Sandia National Laboratories is a multi-program laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000.

SAND No. 2011-XXXXP

MultiscaleMolecularSimulationsofPVEncapsulantsLaurenAbbott1,RossLarsen2,andStanMoore1

1SandiaNationalLaboratories,Albuquerque,NM,2NationalRenewableEnergyLaboratory,Golden,CO

• Automatedgenerationofcandidatemoleculesformaterialsdiscovery• Automatedgenerationofinitialconditionsformolecularsimulations• Automatedextractionofsub-regionsforanalysisandcalculationofproperties• Integrationwithanalysisandvisualizationalgorithms• Object-orientedframeworkforeasyextensionofthefunctionalitytospecificneeds

SimulationToolkitforRenewableEnergyandAdvancedMaterialsModeling

streamm.nrel.gov

• Opensourceclassicalmoleculardynamicscode• Particlesimulationsattheatomic,meso,andcontinuumscales• Inclusionsofcommonpotentialsforhardandsoftmaterials• Efficientparallelsimulationsusingspatialdecompositionofdomainspace• Modularframeworkforeasyextensionwithnewfeaturesandfunctionality

Large-scaleAtomic/MolecularMassivelyParallelSimulator

lammps.sandia.gov

Polymer/ClayNanocomposites• SandiaandTexasA&Mareworkingon

inexpensive,transparentpolymer/claynanocompositeswithlayersofpolymersandorientedclayplatelets

• Tailoringtheclayparticlesandpolymermatrixcontrolsthebarrierproperties,compositeintegrity,andfireretardancy

• Chemically-specificcoarse-grainedmodelsofpolymer/claynanocompositescanyieldinsightintomaterialprocesses(e.g.,encapsulation,intercalation,andexfoliation)andresultingmaterialproperties

Clayplatelets,~50-100nmdiam

*Sandia:MargaretGordon,ErikSpoerke,EricSchindelholtz,KenArmijo,RobSorensen,TexasA&M:JamieGrunlan,KevinHolder,Shuang Qin

neatpolymer

composite

Stress-straincurveLeft:Coarse-grainedmodel;Middle:Exampleofpolymer intercalation;Right:Claydispersion inpolymermatrixwithpolymer notshown

*JLSuter,DGroen,PVCoveney,Adv.Mater.,2015,27,966-984;NanoLett.,2015,15,8108-8113

• Atomisticsimulationselucidatedpossibleeffectsofmorphologyonelectron-transportmechanismsinradicalpolymerelectrodes

• MolecularmodelsandMDsimulationsofPTMAweresetupusingSTREAMM

• Twoprimarydistances(4.5and6.5Å)wereidentifiedascontributingtoaneffectiveelectrontransferdistanceof5.5Å

• Inter-sitecouplingswerecomputedbetween>10,000pairsofsiteswithQMcalculations,showingthatelectrontransfermostlytakesplacebetweensitesondifferentchains

RadicalPolymerElectrodes

Top:N-Nradialdistribution functions;Bottom:Averageinter-siteelectroniccoupling

Figure 3: PMMA and PTMA oligomers 12, 24 and 48 monomerunits, (PMMA38PTMA10), (PMMA2PTMA8)5, (PMMA30PTMA18),(PMMA6PTMA4)5, (PMMA10PTMA38), (PMMA8PTMA8)5

10

MD

QM

PTMA

STREAMM’sautomatedstructuregenerationfrommonomertooligomertobulkmaterial,andautomatedextractionofsub-regionsforQMcalculations

*TWKemper,RELarsen,TGennett,J.Phys.Chem.C,2014,118,17213-17220

PolymerPhotovoltaicActiveLayers• AtomisticMDsimulationsweresetupusingSTREAMMandperformedin

LAMMPSforthreeorganicPVcopolymers(BDT-TPD,PTB7,andPTB7-Th),withvariationstothebackboneandside-chainstructure

• AlignedparallelchainswithpistackingwereobservedinBDT-TPDandPTB7,butnotinPTB7-Thduetosterichindrancefromtheside-chain,whichcontradictsassumptionmadeintheliterature

• TransportinBDT-TPDandPTB7likelyoccursbetweenparallelpistacks,whiletransportinPTB7-Thlikelyoccursbetweenorthogonalpistacks

BDT-TPDPTB7

PTB7-Th

Molecularstructures(top)andsimulation snapshots (bottom)fororganicPVcopolymers

*NREL:TravisKemperandRossLarsen

Ion-ContainingPolymers• Sulfonatedpolyphenylenes(SDAPP)showpromiseforprotonexchange

membranefuelcells(PEMFCs)andvanadiumflowbatteries• Littleisknownaboutthenanoscalestructureoftheseamorphouspolymers,

whichisnoteasilycharacterizedwithcurrentexperimentaltechniques• Atomisticsimulationsyieldeduniqueinsightintothemorphologyoftheionic

domainsformedbytheaggregationofwaterandionicgroups,specificallythesize,shape,andconnectivity,aswellasresultingimplicationsforiontransport

