Surface-driven, one-step chemical vapor deposition of γ-Al4Cu9 complex metallic alloy film

(1)

HAL Id: hal-00966984

https://hal.archives-ouvertes.fr/hal-00966984

Submitted on 27 Mar 2014

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of

sci-entific research documents, whether they are

pub-lished or not. The documents may come from

teaching and research institutions in France or

abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est

destinée au dépôt et à la diffusion de documents

scientifiques de niveau recherche, publiés ou non,

émanant des établissements d’enseignement et de

recherche français ou étrangers, des laboratoires

publics ou privés.

Surface-driven, one-step chemical vapor deposition of

γ-Al4Cu9 complex metallic alloy film

Nathalie Prud’Homme, Thomas Duguet, Diane Samélor, François Senocq,

Constantin Vahlas

To cite this version:

Nathalie Prud’Homme, Thomas Duguet, Diane Samélor, François Senocq, Constantin Vahlas.

Surface-driven, one-step chemical vapor deposition of γ-Al4Cu9 complex metallic alloy film. Applied Surface

Science, Elsevier, 2013, vol. 283, pp. 788-793. �10.1016/j.apsusc.2013.07.019�. �hal-00966984�

(2)

O

pen

A

rchive

T

OULOUSE

A

rchive

O

uverte (

OATAO

)

OATAO is an open access repository that collects the work of Toulouse researchers and

makes it freely available over the web where possible.

This is an author-deposited version published in :

http://oatao.univ-toulouse.fr/

Eprints ID : 11271

To link to this article :

DOI:10.1016/j.apsusc.2013.07.019

URL :

http://dx.doi.org/10.1016/j.apsusc.2013.07.019

To cite this version :

Prud’homme, Nathalie and Duguet, Thomas and Samélor, Diane

and Senocq, François and Vahlas, Constantin Surface-driven,

one-step chemical vapor deposition of γ-Al4Cu9 complex metallic alloy

film. (2013) Applied Surface Science, vol. 283 . pp. 788-793. ISSN

0169-4332

Any correspondance concerning this service should be sent to the repository

administrator:

staff-oatao@listes-diff.inp-toulouse.fr

(3)

Surface-driven,

one-step

chemical

vapor

deposition

of

g-Al

4 Cu

9 complex

metallic

alloy

ﬁlm

Nathalie

Prud’homme

a,b

_,

_Thomas

_Duguet

a,∗

_,

_Diane

_Samélor

a

_,

Franc¸

ois

Senocq

a

_,

_Constantin

_Vahlas

a

a_CIRIMAT,_Université_de_Toulouse_-_CNRS,₄_allée_Emile_Monso,_BP-44362,₃₁₄₃₂_Toulouse_Cedex_4,_France b_Université_Paris-Sud_11,_LEMHE/ICMMO,_Bat_410,₉₁₄₀₅_Orsay_Cedex,_France

a

b

s

t

r

a

c

t

Thepresentpaperisaparadigmfortheone-stepformationofcomplexintermetalliccoatingsby chem-icalvapordeposition.Itgenuinelyaddressesthechallengeofdepositinganintermetalliccoatingwith comparablecontentsofCuandAl.Dependingonprocessingconditions,apureg-Al4Cu9andmulti-phase

Al-Cufilmsaregrownwithwettingpropertiesoftheformerbeingsimilartoitsbulkcounterpart.The depositionprocessanditsparametricinvestigationaredetailed.Twometalorganicprecursorsareused takingintoaccounttheirtransportandchemicalproperties,anddepositiontemperatureranges.Online andexsitucharacterizationsenlightenthecompetitionwhichoccursatthegrowingsurfacebetween molecularfragments,andwhichlimitsgrowthrates.Notably,introducingapartialpressureofhydrogen gasduringdepositionreducesAlgrowthratefromdimethylethylaminealane(DMEAA),bydisplacingthe hydrogendesorptionequilibrium.ThisAlpartialgrowthratedecreaseisnotsufficienttoachieveaCu/Al atomicratiothatishighenoughfortheformationofintermetallicswithcloseAlandCucompositions. Afivefoldincreaseofthefluxofthegaseouscopper(I)cyclopentadienyltriethylphosphineCpCuPEt3,

whereastheDMEAAﬂuxremainsconstant,resultsinthetargetedAl/Cuatomicratioequalto44/56. Nevertheless,theglobalgrowthrateisrenderedextremelylowbythedepositioninhibitioncausedby amassivephosphineadsorption(-PEt3).Despitetheselimitations,theresultspavethewaytowardsthe

conformalcoatingofcomplexsurfacegeometriesbysuchintermetalliccompounds.

1. Introduction

Coating of complex 3-dimensional surfaces with complex

metallic alloys (CMAs) including quasicrystals provides added

valuematerialsforinnovativeapplications.Examplesarelow

stick-ingandlowfrictionglassmoldsthatsustainhightemperatures,

andporouspreformsinﬁltrationfortheproductionofsupported

catalysts([1–3],andreferencestherein).CMAcoatingsshow

excel-lentmechanicalpropertiessuchaslowadhesion,anti-fretting,or

wearresistance,butthedrawbackoflowadhesionisthatitalmost

alwaysresultsintoautodelaminationthatcanonlybeovercome

byanefﬁcientincreaseofinterfacialaccommodationandadhesion.

