Process development for biomass delignification using deep eutectic solvents. Conceptual design supported by experiments

(1)

ContentslistsavailableatScienceDirect

Chemical

Engineering

Research

and

Design

j o u r n a l ho me p a g e : w w w . e l s e v i e r . c o m / l o c a t e / c h e r d

Process

development

for

biomass

delignification

using

deep

eutectic

solvents.

Conceptual

design

supported

by

experiments

Dion

Smink,

Sascha

R.A.

Kersten,

Boelo

Schuur

∗

UniversityofTwente,FacultyofScienceandTechnology,SustainableProcessTechnologyGroup,TheNetherlands

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received10April2020

Receivedinrevisedform26August 2020

Accepted12September2020 Availableonline19September2020

Keywords: Delignification Deepeutecticsolvent Conceptualprocessdesign Liquid–liquidextraction

a

b

s

t

r

a

c

t

Deepeutecticsolvents(DES)havebeenproposedassolventsforbiomassdelignification.This paperdescribesaconceptualprocessdesignforthedelignificationofEucalyptusglobulus usingaDEScomprisedof30wt%cholinechlorideand70wt%lacticacid.Inthisdesign, theligninandhemicelluloseby-productsarerecoveredbyliquid–liquidextractionusing 2-MTHFassolvent.Materialandenergybalancesweremadeandtheenergyusageofthe processwasoptimizedwithadditionalexperiments.TheamountofDESwasreducedto theminimalamountrequiredtofilltheporousbiomass(5kgperkgwood),whichonly reducedthedelignificationfrom94%to87%andincreasedtheyieldfrom57to59%.Direct recyclingoflignin-in-DESmixtureswithoutligninremovalbyliquid–liquidextractiontothe delignificationstagemaysaveenergy,butincreasedrepolymerizationincreasesthelignin’s molarweight,whichdecreasesitsvalueandmakesrecoverybyliquid–liquidextractionmore difficult.Afteroptimization,thetotalheatdutyoftheproposedprocessis7.9GJ/tpulp,which is28%lowerthanthekraftprocess.ThemainbenefitofDESbaseddelignificationprocesses isthepossiblevalorizationofbyproducts,suchasligninandfuransfromhemicellulose.

1. Introduction

Deepeutecticsolvents(DES,plural:DESs)arecompositesolventsthat exhibitdeepeutecticbehavioruponmixingtheconstituents.Various authorsproposedadefinitionforaDES(Zhangetal.,2012;Paivaetal., 2014;Smithetal.,2014;Martinsetal.,2019;Schuuretal.,2019)but nofinaldefinitionorconsensushasbeenachievedyet.Regardlessof whataformaldefinitionofDESsshouldbe,thesecompositesolvents thatexhibittheirlowtransitiontemperatures(Franciscoetal.,2013b) duetohydrogenbonding,attractedwideattentioninacademia.Since thesesolventscaneasilybepreparedinnumerouswaysby combin-ingahydrogenbonddonorandacceptor(Gomezetal.,2018)andare oftenbiocompatible,biodegradable(Luetal.,2012)andcanhavealow toxicity(Macárioetal.,2018),DESshavebeenusedforvarious appli-cations.ExamplesincludeCO2capture(López-Salasetal.,2014;García

etal.,2015;Mirzaetal.,2015),airpollutantremoval(Mouraetal.,2017),

∗ _{Corresponding}_author.

E-mailaddress:b.schuur@utwente.nl(B.Schuur).

extractivedistillation(Panetal.,2019),metalextractions(Söldneretal., 2019),desulfurization(Limaetal.,2018,2019)andbiomass fractiona-tion(Jablonsk ´yetal.,2019;Miˇsanetal.,2019;Sminketal.,2019).

Lignocellulosicbiomassisarenewablerawmaterialsourceandcan beconvertedintocellulosefibersand,amongotherbyproducts,lignin bydelignificationtechnologies.Theobtainedcellulosepulpcanbeused forpaperproduction,productionofothermaterials,orcanbeconverted tobio-ethanolorotherplatformchemicals(Naiketal.,2010;Rosatella etal.,2011;Dusselieretal.,2013).Ligninisanaromaticbiopolymerwith advocatedpotentialforthechemicalindustryandcurrentresearchis focusingonligninvalorization(Rinaldietal.,2016;Westwoodetal., 2016).Otherbyproducts,suchashemicelluloseandextractablescan beconvertedintofuransandturpentinerespectively(Houetal.,2018b; Lietal.,2018).

Mostconventionalcellulosefiberproducinginstallationsmakeuse ofthekraftprocess,inwhichlignocelluloseisdelignifiedusingsodium hydroxideandsodiumsulfidetoformcellulosefiberssuitablefor paper-making.Theligninisdepolymerizedbyanucleophilicsubstitution reactionwiththesulfide,meaningtheligninthatisfinallyobtained ishighlysulfurized.Asaconsequence,tobeabletorecyclethesodium sulfide,theligninhastobecombusted.Kraftmillsmayalsoproduce https://doi.org/10.1016/j.cherd.2020.09.018

(2)

ligninasbyproduct,butthescaleandapplicationsarelimited.Kraft ligninismainlyusedforlow-valueapplications,suchasheat recov-ery(Chen,2015).Alsootherprocesses,suchasorganosolv,canbeused forcelluloseproduction,buttheseprocessesrequirehighamountsof organicsolvents,whichareoftenvolatileandflammable(VanOsch etal.,2017).Othersolvents,suchasionicliquids(Sunetal.,2009)may alsobeconsideredforbiomassdelignification,buttheirhighercosts maylimitindustrialapplicability.

Althoughkraftplantsarehighlyintegratedandenergyeffective plants(RudieandHart,2014),theystillhaveanaverageheatusage around11GJ/t(Suhretal.,2015).Thisheatisprovidedfromthe com-bustionoftheextractedligninandotherpartsofthetreethatare unsuitableofpapermaking,suchasthebark.Mostmodernkraftmills haveanenergysurplus,whichismostoftenusedtoprovidepowerto thegrid.

DESshavebeenreportedforthedelignificationofvarioustypesof lignocellulose.AnoverviewofthesestudiesisshowninTable1.Lignin thatisproducedusingDESsoftenshowsverylittlesignsof condensa-tion(Alvarez-Vascoetal.,2016)andtypicallyhasalowmolarweight (Alvarez-Vascoetal.,2016;Lyuetal.,2018).Manystudieshavebeen per-formedonbiomassdelignificationusingDESandmanyDES,biomass typesandprocessingconditionshavebeenstudied.However,tothe bestofourknowledge,noanalysisoftheoverallprocessenergy effec-tivenesshasbeenmade,andtheconceptualdesignasmadeinthis study,hasnotbeenmadeforbiomassdelignificationusingDESyet.

EspeciallyDESscomprisedoflacticacidandcholinechloridewere veryoften used forbiomass delignification(Jablonsk ´y etal., 2015; Alvarez-Vascoetal.,2016;Changetal.,2016;Kumaretal.,2016;Zhang etal.,2016;Lietal.,2017,2019a,2019b;Houetal.,2018a;Lyuetal.,2018; Mamillaetal.,2019;Muleyetal.,2019;Shenetal.,2019;Sminketal., 2019;Tianetal.,2019).Althoughtheoptimallacticacidtocholine chlo-rideratiosfordelignificationarebetween10:1and50:1,(Sminketal.,

2019)aDEScomprisedof30wt%cholinechlorideand70wt%lactic acidwasusedinthisstudysinceitispossibletoregenerateligninfrom thisDESbyliquid–liquidextraction,whichhaspotentialfor signifi-cantenergysavings(Sminketal.,2020a).Thisequalsa1:3.6Mratioof cholinechloridetolacticacid.

Althoughthetechnicalconceptofbiomassdelignificationusing DESsiswellproven,thelackofaprocessevaluationbasedona con-ceptualdesignhindersmakingestimationsontheeconomicfeasibility oftheconcept.Eventhoughmanydetailsarestillunknown,makinga conceptualdesignwithallthenecessaryestimationsisafirststepthat wepresentheretoidentifywhethertheconceptofDES-based pulp-inghaspotentialtobecomeindustriallyviableornot.Basedonthis study,thereadershallbeabletovaluethepotentialofthe technol-ogy,andinadditionbeabletoindicatewhichadditionaldataisstill requiredtomakeafullydetailedprocessdesign.Inthiswork,a con-ceptualdesignwasmadeaccordingtothemethoddescribedbyDouglas (Douglas,1988),whichiseasytouseandrelevantforindustrial applica-tions(Harmsen,2004).Usingthisapproach,acomplexprocessdesignis dividedintomultiple,muchsimplerproblems.FollowingtheDouglas method,theinput–outputstructureoftheprocesswasdefinedfirst, afterwhichtheseparationsweredefined.Later,themassandenergy balancesweremadeusingrigorouscalculations.Therefore,itis rela-tivelyeasytoscreenalternativesandtherebyoptimizethedesign.In thiswork,theprocesswasoptimizedforenergyusage.Energyusage isaveryimportantparameter,notjustfromanenvironmentalpoint ofviewandfortheoperationalcostsoftheprocess,butenergylosses typicallydictatetheinvestmentcostsforaprocessaswell(Lange,2001).

2. Methods

and

materials

2.1. Materials

Air-dryE.globuluschipsweredonatedbyTheNavigator Com-pany from their mill in Portugal and free from impurities, suchasstones,soilormetal.Thecommerciallysizedchips (typically25–35×10–25×2.5–6mm,L×W×T)weremilledto producewoodmealusingahammermillandsieved(mesh −25/+70) and contained 5.6wt% moisture, as determined accordingtotheNRELmethod(Sluiteretal.,2008a).Thewood contained21.6%lignin,50.6%glucose,14.0%xylose,and1.1% galactose(ovendrybasis),asdeterminedbyacidhydrolysis usingthestandardNRELmethod(Sluiteretal.,2012).Lactic acid(>85%),cholinechloride(>98%)andsulfuricacid(95–98%) werepurchasedfromSigma-Aldrich.2-MTHF(Emplura)was purchasedfromVWR.

