Colloidal gas aphrons based separation process for the purification and fractionation of natural phenolic extracts

(1)

Pleasecitethisarticleinpressas:Spigno,G.,etal., Colloidalgasaphronsbasedseparationprocessforthepuriﬁcationandfractionationof ContentslistsavailableatScienceDirect

Food

and

Bioproducts

Processing

jo u r n 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 / f b p

Colloidal

gas

aphrons

based

separation

process

for

the

puriﬁcation

and

fractionation

of

natural

phenolic

extracts

G. Spigno

a,∗

_,

_D.

_Amendola

a

_,

_F.

_Dahmoune

b

_,

_P.

_Jauregi

c

a_Institute_of_Oenology_and_Agro-Food_Engineering,_Università_Cattolica_del_Sacro_Cuore,_Via_Emilia_Parmense_84,

29122Piacenza,Italy

b_Laboratory_of_{Biomathematics,}_{Biochemistry,}_Biophysics_and_{Scientometrics,}_Abderrahmane_Mira_University_of

Bejaia,Bejaia06000,Algeria

c_Department_of_Food_and_Nutritional_Sciences,_University_of_Reading,_{Whiteknights,}_PO_Box_226,_Reading_RG6_6AP,

UK

a

b

s

t

r

a

c

t

Followingpreviousstudies,theaimofthisworkistofurtherinvestigatetheapplicationofcolloidalgasaphrons (CGA)totherecoveryofpolyphenolsfromagrapemarcethanolicextractwithparticularfocusonexploringthe useofanon-ionicfoodgradesurfactant(Tween20)asanalternativetothemoretoxiccationicsurfactantCTAB. Differentbatchseparationtrialsinaflotationcolumnwerecarriedouttoevaluatetheinfluenceofsurfactanttype andconcentrationandprocessingparameters(suchaspH,drainagetime,CGA/extractvolumetricandmolarratio) ontherecoveryoftotal andspecificphenoliccompounds.Thepossibility ofachievingselectiveseparationand concentrationofdifferentclassesofphenoliccompoundsandnon-phenoliccompoundswasalsoassessed,together withtheinfluenceoftheprocessontheantioxidantcapacityoftherecoveredcompounds.Theprocessledtogood recovery,limitedlossofantioxidantcapacity,butlowselectivityunderthetestedconditions.Resultsshowedthe possibilityofusingTween20withaseparationmechanismmainlydrivenbyhydrophobicinteractions.Volumetric ratioratherthanthemolarratiowasthekeyoperatingparameterintherecoveryofpolyphenolsbyCGA.

Keywords:Antioxidants;Colloidalgasaphrons;Grapemarc;Phenolics;Surfactants

1. Introduction

Naturalphenoliccompoundshavebeeninvestigatedfortheir antioxidantandantimicrobialactivitytoquitesomeextent and this subject still raises a large interest due to the

Abbreviations: ABTS,2,2-azino-bis(3-ethylbenzothiazoline-6-sulfonicacid)diammoniumsalt;AOC,antioxidantcapacityexpressed asabsorbancepercentinhibitionintheABTSassay;CAE,caffeicacidequivalents;CAPy/CLPy,concentrationofcompoundyinthe col-lapsedaphronphase/liquidphaseattheendofseparationtrials;CGA,colloidalgasaphrons;CTAB,cetyltrimethylammonium-bromide; GAETPI/GAEFI,gallicacidequivalentsbasedontotalphenolindex/FolinIndex;ME,malvidin-glucosideequivalents;My/feed/My/liq,total

amountofcompoundyaddedinthefeed/measuredintheliquidphaseinseparationtrials;QE,quercetinequivalents;REy,recoveryof

compoundyintheaphronphaseinseparationtrials;SF,separationfactor;VAP/VLP/VCGA/Vliquiddrained/Vsample,volumeof:aphronphase (measuredaftercompletecollapse)/liquidphase/CGA/liquiddrainedfromcompletecollapseofCGAintheCGAcharacterisation/standard gallicacidsolutionorextractfedintothecolumninseparationtrials;ε/ε_,_gas_hold-up_from_CGA_{characterisation/effective}_gas_hold-up fromseparationtrials.

∗ _{Corresponding}_author._Tel.:₊₃₉₀₅₂₃_599181;_fax:₊₃₉₀₅₂₃_599232. E-mailaddress:giorgia.spigno@unicatt.it(G.Spigno).

complexnatureofthesubstances.Polyphenolsarethebiggest groupofphytochemicals,withmorethan8000phenolic struc-turescurrentlyknown,andamongthemover4000ﬂavonoids identiﬁed(Löfetal.,2011).Thepolyphenolsnameincludes dif-ferentphenolicclasseswithevenlargelydifferentmolecular

http://dx.doi.org/10.1016/j.fbp.2014.06.002

(2)

Pleasecitethisarticleinpressas:Spigno,G.,etal., Colloidalgasaphronsbasedseparationprocessforthepuriﬁcationandfractionationof structures, such as non-ﬂavonoid phenolic acids,

non-flavonoid stilbenes, anthocyanins,flavonols, flavanols, tan-nins, flavones and flavanones. The majority of studies availableinliteraturedealwiththenutritional,biochemical, orchemicalstructureaspectsofnaturalphenoliccompounds whileafew studiesaddress the physico-chemical and col-loidal properties (Tsao, 2010) which are, however, of high importanceintheunderstandingofthebehaviourand func-tionalityofthesubstances.Oneofthemainfactorspreventing efficient and widespread applications of natural phenolic compoundsistheirlimitedsolubilityinaqueousorlipidmedia (dependingonthephenolicclassandmolecular hydrophobic-ity),withthingsevenmorecomplicatedwhendealingwith naturalphenolicextractswhichcontainmixturesofdifferent phenoliccompounds(Amendolaet al.,2010;Menese etal.,

2013; Spigno et al., 2013). Production ofstandard phenolic

compounds from natural sources, including agro-food by-productswhichrepresentauniquelow-costresource,requires also important purification and fractionation steps with quiteexpensivetechnologies(strategiesincludingsequential extraction,liquid–liquidpartitioningandothercomplexand expensiveprocessesbasedonmembraneorchromatographic technologieshavebeenproposed).Asitconcernsthepolluting characterofnaturalphenoliccompounds,besidesthe intrin-sicantimicrobialproperties,limitedsolubilitycanalsoprevent theirefficientbioremediation. Limitedstability duetoeasy thermalandoxidativedegradationofnaturalphenolic com-poundsduringstorageandprocessingisanotherheavylimit totheirapplication.Finally,bindingofsomenaturalphenolic compoundstomacromolecules,primarilyproteins/enzymes andpolysaccharides,butalsominerals(suchasironand cop-per),mayhaveimportantantinutritionaleffectsorlimittheir bioavailability afteringestion, compromising theirso com-monlyclaimedhealthbenefits,orcompromisingfoodquality (seeprecipitationofproteins/polyphenolscomplexesinwine andfruitjuices).

