micromachines Article High-Precision Solvent Vapor Annealing for Block Copolymer Thin Films GunnarNelson,ChloeS.Drapes,MeaganA.Grant,RyanGnabasik,JeffreyWong andAndrewBaruth* ID DepartmentofPhysics,CollegeofArtsandSciences,CreightonUniversity,2500CaliforniaPlaza, Omaha,NE68178,USA;[email protected](G.N.);[email protected](C.S.D.); [email protected](M.A.G.);[email protected](R.G.);[email protected](J.W.) * Correspondence:[email protected];Tel.:+1-402-280-2644 (cid:1)(cid:2)(cid:3)(cid:1)(cid:4)(cid:5)(cid:6)(cid:7)(cid:8)(cid:1) (cid:1)(cid:2)(cid:3)(cid:4)(cid:5)(cid:6)(cid:7) Received:19April2018;Accepted:25May2018;Published:29May2018 Abstract: Despiteitsefficacyinproducingwell-ordered,periodicnanostructures,theintricaterole multipleparametersplayinsolventvaporannealinghasnotbeenfullyestablished. Insolventvapor annealingathinpolymerfilmisexposedtoavaporofsolvent(s)thusformingaswollenandmobile layertodirecttheself-assemblyprocessatthenanoscale. Recentdevelopmentsinboththeoryand experiments have directly identified critical parameters that govern this process, but controlling them in any systematic way has proven non-trivial. These identified parameters include vapor pressure,solventconcentrationinthefilm,andthesolventevaporationrate. Toexploretheirrole, a purpose-built solvent vapor annealing chamber was designed and constructed. The all-metal chamberisdesignedtobeinerttosolventexposure. Computer-controlled,pneumaticallyactuated valvesallowforprecisiontimingintheintroductionandwithdrawalofsolventvaporfromthefilm. Themassflowcontroller-regulatedinlet,chamberpressuregauges,insituspectralreflectance-based thicknessmonitoring,andlowflowmicrometerreliefvalvegivereal-timemonitoringandcontrol duringtheannealingandevaporationphaseswithunprecedentedprecisionandaccuracy.Thereliable andrepeatablealignmentofpolylactidecylindersformedfrompolystyrene-b-polylactide, where cylinders stand perpendicular to the substrate and span the thickness of the film, provides one illustrativeexample. Keywords: blockpolymers;self-assembly;thinfilms;solventvaporannealing;nanolithography 1. Introduction Techniquestoachieveperiodicnanostructuresviatraditional“topdown”methods,including photolithography, have become increasingly challenging within the semiconductor industry. Ultra-small feature production is approaching fundamental resolution limits (193 nm ultraviolet (UV)lithography,forexample,recentlyreachingsub-30nmfeatures)[1–11]. Onepromisingstrategy is investigating “bottom up” approaches that rely on nanoscale self-assembly. In 2007, directed self-assembly was first considered as a potential scaling solution, according to the International TechnologyRoadmapforSemiconductors(ITRS)[11]. Intheir2013report,thedirectedself-assembly ofcomplexstructureswithlowannealtime,lowdefectdensity,andhighreproducibilitywasidentified asoneofthe“GrandChallenges”toextendMoore’slaw[11]. Thedirectedself-assemblyofblock polymer(BP)thinfilmshasbecomeaparticularlystrongcandidatetoachievesub-20nmdimensions, where the size and morphology is controlled by varying the molecular weight of the constituent polymerblocks. Duetothoroughinvestigationsoverdecades,theBPresearchcommunitygenericallyconsiders the bulk behavior of many BPs well known [12,13]. Bulk morphologies are characterized by the Micromachines2018,9,271;doi:10.3390/mi9060271 www.mdpi.com/journal/micromachines Micromachines2018,9,271 2of21 Flory–Hugginsinteractionparameter(χ),thedegreeofpolymerization(N),thevolumefractionsof theconstituentblocks(f),andtheparticulararchitecture[13]. Furthermore,muchisknownabout bulkbehaviorunderavarietyofstimuli,whetherthermal,solvent,ormechanical[12]. Inparticular, measuredorknownquantitiescanwellpredictthealignmentofstructuresandmorphologieswithin thebulk.Onthecontrary,theconfinementofathinfilmandtheassociatedsurfaceenergycontributions introducesadditionalconfounders[14,15]. Inparticular,thecomplexityofsurfaceinteractions(both with the substrate and free surface) impose new, often asymmetric, boundary conditions [16,17]. Regardless,techniquesdesignedtopromoteorderingofBPthinfilmshaveprogressedrapidly,driven byapplicationsrequiringlong-rangelateralordering,uniformityinfeaturesizeandhighplacement precision. MostofthehistoricalapproachestoBPorderinghavebeenlargelyunpredictableandare oftensloworenergy-intensive: includingthermalannealing[18,19],electricfieldalignment[20,21], andincorporationoflowsurfaceenergymidblocks[22,23]. Morerecently,theuseofpre-patterned substrates (chemical or topographical) to act as a guide for BP assembly has been successfully incorporatedbutcanbetime-intensiveandinvolvesmultiplelithographicapproaches[24]. Therefore, itisnotnecessarilycost-effectiveforhigh-throughputapplications. Asaresult,amaturingtechnique issolventvaporannealing(SVA)[25]. SVA was originally introduced as an alternative to thermal annealing for BPs exhibiting thermaldegradation,problematicthermally-driventransitions,orslowdynamicsduetohighmolar mass[26–28]. TheinterestinSVAofBPshasgrownwellbeyondthisinrecentyears. Ithasbeenshown tooptimizeorganizationquicklyduetotheincreasedchainmobility,apossiblydecreasedχ(dependent onsolventpolarity), andtunablesurfaceenergies[25]. Inthisprocess, aBPfilmisexposedtothe vaporsofoneormoreorganicsolvents,offeringdirectcontroloverlyotropictransitions(cylindersto spheres,forexample)whileinthesolvatedstateaswellasduringevaporation[29,30]. Thistechnique alsohastheabilitytoreducedefectdensitydramatically[31],whileimprovinglateralordering,both at the free surface and into the bulk of a film. This ordering is achieved more quickly (by several ordersofmagnitude)andcompletelythanpreviousmethods[32]. Althoughmuchisknownaboutthe interactionsofBPswithsolventinbulk[33],theeffectsinthinfilmsexhibitdifferentbehaviorsdueto thepresenceofconfiningsurfacesandthedynamicexchangewiththesolventvaporatmosphere[17]. Thus,thistechniquecontinuestosufferfromreliabilityproblemsandnostandardizedmethodshave becomeapparent. Itisbecomingincreasinglyclearthatacontinuedunderstandingofthespecific orderingmechanismsofaBPsystemisparamount,wherethequestforgenericunderstandingofany BPthinfilmsystemremainselusive. OfcriticalimportancetomanytechnologicalapplicationsistheabilitytodirectBPthinfilms toformcylindersthatstandperpendiculartothesurface,traversethroughthethicknessofthefilm, andlaterallypackwithhexagonalorder. Arecentlydemonstratedapproachtonanolithographyofa magneticthinfilm,forexample,utilizedsuchBP-basedlithographymasksusingaDamascene-like approach. This approach was able to synthesize hexagonally-packed magnetic nanodots with a diameter of 19 nm with high fidelity and retention of robust ferromagnetism [34]. Furthermore, this approach achieved diameter control, down to 14 nm, and the potential for high-temperature processingwithanadditionalatomiclayerdepositionstepofZnO[35]. Thesetechniquesrelyonthe verticalalignmentofcylinders. Onthecontrary,manycurrentadvancesinBPalignment,including