385 Chapter THERMAL ANALYSIS AND CALORIMETRIC METHODS APPLIED TO PHARMACEUTICAL SOLID FORMS Danièle Giron Contents 8.1. INTRODUCTION ...................................................................................................................................... 387 8.2. BASIC PRINCIPLES OF THERMAL METHODS AND THEIR IMPORTANCE IN PHARMACEUTICAL ANALYSIS ......................................................................................................... 388 8.2.1. Thermal analysis methods ................................................................................................... 388 8.2.2. Microcalorimetry...................................................................................................................... 388 8.2.3. Overview of applications....................................................................................................... 388 8.3. FACTORS TO BE CONSIDERED IN THE INSTRUMENTATION............................................ 391 8.3.1. Heat flux differential scanning calorimetry (DSC) and power compensation DSC ................................................................................................................... 391 8.3.2. Thermogravimetry .................................................................................................................. 397 8.4. WHAT ARE THE AREAS OF APPLICATION FOR THE STUDY OF SOLID FORMS? ...... 400 8.4.1. Thermodynamic data.............................................................................................................. 400
Thermal analysis techniques, in which a physical property is monitored asfunction of temperature or time while the analyte is heated or cooled undercontrolled conditions, are fundamental techniques for the characterisation ofdrug substances and products. They are fundamental as processing or agingconditions may be simulated, and because the methods give access tothermodynamicdata.Duetothedifferentinformationconveyed,thermalanalysismethodsarecomplementarytootheranalyticaltechniques,suchasspectroscopy,chromatography, melting point determination, loss on drying, assays foridentification, purity and quantitation. Thermal analysis techniques are basicmethods in the field of solid state analysis, in physical and chemicalcharacterisationof pure substances, aswell as formixtures.They find thebestapplication inpre‐formulation,processingandcontrolof thedrugproduct.Theintroduction of sophisticated, automated, robust and sensitive instrumentsconsiderably increased the advantages of these methods. New horizons havebeenopenedwiththeavailabilityofcombinedtechniquesandmicrocalorimetry,anemergingtechniquewhichisnowusedroutinely.
Sincechangesintemperatureandmoisturemightinducethechangesinthesolidstate,processingandstoragemayhaveaconsiderableeffectonactivity,toxicityandstabilityof compounds.Thepharmaceutical industry is facedwith thenewchallenges of quicker development and higher performance, in terms oftechnology, reliability and up‐scale in the international current GoodManufacturing Practise (cGMP) environment. Current requirements set by theInternationalConferenceofHarmonisation(ICH)forthecharacterisationandthequantitation of polymorphism in new entities [1], re‐enforce the position ofthermalanalysisandmicrocalorimetrictechniques,whichcandeliverthecorrectinformation concerning the thermodynamic relationships between phases, forthe proper selection of salt and crystal forms. The amorphous state is betterunderstood and determinable. Differential scanning calorimetry (DSC) purityanalysis isa fast,absolute,orthogonalpurity technique fororganiccompounds.Microcalorimetry, DSC and thermogravimetry (TG) are advantageous in theprocessofdesignofdrugproducts.
Whenamaterialisheatedorcooled,thereisachangeinitsstructureorcompo‐sition.Thesetransformationsareconnectedwithaheatexchange.Thefirstappli‐cationofthermalanalysiswaspresentedbyLeChatelierin1887[2].Indiffere‐ntialthermalanalysis(DTA),thetemperatureinducedinthesampleismeasured.DSC,whichdetermines theheat flow intoandoutof the sample, aswellas thetemperatureofthethermalphenomenonduringacontrolledchangeoftempera‐ture,isthebasicthermaltechniqueusedforsolidforms.Thermalanalysistechni‐quesalsocoverallothertechniquesinwhichaphysicalpropertyismonitoredasa function of temperature or time,while the sample is heated or cooled undercontrolledconditions.Forthesolidstate,thephasediagramrulesforsinglecom‐pounds,aswellasformixturesofseveralcomponents,havetobeconsidered.Themostcommonmethodsusedinsolidstatecharacterisationare:DSCandTG,andtheyareoftencombinedwithmassspectrometry,X‐raydiffraction(XRD),IR,Raman spectroscopy and microscopy. Books and reviews dealing with theprinciples,instrumentationandapplicationsaregiveninreferences[3‐10].Specificapplications of thermal analysis and calorimetric methods for pharmaceuticalpolymorphismaregiveninreferences[11‐19].
8.2.2. Microcalorimetry
A calorimeter measures the heat flow into or out of a sample, whereas adifferentialcalorimetermeasurestheheatofasamplerelativetoareference.Adifferentialscanningcalorimetercombinesthesetwomethods,andadditionally,heatsthesamplewithalineartemperatureramp.
Microcalorimetry in isothermalmode is a growing technique, complementary toDSC for the characterisation ofpharmaceuticals [20]. Larger sample volume andhighersensitivitymeanthatphenomenaofverylowenergy,immeasurablebyDSC,may be studied. The output of the instrument is measured by the rate of heatchange(dq/dt),asafunctionoftime,withasensitivitybetterthan0.1W.Micro‐calorimetrycanbeappliedtoisolatedsystemsinspecificatmospheres,orforbatchmode,wherereactantsaremixed in thecalorimeter.Themostusefulapplicationcurrently, thanks to high throughput microcalorimeters [21], is the routinequantitationofundesirableamorphouscontentdownto0.1‐0.3%[22‐25].
Solutioncalorimetrycanbeusedinadiabaticorisoperibolmodesinmicrocalori‐meters at constant temperature. This method is used for polymorphic inter‐pretationandforquantitation[12].
