ultra-performance and ﬂuidchromatography supercritical Generic development chiral method in supercritical ﬂuidchromatography Journal of Chromatography A

(1)

ContentslistsavailableatScienceDirect

Journal of Chromatography A

jou rn al h om ep a g e : w w w . e l s e v i e r . c o m / l o c a t e / c h r o m a

Generic chiral method development in supercritical ﬂuid chromatography and ultra-performance supercritical ﬂuid chromatography

Katrijn De Klerck, Yvan Vander Heyden, Debby Mangelings

^∗

DepartmentofAnalyticalChemistryandPharmaceuticalTechnology,CenterforPharmaceuticalResearch(CePhaR),VrijeUniversiteitBrussel–VUB, Laarbeeklaan103,B-1090Brussels,Belgium

a r t i c l e i n f o

Articlehistory:

Received12April2014

Receivedinrevisedform2June2014 Accepted3June2014

Availableonline9June2014

Keywords:

SFCmethoddevelopment Chiralseparationstrategy

Polysaccharide-basedstationaryphases Ultra-performanceSFC

Methodtransfer

a b s t r a c t

Thedevelopmentofchiralseparationmethodsinpharmaceuticalindustryisoftenaverytedious,labour intensiveandexpensiveprocess.Atrial-and-errorapproachremainsfrequentlyused,giventheunpre- dictablenatureofenantioselectivity.Tospeed-upthisprocessandtomaximizetheefficiencyofmethod development,agenericchiralseparationstrategyforSFCisproposedinthisstudy.Todefinesuchstrat- egy,theeffectofdifferentchromatographicparametersontheenantioselectivityisinvestigatedand evaluated.Subsequently,optimizationstepsaredefinedtoimproveachiralseparationintermsofres- olution,analysistime,etc.ortoinduceseparationwheninitiallynotobtained.Thedefinedstrategy proveditsapplicabilityandefficiencywiththesuccessfulseparationofanovel20-compoundtestset.In asecondstage,themethodtransferfromaconventionaltoanultra-performanceSFCsystemisinvesti- gatedforthescreeningstepoftheseparationstrategy.Themethodtransferprovedtobeveryeasyand straightforward.Similarenantioresolutionvalues,butslightlyshorteranalysistimeswereobtainedonthe ultra-performanceequipment.Nevertheless,evenmorebenefitmaybeexpectedinultra-performance SFCwhencustomizedsub-2␮mchiralstationaryphaseswillbecomeavailable.

1. Introduction

Over the past years, much attention has been paid to sub- and supercritical fluid chromatography (SFC)in the context of chiral separations [1–4]. By exploitingthe benefitsof sub- and supercritical fluids,fastand efficient enantioseparations canbe obtainedinSFC. Simplyreturning toambientconditionsevapo- ratestheprimaryeluent,carbondioxide(CO₂), fromthemobile phaseafteranalysis.Hence,SFCcandelivera significantreduc- tioninwastegenerationand–disposalcomparedtoconventional high-performanceliquidchromatography(HPLC)[5].Thehigher flowrates,thatcanbeappliedinSFC,allowhigherproductivities relativetoHPLC,whichisanimportantassetinapharmaceutical industrialenvironmenttoacceleratethedrugdevelopmentpro- cess[6,7].Giventheseproperties,SFChasbecomeapredominant techniquefor(preparative)enantioresolutions[2,3,6–8].

∗ Correspondingauthor.Tel.:+3224774329;fax:+3224774735.

E-mailaddresses:debby.mangelings@vub.ac.be,Yvan.Vander.Heyden@vub.ac.be (D.Mangelings).

Asforall separationtechniques,chiral methoddevelopment is alsoinSFC quitelabourintensive. Enantioselectivityremains unpredictable and the best way to achieve appropriate separation conditions is by experimental trial-and-error. To make methoddevelopmentmore efﬁcientand faster, generic separa- tionstrategies can beutilized [9–12].Thesestrategies screen a chiral compound ona limitednumber ofcomplementary chro- matographicsystems(stationary+mobilephasecombinations)in ordertoﬁndthemostsuitablesystem,showingthebestenantiose- lectivity.Dependingontheoutcomeofthisscreening,optimization stepsguidetheuserfurthertoobtainthedesiredseparation.In casethedesiredseparationcouldnotbeachieved,oneisreferred toscreeninanotherseparationtechnique.

AgenericscreeningapproachinSFC,thatallowsafastselection of anappropriatechromatographic chiral separationsystemfor diversechiralmixtures,wasproposedearlier[13].Polysaccharide- basedchiralstationaryphases(CSPs)wereusedinthisscreening becauseof theirbroad enantiorecognitioncapabilities and easy availabilities[3,14].However,afterexecutingthisscreening,one mightnothaveachievedthedesiredseparationyet.Inthatcon- text,furthermethodoptimizationstepscanbedeﬁned.Theseaim tooptimizeresolution,selectivity,analysistime,andinrelevant http://dx.doi.org/10.1016/j.chroma.2014.06.011

(2)

casesalsothepeakshape.Aﬁrstpartofthispaperfocusesonthe inﬂuenceofdifferentparametersonachiralSFCseparation.Based onthisinformation,appropriateoptimizationstepsarederivedto completetheentiregenericseparationstrategy.Toevaluatethe performanceofthisstrategy,anovel20-racematessetistested.

Tocatchupwiththestate-of-the-arttechnologyfoundinthe ﬁeldofHPLC,SFC equipmentis becomingbetteradapted,more robustandmorereliabletoachievechromatographicseparations withacceptablerepeatabilityandreproducibility.Inparticular,the mobilephase densitycanbe controlledmuchstricter, which is acrucial aspectinSFC sincethedensityhasadirect impacton themobile-phase strength.Followingthetrend in HPLC,SFC is undergoinganevolutiontoultra-highperformanceSFC(UHP-SFC) [15,16].Withminimalvoidvolumesandmaximalsensitivity,fast separationscanbeachievedwithhighefﬁciencies.Becausecertain parameters aredifferent betweenthedifferent systems,(enantio)separationsmightbeimpactedwhentransferred.Asecondpart ofthisresearchthereforefocusesonthemethodtransferfromcon- ventionalSFCtoUHP-SFC.

