Supercritical CO2 extraction of Tetraclinis articulata: chemical composition, antioxidant activity and mathematical modeling

HAL Id: hal-00881074

https://hal.archives-ouvertes.fr/hal-00881074

Submitted on 7 Nov 2013

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of

sci-entific research documents, whether they are

pub-lished or not. The documents may come from

teaching and research institutions in France or

abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est

destinée au dépôt et à la diffusion de documents

scientifiques de niveau recherche, publiés ou non,

émanant des établissements d’enseignement et de

recherche français ou étrangers, des laboratoires

publics ou privés.

Supercritical CO2 extraction of Tetraclinis articulata:

chemical composition, antioxidant activity and

mathematical modeling

Nejia Herzi, Séverine Camy, Jalloul Bouajila, Philippe Destrac, Mehrez

Romdhane, Jean-Stéphane Condoret

To cite this version:

Nejia Herzi, Séverine Camy, Jalloul Bouajila, Philippe Destrac, Mehrez Romdhane, et al..

Super-critical CO2 extraction of Tetraclinis articulata: chemical composition, antioxidant activity and

mathematical modeling.

Journal of Supercritical Fluids, Elsevier, 2013, vol.

82, pp.

72-82.

To link to this article

: DOI:10.1016/j.supflu.2013.06.007

http://dx.doi.org/10.1016/j.supflu.2013.06.007

This is an author-deposited version published in:

http://oatao.univ-toulouse.fr/

Eprints ID: 9928

To cite this version:

Herzi, Nejia and Camy, Séverine and Bouajila, Jalloul and Destrac,

Philippe and Romdhane, Mehrez and Condoret, Jean-Stéphane

Supercritical CO2 extraction of Tetraclinis articulata: Chemical

composition, antioxidant activity and mathematical modeling. (2013) The

Journal of Supercritical Fluids, vol. 82. pp. 72-82.

O

pen

A

rchive

T

oulouse

A

rchive

O

uverte (

OATAO

)

OATAO is an open access repository that collects the work of Toulouse researchers and

makes it freely available over the web where possible.

Any correspondence concerning this service should be sent to the repository

administrator:

staff-oatao@listes-diff.inp-toulouse.fr

Supercritical

CO

2

extraction

of

Tetraclinis

articulata:

Chemical

composition,

antioxidant

activity

and

mathematical

modeling

Nejia

Herzi

_,

_Séverine

_Camy

_,

_Jalloul

_Bouajila

_,

_Philippe

_Destrac

_,

Mehrez

Romdhane

_,

_{Jean-Stéphane}

_Condoret

a_{UnitéderechercheMACS,ENIG,RoutedeMédenine,6029Gabès,Tunisia}

b_{UniversitédeToulouse,INPT,UPS,LaboratoiredeGénieChimiqueUMRCNRS5503,4,AlléeEmileMonso,F-31030Toulouse,France}

c_{UniversitédeToulouse,LaboratoiredesInteractionsMoléculairesetRéactivitéChimiqueetPhotochimiqueUMRCNRS5623,UniversitéPaul-Sabatier,118}

routedeNarbonne,F-31062Toulouse,France

Keywords:

Supercriticalcarbondioxideextraction Tetraclinisarticulata

Essentialoil Antioxidantactivity Modeling

a

b

s

t

r

a

c

t

OperatingconditionsforextractionfromtheleavesofTetraclinisarticulatausingsupercriticalcarbon diox-ide(SCCO2)werestudiedtofocusonthefeasibilityofobtainingvolatileandnonvolatilefractionsthrough

theuseofdifferentextractionpressures(90,280and1000bar).Inaddition,influenceoftemperature, staticpretreatmentanddynamicextractiondurations,particlesizeandCO2flowratewereinvestigated.

AllextractswereanalyzedbyGC–FID/MSandtheirantioxidantactivitywasmeasuredusingABTS•+_and

DPPH•_{methods.Conventionalhydrodistillation(HD)wasalsoperformedforcomparison.AthighCO} 2

pressure(280and1000bar),theamountofphenolicsintheextractswashigher(respectively102.03 and267.90GAEmg/g)thanforHDandsupercriticalfluidextraction(SFE)at90bar(respectively8.89and 9.70GAEmg/g).Correlatively,highantioxidantactivitywasfoundforhighpressureSFE.Surprisingly,for extractsobtainedbySFEat90bar,despiteverylowphenoliccontent,significantantioxidantactivitywas observed,whileessentialoilobtainedbyHD,whichpresentedalsolowphenoliccontent,exhibitedlow antioxidantactivity.

Physicalaspectswereonlyinvestigatedforthelowpressuresupercriticalextraction(90bar)process. Qualitativeassessmentofkineticcurvestogetherwiththeirmodelingrevealedthattheextraction pro-cesswasmainlylimitedbythethermodynamicequilibriumofeasilyaccessiblesolutesbutwhereaxial dispersionwassignificant.Fromthisresultasimpleextrapolationprocedurewasproposed.

1. Introduction

Inthelast20years,researchuponnaturalsubstancessuchas aromaticandmedicinalplantshasconcentratedontwoprimary areas:identificationoftheirbiologicalactivity(suchasantioxidant activity)anddevelopmentofnovelextractionmethods insuring betterqualityandeco-friendlyoperation.Inthiswork,the Tetra-clinisarticulataspeciewillbeinvestigatedinrespecttoboththese aspects.

T.articulatabelongstothefamilyCupressaceaeandisanative plantofthesouth-westernMediterranean,mainlyNorthAfrica.Itis animportantelementoftheMaghrebvegetation(Morocco, Alge-riaandTunisia).ThisplantisawidelydistributedtreeinNorth Africawhereitcoversatotalareaof10,000km2 _{[1].T.articulata}

∗ Correspondingauthor.Tel.:+33534323713. ∗∗ Correspondingauthor.Tel:+33562256885.

E-mailaddresses:severine.camy@ensiacet.fr(S.Camy), jalloul.bouajila@univ-tlse3.fr(J.Bouajila).

isknownasahighqualityhealthfoodwhichiscommonlyusedin Tunisiantraditionalmedicine.Ithasbeenfoundtohave pharmaco-logicaleffects,includingantioxidantactivitythatmayexplainthe benefitsintreatingcirculatorydisordersperceivedfromtheuse ofthisherboverthecenturies.Itismainlyusedagainstchildhood fevers,respiratoryandintestinalinfections,stomachpain,diabetes andhypertension[2].Therefore,itisinterestingtofindaneffective methodforisolationofbioactivecompoundsfromT.articulata.

Itisknownthattheisolationofnaturalantioxidantsisdifficult becausemostofthemareheat-sensitiveandvolatileinsteam,and canbethereforedestroyedwhenanonadaptedextractionprocess is used [3]. Traditionally, extraction of bioactive compounds fromplantsisperformedbysteamdistillation,hydrodistillation (HD)ororganicsolventextraction[4,5].Usingthesetechniques, long extraction time, toxic solvent residues, labor-intensive operation and degradation of thermo-sensitive compounds are usuallyobserved.Suchdisadvantagescanbeavoidedbyusingthe supercriticalcarbondioxide(SCCO2)technique[6,7].Indeed,this

technique,moregenerallyreferredassupercriticalfluidextraction (SFE), should improve the volatile oil quality by avoiding any

thermal stresstothevolatilecomponent. Thesolventpowerof SCCO2 isdirectlyrelated toits density.So, choiceof thevalues

of operating pressure and temperatureis a critical step in the development of an efficient process. These operating variables determinethevalueofthedensityandconsequentlytheextract yieldand composition, and,therefore, thefunctionalproperties [8,9] of theextract. Also, other factorssuch asduration ofthe staticanddynamicperiodsofextraction,particlesizeandsolvent flowratecaninfluencetheprocessperformance[10],andmustbe investigated.

Intheliterature,toourknowledge,veryfewstudieswere per-formedonT.articulata.Asanexample,Rachedetal.[11]presented resultsobtainedfromaplantfromAlgeriaharvestedinApril2008. But,up tonow, nostudy ontheSFE extractionof T. articulata fromTunisiahasbeenreportedintheliterature.However,very numerousworkshavestudiedtheextractionofvolatileoilfrom otherplantsusingSCCO2[6,7,12,13].Forexample,extractionofthe

volatileoilfromSalviaofficinalishasbeenstudiedbyseveralauthors (Reverchon et al.[14],Aleksovski and Sovová[15],Glisic et al. [16])whousedthesameextractionconditionstorecovervolatile oils(pressuresbetween80and130bar,temperaturesbetween40 and60◦_{Candsolventflowratesbetween5and30g/min}

approxi-mately).OnlythemethodproposedbyReverchonetal.[14]differed fromthefactthattheycarriedouttheseparationoftheextracts usingtwoseparatorsatdifferentconditions,whichyieldedabsence ofwaxesinthecollectedvolatileextract.