importantworkinsolventconcentrationgradients[36],havefocusedsolelyonthefreesurfaceofthe film. Thisisinsufficienttobeadirectsubstitutionformosttraditionallithographicapproaches[37]. Recentresultsrevealthatoptimizedorderingofhexagonally-packedcylinderscanpotentiallyoccurin secondsandextendthroughthefilmusingSVA,butthisapproachcanbesomewhatunreliable[38], wherenear100%successofformingthedesiredmorphologyhasyettobeachieved. Advancesin computersimulationaswellasinsituX-rayandneutronscatteringcontinuetobothfurtherunderstand thisprocessandimproveuponit[39–41]. Whilesuchanunderstandingoftheself-assemblybehavior iscriticaltotheadvancementofthisfield,utilizingastatisticalapproachtoquantifyingSVAwithlarge samplesets,withagoalofovercomingreliabilitybarriers,requirestheabilitytoidentify,measure Micromachines2018,9,271 3of21 anddevelopcontrolsforallofthepertinentvariableswithreliableprecision. Thisincludeschamber pressure,solventexposuretime,solventconcentrationinthefilm,solventevaporationrate,solvent purityorcombinations,ambienttemperature,sampleandsolventtemperature,humidity[42],andfilm thicknesstonameafew. Theseareaddressedinthepresentmanuscript,withaprimaryfocusonthe roleofcontrollingchamberpressureandsolventevaporationratesatfasttimescalesof~10ms.Wenote that this chamber does not actively control sample [41] or solvent temperature [43–45], or include multiplesolvents[46],whichhavebeenshownimportanttocontrollingannealingkinetics. So,these results work to keep these parameters as consistent as possible. To that end, we present here our strictannealingprotocolsandourclimate-controlledSVAchamberwithcomputer-controlledsolvent vaporflowandpressuremanagementandinsituspectralreflectance-basedsolventconcentration (φ)measurements. Thischamberallowsustofixpotentialvariableswhileinvestigatingonlyone. ThisleadstounprecedentedcontrolovertheSVAprocesswithagoalofsystematicstudieswithhigh reliabilityandrepeatabilitythatmayhaveadvantagesasBPSVAalignmentmovesfromitscurrent researchphase(relyingheavilyoninsituX-rayandneutronscattering)intoscaled-upprocessing. One criticalandunprecedentedadvancementistheabilitytostabilizesolventconcentrationwithinaBP thinfilmforanarbitrarilylongtimeatauser-definedchamberpressure,anecessaryprerequisitefor anytemporalstudiesofcrystallization. Todate,therearethreeprimarymethodsofcontrolledSVAintheliteraturewithdifferinglevelsof complexity,asillustratedinFigure1. TheseprocesseswererecentlyreviewedbyPosseltetal. andGu etal.[32,47],whereeachmethodtypicallyincorporatesanopticalinterferometertomonitorthickness (i.e.,swellingduetosolventuptake)inrealtimethatisplacedabovethechamberandshinesdownon thesamplesurface. Briefly,in“jarannealing,”depictedinFigure1a,asolventreservoirisplacedina sealedvesselwiththeBPfilm.Thesolventvaporpressure,andthusthefilmswelling,isparameterized by the ratio of liquid solvent surface area to the chamber volume [48]. The solvent evaporation is eitherdoneviaopeningthelid,whichisdifficulttoquantify,orintroducingaleakuntilallsolvent hasevaporated[48,49]. AnaturalextensionofthisSVAmethodistheinclusionofinlet/outletflow lines,asseeninFigure1b,wheretheincorporationofaninertgasflowcanservetomodifythevapor pressureandBPfilmswellingmoredirectly[50]. However,thereisstillonlylimitedcontroloverthe solventevaporationrate. Thenextextensionisdisplacingthesolventreservoirfrominsidethesealed chamberintoaseparatesealedreservoir,asshowninFigure1c. InthisSVAmethod,solventiscarried inthevaporphasebyacarriergasintothechamber. Boththeabsolutepressureandvaporpressure insidethechamberarecontrolledviatheflowratesofthevaporlineandasecondinletforaninertgas stream. ThisgivessignificantdynamiccontrolofBPfilmswelling,althoughthehighestachievable vaporpressureisreducedcomparedtoFigure1a,bduetoadilutionofvaporfromthecarriergas. ThesolventevaporationrateiscontrolledinasimilarfashiontothemethoddepictedinFigure1b, butwithoutthepresenceofthesolventreservoir,thusofferingmorecontrol. Similarmethodshave proliferated[30,51–53],butonlyfewutilizereal-timecomputercontrol[41,45]andthereiscurrently aninabilitytomaintainafixedsolventconcentrationforanarbitrarilylongtimescale. Thepresent manuscriptdescribesanewevolutionintheSVAmethodthatbuildsontheseexistingmodels,adding additionalcontroloversolventintroductionintoandevaporationoutofaBPfilmwithahighdegree ofreliabilityandreproducibilitybycontrollingthevaporpressurethroughinlet/outletflows. Several recent in situ studies utilizing Grazing Incidence Small Angle X-ray and Neutron Scattering(GISAXSandGISANS)servetomotivatetheabilitytostabilizeaBPfilmwithaspecific solvent concentration [41,54]. These studies have shown to be critical in identifying the role of the order-to-disorder transition in directed self-assembly, as well as the role of solvent polarity and temperature. Their results could be further extended with an increased level of control over solvent concentration during the anneal and during the solvent evaporation. For example, Gu et al. show a high degree of ordering in poly(styrene)-b-poly(2-vinyl pyridine) using GISAXS with an increasing solvent concentration [30]. Upon rapid evaporation, scanning electron micrographs (SEM) of the resultant films exhibit long correlation lengths of in-plane cylinders. Furthermore, Micromachines 2018, 9, x 3 of 20 solvent evaporation rate, solvent purity or combinations, ambient temperature, sample and solvent temperature, humidity [42], and film thickness to name a few. These are addressed in the present manuscript, with a primary focus on the role of controlling chamber pressure and solvent evaporation rates at fast time scales of ~10 ms. We note that this chamber does not actively control sample [41] or solvent temperature [43–45], or include multiple solvents [46], which have been shown important to controlling annealing kinetics. So, these results work to keep these parameters as consistent as possible. To that end, we present here our strict annealing protocols and our climate-controlled SVA chamber with computer-controlled solvent vapor flow and pressure management and in situ spectral reflectance-based solvent concentration (φ) measurements. This chamber allows us to fix potential variables while investigating only one. This leads to unprecedented control over the SVA process with a goal of systematic studies with high reliability and repeatability that may have advantages as BP SVA alignment moves from its current research phase (relying heavily on in situ X-ray and neutron scattering) into scaled-up processing. One critical and unprecedented advancement is the ability to stabilize solvent concentration within a BP thin film for an arbitrarily long time at a user-defined chamber pressure, a necessary prerequisite for any temporal studies of crystallization. To date, there are three primary methods of controlled SVA in the literature with differing levels of complexity, as illustrated in Figure 1. These processes were recently reviewed by Posselt et al. and Gu et al. [32,47], where each method typically