Tabakova ac Krassimir D Danov b Laboratory of Physicochemical Hydrodynamics Institute of Mechanics Bulgarian Academy of Science 1113 Soa Bulgaria Laboratory of Chemical Physics and Engineering Faculty of Chemistry University of Soa 1164 Soa Bulgari ID: 78527
Download Pdf The PPT/PDF document "Effect of disjoining pressure on the dra..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
EffectofdisjoiningpressureonthedrainageandrelaxationdynamicsofliquidlmswithmobileinterfacesSoniaS.Tabakova,KrassimirD.DanovLaboratoryofPhysicochemicalHydrodynamics,InstituteofMechanics,BulgarianAcademyofScience,1113Soa,BulgariaLaboratoryofChemicalPhysicsandEngineering,FacultyofChemistry,UniversityofSoa,1164Soa,BulgariaDepartmentofMechanics,TU-Soa,BranchPlovdiv,4000Plovdiv,Bulgariaarticleinfo Correspondingauthor.Fax:+35929625643.E-mailaddress:(K.D.Danov). JournalofColloidandInterfaceScience336(2009)273 284 ContentslistsavailableatScienceDirectJournalofColloidandInterfaceScience low.Thekinematicboundaryconditionatthelmsurface,seeEq.below,introducesanadditionalfunction,thelocallm.Hence,thekinematicboundaryconditiondoesnotde-creasethenumberofunknownparameters.Todeterminethem,thedynamicboundaryconditionattheinterfacemustbetakenintoaccount.Inlubricationapproximationthenormalstressboundaryconditionrelates=1,2),,and,seeEq.low.However,thetangentialstressboundarycondition,Eq.isautomaticallyfullledbytheleadingordersolutionoftheprob-lem.Forthatreason,thesecondordertermsshouldbeconsidered.Thisapproachisknownintheliteratureastheextendedlubrica-tionapproachapproach .Forsomerecentpracticalapplicationsitisveryimportanttoproducethinliquidlmsfromsurfactant-freephasesorfromphaseswithaverysmallamountofsurfaceactivecomponents.Similarproblemsarisewhenthelmsarestabilizedbynanome-ter-sizedparticles.Inthesecases:(i)thelateraldimensionsaremuchlargerthanthelocallmthickness;(ii)thelmdrainageoc-cursinadynamicregimeatarbitraryvaluesoftheReynoldsnum-ber.Modelingofthedynamicsofsuchsystemsispossibleusingtheextendedlubricationapproachapproach ,formerlyappliedtostudylinearandnonlinearstabilityoflms.Thisapproachhasalsoproveditswideapplicabilityformodelingthedrainageandstabil-ityofevaporatinglmslms andformodelingthesolidicationoffreethinlmsattachedtoaframeandpulledfromameltmelt.OriginallythesemodelsaccountforthevanderWaalsinteractionsbetweeninterfaces.Recentexperimentalandtheoreticalinvestiga-tionsmanifesttheimportanceofanumberofothercomponentsofthedisjoiningpressure.Experimentalmeasurementsshowthat,eveninpureliquid,thegasbubbleshave-potentialof65mVmV,whichforthexylene,dodecane,hexadecane,andperuorom-ethyldecalindropletschangesfrom100mVto20mV,depend-ingontheelectrolyteconcentrationandpHofthesolutionsolution.Thechargeaccumulationonthesurfacesisduetothespontaneousadsorptionofhydroxylions,whichcauseselectrostaticrepulsionbetweenthelminterfaces.Manyexperimentswithemulsionandfoamlmslms suggestthatthelongrangehydrophobicattractionforcescanbeconsiderablylargerthanthevanderWaalsinteractions,andmustbeincludedasadisjoiningpressurecompo-nent.Atclosedistancesbetweenlminterfacesstrongstericinter-actionsbetweenadsorbedparticlescanstabilizelmdrainageandpreventlmrupturerupture .Theaimofthepresentstudyistotakeintoaccounttheroleofdifferenttypesofintermolecularinter-actionsforthelmdrainageandstabilityindynamicconditions,usingtheextendedlubricationapproach.Thepaperisorganizedasfollows.InSection,wedescribetheextendedlubricationapproachappliedforlmswithtangentiallymobilesurfacesintheframeworkofthedisjoiningpressureap-proach.Thedimensionlessnumbersappearinginthemodelequa-tionsforone-dimensionalsymmetriclmsareintroducedin.Theytakeintoaccountthecontributionofinertia,inter-facialtension,vanderWaals,hydrophobic,electrostaticandstericforces.Therein,theboundaryconditionsandpossibleregimesofthelmdrainageandrelaxationarediscussed.TheeffectoftheintermolecularforcesandthemagnitudesoftheReynoldsandWe-bernumbersonlmdrainagearestudiedinSectionandtheirinuenceonlmrelaxation inSection.ThegeneralconclusionsaregiveninSection2.Modelingofthedynamicsofthinliquidfreelms2.1.ConservationofmassandmomentuminthebulkphaseWeconsiderasymmetricthinfoamlmwithalocalthicknessFig.1).Thebulkphaseinthelmisdescribedasanincompress-iblehomogeneousliquidwithvelocityvector,pressuredynamicviscosityanddensity.TheCartesiancoordinatesystem,,isplacedintheplaneoflmsymme-=0,where=1,2)arethelateralcoordinatesandtheverticalcoordinate(Fig.1).Thecharacteristiclateralvelocity,thelaterallengthscaleis,andthenaturalscaleoftime,.Thelmthickness,,ismeasuredwiththeverticallength,andisassumedtobesosmallthattheratio,canbeconsideredasasmallparameterintheproblem.Thedimen-sionlesstime,,thelateralcoordinatesandcomponentsofthevelocityvector,=1,2),aredenedasfollows UTa;xa Xaa;andua Thedimensionlesslmthickness,,theverticalcoordinate,,andtheverticalcomponentofthevelocity,,areintroducedthroughthefollowingrelationships Hb;x3 X3b;andu3 accountsforthefactsthat:(i)inthecaseofthinliquidlmstheverticalcomponentofthevelocityismuchsmallerthantherespectivelateralcomponents;(ii)thecontinuityequationmustbeinvariantlyexpressedintermsofdimensionlessquantities ouboxbþ InEq.andbelowwewillusetheEinsteinsummationconven-tion,wheretheGreeklettersubscripts()indicatesumsre-latedtothelateralcomponentsdenotedwithindexes1and2.ForNewtonianuidsthemomentumbalanceisdescribedbytheNavier Stokesequation.Forfoamlmswithfullymobilesur-faces(seeSection)aviscousscale,,isappropriateformea-suringthepressure.Thedimensionlesspressure,,andtheReynoldsnumber,Re,areintroducedbythedenitions andRe Usingtheaboveexpressions,Eqs.