Effect of reactor configuration on biogas production from wheat straw hydrolysate

Bioresource Technology 100 (2009) 6317–6323

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Effect of reactor configuration on biogas production from wheat straw hydrolysate Prasad Kaparaju 1, María Serrano, Irini Angelidaki * Department of Environmental Engineering, Technical University of Denmark, Building 115, DK-2800, Kgs. Lyngby, Denmark

a r t i c l e

i n f o

Article history: Received 24 April 2009 Received in revised form 26 June 2009 Accepted 27 June 2009 Available online 31 July 2009 Keywords: Wheat straw Hydrolysate Biogas Bioethanol Lignocellulose

a b s t r a c t The potential of wheat straw hydrolysate for biogas production was investigated in continuous stirred tank reactor (CSTR) and up-flow anaerobic sludge bed (UASB) reactors. The hydrolysate originated as a side stream from a pilot plant pretreating wheat straw hydrothermally (195 °C for 10–12 min) for producing 2nd generation bioethanol [Kaparaju, P., Serrano, M., Thomsen, A.B., Kongjan, P., Angelidaki, I., 2009. Bioethanol, biohydrogen and biogas production from wheat straw in a biorefinery concept. Bioresource Technology 100 (9), 2562–2568]. Results from batch assays showed that hydrolysate had a methane potential of 384 ml/g-volatile solids (VS)added. Process performance in CTSR and UASB reactors was investigated by varying hydrolysate concentration and/or organic loading rate (OLR). In CSTR, methane yields increased with increase in hydrolysate concentration and maximum yield of 297 ml/g-COD was obtained at an OLR of 1.9 g-COD/l d and 100% (v/v) hydrolysate. On the other hand, process performance and methane yields in UASB were affected by OLR and/or substrate concentration. Maximum methane yields of 267 ml/g-COD (COD removal of 72%) was obtained in UASB reactor when operated at an OLR of 2.8 g-COD/l d but with only 10% (v/v) hydrolysate. However, co-digestion of hydrolysate with pig manure (1:3 v/v ratio) improved the process performance and resulted in methane yield of 219 ml/g-COD (COD removal of 72%). Thus, anaerobic digestion of hydrolysate for biogas production was feasible in both CSTR and UASB reactor types. However, biogas process was affected by the reactor type and operating conditions. Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction Biomass currently provides around 45 ± 10 EJ per year (9–13%) of the 467 EJ global energy supply (IEA Bioenergy, 2007). Biomass is a versatile raw material that can substitute fossil fuels for production of heat, power, transport fuels, and other economically important bioproducts. When produced and used on a sustainable basis, biomass will not only decouple the food and biofuel production and reduced greenhouse gas emissions from fossil fuels but also ensure a more stable and secured energy supply especially in transport sector. In many countries, including Denmark, wheat straw is an abundant agricultural residue and can also serve as a low cost attractive feedstock for production of 2nd generation bioethanol. Wheat straw contains 35–45% cellulose, 20–30% hemicellulose and 8–15% lignin (Sun et al., 1996). Pretreatment of wheat straw is necessary to break down the lignocellulose into the three major polymeric constituents: cellulose, hemicellulose and lignin (Gray et al., 2006). Different pretreatment methods, such as dilute acid

* Corresponding author. Tel.: +45 45251429; fax: +45 45932850. E-mail address: [email protected] (I. Angelidaki). 1 Present address: Department of Biological and Environmental Science, University of Jyvaskyla, P.O. Box 35, FI 40014, Jyvaskyla, Finland. 0960-8524/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2009.06.101

pretreatment, steam explosion, hydrothermal pretreatment and wet oxidation have been proposed for the fractionation of lignocellulose (Olofsson et al., 2008). However, hydrothermal pretreatment is considered as potential solution for pretreatment of large quantities of straw on commercial scale as this treatment is performed without addition of chemicals or oxidizing agents (Larsen et al., 2008). Depending on the process conditions, most of the hemicellulose and degradation products will be dissolved in the hydrolysate (liquid) while the cellulose and lignin will be recovered in the solid fraction (Thomsen et al., 2008). Fermentation of cellulose (C-6) for bioethanol production is a well-established process and efficiently carried out by Baker’s yeast Saccharomyces cerevisiae (e.g. Gray et al., 2006). However, yeast or the high ethanol yielding bacterium Zymomonas mobilis cannot ferment multiple sugar substrates e.g. xylose and arabinose to ethanol (Olofsson et al., 2008). On the other hand, utilization of both cellulose and hemicellulosic sugars present in typical lignocellulosic biomass hydrolysate is essential for the economical production of ethanol (Hinman et al., 1989). Several organisms have been recently engineered to ferment all the sugars from hemicellulose (C-5) to ethanol with high yields and stability (Chu and Lee, 2007). For instance, a recombinant Escherichia coli (strain FBR5) has been developed to ferment mixed multiple sugars to ethanol (Dien et al., 2000). Alternatively, hemicellulose can also be utilized