Besidestheintensiveuseofemulsiﬁersinthefoodindustry

(HasenhuettlandHartel,2010)someworkshavebeenreported

abouttheinteractionsbetweensurfactantsandphenolic com-pounds.Forexample,afewworkshaveshownthepotential ofspeciﬁcsurfactantsinmicellarformto:solubiliseor precip-itatespeciﬁcphenoliccompounds(Löfetal.,2011);altertheir partitioninginoil-in-wateremulsions(Richardsetal., 2002;

SØrensenet al.,2008);enablephenoliccompounds

analyti-caldeterminationexploitingdifferentafﬁnities(Wangetal., 2007);protectphenoliccompoundsfromoxidation(Linetal., 2007); solubilise aromatic compounds (Yoshidaand Moroi,

2000;Weietal.,2012);improvephenoliccompoundsefﬁciency

intopicalformulations(Scognamiglioetal.,2013;Yutanietal., 2012).Otherworkshaveinvestigatedtheuseofsurfactantsfor therecoveryofnaturalphenoliccompoundsfrom

wastewa-ters(Gortzietal.,2008;Katsoyannosetal.,2012).

Surfactantsand colloidalgas aphrons (CGA), which are surfactant-stabilizedmicrobubbles(10–100␮m)generatedby intense stirring of a surfactant solution at high speeds (>8000rpm) (Jauregiand Dermiki,2010),havebeenusedfor manyseparationprocessessuchasproteinandenzyme

recov-ery (Fudaand Jauregi, 2006;Zidehsaraeietal., 2009;Cheng

andStuckey,2012),carotenoidsrecovery(Dermikietal.,2008;

Alvesetal.,2006);recoveryoftoxicwastesfromsoil,removal

ofdyesfrom wastewaters,strippingofdissolvedgasesand removalofdispersedoildropletsfromwater,bubble-entrained ﬂocﬂotation.Dependingonthe typeofsurfactantusedfor thegenerationofCGA,e.g.cationic,anionicornon-ionic,the

outersurfaceofthemicrobubblemaybepositivelycharged, negativelychargedorneutral,towhichoppositelychargedor non-chargedmoleculeswilladsorb,resultingintheireffective separationfromthebulkliquid,andconsequentlythe selec-tivity ofadsorptioncanbe controlled.CGApossess unique properties:(i)highinterfacialareaduetotheirsmallsize;(ii) highstabilitycomparedtoconventionalfoams;(iii)sufﬁcient stabilitytoallowthemtobepumpedfromthegenerationpoint to thepoint ofuse without lossoftheir originalstructure; (iv)CGAcanbeeasilyseparated fromthe bulkliquid with-outmechanicalaid,asopposedtoconventionalliquid–liquid extractionmethodsandtheaqueoustwophaseseparations thatneedcentrifugationforphaseseparation.Ifitis possi-bletousebiodegradableandnon-toxicsurfactants,thiscould resultinanenvironmentallyfriendlyprocesseswhiletheﬁnal productcouldalsobesafeforhumanconsumption.

Previous research carried out by the authors (Spigno

and Jauregi, 2005; Spigno et al., 2010;) demonstrated that

the phenolic acid gallic acid in its anionic form (at neutral–basicpHconditions)canberecoveredfromaqueous solutionsusingCGAgeneratedfromthecationicsurfactant cetyltrimethylammonium-bromide(CTAB).Gallicacid recov-erywasmainlyaffectedbypH,ionicstrength,surfactant/gallic acidmolarratio,mixingconditionsandcontacttime.In fur-therworkitwasdemonstratedthattotalphenoliccompounds couldberecoveredinhighyieldfromagrapemarc(the wine-makingwasteconsistinginseedsandskins,whichistypically removed afterfermentationforredwines,and after press-ing forwhite wines)ethanolic extract(Spigno et al.,2010). Herewetakethisworkfurtherontoinvestigatingthe applica-tionofCGAtotherecoveryofpolyphenolsfromagrapemarc ethanolicextractwithparticularfocusonexploringtheuseof anon-ionicfoodgradesurfactantTween20asanalternative toamoretoxiccationicsurfactant.Thiscouldofferthe addi-tionaladvantageofextractingthepolyphenolsinasurfactant richsolutionwhichcouldconfertheproductimproved solubil-ityproperties.Inparticularthispaperaimsatgettingabetter insightintothefollowingaspectsoftheapplicationofCGA totherecoveryofnaturalphenoliccompoundsfromacrude ethanolicextract:

• Inﬂuence of surfactant type and concentration and processing parameters (such as pH, drainage time, CGA/extractvolumetricandmolarratio)ontherecoveryof totalandspeciﬁcphenoliccompounds.

• Selectiveseparationandconcentrationofdifferentclasses ofphenoliccompoundsandnon-phenoliccompounds. • Inﬂuence of the process on the functional properties

(antioxidantactivity)oftherecoveredcompounds.

2. Materials

and

methods

2.1. Materials

Gallicacid,caffeicacid,quercetin,CTAB (cetyltrimethylam-monium bromide) and Tween 20 were supplied by Fluka (Milan, Italy); ABTS (2,2 -azino-bis(3-ethylbenzothiazoline-6-sulfonicacid)diammoniumsalt)andmalvidin-glucosideby Sigma (Milan, Italy); potassium persulphate by Carlo Erba (Milan,Italy);allotherchemicalswereofanalyticalgrade.

