Academia.eduAcademia.edu

Outline

L2 Milestone: Simulations as Data. Quantum Molecular Dynamics Contribution

2012

https://doi.org/10.2172/1051092

Abstract

Milestone description: Direct Numerical Simulations and Quantum Molecular Dynamics simulations have been used as one would use physical experiments for cases where the physical data are not available. The goal of this milestone is to assess these simulations in the manner that they are being used and the priority of the specific configurations being planned. Milestone closure criteria: Description of the work done to date and how it has been used. What are the current plans for various timeframes and why were the particular configurations where chosen? Also quantify the physical scaling as appropriate for inclusion in applications.

LA-UR-12-24783 Approved for public release; distribution is unlimited. Title: L2 Milestone: Simulations as Data. Quantum Molecular Dynamics Contribution Author(s): Kress, Joel D. Collins, Lee A. Burakovsky, Leonid Herring, Stuart D. Ticknor, Christopher Crockett, Scott Intended for: Report Disclaimer: Los Alamos National Laboratory, an affirmative action/equal opportunity employer,is operated by the Los Alamos National Security, LLC for the National NuclearSecurity Administration of the U.S. Department of Energy under contract DE-AC52-06NA25396. By approving this article, the publisher recognizes that the U.S. Government retains nonexclusive, royalty-free license to publish or reproduce the published form of this contribution, or to allow others to do so, for U.S. Government purposes. Los Alamos National Laboratory requests that the publisher identify this article as work performed under the auspices of the U.S. Departmentof Energy. Los Alamos National Laboratory strongly supports academic freedom and a researcher's right to publish; as an institution, however, the Laboratory does not endorse the viewpoint of a publication or guarantee its technical correctness. L2  Milestone:    Simulations  at  Data.   Quantum  Molecular  Dynamics  Contribution   J.  D.  Kress,  L.  A.  Collins,  L.  Burakovsky,  S.  D.  Herring,  C.  Ticknor,  S.  Crockett   Group  T-­‐1,  Theoretical  Division,  Los  Alamos  National  Laboratory   September  11,  2012     Milestone  description:    Direct  Numerical  Simulations  and  Quantum  Molecular   Dynamics  simulations  have  been  used  as  one  would  use  physical  experiments  for   cases  where  the  physical  data  are  not  available.    The  goal  of  this  milestone  is  to   assess  these  simulations  in  the  manner  that  they  are  being  used  and  the  priority  of   the  specific  configurations  being  planned.   Milestone  closure  criteria:    Description  of  the  work  done  to  date  and  how  it  has   been  used.    What  are  the  current  plans  for  various  timeframes  and  why  were  the   particular  configurations  where  chosen?    Also  quantify  the  physical  scaling  as   appropriate  for  inclusion  in  applications.   Large-­‐scale  radiation-­‐hydrodynamics  simulations  of  fluids  and  dense  plasmas  under   extreme  conditions  require  knowledge  of  certain  microscopic  properties  such  as   equation-­‐of-­‐state,  mass  diffusion,  viscosity,  opacity,  and  thermal  conductivity.   Matter  at  densities  greater  than  ambient  solid  and  temperatures  greater  than  1  eV   encompasses  a  wide  range  of  important  physical  environments  where  constantly   interacting  clusters,  molecules,  atoms,  positive  and/or  negative  ions,  and  free   electrons  make  simulation  difficult.  The  challenging  physical  environments  lead  to  a   paucity  of  experimental  data.  Thus,  the  need  for  “Simulations  as  Data”.    While   many  modeling  and  simulation  approaches  exist,  Quantum  Molecular  Dynamics   (QMD)  is  the  most  accurate  for  determining  properties  of  these  systems.    In  QMD,   the  ions  are  evolved  with  classical  equations  of  motion  and  the  electrons  are  treated   quantum  mechanically.    From  the  many-­‐electron  wavefunction  (or  density)  and  the   ion  positions,  a  fully  consistent  set  of  static,  dynamical  and  optical  properties  can   be  computed.    Finite-­‐temperature  versions  of  two  major  flavors  of  density   functional  theory  (DFT)  are  employed  for  the  electrons,  the  orbital-­‐based  Kohn-­‐ Sham  (KS)  and  the  density-­‐based  orbital-­‐free  (OF).    Due  to  computational  cost,  the   KSMD  approach  is  limited  to  temperatures  below  about  10  eV  for  matter  at  ambient   solid  densities.    For  higher  temperatures,  we  then  appeal  to  OFMD  that  smoothly   maps  onto  the  extreme  temperature  limit  of  a  plasma  of  classical  ions  and  electrons   thereby  forming  a  natural  bridge  between  the  low-­‐  and  high-­‐temperature  regimes.   