ContentslistsavailableatScienceDirect
Journal
of
Chromatography
A
j ou rn a l h om ep a ge :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
Lewis
acid/base
character
and
crystallisation
properties
of
poly(butylene
terephthalate)
José
M.R.C.A.
Santos
∗,
James
T.
Guthrie
DepartmentofColourandPolymerScience,SchoolofChemistry,TheUniversityofLeeds,WoodhouseLane,Leeds,WestYorkshireLS29JT,UnitedKingdom
a
r
t
i
c
l
e
i
n
f
o
Articlehistory:
Received11September2014 Receivedinrevisedform 12December2014 Accepted15December2014 Availableonline29December2014 Keywords:
Inversegaschromatography Poly(butyleneterephthalate) Lewisacid/baseproperties Crystallisation
a
b
s
t
r
a
c
t
Twogradesofpoly(butyleneterephthalate)wereanalysedbymeansofinversegaschromatography
(IGC)andtheresultscorrelatedwiththerespectivecrystallisationproperties.Thefollowingparameters
weredeterminedbyIGC:thedispersivecomponentofthesurfacetension,theenthalpyandtheentropy
ofadsorptionofselectedpolarandapolarprobes,andtheLewisacidityandbasicityconstants,Kaand
Kbrespectively.TheinterpretationofthevaluesdeterminedforKaandKbisinagreementwiththeFTIR
spectrarelatingtothecarboxylend-groupandthehydroxylend-groupconcentrationsinthesepolymers.
Thedifferencesinthemolecularweightvaluesandintheend-grouptypeandconcentration,between
thetwogradesofPBT,donotcausedifferencesinthecrystallisationactivationenergy.Thisobservation
suggeststhatthereisaleadingcontributionoftheLewisbasicsitestothecrystallisationactivationenergy
ofthegradesofPBTthatwereanalysed.However,thelowervalueofKaandthegreatermolarmassof
oneofthePBTgradesleadtoacorrespondinglowercrystallisationdegree.
©2014ElsevierB.V.Allrightsreserved.
1. Introduction
Duetotheirsemi-crystallinecharacterandhighmelting tem-peratures,poly(butylene terephthalate), PBT, and poly(ethylene terephthalate), PET, are the more widely used thermoplastic polyesters.Common applicationsof PET include fibrespinning, blowmouldingofbottles,injectionmouldingofengineeringparts andthermoformingoftrays.PETis,nevertheless,characterisedby lowcrystallisationrates.Thishaslimiteditsapplicationininjection mouldingappliances.Toovercomethisdrawback,fastcrystallising gradeshavebeendevelopedthatcontainnucleatingagents(e.g. BaSO4,ZnPO4andSb2O3)toincreasethenon-isothermal crystalli-sationtemperatureandreducethesizeofthespherulites,and/or containotherpolymers.PBT,ontheotherhand,exhibitsa signif-icantlyfastercrystallisationrate.The realisedshorter moulding cyclesandthelowerviscosityunderappropriateconditionshave resultedinawideruseofPBTthanPETformouldingapplications. PBTisactuallyoneofthefastercrystallisingpolymers[1–5].The chemicalnatureoftherepeatunitisshowninFig.1.Thehigh crys-tallisationrateisaconsequenceoftheconsiderablemobilitythat
∗ Correspondingauthor.Presentaddress:PolytechnicInstituteofBraganc¸a, Cam-pusSta.Apolónia,5300-253Braganc¸a,Portugal.Tel.:+351273330832; fax:+351273325405.
E-mailaddresses:josesantos@ipb.pt,josemrcasantos@gmail.com
(J.M.R.C.A.Santos).
isprovidedbythebutyleneunitinthechain.Usually,purePBThas acrystallinityextentthatisintherangeof30–40%.Theenthalpy offusion,Hf◦,ofthe100%crystallinePBTis142J/g[4,6–8].The meltingtemperatureTmisabout225◦Candthecrystallisation tem-perature,Tc,is approximately180◦C [9].The valueoftheglass transitiontemperature,Tg,ofthesemicrystallinestateoccursat about40◦C.ThecompletelyamorphousstateofpurePBTis diffi-culttoproduceduetothehighcrystallisationrateofthispolymer. Nevertheless,avalueof−25◦ChasbeendeterminedfortheTgof 100%amorphousPBT[7,10].
Becauseofitssemi-crystallinenature,PBToffersconsiderable chemical/solvent resistance but low dimensional stability, low ductility,low glasstransitiontemperatureand low Izodimpact strength.Consequently,core–shelltypeelastomersaremost fre-quentlyaddedtoPBTtoimproveitstoughness.Theseblendsshow lowmeltviscosityand,thus,areeasytoprocess.Moreover,the blendshavegoodmechanicalpropertiesandelectricalproperties, excellentsolventresistance,andgoodhydrolyticstability.Blends ofPBTwithelastomersandwithotherpolymershavebeenusedfor theinjectionmouldingofexteriorautomotivepartssuchasmirror housingsandbumpers,andareoftendesignedtoreplacemetalsin specificapplications.
Furthermore,inordertotakefulladvantageoftheuseful prop-erties ofPBT, and toovercomethe aforementionedlimitations, PBTisusually combinedwithpolymerssuchaspolycarbonates, PC,andPET(e.g.Xenoy®andValox®fromSABICInnovative Plas-tics,Makroblend®fromBayer,andPocan®fromLanxess,Sabre® http://dx.doi.org/10.1016/j.chroma.2014.12.042
Fig.1. RepeatingunitinPBT.
Table1
Weightaveragemolecularweight(Mw),molecularweightdistribution(D)and
number-averagemolecularweight(Mn),carboxylend-groupconcentrationand
hydroxylend-groupconcentration,forPBTAandPBTB.
Mw(gmol−1) D Mn(gmol−1) −COOH(eq/g)
PBTA 108,500 3.18 34,100 48
PBTB 46,000 2.70 17,000 29
1600fromDowChemicalCompany, Stapron® EfromDSM and
Ultrablend®KRfromBASF).Inthepolymerblendsthusobtained,
controloverthecrystallinepropertiesofthePBTisnecessary.
Dur-ingprocessingatelevatedtemperatures,hydrolytic,thermal,and
oxidativedegradationofPBTcanoccurwiththeformationofnew
carboxylicacidend-groups,alongwiththereductioninthe
molec-ularweightand alterationof themolecularweightdistribution.
