Highly purified mineral oils are used in several pharmaceutical, foods and cosmetics applications. A fast and simple method was developed for the analysis of the total level of residual mineral oil aromatic hydrocarbons (MOAH)in these oils and in the intermediate oils that were sampled during the purification process.
Trang 1jou 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
a University of Amsterdam, Van ‘t Hoff Institute for Molecular Sciences, Analytical Chemistry-Group, P.O Box 94157, 1090 GD, Amsterdam, the Netherlands
b Unilever Research and Development, P.O Box 114, 3130 AC Vlaardingen, the Netherlands
a r t i c l e i n f o
Article history:
Received 26 September 2018
Received in revised form 4 December 2018
Accepted 7 January 2019
Available online 8 January 2019
Keywords:
Mineral oil analysis
Mineral oil aromatic hydrocarbons
MOAH quantification
Gas chromatography
Vacuum ultraviolet detector
a b s t r a c t
(http://creativecommons.org/licenses/by/4.0/)
1 Introduction
Highlyrefinedandpurifiedmineraloilfractionsarewidelyused
forelaborationofconsumerproductssuchasfoodsand
cosmet-ics.Thesemineraloils(MO)arecalled‘whiteoils’andmustmeet
verystrictcriteriaintermsofresiduallevelsofmineraloil
aro-matichydrocarbons(MOAH)[1 Inadditiontothedeliberateuse
ofthesepurifiedoilsasingredientsforcosmeticsorfoods,mineral
oilcanalsofinditswayintoconsumerproductsasacontaminant,
e.g.throughmigrationfromrecycledpackagingmaterialsorfrom
theenvironment[2 Thesecontaminantscaneasilycontainupto
30%andsometimesevenupto50%ofMOAHandtheir
concen-trationinconsumerproductsrangesfrom10to100mg/kg[3,4
ReliablemethodsformonitoringthetotalMOAHlevelsinMOare
neededfortheoptimizationofMOAHremovalprocessesinwhite
oilproduction, aswellasfordetectingMOAHcontaminationin
qualityassessmentofconsumerproducts.Ideallythesemethods
∗ Corresponding author.
E-mail address: a.r.garciacicourel@uva.nl (A.R García-Cicourel).
wouldalsoprovideinformationonthecompositionoftheMOAH fractionasthismightaffectbothremovalefficiencyandtoxicityof theMOAH[5
Biedermannetal.developedafullyautomatedon-linenormal phaseliquidchromatography(NPLC)–gaschromatography-flame ionizationdetection(GC-FID)methodwithasilicacolumnforthe quantitativeanalysisoflowlevelsofMOAHinfoodsandcosmetics
Inthismethod,theNPLCstepwasusedtoperformapreseparation
ofmineraloilsaturatedhydrocarbons(MOSH)andMOAH.Sensitive anduniversalquantificationofthetwofractionswasthenprovided
bytheFID[6,7 TheBiedermannandGrobmethodisnowwidely used.However,becauseinmanylaboratoriestheinstrumentation requiredisnotavailable,alsolessautomatedapproacheshavebeen developed.AmethodusingSolidPhaseExtraction(SPE)forthe pre-separationofMOSHandMOAHwasdevelopedbyMoretetal.in
2011.ToimprovetheMOSH/MOAHseparation,thesilicasorbent wasreplacedbysilver-loadedsilicamakingthedeterminationof thecutpointbetweenMOSHandMOAHless critical[8,9 Still, forsampleswithMOAHcontentsbelow1%interferencesofMOSH readilyoccur.ForsuchsamplestheendoftheMOSHbandmight overlapwiththestartofthelaterelutingMOAHfraction.Because https://doi.org/10.1016/j.chroma.2019.01.015
0021-9673/© 2019 The Authors Published by Elsevier B.V This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ ).
