September 2016 - Allocations-admins - mailman3.cels.anl.gov

[LCRC Accounts] Project Request: LiSx
by accounts＠lcrc.anl.gov 07 Sep '16

07 Sep '16

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: Badri Narayanan Applicant's institution: ANL Applicant's division: NST Project Name: LiSx Project title: Solvation dynamics of lithium polysulfide species in sparingly solvating electrolytes Associated funding: JCESR Other Systems: Carbon (CNM): 500,000 core hours ALCF: 50 million core hours Science: Lithium-sulfur (Li-S) batteries hold tremendous potential for next-generation energy storage devices owing to their high theoretical capacity (1675 mAhg-1), nearly 10 times that of a conventional Li-ion battery (140 mAhg-1), while using inexpensive and abundantly available materials. The Li-S battery system, however, is very challenging, mainly due to the complex redox reactions taking place at the S cathode; these cause reduction of S8 into various soluble polysulfide species that diffuse away from the cathode, and consequently, decreases Coulombic efficiency and capacity retention of the cell, as well as result in undesirable side reactions at the Li anode, termed polysulfide shuttle. Recently, the use of sparingly solvating electrolytes (e.g., acetonitrile with lithium bis(trifluoromethane sulfonyl) imide (LiTFSI)) have shown great promise in impairing solubility of Li-polysulfide species. These studies utilize high concentrations of salt with com mon ions (Li+), which shift the equilibrium between solid and solvated Li+ towards solid and decrease the solubility. This, however, comes at the price of increased viscosity and poor diffusion kinetics. To mitigate this effect, secondary solvents called diluents (e.g., 1,1,2,2-tetrafluoroethyl 2,2,3,3-tetrafluoropropyl ether (TTE)) thought to be inert to the solvate structure are added to reduce the viscosity, and enhance the diffusion kinetics. Although, experimental works by our JCESR collaborators have shown marked success in the use of sparingly solvating primary solvents in combination with diluents for electrolytes, a fundamental understanding of the solvation dynamics is still lacking. A molecular level knowledge of the solvation structure, and dynamics around LiSx species is crucial for designing new electrolytes with exceptional electrochemical performance. Such an understanding is invaluable particularly under conditions necessary for high energy density such as a moderately high S loading (4 mg/cm2), low E/S ratio (< 3 mL/g) and room temperature. In this project, we will employ ab-initio molecular dynamics simulations, solvation free-energy density functional theory calculations, and electronic structure calculations to investigate the solvation structure around various LiSx species in two ACN-LiTFSI (primary solvent), and TTE/HFE diluents. These studies will identify the effect of diluent on electrochemic al performance, electrolyte stability, and physical properties (e.g., viscosity) of the electrolyte system. Furthermore, these will unravel the molecular/electronic origin of the improved performance of electrolyte; these, in turn, will enable computationally-aided discovery of new and better diluent structures. Project description: The first set of simulations will employ ab-initio molecular dynamics to identify the solvation structure around various LiSx species in numerous electrolyte systems. Three LiSx species: 1) S8, 2) Li2S8, and 3) Li2S6 will be investigated in electrolyte systems comprising of (ACN)2-LiTFSI (primary), and two diluents (TTE, HFE). The AIMD simulations will be performed using the plane wave density functional theory package VASP; the k-point will be sampled only at the Γ-point, and modest energy cut-off (300 eV) will be employed. The Perdew-Burke-Ernzerhof pseudopotentials will be employed for exchange correlations within the generalized gradient approximation. Three different ratios of primary/diluent will be studied for each LiSx species and diluent. For each simulation (system size ~200 atoms), AIMD runs for 10 ps will be performed at 300 K and 400 K (operation temperatures); each of these runs are expected to take 100 hours on 96 cores. So we need (3 LiSx species) x (3 ratios) x (2 diluents) x (2 temperatures) x (96 cores ) x (100 hours) = 345600 core hours. >From the AIMD simulations, we will choose 12 solvation structures on which we will perform more accurate electronic structure, and charge transfer calculations. Each of these will take 144 hours on 96 cores. So we need (12 structures) x (96 cores) x (144 hours) = 165888 core hours. In total, require 500,000 core hours. Industry partnership: Project URL: Requested allocation: 499999 Q1: 125000 Q2: 125000 Q3: 125000 Q4: 124999 Justification: Storage requirements: The requester has used undetermined amount 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] Yearly Allocation Request for ElectrolyteRedoxFlow
by accounts＠lcrc.anl.gov 07 Sep '16

07 Sep '16

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Lei Cheng Project Name: ElectrolyteRedoxFlow Division: MSD Project title: discoveries of electrolyte for redox flow batteries Associated funding: JCESR Other Systems: NERSC Science: To meet global energy demand, renewable energy sources such as solar and wind energy are being developed and utilized. The increased use of wind and solar energy will challenge the grid’s ability to provide a stable electrical supply around the clock. Grid energy storage can be used to shift load from peak to off-peak hours and redox flow batteries are of particular interest to perform this load leveling. Redox flow batteries differ from classical batteries in that energy is stored in the liquid electrolyte in separate liquid reservoirs instead of in the electrode. This bypasses common electrode degradation issues observed in lithium ion batteries and makes the battery systems more durable and scalable. However, all flow batteries based on aqueous electrochemical couples are limited by the electrochemical properties of water which is only stable within a small potential window (typically 1.2-1.6 V) outside of which water electrolysis occurs. This means that redox- flow battery suffers from lower energy density. In order to increase the energy density, the current or potential must be increased. Employing non-aqueous electrolytes offers a wider window of electrochemical stability, enabling flow batteries to operate at higher cell potentials (>2 V) and eliminating this low energy density limitation. Among various fundamental science and engineering challenges in the design of more efficient redox flow batteries, discoveries of appropriate redox couples that operate at higher cell voltage will lead to greater system energy densities and higher energy efficiencies. It is also one of the important missions of the energy storage hub Joint Center for Energy Storage Research (JCESR). Project description: In this renewal work, we will focus on stability studies of redox-active molecules for aqueous redox flow battery application. A short list of redox-active molecules for NRF application down-selected through high-throughput screening scheme were synthesized and tested experimentally (Moore and Brushett). The validation work shows that most of the compounds have irreversible redox activity, suggesting their instability in charged states. Similar phenomenon has been observed for molecules screened for aqueous redox flow (ARF) application. For example, catechol molecule can potentially be used as active catholyte materials in ARF battery. The electron-withdrawing group nitro is added to increase the oxidation potential of the catechol base molecule and it was predicted to increase the oxidation potential by 0.2 V. However, the experimental cyclic voltammetry measurements show that the oxidation of nitro-substituted catechol is not reversible (no reduction o bserved) though the oxidation potentials agree well with calculation (courtesy of Zhang and Su, ANL). We are investigating addition reactions (1,2-addition and 1,4-addition) by H2O molecule as possible degradation reaction mechanism on a series of hydroquinone compounds with available experimental reversibility data. This study will shed light on possible stability predictors such as substitution site and electron-activating strength of substituent, as well as designing more stable molecules for this category of compounds. Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 800000 Q1: 200000 Q2: 200000 Q3: 200000 Q4: 200000 Justification: All calculations will be performed using Gaussian. This package is considered very efficient and reliable for quantum chemistry calculations. Gaussian is not very scalable but it provides really robust geometry optimizer and the G3/G4 methods are built-in. The size of molecules we propose to study is very typical for the computational software (Gaussian) so they scale reasonably well for most DFT calculation. The higher level calculations will consume more memory as well as computational time. Storage requirements: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Request: PbTe_thermoelectric
by accounts＠lcrc.anl.gov 07 Sep '16