2. Experimental

2.1. Chromatographicequipment

TheanalyticalSFCmethodstationfromThar^®(Pittsburgh,PA, USA,a Waters^® company)equippedwithaWaters^® 2998-DAD detector(Milford,MA,USA)wasusedfortheﬁrstpartoftheexperi- ments(deﬁnitionoftheseparationstrategy).Theautosamplerwas equippedwitha10␮lloop.Forallanalysespartialloopinjections of5␮lweredone.Dataacquisitionandprocessingwereperformed usingChromscope^®V1.10software(2011)fromWaters^®.

ForthestrategyevaluationandmethodtransfertoUHP-SFC,an AcquityUltraPerformance ConvergenceChromatography(UPC²) fromWaters^®wasused.Thesystemwasequippedwithabinary solventmanager,asamplemanagerwithaﬁxedloopof10␮l,a convergencemanager,anexternalAcquitycolumnovenandaPDA detector.Forallanalysespartialloopinjectionsof5␮lweredone.

Empower^® 3V7.10software(2010,Waters^®,Milford,MA,USA) wasusedfordataacquisitionandprocessing.

Thechromatographicconditionsweredifferentfortheanalyses performedduringtheoptimizationprocess.Forthisreasonthey arespeciﬁedfurther.

2.2. Materials

ThecolumnsChiralpak^®AD-HandChiralcel^®OD-H,OJ-Hand OZ-HwerepurchasedfromChiralTechnologies(WestChester,PA, USA).Lux^® Cellulose-1,-2, and-4werepurchasedfromPheno- menex(Utrecht,TheNetherlands).Toallowafaircomparison,all columnshaddimensionsof250mm×4.6mmi.d.with5␮mpar- ticlesize.

2.3. Chemicals

Methanol (MeOH), ethanol (EtOH) and 2-propanol (2PrOH) wereHPLCgrade andpurchasedfromFisherChemicals(Lough- borough,UK).Isopropylamine(IPA)andtriﬂuoroaceticacid(TFA) werefromAldrich(Steinheim,Germany).CO₂wasusedasadvised bythemanufacturersoftheindividualSFC instruments.Forthe Thar^®equipmentthiswasquality2.7(purity≥99.7%)fromLinde Gas(Grimbergen,Belgium);fortheUPC^2®equipmentquality4.5 (purity≥99.995%)fromMesser(Sint-Pieters-Leeuw,Belgium).

Allpercentagesexpressedinthecontextofmobile-phasecom- positionarevolumepercentages.

2.4. Chiraltestset

Forthedeﬁnitionoftheoptimizationstepsandseparationstrat- egy,agenericchiraltestsetof56pharmaceuticalswasused.Test solutionsofthese56racemateswithaconcentrationof0.5mg/ml weremadeinmethanol.Thesolutionswerekeptat4^◦Cwhennot used.Thetestsetwascomposedofracemateswithdiversestruc- tural,chemical, and pharmacologicalproperties.Because it was usedinearlierresearch,werefertothesepapersfordetailedinfor- mation[17,18].Toevaluatetheproposedseparation strategy, a noveltestsetcomposedof20pharmaceuticalracematesisused (Table1).TheseracemateswerealsodissolvedinMeOHatacon- centrationof0.5mg/mlandkeptat4^◦C.

2.5. Dataprocessing

Forallenantioseparations,theresolution(Rs)iscalculatedusing theEuropeanPharmacopoeiaequationsapplyingpeakwidthsat halfheights[19].Separationsobtainedwitha resolutionhigher than1.5areconsideredasbaselineseparated.Whentheresolu- tionisbetween0and1.5theseparationsaredesignatedaspartial.

Theselectivity(˛)iscalculatedastheratiooftheretentionfactors ofthelastandﬁrstelutingenantiomersofapair[19].Thevoidtime wasmarkedastheﬁrstdisturbanceofthebaselineafterinjection ofsolvent.Theretentiontimeofthelastelutingpeakistakenas theanalysistime.

Microsoft^®Excel(Microsoft^®Corporation,2010)wasusedfor constructingtheplotsandgraphsandforthestatisticalinterpreta- tionofthedata(Studentt-testandANOVA).

3. Resultsanddiscussion 3.1. Screeningstep

A generic chiral screening approach was derived from the evaluationof12polysaccharide-basedchiralstationaryphasesin combination with eight mobile phases (MP) (total of 96 chromatographic systems).The performance in terms of successful enantioseparations,andthecomplementarityofthelattersystems weretakenintoaccount,todefineascreeningsequence(Fig.1) [13].Thescreeningentailsfourexperiments,evaluatingfourcom- plementarypolysaccharide-basedstationaryphases.Thisapproach allowedtheseparationofallcompoundsfromthe56-compound testset.However,noteveryseparationisoptimal,e.g.Rs<1.5(partialseparations)orexcessiveanalysistimecanbeobtained.Inthese casesfurtheroptimizationimposes itselfin order toobtainthe desiredenantioseparation.Becauseanumberoffactorsinfluence enantioseparationinSFC,e.g.organicmodifier,flowrate,pressure, temperature,etc.,theoptimizationisnotalwaysevident.Inafirst partofthiswork,attentionwillbepaidtothesefactorsimpact- ingenantioseparation.Theobtainedinformation willbeusedto definespecificoptimizationstepsinthecontextofagenericchiral separationstrategy.

3.2. FactorsinﬂuencingenantioseparationsinSFC

3.2.1. Organicmodiﬁertype

In most cases, pure CO₂ is not adequate to elute (pharmaceutical) compounds. Most pharmaceutical compounds possess a structure with hydrophobic, hydrogen-bonding donor and - acceptorsites.Thisrequirestheadditionofanorganicmodiﬁerto themobilephasetoincreasethesolventstrength,allowingelution andanalysisoftheserelativelypolarcompounds[4,5].

Itiswell-knownthattheorganicmodiﬁertypeinthemobile phasealterstheenantioselectivityofaCSPtowardscertainrace- mates.Thelipophilicity,polarity,basicity,i.e.thepropertiesofthe

(3)

Table1

Test-setcompoundsusedtoevaluatetheseparationstrategy.

