Evaluation of pretreatment methods on mixed ...

Bioresource Technology 101 (2010) 959–964

Contents lists available at ScienceDirect

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

Evaluation of pretreatment methods on mixed inoculum for both batch and continuous thermophilic biohydrogen production from cassava stillage Gang Luo, Li Xie *, Zhonghai Zou, Wen Wang, Qi Zhou State Key Laboratory of Pollution Control and Resources Reuse, Key Laboratory of Yangtze River Water Environment, College of Environmental Science and Engineering, Tongji University, 1239 Siping Road, Shanghai 200092, PR China

a r t i c l e

i n f o

Article history: Received 2 June 2009 Received in revised form 23 August 2009 Accepted 24 August 2009 Available online 17 September 2009 Keywords: Thermophilic hydrogen production Pretreatments Anaerobic sludge

a b s t r a c t Anaerobic sludges, pretreated by chloroform, base, acid, heat and loading-shock, as well as untreated sludge were evaluated for their thermophilic fermentative hydrogen-producing characters from cassava stillage in both batch and continuous experiments. Results showed that the highest hydrogen production was obtained by untreated sludge and there were significant differences (p < 0.05) in hydrogen yields (varied from 32.9 to 65.3 mlH2/gVS) among the tested pretreatment methods in batch experiments. However, the differences in hydrogen yields disappeared in continuous experiments, which indicated the pretreatment methods had only short-term effects on the hydrogen production. Further study showed that alkalinity was a crucial parameter influencing the fermentation process. When the influent was adjusted to pH 6 by NaHCO3 instead of NaOH, the hydrogen yield increased from about 40 to 52 mlH2/gVS in all the experiments. Therefore, pretreatment of anaerobic sludge is unnecessary for practical thermophilic fermentative hydrogen production from cassava stillage. Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction Nowadays, hydrogen has attracted people’s attention because it is a kind of clean energy and the burning product is only water that will not contribute to the greenhouse effect. In addition, hydrogen has higher energy yield which is about 2.75 times higher than hydrocarbon fuels (Lay et al., 1999). However, traditional methods (such as steam reforming of natural gas, gasification of coal and electrolysis of water) to produce hydrogen are energy-consuming. As an alternative way, hydrogen can be efficiently and economically obtained from dark fermentation by hydrogen-producing bacteria (HPB). In such process, organic wastes, such as palm oil mill effluent (O-Thong et al., 2007) and household solid waste (Liu et al., 2006), can be used as substrates and also less additional energy input is needed compared with traditional methods. Hydrogen production from dark fermentation by pure HPB with higher hydrogen yields (2–4 mol H2/mol hexose) has been reported (Kumar et al., 2004; Schröder et al., 1994). However, mixed inoculum is more practical compared with pure cultures, since non-sterile organic wastes can be used as substrates directly. Anaerobic sludge (Mohan et al., 2008), sewage sludge (Chang and Lin, 2006) and soil (Logan et al., 2002) have been used as mixed inoculum for fermentative hydrogen production. However, the produced hydrogen may get consumed by the hydrogen-consum* Corresponding author. Tel.: +86 21 65982692; fax: +86 21 65986313. E-mail address: [email protected] (L. Xie). 0960-8524/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2009.08.090

ing bacteria in the mixed cultures. Thus, the crucial step is to enrich HPB in the mixed cultures and suppress the activity of hydrogen-consuming bacteria (Nath and Das, 2004). Various pretreatment methods including heat, acid, base, chloroform, sodium 2-bromoethanesulfonate (BESA), aeration and loading-shock have been conducted on the mixed inoculums to enrich HPB (O-Thong et al., 2009; Wang and Wan, 2009). Many studies evaluated the effects of pretreatment methods on fermentative hydrogen production. However, there are disagreements on the optimal pretreatment method to enrich HPB from mixed inoculums. Some researchers reported that heat pretreatment on anaerobic sludge was the most suitable method to prepare HPB from mixed inoculums (Mu et al., 2007; Wang and Wan, 2008). While other pretreatment methods such as loading-shock (O-Thong et al., 2009), alkaline (Zhu and Béland, 2006), chloroform (Hu and Chen, 2007) and BESA (Mohan et al., 2008) were also reported as the most effective method to enrich HPB in different studies. Such differences may be resulted from differences in inoculum sources, substrates types and concentrations, cultivation temperatures, fermentation pH and so on. In addition, most of the above studies were conducted using pure cultures such as glucose or sucrose and also most studies were conducted under mesophilic condition. Thermophilic hydrogen production has been proven to be effective in some studies (Akutsua et al., 2009; Ueno et al., 2007). However, little information on the comparison of the effectiveness of various pretreatments on thermophilic hydrogen production from actual wastewater is available.

960

G. Luo et al. / Bioresource Technology 101 (2010) 959–964

It should be further noticed that all the evaluation on suitable pretreatments were conducted only in batch tests. Usually, a certain method was chosen as the best and applied directly in continuously process without further comparison. However, only directly usage of results obtained from batch experiments may not reflect the real case. For example, though heat pretreatment has been demonstrated to be the best method to enrich HPB in batch experiments (Mu et al., 2007; Wang and Wan, 2008), hydrogen production using heat treatment of the mixed inoculum in a continuous-flow reactor declined substantially after the initial increase (Duangmanee et al., 2007). Heat pretreatment was also applied for hydrogen production from sucrose in a fixed-bed reactor, but the hydrogen production rate drastically decreased after 8 days operation (Chang et al., 2002). Even in batch experiments, Zhu and Béland (2006) still observed the difference on the best pretreatment method between the first batch and subsequent second batch cultivation. The highest hydrogen production was obtained by iodopropane pretreated mixed inoculum in the first batch cultivation, whereas, in the subsequent second-step batch test, the base pretreatment method exhibited the highest hydrogen production activity. Therefore, evaluation of various pretreatment methods in continuous experiment is necessary to investigate their long-term effects on hydrogen production. Based on the above considerations, in this study various pretreated mixed inoculums (acid, base, heat, chloroform and loading-shock) and untreated inoculum were compared and evaluated for their hydrogen production characters in both batch and continuous experiments under thermophilic condition. Cassava stillage, a kind of carbohydrate-rich wastewater, was used as substrate for hydrogen production. 2. Methods

gen production from cassava stillage. Six identical serum bottles were used as the reactors with working volume of 200 ml. Each pretreated (heat, chloroform, base, acid and loading-shock) anaerobic sludge of 40 ml was added to bottles, respectively, with substrate of raw cassava stillage of 140 ml. The working volume was adjusted to 200 ml with distilled water. Hydrogen production from cassava stillage by untreated anaerobic sludge was also conducted as the control for comparison. The fermentative pH value of the mixed solution in each bottle was adjusted to 6 by 2 N NaOH or 2 N HCl. Nitrogen gas was purged into bottles for 5 min to provide anaerobic condition. The capped bottles with rubber stoppers were placed in a reciprocating water bath shaker and rotated at 150 rpm with the preset temperature of 60 °C. The evolved biogas was collected by gas bag. The amount of biogas was determined periodically using syringe and at the same time the composition of the biogas was measured. The above experiments were carried out in triplicate. After the hydrogen production ceased in the last batch, the reactors were switched to semi-continuous mode and operated as continuously stirred tank reactor (CSTR). Equal amount of digested and fresh cassava stillage were removed and added to the reactor periodically and the HRT was maintained at 3 days. In this study, all data were obtained at a steady-state of each operation condition. Steady-state was defined by a sustained biogas production within ±10% deviation, and in this period, parameters including pH, volatile fatty acids (VFA), hydrogen production and soluble carbohydrate were determined for three consecutive days, creating six replicate data points for every measured parameter for the experimental condition under examination. During the continuous operation, the fermentative influent pH was adjusted to 6 by NaOH and NaHCO3 before respectively needed to make a comparison on hydrogen production.

2.1. Feedstock 2.4. Analytical methods Cassava stillage used in this study was obtained directly from Taicang cassava ethanol plant (Jiangsu, China). The main characteristics of cassava stillage were as follows: pH 4.0, total solids 70.9 g/ L, volatile solids 60.1 g/L, total carbohydrate 45.2 g/L and soluble carbohydrate 7.6 g/L. Detailed information of cassava stillage can be found in our previous study (Luo et al., 2009). After collected, the cassava stillage was stored at 4 °C before usage. 2.2. Seed sludge and preparation of pretreated sludge Anaerobic granular sludge acquired from a full-scale mesophilic upflow anaerobic sludge blanket reactor was used as original inoculum for hydrogen production experiments. The concentrations of total suspended solids, volatile suspended solids and the pH of the anaerobic sludge were 70 g/L, 42 g/L and 7.5 g/L, respectively. Five pretreatment methods, heat, base, acid, chloroform and loadingshock, were applied to anaerobic sludge to enrich HPB. For heat pretreatment, the anaerobic sludge was incubated at 90 °C for 1 h. For chloroform pretreatment, the chloroform was added to the cassava stillage with concentration of 0.2%. For base pretreatment, the anaerobic sludge was adjusted to pH 12 by 2 N NaOH and maintained for 24 h. For acid pretreatment, the anaerobic sludge was adjusted to pH 3 by 2 N HCl and also maintained for 24 h. For loading-shock pretreatment, 50 ml anaerobic sludge was mixed with 40 g sucrose and then diluted to 500 ml by distilled water to obtain a higher loading of 80 kgCOD/L, followed by mixing at 200 rpm for 2 days.

The collected samples were centrifuged at 3500 rpm for 10 min, and filtrated through 0.45 lm filters to determine soluble components. Total solid (TS), volatile solid (VS) were analyzed in duplicate in accordance with Standard APHA Methods (APHA, 1995). Total and soluble carbohydrates were determined using phenol– sulfuric acid method (Miller, 1959). The concentrations of ethanol and VFA (C2–C5) were determined by gas chromatograph (HP6890II, USA) equipped with a flame ionization detector and analytical column CPWAX52CB (30 m 0.25 mm 0.25 lm). The temperature of the injector and FID were 200 and 220 °C, respectively. Nitrogen was the carrier gas with a flow rate of 50 mL/ min. The GC oven was programmed to begin at 110 °C and remain there for 2 min, then increased at a rate of 10 °C/min to 220 °C, and hold at 220 °C for an additional 2 min. The sample injection volume was 1.0 lL. The biogas composition was determined using a gas chromatograph (Shimadzu GC-14B, Japan) equipped with thermal conductivity detector and a stainless steel column packed with Carbosive S II (U3.2 mm 2 m). Injector, detector and column temperatures were kept at 100, 105 and 60 °C, respectively. The carrier was nitrogen, and the flow rate was 30 ml/min. The measured biogas volume was adjusted to volume at STP (standard temperature 0 °C and pressure 1 atm). 3. Results 3.1. Thermophilic hydrogen production in batch experiments

2.3. Experimental design and procedure Batch experiments were firstly conducted to compare the effects of different pretreatment methods on seed sludge for hydro-

The batch experiments with anaerobic sludges pretreated by acid, base, heat, loading-shock and chloroform were performed under thermophilic condition. Seed sludge without any pretreatment

961

6000

Ethanol Acetate Propionate Butyrate

Concentration (mg/L)

600 500 400 300 200 100

5000 4000 3000 2000 1000

0

Pretreatment methods

H

ea t-s

ho ck

ho ck Lo ad in gs

A

ci d

Ba se

Ch lo ro fo rm

Co nt ro l

ck ho H ea t-s

ng oc k-

lo

ad i

A ci d Sh

Ch l

or o

fo

tro Co n

Ba se

rm

0 l

Cumulative hydrogen production (ml)



was used as control for comparison. Fig. 1 illustrates the cumulative hydrogen production for different cases. It is quite interesting to notice that the untreated sludge gave the highest cumulative hydrogen production of 558 ml, while the other five pretreated sludges gave relatively low hydrogen production with the sequence of loadingshock (550 ml), base (504 ml), heat (490 ml), acid (390 ml) and chloroform (281 ml). The corresponding hydrogen yields for each treatment were calculated and listed in Table 1. In this study, methane was detected only in untreated and base pretreated cases with small amount of 14 and 24 ml, respectively indicating the low activity of methanogens, while all the other pretreatments fully inhibited the activity of methanogens. A decrease of fermentative pH was observed and the final pH was in the range of 4.7–5.4, lower than the initial pH 6 (Table 1). The acidic pH condition might lead to the lower methane production as observed in control experiment. The concentration and component of aqueous products could reflect the metabolism of hydrogen-producing anaerobes, which has considerable effect on the hydrogen production. As presented in Table 1, the highest amount of VFA/ethanol (7404 mg/L) was obtained in control experiment while the lowest amount of VFA/ethanol (4495 mg/L) was obtained in chloroform pretreatment experiment, which was consistent with the hydrogen production. The dominant species were acetate, butyrate and ethanol. Small amount of propionate was only detected in the system of untreated sludge seed. Their concentrations for each treatment are shown in Fig. 2. Among the species, butyrate was predominant and accounted for more than 75% of the total amount of VFA/ethanol except the case of loading-shock pretreated sludge (56% in this study). Butyrate concentration was observed to be correlated with the hydrogen production. Chen et al. (2004) also found highest hydrogen production was obtained when butyrate was predominate (70–85% of the total VFA/ethanol). 3.2. Thermophilic hydrogen production in continuous experiments To further evaluate the long-term influence of pretreatment methods on hydrogen production from cassava stillage under thermophilic condition, batch mode was switched to semi-continuous

Fig. 2. Distributions of VFA/ethanol in batch operation by various pretreatment methods on mixed inoculum.

700

Hydrogen production (ml)

Fig. 1. Cumulative hydrogen production in batch experiments by various pretreatment methods on mixed inoculum.

pH adjusted by NaHCO3

pH adjusted by NaOH

600 500 400

Control Chlorofom Base Acid Loading-shock Heat-shock

300 200 100 0 0

5

10

15

20

25

30

Time (d) Fig. 3. The time courses of hydrogen production in continuous experiments by various pretreatment methods on mixed inoculum.

mode operation. Fig. 3 shows the production of hydrogen at different cases with continuous operation time of 28 days. During the continuous operation, the influent pH was firstly adjusted to 6 by NaOH in the operation time of 0–15 days. And then, NaHCO3 was used to adjust the influent pH to 6 from the 16th day to increase the buffer capacity of the influent. As shown in Fig. 3, hydrogen production was fluctuated in the first several days, and after about 6 days continuous operation, stable hydrogen production could be obtained. It is interesting to notice that pretreated sludge and untreated sludge gave similar hydrogen production after steady-state was achieved. The hydrogen production performances in the steady-state were summarized and listed in Table 2. An analysis of variation (ANOVA, p > 0.05) further demonstrated that anaerobic sludges with and without pretreatments have no significant effects on hydrogen yield in the long-term operation. In the whole continuous operation process, no methane was detected for all cases. Therefore, pretreatments on anaerobic sludge only had short-term effects on the thermophilic hydrogen production from cassava stillage and the differences among the various pretreatments disappeared after long-period operation.

Table 1 Conclusions of hydrogen production performances in batch experiments by various pretreatment methods on mixed inoculum.

Final pH Hydrogen content (%) Hydrogen yield (mlH2/gVS) Total VFA/ethanol (mg/L)

Control

Chloroform

Base

Acid

Loading-shock

Heat-shock

5.38 ± 0.01 50.7 ± 1.2 65.3 ± 1.6 7404 ± 158

5.12 ± 0.01 56.3 ± 0.9 32.9 ± 1.2 4495 ± 156

5.42 ± 0.02 51.3 ± 1.8 59.0 ± 1.5 7278 ± 210

4.91 ± 0.01 53.1 ± 0.9 46.5 ± 1.4 5111 ± 263

4.69 ± 0.02 57.1 ± 1.5 64.4 ± 1.2 6600 ± 287

5.43 ± 0.01 51.4 ± 1.8 57.4 ± 1.4 6859 ± 247

962


Table 2 Conclusions of hydrogen production performances in continuous experiments by various pretreatment methods on mixed inoculum.

Influent pH adjusted by NaOH pH Hydrogen content (%) Hydrogen yield (mlH2/gVS) Total VFA/ethanol (mg/L) Influent pH adjusted by NaHCO3 pH Hydrogen content (%) Hydrogen yield (mlH2/gVS) Total VFA/ethanol (mg/L)

Control

Chloroform

Base

Acid

Loading-shock

Heat-shock

4.89 ± 0.02 57.2 ± 1.2 39.0 ± 1.6 4893 ± 145

4.89 ± 0.05 56.9 ± 1.8 37.7 ± 1.7 4822 ± 85

4.85 ± 0.02 57.5 ± 0.9 36.4 ± 0.8 4681 ± 267

4.87 ± 0.02 58.4 ± 0.6 37.6 ± 2.2 4539 ± 195

4.88 ± 0.07 58.1 ± 1.2 35.5 ± 0.8 4526 ± 236

4.86 ± 0.02 57.5 ± 1.3 38.0 ± 0.5 4864 ± 315

5.20 ± 0.01 54.3 ± 1.4 52.9 ± 0.8 6203 ± 196

5.21 ± 0.01 55.3 ± 0.9 52.7 ± 0.8 6134 ± 66

5.18 ± 0.02 55.1 ± 1.2 51.8 ± 1.1 5966 ± 89

5.22 ± 0.01 53.8 ± 1.7 52.9 ± 2.2 6321 ± 220

5.21 ± 0.01 55.2 ± 0.8 51.5 ± 2.0 6244 ± 250

5.20 ± 0.01 54.3 ± 1.6 52.5 ± 1.1 6273 ± 132

As shown in Fig. 3, the use of NaHCO3 instead of NaOH to adjust influent pH was beneficial to the hydrogen production. The hydrogen yield sharply increased from below 40 mlH2/gVS to about 52 mlH2/gVS and quickly achieved steady-state. In addition, the effluent pH was increased to 5.2, compared to the pH 4.9 with NaOH addition (Table.2). Correspondingly, the total amount of VFA/ethanol increased from about 5000 mg/L to more than 6000 mg/L. The compositions of VFA/ethanol at the end of fermentation were similar with the results obtained from batch tests that acetate, butyrate and ethanol were the main species (Fig. 4). Butyrate was predominant and accounted for more than 80% of the total VFA/ethanol, including the case using loading-shock pretreated sludge. In addition, small amount of propionate was detected when NaHCO3 was used to adjust pH, which might be attributable to the increased fermentative pH.


6000 5000

Ethanol Acetate Propionate N-butyrate

A

4000 3000 2000 1000

ad

ck ho

Lo

H ea t-s

in gsh oc k

d ci A

Ba se

or of Ch

lo r

Co

nt ro l

m

0



6000 5000

B

4000 3000 2000 1000

-s ea t

in ad Lo

H

g-

sh

ho

ck

oc k

d ci A

se Ba

ro lo Ch

Co

nt

fo

ro

l

rm

0

Pretreatment methods Fig. 4. Distributions of VFA/ethanol in steady-state periods of continuous operation by various pretreatment methods on mixed inoculum (A) Influent pH adjusted by NaOH (B) Influent pH adjusted by NaHCO3.

4. Discussion Though mesophilic anaerobic sludge was used as seed sludge in this study, hydrogen can be still obtained under thermophilic condition. The reason was that mesophilic anaerobic sludge belonged to mixed inoculum and it contained various organisms which could function at different temperatures (Kim et al., 2009). Kotsopoulos et al. (2009) even demonstrated that mesophilic anaerobic sludge can be successfully used as seed sludge for hyper-thermophilic (70 °C) hydrogen production from pig slurry. The anaerobic sludge without any pretreatment exhibited the highest hydrogen production capacity which is not consistent with the previous reports that the pretreatment methods on mixed inoculums could significantly enhance the hydrogen production compared with raw inoculums (Mohan et al., 2008; O-Thong et al., 2009; Wang and Wan, 2008). Two reasons might be accounted for such difference. Firstly, cassava stillage, a kind of actual carbohydrate-rich wastewater was used as substrate in this study and its characteristics were much different from pure substrates like glucose and sucrose used in previous studies. For example, the pH of cassava stillage was as low as 4.0 and the alkalinity could be ignored, while the glucose culture was neutral with higher buffer capacity (NaHCO3 higher than 5000 mg/L) (O-Thong et al., 2009). Therefore, in control experiment, the fermentative pH could naturally decrease from 6 to 5.38 due to the lower buffer capacity of cassava stillage. The lowered fermentative pH is inhibitory to methanogenic activity, but suitable for HPB (Zhu and Béland, 2006). Secondly, pretreatment methods on mixed inoculums provided extreme conditions for the enrichment of spore-forming HPB. Meanwhile, such extreme condition inhibited some nonspore-forming H2 producers like Enterobacter spp. (Kraemer and Bagley, 2007), consequently leading to the decline of total hydrogen production capacity compared with mixed inoculum without any pretreatment. In batch operations, the maximum hydrogen yield was obtained in control experiment and there were significant differences in hydrogen yield by different pretreatment methods (ANOVA, p < 0.05). The difference was correlated with the different metabolic products (as shown in Fig 2), which should be attributed to the diversity of microbial communities. For example, Iyer et al. (2004) identified Closidium acetobutylicum as the major bacterium that is responsible for the butyrate-type fermentation in a heattreated anaerobic sludge while mixed-type fermentation was obtained when Citrobacter, Clostridium and Klebsiella were predominant in acid pretreated sludge (Fang et al., 2002b). More recently, O-Thong et al. (2009) compared five different pretreatment methods of the seed sludge in batch experiments and found that loading-shock and heat shock resulted in the dominance of Thermoanaerobacterium thermosaccharolyticum while base and acid treated seeds were dominated by Clostridium sp. and BESA-treated seeds were dominated by Bacillus sp., leading to the difference in metabolic products and hydrogen production. However, though

963


the pretreatments of anaerobic sludge may kill or inhibit some bacteria and change the microbial community of the seed sludge and further led to the different hydrogen production performances in batch experiments, the hydrogen production performances were not stable until a steady-state was achieved in continuous operation as shown in Fig 3. The results were accordance with Yokoyama et al. (2007) who found that no change of microbial community was observed only after five batches enrichment when fermenting glucose to hydrogen by cow manure. Therefore, evaluating pretreatment method on seed sludge for biohydrogen production must be conducted under continuous condition. It is obvious that pretreatment methods on mixed sludge have only short-term effects on hydrogen production from cassava stillage, and the hydrogen production will not be affected by the pretreatment methods after a steady-state was achieved in the long-term operation. It is the first time to report this phenomenon. Previous studies just evaluated different pretreatment methods on hydrogen production in batch experiments and a certain method was suggested as the best one (Mohan et al., 2008; O-Thong et al., 2009). For further studies, it is necessary to apply the results obtained from batch experiments to continuous experiments to assure the reliability of the results. Compared with maximum hydrogen yield of 65 mlH2/gVS obtained in batch experiments, the hydrogen yields in continuous experiments were much lower with the values below 40 mlH2/ gVS when influent pH was adjusted by NaOH. The effluent pH 4.9 was relatively lower than the suggested optimal fermentation pH of 5.5–6.0 (Fang and Liu, 2002a; Lay, 2000; Lay et al., 2005). Alkalinity was shown to be a crucial parameter that was related to the fermentation pH and adequate alkalinity was necessary to maintain optimal biological activity for hydrogen production (Valdez-Vazquez and Poggi-Varaldo, 2009; Zhu et al., 2009). Thus NaHCO3 was used instead of NaOH to adjust the influent pH in order to increase the alkalinity of influent from 16th day. Correspondingly, the effluent pH increased from about 4.9 to 5.2 due to the higher buffer capacity of influent. At the same time, hydrogen yields increased 30% with values of 52 mlH2/gVS. The results further demonstrate that alkalinity is a key parameter in the fermentation hydrogen process while pretreatments to mixed inoculums have little effect on the thermophilic continuous hydrogen production process. The hydrogen yields in the continuous experiments were still lower than the batch experiments, and it could be further improved by the optimization of influent alkalinity and HRT. In continuous operation, similar hydrogen production performances were obtained in all experiments that were different from the batch experiments, which may attributed to the similar dominant microbial community. Many studies with various inoculums demonstrated that Thermoanaerobacterium sp., Clostridium sp. and Bacillus sp. were usually the dominant bacteria in thermophilic hydrogen production process and they could produce hydrogen efficiently from carbohydrates (O-Thong et al., 2009; Prasertsan et al., 2009; Ueno et al., 2006). In our study, the anaerobic sludge may also contain the above bacteria and they could not be killed by the pretreatments since they are spore-forming bacteria. Though different pretreatments of the seed sludge may inhibit such bacteria at different extent and result in the different microbial communities in the batch experiment, it may be just temporary. After long-period cultivation, the spore-forming HPB in different pretreated anaerobic sludge may recover their activities and new stable microbial communities would be constructed that was adapted to the environment. Studies on continuous hydrogen production demonstrated that dominant bacteria are strongly associated with the substrate type, HRT, OLR and temperature, and the variation of the above parameters may lead to the shift of the dominant bacteria (Iyer et al., 2004; Prasertsan et al., 2009). In our study, with the same operation parameters such as

influent, HRT and temperature, similar dominant microbial communities with different pretreated seed sludges may be constructed after long-period cultivation, which further led to the similar hydrogen production performances. The distributions of VFA/ethanol were similar in all the experiments as shown in Fig 4, which indicated similar microbial metabolic pathway and also further reflect similar dominant microbial community in all experiments. An interesting phenomenon was that butyrate was predominant in the VFA/ethanol, accounting for more than 80%. Taken control experiment as example, the butyrate concentration increased from 4340 to 5476 mg/L while other components were stable except the appearance of propionate with relatively low concentration when the effluent pH increased from 4.89 to 5.20. The butyrate composition and the value of butyrate/acetate were between 88.7% and 88.3% and 12.7 and 14.9, respectively. The butyrate/acetate ratio greater than 2.6 indicated efficient H2 production by anaerobic microflora (Han et al., 2005; Hussy et al., 2003). Though acetate is more favorable for hydrogen production compared with butyrate as can be seen from Eq. (1) and (2), the acetate also can be produced by homoacetogens to consume H2 and CO2 (Eq. (3)). Some studies also found the lower hydrogen production accompanied the higher production of acetate (Kim et al., 2004; Hussy et al., 2003). The lower concentration of acetate in our study suggested the lower activity of homoacetogens. The butyrate composition is high which is suitable for butyrate production from cassava stillage. Butyrate is an important chemical and has many applications in the chemical industry as well as the food and pharmaceutical industries. Fermentative butyrate production by Clostridium tyrobutyricum has been studied by some researchers. However, unexpected acetate was produced in their studies with butyrate/acetate between 6.5 and 9.0 (Jiang et al., 2009), and 5.52 and 8.88 (Liu and Yang, 2006). The butyrate/acetate obtained in our study was higher than the above studies, which indicated the possibility to produce butyrate from wastes by mixed sludge. The lower butyrate concentration (5476 mg/L) was due to the lower soluble carbohydrate (7600 mg/L) in cassava stillage and pretreatments should be applied to the cassava stillage to improve the soluble carbohydrate.

C6 H12 O6 þ 2H2 O—CH3 CH2 CH2 COO þ 2HCO3 þ 3Hþ þ 2H2

ð1Þ

C6 H12 O6 þ 4H2 O—2CH3 COO þ 2HCO3 þ 4Hþ þ 4H2

ð2Þ

4H2 þ

2HCO3

þ

þ H —CH3 COO þ 4H2 O

ð3Þ

In the whole continuous operation processes, no methane was detected, indicating that the batch operations of the reactors were necessary to suppress the activity of methonogens and stimulate the activity of HPB. The result is consistent with the study of Zhu and Béland (2006). They evaluated the effects of five different pretreatment methods on mixed inoculum in batch experiments and methane was only detected in control and base pretreatment experiments in the first batch cultivation. However, it disappeared in the subsequent second batch cultivation. The stable and methane-free hydrogen production observed in our continuous study should be due to the thermophilic fermentation, which may inhibit the activity of hydrogen-consuming bacteria in the whole operation period. Akutsua et al. (2009) also had proven that thermophilic fermentation was preferable to mesophilic condition for inhibiting hydrogen consumers and producing hydrogen stably by mixed inoculum without any pretreatment. Another important parameter is the influent alkalinity. In our study, the suitable influent alkalinity resulted in the effluent pH below 5.5 which might be helpful to suppress the activity of methanogens. Liu et al. (2008) demonstrated pH 5.5 was enough to inhibit methonogens and produce hydrogen stably at 3-day HRT at extreme-temperature. However, with fermentation pH as low as 4.5, Kim et al. (2004) still

964


detected methane in mesophilic hydrogen production process. The combination effects of thermophilic condition and lower pH may lead to the successful operation of hydrogen production by mixed sludge without any pretreatment. 5. Conclusions The results of the present study demonstrated that pretreatments of the anaerobic sludge had only short-term effects on the thermophilic hydrogen production from cassava stillage and after long-term continuous operation, no differences in hydrogen production were observed by different pretreatments on anaerobic sludge and untreated sludge. Therefore, for actual thermophilic hydrogen production from cassava stillage, no pretreatment is needed to inhibit the hydrogen-consuming bacteria in the anaerobic sludge. Instead of pretreatments, influent alkalinity was an important parameter influencing the thermophilic continuous hydrogen production. When the influent pH was adjusted to 6 by NaHCO3 instead of NaOH, the hydrogen yield increased from about 40 to 52 mlH2/gVS. Butyrate was the most abundant byproduct accounting for more than 80% of the total VFA/ethanol and it is an ideal and easier way for fermentative butyrate production from wastes by mixed inoculum. Acknowledgements This research was financially supported by the Bayer Sustainable Development Foundation and the Science Committee Project of Shanghai (Grant no. 062307038, 2008DFA91000); the authors wish to thank Taicang cassava ethanol plant for their raw cassava stillage and valuable practical experience. References Akutsua, Y., Li, Y.Q., Harada, H., Yu, H.Q., 2009. Effects of temperature and substrate concentration on biological hydrogen production from starch. Int. J. Hydrogen Energy 34 (6), 2558–2566. American Pulic Health Association, 1995. Standard Methods for the Examination of Water and Wastewater, 19th ed. APHA, Washington, DC, USA. Chang, F.Y., Lin, C.Y., 2006. Calcium effect on fermentative hydrogen production in an anaerobic up-flow sludge blanket system. Water Sci. Technol. 54 (9), 105– 112. Chang, J.S., Lee, K.S., Lin, P.J., 2002. Biohydrogen production with fixed-bed bioreactors. Int. J. Hydrogen Energy 27, 1167–1174. Chen, C.C., Lin, C.Y., Lin, M.C., 2004. Acid–base enrichment enhances anaerobic hydrogen production process. Appl. Microbiol. Biotechnol. 58 (2), 224–228. Duangmanee, T., Padmasiri, S.I., Simmons, J.J., Raskin, L., Sung, S., 2007. Hydrogen production by anaerobic microbial communities exposed to repeated heat treatments. Water Environ. Res. 79 (9), 975–983. Fang, H.H.P., Liu, H., 2002a. Effect of pH on hydrogen production from glucose by a mixed culture. Bioresource Technol. 82 (1), 87–93. Fang, H.H.P., Zhang, T., Liu, H., 2002b. Microbial diversity of a mesophilic H2 producing sludge. Appl. Microbiol. Biotechnol. 58 (1), 112–118. Han, S.K., Kim, S.H., Shin, H.S., 2005. UASB treatment of wastewater with VFA and alcohol generated during hydrogen fermentation of food waste. Process Biochem. 40 (8), 2897–2905. Hu, B., Chen, S.L., 2007. Pretreatment of methanogenic granules for immobilized hydrogen fermentation. Int. J. Hydrogen Energy 32 (15), 3266–3273. Hussy, I., Hawkes, F.R., Dinsdale, R., Hawkes, D.L., 2003. Continuous fermentative hydrogen production from a wheat starch co-product by mixed microflora. Biotechnol. Bioeng. 84 (6), 619–626. Iyer, P., Bruns, M.A., Zhang, H.S., Ginkel, S.V., Logan, B.E., 2004. H2-producing bacteria communities from a heat-treated soil inoculum. Appl. Microbiol. Biotechnol. 66 (2), 166–173. Jiang, L., Wang, J.F., Liang, S.Z., Wang, X.N., Cen, P.L., Xu, Z.N., 2009. Butyric acid fermentation in a fibrous bed bioreactor with immobilized Clostridium tyrobutyricum from cane molasses. Bioresource Technol. 100 (13), 3403–3409. Kim, I.S., Hwang, M.H., Jang, N.J., Hyun, S.H.S.H., Lee, S.T., 2004. Effect of low pH on the activity of hydrogen utilizing methanogen in bio-hydrogen process. Int. J. Hydrogen Energy 29 (11), 1133–1140. Kim, W., Hwang, K., Shin, S.G., Lee, S., Hwang, S., 2009. Effect of high temperature on bacterial community dynamics in anaerobic acidogenesis using mesophilic sludge inoculum. Bioresource Technol. doi:10.1016/j.biortech.2009.03.029.

Kotsopoulos, T.A., Fotidis, I.A., Tsolakis, N., Martzopoulos, G.G., 2009. Biohydrogen production from pig slurry in a CSTR reactor system with mixed cultures under hyper-thermophilic temperature (70 °C). Biomass Bioenergy 33 (9), 1168–1174. Kraemer, J.T., Bagley, D.M., 2007. Improving the yield from fermentative hydrogen production. Biotechnol. Lett. 29 (5), 685–695. Kumar, N., Ghosh, A., Das, D., 2004. Redirection of biochemical pathways for the enhancement of H2 production by Enterobacter cloacae. Biotechnol. Lett. 23 (7), 537–541. Lay, J.J., 2000. Modeling and optimization of anaerobic digested sludge converting starch to hydrogen. Biotechnol. Bioeng. 68 (3), 269–278. Lay, J.J., Lee, Y.J., Noike, T., 1999. Feasibility of biological hydrogen production from organic fraction of municipal solid waste. Water Res. 33 (11), 2579–2586. Lay, J.J., Tsai, C.J., Huang, C.C., Chang, J.J., Chon, C.H., Fan, K.S., Chang, J.I., Hsu, P.C., 2005. Influences of pH and hydraulic retention time on anaerobes converting beer processing wastes into hydrogen. Water Sci. Technol. 52 (1–2), 123–129. Liu, X.G., Yang, S.T., 2006. Kinetics of butyric acid fermentation of glucose and xylose by Clostridium tyrobutyricum wild type and mutant. Process Biochem. 41 (4), 801–808. Liu, D.W., Liu, D.P., Zeng, R.J., Angelidak, I., 2006. Hydrogen and methane production from household solid waste in the two-stage fermentation process. Water Res. 40 (11), 2230–2236. Liu, D.W., Zeng, R.J., Angelidaki, I., 2008. Effects of pH and hydraulic retention time on hydrogen production versus methanogenesis during anaerobic fermentation of organic household solid waste under extreme-thermophilic temperature (70 degrees C). Biotechnol. Bioeng. 100 (6), 1108–1124. Logan, B.E., Oh, S.E., Kim, I.S., Ginkel, S.V., 2002. Biological hydrogen production measured in batch anaerobic respirometers. Environ. Sci. Technol. 36 (11), 2530–2535. Luo, G., Xie, L., Zhou, Q., 2009. Enhanced treatment efficiency of an anaerobic sequencing batchreactor (ASBR) for cassava stillage with high solids content. J. Biosci. Bioeng. 107 (6), 641–645. Miller, G.L., 1959. Use of dinitrosalicylic acid reagent for determination of reducing sugar. Anal. Chem. 31 (3), 426–427. Mohan, S.V., Babu, V.L., Sarma, P.N., 2008. Effect of various pretreatment methods on anaerobic mixed microflora to enhance biohydrogen production utilizing dairy wastewater as substrate. Bioresource Technol. 99 (1), 59–67. Mu, Y., Yu, H.Q., Wang, G., 2007. Evaluation of three methods for enriching H2producing cultures from anaerobic sludge. Enzyme Microb. Technol. 40 (4–5), 947–953. Nath, K., Das, D., 2004. Improvement of fermentative hydrogen production: various approaches. Appl. Microbiol. Biotechnol. 65 (5), 520–529. O-Thong, S., Prasertsan, P., Intrasungkha, N., Dhamwichukorn, S., Birkeland, N.K., 2007. Improvement of biohydrogen production and treatment efficiency on palm oil mill effluent with nutrient supplementation at thermophilic condition using an anaerobic sequencing batch reactor. Enzyme Microb. Technol. 41 (5), 583–590. O-Thong, S., Prasertsan, P., Birkeland, N.K., 2009. Evaluation of methods for preparing hydrogen-producing seed inocula under thermophilic condition by process performance and microbial community analysis. Bioresource Technol. 100 (2), 908–918. Prasertsan, P., O-Thong, S., Birkeland, N.K., 2009. Optimization and microbial community analysis for production of biohydrogen from palm oil mill effluent by thermophilic fermentative process. Int. J. Hydrogen Energy 34 (17), 7448– 7459. Schröder, C., Selig, M., Schönheit, P., 1994. Glucose fermentation to acetate, CO2 and H2 in the anaerobic hyperthermophilic eubacterium Thermotoga maritima: involvement of the Embden–Meyerhof pathway. Arch. Microbiol. 161 (6), 460– 470. Ueno, Y., Sasaki, D., Fukui, H., Haruta, S., Ishii, M., Igarashi, Y., 2006. Changes in bacterial community during fermentative hydrogen and acid production from organic waste by thermophilic anaerobic microflora. J. Appl. Microbiol. 101 (2), 331–343. Ueno, Y., Fukui, H., Goto, M., 2007. Operation of a two-stage fermentation process producing hydrogen and methane from organic waste. Environ. Sci. Technol. 41 (4), 1413–1419. Valdez-Vazquez, I., Poggi-Varaldo, H.M., 2009. Alkalinity and high total solids affecting H2 production from organic solid waste by anaerobic. Int. J. Hydrogen Energy 34 (9), 3639–3646. Wang, J.L., Wan, W., 2008. Comparison of different pretreatment methods for enriching hydrogen-producing bacteria from digested sludge. Int. J. Hydrogen Energy 33 (12), 2934–2941. Wang, J.L., Wan, W., 2009. Factors influencing fermentative hydrogen production: a review. Int. J. Hydrogen Energy 34 (2), 799–811. Yokoyama, H., Moriya, N., Ohmori, H., Waki, M., Ogino, A., Tanaka, Y., 2007. Community analysis of hydrogen-producing extreme thermophilic anaerobic microflora enriched from cow manure with five substrates. Appl. Microbiol. Biotechnol. 77 (1), 213–222. Zhu, H.G., Béland, M., 2006. Evaluation of alternative methods of preparing hydrogen producing seeds from digested wastewater sludge. Int. J. Hydrogen Energy 31 (14), 1980–1988. Zhu, H.G., Parker, W., Basnar, R., Proracki, A., Falletta, P., Béland, M., Seto, P., 2009. Buffer requirements for enhanced hydrogen production in acidogenic digestion of food wastes. Bioresource Technol. 100 (21), 5097–5102.




Comparative study of mechanical, hydrothermal, chemical and enzymatic treatments of digested biofibers to improve biogas production Emiliano Bruni a,b, Anders Peter Jensen b, Irini Angelidaki a,* a b

Department of Environmental Engineering, Technical University of Denmark, Building 113, 2800 Kgs. Lyngby, Denmark Process Development, Xergi A/S, Hermesvej 1, 9530 Støvring, Denmark

a r t i c l e

i n f o

Article history: Received 30 April 2010 Received in revised form 15 June 2010 Accepted 24 June 2010 Available online 16 July 2010 Keywords: Treatment Lignocellulose Biogas Hydrolysis Enzymes

a b s t r a c t Organic waste such as manure is an important resource for biogas production. The biodegradability of manures is however limited because of the recalcitrant nature of the biofibers it contains. To increase the biogas potential of the biofibers in digested manure, we investigated physical treatment (milling), chemical treatment (CaO), biological treatment (enzymatic and partial aerobic microbial conversion), steam treatment with catalyst (H3PO4 or NaOH) and combination of biological and steam treatments (biofibers steam-treated with catalyst were treated with laccase enzyme). We obtained the highest methane yield increase through the chemical treatment that resulted in 66% higher methane production compared to untreated biofibers. The combination of steam treatment with NaOH and subsequent enzymatic treatment increased the methane yield by 34%. To choose the optimal treatment, the energy requirements relative to the energy gain as extra biogas production have to be taken into account, as well as the costs of chemicals or enzymes. Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction Anaerobic digestion of organic waste and residues combines both sustainable treatment and renewable energy production. Some substrates such as lignocellulosic materials are resistant to anaerobic digestion and can be converted into biogas only to a low extent. The low susceptibility of these materials to conversion into biogas is a result of their composition and structure. Lignocellulose is the complex and rigid matrix of plant cells, it is resistant to enzymatic attack because of the tight association between lignin, cellulose and hemicellulose. Cellulose and hemicellulose (carbohydrates composed of hexoses and mainly pentoses, respectively) can be degraded in biogas processes. Lignin can however not be degraded under anaerobic conditions (Fernandes et al., 2009). In full-scale biogas plants digesting manure, the low digestibility of the biofibers contained in the manure causes a loss of methane production and limits the overall efficiency of the process (Jin et al., 2009). Therefore, treatments facilitating the accessibility of holocellulose (cellulose and hemicellulose) are needed to increase the biogas potential of biofibers in manure. Many treatments for increasing the biodegradability of lignocellulosic material have been reported (Demirbas, 2008). Mechanical treatment (milling) increases the surface available for enzymatic attack and proved to be suitable for applications at full-scale biogas * Corresponding author. Tel.: +45 4525 1429; fax: +45 4593 2850. E-mail address: [email protected] (I. Angelidaki). 0960-8524/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2010.06.108

plants, increasing the methane yield of lignocellulosic substrates by up to 25% (Hartmann et al., 2000). Steam treatment has been optimized mainly for ethanol production. Glucose yields of 98% after enzymatic hydrolysis have been registered for wheat straw steam-treated with H2SO4 (Talebnia et al., 2010). Likewise, steam treatment with H3PO4 hydrolyzed hemicellulose and increased the accessibility of cellulose for enzymatic hydrolysis (Geddes et al., 2010). Steam treatment with NaOH has been reported to increase the biogas production of sorted municipal solid waste by 50% (Wang et al., 2009). Chemical treatments investigated include treatments with acids, bases and oxidants (Taherzadeh and Karimi, 2008). Chemical treatments with NaOH are among those that have been investigated most (Tanaka et al., 1997; Zheng et al., 2009). Alkaline hydrolysis with NaOH has been successfully applied to treat lignocellulosic materials such as straw or hardwood (Sun and Cheng, 2002). Some aerobic microorganisms (white-, brown-, soft-rot fungi) can selectively degrade lignin or hemicellulose. Srilatha et al. (1995) obtained a 33% methane yield increase for orange processing waste by treatment with selected fungi strains. Similarly, Taherzadeh and Karimi (2008) reported improved enzymatic hydrolysis (94% sugar recovery) when treating office paper with selected aerobic bacteria. Commercially available enzymes have been used to treat the substrate for biofuels production. While enzymatic hydrolysis of holocellulose has been widely investigated mainly in connection with bioethanol production (Talebnia et al., 2010), enzymatic hydrolysis or oxidation of lignin for biogas production has not been sufficiently researched.

8714

E. Bruni et al. / Bioresource Technology 101 (2010) 8713–8717

Ligninases such as laccase, lignin peroxidase and manganese peroxidase are reported to delignify lignocellulose (Ohkuma, 2003; Hatakka, 1994). For the present study, we have treated biofibers separated from digested manure with different methods with the aim to increase their methane potential. We tested chemical, biological, physical, hydrothermal treatments and combinations. The treatments considered have been selected due to their low input energy requirements. We evaluated each treatment based on the effects on the measured methane potential. 2. Methods 2.1. Materials Biofibers separated from digested manure were used as the substrate. The biofibers were collected from the effluent of Foulum biogas plant (Denmark) digesting cow and pig manure, maize silage and industrial by-products. The biofibers were separated with a 2-mm sieve and frozen at 18 °C. Batches of these biofibers taken at different times were used for the treatments. For treatments with enzymes, steam and a combination of these, the biofibers were washed before being frozen. 2.2. Treatments 2.2.1. Chemical treatment The biofibers were treated with calcium oxide (CaO) from Faxe Kalk (Denmark). CaO was added to the biofibers to obtain concentrations of 6%, 8%, 10% w/w on wet weight (WW) basis. Controls without CaO addition were included. The treatments were made in duplicates, at 15 °C and lasted 25 days (Table 1). Every fifth day, the biofibers were mixed and samples were retrieved. 2.2.2. Biological treatment 2.2.2.1. Partial aerobic treatment. The aerobic inocula used for the partial aerobic treatment were compost from garden waste from Århus Affaldscenter (Denmark) and fungi collected from straw and maize silage stored outdoor in Foulum (Denmark) for a period of 6 months. Aerobic inoculum and biofibers were mixed in proportions of 2% and 10% w/w on WW basis. Controls without inoculum were included. Atmospheric air was supplied from the bottom of the biofibers–inoculum mixture with a Güde 215/8/24 compressor at a flow of 0 (no aeration) and 280 ml (min kg TS)1 (TS is the total solids content of the mixture). The flow was measured with a flow meter Gallus 2100 G 1.6 TCE. A 2-mm sieve was used to avoid the blocking of the air tube and to ensure homogenous air distribution. Experiments were carried out in duplicates at 27 °C. The treatment lasted 20 days and sampling and mixing took place every other day.

Table 1 Treatments and corresponding conditions.

Chemical

CaO

Biological

Partial aerobic Enzymatic

Physical Steam Combined

Size reduction H3PO4 NaOH Steam + H3PO4, laccase Steam + NaOH, laccase

Procedure

Temperature (°C)

Reaction time

6%, 8%, 10% w/w WW Compost, fungi Laccase, cellulase, hemicellulase 2 mm 4% w/w TS

15

0–25 d

27 37

0–20 d 20 h

– 160

– 15 min

160, 37

15 min, 20 h

2.2.2.2. Enzymatic treatment. The enzymes tested were laccase (E.C. 1.10.3.2) and a mixture of cellulases and hemicellulases with cellulase being the main activity (E.C. 3.2.1.4). The enzymes were commercial products DeniLite II S (laccase), Novozym 51003 (laccase), Novozym 342 (cellulase) from Novozymes (Denmark), Laccase EN204 from JenaBios (Germany). Treatments with DeniLite II S and Novozym 342 were made at concentrations of 0.5, 5.0, 10.0, 20.0 U (g TS)1 and 0.3, 0.5, 1.0, 2.0 U (g TS)1, respectively. The same concentrations were used for treatments with the combination of the two enzymes. To elucidate the effect of the pH on enzymatic treatment, the pH was adjusted to 4.0, 5.5, 6.0, 7.0 by adding H3PO4. An additional experiment with DeniLite II S alone at dosage 66 U (g TS)1 at pH 4.0 was made. The treatments with Novozym 51003 were made at 60 U (g TS)1, with and without the addition of a mediator (syringaldazine 0.8 and 1.5 mmol l1), at pH 5.5 and 7.0. The treatments with Laccase EN-204 were made at 84 U (g TS)1, with and without the addition of a mediator (ABTS 1.7 mmol l1). The pH was adjusted to 5.0 with a citrate–phosphate buffer solution. All the enzymatic treatments were made at 37 °C. Control experiments with enzymes without addition of biofibers and with biofibers but without enzymes were included. The treatments lasted 20 h, under continuous mixing. Oxidative enzymes such as laccase use oxygen as the electron acceptor, therefore the enzymatic treatments were made with continuous oxygen supply bubbling air with a membrane pump. The treated material was separated into a solid and a liquid fraction with a 1-mm sieve. 2.2.3. Physical treatment As physical treatment, the size reduction of the fibers was tested. A kitchen blender Braun K600 was used to reduce the size of the fibers (2 mm, based on an approximate particle size distribution). 2.2.4. Steam treatment The effect of steam treatment was studied in combination with a catalyst. Phosphoric acid (H3PO4) and sodium hydroxide (NaOH) were used as catalysts. Substrate concentration (12.9% TS), temperature (160 °C), time of treatment (15 min), WW of substrate (200 g) and the concentration of the catalyst (4% w/w TS) were the same for all steam treatments. The unit for steam treatment was as described by Bruni et al. (in press). 2.2.5. Steam treatment followed by enzymatic treatment Enzymatic treatment was applied to the solid fraction of the steam-treated material. The combined treatment was compared to a mere enzymatic treatment as control. The enzyme laccase Novozym 51003 was used for the combined treatment. The enzyme dosage for the enzymatic treatment of the steam-treated material with H3PO4, NaOH and for the control experiment was 50, 48 and 59 U (g TS)1, respectively. H3PO4 was used to adjust the pH of the substrate and was dosed until pH 5.5 was reached, under continuous mixing. The enzymatic treatments lasted 20 h, at 37 °C, under continuous mixing and oxygen supply (bubbling air with a membrane pump). 2.3. Methane potential assays Methane potential was determined in batch assays (infusion bottles of 543 ml total volume, 200 ml inoculum, at 52 °C), as described by Angelidaki et al. (2009). The inoculum for the batch assays was the effluent from a thermophilic biogas plant (52 °C) digesting cow manure. For the samples with high TS content (untreated biofibers or solid fraction from treated biofibers), 10 g WW of substrate were used. Batches digesting the complete treated mixture of solid fraction and liquid fraction were prepared

8715


2.4. Analyses and calculations Total nitrogen TKN and ammonium nitrogen NH4-N (Kjeldahl-N method), TS, volatile solids (VS) were measured according to the standard methods (APHA, 1998). The organic nitrogen content (expressed as proteins) was calculated from TKN and NH4-N as:

300

240

ml CH4 (g VS)-1

adding 40 g WW of the solid and liquid fractions with the same proportions as in the treated material (on a WW basis). The methane potential of inactivated enzymes (enzymes heated to 100 °C for 3 h) was measured. Blank batches containing only inoculum and control batches with pure amorphous cellulose as substrate were included. The batch assays were done in triplicates.

180

120

60

0

organic N ¼ ðTKN—NH4 -NÞ 6:25

0

5

10

15

20

25

30

days

3. Results and discussion 3.1. Effects on the methane potential Treatments with CaO and NaOH resulted in the highest methane yield increases (Table 2). We obtained the highest methane yield (239 and 234 ml CH4 (g VS)1) by treating biofibers for 10 days with 6% and 8% CaO, respectively (Fig. 1). The lower methane yield that derived from the treatment with 10% CaO compared to treatments with lower CaO dosages may be due to the formation of calcium–lignin complexes that caused lower lignin removal and consequently lower methane yield increase. Xu et al. (2010) reported a decrease of lignin solubilization because the treatment with lime caused interaction between negatively charged lignin molecules and positively charged calcium ions. Among the two catalysts used for steam treatment, NaOH resulted in the highest methane yield increase (49 ml CH4 (g VS)1) and had the highest conversion rate (Fig. 2). Steam treatment with NaOH may have converted part of the lignin into acetic acid (we detected acetic

Table 2 Effect on methane yield.

Chemical Biological Physical Steam Combined a b

CaO Partial aerobic Enzymatic Size reduction H3PO4 NaOH steam + H3PO4 + laccase Steam + NaOH + laccase

VS of treated material. WW of untreated biofibers.

Variation% of yield ml CH4 (g VS)1a

Variation% of yield m3 CH4 (t WW)1b

+59% No effect No effect +8% +6% +38% +24% +69

+66% No effect No effect +10% +8% +26% +18% +34%

no CaO added

6% CaO

8% CaO

10% CaO

Fig. 1. Chemical treatment with CaO, specific methane yield of treated biofibers.

200 180 160 140

ml CH4 (g VS)-1

The concentration of volatile fatty acids (VFA) in the biofibers was measured adding 20 ml of H3PO4 0.5 mol l1 to 20 g of sample and analyzing the VFA content of 1 ml liquid (Kaparaju et al., 2009). A gas chromatograph equipped with a flame ionization detector (FID) was used to monitor the methane production (Hansen et al., 2004). Gas measurements are reported in STP conditions (Standard Temperature and Pressure, 273 K, 101,325 Pa). Methane and energy yields are expressed as m3 CH4 (t WW)1 and kWh (t WW)1, respectively, where WW indicates the wet weight of untreated biofibers. The thermal energy content of the methane was calculated using the lower calorific value 50.1 MJ (kg CH4)1. The mass balance was made on VS basis as described by Cullis et al. (2004).

120 100 80 60 40 20 0 0

20

40

60

80

100

days untreated biofibers

steam + H3PO4

steam + NaOH

Fig. 2. Steam treatment with catalyst, specific methane yield of whole treated mixture (solid fraction + hydrolysate).

acid, 11.32% and 0.33% of TS in the steam-treated liquid fraction and in the steam-treated solid fraction, respectively), while steam treatment with H3PO4 addition may have only reallocated lignin (Kaparaju and Felby, 2010). It is reported that oxidative treatments in alkaline conditions convert carbohydrates and lignin into carboxylic acids (Schmidt and Thomsen, 1998). In this study, steam treatment with NaOH probably had a similar effect regarding acetic acid formation. The presence of the acetic acid in the steamtreated material explains the high conversion rate of this material into methane, as acetic acid can directly be utilized by aceticlastic methanogens. Steam treatment with H3PO4 addition increased the methane yield of biofibers to a lower extent than the other treatments (8 ml CH4 (g VS)1). Our results showed a lower improvement of the biodegradability of biofibers compared to results

8716


3.3. Considerations for full-scale applications

g NH4-N (kg VS)-1

10.00 8.00 6.00 4.00 2.00 0.00 0

5

10

15

20

25

30

days no CaO added

6% CaO

Fig. 3. Chemical treatment with CaO, NH4-N in the biofibers.

obtained by other researches with H3PO4 on sugar cane bagasse and corn stover (Geddes et al., 2010; Um et al., 2003). This can be explained with differences in the materials treated. In our study, the biofibers had previously been digested in a biogas reactor, thus the easily degradable organic material had already been removed and we used a lower concentration of catalyst compared to some of the previous researches. Combined steam treatment using catalyst followed by treatment with laccase further increased the methane yield by 2.0 ± 0.5 and 1.7 ± 0.4 m3 CH4 (t WW)1 compared to steam treatment alone with H3PO4 and NaOH, respectively, while enzymatic treatment alone did not improve the biodegradability significantly (19.8 ± 0.4 m3 CH4 (t WW)1). This suggests that the tight association of lignocellulose did not allow effective enzymatic treatment. Steam treatment can increase the porosity of the biofibers and can thus make the substrate more accessible for the enzymes. Interestingly, treating the steam-treated material with laccase, without adding a mediator increased the methane yield. Probably, laccase sufficiently oxidized lignin to an extent, which was high enough to improve the biodegradability of the biofibers, although it is reported that laccase in the absence of a mediator can only oxidize small fractions of lignin (Widsten and Kandelbauer, 2008). Similar findings without addition of a mediator were obtained by Palonen and Viikari (2004). It is possible that oxidizing substances, such as solubilized or colloidal lignin, contained in the steam-treated material acted as mediators in the enzymatic reactions (Felby et al., 1997; Grönqvist et al., 2003). 3.2. Overall mass balance Among the treatments tested, only steam treatment with NaOH caused mass losses (10%). This treatment formed the hydrolysate (steam-treated) with the highest VS content (12%), while the hydrolysate from steam treatment with H3PO4 contained approximately 5% of the total steam-treated VS. Untreated biofibers contained 11.2 ± 0.4 g N/(kg WW)1 as organic nitrogen. Steam treatment with H3PO4 and NaOH released 19% and 53% of the organic nitrogen into the hydrolysate, respectively. Untreated biofibers did not contain VFA and we detected acetic acid only in biofibers treated with steam with NaOH. Based on of the overall effect of the treatment on the mass and nitrogen distribution, on the acetic acid formation and on the methane yield increase, steam treatment with NaOH was more aggressive compared to steam treatment with H3PO4. The highly-reactive OH anions that were released with steam treatment with NaOH reacted with different organic molecules, while the protons H+ from steam treatment with H3PO4 could hydrolyze mainly hemicellulose (Mousavioun and Doherty, 2010). Chemical treatment with CaO increased the rate of ammonia volatilization from the biofibers (Fig. 3).

Chemical treatment with CaO has a significant potential to decrease the concentration of ammonia in the substrate. Because of the risk of inhibition of the microorganisms involved in the biogas process, removal of ammonia may be required at full-scale biogas plants digesting substrates such as manure that have a high content of organic nitrogen or ammonia. Although the chemical treatment with CaO resulted in the highest methane yield gain and increased the rate of NH4-N volatilization, the advantage of this treatment has to be evaluated carefully. The analysis will have to take into account the costs of chemicals and the need for extra investments such as mixers (thorough mixing is required to ensure homogenous distribution of CaO on the biofibers) and storage (the storage volume is proportional to the reaction time of the treatment). Steam treatment with NaOH resulted in a lower methane yield increase compared to the chemical treatment with CaO. The low dosage of chemicals and the short reaction time make this treatment however very interesting for full-scale biogas processes. The energy input for steam treatment may be available from waste heat at full-scale biogas plants equipped with gas engines (Pickworth et al., 2006). Considering economical and environmental aspects, the catalyst NaOH is more expensive than CaO and for reasons of soil pollution it is not desirable for the effluent of the biogas process, should this be used as a fertilizer (Wyman et al., 2005). Although the optimal dosage for NaOH would need to be investigated, steam treatment with NaOH addition is preferred to steam treatment with H3PO4, due to the higher methane yield increase. The combined steam treatment with NaOH followed by laccase treatment resulted in extra methane production corresponding to 68 kWh (t WW)1, compared to untreated biofibers, and 16 kWh (t WW)1 compared to biofibers that were treated with steam with catalyst NaOH. Considerations should be made regarding the additional costs of the laccase treatment (separation of the solid fraction after steam treatment, pH adjustment and enzymatic treatment) relative to the extra energy yield. The methane yield increase due to mechanical treatment was lower than the one obtained by Angelidaki and Ahring (2000). This underlines that although mechanical treatment is a straightforward treatment that can be implemented at full-scale biogas plants, the costs of the energy input may be high in relation to the extra energy yield (Taherzadeh and Karimi, 2008). Our treatments with aerobic microorganisms did not improve the methane yield of the biofibers. We tested these microorganisms to aerobically treat the lignocellulose for a short time to initiate decomposition of the lignocellulosic structure and to increase its biodegradability, avoiding that aerobes achieve the oxidation of the holocellulose. Other researchers applied treatments with microorganisms to substrates such as rice straw, office paper, agricultural waste and kitchen waste. Positive results were reported from studies made with pure cultures of microorganisms (Dhouib et al., 2006; Kurakake et al., 2007; Schober and Trösch, 2000; Srilatha et al., 1995), but the use of pure cultures of microorganisms may not be possible for full-scale applications, unless the microorganisms can be cheaply cultivated at the biogas plant.

4. Conclusions We identified different methods to increase the methane yield of biofibers from digested manure. Chemical treatment (CaO) and steam treatment with NaOH resulted in the highest methane yield increases (66% and 26%, respectively). Because steam treatment with NaOH addition released 12% of the VS into the hydrolysate, the whole steam-treated mixture (solid fraction + hydrolysate)


should be digested for biogas production. Steam treatment with NaOH requires further optimization, but it has high potential for full-scale applications. Enzymatic treatment improved the methane yield of biofibers from manure only when combined with a steam treatment. Laccase alone without addition of external mediator increased the methane yield of steam-treated biofibers, enhancing the effect of steam treatment.

Acknowledgements We thank the Danish Agency for Science Technology and Innovation for financially supporting the project.

References Angelidaki, I., Ahring, B.K., 2000. Methods for increasing the biogas potential from the recalcitrant organic matter contained in manure. Water Sci. Technol. 41, 189–194. Angelidaki, I., Alves, M., Bolzonella, D., Borzacconi, L., Campos, J.L., Guwy, A.J., Kalyuzhnyi, S., Jenicek, P., van Lier, J.P., 2009. Defining the biomethane potential (BMP) of solid organic wastes and energy crops: a proposed protocol for batch assays. Water Sci. Technol. 59, 927–934. APHA, 1998. Standard Methods for the Examination of Water and Wastewater, 20th ed. American Public Health Association, Washington, DC. Bruni, E., Jensen, A.P., Angelidaki, I., in press . Steam treatment of digested biofibers for increasing biogas production. Bioresour. Technol., doi: 10.1016/ j.biortech.2010.04.064. Cullis, I.F., Saddler, J.N., Mansfield, S.D., 2004. Effect of initial moisture content and chip size on the bioconversion efficiency of softwood lignocellulosics. Biotechnol. Bioeng. 85, 413–420. Demirbas, A., 2008. Products from lignocellulosic materials via degradation processes. Energy Sources Part A 30, 27–37. Dhouib, A., Ellouz, M., Aloui, F., Sayadi, S., 2006. Effect of bioaugmentation of activated sludge with white-rot fungi on olive mill wastewater detoxification. Lett. Appl. Microbiol. 42, 405–411. Felby, C., Nielsen, B.R., Olesen, P.O., Skibsted, L.H., 1997. Identification and quantification of radical reaction intermediates by electron spin resonance spectrometry of laccase-catalyzed oxidation of wood fibers from beech (Fagus Sylvatica). Appl. Microbiol. Biotechnol. 48, 459–464. Fernandes, T.V., Klaasse Bos, G.J., Zeeman, G., Sanders, J.P., van Lier, J.B., 2009. Effects of thermo-chemical pre-treatment on anaerobic biodegradability and hydrolysis of lignocellulosic biomass. Bioresour. Technol. 100, 2575– 2579. Geddes, C.C., Peterson, J.J., Roslander, C., Zacchi, G., Mullinnix, M.T., Shanmugam, K.T., Ingram, L.O., 2010. Optimizing the saccharification of sugar cane bagasse using dilute phosphoric acid followed by fungal cellulases. Bioresour. Technol. 101, 1851–1857. Grönqvist, S., Buchert, J., Rantanen, K., Viikari, L., Suurnäkki, A., 2003. Activity of laccase on unbleached and bleached thermomechanical pulp. Enzyme Microb. Technol. 32, 439–445. Hansen, T.L., Schmidt, J.E., Angelidaki, I., Marca, E., Jansen, J.C., Mosbæk, H., Christensen, T.H., 2004. Method for determination of methane potentials of solid organic waste. Waste Manag. 24, 393–400.