2.2. Experimentalprocedures

2.2.1. Biomassdelignificationexperiments

1to2geucalyptusmeal(ovendrybasis)wasaddedtogether with10gDEStoaTeflon-linedautoclave.Thisautoclavewas insertedintoadirectcontactheater,whichheatedthe auto-claveto130◦Cinacoupleofminutes.After1h,theautoclave wastakenoutoftheheaterandgentlycooleddowntoroom temperature. Thecontents werewashedusing ethanoland filteredoveraWhatmanglassfiberfilterbeforedryingofthe residuetoaconstantweightat105◦Cinaconvectionoven. Allexperimentswereperformedinduplo.Theyieldwas cal-culatedaccordingtoEq.(1).

yield (%)= Drysolidresidue (g)

Eucalyptusmealaddedtoautoclave (g)∗

1−Moisturecontentineucalpytus

g_g

∗100 (1) Theacidinsolublelignininthesolid residuewas deter-mined by acid hydrolysis according to the NREL method (Sluiteretal.,2012).Thedegreeofdelignificationwas calcu-latedaccordingtoEq.(2)

Delignification (%)=100−Yield (%)∗Lignin_Ligninin_inresidue (%)_{wood (%)} (2) Theerrorinthedelignificationwasdeterminedfromone experimentthatwasrepeated 5times.The95%confidence intervalwasdeterminedusingthet-teststatisticmethod.For duploexperiments,theerrorindelignificationwas1.2%.

2.2.2. Ligninextractionexperiments

3mLlignininDESmixture(seesection2.2.3.forthe prepara-tionprocedure),3mL2-MTHFand0.5–3mLwaterwereadded toaglassvialwithascrewcap.Thevialswereplacedina JulaboSW22shakingbathat50◦Candshakenovernightat 200rpm. The phases were settled under gentle shaking at 20rpmuntiltheywerefullyseparated.Thelignincontentsin bothphasesanalyzedbyGPC.

2.2.3. PreparationoflignininDESmixture

700glacticacidand 300gcholinechloridewereaddedtoa round bottomed flask(2L) with a condenser. The mixture washeatedto130◦Cand200g(ovendrybasis)commercially sizedchipswereaddedtotheflask.Themixturewasstirred gentlyfor3or8handthemixtureswerefilteredwhilehot overa53␮msteelmesh.Thelignincontentinthemixtures

(3)

Table1–OverviewofpreviousworksonbiomassdelignificationusingDES,sortedbybiomasstype.

Woodtype Author Year DES Delignification(%)and/or

remarks

Softwoods

Spruce Wahlströmetal.(2016) ChCl+boricacidand glycerol,betaine+glycerol

Thepretreatmentefficiency ofDESforenzymatic hydrolysisislimited. Schäferetal.(2018) ChCl+3-phenylpropionic

acid

21%,Particlesizehasalarge influenceondelignification Pine

Alvarez-Vascoetal.(2016) ChCl+aceticacid,lactic acid,levulinicacidand glycerol.

58%,delignificationusing ChCl+glycerolineffective. Lietal.(2019a,2019b) ChCl+lacticacid 67%,treatmentusing

microwavereduced delignification Muleyetal.(2019) ChCl+lacticacidandoxalic

acid

90%,microwaveheating reducedprocessingtime

Hardwoods

Beech Jablonsk ´yetal.(2019) Lactic_glycineacid_and+_betaine,alanine, ChCl+ethyleneglycoland glycolicacid

15%,ChClandglycolicacid mosteffective

Mamillaetal.(2019) ChCl+lacticacid,oxalic acid,KOHandurea

OxalicacidandKOHcould dissolvemorewood polymers

Poplar

Alvarez-Vascoetal.(2016) ChCl+aceticacid,lactic acid,levulinicacidand glycerol.

78%,delignificationusing ChCl+glycerolineffective. Tianetal.(2017) Lacticacid+betaine 69%,DEShadahigher

selectivityandrequireda lowertemperaturethan soda/AQororganosolv pulping

Tianetal.(2019) ChCl+formicacid,acetic acidandlacticacid

77%,Producedpulpswere comparabletokraftpulps Lietal.(2019a,2019b) Lacticacid+ChCland

glycine

90%withChClDES,but only58%withglycineDES

Eucalyptus

Changetal.(2016) ChCl+lacticacid 90%,1:9ratiooptimal. Achievedat90◦Cand12h Liangetal.(2019) ChClandethyleneglycol 90%,DESrecoveredusing

membranes

Shenetal.(2019) ChCl+lacticacid 93%,obtainedligninhad lowweightandawell preservedstructure Sminketal.(2019) ChCl+lacticacid 92%,chlorideistheactive

ingredientincholine chloride

Willow Lietal.(2017) ChCl+lacticacid,glycerol andurea

92%,usingChCl+lacticacid at120◦Cand12h

Songetal.(2019) Lacticacid+betaine 53%,moreselectivethan hydrotropes,butrequired highertemperatures.

Agricultural byproductsand grasses

Miscanthus Guoetal.(2019a) ChCl+glycerol+acid catalyst

90%,Suitablepretreatment forenzymaticdigestion Wheat

straw

Jablonsk ´yetal.(2015) ChCl+Urea,malonicacid, lacticacid,malicacid,lactic acidandoxalicacid

58%byChCl+oxalicacid, butChCl+lacticacidhad thehighestselectivity Wahlströmetal.(2016) ChCl+boricacidand

glycerol,betaine+glycerol

Thepretreatmentefficiency ofDESforenzymatic hydrolysisislimited. Zhaoetal.(2018) ChCl+monoethanolamine 71%,suitablepretreatment

forenzymatichydrolysis Jablonsk ´yetal.(2019) Lacticacid+alanine,

glycineandbetaine, ChCl+ethyleneglycoland glycolicacid

24%,lacticacidandalanine wasmosteffective

Ricestraw

Kumaretal.(2016) ChCl+lacticacidand betaine

60%,additionofwater enhanceddelignification Houetal.(2018a) 6lacticacidand13ChCl

basedDESs

51%,withChCl+lacticacid Houetal.(2018b) ChCl+oxalicacid 70%,suitablepretreatment forenzymatichydrolysis

(4)

–Table1(Continued)

Woodtype Author Year DES Delignification(%)and/or

remarks

Limetal.(2019) K2CO3+glycerol AlkalineDESsuitablefor

ricestrawpulping Corncob

Procenteseetal.(2015) ChCl+glycerol,imidazole andurea

88%,DEStreatment enhancedenzymatic digestibility. Zhangetal.(2016) ChCl+7acidsand2

polyalcohols

98.5%,delignification enhanceswithacid strength

Guoetal.(2019b) Lacticacid

+Benzyltrimethylammonium

63%,suitablepretreatment forenzymaticdigestion Chlorideand

benzyltriethy-lammonium chloride

Cornstover Xuetal.(2018) 9ChClbasedDESs 60%,DESsuitablefor

biofuelproduction Songetal.(2019) Lacticacid+betaine 79%,moreselectivethan

hydrotropes,butrequired highertemperatures. wasdetermined byacid hydrolysis,forwhich20mLofthe

lignin-in-DESmixturewasaddedtoaTeflon-linedautoclave, togetherwith60mLwater.3mL72%Sulfuricacidwasadded tothemixtureusingapipetteandtheautoclavewasinserted intoadirectcontactheaterat120◦Candkeptthereforone hour.Themixturewascooleddowntoroomtemperatureand filteredoveraWhatmanglassfiberfilter.Theresiduewasdried inaconvectionovenat105◦Candtheacidinsolublelignin contentwasdeterminedgravimetrically.

2.3. Analysismethods

2.3.1. Acidhydrolysis

Theacidinsolublelignincontent ofthesamplewas deter-minedbyhydrolysisofthesampleaccordingtostandardized NRELprocedure(Sluiteretal.,2012).0.3gSamplewasadded toaTeflon-linedautoclave,togetherwith3mLsulfuricacid (72%)andthemixturewaskeptat30◦Cfor1hwhilestirring every10min.Afterthis,84mLwaterwasaddedandthetube waskeptat120◦Cforanotherhour.Thesolidswerefiltered anddriedovernightat105◦C.Theacidinsolublelignincontent wascalculatedaccordingtoEq.(3).

Acidinsolublelignin (%)

= Residue (g)

sample (g)∗

1−Watercontent

g_g

∗100 (3)

2.3.2. GPC

AnAgilent1200serieswasusedforthegelpermeation chro-matography(GPC)witharefractiveindexdetectorandaUV detectoroperatingat254nmusing3GPCPLgel3␮mMIXED-E columnsinseries.Thecolumnwasoperatedat40◦Canda95:5 (v:v)tetrahydrofuranandwatermixturewasthesolventata flowrateof1mL/min.Molecularweightdistributionswere cal-ibratedusingpolystyrenesolutionshavingmolecularweights rangingfrom162to27,810g/mol.

2.3.3. HPLC

Thelactic acidcontent inthehydrolysisliquidswas deter-minedbyhigh performanceliquidchromatography (HPLC).

AnAgilent1200systemwasequippedwithaHi-Plex-H col-umnoperatedat60◦Candarefractiveindexdetectorat55◦C. 5mMsulfuricacidinwaterwasusedasmobilephasewitha flowrateof0.6mL/min.

2.3.4. Karl–Fischertitration

ThewatercontentoflignininDESmixturewasdetermined by Karl–Fischer titration using a Metrohm 787KF Titrino. Hydranal composite 5 (5mg water/mL) was titrated from a 20mL burette in a 3:1 (v:v) mixture of methanol and dichloromethane.Thesamplewasmeasured induplowith arelativeerror<2%.

2.3.5. Determinationofashcontent

Theashcontentintheeucalyptuswasdeterminedaccording tothestandardNRELprocedure(Sluiteretal.,2008b).1.5–2g biomass was pre-driedin aconvection oven at105◦C and addedtoaporcelaincruciblewhichwaspre-driedat575◦C. Thiscruciblewasinsertedintoanovenat575◦Ctooxidizeall organicmaterial.Theanalysiswasperformedinduplo.The ashcontentwasdeterminedgravimetricallybyEq.(4).

Ashcontent (%)=Residueafteroxidation (g)

Drybiomass (g) ∗100 (4)

2.4. Calculationmethods

Allcalculationsinthisworkarerigorouscalculations,based on literature data, and sometimes based on experimental results reported in this paper. In order to make calcula-tions on the mass balances and energy balances of unit operations,informationisneededonphasebehavior.Where liquid–liquid equilibria of DES and 2-MTHF and lignin are involved,thisinformationistakenfromearlierpublications (Sminketal.,2020a,b).Theliquid–liquidequilibriumandvapor liquid equilibrium data for 2-MTHF and water was taken fromGlassetal.(2016)andStephenson(1992).Furthermore, it is assumed that the DES is completely non-volatile, as shownbyFranciscoetal.(2013a), andforthe enthalpiesof evaporationofwaterfromblackliquorsalsoliteraturedata wasused.For kraftblackliquor,theinformationwastaken fromFerreiraetal.(2011)(Hvap,BL=54.1kJ/mol),whileforthe

(5)

Table2–Assumptionsmadeforthemassbalanceoftheconceptualprocesswiththeirjustifications.

