Biocatalytic axidation of alcohols

(1)

Biocatalytic axidation of alcohols

Puetz, Hendrik; Puchľová, Eva; Vranková, Kvetoslava; Hollmann, Frank

DOI

10.3390/catal10090952

Publication date

2020

Document Version

Final published version

Published in

Catalysts

Citation (APA)

Puetz, H., Puchľová, E., Vranková, K., & Hollmann, F. (2020). Biocatalytic axidation of alcohols. Catalysts,

10(9), 1-30. [952]. https://doi.org/10.3390/catal10090952

Important note

To cite this publication, please use the final published version (if applicable).

Please check the document version above.

Copyright

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons. Takedown policy

Please contact us and provide details if you believe this document breaches copyrights. We will remove access to the work immediately and investigate your claim.

This work is downloaded from Delft University of Technology.

(2)

ContentslistsavailableatScienceDirect

Food

and

Bioproducts

Processing

jo u r n al ho m e p 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

Flavour-improved

alcohol-free

beer

–

Quality

traits,

ageing

and

sensory

perception

D.C.

Gernat

a

_,

_E.R.

_Brouwer

b

_,

_R.C.

_{Faber-Zirkzee}

b

_,

_M.

_Ottens

a,∗

a_Department_of_{Biotechnology,}_Delft_University_of_Technology,_van_der_Maasweg_9,₂₆₂₉_HZ_Delft,_The_Netherlands b_Heineken_Supply_Chain,_Burgemeester_Smeetsweg_1,₂₃₈₂_PH_Zoeterwoude,_The_Netherlands

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received9March2020

Receivedinrevisedform15July2020 Accepted19July2020

Availableonline25July2020

Keywords: Alcohol-freebeer Zeolite Aldehydes Wortflavour Ageing Sensoryevaluation Adsorption

a

b

s

t

r

a

c

t

Theincreasingpopularityofalcohol-freebeers(AFBs)fosterstheindustryinterestin deliver-ingthebestpossibleproduct.Yet,aremainingsensorydefectofAFBsistheover-perception ofwortflavour,causedbyelevatedconcentrationsofsmallvolatileflavourcompounds(i.e. aldehydes).Previously,molecularsieves(hydrophobicZSM-5typezeolites)werefoundmost suitabletoremovetheseflavoursbyadsorptionwithhighselectivityfromtheAFBs.Inthis work,aflavour-improvedbeerisproducedatpilot-scaleusingthisnoveltechnology,and itschemicalcomposition,sensoryprofileandstabilityareevaluatedagainstareference. Aldehydeconcentrationsintheflavour-improvedproductwerefound79–93%lowerthanin thereference.Thedistinctdifferencewasconfirmedwithatrainedsensorypanelandcould beconservedevenafterthreemonthsageingat30◦C.Futureworkwillfocusontheprocess designtoscaleupthistechnology.

1. Introduction

Duetoincreasingawareness,theimportanceofresponsible drinkingaswellasothermotivatorssuchasreligionorfitness, oneofthefastestgrowingcategoriesinthebeverage indus-tryarecurrentlylow-alcoholandalcohol-freedrinks(Liguori etal.,2018;Blancoetal.,2016).Nonetheless,alcohol-freebeers stilllackbehindtheirregularcounterparts,duetoseveral sen-sorydefectsthatcanoccurasaconsequenceoftheiraltered productionprocess (Blanco etal.,2016;Brányiketal., 2012; Mangindaanetal.,2018).Therefore,anewprocesstoproduce aflavour-improvedalcohol-freebeer(AFB)hasbeendeveloped recently (Gernat et al., 2020a): Zeolites, a type of molecu-larsieve,are usedtoadsorptively removewortoff-flavours frombiologicallyproducedalcohol-freebeers,resultingina productwithasignificantlyreducedamountofStrecker alde-hydes,whicharecommonlyknowntocausethe undesired wortoff-flavourinAFBs.Tobespecific,2-and3- methylbu-tanal(2-MBand3-MB)aswellasmethional(Met)havemost

∗ _{Corresponding}_author.

E-mailaddress:m.ottens@tudelft.nl(M.Ottens).

oftenbeenrelatedtothewortflavourperception,butrecent studiesfoundmorecompoundssuchas2-methylpropanal (2-MP)tobeinvolved(BealandMottram,1994;PerpèteandCollin, 1999;Piornosetal.,2019).

Strecker aldehydes are formed through a heat-induced reaction between an amino acid and a reducing sugar (SchonbergandMoubacher,1952),butalsoalternative path-waysthroughnumerousreactivecarbonylsarepossible(Rizzi, 2008;Hidalgoetal.,2013;HidalgoandZamora,2004;Delgado et al., 2015). Itisthereforedifficulttopreventtheir forma-tion. Previous researchers in brewing have addressed this matterindifferentways,forinstancebyadditionofmasking compounds,removalthroughfermentationorco-separation withethanolduringdealcoholisationbythermalormembrane technologies.Sofar,eitherthedegreeinwortflavourdecrease ortheselectivityhasbeenlimiting(Gernatetal.,2020b).The advantageofusinganadsorbentandspecificallyin particu-larzeolitesfortheseparationtaskistheir2-foldselectivity basedonmolecularsizeandhydrophobicity.Inthisway,only small (volatile),hydrophobiccompoundsare removed from theproduct.