Severalsolutionshavebeenproposedtofacethisproblem;one

beingtheuseofabufferlayerofamaterialthatshowsintermediate

electronicandstructuralpropertiesbetweenthoseofthesubstrate

andthecoating[4,5].TheAl-Cusystempresentssuch

intermedi-atephaseswhichhaveproventheirefﬁciencyforaccommodating

icosahedralAl-Pd-MnandAl-Cu-Fequasicrystals[5,6].

∗ Correspondingauthor.

E-mailaddresses:thomas.duguet@ensiacet.fr,doug181@gmail.com(T.Duguet).

Actually,thereisnosatisfactorytechniquefortheprocessing

ofsuchcoatingsoncomplexsurfaces.Byoperatinginthesurface

reactionlimitedregime,chemicalvapordepositioncaninprinciple

meetthisrequirement.Inthispaperwepresentfortheﬁrsttime

thedirect,one-stepprocessingofapureCMA-containingcoating

bymetalorganicchemicalvapordeposition(MOCVD).Speciﬁcally,

wereportontheMOCVDofAl-Cualloyedcoatings,includingthe

g-Al4Cu9 approximantphasewithsimilarpropertiestothebulk

crystal.

PreviousCVDco-depositionexperimentsofAl-Cuﬁlms were

performedintheearlyninetiesinordertoreplacepureAl

intercon-nectionsinverylargescaleintegrationtechnologybya-(Al)solid

solutionswhoseelectromigrationandresistivityarelower.

Tomo-haruetal.[7]andEiichietal.[8]useddimethylaluminumhydride,

DMAH,andcoppercyclopendadienyltriethylphosphine,CpCuPEt3

asprecursorsfortheco-depositionofAlandCu,respectively.They

obtainedpureﬁlmswithhomogeneousdistributionofCuintheAl

matrix,containingAl2CuprecipitatestypicalofhypoeutecticAl-Cu

alloys.Themajordifferencebetweentheseworksandthepresent

oneliesinthelowconcentrationofCuintheformer,whichisin

theorder1wt%.Processingofﬁlmswithcomparablecontentof

(4)

phaseand/orsurfaceinteractionsbetweenthetwoprecursorsor

theirfragments.InaﬁrstapproachtotheMOCVDofAl-Cucoatings,

werecentlypublishedresultsonthesequentialdepositionofAl

andCufollowedbyannealing.Intermetallicsarealsoobtainedbut

whilethemethodissimplerthantheoneproposedhere,itpresents

additionalstepsthataresourceofﬁlmscontaminationandinvolve

higherthermalload[9].

Inaco-depositionprocess,cautionhastobetakenregarding

compatibilitybetweentheselectedmetalorganicprecursors.

Com-patibilityconcernsdepositiontemperatures,chemicalcomposition

(O, F, and C contaminants release), chemical interactions, and

transportinthegasphase.Forthisreason,thepresentworkhas

beenprecededbytheinvestigationoftheMOCVDofunaryﬁlms

of Aland Cu. Cuis obtained by decomposition of CpCuPEt3, a

solidprecursor,thesaturatedvaporpressureofwhichis0.2Torr

at 60◦_C _[10]_. _Its _thermal _{decomposition} _in _the _gas _phase _at

150◦_C_<_T_<₂₇₀◦_C _results _in _the_formation _of _PEt_3gas _and _a

sur-face intermediate {CuCp}surf that rapidlyconvertsinto metallic

Cuand(Cp2)gas.Above270◦C,themechanismispresentedas

fol-lows: CpCuPEt3→Cusurf+HCpgas+other organic fragments [10].

In vacuum, mass spectrometry experiments show an intensity

decrease of themolecular peak at temperaturehigher than ca.

150◦_C, _whereas_{decomposition}_with_hydrogen_is_somehow

sta-bilized and shifted at approx. T>170◦_C. _CuCpPEt

3 is therefore

suitableforMOCVDinthesurfacetemperaturerange150–290◦_C.

In this range, growth rate increases with increasing precursor

feedingrateindicatingthatdepositiontakesplaceinadiffusion

controlledregime.Nonetheless,massspectrometryexperiments

show that the growing surface saturates with adsorbed

phos-phine leading to a continuous decrease of the Cu growthrate

withdepositiontime.Thiscanbecounterbalancedbyincreasing

deposition temperature above 240◦_C _in _order _to _desorb

by-products andofferenoughfreesurfaceforprecursoradsorption

[11].

Ontheotherhand,Alﬁlmsareobtainedbydecompositionof

dimethylethylaminealane,DMEAA–aliquidprecursor.The

mech-anismattributedtoitsdecompositionatsurfacescanbestressed

asfollows:adsorption,breakingoftheAl-Nbond,desorptionof

thedimethylethylamine,cracking ofthealaneAlH3 into

metal-licAlandadsorbedHatomsthatrecombinetoformH2gas[12].

