be drawn based on a variety of simulations specifically involving i) the adsorption of cell-surface receptors and plasma proteins to silica,. polystyrene and asbestos particulates; and, ii) the effect of enzyme confinement on mesoporous silica. This effort involves a synergistic collaboration between experimental and computational teams at PNNL (Computational Biology and Bioinformatics, EMSL and Biological Sciences) and Ohio State University.
• Density Function Theory Studies of Complex Systems: Structure, Dynamics, and Excited States
  Project Lead: L. Rene Corrales
  Project Lead Institution: University of Arizona
  Research Area: Interfacial Phenomena
  FY10 Allocation: 176,000 node-hours
  Abstract
  Our goal is to use DFT approaches to investigate structural and dynamical properties of composite systems by partitioning the underlying elements of such complex systems into the components that are computationally tractable. Projects include solvation studies of actinide ions in polar and non-polar solvents, structure and dynamics of hydrogen storage in nanoparticle frameworks, determination of structure and photoelectron spectrum of molecular systems used in thin-film photovoltaic devices, the role of transition metals in light-harvesting systems, characterization of hybrid organic-inorganic host systems, and luminescent lanthanide complexes used as optical probes for biomedical imaging applications. These projects overlap critical DOE mission objectives that include increasing the knowledge of the underlying factors that control purification processes for actinides (i.e. nuclear fuels), developing predictability capabilities to assist experiment in the characterization and development of renewable energy resources that includes hydrogen storage and light harvesting systems, characterizing the role of host matrices for alternative lighting systems with an emphasis on organic light emitting diodes, and to develop novel lanthanide-based luminescent materials to provide in vitro and in vivo optical imaging. At least half of these projects are in collaboration with experimental projects where the rest are anticipated to reach out to the experimental community for validation and verification. EMSL computing resources are required as the systems of interest are large with respect to the number of atoms, and with respect to the number of electrons. These projects have been initiated with smaller systems and smaller basis sets. Access to teraflop computing allows for more accurate descriptions as a result of using better basis sets for transition metals and actinides, and using larger systems to study kinetic properties (e.g. diffusion and reaction barriers) in condensed systems where response of the local environment plays a critical role. In addition, recent and ongoing theoretical developments are being incorporated into codes to study systems with difficult electronic structure configurations that are strongly affected by the complex electrochemical environments in which they reside.
• Development of accurate force fields for aqueous biological environments using a massively parallel multiscale approach
  Project Lead: Jeff R. Hammond
  Project Lead Institution: University of Chicago
  Research Area: Biological Interaction Dynamics
  Proposal 30794 - FY10 Allocation: 900,000 node-hours
  Abstract
  The ability to calculate accurate energies and properties of large molecules with massively-parallel computers presents an extraordinary opportunity to develop new models for bonded and non-bonded interactions used in classical force fields by directly calculating such parameters from first principles electronic structure calculations. Since those force fields usually contain a large set of parameters, especially those that describe polarization, it becomes increasingly difficult to determine these parameters in a rigorous way using only experimental data. In contrast, quantum chemical calculations accurately determine force constants, charge distributions, and polarizabilities for the relevant moieties with well-known accuracy. The coupled-cluster method is universally accepted as the most accurate method for molecules near their equilibrium geometry, and the functionality to calculate all relevant properties for determining force fields is available within NWChem. Using NWChem on the new EMSL supercomputer, polarizable force field parameters will be determined using first principles for use in both classical and nuclear quantum statistical mechanical simulations. The use of those state-of-the-art quantum chemical methods for force field development will increase the accuracy and generality of existing models and contribute significantly to the development of a hierarchy of increasingly accurate models. The emergence of a Jacobs ladder of models for molecular simulations, in analogy to that in density functional theory [J. P. Perdew and K. Schmidt, in "Density Functional Theory and Its Application to Materials, edited by V. Van Doren, C. Van Alsenoy, and P. Geerlings (AIP, Melville, NY, 2001)], promises to make significant contributions to molecular simulations relevant to all four of EMSL's science themes. The open (non-proprietary) nature of this project ensures that it will have maximum impact on the user community for the respective models developed.
• Electron and Proton Transfer Reactions in Photo-biological H₂ Production
  Project Lead: Thereza A. Soares Da Silva
  Project Lead Institution: Pacific Northwest National Laboratory
  Research Area: Biological Interaction Dynamics
  FY10 Allocation: 273,000 node-hours
  Abstract