Inthiswork,wepresentthequalitytraitsofthisnovel prod-uct,aswell asitssensory characteristics.Furthermore,the https://doi.org/10.1016/j.fbp.2020.07.014

(3)

Table1–Flavourdetectionthresholdsinalcohol-freebeerdeterminedbylogisticregressionandadjustedforfalse positives(Piornosetal.,2019;Saisonetal.,2009).

Compound Flavour/Odour Detectionthreshold(␮g/L)inAFB

Orthonasal Retronasal

2-MP Fruity,grainy,nutty,chocolate 4.32 0.86

2-MB Fruity,sweet,almond,malt 23.4 8.99

3-MB Malty,nutty,chocolate 0.61 0.44

Met Boiledpotato,metallic,wort 0.47 0.73

Table2–Specificationoforiginalextract,alcohol content,pHandcolourinfeed.

Parameter Feed

Originalextract(◦P) 12.12

Ethanol(%v/v) 0.015

pH 4.2

Colour(EBC) 8.54

Totaliso-␣-acids(mg/L) <0.4

ageingprocessoftheAFBisstudiedforselectedcompounds.

Thechemicaland physicalanalysis presentedinthis work

includesacomprehensiverepresentationofchemicalgroups

andqualityindicators.Particularfocuslies,however,onfour

commonStreckeraldehydes,whoseflavourdescriptorsand

orthonasal(forwardorsniffing-smell)and retronasal

(back-wardormouth-smell)thresholdsinAFBaregiveninTable1.

Furthermore,furfural(FF)asarepresentativeforcompounds formedbyheat-inducedreactionsandtrans-2-nonenal(t2N), acommon indicatorforageingwerestudiedinmoredetail (Saisonetal.,2009).

2. Materials

and

methods

2.1. Chemicals

TheadsorbentZSM-5G-360waspurchasedfromACS Mate-rials(United States).Topreparesolutionsinthelaboratory, eitherMilli-Q gradewater (Merck,Millipore, United States) oranalyticalgrade,absoluteethanol(VWRInternationalBV, TheNetherlands)wasutilized.Thepre-isomerizediso-␣-acid solutionwasobtainedfromJoh.Barth&Son(Germany).Ifnot mentionedotherwise,otherchemicalswerepurchasedfrom SigmaAldrich(UnitedStates).

2.2. Preparationofwort-flavourreducedbeer

Theadsorbent was prepared by soaking it in 70% ethanol for1hand applyingthree consequent washing steps with water,toremovelessvolatile hydrophobiccompoundsthat might have adsorbed during storage. It was then dried at 220◦Cuntilastableweightwasreached.Aftercooling,4kg ofdryadsorbentwere filledintoacolumnwithadiameter of20cmandadjustableheight(Evolve200,Astrea Biosepa-rationsLtd,UnitedKingdom)andpurgedwithsterilefiltered nitrogengasfor3htoremoveasmuchoxygenaspossible. Finally,it waswettedbyflushingthecolumnwith oxygen-freeprocesswater.Asasubstrate,anAFBwasproducedby fermentinghop-freewortat2–4◦C,withthecomposition out-linedinTable2.Duetothelowfermentationtemperature,the formationofethanolwasinhibited.Theresultingunhopped, alcohol-freeproductwasstabilized(polyvinylpolypyrrolidone andsilica) andfilteredthroughaBeerMembrane Filtration System(Pentair,TheNetherlands).Thebatchwasthensplit

into twovesselswithaliquidvolumeof200L,respectively, addingaslightover-pressuretoavoidfoaming.Oneofthe ves-selswasthenconnectedtothezeolite-filledcolumn,avoiding the introduction of airasmuch aspossible. TheAFB con-tainedinthisvesselwasthencirculatedoverthecolumnand samplesweretakenfrequently.After42h,thetankwas dis-connected and bothAFBs (flavour-improvedand reference) wereagainfiltered,standardizedto5.3◦P(1◦Pequals10gsugar perkilogramwort),pasteurizedwith50PU(1PUequals pas-teurizationfor1minat60◦C)and bottled into0.3Lbottles madefrombrownglass.

2.3. Ageingofsamples

Thebottledproductswerestoredat30◦Cinadry,darkroom foramaximumof4months.Eachmonth,twobottleswere sampledtoanalyzeforageingindicators.After3months sam-plesweretestedbyasensorypanelasdescribedinSection 2.9.

2.4. Determinationoforiginalextract,oxygencontent, pHandcolour

TheoriginalextractandpHofaproductweredeterminedby adigital densitymetreofthe oscillationtype(Xample510, AntonPaar,Austria)andapHmetre(AccumetBasicsAB15, FisherScientific,TheNetherlands),respectively.Theoxygen contentwasanalyzedwiththeHaffmansCO2/O2

Gehalteme-ter,typec-DGM(Pentair,TheNetherlands)accordingtothe supplierinstructions.Tomeasuretheproductcolour,theEBC method9.6wasemployed,i.e.spectrophotometrically (Spec-trostar Nano,BMGLabtech,Germany)and convertedtothe EBCunitbyEq.(1),whereA430nmistheabsorptionat430nm

anddthedilutionfactor.