Aboveasubstratetemperatureof150◦_C,_Fourier_transform_infrared

spectroscopyrevealedthatDMEAAdecomposesinthegasphase

[13],andchemicalkineticmodelingallowedtoestimatethatonly

ca. 5% of theprecursor participates in the depositionreaction,

with a stickingcoefﬁcient onAl(111)closeto unity [14].Jang

et al. [15] showed that the Al ﬁlms with the lowest electrical

resistivity,highestdensity,andhighestpurity,areobtainedata

substrate temperature of 160◦_C _which _also_corresponds _to_the

highestAlgrowthrateintheinvestigated100–220◦_C_temperature

range.

Obviously, a compromise must be found for co-deposition

involving thetwo aforementionedprecursors. Speciﬁcally,

sub-stratetemperaturemustbechoseninsuchawaythatgrowthrates

arecompatiblewiththetargetedAl-Cuphases’compositionswhile

minimizinghomogeneousdecompositionofprecursors.Itis

reas-suring thoughthatpreliminarymassspectrometryexperiments

during co-deposition show nointeractions betweenprecursors

[11].

Thearticleispresentedasfollows.Theexperimentalprotocol

involvingMOCVDofAl-Cuﬁlmsispresentedindetails,ﬁrst.Then,

thecomposition,theobtainedAl-Cuphasesandthemorphologyof

theﬁlmsarepresentedanddiscussed.Finally,preliminaryresults

ofthewettingbywateroftheg-Al4Cu9bulkandMOCVDprocessed

phasearepresented,priortoprovidingconcludingremarks.

2. Materialsandmethods

Depositionsareperformedintheexperimentalsetupdescribed

indetailsandmodeledinRef. [16].Thesetupiscomposed ofa

stagnantﬂow,cylindrical,stainlesssteelreactor.Thedeposition

chamberfeaturesadoubleenvelopeallowingthemonitoringof

wallstemperaturethroughthecirculationofthermallyregulated

silicon oil. A turbomolecular pumpensures a base pressure of

1.3×10−4_Pa._The_pumping_group_is_protected_from_the_corrosive

by-productsbyaliquidnitrogentrap.Gasisdistributedthrough

ashowerheadsystem,describedandmodeledinRef.[17].Gases

arefedthroughelectropolishedstainlesssteelgaslineswithVCR

ﬁttingsandtheirﬂowrateiscontrolledbycomputerdrivenmass

ﬂowcontrollers.HFcleaned10×5mm2_silicon_coupons_are_used

assubstrates. For each experiment,ﬁvesamplesare positioned

horizontallyona58mmdiametersusceptorandheatedbya

resis-tancecoilgyredjustbelowthesurfaceofthesusceptor.DMEAA

(Epichem)(99%pure)isusedasreceivedandCpCuPEt3(Strem)is

puriﬁedbyvacuumsublimationat70◦_C._N₂_(99.9992%_pure)_and_H₂

(99.995%pure)(airproducts)arefedthroughelectropolished

stain-lesssteelgaslineswithVCRﬁttingswithatotalﬂuxmaintainedat

325standardcubiccentimetersperminute,sccm.Allexperiments

areperformedat10Torr;i.e.areduced,althoughtechnologically

convenientpressure.

Table1isanoverviewofexperimentalconditionsandofresults

thatwillbediscussedinthefollowingsection.Thevaporization

ves-selofDMEAAismaintainedat8◦_C,_{corresponding}_to_a_saturated

vaporpressureof0.5Torr.TransportofDMEAAinvolvesaﬂowrate

ofN2 throughtheprecursorFN2,DMEAA,equalto5sccm.Inthese

conditions,themaximumﬂowrateofDMEAA,FDMEAAiscalculated

asbeingequalto0.26sccminallexperiments[18].Thesublimation

vesselofCuCpPEt3ismaintainedat75±5◦C.Twosetsoftransport

conditionsofCuCpPEt3areinvestigated,correspondingtotwo

dif-ferentﬂowratesofN2throughtheprecursor(30and150sccm).

FCpCuPEt3isestimatedfromweightlossoftheprecursorsublimator

andusingtheanalyticalformulaofRefs.[18,19].Additionally,two

substratetemperatures(200and260◦_C)_are_investigated_as_well_as

depositionwithandwithoutH2intheinputgas.

Electronprobemicroanalysis(EPMA)isusedfordetermining

theatomiccompositionsoftheﬁlms,andX-Raydiffraction(XRD)

isusedforstructuralinvestigation.Growthrate(GR)iscalculated

by weightgain of each sample prior and after depositionover

thedepositionduration. Table1presentsGRsandatomic

com-positions (at±1%) averagedover the 5 samplesplaced in each

experiment.Thiswaytoquantifythegrowthratedoesnotprovide

directaccesstothethicknessof theﬁlm.However,it has been

adoptedsincethemeasurementisstraightforwardanditismore

accuratethaninvestigation ofcrosssections bySEMduetothe

roughnessandporosityofsomesamplesaswillbeshowninthe

nextsections.GRsoftheindividualelementsduringco-deposition

areestimatedwiththefollowingformulas:GR(Al)+GR(Cu)=GR

Table1

Investigatedoperatingconditionsandresultingaveragegrowthrates(inmg/cm2/h),averageelementalcompositionsandstructureoftheﬁlms. Experiment# t(h) T(◦_C) _H₂_(sccm) _F_N