In  previous  fiscal  years,  we  established  the  utility  of  KSMD  and  OFMD  simulations   for  predicting  the  equation  of  state  and  mass  transport  properties  (diffusion  and   shear  viscosity)  for  various  materials  over  a  wide  range  of  densities  from  ambient  to   many-­‐times  compressed  including:    LiH  for  temperatures  below  6  eV  [1,  2];     1   mixtures  of  deuterium-­‐tritium,  DT  [3],  Pu  [4],  and  U  [5]  for  temperatures  from  melt   up  to  5  keV.    In  these  regimes,  we  have  also  predicted  the  optical  properties   (absorption,  electrical  conductivity,  reflectivity,  and  thermal  electronic   conductivity)  for  deuterium  [6],  LiH  [1],  and  for  Be,  DT,  plastic  and  mixtures  thereof   [7].    Additionally,  the  present  work  has  benefitted  greatly  from  the  longstanding   collaboration  between  LANL  and  the  CEA  (including  J.  Clerouin,  G.  Zerah,  S.  Mazevet,   V.  Recoules,  and  F.  Lambert).   During  the  present  fiscal  year,  the  following  was  accomplished.    (For  details,  we   refer  the  reader  to  the  published  work,  the  slides  from  the  presentation,  and   subsequent  sections  of  the  report.):   1. Mass  transport  properties  (shear  viscosity  and  self-­‐diffusion  coefficient)  for   Li6D  and  U,  and  for  mixtures  thereof,  have  been  calculated  using  Orbital  Free   Molecular  Dynamics  (OFMD)  for  densities  between  ambient  and  10-­‐fold   compressed  and  for  temperatures  between  10  to  10,000  eV  [8].    The   tabulated  results  have  been  fitted  to  an  analytic  Dufreche-­‐Clerouin  formula   as  a  function  of  mass  density  and  temperature,  a  form  suitable  for  inclusion   in  radiation-­‐hydrodynamics  codes.   2. These  mass  transport  simulations  have  provided  an  additional  bonus  in  the   form  of  determining  the  equation  of  state  (EOS)  and  shock  Hugoniot  for   Lithium  Deuteride  (Li6D)  with  QMD.    Based  on  these  QMD  results,  a  new   SESAME  EOS  table  has  been  generated  (by  Scott  Crockett,  T-­‐1)  and  provided   to  the  design  community.    [See  below.]   3. OFMD  simulations  for  diffusion  and  viscosity  have  been  performed  for   mixtures  of  LiD  and  U  [9].    These  benchmark  (computationally-­‐demanding)   simulations  of  the  fully  interacting  LiD/U  mixtures  are  compared  with  results   obtained  using  a  pressure  (P)  matching  rule.    In  the  P-­‐matching  scheme,  the   results  for  LiD  [8]  are  mixed  with  the  OFMD  results  for  U  [5].    In  turn,  the  P-­‐ matching  results  were  used  to  generate  simple  interpolation  functions  of   density  and  temperature,  suitable  for  inclusion  in  radiation-­‐hydrodynamics   codes  [9].    [See  addendum  for  details.]   4. The  capability  to  perform  nonequilibrium  (NE)  OFMD  simulations  for  plasma   interfaces  has  been  implemented.  First  demonstrations  include  interfaces  of   hydrogen  isotopes  and  interfaces  of  carbon  and  deuterium/tritium  [10].        In   order  to  carry  out  NE  simulations  on  massively  parallel  platforms,  improved   equation  solvers  have  been  implemented  in  the  OFMD  code.  Up  to  65,536   cores  on  the  Cielo  platform  have  been  utilized.    (This  work  has  been   supported  jointly  with  ASC  under  the  Capability  Computing  Campaigns  2  and   3  and  through  a  Metropolis  postdoctoral  fellowship.)     5. Average-­‐atom  (AA)  calculations  of  EOS  and  electrical  and  thermal   conductivites  in  deuterium/tritium,  Be,  and  plastic  (C/H)  and  mixtures   thereof  have  been  compared  with  benchmark  QMD  simulations  [11].    The  ad     2   hoc  renormalization,  necessary  for  the  AA  calculations,  has  been  calibrated   against  the  QMD  results.    Also,  the  Quantum  Hypernetted  Chain  model,  used   to  introduce  approximately  the  effects  of  ion  distributions  in  many  simplified   prescriptions,  was  validated  against  KSMD  simulations  for  Al  [12].   6. Mass  transport  properties  (shear  viscosity  and  self-­‐diffusion  coefficient)  for   deuterium  for  mass  density  ρ  =  2  g/cm3  has  been  calculated  using  orbital  free   molecular  dynamics  (OFMD)  and  used  to  validate  a  common  approximation   (Landau/Braginskii  theory)  in  transport  models.  [See  below.]   7. The  shear  viscosity  for  liquid  copper  has  been  calculated  using  quantum   molecular  dynamics  (QMD)  and  compared  with  experimental  results   evaluated  with  the  Arrhenius  and  Andrade  equations.    These  results  were   communicated  to  Guy  Dimonte  (XCP-­‐5)  for  use  in  analyzing  interface   instabilities.  [See  below.]   