Racemate Structure Origin

Carprofen Sigma–Aldrich,Steinheim,Germany

Carteolol MadausAG,Köln,Germany

Celiprolol Originunknown

Ceterizine Sigma–Aldrich,Steinheim,Germany

Clopidogrel Originunknown

Cyclopentolate GiftfromPhenomenex

Econazol Janssenresearchfoundation,Beerse,Belgium

Felodipine Hassle(Astra),Sweden

Fluoxetine Sigma–Aldrich,Steinheim,Germany

Indapamide Sigma–Aldrich,Steinheim,Germany

Indoprofen Sigma–Aldrich,Steinheim,Germany

Isradipine Originunknown

(4)

Table1(Continued)

Racemate Structure Origin

Lorazepam Wyeth,NY,USA

Miconazol Janssenresearchfoundation,Beerse,Belgium

d/l-Nebivolol Janssenresearchfoundation,Beerse,Belgium

Ondansetron GlaxoWellcome,Belgium

Temazepam Originunknown

Terazosine Sigma–Aldrich,Steinheim,Germany

Thioridazine Originunknown

trans-Stilbeneoxide Originunknown

organicmodiﬁeraffecttheinteractionsbetweenthesoluteandsta- tionaryphase[20].Consequently,bychangingtheorganicsolvent inthemobilephase,differentenantioseparationscanbeachieved onthesameCSP.InchiralSFC,methanol,2-propanolandethanol aremostoftenusedasmodiﬁers[6,12,7,21–23].Inourexperience,

MeOHisslightlymoresuccessfulonthepolysaccharide-basedCSPs, followedby2PrOHandEtOH.MeOHofferstheadditionaladvan- tagethatitsboilingpointislowerthanthatof2PrOHandEtOH, makingsolventevaporationafteranalysiseasier.Theviscosityof MeOHisalsoloweranditsuseposesthuslessstressontheCSPs.

(5)

Fig.1. Schemeofthescreeningstepasdeﬁnedin[13].Inthetoprowthechiralstationaryphasesarepresented,whilethesecondrowrepresentstheusedmodiﬁer concentrationinthecarbon-dioxidebasedmobilephase.

Inearlierresearch,12 polysaccharide-basedchiralstationary phases wereevaluated witheight MeOH- or 2PrOH-containing mobile phases [13]. On eight of these twelve CSPs, a MeOH- containingmobilephaseprovidedthehighestsuccessrate.Forthis reasonweslightlyfavourMeOHover2PrOH.

Asfarasenantioselectivityisconcerned,itisimpossibletopre- dictwhich solventwill providethemostfavourable separation conditionsforagivenracemate.Earlierweselectedfoursuccess- fulandcomplementarychromatographicsystems,usingageneric compound test set. We included these systems in a screening approach [13]. For most compounds, executing this screening shoulddeliverappropriateselectivitytoachievethedesiredenan- tioseparation.Wewereabletoseparate(baselineorpartially)the entire56-compoundtestsetusingMeOHincombinationwithOZ- H(orLC-2)andOD-H(orLC-1);andusing2PrOHwithAD-Hand LC-4.

However,in caseno(satisfying) separation isobtainedafter thisscreening, itis advisabletoscreen thesameCSPswiththe alternativemodifier,i.e.2PrOH(forOZ-H/LC-2andOD-H/LC-1)or MeOH(forAD-HandLC-4),sincethisbroadenstheenantioselective range.Differentenantioselectivities,wereobservedwhenconsid- eringbothmodifiers.InmostcasesMeOHyieldsmoreseparations (Table2).AD-Hseemsanexceptiontothistrend,since2PrOHis muchmoresuccessfulthanMeOHonthisCSP.Nevertheless,on eachCSP,anumberofuniqueseparationsisprovidedbybothmod- ifiers.Thisexplainsthesecondstepinourscreeningstrategy,which proposestoscreentheselectedCSPswithanalternativemodifier.

Wealsonoticedauniqueenantioselectivityofsomestationary phasesincombinationwithEtOH.Incasenoenantioselectivityis obtainedafterscreeningwithMeOHor2PrOH,EtOHcantherefore betestedasalternativemodifier.However,giventhelowergen- eralsuccessrateofEtOH,itwouldbelessadvisabletoincludethis modifierinafirstscreeningattempt.

3.2.2. Concentrationoftheorganicmodiﬁer

Inlowconcentrations(<2–5%),theorganicmodiﬁercompetes withtheanalytesforinteractionwithresidualsilanolgroupson thestationaryphase.Bysurroundingtheactivesilanolsites,the stationaryphasebecomesmoreuniformintermsofpolarityand consequentlypeak shapesbecomemoresymmetrical.Hence,in

Table2

Numberofseparationsobtainedwiththe56-compoundtestsetusing20%methanol (MeOH)or2-propanol(2PrOH)inthemobilephase(with0.1%isopropylamine and0.1%trifluoroaceticacidaddedtothemodifier).Theseparationsthatareonly obtainedwithonemodifieronagivenstationaryphaseareconsideredasunique separations.

Baselineseparations Partialseparations Uniqueseparations

MeOH 2PrOH MeOH 2PrOH MeOH 2PrOH

OZ-H 27 22 18 10 15 2

AD-H 12 25 11 12 3 17

OD-H 27 25 14 10 7 2

LC-4 28 24 14 12 9 4

thelower concentrationrange, an increase in modifier content is advantageous for theresolution of the separations. Once all silanolsitesarecoveredbymodifiermolecules,afurtherincrease inmodifierconcentrationnegativelyinfluencestheresolutionby impacting the solvent strength of the mobile phase [24]. The separationefficiencyalsotendstodeteriorate,since theanalyte diffusionthroughthecolumnisinhibitedbytheincreasingmobile phaseviscosity[25].

Thesetrendsareclearlyseenintheseparationofclopidogrelon Chiralpak^®AD-H(Fig.2).Whenthemodifiercontentisincreased from5to10%theresolutionincreasesfrom3.8to4.6.Increasingthe modifiercontentabove10%,decreasestheresolution.Ontheother hand,theanalysistimeisimpactedbythemobilephasestrength.A decreasefrom7.98to2.75minoccurswhenthemodifierincreases from5to20%.Furtherincreasingthemodifierinthemobilephase to40%decreasestheanalysistimeto1.72min.However,therela- tionbetweentheanalysistimeandmodifiercontentisnotlinear andtheobserveddecreaseinanalysistimeishigherinthelower concentrationrange(5–20%).

Conclusivelyitcanbestatedthatthemostappropriatemodiﬁer contentinthemobileshouldbeacompromisebetweenanalysis timeand resolution.Inourstrategywe proposetoincrease the modiﬁercontentwhenshorteranalysistimesaredesired.Ifhigher

Fig. 2. Separation results of clopidogrel on Chiralpak^® AD-H with (2PrOH+0.1%TFA+0.1%IPA) in the mobile phase in varying concentrations.