(2.1),(2.2)and(2.4),weobtainthedimensionlessformofthemomentumbalanceequationinthelateraldirections(=1,2) ouaotþ ooxbðubuaÞþ oox3ðu3uaÞ¼ opoxaþ o2uaoxboxbþ 1e2 o2uaox23:ð2:5Þ Fig.1.Sketchofasymmetricthinliquidlmwithalocallmthickness.TheverticalcoordinateaxisoftheCartesiancoordinatesystem,,isperpendic-ulartothelmmiddleplane,=0.Thecharacteristiclaterallength,,ismuchlargerthanthecharacteristicverticallength,.Thepositionoftheupperlmsurfaceisdenedas/2andthepositionofthelowerone as/2.Theuppersurfacenormalandtangentvectorsare,and,respectively.S.S.Tabakova,K.D.Danov/JournalofColloidandInterfaceScience336(2009)273 284 TherespectivedimensionlessformoftheverticalprojectionoftheNavier Stokesequationreads ou3otþ ooxbðubu3Þþ oox3ðu3u3Þ¼ 1e2 opox3þ o2u3oxboxbþ 1e2 TheReynoldsnumber,Re,appearinginEqs.(2.5)and(2.6)andde-nedbyEq.,isassumednottobeverylarge,i.e.ReForexample,forwaterlms(=10=10Pas)withaninitialthickness=1mmandaframelaterallength=1cm=0.1),thelateralvelocity,,atwhichRe==1isequalto1mm/s,thusinthatcasethemodelwouldbeapplicableforsuctionvelocitiessmallerthan1mm/s.TheeffectoftheinertiaonthelmdrainageisdiscussedinSectionFig.3Becauseofthesymmetryoftheconsideredproblematthemiddleplane=0.TakingintoaccountEq.oneobtainstheleadingordersolutionofEqs.(2.5)and(2.6)[22] wherethefunctionsdependontime,,andlateralcoordi-.AftersubstitutingEq.intothecontinuityequa-tion,Eq.,integratingtheobtainedresultwithrespecttousingthesymmetrycondition=0atthemiddleplane,weobtaintheleadingorderexpressionfortheverticalvelocitycomponent: Theunknownfunctions,andaredeterminedfromthekine-maticanddynamicboundaryconditionsatthelminterfaceusinganappropriateasymptoticprocedure(seeSectionAppendixA2.2.KinematicanddynamicboundaryconditionsattheinterfaceUsingtheabovedenitions,Eqs.(2.1)and(2.2),onepresentsthekinematicboundaryconditionatthelmsurfaceinasimpledimensionlessform ohotþub ohoxb¼2u3atx3¼ Throughsubstitutingtheleadingordersolutions,Eqs.(2.7)and,intoEq.weobtaintherespectiveexpressiondescribingtheconservationofmassinthecontinuouslmphase ohotþ Insurfactant-freelmstheinterfacialtension,,isconstantandthesurfaceshearanddilatationalviscositiesarenegligible.There-fore,inthiscase,thebulkviscousfrictionforceatthelmsurfaceiszerozero.Theleadingorderofthetangentialstressboundarycon-ditionat/2iscalculatedinAppendixA,seeEq..AftercombiningtheobtainedresultwiththeintegratedformofEq.AppendixA,wederivetheasymptoticformofthemomen-tumbalanceequationinthelm: ootðhwaÞþ ooxbðhwbwaÞ¼h ooxaqþ owboxbþ ooxbh owboxaþ owaoxbþ2 istheKroneckerdelta.ItisimportanttoclearlyindicateforwhichsystemsthemomentumbalanceEq.,originallyderivedforpureliquids,isapplicable.(i)Ifthesolutioncontainsonlyindifferentelectrolytes(salts),thenthechangeofthesurfacetensionwiththeionicstrengthissmall,theGibbselasticityissmallandtheMarangonistresscanbeneglectedneglected.TheelectrostaticandvanderWaalscomponentsofthedisjoiningpressureappearatsmalllmthicknessandcontrolthelmdrainageandstability.(ii)Inthecaseofsmallamountoflowmolecularweightsurfac-tants,thesurfactantrelaxationtimeiscomparabletothecharacteristictimeoflmdrainageandtheMarangonieffectbecomesevident.Thiseffectsuppressesthesurfacemobilityanddecreasestherateoflmthinningthinning.TheaccountoftheMarangonistressmakestherespectiveana-logueofEq.muchmorecomplex.(iii)Inthecaseoflargeamountoflowmolecularweightsurfac-tants(closeorabovethecriticalmicelleconcentration)thesurfactantrelaxationtimedecreasesconsiderablyanditisoftheorderoforsmallerthanseveralmicrosecondsmicroseconds.InthesesystemsthesurfacetensiondoesnotchangeduetothelmdrainageandtheMarangonistresscanbeneglected.Fromtheliteratureitisknownthatlowmolecularweightsurfactantshaveverysmallvaluesofsurfaceviscos-viscos-andtherefore,theroleofinterfacialrheologyisalsonegligible.ForsuchsystemsEq.isvalid.Inaddi-tiontotheclassicalcomponentsofthedisjoiningpressure,thesteric,oscillatory,etc.non-DLVOsurfaceforcesshouldbeconsidered.(iv)Filmsstabilizedbyproteinsorpolymershavesignicantval-uesofthesurfacedilatationalandshearviscosities,whichconsiderablyincreasetheeffectofthesurfacerheologyonthedrainageandrelaxationdynamics.Forsuchsystemsourmodelisnotapplicable.Forsmalllmthicknesses(below200nm)thesurfaceforcesofintermolecularoriginmustbeaccountedfor.OnetypeofsuchforcesisthevanderWaalsattractionforce,whichischaracter-izedbytheHamakerconstant,constant,.ManyauthorsincludethevanderWaalsinteractionsasabulkpotentialforceintheNavier Stokesequation.Thisapproachisknownintheliteratureasthebodyforceapproachapproach .Thebodyforceapproachap-pliedforionicsolutionsleadstoverycomplexmathematicalcal .Manyothertypesofsurfaceforcesaretheoreticallyandexperimentallyinvestigatedintheliterature(thestericrepulsion,thehydrophobicattraction,theoscillatorystructuralforces,etc.).Inmanycasestheoriginoftheobtainedinteractionsisnotclearlyunderstoodunderstood.Forthatreason,amoreconvenientwaytoaccountfortheroleofthesurfaceforcesistousethesocalleddisjoiningpressureapproachapproach.InthisapproachthebulkforcedensityintheNavier Stokesequationisomittedandallintermolecularinteractionsinthelmaretakenintoconsiderationbyintroducinganaddi-tionaldisjoiningpressureterm,,inthenormalstressboundarycondition.Thedisjoiningpressuredependsonlyonthelocallm,andthephysicochemicalpropertiesofthesolutions.ThespecicexpressionsforthedifferentcomponentsofgiveninSection.Thenormalcomponentofthebulkforceact-ingonthelmsurfaceisthencompensatedbythelmcapillaryanddisjoiningpressuresandthepressureinthegasphase,.InAppendixAwederivetheasymptoticformofthenormalstressboundarycondition.TheobtainedresultreadsS.S.Tabakova,K.D.Danov/JournalofColloidandInterfaceScience336(2009)273 284 qþ owboxbþ e2Ca wherethecapillarynumber,Ca,thedimensionlessdisjoiningpres-,andthepressureinthegasphase,,aredenedas gUr;pdp 2Pa2rb;andpg SubstitutingEq.intoEq.wearrivetothenalequationofthemodel: ootðhwaÞþ ooxbðhwbwaÞ¼ eh2Ca ooxa o2hoxboxbþpdp !þ ooxbh owboxaþ owaoxbþ2 Notethatthetypicalvaluesofthecapillarynumberaresmall.Thustheparameter/CainEq.isoftheorderoforlargerthanErneuxandDavisDavisshowedthatthemodelisanalogoustoarespectiveapproachingasdynamics.Eqs.