Transcript
Bioethanol, biohydrogen and biogas production from wheat strawin a biorefinery concept Prasad Kaparaju a , María Serrano a , Anne Belinda Thomsen b , Prawit Kongjan a , Irini Angelidaki a, * a Department of Environmental Engineering, Technical University of Denmark, Building 115, DK-2800 Kgs. Lyngby, Denmark b Biosystems Department, RIS Ø -DTU, Building 301, DK-4000, Roskilde, Denmark a r t i c l e i n f o Article history: Received 29 June 2008Receivedinrevisedform10November 2008Accepted 11 November 2008Available online 8 January 2009 Keywords: BiorefineryBioethanolBiogasBiohydrogenHydrothermal pretreatment a b s t r a c t The production of bioethanol, biohydrogen and biogas from wheat straw was investigated within a bior-efinery framework. Initially, wheat straw was hydrothermally liberated to a cellulose rich fiber fractionand a hemicellulose rich liquid fraction (hydrolysate). Enzymatic hydrolysis and subsequent fermenta-tion of cellulose yielded 0.41g-ethanol/g-glucose, while dark fermentation of hydrolysate produced178.0ml-H 2 /g-sugars. The effluents from both bioethanol and biohydrogen processes were further usedto producemethane with the yields of 0.324 and0.381m 3 /kg volatile solids (VS) added , respectively. Addi-tionally, evaluation of six different wheat straw-to-biofuel production scenaria showed that either use of wheat straw for biogas production or multi-fuel production were the energetically most efficient pro-cesses compared to production of mono-fuel such as bioethanol when fermenting C6 sugars alone. Thus,multiple biofuels production from wheat straw can increase the efficiency for material and energy andcan presumably be more economical process for biomass utilization. Ó 2008 Elsevier Ltd. All rights reserved. 1. Introduction Lignocellulosic materials from agriculture and forest manage-mentarethelargestsourcesofhexose(C-6)andpentose(C-5)sug-ars with a potential for the production of biofuels, chemicals andother economic by-products. Progress in this area will not onlydecouple the food and biofuel production and reduced CO 2 emis-sionsbutalsoensureamorestableandsecuredenergysupplyespe-cially in transport sector. Biofuels generated globally fromlignocelluloses are estimated at about 30EJ/year, compared to thetotalenergyusedworldwideofover400EJ/year(McKendry,2002).Biorefineries for production of several products and by-prod-ucts such as biofuels, heat and/or electricity have been in focusin the recent years (Chen et al., 2005; Zhang, 2008). In a biorefin-ery, biomasscanbeconvertedtousefulbiomaterialsand/orenergycarriers in an integrated manner and thereby it can maximize theeconomic value of the biomass used while reducing the wastestreams produced (Thomsen, 2005). Development of multiple bio-fuels based biorefinery from lignocellulose is seen as an importantpossibility to increase the efficiency for materials and energy, andreduce the costs of biomass options to mitigate GHG emissions(Sheehan et al., 2003).Lignocellulose composed of cellulose (40–50%), hemicelluloses(25–35%) and lignin (15–20%) is extremely resistant to enzymaticdigestion. Therefore, a thermochemical pretreatment is usuallynecessary to disrupt the plant cell wall (lignin) in order to improveenzymaticdigestibility(Fanetal.,2006).Thethermalpretreatmentof biomass results in two main streams; the solid fraction mainlyconsistingofcellulose(hexose:glucose)(Klinkeetal., 2002)andli-quid phase (hydrolysate) mainly consisting of hemicellulose (pen-tose: xylose and arabinose) (Bercier et al., 2007). Hexoses caneffectively be converted to bioethanol and the process is carriedoutwithhighyield(around0.4–0.51g-ethanol/g-glucose)andpro-ductivity (up to 1.0gL À 1 h À 1 ) by Saccharomyces cerevisiae or re-combinant S. cerevisiae (Hjersted and Henson, 2006; Öhgrenetal.,2006).Wild S. cerevisiae strainsareunabletoutilizepentoses.Several