Pistacia lentiscus leaves as a source of phenolic compounds: Microwave-assisted extraction optimized and compared with ultrasound-assisted and conventional solvent extraction

(1)

ContentslistsavailableatScienceDirect

Industrial

Crops

and

Products

j ou rn a l h o m epa g e :w w w . e l s e v i e r . c o m / l o c a t e / i n d c r o p

Pistacia

lentiscus

leaves

as

a

source

of

phenolic

compounds:

Microwave-assisted

extraction

optimized

and

compared

with

ultrasound-assisted

and

conventional

solvent

extraction

Farid

Dahmoune

a

_,

_Giorgia

_Spigno

b,∗

_,

_Kamal

_Moussi

a

_,

_Hocine

_Remini

a

_,

Asma

Cherbal

c

_,

_Khodir

_Madani

a

a_Laboratory_of_{Biomathematics,}_{Biochemistry,}_Biophysics_and_{Scientometrics,}_Abderrahmane_Mira_University,_Bejaïa_06000,_Algeria

b_Istituto_di_Enologia_e_Ingegneria_{Agro-Alimentare,}_Università_Cattolica_del_Sacro_Cuore,_Via_Emilia_Parmense,_84-29122_Piacenza,_Italy

c_Laboratory_of_Molecular_Biology,_University_of_Jijel,_Jijel_18000,_Algeria

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received10January2014

Receivedinrevisedform17June2014

Accepted22June2014

Keywords:

PistacialentiscusL.leaves

Phenoliccompounds

Microwaveassistedextraction

Responsesurfacemethodology

a

b

s

t

r

a

c

t

Microwaveassistedextraction(MAE)wasinvestigatedforextractionoftotalphenoliccompounds(TPC expressedasgallicacidequivalentsGAE)fromtheleavesofPistacialentiscusL.withmaximizedtotal phe-nolicyieldusingresponsesurfacemethodology(RSM)coupledwithaBox–Behnkendesign.Theoptimal MAEprocessingparameterswere46%ethanol,extractiontime60s,potencydensity17.86W/mL,and liquid/solidratio28:1,withanextractionyieldof185.69±18.35mgGAE/gdw.TheoptimizedMAEwas

comparedwithanotheremergingtechnology(ultrasoundassistedextraction,UAE)andwith conven-tionalsolventextraction(CSE)givinghigherextractionyieldsofTPC,totalﬂavonoidsandtanninsand comparableantioxidantcapacity(accordingtotheradicalABTSassay).

1. Introduction

PistacialentiscusL.(lentisk)isanevergreenbushthatcanreach

3minhigh,belongingtotheAnacardiaceaefamilythatconsists

ofmorethanelevenspecies.Itislargelydistributedin“extreme”

ecosystemsoftheMediterraneancountriesandhasa large

geo-graphicalandbioclimaticaldistribution,extendingfromthehumid

tothearidareas(LoPrestietal.,2008).

InAlgeria,thetreeiswidespreadinforestaloneorassociated

withothertreespeciessuchasterebinth,olivesandcarob,inall

coastalareasupto700mabovesealevelorinseasidestonyareas.

Itsucceedsinanyordinarygardensoil,preferringahotdryposition

infullsun.Itprefersawell-drainedtodrysandyorstonyalkaline

soil,makingitmoreabundantnearthesea.Italsoshowstolerance

torockyareas,droughtandcold(−7◦_C)_in_winter_together_with

resistancetocalcareoussoilandre-growthaftercuttingﬁreinjuries

(Yildirim,2012).

∗ Correspondingauthor.Tel.:+390523599181;fax:+390523599232.

E-mailaddresses:farid.dahmoune@univ-bejaia.dz(F.Dahmoune),

giorgia.spigno@unicatt.it(G.Spigno),moussikamal2012@yahoo.fr(K.Moussi),

hociner06@live.fr(H.Remini),phytopharmaco@yahoo.fr(A.Cherbal),

khodir.madani@univ-bejaia.dz(K.Madani).

AqueousextractofP.lentiscusL.leavesisaverypopulardrink

in North Africacountriesand is becoming increasingly popular

worldwide,partlybecauseofmore documentedevidenceabout

itsbeneﬁcialhealthproperties(Trabelsietal.,2012).P.lentiscus

L. is a rich source of essential oils (96 components) (Lo Presti

etal.,2008),fattyacidssuchasoleic,palmiticandlinoleicacids

(representingthe50.72%,23.16%and21.75%ofthelipidfraction,

respectively) and polyphenols (Trabelsi et al.,2012). Thelatter

representthe7.5%ofleafdryweightandincludemyricetin

glu-curonide,myricetin3-O-rutinosideandmyricetin3-O-rhamnoside

(20%of totalphenolic content), quercetin3-O-rhamnoside,

del-phinidin3-O-glucoside, cyanidine3-O-glucoside, phenolic acids

suchasgallicacid(3.7mg/gdry weight) and5-O-galloylquinic

acid(9.6mg/gdryweight)(Romanietal.,2002).P.lentiscusL.has

agreatnutritionalandindustrialimportance,particularlyinthe

pharmaceutical industry,and thecurrent interest in thehealth

effectsoflentiskhasstimulatedthedevelopmentofnew

extrac-tionprocessesofphenoliccompoundsfractions.Indeed,thequality

ofpolyphenolextractsandtheirantioxidantcapacitydependsnot

only on thequality of the starting biomass (geographic origin,

climaticconditions,harvestingdateandstorageconditions),but

alsoonthetechnologicalprocessesinvolvedintheirmanufacture

(Nkhilietal.,2009).

Furthermore, a recent work (Gratani et al., 2013) has

esti-matedthepotentialbeneﬁttotheglobalcarboncyclethatwould

http://dx.doi.org/10.1016/j.indcrop.2014.06.035

(2)

comefromextendingtheMediterranean shrublandsthanksto

thegoodyearly carbon dioxidesequestrationof someanalysed

species,includingP.lentiscusL.Restorationprojectsand

reintro-ductionactionsofnativespecies,inparticularinmoredegraded

ordryerareasintheMediterraneanregion,couldthen increase

theavailabilityofbiologicalresiduesassociatedtothese

planta-tions,thereforeremunerativeandsustainableprocessesfortheir

valorizationare lookedfor, suchastherecoveryof antioxidant

substances.

Overthepastdecade,variousnovelextractiontechniqueshave

beenintroducedand investigated,mostof whichwere claimed

toimproveefﬁciency,extractquality,extractiontimeandsolvent

consumption.Theemergingtechniquesavailablearemicrowave

assistedextraction(MAE)(Songetal.,2011),ultrasoundassisted

extraction(UAE)(Carreraetal.,2011),supercriticalﬂuid

extrac-tion(SFE)(Santosetal.,2012)andpressurizedsolventextraction

(Xiaoetal.,2012).MAEhasparticularlydrawnsigniﬁcantresearch

attention in various ﬁelds, such as recovering of active

ingre-dients from plant materials. One of the main advantages of

MAEis the signiﬁcantreduction of processing time (from

sev-eralhourstominutes)comparedtoconventional solventliquid

extraction(CSE)andahighyieldofactivesubstances(Chanetal.,

2011).

AsMAEisinﬂuencedbymanyfactorsincludingirradiationtime,

potencydensity,temperature,typeofsolventaswellasthe

inter-actionsofallthesefactors,astatisticaloptimizationneedstobe

adoptedfordeterminationoftheoptimumoperatingconditions.

Responsesurfacemethodology(RSM)isastatisticalmethodthat

usesquantitativedatafromanappropriateexperimentaldesign

todetermineorsimultaneouslysolvemultivariateequation.RSM

canthengenerateamathematicalmodelandtakeintoaccountthe

possibleinterrelationshipsamong thetest variableswhile

min-imisingthenumberofexperiments(Songetal.,2011)which is

particularlyusefulinanexperimentaldesignwithmorethantwo

factorsallowingfor considerablereductionof runningcost and

time (Cheok etal., 2012).In this study, a Box–Behnkendesign

(BBD)wasadopted for theexperimentalplanningsince BBD is

one of the most efﬁcient designs of experiment methods. An

advantageof the BBDis that it doesnot contain combinations

forwhichallfactorsaresimultaneouslyattheirhighest or

low-estlevels.BBDisthenusefulinavoidingexperimentsperformed

under extreme conditions, for which unsatisfactory results are

oftenobtained.

