808J.Wangetal./FoodChemistry106(2008)804–810

3.4.E ectofextractiontimeonextractionoftotalphenoliccompounds

Thecontentsoftotalphenoliccompoundsextractedfromwheatbranatdi erenttimesofsonicationarepre-sentedinFig.4.Amarkedincreaseofthetotalphenolicscontentwasobservedupto30min,remainingconstantuntil50min.

3.5.Optimisationofextractionconditions

Basedontheinvestigationsofthee ectsofethanolcon-centration,extractiontemperatureandtimeonthecontentofphenoliccompoundsofwheatbranextracts,thesevari-ableswereconsideredintheexperimentaldesign.Toopti-

misetheextractionprocessofthetotalphenolicsfromwheatbran,anethanolconcentrationof60%(v/v),anextractiontemperatureof50°C,andanextractiontimeof20minwerechosenasthecentralconditionoftheCCRD.

Table2showstheexperimentalconditionsandtheresultsofextractionaccordingtothefactorialdesign.Themaximumcontentofphenoliccompounds(2.99mgGAE/gbran)wasrecordedundertheexperimentalparam-etersofanethanolconcentrationof60%,anextractiontemperatureof50°Candanextractiontimeof29min.Thelowestcontentofphenoliccompounds(1.83mgGAE/gbran)wasobservedintestsetNo.13,comparedtotheothers.Statisticalanalysisrevealedthatthemostrel-evantvariable(p<0.001)concerningthecontentoftotalphenoliccompoundswastheextractiontime(Table3).Thetotalphenoliccontentwasalsohigh(2.96mgGAE/gbran)withanethanolconcentrationof70%,anextractiontemperatureof60°Candanextractiontimeof25min,cor-respondingtotestsetNo.8.AhighercontentofphenolicsfromtestsetNo.14,comparedtothatfromtestsetNo.13,couldimplythatanincreaseinextractiontimewasmostfavourableforphenoliccompoundsextraction.

Multipleregressionanalysiswasperformedontheexperimentaldataandthecoe cientsofthemodelwereevaluatedforsigni cancewithaStudentt-test.Allthelin-earcoe cientsweresigni cant(p<0.05).Allthecross-productcoe cientswereeliminatedinthere nedequationastheire ectswerenotsigni cant.Thevaluesofthecoef- cientsarepresentedinTable3.TheanalysisofvariancefortheCCRDisshowninTable4.Thecoe cientofdeter-mination(R2)ofthemodelis0.9492,whichindicatedthatthemodelhadadequatelyrepresentedtherealrelationshipbetweentheparameterschosen.Theaverageabsoluterela-tivedeviationofthereducedmodelis2.69%.Neglectingthenon-signi cantterms,the nalpredictiveequationobtainedisasgivenbelow:

Y¼2:85637þ0:110401x1þ0:138602x2þ0:267413x3

À0:091756x222

1À0:101613x2À0:163617x3

ð3Þ

Todetermineoptimallevelsofthevariablesfortheextractionofphenoliccompoundsfromwheatbran,three-dimensionalsurfaceplotswereconstructed,accord-ingtoEq.(3).Fig.5showsthee ectofethanolconcentra-tionandextractiontemperatureonthecontentoftotalphenoliccompounds.Thetotalphenoliccontentincreasedslowlywiththeincreaseofethanolconcentrationata xedextractiontemperature,andnearlyreachedapeakatthehighestethanolconcentrationtested.Similarly,theincreaseinextractiontemperatureata xedethanolconcentrationledtoagradualincreaseinthetotalphenoliccontent,andreachedamaximumatthehighestextractiontemper-Table3

Estimatedcoe cientsofthe ttedsecond-orderpolynomialmodelforthephenoliccontentfromwheatbranaTermCoe cientsestimatedStandarderrortValuepValueb02.856370.05580351.19<0.0001b10.1104010.0302613.650.0065b20.1386020.0302614.580.0018b30.2674130.0302618.84<0.0001b11À0.0917560.031476À2.950.0185b22À0.1016130.031476À3.230.0121b33

À0.163617

0.031476

À5.2

0.0008

a

Onlytermswithp<0.05wereincluded.

Table4

Analysisofvarianceofthesecond-ordertotalphenoliccontentsofwheatbranmodela

DegreeofSumofMeanFpValuefreedom

squaressquareValueTotalmodel91.8686880.94920016.620.0003Linear31.404170.71330037.46<0.0001Quadratic30.4277680.21730011.410.0029Cross-30.036550.0186000.98

0.451

productTotalerror80.0999620.012495Lackof t50.0924620.0184927.40.0651

Pureerror

3

0.0075

0.0025

a

Coe cientofdetermination(R2)=0.9492.