Proposed  work  for  FY13  includes:   1. Calculations  of  equilibrium  static,  dynamical,  and  optical  properties  of   LiD,  U,  and  LiD/U  mixtures  –  includes  EOS,  viscosity,  mutual  diffusion,   opacity,  conductivities  –  with  both  KS  and  OF  QMD  methods,  for   densities  1-­‐10  times  ambient  solid  and  temperatures  from  10  to  1000   eV.   2. Development  of  techniques  to  determine  opacity  and  conductivity   properties  directly  from  the  OF  electron  density  allowing  more   efficient  calculations  across  a  broad  density/temperature  range   without  resort  to  the  KS  code.     3. Continue  development  and  implementation  of  non-­‐equilibrium   simulations  of  low  Z/high  Z  plasma  interfaces  using  both  QMD  and   classical  models.   4. Development  and  implementation  of  new  algorithms  for  enhanced   computation  of  the  electron  density  in  OFMD.   5. Ejecta  modeling  from  continuum  and  MD  methods  for  fission   fragments.         3   References   1. “Quantum  molecular  dynamics  simulations  of  warm  dense  lithium  hydride:     Examination  of  mixing  rules,”  D.  A.  Horner,  J.  D.  Kress,  L.  A.  Collins,  Phys.  Rev.   B  77,  064102-­‐1-­‐9  (2008).   2.  “Transport  properties  of  lithium  hydride  from  quantum  molecular  dynamics   and  orbital  free  molecular  dynamics,”  D.  A.  Horner,  F.  Lambert,  J.  D.  Kress,  L.   A.  Collins,  Phys.  Rev.  B  80,  024305-­‐1-­‐10  (2009).   3. “Viscosity  and  mutual  diffusion  of  deuterium-­‐tritium  mixtures  in  the  warm,   dense  matter  regime,”  D.  A.  Horner,  J.  S.  Cohen,  J.  D.  Kress,  L.  A.  Collins,  Phys.   Rev.  E  82,  036404-­‐1-­‐10  (2010).     4.  “Quantum  molecular  dynamics  simulations  of  transport  properties  in  liquid   and  dense-­‐plasma  plutonium,”  J.  D.  Kress,  J.  S.  Cohen,  D.  P.  Kilcrease,  D.  A.   Horner,  L.  A.  Collins,  Phys.  Rev.  E  83,  026404  (2011).   5. “Orbital-­‐free  molecular  dynamics  simulations  of  transport  properties  in   dense-­‐plasma  uranium,”  J.  D.  Kress,  J.  S.  Cohen,  D.  P.  Kilcrease,  D.  A.  Horner,  D.   A.  Horner,  L.  A.  Collins,  High  Energy  Density  Physics  7,  155  (2011).   6. “Reflectivity  of  warm  dense  deuterium  along  the  principal  Hugoniot,”  L.  A.   Collins,  J.  D.  Kress,  and  D.  E.  Hanson,  Phys.  Rev.  B  85,  233101-­‐1-­‐4  (2012).   7. “Calculations  of  the  thermal  conductivity  of  National  Ignition  Facility  target   materials  at  temperatures  near  10  eV  and  densities  near  10  g/cc  using  finite-­‐ temperature  quantum  molecular  dynamics,”  D.  E.  Hanson,  L.  A.  Collins,  J.  D.   Kress,  and  M.  P.  Desjarlais,  Physics  of  Plasma  18,  082704  (2011).     8. “Transport  properties  of  lithium  hydride  at  extreme  conditions  from  orbital   free  molecular  dynamics,”  L.  Burakovsky,  C.  Ticknor,  J.  D.  Kress,  L.  A.  Collins,   and  F.  Lambert,  to  be  submitted  for  publication,  September,  2012.   9. “Mass  transport  properties  of  LiD-­‐U  mixtures  from  orbital  free  molecular   dynamics  and  a  pressure-­‐matching  mixing  rule,”  L.  Burakovsky,  J.  D.  Kress,  L.   A.  Collins,  LA-­‐UR-­‐12-­‐21806  (2012).   10.  “First  principles  non-­‐equilibrium  plasma  mixing,”  C.  Ticknor,  S.  D.  Herring,  F.   Lambert,  L.  A.  Collins,  and  J.  D.  Kress,  to  be  submitted  for  publication,   October,  2012.   11. “Average  atom  transport  properties  for  pure  and  mixed  species  in  the  hot   and  warm  dense  matter  regime,”  C.  E.  Starrett,  J.  Clerouin,  V.  Recoules,  J.  D.   Kress,  L.  A.  Collins,  D.  E.  Hanson,  Physics  of  Plasmas,  submitted  for   publication,  July,  2012.   12. “The  Quantum  Hypernetted  Chain  Model  of  Warm  Dense  Matter,”  D.  Saumon,   C.  E.  Starrett,  J.  D.  Kress,  J.  Clerouin,  High  Energy  Density  Physics  8,  150-­‐153   (2012).           4   FY12  Publications,  Invited  Presentations,  and  Unpublished  Reports   Publications   1.    “The  Quantum  Hypernetted  Chain  Model  of  Warm  Dense  Matter,”  D.  Saumon,  C.  E.   Starrett,  J.  D.  Kress,  J.  Clerouin,  High  Energy  Density  Physics  8,  150-­‐153  (2012).   2.    “Average  atom  transport  properties  for  pure  and  mixed  species  in  the  hot  and   warm  dense  plasma  regimes,”  C.  E.  Starrett,  J.  Clerouin,  V.  Recoules,  J.  D.  Kress,  L.  A.   Collins,  and  D.  E.  Hanson,  Physics  of  Plasmas,  submitted  for  publication.   3.    “Transport  properties  of  lithium  hydride  at  extreme  conditions  from  orbital  free   molecular  dynamics,”  L.  Burakovsky,  C.  Ticknor,  J.  D.  Kress,  L.  A.  Collins,  and  F.   Lambert,  to  be  submitted  for  publication,  September,  2012.   4.  “First  principles  non-­‐equilibrium  plasma  mixing,”  C.  Ticknor,  S.  D.  Herring,  F.   Lambert,  L.  A.  Collins,  and  J.  D.  Kress,  to  be  submitted  for  publication,  October,  2012.       Invited  Presentations   1.  “Mixtures  in  the  Warm,  Dense  Matter  Regime  (U),”  J.  