《生物质能工程》课程教学资源（阅读材料）5-Comparison of solid-state to liquid anaerobic digestion

Bioresource Technology 124(2012)379-386 Contents lists available at SciVerse ScienceDirect 88恶 Bioresource Technology ELSEVIER journal homepage:www.elsevier.com/locate/biortech Comparison of solid-state to liquid anaerobic digestion of lignocellulosic feedstocks for biogas production Dan Brown,Jian Shi,Yebo Li* Department of Food,Agricultural,and Biological Engineering.The Ohio State University/Ohio Agricultural Research and Development Center,1680 Madison Ave,Wooster, OH 44691-4096.United States HIGH LIGH TS Solid state anaerobic digestion(SS-AD)of eight types of lignocellulosic biomass. Liquid anaerobic digestion (L-AD)of eight biomass feedstocks was compared with SS-AD. No significant difference in methane yield between SS-AD and L-AD. Volumetric biogas productivity of SS-AD was 2-to 7-fold greater than that with L-AD. Methane yields from crop residues were higher than those from woody biomass. ARTICLE IN FO ABSTRACT Article history: Lignocellulosic biomass feedstocks (switchgrass,corn stover,wheat straw,yard waste,leaves,waste Received 20 April 2012 paper,maple,and pine)were evaluated for methane production under liquid anaerobic digestion(L- Received in revised form 11 August 2012 AD)and solid-state anaerobic digestion(SS-AD).No significant difference in methane yield between L- Accepted 13 August 2012 Available online 22 August 2012 AD and SS-AD.except for waste paper and pine,were found.However,the volumetric productivity was 2-to 7-fold greater in the SS-AD system compared with the L-AD system,except for paper.Methane yields from corn stover,wheat straw,and switchgrass were 2-5 times higher than those from yard waste. Keywords: Solid-state anaerobic digestion maple,and pine biomass.Waste paper had a methane yield of only 15 L/kg VS caused by souring during Biogas SS-AD due to organic overloading.Pine also had very low biogas yield of 17 L/kg VS,indicating the need Corn stover for pretreatment prior to SS-AD.The findings of this study can guide future studies to improve the Switchgrass efficiency and stability of SS-AD of lignocellulosic biomass. Lignocellulosic biomass 2012 Elsevier Ltd.All rights reserved. 1.Introduction methanogenic bacteria involved in AD have a low growth rate and are sensitive to inhibitors such as low pH caused by excessive Anaerobic digestion (AD)is a naturally occurring phenomenon concentrations of volatile fatty acids (VFAs).Thus,maintaining a in which organic matter is decomposed by an assortment of mi- balance of the four phases (hydrolysis,acidogenesis,acetogenesis crobes in an oxygen-free environment to produce biogas,composed and methanogenesis)of the AD process is essential.The effect of primarily of methane(CH4)and carbon dioxide(CO2)(Frigon and composition of lignocellulosic biomass on methane yield has al- Guiot,2010).Although the initial applications of AD were for stabil- ready been studied extensively:however,most of these studies ization and treatment of waste sludge,AD can also be a source of were limited to liquid AD (L-AD)which operates at a total solids renewable energy (Yu and Schanbacher,2010).AD not only (TS)content of 15%or less even though solid-state AD (SS-AD). provides an alternative source of energy but is also an alternative which is generally operated at a TS content of 15%or higher route to divert organic wastes and reduce greenhouse gas emissions (Guendouz et al.,2010),would be ideal for feedstocks such as agri- from landfills(Mata-Alvarez et al.,2000;Frigon and Guiot,2010). cultural and municipal solids wastes due to their availability and The presence of lignin,the crystallinity of cellulose,and limited low moisture content(Yu and Schanbacher,2010).SS-AD has many surface availability reduce the biodegradability of lignocellulosic advantages over L-AD including a smaller reactor volume for the biomass,making the hydrolysis step one of the bottlenecks that same solids loading;fewer moving parts:lower energy input for limit the production of methane (Frigon and Guiot,2010).The heating and mixing:easier to handle end product;and a greater acceptance of inputs containing glass,plastics,and grit (Li et al.. Corresponding author.Tel.:+1 330 263 3855. 2011).Problems in L-AD,such as floating and stratification of fats, E-mail address:li.851@osu.edu (Y.Li). fibers,and plastics,are not present in SS-AD(Chanakya et al.,1999). 0960-8524/-see front matter2012 Elsevier Ltd.All rights reserved. http://dx.doi.org/10.1016/j.biortech.2012.08.051

Comparison of solid-state to liquid anaerobic digestion of lignocellulosic feedstocks for biogas production Dan Brown, Jian Shi, Yebo Li ⇑ Department of Food, Agricultural, and Biological Engineering, The Ohio State University/Ohio Agricultural Research and Development Center, 1680 Madison Ave., Wooster, OH 44691-4096, United States highlights " Solid state anaerobic digestion (SS-AD) of eight types of lignocellulosic biomass. " Liquid anaerobic digestion (L-AD) of eight biomass feedstocks was compared with SS-AD. " No significant difference in methane yield between SS-AD and L-AD. " Volumetric biogas productivity of SS-AD was 2- to 7-fold greater than that with L-AD. " Methane yields from crop residues were higher than those from woody biomass. article info Article history: Received 20 April 2012 Received in revised form 11 August 2012 Accepted 13 August 2012 Available online 22 August 2012 Keywords: Solid-state anaerobic digestion Biogas Corn stover Switchgrass Lignocellulosic biomass abstract Lignocellulosic biomass feedstocks (switchgrass, corn stover, wheat straw, yard waste, leaves, waste paper, maple, and pine) were evaluated for methane production under liquid anaerobic digestion (L￾AD) and solid-state anaerobic digestion (SS-AD). No significant difference in methane yield between L￾AD and SS-AD, except for waste paper and pine, were found. However, the volumetric productivity was 2- to 7-fold greater in the SS-AD system compared with the L-AD system, except for paper. Methane yields from corn stover, wheat straw, and switchgrass were 2–5 times higher than those from yard waste, maple, and pine biomass. Waste paper had a methane yield of only 15 L/kg VS caused by souring during SS-AD due to organic overloading. Pine also had very low biogas yield of 17 L/kg VS, indicating the need for pretreatment prior to SS-AD. The findings of this study can guide future studies to improve the efficiency and stability of SS-AD of lignocellulosic biomass. 