Theseeffects resultin higher crystallisationratesand
crystalli-sationextents[4,5,8,11].IntheparticularcaseofPBT/PCblends,
theformation of a PC-PBT copolyester results in a decrease in themolecularweightofthePCandofthePBT,inareductionin thedegreeofcrystallinityandinthecrystallisationtemperature [2,6,7,12–14].
The growingawareness of the importanceof solid surfaces, interfacesandinterphasesindeterminingtheusefulpropertiesof polymericcompositions,hasledtothedevelopmentofinversegas chromatography(IGC)asavaluabletechniqueforevaluatingthe potentialforinteractionofdifferentcomponentsofpolymerblends andcompositesandmulticomponentpolymericsystemsin gen-eral.Duetoitsapplicabilityindeterminingthesurfaceproperties ofsolidsinvariedformssuchasfilms,fibresandpowdersofboth crystallineand amorphousstructures,IGC hasbeenextensively usedforsurfacecharacterisation.TheapplicabilityofIGCin mea-surementsofphysicochemicalpropertiesofvariousmaterialshas beenthoroughlydescribedintheliterature[15–18].Dataobtained fromIGCexperimentsmay,infavourablecases,correlatedirectly withobservedperformancecriteria,suchascolourdevelopment, gloss,rheologicalproperties,adhesionandmechanicalproperties [19–21].
Twopapershavebeenpublishedthatdealtwiththeanalysisof thethermodynamicpropertiesofPBT(surfacefreeenergy,surface LewisacidityandsurfaceLewisbasicity),onethatconcernsastudy carriedoutonaparticulargradeofPBTandasecondthatrelates tosurfacetensionstudiesonPBT,wherethedispersivecomponent ofthesurfacetensionwasquantifiedusingcontactangle measure-ments[22,23].Thespecificcomponentofthesurfacefreeenergy (withoutquantifyingthecontributionoftheLewisacidicsitesand oftheLewisbasicsites)wasdeterminedthereof.
Differencesinend-group“concentrations”alongsidewith dif-ferencesintheaveragemolecularweightofPBTpermitthetailoring of thecontributionof thePBT phase towardstheproperties of thePBT-basedblends,namelytheviscosityandthecrystallinity. Thus,theyinfluencetheprocessabilityandmechanicalproperties ofthesepolymericsystems.Forthisstudy,IGCatinfinitedilution wasusedas atool by whichtoassess thedifferences between thesurfaceLewisacidic/basicpropertiesoftwogradesofPBT, dif-ferentiatedbytheirvaryingend-group“concentration”.Someof theeffectsthatthedifferencesinend-group“concentration”and molecularweighthaveonthecrystallisationpropertieswerealso examined.
Table2
RelevantcharacteristicsofcommonlyusedIGCprobes.
Probemolecule a(d
l)
0.5 (cm2(mJcm−2)0.5)
DN(kJ/mol) AN*(kJ/mol)
n-Hexane 2.21E−16 n-Heptane 2.57E−16 – – n-Octane 2.91E−16 – – n-Nonane 3.29E−16 – – n-Decane 3.63E−16 – – Tetrahydrofuran(THF) 2.13E−16 84.42 2.10 Trichloromethane(TCM) 2.24E−16 0.00 22.68 Dichloromethane(DCM) 1.73E−16 0.00 16.38 Diethylether(DEE) 1.82E−16 80.64 5.88 Acetone(Acet) 1.65E−16 71.40 10.50 Ethylacetate(EtAcet) 1.95E−16 71.82 6.30
Another aspectof relevance to this study wasthe fact that
thespecificcomponentoftheenthalpyofadsorptionofthepolar
probes onthesurface ofPBT A and onthe surfaceof PBT Bis
endothermic.
2. Materialandmethods
2.1. Materials
Twogradesofpoly(butyleneterephthalate),PBTAandPBTB,
werekindlysupplied bySABICInnovativePlastics(formerlyGE
PlasticsEurope),BergenopZoom,TheNetherlands.Theaverage
molecularweight,Mw,molecularweightdistribution,D,
number-averagemolecularweight,Mn,carboxylend-groupconcentration
and hydroxyl end-group concentration of these poly(butylene
terephthalate)sarepresentedinTable1.Thehydroxylend-group
and carboxyl end-group concentrations were determinedusing the –OH absorbance at 3550cm−1 and the –COOHabsorbance at3290cm−1,respectively.ThevaluesindicatedinTable1were obtainedfromthesupplier.Theglasstransitiontemperatureand themeltingtemperatureofbothPBTAandPBTB,are318Kand 503K,respectively(determinedbyDSC).
For the IGC analysis, analytical grade (Sigma–Aldrich Ltd.) probeswereusedwithoutfurtherpurification.Theapolarprobes usedweren-heptane,n-octane,n-nonane,n-decane,andthepolar probestetrahydrofuran(THF,basicprobe),theamphotericprobes acetone (Acet) and diethyl ether (DEE), and the acidic probes chloroform (TCM) and dichloromethane (DCM). In Table 2 are summarisedrelevantpropertiesoftheprobemoleculesmentioned [17–21].Thechemicals usedasprobe moleculeswereobtained fromSigma–AldrichLtd.,Poole,UK.Methane(PhaseSeparations Ltd.,Deeside,UK)wasusedasanon-interactingreferenceprobe andthecarriergasusedwashelium(99.999+%purity,BOCGases Ltd.,Guildford,UK).
2.2. Inversegaschromatography 2.2.1. IGCdataprocessing
Themaindifferencebetweenconventionalgaschromatography (GC)andIGCliesinthefactthatthespeciesofprimaryinterestare notthevolatilecomponentsinjectedbutthematerialactingasthe stationaryphase,typicallyapowder,fibreorfilm.Thismaterialmay bepackeddirectlyintothecolumn,coatedontoasuitablesupport orcoatedontothewallsofthecolumn.Thisallowsthe investi-gationoftheinteractivenatureviathedegreeofinteractionwith well-characterisedvolatileliquids/vapours(“probes”). Quantifica-tionofthisinteractionmaybeachievedbythedeterminationof theretentiontime,tr,foragivenprobe.Inmostuses,thequantity ofprobevapourinjectedintothecarriergasisextremelysmall. Thus,theretentiondatarelatetothethermodynamicinteraction
thatoccursbetweenpolymerandthevapourwhenthepolymeris highlyconcentrated,asinmostpracticalsituations.Furthermore, IGCexperimentsmaybecarriedoutoverappreciabletemperature ranges,sothatthetemperature dependence ofthermodynamic interactionsisnolongerindeterminate.