Symbol Description Value Unit Justification

DB DEStobiomassratio 10 kg/kg Commonratiousedinliterature

(Alvarez-Vascoetal.,2016) lB Ligninfractioninbiomass 0.216 kglignin/kgbiomass Measuredbyacidhydrolysis,see

(Sminketal.,2019) MP,out Totalcelluloseproduction 1000 kg/h Forconveniencea

SF Solventtofeedratio 0.5 kg/kg Estimatedfrom(Sminketal.,

2020a)

WFL Washingfactorlignin 2 kg/kgDES

WFP Washingfactorpulp 3 kg/kg Estimatedfromwaterwashing

datab

XDES,MTHF SolubilityofDESin2-MTHF 0.3 kg/kg Estimatedfrom(Sminketal.,

2020a)

xMTHF,DESrecycle Solubilityof2-MTHFinDES 0.1 kg/kg Estimatedfrom(Sminketal.,

2020a)

xW,B Waterfractioninbiomass 0.5 kg/kg (BownandLasserre,2015)

xW,DESrecycle WaterfractioninDESafter

liquid-liquidextraction

0.5 kg/kg Estimatedfrom(Sminketal., 2020b)

XW,P Waterfractioninpulp 0.67 kg/kg Typicallyachievablebypressing

(Eketal.,2009b) yL Ligninremovalfrombiomass 0.94 kgligninin

biomass/kglignin removed

Measuredinthiswork

yp Solidyieldafterdelignification 0.57 kg/kgbiomass Measuredinthiswork a _Normalized_to₁_t/h_for_convenience._All_numbers_can_be_scaled_linearly_to_any_desired_production_capacity.

b _Only_data_about_water_washings_are_available,_which_state_that_2.5_m3_water₍_Ek_et_al.,_2009b₎_per_t_of_pulp_is_required._Based_on_the_density_of

DES,beinghigherthanwater,3kgDESperkgpulpwasassumedtobetheminimumforDESwashingofpulp.

DESblackliquor,Hvap,DESwascalculatedusingtheClausius

ClapeyrontheorywithVLEdataobtainedbyFranciscoetal. (2013a),Hvap,DES=44.3kJ/mol. Othernecessaryinformation

wasmeasuredinthecasethatnoreasonableestimatecould begiven.Allrelevantequilibriumvaluesforthemassbalance arereportedinTable2(videinfra).

3. Results

and

discussion

3.1. Conceptualprocessdesignandinitialenergy requirementestimation

3.1.1. Input–outputstructure

First, the typeoflignocellulosic biomassmust beselected. Inprinciple,everytypeoflignocellulosicbiomass(hardwood, softwood,grasses,agriculturalwastes)canbeselectedforthe process.WeselectedE.globulusforfurtherstudiesbecausethis speciesisthemostcultivatedspeciesinfastgrowing planta-tions(Dillenetal.,2016)andbecauseeucalyptuswasalsoused inourpreviousstudies(Sminketal.,2019,2020a).However, the methodology applies to majority of the lignocellulosic biomasssources.

Next, the morphology of the feed must be selected. AlthoughtheDESimpregnationrateincreasesiftheparticles aresmallerandtherefore,thecookingwillbefasterandmore homogeneous,cuttingwoodinherentlycutsandthereby dam-agesthecellulosefibersinthe woodandthereforereduces thequalityoftheproducedcellulosefibers.Thus,smallerchip sizesincreasethepulpingrate,butcompromisesthequality oftheproducedcellulosefiber.Inthepaperindustry,where fiberqualityisveryimportant,chipsaregenerally20–30mm long.Forotherapplications, suchascellulosefermentation tobio-ethanol,thefiberqualityisirrelevant.Inthesecases, theusedchipsizeswillbemuchsmaller.Thechipsizeisvery importantforthedesignofthedigester,butnotforfurther

processingofthepulpandDESrecovery.Therefore,thisisno factorintherigorouscalculations.

Feedimpuritiesareimportantforany(chemical)process. Inthecaseoflignocellulosicbiomass,threemajorimpurities arepresent:water,extractablesandash.Presenceofwaterin thefeedisnotexpectedtoposeanyproblemsintheprocess, sincetheamountsofwaterpresentinthewoodare negligi-blecomparedtheamountsthatarerequiredforpulpwashing. AdditionofsomewatertoDESwilldecreasetheviscosityand isbeneficialintherecoveryprocess.However,littleisknown yeton theinfluenceofwateronthepulpingprocess.Also, highwatercontentsmayresultinsignificantevaporationat thepulpingtemperature,whichmayrequiretheuseof pres-surizedequipmentorventingduringpulping.Extractablesare low molar weightcompounds, suchas fatty acids, sterols, terpenoids and waxesthat can beextracted from biomass usingneutralsolvents.Consideringthe hydrophobicnature ofthesecompounds,theywillendupinthe2-MTHFduring theligninextraction.Extractablesmayberemovedby steam-ing biomasspriortofurtherprocessing.Ashpresentinthe feeddissolvesintheDESandwillaccumulateintheDES recy-cleloop.Therefore,itcouldbedesirabletoremoveashprior toDESprocessing,forinstancebyacidleaching(Oudenhoven etal.,2016).IncasesaltsdoendupintheDESrecyclethey willaccumulateandhavetoberemoved,forinstanceusing a bleedstream intherecycle. Purgingofthis bleedstream willbeveryexpensive,sothesaltswillhavetoberemovedby techniquessuchaselectrodialysisorprecipitationinanother solvent, such as ethanol. The ash content inthe eucalyp-tus usedforthis study wasonly0.38wt%. Incase1%ash isallowedintheDESrecyclestream,0.38kgDES hastobe bleadfromtherecycleperkgeucalyptusinthefeed(0.38kg DES*1%salt=3.8gsalt,whichisequaltotheamountin1kg eucalyptus). However,theashcontent canbemuchhigher in other biomasssources. For example,straw can havean ashcontentof6–8%(Oudenhovenetal.,2015).Forthe

(6)

rigor-Fig.1–Theinput-outputstructureoftheconceptual process.

ouscalculationsonlythepresenceofwaterinthefeedwas included.

Theoutletstreamsofthisprocessmustatleastconsistof astreamfortheproducedcelluloseandlignin.Besidesthese products,someofthehemicelluloses willalsobreakdown during DES processing. During treatment with acid DESs, hemicellulosebreaksdowntosugarfractions,whichfurther reacttofurans.Lietal.(2018)andHouetal.(2018b)studied thefractionation ofricestraw using acidbased DESs.Both authors(Houetal.,2018b;Lietal.,2018)couldonlyrecover lessthanhalfofthehemicelluloseremovedaseithersugaror furaniccompound,withtheremainderreferredtoas‘Loss’. Weexpecttheselossestobefuranicsandsugarsthatreacted furthertoformpolymericbyproductscalled‘humin’. Sepa-rationofsugarsfromDES iscomplicated.Bothcompounds arecompletelynon-volatile,excludingseparationby distilla-tion.Furthermore,bothcompoundsarehighlypolar,making affinitybasedseparationverychallenging.Theonlyoptions appeartobeadelicate membrane-basedprocess basedon thesmalldifferenceinsizeofthemoleculesandionsinthe system,another optionistoletthe sugarsreact furtherto furans.Thesefuransmaybeseparatedbyliquid–liquid extrac-tionusingorganicsolvents,suchas2-MTHF(Rosatellaetal., 2011;Duttaetal.,2012;Mikheenkoetal.,2019).Smallsugars producedinthecookingwillforamajorpartalreadybe con-vertedintofurans,andintheproposedprocessitseemslikely thatproducedfuranswillalsobeextractedby2-MTHFand thusleavetheprocesstogetherwiththelignin.Littleisknown about the humins, but withits furanicstructure, it seems possiblethattheseby-productsarealsoextractedby2-MTHF and thus leavethe process together with thelignin. Alter-natively,huminsmayprecipitateduringtheprocessonthe reactionequipment,asinthecommercialproductionof fur-furalfromhemicelluloserichagriculturalwastes(Hoydonckx etal.,2007).Intheusedlignin-in-DESmixtures,only3%of theinitialamountofxylosepresentinthebiomasswasfound backbyHPLC.Fortherigorouscalculations,weassumedthat allhemicellulosethatisremovedfromthebiomassstructure leavestoprocessas‘hemicellulosebyproduct’(HBP)together withthelignin.Thismeansthattheinput–outputstructure looksasdescribedinFig.1.

3.1.2. Batchversuscontinuous

In kraft pulping, continuous digesters have become stan-dard.Although somebatchdigesters are stillin operation, batchpulpinghasbecomeoutofdate.Inatypicalcontinuous digester,woodchipsarefedatthetopofthereactorvesseland slowlysinkthroughcookingliquorwithvariouscompositions. Comparedtothecurrentaqueouspulpingliquors,DESshave

amuchhigherdensityandviscosity(Franciscoetal.,2013c). TheincreaseddensityoftheDESincreasesthebuoyantforces exertedonawoodchipandtheincreasedviscosityincreases the dragforce onawood chip.Thiswilldecrease the flow rate ofchips throughacontinuousdigester filledwithDES and thusgreatlycomplicatesthedesignofsuchadigester. Asaconsequence,muchdatamustbegatheredtomakea detailed design. More than half ofall Europeanpulp mills hadacapacityexceeding200ktperyearin2018(CEPI,2018). Althoughsomeoldermillsstilloperatesemi-batchwise,this ishighlyunusualatthisscale.Therefore,pulpingmustbe exe-cutedcontinuously,justasotheroperations,suchassolvent recoveryandpulpwashing.

3.1.3. Theseparationsystem

Afterthe pulpingreactor, amixtureofcellulosepulp,DES, ligninandhemicelluloseby-productsmustbeseparated.In this mixture, the lignin and hemicellulose by-productsare dissolved intheDES. Itisconvenient toseparatethe solid pulpfromtheliquidDESfirst,forinstancebypressing. Dur-ingpressing,partoftheliquidremainsasstagnantliquidin the fibersin theporous pulpbeds.Therefore,the liquidin thepulpmustbedisplacedbyanotherliquid.Washingwith waterisveryconvenientsincewaterisinexpensive,non-toxic andnon-flammable.However,ifwaterisusedtodisplacethe DESinthecellulosefibers,thewaterwillmixwiththeDES, causinglignintoprecipitateonthefibers,whichisvery unde-sirable.Therefore,theDESinthecellulosefibersmustfirstbe displacedbycleanDES,beforeitcanbedisplacedbywater.

