806J.Wangetal./FoodChemistry106(2008)804–810

Table1

VariablesandtheirlevelsemployedinacentralcompositerotatabledesignforoptimisationofwheatbranextractsVariable

CodedlevelsÀa(À1.68)Naturallevels

Ethanolconcentration(%)Extractiontime(min)

Extractiontemperature(°C)

431133

501540

602050

702560

772967

À1

+1

+a(+1.68)

icancewascheckedbyanF-test.Thesigni cancesoftheregressioncoe cientweretestedbyat-test.2.8.Determinationoftotalphenoliccontent

ThetotalphenoliccontentofthewheatbranextractswasdeterminedwiththeFolin–Ciocalteumethod(Single-ton,Orthofer,&Lamuela-Raventos,1999).Brie y,anali-quotof100llofappropriatedilutionoftheextractswasshakenfor1minwith500lloftheFolin–Ciocalteureagentfreshlypreparedinourlaboratory,and6mlofdistilledwater.Afterthemixturewasshaken,2mlof15%(w/v)sodiumcarbonatewasaddedandthemixturewasshakenagainfor0.5min.Finally,thesolutionwasbroughtupto10mlbyaddingdistilledwater.After2hofreactionatambienttemperature,inggallicacidasstandard,thetotalphenoliccontentofextractswasexpressedasagallicacidequivalent(mggallicacid/gwheatbran).Dataarereportedasmeans±SDforatleastthreereplications.2.9.Statisticalanalysis

Allanalyseswereperformedintriplicate.Theexperi-mentalresultsobtainedwereexpressedasmeans±SD.StatisticalanalysiswasperformedusingthesoftwareSTATISTICA6.0.Datawereanalyzedbyanalysisofvar-iance(p<0.05)andthemeansseparatedbyDuncan’smul-tiplerangetest.

experimentalparameterscorrespondingtothedesignedlevels,whichwerecarriedoutfordevelopingthemodel.Eachexperimentwasperformedintriplicateandtheaver-agetotalphenoliccontentwastakenastheresponse,Y.Regressionanalysiswasperformedbasedontheexper-imentaldataandwas ttedintoanempiricsecondorderpolynomialmodelasshowninthefollowingequation:Y¼b0þ

5Xi¼1

bixiþ

5Xi¼1

biix2i

þ

45XXi¼1j¼iþ1

bijxixjð2Þ

whereYistheresponsevariable,b0,bi,bii,bijaretheregres-sioncoe cientsofvariablesforintercept,linear,quadratic

andinteractionterms,respectively.xiandxjareindepen-dentvariables.

Theresponsesobtainedfromtheexperimentaldesignset(Table2)weresubjectedtomultiplenonlinearregressionusingthesoftwareSTATISTICA6.0,toobtainthecoe -cientsofthesecondpolynomialmodel.Thequalityofthe tofthepolynomialmodelequationwasexpressedbythecoe cientofdeterminationR2,anditsstatisticalsignif-

Table2

Experimentaldesignof ve-level,three-variablecentralcompositerotatabledesigna,bTestset123456789101112131415161718

abc

x1,Ethanolconcentration(%)505050507070707043776060606060606060(À1)(À1)(À1)(À1)(+1)(+1)(+1)(+1)(À1.68)(+1.68)(0)(0)(0)(0)(0)(0)(0)(0)

x2,Extractiontemperature(°C)404060604040606050503367505050505050(À1)(À1)(+1)(+1)(À1)(À1)(+1)(+1)(0)(0)

(À1.68)(+1.68)(0)(0)(0)(0)(0)(0)

x3,Extractiontime(min)152515251525152520202020112920202020(À1)(+1)(À1)(+1)(À1)(+1)(À1)(+1)(0)(0)(0)(0)

(À1.68)(+1.68)(0)(0)(0)(0)

Totalphenolicscontent(mgGAEeq.c/gbran)Experimental2.072.482.262.792.382.572.392.962.372.852.292.881.832.992.852.872.792.91

Predicted1.982.522.262.792.202.742.483.012.412.782.342.801.952.842.862.862.862.86

Experimentaltotalphenoliccontentsareaverageoftriplicates.Averageabsoluterelativedeviation(%)=2.69.Gallicacidequivalent.