2DMEAA/FN2CpCuPEt3 FCpCuPEt3(sccm) GR EstimatedGR(Al) EstimatedGR(Cu) Al/Cu(at.%) Phases

1 2 200 25 5/30 0.07 80 78.1 1.9 99/1 Al+Al2Cu

2 2 260 25 5/30 0.04 30 24.8 5.2 33/67 Al2Cu+Al4Cu9

3 2 260 0 5/30 0.06 50 42.5 7.5 93/7 Al+Al2Cu

(5)

and[GR(Al)/MAl]/[GR(Cu)/MCu]=[Al/Cu](withMEmolecularmass

ofelementE), usingEPMAresults.Oxygen has beensubtracted

fromelementalcompositionsonpurposebecausesystematic

mon-itoringofthegasphaseduringallexperimentsbyonlinemass

spectrometrydoes notshowsigniﬁcantamountofoxygenorof

watervapors,thusensuringthatallexperimentswereconducted

intheabsenceofdetectableleaksordesorptionfromthereactor

walls.

Wettabilitybywaterisperformedbythesessiledropmethod

withaGBXDigidropinstrument.

3. Resultsanddiscussion

3.1. Averagecompositionsandgrowthrates,andcrystallographic

phases

Itiswellknownthatsublimation ofaﬁxedbedisneithera

robustnorareproduciblemeanforfeedingthedeposition

cham-berwithvaporsofa solid precursor.Indeed,statistical analysis

over10experimentsofCuprecursorsublimationwithourreactor

demonstratesthatitisnotpossibletodiscriminatetheresults#1–3

attheFN₂,DMEAA/FN₂,CpCuPEt₃ ﬂuxratioof5/30inTable1.Hardly,

weareabletoconﬁrmthatthe5/150ﬂuxratiodoescorrespond

toahigherfeedingrateofCuprecursor(0.12sccmcomparedto

0.04–0.07sccm).Anadditionalreasonforthisdiscrepancyisthe

observeddegradationoftheprecursoronthereactorwallsand

insidethesublimatorbecauseofitsreducedstabilityaboveroom

temperature.Moreover,itappearsthattheelementalcomposition

oftheﬁlmsdoesnot followthatofthegasphasebasedonthe

estimationofFprecursor.Let’sconsiderexperiments#2and#3at

260◦_C _in_Table₁_,_for_which_the_only_difference_lies_in_the

pres-enceofhydrogenornot.TheestimationsofFCpCuPEt3,at0.04and

0.06sccm,canbeconsideredclosevalues.With(without)

hydro-gen,theﬁlmscontain33at.%Al(93at.%Al)andtheglobalgrowth

rateis30mg/cm2/h(50mg/cm2/h).Amoredirectcomparisoncan

bedonethroughtheindividualgrowthrates.SuppressionofH2

intheinputgasleadstodoublingGR(Al),whereas GR(Cu)only

increasesfrom5.2to7.5mg/cm2/h.Werecallthatthemechanism

ofAldepositionfromDMEAAinvolvesdesorptionofH2gasfrom

thegrowingsurface,afterthealanedecomposition.Asa

rule-of-thumb,wepostulatethattheincreaseofthepartialpressureof

H2inthereactordisplacesthereactionequilibrium,andtherefore

decreasesthereactionrateofDMEAAatthesurface.

Thedifferencesobservedforexperiments#1and#2inTable1

aremorestraightforwardandmainlyrelyonthedeposition

tem-perature. At 200–220◦_C, _the _deposition _conditions _of _Al _from

DMEAAareoptimal.Abovethistemperature,gasphasereactions

consumemorethan97%oftheprecursorandgrowthratedecreases

rapidly[14].Inversely,increasingTfrom200to260◦_C_leads_to_a

highergrowthrateofCufromCuCpPEt3[10].Consecutively,the

individualgrowthratesofCuandAlvaryinoppositedirections

fromT=200◦_C_to_T₌₂₆₀◦_C;_GR(Cu)_being_multiplied_by_≈2.5_while

GR(Al)isdividedby≈3.However,increasingtemperatureto260◦_C

isnotanappropriateoptionforincreasingtheCucontentoftheﬁlm

sinceitdrasticallyaffectstheglobalgrowthratethatdecreasesfrom

80to30mg/cm2/h.ThesolutionforgettingCuenrichmentmust

comefromthegasphaseandnotfromahighersurface

tempera-ture.

Therefore, an attempt has been made (experiment #4 in

Table 1) with a substrate temperature of 200◦_C _and _a _lower

FN2,DMEAA/FN2,CpCuPEt3 ratio,equalto5/150.H2isalsointroduced

inordertofavorCudepositionattheexpenseofAl.Inthese

condi-tions,theestimatedCuprecursorﬂuxhasincreasedto0.12sccm.

Theobtainedglobalgrowthrateof25mg/cm2/histhelowestin

thisseriesofexperiments,mainlybecauseitisgovernedbyGR(Al)

whichislow.DuringCuunarydeposition,ithasbeendemonstrated

bymass spectrometrythat PEt3 isextensively adsorbedonthe

growingsurfacewithsubsequenthinderingofprecursorsupplyto

thedepositionzone,resultinginself-inhibitionofCudeposition

andgradualdecreaseofthegrowthrate[11].Here,thissituationis

validinalltheexperimentsbutweprobablyattainsucha

precur-sorconcentrationinexperiment#4,thatinhibitingeffectsbecome

predominantandaffectthedepositionmechanismsofAl,aswell.