IGC data processing was carried out according to methods describedintheliterature[15,24,28,32].Theretentiontimesof apo-larprobemoleculesandofpolarprobemoleculesweredetermined atspecifictemperatures,andthevaluesoftheretentionvolume,the energyofadsorption,theenthalpyandentropyofadsorption (dis-persiveandspecificcomponents)oftheprobes,andofthesurface Lewisacidityandbasicityconstants,KaandKb,respectively,were computed.
InIGC,aninertcarriergaselutesaminutequantityofaprobe moleculethroughacolumnthatispackedwiththematerialunder study.Duetotheinteractionsbetweenthetwophases,theprobe moleculesareretainedforacertaintime,tr,whichisusedto cal-culatethenetretentionvolume,Vn,accordingtoEq.(1):
Vn=(tr−t0)F·C·J (1)
Here,Vnistheretentionvolume,t0isthedeadretentiontimeofthe markerprobe,Fisthecarriergasflowrate,Cisacorrectionfactor, allowingforthevapourpressureofthewateratthetemperatureof thebubbleflowmetreusedtodeterminetheflowrate,andJisthe correctionfactorforgascompressibility.Theretentiontimewas determinedusingthegeometrictechniqueoutlinedbyConderand Young[29,30].
Assumingthatexperimentstakeplaceatinfinitedilution,the freeenergyofadsorptionoftheprobeonthestationaryphase sur-facepermole,G,canbedeterminedfromtheretentionvolume, Vn,accordingto:
G=−RTln(Vn)+C1 (2)
Here,Ristheidealgasconstant,Tistheabsolutecolumn tempera-tureandC1isaconstant,whichdependsuponthechromatographic columnandthereferencestate[27].ThroughouttheIGCstudies, thestandarddeviationoftheenergyofadsorptionvaluesofthe probemoleculeswascalculatedasbeingtypicallybelow5%.
Considering that the dispersive and specific components, respectivelyGdandGs,areadditive,assuggestedbyFowkes [31],Eq.(2)canberewrittenas:
Gd+Gs=−RTln(Vn)+C1 (3)
Thefreeenergyofadsorptioncanberelatedtoadhesionwork, Wa,accordingto[27]:
−G=N·a·Wa (4)
Here,NisAvogadro’snumberand“a”thecross-sectionalareaof theprobetobetested(Table2).
Ifnonpolarcomponents(suchasn-alkanes)areused,only dis-persiveinteractionsoccurandtheadhesionworkisgivenby: Wa=2(sdld)
1/2
(5) Here,d
s and ld are,respectively,thedispersivecomponentsof surfacetensionofthesolid(stationaryphase)andofthe probe-molecule.
ReplacingEqs.(4)and(5)inEq.(2)leadsto: 2N(sd)
1/2
a(ld)1/2+C1=RT ln(Vn) (6)
Theslopeofthestraightline,referredtoasthereferenceline, obtainedbyplottingRTln(Vn)versus2aN(ld)1/2,forahomologous n-alkaneseries(Fig.4),leadstothedeterminationofd
sforagiven temperature.
Acid–base characteristics of surfaces were determined by analysingtheinteractionofthepolarprobeswiththesolid sur-faceandquantifyingthedeviationfromthereferenceline,leading totheestimationofthespecificfreeenergy,Gs,as:
−Gs=RTln(V
n)−RTln(Vnref) (7)
Here,Vnref istheretentionvolumeestablishedbythen-alkanes referenceline(Eq.(1)),Vnbeingnowtheretentionvolumeofthe polarprobes.ThiscalculationisalsoillustratedinFig.4.
Theadhesionworkbetweenthepolarprobestestedandthe solidsurface,Ws
a,canbeevaluatedfromthespecificfreeenergy, givenbyEq.(7),as Was= NRT·aln
Vn Vnref (8) Bycarryingoutexperimentsatdifferenttemperatures,itwas possibletodeterminetheenthalpyofadsorptionandtheentropy ofadsorption,respectivelyHandS,fromplotsofG/Tversus 1/T,(Fig.6),accordingtothefollowingequation:G T =
H
T −S (9)
Theacidicandbasicconstants,respectivelyKaandKb,were cal-culatedfromtheplotof−Hs/AN*versusDN/AN*,accordingtoEq. (10)[26](Fig.8).
(−Hs) AN∗ =Ka
DN
AN∗+Kb (10)
Here, AN*and DN are,respectively, theGutmann’s modified acceptoranddonornumbersoftheprobestested[25,29](Table2). 2.2.2. IGCexperimentalset-up
InexperimentalworkinvolvingIGC,conventional GC equip-mentisgenerallyused,withsomeadaptations[15].
TheinstrumentusedwasaFisonsGC9100unit(FisonsScientific EquipmentLtd.,Loughborough,UK),equippedwithaFIDdetector. Themarkerprobeusedwasmethane.Typically,thesyringewas filledwith0.1lofeachprobe,flushed10times,inordertoensure thecreationofaHenry’sinfinitedilutionregion,andinjected man-ually.Theinjector washeatedto150◦Cand theFIDdetectorto 180◦C.Theattenuationwassetto1.Theflowratewascontrolled usinganeedlevalvepressureregulatoranddeterminedusinga bubbleflowmetrethatwasequippedwithaheliumtrap[33]and thermometer.Theinletpressure,Pi,wasmeasuredusingapressure gaugeandtheatmosphericpressure,Po,wasobtainedthroughthe BritishAtmosphericDataCentre(www.badc.rl.ac.uk).TheIGCunit waskeptoncontinuouslyduringtheentirecourseofthework.The temperaturewasincreasedincrementally.
Duetothelackofanappropriatesolvent,insteadofcoatinga supportmaterial,thepolymerwasusedasreceived,after grind-ingandsievingtoachieveanappropriateparticlesize.Tothisend, thepolymerparticleswereprocessedinacryogenicgrinderwhile beingcooledwithliquidnitrogen,followedbysievingthematerial through125mand250mfiltergauzesandthecolumnfilled. Thesievingoperationexcludedfineparticlesthatwouldincrease undesirablythepressuredropinthecolumn.