  The biological photoproduction of hydrogen by microbial photosynthetic organism requires light as the energy source, an electron-donating substrate, and a catalyst that combines electrons and protons to generate molecular hydrogen. At the heart of this process lies electron and proton transfer reactions mediated by structural changes at femto to picosecond timescales. Although high-resolution X-ray structures of many of the protein complexes involved in photosynthesis and metabolism are available, the mechanistic details of how proton and electron transfer reactions occur at a molecular level are not well understood. Relevant time and spatial resolution are notoriously difficult to access experimentally, and therefore computational simulations are the methods of choice to investigate the relationship between electron/proton transfer rates and molecular dynamics underlying photo-biological H₂ production.

  Electron and proton transfer reactions will be studied for two constituents of oxygenic photosynthesis, hydrogenases and the photosynthetic reaction center of Blastochloris viridis. We will use electron transfer simulation methodologies based on Marcus' theory, and classical molecular dynamics simulation to sample representative molecular environments for the electron transfer between redox sites, the simulation of proton hopping using the semi-classical QHOP and QM/MM methodologies, and thermodynamic integration to evaluate relative free energies of the possible redox states of the reaction center complex and hydrogenases. Ultimately, our research aims to provide a detailed molecular-level characterization and understanding of microbial processes that may play an important role in biological light harvesting and hydrogen production. The proposed research project will take advantage of the computational chemistry software system NWChem which, in addition to a wide range of electronic structure capabilities for density functional theory and ab initio molecular orbital theory calculations, also supports molecular dynamics (MD) calculations with a variety of empirical (classical mechanical) force fields for the simulation of macromolecular and solution systems.
• Establishing first-principle description of heavy fermion materials
  Project Lead: Sergey Y. Savrasov
  Project Lead Institution: University of California, Davis
  Research Area: Chemistry
  FY10 Allocation: 300,000 node-hours
  Abstract
  Utilizing a state-of-the-art methodology for first-principle simulation of strongly correlated materials, we study rare earth compounds and actinides of current interest including pure Ce, Ce₂O₃, the full rare earth sesquioxide series, CeTIn₅ (T=Co,Rh,Ir), Plutonium and PuCoGa₅. We will also study the narrow band semiconductors CeRhSb and CeRhAs. These rare earth based materials are of interest for use as catalytic converters, solar cells and as environmentally friendly pigments for plastic coloration. The Kondo insulators have highly tunable semiconducting gaps which may find a widespread application in industry. Also the storage of radioactive actinide compounds yielded by nuclear power plants has been one of the biggest issues associated with general DOE mission and the environmental concern. Being equipped with the right theoretical tools we target the problems related to recent DOE energy initiatives.
• First Principles Computations of Interfacial Phenomena for Environment-Friendly Catalysis
  Project Lead: Larry Curtiss
  Project Lead Institution: Argonne National Laboratory
  Research Area: Interfacial Phenomena
  FY10 Allocation: 1,280,000 node-hours
  Abstract

  A fundamental knowledge of the chemical and physical properties at gas-solid or liquid-solid interfaces is essential for the design and discovery of breakthrough materials for cleaner and more efficient catalysts. Computational strategies using high performance computing provide the opportunity to design surfaces and interfaces with desired functionalities. This requires both an understanding of the processes that are occurring at the interfaces as well as relating catalytic properties to key calculated parameters such as adsorption energies, which can be used in screening many potential candidates. This proposal is focused on both the understanding of structure-function relationships as well as use of this insight to help design new materials for processes involving heterogeneous catalysis, electrocatalysis, and photocatalysis. The computational studies will be based on state-of-the-art electronic structure methods including density functional theory, a powerful tool for understanding and design of new catalysts. The proposed computational studies involve collaboration with world-class experimentalists at the Pacific Northwest National Laboratory and Argonne National Laboratory.