Consequently,inadditiontothedisadvantageofusingH2,Al

depo-sitionrateisfurtherreducedbythesaturationofthesurfaceby

adsorbedPEt3.

Tofurthersupporttheabove-mentionedassumptions,wenow

considerexperimentally-determinedgrowthratesofunaryAl

depo-sitions.ThesegrowthratesaredifferentfromtheestimatedGR(Al)

reported in Table1. They result fromdedicated Al depositions

performed without H2(g) witha ﬂow rate of 0.26sccm. In

co-depositionconditionsbeneﬁcialtotheDMEAAprecursor,i.e.either

200◦_C_or_without_H₂_,_growth_rates_of_experiment_#1₍₈₀_mg/cm2_/h)

and#3(50mg/cm2/h)aresimilartoexperimentally-determinedAl

depositionsgrowthrates(128and40mg/cm2/h,respectively).In

co-depositionconditionsbeneﬁcialtotheCpCuPEt3precursor,i.e.

either260◦_C_(experiment_#2)_or_high_F

CpCuPEt3 at200◦C

(exper-iment#4), thepollution ofthesurface byphosphine resultsin

(i) a low globalgrowthrate,and (ii) a low GR(Al),to be

com-paredwiththeexperimentally-determinedAlgrowthrateat200◦_C

(128mg/cm2/h). Therefore, theformation of Cu-rich samplesis

more due totheinefﬁcient depositionof Alrather than tothe

increaseofthegrowthrateofCu.

Fromtheresultsreportedaboveweconcludethatwecan

suc-cessfully obtain intermetallic compounds of interest. Despite a

majordrawbackreferringtothereducedgrowthrateandprobably

theprocessyield,solutionsareavailabletocircumventthose

prob-lemswhichareattributedtothecontroloftheco-depositionby

synergeticphenomena.Fromtheanalysisoftheresultsreported

inTable1itappearsthatbetterconditionsforthegrowthof

Al-CucompoundsfromCuCpPEt3andDMEAAshouldinclude:(i)the

useofbothprecursors–butmoreimportantly ofCpCuPEt3–in

liquidsolutionincombinationwitharegulated deliverysystem

(e.g.directliquidinjection,pulsedinjection,...),(ii)avoidingthe

useofH2formaximizingAlgrowth,(iii)determiningacoupleof

[T;FCpCuPEt₃]parametersthatallowtheproperdesorptionofPEt3

fromthesurface,andpreventthethermaldegradationofDMEAA

inthegasphase,and(iv)elucidatesurfacemechanismsthatare

Fig.1.XRDspectraofsamplesfromexperiments#1–4withgivenidentificationsof phases.Phasesreportedontheright-handsideofthefigurewhereidentifiedwith aidofJCPDS#04-0787forfcc-Al,#25-0012for-Al2Cu,and#24-003forg-Al4Cu9.

(6)

responsibleforthosesynergeticeffectsandtunethesubstrate

reac-tivity,accordingly.

Fig.1showstheXRDpatternforasprocessedsamplesof

exper-iments#1–4.Theﬁlmsarecomposedof3differentphases:fcc-Al,

-Al2Cu,andg-Al4Cu9.ComparisonofthesephaseswiththeAl/Cu

ratios shown inTable1,allows concludingthatAl-rich

compo-sitions(experiments#1and#3)correspondtotheformationof

eitherpureAlorofana-(Al)solidsolutionplus-Al2Cu,

eventu-ally.Indeed,whencompositiongetsricherinCu(fromexperiments

#1–3)thenthecontributionofthe-Al2CuphasetotheXRD

dia-grambecomes larger. If thecompositionin Cu is large enough

thenthetargetedg-Al4Cu9isformed,eitherpureorwith-Al2Cu.

Hence,phaseformation qualitativelyfollowsthesimplemixing

rulethatcanbededucedfromtheAl-Cuphasediagram.Itisworth

(7)

notinghowever,thatthesampleofexperiment#2whichshows

theelementalcompositiontheclosesttothetheoreticalAl4Cu9

composition(Al33Cu67andAl31Cu69,respectively)isnottheone

giving rise to the formation of a single-phase g-Al4Cu9

coat-ing. This is observed for the sample of experiment #4 whose

elementalcompositionissomehowshiftedfromthisideal

com-position.Thisresultisincontrastwithreports,followingwhich

fora given coatingcompositionthecorrespondingpoint in the

phase diagram can beused to predict the ﬁnal phase fraction

[6,20]. It is rather attributed to partial alloying giving rise to

transient phases, mainly-Al2Cu and g-Al4Cu9 because of the

relativelylow activationenergy required fortheir formation in

thinﬁlms [9,21].Moreover,it hasbeenreportedthat,

indepen-dentlyoftheAl/Curatio,thermalannealingofAl-Cubilayersﬁrst

yields-Al2Cu at temperatures aslow as 130◦C followed by a

smallamount ofa secondphase, AlCu3. Thesubsequent phase

formation wasfoundtodepend ontheCu:Al atomic ratio. For

Cuconcentrationshigherthan50at.%,thereactionsubsequently

proceedstowardthephasesg-Al4Cu9 anda-Cu.ForCu

concen-trationlowerthan50at.%,AlCu3transformstoanewhexagonal

phaseAlxCuwhichsubsequentlytransformsbacktoAlCu3before

the reaction proceeds toward the end phases AlCu and Al2Cu

[22].