Theobjectiveofthepresentworkwastoevaluatetheextracting capacityofSCCO2toobtainT.articulatavolatileoilwithpossibly

differentpropertiesandtoinvestigatetheeffectoftheextraction conditions.Itissuspectedthatthevolatilefractioncouldbe selec-tivelyisolatedatlowpressureextraction,wherebythepressures between100and1000barwouldbemorefavorablefor obtain-ingfractionscontainingheaviercompoundswithhighantioxidant characteristics.Tochecktheseassumptions,extractionswere per-formedatpressures90,280and1000barandtemperaturesequal to40and60◦_C.CO

2flowratewasvariedbetween5and20g/min,

anddifferentmeanparticlesizessoasdifferentextraction dura-tionsweretested.Inthisworkcomparisonwithextractsobtained byHDwasalsoproposed.

On account of importance of SFE for industrial applica-tion,extrapolationof laboratoryexperimentaldata isnecessary and requires extensive knowledge of the physical phenomena occurringduringthecourseoftheextraction.Thentheuseof math-ematicalmodelsallowsgivingtheextrapolationrules.Especially, forextrapolationpurposes,itisofgreatinteresttoidentifywhether theextractionprocessislimitedbymasstransferphenomenaor bythermodynamics.Suchinformationcanbeobtainedfromthe studyofextractionkineticsandtheirmodelingusingmathematical modelsproposedintheliterature[17].

2. Materialsandmethods 2.1. Plantmaterial

LeavesfromcultivatedplantsofT.articulatawerecollectedby handduringOctober2010(inthemorning)fromthesame loca-tion,Korbos,locatedintheregionofCapBonat60kilometersfrom Tunis(latitudes36.82/36◦₄₉′₄₄′′_{andlongitudes10.59/10}◦₃₅′₃₆′′_),

Tunisia.Harvestedmaterialwasdriedintheair,protectedagainst directsunlight,untilaconstantweightwasachieved.Theaverage particlesizeofT.articulata,wasobtainedaftersievingofgrinded leaves(usinglaboratoryknifegrinder).Themoisturecontentofthe air-driedplantmaterial,determinedbytheKarlFischer volumet-rictitration,was9%(w/w).Theefficiencyofthedifferentextraction processesisassessedusingtheextractionyield,definedasthemass

ofcollectedextractdividedbyinitialmassofwetplantmaterial(9%, w/wmoisturecontent).

2.2. Chemicals

Allchemicalswereofanalyticalreagentgrade.Allreagentswere purchased from Sigma–Aldrich–Fluka (Saint-Quentin, France). Commercialcarbondioxide(99.99%purity)waspurchasedfrom AirLiquide(Bordeaux,France) fortheextractionofvolatileand nonvolatilecomponentsbySFEprocess.

2.3. Supercriticalfluidextraction(SFE)

Supercriticalextractionswerecarriedoutusingtwodifferent devices.ThefirstonetermedSF200extractionpilotunit,from Sepa-rexChimiefine,(France)isdescribedelsewhere[18,19].Briefly,this apparatusiscomposedofa113mL(16cmheightand3cminternal diameter)tubularextractor(300barmax)withacascadeofthree 15-mLcyclonicseparatorsconnectedtotheextractoroutlet.

Inthepresentstudy,the113mLstainlesssteelextractorwas loadedwith50gofdriedT.articulataleaveswithanaverageparticle sizeof1.5±0.3mm.Thesystemwasoperatedatatemperature between40and60◦_{Candpressureof90barforvolatilefraction}

extraction.PureCO2 waspassedintothecellwiththeflowrate

keptbetween5and20g/mininallruns.

Thesecondapparatus,termedA21(SFE1000bar),from Sepa-rexChimiefine,France,iscomposedofa63mLtubularextractor whichcanbeoperatedupto1000barand200◦_C.Onlyone

sepa-ratorisconnectedtotheextractoroutlet.Pressureintheextractor isadjustedbyabackpressureregulator.Inthiswork,theextractor wasloadedwith28gofdriedT.articulataleaveswithanaverage particlesizeof1.5±0.3mm.Thesystemwasoperatedat280and 1000bar,CO2flowrateof20g/minandtwotemperatures,60and

40◦_C.

Densityofthesolidphasehasbeenexperimentallydetermined andisequaltos=815kg/m3.Thebedvoidfractionwasestimated

byε=1−a/s=0.4575wheresolidapparentdensityisobtained

witha=m0/Vextr(a=442kg/m3)wherem0istheinitialmassof

plantmaterial. 2.4. Hydrodistillation

AconventionalmethodofHDwascarriedouttocomparethe extractionperformanceswithSFE.Thisset-upwascomposedofa 2-Lboiler,acondenserandameasuringtubewithastopcock.Areturn tubefortheaqueouspartofthedistillateallowedthecohobation techniquetobeused.DriedleavesofT.articulata(100g),groundat diameter1.5mm,wereplacedinaClevengerapparatusandmixed withdistilledwaterataratioof1/10(plantmaterial/water).After 180min(correspondingtothedurationwhennomoreessential oilwasobtainedatthecondenser),theessentialoilwascollected, driedoveranhydroussodiumsulfateandstoredat4◦_Cinthedark

untilused.

2.5. Chemicalcomponentsanalysis:GC–FIDandGC–MS

Quantitative and qualitative analysis of the volatile oil was carried out by gas chromatography–flame ionization detection (GC–FID)andgaschromatography–massspectrometry(GC–MS). Gaschromatography analyseswereperformedona VarianStar 3400 Cx chromatograph (Les Ullis, France) fitted with a DB-5MSfusedsilicacapillarycolumn(5%phenylmethylpolysyloxane, 30m×0.25mm,filmthickness0.25mm).Chromatographic condi-tionswereatemperaturerisefrom60◦_Cto260◦_{Cwithagradientof}

5◦_{C/minandthen15minat260}◦_{C.Asecondgradientof40}◦_C/min

Fig.1.SFEyieldofvolatileoilfromT.articulataatWCO2=20g/min,P=90barandT=40

◦

C.(a)Kineticsoftheextractionprocess,(b)influenceofstaticperiodduration (followedby30minofdynamicextraction)and(c)influenceofdynamicextractionduration(withastaticperiodof10min).

analysispurposes,thevolatileoilwasdissolvedinpetroleumether. Samples(1mL)wereinjectedinthesplitmodeataratioof1:10. Helium(purity99.999%)wasusedasthecarriergasat1mL/min. Theinjectorwasoperatedat200◦_{C.Themassspectrometer}

(Var-ianSaturnGC-MS-MS4D)wasadjustedforanemissioncurrentof 10mAandelectronmultipliervoltagebetween1400and1500V. Thetemperatureofthetrapwas150◦_{Candthatofthetransferline}

was170◦_{C.Massscanningwasfrom40to650amu.}

Compoundswereidentified (i)bycomparisonoftheir reten-tionindex(RI),obtainedonanonpolarDB-5MScolumnrelative toC5-C24n-alkanes,withthose providedintheliterature,(ii)by

comparisonoftheirmassspectrawiththoserecordedinNIST08 (NationalInstituteofStandardsandTechnology)andreportedin publishedarticles,(iii)byco-injectionofavailablereference com-pounds.Thesampleswereanalyzedinduplicate.Thepercentage compositionofthevolatileoilwascalculatedbythenormalization methodfromtheGCpeakareas,assumingidenticalmassresponse factorsforallcompounds.Resultswerecalculatedasmeanvalues aftertwo injectionsofvolatileoil,withoutusingcorrection fac-tors.Theidentificationisonlymadeforthevolatilecompounds. Especially,forSFEextracts,somenonvolatilecompoundscannotbe identified.Thedetectionlimitis0.1mg/Lwhilethequantification limitis1mg/L.

2.6. Determinationoftotalphenoliccontent

The phenolic contents of extracts were determined by the Folin–Ciocalteumethod[20]. Adiluted solutionof each extract (0.5mL) was mixed with the Folin–Ciocalteu reagent (0.2M, 2.5mL).Thismixturewaskeptatroomtemperaturefor5minand thenasodiumcarbonatesolution(75g/Linwater,2mL)wasadded.