  New catalysts are needed for making and breaking specific bonds for energy efficient and environmentally friendly catalysts. Recent studies have shown that supported subnanometer clusters exhibit novel catalytic properties for breaking C-H bonds. We will carry out systematic studies to explore the unusual catalytic properties of these clusters and to determine how they might be tailored to make and break specific bonds. These computations will play a key role in the design and discovery of environmentally friendly catalysts for industrial processes. Another interesting area of heterogeneous catalysis is synthesis of nanocarbons, which have many potential applications.

  Electrocatalysis provides a low-temperature, energy efficient alternative to heterogeneous catalysis that has potential uses for chemical transformations as well as the development of alternative energy sources. We will investigate carbon dioxide electroreduction, wherein CO₂ is electrochemically reduced to CO and hydrogenated products in acidic solution. This reaction has the potential to both produce useful chemicals from an essentially limitless feedstock and reduce atmospheric greenhouse gas loadings. Among the most exciting possibilities presented by this reaction is the electrochemical production of methanol. We will also investigate oxygen reduction catalysis, which is important for energy conversion by fuel cells.

  A third focus of the proposed work is photocatalysis. Efficient utilization of solar energy through photocatalysis would be a significant step towards easing current energy demands. TiO₂ is a prototypical photocatalyst used for organic pollutant degradation and also for water splitting (H₂O => H₂ + ½O₂), H₂ being an important fuel source. Despite the importance of TiO₂, there are many fundamental issues that are not well understood, such as how photoexcited holes and electrons interact with surface adsorbates and lead to their catalytic decomposition. Molecular modeling methods provide an excellent way to study surface reactions and provide insights into the intermediates and reaction pathways. We will study the reactivity of organic molecules over TiO₂, better characterize the nature of electrons/holes and their transport properties, and model the interactions of holes/electrons at the surface in the presence of adsorbate molecules.
• High-throughput sequence analysis for communities and environmental sample metagenomics
  Project Lead: Christopher S. Oehmen
  Project Lead Institution: Pacific Northwest National Laboratory
  Research Area: Biological Interaction Dynamics
  FY10 Allocation: 800,000 node-hours
  Abstract
  Metagenomes and multiple genome samples include environmental samples from soil and water, community samples involving multiple species (such as human or termite gut) and others. DOE has recognized the likely importance of analyzing metagenome datasets as a key capability for enabling carbon management, renewable fuels, environmental cleanup and a host of other application domains. Understanding complex interactions between physical and chemical processes in these populations requires computational tools and capabilities for basic analysis such as gene sequence alignment, homolog detection, and others. However, metagenomic datasets in many cases already are too large for conventional analysis. Hence DOE capabilities such as ScalaBLAST and the MSC are uniquely suited to address this significant challenge for DOE and a wider research community. The proposed Science Themes work will provide a bridge to make high-throughput sequence analysis available to a wide spectrum of EMSL users for multiple genome and metagenome research.
• Highly scalable molecular scale software development for environmental sciences
  Project Lead: Theresa L. Windus
  Project Lead Institution: Iowa State University
  Research Area: Chemistry
  FY10 Allocation: 345,000 node-hours
  Abstract
  Understanding the complex chemistry and physical properties of systems ranging from small gas-phase clusters to the solution phase is one of the major challenges for environmental sciences. Many previous efforts to examine the dynamical properties of these systems have involved the use of classical, empirical or semi-empirical potentials for the molecular interactions which have complications due to the fitting of the potentials for bulk properties as well as not being suitable for bond making and breaking processes. Ab initio methods combined with statistical methods, while quite expensive, offer the promise of providing a general framework for examining these systems. In this proposal, development and testing of highly scalable methods will be accomplished using two of the most generally available and used computational chemistry codes - NWChem and GAMESS. In particular, this research will focus on the development of new algorithms for Dynamical Nucleation Theory Monte Carlo (DNTMC), molecular dynamics using the fragment molecular orbital (FMO) method, and a new effective fragment method for long range effects in solvents. These methods will impact various environmental sciences including the formation dynamics of aerosol clusters in the atmosphere, the properties of molecular reactions and solution structure of ions (including actinides and heavy elements) in water and other solutions including ionic liquids used in "Green Chemistry". Multi-level parallelism and automatic dynamic load balancing as well as on the fly surface fitting and component technology will be used to enable a high level of scalability in these algorithms. The computational resources at the MSCF are ideal for this development. The architecture is similar to those which are typically available as a group level resource, but are available at a much larger scale than can be afforded by any researcher. This provides a useful migration path from development of initial software on local resources to large scale development and testing required for the actual production of results that will impact the science. In addition, the staff expertise required for trouble shooting issues is available at the MSC.
• Interfacial Properties of Biomolecular Systems: Mechanism and Kinetics of Association and Melting
  Project Lead: B. Monti Pettitt
  Project Lead Institution: University of Houston
  Research Area: Biological Interaction Dynamics
  FY10 Allocation: 1,275,000 node-hours
  Abstract