2.3. Differentialscanningcalorimetry
Theinstrumentusedwasa DSC2010DSCalorimeter. Allthe studieswerecarriedoutunderanitrogenflowrateof200cm3/min. Thetypicalsamplemasswas3–9mg.Fortheheatingmode anal-ysis,thesampleswereheatedatarateof10◦C/min,from50◦C to440◦C.Thenon-isothermal crystallisationbehaviourof tape-extruded blends was studied using the following temperature programme, (i) heating thesample from roomtemperature to
0 5 10 15 20 25 30 35 40 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
303 K
313 K
-Δ
G
a(
kJ
/m
ol)
Carrier
ga
s flo
w ra
te (cm
3/min)
Fig.2.Influenceofthecarriergasflowrateontheenergyofadsorptionofn-decane onthesurfaceofPBTB.
250◦Cat200◦C/min,(ii)keepingthesampleatthistemperature foroneminute,toreleaseallthestresseswithinthematerialand toerasethethermalhistoryand(iii)coolingthesamplefrom250◦C ataconstantrate(8◦C/min)to162◦C.Inpractice,thesamplewas heldat250◦Cfor2minand40s,duetothetemperature equilibra-tionstagethatwaspriortotheisothermalstep.Thenon-isothermal crystallisationpropertiesthatweredeterminedwerethe crystalli-sationtemperature,Tc,andtheenthalpyofcrystallisation,−Hc.
3. Resultsanddiscussion
3.1. Inversegaschromatography
Carriergasflowrates,rangingfrom3cm3/minto35cm3/min, wereusedtoassesstheinfluenceoftheflowrateonthe reten-tiontimesofn-decane,at303Kand313K,onthesurfaceofPBTB. Thispreliminarystudyisanessentialpre-requisiteinsurface ther-modynamiccharacterisationbyIGC,inordertoensurethatsurface adsorption,andnotbulkabsorption,isthedominantphenomenon. TheresultsarepresentedinFig.2.
Fig.2indicatesthatthecarriergasflowratedoesnotinfluence theadsorptionenergyofn-decanetoasignificantextent.Thus,it canbeconcludedthatsurfaceadsorptionisthedominant mecha-nism,andabsorptionoftheprobemoleculesintothebulkofthe polymercanbeneglected,iftemperatureslowerthantheTgare used,fortheflowraterangestudied.Thus,themeasurement tem-peratureusedinthestudyofthesurfacePBTwasvariedbetween 298Kand318K,inincrementsof5K,underacarriergasflowrate of10cm3/min.
Inordertofurtherconfirmthatsurfaceadsorptionisthe gov-erningmechanismofinteractionsbetweenthepolymerandthe probes,andthattheexperimentalconditionschoseninthisstudy (flowrateandtemperature)aresuitableforsurface characterisa-tion,aretentiondiagramwasdeveloped(Fig.3)byplottingLn(Vg) asafunctionof1/T[34].
InFig.3,nothermalchanges,duetothechangeinthe morphol-ogyofthesemi-crystallinepolymerastemperatureincreases,are observed.Thelinearrelationshipsobservedareanindicationofthe establishmentoftheequilibriumbetweentheprobesandthe sur-faceofPBTduetothehomogeneityofthesurface.Itcanthusbe concludedthatsurfaceadsorptionisthegoverningphenomenon.
Fig.3.RetentiondiagramfortheapolarprobesonthesurfaceofPBTB.
1.50E-016 2.00E-016 2.50E-016 3.00E-016 3.50E-016 4.00E-016
-12 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 C7 C8 C9 C10 TCM DCM Acet THF EtAcet Y = -22.56 + 7.98E16X R = 1.00
-ΔG
asRT
Ln(V
g) [kJ
/m
ol]
a
√(γ
l d) [
cm
2(mJ/cm
2)
0.5]
Fig.4.Surfacefreeenergyofadsorptionversusa
dl,forthesurfaceadsorption
ofn-alkanes,andpolarprobes,onPBTA,at295K.
3.1.1. DispersivecomponentofthesurfacefreeenergyofPBTA andofPBTB
Fig.4illustrates thedeterminationof thedispersive compo-nentofthesurfacetensionofPBTA,at295K,accordingtoFowkes’ approach.
Table3summarisesthevaluesdeterminedforthedispersive component of thesurface tension of PBT A and of PBT B. The valuespresentedindicatethatthedispersivecomponentofthe sur-facetensionremainsconstant,withinexperimentalerror,forthe Table3
ValuesofthedispersivecomponentofthesurfacetensionofPBTAandofPBTB.
PBTA PBTB T(K) d s(mJ/m2) R2 sd(mJ/m2) R2 295 43.9±2.1 1.00 42.4±1.0 1.00 303 40.7±1.9 0.98 42.5±2.8 0.99 308 42.8±2.4 1.00 41.4±2.6 1.00 313 40.0±0.6 1.00 40.7±1.3 1.00 318 40.7±2.1 1.00 44.7±9.9 0.92
TCM DCM Acet THF EtAcet 0 2 4 6 8 10 12
RT
L
n
(V
s) (kJ/m
gol)
Probe
PBT
A
PBT B
Fig.5.Comparisonofthespecificcomponentvaluesoftheenergyofadsorptionof thepolarprobesonthesurfaceofPBTAwiththoserelatingtothesurfaceofPBTB, at295K.
temperaturerangestudied,andequalisto41.6±1.7mJ/m2 and
42.3±1.5mJ/m2,forPBTAandPBTB,respectively.Thevaluesof
d
s forPBTAandforPBTBarepracticallyidentical.Thisleadsto
theconclusionthatthedifferentend-groupcomposition(Table1)
doesnotinfluencetoanoticeableextent,thedispersivecomponent ofthesurfacetension.Bearinginmindthatthemajordifference betweenPBTAandPBTBisintheOHend-groupconcentration, thevaluefoundford
s confirmstheobservationsfoundin litera-ture[35]withregard tothefactthatthisfunctionalgroup,asa high-energysite,contributesmainlytotheformationofspecific intermolecularinteractions.Thevaluesfoundareinaccordwith surfacetensionstudiesonPBTbasedoncontactangle measure-ments[23].