D.  Kress,  L.  Burakovsky  ,  D.   Herring,  C.  Ticknor,  L.  A.  Collins,  R.  A.  Gore,  Nuclear  Explosives  Design  Physics   Conference  (NEDPC),  Los  Alamos,  NM,  October  17-­‐21,  2011.     2.“Mass  Transport  in  the  Warm,  Dense  Matter  and  High-­‐Energy  Density  Regimes,”  J.   D.  Kress,  L.  Burakovsky,  C.  Ticknor,  L.  A.  Collins,  and  F.  Lambert,  American  Physical   Society,  Division  of  Plasma  Physics  meeting,  Salt  Lake  City,  UT,  Nov.  14-­‐18,  2011.       3.  “Mass  Transport  in  the  Warm,  Dense  Matter  and  High-­‐Energy  Density  Regime,”  J.   D.  Kress,  Computational  Challenges  in  Warm,  Dense  Matter  Workshop,  Institute  for   Pure  and  Applied  Math  October,  UCLA,  Los  Angeles,  CA,  May  21-­‐15,  2012.   4.  CECAM  Workshop  on  Orbital-­‐Free  Methods,  Paris  Sept.  5-­‐7,  2012,  L.A.  Collins,  co-­‐ organizer;  “Quantum  Simulations  of  Warm  Dense  Matter”,  plenary  talk.   5.“Interface  Dynamics  in  the  Warm  Dense  Matter  Regime,”  Topic  Talk,  L.A.  Collins,   J.D.  Kress,  C.  Ticknor,  S.  Herring,  L.Burakovsky,  F.  Lambert,  Physics  of  Non-­‐Ideal   Plasmas,  Rostock,  Germany,  Sept.9-­‐14,  2012.     Unpublished  Report     1.“Mass  Transport  Transport  Properties  of  LiD-­‐U  Mixtures  from  Orbital  Free   Molecular  Dynamics  Simulations  and  a  Pressure-­‐Matching  Mixing  Rule,”  L.   Burakovsky,  J.  Kress,  and  L.  Collins,  LA-­‐UR-­‐12-­‐21806,  June  25,  2012.     5   2.  Shock  Hugoniot  for  Li6D.  Equations  of  state  for  Lithium  Hydride  (LiH,  7.941   g/mol)  and  Lithium  6  Deuteride  (Li6D,  8  g/mol)  have  been  calculated  with  both   Kohn-­‐Sham  (KS)  density  functional  theory  (DFT)  molecular  dynamics  (MD)  and   with  Orbital-­‐Free  (OF)  DFT  MD.    The  shock  Hugoniot  for  Li6D  has  been  determined   for  temperatures  between  0.5  and  25  eV  using  the  KSMD  EOS  and  for  temperatures   of  25  eV  and  above  using  the  OFMD  EOS  [1].    KSMD  simulations  here  have  a   practical  upper  limit  of  about  T=25  eV,  since  the  computational  work  scales  as  the   square  of  the  number  occupied  bands  in  the  electron  density  expansion.    On  the   other  hand,  the  OFMD  simulations  have  a  lower  limit  of  about  T=5  eV  since  the   Thomas-­‐Fermi  kinetic  energy  plus  DFT  exchange  and  correlation  energy  provides  a   relatively  poor  description  of  the  electronic  structure  of  condensed-­‐matter  bound   solids.    This  provides  an  overlap  in  temperature  of  about  20  eV  between  the  two   quantum  MD  approaches.    (See  Horner  et  al.  [2]  for  a  description  of  the  KSMD  and   OFMD  techniques  and  for  comparisons  in  LiH  pressures  for  temperatures  between  2   and  6  eV  and  compressions  between  1  and  4  relative  to  ambient  solid.)   In  Fig.  1  we  present  the  shock  Hugoniot  for  Li6D  as  calculated  with  the  KSMD  and   OFMD  EOS  [1].    To  determine  the  zero  of  internal  energy  for  the  KSMD  Hugoniot,  a   standard  KSMD  calculation  for  LiH  solid  at  T=300K  was  performed.    Since  the  LiH   solid  at  T=300K  is  unbound,  or  nearly  so,  at  the  OF  level  of  DFT,  the  zero  of  internal   energy  for  the  OFMD  Hugoniot  was  determined  instead  by  matching  to  the  nuclear   impedence  match  (NIM)  pressure  [3]  for  a  compression  C=ρ/ρ0=3.58.    In  Fig.  1,  we   also  compare  with  experimental  data  from  the  LASL  Shock  compendium  [4]  and   with  shock  Hugoniots  constructed  from  the  7247  and  7245  SESAME  EOS  tables  and   from  a  new,  improved  SESAME  table.      7245  is  older  [5],  dating  back  to  1985,   whereas  7247  is  a  newer  baseline  EOS  [6]  containing  physics  “we  know  and  can   manipulate”.    A  brand  new  table  was  created  using  the  KSMD  and  OFMD  results   generated  over  a  wide  range  of  densities  and  temperatures.    Starting  at  low   pressure  (50  GPa  and  below)  the  7247  and  7245  are  indistinguishable  from  each   other  and  both  go  through  the  LASL  data  to  within  experimental  error  (except  at  77   and  86  GPa).    At  compressions  above  2.5,  the  7245  Hugoniot  is  noticeably  softer   than  7247.  The  7245  Hugoniot  agrees  with  the  three  QMD  results  at  C=2.5,  3.0,  and   3.5  and  with  the  NIM  experimental  point  [3]  at  C=3.58.    At  C=3.5,  the  difference   between  QMD  and  OFMD  is  P=1430  vs.  988  GPa  and  T=8.6  vs.  5.3  eV.    Continuing  to   higher  pressure,  7245  agrees  well  with  the  OFMD  results  up  until  the  end  of  the   7245  table  is  reached  at  about  1000  GPa.    The  OFMD  results  reach  a  maximum   compression  of  4.26  at  T=100eV  and  P=22,000  GPa  as  the  Hugoniot  continues  on  to   higher  pressures.    