  High performance parallel computational resources and scalable software has enabled computational chemistry to model actinides at various interfaces, in the presence of multiple ligands in solution, and to obtain a basic understanding of actinide sorption and redox behavior in the subsurface as well as in solutions critical to minimize the generation of new radioactive waste from fuel reprocessing. We propose to study the influence of the local environments on sorption, redox mechanisms and oxidation state stability of actinides in solution, especially for nanoparticle and colloid formation, and at solution/mineral interfaces. We believe that computational chemistry methods that integrate electronic structure and classical techniques provide invaluable information that will make currently existing surface complexation and field-scale models more accurate, and will provide critical guidance that could enhance the effectiveness of separations schemes. We propose to study the actinides thorium through curium in their relevant oxidation states: 1) with nitrate, carbonate, sulfate and phosphate aqueous co-contaminants, interacting with mineral interfaces that including magnetite, quartz, and goethite, 2) forming colloidal or nanophase actinide hydroxides, hydrous oxides and oxides in aqueous solution, and 3) incorporated in actinide containing oxides and minerals. We propose to study these systems with atomistic simulations, using quantum and classical mechanical models through molecular dynamics. We will use Gaussian and Car-Parinello plane-wave density functional theory (DFT) with relativistic scalar and spin-orbit effects, and will properly account for the influence of the local molecular environment by using a variety of methods including solvent reaction fields, explicit inclusion of solvents, and hybrid point-charge models for extended systems. Ab initio and classical molecular dynamics simulations will be performed to obtain a continuous description of the molecular- and meso-scale reactivity and redox behavior of the actinides in ground water conditions and with co-contaminants as well as for models of reprocessing solutions, and to study the properties of colloids and nanoparticles.