Colour[EBC]=A430nm·25·d (1)

2.5. Analysisofvolatilealdehydes

Aldehydes,otherthanacetaldehyde,wereanalyzedwithan adaptedmethodofVeselyetal.(Veselyetal.,2003)werethe samplewas concentrated byheadspace solid-phase micro-extraction on a PDMS/DVB fibre (57327-U, Supelco, United States) and then injected to a gaschromatograph (Agilent 7890A)equippedwitha30m×0.25mm×0.25␮mVF17MS col-umn and amass spectrometeras the detectorinnegative ionization mode (Agilent 5975C MSD). Inorder To improve theselectivityO-(2,3,4,5,6-pentafluorobenzyl)-hydroxylamine wasusedasderivatisationagent.

2.6. Analysisofothervolatiles

Theconcentrationofethanolwasdeterminedwithan enzy-matic method with the test kit obtained from Thermo

(4)

Table3–Overviewontastingsofflavour-improved(A)andreference(B)AFBsperformedwiththedescriptivepanel.

Session Productsincluded Storage #ofpanellists

1 (I)BaseAFBs(AandB) Fresh 13

2 (II)BaseAFBs(AandB)

(III)BaseAFBsspikedwithfruityflavour mix(AandB)

(IV)BaseAFBsspikedwithfruityflavour mixandiso-␣-acidstoadjustto15BUa_(A andB)

Fresh 13

3 (V)BaseAFBs(AandB) Aged 11

a _BU_(bitterness_unit)₌_50·A

275nm,whereA275nmistheabsorbanceat275nmofabeerisooctaneextract.

FisherScientific(TheNetherlands) accordingtothe recom-mended EBC method 9.3.1 (Analytica, 2013). Diacetyl and 2,3-pentanedione are quantified with an adapted method of Ruehle et al. (2013) by headspace gas chromatography (7820A, Agilent Technologies, The Netherlands) equipped with a fused silica WCOT CP Sil CB wide bore column (50m×0.53mm×1␮m)anddetectedwithanelectroncapture detector.Toincreaseaccuracy,2,3-hexanedionewasusedas internalstandard.

Acetaldehyde,ethylacetate,isoamylacetateandisoamyl alcohols were analyzed by headspace gas chromatography (7820A,AgilentTechnologies,TheNetherlands)equippedwith aflameionizationdetector(7890B,AgilentTechnologies,The Netherlands)accordingtoEBCmethod9.39(Analytica,2000). Thecompoundswereseparatedoverapolarcapillarynarrow borecolumn(DBWaxETR,60m*0.32mmID,1␮mfusedsilica) and4-heptanonand1-butanolwereaddedtoeachsampleas internalstandard.

2.7. Analysisofnon-volatilecompounds

Thetotaloffermentablesugarswascalculatedfromthe sum-mationofglucose,fructose,sucrose,maltoseandmaltotriose. ThesesugarswerequantifiedbyUPLC(WaterAcquity,Milford, UnitedStates)equippedwithaRIdetectorandaBEHAmide column1.7␮m(2.1×150mm)asdescribedinEBCmethod9.27 (Analytica,1997).Todeterminethefree-aminonitrogen(FAN), aCDRBeerlabtestkitfromFoodLab(Italy)wasusedaccording totheproceduredescribedbythemanufacturer.

2.8. Analysisoffoamstabilityandturbidity

Thefoam stabilitywasdeterminedatroomtemperaturein aNIBEM-TmetrefromPentair(TheNetherlands)(Analytica, 2004).FollowingMEBAKmethod2.141.2,thesample’sturbidity wasanalyzed(MEBAK-EV,2013).

2.9. Sensoryevaluation

ToevaluatetheimprovementintheflavouroftheAFBthat hasbeencontactedwiththeselectiveadsorbent,three differ-entsensoryevaluationsessionswereheld.Inthefirstsession, theunhoppedbaseproducts,i.e.theflavour-improved(IA) andthereference(IB)weretasted.Tounderstandtheimpact oftheadsorptiveremovalsteponthefinalproduct,thebase AFBs(bothflavour-improvedAandreferenceB)werespiked witha fruity/esteryflavour mix(sample III)as well asthe fruityflavour mix and iso-␣-acid solution (sample IV) and comparedtothebaseAFB(sampleII)inthesecondsession. Furthermore,thebaseAFBswereagaintastedafter3months ageingat30◦Cinathirdsession.Theoverviewofall

tast-ingsisgiveninTable3.Thesensoryevaluationofproduced AFBswasperformedwithatrainedsensorydescriptivepanel consistingofatotalof16trainedassessorswithamodified Quantitative DescriptiveAnalysis(QDA)(Stoneet al.,2004). First,anattributelistwasdeterminedduringagroup discus-sion.Attributeintensitieswerequantifiedinduplicateona 100-pointlinescaleduringtwoindividualsessions.Panellists werealignedintheirlinescaleusageandthesampleswere offeredone-by-oneinrandomizedorder.Inasubsequent ses-sion,panellistswereseatedinindividualsensoryboothsand received 100mLofeach sample, presentedin black-coated glassesthatarecodedwiththree-digitcodes.Theyweregiven approximately10minutestoevaluateeachsample,allowing toneutralizetheirpalateinbetween.Persession,thepanel evaluatedamaximumofsixsamples.