AftertheseparationbetweentheDESandpulp,theliquid mixturecontainingDES,water,ligninandhemicellulose by-productsmustbeseparated.Inourpreviouspaperwemade abriefcomparisonbetweentheregenerationofligninfrom DESusingcoldwaterprecipitation(Sminketal.,2020b)and concludedthatthe lowmolarweightligninfractionscould not berecovered bycold waterprecipitation.Furthermore, liquid–liquid extraction using 2-MTHFhas the potential to save95%energycomparedtocoldwaterprecipitation(Smink etal.,2020b).Therefore,ligninisrecoveredfromtheDESusing liquid–liquidextraction.Waterisremovedlaterfromthe mix-ture,sinceitaidsliquid–liquidextractionofligninfromthe DES.Weassumethatthehemicellulose-byproductsarealso extractedby2-MTHF(seeSection3.1.1.).The2-MTHFcanbe recoveredbyevaporationandtheligninandhemicellulose by-productsremainintheevaporationresidue,togetherwithany DESthatmayhaveleachedtothe2-MTHFduringextraction. ThisDEScanbewashedfromtheproductsusingwaterand thewashingwaterisrecycledbacktotheextractionstage.

Atlast,the DES mustbe separated from the waterand any 2-MTHF that leached to the DESphase during extrac-tion.Waterand2-MTHFmaybeseparated fromtheDESby evaporation.Aftercondensation,thewaterand2-MTHFcan beseparatedinasettlervesselbecauseoftheimmiscibility between waterand 2-MTHF. TheLLEbetween 2-MTHFand water was reported byGlass et al. (2016) and Stephenson (1992). At70◦C,the solubility of2-MTHFinwateris5wt% andthesolubilityofwaterin2-MTHFis6wt%.Remarkably, thesolubilityof2-MTHFinwaterdecreaseswithincreasing temperatures.Altogether,theflowsheetasshowninFig.2is obtained.

Massbalance

In the conceptual process, biomass is fractionatedinto pulp, lignin and hemi-cellulose by-products. Any other byproductsorlossesarenottakenintoaccountinthisprocess

(7)

Fig.2–Structureoftheconceptualprocess.Inthetopfigure,theprocessisshownbyfunctionandinthebottomfigure,the processisshownbyoperations.Afterpulping,thepulpinusedDESsuspension(3)ispressedandwashedtoformaclean pulpsuspension(6).AllUsedDESandwaterfromthewashingandpressingstagesarecombinedinstream(7),whichis mixedwithwaterfromtheligninwash(13)andtheligninandhemicellulosebyproductpresentinthismix(8)areextracted using2-MTHF(9and10).ThecleanDESrecycle(16)isdriedandthecleanDES(18)isusedagainforpulping(2)andwashing (4).

design.Therefore,theamountofbiomassenteringtheprocess (stream1,Fig.2)(MB,in)isequaltotheamountsofpulp(stream

6,Fig.2)(Mp,out),lignin(ML,out)andhemicellulosebyproducts

(MHBP,out)produced(stream15,Fig.2),asshowninEq.(5).

MB,in=MP,out+ML,out+MHBP,out (5)

Theamountofpulpproducedisthesumoftheamountof biomassenteringtheprocess,multipliedbythepulpyield(yP),

asshowninEq.(6).Thisyieldwasdeterminedexperimentally at57%.

MP,out=MB,in∗yP (6)

Theamount oflignin produced isequal tothe amount ofligninremovedfrom thebiomass.Thisisthe amountof biomassentering theprocess,multipliedbythe lignin frac-tioninthebiomass(xL,B)andthedegreeofdelignification(yL),

asshowninEq.(7).Thesefractionsweredetermined experi-mentallyatrespectively21.6%and94.1%.

ML,out=MB,in∗xL,B∗yL (7)

Forconvenience,thepulpproductionissetto1000kg/hand therewith,thesystemof3equationscanbeusedtosolvethe 3unknowns(MB,in,ML,outandMHBP,out).

TheamountofDESusedforpulpinginthedigester(stream 2,Fig.2)(MDES,pulping)isequaltotheamountofbiomass,

mul-tipliedbytheDEStobiomassratio(DB),asshowninEq.(8).

Priortooptimizingtheflowsheet,thisratioissetto10since thisisacommonratioinliterature(Alvarez-Vascoetal.,2016).

MDES,pulping=MB,in∗DB (8)

Afterpulping,thepulpisfirstwashedusingDEStoremove ligninfromthepulp.Tothebestofourknowledge,nodatais availableonpulpwashingusingDESyet.Therefore,dataused onwaterwashingisusedtogiveanestimateontheamount ofDES requiredforwashing,but morework isrequiredto provewhetherthisassumptionisjustified.Itiscommonto usearound3kgwaterperkgpulpinindustry(Eketal.,2009a). Thiswashingfactor(WF)ismultipliedbytheamountofpulp producedtocalculatetheamountofDESrequiredforwashing (MDES,wash),asshowninEq.(9).

MDES,wash=MP,out∗WFP (9)

Waterisrequiredtoaidliquid–liquidextractionoflignin fromtheDES.Therefore,thewaterfractionintheDES recy-clestream(stream16,Fig.2)(xW,DESrecycle)isdefinedasthe

amountofwaterintherecycle(stream16,Fig.2)(MW,DESrecycle)

overthesumoftheamountsofwaterandDES(frompulping andwashing),asshowninEq.(10).Awatermassfractionof 0.5isassumed,sincethisissufficientforfullligninrecovery usingliquid–liquidextraction(Sminketal.,2020b).

xW,DESrecycle=

MW,DESrecycle

MDES,Pulping+MDES,washing+MW,DESrecycle

(10)

Theamount ofsolventrequiredfortheextractionstage (stream9+10,Fig.2)(MMTHF,extraction)isequaltothetotalmass

(8)

oftheDESandwaterintherecycle(stream16,Fig.2), multi-pliedbythe solventtofeedratio (SF),asshowninEq.(11). Aratioof0.5wasused,sincethisissufficientforfulllignin recovery(Sminketal.,2020b).

xW,DESrecycle=

MW,DESrecycle

MDES,Pulping+MDES,washing+MW,DESrecycle

(11)

SincetheextractionsolventandDESarepartiallymiscible, someoftheextractionsolventleachestotheDESrecycle(from stream9tostream16,Fig.2).Thisamount(MMTHF,leached)is

determinedbythesolubilityfactor(xMTHF,leached)of2-MTHFin

DES,asshowninEq.(12).Thisfactorwasestimatedat0.1from equilibriumdata(Sminketal.,2020a).

MMTHF,extraction

=

MDES,pulping+MDES,washing+MW,DESrecycle

∗SF (12)

BesidesleachingofsolventtotheDESrecycle,someDES willalsoleachtothesolvent(fromstream 8tostream11). ItmustbenotedthatDESsarenotpseudo-purecompounds, andespeciallyduringliquid–liquidextraction,their compo-sition changes sinceevery DES constituent has adifferent leachingcharacteristics.Inthiscase,lacticacidhasamuch highersolubilityin2-MTHFthancholinechloride(Sminketal., 2020a)andthus,relativelymorelacticacidthancholine chlo-ridewillleachtothe2-MTHFphase(fromstream8tostream 11,Fig.2).However,afterextractionanyleachedDESwillend up instream 12 ofthe process,where it iswashed ofthe ligninandhemicellulosebyproductandfurthermorerecycled directlybacktotheextractionstageviastreams13and8(see Fig.2).Therefore,itisnotexpectedthatthecompositionof theleachedDESwillhaveanyinfluenceontheenergyusage oftheprocessandisthereforenottakenintoconsideration. Thisamount(MDES,MTHFleach)isdeterminedfromthe

solubil-ityfactorofDESin2-MTHF(xDES,MTHF),asshowninEq.(13).

Thisfactorwasestimatedat0.3fromequilibriumdata(Smink etal.,2020a).

MDES,MTHFleach=MMTHF,extraction∗xDES,MTHF (13)

Atlast,thewaterbalanceismade.Somewateris natu-rallypresentinthiswood.Sinceitwasdecidednottopre-dry thewood,thisamounthastobetakenintoaccountaswell. Thefractionofwaternaturallypresentinbiomass(xW,B)is0.5

(Bownand Lasserre,2015).Thisisdefinedasde amountof waterinthebiomass(MW,B),dividedbytheamountsofwater

andbiomassenteringtheprocess,asshowninEq.(14).

xW,B=

MW,B

MB,in+MW,B

(14)

Sincepulpisaporousmaterial,somewaterwillalways remaininthepulpafterwashing(stream6,Fig.2).Ifthe pro-cessisintegratedwithforinstanceapapermakingplant,this amountdoesnotneedtoberemovedsincecellulosepulpmust befedtoapapermakingmachineinanaqueoussuspension. Alternatively,thewatercanberemovedifthepulpissoldon themarket.Thewaterfractioninthepulp(xW,P)isdefinedby

theamountofwaterinthepulp(MW,P),dividedbytheamount

ofpulpandwaterinthepulp,asshowninEq.(15).Thewater

fractioninpulpwassetat0.67,sincethisamountcantypically beachievedbypressingofthepulp(Eketal.,2009b).

xW,P=

MW,P

MP,out+MW,P (15)

Afterevaporationof2-MTHFfromtheligninand hemicel-lulosebyproduct,anyDESthatremainswiththeligninand hemicellulosebyproduct(stream12,Fig.2)mustberemoved fromtheligninbywashingwithwater.Theamountofwater required for lignin and hemicellulose byproduct washing (stream14,Fig.2)(MW,Lwash)isdeterminedbytheamountof

DESthatmustbewashedfromtheligninandhemicellulose byproduct,multipliedbyawashinfactor(WFL),asshownin

Eq.(16).Thewashingfactorisestimatedat2,itisassumed that2kgwaterisrequiredtowash1kgDESfromthelignin andhemicellulosebyproduct.