After1hofincubation,theabsorbancewasmeasuredat765nm againstwaterblank.Astandardcalibrationcurvewasplottedusing gallicacid(0–300mg/L).TheresultswereexpressedasmgofGallic AcidEquivalent(GAE)/kgofplantmaterial.

2.7. DPPH•_{freeradicalscavengingactivity}

Antioxidantscavengingactivitywasdeterminedusingthe 1,1-diphenyl-2-picrylhydrazyl free radical (DPPH•_{)as described by}

Blois[21]withsomemodifications;variousdilutionsofthetest materials(ascorbicacidorextracts,1.5mL) weremixed witha 0.2mM methanolicDPPH• _{solution (1.5mL).After 30min}

incu-bationperiod at25◦_{C, theabsorbanceat 520nm wasrecorded}

asA(sample).Acontrolexperimentwasalsocarriedoutby

apply-ingthesame proceduretoa solutionwithoutthetest material and the absorbance was recorded (A(blank)). The free radical

scavenging activity of each solution was then calculated as percentageinhibitionaccordingtothefollowingequation:% inhi-bition=100×[(A(blank)−A(sample))/A(blank)]

ExtractantioxidantactivitywasexpressedasIC50,definedasthe

concentrationofthetestmaterialrequiredtocausea50%decrease ininitialDPPH•_{concentration.Valueswereestimatedusinglinear}

regression.Ascorbicacidwasusedasareference. 2.8. ABTS•+_{radicalscavenging}

The radical scavenging capacity of the samples for the ABTS•+ _(2,2′_{-azinobis-3-ethylbenzo-thiazoline-6-sulfonate) was}

determinedasdescribedbyReetal.[22].ABTS•+_{wasgeneratedby}

mixinga7mMsolutionofABTS•+_{atpH7.4(5mMNaH}

2PO4,5mM

(finalconcentration)andstoredinthedarkatroomtemperature for16hbeforeuse.Themixturewasdilutedwithwatertogive anabsorbanceof0.70±0.02unitsat734nmusing spectrophoto-metry.Foreachsample,adilutedmethanolsolutionofthesample (100mL)wasallowedtoreactwithfreshABTS•+_{solution(900mL),}

andtheabsorbancewasmeasured6minafterinitialmixing. Ascor-bicacidwasusedasareferenceandthefreeradicalscavenging capacitywasexpressedbyIC50(mg/L)values,whichrepresentsthe

concentrationrequiredtoscavenge50%ofABTS•+_{.Thefreeradical}

scavengingcapacityIC50wasdeterminedusingthesameequation

asusedpreviouslyfortheDPPH•_method.

2.9. Statisticalanalysis

Resultsofantioxidantactivityandtotalphenoliccontentwere expressedasmeans±standard deviationsoftriplicate measure-ments.Theconfidence limitswere setatP<0.05. Dataanalysis procedure(ANOVA)wasperformedinordertoassessthedata. 3. Resultsanddiscussion

3.1. InfluenceofparametersforSFE 3.1.1. Influenceofextractionduration

Anexampleofkineticsoftheextractionprocessat90barand 40◦_{CispresentedinFig.1(a).Theglobalextractionyieldincreases}

linearlyuntilabout30minutesandthenaplateauisreached.The initiallinearshapeofthecurvesuggeststhatextractionislimited bythesolubilityofvolatileoilinCO2 orbysolid–fluid

equilib-rium;internaldiffusionlimitationseemsnottobepresentinthese operatingconditions.

Weuseda dynamicextractionmethodwhere theextraction durationisactuallyconstitutedofaperiodofstaticpretreatment (i.e.,noCO2flowrate)followedbyadynamicperiod(i.e.,with

con-stantCO2flowrate).Thestaticpretreatmentallowsequilibration

ofthesolidandthefluidanddoesnotrequireCO2flow,andalso

allowsSCCO2dissolvingintotheoilpresentinsecretorystructures

oftheplanttissue(cells,cavitiesorducts),previouslyopenedby grinding,asdescribedbyStameni ´cetal.[23].

Inthefirstpartofthisstudy,theeffectofstaticpretreatment onSFEefficiencywasstudied.Threestaticpretreatmentdurations (5,10and15min),followedby30mindynamicextraction,were employed.Thepressureandtemperaturewerefixedat90barand 40◦_{Cwithparticlediameter1.5mm.Experimentalglobalyieldsare}

presentedonFig.1(b).Itisseenthatincreasingthestaticextraction period,from5to10min,increasedextractionefficiencyofvolatile oilbymore than40%.However,increasingthestaticextraction durationfrom10to15minhadminimaleffectontheextraction efficiency.Therefore,inallexperiments,10minofstatic pretreat-mentwasused.

Effectofdynamicextractiondurationonextractability ofthe naturalsubstanceswasinvestigatedwhilekeepingother parame-tersfixed.Thedynamicmethodusingthreedurations(20,30,and 60min)wastestedwith10minstaticpretreatmentduration.The pressure,temperatureandCO2 flow ratewere90bar,40◦C and

20g/minforallexperiments.

Fig.1(c) shows how extended dynamic extractiondurations increasetheefficiencyoftheprocess.However,thedependence betweenrecoveryandextractiondurationgraduallybecomesflat. Thereforea dynamicmethodwith30minwaschosenin subse-quentextractions.

3.1.2. Influenceoftemperature

TheinfluenceoftemperatureontheextractionyieldofT. artic-ulataat three temperaturelevels,40, 50and 60◦_{C isshown in}

Fig.2.Ahightemperatureislikelytoimprovemasstransferrates

Fig.2. InfluenceofthetemperatureontheT.articulatavolatileoilyieldfor1.5mm particlesasafunctionoftheextractiontimeatP=90barandWCO2=20g/min.

andthuskineticsofextraction.Neverthelessatlowpressure,the density of CO2 is very sensitive to temperature and decreases

withincreasingtemperature,leadingpossiblytoreducedsolvent powerofSCCO2dependingonthecompetitionbetweendensity

ofSCCO2andvolatilityofthesolute(retrogradesolubilityeffect)

[24].Thismightbethereasonoftheobserveddecreaseoftheyield (1.6–1.48mg/g)whentemperatureisincreasedfrom40to60◦_C,as

seenonFig.2.SuchasimilarresultwasalsoobtainedbydaCruz Franciscoetal.[25]inthecaseoftheextractionofmonoterpenesof eucalyptusoil.Inaddition,itisverylikelythathigh-temperature extractionhaveadetrimentaleffectonthequalityoftheextract becauseofpossiblethermaldegradationreactions.Thepossibility toperformefficientextractionofvolatilecompoundsatlow tem-peratureisindeedoneofthemainadvantagesoftheSFEtechnology whencomparedtotraditionalHD.

3.1.3. Influenceofpressure

Fig.3presentstheinfluenceofpressureontheextractionyield of T. articulata leavesin SC-CO2 at three pressure levelsof 90,

280and 1000barfor 30minextractionduration (1.6, 19.2and 25.5g/kg,respectively).

TheSFEyieldofT.articulatawassignificantlyinfluencedbythe pressure,ascanbeseeninFig.3.Infact,increasingtheextraction pressurefrom90to280barincreasedtheyieldmorethan9times and,from90to1000bar,formorethan16times.ThehighCO2

densityathighpressureincreaseditssolventpowerandtherefore, moresubstanceswereextracted.Also,itcanbenoticedthat, espe-cially,theincreasefrom90to280barisveryefficientupon the yield.

High extraction pressure is likely tofavor the extractionof heavy and more polar compounds suchas phenolics, lipidsor waxes. Therefore, in thepresent study,the phenolic contentis expectedtoincreasewithincreasingpressure(280and1000bar).

Fig.3.InfluenceofthepressureontheT.articulataextractionyieldfor1.5mm particles,aCO2flowrateof20g/min,T=40◦Candextractionduration30min.

Fig.4.T.articulatavolatileoilyieldfor1.5mmparticlesatdifferentCO2flowrates

asafunctionoftheextractiontime(a)andafunctionofamountofCO2referredto

initialamountofplantmaterial(PM)(b)atT=40◦_CandP=90bar.

Nevertheless, increasing the pressure obviously results in a decreaseoftheselectivityoftheextraction[26,27].

3.1.4. InfluenceofCO2flowrate

Asshownin Fig.4(a),forextractionat90barand 40◦_C,the

CO2flowrateexhibitedapositiveandsignificanteffectontheT.

articulatavolatileoilextraction.Thisresultindicatesthat,forCO2

flowrateequalto5g/min,60minofextractionarenotsufficient to achieve complete extraction while 30min are sufficient at 20g/min.Also,atallflowrates,theinitialquasi-linearshapeofthe

Fig.5.T.articulatavolatileoilyieldfordifferentmeanparticlesizesataCO2flow

rateof20g/min,temperature40◦_{Candpressure90bar.}

curvessuggeststhehypothesisthatthisSFEprocessissolubility orsolid–fluidequilibriumlimited.