recombinant candidates for pentose sugars fermentationhave been described and presented as the future solution ( JinandJeffries,2004;Ruohonenetal.,2006;ChuandLee,2007).Addi-tionally, several new bacterial ethanologenic strains have been re-ported (Ahring et al., 1999; Nigam, 2001; Georgieva and Ahring,2007). Meanwhile, none of these organisms were as efficient as S.cerevisiae and improvements would certainly be desirable. Oftenthese microorganisms are suffering from low productivities (e.g. Thermoanaerobacter mathranii with productivity of 0.10gL À 1 h À 1 ),low ethanol tolerance and high sensitivity to inhibitors present inthe hydrolysates from the pretreatment step (Torry-Smith et al.,2003). Therefore, utilization of hemicellulose remains a challengeto be resolved. One alternative prospect for utilization of hemicel-lulose is to produce biohydrogen. Biohydrogen production of sug-ars through anaerobic fermentation is recognized as a verypromising, environmental friendly and feasible process (Hawkeset al., 2007). Several studies for utilization of C5 sugars to 0960-8524/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.doi:10.1016/j.biortech.2008.11.011 * Corresponding author. Tel.: +45 45251429; fax: +45 45932850. E-mail address:
[email protected](I. Angelidaki).Bioresource Technology 100 (2009) 2562–2568 Contents lists available atScienceDirect Bioresource Technology journal homepage:www.elsevier.com/locate/biortech biohydrogen have been reported (Lin et al., 2008; Lo et al., 2008;andWu et al., 2008). The extreme thermophile Caldicellulosiruptor saccharolyticus was used successfully to convert C-5 sugar (xylose)to hydrogen with a relatively high yield of 334.7ml-H 2 /g-sugar,accounting to 67% of theoretical yield of 497.6ml-H 2 /g-sugar(Kádár et al., 2004).Another important challenge for biorefineries, is to handle thelarge amounts of wastewater streams generated by the process.A sustainable solution for removal of the residual organic matterin the effluents from bioethanol and biohydrogen processes is toconvert them to biogas and use the residual effluents as fertilizerson agricultural soil, (Torry-Smith et al., 2003; Liu et al., 2006).Use of wheat straw for single production of bioethanol, biogasand biohydrogen has been demonstrated successfully (Lindeet al., 2007; Fan et al., 2006). Recently, coproduction of bioethanolwith other biofuels such as biogas from energy crops like winterrye, oilseed rape and faba bean was demonstrated (Peterssonet al., 2007). However, there is still need for studies on multibiofu-els (bioethanol, biohydrogen and biogas) production based biore-finery from agricultural residues which could assist evaluation of newbiofuel concepts.Theobjectiveof thisinvestigationwastoex-plorethepossibilityofusingwheatstrawfortheproductionofbio-ethanol (from cellulose), biohydrogen (from hemicellulose) andbiogas (from effluents from bioethanol and biohydrogen produc-tion) in an integrated biorefinery concept to improve the overallefficiency of biomass utilization. Wheat straw and its by-productsgenerated during the integrated bioethanol, biohydrogen and bio-gas production in a biorefinery concept were characterized. Themethane potential from wheat straw and its by-products/effluentsin a biorefinery process was then determined. Finally, the outputenergy for six different wheat straw-to-biofuel scenaria was esti-mated quantitatively to evaluate the efficient utilization of the en-ergy content of wheat straw. 2. Methods 2.1. Raw materials Wheat straw( Triticum aestivum L.) grown and harvested duringsummer2003inDenmarkwasobtainedfromRisøNationalLabora-tories, Technical University of Denmark (Denmark). The straw wascutinto1–5cmpiecesonthefieldbyforageharvesterandstoredincontainers at ambient temperature until further use. For chemicalanalysis and biogas potential assays, wheat straw was milled to<1mm.Thedrymatter(DM)contentwas90–91%(w/w).Thechem-ical composition of the untreated wheat straw is given inTable 1. 