• Asmorewaterisaddedintothesystem,theionicdomainsbecomemorefullypercolatedandthedomainsbecomeslightlylargerandmoresphericalinshape,whichwouldimproveiontransport

increasingwater

Left:MolecularstructureofSDAPP;Right:Ionicdomainswithdisparateclustersshown indifferentcolors

impedance composed of a real component, Z′, and an imaginarycomponent, Z′′. An example of the raw data is shown in Figure3. To compute the membrane proton conductivity from thecomplex impedance response, the impedance line is extrapo-lated to the x-axis. The extrapolated value of the real imped-ance where the imaginary response is zero (Z′ at Z′′ ) 0) isthen taken as the resistance of the membrane, and eq 3 is usedto compute the membrane proton conductivity

where L is the length between the sense electrodes, Z′ is thereal part of the impedance response (extrapolated to Z′′ ) 0),and A is the area available for proton conduction (width ×thickness). All proton conductivities reported here were mea-sured with the film immersed in liquid water at 30 °C duringthe measurement time.

The tensile stress-strain properties of the SDAPP3 mem-branes were compared with Nafion 117 in order to evaluatethe mechanical strength of a typical sulfonated poly(phenylene)vs the state-of-the-art membrane. The samples were testedusing a Com-Ten Industries 95T series load frame equippedwith a 200 lbf load cell and computerized data acquisitionsoftware. Samples of 9 mm width were deformed at a cross-head speed of 5 mm/min with gauge length of 30 mm. Reportedstress-strain measurements are the average of at least threetests.

Results and DiscussionSynthesis and Characterization. Preparation of

Diels-Alder poly(phenylene)s has been described else-where16 and briefly described above (vide supra). Sul-fonation of SDAPP0 has not been reported, but substi-

tuted 1,4-poly(phenylene) has been sulfonated withconcentrated sulfuric acid with levels of sulfonationcontrolled by varying the reaction time.11 Our approachto sulfonating SDAPP0 was the creation of a 6 wt %polymer solution of SDAPP0 in methylene chloride thatwas subsequently homogeneously sulfonated in situwith chlorosulfonic acid.

The repeat unit of SDAPP0 with its six pendentphenyl groups (Scheme 2) provides a number of possiblesites for sulfonation. However, sulfonation is thoughtto occur predominantly at the para positions of thependant phenyl groups due to their positioning17 aboutthe sterically congested, concoplanar, rigid-rod backbone(Figure 1). Therefore, the limits of sulfonation areprojected to be between 0 and 6 sulfonic acid groups perrepeat unit. By varying the ratio of moles of chlorosul-fonic acid to moles of polymer repeat unit charged tothe sulfonation reaction, 0.8-2.1 sulfonic acid groupsper repeat unit were achieved as measured by titration(see below).

Introduction of increasing numbers of sulfonic acidgroups onto polymers improves their ionic conductivity

Figure 2. Schematic of four-point membrane proton conductivity cell.

Figure 3. Impedance response of a typical proton conductingmembrane between 100 kHz and 100 Hz.

κ ) LZ′A

(3)

Scheme 2. Sulfonation of Diels-AlderPoly(phenylene)

5012 Fujimoto et al. Macromolecules, Vol. 38, No. 12, 2005

SDAPP

*Sandia:LaurenAbbott, AmalieFrischknecht,CyFujimoto,ToddAlam,EricSorte

• MolecularsimulationscanprovideimportantinformationaboutnanoscalestructureandphenomenainPVencapsulantmaterialswithatomic-leveldetailnotcurrentlyaccessibletoexperimentaltechniques

• Theatomic-leveldetailofmolecularsimulationsareusefulforstudyinglocalpackingofthepolymer,aswellastheinteractionsanddynamicsofadditives(e.g.,UVabsorbers)andcontaminants(e.g.,H2OorO2)withinthepolymer

• Molecularsimulationsareacost-effectiverouteforefficienthigh-throughputscreeningofpotentialencapsulantmaterials

• Amultiscaleapproachcombinesquantummechanical(QM)calculationsandatomisticandcoarse-grainedmoleculardynamics(MD)simulationstocaptureabroadrangeoflengthandtimescales

• Resultsfromthemolecularsimulationscanbepassedtosimulationmodelsathigherlengthsscales,suchasmeso orcontinuumtechniques

• ThiscapabilityleveragesopensourcetoolslikeSandia’sLAMMPSMDpackageandNREL’sSTREAMMtoolkittosetup,run,andanalyzemolecularsimulations

Overview• MDtechniquesfollowclassicaldynamicsusingNewton’slaw(F =ma),by

updatingthepositionsofparticlesusingthenetforcesontheparticles• Potentialsaredescribedusingforcefieldswithtermsforbondedinteractions

(e.g.,bondsandangles)andnonbondedinteractions(e.g.,vanderWaals)

MolecularModeling

• QMtechniquesapproximatethewavefunctionfollowingSchrödinger’sequation(Hψ =iħ ∂ψ/∂t)toconsiderquantumeffects

• Wecanmovebetweendifferentscalesusingfine-grainingorcoarse-grainingtechniquestocapturedifferentfeaturesandphenomena

• Chemically-specificcoarse-grainedmodelscanbederivedfromatomisticmodelswithtechniqueslikeiterativeBoltzmanninversionandforcematching

U =Ubond +Uangle +Udihedral +Unonbond