8.2.3. Overviewofapplications
The processing of the drug substances and drug products involves solvent(s),temperatureandpressurechanges,aswellasmechanicalstressand,asaresult,
differentsolidphasesmaycoexistinthedrugproduct.Organicsubstancesshowsupersaturationbehaviourandunstablesolidphaseswhichshouldnotexistatadefined temperature,pressureandhumidity,butmaybehave likestable forms.Thesesolidmetastablephases,obtainedoutsideoftheirdomainsofstability,willconvert to their thermodynamically stable forms at given temperatures, pres‐suresandrelativehumidities, in response tochanges inenvironmental conditi‐ons,processing,orovertime.Theseconversions,drivenbythermodynamics,aregoverned by kinetics and are influenced by impurities, particle size, crystaldefectsandpresenceofseeds.
The market withdrawal of Ritonavir in 1998, one year after its launch, wascausedbyithavingarateofdissolutionthatwasfartooslow,duetoaninsolublestable polymorph not found during development [26]. The current focus ofresearch in the solid‐state area is tounderstand theoriginofpolymorphismatthemolecularlevel,andtopredictandpreparethemoststableformatthestartofdevelopment.Theselectionofametastable formshouldresult fromtargetedchoice,ratherthanfromchance.Theadvantageofthermalandmicrocalorimetrictechniquesisthattheirsensitivityallowsasamplesizeofconsiderablylessthan1mg.Table8.1summariestheanalysismethodsdetailed,includingcoupledandcombinedtechniques.
DSCmeasures thetemperatureand theheat flowassociatedwithtransitions inmaterials,asafunctionoftimeandtemperature,inacontrolledatmosphereandrelativetoareference.Thesemeasurementsprovidequantitativeandqualitativeinformation about physical and chemical changes that involve endothermic orexothermicprocesses,orchangesinheatcapacity.
TheprincipleofDSCisasfollows:twoovensareheatedinalinearmanner.Oneovencontainsthesampleinapan;theothercontainsanemptypanasabalance,termedareferencepan.Ifnothingoccurstothesamplethenthesamplepanandreferencepanareat the same temperatureduringheating. If a change, suchasmelting occurs in the sample, then energy is used by the sample and thetemperature remains constant in the sample pan during melting. Therefore, adifferenceoftemperatureoccursbetweenthesamplepanandreferencepan.
Two methods of measurement are used by manufacturers. The first method istermed "heat flux DSC", in which the instrument measures the temperaturedifference(DTA).Throughcalibration,thistemperaturedifferenceistransformedintoheatflow(dq/dt).Theinstrumentcontainsafurnace,whichisablockforbothsample and reference cells. The sampleholder contains the samplepan, and thereferencepanisconnectedbyalowresistanceheatflowpath;thermocouplesareusedfordeterminingthetemperaturedifferencebetweensampleandreference.
Inthesecondmethod,whichistermed"powercompensationDSC",twoindividualheaters are used in order to monitor the individual heating rates of the twoindividual ovens. A system controls the temperature difference between sampleandreference.Ifanydifferenceisdetected,thespecificheaterwillbecorrectedsothatthetemperatureiskeptthesameinbothpans.Thatis,whenanendothermicor exothermic process occurs, the instrument delivers the compensation energywhichmustbegiveninordertomaintainequaltemperaturesinbothpans.
In the firstmethod (heat flux DSC), temperature is primarilymeasured, in thesecond case (power compensationDSC), energy isprimarilymeasured.Despitetheir different modes of operation, both instruments deliver the same infor‐mation:heatflowasafunctionoftemperature(ortime).
Chapter8
392
For first order transitions, such asmelting, crystallisation, sublimation, boilingetc.,integrationofthecurvegivestheenergyinvolvedinthetransition.Melting,boiling, sublimation and desolvation are endothermic, whichmeans they needenergy.Crystallisationisexothermic,whichmeansthatitsuppliesenergy.Solid‐solidphasetransitionanddecompositionmaybeendothermicorexothermic.Forsecondordertransitions,thesignalgivesthechangeinthespecificheat;forex‐ample,intheglasstransitionsofamorphousmaterials.Manufacturersrepresenttheheatflowdifferently:endothermshavethepositivevaluesinpowercompen‐sation DSC, and negative values for heat flux DSC. Figure 8.1 shows typicaltransitions for a single compound: with a power compensation DSC with anendotherm rising, or for heat flux DSC with an endotherm falling. Bothrepresentationsareacceptedifthesignsforendothermorexothermaregivenonthecurves.Generallythetemperatureisplottedonthexabscissa(usualgivenin°C)andtheheatflow(dQ/dt,dq/dtordH/dt)isplottedontheyordinateinJs‐1.Most recent curvesaregiven inmWatts (mW),which is equivalent, and in thischapterbothexpressionsareused.Melting,crystallisationandphasetransitionsarefirstordertransitions.Theintersectionbetweenthebaselineandtheslopeofthe peak is the extrapolated temperature onset of the transition (Te). Thecorrespondingenergy is theareaunderthepeak.Thepeaktemperature(Tm) isvery dependent on instrument andmeasurement parameters. The glass point,whichisasecondordertransition,isdeterminedasaninflexionpoint.
Modern instruments permit the heating, cooling and isotherm curves betweensub‐ambienttemperatures,withuseofacoolingdevice,andhighertemperaturesintherangeof1200‐1500°C.Curvesmadeusingdifferentinstrumentscannotbe compared to one another; correctionsmust be calculated according to eachinstrument.Greatcaremustbegivento thetemperature,withthetemperatureplotted on the abscissa being the programmed temperature, not the realtemperature of the sample. The actual temperature depends on the thermalresistanceandheatingrateoftheinstrument,whichiscalledthe"thermallag"ofthe system. The correlation between apparent melting point and heating ratedependsonthemanufacturer.Generally,pureindium(>99.9999%)isusedwitheach series of measurements. Modern instruments, with data handling, enablethetruetemperatureofeachpointoftheDSCcurvetobecalculated.