07 Sep '16

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: Yi Xia Applicant's institution: ANL Applicant's division: NST Project Name: PbTe_thermoelectric Project title: Atomistic simulations of electronic and thermal transport in nanostructured PbTe Associated funding: Midwest Integrated Center for Computational Materials (MICCoM) at DOE's Argonne National Laboratory Other Systems: Carbon cluster at Center for Nanoscale Materials (CNM) at Argonne: 590,000 core hours Science: Thermoelectrics are materials that can reversibly convert heart to electricity, which can be used for both power conversion from waste heat and for coolers by consuming electric energy [1-4]. In spite of their vast potential applications, their practical use depends on their efficiency, which is characterized by a dimensionless figure of merit called ZT. This merit is proportional to the product of electrical conductivity and square of Seebeck coefficient (defined as power factor) divided by total thermal conductivity. The ZT of commercially available thermoelectric devices is about 1, while ZT above 2 is required for a wider application [4]. The challenge in increasing the ZT is the need to optimize a variety of conflicting properties. To maximize the ZT, a large absolute value of the Seebeck coefficient, high electrical conductivity, and low thermal conductivity are required. However, an increase in electrical conductivity generally leads to a decreased Seebeck coefficient and an increased electronic contribution to thermal conductivity [5]. Therefore, electronic degree of freedoms should be optimized through engineering electronic bands, dopants and carrier concentrations [4, 6]. On the other hand, ZT also can be directly enhanced through reducing lattice thermal conductivity, since its dependence on electronic parameters is generally weak [7-9]. PbTe is an excellent thermoelectric material because of both high power factor and low lattice thermal conductivity [10]. The former is due to a combination of high mobility and valley degency, and the latter arises from strong lattice anharmonicity [11, 12]. Recent studies by Kanatzidis and coworkers showed significant enhancement in the ZT of doped PbxM1-xTe (M=Na, Sr, Mg) by phonon scattering at different length scale [13, 14]. The scattering is caused by dopants at the atomic scale, by precipitates at the nanoscale and grain boundaries at the mesoscale. These nanostructured composite materials have ZTs above 2 around 800 K. Experimental studies revealed improvements in ZT of lead based thermoelectrics using different dopants, alloying elements and structural features of various length scales. However, fundamental understanding about the contribution of these diverse methodologies to the enhancement of ZT is far from satisfactory. In this project, focusing on lead-telluride based systems, we will perform first-principle density functional theory (DFT), Boltzmann transport equation (BTE) and classical molecular dynamics (MD) calculations to understand the effects of dopants and secondary phases on thermoelectric properties. In the first part of this project, we will investigate the effects of alloying elements (Mg, Ca, Sr Ba and Na) on the electronic band structures and transport coefficients of PbTe. In the second part, we will investigate the effects of strain (lattice constant change) and shallow dopants (mass disorder) on lattice thermal conductivity by iteratively solving the phonon BTE after extracting both harmonic and anharmonic interatomic force constants (IFCs) [15]. In the third part, the effect of structural features on the lattice thermal conductivity will be analyzed using classical MD and Green-Kubo method [16]. We will introduce nanoparticles of (Sr, Mg)Te as secondary phases in PbTe matrix to evaluate interface effects at the nanoscale on lattice thermal conductivity. In order to perform accurate MD simulations, we will first develop empirical force-fields for PbTe, SrTe and MgTe that can represent DFT based lattice, elastic constants and phonon spectra. [1] L. E. Bell, “Cooling, heating, generating power, and recovering waste heat with thermoelectric systems”, Science 321, 1457 (2008) [2] F. J. DiSalvo, “Thermoelectric cooling and power generation”, Science 285, 703 (1999) [3] H. Alam and S. Ramakrishna, “A review on the enhancement of figure of merit from bulk to nano-thermoelectric materials”, Nano Energy 2, 190 (2013) [4] “Thermoelectrics Handbook, Macro to Nano”, edited by D. M. Rowe (CRC Press-Taylor and Francis Group, Boca Raton, 2006) [5] G. J. Snyder and E. S. Toberer, “Complex thermoelectric materials”, Nat. Mater. 7, 105 (2008) [6] G. D. Mahan, “Figure of merit for thermoelectrics”, J. Appl. Phys. 65, 1578 (1989) [7] G. Mahan, B. Sales, and J. Sharp, “Thermoelectric materials: New approaches to an old problem”, Phys. Today 50, 42 (1997) [8] G. Joshi, H. Lee, Y. Lan, X. Wang, G. Zhu, D. Wang, R. W. Gould, D. C. Cuff, M. Y. Tang, M. S. Dresselhaus, G. Chen, and Z. Ren, “Enhanced thermoelectric figure-of-merit in nanostructured p-type silicon germanium bulk alloys”, Nano Lett. 8, 4670 (2008) [9] S. Ohta, T. Nomura, H. Ohta, and K. Koumoto, “High-temperature carrier transport and thermoelectric properties of heavily La-or Nb-doped SrTiO3 single crystals”, J. Appl. Phys. 97, 034106 (2005) [10] D. Zhao and G. Tan, “A review of thermoelectric cooling: Materials, modeling and applications”, Appl. Therm. Eng. 66, 15 (2014) [11] Y. Z. Pei, X. Y. Shi, A. Lalonde, H. Wang, L. D. Chen, and G. J. Snyder, “Convergence of electronic bands for high performance bulk thermoelectrics”, Nature 66, 473 (2011) [12] O. Delaire, J. Ma, K. Marty, A. F. May, M. A. McGuire, M-H. Du, D. J. Singh, A. Podlesnyak, G. Ehlers, M. D. Lumsden, and B. C. Sales, “Giant anharmonic phonon scattering in PbTe”, Nat. Mater. 10, 614 (2011) [13] K. Biswas, J. He, I. D. Blum, C.-I Wu, T. P. Hogan, D. N. Seidman, V. P. Dravid, and M. G. Kanatzidis, “High-performance bulk thermoelectrics with all-scale hierarchical architectures”, Nature 489, 414 (2012) [14] M. Ohta, K. Biswas, S.-H. Lo, J. He, D. Y. Chung, V. P. Dravid , and M. G. Kanatzidis, “Enhancement of thermoelectric figure of merit by the insertion of MgTe nanostructures in p-type PbTe doped with Na2Te”, Adv. Energy Mater. 