ContentslistsavailableatScienceDirect
Chemical
Physics
Letters
j o ur na l h o me p a g e :w w w . e l s e v i e r . c o m / l o c a t e / c p l e t t
Investigation
of
structural
and
dynamical
properties
of
hafnium(IV)
ion
in
liquid
ammonia:
An
ab
initio
QM/MM
molecular
dynamics
simulation
Suwardi
1,
Harno
Dwi
Pranowo,
Ria
Armunanto
∗DepartmentofChemistry,FacultyofMathematicsandNaturalSciences,Austrian–IndonesianCentre(AIC)forComputationalChemistry, GadjahMadaUniversity,Indonesia
a
r
t
i
c
l
e
i
n
f
o
Articlehistory: Received4June2015 Infinalform17July2015 Availableonline29July2015
a
b
s
t
r
a
c
t
ThestructureanddynamicsofHf4+ioninliquidammoniahavebeeninvestigatedbyanabinitio
quan-tummechanicsmolecularmechanics(QM/MM)moleculardynamicssimulation.Thestructuraldatawas obtainedintermsofradialdistribution,coordinationnumberandangulardistribution,andthenthe dynamicsinmeanligandresidencetime.TheHf4+ioniscoordinatedbyfiveammoniamoleculesinthe
firstsolvationshellshowingadistortedsquarepyramidalstructurewithanaverageHf4+–Ndistanceof
2.38 ˚A.Noammonialigandwasobservedforexchangeprocessesbetweenthefirstandsecondshells. ©2015ElsevierB.V.Allrightsreserved.
1. Introduction
Interestinstudyingcomplexformationofmetalionshasgrown rapidly.Itisnecessarytounderstandtheinteractionsofmetalions witha ligandinbiochemical, and chemicalprocesses. Research onmetalinteractionswithproteinshasdeveloped intherecent years[1–3].Thequestionthatarisesthenhowtounderstandthe reactivityofthesemetals.Inasolutionsystemreactivityofmetal ionsisaffectedbythecoordinationshellofthemetalions[4].In thiscontext,themoleculardynamicssimulationplaysan impor-tantroletoinvestigatethestructureofsolvationofthemetalion anditsreactivity.Characteristicsofsolvatedionsinwaterorliquid ammoniahavebeenatopicofspecialinterestsincesuchdetailed knowledgeisessentialforunderstandingtheroleoftheseionsin chemicalandbiologicalprocesses[5–7].Liquidammoniahasthe weakesthydrogenbondsinnatureanditsassociatedionsalsoplay anessentialroleinthechemistryoffertilizers,biochemical pro-cessesoftheliver,kidneysandintestines,andavarietyoforganic reactions[8–13].Hafnium(Hf)fromgroup4hasbeenknownas implantmaterialsbutitsbiocompatibilityoflittle-knownone[14]. TheHf4+ioncouldbesolvatedinwaterandalsoinliquidammonia.
Indeed,thestructureinvestigationofHf4+inaqueoussolutionhas
∗Correspondingauthor.
E-mailaddress:ria.armunanto@ugm.ac.id(R.Armunanto).
1 Permanentaddress.DepartmentofChemistryEducation,Facultyof Mathemat-icsandNaturalSciences,YogyakartaStateUniversity,Yogyakarta,Indonesia.
beencarriedoutbythespectroscopyaswellasQuantumMechanics ChargeField(QMCF)methodwhereasthestructuralanddynamical aspectfortheHf4+ioninliquidammoniahasnotbeenreportedso
far[15–21].However,thereisnomanyinvestigationsofmetalions
inliquidammonia.Therefore,itisstillinterestingtoinvestigatethe solvationofHf4+inliquidammonia.
The pentaammoniates of MF4, namely M(NH3)4F4·NH3 (1,
M=Zr; 2,M=Hf)are formedifAg3M2F14 and liquidNH3 were
prolongedinthestorage.Compounds1and2arealsoformedin thereactionoftheMF4withliquidNH3inweeks.TheHf4+–N
dis-tanceof2.383 ˚AhasbeenknownbytheX-raydiffractionmethod
[22].
ThehydrationstructureofsometetravalentionsasU4+,Th4+,
Hf4+,andZr4+hasbeenextensivelyinvestigatedbythediffraction,
spectroscopy,andcomputersimulationmethods[23–27].Hybrid quantum mechanics/molecular mechanics molecular (QM/MM MD)dynamicssimulationmethodshavesuccessfullyinvestigated structuralanddynamicalpropertiesofmetalionsinwateraswell asinliquidammonia.Thesimulationmethodhasthusbecomean alternativetoexperiments,in particularwheretheexperiments reachtheirlimitations[15–17].Determinationsofthestructureand dynamicspropertiesofsolvationionsareverysensitivetothe accu-racyofthesimulationmethods.Itisknownthattheinvestigations ofthesolvationofmetalionsbyQM/MMMDmethodsprovidea highlevelofaccuracy.Therefore,inthepresentwork,weperformed aQM/MMMDsimulationforHf4+inliquidammonia,inorderto
investigatethestructureanddynamicsofthesolvatedHf4+ionin
liquidammoniaat235.15K.
Table1
Hafnium(IV)–ammoniadistances(r),bindingenergiesperammonialigandobtained byquantummechanicalcalculationsofHf4+–ammoniacomplexatHF,MP2,CCSD, andB3LYPlevels.
Method r Bindingenergy
(Å) (kcal/mol)
HF 2.24 −152.06
MP2 2.25 −160.75
CCSD 2.25 −158.95
B3LYP 2.21 −167.51
2. Methods
2.1. Constructionofpotentialfunction
TherearetwostepsthatmustbedonepriortoQM/MM sim-ulationcouldberun, thatis selecting theproperbasis sets for Hf4+,NandHatomsandcalculationmethodappliedinthe quan-tummechanicszone.Accordingtotheliterature,DZPbasis sets forHandNatomscouldbeappliedsuccessfullyandhavebeen selected,therefore,alsointhisinvestigation[15,28].TheLANL2DZ ECPbasissetsofHfwithaminormodification(sandpbasis func-tionswiththe smallestexponent have beenremoved to make thebasissetsmorecompatiblewiththeHf4+ionratherthanthe
hafniumatom)wasselectedforHf4+,includingtherelativistically
inordertobecompatiblewithhafnium(IV)ion[29].Thelevelsof theoryforQMregionnamely,Hartree–Fock(HF),MP2,CCSD,and B3LYPwereappliedinenergycalculationusingGaussian09 pro-gram,optimizingthegeometryofHf4+–NH
3complex.Theresult
showsthatHFcalculatedenergyclosestothemostcorrelatedCCSD
asseeninTable1.Inrecentinvestigationsofalikeionicsystems,
resultsofHFcalculationswereingoodagreementwith experimen-taldata,whereaselectroncorrelationmethodssuchasMP2and CCSDseemedtohavemoreexpensiveandconsumingtime,and eventhecurrenthybridB3LYPfunctionalsometimesprovideapoor resultoratbestresulttakealongcomputationtime[28].Therefore, Hartree–FockmethodwaschosentodescribetheQMpartinthe simulation.
Thetwo-bodyenergies,E2bd,betweenammoniaandHf4+ion wereevaluatedbysubtractingtheabinitio energiesof the iso-latedspeciesEHf4+andENH3 fromthoseofthemonosolvates[30] EHf(NH
3)4+
E2bd=EHf(NH
3)4+−EHf4+−ENH3 (1)
ThenewpairpotentialofHf4+–NH
3 systemwasconstructed.