Trang 2114 A.R García-Cicourel, H.-G Janssen / J Chromatogr A 1590 (2019) 113–120
theGC-FIDquantificationstepprovidesnoadditionalselectivity,
tracesofMOSHcollectedintheMOAHfractionresultinincorrect
MOAHlevels.Theneedforthiscarefulprefractionationmakesthe
methodtime-consuminginroutineanalysisandfaster methods
withlesssamplehandlingstepsaredesired
Recently a novel detector for GC has been introduced, the
VacuumUltraviolet (VUV)detector.This detectormeasuresthe
absorbanceof gas phase compounds in thefar UV wavelength
range from 120to 430nm [10] Essentially almost every
com-poundabsorbsstronglyinthisrangeofwavelengths[11].Indeed
theuniversalnatureofthedetectorhasbeendemonstratedwith
theanalysisof awide varietyofcompoundsin different
matri-ces[12–17].TheVUVdetectorhasattractedsignificantattention
for petrochemical analysis due to its ability to provide
group-typeinformation.In2016,Groegeretal.foundspectraldifferences
betweenaliphaticandaromaticscompoundsindieselfuelanalysis
[18].Atlowwavelengthsallcompounds,aliphaticandaromatic,
aredetected,whereasathigherwavelengthsonlythearomatics
andunsaturatedcompoundsabsorb.Thisisanextremely
attrac-tivefeatureinMOSH/MOAHanalysis.Ifsufficientlyselective,this
wouldmeantheGC-VUVcombinationwouldbeabletoquantify
MOAHdirectly,i.e.withoutpriorseparationofMOSHandMOAH
Thiswouldbehighlyadvantageousintermsofanalysiscostand
methodruggedness Moreover,thechromatogramsand spectra
willcontaininformationonthenumberofaromaticringsorthe
lengthandtypeofalkyl-moietiespresent,whichcouldberelevant
informationfore.g.optimizationoftheMOAHremovalprocesses
inwhiteoilproduction
InthisworkanovelandrapidGC-VUVmethodforthe
deter-mination of the aromatics content in mineral oil samples and
intermediatesfromthewhiteoilproductionprocessisdescribed
QuantitativedataobtainedusingthenewdirectGC-VUVmethod
arecomparedwiththosefromstandard,off-lineSPE-GCmethods
Furthermore,theabilityoftheset-upforprovidingstructural
infor-mationonthemoleculesinthearomaticfractionsfromtheirVUV
spectraisstudied
2 Materials and methods
Hexaneanddichloromethane (DCM),both HPLCgrade,were
purchased from Biosolve BV (Valkenswaard, The Netherlands)
n-undecane (C11), bicyclohexyl (Cycy), n-tridecane (C13), 5-
␣-cholestane (Cho), n-hexyl benzene (6B), 1-methyl naphthalene
(1-MN),biphenyl(BP),1,3,5-tri-tert-butylbenzene(TBB)and
n-nonylbenzene(9B) werefromSigma-Aldrich(Zwijndrecht, the
Netherlands) Silver-nitrate impregnated silica gel (230 mesh)
loaded at approximately 10wt.% was purchased from
Sigma-Aldrich.Mineraloilsamplesfromdifferentstagesinthepurification
processwereobtainedfromalocalsupplierofwhiteoils.The
vis-cositiesoftheoils rangedfrom10.32to153.4cStat 40◦C The
carbonnumbersofthecompoundsrangedfrom15tomorethan
50
2.1 Mineraloilandstandardsolutions
Stocksolutionsofeachmineraloil(at500mg/mL)andofthe
mixtureofstandardcompounds (C11,C13,Cycy,Cho,6B,1-MN,
BP, TBBand 9Bat 1mg/mL percompound), which areused as
markersfortheMOSHandMOAHseparation,toevaluatelosses
ofvolatilecompounds and forquantification, wereprepared in
hexane For theoff-line argentation normal phase liquid
chro-matography–gaschromatography -flame ionizationdetection
(AgNPLC-GC-FID)analyses, samplesfor injectionwereobtained
bycombining200Lofthemineraloilstocksolutionand300L
ofthestandardmixtureandfillingthevialto1mLusinghexane
(finalconcentrationsof100and0.3mg/mL,respectively).Forthe off-lineSPE-GC-FID,a50mg/mLsolutionofeachmineraloilwas preparedfromthestocksolutionand50Lofthissolutionwere combinedwith6Lofthestandardsolutionand444Lofhexane (5and0.012mg/mL,respectively).FortheGC-VUVanalysis, min-eraloilsolutionsof5mg/mLinhexanewerepreparedwithoutthe standardcompounds
2.2 AgNPLCfractionation
LCfractionation ofthe sampleswasconducted ona Waters Alliance 2695 LC instrument with a Waters 996 DAD detector (Waters,Etten-Leur,TheNetherlands).Themethodwasbasedon that described byBiedermann andGrob [19] withsome modi-ficationstoincreasetheseparationgapbetweentheMOSHand MOAHfractions.Insteadofa250×2mmIDcolumnpackedwith silica,twoseriallyconnected100mmx4.6mmIDx5mAgNO3 loadedsilicacolumns(Agilent,Amstelveen,TheNetherlands)were used.TheMOSH/MOAHseparationwasperformedusinga gradi-entstartingwithhexaneheldfor12min,andthenprogrammed
to30%DCMin1min,maintainingthiscompositionuntiltheend
oftherun(25min)ataflowrateof0.3mL/min.Usingthese chro-matographicconditions,theMOSHfractionelutesbetween8and 10.5min,andMOAHfrom13minuntiltheendoftherun.Both frac-tionswereconcentratedto0.5mLpriortoGCanalysis.Toavoid shiftsin theelutionwindow duetocolumn contamination,the column wasreconditioned after every run with100% DCM for
5minat0.5mL/min,followedbya5minrinsewithhexane,also
at0.5mL/min.20Lofthemineraloilsamplewasinjectedinto thesystem
2.3 SPEfractionation EmptyglassSPEcartridges(Sigma-Aldrich)werepackedwith 0.5gofsilverimpregnatedsilicagel.Conditioningwasperformed
byheatingthecartridgesat120◦Cfor2h,washingwith10mLof DCMandwith4mLofhexane.Thevolumeofsampleappliedtothe silver-silicaSPEwas0.5mL.ElutionoftheMOSHwasachievedwith
5mLofhexane.TheMOAHfractionwaselutedwith8mLof hex-ane/DCM(50:50v/v).Bothfractionswereconcentratedto0.5mL priortoGCanalysis
2.4 GC-FID
GCanalysisoftheMOSHandMOAHfractionsobtainedfromSPE
orLCwascarriedoutusinganAgilent6890NCinstrument sys-temwithaFocus-PALautosampler(GLSciences,Eindhoven,The Netherlands).The sample(1L) wasinjected in splitless mode (2minsplitless time)at 350◦C in a 4mmID linerpackedwith glasswool.Thecapillarycolumn,15mx0.32mmx0.1m
DB5-HT(Agilent),wasusedataconstantflowof2mL/minusinghelium
ascarriergas.Thetemperatureprogramranfrom60◦C(3min)to
350◦C(3min)at15◦C/min.TheFIDtemperaturewassetat350◦C andthedatacollectionratewas200Hz
2.5 GC-VUV
AVGA-101VUVdetector(VUVAnalytics,CedarPark,TX,USA) wasconnectedtoanAgilentG1530AGCsystemequippedwithan Optic3injectorandaFocus-PALautosampler(GLSciences).The chromatographicconditionswereidenticaltothosefortheGC-FID analyses.Thetemperatureofthetransferlineandtheflowcellof theVUVdetectorweresetat350◦C.Nitrogenwasusedasmake-up gasatapressureof0.35psi.Thedatacollectionratewas100Hz
Trang 3Fig 1.GC-VUV absorbance spectra of a) C 13 , b) 6B and c) 1-MN.