Grapemarcsamplesfromtwodifferentredgrapevarieties (Barberaandoff-skinsfermentedPinotnoir)werekindly pro-videdbywineriesintheNorthernItaly.Thesampleswereoven

(3)

Pleasecitethisarticleinpressas:Spigno,G.,etal., Colloidalgasaphronsbasedseparationprocessforthepuriﬁcationandfractionationof

Table1–Experimentalconditionsoftheseparationtrialscarriedoutwithcolloidalgasaphrons(CGA).

Surfactant Investigatedvariables Fixedparameters Evaluatedoutputs

CTAB •CTABconcentration(0.5–1–2mM) •Drainagetime(3–25min) •pH(2–6)

•VolumetricratioCGA/sample(1–23)

•Extractconcentration (200mg/LGAETPI)

REGAE,REME,SFGAE,SFME,antioxidant

activityofrecoveredaphronphase; liquiddrainage.

Tween20 •Drainagetime •Extractconcentration (200–5000mg/LGAETPI)

•VolumetricratioCGA/sample(4–16) •InsituCGAgeneration

•Tweenconcentration(10mM) •NotmodiﬁedpH

REandSFof:GAETPI,GAEFI,WAE,CAE,

QE,glucose,fructose,potassium. Antioxidantactivityofrecovered aphronphase;liquiddrainage.

CAE:cinnamicacidsequivalents;FI:FolinIndex;GAE:gallicacidequivalents;ME:malvidin-glucosideequivalents;QE:quercetinequivalents; RE:percentrecovery;SF:separationfactor;TPI:totalphenolicsindex.

driedat60◦Cuntilresidualmoisturecontent<5%andthen milled(particlesize<2mm).Phenolicextractswereobtained bysolventextractionwithaqueousethanolaccordingtothe procedurereportedinAmendolaetal.(2010)slightlymodiﬁed. Brieﬂy,125gofpowderwasextractedwith1Lof60%aqueous ethanolkeepingthemixturestirredat3500rpm(mixer Silver-son,L5M)for2hat60◦C(bymeansofanelectricheatingplate). Afterextraction,themixturewascentrifugedat5000rpmfor 10min(CentrifugeALC4237R)andthesupernatant(extract) recovered.

ForthetrialswithBarberamarc,theextractwas freeze-driedand the powder used toreconstitute a concentrated extractinethanol 60%.Theconcentratedextract wasthen dilutedwithwatertohaveaconcentrationof200mg/Loftotal phenolics(astotalphenolicsindex).Thisstepwascarriedout inordertoreducetheethanolcontenttoabout10%tolimitits antifoameffect(Alvesetal.,2006).

ForthetrialswithPinotnoirmarc,theextractwas concen-tratedundervacuumat40◦C(RotavaporBüchiR-114)until removinghalftheinitialvolumeandthendilutedwithwater untilthedesiredconcentrationintotalphenolics.Thisway,it wassimulatedanindustrialapplicationoftheprocesswhere ethanolneedstoberecovered.

2.2. CGAgenerationandcharacterisation

TheCGAweregeneratedfromasurfactantsolutionstirredat 8000rpmwiththeSilversonmixerfor5min.CGAhadbeen characterisedintermsofgashold-upandstabilityforCTAB andTween20surfactant solutionsinthegeneration vessel

(JauregiandDermiki,2010).Howeverinthisworkthegas

hold-upwasmeasuredtakingintoaccounttheeffectofpumping. Todoso,aftergeneration,100mLoftheCGAwerepumped intoavolumetriccylinderandlettodraincompletely.Thegas hold-up(ε)wasthencalculatedfromthemeasureddrained volume(Vliquiddrained)accordingtoEq.(1):

ε= VCGA−V_Vliquiddrained

CGA

(1)

SurfactantconcentrationinthegeneratedCGAwasestimated combiningthedrainedvolumewiththemolarityofthe sur-factantsolutionusedtogenerateCGA.

Stabilitywasmeasuredbymeasuringthedrainagerateand ashalf-life,whichisthetimetaken forhalfoftheoriginal liquidtodrain.

2.3. SeparationwithCGA

Trialswerecarriedoutinaflotationglasscolumn(i.d.0.25m, totalheight0.4m).First,theextractsampleandthentheCGA werepumpedintothecolumnbyaperistalticpump(Watson Marlow505U)fromthebottom.Thevolumetricflowwas reg-ulatedsothatthemixingtimewouldbe3.5–4min.Oncethe columnwasfilled,themixturewasleftstandingforaselected contact(ordrainagetime)beforepumpingouttheseparated bottomliquidphaseandupperaphronphase.Thelatterwas lettodraincompletely(collapsedaphronphase).Thevolumes oftheseparatedliquidphaseandcollapsedaphronphase(VLP

andVAP,respectively)weremeasured.Percentrecoveryofa

speciﬁccompoundyintheaphronphase(REy)wascalculated

basedonthedifferencebetweenthetotalamountofadded

yinthefeed(My/feed)andtheamountofymeasuredinthe

separatedliquidphase(My/liq).

For some experiments the amount of y in the aphron phasewasalsocalculatedandthemassbalancedeviationwas within10%.

Theseparationfactor(SF)wascalculatedtogiveanideaof theafﬁnityofacompoundtotheaphronphasecomparedto theafﬁnitytotheliquidphase,basedontheconcentrations ofthecompoundyintheaphronandintheliquidphase(CAPy

andCLPy)accordingtoEq.(2):

SF= CAPy CLPy

(2)

Thedifferentcarriedoutseparationtrialsaresummarisedin

Table1.

Thevolumetricratiowascalculatedastheratiobetween thevolumeofCGAandthevolumeofthesamplefedintothe column.

The molarratio was calculated as the moles of surfac-tant tothemoles ofthecompound y fedintothe column; totalphenolsweredeterminedasgallicacidequivalentsand anthocyaninsasmalvidinglucosideequivalentsthereforethe molecularweightsofthesecompoundswereusedtocalculate themolarratioscorrespondingtoeachvolumeratio.