More than 7600 ab initio energy points were generated at Hartree–FocklevelwiththemodifiedLANL2DZECPbasissetsfor Hf4+andDZPbasissetsforNandHatomsusingTurbomoleprogram
[31–33].Thefollowingpairpotentialfunctionwasconstructedand
usedinthesimulation,
E2bdfit =
qHf4+qN
r +
AN
r5 +
BN
r9 + CN
r11+ DN
r12
+
3
i=1
qHf4+qHri
+AH
r4 +
BH
r5 +
CH
r6 +
DH
r12
(2)
ThefittingparametersofA,B,CandDarelistedinTable2.The qHf4+,qNandqHarethechargeofhafnium,nitrogen,and hydro-gen,respectively,andrandriaretheHf4+–NandHf4+–Hdistances,
respectively.Thenetchargesofnitrogenandhydrogenweresetto
−0.8022and0.2674,respectively.Thegeometryofammoniawas
keptconstantthroughoutthewholecalculationatits experimen-talgasphasevalues(N–H=1.0124 ˚A,H–N–H=106.68◦)[15,16].The
corrected3-bodyenergy(E3bdcorr)wascalculatedasthefollowing formula:
E3bdcorr=
EabAMA−EMab−2EAab
−E2bdMA(r1)−E2bdMA(r2)−EAA2bd(r3) (3)
whereab,2bddenoteabinitioandpairenergy;MAandAAindicated ion–ammoniaandammonia–ammoniainteractions;r1,r2,andr3
areaccordingtothedistanceofion-ammonia(1),ion–ammonia(2), andammonia(1)–ammonia(2),respectively.Thecorrected3-body energy was generated at Restricted Hartree–Fock (RHF). The obtainedthree-bodycorrectionfunctionwas
E3bdFit =0.684e0.447(r1+r2)e−0.233r3(CL−r1)2(CL−r2)2 (4)
wherer1 andr2 arethedistancesHf4+–N1andHf4+–N2,
respec-tively,andr3isthedistancebetweenN1andN2.TheCLisacutoff
limitsetto6.0 ˚A,afterwhichthree-bodytermsbecomenegligible. TheQM/MMsimulationsystemisdividedintotwoparts,aregion thatincludestheionandthefirstsolvationshell(QM)aretreated byquantum mechanicsandremainingarea(MM)by molecular mechanics.Thesystemforceisdescribedbythefollowingformula
Ftot=FMMsys +(FQMQM−FQMMM)S(r) (5)
whereFMMsys istheMMforceofthewholesystemandFQMQMandFQMMM areQMandMMforcesintheQMregionwhileFtotisthetotalforce
actingonaparticle. Toensureacontinuouschangeofforces, a smoothingfunctionS(r)isappliedbetweentheradiir0andr1:
S(r)=1, forr≤r1
S(r)=
(r20−r2)2(r02+2r2−3r21)
(r02−r12)3 , forr1<r≤r0 S(r)=0, forr>r0
(6)
FreemigrationofligandsbetweenQMandMMregionispermitted inthisapproach[15–17].
Table2
Theoptimizedparametersoftheanalytical2-bodypotentialfunctionforHf4+–ammoniainteraction.
2-Body AN BN CN DN
(kcal/molA5) (kcal/molA9) (kcal/molA11) (kcal/molA12)
Hf4+–N −13579.1072059 618783.1566433 −2950364.7296899 2609462.3322842
2-Body AH BH CH DH
(kcal/molA4) (kcal/molA5) (kcal/molA6) (kcal/molA12)
Table3
ThestructureparametersofthesolvatedHf4+inliquidammoniadeterminedbythe
classicalandQM/MMMDsimulations.
ClassicalMD QM/MMMD Experimentc
r1Hf4+
–Na 2.51 2.38 2.383
r2Hf4+
–Na 5.27 5.31 –
CN1stb 5.00 5.00 5.00
CN2ndb ∼34.70 ∼30.00 –
N–Hf4+–Nangle(◦) 73.90/146.70 90.00/173.00 –
aFirstandsecondpeakmaximumofHf4+–NRDFinÅ. bCoordinationnumbersofthefirstandsecondsolvationshell.
c BasedonX-raydiffractionmeasurementofthecrystalofHf(NH3)4F4·NH3.
2.2. DetailsofQM/MM-MDsimulation
ThesimulationwascarriedoutinthecanonicalNVT ensem-ble,consistingofoneHf4+ionand215NH
3 moleculesinacubic
boxof20.8 ˚Asidelength,correspondingtothedensityofthe sys-tem0.690g/cm3.Thesimulationtemperaturewaskeptconstant
at235.15K usingtheBerendsen algorithm.The flexible ammo-niamodelincludingintra-andinter-molecularpotentialwasused. Consequently,thetime stepof thesimulationwassetto0.2fs, whichallowsforexplicitmovementofthehydrogen.Acutoffof 10.40 ˚AwassetexceptforN–HandH–Hnon-Coulombic interac-tionswhereitwassetto6.0and5.0 ˚A.Thereactionfieldmethod wasusedtoaccountforlong-rangeelectrostaticinteractions.