3 Results and discussion
3.1 VUVspectroscopicdifferencesofMOSHandMOAH
TheVUVspectraofcompoundscandiffersignificantly
depend-ingonthedifferentfunctionalgroups present inthemolecules
[10].Fig.1showsafewrepresentativeGC-VUVabsorbance
spec-traofanumberofaliphaticandaromaticcompoundscontaining
thetypicalstructuralelementsofMOSH/MOAHcompounds.Clear
differencesareobservedbetweenthedifferentspectra.Forthe
sat-uratedcompound,representingtheMOSHfractionofthemineral
oil,thespectrummonotonouslydecreasesmovingfromthelowest
wavelengthof125nmtolongerwavelengths.Forthesecompounds
noabsorbanceisobservedatwavelengthsabove180nm(Fig.1a)
Thearomaticcompoundsshowadifferentbehaviour(Fig.1 and
c).Forthesecompoundsoneormoredistinctmaximaoccur.The
exactlocationsofthesemaximaandtheirbandwidthdependon
thenumberofaromaticringsandthewaythesearelinked[11]
Moleculeswithoneringshowanabsorbancemaximumaround
185nm Thismaximum shifts tohigher wavelengths whenthe
numberofringsinthemoleculeincreases.ForMOAH,whichare
basicallyalkylatedaromaticcompoundswithoneormoreisolated
orfusedaromaticrings,theabsorbancespectrumwillconsistoftwo
regions,oneadsorptionbandcharacteristicforthearomaticpart,
andamonotonouslydecreasingpartoriginatingfromthealiphatic
substituents
TostudythespectraldifferencesbetweenrealMOSHandMOAH
fractions,eightdifferentmineraloilsampleswereseparatedusing
SPE Fig.2shows theoverall averagedVUV absorbance spectra
obtainedfor the MOSH and MOAHfractions from these8 oils
Sincetheabsorbancespectraofallthemineraloilsanalyzeddid
notpresentanysignificantresponseatwavelengthshigherthan
240nm,onlydatauptothisvaluearepresented.Becausethe
lev-elsof MOSH andMOAH in thesamplesweredifferent, spectra
werenormalizedtoalloweasiercomparison,i.e.intensitiesata
specificwavelengthweresetequal.Normalizationwasperformed
at125nmforMOSH,andat125nmor195nmforMOAH.From thecomparisonoftheFigs.1and2itisclearthattheMOSHand MOAHfractionsshowsimilarabsorbancespectraasthe individ-ualmodel molecules Thealiphatic fractionsshow noresponse above180nmandtheshapeofthespectrafortheeightMOSH fractionsisverysimilar(Fig.2a).Whencomparingthespectraof thearomaticfractionsfromtheeightdifferentmineraloils,two maindifferencesbecomenoticeable.Inthespectranormalizedat
125nm,forthecomparisonofthealiphaticzone(125–180nm), dif-ferencesoccurinthe160–180nmregion.Normalizationat195nm, forthecomparisonofthearomaticzone,emphasizesthe differ-encesintheregionof190–240nm.Thesetwodifferencesamong theeightMOAHfractionsarerelatedtothestructuraldifferences
inthemoleculespresentinthedifferentfractions.Thefirstregion, reflectingthealiphaticsubstituentsofthearomaticrings,differs becauseofdifferencesinthestructure(linear,cyclicorbranched) andlengthofthealkylsubstituents.Thesecondregionwhere dif-ferencesoccur,i.e.the190–240nmregion,reflectsdifferencesin the number and interconnectivity of the aromatic ringsin the molecules.Mono-,di-orpolyaromaticcompoundswithfusedrings
orringsclosely togetherallhave (slightly)differentabsorbance maxima
3.2 StructuralinformationofMOAH
In the previous section, clear differences were observed betweentheVUVspectraofMOAHfractionsfromdifferent min-eraloils.Inasecondseriesofexperiments,changesinthespectra werestudiedovertheGCelutionwindowofaMOAHfraction.This wasdonebyexaminingspectracollectedatdifferenttimesinthe chromatographicrun.Thespectraobtainedareshownin Fig.3 Theseabsorbancespectracanprovidestructuralinformation on thecompoundsformingthearomaticfraction.Themaindifferences seenarerelatedtothealkylpartoftheMOAH.Inthespectra col-lectedatlongerretentiontimes,i.e.fortheheavieranalytes,the aliphaticresponse(125–180nm)increases.Thisisanindicationof
Trang 4116 A.R García-Cicourel, H.-G Janssen / J Chromatogr A 1590 (2019) 113–120
Fig 2.GC-VUV absorbance spectra of a) MOSH normalized at 125 nm, b) MOAH normalized at 125 nm and c) MOAH normalized at 195 nm from eight different mineral oils The noisy nature of the spectra (below 140 nm) is probably caused by variations in the pressure of the nitrogen in the flow cell.