2.10. Statisticalanalysis

The standard deviation of the sample and the standard errormofeachmeasurementweredeterminedaccordingto Eqs.(2)and (3),respectively. Thepropagated errormQ was determined as described elsewhere (Young, 1962). Thereby, thestatisticalerrorofthesample,aswellasthesystematic erroroftheregressedcalibrationparameter,weretakeninto account.Allerrorsintheresultsectionrepresentthestandard propagatederror,ifnotmentionedotherwise.

=

1 (n−1) n

i=1 (xi−x2) (2) m= √ n (3)

Thedatacollectionforthesensoryevaluationisdonein EyeQuestion4.11.32.EyeOpenR(partofEyeQuestion)isused forstatisticalanalysis.EyeOpenRisbasedonthestatistical language R.Theanalysisisverifiedand documentedbyQi StatisticsLtd.

3. Results

and

discussion

3.1. Flavour-improvedproduct

During the adsorptiveremovalthroughrecirculation ofthe alcohol-freebeeroverthezeolite-filledcolumn,sampleswere takenfrequentlytomonitorthereductioninwortflavourover time.ThenormalizedresultsaredepictedinFig.1.The con-centrationdecreasesexponentiallyovertime,equilibratingat a plateau of6–10% oftheir initial concentration, thus dis-tinctlydecreasingthewortflavourintheproduct.According

(5)

Fig.1–Evolutionofaldehydeconcentrationsduringthe

adsorptiveremovalstep.

topreviouslyreported data (Gernat et al., 2020a), however, the theoretical capacity ofthe adsorbent should allow for a decrease of freealdehydes tonear zero. There could be severalreasons forthisobservation, but,whenconsidering theabsoluteconcentrationrange(<15␮g/L)andthefactthat theoxygenlevelwassignificantlyincreasedduringthetrial (∼210–1000␮g/L) onecouldsuspectthatoxygen causedthe formation ofnew Strecker aldehydes forinstance through theformationofreactivedicarbonyls.Furfural’sconcentration showsadifferentbehaviour,decreasingto2%ofitsoriginal amount,whilestill furtherreducing inconcentration after 42h.ContrarytotheStreckeraldehydes,2-MP,2-MB,3-MBand Met,whoseformationarecloselyrelatedtooxygen(Wietstock etal.,2016),furfuraloriginatesfromthereactionofapentose withanamineoraminogrouptoaniminecalledSchiffbase, whichatpH<5furtherreactsto3-deoxyosone.Duringa con-densationreaction,ringformationoccursandfurfuralisthe product(Vanderhaegen etal.,2006a).Thus,sinceoxygen is hencenotinvolvedintheformationmechanismoffurfural, theobserveddifferencetoStreckeraldehydegivesevidence tosupportthetheoryofoxygenbeingthecauseforthelower limitofStreckeraldehyderemoval.

Overall,however,theconcentrationinwortflavoursis sig-nificantlydecreased.Fig.2showstheabsoluteconcentration oftheselectedaldehydes intheend-productofthetreated productandthereference.Theachievedreductionamounts to79–93%.Thisreductionislesspronouncedthanwhatwas measuredduringthetest,duetothe intermediate pasteur-ization step, where the aldehyde concentration is slightly increasedagain(datanotshown).

Togain a fullunderstanding of the effectof the adsor-bentontheproductquality,athoroughanalysisofnumerous qualityindicatorswasperformed.Themostinformativedata issummarized in Table4. Both products were adjusted to asimilaroriginalextracttomakethe analysiscomparable. Duetotheprocess-relatedhandling,theoxygenconcentration intheflavour-improvedproductwasincreasedsignificantly. However,whenscaling up,it isexpectedthatwith amore sophisticated set-up, the introduction of oxygen can be avoided.Themeasurementsconfirmthattheoptical appear-anceofthe aroma-improved AFB isunchanged.It exhibits the same colour and turbidity as the reference and even hasaslightlyimprovedfoamstability.Relativelypolar com-pounds such as sugars and ethanol are not removed and alsotheFANcontentissimilartothereference.Accordingto

Fig.2–Concentrationof2-methylpropanal,2-and

3-methylbutanal,methionalandfurfuralinalcohol-free

beerdirectlyafterpasteurization.

expectation,theadsorbentremovessmallvolatilecompounds suchasacetaldehyde,diacetylorethylacetate.Isoamylacetate (<0.05mg/L),amylalcohols(<1.1mg/L)and2,3-pentanedione (<2␮g/L)werefoundbelowtheirrespectivedetectionlimitsfor bothsamplesandcouldhencenotbeincludedintheanalysis. Toconfirmthattheanalyticalresultsarealsoperceivable duringconsumptionandtoestablishwhetherthereduction inwortflavourissignificantlylowerasintended,aQDAwas performedwithatraineddescriptivepanel.Inthefirstsetof tasting,onlythe hop-freebaseproducts(I Aand IB)were tasted bythepanellists.Theoutcomeisrepresentedinthe diagramofFig.3.Thepanellistfoundacleardifferenceinthe perceivedwortflavour(−8points)betweenthetwosamples. Particularly,raisinsandryebreadflavourdiminished,whichis desiredforAFBs.Furthermore,thebitterandsour(after)taste wasreduced.Asaconsequence,thetreatedsamplewasfound morewateryandexhibitedalowerodourandtotalintensity (−6/−5points).

