Lubrication in tablet formulations

EuropeanJournalofPharmaceuticsandBiopharmaceuticsjournalhomepage:www.elsevier.com/locate/ejpbReviewarticle

Lubricationintabletformulations

JenniferWanga,*,HongWenb,1,DivyakantDesaia,2abBristol-MyersSquibbCompany,NewBrunswick,NJ,UnitedStatesNovartisPharmaceuticalsCorporation,EastHanover,NJ,UnitedStatesarticleinfoabstract

Theoreticalaspectsandpracticalconsiderationsoflubricationintabletcompressionarereviewedinthispaper.Propertiesofthematerialsthatareoftenusedaslubricants,suchasmagnesiumstearate,intabletdosageformaresummarized.Themanufacturingprocessfactorsthatmayaffecttabletlubricationarediscussed.Asimportantasthelubricantsintabletformulationsare,theirpresencecancausesomechangestothetabletphysicalandchemicalproperties.Furthermore,adetailedreviewisprovidedonthemethodologiesusedtocharacterizelubricationprocessduringtabletcompressionwithrelevantprocessanalyticaltechnologies.Finally,theQuality-by-Designconsiderationsfortabletformulationandprocessdevelopmentintermsoflubricationarediscussed.Ó2010ElsevierB.V.Allrightsreserved.Articlehistory:Received11October2009Acceptedinrevisedform12January2010Availableonline22January2010Keywords:LubricationLubricantsTabletformulationTabletcompressionMagnesiumstearateQuality-by-DesignProcessanalyticaltechnology1.TheoreticalaspectsoflubricationduringtabletcompressionTabletcompressionleadstotheconsolidationofparticlesintoapelletofspecificstrength.Tabletcompressionnormallyresultsinparticlerearrangement,deformationofparticles,interparticulatebondformation,andelasticrecoveryuponejectionofthecompactfromthedie[1].Thepenultimatestepinatabletcompressionpro-cessisejection.Theejectionforceistheforceneededtopushthetabletoutofthedie.Significantdecreaseintheoverallejectionforceisobservedwhenthematerialand/orthedieareproperlylubricated.Theextentoflubricationalsobecomesimportantinthelaststepduringtabletcompressionwhenthetabletleavesthelowerpunch.Thetake-offforceistheforcerequiredtoscrapetheformedtabletoffthelowerpunchfaceafteritisejectedfromthedie.Lubricationismostrelevanttothetabletejectionandtab-lettake-offstepsasthelubricanthelpstoreducethefrictionbe-tweenthetabletandthemetalsurface,makingtheoveralltabletcompressionprocessmuchsmoother.Thephysicalbasisofeitherejectionforceortake-offforceduringtabletcompressionisintermolecularinteractionsofthepowderblend.Duetothethermodynamicnatureoftheintermo-lecularandinterparticulateinteractions,itiseasiertounderstandthemthroughenergyterms.However,itistheinteractionforcesbetweenmacroscopicbodiesthatareofteneasiertomeasureexperimentally.Forexample,interactionsbetweenthesidesofthetabletandthediewallaredeterminedbymeasuringthemag-nitudeofthetabletejectionforce,nottheejectionenergy.Thesameistrueforthetablettake-offprocess.Boththetabletejectionandtake-offforcesaremeasuresoftheadhesiveinteractionsbe-tweentwosurfaces.ArigoroustheoreticaltreatmentoftheadhesiveinteractionsofelasticsphereswasproposedbyJohnson,KendallandRoberts[2]basedontheassumptionthatrealparticles(surfaces)arenotcom-pletelyrigid.Sincethen,thetheoreticaltreatment,knownasthe‘‘JKRtheory”,hasformedthebasisofmuchofthemoderntheoriesofadhesionmechanics.Whiledifferentpharmaceuticalpowdersmaytendtowardelastic,plasticdeformation,orbrittlefragmenta-tion,mostpharmaceuticalingredients,however,exhibitmixedbehavior.AsindicatedbytheJKRmodel,theradiusofthecontactarea,a,canbecalculatedby:󰀅󰀃qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi󰀄󰀆1=3Ra¼Fþ3pRW12þ6pRW12Fþð3pRW12Þ2Kð1Þ*Correspondingauthor.Bristol-MyersSquibbCompany,OneSquibbDrive,NewBrunswick,NJ003,UnitedStates.Tel.:+17322275684.E-mailaddresses:jennifer.wang@bms.com(J.Wang),hong.wen@novartis.com(H.Wen),divyakant.desai@bms.com(D.Desai).1Tel.:+18627785558.2Tel.:+17322276548.0939-11/$-seefrontmatterÓ2010ElsevierB.V.Allrightsreserved.doi:10.1016/j.ejpb.2010.01.007whereKistheelasticconstant,W12istheinterfacialsurfaceenergy,Ristheradiusofthesphere,andFistheexternalforce.Ifzeroexter-nalforceisappliedtothesphere,thecontactradius,a0,isgivenby: a0¼6pR2W12K!1=3ð2Þ2J.Wangetal./EuropeanJournalofPharmaceuticsandBiopharmaceutics75(2010)1–15Becauseofadhesion,thespherecannotbeseparatedfromthesurfaceuntilthepull-offforcereachesacriticalvalue,FS,asgivenbythefollowingequation(Fig.1[3]):FS¼32pRW12ð3ÞInthecaseoftabletejection,insteadofbreakingtheadhesiveinteractionsbya‘‘pull-off”forceforwhichaverticalforceisex-ertedtopulloneparticleapartfromanother,itisassumedthatthe‘‘bond”betweenthesidesofthetabletandthediewallisbro-kenupbypushingthetabletoutofthedieduringtabletejection.Theseadhesiveinteractionsareoftenexpressedintermsof‘‘fric-tion”,whichcanbecharacterizedbythe‘‘coefficientoffriction”.Forthetablettake-offprocess,theadhesiveinteractionsarebrokenupbythescrapingaction.So,ifamaterialhastheabilitytoreducetheadhesiveinteractionsbetweenthetabletandthediewallorlowerpunchsurface,suchamaterialhasthepotentialofbeingalubricant.Therearetwobasicmechanismsbywhichalubricantaffectstabletcompression[4,5].Thefirstisbymeansofacontinuousfluidthinlayerinwhichthelubricant,suchasmineraloil,separatesthetabletsurfaceandthemetalsurface.Thecoefficientoffrictionofsuchfluidlubricantscanbeaslowas0.001.Thesecondmecha-nismisbymeansof‘‘boundarylubrication”.Thelubricantparticlesformacontinuousornon-continuousresistantlayerorfilmonthetabletingmaterialormetalsurfaces.Thecoefficientoffrictionofboundarylubricantscanrangefrom0.15to0.5.Thefluidlubricantsneedtobeusedinlargerquantitythanboundarylubricants,oftenmakingtheiruseimpracticalintabletformulations.AccordingtotheJKRtheory(seeEq.(3)),therearetwomainap-proachestoreducingadhesiveinteractionsduringtabletejectionortake-offwiththeboundarylubricationmechanism.Theyareasfollows:(1)reductionintheintrinsicadhesiveinteractions,i.e.,theinterfacialinteractions,betweenthetabletingpowdermate-rialandthemetalsurfaceand(2)reductioninthecontactareabetweenthetabletpowderparticlesandthediewallsurfaceorbetweenthetabletfaceandpunchsurface.Thecommonlyusedboundarylubricantsusuallyhaveoneorbothfunctions.Aneffective‘‘boundary”lubricanttypicallyhasFig.1.Criticalpull-offforceduetoadhesionofasphere(1)onaflatsurface(2).suchattributesaslowshearstress,relativelyhighmeltingpoint,largespecificarea/smallparticlesize,amphiphilicactivity,andfilm-formingtendency.Hiraietal.[6]havealsoproposedthatthejunctionsbetweentheparticlesandthediewallgrowinsizeduringthecourseofsliding,andthelubricantspreventthisgrowthresultinginadecreaseinthefrictionandadhesionforces.Theaforementionedattributesofalubricantareuniqueastheyarenotnormallyfoundwithotherexcipients,suchasfillers,binders,ordisintegrants,inatabletformulation.Forexample,thelubricationfunctionofmagnesium(Mg)stearate,awell-knownlubricantusedfortabletcompression,arisesfromthesecharac-teristics.AccordingtotheUnitedStatesPharmacopeia(USP32-NF27)andEuropeanPharmacopeia(6thEdition),Mgstearateisamixturecontainingmainlymagnesiumstearateandmagnesiumpalmitatewithstearatecontentnotlessthan40%andthesumofthestearateandpalmitatenotlessthan90%ofthetotalofallfattyacidester.Itsphysicalpropertiesarewidelyreportedintheliterature.Firstofall,Mgstearatehasaremarkablylowmax-imumshearstressof85kg/cm2,whichisdeterminedbyitscoef-ficientoffrictiononthediewallsurface[7,8].ThelowershearstressindicatesthatMgstearatehaslittleaffinityforthemetalsurface.Waterand/orgasmoleculesfromtheenvironmentmayalsoenterthelonglatticeoftheMgstearatecrystalstructure,spreadingwithinspaces,anddecreasingtheinteractiveforceofthecrystallattice.ThisreducestheshearingforcerequiredtocleavethecrystallineparticlesofMgstearate[9].Secondly,Mgstearatehasverysmallparticlesizeandlargesurfacearea.Awiderangeofphysicalpropertieshavebeenreported[10,11]forMgstearatewithmostofthevaluesfallingintherangesprovidedinparenthesis:meltingpoint94–150°C(125–127°C);specificsurfacearea1.3–10.5m2/g(4–6m2/g);particlesize2–15lmandmoisturecontent,byKarlFischer,4.8–5.2%.Thirdly,ithasbeenreported[5]that,duetoitsamphiphilicproperty,Mgstea-rateadherestometalsurfaceswithitspolarheadleavingthecar-bohydratetailgrouptostickouttoformboundaryfilmlubrication.ConvincingevidencethatMgstearateformsafilminordertobeaneffectivelubricantwasprovidedbyBolhuisetal.[12]in1975.MgstearatefilminasodiumchloridesolutionisshowninFig.2.Inthisclassicexperiment,onecrystallinepar-ticleofsodiumchloride,outofawell-mixedblendofsodiumchloridewith0.5%Mgstearate,wasplacedonawatersurfacesaturatedwithMgstearate.Oncesodiumchloridestartedtodis-solve,theMgstearatefilmunfoldedasafilmmaintainingtheshapeofthesaltcrystal.Ithasalsobeenshownbyaflow-throughdissolutiontechnique[13,14]thatMgstearateformsahydrophobicthinfilmonthesurfaceofthecarriermaterial.Fig.2.Unfoldedmagnesiumstearateenvelopesonwater,isolatedfromsodiumchloridecrystals,aftermixingwithmagnesiumstearate(approx.30Âand60Â).J.Wangetal./EuropeanJournalofPharmaceuticsandBiopharmaceutics75(2010)1–1532.TypesoflubricantsMostofthelubricantsusedforpharmaceuticalprocessesareboundarylubricants.Theyarechemicallyinert,odorless,andtaste-less.Somecommonlyusedlubricantsfortabletcompressionaredescribedinthefollowingparagraphs.2.1.MetallicsaltsoffattyacidsMgstearate,aluminumstearate,calciumstearate,sodiumstea-rate,andzincstearateareafewexamplesinthisgroupoflubricants.Theselubricantsarecommonlyusedintherangeof0.25–1.0%(w/w)fortabletcompression[15].Mgstearatecanbereplacedwithso-diumstearateorcalciumstearateifMgstearatecannotbeusedintheformulationduetochemicalstabilityconcerns[16].Mgstearateisverywidelyusedfortabletcompression[17],soitisnotsurprisingthatithasbecomethemoststudiedlubricant.Stearicacidfromfood-gradecrudeoilisthekeyrawmaterialtoproduceMgstearate.Theoilisinitiallyhydrogenatedandthenthermallysplitintofattyacidandglycerin.Thefattyacidandglyc-erinareseparatedbydistillation.Mgstearateisthenproducedbyusingsodiumhydroxidetosaponifythefattyacidtothesolubleso-diumsaltform,followedbytheadditionofMgsulfatesolutiontoprecipitatetheMgstearate.IntermsofthesourceofMgstearate,therehasbeenincreasingconcernsaboutthepotentialriskofthetransmissiblespongiformencephalopathies(TSE)includingthebovineformoccurringincows(BSEormad-cow-disease),thescrapiediseaseinsheep,aswellasthehumanform,thevariantCreutzfeldt–Jakobdisease.Inordertominimizetherisks,materialswithapotentialriskofBSE/TSE-infectivitymustbeeliminatedfromthehumanandani-malfood-chain.Therefore,thevegetable-derived,notanimal-de-rived,Mgstearatehasincreasinglygaineditspopularityinthepharmaceuticalindustry.Inaddition,Mgstearatehasbeenin-cludedinvariouscompendiasuchasUSP-NF,Eur.Ph.,BP,andJP.WhileUSP-NFhaspublishedGeneralChapterssuchasGoodMan-ufacturingPracticesforBulkPharmaceuticalExcipientsh1078iandBulkPharmaceuticalExcipients-CertificateofAnalysish1080i,theharmonizationoftheinternationalexcipientstandards(e.g.,theexcipientmasterfileguidebytheInternationalPharmaceuticalExcipientsCouncil)isstillanongoingeffort[18].TheparticlesizeandspecificsurfaceareaofMgstearatemaybethekeyfactorsinfluencingitslubricationefficiency.Inpractice,theparticlesize,insteadofthesurfacearea,ismuchmoreoftenusedasameasureofthelubricantcharacteristics.Itismainlyduetothefactthatparticlesizemeasurementofapowdermaterialisfas-ter,easier,andlesscontroversial.ThreebatchesofMgstearatedif-feringinmorphology,specificsurfacearea,bulkdensity,andparticlesizewerecomparedinthesameformulation[10].Althoughtheywereinthesameamount,theygaverisetotabletsdifferinginhardness,disintegration,anddissolution[10].ItwassuggestedthatthespecificsurfaceareaandnottheamountofMgstearateshouldbeusedtodetermineitsusagesincethediffer-encebetweenMgstearatebatchescouldbeexplainedtoagreatextentbydifferencesinsurfaceareas[19].Forexample,ifthecon-centrationofMgstearateincreasesfromzeroto0.2%intablets,thecoefficientoffrictionondiewalldecreasesbymorethan50%[20].Whenthelubricantconcentrationreachesacertainconcentration,itdidnotmakemuchdifferencewhetherthelubricantisingran-ularformorpowderformintermsofhowitperformeditslubrica-tionfunction[21].Sadeketal.[13]deducedtheminimumamountofglidant/lubri-cantinaformulationasfollows:Minimumamountð%Þ¼6ÂdÁqgDÁqÂ100ð4Þpwheredisthediameteroftheglidant/lubricantparticle,Disthediameterofthehostparticle,qgisthetrueglidantdensity,andqpisthebulkdensityofthepowder.Theirstudysuggestedthattheuseofanexcessiveamountwouldhinderflowabilitybecausetheforcesofinterparticulateadhesionwouldbegreaterthantheinterp-articulatefrictionbetweenthehostparticles.Thecrystallinestruc-tureofMgstearateisalsoimportant.MonohydrateanddihydrateformsofneatMgstearatehavebeenshownbySEM,XRPD,NIR,DSC,andTGAtechniques[22]tohavedistinctphysicochemicalprop-erties.Ingeneral,thecommerciallyavailableMgstearateconsistsofthedihydrateoramixtureofmonohydrate,dihydrate,andtrihy-drate,withthematerialfromsomevendorscontainingmoreamor-phouscomponents[23].TabletcompactionstudiesshowedthattheblendslubricatedwithMgstearatemonohydraterequiredhighercompressionforces,ejectionforce,andtake-offforcethanthosewiththedihydrateform.Anotherimportantpointtonoteisthatmaterialssuppliedbydifferentvendorsareunlikelytobeofexactlythesamephysicalpropertiesbutlot-to-lotvariabilityofmaterialsobtainedfromthesamevendorislesslikelytopresentaproblem.2.2.Fattyacids,hydrocarbons,andfattyalcoholsIngeneral,fattyacidsaremoreeffectivedielubricantsthanthecorrespondingalcohols,andthealcoholsarebetterthanthecorre-spondinghydrocarbons.Amongfattyacidswithcarbonatomcon-tentofC10toC24;alcoholswithcarbonatomcontentofC12toC22;andhydrocarbonswithcarbonatomcontentofC16toC22,thelubricationeffectivenessincreasesasthelengthofthemolecu-larcarbonchainincreasestoacertainpoint.Forexample,stearicacid(C18)offergreaterlubricationthansuchshortercarbonchaincompoundsasdecanoic(C10)anddodecanoic(C12)acidsorlongercarbonchaincousinssuchaseicosanoic(C20),docosanoic(C22),andtetracosanoicacids(C24)[24–27].Similarly,octadecanol(C18)providesamoreeffectivelubricationeffectthanalcoholswithlongerorshortercarbonchains.Thelubricationeffectofocta-decanol,however,decreaseswhentabletcompressionrunispro-longed.Hydrocarbonssuchashexadecane(C16)andoctadecane(C18)donotprovideanybetterlubricationinformulationswith-outlubricantsconfirmingtheirlackoflubricatingcapability.Octa-cosane(C28),ontheotherhand,maintainsitslubricationefficiencyatthebeginningandduringextendedperiodsoftabletcompression.Thelowermeltingpointofthematerialsinthiscat-egoryisreportedtobethereasonfortheirlesslubricationeffec-tivenesscomparedtothatofmetalstearates[28].Stearicacid(C18)isthemostcommonlyusedlubricantinthiscategory,typicallyatlevelsof2.5%(w/w).Physicalpropertiesreportedforstearicacidinclude[11]:󰀄specificsurfacearea1.1–1.9m2/g,󰀄geometricmeandiameter43.8–56.8lm,󰀄bulkdensity0.38–0.46g/mL,󰀄tappeddensity0.48–0.54g/mL,and󰀄meltingpoint52.2–54.7°C.Threepolymorphsofstearicacid,FormsA,B,andC,havebeenmadeusingdifferentorganicsolventsunderdifferentcrystalliza-tionconditions[29].FormCismoststable.FormBhasanirrevers-ibleendothermicphasetransitiontoFormCat54°C,whileFormAtransformstoFormCat°C.TheDSCandTGAthermogramsindi-catethatstearicacidfromdifferentvendorsshowverylittlebatch-to-batchormanufacturer-to-manufacturervariability[29,30].2.3.FattyacidestersAvarietyoffattyacidestershavebeenusedaslubricantsfortabletcompression.Someexamplesarelistedbelow[31–33]:4J.Wangetal./EuropeanJournalofPharmaceuticsandBiopharmaceutics75(2010)1–15󰀄sodiumstearylfumarate(PruvÒ),1–3%(w/w),󰀄sodiumlaurylsulfate,1.0%(w/w),󰀄magnesiumlaurylsulfate,󰀄glycerylbehenate(CompritolÒ888),1.5–3%(w/w),󰀄dodecanoictriglyceride,1%(w/w),󰀄glyceryl-palmitostearate(PrecirolÒATO),0.5%(w/w),󰀄sucrosemonopalmitate,󰀄sucrosemonolaurate,0.12%(w/w),󰀄samariumstearate.Sodiumstearylfumarate(PruvÒ)andglycerylbehenate(CompritolÒ888)arethemostcommonlyusedamongtheabove-mentionedlistoffattyacidesters.ComparedtoMgstearate,theselubricantsshowlessinterferencewithtabletstrengthandhavealessnegativeeffectontabletdissolution.About0.5%ofPruvÒ,incombinationwithPEG4000orPEG6000,actedasasuc-cessfullubricatingsystemforsomeeffervescenttablets[34].PruvÒwasalsoneededinlowerconcentrationthantalcfortheoptimallubricationinpentobarbitoltablets[35].Anamountequalto2%(w/w)ofCompritolÒ888showedthesamelubricationeffective-nessas0.75%(w/w)ofMgstearateinmicrocrystallinecellulosetablets[36].AlthoughCompritolÒ888wasusedinagreateramount,itdidnotnegativelyaffectthetablethardnessordisinte-grationtime.Therankorderforanti-adherent(stickingandpick-ing)effectofthethreestudiedlubricantsinsalicylicacidtablets[37]wasMgstearatePsodiumstearylfumarate>glycerylbehenate.Someoils,suchashydrogenatedvegetableoil(LubritabÒ),hydrogenatedcastoroil(CutinaÒHR),hydrogenatedcottonseedoil(SterotexÒK),arealsousedaslubricantsinamountsrangingfrom0.5%to2%(w/w).Forexample,theeffectivenessof2.5–3%(w/w)ofLubritabÒwasveryclosetothatof0.5%(w/w)Mgstearate[38].LackingthedisadvantageousmixingtimesensitivityofMgstearate,formulationslubricatedwithLubritabÒexhibitedmuchhighertabletstrengthandfasterdissolutioninprednisolonetab-lets[39].2.4.AlkylsulfateMagnesiumlaurylsulfateandsodiumlaurylsulfate(mainlyusedasasurfactant)arebothwater-solublelubricants.Magne-siumlaurylsulfateisbetterthansodiumlaurylsulfateandwasanequallyeffectivelubricantasMgstearateinlithiumcarbonatetablets.ItalsopossessesthelubricatingpropertiesofMgstearatebutwithoutitswaterproofingliability.However,theuseof0.5%(w/w)magnesiumlaurylsulfateinthedirectcompressionofatab-letcontaininganinsolublecompoundindicatedthatithasmoreretardingeffectthan0.5%(w/w)ofMgstearate[40].Thedisinte-grationtimewasalsomuchhigher,75sversus25sforMgstea-rate,atcertaintabletcompressionpressures.2.5.InorganicmaterialsIfMgstearateisnotappropriateforacompoundbecauseofchemicalinstability,itmaybereplacedwithtalcasaglidantandlubricant.Talcishydrousmagnesiumsilicate,sometimescontain-ingsmallamountofaluminumsilicate[41].Thephysicalproper-tiesoftalcfromseveralvendorsfallwithinthefollowingranges:󰀄specificsurfacearea3.5–10m2/g,󰀄geometricmeandiameter9.4–18.3lm,󰀄bulkdensity0.41–0.45g/mL,and󰀄tapdensity0.74–0.79g/mL.Mostinorganicmaterialsusedaslubricantscomeinvariousmixtureofsizescharacterizedaslaminatedflakes(2–5lm)andaggregatesofflakes(50–150lm)[11].Althoughthereisonlysmallbatch-to-batchvariabilityinthephysicalpropertyofthematerialfromthesamemanufacturer,thedifferencesobservedfromdiffer-entmanufacturerscanbelarge.Inexperimentswithacetamino-phentablets[42],1%(w/w)talcwasfoundnottobesignificantlydifferentfrom0.25%(w/w)Mgstearateintermsofgranulationflowabilityandejectionforce,andthetabletswerealsoharderandlessfragile.Ontheotherhand,talcwasalsofoundtobeaverysuitablelubricantataconcentrationof0.5–3%andupto5%(w/w)foraspirintablets[43,44].2.6.PolymersIfMgstearatecannotbeusedduetoproblemsofcompaction,lubrication,chemicalinstability,orforotherbiopharmaceuticalrea-sons,somepolymersmaybeusedasthetabletlubricantofchoice.PEG4000,PEG6000(CarbowaxÒ6000),polyoxyethylene–polyoxy-proprylenecopolymer(LutrolÒF68),andpolytetrafluoroethylene(FluonÒL169),forexample,havesuccessfullybeenusedinvarioustablets[43,45].FluonÒL169hasapproximatelythesamelubricatingpropertiesasMgstearateinacetylsalicylicacidtablet,butitdidnoteliminatetheelectrostaticchargesoftheformulationaswasobservedwithsmallpercentagesofMgstearate[46].3.Processingfactorsthatmayaffectlubrication3.1.InternallubricationLubricant(s)usedfortabletsareoftenincorporatedintothefor-mulationasthelaststeppriortotabletcompression.Itisusuallyblendedwiththemixtureconsistingofalloftheotheringredientsingranularorpowderformsinablender.Thistypeofprocedureisoftencalledinternallubrication.Theselectionoflubricant(s)affectthetypeofmixingequipmentandprocesstouseandthese,inturn,mayaffectthelubricationprocessandconsequentlythetabletproperties.Thisisexemplifiedbythestudy[47]inwhichthedependenceoftablettensilestrengthonlubricantmixingtime,pre-compression,andmaincompressionforcesweremeasuredwithmicrocrystallinecellulosecontaining0.5%(w/w)Mgstearate.Bymeasuringtheadhesionoftabletsonthelowerpunchsurfaceusinganinstrumentedrotarytabletpress,theadhesionofmicro-crystallinecellulosetabletsappearedtodecreasewithanincreaseinblendingtimeorintensityofblendingwithMgstearateatanygivencompressionforce[48].Thetabletejectionforcealsode-creasedwithlongerandmorevigorousblending.Tablethardness,however,decreasedwiththeincreaseinblendingtimeandinten-sityofblending.Inanotherstudy[49],powerconsumptionduringmixingofadirectcompressionblendwasmeasured.Fig.3showsthepowerconsumptionasafunctionofMgstearateconcentration.Thechangeinpowerconsumptionalsocorrelatedwellwiththechangesintabletejectionforce,tabletcrushingstrength,andtabletdissolution.Bolhuisetal.[50]alsofoundthattheeffectofMgstea-rate,whenmixedwithalactose/microcrystallinecellulosefor-mula,wasstronglydependentonthetype,size,androtationspeedofthemixer.Thedecreaseintabletcrushingstrengthwasmuchsteeperforthelargeindustrialtypemixersthanforsmalllabmixerswhentheyareoperatedatthesamerotationspeed.Inotherwords,atindustrialscale,tabletcrushingstrengthwasmoresensitivetothemixerrotationspeedthanthetype,size,andtheloadofthemixer.Theadverseeffectsofover-mixingofMgstearateontabletejec-tionforce,tablethardness,anddisintegrationtimehavebeenshownbyKikutaetal.[51].Theyalsofoundthatasemi-logarith-micplotofejectionforce,hardness,ordisintegrationtimeversusmixingtimeallyieldedtwo-phasestraightlineswiththeearlyJ.Wangetal./EuropeanJournalofPharmaceuticsandBiopharmaceutics75(2010)1–155Fig.3.Plotofpowerconsumptionversusmagnesiumstearateconcentration.phasehavingalargerfirst-orderrateandcontinuedtodecrease,inthesecondphase,atasmallerfirst-orderrate.AsshowninFig.4,theseresultssuggestedthatthemixingtimeinthesecondphaseshouldbeselectedtooptimizeMgstearateconcentrationsuchthatthechangeintablethardnessislesspronouncedonthetabletproperties.TheeffectofmixingtimeonlubricationanditsnegativeimpactontabletcrushingstrengthisnotlimitedtoMgstearate[52].Similareffectswereobservedonotherlubricants,