2.2. Pretreatment of wheat straw Hydrothermal pretreatment of wheat straw carried out in pilotplant(100kg/hcapacity)wasdescribedelsewhere(Thomsenetal.,2008).Briefly,wheatstrawattherateof120–150kg-DM/hwasfedcountercurrentlywithwater at aflowrateof 400–600L/htothreeserial reactors. The first step being a soaking step was operated attemperature of 80 ° C and residence time around 6min. The pre-soaking wheat straw was then heated up in stage two to approx.180 ° C for 15min. followed by heating at 190 ° C for 3min. in athird stage. Hydrothermal pretreatment of wheat straw resultedin a liquid fraction called hydrolysate, containing mainly hemicel-luloses and a solid fraction rich in cellulose. These fractions werekindly supplied by Dong Energy, Denmark. All materials were col-lected and stored at À 20 ° C until further use. Glucose and pentoserecovery were calculated as follows: Glucose recovery ð % Þ¼ Total glucose solid fraction þ Total glucose liquid fraction Total glucose straw 100 ð 1 Þ Pentose recovery ð % Þ¼ Total pentose solid fraction þ Total pentose liquid fraction Total pentose straw 100 ð 2 Þ 2.3. Bioethanol production Prehydrolysis(liquefaction)andfermentationofsolidfiberfrac-tionwastreatedbyfirstsaccharificationandsubsequentlyfermen-tation by Bakers yeast ( S. cerevisiae ) as described elsewhere(Thomsen et al., 2006). Briefly, the experiment was performed in200ml fermentation flasks. Eight grams of the solid fiber fraction Table 1 Chemical characterization of wheat straw and its by-products during bioethanol and biohydrogen production. (The values are a mean of triplicate measurements.) Wheat straw Hydrolysate Fiber fraction Hydrogen effluents Stillage (Risø) b Stillage (Sweden) b pH – 4.9±0.1 – 5.6±0.1 3.6±0.1 4.0±0.1TS (%) 91.6±0.02 4.4±0.01 32.5±0.57 3.5±0.01 12.0±0.03 19.6±0.18VS (%) 87.5±0.02 3.3±0.01 29.5±0.40 2.8±0.01 10.2±0.03 17.8±0.18Ash content (%) 4.1±0.02 1.1±0.01 3±0.49 0.7±0.01 1.8±0.03 1.8±0.18COD (g/l) – 37.9±1.31 – 24.2±0.30 149.8±3.59 170.7±0.38SCOD (g/l) – 32.05±2.22 – 8.9±1.23 60.9±4.36 85.08±0.13VFA (g/l) 0.13±0.02 0.7±0.14 0.61±0.04 0.6±0.04 0.18±0.02 0.37±0.02Ethanol (g/l) N.D. N.D. N.D. N.D. 2.3±0.13 0.8±0.10Total nitrogen (g/l) 1.3±0.04 0.2±0.01 0.40±0.01 0.32±0.02 1.4±0.02 6.2±0.20Ammonia (g/l) 0.31±0.01 0.03±0.01 0.03±0.01 0.15±0.01 0.16±0.01 1.3±0.02Proteins c (g/l) 6.5±0.17 1.1±0.03 2.3±0.2 1.1±0.04 7.7±0.09 38.8±1.15Lipids (%) 1.5± 0.24±0.01 1.2±0.04 0.02±0.01 0.99±0.02 0.93±0.12Carbohydrates d (g/l) 853.1 30.5 265.6 14.2 84.5 129.3Furfurals (g/l) – 0.25±0.04 – N.D. N.D. N.D.HMF (g/l) – 0.14±0.02 – N.D. N.D. 0.02±0.01Phenols (g/l) – 0.14±0.12 – 0.015±0.23 0.06±0.01 0.08±0.02Klason lignin (g/l) 19.3±0.10 a N.A. 25.7±0.12 a N.A. 24.6±0.21 a 16.5±0.13 a Arabinose (g/l) 2.6±0.19 a 1.3±0.05 0.4±0.01 a 0.02±0.01 N.D. a 6.9±0.04 a Xylose (g/l) 21.3±0.30 a 11.3±0.15 8.3±0.12 a 0.11±0.05 8.2±0.03 a 21.3±0.12 a Glucose (g/l) 35.9±0.03 a 2.9±0.21 49.6±0.21 a 0.00±0.01 0.1±0.11 a 29.9±0.21 aa Value expressed in g/100gDM. b After stripping the ethanol; Swedish stillage obtained from industrial plant producing ethanol from wheat straw and grains. Risø stillage obtained from lab-scalefermentor producing ethanol from wheat straw alone; N.D., not detected. c Proteins=(TKN À NH þ 4 ). d CHO=VS À proteins À lipids À VFA. P. Kaparaju et al./Bioresource Technology 100 (2009) 2562–2568 2563 were mixed with 60ml of a 0.2M acetate buffer (pH 4.8). Prehy-drolysis of the solid fraction was performed at 50 ° C for 24h atan enzyme loading of 15FPU/g DM filter cake using Cellubrix L.After liquefaction, the flasks were supplemented with a seconddose of Cellubrix L enzyme at a loading of 20FPU/g DM and0.2ml of urea (24%). The suspensions were