MW,Lwash=MDES,MTHF∗WFL (16)

Inaddition,theamountofwaterrequiredforthewashing ofpulpmustbedetermined.Industrially,3kgwaterisused towash1kgpulp.Afterthepulpwashingstep,theDESand watermixtureissenttoaliquid–liquidextractionoperation toremovelignin.Inthis step,waterisalsorequiredforan efficientligninextraction.Itisexpectedthattheamountof waterpresentinthismixtureisnotsufficientforaneffective ligninextraction.Therefore,extrawatermustbeaddedtothis mixture.Itispossibletodothisattheextractionstage,butitis alsopossibletousemorewaterduringthewashingstep.This amountwillthen(viastream7and8,Fig.2)alsoendupinthe extractionstage.Sinceusingextrawaterinthewashingstage willeasethewashingoperation,thisispreferredoverthefirst option.Therefore,theamountofwaterusedforpulpwashing (MW,Pwash,stream5)iscalculatedfromamassbalanceonthe

waterrequiredintheDESrecycleforefficientligninextraction (MW,DESrecycle,stream16).Theamountsofwaterthatalready

endedupinthisstreamfromthewaternaturallypresentin thebiomass(MW,B,stream1)andtheamountthatisadded

fromtheligninwash(MW,Lwash,stream14)arereducedfrom

thisamountandtheamountofwaterthatwaterthatleaves theprocessinthecellulosepulp(MW,P,stream6)mustbemade

upfor.ThefinalequationisshowninEq.(17).

MW,Pwash=MW,DESrecycle−MW,B−MW,Lwash+MW,P (17)

Theamountofwaterusedfrompulpwashingmustalways behigherorequaltotheminimumamountofwaterrequired forpulpwashing.Thisamountis3kgwaterperkgpulpand calledthewashfactor(WFP),showninEq.(18).Incasethe

amountofwatercalculatedbyEq.(17)islowerthanthelower boundary stated in Eq.(18), the amount ofwaterrequired forwashingwillbedeterminedaccordingtoEq.(18)andthe waterfractionintheDESrecycle(calculatedfromEq.(10))will increase.

MW,Pwash≥MP,out∗WFP (18)

UsingEqs.(5)–(18)andthementionedassumptions,which aresummarizedinTable2,themassbalanceoftheprocess canbecalculated,whichisshowninTable3.

(9)

Table3–Massbalanceoftheconceptualprocess.

Materialsinprocessstreams(kg/h) Processstream

1 2 3 4 5 6 7 8 9 10 Wood 1754 DES 17,544 17,544 3000 20,544 25,979 Water 1754 1754 5065 2000 8819 19,689 2-MTHF 18,116 3623 Pulp 1000 1000 Lignin 356 356 356 HBP 398 398 398

Materialsinprocessstreams(kg/h) Processstream

11 12 13 14 15 16 17 18 19 20 Wood DES 5435 5435 5435 20,544 20,544 Water 10,870 10,870 15,689 15,689 194 20,350 2-MTHF 18,116 3623 3623 3038 1071 Pulp Lignin 356 356 356 HBP 398 398 398 3.2. Energybalance

Duringpulping,themixtureofDESandwoodiskeptat130◦C. Afterpulping,thepulpisfirstwashedusingcleanDESfrom theevaporationplant.ThecleanDESisnotheated,norcooled beforethewashingstage.AfterthecleanDESwash,thepulp iswashedusingfreshwater.Theuseofhotwaterisbeneficial becauseitincreasesdiffusionratesanddecreasesthe viscos-ityoftheliquidmixtures.Franciscoetal.(2013a).determined theviscosityofasimilarDES, comprisedoflactic acidand cholinechlorideina2:1Mratioat13temperaturesfor11 mix-turesofthisDESwithwater.Itisassumedthatwaterenters thewashingplantat80◦C.Atthistemperature,theviscosityof thecomparableDESusedbyFranciscoetal.(2013a)is25mPas andisreducedtoapproximately 1mPasina50wt%DESin watersolution(∼0.14mol/mol).Furthermore,nopressurized equipmentisrequiredatthistemperatureanditcaneasilybe achievedusingwasteheatoftheprocess.Eitherwaterfrom theevaporationplantmay beuseddirectly inthe washing stage,or the heat in this stream may beused topre-heat waterforthewashingstage.Itisassumedthatthe suspen-sionofcellulosefibersinwaterexitsthewashingplantatthe inlettemperatureofthewashingwater(80◦C)andthatthe restoftheliquidsfromthewashingplantaremixed.Before liquid–liquidextraction,thisstream(stream7,Fig.2)iscooled downto75◦C.Thisisslightly lower thanthe boilingpoint ofpure2-MTHF(whichis80◦C),andbasedontheVLEdata reportedbyGlassetal.(2016),weestimatethatnopressurized equipmentisrequiredforliquid–liquidextraction.

Afterliquid–liquidextraction,waterisremovedfromthe DESrecycle(stream16,Fig.2.)bymulti-effectflash evapora-tion.Itisassumedthatthesemulti-effectflashevaporators havethesamecharacteristicsasthemulti-effectflash evap-oratorscurrentlyusedinthekraftprocess.Furthermore,itis assumedthattheDESiscompletelynon-volatile,asshown byFranciscoetal.(2013a).Typically,theseevaporatorsusea 7effectevaporationtrainandhaveasteamefficiencyof5.95 (Valmet,2015)(1tonofsteamisusedtoevaporate5.95tonof water).However,theheatrequiredtoevaporatewaterfrom kraftblackliquor(Hvap,BL)isdifferentthantheheatrequired

to evaporatewater from the DES (Hvap,DES). Hvap,BL was

determinedbyFerreiraetal.(2011)andis54.1kJ/mol.Hvap,DES

wascalculatedusingtheClausiusClapeyrontheorywithVLE dataobtainedbyFranciscoetal.(2013a)andequals44.3kJ/mol. Ifitisassumedthattheheatcontentofsteamis2.2GJ/t[75], theheatdutyofthewaterevaporationstepcanbecalculated usingEq.(19). Qwaterevaporation=2.2∗ MW,DESrecycle steamefficency∗ Hvap,DES Hvap,BL (19)

Inmultieffectflashevaporation,thefirststageorstages areheatedusingdirectsteam.Thesteamthatevaporatesfrom thisliquidisusedtoheatthenextstage,andsoforth.TheDES entersatthelasteffectandisthusheatedfurtherwitheach consecutivestage.Itisassumedthatthetotalheatincrease isequaltotheheatincreaseinakraftevaporator,whichis typically 40◦C [76]. Thismeans that the dry DES exits the evaporationtrainat75+40=115◦C.

Afterliquid–liquidextraction,the2-MTHFisalsorecovered using multi-effectflashevaporation.Itisassumedthat the setupofthemultistage flashevaporationissimilar tothe setupofthewaterevaporation.However,theheatof evap-oration of2-MTHFisonly 0.40MJ/kg, whichismuch lower thantheheatofevaporationofwater,whichis2.2MJ/kg.This meansthatinthefirststage,1kgsteamisrequiredto evapo-rate2.2/0.4=5.5kg2-MTHFandthustheheatdutyisreduced bythisfactor.The2-MTHFvaporfromthefirststageisthan usedtoheatthenextstage,andsoforth.Itisassumedthat thesameevaporationefficiencyisachievedasforthe evap-orationofwaterfromDES,thustheheatdutyisreducedby thesameefficiencyfactorof5.95.Thetotalheattoevaporate the 2-MTHFafterliquid–liquid extractioncanbecalculated accordingtoEq.(20). QMTHFevaporation=2.2∗ 0.40 2.2 ∗ MMTHF,extraction Evaporationefficiency (20)

Aftertheevaporationtrain,theDESiscleanandreadytobe usedagaininthepulpingstep.TheDESandwoodchipsare addedtoadigester.For theenergycalculationsweassume thatthewoodchipsenterthedigesteratroomtemperature. Both the DES and the chips(withwater inthem)must be heatedto130◦Cinthedigester.Thisheatcanbecalculated usingtheheatcapacitiesofwood,waterandDES,according

(10)

Table4–Energyusageintheconceptualprocessbefore andafteroptimization.

Operation Heatbefore optimization(GJ/t) Heatafter optimization(GJ/t) Digester 1.7 1.3 Waterevaporation 6.2 2.7 2-MTHFevaporation1.4 0.7 Pulpdrying 3.0 3.0 Total 12.3 7.7 toEq.(21).

QDigester= (CP,wood∗MB,in+CP,water∗MW,B)∗

Tdigester−Troom

+CP,DES∗MDES,pulping∗

Tdigester−Tevaporation,end

(21)

LittleisknownabouttheheatsofreactioninDES delig-nification.Tothebestofourknowledge,onlyonestudywas performedon the heats ofkraft delignification(Courchene etal.,2005).However,theseresultscannotbeusedtomake anestimation forDES delignification since the delignifica-tionchemistryiscompletelydifferent.Courcheneetal.(2005) found that the heat released during kraft pulping indicat-ing an exothermic process,is due to the neutralization of acidiccarbohydratebreakdownproductswithNaOH,a reac-tionthatdoesnotoccurinDESdelignification.Thequestion whether biomassconversion processesare endothermicor exothermicalsoplaysinotherfields,suchaspyrolysis.Inthe pyrolysisprocess,manyendothermicandexothermicreaction zoneswereobserved.Overall,itisbelievedthatpyrolysisof woodisslightlyendothermicwithaheatofreactionaround 0.6MJ/kg(Sinhaetal.,2000),althoughitmustalsobenoted thatligninbreakdown reactions aremoreexothermic than cellulosebreakdown reaction(Roberts,1971)(whichremain mostlyinertduringpulping).Furthermore,thefractionofthe polymersthatreactsinthepulpingismuchsmallerthanin pyrolysis.Taken together,itappearsjustifiedtoneglectthe heatthatisproducedorrequiredduringDESpulping,but fur-therresearchisnecessarytofindoutwhetherthisassumption isrealistic.

Atlast,thepulpmaybedried.Formanycellulose applica-tions,suchaspapermakingorenzymatichydrolysisforthe productionof other chemicals, the celluloseis used in an aqueousslurry.Whenthisplantisintegratedwiththe pulp-ingplant,dryingofthepulpisunnecessary.Inthisstudythe heatofdryingistakenintoaccount,sincewestudya non-integratedpulpmill.Itisassumedthattheheatrequiredfor pulpdryingisthesameasinacommercialpulpmill,which is3GJ/t(Suhretal.,2015).Inshort,heatisrequiredforthe digesterandtoevaporatewaterand2-MTHFandcoolingofthe DESisrequiredbeforeliquid–liquidextraction.Anoverview oftherequiredheatfortheprocessisshowninTable4. Fur-thermore,theproducedpulpmaybedried.Unfortunately,the heatsformulti-effectflashevaporationandpulpingmustbe providedathightemperatures(>120◦C)andthus,thereare littleopportunitiesforheatintegration.