(a)Obtainedresolutionsand(b)thetotalanalysistimeinfunctionofthepercentage modiﬁercontent.

(6)

Fig.3. ResultsoftheenantioseparationofcetirizineonChiralpakAD-Hwith20%(2PrOH+0.1%IPA+0.1%TFA)inthemobilephaseasafunctionoftheﬂowrate.(a)Overlay oftheobtainedchromatograms;(b)analysistime;(c)resolutionandselectivityasafunctionoftheﬂowrate.

resolutionsaredesiredweadvisetheopposite.Asacompromise 20%modiﬁerisusedinthescreening.

3.2.3. Flowrate

Supercriticalﬂuidchromatographyissuitableforfastanalyses.

Becausethesub-orsupercriticalmobilephasehasalowviscosity andhighdiffusivity,higherflow ratescanbeusedcomparedto HPLC.Flowratesupto5.0mlperminarenoexceptioninanalytical SFC.Increasingtheflowratewillfastenananalysissignificantly, withoutcompromisingtheseparationefficiencytoodrastically.

Forexample,whentheﬂowratefortheenantioseparationof cetirizineisincreasedfrom1to6ml/min,theanalysistimereduces with84%(from16.8to2.6min), theRsdecreaseslessthan50%

(from12.42to6.41),while˛remainsalmostunchanged(Fig.3).

Theseparationat6ml/minisstilllargelyacceptable,andrequires 14minlessthanthatatﬂowrate1ml/min.Increasingtheﬂowrate above6ml/minwasnotpossibleduetopressurelimitationsofthe CSP.

Aboveanexampleisshownwhichactuallyisvalidforallchi- ralSFC-separations.Thisisexplainedbytheflatterprofileofthe VanDeemtercurveinSFCcomparedtoHPLC,allowinganalysesat highermobilephasevelocitieswithoutasubstantiallossineffi- ciency[5]. Hence, when optimizing analysistimes in SFC, it is advisabletoincreasetheflowrate,sincetheimpactonthereso- lutionremainsratherlimited.Thelimitingfactorsinthisapproach arethepressurerestrictionsimposedbytheequipmentandthe chromatographiccolumn.

3.2.4. Backpressure

Toguaranteeaconstantmobile-phasedensity,aback-pressure regulatorisemployedinSFCcontrollingthepressure.Themobile- phasedensityhasadirectimpactonthemobile-phasestrength, thus on the (enantio)selectivity and retention. A higher back

pressuremeansahighermobile-phasedensity,and-strength,and shorterretentiontimes.Asaconsequence,theselectivitymight alsodecrease.

However,whenexploringa pressurerangein thesearchfor optimalseparationconditions,auserisrestrictedbythelimita- tionsofthepolysaccharide-basedcolumnandtheequipment.In practice,backpressuresbetween125and250bararecommonly usedforchiralSFCseparations.Usinglowerpressuresharmsthe chromatographicresultssigniﬁcantlysincethesub-criticalstateof themobilephaseisnolongerguaranteed[26].

Inthispressurerange(125–250bar),theactualimpactofthe backpressure ontheretention andselectivityis rather limited andconsiderablylowerthanthatoftheorganicmodiﬁercontent.

In otherwords,whena largechangein retentionor selectivity isdesired,thefirststepshouldbetoadoptthemodifiercontent intheMP.Whenfine-tuningaseparation,thebackpressurecan bechanged.For shorter retention/analysistimes thebackpres- sureshouldbeincreased,whiledecreasingisadvisablewhenthe selectivityshouldbeimproved.

Fortheseparationofeconazole,adoublingofthebackpressure from125to250bardecreasestheretentionofthelastelutingpeak from8.5to6.4min(Fig.4).Asaconsequence,thepartialresolution islostwhenthebackpressureiselevatedabove200bar.

Forscreeningpurposes,itisproposedtosetthebackpressure at150barasacompromisebetweenretentiontimeandenantio- selectivity.Consequently,reducingthebackpressuretothelower limitof125barwouldonlyresultinaminimalgaininenantioselec- tivity.Thereforethisstepisnotincludedinthepartialseparation branchofthestrategy(seefurther).Ontheotherhand,tospeed uptheanalysis,itismoreeffectivetoincreasetheﬂowrateand/or modiﬁercontentthanthebackpressure.Therefore,anincreasein backpressureisonlyrecommendedasathirdchoicetoreducethe analysistimeofbaselineseparations(seefurther).

(7)

Fig.4.OverlayofthechromatogramsofeconazoleonChiralcel^®OZ-Hwith20%(MeOH:IPA:TFA,100:0.1:0.1,v/v/v)inthemobilephase.Aﬂowrateof3ml/minand temperatureof30^◦Cwasused.Thebackpressureswere(1)125bar;(2)150bar;(3)175bar;(4)200bar;(5)225bar;and(6)250bar.(ResultsgeneratedwiththeUPC² system.)

3.2.5. Temperature

Temperature also inﬂuences the mobile-phase density. An increase resultsin a decreaseof the mobile-phase density and hastheabove-mentionedconsequences.Itisimportanttorealize thatbyreducingthetemperature,thechromatographicconditions deviatefurtherfromthesuper-intothesubcritical region.This doesnot createpractical issues untilthe subcritical stateturns intoa two-phasestate,whichwould deterioratethechromato- graphic results signiﬁcantly and prevents proper analyses.The vapour–liquidcurveofthepressure–temperaturephasediagram separatesthetwo-phase regionfromthesubcritical region.For (chiral)SFCseparationsitisthusimportanttoremainabovethat vapour–liquidcurve,buttherearenofurtherrestrictionstothe chosen conditions.SFC separations can thus alsobeperformed below31^◦C,i.e.thecriticaltemperatureofpurecarbondioxide[26].

For polysaccharide-basedcolumns,the temperaturerange is limitedfrom5to40–50^◦C,varyingbycolumn-manufacturerinfo.

Theactualimpactofthetemperatureontheretentionandselectiv- ityinthisworkablerangeisratherlimited.Whenthetemperature isincreasedfrom10to45^◦C(a350%increase),theretentionofthe lastelutingpeakofcarprofenonlydecreasesfrom2.70to2.56min (adecreaseof 5%)(Fig.5).Theresolution andselectivityofthe separationarehardlyaffectedbythistemperaturechange.