(2.10)and(2.14)spondtothetwo-dimensionalviscous-owequationsifweiden-tifythethicknessasthedensityofthecompressiblegas,thethicknessastheshearviscosity.Thersttermintheright-handsideofEq.canbeconsideredasthepressure),whichadmitsthefollowingcomplexequationofÞ¼ eq2Ca o2qoxaoxbþ e4Ca2Z1qn dpdpðnÞdndnþ oqoxc generalizedforanarbitraryfunction).Forsmallvaluesofthelmthickness,,thedensityinEq.isverysmallandthesoundspeedbecomesequalto eq2Ca showsthatifd0thesoundspeedisimaginary,whichcorrespondstoinstability.Incontrast,ifd0thesys-temhasarealsoundspeed,correspondingtowavepropagation.Thisfactiswellknownintheliterature,i.e.d-241;0destabilizesandd0stabilizesthedrainageofthinliquidlmslms.3.One-dimensionalthinliquidlmsattachedtoframesWeconsiderthesimplecaseofone-dimensionalthinliquidlmattachedtoaframe(Fig.2).Theparametersoftheframeare:width;left-andright-hand-sideheights,respectively.TheowratesfrombothframechannelsarecharacterizedbytheFig.2),thelmsurfacesaresymmetricwithrespecttothemiddleplaneandhavexedcontactlinesattheframeborders.Atinitialtime,=0,thelmthicknessisgivenbythefunction),whichdeterminestheinitialliquidvolumecaughtintheframe.Themodelequations,Eqs.(2.10)and(2.14)describingthedynamicsofsuchlms,aresimpliedtothe ohotþ ooxðhwÞ¼0;ð3:1Þ owotþw owox¼ 1We o3hox3þ 4Re o2wox2þ 4Re owox olnhoxþ oox isthedimensionlessvolumeoftheliquidcaughtinthelm,perunitchannellength;,andarethedimensionlesslat-eralcoordinate,thicknessandvelocity,respectively;andtheclassi-calWebernumberisrescaledwiththesmallparameterasfollows 2eCaRe¼2 qa2U2rb¼ Atinitialtime,=0,theliquidisatrest,thelmproleisaknown),andtheliquidvolumeis.Therefore,theinitialconditionsfornumericalcalculationofEqs.(3.1)and(3.2)Thelmsurfaceshavexedcontactlineswiththeframe,whichas-signsthefollowingboundaryconditionsforthelmthickness:Theowratesfrombothsidesoftheframeareknown(Fig.2);theydenetheboundaryconditionsforvelocityisthecharacteristiclateralvelocity(seeSection).TheminussignintherstequationofEq.accountsforthedirec-tionoftheow(Fig.2).InSectionwewillstudytheprocessoflmdrainagetothepointofminimumpossiblelmthickness,Inthiscasetheowratesarenon-zeroduringthewholeprocess.Inthelmthinsundertheactionofthesuctionvelocities,,foratime,,necessarytoreachagivenvolume,.Atsuctionisstoppedandthelmisallowedtorelaxtoitsnalshape.Inthiscasetheright-handsidesofbothequationsinEq.aresettozerofor,whichdoesnotchangethenumericalschemede-scribedinAppendixBThedisjoiningpressureappearinginEq.canbepresentedasasumofthevanderWaalsandhydrophobicattraction,,elec-trostaticrepulsion,,stericrepulsion,,andotherinteractions.Hereweconsideronlytherstthreecomponents:.Therespectivedimensionlesscomponentsofthedisjoiningpressurearedenotedas,and.ThevanderWaalsandhydrophobiccomponentsofthedisjoiningpressurearecalculatedbytheexpression),whereaneffectiveconstantconstant.ItisimportanttonotethatrecentexperimentssuggestthehydrophobicinteractionsareinmanycasesstrongerthanthevanderWaalsinteractioninteraction .TheauthorsshowthatthehydrophobicinteractionsaredescribedbythesameexpressionasthatforthevanderWaalscomponentofthedisjoiningpressure,butwithmuchlargerinteractionconstant,,thantheHamakerconstant,.Forthatreasontheeffectiveisusedbelow.FromEqs.(2.13)and(3.3) Fig.2.Sketchofone-dimensionalsymmetricthinliquidlmattachedtoaframewithaleft-hand-sideheight,right-hand-sideheight,andwidth2.Theinitiallmthicknessis),theowratesattheleftandrightbordersoftheframearecharacterizedbyvelocities,respectively.S.S.Tabakova,K.D.Danov/JournalofColloidandInterfaceScience336(2009)273 284 theattractivecomponentofthedisjoiningpressureispresentedinthefollowingform: p1We¼ B1h3andB1 isthedimensionlessinteractionconstant.Inthecaseof1:1electrolytestheelectrostaticcomponentofthedisjoiningpressure,,iscalculatedfromtheexpression1),whereisthethermalenergy,istheionicstrengthofthesolution,andisthedimensionlesselectricpo-tentialatthemiddleplaneplane.FromEqs.(2.13)and(3.3)weob-taintherelationshipbetweentheelectrostaticcomponentofthedimensionlessdisjoiningpressureandthedimensionlessionic p2We¼B2ðcoshUm1ÞandB2 ItiswellknownintheliteratureliteraturethatthesolutionofthePois-son Boltzmannequationfortheelectricpotentialinthelmleadstothetranscendentalequation, 1jbZUsUm whichrelatestheelectricpotentialatthemiddleplane,,withthesurfaceelectricpotential,,thelocallmthickness,,andtheDebyescreeninglength,.Foragivensystemthesurfaceelectricpotentialchangesinsignicantlyduringthelmdrainage.Onlysmallvariationsofarepossibleforverythinlmswhenthelocallmthicknessissmallerthan10nmnm.Belowwewillas-sumethat:(i)theconstantvaluesofareknown;(ii)thenumericalsolutionofEq.providesthedependence);(iii)Eq.).VarioussimplerapproximationsofEqs.(3.8)and(3.9)areknownintheliteratureliterature.Theseapproximationsarevalidfor1or1andcannotbeappliedforthewholelmprolewheresignicantly.Thepowerlawexpressionsforthestericcomponentofthedis-joiningpressure,widelyusedintheliteratureliteratureforproteinandpolymersolutions,arenotconsistentwithourmodel,asseeninthediscussionafterEq..Inthecaseoflowmolecularweightsurfactants,theadsorbedmolecules,forwhichourmodelcanbeapplied,oftenpossessbulkyhydrophilicheads,e.g.,whenthepolarpartoftheamphiphileconsistsofoxyethylenechainschains .Asthelmthins,thetwosurfacespackedwithsurfactantmoleculescloselyapproachoneanotherandthehydrophilicheadsmaybegintooverlap.Letbethecharacteristicthresholddistanceatwhichthestericinteractionbecomesactive.Therepulsiveforcewillbedescribedbymeansofamodelexpressionfortherespec-tivedisjoiningpressurecomponent,),whereistheinteractionconstant.Thevalidityofthisasymptoticequationisdiscussedbymanyauthorsauthors.FromEqs.(2.13)and(3.3)oneobtainstheexpressionsforandfortheinteractionconstant, p3We¼B3exp hbdandB3 Thenonlinearsystemofequations,Eqs.