3.3. Processimprovement

Theamountofheatrequiredforanon-integratedDESbased pulping process was estimated at 12.3GJ/t using rigorous calculations.Toputthis numberinto perspective;atypical

Fig.3–DelignificationofeucalyptusasfunctionoftheDB ratio.Thethreeexperimentaldelignificationsofeucalyptus wereobtainedbytreatingthewood1hat130◦C.The regioninwhichtheDBratioistoolowtofilltheporous biomassbedwithDESisindicatedinredblocksandthe regionbelowthefiberliberationpointisindicatedwith orangestripes.(Forinterpretationofthereferencesto colourinthisfigurelegend,thereaderisreferredtothe webversionofthisarticle).

kraftprocess uses11.0GJ/t inheat. InordertomakeaDES based pulpingprocess competitive tothe kraft,the energy usageshould belower than inthekraftprocess. Consider-ing thatmanydecadesofoptimizationhavetakenplacein the development ofthe kraftprocess, alsooptimizationof the DES-based process should be done, and startingfrom thevaluestakenfromlaboratoryobservations,afirstsetof improvementsisconsideredbasedonacombinationof exper-imentsandcalculations.

3.3.1. DEStobiomassratio

Themostobviousparametertochangeintheprocessisthe DBratio.UsinglessDESwillreducethedutyinthedigester andrequirelesswaterand2-MTHFforextraction.However, a decreasedDBratio willincreasethe ligninconcentration intheDES,whichmaydecreasethedelignificationrate. Fur-thermore,thereisaminimumDBratiobecausebiomassitself isaporousstructureandbiomassinadigesterhasa pack-ingdensity.Thistypicallyresultsinapackingdensityaround 230kg/m3_for_hardwoods₍_Ek_et_al.,_2009b_)._Assuming_the

den-sityofwoodis1500kg/m3₍_Ek_et_al.,_2009b₎_and_the_density

oftheDESappliedinthisstudyis1140kg/m3₍_Francisco_et_al.,

2013c),atleast4.2kgDESisrequiredtofilltheemptyspacesin thewoodpacking.Forconvenience,theDBratiowasreduced from10:1to5:1kg/kg.Itwasfoundthatdelignificationdegree decreasedfrom94.1%to87.0%andtheyieldincreasedfrom 57%to59%.Althoughthedelignificationdegreedecreased,it isstillwellabovethefiberliberationpoint,whichisat80% delignification(Brännvall,2017).Thisisschematicallyshown inFig.3(Brännvall,2017).Thetotalenergyuseintheprocess decreasedto8.6GJ/tpulp.

3.3.2. Wateraddition

Reducing the amount of water used in the liquid–liquid extraction ofligninfrom the DES willdecrease the energy requirementinthewaterevaporator.Thewatercontentinthe DESmustbebetween25and50wt%forfullligninrecovery

(11)

Fig.4–SinglestageligninrecoveryfromDESusing2-MTHF forvariousamountsofwaterintheDESraffinateat equilibrium.Thewatercontentsintheraffinatethatare obtainedintheproposedprocessbymixingtheminimum amountsofwater(seeEq.(18))areindicatedbydashed linesfortwoDBratios.Thesearetheminimumamountsof waterthatcanbereachedfromaprocesspointofview. (Sminketal.,2020a).Theligninrecoveriesinoneextraction stageweredeterminedfor6watercontentsintheraffinates between20 and 55wt%. Itwasfoundthat the singlestage recoverydidnotimprove muchwhen thewatercontent in theraffinateincreasedmorethan40%.

Waterhastwofunctionsintheprocess,itisrequiredfor fullextractionoftheligninfromtheDES,anditisrequired forthewashingofthecellulosefibersandlignin.Thewater thatisrequiredforwashingthecelluloseandligninis recy-cledtotheextractionstage.Whentheminimumamountsof washingwateraremixedwiththeDES,thewatercontentsin theraffinatearebetween38%and43%foraBDof1:10and1:5 respectively.ThesevaluesareshowninFig.4asthedashed anddash-dotlines.Itcanbeseenfromthefigurethatthe sin-glestagerecoverydoesnotincreasemuchfurtherwhenmore waterisaddedthanthewaterthatisalreadypresentfrom therecyclingofthewaterwashings.Therefore,theamount ofwaterrequiredforthe processislimited bytheamount requiredforthewashingstages,ratherthanthelignin recov-ery.AtaDBratioof5:1and43%waterintheraffinate,thetotal energyusageoftheprocessdropsto7.7GJ/tcellulosepulp.

3.3.3. RecyclingofwashingDES

TheDESthatiscurrentlyusedforwashingissentdirectlyto theliquid–liquidextractionstage,whereligninand hemicel-lulosebyproductareremovedfromit.Thelignincontentsin theeluentstreamsarehoweverlowerthanthelignincontents intheDESthatexitsthedigester.Thus,theligninthatexits thewashingstreammaybeconsideredasnotfullysaturated. Therefore,itmaybeconsideredtosendtheeluentfromDES washingwithalowlignincontentbackdirectlytothe pulp-ingstage.IfthisDESisrecycledbackdirectlytothedigester, the total energyusage can befurther reduced from 7.7 to 6.8GJ/t.However,whenmoreligninispresentinthedigester, thismaycauseligninandhemicellulosepolymerization reac-tions,whichmayinhibitthedelignificationrateandincrease theligninmolarweight.Toinvestigateontheseverenessof therepolymerization, delignificationexperimentswere car-riedoutwithpreviouslypreparedlignin-in-DESmixtures.

FromFig.5itcanbeseenthatthedelignificationdecreases withthe initiallignincontent.However, thedelignification

Fig.5–Delignificationofeucalyptusasfunctionofthe initiallignincontentoftheDES.Eucalyptuswastreated1h at130◦C.

staysabove80%,evenwhentheinitiallignincontentexceeded 2.5wt%. The lignin molar weightdistribution in these liq-uids was also investigated by GPC. First, the lignin molar weightdistributionsinthepreviouslypreparedlignin-in-DES mixture,andintheDESthatwasobtainedafterthe delignifi-cationexperimentusingcleanDESweredetermined.When the lignin-in-DES mixture is used for a consecutive delig-nification experiment, the lignin that is added to the DES duringthedelignificationexperimentisaddedtothelignin thatwasalreadypresent inthemixture.Ifweassumethat theligninthatwasinitiallypresentinthelignin-in-DES mix-tureisinertduringthedelignificationexperiment,itwouldbe expectedthattheligninobtainedbythedelignification exper-imentusingthelignin-in-DESmixtureisthesumofthelignin inthe initiallignin-in-DES mixtureandtheligninobtained usingthecleanDES.Wecanthenexpectthemolarweightof theligninobtainedfromthedelignificationexperimentusing thelignin-in-DESmixturetobetheweightedaverageofthe lignininthe lignin-in-DESmixtureand theligninobtained by delignification using a clean DES. This expected molar weightdistributionwascomparedtothedistributionthatwas actuallyobtained inthe delignificationexperiment usinga lignin-in-DES mixture, as shownin Fig. 6. In this figure, it canclearlybeseenthattheobtainedmolarweightismuch higherthantheexpectedmolarweightandthus,additional condensationreactionsmusthavetakenplacewiththelignin thatwasinitiallypresentinthemixture.Thiswillnotonly makeithardertovalorizetheobtainedlignin,butitwillalso makeligninrecoverybyliquid–liquidextractionmuchharder. Therefore,recyclingofwashingDESwithoutligninrecoveryis notconsideredfortheconceptualdesign.

3.4. Solventlosses

Lactic acid can form esters withhydroxyl groups in other molecules,justlikeanyothercarboxylicacid.Thismeansthat lacticacidmayalsoformesterswithhydroxygroupsinlignin andcellulose.Forexample,Kajimotoetal.(2008)produced cel-lulosefibersfromSugipineusinglacticacidandfoundthatthe producedfiberscontained6–9%lacticacid,whichwas poly-merizedtothecellulosefibers.Thelacticacidfractioninthe ligninandcelluloseproducedbyDESwasdeterminedbyacid hydrolysis.Itwasfoundthatthecellulosecontained3.6wt% lacticacid,whilethelignincontained6.4wt%lacticacid.By massbalancethismeansalossof36kglacticacidinthepulp

(12)

Fig.6–Obtainedandexpectedmolarweightdistributions foradelignificationexperimentusingalignin-in-DES mixture.TheobtainedgraphistheGPCspectrumfromthe trialwith2.6wt%initiallignin.Theexpectedmolarweight graphistheweightedaverageoftheGPCspectrumofthe lignin-in-DESmixtureandtheGPCspectrumobtainedfrom thedelignificationexperimentusingcleanDES.

and20kgintheligninpertcelluloseproduced,addinguptoa totallossof56kglacticacidpertcellulosepulp.Possibly,lactic acidcanberemovedfromcellulosefibersbyozonebleaching (NeumannandBalser,1993).

Moreresearchisrequiredtodeterminethelossesby sol-ventdegradationduringDESprocesses.Rodriguezetal.(2019) researcheddegradation byesterification betweenthe lactic acidandcholinechloride.However,theseesterification reac-tionsareequilibriumreactions.Thismeansthattheformed degradationproduct(theesterbetweenlacticacidandcholine chloride)reactsbacktotheinitialDES,andespeciallyby oper-atingthepulpingstagewet,thepresenceofwaterdramatically reducestheesterificationoflacticacid.Hazetal.(2016) deter-minedthataDEScomprisedoflacticacidandcholinechloride wasstableupto197◦C,meaningtheusedDESisthermally stableattheoperatingconditionsusedinthiswork.

4. Outlook

Theenergyuseinthisprocessismainlyduetotheevaporation ofwaterusedforwashingthecellulosefibers.Thismeansthat thereisrelativelylittleroomforanimprovementinenergy usebythe optimizationofpulpingandregeneration condi-tions.Nevertheless,theheatdutyoftheproposedDESprocess is28%lowerthanthekraftprocess.Washingwithother sol-vents,suchasethanolinsteadofwatermayfurtherdecrease theenergyusageoftheprocess sincetheheat of evapora-tionofethanolismuchlowerthan water.Anotherpossible reductionoftheheatdutycanbeobtainedbyreducingthe evaporationofwater,by(atleastpartially)dewatertheDES usingamembrane.Furtherresearchonthissubjectis nec-essarytoaccuratelyestimatethetechnicalfeasibilityofsuch processandtoestimatethepossiblesavingsinenergyuse.

