Effects of incident cluster size, substrate temperature, and incident energy on bombardment of Ni clusters onto Cu (0 0 1) surface studied using molecular dynamics simulation

AppliedSurfaceScience258 (2012) 5892–5897

ContentslistsavailableatSciVerseScienceDirect

Applied

Surface

Science

j o ur na l ho me p age :w w w . e l s e v i e r . c o m / l o c a t e / a p s u s c

Effects

of

incident

cluster

size,

substrate

temperature,

and

incident

energy

on

bombardment

of

Ni

clusters

onto

Cu

(0

0

1)

surface

studied

using

molecular

dynamics

simulation

Shiang-Jiun

Lin

_,

_Cheng-Da

_Wu

_,

_Te-Hua

_Fang

_,

_Guan-Hung

_Chen

a_{DepartmentofMoldandDieEngineering,NationalKaohsiungUniversityofAppliedSciences,Kaohsiung807,Taiwan}

b_{DepartmentofMechanicalEngineering,NationalKaohsiungUniversityofAppliedSciences,Kaohsiung807,Taiwan}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received9January2012

Receivedinrevisedform25February2012

Accepted26February2012

Available online 5 March 2012 Keywords: Bombardment Temperature Clustersize Kineticenergy Moleculardynamics

a

b

s

t

r

a

c

t

ThebombardmentprocessofaNiclusterontoaCu(001)surfaceisstudiedusingmolecular dynam-ics(MD)simulationsbasedonthetight-bindingsecond-momentapproximation(TB-SMA)many-body potential.Theeffectsofincidentclustersize,substratetemperature,andincidentenergyareevaluated intermsofmoleculartrajectories,kineticenergy,stress,self-diffusioncoefficient,andsputteringyield. ThesimulationresultsclearlyshowthatthepenetrationdepthandCusurfacedamageincreasewith increasingincidentclustersizeforagivenincidentenergyperatom.Theself-diffusioncoefficientandthe penetrationdepthofaclustersignificantlyincreasewithincreasingsubstratetemperature.Anincident clustercanbescatteredintomoleculesoratomsthatbecomeembeddedinthesurfaceafterincidence. Whentheincidentenergyisincreased,thenumberofvolcano-likedefectsandthepenetrationdepth increase.Ahighsputteringyieldcanbeobtainedbyincreasingtheincidentenergyathightemperature. Thesputteringyieldsignificantlyincreaseswithclustersizewhentheincidentenergyisabove5eV/atom.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

Anunderstandingofthedynamicinteractionsofenergetic par-ticlesandclusterswithsolidsurfacesisnecessaryforimproving nanoscale processes, suchas thin-filmdeposition [1,2], surface modification[3],andsurfacesmoothing[4,5],andcharacterizing surfacesthroughvarious spectroscopytechniques.Particlesand clustersrequireacertainlevelofincidentkineticenergy,which varieswiththetypeofprocessandtheinteractionmechanism (sin-gle,binary,ormulti-collision),tointeractwithasolidsurface.For example,asmallincidentkineticenergyisrequiredforthin-film depositiontoallowtheincidentparticlestostayonthesurfaceand grow;incontrast,ahighincidentenergyisoftenappliedforsurface modificationandsurfacesmoothingprocessesinordertobreakthe surfacebondsandpenetratethesurface.

Thecollisionsofparticlesandclustersof10–104 _atomswith

solidsurfacesoccuratsmallspatialandtemporal (pico-second) scales,makingthemverysuitableforcharacterizationandanalysis byatomistic-levelmodeling.Moleculardynamics(MD)isa pow-erfulatomisticsimulationtoolwhich avoidsexperimentalnoise andturbulenceproblems,andcanbeusedtoanalyzemolecular

∗ Correspondingauthor.Tel.:+886738145265336.