(3.1)and(3.2),withini-tialandboundaryconditions,Eqs.,canbesolvednumericallybyappropriateconservativenitedifferenceschemeonstaggeredgridsanditerativealgorithmsalgorithms.Here,wedevelopafasternumericalmethodbrieydescribedinAppendixB.Wechosearegularspaceandtimegridwithalengthstep,,andtimestep,.Theprecisionofthenumericalscheme()is),whichguaranteeshighaccuracyandefciencyofthealgorithm.Allcalculationsbelowareperformedwith=10.Thesesmallvaluesofmakeitpossibletocalcu-latetheformulatedproblemuptotheminimumlmthickness,,oftheorderoforlargerthan10.Moreprecisely,thevaluedependsontheprecisionofthedifferenceformulausedfortherstderivativeofdisjoiningpressure,Eq..Inthecaseofacentraldifferenceschemewithprecision)therelativeerrorofthederivativecalculatedat=10is3.4%.Thiserrorde-creaseswiththeincreaseof:for=20therelativeerroris0.84%;for=30itis0.37%,etc.4.Drainageofone-dimensionalthinliquidlms4.1.NodisjoiningpressureWhenthedisjoiningpressurehasnoeffectonthelmdynam-ics,themodel,givenbyEqs.(3.1)and(3.2),containstwodimen-sionlessnumbers,ReandWe,denedbyEqs.(2.4)and(3.3)Thesenumberschangeinadifferentwayforgivenphysicochemi-calparametersoftheliquidphasedependingonthevelocity,andontheratio.Forexample,ifforagivencongurationRe=1andWe=1(solidlinesinFig.3a),thenthevetimesin-creaseofthevelocity,,leadstoRe=5andWe=25(dashedlinesFig.3a),whilethevetimesdecreaseofreducesthesenum-berstoRe=0.2andWe=0.04(solidlinesinFig.3b).Foraxedva-lueofRetheWebernumberincreaseswiththeincreaseof Fig.3.Timeevolutionofthelmprolesinthecaseofsymmetricboundary,andaninitialcondition)=1:(a)Re=1andWe=1(solidlines),Re=1andWe=10(dash-dottedlines),andRe=5andWe=25(dashedlines);(b)Re=0.2andWe=0.04(solidlines)andRe=1andWe=0.1(dashedlines).S.S.Tabakova,K.D.Danov/JournalofColloidandInterfaceScience336(2009)273 284 ,andWedecreaseswiththeincreaseofdensity,frameheight,,andsurfacetension,,seeEq..Notethatthepossiblechangesofthesurfacetensionoffoamlmsarefrom72mN/m(surfactant-freesystems)to25mN/m(closeorabovethecriticalmicelleconcentration),whiletheotherparameters()canvaryinordersofmagnitude.Theexamplesfortheevo-lutionofthelmprole,),inthecaseofRe=1,We=10andRe=1,We=0.1areillustratedinFig.3a(dash-dottedlines)andFig.3b(dashedlines),respectively.Forsymmetricboundaryconditions,,andaconstantvalueoftheinitiallmthickness,)=1,theinitiallmvolumeis=2andthedimensionlesstimeneededforthecom-pletesuctionofthisvolumeisequalto1.ForRe=1andWe=1thelmdrainsupto=0.7withoutdimpleformationandwithavirtuallyxedlengthofitsplane-parallelregion(Fig.3a solidlines).Subsequentlythelmringspreadstotheborders,apro-nounceddimpleisformed,andat=0.9thelmrupturesbecausethemenisciatthelmperipherytouchtheframeborders.There-fore,theruptureisnotaresultofreachingtheminimalpossible.Withthevetimesincreaseoftheowrateawellpronounceddimpleisformedatanearlystage,thedimplegrows,thelmringspreads,andat=0.6thelmrupturesbeforereach-Fig.3a dashedlines).Atthenalmomentthedimpleheightat=0hasquitealargevalueof0.6,thusthedimplecovers40%oftheinitiallmvolume.IncreasingtheWebernumber(Re=1andWe=10)resultsinthelmsurfacesbecomingmoredeformable(Fig.3a dash-dottedlines).Inthiscasetheevolution)issimilartothecasewiththetentimessmallerWebernumber:thethicknessesofthetwolmsat=0areclosetoeachother;thelmringforWe=10isinitiallylargerthanthatforWe=1,butatthenalstagebothlmprolesarevirtuallyidenti-cal.Therefore,theWebernumbercanbeconsideredasacharacter-isticofthedeformabilityoflmsurfaces.Fromapracticalviewpointitisveryimportanttoproducelmswiththicknessasuniformaspossibleandwithalargeplane-par-allelarea.Onewayistoreducetheowrate:asitisshowninFig.3b(solidlines),thevetimesdecreaseoftheowrate(Re=0.2,We=0.04)leadstoaregularlmthinningwithacontin-uousincreaseoftheplane-parallelregionupto=0.9whenthelmthicknessat=0is4.25andthelmringisplacedapproximatelyat0.8.Anotherwayistodecreasethedeforma-bilityoflmsurfaces:forRe=1thetentimesdecreaseoftheWe-bernumber,We=0.1,stabilizestheprocessoflmdrainageFig.3b dashedlines).Suchlmshaveregularproles,),dur-ingthewholeprocess.Atmoment=0.9theirthicknessat=0is,i.e.smallerthanthatforthecaseofRe=0.2andWe=0.04,andthelmringisplacedapproximatelyatthesameInthecaseofunsymmetricalboundaryconditionstheproduc-tionoflmswithlargeplane-parallelareasisamorecomplicatedtask.Onecanexpectthattheincreaseofthesurfacetensionwillhelptosolvethisproblem.OurcalculationsforRe=1,We=0.1,)=1,andwithoutaliquidowfromtheleft-hand-sideboundary(Fig.4a)showthatinthiscaseadimpledoesnotform,i.e.alargeamountofliquidisleftcapturedclosetotheleft-hand-sideboundary.Becauseofthetwicesmallerowrate(onlyfromoneside)thedimensionlesstimenecessaryforthecompletesuctionoftheinitialvolumeisequalto2.Thelmreachesitsmin-imalthicknessslightlyafter=1.4withacomplexshapewithoutplane-parallelregions.Theminimallmthicknessislocatedat=0.75.Theprocessoflmdrainageisregular,withoutwavesandinstabilities.WehaveperformedcalculationsforthesamecasedecreasingtheWebernumbertoWe=0.01.ThelmproleatthemomentofreachingofisplottedinFig.4a(dashedline),whereitcanbeseenthatislocatedat=0.38.Thefollowingconclusionscanbedrawn:theselmsreachtheminimalthicknessatapproximatelythesametime;thedecreaseofdeformabilityoflmsurfacesleadstomoresymmetriclmthicknessesbuttheplane-parallelareasofthelmsareverysmallforunsymmetricalboundaryconditions.Toextendtheplane-parallellmregionwehavedecreasedvetimestheheightoftheleft-hand-sideborderassumingthat=0.2andhaveusedthesameconditionsasinFig.4awithasmal-lervalueoftheWebernumber,We=0.01(Fig.4b).Startingwithaninitialcondition,)=0.6+0.4,correspondingtoaninitialdimensionlessvolume,=1.2,wehaveobtainedthattheevolu-tionofsuchlmsisregularuntiltheminimallmthickness,isreachedat=0.9and=0.23.Therefore,theconsiderablede-creaseofthedeformabilityoflmsurfacesandtheframeheight,,isnotsufcienttoproducelargeplane-parallellmareas.ThecalculationsillustratedinFigs.3and4showthatthemostefcientwaytoaccomplishthistechnologicaltaskistousesymmetricowratesandsymmetricgeometricalparametersoftheframesasmuchaspossible.4.2.EffectofvanderWaalsandhydrophobicattractionThevanderWaalsandhydrophobicattractions,accountedforbythedisjoiningpressurecomponent,arealwaysactiveandhavethelongestrangecomparedtotheothercomponentsofTheroleofthiskindofattractionforcesontheevolutionofthelmprolesinthecaseofsymmetricboundaryconditionsforRe=1, Fig.4.Plotsof)forlmdrainageintheabsenceofaowfromtheleft-hand-sideborder:(a)andinitialcondition)=1atRe=1,We=0.1,till=1.4(solidlines)andatRe=1,We=0.01,=1.4(dashedline);(b)Re=1,We=0.01,=0.2,andinitialcondition)=0.6+0.4S.S.Tabakova,K.D.Danov/JournalofColloidandInterfaceScience336(2009)273 