Electrode material studies and cell voltage characteristics of the in situ water electrolysis performed in a pH-neutral electrolyte in bioelectrochemical systems

(1)

This is a version of a publication

in

Please cite the publication as follows:

DOI:

Copyright of the original publication:

This is a parallel published version of an original publication.

This version can differ from the original published article.

published by

electrolysis performed in a pH-neutral electrolyte in bioelectrochemical systems

Givirovskiy Georgy, Ruuskanen Vesa, Ojala Leo S., Lienemann Michael, Kokkonen Petteri, Ahola Jero

Georgy Givirovskiy*, Vesa Ruuskanen, Leo S. Ojala, Michael Lienemann, Petteri Kokkonen, Jero Ahola. Electrode material studies and cell voltage characteristics of the in situ water electrolysis performed in a pH-neutral electrolyte in bioelectrochemical systems. Heliyon, 5(2019) e01690, p.1-10. DOI: https://doi.org/10.1016/j.heliyon.2019.e01690

Publisher's version Elsevier

Heliyon

10.1016/j.heliyon.2019.e01690

© 2019 The Authors.

(2)

Heliyon 5 (2019) e01690

Contents lists available atScienceDirect

Heliyon

www.elsevier.com/locate/heliyon

Electrode material studies and cell voltage characteristics of the in situ water electrolysis performed in a pH-neutral electrolyte in

bioelectrochemical systems

Georgy Givirovskiy

^a,∗

, Vesa Ruuskanen

^a

, Leo S. Ojala

^b

, Michael Lienemann

^b

, Petteri Kokkonen

^b

, Jero Ahola

^a

aLUTUniversity,P.O.Box20,FI-53851,Lappeenranta,Finland

bVTTTechnicalResearchCentreofFinlandLtd.,P.O.Box1000,02044VTT,Finland

A R T I C L E I NF O A B S T R A C T

Keywords:

Bioengineering Electrochemistry Materialschemistry

Hydrogen-oxidizingbacteria(HOB)havebeenshowntobepromisingmicro-organismsforthereduction of carbondioxidetoa wide rangeof value-addedproductsin bioelectrochemical systemswithinsitu water electrolysisofthecultivationmedium,alsoknownasahybridbiological-inorganicsystems(HBI).However, scalingupof thisprocessrequires overcomingtheinherentconstraintsof thelow energyefficiencypartly associatedwiththepH-neutralelectrolytewithlowconductivity.Mostoftheresearchinthefieldisconcentrated onthebacterialcultivation,whereastheanalysisandevaluationoftheelectrodematerialperformancehave receivedlittleattentionintheliteraturesofar.Therefore,inthepresentwork,in situelectrolysisofa pH- neutralmediumforHOBcultivationwasperformedwithdifferentcombinationsofelectrodematerials.Besides conventionalelectrodetypes,electrodeswithcoatingsmadeofearth-abundantcobaltandanickel-ironalloy, knownfortheircatalyticactivityforthekineticallysluggishoxygenevolutionreaction(OER),wereprepared andtested aspotential substitutes forcatalysts madeof precious metals.The cultivation of HOB within situwaterelectrolysishasbeensuccessfullytestedinasmallscaleelectrobioreactorinordertosupportthe experimentalresults.Asimplifiedwaterelectrolysismodelwasdevelopedandappliedtoevaluatethecurrent- voltagecharacteristicsofanbioelectrochemicalsystemprototype.Applicationofthedevelopedmodelallows quantitativeevaluationandcomparisonofreversible,ohmic,andactivationovervoltagesofdifferentelectrode sets.Themodelingresultswerefoundtoagreewellwiththeexperimentaldata.Thedevelopedmodelandthe datagatheredcanbeappliedtofurtherinvestigation,simulation,andoptimizationofHBIsystems.

1. Introduction

The rapid economic growth and the increasing consumption of fossil-fuel-based energy have led to higher concentrations of pollu- tantgasesintheatmosphere,depletionofnaturalresources,adverse climateimpacts,andgeopoliticaltensions.Theglobalshiftfromafossil- fuel-basedeconomytoarenewable-energy-basedonehasthepotential totackle theaforementioned problems[1, 2]. Electrical energypro- ducedfromabundantrenewableenergysources,suchassolarandwind power,isconsideredtobethecleanestformofenergy.However,the ﬂuctuating nature of these sources leads totechnical challenges as- sociated with the storage of the generated electricity [3]. Recently, hydrogen,whichisthesimplestandlightestelement,hasbeenshown

*

Correspondingauthor.

E-mailaddress:georgy.givirovskiy@lut.ﬁ(G. Givirovskiy).

tobeasustainableandpromisingenergycarrierintheHydrogenEcon- omyConcept[4].Eventhoughthecurrentlydominatingtechnologiesof hydrogenproductionaresteamreforming,partialoxidationofhydro- carbons,andcoalgasiﬁcation,thedevelopmentofadvancedtechnolo- giesforrenewable-energy-basedhydrogenproductionisgivenahigh priority,andthetopicisattractingscientiﬁcinterest.Oneofthemost maturetechnologiesofrenewablehydrogenproductioniselectrolysis ofwater[5].Bythismethod,surpluspeakelectricityfromrenewable energysourcesisappliedtogeneraterenewablehydrogen,whichcan befurtherusedinPower-to-Xprocessestoproducenetcarbon-neutral fuelsandchemicals[6,7].

Oneapproachattractingscientiﬁcinterestinthiscontextismicro- bial electrosynthesis(MES), electricity-driven synthesis of chemicals

https://doi.org/10.1016/j.heliyon.2019.e01690

Received19October2018;Receivedinrevisedform5April2019;Accepted7May2019

2405-8440/© ²⁰¹⁹^TheÂuthors. ^Published^byÊlsevier^Ltd. ^Thisîsânôpenâccessârticleûnder^the^CC^BY-NC-ND^license(http://creativecommons.org/li- censes/by-nc-nd/4.0/).