suchashydrogenatedvegetableoil,glycerides,talc,andPTFE[52].Theeffectofmixingonlubricationhasalsobeenstudiedwithartificialneuralnetworkandthepolynomialregressionmethods[53].Usinghydrochlorothiazideasatestcompound,themodelssuccessfullyshowedthatthemixingtimewasthedominantfactorthatdecreasedtabletcrushingstrengthwhenmixingwithMgstearatelongerthan10min.Themodelsalsoshowedthatthecrushingstrengthoftabletscontainingglycerylbehenate(Compr-itolÒ888)wasnotsignificantlyaffectedbythemixingtime.3.2.ExternallubricationAsopposedtomixinglubricantwithformulationingredients,externallubricationoftenreferstoalubricationprocessinwhichonlythelowerpunchanddie,notthefinalblendmaterial,arelubricated[54].Thistypeoflubricationproceduremaybeusedwhentabletpropertiesareverysensitivetolubricants.AsshownFig.4.Semi-logarithmicrelationshipofthetablethardnessandmagnesiumstearateconcentration.Key:opencircle–0.1%,half-opencircle–0.3%,solidcircle–0.5%inFig.5,asuspensionof5%(w/w)Mgstearateinliquidpetroleumwastransferredthroughatubetothefoamrubberringssurround-ingthelowerpunch.Aneffervescenttabletcontainingnicotinicacidandsodiumbicarbonatewassuccessfullycompressedusingthislubricationmethod.Asimilartechniquewasusedtomakeanorallydisintegratingtablet[55].Usingsimilarautomatedtech-niqueduringrotarytabletcompression,Yamamuraetal.investi-gatedtheeffectsofexternallubricationontabletpropertiesoftheeprazinonehydrochloridetablets[56].Theamountoflubricantrequiredtopreventstickingwithexternallubricationwasonly0.08%ofthatwithinternallubrication.Inthemeanwhile,externallubricationgave40%highertabletcrushingstrengthwithoutpro-longingtabletdisintegrationtime.Trypsintabletshavebeenmadewithbothinternalandexternallubrication[57].Comparedtotabletsmadewithinternallubrica-tion,tabletsmadewithexternallubricationneededlowercom-pressionenergybuthigherejectionenergy.Thetabletsalsoshowedhigherhardness,lesstotalporevolume,fasterdissolutionandhighertrypsinactivity.Whentablettensilestrengthordissolutionaresusceptibletothelubrication,itisperhapsmoreimportanttoconsidertheoptionofexternallubrication.Astheadverseeffectsoflubricationontab-letpropertiesareoftenexacerbatedwhenthemixingoperationisconductedatlargerscaleforinternallubrication,theusageofexternallubricationshouldhelpavoidsuchproblems.Althoughnotcost-effective,itispossibletotreatthetoolingsurfacesofatabletpresswithhighlypolishedchromiumcoatinginordertoreducetheircoefficientoffriction.Additionally,thedieissometimestreatedwithMgstearatepowderoritssolutioninorganicsolventtoprovidelubricationintabletcompactionresearch.4.CharacterizationofthelubricationprocessTrial-and-errormethodshavebeenusedforaverylongtimetodeterminewhetheralubricantshouldbeused,andifso,whatlu-bricanttouseandhowmuchwouldbetheoptimalamount.Withrecentadvancementsinanalyticaltechniques,tabletpressinstru-mentation,andotherprocessanalyticaltechnologies,manyresearchershavemadetheefforttounderstandthefundamentalworkingsoflubricantsbycharacterizingthelubricationprocess.4.1.CoefficientoffrictionduringtabletejectionDuringtabletcompression,becausearesidualdiewallforceisgenerated,anejectionforceisneededtoovercomethisforceinordertopushthetabletoutofthedie.Intuitively,thecoefficientFig.5.Diagramofthearrangementforthelubricationofthelowerpunchanddie(externallubrication).6J.Wangetal./EuropeanJournalofPharmaceuticsandBiopharmaceutics75(2010)1–15offrictionshouldbeagoodindicatoroftheeffectivenessoflubri-cation.HölzerandSjogren[17]investigatedthecoefficientoffric-tionamongafewotherparametersduringtabletcompression.Bycomparingseverallubricants,theyfoundthattherewasnocorre-lationbetweenthecoefficientoffrictionofthelubricantitselfandthelubricatingeffectinsodiumchloridetabletcompression.Ontheotherhand,thenegativeeffectsontabletstrengthanddisinte-grationtimecorrelatedquitewellwiththelubricatingability.Althoughthestudywasquitecomprehensive,theremaybeaseri-ousflawinattemptingtoobtainthecoefficientoffrictionbysimul-taneouslymeasuringboththeaxialandradialforcesatdifferentcompressionforcesduringtabletcompression.Toovercometheshortfallsofthismethod,whichoperatesontheassumptionthatvaryingcompressionforcesmaycausechangesofthestructureofthetablets,KikutaandKitamori[58]developedamethodtoestimatethecoefficientoffrictionandtheadhesiveforceatconstanttabletcompressionforce.Therela-tionshipbetweenthetabletejectionforce,radialforceatgivencompressionforceisillustratedbyCoulomb’sequation:Fe¼lFrþCð5ÞwhereFeistheejectionforce,Fristheradialforce,listhecoeffi-cientoffrictionbetweenthetabletsideandthediewall,andCisanindicatorofadhesiveforce.Usinglactoseasamodelmaterial,theauthorsprovedthatthelinearrelationshipheldfortheejectionforceandradialforce(Fig.6).Theejectionforcesweredeterminedundervariousradialforcesproducedbyregularcompressionandthesubsequentre-compression(withoutsurpassingtheregularcompressionforce),orrelaxationofthetabletinthedie.Fig.7showsthat,usingthismethod,thecoefficientsoffrictionofMgstearateandcalciumstearategavethelowestvalue(0.04)andtalcthehighestvalue(0.25),indicatingbetterlubricationpropertiesofMgstearateandcalciumstearatecomparedtotalc[59].Theaffinityoflubricantstothediewallwasstudiedbyconditioningthediewallwitheachlubricantfollowedbycompressinglactosetablets.AsshowninFig.8,thelubricationeffectoftalcandcornstarchwaslostafterseveraltablets,whileMgstearatemaintaineditslubricationeffectevenaftertenormoretablets.Inauthors’opinion,eventhoughthismethodistime-consumingandrequirescustom-builtequipment,itisthemostconvincingmethodforcharacteriz-inglubricantsandtheirlubricationeffectsintabletcompressionatmacroscopiclevel.Inanotherwork[60],thecoefficientsoffric-tionofsomelubricantsonasteelsurfaceweremeasuredwithamodifiedannularshearcelltorankordertheirlubricationeffective-Fig.6.Relationshipbetweenejectionandradialforcesaftersixthejection.Fig.7.Relationshipbetweenejectionandradialforcesforthecompactsofsomelubricants.Fig.8.Changeinejectionforceinserialtabletingofun-lubricatedlactosegranulatesaftertabletingofeachlubricant.ness.Inallcomparisons,Mgstearatehadthelowestcoefficientoffriction.Divalentsaltsofstearicacid,suchascalciumstearate,appearedtobebetterthantheotherlubricants,suchasaluminummonostearate,glycerylbehenate,andPEG8000.4.2.OtherparametersorindicesduringtabletcompressionCompactionparameters,suchastheratiooftheresidualdiewallpressuretomaximumdiewallpressure,orRDP/MDP,hasbeenusedtoevaluatetheeffectivenessoflubrication[61].ShowninFig.9arethedifferencesinRDP/MDPvaluesforlactoseandmannitolwithoutMgstearateversusthelubricantisaddedinter-nallyorexternally.Potassiumchloride(KCl),awell-knownplasti-callydeformablematerial,wasstudiedtofurtherdemonstratetheusefulnessoftheRDP/MDP.AsshowninFig.10,RDP/MDPvaluesdecreasedwithincreasingMgstearateconcentration,suggestingreducedcompactability.Tablettensilestrength,however,de-creasedasthelubricantamountexceeded0.5%.TheresidualdieJ.Wangetal./EuropeanJournalofPharmaceuticsandBiopharmaceutics75(2010)1–157Fig.9.TheeffectofexternallubricationontheRDP/MDPoflactoseormannitol.Compressionpressureis100MPa.Fig.10.TheeffectsoflubricantcontentandlubricationmethodontheRDP/MDPandtensilestrengthofpotassiumchloridetablet.wallpressureofmicrocrystallinecellulose–lactosetabletsin-creasedasthedegreeofcurvatureofthetabletfaceincreased,intheorderofflat