Tothebestofourknowledge,noliteraturereportexistsonthe

optimizationofMAEprocedurefortheextractionoftotalphenolic

compounds(TPC)fromP.lentiscusL.leaves.Therefore,the

objec-tivesofthecurrentstudywereto:

• investigatetheeffectsofdifferentparametersontheextraction

efﬁciency(interms ofTPCrecovery) byMAEprocessthrough

preliminarysingle-factorexperiments;

• optimizetheMAEconditionsbyRSM;

• comparetheoptimizedMAEprocesswithCSEprocessandwith

UAEasanotheremergingnon-conventionalprocess(intermsof

TPC,ﬂavonoids,condensedtanninsrecovery,antioxidant

capac-ityof therecovered compounds,and process effectontissue

damage).

Furthermore,sinceantioxidantcompoundsrepresent,usually,

aminor fractionofthebiomass, anintegral exploitation ofthe

residueafterextractionshouldbelookedfor,accordingtoa

biore-ﬁneryapproach.Ageneralchemicalcharacterizationofthebiomass

was,then,carriedouttoevaluatetheﬁbrecontent.Lignocellulose

fractionation,infact,couldbeproposedtorecoverthethreemain

fractionshemicellulose,celluloseandlignin.

Finally,foramorecompleteevaluationofMAEimplementation

asapotentialgreentechnology,calculationofitsenergy

consump-tioncomparedtoCSEandUAEwasalsoaddressed.

2. Materialsandmethods

2.1. Plantmaterial

TheleavesofP.lentiscusL.wereharvestedinJune2012from

spontaneous plants in Oued Ghir, Bejaia, located in the North

EastofAlgeria.Thecollectedsampleswereidentiﬁedbythe

Veg-etableEcologicalLaboratoryandthevoucherspecimenshavebeen

depositedattheHerbariumoftheAlgiersUniversity,Algeria.The

sampleswerewashedwithtapwateranddistilledwater,driedin

astaticovenat40◦Cforaboutoneweek,andthengroundinan

electricalgrinder(IKAmodelA11Basic).Thepowderwaspassed

throughstandard125␮msieveandstoredinairtightbagsunder

darkness untiluse. Moisturecontent wasassessed by constant

weightat105◦Candwas5.0±0.5%.Contentinethanol–toluene

extractives, lignin, holocellulose, cellulose and hemicelluloses,

givenonanovendryweightbasis,wasdeterminedaccordingto

themethodsreportedbyAmendolaetal.(2012).

2.2. Reagents

Sodiumcarbonate(Na2CO3),Folin–Ciocalteu’sphenolreagent,

disodium hydrogen phosphate (Na2HPO4), hydrochloric acid

(HCl)andaluminiumchloride(AlCl3·6H2O) wereobtainedfrom

Prolabo (CE); 2,2-azino-bis(3-ethylbenzothiazoline-6-sulphonic

acid) (ABTS), polyvinyl polypyrrolidone (PVPP) and FeSO4

from Sigma–Aldrich (Germany), and gallic acidfrom

Biochem-chemopharma(UK).Allsolventsusedwereofanalyticalgradeand

purchasedfromProlabo(CE).

2.3. Experimentalwork

ForoptimizationoftheMAEprocedure,theinﬂuencesofthe

processparameterswereﬁrstlyseparatelyinvestigatedin

single-factorexperimentstolimitthetotalexperimentalwork(Table1).

Whenonevariablewasnotstudied,itwaskeptconstant.The

con-stantvaluesforirradiationtime,solvent-to-solidratio(S/Sratio),

ethanolconcentrationandmicrowavepowerwere120s,20mL/g,

50%and500W,respectively.Sinceinallthetrialstheleavespowder

amountwaskeptconstantat1g,theS/SratiogivesthemLof

sol-ventandcanbeusedtocalculatethepowerdensity(W/mL)which,

asaconsequence,variedbothinthemicrowavepowerexperiments

andintheS/Sratioexperiments.Onthebasisofthesingle-factor

experimentalresults,majorinﬂuencefactorswereselected.

Forboththesingle-factorexperimentsandthesuccessiveRSM

optimization,only theTPC yield was considered.In fact, since

theFolin assayusedtoquantifyTPC(seeSection2.4.1)actually

measuresthereducingpowerofthecompounds,theresultsare

inﬂuenced bytheoxidative status ofthesample and are,then,

relatedtotheantioxidantqualityoftheextract.

Then, an RSM based on a Box–Behnken design (BBD) was

conductedtooptimizetheprocesses(Table2).A

three-level-four-factorBBDexperimentaldesign wasappliedandthenumberof

experiments(N)requiredwasdeﬁnedaccordingtoEq.(1):

N=2k(k−1)+C0 (1)

wherekisthenumberoffactorsandC0isthenumberofcentral

points(3).

Regressionanalysisofthedatatoﬁt asecond-order

(3)

Table1

Resultsofsingle-factorexperimentsformicrowaveassistedextractionfromP.lentiscusleaves.Resultsarereportedasmeans±SD.Samelettersinthesamecolumnreferto

meansnotstatisticallydifferentaccordingtoANOVAandTukey’stest.TPC,totalphenoliccompounds;GAE,gallicacidequivalents;dw,dryweightofleaves.

Ethanolconcentration Extractiontime Powerdensity Solvent-to-solidratio

%,v/v TPCyield(mgGAE/gdw) s TPCyield(mgGAE/gdw) W W/mL TPCyield(mgGAE/gdw) mL/g W/mL TPCyield(mgGAE/gdw)

20 144.94±2.06b ₃₀ _149.39_±_8.11c ₃₀₀ ₁₅ _158.27_±_4.72b ₁₀ ₅₀ _82.34_±_4.30c 40 160.46±4.63a ₆₀ _179.22_±_6.43a ₄₀₀ ₂₀ _157.67_±_4.75b ₂₀ ₂₅ _163.40_±_5.64ab 60 154.58±3.41ab ₉₀ _168.52_±_5.13ab ₅₀₀ ₂₅ _179.22_±_6.43a ₂₅ ₂₀ _171.98_±_4.49a 80 152.91±4.55ab ₁₂₀ _144.72_±_4.08c ₆₀₀ ₃₀ _166.26_±_4.76ab ₃₀ _16.7 _160.80_±_4.06ab 100 150.14±4.63ab ₁₅₀ _147.28_±_7.87c ₇₀₀ ₃₅ _168.07_±_5.22ab ₄₀ _12.5 _152.70_±_4.98b 180 160.38±4.17bc ₉₀₀ ₄₅ _166.26_±_5.70ab 210 157.37±3.59bc

followinggeneralequation(Eq.(2))whichwas,then,usedto

pre-dicttheoptimumconditionsofextractionprocess.

Y=B0+ k

i=1 Bixi+ k

i=1 Biix2+ k

i>j Bijxixj+E (2)

where Yrepresentsthe responsefunction (inourcase the TPC

yield);B0isaconstantcoefﬁcient;Bi,BiiandBijarethecoefﬁcients

ofthelinear,quadraticandinteractiveterms,respectively,andxi

andxjrepresentthecodedindependentvariables.Thefactorlevels

werecodedas−1(low),0(centralpointormiddle)and1(high),

respectively.Thevariableswerecodedaccordingtothefollowing

equation(Eq.(3)):

xi=Xi−X0

X (3)

wherexiisthe(dimensionless)codedvalueofthevariableXi;X0is

thevalueofXatthecentrepointandXisthestepchange.

Accordingtotheanalysisofvariance,theregressioncoefﬁcients

ofindividuallinear,quadraticandinteractiontermswere

deter-mined.Inordertovisualizetheeffectsofindependentvariables

andtheirmutualinteraction,theregressioncoefﬁcientswereused

togenerate3-Dsurfaceplotsfromtheﬁttedpolynomialequation.