2012 Elsevier Ltd. All rights reserved. 1. Introduction Anaerobic digestion (AD) is a naturally occurring phenomenon in which organic matter is decomposed by an assortment of mi￾crobes in an oxygen-free environment to produce biogas, composed primarily of methane (CH4) and carbon dioxide (CO2) (Frigon and Guiot, 2010). Although the initial applications of AD were for stabil￾ization and treatment of waste sludge, AD can also be a source of renewable energy (Yu and Schanbacher, 2010). AD not only provides an alternative source of energy but is also an alternative route to divert organic wastes and reduce greenhouse gas emissions from landfills (Mata-Alvarez et al., 2000; Frigon and Guiot, 2010). The presence of lignin, the crystallinity of cellulose, and limited surface availability reduce the biodegradability of lignocellulosic biomass, making the hydrolysis step one of the bottlenecks that limit the production of methane (Frigon and Guiot, 2010). The methanogenic bacteria involved in AD have a low growth rate and are sensitive to inhibitors such as low pH caused by excessive concentrations of volatile fatty acids (VFAs). Thus, maintaining a balance of the four phases (hydrolysis, acidogenesis, acetogenesis and methanogenesis) of the AD process is essential. The effect of composition of lignocellulosic biomass on methane yield has al￾ready been studied extensively; however, most of these studies were limited to liquid AD (L-AD) which operates at a total solids (TS) content of 15% or less even though solid-state AD (SS-AD), which is generally operated at a TS content of 15% or higher (Guendouz et al., 2010), would be ideal for feedstocks such as agri￾cultural and municipal solids wastes due to their availability and low moisture content (Yu and Schanbacher, 2010). SS-AD has many advantages over L-AD including a smaller reactor volume for the same solids loading; fewer moving parts; lower energy input for heating and mixing; easier to handle end product; and a greater acceptance of inputs containing glass, plastics, and grit (Li et al., 2011). Problems in L-AD, such as floating and stratification of fats, fibers, and plastics, are not present in SS-AD (Chanakya et al., 1999). 0960-8524/$ - see front matter 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biortech.2012.08.051 ⇑ Corresponding author. Tel.: +1 330 263 3855. E-mail address: li.851@osu.edu (Y. Li). Bioresource Technology 124 (2012) 379–386 Contents lists available at SciVerse ScienceDirect Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

380 D.Brown et aL/Bioresource Technology 124 (2012)379-386 Feedstocks for AD are influenced by accessibility and availabil- et al.,2012).Well mixed materials were loaded into 1-L glass reac- ity due to costs associated with collection and transportation(Fri- tors and incubated in a walk-in incubation room for up to 30 days gon and Guiot,2010:Li et al,2011).Lignocellulosic biomass can be at 37t 1 C.Duplicate reactors were run for each condition.Inocu- harvested at a high TS content,which results in lower transporta- lum without any feedstock addition was used as a control.Biogas tion costs per unit of solids compared to low TS feedstocks.Crop was collected in a 5-L Tedlar gas bag (CEL Scientific,Santa Fe residues,such as corn stover and wheat straw,are widespread Springs,CA,USA)attached to the outlet of the reactor.Biogas com- through much of the US.Switchgrass has the potential to be grown position and volume were measured every 2-3 days during the 30- in areas where common crops experience low production,such as day SS-AD. areas currently in the Conservation Reserve Program through the Natural Resources Conservation Service.Yard waste,which in- 2.3.Liquid anaerobic digestion (L-AD) cludes grass,leaves,and branches,is a major lignocellulosic source generated from households,municipalities,and landscaping com- Each feedstock was mixed with deionized(DI)water and efflu- panies.Although yard waste can be composted,the stored energy ent to obtain a mixture of 5%TS and an F/E ratio of 0.5 based on dry is lost in the form of respiration heat(Koch et al.,2010).The use of volatile solids (VS).The mixtures were loaded in 2-L glass jars and yard waste,leaves,maple,and pine may improve the overall eco- incubated in a walk-in incubation room on an Innova model 2300 nomics of SS-AD due to the low or zero cost associated with acquir- platform shaker (New Brunswick Scientific;Enfield,CT.USA)at ing these feedstocks (Jagadish et al..1998).Waste paper is paper 150 rpm for up to 30 days at 37+1C.Duplicate reactors were that either cannot be recycled and/or is used in some other way run for each condition.Inoculum without any feedstock addition such as a landfill cover.Paper is treated both chemically and ther- was used as a control.Biogas generated was collected in a 5-L Ted- mally during the paper-making process,effectively serving as a lar gas bag(CEL Scientific,Santa Fe Springs,CA,USA)attached to pretreatment and allows the breakdown of cellulose to occur.A the outlet of the reactor.Biogas composition and volume were number of these lignocellulosic biomass sources have been tested measured every 2-3 days during the 30-day L-AD. for methane production from L-AD systems (Turick et