(3)

andfuels.Microbialelectrosynthesis(MES)isanemergingtechnology capableofusingwaterelectrolysisandvariousmicroorganismsdirectly forthereductionofcarbondioxidetovalue-addedcompoundsinbio- electrochemicalsystems(BESs).TheconceptwasﬁrstprovenbyNevin et al.[8],whowereabletoreducecarbondioxidetoacetateandsmall amounts of 2-oxobutyrateby applying electriccurrent toacetogenic microorganisms. Thesubsequentresearchrevealedanopportunityof applyinginsituwaterelectrolysisandmicrobesfortheeﬃcientproduc- tionof othervalue-addedcommodities.Informationabout chemicals thatcanbeproducedinbioelectrochemicalsystemscanbefoundin[9]

and[10].Hydrogen-oxidizingbacteria(HOB),themetabolicgrowthof whichisbasedontheuseofhydrogenasanelectrondonorandoxygen asanelectronacceptor,wereshowntobepromisingmicroorganismsfor thereductionofcarbondioxidetoawiderangeofvalue-addedproducts [11].Volovaet al.[12] foundthatthebiologicalvalueofproteinssyn- thesizedbydifferentstrainsofhydrogen-oxidizingbacteriaissufficient toconsiderthemasapotentialproteinsourceforhumanandanimal nutrition.Moreover,researchiscurrentlycarriedoutintoHOB-based singlecellproteinproduction.Forinstance,Matassaet al.[13] used autotrophichydrogen-oxidizingbacteriatorecycleammoniarecovered byairstrippingfromawastewatertreatmentplantandcapturedCO₂, togetherwithhydrogenandoxygenproducedbywaterelectrolysis,to foodandfeed[13,14].Furthermore,apilotplanthasbeenconstructed inBelgiumwithintheframeworkofaPower-to-Proteinproject,which producessinglecellproteinwithatargetedcapacityof1 kg–2 kgper day[15].However,theseprocessesrequireexternalsupplyofhydrogen andoxygentothebioreactorswheretheHOBarebeingcultivated.The applicationofbioelectrochemicalsystemwithinsituwaterelectrolysis couldprovideasolutionforovercomingthemasstransferlimitations ofthisprocess,andcouldthusbeconsideredaprospectivestrategyfor renewableelectricalenergystorage.Torellaet al.[16] reporteddevel- opmentofascalableintegratedbioelectrochemicalsystemusingHOB forcarbondioxideconversionintobiomassandisopropylalcoholwith maximumbioelectrochemicalefficienciesof 17.8%and3.9%,respectively.AdistinctivefeatureofthisBESwastheuseofacobaltphosphate (CoPi)anode,whichiscapableofperformingoxygenevolutionreaction (OER)atlowoverpotentialsataneutralpH.Thesameanodematerial wasusedincombinationwithacobalt-phosphorus(Co-P)alloycathode inthestudiesofLiuet al.[17] toestablishaneffectivewatersplitting systemforHOBconversion intobiomass atanefficiencyof approxi- mately55%withinaperiodofsixdaysatanappliedpotentialof2.0 V.

Inadditiontobiomass,polyhydroxybutyrate(PHB),whichisconsidered anintermediatecompoundinmicrobialassimilationofcarbondioxide, wassynthesizedwitha36%energyefficiency.Furthermore,different fuselalcohols wereproducedwithefficienciesrangingfrom approxi- mately15%to30%.Hybridbiological-inorganic(HBI)systems,which couplemicroorganismswithchemicalcatalyststoderivevalue-added products,havealsobeenapplied,forexample,toammonia[18] and bacterialbiomassproduction[19].

Nevertheless,upscalingofbioelectrochemicalprocessesforHOBcul- tivation requires overcoming theinherent constraints of low energy efficiency.Thetargetofthepresentstudyistodevelopascalableen- ergyefficientsystemforcultivationofhydrogen-oxidizingbacteria.The effectsofvariousoxygenevolution(OER)catalystsareextensivelyre- portedin theliteratureforalkalinewaterelectrolyzers,butthereare onlyafewstudiesofelectrolyzercellperformanceinbioreactorswith pH-neutralconditionssofar.Further,asimplifiedmathematicalmodel is introduced,basedon modelsdeveloped fortraditionalwaterelec- trolyzers.Themodelparametersaretunedandthemodelisverifiedby experimental results.Themodelis appliedtoquantitativelyevaluate andcomparepossibleovervoltagesourcesinthesystemwithvarious electrodematerials.

Thispaperisorganizedasfollows.Thecharacteristicsoftheinsitu waterelectrolysis,initialHOBcultivationresultswithinsituwaterelec- trolysis,theexperimentalsetupusedforelectrodematerialtests,the proceduredescribinginsituformationofcoatings,andthesimpliﬁed

electrolyzercellvoltagemodelareintroducedin Section2. Thecell voltagemodelparametersareﬁttedbyexperimentalresults,andthe modelisappliedtodescribetheperformanceoftheselectedelectrode materialsinSection3.Finally,Section4concludesthepaper.

2. Materials& methods

This sectionﬁrst deﬁnes thespecial characteristicsof thein situ waterelectrolysiscomparedwithtraditionalwaterelectrolyzers.HOB cultivationresultswithinsituwaterelectrolysisareshown.Further,the experimentalsetupandmethodsappliedforelectrolyzercellstudiesin thispaper,includingtheanalyticalmodelusedtodescribetheoperation characteristicsoftheelectrolyticcell,areintroduced.

2.1. Insituwaterelectrolysischaracteristics

One of the key issues of the gas-fermentation-based hydrogen- oxidizingbacteriaproductionisthemasstransferofthehydrogengas tothecultivationmedium,eventhoughahydrogengasconversionef- ﬁciencyupto81%hasbeenreported[13].Themasstransferproblems canbeeﬀectivelyavoidedbyBES,wherethein situwaterelectroly- sistakesplacedirectlyinthecultivationmedium.However,theinsitu waterelectrolysisimposessomeconstraintsonthesystem.Firstly,the temperaturesandpressuresmustbeinafavorablerangefortheHOB.

Secondly,thecurrentdensitiesappliedtothewaterelectrolysismustbe limitedtolevelsnotharmingtheHOB.Finally,numerousrequirements aresetontheelectrolyte,whichalsoactsasacultivationmedium,and theelectrodes themselves. Contrarytothetraditionalalkalinewater electrolysis,thecultivationmedium mustofferanalmostpH-neutral environmentfor thebacteria.This constraintis connectedtotheki- netically sluggish oxygen evolution reaction (OER), which produces a highactivation overvoltage. Further,the side reactions producing toxiccompoundsmustbeprevented.Inpractice,theselimitationslead toasignificantlylowerconductivityoftheelectrolytecomparedwith thetraditionalalkaline electrolysis.Therefore, toachievean accept- ableenergyefficiencyofthewaterelectrolysis,relativelylowcurrent densitieshavetobeapplied,whichleadstolargeelectrodeareas,yet thedistancebetweentheelectrodesisminimized.Becauseofthelarge electrodearea,low-costelectrodematerialsarepreferred.Finally,the electrodematerialsmustbecorrosionresistantnottoreleaseanytoxic compoundstothecultivationmedium.