  All-atom molecular dynamics simulations will be used to study the melting of DNA. For the stress studies the calculations will be composed of a DNA duplex with at least three complete turns, in solution. Periodic boundary conditions will be used to connect the ends of the duplex at the edge of the simulation box and the linking number is altered by either adding base pairs to underwind the DNA or removing base pairs to overwind relative to the relaxed structure. The now familiar multiscale approach which uses data from calculations on a shorter length and time scale as input parameter for simulations at a longer time scale will be used to study the protein-DNA association. Brownian dynamics simulations coupled with all-atom molecular dynamics simulations will be used to understand the protein recognition and binding with DNA. Finally, the RNA hairpin folding studies will be performed with temperature dependent replica exchange molecular dynamics simulations.
• Large-scale computational modeling of the chemical behavior of actinide elements at interfaces
  Project Lead: Wibe A de Jong
  Project Lead Institution: Pacific Northwest National Laboratory
  Research Area: Geochemistry and Interfacial Phenomena
  Proposal 29990 - FY10 Allocation: 1,000,000 node-hours
  Abstract
  High performance parallel computational resources and scalable software has enabled computational chemistry to model actinides at various interfaces, in the presence of multiple ligands in solution, and to obtain a basic understanding of actinide sorption and redox behavior in the subsurface as well as in solutions critical to minimize the generation of new radioactive waste from fuel reprocessing. We propose to study the influence of the local environments on sorption, redox mechanisms and oxidation state stability of actinides in solution, especially for nanoparticle and colloid formation, and at solution/mineral interfaces. We believe that computational chemistry methods that integrate electronic structure and classical techniques provide invaluable information that will make currently existing surface complexation and field-scale models more accurate, and will provide critical guidance that could enhance the effectiveness of separations schemes. We propose to study the actinides thorium through curium in their relevant oxidation states: 1) with nitrate, carbonate, sulfate and phosphate aqueous co-contaminants, interacting with mineral interfaces that including magnetite, quartz, and goethite, 2) forming colloidal or nanophase actinide hydroxides, hydrous oxides and oxides in aqueous solution, and 3) incorporated in actinide containing oxides and minerals. We propose to study these systems with atomistic simulations, using quantum and classical mechanical models through molecular dynamics. We will use Gaussian and Car-Parinello plane-wave density functional theory (DFT) with relativistic scalar and spin-orbit effects, and will properly account for the influence of the local molecular environment by using a variety of methods including solvent reaction fields, explicit inclusion of solvents, and hybrid point-charge models for extended systems. Ab initio and classical molecular dynamics simulations will be performed to obtain a continuous description of the molecular- and meso-scale reactivity and redox behavior of the actinides in ground water conditions and with co-contaminants as well as for models of reprocessing solutions, and to study the properties of colloids and nanoparticles.
• Mechanistic Modeling of Subsurface Multifluid Flow and Biogeochemical Reactive Transport
  Project Lead: Steve B. Yabusaki
  Project Lead Institution: Pacific Northwest National Laboratory
  Research Area: Subsurface Modeling
  FY10 Allocation: 750,000 node-hours
  Abstract
  Our goal is to use DFT approaches to investigate structural and dynamical properties of composite systems by partitioning the underlying elements of such complex systems into the components that are computationally tractable. Projects include solvation studies of actinide ions in polar and non-polar solvents, structure and dynamics of hydrogen storage in nanoparticle frameworks, determination of structure and photoelectron spectrum of molecular systems used in thin-film photovoltaic devices, the role of transition metals in light-harvesting systems, characterization of hybrid organic-inorganic host systems, and luminescent lanthanide complexes used as optical probes for biomedical imaging applications. These projects overlap critical DOE mission objectives that include increasing the knowledge of the underlying factors that control purification processes for actinides (i.e. nuclear fuels), developing predictability capabilities to assist experiment in the characterization and development of renewable energy resources that includes hydrogen storage and light harvesting systems, characterizing the role of host matrices for alternative lighting systems with an emphasis on organic light emitting diodes, and to develop novel lanthanide-based luminescent materials to provide in vitro and in vivo optical imaging. At least half of these projects are in collaboration with experimental projects where the rest are anticipated to reach out to the experimental community for validation and verification. EMSL computing resources are required as the systems of interest are large with respect to the number of atoms, and with respect to the number of electrons. These projects have been initiated with smaller systems and smaller basis sets. Access to teraflop computing allows for more accurate descriptions as a result of using better basis sets for transition metals and actinides, and using larger systems to study kinetic properties (e.g. diffusion and reaction barriers) in condensed systems where response of the local environment plays a critical role. In addition, recent and ongoing theoretical developments are being incorporated into codes to study systems with difficult electronic structure configurations that are strongly affected by the complex electrochemical environments in which they reside.
• Quantum Calculations as a Tool in Structural Biology
  Project Lead: Michael E. Green
  Project Lead Institution: City College of New York
  Research Area: Biological Interaction Dynamics
  FY10 Allocation: 675,000 node-hours
  Abstract