3.1.2. AdsorptionofpolarprobesonPBTAandonPBTB InFig.5arecomparedthevaluesof−Gs
aforthepolarprobes onthesurfaceofPBTAwiththoseconcerningthesurfaceofPBTB, at295K.TheretentiontimesforDEEonthesurfaceofPBTAwere verylowand,therefore,pronetolargeexperimentalerrors. Con-sequently,thedeterminationoftheenergyofadsorption,andof thespecificcomponentoftheenthalpyofadsorptionofDEE,was notpossible.Thislowretentiontimeprobablyresultsfromsteric hindrance,thatis,structuralrestrictions,fromboththeadsorbate andtheadsorbent,hinderingthesemoleculesfromspatial con-formationsbecomingeffectiveintermsofspecificintermolecular interactions.Theoxygenatom(Lewisbasiccentre)ofdiethylether (DEE)ismoresusceptibletoshieldingbytheneighbouring hydro-genatoms[36]thanisthatof,e.g.acetone(Acet),whichisreadily accessibleforinteraction.Similarresultshavebeenobservedwhen CCl4wasusedasaprobe[37].
ForbothPBTAandPBT B,theadsorptionoftheLewisacidic probesisasstrongas,orstrongerthan,theadsorptionofLewis basic/Lewisamphotericprobes.Bearinginmindtherelativelylow acidityoftheacidicprobes,whencomparedtothebasicityofthe basicprobes(e.g.THF),itcanbeconcludedthatthesurfacesofboth PBTAandPBTBareamphoteric,beingstronglyLewisbasic.Fig.5 alsoshowsthatthesurfaceofPBTAiscomparabletothatofPBTBin termsofLewisacidity/basicity.Ananalysisofthespecific compo-nentoftheenthalpyofadsorptionwillprovideabetterinsightinto thedifferencesbetweentheLewisacidic/basicpropertiesofthe twoPBTs.Thevaluesoftheenthalpy,andentropy,ofadsorption ofthepolarprobes,alongwiththecorrespondingdispersiveand
Table4
Dispersivecomponentsoftheenthalpyofadsorptionandoftheentropyof adsorp-tion,Hda,andSda,respectively,ofthepolarprobes,onthesurfaceofPBTA.
Probemolecule −Hd a(kJ/mol) Sda(J/molK) R2 TCM 33.0 −127.0 0.91 DCM 26.2 −120.0 0.81 Acet 27.1 −121.0 0.83 THF 31.8 −126.0 0.89 EtAcet 29.7 −124.0 0.87 Table5
Specificcomponentsoftheenthalpyofadsorptionandoftheentropyofadsorption, Hs
aandSda,Ssa,respectively,ofthepolarprobes,onthesurfaceofPBTA. Probemolecule −Hsa(kJ/mol) Ssa(J/molK) R2
TCM −38.4 147.0 0.94
DCM −23.1 138.0 0.82
Acet −20.9 101.0 0.89
THF −23.0 93.7 0.81
EtAcet −17.2 85.8 0.92
specificcomponents,aresummarisedinTables4and5forPBTA,
andtheirdeterminationisillustratedinFig.6forthispolymer.The valuesrelatingtoPBTBhavealreadybeenreportedinaprevious publication[22].
The specific component of the enthalpy of adsorption of polarprobesonthesurfaceof PBT Ais endothermic. This con-firms the reported thermodynamic analysisof the surface and bulk PBT B [22]. Endothermic values of energy of adsorption have alsobeen reported in theliterature for theadsorption of polar probe molecules on titanium dioxide pigments [37,38], on a vinyl acetate-vinyl alcohol copolymer [39], and on 2-(N-morpholino)ethylmethacrylate [40]. A rearrangement of the surfaceuponthechemisorptionofpolarmoleculesisthoughtto occur,thusinitiatinganincreaseintheentropyofthesystem.This rationaleisconfirmedbythepositivevaluesofSs
a,andsupported bytheexperimentallyverifiedphenomenonoftherearrangement ofthesurfacelayersduetochemisorption[41].
The enthalpy of formation of individual hydrogen bonds is alwaysnegative.However,thetotalenthalpy(andentropy)related totheformation of hydrogenbondsis theresult of three con-tributions[42]:apositivecontributionthatresultsfrombreaking hydrogenbondsintheself-associatingpolymer,anegative con-tributionthatresultsfromforminghydrogenbondsbetweenthe self-associatingpolymerandtheadsorbentmolecule,and contrib-utionsfromvanderWaalsanddipoleforces.Whenpolarmolecules
0.0031 0.0032 0.0033 0.0034 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 DCM TCM Acet THF EtAcet -Δ Ga s /T [kJ /(m ol .K)] 1/T [K-1]
Fig.6. Determinationofthespecificcomponentoftheenthalpyandoftheentropy ofadsorption,ofpolarprobesonthesurfaceofPBTA.
TCM DCM Acet THF EtAcet 0 5 10 15 20 25 30 35 40 45 50 55 60
H
s(kJ/mol)
aProbe
PBT A
PBT B
Δ
Fig.7. Comparisonofthespecificcomponentvaluesoftheenthalpyofadsorption ofthepolarprobesonthesurfaceofPBTAwiththoserelatingtothesurfaceofPBT B.
Table6
ValuesofKaandKbdeterminedforthesurfacesofPBTAandPBTB.
Ka Kb R2
PBTA −0.24 −0.97 0.97
PBTB −0.49 −0.96 0.96
adsorbonPBT,thedominantfactoristhepositivecontribution
aris-ingfromthebreakingofhydrogenbondsintheself-associating
polymer.Thisresultsinanendothermicenthalpyofadsorptionand
anincreaseintheentropyofthesystem[22].
Itshouldbenoticed(Tables4and5)thatthedispersive compo-nentoftheenthalpyofadsorption,Hd
a,isexothermic,thatthe correspondingchangeinentropy,Sd
a,negative,andthatthe val-uesaresimilarformostofthepolarprobes.Furthermore,thevalue of−Hd
aincreaseswithincreasinga×
d l.
InthecaseofPBTA,ananalysisofthespecificcomponentof theenthalpyof adsorption, leadstothefollowing ranking (the more negative, the stronger the interaction, as it is endother-mic):TCM>DCM=THF>Acet>EtAcet.ThisindicatesthatPBTAis amphotericandKboughttobehigh,inlinetheanalysisofthe spe-cificcomponentoftheenergyofadsorptionofthepolarprobes. Thechangeinentropyuponadsorptionisgreaterforthoseprobes whoseadsorptionischaracterisedbythegreaterspecific compo-nentoftheenthalpyofadsorption.Theanalysisof−Hs
aforPBT Aisalsoinlinewiththeanalysisoftheenergyofadsorption,and withtheresultsrelatingtothesurfaceandbulkPBTB[22].
