Rolling reboots of all clusters, starting from 8 AM Tuesday, June 19, 2018

Statewide Users Group Meeting Agenda - Dec 4, 2014

Wednesday, Dec. 3th

4:00  Allocations Committee Meeting

Thursday, Dec. 4th

10:30 - 11:30

Software Committee Meeting

Hardware Committee Meeting

Face:Face OSCHelp

11:30 - 12:30 Lunch
12:30 - 1:15

Keynote Address
NSF support for High Performance Computing and Cyberinfrastructure
Dr. Evelyn M. Goldfield
Program Director, Chemistry Division, NSF

1:15 - 1:45

Invited Talk
Total Sim USA and the Ohio Supercomputer Center
Ray Leto
President of Total Sim USA

1:45 - 2:00 Break
2:00 - 3:00 OSC Perspective
3:00 - 4:00 Flash Talk Contest
4:00 - 6:00

Poster Contest

Interface Lab Open House

6:00 Social Outing at Hofbräuhaus

Flash Talks

  • Jonathan Brown
  • Ginevra Cochran
  • Frank King
  • Matthew McMahon
  • Youngmi Seo
  • Joy-El Talbot
  • Mahendra Thapa


Poster Map

  • Ayse Arslanargin
  • Nikolas Antolin
  • Molly Ball
  • Katharine Cahill
  • Ginevra Cochran
  • Sukirth Ganesan
  • Michael Gibbons
  • Mark Hornak
  • Frank King
  • Matthew McMahon
  • Travis P. Pollard
  • August Powers
  • Rodney Richardson
  • David Riegner
  • Janani Sampath
  • Youngmi Seo
  • Anne Shim
  • Joy-El Talbot
  • Mahendra Thapa
  • Travis Withrow
  • Xing Zhang

Flash Talk and Poster Topics

Modeling ion solvation in ethylene carbonate and propylene carbonate
Ayse Arslanargin, University of Cincinnati

Non-aqueous solvents are widely used as liquid electrolytes in lithium-ion batteries. Understanding the solvent structure is essential for battery performance enhancements. This work investigates the thermodynamics of ion solvation in ethylene carbonate and propylene carbonate. Free energy and enthalpy of solvation calculations have been conducted employing different force fields. Simulated annealing calculations have been performed to alter non-bonded energy parameters. The new parameters result in good agreement with the experimental free energy of solvation values, while the enthalpy of solvation results show deviations from the experimental data. These results suggest that classical models often do not accurately predict basic interactions in ion-solvent systems.

Phonon induced magnetism in diamagnetic materials
Nikolas Antolin, The Ohio State University

Thermoelectric phenomena in magnetic materials create exciting possibilities in future  spin caloritronicdevices by manipulating spin information using heat. An accurate understanding of spin-lattice interactions, i.e. coupling between magnetic excitations (magnons) and phonons (lattice vibrations), holds the key to unraveling their underlying physics. Here we present ab-initio frozen-phonon calculations of CsI and InSb that result in non-zero magnetization when degeneracy between spin-up and spin-down electronic density of states is lifted for certain phonon displacements. We discuss mechanisms of this previously unobserved effect as well as present data collected by collaborating OSU researchers measuring the magnitude of this response in InSb.

First principles study of the origin of Strain-tunable extraordinary magnetocrystalline anisotropy in Sr2CrReO6 epitaxial films
Molly Ball, The Ohio State University

We report the discovery of extraordinarily large, strain-tunable magnetocrystalline anisotropy (MCA) in Sr2CrReO6 epitaxial films. These films grown on several different substrates undergo dramatic changes in MCA shown by a shift in easy-axis from in-plane to out-of plane. We determine the strain-induced structural distortions by performing density functional theory (DFT) calculations using VASP. The change in easy-axis under strain can be examined through the energy differences for magnetic orientation along crystalline axes known as the magnetic anisotropy energy (MAE). We are able to establish the origin of this large strain-tunable MCA by directly relating the MAE to the orbital moments.

Co-Authors:  M. R. Ball, J. M. Lucy, OD. Restrepo, A. J. Hauser, J. R. Soliz, J. W. Freeland, P. M. Woodward, W. Windl, and F. Y. Yang

Phase Behavior of Tapered Diblock Copolymers from Self-Consistent Field Theory
Jonathan Brown, The Ohio State University

Tapered block copolymers are similar to AB diblock copolymers, but with a statistical A-to-B (normal) or B-to-A (inverse) gradient “taper” between the A and B blocks. Depending on the sequence of monomers along the chain and the segregation strength χN, the A and B monomers microphase separate to form various ordered morphologies. We map the phase diagrams of model tapered polymers using self-consistent field theory (SCFT). The normal tapered systems are similar to diblocks, but show a wider region of network structures. In inverse tapered systems, the polymer can fold across the interface and new ordered structures are found.