WhenplottingtheyieldversustheamountofCO2(Fig.4(b)),

thecurvesmergeandthisconfirmstheoccurrenceofsolubilityor equilibriumlimitationfortheextractionprocess.Theslopeofthese curvesgavethesoluteconcentrationintheoutputfluid(C0)ata

valueof0.20×10−2_g/gCO

2,avaluewhichisbelowthevalueof

sol-ubilityofusualterpenes(a-pineneforinstanceis1.051×10−2_g/g

CO2atT=40◦CandP=90bar[28]).Whenhighmasstransfer

lim-itationisnotsuspected,this indicatesvery probableadsorption phenomenawhichlimittheextractionprocess.

3.1.5. Influenceofparticlesize

TheeffectofparticlesizeonextractionrateisshowninFig.5. Despitesomedispersionin theexperimentalpoints,thecurves aresimilar,andespeciallythesamefinalyieldisreachedafterthe same extractionduration (around30min) for allparticle sizes, indicatingsimilarkineticbehavior.Thisresultisinaccordancewith thesuspectedpredominantequilibriumlimitationwhichpredicts kineticsalmostindependentofinternalandexternalmasstransfer, andthereforeofparticlesize.Thelowimpactofparticlesizeon extractionkineticsistypicalofglandularstructurescorresponding tosecretoryducts[23,29],asitisthecaseforplantsofthe Aster-aceae,Apiaceae,Hypericaceae,PinaceaeorConiferaefamily.These ductsareopenedbygrindingandtheoiliseasilyaccessibleforCO2.

CO2dissolvesintheearlystageofextraction,leadingtoanincrease

ofthevolumeoftheoilthatwets thesurfaceoftheplant.This

Table1

Chemicalcomposition(GC–MS)andglobalyieldofT.articulataextractsobtainedbySFEatdifferentpressuresandbyHD. Typeofextractiontechnique

SFE HD

Pressure(bar) 90 280 1000 –

Temperature(◦_{C) 40 40 40 100}

Flowrate(g/min) 20 20 20 0.02mL/s

Solvent CO2 CO2 CO2 Water

Extractiontime(min) 30 30 30 180

Globalyield(g/kg) 1.6 19.2 25.5 0.61 RI Compounds % % % % 906 Santolinatriene 0.85 0.39 931 Artemesiatriene 0.8 0.6 936 a-Pinene 31.32 24.9 940 Cumene 0.33 953 a-Fenchene 0.22 0.14 958 Camphene 0.49 0.45 967 b-Thujene 2.64 1.99 973 1-Ethyl-4-methylbenzene 0.43 990 Myrcene 0.17 0.48 1004 Pseudolimonene 0.26 0.37 1008 3-Carene 3.39 3.44 1012 a-Terpinene 0.12 990 Myrcene 0.17 0.48 1037 Limonene 0.16 0.18 1047 3-Methyl-1-decene 0.13 1068 Isoterpinolene 0.26 0.46 1077 Ethylmaltol 0.26 1079 Artemisiaalcohol 0.19 1089 a-Terpinolene 1108 cis-Thujone 0.22 0.35 1122 trans-2,8-Menthadienol 0.17 0.38 1128 a-Campholenal 3.44 0.1 0.43 3.84 1132 Isothujol 0.17 1148 Camphenehydrate 1.25 4.36 1157 Isoborneol 1159 b-Terpineol 0.18 0.73 1172 3,5,5-Trimethyl-1,4-cyclohexanedione 0.09 0.14 1.41 1193 Myrtenal 0.23 0.08 1200 Z-Dihydrocarvone 0.18 0.34 1232 cis-Carveol 0.15 1266 Linaloolacetate 18.18 1.72 21.44 1279 p-sec-Butylphenol 0.58 1318 (Z)-3-Hexenyl2-methyl-(E)-2-butenoate 0.38

1327 Trans-Pulegoneoxide[a] 0.91 0.31 1.53

1356 a-Cubebene 1.11 0.47 0.26 1370 Cyclosativene 0.72 0.16 1383 b-Bourbonene 1.74 0.44 1400 g-Caryophyllene 4.16 2.05 2.53 1409 Aromadendrene 1435 a-trans-Bergamotene 1.29 0.59 1.04 1444 a-Caryophyllene 0.13 1455 a-Patchoulene 0.48 0.18 1461 Alloaromadendrene 7.55 1.41 1.14 1474 g-Muurolene 0.35 1479 g-Curcumene 0.58 0.43 1494 a-Zingiberene 2.14 2.38 0.53 1501 a-Amorphene 1.39 1.64 1516 a-Selinene 0.32 1534 incisol 0.17 0.59 1557 Caryophyllenealcohol 1.06 1568 Caryophylleneoxide 1.82 2.45 4.24 1588 b-Oplopenone 0.38 1594 Cedrenol 0.48 1.03 2 1613 1-epi-Cubenol 0.52 2.12 2.83 1619 trans-Isolongifolanone 0.42 1626 g-Eudesmol 0.58 2.56 1642 Spathulenol 0.43 0.77 2.37 1660 14-Hydroxy-9-epi-(E)-caryophyllene 0.14 0.5 0.97 1673 Cedr-8-en-13-ol 0.23 1678 cis-Alpha-santalol 0.44 1.24 1.74 1717 (Z)-2-Heptadecenea _0.22 1761 1,10-Dihydronootkatone 0.88 0.08 1928 16-Hexadecanolide 0.43 0.15 1962 Palmiticacid 2.98 36.19 1991 Manoyloxide 0.13 0.26 2008 Kaur-16-ene 0.14 0.1

Table1(Continued)

Typeofextractiontechnique

SFE HD 2053 abietatriene 0.32 2095 Heneicosane 0.4 3.66 0.09 2135 1R-pimaral 0.13 6.09 0.27 2143 Cubitene 0.22 2187 CembreneAketone 0.25 1.37 0.14 2216 15-Ripperten-3alpha-ol 0.41 2.11 0.52 2223 trans-Totarol 0.12 2274 Larixola _2.17 2288 2-Methyltricosane 10.29 2311 trans-Totarol 4.85 0.05 0.07 2368 androst-5-en-17-ol,4,4dimethyl 2.28 2.49 2383 Podocarp-7-en-3-one,13a-methyl-13-vinyl 7.77 25.53 2473 Docosylacetate 14.4 1.35 2922 Abieticacid 0.31 2925 ni 1.05 2930 Podocarp-8(14)-en-15al,13a-methyl-13-vinyl 10.79 ni(M=410) 0.73 ni(M=552) 2.14 ni(M=618) 8.96 ni(M=296) 7.34 7.98 Numberofcompounds 56 29 14 56 Monoterpenehydrocarbons 40.73 0 0 34 Monoterpeneoxygenated 6.92 0.41 0.43 12.36 Sesquiterpeneshydrocarbons 21.48 6.9 0 8.83 Sesquiterpenesoxygenated 5.64 8.99 0 18.41 Others 21.14 67.55 97.71 25.03 Total 96.24 83.85 98.14 98.63

ni,notidentified.

a_{Tentativeidentificationsupportedbygoodmatchofmassspectrum.}

hypothesis could explain the predominance of equilibrium phenomenaascomparedtointernalmasstransferintheextraction process.

3.1.6. EffectofSCCO2extractiononthestructureofvegetal

material

Optical microscope or Scanning Electron Microscope (SEM) analysesofthematerialweredonetovisualizeitsstructureand theeffectoftheextractionprocessuponthisstructure.Theinitial structureofthegroundleavesfromT.articulata,observedbySEM, isshowninFig.6(a).FromcomparisonwithFig.6(b),itisseenthat SFEhadalmostnoeffectonthestructureoftheplant,atleaston avisualpointofview.Moreover,SEMimagestendtoconfirmthe hypothesisofsecretoryducts,sincetheplantisactuallymadeofa coreofinterconnectedchannelssurroundedbyadensebark. 3.2. CharacterizationofextractsandcomparisonwithHD 3.2.1. Chemicalcompositionoftheextracts

TheglobalyieldandtheresultsoftheGC–MSanalysisofvolatile oilobtainedfromT. articulata by SFEand HD are presented in Table1.

Asexpected,therewasasignificantdifferenceintheextraction yieldsbetweentheextractsobtainedusingHDandSFEatthethree pressures.Extractionathighpressurewasfoundtobemore effi-cientthanotherconditionsstudiedinextractingtheantioxidant componentspresentintheT.articulataleaves.