The  new  table  agrees  well  with  all  of  the  experimental  data  and   with  the  OFMD  results  all  the  way  up  to  the  last  point  at  P=230,000  GPa.     6     Figure  1.  Shock  Hugoniot  for  Li6D.       The  new  EOS  table  was  calculated  in  a  similar  way  as  7247.      The  new  EOS  is  based   on  a  three-­‐component  free  energy.  The  three-­‐components  consist  of  a  cold  curve,  an   ion  thermal,  and  an  electron  thermal  part.    The  electron  thermal  part  is  simply  an   additive  volume  mixture  of  the  total  TFD  (Thomas-­‐Fermi-­‐Dirac)  electron   contribution  of  each  element.    The  additive  volume  mixture  procedure  assumes   pressure-­‐temperature  equilibrium.  After  the  mixture  is  performed,  the  zero   temperature  component  is  subtracted  off  leaving  only  the  thermal  contribution.   Compared  to  the  older  EOS,  the  cold  curve  model  was  simplified.    The  cold  curve   was  inferred  by  using  Hugoniot  data  and  a  Mie-­‐Gruneisen  approximation.  In  the   previous  work  the  DAC  (diamond  anvil  cell)  data  was  used  directly  to  infer  the  cold   curve.  To  no  surprise,  the  two  methods  produced  almost  identical  results.    The  ion     7   model  used  the  same  parameters  for  the  Debye  model  as  7247  with  only  one   exception:  a  change  to  the  derivative  of  the  Gruneisen  γ  with  respect  to  density.    In   the  case  of  7247,  the  pressure  contribution  from  the  ion  model  was  too  strong  due   to  having  a  nearly  constant  γ  with  respect  to  compression.  By  utilizing  the  KSMD  and   OFMD  results  to  guide  our  parameter  choices  for  the  ion  model,  the  new  EOS  overall   is  a  much  better  match  to  melt  and  NIM  shock  data.   In  conclusion,  the  shock  Hugoniot  constructed  with  the  7245  SESAME  EOS  is  more   consistent  with  the  KSMD  and  OFMD  results  as  compared  to  the  more  modern  7247   EOS.    A  new,  improved  SESAME  table  has  been  created  that  incorporates  the  results   of  the  KSMD  and  OFMD  simulations.  This  work  was  performed  in  collaboration  with   M.  P.  Desjarlais  at  Sandia  National  Laboratories.    We  thank  Carl  Greeff  and  Rob  Gore   for  stimulating  discussions  on  these  matters.   References     1. L.  Burakovsky,  D.  A.  Horner,  J.  D.  Kress,  and  L.  A.  Collins,  “Final  Report:     Quantum  Molecular  Dynamics  Simulations  of  Transport  Properties  in  LiH,”   unpublished  (October,  2009);  L.  Burakovsky,  J.  D.  Kress,  L.  A.  Collins,  and  F.   Lambert,  “Transport  properties  of  lithium  hydride  at  extreme  conditions   from  orbital  free  molecular  dynamics,”  submitted  for  publication.   2. D.  A.  Horner,  F.  Lambert,  J.  D.  Kress,  and  L.  A.  Collins,  Phys.  Rev.B  80,  024305   (2009).   3. C.  E.  Ragan  III,  Phys.  Rev.  A  29,  3  (1984).   4. S.  P.  Marsh,  “LASL  Shock  Hugoniot  Data,”  U.  of  Cal.  Press  (1980).   5. S.  P.  Lyon  and  J.  D.  Johnson,  “The  Los  Alamos  National  Laboratory  Equation  of   State  Database,”  LA-­‐UR-­‐92-­‐3407  (1992).   6. S.  Crockett,  “Analysis  of  New  Lithium  Hydride  and  Lithium  Deuteride   Equations  of  State,”  unpublished,  LA-­‐UR-­‐06-­‐8403  (2006).     8   4.  Non-­Equilibrium  Quantum  Molecular  Dynamics  Simulations  of  Carbon/   Deuterium-­Tritium  Plasma  Interfaces.    The  dynamics  across  interfaces   separating  materials  at  different  conditions  play  important  roles  in  determining  the   basic  physical  state  at  boundary  zones  found  between,  for  example,  the  core  and   atmosphere  of  planets  and  stars  and  between  the  ablator  and  fuel  in  inertial   confinement  fusion  (ICF)  capsules.  In  the  ICF  case,  the  interpenetration  of  the   ablator  into  the  DT  fuel  can  have  significant  effects  on  the  burn  and  yield  of  the   resulting  fusion  reaction.  Although  the  interface  has  macroscopic  extent,  the  basic   dynamical  processes  involve  the  microscopic  motion  of  one  species  into  another   under  non-­‐equilibrium  and  highly  transient  conditions.  In  addition,  the  medium   resides  in  the  warm  dense  matter  regime  characterized  by  a  highly-­‐compressed,   elevated  temperature  state  of  transient  populations  of  electrons  and  partially-­‐ ionized  atoms  that  requires  to  some  degree  a  quantum  mechanical  treatment.  As  an   example  of  the  complexity  of  the  boundary  dynamics,  we  display  in  Fig.1  three   representative  snap  shots  for  the  mixing  of  hydrogen  and  tritium  initially  at   different  densities  separated  by  an  interface.     In  order  to  simulate  the  dynamics  at  such  an  interface,  we  have  developed  a  large-­‐ scale  computational  simulation  technique  based  on  an  Orbital-­‐Free  Density   Functional  Theory  molecular  dynamics  (OFMD)  approach,  which  treats  the   electrons  quantum  mechanically  and  the  ions  classically.  