2.2. HOBcultivationexperimentswithinsituwaterelectrolysis

ThecultivationofHOBwithinsituwaterelectrolysishasbeensuc- cessfullytested in asmall scaleelectrobioreactor[20]. Theresearch utilizedaBESwithinternalliquidvolumeof60 mlasshowninFig.1a.

CO₂ gas wasfed tothereactor,while hydrogenandoxygenfor the microbialgrowthandCO₂ﬁxationweregeneratedinsidethereactor vesselatastainlesssteelcathodeandaniridiumoxidecoatedtitanium anode.Theelectrodesweremanufacturedfromwiresofaforementioned materialswhichwereloopedincoilssothatthesurfaceofeachelec- trodewas13 cm².

InFig.1b,thebiomassincreaseofahydrogenenrichmentculture ispresented.Theculture wasamixedpopulation ofyetunidentiﬁed species,whichhadevolvedatleastsomeresistancetowardstheBESen- vironment.Thebioreactorwasfedwith0.13 g h⁻¹gaseousCO₂,and suppliedwithelectrolysiscurrentof18 mA,whichroughlyequalscur- rentdensityof1 mA cm⁻²atthesurfaceoftheelectrodes,withaverage cellvoltageof2.31 V.

Thecellmassincreasesin linearfashion asthegrowthis limited bytheavailabilityofhydrogen. Assumingfaradiceﬃciencyof unity fortheelectrolysisofwaterandcompleteconsumptionofthehydro- gen,theapparentbiomassyieldfromhydrogenwascalculatedtobe 2.5 g_biomass/mol_H

2.Matassaet al.havecollectedbiomasstohydrogen

(4)

G. Givirovskiy et al. Heliyon 5 (2019) e01690

Fig. 1.Experimentalsetupusedforthecultivationtests:(a)schemeofthesmall-scalein-situelectrolysisbioelectrochemicalsystem,and(b)biomassincreaseofa mixedhydrogenenrichmentculture.

Fig. 2.Experimental setup used for the electrolysis tests: (a) cross section of the electrolyzer cell, (b) photo of the experimental setup.

yieldsforvariousHOBspeciescultivatedwithgaseoushydrogenfeed [13].

The published values range between 1.12 g_biomass/mol_H

2– 4.64 g_biomass/mol_H

2,thereforetheHOBcultivationwithinsituelectrol- ysisgivesbiomassyieldcomparabletothegaseousH₂feedcultivation, butwithouttheneedforhandlingandstorageofﬂammable,andpo- tentiallyexplosive,hydrogengasandhydrogen –oxygengasmixtures.

Thevolumetricproductivityofbiomassduringthecultivationtest isbelow15 mg l⁻¹h⁻¹whilethehydrogenproductionisthelimiting factorforthegrowth.Therefore,thehydrogenproductionratemustbe improvedtoenhancethevolumetricproductivityoftheelectrobioreac- tor.Therefore,thecurrentdensityorelectrodeareamustbeincreased toimprovetheproductivity.Inthisarticle,theelectrodematerialsare studiedtoenhancethecurrentdensitywithoutloweringtheeﬃciency.

2.3. Experimentalelectrodematerialstudysetup

TheexperimentalsetupispresentedinFig.2.Thesetupconsistsof thefollowingelements: (i)anelectrolyzercellwithacross-sectional areaof2.6 cm²andaninitialdistanceof3 mmbetweentheelectrodes, (ii)aWaveNowpotentiostattoconductelectrochemicalmeasurements, (iii)awaterbathwithasubmergedLaudaheatertokeepthesystem optimalforthebacterialcultivationtemperatureof33^◦C,and(iv)a constantﬂowpumptocirculatethemediumthroughtheexternalvessel equippedwiththetemperaturemeasurement.Diﬀerentcombinationsof electrodematerials,suchasstainlesssteel(SS),nickel(Ni),graphite(C),

platinum(Pt),cobaltphosphate(CoPi),nickel-iron(NiFe),andiridium dioxide(IrO₂)depositedontoatitaniumsubstrate,weretested.Stain- lesssteeliswidelyusedmaterialbecauseoftherelativelylowcostand highcorrosionresistanceinmostenvironments.ApplicabilityofSS304 materialforHOBcultivationswasfirststudiedby[16],whiletheef- fectof stainlesssteelor carbonsurfacemodificationbyCoPi orCoP electrocatalystswasfurtherinvestigatedinthesubsequentstateofthe artstudiesof thesameresearch group[17, 18, 19]. However,HOB haveshowedtohaveeffectonthecorrosionofthelowcarbonsteels [21].Further,theselected316Lhasbeenmentionedtobevulnerable tomicrobialcorrosion,andsomeothersteelshouldbeselectedifun- coatedelectrodesareusedforlongerperiodsoftime[22].Platinum iswidelyusedaselectrodematerialbecauseofitsstabilitydespitethe highcost.Nickelbasedmetalsarewidelyusedinalkalinewaterelec- trolyzers,andtherefore,usedasareferencefortheothermaterials[23].

Graphiteisalsostable,butnothighlycatalyticmaterial.CoPicoatings areshowntobeselfhealingandbiocompatibleintheliterature[17].

IrO₂coatedanodeisfoundtobeapromisingcandidateintheHOBcul- tivationexperimentsdescribedabove.Linearsweepvoltammetry(I–V) wasappliedtomeasurethecellvoltageasafunctionofcellcurrent.

Thesweeprateof thelinearsweepvoltammetrywas selected tobe 10 mV s⁻¹tomitigatetheeﬀectofcellcapacitances.

Themineralmedium,used for thebioelectrochemicalcultivation ofhydrogen-oxidizingbacteriapreparedaccordingtotheDSM-81-LO4 recipeattheVTTTechnicalResearchCentreofFinland,wasapplied asanelectrolyteinthestudy.Oneliterofthemediumsolution con- 3

(5)

Fig. 3.Scanningelectronmicroscope(SEM)imagesofthecobaltphosphate(CoPi)coatingonto(a)graphitesubstrate,(b)stainlesssteelsubstrate,and(c)nickel-iron (NiFe)coatingontostainlesssteelsubstrate.

taineddistilledwater,50 mlofphosphatebuﬀer,2.3 g(KH₂PO₄),2.9 g (Na₂HPO₄),2 ml(NH₄)(Fe)(citrate),0.005 gofferricammoniumcitrate (16%Fe),10 mlof(NaHCO₃)solution,0.5 g(NaHCO₃),mineralsalts, 5.45 g(Na₂SO₄),1.19 g((NH₄)₂SO₄),0.5 g(MgSO₄⋅5H₂O),0.0117 g (CaSO₄⋅2H₂O),0.0044 g(MnSO₄⋅1H₂O),0.005 g(NaVO₃),and5 ml oftraceelementsolution.500 mloftraceelementstocksolutionwas madefrom0.05 g(ZnSO₄⋅7H2O),0.15 g(H₃BO₃),0.1 g(CoCl₂⋅6H2O), 0.005 g(CuCl₂⋅2H₂O),0.01 g(NiCl₂⋅6H₂O),and0.015 g(Na₂MoO₄).