EPMAandRutherfordbackscatteringexperimentsrevealthat

oxygen concentration in the ﬁlms is high, ranging from 13

to 61%. Taking into account that on line mass spectrometry

unambiguouslyshowsbackgroundlevel forall oxygen

contain-ing species, it is concluded that oxidation of the ﬁlms occurs

ex situ.Based onSEM observations illustratedin the next

sec-tion,thehighsurfacetovolumeratiois mostlikelyresponsible

for surface oxidation in addition to O2 and water

adsorp-tion.

3.2. Morphology

Fig.2regroupssurfaceandcross-sectionSEMimages.Theﬁrst

obvious observation is that rough morphologies correspond to

Al-richcompositions(experiment#1and#3)whereassmoother

morphologiesarefoundonCu-richsamples(experiment#2and

#4).However,sinceAl-richﬁlmsaremuchthickerbasedonweight

gainandonSEMcross-sectionsanalysis,itisunclearwhetherhigh

thicknessorAlgrowthorthecombinationofbothisresponsible

forroughnessandporosity.Agoodindicationcomesfrom

deposi-tiontemperatures.Theﬁlmofexperiment#1wasgrownat200◦_C

whereastheﬁlmofexperiment#3wasgrownat260◦_C._At₂₀₀◦_C,

growthrateishigherbecauselessprecursorisconsumedin

homo-geneousreactions.Fromthesurfacepointofview,itcorrespondsto

ahigherﬂuxofprecursor.Inversely,ahighertemperatureresultsin

thecombinationofalowerprecursorﬂuxandhigherbulkand

sur-facediffusionrates;bothmechanismsarebeneﬁcialtothesurface

smoothness.

The surface of the sample of experiment #2 is composed

of a dense ﬁlm with equiaxed grains on top of which

quasi-unidirectionalﬁlamentsarepointingawayfromthesurfaceplane.

In Fig. 3, similar ﬁlaments are found for experiments #2–4.

It was not possible to determine their structure and

compo-sition due to their size which is not compatible with SEM,

XRD or EPMA. Experiment #4 is the most promising

applica-tion wise:the ﬁlm is single-phase g-Al4Cu9 (within thelimits

of XRD characterization) and it shows large surface regions

like the one illustrated in Fig. 3 with no parasite ﬁlament

grains.

Thissampleprovides theﬁrstexperimental proofofconcept

accordingto which MOCVD co-deposition canform a complex

metallicalloycoatinginonestep.

3.3. Wettingof -Al4Cu9surfacesbywater

The useof MOCVD ﬁlms containing approximant phases as

buffer layers for the interfacial accommodation and adhesion

betweenametallicsubstrateandaquasicrystallinecoatingis

sub-jectedtotheirsurfaceenergy,whichinturnimpactstheirwetting

byliquids,e.g.water.Itiswellknownthatinambientair,typical

valuesforaquasicrystalofhighlatticeperfectionlikeanannealed

singlegrainicosahedralAl-Pd-Mnareintherange90◦_<_<₁₀₀◦_,

whereaspurealuminummetalshowsvaluesaround70◦ _[23]_._In

thiscontext,itisexpectedthatag-Al4Cu9 single-phasedsurface

shouldpresentacontactanglewithwaterbetween70◦ _and₉₀◦_.

Inordertoverifythispointaseriesofwettingexperimentswere

performedwithultrapurewater(resistivity18.2MOhm/cm)onthe

surfaceofsample#4(pureg-Al4Cu9)andonthesurfaceofapure

g-Al4Cu9bulksample.Beforewettingexperiments,thesurfaceof

thebulksamplewaspolishedwithSiCdisks.Threesurfaceﬁnishes

wereobtainedwith600,2400and4000grit,providingsurfaces

withmeanroughnessRaof246nm,16nmand7nm,respectively.

Fourmeasurementsofcontactanglewithwaterwereperformedon

eachsurfaceﬁnish.Theobtainedvaluesare82◦_,₈₄◦_,_and₈₇_(±2◦₎

fromtheroughertothesmoothersurface.Theseresultsconﬁrmthe

intermediatevalueofthecontactanglewithwateronthesurface

oftheg-Al4Cu9approximantbetweenthoseobtainedonthe

sur-faceofaluminummetalandoficosahedralcrystals.Theyarealso

consistentwithWenzel’stheory,followingwhichanintrinsically

hydrophilicsurfacebecomesmorehydrophilicwhenitsroughness

increases[24].Followingthesameprotocol,fourmeasurements

ofcontactanglewithwaterwerealsoperformedonthesurface

oftheMOCVDprocessedpureg-Al4Cu9ﬁlm.Theobtainedmean

valueis89(±2◦_)._Taking_into_account_that_R_a_of_the_latter

sam-plewasmeasuredtobe5nm,thisresultistobecomparedwith

theoneof87◦_obtained_on_a_bulk_g-Al₄_Cu₉_surface_with

compara-bleroughness.ItisconcludedthatMOCVDprocessedapproximant

phasespresentsimilarsurfacecharacteristicstothose of

equiv-alentbulkcrystals.Inamore prospectiveview,itispossibleto

engineerthesurfaceenergyofMOCVDAl-Cuﬁlmbyproper

selec-tionofdepositionconditionsleadingtoamixtureofintermetallic

phasesincludingapproximantones.