Summarized,it canbestatedthat althoughthetemperature hasanimportantimpactonSFCseparations,theworkabletem- peraturerangewithpolysaccharide-CSPsistoolimitedtohavea signiﬁcantgain inanalysistime orselectivity.Forthis reason,a temperatureoptimizationisnotincludedintheﬁnalseparation strategy(seefurther).Thetemperaturewasthereforesetat30^◦C forallexperiments,basedonthestudyofMaftouhetal.[12].

3.2.6. Additives

Inthescreening,0.1% isopropylamine(IPA) and0.1%triﬂuo- roaceticacid(TFA)areaddedtothemodiﬁer,ofwhichonly20%

inusedinthemobilephase.Hence,theﬁnalconcentrationinthe MPis0.02%IPAandTFA.Nevertheless,theiraddition,eveninthese lowconcentrations,affectstheinteractionsbetweentheanalytes andthestationaryphase.Withoutthepresenceofadditivesinthe MP,chromatographicresultstendtodeterioratesigniﬁcantly.IPA

and otherbasic amine-additivesshield silanol sitesonthe sta- tionaryphase, decreasingthenon-speciﬁcretentionofanalytes.

Theyalsocompetewiththebasicfunctional groupsof analytes forinteractionswithspeciﬁcsitesonthestationaryphase.These additivesalsoneutralizechargedgroupsofbasicanalytes,whichis

Fig.5. ChromatogramsoftheenantioseparationofcarprofenonChiralcel^®OZ-H with20%(2PrOH:IPA:TFA,100:0.1:0.1,v/v/v)inthemobilephase.Atotalﬂowrate of4ml/minandbackpressureof150barwasused.Thetemperatureswere(a)10^◦C;

(b)15^◦C;(c)20^◦C;(d)25^◦C;(e)30^◦C;(f)35^◦C;(g)40^◦C;and(h)45^◦C.

(8)

Fig.6.Chiralseparationstrategyforpolysaccharide-basedcolumnsinSFC.

importantfortheinteractionswithneutralchiralselectors,suchas polysaccharide-derivatives[18,27].Acidicadditives,suchasTFA, suppresstheionizationofacidicanalytes.

Forpolysaccharide-basedchiralcolumns,theseeffectsdonot seemdirectlyrelatedtotheconcentrationoftheadditivesintheMP

[28].Weinvestigateddifferentadditiveconcentrationsinarange from0.1to0.25%andsawonlyaminorimpactontheretentionor resolution.Peakshapestendtobeslightlysharperwithincreasing additiveconcentrations.Ontheotherhand,addinglessthan0.1%

tothemodiﬁerwasnotsufﬁcienttoinducethedesiredeffect;peak

Table3

Forseparationstrategy:separationresultsandoptimalseparationconditionsforthe20compoundsfromthetestset.

Separationresults Selectedoptimalseparationconditions

Rs ˛ AT(min) CSP Flowrate(ml/min) Modiﬁer(%) Modiﬁertype

Carprofen 1.6 1.2 2.6 OZ-H 4.0 20 MeOH:IPA:TFA,100:0.1:0.1,v:v:v

Carteolol 2.6 6.5 1.3 OD-H 4.0 30 MeOH:IPA:TFA,100:0.1:0.1,v:v:v

Celiprolol 1.5 1.3 3.8 AD-H 4.0 15 EtOH:IPA:TFA,100:0.1:0.1,v:v:v

Ceterizine 5.5 2.4 2.2 AD-H 4.0 35 2PrOH:IPA:TFA,100:0.1:0.1,v:v:v

Clopidogrel 2.5 1.5 1.4 AD-H 4.0 35 2PrOH:IPA:TFA,100:0.1:0.1,v:v:v

Cyclopentolate 4.8 1.7 2.8 AD-H 3.0 20 2PrOH:IPA:TFA,100:0.1:0.1,v:v:v

Econazole 1.6 1.1 5.3 OZ-H 4.0 20 MeOH:IPA:TFA,100:0.1:0.1,v:v:v

Felodipine 2.0 1.2 4.9 AD-H 4.0 10 2PrOH:IPA:TFA,100:0.1:0.1,v:v:v

Fluoxetine 1.3 1.1 15.0 OZ-H 2.0 5 MeOH:IPA:TFA,100:0.1:0.1,v:v:v

Indapamide 1.5 1.3 3.6 OD-H 4.0 30 MeOH:IPA:TFA,100:0.1:0.1,v:v:v

Indoprofen 2.7 1.2 4.5 AD-H 4.0 35 2PrOH:IPA:TFA,100:0.1:0.1,v:v:v

Isradipine 1.6 1.1 7.2 LC-4 3.0 10 2PrOH:IPA:TFA,100:0.1:0.1,v:v:v

Lorazepam 3.0 1.4 2.9 OZ-H 4.0 35 MeOH:IPA:TFA,100:0.1:0.1,v:v:v

Miconazol 2.0 1.2 5.4 AD-H 4.0 15 2PrOH:IPA:TFA,100:0.1:0.1,v:v:v

d/l-Nebivolol 2.2 1.5 1.9 OZ-H 4.0 25 MeOH:IPA:TFA,100:0.1:0.1,v:v:v

Ondansetron 3.4 1.4 3.0 OD-H 4.0 40 MeOH:IPA,100:0.1,v:v

Temazepam 2.0 1.2 4.2 OZ-H 4.0 35 MeOH:IPA:TFA,100:0.1:0.1,v:v:v

Terazosine 1.7 1.2 3.7 AD-H 4.0 30 MeOH:IPA,100:0.1,v:v

Thioridazine 1.8 1.2 3.4 OZ-H 4.0 35 MeOH:IPA:TFA,100:0.1:0.1,v:v:v

trans-Stilbeneoxide 4.4 1.6 1.9 OZ-H 3.0 20 MeOH:IPA:TFA,100:0.1:0.1,v:v:v

(9)

Fig.7.Separationstrategyappliedontheracematethioridazine.Chromatogramsa–d:experimentsfromthescreeningstep,e:optimizedconditions.

shapesandchromatographicresultswereunacceptable.Hence,in thescreening,theadditiveconcentrationissetat0.1%IPAandTFA inthemodiﬁer.