E-mailaddress:[email protected](T.-H.Fang).

trajectoriesandthermodynamicproperties.MDhasbeenutilized toinvestigatethemechanicalcharacteristicsofC60-filledcarbon

nanotubes[6],dip-pennanolithography [7,8],and nanoimprint-ing[9,10].Hongetal.[11] studiedthecopper(Cu)filmgrowth processusingMDandfoundthatthegrowthmodes(epitaxy, mix-ing,andsputtering)dependontheaccelerationenergyofparticles. AokiandMatsuo[12]studiedtheeffectsoftheparametersofthe glancing-anglegasclusterionbeam(GCIB)techniqueonsurface modificationandfoundthattheincidentangleandenergyofagas clusterarecriticalforsurfacedamage,sputtering,andsmoothing.

Niisfrequentlyusedasasputteringtargetmaterial.Nicanbe dopedintoCusubstratestoimprovetheirsurface strengthand it hasexcellent catalyticcapabilityfor carbonnanotube growth [13,14]. This work investigates the effects of incident cluster size, substratetemperature, and incidentkinetic energy onthe bombardmentof Niclustersonto aCu (001)surface usingMD simulation.Theresultsarediscussedintermsofatomic trajecto-ries,kineticenergy,stress,self-diffusioncoefficient,andsputtering yield.

2. Methodology

Fig.1showsaschematicMDmodelofNiclustersbombardinga Cusubstrate.Incidentclusterswith10,107,and214Niatomswere usedatnormalincidence,respectively.ThedimensionsoftheCu

0169-4332/$–seefrontmatter © 2012 Elsevier B.V. All rights reserved.

S.-J.Linetal./AppliedSurfaceScience258 (2012) 5892–5897 5893

Fig.1.SchematicMDmodelofNiclustersbombardingCu(001)surface.

substrate,whichismadeupofperfectface-centeredcubic (fcc) atoms,are7.3(length)_{× 7.3}(width)_{× 7.3}(height)nm.TheCu sub-strateconsistsoftwofixedlayersofatomsatthebottomtosupport theentiresystem,fourthermostatlayersabovethem,andfree lay-ersonthesurface.Theinitialsystemiscontrolledtobeatroom temperature.Aperiodicboundarycondition(PBC)isimposedon theX-andY-axesofthesurfaceplane.Inmathematicalmodels andcomputersimulations,aPBC[15]isoftenusedtosimulatea largesystembymodelingasmallpartthat isfarfromitsedge. Theincidentenergyoftheclusterandthesubstratetemperature arevariedintherangesof1–10eV/atomand300–1000K, respec-tively,toinvestigatetheirinfluencesonthebombardmentprocess. Beforethebombardmentsimulationwasstarted,aseparation dis-tanceof2nmwassetbetweentheclusteratomsandthesubstrate topreventinteractionsintheearlystages,andanMDequilibrium runof5pswasusedtoachieveenergyrelaxation.

The tight-binding second-moment approximation (TB-SMA) many-bodypotential model [16] is adopted for simulating the interactive behavior between substrate–substrate (Cu–Cu) and substrate–cluster(Cu–Ni)atoms.TheTB-SMApotentialhasbeen proven to bemore accuratethan the embedded-atom method (EAM)many-body potential[17,18]. Thestress descriptionof a substrateatatomicscale[19]is

mn₌₋1 Ns N_S



⎡

⎣

v

v



⎤

⎦

wheremn _{representsthestressoccurredinthedirection‘n’on}

theplane‘m’;Nsrepresentsthenumberofatomsinasubstrate;

Nrepresentsthenumberofatomsinasubstrateandaincident cluster;Mirepresentsthemassofatomi;

v

ofatomiinthedirection‘m’;(rij)representstheTB-SMApotential

function;rijrepresentstherelativedistancefromatomitoj;xijm

representstherelativedistancefromatomitojinthedirection‘m’; Viistheregionalatomicvolumetoatomiandgivenby

Vi=4ai 3

3 (2)

whereairepresentstheregionalparticleradiusandgivenby

ai=



rij−1

2



(3) ThetimeintegrationofmotionwasperformedusingGear’s fifth-orderpredictor-correctormethod[15]withatimestepof1fs.To increasecalculationefficiency,theVerletneighbor-listmethod[15]

wasused.Inthesimulation,thelistsofneighboratomswere cal-culatedevery10timesteps.