6318

P. Kaparaju et al. / Bioresource Technology 100 (2009) 6317–6323

for production of hydrogen through dark fermentation (Kádár et al., 2004; Kongjan et al., 2008) and was recognized as the most environmental friendly and feasible process (Hawkes et al., 2007). However, the hydrogen yield obtained by dark fermentation is relatively low (amounting of approximately 20–30% of the energy content in the organic matter) while the remaining organic mater is converted to volatile fatty acids and/or lost with the effluent unutilized (Kongjan et al., 2008). Therefore, the most sustainable solution would be to use wheat straw hydrolysate for biogas production through anaerobic digestion, a process that has the possibility to utilize all types of organic molecules. The produced biogas can be used for heat, electricity and/or can be upgraded to methane for application as vehicle fuel. Biogas reactors of various reactor configurations such as continuous stirred tank reactor (CSTR), anaerobic contact reactor, fluidized bed reactor, anaerobic fixed film reactor and up-flow anaerobic sludge blanket (UASB) reactor have been developed for treating high strength wastes (Rao and Bapat, 2006). Among these configurations, CSTR and UASB reactors are the most extensively used due to several advantages (Wilkie et al., 2004). Conventional CSTR is simple to operate but less efficient in terms of the effluent quality compared to other reactor configurations such as UASB reactor (Azbar et al., 2001). The limitations of CSTR are the need to retain bacteria within the reactor that catalyze the process (Boe, 2006) and loss of undegraded particulate matter due to short-circuit (Kaparaju et al., 2008). On the other hand, UASB systems are commonly used for treating complex wastewaters including the treatment of lignocellulosic wastewater from forest industry (Torry-Smith et al., 2003). UASB reactor in general has several advantages over the CSTR as the former reactor type can be operated at a much shorter hydraulic retention times (HRT) thereby allowing for a faster treatment and reducing storage costs. Use of C-6 fraction of hydrothermally pretreated wheat straw for bioethanol (Larsen et al., 2008; Thomsen et al., 2008; Kaparaju et al., 2009) or for biogas production (Gray et al., 2006; Kaparaju et al., 2009; Linde et al., 2007) has been demonstrated successfully. However, little or no information is available in the literature on anaerobic digestion of wheat straw hydrolysates (C-5 sugars) for biogas production. In the present study, the feasibility of using wheat straw hydrolysate for biogas production was investigated. Firstly, methane potential of the wheat straw hydrolysate was

determined in batch assays. Process performance during semi-continuous digestion of hydrolysate alone or co-digestion with manure in CSTR and UASB reactors at varying hydrolysate concentration and/or organic loading rate (OLR) was investigated.

2. Methods 2.1. Substrates 2.1.1. Wheat straw hydrolysate Wheat straw hydrolysate was obtained from Risø National Laboratories, Technical University of Denmark (Denmark). Hydrothermal pretreatment of wheat straw was carried out in a pilot plant (100 kg/h capacity) as described elsewhere (Thomsen et al., 2008). Briefly, wheat straw at the rate of 120–150 kg-DM/h was fed in a counter-current with a water flow rate of 400–600 l/h through three reactors. In the first step, being a pre-soaking step, reactor was operated at temperature of 80 °C and residence time of 6 min. The pre-soaked wheat straw was then heated in stage two at approximately 180 °C for 15 min followed by at 190 °C for 3 min in stage three. Pretreated biomass was then decanted into a liquid fraction called hydrolysate, containing mainly hemicelluloses, monomeric sugars, salts and degradation products and a solid fibre fraction rich in cellulose and lignin. Cellulose and hemicellulose recovery after pretreatment and hydrolysis yield were outlined by Thomsen et al. (2006). The produced wheat straw hydrolysate was stored at +4 °C until further use. The characteristics of the wheat straw hydrolysate are shown in Table 1. 2.1.2. Manures Fresh cow manure collected from a centralized biogas plant (Hashøj, Denmark) was used in CSTR experiments. The manure was filtered (10 mm) in order to avoid clogging of feed tubes. On the other hand, pig manure collected from Hegndal biogas plant (Hemmet, Denmark) was used for UASB experiments as described elsewhere (Karakashev et al., 2008). It was a mixture of pig manure (90% w/w) and fish-processing industrial waste (10% w/w). Filtered pig manure was diluted with water (1:1) on volume basis. The characteristics of the prepared feeds are presented in Table 1.

Table 1 Characterization of wheat straw hydrolysate and manures used in the reactor experiments (standard deviations on triplicate samples). Parameter

Raw wheat straw

Hydrolysate

Filtered cow manure

Filtered and diluted pig manure

pH TS (%) VS (%) Ash content (%) SS (g/L) VSS (mg/L) COD (g/L) SCOD (g/L) VFA (g/L) Acetate (g/L) Total nitrogen (g/L) Ammonia (g/L) Proteins (g/L) Lipids (g/L) Carbohydrates (g/L) Furfurals (g/L) HMF (g/L) Phenols (g/L) Arabinose (g/L) Xylose (g/L) Glucose (g/L)

6.2 (1:40) ± 0.2 91.6 ± 0.02 87.5 ± 0.02 4.1 ± 0.02 N.D. N.D. N.D. N.D. 0.13 ± 0.02 N.D. 1.3 ± 0.04 0.31 ± 0.01 6.5 ± 0.17 1.5± 853.1 N.D. N.D. N.D. 2.6 ± 0.19a 21.3 ± 0.30a 35.9 ± 0.03a

4.9 ± 0.1 4.4 ± 0.01 3.4 ± 0.01 1 ± 0.01 0.16 ± 0.05 0.32 ± 0.1 38.0 ± 1.31 32.1 ± 2.22 0.70 ± 0.14 0.34 ± 0.03 0.20 ± 0.01 0.03 ± 0.01 1.1 ± 0.03 0.24 30.5 ± 1.8 0.25 ± 0.04 0.14 ± 0.02 0.14 1.3 9.3 2.9

7.5 ± 0.3 4.5 ± 0.4 3.3 ± 0.2 1.2 ± 0.2 N.D. N.D. N.D. N.D. 2.0 ± 0.3 1.5 ± 0.3 N.D. 0.7 ± 0.1 N.D. N.D. N.D. N.D. N.D. N.D. N.D. N.D. N.D.