Inthetrialswith“in-situ”CGAgeneration,thesurfactant solutionwasmixedwiththeextract,stirredat8000rpmfor 5min and then pumped into the column and the separa-tion was performed asabove. Theinﬂuence ofmixing the feed(extract)withtheCGAonthegashold-upwasassessed calculating an effective gashold-up (ε₎ _from _Eq. ₍₁₎ _using

asVliquiddrainedthedifferencebetweenthetotaldrained

vol-ume and the volume of the sample fed into the column (VLP+VAP−Vsample).Allexperimentswereperformedatleast

(4)

Pleasecitethisarticleinpressas:Spigno,G.,etal., Colloidalgasaphronsbasedseparationprocessforthepuriﬁcationandfractionationof in duplicate and the results in this work are reported as

means±SD.Signiﬁcance(P<0.01)oftheeffectofinvestigated variable on evaluatedoutputs was establishedtreating the dataaccordingtoagenerallinearmodel(univariateANOVA) andapplyingtheposthocTukey’stesttodiscriminatebetween means(IBMSPSSStatistics19).

2.4. Analyticaldeterminations 2.4.1. Phenoliccompounds

Thegrapemarcextractsandtherecoveredliquidandaphron phaseswerecharacterisedforthetotalphenolicscontentand forthecontentofspeciﬁcphenolicclassesaccordingtothe followinganalyses.

• Totalphenolicscontentbytotalphenolindex(TPI)basedon absorbancereadingat280nmandexpressingtheresultsas gallicacidequivalents(GAETPI)bymeansofacalibration

curvewithstandardgallicacid(Amendolaetal.,2010). • Totalphenolicscontent byFolin–Ciocalteuanalysis(Folin

Index),expressingtheresultsasGAEFIbasedona

calibra-tioncurvewithstandardgallicacid(Amendolaetal.,2010). • Total anthocyanins content by absorbance reading at 538nm of the sample diluted with chloridric ethanol and expressing the results as ME (Malvidin-glucoside Equivalents) based on a calibration curve with standard malvidin-glucoside(Amendolaetal.,2010).

• Total cinnamic acids content by absorbance reading at 320nm and expressing the resultsas caffeicacid equiv-alents(CAE) based on acalibration curve withstandard caffeicacid(Spignoetal.,2007).

• Totalﬂavonolscontentbyabsorbancereadingat370nmand expressingtheresultsasquercetinequivalents(QE)(Spigno

etal.,2007).

For all the abovespectrophotometric analyses,different calibrationcurveswithaqueousethanol,CTABorTween solu-tions were prepared and used depending on the analysed sample(initialextract,liquidandaphronphase).

2.4.2. Non-phenoliccompounds

Thegrapemarcextractsandtherecoveredliquidandaphron phaseswerecharacterisedalsoforthecontentofcompounds otherthanphenolics.Afterpreliminarycharacterisation,the followingcompoundswereselectedbasedontheirlevelinthe extracts.

• GlucoseandfructosecontentwasevaluatedbyaMegazyme enzymatickit(K-FRUGL).

• Potassiumcontent wasevaluated byAtomicAbsorptions Spectroscopy(PerkinElmerAAnalyst300,Norwalk,CT,USA) basedoncalibrationcurvewithstandardK.

2.4.3. Antioxidantcapacity

Theantioxidantcapacityofthegrapemarcextractsandthe recoveredliquidandaphronphaseswereassessedbythe rad-icalABTSassaywhichisbasedontheabilityofantioxidants to reduce the radical decreasing its absorbance at 734nm

(Amendolaetal.,2010).Antioxidantcapacitywasexpressed

asabsorbancepercentinhibition(AOC).

3. Results

and

discussion

3.1. SeparationwithCTAB

In previous work bythe authors it was demonstrated that CGAgeneratedfrom thecationicsurfactant CTAB couldbe appliedsuccessfullytotherecoveryofthestandard pheno-licacidgallicacidfromwatersolutions(Spignoetal.,2010).In thisworkitwasfoundthatoptimumrecoverywasobtained withCGAgeneratedfroma1mMCTABaqueoussolutionat neutral to basic pH’s, without buffer salts as these would increase ionic strength which will in turn reduce interac-tions between gallic acidand surfactant resulting in lower recoveries.

HereweappliedtheoptimumconditionsabovetoaBarbera marc extract diluted to 200mg/L GAETPI as in the

experi-mentswithgallicacid.Similarrecoverieswereobtainedfor the extractasforthestandardgallicacid (Fig.1A).Slightly higherrecoverywasobtainedforanthocyaninsthanfortotal phenolics basedonthe volumetricratio(Fig.1A),However, particularlyathighvolumetricratios,higherseparation fac-torswereobtainedforanthocyaninsthanfortotalphenolics whichsuggeststhatanthocyaninshaveahigherafﬁnityfor CGAthantotalphenols(Fig.1B).

Thehigherrecoveriesobtainedwiththerealextract,might havebeenduetothestabilisingeffectofthesampleonthe CGA, as evidentfrom the lower volumes ofdrained liquid phase(Fig.1C).Othercompoundspresentintherealextract suchassugarsmostprobablycontributedtoanincreasein theviscosityoftheliquidwithintheaphronswhichresulted inreduceddrainagerate.

Further separation trials at constant volumetric ratio (15.33) and variable drainage time were conducted(Fig. 2) in order to see if increased drainage led to an increase concentrationoftotalphenolicsand/oranthocyaninsinthe aphron phaseasobservedinourpreviousworkfor asthax-antin(Dermikietal.,2008).Recoverydecreasedatincreased drainage time following a linear trend (r2 _0.985 _and _0.978

for MEand GAE,respectively)which impliedthat phenolic compoundsdrainedtogetherwiththeliquidphase(Fig.2A). However,theseparationfactorincreasedwithdrainagetime particularlyforanthocyaninsfollowinganonlineartrend.A verysimilartrendwasfollowedbytheratioofdrainedliquid phasetothevolumeofcollapsedaphronphase(Fig.2B).This againsupportsthatanthocyaninshavemoreafﬁnityforthe CGAthanpolyphenols.