PulpingwithDESsofferspotentialadvantagesoverthe cur-rentkraftprocessfromtheviewpointofthe12principlesof greenchemistry(AnastasandEghbali,2010).First,the poten-tialtoproducevaluablebyproductsfromthefeedstockismuch higher,meaningmoreproductscanbemadefromthesame feedstock,whichincreasestheatomeconomy.Reasonforthe

higherpotentialonvaluablebyproductsisthatnonucleophlic substitutionusing bisulfidehastobeusedtodelignifiy the biomass(DaCostaLopesetal.,2020).Thismeansthatthe pro-cessislessdestructivetothelignin,whichmaybeobtained withhighervalue.Second,theenergyefficiencyofthe pro-cessispotentiallyimprovedsignificantlyasfollowsfromour energyanalysis.Third,becausenopressurizedequipmentis requiredtooperatetheprocess,thecookingissafer.However, therearealsosomedrawbacksrelatedtotheuseof2-MTHF, whichisusedintherecovery.Although2-MTHFabio-based solvent, it isveryvolatile and flammable.Althoughits use seemsjustifiedbythesignificantenergysavings,itis desir-abletofindnon-flammablealternativesto2-MTHF.Allinall, weconsiderthebenefitsoftheproposedprocess,intermsof the12principlesofgreenchemistry,tobemuchbiggerthan thedrawbackscomparedtothekraftprocess.

Itseemsthatalloperationscanbeperformedusingreadily availabletechnology.Furthermore,nopressurizedequipment isrequiredsincetheproposedDESisnon-volatile.Thismeans thattheprocessmayhaveaverylowcapitalexpenditure. Fur-thermore,theenergyusageoftheproposedprocessisvery low,whichisnotonlyfavourablefortheoperationalcosts,but sincetheinvestmentcostsofchemicalprocessesaretypically dictatedbyenergylosses(Lange,2001),thisisalsofavourable fortheinvestmentcostsoftheproposedprocess.However,the biggestopportunityforDESbasedproductsisthevalorization ofby-products.Nexttotheproducedcellulose,ligninsmaybe producedalwell.Furthermore,theprocessmaybeadjusted toproducevaluablefuranicproductsfromthehemi-cellulose, suchas(hydroxymethyl)furfural(Huetal.,2008;Zhangetal., 2014).Thesechemicalmaybeusedasfeedstockforchemical industryasareplacementoffossilresources,asschematically showninFig.7.Allinall,theproppedDESbaseddelignification processshowsagreateconomicpotential.

In the proposed process,DES losses are significant due topolymerizationoflacticacidtoligninandcellulose.This means that the process either has high solvent losses, or expensivepost-treatmentstorecoverDES.Furthermore,some 2-MTHFleaches tothe DESduringliquid–liquid extraction. This is evaporated from the DES together with the water andcanbeseparatedusingasimplesettlersince2-MTHFis immisciblewithwater.However,2-MTHFandwaterdohave amutualsolubility (Glasset al., 2016). Asolubility calcula-tionusingthedatafromGlassetal.(2016)resultsinalossof 456kg2-MTHFpertcellulose.Thisamountmustalsobe recov-eredfromthewater,forinstancebyextractionusingeither ahighboilingsolvent,suchasdodecane,orsolvent impreg-natedresins.Inthiscase,the2-MTHFcanberecoveredfrom theextractantbyflashevaporation,meaningheatof evapo-rationof2-MTHFmustbeaddedtothetotalheatduty.This amountis0.456*0.4=0.18GJ/tcellulose.Alternatively, mem-braneseparationsmaybeconsideredforthisseparation,but moreworkisrequiredtofindoutwhetheramembrane sep-arationistechnicallyfeasible.Membraneseparationsdonot requireadditionalheat,butwillrequiresomeadditional elec-tric power (Kujawski,2000). The kraft process is currently favoredoverotherbiomassfractionationprocessessincethe cellulosefibersproducedbythisprocesshavegoodproperties forpapermaking.IthasnotyetbeenproventhatDESbased delignificationprocessescanproduceasimilarfiberqualityto thekraftprocess,whichmaylimitpossibleindustrial appli-cation.ConsideringthemanypossibleDESconstituentsand ratios,whichwillallaffectthefiberquality,wethinkthatit mustbepossibletofindaDESthatcanproducecellulosefibers

(13)

Fig.7–ConceptualcomparisonofthefeedstocksandproductsinthecurrentindustrialsituationandthedesiredDES process.PicturewasadaptedfromtheProviDESfinalreport(DeepEutecticSolventsinthepaperindustry,Institutefor SustainableProcessTechnology,2018).

withtherightpropertiesforpapermaking,butmoreresearch isrequiredtoproofthis.

5. Conclusions

Aconceptualprocessdesignwasmadeforthedelignification oflignocellulosicbiomassusingadeepeutecticsolvent com-prisedoflactic acidand cholinechloride.Massand energy balancesweremadeusingrigourouscalulations.Optimizing theamountsofwaterandDESusedintheprocessreduced theestimatedheatdutyfrom12.3to7.9GJ/tcellulose,which is28%lowerthan thekraftprocess.TheamountofDESin theprocess could bereduced tothe amountthat is physi-callyrequiredtofilltheporousbiomassbed.Theamountof waterrequiredforsufficientligninrecoverybyliquid–liquid extractionisalreadyachievedwhenthewaterusedfor wash-ingofthecelluloseandligninarerecycledtotheliquid–liquid extractionstage.Recyclingoflignin-in-DESmixturesdirectly backtothedelignificationstagemaysaveaminoramountof energy,butwillincreasethemolarweightofthelignin,which decreasesits potentialforvalorizationand makesrecovery byliquid–liquidextractionmoredifficult.Valorizationof by-products,suchasligninandfuranicsfromhemi-celluloseis keyinDESbaseddelignificationprocesses.

Declaration

of

interests

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

Acknowledgements

Theauthors would like tothankthe members ofthe ISPT “DeepEutecticSolventsinthepulpandpaperindustry” con-sortium for their financial and in kind contribution. This cluster consists of the following organisations: CTP, CTS Twente,ISPT,Mayr-MelnhofKarton,MidSwedenUniversity, Mondi, Sappi, Stora Enso, University of Aveiro, University ofTwente,ValmetTechnologiesOy,VTTTechnicalResearch CentreofFinlandLtd,WEPAandZellstoffPöls.Thisproject receivedfundingfromTKIE&Iwiththesupplementarygrant ’TKI-Toeslag’forTopconsortiaforKnowledgeandInnovation (TKI’s)ofthe<GS1>MinistryofEconomicAffairsandClimate Policy.

Eucalyptuschipswere kindlyprovidedbyTheNavigator Company.

Appendix

A. Supplementary

data

Supplementary material related to this article can be found,inthe onlineversion,atdoi:https://doi.org/10.1016/j. cherd.2020.09.018.

References

Alvarez-Vasco,C.,Ma,R.,Quintero,M.,Guo,M.,Geleynse,S., Ramasamy,K.K.,Wolcott,M.,Zhang,X.,2016.Unique low-molecular-weightligninwithhighpurityextractedfrom woodbydeepeutecticsolvents(DES):asourceofligninfor valorization.GreenChem.18,5133–5141,

http://dx.doi.org/10.1039/c6gc01007e.

Anastas,P.,Eghbali,N.,2010.Greenchemistry:principlesand practice.Chem.Soc.Rev.39,301–312,

http://dx.doi.org/10.1039/B918763B.

Bown,H.E.,Lasserre,J.-P.,2015.Anair-dryingmodelforpiledlogs ofEucalyptusglobulusandEucalyptusnitensinChile.New Zeal.J.For.Sci.45,1–9,

http://dx.doi.org/10.1186/s40490-015-0047-6.

Brännvall,E.,2017.Thelimitsofdelignificationinkraftcooking. BioResources12,2081–2107,

http://dx.doi.org/10.15376/biores.12.1.2081-2107. CEPI,2018.KeyStatistics,EuropeanPulpandPaperIndustry.

ConfederationofEuropeanPaperIndustries,Brussels. Chang,J.,Liu,J.,Guo,S.J.,Wang,X.F.Y.,2016.Investigationinto

selectiveseparationoflignininnoveldeepeutecticsolvent.J. SouthChinaUniv.Technol.44,14–20,

http://dx.doi.org/10.3969/j.issn.1000-565X.2016.06.003. Chen,H.,2015.3—Lignocellulosebiorefineryfeedstock

engineering.In:Chen,H.(Ed.),LignocelluloseBiorefinery Engineering,PrinciplesandApplications.Woodhead Publishing,Sawston,Cambridge,pp.37–86,

http://dx.doi.org/10.1016/B978-0-08-100135-6.00003-X. Courchene,C.,McDonough,T.,Hart,P.,Malcom,E.,Carter,B.,

2005.Determiningtheheatofreactionofkraftpulping.Tappi J4,9–13.

DaCostaLopes,A.M.,Gomes,J.R.B.,Coutinho,J.A.P.,Silvestre, A.J.D.,2020.Novelinsightsintobiomassdelignificationwith acidicdeepeutecticsolvents:amechanisticstudyof␤-O-4 etherbondcleavageandtheroleofthehalidecounterionin thecatalyticperformance.GreenChem.22,2474–2487, http://dx.doi.org/10.1039/C9GC02569C.

Dillen,J.R.,Dillén,S.,Hamza,M.F.,2016.Pulpandpaper:wood sources.In:ReferenceModuleinMaterialsScienceand MaterialsEngineering.Elsevier,

http://dx.doi.org/10.1016/B978-0-12-803581-8.09802-7. Douglas,J.M.,1988.ConceptualDesignofChemicalProcesses.

McGraw-Hill,NewYork.

Dusselier,M.,VanWouwe,P.,Dewaele,A.,Makshina,E.,Sels,B.F., 2013.Lacticacidasaplatformchemicalinthebiobased economy:theroleofchemocatalysis.EnergyEnviron.Sci.6, 1415–1442,http://dx.doi.org/10.1039/C3EE00069A.

Dutta,S.,De,S.,Saha,B.,Alam,M.I.,2012.Advancesin conversionofhemicellulosicbiomasstofurfuraland

(14)

upgradingtobiofuels.Catal.Sci.Technol.2,2025–2036, http://dx.doi.org/10.1039/C2CY20235B.

Ek,M.,Gellerstedt,G.,Henriksson,G.,2009a.WoodChemistry andWoodBiotechnology.DeGruyter,Berlin,

http://dx.doi.org/10.1515/9783110213409.