Earlier,weobservedasignificantdifferenceinenantioselectiv- itybetweenthesimultaneoususeofIPAandTFAandtheindividual useoftheseadditives[18].Inthelattercase,TFAisusedforacidic compoundsandIPAforallothercompounds.Sincethesuccessrate tendedtobehigherwhencombiningtheadditives,weadviseusing thisapproachinascreeningstage[18].Moreover,thebenefitis thatthescreeningconditionsarethesameforallcompounds,inde- pendentoftheirchemicalproperties.However,incasethedesired enantioseparationisnotachieved,itcanbeusefultotryonlyone singleadditiveinthemodifier.Thisisthereforerecommendedin thepartialseparationbranchofthestrategy(seefurther).

3.3. Separationstrategy

Basedontheaboveinformationandearlierexperience,asepa- rationstrategy wasdeﬁned(Fig.6).Thisstrategywasevaluated withanoveltest setof 20pharmaceuticalracemates(Table1).

Afterexecutingthescreeningexperiments,18/20compoundswere separated.Afterapplyingtheentirestrategy,allcompoundswere baselineseparated,withtheexceptionof ﬂuoxetine,which was partiallyseparated(Rs=1.3)(Table3).

Analysistimefortheseoptimizedseparationswasin16/20cases below5min,for19/20below10minandforﬂuoxetine15min.

Theseparationstrategyappliedontworacemates,i.e.thiori- dazineandclopidogrelispresentedinFigs.7and8,respectively.

Thechromatograms(a–d)clearlyshowthecomplementarityofthe

Fig.8. Separationstrategyappliedonclopidogrelracemate.Chromatogramsa–daretheresultsfromthescreeningstep,eistheresultafteroptimization.

(10)

Fig.9.TransferofthechromatographicconditionsfromconventionalSFC(Tharequipment)toultraperformanceSFC(UPC²equipment).TheseparationsareobtainedonLux Cellulose-2,with20%(MeOH:IPA:TFA,100:0.1:0.1,v:v:v)inthemobilephase,ﬂowrate3.0ml/min,30^◦C,detectionat220nm,andabackpressureof150bar.(a)Mepindolol, (b)naringenin,(c)mianserine.

chromatographicsystemsincludedinthescreeningstep.Afterthe optimizationsteps,goodbaselineseparationswithsatisfyingpeak shapesandshortanalysistimesareobtained.

3.4. MethodtransferfromconventionalSFCtoUHP-SFC

Thescreening conditions from the separation strategy were transferred from a conventional SFC to an ultra-performance (UPC²)SFCequipment.Toevaluatethetransfer,the56-compound testsetusedasforthedeﬁnitionofthescreening wasapplied.

We refer to these earlier papers for more information on its composition[13,17,18].Thefourchromatographicsystemsfrom the screening were evaluated, i.e. OZ-H and OD-H, with 20%

(MeOH:IPA:TFA,100:0.1:0.1,v:v:v),andAD-HandLC-4,with20%

(2PrOH:IPA:TFA,100:0.1:0.1,v:v:v)intheMP.Thesamecolumns andconditionswereusedonbothinstruments.

Generally the method transfer from conventional to ultra- performance SFC seems rather straightforward. Usually similar separation results are achievedwhen applying thesame chro- matographicconditionsinconventionalandultra-performanceSFC (Fig.9).

However,the success rates on all chromatographic systems obtainedwith theultra-performance system are slightly lower (Fig.10).Inthiscontext,itisimportanttoanalyzetheresultsfur- thersincethedifferenceinsuccessratemayoriginatefromsmall differencesin resolution. A partialseparation is anyseparation witharesolutionhigherthanzero,whilebaselineseparationshave Rs>1.5.Hence,incaseaseparationwithresolution0.2isobtained ononeinstrument,asmalldecreaseinRsontheothermayresult inalossoftheseparation.

Wethuscomparedtheresolutionsandanalysistimes(AT)of the56compounds.TheobtainedRsandanalysistimesaresimilar

buttendtobeslightlylowerontheUPC²thanontheconventional equipment(Fig.11).Theselowerresolutionsarereﬂectedinthe lowersuccessratesontheUPC².Toassessthesigniﬁcanceofthe differenceinRsand ATbetweenbothinstruments,atwo-tailed pairedStudentt-testwasperformed.

Table4summarizestheresultsinterms ofthecalculated t- andp-values.Fortwochromatographicsystems,i.e.AD-HandLC- 4with2-propanolinthemobilephase,theresolutionswerenot signiﬁcantlydifferentontheconventionalTharSFCandUPC².For OD-HandOZ-Hwithmethanol,thedifferencewasdeterminedto besigniﬁcant.

TheanalysistimeswereslightlylowerontheUPC²thanonthe conventionalTharinstrument.Thisdifferencewasdeterminedto besigniﬁcantforallchromatographicsystems,withtheexception ofOD-Hwithmethanolinthemobilephase.

Fig.10.Numberofbaseline(Rs>1.5)andpartial(0<Rs<1.5)separationsachieved withtheTharSFCandUPC²systemsonthefourchromatographicsystemsofthe screening.

(11)

Table4

Two-tailedpairedStudentt-testappliedonthedataobtainedforthe56compounds(58enantiomerpairs)ontheTharandUPC². ChiralcelOZ-H20%

MeOH:IPA:TFA

ChiralpakAD-H20%

2PrOH:IPA:TFA

ChiralcelOD-H20%

MeOH:IPA:TFA

LuxCellulose-420%

2PrOH:IPA:TFA

Rs t-Value 3.11 1.66 3.10 0.41

p-Value 1.46×10⁻³ 5.15×10⁻² 1.50×10⁻³ 3.42×10⁻¹

AT t-Value 2.06 1.71 1.50 4.43

p-Value 2.20×10⁻² 4.62×10⁻² 6.93×10⁻² 2.14×10⁻⁵

WithRstheresolutionandATtheanalysistimeasresponses.NullhypothesisH0:XThar=X_UPC2,withXagivenresponse(RsorAT).t57,˛=0.05=1.67.Signiﬁcantt-andp-values aremarkedinbold.

Hence,toconcludeitcanbestatedthat,ingeneral,theanalysis timesontheUPC²areshorterthanontheconventionalSFCsys- tem.Thiscanberelatedtotheminimizationofthevoidvolume inthisequipment,resultinginalowervoidtime.However,this isnottranslatedintoseparationswithhigherresolutions.Inmost cases,theresolutionswereslightlylowerontheultra-performance SFCsystem,ontwoofthefoursystems,thisdecreasewassigniﬁ- cant.Thus,theresolutionsarerathercomparablebetweenthetwo systems,whileagaininanalysistimeisobtainedwiththeultra- performancesystem.