Thephosphatebuﬀer,theammoniumiron(III)citrate,themineralsalts, andthetraceelementsolutionswereautoclavedseparately.Thevita- minsolution(NaHCO₃)wasﬁltersterilized.Thesolutionswerecom- binedasepticallyatroomtemperature.ThepHandconductivityofthe medium,measuredbeforeandaftertheelectrolysistests,were7and 12 mS cm⁻¹,respectively.

2.4. Insitucatalystformation

Electrodeposition of coatings based on earth-abundant first-row transitionmetalssuchasCoandFe–Niisconsideredanefficientmethod for theelectrodesurfacestructure modificationandenhancementof theelectrochemicalactivity.Inthepresentstudy,insitupreparation ofcoatingswasperformedintheexperimentalsetupdescribedinthe previoussectionbasedontheelectrodepositionstrategiesadoptedfrom [16] and[24].Cobaltphosphate(CoPi)coatingwaselectrodeposited ontographiteandstainlesssteelplates(substrates)inasolutioncon- taining0.1MKH₂PO₄and0.5 mMCo(NO₃)₂⋅6H2Oand250 mlofdis- tilleddeionizedwater.Pretreatmentoftheelectrodesamplesincluded polishingwithsandpaperandrinsingwithacetoneanddeionizedwa- ter.Electrolyticdepositionwascarriedoutbybulkelectrolysisat2 Vfor 5 hforthegraphitesubstrateandfor3 hforthestainlesssteelsubstrate.

Graphiteandstainlesssteelwereusedastheauxiliaryandreference electrodeforthecorrespondingexperimentsinatwo-electrodesystem.

SolutionwithtwotimesincreasedconcentrationofCo₂⁺ wasalsode- positedontothestainlesssteelsubstratetoinvestigatetheinﬂuenceof theincreasedcobaltmassonthecoatingstructureandtheelectrochem- icalperformanceofthesynthesizedcatalyst.

Insituformationofnickel-iron(NiFe)coatingwascarriedoutbythe bulkelectrolysismethodat2.8 Vfor15 mininthesolutioncontaining 0.1MNa₂SO₄,0.25MNiSO₄⋅6H₂O,0.25MFeSO₄⋅7H₂O,and250 ml ofdistilleddeionizedwater.AsmallamountofH₂SO₄wasaddedtothe solutiontoadjustthepHto2.Astainlesssteelplatewiththeafore- mentioned pretreatmentwas used as a substratefor theelectrolytic depositionofthenickel-iron(NiFe)ﬁlm.Scanningelectronmicroscope (SEM)imagesoftheobtainedcobaltphosphate(CoPi)andnickel-iron (NiFe)structuresarepresentedinFig.3.

2.5. Cellmodel

Inneutralconditions(pH=7),thewaterelectrolysisisdescribedby thefollowingelectrochemicalreactions[7].Oxidationhalf-reactionat theanode–oxygenevolutionreaction(OER):

2H₂O⟶O₂+ 4H⁺+ 4e⁻, 𝐸0= 0.817V (1) Reductionhalf-reactionatthecathode–hydrogenevolutionreaction (HER):

4H₂O+ 4e⁻⟶2H₂+ 4OH⁻, 𝐸0= −0.413V (2) Theoverallreactionintheelectrolyticcell:

2H₂O+electrical energy⟶O₂+ 2H₂, 𝐸0= −1.23V (3) Theaboveequationsdemonstratethattheequilibriumorreversible cellvoltage,whichisthelowestpotentialrequiredfortheelectrolysis totakeplaceat25^◦C and1atm,isequalto1.23 V.However,inpractice,highervoltagesarerequiredtodissociatewater;thisisduetothe additionalovervoltagespresentedinthefollowingequation:

𝑈cell=𝑈rev+𝑈ohm+𝑈act+𝑈con, (4) where𝑈cellisthecellvoltage,𝑈revisthereversibleopencircuitvoltage, 𝑈ohmistheovervoltagecausedbyohmiclossesinthecellelements,𝑈act

istheactivationovervoltagecausedbyelectrodekinetics,and𝑈con is theconcentrationovervoltagecausedbymasstransportprocesses[1].

(6)

Inelectrolysis,theproductionofhydrogenandoxygenisdirectly proportionaltothemeanvalueofthecurrentﬂowingthroughtheelec- trolyzercell.Thus,thehydrogenandoxygenproductionrates(mol s⁻¹) ofasingleelectrolyticcellcanbeexpressedas:

𝑓H₂=𝜂F

𝑖cell𝐴cell

𝑧𝐹 , (5)

where𝑧 (𝑧=2and4for hydrogenandoxygen,respectively)is the numberofmolesofelectronstransferredinthereaction,𝐹 istheFara- dayconstant(9.6485×10⁴C mol⁻¹),𝑖cellisthecurrentdensity(A cm⁻²), 𝐴cellistheeffectivecellarea(cm²),and𝜂𝐹 istheFaradayefficiency, alsoknownasthecurrent efficiency.Inthis study,theFaraday effi- ciencycanbeassumedtobeunitybecausethereshouldbenoleakage currents,andfurther,astheproductgasisamixtureofhydrogenand oxygen,thereisnoleakageofhydrogentotheoxygenlineasintradi- tionalelectrolyzers[25].Therefore,thehydrogenproductionratecan bedirectlyestimatedbasedoncurrent,andthevoltageeventuallyde- scribestheenergyefficiencyofthecell.

Asimpliﬁedmodeltodescribetheelectrolyticcellvoltagebehavior asafunctionofcurrentisintroduced.Theopen-circuitvoltagecanbe describedusingtheNernstequation[26]

𝑈rev=𝑈rev⁰ +𝑅𝑇el

𝑧𝐹 ln

⎛⎜

⎜⎝ 𝑝H₂⋅𝑝^1∕2_O₂

𝑝H₂O

⎞⎟

⎟⎠, (6)

where𝑈_rev⁰ isthereversiblecellvoltage,𝑅istheuniversalgasconstant (8.3144621 J mol⁻¹K⁻¹),and𝑇elisthetemperature.Further,𝑝H₂,𝑝O₂, and𝑝H₂Oarethehydrogen,oxygen,andwaterpartialpressures.