Overallahigherselectivityforanthocyaninsthanfortotal phenolicswasfound.AlthoughatthepHconditionsofour experiments (pH of the recovered liquid phase 3.5–4.36 in therangeofvolumetricratiosstudied)anthocyaninsarenot expectedtobechargedsincetheﬂaviliumform(cationicform) existsonlyatpH≤2thesemoleculesarepolarisedwithhigh electrondensityinthearomaticringswhichmayexplaintheir afﬁnityforthepositivelychargedCGAgeneratedfromCTAB. However,itisexpectedthathydrophobicinteractionswillalso playanimportantroleintheirseparation.

ToverifythepHeffect,additionaltrialswerecarriedoutat pH2(with1MHCl)inboththeCTABsolutionandtheextract. ThepHreductionledtoasmallincreaseinrecoveryoftotal phenolicsand areductioninthe recoveryofanthocyanins, particularlyatlowvolumetricratios(Fig.3A).Theseparation factorincreasedfortotalphenolicsbutnotforanthocyanins (Fig. 3B). A reduction in recovery can be explained based onrepulsiveelectrostaticinteractionsbetweenthepositively

(5)

Fig.1–Resultsofseparationtrialscarriedoutwithcolloidalgasaphrons(CGA)(generatedfromCTAB1mM)andsample

consistinginstandardgallicacid(GA)solutionsorBarberamarcextract(GAE:gallicacidequivalentsastotalphenolics

index;ME:malvidin-glucosideequivalents).(AandB)Recoveryasafunctionofvolumetricratioormolarratio,respectively.

(CandD)Separationfactororvolumeofdrainedliquidphaseasafunctionofvolumetricratio,respectively.Errorbars

indicate±SD.

chargedcationicsurfactantandanthocyaninsatthispH. How-ever,therecoveryandseparationfactorwerestillhighwhich shows that separation of anthocyanins occurs also due to hydrophobicinteractions;atthispHaproportionof antho-cyaninmoleculeswillnotbechargedandwillbecomplexed withotherphenoliccompounds.

TrialscarriedoutatpH2and0.5mMCTAB,conﬁrmedthe aboveresults,alower recoveryofanthocyaninsatthispH. However,therecoveryofthetotalphenolicsdecreasedalso contrarytowhatwasobservedatpH2with1mMCTAB.This

islikelytobenotapHeffectbut aneffectofCGAstability. Infact,0.5mMisbelowthecriticalmicellarconcentrationof CTAB (reportedtobe0.8–1mMat25◦C),withaconsequent decreaseinCGAstability,ahigherdrainagerateandalower effectivegashold-up.

Furthermore,inallthetrials,itwasfoundthatthe antioxi-dantactivityoftotalphenolicsrecoveredintheaphronphase isreducedcomparedtototalphenolicsdrainedintheliquid phase(Fig.4).Thiscanbeduetooxidationduringrecoveryor toaselectivepartitionofphenolicsbetweenthetwophases.

Table2–Results(intermsofrecoveryefﬁciencyRE)fortheseparationtrialscarriedoutwithcolloidalgasaphrons generatedfromTween2010mM,atdifferentvolumetricratios(mlCGA/mlextract)anddifferentmolarratios

(molTween/molGAETPI).VAPandVLPistherecoveredvolumeofaphronandliquidphase,respectively.

mlCGA/mlextract 16 16 16 8 4

gGAETPI/Lextract 0.3 1.0 5.4 1.0 1.0

molTween/molGAE 90.7 27.2 4.8 13.6 6.8

VLP/VAP 1.25 1.00 0.83 1.25 2.50

Effectivegashold-up 0.53 0.60 0.60 0.60 0.55

RE GAETPI 77.29c 75.59d 78.47b 61.75c 48.25c GAEFI 69.42b 72.86bcd 76.33b 55.40b 41.21b CAE 75.92c _78.90e _84.51c _60.66bc _54.56d ME NE 69.76b _76.36b _68.90d _24.56a QE 70.75b _79.28e _84.28c _60.18bc _54.03d Glucose 44.52a _73.03cd _70.98a _66.09cd _46.99c Fructose 46.20a _71.30bc _70.62a _46.09a _45.34bc Potassium 70.88b _52.38a _72.21a _68.40d _43.92bc

FI/TPI(Aphronphase) 1.88 1.76 2.02 1.12 0.76

FI/TPI(liquidphase) 2.82 1.52 1.78 1.46 1.01

GAE:gallicacidequivalentsaccordingtototalphenolindex(TPI)orFolinIndex(FI).CAE:caffeicacidequivalents.ME:malvidin-glucoside equivalents.QE:quercetinequivalents.NE:notevaluated.Samesuperscriptlettersinthesamecolumnindicatemeansnotstatisticallydifferent accordingtoANOVAandTukey’sposthoctest.

(6)

Fig.2–Resultsofseparationtrialscarriedoutwithcolloidal

gasaphrons(CGA)(generatedfromCTAB1mM)and

Barberamarcextract(GAE:gallicacidequivalentsastotal

phenolicsindex;ME:malvidin-glucosideequivalents)with

aconstantvolumetricratioof15.33mLCGA/mLextract.(A):

recoveryasafunctionofdrainagetime.(B):thevolumetric

ratioofdrainedliquidphasetodrainedaphronphaseasa

functionofdrainagetime.Errorbarsindicate±SD.

3.2. SeparationwithTween20

CGAweregeneratedfromthenon-ionic,food-gradesurfactant Tween20 at a ﬁxedconcentration of10mM. This concen-trationwas used based on characteristics ofCGA suchas, stability(half-life)=407sandgashold-up=60%which,were similartothose ofthe CGAgeneratedfromCTAB,stability (half-life)=493sandgashold-up=63%.