Toverifytheadequacyofthemodel,additionalextractiontrials

werecarriedoutattheoptimalconditionspredictedwiththeRSM

andtheobtainedexperimentaldatawerecomparedtothevalues

predictedbytheregressionmodel.

OptimizedMAEconditionswerethencomparedtoareference

CSEprocedureandtoanotheremergingnon-conventional

extrac-tiontechnique(UAE).

Finally, in order to investigate the inﬂuence of ultrasound,

microwaveandsimplesolventmacerationonthemicrostructureof

thesamplespowder,thepeelpowdersamples(beforeandafterthe

differentextractionprocesses)wereobservedbyscanningelectron

microscopy(SEM).

2.3.1. Microwave-assistedextraction

Adomesticmicrowaveoven(NN-S674MF,Samsung,Malaysia)

with cavity dimensions of 22.5cm×37.5cm×38.6cm and

2450kHz working frequency was used. The apparatus was

equippedwitha digitalcontrol systemforirradiationtime and

microwavepower(thelatterwaslinearlyadjustablefrom200to

1000W).Theovenwasmodiﬁedinordertocondensateintothe

samplethevapoursgeneratedduringextraction.

For extraction, 1gof P. lentiscus L. powder was placed in a

roundbottomﬂaskof250mL(mediumneckof45mm)containing

Table2

Box–Behnkendesignwiththeobservedresponsesandpredictedvaluesforyieldoftotalphenoliccompounds(TPC)referredtodryweight(dw)ofP.lentiscusleavesusing

microwaveassistedextraction.GAE,gallicacidequivalents;S/S,solventtosolid.

Run X1Ethanolconcentration(%,v/v) X2 Power X3Irradiationtime(s) X4S/Sratio(mL/g) TPCyield(mgGAE/gdw)

W W/mL Experimental Predicted 1 40 400 10.00 60 40 175.15 171.62 2 60 600 20.00 60 30 163.70 168.41 3 40 400 13.33 90 30 187.96 193.33 4 60 500 16.67 90 30 175.53 174.48 5 60 500 16.67 30 30 164.75 169.77 6 20 600 20.00 60 30 151.12 146.55 7 40 600 30.00 60 20 175.15 177.67 8 40 500 25.00 30 20 171.01 167.49 9 40 500 12.50 30 40 166.86 162.28 10 60 500 12.50 60 40 147.31 141.65 11 20 500 12.50 60 40 114.96 123.57 12 60 500 25.00 60 20 165.38 159.40 13 40 600 15.00 60 40 158.27 161.54 14 20 500 16.66 90 30 156.17 150.15 15 20 500 25.00 60 20 113.38 119.10 16 40 500 12.50 90 40 168.52 170.42 17 40 600 20.00 30 30 185.92 183.17 18 40 500 25.00 90 20 175.52 178.56 19 40 600 20.00 90 30 202.42 199.34 20 40 400 20.00 60 20 173.04 168.80 21 20 400 13.33 60 30 146.15 139.92 22 20 500 16.67 30 30 135.68 135.75 23 40 400 13.33 30 30 184.49 190.50 24 60 400 13.33 60 30 173.34 176.26 25 40 500 16.67 60 30 179.61 182.70 26 40 500 16.67 60 30 176.32 182.70 27 40 500 16.67 60 30 178.41 182.70

(4)

water–ethanolmixtureatdifferentethanolconcentrations(20,40, 60,80,100%(v/v)).Thesuspensionwasirradiatedatregular inter-valsaccordingtoovenoperation.Dependingonthetrial,adifferent solvent,irradiationtime,microwavepowerandsolvent-to-solid ratiowereused(Tables1and2).Attheendofmicrowave

irradia-tion,thevolumetricﬂaskwasallowedtocooltoroomtemperature

(Dahmouneetal.,2013).Afterextraction,theextractwasrecovered

byﬁltrationonaBüchnerfunnelthroughNo.1Whatmanpaper

andcollectedinavolumetricﬂask.Theextractwasstoredat4◦C

untiluseandanalysedfortheTPC.Theextractobtainedunderthe

optimumconditionsbyRSMwasanalysedalsoforthecontentof

ﬂavonoids,condensedtanninsandfortheantioxidantcapacity.

2.3.2. Ultrasoundassistedextraction

Anultrasonicapparatus(SONICSVibracell,VCX130PB,Stepped

microtipsandprobes,No.630-0422)wasusedforUAEwith

work-ingfrequencyﬁxedat20kHz.Theenergyinputwascontrolledby

settingtheamplitudeofthesonicatorprobe.

Fortheextraction,1gofﬁnepowderwasplacedina250mL

amber glass bottle (Ø×H: 45mm×140mm and cap size of

28mm)containingwater–ethanolmixture;theobtained

suspen-sionwasexposedtoacousticwavesfor15min.Thetemperature

(27±2◦C)wascontrolledcontinuouslybycirculatingexternalcold

waterandcheckingthetemperatureusingaT-typethermocouple

(Cooking,Thermo-Timer,China).Theacousticenergydensity(AED)

employedinthisexperimentwas0.01W/mL.AEDwascalculated

atanamplitudelevelof45␮m,withtemperaturerecordedduring

15minofexperiment(Rawsonetal.,2011).Aftertheextraction,

theextractwasrecoveredandanalysedasreportedinSection2.3.1

fortheoptimizedMAEextract.

2.3.3. Conventionalsolventextraction

Fortheconventionalsolventextraction,1gofﬁnepowderwas

placedinaconicalﬂaskof250mL(Ø×H:51×150mmandcapsize

of38mm),and50mLof60%(v/v)EtOHwereadded.Themixture

waskeptinathermostaticwaterbath(mod.WNB22,Memmert)

withshakingspeed of 110strokes perminute, at60◦C for2h,

accordingtothemethodrecommendedbySpignoetal.(2007).

TheextractwasthenrecoveredandanalysedasreportedinSection

2.3.1fortheoptimizedMAEextract.

2.4. Analyticaldeterminations

2.4.1. Totalphenoliccompounds

ThecontentinTPCoftheP.lentiscusL.leavesextractwas

deter-minedbyFolin’sassayasreportedinliterature(Jaramillo-Flores

etal.,2003).Brieﬂy,100␮Loftheextractwasmixedwith750␮Lof

10%dilutedFolin–Ciocalteureagent.Thesolutionsweremixedand

incubatedatroomtemperaturefor5min.Afterincubation,750␮L

of7.5%sodiumcarbonate(Na2CO3)solutionwasadded.After

incu-bationat25◦Cfor90mintheabsorbancewasmeasuredat725nm

(1cmopticalpath)againstablank(madeasreportedforthesample

butwith100␮Lofsamplesolvent)usingaSpectroScan50UV-Vis

spectrophotometer.

Gallicacidhydratewasusedasstandardforthecalibrationcurve

toexpresstheTPCconcentrationofthesampleasmg/Lofgallicacid

equivalents(GAE).TPyieldwasthencalculatedbasedonthe

sam-pleconcentration,extractvolumeandleavespowderdryweight,

accordingtothefollowingequation(Eq.(4)):

TPyield= mgGAE/L·LExtract

g_dw_leaves_powder (4)

2.4.2. Condensedtannins

Condensed tannins (CT) were estimated according to the

PVPP method (Ribéreau-Gayon et al., 2000). Brieﬂy, 0.2g of

polyvinylpolypyrrolidone (PVPP) was added to 5mL of extract

(diluted25times)andmixedwith15mLofdistilledwater

(acid-iﬁedtopH3withHCl12N).Aftercentrifugationat4600×gfor

10min,theprecipitatewasrinsedwith10mLofdistilledwater,

thenmixedwith20mLofaBuOH–HCl(12N)(1:1,v:v)mixture

containing150mg/LofFeSO4.Afterheatingfor30mininawater

bathat95◦C,theopticaldensity(d1)at550nmwasmeasuredin

a1cmopticalpathcuvette.Theopticaldensityofacontrol(d0),

preparedunderthesameconditionsbutnotheated,wasalso

mea-sured.Thecondensedtanninconcentration(CTC)wascalculated

bythefollowingequation(Eq.(5)):

CTC=273·(d1−d0)=mg/L (5)

Asreportedfortotalphenols,theyieldofcondensedtanninswas

calculatedbasedonthesampleconcentration,extractvolumeand

leavespowderdryweightasreportedin(Eq.(6)):

CTyield= CTC·LExtract

g_dw_leaves_powder (6)

2.4.3. Totalﬂavonoids

Thecontentoftotalﬂavonoids(TF)wasestimatedbytheAlCl3

method(Quettier-Deleuetal.,2000).Brieﬂy,1mLofextractwas

addedto1mLof2%methanolicAlCl3·6H2Oincubatedfor10min

atroomtemperature.Theabsorbancewasthen readat 415nm

(1cmopticalpath).Theresultswereexpressedinmgquercetin

equivalents/gdryweightofﬁnepowderofP.lentiscus(mgQE/gdw).