7.6 ± 0.4 21.0 ± 0.9 16 ± 1.0 N.D. N.D. N.D. 9.0 ± 1.5 7.5 ± 0.6 0.9 ± 0.2 0.61 ± 0.03 4.2 ± 0.08 3.5 ± 0.4 N.D. N.D. N.D. N.D. N.D. N.D. N.D. N.D. N.D.

N.D. non determined. Protein = 6.25  (TKN–NHþ 4 –N); carbohydrate = VS – protein – lipids – VFA.

6319

P. Kaparaju et al. / Bioresource Technology 100 (2009) 6317–6323

2.1.3. Inocula Two different inocula were used in the present study. For CSTR and methane potential experiments, thermophilically digested cow manure from a pilot-scale plant treating cow manure was used (Kaparaju et al., 2008). For UASB reactor experiments, thermophilic granules from a potato-processing wastewater treatment plant (Kruiningen, The Netherlands) were used as inoculum (organic matter content of 0.08 g/g sludge). In addition, basic anaerobic medium containing vitamins, carbonate buffer, phosphate, ammonium and trace metals as described by Angelidaki et al. (1990) was used as synthetic media in UASB reactor. 2.2. Biological methane potential Biological methane potential assays were performed according to Angelidaki et al. (2009). The experiments were performed in 118 ml serum glass bottles with working volume of 40 ml. To each bottle, 30 ml of inoculum and 10 ml of substrate at a substrate concentration of 4.25 and 8.5 g VS/l was added. The headspace in the bottle was flushed with pure N2 for 3–5 min, before adding 2–3 drops of sodium sulphide to ensure anaerobic conditions. The prepared bottles were then sealed with rubber stoppers and aluminium crimps, and incubated statically at 55 °C. The experiment was conducted in triplicates. Assays with digested manure alone were used as controls. Methane produced from inoculum was subtracted from the sample assays. 2.3. Reactor experiments set-up The effect of reactor configuration on biogas production from wheat straw hydrolysate was evaluated in CSTR and USAB reactors.

Table 2 Main parameters and methane yields during semi-continuous anaerobic digestion of wheat straw hydrolysate in CSTR reactor at 55 °C. Steps

Days of operation Feed ratio v/v (hydrolysate:manure) HRT (days) Working volume (l) OLR (g VS/l d) Biogas prod. (ml/mlfeed) Methane yield (ml/g VSfed) Relative yield (%) Methane content (%) VFA (g/l) pH

Start-up

1

2

3

0–22 0:100 20 3.5 1.65 10.0 207 – 67.5 0.71 8.2

23–51 20:80 20 3.5 1.40 10.2 226 9.6 62.3 0.78 8.2

52–82 50:50 20 3.5 1.38 10.0 242 17.3 66.6 0.92 8.1

83–102 100:0 20 3.5 1.35 19.1 399 93.6 58.9 0.93 7.4

2.3.1. CSTR experiments A stainless steel CSTR (5 l) with a working volume of 3.5 l was operated semi-continuously with HRT of 20 d. Temperature in the reactor was maintained at 55 °C by using a self-limiting heating coil (maximum temperature 80 °C) wrapped around the reactor wall. Reactor was fed (175 ml) twice a day using peristaltic pump. The produced biogas was registered using gas meter. Reactor was stirred by mechanical mixer operated on a cycle of 15 min on/off. During start-up, reactor was filled with 3.4 l of inoculum and 100 ml of raw cow manure. As the inoculum originated from a biogas plant treating cow manure, the semi-continuous feeding was initiated with cow manure alone (days 0–22). Data obtained during this period was used as reference period (control). Hydrolysate was introduced in a step-wise manner. On day 23, co-digestion of the hydrolysate with cow manure was initiated at a feed ratio of 20:80 (v/v) between days 23 and 51 and at feed ratio of 50:50 (v/v) between days 52 and 82. During the last period (days 83– 102), hydrolysate alone was used (hydrolysate to manure ratio of 100:0 v/v). During each run, a minimum of 1 HRT was maintained and OLR was changed only when steady-state condition, characterized by a constant gas production rate (±5%), was achieved. Operating conditions are shown in Table 2. 2.3.2. UASB experiments A lab-scale glass UASB reactor (334 ml) was operated with a working volume of 255 ml and HRT of 48 h. Reactor temperature was maintained at 55 °C by circulating hot water in the heat jacket. The prepared feed, stored at 10 °C, was fed semi-continuously at a feed rate of 128 ml/d. Reactor was fed 24 times per day at a feeding rate of 5.3 ml per time. The produced biogas was measured by gas meters. During the start-up, the reactor was inoculated with 100 ml of anaerobic granular sludge. During the initial run, filtered and diluted pig manure was fed at an OLR of 2.26 g-COD/l d and HRT of 48 h. Data collected during this period were used as control (days 0–18). After reaching the steady-state, hydrolysate was carefully introduced. The feed contained wheat straw hydrolysate diluted with water (1:9 v/v) and basic anaerobic medium (1:1 v/v). OLR during this period was gradually increased in a step-wise manner from the initial 2.41 to final 11.9 g-COD/l d (days 19–71). When methanogenesis was inhibited, pH dropped significantly and the process was recovered by adding sodium bicarbonate at 5 g/l reactor volume. As an attempt to reverse toxicity caused by hydrolysate, co-digestion of hydrolysate with filtered pig manure was carried out at an OLR of 8.5–14.8 g-COD/l d and HRT of 48 h during days 72–90. Similar to CTSR, each OLR was operated until a steadystate condition, characterized by a constant gas production rate (+5%), was noticed. Operating conditions are shown in Table 3.

Table 3 Operating conditions and process performance during semi-continuous anaerobic digestion of wheat straw hydrolysate in UASB reactor at 55 °C. Steps Start-up

1

2

Substrate

Filtered and diluted pig manure

Hydrolysate

Days Hydrolysate concentration (% (v/v)) HRT (h) OLR (g COD/l d) SCOD removal (% (w/w)) Biogas production (ml/mlfeed) Methane yield (ml-CH4/g COD) Methane content (%) VFA (g/l) pH

0–18 – 48 2.26 51.6 2.1 88.6 60.6 0.22 8.2

19–43 5% 48 2.41 54.4 1.5 171.6 66.5 0.03 7.0

44–50 10% 48 2.80 71.2 2.7 267.2 66.9 0.04 7.2

3

4

5

6

51–57 25% 48 6.97 78.7 5.2 174.1 57.3 0.04 7.1

58–68 40% 48 9.80 59.8 7.3 167.8 54.6 0.16 7.1

69–71 50% 48 11.99 48.7 6.8 101.4 42.0 0.78 5.2

74–79 50% 48 14.82 21.4 6.3 69.2 42.0 1.25 7.3

7

Codigestion (hydrolysate–pig manure) 80–90 25% 48 8.53 71.2 6.3 219.2 65.0 0.3 7.9

P. Kaparaju et al. / Bioresource Technology 100 (2009) 6317–6323

3.1. Substrate characteristics The composition of the wheat straw hydrolysate is presented in Table 1. Hydrolysate had TS and VS content of 4.4% and 3.4%, respectively, with a pH of 4.9. The low pH of hydrolysate was probably due to the production of organic acids e.g. acetic acid produced during hydrothermal pretreatment as a result of cleavage of acetyl groups contained on the side chains of hemicelluloses. This rather low pH was of concern as the optimum pH for methanogenesis is between 6.5 and 8 (Nielsen, 2006). Nevertheless, acetate, the main VFA, and ammonia levels were less than 1 and 2 g/l, respectively and were below the concentrations reported to cause any process inhibition (Sung and Liu, 2002). Analyses also revealed that xylose was the main sugar and accounted for 72.9% of total sugars. Glucose and arabinose were also found but at a very low concentration (Table 1). The low glucose recovery in hydrolysate was probably due to its crystalline and thermally stable structure (Klinke et al., 2002). Hydrothermal pretreatment of wheat straw also resulted in the production of degradation products such as phenolic compounds, 5-HMF and furfural in the hydrolysate (Table 1). 5-HMF and furfural are derived from the degradation of glucose and xylose respectively, whereas, phenolic compounds are produced due to partial degradation of lignin (Mussatto et al., 2005). The relatively low concentration of 5-HMF (0.14 g/l), furfural (0.25 g/l) and phenolic compounds (0.14 g/l) in hydrolysate, considered to be inhibitory to most microorganisms during hydrolysate fermentation (Thomsen et al., 2006), was probably due to their highly volatile nature and thus may not have been captured well in hydrolysate (Kaparaju et al., 2009). The main phenolic compounds identified were vanillin, 2-furoic acid, coumaric acid and ferulic acid and occurred in a concentration of 0.101 g/l. However, the concentration of 2furoic acid, the predominant phenolic compound, was less than the level expected to cause inhibition to methanogenis (data not shown). Previous researchers have demonstrated that compounds

Anaerobic biodegradability of hydrolysate was evaluated in batch digestions and terminated on day 63 when no more methane was evolved. Methane production started immediately in all assays and maximum methane production of 384 ml/g VSadded was obtained after 50 d of incubation (data not shown). Results also showed no significant difference in methane yields at substrate concentrations of 4.25 or 8.5 g VS/l. The obtained experimental methane yield, which was 83.7% of theoretical methane yield of 459 ml/g VSadded (calculated based on the COD content), indicates that a small part of the organic matter was recalcitrant. 3.3. Reactor experiments 3.3.1. CSTR experiments The performance of the reactor at steady-state is presented in Table 2 and Fig. 1. After the initial start-up, reactor was fed with cow manure alone (control) for 22 d. Mean methane production during anaerobic digestion of cow manure alone at an OLR of 1.65 g VS/l d was 207 ml/g VSfed. Co-digestion of hydrolysate with cow manure at an OLR of 1.4 g VS/l d and at a feed ratio (v/v basis) of 20/80 (days 23–51) or 50/50 (days 52–82) resulted in methane yields of 226 and 242 ml/g VSfed, respectively. The increase in methane yields during co-digestion were 9.6% and 17.3%, respectively, compared to cow manure alone, although the biogas production per ml feed was more or less the same during days 0–82 (around 10 ml/ml-feed). It should be noted that the organic content in the hydrolysate was only 3.4%, while the organic load in manure was 16%. The methane content in the biogas was affected by the concentration of hydrolysate in the feed. The methane percentage varied from a high 67.5% during control period to 62.3–66.6% during codigestion and then decreased to 59% when reactor was fed with 100% hydrolysate. The lower methane content during co-digestion or 100% hydrolysate is attributed to the low pH caused by the use of acidic hydrolysate (Table 2). Low pH could have resulted in removal of a small part of the produced CO2 as bicarbonate through the effluent and not as CO2 together with the methane. The pH during cow manure digestion was approx. 8.2, while the pH during digestion of hydrolysate alone was 7.4. 800

2.0 Manure

600

ManureHydrolysate (80-20%)

ManureHydrolysate (50-50%)

Hydrolysate (100%)

1.5

400

1.0

200

0.5

0

0.0

4

9.0

3

8.5

2

8.0

1

7.5

0 0

10

20 30

40 50

60 70

OLR (g-VS/l.d)

3. Results and discussion

3.2. Biological methane potential

pH

Total solids (TS), volatile solids (VS) and suspended solids were determined according to the Standard Methods (APHA, 1998). Volatile fatty acids (VFA) concentrations and methane content in biogas were determined by gas chromatograph with flame ionization detector. Ammonia and total Kjeldahl nitrogen (TKN) were analysed by Kjeldahl method (APHA, 1998). Chemical oxygen demand, total (TCOD) and soluble (SCOD) were analysed according to the Danish Standards (Danish Standards Association, 1991). SCOD samples were filtered through glass fibre filter paper (U90 mm, GF50, Schleicher and Schuell). To determine the sugars (glucose, xylose and arabinose) content in raw and pretreated straw and liquid fractions, strong acid (72% w/w H2SO4) hydrolysis of solid fraction and weak acid (4% w/w H2SO4) of liquid fraction was applied (Thomsen et al., 2006). Sugars were quantified on high performance liquid chromatography (HPLC) system HP 1100 (Agilent 1100) equipped with a BioRad Aminex HPX-87 H at 63 °C and a refractive index (RI) detector (RID 1362A) using 0.6 ml/min of 4 mM H2SO4 as eluent. The analysis detection limits for glucose, xylose and arabinose were 0.011, 0.002 and 0.014 g/l, respectively. Klason lignin content was determined as the weight of the filter cake subtracted the ash content. Furfural and 5-hydroxymethyl-2furaldehyde (5-HMF) were quantified using a HPLC as described elsewhere (Klinke et al., 2001). Lipid extraction was carried out through Soxhlet Method (AOCS, 1997).

such as 2-furoic acid may be utilized as a sole substrate by the methanogenic consortia (Torry-Smith et al., 2003).

Methane yield (ml/gVS fed )

2.4. Analytical methods

VFA (g/l)

6320

7.0 80 90 100 110

Time (days) Fig. 1. Process performance during semi-continuous anaerobic digestion of wheat straw hydrolysate in CSTR reactor at 55 °C: methane yield (s), organic loading rate (-) VFA (h) and pH (d).

P. Kaparaju et al. / Bioresource Technology 100 (2009) 6317–6323

The VFA levels reached from the initial 0.71 g/l noticed during control period to 0.78–0.92 g/l during co-digestion to 0.93 g/l when 100% hydrolysate was used. Acetate was the main VFA (80% of the total VFA) but was well below the levels reported to induce process inhibition (Larsson et al., 1999). The increase in VFA levels (days 83–102) with a corresponding decrease in pH suggests stabilisation of the process and possible adaptation to inhibitor compounds such as furfural and 5-HMF, and phenolic compounds or the ability of the anaerobic bacteria to utilize the hydrolysate as substrate for methanogenesis. This also explains the reason for the sharp increase in methane yields when fed with 100% hydrolysate (Fig. 1). These results are in accord to Torry-Smith et al. (2003) who also reported that compounds such as furfural and 5-HMF, and phenolic compounds are known to have inhibitory effect on biogas process but can be utilized as substrate for methanogenic bacteria. 3.3.2. UASB experiments The performance of the reactor at steady-state is presented in Table 3 and Fig. 2. The mean methane production during anaerobic digestion of pig manure alone (control) was 136 ml-CH4/g-COD with a SCOD removal of 51.6% (days 0–18). The low methane production from pig manure alone was obviously due to the use of filtered and diluted pig manure as feed. Results also showed that process performance, biogas production and methane yields were affected by the OLR and concentration of hydrolysate in the feed. In general, an increase in hydrolysate concentration from 5% to 50% v/v resulted in a decrease in methane yields. Maximum methane yields of 267 mlCH4/g-COD was noticed when reactor was operated at an OLR of 2.8 g COD/l d and with hydrolysate concentration of 10% (days 44–50). However, an increase in OLR to 12 g COD/l d (50% v/v hydrolysate) resulted in complete process failure (day 71). Upon recovering the process (days 71–73), co-digestion of hydrolysate with pig manure was carried out at 50:50 v/v ratio (days 80–90). Results showed that co-digestion improved the methane yield (219 ml-CH4/g-COD) but only at a maximum OLR of 8.53 g COD/ l d and hydrolysate concentration of 25% (v/v). The methane yield per removed COD achieved was in agreement with the theoretical methane yield of approx. 400 ml/g-COD, assuming that the measurement temperature was around 35–40 °C (350 ml/g-COD at 0 °C). Thus, the maximum experimental methane yield of 297 ml/g CODadded at a COD removal efficiency of 75% would correspond to a methane yield of 396 ml/g-CODremoved which is

Control

Hydrolysate 5%

10% 25%

Codigestion 40%

50% 25%

10 200

5

0

0

3

9

2

7

1

5

pH

Methane Yield (mL/g-COD)

VFA (g/L)

15

50%

400

OLR (g-COD/L.d)

20

600

3

0 0

10

20

30

40

50

60

70

80

90

100

Time (days) Fig. 2. Process performance during semi-continuous anaerobic digestion of wheat straw hydrolysate in UASB reactor at 55 °C: methane yield (s), organic loading rate (-) VFA (h) and pH (d).

6321

in agreement with the calculated theoretical yield of approx. 400 ml/g CODremoved. The decrease in methane yield without affecting the biogas production upon increase in OLR from 2.8 to 6.97 g-COD/l d or hydrolysate concentration from 10% to 25% (v/v) was probably due to production of other gasses. Torry-Smith et al. (2003) also reported that degradation of some phenolic compounds might not be mineralized to methane and carbon dioxide but could be transformed to other intermediates. In fact, hydrogen was detected at significant concentrations when hydrolysate concentration in the present study was increased to 40% or 50% v/v (data not shown). This was evident from VFA build-up and reduction in pH (Fig. 2). The inhibition caused at high hydrolysate concentration was probably due to presence of low molecular weight lignin compounds (Sierra-Álvarez et al., 1994). However, analyses of effluent revealed that furfurals and 5-HMF were below the detection limit (data not shown) suggesting that these compounds were degraded during the anaerobic process. The results from the present study in practice suggest that reactor configuration has a profound influence on the anaerobic digestion of wheat straw hydrolysate. The process was found to be more efficient in CSTR than in UASB as the methane yields obtained in the former reactor type (297 ml/g-COD) were slightly higher than the latter reactor type (267 ml/g-COD). Moreover, the process was less stable and showed to be inhibited in UASB reactor when OLR was >2.8 g COD/l d or hydrolysate concentration >10% (v/v) in the feed. Both these results confirms that CSTR reactor configuration is more simple to operate but less efficient than UASB reactor. Furthermore, the long HRT in CSTR might have helped the bacteria to adapt to the inhibitors and thus to avoid inhibition when hydrolysate concentration was increased up to 100%. On the other hand, the OLR in UASB reactor (OLR of 2.8 g COD/l d) was much higher than for the CSTR reactor (OLR of 1.65 g COD/ l d), which can be the reason for the earlier failure of UASB reactor. Previous studies using bioethanol effluents from wet oxidized wheat straw hydrolysate showed slight initial inhibition in UASB reactors due to phenolic compounds of low molecular weight (Torry-Smith et al., 2003). However, inhibition noticed in the present study was stronger than that reported by Torry-Smith et al. (2003). The probable reason for this difference is that higher amount of inhibitors are produced during hydrothermal pretreatment compared to that of wet oxidation pretreatment (Klinke et al., 2001). For instance, the concentrations of phenolic compounds in the present study were 0.14 g/l and considered to be much higher than the concentration of 0.02–0.07 g/l reported by Klinke et al. (2001) during wet oxidation of wheat straw hydrolysate. Furfural and 5-HMF, and phenolic compounds are known to have inhibitory effect on ethanol fermentation (Thomsen et al., 2006) and biogas production (Torry-Smith et al., 2003). The concentration of these inhibitory compounds is however dependent on the substrate used, pretreatment conditions and type of biofuel produced. Thus, in order to optimize biogas production from UASB reactor, retention time should be increased or substrate should be diluted to reduce the possible inhibition effect. Alternatively, use of inoculum that has been adapted to degrade phenols and furfurals or other inhibitory compounds present in the hydrolysate should be sought for. Several detoxification methods like neutralization, over liming with calcium hydroxide (Carvalheiro et al., 2005) and enzymatic detoxification using laccase (Jönsson et al., 1998) have been reported earlier. Based on the data from FAO, 4.5 M t of wheat were produced in the Denmark in 2007 (FAO, 2008). Calculating at an average of 1.3– 1.4 kg wheat straw per kg of wheat grain produced (Montane et al., 1998), approximately 6.3 M t of wheat straw was produced in Denmark. The produced wheat straw if burned in a power plant would produce approximately 105  106 MJ of useful energy (fuel

