3D打印英文文献 翻译.docx

JournaloftheEuropeanCeramicSociety31(2011)2543-2550Inkjetprintingceramics:FromdropstosolidB.DerbySchoolofMaterials,UniversityofManchester,OxfordRoad,ManchesterM139P1.,UKAvailableonline16February2011AbstractInkjetprintingisapowerfulInicrofabricationtoolthathasbeenappliedtothemanufactureofceramiccomponents.Tosuccessfullyfabricateceramicobjectsanumberofconditionsmustbesatisfiedconcerningfluidpropertiesanddropplacementaccuracy.Ithasbeenproposedthatfluidsareprintablewithinthebounds1<Z<10(whereZistheinverseoftheOhnesorgenumber)andtheselimitsareshowntobeconsistentwithceramicsuspensionsdeliveredbypiezoelectricdrop-on-demandinkjetprinters.Thephysicalprocessesthatoccurduringdropimpactandspreadingarereviewedandtheseareshowntodefinetheminimumfeaturesizeattainablef()ragivenprinteddropdiameter.Finallythedefectsthatcanoccurduringthedryingofprinteddropsarereviewed(coffeestaining)andmechanismsandmethodologiestoreducethisphenomenonarediscussed.Keywords:Inkjetprinting;Shaping;Drying;Suspensions1. IntroductionInkjetprintinghasmajorcommercialapplicationsingraph-icsoutputandotherconventionalprintingoperations.However;therehasbeendevelopinginterestinusinginkjetprintingtomanufacturecomponentswithapplicationsfor:displays,1plas-ticelectronics,2rapidprototyping,3tissueengineering,4andceramiccomponentmanufacture.5Asignicantandfundamen-taldifferencebetweenthesenewapplicationsandthemorewidespreadapplicationofprintingtextorimagesisthebehaiouroftheprintedinkdropletsontheprintedsubstrate.Mostimagesareconstructedbythedepositionofdiscretedropletsand,inordertooptimiseresolutionandcontrast,thesedropletsareiso-latedanddonotcontacteachother.Incontrast,manyofthenewapplicationsforinkjetprintingenvisagethemanufactureofcontinuous1-,2-,or3-dimensionalstructures(1-,2-,or3-D).Suchstructuresrequireacontinuousdistributionofmaterialandthisnecessitatescontactandadhesionbetweenindividualdropsafterprinting.Inkjetprintingconstructsobjectsbytheprecisionplacementofpicolitrevolumesofliquidandthustheinitialinterac-tionbetweenprintedmaterialandasubstrateisaliquid/solidinteraction.Ultimately,theprinteddepositundergoesasolid-ificationprocessthatcanoccurthroughsolventevaporation,temperatureinducedsolidification/gelationorchemicalreac-tion.Considerationsoftherelativetimescalesofdropspreadingandsolidificationindicatethattherewillbeasignificantperiodoftimeafterprintingwhenaliquidispresentonasurface6andthusthemorphologicalstabilityofcoallescingliquidfilmsmustbeexamined,asmusttheeffectsofthesolidificationprocess.Therehasbeenaconsiderablenumberofpublicationsontheuseofinkjetprintinginthemanufactureofceramics.7-17Thesepriorstudieshaveusedallinkjetdropgenerationtechnologies(continuous,thermaldrop-on-demandandpiezoelectricdrop-on-demand)tosuccessfullyproduceceramicobjectsusingbothsolventevaporationandphase-changesolidification.Industrialinkjetprintingtechnologynowusespiezoelectricdrop-on-demand(DOD)generationtechnologyandthisisthechosenmethodformostapplicationsinprintingfunctionalmaterials.ThephysicaloperationofthesedifferentprintingtechnologiesandthereasonsforthechoiceofpiezoelectricDODprintinghavebeendiscussedindetailelsewhere6,18;henceherewewillconfineourconsiderationstothistechnology.Wewillalsoonlyconsidertheprintingofceramicinksthatsolidifythroughsol-ventevaporation.Despiteearlierworkdemonstratingthatitispossibletosuccessfullyprintcmscaleobjectsusingawaxbasedphasechangeceramicink,11-13ceramicinkscontainrelativelylowvolumefractionsofsolidandthusthereisconsiderableshrinkageandpotentialfordistortionduringdewaxingandsintering.14Inordertofabricateceramicobjectsusinginkjetprinting,itisnecessarytosatisfyanumberofrequirements.Firstthereisaneedtoproducestableceramicsuspensionswithdefinedfluidpropertiessuchthattheycanbepassedthroughadropletgener-atorandformregulardrops.Second,thesesuspensionsneedtobedeliveredontoasubstrateorontoapreviouslyprintedlayerofsolidiedceramicink,withdropsinsufficientproximitytoeachothertoallowthemtointeractandformdesired2-Dfeatures.Next,theprintedceramicinkmustundergophasetransitiontoasoliddeposit.Finally,toproduce3-Dstructuresthedeposi-tionanddrying/solidificationprocessesneedtoberepeatedonalayerofpre-depositedanddriedmaterial.Herewewillconsidereachoftheserequirementsandtheiroptimizationforthedirectprintingofceramics.2. CeramicinksManufacturersofDODinkjetprintingequipmentnormallystatearangeofviscosityandsurfacetensionwithinwhichinksmaybesuccessfullyprinted.However;thisinformationisnor-mallyprovidedforthebenefitofformulatinggraphicsinksandmaynotbedirectlyapplicabletothedevelopmentofceramicinks.Thisisbecauseinkscontainingasignificantvolumefrac-tionofceramicparticlesinsuspensionhavemuchhigherdensityvaluesthantypicalgraphicsinks,whichtypicallyhavedensitiesintherange800-1000kgm-3andthebehaviourofafluidduringprintingdependsstronglyonitsinertialbehaviour.Thefluidrheologicalrequirementsforaprintableinkaredeterminedbythephysicsandfluidmechanicsofthedropgen-erationprocess.6,18ThebehaviouroffluidsduringinkjetprintingcanberepresentedbytheReynolds,WeberandOhnesorgenum-bers(Re,We,Oh):Re=vpa,(la)We=v2pa(Ib)Oh=We=(Ic)Re(pa)l2wherep,andarethedensity,dynamicviscosityandsurfacetensionofthefluidrespectively,visthevelocityandaisacharacteristiclength.FrommidentifiedtheOhnesorgenumber,Oh,astheappro-priategroupingofphysicalconstantstocharacterisedropgenerationinaninkjetprinter.19Ohisindependentoffluidvelocityandiscommonlyusedinanalysesdescribingthebehaviourofliquiddrops.However,inFromm,spublication,hedefinedtheparameterZ=1/Ohandfromasimplemodeloffluidflowinadropgeneratorofsimplifiedgeometry,hepro-posedthatZ>2forstabledropgeneration.19Reisextendedthisthroughnumericalsimulationandproposedthefollowingrange,10>Z>1,forstabledropformation.20IfZ<1,viscousdissipa-tionpreventsdropejectionfromtheprinterandifZ>10,dropletsareaccompaniedbyunwantedsatellitedrops.Jangetal.studiedtheDODprintabilityofanumberoffluidmixturesofethanol,waterandethyleneglycol.ThroughthistheyexploredarangeofvaluesofOhanddeterminedthattherangeofprintabilitywasFig. 1. Fromm,sparameterZ(Z=1/Oh)influencestheprintabilityoffluids.DashedlinesidentifythelimitsforprintabilityproposedbyReisetal.20Experi-mentalpointsareplottedforanumberofceramicsuspensions/inks:greysymbolsindicatesuccessfulinkjetprinting,blacksymbolsindicatethatnodropswereformed,andwhitesymbolsindicatethepresenceofsatellitedropsalongwiththemainprinteddrop.4<Z<14,21whichisVerysimilartothatdeterminedbyReis,snumericalsimulation.Thereisnowasubstantialbodyofliteraturedescribingtheinkjetprintingofanumberofceramicsuspensionsandotherfluidsfornon-graphicsapplications;unfortunatelynotallpublicationsreportsufficientinformationontherheologi-calpropertiesoftheceramicsuspensionstotestthisproposedcriterionforprintabilityinallcases.Fig.1presentssuchdatathateitherreportedthevalueofOh(orZ)orreportedsufficientdatathatitiseasilycalculated.TheverticaldashedlinesonthefigureatOh=1andOh=10representthelimitsforstableinkjetprintingcalculatedbyReis.20Theexperimentaldataispresentedfromeightfluidsystemswithagreysymbolindicat-ingthesuccessfulprintingofindividualdrops,ablacksymbolindicatesthatfluidswiththesepropertiescouldnotbeprinted,andfinallyawhitesymbolshowsthecaseswhereafluiddropwassuccessfullyejectedbutaccompaniedbyoneormoresatellitedrops.Itisusefultoseparatethesedataintotwosets:fluidsystems1-6weredeliveredusingpiezoelectricDODprinters,whilefluidsystems7and8weredeliveredusingathermalDODprinter.ThedataobtainedfromexperimentsusingpiezoelectricDODprintingshowsreasonablygoodagreementwithReis,smodel,howeverthatobtainedintheonestudyusingathermalDODprintershowsverypooragreement,17atleastwiththeupperboundforthepredictionoftheonsetofsatellitedropfor-mation.OzkolconsideredthatonereasonforthediscrepancybebveenReis,spredictionandtheirresultscouldbethediffer-enceinactuationbetweenpiezoelectricandthermalDODinkjetdropletgenerators.17Thehypothesisthatchangesinactuationexplainthedif-ferentbehaviourobservedbetweenthermalandpiezoelectricDODinkjetprintingissupportedbyanexperimentalstudyofdropandsatelliteformationinapiezoelectricDODprinterbyDongetal.22Theyfoundthatthedropformationmechanismandtheconditionsunderwhichagivenfluidformedsatellitesisalsocontrolledbytheshapeandamplitudeofthedrivingpulseappliedtothepiezoelectricactuator.ThedrivingpulseinDODprintingisalsoknowntocontrolboththesizeoftheejecteddropanditsvelocity.12,22,23Reisdemonstratedthatfortheformationofdropsusinghighlyloadedceramicsuspensions,acousticphenomenaareimportantandthattherearemaximaininkjetperformancethatcorrelatewithacousticresonancesintheprinthead.23TheseareparticularlyimportantconsiderationsgiventhattypicalindustrialDODprintheadsoperateinthekHzregime.Otherstudiesofinkjetprintingforapplicationsingraphicsalsoemphasisetheimportanceofacousticphenom-enaandtheneedforthesetodampbeforethedropgeneratorisrefilledpriortodeliveringsubsequentdrops.18IndeedtheshapeandformoftheactuatingwaveformisconsideredanimportantaspectofthedesignofpiezoelectricDODprintingsystems.However;fromFig.I9wecanseethatforthestudiesthatusedpiezoelectricDODprinters,Reis,scriterionforaprintablefluid20seemstoshowreasonableagreementwithdataanditisalsoinbroadagreementwiththeonlyexplicitstudyofinkjetprintabilityoffluidsbyJangetal.21Thusdespiteapossibleoversimplificationoftheconditionsthatleadtotheformationofsatellitedrops,wesuggestthecondition10>Z>1(whereZ=lOh)canbeusedasaguidetothedevelopmentoffluidsforinkjetprinting.ThesuitabilityofafluidforinkjetprintingcanberoughlyassessedbyitsOhnesorgenumber.Howeverthereareotherlim-itsoffluidbehaviourthatimposeadditionallimitstopracticaldropgeneration.Inordertogenerateasmallradiusdrop,thesur-facetensionandassociated1.aplacepressuremustbeovercomebeforeadropcanbeejectedfromaprinter.DuineveldproposedthatthiscanbedescribedbyaminimumvalueoftheWebernumber,We>4,belowwhichthereisinsufficientfluidflowtoovercomesurfacetension.24Afinalboundtoprintabilityisgivenbytheonsetofsplashingthatoccursifadrophitsthesubstratewithvelocityaboveacriticalthreshold.FromtheworkofStoweandHadfield,25thisoccurswhenWel2Rel4>50.Theselimit-ingboundsdefinearegionoftheparameterspaceofWeandRe,withinwhichDODinkjetprintingispossible.5,6Fig.2showsFig. 2. Inkjetprintingispracticalforalimitedrangeoffluidsandprintingcon-ditions.ThisisillustratedhereinaparameterspacedefinedbyaxesofReynoldsandWebernumbers.BasedonadiagramoriginallypublishedinRef.5.thisparameterspaceandtheregionsuitableforDODinkjetprinting.Dropvelocityincreasesdiagonally,asindicatedandhaslowerandupperboundsthataredefinedbytheappropriatelimitsofdropejectionandsplashing,orthogonaltovelocityistheOhnesorgenumber;whichdefinesthelimitsofthefluidprop-erties,thusFig.2canbeconsideredrepresentingaguidetothelimitsofbothfluidcharacteristicsanddropdynamicsconsistentwiththepracticaluseofpiezoelectricDODinkjetprinting.3. Dropimpact,spreadingandcoalescenceAsdiscussedearlier,animportantaspectofinkjetprintinginmanufacturingtechnologyistheprocessbywhichadjacentdropsinteracttoformasolid.Inallcasestheliquiddropwillinteractwithasolidsubstrate.Followingdepositiontherewillbeaperiodwhenthedrop,sshapeiscontrolledbyfluidpro-cessespriortosolidification.Thusanimportantconsiderationistheappropriatetimeconstantsthatapplytothemechanismsofsurfacespreadingandsolidication.Hereweareconfiningourdiscussiontosolidificationthroughevaporation.Giventhatdropletsolidificationtimescalesarenormallyintheregimeofaround1sanddropletdepositionratesare>1kHz,weneedtoconsidertheinteractionbetweenmanyliquiddropletsonthesurfaceofthesubstrate.Itispossibletouseinterlacingandsequentialprintingpassestodepositisolateddrops,allowthemtosolidifyandthenfillinthegapstoproduceaprintedplane.However,thismethodologyproducesanirregulardepositwithpoorsurfaceroughnessforeachprintedlaye9withaconse-quentriskofdefectsfrompoorpenetrationoftheliquid.Ifprintingoccurswithappropriatedropspacingtoallowover-lapbeforesolidication,theinteractionbetweenadjacentliquiddropsandtheconsequentinfluenceofsurfacetensionwilltendtoproducesmoothsurfacesandeliminatepossibledefectsbetweensolidifieddrops.Whenaliquiddropimpactsaplanarsubstrateitwilldeformandspreadtocoverthesubstrate,ultimatelyachievinganequi-Iibriumsessiledropconfiguration.YarinhasrecentlyreviewedtheimpactofdropsoverasizeandvelocityrangethatintersectsthoserelevanttoDODprinting.26Thetypicalrangeofdropsize(radiusfrom5to50mm)andvelocity(1<v<10ms-l)issuchthattheinitialdeformationofthedropwillbecontrolledbydynamicimpactandviscousdissipationprocesses.6,18,26However,thisinitialstageofdropdeformationisexpectedtohavefinishedafterafewmsandsubsequentspreadingtoequilibriumwillbedrivenbycapillaryforces.27Aschematicrepresenta-tionofthetimescalesassociatedwithdropdeformationafterimpactispresentedinFig.3.Thedynamicprocessesofdropimpactoccuroveratimescaleofms.26Firstthedropdeformsonimpactwithitskineticenergyconvertedintonewsurfaceareaasthedropdeforms,someenergyisabsorbedthroughviscousdissipation.Iftheimpactconditionsaresuchthatsplashingdoesnotoccur(asisnormalwiththeconditionsofinkjetprinting),thedropmayrecoilafterexpansionandoscillatebrieflydissipatingenergy.Meanwhiletheprocessofcapillaryspreadingwilloccur,thishasatimescaleofmsfordropdimensionsconsistentwithinkjetprintingandthefinalequilibriumdropshapeisnormallycontrolledbythisprocess.2546Fig.3.Schematicillustrationofthetimescalesappropriatetotheprocessesofdropdeformationandspreadingonasubstrateafterimpact.Axesofdiameterandtimearetoarbitrarynon-linearscales.ReproducedwithpermissionfromRef.6.Onceequilibriumhasbeenreached,thedropcanbemodelledasasphericalcapbecausetheBondnumberissufficientlylowthatwemayignoretheinfluenceofgravity.Inwhichcasetheequilibriumcontactdiameterofthedrop,deqm,canbecalculatedwith(2a)deqm=d,where=2taneqm3+tan2eqm-1/3(2b)wheredisthediameterofthedropinflightandeqmistheequilibriumcontactangle.Foranisolateddropofpuresolvent,wewouldexpectthedropdiametertodecreaseandthecontactlinetoretractataconstantrecedingcontactangleduringevap-oration.However,forthecaseofparticlesinsuspension,thebehaviourofaliquiddropisdifferent.Solventevaporationdoesnotoccuruniformlyfromthesessiledrop.Atlowcontactangles,thefluidclosetothecontactlineisadjacenttoalargedrysurfaceandthisenhancesthetransportofthesolventvapourpromotingfasterevaporation.Thisleadstoaringofparticlescomingoutofsuspensionandthepresenceofthisdrieddepositpinsthecontactlineandpreventsitretracting.Thiscontactlinepinningresultsintherecedingcontactangledecreasingassolventisremoved.Itcanalsoresultinaflowofparticlestothecontactline,leadingtosuspensionsegregationandaringdeposit;thisisaphenomenonknownasthecoffeestaineffect.28Wewillreturntothecoffeestainphenomenonlaterinthisarticle.Oneeffectofcontactlinepinningduringdryingisthatthefootprintorequi-libriumdiameterofthespreaddropofceramicinkwilldefinethediameterofthedrieddepositonthesurfaceaftersolventevaporation.Inordertoprinttwo-dimensionalpatternsitisnecessarytoallowadjacentdropletstointeractandcoalesce.Itisadvanta-geousforthesedropstointeractwhileinaliquidstatebecausesurfacetensionforceswillresultinasmo