Thereversiblecellvoltageisdeﬁnedasafunctionoftemperature;

for example,fora PEMelectrolyzer cellin [27] andfor analkaline electrolyzer cellwith theKOH electrolytein [28]. However,in this simpliﬁedcase,theopen-circuitcellvoltageunderconstantoperating temperatureandatmosphericpressureisconsideredasoneparameter tobefoundbythecurveﬁttingofthemeasureddata.

Theohmicoverpotentialismainlycausedbythevoltageacrossthe cultivationmediumwiththeconductivityintherangeof10 mS cm⁻¹ astheconductivityoftitaniumorstainlesssteelelectrodesisroughly 2.5 kS cm⁻¹.Therefore,theohmicoverpotentialcanbeexpressedas 𝑈ohm=𝛿m𝑖cell

𝜎m , (7)

where𝛿misthedistancebetweentheelectrodesin(cm),and𝜎misthe conductivityofthemediumin(S cm⁻¹).

The activation overpotential is typically described by using the Butler–Volmerequation[29]

𝑈act=𝑅𝑇el

𝛼an𝐹arcsinh ( 𝑖cell

2𝑖o,an

) + 𝑅𝑇el

𝛼cat𝐹arcsinh ( 𝑖cell

2𝑖o,cat

)

, (8)

where𝛼isthechargetransfercoefficientfortheanodeandthecathode separately,and𝑖oistheexchangecurrentdensityontheelectrodesur- faces.Thechargetransfercoefficientsandtheexchangecurrentdensi- tiesareexperimentallydefinedasafunctionoftemperatureforexample in[30].

Finally,thesimpliﬁed modelfor thecellvoltageasafunctionof currentcanbeexpressedas

𝑈cell=𝑈rev+𝛿m𝑖cell

𝜎m

+𝛼arcsinh (𝑖cell

2𝑖0

)

, (9)

where𝑈rev,𝜎m,𝛼,and𝑖0aretheparameterstobeﬁttedbytheexperi- mentaldata.

3. Results& discussion

Graphitewasusedasanelectrodematerialfortheﬁrstbioelectro- chemicalcultivationtestsofanacetogenicmicroorganismin[8].Nickel and stainless steel have traditionally been used with alkaline elec- trolyzerswhereasnoblemetalsandtheiroxides,suchasplatinumand

Fig. 4.Experimentalresultsofwaterelectrolysiswithstainlesssteel(SS)electrodes,obtainedwithavariabledistance,andthemodelingresultswithEq.

(9).

iridiumdioxide,areknownfortheirhighcatalyticactivity.Therefore, theperformanceoftheaforementionedmaterialswasstudiedforthe electrolyteintroducedinSection2.Thepotentialofcoatedelectrodes preparedbyelectrodepositionofCoandaFe–Nialloyasapossiblesub- stituteforelectrodesmadeofpreciousmetalswasalsoevaluated.

BothhydrogenandoxygenareimportantinthecultivationofHOB, andthus,amembrane-freeelectrolyzercellprototypewasusedforthe electrolysistests.Theabsenceofamembranemakesitpossibletode- creasethedistancebetweentheelectrodesandincreasetheelectrical eﬃciency,whichisespeciallyimportantinpH-neutralconditions.How- ever,theﬂowofbiomassthroughtheelectrolyzercanbecomeanissue atverylowdistancesbetweentheelectrodes.Hence,theelectrodema- terialsweretestedatdistancesvaryingfrom3 mmto16 mmtocollect thevoltage-currentcharacteristicsoftheelectrolyzercellasafunction ofdistance betweentheelectrodes.Inthepresentsection, thelinear sweepvoltammetryresultsforvariousanodeandcathodematerialsets oftheinsitu waterelectrolysisarepresentedandanalyzedwiththe developedcellmodel.

3.1. Stainlesssteelelectrodes

First, stainless steel electrodes were used as the anode and the cathode.Themainsolutesof theSanmac316Lalloyperweightare:

chromium17.0%, nickel10.1%, molybdenum 2.0%, andmanganese 1.6%.Thedistanceoftheelectrodeswasvariedtostudytheeffectof distanceonthecellvoltage.Further,theresultsareusedtoverifythe simplifiedcellmodel.Asonlythedistancebetween theelectrodes is changedandresistiveconductionlossesaredescribedbythemedium conductivity,allthemodelparametersshouldmatcheachotherinall cases.Aminimum distance of3 mmbetween theelectrodeswas se- lectedtolimitthe flowresistance of theelectrolyte. Further,itwas assumedthatdistances exceeding10 mmcannotbe usedbecauseof thelowconductivityoftheelectrolyte.Thecellvoltageasafunctionof currentdensityisshowninFig.4.

Thedistancebetweentheelectrodeshasasignificantimpactonvolt- ageowingtothehighohmiclossescausedbythelowconductivityof themedium.Ifthevoltageefficiencyoftheelectrolysisisrequiredtobe higherthan50%,consideringthethermoneutralvoltageof1.48 V,the currentdensitycannotexceedthevalueof10 mA cm⁻²atthedistance of3 mmbetweentheelectrodesasthecurrentdensitiesincommercial alkalineelectrolyzersareupto500 mA cm⁻²[31].Atgreaterdistances theallowedcurrentdensitywouldbeevenlower.Therefore,itcanbe concludedthatthedistancebetweentheelectrodesshouldbeasshort aspossibletoachieveahighefficiencyandacompactstructure.The parameters𝑈rev,𝜎m, 𝛼,and𝑖0 in Eq.(9) were determinedusing ex- perimentalvoltageandcurrentdataandthemethodofnonlinearleast squareregression,andpresentedinTable1.Further,thereversiblevoltage,theohmicvoltage,andtheactivationvoltagetermsarepresented separatelyinFig.5.

5

(7)

Fig. 5.Reversiblevoltage,ohmicovervoltage,andactivationovervoltageasafunctionofcurrentdensityforthewaterelectrolysisexperimentswithavariable distancebetweenthestainlesssteelelectrodes:(a)𝛿m= 3mm,(b)𝛿m= 5mm,(c)𝛿m= 9mm,and(d)𝛿m= 16mm.

Table 1

Experimentallyﬁttedparametersofthesimpliﬁedcellmodelwithstainlesssteel electrodes.