then inoculated with0.2g yeast after cooling down to room temperature. The flaskswere sealed with a loop trap filled with glycerol and incubated at32 ° C for 6–8h. The CO 2 -production was followed by measuringtheweightlossindicatingtheethanolyield(=0.51g à CO 2 ).Thefinalethanol yield was also measured by HPLC. Glucose yield solid fraction was calculated as follows: Glucose yield solid fraction ð % Þ¼ Free glucose after enzymatic hydrolysisTotal glucose solid fraction 100 ð 3 Þ 2.4. Biohydrogen production At the same time, hydrolysate was used for biohydrogenproduction through dark fermentation as described in detail byKongjan et al. (2008). The experiment was carried out in a 1L con-tinuously stirred tank reactor (CSTR) reactor with 700ml workingvolume and hydraulic retention time (HRT) of 72h. The feed wascomposed of BA medium and 40% hydrolysate solution (1:1 ratioV/V). The reactor was first started-up in a batch fed mode with140ml of inoculum from batch cultures and 560ml of 25% hydro-lysate diluted with BA medium. Continuous mode was startedwhen hydrogen content in the gas phase during the batch modereached maximum. The reactor temperature was controlled at70 ° C by circulating hot water inside the reactor jacket. 2.5. Biological methane potential Methane production was then investigated comparatively byusingstillages,theeffluentsfrombioethanolfermentation,obtainedfrom the previous experiments in Section2.3and from a Swedishindustrialplantproducingethanolfrombothwheatstrawandfromgrainandtheeffluentsofbiohydrogenproduction.Allsubstratesformethane production were also kept at À 20 ° C until digestion. Bio-logical methane potentials were evaluated in batch experiments.Theexperimentswereperformedin118mlserumglassbottleswithworkingvolumeof40mlconsistingof30mlofinoculumand10mlof substrate. Substrate was diluted with distilled water to attain asubstrateconcentrationof5,50or100%.TheheadspaceinthebottlewasflushedwithpureN 2 for 3–5min. Inaddition, 2–3dropsof so-diumsulphide was added to ensure anaerobic conditions. The bot-tles were sealed with rubber stoppers and aluminium crimps. Theexperimentwasconductedintriplicates.Thepreparedbottleswereincubated statically at 55 ° C. Assays with digested manure alonewere used as controls. Methaneproduced frominoculumwas sub-tracted from the assays. Digested manure from a pilot-scale planttreating cow manure at 55 ° C was used as inoculum (Kaparajuet al., 2008). Theoretical methane yield (m 3 /kg-VS) was calculatedbased on the stoichiometric conversion of organic matter to meth-aneandcarbondioxideasfollows: 2.6. Analytical methods Total and soluble Chemical oxygen demand (TCOD and SCOD),Total solids (TS), volatile solids (VS), ash content, suspended solids(SS), ammonia and total Kjeldahl nitrogen(TKN) were determinedaccording to the Standard Methods (APHA, 1998). Standards SCODsamples were filtered through glass fiber filter paper ( U 90mm,GF50, Schleicher & Schuell). Lipid extraction was carried outthrough Soxhlet Method. pH was measured using pH meter(PHM92 LAB).Methane production was measured by gas chromatography(GC) using flame ionization detection (FID) fitted with a Porapak60/80molsieve column. Nitrogen is used as carrier gas with apressure of 2.0kg/cm 2 . The injection temperature is set to 110 ° C.The detector and oven temperature is 160 ° C.ForVFAandalcoholdetermination,1mlofsamplewasacidifiedwith 50 l L of 34% of phosphoric acid and then centrifuged at12,000rpm for 10min and measured on GC (Hewlett Packard, HP5890 series II) equipped with a flame ionization detector (FID)and HP FFAP column (dimensions 30m  0.53mm  1.0mm).The temperature program for the column was increased from50 ° Cto190 ° Cwitharateof15 ° C/min. Thetemperaturesofinjec-tion port and detector were 200 and 150 ° C, respectively. Nitrogenwas used as the carrier gas at a flow rate of 