incorporates an optical interferometer to monitor thickness (i.e., swelling due to solvent uptake) in real time that is placed above the chamber and shines down on the sample surface. Briefly, in “jar annealing,” depicted in Figure 1a, a solvent reservoir is placed in a sealed vessel with the BP film. The solvent vapor pressure, and thus the film swelling, is parameterized by the ratio of liquid solvent surface area to the chamber volume [48]. The solvent evaporation is either done via opening the lid, which is difficult to quantify, or introducing a leak untiMl iacrlolm saochlinvees2n0t1 8h,9a,s27 e1vaporated [48,49]. A natural extension of this SVA method is t4hoef 2i1nclusion of inlet/outlet flow lines, as seen in Figure 1b, where the incorporation of an inert gas flow can serve to modifrye ctehnet rvesaupltosrin pdriceastseuthraet aapnpdr oBaPch ifnilgma ssowlveelnltincogn cmenotrraet idonirceocntslyis te[5n0tw]. iHthothweeovrdeerr,- ttoh-deriseo ridse rstill only limited ctornantrsoitli oonvleera dthseto stohlevelanrgt eesvtacpororrealattiioonn lreantget.h Tsh[3e2 n].eTxht iesxstoelvnesniotnco insc denistrpaltaiocninisg atlhsoe csoonlvsiestnetn rteservoir with enhanced defect removal [31]. Annealing at solvent concentrations above the order-disorder from inside the sealed chamber into a separate sealed reservoir, as shown in Figure 1c. In this SVA transitionwill,instead,resultinapotentiallydisorderedfilm,aspreviouslyshown[34]. Inaddition, method, solvent is carried in the vapor phase by a carrier gas into the chamber. Both the absolute Sintureletal.revealedsimilarorderinginasolvatedpoly(styrene)-b-poly(lactide)film,whereGISAXS pressure and vapor pressure inside the chamber are controlled via the flowrates of the vapor line wasmonitoredwithincreasingsolventconcentration[29]. Thedatashowsthatthecorrelationlength and a secoofnpder pinenledti cfuolra ralyn ailnigenretd gcaysl isntdreerasmin.q Tuhiciksl ygidvrieesd sfiilgmnsifisiclaarngte dstyfonraamhiicg hcosonlvtreonlt ocofn BcePn tfrialmtio nswelling, although( btehloew hitghehoersdt earc-tho-ideivsaorbdleer vtraapnsoirti opnr)e.sFsinuarlely i,sB erreedzukicneedt aclo.mrecpeanrtleyds htoow Feidguhorew 1mai,xbe ddusoel vteon ats dilution andtheuseofelevatedtemperatureinacontrolledSVAchamber(similartoFigure1c)furtherenhance of vapor from the carrier gas. The solvent evaporation rate is controlled in a similar fashion to the controloverfinalmorphologies. TheseGISAXSstudies,whileextremelyinsightful,relyonsomewhat method depicted in Figure 1b, but without the presence of the solvent reservoir, thus offering more uncontrolled solvent exposure methods, as presented in Figure 1. In each case, no static solvent control. Similar methods have proliferated [30,51–53], but only few utilize real-time computer concentration(oftendescribedasswellingratio)wasachievedoveranyappreciableamountoftime. control [S4u1c,4h5c]o natnrodl itshaeprere riesq cuuisritreefnotrlyan aynte imnpaobriallitsytu tdoy mofaoirndteariinng ak ifniexteicds asnodlvfuerntht ecroonpctiemnitzraattioionn for an arbitrariloyf alnonangn teiamlineg spcraolteo.c oTl.hMe opvrinegsefonrtw maradn,eusspcerciipaltl ydwehscenricboenss iad enrienwg secavlionlguutipotnh eisne ttehceh nSiqVuAes method towardsmassproduction,creatingahighlycontrolledSVAchambercouldassistinavoidingtheuse that builds on these existing models, adding additional control over solvent introduction into and ofcontinuedinsituX-rayorneutronscatteringtechniquesthatarenoteasilyextensibletoscalingup. evaporation out of a BP film with a high degree of reliability and reproducibility by controlling the Instead,ahighlycontrolledSVAchambercouldestablishreliabilityandrepeatabilitythatincorporates vapor pressure through inlet/outlet flows. thecriticalparametersthathavebeenestablishedfromthesescatteringtechniques. Figure1.Primarymethodsofsolventvaporannealing;(a)“jarannealing,”whereasealedchamber Figure 1. Primary methods of solvent vapor annealing; (a) “jar annealing,” where a sealed chamber containsthesampleandasolventreservoir;(b)“jarannealing”withinlet/outletlinesforinertgasflow; contains the sample and a solvent reservoir; (b) “jar annealing” with inlet/outlet lines for inert gas (c)solventvaporflowviaacarriergasthroughinlet/outletlines.Blockpolymerfilmsaredepictedin flow; (c) bsluoel,vweintht lviqaupidorso flvloenwt dvepiaic tae dcianrgrrieeern .gas through inlet/outlet lines. Block polymer films are depicted in blue, with liquid solvent depicted in green. Inadditiontoacontrolledstaticsolventconcentration,theevaporationrateservesasanother complicating factor to be addressed. Solvent evaporation rate has been shown to strongly affect the surface morphology of BP films [52], where the evaporation has been shown to produce an orderingfrontthatpropagatesthroughthefilmfromthefreesurfacetothesubstrate[55]. Thetiming of solvent removal dominates this effect. Enviable results were reported in 2004 from Kim et al. revealingdefect-freeorderingofcylindersatthefreesurfaceoverlargelaterallengthscales,where GISAXS data aimed to expose the role of this ordering front [55]. The ordering depends strongly on the exact trajectory through the BP phase diagram as solvent is removed [56], with changes in polymerdynamicsandorder–ordertransitionsservingaspotentialcomplicatingfactors[57]. Assuch, conflicting results dictate that either slow [51,52] or fast [58] evaporation times may lead to ideal morphologies. Recently,studiesofsolventconcentrationprofiles[58]havesuggestedthatthecylinder growthrateduringevaporationisaproductofthepolymerchainmobilityandthedrivingforceto phaseseparate. DynamicalfieldtheorysimulationsfromParadisoetal. furtherexposetheroleof rapidsolventevaporationandsegregationstrength(χ)oncylinderorientation[57]. Itisincreasingly clearthattheabilitytosystematicallyremovesolventfromthefilm,includingatrapidtimescales,is importanttounderstandingfinalfilmmorphologies. Inparticular,revealinghowagivenmorphology propagates into the film during solvent evaporation, which is a necessary requirement for many nanolithographicapplications. Itisimportanttonotethattheroleofsolventremovalonrapidtime scales(lessthan100ms)hasnotbeensuccessfullymonitoredusingGISAXS,whereintegrationtimes Micromachines2018,9,271 5of21 canbesomewhatlong(1–10s)[32]; however, recentuseofcharge-coupleddevice(CCD)cameras havebeenshowntobeeffectiveforstudyingfastkinetics[47]. Thisopensthedoorforquantifying the effects of solvent removal in real time. Presently, in many of these studies, combining known structuralpropertiesfromGISAXSinthesolvatedstatewithfinalmorphologiesimagedwithatomic forcemicroscopy(AFM)orSEMfollowingrapidevaporation,asaddressedabove,havebeenusedto revealimportantrelationshipsbetweenevaporationtrajectoriesandfinalmorphologies. Thepresent manuscript describes a systematically improved SVA method that offers a potential to elucidate multiplekeyvariablesintheSVAprocessandoffersalevelofcontrol,reliabilityandreproducibility toenhancetheunderstandingoforderingkineticsbothduringannealingand,critically,evaporation. Specifically,thisimprovedSVAmethodmonitorsandcontrolschamberpressure,solventexposure time,solventconcentrationinthefilm,solventpurityandsolventevaporationrate,whileconcurrently monitoringambienttemperatureandhumidity. 