8717

Hartmann, H., Angelidaki, I., Ahring, B.K., 2000. Increase of anaerobic degradation of particulate organic matter in full-scale biogas plants by mechanical maceration. Water Sci. Technol. 41, 145–153. Hatakka, A., 1994. Lignin-modifying enzymes from selected white-rot fungi: production and role in lignin degradation. FEMS Microbiol. Rev. 13, 125–135. Jin, Y., Hu, Z., Wen, Z., 2009. Enhancing anaerobic digestibility and phosphorus recovery of diary manure through microwave-based thermochemical pretreatment. Water. Res. 43, 3493–3502. Kaparaju, P., Felby, C., 2010. Characterization of lignin during oxidative and hydrothermal pre-treatment processes of wheat straw and corn stover. Bioresour. Technol. 101, 3175–3181. Kaparaju, P., Serrano, M., Thomsen, A.B., Konjan, P., Angelidaki, I., 2009. Bioethanol, biohydrogen and biogas production from wheat straw in a biorefinery concept. Bioresour. Technol. 100, 2562–2568. Kurakake, M., Ide, N., Komaki, T., 2007. Biological pretreatment with two bacterial strains for enzymatic hydrolysis of office paper. Curr. Microbiol. 54, 424–428. Mousavioun, P., Doherty, W.O., 2010. Chemical and thermal properties of fractionated bagasse soda lignin. Ind. Crops Prod. 31, 52–58. Ohkuma, M., 2003. Termite symbiotic system: efficient bio-recycling of lignocellulose. Appl. Microbiol. Biotechnol. 61, 1–9. Palonen, H., Viikari, L., 2004. Role of oxidative enzymatic treatments on enzymatic hydrolysis of softwood. Biotechnol. Bioeng. 86, 550–557. Pickworth, B., Adams, J., Panter, K., Solheim, O.E., 2006. Maximising biogas in anaerobic digestion by using engine waste heat for thermal hydrolysis pretreatment of sludge. Water Sci. Technol. 54, 101–108. Schmidt, A.S., Thomsen, A.B., 1998. Optimization of wet oxidation pretreatment of wheat straw. Bioresour. Technol. 64, 139–151. Schober, G., Trösch, W., 2000. Degradation of digestion residues by lignolytic fungi. Water Res. 34, 3424–3430. Srilatha, H.R., Nand, K., Babu, K.S., Madhukara, K., 1995. Fungal pretreatment of orange processing waste by solid-state fermentation for improved production of methane. Process Biochem. 30, 327–331. Sun, Y., Cheng, J., 2002. Hydrolysis of lignocellulosic materials for ethanol production: a review. Bioresour. Technol. 83, 1–11. Taherzadeh, M.J., Karimi, K., 2008. Pretreatment of lignocellulosic wastes to improve ethanol and biogas production: a review. Int. J. Mol. Sci. 9, 1621–1651. Talebnia, F., Karakashev, D., Angelidaki, I., 2010. Production of bioethanol from wheat straw: an overview on pretreatment, hydrolysis and fermentation. Bioresour. Technol. 101, 4744–4753. Tanaka, S., Kobayashi, T., Kamyiama, K., Bildan, M.L.N.S., 1997. Effects of thermochemical pretreatment on the anaerobic digestion of waste activated sludge. Water Sci. Technol. 35, 209–215. Um, B., Karim, M., Henk, L., 2003. Effect of sulfuric and phosphoric acid pretreatments on enzymatic hydrolysis of corn stover. Appl. Biochem. Biotechnol. 105–108, 115–125. Wang, H., Wang, H., Lu, W., Zhao, Y., 2009. Digestibility improvement of sorted waste with alkaline hydrothermal pretreatment. Tsinghua Sci. Technol. 14, 378–382. Widsten, P., Kandelbauer, A., 2008. Laccase applications in the forest products industry: a review. Enzyme Microb. Technol. 42, 293–307. Wyman, C.E., Dale, B.E., Elander, R.T., Holtzapple, M., Ladisch, M.R., Lee, Y.Y., 2005. Coordinated development of leading biomass pretreatment technologies. Bioresour. Technol. 96, 1959–1966. Xu, J., Cheng, J.J., Sharma-Shivappa, R.R., Burns, J.C., 2010. Lime pretreatment of switchgrass at mild conditions for ethanol production. Bioresour. Technol. 101, 2900–2903. Zheng, M., Li, X., Li, L., Yang, X., He, Y., 2009. Enhancing anaerobic biogasification of corn stover through wet state NaOH pretreatment. Bioresour. Technol. 100, 5140–5145.




Effects of pretreatment methods for hazelnut shell hydrolysate fermentation with Pichia Stipitis to ethanol Yesßim Arslan *, Nurdan Eken-Saraçog˘lu Department of Chemical Engineering, Gazi University, Maltepe 06570, Ankara, Turkey

a r t i c l e

i n f o

Article history: Received 19 November 2009 Received in revised form 18 May 2010 Accepted 26 May 2010

Keywords: Pichia stipitis Hazelnut shell hydrolysate Fermentation Ethanol Pretreatment

a b s t r a c t In this study, we investigated the use of hazelnut shell as a renewable and low cost lignocellulosic material for bioethanol production for the first time. High lignin content of hazelnut shell is an important obstacle for such a biotransformation. Biomass hydrolysis with acids yields reducing sugar with several inhibitors which limit the fermentability of sugars. The various conditioning methods for biomass and hydrolysate were performed to overcome the toxicity and their effects on the subsequent fermentation of hazelnut shell hydrolysate by Pichia stipitis were evaluated with shaking flasks experiments. Hazelnut shells hydrolysis with 0.7 M H2SO4 yielded 49 g l1 total reducing sugars and fermentation inhibitors in untreated hydrolysate. First, it was shown that several hydrolysate detoxification methods were solely inefficient in achieving cell growth and ethanol production in the fermentation of hazelnut shell hydrolysates derived from non-delignified biomass. Next, different pretreatments of hazelnut shells were considered for delignification and employed before hydrolysis in conjunction with hydrolysate detoxification to improve alcohol fermentation. Among six delignification methods, the most effective pretreatment regarding to ethanol concentration includes the treatment of shells with 3% (w/v) NaOH at room temperature, which was integrated with sequential hydrolysate detoxification by overliming and then treatment with charcoal twice at 60 °C. This treatment brought about a total reduction of 97% in furans and 88.4% in phenolics. Almost all trialed treatments caused significant sugar loss. Under the best assayed conditions, ethanol concentration of 16.79 g l1 was reached from a hazelnut shell hyrolysate containing initial 50 g total reducing sugar l1 after partial synthetic xylose supplementation. This value is equal to 91.25% of ethanol concentration that was obtained from synthetic D-xylose under same conditions. The present study demonstrates that Pichia stipitis is able to grow and ferment sugars to ethanol in detoxified hazelnut hydrolysate derived from delignified biomass. Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction With the depletion of the world’s petroleum supply, there has been an increasing worldwide interest in alternative, non-petroleum based sources of energy. Ethanol derived from biomass has the potential to be renewable transportation fuel that can replace gasoline (Kim and Holtzapple, 2005). The use of bioethanol as a source of energy would be more than just complementing solar, wind and other renewable energy sources in the long run (Lin and Tanaka, 2006). Moreover bioethanol can play an important role in reducing green house gas emission. Ethanol use will increase because of its biodegradable, renewable and performance qualities (Kumar et al., 2009). Ethanol is a high performance fuel in internal combustion engines and burns relatively cleanly, especially as the amount of gasoline with which it is blended decreases (Lynd, 1996). * Corresponding author. E-mail address: [email protected] (Y. Arslan). 0960-8524/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2010.05.085

The largest potential feedstock for ethanol is lignocellulosic biomass, which includes materials such as agricultural residues, forest residues, wood, grass, waste paper and municipal wastes (Noureddini and Byun, 2010). Basically, the lignocellulosic biomass composed of cellulose, hemicellulose and lignin. Cellulose and hemicellulose are composed of mixture of carbohydrate polymers. Lignocellulosic biomass is an attractive material for bioconversion to ethanol because they are renewable, widespread and cheap. One of the advantages of bioconversion with lignocellulosic materials is the opportunity to create biorefinery producing value-added co-products plus fuel bioethanol (Balat et al., 2008). The main processing challenge in the ethanol production from this resource is the feedstock pretreatment (Sánchez Óscar and Cardona, 2008). The biological process for converting the lignocellulose to fuel ethanol requires: (1) delignification to liberate cellulose and hemicellulose from their complex with lignin (2) depolimerization of the carbohydrate polymers to produce free sugars and (3) fermentation of mixed hexose and pentose sugars to produce ethanol. The

Y. Arslan, N. Eken-Saraçog˘lu / Bioresource Technology 101 (2010) 8664–8670

hydrolysis of lignocellulosic materials is usually catalyzed by acids and enzymes. In general acid hydrolysis of lignocellulose is conducted with mineral acids such as H2SO4 or HCl (Laopaiboon et al., 2010). In acid hydrolysis, the hemicellulose fraction is depolymerized at lower temperature than the cellulosic fraction to xylose and other sugars (Chandel et al., 2007). The use of xylosefermenting yeasts such as Pichia stipitis that are able to co-ferment xylose, glucose, mannose, galactose and cellobiose offers an opportunity for efficient utilization of hemicellulose derived hydrolyzates (Hernandez-Salas et al., 2009). Acid hydrolysis has some limitations. If higher temperatures or longer residence time are applied, the hemicellulosic derived monosaccharides will degrade and give rise to fermentation inhibitors. The major toxic compounds include furfural and hydroxymethylfurfural (sugar degradation products), acetic acid (released from the hemicellulosic structure), several aromatic and phenolic compounds (lignin degradation products) and metallic ions (Klinke et al., 2004). These compounds affect microbial metabolism and fermentative process (Carvalho et al., 2004). The type and concentration of these compounds in hemicellulose hydrolysates depend upon the type of raw material and hydrolysis conditions employed and their toxicity is the major factor limiting the bioconversion of lignocelluose material (Hendriks and Zeeman, 2009; Laopaiboon et al., 2010). Physical, physico-chemical, chemical and biological processes have been used for pretreatment of lignocellulosic materials (Sun and Cheng, 2002). Several techniques have been reported to overcome the inhibitory effect of these compounds during fermentation by yeasts, including adaptation of microorganism to the medium (Roberto et al., 1991), treatments with molecular sieves, ion-exchange resins or charcoal (Gong et al., 1993), steam stripping and over titration (Roberto et al., 1994; 1995). Pretreatment must improve the formation of sugars, avoid degradation or loss of carbohydrates, avoid the formation of by-products inhibitory to the subsequent hydrolysis and fermentation processes and be costeffective. The utility of various agricultural residues for the production of organic fuels and chemicals has enormous potential for the commercial applications. Numerous studies for developing largescale production of ethanol from lignocellulosics have been carried out (Sánchez Óscar and Cardona, 2008). The possible utilization of inexpensive new feedstock requires the extensive evaluation of the pretreatment conditions which is related to the nature and composition of lignocellulosic material. Studies have shown that pretreatment is the most significant step for the cost and the success of the cellulosic bioethanol technology (Balat et al., 2008). Hazelnut shell can be one of the most important types of biomass, as it is an abundant, important agricultural and commercial material in Turkey. At present, two-thirds of the world production capacity of hazelnuts is provided by Turkey, with around 250 thousand tons of hazelnut shells per year (equivalent to 4.63 109 MJ) produced in the Black-Sea region of Turkey alone (Dog˘ru et al., 2002). Elemental analysis of hazelnut shells was reported as 45.59% C, 4.59% H, 38.14% O, 1.26% ash and 10.07% moisture (Midilli et al., 2000). Proximate analysis indicates that hazelnut shells have 43.1% lignin, 27.5% hemicellulose, 24.7% cellulose, 3.4% alcohol-benzene extractives and 1.4% ash. Today, its main utilization remains as boiler fuel. Burning agricultural residues may causes air pollution, soil erosion and a decrease in soil biological activity (Çöpür et al., 2007). Therefore any possible usage of hazelnut shells will yield economic as well as environmental dividends. The conversion of hazelnut shell to useful chemicals such as acetic acid, methanol (Asßık et al., 1977), ammonia (Corlett, 1975), furfural (Demirbasß, 2006) and hydrogen (Midilli et al., 2000) has also been investigated. No known effort has been made to utilize hazelnut shell as a renewable and low cost lignocellulosic material for bioethanol production. High lignin content of hazelnut shell is an important obstacle for such a biotransformation.

8665

The main objective of this work was to study effects of pretreatment methods on hazelnut shell hydrolysate fermentation to ethanol by Pichia stipitis. First the effect of hydrolysate detoxification methods on the fermentability of hazelnut shell hydrolysate derived from non-delignified biomass was investigated. Next different pretreatments of hazelnut shells for delignification were also employed before hydrolysis in conjunction with hydrolysate detoxification to improve alcohol fermentation. 2. Methods 2.1. Microorganism and growth media Pichia stipitis NRRL Y-7124 was grown at 30 °C on agar slants. The medium described by Slininger et al., (1982) was used for growth. Inocula were prepared by transferring organism by loop from two days slants to 250 ml Erlenmeyer flasks containing 100 ml of growth medium. The inoculum medium consisted of (g l1) 6.4 urea, 1.2 KH2PO4, 0.18 Na2HPO4, 10 yeast extract and 50 D-xylose (pH = 4.5). CaO, ZnO, FeCl3.6H2O, MgO, CuSO4.5H2O, CoCl2.6H2O, H3BO3 were also included in the growth medium as trace metals. The yeast was incubated aerobically with a magnetic stirrer (600 rpm) at 30 °C for 27 h prior to use. 2.2. Raw material Hazelnut shells used in experiments were obtained from a local plant in Düzce province in Turkey. Hazelnut shells were milled into fine particles and screened into fractions (1.4–0.63 mm) to easy reaction with acid. To reduce water content hazelnut shells were dried in an oven at 105 °C then 5.24% moisture content was measured. The Standard TAPPI method (Tappi, 1978) was applied to determine the pentosan, which was found on dry basis as 29.26% (w/w). 2.3. Delignification of hazelnut shell To minimize the concentrations of fermentation inhibitors, hazelnut shells were submitted to one of six delignification methods before acid hydrolysis (Table 1). 2.4. Preparation of hydrolysate samples Hazelnut shell hemicellulosic hydrolysate was prepared from non-delignified and delignified hazelnut samples in a glass reactor equipped with a mechanically driven glass stirrer. The acid hydrolysis was conducted for 220 min with 0.7 M H2SO4 and a solid/liquid ratio of 1/5 (w/v). The temperature of hydrolysis was 90 °C and agitation was 100 rpm. At the end of hydrolysis, two fractions were obtained and separated by filtration: a solid residue and a

Table 1 Delignification methods. Treatment

Delignification methods

I II III IV V VI

1% NaOH pretreatment at room temperaturea 1% NaOH pretreatment in an autoclaveb 3% NaOH pretreatment at room temperaturea 3% NaOH pretreatment in an autoclaveb 10% Hypochlorite pretreatment at room temperaturea 10% Hypochlorite pretreatment in an autoclaveb

a Hazelnut shells were mixed with 1%(w/v) NaOH or 3%(w/v) NaOH or 10%(w/v) hypochlorite solution by stirring at 1000 rpm and 18 °C for 2 h. Then filtrated hazelnut shells were washed until the wash water was neutral. b Hazelnut shells were mixed with 1%(w/v) NaOH or 3%(w/v) NaOH or 10%(w/v) hypochlorite solution and kept in an autoclave for 2 h at 129 °C and 2.57 atm. Then filtrated hazelnut shells were washed until the wash water was neutral.


8666

hydrolysate. Recovery of sugars from lignocellulose solubilization was measured as the reducing sugar which primarily may contain xylose and arabinose from hemicellulose and also glucose and cellobiose from cellulose. The untreated hydrolysate resulting from non-delignified biomass contains 49 g total reducing sugar l1 before any detoxification of hydrolysate. Delignification treatments also cause degradation or decomposition of polysaccharides in biomass which lowers the sugar recovery in hydrolysis step (Hendriks and Zeeman, 2009). Detoxification treatment of hydrolysates prior to fermentation also results in different sugar loss. To see exact effect of the remaining inhibitory constituents and to produce a hydrolysate with a suitable sugar concentration for fermentation, initial reducing sugar level was adjusted to 50 g l1 with synthetic xylose supplementation into hydrolysates before each fermentation. The treated hydrolysates were further supplemented with additional nutrients to maintain the growth medium composition. The hazelnut shell hydrolysate and the remaining nutrients were separately autoclaved and combined after sterilization. Synthetic media were prepared by dissolving known amounts of xylose in water and complemented with the same nutrients as hemicellulosic hydrolysates to compare the fermentation parameters.

2.7. Analytical methods Biomass concentration was determined turbidiometrically at 600 nm and converted to the biomass dry weight. Total reducing sugars were determined colorimetrically using dinitrosalicylic acid reagent (Miller, 1959). Ethanol was measured by the dichromate oxidation method which is based on the complete oxidation of ethanol by dichromate in the presence of sulfuric acid (Horwitz, 1980). Total furans in hydrolysate samples were estimated by a spectrophotometric method based on the difference in absorbance at 284 and 320 nm using a Hach DR/4000 spectrophotometer (Hach Company, PO Box 389, Loveland, CO 8059) (Martinez et al., 2000) before and after any pretreatment. Absorbances of each hydrolysates at 280 nm were also measured with a spectrophotometer and compared with the absorbance of the untreated hydrolysate. From the results, R280 were calculated by Eq. (1) as shown below. From this value the effect of any treatment on removing the furan and phenolic compounds from the hydrolysate in comparison to untreated hydrolysate can be approximately evaluated (Miyafuji et al., 2002).

R280 ¼

Absorbance of the hydrolysate after any treatment Absorbance of the untreated hydrolysate

ð1Þ

2.5. Detoxification of hydrolyzates and integrated treatments The hydrolysates prepared from non-delignified biomass were subjected to many detoxification methods (Treatments A, B, C, D, E, F and G) before fermentation. These seven treatments were summarized in Table 2. Two step treatment methods (Treatments I–VI) as cleaning of hazelnut shells before hydrolysis followed by detoxification of hydrolysate were given in Table 3. Six different delignification treatments were performed for conditioning of hazelnut shells prior to acidic hydrolysis. For hydrolysates derived from delignified hazelnut shells, a single detoxification method was employed in sequential manner after delignification of biomass. Each hydrolysate was treated with CaO (76.66 g l1) until pH 10 and filtered and acidified to pH 6 with concentrated sulfuric acid. After that overlimed hydrolysate was mixed with charcoal (20 g l1) and stirred during one hour at 60 °C twice. The hydrolysate was recovered with filtration.

2.6. Shaking flask experiments All treated hydrolysates containing additional nutrients were fermented by Pichia stipitis at 30 °C in 250 ml shake flasks having 135 ml medium at 150 rpm and pH 6. Hydrolysate fermentation medium was inoculated with 10% (v/v) growth culture.

3. Results and discussion 3.1. Effect of hydrolysate detoxification Several by-products of sugar and lignin degradation formed during the hydrolysis process are toxic to the microbial metabolism and have negative effect on the fermentation efficiency. The inhibitors such as furfural, 5-HMF (hydroxymethylfurfural) and total phenols were considered toxic to microorganism. Acid hydrolysis of non-delignified hazelnut shell caused a release of reducing sugar of 49 g l1 along with fermentation inhibitors (Table 2) such as furans 116 mg l1 in untreated hydrolysate. R280 of untreated hydrolysate was assigned as unity to see the effect of any treatment on removing the furans and phenolic compounds. In order to overcome adverse effects, in a first series of experiments, various detoxifications of hydrolysate samples were performed following acidic hydrolysis with non-delignified biomass. Recovered reducing sugar level, total furan left and R280 after each detoxification were summarized in Table 2. It is well known that the effectiveness of a detoxification method depends on the type of toxic constituents present in hydrolysate which varies according to the raw and hydrolysis conditions (Carvalho et al., 2006). It seems that overliming (Treatment B) led to a drastic change in furans and phenolics comparing to neutralization (Treatment A).

Table 2 Detoxification methods of hydrolysates prepared from non-delignified hazelnut shells.

Untreated Treatment Treatment Treatment Treatment Treatment Treatment Treatment Treatment

A B C D E F G H

Hydrolysate detoxification

Recovered reducing sugar after hydrolysis (g l1)

Total furan (mg/l)

R280

None Neutralization to pH 6 with CaO Overliming with CaOa 10 g l1 charcoal treatment one hour at 30 °C, pH 2 and overliming 20 g l1 charcoal treatment one hour at 30 °C, pH 2 and overliming 30 g l1 charcoal treatment one hour at 30 °C, pH 2 and overliming Overliming and 20 g l1 charcoal treatment one hour at 30 °C, pH 6 Overliming and ethyl acetate treatment one hour with a volume ratio of 1:1 Overliming and 20 g l1 charcoal treatment one hour at 60 °C, pH 6 twice

49.00 43.00 38.00 35.40 35.00 35.00 32.52 33.57 26.30

116.0 97.0 57.0 29.0 7.8 9.7 5.0 25.8 1.1

1.000 0.896 0.567 0.302 0.134 0.142 0.056 0.288 0.021

a CaO addition until pH 10 and filtered and acidified to pH 6 with concentrated sulfuric acid. Analysis were carried out in triplicate. The values are mean of three replicates. Standard deviation was within 10%.


8667

Table 3 Delignification methods of hazelnut shells and effects of different treatments. Hazelnut shell treatment

Treatment I Treatment II Treatment IIIb Treatment IV Treatment V Treatment VI

1 wt.% NaOH pretreatment at room temperature 1 wt.% NaOH pretreatment in an autoclave 3 wt.% NaOH pretreatment at room temperature 3 wt.% NaOH pretreatment in an autoclave Hypochlorite pretreatment at room temperature Hypochlorite pretreatment in autoclave

Recovered reducing sugar after hydrolysis (mg l1)

Total furan (mg l1)

R280

Recovered reducing sugar after detoxification (g l1)

a

Total furan in hdrolysate after detoxification (mg l1)

R280

27.54

60.4

0.64

21.30

2.55

0.029

28.86

78.4

0.80

13.45

3.42

0.035

26.40

82.6

0.83

13.61

2.70

0.027

27.17

93.0

0.95

17.10

2.94

0.037

49.00

114.0

1.00

47.24

97.35

0.465

40.00

95.0

1.00

33.73

9.03

0.100

a

Overliming with CaO and charcoal treatment at 60 °C twice. Treatment III experiment was performed in duplicate. All analysis was carried out in triplicate. The values are mean of three replicates. Standard deviation was within 10%. b

Combination of overliming method with other treatments (in Treatments C–H) seems to yield a markedly beneficial effect in the removal of inhibitors. Furans concentrations and R280 values in hydrolysate were found to be decreased in different amounts with the sequential use of charcoal treatment before and after overliming. Table 2 shows that the efficiency of charcoal treatment in removal of furan and phenolic compounds (R280) is changing with charcoal amount, order of treatments, treatment temperature and treatment number (Treatments C, D, E, F, and H). It is clear that ethyl acetate extraction (Treatment G) wasn’t as effective as charcoal in removing of phenolics in hazelnut shell hydrolyzate. After each treatment significant sugar loss took place. Sugar removals varied for each detoxification procedure employed. All lime additions (Treatments B–H) caused a progressive decline in sugar content, which correlated with an increase in pH after overliming. Above pH 9.0 the rate of sugar loss increased dramatically. Amartey and Jeffries (1996) found overliming resulted in the loss of D-glucose (14%), D-xylose (4%), L-arabinose (8%) and acetic acid (43%). Treatment H which is detoxification with overliming in combination with twice charcoal treatment promotes the highest inhibitor removal and sugar loss. Even though, at the beginning of the fermentations of hydrolysates subjected to Treatments B, C, D, E, F, and G the concentration of furan was below the thresholds of inhibition (Almeida et al., 2007) and phenolic compounds content were reduced to 2.1–56.7% of untreated hydrolysate level no significant cell growth and ethanol production were observed with very little sugar metabolization. Only, fermentation of hazelnut hydrolysate subjected to Treatment H gave 5 g l1 ethanol while cell growth was nearly inhibited and 48% sugar consumption occurred (Arslan, 2007). Twenty-one other hydrolysate detoxification methods performed (Arslan, 2007) not mentioned here had no positive impact on fermentability of hazelnut shell hydrolysate. Vogel-Lohmeier et al. (1998) reported the high sensitivity of Pichia stipitis to inhibitors and the effect of toxic compounds on xylose metabolism of yeasts is very complex. They found some evidence of synergistic inhibition effect of toxic compounds to cell metabolism. Other researchers also demonstrated synergistic effects of HMF and furfural (Taherzadeh et al., 2000). However recently it was reported that Pichia stipitis reduces the aldehyde group in the furan ring of HMF and furfural (Liu et al., 2005). The inhibition could probably be attributed to phenolic compounds. Lignin derived phenolic compounds such as vanillin and syringaldehyde (Tran and Chambers, 1986) were identified as some of the most toxic components.

Phenolic inhibitors may act on biological membranes, causing loss of integrity (Heipieper et al., 1994) which can explain the absence of growth in hydrolysates derived from hazelnut shells with very high lignin content while furan levels were below 100 mg l1 in our study. This result brought the delignification of hazelnut shells before hydrolysis into consideration. 3.2. Effect of combination of delignification methods and detoxification Lignin content of hazelnut shell is very high (43.1% w/w). Fermentation trials in previous section indicated that the lignin content in hazelnut shell is relatively more important than other structural feature on fermentation performance. In order to decrease the influence of toxins coming from lignin, hazelnut shells were conditioned before acidic hydrolysis. Six different delignification treatments of hazelnut shells were performed (Table 1). Hydrolysis conditions (acid concentration, temperature, solid/liquid ratio, time) of delignified biomass were same with non-delignified biomass. Comparison of recovered total reducing sugars in acidic hydrolysate (26.40–28.86 g reducing sugar l1) derived from alkaline treated biomass (Treatments I–IV) and with total reducing sugars in untreated hydrolysate (49 g reducing sugar l1) showed a strong evidence of polysaccharides loss as well as lignin solubilization (Hendriks and Zeeman, 2009). Xylans can be selectively removed by peeling and hydrolytic reactions in alkaline treatment of hazelnut shell which leads to losses of fermentable sugar for ethanol production. Prior to fermentation, all hydrolysate samples were detoxified with Treatment H which promoted the best results in fermentations with non-delignified biomass hydrolysates in previous section. Sequential application of pretreatment methods caused severe reduction in total sugar left in hydrolysates. Fig. 1 shows sugar loss as percent of reducing sugar (49.0 g l1) in original untreated hydrolysate after each treatment stage. Total loss in reducing sugar concentration varied between 3.57% and 72.63%. It is noteworthy that the most reducing sugar loss occurred in Treatments II and III while Treatment V led to minimal substrate diminishment. It was clear that the loss of fermentable sugars is the major drawback of delignification and detoxification treatments. Fig. 1 also show total furan and phenolic removals caused by various treatments. All delignification methods were capable of reducing notably the concentrations of furan except hypochlorite pretreatment at room temperature in Treatment V. It can be seen that as a means of total furan and phenolics removal the best


Sugar Loss or Furan Removal or Removal of phenolics ( % of untreated hydrolysate)

8668

100 Detoxification De-lignification

90 80 70 60 50 40 30 20 10 0 I

II

III

IV

V

VI

Treatments Fig. 1. Loss of reducing sugar concentration, removal of total furan and phenolics.

delignification treatment was found as 1% NaOH pretreatment at room temperature in Treatment I. Delignification method in Treatment V was inefficient both for furan and phenolic removals. However, when hypochlorite used in autoclave furan removal was observed (Treatment VI). This indicated that hypochlorite treatment was only effective with high temperatures for furan removal. Integration of delignification and detoxification in Treatments I, II, III, and IV provided the best results. 3.3. Fermentation of Hazelnut Hydrolysate Before the fermentation all treated hydrolysates were supplemented with synthetic xylose to reach 50 g l1 reducing sugar concentrations. Composition of different fermentation media used in this report were described in Table 4. Figs. 2–4 show time course for biomass growth, total reducing sugar concentration and ethanol level in treated hydrolysates and synthetic medium with Pichia stipitis. As expected, Treatment V and Treatment VI were inefficient for biomass growth and ethanol production with very low sugar consumption. This result emphasizes the importance of the effec-

tiveness of feedstock pretreatment before hydrolysis. Results show that cell growth was close for hydrolysates when Treatments I–IV were employed and inferior compared to the synthetic medium (Fig. 2). Independent of the treatments, similar patterns of reducing sugar consumption were observed for hydrolysates subjected to Treatments I–IV (with mixed sugars) and for control run (with xylose only) (Fig. 3). It should be pointed out that delignification of hazelnuts prior to hydrolysis and detoxification of hydrolysates before fermentation in Treatments I–IV results in a better fermentability but still less than synthetic medium (Fig. 4). The fermentation performance of Pichia stipitis in treated hydroysates were also compared in Table 4. In studying ethanol production by Pichia stipitis, the fermentation of hydrolysate prepared from hazelnut shells delignified with 3% NaOH at room temperature and detoxified with overliming and charcoal (Treatment III) resulted maximum ethanol concentration as 16.79 g l1. This treatment brought about a total reduction of 88.4% in furans and 97% in phenolics. Maximum alcohol level reached with hydrolysate is equal to 91.25% of ethanol concentration that was obtained from synthetic xylose under the same conditions (18.40 g l1) with this

Table 4 Fermentation parameters. Run No

a b c

Treatment

1

I

2

II

3

IIIc

4

IV

5

V

6

VI

Control

–

Hydrolysatea Identity

A mixture of hazelnut hydrolysate (21.30 g reducing sugar l1) and externally added xylose (28.70 g l1) A mixture of hazelnut hydrolysate (13.45 g reducing sugar l1) and externally added xylose (36.55 g l1) A mixture of hazelnut hydrolysate (13.61 g reducing sugar l1)and externally added xylose (36.39 g l1) A mixture of hazelnut hydrolysate (17.10 g reducing sugar l1) and externally added xylose (32.9 g l1) A mixture of hazelnut hydrolysate (47.24 g reducing sugar l1)and externally added xylose (2.76 g l1) A mixture of hazelnut hydrolysate (33.73 g reducing sugar l1)and externally added xylose (16.27 g l1) Xylose (50 g l1) 1

Ratio of reducing sugar originated from hazelnut shell

Ethanol produced

0.43

Yp/s (max ethanol/consumed reducing sugar)

Qp (max ethanol conc /time)

10.12

0.282

0.112

0.27

16.10

0.415

0.179

0.27

16.79

0.432

0.186

0.34

15.64

0.449

0.174

b

(g l1)

0.95

4.440

0.218

0.035

0.68

4.67

0.334

0.052

0.44

0.194

0.00

18.4

Total reducing sugar: 50 g l . Cultivation conditions employed: 135 ml media in 250 ml flasks, 150 rpm, and pH 6, 30 °C. Treatment III and Control run experiments were performed in duplicate. The values are mean of two replicates. Standard deviation was within 10%.


I

II

III

IV

V

VI

8669

control

14 12

Biomass (g/l)

10 8 6 4 2 0 0

24

48

67

90

125

t (h) Fig. 2. Effect of detoxification methods employed on growth.

I

II

III

IV

V

IV

control

Total reducing sugar (g/l)

60 50 40 30 20 10 0 0

24

48

67

90

t (h) Fig. 3. Effect of detoxification methods employed on reducing sugar consumption.

I

II

III

IV

V

VI

control

25

Ethanol (g/l)

20 15 10 5 0 0

24

48

67

90

125

t (h) Fig. 4. Effect of detoxification methods employed on ethanol formation.

method ethanol yield was 0.432 g g1. Fermentation results of hydrolysates subjected to Treatment II and IV were comparable to Treatment III. The fermentation parameters such as ethanol yields (YP/S) and productivity (QP) are also given in Table 4. Treatment IV gave best outcome regarding to ethanol yield which is 0.449 gg1. The most efficient treatment for productivity Treatment III with this method productivity was 0.186 g l1 h1. Same conditions with synthetic medium, productivity was 0.194 g l1 h1. It appears in Table 4 that ethanol levels as well as yields and productivities for

hydrolysate fermentations may depend on substrate composition in hydrolysate. Even though initial reducing sugar concentration was maintained at 50 g l1 by supplementation of xylose in each hydrolysate prior to fermentation, carbohydrate constituents were not same. Slininger et al. (2008) in their work with Pichia sitipitis suggested carbon sources may change the ability of a culture to resist and adapt to inhibitor stress additional to other fermentation conditions. It is not clear that beside very low residual inhibitory compounds whether initial high concentration of xylose in the

8670


hydrolysates treated with Treatments II, III, IV played a role in the direction of carbon toward ethanol production. The present study demonstrates that some pretreatments of hazelnut shell before the hydrolysis result in a better growth of microorganism and fermentability of hydrolysate. However, the fermentation parameters for these treated hydrolysates were generally lower than those obtained with synthetic medium. The raw material and process steps before fermentation will strongly influence the composition and the fermentability of the hydrolysate. It is difficult to make comparison but the result of this study is in the range of those obtained by other researchers. In literature, ethanol yields of 0.13–0.44 g g1 and maximum ethanol concentration of 3.1–18 g l1 were reported by different investigators (Wilson et al., 1989; Nigam, 2002; Agbogbo and Wenger, 2007; Kumar et al., 2009). 4. Conclusions For the first time in this study, hazelnut shells containing high lignin content were evaluated as substrates for ethanol production. The results of fermentation of hazelnut shell hydrolysate by Pichia stipitis were presented regarding to the effect of pretreatment methods of hazelnut shells and hazelnut shell hydrolysate. The present study demonstrates that Pichia stipitis is able to grow and ferment sugars to ethanol in detoxified hazelnut shell hydrolysate derived from delignified biomass. Removing of inhibitors with pretreatments is essential and key factor in the biological conversion of hazelnut shells to ethanol. A wide range of pretreatments targeting for conditioning of hydrolysate or hazelnut shells have been trialed. The results indicated that hydrolysate detoxification such as overliming and charcoal shaking was seen an effective way to remove some toxins but not solely enough for the growth of yeast and alcohol production in hazelnut shell hydrolysate. A good fermentability could be achieved if lignin content of shells was removed with 3% NaOH treatment before hydrolysis of biomass. These delignification and detoxification treatments were carefully integrated to realize better cellulosic ethanol production from hazelnut shells. Ethanol yield obtained from this method was 0.084 g ethanol/g hazelnut shell. Acknowledgements We would like to thank Dr.Mübeccel Ergun and Müjgan TelliOkur, Faculty of Engineering, Gazi University for their valuable efforts in the review of this paper and helpful suggestions. References Agbogbo, F.K., Wenger, K.S., 2007. Production of ethanol from corn stover hemicellulose hydrolysate using Pichia stipitis. J. Ind. Microbiol. Biotechnol. 34, 723–727. Almeida, J.R.M., Modig, T., Petersson, A., Hähn-Hägerdal, B., Liéden, G., GorwaGrauslund, M.F., 2007. Increased tolerance and conversion of inhibitors in lignocelllulosic hydrolysates by Saccharomyces cerevisiae. J. Chem. Technol. Biotechnol. 82, 340–349. Amartey, S., Jeffries, T., 1996. An improvement in Pichia stipitis fermentation of acid hydrolyses hemicellulose achieved by overliming (calcium hydroxide treatment) and strain adaptation. World J. Microbiol. Biotechnol. 12, 281–283. Arslan, Y., 2007. The utilization of hazelnut shell for ethanol production. Ph.D. Thesis, Gazi University Instıtue of Science and Technology. Asßık, M., Deymer, J., Gülensoy, H., 1977. Utilization of hazelnut shell. Chim. Acta. Turc. 5, 27–42. Balat, M., Balat, H., Öz, C., 2008. Progress in bioethanol processing. Progr. Energ. Combust. Sci. 34, 551–573. Carvalho, W., Canilho, L., Mussatto, S.F., Dragone, G., Morates, M.L.V., Solenzal, A.I.N., 2004. Detoxification of sugarcane bagasse hemicellulosic hydrolysate with ionexchange resins for xylitol production by calcium alginate-entrapped cells. J. Chem. Technol. Biotechnol. 79, 863–868.

Carvalho, G.B.M., Musasatto, S.I., Cândido, E.J., Silva, J.B.A., 2006. Comparison of different procedures for the detoxification of eucalyptus hemicellulosic hydrolysate for use in fermentative processes. J. Chem. Technol. Biotechnol. 81, 152–157. Chandel, A.K., Kapoor, R.K., Singh, A., Kuhad, R.C., 2007. Detoxification of sugarcane bagasse hydrolysate improves ethanol production by Candida shehatae NCIM 3501. 2007. Bioresour. Technol. 98, 1947–1950. Corlett, R.F., 1975. Conversion of seattle’s solid waste to methanol or ammonia. The trend in engineering, University of Washington D. Çöpür, Y., Güler, C., Akgül, M., Tasßcıog˘lu, C., 2007. Some chemical properties of hazelnut husk and its suitability for particleboard production. Build. Environ. 42, 2568–2572. Demirbasß, A., 2006. Furfural production from fruit shells by acid catalyzed hydrolysis. Energ. Sourc. 28, 157–165. Dog˘ru, M., Howarth, C.R., Akay, G., Keskinler, B., Malik, A.A., 2002. Gasification of hazelnut shells in a down draft gasifier. Energy 27, 415–427. Gong, C.S., Chen, C.S., Chen, L.F., 1993. Pretreatment of sugar cane bagasse hemicellulose hydrolyzate for ethanol production by yeast. Appl. Biochem. Biotechnol. 39 (40), 83–88. Heipieper, H.J., Weber, F.J., Sikkema, J., Keweloh, H., De Bont, J.A.M., 1994. Mechanisms of resistance of whole cells to toxic organic solvents. Trends Biotechnol. 12, 409–415. Hendriks, A.T.W.M., Zeeman, G., 2009. Pretreatments to enhance the digestibility of lignocellulosic biomass. Bioresour. Technol. 100, 10–18. Horwitz, W., 1980. Dicromate oxidation methods. J. of AOAC. 13, 25–27. Kim, S., Holtzapple, M.T., 2005. Lime pretreatment and enzymatic hydrolysis of corn stover. Bioresour. Technol. 96, 1994–2006. Klinke, H.B., Thomson, A.B., Ahring, B.K., 2004. Inhibition of ethanol-producing yeast and bacteria by degradation products produced during pre-treatment of biomass. Appl. Microbiol. Biotechnol. 66, 10–26. Kumar, A., Singh, L.K., Ghosh, S., 2009. Bioconversion of lignocellulosic fraction of water- hyacinth (Eichhornia crassipes) hemicellulose acid hydrolysate to ethanol by Pichia stipitis. Bioresour. Technol. 100, 3293–3297. Laopaiboon, P., Thani, A., Leelavatcharamas, V., Laopaiboon, L., 2010. Acid hydrolysis of sugarcane bagasse for lactic acid production. Bioresour.Technol. 101, 1036– 1043. Lin, Y., Tanaka, S., 2006. Ethanol fermentation from biomass researches: current state and prospects. Appl. Microbiol. Biotechnol. 69, 627–642. Liu, Z.L., Slininger, P.J., Gorsick, S.W., 2005. Enhanced biotransformation of furfural and hydroxymethylfurfural by newly developed ethanologenic yeast strains. Appl. Biochem. Biotechnol. 121 (124), 451–460. Lynd, L.R., 1996. Overview and evaluation of fuel ethanol from cellulosic biomass: technology, economics, the environment and policy. Annu. Rev. Energ. Environ. 21, 403–465. Martinez, A., Rodriguez, M.E., York, S.W., Preston, J.F., Ingram, L.O., 2000. Use of UV absorbance to monitor furans in dilute acid hydrolysates of biomass. Biotechnol. Progr. 16, 637–641. Midilli, A., Rzayev, P., Olgun, H., Ayhan, T., 2000. Solar hydrogen production from hazelnut shells. Int. J. Hydrogen Energ. 25, 723–732. Miller, G.L., 1959. Use of dinitrosaliciyle acid reagent for reducing sugar. Anal. Chem. 31, 426–430. Miyafuji, H., Danner, H., Neuretier, M., Thoasser, C., Bvochora, J., Szola, O., Braun, R., 2002. Detoxification of wood charcoal for increasing the fermentability of hydrolysates. Enzyme Microb. Technol. 6234, 1–5. Nigam, J.N., 2002. Bioconversion of water-hyacinth (Eichhornia crapsipes) hemicellulose acid hydrolysate to motor fuel ethanol by xylose-fermenting yeast. J. Biotechnol. 97, 107–116. Noureddini, H., Byun, J., 2010. Dilute-acid pretreatment of distillers’ grains and corn fiber. Bioresour. Technol. 101, 1060–1067. Roberto, I.C., Mancilha, I.M., Souza, C.A.D., Felipe, M.G.A., Sato, S., Castro, H.F.D., 1994. Evaluation of rice straw hemicellulose hydrolysate in the production of xylitol by Candida guilliermondii. Biotechnol. Lett. 16, 1211–1216. Roberto, I.C., Felipe, M.G.A., de Mancilha, I.M., Vitola, M., Sato, S., Silva, S.S., 1995. Xylitol production by Candida guilliermondii as an approach for the utilization of agroindustrial residues. Bioresour. Technol. 51, 255–257. Sánchez Óscar, J., Cardona, C.A., 2008. Trends in biotechnological production of fuel ethanol from different feedstocks. Bioresour. Technol. 99, 5270–5295. Slininger, P.J., Bothast, R.J., Cau Wenberge, J.E.V., Kurtzman, C.P., 1982. Conversion of D- xylose to ethanol by the yeast Pachysolen tannophilus. Biotechnol. Bioeng. 24, 371–384. Slininger, P.J., Gorsich, S.W., Liu, Z.L., 2008. Culture nutrition and physiology impact the inhibitor tolerance of the yeast Pichia stipitis NRRL Y-7124. Biotechnol. Bioeng. 102, 778–790. Sun, Y., Cheng, J., 2002. Hydrolysis of lignocellulosic materials for ethanol production a review. Bioresour. Technol. 83, 1–11. Taherzadeh, M.J., Gustafsson, L., Niklasson, C., Liéden, G., 2000. Physiological effects of 5- hydroxymethylfurfural on Saccharomyces cerevisiae. Appl. Microbiol. Biotechnol. 53, 701–708. Tappi, 1978. Pentosans in wood and pulp, Tappi standart. Tran, A.V., Chambers, R.P., 1986. Ethanol fermentation of red oak acid prehydrolysate by yeast Pichia stipitis NRRL Y-7124. Enzyme Microb. Technol. 8, 439–444. Wilson, J.J., Deschalelets, L., Nishikawa, N.K., 1989. Comparative fermentability of steampretreated aspenwood hemicellulose by Pichia stipitis CBS 5776. Appl. Microb. Biotechnol. 31, 592–596.


Contents lists available at SciVerse ScienceDirect


Process and technoeconomic analysis of leading pretreatment technologies for lignocellulosic ethanol production using switchgrass Ling Tao a,⇑, Andy Aden a, Richard T. Elander a, Venkata Ramesh Pallapolu b, Y.Y. Lee b, Rebecca J. Garlock c,d, Venkatesh Balan c,d, Bruce E. Dale c,d, Youngmi Kim e, Nathan S. Mosier e, Michael R. Ladisch e, Matthew Falls f, Mark T. Holtzapple f, Rocio Sierra g, Jian Shi h, Mirvat A. Ebrik h, Tim Redmond h, Bin Yang h, Charles E. Wyman h, Bonnie Hames i, Steve Thomas i, Ryan E. Warner j a

National Bioenergy Center, National Renewable Energy Laboratory, 1617 Cole Blvd., Golden, CO 80401, USA Department of Chemical Engineering, Auburn University, 212 Ross Hall, Auburn, AL 36849, USA c Biomass Conversion Research Laboratory, Department of Chemical Engineering and Materials Science, Michigan State University, 3900 Collins Road, Lansing, MI 48910, USA d Great Lakes Bioenergy Research Center, Michigan State University, East Lansing, MI, USA e LORRE, Department of Agricultural and Biological Engineering, Purdue University, 500 Central Dr., West Lafayette, IN 47907, USA f Department of Chemical Engineering, Texas A&M University, 3122 TAMU, College Station, TX 77843-3122, USA g Universidad de los Andes Chemical Engineering Department Grupo de Conversion de Energia, Bogotá, Colombia h Center for Environmental Research and Technology, Department of Chemical and Environmental Engineering, Bourns College of Engineering, University of California at Riverside, 1084 Columbia Avenue, Riverside, CA 92507, USA i Ceres, Inc., 1535 Rancho Conejo Blvd., Thousand Oaks, CA 91320, USA j Genencor, A Danisco Division, 925 Page Mill Road, Palo Alto, CA 94304, USA b

a r t i c l e

i n f o

Article history: Available online 4 August 2011 Keywords: Pretreatment Enzymatic hydrolysis Biomass Switchgrass Process economics

a b s t r a c t Six biomass pretreatment processes to convert switchgrass to fermentable sugars and ultimately to cellulosic ethanol are compared on a consistent basis in this technoeconomic analysis. The six pretreatment processes are ammonia fiber expansion (AFEX), dilute acid (DA), lime, liquid hot water (LHW), soaking in aqueous ammonia (SAA), and sulfur dioxide-impregnated steam explosion (SO2). Each pretreatment process is modeled in the framework of an existing biochemical design model so that systematic variations of process-related changes are consistently captured. The pretreatment area process design and simulation are based on the research data generated within the Biomass Refining Consortium for Applied Fundamentals and Innovation (CAFI) 3 project. Overall ethanol production, total capital investment, and minimum ethanol selling price (MESP) are reported along with selected sensitivity analysis. The results show limited differentiation between the projected economic performances of the pretreatment options, except for processes that exhibit significantly lower monomer sugar and resulting ethanol yields. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction The objective of biomass pretreatment is to make the cellulose and hemicellulose accessible to enzymes without unnecessary degradation of sugars to unusable compounds. As the first step in the biochemical conversion process to produce bioethanol, pretreatment plays a critical role in preparing biomass for enzymatic conversion to C5 and C6 sugars and, in some processes, directly hydrolyzing a portion of structural carbohydrates to oligomeric and monomeric sugars. The Consortium for Applied Fundamentals and Innovation was formed in 2000 (Elander et al., 2009) to collaboratively develop and publish data on leading biomass pretreatment options. The pretreatment step has been projected to be

⇑ Corresponding author. E-mail address: [email protected] (L. Tao). 0960-8524/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2011.07.051

one of the most expensive capital investments in the biochemical conversion process and has significant impacts on the downstream conversion steps, such as enzymatic hydrolysis and fermentation. Technoeconomic analysis (TEA) has been used as an important tool to facilitate research and development within collaborative projects conducted by the CAFI team. The first CAFI project (CAFI 1) used a consistent source of corn stover to determine glucose and xylose yields from various pretreatment processes upon pretreatment and subsequent enzymatic hydrolysis. Each pretreatment process TEA was conducted using these reported sugar yields (Eggeman and Elander, 2005). The CAFI project team recently extended this approach to switchgrass in the CAFI 3 project. In this paper, a process economic analysis of the six pretreatment processes studied in CAFI 3 is reported, including glucose and xylose yields, total capital investment, and MESP for an integrated, commercial-scale lignocellulosic ethanol process based on each pretreatment.

11106

L. Tao et al. / Bioresource Technology 102 (2011) 11105–11114

Enzyme Biomass

Hydrolyzate

Beer

Products

Fig. 1. Simplified flow diagram of the overall process.

2. Methods 2.1. Process design Developing process economics requires significant amounts of experimental data, modeling toolsets, and engineering expertise. Performing such a process economics study requires a wide range of detailed information, including (1) a conceptual level of process design to develop a detailed process flow diagram (based on rigorous conversion data), (2) rigorous material and energy balance calculations via commercial simulation tools such as Aspen Plus (ApsenTech, Cambridge, MA), (3) capital and project cost estimation (via in-house model using spreadsheets), (4) a discounted cash flow economic model (via in-house model using spreadsheets), and (5) a final calculation of minimum biofuel selling price. 2.1.1. Process overview of 2002 National Renewable Energy Laboratory (NREL) biochemical conversion design This process (Aden et al., 2002) uses various pretreatment technologies on lignocellulosic biomass (switchgrass), followed by enzymatic hydrolysis (saccharification) of the remaining cellulose (and possibly the remaining hemicelluloses) and fermentation of the resulting glucose and xylose to ethanol. The process design (shown in Fig. 1) also includes feedstock handling and storage, product purification, wastewater treatment, lignin combustion, product storage, and all other required utilities. In all, the process is divided into nine areas (from Area 100 to 900). 2.1.1.1. Area 100 feed handling. The feedstock, in this case milled switchgrass, is delivered to the feed handling area. Only minimum storage and feed handling are required in the current biochemical conversion process. Feedstock composition (Kim et al., in press) significantly influences the overall analysis, based on 35.0% cellulose, 22.5% xylan, 22.6% lignin dry weight. 2.1.1.2. Area 200 pretreatment and conditioning. Depending on the selected pretreatment process, the biomass is treated and then washed and/or neutralized prior to enzymatic hydrolysis. Hydrolyzate conditioning is not considered in the conceptual level of design but will be in the detailed design stage. We assumed post-washing would adjust the pH to 5.0 for all cases, so neutral-

ization or conditioning is not considered in those scenarios. No all methods require post-washing, however, to be consistent in this analysis post-washing is included. We chose a plant size of 2000 metric tons per day and 8406 operating hours per year. Equipment for Area 200 was sized from the material and energy balance calculated by simulation. All other areas are derived from the NREL 2002 design report (Aden et al., 2002) with minimal changes, except for the conversion yields in each pretreatment/ enzymatic hydrolysis combination, utilities, and related storage and handling equipment for pretreatment chemicals. Pretreatment chemical recovery systems were also included in the analyses when recycling results in an economic benefit for the overall process. Horizontal screw-feeder reaction system are assumed for DA, SO2, AFEX, and SAA pretreatment reactions, but are operated at different reaction temperature, pressure and residence time depending on the technology. The cost also various by different material of construction. For instance, DA pretreatment reactor requires Incoloy, while SAA only requires stainless steel. For LHW system, the biomass is fed (pumped) at 15–20% total solids into a long tubular reactor to supply 10–20 min of residence time at 200 °C. For lime pretreatment, high-pressure oxygen (100 psi) is charged to a closed stir tank with biomass for reduced residence times (4 h), comparing to CAFI 1 lime ‘‘pile’’ reactors.

2.1.1.3. Area 300 enzymatic hydrolysis and fermentation. Enzymatic hydrolysis is initiated in a continuous reactor using a purchased cellulase enzyme. The enzymatic hydrolysis solids level is managed at 20%1 total solids, to be kept consistent even as the pretreatment solids level varies with different technologies. The partially hydrolyzed slurry is next batched to a system of parallel anaerobic bioreactors. Hydrolysis is completed in the batch reactor, and then the slurry is cooled to 30 °C and inoculated with the glucose–xylose co-fermenting organism Zymomonas mobilis for fermentation. After 7 days of enzymatic hydrolysis and fermentation, most of the cellulose and xylose will be converted to ethanol. The resulting ethanol water mixture (beer) is sent to the product recovery train (Area 500). Oligomeric sugars (mainly gluco-oligmer and xylo-oligomer) are not converted to monomers in the enzymatic hydrolysis stage 1 Unless otherwise stated in this paper, the measurement of composition, percentage of solid, and conversion data are by weight.

11107


in the base case for all six pretreatments. Therefore in all the base cases, it is assumed that oligomer sugars pass through fermentation unreacted and end in the boiler, where their combustion heating value is recovered. Only in the study of MESP with oligomer credits are oligomer sugars assumed to be fermentable to produce ethanol (assuming no additional enzyme usage). The fermentation condition is consistent for all six pretreatments, assuming glucose yield is 92% and xylose yield is 85%. Minor sugars (arabinose, galactose, and mannose) are assumed to be non-fermentable, but heating values for the minor sugars are recovered in the boiler section in Area 800. The purchased enzyme usage to generate the experimentally achieved sugar yield for all the pretreatments was 15 FPU cellulase in Spezyme CP plus 30 CBU b-glucosidase in Novozyme 188 per g glucan in untreated raw switchgrass, equivalent to 27 mg protein per g glucan. 2.1.1.4. Area 500 product recovery. The beer is separated into ethanol, water, and residual solids by distillation and solid–liquid separation. Ethanol is distilled to a nearly azeotropic mixture with water and then purified to 99.5% using vapor-phase molecular sieve adsorption. Solids from the distillation bottoms are separated and sent to the combustor (Area 800) while the liquid is sent to wastewater treatment (Area 600). The distillation bottoms liquid is concentrated in a series of evaporations using recycled heat. The evaporated condensate is returned to the process. 2.1.1.5. Area 600 wastewater treatment. Plant wastewater streams are treated by anaerobic and aerobic digestion. The biogas (high in methane) from anaerobic digestion is sent to the combustor (Area 800). The treated water is suitable for recycling and is returned mostly to the pretreatment area (Area 200). 2.1.1.6. Area 700 storage. This area provides bulk storage for chemicals used and produced in the process, including corn steep liquor (CSL), ammonia (as well as sulfuric acid, lime, and so forth, if needed) nutrients, water, and ethanol. 2.1.1.7. Area 800 combustor, boiler, and turbogenerator. The solids from distillation and biogas from anaerobic digestion are burned in a fluidized bed combustor to produce high-pressure steam for electricity production and process heat. The majority of the process steam demand is in the pretreatment reactor and the distillation area. The boiler produces excess steam that is converted to electricity for use in the plant and for sale to the grid as a co-product. 2.1.1.8. Area 900 utilities. This area includes a cooling water system, chilled water system, process water manifold, and power systems. 2.2. Process economic analysis The process economic analysis includes the following parts. 2.2.1. Variable operating cost The variable operating cost is based on material and energy balance calculations from process modeling using Aspen Plus simulations. Raw materials include switchgrass, pretreatment and neutralization chemicals, nutrients (CSL and potassium salts), wastewater treatment chemicals and polymers, and diammonium phosphate. Utility includes steam (both low and medium pressure steam), power, water, and nitrogen gas. Major raw material and utility costs for all six pretreatment processes are listed in Table 1. Switchgrass delivery cost to the conversion process is assumed to be $69.5 per dry ton (MYPP, 2010) for cost year 2007. The overall enzyme cost is assumed to be $0.25/gal of produced ethanol, realizing that each process has different overall ethanol yields. This enzyme price does not reflect any commercial enzyme costs but is

Table 1 Raw material and utility costs. Raw material cost

Price ($2007)

Switchgrass CSL Cellulase enzyme Diammonium phosphate Propane Sulfuric acid (93%) NH3 Lime SO2 Water Power (byproduct credit)

$69.50/dry ton $205.00/ton $0.25/gal ethanol $182.30/ton $5.90/ton $32.10/dry ton $300.00/ton $89.80/ton $400.00/ton $0.40/ton $0.04/kWh

solely an estimated contribution of enzyme cost to the operating cost of ethanol production. 2.2.2. Fixed operating cost Salaries are inflated to 2007 dollars using NREL’s 2002 design basis (Aden et al., 2002). In addition to salary, general overhead is a factor of 60% applied to the salary total and covers items such as safety, general engineering, general plant maintenance, and payroll overhead. Annual maintenance materials were estimated at 2% of the total project investment, and property insurance and local property tax were estimated at 1.5% of the total project investment based on standard literature assumptions (Peters and Timmerhaus, 1991). 2.2.3. Equipment cost and installed cost Based on the material and energy balances, the equipment can be sized appropriately. Equipment costs are developed using a number of sources, including past vendor quotations (for more specialized equipment) from the corn stover ethanol biochemical design report (Aden et al., 2002), the CAFI 1 process designs (Eggeman and Elander, 2005), the United State Department of Agriculture (USDA) corn ethanol model (Tao and Aden, 2009), and the NREL equipment database, as well as costing software estimates (for simpler equipment such as distillation columns, pumps, and tanks), chemical engineering textbooks (Peters and Timmerhaus, 1991), and other database information. Equipment costs for areas outside of Area 200 were derived from the NREL 2002 design case using the power law to adjust for changes in capacity. The scaling exponent for the power law was obtained from the NREL 2002 design case (Aden et al., 2002) for most of the equipment. For equipment not listed in the NREL 2002 design case and for which we are not able to get vendor’s guidance, the exponent term is assumed 0.60. Standard NREL factors were used to obtain the total project investment from the purchased equipment costs. However, pretreatment reactor installation factors have been re-evaluated to reflect both proper scaling up and consideration of material of construction, based on common engineering judgments. Operating costs, revenues, and discounted cash flows were obtained by modifying the NREL 2002 nth plant design. 2.2.4. Discounted cash flow analysis The method for the discounted cash flow calculation in this study assumes 100% equity financing and 2.5 years construction plus 0.5 years start-up. The plant life is 20 years. The income tax is 39%. Working capital is 5% of fixed cost investment. The MESP is the minimum price that the ethanol must sell for in order to generate a net present value of zero for a 10% internal rate of return. This makes the MESP higher than a true cost of production. The cost year is 2007 (US dollars). Equipment quotations obtained in earlier or later years would be inflated or deflated to year 2007 using chemical indices. All cost numbers stated here are 2007 US

11108


dollars. For each pretreatment process, a conceptual-level process is defined and Aspen Plus simulation is performed to develop material and energy balances. Process economics are then determined based on Aspen Plus simulation results combined with the financial assumptions to calculate MESP. Several sensitivity cases are analyzed in comparison with the base case scenario for each pretreatment. It should be emphasized that a certain percentage of uncertainty exists around conceptual cost estimates. These values are best used for relative comparisons against technological variations or process improvements. Use of absolute values without a detailed understanding of the basis behind them can be misleading.

quor, is then added to dilute the total solids content to 20% for enzymatic hydrolysis. The overall conversion (stage 1 – pretreatment and stage 2 – enzymatic hydrolysis) glucose is 76.0% and the conversion of xylose is 76.1%, shown in Table 2. Overall monosugar yield is shown in Fig. 2 for stages 1 and 2. The MESP comparison is shown in Fig. 3 for AFEX with and without post-washing. The higher MESP for AFEX without postwashing is mainly due to lower achieved overall monomer sugar yield, even with lower pretreatment process (Area 200) cost. Because the reactor contributes to more than 60% of the pretreatment capital cost for the AFEX process, the removal of the washing step does not result in significant reduction in overall capital costs. 3.2. DA pretreatment process

3. Results and discussion 3.1. AFEX pretreatment process Normally AFEX uses anhydrous ammonia (NH3) as the pretreatment chemical. Due to the high loading of NH3 to dry-weightbased biomass that is required, recycling of anhydrous NH3 in the liquid phase is a crucial aspect of this pretreatment process. In this process, the biomass is fed directly to the pretreatment reactor, an extruder-like piece of equipment that mixes the biomass with injected liquid water and liquid NH3. The targeted water to dry biomass weight ratio is 0.81. Using the recycled NH3, the NH3 make-up flowrate is adjusted to a total NH3 to dry biomass weight ratio of 1.52. It is common practice to inject low pressure steam in the feed hopper to reduce the uptake of non-condensable gases (e.g., air). If both total solid level and NH3 loading are satisfied with the water to biomass ratio, indirect high-pressure steam then is used to further heat up the pretreatment reactor to the targeted temperature. The pretreatment reactor is operated at 150 °C and 30 min residence time. The effluent from the AFEX reactor is depressurized first into the knock out (KO) drum. The liquids flash and the biomass fibers are disrupted at 105 °C. The overhead vapor, which contains concentrated NH3, is compressed and condensed to a liquid and then stored in an NH3 day tank. The compressor is a multistage centrifugal compressor connected with inter-stage coolers. The inter-stage cooler is needed to keep the process stream temperatures below 150 °C, which is the limit typically imposed to prevent thermal breakdown of the compressor oils. The bottom of the KO drum is sent to a flash tank before being sent to the NH3 recovery column. Ammonia from the reactor effluent is concentrated by an absorber column of 10 theoretical trays. The recovered NH3 from the column overhead is condensed by indirect cooling and direct water quenching, then recycled back to the NH3 day tank. The bottom of the NH3 column is sent to the wastewater treatment system in Area 600. It is important to have liquid NH3, since gaseous NH3 can have minimal effect on the biomass chemistry or structure and hence is ineffective in promoting enzymatic hydrolysis (Sendich et al., 2008). Certain process condition is needed to allow hot ammonia gas to condensate in order to contact the biomass surface to achieve similar pretreatment results of liquid ammonia. Numerous pumps and compressors are utilized to compress NH3 to liquid, as well as to achieve the targeted reactor pressure. The power is integrated with the overall plant utility in Area 900. Steam consumption by indirect steam to the reactor and NH3 distillation column are integrated with boiler section in Area 800. The AFEX process with a post-washing step is used as the base case for this study. The hydrolyzate slurry (containing all the solids) from the KO drum and flash tank is separated by a pressure belt filter. The separated solid stream is washed by process water and sent to the mix tank. Additional water (evaporator condensate from Area 500), combined with liquid hydrolyzate and washed li-

Normally DA pretreatment uses dilute sulfuric acid as the pretreatment chemical at a liquid phase concentration of 0.5–2% and reaction temperatures of 120–200 °C, with a relatively short residence time (less than 60 min). This pretreatment effectively removes and recovers most of the hemicellulose as dissolved sugars, and subsequent enzymatic hydrolysis yields of glucose from cellulose increase with the increasing removal of hemicellulose, to well above 90% upon complete hemicellulose hydrolysis in pretreatment (Knappert et al., 1981). Only a small portion of lignin is typically soluble after DA pretreatment, but the lignin structure is found to be disrupted by scanning electron microscopy analysis (Donohoe and Johnson, 2010). The DA pretreatment process converts most of the hemicellulose carbohydrates in the feedstock to soluble sugars (xylose, mannose, arabinose, and glucose) by hydrolysis reactions. These sugars are primarily recovered as monomers, although less severe conditions can result in a higher relative proportion of soluble oligomeric hemicellulosic sugars. Glucan in the hemicellulose and a small portion of the cellulose are converted to glucose. In the present design, this hydrolysis conversion is accomplished using dilute sulfuric acid and heat from steam. DA pretreatment also liberates acetic acid from the hemicellulose and can form sugar degradation products like furfural and 5-hydroxymethyl furfural. In sufficiently high concentrations, these compounds can have adverse effects on the fermenting organisms. DA reaction is carried out by a horizontal screw-feed reactor followed by a blowdown tank to ambient atmosphere pressure. The pretreatment reactor system includes biomass feeders followed by a vertical section for steam heating and acid impregnation of the biomass, a horizontal screw reactor for 40 min residence time at 140 °C, and discharge into a blowdown tank. The plug screw feeder is a rugged, high-compression screw device designed to form a pressure-tight plug of feed material by axial compression. The compression action in each plug screw feeder may also squeeze out water and some of the water-soluble extractive components, depending on the incoming feedstock moisture content. Sulfuric acid is injected at the discharge spool of the plug screw feeder. The plug screw feeders meter feedstock to the pretreatment reactor and control the pressure and temperature. Low-pressure steam (appropriate to achieve the 40 psig saturated steam pressure for a 140 °C pretreatment temperature) is injected into this vessel to maintain proper temperature. The total feedstock mixture, including the injected steam, is targeted at 30% total solids. If the constraints of the reaction temperature and total solids level cannot both be satisfied by using low-pressure steam, high-pressure steam will be used. The pretreatment reactor is discharged through a reactor discharge scraper to a blowdown tank at a flash temperature of 100 °C. The flash vapor from the blowdown tank is then sent to the wastewater treatment area. The hydrolyzate slurry is discharged to the washing step using recycled process water and then goes through a solid–liquid separation step before post-wash-

11109


Table 2 Summary of conversion data for all six pretreatment processes. Method

Monomer sugar yield (%)

Glucan ? glucose yield (%)

Xylan ? xylose yield (%)

Ethanol production (MM gal/year)

Ethanol yield (gal/dry US ton feedstock)

% of Theoretical ethanol yield

AFEX DA Lime LHW SAA SO2

76.0 76.0 70.0 61.4 52.0 79.2

76.0 74.5 80.9 78.8 65.5 80.5

76.1 78.4 52.7 33.0 30.8 77.1

50.8 49.9 46.8 40.1 34.3 52.2

65.8 64.6 60.5 52.0 44.4 67.7

58.0 56.9 53.3 45.8 39.1 59.6

Fig. 2. Monomer sugar yields of stages 1 and 2.