al.,1991: Gunaseelan,1997);however,the suitability of lignocellulosic bio- 2.4.Enzyme hydrolysis of lignocellulosic feedstocks mass as feedstocks and factors affecting methane production dur- ing SS-AD have not been studied. Cellulase (Spezyme CP)(Lot#4900857805,protein content Therefore,the objectives of the present study were to 82 mg/ml with activity of approximately 50 FPU/ml),and Multifect determine:(1)the methane yield and volumetric productivity of xylanase(Lot#4901063047,protein content 27 mg/ml with activ- switchgrass,corn stover,wheat straw,yard waste,leaves,maple, ity of approximately 25000 OSX/ml).were obtained from Genen- pine,and waste paper in L-AD and SS-AD systems using a single cor,now DuPont Industrial Biosciences (Palo Alto,CA,USA).All source of inoculum:and(2)the effect of the composition and enzy- enzymatic hydrolysis experiments were run in duplicate following matic digestibility of lignocellulosic biomass on methane yield the National Renewable Energy Laboratory(NREL)Laboratory Ana- from SS-AD. lytical Procedures(LAP)(Selig et al.,2008)at 50C with 0.05 M cit- rate buffer,pH 4.8,solid loading of 2%,and cellulase loading of 15 FPU/g solids supplemented with xylanase at a loading of 3750 OSX/ 2.Methods g solids on a rotary shaker (150 rpm for 72 h).The hydrolysate sample was filtered through a 0.2 um nylon membrane filter for 2.1.Feedstock and inoculum sugar analysis by HPLC as described in Section 2.5.The overall glu- cose and xylose yields of enzymatic hydrolysis (i.e.,enzymatic Corn stover,wheat straw,and switchgrass were collected in digestibility)were defined according to Cui et al.(2011). October 2009 from farms operated by the Ohio Agricultural Re- search and Development Center (OARDC)in Wooster,and Jackson, 2.5.Analytical methods OH.Fallen tree leaves were collected in October 2010 from the OARDC Wooster campus.Yard waste containing leaves and tree The extractive content of feedstock was measured according to branches was obtained in June 2011 from the OARDC Wooster the NREL Laboratory Analytical Procedure (Sluiter et al.,2008a). campus.Pine and maple wood with bark were obtained in June The extractive-free biomass and solid fractions before and after 2011 from the OARDC Wooster campus.The shredded waste paper digestion were further fractionated using a two-step hydrolysis used was classified as alternative daily cover for landfills and col- method based on the NREL Laboratory Analytical Procedure(Sluit- lected from the Solid Waste Management facility of Wayne County. er et al.,2008b).Monomeric sugars (glucose,xylose,galactose OH.These feedstocks were oven-dried at 40C for 48 h in a convec- arabinose,and mannose)and cellobiose in the acid hydrolysate tion oven (Precision Thelco Model 18,Waltham,MA)to obtain a were measured by an HPLC instrument(Shimadzu LC-20AB,MD. moisture content of less than 10%,ground with a hammer mill USA)equipped with a Biorad Aminex HPX-87P column and a (Mackisik,Parker Ford,PA)to pass through a 5 mm screen,and refractive index detector (RID).A deionized water flow rate of stored in air tight containers. 0.6 ml/min was used as the mobile phase.The temperature of Effluent from a mesophillic liquid anaerobic digester in Colum- the column and detector were maintained at 80 and 55 C,respec- bus,OH fed with municipal waste water sludge and food waste tively.The Standard Methods for the Examination of Water and was used as inoculum.The inoculum was kept in air-tight buckets Wastewater were used to analyze the TS and VS contents of feed- at 4C.Prior to use,the inoculum was acclimated at 37C for stocks,inoculum,and digestate taken at the beginning and end of 1 week. the AD process(Eaton et al.,2005).Total carbon and nitrogen con- tents in feedstock,effluent and digestion materials were deter- 2.2.Solid-state anaerobic digestion (SS-AD) mined using an elemental analyzer (Elementar Vario Max CNS, Elementar Americas,Mt.Laurel,NI.USA).VFAs and alkalinity were Each lignocellulosic biomass feedstock was mixed with the measured using a two-step titration method(McGhee,1968).Sam- inoculum using a hand-mixer(Black Decker,250-watt mixer, ples for pH,VFA,and alkalinity measurements were prepared by Towson,MD,USA)to obtain a feedstock to effluent (F/E)ratio of diluting a 5-g sample with 50 ml of DI water.The mixture was ana- 3(based on dry volatile solids)and TS content of 18-19%(Liew lyzed using an auto-titrator(Mettler Toledo,DL22 Food Beverage

Feedstocks for AD are influenced by accessibility and availabil￾ity due to costs associated with collection and transportation (Fri￾gon and Guiot, 2010; Li et al., 2011). Lignocellulosic biomass can be harvested at a high TS content, which results in lower transporta￾tion costs per unit of solids compared to low TS feedstocks. Crop residues, such as corn stover and wheat straw, are widespread through much of the US. Switchgrass has the potential to be grown in areas where common crops experience low production, such as areas currently in the Conservation Reserve Program through the Natural Resources Conservation Service. Yard waste, which in￾cludes grass, leaves, and branches, is a major lignocellulosic source generated from households, municipalities, and landscaping com￾panies. Although yard waste can be composted, the stored energy is lost in the form of respiration heat (Koch et al., 2010). The use of yard waste, leaves, maple, and pine may improve the overall eco￾nomics of SS-AD due to the low or zero cost associated with acquir￾ing these feedstocks (Jagadish et al., 1998). Waste paper is paper that either cannot be recycled and/or is used in some other way such as a landfill cover. Paper is treated both chemically and ther￾mally during the paper-making process, effectively serving as a pretreatment and allows the breakdown of cellulose to occur. A number of these lignocellulosic biomass sources have been