Co-Authors:  Jonathan Brown, Scott W. Sides, Lisa M. Hall

Molecular docking study of Organophosphorus pesticides with G3C9 and its variants
Katharine Cahill, The Ohio State University

Organophosphorus (OP) compounds are highly toxic chemicals capable of inhibiting the hydrolysis of the neurotransmitter acetylcholine by acetylcholinesterase. Catalytic hydrolysis of OPs with enzymatic bio-scavengers, such as paraoxonase (PON1), is an active avenue of investigation towards the treatment of OP exposure. G3C9 is a recombinant PON1 enzyme which was developed for its improved solubility and has some activity against OP pesticides. In this study, molecular docking simulations were performed on G3C9 and several of its variants. Docking analysis shows that, the V346A mutation significantly improves OP binding to the active site compared to G3C9. Several OP compounds with bulky leaving groups, including paraoxon and diazoxon, were studied to understand both the efficiency of binding as well as the orientation of the guest in the active site.

Co-Authors:  Katharine J. Cahill, Kiran Doddapaneni, Shameema Oottikkal, Thomas J. Magliery and Christopher M. Hadad

Convergence criteria for PIC simulations of electrons in an ultraintense laser field
Ginevra Cochran, The Ohio State University

We present a study of particle-in-cell (PIC) simulation error in modeling an electron in an ultraintense laser field. We find an unexpectedly small timestep is required to resolve the electron motion, decreasing with increasing laser intensity. We consider several sources of PIC error, and find that the error from the particle velocity and position advance is dominant. We derive the timestep constraint and find that it decreases with laser intensity. We find the particle advance error accumulates when the electron is at rest and the laser fields are strong, and present a sub-cycled particle advance which reduces error.

Co-Authors:  Alexey V. Arefiev, Douglass W. Schumacher, A.P.L. Robinson

What’s In A Cigarette? Nicotine, Not Metals, Affects Biofilm Transcription
Sukirth Ganesan, The Ohio State University

Dysbiotic oral microbial communities underlie the etiology of cancer, caries and periodontitis. Since smoking is a primary risk factor for these bacterially-driven diseases, a metatranscriptomic approach was used to examine the effect of smoke on gene transcription within oral biofilms. Polymicrobial biofilms were generated in smoke-free, smoke-rich, nicotine-conditioned and heavy-metal-depleted environments. Enriched mRNA was sequenced and analyzed using the computational tools incorporated within the MG-RAST pipeline. The pipeline was offloaded to the Oakley Cluster at OSC, allowing for expedited processing of immense datasets. Smoke significantly modulates transcription in oral biofilms with a majority of the effects attributable to nicotine.

Co-Authors: SM Dabdoub, W. Tang,  J. Bischof, F. Meyer, PS Kumar

FFT modeling of deformation behavior in bulk metallic glass composites
Michael Gibbons, The Ohio State University

Bulk metallic glass composites (BMGCs), consisting of an amorphous matrix and a homogeneous distribution of crystalline dendrites, offer a promising solution to the low fracture toughness and brittleness of pure bulk metallic glasses. To date, the interaction between the crystalline and amorphous phase during deformation is not well understood and will be examined here with fast Fourier transform (FFT) based continuum modeling. Given appropriate computational resources, the efficiency of this FFT approach allows us to study the mesoscale deformation behavior of these composites in 3D, using spatial and time resolutions high enough to obtain meaningful insight.

Determining Point Defect Energetics and Optics in Sapphire from First Principles
Mark Hornak, The Ohio State University

Single crystal α-Al2O3 (sapphire) is a potential candidate for optical fibers and sensors in extreme high-temperature radiation environments.  Transmission of light under such conditions can be impeded by the generation of defects within the material.  In order to validate sapphire, we determine the stable point defects, and their charge states, in stoichiometric sapphire.  The increase in attenuation due to individual point defects was calculated using density functional theory and hybrid functional mixing.  We find that oxygen and aluminum vacancies are dominant, and attenuate light in the 200-300nm range.  Oxygen divacancies also show significant attenuation in the 100-200nm and 300-450nm ranges.

Simulations of High Intensity Short Pulse Lasers Incident on Reduced Mass Targets Using LSP
Frank King, The Ohio State University

We present the results of a series of fully kinetic 2D and 3D simulations using the Particle-In-Cell code LSP for the study of the heating and deformation of micron scale targets.  These simulations model an experimental laser pulse incident on a realistic several micron thick copper target as a function of intensity, spot size, pre-plasma, and target lateral extent and thickness.  We observe that the target deformation and heating has a strong dependence on intensity of the laser pulse and creation of a shock in the target.