Quercetin wasused asstandard for thecalibrationcurveto

expresstheTFconcentrationofthesampleasmg/Lofquercetin

equivalents(QE).TFyieldwasthencalculatedbasedonthesample

concentration,extractvolumeandleafpowderdryweight,

accord-ingtothefollowingequation(Eq.(7)):

TFyield=mgQE/L·LExtract

gdwleavespowder

(7)

2.4.4. Antioxidantcapacity

The antioxidant capacity (AOC) of the P. lentiscus L. leaves

extractswasassessedbytheABTSassay(SpignoandDeFaveri,

2009),whichisbasedontheabilityofantioxidantsofinteracting

withtheABTSradical,decreasingitsabsorbanceat734nm.Brieﬂy,

aradicalsolution(7mMABTSand2.45mMpotassiumpersulfate)

waspreparedandincubatedinthedarkatroom-temperaturefor

12–16hbeforeuse. Thissolutionwasthendilutedwithethanol

50%toanabsorbanceof0.705±0.02at734nmandequilibratedat

30◦C.Control,blankandextractsampleswerepreparedconsisting,

respectively,in:2mLofradicalsolution,2mLofradicalsolution

mixedwith20␮Lofsamplesolventand2mLofradicalsolution

mixedwith20␮Lofextract.Absorbanceofthethreesampleswas

readafter6minat734nmagainstethanol50%andtheAOCwas

calculatedaspercentageabsorbancedecreaseaccordingtoEq.(8):

AOC=ABlank·AExtract

AControl ·

100 (8)

Theextracthadtobedilutedwithwaterinordertohaveaﬁnal

AExtract above 0.10. Extract samples were,then, alwaysdiluted

tohaveaTPCcontentintherange40–310␮gGAE/mL.Fromdata

elaboration(AOCplottedversustotalphenolsconcentration),the

concentrationofTPCrequiredtoreach50%radicalinhibition(IC50)

wascalculated.

2.4.5. Scanningelectronmicroscopy(SEM)analysis

Thepowder(beforeandafterextraction)wasobservedunder

SEM(Quanta200,FEIcompany)formorphologicalcharacterization

(Dahmouneetal.,2013).Foursamplesofpowders(untreatedand

(5)

UAE)werecollectedanddrieduntilconstantmassinovenat60◦C

beforeSEManalysis.

Sampleparticleswerefixedonthespecificcarbonfilmsupport,

andtheirshapeandsurfacecharacterswereobservedusingGSED

detectorwithenvironmentalmode(ESEM).

2.5. Statisticalanalysis

Eachextractiontrialandalltheanalyseswerecarriedoutin

trip-licate(induplicatethechemicalcharacterizationofdryleaves)and

allthedatainthispaperhavebeenreportedasmeans±SD.

Inﬂu-enceofeachfactorontheTPCyieldinthesingle-factorexperiment

fortheMAEwasstatisticallyassessedbyANOVAandTukey’spost

hoctestwith95%conﬁdencelevel.

DataobtainedfromtheBBDtrialsfortheMAEwerestatistically

analysedusingANOVAfortheresponsevariableinordertotestthe

modelsigniﬁcanceandsuitability.p<0.05andp<0.01weretaken

assigniﬁcantandhighlysigniﬁcantlevel,respectively.

TheJMP (Version 7.0, SAS)and Design-Expert (Trialversion

8.0.7.1)softwarewereusedtoconstructtheBBDandtoanalyse

alltheresults.

Inﬂuenceofextractiontechnique(MAE, UAEorCSE)onTPC,

TF,CTyieldsandextractAOCwasassessedbyunivariateANOVA

andTukey’sposthoctestformeansdiscrimination(95%conﬁdence

level).

3. Resultsanddiscussion

Biomass characterization gave the following composition of

dryleaves(onadryweightbasis):21.21±0.11%ethanol–toluene

extractives; 36.07±1.15% lignin like fraction; 25.66±1.37%

holocellulose;10.77±0.58%celluloseand14.89±0.80%

hemicel-luloses.Thisis inagreementwithliterature,whichreports that

leavesare tissueshighly specialized for carbohydrate synthesis

(Chapman et al.,2013).The residualbiomassafter polyphenols

extractioncouldthenbeexploitedforthefractionationoftheﬁbre

components(Spignoetal.,2013).

3.1. Microwave-assistedextraction

3.1.1. Single-factorexperiments

Theeffectsofvariousextractionparameterssuchasethanol

per-centage,irradiationtime,microwavepowerandpowerdensity,and

S/SratioontheTPCyieldwerestudiedusingthe

one-variable-at-a-timeapproach.

Regardingtheinﬂuenceofethanolconcentration,selectionof

extractionsolventiscriticalasitwilldeterminetheamountand

typeofextractedphenoliccompounds.Acetoneandaqueous

alco-hols,particularlyethanolandmethanol,arethemostcommonly

employedsolventsinphenolicextractionfrombotanical

materi-als,morethanthecorrespondingmono-componentsolventsystem

(Inglettet al.,2010; Spignoet al.,2007).However,organic

sol-vents,suchasmethanolandacetone,aretoxicandnotacceptable

forfoods.Theuseofethanolhasseveraladvantagesovertheuse

ofothersolvents,includinghigherextractionefﬁciency,

environ-mentalcompatibilityand lowertoxicity andcost.However,the

percentageofethanolinwaterasanextractionsolventcanaffect

theextractionefﬁciency(TabarakiandNateghi,2011;Wuetal.,

2011).Thatiswhyaqueousethanolatdifferentpercentageswas

testedinthisstudy.

Table1showsthattheTPCyieldraisedsigniﬁcantlyonlywhen

theconcentrationofethanolincreasedfrom20to40%.Athigher

concentrationstheyieldswerestatisticallythesameasthatatboth

20and40%.Asimilareffectwasreportedfortheextractionof

phe-noliccompoundsfromotherplantsources(Spignoetal.,2007;Li

etal.,2012;Panetal.,2003).Thisphenomenoncouldbeexplained

bythefactthatwiththeincreaseinEtOHconcentration,solvent

polaritydeclinedandmolecularmovementdecreased,whichled

tolightdissolutionofphenoliccompoundsfortheloweringof

dif-fusioncoefﬁcientandthedecreaseofsolubility(Yangetal.,2009).

Relatedtopolarityandthenmicrowaveheatingmechanism,the

dielectricconstantofwaterishigherthanethanol(ˆε=80.4and

24.3,respectively),whilethedielectriclossfactorislower(22.8

comparedto8.52at25◦C)which indicatesaslowermicrowave

energyabsorptionandahigherpenetrationdepth.Therefore,

glob-allythepositiveeffectsofintermediateethanolconcentrationscan

beattributedtobothanincreasedsolubilityofphenoliccompounds

andareducedheatingofthemixturewithalimitedthermal

degra-dationoftherecoveredcompounds.Duetotheanalyticalprinciple

oftheFolinassayusedtoquantifyTPC(ascommentedinSection

2.3), lower valuesmaysuggest thatintense heatinghascaused

degradationofthebioactivecompounds.

Based ontheseresults,theconcentrationrange 40–60%was

selectedfortheRSMtrialsand50%wasﬁxedforthenext

single-factorexperimentsontheinﬂuenceofmicrowaveirradiationtime.