The  method  accurately   determines  both  static  (equation-­‐of-­‐state)  and  dynamical  (diffusion/viscosity)   properties  for  both  pure  and  mixed  systems  across  a  broad  temperature  and  density   range  from  warm  fluids  to  hot,  dense  plasmas.      The  interface  poses  new  challenges   in  that  its  non-­‐equilibrium  nature  requires  much  larger  atomic  samples  than  for   extracting  equilibrium  properties  such  as  pressure  or  viscosity.  However,  the  advent   of  new  parallel  algorithms  and  the  access  to  very  large  distributed-­‐processor   platforms  such  as  Cielo  permit  the  effective  modeling  of  such  physical   environments.    As  an  illustrative  example  with  applications  to  ICF,  we  present   results  for  carbon  (C)  and  deuterium-­‐Tritium  (DT)  initially  separated  at  a  boundary   and  explicitly  examine  the  rate  at  which  the  atomic  species  intermix  and  the  clean   interface  deteriorates.    This  simulation  resembles  the  ICF  case  in  which  carbon   (plastic  or  diamond)  interpenetrates  the  DT  fuel  and  in  turn  reduces  the  fusion   yield.       9     Fig. 1. An example of H/T mixing. From these simulations mixing rates and diffusion coefficients can be extracted. The C/DT simulations are much larger. The  non-­‐equilibrium  simulations  of  the  interface  uses  4000  DT  atoms  and  1132  C   atoms  at  a  total  density  of  2.94  g/cc  (2.5g/cc  for  the  DT  and  3.4  g/cc  for  the  C)  in  a   simulation  box  length  of  16  nm  with  a  1  nm  by  1  nm  cross  section.    We  have  also   simulated  a  box  with  a  cross  section  of  2  nm  by  2  nm  with  a  simulation  containing   over  10,200  particles.    We  have  conducted  simulations  at  many  temperatures:  10,   40,  60,  100,  and  1000  eV.    From  these  OFMD  simulations  we  are  able  to  extract  the   mixing  rates  and  study  the  diffusive  behavior  of  the  interface.  The  high  temperature   cases  show  the  ballistic  increase  in  width  while  for  the  lower  temperatures,  the   width  saturates,  indicative  of  collision-­‐dominated  diffusion  behavior.    From  this   behavior  we  can  extract  out  the  diffusion  coefficients  and  compare  them  to  diffusion   coefficients  from  an  equilibrium  mixed  simulation.    The  agreement  is  good.     Additionally,  we  have  conducted  classical  Yukawa  MD  simulations  at  similar   conditions  and  found  long  time  agreement  in  the  mixing  rates.  However,  at  short   time  there  is  noticeable  disagreement;  this  is  an  area  of  future  study.    The  quantum   simulations  reach  unprecedented  size,  for  example,  2  to  3  days  CPU  time  on  16384   cores  on  Cielo  for  a  typical  interface  simulation.  The  calculations  were  performed   under  the  Capability  Computing  Campaign  (CCC)-­‐2  and  -­‐3.      C.  Ticknor  was  partially   supported  by  ASC  through  a  Metropolis  Postdoctoral  Fellowship.       10     6.  Deuerium  Mass  Transport  for  ρ  =  2  g/cm3.    Mass  transport  properties  (shear   viscosity  and  self-­‐diffusion  coefficient)  for  deuterium  plasmas  for  ρ  =  2  g/cm3  have   been  calculated  using  orbital-­‐free  molecular  dynamics  (OFMD)  [1]  and  compared   with  the  results  evaluated  using  Eqs.  (17)  and  (21)  in  Vold  [2].  Further  comparisons   are  made  with  Landau  theory  [3]  evaluated  with  two  popular  parameterizations  [3,   4]  for  the  Coulomb  logarithm,  ln(Λ).    For  T=10  eV,  the  temperature  is  too  low   (plasma  coupling  constant  Γ  is  too  large)  for  a  reliable  evaluation  with  ln(Λ).    In   Table  I,  for  T=100  eV  and  above,  the  OFMD  viscosity  is  2  to  4  times  larger  than  the   Vold  values.  Note,  however,  the  size  of  the  OFMD  statistical  error  for  T=333  and   1000  eV.    In  Table  II,  for  T=100  eV  and  above,  the  Landau  viscosity  evaluated  with   both  the  Dimonte-­‐Daligault  (DD)  [4]  and  Clerouin-­‐Cherfi-­‐Zerah  (CCZ)  [3]   parameterization  for  the  Coulomb  logarithm,  ln(Λ)  yield  consistent  shear  viscosities   (within  10%  or  better).    In  a  comparison  between  the  Vold  values  and  the  Landau   CCZ  results,  at  T=100  eV,  the  CCZ  is  about  30%  larger  than  Vold;  at  T=333  eV,  the   CCZ  value  is  twice  as  large  as  Vold;  and  at  T=1000  eV,  CCZ  and  Vold  agree  quite  well.     This  inconsistency  in  temperature  trends  between  the  Vold  and  the  Landau  results   deserves  further  study  and  analysis.     References   1. J.  D.  Kress,  J.  S.  Cohen,  D.  A.  Horner,  F.  Lambert,  and  L.  A.  Collins,  Phys.  Rev.  E   82,  036404  (2010).   