Fig 3.Absorbance spectra of MOAH fractions from four different mineral oil Each spectrum is the average of a 1 min slice of the chromatogram For mineral oil A the hump starts at 9.29 min (≈C 17 ) and finishes at 24.02 min (≈C 52 ), for mineral oil E from 10.4 min (≈C 18 ) to 23.21 min (≈C 48 ), for mineral oil H from 7.46 min (≈C 14 ) to 16.77 min (≈C 31 ) and for mineral oil C from 13.66 min (≈C 24 ) to 21.69 min (≈C 44 ).
Trang 5Fig 4. GC-VUV absorbance spectra of different one-ring aromatic compounds.
thepresenceoflongerchainsorahigherdegreeofsubstitutionon
thearomaticcoreforthelaterelutinganalytes.Furthermore,for
thesamplesA,EandCdifferencesarealsoseeninthearomatic
part.Thespectraobtainedatlongerretentiontimesshowhigher
responsesinthe190–240nmregionthantheonesobtainedatthe
beginningofthechromatogram.Thiscanbeanindicationofthe
presenceofalkylatedpolyaromaticcompounds
InordertounderstandhowthestructureoftheMOAH
com-poundsaffects theabsorbancespectrum,VUVlibraryspectraof
differentaromaticcompoundswerecompared.Fig.4ashowsthe
spectraofone-ringaromaticspecieswithincreasingstraightalkyl
chainlength.Asteeperslopeinthealiphaticregionisrelatedto
theincrementofthealkylchainlength.Thearomaticregionisnot
affectedbythesubstituentlength.Ontheotherhand,Fig.4 shows
thespectraldifferencesoffourone-ring isomerswithone alkyl
chainwithdifferentlevelsofbranching.Smallchangesareseenin
thealiphaticpartofthespectrumdependingonthebranching
posi-tionsinthealkylchain.Branchedalkylsubstituentsgiveaslightly
lowersignalinthealiphaticpartofthespectrumcomparedtothe
straightchainmolecules.Ifthemethylgroupsareclosertothe
aro-maticcore,thealiphaticsignalintensityisreduced.Finally,Fig.4c
showsthespectrumofone-ringisomerswithincreasingnumbers
ofalkylchainsattachedtothering.Fortheseisomersthearomatic
partisaffectedmorethanthealiphaticpart.Forthecompounds
withmorethanonealkylsubstituenttheabsorbancemaximumof
thearomaticpartofthespectrumshiftstohigherwavelengths.In
otherwords,thelocationofthemaximumcouldbeanindicator
forthedegreeofsubstitutionofthearomaticcompounds
How-ever,it isimportanttoemphasizethatthepresenceofmultiple
aromaticringsinthestructurewillalsoshifttheabsorbance
maxi-mumofthearomaticstohigherwavelengths,asisnoticedinFig.4d
Theabsorbancespectraofbenzene,naphthaleneandanthracene
demonstratethatanincreaseinthenumberofaromaticringsshifts
theabsorbancemaximumtolongerwavelengths[11].Differences
arealsoobservedamongmoleculeswiththesamenumberof
aro-maticrings.Naphthalene,biphenyland1-methylnaphthaleneall
containtworings;however,theirabsorbancespectraaredifferent Alkylsubstitutionsaffecttheabsorbancespectruminthesameway
aswiththemono-aromatics.1-MNshowsashiftoftheabsorbance maximumtolongerwavelengthsandhasabroadersignalinthe aromaticzonecomparedtonaphthalene.Ontheotherhand,the biphenylabsorbancespectrumismoresimilartothoseobtained forone-ringspecieswithanabsorbancemaximumat195nm Basedontheinformationdiscussedabove,itispossibletodraw somepreliminaryconclusionsonthetypeofMOAHspeciespresent
inthedifferentoilsamplesanalyzedinFig.3.Fromthespectraitcan
beconcludedthatthearomaticcoresofthedifferentmineraloils arelargelyidentical.Fromtheratiosofthearomaticandaliphatic intensities(aromatic/aliphatic),itcanbeconcludedthatmineraloil
A(0.820-0.426)containsthelongestalkylchainsassubstituents attachedtothearomaticcore,whilemineraloilH(0.938–0.563) hastheshortestchains.Thisstatementisconfirmedbytheviscosity valuesofthemineraloils(91.21and10.32cStat40◦C,respectively) FormineraloilA,EandC,anabsorbanceincreaseisobservedinthe regionofthetwo-ringaromaticcompounds,205–240nm,atlonger