tested for methane production from L-AD systems (Turick et al., 1991; Gunaseelan, 1997); however, the suitability of lignocellulosic bio￾mass as feedstocks and factors affecting methane production dur￾ing SS-AD have not been studied. Therefore, the objectives of the present study were to determine: (1) the methane yield and volumetric productivity of switchgrass, corn stover, wheat straw, yard waste, leaves, maple, pine, and waste paper in L-AD and SS-AD systems using a single source of inoculum; and (2) the effect of the composition and enzy￾matic digestibility of lignocellulosic biomass on methane yield from SS-AD. 2. Methods 2.1. Feedstock and inoculum Corn stover, wheat straw, and switchgrass were collected in October 2009 from farms operated by the Ohio Agricultural Re￾search and Development Center (OARDC) in Wooster, and Jackson, OH. Fallen tree leaves were collected in October 2010 from the OARDC Wooster campus. Yard waste containing leaves and tree branches was obtained in June 2011 from the OARDC Wooster campus. Pine and maple wood with bark were obtained in June 2011 from the OARDC Wooster campus. The shredded waste paper used was classified as alternative daily cover for landfills and col￾lected from the Solid Waste Management facility of Wayne County, OH. These feedstocks were oven-dried at 40 C for 48 h in a convec￾tion oven (Precision Thelco Model 18, Waltham, MA) to obtain a moisture content of less than 10%, ground with a hammer mill (Mackisik, Parker Ford, PA) to pass through a 5 mm screen, and stored in air tight containers. Effluent from a mesophillic liquid anaerobic digester in Colum￾bus, OH fed with municipal waste water sludge and food waste was used as inoculum. The inoculum was kept in air-tight buckets at 4 C. Prior to use, the inoculum was acclimated at 37 C for 1 week. 2.2. Solid-state anaerobic digestion (SS-AD) Each lignocellulosic biomass feedstock was mixed with the inoculum using a hand-mixer (Black & Decker, 250-watt mixer, Towson, MD, USA) to obtain a feedstock to effluent (F/E) ratio of 3 (based on dry volatile solids) and TS content of 18–19% (Liew et al., 2012). Well mixed materials were loaded into 1-L glass reac￾tors and incubated in a walk-in incubation room for up to 30 days at 37 ± 1 C. Duplicate reactors were run for each condition. Inocu￾lum without any feedstock addition was used as a control. Biogas was collected in a 5-L Tedlar gas bag (CEL Scientific, Santa Fe Springs, CA, USA) attached to the outlet of the reactor. Biogas com￾position and volume were measured every 2–3 days during the 30- day SS-AD. 2.3. Liquid anaerobic digestion (L-AD) Each feedstock was mixed with deionized (DI) water and efflu￾ent to obtain a mixture of 5% TS and an F/E ratio of 0.5 based on dry volatile solids (VS). The mixtures were loaded in 2-L glass jars and incubated in a walk-in incubation room on an Innova model 2300 platform shaker (New Brunswick Scientific; Enfield, CT, USA) at 150 rpm for up to 30 days at 37 ± 1 C. Duplicate reactors were run for each condition. Inoculum without any feedstock addition was used as a control. Biogas generated was collected in a 5-L Ted￾lar gas bag (CEL Scientific, Santa Fe Springs, CA, USA) attached to the outlet of the reactor. Biogas composition and volume were measured every 2–3 days during the 30-day L-AD. 2.4. Enzyme hydrolysis of lignocellulosic feedstocks Cellulase (Spezyme CP) (Lot#4900857805, protein content 82 mg/ml with activity of approximately 50 FPU/ml), and Multifect xylanase (Lot#4901063047, protein content 27 mg/ml with activ￾ity of approximately 25000 OSX/ml), were obtained from Genen￾cor, now DuPont Industrial Biosciences (Palo Alto, CA, USA). All enzymatic hydrolysis experiments were run in duplicate following the National Renewable Energy Laboratory (NREL) Laboratory Ana￾lytical Procedures (LAP) (Selig et al., 2008) at 50 C with 0.05 M cit￾rate buffer, pH 4.8, solid loading of 2%, and cellulase loading of 15 FPU/g solids supplemented with xylanase at a loading of 3750 OSX/ g solids on a rotary shaker (150 rpm for 72 h). The hydrolysate sample was filtered through a 0.2 lm nylon membrane filter for sugar analysis by HPLC as described in Section 2.5. The overall glu￾cose and xylose yields of enzymatic hydrolysis (i.e., enzymatic digestibility) were defined according to Cui et al. (2011). 2.5. Analytical methods The extractive content of feedstock was measured according to the NREL Laboratory Analytical Procedure (Sluiter et al., 2008a). The extractive-free biomass and solid fractions before and after digestion were further fractionated using a two-step hydrolysis method based on the NREL Laboratory Analytical Procedure (Sluit￾er et al., 2008b). Monomeric sugars (glucose, xylose, galactose, arabinose, and mannose) and cellobiose in the acid hydrolysate were measured by an HPLC instrument (Shimadzu LC-20AB, MD, USA) equipped with a Biorad Aminex HPX-87P column and a refractive index detector (RID). A deionized water flow rate of 0.6 ml/min was used as the mobile phase. The temperature of the column and detector were maintained at 80 and 55 C, respec￾tively. The Standard Methods for the Examination of Water and Wastewater were used to analyze the TS and VS contents of feed￾stocks, inoculum, and digestate taken at the beginning and end of the AD process (Eaton et al., 2005). Total carbon and nitrogen con￾tents in feedstock, effluent and digestion materials were deter￾mined using an elemental analyzer (Elementar Vario Max CNS, Elementar Americas, Mt. Laurel, NJ, USA). VFAs and alkalinity were measured using a two-step titration method (McGhee, 1968). Sam￾ples for pH, VFA, and alkalinity measurements were prepared by diluting a 5-g sample with 50 ml of DI water. The mixture was ana￾lyzed using an auto-titrator (Mettler Toledo, DL22 Food & Beverage 380 D. Brown et al. / Bioresource Technology 124 (2012) 379–386