Comparedto the trials withCGA generated from CTAB 1mM (withoutpHcorrection),lower anthocyaninsrecovery andseparationfactorwereobtainedwithCGA(Tween20)than CGA(CTAB),butonlyslightlylowerrecoveryandhigher sepa-rationfactoroftotalphenolics(Fig.5AandB).Theseresults conﬁrmthattheseparationofpolyphenolsisdrivenbyboth electrostaticandhydrophobicinteractionsinthecaseofCGA generatedfromCTABwhilstwithCGAgeneratedfromTween 20,onlyhydrophobicinteractionsdrivetheseparationhence thelowerrecoverywithTween.CGAgeneratedfromTween 20hadalsolowerstabilityandgashold-upthanCGA gener-atedfromCTAB(Fig.5C).Particularlyatlowvolumetricratio(4) Tween20hadveryloweffectivegashold-up(about0.2)which coincideswithlowrecoveryofpolyphenols.Thereductionin gashold-upleadstoareductionininterfacialareahencelower recovery;asimilarobservationwasmadebytheauthorson therecoveryofastaxanthin(Dermikietal.,2010).

Fig.3–Resultsofseparationtrialscarriedoutwith

colloidalgasaphrons(CGA)(generatedfromCTAB1mM

and0.5mM,modifyingornotthepHto2)andBarbera

marcextract(GAE:gallicacidequivalentsastotalphenolics

index;ME:malvidin-glucosideequivalents).(A):recovery;

(B):separationfactorasafunctionofvolumetricratio.Error

barsindicate±SD.

Fig.4–Comparisonbetweentheantioxidantcapacityasa

functionoftotalphenolicsconcentration(GAE:gallicacid

equivalentsastotalphenolicsindex)fortheinitialBarbera

marcandtherecoveredliquidandaphronphase,for

separationtrialscarriedoutwithcolloidalgasaphrons

generatedfromCTAB(withoutpHmodiﬁcation)and

(7)

Fig.5–Resultsofseparationtrialscarriedoutwith

colloidalgasaphrons(generatedfromCTAB1mMorTween

2010mM)andBarberamarcextract(GAE:gallicacid

equivalentsastotalphenolicsindex;ME:

malvidin-glucosideequivalents).(A):recovery;(B):

separationfactorand(C)thevolumetricratioofdrained

liquidphasetodrainedaphronphaseandtheeffectivegas

hold-upasafunctionofvolumetricratio.Errorbars

indicate±SD.

ContrarytowhatwasobservedwithCGAgeneratedfrom CTABnolossofantioxidantactivitywasobservedinrecovered fractionswithCGAgeneratedfromTween20(datanotshown here).

Overall16wasfoundtobeanoptimumvolumetricratio where recovery of both total phenolics and anthocyanins reachedaplateau.Furtherexperimentswerecarriedout in ordertoinvestigatetheapplicabilityoftheprocessto differ-entphenolicextracts,totalphenolicsloadandselectivityof

different classes ofphenoliccompounds and non-phenolic compounds.

ComparedtoBarberaextract,thePinotextractwas charac-terisedbya1.6timeshighercontentoftotalphenolics,but a5times loweramount inanthocyanins(Barberagrapeis, infact,well-knownforitshighanthocyaninscontent). Over-all composition ofPinot extract was: 5.43±0.16g/LGAETPI;

8.74±1.32g/L GAEFI; 2.54±0.03g/L CE; 161.46±2.28mg/L

ME; 8.10±0.086g/L glucose; 10.42±0.16g/L fructose and 767.5±0.14mg/Lpotassium.

Separation experiments were carried out at differ-ent volumetric ratios and different dilutions of the feed extract (0.3–1.0–5.4g/L GAETPI). The results are reported in

Tables2and3.

Very similarrecoveryoftotalphenolicswasobtainedto thatobtainedwithBarberaextract.

When the volumetric ratio was kept constant, it was observed,foralmostallthecompounds,ageneralincreasein therecoverywithdecreasingmolarratio(Table2).Thismight havebeenduetothehigherCGAstabilityobservedwhena moreconcentratedextractwasused;withthestabilisingeffect possiblyduetoothercompoundsintheextractasobserved previouslyfortheBarberaextract.Furthermore,contraryto whatwasobservedwithCGAgeneratedfromCTAB,the volu-metricratiohasamajoreffectontherecoveryinsteadofthe molarratio. Forexample inexperimentsataconstant vol-umetric ratioof16 but varyingmolarratio,the recoveryis almostthesame(Table2).However,experimentsatincreased volumetricratio(4–16)atconstantfeedconcentrationshowan importantincreaseinrecovery(48.25–75.59)ofGAE.Inthe tri-alswithconcentratedextract(molarratio4.8),theformation ofaggregateswasobservedintherecoveredaphron phase. Sincethese aggregatedcouldnotbe completelysolubilised duringsampleanalysis,anunderestimationoftherecovery wasprobablydone.

Insummarymaximumrecoveryisobtainedathigh volu-metricratioandconcentratedextract,whichresultsoptimal also for the selectivity of the process. In fact, looking at the separation factor (Table 3), in these trials it was observedahigheraffinityfortheaphronphasebycinnamic acids and flavonols, which tend to be smaller molecules comparedtoanthocyaninsand otherphenoliccompounds. However,itwasstillnotpossibletoselectivelyseparate antho-cyanins (confirming their probable association with other molecules),ornon-phenoliccompoundssuchassugarsand minerals.

Analysis of antioxidant activity (Fig. 6) conﬁrmed that only a slight reduction in the antiradical activ-ity of the phenolic compounds could be achieved using Tween20.

Theoccurrence ofa certainselectivity and/oroxidation of the compounds duringthe separation is shown alsoby the ratio between the total phenolic index and the Folin Index. In fact, even though both methods are used to quantify total phenolics, the total phenolic index is based on the characteristic absorption of the aromatic ring at 280nm(withdifferentmolarextinctioncoefﬁcients depend-ing on the molecular structure), while the Folin Index is based on the abilityof the compound to reduce the Folin reagentwhichdependsonthemolecularstructureand oxida-tive status of the antioxidant. It follows that a change in the Folin Index/total phenolic index ratio indicates a changeinthecompositionand/oroxidationofthephenolic compounds.