Fig. 3. MESP sensitivity studies for AFEX, LHW and SAA pretreatments.

ing of the solids. Liquor from solid–liquid separation is sent to the neutralization tank, where ammonium hydroxide is used to raise the pH of the slurry from 1 to 5. The solids are washed by process water and then sent to enzymatic hydrolysis. The cost of the DA pretreatment reactor is typically quite high due to the requirement of corrosion resistant alloys. For the pretreatment reactor system, all parts in contact with acid (pretreatment reactor, pressurized transporter, and plug screw feeder) are constructed of Incoloy 825-clad steel. This reactor system could be overdesigned for the pretreatment process conditions considered here and is capable of operating at a significantly higher pre-

treatment severity. We chose not to adjust the cost of the reactor to account for the lower temperature or lower acid loading. It is noted that this assumption may penalize the DA pretreatment cost, especially given the high cost material of construction. For instance, it is questionable that the wall thickness and resulting reactor costs for a 40 psig reactor are the same as assumed in this study, derived from the NREL 2002 report rating of 185 psig. For DA pretreatment of switchgrass, glucose yield is 74.5% and xylose yield is 78.4%, shown in Table 2. The other sugars (such as oligomers of all sugars, monomers of minor sugars) are not considered fermentable and are burned in the combustion area to collect

11110


heating values. The current design uses ammonium hydroxide solution to neutralize the hydrolyzate. Although not considered in the base case of DA pretreatment, the high miscibility of NH3 additionally permits treatment of the whole hydrolyzate slurry and can potentially eliminate the hydrolyzate solid–liquid separation. For consistency, NH3 is only used to neutralize the hydrolyzate to pH 5. The ammonium hydroxide usage of 16 g/L of hydrolyzate reflects a stoichiometric amount of NH3 needed to neutralize all the acetic acid liberated from the feedstock and sulfuric acid when loaded at 90 mg acid per g of dry feedstock. This is based upon laboratory studies at low solids loadings and would be significantly lower at higher solids loadings in larger-scale, continuous pretreatment systems. The consumption rate of NH3 is only slightly higher than what is needed to reach pH 5. About 20% of NH3 is used to neutralize acetic acid and the rest is used to neutralize sulfuric acid. 3.3. Lime pretreatment process In the lime pretreatment process, the feedstock is mixed with lime and water. Pure oxygen (O2) is applied to the reactor to enhance delignification. In commercial scale production, high-pressure O2 (about 85 psig) is purged into the pretreatment reactor vessel, maintained at a constant pressure, and blankets the head space of the vessel. Lime pretreatment can be performed at a wide range of reaction temperatures, ranging from ambient temperature to 130 °C. To have a shorter reaction residence time, high temperature is preferred but requires use of a pressurized reactor vessel. In the CAFI 1 corn stover project, lime pretreatment uses the ‘‘pile’’ model with extended reaction time. The biomass is mixed with water and lime, then stored in a pile at ambient temperature, assumed to be 25 °C. Each pile was assumed to take 1 week to construct, 1 week to deconstruct, and to have 6 weeks of hold time; thus a minimum of eight piles are needed to maintain continuous operation. The lime to biomass (dry) ratio in the pile was set at 0.1. The water content was adjusted to give a solid content of 15% for the material leaving the pile for downstream processing. In this design, the short term lime pretreatment (Sierra et al., 2009a,b) technology studied in the CAFI 2 hybrid poplar project is used with high temperature and short residence time. NREL’s thermodynamics package was used for the thermodynamic model in Area 200. No attempt was made to model electrolyte behavior for this pretreatment. For the lime pretreatment process analyzed in the study, the switchgrass feedstock is mixed with water and lime and is then charged into the pretreatment reactor. The lime to biomass (dry) mass ratio in the reactor is set at 1.0. The water content is adjusted to give a solid level of 20% for the material leaving the pretreatment reactor for downstream processing. Total reaction time is 4 h at 120 °C. For this pretreatment chemistry, Stainless Steel 304 is chosen as the reactor material of construction. Agitation is also needed for the reactor based on bench scale experiments. Pure O2 is charged into the pretreatment reactor to improve yields. After pretreatment, the biomass slurry is washed using water. Washing is assumed sufficient to decrease pH to 7.0. The calcium carbonate left in the pretreated stover is sent to fermentation and then is processed along with the insoluble solids through Area 500. For this design, the calcium carbonate is combusted in the boiler with lignin wet cake and other fermentation residues. There is no need to add a separate calcium kiln for converting calcium carbonate to quicklime (CaO). This stream (containing mainly ash and CaO) is then separated and cooled. It then is mixed with water and then slaked to produce milk of lime (i.e., a Ca(OH)2 slurry) in the slaker in Area 800. The amount of water used for slaking was adjusted to give a 20% slurry. If the entire slurry is recycled to Area 200 in order to regenerate all the lime, then the ash is recycled as

well and will accumulate. The solubility of lime in water is relatively low (about 1.5–1.6% in water at ambient conditions). Therefore, recycling only the liquid fraction of the slurry is also not applicable due to significant loss of lime in the solid stream due to the very fine hydrated lime particles. When the slurry is agitated, the lime particles can stay in the suspension for a relatively long time even without agitation. A decanter is used in this design to separate the suspended lime slurry to be recycled to the pretreatment reactor, assuming 20% of the lime flow was to be disposed with the ash at a landfill. The disadvantage of this decanter design is the high water consumption required to carry relatively pure lime back to the pretreatment area. Additional large scale tests should be demonstrated to further prove the process feasibility of lime recovery. The feed O2 to the combustor is designed to have 17% excess. A pressure swing adsorption (PSA) unit is used for commercial-scale O2 recovery, and the cost of the PSA unit adds significantly to capital costs associated with lime pretreatment. In addition to the plant air compressors, two compressors are used for CO2 flow to the slaker and enriched O2 flow to the lime pretreatment reactor. The monosugar yields are shown in Table 2. 3.4. LHW pretreatment process Pretreatment using pressurized hot water at a controlled pH minimizes hydrolysis of the oligosaccharides while causing hydration of the biomass structure by liquid water at a pressure above saturation vapor pressure of water at the selected temperature (Wyman et al., 2005). The intent of liquid hot water (LHW) pretreatment is to avoid the formation of monomer sugars that could degrade to aldehydes at severe conditions in pretreatment and to utilize enzymatic hydrolysis to convert cellulose and soluble oligosaccharides to fermentable sugars. Advantage of LHW is low pretreatment reactor cost, but that is often counterbalanced by lower monomeric xylose yield. In the CAFI 1 corn stover study, LHW resulted in the highest MESP (Eggeman and Elander, 2005) due to low sugar and ethanol yields as a result of a high proportion of oligomers generated from hemicellulose. The switchgrass is fed to a tank where the total solids content is lowered to 20% by dilution with recycle water from the bottoms of the first evaporator. The diluted switchgrass is heated by a cross exchanger. The exchanger’s design is close to a spiral exchanger; however, multiple shells will likely be required. After the cross exchanger, the stover slurry is then heated to 200 °C by indirect steam heating in the pretreatment trim heater. The pretreatment reactor is a long tube reactor designed for 10 min of residence time. The reactor is quite similar to the retention ‘‘U-tube’’ or ‘‘trombones’’ used for starch cooking. Sufficient backpressure is maintained to keep the liquor from flashing in the reactor tube. After the specified hold time is attained in the reactor, the effluent is cooled by cross exchange with the feed. The cooled effluent is held in a flash tank, where any non-condensable gas is vented, and then the slurry is dispatched to the multiple downstream enzymatic hydrolysis and fermentation trains. The resulting glucose yield is 78.8% and xylose yield is 33.0%, shown in Table 2. A significant proportion of soluble oligomeric xylose is generated in the LHW process. Because post-washing is considered in the base case study, a wash tank, solid–liquid separator, and reslurry tank are needed prior to enzymatic hydrolysis. The process alternative without a post-washing step is studied in comparison with the LHW pretreatment base case, shown in Fig. 3. 3.5. SAA pretreatment process Soaking in aqueous ammonia (SAA) pretreatment utilizes aqueous NH3 in a fairly simple pretreatment reactor configuration.


Switchgrass is pretreated with aqueous NH3, resulting in a significant degree of delignification. About 70% of the lignin is soluble in the aqueous solution, leaving insoluble carbohydrate biomass components almost ‘‘untouched’’ to be processed in enzymatic hydrolysis. In an earlier version of this approach, known as NH3 recycle percolation (ARP), the reaction temperature is typically in a range of 150–180 °C using an aqueous NH3 concentration of 10– 15%. At high temperatures, aqueous NH3 swells biomass, depolymerizes lignin, and breaks lignin–hemicellulose bonds without degrading carbohydrates (Wyman et al., 2005), and it enhances the cellulose digestibility in the downstream step of enzymatic hydrolysis. The SAA process often uses lower temperatures than the ARP process, typically 60–180 °C. Biomass is first soaked in the aqueous NH3 solution for an extended period of time (from hours to days) and then pretreated at high temperature and pressure for a relatively short residence time. The SAA retains most of the hemicellulose in the solids, eliminating the need to separately process hemicellulose and cellulose sugars (Wyman et al., 2005). One significant impact is that a much higher demand of xylanase and/or other accessory enzymes may be required to further convert xylan to xylose monomer in the SAA pretreated solids. In the base case of the SAA process, a high percentage of xylose oligomer remains in the fermentation broth. Because the baseline study assumes oligomer sugars are not fermentable, the SAA process results in a low overall xylose yield and therefore higher MESP due to relatively low overall ethanol yields. The SAA reactors are screw-like reactors, similar to the DA pretreatment reactor design, including the feeder, the reaction section, and the discharger. Low-pressure steam is injected into the feeder to preheat the biomass. The preheated biomass, aqueous NH3, high-pressure steam, and additional dilution water are then fed to the pretreatment reactor section. The NH3 rate is set at 1.35 g pure NH3 per g dry biomass. Water content is adjusted to 2.47 g water per g dry biomass. The reactor operates at 160 °C and 465 psig for 60 min with 20% total solids. The material of construction is Stainless Steel 304, so the cost of such a reactor is less than that of the DA pretreatment reactor. High-pressure steam is injected into this vessel to maintain proper temperature. The total feedstock mixture, including the injected steam, is targeted at 20% total solids and 30% NH3 concentration. A flash tank is used to flash vapor at near atmospheric pressure to recover concentrated NH3 to the NH3 day tank. No additional separation is needed to further purify NH3 for SAA pretreatment, which is different from AFEX pretreatment. The flash vapor contains mostly NH3 and is compressed using a centrifugal compressor with two stages of inter-cooling plus after-coolers, recovering 93% of the NH3. The condensed NH3 and recovered NH3 are sent to the NH3 day tank to be re-injected into the SAA pretreatment reactor. The flash vapor contains about 20% water. A belt filter is used to separate the pretreated slurry after the flash tank. The liquid stream from the belt filter is sent to the NH3 distillation column to further recover NH3. The solids are washed by process water and reslurried with the filtrate (containing soluble sugars and soluble lignin) in a mix tank. The solids content is adjusted using NH3 recovery column bottoms in the mix tank to achieve a total solids of 20% before sending the pretreated mass to downstream enzymatic hydrolysis and fermentation. Only 5% of NH3 from the pretreatment reactor effluent is concentrated by an absorber column containing 10 theoretical trays. The recovered NH3 is recycled back to the NH3 day tank. The bottom of the NH3 column is sent to the wastewater treatment system in Area 600. An estimated 2% of NH3 is considered unrecovered in the pretreatment area, so make-up NH3 is required to feed the NH3 day tank. This loss represents about 75% of total NH3 consumption, while the remaining 25% of NH3 reacts with released acetic acid to form ammonium acetate. The ratio of total NH3 consumption is

11111

consistent with published findings for AFEX pretreatment (Chundawat et al., 2010). A lignin recovery system, as was used in the ARP design in the CAFI 1 project, is not included in this SAA design. In this process, the soluble lignin flow goes through the NH3 recovery column, to the wastewater area, then to the combustor area with the sludge of the wastewater treatment. The residual lignin heating value is recovered in the combustor area. If a lignin (coproduct) recovery system is employed, lignin can potentially be recovered as a more valuable by-product as compared to combustion. The yield data are shown in Table 2. Because production cost is driven significantly by ethanol yield, the MESP of SAA pretreatment is the highest among all six pretreatments due to the low sugar and resulting ethanol yields. However, other related work suggests that SAA pretreatment may be more effective with earlier-harvest switchgrass feedstock. It may be that SAA pretreatment cannot effectively disrupt the lignin structure found in later-harvest, aged switchgrass. A sensitivity analysis using three switchgrass species has been performed. As shown in Fig. 3, higher achieved sugar yields using Alamo and Shawnee switchgrass result in significantly lower MESP values that are comparable to the other pretreatments. 3.6. SO2 pretreatment process Steam explosion pretreatment with SO2 is related to a wellknown technology known as sulfite pulping. It can achieve both high glucose and high xylose yields when combined with enzymatic hydrolysis. High temperature and short time in the pretreatment step have been shown to favor a high enzymatic glucose yield, and low temperature and longer pretreatment time tend to favor high xylose yields (Bura et al., 2002; Ohgren et al., 2005). SO2 pretreatment enhances glucose and xylose yields in a way similar to DA pretreatment, but the purchased cost of SO2 is relatively high due to costs of shipping SO2 safely. Onsite production of SO2 is believed to be a more cost-effective in a large-scale cellulosic ethanol process and should be evaluated in future processes. The SO2 pretreatment process converts most of the hemicellulose carbohydrates in the feedstock to soluble sugars (xylose, mannose, arabinose, and glucose) by hydrolysis reactions, in a manner similar to DA pretreatment. Glucan in the hemicellulose and a small portion of the cellulose are converted to glucose, but most glucan remains in an insoluble form that requires subsequent enzymatic hydrolysis. Biomass is first impregnated with SO2 and then heated by high-pressure steam to the targeted temperature in the pretreatment reactor. The reactor design uses a horizontal screw-feed reactor followed by a blowdown tank. The pretreatment reactor system is similar to the DA reactor and includes biomass feeders followed by a vertical vessel with an appropriate residence time for steam-heating and acid impregnation of the biomass, a horizontal screw-feed reactor sized for a 10-min residence time at 180 °C, and a discharge unit to a blowdown tank. Highpressure steam is used to heat incoming biomass to the targeted reactor temperature quickly as well as to shorten the total residence time in the pretreatment system. Pressurized SO2 is injected at the discharge spool of the plug screw feeder. The plug screw feeders meter feedstock to the pretreatment reactor and control the pressure and temperature difference between it and the preimpregnation chamber. A portion of the SO2 stream may be added at the discharge of each plug screw feeder, some or all of which may be pressate from the plug screw feeders themselves. High-pressure steam is injected into this vessel to maintain proper temperature. The total feedstock mixture, including the injected steam, is targeted at 30% total solids. The reactor pressure is held just at the bubble point for the mixture. The residence time in the pretreatment reactor is 10 min at 180 °C. For the pretreatment

11112


reactor system, all parts in contact with gaseous acid (pretreatment reactor, pressurized transporter, and plug-screw feeder) are constructed of Incoloy 825-clad steel. The pretreatment reactor is discharged through a reactor discharge scraper to a blowdown tank. The pressure of the blowdown tank is controlled such that the flash operates at ambient pressure. The flash vapor from the blowdown tank is then sent to the wastewater treatment area. The hydrolyzate slurry is discharged to the washing step using recycled process water. A pressure belt filter is used in the washing area to separate liquid and solids. The liquor from the belt filter is sent to the neutralization tank where it is conditioned with ammonium hydroxide to raise the pH of the slurry from 1 to 5, suitable for enzymatic hydrolysis and fermentation. In this design, the glucose yield is 74.5%, the xylose yield is 78.4% (from both pretreatment and enzymatic hydrolysis), shown in Table 2. Overall monosugar yield is shown in Fig. 2.

3.7. Comparison of sugar and ethanol yields The total sugar yields from glucan and xylan for each stage are summarized in Fig. 2. In the base cases, only monomeric glucose and xylose are assumed as fermentable sugars to produce ethanol, so the overall ethanol yields and associated annual production levels are dependent on the monomer sugar yields. Fig. 2 summarizes monomer sugar yields of stages 1 and 2 for each case. Higher ethanol production generally results in lower MESP, so the MESP values are highly dependent on the monomer sugar yields. For instance, the SAA pretreatment has the lowest total monomer sugar yields for both glucose and xylose and results in the highest MESP (representing ethanol costs on a per gallon basis). The yearly ethanol production for the SAA case is only 34.3 million gallons, the lowest among all six cases, shown in Table 2. The SO2 pretreatment results in the highest overall monomer sugar yields (see Fig. 2) and the highest ethanol yield per dry ton of switchgrass (see Table 2), as well as low total capital on pretreatment (see Table 3). However, the MESP for the SO2 process is higher than for AFEX and DA, due to significant costs associated with handling and usage of SO2. If SO2 can be produced onsite cheaply, the MESP can be further improved. Although LHW pretreatment has a relatively high total sugar yield and the lowest pretreatment process capital cost, its MESP is not the lowest. This is because a significant amount of sugar released from the LHW pretreatment is in oligomeric form and is assumed unfermentable in the base case process (see Table 3). Based on the switchgrass composition, the theoretical ethanol yield is 87.7 million gallons per year if all the cellulose and hemicellulose is ultimately converted to ethanol. The percentage of theoretical ethanol yield for each pretreatment method is shown in the last column of Table 2. Improvement of theoretical ethanol yield should focus on the development of microorganisms that can ferment oligomer sugars along with minor sugars (arabinose, galactose, and mannose).

3.8. Comparison of pretreatment capital costs The direct capital for the pretreatment section depends strongly on the processing conditions, such as reaction temperature, residence time, solids levels, and any needed pretreatment chemical recovery strategies. The reaction temperature and residence time are defined by the CAFI research teams for each technology based on extensive bench-scale studies. The total solids levels are defined based on the bench scale process conditions but also considering reasonable process feasibility in a conceptual large scale continuous process configuration. For instance, LHW pretreatment using a tube reactor requires lower total solids level, but an assumption was made to model this pretreatment at 20% total solids. A similar assumption was made for lime pretreatment. The pretreatment chemical loading for each case is based on the bench-scale CAFI 3 pretreatment conditions. Further optimization of catalyst loading for commercial scale operation is needed for a detailed engineering design. Dilution water is calculated based on pretreatment chemical usage and assumed total solids loading in pretreatment. It should be noted that lime, SO2, and DA have a relatively higher loading of dilution water than the other pretreatments do. The dilution water loading for lime pretreatment is due to the lower total solids requirement for effective pretreatment, but for DA and SO2 pretreatments it is due to much lower chemical loading than all other cases. The solids and pretreatment chemical loadings of each pretreatment also impact the final concentration of ethanol in the fermentation broth, plant steam usage, and plant power balance, as well as capital costs of downstream processes of fermentation and ethanol recovery areas. Table 3 shows the plant steam and power usages for all six pretreatments, with all data presented on a per gallon ethanol basis except for the heating values of biomass to the boiler. SAA pretreatment has significantly higher power usage and production because the total ethanol yield is significantly lower than the other cases. Total pretreatment area capital costs, including any required pretreatment chemical recovery and recycle systems, are shown in Table 3. Lime, DA, and SAA pretreatments have higher pretreatment capital than the others do, and LHW has the lowest capital among all six pretreatment processes. Total capital for the lime pretreatment is higher than for all other cases. This is due to the cost of the lime recovery system, which uses an expensive slaking operation, and the cost of the O2 PSA unit, in addition to high cost of lime pretreatment reactors. Although one lime pretreatment reactor unit is not expensive compared to the DA pretreatment reactor, a significant number of reactors is required due to the long residence time (4 h). Because the lignin content of switchgrass is relatively high, somewhat severe pretreatment conditions are required (Sierra et al., 2009a,b), which requires both relatively higher reaction temperature and oxidative conditions of gas pressure at or above 300 psig. The longer residence time and high pretreatment reactor costs are the main causes for the high DA pretreatment capital, contributing 76% of the $45M of capital cost in the pretreatment area. The SAA pretreatment area has not only pretreatment reactors but

Table 3 Pretreatment capital cost and steam and power usages for all six pretreatments.

AFEX DA Lime LHW SAA SO2

Pretreatment capital ($MM)

% Reactor cost in pretreatment capital

% Pretreatment area in overall capital

Total installed cost ($MM)

Total capital ($MM)

Plant electricity use (kWh/gal)

Excess electricity (kWh/gal)

Plant steam use (kg steam/gal)

$31 $45 $57 $20 $45 $35

57 76 44 18 38 65

16 23 27 11 23 19

$191 $192 $212 $179 $200 $187

$348 $349 $385 $325 $364 $340

3.3 2.4 3.8 3.0 4.4 2.4

9.6 9.0 9.9 8.6 15.1 8.9

16.5 21.0 20.8 61.7 41.4 19.6


11113

Fig. 4. Total monomer and oligomer sugar yields.

also two solid–liquid separation units, a compressor, and an NH3 column to enhance the recovery of liquid NH3. On the other hand, the AFEX, SAA, and lime pretreatment processes have significant equipment requirements related to recovery of pretreatment chemicals, which contribute significantly to the overall pretreatment system capital costs, as one-pass use of the pretreatment chemical is not realistic due to high pretreatment chemical usage levels in these processes. However, the design of these pretreatment chemical recovery and recycle systems is preliminary. Further development of efficient pretreatment chemical recovery configurations may lead to lower overall pretreatment system capital costs. The pretreatment direct fixed capital for LHW is again significantly lower than for the other five cases, similar to the capital estimation in the CAFI 1 study (Eggeman and Elander, 2005). The cost of the pretreatment reactor and pretreatment chemical recovery system and the influence of downstream process capital due to different solid levels in the pretreatment step are distributed differently in each case, but result in only relatively small differences in the total direct capital costs for the overall process. Depending on the complexity and major equipment costs of each pretreatment process, the percentage of pretreatment system process capital costs varies from 11% of the

total capital investment in LHW pretreatment to 27% in lime pretreatment. The total capital costs (shown in Table 3), including both direct and indirect capital costs, indicate that the switchgrass to ethanol conversion processes are capital-intensive, ranging from $325M to $385M. Total fixed capital includes both direct cost and indirect cost (also known as added capital). The indirect costs are factored based on the direct capital costs, so it is necessary to examine the direct costs in more detail. The pretreatment, enzymatic hydrolysis, fermentation, and recovery sections of the plant are responsible for slightly less than half of the total direct fixed capital. The boiler system to recover heating values of residual biomass and lignin is about one-third of total direct fixed capital cost. 3.9. Comparison of MESP results Fig. 5 summarizes the contributions of the feedstock, other variable costs, fixed cost, depreciation, income tax, and return on capital to the overall MESP for all six base cases. All these costs are divided by the total annual ethanol production to calculate cost per gallon of produced ethanol. The other variable costs account for the cost of pretreatment chemicals, enzyme, nutrients, and

Fig. 5. MESPs with and without oligomer credits.

11114


other chemicals, as well as net electricity generation credits. For most of the cases, the other variable costs are positive due to electricity by-product credits. The SO2 pretreatment process has the highest other variable costs because the cost of SO2 usage is significant. Fixed costs include labor, maintenance, insurance, and other costs not tied to production rate. The feedstock cost contributes 45–53% of the MESP for the six cases. If the monomer sugar yields and resulting ethanol yield were higher, then the switchgrass cost per gallon ethanol produced would be lower. The MESP contribution from return on capital is high, ranging from $0.59/gal for SO2 pretreatment to $0.94/gal for SAA pretreatment. The capital depreciation is the third highest value in the MESP distribution for all six cases. The overall MESP ranges from $2.74/gal for AFEX pretreatment (with closely comparable MESPs for DA and SO2 pretreatment) to $4.07/gal for SAA pretreatment. The monomer sugar yields are the most important factor in process economics estimates. 3.10. MESP with oligomer credits If oligomer sugars can be fermented to ethanol (or readily converted to monomeric sugars at no added cost), the MESP for pretreatment processes that produce large fractions of oligomeric sugars can be improved significantly. It is possible that the xylanase activity of the enzyme preparation used for enzymatic hydrolysis could be increased with little additional enzyme cost, or that an engineered microorganism could ferment soluble oligomer sugars in the fermentation step, but at this point neither has been fully demonstrated in commercially viable lignocellulosic ethanol process configurations. The monomer and oligomer sugar yields for each CAFI base case are compared in Fig. 4. Significant amounts of total soluble sugars are found as oligomers for lime, LHW, and SAA pretreatment. If we assume that all soluble xylose and glucose sugars, both monomeric and oligomeric, can be fermented to ethanol (or oligomer sugar can be hydrolyzed to monomers via use of an enzyme preparation with appropriate oligomer-hydrolyzing activities), the MESP results for the six pretreatment options may be significantly different. The effect of oligomer conversion on MESP is shown in Fig. 5. If oligomer yields are included, significant yield improvements can be achieved in SAA, LHW, and lime pretreatments, but only minor improvements to DA, SO2, and AFEX pretreatment are found. The MESP with oligomer credits is lower than $3.00 per gallon for all cases, with differences among the pretreatments within a range of $0.62/gal. These findings indicate that if oligomeric sugars can be converted, much less differentiation exists between the projected economic performances of the CAFI 3 pretreatment options, with the possible exception of SAA pretreatment. Additional process performance data may lead to better representation of evaluation of process economics, including optimal enzyme blends for each case, operational conditions of key process steps, required additional hydrolyzate conditioning, actual fermentation performance of hydrolyzate at relevant concentrations, as well as optimal energy and water integration are not addressed in detail in this study.

4. Conclusions Six pretreatment processes for converting switchgrass to ethanol are compared in this technoeconomic analysis within the framework of the 2002 NREL biochemical design report model (Aden et al., 2002). The resulting MESPs range from $2.74 to $4.09 per gallon, and from $2.32 to $2.94 per gallon if oligomer sugar can be fermented. The six pretreatment technologies vary greatly in terms of their process design and projected total capital investment. Overall ethanol yield, which is largely based on the overall sugar yield achieved in pretreatment and subsequent enzymatic hydrolysis steps, is the single-most important factor in determining projected MESP. Acknowledgements This research was funded under the Office of the Biomass Program of the United States Department of Energy through Contract No. DE-FG36-04GO14017. We would like to acknowledge all the generous help from each CAFI 3 project PI to establish the design basis for each pretreatment method. References Aden, A., Ruth, M., et al., 2002. Lignocellulosic Biomass to Ethanol Process Design and Economics Utilizing Co-Current Dilute Acid Prehydrolysis and Enzymatic Hydrolysis for Corn Stover. Other Information. PBD: 1 June 2002; Medium: ED; Size: p. 154. Bura, R., Mansfield, S.D., et al., 2002. SO2-catalyzed steam explosion of corn fiber for ethanol production. Applied Biochemistry and Biotechnology 98, 59–72. Chundawat, S.P.S., Vismeh, R., et al., 2010. Multifaceted characterization of cell wall decomposition products formed during ammonia fiber expansion (AFEX) and dilute acid based pretreatments. Bioresource Technology 101 (21), 8429–8438. Donohoe, B.K.M., Johnson, David, 2010. Structured Biomass Image Data Management to Enable Systematic Quantitative Image Analysis. NREL Report. Eggeman, T., Elander, R.T., 2005. Process and economic analysis of pretreatment technologies. Bioresource Technology 96 (18), 2019–2025. Elander, R.T., Dale, B.E., et al., 2009. Summary of findings from the Biomass Refining Consortium for Applied Fundamentals and Innovation (CAFI): corn stover pretreatment. Cellulose 16 (4), 649–659. Knappert, D., Grethlein, H., et al., 1981. Partial acid-hydrolysis of poplar wood as a pretreatment for enzymatic-hydrolysis. Biotechnology and Bioengineering (Suppl. 11), 67–77. Kim, Y., Mosier, N.S., et al., in press. Comparative study on enzymatic digestibility of switchgrass varieties and harvests processed by leading pretreatment technologies. Bioresource Technology. doi:10.1016/j.biortech.2011.06.054. MYPP, 2010. Biomass Multi-Year Program Plan Office of the Biomass Program, Energy Efficiency and Renewable Energy, U.S. Department of Energy. Available from: . Ohgren, K., Galbe, M., et al., 2005. Optimization of steam pretreatment of SO2impregnated corn stover for fuel ethanol production. Applied Biochemistry and Biotechnology 121, 1055–1067. Peters, M., Timmerhaus, K., 1991. Plant Design and Economics for Chemical Engineers. McGraw-Hill, New York City. Sendich, E., Laser, M., et al., 2008. Recent process improvements for the ammonia fiber expansion (AFEX) process and resulting reductions in minimum ethanol selling price. Bioresource Technology 99 (17), 8429–8435. Sierra, R., Granda, C., et al., 2009a. Short-term lime pretreatment of poplar wood. Biotechnology Progress 25 (2), 323–332. Sierra, R., Granda, C.B., et al., 2009b. Biofuel: Methods and Protocols. Lime Pretreatment 581, 115–124 (Chapter 9). Tao, L., Aden, A., 2009. The economics of current and future biofuels. In Vitro Cellular and Developmental Biology – Plant 45 (3), 199–217. Wyman, C.E., Dale, B.E., et al., 2005. Coordinated development of leading biomass pretreatment technologies. Bioresource Technology 96 (18), 1959–1966.




Short Communication

The effects of four different pretreatments on enzymatic hydrolysis of sweet sorghum bagasse Jingzhi Zhang, Xingxing Ma, Jianliang Yu, Xu Zhang ⇑, Tianwei Tan Beijing Key Lab of Bioprocess, College of Life Science and Technology, Beijing University of Chemical Technology, Beijing 100029, China

a r t i c l e

i n f o

Article history: Received 12 September 2010 Received in revised form 22 December 2010 Accepted 23 December 2010 Available online 30 December 2010 Keywords: Sweet sorghum bagasse Cellulase FTIR NMR XRD

a b s t r a c t Four pretreatment processes including ionic liquids, steam explosion, lime, and dilute acid were used for enzymatic hydrolysis of sweet sorghum bagasse. Compared with the other three pretreatment approaches, steam-explosion pretreatment showed the greatest improvement on enzymatic hydrolysis of the bagasse. The maximum conversion of cellulose and the concentration of glucose obtained from enzymatic hydrolysis of steam explosion bagasse reached 70% and 25 g/L, respectively, which were both 2.5 times higher than those of the control (27% and 11 g/L). The results based on the analysis of SEM photos, FTIR, XRD and NMR detection suggested that both the reduction of crystallite size of cellulose and cellulose degradation from the Ia and Ib to the Fibril surface cellulose and amorphous cellulose were critical for enzymatic hydrolysis. These pretreatments disrupted the crystal structure of cellulose and increased the available surface area, which made the cellulose better accessible for enzymatic hydrolysis. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction Sweet sorghum is presently considered as a promising alternative crop for fuel ethanol production mainly because it can yield biomass and fermentable sugars. Sweet sorghum juice is a good and economical substrate for ethanol production (Yu et al., 2009). A large amount of sweet sorghum bagasse could also be hydrolyzed to fermentable sugars for ethanol production. FTIR, XRD, SEM, 13C NMR have been used to investigate the structure of cellulose in previous studies. The results indicated that the cellulose accessibility, the crystallinity, the size of crystallite and other structure factor of lignocelluloses may influence the enzymatic hydrolysis of cellulose (Sun et al., 2009). To increase the cellulose surface area and improve accessibility to cellulase for enzymatic hydrolysis, lignocellulose might be either physically or chemically pretreated before enzymatic hydrolysis (Mosier et al., 2005). Recently, studies of pretreatments for enzymatic hydrolysis were mainly on ionic liquids (Feng and Chen, 2008), steam explosion (Kaar et al., 1998; Chen and Liu, 2007), lime (Saha and Cotta, 2008), and dilute acid treatment (Sun and Cheng, 2005). These four pretreatment approaches improved the enzymatic hydrolysis of lignocellulose. However, the different effects of the four pretreatment approaches on lignocellulose have not been comparatively investigated. This study investigated the improvement of four pretreatment approaches on enzymatic hydrolysis of sweet sorghum bagasse and discussed the effects of the changes of cellulose structure on ⇑ Corresponding author. Tel.: +86 10 6445 0593; fax: +86 10 6441 6428. E-mail address: [email protected] (X. Zhang). 0960-8524/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2010.12.093

enzymatic hydrolysis. These results provide useful information to the commercial utilization of sweet sorghum bagasse for large scale enzymatic hydrolysis. 2. Methods 2.1. Raw material Sweet sorghum ‘‘Chuntian No. 2’’, bred by Chinese Academy of Agricultural Science, was harvested in Changping District (Beijing), October 2008. After squeezing of sweet sorghum stem, the bagasse was washed with hot water to remove residual sugars and dried at 75 °C to a constant weight. It was milled to 18 meshes and stored at room temperature. The cellulase was provided by the Tianfeng Bioengineering Corporation in Hebei province, China. The filter paper activity (FPA) of the cellulase was 78 FPU/ml. 2.2. Pretreatment 2.2.1. Dilute acid pretreatment The sweet sorghum bagasse (10% w/v) was mixed with 1% (w/v) H2SO4 for 20 min at room temperature and then kept for 10 min separately at 180 °C in an autoclave with saturated steam. The dilute sulfuric acid pretreated sweet sorghum bagasse was washed with neutral pH distilled water and dried at 75 °C to a constant weight. 2.2.2. Lime pretreatment The sweet sorghum bagasse (10% w/v) and lime (1.5% w/v) were slurried in water, mixed and autoclaved at 121 °C for 1 h.

4586

J. Zhang et al. / Bioresource Technology 102 (2011) 4585–4589

The treated bagasse was then washed and dried as a dilute acid pretreatment.

the integrals of the C4 peaks at 86–92 (a) and 80–86 ppm (b) (Oh et al., 2005), which were calculated by Eq. (3).

2.2.3. Steam-explosion pretreatment The steam-explosion pretreatment was performed in a 5 m3 reactor with 0.25 t/h high pressure, at 160 °C for 5 min treatment time. 40% (w/w) moisture bagasse (200 kg dry weight) was fed to the reactor, when pressure of the reactor reached the target value the bagasse was exploded into the storeroom. At last the treated bagasse was washed and dried process similar to dilute acid pretreatment.

Cr:I ¼

2.2.4. Ionic liquids pretreatment (ILs) A cellulose solution of 10% w/w was prepared by mixing 5 g bagasse 45 g [BMIM] Cl in a 110 °C oil bath at 120 rpm for 1 h. After that, 200 ml of deionized water was rapidly added to the solution. The precipitated cellulose was collected by filtration. When the temperature of the solution cooled to room temperature, the treated bagasse was washed and dried as a dilute acid pretreatment. 2.3. Enzymatic hydrolysis

Fc 100% Fc þ Fa

ð1Þ

where Fc and Fa were the areas of the crystal and noncrystalline regions, respectively.

DðhklÞ ¼

kk b0 cos h

ð2Þ

D (hkl) is the size of the crystallite (nm), k is the Scherrer constant (0.94), k is X-ray wavelength (copper, 0.15418 nm). b0 is the full-width at half-maximum of the reflection hkl, measured at 2h which is the corresponding Bragg angle (Focher et al., 2001).

Cr:I ¼

a 100% aþb

ð3Þ

3. Results and discussion 3.1. Compositions of treated and untreated sweet sorghum bagasse

The enzymatic hydrolysis was performed at 50 °C in 250 ml shake flasks with 100 ml 0.05 M citrate buffer (pH 4.8). The enzymatic hydrolysis started with a solid loading of 10% and the enzyme loading was at 20 FPU/g dry substrates. Samples were withdrawn at 6, 12, 24, 36, 48, 60 and 72 h for analysis. 2.4. Analytical methods The method for analyzing the concentration of cellulose, hemicellulose and lignin was first described by Van Soest (1963). The reducing sugars were determined using 3,5-dinitrosalicylic acid (DNS) method (Miller, 1959). Glucose were determined by highperformance liquid chromatography (HPLC) in a RID refractive index detector (Shimadzu LC-10A, Japan) with an Aminex HPX-87P carbohydrate analysis column (Bio-Rad Labs, USA) operated at 80 °C by deionized water as mobile-phase (0.6 mL/min). (Cara et al., 2007). The activity of the cellulase was measured by FPA assay according to the method of standard IUPAC procedures (Ghose, 1987). Samples were imaged using a Hitachi S4700 SEM with an accelerating voltage of 20 kV. The FTIR was recorded between 400 and 2000 cm1 using a VARIAN-3100 gas chromatography. Discs were prepared with 2 mg dried sample and 200 mg KBr. X-ray diffraction was conducted using a Rigaku D/Max 2500 VB2+/ PC, with 2° min1 scan speed. Copper radiation (k = 0.154184 nm) was generated at a voltage of 40 kV. The scan scope was 5–45°. The degree of crystallinity was calculated by Eq. (1). The crystallite size was calculated from the Scherrer equation (2). 13C solid-state NMR experiments were performed on a Bruker AV-300 spectrometer (Bruker BioSpin, Karlsruhe, Germany) operating at a frequency of 75.47 MHz at 25 °C. The MAS speed was 5 kHz. The delay time after the acquisition of the free induction decay signal was 3 s. The crystallinity (Cr.I.) indices were calculated by a percentage of

An analysis of compositions on untreated and pretreated sweet sorghum bagasse is shown in Table1. For untreated materials, the contents of hemi-cellulose, cellulose and lignin were 26.3%, 45.3%, and 15.2%, respectively. For the treated materials, the contents of hemicelluloses decreased and cellulose increased compared with untreated bagasse. While the content of hemi-cellulose of dilute acid treated bagasse decreased significantly to 1.8%, which was much lower than those of other treated bagasse. The lignin content of lime treated bagasse decreased to 8.8%.By contrast, the ligin contents of other treated bagasse increased to about 25%. 3.2. Enzymatic hydrolysis of sweet sorghum bagasse The results of enzymatic hydrolysis of untreated and pretreated sweet sorghum bagasse are shown in Fig. 1. From the beginning to 60 h, all the enzymatic hydrolysis of untreated and pretreated bagasse experienced rapid increased on cellulose conversion and glucose concentration. At 60 h, the maximum conversion of cellulose and glucose concentration by enzymatic hydrolysis of steam explosion treated bagasse reached 70% and 25 g/L, respectively, which were 2.5 times higher than that of untreated bagasse (27% and 11 g/L). The maximum conversion of cellulose by enzymatic hydrolysis of dilute acid treated bagasse reached 50%, and that of ILs and lime treated bagasse reached about 40%. Thus, the improvements resulting from enzymatic hydrolysis using the four pretreatment approaches were significant. The enzymatic hydrolysis of steam explosion and dilute acid treated bagasse showed higher cellulose conversion rates than that of ILs and lime treated bagasse. For glucose conversion of steam explosion treated bagasse, the value of 70% was much higher than the optimum results obtained by Kaar et al. (1998).

Table 1 Compositions of treated and untreated sweet sorghum bagasse (of % dry weight).

Hemi-cellulose (%) Cellulose (%) Lignin (%) Others (%)

Untreated

ILs

Steam explosion

Lime

Dilute acid

26.3 ± 0.1 45.3 ± 0.3 15.2 ± 0.2 17.4 ± 0.1

16.7 ± 0.5 48.8 ± 0.3 25.3 ± 0.4 9.2 ± 0.1

12.5 ± 0.3 48.2 ± 0.2 23.4 ± 0.5 15.9 ± 0.5

15.1 ± 0.1 55.2 ± 0.4 8.8 ± 0.3 16.1 ± 0.2

1.8 ± 0.1 59.4 ± 0.1 26.5 ± 0.3 10.2 ± 0.4


4587

3.4. The FTIR analysis The FTIR spectra of sweet sorghum bagasse is shown in Supplemental Fig. 2(a). The representative bands of the figure are as follows: the absorption at 1737 cm1 is related to the stretching of C@O in hemicelluloses. The band at 1636 cm1 is attributed to the bending mode of the absorbed water and the stretching of C@O in lignin. 1516 cm1 represents the stretching of the phenyl ring. Two absorption bands at 1158 cm1 and 901 cm1 stem from CAOAC stretching at the b-(1–4)-glycosidic linkages. The in-plane ring stretching gave a shoulder at 1114 cm1. Strong peaks at 1061 cm1 and 1033 cm1 were indicative of CAO stretching at C-3, and CAC stretching and CAO stretching at C-6. (Oh et al., 2005; Cao and Tan, 2004). Compared with untreated sweet sorghum bagasse, no new peaks on the FTIR spectra of treated bagasse were observed. Therefore, there was no formation of new functional groups within the bagasse by any of the four pretreatments. For the four treated bagasse samples, the absorptions at 1737 cm1 were lower than that of untreated sample and the peak intensity of dilute acid treated and steam explosion treated bagasse also decreased obviously. These indicated that the bonds in the hemicelluloses were broken. At the absorption of 1516 cm1, the intensity of peak for dilute acid treated bagasse increased sharply. This indicates that the stretching of the phenyl ring in lignin was enhanced. For the dilute acid treated and steam explosion treated bagasse, the absorption of 1158, 1105 and 1033 cm1 increased strongly. This indicates that the CAO and CAC bonds of cellulose were exposed by the two pretreatment methods. The bonds of the hemicelluloses and cellulose were changed by dilute acid and steam explosion treatment. These two approaches improved enzymatic hydrolysis greatly. 3.5. XRD analysis

Fig. 1. Effect of different pretreatments on hydrolysis of cellulose (a) Time course of cellulose conversion; (b) Time course of glucose concentration.

3.3. SEM photos analysis As shown in Supplemental Fig. 1, the surface of the untreated bagasse was glazed, and there were no holes or damage in the bagasse. The structure of the ILs treated bagasse was loose, the bagasse surface was cracked and rough. The ILs treatment had a significant effect on the surface of the bagasse. The steam explosion treated bagasse was flocculent with many deep longitudinal cracks and micro-holes. The steam explosion treatment showed great effect on the change of external and internal structure of the bagasse. The surface of the lime treated bagasse showed many grooves. There was distinct change in the dilute acid treated bagasse as shown by the SEM photo. The photo shows an array of micro-holes in the bagasse. This increased the accessibility of the cellulose. The analysis of SEM photos showed that the internal and external structures of the bagasse were changed by the treatment of steam explosion and dilute acid, thus increasing the cellulose accessibility by cellulase. According to the results of enzymatic hydrolysis, the structure of bagasse and cellulose accessibility showed significant improvements on enzymatic hydrolysis.

The XRD analysis of treated and untreated sweet sorghum bagasse is shown in Supplemental Fig. 2(b). For the XRD spectra of untreated bagasse, a quite flat diffraction pattern was obtained. This indicated the presence of highly amorphous cellulose. All the XRD patterns were fit with a gauss function to calculate the amorphous and crystalline zone diffraction peaks (Supplemental Fig. 3). The typical diffraction angles were defined as crystalline 0 0 2) and amorphous zone. The results of cryszone (1 0 1, 1 0 1, tallinity (Cr.I) and crystallite size (Dhlk) are shown in Table 2. According to the results of hydrolysis, the cellulose conversions of untreated and treated bagasse were different. However, the calculated Cr.I changed weakly, which indicated the hydrolysis occurred simultaneously in the crystalline and amorphous zones. The Cr.I from XRD showed a weak effect on the enzymatic hydrolysis of the bagasse. At the same time, all of the crystallite sizes of treated bagasse decreased and this supported the observation that the crystalline size change had an effect on the hydrolysis of crystalline cellulose though the differences were not obvious. 3.6.

13

C NMR analysis

The 13C solid-state NMR spectra of treated and untreated bagasse was shown in Supplemental Fig. 4. The main 13C peak assignments of C1, C2, C3, C4, C5 and C6 are labeled in Supplemental Fig. 4. Though the cellulose structure differed widely in samples, the remarkable tendency was in the difference between spectral details of strict ordered and loose ordered cellulose. A clear observation from the Lorentzian fitting of C4 regions and seven zones was that there was a separation of the overlapping peaks of C4 (Ia, Ib, I(a + b), Para-crystalline amorphous and fibril surface cellulose). The fitting results are shown in Supplemental Fig. 5 and the

4588


Table 2 Crystallization index, characteristic peaks and crystalline size of treated and untreated sweet sorghum. Assignment

Untreated

Bragg angle 2h (°)

101 101 002 101 101 002

Crystalline size (nm)

Treatment

11.273 ± 0.01 15.670 ± 0.01 22.015 ± 0.05 3.6 ± 0.01 2.9 ± 0.01 4.5 ± 0.01 71.4

Cr.I (%)

ILs

Steam explosion

Lime

Dilute acid

11.322 ± 0.03 15.789 ± 0.01 22.098 ± 0.01 2.6 ± 0.01 2.9 ± 0.04 2.8 ± 0.01 69.6

11.316 ± 0.01 15.578 ± 0.01 22.376 ± 0.03 2.9 ± 0.03 2.4 ± 0.01 3.3 ± 0.01 64.7

11.395 ± 0.01 15.916 ± 0.03 22.189 ± 0.01 2.5 ± 0.01 2.6 ± 0.01 3.1 ± 0.02 62.3

11.155 ± 0.02 15.530 ± 0.02 22.392 ± 0.01 2.7 ± 0.01 2.1 ± 0.01 3.0 ± 0.05 62.5

Table 3 Crystallization Index of sweet sorghum and Chemical shift and intensity of fitting peak in Assignments Untreated ILs Steam explosion Lime Diluted acid

Chemical shift Intensity (%) Chemical shift Intensity (%) Chemical shift Intensity (%) Chemical shift Intensity (%) Chemical shift Intensity (%)

(ppm) (ppm) (ppm) (ppm) (ppm)

13

C NMR spectra.

Ia

Ib

I (a + b)

Para-crystalline

Amorphous

Fibril-surface

89.41 ± 0.05 15.8 ± 0.03 89.46 ± 0.01 2.0 ± 0.01 90.54 ± 0.01 2.1 ± 0.05 90.69 ± 0.02 3.0 ± 0.02 89.38 ± 0.01 2.9 ± 0.01

88.57 ± 0.01 31.8 ± 0.01 88.946 ± 0.05 13.1 ± 0.02 88.94 ± 0.01 3.8 ± 0.01 88.97 ± 0.01 1.4 ± 0.01 88.80 ± 0.01 8.2 ± 0.05

88.08 ± 0.01 8.9 ± 0.01 88.32 ± 0.02 20.1 ± 0.05 88.46 ± 0.01 18.4 ± 0.03 88.36 ± 0.01 14.2 ± 0.01 88.30 ± 0.03 7.3 ± 0.02

87.44 ± 0.01 10.9 ± 0.02 87.764 ± 0.01 4.6 ± 0.01 88.00 ± 0.04 10.4 ± 0.01 87.88 ± 0.01 7.6 ± 0.01 87.65 ± 0.01 14.1 ± 0.01

84.32 ± 0.02 14.2 ± 0.01 83.77 ± 0.01 21.5 ± 0.04 84.17 ± 0.02 21.4 ± 0.01 83.43 ± 0.01 41.8 ± 0.01 83.87 ± 0.01 32.1 ± 0.01

82.72 ± 0.05 12.8 ± 0.05 81.22 ± 0.03 17.7 ± 0.01 81.43 ± 0.01 19.6 ± 0.01 80.58 ± 0.01 6.2 ± 0.01 81.15 ± 0.05 8.6 ± 0.01

calculated peak intensity and crystallinity (Cr.I) are shown in Table 3. Because of the 13C NMR was more sensitive for orders of internal cellulose, the Cr.I calculated from NMR was different from XRD. Additionally, the calculated Cr.I from NMR more accurately reflected the real order of crystalline cellulose. According to results of hydrolysis, the degree of decreased Cr.I did not simultaneously relate to the increase of cellulose conversion. This indicated that the Cr.I was not a main factor contributing to enzymatic hydrolysis, though there may have been some relationship between decreased Cr.I and improved hydrolysis. According to previous studies, orderly cellulose was transformed to another state by pretreatment. The results in Table 3 shows that each pretreatment method transformed Ia and Ib to other bagasse forms and this resulted in an improvement of the hydrolysis of cellulose. For the ILs and lime treated bagasse, the composition of Para-crystalline decreased, while acid and steam explosion treated bagasse increased. This indicated the transformation of Para-crystalline from Ia and Ib improved hydrolysis. The large increase of amorphous and fibril surface cellulose from steam explosion treated bagasse also indicated that the hydrolysis occurred simultaneously in the crystalline and amorphous zones, which was the same as the results of XRD. It appears that the XRD and NMR analysis to determine cellulose crystallinity, structure and crystallite size were performed on whole biomass and not on cellulose isolated from it. In this case, there will be overlapping of hemicellulose peaks with cellulose in the C4 region, which affect the crystallinity and structure determination. Maybe spectral edit can mitigate this problem. It is worth to further study. The Cr.I values from NMR show greater than 50% decrease after pretreatment, while those from XRD only decrease by 13.8% yield, AP was superior to the other two pretreatments, as was also seen with the commercial enzymes (Table 1). Corn stover gave the highest yields of any substrate, followed by switchgrass. Comparing Glc yields obtained with the commercial enzymes to the core set optimized for each pretreatment/substrate combination, the core set mixtures performed about as well as Spezyme CP for most substrate/ pretreatment combinations, although yields with the core set were higher (25.9% yield vs. 16.4%) for 0.25% NaOH-switchgrass and lower for both AFEX-Miscanthus (22.8% vs. 36.2%) and AFEX-DDGS (22.6% vs. 32.6%). Accellerase 1000 consistently outperformed the core set for Glc yield but gave about the same yield of Xyl, consistent with our earlier results [4]. With regard to the optimal proportions of the core enzymes for Glc release from the different pretreatment/substrate combinations, CBH1 ranged from 16% (for AP-DDGS) to 49% (for 0.25% NaOH-corn stover). A noticeable trend was that the optimal proportions of EX3 were higher for AFEX-pretreated corn stover, switchgrass, and DDGS than for the other two pretreatments. For all substrates treated with 0.25% NaOH, EX3 was required only at the lowest proportion (4%-5%). This trend did not follow for Miscanthus, perhaps because Miscanthus has the lowest Xyl content of the three grass feedstocks (Additional file 1, Table S2). The elevated EX3 requirement was generally compensated by a lower loading of CBH1, i.e., CBH1 was present at a

Banerjee et al. Biotechnology for Biofuels 2010, 3:22 http://www.biotechnologyforbiofuels.com/content/3/1/22

Page 5 of 15

Table 2 Optimized proportions of six core enzymes for release of Glc from 15 pretreatment/substrate combinations* Optimized enzyme proportions (%)

Glc yield (%)

Feedstock

Pre-treatment

CBH1

BG

EG1

BX

EX3

CBH2

MP

Exptl.

Corn stover

AFEX

35

12

26

4

19

4

42.9

43.9 ± 0.5 40.7 ± 1.0

Switchgrass

Miscanthus

DDGS

Poplar

0.25% NaOH

49

5

34

4

4

4

42.0

Alk. peroxide

43

8

30

4

11

4

58.9

58.2 ± 0.2

AFEX

31

4

23

4

35

4

23.0

24.4 ± 0.6 25.9 ± 0.8

0.25% NaOH

46

12

30

4

4

4

27.0

Alk. peroxide

47

4

19

4

22

4

38.0

39.1 ± 0.5

AFEX 0.25% NaOH

42 36

4 4

42 46

4 4

4 5

4 4

23.0 17.0

22.8 ± 1.3 17.5 ± 0.7

Alk. peroxide

48

4

32

4

7

4

30.0

32.1 ± 1.2

AFEX

28

20

23

4

20

4

21.3

22.6 ± 0.8 23.8 ± 0.5

0.25% NaOH

27

27

34

4

4

4

23.1

Alk. peroxide

16

43

29

4

4

4

31.0

29.8 ± 0.5

AFEX

44

4

40

4

4

4

14.5

13.8 ± 0.1

0.25% NaOH

47

4

36

4

4

4

10.0

9.8 ± 0.1

Alk. peroxide

47

4

37

4

4

4

10.0

10.5 ± 0.2

*MP, model prediction. Exptl., experimental results. Data are expressed as means ± 1 SD (n = 8). All enzyme loadings were 15 mg/g glucan.

were different from those for corn stover, switchgrass, or Miscanthus, which is consistent with its very different sugar composition (Additional file 1, Table S1). For DDGS, a higher proportion of b-glucosidase and a lower proportion of CBH1 were optimal across all pretreatments (Table 2). This might be due to a lower degree of polymerization (DP) of the glucan of DDGS or to a lower degree of crystallinity. EG1, BX, and EX3 were the most important enzymes for optimal Xyl yield, except in the case of AP-treated switchgrass, for which EG1 was needed only at the

lower proportion for AFEX-pretreated corn stover, switchgrass and DDGS compared to other pretreatments. Owing to the limited importance of EX3 for releasing Glc from Miscanthus, CBH1 and EG1 constituted > 80% of the enzyme cocktail for this substrate, with the other enzymes needed only at their lowest limit. (Because the minimum proportion of each core enzyme was set at 4%, it cannot be determined whether a particular enzyme whose “optimum” is listed in Table 2 as 4% is needed at this level or at a lower level.) The optimum proportions of core enzymes for DDGS

Table 3 Optimized proportions of six core enzymes for release of Xyl from 15 pretreatment/substrate combinations* Optimized enzyme proportions (%) Feedstock Corn stover

Switchgrass

Miscanthus

DDGS

Poplar

Pretreatment

Xyl yield (%)

CBH1

BG

EG1

BX

EX3

CBH2

MP

Exptl. 27.8 ± 0.7

AFEX

7

4

20

26

39

4

28.2

0.25% NaOH

23

4

23

23

23

4

29.0

30.1 ± 0.5

Alk. peroxide AFEX

4 4

4 4

30 18

35 23

23 47

4 4

40.4 26.2

40.0 ± 2.0 24.8 ± 0.5 26.3 ± 0.6

0.25% NaOH

17

4

23

31

21

4

27.0

Alk. peroxide

4

4

4

37

47

4

40.5

39.5 ± 0.3

AFEX

14

4

20

30

28

4

37.2

35.5 ± 1.0

0.25% NaOH

20

4

28

25

19

4

22.6

23.0 ± 0.0

Alk. peroxide

8

4

24

29

31

4

42.3

40.6 ± 0.9

AFEX

-

-

-

-

-

-

-

< 7.0

0.25% NaOH Alk. peroxide

-

-

-

-

-

-

-

< 7.0 < 7.0 21.0 ± 0.0

AFEX

4

4

11

28

47

5

19.6

0.25% NaOH

7

4

25

16

43

4

23.0

23.8 ± 1.1

Alk. peroxide

4

4

23

19

46

4

25.2

26.0 ± 0.5

*MP, model prediction. Exptl., experimental results. Data are expressed as means ± 1 SD (n = 8). All enzyme loadings were 15 mg/g glucan. Xyl release from all DDG/pretreatment combinations was too low to construct statistically significant models.


lowest limit (Table 3). The importance of EG1 in Xyl release might be attributed to the fact that EG1also hydrolyzes b1,4-xylan [4,6]. In the case of DDGS, reliable models could not be derived for any of the pretreatment conditions because Xyl yield never exceeded 7%. The low yield of Xyl might be related to the fact that DDGS has the highest percentage of Ara (6.3%) and uronic acid (3%). If these sugars are attached to xylose residues, as is typically found in cereal glucuronoarabinoxylan, then little free Xyl can be released in the absence of Abf2 and a-Glr. This hypothesis is supported by the fact that high levels of Abf2 and a-Glr are needed for optimized Xyl release from DDGS in the 16component synthetic mixture (see next section). The statistical analyses of all models are shown in Additional file 1, Tables S5 and S6. Sixteen-component mixtures

As a further test of the use of GENPLAT for enzyme mixture optimization, three pretreatment/substrate combinations were studied using more complex cocktails. Since DDGS contains significant levels of starch (~5%) and Man (~2.5%) (Additional file 1, Table S2) [18], amyloglucosidase (g-amylase) and b-mannanase were added to the cocktails. The Megazyme b-mannanase was confirmed by MS-based proteomics to be pure GenBank XP_001390707 (JGI Aspni5_50378), which is an A. niger protein in CAZy GH family 5 (data not shown). Three proteins that had previously been tested but found to have no effect on Glc or Xyl yield from AFEX-corn stover were also included, namely, Cip1, Cip2, and Cel12A [5]. The rationale for including them was that these proteins might be important in the presence of the other new enzymes (i.e., amyloglucosidase and b-mannanase) or on the new substrates (i.e., AP-pretreated corn stover and AFEX-pretreated DDGS). These mixtures thus contained a total of 16 proteins. All three pretreatment/substrate combinations were analyzed with all 16 enzymes. Cip1, Cip2, and Cel12A did not contribute to Glc or Xyl release from AFEX-treated corn stover as was found earlier (Figure 1) [5]. Amyloglucosidase and b-mannanase were also unnecessary for Glc or Xyl release from this substrate, which is consistent with its low levels of starch and Man (Figure 1; Additional file 1, Table S2). CBH1, EG1, Cel61A, and EX2 were the most important enzymes for Glc yield from AFEX-corn stover (Figure 1a), as found earlier [5], although there were some differences from the earlier results. In particular, the optimum levels of EG1 were now found to be 19%, compared to 8% in the earlier work, and the optimum level of Cel5A was 2%, compared to 11% in the earlier work. Possible reasons for this are considered in the Discussion.

Page 6 of 15

Comparing AFEX- and AP-corn stover, there were some rather large differences in the optimal proportions of some of the enzymes for Glc release (Figures 1a and 2a). In particular, the optimal proportion of CBH1 for AP-corn stover was 31% compared to 22% for AFEX-corn stover, and CBH2 was also higher for AP-corn stover (12% vs. 4%). Therefore, total exo-b1,4-glucanase (CBH1 + CBH2) was 17% higher for AP-corn stover than for AFEX-corn stover. In contrast, the optimal proportions were lower for EG1 (8% vs. 19%) and Cel5A (0% vs. 2%) for AP-corn stover than for AFEX-corn stover. Therefore, total endob1,4-glucanase (EG1 + Cel5A) was 13% lower for AP-corn stover. Expressed in another way, the exoglucanase to endoglucanase ratio increased from 1.2:1 for AFEX-corn stover to 5.4:1 for AP-corn stover. This change in ratio indicates either that the exoglucanases are more important for AP-corn stover digestion compared to AFEX-corn stover or that the endoglucanases are less important. In the latter case, and if the exo-b1,4-glucanase activities are the limiting factor for crystalline cellulose hydrolysis, a reduced requirement for endoglucanases permits a higher proportion of the limiting factor, and this in turn could be the explanation for the enhanced Glc yields seen following AP pretreatment (Tables 1,2). The amount of EX2 necessary for optimal Glc release decreased from 13% to 3% when comparing AP- to AFEXcorn stover (Figures 1a and 2a). Possible reasons for this are considered in the Discussion. The enzyme proportions for maximum release of Xyl from AFEX-corn stover were similar to previous results [5] and similar between AFEXand AP-corn stover (Figures 1b and 2b). Consistent with its different monosaccharide composition (Additional file 1, Table S2), the optimal proportions of the 16-component mixture were very different for AFEX-DDGS compared to AFEX- or AP-corn stover for release of both Glc and Xyl (Figure 3). The requirement for a high proportion of amyloglucosidase is consistent with the presence of significant residual starch in DDGS (Additional file 1, Table S2). Surprisingly, b-mannanase is the second most important enzyme for the release of Glc from DDGS (Figure 3). Possible reasons for this are considered in the Discussion section. EX2 is also necessary (at an optimum proportion of 16%), but the other enzymes are needed only at or near their lowest set levels (Figure 3). Abf2, a-Glr and BX were the most important enzymes for Xyl release from AFEX-DDGS (Figure 3b). The importance of the first two of these is consistent with the Xyl in DDGS being heavily substituted with Ara and glucuronic acid. Experimental yields of Glc and Xyl from all three pretreatment/substrate combinations with the optimized 16-component mixtures are shown in Figure 4. For ease of comparison, the yield results for Spezyme CP and Accellerase 1000 for AFEX and AP-corn stover and


Page 7 of 15

Figure 1 Enzyme proportions for a 16-component synthetic mixture optimized for (a) Glc or (b) Xyl release from AFEX-corn stover. Numbers are percentages of each enzyme in the optimized mixture. Enzyme loading was 15 mg/g glucan. Yields are shown in Figure 4. g-Amylase is also known as amyloglucosidase.