D.Brown et al/Bioresource Technology 124(2012)379-386 381 Analyzer,Columbus,OH,USA).The VFA/alkalinity ratio was calcu- which are easily degradable and can potentially contribute to bio- lated using the empirical formula to determine the risk of acidifi- gas generation(Tong et al.,1990;Zheng et al.,2009). cation,a measure of the process stability (Anderson and Yang. 1992) Biogas volume and composition were determined every 2- 3.2.Biogas production 3 days.The volume of biogas collected in a 5-L Tedlar bag was mea- Fig.1a shows the daily methane yield during the 30-day SS-AD sured with a drum-type gas meter(Ritter,TG 5,Germany)and the composition of biogas(CO2.CH4.N2.and O2)was analyzed using a period.Corn stover and wheat straw had significantly higher peak daily methane yields (~12 L/kg VS)compared to the rest of the GC(Agilent Technologies,HP 6890,DE,USA)equipped with a Ther- feedstocks.Leaves,yard waste,and waste paper had peaks be- mal Conductivity Detector at 200C and a 10-ft stainless steel col- tween 8.7 L/kg VS(switchgrass)and 6.5 L/kg VS(maple)and pine umn 45/60 Molecular Sieve 13X.Helium was used as a carrier gas had a significantly lower peak of 2.5 L/kg VS.Different feedstocks at a flow rate of 5.2 ml/min.The temperature of the column oven peaked at different times.Waste paper peaked at day 4;switch- was initially programmed at 40C for 4 min,then elevated to 60 C at 20C/min and held for 5 min. grass,leaves,yard waste,and maple all peaked at day 6;corn stover peaked at day 8:wheat straw peaked at day 9;and pine Methane yield expressed in L/kg VSfeedstock was calculated as the peaked at day 13.Waste paper had two significant peaks and volume of methane gas produced per kg of VS loaded into the reac- tor at start-up corrected by subtracting the methane yield obtained stopped producing methane by day 13. Fig.1b shows the daily methane yield observed during the 30- from the control reactor with only inoculum then adjusted linearly (Angelidaki et al.,2009).Volumetric methane productivity of ligno- day L-AD period.Waste paper achieved the highest daily methane yields of 19.2 L/kg VS.The rest of the feedstocks had peaks between cellulosic biomass is expressed in Vmethane/Vwork:volume of meth- 11.6 L/kg VS(wheat straw)and 6.3 L/kg VS(yard waste and maple). ane gas produced (Vmethane)per unit working volume of reactor Yard waste peaked at day 4;corn stover,wheat straw,and maple (Vwork). peaked at day 7:waste paper peaked at day 11:pine at day 13: and leaves and switchgrass at day 14.However,these peaks are 2.6.Statistical analysis not as well defined,that is,do not show clear increases in daily methane yield over a couple of days as the peaks in SS-AD,since Data were analyzed using PROC GLM for means and variances a much smaller amount of lignocellulosic biomass was used in L- and statistical significance was determined by analysis of variance AD.The high concentration of lignocellulosic biomass in SS-AD (ANOVA)using SAS software (Version 8.1,SAS Institute Inc.,Cary. may have resulted in the enrichment of hydrolytic and acidogenic NC,USA)with a threshold p-value of 0.05 microbes in early stages of the digestion,and thus led to high VS destruction and,subsequently,a more apparent peak of biogas 3.Results and discussion production than for L-AD. The total methane yield of the lignocellulosic biomass during 3.1.Composition of inoculum and lignocellulosic biomass 30 days of SS-AD and L-AD are presented in Fig.2a.Compared to SS-AD,total methane yields from L-AD were similar with the Table 1 shows the TS,VS,C/N ratio,extractives,lignin,cellulose, exception of waste paper and pine.Methane yields of paper and hemicellulose,pH,and VFA/alkalinity ratio,of the inoculum and pine were significantly lower in the SS-AD compared to the L-AD. lignocellulosic biomass tested in this study.The C/N ratio of the It is speculated that the highly digestible waste paper quickly leaves of 26.1 was considerably lower than of the other materials. caused acidification and inhibition to methanogenesis during SS- The cellulose content was highest in waste paper(60.0%):but low- AD and led to a methane yield of only 15 L/kg VS;however,due est in leaves (11.7%).Switchgrass had the highest hemicellulose to the lower organic loading and higher buffering capacity in L- content,which is the sum of the xylan,galactan,arabinan,and AD,a much higher methane yield of 312.4 L/kg VS was achieved mannan,among the tested feedstocks.The highest lignin content In contrast,pine,a soft wood notorious for its resistance to micro- was found in pine.The paper's lignin content was very low at bial degradation,resulted in a very low biogas yield for SS-AD com- 2.1%owing to the lignin removal during the thermo-chemical pa- pared with that from the other feedstocks,indicating the need for per-making process.Leaves contained the highest amount of pretreatment prior to feeding to SS-AD. extractives.In general,extractives include compounds such as free Volumetric productivities of lignocellulosic biomass in both sugars,oligosaccharides,and organic acids (Chen et al.,2007) L-AD and SS-AD are shown in Fig.2b.Comparison of volumetric Table 1 Characteristics of inoculum and lignocellulosic biomass. Parameters Inoculum Corn stover Switchgrass Wheat straw Leaves Yard waste Maple Waste paper Pine Total solids ( 7.7±0.0 91.6±03 93.0±03 90.0±0.3 93.0±0.1 943±0.1 93.6±0.5 94.2±02 93.0±0.1 Volatile solids 4.1±0.0 89.0±0.1 89.9±0.3 83.4±0.3 86.9±02 91.7±0.2 92.2±0.3 83.9±0.2 90.5±0.3 C:N ratio 3.0±0.0 746±5.1 898±0.1 60.0±00 26.1±00 553±32 1225±18 323.5±12.6 933±2.4 Extractives ( ND 65±0.1 11.9±0.3 17.0±0.1 35.1±1.1 14.7±0.5 5.1±0.6 9.9±1.1 14.2±0.6 Lignin(%泸 ND 17.1±0.3 17.8±0.2 15.2±0.1 22.8±1.2 22.1±0.9 22.0±0.3 2.1±0.1 28.3±0.6 Glucan ( ND 39.2±1.1 32.3±0.5 33.1±0.7 11.7±2.3 27.4±1.3 36.5±0.6 60.0±0.1 26.0±0.4 Xylan ( ND 15.3±0.6 16.7±0.1 15.2±0.3 42±0.5 9.0±0.7 9.2±0.2 8.7±0.5 4.4±0.3 Galactan ( ND 0.0±0.0 0.0±0.0 0.0±0.0 2.1±03 0.9±0.1 0.0±0.0 0.0±0.0 39±0.5 Arabinan ( ND 1.6±0.4 2.0±0.2 15±0.1 18±0.4 1.1±0.1 09±0.2 0.0±0.0 15±0.2 Mannan ( ND 0.0±0.0 0.0±0.0 10±0.1 1.0±0.0 1.2±0.1 12±02 22±0.0 4.4±0.6 pH 7.8±0.0 ND ND ND ND ND ND ND ND VFA/alkalinity ratio 1.00±021 ND ND ND ND ND ND ND ND a Data shown are the average and standard deviation based on duplicate runs:ND=not determined. b based on the total solids of sample. based on the total weight of sample. mg HAceq/mg CaCO3