2. MaterialsandMethods 2.1. SolventVaporAnnealing Optically polished substrates were single-crystalline silicon <111> wafers (n-type, 5 Ω·cm) with a native oxide layer (University Wafer, South Boston, MA, USA). No attempt was taken to remove the native oxide. Toluene (ACS Certified), tetrahydrofuran (THF) (Certified, contains 0.025% butylated hydroxytoluene as a preservative), acetone (99.5%), methanol (ACS Certified), and sodium hydroxide granules were all purchased from Fisher Chemical, Pittsburgh, PA, USA. 1,1,1,3,3,3-Hexamethyldisilazane(HMDS)(98%)and3Å,4to8mesh,3333werepurchasedfromAcros Organics,Belgium,WI,USA. Construction of the solvent vapor-annealing chamber (SAC) required chemically inert stainless-steel tubing and valves with Swagelok tube fittings. Pneumatically actuated stainless steelballvalves(1/4”and3/4”)utilizingDowCorningM111(heavy-consistencydimethylsilicone compound)lubricantandModifiedPTFEpacking(SS-T12-S-065-20andSS-45S8-33C),stainlesssteel quarter-turnplugvalves(1/4”)withKalrezO-rings(SS-4P4T),astainlesssteellow-flowmetering valve(1/4”)utilizingaKalrezO-ring(SS-SS4-KZ-VH),astainlesssteellow-pressure(5psig/34.5kPa) unfilledpressuregauge(LP1-SS-254-5PSI),astainlesssteelunfilledpressuregauge(60psig/413.7kPa, PGI-63C-PG60-LAOX),and1/4”and3/4”stainlesssteeltubing(SS-T4-S-035-20andSS-T12-S-065-20, respectively) were purchased from the Swagelok Company, Saarland, OH, USA. The body and bubblerportionsoftheSACutilizedvacuum-gradeConflatflangeswithoxygen-freecoppergaskets. Stainless steel tees (275-150-CFT), 6-way cube (275-CUBE-OS), and all Conflat blanks (275-000-T) andSwagelokadaptorswerepurchasedfromLDSVacuumProducts,Inc.,Longwood,FL,USA,or modifiedin-house. ConflatflangedzerolengthdeepUVquartz(CorningHPFS7980FusedSilica) viewportwaspurchasedfromtheKurtJ.LeskerCompany,JeffersonHills,PA,USA(VPZL-275Q).Dry N inletwasachievedwithaDrieritegaspurifier,andwascontrolledviaanApex500SCCMmass 2 flowcontrollerwithRS-232digitalcontrol(SchoonoverInc.,Canton,GA,USA).DryN pressurefrom 2 thegastankweremonitoredandcontrolledwithadual-stagegasregulator(0–344.7kPa)andflowrates weremonitoredandcontrolledwithaPanel-MountFlowmeter(OMEGAEngineering,Stanford,CT, USA)forair,withabrassvalve,withtwodifferentflowranges(3–30SCFHand30–300SCFH)from McMaster-Carr, Elmhurst, IL, USA. Pneumatic valve control was achieved with 3-way 1/4” NPT, normallyclosed,120V,100psisolenoids,variousOne-touchfittings(1/4”)and1/4”nylontubing purchased from the Swagelok Company, Saarland, OH, USA. Electric control of pneumatics was achievedwithaNationalInstrumentsUSB,8input,12-Bit,10kS/s,MultifunctionDAQ;25A,250V solidstaterelays;andLabVIEW2016software(NationalInstruments,Austin,TX,USA).Finally,film thicknesswasdeterminedinsituwithspectralreflectanceviaaFilmetricsF20-UV,SanDiego,CA, USA,general-purposefilmthicknessmeasurementsystemwithbothhalogenanddeuteriumsources. Micromachines2018,9,271 6of21 2.2. SynthesisofPoly(Styrene)-Block-Poly(Lactide) Thesynthesisofpoly(styrene)-block-poly(lactide)(PS-b-PLA)isdescribedfullyelsewhere[34]. Succinctly,hydroxyl-terminatedPS(Mn=42.5kDa)wassynthesizedvialivinganionicpolymerization. ThesubsequentPLAwassynthesizedviaringopeningtransesterificationpolymerization(ROTEP)of D,L-lactideindichloromethaneatroomtemperatureusing1,8-diazabicyclo[5.4.0]undec-7-ene(DBU)as acatalystfor1h. PS-b-PLAwasobtainedbyprecipitatinginmethanolafterterminationwithbenzoic acid. ThefinalPS-b-PLAhadatotalMn=63kDa,withaPLAvolumefractionof0.28(byvolume) yieldingacylindricalmorphologywithapolydispersityindexof1.08,asdeterminedwith1HNuclear MagneticResonance(NMR)andSize-ExclusionChromatography(SEC)[34]. 2.3. ThinFilmPreparation Typicalsolutionsof1.5%(w/v)PS-b-PLAintoluene(non-selectivesolvent)werespincoated ontoHMDStreated,natively-oxidizedsiliconwafers(20mm×20mm). HMDStreatmentoftheSi waferswascarriedoutbyultrasonicallycleaningsubstratesinorganicsolvents(acetonefollowedby methanol),treatingthemina1:5(v/v)HMDS:toluenesolutionfor16h,thenrinsingintolueneto removeanyexcessHMDSthathadnotgraftedfullytothesubstrate,andblowingdrywithN gas. 2 Treatedwaferswerethenplacedina75◦Coventoremoveanyresidualsolvent. Thefilmswerespin coatedat1000–3000rpmfor30s,dicedinto12–16pieces(~5mm×5mm),andimmediatelyplaced ina75◦C(belowtheglasstransitiontemperatureofeitherblock[34])ovenfordrying. Thisprocess yielded~60nmthickfilms,dependentonexactspinspeedandsolutionconcentration. Sampleswere driedforaminimumofoneday. Periodicatomicforcemicrographsindicatenoapparentagingissues orannealingeffectswhilestoringfilmsatthistemperature. Forthepresentexperiment,thinfilmswerehot-loadedintoaN purged,over-pressurizedsolvent 2 annealingchamberandimmediatelysealedtoavoidwatercontamination. Thesampleswereallowed tocooltoroomtemperatureandnofurthertemperaturecontrolwasimplemented. Thisisdescribed infulldetailbelow. THFliquidwasstoredinasealedErlenmeyerflaskoveractivated3Åmolecular sievesforseveraldaysbeforeintroducingittotheSAC,resultinginwaterlevelsof4.1ppmorless[59]. AdditionalactivatedmolecularsievesresideinthesolventreservoiroftheSACtocontinueactive dryingduringsubsequentsolventvaporanneals. Followingsolventvaporexposureandcomplete solvent evaporation from the film, the samples were immediately transferred to a 0.05 M NaOH solution(H O:CH OH=6:4byvolume)forPLAminoritydomaindegradationandlefttosoakfor 2 3 45min,wherethedegradationrateissensitivetomolarity[60]. Afterremoval,filmswerewashedwith deionizedwater/methanolfor5min. Byremovingtheminoritycomponentimmediately,thisserves toensurethemorphologyofthefilmisimmobilizedforsubsequentimaging. Furthermore,toremove anyadditionalsurfacecontamination,a150W,50KHzO reactiveionetch(PE-50,PlasmaEtch,Inc., 2 Carson,NV,USA)isemployedfor10sat100mTorr. Thisprocessremoves~2–3nmoforganicmaterial. Samplesatthisstagewereimmediatelyimagedwithatomicforcemicroscopy(AFM),withoutany furthermodification. 2.4. MeasuringFilmThickness FilmthicknesswasdeterminedwithaFilmetricsF20-UV(SanDiego,CA,USA)general-purpose filmthicknessmeasurementsystemwithbothhalogenanddeuteriumsources. Spectralreflectance datawastakenatdifferingtimeintervals(asdiscussedbelow),between10ms–3switha10–250ms integrationtime.Experimentaldatawasmodeledoveraspectralrangeof270–900nmwithathree-layer model(Si+PS-b-PLA+air).Wedevelopedanticipatedrefractiveindexprofilesbasedonknownvalues (e.g.,PS,n=1.59;PLA,n=1.482;THF,n=1.407). Therefore,weexpectedanindexofrefractionof 1.55fortheneatfilm,droppingto1.45withincreasingsolvent(THF)concentrationuptoφ=0.55[34]. Samplesnotfollowingthistrendinrefractiveindexwithincreasingsolventwereabortedanddisposed. Micromachines2018,9,271 7of21 2.5. AtomicForceMicroscopy(AFM) TappingmodeAFMwasperformedonanAgilent5420microscope(SantaClara,CA,USA)under ambientconditionsusingengagementsetpointsbetween0.9–0.95ofthefreeamplitudeoscillation. Thetappingmodecantilevers(BudgetSensors,Sofia,BulgariaandBruker,Billerica,MA,USA)had a resonant frequency of 300 kHz and a force constant of 40 N/m. For imaging the film/substrate interface,thePLA-removedthinfilmswereplacedupsidedownondouble-sidedtransparenttape (ScotchBrand,St. Paul,MN,USA)andplacedinliquidN for30s. FollowingliquidN exposure, 2 2 theSiwaferwaspeeledawayfromthefilmprovidingaccesstotheunderside. Thesefilmswereagain exposedtoa10–20sO reactiveionetch(150W50KHzin100mTorr)ontheundersidetoremoveany 2 HMDS,athinPSwettinglayeroradhesiveresidue[34,61]. 