Theclassical2-bodypotentialmoleculardynamicssimulation hasbeenperformedfirstfor100psandcontinuedwith2-body+ 3-bodypotentialfor100ps.Then,a90psoftheQM/MMsimulation wascarriedoutfromtheequilibriumconfigurationoftheclassical simulation.ToensurethefullinclusionofthefirstshellintotheQM zonetheradiusoftheQMspherewassetto3.9 ˚Ainaccordancewith theHf4+–NRDFobtainedfromtheclassicalsimulation.Theworkof
theMDsimulationswereperformedinAustrian–Indonesian Cen-ter(AIC)forComputationalChemistry,GadjahMadaUniversity, Yogyakarta,Indonesia.
3. Resultsanddiscussion
3.1. Structure
ThestructureofsolvationofHf4+inliquidammoniahasbeen
obtainedby QM/MMsimulation at theHF level.The structural propertiesareconfirmedonseveralparameterssuchasRadial Dis-tributionFunction(RDF),CoordinationNumberDistribution(CND), AngleDistributionFunction(ADF).Theirparametervaluesobtained fromtheclassicalaswellasQM/MMsimulationarepresentedin
Table3. The Hf4+–N and Hf4+–H RDFs aredepicted in Figure1
thatobtainedbytheQM/MMsimulationatHartree–Focklevel.The maximumpeakofHf4+–NRDFislocatedat2.38 ˚Ainthefirstshell
whilethefirstmaximumpeakinclassical2-body+3-bodypotential simulationsisobservedat2.51 ˚A.TheHf4+–Ndistanceof2.38 ˚Ais
consistentwithexperimentaldata2.383 ˚AobtainedbyKrausetal.
[22].TheprobabilitiesgHf4+–N(r)betweenthefirstandsecondshell equaltozerothatindicatenoligandexchangetooccurbetweenthe twoshellsduringsimulation.Abroadpeaklocatedbetween4.2 and6.4 ˚Awithamaximumat5.31 ˚Aindicatedforhighflexibilityof ammoniamoleculeswithinthisshell.
Coordinationnumber distributionfor solvatedHf4+ in liquid
ammoniaderivedfromtheclassicalandQM/MMsimulationswere displayedinFigure2.AccordingtotheQM/MMsimulation, coor-dinationnumberof5observedinthefirstsolvationshellwitha 100%occurrenceisinagreementwiththeexperimentaldatawhile thecoordination numberof10isobtainedfromtheclassical 2-bodypotentialsimulations.Givingcorrection3-bodyeffecttothe 2-bodypotentialshowedadecreasetothecoordinationnumber5
Figure1. TheradialdistributionfunctionsofHf4+–NandHf4+–Handtheirrunning integrationnumbersobtainedbyQM/MM-MDsimulation.
withanearly100%occurrence.Thecoordinationnumberin sec-ondshellrangedfrom28to34fortheclassical2-body+3-body potentialand27–33forQM/MM-MDsimulations(average:30.2) whileifonlyusing2-bodypotentialthebroadcoordination num-berdistributionwithahighoccurrenceofabout27wasobtained.As comparison,thesecondsolvationshellcontainsabout30ammonia moleculesinthecaseofsolvationLi+inliquidammonia[34].
Thesolvationstructurecouldbecharacterizedonthebasisof angular distributionfunction. Angular distributionof N–Hf4+–N
anglesinthefirstsolvationshellwasdepictedinFigure3a.Inthe distributionplotsoftheN–Hf4+–Nangles,theobtainedtwopeaks
bytheclassicalsimulationarelocatedat73.90◦and146.70◦.The
changesoftheN–Hf4+–Nanglesarefoundafterthemany-body
Figure3.(a)TheangulardistributionofN–Hf4+–Nanglesuptothefirstminimum oftheHf4+–NRDFs,(b)distortedsquarepyramidalstructureofthesolvationofHf4+ inliquidammonia(snapshottakenbyTmolex).
correctionshavebeenincluded.Asharppeakwasobservedat90◦
whilethebroadpeakappearedat173◦andaminimumoccurring
at150◦.Theexistenceofthefirstpeakat90◦withahighprobability
andthesecondpeakat173◦(almost180◦)indicatedthestructureof
Figure4.Noammonialigandsmigrationbetweenfirstandsecondsolvationshell areobserved.Theammonialigandexchangesareobservedbetweensecond solva-tionshellandbulk.