Analyzer, Columbus, OH, USA). The VFA/alkalinity ratio was calcu￾lated using the empirical formula to determine the risk of acidifi- cation, a measure of the process stability (Anderson and Yang, 1992). Biogas volume and composition were determined every 2– 3 days. The volume of biogas collected in a 5-L Tedlar bag was mea￾sured with a drum-type gas meter (Ritter, TG 5, Germany) and the composition of biogas (CO2, CH4, N2, and O2) was analyzed using a GC (Agilent Technologies, HP 6890, DE, USA) equipped with a Ther￾mal Conductivity Detector at 200 C and a 10-ft stainless steel col￾umn 45/60 Molecular Sieve 13X. Helium was used as a carrier gas at a flow rate of 5.2 ml/min. The temperature of the column oven was initially programmed at 40 C for 4 min, then elevated to 60 C at 20 C/min and held for 5 min. Methane yield expressed in L/kg VSfeedstock was calculated as the volume of methane gas produced per kg of VS loaded into the reac￾tor at start-up corrected by subtracting the methane yield obtained from the control reactor with only inoculum then adjusted linearly (Angelidaki et al., 2009). Volumetric methane productivity of ligno￾cellulosic biomass is expressed in Vmethane/Vwork: volume of meth￾ane gas produced (Vmethane) per unit working volume of reactor (Vwork). 2.6. Statistical analysis Data were analyzed using PROC GLM for means and variances and statistical significance was determined by analysis of variance (ANOVA) using SAS software (Version 8.1, SAS Institute Inc., Cary, NC, USA) with a threshold p-value of 0.05. 3. Results and discussion 3.1. Composition of inoculum and lignocellulosic biomass Table 1 shows the TS, VS, C/N ratio, extractives, lignin, cellulose, hemicellulose, pH, and VFA/alkalinity ratio, of the inoculum and lignocellulosic biomass tested in this study. The C/N ratio of the leaves of 26.1 was considerably lower than of the other materials. The cellulose content was highest in waste paper (60.0%); but low￾est in leaves (11.7%). Switchgrass had the highest hemicellulose content, which is the sum of the xylan, galactan, arabinan, and mannan, among the tested feedstocks. The highest lignin content was found in pine. The paper’s lignin content was very low at 2.1% owing to the lignin removal during the thermo-chemical pa￾per-making process. Leaves contained the highest amount of extractives. In general, extractives include compounds such as free sugars, oligosaccharides, and organic acids (Chen et al., 2007) which are easily degradable and can potentially contribute to bio￾gas generation (Tong et al., 1990; Zheng et al., 2009). 3.2. Biogas production Fig. 1a shows the daily methane yield during the 30-day SS-AD period. Corn stover and wheat straw had significantly higher peak daily methane yields (12 L/kg VS) compared to the rest of the feedstocks. Leaves, yard waste, and waste paper had peaks be￾tween 8.7 L/kg VS (switchgrass) and 6.5 L/kg VS (maple) and pine had a significantly lower peak of 2.5 L/kg VS. Different feedstocks peaked at different times. Waste paper peaked at day 4; switch￾grass, leaves, yard waste, and maple all peaked at day 6; corn stover peaked at day 8; wheat straw peaked at day 9; and pine peaked at day 13. Waste paper had two significant peaks and stopped producing methane by day 13. Fig. 1b shows the daily methane yield observed during the 30- day L-AD period. Waste paper achieved the highest daily methane yields of 19.2 L/kg VS. The rest of the feedstocks had peaks between 11.6 L/kg VS (wheat straw) and 6.3 L/kg VS (yard waste and maple). Yard waste peaked at day 4; corn stover, wheat straw, and maple peaked at day 7; waste paper peaked at day 11; pine at day 13; and leaves and switchgrass at day 14. However, these peaks are not as well defined, that is, do not show clear increases in daily methane yield over a couple of days as the peaks in SS-AD, since a much smaller amount of lignocellulosic biomass was used in L￾AD. The high concentration of lignocellulosic biomass in SS-AD may have resulted in the enrichment of hydrolytic and acidogenic microbes in early stages of the digestion, and thus led to high VS destruction and, subsequently, a more apparent peak of biogas production than for L-AD. The total methane yield of the lignocellulosic biomass during 30 days of SS-AD and L-AD are presented in Fig. 2a. Compared to SS-AD, total methane yields from L-AD were similar with the exception of waste paper and pine. Methane yields of paper and pine were significantly lower in the SS-AD compared to the L-AD. It is speculated that the highly digestible waste paper quickly caused acidification and inhibition to methanogenesis during SS￾AD and led to a methane yield of only 15 L/kg VS; however, due to the lower organic loading and higher buffering capacity in L￾AD, a much higher methane yield of 312.4 L/kg VS was achieved. In contrast, pine, a soft wood notorious for its resistance to micro￾bial degradation, resulted in a very low biogas yield for SS-AD com￾pared with that from the other feedstocks, indicating the need for pretreatment prior to feeding to SS-AD. Volumetric productivities of lignocellulosic biomass in both L-AD and SS-AD are shown in Fig. 2b. Comparison of volumetric Table 1 Characteristics of inoculum and lignocellulosic biomass.a Parameters Inoculum Corn stover Switchgrass Wheat straw Leaves Yard waste Maple Waste paper