Toassesswhichimpacttheflavourimprovementhason thefinalproduct,i.e.aproductexhibitingfruitybeerflavour (notcontainingwortflavour)andbitterness,thesecondsetof tastingswasexecuted.Forthispurpose,thebaseAFBs(IIA andIIB),aflavouredAFB(IIIAandIIIB)andaflavouredAFB adjustedto15BUinbitterness(IVAandIVB)weretasted.The resultsfortheQDAarelistedinTable5.Atfirstglance,itis prominentthatthedifferenceinwortflavourbetween match-ingsamplespersists,althoughitisperceivedlesspronounced. Forinstance,ifsampleIIAandIIBarecompared,theydiffer in3pointsinwortflavourcontrarytothe8pointswhenthe samesamples(IAandIB)weretastedinaseparatesession. ItshouldbenotedthatsamplesIandIIarethesamesample, buttastedindifferentsessions.Therelativedifferent percep-tionofwortinessreductionisnotrelatedtotheuncertaintyof thesensoryevaluation.Itrathermeansthatthepresenceof fruityflavoursandbitternessinoneormoresamplestastedin thesamesessioninfluencesthepanellist’sfocusonthewort character.Nonetheless,asignificantdifferenceinparticularin ryebreadflavourisdetected.InbetweensamplesII,IIIandIV onecanclearlyobservethemaskingeffectofaddingafruity flavouringandbitternesswithrespecttothewortcharacterof theproducts.Interestingly,thepanellistsperceivedflavoured samplesoftheimprovedbasehigherintropicalfruitflavour than thereference.Asaconsequence,flavourdosingcould bereducedinthefinalproducttoachievethesamesensory profile.

(6)

Table4–Specificationsandtheirstandarddeviationofbottledproducts(referenceandflavour-improved).

Flavour-improved Reference

Adjusted OE(%m/m) 5.23±0.02 5.30±0.02

Processrelated Oxygen(␮g/L) 61.1±3.2 16.3±1.3

Unchanged Colour(EBC) 3.4±0.0 3.3±0.0

Ethanol(%V/V) 0.01±0.00 0.01±0.00

FAN(mg/L) 88±2 90±2

pH 4.76±0.02 4.74±0.02

Totalfermentablesugars(g/L) 34.0±1.5 33.9±1.5

Turbidity(EBC) 0.2±0.0 0.1±0.0

Turbidityafter7daysat57◦C 0.1±0.0 0.2±0.0

Increased FoamStability(s) 256±5 217±4

Decreased Acetaldehyde(mg/L) 1.4±0.1 2.3±0.1

Diacetyl(␮g/L) 3.9±0.3 15.9±1.1

Ethylacetate(mg/L) <0.2±n/a 0.3±0.0

Fig.3–SensoryprofileofbaseAFB,flavour-improved(solidline)andreference(dashedline).Attributeswhodifferwitha significancelevelof95%arecircledinblackandwithasignificancelevelof90%ingrey.

Table5–Sensoryattributesquantifiedbythedescriptivepanel.

IIA IIB IIIA IIIB IVA IVB p-Value

Odourintensity 37 40 39 40 38 38 0.14 Sparklingmouthfeel 31 31 32 32 31 29 0.19 Bittertaste 24 24 24 24 35 35 0.00 Sourtaste 29 29 28 29 30 29 0.72 Sweettaste 22 23 22 22 21 22 0.19 Wortflavour 34 37 28 32 29 31 0.00 Grainyflavour 24 25 18 22 18 21 0.00

Grasssilageflavour 16 19 14 17 18 16 0.45

Raisinsflavour 12 17 9 12 9 9 0.04

Ryebreadflavour 14 17 5 12 5 11 0.00

Fruity/esteryflavour 26 25 39 34 37 35 0.00

Peardropflavour 23 22 34 31 34 32 0.00

Tropicalfruitflavour 8 7 20 12 18 15 0.00

Artificialsweetenerflavour 16 14 14 15 12 12 0.23

Totalintensity 41 41 43 40 43 43 0.02

Watery(lackoftaste/flavours) 41 41 41 42 39 39 0.35

Thicknessmouthfeel 22 22 21 22 21 22 0.48

Astringentmouthfeel 22 22 22 19 24 23 0.02

Bitteraftertaste 20 20 20 21 34 33 0.00

Souraftertaste 24 24 23 23 25 22 0.04

(7)

Fig.4–Concentrationofselectedcompoundsoveracoursefor4monthsfortreated(dark)andreferenceproduct(white).

Table6–Formationrateofaldehydesinflavour-improvedandreferencebasebeerduringstorageat30◦Cand comparisonwithliteraturerange.