  We are planning to continue work on understanding ion channels by application of quantum calculations to these critically important proteins. Ion channels play a key role in essentially every living cell. Malfunctioning ion channels produce numerous diseases, not limited to diseases of excitable tissue, such as nerve and heart. Second, we will have prepared the tools for a much more accurate calculation of protein structure and function by improving hydrogen bond and salt bridge potentials so that they can be used in simulations, which allow many more atoms to be included in calculations. This applies to both biological problems and surface chemistry science themes.

  We have already completed enough of these calculations to be confident that the results are significant. Two questions have been raised about whether the results would accord with experiment: i) quantum optimizations give results at 0 K, not room temperature; ii) they require truncation of the system. One aim of the next year's work is to answer these comments. Fortunately, we already have results for enough systems, with comparison to experimental work, to have a high degree of confidence that these objections can be answered successfully; two new collaborative agreements will make possible more directed experiments, and a third will allow incorporation of some of the results into improved simulations. We are already doing frequency calculations that allow us to bring the optimizations to room temperature for thermodynamic quantities. A major part of the next year's work will be devoted to demonstrating the quality of the calculations by further comparison with experimental results. Also, we are developing potentials for hydrogen bonds and salt bridges that include the effects of surroundings, for use in simulations. We expect to add an extremely valuable technique to those available for understanding the functioning not only of ion channels, but of proteins more generally.

  Standard molecular dynamics (MD) potentials are parameterized for bulk solutions; we have shown that these can be significantly in error in confined spaces. The potentials that we have already found are being prepared for use in Monte Carlo simulations; we expect to complete this in the next year. The necessary neighbor tracking program has been written, and it should be in usable form by the start of the first year of this project. Once this is available, QM/MM methods will also be more accurate; we will adapt those as well to our problem.

  We have proposed a mechanism for gating (opening and closing) of ion channels that includes a non-standard characteristic, proton transport as gating current; as evidence grows, our proposal seems more and more viable. Calculations on the voltage sensing domain (VSD) of an ion channel are needed; a very large number of experimental groups have worked on channel gating for several decades, with results that still leave the mechanism ambiguous. More recently, we have worked on the problem of how a channel chooses between K+ and Na+ (selectivity), second of the three fundamental questions in understanding ion channels (the third is conduction). The results so far, on the channel "cavity", point the way to what is needed for understanding selectivity; the "selectivity filter" itself remains to be done.

  In the next year, therefore, we will be able to test some of the main ideas on which we have been working. Two additional years may make it possible to completely reap the benefits of the past five years of effort, and the computer time so far invested.
• Reliable Electronic Structure Prediction of Molecular Properties
  Project Lead: Jerome D. Fast
  Project Lead Institution: Pacific Northwest National Laboratory
  Research Area: Aerosol Modeling
  Proposal 30993 - FY10 Allocation: 480,000 node-hours
  Abstract

  The direct (scattering and absorption of radiation) and indirect (cloud-aerosol interactions) effects of aerosols predicted by climate models still contain large uncertainties. Many of these uncertainties relate to the inability of current models to accurately simulate the evolution of particulate mass, composition, size distribution, hygroscopicity, and optical properties and the current haphazard approach of developing new aerosol process modules. To address these issues, we are developing an Aerosol Modeling Testbed (AMT) designed to streamline the process of testing and evaluating refined aerosol process modules over a wide range of spatial and temporal scales and improve the scientific accuracy and computational speed of aerosol process modules used in climate models. The AMT consists of a modular version of a fully-coupled meteorology-chemistry-aerosol model (WRF-chem) and a suite of tools to automatically evaluate the performance of aerosol process modules via comparison with a wide range of field measurements. A modular model enables various treatments of specific aerosol processes to be systematically compared, while all other atmospheric processes (emissions, gas chemistry, meteorology, other aerosol processes, etc.) remain the same. Examples of specific aerosol processes that require further refinements to better simulate particulate evolution in climate models include new particle formation, representing the aerosol size distribution, formation of secondary organic aerosols, aerosol optical properties, aerosol-cloud interactions, and better coupling of meteorology with primary particulate emissions (dust and sea-salt) and deposition.