2, 1117 (2012) [15] F. Zhou, W. Nielson, Y. Xia, and V. Ozolins, “Lattice anharmonicity and thermal conductivity from compressive sensing of first-principles calculations”, Phys. Rev. Lett. 113, 185501 (2014) [16] M. S. Green, “Markoff random processes and the statistical mechanics of time‐dependent phenomena. II. Irreversible processes in fluids”, J. Chem. Phys. 22, 398 (1954); R. Kubo, “Statistical-mechanical theory of Irreversible processes. I. General theory and simple applications to magnetic and conduction problems”, J. Phys. Soc. Japan 12, 570 (1957) Project description: The calculations will involve the items below. (1) We will substitute Sr, Mg, Ca, Ba and Na in place of Pb with 2 different concentrations (1.6% and 3.2%) in a 2x2x2 PbTe supercell (having 64 atoms) and perform DFT calculations. For each 3.2% doped sample, we will try 3 different atomic configurations. The structures will be fully relaxed and band energy eigenvalues will be calculated. (2) BTE calculations for electron transport will be performed using the Wannier90 interpolated electronic band structures within relaxation time approximation for the doped systems in (1). (3) Interatomic force constants will be extracted using compressive sensing lattice dynamics (CSLD) method for 5 different lattice parameters (-1.0%, -0.5%, 0.0%, 0.5% and 1% deviation from the value at room temperature). For each lattice parameter, calculations of twenty 6x6x6 PbTe supercells (432 atoms) with random atomic displacements will be performed. (4) BTE calculations for phonon transport will be performed using ShengBTE for each lattice parameter indicated above from 200 to 800 K. In addition, mass disorder induced phonon scattering will be included for above listed 5 dopants and 2 concentrations. (5) The lattice parameters, bond distances, elastic constants and phonon dispersions will be calculated for PbTe, SrTe and MgTe from DFT simulations. (6) Morse + Coulomb type interatomic potentials will be parameterized for PbTe, SrTe and MgTe to reproduce DFT based training sets. The parameterization will be carried out using a multi-objective genetic algorithm. Coupling parameters for the different potentials will be obtained from Lorentz–Berthelot mixing rules. (7) Successful potentials will be used to test the convergence of thermal conductivity in pure bulk systems with respect to dimensions of the periodically repeated simulation box size. From previous experience we estimate a size between 20,000-30,000 atoms will be reasonable to capture long phonon wavelengths that are important for thermal transport. Here, Green-Kubo approach will be used to estimate thermal conductivity in 3 structure sizes to make sure the obtained values are well converged with respect to simulation box size. Due to statistical nature of Green-Kubo method we will repeat the calculations 6 times starting from different initial conditions (initial atom velocities). The simulation time step will be 1 fs. Heat current will be evaluated for 3 million steps (corresponding to 3 ns) before calculation of thermal conductivity. This procedure will be repeated for all tellurides. (8) We will investigate the effect of SrTe and MgTe nanoparticles on the thermal conductivity of PbTe matrix with 2 different particle sizes (diameters) namely 2 and 4 nm. The separation between the nanoparticles will be set by increasing the box size, effectively increasing the concentration PbTe in the system. 3 different concentrations (1%, 2% and 4%) of (Sr, Mg)Te will be used in accordance with experimental studies in this chemistry. The largest systems in these configurations will include ~100,000 atoms. Again, each of MD calculation will start from 6 different initial conditions (different distribution of initial velocities) to obtain a reasonable ensemble average for the thermal transport properties. 1 fs time step will be used and the heat current will be collected for 3 million steps. The DFT calculations will be performed using the plane wave code VASP, with GGA-PBEsol exchange correlation functional, projector augmented wave (PAW) treatment of core electrons, and full relativistic spin-orbit coupling. The VASP code is tested in Blues with a 350 electron system and 65%-80% scaling efficiency is obtained 64 to 128 cores compared to 16 cores. In Fusion where the code has been tested for systems having up to 7000 electrons, the code showed 70%-80% scaling efficiency for 64 to 128 cores compared to 8 cores. Given this performance, the VASP code has the efficiency to simulate large number configurations in reasonable amount of time. Electron BTE will be solved using BoltzWann implemented in Wannier90 package. The electronic band structure will be accurately interpolated using constructed localized Wannier functions. Wannier90 is a serial code and consumes negligible cpu time. IFCs will be extracted using CSLD code, which also consumes negligible cpu time. Phonon BTE will be solved using ShengBTE package. For the optimization process in force field fitting, we will use a multi-objective genetic algorithm toolbox which is developed at University of Illinois Urbana-Champaign. The code has a 100% scaling efficiency in single node (8 cores on Fusion and 16 cores on Blues). Molecular dynamics calculations will be carried out with LAMMPS code. This code is tested on Fusion for a ~13000 atom system using a Buckingham + point charge force field. A scaling efficiency of 87% is obtained for 64 cores compared to 8 cores. Preliminary testing shows that an MD simulation of 30,000 atoms and 1 million time steps require 30 hours on 64 cores. Therefore, we estimate that thermal conductivity calculations require 2000 core-hours on average. The time estimates for the various parts above, based on benchmark results with VASP, Wannier90, BoltzWann, ShengBTE and LAMMPS running on LCRC and NERSC, are listed below. (1) 5 dopants x 4 calculations x 2000 core hours each = 40,000 (2) 5 dopants x 4 calculations x 100 core hours each = 2,000 (3) 5 lattice parameters x 20 calculations x 1200 core hours each = 120,000 (4) 5 dopants x 2 concentrations x 5 lattice parameters x 8 temperatures x 40 core hours each = 16,000 (5) 3 compounds x 2000 core hours each = 6,000 (6) 3 compounds x 20,000 core hours = 60,000 (7) 3 compounds x 3 sizes x 6 initial conditions x 2000 core hours each = 108,000 (8) 2 nanocomposites x 2 nanoparticle sizes x 3 concentrations x 6 initial conditions x 5000 core hours each = 360,000 The total request is 712,000 core-hours. Dr. Yi Xia and Dr. Maria Chan will be responsible for carrying out the simulations. Industry partnership: Project URL: Requested allocation: 712000 Q1: 200000 Q2: 200000 Q3: 200000 Q4: 112000 Justification: Since most core hours will be used for large-system (more than 400 atoms) calculations performed using VASP and LAMMPS in our project, we provide extra scaling data and CPU performance analysis in addition to those listed above. VASP is tested on XSEDE (Extreme Science and Engineering Discovery Environment) HPC system (300 nodes, 8 cores per node) with number of electrons from ~1000 to ~6000 (~100 to ~600 atoms) using PBE exchange-correlation functional and kinetic cutoff of 350 eV. The scaling efficiency of 80% to150% is obtained for 128 to 1024 cores compared to 16 cores. LAMMPS is tested on XSEDE HPC system (724 nodes, 16 cores per node) with number of atoms from ~60000 to ~260000 using Buckingham+Coulomb force fields. The scaling efficiency of 68% to 115% is obtained for 64 to 512 cores compared to 16 cores. Given this performance, both the VASP and LAMMPS codes have the efficiency to simulate large number of configurations in reasonable amount o f time. Storage requirements: The requester has used undetermined amount 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] Yearly Allocation Request for STOandPlus
by accounts＠lcrc.anl.gov 07 Sep '16