4. Conclusions

We present a one-step process for the deposition of Al-Cu

intermetallic coatings by metalorganic chemical vapor

deposi-tion.Dimethylethylaminealane(DMEAA),aliquidprecursor,and

copper(I)cyclopentadienyltriethylphosphine(CpCuPEt3),asolid

precursor,arechosenasprecursorsfortheco-depositionofAland

Cubecausetheycontainnooxygen,theyarechemicallycompatible,

andtheirdepositiontemperaturesaresimilar.

Theelementalcompositionoftheﬁlmsdoesnotfollowthatof

thegasphase.Thepresenceofhydrogengasplaysanimportantrole

intheco-depositionprocessbydisplacingthereactionequilibrium

ofDMEAAatthesurface,thereforedecreasingtheAlgrowthrate.

Additionally,whentheCpCuPEt3gasﬂuxishighenoughfor

achiev-ingAl/Cucompositionsofinterest,theexperimentalgrowthrateof

Al-Cuﬁlmsislimitedbytheadsorbedphosphineoriginatingfrom

CuCpPEt3decompositiononthegrowingsurface.Encouragingly,

solutionsareavailabletocircumventthoseproblemswhich are

attributedtothecontroloftheco-depositionbysynergeticsurface

phenomena.

Thesingle-phaseg-Al4Cu9isformedwhentheCucomposition

intheAl-Cuﬁlmsishighenough.Thesurfaceroughnessofsuch

ﬁlmsisverylow,withaRavalueequalto5nm.Wetting

measure-mentswithwateronthissurfaceprovideacontactangleof89±2◦_;

i.e.roughlythesameastheoneonabulkpolycrystallineg-Al4Cu9of

(8)

approximantphasepresentsimilarsurfacepropertieswiththose

ofequivalentbulkcrystalsandcan,thereforebeusedas

interfa-ciallayerbetweenmetallicsubstratesandquasicrystallinecoatings

accommodatingthemismatchofsurfaceenergybetweenthem.

ThepresentworkdemonstratesthatitispossiblebyMOCVD

toprocessinasinglestepandatrelativelylowtemperature,ﬁlms

containingintermetallicphasesincludingapproximantones.This

workmaybeconsideredasapreliminaryapproachpavingtheway

totheconformalcoatingofcomplexsurfacegeometriesbysuch

intermetalliccompounds.

Acknowledgments

WeareindebtedtoLyacineAlouiandtoMaëlennAufray,both

atCIRIMAT,Toulouseforperformingthe,andforadviceon,

con-tact angle measurements,respectively, to Philippe de Parseval,

Observatoire Midi-Pyrénées,Toulouse,forEPMAanalysisandto

Marie-CéciledeWeerd,InstitutJeanLamour,Nancy,forproviding

theg-Al4Cu9crystal.Thisworkwassupportedbythe6th

Frame-workEUNetworkofExcellence‘ComplexMetallicAlloys’(Contract

No.NMP3-CT-2005-500140),andbytheFrenchAgenceNationale

delaRecherche(ANR)undercontractNo.NT05-341834.Itwould

neverhavebeencompletedwithoutthepersistentsupportof

Jean-MarieDubois,CNRS,Nancy.

References

[1]J.-M.Dubois,P.Brunet,W.Costin,A.Merstallinger,Frictionandfrettingon quasicrystalsundervacuum,JournalofNon-CrystallineSolids334–335(2004) 475.

[2]J.-M.Dubois,M.-C.DeWeerd,J.Brenner,M.Sales,G.Mozdzen,A.Merstallinger, E.Belin-Ferré,Surfaceenergyofcomplex–andsimple–metalliccompounds asderivedfromfrictiontestinvacuum,PhilosophicalMagazine86(2006)797. [3]B.Ngoc,C.Geantet,J.Dalmon,M.Aouine,G.Bergeret,P.Delichere,S.Raffy, S.Marlin,Quasicrystallinestructuresascatalystprecursorsforhydrogenation reactions,CatalysisLetters131(2009)59–69.

[4]M.Bielmann,A.Barranco,P.Rufﬁeux,O.Gröning,R.Fasel,R.Widmer,P. Grön-ing,FormationofAl4Cu9onthe5foldsurfaceoficosahedralAlPdMn,Advanced EngineeringMaterials7(2005)392.

[5]T.Duguet,J.Ledieu,J.M.Dubois,V.Fournée,Surfacealloysasinterfaciallayers betweenquasicrystallineandperiodicmaterials,JournalofPhysics:Condensed Matter20(2008)314009.