6322

P. Kaparaju et al. / Bioresource Technology 100 (2009) 6317–6323

value per kg of wheat straw is 19.1  106 J). Alternatively, if the produced wheat straw is used for production of 2nd generation bioethanol, then approx. 756  106 l of bioethanol could be produced through simultaneous saccharification and fermentation (SSF) of C-6 sugars. The produced bioethanol could replace 540  106 l of gasoline when used as E85 biofuel for a midsize passenger vehicle. On the other hand, inclusion of a xylose fermenting microorganism in the bioethanol process could increase bioethanol production up to 5% (794  106 l). In addition, the lignin, which is not affected in the SSF process, can be separated and used as a solid biofuel through combustion in a boiler. The produced energy could be used either in the pretreatment of wheat straw or internally for energy requirements of the biorefinery. Despite low energy output from ethanol obtained during SSF of C-6 sugars, production of bioethanol has several advantages over biogas. For instance, bioethanol, being a liquid biofuel, can readily be integrated into existing fuel infrastructure and directly substitute fossil fuel in the transport sector. However, bioethanol is only utilizing a narrow spectrum of substrates (efficiently only C-6 sugars), while other molecules, such as C-5 sugars, organic acids, proteins etc. are not utilized fully. Therefore, in order to improve economic feasibility of the bioethanol process, efficient methods for utilization of the rest organic matter should be used/developed. Anaerobic digestion is a very efficient process, which is able to utilize versatile types of organic molecules. Thus, combination of biogas and bioethanol production technologies has a tremendous potential for full recovery of energy potential in biomass. 4. Conclusions Results from the present study showed that anaerobic digestion of wheat straw hydrolysate as a sole substrate was feasible in CSTR and UASB reactors. With CSTR, maximum methane yield of 297 ml CH4/g-COD could be obtained at an OLR of 1.9 g-COD/l d and HRT of 20 d. On the other hand, the same process in UASB reactor was less efficient at a substrate concentration of >25% (v/v) and could completely inhibit the process at hydrolysate concentrate of >50% (v/v) and/or OLR of 12 g-COD/l d. Co-digestion of hydrolysate with pig manure could improve methane yields in UASB. However, maximum concentration of hydrolysate should not exceed 25% (v/ v). Thus, anaerobic digestion of hydrolysate for biogas production would improve the overall energy and economy of biorefinery concept. Acknowledgements Authors thank Héctor García, Jens Shaarup Sørensen and Asst. Prof. Markus Emili from Technical University of Denmark for setting up the experiments and furfural analyses. Tomas Fernqvist and Ingelis Larsen from Risø National Laboratories (Denmark) are greatly acknowledged for the help rendered during sugars and phenols analyses. The study was supported by the Danish Innovation and Research Council, DSF project no. 2104-06-0004. References Angelidaki, I., Peterden, S.P., Ahring, B.K., 1990. Effects of lipids on thermophilic anaerobic digestion and reduction of lipid inhibition upon addition of bentonite. Applied Microbiology and Biotechnology 33, 469–472. Angelidaki, I., Alves, M., Bolzonella, D., Borzacconi, L., Campos, J.L., Guwy, A.J., Kalyuzhnyi, S., Jenicek, P., van Lier, J.B., 2009. Defining the biomethane potential (BMP) of solid organic wastes and energy crops: a proposed protocol for batch assays. Water Science and Technology 59, 927–934. American Oil Chemists’ Society (AOCS), 1997. Official Methods and Recommended Practices. APHA, 1998. Standard Methods for the Examination of Water and Wastewater, 20th ed. American Public Health Association, Washington, DC.