3. Resultsanddiscussion 3.1. Effectofclustersize

Theincidentclustersconsistof10,107,and214Niatoms, corre-spondingtoclusterdiametersof0.1,1.1,and2.4nm,respectively. Fig.2(a)–(c)and(d)–(f)showaseriesofsnapshotsfortheNi clus-tersimpactingtheCu(001)surfacewith1eVofkineticenergy peratomatroomtemperatureat2000fsand10,000fs, respec-tively.At2000fs,alittlerippleformsontheCusurfacewherethe impactoccurs.ThisisduetotheclusterpenetratingtheCu sur-face.Foragivenincidentenergyperatom,theextentoftheripple ontheCusurfaceincreaseswithincreasingincidentclustersize (numberofincidentatoms).Withincreasingsimulationtime,the penetrationdepthincreasesuntilthekineticenergyofthe clus-tergetscloseto0eV.WhenaclusterpenetratestheCusurface,it separatesintomoleculesoratoms,whichareadsorbedontothe Cusurfaceviastrongattractiveforces.Boththepenetrationdepth andCusurfacedamageincreasewithincreasingclustersizedue tothelargermassand largertotal kineticenergy ofthe projec-tile,asshownin Fig.2(d)–(f).For example,theincidentcluster with10NiatomsisdirectlyadsorbedontotheCusurface,which doesnotexhibitcleardamageafterbombardment,whereasthat with214Niatomspenetratesthesubstrateandcreatesobvious surfacedefects.Fig.3showsthevariation ofthekineticenergy ofclusterswithtime. Thekineticenergyofaclusteristhesum ofthekineticenergiesofitsconstituentatoms.Before bombard-ment(0–1200fs),thekineticenergyofaclusterisatitshighest levelandstable.Thekineticenergyofaclusterincreaseswiththe numberofatoms.Attheinstantofcollision,thekineticenergyof theclusterquicklydecays(1200–5000fs).Theextentofdecayof thekineticenergyincreaseswithclustersizeandthuspenetration depth.After5000fs,thekineticenergyoftheclusterdecaystoa valuenear0eV,whichindicatesthatthebombardmentprocesshas ended(i.e.,theclusteratomsaremotionless).Fig.4showsthe vari-ationoftheresidualstress(equivalentstress)oftheCusubstrate withbombardmenttimeforincidentclusterswithvarioussizes.At theinstantofcollision,thesubstratestressincreaseswithincident clustersizeduetotheincreaseinincidentenergy.Withincreasing time,theinternalstressofthesubstrategraduallyrelaxes;therate ofrelaxationishighestforthesmallestincidentcluster(with10 Niatoms).Forsubstratesbombardedbyclusterswith107and214 atoms,respectively,thestresscurvesaresimilar,whichindicates similarstress-relaxationprocesses.Theeffectofclustersizeonthe residualstressoftheCusubstrateisnotsignificantduetothelow incidentenergy.

3.2. Effectoftemperature

Thesubstratetemperaturewasvariedtoinvestigateitseffecton thebombardmentprocess.Anincidentclusterwith107Niatoms withakineticenergyof1eV/atomwasused.Fig.5(a)–(b)shows snapshotsofthebombardmentprocessat2000fsforsubstrate tem-peraturesof600and1000K,respectively.Themorphologyofthe Cusurfaceisirregularathightemperature.Ingeneral,thekinetic energyofatomsincreaseswithincreasingtemperature.Thefigures alsoindicatethattheatomsonthesurfacebecomemoreactive (fastermovement)athighertemperatures.Acomparisonwiththe bombardmentbehavioratroomtemperature(Fig.2(b))indicates thatthepenetrationdepthsignificantlyincreaseswithincreasing substratetemperature.Figu.5(c)and(d)showssnapshotsofthe bombardmentprocessat10,000fsforsubstratetemperaturesof

S.-J.Linetal./AppliedSurfaceScience258 (2012) 5892–5897 5897

4. Conclusion

ThisstudyconductedanMDsimulationtoinvestigatetheeffects ofincidentclustersize,substratetemperature,andincidentenergy onthebombardmentofNiclustersontoaCu(001)surface.The followingconclusionswereobtained:

(1)Foragivenincidentenergyperatom,thepenetrationdepth and Cu surface damage increase with increasing cluster size.