GCretentiontimes.Thisincrementismoreremarkableformineral oilEindicatingahighercontentofpolyaromaticcompoundsinthis mineraloil.Onthecontrary,basedonthespectra,mineraloilH seemstohavethelowestcontentofpolyaromaticcompounds Fig 5 shows a comparison of spectraobtained at thesame retentiontime(15min)forfourmineraloilsamplesfrom differ-entstagesin thepurificationprocess.Becausethespectrawere recordedatthesameGCretentiontime,theybelongtocompounds withsimilarvolatility,andthereforesimilarmolecularweight.The aliphaticresponseishighestformineraloilsAandC.This indi-catesthepresenceofmultiplestraight-chainalkylsubstituentsin theseoilsascomparedtothemineralsoilHandE.Basedonthe aliphaticresponse,mineraloilEseemstocontainmorebranched alkylchains.Furthermore,thearomaticregionforthismineraloil showsashiftintheabsorbancemaximumtohigherwavelengths andtheabsorbancebandiswiderthanfortheotherthreesamples Theseareadditionalindicationsforahigherdegreeofsubstitution
Trang 6118 A.R García-Cicourel, H.-G Janssen / J Chromatogr A 1590 (2019) 113–120
Fig 5. GC-VUV absorbance spectra of the MOAH fraction of four mineral oils The
spectra were recorded at the same GC retention time.
Fig 6. Chromatogram of mineral oil H analysed on the GC-VUV system without
pre-vious prefractionation The total mineral oil response was obtained using the entire
wavelength range (125–240 nm), whereas the aromatic response was obtained with
a 190–240 nm filter The solvent peak (hexane) in both wavelength filters elutes at
2.5 min The signal in the aromatic filter is either due to impurities in the solvent or
limited detector selectivity The sharp peak at 14.5 min is an impurity.
Apparently,sampleEhasmorebutshorteralkylchainsonthe
aro-maticringthantheoilsA,CandH.Regardingthearomaticregion
thiscorroboratestheinformationpreviouslymentionedaboutthe
type ofaromatic species for each mineraloiland confirms the
addedvalueofVUVdetection.Thecontentofpolyaromaticspecies
ishigherformineraloilAandE,andlowforoilH.However,for
accuratelyestimatingthepercentagesofthesecompoundsinthe
mineraloilotheranalyticalmethodsareneededlike
comprehen-siveLCxGC-MS[20]
3.3 DirectdeterminationoftheMOAHcontentbyGC-VUV
Thepresenceofabsorbancebandsspecificforaromaticsinthe
VUVspectrainprincipleenablesdirectquantitativeassessmentof
MOAHlevelsinmineraloilswithouttheneedforaMOSH/MOAH
preseparation.ToobtaintheMOAHcontentfromtheGC-VUVdata,
twospectral filters wereapplied.Thefirstfiltercoversthe190
to240nm region and detects only the MOAH species (Fig 6)
Thesecondfiltercoverstheentirerangeofwavelengthsrecorded
(125–240nm)andgivesa(semi-)quantitativenumberforthetotal
GCresponseofthemineraloil(MOSH+MOAH).Thissecondfilter
isappliedtobeabletocorrectthedatafordifferencesintotal
min-eraloilresponsesduetosamplediscriminationintheGCinjection
whichwasevaluatedcomparingtheresponseofdifferentmineral
oilsatthesameconcentration
TobeabletocalculatetheMOAHpercentagedirectlyfromthe
GC-VUVchromatogram,theMOSHamountfirstneedstobe
cal-culated.Thiscanbedonefromthepeakareainthe125–240nm
region,correctedforthecontributionofthealiphaticpartsofthe
MOAHtothisarea.Agoodestimateofthecorrectionfactorcan
beobtained fromthe MOAH spectraof the eight differentoils showninFig.2.Theseoilsrepresentthetypicalmineraloilsand intermediatesencounteredinindustry.Inthesespectra,thetotal areais1.38±0.11timeshigherthantheareaintheMOAHregion (190–240nm).Thus,thefollowingequationcanbederived:
BecauseVUVspectroscopyfollowstheadditivityprinciplesof theBeer-Lambertlaw,theMOSHresponsecanbecalculatedfrom: MOSH125 −240=TR125 −240−MOAH125 −240 (2)
IntheseequationsMOSH125-240,MOAH125-240andTR125-240are thepeakareasinthechromatogramoftheMOSH,MOAHandtotal mineraloil,respectively,allcalculatedfromthe125–240nm sig-nal.MOAH190-240 isthecorrespondingareacalculatedfromthe 190–240nmsignal.Theaboveareascanbeconvertedintomass percentagesofMOSHandMOAHinthemineraloilusingtheEqs (3)and(4):
MOAH % = 100 x MOAH125−240 × RRFMOAH
(MOSH125−240× RRFMOSH) + (MOAH125−240× RRFMOAH) (3)
HereRRFsaretherelativeresponsefactorsforMOSHandMOAH relativetomethane.RRFsvaluesforindividualalkaneswere deter-minedinpreviousstudiesanditwasfoundthattheyareremarkably independentofthelengthandstructureofthealkylchain[21].In thiswork,anRRFMOSHof0.775wasusedbasedonpreviousdata [11].ForMOAH,anRRFMOAHof0.425±0.055wascalculated.This valuewasobtainedasanaverageoftheRRFvaluesfortheeight differentMOAHfractionsdescribedpreviously.EachRRFwas cal-culatedfollowingEq.(5),whereAMOSH/AMOAHistheratioofthe areasofthetwofractions:
RRFMOAH= MMMOAH
MOSH
AMOSH(125 −240 nm)
AMOAH(125 −240 nm)RRFMOSH (5)
OncetheseRRFareestablished,theycanbeusedforthedirect GC-VUV MOAH quantitation in unknown mineral oil samples withoutthenecessityofanyMOSH/MOAHpreseparation,i.e.the laboriousLCorSPEstepcanbeavoided.Thisisreflectedinthe anal-ysistime.Whilefortheconventionalmethodsaruntakes46min, theproposedGC-VUVmethodisonly25min
ToevaluatetheperformanceofthedirectGC-VUVmethod,some performancecharacteristicssuchasselectivity,linearity, repeata-bility,limitofdetection(LOD) andlimitof quantification(LOQ) wereassessedusingthetwowavelengthsrangesdescribedabove andfollowingtheEurachemguideformethodvalidation[22].This evaluationwasdoneforthreedifferentmineraloils.Onemineral oilwasvirtually freeof aromatics (mineral oilN),whereasthe othertwocontainedsignificantMOAHlevels(mineraloilAand H).Alsotheboilingpointrangeofthethreemineraloilswas dif-ferent(Fig.7).Theperformancecharacteristicsofthenewmethod areshowninTable1.Theselectivitywasevaluatedbycomparing thepeakareasofequalinjectedamountsofMOSHandMOAH, iso-latedbySPE,intheMOAHregion(i.e.195–240nmregion).Equal amountsofMOSHandMOAHresultedinanapproximately3300 timeshigherresponseforMOAHthanforMOSH.Or,phrased differ-ently,foraMOAHfreeoilourdirectGC-VUVmethodwouldreport
aMOAHlevelof 0.03%.AgoodlinearityofVUV absorbance(R2
>0.989)versusconcentrationwasseenatbothwavelengthranges TheLODandLOQvaluesatthecurrentGCinjectionsettingsare similarformineraloilNandH,bothhavingarathernarrowboiling pointrange.SlightlypoorervalueswereobtainedformineraloilA duetothebroadervolatilityrange,andhencebroader chromato-graphic‘hump’ofthisoil.TheLODandLOQvaluesfortheMOAH fractionwerecalculatedbasedontheconcentrationofMOAHin mineraloilAandH(25and30%respectively).Thus,consideringa
Trang 7Table 1
Performance characteristics for three different mineral oil samples MO 2 is aromatic free while MO A and H contain MOSH and MOAH.
Total mineral oil 125 – 240 nm MOAH 190 – 240 nm
Fig 7.GC-VUV chromatograms of the three mineral oils selected for the method
validation The chromatograms were obtained using the entire wavelength range
(125 to 240 nm) and at the same concentration 10 mg/mL The boiling point range
of the mineral oils is different Mineral oil N ranges from C 14 to C 40 (centered around
C 23 ), mineral oil A ranges from C 17 to C 52 (centered around C 33 ) and mineral oil H
ranges from C 14 to C 31 (centered around C 23 ).
sufficientlysensitiveforallbutthehighestpuritysamples.Finally,
therepeatabilityofthemethodresultedinRSDvaluesbetterthan
8%.Unlikeinthestandardmethods,on-lineoroff-lineSPE-GC-FID
orLC-GC-FID,thesevaluesareonlyrelatedtotheintegration
diffi-cultiesanddeviationsinthesampleintroductionbecausenoother
samplepreparationstepsarepresent
3.4 Analysisofdifferentsamples
Todeterminethepracticalapplicationlimitsofthenewly
pro-posedrapidGC-VUVmethodforMOAHanalysis,aseriesofstarting
samplesand intermediatesof white oilproduction of different
origins and differentMOAH levelswere analysed Fig.8 shows
theMOAHcontentobtainedfor18differentmineraloilsamples
usingGC-VUV.Datafromthestandardmethods(SPE-GC-FIDand
LC-GC-FID)areincludedtoprovideaframeofreference.The
GC-VUVresultsarenotstatisticallysignificantlydifferentfromthose
obtainedbythetwostandardmethods,whichwasevaluatedby
applyingananalysisofvariance(ANOVA)toeachsample.Atthis
point,itisimportanttomentionthattherepeatabilityoftheMOAH
measurementsis much betterfor theGC-VUVmethodas
com-paredtothestandardmethods.Thisisduetotheabsenceofthe
complexmanualsamplepreseparationstep.Thesamplesanalyzed
inthecomparisonhaveMOAHcontents fromlessthan0.13%to
almost50%.Theirsuccessfulanalysisclearlyshowsthe
applicabil-ityofthemethodtooilswithbothhighandlowaromaticscontent
andfractionsfromdifferentstagesinthepurificationprocess
Min-eraloilJshowsthelimitationofthestandardmethods.Thismineral
Fig 8. Comparison of the aromatic percentage using GC-VUV and the two standard methods The GC-VUV method presented the smallest variation due to the absence
of pre-separation.