Azbar, R., Ursillo, P., Speece, R.E., 2001. Effect of process configuration and substrate complexity on the performance of anaerobic processes. Water Research 35 (3), 817–829. Boe, K., 2006. Online Monitoring and Control of the Biogas Process. Institute of Environment and Resources. Technical University of Denmark (DTU). Ph.D. Thesis. Carvalheiro, F., Duarte, L.C., Lopes, S., Pajaró, J.C., Pereira, H., Gírio, F.M., 2005. Evaluation of the detoxification of brewery’s spent grain hydrolysate for xylitol production by Debaryomyces hansenii CCMI 941. Process Biochemistry 40, 1215–1223. Chu, B., Lee, H., 2007. Genetic improvement of Saccharomyces cerevisiae for xylose fermentation. Biotechnology Advances 25 (5), 425–441. Danish Standard Association, 1991. DS 217: Water Examination. Determination of Chemical Oxygen Demand in Water CODCr with Dichromate. Dien, B.S., Nicholsm, N.N., O’Bryan, P.J., Bothast, P.J., 2000. Development of new methanologenic Escherichia coli strains for fermentation of lignocellulosic biomass. Applied Biochemistry and Biotechnology 84–96, 181–186. Food and Agriculture Organization of the United Nations, 2008. webpage on-line http://aaas.fao.org/page/collections (visited Jan 2009). Gray, K.A., Zhao, L., Emptage, M., 2006. Bioethanol. Current Opinion in Chemical Biology 10, 1–6. Hawkes, F., Hussy, I., Kyazze, G., Dinsdale, R., Hawkes, D., 2007. Continuous dark fermentative hydrogen production by mesophilic microflora: principles and progress. International Journal Hydrogen Energy 32 (2), 172–184. Hinman, N.D., Wright, J.D., Hoagland, W., Wyman, C.E., 1989. Xylose fermentation and economic analysis. Applied Biochemistry and Biotechnology 20 (21), 319– 401. IEA Bioenergy, 2007. Potential Contribution of Bioenergy to the World’s Future Energy Demand. http://www.idahoforests.org/img/pdf/PotentialContribution. pdf (visited 03.04.03.). Jönsson, L.J., Palmqvist, E., Nilvebrant, N.-O., Hahn-Hägerdal, B., 1998. Detoxification of wood hydrolysate with lacasse and peroxidise from the white-rot fungus Trametes versicolor. Applied Microbiology and Biotechnology 49, 691–700. Kádár, Z., de Vrije, T., van Noorden, G., Budde, M., Szengyel, Z., Réczey, K., Claassen, P., 2004. Yields from glucose, xylose, and paper sludge hydrolysate during hydrogen production by the extreme thermophile Caldicellulosiruptor saccharolyticus. Applied Biochemistry and Biotechnology 114 (1), 497–508. Kaparaju, P., Buendía, I., Ellegaard, L., Angelidaki, I., 2008. Effects of mixing on methane production during thermophilic anaerobic digestion of manure: labscale and pilot-scale studies. Bioresource Technology 97 (11), 4919–4928. Kaparaju, P., Serrano, M., Thomsen, A.B., Kongjan, P., Angelidaki, I., 2009. Bioethanol, biohydrogen and biogas production from wheat straw in a biorefinery concept. Bioresource Technology 100 (9), 2562–2568. Karakashev, D., Schmidt, J.E., Angelidaki, I., 2008. Innovative Process Scheme for Removal of Organic Matter, Phosphorus and Nitrogen from Pig Manure. Klinke, H.B., Thomsen, A.B., Arhing, B.K., 2001. Potential inhibitors from wet oxidation of wheat straw and their effect on growth and ethanol production by Themoanaerobacter mathranii. Applied Microbiology and Biotechnology 57, 631–638. Klinke, K., Ahring, B., Schmidt, A., Thomsen, A., 2002. Characterization of degradation products from alkaline wet oxidation of wheat straw. Bioresource Technology 82 (1), 15–26. Kongjan, P., Kotay, S.M., Min, B., Angelidaki, I., 2008. Biohydrogen Production from Wheat Straw Hemicellulosic Hydrolysate (WSHH) by extreme Thermophilic Mixed-culture Fermentation. Prepared paper. Larsen, J., Petersen, M.Ø., Thirup, L., Li, H.W., Iversen, F.K., 2008. The IBUS processlignocellulosic bioethanol close to a commercial reality. Chemical Engineering and Technology 31 (5), 765–772. Larsson, S., Palmqvist, E., Hahn-Hagerdal, B., Tengborg, C., Stenberg, K., Zacchi, G., Nilvebrant, N.O., 1999. The generation of fermentation inhibitors during dilute acid hydrolysis of softwood. Enzyme Microbiology Technology 24, 151–159. Linde, M., Jakobsson, E., Galbe, M., Zacchi, G., 2007. Steam pretreatment of dilute H2SO4-impregnated wheat straw and SSF with low yeast and enzyme loadings for bioethanol production. Biomass and Bioenergy 32 (4), 326–332. Montane, D., Farriol, X., Salvado, J., Jollez, P., Chornet, E., 1998. Fractionation of wheat straw by steam explosion pretreatment and alkali delignification. Cellulose pulp and by-products from hemicellulose and lignin. Journal of Wood Chemistry and Technology 18, 171–191. Mussatto, S.I., Dragone, G., Roberto, I.C., 2005. Influence of the toxic compounds present in brewer’s spent grain hemicellulosic hydrolysate on xylose-to-xylitol bioconversion by Candida guilliermondii. Process Biochemistry 40 (12), 3801– 3806. Nielsen, H.B., 2006. Control Parameters for Understanding and Preventing Process Imbalaces in Biogas Plants: Emphasis on VFA Dynamics. PhD dissertation, BioCentrum-DTU, Technical University of Denmark. Olofsson, K., Bertilsson, M., Lidén, G., 2008. A short review on SSF – an interesting process option for ethanol production from lignocellulosic feedstocks. Biotechnol Biofuels 1: 7 (doi: 10.1186/1754-6834-1-7). Rao, A.G., Bapat, A.N., 2006. Anaerobic treatment of pre-hydrolysate liquor (PHL) from a rayon grade pulp mill: pilot and full-scale experience with UASB reactors. Bioresource Technology 97 (18), 2311–2320. Sierra-Álvarez, R., Fiel, J.A., Kortekaas, S., Lettinga, G., 1994. Overview of the anaerobic toxicity caused by organic forest industry wastewater pollutants. Water Science Technology 20, 353–363.

P. Kaparaju et al. / Bioresource Technology 100 (2009) 6317–6323 Sun, R.C., Lawther, J.M., Banks, W.B., 1996. Fractional and structural characterization of wheat straw hemicelluloses. Carbohydrate Polymers 29 (4), 325–331. Sung, S., Liu, T., 2002. Ammonia inhibition on thermophilic aceticlastic methanogens. Water Science Technology 45, 113–120. Thomsen, M.H., Thygesen, A., Thomsen, A.B., 2008. Hydrothermal treatment of wheat straw at pilot plant scale using a three-step reactor system aiming at high hemicellulose recovery, high cellulose digestibility and low lignin hydrolysis. Bioresource Technology 99 (10), 4221–4228. Thomsen, M.H., Thygesen, A., Jørgensen, H., Larsen, J., Christensen, B.H., Thomsen, A.B., 2006. Preliminary results on optimisation of pilot scale pre-treatment of

6323

wheat straw used in co-production of bioethanol and electricity. Applied Biochemistry Biotechnology 129–132, 448–460. Torry-Smith, M., Sommer, P., Ahring, B.K., 2003. Purification of bioethanol effluent in a UASB reactor system with simultaneous biogas formation. Biotechnology Bioengineering 84 (1), 7–12. Wilkie, A.C., Castro, H.F., Cubisnki, K.R., Owens, J.M., Yan, S.C., 2004. Fixed-film anaerobic digestion of flushed dairy manure after primary treatment: wastewater production and characterization. Biosystems Engineering 89 (4), 457–471.