AFEX-DDGS from Table 1 are included in Figure 4. The 16-component mixture surpassed Spezyme CP and equaled Accellerase 1000 for Glc yield from AFEX-corn stover (Figure 4a). This result is similar to earlier results with 11-component mixtures [5], which was expected because none of the new enzymes influenced Glc yield from this substrate (Figure 1a). The 16-component mixture was also equal to Accellerase 1000 and superior to Spezyme CP for release of Glc from AP-corn stover (Figure 4a). Glc yields from AFEX-DDGS were higher

with the 16-component synthetic mixture compared to either commercial preparation (Figure 4a). For Xyl, the 16-component mixture yielded more from AFEX-corn stover than either commercial preparation (Figure 4b), as previously reported [5]. This was also true for APcorn stover (Figure 4b). Although the synthetic mixture was the best at releasing Xyl from AFEX-DDGS, it still only released 14% of the available Xyl (Figure 4b). Neither Accellerase 1000 nor Spezyme CP was designed to degrade DDGS, and therefore their poor


Page 8 of 15

Figure 2 Enzyme proportions for a 16-component mixture optimized for (a) Glc or (b) Xyl release from AP-corn stover. Numbers are percentages of each enzyme in the optimized mixture. Enzyme loading was 15 mg/g glucan. Yields are shown in Figure 4.

performance on this substrate is not surprising. It has been shown that mixtures of commercial enzymes can be more effective than individual ones when compared at equal protein loading (e.g., [19]). To test this, an optimization experiment was performed with four commercial enzymes on AFEX-DDGS (Figure 5). The optimized four-component mixture contained 47% Accellerase 1000, 27% Multifect Pectinase, 22% Novozyme 188 and 4% Multifect Xylanase, and outperformed all other

enzyme preparations, releasing 57.2 ± 1.2% of the total Glc (Figures 4 and 5). The resulting ternary diagram is shown in Figure 5, and the underlying experimental data are shown in Additional file 1, Table S3. Although Novozyme 188 is frequently used as a supplementary source of b-glucosidase, amyloglucosidase (glucoamylase) is actually the most abundant protein in this product, as shown experimentally by proteomics (Additional file 1, Table S4). The results in Figure 3a,


Page 9 of 15

Figure 3 Enzyme proportions for a 16-component mixture optimized for (a) Glc or (b) Xyl release from AFEX-DDGS. Numbers are percentages of each enzyme in the optimized mixture. Enzyme loading was 15 mg/g glucan. Yields are shown in Figure 4.

which indicate that amyloglucosidase is the single most important enzyme for releasing Glc from DDGS, suggest that the superiority of the optimized four-component mixture to Accellerase 1000 alone is due, at least partially, to the high levels of amyloglucosidase present in Novozyme 188. A mixture of the same four commercial enzymes optimized for Xyl also yielded more Xyl (29.1% ± 0.7%) from AFEX-DDGS than any of the other tested

preparations (i.e., Spezyme CP, Accellerase 1000, or the 16-component synthetic mixture) (Figures 4 and 5). The mixture optimized for Xyl contained 69% Multifect Pectinase, 31% Accellerase 1000, 0% Multifect Xylanase, and 0% Novozyme 188. The paradoxical unimportance of Multifect Xylanase in this experiment might be because any necessary xylanase activity is being supplied by Multifect Pectinase [13]. The ternary diagram and experimental data for Xyl yield are shown in Figure 5


Page 10 of 15

and Additional file 1, Table S3, respectively. Despite optimization, the low maximum yield of Xyl from DDGS indicates that there is still considerable scope for improvement of enzyme degradation of this substrate [13]. Influence of loading and time on digestion of AP-corn stover

Figure 4 Yields of Glc (a) and Xyl (b) for Spezyme CP, Accellerase 1000, or the 16-component optimized mixture on three pretreatment/substrate combinations. Enzyme loadings were fixed at 15 mg/g glucan and digestions were for 48 h. Yields (as a percentage of total Glc or Xyl) are shown above the data bars. Values for Spezyme CP and Accellerase 1000 are taken from Table 1. Error bars indicate ±1 SD of the mean (n = 8).

All of the preceding experiments were performed at a fixed enzyme loading (15 mg/g glucan) and a single digestion time (48 h). It is possible that synthetic enzyme mixtures show different efficiency on different substrates as a function of loading or incubation time. In particular, the higher yields seen with AP vs. AFEX-corn stover (Figure 4) led us to study the effect of loading and time on digestion of corn stover pretreated with AP. In earlier work with AFEX-corn stover, Accellerase 1000 gave a better yield of Glc at 6 h, but Spezyme CP and an 11-component synthetic mixture were as effective as Accellerase 1000 at 48 h. At 48 h, as loading increased, Glc yields tended to plateau at ~20 mg/g glucan at a maximum yield of ~52% [5]. Because none of the extra components in the 16-component mixture used in the current paper, namely, Cip1, Cip2, Cel12A, b-mannanase, and amyloglucosidase, made any contribution to Glc yield from corn stover (Figure 2a), the 16component results in this paper can be compared to the earlier 11-component results [5]. As shown in Figure 6, at 12 h Accellerase 1000 was superior to Spezyme CP, the core set or the 11-component mixture for Glc release from AP-corn stover. At 48 h, there was less difference between the four enzyme preparations at higher

Figure 5 Ternary diagrams of optimization of mixtures of four commercial enzyme preparations for release of Glc (left) or Xyl (right) from AP-treated DDG. In these graphical representations, Multifect Xylanase was kept constant. Loadings were constant at 15 mg/g glucan. The experimental data on which the models were based are shown in Additional file 1, Table S3.


Page 11 of 15

more than 60% of the available Glc in 12 h, and at a loading of 5 mg/g glucan released more than 60% of the Glc in 48 h (Figure 6). That is, under these experimental conditions, AP-pretreatment resulted in the release of more Glc in a shorter time at a lower enzyme loading than AFEX pretreatment. At 6 h, Xyl release from AP-corn stover by the two synthetic enzyme mixtures tended to reach a plateau above ~10 mg/g glucan, whereas Xyl release by the two commercial enzyme loadings was close to linear over the whole tested range of enzyme loadings (Figure 7a). This was similar to earlier results with AFEX-corn stover [5]. By 48 h, the 11-component synthetic mixture was superior to

Figure 6 Glc release from AP-corn stover as a function of enzyme loading. (a) 6 h, (b) 12 h, (c) 48 h. The “core set” and “core set + accessory” mixtures were identical to those used in Figures 2 and 3, i.e., optimized for 15 mg/g glucan loading and 48h digestion. Error bars, which are sometimes obscured by the data symbols, represent ± 1 SD of the mean (n = 8).

loadings. At lower loadings (< 10 mg/g glucan), Accellerase 1000 and the 16-component mixture were superior to the core set or Spezyme CP (Figure 6). Glc release from AP-corn stover in response to Accellerase 1000, the core set, or the 16-component mixture reached a loading plateau at a lower level than Glc release from AFEX-corn stover (< 10 mg/g glucan vs. ~20 mg/g glucan) and at a higher yield (~70% vs. ~52%). This trend was especially notable for Accellerase 1000, which reached its plateau on AP-corn stover even at ~5 mg/g (Figure 6c). On AP-corn stover, therefore, Accellerase 1000 at a loading of 15 mg/g glucan released

Figure 7 Xyl release from AP-corn stover as a function of enzyme loading. (a) 6 h, (b) 12 h, (c) 48 h. The “core set” and “core set + accessory” mixtures were identical to those used in Figures 2 and 3, i.e., optimized for 15 mg/g glucan loading and 48-h digestion.


all of the others by a few percentage points (Figure 7c), which is also similar to the earlier results. For Xyl at 48 h, all of the enzymes showed a tendency to reach a plateau with increased loading, but it was not as pronounced as it was for Glc (Figures 6 and 7). As was observed for Glc yield, AP gave higher yields of Xyl than AFEX at 15 mg/g glucan and 48 h (55% vs. 41% Xyl yield with an 11-component mixture) (Figure 4b). The data for 15 mg/g glucan loading from Figures 6 and 7 were replotted as a function of time (Figure 8). The trends with AP-corn stover were similar to those seen with AFEX-corn stover. At 12 h, Accellerase 1000 performed the best for Glc, and the multi-component synthetic mixture performed the best for Xyl. These results suggest that the synthetic mixture still lacks one or more enzymes needed for Glc release at early time points compared to Accellerase 1000. A caveat to this is to recognize that the synthetic mixture was optimized for 48 h and might be capable of performing as well as or better than Accellerase 1000 if it were reoptimized for 12 h.

Figure 8 Time course of release of Glc and Xyl from AP-corn stover. (a) Glc, (b) Xyl. All enzyme loadings were 15 mg/g glucan. Data are from the same experiment shown in Figures 6 and 7. The “core set” and “core set + accessory” mixtures were identical to those used in Figures 2 and 3, i.e., optimized for 15 mg/g glucan loading and 48-h digestion.

Page 12 of 15

For Glc, the advantage of Accellerase 1000 disappeared at later time points as it tended to plateau more dramatically than the other enzyme mixtures (Figure 8a). For Xyl, the 16-component mixture was superior at all time points. The fact that yields of Glc and Xyl were still increasing from 12 to 48 h, in some cases almost linearly (e.g., Spezyme CP and 11-component mixture in Figure 8a and Accellerase 1000 in Figure 8B), suggests that the enzymes were still active and that the plateau effect seen at 48 h in Figures 6c and 7c was not due to enzyme inactivation.

Discussion The future of the lignocellulosic ethanol industry will depend on the development of economical and effective pretreatments and compatible enzyme cocktails. Many different feedstocks and pretreatments are currently being actively investigated. Insofar as regions differing in geography and climate favor the production of different feedstocks, and because different feedstocks require different pretreatments, a mature lignocellulosic ethanol industry may need to utilize a wide variety of pretreatment/feedstock combinations. Since it is unlikely that only one or a few enzyme cocktails will be sufficiently effective in all cases, it will be necessary to develop custom cocktails. GENPLAT (or a scaled-up version of it) should prove to be useful for the development of such custom cocktails. In this paper, we show the use of GENPLAT to rapidly determine the optimal proportions of enzymes in complex synthetic mixtures for the digestion of a variety of pretreated biomass feedstocks. Most of these results are consistent with other studies and with our earlier results, e.g., the importance of core enzymes such as CBH1, EG1, and BG for Glc release; the importance of EX, BX, Abf2, and a-Glr for Xyl release; the importance of EX for Glc release; the importance of the cryptic endoglucanase known as Cel61 for Glc release; and the involvement of amyloglucosidase for Glc release from DDGS [4,5]. However, other results reported here were unexpected from previous studies. First, we found that the optimal proportions of several enzymes were significantly different in the current experiments compared with those we reported earlier [5]. In particular, in the 16-component experiments (which are comparable to the earlier 11-component experiments [5]), the new optimized model predicted a large shift in the optimal proportions of EG1 (i.e., an increase from 8% to 19%) and of Cel5A (i.e., a decrease from 11% to 2%). A possible explanation for this observation might lie in the fact that EG1 and Cel5A are both endo-b1,4-glucanases and therefore have overlapping, or even identical, functions in the degradation of lignocellulose. Large changes in the relative proportions of these two, while maintaining their sum constant


(i.e., the sum of EG1 + Cel5A was 19% in the earlier experiments and 21% in the current ones) might result in a very small shift in Glc yield, resulting in different but equally valid models. That is, the topology of the response surface in the EG1/Cel5A region might be relatively flat, and minor experimental variations could change their relative proportions greatly without having a major effect on Glc yield. If these two enzymes are essentially interchangeable, the choice of which one to use in synthetic mixtures might not make much difference and could be dictated by other factors such as cost. Another example of such “equivalence switching” might occur in the 16-component experiments comparing AFEX-corn stover and AP-corn stover (Figures 1 and 2). As found earlier [5], for both substrates a combination of two endo-b1,4-xylanases (EX2 and EX3) is superior to either one alone (Figures 1 and 2). The reason for this is not clear, but it indicates that the functions of the two enzymes are not completely overlapping. In the current experiments, shown in Figures 1 and 2, AP-corn stover requires more EX3 (14% vs. 9%) but less EX2 (3% vs. 13%). The sum of EX2 + EX3, however, changes only from 22% to 17%. Compared to the two endo-b1,4-glucanases, however, in this case the situation is complicated by the fact that the two endoxylanases are not catalytically equivalent. EX2 (GH family 11) is specific for xylan, whereas EX3 (GH 10) has a broader range of substrates, including b1,4-glucan oligomers [20]. In contrast to EG1 and Cel5A, the optimum proportions of Cel61A for Glc release stayed fairly constant throughout the experiments. Optimum proportions were 17% (AFEX-corn stover) [5], 18% (AFEX-corn stover; Figure 1a) and 23% (AP-corn stover; Figure 2a). Cel61A was not needed for Xyl release from corn stover nor for either Glc or Xyl release from AFEX-DDGS (Figure 3). Although the catalytic function of Cel61A is still uncertain, especially on complex polysaccharides, we have found that it can hydrolyze cellopentaose to cellobiose and cellotriose, and cellohexaose to cellobiose (unpublished observations). Nonetheless, our experiments indicate that Cel61A is not redundant with other endo-b1,4-glucanases, which can also act on these same cello-oligosaccharides, and that it therefore probably plays some other, unknown role in cell wall degradation [21]. Our empirical determination of the importance of Cel61A illustrates that it can be difficult to predict which enzymes will be important, and at what proportions, based solely on our (imperfect) knowledge of the structures of plant cell walls and our (imperfect) knowledge of the full range of enzymatic activities of the large number of proteins secreted by lignocellulolytic microorganisms [3]. Another unexpected result was the need for a high level of b-mannanase to release Glc from AFEX-DDGS

Page 13 of 15

(Figure 3a). Although DDGS contains only 2.5% Man, polymers of Man apparently play an important role in the cell wall structure of DDGS. Hägglund et al. [22] found that the CBM module of fungal b-mannanase binds to cellulose and promotes the hydrolysis of cellulose in mannan/cellulose complexes, but plays no role in the hydrolysis of pure cellulose. There is evidence that mannan in plant cell walls is tightly bound to cellulose microfibrils [23,24]. From these observations, a plausible explanation for the importance of b-mannanase is that in DDGS, and perhaps some other lignocellulosic materials, polymers of Man must be hydrolyzed before the true cellulases can access cellulose. It is difficult to speculate further about the importance of b-mannanase in the degradation of DDGS because little is known about the cell wall structure of corn DDGS. Most analyses of the “glucan” in DDGS have not discriminated between the different polymers in which Glc could be present. These include not only crystalline cellulose and starch but also amorphous cellulose, mixed-linked glucan, and glucomannan. For example, Kim et al. [25] assumed that all nonstarch Glc in DDGS is derived from “cellulose,” but it is known that the endosperm tissues of barley and other cereals can contain high levels of Glc in the form of mixed-linked glucan [26]. Another paradox about DDGS is our observation that the core set of enzymes (which includes EX3) was much less effective at releasing Xyl from DDGS than from AFEX-corn stover [4]. It is also surprising that the exo-b1,4-glucanases and the endo-b1,4-glucanases were relatively unnecessary for releasing Glc from DDGS compared to the stovers (Figure 3a), even though total Glc yields from DDGS were comparable to those of AFEX-corn stover (Figure 4a). Collectively, our enzyme optimization results suggest that our knowledge of the structures and interactions of polysasccharides in DDGS is still very imperfect. The importance of xylanases (in the case of grass stovers) and b-mannanase (in the case of DDGS) for Glc yield are consistent with hemicelluloses in the cell wall having a major effect on limiting access of cellulases to cellulose. Collectively, enzymes that degrade the hemicelluloses thus play two roles in industrial lignocellulose degradation: promotion of access of cellulases to cellulose and release of fermentable hemicellulosic monosaccharides.

Conclusions Improvement in the efficiency of enzyme mixtures, either mixtures of commercial enzymes or synthetic mixtures of pure enzymes, would greatly improve the economic viability of lignocellulosic ethanol. Results from many laboratories have contributed to the identification of the critical enzymes needed for release of Glc


and Xyl from pretreated biomass. However, we predict that there are additional critical enzymes that are either absent or present at suboptimal levels in commercial enzyme preparations. GENPLAT provides a technology by which such enzymes can be identified and validated. Our results also illustrate the difficulty of predicting which enzymes are necessary to degrade a particular pretreatment/biomass substrate based on either monosaccharide or polysaccharide abundance. Therefore, the development of better enzyme cocktails will require empirical approaches.

Additional material Additional file 1: Supplementary supporting data. Supplementary Table S1. Optimized proportions of the core set for a 1:1 yield of Glc and Xyl. Supplementary Table S2. Monosaccharide and lignin composition of feedstocks used in this paper. Supplementary Table S3. Experimental results for optimization of digestion of AP-treated DDGS with mixtures of four commercial enzyme preparations. Supplementary Table S4. Proteomic analysis of the commercial enzyme product Novozyme 188. Supplementary Table S5. Statistical analysis for Glc optimization from pretreatment/substrate combinations. Supplementary Table S6. Statistical analysis for Xyl optimization from pretreatment/substrate combinations.

Page 14 of 15

4.

5.

6.

7.

8.

9.

10. 11.

12.

Abbreviations Abf2: arabinosidase 2; AFEX: ammonia fiber expansion; a-Glr: aglucuronidase; AP: alkaline peroxide; Ara: arabinose; BG: b-glucosidase; BX: bxylosidase; CBH: cellobiohydrolase; DDGS: dried distillers’ grains plus solubles; EG: endo-b1,4-glucanase; EX: endo-b1,4-xylanase; Glc: glucose; Man: mannose; Xyl: xylose. Acknowledgements We thank Cliff Foster, Michigan State University and the Great Lakes Bioenergy Research Center, for the wall carbohydrate analyses, and Derek Marshall of the Biomass Conversion Research Laboratory, Department of Chemical Engineering, Michigan State University, for the AFEX treatments. This work was funded by the Department of Energy Great Lakes Bioenergy Research Center (GLBRC) (DOE Office of Science BER DE-FC02-07ER64494). Authors’ contributions GB and SC performed the alkaline peroxide and dilute base pretreatments, expressed some of the proteins, designed and executed the digestion experiments, performed the statistical analyses and helped write the paper; JSC and MB cloned genes and expressed proteins and helped write the paper; JDW contributed to experimental design and drafted the final manuscript. All authors provided input to the manuscript and read and approved the final manuscript. Competing interests The authors declare that they have no competing interests.

13.

14.

15.

16.

17. 18.

19.

20. 21.

Received: 18 August 2010 Accepted: 12 October 2010 Published: 12 October 2010 References 1. Lynd LR, Laser MS, Bransby D, Dale BE, Davison B, Hamilton R, Himmel M, Keller M, McMillan JD, Sheehan J, Wyman CE: How biotech can transform biofuels. Nature Biotechnol 2008, 26:169-172. 2. Banerjee G, Scott-Craig JS, Walton JD: Improving enzymes for biomass conversion: a basic research perspective. Bioenerg Res 2010, 3:82-92. 3. Nagendran S, Hallen-Adams HE, Paper JM, Aslam N, Walton JD: Reduced genomic potential for secreted plant cell-wall-degrading enzymes in the

22.

23. 24.

ectomycorrhizal fungus Amanita bisporigera, based on the secretome of Trichoderma reesei. Fung Genet Biol 2009, 46:427-435. Banerjee G, Car S, Scott-Craig JS, Borrusch MS, Aslam N, Walton JD: Synthetic enzyme mixtures for biomass deconstruction: production and optimization of a core set. Biotechnol Bioengineer 2010, 106:707-720. Banerjee G, Car S, Scott-Craig JS, Borrusch MS, Bongers M, Walton JD: Synthetic multi-component enzyme mixtures for deconstruction of lignocellulosic biomass. Bioresour Technol 2010, 101:9097-9105. Gao D, Chundawat SP, Krishnan C, Balan V, Dale BE: Mixture optimization of six core glycosyl hydrolases for maximizing saccharification of ammonia fiber expansion (AFEX) pretreated corn stover. Bioresour Technol 2010, 101:2770-2781. Selig MJ, Knoshaug EP, Adney WS, Himmel ME, Decker SR: Synergistic enhancement of cellobiohydrolase performance on pretreated corn stover by addition of xylanase and esterase activities. Bioresour Technol 2008, 99:4997-5005. Selig MJ, Knoshaug EP, Decker SR, Baker JO, Himmel ME, Adney WS: Heterologous expression of Aspergillus niger β-D-xyosidase (XlnD): characterization on lignocellulosic substrates. Appl Biochem Biotechnol 2008, 146:57-68. Garlock RJ, Chundawat SPS, Balan V, Dale BE: Optimizing harvest of corn stover fractions based on overall sugar yields following ammonia fiber expansion pretreatment and enzymatic hydrolysis. Biotech Biofuels 2009, 2:29. Tilman D, Hill J, Lehman C: Carbon-negative biofuels from low-input high-diversity grassland biomass. Science 2006, 314:1598-1600. Bals B, Rogers C, Jin M, Balan V, Dale B: Evaluation of ammonia fibre expansion (AFEX) pretreatment for enzymatic hydrolysis of switchgrass harvested in different seasons and locations. Biotechnol Biofuels 2010, 3:1. Murnen HK, Balan V, Chundawat SPS, Bals B, da Costa Sousa L, Dale BE: Optimization of ammonia fiber expansion (AFEX) pretreatment and enzymatic hydrolysis of Miscanthus × giganteus to fermentable sugars. Biotechnol Prog 2007, 23:846-850. Dien BS, Ximenes EA, O’Bryan PJ, Moniruzzaman M, Li XL, Balan V, Dale B, Cotta MA: Enzyme characteristics for hydrolysis of AFEX and liquid hotwater pretreated distillers’ grains and their conversion to ethanol. Bioresour Technol 2008, 99:5216-5225. Balan V, da Costa Sousa L, Chundawat SPS, Marshall D, Sharma LN, Chambliss CK, Dale BE: Enzymatic digestibility and pretreatment degradation products of AFEX-treated hardwoods (Populus nigra). Biotechnol Prog 2009, 25:365-375. Santoro N, Cantu SL, Tornqvist CE, Falel TG, Bolivar JL, Patterson SE, Pauly M, Walton JD: A high throughput platform for screening milligram quantities of plant biomass for lignocellulose digestibility. Bioenergy Res 2010, 3:93-102. Albersheim P, Nevins DJ, English PD, Karr A: A method for the analysis of sugars in plant cell wall polysaccharides by gas-liquid chromatography. Carbohydr Res 1967, 5:340-345. Updegraff DM: Semimicro determination of cellulose in biological materials. Anal Biochem 1969, 32:420-424. Wu YV: Determination of neutral sugars in corn distillers’ dried grains, corn distillers’ dried solubles, and corn distillers’ dried grains with solubles. J Agr Food Chem 1994, 42:723-726. Kumar R, Wyman CE: Effect of xylanase supplementation of cellulase on digestion of corn stover solids prepared by leading pretreatment technologies. Bioresour Technol 2009, 100:4203-4213. Collins T, Gerday C, Feller G: Xylanases, xylanase families and extremophilic xylanases. FEMS Microbiol Rev 2005, 29:3-23. Harris PV, Welner D, McFarland KC, Re E, Poulsen JCN, Brown K, Salbo R, Ding H, Vlasenko E, Merino S, Xu F, Cherry J, Larsen SY, Lo Leggio L: Stimulation of lignocellulosic biomass hydrolysis by proteins of glycoside hydrolase family 61: Structure and function of a large, enigmatic family. Biochemistry 2010, 49:3305-3316. Hägglund P, Eriksson T, Collén A, Nerinckx W, Claeyssens M, Stålbrand H: A cellulose-binding module of the Trichoderma reesei β-mannnase Man5A increases the mannan-hydrolysis of complex substrates. J Biotechnol 2003, 101:37-48. Åkerholm M, Salmén L: Interactions between wood polymers studied by dynamic FT-IR spectroscopy. Polymer 2001, 42:963-969. Wilkie KCB: The hemicelluloses of grasses and cereals. Adv Carbohydr Chem Biochem 1979, 36:215-264.


Page 15 of 15

25. Kim Y, Mosier NS, Hendrickson R, Ezeji T, Blaschek H, Dien B, Cotta M, Dale B, Ladisch MR: Composition of corn dry-grind ethanol by-products: DDGS, wet cake, and thin stillage. Bioresour Technol 2008, 99:5165-5176. 26. Burton RA, Fincher GB: (1,3;1,4)-β-D-Glucans in cell walls of the Poaceae, lower plants, and fungi: a tale of two linkages. Mol Plant 2009, 2:873-882. doi:10.1186/1754-6834-3-22 Cite this article as: Banerjee et al.: Rapid optimization of enzyme mixtures for deconstruction of diverse pretreatment/biomass feedstock combinations. Biotechnology for Biofuels 2010 3:22.

Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color figure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit




Comparison of the effects of five pretreatment methods on enhancing the enzymatic digestibility and ethanol production from sweet sorghum bagasse Weixing Cao, Chen Sun, Ronghou Liu ⇑, Renzhan Yin, Xiaowu Wu Biomass Energy Engineering Research Centre, School of Agriculture and Biology, Shanghai Jiao Tong University, 800 Dongchuan Road, Shanghai 200240, People’s Republic of China

a r t i c l e

i n f o

Article history: Received 21 November 2011 Received in revised form 29 January 2012 Accepted 7 February 2012 Available online 16 February 2012 Keywords: Sweet sorghum bagasse Enzymatic hydrolysis Pretreatment Bio-ethanol

a b s t r a c t To improve the enzymatic digestibility of sweet sorghum bagasse and bioethanol production, five pretreatment methods have been investigated and compared, including (1) dilute NaOH solution autoclaving pretreatment, (2) high concentration NaOH solution immersing pretreatment, (3) dilute NaOH solution autoclaving and H2O2 immersing pretreatment, (4) alkaline peroxide pretreatment and (5) autoclaving pretreatment. Among them, the best result was obtained when sweet sorghum bagasse was dilute NaOH solution autoclaving and H2O2 immersing pretreatment. The highest cellulose hydrolysis yield, total sugar yield and ethanol concentration were 74.29%, 90.94 g sugar/100 g dry matter and 6.12 g/L, respectively, which were 5.88, 9.54 and 19.13 times higher than the control. Moreover, the FTIR and SEM analysis illustrated significant molecule and surface structure changes of the sweet sorghum bagasse after pretreatments. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Sweet sorghum is a high photosynthetic efficiency energy crop, and it is adapted to growing in nutrient-poor soils, making it one of the most promising energy-crop for bio-ethanol production. Both its grain and stalk are ideal raw material for bioethanol production due to its high biomass and sugar-yielding (Billa et al., 1997). The starches in the grain can be hydrolyzed to fermentable sugars by amylase, while the cellulose and hemicellulose in the bagasse require pretreatment prior to enzymatic hydrolysis and further conversion to ethanol (Lin and Tanaka, 2006). Bio-ethanol production from sweet sorghum bagasse is more attractive in terms of energy balances and emissions, and it may be the case that the bagasse will become the vital supplementary material for ethanol production. The main components of sweet sorghum bagasse are cellulose, hemicellulose and lignin (Sipos et al., 2009), of which the intricate structure severely restricts the enzymatic hydrolysis. In order to improve the accessibility of the enzyme to cellulose, the studies on lignocellulose pretreatment are needed (Taherzadeh and Karimi, 2008). The pretreatment methods mainly include physical methods such as mechanical or thermal (Mais et al., 2002; Silva et al., 2010), chemical methods (Saha et al., 2005; Yamashita et al., 2010; Saha and Cotta, 2006; Beukes and Pletschke, 2011), biological methods and combination of these methods. The mechanism of physical pretreatment is to increase the accessible surface area and decrease the crystallinity degree of lignocellulose by chipping, milling, grinding ⇑ Corresponding author. Tel./fax: +86 21 34205744. E-mail address: [email protected] (R. Liu). 0960-8524/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2012.02.034

or irradiation (Fan et al., 1980; Taherzadeh and Karimi, 2008; Mais et al., 2002; Silva et al., 2010). It is also effective for improving the enzymatic hydrolysis using thermal pretreatment such as steam explosion and liquid hot water pretreatment which can remove most of the hemicellulose (Shen et al., 2011; Laser et al., 2002). The typical conditions of liquid hot water pretreatment are high temperature and pressure (160–260 °C, 0.69–4.83 MPa) at short time (e.g. a few seconds to several minutes) or ambient temperature and normal pressure at relatively long time (several hours to several days) (Dien et al., 2006; Sreenath et al., 1999). Generally, the shortcoming of these physical pretreatment methods is high energy requirement. Alkaline (e.g. NaOH) and alkaline peroxide pretreatments which belong to chemical methods are effective processes for pretreating lignocellulose materials. NaOH is widely used for lignocellulose pretreatment. It can remove partial lignin and hemicellulose in the biomass by fracturing the ester bonds thereby increasing the porosity of the biomass (Xu et al., 2010; Zhang et al., 2011). Alkaline peroxide is used for lignocellulose pretreatment in recent years. It can improve the enzymatic hydrolysis by delignification. Mishima et al. (2006) also showed that the alkaline peroxide pretreatment method was one of the most effective methods for improving the enzymatic hydrolysis. In addition, no measurable furfural and hydroxymethylfurfural (HMF) which are harmful for yeast were detected in the process. That indicates it is easier for yeast to ferment the hydrolyzate compared to dilute-acid pretreatment using alkaline peroxide pretreatment method. Although several pretreatment methods have shown the effectiveness in many researches (Taherzadeh and Karimi, 2008), the combination of two

216

W. Cao et al. / Bioresource Technology 111 (2012) 215–221

or more pretreatment methods may be more effective than separate single one. Moreover, selection of pretreatment process for a certain material depends on the biomass material type. Besides, the enzymatic hydrolysis still needs to be improved when sweet sorghum bagasse is concerned. This study aims to compare the effects of five pretreatment methods on improving enzymatic hydrolysis of sweet sorghum bagasse and the ethanol production from the hydrolyzate. In addition, the lignocellulose compositions and structure changes of the sweet sorghum bagasse were also investigated. 2. Methods 2.1. Sweet sorghum Sweet sorghum (Chongming No. 1) was harvested in Qibao campus, Shanghai Jiao Tong University, China. The stalks were squeezed by a three-roller mill to obtain the liquid phase and bagasse separately. The bagasse was dried in the air and then ground to pass through sieve with 40 meshes. After that it was washed using boiling water three times to substantially remove the major soluble sugars (sucrose, glucose, fructose, etc.) present in the stalk liquid phase. The ratio of bagasse to water was 1:1 on a weight basis. Finally it was dried at 60 °C and stored in plastic bag at room temperature to avoid possible interference in the evaluation of the enzymatic hydrolysis of the cellulose (Wu et al., 2011). 2.2. Pretreatment process Five pretreatment methods were used in this study, labeled as A, B, C, D and E. Untreated bagasse was used as the control in the following experiments. Each pretreatment was conducted duplicated and the results were averaged. 2.2.1. Dilute NaOH solution autoclaving pretreatment (A) Ten grams of dry sweet sorghum bagasse was slurried with 100 mL 2% (w/v) sodium hydroxide solution for 5 min in a 500 mL flask with a silicone stopple and then autoclaved at 121 °C for 60 min. The residues were centrifuged and washed with distilled water until neutral pH was achieved and dried at 60 °C. The yield of product was 5.978 ± 0.212 g. 2.2.2. High concentration NaOH solution immersing pretreatment (B) Ten grams dry sweet sorghum bagasse was slurried with 100 mL 20% (w/v) sodium hydroxide solution for 5 min, and then stood for 2 h in a 500 mL flask with a silicone stopple. The residues were centrifuged and washed with distilled water until neutral pH was achieved and dried at 60 °C. The yield of product was 1.642 ± 0.141 g. 2.2.3. Dilute NaOH solution autoclaving and H2O2 immersing pretreatment (C) Ten grams dry sweet sorghum bagasse was slurried with 100 mL 2% (w/v) sodium hydroxide solution for 5 min in a 500 mL flask with a silicone stopple, and then autoclaved at 121 °C for 60 min. The 5% (w/v) hydrogen peroxide was mixed into the pretreated slurry after cooling down to room temperature. The mixture was kept airtight and in dark place for 24 h. The residues were centrifuged and washed with distilled water until neutral pH was achieved and dried at 60 °C. The yield of product was 5.432 ± 0.038 g. 2.2.4. Alkaline peroxide pretreatment (D) Ten grams dry sweet sorghum bagasse was slurried with 100 mL 2% (w/v) sodium hydroxide solution for 5 min, and then

stood for 2 h in a 500 mL flask with a silicone stopple. The 5% (w/v) hydrogen peroxide was mixed into the pretreated slurry. The mixture was kept airtight and in dark place for 24 h. The residues were centrifuged and washed with distilled water until neutral pH was achieved and dried at 60 °C. The yield of product was 6.856 ± 0.014 g. 2.2.5. Autoclaving pretreatment (E) Ten grams dry sweet sorghum bagasse was slurried with 100 mL distilled water for 5 min in a 500 mL flask with a silicone stopple, and then autoclaved at 121 °C for 60 min. After cooling down to room temperature, the residues were centrifuged and washed with distilled water until neutral pH was achieved and dried at 60 °C. The yield of product was 8.824 ± 0.156 g. 2.3. Enzymatic assay and hydrolysis A commercial cellulase (Celluclast 1.5L, Sigma Aldrich) supplemented with b-glucosidase (Novzymes 188, Denmark) was used for enzymatic hydrolysis of the sweet sorghum bagasse. The activities of Celluclast 1.5L and b-glucosidase were 45.8 FPU/mL and 302.1 U/mL, respectively. One unit of cellulase activity is defined as the amount of the enzyme that releases 1 mg of glucose per minute in the reaction mixture at 50 °C and pH 4.8. One unit of b-glucosidase activity is defined as the amount of the enzyme that releases 1 lmol of p-nitrophenol per minute in the reaction mixture at 50 °C and pH 4.8 (Saha and Cotta, 2006). The mixture was hydrolyzed in sodium citrate buffer (50 mM, pH 4.8) with the substrate loading of 2%. The enzyme loadings of Celluclast 1.5L and b-glucosidase were 20 FPU/g dry biomass and 40 IU/g dry biomass, respectively. The hydrolysis was conducted in a shaker at 50 °C. Samples (1 mL) were withdrawn at 12, 24, 48, 72 and 96 h and centrifuged at 10,000 rpm for 10 min, and the supernatants were preserved at 20 °C until they were used for sugar analysis. 2.4. Microorganism and batch fermentation Active dry yeast was bought from Angel Yeast Company of Hubei Province in China for ethanol fermentation. The medium was as follows: glucose 50 g/L, yeast 5 g/L, peptone 5 g/L, MgSO47H2O 1 g/ L, K2HPO4 1 g/L. 6 M HCl or NaOH solution was used for pH control. The medium was autoclaved at 121 °C for 20 min. The hydrolyzate was sterilized by autoclaving at 121 °C for 20 min before it was inoculated with the yeast medium at the volume ratio of 1:10 of the fermentation broth aseptically. The fermentation tests were conducted at 30 ± 0.5 °C in a shaker at 150 rpm. 2.5. Analytical methods The contents of cellulose, hemicellulose and lignin were determined according to the reference (Van Soest, 1963). The sugars in hydrolyzate including glucose, xylose, arabinose, galactose and mannose were measured with a HPLC system (Shimadzu LC-10A, Japan) with an Aminex HPX-87P column (7.8 mm I.D. 30 cm, Bio-Rad, USA) (Cara et al., 2007). Deionized water was used as the mobile phase with a flow rate of 1 mL/min and column temperature of 80 °C. The activities of the cellulase and b-glucosidase were measured according to the references (Xu et al., 2010). Ethanol concentration was analyzed by gas chromatography (Agilent 7890A GC system, USA) with a flame ionization detector and isopropanol was used as an internal standard (Liu and Shen, 2008). A scanning electron microscope (SEM, Siron 200, FEI Company, USA) was used to detect the microscopic structure of the sweet sorghum bagasse samples. The Fourier Transform Infrared Raman Spectroscopy (FTIR, EQUINOX 55, BRUKER Company, Germany)

217


was recorded between 400 and 4000 cm1. Discs were prepared with 2 mg dried sample and 200 mg KBr (Zhang et al., 2011). The cellulose hydrolysis yield, total sugar yield and dry matter loss are defined as following three equations (Wu et al., 2011).

R1 ¼ C 1 =ðm W 1:11Þ 100ð1ÞR2 ¼ C 2 =ðm 100Þð2ÞR3 ¼ ðm0 mt Þ=m0 100ð3Þ where R1 is cellulose hydrolysis yield, %. C1 is the glucose mass in the hydrolyzate, g. m is the dry sweet sorghum bagasse mass, g. W is the cellulose content in dry sweet sorghum bagasse, %. 1.11 in Eq. (1) is the theoretical conversion from cellulose to glucose. R2 is the total sugar yield, g sugar/100 g dry matter. C2 is the total sugar mass in the hydrolyzate, g. R3 is the dry matter loss, %. m0 is the dry sweet sorghum bagasse mass before pretreatment, g. mt is the dry sweet sorghum bagasse after pretreatment, g. 2.6. Statistical analysis Data were analyzed for statistical significance by a one-way analysis of variance (ANOVA). A 5% probability level (p = 0.05) was used to accept or reject the null hypothesis. Duncan’s multiple range tests at the level of 5% were used to analyze the significances of different pretreatment methods. In this study, means in the tables followed by the same letter with a same column means not significantly different using Duncan’s multiple range tests at the level of 5% (Duncan, 1955). 3. Results and discussion 3.1. The effects of different pretreatment methods on the composition of sweet sorghum bagasse The aim of pretreatments was to change raw material properties, remove or dissolve lignin and hemicellulose and reduce the crystallinity of cellulose (Kumar et al., 2009). To be specific, with perfect pretreatment, the lignin will be mostly degraded while the cellulose and hemicellulose will be retained. The best pretreatment method and condition usually depend mainly on the type of lignocelluloses (Taherzadeh and Karimi, 2008). The dry matter loss has also been applied to evaluate the pretreatment effect in this study. The less dry matter loses, the more potential fermentation substrate is retained (Zhu et al., 2005). Table 1 shows the main composition, acid detergent lignin (ADL) removal and dry matter loss of sweet sorghum bagasse. The results of Table 1 showed that the orders of cellulose content, hemicellulose content, ADL content, ADL removal and dry matter loss were C > A > D > B > E > control, control > E > D > A > B > C, control > E > B > D > A > C, C > A > D > B > E and B > C > A > D > E, respectively. As it could be seen from Table 1, the cellulose content of the bagasse pretreated by method C was the highest one, while its ADL content and hemicellulose content were the lowest one. Dilute NaOH pretreatment of lignocellulose materials has been

found to cause swelling, leading to an increase in internal surface area of the sweet sorghum bagasse and disruption of the lignin structure. Millet reported (1976) that the NaOH pretreatment could decrease lignin content from 24% to 55%. But no effect was observed for softwoods with lignin content greater than 25% at the normal ambient conditions. Elevated temperatures can enhance lignin removal (Taherzadeh and Karimi, 2008). Silverstein et al. (2007) also reported that 2% NaOH in 90 min at 121 °C was the best pretreatment condition, resulting in 65% of delignification. In this study, both methods A and C were conducted at 121 °C, and the pretreatment time was 60 min, which resulted in more than 80% of ADL removal. Thus lignin removal increases enzyme effectiveness by eliminating nonproductive adsorption sites and by increasing access to cellulose and hemicellulose (Kumar et al., 2009). On the other hand, sweet sorghum bagasse pretreated by method A has higher cellulose content and lower ADL lignin than the others except method C. However, the hemicellulose content with method A was significantly higher than method C (p < 0.05). Meantime, the dry matter losses of bagasse pretreated by methods A and C were 40.75% and 46.10%, respectively. Both of which belonged to the medium level among all the pretreatment methods. The pretreatment methods A and C led to a comparatively low dry matter loss but significantly reduced the ADL lignin content and increased the cellulose content, which meant 2% sodium hydroxide pretreatment under autoclaving condition had shown sound effectiveness. The dilute NaOH pretreatment was beneficial for the follow-up enzymatic hydrolysis of the bagasse, which was supported by the results of the references (Wu et al., 2011; Curreli et al., 1997). By comparing methods A and C, it can be deduced that applying hydrogen peroxide during alkaline pretreatment will improve the effectiveness of hemicellulose and ADL lignin removal and cellulose retaining. But the cost of applying hydrogen peroxide is the increase of dry matter loss. The ANOVA showed that there was no significant differences (p > 0.05) in terms of the hemicellulose content between method B and C. Therefore, applying considerable amount of sodium hydrogen will do the same work in degrading and retaining hemicellulose as using small quantity of sodium hydrogen under autoclaving condition plus hydrogen peroxide immersing. However, the ADL lignin content of bagasse pretreated by method B was only superior to the control, which would definitely affect the enzyme hydrolysis process. In addition, the dry matter loss of method B strikingly came up to 83.7%, which would be disadvantageous for potential productivity of the follow-up fermentation. Also, this means that the concentration of 20% NaOH was too high for pretreatment due to the great loss in dry matter. Method D could be considered as the moderate method because of the middle level of its cellulose, hemicellulose, ADL lignin content and the dry matter loss. Compared with method C, method D pretreated raw material in rather low temperature and pressure. So it can be concluded that high temperature and pressure during oxydic alkaline pretreatment will be beneficial for retaining cellulose, removing hemicellulose and ADL lignin, but not good for retaining dry matter. On the other

Table 1 The main composition, ADL removal and dry matter loss of sweet sorghum bagasse. No. Control A B C D E A B

Cellulose (%) f,B

49.78 ± 0.86 78.44 ± 0.35b 69.90 ± 0.09d 82.08 ± 0.48a 72.45 ± 0.15c 54.40 ± 0.27e

Hemicellulose (%) a

27.72 ± 0.08 15.09 ± 0.91d 13.94 ± 0.35d 9.45 ± 0.57e 17.52 ± 0.91c 22.44 ± 0.77b

ADLA (%)

ADL removal (%)

Dry matter loss (%)

10.83 ± 0.18a 1.68 ± 0.16d 7.50 ± 0.29b 0.97 ± 0.12e 2.29 ± 0.01c 10.71 ± 0.06a

Null 84.52 ± 1.24b 30.82 ± 1.48d 91.02 ± 0.98a 78.84 ± 0.23c 1.16 ± 1.11e

Null 40.75 ± 2.19c 83.70 ± 1.56a 46.10 ± 0.57b 32.25 ± 0.21d 13.00 ± 1.70e

Acid detergent lignin. Means in the tables followed by the same letter with a same column means not significantly different using Duncan’s multiple range tests at the level of 5%.

218


hand, when methods A and D are compared, it can be deduced that method A is more effective than method D in retaining cellulose and removing hemicellulose and ADL lignin than hydrogen peroxide. The lowest dry matter loss appeared in method E, but ANOVA showed that compared with control, there was no significant difference on ADL lignin removal (p > 0.05). Liquid hot-water (LHW) pretreatment is an environmental friendly pretreatment method with no addition of chemicals. Dien et al. (2006) reported that 75% of the xylan was dissolved at 160 °C, for 20 min, and higher temperatures, e.g. 220 °C, which can dissolve hemicellulose completely and remove lignin partially with 2 min (Dien et al., 2006; Sreenath et al., 1999). Laser et al. (2002) also showed that most hemicellulose was removed at 170–230 °C for 1–46 min. The pretreatment temperature of method E in this study was 121 °C, which is not sufficient to dissolve the hemicellulose and remove the lignin. Therefore, cellulose and hemicellulose might still be bundled by a great amount of lignin, which would harm the follow-up enzyme hydrolysis and ethanol production (Hendriks and Zeeman, 2009). So when compared with method E and control, it is clear that mere high temperature and high pressure will do limited work on effective pretreatment without alkali being involved. In brief, from the aspect of contents of cellulose, hemicellulose, ADL lignin, and the dry matter loss, methods A and C would be advisable as the pretreatment methods before enzymatic hydrolysis and ethanol fermentation. On the other hand, considering that not all of the yeast for ethanol production can utilize pentose such as xylose. And in such condition, hemicellulose would turn out to be another physical barrier which surrounds the cellulose fibers and can protect the cellulose from enzymatic attack (Wu et al., 2011; Curreli et al., 1997). Therefore, method C would be the best one for the follow-up enzymatic hydrolysis and ethanol formation.

3.2. The effects of different pretreatment methods on sugar concentration

16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

3.2.2. The effects of different pretreatment methods on the xylose concentration Fig. 2 shows the effects of different pretreatment methods on xylose concentration during the enzymatic hydrolysis. As it could be seen from Fig. 2, the variation tendency of xylose concentration in the hydrolyzate was similar to that of glucose concentration, but the xylose concentration was much lower than the glucose concentration in all the pretreatment samples. This was because hemicellulose content in both pretreated and unpretreated bagasse was lower than the cellulose content. Generally speaking, the majority of hydrolyzate of hemicellulose in ligno-cellulosic waste is xylose, and there are minor amounts of arabinose in the hydrolyzate of the bagasse (Taherzadeh and Karimi, 2008). In this sense, the hydrolysis of hemicellulose is related to the xylose concentration in the hydrolyzate. The order of final xylose concentration in hydrolyzate was A > D > C > B > E > control. According to Table 1, there was more hemicellulose while comparatively less lignin remained in

6

Control A B C D E

0

12

24

36

48

60

72

84

96

Hydrolysis time (h) Fig. 1. The effects of different pretreatment methods on glucose concentration during the enzymatic hydrolysis.

Concentration of xylose (mg/mL)

Concentration of glucose (mg/mL)

3.2.1. The effects of different pretreatment methods on glucose concentration Fig. 1 shows the effects of different pretreatment methods on glucose concentration during the enzymatic hydrolysis. As it could be seen from Fig. 1, the glucose concentration in the hydrolyzate increased as the hydrolysis time prolonged. The ANOVA analysis showed that the glucose concentration in the hydrolyzate for all pretreatment methods and control reached a plateau within 24 h (p > 0.05), which indicated that all the glucose production potential had been achieved in this short period (Saha and Cotta, 2006). The order of glucose concentration in the hydrolyzate after 96 h with

different pretreatment methods was C > B > A > D > E > control. Specifically, there was no significant difference of the glucose concentration between method E and the control after hydrolysis for 24 h, although the time for method E to achieve constant concentration of glucose was much shorter than the control. With regard to method C, the glucose concentration was the highest one among all samples after 12 h, which indicated that the pretreatment method C was the most effective one on converting the cellulose in sweet sorghum bagasse to glucose in these five pretreatment methods. The final glucose concentration in method C was up to 14.16 mg/mL, which was 9.8 times as much as the control. The glucose concentration in the hydrolyzate of bagasse pretreated with method A was higher than method B during the first 24 h, but the trend was reverse in the following 24 h. It demonstrated that the pretreatment with low concentration of sodium hydrogen plus autoclaving was much better than mere high concentration of alkali pretreatment for increasing the glucose concentration in hydrolyzate during 48 h. Although the glucose concentration in the hydrolysis pretreated by method D was almost constant after 12 h, the glucose concentration in method D was 26.5% lower than method C. This proved once again that it is much better for the cellulose hydrolysis to use relatively high temperature and pressure than the normal temperature and pressure during the pretreatment process. Both method C and D used alkaline hydrogen peroxide, which is good for improving the cellulose conversion by the removal of hemicellulose and lignin and increasing the accessibility of cellulose (Sreenath et al., 1999). In short, pretreatment method C yielded the highest glucose concentration in the hydrolyzate.

5 4

Control A B C D E

3 2 1 0 0

12

24

36 48 60 72 Hydrolysis time (h)

84

96

Fig. 2. The effects of different pretreatment methods on xylose concentration during the enzymatic hydrolysis.

219


methods A and D than the others, except for method C. And both the hemicellulose and ADL lignin contents of method A were less than method D. But according to Fig. 2, firstly, method A could release xylose much faster than method D, and secondly, the xylose concentration in the hydrolyzate with pretreatment method A was the highest one. So lignin amount left after pretreatment is key factor for hemicellulose degrading during the enzymatic hydrolysis in the whole process. The same conclusion could also be deduced from the results of method B and C. Besides, the xylose concentration in method E was close to the control, and both method E and the control had less xylose in the hydrolyzate. However, both of which had more hemicellulose. This indicated that less hemicellulose in the bagasse of method E and control was hydrolyzed, because much of the cellulose and hemicellulose is unreachable for enzyme under the wrap of remained lignin (Beukes and Pletschke, 2011). Accordingly, xylose concentration with method A was the highest one in all pretreatment methods due to the reasons that method A yields a relatively high content of hemicellulose and low content of ADL lignin. 3.3. The effects of different pretreatment methods on enzymatic hydrolysis and ethanol fermentation of sweet sorghum bagasse Table 2 shows the effects of different pretreatment methods on cellulose hydrolysis yield, total sugar yield and ethanol concentration of sweet sorghum bagasse. The cellulose hydrolysis yield implies the efficiency of converting cellulose to glucose. The order of cellulose hydrolysis yield of sweet sorghum bagasse with different pretreatment methods for 96 h hydrolysis was C > B > A > D > control > E. This order was almost the same as that of glucose concentration in the hydrolyzate after 96 h. It indicated that alkaline pretreatment was good for improving the cellulose conversion. The cellulose hydrolysis yield of method C was 74.29%, which was 5.88 times higher than that of the control. It was higher than the result (70%) obtained from the steam pretreatment of Zhang’s research (Zhang et al., 2011). That may be due to the reasons that the content of hemicellulose and lignin was lower in the pretreated bagasse, and more cellulose was converted to glucose. Different pretreatment conditions can result in different pretreatment effects. In low alkaline concentration processes, with the concentration of 0.5– 4% NaOH at high temperature and pressure, the structure of lignocellulose can be destroyed in the pretreatment process while NaOH pretreatment at high temperature, the majority of the lignin and hemicellulose can be removed from the solid phase (Mirahmadi et al., 2010). On the other hand, the high concentration NaOH pretreatment (6–20%) is usually used at ambient conditions. Lignin was not significantly removed from the biomass (Mirahmadi et al., 2010). The ANOVA showed there was no significant difference of the cellulose hydrolysis yield between method E and the control. Also, it showed that mere autoclaving with distilled water cannot significantly improve the cellulose hydrolysis yield. Method E in this study was not similar to LHW pretreatment, because the pretreatment temperature (121 °C) of method E in this study was lower than the LHW pretreatment (160–260 °C). Therefore, it was not enough for method E to dissolve the hemicellulose and remove the lignin

for improving the cellulose hydrolysis yield (Sreenath et al., 1999). Totally, the pretreatment method C was the most effective one on increasing the cellulose hydrolysis yield in five pretreatment methods. During the enzymatic hydrolysis of sweet sorghum bagasse, the cellulose was converted to glucose, and the hemicellulose was converted to xylose, arabinose, mannose, galactose and other sugars. Table 3 shows the sugar concentration in the hydrolyzate after enzymatic hydrolysis for 96 h. As it could be seen from Table 3, there were minor amounts of arabinose in the hydrolyzate of the bagasse pretreated by methods A, B, C and D, while no arabinose was detected in the hydrolyzate of bagasse pretreated by method E and the control. The order of total sugar concentration in the hydrolyzate is C > A > D > B > E > control. According to Table 2, the order of total sugar yield of different pretreatment method was C > A > D > B > E > control. This order was exactly the same as that of cellulose contents after pretreatments. The total sugar yield with the method C was 9.54 times higher than the control. It is indicated that the total sugar yield is relevant to the residual cellulose content after pretreatment. Besides, total sugar yields of method A, D, and B were all more than 69.61 g sugar/100 g dry matter. Thus pretreatment method A, B and D can effectively improve the total sugar yield of sweet sorghum bagasse compared to others. The total sugar yield of method E was only 10.12 g sugar/100 g dry matter and there was no significant difference between method E and the control. Because of the addition of 2% NaOH solution, most of the hemicellulose and lignin in the bagasse with pretreatment method C, A and D were removed and the cellulose can be more easily converted to glucose by cellulase (Taherzadeh and Karimi, 2008). As a result, the total sugar yield was increased. The total sugar yield of bagasse pretreated by method B (69.61 g sugar/100 g dry matter) showed that relatively high concentration of NaOH solution (20%) had positive effect on the saccharification of sweet sorghum bagasse. But it had relatively high dry matter loss according to Table 1 (83.70%) which was not an ideal pretreatment method compared with the others. The objectives of sweet sorghum bagasse pretreatment were to improve the cellulose hydrolysis yield and total sugar yield, and to produce more glucose for the follow-up ethanol fermentation. As can be seen from Table 2, the order of the ethanol concentration in fermentation broth is C > B > A > D > E > control. This order was the same as those of glucose concentration in the hydrolyzate and cellulose hydrolysis yield. The ethanol concentration with the method C was 19.13 times higher than the control. So it can be inferred that the amount of cellulose content remained in the pretreatment residue may be a kind of criterion for telling the cellulose hydrolysis yield as well as the amount of ethanol concentration in the broth after fermentation. Due to high cellulose hydrolysis yield and high total sugar yield, the ethanol concentration of method C inevitably was the highest one in all methods. 3.4. SEM photos analysis Scanning electron microscope photos of unpretreated and pretreated sweet sorghum bagasse is shown in Supplemental Fig. 1.

Table 2 The effects of different pretreatment methods on cellulose hydrolysis yield, total sugar yield and ethanol concentration of sweet sorghum bagasse.

A

No.

Cellulose hydrolysis yield (%)

Total sugar yield (g sugar/100 g dry matter)

Ethanol concentration (g/L)

Control A B C D E

12.64 ± 0.19e,A 69.94 ± 0.93c 72.89 ± 0.12b 74.29 ± 0.81a 62.46 ± 0.56d 11.78 ± 0.19e

9.53 ± 0.02e 86.83 ± 0.68b 69.61 ± 0.90d 90.94 ± 0.39a 76.06 ± 0.17c 10.12 ± 0.03e

0.32 ± 0.01e 5.55 ± 0.23b 5.73 ± 0.06b 6.12 ± 0.06a 4.93 ± 0.16c 0.89 ± 0.04d

Means in the tables followed by the same letter with a same column means not significantly different using Duncan’s multiple range tests at the level of 5%.

220


Table 3 The sugar concentration in the hydrolyzate after enzymatic hydrolysis for 96 h. No.

Control A B C D E A

Sugar concentration (mg/mL) Glucose

Xylose

Arabinose

Mannose

Galactose

Total sugar

1.45 ± 0.00 12.55 ± 0.41 12.45 ± 0.20 14.16 ± 0.00 10.41 ± 0.03 1.46 ± 0.03

0.53 ± 0.00 4.98 ± 0.15 2.56 ± 0.11 4.57 ± 0.00 4.98 ± 0.06 0.62 ± 0.03

NDA 0.35 ± 0.00 0.30 ± 0.00 0.28 ± 0.01 0.36 ± 0.00 ND

ND ND ND ND ND ND

ND ND ND ND ND ND

1.97 ± 0.00 17.88 ± 0.56 15.30 ± 0.09 19.01 ± 0.01 15.75 ± 0.10 2.08 ± 0.00

ND means not detected.

The result of Supplemental Fig. 1 indicated that there was an apparent net structure in sweet sorghum bagasse pretreated by method E and the control. This complex net structure in the bagasse will restrict the cellulose being attacked by cellulase (Fan et al., 1980), while it cannot be found in other samples. Autoclaving pretreatment of method E changed the structure a little, but the net structure could not be changed constitutionally. The alkali autoclaving pretreatment such as methods A and C could degrade the lignin and hemicellulose in the bagasse. The cellulose hydrolysis yield is closely related to the removal of hemicellulose and lignin (Yu et al., 2010). The structure of bagasse pretreated by method B using 20% NaOH solution was destroyed thoroughly because relatively high concentration of NaOH could dissolve the hemicellulose. The SEM photo of the bagasse pretreated by method D seemed different from those of method A, B and C, and it showed a smooth surface of the bagasse. This suggested that the pretreatment temperature or time might be not enough to destroy the bagasse thoroughly. Also, bigger pores could be found in bagasse pretreated by method C and smaller pores could found in bagasse pretreated by method D. Both of which were applied hydrogen peroxide. The analysis of the SEM photos indicated that there were distinct changes in the bagasse pretreated by method A, B, C and D. This was also corresponding to the analysis of the cellulose hydrolysis yield and total sugar yield.

lulose hydrolysis yield, total sugar yield and ethanol concentration were 74.29%, 90.94 g sugar/100 g dry matter and 6.12 g/L, respectively, which were 5.88, 9.54 and 19.13 times higher than the control. Some significant structure and chemical bonds changes after pretreatment were found. Cellulose content and total sugar yield were related to each other, so were glucose concentration, cellulose hydrolysis yield and ethanol concentration. Acknowledgements The authors are grateful for the support provided by the Ministry of Agriculture of the People’s Republic of China under the 948 Project entitled ‘‘Introduce and industrialization of advanced agriculture biomass energy technologies’’ (Grant No. 2008 G2) and Liaoning Natural Science Foundation (Grant No. 20092031). In addition, Yue Zhang from University of Southampton of UK is greatly acknowledged for her valuable suggestion and correction of the manuscript. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.biortech.2012.02.034. References

3.5. FTIR analysis Supplemental Fig. 2 shows the FTIR spectra of untreated and pretreated sweet sorghum bagasse. As can be seen from Supplemental Fig. 2, there were no new peaks in the spectra of unpretreated and pretreated sweet sorghum bagasse, but there were some differences in parts of peak transmittance. The absorption at 3400 cm1 represents the stretching of hydroxyl and phenol in the bagasse. After pretreatment, the absorption reduced, which indicated that less hydroxyl and phenol existed compared to the control. The absorption at 2910 cm1 represents the stretching of –CH3 and –CH2. The transmittances of the bagasse pretreated by method A, B and C were less than others, which were indicators of fracture of carbon chains. The absorption at 1630 cm1 represents the bending mode stretching of the absorbed water and stretching of C@O in lignin (Zhang et al., 2011). The strong absorption at 1057 cm1 represents the stretching of C–O in cellulose and hemicellulose (Oh et al., 2005; Cao and Tan, 2004). The transmittances of the bagasse pretreated by method C, B and A were less than others, which are indicators of the structural destroy of the bagasse after pretreatment. 4. Conclusions In this study, dilute NaOH solution autoclaving and H2O2 immersing pretreatment was considered as the most suitable method for sweet sorghum bagasse pretreatment. The highest cel-

Billa, E., Koullas, D.P., Monties, B., Koukios, E.G., 1997. Structure and composition of sweet sorghum stalk components. Ind. Crop. Prod. 6, 297–302. Beukes, N., Pletschke, B.I., 2011. Effect of alkaline pre-treatment on enzyme synergy for efficient hemicellulose hydrolysis in sugarcane bagasse. Bioresour. Technol. 102, 5207–5213. Curreli, N., Fadda, M., Rescigno, A., Rinaldi, A.C., Soddu, G., Sollai, F., Vaccargiu, S., Sanjust, E., Rinaldi, A., 1997. Mild alkaline/oxidative pretreatment of wheat straw. Process. Biochem. 32, 665–670. Cao, Y., Tan, H., 2004. Structural characterization of cellulose with enzymatic treatment. J. Mol. Struct. 705, 189–193. Cara, C., Moya, M., Ballesteros, I., Negro, M., González, A., Ruiz, E., 2007. Influence of solid loading on enzymatic hydrolysis of steam exploded or liquid hot water pretreated olive tree biomass. Process. Biochem. 42, 1003–1009. Duncan, D.B., 1955. Multiple range and multiple F tests. Biometrics 11, 1–42. Dien, B.S., Li, X.L., Iten, L.B., Jordan, D.B., Nichols, N.N., O’Bryan, P.J., Cotta, M.A., 2006. Enzymatic saccharification of hot-water pretreated corn fiber production of monosaccharides. Enzyme Microb. Tech. 39, 1137–1144. Fan, L.T., Lee, Y., Beardmore, D.H., 1980. Mechanism of the enzymatic hydrolysis of cellulose: effect of major structural features of cellulose on enzymatic hydrolysis. Biotechnol. Bioeng. 22, 177–199. Hendriks, A.T.W.M., Zeeman, G., 2009. Pretreatments to enhance the digestibility of lignocellulosic biomass. Bioresour. Technol. 100, 10–18. Kumar, P., Barrett, D.M., Delwiche, M.J., Stroeve, P., 2009. Methods for pretreatment of lignocellulosic biomass for efficient hydrolysis and biofuel production. Ind. Eng. Chem. Res. 48, 3713–3729. Laser, M., Schulman, D., Allen, S.G., Lichwa, J., Antal, M.J., Lynd, L.R., 2002. A comparison of liquid hot water and steam pretreatments of sugarcane bagasse for to ethanol. Bioresour. Technol. 82, 33–34. Lin, Y., Tanaka, S., 2006. Ethanol fermentation from biomass resources: current state and prospects. Appl. Microbiol. Biot. 69, 627–642. Liu, R., Shen, F., 2008. Impacts of main factors on bioethanol fermentation from stalk juice of sweet sorghum by immobilized Saccharomyces cerevisiae (CICC1308). Bioresour. Technol. 99, 847–854. Millet, M.A., Baker, A.J., Scatter, L.D., 1976. Physical and chemical pretreatment for enhancing cellulose saccharification. Biotech. Bioeng. Symp. 6, 125–153.