(8)

Table3–Results(intermsofseparationfactor,SF)fortheseparationtrialscarriedoutwithcolloidalgasaphrons generatedfromTween2010mM,atdifferentvolumetricratios(mlCGA/mlextract)anddifferentmolarratios

(molTween/molGAETPI).VAPandVLPistherecoveredvolumeofaphronandliquidphase,respectively.

mlCGA/mlextract 16 16 16 8 4

gGAETPI/Lextract 0.3 1.0 5.4 1.0 1.0

molTween/molGAE 90.7 27.2 4.8 13.6 6.8

VLP/VAP 1.25 1.00 0.83 1.25 2.50

Effectivegashold-up 0.53 0.60 0.60 0.60 0.55

SF SF SF SF SF GAETPI 4.10c 3.00c 3.78b 1.91bc 2.27c GAEFI 2.71b 2.60bc 3.35ab 1.48ab 1.71b CAE 3.82c _3.64d _5.70c _1.84bc _2.93d ME NE 2.24b _3.43b _2.66d _0.79a QE 2.88b _3.74d _6.60d _1.79b _2.88d Glucose 0.95a _2.62bc _3.03a _2.32cd _2.18bc Fructose 1.02a _2.41bc _3.03a _1.01a _2.03bc Potassium 2.89b _1.06a _3.21a _2.56d _1.91bc

Fig.6–Comparisonbetweentheantioxidantcapacityasa functionoftotalphenolicsconcentration(GAE:gallicacid equivalentsasFolinIndex)fortheinitialPinotnoirmarc extractandtherecoveredliquidandaphronphase,forthe differentseparationtrialscarriedoutwithcolloidalgas aphronsgeneratedfromTween20.

3.2.1. InsituCGAgeneration

The last trials were carried out by generating the CGA directly from a mixture of surfactant solution and extract withavolumetricratioof8mLTween/mLextract(corresponding to a theoretic volumetric ratio of 18.56mLCGA/mLextract, based on the gas hold-up of Tween 20 10mM). Two different extract concentrations (0.3 and 5.4g/L GAETPI) were used, corresponding to molar ratio of 45.3 and 2.5molTween/molGAE, respectively. Results are reported

in Table 4. Decidedly lower recovery and selectivity,

com-paredto the conventionalseparation trials, were obtained. In particular, the trials withthe undiluted extract showed a uniform partition of all the analysed compounds in the two phases (separation factors around 1 and the same ratio FolinIndex/totalphenolicindex),whichsuggests that when CGA are generated in the presence of the extract, the solubilisation of polyphenols by the micelles com-petes strongly with the ﬂotation of polyphenols by the CGA.

Table4–Results(intermsofrecoveryefﬁciencyREandseparationfactorSF)fortheseparationtrialscarriedoutwith colloidalgasaphronsgeneratedfromTween2010mMmixedwithPinotnoirmarcextract,atdifferentmolarratios (molTween/molGAETPI).VAPandVLPistherecoveredvolumeofaphronandliquidphase,respectively.

gGAETPI/Lextract 0.3 5.4

molTween/molGAE 45.3 2.5

VLP/VAP 1.69 0.77 RE SF RE SF GAETPI 41.96ab 1.23ab 55.07a 0.94a GAEFI 50.94c 1.78c 55.65a 0.97a CAE 41.69ab _1.21ab _57.12a _1.03a ME 52.44c _1.87c _55.91a _0.98a QE 45.83b _1.43b _55.07a _0.95a Glucose 37.98a _1.04a _58.24a _1.09a Fructose 38.51a _1.06a _48.27b _0.72b

FI/TPI(Aphronphase) 2.12 1.50

FI/TPI(liquidphase) 1.46 1.46

(9)

4. Conclusions

InthisworkwehaveshownthatCGAgeneratedfromCTAB caneffectivelyextractphenoliccompounds. Slightlyhigher recoveriesthan whenappliedtostandard solutionofgallic acidareobtainedwhichisthoughttobeduetoan stabilis-ingeffectbythefeed.Furthermoreslightlyhigherrecoveries andseparationfactorsareobtainedforanthocyanins show-ingaslightlyhigherafﬁnityofCGAgeneratedfromCTABfor thesecompounds.Thisresultstogetherwiththoseobtained atpH=2suggestthatpolyphenolspartitiontotheCGAmainly duetohydrophobicinteractionsbutelectrostaticinteractions enhancethepartitioningasshownbyahigherrecoveryand separationfactor obtainedforanthocyaninsatpH>2. This couldalsoexplainwhylowerrecoveryandseparationfactor wereobtainedforphenoliccompoundswithCGAgenerated fromTween20thanwithCGAgeneratedfromCTABaswith the former partitioning into the CGA phase is driven by hydrophobicinteractionsonly.Maximumrecovery(78%GAE and76% ME)and separationfactor (4 GAEand 3ME)were obtainedatthehighestvolumetricratioandhigherextract concentrationwithCGAgeneratedfromTween20.Moreover, theextractionwithCGAgeneratedfromTween20didnotlead toasubstantiallossofantioxidantcapacitycontrary tothe extractionwithCGAgeneratedfromCTAB.Overall,hereithas beenshownthattheextractionofphenoliccompoundswith CGAgeneratedfromTween20couldbeapromisingalternative tootherseparationssuchasorganicextraction.

Acknowledgment

ThisresearchwaspartlysupportedbyProjectAger,grantn◦ 2010-2222.

References

Alves,R.W.,UlsondeSouza,A.A.,UlsondeSouza,S.M.,Jauregi,P., 2006.Recoveryofnorbixinfromarawextractionsolutionof

annattopigmentsusingcolloidalgasaphrons(CGAs).Sep.

Purif.Technol.48,208–213.

Amendola,D.,DeFaveri,D.M.,Spigno,G.,2010.Grapemarc

phenolics:extractionkinetics,qualityandstabilityof

extracts.J.FoodEng.97,384–392.

Cheng,S.,Stuckey,D.C.,2012.Proteinprecipitationusingan

anionicsurfactant.ProcessBiochem.47(5),712–719.

Dermiki,M.,Gordon,M.H.,Jauregi,P.,2008.Theuseofcolloidal

gasaphronsasnoveldownstreamprocessingfortherecovery

ofastaxanthinfromcellsofPhafﬁarhodozyma.J.Chem.