Althoughatotalof86compoundswereidentified,forthesake ofsimplicity,thediscussionhasbeenrestrictedtothemajor com-ponentsidentifiedforeachmethod.Thesemajorcomponentsare presentedinTable2,whereatotalof28componentswitha per-centagehigherthan2%aregathered.Resultsarepresentedinterms ofrelativeproportion(%ofareaofthepeakinthechromatograms) andamountinmg/kgofextract.CompoundsappearinTable2inthe orderofdecreasingvolatility.Althoughthiscouldleadto approx-imateresults,becauseterpenesarethemaincomponents,ithas

beenconsideredherethatpalmiticacid,abieticacidandterpenes havesimilarresponsefactor.

Whencomparingcompositionofessentialoil(HD)andvolatile oilobtainedbySFEat90bar,nogreatdifferenceswereobserved anda-pineneandlinaloolacetatearepredominantintheextracts obtainedwiththesemethods.However,thecompoundsinextracts obtainedbySFEat280and1000bararesignificantlydifferent com-paredtothoseobtainedbyHDorSFEat90bar.

ItappearsthatthehigherthepressureofSFE,thelowerthe numberofdetectedcompounds,whichinthatcasecorresponds toheavycompounds.For example,only 14volatilecompounds havebeendetectedbyGC–MSintheextractobtainedbySFEat 1000bar(only10havebeenreportedinTable2).Notealsothat, duetothegaschromatographymethod,onlyhighmolecularmass compoundswithretentionindex(RI)valuegreaterthan1900were detected.Ontheotherside,theessentialoil(HD),volatileoilofSFE (at90bar)andextractfromSFEat280barwerecomposedof56,56 and29compounds,respectively.Thisresultissurprisingbecause highpressureoperationallowsextractingmorecompoundsdueto theincreasedsolventpowerofSCCO2.Indeed,thisresultsfroman

“artifact”oftherecoverysystemoftheA21(SFE1000bar)apparatus whichwasnotveryeffectiveoratleastincorrectlyused.This appa-ratushasonlyoneseparator(converselytoSF200extractionwhich has3separatorsinseries).Experimentsat280and1000bar(where volatilecomponentswerenotdetected)weredoneonA21 appara-tuswhileexperimentsat90barweredoneontheSF200extraction (wherethesevolatilecomponentswererecoveredanddetected). SoitisveryprobablethattheuniqueseparatorofA21hasfavored theselectivelossofthevolatilecomponentsthatobviouslywere extractedathighpressurebutnotrecoveredinthelowpressure recoveryzone.Thisexplainsthatlowmolecularmasscompounds donotappearinthecomposition.

In volatileoils obtained by HD and SFE at 90bar,the main constituents were a-pinene (24.90–31.32%), linalool acetate (21.44–18.18%), alloaromadendrene (1.14–7.55%), camphene hydrate(4.36–1.25%)andg-caryophyllene(2.53–4.16%).Themain

Table2

ChemicalcompositionobtainedbyGC–MS(abundance>2%)andglobalyieldofT.articulataextractsobtainedbySFEatdifferentpressuresandbyHD. Typeofextractiontechnique

SFE HD

Flowrate(g/min) 20 20 20 0.02mL/s

Temperature(◦_{C) 40 40 40 100}

Pressure 90 280 1000

-Solvent CO2 CO2 CO2 Water

Extractiontime(min) 30 30 30 180

Globalyield(mg/kg) 1.63 19.2 25.9 0.61 RI Compounds % mg/kg % mg/kg % mg/kg % mg/kg 936 a-Pinene 31.32 511 24.9 152 967 b-Thujene 2.64 43 1.99 12 1004 Pseudolimonene 0.26 4 0.37 2 1008 3-Carene 3.39 55 3.44 21 1128 a-Campholenal 3.44 56 0.1 19 0.43 111 3.84 23 1148 Camphenehydrate 1.25 20 4.36 27 1266 Linaloolacetate 18.18 296 1.72 330 21.44 131 1400 g-Caryophyllene 4.16 68 2.05 394 2.53 15 1461 Alloaromadendrene 7.55 123 1.41 271 1.14 7 1494 a-Zingiberene 2.14 35 2.38 457 0.53 3 1568 Caryophylleneoxide 1.82 30 2.45 470 4.24 26 1613 1-epi-Cubenol 0.52 8 2.12 407 2.83 17 1626 g-Eudesmol 0.58 9 2.56 16 1642 Spathulenol 0.43 7 0.77 148 2.37 14 1962 Palmiticacid 2.98 572 36.19 9373 2095 Heneicosane 0.4 7 3.66 703 0.09 1 2135 1R-pimaral 0.13 2 6.09 1169 0.27 2 2216 15-Ripperten-3alpha-ol 0.41 7 2.11 405 0.52 3 2274 Larixola _{2.17 417} 2288 2-Methyltricosane 10.29 1976 2311 trans-Totarol 4.85 931 0.05 13 0.07 2368 Androst-5-en-17-ol,4,4dimethyl 2.28 438 2.49 645 2383 Podocarp-7-en-3-one,13a-methyl-13-vinyl 7.77 1492 25.53 6612 2473 Docosylacetate 14.4 2765 1.35 350 2930 Podocarp-8(14)-en-15al,13a-methyl-13-vinyl 10.79 2795 ni(M=552) 2.14 554 ni(M=618) 8.96 2321 ni(M=296) 7.34 1409 7.98 2067 Total 78.62 1282 76.94 14,772 95.91 24,841 77.49 473

ni,notidentified.

a_{Tentativeidentificationsupportedbygoodmatchofmassspectrum.}

compoundsofextractsobtainedbySFEat280and1000barwere palmitic acid (2.98–36.19%), docosyl acetate (14.40–1.35%), podocarp-7-en-3-one, 13a-methyl-13-vinyl (7.77–25.53%), 2-methyltricosane (10.29–0%), podocarp-8(14)-en-15al, 13a-methyl-13-vinyl(0–10.79%)and1R-pimaral(6.09–0%).Also,some compounds with high molecular mass were detected but not identified.Forinstance,theamountofsomeoftheseunidentified compounds was significant in SFE extract at 1000bar (8.96% (M=618g/mol)and7.98%(M=296g/mol)).

On the other hand, important differencebetween the com-poundsofeachextractscanbeseenintermofquantity(mg/kg). Itisthecaseforexampleofa-pinenewhichwasidentifiedasthe majorcompoundinHDandSFEat90bar,wherequantities deter-minedinHDandSFEextracts,were152mg/kg(for24.9%ofglobal oil)and511mg/kg(for31.31%ofglobaloil)respectively.

Itisinterestingtocomparetheseresultswiththefewstudies thathavebeenconductedontheessentialoilsofT.articulata.Ben HadjAhmedetal.[30]havestudiedessentialoilfromT.articulata plantsharvestedin theregionofSnedGafsa(Tunisia)andhave onlyreportedaboutantioxidantactivityoftheessentialoiland notaboutitschemicalcomposition.Tekaya-Karouietal.[31]have workedonT.articulataplantsharvestedintheregionofZaghouas (Tunisia)andhaveshownthatchemicalcompositionwasdifferent, dependingonthepartoftheplant.Intheessentialoilobtained from branches, themajor compounds were camphene (43.2%), Z-muurolène(29.0%)and p-cymene-8-ol(10.4%) whileessential oilfrom theroots was richer in camphene(70.2%). A study of Barreroetal.[32]uponthechemicalcompositionofessentialoil

fromleavesandwoodofT.articulataplantsharvestedintheregion ofAmsaTéouan(Morocco)hasshownthattheleaveswererich inbornylacetate(16.5%),camphor(19.1%)andborneol(9.6%)and thattheessentialoilfromthewoodwasrichincedrol(28.2%)and 1,7-di-epi-cedrol(17.9%).TheresultsreportedbyAitigri[33]have shownthatessentialoilfromwoodofT.articulataplantsharvested in the region of Rabat (Morocco) were rich in carvacrol and a-cedrol.AnotherstudyofZriraetal.[34]uponT.articulataplants fromtheregionofKhemissetandAoulouz(Morocco)hasreported thepredominant presence ofcarvacrol (21.3–36.4%), a-cedrene (10.1–13.1%),cedrol(1–7.3%)andterpinen-4-ol(2.8–6%).Inastudy fromBuhagiaretal.[35],18compoundswereidentifiedinaerial partsofT.articulataplants,wherea-pinene,limonene,camphor andbornylacetatewerethemaincompounds.Fromthissurvey,it isseenthatthechemicalcompositionsofT.articulataessentialoils isverydependentontheharvestingplaceandonthespecificpart oftheplant.Tothebestofourknowledge,nostudyuponchemical compositionofSFEextractsofT.articulatahasyetbeenproposed. 3.2.2. Phenolicscontentandantioxidantactivityoftheextracts

Theconcentrationofphenolicsintheextracts,expressedas Gal-lic AcidEquivalent(GAE) isshown inTable3.Asexpected,the amountofphenoliccompoundsintheSFEextractsathighpressure wasthehighest(102.03±4.57and267.90±8.06mgGAE/gplant materialfor280barand1000bar,respectively),followedbySFEat 90bar9.70±0.57andHD8.89±0.16mgGAE/gdryplantmaterial. TheresultsfromTable3showthat,althoughthequantityof phenolicsislowandalmostidenticalfortheHDessentialoiland

Table3

TotalphenolicscontentandantioxidantactivityofT.articulataextractsobtainedby SFEatdifferentpressuresandbyHD.