𝛿m(mm) 𝑈rev(V) 𝜎m(S cm⁻¹) 𝛼(-) 𝑖0(A cm⁻¹)

3 1.905 0.012 0.393 0.0010

6 2.058 0.012 0.425 0.0021

9 2.132 0.012 0.530 0.0036

16 1.92 0.012 0.278 0.0007

Wecanseethattheohmicoverpotentialbecomeshigherthanthe activation overpotential at relatively moderate current densities of 3 mA cm⁻¹–25 mA cm⁻¹dependingonthedistancebetweentheelec- trodes.Atgreatdistancesbetweentheelectrodestheohmicoverpoten- tialevenexceedsthereversiblevoltage.Thereversiblevoltageandthe activationoverpotentialarealmostthesamewithalldistancesbetween theplates,assupposed,thatsupportstheuseofthesimpliﬁedmodel.

3.2. Anodematerialcomparison

According toEq. (1), thepotential of the anode half reaction is higher thanthepotential ofthecathodehalfreaction. Therefore,all thestudiedmaterialswereappliedtotheanodeasthecathodeismade ofstainlesssteel.Thecellvoltageswithdiﬀerentanodematerialswith theelectrodedistanceof3 mmarepresentedasafunctionofcurrent densityinFig.6.

Ascan beseen inFig.6,theanodematerialhasasignificantef- fectonthecellvoltage,especiallyathighercurrentdensities.Graphite clearlyexhibitstheworstperformancewiththehighestcellvoltage,and thenickelandplatinumanodeshavevoltagesrelativelyclosetoeach other.Theiridium-dioxide-coatedanodeisobviouslythemostfavorable anodematerialofthestudiedmaterials.Withtheiridiumdioxide,acur- rentdensityof15 mA cm⁻²canbeachievedwithavoltageefficiency of50%.Thereversiblevoltage,theohmicvoltage,andtheactivation voltagetermsasafunctionofcurrentdensitywithdifferentanodema-

Fig. 6.Cellvoltageasafunctionofcurrentdensitywithvariousanodematerials andastainlesssteelcathode.Thesolidlinesindicatethemeasureddataandthe dashedlinesrepresentthesimpliﬁedmodel.

Table 2

Experimentallyﬁttedparametersofthesimpliﬁedcellmodelwithvariousan- odematerials.

Anode 𝑈rev(V) 𝜎m(S cm⁻¹) 𝛼(-) 𝑖0(A cm⁻¹)

C 2 0.012 0.455 0.0010

Ni 2 0.012 0.338 0.0007

Pt 1.975 0.012 0.332 0.0007

IrO₂ 1.766 0.012 0.351 0.0013

terialsarecomparedwitheachotherinFig.7andthemodelparameters areshowninTable2.

The material selection signiﬁcantly aﬀects the reversible voltage andtheactivationvoltage.The iridiumoxideyieldsaslightlylower reversiblevoltagecompared withtheothermaterials.Theactivation overpotentialishighestinthecaseofthegraphiteanodeastheactiva- tionoverpotentialswiththeothermaterialsareinthesamerangewith eachother.Furthermore,theresistivevoltagelossismainlycausedby

(8)

Fig. 7.Separatedovervoltagesforthewaterelectrolysisexperimentswithdiﬀerentanodematerials:(a)graphite(C),(b)nickel(Ni),(c)platinum(Pt),and(d) iridiumdioxide(IrO₂).

Fig. 8.Cellvoltageasafunctionofcurrentdensitywithcoatedelectrodesused astheanodesandstainlesssteelasthecathode.

thelow-conductivityelectrolytemedium,andthus,theelectrodemate- rialhaspracticallynoimpactonit.

3.3. Coatedelectrodes

Subsequently,coatedelectrodeswereappliedtotheanodeandstain- less steel was used as the cathode. The cellvoltages with diﬀerent coatedanodematerialswiththeelectrodedistanceof 3 mmarepre- sentedasafunctionofcurrentdensityinFig.8.

ItcanbeclearlyseenfromFig.8thatcoatedelectrodescanbecon- sideredanattractivealternativeforcatalystsmadeofpreciousmetals.

Electrodepositionof CoandtheNi-Fealloyenablessubstratesurface structure modification by enhancement of the electrochemically ac- tivesurfacearea,whichiswellshowninFig.3.Theobtainedhighly orderedCoPicoatingsexhibitedabetterperformancethanthePtan- ode,whereastheperformanceoftheNi-Fefilmwascomparablewith theIrO₂ anode.Acurrentdensity ofapproximately14 mA cm⁻² was achievedwithavoltageefficiencyof50%whenusingstainless steel coatedwiththeNi-Fealloy.Itwasalsofoundthatthesubstratematerial

Table 3

Experimentally ﬁtted parameters of the simpliﬁed cellmodel with various coatedelectrodesusedastheanodematerials.

Anode 𝑈rev(V) 𝜎m(S cm⁻¹) 𝛼(-) 𝑖0(A cm⁻¹)

C(CoPi-sol.1) 1.790 0.012 0.443 0.0014

SS(CoPi-sol.1) 1.630 0.012 0.359 0.0006

SS(CoPi-sol.2) 1.695 0.012 0.370 0.0009

SS(NiFe) 1.449 0.012 0.338 0.0004

hadaneﬀectontheelectrochemicalperformanceoftheelectrode.The performanceoftheCoPicoatingonthegraphitesubstratewasslightly lowerthantheperformanceofthesamecoatingelectrodepositedonto stainlesssteelsubstrates.TheperformancesoftheCoPi coatingselec- trodepositedontothestainlesssteelsubstratefromsolution1andso- lution2with0.5and1mMconcentrationsofCo₂+,respectively,were similar.

Thereversiblevoltage,theohmicvoltage,andtheactivationvoltage termsasafunctionofcurrentdensitywithdiﬀerentcoatedanodesare comparedwitheachotherinFig.9andthemodelparametersareshown inTable3.

3.4. Cathodematerialcomparison

Finally,themostpromisinganodematerialswerealsousedasthe cathodematerialtoseeiftheperformancecanbefurtherimproved.The cellvoltageswithdiﬀerentanodeandcathodematerialcombinations arepresentedasafunctionofcurrentdensityinFig.10,andthemodel parametersaresummarizedinTable4.