AsshowninTable1,a signiﬁcantincreasein extraction

efﬁ-ciency wasobserved as the extractiontime increasedfrom 30

to 60/90sto be followed by a signiﬁcantdecrease after120s.

Longerirradiationexpositionwithouttemperaturecontrol

prob-ablyinducedthermaldegradationofphenoliccompounds(Yang

et al.,2009).Since shorter extractiontime is alsofavourableto

reduceenergycosts,the30–90srangewasselectedfortheRSM

trials,while120swaskeptforthenextsingle-factortrials.

Intheevaluationoftheeffectofadifferentirradiationpower,

thesameS/Sratiowasmaintainedgivingthecorrespondingpower

density reported in Table 1. Power density signiﬁcantly

inﬂu-encedtheTPC yieldinthetestedcondition.Theyieldincreased

withincreasingmicrowavepowerfrom300to500Wandthenit

remainedconstantorslightlydecreasedforhigherpowers

(aver-agevaluesat600/700/900Wwerestatisticallynotdifferentfrom

bothvaluesat400and500W).Itmustbeunderlinedthatthe

oper-atingtemperaturecouldnotberegulatedintheusedequipment

and thatfora constantsample sizetemperatureincreaseswith

microwavepower(SpignoandDeFaveri,2009).Sinceithasbeen

reportedthatthemaineffectofmicrowaves,andinmanycasesthe

only,istheheatingeffect(Kappeetal.,2013),theobtainedresults

wereprobablyduetoanenhancementofphenolsrecoverydueto

aheatingeffect,withconsequentincreaseofmasstransfer

phe-nomena,uptoacertainpowerdensityvalueand,then,tothermal

degradationofbioactivecompoundsathigherdensities(Lietal.,

2012).

Therange400–600WwasselectedfortheRSMstudy,whilethe

500Wpowerwasusedforthelastsingle-factortrialsdealingwith

avaryingS/Sratio,which,broughtalsotohaveavaryingpower

density.TheresultsataS/Sof10conﬁrmedwhatpreviously

sup-posedsincethepowerdensitywashigher(50W/mL)andtheTPC

yieldwasverylow.DuetothesmallsizeofthesamplewiththisS/S,

itswidthwasverycloseorevenlowerthanthemicrowave

pene-trationdepth(apenetrationdepthof1.4and0.42cmisreported

forwaterandethanolrespectivelyat25◦C(www.pueshner.com;

Horikoshietal.,2009).AsalreadyreportedbySpignoandDeFaveri (2009)thiscausessigniﬁcantamountsofelectricﬁeld reachthe

backfaceofthesamplesandarereﬂected,sothattheelectricﬁeld

inthesampleisincreasedmanyfoldsandthesolventheatingrate

increasesdramaticallywithaconsequentthermaldegradationof

thecompounds,whichexplainsthelowmeasuredTPCyield.For

alltheothertestedvalues,theTPCyieldswerealmostnon

statisti-callydifferentwithanincreasingtrendforS/Sratiofrom20to25

andadecreasingtrendfrom25to40.Theobservedresultscanbe

explainedbythecontemporaryvariationofthetwofactorsS/Sratio

andpowerdensity.AsalsoreportedbySpignoandDeFaveri(2009),

(6)

duetoanincreasedconcentrationgradientwhichdrivesthemass

transferofthesolutesfromthesolventimpregnatedonthesolid

particlesintotheexternalsolvent(thisisthesameeffectthatcan

beobservedinCSE).However,inMAE,whenthesolventamount

increasesalsothesamplesizeincreasesandifthemicrowavepower

iskeptconstant,thepowerdensityisreducedandthetotal

sam-pleheatingisreducedaswell.Sincetemperaturedecreaseleadsto

reductioninmasstransfercoefﬁcients,atacertainliquid/solidratio

thiseffectcouldnot becompensatedanymorebytheincreased

concentrationgradient.Thisresultunderlinestheimportanceof

consideringinmicrowaveheatingnotonlythemicrowavepower

butalsothesamplesizewhencombinedwithpower,determines

themoreimportantfactorofpowerdensity.

The20–40S/SratiorangewasﬁnallyselectedfortheRSMtrials.

3.1.2. OptimizationbyRSM

Table2showstheresultsofTPCrecoveryobtainedintheBBD

experimentsandthecorrespondingpredictedvaluesaccordingto

theappliedsecond-orderregressionmodel.

ThemathematicalEq.(9)correlatingtherecoveryofTPCwith

MAEprocessvariablesis givenbelowin termsof codedfactors

excludingnon-signiﬁcantterms:

TPCyield=178.11+13.85x1+4.26x3−27.46x21+5.95x23

−14.55x24 (9)

Inordertodeterminewhetherthequadraticmodelis

signif-icant,it is necessarytorun ANOVAanalysis.The resultsofthe

secondorderresponsesurfacemodelﬁttingintheformofANOVA

aregiveninTable3.TheANOVAdemonstratedthemodeltobe

sig-niﬁcantasevidentfromFvalueandpvalue(Yangetal.,2009).The

determinationcoefﬁcient(R2₌_0.95)_indicates_that_only_5%_of_the

totalvariationsisnotexplainedbythemodel.Foragoodstatistical

model,theadjusteddeterminationcoefﬁcientR2

adjshouldbeclose

toR2_._In_our_model_it_was_0.90_and_then_quite_close_to_R2_.

More-over,R2

pred0.81isinreasonableagreementwithR2adjandconﬁrms

thatthemodelishighlysigniﬁcant.Theregressioncoefﬁcientfrom

experimentaldataandtheadjustedonewerereasonablycloseto1,

whichindicatedahighdegreeofcorrelationbetweentheobserved

andpredictedvalues.Atthesametime,alowvalueofthe

coefﬁ-cientofvariationindicatedahighdegreeofprecisionandagood

dealofreliabilityoftheexperimentalvalues.Thelackofﬁt test

determinesifthemodelisadequatetodescribetheexperimental

dataorwhetheranothermodelshouldbereselected.Thevalueof

lackofﬁttest(0.0539)ishigherthan0.05whichisnotsigniﬁcant

relativetothepureerrorandindicatesthattheﬁttingmodelis

adequatetodescribetheexperimentaldata.Anadequateprecision

isameasureofthesignaltonoiseratio,whichgreaterthan4is

consideredtobedesirable(Canettierietal.,2013).Inourmodel

thevalueofadequateprecisionwas15.98,demonstratingan

ade-quatesignal.Atthesametime,arelativelylowvalueofcoefﬁcient

ofvariation(CV)(2.91)indicatesa betterprecisionand

reliabil-ity.Therefore,theobtainedmodelisadequateforpredictioninthe

rangeofexperimentalvariables.

Thesigniﬁcanceofeachcoefﬁcientmeasuredusingp-valueand

F-valueislistedin Table3.Smallerp-valueand greaterF-value

meanthecorrespondingvariableswouldbemoresigniﬁcant.The

p-valueofthemodelislessthan0.0001,whichindicatesthatthe

modelissigniﬁcantand canbeusedtooptimizetheextraction

variables.

Theeffectsoftheindependentvariablesandtheirmutual

inter-actionsontheTPCyieldcanbevisualizedonthethreedimensional

responsesurfaceplotsandtwodimensionscontourplotsshown

inFigs.1and2,respectively.Theplotsweregeneratedbyplotting

theresponseusingthez-axisagainsttwoindependentvariables

whilekeepingtheothertwoindependentvariablesattheirzero

level(Hayatetal.,2009).Each3Dplotrepresentsthenumberof

combinationsofthetwo-testvariable. 3Dresponsesurface and

2Dcontourplotsarethegraphicalrepresentationsofregression

equationandarevery usefultojudgetherelationship between

independentanddependentvariables.Differentshapesofthe

con-tourplotsindicatewhetherthemutualinteractionsbetweenthe

variablesaresigniﬁcantornot.Circularcontourplotmeansthe

interactionsbetweenthecorrespondingvariablesarenegligible,

whileellipticalcontoursuggeststheinteractionsbetweenthe

cor-respondingvariablesaresigniﬁcant(Liuetal.,2013).