oilisfreeofaromatics,buteventhoughahighlyselective silver-loadedsilicastationaryphasewasused,stillsomeMOSHeluted
inthetimewindowwheretheMOAHwouldelute.Thisproblem
isrelatedtothecomplexityofthesample.SampleJcontainsvery highmolecularweightcompoundsand,additionally,hasalow con-tentofaromatics.Becauseofthepresenceofthepoorlysoluble highmolecularweightspeciesthereisanincreasedriskofMOSH compoundselutingintheMOAHfraction.Becauseinthenextstep
a non-selectiveflameionizationdetectoris usedfor quantifica-tion,thisimmediatelyresultsinanoverestimationoftheMOAH level.ThisisnotanissueintheGC-VUVanalysisbecauseno pre-separationisrequired
4 Conclusions
AdirectGC-VUVmethodforthedeterminationoftheMOAH contentofpurifiedmineraloilsampleswasdeveloped.Themethod showedtobeagoodalternativefordeterminingthearomatics con-tentinmineraloilsamples.Detectionlimitswerearound0.13%, makingthemethodsufficientlysensitivefor allbutthehighest puritysamples.TheVUVspectraprovidesomestructural informa-tionofthemoleculespresentinthearomaticsfraction.Thanksto theselectivityprovidedbytheVUV detectornosample presep-arationisneeded,whichshortenstheanalysistimesignificantly andeliminatesdifficultexperimentalsetupsand/ormanual sam-plehandlingsteps.Additionally,thenewmethodreducestherisks forerrors,particularlywhencomplexsamples,i.e.sampleswith
alowcontentofaromaticsandahighmolecularweight,haveto
beanalyzed.Thisnewmethodcouldpotentiallyalsobeusedfor analysisofmineraloilinotherconsumerproductssuchasfood However,sample preparation for reducing matrix interferences frome.g.triglyceridesandolefinswillstillbeneeded
Trang 8120 A.R García-Cicourel, H.-G Janssen / J Chromatogr A 1590 (2019) 113–120
Acknowledgements
TheauthorsthanktheConsejoNacionaldeCienciayTecnología
deMéxico(CONACyT)forthescholarshipawardedtoA.R
García-Cicourel(No.410740)
Thisresearchdidnotreceiveanyspecificgrantfromfunding
agenciesinthepublic,commercial,ornot-for-profitsectors.H.-G
JanssenwasawardedanacademicresearchgrantbyVUVAnalytics,
IncformethoddevelopmentforMOSHandMOAHanalysisinfood
products
References
[1] M Niederer, T Stebler, K Grob, Mineral oil and synthetic hydrocarbons in
cosmetic lip products, Int J Cosmet Sci 38 (2016) 194–200, http://dx.doi.org/
10.1111/ics.12276
[2] S Moret, K Grob, L.S Conte, On-line high-performance liquid
chromatography-solvent evaporation - high-performance liquid
chromatography - capillary gas chromatography - flame ionisation detection
for the analysis of mineral oil polyaromatic hydrocarbons in fatty foods, J.
Chromatogr A 750 (1996) 361–368,
http://dx.doi.org/10.1016/0021-9673(96)00453-0
[3] R Lorenzini, K Fiselier, M Biedermann, M Barbanera, I Braschi, K Grob,
Saturated and aromatic mineral oil hydrocarbons from paperboard food
packaging: estimation of long-term migration from contents in the
paperboard and data on boxes from the market, Food Addit Contam - Part A
Chem Anal Control Expo Risk Assess 27 (2010) 1765–1774, http://dx.doi.
org/10.1080/19440049.2010.517568
[4] K Fiselier, D Fiorini, K Grob, Activated aluminum oxide selectively retaining
long chain n-alkanes: Part II Integration into an on-line high performance
liquid chromatography-liquid chromatography-gas chromatography-flame
ionization detection method to remove plant paraffins for the determination
of mineral oil paraffins in foods and environmental samples, Anal Chim Acta
634 (2009) 102–109, http://dx.doi.org/10.1016/j.aca.2008.12.011
[5] M Biedermann, K Grob, Comprehensive two-dimensional GC after HPLC
preseparation for the characterization of aromatic hydrocarbons of mineral
oil origin in contaminated sunflower oil, J Sep Sci 32 (2009) 3726–3737,
http://dx.doi.org/10.1002/jssc.200900366
[6] M Biedermann, K Fiselier, K Grob, Aromatic hydrocarbons of mineral oil
origin in foods: method for determining the total concentration and first
results, J Agric Food Chem 57 (2009) 8711–8721, http://dx.doi.org/10.1021/
jf901375e
[7] M Biedermann, C Munoz, K Grob, Update of on-line coupled liquid
chromatography – gas chromatography for the analysis of mineral oil
hydrocarbons in foods and cosmetics, J Chromatogr A 1521 (2017) 140–149,
http://dx.doi.org/10.1016/j.chroma.2017.09.028
[8] S Moret, L Barp, G Purcaro, L.S Conte, Rapid and sensitive solid phase
extraction-large volume injection-gas chromatography for the analysis of
mineral oil saturated and aromatic hydrocarbons in cardboard and dried
foods, J Chromatogr A 1243 (2012) 1–5, http://dx.doi.org/10.1016/j.chroma.