Co-Authors: Chris Orban, Kramer U. Akli, Douglass Schumacher

First PIC simulations modeling the interaction of ultra-intense lasers with sub-micron, liquid crystal targets
Matthew McMahon, The Ohio State University

We recently introduced liquid crystal films as on-demand, variable thickness (50 – 5000 nanometers), inexpensive targets for intense laser experiments. Here we present the first particle-in-cell (PIC) simulations of short pulse laser excitation of liquid crystal targets using the PIC code LSP. In order to accurately model the target evolution, a low starting temperature and field ionization model are employed. This is essential as large starting temperatures lead to expansion of the target causing significant reduction of the target density before the laser pulse can interact. We also present an investigation of the modification of laser pulses by very thin targets

The thermodynamics of proton hydration and the electrochemical surface potential
Travis P. Pollard, University of Cincinnati

A major roadblock in refining our understanding of the role of ions in chemistry and biology is the inability of experiment to resolve single-ion contributions to the hydration thermodynamics of salts. Numerous methods have been proposed to decompose the whole into a sum of its parts. These methods agree on the thermodynamics of various salts but not on individual ion quantities which appear systematically shifted from one another by a charge-dependent amount (q*x). The poster highlights our recent work, which concludes that a solvent-specific surface potential (x) of -11.6 to -9 kcal/mol-e may account for the observed differences.

Co-Author: Thomas L. Beck

Investigating Classical Models of Non-Aqueous Ion Solvation
August Powers, University of Cincinnati

Modeling energy storage systems is an important avenue in the optimization and advance of these technologies. This work investigates the classical models used in describing ion solvation in non-aqueous systems relevant to lithium-ion batteries and supercapacitors. Quantum mechanical data was used to try improving the van der Waals parameters for ions solvating in ethylene carbonate; improvements (relative to experimental data) were seen for the solvation free energies, but not for the solvation enthalpies. Additional symmetry-adapted perturbation theory results suggest the discrepancies may lie in the description of the non-electrostatic interactions within the ions' first solvation shells.

Co-Authors: Ayse Arslanargin, Thomas Beck

A comparative study of ITS2 metabarcoding and traditional microscopic palynology as methods of identifying taxonomic origins of bee collected pollen
Rodney Richardson, The Ohio State University

Honey bees, Apis mellifera, display high floral fidelity as they collect pollen from across large geographical areas.  Honey bees pack pollen into corbicular pellets for transport back to the colony. Identifying the flowers used by bees is important for understanding both bee and plant biology. Traditionally, microscopic palynology has been employed to identify floral sources. We developed a novel, molecular strategy for determining the floral sources of bee collected pollen, which involves amplifying the ITS2 locus using universal primers (Chen et al., 2010), followed by amplicon sequencing on the Illumina MiSeq platform

Co-Authors: Chia-Hua Lin, Juan QuijiaPillajo, Douglas B. Sponsler, Karen Goodell and Reed Johnson

Molecular Dynamics Simulations of Al-La Glass and Liquid
David Riegner, The Ohio State University

Metallic glass, a metal without an ordered atomic structure, may represent the next generation of engineering material. Design and application of these materials is hampered by long-standing challenges in fabrication and characterization. Molecular dynamics simulations can probe properties of undercooled (below their melting temperature) metallic liquids as they transition from liquid to glass. Characteristics and mechanisms at work during this transition, some attainable only through computer simulations, will narrow future experiments to target only systems and compositions that most readily assume a glassy state.

Effect of Aggregation on the Mechanical Properties of Ionomers From Molecular Dynamics Simulations
Janani Sampath, The Ohio State University

Ionomers are polymers with a small fraction of charged monomers that have a wide range of applications. We consider dense melts of ionomers and counterions with no solvent; an important aspect of their performance is the aggregation of ions, which holds polymer chains together like temporary cross-links. Because of the size scales involved, it is difficult to obtain a complete 3D microscopic picture of polymer aggregation. By performing MD simulations of ionomers of various architectures, we will show aggregate morphology and scattering profiles. Connecting these results with observed mechanical features will suggest how to design new ionomers with improved properties.

Co-Author: Lisa M. Hall

Molecular Dynamics Simulations of Microphase Separating Tapered Diblock Copolymers
Youngmi Seo, The Ohio State University

Tapered AB copolymers consist of pure A and B blocks separated by a midblock whose composition is a statistical gradient from A to B (B to A for an inverse taper); they can microphase-separate into various ordered phases. Taper length can be used as a tuning parameter to control microphase separation, but better physical understanding is needed for experimentalists to use this new parameter effectively. Using coarse-grained molecular dynamics simulations, we study structure and dynamics of these materials. Among other results, we show that inversely tapered polymers fold across the microphase interface, leading to significantly shorter domain spacing than diblocks.