Figure 8.2 exemplifies the effect of the heating rate on the DSC curves. Anexample of the DSC curves of indium using different heating rates on a DSC‐2Perkin‐ElmerinstrumentisgiveninFigure8.2a.Sinceenergyistheintegrationofthe signal as a function of the time, the DSC curves are not identical at everyheating rate. The “thermal lag” is instrument specific. Since the meltingtemperatureofindiumis156.6°C,theextrapolatedonsethastobecorrectedforeach heating rate. Figure 8.2b shows the plot of the extrapolated onset of thecurvesforindiumversusheatingrateat1.25,2.0,5.0,10.0and20°Cmin‐1.Thiseffecthastobetakenintoconsiderationwhencomparingcurvesobtainedusingdifferentheatingrates,andespeciallyifmeasuredwithdifferentinstruments.
Figure8.3showstheinfluenceofheatingrateontheresolutionofpolymorphsoftemazepam.Thebest resolution isobtainedwitha lowheating rate. Figure8.4exemplifies the influence of the sample mass on the resolution for the samecompound.Theheatingrate(sensitivityoftheinstrument),andthesamplemasshavetobechosencarefullytodetectthermaleffectswhichareveryclosetooneanother.
Figure8.5dealswiththemeasurementof theglasstransitionwithandwithoutthermal effect. The thermal effect appearing as a maximum on DSC curvedependsonthethermalhistoryofthesample,andiscommonlyusedinthestudyofamorphousmaterialsandpolymers.Athighheatingratesthesignalishigherbecause,fortheglasstransitionthechangeincpisdq/dtordH/dt,andiseasiertomeasure.
Great effortshavebeenmade in recent years to validatedifferent instruments,notonlyincomparingprinciplesandresults,butalsoindeterminingheatingandcoolingrates,particlesize,weight,resolution,atmosphereandtypeofpans.Thetypeand thesizeofpans isofgreat importance forcomparingresults; typesofpanmayincludeopenpan,crimpedpan,sealedpanorpinhole.Thesametypeofpansmustbeusedforbothsampleandreference.Ifpansarehermeticallyclosedthenthesolventcannotescape,andthecurvesarenotthesameasforpanswithapinhole.Figure8.6exemplifiestheproblemforamonohydrate.Inasealedpan,theDSCmeltingpeakof themonohydrate isobserved.Whereas, ifapinhole ismade in the cover of the pan, water can escape and the transition into theanhydrousformisobservedasendothermic/exothermicphenomenonbeforethemelting of the anhydrous form occurs. Unlike the curves formonohydrate, theanhydrousformisnotaffectedbythetypeofthepan.
Theuseof flowingpurgegas (nitrogenor argon) is requiredas thewastepro‐ductsfromsublimationordecompositionhavetoberemoved.Duetothelubri‐cantusedinthemanufactureofaluminiumpans,thesepansmustbepreheatedinordertoavoidartefacts.Thehomogeneityofthesample,thesampleprepara‐tion,anycontaminations,thetimebetweenthemanufactureofthesampleanditsmeasurement,thesamplemassandthermallagarethemainsourceoferrors.Inpharmaceutical development, instruments must be calibrated and routinelychecked. It shouldbeemphasised thatmeltingpointsofpolymorphsmaydifferbylessthan1°C,andthatdifferencesinmeltingenergyareoftenlessthan5%.Therefore,allvariablesofanautomateddevicehavetobecontrolled.
Table8.2dealswithexamplesofstandards,includingIUPACandWHOstandards,proposedfortemperaturecalibration.Thestandardsmustbeofcertifiedpurity.For heat calibration, the melting energies of certified substances with meltingpointsbetween60°Cand230°CareproposedinTable8.2.
InmodulatedDSC (MDSC), thenormally linearheating ramp is overlaidwith asinusoidal function, or an oscillating function defined by a frequency and anamplitude, to produce a sine wave shaped temperature versus time function.UsingFouriermathematics,theDSCsignalisthensplitintotwocomponents:onereflecting non‐reversible events (kinetic), and the other reversible events. Thismethod is generally not applied in the characterisationof solidpharmaceuticalforms.Fordetailsonthisapplication,see[3‐8].
8.3.2. Thermogravimetry
Inthermogravimetry(TG),thechangeinsamplemassisdeterminedasafunctionoftemperatureand/ortime.Theinstrumentusedisathermobalancethatallowsthecontinuousweighingofasampleasafunctionoftime.Thesampleholderanda reference holder are bonded to each side of amicrobalance, in a vertical orhorizontal design. The sample holder is situated in a furnace, without directcontactwiththesample,thetemperatureofwhichiscontrolledbyatemperatureprogrammer.Thebalanceportion ismaintainedat a constant temperatureandthe instrument is able to record the loss or gain in mass of the sample as afunctionoftemperatureandtime[m=f(T)].Frequently,themasschangeisgivenaspercentageof thesample.Most instrumentsalsorecordthederivativeof themasschange[dm/dt=f(T)].
Balanceisadjustedwithcertifiedmassatambienttemperatureanditisrecomme‐nded to check themass at temperatures of use. Therefore, the author proposesthreestandards:calciumoxalatedihydrate(watercontent12.3%)(usedforTGinEuropean Pharmacopoeia), disodium tartrate dihydrate (water content15.7%)(used as standard for Karl Fischer analysis) and copper sulphate pentahydrate(watercontent36.1%).Figure8.7ashowssuchcalibrationcurvesat20°Cmin‐1heatingrate.Thedehydrationscoveratemperaturerangeof50to270°C.Inthisfigureweplotforcomparisonthecommonexcipientlactosemonohydrate.
Fortemperaturecalibration,thestandardsrecommendedbyICTAareferromag‐neticstandardsexhibitinglossofferromagnetismattheirCuriepointtemperatu‐re within a magnetic field: Nickel (354°C), Permanorm 3 (266°C), Numetal(386°C),Permanorm5(459°C)andTrafoperm(754°C).Athermobalancewithhorizontal plates is proposed by manufacturers as the melting of standards iseasierfortheuser.