284 We=1,and=0.1isillustratedinFig.5a(solidlines).Thecom-parisonbetweenFigs.3aand5a(solidlines)showsthatuptotime=0.7thedisjoiningpressuredoesnotconsiderablyaffectthepro-cessofthelmdrainage.Whenthelocallmthicknessdecreasestoacertaincriticalvalue,theattractionforcesbecomesignicantandtheareaofthelmringthinsfaster.Attime=0.818thelmatposition=0.984andbreaks.Thenaldimplecov-ers1.82%oftheinitialliquidvolume.Oneexpectsthatthedecreaseoftheowratealwaysmakesthelmdrainagemoreregular(seeFig.3).Whenkeepingalltheotherparametersconstantanddecreasingvetimestheowrate,thevalueofincreases25times,seeEq..TheobtainednumericalresultsforRe=0.2,We=0.04,and=2.5areplottedinFig.5a(dashedlines).ThecomparisonoftheseresultswiththoseplottedinFig.3b(solidlines)inthecasewithoutdisjoiningpressuretakenintoaccountshowsthatbothlmsfollowthesameregularbehavioruptotime=0.4.Atsubsequentmomentstheattractionmolecularforcessig-nicantlyacceleratethelmthinninginthecentralzone,(comparetherespectiveprolesfor=0.5,0.55,0.58,and0.591).=0.591atypicalinstabilityappearsandthelmbreaks.Thiseffectisknownintheliteratureaspimpleformationasopposedtothecaseofthelmbreakageattheringring.4.3.EffectofelectrostaticandstericrepulsionOnepossibilitytoreducethedestabilizingroleoftheattractivecomponentsofthedisjoiningpressureistouseionicsurfactantsandthusincreasetheelectrostaticcomponentof.Itisprovenintheliteraturethatevenwithoutaddedsurfactantsthesurfacepotentialisdifferentfromzero:itsoriginforsuchsystemsisthesurfacechargeduetothespontaneousadsorptionofhydroxylionsions.Inthecaseoflargevaluesoftheasymptoticexpressionfortheelectrostaticcomponentofthedisjoiningpressure,=644.Therefore,amorecon-venientdimensionlessparameter,characterizingthemagnitudeofthisforceis/4),seeEq..InFig.3(dashedlines)wehaveillustratedthatatRe=1thedecreaseoftheWebernumbertoWe=0.1leadstoaregulardrainageofsuchlms.Ifwetakeintoaccounttheactionoftheattractivecompo-nentofthedisjoiningpressurebyassumingthat=1andcalcu-latethelmevolution,weobtainthesolidlinesinFig.5b.Itisclearlyseenthatafter=0.4theproles)accelerateintime,=0.537apimpleinstabilityappearsatthecentralzoneandsubsequentlythelmbreaks.NotethatthelminFig.3bobtainedatthesameprocessparametersbutwithouttheactionofhasawideplane-parallelregionat=0.9.InordertohaveanideaaboutthemagnitudeofonecalculatesitsvalueforwatersolutionwithoutsurfactantsatpH=9androomtemperature:=45We).TakingthevaluesWe=0.1,m,and=5mmweob-tainthat=1.1andtherespectivevalueofis10.3.ExperimentsforlmdrainageandequilibriumthicknessofsimilarsystemsarereportedinRef.Ref..Theincreaseofthesurfacepoten-tialandionicstrengthincreases.Thelm,whichisunstablewithoutaccountingfortheelectrostaticrepulsion,be-comesmorestablefor=10=10,seeFig.5b(dashedlines).ThecomparisonbetweenFigs.3band5b(dashedlines)showsthatinthepresenceofdisjoiningpressuretheelectrostaticrepulsioncounterbalancestheattractionforcesinawideplane-parallelcentralzone.Nevertheless,atthelateststageoflmdrain-ageawellpronouncedlmringat=0.864isformedandsubse-quentlythelmbreaksdown.Thedecreaseofthesurfacepotential,,suppressestheelec-trostaticinteractions.Forexample,ifwetaketheabovesystemandchangepHto5whilekeepingalltheotherparameterscon-stants,thevalueofdecreasestentimes.Thesmallerrepulsiveforceisnotsufcienttopreventthepimpleformation,seeFig.6(solidline).ThecomparisonbetweenFigs.5band6(solidlines)illustratesthattheelectrostaticcomponentofdeceleratestheappearanceofsuchkindofinstabilities.Thelm(Fig.6,solidline)thinsregularlyforalongertime=0.65comparedtothatwithout,whichthinsinasimilarmannerupto=0.52(Fig.5b,solidline).Nevertheless,at=0.682(Fig.6,solidlines)theattractivecomponentsofthedisjoiningpressureprevailandthelmrup-turesatitscenterline.Inaddition,theshorterrangerepulsiveforcecanbeincluded,namelytheeffectofstericinteractionscanbeinvestigated.Asarulerule forsuchforcesismuchlarger,exceptforthecasesoflargeionicstrengths,andtheinter-actionconstant,,isverylarge,seeEq..InFig.6lines)wecalculatethelmevolutionadditionallyaccountingforthestericinteractionsat=20and.Inthiscasethedrainageprocessstabilizesandregularlmthinningisob-servedupto=0.9.Themechanismofthelmrupturechangestoaringformationatpositionandtime,=0.916and=0.921,respectively,andalargeplane-parallelcentralzonewiththickness0.064isformed.Figs.5band6showthatthedynamiceffectscan-notbesuppressedbytheactionoftherepulsivedisjoiningpressure.5.Relaxationofone-dimensionalthinliquidlmsIfonewantstoproducealmwithagivenvolume,,theowratesarestoppedaftertime,.Dependingonthemagnitudesof Fig.5.Roleofdisjoiningpressureonthedrainageoflmswithsymmetricboundaryconditionsandinitialcondition)=1:(a)Re=1,We=1,=0.1(solidlines)andRe=0.2,We=0.04,=2.5(dashedlines);(b)Re=1,We=0.1,(solidlines)andadditionally=10,and=10(dashedlines).S.S.Tabakova,K.D.Danov/JournalofColloidandInterfaceScience336(2009)273 284 interactionforcesandregimesoflmdrainagethelmsurfaceswillrelaxtothesteadystatelmprole,),oreventuallythelmmayrupturebeforereaching)(seebelow).Inthecaseofthicklmswhentheroleofthedisjoiningpressurecanbene-glected,thestaticvariantofEq.,simpliestod=0.Thesolutionforlmswithsymmetricboundarycondi-(±1)=1,andaxedvolume,,is 3vf24þ Therefore,theminimumpossibledimensionlessvolume(minimumstaticequilibriumvolume)is2/3andif2/3theLaplaceequationofcapillarityhasnosolution,thusnostaticequilibriumexists.Therefore,ifonestopstheowratesat-248;2/3inthecaseshownFig.3therespectivelmswillbreak.Whenthedisjoiningpres-sureistakenintoaccount,thestaticvariantofEq.hasnoana-lyticalsolutionand)isobtainedasalimitof)aftercompleterelaxation.Therststepinourstudyistoobtainthecharacteristicrelaxa-tiontimeandthetypeoflmprolerelaxationinthesimplestcase,withoutthedisjoiningpressureeffectbeingtakenintoac-count.WeconsidertwocasesforRe=1,We=1,andvolume=1,largerthantheminimumpossibleone,2/3:initiallmpro-leis)=1(atsurfaces)and=0.5,andparabolicinitialpro-)=(5+3)/8with=1.5and=0.25(Fig.7).ThedrainageoftheselmsisillustratedinFig.3a(solidlines)andFig.7a(dashedlines);therespectiverelaxationcurvesshowasim-ilarbehavior.ThesolidlinesinFig.7ashowtheevolutionoflm),duringtheprocessofrelaxationfor.Thestableregularapproachtotheequilibriumstate)=(1+3)/4iswellillustrated.Thechangesofthethicknessdifference,withtimeforbothsetsofinitialconditionsareshowninFig.7Oneseesthatinthesecasestheinitialconditionsdonotaffectthetypeofrelaxationforallvaluesofthelateralcoordinate.IntheinsetofFig.7bwepresentthechangesofthicknessdifferenceinthecentralzone,(0),inalogarithmicscaleasafunc-tionof.Theexcellentparallelstraightlineswithaslopeof1.5provesthattherelaxationobeysanexponentialrule,),forbothkindsofinitialconditions.Therefore,theelasticforcearisingfromtheactionofthesurfacetensionisstrongenoughtostabilizesuchlms.Theattractioncomponentofthedisjoiningpressuredestabi-lizesthelmsnotonlyinthelmdrainageprocess(seeFig.5butalsoduringthelmrelaxation.IfwetakethelmsgiveninFig.5aandstoptheowratesat=0.45,whentheliquidvolume=1.1islargerthanthatforthelmsinFig.7,thelmscannotreachthesteadystateandthereforerupture,asseenonFig.8.InthecaseofRe=1,We=1,and=0.1thechangeof)for0.45withatimestep0.1isshowninFig.8.Thepimpletypeinstabilityappearsat=2.245andsubsequentlythelmrupturesinitscentralzone.TheinsetofFig.8demonstratesthedecreaseof,0)withtimeforthislm(solidline)andforthelmwiththesamephysicochemicalparameters,butwith5timessmallersuc-tionvelocity(dashedline).Oneseesthatinbothcasestheinstabil-ityappearsat=0andforsmallerowratesthedimensionlessrupturetimeof0.707forRe=0.2islowerthanthatof2.245forRe=1.Takingintoaccountthedenition,Eq.,oneconcludesthatduetothe5timessmallerowratetherealrupturetime,forRe=0.2becomeslonger(50.707=3.535)thanthatforRe=1.Thisfactcanbeexplainedbythelongertimeneededtosup-pressthelargerlocalsurfacevelocitiesafterstoppingthelmdrainageatlargerowrates.Asdiscussedabove,theproductionofplane-parallellmswithsmallvolumes,2/3,andthicknessesisimpossiblewithoutusingadditives,whichincreasetherepulsivecomponentsofthedisjoiningpressure.However,theattractivedisjoiningpressureal-waysappearsandcannotbeexcludedfromconsideration.ThedashedlinesinFigs.5band6representregularlmdrainagein Fig.6.Roleofelectrostaticandstericcomponentsofthedisjoiningpressureforthedrainageoflmswithsymmetricboundaryconditionsandinitialcondition)=1forRe=1,We=0.1,=1,=10=10(solidlines),and=20(dashedlines). Fig.7.RelaxationoflmsforRe=1,We=1,and=1:(a)Locallmthickness,),withinitialprole,)=(5+3)/8,andinitialvolume,=1.5,=0.25(solidlines).For0.35solidlinescorrespondtotimestep0.2;(b)thickness),linesplottedwithtimestep0.1for)=1,=2,=0.5(solidlines)andforcase(a)(dashedlines).Theinsetshowsthedifference(0)intime,,forthetwocasesafterthelmsuctionisstopped.S.S.Tabakova,K.D.Danov/JournalofColloidandInterfaceScience336(2009)273 284 thepresenceof.Wetookthesesystemsandstoppedtheowratesattwodifferenttimes,at=0.7when=0.6,andat=0.85whenthevolumeissmaller,=0.3.Forbothsystemsandtimesthenallmproles,),areshowninFig.9.Oneseesthatthelmscontainlargealmostplane-parallelcentralzonesandsteepmenisciaroundthem.Theselmsarestablebecausetherepulsivecomponentsofthedisjoiningpressure,,prevailtherespectivevanderWaalsandhydrophobiccomponents,TheinsetofFig.9illustratesthetypeofrelaxationoftheselms inallcasesthelmthickness,,0),foroscillatesaroundtheequilibriumthickness,(0),buttheamplitudesofoscillationsexponentiallydecreasewithtime.Byapplyinglinearstabilityanal-ysis,analogoustothatgiveninRef.Ref.,forthedisturbancesofthelmsurfacesinacentralplane-parallelzoneoneprovesthatthecriticalwavelengthsforthecasesillustratedinFig.9aremuchlar-gerthanthelateraldimensionoftheframe.Therefore,thenallmthicknessescorrespondtoastableequilibrium.Theseconclusionsareimportantforpracticalapplicationsbe-causetheygiveonepossiblewayforacontrollableproductionoflargethinlms.6.SummaryandconclusionsAmodeldescribingthedrainageandrelaxationdynamicsofsymmetricthinliquidlmswithtangentiallymobilesurfacesisdeveloped.Theobtainedequations,Eqs.(2.10) (2.12),areapplica-bleforlaminarowswhenthelmthicknessesaremuchsmallerthanthecharacteristiclateraldimensions.Themodeldescribesthelmprole,surfacevelocityandpressureforarbitraryvaluesoftheReynoldsandcapillarynumber,accountingfordifferentcompo-nentsofthedisjoiningpressure.Inthecaseofone-dimensionallmattachedtoaframetheproblemissimpliedinSectionandthevanderWaals,hydrophobic,electrostaticandstericinter-actionsbetweenlmsurfacesareincludedinthedisjoiningpres-sure.Afastalgorithmfornumericalsolutionofthisproblem,whichhasasecondorderprecisionintimeandspace,isdevelopedAppendixBandappliedinSections4and5Inthecaseofunsymmetricalboundaryconditions(differentowratesorframeheights)thelmsdrainwithcomplexshapesoftheirsurfaces,thereforetheformationoflargeplane-parallelareasbycontrollingtheowratesisdifcult(Fig.4).Forlmswithsymmet-ricalboundaryconditionstheincreaseofowratesleadstoapro-nounceddimpleformationwitharingclosetotheframe(Fig.3Theprocessoflmdrainagebecomesmoreregularwithalmostnodimpleformationbydecreasingtheowrateorthedeformabilityofthelmsurfaces,i.e.forsmallervaluesoftheReynoldsandWebernumbers(Fig.3b).Theattractivedisjoiningpressurefavorsthelmbreakageintwodifferentways:forlargerowratesthedimplerup-turesatitsring;forsmallerowratesthelmbreaksinitscentralzone(Fig.5).TheelectrostaticandstericrepulsivecomponentsofthedisjoiningpressurecansuppressthevanderWaalsandhydro-phobicattraction,makingtheprocessoflmdrainageregular,withawidecentralplane-parallelzone(Figs.5and6).Thetypeoflmdrainagecanbeefcientlycontrolledbyvaryingthemagnitudesofdifferentcomponentsofthedisjoiningpressure.Whentheowratesarestopped,thelmsurfaceseitherrelaxtotheirnalequilibriumprole,orthelmruptures.ForagivenliquidvolumetheLaplaceequationofcapillaritydescribesthesteadylm).Withoutaccountingfortherepulsivecomponentsofthedisjoiningpressure,)hasaparabolicorpimpleformwith-outplane-parallelareas.Thelmscanbeeitherstablewithanexpo-nentialdecayofsurfacecorrugations(Fig.7)orunstableifattractivesurfaceforcesarepresent(Fig.8).Theonlypossiblewaytoproducelmswithsmallvolumesandthicknessesistocontrolthemagni-tudesofelectrostaticandstericinteractions(Fig.9).Theselmsarestable:theirthicknessesoscillatearoundtheequilibriumpro-),andtheamplitudesofsurfacecorrugationsexponentiallydecreasewithtime.Theequilibriumprolescontainlargealmostplane-parallelcentralzonesandsteepmenisciclosetotheframes.ThemodeldescribedinSectioncanbeappliedtostudythedrainageandrelaxationdynamicsofcircularlms.Howeverinthiscasethelmsthinundertheactionofaconstantpressurediffer-ence,whichchangestheboundaryconditions.Thechangeofgeometryandboundaryconditionsmustbetakenintoaccountinthemodelequationsandnumericalscheme.Thisproblemisanaimforourfuturepublication.AcknowledgmentThisstudyispartiallyfundedbyprojectINZ01/01024oftheNa-tionalScienceFundofBulgaria(programIntegratedResearchCentersintheUniversities). Fig.8.RelaxationoflmsurfacesinthecaseofsymmetricboundaryconditionsforRe=1,We=1,=0.1,and=0.45.Solidlinesareplottedwithtimestep0.1andthemomentbeforethelmruptureis2.245.Theinsetcomparestherelaxationof,0)forthislm(solidline)withthecaseofrelaxationofthesamelmbutwithvetimessmallerowrateRe=0.2,We=0.04,=2.5,and=0.45(dashedline). Fig.9.EquilibriumlmthicknessesforRe=1,We=0.1,and=1correspondingto=0.7and=0.85.Solidlinesareplottedfor=10=10withoutaccountingforthestericinteractions(seeFig.5b)anddashedlines for=10=10,,and=20(seeFig.6).Theinsetillustratestherespectiverelaxationsof,0)afterstoppingtheowrates.S.S.Tabakova,K.D.Danov/JournalofColloidandInterfaceScience336(2009)273 284 AppendixA.DerivationoftheleadingorderofthedynamicboundaryconditionsatthelmsurfaceToobtainthedensityoftheforce,,actingfromthebulkuidontotheupperlminterfaceoneneedstocalculatethenormalprojectionofthebulkstresstensor,,atthesurface.Wewillusethefollowingnotations:=1,2,3)aretheunitvectorsoftheCartesiancoordinatesystems=1,2)arethetangen-tialvectorsattheupperlmsurface;istherespectivesurfacenormalvector.Therefore, 12 oHoXae31þ 1 oHoXa2"#1=2ða¼1;2Þ;ðA:1Þn¼ 1 oHoXbebþe31þ 1 oHoX12þ 1 ForincompressibleNewtonianuidsthestresstensor,,isgivenbythedenition oUioXkþ istheKroneckerdelta.FromEqs.onederivestheexpressionfortheforcedensity, Pkb2 oHoXbþPk31þ 1 oHoX12þ 1 whereallparametersarecalculatedat/2and=1,2,3.FromEqs.(A.1)and(A.4)weobtaintheformulaforthetangen-tialcomponentoftheforce: P332 oHoXa Pab2 oHoXb P3b4 oHoXb oHoXa1þ 1 oHoX12þ 1 oHoX22"#1=21þ 1 oHoXa2"#1=2atX3¼ Thetangentialstressboundarycondition(seeSection)forfreesurfacesreads=0.Substitutingthedenitionofthestresstensor,Eq.,intoEq.onereducesthetangentialstressboundaryconditiontothefollowingrelationship oUaoX3þ oU3oXaþ4 oU3oX3 oHoXa2 oUaoXbþ oUboXa oHoXb oU3oXbþ oUboX3 oHoXb 0at FromEqs.(2.1),(2.2)and(A.6)weobtainthedimensionlessformofthisboundarycondition: 1e2 ouaox3¼ ou3oxa ou3ox3 ohoxaþ 1 ouaoxbþ ouboxa ohoxbþ 1e2 ou3oxbþ oubox3 ohoxb ohoxaatx3¼ Usingtheleadingordersolutionoftheconsideredproblem,Eqs.(2.7)and(2.8),wesimplifytheright-handsideofEq.totheasymptoticexpression: 1e2 ouaox3x3¼h=2¼ h2 o2wboxaoxbþ ohoxa owboxbþ 12 owaoxbþ owboxa SubstitutingEq.intoEq.wereduceittotheasymptoticrelationship: 1e2 o2uaox23¼ ooxaq owboxb o2waoxboxbþRe owaotþ ooxbðwbwaÞþ IntegratingEq.withrespecttofrom0to/2andaccountingforthekinematicboundarycondition,Eq.,onederives 1e2 ouaox3x3¼h=2¼ h2 oqoxa h2 o2wboxaoxb h2 o2waoxboxbþRe oot h2waþ ooxb Therefore,theright-handsidesofEqs.(A.8)and(A.10)mustbeequalandaftersimpletransformationswearrivetotheresultgivenbyEq.inSectionFromEqs.(A.2)and(A.4)onecalculatestheexpressionforthenormalcomponentoftheforcedensity: Pab4 oHoXa oHoXb oHoXbPb3þP331þ 12 oHoX12þ 1 whereallparametersarecalculatedat/2.Bysubstitutingthedenitionsofthestresstensor,Eq.,andofthedimensionlessparameters,Eqs.(2.1),(2.2)and(2.4),inEq.wederivetherelationship gUapþ gUa2 ouboxbþ ohoxb oubox3þe2 ohoxb ou3oxb e24 ouaoxbþ ouboxa ohoxa ohoxb1þe2 1 ohox12þe2 1 writtenatFromtheleadingordersolution,Eq.,thetangentialstressboundarycondition,Eq.,andEq.oneobtains: gUaqþ Thenormalstressboundaryconditionreads(seeSectionSection)Fs\bnþPcþP¼Pg;ðA:14ÞwherePgisthepressureinthegasphase,isthelmcapillarypressure,andisthedisjoiningpressure.FromEqs.(A.13)andtheleadingorderformofthenormalstressboundarycondi-tionisreducedto owboxbþ rb2gUa o2hoxboxbþ 2Pa2rb BydeningthecapillarynumberandthedimensionlessdisjoiningandgaspressuresbyEq.wetransformEq.intothe-nalresultgivenbyEq.AppendixB.NumericalschemeToconstructthenumericalschemewedenotethevaluesofallparametersattimewiththesuperscript(-)andthosecalculatedattimewiththesuperscript(+),whereisthelocalnumer-icaltimestep.Forexample,aretakenattimelevelS.S.Tabakova,K.D.Danov/JournalofColloidandInterfaceScience336(2009)273 284 atthesubsequenttimelevel.Therespec-tivetimedifferences,,aredenedas.Eqs.(3.1)and(3.2)areconsideredatmo-/2andthetimederivativesappearingthereinareapproximatedwithprecision DhDt¼RhjtþDt=2þOðDt2Þand wheretheoperators,,aredenedbytheexpressions ooxðhwÞ;ðB:2ÞRw oox w22þ 1We o3hox3þ oox pdpWeþ 4Re o2wox2þ 4Re owox UsingtheCrank-Nicolsonmethoddtheright-handsidesoftheequationsinEq.arereplacedwiththemeanvaluesofoperatorsattimelevelswithaccuracy),i.e. DhDt RðþÞhRðÞ2¼RðÞþOðDt2Þand DwDt InordertoapplytheNewtonmethodforlinearizationtionwecalculatethevariationsofwithsecondor-derprecision)usingtheabovedenitions,Eqs.(B.2)and.Aftersimpletransformationsonearrivestotherelationships DhDtþ 12 oox½hðÞDwþwðÞDh\t¼RðÞh;ðB:5Þ DwDtþ oox wðÞDw2 12We o3Dhox3 oox ddhð pdpWeÞðÞ Dh2() 2Re o2Dwox2 2Re olnhðÞox oDwox 2Re owðÞox oox TosolvenumericallyEqs.