Pine Total solids (%) 7.7 ± 0.0 91.6 ± 0.3 93.0 ± 0.3 90.0 ± 0.3 93.0 ± 0.1 94.3 ± 0.1 93.6 ± 0.5 94.2 ± 0.2 93.0 ± 0.1 Volatile solids (%)b 4.1 ± 0.0 89.0 ± 0.1 89.9 ± 0.3 83.4 ± 0.3 86.9 ± 0.2 91.7 ± 0.2 92.2 ± 0.3 83.9 ± 0.2 90.5 ± 0.3 C:N ratioc 3.0 ± 0.0 74.6 ± 5.1 89.8 ± 0.1 60.0 ± 0.0 26.1 ± 0.0 55.3 ± 3.2 122.5 ± 1.8 323.5 ± 12.6 93.3 ± 2.4 Extractives (%)b ND 6.5 ± 0.1 11.9 ± 0.3 17.0 ± 0.1 35.1 ± 1.1 14.7 ± 0.5 5.1 ± 0.6 9.9 ± 1.1 14.2 ± 0.6 Lignin (%)b ND 17.1 ± 0.3 17.8 ± 0.2 15.2 ± 0.1 22.8 ± 1.2 22.1 ± 0.9 22.0 ± 0.3 2.1 ± 0.1 28.3 ± 0.6 Glucan (%)b ND 39.2 ± 1.1 32.3 ± 0.5 33.1 ± 0.7 11.7 ± 2.3 27.4 ± 1.3 36.5 ± 0.6 60.0 ± 0.1 26.0 ± 0.4 Xylan (%)b ND 15.3 ± 0.6 16.7 ± 0.1 15.2 ± 0.3 4.2 ± 0.5 9.0 ± 0.7 9.2 ± 0.2 8.7 ± 0.5 4.4 ± 0.3 Galactan (%)b ND 0.0 ± 0.0 0.0 ± 0.0 0.0 ± 0.0 2.1 ± 0.3 0.9 ± 0.1 0.0 ± 0.0 0.0 ± 0.0 3.9 ± 0.5 Arabinan (%)b ND 1.6 ± 0.4 2.0 ± 0.2 1.5 ± 0.1 1.8 ± 0.4 1.1 ± 0.1 0.9 ± 0.2 0.0 ± 0.0 1.5 ± 0.2 Mannan (%)b ND 0.0 ± 0.0 0.0 ± 0.0 1.0 ± 0.1 1.0 ± 0.0 1.2 ± 0.1 1.2 ± 0.2 2.2 ± 0.0 4.4 ± 0.6 pH 7.8 ± 0.0 ND ND ND ND ND ND ND ND VFA/alkalinity ratiod 1.00 ± 0.21 ND ND ND ND ND ND ND ND a Data shown are the average and standard deviation based on duplicate runs; ND = not determined. b based on the total solids of sample. c based on the total weight of sample. d mg HAceq/mg CaCO3. D. Brown et al. / Bioresource Technology 124 (2012) 379–386 381

382 D.Brown et aL/Bioresource Technology 124(2012)379-386 14 -◇-5 witchgrass Corn Stover ((Aea 12 ★-Wheat Straw Leaves 10 --Yard Waste -Waste Paper 8 +Maple -Pine o 30 Time(Days) 25 -Switchgrass ■-Corn Stover t一Wheat Straw Leaves -米-Yard Waste 15 Waste Paper ...Maple 10 Pine 5 0 10 30 Time(Days) Fig.1.Daily methane production during 30-day AD in (a)SS-AD and (b)L-AD. productivities(Lmethane/Lwork)of L-AD to SS-AD showed increases of 11.5,15.1,15.0,and 15.0,respectively.The C/N ratios of SS-AD of 240-730%ofr L-AD while that of paper decreased by 21%.The high- switchgrass,corn stover,wheat straw,yard waste,leaves,waste er volumetric productivities of SS-AD indicate a major advantage paper,pine,and maple were43.0.39.0.36.0.31.5,21.0.64.1, over L-AD as a smaller reactor volume is required for the same sol- 46.2.and 50.3,respectively.Addition of nitrogen in SS-AD may ids loading:however,due to the difficulty in continuous loading eliminate the nutrient limitation placed on the bacteria consortium and unloading of solid feedstock,continuous operation of large but will add cost at commercial scale.This leads to the need for co- scale solid state digesters is challenging.In order to compare the digestion of lignocellulosic biomass at commercial scale with a methane yields of lignocellulosic biomass under both L-AD and nitrogen-rich substrate. SS-AD,this research was conducted at batch mode,but further In the present study,pH and VFA/alkalinity ratios were used as development of SS-AD technologies with continuous or semi-con- indicators of reactor performance (Lossie and Putz,2008).Initial tinuous configurations is favored for commercial systems. and final pH and VFA/alkalinity ratios of both L-AD and SS-AD can be found in Table 2.Initial pH of the lignocellulosic biomass 3.3.Reactor characteristics in SS-AD ranged from 7.1 for leaves to 8.3 for wheat straw.The fi- nal pH values were all considered healthy (above 7)except for Organic loading,C/N ratio,and the strength and amount of inoc- reactors fed with waste paper which had a pH of 6.2,indicating ulum are crucial factors affecting the performance of AD.In case of acidification due to VFA accumulation.Initial VFA/alkalinity ratios an imbalanced digester caused by high organic loading and C/N ra- were below 0.8 except for leaves and the final VFA/alkalinity ratios tio,accumulation of VFA and low pH become inhibitory to metha- were below 0.7 except for waste paper which,when coupled with a nogenesis(Frigon and Guiot,2010). pH of 6.2,indicates souring within the reactor.The L-AD reactors A balanced C/N ratio is crucial for growth of bacteria.The C/N had initial and final pH values in the range of 7.5-7.9 and initial ratios of L-AD of switchgrass,corn stover,wheat straw,yard waste, and final VFA/alkalinity ratios in the range of 0.27-0.60,indicating leaves,waste paper,pine,and maple were 14.3,13.2,13.5,11.8, healthy reactors (Lossie and Puitz,2008)

productivities (Lmethane/Lwork) of L-AD to SS-AD showed increases of 240–730% ofr L-AD while that of paper decreased by 21%. The high￾er volumetric productivities of SS-AD indicate a major advantage over L-AD as a smaller reactor volume is required for the same sol￾ids loading; however, due to the difficulty in continuous loading and unloading of solid feedstock, continuous operation of large scale solid state digesters is challenging. In order to compare the methane yields of lignocellulosic biomass under both L-AD and SS-AD, this research was conducted at batch mode, but further development of SS-AD technologies with continuous or semi-con￾tinuous configurations is favored for commercial systems. 