W. Cao et al. / Bioresource Technology 111 (2012) 215–221 Mais, U., Esteghlalian, A.R., Saddler, J.N., Mansfield, S.D., 2002. Enhancing the enzymatic hydrolysis of cellulosic materials using simultaneous ball milling. Appl. Biochem. Biotech. 98, 815–832. Mishima, D., Tateda, M., Ike, M., Fujita, M., 2006. Comparative study on chemical pretreatments to accelerate enzymatic hydrolysis of aquatic macrophyte biomass used in water purification processes. Bioresour. Technol. 97, 2166– 2172. Mirahmadi, K., Kabir, M.M., Jeihanipour, A., Karimi, K., Taherzadeh, A., Mohammad, J., 2010. Alkaline pretreatment of spruce and birch to improve bioethanol and biogas production. Bioresources 5, 928–938. Oh, S.Y., Yoo, D.L., Kim, H.C., Kim, H.Y., Chung, Y.S., et al., 2005. Crystalline structure analysis of cellulose treated with sodium hydroxide and carbon dioxide by means of X-ray diffraction and FTIR spectroscopy. Carbohydr. Res. 340, 2376– 2391. Sreenath, H.K., Koegel, R.G., Moldes, A.B., Jeffries, T.W., Straub, R.J., 1999. Enzymatic saccharification of alfalfa fiber after liquid hot water pretreatment. Process Biochem. 35, 33–41. Saha, B.C., Iten, L.B., Cotta, M.A., Wu, Y.V., 2005. Diluted acid pretreatment, enzymatic saccharification and fermentation of wheat straw to ethanol. Process Biochem. 40, 3693–3700. Saha, B.C., Cotta, M.A., 2006. Ethanol Production from alkaline peroxide pretreated enzymatically saccharified wheat straw. Biotechnol. Prog. 22, 449–453. Silverstein, R.A., Chen, Y., Sharma-Shivappa, R.R., Boyette, M.D., Osborne, J., 2007. A. comparison of chemical pretreatment methods for improving saccharification of cotton stalks. Bioresour. Technol. 98, 3000–3011. Sipos, B., Réczey, J., Somorai, Z., Kádár, Z., Dienes, D., Réczey, K., 2009. Sweet sorghum as feedstock for ethanol production: enzymatic hydrolysis of steampretreated bagasse. Appl. Biochem. Biotech. 153, 151–162.

221

Silva, A.S., Inoue, H., Endo, T., Yano, S., Bon, E.P., 2010. Milling pretreatment of sugarcane bagasse and straw for enzymatic hydrolysis and ethanol fermentational. Bioresour. Technol. 101, 7402–7409. Shen, F., Saddler, J.N., Liu, R., Lin, L., Deng, S., Zhang, Y., Yang, G., Xiao, H., Li, Y., 2011. Evaluation of steam pretreatment on sweet sorghum bagasse of enzymatic hydrolysis and bioethanol production. Carbohyd. Polym. 86, 1542–1548. Taherzadeh, M.J., Karimi, K., 2008. Pretreatment of lignocellulosic waster to improve ethanol and biogas production: a review. Int. J. Mol. Sci. 9, 1621–1651. Van Soest, P.J., 1963. Use of detergents in the analysis of fibrous feeds. II. A rapid method for the determination of fiber and lignin. Ass. Offic. Agr. Chem. 46, 829– 835. Wu, L., Arakane, M., Ike, M., Wada, M., Takai, T., Gau, M., Tokuyasu, K., 2011. Low temperature alkali pretreatment for improving enzymatic digestibility of sweet sorghum bagasse for ethanol production. Bioresour. Technol. 102, 4793–4799. Xu, J., Cheng, J.J., Sharma-Shivappa, R.R., Burns, J.C., 2010. Sodium hydroxide pretreatment of switchgrass for ethanol production. Energy Fuels 24, 2113– 2119. Yamashita, Y., Shono, M., Sasaki, C., Nakamura, Y., 2010. Alkaline peroxide pretreatment for efficient enzymatic saccharification of bamboo. Carbohyd. Polym. 79, 914–920. Yu, Q., Zhuang, X., Yuan, Z., Wang, Q., Qi, W., Wang, W., 2010. Two-step liquid hot water pretreatment of Eucalyptus grandis to enhance sugar revpvery and enzymatic digestibility of cellulose. Bioresour Technol. 101, 4895–4899. Zhu, S., Wu, Y., Yu, Z., Liao, J., Zhang, Y., 2005. Pretreatment by microwave/alkali of rice straw and its enzymic hydrolysis. Process Biochem. 40, 3082–3086. Zhang, J., Ma, X., Yu, J., Zhang, X., Tan, T., 2011. The effects of four different pretreatments on enzymatic hydrolysis of sweet sorghum bagasse. Bioresour. Technol 102, 4585–4589.

Waste Biomass Valor DOI 10.1007/s12649-012-9183-x

ORIGINAL PAPER

Effect of Pretreatment of Sweet Sorghum Biomass on Methane Generation Georgia Antonopoulou • Gerasimos Lyberatos

Received: 20 July 2012 / Accepted: 24 November 2012 Ó Springer Science+Business Media Dordrecht 2012

Abstract In this work, the effect of pretreatment on the biochemical methane potential (BMP) of sweet sorghum biomass was determined. Various pretreatment methods, such as thermal (1 h at 121 °C), enzymatic [through the addition of the enzyme Celluclast 1.5L (Cellulase from Trichoderma reesei, ATCC 26921) or by the addition of a mixture of Celluclast 1.5L and Novozyme 188 (Cellobiase from Aspergillus niger) at a ratio of (3:1)], chemical [through alkali (NaOH) or acid (H2SO4) addition, at concentrations of 0–2 % w/v] or combination of the above methods (thermal acid and thermal alkaline) were tested, in order to evaluate their effect on carbohydrate solubilization (saccharification) and on the methane yield. The experimental results showed that thermal acid treatment and enzymatic treatment for all enzymes concentrations tested improved saccharification. Under thermal alkaline treatment at NaOH concentrations above 0.5 % w/v, a significant decrease in soluble carbohydrates concentration was observed meaning that a high portion of sugars also contained in sorghum biomass was degraded or was transformed into other components. BMP experiments showed that the chemical pretreatment methods did not enhance methane generation compared to the raw substrates. This could be attributed either to inhibitory compounds released during pretreatment, or to high salts (cations of sodium during alkaline treatment) concentration, causing in both cases methanogenic bacteria inhibition. On the other hand, G. Antonopoulou (&) G. Lyberatos Institute of Chemical Engineering Sciences, ICE-HT/FORTH, P.O. Box 1414, 26504 Patras, Greece e-mail: [email protected] G. Lyberatos School of Chemical Engineering, National Technical University of Athens, 15780 Athens, Greece

thermal treatment improved the methane yield from 253 to 288 LCH4/kg sorghum. During enzymatic pretreatment, the methane production was enhanced either with only one or with the mixture of enzymes. Keywords Pre-treatment Chemical Enzymatic Anaerobic digestion Methane Sweet sorghum biomass

Introduction Sweet sorghum (Sorghum bicolor) is a C4, highly productive crop, with a high photosynthetic efficiency. It is a heat- and drought- tolerant plant with low cultural requirements, making efficient use of soil moisture [1]. It has the ability to withstand extended dry periods and to be tolerant of poor soil conditions and low nutrient levels, without affecting its quantitative and qualitative characteristics [2]. Sweet sorghum is characterized as a suitable feedstock for a variety of bioprocesses, mainly because of its high yields of both lignocellulosic biomass and fermentable saccharides [3]. Sweet sorghum can be fractionated via sugar extraction of the stalks to a liquid fraction, rich in fermentable sugars (sorghum extract or juice) and a solid fraction, rich in cellulose and hemicellulose (sorghum bagasse) [4–6]. To date, ethanol and hydrogen are the best-known microbial products produced from sweet sorghum extract [5, 7–10]. In addition, methane generation from the effluent of a hydrogen producing reactor using sorghum extract as substrate, in a two-stage process has also been demonstrated [5, 11]. Fermentative production of hydrogen from sweet sorghum bagasse using the extreme thermophilic bacterium Caldicellulosiruptor saccharolyticus has recently been studied

123

Waste Biomass Valor

[12]. However, little attention has been given so far to the utilization of the whole lignocellulosic material of sweet sorghum biomass for biofuels production. Ivanova et al. [13] studied the production of hydrogen from homogenized whole plants of sweet sorghum using the C. saccharolyticus. The main obstacle using the whole plant or the bagasse of sweet sorghum instead of the extract is the low efficiencies and yields attained, due to the recalcitrant nature of their lignocellulosic content. With the application of a pretreatment process, the major compounds, cellulose and hemicellulose, can be hydrolyzed into fermentable sugars, which are actually the key molecules for subsequent bioconversions to biofuels. Microwave [14] and steam pretreatment [15] are among the thermal treatment methods that have been applied so far on sweet sorghum bagasse using high temperatures. It was found that during both methods, by-products inhibitory to fermentation such as furfural, acetic acid, and formic acid were generated. Chemical pretreatment such as dilute acid treatment has also been tested in sweet sorghum bagasse, causing an increase in hemicellulose hydrolysis rate and lignin solubilization, at lower temperatures (80–100 °C). Acids such as sulphuric [16–18], hydrochloric [19], and phosphoric acid, respectively [20–22] have been used so far for the treatment of sorghum bagasse, achieving hemicellulose degradation between 70 and 80 %, depending on the treatment conditions. Furthermore, alkaline pretreatment by sodium hydroxide [23–25] or lime addition [26] was applied on sorghum residues such as bagasse or field wastes. In addition to these methods, sugars’ solubilization through de-polymerization of polysaccharides, particularly cellulose, could be accomplished through enzymatic pretreatment/hydrolysis. For example, Dogaris et al. [14] used an enzymatic scheme with cellulases derived from both Fusarium oxysporum and Neurospora crassa which hydrolyzed sorghum bagasse giving high sugar yields. It should be emphasized that a particular pretreatment type, while appropriate for a particular bioconversion (such as alcoholic fermentation), may be completely inappropriate for another (such as anaerobic digestion), due to the varying sensitivity of the key microbial species involved to the inhibitory intermediates that are generated. In this work, the whole plant of sweet sorghum biomass was used for biogas (methane) production through a batch anaerobic digestion process and alternative pretreatment methods, such as thermal, enzymatic, chemical (through alkali or acid addition) or combinations thereof were applied. It is the first time that alternative pretreatment methods are examined for methane generation from the lignocellulosic part of sweet sorghum biomass. The effectiveness of each method for carbohydrates’ solubilization and methane productivity and yield was evaluated.

123

Materials and Methods Analytical Methods The measurements of total chemical oxygen demand (T.COD), dissolved chemical oxygen demand (d.COD), total solids (TS) and volatile solids (VS), total suspended solids (TSS) and volatile suspended solids (VSS), were carried out according to Standard Methods [27]. The methane content of the produced gas was quantified with a gas chromatograph (SRI 8610c MG#1) (two columns in series: molecular sieve column, 6 ft., O.D. 1/8 in., I.D. 2.1 mm and silica gel column, 6 ft., O.D. 1/8 in.) equipped with a TCD (thermal conductivity detector). The column oven temperature was 80 °C, the injector valve 90 °C and the TCD oven 100 °C. Helium was used as carrier gas at 20 mL/min. For the determination of carbohydrates, a colored sugar derivative was produced through the addition of L-tryptophan and sulphuric and boric acids and subsequently measured colorimetrically at 520 nm [28]. Inoculum Anaerobic sludge from the anaerobic digester of the Patras wastewater treatment plant, treating municipal sewage sludge and operating at steady state at an HRT of 15 days, was used as inoculum. The main characteristics of the sludge were: pH: 7.7, T.COD: 34.3 g/L, d.COD: 0.45 g/L, TSS: 25.7 g/L and VSS: 12.8 g/L. Feedstock Sweet sorghum biomass (Sorghum bicolor L. Moench) was used in the present study. After plant harvesting, the fresh stems were stripped from their leaves and the stalks were chopped to a size of 20 cm, stored in the freezer at -20 °C and milled by a laboratory grinder to an average particle size of 1–2 mm. A physicochemical characterization of sweet sorghum biomass was carried out and the main measured characteristics are presented in Table 1. Prior to use, samples were grinded to the desired particle size by a small laboratory scale mill and were sieved to powder of \1 mm diameter. In the sequel, sweet sorghum biomass powder was dried at 55 °C and used for the experiments.

Table 1 The main characteristics of sweet sorghum biomass used in this study Characteristic

Value

Dry matter (wt%)

90.2

Volatile solids (wt% dry basis)

76.2

Chemical oxygen demand (g O2/g dry basis)

1.0

Waste Biomass Valor

Pretreatment For all pretreatment methods the mass/volume ratio of solid (g) to liquid (mL) was 5:100. Thermal pretreatment was conducted at 121 °C for 60 min in an autoclave. Acid or alkali pretreatment of the feedstock was conducted by the addition of aqueous solution of H2SO4 or NaOH at different concentrations [0.1, 0.5, 1, 1.5 and 2 % (w/v), respectively] for 60 min at a temperature of 25 °C (room temperature) (acid and alkalkine pretreatment, respectively), or at 121 °C for 60 min in an autoclave (thermal acid or thermal alkali pretreatment, respectively). The enzymatic pretreatment was performed using commercial enzymes, either Celluclast 1.5L (Cellulase from Trichoderma reesei, ATCC 26921) or a mixture of Celluclast 1.5L and Novozyme 188 (Cellobiase from Aspergillus niger) at a ratio of (3:1) [29]. The filter paper activity (FPU) of cellulase (Celluclast 1.5L) was measured with the improved method developed by Ghose [30]. The effect of enzymatic pretreatment on saccharification of sorghum biomass was evaluated for different enzyme concentrations (20, 40, 60, 80, 100 and 150 FPU of Celluclast 1.5L/gTS of sorghum biomass, respectively) by measuring soluble carbohydrates in the hydrolyzate. In the case that the mixture of the enzymes was used, the aforementioned concentrations of Celluclast 1.5L were used and the ratio of Celluclast 1.5L and Novozyme 188 of 3:1 was kept constant. The samples containing sorghum biomass and the enzymes were incubated for 1 day at 50 °C and at a pH value of 4.8, which is suggested as optimum for the enzymes (Novozyme). All pretreatment experiments were carried out in duplicates and at the end of each experiment, soluble carbohydrates were measured and carbohydrate solubilisation was expressed in glucose equivalents. Biochemical Methane Potential (BMP) Assays BMP experiments were carried out in duplicate at 35 °C according to Owen and Chynoweth [31]. Serum bottles of 160 mL were seeded with 20 mL of mixed anaerobic culture, 76 mL water and the 4 mL of the mixed liquor of the enzymatically, thermally, acid- or alkaline-pretreated feedstocks composed of 0.2154 g sweet sorghum (SS) and the mixture of enzymes, water, or H2SO4 or NaOH, (depending on the experiment) thermally or not thermally treated. The final solids content in the vials was 2 g TS/L. The microbial culture was supplemented with 10 mL/L of a solution of (NH4)2HPO4 (7.21 g/L), 10 mL/L of a solution of FeSO47H2O (0.7 g/L) and 10 mL/L of a solution with trace metals [32]. Prior to the BMP experiments, the acid or alkali treated feedstocks (thermally treated or not) were neutralized with concentrated NaOH or H2SO4 (6 N) respectively, to a final pH of 7. Control experiments using

glucose instead of sorghum biomass were carried out for checking the methanogenic biomass activity. Blank experiments were also carried out in order to determine the background gas productivity of the inoculum. The content of the vials was gassed with a gas mixture of N2/CO2 (80/20) in order to secure anaerobic conditions. The vials were sealed with butyl rubber stoppers and aluminum crimps and methane production was monitored as a function of time according to Owen and Chynoweth [32].

Results and Discussion Effect of Acid Pretreatment While acid pretreatment is a common strategy for bioethanol production, up to now few studies have been published on the impact of acid pretreatment on the anaerobic digestion of lignocellulosic biomass. The main reaction during acid pretreatment is the hydrolysis of hemicellulose into xylose [33], which is the main component of hemicellulose [34]. In this study, acid pretreatment was carried out through the addition of H2SO4 at a concentration range of 0.1–2 % w/v. Acid treatment was conducted for 1 h, either at 25 °C (room temperature) (acid treatment) or at 121 °C (thermal acid pretreatment). The effect of acid pretreatment as well as of thermal acid pretreatment on soluble carbohydrates’ concentration expressed in glucose equivalents (accompanied by their standard deviations), is shown in Fig. 1. Acid pretreatment at mild conditions (25 °C for 1 h) did not affect carbohydrates’ solubilization even at high acid concentrations. However, from Fig. 1b it is obvious that the combination of acid with thermal treatment caused an increase in soluble carbohydrates’ concentration, reaching, for example, for 1 % w/v H2SO4 an increase in sugars concentration of 32 % compared to the untreated sorghum biomass. The effect of both pretreatment methods on methane yields expressed in L methane per kg of sweet sorghum (accompanied by their standard deviations), are presented in Fig. 2. Both pretreatment methods caused a decrease in methane production, with the decrease being higher when acid treatment was carried out at high temperatures. Moreover, the higher the acid concentration, the lower was the observed methane yield. The concentration of 0.1 % w/v H2SO4 did not affect methane production, compared to the untreated sample. The decrease was high for acid concentrations above 0.5 %w/v. This means that using even higher sulfuric acid concentrations than those which were tested (higher than 2 % w/v), the methane yields would be much lower than the observed ones. Thermal treatment boosted the observed trend, since the methane yield upon treatment with 2 % w/v H2SO4 at room temperature was

123

o

H2SO4 at 25 C

400

300

200

100

0.1 %

0.5 %

1%

1.5 %

o

b 400

300

200

100

0

TT

0.1%

0.5%

1%

1.5%

2%

120.6 ± 24.4, while at 121 °C it was 60.9 ± 9.9 LCH4/kg sorghum biomass. It is well known that acid hydrolysis of hemicellulose is accompanied by furfural and hydoxymethyl furfural (HMF) production due to dehydration of pentoses and hexoses. These compounds along with lignin polymers such as vanillin, or aliphatic acids such as formic and levulinic acids, may have a toxic effect on bacteria like methanogens [35, 36]. The concentration and diversity of these toxic compounds vary according to the type of raw material and to the pretreatment conditions applied [37]. For example, Vazquez et al. [20] observed high levels of furfural production during phosphoric acid pretreatment of sweet sorghum straws. In our case, the decrease in methane yields could be attributed to the inhibitory compounds, which probably were released during the tested pretreatment conditions, causing methanogenic bacteria inhibition. Similar results were also observed by Antonopoulou et al. [38], who applied acid pretreatment (H2SO4 2 %w/v, 1 h either at room temperature or at 121 °C) on rapeseed and

SSB

0.1 %

0.5 %

1%

b

1.5 %

2% o

H2SO4 at 121 C

300

200

100

0

SSB

Fig. 1 The effect of acid pretreatment (H2SO4 at 25 °C) and thermal acid prtreatment (H2SO4 at 121 °C) of sweet sorghum biomass on soluble carbohydrates concentration expressed in glucose equivalents. SSB means untreated sweet sorghum biomass and TT thermal treated sweet sorghum, respectively

123

100

2%

H2SO4 at 121 C

o

H2SO4 at 25 C

200

400

500

a

300

0

0

SSB

Soluble carbohydrates (mg/g SSB)

400

a

Methane yield (LCH4 / kg SSB)

500

Methane yield (LCH4 /kg SSB)


Waste Biomass Valor

SSB

TT

0.1%

0.5%

1%

1.5%

2%

Fig. 2 The effect of acid pretreatment (H2SO4 at 25 °C) and thermal acid prtreatment (H2SO4 at 121 °C) of sweet sorghum biomass on methane yield. SSB means untreated sweet sorghum biomass and TT thermal treated sweet sorghum, respectively

sunflower straws, which are the residues after plants harvesting, and rapeseed and sunflower meals, which are the solid wastes from an oil extraction process. The experiments showed that neither acid treatment at room temperature nor acid treatment at high temperature improved the methane yield from all substrates. This could be attributed to a possible inhibition of methanogens by toxic compounds released during the pretreatment. During acid treatment at high temperatures, the liberation of such toxic compounds is more intense, justifying the even lower methane yields. Similar results were reported in other studies, using different acids or different substrates for studying the acid treatment effect on methane production and yield. Fernandes et al. [39] studied the effect of maleic acid (150 °C, 30 min) on the biodegradability, methane yield and hydrolysis rate of three plant species; hay, straw and bracken. The experiments showed that both the biodegradability and the methane production of hay and straw were not enhanced by the treatment with maleic acid, while the acid treatment, only in the case of bracken improved methane production. On the contrary, Nieves et al. [40]

Effect of Alkaline Pretreatment Alkaline pretreatment removes lignin and a part of the hemicellulose, and thus efficiently increases the accessibility of enzymes to the cellulose. The effectiveness of alkaline treatment depends on the lignin content of the biomass [43]. In general, alkali pretreatment can result in a sharp increase in saccharification yields due to the destruction of links between lignin and other polymers and due to lignin breakdown [44]. Up to now, the alkaline treatment has been applied as a pretreatment method for biogas production from different kinds of lignocellulosic biomasses [45–47]. The effect of alkaline pretreatment as well as of thermal alkaline pretreatment on carbohydrates’ concentration is shown in Fig. 3. Alkaline pretreatment at 25 °C at all NaOH concentrations did not enhance carbohydrates’ solubilization of sorghum biomass, since the concentration of sugars was not affected at all. The combination with thermal treatment (Fig. 3b) caused a significant decrease in soluble carbohydrates concentration, especially at concentrations of NaOH of 0.5–2 %w/v. This means that a high portion of sugars also contained in sorghum biomass, was degraded or converted into other components under these conditions. Similar behavior was observed by McIntosh and Vancov [23], who applied alkaline pretreatment with sodium hydroxide on grain sorghum stover. They found that during alkaline treatment at 121 °C, little of the sucrose initially present was degraded, indicating that thermal alkaline pretreatment of materials which contain relatively high levels of soluble sugars, such as sweet sorghum and its bagasse is inappropriate.

400

a

o

NaOH at 25 C

300

200

100

0


achieved a 40 % improvement in the methane yield when using phosphoric acid for the treatment of oil palm empty fruit bunches, a lignocellulosic by-product of vegetable oil production industries. This means that different acid pretreatment methods affect different substrates in a different way. In this respect, a pretreatment method that is routinely used in biotechnological processes, such as bioethanol production, may be inappropriate as a step prior to anaerobic digestion. From Fig. 2, it becomes obvious that only thermal treatment, without the addition of chemicals led to a 13.8 % increase of methane yield (from 253 to 288 LCH4/ kg sorghum). During thermal treatment, the sugars’ concentration increase was approximately 3 %. However, when the lignocellulosic biomass is thermally pretreated, part of the hemicellulose is hydrolyzed, forming acids which subsequently act as catalysts for further hydrolysis of hemicelluloses [41, 42]. These dissolved compounds could be converted to methane, justifying the higher methane yields.


Waste Biomass Valor

SSB

0.1 %

0.5 %

1%

1.5 %

2%

400

b

NaOH at 121o C

300

200

100

0 SSB

TT

0.1 %

0.5 %

1%

1.5 %

2%

Fig. 3 The effect of alkaline pretreatment (NaOH at 25 °C) and thermal alkaline prtreatment (NaOH at 121 °C) of sweet sorghum biomass on soluble carbohydrates concentration expressed in glucose equivalents. SSB means untreated sweet sorghum biomass and TT thermal treated sweet sorghum, respectively

The methane yields during alkaline and thermal-alkaline pretreatment are presented in Fig. 4. It is obvious from Fig. 4a that 0.1–1 % w/v NaOH at 25 °C did not affect the methane yield, while at higher alkali concentrations the methane yield decreased. The decrease was higher with a significant increase of NaOH concentration to 1.5 and 2 %. At these NaOH loads the concentration of sodium cations is high (8.6 and 11.5 g Na?/L, respectively). It is well known that cations of sodium at high concentrations (above 8 g/L) are inhibitory to the anaerobic digestion process [48]. In this respect, the decrease in methane yields at these high NaOH loads could be attributed to a partial inhibition or toxicity of methanogens by high sodium ion concentrations. Similar results were obtained by Antonopoulou et al. [38] who applied alkaline treatment (NaOH 2 % w/v, 1 h at room temperature) on rapeseed and sunflower straws and rapeseed and sunflower meals. They found that for all substrates, alkaline treatment did not improve methane yields, which was attributed to the high concentrations of sodium that caused an inhibition to methanogens.

123

Waste Biomass Valor

a

o

NaOH at 25 C

300

200

100

0 SSB

0.5 %

1%

b

1.5 %

2%

o

NaOH at 121 C

300

200

100

0 SSB

TT

0.1 %

0.5 %

1%

1.5 %

2%

Fig. 4 The effect of alkaline pretreatment (NaOH at 25 °C) and thermal alkaline prtreatment (NaOH at 121 °C) of sweet sorghum biomass on methane yield. SSB means untreated sweet sorghum biomass and TT thermal treated sweet sorghum, respectively

To overcome a possible inhibition, the use of another alkali (other than sodium, e.g. potassium) could be a solution. However, Yang et al. [49] found that among the three kinds of alkali (NaOH, KOH, lime) tested for the pretreatment of rice straw, NaOH was the most effective for lignin removal and biogas production. During thermal alkaline pretreatment, the disruption of the lignin structure may result to the formation of soluble compounds with toxic effect on methanogenic microorganisms. Gossett et al. [50] found that alkaline heattreated lignin in concentrations above 1 g/L, had a major inhibitory effect on the methanogens. This, in combination with sugars degradation and high sodium concentrations, caused a decrease in the methane yield, especially at higher concentrations of NaOH (0.5–2 % w/v). Effect of Enzymatic Pretreatment The effect of enzymatic pretreatment on the saccharification of sorghum biomass, for different enzyme concentrations (20, 40, 60, 80, 100 and 150 FPU of Celluclast 1.5L/gTS of

123



400

0.1 %

sorghum biomass, respectively) when Celluclast 1.5L or Celluclast and Novozyme at a ratio of 3:1 were used, is shown in Fig. 5a, b, respectively. With the addition of Celluclast 1.5L, the sugars’ concentration increased to the same extent and was about 14.5 % for all Celluclast concentrations. This means that the lowest concentration of 20gFPU/gTS is prefereable due to the high cost of the enzymes. For the mixture of enzymes, the soluble carbohydrates’ concentration increased with the concentration of Celluclast. The increase was higher (25 %) when the Celluclast concentration was 40 FPU/g TS of sorghum biomass. Higher enzymes concentrations did not cause further increase in soluble carbohydrates concentration. Sugars concentration might have been higher if a sterilization step had been applied, since free sugars released during enzymatic hydrolysis were probably consumed by indigenous sweet sorghum microorganisms. This is in agreement with Quemeneur et al. [37] who studied the

400

a

300

200

100

0

SSB

20

40

60

80

100

150

Celluclast 1.5L (FPU/g TS of SSB) 400


Methane yield (LCH4/kg SSB)

400

b 300

200

100

0

SSB

20

40

60

80

100

150

Celluclast 1.5L (FPU/g TS of SSB) Fig. 5 The effect of enzymatic pretreatmant with Celluclast 1.5L at different concentrations (a) and Celluclast and Novozyme at a ratio of 3:1 (b) on carbohydrates concentration of sorghum biomass. SSB means untreated sweet sorghum biomass

Waste Biomass Valor

up, it is necessary to take into account both economic (enzymes and energy costs) and technical (methane yield) aspects.

400


40 FPU/gTS of SSB

300

Conclusions 200

100

0

SSB

CL

CN

Fig. 6 The effect of enzymatic pretreatment with Celluclast 1.5L (CL) and Celluclast and Novozyme at a ratio of 3:1 (CN) on methane yield of sorghum biomass. SSB means untreated sweet sorghum biomass

enzymatic hydrolysis of wheat straw and found that sugars were degraded by indigenous bacteria. In order to investigate the effect of enzymatic pretreatment on methane yield, the concentration of 40 FPU of Celluclast 1.5L or Celluclast 1.5L and Novozyme (3:1) per gTS of sweet sorghum biomass were tested (Fig. 6). The experiments showed that the methane yield increased with the addition of enzymes. Specifically, the increase of carbohydrates solubilisation led to a tantamount increase of the respective methane yields. The methane yield increased by 17 % (295 L/kg sorghum), when only Celluclast was used and the respective increase was approximately 20 % (303 L/kg sorghum biomass) when the mixture of Celluclast and Novozyme was used. In this study commercial enzymes, without any combination with other pretreatment schemes were used, as a unique pretreatment step prior to anaerobic digestion. In the literature, it has been reported that enzymatic hydrolysis is accomplished after a thermochemical pretreatment step, followed by microbial fermentations such as a fermentative hydrogen or a bioethanol production process [51, 52]. The experimental results in this study showed that enzymatic pretreatment was the sole pretreatment scheme which led to enhanced methane yields. This was due to the fact that during the enzymatic depolymerization of polysaccharides, particularly cellulose, the procedure is accomplished at mild conditions, where no toxic-inhibitory compounds for methanogens are released. In addition, mild conditions, which are required during enzymatic treatment (room temperature without the use of any chemical agents), require less energy than thermochemical processes [53]. On the other hand, the high cost of commercial enzymes contribute significantly to the total process cost and in order to decide for process scale

In this study, various pretreatment methods, such as thermal, enzymatic through the addition of commercial enzymes, chemical through alkali (NaOH) or acid (H2SO4) at concentrations of 0–2 % w/v addition, or combination of the above methods (thermal acid and thermal alkaline) were applied the stalk of sweet sorghum, in order to evaluate the effect of each method on carbohydrate solubilization (saccharification) and on the methane yield. Only thermal acid and enzymatic treatment improved saccharification, while under thermal alkaline treatment at high NaOH concentrations, a high portion of sugars also contained in sorghum biomass was degraded or was transformed into other components. BMP experiments showed that the chemical pretreatment methods did not enhance methane generation compared to the raw substrates, which could be attributed either to toxic for methanogens compounds released during pretreatment or (in the case of alkali treatment) to high sodium concentration, causing in both cases methanogenic bacteria inhibition. Only during enzymatic and thermal pretreatment, the methane production was enhanced. Acknowledgments The authors wish to thank the Greek General Secretariat for Research and Technology for the financial support of this work under ‘‘Supporting Postdoctoral Researchers Projects’’— Pretreatment of lignocellulosic wastes for 2nd generation biofuels [POSTDOC_PE8(1756)].

References 1. Dercas, N., Liakatas, A.: Sorghum water loss in relation to irrigation treatment. Water Resour. Manag. 13, 39–57 (1999) 2. Almodares, A., Hadi, M.R.: Production of bioethanol from sweet sorghum: a review. Afr. J. Agric. Res. 4, 772–780 (2009) 3. Whitfield, M.B., Chinn, M.S., Veal, M.W.: Processing of materials derived from sweet sorghum for biobased products: review. Ind. Crops Prod. 37, 362–375 (2012) 4. Billa, E., Koullas, D.P., Monties, B., Koukios, E.G.: Structure and composition of sweet sorghum stalk components. Ind. Crops Prod. 6, 297–302 (1997) 5. Antonopoulou, G., Gavala, H.N., Skiadas, I.V., Angelopoulos, K., Lyberatos, G.: Biofuels generation from sweet sorghum: fermentative hydrogen production and anaerobic digestion of the remaining biomass. Bioresour. Technol. 99, 110–119 (2008) 6. Miller, A.N., Ottman, M.J.: Irrigation frequency effects on growth and ethanol yield in sweet sorghum. Agron. J. 102, 60–70 (2010) 7. Mamma, D., Koullas, D., Fountoukidis, G., Kekos, D., Makris, B.J., Koukios, E.: Bioethanol from sweet sorghum: simultaneous

123

Waste Biomass Valor

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21. 22.

23.

24.

25.

saccharification and fermentation of carbohydrates by a mixed microbial culture. Process Biochem. 31, 377–381 (1996) Ntaikou, I., Gavala, H.N., Kornaros, M., Lyberatos, G.: Hydrogen production from sugars and sweet sorghum biomass using Ruminococcus albus. Int. J. Hydrogen Energy 33, 1153–1163 (2008) Antonopoulou, G., Gavala, H.N., Skiadas, I.V., Lyberatos, G.: Influence of pH in fermentative hydrogen production from sweet sorghum extract. Int. J. Hydrogen Energy 35, 1921–1928 (2010) Antonopoulou, G., Gavala, H.N., Skiadas, I.V., Lyberatos, G.: Effect of substrate concentration on fermentative hydrogen production from sweet sorghum extract. Int. J. Hydrogen Energy 36(8), 4843–4851 (2011) Antonopoulou, G., Gavala, H.N., Skiadas, I.V., Lyberatos, G.: ADM1-based modeling of methane production from acidified sweet sorghum extract in a two stage process. Bioresour. Technol. 106, 10–19 (2012) Panagiotopoulos, I.A., Bakker, R.R., Vrije, T., Koukios, E.G., Claassen, P.A.M.: Pretreatment of sweet sorghum bagasse for hydrogen production by Caldicellulosiruptor saccharolyticus. Int. J. Hydrogen Energy 35, 7738–7747 (2010) Ivanova, G., Rakhely, G., Kovacs, K.L.: Thermophilic biohydrogen production from energy plants by Caldicellulosiruptor saccharolyticus and comparison with related studies. Int. J. Hydrogen Energy 34, 3659–3670 (2009) Dogaris, I., Karapati, S., Mamma, D., Kalogeris, E., Kekos, D.: Hydrothermal processing and enzymatic hydrolysis of sorghum bagasse for fermentable carbohydrates production. Bioresour. Technol. 100, 6543–6549 (2009) Sipos, B., Reczey, J., Somorai, Z., Kadar, Z., Dienes, D., Reczey, K.: Sweet sorghum as feedstock for ethanol production: enzymatic hydrolysis of steam-pretreated bagasse. Appl. Biochem. Biotechnol. 153, 151–162 (2009) Corredor, D., Salazar, J., Hohn, K., Bean, S., Bean, B., Wang, D.: Evaluation and characterization of forage sorghum as feedstock for fermentable sugar production. Appl. Biochem. Biotechnol. 158, 164–179 (2009) Yu, J., Zhong, J., Zhang, X., Tan, T.: Ethanol production from H2SO3-steam pretreated fresh sweet sorghum stem by simultaneous saccharification and fermentation. Appl. Biochem. Biotechnol. 160, 401–409 (2010) Xu, F., Shi, Y.-C., Wu, X., Theerarattananoon, K., Staggenborg, S., Wang, D.: Sulfuric acid pretreatment and enzymatic hydrolysis of photoperiod sensitive sorghum for ethanol production. Bioprocess Biosyst. Eng. 34, 485–492 (2011) Herrera, A.: Effect of the hydrochloric acid concentration on the hydrolysis of sorghum straw at atmospheric pressure. J. Food Eng. 63, 103–109 (2004) Vazquez, M., Oliva, M., Tellez-Luis, S.J., Rammrez, J.A.: Hydrolysis of sorghum straw using phosphoric acid: evaluation of furfural production. Bioresour. Technol. 98, 3053–3060 (2007) Ban, J., Yu, J., Zhang, X., Tan, T.: Ethanol production from sweet sorghum residual. Front. Chem. Eng. China 2, 452–455 (2008) Goshadrou, A., Karimi, K., Taherzadeh, M.J.: Bioethanol production from sweet sorghum bagasse by Mucor hiemalis. Ind. Crops Prod. 34, 1219–1225 (2011) McIntosh, S., Vancov, T.: Enhanced enzyme saccharification of Sorghum bicolor straw using dilute alkali pretreatment. Bioresour. Technol. 101, 6718–6727 (2010) Wu, L., Arakane, M., Ike, M., Wada, M., Takai, T., Gau, M., Tokuyasu, K.: Low temperature alkali pretreatment for improving enzymatic digestibility of sweet sorghum bagasse for ethanol production. Bioresour. Technol. 102, 4793–4799 (2011) Sathesh-Prabu, C., Murugesan, A.G.: Potential utilization of sorghum field waste for fuel ethanol production employing Pachysolen tannophilus and Saccharomyces cerevisiae. Bioresour. Technol. 102, 2788–2792 (2011)

123

26. Zhang, J., Ma, X., Yu, J., Zhang, X., Tan, T.: The effects of four different pretreatments on enzymatic hydrolysis of sweet sorghum bagasse. Bioresour. Technol. 102, 4585–4589 (2011) 27. APHA, AWWA, WPCF: In: Clesceri, L.S., Greenberg, A.E., Trussell, R.R. (eds.) Standard methods for the examination of water and wastewater, 7th edn. American Public Health Association, Washington, DC (1995) 28. Joseffson, B.: Rapid spectrophotometric determination of total carbohydrates. In: Grasshoff, K., Kremling, K., Ehrhardt, M. (eds.) Methods of seawater analysis, pp. 340–342. Wiley-VCH Verlag Chemie GmbH (1983) 29. Panagiotou, G., Olsson, L.: Effect of compounds released during pretreatment of wheat straw on microbial growth and enzymatic hydrolysis rates. Biotechnol. Bioeng. 96(2), 250–258 (2007) 30. Ghose, T.K.: Measurement of cellulase activities. Pure Appl. Chem. 59(2), 257–268 (1987) 31. Owens, J.M., Chynoweth, D.P.: Biochemical methane potential of municipal solid waste (MSW) components. Wat. Sci. Technol. 27, 1–14 (1993) 32. Skiadas, I.V., Lyberatos, G.: The periodic anaerobic baffled reactor. Water Sci. Technol. 38(8–9), 401–408 (1998) 33. Cui, M., Yuan, Z., Zhi, X., Shen, J.: Optimization of biohydrogen production from beer lees using anaerobic mixed bacteria. Int. J. Hydrogen Energy 34, 7971–7978 (2009) 34. Thomsen, A.B., Thygesen, A., Bohn, V., Nielsen, K.V., Pallesen, B., Jorgensen, M.S.: Effects of chemical–physical pre-treatment processes on hemp fibres for reinforcement of composites and for textiles. Ind. Crops Prod. 24(2), 113–118 (2006) 35. Ramos, L.P.: The chemistry involved in the steam treatment of lignocellulosic materials. Quim. Nova 26(6), 863–871 (2003) 36. Fengel, D., Wegener, G.: Wood: Chemistry, Ultrastructure, Reactions. De Gruyter, Berlin (1984) 37. Quemeneur, M., Bittel, M., Trably, E., Dumas, C., Fourage, L., Ravot, G., Steyer, J.-P., Carrere, H.: Effect of enzyme addition on fermentative hydrogen production from wheat straw. Int. J. Hydrogen Energy 20, 10639–10647 (2012) 38. Antonopoulou, G., Stamatelatou, K., Lyberatos, G.: Exploitation of rapeseed and sunflower residues for methane generation through anaerobic digestion: the effect of pretreatment. Chem. Eng. Trans. 20, 253–258 (2010) 39. Fernandes, T.V., Klaasse Bos, G.J., Zeeman, G., Sanders, J.P.M., van Lier, J.B.: Effects of thermo-chemical pre-treatment on anaerobic biodegradability and hydrolysis of lignocellulosic biomass. Bioresour. Technol. 100(9), 2575–2579 (2009) 40. Nieves, D.C., Karimi, K., Horva´th, I.S.: Improvement of biogas production from oil palm empty fruit bunches (OPEFB). Ind. Crops Prod. 34(1), 1097–1101 (2011) 41. Liu, C., Wyman, C.E.: The effect of flow rate of compressed hot water on xylan, lignin and total mass removal from corn stover. Ind. Eng. Chem. Res. 42, 5409–5416 (2003) 42. Zhu, Y., Lee, Y.Y., Elander, R.T.: Optimization of dilute-acid pretreatment of corn stover using a high-solids percolation reactor. Appl. Biochem. Biotechnol. A Enzyme Eng. Biotechnol. 124, 1045–1054 (2005) 43. Sun, Y., Cheng, J.: Hydrolysis of lignocellulosic materials for ethanol production: a review. Bioresour. Technol. 83, 1–11 (2002) 44. Fan, L.T., Gharpuray, M.M., Lee, Y.-H.: Cellulose Hydrolysis Biotechnology Monographs, p. 57. Springer, Berlin (1987) 45. Fox, M.H., Noike, T., Ohki, T.: Alkaline subcritical-water treatment and alkaline heat treatment for the increase in biodegradability of newsprint waste. Water Sci. Technol. 48, 77–84 (2003) 46. Qingming, L., Xiujin, L., Baoning, Z., Dongyan, Y., Laiqing, L.: Anaerobic biogasification of NaOH-treated corn stalk. Trans. CSAE 21, 111–115 (2005) 47. Mirahmadi, K., Kabir, M.M., Jeihanipour, A., Karimi, K., Taherzadeh, M.J.: Alkaline pretreatment of spruce and birch to

Waste Biomass Valor improve bioethanol and biogas production. BioResources 5(2), 928–938 (2010) 48. McCarty P.L.: Anaerobic Waste Treatment Fundamentals. Part III, Toxic Materials and Their Control, pp. 91–94. Public Works 95 (1964) 49. Yang, D., Zheng, Y., Zhang, R.: Alkali pretreatment of rice straw for increasing the biodegradability. In: An ASABE International Meeting, Reno, Nevada, USA, 21–24 June, American Society of Agricultural Engineers. Paper no.: 095685 (2009) 50. Gossett, J.M., Stuckey, D.C., Owen, W.F., Mccarty, P.L.: Heat treatment and anaerobic digestion of refuse. J. Environ. Eng. Div. 108, 437–454 (1982)

51. Pan, C.-M., Ma, H.-C., Fan, Y.-T., Hou, H.-W.: Bioaugmented cellulosic hydrogen production from cornstalk by integrating dilute acid-enzyme hydrolysis and dark fermentation. Int. J. Hydrogen Energy 36(8), 4852–4862 (2011) 52. Chen, C., Boldor, D., Aita, G., Walker, M.: Ethanol production from sorghum by a microwave-assisted dilute ammonia pretreatment. Bioresour. Technol. 110, 190–197 (2012) 53. Ballesteros, M.: Enzymatic hydrolysis of lignocellulosic biomass. In: Waldron, K. (ed.) Bioalcohol Production: Biochemical Conversion of Lignocellulosic Biomass, Woodhead Publishing Series in Energy, pp. 159–177. CRC Press, Woodhead Pub., Boca Raton (2010)

123

Volume 2 No.9, September 2012

ISSN 2224-3577

International Journal of Science and Technology

©2012 IJST. All rights reserved

http://www.ejournalofsciences.org

Enhancing the Production of Reducing Sugars from Cassava Peels by Pretreatment Methods Olanbiwoninu.A.A, Odunfa S.A Department of Microbiology, University of Ibadan, Ibadan, Oyo State, Nigeria

ABSTRACT Cassava peel is one of the major biomass wastes in Nigeria obtained from production of cassava tuber for human consumption, starch production and industrial uses. The objective of this work was to investigate the optimal condition for pretreating cassava peel with dilute sulphuric acid, methanol with catalyst (organosolv) and alkali prior to microbial enzymatic hydrolysis for the production of fermentable sugars. The pretreated samples reducing sugar yield was measured after enzymatic hydrolysis. The result shows that acid hydrolysis using sulphuric acid at a concentration of 0.1M at 120oC for 30 min gave a maximum reducing sugar yield of 88.8% and 98%, followed closely by methanol treated peels (78 and 98% ) while alkali pretreated peels produce the least (66 and 88%) for Pseudomonas flourescens and Aspergillus terreus respectively. In this study, H2SO4 and methanolysis treated peels prior to enzymatic hydrolysis had a greater capacity for hydrolyzing cassava peels than NaOH and also combination of pretreatments method with enzymatic treatment is an alternative to improve efficiency of reducing sugar production from cassava peel. Keywords: Acid hydrolysis, Alkali hydrolysis, organosolv, Cassava peels, enzymatic hydrolysis, Pretreatment.

1. INTRODUCTION Modernisation has open up an access to efficient utilisation of agro-industrial by-products, aiming at the obtainment of value added products like biofuels, biochemicals and biomaterials.Bioprocessing of agroindustrial residues can help solve environmental problems associated with the disposal of these materials. Cassava (Manihot esculenta Crantz) is a perennial woody shrub, grown as an annual mainly for its starchy roots. It is a cheap source of carbohydrates for human populations in the humid tropics [1, 2, 3, and 4]. The largest producer of cassava world-wide is Nigeria, followed by Brazil, Thailand, Zaire, and Indonesia [5, 6]. It is the staple food for over 500 million people in western and central Africa [7, 8, and 9] with an average consumption of approximately 500 cal/day [10]. In the processing of cassava, the roots are normally peeled to rid them of two outer coverings, i.e. a thin brown outer covering and a thicker leathery parenchymatous inner covering. The peels constitute about 20-35% of the weight of the tuber, especially in the case of hand peeling [11]. Consequently, a large amount of cassava peel waste is generated annually. In Nigeria, cassava peels produced were about 450,000 tons annually with an increasing trend [6]. These peels are regarded as waste and are usually discarded and allowed to rot. Vegetation and soil around the heaps of the peels are rendered unproductive and devastated due to biological and chemical reactions that take place between the continuously fermenting peels, soil and the surrounding vegetation. Besides their pollution and hazardous aspects, in many cases, these peels might have potentials for recycling raw materials or for conversion into useful products of higher value or even as

raw material for other industries or for their use as food or feed after biological treatment [12]. Cassava peel is a lignocellulolytic material and because of this it is of interest to be used as an alternative substrate for ethanol production. To achieve this maximally ,cassava peel has to be pretreated so as to remove lignin and hemicellulose, reduce cellulose crystallinity, and increase the porosity of the lignocellulosic materials and also to enhance sugar production by reducing the formation of byproducts that have inhibitory to the enzymatic hydrolysis and reducing the possibilities of loss of carbohydrates [13]. A mature cassava root possesses three distinct regions: a central vascular core, the cortex (flesh), and the phelloderm (peels). The peel is 1-4 mm thick and may account for 10-12% of the total dry matter of the root [14]. Cassava peels have been evaluated as a feedstuff for animals [15, 16 and 17]. The aim of the study was to obtain soluble reducing sugars by using different pretreatment methods prior to microbial enzymatic hydrolysis of cassava peels and to determine the optimal conditions of each treatment and measure the reducing sugar produced.

2. MATERIALS AND METHODS A. Substrate Preparation Cassava peels from the factory of cassava processing site in Odogbo barracks, Ojo, Ibadan, Oyo State, Nigeria were collected and washed thoroughly in tap water. It was then air dry for 24hours after which it was milled by a blender machine and dried overnight at 55oC in a hot-air oven. The moisture content was found to be 11.2%.

650


ISSN 2224-3577




B. Cassava Peel Pretreatments

C. Microorganism

i.

Alkali pretreatment by NaOH: Ten grams of cassava peel was suspended in 90 ml of 0.01M to 0.25M sodium hydroxide (NaOH) and placed in an autoclave for fifteen minutes at 1210 C. The solid residues were collected and washed extensively with tap water until neutral pH was reached prior to enzymatic hydrolysis.

Aspergillus terreus SC1 and Pseudomonas fluorescens B9 used in this research work were obtained from rotten cassava peels. The fungi was maintained on PDA agar slant and kept at 4 oC for further use while the bacteria was maintained on nutrient agar slant and kept the same way.

ii.

Acid pretreatment by sulfuric acid (H2S04): Ten grams of cassava peel was suspended in 90 ml 0.01M to 0.25M sulphuric acid (H2SO4) and heated at 121 ºC for 15 min in a 500 ml beaker. The solid residues were collected and neutralize with 2M NaOH and then washed extensively with tap water until neutral pH was reached prior to enzymatic hydrolysis.

D. Enzymatic Hydrolysis

iii.

Organosolv pretreatment: Ten grams of cassava peel was suspended in 100mls of methanol with varying concentration of Sodium Acetate (0.01M to 0.25M).Both hot and cold treatment was applied, after which the solid residue were collected and washed extensively with tap water prior to enzymatic hydrolysis.

Pretreated cassava peels (1.5%w/v) were placed in a jar containing 100mls of minimal basal medium (1g/l CaCl 2.7H2O, 1g/l MgSO4, 2g/l (NH4)2SO4, and 0.5g/l KH2SO4), the medium was sterilized and inoculated with 3.0x108 cfu of Pseudomonas flourescens B9 and 9mm plug of Aspergillus terreus SC1 respectively. The mixture was then incubated at 50 ºC for 24 h for bacteria and 72hrs for fungi. After the appropriate timing, the reducing sugar content of the hydrolysates was measured quantitatively using the DNS method [18] The concentration that gave a maximum reducing sugar was chosen. The selected concentration was applied to determine the optimum temperature for cassava peel hydrolysis. A temperature range used in this study was between 115 0C and 1300C and hydrolysis time was from 15 to 60 min.

% Peel Hydrolysis= Reducing Sugars produced by growth – Reducing sugar in control x 100 Reducing sugar in control

3. RESULTS AND DISCUSSION Cassava peels pretreated with dilute acid, alkali and methanol with sodium acetate as catalyst (organosolv) were hydrolyzed using microbial crude enzymes under varying conditions and the % reducing sugar yield was determined for both bacteria and fungi used. Figure 1 shows that optimal % reducing sugar yields for 0.01M alkali and organosolv treated peels were 55.5 and 66.6% respectively while 0.1M dilute acid treated peels produce optimal % reducing sugar of 77.7% using Pseudomonas flourescens B9. Figure 2 shows that optimal reducing sugar yield for 0.05M alkali and organosolv treated peels were 66.6 and 94.4% respectively while 0.1M dilute acid treated peels produce optimal % reducing sugar of 88.8% using Aspergillus terreus.

The concentration of acid, alkali and methanol that gave a maximum reducing sugar was chosen to determine the optimum temperature and hydrolysis time. For organosolv, both isolates produce high concentration of reducing sugar (78% and98% for bacteria and fungi respectively) after 45 minutes of exposure at 0.05M concentration. For dilute acid treated peels, the highest concentration of reducing sugar was produced at 120 0 C after 30 minutes of exposure; the same is applicable to NaOH treated cassava peel. From this study, dilute acid pretreated peels was found to be suitable for highest reducing sugar production and this was supported by the work of Teerapart and Palmarola [19, 20].

651


ISSN 2224-3577




652


ISSN 2224-3577




653


ISSN 2224-3577




654


ISSN 2224-3577




655


ISSN 2224-3577




4. CONCLUSION Experimental study to determine the appropriate method for pretreatment of cassava peels was carried out using acid, alkali and organosolv prior to microbial enzymatic hydrolysis. Different pretreatment methods were evaluated for their ability to produce reducing sugars from cassava peels. Hydrolysing cassava peels with dilute sulphuric acid pretreatment gave a higher reducing sugar production than organosolv and alkali pretreatment method. Pretreatment of cassava peels using 0.1M H2SO4 at 120OC for 30minutes prior to enzymatic hydrolysis yielded % reducing sugar of 88% and 98% for Pseudomonas fluorescens and Aspergillus terreus respectively, this is in agreement with the work of Jirasak and Kalaya, 2006 [21] who obtain 68% of reducing sugar using 0.15M dilute sulphuric acid. Organosolv pretreated cassava peels using methanol with 0.05M Na2CO3 as catalyst for 45 minutes prior to enzymatic hydrolysis yielded % reducing sugar of 78 and 98% for bacteria and fungi isolates respectively, this is close to the results obtain from dilute acid, therefore this method of pretreatment could be apply so as reduce the environmental hazard of using acid for pretreatment. While 0.05M NaOH treated peels produced 66 and 88% reducing sugar at 120OC for 30minutes for bacteria and fungi isolate respectively. It can thus be concluded that organosolv pretreatment method can be used as an alternative to acid pretreatment method in combination with Aspergillus terreus SC1 crude enzymes for the bioconversion of cassava peels. Also, using microbial crude enzymes for saccharification of cassava peels produced more soluble sugars effectively than the pretreatment method alone and also makes the process cost effective in comparison to using commercial enzymes.

[4] Onwueme, I.C., 2002. Cassava in Asia and the Pacific. In: Hillocks, R.J., J.M. Thresh and A.C. Bellotti (eds.). Cassava: Biology, Production and Utilization. CABI Publishing Oxon, UK and New York, USA, 55-65p.

REFERENCES

[10] Iglesias, C., J. Mayer, L. Chávez and F. Calle, 1997. Genetic potential and stability of carotene in cassava roots. Euphytica 94: 367-373.

[1] Lin HJ, Xian L, Zhang QJ, Luo XM, Xu QS, Yang Q, Duan CJ, Liu JL, Tang JL, Feng JX. Production of raw cassava starch-degrading enzyme by Penicillium and its use in conversion of raw cassava flour to ethanol. J Ind Microbiol Biotechnol. 2011;38(6):733–742. doi: 10.1007/s10295-010-0910-7.[PubMed][Cross Ref] [2] Henry, G. and C. Hershey, 2002. Cassava in South America and the Caribbean. In: Hillocks, R.J., J.M. Thresh and A.C. Bellotti (eds.). Cassava: Biology, Production and Utilization. CABI Publishing Oxon, UK and New York, USA. 17-40p. [3] Hillocks, R. J., 2002. Cassava in Africa. In: Hillocks, R.J., J.M. Thresh and A.C. Bellotti (eds.). Cassava: Biology, Production and Utilization. CABI Publishing Oxon, UK and New York, USA, 40-54p.

[5] Phillip, T.A., D.P. Taylor, L.O. Sanni, M. Akoroda, C. Ezedinma, R. Okechukwu, A. Dixon, E. Okoro and G. Tarawali, 2005. The Nigerian cassava industry statistical handbook. IITA, Nigeria. International Institute of Tropical Agriculture, Ibadan, Nigeria, 56p. [6] FAO, 2006. Statistical database of the Food and Agricultural Organization of the United Nations. Available at http://faostat.fao.org/ (accessed 14 August 2007; verified 2 August 2008). FAO, Rome [7] FAO, 1996. Food requirements and population growth. Technical Background Document. No.4. Rome. [8] FAO, 2005. Statistical database of the Food and Agricultural Organization of the United Nations. Available at http://faostat.fao.org/ (accessed 14 August 2006; verified 2 April 2007). FAO, Rome. [9] Egesi, C.N, P. Ilona, F.O. Ogbe, M. Akoroda and A. Dixon, 2007a. Genetic variation and genotype x environment interaction for yield and other agronomic traits in cassava in Nigeria. Agronomic Journal 99: 11371142.

[11] Obadina, A.O., O.B. Oyewole, L.O. Sanni and S.S. Abiola. 2006. Fungal enrichment of cassava peels proteins. African J. Biotechnol. 5(3): 302–304. [12] Ajao,A.T, Abdullahi H.J, Atere,T.G, Kolawole, O.M, 2009: Studies on the biodegradation and utilisation of selected tuber wastes by Penicillium expansum,Bioresource Research Communication, Vol.21, No 5. [13] Godliving Y. S. Mtui Recent advances in pretreatment of lignocellulosic wastes and production of value added products African Journal of Biotechnology Vol. 8 (8), pp. 1398-1415, 20 April, 2009

656


ISSN 2224-3577




[14] Nartey,F. 1979. Studies on cassava cynogenesis, and biosynthesis of cynogenic glucoside in cassava (Manihot spp.), In: Nestel, B., and R. Maclntyre. (Eds.), Chronic cassava toxicity, IDRC 010e: 73-87. [15] Adegbola, A.A.,and V.O. Asaolu. 1986. Preparation of cassava peels for use in small ruminant production in Western Nigeria. In: Toward Optimum Feeding of Agricultural By-products to Livestock in Africa. Preston, T.R., Nuwanyakpa, M.Y.(Eds.), pp. 109-115.Proc. of Workshop held at the University of Alexandria, Egypt. Oct. 1985. ILCA, Addis Ababa, Ethiopia. [16] Obioha, F.C. and P, C.N. Anikwe. 1982. Utilization of ensiled and sun-dried cassava peels by growing swine. Nutr. Rep. Int. 26: 961-972. [17] Osei, S.A., M. Asiamah and C.C. Atuahene. 1990 Effects of fermented cassava peel meal on the performance of layers. Animal Feed Sci. Technol. 29(3– 4): 295–301.

[18] Miller L.G.,, “Use of Dinitrosalisalic acid reagent for determination of reducing sugar,” Analytical Chemistry, vol 31, Mar. 1959. [19] Teerapart Srinorakutara, Lerdluk Kaewvimol and La-aied Saengow, 2005: Approach of Cassava waste pretreatments for fuel ethanol production in Thailand. [20] Palmarola B, Choteborska P, Galbe M and Zacchi G, 2005: Ethanol production from non starch carbohydrates of wheat bran. Bioresource Technology, vol 96, pp 843850. [21] Jirasak Kongkiattikajorn* and Kalaya Yoonan 2006. Conversion of Cassava Industry Waste to Fermentable SugarsThe 2nd Joint International Conference on “Sustainable Energy and Environment (SEE 2006)” E031 (P) 21-23 November 2006, Bangkok, Thailand.

657




Comparison of some new pretreatment methods for second generation bioethanol production from wheat straw and water hyacinth Yadhu Nath Guragain a, Joelle De Coninck b, Florence Husson b, Alain Durand c, Sudip Kumar Rakshit a,⇑ a

Asian Institute of Technology, 58 Moo 9, Km. 42, Paholyothin Highway, Klong Luang, Pathumthani 12120, Thailand GPMA Laboratory, AGROSUP Esplanade Erasme, 21000 Dijon, France c Welience – Platform for Development in Biotechnology 17, rue Sully BP 86510, 21065 Dijon Cedex, France b

a r t i c l e

i n f o

Article history: Received 7 August 2010 Received in revised form 26 November 2010 Accepted 30 November 2010 Available online 6 December 2010 Keywords: Bioethanol Pretreatment Hydrolysis Crude glycerol Ionic liquid

a b s t r a c t Pretreatment of lignocellulosic residues like water hyacinth (WH) and wheat straw (WS) using crude glycerol (CG) and ionic liquids (IL) pretreatment was evaluated and compared with conventional dilute acid pretreatment (DAT) in terms of enzymatic hydrolysis yield and fermentation yield of pretreated samples. In the case of WS, 1-butyl-3-methylimidazolium acetate pretreatment was found to be the best method. The hydrolysis yields of glucose and total reducing sugars were 2.1 and 3.3 times respectively higher by IL pretreatment than DAT, while it was 1.4 and 1.9 times respectively higher with CG pretreatment. For WH sample, CG pretreatment was as effective as DAT and more effective than IL pretreatment regarding hydrolysis yield. The fermentation inhibition was not noticeable with both types of pretreatment methods and feedstocks. Besides, CG pretreatment was found as effective as pure glycerol pretreatment for both feedstocks. This opens up an attractive economic route for the utilization of CG. Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction Bioethanol is considered as a sustainable alternative to gasoline to mitigate the global energy problem due to depletion of fossil fuel and also to reduce greenhouse gas emissions. The food and fuel conflict due to the production of first generation bioethanol from sugar and starchy food materials is an important issue from the food security point of view. In order to avoid the competition with food, use of abundantly available and non-edible parts of plants, including agricultural wastes and fast growing aquatic plants, as feedstock is being attempted (Hu et al., 2008). However, the bioethanol production from lignocellulosic residues is still not economically viable using existing technologies in the context of current petroleum price (Hendriks and Zeeman, 2009; Hu et al., 2008; Mosier et al., 2005; Laus et al., 2005; Karunanithy et al., 2008). There are four major steps for ethanol production from the lignocellulosic biomass. First step is pretreatment of biomass to break down lignin–hemicelluloses–cellulose complex to make it more susceptible for hydrolysis. The second step is hydrolysis to break down the cellulose and hemicelluloses into monomer sugars. The third step is fermentation of these sugars to ethanol. The final step is product recovery and concentration by distillation (Hu et al., 2008). Each

⇑ Corresponding author. Tel.: +66 2 524 5089; fax: +66 2 524 5003. E-mail address: [email protected] (S.K. Rakshit). 0960-8524/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2010.11.125

step of its production process has lots of challenges that have to be overcome. Among them, pretreatment is a major bottleneck because it is the highest cost for the bioethanol industry, excluding capital cost of the plant (Sousa et al., 2009; Shi et al., 2009; Tomas-Pejo et al., 2008). There are number of biological, physical and chemical technologies available for the pretreatment of biomass, such as use of enzyme, ball milling, steam explosion, acid, alkali, lime and wet oxidation. Many new technologies are also at the different stages of development throughout the world (Hu et al., 2008). Some pretreatment methods including acid pretreatment, wet oxidation and lime pretreatment seem to be economically more feasible than others. However, environmental concern and production of inhibitors to fermenting yeasts during these pretreatment processes are the major hurdles that has to be overcome for commercial production of the bioethanol (Hendriks and Zeeman, 2009; Hu et al., 2008; Mosier et al., 2005). Very slow action of the biological based pretreatment processes (Hu et al., 2008) and high cost of ammonia fiber explosion process, hot water pretreatment and steam explosion process make them economically infeasible (Mosier et al., 2005; Karunanithy et al., 2008). Therefore, the development of efficient, cost effective and environmentally friendly pretreatment method is important (Li et al., 2009). Some new pretreatment technologies have attracted much attention in last few years. Crude glycerol pretreatment and the dissolution of biomass in ionic liquids are some of these emerging pretreatment technologies. Crude glycerol is a major by-product of

Y.N. Guragain et al. / Bioresource Technology 102 (2011) 4416–4424

biodiesel industry. It is neither suitable for fuel components nor its purification is technically and economically feasible to use it in food, pharmaceutical and cosmetics products. The rapid growth of global biodiesel production (Bournay et al., 2005) indicates that the crude glycerol of biodiesel industry will be a costly waste rather than a valuable byproduct in the future. This also could be a serious environmental problem (Sun and Chen, 2008). Therefore, the exploitation of crude glycerol for pretreatment of biomass could be an attractive economic route for the utilization of the byproduct of biodiesel industry, leading to commercial success of both bioethanol and biodiesel. Ionic liquids are salts that remain in liquid state at room temperature. They are stable up to 300 °C and their extreme non-volatility is also expected to have minimum environmental impact (Mousdale, 2008). The microcrystalline cellulose can be readily solubilized in ionic liquids, which was recovered later by addition of the anti-solvents. The crystallinity of the regenerated cellulose is much less than untreated cellulose, which makes them more susceptible to enzymatic hydrolysis (Zhao et al., 2009). The ionic liquids are considered as the potential solvents for the pretreatment of lignocellulosic materials because of their unique solvent properties, fluidity at room temperatures and negligible environmental impact (Dadi et al., 2007; Liu and Chen, 2006). However, most of research works have carried out on chloride-based ionic liquids, which are extremely hygroscopic, corrosive, and also have some toxicity (Zhao et al., 2009). It is reported that the phosphate-based and the acetate-based ionic liquids are less viscous and more thermally stable than the chloride-based (Li et al., 2009; Zhao et al., 2009). Therefore, it is very important to further evaluate the efficiency of these ionic liquids for the pretreatment of abundantly available lignocellulosic biomass like wheat straw and water hyacinth. The cost and availability of feedstock is also equally important to make bioethanol commercially viable. In this study, wheat straw (Triticum aestivum L.) was used as the first substrate as it is one of the major agricultural wastes in the world. The total global production of wheat straw was about 850 billion kg in 2008, out of which approximately 430 billion kg wheat straw is annually available for production of bioethanol (Talebnia et al., 2009; Kim and Dale, 2004). Water hyacinth (Eichhornia crassipes), a prolific aquatic pest, was also used as the second substrate as it is abundantly available in many countries and it is a major problem in the waterways (Gunnarsson and Petersen, 2007). The utilization of water hyacinth as the feedstock for bioethanol production has number of advantages. The commercial cultivation of this plant in controlled condition leads the sustainable solution of some environmental problems created in many countries due to its proliferation in waterways. Being an aquatic plant, it does not compete with food crops for arable lands. Its very high growth rate, 60– 100 t ha 1 yr 1, is also favorable for its commercial cultivation (Mishima et al., 2008). In this study, an attempt was made to determine the best condition of ionic liquids and crude glycerol pretreatment in terms of enzymatic hydrolysis yield of glucose and total reducing sugars, using wheat straw and water hyacinth as feedstocks. The hydrolysis of pretreated samples was carried out using cellulase enzyme from Trichoderma reesei. Then, the effectiveness of these pretreatment methods was compared with dilute acid pretreatment. Enzymatic hydrolysis yield of glucose and total reducing sugars as well as fermentation yield of ethanol was taken as a measure of effectiveness of these pretreatment methods. The fermentation of hydrolysates was carried out using Saccharomyces cerevisiae. The pure glycerol pretreatment was also carried out to determine the effect of impurities present in crude glycerol during enzymatic hydrolysis and fermentation steps of bioethanol production process.