(B.5)and(B.6)weusearegulargridforthespacecoordinate,,denedas=0,1,2,...isthenumberofintervalswithalength=2/.Theval-uesofthefunctionsinthenodesofthegridaredenotedwiththe.Forexample,thecentraldifferenceinspacewithpreci-)appliedtoEq. wðÞk14DxDhk1 hðÞ14DxDwk1þ DhkDtþ wðÞþ14DxDhkþ1þ hðÞþ14DxDwkþ1¼RðÞh;k;ðB:7ÞwhereRðÞ;k¼ =1,2,Therespectivedifferenceschemewithprecision)appliedtoEq. lnhðÞþ1lnhðÞ12ReDx2 wðÞ14Dx 2ReDx2"#Dwk1þ wðÞkþ1wðÞ12ReDx2hðÞ1þ 14Dx ddh pdpWeðÞk1()Dhk1þ 1Dtþ 4ReDx2Dwk 12WeDkðDhÞþ wðÞkþ14Dx 2ReDx2 lnhðÞþ1lnhðÞ12ReDx2"#Dwkþ1þ 12WeDx3 14Dx ddh pdpWeðÞkþ1 wðÞkþ1wðÞ12ReDx2hðÞkþ1()Dhkþ1¼RðÞ;k;ðB:9ÞwhereRðÞ;k¼ wðÞ1wðÞ1wðÞþ1wðÞþ14Dxþ 1WeDk½hðÞ\tþ pðÞdp;kþ1pðÞ;k12WeDxþ 4Re wðÞþ12wðÞþwðÞ1Dx2þ isthedifferenceoperatorrepresentingthethirdderivativewithrespecttothespacecoordinateBecauseofthethirdderivativewithrespectto,thedifferenceoperator,,inthenodes=2,3,2hasdifferentformthanthosewrittenforthenodes1and1.For=2,3,2weusethefollowingcentraldifferenceexpressionfor isdenedbytherelationship iscalculatedusingthevaluesofthefunctioninthenodes1,andthroughequation Thematrixoftheobtainedlinearsystemofequations,Eqs.and(B.9),completedwiththeboundaryconditions,Eqs.(3.5)and,writtenintermsoftheunknownvector(),contains5upperand7lowerdiagonals.Thenumericalsolutionofthissystemisper-formedefcientlyusingtheThomasalgorithmalgorithm.Thedescribednumericalschemeisstable,ofthesecondorderoftime,andistherefore,applicablefor[1]J.Sjoblom(Ed.),EmulsionsandEmulsionStability,seconded.,CRCTaylorandFrancis,BocaRaton,2006.[2]D.Exerowa,P.M.Kruglyakov,FoamandFoamFilms:Theory,Experiment,Application,ElsevierScience,Amsterdam,1998.[3]I.B.Ivanov,K.D.Danov,P.A.Kralchevsky,ColloidsSurf.A152(1999)161 182.[4]I.B.Ivanov,P.A.Kralchevsky,ColloidsSurf.A128(1997)155 175.[5]I.B.Ivanov(Ed.),ThinLiquidFilms,MarcelDekker,NewYork,1988.[6]B.V.Derjaguin,N.V.Churaev,V.M.Muller,SurfaceForces,PlenumPress,NewYork,1987.[7]K.S.Birdi(Ed.),HandbookofSurfaceandColloidChemistry,thirded.,CRCTaylorandFrancis,BocaRaton,2008.[8]J.Blawzdziewicz,E.Wajnryb,M.Loewenberg,J.FluidMech.395(1999)29 59.[9]A.K.Chesters,I.B.Bazhlekov,J.ColloidInterfaceSci.230(2000)229 243.[10]J.Blawzdziewicz,V.Cristini,M.Loewenberg,J.ColloidInterfaceSci.211(1999)[11]M.D.Ida,M.J.Miskis,SIAMJ.Appl.Math.58(1998)456 473.[12]M.D.Ida,M.J.Miskis,SIAMJ.Appl.Math.58(1998)474 500.[13]I.B.Ivanov,PureAppl.Chem.52(1980)1241 1262.[14]A.K.Malhotra,D.T.Wasan,AIChEJ.33(1987)1533 1541.[15]D.E.Tambe,M.M.Sharma,J.ColloidInterfaceSci.147(1991)135 151.[16]D.S.Valkovska,K.D.Danov,J.ColloidInterfaceSci.241(2001)400 412.[17]P.deRoussel,D.V.Khakhar,J.M.Ottino,Chem.Eng.Sci.56(2001)5511 5529.[18]S.D.Stoyanov,N.D.Denkov,Langmuir17(2001)1150 1156.[19]D.S.Valkovska,K.D.Danov,I.B.Ivanov,Adv.ColloidInterfaceSci.96(2002)[20]M.B.Williams,S.H.Davis,J.ColloidInterfaceSci.90(1982)220 228.[21]M.Prévost,D.Gallez,J.Chem.Phys.84(1986)4043 4048.[22]T.Erneux,S.Davis,Phys.FluidsA5(1993)1117 1122.[23]A.deWit,D.Gallez,C.I.Christov,Phys.Fluids6(1994)3256 3266.[24]D.Vaynblat,J.R.Lister,T.P.Witelski,Phys.Fluids13(2001)1130 1140.[25]J.P.Burelbach,S.G.Bankoff,S.H.Davis,J.FluidMech.195(1988)463 494.[26]D.Hatziavramidis,J.MultiphaseFlow18(1992)517 530.[27]K.D.Danov,N.Alleborn,H.Raszillier,F.Durst,Phys.Fluids10(1998)131 143.[28]K.D.Danov,V.N.Paunov,N.Alleborn,H.Raszillier,F.Durst,Chem.Eng.Sci.53(1998)2809 2822.[29]K.D.Danov,V.N.Paunov,S.D.Stoyanov,N.Alleborn,H.Raszillier,F.Durst,Chem.Eng.Sci.53(1998)2823 2837.S.S.Tabakova,K.D.Danov/JournalofColloidandInterfaceScience336(2009)273 284 [30]V.N.Paunov,K.D.Danov,N.Alleborn,H.Raszillier,F.Durst,Chem.Eng.Sci.53(1998)2839 2857.[31]S.Tabacova,L.Carotenuto,MicrogravityQ4(1994)55 61.[32]A.Graciaa,G.Morel,P.Saulner,J.Lachaise,R.S.Schechter,J.ColloidInterfaceSci.172(1995)131 136.[33]K.G.Marinova,R.G.Alargova,N.D.Denkov,O.D.Velev,D.N.Petsev,I.B.Ivanov,R.P.Borwankar,Langmuir12(1996)2045 2051.[34]V.N.Paunov,S.I.Sandler,E.W.Kaler,Langmuir17(2001)4126 4128.[35]L.Wang,R.-H.Yoon,Langmuir20(2004)11457 11464.[36]J.K.Angarska,B.S.Dimitrova,K.D.Danov,P.A.Kralchevsky,K.P.Ananthapadmanabhan,A.Lips,Langmuir20(2004)1799 1806.[37]L.Wang,R.-H.Yoon,ColloidsSurf.A263(2005)267 274.[38]L.Wang,R.-H.Yoon,ColloidsSurf.A282 283(2006)84 91.[39]M.Bosröm,D.R.M.Williams,B.W.Ninham,Phys.Rev.Lett.87(2001)168103.[40]J.N.Israelachvili,H.Wennerström,J.Phys.Chem.96(1992)520 531.[41]K.D.Danov,I.B.Ivanov,K.P.Ananthapadmanabhan,A.Lips,Adv.ColloidInterfaceSci.128 130(2006)185 215.[42]L.G.Leal,LaminarFlowandConvectiveTransportProcesses:ScalingPrinciplesandAsymptoticAnalysis,Butterworth-Heinemann,Stoneham,MA,1992.[43]M.Boström,W.Kunz,B.W.Ninham,Langmuir21(2005)2619 2623.[44]N.C.Christov,K.D.Danov,P.A.Kralchevsky,K.P.Ananthapadmanabhan,A.Lips,Langmuir22(2006)7528 7542.[45]J.T.Petkov,N.D.Denkov,K.D.Danov,O.D.Velev,R.Aust,F.Durst,J.ColloidInterfaceSci.172(1995)147 154.[46]J.T.Petkov,K.D.Danov,N.D.Denkov,R.Aust,F.Durst,Langmuir12(11)(1996)[47]J.N.Israelachvili,IntermolecularandSurfaceForces,AcademicPress,London,[48]B.U.Felderhof,J.Chem.Phys.48(1968)1178 1185.[49]B.U.Felderhof,J.Chem.Phys.49(1968)44 51.[50]S.Sche,H.M.Fijnaut,Surf.Sci.76(1978)186 202.[51]A.A.Mohamed,E.F.Elshehawey,M.F.El-Sayed,J.ColloidInterfaceSci.169(1995)65 78.[52]L.Popova,G.Gromyko,S.Tabakova,Diff.Equat.39(2003)1037 1043.[53]D.S.Valkovska,K.D.Danov,I.B.Ivanov,ColloidsSurf.A156(1999)547 566.[54]S.I.Karakashev,E.D.Manev,R.Tsekov,A.V.Nguyen,J.ColloidInterfaceSci.318(2008)358 364.[55]D.A.Edwards,H.Brenner,D.T.Wasan,InterfacialTransportProcessesandRheology,Butterworth-Heinemann,Boston,1991.[56]C.A.J.Fletcher,ComputationalTechniquesforFluidDynamics.1.FundamentalandGeneralTechniques,Springer-Verlag,NewYork,1991.[57]C.A.J.Fletcher,ComputationalTechniquesforFluidDynamics.2.SpecicTechniquesforDifferentFlowCategories,Springer-Verlag,NewYork,1991.[58]M.Abramowitz,I.A.Stegun,HandbookofMathematicalFunctions,NationalBureauofStandards,Washington,DC,1984.[59]W.H.Press,B.P.Flannery,S.A.Teukolsky,W.T.Vetterling,NumericalRecipesinFORTRAN:TheArtofScienticComputing,CambridgeUniv.Press,NewYork,S.S.Tabakova,K.D.Danov/JournalofColloidandInterfaceScience336(2009)273 284