07 Sep '16

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Alejandro Lopez Bezanilla Project Name: STOandPlus Division: MSD Project title: Ab initio calculations for transition metal oxides Associated funding: DOE Other Systems: Science: The sensitive dependence of correlated electron properties on electron concentration and crystal structure has motivated the exploration of new systems that provide access to different regimes of the structure/concentration phase space. Oxides involving components with correlated electron properties are of particular interest because they present the possibility of controlled synthesis of correlated materials with specifically designed properties. In the future will pursue new investigations on both pristine and multilayered lattices composed of BaSnO3. Hybrid functionals will be used to accurately describe the band structure of this insulating material. The possibility of reducing the gap by introducing O vacancies will be explored. The importance of cubic vs other crystalline arrangements will be determined with phonon calculations. The possibility to introduce correlated atoms will be also considered. Project description: First-principles code VASP will be used for electronic structure calculations. Phonopy packaged will help in preparing the input files for phonon spectrum calculations.dope Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 495000 Q1: 125000 Q2: 125000 Q3: 120000 Q4: 125000 Justification: Storage requirements: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for Drude
by accounts＠lcrc.anl.gov 07 Sep '16

07 Sep '16

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Benoit Roux Project Name: Drude Division: CNM Project title: Development and validation of new polarizable force field for molecular dynamics Associated funding: LDRD Project Number: 2014-161-N0, "Bridging the Electronic and Atomistic Scales: Force Field Development for Reactive Interfaces from First Principles" (Principal Investigators: Subramanian Sankaranarayanan, Maria Chan, Stephen Gray, Michael Davis, and Be Other Systems: BG/P, Anton, Midway (from RCC at UC) Beagle (from CI at UC) Science: The functional form and parameters of a new force field accounting for induced polarization in molecular dynamics simulations must be developed, optimized, and validated for various biomolecular systems. We are interested in ions binding to proteins and ion channels. Project description: Detailed understanding of the fundamental physical forces driving the structure and dynamics of proteins and membranes is essential for the engineering of these biologically key molecules for commercial and medical purposes. Towards such understanding, theoretical studies of proteins and lipids based on empirical force fields (FF) have made important contributions. To further these efforts, improvements in FF accuracy based on extensions of the underlying energy function are required, including the explicit inclusion of induced electronic polarizability. In the proposed study we will investigate the fundamental forces driving a range of phenomena in both proteins and lipids using the computationally efficient polarizable FF based on the classical Drude oscillator, allowing for unbiased simulations on the microsecond time scale as well as simulations exploiting a range of enhanced sampling technologies. Investigations will focus on how the explicit treatm ent of electronic polarization underlies a range of experimental observables. These efforts will yield a better understanding of the fundamental forces driving the structure and dynamics of biomolecular systems while also helping to validate the current generation of the Drude FF. Results from these studies will guide efforts aimed at further improving the model for proteins, lipids, metal ions, and drug-like molecules. We have made significant progress towards improvements in the CHARMM36 additive protein and lipid FF, development of methods to facilitate FF optimization, and the 1st generation Drude polarizable FF for atomic ions, proteins and the lipid DPPC. We have subsequently used these models in large-scale simulations, demonstrating quantitative agreement with experimental data associated with the improved model of the fundamental forces due to the explicit treatment of electronic polarizability. Preliminary results already show that new physical mechanisms emerge when explicitly accounting for induced polarization. In the near future we will extend these efforts, focusing on selected peptides and proteins, selected membranes and membrane-protein interactions and on ligand-protein interactions including ions and drug-like molecules. Specific improvements will be made in the intra- and intermolecular electrostatic and Lennard-Jones (LJ) terms, the treatment of non-bonded interactions between different classes of functional groups not specifically targeted until now, and the extension of the energy function to include charge transfer between selected moieties in the FF. To use FF methods it is necessary to automatically generate parameters for a wide range of chemical entities of interest at a level of accuracy required for reliable predictions. Our previous experience towards this goal with the development of the General Automated Atomic Model Parametrization (GAAMP) web server (see http://gaamp.lcrc.anl.gov) is encouraging. GAAMP was used to optimize electrostatic parameters for 217 molecules within the polarizable Drude model, with the resulting model yielding a R2 with respect to experiment of 0.87 and an AUE of 0.92 kcal/mol. These results compare to values of R2=0.90 and AUE=0.85 for GAFF/AM1-BCC model despite that GAFF-adapted LJ parameters were used rather than LJ parameters optimized specifically for the Drude electrostatic model. To improve the treatment of ligands and drug-like molecules we will extend and optimize the Drude LJ parameters. Work will involve the implementation of a systematic atom typing algorithm together with t he development of new LJ parameters targeting neat liquid properties and relative interaction energies from high level QM calculations, expanding on methods developed in the MacKerell laboratory. Utilities such as the GAAMP will be extended for the automated optimization of Drude parameters for drug-like molecules in anticipation of the application of the Drude model in the area of drug discovery. These efforts will lay the groundwork for a readily extendible Drude FF for drug-like molecules, that will be put into an engine similar to that in place for the additive small molecule CGenFF. Specifically, in the coming year we will work on: 1. Fundamental forces at play in membranes and membrane-protein interactions. Investigations will focus on the impact of polarization of the nonpolar hydrocarbon core and the internal dielectric structure of membranes on the reproduction of experimental data including 2D-IR spectroscopy, solution and solid state NMR, scattering, ion binding to charged lipids, permeation of small species and cell penetrating peptides, and voltage-sensitive membrane-bound chromophores. Results will be used to drive additional lipid FF optimization, and will include extension to unsaturated and anionic lipids, cholesterol and sphingomyelin. 2. Fundamental forces at play in ligand binding. The impact of polarization on the binding of ions and ligand to proteins will be investigated. The allosteric regulation of thrombin by Na+ binding will be elucidated. The fundamental forces governing the binding of charged ligands to cytochrome c peroxidase and the acetylcholine-binding protein (AChBP) will be investigated. An automated parameter optimization approach will be extended and applied to a collection of model compounds with broad chemical functionalities representative of drug molecules, yielding an engine that will allow for the objective generation of topologies and parameters for drug-like compounds that will be compatible with the remainder of the Drude FF. Towards the goal of providing a comprehensive and accurate description for the behaviors of ions interacting with macromolecule, in the next allocation cycle, we will focus on improving polarizable ion force fields based on the classical Drude oscillator models. The Drude polarizable force field has been implemented in the program NAMD. It runs efficiently, at a computational cost comparable to that of non-polarizable force fields (slower only by a factor of about 1.4). This study will provide a systematic protocol by employing a wide range of target data. Besides the conventionally employed benchmarks in parameterization such as the gas-phase ion-small molecular cluster interactions and the experimental measured condensed-phase properties, the proposed study will also incorporate QM binding energy calculations for a wide spectrum of proteins. Specifically, we will improve models for biologically important ions interacting with proteins, including the alkali cation Na+, K+, a lkaline earth metal cations Ca2+, Mg2+, and transition metals Zn2+ and Ba2+. In addition, we hope that the approach employed will help initiating and facilitating a systematical approach for evaluating force field performances, thus allowing scientists resolving the functions and behaviors of ions in biological systems more efficiently and with higher confidence. In the propose study, we will perform MD simulations and QM interaction energy calculations for a large number of ion-binding proteins (10 different proteins for each ion type) and utilize these data as the benchmarks in the parameter optimization process. The work will yield novel insights on the fundamental contribution of electronic polarization to biological phenomena such as peptide folding, membrane permeation and ligand-protein interactions. This work will allow us to validate and improve the accuracy of the Drude FF, which is currently the **only** polarizable FF that allows for computationally efficient MD simulations on the microsecond time scale for large biomolecular systems. An allocation from LCRC of the Argonne National Lab is therefore requested to accomplish this project. Industry partnership: Project URL: http://thallium.bsd.uchicago.edu/RouxLab/ Current FY Hours Used: undetermined amount New FY Requested allocation: 2000000 Q1: 500000 Q2: 500000 Q3: 500000 Q4: 500000 Justification: We are mainly using NAMD for molecular dynamics simulations with a polarizable force field based on Drude oscillators. This HPC code is well documented and scales very well on all platforms. We also use Gaussian (in GAAMP) to compute energies from sets of configurations previously generated with classical molecular dynamics. On average the configurations used in Gaussian have 50-100 atoms, so the QM calculations are substantial. While the computations are extensive, Gaussian runs well on fusion and it well optimized. Storage requirements: 1 TB Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for orfpga
by accounts＠lcrc.anl.gov 07 Sep '16