4417

2. Methods 2.1. Materials Fresh water hyacinth (Eichhornia crassipes) samples were collected from canals of Asian Institute of Technology (AIT), Bangkok (Thailand). It was washed, chopped and sun dried to about 55% moisture content. The partially dried sample was ground and dried at 60 °C until the final moisture content reached to less than 10%. The dried sample was then ground and sieved to a size of 750 lm. The grinder used in this experiment had the provision to fix a sieve with the desire pore size just before the outlet of sample such that the sample could be ground till it passes through the sieve. The ground sample was then packed in plastic bags. The dry wheat straw (Triticum aestivum L.) samples were collected from a farm at Dijon (France). It was first ground to smaller pieces using sieve size of 1 cm and then further ground using sieve size of 750 lm. The ground sample was then packed in plastic bags. The crude glycerol was procured from a biodiesel industry in France. Two types of ‘Aldrich’ brand ionic liquids, 1-butyl-3-methylimidazolium acetate (BMIMA) and 1-ethyle-3-methylimidazolium diethyl phosphate (EMIMDP), and the Sigma brand cellulase enzyme from Trichoderma reesei was purchased from the market. The Saccharomyces cerevisiae available at the Bioprocess Technology Laboratory of AIT was preserved by sub-culturing in yeast extract peptone dextrose (YPD) agar slant in every two months and this yeast was used throughout this study.

2.2. Pretreatment of samples For the comparative study of different pretreatment methods, the best pretreatment condition of each method was first selected on the basis of enzymatic hydrolysis yield of glucose and total reducing sugars of the pretreated sample. These pretreatment methods were then compared with dilute acid as well as pure glycerol pretreatment. Previous work on pretreatment of biomass by physiochemical process showed that time and temperature are the major factors that influence the pretreatment process measured by the enzymatic hydrolysis yield of pretreated sample. Therefore, a 22 factorial design, by using the Minitab 2001 software, was chosen to study the effect of these two factors and their combinations during pretreatment of biomass. The minimum and maximum levels for these two factors were first fixed to apply factorial design as explained in Section 3.2.

2.2.1. Crude glycerol pretreatment Ten grams of samples to be pretreated was mixed with 200 ml crude glycerol in a two-necked round bottom flask. A condenser was fitted in one neck of the flask to reflux and thermocouple wire was passed through other neck to measure temperature. Magnetic stirrer was used to stir the mixture. The flask was heated for different combinations of time and temperature ranging from 180 to 230 °C for 1–4 hours. In order to avoid the formation of excessive foam, the sample was heated slowly so that the desire temperature was reached in 30–45 min. After heating for desire time and temperature, the mixture was allowed to cool to about 100 °C. Then, 400 ml dilute crude glycerol (water/crude glycerol = 1/1) was added in it. The mixture was stirred vigorously and filtered through a cotton cloth. The residue was washed with excess of distilled water until the filtrate became clear. Then, the pretreated samples were dried overnight at room temperature (23 °C and 57% RH) until final moisture content became less than 10% and packed in plastic bags. This method is modified version of the method used by Sun and Chen (2008).

4418


The pure glycerol pretreatment was also carried out to evaluate the effect of impurities present in crude glycerol during bioethanol productions processes. The best condition of time and temperature for the crude glycerol pretreatment was used for pure glycerol pretreatment. 2.2.2. Ionic liquid pretreatment In a round bottom flask, 240 mg sample was mixed with 6 g ionic liquids. The flask was heated for different combinations of time and temperature ranging from 100 to 150 °C for 10 to 60 min to dissolve the sample in the ionic liquids. Magnetic stirrer was used to stir the mixture. After heating for desire time and temperature, the mixture was allowed to cool to about 60 °C. Equal volume (6 ml) of distilled water was added in it, as an anti-solvent. This solution was left for about 30 min for precipitation of dissolved cellulose. The mixture was then stirred properly and filtered through a cotton cloth. The residue was washed with excess of distilled water until the filtrate became clear. The pretreated sample was dried overnight at room temperature (23 °C and 57% RH) and then further dried at 60 °C for 48 h. The second drying step was carried out to skip moisture determination of pretreated sample because the samples were not enough for it. It was because the moisture meter used in this experiment required minimum 100 mg sample for the determination of the moisture content and the sample used for moisture determination could not be used for the enzymatic hydrolysis. The dry samples were then packed in air tight plastic bags. This method is modified version of the method used by Li et al. (2009). 2.2.3. Dilute acid pretreatment The effectiveness of these new methods of pretreatment was evaluated by comparing them with the conventional dilute acid pretreatment method. In this method, 15 g sample was mixed with 300 ml 1% (v/v) sulfuric acid (H2SO4) and autoclave at 140 °C for 40 min. The sample was then washed with excess of water until the washed water became acid free and had a pH of 7. In case of water hyacinth sample, dilute sodium hydroxide solution was also used to neutralize the acidity of the mixture. This method is modified version of the method used by Karunanithy et al. (2008). 2.3. Enzymatic hydrolysis 2.3.1. Optimization of cellulase concentration and hydrolysis time The biomass samples were pretreated using crude glycerol pretreatment at maximum possible condition (at 230 °C for 4 h). Then, 100 mg dry pretreated sample and 5 mg acetate buffer (0.05 M) pH 5.0 were taken in 50 ml flasks. The slurries containing 2% solid were maintained to reduce the end-product inhibition for cellulose hydrolysis (Sun and Chen, 2008). In these flasks, four different concentrations of enzyme, 50, 100, 250 and 500 ll (1 ll = 0.84 U cellulase), were added to prepare four set of the flasks, each in triplicate. One flask was prepared for each set without addition of substrate to serve as a blank for the determination of total reducing sugars and glucose content during hydrolysis. These flasks were incubated at 50 °C, shaking at 150 rpm. Approximately 250 ll samples were drawn periodically from each flask and centrifuged at 13,000 rpm for 15 min. The supernatants with suitable dilution were used for the determination of glucose and total reducing sugars content. 2.3.2. Hydrolysis of pretreated samples The hydrolysis of the pretreated samples was carried out after optimization of the enzyme concentration and the hydrolysis time as described in the previous section. These values were found 250 ll enzyme (210 U cellulase) per 100 mg pretreated sample for enzyme concentration and 2.5 h for incubation time. One hundred milligram pretreated samples were taken in 50 ml flask. Five

milliliter acetate buffer (0.05 M) pH 5.0 and 250 ll cellulase enzyme were added in each flask. These flasks were incubated at 50 °C, shaking at 150 rpm. The glucose and total reducing sugars were determined at the beginning and at the end that is after 2.5 h of incubation. The hydrolysis yield of glucose and total reducing sugars was statistically analyzed to determine the best conditions of each methods of pretreatment as well as to compare the different methods of pretreatments, at their selected conditions. For the determination of the best conditions of each methods of pretreatment, the Minitab 2001 software was used, in which a 22 factorial design was chosen at 95% confidence level taking time and temperature as two factors. Similarly, for the comparison of the different methods of pretreatments at their selected conditions, the least significant difference (LSD) test was carried out at 95% confidence level using SPSS software. All the experiments were carried out in triplicate. This method is modified version of the method used by Li et al. (2009) and Sun and Chen (2008). 2.4. Fermentation 2.4.1. Inoculum preparation The culture medium (YPD) was prepared by dissolving 1 g glucose (20 g/L), 1 g peptone (20 g/L) and 0.5 g yeast extract (10 g/L) in 50 ml distilled water in 250 ml conical flask. The flask was autoclaved at 121 °C for 15 min. One loop of cell of Saccharomyces cerevisiae was added into the medium and incubated at 30 °C, shaking at 150 rpm, for 24 h (Li et al., 2009). The prepared inoculum was then added directly to the fermentation medium. 2.4.2. Fermentation of hydrolysate One gram of each pretreated sample was hydrolyzed with 1 ml cellulase enzyme in 25 ml acetate buffer (0.05 M) pH 5.0 in 125 ml flask for 48 h at 50 °C, shaking at 100 rpm. After hydrolysis, the flasks were heated in water bath at 95 °C for 15 min and then centrifuged at 12,000 rpm for 15 min. The supernatant of these hydrolysates was either vacuum concentrated using Rotavapour or diluted in acetate buffer such that the glucose concentration would be 15 ± 2 mg/ml in the final fermentation medium. The final fermentation medium contained 4 ml concentrated/diluted hydrolysate, 1 ml supplementary nutrition solution and 0.5 ml (10%, v/v) inoculums. The supplementary nutrition solution was prepared in acetate buffer in such a way that the final fermentation medium contained 10 g/L yeast extract, 5 g/L NH4Cl, 1 g/L MgSO4 7H2O and 2 g/L KH2PO4. A control medium was also prepared by using glucose solution of similar concentration instead of hydrolysate. The pH of these media was maintained at 6. These media were prepared in 25 ml flasks and sterilized at 121 °C for 15 min. Then, 0.5 ml fresh inoculum was aseptically added in each flask, followed by incubation at 30 °C, shaking at 200 rpm. Certain amount of sample was drawn periodically from each flask to determine ethanol, glucose and total reducing sugars. All the experiments were carried out in triplicate. This method is modified version of the method used by Li et al. (2009) and Larsson et al. (1999). 2.5. Analytical procedures The glucose was determined using commercially available Biomerieux brand (REF 61/270) glucose oxidase–peroxidase–chromogen reagent. The total reducing sugars were determined by a modification of the standard DNS method (Miller, 1959). The lignin content was determined by two step acid hydrolysis method, as described by Sluiter et al. (2008) in National Renewable Energy Laboratory (NREL)/TP-510-42618. The wavelength of 320 nm was used for the determination of acid soluble lignin. The ethanol concentration was measured by using Gas Chromatography.


2.6. Statistical analysis The establishment of the experimental design and the analysis of all the results have been carried out using Minitab and SPSS software. 3. Results and discussions 3.1. Optimization of concentration of enzyme and incubation time Initial experiments were done to select the best condition of each pretreatment method and also to compare the effectiveness of different pretreatment methods. The concentration of enzyme and incubation time was first optimized for the hydrolysis of pretreated sample. Fig. 1 shows that enzymatic hydrolysis yield of the glucose from pretreated wheat straw sample increased linearly with incubation time until 3 h while its rate of increment reduced considerably thereafter. On this basis, the incubation time of 2.5 h was considered the best time period for the comparative study of different pretreatment method on the basis of hydrolysis yield. It is also found that the hydrolysis yield of glucose increased with increase in enzyme concentration from 50 to 250 ll enzyme per 100 mg sample whereas there was no significant increase in glucose yield by increasing enzyme concentration from 250 to 500 ll enzyme per 100 mg sample. Similar curves were found for hydrolysis yield of total reducing sugars from wheat straw, and glucose and total reducing sugars from water hyacinth sample (figures not given here). Therefore, incubation time of 2.5 h and enzyme concentration of 250 ll enzyme per 100 mg sample were selected as the standard condition for the comparative study of hydrolysis yield of different samples. 3.2. Determination of the best pretreatment condition First of all, the minimum and maximum levels for time and temperature of the pretreatment methods were fixed to apply factorial design for the determination of the best pretreatment condition. In this study, we found that crude glycerol pretreatment, taking solvent and sample in the ratio of 200 ml:10 g, could not be carried out above 230 °C because of the formation of excessive foam above this temperature. In the similar experiment, Sun and Chen (2008) found that the optimum condition for 70% aqueous glycerol pretreatment of biomass is 220 °C for 3 h. Therefore, the range of crude glycerol pretreatment conditions for time and temperature chosen in this study was 180–230 °C for 1–4 h. Similarly, we found that ionic liquid pretreatment, taking solvent and

4419

sample in the ratio of 25:1 (w/w), could not be done above 150 °C and more than 1 h. It was because the mixture became very thick and hence the magnetic stirrer could not stir the mixture beyond that condition. In the similar experiment, Li et al. (2009) found that the optimum condition for ionic liquid pretreatment is 130 °C for 30 min. Therefore, the range of ionic liquid pretreatment conditions for time and temperature chosen in this study was 100–150 °C for 10–60 min. In order to test the repeatability of the results at 95% confidence level, the pretreatment of each biomass samples was first carried out at the central point in triplicate. The central points were 205 °C for 2.5 h for crude glycerol pretreatment and 125 °C for 35 min for ionic liquid pretreatment. The data for enzymatic hydrolysis yield of glucose and total reducing sugars were then analyzed. These results were found acceptable repeatability for all methods of pretreatment at central point. Further experiments were then carried out at four extreme points within the chosen range of time and temperature for each method of pretreatment. The glucose and total reducing sugars were measured after conducting enzymatic hydrolysis of these pretreated samples. From the response curves for glucose yield (Fig. 2), it would seem that the optimum pretreatment condition may be beyond the chosen range of experiment in some cases. Similar curves were obtained for hydrolysis yield of total reducing sugars (figures not given here). However, from a practical point of view, the experiments could not be carried out beyond the chosen scale because of the formation of excessive foam in crude glycerol pretreatment and inability to stir the sample by magnetic stirrer due to excessive thickening of mixture in ionic liquid pretreatment. Therefore, it was concluded from these results that the best condition of crude glycerol pretreatment is 230 °C for 4 h for wheat straw and 230 °C for 1 h for water hyacinth. Similarly, 150 °C for 1 h was found as the best condition for ionic liquid pretreatment for both wheat straw and water hyacinth.

3.3. Enzymatic hydrolysis of pretreated samples After selection of the best pretreatment condition, within the chosen range for each pretreatment method, the enzymatic hydrolysis of the samples pretreated under the selected condition were then evaluated and compared with dilute acid as well as pure glycerol pretreatment. It is very important to note that the yield of glucose and total reducing sugars in Figs. 3 and 4 is not total yield but only yield during which the action of cellulase enzyme is linear and hence at fastest rate with hydrolysis time. Based on the data described in Section 3.1, the optimum hydrolysis time was taken to be 2.5 h.

Fig. 1. Production of glucose at different enzyme concentrations during hydrolysis of 100 mg crude glycerol pretreated wheat straw samples in 5 ml buffer. (1 ll = 0.84 U cellulase enzyme from Trichoderma reesei.)

4420


Fig. 2. Hydrolysis yield of glucose (mg/mg pretreated sample) under different conditions and pretreatment methods. (a) BMIMA pretreated wheat straw. (b) EMIMDP pretreated wheat straw. (c) Crude glycerol pretreated wheat straw. (d) BMIMA pretreated water hyacinth. (e) Crude glycerol pretreated water hyacinth. (BMIMA and EMIMDP are ionic liquids. Hydrolysis was carried out for 2.5 h at 50 °C using 250 ll cellulase enzyme for 100 mg pretreated sample in 5 ml buffer.)

Fig. 3. Hydrolysis yield of glucose and total reducing sugars of wheat straw samples pretreated by different pretreatment methods. (Hydrolysis was carried out for 2.5 h at 50 °C using 250 ll cellulase enzyme for 100 mg pretreated sample in 5 ml buffer. EMIMDP and BMIMA are ionic liquids used in this study.)

3.3.1. Wheat straw Fig. 3 shows that use of 1-butyl-3-methylimidazolium acetate (BMIMA) is the most effective method of pretreatment for wheat straw sample to improve enzymatic hydrolysis. This pretreatment

method led to the production of more than double amount of glucose and more than triple amount of total reducing sugars than the conventional dilute acid pretreatment. This is due to the fact that the ionic liquid pretreatment leads to not only break down of


4421

Fig. 4. Hydrolysis yield of glucose and total reducing sugars of water hyacinth samples pretreated by different pretreatment methods. (Hydrolysis was carried out for 2.5 h at 50 °C using 250 ll cellulase enzyme for 100 mg pretreated sample in 5 ml buffer. EMIMDP and BMIMA are ionic liquids used in this study.)

lignin–hemicelluloses–cellulose complex, but also leads to formation of more amorphous structure of cellulose than that of original biomass (Zhao et al., 2009; Dadi et al., 2007; Zhu et al., 2006). However, the yield of total reducing sugars in phosphate-based ionic liquid that is 1-ethyle-3-methylimidazolium diethyl phosphate (EMIMDP) pretreatment was only 27% more than dilute acid pretreatment and glucose yield was even less. This was because of the incomplete dissolution of the sample in EMIMDP, whereas almost complete dissolution of the sample was observed in BMIMA. Similar result was shown by Zavrel et al. (2009). They showed that the acetate-based ionic liquid is better than phosphate-based ionic liquid to dissolve lignocellulosic biomass as well as pure crystalline cellulose. Fig. 3 also showed that yield of glucose and total reducing sugars after crude glycerol pretreatment were 8% and 14%, respectively, less than pure glycerol pretreatment. The hydrolysis yield of crude glycerol pretreatment was found to be better than a similar study by Sun and Chen (2008). In their study, hydrolysis yield of crude glycerol pretreated wheat straw sample was 21% less (37% vs. 47%) than 95% pure glycerol pretreated sample (Sun and Chen, 2008) whereas in our experiment, it was only 14% less in crude glycerol pretreated sample than 99% pure glycerol pretreated sample. Sun and Chen (2008) also showed that 95% glycerol pretreatment removed much higher lignin than crude glycerol pretreatment (64% and 19% lignin removal from original biomass with 95% pure glycerol and crude glycerol, respectively) whereas in our study, crude glycerol was found to be as effective as pure glycerol for the delignification of biomass. This is because the lignin content of the crude glycerol pretreated sample (11.4 ± 1.2%) was not significantly different than pure glycerol pretreated sample (11.2 ± 0.7%). Sun and Chen (2008) also found that there was 28% lipophilic extracts in crude glycerol pretreated sample whereas it was almost nil in 95% glycerol pretreated sample. These lipophilic compounds lead to formation of so-called ‘‘pitch deposition’’ on the pretreated sample, leading to decrease in permeability of the pretreated sample in the aqueous phase (Gutierrez and Rio, 2005). This phenomenon adversely affects in both lignin removal and cellulase action. The crude glycerol used in our study probably had less lipophilic compounds than the crude glycerol used by Sun and Chen (2008), which led to less pitch deposition in pretreated sample, more lignin removal and high enzymatic hydrolysis yield. This is because the quality of the crude glycerol depends on the processing method as well as the catalysts used for the production of biodiesel and determines the amount and type of the impurities present in crude glycerol (Bournay et al., 2005). However, the

determination of the impurities in crude glycerol and their effect on enzymatic hydrolysis was out of the scope of this study. 3.3.2. Water hyacinth The hydrolysis yield of both glucose and total reducing sugars in crude glycerol pretreated water hyacinth samples was not significantly different than pure glycerol as well as dilute acid pretreated samples at 95% confidence level, as indicated in Fig. 4. However, use of the crude glycerol is more advantageous because of its cost, availability and byproduct utilization point of view, when biodiesel production is increased in future. Moreover, requirement of alkali to neutralize the dilute acid pretreated sample and their environmental issue further discourage the use of dilute acid for the pretreatment of biomass. The environmental problem due to use of crude glycerol is less severe as compared to dilute acid. On the one hand, the crude glycerol pretreatment does not produce a new waste like dilute acid pretreatment because crude glycerol itself is a waste of biodiesel industry. On the other hand, crude glycerol could be converted to ethanol and other fermentative products using some microorganism, for example, mutated strain of Enterobacter aerogenes (Ito et al., 2005) and the Klebsiella planticola, isolated from rumen of red deer (Yazdani and Gonzalez, 2007). However, further research work is needed to explore the possibility of co-fermentation of the hydrolysates and glycerol. It was also found that the lignin content in crude glycerol and pure glycerol pretreated samples were not significantly different and were 10.1 ± 0.5% and 9.6 ± 1.1%, respectively. This could be the reason for almost equal enzymatic hydrolysis yield in pure and crude glycerol pretreated samples. The ionic liquid pretreatment of water hyacinth sample was found less effective than dilute acid pretreatment to improve hydrolysis yield for both types of ionic liquids, EMIMDP and BMIMA. This result is completely different from the results obtained with wheat straw sample. One of the reasons for such result could be due to solubilization of almost all hemicelluloses in dilute acid pretreatment (Alvira et al., 2009) whereas there could be significant amount of non-hydrolyzed as well as regenerated and/or undissolved hemicelluloses in ionic liquid pretreated sample. Since equal amount of pretreated biomass was hydrolyzed using cellulase enzyme, presence of higher proportion of hemicelluloses in the biomass reduced the hydrolysis yield of pretreated sample. Water hyacinth contains more hemicelluloses than cellulose (Nigam, 2002), which is contrary to the composition of wheat straw, which contains more cellulose than hemicelluloses (Talebnia et al., 2009). Therefore, the effect of hemicelluloses present in

4422


pretreated sample might be more significant in water hyacinth than in wheat straw. This result showed that there is no universal pretreatment method that could be applied to all lignocellulosic materials. There are number of substrate-related factors that must be taken into consideration for the selection of the best pretreatment method which would lead to improvement of hydrolysis yield without producing fermentation inhibitors. Some of the parameters include polymer composition of biomass, degree of polymerization and crystallinity of cellulose (Zhao et al., 2009), lignin composition, extend of side-chain branching of hemicelluloses as well as ferulate and coumarate cross linking (Sousa et al., 2009). Therefore, each material is to be optimized for suitable pretreatment method (Mishima et al., 2006). 3.4. Fermentability of the hydrolysates The pretreatment methods were finally compared based on their fermentability to ethanol. The pretreated samples were hydrolyzed using cellulase at 50 °C for 48 h. The initial glucose concentration in fermentation media, after addition of supplementary nutrition and inoculum, was maintained 15 ± 2 mg/ml for all experiments. This concentration was achieved either by diluting the hydrolysates with acetate buffer or by vacuum concentrating it using Rotavapour. Glucose solution in acetate buffer, with similar concentration, was taken as a control to evaluate the fermentability of the hydrolysates. 3.4.1. Wheat straw The hydrolysates of the wheat straw samples, pretreated by different methods of pretreatments, were concentrated at reduced pressure, using Rotavapour, to the desire concentration. Assuming only glucose as the fermentable sugar, the maximum ethanol yield in all of these samples was not found significantly different than control at 95% confidence level, as indicated in Fig. 5. However, the yields of ethanol in all samples were significantly lower than control at first 3 h of fermentation. The maximum ethanol yield was reached in 6 h of fermentation, excluding EMIMDP pretreated sample. This yield was achieved only in 12 h of fermentation with the latter sample. The maximum ethanol yield was found when glucose concentration in the fermentation media was almost zero. The data for glucose concentration during fermentation is not shown here. It is also found from Fig. 5 that the yields of ethanol in some of these samples were even slightly more than control. This was due to the utilization of some reducing sugars other than glucose that were present in the hydrolysates.

Fig. 5 also shows that there could be some fermentation inhibitors in the hydrolysates that were responsible for less yield of ethanol at the beginning of the fermentation in all samples as compared to control. However, its effect was not significant when 6 h fermentation was carried out, except in the EMIMDP pretreated sample. This effect was found more severe in EMIMDP pretreated sample because the hydrolysate of this sample had lower concentration of glucose than other samples. The EMIMDP pretreated sample had 7 mg/ml glucose whereas all other samples had 11– 13 mg/ml glucose. This hydrolysate was, therefore, concentrated more than others to acquire the desire glucose level in the fermentation medium. This might have led to an increase in the concentration of inhibitors in the fermentation medium. Due to this, the maximum ethanol yield was achieved only after 12 h of fermentation in EMIMDP pretreated sample whereas it was achieved only in 6 h in other samples. Talebnia et al. (2009) found that the significant amount of fermentation inhibitors is produced in dilute H2SO4 pretreated wheat straw only when pretreatment temperature is reached to 180 °C. In our study, dilute acid pretreatment was carried out only at 140 °C for 40 min, as the reference method of pretreatment. Therefore, it could be assumed that the reference method of pretreatment of our study did not produce significant amount of inhibitors. The similarity in ethanol yield between the reference method and the tested method of pretreatment showed that both BMIMA and crude glycerol pretreatment do not produce fermentation inhibitor for Saccharomyces cerevisiae. Similar result was reported by Li et al. (2009) regarding production of fermentation inhibitor for Saccharomyces cerevisiae during ionic liquid pretreatment of wheat straw. Fig. 5 also found that the yield of ethanol from crude glycerol pretreated and pure glycerol pretreated wheat straw sample were not significantly different. This showed that the impurities present in crude glycerol were not inhibitors for Saccharomyces cerevisiae neither produce any fermentation inhibitors during pretreatment of wheat straw. 3.4.2. Water hyacinth The hydrolysates of the EMIMDP and BMIMA pretreated water hyacinth samples were concentrated, at reduced pressure, using Rotavapour, while hydrolysates of other samples were diluted to the desire concentration. In this study, the maximum ethanol yield was not found significantly different between samples and control, except in the BMIMA pretreated sample, at 95% confidence level, as indicated in Fig. 6. The yield of ethanol in first 3 h of fermentation was significantly lower in ionic liquid pretreated samples than control, while it was not true for dilute acid and crude glycerol

Fig. 5. Yield of ethanol with wheat straw samples pretreated by different methods on fermentation using Saccharomyces cerevisiae. (The fermentation media were prepared after enzymatic hydrolysis of pretreated samples for 48 h at 50 °C; followed by the concentration of hydrolysates to a glucose concentration level of 15 ± 2 mg/ml. Yield was calculated assuming only glucose as the fermentable sugar. EMIMDP and BMIMA are the ionic liquids used in this study. Control is glucose solution in buffer.)


4423

Fig. 6. Yield of ethanol with water hyacinth samples pretreated by different methods on fermentation using Saccharomyces cerevisiae. (The fermentation media were prepared after enzymatic hydrolysis of pretreated samples for 48 h at 50 °C; followed by the concentration or dilution of hydrolysates to a glucose concentration level of 15 ± 2 mg/ml. Yield was calculated assuming only glucose as the fermentable sugar. EMIMDP and BMIMA are the ionic liquids used in this study. Control is glucose solution in buffer.)

pretreated samples. It could be due to the same reason that was explained above for wheat straw that is concentration of hydrolysates to acquire the desire glucose level, which led to higher levels of inhibitors in the fermentation medium. This result also shows that the maximum yield of ethanol in some of these samples were even slightly more than control. Just as in the case of wheat straw, this was because of utilization of some reducing sugars other than glucose that were present in the hydrolysates. It was also found from this result that both crude glycerol and ionic liquid pretreatment of water hyacinth sample did not produce significant amount of fermentation inhibitors. Moreover, crude glycerol pretreatment was found to be better than both types of ionic liquids for fermentation yield, which is similar to the enzymatic hydrolysis yield for water hyacinth sample. Similar to the wheat straw sample, this study also found that the impurities present in crude glycerol neither inhibit Saccharomyces cerevisiae nor produce any fermentation inhibitors during pretreatment of water hyacinth sample. It is because the yield of ethanol in pure and crude glycerol pretreated samples was not significantly different. 4. Conclusion For the pretreatment of wheat straw, BMIMA pretreatment was the best method regarding enzymatic hydrolysis yield. Crude glycerol pretreatment was also better than conventional dilute acid pretreatment. For water hyacinth, crude glycerol pretreatment was as effective as conventional pretreatment and more effective than ionic liquid pretreatment. The fermentation inhibition was not noticeable with both types of pretreatment methods and feedstocks. Moreover, crude glycerol pretreatment was as effective as pure glycerol pretreatment for both feedstocks. This opens up an attractive alternative route for the utilization of crude glycerol, leading to more economic routes for simultaneous production of bioethanol and biodiesel. Acknowledgement This work was possible due to the financial support of Joint Japan World Bank Graduate Scholarship Program to first author, Y.N. Guragain, for his Master study in Food Engineering and Bioprocess Technology in Asian Institute of Technology (AIT), Thailand. The AUF (Agence Universitaire de la Francophonie) project on ethanol also provided the mobility fund to do some parts of the work at Welience

– Platform for Development in Biotechnology, Dijon (France). The authors are deeply grateful to GPMA laboratory AgroSup Dijon (France) for her technical inputs during the work at Dijon. References Alvira, P., Pejo, T., Ballesteros, M., Negro, M.J., 2009. Pretreatment technologies for an efficient bioethanol production process based on enzymatic hydrolysis: a review. Bioresour. Technol. 101 (13), 4851–4861. Bournay, L., Casanave, D., Delfort, B., Hillion, G., Chodorge, J.A., 2005. New heterogeneous process for biodiesel production: a way to improve the quality and the value of the crude glycerin produced by biodiesel plants. Catal. Today 106, 190–192. Dadi, A.P., Schall, C.A., Varanasi, S., 2007. Mitigation of cellulose recalcitrance to enzymatic hydrolysis by ionic liquid pretreatment. Appl. Biochem. Biotechnol. 136–140, 407–421. Gunnarsson, C.C., Petersen, C.M., 2007. Water hyacinths as a resource in agriculture and energy production: a literature review. Waste Management 27, 117–129. Gutierrez, A., Rio, J.C., 2005. Chemical characterization of pitch deposits produced in the manufacturing of high-quality paper pulps from hemp fibers. Bioresour. Technol. 96, 1445–1450. Hendriks, A.T.W.M., Zeeman, G., 2009. Pretreatments to enhance the digestibility of lignocellulosic biomass. Bioresour. Technol. 100, 10–18. Hu, G., Heitmann, J.A., Rojas, O.J., 2008. Feedstocks pretreatment strategies for producing ethanol from wood, bark and forest residues. BioResources 3 (1), 270–294. Ito, T., Nakashimada, Y., Senba, K., Matsui, K., Nishio, N., 2005. Hydrogen and ethanol production from glycerol-containing wastes discharged after biodiesel manufacturing process. J. Biosci. Bioeng. 100 (3), 260–265. Karunanithy, C., Muthukumarappan, K., Julson, J.L., 2008. Enzymatic hydrolysis of corn stover pretreated in high shear bioreactor. In: ASABE Annual International Meeting, Rhode Island. Kim, S., Dale, B.E., 2004. Global potential bioethanol production from wasted crops and crop residues. Biomass Bioenerg. 26, 361–375. Larsson, S., Palmqvist, E., Hahn-Hägerdal, B., Tengborg, C., Stenberg, K., Zacchi, G., Nilvebrant, N.O., 1999. The generation of fermentation inhibitors during dilute acid hydrolysis of softwood. Enzyme Microb. Technol. 24, 151–159. Laus, G., Bentivoglio, G., Schottenberger, H., Kahlenberg, V., Kopacka, H., Röer, T., et al., 2005. Ionic liquids: current developments, potential and drawbacks for industrial applications. Lenzinger Berichte 84, 71–85. Li, Q., He, Y.C., Xian, M., Jun, G., Xu, X., Yang, J.M., et al., 2009. Improving enzymatic hydrolysis of wheat straw using ionic liquid 1-ethyl-3-methyl imidazolium diethyl phosphate pretreatment. Bioresour. Technol. 100, 3570–3575. Liu, L.Y., Chen, H.Z., 2006. Enzymatic hydrolysis of cellulose materials treated with ionic liquid [BMIM]Cl. Chin. Sci. Bull. 51 (20), 2432–2436. Miller, G.L., 1959. Use of dinitrosalicylic acid reagent for the determination of reducing sugars. Anal. Chem. 31, 426–428. Mishima, D., Kuniki, M., Sei, K., Soda, S., Ike, M., Fujita, M., 2006. Ethanol production from candidate energy crops: water hyacinth (Eichhornia crassipes) and water lettuce (Pistia stratiotes L.). Bioresour. Technol. 99 (7), 2495–2500. Mishima, D., Kuniki, M., Sei, K., Soda, S., Ike, M., Fujita, M., 2008. Ethanol production from candidate energy crops: water hyacinth (Eichhornia crassipes) and water lettuce (Pistia stratiotes L.). Bioresour. Technol. 99, 2495–2500. Mosier, N., Wyman, C.E., Dale, B.E., Elander, R., Lee, Y.Y., Holtzapple, M., et al., 2005. Features of promising technologies for pretreatment of lignocellulosic biomass. Bioresour. Technol. 96 (6), 673–686.

4424


Mousdale, D.M., 2008. Biofuels: Biotechnology, Chemistry, and Sustainable Development. CRS Press, Boca Rato, London, New York. Nigam, J.N., 2002. Bioconversion of water-hyacinth (Eichhornia crassipes) hemicellulose acid hydrolysate to motor fuel ethanol by xylose–fermenting yeast. J. Biotechnol. 97, 107–116. Shi, J., Sharma-Shivappa, R.R., Chinn, M.S., 2009. Microbial pretreatment of cotton stalks by submerged cultivation of Phanerochaete chrysosporium. Bioresour. Technol. 100, 4388–4395. Sluiter, A., Hames, B., Ruiz, R., Scarlata, C., Sluiter, J., Templeton, D., Crocker, D., 2008. Determination of structural carbohydrates and lignin in biomass. Laboratory Analytical Procedure (LAP). National Renewable Energy Laboratory (NREL), NREL/TP-510-42618. Sousa, L.C., Chundawat, S.P.S., Balan, V., Dale, B.E., 2009. ‘Cradle-to-grave’ assessment of existing lignocelluloses pretreatment technologies. Curr. Opin. Biotechnol. 20, 1–9. Sun, F., Chen, H., 2008. Enhanced enzymatic hydrolysis of wheat straw by aqueous glycerol pretreatment. Bioresour. Technol. 99, 6156–6161.

Talebnia, F., Karakashev, D., Angelidaki, I., 2009. Production of bioethanol from wheat straw: an overview on pretreatment, hydrolysis and fermentation. Bioresour. Technol. 101, 4744–4753. Tomas-Pejo, E., Oliva, J.M., Ballesteros, M., 2008. Realistic approach for full-scale bioethanol production from lignocellulose: a review. J. Sci. Ind. Res. 67, 874– 884. Yazdani, S.S., Gonzalez, R., 2007. Anaerobic fermentation of glycerol: a path to economic viability for the biofuels industry. Curr. Opin. Biotechnol. 18, 213– 219. Zavrel, M., Bross, M., Funke, M., Büchs, J., Spiess, A.C., 2009. High-throughput screening for ionic liquids dissolving (ligno-)cellulose. Bioresour. Technol. 100, 2580–2587. Zhao, H., Jonesa, C.L., Bakerb, G.A., Xia, S.Q., Olubajo, O., Persona, V.N., 2009. Regenerating cellulose from ionic liquids for an accelerated enzymatic hydrolysis. J. Biotechnol. 139, 47–54. Zhu, S., Wu, Y., Chen, Q., Yu, Z., Wang, C., Jin, S., et al., 2006. Dissolution of cellulose with ionic liquids and its application: a mini-review. Green Chem. 8, 325–327.


A comparison of chemical pretreatment methods for improving saccharification of cotton stalks Rebecca A. Silverstein a, Ye Chen a, Ratna R. Sharma-Shivappa Michael D. Boyette a, Jason Osborne b a

a,*

,

Department of Biological and Agricultural Engineering, Campus Box 7625, North Carolina State University, Raleigh, NC 27695-7625, USA b Department of Statistics, Campus Box 8203, North Carolina State University, Raleigh, NC 27695-7625, USA Received 10 October 2006; accepted 12 October 2006 Available online 8 December 2006

Abstract The effectiveness of sulfuric acid (H2SO4), sodium hydroxide (NaOH), hydrogen peroxide (H2O2), and ozone pretreatments for conversion of cotton stalks to ethanol was investigated. Ground cotton stalks at a solid loading of 10% (w/v) were pretreated with H2SO4, NaOH, and H2O2 at concentrations of 0.5%, 1%, and 2% (w/v). Treatment temperatures of 90 C and 121 C at 15 psi were investigated for residence times of 30, 60, and 90 min. Ozone pretreatment was performed at 4 C with constant sparging of stalks in water. Solids from H2SO4, NaOH, and H2O2 pretreatments (at 2%, 60 min, 121 C/15 psi) showed significant lignin degradation and/or high sugar availability and hence were hydrolyzed by Celluclast 1.5 L and Novozym 188 at 50 C. Sulfuric acid pretreatment resulted in the highest xylan reduction (95.23% for 2% acid, 90 min, 121 C/15 psi) but the lowest cellulose to glucose conversion during hydrolysis (23.85%). Sodium hydroxide pretreatment resulted in the highest level of delignification (65.63% for 2% NaOH, 90 min, 121 C/15 psi) and cellulose conversion (60.8%). Hydrogen peroxide pretreatment resulted in significantly lower (p 6 0.05) delignification (maximum of 29.51% for 2%, 30 min, 121 C/15 psi) and cellulose conversion (49.8%) than sodium hydroxide pretreatment, but had a higher (p 6 0.05) cellulose conversion than sulfuric acid pretreatment. Ozone did not cause any significant changes in lignin, xylan, or glucan contents over time. Quadratic models using time, temperature, and concentration as continuous variables were developed to predict xylan and lignin reduction, respectively for sulfuric acid and sodium hydroxide pretreatments. In addition, a modified severity parameter (log M0) was constructed and explained most of the variation in xylan or lignin reduction through simple linear regressions. 2006 Elsevier Ltd. All rights reserved. Keywords: Delignification; Bioethanol; Modeling; Lignocellulose; Enzymatic hydrolysis

1. Introduction Growing concerns over the environmental impact of fossil fuels and their inevitable depletion have led to intense research on the development of alternative energy sources that can reduce the United States dependence on foreign oil imports. Biomass, which includes animal and human waste, trees, shrubs, yard waste, wood products, grasses, and agricultural residues such as wheat straw, corn stover,

*

Corresponding author. Tel.: +1 919 515 6746; fax: +1 919 515 7760. E-mail address: [email protected] (R.R. Sharma-Shivappa).

0960-8524/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2006.10.022

rice straw, and cotton stalks, is a renewable resource that stores energy from sunlight in its chemical bonds (McKendry, 2002). It can be processed either chemically or biologically by breaking the chemical bonds to extract energy in the form of biofuels such as bioethanol, biodiesel, and methane. Currently, corn is the primary raw material for ethanol production in the United States. Starch, which constitutes about 70% of the corn kernel, is easily broken down into glucose that is then fermented to ethanol. However, the corn to ethanol industry draws its feedstock from a food stream and is quite mature with little possibility of process improvements (Ingram and Doran, 1995). Lignocellulosic

R.A. Silverstein et al. / Bioresource Technology 98 (2007) 3000–3011

feedstocks, which have the potential to reduce the cost of producing ethanol because they are less expensive than corn and available in large quantities, offer a plausible alternative. One promising technology is to convert this abundant and renewable lignocellulosic biomass to ethanol through an enzyme-based process (Schell et al., 2003). The conversion of lignocellulosic biomass to ethanol is, however, more challenging than corn due to the complex structure of the plant cell wall. Pretreatment is required to alter the structural and chemical composition of lignocellulosic biomass to facilitate rapid and efficient hydrolysis of carbohydrates to fermentable sugars (Chang and Holtzapple, 2000). A variety of physical (comminution, hydrothermolysis), chemical (acid, alkali, solvents, ozone), physico-chemical (steam explosion, ammonia fiber explosion), and biological pretreatment techniques have been developed to improve the accessibility of enzymes to cellulosic fibers (Moiser et al., 2005). Acid pretreatment involves the use of sulfuric, nitric, or hydrochloric acids to remove hemicellulose components and expose cellulose for enzymatic digestion (Schell et al., 2003). Agricultural residues such as corncobs and stovers have been found to be particularly well suited to dilute acid pretreatment (Torget et al., 1991). Alkali pretreatment refers to the application of alkaline solutions to remove lignin and various uronic acid substitutions on hemicellulose that lower the accessibility of enzyme to the hemicellulose and cellulose (Chang and Holtzapple, 2000). Generally, alkaline pretreatment is more effective on agricultural residues and herbaceous crops than on wood materials (Hsu, 1996). Peroxide pretreatment enhances enzymatic conversion through oxidative delignification and reduction of cellulose crystallinity (Gould, 1985). Increased lignin solubilization and cellulose availability were observed during the peroxide pretreatment of wheat straw (Martel and Gould, 1990), Douglas fir (Yang et al., 2002), and oak (Kim et al., 2001). Ozonation is another attractive pretreatment method that does not leave strong acidic, basic, or toxic residues in the treated material (Neely, 1984). The effect of ozone pretreatment has been found to be essentially limited to lignin degradation. Hemicellulose is slightly attacked, while cellulose is hardly affected (Sun and Cheng, 2002). Ozonation has been widely used to reduce the lignin content of both agricultural and forestry wastes (Neely, 1984). Cotton (Gossypium hirsutum), which is one of the most abundant crops in the southern United States, apart from being invaluable for the textile industry is a significant source of lignocellulosic biomass. In 2003, nearly 13.2 million acres of cotton were planted nationwide as a result of increased world demand for cotton (USDA, 2004). The increase in cotton planting is highly beneficial for economic development, but it also raises concerns about the disposal of cotton stalks left in the field (TBWEF, 2004) that serve as breeding ground for pests. Cotton stalks, which mainly contain lignocellulose, have the potential to serve as a low-cost feedstock to increase the production of fuel

3001

ethanol through proper pretreatment, hydrolysis, and fermentation. Conversion of this agricultural waste into a value-added product can provide an environmentally sound method of disposal and simultaneous destruction of feeding and fruiting sites of boll weevils and other insects. To fully utilize cotton stalk as a feedstock for ethanol production, optimal pretreatment is required to render the cellulose fibers more amenable to the action of hydrolytic enzymes. This study was therefore initiated to: (1) investigate the effect of sulfuric acid, sodium hydroxide, hydrogen peroxide, and ozone pretreatments of cotton stalks, (2) develop models to predict lignin degradation and xylan solubilization percentage during sulfuric acid and sodium hydroxide pretreatments, and (3) identify pretreatment(s) which provide the highest cellulose to glucose conversion during subsequent enzymatic hydrolysis. 2. Methods 2.1. Biomass Cotton stalks, harvested in early October 2003, were obtained from Cunningham Research Station in Kinston, NC. The stalks were shredded and bailed in the field soon after the cotton was picked, and then transported to North Carolina State University in Raleigh, NC. Prior to composition analysis, the biomass which consisted primarily of stalks, leaves, cottonseed, and cotton residue, was ground to a 40 mesh particle size. The feedstock was ground to pass a 3 mm sieve in a Thomas Wiley Laboratory Mill (Model No. 4) and stored in sealed plastic bags at room temperature until use for pretreatment. 2.2. Analysis methods The total solids, acid soluble lignin, and acid insoluble lignin (acid-insoluble material) content of the untreated cotton stalks and the solid fraction remaining after pretreatment were determined by Laboratory Analytical Procedures (LAP) from the National Renewable Energy Laboratory (NREL) (Ehrman, 1994, 1996; Templeton and Ehrman, 1994). Ash content, extractives, and holocellulose (combination of hemicellulose and cellulose), were determined for the untreated stalks by the gravimetric methods developed by Han and Rowell (1997). The carbohydrate contents of the untreated cotton stalks and pretreated solids were determined by measuring the hemicellulose (xylan, galactan, and arabinan) and cellulose (glucan) derived sugars. The composition of the hydrolysate from enzymatic hydrolysis was determined by measuring glucose and xylose using high performance liquid chromatography (HPLC). The LAP-002 analysis procedure from NREL was modified for use with a Dionex-300 HPLC system (Ruiz and Ehrman, 1996). The HPLC system was equipped with a CarboPactrade PA10 (4 · 250 mm) anion exchange column, a guard column

3002


(4 · 50 mm), an automated sampler, a gradient pump, and a pulsed amperometric detector with a gold working electrode (Dionex Corp., Sunnyvale, CA). The mobile phase used was 10 mM NaOH at a flow rate of 1 mL/min. Monomeric sugars (arabinose, galactose, glucose, xylose) with the concentrations of 0, 10, 30, and 50 mg/l were used as standards. Prior to HPLC injection, all samples (derived from solids and hydrolysate) were neutralized with barium hydroxide, centrifuged at 5000g for 10 min, and filtered through 0.45 lm Millipore filters. Fucose was added as an internal standard for the samples analyzed.

at 5000g for 10 min. The supernatant was removed for sugar content analysis (Yang and Wyman, 2004). The percent glucan conversion was calculated as follows: % glucan conversion ¼

%GH 100 %GP

ð1Þ

where GH is the dry-weight percentage of glucose in enzyme hydrolysis supernatant (g glucose/g solids hydrolyzed %), GP is the dry-weight percentage glucan in pretreated solids (g glucose/g solids pretreated %). A similar equation was used to determine percent xylan conversion.

2.3. Pretreatments Sulfuric acid (H2SO4), sodium hydroxide (NaOH), and hydrogen peroxide (H2O2) at concentrations of 0.5%, 1%, and 2% (w/v) were used to pretreat 10 g ground cotton stalk samples at a solid loading of 10% (w/v). Treatments were performed in triplicate at 90 C and in an autoclave at 121 C with 15 psi (103.4 kPa) pressure for residence times of 30, 60, and 90 min. Ozone pretreatment was performed by continuously sparging ozone gas through a 10% (w/v) mixture of cotton stalks and deionized water for 30, 60, and 90 min. The reaction temperature was controlled at 4 C by placing the reaction flask in a water bath. Ozone gas was generated on site by passing 5 L/min of oxygen through an ozonator (AOS-1M/MS, Applied Ozone Systems, CA). Ozone concentrations in pure deionized water sparged with ozone were determined by measuring the absorbance using a Shimadzu PharmaSpec UV-1700 spectrophotometer (Columbia, MD) at 258 nm (Sharma et al., 2002). The pretreated solids were washed with 750 mL of hot deionized water and used for determination of total solids, acid insoluble lignin, and glucan and xylan prior to storage at 4 C for enzymatic hydrolysis. 2.4. Enzymatic hydrolysis Cellulase from Trichoderma reesei (Celluclast 1.5 L, Sigma Co., St. Louis, MO) with an activity of 96.1 filter paper unit FPU/mL enzyme solution, supplemented with cellobiase from Aspergillus niger (Novozyme 188, EC No. 232-589-7, Sigma Co., St. Louis, MO) at a ratio of 1:1.75 was used for hydrolysis experiments. The protein contents of Celluclast 1.5 L and Novozym 188 have been reported to be 191 and 143 mg/mL, respectively (Shoemaker, 2004). Pretreated cotton stalks at 5% solids loading (grams dry weight per 100 mL) in 50 mM acetate buffer (pH 4.8) containing 40 lg/mL tetracycline (an antibiotic added to avoid microbial contamination) were preincubated in flasks in a shaking water bath at 50 C and 150 rpm for 10 min. The hydrolysis, conducted at a cellulase activity of 40 FPU/g cellulose, was initiated by adding 2.18 mL and 3.82 mL of cellulase and cellobiase, respectively. Aliquots of 2.0 mL were taken at the termination of enzymatic hydrolysis after 72 h, immediately chilled on ice, and centrifuged

2.5. Data analysis and modeling The experimental design was crossed and complete with respect to temperature, time and concentration. Factorial effects models with main and interaction effects, on lignin reduction and xylan and glucan solubilization by sulfuric acid, sodium hydroxide, hydrogen peroxide, and ozone pretreatments, were fit using PROC GLM in SAS (SAS Institute, Cary, NC). Multiple comparisons among treatment means were carried out using Tukey’s procedure to control the experiment wise error rate at 0.05 for each response variable. In cases where interaction effects were significant, the SLICE option was used with the LSMEANS command of the GLM procedure to test for simple treatment effects of one factor while holding the other two factors constant. Empirical quadratic models with time, temperature, and concentration as continuous numeric variables were developed using SAS to predict percent lignin reduction for sodium hydroxide pretreatment and xylan solubilization for sulfuric acid pretreatment and were of the form y ¼ b0 þ b1 T þ b2 t þ b3 C þ b4 Tt þ b5 CT þ b6 CTt þ b7 Ct þ b8 t2 þ b9 C 2

ð2Þ

where T is the temperature (C); t is the time (min); C, the concentration (%); bn, the estimated regression coefficients, n = 0, 1, . . . , 9. The squared temperature term (T2) was not included in the model because only two temperatures were used during the experiments. This did not provide a sufficient number of degrees of freedom to estimate a regression coefficient for a squared term. Additionally, since the data set was relatively small to quantify the predictability of the models, the focus was on model development rather than assessment of predictive ability of the models. In addition, modeling based on combining the effects of time, temperature, and concentration into one single parameter was used to develop a linear model expressing the relationship between pretreatment severity and lignin reduction or xylan solubilization. Overend and Chornet (1987) initially defined this severity parameter to relate temperature and time for steam explosion pretreatment based on the assumption that pretreatment affects follow


first-order kinetics and obey the Arrhenius equation. Using this relationship they defined a reaction ordinate (R0, min) ðT r T b Þ R0 ¼ t exp ð3Þ 14:75 where t is the residence time (min), Tr is the reaction temperature (C), Tb is the base temperature (100 C) and 14.75 is the conventional energy of activation assuming the overall reaction is hydrolytic and the overall conversion is first order. The logarithm of the reaction ordinate defines the severity during steam explosion pretreatment such that severity is equal to log (R0). A modified severity parameter was later developed by Chum et al. (1990) for use with sulfuric acid pretreatment Tr Tb M 0 ¼ t C n exp ð4Þ 14:75 where M0 is the modified severity parameter; t is the residence time (min); C is the chemical concentration (wt%); Tr is the reaction temperature (C); Tb is the base temperature (100 C); n is an arbitrary constant. This equation was adapted for application to sodium hydroxide pretreatment by replacing the acid concentration with the alkali concentration and calculating a different n-value. 3. Results and discussion 3.1. Characterization of cotton stalk The chemical composition of cotton stalks varies depending on the growing location, season, harvesting methods, as well as analysis procedures (Agblevor et al., 2003). The composition of the cotton stalks used in this study is presented in Table 1. Based on the HPLC carbohydrate analysis, the sugar fraction was 41.8% of the dry biomass. Glucan, which is derived from both the cotton fiber and plant cell wall, was the major component at 31.1%. Xylan, which is the major hemicellulose constituent, was 8.3%. Arabinan and galactan accounted for only a small portion of the biomass, while mannan was not detected.

Table 1 Summative composition of untreated cotton stalks Component

Percentage (%)a

Holocellulose Glucan Xylan Arabinan Galactan Acid-insoluble lignin Acid-soluble lignin Extractives Ash Other

41.8 31.1 8.3 1.3 1.1 27.9 2.2 9.0 6.0 13.1

a

Composition percentages are on a dry-weight basis.

3003

Both glucan and xylan content were lower than the reported values of 40–50% glucan and 15–35% xylan for other agricultural residues and hardwoods (Milne et al., 1992). The holocellulose fraction, determined by the procedure of Han and Rowell (1997), was 51.1% of the total biomass. The discrepancy between the holocellulose content and total sugars is probably due to sugar degradation during the intense hydrolysis with sulfuric acid (Badger, 2002) used for the carbohydrate analysis procedure. Since the carbohydrate content of cotton stalks (based on monomeric sugars) determined by HPLC analysis was more likely to represent the actual sugars available after the treatments, subsequent calculations and analysis of data in this study were performed on the basis of HPLC measurements. The acid-insoluble material content of the cotton stalks (27.9%) was higher than expected. It was comparable to the acid-insoluble material content of hardwoods (18–25%), rather than that of herbaceous species and agricultural residues (10–20%) (McMillan, 1994). The acid insoluble material from woody biomass is normally classified as lignin. However, it would be incorrect to classify all the acid insoluble material from cotton stalks as lignin. A possible source of non-lignin, acid-insoluble material is the cottonseed. Cottonseed is composed of 32% hull, 23% protein, 12% fibers, 20% oil, and 14% carbohydrates. Upon analysis of the cottonseed from the Emporia gin in Virginia, Agblevor et al. (2003) discovered that the cottonseed contained 34% acid-insoluble material. The hull, which is lignocellulosic, and thus the only source of lignin, makes up only 32% of the cottonseed. Thus, the acid-insoluble material is expected to be composed of lignin and other condensable compounds. Since it is known that proteins condense and become insoluble in concentrated sulfuric acid (Agblevor et al., 1994), it could be surmised that high acid-insoluble material of the cottonseed, and in turn, the cotton stalks, is a combination of lignin and condensed proteins (Agblevor et al., 2003). However, because the majority of the acid insoluble material is lignin, it has been referred to as such in this study to limit confusion. The composition of cotton stalks used in this study agreed with that of cotton gin residue (immature bolls, cottonseed, hulls, sticks, leaves, and dirt) analyzed by Agblevor et al. (2003). The residue was sampled two to three times on different days from five different cotton gins across Virginia. The composition varied depending on the discharge date and the gin location, with approximate ranges of each component being 21–38% glucan, 3–12% xylan, 0.5–3% each of mannan, galactan, and arabinan, 5–13% extractives, 18–26% acid-insoluble lignin, and 7–14% ash. 3.2. Effect of sulfuric acid pretreatment Dilute acid pretreatment of lignocellulosic biomass is one of the most effective pretreatment methods which predominantly affect hemicellulose with little impact on lignin degradation. The lignin reduction, xylan and glucan

3004


solubilization during sulfuric acid pretreatment of cotton stalks is shown, respectively in Fig. 1a–c and the solids recovery after pretreatment is presented in Table 2. ANOVA tables providing information on which treatment parameters had significant impact during pretreatment are not presented in this text but are available from Silverstein (2004). Elevated temperature, residence time, and acid concentration reduced solids recovered after pretreatment. The acid insoluble lignin, remaining after acid pretreatment varied from 28.72% (1%, 30 min, 90 C) to 40.68% (2%, 60 min, 121 C/15 psi). The xylan and glucan contents of the sulfuric acid pretreated samples ranged from 0% (2%, 60 min, 121 C/15 psi) to 10.24% (0.5%, 30 min, 90 C) and 33.74% (2%, 30 min, 90 C) to 46.3% (2%, 60 min, 121 C/15 psi) respectively. The reduction of lignin, based on a comparison between the weight of lignin in the initial

10 g (dry-weight) sample before pretreatment and the weight of lignin in the solids remaining after pretreatment, ranged from 2.27% to 24.16%. Concentration had a significant (p 6 0.05) effect on delignification for treatments at 90 C for 60 min and 121 C/15 psi for 30, 60, and 90 min. Increasing the temperature from 90 C to 121 C/ 15 psi significantly increased delignification with pretreatment for 60 min at 2% H2SO4 and 90 min at 0.5%, 1%, and 2% H2SO4. The filtrate from the lignin analysis was used to quantify the remaining hemicellulose and cellulose. The xylan content, which makes up the largest portion of hemicellulose in the cotton stalks, is the most important indicator of pretreatment effectiveness. Arabinan and galactan, although making up 1.3% and 1.1%, respectively of the untreated sample, were below the HPLC detection limit after

Lignin reduction (%)

35 30 25 20 15 10 5 0 30

60

90

30

90

60

90

Time (min) Temp. (o C)

121/15 psi

Xylan solubilization (%)

120 100 80 60 40 20 0 30

60

90

30

90

60

90


121/15 psi

Glucan solubilization (%)

30 25 20 15 10 5 0 30

60 90

90

30

60 121/15 psi

90


Fig. 1. (a) Lignin reduction, (b) xylan solubilization, and (c) glucan solubilization in sulfuric acid pretreated samples as a function of residence time, temperature, and concentration.

R.A. Silverstein et al. / Bioresource Technology 98 (2007) 3000–3011 Table 2 Percent solids recovery after pretreatments Time (min), conc. (%), temp. (C)

Solids recovered after pretreatment (%)a,b Sulfuric acid

Sodium hydroxide

Hydrogen peroxide

30, 30, 30, 60, 60, 60, 90, 90, 90, 30, 30, 30, 60, 60, 60, 90, 90, 90,

80.27 85.05 81.18 83.32 83.56 73.42 83.83 77.87 75.02 75.45 67.71 62.16 73.07 64.44 56.93 68.23 60.56 56.95

75.47 70.42 62.22 70.79 60.64 53.40 74.17 63.57 55.39 71.56 68.24 60.02 69.16 59.07 55.14 72.95 58.11 54.50

83.16 82.27 83.51 86.02 84.28 81.93 85.59 83.21 80.44 84.12 81.64 76.83 83.97 81.15 74.42 85.04 79.26 72.59

0.5, 1.0, 2.0, 0.5, 1.0, 2.0, 0.5, 1.0, 2.0, 0.5, 1.0, 2.0, 0.5, 1.0, 2.0, 0.5, 1.0, 2.0,

90 90 90 90 90 90 90 90 90 121/15 psi 121/15 psi 121/15 psi 121/15 psi 121/15 psi 121/15 psi 121/15 psi 121/15 psi 121/15 psi

RMSE R-square Tukey’s HSD

2.59 0.95 8.04

(3.86) (2.81) (4.24) (2.61) (0.36) (3.28) (1.61) (4.96) (2.56) (1.12) (1.97) (0.41) (0.73) (1.88) (3.04) (3.04) (1.98) (0.99)

3.22 0.88 10.02

(3.92) (1.28) (3.30) (4.58) (2.50) (1.69) (1.44) (1.38) (3.98) (7.97) (4.87) (2.03) (0.83) (1.49) (1.29) (1.32) (1.82) (1.02)

(1.57) (1.41) (0.52) (1.37) (1.61) (2.21) (5.83) (1.39) (2.43) (1.36) (0.81) (0.82) (2.19) (4.83) (2.90) (1.55) (1.76) (0.60)

2.38 0.78 7.28

a

Percentages calculated from value on a dry-weight basis. Data are averages of three replicates. Numbers in parentheses represent standard deviations. b

pretreatment, and therefore xylan was the only hemicellulose sugar determined hereafter. Sulfuric acid pretreatment effectively solubilized 14.57% of the xylan for the least severe pretreatment (0.5%, 30 min, 90 C) and 95.2% for the most severe treatment (2%, 90 min, 121 C/15 psi) (Fig. 1b). Results from this study are comparable to those obtained by Varga et al. (2002), who observed 85% solubilization of hemicellulose in corn stover at 121 C/15 psi for 1 h with 5% H2SO4. Increasing temperature had most pronounced effect on xylan solubilization. The temperature effect was significant for all combinations of time and concentration, showing that 121 C/15 psi is more effective for xylan solubilization than 90 C. No significant concentration effect on xylan solubilization was detected at 90 C for 30 min or 60 min treatment (p > 0.05). This indicated increasing acid concentration for the two lowest combinations of time and temperature did not increase the amount of xylan solubilization. In addition, there was no significant time effect at 90 C, 0.5% acid, indicating that the severity of the treatment at the lowest concentration and temperature does not show any significant improvement with an increase in time from 30 to 90 min. At 2% acid, 60 min, and 121 C/15 psi no xylose was detected. The results obtained at 2% acid, 90 min, and 121 C/15 psi were similar, with two of the three replicates showing no detectable levels of xylan and the third sample possessing only 2.44% xylose. Possible explanations for this could be that (1) there was complete solubilization of xylan during pretreatment and/or (2) the amount of xylan remaining in the sample,

3005

treated at 121 C/15 psi for 60 and 90 min and used for analysis was lower than the detection limit for HPLC analysis. The percentage of glucan solubilization due to sulfuric acid pretreatments was between 10.00% (0.5%, 60 min, 121 C/15 psi) and 23.88% (2%, 90 min, 121 C/15 psi) (Fig. 1c). Temperature and concentration had significant (p 6 0.05) effect on glucan solubilization. During pretreatment it is desirable that the cellulose portion of the biomass remain virtually unaffected. However, during acid pretreatment of cotton stalks slightly higher glucan reduction was observed (Kim et al., 2001) because the cellulose rich, loose cotton fiber, in the stalks, is not imbedded in lignin and hemicellulose. The acid therefore has direct access to the cellulose during pretreatment and can cause more glucan degradation than is usually the case with other feedstocks. 3.3. Effect of sodium hydroxide pretreatment Using sodium hydroxide to pretreat lignocellulosic materials is an alternative to sulfuric acid pretreatment. The main effect of sodium hydroxide pretreatment on lignocellulosic biomass is delignification by breaking the ester bonds cross-linking lignin and xylan, thus increasing the porosity of biomass (Tarkov and Feist, 1969). As in acid pretreatment, elevation in temperature, residence time, and alkali concentration increased the loss of solids during NaOH pretreatment (Table 2). The amount of lignin in the solids after NaOH pretreatment ranged from 23.31% (30 min, 90 C) to 25.22% (30 min, 121 C/ 15 psi) for 0.5% NaOH, 19.46% (60 min, 121 C/15 psi) to 21.90% (30 min, 90 C) for 1% NaOH, and 17.64% (90 min, 121 C/15 psi) to 20.94% (30 min, 121 C/15 psi) for 2% NaOH. Changes in concentration caused significant (p 6 0.05) decrease in lignin. The xylan content of pretreated solids ranged from 7.91% (0.5%, 30 min, 90 C) to 13.00% (1%, 90 min, 121 C/15 psi) and the glucan content ranged from 35.54% (0.5%, 30 min, 90 C) to 50.33% (2%, 60 min, 121 C/15 psi). The maximum reduction in lignin was 65.63% with 2% NaOH treatment for 90 min at 121 C/15 psi (Fig. 2a). Varga et al. (2002) reported 95% reduction in lignin content as a result of pretreatment of corn stover with 10% NaOH for 1 h in the autoclave. The high reduction level may be attributed to a higher NaOH concentration of 10%, which in this study was limited to 2%. An increase in the concentration of NaOH significantly improved delignification at all combinations of temperature and time (p 6 0.05). There was no significant (p > 0.05) effect of time with 0.5% NaOH at either temperature, indicating that 0.5% NaOH is too low to affect delignification for treatment times up to 90 min and temperatures up to 121 C in the autoclave. The effect of temperature for sodium hydroxide pretreatment was significant (p 6 0.05) only when the residence time was 90 min at 1% and 2% NaOH. This indicates that increasing the temperature only improved the amount of lignin removal for longer times and higher concentrations.