TherecoveryofTPCmainlydepends ontheethanol

concen-trationasitsquadraticandlineareffectswerehighlysigniﬁcant

(p<0.001),conﬁrmingthesingle-factorexperimentresults.

Extrac-tionofphenolicscouldbeenhancedusinganaqueousethanolover

alimitedcompositionalrange.Thisﬁndingwasinagreementwith

ourpreviousconclusionsonextractionofnaturalphenolic

com-pounds fromCitrus limonpeels(Dahmouneet al., 2013).Other

literatureworkshaveshownthesigniﬁcantinﬂuenceoftheethanol

percentageonthephenolicextractionfromplantmaterials,suchas

Rosmarinusofﬁcinalis(ˇSvarc-Gajicetal.,2009)andgreentealeaves

(Panetal.,2003).Inthelattercase,theextractionwasincreased

Table3

Analysisofvariance(ANOVA)fortheexperimentalresultsobtainedbyusingmicrowaveassistedextraction.

Source Sumofsquares Degreesoffreedom Meansquare F-value p-ValueProb>F

Model 10,293.45 14 735.24 17.41 <0.0001 X1-Solvent 2303.64 1 2303.64 54.57 <0.0001 X2-Power 1.04 1 1.04 0.02 0.8776 X3-Time 217.83 1 217.83 5.16 0.0423 X4-Ratio 149.97 1 149.97 3.55 0.0839 X1X2 53.36 1 53.36 1.26 0.2828 X1X3 63.94 1 63.94 1.51 0.2420 X1X4 96.50 1 96.50 2.28 0.1564 X2X3 42.51 1 42.51 1.01 0.3354 X2X4 90.15 1 90.15 2.13 0.1696 X3X4 2.04 1 2.04 0.04 0.8295 X2 1 4024.13 1 4024.13 95.34 <0.0001 X2 2 259.10 1 259.10 6.13 0.0291 X2 3 188.87 1 188.87 4.47 0.0560 X2 4 1128.61 1 1128.61 26.73 0.0002 Residual 506.49 12 42.20 Lackofﬁt 500.91 10 50.09 17.96 0.0539 Pureerror 5.57 2 2.79 Cortotal 10,799.95 26 R2₌_0.95 _Adj_R2₌_0.90 _Pred_R2₌_0.81 _C.V._%₌_2.91

(7)

Fig.1. ResponsesurfaceanalysisforthetotalphenolicyieldfromP.lentiscusleaveswithmicrowaveassistedextractionwithrespecttoethanolpercentageandmicrowave power(A);irradiationtimeandethanolpercentage(B);solvent-to-solidratioandethanolpercentage(C);microwavepowerandirradiationtime(D)solvent-to-solidratio andmicrowavepower(E)solvent-to-solidratioandirradiationtime(F).

whentheproportionofethanol/watersolventwaslowerthan50% (v/v)anddecreasedwhenhigherthan50%(v/v).

Noneoftheinteractivetermsofthemodelwassigniﬁcantand Figs.1and2showthattheTPCcouldbemaximizedusingabout45%

ethanoloverarangeoftheotheroperationalfactors(microwave

power,irradiationtimeandS/Sratio).

The linear term of irradiation power was signiﬁcant, while

(8)

Fig.2.ContourplotsforthetotalphenolicyieldfromP.lentiscusleaveswithmicrowaveassistedextractionwithrespecttoethanolpercentageandmicrowavepower

(A);irradiationtimeandethanolpercentage(B);solvent-to-solidratioandethanolpercentage(C);microwavepowerandirradiationtime(D)solvent-to-solidratioand

microwavepower(E)solvent-to-solidratioandirradiationtime(F).

signiﬁcantagaininagreementwiththepreliminarytrials.Thisis

particularlyevidentforthemicrowavepower(p=0.8776),since,

aspreviouslycommented,thisparametershouldbemoreproperly

consideredincombinationwithsamplesize,thatistosaylooking

atthepowerdensity.Thequadratictermsweresigniﬁcantforall

thevariables,exceptirradiationtime.

Alsoinanotherliteraturestudyonanalytical-scaleMAE(Wang

etal.,2010)itwasfoundthattheinteractionsbetweenmicrowave

powers,extractiontimeandethanolproportionwerenot

signif-icant butthe tendencywas reversed withlinear andquadratic

effect.

In conclusion, using an adequateethanol concentration,the

otherfactorscansetatdesired levelsinorder tominimizethe

energyconsumptionandenvironmentalimpactoftheprocess

(sol-ventamount,irradiationpowerandtime)whichisfundamentalfor

thepotentiallargescaleapplicationofMAE.

Usingthederivedmodel(Eq.(9)),theoptimalMAEconditions

fortheTPCyieldwereobtained:ethanolconcentration46%;

irra-diationtime60s;microwavepower500WandS/Sratio28mL/g

(correspondingtoa powerdensityof 17.86W/mL).Under

opti-mal conditions, the model predicted a maximum response of

180.43±18.05mgGAE/gdw.

To validate the predictability of the established model, the

optimizedparametersweretestedinanadditionalexperiment.A

meanvalueof185.69±18.35mgGAE/gdwwasobtainedwhichwas

foundtobenotsigniﬁcantlydifferentthanthepredictedoneat

p>0.05usingapairedt-test(Hossainetal.,2012),conﬁrmingthat

themodelwasadequateforreﬂectingtheexpectedoptimization

(Wangetal.,2010).

3.2. ComparisonbetweenMAE,UAEandCSE

TheoptimizedMAEconditionswerecompared withanother

emergingextractiontechnique(UAE)andwithaCSE.Theoperating

conditionsselectedforUAE(15min,40%ethanol,energydensity

2.6W/mLandS/Sratio50)comefromnotyetpublishedtrialson

theapplicationofultrasoundsextractiontoP.lentiscusL.leaves.

TheparametersadoptedinCSE(120min,60%ethanol,S/Sratio50)

weretakenbythemethodofSpignoetal.(2007).Theethanol

con-centrationwasnotexactlythesame,butintheMAEsingle-factor

experiments;therewasnotahighlysigniﬁcantdifferencebetween

40and60%.Inaddition,theS/SratiowaslowerintheMAEbut

itmustbesaidthat intheUAEandCSEprotocolsthe

tempera-turewaskeptconstant;therefore,ahigherS/Sratioshouldonly

enhancephenolsextraction.Furthermore,sinceoneofthemost

commonlyappreciatedadvantagesforMAEisthereductionin

sol-ventconsumption,wepreferredtohavealowerS/Sratioforthis

technology.

Theresultsshowthat thethreeinvestigatedextraction

tech-niquesgave statistically comparable TPCyields (185.69_±18.35,

142.76±19.98,178.00±19.80mgGAE/gdwforMAE,UAEandCSE

respectively).TheTFyield wassigniﬁcantlydifferentforallthe

threeprocessesandhigherwithMAE(5.16±0.22mgQE/gdw)than

withUAEandCSE(4.61±0.02and4.79±0.03mgQE/gdw,

respec-tively). Also the CT yield was signiﬁcantly higher with MAE

(40.21±1.76mg/gdw)thanwithUAE(35.94±1.13mg/gdw)orwith

CSE(31.15±3.88mg/gdw).

Different total phenolic yield extraction values from leaves

(9)

Fig.3.ScanningelectronmicroscopeimagesofP.lentiscusleavespowderbefore(A)andafterextractionbyultrasoundassistedextraction(B),conventionalsolventextraction

(C)andmicrowaveassistedextraction(D).

appliedextractiontechniqueandsolvent.Karabegovi ´cetal.(2014)

found,inagreementwiththepresentwork,thattheextractsfrom

cherrylaurelleavesobtainedbyMAEcontainedthehighestamount

ofphenolicandﬂavonoidcompoundscomparedtoUAEandCE.

Theyrecoveredupto30mgGAE/gdw.Nouretal.(2014)extracted

about 40mgGAE/gdw of TPC from blackcurrant leaveswith 40%

aqueousethanol.LowerTPCyield,6.2mgGAE/gdw,wasreportedby

López-Mejíaetal.(2014)fortheTPCextractionofamaranthleaves.