2. E.  L.  Vold,  “Momentum  transport  and  associated  scale  lengths  in  an  ICF   plasma,”  unpublished  draft  report.    The  shear  viscosity  values  from  Table  1  in   this  report  have  been  multiplied  by  2.2  (E.  Vold,  private  communication).   3. Clerouin,  Cherfi,  and  Zerah,  Europhys.  Lett.  42,  37  (1998).   4. Dimonte  and  Daligault,  Phys.  Rev.  Lett.  101,  065003  (2008).     11     7.  Shear  Viscosity  in  Liquid  Copper.    November  15,  Guy  Dimonte  communicated   his  interest  in  the  shear  viscosity  of  Cu  (and  other  fcc  metals)  below  and  above  melt   temperatures.  Directing  him  to  an  excellent  review  paper  [1],  we  recommended  the   use  of  the  Arrhenius  and  Andrade  equation  to  represent  the  temperature  and   density  dependence,  respectively.    Quantum  Molecular  Dynamics  (QMD)   simulations  were  then  performed  with  the  Vienna  ab  initio  Simulation  Package   (VASP)[2].  VASP  provides  a  fully  quantum  mechanical  treatment  of  the  electrons  by   employing  a  plane-­‐wave,  Finite  Temperature  Density  Functional  Theory  (FTDFT)   solution  to  the  Schroedinger  equation.      See  our  publication  on  Pu  [3]  for  details  on   the  QMD  approach.    A  total  Cu  mass  density  of  ρ=8  g/cm3  was  simulated  using  54  Cu   atoms  in  a  cubic  periodic  box.  The  shear  viscosity  (η  )  and  self-­‐diffusion  coefficient   (D),  presented  in  Table  I,  was  calculated  from  the  appropriate  autocorrelation   functions  constructed  from  the  MD  trajectories.      In  Fig.  1  we  present  an  Arrhenius   plot  of  the  shear  viscosity,  ln(η)  vs.  1/T.      The  results  in  Fig.  1  are  linear,  and  thus,   are  described  well  by  η  = η  0  exp(Ea/RT),  the  Arrhenius  equation,  where  η  0  =  0.91   mPa-­‐sec,  Ea=  4  kcal/mol,  R=gas  constant.    The  activation  energy  Ea  is  about  40%   and  80%  smaller  that  the  two  different  experimental  values,  5.7  and  7.3  kcal/mol,   cited  in  Ref.  [1].    However,  extrapolation  of  the  QMD  result  to  the  experimental   melting  temperature  of  1356  K  yielded  η(1356K)=4.0  mPa-­‐sec,  in  excellent   agreement  with  the  experimental  values,  4.38  and  4.0  mPa-­‐sec,  cited  in  Ref.  [1].     Table  I.    QMD  simulations  of  Cu  shear  viscosity  and  self-­‐diffusion  for  ρ=8  g/cm3.       Autocorrelation  functions  fit  to  a  Gaussian.    Relative  statistical  error  is  given.   T(K)     η  (mPa-­‐sec)   error   D  (cm2/sec)   2000     2.48     5%   4.7x10-­‐5     2500     2.08     2%   7.2x10-­‐5   3000     1.76     4%   9.3x10-­‐5     3500     1.63     3%   10.4x10-­‐5     4000     1.50     3%   11.0x10-­‐5         12     Figure  1.  Arrhenius  plot  for  QMD  simulations  of  Cu  viscosity  for  ρ=8.0  g/cm3  .   References   1.  R.  F.  Brooks,  A.  Dinsdale,  and  P.  N.  Quested,  Meas.  Sci.  Tech.  16,  354  (2005).   2.  G.  Kresse  and  J.  Hafner,  Phys.  Rev.  B  47,  558  (1993).   3.  J.  D.  Kress,  J.  S.  Cohen,  D.  P.  Kilcrease,  D.  A.  Horner,  and  L.  A.  Collins,  Phys.   Rev.  E  83,  026404  (2011).       13  

References (36)

  1. "Quantum molecular dynamics simulations of warm dense lithium hydride: Examination of mixing rules," D. A. Horner, J. D. Kress, L. A. Collins, Phys. Rev. B 77, 064102--1--9 (2008).
  2. "Transport properties of lithium hydride from quantum molecular dynamics and orbital free molecular dynamics," D. A. Horner, F. Lambert, J. D. Kress, L. A. Collins, Phys. Rev. B 80, 024305--1--10 (2009).
  3. "Viscosity and mutual diffusion of deuterium--tritium mixtures in the warm, dense matter regime," D. A. Horner, J. S. Cohen, J. D. Kress, L. A. Collins, Phys. Rev. E 82, 036404--1--10 (2010).
  4. "Quantum molecular dynamics simulations of transport properties in liquid and dense--plasma plutonium," J. D. Kress, J. S. Cohen, D. P. Kilcrease, D. A. Horner, L. A. Collins, Phys. Rev. E 83, 026404 (2011).
  5. "Orbital--free molecular dynamics simulations of transport properties in dense--plasma uranium," J. D. Kress, J. S. Cohen, D. P. Kilcrease, D. A. Horner, D. A. Horner, L. A. Collins, High Energy Density Physics 7, 155 (2011).
  6. "Reflectivity of warm dense deuterium along the principal Hugoniot," L. A. Collins, J. D. Kress, and D. E. Hanson, Phys. Rev. B 85, 233101--1--4 (2012).
  7. "Calculations of the thermal conductivity of National Ignition Facility target materials at temperatures near 10 eV and densities near 10 g/cc using finite-- temperature quantum molecular dynamics," D. E. Hanson, L. A. Collins, J. D. Kress, and M. P. Desjarlais, Physics of Plasma 18, 082704 (2011).
  8. "Transport properties of lithium hydride at extreme conditions from orbital free molecular dynamics," L. Burakovsky, C. Ticknor, J. D. Kress, L. A. Collins, and F. Lambert, to be submitted for publication, September, 2012.