Currentdensitiesof 25,15,and10 mA cm⁻²wereachievedwith thevoltageeﬃciencyof50%for IrO₂,Pt, andSS usedforboth the anodeandthecathode.Inthepreviousresearch[16],HOBmanagedto tolerateandgrowatcurrentdensitiesuptoapproximately4 mA cm⁻² with2.5 Vcellpotential.Furtherincreaseofthedrivingvoltageupto 3 Vresulted intheexponentialincrease ofthecelldensitiesandthe 7

(9)

Fig. 9.Separatedovervoltagesforthewaterelectrolysisexperimentswithdiﬀerentcoatedelectrodesusedastheanodematerials:(a)graphitecoatedwithCoPi usingsolution1,(b)stainless steelcoatedwithCoPiusingsolution1,(c)stainless steelcoatedwithCoPiusingsolution2,and(d)stainless steelcoatedwithNiFe.

Fig. 10.Cellvoltageasafunctionofcurrentdensitywiththebest-performing anodeandcathodematerials.

Table 4

Experimentallyﬁtted parametersofthesimpliﬁedcellmodel withthebest- performingelectrodematerials.

An./Cath. 𝑈_rev(V) 𝜎_m(S cm⁻¹) 𝛼(-) 𝑖0(A cm⁻¹)

Pt-Pt 1.4 0.012 0.255 0.0002

IrO₂-IrO₂ 1.366 0.012 0.227 0.0004

highestreportedvalueforthecurrentdensity,whichbacteriamanaged totolerate,was11 mA cm⁻².

It is importanttonote,thatthere was nosubstantial increase in theperformancewhenusingCoPiandNi-Fecoatedcatalystsforboth electrodesincomparisonwiththeexperimentswhereSSwasusedas thecathodematerial.Thus,wecanstatethattheaforementionedcoat- ings arecatalyticallyactive foroxygenevolutionreaction (OER)but donotexhibithighcatalyticactivityforhydrogenevolutionreaction (HER).EventhoughIrO₂showsthebestperformanceastheanodeand cathodematerial,theSSperformanceisstillacceptablewhenconsider- ingthehighmanufacturingcostofcatalystsmadeofpreciousmetals.

Furthermore,theSScan beconsideredapotentialcost-eﬀectivesub-

stratematerial forelectrodepositionof coatings.Itcan beconcluded thatinaneutralenvironmentthecathodematerialalsohasasignif- icant eﬀect on the water electrolysis performance. The celloverpo- tentialswithdiﬀerent electrodematerial combinationsareshown in Fig.11.

Thereversiblevoltagewithboththeplatinum- andiridium-dioxide- coated cathodes is signiﬁcantly lower than with the stainless steel cathode.Further,theiridium-dioxide-coatedcathodeexhibitsalower activationoverpotentialthanplatinum.

4. Conclusions

Inthepresentpaper,asimpliﬁedcellmodelwasproposed tode- scribe thecellvoltagecomponents asa function of current density.

Itisnoteworthythatthechemicalformulationoftheelectrolytesig- nificantly affectsthe electrical resistanceof theelectrolysis celland thereby the energy efficiency of the whole process. The developed modelwasimplementedtoanalyzetheapplicabilityofnumerouselec- trode materials for the in situ electrolysis of a pH-neutral medium forbioelectrochemicalcultivationofhydrogenoxidizingbacteria.The modelenablesquantitiveevaluationof thereversiblevoltage,ohmic overpotential,andactivationoverpotentialfordifferentsetsofelectrode materials.

TheobtainedhighlyorderedCoPiandNi-Fecoatingsexhibitedan oxygenevolution reaction (OER) performanceexceeding thatof the Ptanode andbeing comparablewith theIrO₂ anode. However, the aforementionedcoatingsdid not show asubstantial improvementin performanceforthehydrogenevolutionreaction(HER)comparedwith thestainlesssteelcathode.Basedonthisobservation,wecanconclude thatadditionalresearchisrequiredtoﬁndsuitablecoatingswithhigh electrocatalyticperformancefortheHER.

Thelowestcellvoltageasafunctionofcurrentdensitywasreached withtheIrO₂coatingbothattheanodeandthecathode.Withthestain- lesssteelelectrodes,thesamevoltagelevelwasachievedatroughly 50%lowercurrentdensitiescomparedwiththeIrO₂-coatedelectrodes.

(10)

Fig. 11.Separatedovervoltagesforthewaterelectrolysisexperimentsofthebestelectrodematerialcombinations:(a)platinum-platinum(Pt–Pt),and(b)iridium- iridium(IrO₂–IrO₂).

Despite this, the stainless steel can be considered a potential cost- eﬀectivesubstratematerialforpreparationofcoatingsinelectrobiore- actorswithinsituelectrolysisofmedia.

Adetailedenergyeﬃciencyanalysisofthebioelectrochemicalsys- temandananalysisoftheeﬀectsoftheinsituwaterelectrolysison themicrobialgrowth,e.g.maximumallowablecurrentdensity,willbe conductedinthefurtherresearchintothetopic.

Declarations

Authorcontributionstatement

GeorgyGivirovskiy,VesaRuuskanen,JeroAhola,LeoOjala,Michael Lienemann,PetteriKokkonen:Conceivedanddesignedtheexperiments;

Performed theexperiments;Analyzedandinterpretedthedata;Con- tributedreagents,materials,analysistoolsordata;Wrotethepaper.

Fundingstatement

This work was supported by Finnish Academy of Science for

“MOPED–MicrobialOilandProteinsfrom Airby Electricity-Driven Microbes”projectfundingundernumber295866.

Competingintereststatement

Theauthorsdeclarenoconﬂictofinterest.

Additionalinformation

Noadditionalinformationisavailableforthispaper.

References

[1]A.Ursúa,L.Gandía,P.Sanchis,Hydrogenproductionfromwaterelectrolysis:cur- rentstatusandfuturetrends,Proc.IEEE100 (2)(2012)410–426.

[2]I.Dincer,C.Acar,Reviewandevaluationofhydrogenproductionmethodsforbetter sustainability,Int.J.Hydrog.Energy40 (34)(2014)11094–11111.

[3]T.Kousksou,P.Bruel,A.Jamil,T.ElRhaﬁki,Y.Zeraouli,Energystorage:applica- tionsandchallenges,Sol.EnergyMater.Sol.Cells120 (PARTA)(2014)59–80.

[4] UnitedNationsEnvironmentProgramme,Thehydrogeneconomy:anon-technical review, https://www.unenvironment.org/resources/report/hydrogen-economy- non-technical-review,2006.

[5]S.E.Hosseini,M.A.Wahid,Hydrogenproductionfromrenewableandsustainable energyresources:promisinggreenenergycarrierforcleandevelopment,Renew.

Sustain.EnergyRev.57(2016)850–866.

[6]M.Lehner,R.Tichler,H.Steinmüller,M.Koppe,Power-to-Gas:Technologyand BusinessModels,SpringerInternationalPublishing,NewYork,2014.