The AOC of the obtained extract was comparable for

MAE and CSE extracts (IC50 of 126.19±1.88 and 130.59±

1.33␮gGAE/mL,respectively)and higherthanUAEextracts(IC50

189.69±2.18␮gGAE/mL).TheloweractivityofUAEextractcould

betheresultofsomeultrasoundrelatedeffect.

MAEhastheadvantageofareduceduseofsolventandashorter

extractiontimeparticularlycomparedtoCSE.Aroughestimation

ofenergyconsumptioninthethreesystemswascarriedout.

InthecaseofMAEandUAE,thenominalpowerdensities(17.9

and2.6W/mL,respectively)andextractiontimeswereconsidered

togetherwithasolventvolumeof50mL,obtainingsimilarenergy

consumptions:107.4and117kJforMAEandUAE,respectively.For

CSEtheenergyneedstoheatinitiallythe50mLofsolventcanbe

estimatedin6.7kJ,basedonasolventdensityof0.887g/mL,a

sol-ventspeciﬁcheatof3.77Jg−1 ◦C−1 (PerryandGreen, 1998)and

aninitialsolventtemperatureof20◦C.However,theefﬁciencyof

theusedheatingsystem,plustheenergyconsumptionsfor

sam-plemixingandtemperaturemaintenanceduringthe2hextraction

shouldbealsoadded,thereforeitisverydifﬁcultanaccurateenergy

costscomparisonfortheCSE.

Finally,theplantmaterialstreatedbyMAE,UAEandCSEwere

examinedbySEMtoinvestigatetheeffectofthedifferent

extrac-tionmethodsonthephysicalstructureoftheﬁnepowder(Fig.3).

Obviousfracturalchangesinthesurfacemorphologybecame

evi-dentafter60sofmicrowaveirradiationenergy(Fig.3D)and15min

ofultrasound treatment(Fig.3B); whilenoseverefracture was

observedduringtheconventionalextraction(Fig.3C)exceptfew

slightrupturesonthesurfaceofthesample.Thesurfaceofthe

sam-pleafterMAEwasfoundgreatlydestroyedsuggestingmicrowave

irradiationplayed animportantrolein breakingupvegetalcell

walls. In addition, this phenomenon suggests that microwave

actionaffectsthephysicalstructureofthecellduetothe

molec-ularmovementandrotationofliquidswithapermanentdipole,

leadingtospeedy heatingoftheethanol/water solvent.Sudden

temperatureriseandinternalpressureincreaseisdifferentfrom

thatofCSEwhichdependsonaseriesofpermeationand

solubi-lizationprocessestobringthechemicalsoutofthematrix.Incase

ofUAE,theswellingandsofteningprocessofthecellwallwasalso

observed(Fig.3B)probablyviathehydrationofpectinousmaterial

frommiddlelamella,whichmightleadtothebreak-upofvegetal

tissueduringsonication(Maetal.,2008).

4. Conclusions

ThispresentstudyindicatesthatP.lentiscusL.leavescanbe

consideredasagoodsourceofphyto-pharmaceuticalinterest

com-poundssinceitwaspossibletorecoverupto18.6%(ondryweigh

(10)

MAEwasoptimized througha RSMbasedapproach andthe

optimaldeterminedconditionsresulted46%aqueousethanolas

solvent,17.86W/mLaspowerdensity,60sasirradiationtimeand

28:1S/Sratio.Theextractobtainedundertheseconditionsshowed

atotalphenoliccompoundsyieldcomparabletotheCSEandUAE

extracts,anaverage10%highertotalﬂavonoidsyield,anda10–30%

highercondensedtanninsyieldcomparedtoUAEandCSE,

respec-tively.TheantioxidantcapacitywascomparabletothatoftheCSE

extract,buta34%higherthanfortheUAEextract.

SEMobservationoftheextractionresiduesshowedthatplant

tissuewasgreatlydestroyedafterMAE.Thisprobablyenhanced

thefastdiffusionofthecompoundsintothesolvent.

EventhoughtheoptimizedMAEprocessallowsfordeﬁnitely

shorterworkingtimeandlowerS/SratiothanCSEandUAE,this

doesnot automatically mean energy savings, since microwave

heatingneedselectricalenergythatisitsmostexpensiveform,so

apunctualenergycostanalysisshouldberequiredtocomparethe

conventionaland MAEextractionsystems(aroughenergy

con-sumptionestimationshowedsimilarvaluesforMAEandUAEbut

itwasnotpossibletoaccuratelyevaluatethevaluefortheCSE).

Anyway,fromthepointofviewofindustrialapplication,this

researchcouldbethebasisforfurtherpilot-planttrialsofMAEas

agreenextractiontechnologyfortherecoveryofhigh-addedvalue

compoundsfrombiomassresidues.Implementationofthefound

optimalconditionsinbatchmicrowaveextractorsequippedwith

mixingdeviceorincontinuousmicrowaveapplicators,shouldbe

investigated.

References

Amendola,D.,DeFaveri,D.M.,Egües,I.,Serrano,L.,Labidi,J.,Spigno,G.,2012.

Autohydrolysisandorganosolvprocessforrecoveryofhemicelluloses,phenolic compoundsandligninfromgrapestalks.Bioresour.Technol.107,267–274.

Canettieri,E.V.,Rocha,G.J.M.,Carvalho,J.A.,Silva,J.B.A.,2013.Optimizationofacid

hydrolysisfromthehemicellulosicfractionofEucalyptusgrandisresidueusing responsesurfacemethodology.Bioresour.Technol.98,422–428.

Carrera,C.,Ruiz-Rodríguez,A.,Palma,M.,Barroso,C.G.,2011.Ultrasoundassisted

extractionofphenoliccompoundsfromgrapes.Anal.Chim.Acta32,100–104.

Chan,C.H.,Yusoff,R.,Ngoh,G.C.,Kung,F.W.,2011.Microwave-assistedextractions

ofactiveingredientsfromplants.J.Chromatogr.A1218,6213–6225.

Chapman,K.D.,Dyer,J.M.,Mullen,R.T.,2013.Commentary:Whydon’tplantleaves

getfat?PlantSci.207,128–134.

Cheok,C.Y.,Chin, N.L.,Yusof,Y.A.,Talib,R.A.,Law,C.L.,2012. Optimizationof

totalphenoliccontentextractedfromGarciniamangostanaLinn.hullusing responsesurfacemethodologyversusartiﬁcialneuralnetwork.Ind.CropsProd. 40,247–253.

Dahmoune,F.,Boulekbache,L.,Moussi,K.,Aoun,O.,Spigno,G.,Madani,K.,2013.

ValorizationofCitruslimonresiduesfortherecoveryofantioxidants:evaluation andoptimizationofmicrowaveandultrasoundapplicationtosolventextraction. Ind.CropsProd.50,77–87.

Gratani,L.,Varone,L.,Ricotta,C.,Catoni,R.,2013.Mediterraneanshrublandscarbon

sequestration:environmentalandeconomicbeneﬁts.Mitig.Adapt.Strat.Global Change18,1167–1182.

Hayat,K.,Hussain,S., Abbas,S.,Farooq,U.,Ding,B.,Xia,S.,2009. Optimized

microwave-assistedextractionofphenolicacidsfromcitrusmandarinpeelsand evaluationofantioxidantactivityinvitro.Sep.Purif.Technol.70,63–70.

Horikoshi,S.,Abe,M.,Serpone,N.,2009.Inﬂuenceofalcoholicandcarbonylfunctions

inmicrowave-assistedandphoto-assistedoxidativemineralization.Appl.Catal. B89,284–287.

Hossain,M.B.,Brunton,N.P.,Patras,A.,Tiwari,B.,O‘Donnell,C.P.,Martin-Diana,A.B.,

Barry-Ryan,C.,2012.Optimizationofultrasoundassistedextractionof

antioxi-dantcompoundsfrommarjoram(OriganummajoranaL.)usingresponsesurface methodology.Ultrason.Sonochem.19,582–590.

Inglett,G.E.,Rose,D.J.,Chen,D.,Stevenson,D.G.,Biswas,A.,2010.Phenolic

con-tentandantioxidantactivityofextractsfromwholebuckwheat(Fagopyrum esculentumMöench)withorwithoutmicrowaveirradiation.FoodChem.119, 1216–1219.