Ek,M.,Gellerstedt,G.,Henriksson,G.,2009b.PulpingChemistry andTechnology.DeGruyter,Berlin,

http://dx.doi.org/10.1515/9783110213423.

Ferreira,A.,Egas,A.,Domingues,A.,Fernandes,C.,Lobo,L.,2011. VaporpressureandboilingpointelevationofEucalyptuskraft blackliquors.SilvaLusit.

Francisco,M.,González,A.S.B.,GarcíadeDios,S.L.,Weggemans, W.,Kroon,M.C.,2013a.Comparisonofalowtransition temperaturemixture(LTTM)formedbylacticacidandcholine chloridewithcholinelactateionicliquidandthecholine chloridesalt:physicalpropertiesandvapour–liquidequilibria ofmixturescontainingwaterandethanol.RSCAdv.3, 23553–23561,http://dx.doi.org/10.1039/C3RA40303C. Francisco,M.,vandenBruinhorst,A.,Kroon,M.C.,2013b.

Low-transition-Temperaturemixtures(LTTMs):anew generationofdesignersolvents.Angew.Chem.Int.Ed.52, 3074–3085,http://dx.doi.org/10.1002/anie.201207548. Francisco,M.,vandenBruinhorst,A.,Zubeir,L.F.,Peters,C.J.,

Kroon,M.C.,2013c.Anewlowtransitiontemperaturemixture (LTTM)formedbycholinechloride+lacticacid:

characterizationassolventforCO2capture.FluidPhase

Equilib.340,77–84,

http://dx.doi.org/10.1016/j.fluid.2012.12.001.

García,G.,Aparicio,S.,Ullah,R.,Atilhan,M.,2015.Deepeutectic solvents:physicochemicalpropertiesandgasseparation applications.EnergyFuels29,2616–2644,

http://dx.doi.org/10.1021/ef5028873.

Glass,M.,Aigner,M.,Viell,J.,Jupke,A.,Mitsos,A.,2016. Liquid-liquidequilibriumof2-methyltetrahydrofuran/water overwidetemperaturerange:measurementsandrigorous regression.FluidPhaseEquilib.433,212–225,

http://dx.doi.org/10.1016/j.fluid.2016.11.004.

Gomez,F.J.V.,Espino,M.,Fernández,M.A.,Silva,M.F.,2018.A greenerapproachtopreparenaturaldeepeutecticsolvents. ChemistrySelect3,6122–6125,

http://dx.doi.org/10.1002/slct.201800713.

Guo,Z.,Zhang,Q.,You,T.,Ji,Z.,Zhang,X.,Qin,Y.,Xu,F.,2019a. Heteropolyacidsenhancedneutraldeepeutecticsolvent pretreatmentforenzymatichydrolysisandethanol

fermentationofMiscanthus×giganteusundermildconditions. Bioresour.Technol.293,122036,

http://dx.doi.org/10.1016/j.biortech.2019.122036. Guo,Z.,Zhang,Q.,You,T.,Zhang,X.,Xu,F.,Wu,Y.,2019b.

Short-timedeepeutecticsolventpretreatmentforenhanced enzymaticsaccharificationandligninvalorization.Green Chem.21,3099–3108,http://dx.doi.org/10.1039/c9gc00704k. Harmsen,G.J.,2004.Industrialbestpracticesofconceptual

processdesign.Chem.Eng.Process.ProcessIntensif.43, 671–675,http://dx.doi.org/10.1016/j.cep.2003.02.003.

Haz,A.,Stri ˇzincová,P.,Majová,V.,Andrea,S.,Jablonsky,M.,2016. Thermalstabilityofselecteddeepeutecticsolvents.Int.J.Sci. Res.7,14441–14444.

Hou,X.-D.,Li,A.-L.,Lin,K.-P.,Wang,Y.-Y.,Kuang,Z.-Y.,Cao,S.-L., 2018a.Insightintothestructure-functionrelationshipsof deepeutecticsolventsduringricestrawpretreatment. Bioresour.Technol.249,261–267,

http://dx.doi.org/10.1016/j.biortech.2017.10.019.

Hou,X.-D.,Lin,K.-P.,Li,A.-L.,Yang,L.-M.,Fu,M.-H.,2018b.Effect ofconstituentsmolarratiosofdeepeutecticsolventsonrice strawfractionationefficiencyandthemicro-mechanism investigation.Ind.CropsProd.120,322–329,

http://dx.doi.org/10.1016/j.indcrop.2018.04.076.

Hoydonckx,H.E.,VanRhijn,W.M.,VanRhijn,W.,DeVos,D.E., Jacobs,P.A.,2007.Furfuralandderivatives.In:Ullmann’s EncyclopediaofIndustrialChemistry.Wiley,Weinheim, http://dx.doi.org/10.1002/14356007.a12119.pub2.

Hu,S.,Zhang,Z.,Zhou,Y.,Han,B.,Fan,H.,Li,W.,Song,J.,Xie,Y., 2008.Conversionoffructoseto5-hydroxymethylfurfural

usingionicliquidspreparedfromrenewablematerials.Green Chem.10,1280–1283,http://dx.doi.org/10.1039/b810392e. Jablonsk ´y,M., ˇSkulcová,A.,Kamenská,L.,Vrˇska,M., ˇSíma,J.,

2015.Deepeutecticsolvents:fractionationofwheatstraw. BioResources10,8039–8047,

http://dx.doi.org/10.15376/biores.10.4.8039-8047. Jablonsk ´y,M.,Haz,A.,Majova,V.,2019.Assessingthe

opportunitiesforapplyingdeepeutecticsolventsfor fractionationofbeechwoodandwheatstraw.Cellulose26, 7675–7684,http://dx.doi.org/10.1007/s10570-019-02629-0. Kajimoto,T.,Tachibana,Y.,Maeda,Y.,Kubota,S.,Hata,T.,

Imamura,Y.,2008.Characterizationofthepulp-likefibers separatedfromSugiwithl-lacticacid.MokuzaiGakkaishi54, 319–326,http://dx.doi.org/10.2488/jwrs.54.319.

Kujawski,W.,2000.Pervaporativeremovaloforganicsfromwater usinghydrophobicmembranes.Binarymixtures.Sep.Sci. Technol.35,89–108,http://dx.doi.org/10.1081/SS-100100145. Kumar,A.K.,Parikh,B.S.,Pravakar,M.,2016.Naturaldeepeutectic

solventmediatedpretreatmentofricestraw:bioanalytical characterizationofligninextractandenzymatichydrolysisof pretreatedbiomassresidue.Environ.Sci.Pollut.Res.23, 9265–9275,http://dx.doi.org/10.1007/s11356-015-4780-4. Lange,J.-P.,2001.Fuelsandchemicalsmanufacturing;guidelines

forunderstandingandminimizingtheproductioncosts. CATTECH5,82–95,http://dx.doi.org/10.1023/A:1011944622328. Li,T.,Lyu,G.,Liu,Y.,Lou,R.,Lucia,L.A.,Yang,G.,Chen,J.,Saeed,

H.A.M.,2017.Deepeutecticsolvents(DESs)fortheisolationof willowlignin(Salixmatsudanacv.Zhuliu).Int.J.Mol.Sci.18, 2266,http://dx.doi.org/10.3390/ijms18112266.

Li,A.L.,Hou,X.D.,Lin,K.P.,Zhang,X.,Fu,M.H.,2018.Ricestraw pretreatmentusingdeepeutecticsolventswithdifferent constituentsmolarratios:biomassfractionation, polysaccharidesenzymaticdigestionandsolventreuse.J. Biosci.Bioeng.126,346–354,

http://dx.doi.org/10.1016/j.jbiosc.2018.03.011.

Li,L.,Yu,L.,Wu,Z.,Hu,Y.,2019a.Delignificationofpoplarwood withlacticacid-baseddeepeutecticsolvents.WoodRes.64, 499–514.

Li,P.,Zhang,Q.,Zhang,X.,Zhang,X.,Pan,X.,Xu,F.,2019b. Subcellulardissolutionofxylanandligninforenhancing enzymatichydrolysisofmicrowaveassisteddeepeutectic solventpretreatedPinusbungeanaZucc.Bioresour.Technol. 288,121475,http://dx.doi.org/10.1016/j.biortech.2019.121475. Liang,X.,Fu,Y.,Chang,J.,2019.Effectiveseparation,recoveryand

recyclingofdeepeutecticsolventafterbiomassfractionation withmembrane-basedmethodology.Sep.Purif.Technol.210, 409–416,http://dx.doi.org/10.1016/j.seppur.2018.08.021. Lim,W.L.,Gunny,A.A.N.,Kasim,F.H.,AlNashef,I.M.,Arbain,D.,

2019.Alkalinedeepeutecticsolvent:anovelgreensolventfor lignocellulosepulping.Cellulose26,4085–4098,

http://dx.doi.org/10.1007/s10570-019-02346-8.

Lima,F.,Gouvenaux,J.,Branco,L.C.,Silvestre,A.J.D.,Marrucho, I.M.,2018.Towardsasulfurcleanfuel:deepextractionof thiopheneanddibenzothiopheneusingpolyethylene glycol-baseddeepeutecticsolvents.Fuel234,414–421, http://dx.doi.org/10.1016/j.fuel.2018.07.043.

Lima,F.,Dave,M.,Silvestre,A.J.D.,Branco,L.C.,Marrucho,I.M., 2019.Concurrentdesulfurizationanddenitrogenationoffuels usingdeepeutecticsolvents.ACSSustain.Chem.Eng.7, 11341–11349,

http://dx.doi.org/10.1021/acssuschemeng.9b00877.

López-Salas,N.,Jardim,E.O.,Silvestre-Albero,A.,Gutiérrez,M.C., Ferrer,M.L.,Rodríguez-Reinoso,F.,Silvestre-Albero,J.,Del Monte,F.,2014.Useofeutecticmixturesforpreparationof monolithiccarbonswithCO2-adsorptionandgas-separation

capabilities.Langmuir30,12220–12228, http://dx.doi.org/10.1021/la5034146.

Lu,Q.,Liu,W.,Yang,L.,Zu,Y.,Zu,B.,Zhu,M.,Zhang,Y.,Zhang, Xiunan,Zhang,R.,Sun,Z.,Huang,J.,Zhang,Xiaonan,Li,W., 2012.Investigationoftheeffectsofdifferentorganosolv pulpingmethodsonantioxidantcapacityandextraction efficiencyoflignin.FoodChem.131,313–317,