3. Results 3.1. DesignofaPurpose-BuiltSolventVaporAnnealingChamber Thereproducibilityofafinalmorphologyanditspropagationintosolventvapor-annealedBP filmsremainssomewhatelusive,despitenumerousadvancesinthefield. Todirectlyaddressthisissue, wehavedesigned,constructedandtestedapurpose-builtsolventvaporannealingchamber(SAC) thatprovidesunparalleledcontroloverintroducingandmaintainingprecisesolventconcentrations within the film during the annealing process as well as during the evaporation of solvent from thefilm. Concurrently,wemaintaincontroloverseveraladditionalnecessaryparameters(Table1), including an exceptionally low dew point in the sample cell with active dry N purging, solvent 2 vaporflowrate,filmthicknessasaproxyforsolventconcentrationinthefilm[34],chamberpressure, solventselectivity,andsubstratesurfacepreparation. Thepresentinvestigationuses~60nmthick films of a prototypical BP, PS-b-PLA, that adopts a cylindrical morphology in the bulk. We use at roomtemperaturearelativelyneutralsolvent,THF,fortheanneal[29,34,62];aPS-selectivesubstrate, HMDS-functionalized Si; and maintain a low dew point (−100 ◦C) for the N carrier gas, due to 2 thehygroscopicnatureofTHFandPLA.ThefollowingSACdescriptionisasignificantevolution overinitialworkindevelopingtheseprotocols, includingtheadditionofpneumatically-actuated, computercontrolledsolventflowratesandvalveactuation[34,38]. Inourmethod,acopper-gasket sealed,all-metalchambercontrolstheSVAclimate(i.e.,humidity,chamberpressure,andevaporation times). Theall-metalconstruction,exceptforchemicallyinertmodifiedpolytetrafluoroethylenesealed ballvalvesandaperfluoroelastomersealedlow-flowmeteringvalve,ensuresnegligibleinteraction withthesolventduringtheannealingprocess. Inparticular,THFcanbeparticularlyaggressiveon traditional organic gaskets and lubricants. Following hundreds of hours of exposure to THF, no degradationtoanyvalvesinthechamberisevident. Thechamber,showninFigure2(additionalimagesareavailableinAppendixA),isconnectedto anactivelydriedN2line(dewpointguaranteedto−100◦C)(cid:13)1 ,whichsplitsintotwo1/4”stainless steeltubingpaths.Path1(inred)purgesthesamplespacebeforeannealingandduringsampleloading; thisensuresalowdewpointduringtheSVA.Additionally,BPfilmsarehot-loadedfroma75◦Coven intothepurged,over-pressurizedchamberviaaConflat-flangeddoorandimmediatelysealedtoavoid watercontaminationandallowedtocooltoroomtemperature. Path1isalsovitaltothecontrolled evacuationofsolventvaporfromtheSACduringtheevaporationphase. N flowispassedthroughan 2 acrylic,block-styleflowmeter(3–30SCFHor30–300SCFH,dependentonflowratechosen)tocontrol andmeasureflowratesduringsolventevaporation. Micromachines2018,9,271 8of21 Table1. Acompilationofvariablesconsideredinthedesignofsolventvaporannealingprotocols, alongwiththeirimportanceandthestepstakeninthedesigntoaddressthem. Variable Importance ProtocolsinDesign Samplesarestoredin75◦Coven Samplesarehotloadedintochamber Waterisapolarsolvent,which Vacuum-grade,coppergasketsand willmodifythesolubility. Swageloksealsthroughoutchamber Humidity Forexample,wateris Molecularsieve-driedsolvent PLAselective (tetrahydrofuran(THF)) Activelypurged(dryN )samplecell 2 DrieriteGasPurifier(−100◦Cdewpoint) Flowrateisproportionalto Computer-controlledmassflowcontroller SolventVaporFlowRate solventuptakeinfilm Low-flowmeteringoutletvalve Solventconcentrationinfilm Insituopticaldetectionofsolvent SolventConcentration duringsolventvaporannealing concentration,assumingproportionalto (SVA)modifiesmobility filmthickness,with10–20msresolution Evaporationrateinlinkedto Computer-controlledpneumaticvalves SolventEvaporationRate morphologyalignment VariableflowN purgeline 2 Filmsareallspuncastataconstantspin Roleofcommensuratethickness InitialThickness speedfromthesamesolution onmorphology concentration VaporpressureduringSVA Highandlowpressuregauges VaporPressure modifiessolventuptakeand Low-flowmeteringoutletvalvecanfinely evaporation adjustvaporpressure Solventselectivitymodifies THFisarelativelyneutralsolventforPS SolventSelectivity surfaceenergyand andPLA.IthasslightPSselectivity polymer-polymerinteractions Substratepreparationmodifies HMDS-functionalizedSisubstratesurface SurfaceSelectivity surfaceenergy promotesPS(majorityblock)adhesion Path2(ingreen)flowisgovernedbyamassflowcontroller(cid:13)2 connectedtoaLabVIEW-enabled computer(NationalInstruments,Austin,TX,USA).Flowratesrangefrom0–500SCCM,dependenton intendedannealconditions. Theflowofthemetered,activelydriedN gasiscomputercontrolledvia 2 solenoidactuatedpneumaticvalves(cid:13)3 thatcontrol80psi(551kPa)airflowtoeachoffourpneumatic processvalves(cid:13)4 . FollowingtheflowcontMroicrlolmearch,intehs e201N8, 92, xg asispassedthroughaprimarysafetyvalve 9 of 20 (cid:13)4 (this protects the mass flow controller from liquid solvent exposure). The controlled gas flow continuesonitswaytoasealed,moleculacro-msipeuvteer- cdonritreodllesdo vliva esonletnroeids earctvuoatierd(cid:13) p5n,euamsattaici nvalelvsess ③ste tehlatC coonntrflola 8t0 psi (551 kPa) air flow to each of four pneumatic process valves ④. Following the flow controller, the N2 gas is passed teewithviewwindow(FigureA5). Thisrtehsroeurgvho iar pcroimnatrayi nsasfeatyn vaadlvde it④io n(tahils inprloettecttus bthee wmiatshs falonwo cromntarollllyer from liquid solvent closedplugvalvethatisusedforliquidsolevxepnostulroe)a. dTihne gcowntirtohlloedu gtabs rfeloawk icnongtiannueys Cono intsfl waatys teoa als seFailgedu, rmeoAlec6u)l.ar-sieve-dried solvent The solvent reservoir is backed by a saferteyserrveosier r⑤vo, air st(cid:13)a6in,leass sstteaeiln Cloensfslats tteeee wliCtho vnieflwa wtitnedeoww (Fitighurve iAew5). iTnhgis reservoir contains an additional inlet tube with a normally closed plug valve that is used for liquid solvent loading window. Thesafetyreservoirwillcollectfluidsolventifbackpressureispresent,avoidingexposure without breaking any Conflat seals Figure A6). The solvent reservoir is backed by a safety reservoir tothemassflowcontroller. Thesafetyres⑥er,v ao sitarincloesns tsateienl sCoannflaat dtede iwtiitohn vaielwtiungb weiwndiotwh. Tahne osarfmetya rlelsyercvoloirs weidll collect fluid solvent if backpressure is present, avoiding exposure to the mass flow controller. The safety reservoir contains