  The simulations performed by WRF-chem are designed to resolve the spatial and temporal variations of particulate properties observed in the atmosphere. In this way, aerosol process module performance can be assessed using extensive measurements made during recent field campaigns. The design of the numerical experiments are consequently computationally expensive and require large amounts of storage to manage a large number of simulations; therefore, a use of the AMT by the scientific community requires a high performance computing facility. This proposal describes the tasks required to implement the AMT operationally on MSC's Chinook supercomputer including porting the WRF-chem code, testing parallel I/O, setting up the components of the AMT, developing data management strategies, performing beta tests of the AMT, and enabling long-term use of the AMT by the aerosol modeling community.
• Simulating surface-mediated proton transport at decorated surfaces
  Project Lead: Volkhard Helms
  Project Lead Institution: Saarland University, Germany
  Research Area: Biological Interaction Dynamics
  FY10 Allocation: 340,000 node-hours
  Abstract

  Proton translocation in biological membrane proton pumps is an essential part of many bio-energetic processes. Many membrane proton pumping proteins transport protons in a very efficient manner. The proton conductance via such proteins can be one order of magnitude faster than proton diffusion in bulk even when there exists a proton gradient against which the protons need to be transported. This fast proton transport requires an efficient proton uptake from the environment (e.g. the cellular solution). The proton uptake process is essentially a directed proton transfer near the protein-membrane / bulk solution interface. Understanding the proton diffusion mechanism near surfaces with different chemical composition (e.g. the protein/membrane surface) is not only important for deeper insight in basic biological phenomena but also has relevance in related application fields like the design of new nanoscale instrument like proton filer/pumps or more efficient hydrogen-fuel cells.

  In this project, we will extend our previous study on the proton-collecting antenna effect at protein surfaces to the study of surface-mediated proton transport at surfaces with different properties (e.g. hydrophobic surfaces, hydrophilic surfaces) and with different decorations (e.g. amino acid residues with different pKa). We hope to find the relation between proton uptake rate and the surface properties and mechanism of proton diffusion/uptake near different surfaces at atomic level. The main computational technique that will be applied in this project is the Q-HOP molecular dynamics simulation, which has been developed in the Helms' group and has been implemented in the NWChem package.
• Theoretical Modeling of Fluorescence Properties in Biological Systems
  Project Lead: Spiridoula Matsika
  Project Lead Institution: Temple University
  Research Area: Biological Interaction Dynamics
  FY10 Allocation: 450,000 node-hours
  Abstract

  In this project high level ab initio methods will be used to study photophysical properties of nucleic acid bases and their fluorescent analogs. The natural nucleobases have ultrashort excited state lifetimes and very short quantum yields for fluorescence. Small modifications in their structure renders them fluorescent. We seek to understand these effects by calculating potential energy surfaces (PESs) of the excited states of these systems and comparing their features. Conical intersections are of particular interest since they facilitate radiationless decay and fluorescence quenching. Multireference configuration interaction as implemented in the COLUMBUS suite of programs and completely renormalized equation-of-motion coupled cluster techniques as implemented in NWChem will be used to obtain accurate energies on PESs. Due to a large number of single point calculations that need to be performed in order to obtain a reliable characterization of the excited-state PESs for a wide variety of internuclear geometries, the use of highly scalable software is of paramount importance. Both of the computational packages that will be used satisfy this requirement. Monomers and pi-stacked dimers will be considered. A QM/MM approach will also be used to account for solvent and other environmental effects.