However,themaximalpotentialoftheUHP-SFCsystemmight notbeachievedwiththe5␮mparticlecolumnsusedinthisstudy.

Reducingtheparticlesizetosub-2␮mdimensions,wouldpossi- blyincreasetheseparationefﬁciencysigniﬁcantly[15,16].Sofar, nosub-3␮m chiral polysaccharide-based stationary phases are commerciallyavailable.Thecoatingofthepolysaccharide-based selector on the silica and the uniform and reproducible pack- ingofthesesmallerparticlesappearstobeverytedious.Hence, morepotentialliesinUHP-SFCforchiralseparationsprovidedthat adaptedCSPbecomeavailable.

Forthisstudy,wherethesamecolumnswereused,themethod transferfromtheconventional totheultra-performancesystem wasveryeasyandstraightforward.

3.5. Precisionstudy:conventionalSFCvsUPC²

Toevaluatetheprecisionofexperimentsonbothsystems,six enantioseparations;bopindolol,mepindolol,methadone,mianserine,naringenin,andsotalol,wereselectedandrepeatedtwiceover sixconsecutivedays.Thesamechromatographicconditionswere usedonbothsystems.LuxCellulose-2wasusedasstationaryphase,

with20%MeOH:IPA:TFA(100:0.1:0.1)in themobilephase.The totalﬂowratewas3.0ml/min,thetemperature30^◦Candbackpres- sure150bar.Detectionwasdoneat220nm.Thesampleloopwas 10␮landpartialinjectionsof5␮lweredoneforeachsample.

Theinter-andintra-dayvariabilitiesandtheintermediatepre- cision(expressedinvariance)wereestimatedforeachseparation usingANOVA.Table5showstheresultsforallseparationsonboth systems.Tworesponseswereconsidered:theresolutionandthe analysistime(AT).Thevariancesobtainedwithbothsystemswere comparedwithanF-test.

Theintra-dayvarianceontheRswasnotsignificantlydifferent betweentheUPC²andTharforthreecompounds.Formethadone and sotalol the variance was smalleron the Thar than on the UPC²,formepindololtheoppositewasseen.Theinter-dayvari- abilitywasnotsignificantlydifferentforthreecompounds,while formepindolol,mianserine,andsotalolitwaslowerontheUPC². Theintermediateprecisionwassignificantlydifferentfortwosep- arations:thevarianceformepindololwaslowerontheUPC²and formethadoneontheTharsystem.Theseresultsindicatethatthere isnodistinctbenefitofonesystemovertheotherconcerningthe repeatabilityofexperimentswhenconsideringtheresolutionas response.

Next,weconsideredtheanalysistimeasresponse.Threesep- arationsyieldedasigniﬁcantlydifferentintra-dayvariability,i.e.

bopindolol,methadoneandnaringenin.Theﬁrsttwoseparations showedalowervariabilityontheTharsystem,whiletheopposite situationwasseenforthelast.Theinter-dayvariabilitywassignif- icantlylowerontheTharsystemforbopindololandmethadone and on the UPC² for mianserine, sotalol and naringenin. The intermediateprecisionontheAT,waslowerforbopindololand methadoneontheTharsystemandfornaringeninontheUPC².

Table5

ResultsofthesixprecisionstudiesoftwosampleinjectionsonsixconsecutivedaysontheUPC²andTharsystems,expressedinvariances.

Intra-dayvariability Inter-dayvariability Intermediateprecision

UPC² Thar UPC² Thar UPC² Thar

Bopindolol

Rs 3.86×10⁻² 2.03×10⁻² 1.14×10⁻¹ 5.85×10⁻² 1.53×10⁻¹ 7.88×10⁻²

AT 2.47×10⁻² 6.23×10⁻³ 4.19×10⁻³ 6.58×10⁻⁴ 2.89×10⁻² 6.88×10⁻³

Mepindolol

Rs 1.57×10⁻³ 1.44×10⁻² 4.98×10⁻³ 1.78×10⁻¹ 6.56×10⁻³ 1.93×10⁻¹

AT 2.06×10⁻² 7.67×10⁻³ 1.18×10⁻² 1.89×10⁻³ 3.24×10⁻² 9.56×10⁻³

Methadone

Rs 9.9×10⁻⁴ 1.89×10⁻⁵ 2.99×10⁻⁴ 4.06×10⁻⁴ 1.29×10⁻³ 4.25×10⁻⁴

AT 2.06×10⁻² 3.87×10⁻³ 1.18×10⁻² 9.17×10⁻⁵ 3.24×10⁻² 3.96×10⁻³

Mianserine

Rs 1.88×10⁻³ 8.06×10⁻⁴ 7.03×10⁻⁵ 1.39×10⁻³ 1.95×10⁻³ 2.20×10⁻³

AT 8.78×10⁻⁴ 1.08×10⁻³ 6.33×10⁻⁶ 2.50×10⁻⁵ 8.84×10⁻⁴ 1.11×10⁻³

Sotalol

Rs 2.03×10⁻³ 3.14×10⁻⁴ 5.30×10⁻⁵ 7.43×10⁻⁴ 2.08×10⁻³ 1.06×10⁻³

AT 3.46×10⁻³ 2.80×10⁻³ 7.17×10⁻⁶ 4.17×10⁻⁵ 3.46×10⁻³ 2.84×10⁻³

Naringenin

Rs 2.48×10⁻³ 7.13×10⁻² 2.03×10⁻⁴ 1.33×10⁻³ 2.68×10⁻³ 7.26×10⁻²

AT 8.87×10⁻⁴ 4.15×10⁻² 3.75×10⁻⁶ 2.65×10⁻⁴ 8.91×10⁻⁴ 4.18×10⁻²

WithRstheresolutionandATtheanalysistime.TheresultsobtainedontheUPC²andThararecomparedwithanF-test,thesmallestvarianceofbothismarkedinboldif thedifferenceiscalculatedtobesigniﬁcant,F11,11;˛=0.05=2.82.

(12)

Fig.11.Comparisonofthescreeningresultsofthe56-compoundtestsetonthe ultra-performanceandconventionalSFCequipment.(a)Resolutions(Rs),(b)analysistimes.Straightline=lineofequality.