10ml/min.Todeterminethesugars(glucose,xyloseandarabinose)contentin raw and pretreated straw and liquid fractions Strong acid (72%W/W H 2 SO 4 ) hydrolysis of solid fraction and weak acid (4% W/WH 2 SO 4 )ofliquidfractionwasapplied(Thomsenetal.,2006).Sugarswerequantifiedonhighperformance liquid chromatography HPLCsystem HP 1100 (Agilent 1100) equipped with a BioRad AminexHPX-87H at 63 ° C and a refractive index (RI) detector (RID1362A)using0.6ml/minof4mMH 2 SO 4 aseluent.Manoseandgal-actose could not be separated clearly from glucose and xylose be-cause the manose retention time (10.02min) was very close tothe glucose retention time (10.16min) while, the galactose reten-tion time (10.49min) was also very close to the xylose retentiontime (10.39min). That is, although manose and galactose couldnot be separately measured, possible presence of these two sugarswouldhavebeenincludedintheHPLCpeakstogetherwithglucoseandxylose.Additionally,glucose,xyloseandarabinosearethemainmonomersinwheatstraw,monomersdetectedwerethereforerep-resent in total glucose, xylose and arabinose (Klinke et al., 2002;Bercieretal.,2007).Theanalysisdetectionlimitsforglucose,xyloseand arabinose were 0.011, 0.002 and 0.014g/l, respectively.Klason lignin in solid fraction was determined as the weight of the filter cake (generated during the strong acid hydrolysis of solidfraction) fromsubtracted the ash content. Furfural and 5-hydroxy-methyl-2-furaldehyde (HMF) in liquid fraction were measured onHPLC used for sugars analysis fitted with ultraviolet (UV) detector(G1314A).Phenoliccompoundsderivedfrompretreatment ofwheatstrawwere quantified by GC equipped by FID. Compounds were firstlyisolated from the liquid fraction at pH 2 by solid-phase extractionon polystyrene divinylbenzene polymer columns (Klinke et al.,2002). 2.7. Energy output of a biorefinery concept Six different wheat straw to biofuel energy production scenariawere considered to evaluate the most energetically efficient pro-cess in a biorefinery (1) Untreated wheat straw ? Incineration(2) Untreated wheat straw ? Biogas (3) Pretreated wheatstraw ? Biogas (4) Pretreated wheat straw ? Bioethanol (5) Pre-treated wheat straw ? Bioethanol ? Biogas (6) Pretreated wheat B o : th ¼ 0 : 415Carbohydrates þ 0 : 496Proteins þ 1 : 014Lipids þ 0 : 373Acetate þ 0 : 530Propionate ð Carbohydrates þ Proteins þ Lipids þ Acetate þ Propionate Þð 4 Þ 2564 P. Kaparaju et al./Bioresource Technology 100 (2009) 2562–2568 straw ? Bioethanol ? Biohydrogen ? Biogas. Data on the chemi-cal composition and bioethanol, biohydrogen, methane yields of substrateand/oreffluentfromeachprocessobtainedinthepresentstudy were used to estimate quantitatively the energy output foreach scenario. Energy was expressed in MJ.Mass flow in the studied biorefinery concept was calculatedbased on the amount of sugars and their conversion to differentbiofuels/products (Fig. 1). Pretreatment often results in loss of drymatter.ThiscanbeduetoformationofgasseslikeCO 2 andace-tic acid at high temperature and/or loss of dry matter along withwater as biomass is washed with water in order to remove theinhibitorymaterialsand/orotherwater-solublehemicellulosepro-ducedduringthermal hydrolysis. Losses of pretreatedwheat strawthrough steam explosion was reported to be up to 20% (Bjerre andSchmidt, 1997). In the present study, a dry matter loss of 5% wasassumed, as the system was well closed. From the remaining drymatter, around 75% of the dry matter was transferred to the solidfraction containing mainly cellulose and, lignin and the remaining25% of the dry matter was retained in the liquid fraction.Finally, the energy output for each studied scenario was esti-mated by multiplying the amount of individual biofuel producedwith its lower calorific value (LCV). The LCV for ethanol, methane,hydrogen, dry lignin and wheat straw were 26.72, 50.1, 122, 20.95and 19.1MJ/Kg, respectively (ORNL, 2006). 