[7]Y.Cheng,S.P.Jiang,Advancesinelectrocatalystsforoxygenevolutionreactionof waterelectrolysis-frommetaloxidestocarbonnanotubes,Prog.Nat.Sci.25 (6) (2015)545–553.

[8] K.P.Nevin,T.L. Woodard,A.E. Franks,Z.M. Summers,D.R.Lovley,Microbial electrosynthesis:feedingmicrobeselectricitytoconvertcarbondioxideandwa- tertomulticarbonextracellularorganiccompounds,mBio1 (2)(2010),http://

mbio.asm.org/content/1/2/e00103-10.full.pdf.

[9]S.Bajracharya,M.Sharma,G.Mohanakrishna,X.DominguezBenneton,D.P.Strik, P.M.Sarma,D.Pant,Anoverviewonemergingbioelectrochemicalsystems(BESs):

technologyforsustainableelectricity,wasteremediation,resourcerecovery,chem- icalproductionandbeyond,Renew.Energy98(2016)153–170.

[10]G.Kumar,R.G.Saratale,A.Kadier,P.Sivagurunathan,G.Zhen,S.H.Kim,G.D.

Saratale,Areviewonbio-electrochemicalsystems(BESs)forthesyngasandvalue addedbiochemicalsproduction,Chemosphere177(2017)84–92.

[11]J.Yu,A.Dow,S.Pingali,Theenergyeﬃciencyofcarbondioxideﬁxationbya hydrogen-oxidizingbacterium,Int.J.Hydrog.Energy38 (21)(2013)8683–8690.

[12]T.G.Volova,V.A.Barashkov,Characteristicsofproteinssynthesizedbyhydrogen- oxidizingmicroorganisms,Appl.Biochem.Microbiol.46 (6)(2010)574–579.

[13]S.Matassa,W.Verstraete,I.Pikaar,N.Boon,Autotrophicnitrogenassimilationand carboncaptureformicrobialproteinproductionbyanovelenrichmentofhydrogen- oxidizingbacteria,WaterRes.101(2016)137–146.

[14]S.Matassa,N.Boon,I.Pikaar,W.Verstraete,Microbialprotein:futuresustain- ablefoodsupplyroutewithlowenvironmentalfootprint,Microb.Biotechnol.9 (5) (2016)568–575.

[15]F.Oesterholt,S.Matassa,L.Palmen,K.Roest,W.Verstraete,Pilotscaleproduction ofsinglecellproteinsusingthepower-to-proteinconceptfutureglobalchallenges, in:2ndInternationalResourceRecoveryConference,2018,pp. 1–16.

[16]J.P.Torella,C.J.Gagliardi,J.S.Chen,D.K.Bediako,B.Colón,J.C.Way,P.A.Sil- ver,D.G.Nocera,Eﬃcientsolar-to-fuelsproductionfromahybridmicrobial–water- splittingcatalystsystem,Proc.Natl.Acad.Sci.112 (8)(2015)2337–2342.

[17]C.Liu,M.Ziesack,P.A.Silver,Watersplitting –biosyntheticsystemwithCO₂reduc- tioneﬃcienciesexceedingphotosynthesis,Science352 (6290)(2016)1210–1213.

[18]C.Liu,K.K.Sakimoto,B.C.Colón,P.A.Silver,D.G.Nocera,Ambientnitrogenreduc- tioncycleusingahybridinorganic–biologicalsystem,Proc.Natl.Acad.Sci.114 (25) (2017)6450–6455.

[19]C.Liu,B.E.Colón,P.A.Silver,D.G.Nocera,Solar-poweredCO₂reductionbyahy- bridbiological∣inorganicsystem,J.Photochem.Photobiol.A,Chem.358(2018) 411–415.

[20] M.Wuokko,ChemolithoautotrophicGrowthofKnallgasBacteriainanElectrobiore- actorUsinginsituWaterElectrolysis,Master’sthesis,2017,http://urn.ﬁ/URN:NBN: ﬁ:aalto-201705114594.

[21]R.Moreira,M.K.Schütz,M.Libert,B.Tribollet,V.Vivier,Fluenceofhydrogen- oxidizingbacteriaonthecorrosionoflowcarbonsteel:localelectrochemicalinves- tigations,Bioelectrochemistry97(2014)69–75.

[22] Sandvik, Microbiologically influenced corrosion (MIC), https://www.materials. sandvik/fi-fi/tietopankki/korroosiotietoja/wet-corrosion/microbiologically- influenced-corrosion-mic/,2019.(Accessed 4April2019).

[23]M.Schalenbach,A.R.Zeradjanin,O.Kasian,S.Cherevko,Aperspectiveonlow- temperaturewaterelectrolysis–challengesinalkalineandacidictechnology,Int.J.

Electrochem.Sci.13(2018)1173–1226.

[24]K.H.Kim,J.Y.Zheng,W.Shin,Y.S.Kang,PreparationofdendriticNiFeﬁlmsby electrodepositionforoxygenevolution,RSCAdv.2(2012)4759–4767.

[25]M.Schalenbach,M.Carmo,D.Fritz,J.Mergel,D.Stolten,PressurizedPEMwater electrolysis:eﬃciencyandgascrossover,Int.J.Hydrog.Energy38 (35)(2013) 14921–14933.

[26]A.Awasthi,K.Scott,S.Basu,Dynamicmodelingandsimulationofaprotonex- changemembraneelectrolyzerforhydrogenproduction,Int.J.Hydrog.Energy 36 (22)(2011)14779–14786.

[27]K.Harrison,E.Hernándec-Pacheco,M.Mann,H.Salehfar,Semiempiricalmodelfor determiningPEMelectrolyzerstackcharacteristics,J.FuelCellSci.Technol.3 (2) (2005)220–223.

[28]R.LeRoy,C.Bowen,D.LeRoy,Thethermodynamicsofaqueouswaterelectrolysis, J.Electrochem.Soc.127 (9)(1980)1954–1962.

9

(11)

[29]J.Larminie,A.Dicks,FuelCellSystemsExplained,JohnWiley&SonsLtd.,England, 2003.

[30]C.Biaku,N.Dale,M.Mann,H.Salehfar,A.Peters,T.Han,Asemiempiricalstudyof thetemperaturedependenceoftheanodechargetransfercoeﬃcientofa6 kWPEM electrolyzer,Int.J.Hydrog.Energy33 (16)(2008)4247–4254.

[31]B.Decourt,B.Lajoie,R.Debarre,O.Soupa,TheHydrogen-BasedEnergyConversion FactBook,TheSBCEnergyInstitute,2014.