Hence,theintra-andinter-dayvariabilityandintermediatepreci- sionoftheanalysistimesbetweenbothsystemsarecomparable, andnodistinctadvantageofonesystemovertheotherwasseen.

Conclusively,theseexperimentsshowedthatintermsofpreci- siontheperformanceofbothsystemsweresimilar.

4. Conclusions

Todefineagenericseparationstrategy,theimpactofdifferent parametersonchiralSFCseparationswasinvestigated.Theinflu- enceoforganicmodifiertypeand–concentration,flowrate,back pressure,temperatureandadditives,wereconsidered.

Whendissimilar enantioselectivity is sought, it is advisable toscreendifferentmodifiersinthemobilephase.Methanolwas favouredover2-propanolandethanol,sincethismodifiertended togeneratehighersuccessratesonthepolysaccharide-basedCSPs, although a broad complementarity exists between MeOH and 2PrOH.Toextendtheenantioselectiverecognition,itisthusadvis- abletoscreenaCSPwithbothmodifiers.

Whenhigherresolutionsaredesired,themodifierconcentra- tioncanbedecreased.Whenaimingtodecreasetheanalysistime, theflowratecanbeincreasedwithoutcompromisingtheefficiency much.Thebackpressureandtemperatureonlyexertminorinflu- encesontheresolutionoranalysistimeofchiralSFCseparations onpolysaccharide-derivatives.Thelatterinformationwasusedto definea separation strategy, which applicability wasevaluated withanoveltestsetof20pharmaceuticalracemates.Allracemates

couldbebaselineseparated,withtheexceptionofﬂuoxetine.Anal- ysistimeswerebelow10minforallseparatedcompounds.

Thedevelopedapproachwastransferredfromaconventional toanultra-performanceSFCsystem.Similarseparationresultsin termsofRsweregeneratedbybothsystems,whiletheanalysis timeswereslightlylowerontheultra-performancesystem.The methodtransferthusprovedtobeveryeasyandstraightforward.

AprecisionstudywasperformedforsixseparationsontheThar andUPC²system.Resultsshowednodistinctadvantageofonesys- temovertheotherconcerningtheintra-,andinter-dayvariabilities ortheintermediateprecisionoftheresolutionandanalysistimeof theseparations.

Moreefﬁcientseparationscouldpotentiallybeachievedusing sub-2␮m columns. However, sofar,no CSPsare commercially availablewiththeseparticledimensions.Undoubtedly,therestill remainsawholeunexploreddomaininthiscontextforchiralsep- arations.

Conﬂictofinterest

Theauthorsdeclarednoconﬂictofinterest.

Acknowledgement

ThisworkwasﬁnanciallysupportedbytheResearchFoundation FlandersFWO(projects1.5.114.10N/1.5.093.09N.00).

References

[1]G.Terﬂoth,J.Chromatogr.A906(2001)301.

[2]G.Guiochon,A.Tarafder,J.Chromatogr.A1218(2011)1037.

[3]B.Chankvetadze,J.Chromatogr.A1269(2012)26.

[4]K.W.Phinney,Anal.Chem.72(2000)204A.

[5]Z.Wang,in:S.Ahuja(Ed.),ChiralSeparationMethodsforPharmaceuticaland BiotechnologicalProducts,Wiley,Calabash,NC,USA,2011,p.299.

[6]L.Miller,J.Chromatogr.A1250(2012)250.

[7]L.Miller,M.Potter,J.Chromatogr.B:Analyt.Technol.Biomed.LifeSci.875 (2008)230.

[8]E.R.Francotte,J.Chromatogr.A906(2001)379.

[9]H.Ates,D.Mangelings,Y.VanderHeyden,in:S.Ahuja(Ed.),ChiralSeparation MethodsforPharmaceuticalandBiotechnologicalProducts,Wiley,Calabash, NC,USA,2011,p.383.

[10]H.Ates,D.Mangelings,Y.VanderHeyden,J.Pharm.Biomed.Anal.48(2008) 288.

[11]D.Mangelings,Y.VanderHeyden,Adv.Chromatogr.46(2008)175.

[12]M.Maftouh,C.Granier-Loyaux,E.Chavana,J.Marini,A.Pradines,Y.Vander Heyden,C.Picard,J.Chromatogr.A1088(2005)67.

[13]K.DeKlerck,D.Mangelings,Y.VanderHeyden,J.Supercrit.Fluids80(2013)50.

[14]T.Ikai,Y.Okamoto, in:A.Berthod(Ed.),ChiralRecognitioninSeparation Methods–MechanismsandApplications,Springer-Verlag,Berlin-Heidelberg, Germany,2010,p.33.

[15]T.A.Berger,Chromatographia72(2010)597.

[16]C.Sarazin,D.Thiebaut,P.Sassiat,J.Vial,J.Sep.Sci.34(2011)2773.

[17]K.DeKlerck,G.Parewyck,D.Mangelings,Y.VanderHeyden,J.Chromatogr.A 1269(2012)336.

[18]K.DeKlerck,D.Mangelings,D.Clicq,F.DeBoever,Y.VanderHeyden,J.Chro- matogr.A1234(2012)72.

[19]TheEuropeanPharmacopoeia,DirectoratefortheQualityofMedicinesofthe CouncilofEurope(EDQM),Cedex,France,2005.

[20]G.O.Cantrell,R.W.Stringham,J.A.Blackwell,Anal.Chem.68(1996)3645.

[21]L.Miller,J.Chromatogr.A1256(2012)261.

[22]C.Hamman,M.Wong,M.Hayes,P.Gibbons,J.Chromatogr.A1218(2011) 3529.

[23]P.Franco,T.Zhang,J.Chromatogr.B:Analyt.Technol.Biomed.LifeSci.875 (2008)48.

[24]D.Leyendecker,in:R.M.Smith(Ed.),SupercriticalFluidChromatography,Royal SocietyofChemistry,Loughborough,UK,1988,p.53.

[25]K.W.Phinney,in:G.Subramanian(Ed.),ChiralSeparationTechniques,aPrac- ticalApproach,Wiley,Weinheim,Germany,2001,p.301.

[26]A.Tarafder,G.Guiochon,J.Chromatogr.A1265(2012)165.

[27]Y.K.Ye,K.G.Lynam,R.W.Stringham,J.Chromatogr.A1041(2004)211.

[28]K.W.Phinney,L.C.Sander,Chirality15(2003)287.