Compounds Formationrate(␮g/month) Approximatedformationratea

rangefromJaskula-Goirisetal. (2011)andBaert(2015) (␮g/month) Improved Reference 2-MP 1.2±0.1 1.8±0.2 4.7–28.8 2-MB 0.7±0.0 0.8±0.1 0.6–1.9 3-MB 2.2±0.3 1.3±0.4 1.4–3.2 FF 53±9 69±15 63–99 Met 1.3±0.3 −0.1±0.7 0.9–2.1 t2N −0.006±0.004 0.016±0.026 0.003–0.033

a _Visual_estimation_from_plotted_data.

3.2. Ageing

Aldehydesarenotonlyknownwortoff-flavoursinAFBs,but alsorelatedtoageingduringstorage.Thereisavastvariety ofliteratureavailable,investigatingflavoursandtheir chemi-calpathwaysinvolvedinageingofbeers(Vanderhaegenetal., 2006a;Baertetal.,2012).Inthisstudy,wefocusonaselected

setofaldehydes.Themostrecentstudiesinregularbeer indi-cate thatdenovo formationofaldehydesduringtheageing process islimitedand thatthemostdominantmechanism causingtheincreaseinaldehydeconcentrationovertimeis thereleasefromcysteineandbisulphiteadducts(Baertetal., 2018; Baertet al., 2015). Toour knowledge,this isthe first studytomeasureageinginAFB.Fig.4andTable6depictthe

(8)

Fig.5–SensoryprofileofbaseAFB,flavour-improved(VA,solidline)andreference(VB,dashedline)agedfor3monthsat

30◦C.Attributesdifferingwithasignificancelevelof95%arecircledinblackandwithasignificancelevelof90%ingrey.

time-wiseconcentrationprofileoftheflavour-improvedand referencebeerforallstudiedaldehydesduringfourmonths and their regressedformation rates, respectively. The first remarkable observation isthat the formation rate, i.e. the slope,oftheflavour-improvedandthereferenceAFBare rel-atively similar inthe case of2-MP, 2-MB and FF, although the formation rate of the treated AFB is generally lower. This means that the difference in concentration between theimprovedandreferenceAFBwillremainsimilaroreven increaseoverthestoragetime.Thisisalsothecasefort2N,but herethetreatedproductslightlydecreasesintheageing com-pound.Thisisdifferentfor3-MB:heretheformationrateinthe improvedproductishigherandhencethedifferenceis less-eningduringstorage.Methional’sconcentrationisinclining intheimprovedproduct,however,barelyseemstochangein thereference.Thereasonbehindthisdifferentbehaviourcan onlybespeculatedandarehencegroundforfurtherresearch. Consideringthatthestartingconcentrationofmethionalin thereferenceisthehighestofallmeasuredStrecker aldehy-des,onecouldpostulatethatitalreadyisinequilibriumwith itseducts.Anotherpossibleexplanationisthatitis simultane-ouslyformedandconvertedtodegradationproductssuchas methanethiolandacreolinaspreviouslyreportedinliterature (Gijsetal.,2000;Wainwrightetal.,1972;Ballance,1961).

Sincethebeerisfermentedatlowtemperaturesresulting inminimalyeastactivity,thestudiedbeerhasseveral proper-ties,whichmakesthiscaseinterestingtostudyandcompare withdatareportedinliterature:

1. Thestartingconcentrationinthefreshbeerishigherthan inregularbeer,particularlyinthereferenceproduct. 2. Theethanolcontentis<0.01vol.%inethanol. 3. Nosulphiteisproducedbyyeast.

4. Thefermentablesugarconcentrationishigherthanin reg-ularbeersincetheyarenotconsumedduringfermentation.

Furthermore, no compounds originating from hops are containedintheAFB.Jaskula-Goirisetal.(2011)studiedthe beer stability ofpale lager beers from different breweries. Surprisingly, despitethe numerousdifferences inthe beer composition, theformationrateinaldehydesisquite com-parable(seeTable6).Theyfoundthatstalingwaspositively correlatedwiththeresidualFANcontent,andthusthe con-centrationofaminoacidsinbeer.Intheirstudy,freshbeers contained between 49.1 and 141.2mg/L FAN, wherein our productscomparablelevelsof88–90mg/Lweredetermined. Thisobservationsupportsthehypothesisthataldehyde for-mationisstronglylinkedtothepresenceofaminoacidsin beer.Otherauthorsreportthatunhoppedbeerbarelyproduces anyagedflavour(HashimotoandEshima,1979),contraryto whatisobservedinthisanalysis.

Whencomparingthesensoryprofileoftheagedproducts witheachother,asshowninFig.5,thedifferencebetweenthe flavour-improvedandthereferenceproductgenerallybecome lessdistinct,i.e.inodour/totalintensityandwortflavouris stillfoundtobesignificantbetweenthetwosamples,however, thisdifferenceislesspronounced.Thepanellistsalsocannot perceivecleardifferencesinraisinsandryebreadflavour any-more.Attributesthatwerefoundslightlydivergentinthefresh productsuchasthebittertasteandaftertaste arenow per-ceivedsimilar.Atacloserlook,itisobservedthattheidentified attributesslightlydifferfromthetastingofthefreshproduct. Whiletheartificialsweetenerflavourisnotrecognized any-more,thesweetaftertasteofthesampleisbecomingmore pronouncedinthereferencebeerandasyrupflavourisnoted bythepanellists.