Technol.Biotechnol.83,174–182.

Dermiki,M.,Bourquin,A.,Jauregi,P.,2010.Separationof

astaxanthinfromcellsofPhafﬁarhodozymausingcolloidalgas

aphrons(CGA)inaﬂotationcolumn.Biotechnol.Prog.26(2),

477–487.

Fuda,E.,Jauregi,P.,2006.Aninsightintothemechanismof

proteinseparationbycolloidalgasaphrons(CGA)generated

fromionicsurfactants.J.Chromatogr.B.843,317–326.

Gortzi,O.,Lalas,S.,Chatzilazarou,A.,Katsoyannos,E., Papaconstandinou,S.,Dourtoglou,E.,2008.Recoveryof

naturalantioxidantsfromolivemillwastewaterusing

Genapol-X080.J.Am.OilChem.Soc.85,133–140.

Hasenhuettl,G.L.,Hartel,R.W.,2010.FoodEmulsiﬁersandTheir

Applications,seconded.Springer,NewYork.

Jauregi,P.,Dermiki,M.,2010.Separationofvalueadded

bioproductsbycolloidalgasaphrons(CGA)ﬂotationand

applicationsintherecoveryofvalue-addedfoodproducts.In:

Rizvi,S.S.H.(Ed.),Separation,ExtractionandConcentration

ProcessesinFood,BeverageandNutraceutical

Industries.WoodheadPublishingLimited,Cambridge,

pp.284–313.

Katsoyannos,E.,Gortzi,O.,Chatzilazarou,A.,Athanasiadis,V., Tsaknis,J.,Lalas,S.,2012.Evaluationofthesuitabilityoflow

hazardsurfactantsfortheseparationofphenolsand

carotenoidsfromred-ﬂeshorangejuiceandolivemill

wastewaterusingcloudpointextraction.J.Sep.Sci.35(19),

2665–2670.

Lin,Q.L.,Wang,J.,Qin,D.,Björn,B.,2007.Infuenceofamphiphilic

structuresonthestabilityofpolyphenolswithdifferent

hydrophobicity.Sci.ChinaSer.B:Chem.50,121–126.

Löf,D.,Schillén,K.,Nilsson,L.,2011.Flavonoids:precipitation

kineticsandinteractionwithsurfactantmicelles.J.FoodSci.

76(3),N35–N39.

Menese,N.G.T.,Martins,S.,Teixeira,J.,Mussatto,S.I.,2013.

Inﬂuenceofextractionsolventsontherecoveryof

antioxidantphenoliccompoundsfrombrewer’sspentgrains.

Sep.Purif.Technol.108,152–158.

Richards,M.P.,Chaiyasit,W.,McClements,D.J.,Decker,E.A.,2002.

Abilityofsurfactantmicellestoalterthepartitioningof

phenolicantioxidantsinoil-in-wateremulsions.J.Agric.Food

Chem.50,1254–1259.

Scognamiglio,I.,DeStefano,D.,Campani,V.,Mayol,L., Carnuccio,R.,Fabbroccini,G.,Ayala,F.,LaRotonda,M.I.,De Rosa,G.,2013.Nanocarriersfortopicaladministrationof

resveratrol:acomparativestudy.Int.J.Pharm.440,179–187.

SØrensen,A.-D.M.,Haahr,A.-M.,Becker,E.M.,Skibsted,L.H., Bergenståhl,B.,Nilsson,L.,Jacobsen,C.,2008.Interactions

betweeniron,phenoliccompounds,emulsiﬁers,andpHin

omega-3-enrichedoil-in-wateremulsions.J.Agric.Food

Chem.56,1740–1750.

Spigno,G.,Jauregi,P.,2005.Recoveryofgallicacidwithcolloidal

gasaphrons(CGA).Int.J.FoodEng.1(4),Article5.

Spigno,G.,Dermiki,M.,Pastori,C.,Casanova,F.,Jauregi,P.,2010.

Recoveryofgallicacidwithcolloidalgasaphronsgenerated

fromacationicsurfactant.Sep.Purif.Technol.71,56–62.

Spigno,G.,Donsì,F.,Amendola,D.,Sessa,M.,Ferrari,G.,De Faveri,D.M.,2013.Nanoencapsulationsystemstoimprove

solubilityandantioxidantefﬁciencyofagrapemarcextract

intohazelnutpaste.J.FoodEng.114,207–214.

Spigno,G.,Tramelli,L.,DeFaveri,D.M.,2007.Effectsofextraction

time,temperatureandsolventonconcentrationand

antioxidantactivityofgrapemarcphenolics.J.FoodEng.81,

200–208.

Tsao,R.,2010.Chemistryandbiochemistryofdietary

polyphenols.Nutrients2,1231–1246.

Wang,X.-K.,He,Y.-Z.,Qian,L.-L.,2007.Determinationof

polyphenolcomponentsinherbalmedicinesbymicellar

electrokineticcapillarychromatographywithTween20.

Talanta74,1–6.

Wei,J.,Huang,G.,Zhu,L.,Zhao,S.,An,C.,Fan,Y.,2012.Enhanced

aqueoussolubilityofnaphthaleneandpyrenebybinaryand

ternaryGeminicationicandconventionalnon-ionic

surfactants.Chemosphere89,1347–1353.

Yoshida,N.,Moroi,Y.,2000.Solubilizationofpolycyclicaromatic

compoundsinton-decyltrimethylammonium

perﬂuorocarboxylatemicelles.J.ColloidInterfaceSci.232,

33–38.

Yutani,R.,Morita,S.,Teraoka,R.,Kitagawa,S.,2012.Distribution

ofpolyphenolsandasurfactantcomponentinskinduring

aerosolOTmicroemulsion-enhancedintradermaldelivery.

Chem.Pharm.Bull.60,989–994.

Zidehsaraei,A.Z.,Moshkelani,M.,Amiri,M.C.,2009.An

innovativesimultaneousglucoamylaseextractionand

recoveryusingcolloidalgasaphrons.Sep.Purif.Technol.67,