Tables8.3a and8.3b showcalibrationsof a thermobalance at different heatingrates. In order to check the stability of the system, a baseline at the highestsensitivityhastobeperformedforallheatingratesinthetemperaturerangeofanalysis;thehighestdeviationwillbeobservedatthehighestheatingrate.
Figure8.7bshowstheuseofDTGforcoppersulphatepentahydrate,atthesameheating rate as given in Figure 8.7a. The DTG determines precisely the end ofeach step at rapid heating rates, which is an advantagewhen desolvation anddecomposition overlap. Figure 8.8 exemplifies the effect of heating rate on TG
curves of disodium maleate dihydrate. The results obtained at the 5 °C min‐1heatingratecomplywiththetheoreticalvaluesofwatercontentofthedihydrate(18.45%).
Polymorphismistheabilityofasubstancetocrystallise intodifferentcrystallinephasesthathavedifferentarrangements,and/orconformationsofthemolecules,inthecrystallattice.Thesecrystallineformsarecalledpolymorphs,orcrystallinemodifications. Polymorphs have the same liquid or gaseous state, but theybehave differently in the solid state. The amorphous state is characterised bysolidificationinadisorderedorrandommanner,structurallysimilartotheliquidstate. The expression pseudo‐polymorphism applies to hydrates and solvates.Hydratesorsolvatesdonothavethesamechemicalstructureastheanhydrousforms,thoughpolymorphismbetweenhydratesorsolvatesdoesoccur.
For each solid form, there is a solid‐liquid equilibrium curve and a solid‐vapourequilibriumcurve.Inthecaseofenantiotropy,thereisanequilibriumcurve(BFinFigure 8.9a), where both polymorphs are in equilibrium and undergo reversibletransitionataspecifictemperatureandatafixedpressure.Inthecaseofmonotropy(Figure8.9b),thereisnothermodynamictransitionbetweentwophases,sinceonlyone solid form is thermodynamically stable. The dashed lines correspond to themetastableequilibriumcurveswhichhavetobetakenintoconsideration.
At 0 Kelvin, G = H. Since the entropy S is always positive, G decreases withincreasing temperature. The energy diagram,G versus temperature at a givenpressure, reflects the transition observed between both solid phases, andbetween solid and liquid phases. If a transition betweenphases occurs then atthistemperaturebothphaseshavethesameGibbsenergy,G.
Ingeneral,thethermodynamicrelationshipbetweentwopolymorphicphasesisrepresentedbyplottingGibbsenergyasafunctionoftemperatureforeachform(Figure 8.10). If the two curves intersect below the melting point of eachpolymorph, a reversible transition occurs at the temperature Tt of theintersection.AttemperaturesbelowTt,polymorphAhasthelowerGibbsenergyandisthereforethethermodynamicallystableform,whileattemperaturesaboveTt ,polymorphB isstable.Thetransitionpointcanbe low,evenbelowambienttemperature, or in other cases, very high (above 200 °C). In the case ofmonotropy, there is no intersection of theGibbs energyof both forms, and thehighermeltingformisalwaysthethermodynamicallystableform.
Burger [27‐28], proposed to plot energy/temperature diagrams of the Gibbsenergy(G)andtheenthalpy(H),asfunctionsoftemperature.Thisproposalisthefundamentaltoolforthesolutionofcomplexpolymorphicsystems.AsshowninFigures 8.10a and 8.10b, a notable difference between enantiotropy andmonotropy is the melting enthalpy of the higher melting form. In the case ofenantiotropy,thehighermeltingformhasthelowermeltingenthalpy.Inthecaseofmonotropy,thehighermeltingformhasthehighermeltingenthalpy.
DSC,whichmeasureseveryheatflowchangeuponheatingorcooling,isthemostappropriate technique, since the method allows us to determine the meltingpoints and the melting enthalpies, as well as the transition points and thecorresponding energy. The influence of kinetics can be followed by usingdifferentheatingratesandtemperinginsitu[12].
The temperature of transition between two enantiotrops can be calculatedaccordingtoEquation8.6byneglectingthedifferencesinspecificheat.
Tt=TATB(HfAHfB)/(HfATB‐HfBTA) (8.6)
In the case of enantiotropy, the lowmelting form is the thermodynamic stableformbelowthetransitionpoint,andabovethispointthehighmeltingformisthethermodynamicstableform.Thetransitionpointcanbelow;closeto40°Cinthecaseoftolbutamide,closeto100°Cinthecaseofpropyphenazoneorevenhigherthan200°C.Figure8.11illustratesthecaseofthereversibletransitionbetweenpolymorphsoftolbutamide,withreversibleDSCheatingandcoolingcurves.
In thecaseofmonotropy, there isonly thehighmelting formpresent,which isthe thermodynamic stable form within the entire range of temperatures.Sometimes ametastable form can behave as a stable form, and the DSC curveshowsanexothermictransitionbeforethemeltingofthestableform,helpingtodefinethethermodynamicallystableanhydrousform(seeFigure8.12).
Duetokinetics,metastableformscanappearandbecausedifferentcaseshavetobe taken into account, it is mandatory to use different heating rates for DSCexperiments(seeSection8.4.3).Forsolidformselectionitisabsolutelynecessarytoknowthetemperatureofthetransitionofenantiotropicsystems,sinceit isathermodynamicvalue.Doesthetransitiontakeplaceat20°C,30°C,70°Corat120°C?Whatare the consequences fordrugdevelopment?The temperatureofthe process, the temperature of storage and transport are important factors inproduction.
Duetokineticbehaviour,thetemperatureoftransitionobservedinDSCishigherthan the temperature observed by solubility measurements or slurryexperimentsinthepresenceofsolvents.Table8.4showssomedataobtainedbycalculation,accordingtoEquation8.6,comparedtoexperimentaldata.