Inliterature,itisreportedthatoverthewholecourseof age-ing,sweetnessandtoffee-likeorcaramelaromaincreasesand bitterness decreases (Vanderhaegen et al.,2006b;Dalgliesh, 1977). Ribes and cardboardflavour may appearduring the ageingprocess,butafterreachingapeakatabout4weeks dis-appearagain(Baertetal.,2012;Zufalletal.,2005).Thus,the abovefindingsresonatewellwiththeexpectationsfrom

(9)

liter-ature.Althoughthedifferenceintheactualconcentrationin wortflavourbetweentheflavour-improvedandthereference productisstillverysimilarasinthefreshproduct,thesame degreeofwortflavourreductionisnotreflectedinthe sen-soryevaluation.Onereasonforthismightbethatthecurveof concentrationandresponseinflavourintensityfollowsa sig-moidalshape,henceisnearlylinearintherangeclosetothe flavourthreshold,but,athigherconcentrationsthisresponse declinesand mightevensaturate(Breslin,1996).Thesame absolutedifferenceinconcentrationintwosamplesishence lessobviousathigherconcentrations.Additionally,synergetic effectscanenhancethisphenomenon(Singhetal.,2019).

4. Conclusion

Inthis work,aflavour-improved AFBwas producedwith a significantlydecreased concentrationincharacteristicwort flavours,resulting ina clean-tastingbasebeer. This differ-encewasconfirmedduringsensoryevaluationwithatrained panel. Due to the high selectivity of the adsorbent, only smallvolatilecompoundswereremoved,whileother param-eters suchas colour, foam stability or sugar content were similar.Additionoffruityflavoursandbitternessdecreased therelativeperceiveddifferenceinwortflavourbetweenthe improved AFB and the reference, but a significantly lower wortinessremained.Evenafterthreemonthsofforced age-ing,qualitydifferenceswerestilldetectable.Futureworkwill focusondesigningascalableandeconomicallyfeasible pro-cessunitoperationandinvestigatetheregenerabilityofthe absorbent.Furthermore,moredetailedinvestigationsofthe ageingbehaviourforinstancewithhoppedAFBarenecessary toimprovetheunderstandingoftheinvolvedmechanisms.

Declaration

of

interests

The authors declare that they have no known competing financialinterestsorpersonalrelationshipsthatcouldhave appearedtoinfluencetheworkreportedinthispaper.

Acknowledgements

TheauthorsacknowledgeHeinekenSupplyChainB.V.forthe fundingprovidedforthisprojectandtheopportunitytocarry outthepilotscaletestsattheirfacilities.

References

Analytica,1997.EBC9.27:FermentableCarbohydratesinBeerby HPLC(IM).

Analytica,2000.EBC9.39:DimethylSulphideandOtherLower BoilingPointVolatileCompoundsinBeerbyGas

Chromatography.

Analytica,2004.EBC9.42.1:TheDeterminationoftheFoam StabilityofBeerUsingtheNIBEM-TMeter.

Analytica,2013.EBC9.3.1:EthanolinAlcohol-FreeandLow AlcoholBeers:EnzymaticMethod(IM).

Baert,J.J.,2015.UnravellingtheRoleofFreeandBound-state AldehydesinBeerFlavorInstability.FacultyofBioscience Engineering,KULeuven,Leuven.

Baert,J.J.,etal.,2012.Ontheoriginoffreeandboundstaling aldehydesinbeer.J.Agric.FoodChem.60(46),11449–11472. Baert,J.J.,etal.,2018.Exploringaldehydereleaseinbeerby

4-vinylpyridineandtheeffectofcysteineadditiononthe beer’spoolofboundaldehydes.J.Am.Soc.Brew.Chemists76 (4),257–271.

Baert,J.J.,etal.,2015.Furtherelucidationofbeerflavorinstability: thepotentialroleofcysteine-boundaldehydes.J.Am.Soc. Brew.Chemists73(3),243–252.

Ballance,P.E.,1961.Productionofvolatilecompoundsrelatedto theflavouroffoodsfromtheStreckerdegradationof DL-methionine.J.Sci.FoodAgric.12(7),532–536.

Beal,A.D.,Mottram,D.S.,1994.Compoundscontributingtothe characteristicaromaofmaltedbarley.J.Agric.FoodChem.42 (12),2880–2884.

Blanco,C.A.,Andres-Iglesias,C.,Montero,O.,2016.Low-alcohol beers:flavorcompounds,defects,andimprovement strategies.Crit.Rev.FoodSci.Nutr.56(8),1379–1388. Brányik,T.,etal.,2012.Areviewofmethodsoflowalcoholand

alcohol-freebeerproduction.J.FoodEng.108(4),493–506. Breslin,P.A.S.,1996.Interactionsamongsalty,sourandbitter

compounds.TrendsFoodSci.Technol.7(12),390–399. Dalgliesh,C.E.,1977.Flavourstability.Proceedingsofthe

EuropeanBreweryConventionCongress,623–659. Delgado,R.M.,Zamora,R.,Hidalgo,F.J.,2015.Contributionof

phenoliccompoundstofoodflavors:Strecker-type degradationofaminesandaminoacidsproducedbyo-and p-diphenols.J.Agric.FoodChem.63(1),312–318.

Gernat,D.C.,etal.,2020a.Selectiveoff-flavorreductionby adsorption:acasestudyinalcohol-freebeer.FoodBioprod. Process.121,91–104.

Gernat,D.C.,Brouwer,E.,Ottens,E.,2020b.Aldehydesaswort off-flavoursinalcohol-freebeers–originandcontrol.Food BioprocessTechnol.13(2),195–216.

Gijs,L.,etal.,2000.3-Methylthiopropionaldehydeasprecursorof dimethyltrisulfideinagedbeers.J.Agric.FoodChem.48(12), 6196–6199.

Hashimoto,N.,Eshima,T.,1979.Oxidativedegradationof isohumulonesinrelationtoflavourstabilityofbeer.J.Inst. Brew.85(3),136–140.

Hidalgo,F.J.,Delgado,R.M.,Zamora,R.,2013.Intermediateroleof ␣-ketoacidsintheformationofStreckeraldehydes.Food Chem.141(2),1140–1146.

Hidalgo,F.J.,Zamora,R.,2004.Strecker-typedegradation producedbythelipidoxidationproducts

4,5-epoxy-2-alkenals.J.Agric.FoodChem.52(23),7126–7131. Jaskula-Goiris,B.,etal.,2011.Detailedmultivariatemodelingof beerstalingincommercialpalelagers.Brew.Sci.64(11–12), 119–139.

Liguori,L.,etal.,2018.Productionoflow-alcoholbeverages: currentstatusandperspectives.In:Grumezescu,A.M., Holban,A.M.(Eds.),FoodProcessingforIncreasedQualityand Consumption.AcademicPress,pp.347–382(Chapter12). Mangindaan,D.,Khoiruddin,K.,Wenten,I.G.,2018.Beverage

dealcoholizationprocesses:past,present,andfuture.Trends FoodSci.Technol.71,36–45.

MEBAK-E.V,2013.2.14.1.2.HazeFormationOpticalMethod. Freising,Germany.

Perpète,P.,Collin,S.,1999.Contributionof

3-methylthiopropionaldehydetothewortyflavorof alcohol-freebeers.J.Agric.FoodChem.47(6),2374–2378. Piornos,J.A.,etal.,2019.Orthonasalandretronasaldetection

thresholdsof26aromacompoundsinamodelalcohol-free beer:effectofthresholdcalculationmethod.FoodResInt.123, 317–326.

Rizzi,G.P.,2008.TheStreckerdegradationofaminoacids:newer avenuesforflavorformation.FoodRev.Int.24(4),416–435. Ruehle,G.,etal.,2013.Headspacegaschromatography/electron

capturedetectoranalysisoftotalvicinaldiketonesinbeer.J. Am.Soc.Brewi.Chemists71(4),274–275.

Saison,D.,etal.,2009.Contributionofstalingcompoundstothe agedflavouroflagerbeerbystudyingtheirflavourthresholds. FoodChem.114(4),1206–1215.

Schonberg,A.,Moubacher,R.,1952.TheStreckerdegradationof ␣-aminoacids.Chem.Rev.50(2),261–277.

Singh,V.,etal.,2019.Competitivebindingpredictsnonlinear responsesofolfactoryreceptorstocomplexmixtures.Proc. Natl.Acad.Sci.U.S.A.116(19),9598–9603.

(10)

Stone,H.,etal.,2004.Sensoryevaluationbyquantitative descriptiveanalysis.In:DescriptiveSensoryAnalysisin Practice.

Vanderhaegen,B.,etal.,2006b.Thechemistryofbeeraging–a criticalreview.FoodChem.95(3),357–381.

Vanderhaegen,B.,etal.,2006a.Thechemistryofbeeraging–a criticalreview.FoodChem.95(3),357–381.

Vesely,P.,etal.,2003.Analysisofaldehydesinbeerusing solid-phasemicroextractionwithon-fiberderivatizationand gaschromatography/massspectrometry.J.Agric.FoodChem. 51(24),6941–6944.

Wainwright,T.,McMahon,J.F.,McDowell,J.,1972.Formationof methionalandmethanethiolfrommethionine.J.Sci.Food Agric.23(7),911–914.

Wietstock,P.C.,Kunz,T.,Methner,F.-J.,2016.Relevanceofoxygen fortheformationofStreckeraldehydesduringbeer

productionandstorage.J.Agric.FoodChem.64(42), 8035–8044.

Young,H.D.,1962.StatisticalTreatmentofExperimentalData. M.-H.B.C.Inc.,UnitedStatesofAmerica.

Zufall,C.R.G.,Gasparri,M.,Franzquiz,J.,2005.Flavourstability andageingcharacteristicsoflight-stablebeers.In:30th EuropeanBreweryConvention,FachverlagHansCarl, Nürnberg,Germany.

Biocatalytic axidation of alcohols