  9. "Mass transport properties of LiD--U mixtures from orbital free molecular dynamics and a pressure--matching mixing rule," L. Burakovsky, J. D. Kress, L. A. Collins, LA--UR--12--21806 (2012).
  10. "First principles non--equilibrium plasma mixing," C. Ticknor, S. D. Herring, F. Lambert, L. A. Collins, and J. D. Kress, to be submitted for publication, October, 2012.
  11. "Average atom transport properties for pure and mixed species in the hot and warm dense matter regime," C. E. Starrett, J. Clerouin, V. Recoules, J. D. Kress, L. A. Collins, D. E. Hanson, Physics of Plasmas, submitted for publication, July, 2012.
  12. "The Quantum Hypernetted Chain Model of Warm Dense Matter," D. Saumon, C. E. Starrett, J. D. Kress, J. Clerouin, High Energy Density Physics 8, 150--153 (2012).
  13. "The Quantum Hypernetted Chain Model of Warm Dense Matter," D. Saumon, C. E. Starrett, J. D. Kress, J. Clerouin, High Energy Density Physics 8, 150--153 (2012).
  14. "Average atom transport properties for pure and mixed species in the hot and warm dense plasma regimes," C. E. Starrett, J. Clerouin, V. Recoules, J. D. Kress, L. A. Collins, and D. E. Hanson, Physics of Plasmas, submitted for publication.
  15. "Transport properties of lithium hydride at extreme conditions from orbital free molecular dynamics," L. Burakovsky, C. Ticknor, J. D. Kress, L. A. Collins, and F. Lambert, to be submitted for publication, September, 2012.
  16. "First principles non--equilibrium plasma mixing," C. Ticknor, S. D. Herring, F. Lambert, L. A. Collins, and J. D. Kress, to be submitted for publication, October, 2012. Invited Presentations
  17. "Mixtures in the Warm, Dense Matter Regime (U)," J. D. Kress, L. Burakovsky , D. Herring, C. Ticknor, L. A. Collins, R. A. Gore, Nuclear Explosives Design Physics Conference (NEDPC), Los Alamos, NM, October 17--21, 2011.
  18. "Mass Transport in the Warm, Dense Matter and High--Energy Density Regimes," J. D. Kress, L. Burakovsky, C. Ticknor, L. A. Collins, and F. Lambert, American Physical Society, Division of Plasma Physics meeting, Salt Lake City, UT, Nov. 14--18, 2011.
  19. "Mass Transport in the Warm, Dense Matter and High--Energy Density Regime," J. D. Kress, Computational Challenges in Warm, Dense Matter Workshop, Institute for Pure and Applied Math October, UCLA, Los Angeles, CA, May 21--15, 2012.
  20. CECAM Workshop on Orbital--Free Methods, Paris Sept. 5--7, 2012, L.A. Collins, co-- organizer; "Quantum Simulations of Warm Dense Matter", plenary talk.
  21. "Interface Dynamics in the Warm Dense Matter Regime," Topic Talk, L.A. Collins, J.D. Kress, C. Ticknor, S. Herring, L.Burakovsky, F. Lambert, Physics of Non--Ideal Plasmas, Rostock, Germany, Sept.9--14, 2012. Unpublished Report
  22. "Mass Transport Transport Properties of LiD--U Mixtures from Orbital Free Molecular Dynamics Simulations and a Pressure--Matching Mixing Rule," L. Burakovsky, J. Kress, and L. Collins, LA--UR--12--21806, June 25, 2012. References
  23. L. Burakovsky, D. A. Horner, J. D. Kress, and L. A. Collins, "Final Report: Quantum Molecular Dynamics Simulations of Transport Properties in LiH," unpublished (October, 2009);
  24. L. Burakovsky, J. D. Kress, L. A. Collins, and F. Lambert, "Transport properties of lithium hydride at extreme conditions from orbital free molecular dynamics," submitted for publication.
  25. D. A. Horner, F. Lambert, J. D. Kress, and L. A. Collins, Phys. Rev.B 80, 024305 (2009).
  26. C. E. Ragan III, Phys. Rev. A 29, 3 (1984).
  27. S. P. Marsh, "LASL Shock Hugoniot Data," U. of Cal. Press (1980).
  28. S. P. Lyon and J. D. Johnson, "The Los Alamos National Laboratory Equation of State Database," LA--UR--92--3407 (1992).
  29. S. Crockett, "Analysis of New Lithium Hydride and Lithium Deuteride Equations of State," unpublished, LA--UR--06--8403 (2006).
  30. J. D. Kress, J. S. Cohen, D. A. Horner, F. Lambert, and L. A. Collins, Phys. Rev. E 82, 036404 (2010).
  31. E. L. Vold, "Momentum transport and associated scale lengths in an ICF plasma," unpublished draft report. The shear viscosity values from Table 1 in this report have been multiplied by 2.2 (E. Vold, private communication).
  32. Clerouin, Cherfi, and Zerah, Europhys. Lett. 42, 37 (1998).
  33. Dimonte and Daligault, Phys. Rev. Lett. 101, 065003 (2008). References
  34. R. F. Brooks, A. Dinsdale, and P. N. Quested, Meas. Sci. Tech. 16, 354 (2005).
  35. G. Kresse and J. Hafner, Phys. Rev. B 47, 558 (1993).
  36. J. D. Kress, J. S. Cohen, D. P. Kilcrease, D. A. Horner, and L. A. Collins, Phys. Rev. E 83, 026404 (2011).
About the author
Los Alamos National Laboratory, Faculty Member

This is me

Papers
294
Followers
21
View all papers from Joel Kressarrow_forward