The phase diagrams of salts, solvates and hydrates aremore complex becausebinary mixtures are often of different compositions. The new compound mayundergo congruent melting (Figure 8.13a), or non‐congruent melting (Figure8.13b).Throughheating,themeltingofthesolvatemaybeobserved,followedbytransformation to an anhydrous form, or the solvent may be involved in anendothermic transition into the anhydrous form.A series of suchbinaryphasediagramshastobeconsideredifseveralcompoundsareformed.Thesediagramsarefundamentaltotheunderstandingofcrystallisationanddryingsteps.Analyti‐cal investigations of solvates are not possible without thermogravimetry.ExamplesofDSCandTGcurvescorrespondingat thesebehavioursaregiveninFigures8.14and8.15.
The effect of pressure on dehydration steps is exemplified by Figure 8.16 forcoppersulphatepentahydrate.Dependingonthepressure,dehydrationcangivetheanhydrousform,themonohydrateorthetrihydrate.Thisexplainsthecoexi‐stenceofdifferentstepsofhydrationduringdrying,millingortableting.
Ifaphysicalpropertyofacrystallinesubstanceisplottedagainsttemperature,asharpdiscontinuityoccursatthemeltingpoint.Foramorphoussubstancesthereisnomeltingpoint,andachangeofthebaselineoftheDSCcurveoccursattheso‐called glass transition temperatureTg. Below this temperature, the amorphousphasehascertainpropertiesofacrystallinesolid(e.g.plasticdeformation)andistermed “glassy”. At temperatures above Tg, the substance retains some of thepropertiesofaliquid,e.g.molecularmobility,andistermed“rubbery”.AboveTg,theincreaseinmolecularmobilityfacilitatesspontaneouscrystallisation,withanexothermicenthalpychangeaftertheglasstransition(seeFigure8.17).Theglasstransition temperature is decreased bywater, or other solvents and additives,facilitatingcrystallisation.Theamorphousstateisunstable,andthestudyoftheglasstransitionwithexcipientsunderhumidity, isapartofthepre‐formulationprocess[30].Amorphoussolidphasesareobtainedeitherbymeltquenching, ifno degradation occurs during themelting, or by lyophilisation. They are easilydetected by DSC, since the glass transition can be followed by the exothermiccrystallisation into thecrystallinestate,or canbedetectedbyXRD.Figure8.18shows a casewhere a very small amount of amorphous formwas detected byDSCalone.
Asalreadydescribed, themethodforpuritydeterminationpreviouslyproposed[31] is particularly attractive as it does not require a reference standard. Thismethoddeliversabsolutepurityorassayvalueand resultsareobtained in lessthan 30minutes. The determination of purity bymeans of DSC is based on theassumptionthatimpuritiesdepressthemeltingpointofapurematerialaccordingtoeutecticphasediagrambehaviour.Figure8.19ashowsthephasediagramforthetwo componentmixturewith a eutectic point. At the eutectic point, for example40%A,60%B, thecrystalsAandBmelt togetherat the temperatureTE (Figure8.19b),belowthemeltingtemperatureofthepurecompounds.IfamixtureofAandB(containinge.g.90%A) isheated, themeltingof theeutecticmixture(which is40% A) is observed initially, until all of B is melted. During the melting of theeutectic mixture (40%A, 60%B) a part of A is melted with B, with thecorresponding amount of A (2/3 x 10%, i.e. 6.66 % of A). As the temperatureincreases,pureAmeltsbetweenTEandTm(Tmbeingtheendofthemelting).ForthecorrespondingDSCcurve,ifanendothermicheatflowattheeutectictemperatureisobserved,thenthemeltingofcrystalsofAoccurs.
TheeffectofimpurityontheDSCcurveisameltingdepressionandabroadeningofthemelting curve. The amount of impurities is calculated from themelting pointdepression∆T=To‐Tm.Thevan'tHoff'slawfordilutedsolutionsis:
f2
0
- T Hx =
RT (8.7)
He
at
flo
w
En
do
T T
100% A0% A
A + B
B + Liq. A + Liq.
Tm Tm
TEx2
TE TE
T0
Liq.
Temperature, oCComposition
Heat flow Endo
a b
Chapter8
410
wherex is themole fractionof impurities,∆T themeltingpointdepression,∆Hfthemeltingpointofpurematerial,Tmthemeltingof theanalyte,To themeltingpointofthepurecompoundandRthegasconstant.
TheDSCproceduredoesnotdirectlymeasure∆T,butcanbeusedtocalculateitfromthemeltingcurve.AttheeutecticpointallofBisintheliquidphase.Duringthe melting of A, after the eutectic point, the concentration of B varies in theliquidphase.ThiscausesthebroadeningoftheDSCcurve(Figure8.19c).Withnosolidsolutionformation,theconcentrationofimpurityintheliquidphaseatanytemperatureduringthemelting is inverselyproportional to the fractionmeltedatthattemperature,andthemelting‐pointdepressionisdirectlyproportionaltothe mole fraction of impurity. A plot of the observed analyte temperature,Ti,versusthereciprocalofthefractionmelted1/FiattemperatureTi,shouldyielda straight line with the slope equal to the melting‐point depression (To ‐ Tm)(Figure8.19d).The theoreticalmeltingpointof thepurecompound isobtainedbyextrapolationtol/Fi=0:
2o
ii 0
f
1
RTF
T T xH
(8.8)
Thecorrection,K,hastobemadecorrespondingtothepartalreadymeltedintheeutectic, and the detected beginning of themelting curve. Figure 8.19d depictssuch a determination. Software from manufacturers mostly uses iterativelinearisationwhichgivesthebestvalueofK.
Thecalculationofpurityallowstheinterpretationofthecomplexcurvesgivenbyendotherms preceding the melting. It has been used successfully in stabilitystudies at the beginning of development. An example is given in Section 8.4.4,Table8.5.