plugvalvetoremoveanyliquidsolventwithoutopeninganyConflatseals. Ultimately,thedryN is 2 an additional tube with a normally closed plug valve to remove any liquid solvent without opening bubbledthroughthesolventreservoir(cid:13)5 ,wanhy iCchonsfluatb sseeaqlsu. eUnlttimlyatcealyr, rtihees dsoryl vNe2n its ibnubthbleedv athproourgph hthaes esoilnvetnot reservoir ⑤, which thesamplespace(cid:13)7 ,a6-wayConflatcubseub(7se0qumenmtly ×car7ri0es msolmven×t in7 t0hem vmapo),r pvhiaasea inctoom thpe suatmeprl-ec sopnacter o⑦ll,e ad 6,-way Conflat cube (70 pneumatically-actuatedchambervalve(cid:13)4 .mAml l×fl 7o0 wmmin ×t o70t hmems),a vmia pa lceomsppautceer-ceoxnittrsoltlehdr,o pungeuhm3a/tic4a”llys-taactiunalteesds chamber valve ④. All flow into the sample space exits through 3/4” stainless steel tubing via a pneumatically-actuated ball steeltubingviaapneumatically-actuatedbvaalllvve a⑧lv eo(cid:13)r8 ao rloawl-oflwow-fl omwetemrinegt evrailnvge v⑨al vteo (cid:13)9a tfuomae fuhmooed. hTohoed .computer controlled, Thecomputercontrolled,pneumatically-apcnteuumataetidcalvlya-alvcteuaste(cid:13)d4 v,awlvietsh ④c,o wmithp croemspseredsseadi raiirn inlelett ⑩ aanndd 12102 V0 soVlenoid valves ③, allow solenoidvalves(cid:13)3 ,allowustoquicklyinituias ttoe qaunicdklyte irnmitiaitne aatned ttheremSinVatAe twhe iStVhAa wsipthe ca isfipeecdifileedv leevleol foft tiimmiinngg. .Of critical importance, initial testing indicates a controlled SVA evaporation time down to 15 ms or any time longer with Ofcriticalimportance,initialtestingindicatesacontrolledSVAevaporationtimedownto15msor 10–20 ms temporal resolution. To our knowledge, this is the fastest recorded SVA evaporation time anytimelongerwith10–20mstemporalrefsoor lau BtPio tnhi.nT fiolmo. urknowledge,thisisthefastestrecordedSVA The BP sample resides in the sample space ⑦ on a custom-built mount (Figure A7). The sample evaporationtimeforaBPthinfilm. mount features two recessed ports, which speeds sample loading and helps avoid direct flow from Paths 1 and 2 that can shift sample position. One recessed port holds the BP film, while the other holds a blank Si wafer for use as an optical standard. Recess ports are sized to hold up to an 8 mm × 8 mm BP film and a blank Si wafer. The chamber offers direct optical access to the sample space through a fused silica viewport ⑪, enabling continuous monitoring of the sample chamber with high precision pressure gauges ⑫ to directly measure chamber pressure and in situ spectral reflectance-based measurements of film thickness with 0.1 nm thickness and 10–20 ms temporal resolution. These thickness measurements are directly related to solvent volume concentration (φ) within the film by: 𝑉 −𝑉 𝑡 −𝑡 𝑡 ϕ= 𝑠𝑜𝑙𝑣𝑒𝑛𝑡+𝑓𝑖𝑙𝑚 𝑓𝑖𝑙𝑚= 𝑠𝑜𝑙𝑣𝑒𝑛𝑡+𝑓𝑖𝑙𝑚 𝑓𝑖𝑙𝑚=1− 𝑓𝑖𝑙𝑚 (1) 𝑉 𝑡 𝑡 𝑠𝑜𝑙𝑣𝑒𝑛𝑡+𝑓𝑖𝑙𝑚 𝑠𝑜𝑙𝑣𝑒𝑛𝑡+𝑓𝑖𝑙𝑚 𝑠𝑜𝑙𝑣𝑒𝑛𝑡+𝑓𝑖𝑙𝑚 where φ is solvent concentration, V denotes volume and t is film thickness. As supported by direct observation, the areas for the two film states (i.e., swollen versus dry) are taken to be nominally identical and thus only a thickness measurement is required to obtain a real-time in situ probe of φ during SVA. 3.2. Theory of Operation Through the active control of solvent vapor inlet and outlet flows, the solvent concentration within the BP film is controlled and monitored as a function of time. As shown in Figure 3, using an in situ measurement of φ, we divide the SVA process of a 60 nm PS-b-PLA film into four distinct time regimes. During the first three time regimes, thickness data was taken every 3 s with a 249 ms integration time. The initial regime (blue) includes the opening of the solvent reservoir to the sample chamber and an initial, constant N2 inlet flow of 30–100 SCCM (dependent on desired chamber pressure, 50 SCCM in the present study), resulting in an exponential increase in ϕ. The thickness data is well modeled assuming a copolymer refractive index of 1.55, consistent with the neat PS-b-PLA film (Appendix B). This exponential region typically lasts ~60 s where the time dependence is well modeled with a single rate constant (0.03–0.1 s−1) with 0.088 s−1 for Figure 3 (see Figure A8 for the exponential fit), which is dependent on exact inlet and outlet flow rates. Outlet flow rates are governed by a low flow micrometer relief valve, which is set to a flow coefficient, Cv, of 0.0002–0.002 (1–6 turns), with 0.0004 being used in Figure 3. This, along with inlet flow, dictates the Micromachines 2018, 9, x 8 of 20 Vapor pressure during SVA High and low pressure gauges Vapor modifies solvent uptake and Low-flow metering outlet valve can finely Pressure evaporation adjust vapor pressure Solvent selectivity modifies surface Solvent THF is a relatively neutral solvent for PS energy and polymer-polymer Selectivity and PLA. It has slight PS selectivity interactions Surface Substrate preparation modifies HMDS-functionalized Si substrate surface Selectivity surface energy promotes PS (majority block) adhesion The chamber, shown in Figure 2 (additional images are available in Appendix A), is connected to an actively dried N2 line (dew point guaranteed to −100 °C) ①, which splits into two 1/4” stainless steel tubing paths. Path 1 (in red) purges the sample space before annealing and during sample loading; this ensures a low dew point during the SVA. Additionally, BP films are hot-loaded from a 75 °C oven into the purged, over-pressurized chamber via a Conflat-flanged door and immediately sealed to avoid water contamination and allowed to cool to room temperature. Path 1 is also vital to the controlled evacuation of solvent vapor from the SAC during the evaporation phase. N2 flow is passed through an acrylic, block-style flowmeter (3–30 SCFH or 30–300 SCFH, dependent on flow Micromachines2018,9,271 9of21 rate chosen) to control and measure flow rates during solvent evaporation. Micromachines 2018, 9, x 9 of 20 Micromachines 2018, 9, x 9 of 20 computer controlled via solenoid actuated pneumatic valves ③ that control 80 psi (551 kPa) air flow computer controlled via solenoid actuated pneumatic valves ③ that control 80 psi (551 kPa) air flow to each of tfoo uera cphn eouf mfoautirc ppnreoucmesas tivca lpvreosc e④ss . vFaolvlleosw ④ing. Fthoell oflwowin gc othnetr ofllolewr, ctohne trNo2l legra,s thise pNas2 sgeads is passed through a primary safety valve ④ (this protects the mass flow controller from liquid solvent through a primary safety valve ④ (this protects the mass flow controller from liquid solvent exposure). The controlled gas flow continues on its way to a sealed, molecular-sieve-dried solvent exposure). The controlled gas flow continues on its way to a sealed, molecular-sieve-dried solvent reservoir ⑤, a stainless steel Conflat tee with view window (Figure A5). This reservoir contains an reservoir ⑤, a stainless steel Conflat tee with view window (Figure A5). This reservoir contains an additional inlet tube with a normally closed plug valve that is used for liquid solvent loading additional inlet tube