Samples Phenolics (GAEmg/g) DPPH ABTS GAEmg/g ± IC50 ± IC50 ± SFE90bar 9.70 0.57 146.02 1.99 40.91 0.53 SFE280bar 102.03 4.57 120.21 3.65 33.55 0.40 SFE1000bar 267.90 8.06 108.16 3.07 29.77 1.30 HD 8.89 0.16 3681.49 69.33 324.45 14.21

GAE,GallicAcidEquivalent.

SFEextract at90bar, theirantioxidantactivityis different.The SFEextractat90bar(IC50(mg/ml)=40.91±0.53)isabout8times

moreactivethanessentialoil(IC50(mg/ml)=324.45±14.21).These

resultssuggestthattheSFEvolatileoilfromT.articulata,ismore concentratedwithantioxidantsandcouldbeusedasapotential sourceofnaturalantioxidantswithpossibleapplicationsinfood systems.Thepresenceofantioxidantsinthis plantisconfirmed byapreviouspublication[11]wherephenolics andantioxidant activityofAlgerianT.articulataleavesextractweredetermined. Theseauthors showedthat extract ofT. articulata fromAlgeria obtainedbyheatrefluxwithdistilledwaterexhibitedgood antiox-idantactivity(IC50(mg/ml)=9.519±0.521andhightotalphenolic content(mgGAEg−1_{)=206.187±16.612).AlthoughGC–MSisable}

toidentifysomeof thesemolecules,theywerenot detectedin ourcase.Thereforethehighmeasuredantioxidantactivitycanbe explainedbythepresenceofhighmolecularweightphenolics (tan-nins,anthocyanins,etc.)whichcannotbedetectedusingGC–MS.

Anotherexplanationcouldbetheoccurrenceoflowquantities ofhighlyactivephenoliccompounds.ForexampleHarishNayaka etal.[36]reportedthat,fromvaluesofIC50obtainedforphenolic

acidextractsofswallowroots,vanillicacidwas45timeslessactive thangallicacid.

3.3. Mathematicalmodeling

3.3.1. Preliminaryanalysisofthephysicalphenomena

Because vegetal matter is very diverse, numerous models describingsupercriticalextractionprocessareavailableinthe liter-ature,fromthesimpleempiricalonestothemostcomplexthatare abletohandleallphysicalphenomenaoccurringduringextraction. Thechoiceisalwaysdifficultandthemostcomplexmodelmight notbethemostusefulbecauseitisassociatedwiththeestimation ofnumerousparameters, whilesomeofthemmayhavea neg-ligibleinfluence.Moreovertheircomplexitypreventssometimes fromeasyunderstandingwhichisnecessarytoselecttheoptimal conditionsfortheextractionprocess,especiallywhenscale-upis envisaged.So,qualitativeinterpretationofkineticcurvestogether withtheuseofeasy-to-implementsimplifiedmodelsisvery use-ful.Especially,qualitativestudyoftheextractioncurvesatdifferent flow-ratesanddifferentparticlesizesisveryinstructive.Suchan approach hasbeenrecently recommended bySovová [17] in a recentpaperwhichprovidesasimplifiedmethodbasedontheuse ofcharacteristictimesforallinvolvedphysicalprocesses. Compari-sonofthesecharacteristictimesgreatlyhelpsfordiscriminatingthe prominentparameters.However,caremustbetakentoinsurethat, whenscalingup,thechoiceoftheseparametersremainspertinent. We haverestricted ourapproach tovolatile oilsupercritical extraction and therefore the modeling is proposed for results obtainedat90barand40◦_C.

In the case presented here, from the overlapping of curves of Fig. 4(b), as indicated in Section 3.1.4, a limitation by the adsorption phenomenon is strongly suspected. In this case, when external mass transferis not involved, estimation of the

adsorptioncoefficientisobtainedfromtheinitialslopeofFig.4b. Thehypothesisofnegligibleexternalmasstransferlimitationcan beassessed by computingexternal mass transfer characteristic timetf[17].Externalmasstransfercharacteristictimetfdepends

on kf (external mass transfer coefficient),  which is the

vol-ume/surfaceratiooftheparticle(=R/3forsphericalparticles)and ε,thevoidfractionoftheparticlebed.

tf =₁ε −ε  kf (1) t1= G Km·q′·(1−exp(−1/f)) (2) Theextractiontimet1correspondingtotheendofthelinearpart

oftheextractioncurveisthusafunctionofq′_whichistheCO 2

flow-ratereferredtothemassofsolid,G,thefractionofeasilyaccessible cells,andKmwhichisthesolutepartitioncoefficient(kgplant/kg

solvent)definedasC=Km·q,whereCistheconcentrationofvolatile

oilinthefluidphase(gsolute/gCO2)andqistheconcentrationof

volatileoilinthesolidphase(gsolute/gplant).f,istheratioof

theexternaltransfercharacteristictimetftotheresidencetimetr.

Thislatterisgivenby: tr=

q′ (3)

where istheratiobetweeninitialCO2massandsolidmassinthe

extractor.

To providea quantitativeassessment of these characteristic times,extractionat thehigher flow-rate,20g/min,for assumed spherical 1.5mm diameterparticles, is consideredbecause this caseisthemostlikelytoexhibitinfluenceofexternalmass trans-fer.Consideringavoidfractionε=0.457,thecharacteristic time for external transfertf is computedat 1.05susing the

correla-tion of Tan et al. [37] for the evaluation of kf (which yielded

kf=1.61×10−4ms−1)andtrisequalto75s.Thevalueoftfbeing

lowascomparedtotrconfirmsthatexternalmasstransfer

resis-tancecanbeneglected.Thisresultmakesitpossibletouseasimple equilibriummodelforthelinearpartoftheextractioncurve.Such anapproachwasproposedbyReis-Vascoetal.[38]forthe extrac-tionofpennyroyalessentialoil.Visually,fromFig.4(a),thevalue G=0.7fortheaccessiblefractionisestimated,correspondingtothe endofthelinearpartofthecurveswhichoccursatt1=750s.

Theinterpretation of thecurved final partof the extraction curvesistrickier.Inafirstapproach,assuggestedbySovová,this partwouldcorrespondtoaninternalmasstransferlimitedprocess, concerningtheextractionofthenonaccessiblepartofthesolute. Ifthiswastrue,finalpartsofcurvesofFig.4(b)(whichrelatesthe extractionyieldtothemassofCO2used)shouldexhibitinfluence

oftheflowrate,whilethisisnotthecasebecausecurvesoverlapup totheendofextraction.Inaddition,aconfirmationofthis hypoth-esisisfoundwhenattemptingtofittheexperimentaldatawiththe simplifiedmodelforplug-flowwithadsorptiongiveninSovová’s paper[17]: e=xu−(1−G)xuexp





wherexuistheinitialsolutecontentofthesolidbeforeextraction

(itisdeterminedfromthevalueofthefinalyieldoftheextraction curvesatavaluexu=1.6×10−2_{g/g).Thecombinedcharacteristic}

timeofmasstransfert_comb,iisobtained,asafunctionoft_f,Kmandti

(internalmasstransfercharacteristictime)thatinvolvesthe effec-tiveinternaleffectivediffusioncoefficientDe.Thesecharacteristic

timesaregivenby: ti=

R2

tcomb,i=ti+

tf

·Km (6)

Whenusingthisapproachforthefinalpartoftheexperimental curvesobtainedatdifferentCO2flow-rates,parameter

identifica-tionyieldeddifferentvaluesoftheinternaldiffusioncoefficientDe

foreachflow-rate,whilethisparametershouldnotbeaffectedby theflow-rate.