8.4.3. Kineticaspects
EveryDSCstudyshouldincludescansatdifferentheatingratesduetoDSCbeingadynamicmethod,andsolid‐statetransformations,whilebeingthermodynami‐cally driven, are kinetically controlled. The DSC scans will differ if the sampleunderstudyisstableormetastableatambienttemperature.AtypicalsetofDSCscansillustratesthestudyofthepolymorphicrelationshipoftwoformsAandB.
Scan3.ThesamplestudiedisthestableformAwhichmelts.FormBcrystallisesfrom the melt with an exothermic peak and form B melts at a highertemperature.
Scan4.ThesamplestudiedisthemetastableformB,whichbecomesstableatatemperature above the transition temperature. An exothermic peakcorresponds to the solid transformationB A, followed by successivetransformationAB,andmeltingofB.
Scan5.The sample studied is the metastable form B. The DSC scan shows itsmeltingendotherm.
Scan1.The sample studied is the stable form A, and its melting endotherm isobserved.
Scan2.The sample studied is the metastable form B, which transformsexothermicallyinthesolidstateintothestableformA.FormAmeltsatahighertemperature.
Scan3.The sample studied is themetastable form B, which does not transformintoA,butmeltsendothermically.FromthemeltthestablecrystallineformAappears,withanexothermicpeak,thenAmeltsatahighertemperature.
a b
Heat flow dq/dt
Endo
Heat flow dq/dt
Endo
Temperature
Temperature
Chapter8
412
Similar interpretationsapply to allmethods that involveheating (e.g. hot stageopticalmicroscopy,hotstageinfraredorRamanmicroscopy,temperatureresolv‐edorvariabletemperatureXRD).ItisdifficulttodistinguishbetweenenantiotropyandmonotropyinthecasesofFigure8.20a,Scan5andFigure8.20b,Scan1,aswellasinFigure8.20a,Scan3andFigure 8.20b, Scan 3. The interpretation of theDSC curves is facilitatedbyBurger’senthalpyoffusionrule[27‐28]:ifthehighermeltingformhasthelowermeltingenthalpy,thenbothformsarerelatedenantiotropically.AsdemonstratedinFigure8.21,forabenzisoquinolinehydrochloride[12],themeltingenthalpyofthehighermeltingformBislowerthanthemeltingenthalpyofA.Therefore,thetwo forms are enantiotropically related, with form A being stable below thetransitionpoint.ModificationBishygroscopicandundergoesasolvent‐mediatedtransitiontoformAinalcoholsatambienttemperature.
Figure 8.22 shows DSC curves of propyphenazone. Samples from two manu‐facturerswere investigated in our laboratory since the second sample did notcomplywithmeltingrequirements.Anunexpectedlyhighmeltingpointwasnotdue to higher purity, but to the enantiotropic behaviour. The second samplecontainedtracesofahighmelting form.Aslowerheatingrateshowedforbothsamplestwoendotherms.DSCexperimentsallowedthemanufactureinsituofthe
high melting form, and to characterise it completely including solubility anddissolutionrate[12,13].Forexamplesofkineticeffectsduetopolymorphismandsalts, see references [32,33]. Kinetics also governs the steps of hydration anddehydration in the solid state, giving rise to complex conversion relationships[34‐36],(seeexampleinFigure8.25).
Mostdrugsubstancescrystalliseintheformofsolvatesorhydrates.TheDSC/TGcurvesallowustofollowthedesolvationprocess.ItispossibletomeasurewatersorptionisothermsbyTGincontrolledatmospheres.Oftentheinstrumentusedisanautomaticmulti‐vapourgravimetricsorptionanalyser,aso‐called“dynamicvapour system” (DVS), e.g. those made by Surface Measurement Systems Ltd.Figure 8.23 shows such curves for an enantiotropic drug substance whichabsorbs water to form a trihydrate. The desorption occurs with hysteresis,allowingustodefinetherelativehumidityrangeforeachform.Accordingtothecurves, the trihydrate formwould be possible to develop.However, itwas notpossibletomanufactureitasapureformintheprocess.Theenantiotropicform,which was stable at room temperature, could be manufactured and stored intightcontainers[14].
98 100 102 104 106
Manufacturer 1
Manufacturer 2
Temperature, oC
Hea
tfl
ow
En
do
100 101 102 103 104 105
Temperature, oC
a) Heating rate 2.5 C min-1 b) Heating rate 1.2 C min-1
Figure8.24showstheuseofsorptionisothermmeasurementforasaltselection.Thehydrochloridefirstselectedwastoohygroscopicforeasydevelopment,andtherefore new saltswere studied. TheDSC puritywas determined for stabilitycomparison (Table 8.5). The hydrogen fumarate and hydrogen tartrate wereattractiveduetotheirgoodhygroscopicbehaviour,howeverthestabilityresultsclearly showed that hydrogen tartrate was chemically stable. This exampledemonstratestheadvantageoususeofthermaltechniquesforquickdecisionsinearlydevelopment.
Itisrecommendedthatseveralcyclesofsorptionanddesorptionareperformedwhenahydrate isproduced.Figure8.25shows thekineticeffectof seeds foradrug substance. The anhydrous form is converted to the monohydrate at arelativehumidityofRH>90%.Afterthedesorption,tracesofmonohydratearepresentandthesubstanceabsorbswaterveryfast.Thebehaviourwasalsofoundinstabilitystudiesofbatchesofthedrug[32].
Recentlyintroducedautomaticsystemsbringafasterunderstandingofcomplexhydration and dehydration behaviour, aswell as their kineticswithmonitoredrelativehumidity (RH%). Instruments for researchhave0.1µgsensitivity,andlessthan10mgofsamplecanbeused.Byusingsolventvapours,solvatedformscan be manufactured and studied. These studies are a prerequisite for themicrocalorimetricanalysisofamorphouscontent.