Table4
Meanligandresidencetimes(MRT),inps,numberofaccountedexchangeevents (Nex)obtainedbydirectmethodasafunctionoft*,andsustainabilitycoefficient
(Sex).
tsim t*=0ps t*=0.5ps Sex 1/Sex
ps N0ex Nex0.5
Secondshell 90 1398 1.943 533 5.096 0.381 2.625
Hf(NH3)54+complextendstoasquarepyramidalstructurewiththe Hf4+ionliftedabovetheaveragenitrogenplane(Figure3b)whereas foranidealsquarepyramidalstructurehas90◦and180◦.As
com-parison,structureofsolidcopper(II)complexwithammoniahas beenreported,includingcoordinationoffour(squareplanar)and five(squarepyramidal)nitrogensandsix(distortedsquare bipyra-midal)nitrogens[2,35]whileforNa(NH3)5+couldbeassignedto
twomainstructures,namelyatrigonalbipyramidalandasquare pyramidal[36].
3.2. Dynamics
Ligandexchangeofammoniabetweenthefirstandsecondshell isnotobservedfor90pssimulationasdisplayedinFigure4,while theligandexchangesareoccurredbetweenligandsinthesecond solvationshellandbulk.Ultrafastligandexchangeisimportantto indicatethereactivityof Hf4+ ion.It ispossibletomeasurethe
numberofexchangeeventsleadingtoalonger-lastingchangein thesolvation structure bycomparingthe number ofaccounted exchangeeventswitht*=0.5ps
Nex0.5
andt*=0psN0ex,definingasustainabilitycoefficient:
Sex= N
0.5 ex
Nex0
(7)
Itsinverse(1/Sex)accountshowmanyborder-crossingattemptsare
neededtoproduceonelonger-lastingchangeinthesolvation struc-tureofanindividualion[21,37].Somedynamicpropertiessuchas thenumberofligandexchange,meanligandresidencetime(MRT), andthesustainabilityofthemigrationprocessarelistedinTable4. Themeanligandresidencetimeinthesecondsolvationshellof Hf4+was5.096ps.Incontrast,theMRTinthesecondshellis
sig-nificantlysmallerthaninwater(0.5ps=15.5ps)[19],indicatingfor anincreasedliabilityofthesecondshellinthecaseofammonia.On theotherhand,thesustainabilitycoefficientSexofligandmigration
hasavalueof0.381,thecorresponding1/Sexis2.625,whichmeans
thatlessthanthreeattemptstoleaveorenterthesecondsolvation shellareneeded,toachieveoneexchangeprocess,whichlastsat least0.5ps.
4. Conclusion
ThesimulationHf4+ioninliquidammoniahasbeenperformed
successfully by the QM/MM method. The QM/MM simulation resultsindicatedthatthestructureofthefirstsolvationshell con-sistsoffiveammoniamoleculestendstoformasquarepyramidal structure.TheHf4+–Ndistanceof2.38 ˚Aisinaccordancewiththe
experimentalX-raydiffractiondata.Thereisnoligandexchange betweenthesecondshellandthefirstshellduringthesimulation of90ps.Theresidencetimeoftheligandinthesecondsolvation shellis5.096ps.Thisvalueissmallerthaninwater,indicatinga highflexibilityofthesecondshellinthecaseofammonia.
Acknowledgements
Indonesiawhereasthesoftwareandhardwareweresupportedby theAustrian–Indonesian Center(AIC)for Computational Chem-istry,GadjahMadaUniversityaregratefullyacknowledged.
References
[1]E.Rezabal,J.M.Mercero,X.Lopez,J.M.Ugalde,J.Inorg.Biochem.100(2006) 374.
[2]H.D.Pranowo,B.M.Rode,J.Phys.Chem.A103(1999)4298.
[3]M.Vyalov,M.Kiselev,T.Tassaing,J.C.Soetens,A.Idrissi,J.Phys.Chem.B114 (2010)15003.
[4]S.Varma,S.B.Rempe,Biophys.Chem.124(2006)192.