Alltheseelementsindicatethatthefinalcurvatureofthe extrac-tioncurvescorrespondsmoreprobablytotheinfluenceofaxial dispersion(whichisnotaccountedforinthesimplifiedmodel). Eventually,thissuggestsanadsorptionlimitedextractionprocess where allthesoluteis accessiblewithnegligibleexternal mass transferlimitation,butwheresignificantaxialdispersionispresent. Thecomputation of thePeclet number,Pe=h·u/(DL·ε),withthe

relation of Funakuzuri [39] for theaxial dispersioncoefficient, yieldedPe=19whichisquitelowandascertainsthehypothesis ofanon-negligible axialdispersioninfluence(seedetailsbelow intheparagraphaboutmodeling).Thismakesitpossibletousea simplifiedmodelwhichindeedcorrespondstothefirstpartofthe Reis-Vasco’smodel,i.e.,differentialmassbalanceequations corre-spondingtothecaseofextractionlimitedbythesoluteadsorption anddescribingtheflowpatternasapistonflowwithaxial disper-sion[38].

3.3.2. Descriptionoftheequationsofthemodel

Massbalanceequationsinthatcasecorrespondtothefollowing assumptions:(1)thesuperficialvelocityofSCCO2,u,isconstant

duringtheprocess,(2)theprocessisisothermalandpressuredrop isneglected,(3) volatileoilisdescribedbyasinglecomponent, the“solute”and(4)alinearequilibriumrelationshipisassumed definedasC=Km·q.Thecorrespondingmodelequations,boundary

andinitialconditionsare(Reis-Vascoetal.)[38]: fε∂C ∂t +s(1−ε) ∂q ∂t =−fu ∂C ∂z +DLfε ∂2_C ∂z2 (7) C=Km·q (8) t=0, C=C0 and xu=q0+ C0 z>0 (9) z=0, u εC−DL ∂C ∂z =0 t>0 (10) Z=L, ∂C ∂z =0 t>0 (11)

Eq.(7)canbere-writtenas:





wheretisextractiontime,zistheaxialcoordinateofthe extrac-tionbed,C0istheconcentrationofvolatileoilinthefluidphaseat

thebeginningoftheextractionprocesswhichcorrespondstothe equilibriumwiththeinitialsolidphaseconcentrationq0,q0 and

C0beingrelatedbyEq.(9),εisthebedvoidfraction,sdensityof

vegetablematter,andDListheaxialdispersioncoefficient(m2/s).

ThesystemofequationswassolvedwiththeMatlab®_software.

InitialconcentrationofsoluteinCO2,C0,wascomputedfromthe

initialslopeofFig.4(b)and,thevalueC0=0.2×10−2gsolute/gCO2

wasobtained.Equilibriumadsorption constantKm (=C0/q0)was

thencomputedatKm=0.14kgplant/kgCO2.

At90barand40◦_{C,densityandviscosityofCO}

2wereevaluated

atf=448.45kg/m3andf=40.1×10−6Pas.Theaxialdispersion

Fig.7. T.articulatavolatileoilyieldfor1.5mmparticlessize,pressure90barand temperature40◦_{CatdifferentCO}

2flowrateswithReis-Vascomodel(firstpartonly).

coefficientDLwasdeterminedusingthecorrelationproposedby

Funazukuri[39]: DL=1.317(ReSc)1.392





The binary diffusion coefficient D12 was estimated by the

Wilke–ChangequationasproposedbyFunazukuri: D12=7.410−12

T

p

fV_2bp0.6

(16) whereTisthetemperature(K),˚istheassociationfactorofthe solvent(˚=1 forCO2), Misthemolarmassofthesolventand

V2bpisthemolarvolumeofthesoluteatitsnormalboilingpoint

expressedincm3_{/mol.Asa-pineneisacompoundpresentinall}

volatileextracts,its molarvolume V2bp=178.63×10−3m3/kmol

wasusedin(15)[17].AvalueofD12=1.7×10−8_m2_{/swasobtained}

for90bar,40◦_{C.Axialdispersioncoefficientwasfoundtobeequal}

to1.9×10−5_m2_{/sandPe=19inthecaseofCO}

2flow-rateequalto

20g/min.

In Fig.7, experimentalcurves arecompared withcalculated yieldsusingthismodelandagoodagreementisobservedforthe threedifferentCO2flow-rates.Togetherwiththequalitative

anal-ysiswe have developed in the preliminaryanalysis,this result validatesthechoiceoftheproposedmodel.

3.3.3. Useofthemodelingforscaling-uptheextractionprocess Aswementioned,significantaxialdispersionwaspointedoutat laboratoryscale(Pe=19for20g/minCO2flow-rate).Nevertheless

itsinfluenceontheeffectivedurationforextractingthesoluteisnot verystrong.Forinstance,at20g/min,durationfortheobtaining of90%yield ispredictedat16minwiththeSovova’ssimplified model(whichneglectsaxialdispersion,t1obtainedbyEq.(2))while

ourmodeling,whichtakesintoaccountaxialdispersion,indicates avaluecloseto20min.Thisrathermoderatedifferencemakesit possibletoproposesimplifiedscale-upbymaintainingthespecific flow-rateq′ _{andtheratioheighttodiameteroftheextractor (5}

inourcase).Withthisextrapolationrule,forinstancefora1000 foldgreatervegetalmasstoextract,simple computationresults ina113Lextractor(diameter0.3m,height1.6m)percolatedby

240kg/hofCO2(superficialvelocity10mm/s)inwhich90%yield

isobtainedafter16min.Inthiscase,itcanbecomputedthatthe Pecletnumbervalueisnow82,whichismorefavorablethaninthe laboratoryscaleextractor(wherePe=19)andindeedreinforcesthe validityofthissimpleextrapolationprocedure.

4. Conclusions

Recovery of bioactive compounds from T. articulata leaves was obtainedusing SCCO2 extraction. For volatile compounds,

optimalconditions were found to be90bar,40◦_{C, and 30min}

duration. For non-volatile fraction, suitable conditions were found to be 1000bar, 40◦_{C and 30min duration. Also,}

quan-titative assessment of the extract antioxidant power and the enrichment of antioxidant at different extraction conditions were performed. Interesting selectivity for compounds with highantioxidant activity(ABTS•+ _{essay (29.77±1.3, 33.55±0.4}

and 40.91±0.53mg/L) and phenolic content (267.90±8.06, 102.03±4.57and9.70±0.57GAEmg/gmaterial))wereobserved for SFE extracts at different pressures (1000, 280, and 90bar), respectively. Conventional HD was tested and essential oils obtained by this technique revealed low antioxidant activity (324.5±14.21mg/L).

Also, different chemical compositions of the extracts were found,dependingontheSFEpressureextraction.Physicalaspects ofthelow pressuresupercritical extraction(90bar,40◦_C)were

investigatedandmodelingoftheextractionkineticsusinga sim-plifiedformoftheReis-Vasco’smathematicalmodelprovedtobe adequate.Inourrangeofoperatingconditions,extractionofT. artic-ulatawasfoundtobegovernedbyadsorptionphenomenonand significantaxialdispersionwaspointedout.Thismadeitpossibleto proposesimpleextrapolationprocedure.Itisnoteworthythat pro-ductivitycanbeincreasedproportionallytoCO2flow-rate,aslong

asmasstransfereffectsremainnegligible.Also,adsorptionconstant canberenderedmorefavorablebyincreasingthesolventpowerof CO2 usinghigherpressurebuta correlativelossofselectivityis

expected. References

[1]M.J.Nicolás,M.A.Esteve,J.A.Palazón,J.J.LópezHernández,Modelosobrelas preferenciasdehábitataescalalocaldeTetraclinisarticulata(Vahl)Mastersen unapoblacióndellímiteseptentrionaldesuáreadedistribución,Analesde Biología26(2004)157–167.

[2]M.Bourkhiss,M.Hnach,J.Paolini,J.Costa,A.Farah,B.Satrani,Propriétés antioxydantesetanti-inflammatoiresdeshuilesessentiellesdesdifférentes partiesdeTetraclinisarticulata(Vahl),BulletindelaSociétéRoyaledesSciences deLiège79(2010)141–154.

[3]A.B.Valenzuela,S.K.Nieto,Syntheticandnaturalantioxidants:foodquality protectors,GrasasyAceites47(1996)186–196.

[4]A.Bernhoft,Bioactivecompoundsinplants–benefitsandrisksformanand animals,in:ProceedingsfromaSymposiumHeldatTheNorwegianAcademy ofScienceandLetters,Novusforlag,Oslo,2010.