XRDinstrumentsareequippedwithheatingcellsandareroutinelyusedinmostresearchlaboratories.X‐raypatternsofpureformsobtained insitucanbeusedformodelling,andinfavourablecases,forcrystalstructuredetermination.Figure8.26exemplifies thecombinationofDSC/XRD foranenantiotropic systemwithheatingandcoolingcurves.Theadvantageofcoupledinstrumentsdevelopedinaresearch laboratory has been demonstrated [37]. The trehalose monohydratedehydration and hydration were studied in a humid atmosphere as anapplicationofanewcommercialinstrumentation[38].
AnexampleofTG‐MSisgivenforcalciumoxalateinFigure8.27,wherewater,COand CO2 are positively identified. Figure 8.28 is an example of the successfulinterpretation of a crystallisation into solvates followed by desolvation to ametastable form.TheDSCandTG‐MSinthis figurecorrespondtothemethanolsolvate. Similar pictures are obtained with other solvates e.g. acetone andethanol.TheDSC/TG/TG‐MSandtemperatureresolvedXRDexperiments,aswellas slurry experiments,were necessary to understand the crystallisation of thisdrugsubstanceandtochangethesolventofcrystallisation[32].
Figure8.29a shows theDSC/TG for aspartamehemihydrate andcorrespondingIR spectra (Figure 8.29b). The first endotherm at around 130 °C is the loss of
An example is given in Figure 8.30 of the interpretation of the behaviour of amalonate using several combined technologies: DSC, TG, TG‐MS, temperatureresolved XRD and IR. Through drying, themalonic acid can sublimate and the
base formed [14].Anotherexample, given inFigure8.31,dealswith the lossofwaterfromadrugsubstanceduringdrying.Thedecompositionproductformedwas at first believed to be a new form. DSC/TG/XRD and IR were thensuccessfully used for the identification of the substancewith two enantiotrops[15]. The DSC of the stable form shows an enantiotropic transition. After themeltingofthesecondform,adegradationintothelactamwiththelossofamolarquantityofwaterwasdetectedbyTG‐MSandtemperatureresolvedXRD.TheIR‐heatingcellallowedthestructureofthelactamtobeconfirmed.
Thedecreaseintheglasstransitiontemperature,Tg,bywaterandsolvents,istheprinciple behind the microcalorimetric method of amorphous contentdetermination. The substance is subjected to vapours ofwater, or organic sol‐vent, in an isothermal microcalorimeter (Figure 8.32). The heat flow ofcrystallisationwhichisproportionaltotheamorphouscontentismeasured.Thesorption‐desorption isothermpresented inFigure8.33 exemplifies theprocess.The un‐milled sample is not hygroscopic but after micronisation, the powderabsorbswater(cycle2),andanincreaseinmassisobserved.Anabruptdecreaseisobservedincycle3asthedrugcrystallisesandloseswater.Thecorrespondingexothermicenergyofcrystallisationcanthenbemeasured.
Figure8.34 showsan exampleof thedeterminationof amorphous content in abatch of micronised substance for inhalation, by using two different solventvapours, and themethodswere validated. The same result is obtained for this
Type: Thermal Activity Monitor,Thermometrics
Sample volume: 3 ml, glass ampoule
Temperature: 20 – 80 oCPrecision: 0.1 % (at 300 W)Baseline stability: 0.2 W (8 h)Humidity control: e.g. 57 % RH
batch: 12.9% compared to 12.7 % amorphous content. However, as demon‐stratedinFigure8.34,thetimeofcrystallisationisquitedifferent.Withvapoursfromamixtureofethanolandwater,thecrystallisationisrapid,butthestartofcrystallisation cannot be measured for low amounts of amorphous form. Thismethod can be used in the range 7 ‐ 15% of amorphous content. In order toanalyse lower contents, dimethylformamide (DMF) used as vapour, can actslowly, and themethod has been validated in the range 2 – 30% amorphouscontent.Anautomatic instrumentwith48channels(TAinstruments),allowsustoperformroutineanalysiswithalimitofdetectionatthe1%level.
SolutioncalorimetryhasbeendescribedinUnitedStatesPharmacopeia(USP)forthe determination of the amorphous form, but has also been used forquantitationofpolymorphs[14].Thismethodsupposesthatnoothercrystallineformmayinterfere.Inthecaseofenantiotropy:
Thismethodcanbeusedforpolymorphs,asdemonstratedinFigure8.35foranenantiotropicdrug.ThesameresultsfortheheatoftransitionarefoundbyDSCandsolutioncalorimetry[12,16,18].Solutioncalorimetryisveryhelpfulwhenthesubstancedecomposes onmelting. This approach allowsus to choosebetweenformsandselectthemoststable[14,18].
The following example, given in Figure8.36, is a situation often found in earlydevelopment.Wereceivedasample,obtainedbyprecipitation,forwhichtheDSCscan exhibits dual melting. By using different heating rates it was possible to
calculate the melting enthalpies and it was concluded that the sample is anenantiotropicsystem.Thepolymorphismstudyofslurriesindifferentsolventsofa purer sample gave a third form. According to Burger’s rule, and theequilibrationstudy,Cisthestableform,AandBareenantiotrops,andbotharemonotropstoC.Consequently,formCwasselectedforfurtherstudy[32].
Solvatesareveryoftenformedinthepresenceofvolatilesolvents.Byisolationinthe air they easily transform into anhydrous forms, stable or unstable, inamorphouspowder,orinhydratedforms.Thebestwaytodetectthemwouldbedirectmeasurementinthecrystallisationvessel.Figure8.38showstheresultsofalargenumberofexperimentstounderstandthecomplexsituationfoundwithadrugsubstance.