with a normally closed plug valve that is used for liquid solvent loading without breaking any Conflat seals Figure A6). The solvent reservoir is backed by a safety reservoir without breaking any Conflat seals Figure A6). The solvent reservoir is backed by a safety reservoir ⑥, a stainless steel Conflat tee with viewing window. The safety reservoir will collect fluid solvent if ⑥, a stainless steel Conflat tee with viewing window. The safety reservoir will collect fluid solvent if backpressure is present, avoiding exposure to the mass flow controller. The safety reservoir contains backpressure is present, avoiding exposure to the mass flow controller. The safety reservoir contains an additional tube with a normally closed plug valve to remove any liquid solvent without opening an additional tube with a normally closed plug valve to remove any liquid solvent without opening any Conflaatn yse aClos.n fUlaltt imseaatlesl.y ,U tlhtiem adtreyly N, t2h ies dbruyb bNle2d i st hbruobubglhe dt hteh rsooulgvhen tth ree sseorlvvoeinrt ⑤res, ewrvhoiicrh ⑤, which subsequently carries solvent in the vapor phase into the sample space ⑦, a 6-way Conflat cube (70 subsequently carries solvent in the vapor phase into the sample space ⑦, a 6-way Conflat cube (70 mm × 70 mmmm × × 7 07 0m mmm), ×v i7a0 a m Fcomimg)u,p vruietae ra-2 cc.o onm(tarpo)u llteeCdr-o,c pmonnpetruuomltleeardti-,ca apildnlyee-udamc taudtaiecteasdlilg ycn-ha acmt(uCbaetAerd Dv ac)hl vaeam n④bde. r Av(blall )v er ④ea.l Ailml age of a computer-controlled, flow into the sampleF sipgaucere ex2i.ts( ath)rCouogmh p3/u4t”e srt-aainidleesds sdteeesli tgunbi(nCgA viDa )a apnndeu(mba)triceaalllyi-macatugaeteodf baacllo mputer-controlled,pneumatically flow into the sample space exitsp tnherouumgha t3i/c4a”l slyta iancletussa stteedel stuoblvinegn vti av aa pponre uamnanteicaallilyn-ga cctuhaatmedb bearl.l More images can be found in Appendix A. valve ⑧ voarl vea ⑧low -oflro wa lmoawect-etfrulionawgte dvmaselovteelrv ine⑨gn t vvtaoal vpeao r⑨faunm neto e ahlaoi ondfgu. mcTheh aemh obcooedmr. .pMTuhtoeer r eccoiommntparuogtleelesrd c, caonntbroellfeodu, ndinAppendixA.Numbersand pneumatically-actuaNteud mvablveerss ④ an, dw ictho lcoormepdr easrsreodw aisr ianrleet d⑩es acnrdib 1e2d0 Vin s othleen oteidx vt.a lves ③, allow pneumatically-actuated valves c④ol,o wreitdh caormropwresssaerde adire isnclerti b⑩ed anidn 1t2h0e Vt esxolte.noid valves ③, allow us to quickly initiate and terminate the SVA with a specified level of timing. Of critical importance, us to quickly initiate and terminate the SVA with a specified level of timing. Of critical importance, initial testiningi tiinald ticeasttiensg a incodnictraoPtelalse tdah cS oVn2Atr oelv(lieandp oSrVagtArioe neev tnaimp) oer afdtloioowwnn t itmoi es1 5d omgwson vo rteo ar n1n5ye mtdims eob rl oyann gyea rt i mwmeit hlao nsgse r fwloitwh controller ② connected to a 10–20 ms te1m0–p2o0r mal sr etesmolLuptoairobanlV. rTTeIoshEo oeWluuBtri- oPkennn.so aTwaobml eloedupdgrle ek,c ntorhoeimwss liipesdd tueghtesee, fitranh si(tseNt hsista e rtthesiceoao fnrmadsaetpled sl ItSen rVessActpor earudvcmaeepdoe (cid:13)Sr7nVattAioso ,nne vA taaimpucoesur taistnitoo,n mT tiXm-b,e u UilStAm).o Fulnotw(F riagtuesre raAn7g).e Tfrhoemsa 0m–5p0le0 for a BP thin film. for a BP thin film. mountfeaturestworecessedports,whichspeedssampleloadingandhelpsavoiddirectflowfrom The BP samTphlee BrePs isdaSemCs piCnle tM rhees, isddaemesp pinlee t nshpeda secaenm ⑦tp ol eon nsp aian ccuete s⑦tnomd oen-b dau iclaut nsmtnoomeuan-btl u (cFiloitg mnurdoeui Atnit7o )(.Fn Tisgh.ue rT sea hAme7p ).fl elT ohwe s aomfp tlhe e metered, actively dried N2 gas is mount featmuroeus nttw foe arteucreeP sssa etwtdh ops or1ertcsea,s nwsedhdi c2pho tsrhptsae, etwdchsa iscnahm ssphpleieef tdlossa asdamimngpp allene dlop ahodeslipnitgsi aoavnnod.i dhO edlniprese cartv efolcoiedws d sfireroedmct p floowrt fhroomld stheBPfilm,whiletheotherholds Paths 1 andP a2t htsh a1t acnadn 2sah tihbfta ltsa acnmaknp lsSeh iipfwto ssaiatfmioenpr.l feOo pnroeus irtseiecoenas.s sOedan nep oroerctp ehtsioscledadsl ptshoterat nBhPdo lafdirlsmd t,h. weR BheiPlce ef itslhmse, p owtohhreitrlse tahree osthizere dtoholduptoan8mm×8mmBP holds a blank Si wafer for use as an optical standard. Recess ports are sized to hold up to an 8 mm × 8 holds a blank Si wafer fofir ulmse aasn adn oaptbiclaal nstkanSdiarwd.a Rfeecre.ssT phoertsc harae msizbeedr too hfofeldr supd tior eanc t8 ompmti ×c a8 laccesstothesamplespacethroughafused mm BP film and a blank Si wafer. The chamber offers direct optical access to the sample space mm BP film and a blank Si wafer. The chamber offers direct optical access to the sample space through a fusseidli csialicva iveiwewppoorrtt ⑪,, eennaabblinlign cgonctoinnutoiunsu mouonsitmorionng iotof trhine gsamofptleh cehsaammbepr lweitchh amberwithhighprecisionpressure through a fused silica viewport ⑪, enabling continuous monitoring of the sample chamber with high precision pressure gauges ⑫ to directly measure chamber pressure and in situ spectral high precision pressureg gaauuggeess ⑫ ttoo ddiirreecctltyl ymmeaesuarseu crheacmhbaerm pbreesrsuprree asnsdu rien asnitud sipnecstirtaul spectralreflectance-basedmeasurementsof reflectance-based measurements of film thickness with 0.1 nm thickness and 10–20 ms temporal reflectance-based measurements of film thickness with 0.1 nm thickness and 10–20 ms temporal filmthicknesswith0.1nmthicknessand10–20mstemporalresolution. Thesethicknessmeasurements resolution. These thickness measurements are directly related to solvent volume concentration (φ) resolution. These thickness measurements are directly related to solvent volume concentration (φ) within the filma rbey:d irectlyrelatedtosolventvolumeconcentration(φ)withinthefilmby: within the film by: 𝑉 −𝑉 𝑡 −𝑡 𝑡 ϕ=𝑉𝑠𝑜𝑙𝑣𝑉𝑒𝑠𝑛𝑜𝑡ϕ𝑙+𝑣𝑒𝑓=𝑛𝑖𝑙𝑡𝑚+𝑓𝑠−𝑜𝑖𝑙𝑙𝑚𝑣𝑉𝑉𝑒𝑓𝑠𝑛𝑖𝑜𝑙𝑡𝑙𝑚+𝑣𝑒𝑓𝑛=𝑖𝑙𝑡𝑚+𝑡𝑓𝑠𝑖𝑜𝑙φ𝑚𝑙𝑣𝑡𝑒𝑓𝑠𝑛𝑖𝑜=𝑙𝑡𝑙𝑚+𝑣𝑒𝑓𝑛=𝑖V𝑙𝑡𝑚+s𝑓𝑠−o𝑖𝑜𝑙l𝑙𝑚𝑣v𝑡𝑡𝑒𝑓e𝑠𝑛𝑖𝑜n𝑙𝑡𝑙𝑚t+𝑣+𝑒𝑓𝑛=𝑖𝑙𝑡f𝑚+i1l𝑓m𝑖−𝑙𝑚−𝑡𝑓𝑠𝑖𝑜𝑙𝑚𝑙V𝑣𝑡𝑒𝑓=𝑛𝑖f𝑙𝑡i𝑚+1lm𝑓−𝑖𝑙𝑚𝑡= 𝑠 𝑜𝑙𝑣t𝑒𝑓𝑛s𝑖𝑙o𝑡𝑚+lv𝑓𝑖e𝑙n𝑚t +f(il1m) −tfil(m1) =1− tfilm (1) where φ isw sohlevreen φt cios nscoelnvternatt icoonn, cVen dtreantoiotens, vVo dluemnoet easn dvo tl uiVsm sfoiell mvane ntdht i+tc kifsni lfemislms. Athsi cskunpepsos.r tAedst ssbouylp vdpeinortre+tcetd fi lbmy direct tsolvent+film observation, the areas for the two film states (i.e., swollen versus dry) are taken to be nominally observation, the areas for the two film states (i.e., swollen versus dry) are taken to be nominally identical anidde tnhtuicsa ol nanlyd a wt hthuhisce kornneelysφs a m ithseiacsskounrleevsmse enmntet aicsso urrneeqcmueiernnettd ri stao rt eoiqobutnairi,neVd a trode aoelbn-ttaiomitnee a si nrve saiolt-ultu ipmmreo beine asoinft uφd ptroibse fiofl mφ thickness. Assupportedbydirect during SVA. during SVA. observation, the areas for the two film states (i.e., swollen versus dry) are taken to be nominally 3.2. Theory of Oidpeernattioinc