(2)Theself-diffusioncoefficientandthepenetration depthof a clustersignificantlyincreasewithincreasingsubstrate temper-ature.Anincidentclustercanseparateintomanymoleculesor atoms,whichbecomeembeddedinthesurfaceafterincidence. (3)Thenumberofvolcano-likedefectsandthepenetrationdepth

increasewithincreasingincidentenergy.

(4) Highsputteringyieldcanbeobtainedbyincreasingtheincident energyathightemperature.

Acknowledgment

ThisworkwassupportedbytheNationalScienceCouncilof TaiwanundergrantsNSC100-2628-E-151-003-MY3andNSC 100-2221-E-151-018-MY3.

References

[1] A.Kirkpatrick,Nucl.Instrum.MethodsB206(2003)830.

[2] Z.H.Hong,S.F.Hwang,T.H.Fang,J.Appl.Phys.103(2008)124313.

[3] I.Yamada,J.Matsuo,N.Toyoda,T.Aoki,E.Jones,Z.Insepov,Mater.Sci.Eng.A

253(1998)249.

[4] S.Kakuta,S.Sasaki,K.Furusawa,T.Seki,T.Aoki,J.Matsuo,Mater.Res.Soc.

Symp.Proc.843(2005)T.5.5.

[5]E.Bourelle,A.Suzuki,A.Sato,T.Seki,J.Matsuo,Nucl.Instrum.MethodsB241

(2005)622.

[6] C.D.Wu,T.H.Fang,C.Y.Chan,Carbon49(6)(2011)2053.

[7] C.D.Wu,T.H.Fang,J.F.Lin,Langmuir26(5)(2010)3237.

[8] C.D.Wu,T.H.Fang,J.F.Lin,J.ColloidInterfaceSci.361(1)(2011)316.

[9] T.H.Fang,C.D.Wu,W.J.Chang,S.S.Chi,Appl.Surf.Sci.255(11)(2009)6043.

[10] C.D.Wu,J.F.Lin,T.H.Fang,NanoscaleRes.Lett.4(8)(2009)913.

[11] Z.H.Hong,S.F.Hwang,T.H.Fang,Comput.Mater.Sci.41(1)(2007)70.

[12]T.Aoki,J.Matsuo,Nucl.Instrum.MethodsB257(2007)645.

[13]G.Atthipalli,R.Epur,P.N.Kumta,B.L.Allen,Y.Tang,A.Star,J.L.Gray,ThinSolid

Films519(2011)5371.

[14]F.Bonnet,F.Ropital,Y.Berthier,P.Marcus,Mater.Corros.54(2003)870.

[15]J.M.Haile,MolecularDynamicsSimulation:ElementaryMethods,Wiley,New

York,1992.

[16]F.Cleri,V.Rosato,Phys.Rev.B48(1993)22.

[17]C.G.Hwang,J.G.Chang,G.J.Huang,S.H.Huang,J.Appl.Phys.92(2002)5904.

[18]M.A.Karolewski,Radiat.Eff.DefectsSolids153(2001)239.

[19] Z.C.Lin,J.C.Huang,Nanotechnology15(2004)1509.

[20]C.Anders,G.Ziegenhain,S.Zimmermann,H.M.Urbassek,Nucl.Instrum.

Meth-odsB267(2009)3122.

[21]Z.J.Schiffer,P.E.Kennedy,Z.Postawa,B.J.Garrison,J.Phys.Chem.Lett.2(2011)

2635.