[5]S.Martins,S.I.Mussatto,G.Martínez-Avila,J.Monta ˜nez-Saenz,C.N.Aguilar,J.A. Teixeira,Bioactivephenoliccompounds:productionandextractionby solid-statefermentation,BiotechnologyAdvances29(2011)365–373.

[6]C.G.Pereira,M.A.A.Meireles,Supercriticalfluidextractionofbioactive com-pounds: fundamentals,applicationsandeconomic perspectives,Food and BioprocessTechnology3(2010)340–372.

[7]L.-E.Shi,Z.-L.Zhang,L.-Y.Xing,D.-D.Yang,Y.-P.Guo,X.-F.Guo,L.-M.Zhao,Z.-X. Tang,AntioxidantsextractionbysupercriticalCO2,J.MedicinalPlantsResearch

5(2011)300–308.

[8]E.Reverchon,I.deMarco,Supercriticalfluidextractionandfractionationof naturalmatter,J.SupercriticalFluids38(2006)146–166.

[9]E.Reverchon,J.Daghero,C.Marrone,M.Mattea,M.Poletto,Supercritical fractionalextractionoffennelseedoilandessentialoil:experimentsand math-ematicalmodeling,IndustrialandEngineeringChemistryResearch38(1999) 3069–3075.

[10]E.Reverchon,C.Marrone,ModelingandsimulationofthesupercriticalCO2

extractionofvegetableoils,J.SupercriticalFluids19(2001)161–175.

[11]W.Rached,H.Benamer,M.Bennaceur,A.Marouf,Screeningoftheantioxidant potentialofsomeAlgerianindigenousplants,J.BiologicalSciences10(2010) 316–324.

[12]K.A.Abbas¸,A.Mohamed,A.S.Abdulamir,H.A.Abas,Areviewon supercriti-calfluidextractionasnewanalyticalmethod,AmericanJ.Biochemistryand Biotechnology4(2008)345–353.

[13]M.S.T. Barroso, G. Villanueva,A.M.Lucas,G.P. Perez, R.M.F. Vargas,G.W. Brun,E.Cassel,Supercriticalfluidextractionofvolatileandnon-volatile com-pounds from SchinusmolleL, Brazilian J. ChemicalEngineering 28(2011) 305–312.

[14]E.Reverchon,R.Taddeo,G.DellaPorta,Extractionofsageoilby supercriti-calCO2:influenceofsomeprocessparameters,J.SupercriticalFluids8(1995)

302–309.

[15]S.A.Aleksovski,H.Sovová,SupercriticalCO2extractionofSalviaofficinalisL,J.

SupercriticalFluids40(2007)239–245.

[16]S.Glisic,J.Ivanovic,M.Ristic,D.Skala,Extractionofsage(Salviaofficinalis L.)bysupercriticalCO2:kineticdata,chemicalcompositionandselectivityof

diterpenes,J.SupercriticalFluids52(2010)62–70.

[17]H.Sovová,Stepsofsupercriticalfluidextractionofnaturalproductsandtheir characteristictimes,J.SupercriticalFluids66(2012)73–79.

[18]S.Camy,J.S.Condoret,Modellingandexperimentalstudyofseparatorsfor co-solventrecoveryinasupercriticalextractionprocess,J.SupercriticalFluids38 (2006)51–61.

[19]S.Camy,J.S.Condoret,Dynamicmodellingofafractionationprocessfora liquidmixtureusingsupercriticalcarbondioxide,ChemicalEngineeringand Processing40(2001)499–509.

[20]V.L.Singleton,R.Orthofer,R.M.Lamuela-Raventós,Analysisoftotalphenols andotheroxidationsubstratesandantioxidantsbymeansofFolin–Ciocalteu reagent,MethodsinEnzymology299(1998)152–178.

[21]M.S.Blois,Antioxidantdeterminationsbytheuseofastablefreeradical,Nature 181(1958)1199–1200.

[22]R.Re,N.Pellegrini,A.Proteggente,A.Pannala,M.Yang,C.Rice-Evans, Antiox-idantactivityapplyinganimprovedABTSradicalcationdecolorizationassay, FreeRadicalBiologyandMedicine26(1999)1231–1237.

[23]M.Stameni ´c,I.Zizovic,A.Orlovi ´c,D.Skala,Mathematicalmodellingofessential oilSFEonthemicro-scale–classificationofplantmaterial,J.SupercriticalFluids 46(2008)285–292.

[24]G.Brunner,GasExtraction:AnIntroductiontoFundamentalsofSupercritical FluidsandtheApplicationtoSeparationProcesses,Springer,NewYork,1994, pp.176–250.

[25]J.daCruzFrancisco,B.Sivik,Solubilityofthreemonoterpenes,theirmixtures andeucalyptusleafoilsindensecarbondioxide,J.SupercriticalFluids23(2002) 11–19.

[26]H.Sovová,R.Komers,J.Kucera,J.Jez,Supercriticalcarbondioxideextraction ofcarawayessentialoil,ChemicalEngineeringScience49(1994)2499–2505. [27]P.B.Gomes,V.G.Mata,A.E.Rodrigues,Productionofrosegeraniumoilusing

supercriticalfluidextraction,J.SupercriticalFluids41(2007)50–60. [28]M.Akgun,N.A.Akgun,S.Dincer,Phasebehaviourofessentialoilcomponents

insupercriticalcarbondioxide,J.SupercriticalFluids15(1999)117–125. [29]I.Zizovic,M.Stamenic,A.Orlovic,D.Skala,Supercriticalcarbondioxide

extrac-tionofessentialoilsfromplantswithsecretoryducts:Mathematicalmodelling onthemicro-scale,J.SupercriticalFluids39(2007)338–346.

[30]S.BenHadjAhmed,R.M.Sghaier,F.Guesmi,B.Kaabi,M.Mejri,H.Attia,D. Laouini,I.Smaali,Evaluationofantileishmanial,cytotoxicandantioxidant activitiesofessentialoilsextractedfromplantsissuedfromthe leishmaniasis-endemicregionofSned(Tunisia),NaturalProductResearch:FormerlyNatural ProductLetters25(12)(2011)1195–1201.

[31]A.Tekaya-Karoui,N.Boughalle,S.Hammami,H.BenJannet,Z.Mighri, Chem-icalcompositionandantifungalactivityofvolatilecomponentsfromwoody terminalbranchesandrootsofTetraclinisarticulata(Vahl)Mastersgrowingin Tunisia,AfricanJ.PlantScience5(2011)115–122.

[32]A.F.Barrero,M.M.Herrador,P.Arteaga,J.Quilez,M.Akssira,F.Mellouki,S. Akkad,Chemicalcompositionoftheessentialoilofleavesandwoodof Tetra-clinisarticulata(Vahl)Masters,J.EssentialOilResearch17(2005)166–168. [33]M.Aitigri,M.Holeman,A.Ilidrissi,M.Berrada,Contributiontothechemical

studyofessentialoilsfromthetwigsandwoodofTetraclinisarticulata(Vahl) Masters,PlantesMédicinalesetPhytothérapie24(1990)36–43.

[34]S.Zrira,B.Benjilali,A.Elmrani,Chemicalcompositionofthesawdustoilof moroccanTetraclinisarticulata(Vahl),J.EssentialOilResearch17(2005)96–97. [35]J.Buhagiar,M.T.C.Podestat,P.L.Cioni,G.Flamini,L.Morelli,Essentialoil com-positionofdifferentpartsofTetraclinisarticulata,J.EssentialOilResearch12 (2000)29–32.

[36]M.A.HarishNayaka,U.V.Sathisha,A.M.Dharmesh,Cytoprotectiveand antiox-idantactivityoffree,conjugatedandinsoluble-boundphenolicacidsfrom swallowroot(Decalepishamiltonii),FoodChemistry119(2010)1307–1312. [37]C.S.Tan,S.K.Liang,D.C.Liou,Fluid–solidmasstransferinasupercriticalfluid

extractor,ChemicalEngineeringJ.38(1988)17–22.

[38]E.M.C.Reis-Vasco,J.A.P.Coelho,A.M.F.Palavra,C.Marrone,E.Reverchon, Math-ematicalmodelingandsimulationofpennyroyalessentialoilsupercritical extraction,ChemicalEngineeringScience55(2000)2917–2922.

[39]T.Funazukuri,C.Kong,S.Kagei,Effectiveaxialdispersioncoefficientsinpacked bedsundersupercriticalconditions,J.SupercriticalFluids13(1998)169–175.