07 Sep '16

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Azamat Mametjanov Project Name: orfpga Division: MCS Project title: Autotuning FPGA designs Associated funding: NASA Other Systems: Science: Designing FPGAs for space applications is a complex iterative process involving simultaneous optimization of timing score, device utilization, and power consumption cite{Gaisler1,Gaisler2}. Achieving near optimal performance is challenging even for the most experienced engineers. The number of vendors offering FPGAs (e.g., Xilinx, Altera, QuickLogic, Actel/Microsmei) for space applications has also increased. Each of these vendors base their FPGAs on custom architectures and thus, the best performing design for a given application varies depending on the chosen FPGA architecture. In addition, the optimizations parameters, techniques, and best practices often differ for each vendor's FPGA. For organizations that use FPGAs from multiple vendors, this exacerbates the already lengthly FPGA design optimization process. The FPGA design techniques have evolved to provide abstractions such as IPCores (Intellectual Property Cores), which provide individual functions that can be composed to build larger designs. Therefore, in order to reduce development effort and increase code reuse, IPCores are often used in FPGA designs. However, IPCores have many tunable parameters cite{Junchao} that affect the performance of an FPGA design. In addition, the designer wants to experiment with various HDL (VHDL, Verilog etc.) and HLS (high level synthesis) code constructs. For example, the user might want to try different pipeline depths for a given piece of HLS code. Lastly, the designer also needs to experiment with different input parameters and random seeds to the synthesis and implementation (Mapping, Placement and Routing (PAR)) tools. These user settings can have dramatic impact on design performance because they influence the usage of underlying optimizations such as merging, retiming, and pipelining cite{SE1}. Thus the user tunable parameter space includes: 1) user customizable IPCores parameters, 2) various HDL and HLS code constructs, and 3) input parameters to the Synthesis, Mapping, and Placement and Routing tools. The choice of these parameters and the underlying FPGA architecture significantly affects the performance of an FPGA design. Manually exploring the aforementioned FPGA parameter space is a time consuming task and impedes designer productivity. The existing FPGA design tools (e.g., Xilinx Vivado/ISE Design Suites, Altera Quartus II) are unable to automatically deliver the best possible design for an application through the underlying optimization algorithms. Further HDL/HLS code optimization, and tuning of parameters is required at the designer level to meet the performance goals. This situation results in a potentially large gap between the achieved FPGA performance and best achievable performance on a given FPGA architecture. There is a need among NASA FPGA developers for next generation tools that can generate performance models and assist in deducing the near optimal designs through higher level inputs (performance goals, possible parameter values, different code variants etc.). Such performance optimization tools cite{Vuduc,Hart2009Orio,OrCuda2012, WN147,Tiwari2011,BalWilNor2011} are very prominent in high performance computing (HPC) domain. For scientific computing codes, these autotuning tools greatly assist performance optimizations and improve developer productivity. For example, the Orio tool cite{Hart2009Orio, OrCuda2012} (developed by ANL) takes annotated C or FORTRAN source code as input, generates optimized code variants from the annotated code, and empirically evaluates the performance of the generated codes. Orio then selects the best performing version to use for production runs. Because the search space of all possible optimized code versions can be huge and brute-force search strategy may not always be feasible, Orio provides various search methods (e.g., random search, simulated annealing) for reducing the search space and thus the empirical testing time. Significant performance gains in real scientific applications (an order of magnitude speedups for plasma physics code and climate modeling applications cite{Chung}) have been realized using the autotuning tools. There are no similar autotuning tools for FPGA design. In order to apply similar principles to FPGA designs, additional research and development is necessary. Significant gaps must be filled to automate the steps required for creating performance models on which the FPGA design decisions can be made. Building FPGA designs (synthesis and implementation) suffers from long execution times resulting in enormous compute time for empirical based autotuning. To be suitable for use in NASA labs, it is necessary to investigate new methodologies for optimizing the turnaround time. New automation tools need to be developed targeting the annotation of HDL/HLS code and FPGA design tool scripts. Finally, it is also required to develop a set of robust easy-to-use tools that interface with the existing FPGA design tools. By filling these gaps, the autotuning techniques from the HPC community can be made desirable and beneficial to NASA FPGA developers and the wider hardware development community. Project description: RNET and ANL are proposing to fully develop an empirical performance optimization tool called OrFPGA that efficiently explores the ``user tunable'' parameter space of an FPGA design and assists in deducing the near optimal design in terms of timing score, device utilization, and power consumption. The tunable parameter space will include IPCore parameters, HDL and HLS code constructs, and parameter settings for the vendor's design tools. Special automation tools will be devleoped to facilitate annotation of HDL/HLS code and design tool scripts. The computational magnitude of empirical performance tuning of FPGA designs will be addressed by novel machine learning based search algorithms (requiring minimal empirical evaluations), computational steering, leveraging intermediate performance analysis results, and parallelization techniques. The tool will support specification of prioritized performance metrics, easy-to-use interfaces for defining the parameter space, and intuitive visualization of performance models. The user will be able to accurately identify the optimal power consumption by including the environmental settings in search space and through real-time power monitoring. The benefits of the tool to NASA will be demonstrated in terms of performance metrics and cost benefits (user productivity) using real NASA designs that are used in space missions. Industry partnership: RNet Technologies, Inc. Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 200000 Q1: 50000 Q2: 50000 Q3: 50000 Q4: 50000 Justification: Storage requirements: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for glassforms
by accounts＠lcrc.anl.gov 07 Sep '16