Fig.7givesa comparisonof thevaluesofHs
a of thepolar probesonthesurfaceofPBTAwiththoseconcerningthesurface ofPBTB.
Fig.7suggeststhatthesurfaceofPBTAshouldbelessLewis acidicandlessLewisbasic,thanthatofPBTB.Thedetermination ofKaandofKbprovidesamorecompletedescriptionofthesurface Lewisacidity/basicity.
3.1.3. DeterminationofKaandKbforthesurfaceofPBTAandfor thesurfaceofPBTB
ThedeterminationofthesurfaceLewisacidityconstant,Ka,and ofthesurfaceLewisbasicityconstant,Kb,isillustratedinFig.8for PBTA.ThevaluesobtainedforbothPBTAandPBTBaresummarised inTable6. -30 -25 -20 -15 -10 -5 0 0 5 10 15 20 25 30 35 40 45 50 DCM THF Acet EtAcet TCM TCM
DN/AN*
Y = -0.97 - 0.24X R2 = 0.97-Δ
H
a s/AN
*
Fig.8.DeterminationoftheKaandoftheKbofthesurfaceofPBTA.
ItshouldbenotedthattheKaandKbvaluesarenegativedueto theendothermicadsorptionoftheprobemolecules.Thesurfacesof PBTAandPBTBareconcludedtobeLewisamphotericandstrongly Lewisbasic,inlinewiththeanalysisoftheenergyandenthalpyof adsorptionofpolarprobes.ThevaluesdeterminedforKaandKb areconsistentwiththeanalysisofthestructureofthismolecule (Fig.1).Thus,theLewisacidicsiteshavetheirorigininthehydrogen atomsofthe–O–CH2–segments,andinthehydrogenatomsofthe hydroxylend-groupand ofthecarboxylicend-group.TheLewis basicsiteshavetheiroriginsintheestermoiety,andintheoxygen atomsofthecarboxylicend-groupandofthehydroxylend-group [43].
The poly(butyleneterephthalate), PBT A, is significantly less LewisacidicthanPBTBandasLewisbasicasPBTB.Thiscanbe rationalisedonthebasisofananalysisofthecarboxylend-group andhydroxylend-groupconcentrationsinbothPBTAandPBTB, asdeterminedbyFTIR(Table1).Thehydroxylend-group concen-trationofPBTAislowerthanthatofPBTB,whichisreflectedin theloweracidiccharacterofPBTA.Thecarboxylend-group con-centrationdifferencesbetweenPBTAandPBTBdonotinfluence significantlytheLewisbasicity,asquantifiedbyKb.Thisobservation arisesfromthedominantcontributionoftheesterfunctionalityto thesurfacebasicity,alongsidewiththecontributionofthebasic oxygenatominthehydroxylgroups.ThefactthattheKbvaluefor thetwoPBTsdoesnotfullycorrelatewiththevalueofthespecific componentoftheenthalpyofadsorptionofDCMandofTCMmay berelatedtotherelativelylowvalueofAN*forDCMandTCM,when comparedtothevalueofDNfortheremainingpolarprobes.This resultstresses(i)theimportanceofusingasmanypolarprobesas possibleinthedeterminationofKaandKb,and(ii)theneedfora systematicandholisticanalysisofGs
a,−Hsa,KaandKb,in con-trasttoanindividualanalysisofthesethermodynamicparameters (asisusuallythecaseintheIGCliterature).
3.2. StudyofthecrystallisationpropertiesofPBTAandPBTB ThevaluesofKaandKbforbothPBTAandPBTBindicatethat theintramolecularinteractionsandtheintermolecularinteractions betweenPBTmoleculesarestrong.Thismoleculeischaracterised bythenon-existenceofbulkyside-groups,bythechemically regu-larstructure,andbyhighmobility(duetothebutyleneunitinthe chain).Thus,boththethermodynamicandthestructural require-mentsforcrystallinityexistinthecaseofPBTmolecules.
0 20 40 60 80 100 450 455 460 465 470 475 480
T
c(K)
Weight fraction of PBT A (%)
Fig.9.Non-isothermalcrystallisationtemperatureversusPBTAloadinginsamples ofPBTgranulate.
Toassesstheinfluenceoftheend-grouptypeand concentra-tion,and of molecularweightdifferences onthecrystallisation propertiesof PBT, a seriesof PBT A plusPBT Bgranulate sam-pleswaspreparedwithincreasingamountsofPBTA.Thevalues determinedforthecoldcrystallisationtemperature,Tc,andforthe enthalpyofcrystallisation,−Hc,ofthesesamplesarepresented inFigs.9and10,respectively.
ThecrystallinitydegreeofPBTAiscalculatedas56%andthat ofPBTBas65%(consideringthevalueof142J/gforthe−Hcof 100%crystallinePBT).Theextentofcrystallisationdecreases lin-earlywithincreasingamountofPBTAinthePBTgranulatesample. ThegreaterextentofcrystallinityofPBTBisinterpreted,inview ofthevalues determinedforKaand Kb,tobea consequenceof greaterstrengthofspecificLewisacid/baseintermolecular inter-actionsthan in PBT A. Thisis in accordance withthe fact that highenergy sitesin polymericmaterials(asassessedby IGCat infinitedilution)arewellknowntoactasnucleating spotsthat initiatethecrystallisationprocess[44,45].Thedifferencesinthe averagemolecularweightofPBTAandPBTBalsoplayaroleinthe crystallisationpropertiesofthispolymer.Lowermolecularweight valueswouldbeexpectedtoleadtogreatervaluesofcrystallinity degree.However,inthisstudyitwasnotpossibletodistinguishthe
0 20 40 60 80 100 76 78 80 82 84 86 88 90 92 94
-H
c(J/g)
Weight fraction of PBT A (%)
Δ
Fig.10.EnthalpyofcrystallisationversusPBTAloadinginsamplesofPBTgranulate.
relativecontributionofmolecularweightdifferencesandof end-grouptypeandconcentrationdifferences.Thisisduetothefact thatthegreater–OHend-groupconcentrationisdirectlyrelated withtheloweraveragemolecularweightofthispolymer.