Jaramillo-Flores,M.E.,González-Cruz,L.,Cornejo-Mazón,M.,Dorantes-Alvarez,L.,

Gutiérrez-López,G.F.,Hernández-Sánchez,H.,2003.Effectofthermaltreatment

ontheantioxidantactivityandcontentofcarotenoidsandphenoliccompounds ofcactuspearcladodes(Opuntiaﬁcus-indica).FoodSci.Technol.Int.9,271–278.

Kappe,C.O.,Pieber,B.,Dallinger,D.,2013.Microwaveeffectsinorganicsynthesis:

mythorreality?Angew.Chem.Int.Ed.Engl.52,1088–1094.

Karabegovi ć,I.T.,Stojiˇcevi ć,S.S.,Veliˇckovi ć,D.T.,Todorovi ć,Z.B.,Nikoli ć,N. ˇC.,Lazi ć,

M.L.,2014.Theeffectofdifferentextractiontechniquesonthecompositionand

antioxidantactivityofcherrylaurel(Prunuslaurocerasus)leafandfruitextracts. Ind.CropsProd.54,142–148.

Li,H.,Deng,Z.,Wu,T.,Liu,R.,Loewen,S.,Tsao,R.,2012.Microwave-assisted

extrac-tionofphenolicswithmaximalantioxidantactivitiesintomatoes.FoodChem. 130,928–936.

Liu,Y.,Weia,S.,Liao,M.,2013.Optimizationofultrasonicextractionofphenolic

compoundsfromEuryaleferoxseedshellsusingresponsesurfacemethodology. Ind.CropsProd.49,837–843.

López-Mejía,O.A.,López-Malo,A.,Palou,E.,2014.Antioxidantcapacityofextracts

fromamaranth(AmaranthushypochondriacusL.)seedsorleaves.Ind.CropsProd. 53,55–59.

LoPresti,M.,Sciarrone,D.,Crupi,M.L.,Costa,R.,Ragusa,S.,Dugo,G.,Mondello,L.,

2008.EvaluationofthevolatileandchiralcompositioninPistacialentiscusL.

essentialoil.FlavourFrag.J.23,249–257.

Ma,Y.,Ye,X.,Hao,Y.,Xu,G.,Liu,D.,2008.Ultrasound-assistedextractionof

hes-peridinfromPenggan(Citrusreticulata)peel.Ultrason.Sonochem.15,227–232.

Nkhili,E.,Tomao,V.,ElHajji,H.,ElBoustani,E.S.,Chemat,F.,Dangles,O.,2009.

Microwave-assistedwaterextractionofgreenteapolyphenols.Phytochem. Anal.20,408–415.

Nour,V.,Trandaﬁr,I.,Cosmulescu,S.,2014.Antioxidantcapacity,phenolic

com-pounds and minerals content ofblackcurrant (Ribes nigrumL.) leaves as inﬂuencedbyharvestingdateandextractionmethod.Ind.CropsProd.53, 133–139.

Pan,X.,Niu,G.,Liu,H.,2003.Microwave-assistedextractionofteapolyphenolsand

teacaffeinefromgreentealeaves.Chem.Eng.Process.42,129–133.

Perry,R.H.,Green,D.W., 1998.Perry’s ChemicalEngineers’ Handbook,7th ed.

McGraw-Hill,Australia.

Quettier-Deleu,C.,Gressier,B.,Vasseur,J.,Dine,T.,Brunet,C.,Luyckx,M.,Cazin,

M.,Cazin,J.,Bailleul,F.,Trotin,F.,2000.Phenoliccompoundsand

antioxi-dantactivitiesofbuckwheat(FagopyrumesculentumMöench)hullsandﬂour. J.Ethnopharmacol.72,35–42.

Rawson,A.,Tiwari,B.K., Tuohy,M.G.,O’Donnell,C.P.,Brunton,N.,2011.Effect

ofultrasoundandblanchingpretreatmentsonpolyacetyleneandcarotenoid content of hot air and freeze dried carrot discs.Ultrason. Sonochem. 8, 1172–1179.

Ribéreau-Gayon,P.,Glories,Y.,Maujean,A.,Dubourdieu,D.,2000.Handbookof

Enol-ogy.TheChemistryofWineStabilizationandTreatments,vol.2.JohnWileyand SonsLtd,WestSussex,England(Chapter6).

Romani,A.,Pinelli,P.,Galardi,C.,Mulinacci,N.,Tattini,M.,2002.Identiﬁcationand

quantiﬁcationofgalloylderivatives,ﬂavonoidglycosidesandanthocyaninsin leavesofPistacialentiscusL.Phytochem.Anal.13,79–86.

Santos,S.A.O.,Villaverde,J.J.,Silva,C.M.,Neto,C.P.,Silvestre,A.J.D.,2012.

Supercriti-calﬂuidextractionofphenoliccompoundsfromEucalyptusglobulusLabillbark. J.Supercrit.Fluids71,71–79.

Song,J.,Li,D.,Liu,C.,Zhang,Y.,2011.Optimizedmicrowave-assistedextractionof

totalphenolics(TP)fromIpomoeabatatasleavesanditsantioxidantactivity. Innov.FoodSci.Emerg.12,282–287.

Spigno,G.,DeFaveri,D.M.,2009.Microwave-assistedextractionofteaphenols:a

phenomenologicalstudy.J.FoodEng.93,210–217.

Spigno,G.,Maggi,L.,Amendola,D.,Dragoni,M.,DeFaveri,D.M.,2013.Inﬂuenceof

cultivaronthelignocellulosicfractionationofgrapestalks.Ind.CropsProd.46, 283–289.

Spigno,G.,Tramelli,L.,DeFaveri,D.M.,2007.Effectsofextractiontime,temperature

andsolventonconcentrationandantioxidantactivityofgrapemarcphenolics. J.FoodEng.81,200–208.

ˇSvarc-Gajic,J.,Stojanovic,Z.,Carretero,A.S.,Román,D.A.,Borrás,I.,Vasiljevic,I.,2009.

Developmentofamicrowave-assistedextractionfortheanalysisofphenolic compoundsfromRosmarinusofﬁcinalis.J.FoodEng.119,525–532.

Tabaraki,R.,Nateghi,A.,2011.Optimizationofultrasonic-assistedextractionof

nat-uralantioxidantsfromricebranusingresponsesurfacemethodology.Ultrason. Sonochem.18,1279–1286.

Trabelsi, H., Cherif, O.A.F., Sakouhi, F., Villeneuve, P., Renaud, J., Barouh, N.,

Boukhchina, S., Mayer, P., 2012. Total lipid content, fatty acids and

4-desmethylsterols accumulation in developing fruit of Pistacia lentiscus L. growingwildinTunisia.FoodChem.131,434–440.

Wang,J.,Zhang,J.,Zhao,B.,Wang,X.,Wu,Y.,Yao,J.,2010.Acomparisonstudy

onmicrowave-assistedextractionofPotentillaanserinaL.polysaccharideswith conventionalmethod:moleculeweightandantioxidantactivitiesevaluation. Carbohyd.Polym.80,84–93.

Wu,T.,Yan,J.,Liu,R.,Marcone,M.F.,Aisa,H.A.,Tsao,R.,2011.Optimizationof

microwave-assistedextractionofphenolicsfrompotatoanditsdownstream wasteusingorthogonalarraydesign.FoodChem.133,1292–1298.

Xiao,X.,Song,W.,Wang,J.,Li,G.,2012.Microwave-assistedextractionperformed

inlowtemperatureandinvacuofortheextractionoflabilecompoundsinfood samples.Anal.Chim.Acta712,85–93.

Yang,L.,Jiang,J.G.,Li,W.F.,Chen,J.,Wang,D.Y.,Zhu,L.,2009.Optimumextraction

processofpolyphenolsfromthebarkofPhyllanthusemblicaL.basedonthe responsesurfacemethodology.J.Sep.Sci.32,1437–1444.

Yildirim,H.,2012.MicropropagationofPistacialentiscusL.fromaxenic