Water is present in the air and in excipients, and is oftenused inmixtures forcrystallisation purposes. The TG curves of hydrates can be consideredfingerprints. Figure 8.39 is an example of the coexistence ofmonohydrate andtrihydrate in the solvent mixture of a process. The TG curve permitted theevaluation of thepresenceof the trihydrate, since trihydrate andmonohydratehadseparatedstepsforthelossofwater[16,18].Figure8.40isanexampleoftherelationshipsneededtoselectthermodynamicwindowsforhydrates[12].
Crystallisation in ethanol <40 oC form A, >50 oC form BWhich form to be selected?
Burger’s rule? A: one DSC peak B: two DSC peaks with decompositionX‐ray heating cell: B A before melting and decomposition
Slurry experiments form A, form B and mixture A + B at 10, 25, 40, 50 and 60 oC at different times and solvents. Analysis of solids
Inconclusion,thermalanalysisandcalorimetrictechniquesarenecessaryfortheselection and the characterisation of solid forms. They are versatile and offerquickresults.Asexemplified inFigure8.41,characterisationof therelationshipbetweenallformsistheobjectiveofdevelopment.Forthischallengingpurposearange of methods is necessary, and when this is achieved, process analyticaltechnologywillbethefutureforupscaleandmanufacture.
1. Guideline SpecificationQ6A. Investigating the need to set acceptance criteria forpolymorphismindrugsubstancesanddrugproducts.InternationalConferenceonHarmonization(ICH)(1999).
6. D.Q.M.Craig,M.Reading.ThermalAnalysisofPharmaceuticals.CRCPress,UK2006.7. W‐P. Pan, L. Judovits. Techniques in thermal Analysis, hyphenated techniques,
ThermalAnalysisofsurface,fastrateanalysis.ASTM,2007.8. D. Giron, Thermal Analysis of Drugs and Drug Products. In: Encyclopedia of
PharmaceuticalTechnology.(Eds.J.Swarbrick,J.C.Boylan)MarcelDekker,2006.9. D. Giron. Thermal Analysis and microcalorimetric methods in the industry:
essential techniques for proper development of pharmaceuticals. EuropeanPharmaceuticalReview2006(5)84‐90.
10. D.Giron.LatestDevelopment in theareaof thermalanalysisandcalorimetrythatareimpactingonpharmaceuticalmanufacturing.EuropeanPharmaceuticalReview2008(2)71‐76.
11. B.Rodriguez‐Spong,C.P.Price,A. Jayasankar,A.J.Matzger,N.Rodriguez‐Hornedo.General principles of pharmaceutical solid polymorphism: A supramolecularperspective.AdvancedDrugDeliveryReviews56(3)(2004)241‐274.
12. D. Giron. Thermal analysis and calorimetric methods in the characterization ofpolymorphsandsolvates.ThermochimicaActa248(1995)1‐59.
13. D. Giron. Le polymorphisme. Labo‐Pharma Problèmes et techniques. 307 (1981)151‐160.
14. D. Giron. Investigations of polymorphism and pseudo‐polymorphism in pharma‐ceuticals by combined thermoanalytical techniques. Journal of Thermal AnalysisandCalorimetry64(1)(2001)37‐60.
15. D. Giron. Monitoring of polymorphism from detection to quantification. Engine‐eringinLifeSciences3(3)(2003)103‐112.
18. D. Giron, S. Monnier, M. Mutz. Challenging characterization and monitoring ofpolymorphs. Latest developments. Lecture presented at AFCAT‐GEFTA‐STKmeeting,Mulhouse,2008.
21. M. Mutz, A. Motreff, M. Monnier, T. Buser, P. Schwab, D. Giron. Use of highthroughputmicrocalorimeterforfasterdeterminationofamorphouscontent.STKannualmeeting,June2006,Fribourg,Switzerland.
26. S.R.Chemburkar,J.B.K.Deming,H.Spiwek,K.Patel,J.Morris,R.Henry,S.Spanton,W.Dziki,W.Porter,J.Quick,P.Bauer,J.Donaubauer,B.A.Narayanan,M.Soldani,D.Riley,K.McFarland.Dealingwith the impactof ritonavirpolymorphson the latestages of bulk drug process development. Organic Process Research andDevelopment4(5)(2000)413‐417.
27. A.Burger, R. Ramberger. On the polymorphism of pharmaceuticals and othermolecular crystals. Theory of thermodynamic rules. Mikrochimica Acta [Wien] II72(3)(1979)259‐271.
28. Burger,U.J.Griesser.ScientiaPharmaceutica.58(1990)423‐430.29. M. Soustelle. In: Handbook of powder technology, Vol.9.PowderTechnologyand
pharmaceuticalsystems.JournalofPharmaceuticalScience86(1)(1997)1‐12.31. D. Giron, C. Goldbronn. Place of DSC purity analysis in pharmaceutical develop‐
ment.JournalofThermalAnalysisandCalorimetry44(1)(1995)217‐251.32. D.Giron.Polymorphism:Thermodynamicandkinetic factors tobeconsidered in
33. D. Giron. Characterisation of salts of drug substances. JournalofThermalAnalysisandCalorimetry73(2)(2003)441‐450.
34. D. Giron, C. Golbronn, M. Mutz, S. Pfeffer, P. Piechon, P. Schwab. Solid statecharacterization of pharmaceutical hydrates. Journal of Thermal Analysis andCalorimetry68(2)(2002)453‐465.
35. W.Chongcharoen, S.R. Byrn, N. Sutanthavinu. Solid state interconversion ofnorfloxacinhydrates.JournalofPharmaceuticalScience97(1)(2008)473‐487.
39. Rodriguez, D.E. Bugay.Characterization of pharmaceutical solvates by combinedgravimetricandinfraredanalysis.JournalofPharmaceuticalScience86(2)(1997)263‐266.
41. D.Giron,M.Draghi,C.Goldbronn,S.Pfeffer,P.Piechon.Studyofpolymorphismoftetracaine hydrochloride. Journal of Thermal Analysis and Calorimetry 49(2)(1997)913‐927.