07 Sep '16

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Peter Zapol Project Name: glassforms Division: MSD Project title: First-principles Calculations of Mineral and Glass Dissolution Associated funding: DOE-NE Other Systems: Science: The dissolution of glass forms in nuclear waste disposal determines their long term stability, which is critical for evaluating disposal scenarios. The dissolution behavior of minerals and glasses has been found to strongly depend on the pH value of the aqueous environment. The delicate balance of ions at the solid-water interface in aqueous solutions is still a big challenge both for accurate computational modeling and for unambiguous experimental measurements. This project aims to develop physical models to simulate the water-borosilicate glass forms interface at the atomic level under various pH conditions and to calculate the rates of dissolution for Si, Al and other constituting elements under acidic, neutral and basic conditions. These studies are based on first-principles calculations and kinetic Monte Carlo simulations. The calculated reaction barriers by using DFT calculations can then be used as input in kinetic Monte Carlo simulations to estimate the over all dissolution rates and compare with experimental values. These investigations will be very important to better understanding of the dissolution behavior and prediction of dissolution rates. Project description: The first-principles calculations will be carried out using the electronic-structure code VASP. NEB calculations will be performed to calculate the reaction pathways for a variety of composing elements under acidic, neutral and basic conditions. A single optimization calculation takes 48 hours on 32 processors. Last year, we have performed calculations on glass with B, Si and Na and made structures with Al and started hydrolysis reaction calculations with Al-containing glass. This year, we will continue with calculations of reaction barriers. We are considering 60 supercell samplings of silicate glasses including multiple components (Si, Al, B, Na, K). The calculation for possible reactants, intermediates, products will take about 200,000 core hours. It takes about 50 hours on 160 processors (5 images) to calculate a reaction barrier along the reaction pathway. The total estimated computer time for barrier calculation is 640,000 core hours. kMC simula tion will be performed to simulate the dissolution process of glass for 4 different temperatures and 4 different pH values for 72 hours (each simulation should run 50 times in order to get the statistical results). That requires 70,000 core hours. In sum, the total requested allocation for this project is 910,000 core hours. Industry partnership: Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 910 Q1: 230 Q2: 230 Q3: 220 Q4: 230 Justification: This project will primarily use the density functional theory (DFT) implemented in electronic structure code VASP. The system sizes of our calculations ranges from 100 to 300 atoms. VASP has been tested to scale very well in parallel processing on Blues. The optimum number of processors used for these systems is 16-64 processors per calculation. The kMC simulation code is written in Fortran. Storage requirements: Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for protein_e-wire2
by accounts＠lcrc.anl.gov 07 Sep '16