2012.04.040
[9] S Moret, L Barp, K Grob, L.S Conte, Optimised off-line SPE-GC-FID method
for the determination of mineral oil saturated hydrocarbons (MOSH) in
vegetable oils, Food Chem 129 (2011) 1898–1903, http://dx.doi.org/10.1016/
j.foodchem.2011.05.140
[10] K.A Schug, I Sawicki, D.D Carlton, H Fan, H.M McNair, J.P Nimmo, P Kroll, J Smuts, P Walsh, D Harrison, Vacuum ultraviolet detector for gas
chromatography, Anal Chem 86 (2014) 8329–8335, http://dx.doi.org/10 1021/ac5018343
[11] B.M Weber, P Walsh, J.J Harynuk, Determination of hydrocarbon group-type
of diesel fuels by gas chromatography with vacuum ultraviolet detection, Anal Chem 88 (2016) 5809–5817, http://dx.doi.org/10.1021/acs.analchem 6b00383
[12] I.C Santos, K.A Schug, Recent advances and applications of gas chromatography vacuum ultraviolet spectroscopy, J Sep Sci 40 (2017) 138–151, http://dx.doi.org/10.1002/jssc.201601023
[13] C Qiu, J Cochran, J Smuts, P Walsh, K.A Schug, Gas chromatography-vacuum ultraviolet detection for classification and speciation of polychlorinated biphenyls in industrial mixtures, J Chromatogr A 1490 (2017) 191–200,
http://dx.doi.org/10.1016/j.chroma.2017.02.031 [14] L Skultety, P Frycak, C Qiu, J Smuts, L Shear-Laude, K Lemr, J.X Mao, P Kroll, K.A Schug, A Szewczak, C Vaught, I Lurie, V Havlicek, Resolution of isomeric new designer stimulants using gas chromatography – vacuum ultraviolet spectroscopy and theoretical computations, Anal Chim Acta 971 (2017) 55–67, http://dx.doi.org/10.1016/j.aca.2017.03.023
[15] M Zoccali, K.A Schug, P Walsh, J Smuts, L Mondello, Flow-modulated comprehensive two-dimensional gas chromatography combined with a vacuum ultraviolet detector for the analysis of complex mixtures, J Chromatogr A 1497 (2017) 135–143, http://dx.doi.org/10.1016/j.chroma 2017.03.073
[16] I.C Santos, J Smuts, K.A Schug, Rapid profiling and authentication of vanilla extracts using gas chromatography-vacuum ultraviolet spectroscopy, Food Anal Methods 10 (2017) 4068–4078, http://dx.doi.org/10.1007/s12161-017-0976-1
[17] J Schenk, G Nagy, N.L.B Pohl, A Leghissa, J Smuts, K.A Schug, Identification and deconvolution of carbohydrates with gas chromatography-vacuum ultraviolet spectroscopy, J Chromatogr A 1513 (2017) 210–221, http://dx.doi org/10.1016/j.chroma.2017.07.052
[18] T Gröger, B Gruber, D Harrison, M Saraji-Bozorgzad, M Mthembu, A Sutherland, R Zimmermann, A vacuum ultraviolet absorption array spectrometer as a selective detector for comprehensive two-dimensional gas chromatography: concept and first results, Anal Chem 88 (2016) 3031–3039,
http://dx.doi.org/10.1021/acs.analchem.5b02472 [19] M Biedermann, K Grob, On-line coupled high performance liquid chromatography-gas chromatography for the analysis of contamination by mineral oil Part 1: method of analysis, J Chromatogr A 1255 (2012) 56–75,
http://dx.doi.org/10.1016/j.chroma.2012.05.095 [20] S De Koning, H.G Janssen, U.A.T.T Brinkman, Group-type characterisation of mineral oil samples by two-dimensional comprehensive normal-phase liquid chromatography-gas chromatography with time-of-flight mass spectrometric detection, J Chromatogr A 1058 (2004) 217–221, http://dx.doi.org/10.1016/j chroma.2004.08.083
[21] P Walsh, M Garbalena, K.A Schug, Rapid analysis and time interval deconvolution for comprehensive fuel compound group classification and speciation using gas chromatography-vacuum ultraviolet spectroscopy, Anal Chem 88 (2016) 11130–11138, http://dx.doi.org/10.1021/acs.analchem 6b03226
[22] Eurachem, Eurachem Guide: The Fitness for Purpose of Analytical Methods –
A Laboratory Guide to Method Validation and Related Topics, 2014, doi:978-91-87461-59-0.
[23] Bundesinstitut für Risikobewertung, Highly Refined Mineral Oils in Cosmetics: Health Risks Are Not to Be Expected According to Current Knowledge, 2018, http://dx.doi.org/10.17590/20180702-124741-0