[5]A.Tongraar,T.Kerdcharoen,S.Hannongbua,J.Phys.Chem.A110(2006)4924. [6]A.Tongraar,B.M.Rode,Chem.Phys.219(2005)279.
[7]P.Ji,J.H.Atherton,I.P.Michael,J.Org.Chem.76(2011)3286. [8]Y.Liu,M.E.Tuckerman,J.Phys.Chem.B105(2001)6598.
[9]S.Sarkar,A.K.Karmakar,R.N.Joarder,J.Phys.Chem.A101(1997)3702. [10]S.Tongraar,Hannongbua,J.Phys.Chem.B112(2008)885.
[11]E.A.Orabi,G.Lamoureux,J.Chem.TheoryComput.9(2013)2035.
[12]H.Thompson,J.C.Wasse,N.T.Skipper,S.Hayama,D.T.Bowron,A.K.Soper,J. Am.Chem.Soc.125(2003)2572.
[13]G.C.Miller,J.Chem.Educ.58(1981)425.
[14]H.Matsuno,A.Yokoyama,F.Watari,M.Uo,T.Kawasaki,Biomaterials22(2001) 1253.
[15]R.Armunanto,C.F.Schwenk,B.R.Randolf,B.M.Rode,Chem.Phys.305(2004) 135.
[16]R.Armunanto,C.F.Schwenk,B.M.Rode,J.Am.Chem.Soc.126(2004)9934. [17]R.Armunanto,C.F.Schwenk,B.R.Randolf,B.M.Rode,Chem.Phys.Lett.388
(2004)395.
[18]C.B.Messner,T.S.Hofer,B.R.Randolf,B.M.Rode,Chem.Phys.Lett.501(2011) 292.
[19]C.B.Messner,T.S.Hofer,B.R.Randolf,B.M.Rode,Phys.Chem.Chem.Phys.13 (2011)224.
[20]N.Prasetyo,L.R.Canaval,K.Wijaya,R.Armunanto,Chem.Phys.Lett.619(2015) 158.
[21]B.M.Rode,T.S.Hofer,B.R.Randolf,C.F.Schwenk,D.Xenides,V. Vchirawongk-win,Theor.Chem.Acc.115(2006)77.
[22]F.Kraus,S.A.Baer,M.B.Fichtl,Eur.J.Inorg.Chem.(2009)442.
[23]V.Vchirawongkwin,A.Tongraar,C.Kritayakornupong,Comput.Theor.Chem. 1050(2014)74.
[24]S.Hannongbua,T.Remsungnen,M.Kiselev,K.Heinzinger,Condens.Matter Phys.6(2003)459.
[25]T.Yamaguchi,M.Niihara,T.Takamuku,H.Wakita,H.Kanno,Chem.Phys.Lett. 274(1997)485.
[26]R.J.Frick,A.B.Pribil,T.S.Hofer,B.R.Randolf,A.Bhattacharjee,B.M.Rode,Inorg. Chem.48(2009)3993.
[27]L.R.Canaval,A.K.H. Weiss,B.M.Rode,Comput.Theor.Chem.1022(2013) 94.
[28]T.S.Hofer,H.Scharnagl,B.R.Randolf,B.M.Rode,Chem.Phys.327(2006) 32.
[29]T.Remsungnen,B.M.Rode,Chem.Phys.Lett.385(2004)491.
[30]J.C.Rasaiah,R.M.Lynden-Bell,Phil.Trans.R.Soc.Lond.A359(2001)1545. [31]R.Ahlrichs,M.Br,H.Horn,M.Hser,C.Klmel,Chem.Phys.Lett.162(1989)
165.
[32]R.Ahlrichs,M.vonArnim,MethodsandTechniquesinComputational Chem-istry:METECC-95,Cagliari,1995.
[33]M.vonArnim,R.Ahlrichs,J.Comput.Chem.19(1998)1746.
[34]S.Hayama,N.T.Skipper,J.C.Wasse,H.Thompson,J.Chem.Phys.116(2002) 2991.
[35]M.S.Valli,Matsuo,H.Wakita,T.Yamaguchi,M.Nomura,Inorg.Chem.35(1996) 5642.
[36]T.Kerdcharoen,B.M.Rode,J.Phys.Chem.A104(2000)7077.