Fig.9alsoshows,veryinterestingly,thatthenon-isothermal crystallisationtemperatureisnotsignificantlyinfluencedbythe weightfractionofPBT Ainthesample.Thus,differencesinthe end-grouptypeandconcentrationandintheaveragemolecular weightofthesePBTsdonotcausechangesinthecrystallisation activationenergy.InviewofthevaluesfoundforKaandKb,one canarguethatthecrystallisationactivationenergyis,inthiscase, dominantlydefined bytheenergyof theLewisbasicsites.This observationderivesfromthesignificantlygreatervalueofKbwhen comparedtothevalueofKa,forbothPBTAandPBTB,andfromthe similarvalueofKbforthesegradesofPBT.
4. Conclusions
TheabilityofPBTtointeractthroughdispersiveintermolecular forcesisnotinfluencedbytheend-grouptypeandconcentration orbythemolecularweight.Theenthalpyofadsorptionofpolar moleculesonPBTwasconfirmedtobeendothermic. Thisresult wasinterpretedasthedominantcontributiontotheenthalpyand entropyof adsorptionof breakinghydrogenbondsin the poly-mer,comparedwiththecontributionofforminghydrogenbonds betweenthepolarprobemoleculesandthePBTmolecules.
IGC characterisation of the surface Lewis acidic/basic prop-erties ofPBT A and of PBT Bcorrelateswell withthecarboxyl end-groupconcentrationandhydroxylend-groupconcentration, asdeterminedbyFTIR.Furthermore,theanalysisofthesurface LewisacidityandsurfaceLewisbasicity,asquantifiedbyKaand KbanddeterminedbyIGC,alongsidewiththeanalysisofthe phys-icalstructureofthesepolymers,isusefulintheinterpretationof thecrystallinityand,thus,oftheexcellentsolventresistanceand thermalstabilityofPBT.
Thedifferencesintheend-grouptypeandconcentration,and intheaveragemolecularweight,betweenPBTAandPBTB,donot causedifferencesinthecrystallisationactivationenergy,andthus, inthecrystallisationtemperature.Thisevidenceisinterpretedin termsofdominantcontributionoftheLewisbasicsitestothe crys-tallisationactivationenergy.Theextentofcrystallisationislower forPBTAthanitisforPBTB.ThisisaconsequenceofweakerLewis acid/baseintramolecularandintermolecular interactionsand of greatermolecularweightinthecaseofPBTA.
Acknowledgements
Theauthorswishtoacknowledgethesupportandthe collabora-tionofSABICInnovativePlastics,BergenopZoom,TheNetherlands, inthisproject.
References
[1]V.M.Nadkarni,J.P.Jog,Crystallizationbehaviourinpolymerblends,in:L.A. Utracki (Ed.),Two-Phase Polymer Systems, Munich, Germany, 1991, pp. 213–239.
[2]C.C. Huang,F.-C. Chang,Reactivecompatibilizationofpolymerblends of poly(butyleneterephthalate)(PBT)andpolyamide-6,6(PA66):1.Rheological andthermalproperties,Polymer38(1997)2135–2141.
[3]S.P.Mishra,P.Venkidusamy,StructuralandthermalbehaviorofPC/PBTblends, J.Appl.Polym.Sci.58(1995)2229–2234.
[4]Y.-Y.Cheng,M.Brillhart,P.Cebe,M.Capel,X-rayscatteringandthermal anal-ysisstudyoftheeffectsofmolecularweightonphasestructureinblendsof poly(butyleneterephthalate)withpolycarbonate,J.Polym.Sci.B:Polym.Phys. 34(1996)2953–2965.
[5]I. Pillin, S. Pimbert, J.F. Feller, G. Levesque, Crystallization kinetics of poly(butyleneterephthalate)(PBT):influenceofadditivesandfreecarboxylic acidchainends,Polym.Eng.Sci.41(2001)178–191.
[6]G.Pompe,L.Haubler,W.Winter,Investigationsoftheequilibriummelting temperatureinPBTandPC/PBTblends,J.Polym.Sci.B:Polym.Phys.34(1996) 211–219.
[7]G.Pompe,InfluenceoftransesterificationonthethermalpropertiesofPC/PBT blend-miscibilityofPCandPBT,RecentRes.Dev.Polym.Sci.1(1997)109–147.
[8]G.Pompe,E.Meyer,H.Komber,H.Hamann,InfluenceofPBTcrystallizationon miscibilitydegreeofamorphousphaseinPC/PBTmeltblends,Thermochim. Acta187(1991)185–200.
[9]M.E.J.Dekkers,S.Y.Hobbs,I.Bruker,V.H.Watkins,Migrationofpolymerblend componentsduringmeltcompounding,Polym.Eng.Sci.30(1990)1628–1632.
[10]M.Ratzsch,G.Haudel,G.Pompe,E.Meyer,Interactionbetweenpolymers,J. Macromol.Sci.AChem.A27(1990)1631–1655.
[11]H.Bertilsson,B.Franzen,J.Kubat,AgeingofPC/PBTblends.II:Phaseseparation effects,PlasticsRubberProcessAppl.10(1988)145–153.
[12]A.N.Wilkinson,S.B.Tattum,A.J.Ryan,Melting,reactionandrecrystallizationin areactivePC-PBTblend,Polymer38(1997)1923–1928.
[13]G.Pompe,L.Haubler,InvestigationsoftransesterificationinPC/PBTmeltblends andtheproofofimmiscibilityofPCandPBTatcompletelysuppressed trans-esterification,J.Polym.Sci.BPolym.Phys.35(1997)2161–2168.
[14]D.G. Hamilton, R.R. Gallucci, The effects of molecular weight on polycarbonate–polybutylene terephthalate blends, J. Appl.Polym. Sci. 48 (1993)2249–2252.
[15]J.M.R.C.A.Santos,J.T.Guthrie,Analysisofinteractionsinmulticomponent poly-mericsystems:thekey-roleofinversegaschromatography,Mater.Sci.Eng.R: Rep.50(2005)79–107.
[16]A.Voelkel,Physicochemicalmeasurements(inversegaschromatography),in: C.Poole(Ed.),GasChromatography,Elsevier,Amsterdam(TheNetherlands), 2012,pp.477–494.
[17]A. Voelkel,B. Strzemiecka,K.Adamska, K.Milczewska, Inversegas chro-matographyasasourceofphysiochemicaldata,J.Chromatogr.A1216(2009) 1551–1566.
[18]A.Voelkel,K.Adamska,Propertiesofmaterialsasdeterminedbyinversegas chromatography,Ann.UMCSChem.64(2009)169–183.
[19]L.H.Lee,Relevanceofthedensity-functionaltheorytoacid-baseinteractions andadhesioninsolids,in:K.L.Mittal,J.Anderson(Eds.),Acid–BaseInteractions: RelevancetoAdhesionScienceandTechnology,VSP,Utrecht,TheNetherlands, 1991,pp.25–46.
[20]P.Mukhopadhyay,H.P.Schreiber,Inversegaschromatographyforpolymer surface characterization above andbelow Tg, Macromolecules 26(1993) 6391–6396.
[21]J. Schultz,L.Lavielle,Interfacialpropertiesofcarbonfiber-epoxy, in:D.R. Lloyd,T.C.Ward,H.P.Schreiber(Eds.),InverseGasChromatography, Char-acterization of Polymers and Other Materials, A.C.S., Washington, 1989, pp.185–202.
[22]J.M.R.C.A.Santos,K.Fagelman,J.T.Guthrie,CharacterisationofthesurfaceLewis acid–basepropertiesofpoly(butyleneterephthalate)byinversegas chro-matography,J.Chromatogr.A969(2002)111–118.
[23]S.Y.Hobbs,M.E.J.Dekkers,V.H.Watkins,Effectofinterfacialforcesonpolymer blendmorphologies,Polymer29(1988)1598–1602.
[24]S.Mohammadi-Jam,K.E.Waters,Inversegaschromatographyapplications:a review,Adv.ColloidInterfaceSci.212(2014)21–44.
[25]D.P.Kamdem,S.K.Bose,P.Luner,Inversegaschromatographycharacterization ofbirchwoodmeal,Langmuir9(1993)3039–3044.
[26]H. Chtourou, B. Riedl, B.V. Kokta, Surface characterizations of modified polyethylenepulpandwoodpulpsfibersusingXPSandinversegas chromatog-raphy,J.AdhesionSci.Technol.9(1995)551–574.
[27]J.M.R.C.A.Santos,M.H.Gil,A.Portugal,J.T.Guthrie,Characterisationofthe sur-faceofacellulosicmulti-purposeofficepaperbyinversegaschromatography, Cellulose8(2001)217–224.
[28]J.M.R.C.A.Santos,K.Fagelman,J.T.Guthrie,Characterisationofthesurface Lewisacid–basepropertiesofthecomponentsofpigmented,impact-modified, bisphenolApolycarbonate–poly(butyleneterephthalate)blendsbyinversegas chromatography–phaseseparationandphasepreferences,J.Chromatogr.A 969(2002)119–132.
[29]D.P.Kamdem,B.Riedl,Inversegaschromatographyoflignocellulosicfibers coatedwithathermosettingpolymer:useofpeakmaximumandconderand youngmethods,J.ColloidInterfaceSci.150(1992)507–516.
[30]P.L.Jackson,M.B.Huglin,A.Cervenka,Useofinversegaschromatography toquantifyinteractionsinaminecuredepoxyresins,Polym.Int.35(1994) 119–133.
[31]F.M.Fowkes,J.A.Manson,T.B.Lloyd,D.O.Tischler,B.A.Shah,Inversegas chro-matographyin:D.M.Mattox(Ed.),Mater.Res.Soc.Symp.,vol.119,1988,pp. 223–234.
[32]J.Riddle,F.M.Fowkes,Spectralshiftsinacid–base chemistry.1.Vander Waalscontributions toacceptor numbers, J. Am.Chem. Soc.112 (1990) 3259–3264.
[33]A.E.Bolvari,etal.,ExperimentaltechniquesinIGC,in:D.R.Lloyd,T.C.Ward,H.P. Schreiber(Eds.),InverseGasChromatography,CharacterizationofPolymers andOtherMaterials,A.C.S.,Washington,1989,pp.12–19.
[34]M.Kunaver,etal.,Inversegaschromatography—adifferentapproachto char-acterizationofsolidsandliquids,ActaChim.Slov.51(2004)373–394.
[35]W.Zhang,A.I.Leonov,IGCstudyoffiller–fillerandfiller–rubberinteractionsin silica-filledcompounds,J.Appl.Polym.Sci.81(2000)2517–2530.
[36]A.C.Tiburcio,J.A.Manson,Acid–baseinteractionsinfillercharacterizationby inversegaschromatography,J.Appl.Polym.Sci.42(1991)427–438.
[37]C.R.Hegedus, I.L.Kamel, Thermodynamic analysis ofpigmentand poly-mersurfacesusinginversegaschromatography,J.Coat.Technol.65(1993) 31–43.
[38]Y.J.Lee,I.Manas-Zloczower,D.L.Feke,Characterizationofsolventinteractions withtreatedtitaniumdioxidepowders,PowderTechnol.73(1992)139–146.
[39]F. Chen, Study of acceptor–donor interactions at the polymer interface byinverse gaschromatography dataanalysis,Macromolecules 21(1988) 1640–1643.
[40]A.Askın,D.T.Yazıcı,Astudyofthesurfaceanalysisofsomewater-soluble polymersbyinverse gaschromatography,Surf.InterfaceAnal.40 (2008) 1237–1241.
[41]V.Ponec,Z.Knor,S.Cerny,AdsorptiononSolids,Butterworths,London,1974.
[42]M.M.Coleman,J.F.Graf,P.C.Painter,SpecificInteractionsandtheMiscibilityof PolymerBlends,TechnomicPub.Co.,Lancaster,PA,U.S.A.,1991.
[43]P.L.Huyskens,W.A.P.Luck,T.Zeegers-Huyskens,IntermolecularForces,An IntroductiontoModernMethodsandResults,Springer-Verlag,London,1991.
[44]M.Soleimani,L.Tabil,S.Panigrahi,I.Oguocha,Crystallizationandthermal propertiesofbiofiber-polypropylenecomposites,in:El-SonbatiAdel(Ed.), Thermoplastic—CompositeMaterials,InTech,Croatia,2012,pp.131–146.
[45]D.Page,T.Gopakumar,Propertiesandcrystallizationofmaleated polypropy-lene/graphiteflakenanocomposites,Polym.J.38(2006)920–929.