alandthusonlyathicknessmeasurementisrequiredtoobtainareal-timeinsituprobeofφ 3.2. Theory of Operation Throughd thuer ianctgiveS VcoAnt.rol of solvent vapor inlet and outlet flows, the solvent concentration Through the active control of solvent vapor inlet and outlet flows, the solvent concentration within the BP film is controlled and monitored as a function of time. As shown in Figure 3, using an within the BP film is controlled and monitored as a function of time. As shown in Figure 3, using an in situ measurement of φ, we divide the SVA process of a 60 nm PS-b-PLA film into four distinct in situ measurement of φ, we divide the SVA process of a 60 nm PS-b-PLA film into four distinct time regimes. During the first three time regimes, thickness data was taken every 3 s with a 249 ms time regimes. During the first three time regimes, thickness data was taken every 3 s with a 249 ms integration time. The initial regime (blue) includes the opening of the solvent reservoir to the sample integration time. The initial regime (blue) includes the opening of the solvent reservoir to the sample chamber anchda manb einr iatinadl, caonn isntaitniat l,N c2o innslteatn ftl oNw2 ionfl e3t0 –fl1o0w0 SoCf C30M–1 (0d0e pSCenCdMen t( doenp ednedseirnetd o cnh admesbireerd chamber pressure, 50 SCCM in the present study), resulting in an exponential increase in ϕ. The thickness pressure, 50 SCCM in the present study), resulting in an exponential increase in ϕ. The thickness data is well modeled assuming a copolymer refractive index of 1.55, consistent with the neat data is well modeled assuming a copolymer refractive index of 1.55, consistent with the neat PS-b-PLA film (Appendix B). This exponential region typically lasts ~60 s where the time PS-b-PLA film (Appendix B). This exponential region typically lasts ~60 s where the time dependence is well modeled with a single rate constant (0.03–0.1 s−1) with 0.088 s−1 for Figure 3 (see dependence is well modeled with a single rate constant (0.03–0.1 s−1) with 0.088 s−1 for Figure 3 (see Figure A8 for the exponential fit), which is dependent on exact inlet and outlet flow rates. Outlet Figure A8 for the exponential fit), which is dependent on exact inlet and outlet flow rates. Outlet flow rates afrloe wgo rvaetersn eadre b gyo av elornwe dfl obwy am liocwro fmloewte rm rieclrioefm veatlevre r, ewlihefi cvha livs es,e wt thoi ach f liosw se cto teof fai cfileonwt ,c Cove,f ofifc ient, Cv, of 0.0002–0.002 (1–6 turns), with 0.0004 being used in Figure 3. This, along with inlet flow, dictates the 0.0002–0.002 (1–6 turns), with 0.0004 being used in Figure 3. This, along with inlet flow, dictates the Micromachines2018,9,271 10of21 3.2. TheoryofOperation Through the active control of solvent vapor inlet and outlet flows, the solvent concentration withintheBPfilmiscontrolledandmonitoredasafunctionoftime. AsshowninFigure3, using aninsitumeasurementofφ,wedividetheSVAprocessofa60nmPS-b-PLAfilmintofourdistinct timeregimes. Duringthefirstthreetimeregimes,thicknessdatawastakenevery3switha249ms integrationtime. Theinitialregime(blue)includestheopeningofthesolventreservoirtothesample chamber and an initial, constant N inlet flow of 30–100 SCCM (dependent on desired chamber 2 pressure,50SCCMinthepresentstudy),resultinginanexponentialincreaseinφ. Thethicknessdata iswellmodeledassumingacopolymerrefractiveindexof1.55, consistentwiththeneatPS-b-PLA film(AppendixB).Thisexponentialregiontypicallylasts~60swherethetimedependenceiswell modeledwithasinglerateconstant(0.03–0.1s−1)with0.088s−1 forFigure3(seeFigureA8forthe exponentialfit),whichisdependentonexactinletandoutletflowrates. Outletflowratesaregoverned byalowflowmicrometerreliefvalve,whichissettoaflowcoefficient,C ,of0.0002–0.002(1–6turns), v with0.0004beingusedinFigure3. This,alongwithinletflow,dictatesthechamberpressure,which wasmaintainedbelow3.5kPatoobtainFigure3. FollowingthedisplacementofresidualN inthe 2 chamberwithsolventvapor,thesolventuptakeentersasecondregime(red). Thisregimeinvolvesthe metereduptakeofsolventintothefilm,indicatedbyacontrolled,linearincreaseinthicknessover aperiodof2min(highlytunable,basedonrelativeflowrates),untilthefilmreachesthetargetedφ. Theinletandoutletflowratesremainedthesameforthesecondregimeasgivenforthefirstregimein Figure3,demonstratingthischangeinsolventuptake. Thethicknessdataiswellmodeledassuminga copolymerrefractiveindexthatgraduallyapproaches1.45,consistentwithavolumetriccombination of neat PS-b-PLA and THF (Appendix B). During this phase, two possible methods drive solvent into the film and increase the thickness. If the inlet flow is higher than outlet flow, the increasing pressure in the chamber will increasingly force solvent into the film (a relatively fast mechanism). Iftheinletandoutletflowsarecomparable,theincreaseinrelativesolventconcentrationwithinthe chamberleadstoanincreaseduptakeofsolventintothefilm(arelativelyslowmechanism). Therefore, itispossibletoswellafilmwithachamberpressurethatisnominallyatmospheric. Dependenton desiredchamberpressureandultimatethickness,theinletandoutletflowratesaretuned. Precise controlisbestachievedbyadjustingthemassflowcontroller-regulatedinlet. Wehavenotseenany impactwiththerateofswellinginthissecondregimeonthefinalmorphologyofthefilm,otherthan a potential dependence on the pressure in the chamber (faster swelling is typically accomplished withanincreaseinchamberpressure). Thethicknessdatainthisregimeiswellmodeledassuminga copolymerrefractiveindexof1.45(FigureA10),consistentwithasolvatedPS-b-PLAfilmcontaining φ=0.55. Asthetargetφisapproached,decreasingtheinletfloworincreasingtheoutletflowcausesφto leveloff. Thethirdregime(green)ischaracterizedbyanextremelyconstant(standarddeviationis regularlylessthan∆φ=0.002)solventconcentration. Thisconstantconcentrationcanbemaintained for nearly any specified anneal time (3 min is shown in Figure 3, but we have maintained similar consistencyformorethananhour). Itismaintainedthroughslightmanualvariationsininlet/outlet flowrates,withthefuturepotentialofcomputerfeedbackcontrol. Figure3wasachievedwithafixed outletflow,consistentwithregimes1and2andbycontrollingtheinletflowbetween10–30SCCMto achieveaconstantthickness(solventconcentration). Inthefinalsecondsofthisperiod,theintegration time and thickness acquisition interval are switched to 10 ms and 0 s, respectively. This is the fastestwecanacquirespectralreflectancedataandstillgetahigh-fidelitymodeltoextractthickness. This rapid data acquisition allows for close examination of the solvent evaporation period. In the fourth regime (magenta), evaporation of the solvent from the film occurs. Computer-controlled, pneumatically-actuatedvalvesopenPath1andclosePath2. Throughthereleaseofpressureinthe chamber,intandemwithN flow,thesolventvaporisreleasedfromthefilmandevacuatedfromthe 2 chamber. Detailsontimingcontrol,whicharedependentonPath1flowratesandchamberpressure, are given below. The initial evaporation of solvent from the film is not complete, where φ ≈ 0.2
Description: