March 2011 - Allocations-admins - mailman3.cels.anl.gov

[LCRC Accounts] Project Allocation Request
by accounts＠lcrc.anl.gov 15 Mar '11

15 Mar '11

Hello, A change in allocation has been requested: Requester: jgreeley (Jeffrey Greeley) Project: DFT_corrosion Title: First principles studies of lithiation and corrosion of metal and oxide surfaces Description: For sustainable future energy needs, truly transformational technologies are needed to power both small-scale portable electronic devices, as well as medium-large scale transportation devices. While battery technology has seen improvement in the last 2-3 decades with the advent of the Li-ion battery for portable electronics, significant enhancements in battery energy density are still needed for transportation and other applications. An exciting alternative to this technology, particularly for large systems, is the Li-air battery, which may offer energy storage densities of comparable levels to that of gasoline. However, significant technical challenges relating to battery power and efficiency, battery recharge ability, and electrolyte stability must be addressed before Li-air batteries can truly compete with existing technologies. A particular catalytic challenge associated with Li-air systems concerns the performance of the Li2O2/Li2O cathode. At this e lectrode, oxygen combines with lithium ions and electrons migrating from the anode to produce either lithium peroxide or lithium oxide. While the formation of these oxidic species (termed “discharge” of the battery) is rapid and proceeds at voltages close to the appropriate equilibrium potential, the reverse process (decomposition of the oxidic species, or “charging” of the battery) suffers from significant kinetic inhibitions and requires large overpotentials. Most catalytic studies of Li-air focus, therefore, on finding catalytic electrode materials that facilitate oxide/peroxide decomposition and avoid fouling of the cathode due to excessive solid oxide/peroxide accumulation. In spite of a decade of research in this area, however, the fundamental atomic-scale mechanisms of Li-O2 reactions with catalytic species is poorly understood, and there is an urgent need to perform careful studies on model catalytic systems to elucidate the basic chemistry and physics that governs these systems. We intend to use periodic DFT calculations to study the surface catalytic chemistry of Li air batteries on transition metal surfaces. The periodic DFT calculations, in conjunction with physical-chemical descriptor models, to characterize the reaction thermodynamics and kinetics of Lix-Oy based species with single-component and binary alloys of transition metal (TM) surfaces. Binding energies and reaction barriers for a variety of intermediate and product species will be determined on the TM surfaces; these species will include Li, O, LiO, O2, LiO2, Li2O2, Li2O, Li3O2, Li2O4, Li3O4, Li4O4, and Li2. We will perform such calculations on both the low index, defect-free fcc(111) or hcp(0001) surfaces of Co, Ru, Os, Cu, Ni, Ag, Pd, Rh, Ir, Pt, and Au. Additionally, we will augment such calculations with calculations on stepped fcc(211) surfaces; these calculations will allow us to understand the effects on reaction energies of representative structures (primarily facets and edg es) of metallic nanoparticles. The DFT calculations, performed at the vacuum-solid interface, will be supplemented with simple thermodynamic corrections to estimate reaction free energy diagrams as a function of the electrode potential. The results will be used to propose rate-limiting steps and selective product distributions on the various surfaces, together with screening criteria for improved materials for control of Lix-Oy species production and decomposition. Finally, we will take calculations based on the resulting “best materials” identified in our analysis further; in this further work, we will extend our calculations to identify the thermodynamic properties of Lix-Oy-TM cluster, film, island, and bulk-like nanoparticle growth on the TM surfaces. Such work will be invaluable in identifying key mechanistic phenomena on model catalysts for lithium-air batteries. An important secondary goal of the proposed work will involve studies of oxidation and dissolution of metal alloy surfaces. This work, which is relevant to the long-term durability of both Li air and fuel cell electrodes under harsh electrochemical operating conditions, is an extension of earlier studies on the competition between oxidation and dissolution of platinum surfaces. The first step will be to determine the binding energies of atomic O and OH, as a function of coverage, on a single-crystal kinked metal surfaces. We will focus on metal overlayers (wherein a single layer of an alloying metal is deposited on a bulk metallic substrate) and will consider primarily fcc(763) kinks. Numerous configurations will be considered for each adsorbate coverage, and the lowest-energy configurations at each coverage will be input into an atomistic thermodynamic formalism (essentially a simplified Grand Canonical Monte Carlo technique) to determine the O coverage as a function of e lectrode potential. These results will then be compared to tabulated values of the equilibrium dissolution potential of the relevant overlayer metals as a function of the metal ion concentration in the surrounding solution; with this analysis, it will be possible to estimate the approximate coverages of surface O and OH that are present at the thermodynmically reversible potential on these surfaces, thereby providing important insights into the interplay between surface oxidation and platinum loss by dissolution. By studying how O and OH decorate the step edges at these higher potentials, it will also be possible to obtain semi-quantitative insight into passivation and place exchange mechanisms that may be operative in various potential ranges. These computational studies will be performed in collaboration with Nenad Markovic (ANL/MSD). The computational requirements of the proposed research will be substantial. Test calculations with the Dacapo code have indicated that the p(3x3) unit cells to be used for the Li air studies require approximately 5 hours per total geometric optimization on 24 cores on computer clusters similar to Fusion. We estimate that a total of 10 transition metals, with about 250 total optimizations required per metal, will be considered, giving a total of 300,000 requested core hours. The explicitly corrosion component of the project will also be substantial, likely requiring about 200,000 core hours. We are therefore requesting 500,000 core hours for the proposed project. Dacapo and VASP have been running successfully on Fusion, and GPAW has been recently installed on the system by John Low. Current: undetermined amount Justification: The computational requirements of the proposed research will be substantial. Test calculations with the Dacapo code have indicated that the p(3x3) unit cells to be used for the Li air studies require approximately 5 hours per total geometric optimization on 24 cores on computer clusters similar to Fusion. We estimate that a total of 10 transition metals, with about 250 total optimizations required per metal, will be considered, giving a total of 300,000 requested core hours. The explicitly corrosion component of the project will also be substantial, likely requiring about 200,000 core hours. We are therefore requesting 500,000 core hours for the proposed project. Requested: 400000 A specific reason has been given: We have made good progress on the proposed work, but we have also found that the calculations of Li air battery intermediates on stepped metal surfaces require unit cells approximately twice as large as originally planned, leading to a ~8-fold increase in the required computational time for these calculations. A similar situation has been encountered with the corrosion calculations on kinked alloy metal surfaces, again leading to significant increases in the required compute time. This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Allocation Request
by accounts＠lcrc.anl.gov 15 Mar '11

15 Mar '11

Hello, A change in allocation has been requested: Requester: bliu (Bin Liu) Project: dft_biomass Title: DFT studies of catalytic biomass conversion Description: As part of the IACT (short for the Institute for Atom-efficient Chemical Transformations) supported by Energy Frontier Research Center at Argonne, the DFT calculations will be performed using the Vienna ab initio Simulation Program (VASP), which implements DFT in the periodic Bloch-type plane-wave, self-consistent Kohn-Sham framework. DFT studies conducted on Fusion cluster will focus on obtaining the thermodynamics and kinetics of the intermediate species derived from glycerol decomposition on Pt(111) and furfural hydrogenation on Pd(111) with an ultimate goal of constructing relevant free energy surfaces for their respective reaction network. In both cases, the research will involve a thorough investigation of the most stable adsorption configurations of the intermediate and searches for key transition states (using the Nudged Elastic Band algorithm) to obtain elementary reaction energy barriers. The computational demand of such investigation is substantial, given the relatively large unit cells (either a 2 or 3-layer p(4x4) periodic slab) required, together with the large numbers of proposed reaction intermediates of interest (on the order of hundreds by considering all the possible intermediates). However, it is possible to reduce the computational workload by using insights gained from our preliminary investigations on the stable geometrical motifs of some of these intermediates, together with approximate yet useful correlations that can be used to estimate the binding energies and reaction energy barriers to make the work tractable and practical within the time requested (see below). The classical MD simulations of amorphous alumina and zirconia on Fusion, although also being of great interest to the IACT, will require significantly less time than that requested for the DFT work; we estimate that 5-10% of the total computing time will be needed to run these classical MD simulations with DL-POLY (we have already verified the accuracy and suitability of our potentials with DL-POLY. These jobs will typically run on 4~8 Fusion nodes for large-scale production). Dr. Bin Liu of CNM will be the primary invesitgator of this project with Dr. Jeff Greeley as the co-investigator. To complete the above work, we are requesting 600,000 core hours for FY2011. This is an estimate based on our preliminary work running on Fusion in the past 3 months, during which time approximately 120,000 core hours have been used. We anticipate that our need for Fusion resources will increase somewhat in the coming FY due an even greater level of scientific activity (the project has been ramping up but is now at steady state), as well as due to an oversubscription of other compute clusters (e.g., the Carbon cluster at the CNM). Current: undetermined amount Justification: Recent improvements in the parallelization of VASP combined with the improved hardware capabilities (e.g. the expanded memory) allow for reasonably efficient near-linear scaling for jobs running on up to 80 cores (10 nodes). Typically, a DFT job uses 4~8 nodes at a timescale of 48 hours. Furture usage will require larger number of nodes (8~10 nodes) due to a shift of our focus on transition state calculations. As the computation time for a single job of such calculation can be a week or even more. Large allocation is preferred to guarantee the completion of this investigation. In the mean time, we will efficiently use the requested allocation by carefully selecting the key reactions to produce outstanding science. We hope to continue to use Fusion cluster to satisfy our scientific interest. Requested: 200000 A specific reason has been given: For the project of DFT_biomass, we were originally allocated 480K core hours, 240K hours for the first half of FY2010. At the end of the first quarter, an additional 80K hours were requested and approved. As of today, we have used all 320K hours (for the first half of FY2010). Based on this rate of usage, we would like to request an additional 160K~200K hours for the 2nd half of FY2010. Thanks very much. This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Allocation Request
by accounts＠lcrc.anl.gov 15 Mar '11

15 Mar '11

Hello, A change in allocation has been requested: Requester: ssom (Sibendu Som) Project: cfd_enginemodeling Title: Computational Fluid Dynamics Modeling of Petrodiesel and Biodiesel for Compression Ignition engine applications Description: 1) Finite volume schemes 2) LES and RANS turbulence models used 3) Detailed chemical kinetic models for diesel and biodiesel fuels 4) CONVERGE software is being used - 24 licenses available 5) Scales well up-to 24 processors 6) Detailed spray models used for different spray processes Current: undetermined amount Justification: Requested: 250000 A specific reason has been given: We plan to increase the number of licenses for CONVERGE software. Hence, we are requesting additional allocation This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System

1 0

Re: [allocations-admins] [LCRC Accounts] Project Request: SOFC
by John Valdes 10 Mar '11

10 Mar '11

On Thu, Mar 10, 2011 at 09:55:01AM -0600, Bair, Raymond A. wrote: > > If we just want to launch parallel jobs on the cluster, can we buy the > Server license without buying the Parallel Toolbox license? I'm pretty sure that we'd need at least one license for the Parallel Toolbox, as that toolbox provides all the commands for running tasks in parallel and/or in batch. John

1 0

[LCRC Accounts] Project Allocation Request
by accounts＠lcrc.anl.gov 10 Mar '11

10 Mar '11

Hello, A change in allocation has been requested: Requester: emconsta (Emil M Constantinescu) Project: uncertainty-climate Title: Uncertainty in weather forecasting Description: The aim is to create very accurate uncertainty estimates of the weather and climate forecasts to be used for optimal management of energy systems such as wind- and solar-power generation, building management and power grid planning. The purpose of using Fusion is to execute several scenarios and obtain forecast and uncertainty information about ambient variables by running high resolution weather/climate simulation codes (such as WRF). This information is used in a stochastic optimization control loop that provides an optimal operational point for energy systems. For the practical part of the project, we will develop and enhance an uncertainty quantification framework (C or FORTRAN, MPI, XML, Python) to run WRF in a realistic fashion and illustrate theoretical findings and provide realistic results. We consider running several WRF instances (each on one node) at the same time. Partial results are collected on a separate node and processed before they are sent back to each specific instance. The instances will continue to run and produce new results. In terms of parallelism, this application framework is embarrassingly parallel since each instance run independently. In FY11, we need to run data assimilation experiments with high resolution and obtain more samples in order to achieve a better level of accuracy. The improved weather forecast accuracy simulations allow us to find a more robust operational point for the operation of energy systems (wind power for electricity generation and building heating and cooling). Current: undetermined amount Justification: Note: I used the FY10 form to make updates for renewal and saved them. After that I realized that i need to use a different form... so now renewal for FY11 and FY10 are the same. Sorry about that. Requested: 200000 A specific reason has been given: In this project we were planning to run the weather model and perform model calibration on it. However, we hit a few theoretical problems with the model calibration strategy, and therefore the software implementation was not ready to run. We anticipate to solve all our problems and be ready to use Fusion by May 2011, and I am therefore requesting a renewal with a similar allocation quota for the next period. This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System

1 0

Re: [allocations-admins] [LCRC Accounts] Project Request: SOFC
by Bair, Raymond A. 10 Mar '11

10 Mar '11

On 2/24/11 1:42 PM, "John Valdes" <valdes(a)mcs.anl.gov> wrote: > We would need a (floating) license for the Parallel Computing Toolbox. > Each license for the toolbox provides for 8 worker processes on the > local host, so this could be used, for example, by running an > interactive job on 1 node and then running Matlab on that node. In > addition to the Parallel Computing Toolbox, one can get the Matlab > Distributed Computing Server, which allows one to run Matlab workers > on a cluster either interactively or in batch. The way it was > explained to me is that you could startup Matlab on a login node, for > example, to setup the problem which you then submit to the job queues > from within Matlab. The submitted job will then use worker licenses > from the Distributed Computing Server while it runs on the compute > nodes. While it runs, you can quit Matlab on the login node, thus > freeing up the license used by that Matlab instance. If we just want to launch parallel jobs on the cluster, can we buy the Server license without buying the Parallel Toolbox license? ----------------------------------------- Ray Bair Computing, Environment, and Life Sciences Argonne National Laboratory and the University of Chicago TCS Building 240, Room 4126 9700 South Cass Avenue Argonne, IL 60439 email: rbair(at)anl.gov Phone: (630)252-5751

1 0

[LCRC Accounts] Project Allocation Request
by accounts＠lcrc.anl.gov 08 Mar '11

08 Mar '11

Hello, A change in allocation has been requested: Requester: fonseca (James Fonseca) Project: dynamics Title: Dynamics of biophysical systems. Description: This project focuses on a range of computational studies of electrolyte and electrolyte-protein solutions. We investigate the fundamental behavior of *confined* electrolytes outside of equilibrium, which is of paramount importance to the understanding of the functioning of many biological and physical systems. The primary mode of investigation is Monte-Carlo (equilibrium) and molecular dynamics (out-of-equilibrium) all-atom simulations of protein-electrolyte mixtures (in the primitive or restricted primitive model of the electrolyte). The system representative of our investigation is a membrane channel protein operating under a concentration and voltage gradient. The primary function of channel proteins is to pass ionic current in and out of the cell *selectively* (i.e., favoring some types of ions over others). This selectivity mechanism is crucial to the functioning of the cell and the overall organism, but it is still relatively poorly understood. In particular, since a channel can be sensitive to trace amounts of the selected ion, the permeation and selectivity mechanisms are subtly related to the atomic structure of the protein. At the same time, since the ionic current is largely independent of the thermal fluctuations of the protein (in the open configuration), the channel function is rather robust. Thus, there is a need to isolate the relevant features of the channel geometry and, most importantly, permanent charge distribution, that are responsible for this sensitive yet robust selective beha vior. This problem is one of the most important open biological questions, but the results have implications far beyond the physiological setting and include, for example, the design of artificial channels that can be used in fast DNA recognition, bioremediation of toxic waste, to name but a few. Due to the non-equilibrium nature of the system, Monte-Carlo (MC) simulations (whose importance cannot be overemphasized) must be supplemented by lengthy molecular dynamics (MD) studies. These MD simulations must be as long as possible to approach physiological timescales (~msec), which justifies the use of high-performance resources, such as Fusion. At the same time, we require a fairly quick turnaround time for the processing of the simulation data, in order to adjust the simulation setup, if necessary. This explains a fairly limited resource request. We anticipate that once we gain more experience with the new types of simulations (MC and MD), more extensive allocations may be requested. Current: undetermined amount Justification: Requested: 240000 A specific reason has been given: The simulation results we have obtained thus far from Fusion have been very promising and we would like to continue our work. Some of the initial simulation processor hour requirements were underestimated because under newly-investigated parameters, our model is operating in a high charge density regime and this requires extended computation time to produce acceptable stochastic results. A typical run is about 50 hours on 24 processors, but some may require double that time. We would greatly appreciate the extended allocation. Thank you, Jim Fonseca This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Request: MD_of_pnipam_brush
by accounts＠lcrc.anl.gov 07 Mar '11

07 Mar '11

Hello, A new project on the LCRC cluster has been requested. Please forward the information on to the LCRC Allocation sub-committee. Applicant's name: Sanket Deshmukh Applicant's institution: ANL Applicant's division: CNM Project Name: MD_of_pnipam_brush Project title: Structural and dynamic properties of thermo-sensitive polymer brushes: Molecular Dynamics study. Associated funding: Other Systems: Carbon (CNM) Science: Grafting of polymer chains on the surface is challenging but has great importance in many areas of science and engineering e.g. colloidal stabilization, biotechnology, rheology etc.1 Such a grafted structure is called as “polymer brush” and is of both practical and theoretical interest. Theoretical interest in such systems arises due to the fact that single/free chains have different configurations to the chains grafted on the surface or chains near the surface. For most of these systems, in most situations, a solvent is present and grafted polymeric chains are not at melt density. The conformation of such polymers in a solvent can dramatically change with graft density; at low densities, these chains assume a “mushroom” conformation with the coil dimension similar to that of free chains. However, brushes with high grafting density and height (polymer chain length) have characteristic structure and properties quite different and unpredictable from those of low graft density brushes. Polymer brushes can be divided into two concentration densities, namely “semi-dilute” and “concentrated”. Poly(N-isopropylacrylamide) PNIPAM is a thermo-sensitive polymer that has a lower critical solution temperature (LCST) around 305K in aqueous solution. As the LCST of PNIPAM is close to human body temperature, it is a serious contender for a controlled drug delivery system. Coil-to-globule transition occurs in PNIPAM when temperature is raised above the LCST. It is believed that this occurs due to the change in hydrophilic interaction to the hydrophobic interaction. Concentration and molecular weight of PNIPAM affects the LCST by less than 1K (304–306K). Lower molecular weight samples appear to have slightly higher LCSTs. Presence of salts ions in the solution lowers the LCST significantly.3 In the case of PNIPAM brushes, LCST depends upon the grafting density and molecular weight. In this work, we propose to use molecular dynamics (MD) simulation techniques to understand how the grafting density and local forces acting on the polymer bonds affect PNIPAM brushes at fully atomistic level. Behavior of a single chain in an aqueous solution, across the LCST, in the brush will be studied using MD simulations. This information can be useful for the development of composite system, where functional group or nano-particles are attached to the free ends of chain. It is also proposed to study the structural properties and change in the LCST of PNIPAM, if any, when two polymer chains, namely PNIPAM and poly(ethylene oxide) (PEO), with different LCSTs are grafted together in presence of water. Project description: In this work we propose to use the Consistent Valence force field (CVFF) to understand the structural changes in the polymer brush at global as well as local level using an atomistic level model. The preliminary studies we carried out with CVFF on a single chain made up of 30 monomer units of PNIPAM in presence of pure water predicted the LCST close to 305K, which is in good agreement with experimentally observed value. To carry out MD simulations we will be using LAMMPS MD simulation package. Fully atomistic polymer chains with ~25 monomer units will be generated. Each chain will be end-grafted to form a 5 x 5 grid on flat surface, which will define the x-y plane of the system. Grafting density will be varied from 0.05 to 0.5 to study the structural changes with respect to grafting density. Depending upon the number of polymer chains TIP4P water molecules (~20,000 to ~100,000) will be added to simulation cell. To study the structural changes across the LCST of PNIPAM, simulation system will then be equilibrated up to 40-50ns using LAMMPS at 278, 310 and 360K. To study the change in the LCST of PNIPAM in presence of another thermo-sensitive polymer, if any, we propose to use PNIPAM and PEO grafted and equilibrated in similar manner as mentioned above. The computational requirements for the proposed study will be substantial. To study the phase transition across the LCST in the case of single PNIPAM chain, simulations should be carried out for ~50ns. From preliminary results of the simulations we carried out on the CNM’s cluster Carbon, we observed a clear phase transition across the LCST, for polymer chains with 30 monomer units (572 atoms) in presence of 9,000 water molecules. This system with ~30,000 atoms required ~145 hrs of simulation time for 10ns on 96 cores. In the proposed work we plan to study, PNIPAM ~25 mer (475 atoms) brush made up of ~20 polymer chains (~9000 atoms) in presence of ~30,000 water molecules. Simulations will be carried out at 278, 310 and 360K for ~50ns (total simulation time = 150ns). We also propose to study PNIPAM ~25 mer (475 atoms) and PEO ~25 mer (150 atoms) polymer chains. In the initial stages of this project we plan to use 5 and 10 chains of both PNIPAM (2375 and 4750 atoms) and PEO (1500 and 3000 atoms) in presence of 30,000 water molecules. Total system will be consisted of ~100,000 atoms and simulations will be carried out at 278, 310 and 360K for ~50ns (total simulation time = 300ns). The estimated time to carry out the mentioned calculations in this particular work would be ~500,000 core hours. Project URL: Requested allocation: 600000 Justification: To study the phase transition across the LCST in the case of single PNIPAM chain, simulations should be carried out for ~50ns. From preliminary results of the simulations we carried out on the CNM’s cluster Carbon, we observed a clear phase transition across the LCST, for polymer chains with 30 monomer units (572 atoms) in presence of 9,000 water molecules. This system with ~30,000 atoms required ~145 hrs of simulation time for 10ns on 96 cores. The requester has used 0 hours of their initial startup project. In addition to approving an initial amount, please specify a Category and Subcategory for this project. For a list of the current selection of approved categories, please see: https://wiki.lcrc.anl.gov/wiki/Processes/Categories Once the Allocation committee has approved the project, please go to the Project Management page to create it: https://accounts.lcrc.anl.gov/projects.php Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Allocation Request
by accounts＠lcrc.anl.gov 07 Mar '11