07 Sep '16

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: David Tiede Project Name: protein_e-wire2 Division: CSE Project title: Solar energy conversion in biomimetic molecular architectures Associated funding: This project is supported by three funding programs within the DOE-BES Division of Chemical Sciences, Geosciences, and Biosciences: i) Solar Photochemistry; ii) Photosynthetic Systems; iii) Energy Frontier Research Center: Argonne Northwestern Solar En Other Systems: N/A Science: This computational project will supplement our experimental work on construction and characterization of water-splitting catalysts and biomimetic molecular architectures for solar energy conversion. The first aim is to explore replica-exchange all-atom Molecular Dynamics (MD) simulations as a tool to achieve a better sampling and to predict impact of protein mutations and chemical modifications on conformational dynamics and structural stability with various photosensitizers and photosensitizer-catalyst dyads covalently linked to PpcA, a multi-heme c-type cytochrome from Geobacter sulfurreducens. These results will facilitate interpretation of kinetic electron transfer rates as well as Small- and Wide-Angle X-ray Scattering (SAXS, WAXS) as well as NMR data to provide an insight into strategies for optimization of electron transfer rates within biomimetic hybrids. The second aim is to use replica-exchange MD to simulate equilibrium positions and dynamics of mutation s needed in order to tune redox potentials of PpcA hemes. Our third goal is to use several types of ligand docking simulations to screen and optimize ligand binding sites on PpcA protein surface. Project description: We are developing biomimetic molecular architectures for efficient solar energy conversion using artificial and natural photosensitizers combined with natural and genetically engineered host systems capable to support long-lived charge-separated states and conduct charges away from the photosensitizers. We recently developed and characterized a series of six cysteine mutants of PpcA, a multiheme redox protein from Geobacter sulfurreducens, which demonstrate picosecond photo-induced electron transfer rates. Further, at least 2 attachment sites showed negligible activation energy barriers resulting in similar apparent reaction rates at 2.7 and 295K. These are remarkable results considering that neither picosecond photo-induced electron transfer rates nor activationless reaction rates were observed in comparable synthetic systems, though they are common in natural photosynthesis. However, any attempts in understanding how and why our experimental system rep licates Nature will require detailed structural information in order to interpret the observed kinetic data and to guide our future developments. Obtaining high quality crystals in order to solve structures of labeled PpcA mutants so far was elusive. The molecular dynamics simulations combined with in silico mutation followed by validating computational data with experimental X-ray scattering work and HSQC NMR provide us with a valuable tool to visualize structural changes, generate hypotheses about the mechanisms and factors controlling charge transfer and to develop experimental testing approaches. In the past we have developed CHARMM force field parameters for several photosensitizers and obtained promising preliminary results for some of PpcA mutants in their oxidized form linked to Ru(bpy)3 which closely matched X-ray scattering data. In FY2017 our main focus will be on simulations of the most promising PpcA mutants with several new classes of photosensitizers with a particular focus on compounds using earth-abundant elements. Based on our previous experience, we expect that triplicate 300-500 ns all-atom MD simulations of selected PpcA biohybrids with explicit solvent will provide a sufficient sampling of protein conformational space and help us to better rationalize the collected kinetic data on charge transfer rates and SAXS/WAXS scattering profiles as well as the observed perturbations in the NMR data. At the initial stages of the project we will focus on the systems with all three PpcA hemes either oxidized or reduced. This is particularly interesting as it resu lts in a significant change of the total protein change from +4 to +1. All MD simulations will be performed with NAMD 2.8 and CHARMM force field. We will use standard for protein MD conditions: nPT, PME electrostatics, explicit TIP3 water boxes sufficiently large to keep interaction energy negligible between protein mirror images. Later we repeat some simulation with polarizable forcefields and compare results. The analysis of MD trajectories, computational docking and calculation of X-ray scattering profiles will be carried out on our lab computers. The second major goal is to use several rigid and flexible ligand docking approaches in order to predict and optimize binding sites for a new generation of photosensitizers and photosensitizer-catalyst dyads. We will continue to use an MPI version of DOCK6. We previously successful used this code on Fusion and it already resulted in one peer-reviewed publication. Industry partnership: Project URL: http://blogs.anl.gov/solar-energy/ Current FY Hours Used: undetermined amount New FY Requested allocation: 1000000 Q1: 250000 Q2: 250000 Q3: 250000 Q4: 250000 Justification: NAMD is a highly scalable molecular dynamics simulations code. Our past benchmarks on Fusion with PpcA in a water box (~13,000 atoms) showed more than 99% speed-up efficiency with ~1,600 atoms/CPU core and ~80% efficiency with ~400 atoms/core and yielded 5.1 and 15.7 ns/day on 8 and 32 cores, respectively. DOCK 6 typically had more than 95% scaling efficiency on 32-128 cores used in our previous work and we expect to get similar scaling figures with new simulations Storage requirements: 1TB Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Yearly Allocation Request for StructJuMP
by accounts＠lcrc.anl.gov 07 Sep '16

07 Sep '16

Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Feng Qiang Project Name: StructJuMP Division: MCS Project title: Parallel modelling of stochastic power grid in Julia Associated funding: ASCR,OE, GMLC Other Systems: Science: The objective of this project is to develop and test a parallel, memory-distributed, modeling framework for stochastic optimization problems arising in the optimization of power grid models under uncertainty. We plan to prove scalability and real-time capabilities that make the framework feasible to existing operational practices in power grid industry. Project description: Our previous parallel runs showed that the solver scales very well (>90% parallel efficiency on Intrepid, Mira and Fusion) but determined that the instantiation of the model is far beyond this efficiency. To remove this bottleneck we have moved from AMPL (who is an outdated modeling language for mathematical optimization with virtually no parallel capabilities) to Julia, which has been used by collaborators from MIT to develop JuMP, a modeling framework for optimization. We have build the StructJuMP, a JuMP extension for modeling structured optimization problems. StructJuMP can be used for modeling both linear and nonlinear programming and also fully integrated with our solver, so that we have successfully demonstrate the ability to model and solve stochastic optimization problems from power grid industry in parallel on HPC cluster. The parallelism in the instantiation of such problems comes from the particular structure these problem have (which has been very successfully exploited by the solver, as we mentioned above). Our preliminary results from a test problem on the Blues/Fusion cluster shows that the efficiency of the parallel problem generation is even better than the efficiency of the solver. However, we will need to investigate more into the model setup up time which currently doesn't seem to be scale well. We also plan to solve much larger problem instances than our testing problem. We expect this to scale up to the full Blues/Fusion. We expect a few full machine runs and other smaller runs that are needed to investigate scaling efficiency. Since we want to show a full integration with the solver PIPS, we will need to solve these problems to optimality. For this we expect to use around 80 thousand core hours in total for this year. Running our modeling and solver framework requires MPICH2 library. It would be great if Julia is pre-installed, however, we are also be able to build our own Julia executable from source. I expect the project will have 3 members (collaborators + summer student(s)). Industry partnership: Project URL: http://github.com/StructJuMP/StructJuMP.jl Current FY Hours Used: undetermined amount New FY Requested allocation: 80000 Q1: 10000 Q2: 20000 Q3: 20000 Q4: 30000 Justification: Storage requirements: 1TB will be sufficient Thank You, The LCRC Accounts System