Co-Authors: Jonathan R. Brown and Lisa M. Hall

Micellular encapsulation of nanoparticles from dissipative particle dynamics simulations
Anne Shim, The Ohio State University

Polymer-protected nanoparticles are of interest for drug delivery and medical imaging. Experimentalists need to generate uniform size micelles containing a predetermined number of particles for the particles to be useful commercially. We aim show on a molecular scale how the polymer micellization around hard nanoparticles occurs and thus can be controlled experimentally. We perform dissipative particle dynamics simulations of polymers and nanoparticles in solution. The solvent interaction strength varies over time to account for changing solvent concentration, which allows the micelles to form. We observed how changing polymer concentration, polymer length, and particle size affect the system.

Co-Authors: Jonathan Brown and Lisa Hall

RNA polymerase IV affects nascent transcription in maize
Joy-El Talbot, UC-Berkeley and The Ohio State University

Three RNA polymerases transcribe eukaryotic genomes into RNA. The roles of additional plant RNA polymerases like Pol IV remain unclear, although transcriptional silencing of transposons and normal maize development require Pol IV. Seedling nascent transcriptomes identify a novel role for Pol IV: regulating transcription at gene boundaries. While most transposons remain untranscribed in Pol IV mutants, specific near-genic transposons may promote transcription. To test this hypothesis, I will collapse multi-mapping nascent transcriptome data (500Gb/library) into locus- and read-specific SQLite databases on OSC systems. The maize genome is >85% repetitive, therefore including multi-mapping reads will increase understanding of Pol IV.

Co-Authors: Karl F. Erhard, Jr. and Jay B. Hollick

Comparison of Side-chain Motion of the Protein Calbindin D9k in its Four Calcium Binding States using Molecular Dynamics Simulations
Mahendra Thapa, University of Cincinnati

Molecular Dynamics (MD) simulation of a protein helps to study motion and its development with time which may not be studied experimentally. The side chains of a protein play important role in folding ,ligand binding and interactions. We used AMBER 12.0 software suit on GPU to simulate a protein Calbindin D9k (CAB), which is involved in the uptake and transport of the calcium, on its four calcium binding states: a doubly-loaded state, two singly-loaded states and an apo state. Force field ff12SB and water model TIP3P at NVT condition were chosen in the simulation. Experimental and computational studies on the dynamics of backbone atoms of the protein confirmed that calcium binding occur in a positive cooperative fashion. Studies of the doubly loaded state of the protein by molecular dynamics simulation and NMR experiment further enhance the point. To further investigate by computation, MD approach has been used to study the side chain dynamics of all these states of the protein.

Co-Author: Dr. Mark Rance

Atomistic Simulation for a Better Atom Probe Algorithm
Travis Withrow, The Ohio State University

Atom probe tomography holds a lot of promise as a technique allowing materials engineers to examine a structure in an atom-by-atom basis but is hampered by a simplistic reconstruction algorithm that overly relies on instrumentalist interpretation and fails to take into account much of the physics of sample behavior in the machine. Supercomputer simulations of the atomistic behavior of samples under measurement conditions will allow us to improve the current reconstruction algorithm, remove much of the human error and allow for true error propagation based on the physics of atomic interactions within the device.

Simulations of Energy Transfer in a Self-Assembling Organic Nanotube
Zhi-Qiang You, The Ohio State University

Recent experimental work at Ohio State has led to the synthesis and atomic-resolution structural model for a self-assembling organic nanotube consisting of several hundred naphthalenediimide chromophore units.  Experimentally, this system exhibits rapid (sub-ps) fluorescence depolarization,

with rapid excited-state energy transfer dynamics, and the atomic detail of the structural model facilitates detailed electronic structure calculations that are presented here.  Calculated energy-transfer rates are consistent with experimental estimates, and significant excited-state delocalization is predicted.

Co-Authors: Jon Parquette, Christopher P. Jaroniec, John M. Herbert

Analytic Derivative Couplings for Spin-Flip TDDFT:  An Inexpensive and Topologically-Correct Approach to Simulating Photochemistry
Xing Zhang, The Ohio State University

We have derived, implemented, and tested analytic derivative couplings for the “spin-flip” variant of time-depdendent density functional theory (TDDFT), within the Q-Chem electronic structure code.  Unlike traditional TDDFT, the spin-flip version exhibits correct topology in the vicinity of conical intersections and can therefore correctly describe the “conical funnels” that are responsible for photochemistry and photophysics, but at DFT cost. This should facilitate both non-adiabatic molecular dynamics simulations and the mapping of excited-state reaction pathways in large molecules.  Preliminary applications will be presented.

Co-Author: John M. Herbert