07 Mar '11

Hello, A change in allocation has been requested: Requester: walkosz (Weronika Walkosz) Project: Metastable_Materials Title: First-principles studies of growth mechanisms for metastable materials Description: We will perform first-principles calculations for the adsorption and diffusion of NH3 on GaN, InN, and In surfaces through reaction intermediates such as NH2, NH, N, and H. The simulations will be performed with Vienna Ab-initio Simulation Package (VASP), using plane wave basis sets and PAW potentials. We will first model surfaces of GaN, InN, and In metal in the presence of NH3 to determine the most stable surface sites/configurations for the absorbed NH3 and its decomposed species. Next, diffusion and dissociation pathways of NH3 on various surfaces will be studied by calculating energy barriers using the Nudged Elastic Band Method. The scale of the proposed calculations exceeds small computer cluster capacity. Therefore, we are requesting the LCRC resources to perform a computational study of NH3 on GaN, InN, and In metal. The speed and reliability of the Fusion cluster will allow us to finish the work in a timely manner. The calculations of adsorption a nd decomposition of NH3 on different surfaces will take approximately 48 hours each when ran on 64 processors. At least 50 different structures will be calculated. The calculations of the reaction energy barriers (NEB), on the other hand, will require about 96 hours each when ran on 80 processors as determined by on our previous experience with these types of calculations on semiconductors. At least 42 reaction barriers need to be calculated to validate a diffusion model. Therefore, a total allocation of ~ 475,000 processor hours will be needed to perform the proposed calculations. Current: undetermined amount Justification: Requested: 400000 A specific reason has been given: I have started using new state-of-the-art DFT methods, which are taking considerably more time. Also, I am currently running vibrational frequency calculations, which are computationally very demanding, but are of crucial importance to my project. I did not anticipate these calculations when requesting the allocation units a few months ago. This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Allocation Request
by accounts＠lcrc.anl.gov 07 Mar '11

07 Mar '11

Hello, A change in allocation has been requested: Requester: walkosz (Weronika Walkosz) Project: Metastable_Materials Title: First-principles studies of growth mechanisms for metastable materials Description: We will perform first-principles calculations for the adsorption and diffusion of NH3 on GaN, InN, and In surfaces through reaction intermediates such as NH2, NH, N, and H. The simulations will be performed with Vienna Ab-initio Simulation Package (VASP), using plane wave basis sets and PAW potentials. We will first model surfaces of GaN, InN, and In metal in the presence of NH3 to determine the most stable surface sites/configurations for the absorbed NH3 and its decomposed species. Next, diffusion and dissociation pathways of NH3 on various surfaces will be studied by calculating energy barriers using the Nudged Elastic Band Method. The scale of the proposed calculations exceeds small computer cluster capacity. Therefore, we are requesting the LCRC resources to perform a computational study of NH3 on GaN, InN, and In metal. The speed and reliability of the Fusion cluster will allow us to finish the work in a timely manner. The calculations of adsorption a nd decomposition of NH3 on different surfaces will take approximately 48 hours each when ran on 64 processors. At least 50 different structures will be calculated. The calculations of the reaction energy barriers (NEB), on the other hand, will require about 96 hours each when ran on 80 processors as determined by on our previous experience with these types of calculations on semiconductors. At least 42 reaction barriers need to be calculated to validate a diffusion model. Therefore, a total allocation of ~ 475,000 processor hours will be needed to perform the proposed calculations. Current: undetermined amount Justification: Requested: 400000 A specific reason has been given: I have started using new state-of-the-art DFT methods, which are taking considerably more time. Also, I am currently running vibrational frequency calculations, which are computationally very demanding, but are of crucial importance to my project. I did not anticipate these calculations when requesting the allocation units a few months ago. This needs to be approved and the final allocation amount decided upon. Thank You, The LCRC Accounts System

1 0