1 0

[LCRC Accounts] Project Request: perfect
by accounts＠lcrc.anl.gov 07 Sep '16

07 Sep '16

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: John J. Low Applicant's institution: ANL Applicant's division: EGS, CLS, LCF, APS, MSD Project Name: perfect Project title: The Perfect Thermodynamics of Imperfect Materials Associated funding: LDRD Other Systems: ALCF/INCITE Science: We focus on high-k dielectric materials for complementary metal-oxide-semiconductor (CMOS), which are of particular importance for creating Dynamic Memory Allocation (DRAM) devices. Many CMOS properties strongly depend on material defects such as vacancies, interstitials, defect clusters that occur during synthesis, and thermal treatment. Inclusion of other chemical elements (e.g., dopants) in CMOS can significantly change physical properties such as thermal conductivity, electrical conductivity, and magnetism. Far from representing a drawback, the understanding imperfections, including defects and dopant, leads to prediction and control of properties and the design of post-CMOS materials for the next generation of electronic devices. There is, however, a challenge: the thermodynamic stability of CMOS phases, a key element for any application, is difficult to evaluate, especially at high temperatures, close to the phase transformations or the melting point. In this project, we develop the theoretical, experimental, and computational knowledge and tools for predicting the thermodynamic and kinetic stability/metastability of defected CMOS phases at temperatures ranging from 0K to melting. As a proof of principle, we will study the high k dielectric in complementary metal–oxide–semiconductor (CMOS), using as a prototype hafnia, HfO2. Hafnia has unusual electronic and magnetic properties that make it attractive to both science and technology. One of the most important applications is based on hafnia’s high-k dielectric properties, which are of particular importance for creating Dynamic Memory Allocation (DRAM) devices. In 2007, Intel decided to replace silicon oxide with hafnia as gate insulator for field-effect transistors. Current and future technology (sub 14nm) are using “hafnia based” materials (including additional elements). Many CMOS properties strongly depend on materials defects such as vacancies, interstitials, defect clusters (Figure 1), and how the material was synthesized and pretreated. Inclusion of other chemical elements (dopants) in CMOS can significantly change physical properties such as thermal conductivity, electrical conductiv ity, and magnetism. Far from representing a drawback in the integration of CMOS in the next generation of electronic devices, understanding imperfections (defects and dopants) can lead to prediction and control of important magnetic and electronic properties. Project description: Our initial studies will be focused on defects in Monoclinic HfO2. This will allow us to compare our results to the available experimental and theoretical data. We will focus on oxygen vacancies, hydrogen interstitials, nitrogen dopants and grain boundaries. These are common defects in HfO2 high K-dielectrics. Additional work will done modeling the crystallization of amorphous hafnia. Since we will need to study amorphous high temperature phases, the standard method of estimating entropy from phonon bands will not be adequate. We propose to use Bennett methods (Bennett, J. Comp. Phys. 22(1976)245) to estimate thermodynamics of defect formation. This approach involves computing the overlap of two molecular dynamics trajectories to evaluate the free energy. This approach will correctly include the large amplitude motion important at high temperatures near defects and phase transitions. Published reactive force-fields (Phillipot, Phys. Rev. B 81(2010)125328) will be use to establish the length-scale and time-scale required to get converged thermodynamics with Bennett Methods. We estimate that the preliminary force-field studies will require 100,000 core-hours. The SCAN metafunctional [Sun, et al., Nature Chem. 8(2016)831 and ACBN0 pseudo-hybrid functional [Nardelli, et al. Phys. Rev. B 93(2016)0351104] will be evaluated in this work. Both of these recently develop functionals have been shown to yield accurate bond gaps, lattice constants and bulk moduli. All of which is required for accurate predictions of energetics for band gap formation. SCAN has been implemented in VASP and libxc (Abinit and CP2K). ACBN0 has been implemented in Quantum Espresso and will be obtained through a collaboration with Nardelli (Unv. N. Texas). The SCAN functionals are about 30% more expensive than standard LDA and GGA functionals, over an order of magnitude less computationally intense than hybrid functionals and correctly predict that Mott insulators are insulators. ACNB0 is a DFT+U method with no empirical parameters. These functionals will enable the accurate modeling for defects with ab initio molecular dynamics from first principles. We pla n to calculate lattice constants, bulk module, band gaps and energies for defect formation at 0K to validate the SCAN and ACBN0 functionals. This work will be done with 150,000 core-hours of LCRC resources before doing production runs using ALCF resources. Hafnia in semiconductors is actually a thin (10-100 nm) amorphous layer between layers of Si. The structure of this nanolayer is determined by how is was grown. Typically, atomic layer deposition or chemical vapor deposition is used to fabricate these monolayers. We will follow the procedure of Shirazi and Elliot [J. Comp. Chem. 35(2014)244] to use Kinetic Monte Carlo to predict the structure of the layer of amorphous HfO2 grown on Si during ALD or CVD. The predicted reaction rates, needed by KMC, will be derived from density functional calculations. We will be evaluate Eon [Henkelman and Jonsson, JCP 115(2001)9 657] and SPPARKS [Plimpton, et al. JCP 128(2008)205101] to solve the KMC equations. We estimate that 250,000 core-hours will be required for this evaluation. A critical parameter is the crystallization temperature of the amorphous layer of hafnia. Grain boundaries will form between nanocrystals during crystallization. The grain boundaries will carry current which is not desirable in a dielectric. The processing temperature of the DRAM chip must be less than crystallization temperature. Since the amorphous phase is metastable the standard process of estimating free energies from energies and phonon bands to predict phase transitions is not adequate. We propose to use Forward Flux Sampling (FFS) [Allen, et al. J. Phys.: Condens. Matter. 21(2009)463102] to model the crystallization process, predict rates and the crystallization temperature. We will use reactive force-fields in LAMMPS in the FFS simulations. We will develop LAMMPS scripts to conduct FFS through collaboration with Art Voter (LANL). We estimate that this study will require 100,000 core-hours. All data resulting from theory and calculations will be evaluated using Bayesian Analysis (BA) to determine the optimal models (i.e., mathematical expressions) that relate the free energy of the defected oxide phases to temperature and defect concentration (first and second year). To this end, we will use a Non-Naive Bayesian Network machine learning algorithm and software to process that will analyze the training and validation data sets in an unsupervised mode. The data generated by the DFT and kMC computations described above will enrich the diversity of the samples and reduce the final error in our estimations. The computationally intensive calculations wit evaluate the stability and meta-stability of phases with uncertainty intervals, using a methodology developed by the PI and co-authors (M. Stan and B. Reardon., CALPHAD, 27 (2003) 319). We estimate that 100,000 core-hours will be required for these calculations. We expect up to four project members. John Low, Marius Stan, a postdoc and summer student. Industry partnership: Project URL: Requested allocation: 700000 Q1: 175000 Q2: 175000 Q3: 175000 Q4: 175000 Justification: LAMMPS is a high performance code which has been optimized to run on Haswell processors, Nvidia GPUs, and Xeon Phi processors. Molecular dynamics for thousands of atoms will be used to model crystallation which will scale well to thousands or cores. See http://lammps.sandia.gov/bench.html for more details. Quantum Espresso can be built as a hybrid MPI/openMP executable and has been demonstrated good parallel scaling to 1 core for every 2 atoms and fair parallel scaling to 1 core for every atom. Since we are planning to use on the order of hundred atoms for our validation studies, our jobs will scale to about a hundred. We will profile QE with the Intel and Allinea tools to locate potential bottlenecks and opportunities for performance enhancement. See http://www.quantum-espresso.org/benchmarks/ for more details. CP2K is a high performance code which typically scales to one atom per core. It can exploit Haswells, GPUs and the Xeon PHI architecture as well. It should scale to hundred of cores for the models we will be using. See https://www.cp2k.org/performance#h20-64 for more details. Eon and SPPARKS are both designed to run parallel. Eon with run simulataniously run multiple LAMMPS jobs simultaneously. Since multiple jobs run in parallel with be embarrassingly parallel, the Kinetic Monte Carlo has good parallel scaling to a thousand cores or more. See http://henkelmanlab.org/pubs/chill14_055002.pdf for details. Storage requirements: The requester has used undetermined amount 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