3006



70 60 50 40 30 20 10 0 30

60

90

30

90

60

90

Time (min) Temp. (oC)

121/15 psi


80 60 40 20 0 30

60

90

30

Glucan solubiliza tion (%)

90

60

90


121/15 psi

50 40 30 20 10 0 30

60

90

90

30

60 121/15 psi

90


Fig. 2. (a) Lignin reduction, (b) xylan solubilization, and (c) glucan solubilization in sodium hydroxide pretreated samples as a function of residence time, temperature, and concentration.

Although xylan solubilization due to sodium hydroxide pretreatment was lower than that by sulfuric acid pretreatment (Fig. 2b), it is expected that solubilization of xylan in conjunction with substantial lignin reduction can improve enzymatic hydrolysis. Sodium hydroxide pretreatment resulted in xylan solubilization in the range of 13.90% (0.5%, 90 min, 90 C) to 40.02% (2%, 90 min, 90 C). Concentration, time, and temperature did not cause significant (p > 0.05) differences in percent xylan solubilization for any of the treatment combinations. The solubilization of glucan during NaOH pretreatment was between 12.82% (1%, 30 min, 121 C/15 psi) and 30.14% (2%, 60 min, 90 C) as illustrated in Fig. 2c. Glucan solubilization increased significantly with increasing concentration for 90 C at 90 min and the temperature effect was significant for 2% NaOH for 30 and 60 min. However,

the standard deviations among some replicates were rather large. This could be attributed to the heterogeneous nature of cotton stalks and the fact that amount of free cotton fiber could vary from one sample to the other.

3.4. Effect of hydrogen peroxide pretreatment Hydrogen peroxide pretreatment utilizes oxidative delignification to detach and solubilize the lignin and loosens the lignocellulosic matrix thus improving enzyme digestibility (Martel and Gould, 1990). The lignin reduction and xylan and glucan solubilization due to H2O2 pretreatment in this study are shown in Fig. 3a–c and the solids recovered after pretreatment are presented in Table 2. There was no evidence (p > 0.05) of any effects of either


3007


40 30 20 10 0 30

60

90

30

90

60

90


121/15 psi


40 30 20 10 0 30

60

90

30

90

60

90

Time (m in) Te m p. (o C)

121/15 psi


40 30 20 10 0 30

60

90

90

30

60 121/15 psi

90


Fig. 3. (a) Lignin reduction, (b) xylan solubilization, and (c) glucan solubilization in hydrogen peroxide pretreated samples as a function of residence time, temperature, and concentration.

of the treatment factors on the lignin content of pretreated solids. Hydrogen peroxide pretreatment led to 6.22% (0.5%, 90 min, 90 C) to 32.01% (2%, 60 min, 121 C/15 psi) delignification (Fig. 3a). These lignin degradations are lower than those reported in literature at alkaline conditions where pretreatment of sugar cane bagasse with 2% alkaline H2O2 resulted in 50% decrease in lignin and solubilization of most of the hemicellulose within 8 h at 30 C (Azzam, 1989). Determination of simple treatment effects for delignification showed that increasing the concentration from 0.5% to 2% did not significantly increase delignification for 30 min at 90 C probably because the residence time was too short at the lower temperature. The simple time effect was significant for 121 C/15 psi at 0.5% and 1% H2O2, which indicates that increasing the residence time

from 30 to 90 min showed significant improvements only for the two lower concentrations at the higher temperature. Temperature played a significant role in improving delignification for 0.5% at 60 min and 2% at 30 and 60 min but an increase in temperature significantly reduced the mean delignification for 0.5% at 90 min. The most severe pretreatment in the autoclave at 121 C for 90 min with 2% H2O2 had lower levels of delignification than the treatments at 30 and 60 min at 0.5% and 1%. This could be attributed to the decomposition of H2O2 at high temperature thus diminishing its oxidative delignification potential and to the long residence time which could result in recondensation or repolymerization of solubilized lignin. The solubilization of xylan due to H2O2 pretreatment averaged between 8.18% (0.5%, 60 min, 90 C) and 30.56% (2%, 30, 121 C/15 psi) (Fig. 3b) while the xylan

3008


content ranged from 8.69% (2%, 30 min, 121 C/15 psi) to 10.87% (2%, 90 min, 90 C). Concentration had a significant effect (p 6 0.05) on xylan solubilization for 90 min, 90 C treated samples and 30 min, 121 C treated samples. The simple temperature effect was significant for xylan solubilization for 0.5% and 1% at 60 min and 2% at 30, 60, and 90 min. The percentage of glucan in the pretreated solids remaining as a result of H2O2 pretreatment ranged from 28.4% (1%, 30 min, 90 C) to 34.1% (2%, 90 min, 90 C). Glucan solubilization on average was between 14.91% (0.5%, 60 min, 90 C) to 29.10% (2%, 30 min, 121 C/ 15 psi) as presented in Fig. 3c. Concentration did not have a significant effect on glucan solubilization. Significant differences (p 6 0.05) between percent glucan solubilization due to changes in temperature were noted for 2% at 90 min, while time played a significant role for 0.5 and 1% H2O2 at 90 C. 3.5. Effect of ozone pretreatment Pretreatment of lignocellulosic biomass with ozone gas has been reported to reduce both the lignin and hemicellulose contents of the treated materials (Ben-Ghedalia et al., 1980). The most substantial effect of ozone pretreatment is on degradation of the lignin polymer, followed by hemicellulose and cellulose solubilization (Quesada et al., 1999). In this study, ozone pretreatment reduced lignin in the range of 11.97–16.6% with no significant difference (p > 0.05) noted for treatment times of 30, 60, and 90 min (Table 3). The amount of xylan solubilized during ozone treatment ranged from 1.9% to 16.7%, while the amount of glucan solubilized was between 7.2% and 16.6%. The percent solubilization of xylan and glucan for 90 min treatment was significantly (p < 0.05) lower than the solubilization for 30 and 60 min. The concentrations of ozone measured in pure water after sparging for 30, 60, and 90 min were 16.96, 17.74, and 18.52 ppm, respectively. Ben-Ghedalia et al. (1980) reported a 50% decrease in both lignin and hemicellulose in ozone treated cotton stalks. Possible explanations for the differences between the results from this study and those from past studies include insufficient treatment times, inadequate ozone concentration, and

poor distribution of ozone gas throughout the cotton stalks because of inefficient sparging. 3.6. Enzymatic hydrolysis Acid pretreated samples resulting in maximum glucose availability (2% H2SO4, 60 min, 121 C/15 psi) were chosen for enzyme hydrolysis. This selection criterion was based on the fact that acid pretreatment has little effect on lignin degradation and the main treatment effect is on hemicellulose and cellulose solubilization. Alkali pretreatment caused delignification and glucan solubilization. The selection for NaOH pretreated samples was based on a compromise between having the lowest percentage of lignin in the pretreated solids, while maintaining a high percentage of glucan (2% NaOH, 60 min, 121 C/15 psi). For hydrogen peroxide, there were no significant differences between percentage glucan, xylan, or lignin in the pretreated solids for any of the treatments. Hence, the treatment with the highest percentage of glucan and the lowest percentage of lignin was chosen (2% H2O2, 60 min, 121 C/15 psi). Cellulose conversion of pretreated samples after 72 h of enzymatic hydrolysis is shown in Table 4. Sodium hydroxide pretreated samples had the highest cellulose conversion of 60.8%, followed by hydrogen peroxide (49.8%) and then sulfuric acid (23.8%). Differences in mean cellulose conversions for all the treatments were statistically significant (p 6 0.05). Hydrolysis of sodium hydroxide pretreated samples resulted in the highest xylan to xylose conversion (Table 4) at 62.57%, whereas hydrogen peroxide averaged 7.78% conversion. For the acid pretreated samples, no xylan was detected in the solids during the initial carbohydrate analysis, but an average of 14.3 mg xylose/g dry biomass was detected in the supernatant after enzymatic hydrolysis. This confirms the hypothesis that there was xylan in the stalks after pretreatment, but the amount was below the detection limit during sugar analysis. Detection of xylose in the hydrolysate may be attributed to a higher sugar concentration resulting from the hydrolyzed sample (5 g) being larger than that analyzed for carbohydrate content of pretreated solids (0.3 g). The difference in cellulose conversion during enzymatic hydrolysis is largely dependent on the difference in lignin

Table 3 Effect of ozone pretreatment on cotton stalks Time (min)

30 60 90c RMSE R-square Tukey’s HSD a b c

Reduction (%)a,b

Solids recovery (%)

Lignin

Xylan

Glucan

11.97 (2.91) 16.63 (2.60) 15.15 (3.02)

16.76 (7.32) 10.61 (8.12) 1.92 (2.89)

16.62 (7.80) 13.74 (3.64) 7.19 (0.36)

7.04 0.52 20.19

5.45 0.42 15.63

2.82 0.46 8.08

Percentages calculated from values on a dry-weight basis. Data are averages of three replicates. Numbers in parentheses represent standard deviations. Only two samples were used for 90 min treatment because the third replicate was an outlier.

90.44 (2.47) 88.66 (2.82) 91.66 (0.47) 2.18 0.32 5.46


3009

Table 4 Glucan and xylan conversion after enzymatic hydrolysis Pretreatment agent

Composition of hydrolysis supernatant and pretreated solida,b,c

Sulfuric acid

– 40.68 (1.44)

1.43 (0.16) 0.00 (0.00)

11.03 (0.66) 46.3 (2.89)

23.85 (1.21)

Sodium hydroxide

– 18.40 (0.16)

8.34 (0.15) 12.13 (0.40)

30.57 (0.56) 50.33 (1.84)

60.79 (2.75)

62.57 (2.57)

Hydrogen peroxide

– 25.59 (2.30)

0.90 (0.14) 10.00 (0.26)

17.21 (0.84) 34.53 (0.86)

49.82 (1.40)

7.78 (1.13)

Lignin

a b c d

Xylose

Glucan conversion (%)

Xylan conversion (%)

Glucose 0.00 (0.00)d

Composition percentages calculated from values on a dry-weight basis. Data are averages of three replicates. Numbers in parentheses represent standard deviations. Compositions of xylose and glucose in the hydrolysis supernatant are in upper rows while compositions of pretreated solids are in bottom rows. See text for explanation.

composition. The sulfuric acid and hydrogen peroxide pretreated samples had 2.2 times and 1.4 times the amount of lignin, respectively, compared to sodium hydroxide pretreated samples. Lignin is not attacked by the enzymes and therefore shields the cellulose during hydrolysis (Mansfield et al., 1999). Solubilization of xylan, on the other hand, seems to have a limited impact on cellulose digestibility. 3.7. Modeling Empirical quadratic models using time, temperature, and concentration as continuous variables and linear models relating a modified severity parameter to these variables were developed to predict xylan solubilization in sulfuric acid pretreatment and lignin reduction in sodium hydroxide pretreatment. These two treatment agents were chosen for modeling because they have been widely studied and seem to be the most promising pretreatments for use on cotton stalks. After eliminating the insignificant terms (p > 0.05) from the model based on the p-values from the Type III Sum of Squares ANOVA table (data not shown), the reduced empirical quadratic model used to quantify the percentage of xylan solubilized from the cotton stalks during sulfuric acid pretreatment was

The appropriate values for C, T, and t were plugged into the equations and the plots between fitted vs. observed values, for both percent xylan solubilization (Eq. (5)) and percent lignin reduction (Eq. (6)), had slopes of 0.97. Both models had high R2-values and slopes close to 1 thus indicating good agreement between the experimental data and the models. Linear models relating a modified severity parameter that combines the effects of time, temperature and concentration to the percentage solubilization of xylan by sulfuric acid pretreatment and to the reduction in lignin by sodium hydroxide pretreatment resulted in R2 values of 0.89 and 0.78, respectively. The n-values for sulfuric acid and sodium hydroxide pretreatments that provided the best model fits while keeping log (M0) positive were 0.849 and 3.90, respectively. The resulting equations were T r 100 0:849 exp M 0 ðsulfuric acidÞ ¼ tC ð7Þ 14:75 T r 100 ð8Þ M 0 ðsodium hydroxideÞ ¼ tC 3:90 exp 14:75 The model equation for determination of xylan solubilization during sulfuric acid pretreatment was developed by plotting log (M0) vs. % xylan solubilization (Fig. 4).

120

þ 0:2644t 22:6728C þ 0:6347CT 11:0451C

2

ð5Þ

The square of the correlation coefficient (R2) for the xylan solubilization model was 0.964. The percent lignin reduction model for sodium hydroxide containing significant terms from the Type III Sum of Squares ANOVA table (data not shown) had an R2 of 0.924 and was given by % lignin reduction ¼ 1:3705 þ 0:0002T þ 0:5554t

y = 53.508x - 55.043 R2 = 0.8926

100 80 60 40 20 0 0.8

1.3

1.8

2.3

2.8

-20 log Mo

þ 49:6254C þ 0:0904Ct 15:9216C 2 0:0047t2

Xylan Solubilization (%)

Xylan solubilization ð%Þ ¼ 117:6194 þ 1:0798T

ð6Þ

Fig. 4. Percent xylan solubilization vs. log (modified severity parameter) for sulfuric acid pretreatment.

3010


70

y = 8.6438x + 33.68 R2 = 0.7826

Lignin Reduction (%)

65 60 55 50 45 40 35 30 25 0

0.5

1

1.5

2 log Mo

2.5

3

3.5

4

Fig. 5. Percent lignin reduction vs. log (modified severity parameter) for sodium hydroxide pretreatment.

M0 was calculated using n = 0.849 and the model is represented by Eq. (9) % xylan solubilization ¼ 53:508 logðM 0 Þ 55:043

ð9Þ

The model equation for the reduction of lignin during sodium hydroxide pretreatment, using n = 3.90 to calculate M0, was obtained from Fig. 5 and is represented as % lignin reduction ¼ 8:6438 logðM 0 Þ þ 33:68

to water at high temperatures. Ozone pretreatment also did not perform as effectively as expected. Possible explanations include insufficient time, low ozone concentration, or uneven distribution of ozone throughout the sample. Compared with other pretreatments, sodium hydroxide pretreatment resulted in significantly (p < 0.05) higher cellulose conversion during the subsequent enzymatic hydrolysis. The empirical quadratic models successfully predicted percent xylan solubilization and percent lignin reduction and may be used in the development of better estimation tools. In addition, this study can serve as a step towards the optimization of pretreatment of cotton stalks. Nevertheless, different combinations of treatment factors, perhaps using higher temperatures or concentrations and application of higher pressures could be investigated. In addition, enzymatic hydrolysis using optimized pretreatment factors and ethanol fermentation need to be studied for bioethanol production since they could not be addressed in this study.

ð10Þ

The modified severity parameter model was validated by plotting the experimental values of percent xylan solubilization and percent lignin reduction against the model predicted values (Silverstein, 2004). The R2 from the plot of experimental vs. predicted % xylan solubilization was 0.88 with a slope of 0.95 indicating good predictive ability of the model. Predicted and experimental values for % lignin reduction during sodium hydroxide pretreatment resulted in an R2 of 0.72 and a slope of 0.99. Variation in predicted and experimental values may have likely been due to heterogeneity of cotton stalks and inability of the modified severity parameter to fully capture dependence of response variables on independent variables in the absence of variables such as stalk to cotton fiber ratio and solids loading. 4. Conclusions Sulfuric acid pretreatment substantially solubilized xylan in cotton stalks and temperature had the most significant effect on xylan solubilization. Data analysis indicated that there is a linearly increasing relationship between xylan solubilization and pretreatment severity. The most significant effect of sodium hydroxide pretreatment was on delignification with concentration of sodium hydroxide being the significant factor. Lignin reduction increased linearly with increase in pretreatment severity of sodium hydroxide. Hydrogen peroxide pretreatment resulted in lower lignin and xylan solubilization than expected. This was probably due to decomposition of hydrogen peroxide

References Agblevor, F.A., Evans, R.J., Johnson, K.D., 1994. Molecular-beam massspectrometric analysis of lignocellulosic materials. I. Herbaceous biomass. J. Anal. Appl. Pyrol. 30, 125–144. Agblevor, F.A., Batz, S., Trumbo, J., 2003. Composition and ethanol production potential of cotton gin residues. Appl. Biochem. Biotechnol. 105, 219–230. Azzam, A.M., 1989. Pretreatment of cane bagasse with alkaline hydrogen peroxide for enzymatic hydrolysis of cellulose and ethanol fermentation. J. Environ. Sci. Health B 24, 421–433. Badger, P.C., 2002. Ethanol from cellulose: a general review. In: Janick, J., Whipkey, A. (Eds.), Trends in New Crops and New Uses. ASHS Press, Alexandria, VA, pp. 17–21. Ben-Ghedalia, D., Shefet, G., Miron, J., 1980. Effect of ozone and ammonium hydroxide treatments on the composition and in vitro digestibility of cotton straw. J. Sci. Food Agric. 31, 1337–1342. Chang, V., Holtzapple, M., 2000. Fundamental factors affecting biomass enzymatic reactivity. Appl. Biochem. Biotechnol. 84–86, 5–37. Chum, H.L., Johnson, D.K., Black, S.K., Overend, R.P., 1990. Pretreatment-catalyst effects and the combined severity parameter. Appl. Biochem. Biotechnol. 24–25, 1–14. Ehrman, T., 1994. Method for determination of total solids in biomass. In: Laboratory Analytical Procedures No. 001. Golden, CO, National Renewable Energy Laboratory. Ehrman, T., 1996. Method for determination of acid-soluble lignin in biomass. In: Laboratory Analytical Procedures No. 004. Golden, CO, National Renewable Energy Laboratory. Gould, J.M., 1985. Studies on the mechanism of alkaline peroxide delignification of agricultural residues. Biotechnol. Bioeng. 27, 225–231. Han, J., Rowell, J., 1997. Chemical composition of fibers. In: Rowell, R., Young, R., Rowell, J. (Eds.), Paper Composites from Agro-Based Resources. CRC Lewis Publisher, New York, pp. 83–134. Hsu, T.A., 1996. Pretreatment of biomass. In: Wyman, C.E. (Ed.), Handbook on Bioethanol: Production and Utilization. Taylor & Francis, Washington, DC, pp. 179–195. Ingram, L.O., Doran, J., 1995. Conversion of cellulosic materials to ethanol. FEMS Microbiol. Rev. 16, 235–241. Kim, S.B., Um, B.H., Park, S.C., 2001. Effect of pretreatment of reagent and hydrogen peroxide on enzymatic hydrolysis of oak in percolation process. Appl. Biochem. Biotechnol. 91–93, 81–94.

R.A. Silverstein et al. / Bioresource Technology 98 (2007) 3000–3011 Mansfield, S.D., Mooney, C., Saddler, J.N., 1999. Substrate and enzyme characteristics that limit cellulose hydrolysis. Biotechnol. Prog. 15, 804–816. Martel, P., Gould, J.M., 1990. Cellulose stability and delignification after alkaline hydrogen-peroxide treatment of straw. J. Appl. Poly. Sci. 39, 707–714. McKendry, P., 2002. Energy production from biomass (part 1): overview of biomass. Bioresour. Technol. 83, 37–46. McMillan, J.D., 1994. Pretreatment of lignocellulosic biomass. In: Himmel, M.E., Baker, J.O., Overend, R.P. (Eds.), Enzymatic Conversion of Biomass for Fuels Production. American Chemical Society, Washington, DC, pp. 292–324. Milne, T.A., Chum, H.L., Agblevor, F.A., Johnson, D.K., 1992. Standardized analytical methods. Biomass Bioenergy 2, 341–366. Moiser, N., Wyman, C., Dale, B., Elander, R., Lee, Y.Y., Holtzapple, M., Ladisch, M., 2005. Features of promising technologies for pretreatment of lignocellulosic biomass. Bioresour. Technol. 96, 673–686. Neely, W.C., 1984. Factors affecting the pretreatment of biomass with gaseous ozone. Biotechnol. Bioeng. 26, 59–65. Overend, R.P., Chornet, E., 1987. Fractionation of lignocellulosics by steam-aqueous pretreatments. Philos. Trans. R. Soc. Lond. 321, 523–536. Quesada, J., Rubio, M., Gomez, D., 1999. Ozonation of lignin rich solid fractions from corn stalks. J. Wood Chem. Tech. 19, 115–137. Ruiz, R., Ehrman, T., 1996. Determination of carbohydrates in biomass by high performance liquid chromatography. In: Laboratory Analytical Procedures No. 002. Golden, CO, National Renewable Research Laboratory. Schell, D.J., Farmer, J., Newman, M., McMillan, J.D., 2003. Dilutesulfuric acid pretreatment of corn stover in pilot-scale reactor – investigation of yields, kinetics, and enzymatic digestibilities of solids. Appl. Biochem. Biotechnol. 105, 69–85. Sharma, R.R., Demirci, A., Beuchat, L.R., Fett, W.F., 2002. Inactivation of Escherichia coli O157:H7 on inoculated alfalfa seeds with ozonated water and heat treatment. J. Food Prot. 65, 447–451.

3011

Shoemaker, S., 2004. Advanced biocatalytic processing of heterogeneous lignocellulosic feedstocks to a platform chemical intermediate (lactic acid ester). Final report for award number DE-FC02-99CH11007. (accessed September, 2006). Silverstein, R., 2004. A comparison of chemical pretreatment methods for converting cotton stalks to ethanol. MS thesis, North Carolina State University. Available from: . Sun, Y., Cheng, J.J., 2002. Hydrolysis of lignocellulosic materials for ethanol production: a review. Bioresour. Technol. 83, 1–11. Tarkov, H., Feist, W.C., 1969. A mechanism for improving the digestibility of lignocellulosic materials with dilute alkali and liquid ammonia. Adv. Chem. Ser. 95 (1), 197–218. TBWEF. 2004. Texas Boll Weevil Eradication Foundation (TBWEF). (accessed 29.09.04). Templeton, D., Ehrman, T., 1994. Determination of acid-insoluble lignin in biomass. In: Laboratory Analytical Procedures No. 003. Golden, CO, National Renewable Energy Laboratory. Torget, R., Walter, P., Himmel, M., Grohmann, K., 1991. Dilute acid pretreatment of corn residues and short-rotation woody crops. Appl. Biochem. Biotechnol. 28–29, 75–86. USDA 2004. National Agricultural Statistics Service Crop Production. (accessed May, 2005). Varga, E., Scengyel, Z., Recaey, K., 2002. Chemical pretreatments of corn stover for enhancing enzymatic digestibility. Appl. Biochem. Biotechnol. 98–100, 73–87. Yang, B., Wyman, C.E., 2004. Effect of xylan and lignin removal by batch and flowthrough pretreatment on the enzymatic digestibility of corn stover cellulose. Biotechnol. Bioeng. 86, 88–95. Yang, B., Boussaid, A., Mansfield, S.D., Gregg, D.J., Saddler, J.N., 2002. Fast and efficient alkaline peroxide treatment to enhance the enzymatic digestibility of steam-exploded softwood substrates. Biotechnol. Bioeng. 77, 678–684.




Comparative material balances around pretreatment technologies for the conversion of switchgrass to soluble sugars Rebecca J. Garlock a,b,⇑, Venkatesh Balan a,b, Bruce E. Dale a,b, V. Ramesh Pallapolu c, Y.Y. Lee c, Youngmi Kim d, Nathan S. Mosier d, Michael R. Ladisch d, Mark T. Holtzapple e, Matthew Falls e, Rocio Sierra-Ramirez f, Jian Shi g, Mirvat A. Ebrik g, Tim Redmond g, Bin Yang g, Charles E. Wyman g, Bryon S. Donohoe h, Todd B. Vinzant h, Richard T. Elander h, Bonnie Hames i, Steve Thomas i, Ryan E. Warner j a

Biomass Conversion Research Laboratory, Department of Chemical Engineering and Materials Science, Michigan State University, 3900 Collins Road, Lansing, MI 48910, USA Great Lakes Bioenergy Research Center, Michigan State University, East Lansing, MI, USA c Department of Chemical Engineering, Auburn University, 212 Ross Hall, Auburn, AL 36849, USA d LORRE, Department of Agricultural and Biological Engineering, Purdue University, 500 Central Dr., West Lafayette, IN 47907, USA e Department of Chemical Engineering, Texas A&M University, 3122 TAMU, College Station, TX 77843-3122, USA f Universidad de los Andes Chemical Engineering Department Grupo de Conversion de Energia, Bogotá, Colombia g Center for Environmental Research and Technology, Department of Chemical and Environmental Engineering, Bourns College of Engineering, University of California at Riverside, 1084 Columbia Avenue, Riverside, CA 92507, USA h Chemical and Biosciences Center, National Renewable Energy Laboratory, 1617 Cole Blvd., Golden, CO 80401, USA i Ceres, Inc., 1535 Rancho Conejo Blvd, Thousand Oaks, CA 91320, USA j Genencor, A Danisco Division, 925 Page Mill Road, Palo Alto, CA 94304, USA b

a r t i c l e

i n f o

Article history: Received 8 January 2011 Received in revised form 1 April 2011 Accepted 1 April 2011 Available online 7 April 2011 Keywords: Cellulosic ethanol Enzymatic hydrolysis Material balance Pretreatment Switchgrass

a b s t r a c t For this project, six chemical pretreatments were compared for the Consortium for Applied Fundamentals and Innovation (CAFI): ammonia fiber expansion (AFEX), dilute sulfuric acid (DA), lime, liquid hot water (LHW), soaking in aqueous ammonia (SAA), and sulfur dioxide (SO2). For each pretreatment, a material balance was analyzed around the pretreatment, optional post-washing step, and enzymatic hydrolysis of Dacotah switchgrass. All pretreatments + enzymatic hydrolysis solubilized over two-thirds of the available glucan and xylan. Lime, post-washed LHW, and SO2 achieved >83% total glucose yields. Lime, post-washed AFEX, and DA achieved >83% total xylose yields. Alkaline pretreatments, except AFEX, solubilized the most lignin and a portion of the xylan as xylo-oligomers. As pretreatment pH decreased, total solubilized xylan and released monomeric xylose increased. Low temperature-long time or high temperature-short time pretreatments are necessary for high glucose release from late-harvest Dacotah switchgrass but high temperatures may cause xylose degradation. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction The world supply of fossil fuels is limited and will eventually fail to meet the global demand for energy, which continues to increase each year (Asif and Muneer, 2007). Because of this, it is necAbbreviations: 5-HMF, 5-hydroxymethylfurfural; AFEX, ammonia fiber expansion pretreatment; CAFI, Consortium for Applied Innovation and Fundamentals; CBU, cellobiase unit; DA, dilute sulfuric acid pretreatment; DP, degree of polymerization; DBP, dry biomass entering pretreatment; FPU, filter paper unit; Glc, glucose; GO, gluco-oligomers; LHW, liquid hot water pretreatment; SAA, soaking in aqueous ammonia pretreatment; SO2, sulfur dioxide pretreatment; XO, xylooligomers; Xyl, xylose. ⇑ Corresponding author at: Biomass Conversion Research Laboratory, Department of Chemical Engineering and Materials Science, Michigan State University, 3900 Collins Road, Lansing, MI 48910, USA. Tel.: +1 517 432 0157; fax: +1 517 337 7904. E-mail address: [email protected] (R.J. Garlock). 0960-8524/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2011.04.002

essary to research and develop alternative fuel sources before supplies become severely constrained. Bioethanol is one of the possible renewable alternatives to liquid fossil fuels. The majority of current world-wide production is derived from starch-based (e.g. corn) or sugar-based materials (e.g. sugar cane), but it is also possible to use lignocellulosic materials, such as agricultural and forestry residues, grasses, and trees as feedstocks. For the biological conversion route, a three-step process is necessary to adequately convert the cell wall sugars in these materials to ethanol: pretreatment followed by enzymatic hydrolysis and fermentation. Pretreatment is a process, either biological, chemical, physical, thermal, or some combination of these, which disrupts the cell wall structure and increases enzyme access to the cell wall carbohydrates, the substrate for lignocellulosic ethanol (Alvira et al., 2010; da Costa Sousa et al., 2009; Mosier et al., 2005; Yang and

11064

R.J. Garlock et al. / Bioresource Technology 102 (2011) 11063–11071

Wyman, 2008). The choice of pretreatment can have a significant impact on biorefinery costs (Aden and Foust, 2009; Eggeman and Elander, 2005) and most other processing decisions including feedstock selection, choice of enzymes and microbes, and waste treatment applications (Yang and Wyman, 2008). Because of the cost and pervasive impact of pretreatment on all aspects of the process, the choice of pretreatment method is extremely important. But this decision is hardly straightforward as there are a large number of pretreatment options currently available, each of which has certain advantages and disadvantages and some of which lend themselves better to certain feedstocks (Alvira et al., 2010; da Costa Sousa et al., 2009; Yang and Wyman, 2008). In order to effectively compare different pretreatment methods, it is important to conduct an accurate material balance, tracking the fate of cellulose and hemicellulose throughout the process and generating accurate yields. Many pretreatment methods result in liquid streams and the amount and type of components that are solubilized are dependent on the method. So if process yields are calculated based on the initial biomass composition without taking into account any pretreatment mass losses which are due to solid– liquid separation or post-washing, the results will be erroneous. One mass balance method is the carbon balance which tracks all of the carbon-based compounds in all of the process streams (Hatzis et al., 1996). Closing this balance can be difficult because of the complexity of measuring all of the carbon-based compounds. One study reports significant error-related issues when closing this balance for cellulase production (Schell et al., 2002). Another method that has been employed by previous CAFI projects and other pretreatment researchers is to measure the individual components, focusing on only those of interest such as glucose/ glucan, xylose/xylan, other sugars and lignin (Zhang et al., 2009; Zhu et al., 2010). Another option is to calculate a total biomass mass closure based on the solids content of each stream. But this is a less reliable method because of the difficulty involved in quantifying the solids content of dilute liquid or solution streams. It is often difficult to compare pretreatment methods based on literature because of the inconsistency in materials and methods used. The intent of the Consortium for Applied Fundamentals in Innovation (CAFI) projects was to provide a consistent basis for comparing a number of different thermochemical pretreatment methods (Wyman et al., 2005b). For these comparisons, each pretreatment method was carried out on the same feedstock, used the same enzymes and microbes during enzymatic hydrolysis and fermentation, and employed the same, consistent protocols wherever applicable throughout the process. The feedstocks used for the previous two CAFI projects were corn stover – an agricultural residue (Wyman et al., 2005a), and poplar – a hardwood (Wyman et al., 2009). This third round of the CAFI project examined and compared the effect of pretreatment and enzymatic hydrolysis of different varieties of switchgrass. There have been a number of recent papers that have looked at the feasibility of processing switchgrass for bioethanol (Bals et al., 2010; Himmelsbach et al., 2009; Xu et al., 2010; Yang et al., 2009) and also a recent review that discusses some of the older papers (Keshwani and Cheng, 2009). The goal of the portion of the CAFI III project reported in this manuscript was to conduct material balances around six thermochemical pretreatment methods: ammonia fiber expansion (AFEX) at Michigan State University, dilute sulfuric acid (DA) and sulfur dioxide (SO2) at University of California-Riverside, lime at Texas A&M University, liquid hot water (LHW) at Purdue University, and soaking in aqueous ammonia (SAA) at Auburn University. The objective was to compare the process yields and stream characteristics for the combined stages of pretreatment and enzymatic hydrolysis of switchgrass, using the same feedstock (Dacotah switchgrass), enzymes, and analytical methods. Because washing following pretreatment may not be necessary to improve digest-

ibility for all pretreatment methods with all feedstocks, postwashing was analyzed separately from pretreatment. 2. Methods 2.1. Dacotah switchgrass Dacotah switchgrass, an upland variety, was planted in 1999 in Pierre, SD by Ceres, Inc. (Thousand Oaks, CA) The material used for these experiments was produced in 2007 and harvested in late winter of 2008 after standing on the field over winter. Composition analysis of the switchgrass was performed by Ceres according to the standard NREL protocols (Sluiter et al., 2010). The samples provided by Ceres had initially been milled to pass through a 0.25 in. (6.35 mm) screen and were shipped to each participating university. At each university, prior to performing the pre-wash, the switchgrass was milled to pass through a 2 mm screen using either a knife mill or coffee grinder. This secondary size reduction was performed so the biomass would be the appropriate particle size for composition analysis throughout the process. 2.2. Pre-wash A pre-wash step was performed at each university to remove any soluble sugars which could mask the solubilization of cell wall sugars. Batches of Dacotah switchgrass (200 g each) were soaked in 2 L of 80–90 °C distilled water for 10–15 min. The switchgrass slurry was vacuum-filtered through Whatman No. 1 filter paper (Whatman Ltd.). This process was repeated three times and after each wash step, a portion of the filtrate was retained for oligomeric sugar analysis. The washed solids were dried in a 45 °C oven. The extracted weight loss of the switchgrass was determined by subtracting the dry weight of the washed switchgrass and the dry mass loss to the filter paper from the initial dry weight. 2.3. Solids composition The composition of the pre-washed switchgrass (structural sugars and lignin) was determined using the standard two-stage extraction followed by a two-step acid hydrolysis (Sluiter et al., 2010). The composition of the solids following the pretreatment, post-wash, and enzymatic hydrolysis was determined using the same method, but the extraction step was not performed due to the potential loss of soluble biomass components. 2.4. Soluble total and oligomeric sugar analysis Oligomeric sugar analysis was conducted using either the standard NREL method for oligomeric sugar determination of liquid streams (Sluiter et al., 2008) which uses an autoclave based acid hydrolysis, or a modified version of this method. The modified method was identical except that it was scaled down to use 2 mL of sample and assays were run in duplicate in 10 mL screw-cap culture tubes. The tubes were incubated in a 121 °C bench-top hot plate for one hour, cooled on ice, and filtered into HPLC vials. The oligomeric sugar concentration was determined by subtracting the monomeric sugar concentration of the non-hydrolyzed samples from the total sugar concentration of the acid hydrolyzed samples. 2.5. HPLC analysis The monomeric and total (monomeric + oligomeric) sugar concentrations were determined for the pre-wash liquid, pretreatment liquor, post-wash liquid, and enzymatic hydrolysate. HPLC samples were analyzed using a Bio-Rad (Hercules, CA, USA) Aminex

11065


HPX-87H column equipped with appropriate guard columns. Degassed 5 mM H2SO4 was used as the mobile phase and the column temperature was held at 60 °C. Glucose, xylose (plus galactose and mannose), and arabinose concentrations were determined for each liquid stream. Because the xylose, galactose, and mannose peaks cannot be separated using the HPX-87H column (Irick et al., 1988), any results reported for xylose also includes mannose and galactose. For grasses the galactose and mannose contents tend to be very low – in sum less than 1.5% of the total biomass (Biomass, 2006). 2.6. Pretreatment and post-washing The pretreatment and post-wash conditions for all six pretreatment methods are listed in Table 1. The basis for the mass balance was 100 kg dry biomass entering pretreatment (DBP) in stream A. Water use and catalyst loading for pretreatment are also reported on this basis. The experimental details on each pretreatment method are reported in Supplementary material (Annex 1). Except for AFEX and SAA where pretreatment conditions were chosen to limit hemicellulose degradation, pretreatment conditions were chosen to optimize sugar yields based on preliminary experiments with Dacotah switchgrass which are not detailed here. For pretreatments which produced a slurry, a solid–liquid separation was performed following pretreatment (except for SAA where washing is integrated with the pretreatment). For the whole slurry without post-washing, the entire slurry was weighed following pretreatment. While keeping the sample well-mixed, a sample was removed for enzymatic hydrolysis. The weight and moisture content of the remaining slurry was determined. The slurry was vacuum-filtered through Whatman No. 41 or No. 4 filter paper. The volume of the filtrate was determined and samples were taken for monomeric and oligomeric sugar analysis. The retained solids were washed with 500 mL of distilled water (20–30 °C) per 10 g of dry solids. The volume of the filtrate was determined and samples were taken for monomeric and oligomeric sugar analysis. The values for monomeric and oligomeric sugars solubilized by the pretreatment were calculated as the sum of the sugars from the initial slurry filtrate and the washed solids filtrate. The solids were dried overnight at 50 °C. The total wet weight and the %-solids content of the retained solids and filter paper were determined. Essentially the same method used for unwashed samples was employed for washed, pretreated solids undergoing enzymatic hydrolysis, but the methods differed in three details. Samples were not removed for enzymatic hydrolysis prior to filtration, the retained solids were washed with the wash water after the first filtration step as specified by the pretreatment method, and the washed solids were not dried prior to enzymatic hydrolysis. 2.7. Enzymatic hydrolysis Enzymatic hydrolysis was conducted in triplicate in 250 mL Erlenmeyer flasks. The pretreated biomass solids or slurry was

loaded at 1% glucan loading followed by enough distilled water to bring the total volume to roughly 80% of the final volume (120 mL), including the water already present in the biomass/ slurry. The pH was adjusted to between 4.5 and 5.0 using 50 mM sodium citrate buffer after which antibiotics were added to each flask (600 lL of 10 g/L tetracycline and 450 lL of 10 g/L cycloheximide). Distilled water was added to each flask to bring the final volume (after addition of the enzymes) to 150 mL. The flasks were tightly sealed with rubber stoppers and secured with tape before being placed in a shaking incubator which was set at 200 rpm and 50 °C. After the temperature in the flasks reached 50 °C, the flasks were removed and the enzymes were added based on the glucan content in the pre-washed, untreated biomass. SpezymeÒ CP (Batch: 301-05330-206; Genencor Division of Danisco US, Inc, NY, USA), with a protein content of 82 mg/mL and specific activity of 50 FPU/mL was loaded at 15 FPU g1 glucan in untreated biomass. b-Glucosidase (NovozymeÒ 188, Novozymes Corp.) with a protein content of 67 mg/mL and specific activity of 600 CBU/mL was loaded at 30 CBU g1 glucan in untreated biomass. The protein content of the enzymes was determined from total N analysis using the Dumas method for combustion of nitrogen to NOx following trichloroacetic acid (TCA) precipitation to remove non-protein nitrogen. Because the enzymes were added based on the glucan content in the untreated biomass, the enzyme loading based on the glucan in the pretreated biomass was variable for each feedstock. Table 2 shows the protein loading in terms of the both glucan content and the (glucan + xylan) content of each pretreated feedstock. After adding the enzymes, the flasks were re-sealed and placed in the incubator for 168 h. Following the incubation period, the flasks were removed from the incubator and the flask contents were transferred to disposable centrifuge tubes. The samples were centrifuged at 10,000 rpm for Table 2 Enzyme loadings for pretreated solids on the basis of the polymeric sugars in the pretreated and washed biomass (where applicable). Enzyme loadings for all samples were on the basis of 15 FPU Spezyme CP/g glucan in dry biomass entering pretreatment (DBP) and 30 CBU Novozyme 188/g glucan in DBP. Enzyme loading (mg protein/g polymeric sugar in pretreated [washed] biomass) Glucan

Glucan + Xylan

Spezyme CP

Novo 188

Spezyme CP

Novo 188

Washed AFEX DA Lime LHW SAA SO2

24.1 24.4 21.0 29.8 21.2 23.3

3.2 3.5 3.4 3.1 3.4 3.3

16.4 22.4 15.5 28.4 15.2 21.7

2.2 3.2 2.5 2.9 2.4 3.1

Unwashed AFEX LHW

22.9 29.6

3.0 3.0

14.4 27.8

1.9 2.8

Table 1 Pretreatment and post-wash comparison. Method

a b

Pretreatment

Post-wash

Temp (°C)

Time (min)

Catalyst

Water/solids loading

Catalyst loadinga

Water useb

AFEX DA Lime

150 140 120

30 40 240

2 g H2O/g DBP 10% solids (w/w) 15 g H2O/g DBP

1174 3000 4655

100 20–25 20–25

200 160 180

10 60 10

152 9 100 100 psi N/A 135 5

200 891 1468

LHW SAA SO2

Anhydrous NH3 1% H2SO4 Ca(OH)2 O2 Water 15% Aqueous NH3 5% SO2

663 765 895

3069 10,000 3000

80–90 20–25 20–25

Catalyst loading: kg/100 kg DBP. Water use: L/100 kg DBP.

15% solids (w/w) 10% solids (w/v) 10% solids (w/w)

Post-wash water useb

Water temp (°C)

11066


average of any replicates and included in Supplementary material (Tables S1–S8). Because of space limitations standard deviations are not included in the tables. Previous CAFI publications (Wyman et al., 2005a, 2009) have divided the mass balance into two stages, and for the purpose of comparison, process yields for this paper have been reported in the same manner. Stage 1 consists of the combined pretreatment and post-washing and Stage 2 is enzymatic hydrolysis (Fig. 1). No data are reported for streams 3, 7 and C when no post-wash step was performed. Gluco- (GO) and xylo-oligomers (XO) were reported in monomeric equivalents. Process sugar yields for each stage were calculated based on the pre-washed, dry biomass entering pretreatment (DBP) in stream A using the following equations with subscripts to indicate the stream. The equations are simplified by stating the ratio of the molecular weights of glucose to glucan (180/162) as (1/0.9) and xylose to xylan (150/132) as (1/0.88). Because sucrose that is present in the biomass can also contribute glucose to the liquid streams, this amount was determined by multiplying the amount of sucrose in the biomass by (180.2/342.3), the ratio of the molecular weights of glucose and sucrose.

30 min. Afterward, the volume and weight of the supernatant were recorded and samples were taken for monomeric (aliquot 1) and oligomeric sugar analysis (aliquot 2). The flasks were washed with 25 mL of distilled water to remove any residual solids. The wash liquid was transferred into the original centrifuge tubes containing the hydrolysis solids and the mixture was re-suspended, following which the samples were centrifuged a second time. The weight and volume of the supernatant was determined and samples were taken for monomeric (aliquot 3) and oligomeric (aliquot 4) sugar analysis. Aliquots 1 and 3 were heated at 100 °C for 15 min to denature the enzymes and cooled in a freezer for 15 min. After cooling, the samples were transferred into HPLC shell vials and stored at 20 °C until HPLC analysis. Acid hydrolysis was performed on aliquots 2 and 4 for oligomeric sugar determination. The values of the monomeric and (monomeric + oligomeric) sugars were calculated as the sum from aliquots 1 + 3 and aliquots 2 + 4, respectively. The centrifuge tubes with the solids were placed in the freezer overnight. The next day, the solids were removed from the tubes and allowed to thaw. The total weight and moisture content of the solids was recorded. The solids were dried at 50 °C overnight following which acid hydrolysis was performed to determine the solids composition.

Stage 1 Glucose Yield ð%Þ ¼

Glucose6 þ Glucose7 þ GO6 þ GO7 100% ðGlucanA =0:9Þ þ GlucoseA þ SucroseA ð180:2=342:3Þ ð1Þ

2.8. Mass balance calculations

Stage 1 Xylose Yield ð%Þ ¼

For the mass balance, an inventory of key system components, including water, was compiled for all streams when possible. The actual data generated for each pretreatment is reported as the A

Stage 1 Biomass & Water

Inputs

Water/ Catalyst

1

Outputs

5

8

Wash Stream

Soluble Sugars

E

H2SO4 & Water

Water

B

AFEX 6

A

C

Water

A

Post-Wash

6

Ammonia & Water 2 A

SAA

3 B

Liquid Hot Water

7

Slurry

C

Optional Post-Wash

6

Wash Stream

Pretreatment Liquor

Water

2 C

Lime

Wash Stream

F

3

A

7

Pretreatment Liquor

Water/ HCl

2

C

Post-Wash

6

Wash Stream

Lime, Water & Oxygen

3

Dilute Sulfuric Acid

7

Excess NH3 & H2O to Recycle

Insoluble Residue

Water

2

Optional Post-Wash

Solids

9

7

3

A

Enzymatic Hydrolysis

Post-Wash

6

2

D

C

Pretreatment Liquor

Ammonia & Water

C

4

3 B

Pretreatment

Soluble Sugars

B

Stage 2 Enzymes,Water, Buffer and Antibiotics

Water/ Chemicals

2 A

Pre-Wash & Drying

Xylose6 þ Xylose7 þ XO6 þ XO7 100% ðXylanA =0:88Þ ð2Þ

7 Wash Stream

Pretreatment Liquor

G SO2

Water

CounterCurrent Leaching 7 PT Liquor & Wash Stream

Water

2

3 C

2

Water 3

SO2

A

C

SO2 Thermal Impregnation Pretreatment 6 Excess SO2 to Recycle

6 Pretreatment Liquor

Post-Wash 7 Wash Stream

Fig. 1. Pretreatment input–output diagrams. (A) General pretreatment diagram. Inputs and outputs to the process are indicated by numbered streams while lettered streams indicate streams internal to the process. The pre-wash step and enzymatic hydrolysis step were common to all pretreatments. (B–F) Specific pretreatment diagrams for AFEX, DA, lime, LHW, SAA and SO2. Only the streams which were reported in the mass balance tables are included in the input–output diagrams (with the exception of the excess NH3 and H2O stream for AFEX (6) and the excess SO2 stream in SO2 pretreatment (6) which were not reported due to difficulties in measuring gas stream data). AFEX and LHW pretreatments reported two sets of data, one set with a post-wash step and one set without.

11067


Stage 2 Glucose Yield ð%Þ

sucrose should be included in the material balance. The range in glucan content of the pre-washed switchgrass which was reported by each university was 2.7 kg/100 DBP and the range in xylan content was 3.9 kg/100 DBP. The large range in results is likely indicative of slight differences in feedstock, equipment, and processes between universities. Pretreatment and post-wash conditions were highly variable between the different methods (Table 1). The pretreatments can be grouped based on their temperature–time combinations: high-temp/short-time (LHW and SO2); moderate-temp/moderatetime (AFEX, DA, and SAA); and low-temp/long-time (lime). Water use by the pretreatments was lowest for AFEX, highest for lime, with the other pretreatments within a similar range. Catalyst use also varied, with no use by LHW which relies on hydrothermal breakdown of the cell wall structure, a small amount used by DA and SO2, and the greatest amounts used by the alkaline pretreatments (AFEX, lime, and SAA). Because of this high use, catalyst recycle is considered a necessary part of these processes and would add additional capital cost to the alkaline pretreatment systems (Eggeman and Elander, 2005). Solids recovery following pretreatment with or without postwashing was around 60% for most of the methods (Table 4), except for lime and washed AFEX which were 10% and 20% higher, respectively, and unwashed AFEX which retained all of the biomass with a small increase due to ammonia incorporation. The composition of the LHW pretreated biomass was not strongly impacted by washing, which is not surprising given the similarity in the solids recovery between the washed and unwashed biomass. However, for AFEX pretreatment the relative proportion of glucan and lignin increased as hemicellulose was removed. The composition of the washed and pretreated solids from the acidic pretreatments (DA, SO2 and LHW) had higher glucan and lignin contents due to greater solubilization of xylan during Stage 1 (Fig. 2B) and the alkaline pretreatments (lime, SAA) had a higher glucan content due to hemicellulose and lignin removal (Fig. 2C).

Glucose8 þ GO8 ¼ 100% ðGlucanA =0:9Þ þ GlucoseA þ SucroseA ð180:2=342:3Þ ð3Þ

Stage 2 Xylose Yield ð%Þ ¼

Xylose8 þ XO8 100% ðXylanA =0:88Þ

ð4Þ

The overall mass closure for glucose and xylose was based on the process from pretreatment through enzymatic hydrolysis (pre-washing was not included). Mass closure was calculated as the soluble sugars (monomers and oligomers as monomeric equivalents) in the liquid streams (6, 7, and 8) plus the glucan or xylan in the hydrolysis solids (stream 9 – converted to monomeric equivalents), divided by the amount of polymeric sugars (as monomeric equivalents) and soluble sugar in the dry biomass in stream A. Lignin composition is difficult to determine for liquid streams, therefore the amount of removed lignin was calculated as the difference between the lignin in the solid residue entering and leaving the stage (either Stage 1 or Stage 2), divided by the lignin in stream A. 3. Results and discussion 3.1. Pre-wash and stage 1 – pretreatment and post-wash A pre-wash step may be a desirable step for feedstocks which have a high soluble sugar content that could mask the solubilization of cell wall sugars. However, Dacotah switchgrass is a mature grass sample, and the initial soluble sugar content was quite low (Table 3) so the amount of glucose detected in the hydrolyzed wash stream ranged from only 0.4 to 1.1 kg glucose per 100 kg DBP. For similarly mature samples, a pre-wash step may not be necessary; however, even a small amount of soluble glucose or

Table 3 Composition of unwashed and pre-washed Dacotah switchgrass (% of total DM). Component

Glucan Xylan Arabinan Acid-Insoluble Lignin Sucrose Othera

Unwashed

35.0 21.8 3.5 21.4 1.5 15.6

Pre-washed Dacotah Switchgrass AFEX

DA

Lime

LHW

SAA

SO2

37.1 25.5 3.0 23.4 – 11.0

36.5 22.7 3.2 20.7 – 16.9

37.2 23.7 2.5 20.8 – 15.7

35.6 22.6 3.1 22.8 – 15.9

34.8 22.1 3.4 21.1 – 18.6

36.5 22.7 3.2 20.7 – 16.9

a ‘‘Other’’ for the unwashed biomass includes extractives, ash and acetyl, as determined and provided by Ceres. ‘‘Other’’ for the pre-washed biomass was calculated as the difference between the sum of the listed values and 100%.

Table 4 Solids yields and composition following pretreatment and post-washing (where applicable) for the different pretreatment technologies. Component

Untreateda

Post-Washed AFEX

Unwashed DA

Lime

LHW

SAA

SO2

AFEX

LHW b

Solids recovery (kg/100 kg DBP)

–

83.3

60.4

74.4

59.0

64.2

60.5

101.3

Pretreated solids composition (%) Glucan Xylan Arabinan Lignin

36.3 23.2 3.1 21.6

45.4 21.1 3.1 23.3

52.3 4.5 0.0 29.5

48.4 17.3 2.0 12.8

50.0 2.5 0.0 31.5

53.7 21.2 2.3 13.4

54.5 4.1 0.0 26.8

39.2 23.3 2.8 20.2

61.3c 48.5 3.1 0.0 32.3

a The value for the untreated biomass is provided for the purpose of comparison and was calculated as the average of the composition values provided by all of the universities for the pre-washed biomass (Table 3). b The increase in the solids content for AFEX pretreated biomass is due to the incorporation of nitrogen into the biomass during the pretreatment via reactions with ammonia. c The solids recovery for the unwashed LHW whole slurry represents the solids fraction of the pretreatment slurry, all of which goes into enzymatic hydrolysis.

11068


A

Glucose Yield

(% solubilized of total available in washed, untreated biomass)

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% No Wash Wash

Glu GO

No Wash Wash

Stage 2 Stage 1

AFEX

DA

Lime

LHW

SAA

SO2

B

(Fig. 2B). As the pH decreased from a strongly acidic pH (DA and SO2), to a pH closer to neutral (LHW), to an alkaline pH (AFEX, lime, and SAA), the total amount of solubilized xylan decreased and the proportion of the solubilized sugars that were in oligomeric form increased (Table 6). It is known that as pretreatment severity increases (due to increased temperature, time and/or decreased pH) and more xylan is removed from the biomass, the solubilized xylo-oligomers are simultaneously deconstructed from higher to lower degrees of polymerization (DP), eventually resulting in monomeric xylose and degradation products (Kabel et al., 2007). Lignin solubilization follows the opposite trend, as a greater amount was removed at an alkaline pH (Fig. 2C). This pattern of lignin and hemicellulose solubilization with respect to pretreatment pH has been reported elsewhere (Pedersen and Meyer, 2010; Wyman et al., 2005a, 2009). This pattern is represented by a simple model in Fig. 3. For simplicity, this model does not show changes which may occur to cellulose with respect to pH.

Xylose Yield

(% solubilized of total available in washed, untreated biomass)

100%

3.2. Stage 2 – enzymatic hydrolysis

90% 80% 70% 60% 50% 40% 30% 20% 10% 0%

Xyl XO

No Wash Wash

No Wash Wash

Stage 2 Stage 1

AFEX

DA

Lime

LHW

SAA

SO2

SAA

SO2

C Klason Lignin Removal

(% removed of total available in washed, untreated biomass)

100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Lignin

No Wash Wash

No Wash Wash

Stage 2 Stage 1

AFEX

DA

Lime

LHW

Fig. 2. Glucose and xylose yields and Klason lignin removal from Dacotah switchgrass for Stage 1, Stage 2 and the overall process. (A) Glucose Yield, (B) Xylose Yield, (C) Klason Lignin Removal. Yields include all solubilized monomers and oligomers (in monomeric equivalents) and are expressed as a percentage of the sugar present in the pre-washed, untreated dry biomass. Standard deviations were not reported for DA or SO2, lime (except for Stage 1 lignin), or LHW (lignin Stages 1 and 2). The Stage 1 lignin values for the unwashed LHW and AFEX material is the lignin that has become more readily removed during the acid hydrolysis quantification method. This material, while made more soluble during Stage 1, is actually removed into the enzymatic hydrolysate during Stage 2. Glu, glucose; GO, glucooligomers; Xyl, xylose; XO, xylo-oligomers.

Very little glucan was solubilized during any of the pretreatments (Table 5). The only appreciable amounts solubilized were by DA, SO2, and LHW pretreatments, but in all cases this was 67% of the total glucose in the untreated biomass. Following pretreatment and post-washing, more than two-thirds of the xylan was removed from biomass pretreated via DA, LHW, and SO2, while less than one-third was released from AFEX, lime, and SAA

For all pretreatments, the majority of the glucose was solubilized during Stage 2 – enzymatic hydrolysis (Table 5 and Fig. 2A). Of the pretreatments, only the lime and SAA enzymatic hydrolysates contained any measurable amounts of gluco-oligomers. This amount was fairly low and may indicate either some inadequacy in the enzymes used or some error with respect to the oligomeric sugar quantification. The hydrolysate from unwashed and washed AFEX, unwashed LHW, and SAA all contained large amounts of oligomeric xylose (Table 6 and Fig. 2B). This indicates that at the enzyme loadings used, Spezyme CP and Novo188 were not able to adequately break down all of the hemicellulose, which is not unexpected as these enzymes have been shown to possess low hemicellulase activity and are particularly slow at degrading xylo-oligomers (Qing and Wyman, 2011). The addition of hemicellulases to the enzyme mixture has been shown to increase glucose yields from pretreated materials that have higher hemicellulose content, such as those produced by AFEX and ammonia recycle percolation (ARP), the precursor to SAA (Kumar and Wyman, 2009b). In another study on AFEX pretreated switchgrass, supplementation with hemicellulases was necessary to achieve the highest sugar yields during enzymatic hydrolysis (Bals et al., 2010). A more optimal enzyme mixture could increase the release of sugars from these materials. Post-washing of AFEX-treated and LHW-treated biomass, which solubilized roughly 20–30% of the lignin as well as large quantities of xylo-oligomers (30–50% of the total xylose), had a marked effect on glucose yields, increasing them by around 10–15%. This increase could be due to a number of reasons including increasing enzyme accessibility to the cell wall structure and removal of enzyme inhibitors such as lignin-based compounds, sugar degradation products, xylose, and xylo-oligomers. However, the reason for the increase in yields is most likely not due to the removal of low molecular weight lignin-based inhibitors. At low solids loadings such as those used here, these compounds have not been shown to strongly inhibit enzymes during enzymatic hydrolysis (Hodge et al., 2008). Another possibility is that the hot water washing removed inhibitory xylo-oligomers (Kumar and Wyman, 2009a; Qing et al., 2010), and additionally improved the accessibility of the biomass to the enzymes by removing additional biomass components. None of the pretreatment methods was able to solubilize over 90% of the available glucose or xylose into oligomeric and monomeric form. However, it is apparent that either the combination of low temperature/long residence time or high temperature/short residence time is important for releasing glucose from the substrate. The glucose yields were highest (>80%) for the

11069


Table 5 Glucose solubilization at each process stage expressed as the amount of glucose released in terms of the amount present in pre-washed, untreated dry biomass (kg/100 kg DBP). Stage 1

a b c d e

Stage 2

Pretreatment

Post-wash

Glca

GOb

Glca

GOb

AFEX No Washc Washd

N/A N/A

N/A N/A

N/A 0.0

N/A 0.5

DA Washd

2.0

0.2

0.5

Lime Washd

0.1

0.1

LHW No washc Washd

N/A 0.2

SAAe Washd SO2 Washd

Stage 1 total

Overall solubilization

Enzymatic hydrolysis Glca

GOb

N/A 0.5

26.4 30.9

0.0 0.0

0.1

2.8

27.7

0.0

0.4

0.6

N/A 2.5

N/A 0.0

N/A 0.0

–

–

0.0

0.8

0.8

0.2

Glucose mass closure (%)

Stage 2 total Glca

GOb

Total

26.4 30.9

26.4 30.9

0.0 0.5

26.4 31.4

100 108

0.0

27.7

30.2

0.3

30.5

92

33.4

2.0

35.4

33.5

2.5

36.0

108

N/A 2.7

28.8 31.0

0.1 0.0

28.9 31.0

28.8 31.2

0.1 2.5

28.9 33.7

99 95

0.1

0.1

25.3

0.8

26.1

25.3

0.9

26.2

94

0.2

2.0

31.7

0.0

31.7

32.7

1.0

33.7

91

Glc, solubilized glucose. GO, solubilized gluco-oligomers, reported in monomeric equivalents. No wash, no post-wash step following pretreatment. Wash, post-wash step following pretreatment. SAA pretreatment liquor was not sampled – the post-wash value is combined pretreatment liquor and post-wash liquid.

Table 6 Xylose solubilization at each process stage expressed as the amount of xylose released in terms of the amount present in pre-washed, untreated dry biomass (kg/100 kg DBP). Stage 1

a b c d e

Stage 2

Pretreatment

Post-Wash

Xyla

XOb

Xyla

XOb

AFEX No Washc Washd

N/A N/A

N/A N/A

N/A 0.0

N/A 7.3

DA Washd

14.6

0.9

3.5

Lime Washd

0.0

7.2

LHW No Washc Washd

N/A 3.0

SAAe Washd SO2 Washd

Stage 1 total

Overall solubilization

Enzymatic hydrolysis Xyla

XOb

N/A 7.3

19.1 14.8

3.5 2.0

0.2

19.2

2.2

0.0

1.7

8.9

N/A 7.0

N/A 2.7

N/A 4.3

-

-

0.6

14.6

0.8

3.2

Xylose mass closure (%)

Stage 2 total Xyla

XOb

22.6 16.8

19.1 14.8

3.5 9.3

22.6 24.1

95 98

0.0

2.2

20.3

1.1

21.4

86

14.2

0.5

14.7

14.2

9.4

23.6

97

N/A 17.0

17.9 2.8

2.7 0.7

20.6 3.5

17.9 8.5

2.7 12.0

20.6 20.5

87 81

5.7

6.2

7.2

4.5

11.7

7.7

10.2

17.9

80

0.2

18.8

2.1

0.0

2.1

19.9

1.0

20.9

84

Total

Xyl, solubilized xylose. XO, solubilized xylo-oligomers, reported in monomeric equivalents. No wash, no post-wash step following pretreatment. Wash, POST-wash step following pretreatment. SAA pretreatment liquor was not sampled – the post-wash value is combined pretreatment liquor and post-wash liquid.

pretreatments which operated at either of these temperature/time combinations (lime, washed LHW and SO2). The lowest yields (83%) from lime, washed AFEX, and DA, although all of the pretreatments solubilized more than 80% of the total xylose. While these values may seem low, it is important to keep in mind that the Dacotah switchgrass used for these experiments was harvested in the spring after over-wintering on the field, which can have a strong negative impact on the digestibility of herbaceous biomass (Le Ngoc Huyen et al., 2010). The glucose mass closure values for all pretreatments were 100 ± 9%. The extreme values were for DA and SO2 (low), and washed AFEX and lime (high), and may be due to compounded errors within the method, particularly with respect to acid hydrolysis

used for composition data. It is possible that yields may be underestimated for the samples with low mass closure and overestimated for the samples with high mass closure. The xylose mass closure was between 80% and 98%, slightly lower than calculated for glucose. This may be due to high temperature degradation of xylose and xylo-oligomers into other compounds. While it is less of an issue for alkaline pretreatments, degradation of xylose (Kabel et al., 2007; Lloyd and Wyman, 2005) and production of inhibitory furans, such as furfural from pentoses and 5-hydroxymethyl furfural (5-HMF) from hexoses, can be significant for pretreatments such as DA, SO2, and LHW that operate at low pH and high temperatures (Chen et al., 2007; Du et al., 2010; Kabel et al., 2007). For the 5 min residence time LHW pretreatment without post-washing, the furfural concentration in the pretreatment liquor was

11070

A

B


Hemicellulose

Cellulose

Low pH

Lignin

High pH

yields. Washing improved glucose release from LHW and AFEX pretreated switchgrass. Pretreatment pH effects the solubilization of biomass components. As pH decreases, solubilized lignin decreases, while total solubilized xylan and released monomeric xylose increases. Differences in pretreatment solubilization impact other processing areas and the process economics. Low temperature-long time or high temperature-short time pretreatment is necessary for high glucose release from late-harvest Dacotah switchgrass, but high temperatures may cause xylose degradation. Acknowledgements

Hemicellulose Oligomeric Sugars Hemicellulose Monomeric Sugars

Lignin and Derived Compounds

Fig. 3. Cell wall model showing the general effect of pH on solubilization of hemicellulose and lignin. (A) Untreated cell wall and (B) cell wall during pretreatment. Cellulose can also be degraded under extremely acidic conditions; however that is not portrayed in this diagram. Designed based on figures from Mosier et al. (2005) and Pedersen and Meyer (2010).