[6]T.Duguet,S.Kenzari,V.Demange,T.Belmonte,J.M.Dubois,V.Fournee, Struc-turallycomplexmetalliccoatingsintheAl-Cusystemandtheirorientation relationshipswithanicosahedralquasicrystal,JournalofMaterialsResearch 25(2010)764–772.

[7]K.Tomoharu,K.Eiichi,T.Nobuyuki,N.Tadashi,Y.Hiroshi,O.Tomohiro, Met-alorganicchemicalvapordepositionofaluminum-copperalloyﬁlms,Japanese JournalofAppliedPhysics32(1993)L1078.

[8]K.Eiichi,K.Yumiko,T.Nobuyuki,O.Tomohiro,Interconnectionformation by doping chemical-vapor-deposition aluminum with copper simulta-neously: Al-CuCVD, Journal of TheElectrochemical Society 141 (1994) 3494–3499.

[9]L. Aloui, T. Duguet, F. Haidara, M.-C. Record, D. Samélor, F. Senocq, D. Mangelinck,C.Vahlas,Al–Cuintermetalliccoatingsprocessedby sequen-tialmetalorganicchemicalvapourdepositionandpost-depositionannealing, AppliedSurfaceScience258(2012)6425–6430.

[10]F.Senocq,A.Turgambaeva,N.Prud’homme,U.Patil,V.V.Krisyuk,D.Samélor, A. Gleizes, C. Vahlas, Thermal behaviour of CpCuPEt3 in gas phaseand

Cu thin ﬁlms processing, Surface and Coatings Technology 201 (2007) 9131–9134.

[11]A.E.Turgambaeva,N.Prud’homme,V.V.Krisyuk,C.Vahlas,Mass Spectro-metricmonitoringofthegasphaseduringtheCVDofcopperfromcopper cyclopentadienyl triethylphosphine,ChemicalVapor Deposition18(2012) 209–214.

[12]A.Ludviksson,D.W.Robinson,J.W.RogersJr.,Theinteractionof dimethylethy-laminealaneandammoniaoncleanandoxidizedAl(111):atomiclayergrowth ofaluminumnitride,ThinSolidFilms289(1996)6–13.

[13]J.-H.Yun,M.-Y.Park,S.-W.Rhee,Fouriertransforminfrareddiagnosticsofgas phasereactionsinthemetalorganicchemicalvapordepositionofaluminum fromdimethylethylaminealane,JournalofVacuumScience&TechnologyA: Vacuum,Surfaces,andFilms16(1998)419–423.

[14]T.C.Xenidou,N.Prud’homme,C.Vahlas,N.C.Markatos,A.G.Boudouvis, Reac-tionandtransportinterplayinAlMOCVDinvestigatedthroughexperiments andcomputationalﬂuiddynamicanalysis,JournalofTheElectrochemical Soci-ety157(2010)D633–D641.

[15]T.W.Jang, W.Moon,J.T. Baek, B.T. Ahn,Effectof temperature and sub-strate on the growth behaviors of chemical vapor deposited Al ﬁlms with dimethylethylamine alane source, Thin Solid Films 333 (1998) 137–141.

[16]T.C. Xenidou, A.G. Boudouvis, N.C. Markatos, D. Samélor, F. Senocq, N. Prud’Homme,C.Vahlas,Anexperimentalandcomputationalanalysisofa MOCVDprocessforthegrowthofAlﬁlmsusingDMEAA,SurfaceandCoatings Technology201(2007)8868–8872.

[17]T.C.Xenidou,N.Prud’Homme,L.Aloui,C.Vahlas,N.C.Markatos,A.G.Boudouvis, Shapeoptimizationofashowerheadsystemforthecontrolofgrowth uni-formityinaMOCVDreactorusingCFD-basedevolutionaryalgorithms,ECS Transactions25(2009)1053–1060.

[18]S.D.Hersee,J.M.Ballingall,Theoperationofmetalorganicbubblersatreduced pressure,JournalofVacuumScience&TechnologyA:Vacuum,Surfaces,and Films8(1990)800–804.

[19]C.Vahlas,B.Caussat,F.Senocq,W.L.Gladfelter,L.Aloui,T.Moersch,Adelivery systemforprecursorvaporsbasedonsublimationinaﬂuidizedbed,Chemical VaporDeposition13(2007)123–129.

[20]J.M.Vandenberg,R.A.Hamm,AninsituX-raystudyofphaseformationinCu-Al thinﬁlmcouples,ThinSolidFilms97(1982)313.

[21]H.G.Jiang,J.Y.Dai,H.Y.Tong,B.Z.Ding,Q.H.Song,Z.Q.Hu,Interfacialreactions onannealingCu/Almultilayerthinﬁlms,JournalofAppliedPhysics74(1993) 6165.

[22]F.Haidara,M.-C.Record,B.Duployer,D.Mangelinck,Investigationofreactive phaseformationintheAl–Cuthinﬁlmsystems,SurfaceandCoatings Technol-ogy206(2012)3851–3856.

[23]J.M.Dubois,Amodelofwettingonquasicrystalsinambientair,Journalof Non-CrystallineSolids334–335(2004)481–485.

[24]R.N.Wenzel,Resistanceofsolidsurfacestowettingbywater,Industrial& EngineeringChemistry28(1936)988–994.