3. Results and discussion 3.1. Characteristics of wheat straw and its by-products/effluents in abiorefinery 3.1.1. Wheat straw The chemical composition of wheat straw and effluents formedduring bioethanol and biohydrogen fermentationof the pretreatedstraw streams; i.e. solids fraction or liquid fraction (hydrolysate),respectively, are shown inTable 2. Untreated wheat straw haddry matter content of 91.6%. Cellulose, hemicellulose and lignincontents were 35.9, 23.9, and 19.3g/100g-DM, respectively. Theseresults are in accordance with the typical composition of wheatstraw reportedThomsen et al. (2006). 3.1.2. Solid and hydrolysate fraction Composition of fiber fraction (Table 1) showed that hydrother-mal pretreatment affected the degradation and solubilization of individual sugars. Glucan and xylan were noticed while arabinanwas found in trace amount. The solid fraction had rather highDM of 32.5%-TS. It has been stated that a high DM concentrationof 32.5%-TS in the solid fraction could be possible to perform eth-anol fermentation ( Jorgensen et al., 2007). The high DM content(>20%-TS) in the solids fraction would give sugar and ethanol con-centration above 8% and 4% w/w, respectively during the enzy-matic hydrolysis and fermentation (Larsen et al., 2008).Significant reduction of the distillation costs could be therebyachieved when the ethanol from fermentation broth is above 4%(w/w) (Zacchi and Axelsson, 1989).Analysis of hydrolysate (Table 1) revealed that xylose was themain sugar accounting for 72.9% of sugars which could be con-verted to hydrogen in the subsequent dark fermentation. Glucoseand arabinose were also found in the hydrolysate but at a verylow concentration. The low glucose concentration in hydrolysatewas probably due to its crystalline and thermally stable structure.The relatively low concentration of degradation products such asacetic acid, hydroxymethylfurfural (HMF) furfural and phenoliccompounds, which were only found in the hydrolysate (Table 1)could be an advantage for the subsequent fermentation. Thesecompounds generated during the hydrothermal pretreatment of straw are considered to be inhibitory to most microorganisms(Thomsen et al., 2006). The rather low pH of hydrolysate(pH=4.9) was properly due to high amount acetic acid contentinhydrolysate,accountingfor80%oftotalVFAs.Aceticacidwasre-leased fromthe hydrolysis of acetyl groups contained in the hemi- Fibers DM(Kg)Enz. HydrolysateDM(Kg)Ferm. BrothDM(Kg) Stillage DM(Kg) 652,7692,65416,5296,1 Cellulose323,7Cellulose32,4Cellulose32,4Glucose 2,5Hemic.56,8Hemic.56,8Hemic.56,8XyloseLignin167,7Glucose291,3Ethanol133,7Lignin167,7AshLignin167,7CO2128,2Ash 19,6Wheat straw (DM (Kg)Other 84,8Ash 19,6Lignin167,7Ethanol13,4 1000916,0 Other 124,8Ash 19,6Other140,0Cellulose359,7Other 140,0 Ethanol120,4 Hemic.239,5 CO 2 128,2 Lignin193,3AshOther 85,9HydrolysateDM(Kg)Effluent DM (Kg) 217,6173,051 Glucose14,3Glucose 0Hemic.62,3Hemic. 0,6AshAsh 5,4Other 135,5Other164,9Hydrogen2,0 R1 (80 o C, 20 min)R2 (180 o C 15 min)R3 (195 o C 3 min)PRETREATMENTENZYMATICHYDROLYSISFERMENTATIONDISTILLATIONBIOHYDROGENPRODUCTION ab BIOGASPRODUCTIONWater400-600 l/hEnzyme40 kg19,637,65,456,8 Fig. 1. Mass flow in the biorefinery process. Table 2 Methane potentials of wheat straw and its by-products obtained in a biorefinery concept. Substrate Substrate concentration (g-VS/l) Dilution (%) Ultimate CH 4 yield ( B o , m 3 /kg-VS) Theoretical CH 4 yield ( B u , m 3 /kg-VS) Biodegradability ( B o / B u )Wheat Straw 21.9 50 0.297±0.01 0.426 0.69Hydrolysate 8.5 50 0.384±0.08 0.459 0.84Solid fraction 6.7 95 0.386±0.02 0.418 0.87H 2 effluent 2.1 50 0.381±0.01 0.427 0.89Risø stillage 12.8 50 0.324±0.03 0.479 0.68Swedish stillage 21.3 50 0.485±0.01 0.464 >1.0 P. Kaparaju et al./Bioresource Technology 100 (2009) 2562–2568 2565