3.3. Reactor characteristics Organic loading, C/N ratio, and the strength and amount of inoc￾ulum are crucial factors affecting the performance of AD. In case of an imbalanced digester caused by high organic loading and C/N ra￾tio, accumulation of VFA and low pH become inhibitory to metha￾nogenesis (Frigon and Guiot, 2010). A balanced C/N ratio is crucial for growth of bacteria. The C/N ratios of L-AD of switchgrass, corn stover, wheat straw, yard waste, leaves, waste paper, pine, and maple were 14.3, 13.2, 13.5, 11.8, 11.5, 15.1, 15.0, and 15.0, respectively. The C/N ratios of SS-AD of switchgrass, corn stover, wheat straw, yard waste, leaves, waste paper, pine, and maple were 43.0, 39.0, 36.0, 31.5, 21.0, 64.1, 46.2, and 50.3, respectively. Addition of nitrogen in SS-AD may eliminate the nutrient limitation placed on the bacteria consortium but will add cost at commercial scale. This leads to the need for co￾digestion of lignocellulosic biomass at commercial scale with a nitrogen-rich substrate. In the present study, pH and VFA/alkalinity ratios were used as indicators of reactor performance (Lossie and Pütz, 2008). Initial and final pH and VFA/alkalinity ratios of both L-AD and SS-AD can be found in Table 2. Initial pH of the lignocellulosic biomass in SS-AD ranged from 7.1 for leaves to 8.3 for wheat straw. The fi- nal pH values were all considered healthy (above 7) except for reactors fed with waste paper which had a pH of 6.2, indicating acidification due to VFA accumulation. Initial VFA/alkalinity ratios were below 0.8 except for leaves and the final VFA/alkalinity ratios were below 0.7 except for waste paper which, when coupled with a pH of 6.2, indicates souring within the reactor. The L-AD reactors had initial and final pH values in the range of 7.5–7.9 and initial and final VFA/alkalinity ratios in the range of 0.27–0.60, indicating healthy reactors (Lossie and Pütz, 2008). Fig. 1. Daily methane production during 30-day AD in (a) SS-AD and (b) L-AD. 382 D. Brown et al. / Bioresource Technology 124 (2012) 379–386

D.Brown et al/Bioresource Technology 124(2012)379-386 383 a 350 口SS-AD 300 ☒Liquid AD 250 200 150 100 50 Corn Stover Wheat Switchgrass Yard Waste Leaves Waste Maple Pine Straw Paper b 16 OSS-AD 14 Liquid AD 10 8 Corn Stover Wheat Switchgrass Yard Waste Leaves Waste Maple Pine Straw Paper Fig.2.(a)Total methane yield and(b)volumetric productivity during 30-day L-AD and SS-AD. Table 2 Changes of pH and VFA/alkalinity ratio during 30-day L-AD and SS-AD. Materials SS-AD L-AD pH VFA/alkalinity ratio pH VFA/alkalinity ratio Initial Final Initial Final Initial Final Initial Final Switchgrass 8.0±0.0 8.3±0.0 0.77±0.09 0.48±0.03 7.8±0.0 7.6±0.1 0.53±0.03 0.31±0.01 Corn stover 8.1±0.2 8.5±0.0 0.63±020 0.48±0.01 7.8±0.0 7.4±0.1 0.60±0.01 0.35±0.00 Wheat straw 8.3±0.0 8.3±02 0.56±0.07 0.36±0.03 7.9±0.0 7.6±0.0 0.43±0.00 0.27±0.02 Yard waste 7.2±0.0 8.4±0.1 072±0.00 051±0.18 7.8±0.1 7.7±0.0 0.33±004 034±000 Leaves 7.1±0.0 8.3±0.0 1.02±0.08 0.70±0.04 7.5±0.1 7.6±0.2 0.53±0.04 0.34±0.01 Maple 7.6±0.1 8.3±0.0 0.68±0.03 0.35±0.06 7.6±0.1 7.7±0.1 0.47±0.08 0.36±0.04 Pine 7.2±0.0 82±0.0 0.76±0.06 0.70±0.12 7.6±0.1 7.6±0.1 0.53±0.06 0.34±0.03 Waste paper 7.9±0.0 6.2±0.0 0.45±0.07 3.24±0.03 7.9±0.0 7.5±0.0 0.50±0.03 0.35±0.00 a Data shown are the average and standard deviation based on duplicate runs. 3.4.First-order kinetic model of methane production As shown in Table 3,there is a linear relationship between log- arithmic methane production and reaction time in both L-AD and First-order kinetic models were used to characterize the meth- SS-AD of switchgrass,corn stover,wheat straw,leaves,yard waste, ane production of lignocellulosic biomass in L-AD (Chynoweth and maple,(r between 0.90 and 0.98)indicating that the methane et al.,1993:Jewell et al.,1993;Angelidaki et al.,2009).The first-or- production could be explained by the simple first-order kinetic der kinetic model can be linearized as In()=kt.where:t=time model.Pine showed a linear relationship in the L-AD(r=0.96) in days:Mu=methane yield obtained in 30 days,L/kg VS feedstock: but not in the SS-AD(r=0.77).This indicates a pattern of methane M=methane yield obtained at time t.L/kg VSreedstock:and production in SS-AD of pine that is not experienced in L-AD.It was M=methane yield potential that remained at time,M=Mu-Mr. observed that pine particles settled to the bottom of the reactor

3.4. First-order kinetic model of methane production First-order kinetic models were used to characterize the meth￾ane production of lignocellulosic biomass in L-AD (Chynoweth et al., 1993; Jewell et al., 1993; Angelidaki et al., 2009). The first-or￾der kinetic model can be linearized as ln Mu M ¼ kt, where: t = time in days; Mu = methane yield obtained in 30 days, L/kg VS feedstock; Mt = methane yield obtained at time t, L/kg VSfeedstock; and M = methane yield potential that remained at time, M = MuMt. As shown in Table 3, there is a linear relationship between log￾arithmic methane production and reaction time in both L-AD and SS-AD of switchgrass, corn stover, wheat straw, leaves, yard waste, and maple, (r 2 between 0.90 and 0.98) indicating that the methane production could be explained by the simple first-order kinetic model. Pine showed a linear relationship in the L-AD (r 2 = 0.96) but not in the SS-AD (r 2 = 0.77). This indicates a pattern of methane production in SS-AD of pine that is not experienced in L-AD. It was observed that pine particles settled to the bottom of the reactor Fig. 2. (a) Total methane yield and (b) volumetric productivity during 30-day L-AD and SS-AD. Table 2 Changes of pH and VFA/alkalinity ratio during 30-day L-AD and SS-AD.a Materials SS-AD L-AD pH VFA/alkalinity ratio pH VFA/alkalinity ratio Initial Final Initial Final Initial Final Initial Final Switchgrass 8.0 ± 0.0 8.3 ± 0.0 0.77 ± 0.09 0.48 ± 0.03 7.8 ± 0.0 7.6 ± 0.1 0.53 ± 0.03 0.31 ± 0.01 Corn stover 8.1 ± 0.2 8.5 ± 0.0 0.63 ± 0.20 0.48 ± 0.01 7.8 ± 0.0 7.4 ± 0.1 0.60 ± 0.01 0.35 ± 0.00 Wheat straw 8.3 ± 0.0 8.3 ± 0.2 0.56 ± 0.07 0.36 ± 0.03 7.9 ± 0.0 7.6 ± 0.0 0.43 ± 0.00 0.27 ± 0.02 Yard waste 7.2 ± 0.0 8.4 ± 0.1 0.72 ± 0.00 0.51 ± 0.18 7.8 ± 0.1 7.7 ± 0.0 0.33 ± 0.04 0.34 ± 0.00 Leaves 7.1 ± 0.0 8.3 ± 0.0 1.02 ± 0.08 0.70 ± 0.04 7.5 ± 0.1 7.6 ± 0.2 0.53 ± 0.04 0.34 ± 0.01 Maple 7.6 ± 0.1 8.3 ± 0.0 0.68 ± 0.03 0.35 ± 0.06 7.6 ± 0.1 7.7 ± 0.1 0.47 ± 0.08 0.36 ± 0.04 Pine 7.2 ± 0.0 8.2 ± 0.0 0.76 ± 0.06 0.70 ± 0.12 7.6 ± 0.1 7.6 ± 0.1 0.53 ± 0.06 0.34 ± 0.03 Waste paper 7.9 ± 0.0 6.2 ± 0.0 0.45 ± 0.07 3.24 ± 0.03 7.9 ± 0.0 7.5 ± 0.0 0.50 ± 0.03 0.35 ± 0.00 a Data shown are the average and standard deviation based on duplicate runs. D. Brown et al. / Bioresource Technology 124 (2012) 379–386 383