0.23 g/L. For the 10 min residence time pretreatment, the concentration increased to 2.75 g/L – an amount equal to the degradation of 7.4% of the xylose initially present in the biomass. When this amount is included in the mass balance as an output, the xylose mass closure for LHW pretreatment increased to 95%. Compared to furfural, the amount of HMF produced after the 10 min residence time was very low (0.23 g/L) and accounted for only 0.4% of the glucose initially present in the untreated biomass. For this project, quantification of furfural and HMF for all of the pretreatments was unintentionally overlooked; however, in future work on pretreatment mass balances, particularly those at low pH and high temperatures, at least these two compounds should be included due to their potential impacts on the glucose and xylose mass closure. Apart from the differences in sugar yields, the differences in solubility profiles for the different pretreatments could have a large impact on the process economics. One issue with respect to pretreatment wash streams is whether it is economically worthwhile or possible to recover the hemicellulose sugars for use later in the process. Pretreatments which solubilize the hemicellulose sugars as oligomers and wish to recover them for fermentation will either require a subsequent step to convert them to a monomeric form or use a micro-organism which can utilize oligomeric sugars. However for those pretreatments that retain the hemicellulose, there are also challenges associated with co-fermentation of a mixture of glucose and xylose (Jin et al., 2010). The solubilization of lignin is also important as the lignin solids are generally modeled as the source of biorefinery electricity and energy for steam production (Aden and Foust, 2009; Eggeman and Elander, 2005). Greater solubilization of lignin could lead to a decreased energy production and increased wastewater treatment costs. However, when conducting simultaneous saccharification and fermentation, retention of the lignin with the biomass can have a negative effect on the fermentation microbes (Jin et al., 2010). When comparing pretreatments, the overall sugar yields are just one factor that should be considered and in the end, there are many tradeoffs with respect to the different methods.

4. Conclusions All pretreatments solubilized > 2/3 of the available glucan and xylan. Lime, washed LHW, and SO2 achieved > 83% total glucose yields. Lime, washed AFEX, and DA achieved > 83% total xylose

This research was funded under the Office of the Biomass Program of the United States Department of Energy (Contract: DEFG36-07GO17102). We would like to acknowledge the many undergraduate and graduate students, post doctoral candidates, and technicians at their respective institutions for their vital role in obtaining and compiling this information. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.biortech.2011.04.002. References Aden, A., Foust, T., 2009. Technoeconomic analysis of the dilute sulfuric acid and enzymatic hydrolysis process for the conversion of corn stover to ethanol. Cellulose 16 (4), 535–545. Alvira, P., Tomás-Pejó, E., Ballesteros, M., Negro, M.J., 2010. Pretreatment technologies for an efficient bioethanol production process based on enzymatic hydrolysis: a review. Bioresour. Technol. 101 (13), 4851–4861. Asif, M., Muneer, T., 2007. Energy supply, its demand and security issues for developed and emerging economies. Renew. Sust. Energy Rev. 11 (7), 1388– 1413. Bals, B., Rogers, C., Jin, M., Balan, V., Dale, B., 2010. Evaluation of ammonia fibre expansion (AFEX) pretreatment for enzymatic hydrolysis of switchgrass harvested in different seasons and locations. Biotechnol. Biofuels 3 (1), 1. Biomass feedstock composition and property database, 25 January 2006. US Department of Energy: Biomass Program. (accessed: 12 March 2010). Chen, S.-F., Mowery, R.A., Chambliss, C.K., van Walsum, G.P., 2007. Pseudo reaction kinetics of organic degradation products in dilute-acid-catalyzed corn stover pretreatment hydrolysates. Biotechnol. Bioeng. 98 (6), 1135–1145. da Costa Sousa, L., Chundawat, S.P.S., Balan, V., Dale, B.E., 2009. ‘Cradle-to-grave’ assessment of existing lignocellulose pretreatment technologies. Curr. Opin. Biotechnol. 20 (3), 339–347. Du, B., Sharma, L.N., Becker, C., Chen, S.-F., Mowery, R.A., van Walsum, G.P., Chambliss, C.K., 2010. Effect of varying feedstock-pretreatment chemistry combinations on the formation and accumulation of potentially inhibitory degradation products in biomass hydrolysates. Biotechnol. Bioeng. 107 (3), 430–440. Eggeman, T., Elander, R.T., 2005. Process and economic analysis of pretreatment technologies. Bioresour. Technol. 96 (18), 2019–2025. Hatzis, C., Riley, C., Philippidis, G., 1996. Detailed material balance and ethanol yield calculations for the biomass-to-ethanol conversion process. Appl. Biochem. Biotechnol. 57–58 (1), 443–459. Himmelsbach, J.N., Isci, A., Raman, D.R., Anex, R.P., 2009. Design and testing of a pilot-scale aqueous ammonia soaking biomass pretreatment system. Appl. Eng. Agric. 25 (6), 953–959. Hodge, D.B., Karim, M.N., Schell, D.J., McMillan, J.D., 2008. Soluble and insoluble solids contributions to high-solids enzymatic hydrolysis of lignocellulose. Bioresour. Technol. 99 (18), 8940–8948. Irick, T., West, K., Brownell, H., Schwald, W., Saddler, J., 1988. Comparison of colorimetric and HPLC techniques for quantitating the carbohydrate components of steam-treated wood. Appl. Biochem. Biotechnol. 17 (1), 137– 149. Jin, M., Lau, M.W., Balan, V., Dale, B.E., 2010. Two-step SSCF to convert AFEX-treated switchgrass to ethanol using commercial enzymes and Saccharomyces cerevisiae 424A(LNH-ST). Bioresour. Technol. 101 (21), 8171–8178. Kabel, M.A., Bos, G., Zeevalking, J., Voragen, A.G.J., Schols, H.A., 2007. Effect of pretreatment severity on xylan solubility and enzymatic breakdown of the remaining cellulose from wheat straw. Bioresour. Technol. 98 (10), 2034–2042. Keshwani, D.R., Cheng, J.J., 2009. Switchgrass for bioethanol and other value-added applications: a review. Bioresour. Technol. 100 (4), 1515–1523. Kumar, R., Wyman, C.E., 2009a. Effect of enzyme supplementation at moderate cellulase loadings on initial glucose and xylose release from corn stover solids pretreated by leading technologies. Biotechnol. Bioeng. 102 (2), 457–467.

R.J. Garlock et al. / Bioresource Technology 102 (2011) 11063–11071 Kumar, R., Wyman, C.E., 2009b. Effect of xylanase supplementation of cellulase on digestion of corn stover solids prepared by leading pretreatment technologies. Bioresour. Technol. 100 (18), 4203–4213. Le Ngoc Huyen, T., Rémond, C., Dheilly, R.M., Chabbert, B., 2010. Effect of harvesting date on the composition and saccharification of Miscanthus x giganteus. Bioresour. Technol. 101 (21), 8224–8231. Lloyd, T.A., Wyman, C.E., 2005. Combined sugar yields for dilute sulfuric acid pretreatment of corn stover followed by enzymatic hydrolysis of the remaining solids. Bioresour. Technol. 96 (18), 1967–1977. Mosier, N., Wyman, C., Dale, B., Elander, R., Lee, Y.Y., Holtzapple, M., Ladisch, M., 2005. Features of promising technologies for pretreatment of lignocellulosic biomass. Bioresour. Technol. 96 (6), 673–686. Pedersen, M., Meyer, A.S., 2010. Lignocellulose pretreatment severity – relating pH to biomatrix opening. New Biotechnol. 27 (6), 739–750. Qing, Q., Wyman, C.E., 2011. Hydrolysis of different chain length xylooliogmers by cellulase and hemicellulase. Bioresour. Technol. 102 (2), 1359–1366. Qing, Q., Yang, B., Wyman, C.E., 2010. Xylooligomers are strong inhibitors of cellulose hydrolysis by enzymes. Bioresour. Technol. 101 (24), 9624–9630. Schell, D., Sáez, J., Hamilton, J., Tholudur, A., McMillan, J., 2002. Use of measurement uncertainty analysis to assess accuracy of carbon mass balance closure for a cellulase production process. Appl. Biochem. Biotechnol. 98–100 (1), 509–523. Sluiter, A., Hames, B., Ruiz, R., Scarlata, C., Sluiter, J., Templeton, D., 2008. Determination of sugars, byproducts, and degradation products in liquid fraction process samples. Laboratory Analytical Procedures (LAPs). National Renewable Energy Laboratory, Golden, CO. Sluiter, J.B., Ruiz, R.O., Scarlata, C.J., Sluiter, A.D., Templeton, D.W., 2010. Compositional analysis of lignocellulosic feedstocks. 1. Review and description of methods. J. Agric. Food Chem. 58 (16), 9043–9053.

11071

Wyman, C.E., Dale, B.E., Elander, R.T., Holtzapple, M., Ladisch, M.R., Lee, Y.Y., 2005a. Comparative sugar recovery data from laboratory scale application of leading pretreatment technologies to corn stover. Bioresour. Technol. 96 (18), 2026– 2032. Wyman, C.E., Dale, B.E., Elander, R.T., Holtzapple, M., Ladisch, M.R., Lee, Y.Y., 2005b. Coordinated development of leading biomass pretreatment technologies. Bioresour. Technol. 96 (18), 1959–1966. Wyman, C.E., Dale, B.E., Elander, R.T., Holtzapple, M., Ladisch, M.R., Lee, Y.Y., Mitchinson, C., Saddler, J.N., 2009. Comparative sugar recovery and fermentation data following pretreatment of poplar wood by leading technologies. Biotechnol. Progr. 25 (2), 333–339. Xu, J., Cheng, J.J., Sharma-Shivappa, R.R., Burns, J.C., 2010. Lime pretreatment of switchgrass at mild temperatures for ethanol production. Bioresour. Technol. 101 (8), 2900–2903. Yang, B., Wyman, C.E., 2008. Pretreatment: the key to unlocking low-cost cellulosic ethanol. Biofuel. Bioprod. Bior. 2 (1), 26–40. Yang, Y., Sharma-Shivappa, R., Burns, J.C., Cheng, J.J., 2009. Dilute acid pretreatment of oven-dried switchgrass germplasms for bioethanol production. Energy Fuels 23 (7), 3759–3766. Zhang, Y.-H.P., Berson, E., Sarkanen, S., Dale, B., 2009. Sessions 3 and 8: pretreatment and biomass recalcitrance: fundamentals and progress. Appl. Biochem. Biotechnol. 153 (1), 80–83. Zhu, J., Zhu, W., Obryan, P., Dien, B., Tian, S., Gleisner, R., Pan, X., 2010. Ethanol production from SPORL-pretreated lodgepole pine: preliminary evaluation of mass balance and process energy efficiency. Appl. Microbiol. Biotechnol. 86 (5), 1355–1365.


Process and economic analysis of pretreatment technologies Tim Eggeman

a,*

, Richard T. Elander

b

a

b

Neoterics International, 2319 S. Ellis Ct., Lakewood, CO 80228, USA National Renewable Energy Laboratory, 1617 Cole Blvd., Golden, CO 80401, USA Available online 10 March 2005

Abstract Five pretreatment processes (dilute acid, hot water, ammonia fiber explosion (AFEX), ammonia recycle percolation (ARP), and lime) for the liberation of sugars from corn stover are compared on a consistent basis. Each pretreatment process model was embedded in a full bioethanol facility model so that systematic effects for variations in pretreatment were accounted in the overall process. Economic drivers influenced by pretreatment are yield of both five and six carbon sugars, solids concentration, enzyme loading and hemicellulase activity. All of the designs considered were projected to be capital intensive. Low cost pretreatment reactors in some pretreatment processes are often counterbalanced by higher costs associated with pretreatment catalyst recovery or higher costs for ethanol product recovery. The result is little differentiation between the projected economic performances of the pretreatment options. Additional process performance data, especially involving the identification of optimal enzyme blends for each pretreatment approach and conditioning requirements of hydrolyzates at process-relevant sugar concentrations resulting from each pretreatment may lead to greater differentiation in projected process economics. 2005 Elsevier Ltd. All rights reserved. Keywords: Pretreatment; Economics; Bioethanol

1. Introduction Process engineering and economic analysis for the Biomass Refining Consortium for Applied Fundamentals and Innovation (CAFI) USDA Initiative for Future Agriculture and Food Systems (IFAFS) program were conducted via support from the US Department of EnergyÕs Office of the Biomass Program. The material balance and technoeconomic models were developed early in the USDA IFAFS project for each pretreatment technology in collaboration with each CAFI researcher. Initially, these models were populated with either assumptions or data generated in previous work, if applicable. The models were updated throughout the

*

Corresponding author. Tel.: +1 303 358 6390. E-mail address: [email protected] (T. Eggeman).


course of the IFAFS project as process performance data was generated and thus provided important information for guiding the selection of experimental conditions and the interpretation of experimental results. A series of sensitivity cases were also developed for each pretreatment approach to identify the economic impact of sugar and ethanol yields, enzyme loading and cost, capital costs, and other relevant parameters. The details of the various sensitivity analyses are not covered in this paper, but have been provided to each CAFI researcher. The data generated in the IFAFS project were primarily focused upon determining glucose and xylose sugar yields upon pretreatment and enzymatic hydrolysis using a standard cellulase loading. Less emphasis was placed on downstream process characterization and optimization, such as identifying improved enzyme preparations for each pretreatment or determination of conditioning requirements on hydrolyzates resulting

2020

T. Eggeman, R.T. Elander / Bioresource Technology 96 (2005) 2019–2025

from each pretreatment to allow for efficient fermentation at process-relevant sugar concentrations. Therefore, this paper is focused on identifying the process economic impact of the different pretreatment approaches as related to capital and operating cost investment and baseline glucose and xylose sugar yields from each pretreatment.

2. Methods An ASPEN Plus 10 (Aspen Technology, Inc., Cambridge, MA) simulation model was assembled for each pretreatment process using performance data supplied by each CAFI researcher. Detailed descriptions for each pretreatment process can be found in this volume (Kim and Lee, 2005; Teymouri et al., 2005; Mosier et al., 2005; Lloyd and Wyman, 2005; Kim and Holtzapple, 2005). Appropriate pretreatment reactor design and materials of construction for each pretreatment technology were developed that are consistent with the pretreatment chemistry, corrosion potential, feedstock solids loading, and residence time. Any necessary pretreatment catalyst recovery and recycle equipment were also included in the process design. The designs are best characterized as conceptual since there is still quite a bit of uncertainty in process performance and optimal pretreatment process flowsheet configuration. This is especially true for any necessary pretreatment catalyst recovery and recycle systems. Two additional pretreatment cases were considered. The first, called the ‘‘No Pretreatment’’ case, is the simple case in which the only action performed in pretreatment is dilution of the biomass feedstock to 20 wt% solids prior to enzymatic hydrolysis. Pretreatment related capital and operating costs were assumed to be zero. All yield in the no pretreatment case is attributable to enzymatic action on the native biomass (Lee, 2004). The second additional case is called ‘‘Ideal Pretreatment’’. Again the biomass feed is diluted to 20 wt% prior to hydrolysis, and zero capital and operating costs were assumed for pretreatment. However, in this case, the yield of glucose and xylose sugars after enzymatic hydrolysis were assumed to be 100% of theoretical. Each pretreatment model was then inserted into an Aspen simulation of a full bioethanol production facility, shown in Fig. 1. The 2001 NREL process engineering design case (Aden and Ruth, 2001), less the pretreatment section, was used as the template for the full bioethanol facility. The model assumes a 2000 metric ton (dry) per day corn stover feed rate, which corresponds to nominally 50 MMgal/yr of ethanol production for the assumptions used in the models. Some variation from this nominal ethanol production rate is caused by yield differences among the different pretreatment approaches. Simultaneous saccharification

Enzymes

Stover

Feed Handling

Pretreatment

CO2

Hydrolysis + Fermentation

Water

Recovery

EtOH

Syrup + Solids Chemicals

Water

Boiler + Generator

Steam Power

Fig. 1. Block flow diagram for bioethanol facility.

and fermentation (SSF) is assumed with an enzyme loading of 15 FPU/g cellulose in untreated corn stover (58 mg protein/g cellulose in untreated corn stover). The 15 FPU/g cellulose loading was chosen as a standard loading throughout the experimental work performed by the CAFI group. Hydrolysis performance was assumed to be the same as the laboratory results using Spezyme CP (Lot 30100348-257) (Genencor International Inc., Rochester, NY). An unspecified organism that is capable of metabolizing both monomeric xylose and monomeric glucose is assumed for fermentation. The coproducts for most corn based ethanol facilities are animal feed ingredients. For example, corn dry mills typically produce DDGS while wet mills produce corn gluten feed, corn gluten meal, corn germ and other related coproducts. In contrast, for corn stover based facilities, the recovered syrup and solids has limited feed value, so the model assumes this material is burned and the heat released is used to raise process steam and electricity via a bottoming cycle. The economic model consists of four parts: Capital cost estimate—The capital cost estimate is a factored estimate. To generate the capital costs, the process model is used to establish the flows for each major piece of equipment, the equipment is then sized using standard engineering methods, and purchased costs are estimated using a combination of in-house methods and Questimate (Aspen Technologies, Inc., Cambridge, MA). The total fixed capital then built by using standard factors for both directs and indirects. Operating cost estimate—Variable operating costs are estimated using material balances from the process model. Corn stover pricing is assumed to be $35/metric ton (dry) and represents a target price in a future process for which improvements in the costs of corn stover collection over currently available methods have been achieved. Enzyme pricing is assumed such that the total contribution of enzymes to production costs is about $0.15/gal of ethanol with some variation depending upon actual ethanol yields resulting from the particular pretreatment approach. This enzyme price does not reflect current commercial enzyme prices but instead is a reasonable estimate of the contribution of enzymes to the operating costs for future lignocellulosic based bior-


efineries. Fixed operating costs are estimated from manpower, maintenance, insurance, etc. requirements of ethanol facilities of similar size. Revenue summary—Ethanol and electricity sales are the two revenue streams. Power generated in excess of plant needs is sold to the grid at an assumed price of $0.04/kW h. Discounted cash flow calculations—The discounted cash flow calculations assume 2.5 years of construction, 0.5 years of start-up and 20 years of operations. One hundred percent equity financing and no subsidy credits are assumed. Ethanol pricing is done on a rational pricing basis rather than a market pricing basis. In other words, this is a cost-plus type of analysis, so rather than comparing net present values we use minimum ethanol selling price (MESP) as a performance measure. Minimum ethanol selling price is defined as the ethanol sales price required for a zero net present value for the project when the cash flows are discounted at 10% real-after tax.

3. Results and discussion Table 1 compares the capital costs for each case. The pretreatment area direct fixed capital for the dilute acid, AFEX, ARP, and lime cases are roughly the same. The contribution of the pretreatment reactor dominates pretreatment area cost for the dilute acid case, whereas for AFEX, ARP and lime, other equipment items dominate, with the pretreatment reactor cost being significantly lower than for dilute acid. Much of this other equipment is related to recovery of the pretreatment catalyst, which is necessary in these processes because one-pass use of the catalyst is impractical. As previously mentioned, the design of the various catalyst recovery and recycle systems is very preliminary, which may lead to opportunities for development of more efficient recovery systems. Pretreatment direct fixed capital for hot water pretreatment is significantly lower than for the other cases. However, total capital for the hot water case is roughly in line with most of the other cases. This particular version of hot water pretreatment has limitations on the

2021

concentration of solids that can be processed during pretreatment. The result is a lower solids concentration in the feed to enzymatic hydrolysis, so all of the downstream equipment is larger for the hot water case to accommodate the increased water load. Total capital for the lime case is significantly lower than other cases. The energy balance for this case is significantly different. The fermentation residues are burned to calcine calcium carbonate, converting it to lime for recycle in pretreatment. The calciner also generates steam for the plant, however, the amount of excess heat available after meeting the calciner and plant steam requirement is not enough to justify installation of power generation equipment. The lime case does not generate electricity, thus the reduction in total capital. The last two columns of Table 1 compare yield and capital requirements per annual gallon of capacity. The no pretreatment case has extremely poor yield, giving a very high value for the total fixed capital per annual gallon of capacity. All of the actual pretreatment cases show higher yield and lower capital requirements per annual gallon of capacity as compared to the no pretreatment case. However, all of the cases including the ideal pretreatment case appear to be capital intensive. As a comparison, todayÕs new generation of ethanol plant based on corn dry milling technology have capital investment requirements of $1.00–1.50/gal of annual capacity (BBI International, 2003). While capital investment for a lignocellulose-to-ethanol plant may not need to be quite as low as for a corn dry mill due to the lower expected feedstock cost for a lignocellulose plant, significant capital investment improvements for processes based upon any of these pretreatment approaches are needed. Fig. 2 presents a breakdown of capital investment using the dilute acid case as an example. Total fixed capital includes both directs and indirects. The indirect costs are factored off the directs, so it is only necessary to examine the direct costs in more detail. The pretreatment, fermentation (including enzymatic hydrolysis), and recovery sections of the plant are responsible for slightly less than half of the total direct fixed capital. The steam and power system is responsible for about one-third of total direct fixed capital. With the exception

Table 1 Capital costs

Dilute acid Hot water AFEX ARP Lime No pretreatment Ideal pretreatment

Pretreatment direct fixed capital, $MM

Pretreatment breakdown, % Reactor/% other

Total fixed capital, $MM

Ethanol production, MMgal/yr

Total fixed capital, $/gal annual capacity

25.0 4.5 25.7 28.3 22.3 0 0

64/36 100/0 26/74 25/75 19/81 – –

208.6 200.9 211.5 210.9 163.6 200.3 162.5

56.1 44.0 56.8 46.3 48.9 9.0 64.7

3.72 4.57 3.72 4.56 3.35 22.26 2.51

2022


Total Fixed

Direct Fixed Recovery $21.2MM

Start-Up $19.0MM

Water Treatment $2.1MM

Project Contingency $3.8MM

Storage $1.5MM

Fermentation $16.0MM

Home Office& Construction Fee $32.0MM

Steam & Power $41.8MM

Pretreatment $25.0MM

Feed Handling $7.3MM

Field Expenses $25.7MM

Other OSBL $8.1MM

Other Utilities $5.1MM

Dilute Acid Direct Indirect Start-Up & Contingency Total Fixed Capital

$MM 128.1 57.7 22.8 208.6

Fig. 2. Breakdown of capital costs for dilute acid pretreatment case.

of the lime case discussed earlier, the models assume a circulating fluidized bed boiler is used to combust insoluble lignin-rich residues to generate high pressure steam (8.62 kPa = 1250 psig, 510 C = 950 F), which is let down across a condensing turbine system to produce electricity. Fig. 3 compares the plant level cash costs and MESP across the pretreatment cases using the fourth year of operation as the proof year. The no pretreatment case is not displayed since the cash costs ($2.43/gal) and MESP ($6.45/gal) would distort the graph.

The plant level cash cost is also the same as the lowest ethanol price at which the plant will stay operational, even though the plant would be losing money at these market conditions. As such, it defines the competitive position of the proposed facility within the existing ethanol market. In this analysis, cash cost is comprised by three components: net stover, other variable costs, and fixed costs without depreciation. Net stover, by analogy with the net corn concept used in corn processing, is defined as the cost of stover feedstock less the value of the electricity coproduct. Other variable costs accounts for

Fig. 3. Cash costs and MESP comparison.


2023

Table 2 Yields Pretreatment

Xylose yields, % of theoreticala

Glucose yields, % of theoreticala

After pretreatment

After enzymatic hydrolysis

After pretreatment

After enzymatic hydrolysis


90.2/89.7 50.8/7.3 0/0 47.2/0 24.3/0.8 0 –

95.6/95.1 81.8/38.3 92.7/77.6 88.3/41.1 75.3/51.8 8.5 100

8.0/7.5 4.5/2.0 0/0 1.4 1.6/0.5 0 –

85.1/84.6 90.5/88.0 95.9 90.1 92.4/91.3 15.7 100

a

Cumulative soluble sugars as (oligomers + monomers)/monomers. Single number = just monomers.

the cost of enzymes, chemicals, etc. in which the quantities required are tied to the plant production rate. Fixed costs include labor, maintenance, insurance, and other costs not tied to production rate. Projected cash costs range from $0.54/gal for the ideal pretreatment case to $1.05 for lime pretreatment. The projected cost for the lime case is higher than the others because this case imports electricity, giving a large net stover contribution. The MESP includes additional charges related to depreciation, income taxes and return on capital. The ideal pretreatment case has an MESP of $0.99/gal, while the other cases range from $1.34 to $1.67/gal. The gap between the ideal pretreatment and the other cases is measure of how much improvement could ideally be obtained by future R&D efforts focused just on pretreatment. The lime pretreatment case has zero income tax in the chosen proof year. The main reason for this is that the lime kiln, which also produces steam for the facility, was classified as a piece of process equipment rather than a power generation system for depreciation pur-

poses. The economic model assumes general process equipment is depreciated using the Modified Accelerated Cost Recovery System (MACRS) method with a seven year class life, while power systems are depreciated using the MACRS method with a 20 year class life. Classifying the lime kiln as a piece of process equipment gives a faster rate of depreciation, which in turn delays the start of income taxes in the cash flow calculations. A closer look at the models shows that MESP is sensitive to yield of ethanol from both five and six carbon sugars present in the starting biomass. Table 2 compares of both oligomeric and monomer xylose and glucose for the pretreatment. It is important underscore that the values shown previously in Fig. 3 assumed only conversion of monomer sugars to ethanol. Looking at the glucose data in Table 2 we see that after enzymatic hydrolysis, almost all of the soluble sugars are present in the monomeric form. However, the xylose data in Table 2 shows that as pH of the pretreatment increases, the amount of soluble xylose in the form of oligomers becomes significant. It is possible that

MESP, $/gal EtOH

1.75

1.50

1.25

1.00 Dilute Acid

Hot Water w/o Oligomer Credit

AFEX

ARP w/ Oligomer Credit

Fig. 4. Effect of oligomer credit.

Lime

2024


Table 3 Energy usage Pretreatment


Pretreatment conditions

Pretreatment effluent, wt% liquid

Fermentation beer, wt% of ethanol in liquid

Process steam usage

Electrical power

HP steam, kg/h

LP steam, kg/h

VLP steam, kg/h

Total generated, MWe

Process needs, MWe

1 wt% acid, 140 C 13.9 wt% insolubles, 180 C Stover:NH3:H2O = 1:1:0.6 (weight), 90 C Liquid loading = 3.185 g/g stover, 170 C Lime = 0.08 g as CaO/g stover, 55 C – –

66.8 83.7 61.2 74.5 80.0 80.0 80.0

5.07 3.09 4.62 5.06 3.16 0.92 5.86

14,663 34,281 0 112,330 0 0 0

83,377 146,580 154,980 56,972 160,540 59,697 71,230

16,426 48,374 31,932 11,469 42,265 51,477 20,160

42.1 42.2 35.6 38.0 0 88.8 45.2

14.0 13.5 17.4 11.9 14.4 13.9 32.4

an increase in the xylanase activity of the enzyme preparation used for hydrolysis could be done at little additional enzyme cost, but this has not been fully demonstrated. Fig. 4 shows the changes in the resulting MESPÕs under the assumption that all soluble xylose and glucose sugars, both monomeric and oligomeric, contribute to ethanol production at no additional cost than for the baseline cellulase loading. The result is that there is very little economic differentiation between the pretreatment options after customizing the enzyme formulations to the needs of the process in this manner. Solvent loading in pretreatment, whether the solvent is water or some other chemical, is an important model parameter since it affects the overall plant energy balance and capital costs for fermentation and downstream recovery equipment. Table 3 shows the solvent loadings assumed for each pretreatment, the concentration of ethanol in the beer, the plant steam usage, and the plant power balance. HP steam is supplied at 1317 kPa and 268.2 C; LP steam is supplied at 448 kPa and

163.5 C; VLP steam is supplied at 170 kPa and 115.2 C. Fig. 5 shows the total process steam usage is proportional to the solvent concentration in the pretreatment reactor effluent, not counting the special cases (i.e. no pretreatment and ideal pretreatment).

4. Conclusions The pretreatment processes were compared on a consistent basis. Each pretreatment process model was embedded in a full facility model so that systematic effects for variations in pretreatment were accounted in the overall process. Economic drivers influenced by pretreatment are yield of both five and six carbon sugars, solids concentration, enzyme loading and hemicellulase activity. All of the designs considered were projected to be capital intensive. Low cost pretreatment reactors are often counterbalanced by higher costs associated with pretreatment catalyst recovery or higher costs for ethanol

Fig. 5. Effect of solvent loading in pretreatment on process steam usage.


product recovery. The result is little differentiation between the projected economic performances of the pretreatment options. This is especially true when credit is taken for availability of the oligomer sugars generated in the non-acidic pretreatment processes. The designs generated during this study are best characterized as conceptual. Their accuracy is sufficient to guide research but should not be taken as a basis for an actual construction project. No differentiation was made for variations in the state of development between the pretreatments. It is not completely fair to made economic comparisons between the pretreatment options without additional financial modeling since there is a wide range in the current state of development. A real options analysis (Glantz, 2000) that uses the discounted cash flow results of this work as one of the inputs is one way to formally adjust for the differences in state of development. Real option analyses could also be formulated to formally handle other less tangible issues such as differences in process complexity and reliability, differing potential for creating environmental and safety issues, etc. Acknowledgements • The United States Department of Energy—Office of the Biomass Program. • The United States Department of Agriculture—Initiative for Future Agricultural and Food Systems/ Cooperative State Research, Education and Extension Service (Contract Number 00-52104-9663).

2025

• Biomass Refining Consortium for Applied Fundamentals and Innovation (CAFI).

References Aden, A., Ruth, M., 2001. Process Design Update: C Milestone Completion Report, NREL ID #: FY01-238, March 2001. Superseded by: Aden, A., Ruth, M., Ibsen, K., Jechura, J., Neeves, K., Sheehan, J., Wallace, B., Montague, L., Slayton, A., Lukas, J., Lignocellulosic Biomass to Ethanol Process Design and Economics Utilizing Co-Current Dilute Acid Prehydrolysis and Enzymatic Hydrolysis for Corn Stover, NREL/TP-510-32438, June 2002. BBI International, 2003. Ethanol Plant Development Handbook, 4th ed. BBI International. Glantz, M., 2000. Scientific Financial Management: Advances in Intelligence Capabilities for Corporate Valuation and Risk Assessment. American Management Association, New York. Kim, S., Holtzapple, M.T., 2005. Lime pretreatment and enzymatic hydrolysis of corn stover, this volume. Kim, T.H., Lee, Y.Y., 2005. Pretreatment and fractionation of corn stover by ammonia recycle percolation (ARP) process, this volume. Lee, Y.Y., 2004. Personal communication. Lloyd, T.A., Wyman, C.E., 2005. Combined sugar yields for dilute sulfuric acid pretreatment of corn stover followed by enzymatic hydrolysis of the remaining solids, this volume. Mosier, N., Hendrickson, R., Ho, N., Sedlak, M., Ladisch, M.R., 2005. Optimization of pH controlled liquid hot water pretreatment of corn stover, this volume. Teymouri, F., Laureano-Perez, L., Alizadeh, H., Dale, B.E., 2005. Optimization of the ammonia fiber explosion (AFEX) treatment parameters for enzymatic hydrolysis of corn stover, this volume.




Effects of different pretreatment strategies on corn stalk acidogenic fermentation using a microbial consortium Peng Guo a,b,c, Kazuhiro Mochidzuki c, Wei Cheng a, Ming Zhou a, Hong Gao a, Dan Zheng d, Xiaofen Wang b, Zongjun Cui b,⇑ a

Institute of Agricultural Products Processing and Nuclear Agriculture Technology Research, Hubei Academy of Agricultural Sciences, Wuhan 430064, China College of Agronomy and Biotechnology/Center of Biomass Engineering, China Agricultural University, Beijing 100193, China Institute of Industrial Science, The University of Tokyo, Tokyo 153-8505, Japan d Guangxi Soil and Fertilizer Station, Nanning 530007, China b c

a r t i c l e

i n f o

Article history: Received 26 January 2011 Received in revised form 19 April 2011 Accepted 25 April 2011 Available online 7 May 2011 Keywords: Microbial consortium Lignocellulose degradation Pretreatments Acidogenic fermentation Volatile fatty acid (VFA)

a b s t r a c t The effects of sulfuric acid, acetic acid, aqueous ammonia, sodium hydroxide, and steam explosion pretreatments of corn stalk on organic acid production by a microbial consortium, MC1, were determined. Steam explosion resulted in a substrate that was most favorable for microbial growth and organic acid productions. The total amounts of organic acids produced by MC1 on steam exploded, sodium hydroxide, sulfuric acid, acetic acid, and aqueous ammonia pretreated corn stalk were 2.99, 2.74, 1.96, 1.45, and 2.21 g/l, respectively after 3 days of fermentation at 50 °C. The most prominent organic products during fermentation of steam-exploded corn stalks were formic (0.86 g/l), acetic (0.59 g/l), propanoic (0.27 g/l), butanoic (0.62 g/l), and lactic acid (0.64 g/l) after 3 days of fermentation; ethanol (0.18 g/l), ethanediol (0.68 g/l), and glycerin (3.06 g/l) were also produced. These compounds would be suitable substrates for conversion to methane by anaerobic digestion. Ó 2011 Published by Elsevier Ltd.

1. Introduction Corn stalk is a promising renewable feedstock for biological conversion to fuels and chemicals. Although microbial decomposition of such lignocellulosic materials has been studied extensively, most of these studies have focused on pure cultures of microorganisms (Lynd et al., 2002; Wong et al., 1988). Microorganisms in pure culture regularly demonstrate unsatisfactory lignocellulolytic activities (Kim et al., 2006; Koullas et al., 1992; Osborne and Dehority, 1989), and few of them are able to decompose natural lignocellulose with the complex composition and structure found in corn stalks. MC1 is thermophilic cellulose degrading consortium (Haruta et al., 2002) that has not been fully characterized but is known to contain Clostridium straminisolvens CSK1, Clostridium sp. strain FG4b, Pseudoxanthomonas sp. strain M1–3, Brevibacillus sp. strain M1–5 and Bordetella sp. strain M1–6 (Kato et al., 2005). MC1 is capable of degrading rice straw (Haruta et al., 2002) and corn stalk (Peng et al., 2008) and of producing organic acids. The conversion of compounds present in such straws into organic acids is a prerequisite for biogas production (Kaparaju et al., 2009a,b; Weiland, 2010). ⇑ Corresponding author. Tel.: +86 10 62731857, fax: +86 10 62733437. E-mail address: [email protected] (Z. Cui). 0960-8524/$ - see front matter Ó 2011 Published by Elsevier Ltd. doi:10.1016/j.biortech.2011.04.083

Since the conversion of lignocellulosic biomass such as straw and corn stalk is relatively recalcitrant to microbial degradation, appropriate pretreatment is a crucial prerequisite for bioconversion of lignocellulosic feedstock (Alvira et al., 2010; Chen et al., 2011; Himmel et al., 2007; Mosier et al., 2005). The current study measured biomass and fermentation products of MC1 cultured on corn stalk pretreated with acids, bases or by steam explosion to evaluate the potential for subsequent biogas production. 2. Methods 2.1. Lignocellulosic materials Corn stalk (variety CAU 80, bred by the Centre for National Maize Improvement at China Agricultural University in Beijing, China), the lignocellulosic material, was obtained locally after corn harvest from experimental fields at China Agricultural University, Beijing, China, and dried at 80 °C. The corn stalks were cut into pieces approximately 3 cm in length for further use. 2.2. Pretreatment 2.2.1. Acid pretreatment Fifty grams of dried corn stalks were soaked in 1 l H2SO4 (1 M) or acetic acid (1 M) under static condition at room temperature,

P. Guo et al. / Bioresource Technology 102 (2011) 7526–7531

resulting in 5% (w/v) dry straw solids loading. After 24 h, the corn stalks were collected and washed with tap water (approximately pH 6.0) and then with 1 N NaOH to reach a neutral pH before oven-drying to constant weight at 80 °C and milling to pass through 2 mm screens.

7527

filter bag (F57, ANKOM Technology, USA). The components of residual lignocellulosic materials were analyzed according to Goering and Van Soest (1970) using a fiber analyser (Model ANKOM220, USA) as described by Guo et al. (2010). 2.6. GC–MS analysis of volatile products

2.2.2. Alkaline pretreatment Fifty grams of dried corn stalk was soaked in 1 l of 0.4 N of NaOH solution or NH3H2O (1 M) under static condition at room temperature, resulting in 5% (w/v) dry straw solids loading. The stalk/alkali mixture was placed at room temperature. After 24 h, the corn stalks were collected and washed with tap water (approximately pH 8.0) and then with 1 N HCl to reach a neutral pH, ovendrying to constant weight at 80 °C and milling to pass through 2 mm screens. 2.2.3. Steam explosion pretreatment The steam explosion pretreatment was carried out using laboratory scale equipment, consisting of a steam generator and a 1 l pressurized reactor. The reactor was filled with 20 g of feedstock per batch (the dried corn stalk was previously cut into 1 cm pieces) and was then heated to the desired temperatures to reach a pressure in the range of 0.65–0.75 MPa with saturated steam (165 °C), the corn stalks remained in the reactor with saturated steam for 30 s. After a predetermined cooking time (3 min), the valve was opened to release the pressure abruptly, and then the exploded samples were collected and washed with distilled water three times to remove the soluble substances generated in the explosion process. Samples were then oven-dried to constant weight at 80 °C and milled to pass through 2 mm screens. 2.2.4. Without pretreatment The corn stalks were not pretreated with any chemicals before drying to constant weight at 80 °C and milling to pass through 2 mm screens. 2.3. Cultivation with consortium MC1 MC1 from frozen stock was inoculated into 125 ml of sterile peptone cellulose solution (PCS) pH 7.0 ± 0.2 containing (g/l), peptone, 5; yeast extract, 1; CaCO3, 2; NaCl, 5; printer paper, and allowed to grow statically at 50 °C for 3 days. Two-hundred-fifty ml of PCS medium containing 10 g of treated or untreated corn stalk instead of printer paper was inoculated with 5% (v/v) of the 3-day-old MC1 culture and incubated in 500 ml flask under static conditions at 50 °C for 15 days. Samples were taken on days 0, 1, 3, 6, 9, 12, and 15 for analysis. Unless otherwise specified, the data presented are the average of three replications.

On days 0, 1, 3, 6, 9, 12, and 15, 0.5 ml of fermentation broth was centrifuged at 12,000g for 10 min, and the supernatant was filtered using an aperture of 0.22 lm. The filtrate was analyzed by GC–MS (model QP-2010, Shimadzu, Japan) on line with capillary column CP-Chirasil-Dex CB (25 0.25 mm). Conditions were as follows: initial temperature of 60 °C, 1 min; linear ramp up to 100 °C at 7 °C/min and to 195 °C at 18 °C/min for 3 min; injector temperature: 190 °C; ion source temperature: 200 °C; carrier gas: He (60 kPa); rate of flow: 34 ml/min; splitter ratio: 1/20; voltage of detector: 0.7 kv; sample volume: 1 ll. Qualitative identification of the resulting peaks was accomplished using the NIST database (Wang et al., 2006). Furthermore, dilutions of the corresponding compounds were injected as standards to quantify the compounds and confirm the peak positions. 3. Results and discussion 3.1. Effect of pretreatments on solubilization and composition of corn stalk The percent solubilization of corn stalks under different pretreatments is shown in Fig. 1. Steam explosion pretreatment and sodium hydroxide pretreatment produced greater solubilization of corn stalk than the other pretreatments, and alkali pretreatment (sodium hydroxide or aqueous ammonia pretreatment) gave greater solubilization than acid pretreatment (sulfuric acid or acetic acid); no significant difference (P = 0.178) in solubilization between sodium hydroxide pretreatment and steam explosion pretreatment was detected. The compositional changes in under different pretreatments are summarized in Table 1. The corn stalk mainly lost hemicellulose after acid treatment, and the relative contents of cellulose and lignin increased. The corn stalk mainly lost lignin after alkaline treatment, in addition, the hemicelluloses content decreased, while the relative content of cellulose increased noticeably. Hemicellulose decreased after steam explosion, and the relative cellulose and lignin content increased.

2.4. Measurements of microbial growth and culture pH MC1 growth was based on bacterial protein evaluation as previously described (Bensadoun and Weinstein, 1976; Brown et al., 1989). The pH was determined using a HORIBA Compact pH meter (Model B-212, Japan). 2.5. Determination of lignocellulose component and weight loss of corn stalk The cultures were subjected to centrifugation at 12,000g for 10 min, and the pellets were washed with acetic acid/nitric acid reagent followed by a water rinse to remove non-cellulosic materials (Updegraff, 1969). Uninoculated medium served as the control. The weight loss of residual substrates was determined as described by Peng et al. (2008). Residual corn stalk materials were passed through 1 mm screens, and a 0.5-g sample was transferred into a

Fig. 1. Solubilization of corn stalk by different pretreatments. Solubilization is expressed as the percentage of raw material dissolved by pretreatment on the basis of dry weight. Bars represent the standard deviation on the set of values. Values (means of three replicates) not sharing common letters are significantly different (one-way ANOVA with Student–Newman–Keuls method, P < 0.05).

7528


Table 1 Composition of different pretreated corn stalks (%, dry matter). Different pretreatment

Celluloseb

Hemicellulose

Lignin

Ash

SSc

Without pretreatment H2SO4 (1 M)a CH3COOH (1 M) NH3H2O (1 M) NaOH (0.4 N) Steam explosion

37.1 ± 0.5A 49.7 ± 0.8B 46.5 ± 0.5C 48.4 ± 0.6BD 52.4 ± 0.7E 49.3 ± 0.5BF

24.1 ± 0.5A 14.8 ± 0.4B 17.2 ± 0.3C 18.5 ± 0.3D 17.6 ± 0.2CE 12.2 ± 0.1F

12.1 ± 0.2AB 12.7 ± 0.4BC 12.3 ± 0.3AC 9.3 ± 0.2D 5.7 ± 0.1E 14.4 ± 0.3F

5.8 ± 0.1A 6.2 ± 0.3A 6.5 ± 0.2A 6.4 ± 0.1A 7.1 ± 0.3B 6.2 ± 0.2A

14.3 ± 0.2A 10.5 ± 0.1B 11.2 ± 0.2C 11.6 ± 0.2CDE 12.3 ± 0.3AE 12.2 ± 0.2ADF

a

Concentration of chemical reagent indicated in parentheses. Mean ± standard deviation of three replicate measurements. Values (means of three replicates) in the same column not sharing common letters are significantly different (one-way ANOVA with Student–Newman–Keuls method, P < 0.05). c SS: Soluble substance, includes soluble saccharides, starch, and a small amount of protein. b

3.2. Effects of pretreatments on MC1 growth As a proxy for the amount of microbial growth, the levels of bacterial proteins from MC1 were measured in under different pretreatments. The growth of MC1 was significantly different under the different pretreatments (Fig. 2). At the start of acidogenic fermentation (before day 3), the concentration of the bacterial protein with unpretreated corn stalk as the substrate was significantly higher than that of the pretreated corn stalk, which would be that the bacteria utilized the peptone and some solubles (Table 1) from the unpretreated corn stalk, and that there were inhibitory substances in the pretreated samples that could have delayed growth until used by some component of MC1. After 3 days of fermentation, bacterial protein content remained virtually unchanged using the unpretreated corn stalk as the substrate; however, the bacterial protein levels indicated maintenance of a rapid growth rate on pretreated corn stalks. The protein concentrations of MC1 bacteria growing on steam exploded corn stalk were significantly higher than those of other pretreatments. Nevertheless, because the unpretreated corn stalk is more recalcitrant than the pretreated corn stalk, the microbes reproduced more slowly on the unpretreated corn stalk than on the pretreated corn stalks after day 3 when the soluble substances were initially consumed. Microbe growth on corn stalks also increased more quickly with steam explosion pretreatment and alkali pretreatments than with acid pretreatment. Microbes also grew more quickly when using the alkali pretreated corn stalks than the acid pretreated corn stalk as the substrates. Fig. 2 also indicates that MC1 could grow in the medium without corn stalk as carbon source. It is apparent that, compared with the control (MC1 growth in basal medium without corn stalk), MC1 can grow more significantly on corn stalks than in basal medium.

Fig. 2. Effects of pretreatments on MC1 bacterial protein concentration during growth with different pretreated corn stalk. Bars represent the standard deviation on the set of values. Values (means of three replicates) not sharing common letters are significantly different (one-way ANOVA with Student–Newman–Keuls method, P < 0.05).

Previous investigations indicated that the same MC1 bacteria can be found in each subculture during the degradation process, although some bacteria may be found at low numbers (Haruta et al., 2002; Peng et al., 2008). However, the microbial consortium will likely experience dynamic changes when the lignocellulosic substrates change. Our future studies will focus on these dynamic changes in the bacterial consortium growing on different pretreated crop straws and different lignocellulosic substrates (waste paper, rice straw, corn stalk and wheat straw) using methods of molecular analysis (e.g. real-time quantitative PCR).

3.3. Effects of pretreatments on corn stalk degradation Fig. 3 shows that the unpretreated corn stalk weight decreased most quickly during the first day, but did not significantly change after day 3. The weight loss of steam exploded and sodium hydroxide pretreated corn stalk was significantly higher than that of the sulfuric acid, acetic acid, aqueous ammonia pretreated and unpretreated corn stalk after day 6. The weight loss of steam exploded corn stalk was significantly higher than that of the sodium hydroxide pretreated corn stalk, and the weight loss with aqueous ammonia pretreatment was significantly higher than with sulfuric acid or acetic acid pretreatment at the end of the fermentation. Lignin and hemicellulose removal are two key outcomes of pretreatment (Hendriks and Zeeman, 2009). In this study, alkali pretreatment gave greater removal of lignin than acid pretreatment, mainly removing hemicelluose (Table 1). From the results of chemical pretreated (acid pretreated and alkali pretreated) corn stalk degradation by MC1 and the composition (Table 1) and solubilization (Fig. 2) of corn stalk after acid pretreatment and alkali pretreatment, we concluded that lignin removal is more effective

Fig. 3. Effects of pretreatments on corn stalk degradation by MC1. Bars represent the standard deviation on the set of values. Values (means of three replicates) not sharing common letters are significantly different (one-way ANOVA with Student– Newman–Keuls method, P < 0.05).


than hemicellulose removal for MC1 acidification of corn stalk. However, from the results of sodium hydroxide pretreatment and the steam explosion pretreatment, although solubilization of corn stalk with sodium hydroxide pretreatment and steam explosion pretreatment were at the same level (Fig. 1), sodium hydroxide pretreatment was not as effective in promoting corn stalk degradation by MC1 and bacteria growth as steam explosion. Steam explosion can break down the links between the crystalline cellulose and amorphous cellulose, rendering the cellulose more easily accessible to enzymes and microbes (Chen et al., 2011; Laser et al., 2002; Mosier et al., 2005). The integrated analyzes of the results suggest that solubilization is just one of the indicators for evaluating the effects of the pretreatments; the destruction of the molecular structure of lignocellulose might be another important indicator for evaluation. In addition, it is worth mentioning that very few cultivated microorganisms can degrade lignocellulosic biomass without pretreatment, and we show here that MC1, a microbial consortium, not only can efficiently utilize pretreated corn stalk, but it can also degrade corn stalk without pretreatment.

3.4. Effects of pretreatments on pH during degradation The pH values of the MC1 fermentation broth with different pretreatments were determined. Fig. 4 shows that the pH value of the fermentation broth with the unpretreated corn stalk declined more significantly before day 1 and increased more dramatically after day 3 than the pretreated corn stalk during degradation. The results also indicate that the pH value of fermentation broth without corn stalk declined gently after 1 day. The results also suggest that the pH values of the fermentation broth of acid (sulfuric acid and acetic acid) pretreated corn stalk were still below 7.0 on day 6, however, the pH values of the fermentation broth of alkaline (sodium hydroxide and aqueous ammonia) pretreated corn stalk and steam exploded corn stalk were over 7.0 on day 6. However, at the end of the fermentation, no significant difference in pH values of all of the pretreatments was detected, which suggested that the under all these pretreatments MC1 has the ability recover the pH, which is a known characteristic of MC1 (Liu et al., 2006). Microbes degrading ability is frequently inhibited by the imbalance in pH late in the degradation process (Juhász et al., 2004; Kang et al., 1994). The autorecovery of pH of the fermentation broth is because of the presence of acidophilus strains of MC1 (Kato et al., 2005) preventing the pH from dropping too rapidly to inhibit the degrading activities of the bacteria.

Fig. 4. Effects of pretreatments on pH value in corn stalk degradation by MC1. Bars represent the standard deviation on the set of values. Values (means of three replicates) not sharing common letters are significantly different (one-way ANOVA with Student–Newman–Keuls method, P < 0.05).

7529

3.5. Analyzes of the volatile organic products (VOPs) of corn stalk degradation GC–MS analyzes showed that formic acid, acetic acid, propanoic acid, butanoic acid, lactic acid, ethanol, ethanediol, and glycerin were the predominant VOPs in the fermentation broth during corn stalk degradation. As shown in Fig. 5, the concentration of acetic acid (Fig. 5B), propanoic acid (Fig. 5C), butanoic acid (Fig. 5D), lactic acid (Fig. 5E), ethanol (Fig. 5F), ethanediol (Fig. 5G), and glycerin (Fig. 5H) produced from unpretreated corn stalk are significantly higher than from the sodium hydroxide, aqueous ammonia, sulfuric acid, acetic acid pretreated and steam exploded corn stalks on day 1, and they decrease remarkably after day 1. The concentration of formic acid, acetic acid, propanoic acid, butanoic acid, lactic acid, ethanediol, and glycerin reached peak values with the five pretreated corn stalks as substrate on day 3. The concentration of the five VFAs produced from steam exploded and sodium hydroxide pretreated corn stalks were significantly higher than from the other pretreated corn stalks on day 3, and the concentration of formic acid (0.862 g/l), acetic acid (0.594 g/l), propanoic acid (0.273 g/ l), butanoic acid (0.620 g/l), lactic acid (0.636 g/l) from steam exploded corn stalks were significantly higher than from sodium hydroxide pretreated corn stalks; however, no significant difference in concentration of acetic acid and butanoic acid between steam exploded corn stalk and sodium hydroxide pretreated corn stalk was detected. Based on the concentration of VFAs, we predicted the theoretical yield of methane with the empirical formula (CnHaOb + (n a/4 b/2) H2O = (n/2 a/8 + b/4) CH4 + (n/2 + a/8 b/ 4)CO2) (Buswell and Sollo, 1948); over 6.2 l (under the standard conditions) of methane might be produced from the five main VFAs produced from the steam exploded corn stalks in 1 l of the fermentation broth on day 3. In contrast to the production of the five VFAs, the concentration of ethanol peaks on day 6, and the concentration of ethanol produced from the sodium hydroxide or aqueous ammonia pretreated corn stalk is higher than from other pretreated corn stalk. The production of ethanediol and glycerin peaked on day 3, and was highest with steam explosion compared to the other pretreatments. After a 9-day fermentation, only ethanediol and glycerin were remarkably detectable. Because of the complex microbial composition of MC1 (Haruta et al., 2002; Kato et al., 2005), the concentrations of formic acid, acetic acid, propanoic acid, and butanoic acid, which might be consumed by MC1 itself, declined remarkably after day 3. However, the concentration of ethanediol and glycerin did not decline significantly, which might be due to the fact that ethanediol and glycerin not being further utilized by MC1. On the first day, the pH value of the fermentation broth of unpretreated corn stalk decomposition declined more dramatically than that of the pretreated corn stalks (Fig. 4); simultaneously, the VFAs increased more sharply than the pretreated corn stalks. This is relevant to the pH results, which might be caused by the soluble substance content of the unpretreated corn stalk being significantly higher than those of the five pretreatments (Table 1). The pH value of the fermentation broth of the steam exploded, sodium hydroxide pretreated, aqueous ammonia pretreated, sulfuric acid pretreated, and acetic acid pretreated corn stalk reached a minimum (Fig. 4) on day 3, the VFAs levels peaked, and the weight of the corn stalks decreased sharply in the first 3 days (Fig. 3). The pH values of the fermentation broth of acid pretreated corn stalk were still below 7.0 on day 6, but increased to 7.0 with alkaline pretreated and steam exploded corn stalk as substrate. This is relevant to VFAs production during day 3–6, as the VFAs produced from the alkaline pretreated or steam exploded corn stalks declined more steeply than the acid pretreated corn stalk. This also indicated that acid pretreated corn stalks were more difficult for MC1 to degrade than the alkaline pretreated or steam exploded

7530


Fig. 5. Effects of pretreatments on predominant volatile organic products production during corn stalk degradation by MC1. Bars represent the standard deviation on the set of values. Values (means of three replicates) not sharing common letters are significantly different, asterisk (⁄) indicates significant difference between without pretreatment and other pretreatments on the same day (one-way ANOVA with Student–Newman–Keuls method, P < 0.05).

corn stalks. Meanwhile, the pH values did not decrease below 6.5 during the degradation of acid pretreated corn stalk, which indicated that the degradation of acid pretreated corn stalk was not very vigorous and the accumulation of organic acids was limited. In addition, compared with MC1 growing on the corn stalk, only small amounts of acetic acid (Fig. 5B), butanoic acid (Fig. 5D), ethanediol (Fig. 5G), and glycerin (Fig. 5H) were produced by MC1 growing in the basal media, which indicated that the VOPs were mainly produced from the corn stalks acidification by MC1. The formic, acetic, propanoic, butanoic, and lactic acid are important organic acids which can be used for methane fermenta-

tion. Because the VFAs decreased remarkably after day 3, the fermentation broth should be used for methane fermentation before day 3. Ethanediol and glycerin accumulated remarkably, which can be fermented by anaerobic digestion (Basri et al., 2010; Yang et al., 2008); in addition, glycerol can improve methane production during anaerobic digestion (Amon et al., 2006). There is no appreciable accumulation of fermentable sugars during lignocellulose (corn stalk or rice straw) (Haruta et al., 2002; Peng et al., 2008) degradation by MC1 because of the complex microbial composition (Haruta et al., 2002; Kato et al., 2005); however, MC1 can effectively degrade natural lignocelluose to produce VFAs; therefore,


it can be used, except for saccharification, to overcome the bottleneck in the conversion of lignocellulose to biofuels, including acidogenic fermentation to improve the efficiency of methane fermentation. 4. Conclusions Steam explosion and sodium hydroxide pretreatments were more effective in corn stalk solubilization and promoting MC1 growth than other pretreatments. Steam explosion was the best pretreatment for corn stalk acidification by MC1. Formic, acetic, propanoic, butanoic, and lactic acid were the predominant VFAs produced from steam exploded corn stalks during the acidification, which significantly accumulated on day 3, are good for potential methane fermentation; the predicated theoretical yield of methane from the fermentation broth might reach over 6.2 l of methane/l of fermentation broth. This study will be of great significance in biogas production from agricultural wastes. Acknowledgements This work was supported by the National Key Technology Research and Development Program of China during the 11th FiveYear Plan Period (No. 2006BAD07A01, 2008BADC4B01). References Alvira, P., Tomás-Pejó, E., Ballesteros, M., Negro, M.J., 2010. Pretreatment technologies for an efficient bioethanol production process based on enzymatic hydrolysis: a review. Bioresource Technology 101, 4851–4861. Amon, T., Amon, B., Kryvoruchko, V., Bodiroza, V., Pötsch, E., Zollitsch, W., 2006. Optimising methane yield from anaerobic digestion of manure: effects of dairy systems and of glycerine supplementation. International Congress Series 1293, 217–220. Basri, M., Yacob, S., Hassan, M., Shirai, Y., Wakisaka, M., Zakaria, M., Phang, L., 2010. Improved biogas production from palm oil mill effluent by a scaled-down anaerobic treatment process. World Journal of Microbiology and Biotechnology 26, 505–514. Bensadoun, A., Weinstein, D., 1976. Assay of proteins in the presence of interfering materials. Analytical Biochemistry 70, 241–250. Brown, R.E., Jarvis, K.L., Hyland, K.J., 1989. Protein measurement using bicinchoninic acid: elimination of interfering substances. Analytical Biochemistry 180, 136– 139. Buswell, A.M., Sollo Jr., F.W., 1948. The mechanism of the methane fermentation. Journal of the American Chemical Society 70, 1778–1780. Chen, W.H., Pen, B.L., Yu, C.T., Hwang, W.S., 2011. Pretreatment efficiency and structural characterization of rice straw by an integrated process of dilute-acid and steam explosion for bioethanol production. Bioresource Technology 102, 2916–2924. Goering, H.K., Van Soest, P.J., 1970. Forage fiber analyses. USDA Agriculture, Handbook No. 379. Guo, P., Zhu, W., Wang, H., Lü, Y., Wang, X., Zheng, D., Cui, Z., 2010. Functional characteristics and diversity of a novel lignocelluloses degrading composite

7531

microbial system with high xylanase activity. Journal of Microbiology and Biotechnology 20, 254–264. Haruta, S., Cui, Z., Huang, Z., Li, M., Ishii, M., Igarashi, Y., 2002. Construction of a stable microbial community with high cellulose-degradation ability. Applied Microbiology and Biotechnology 59, 529–534. Hendriks, A.T.W.M., Zeeman, G., 2009. Pretreatments to enhance the digestibility of lignocellulosic biomass. Bioresource Technology 100, 10–18. Himmel, M.E., Ding, S.-Y., Johnson, D.K., Adney, W.S., Nimlos, M.R., Brady, J.W., Foust, T.D., 2007. Biomass recalcitrance: engineering plants and enzymes for biofuels production. Science 315, 804–807. Juhász, T., Szengyel, Z., Szijártó, N., Réczey, K., 2004. Effect of pH on cellulase production of Trichoderma ressei RUT C30. Applied Biochemistry and Biotechnology 113, 201–211. Kang, S.W., Kim, S.W., Kim, K., 1994. Production of cellulase and xylanase by Aspergillus niger KKS. Journal of Microbiology and Biotechnology 4, 49–55. Kaparaju, P., Serrano, M., Angelidaki, I., 2009a. Effect of reactor configuration on biogas production from wheat straw hydrolysate. Bioresource Technology 100, 6317–6323. Kaparaju, P., Serrano, M., Thomsen, A.B., Kongjan, P., Angelidaki, I., 2009b. Bioethanol, biohydrogen and biogas production from wheat straw in a biorefinery concept. Bioresource Technology 100, 2562–2568. Kato, S., Haruta, S., Cui, Z.J., Ishii, M., Igarashi, Y., 2005. Stable coexistence of five bacterial strains as a cellulose-degrading community. Applied and Environmental Microbiology 71, 7099–7106. Kim, T., Lee, Y., Sunwoo, C., Kim, J., 2006. Pretreatment of corn stover by low-liquid ammonia recycle percolation process. Applied Biochemistry and Biotechnology 133, 41–57. Koullas, D.P., Christakopoulos, P., Kekos, D., Macris, B.J., Koukios, E.G., 1992. Correlating the effect of pretreatment on the enzymatic hydrolysis of straw. Biotechnology and Bioengineering 39, 113–116. Laser, M., Schulman, D., Allen, S.G., Lichwa, J., Antal Jr., M.J., Lynd, L.R., 2002. A comparison of liquid hot water and steam pretreatments of sugar cane bagasse for bioconversion to ethanol. Bioresource Technology 81, 33–44. Liu, J.B., Wang, W.D., Yang, H.Y., Wang, X.F., Gao, L.J., Cui, Z.J., 2006. Process of rice straw degradation and dynamic trend of pH by the microbial community MC1. Journal of Environmental Sciences 18, 1142–1146. Lynd, L.R., Weimer, P.J., van Zyl, W.H., Pretorius, I.S., 2002. Microbial cellulose utilization: fundamentals and biotechnology. Microbiology and Molecular Biology Reviews 66, 506–577. Mosier, N., Wyman, C., Dale, B., Elander, R., Lee, Y.Y., Holtzapple, M., Ladisch, M., 2005. Features of promising technologies for pretreatment of lignocellulosic biomass. Bioresource Technology 96, 673–686. Osborne, J.M., Dehority, B.A., 1989. Synergism in degradation and utilization of intact forage cellulose, hemicellulose, and pectin by three pure cultures of ruminal bacteria. Applied and Environmental Microbiology 55, 2247–2250. Peng, G., Xiaofen, W., Wanbin, Z., Hongyan, Y., Xu, C., Zongjun, C., 2008. Degradation of corn stalk by the composite microbial system of MC1. Journal of Environmental Sciences 20, 109–114. Updegraff, D.M., 1969. Semimicro determination of cellulose in biological materials. Analytical Biochemistry 32, 420–424. Wang, X., Haruta, S., Wang, P., Ishii, M., Igarashi, Y., Cui, Z., 2006. Diversity of a stable enrichment culture which is useful for silage inoculant and its succession in alfalfa silage. FEMS Microbiology Ecology 57, 106–115. Weiland, P., 2010. Biogas production: current state and perspectives. Applied Microbiology and Biotechnology 85, 849–860. Wong, K.K.Y., Tan, L.U.L., Saddler, J.N., 1988. Multiplicity of b-1,4-xylanase in microorganisms: functions and applications. Microbiology and Molecular Biology Reviews 52, 305–317. Yang, Y., Tsukahara, K., Sawayama, S., 2008. Biodegradation and methane production from glycerol-containing synthetic wastes with fixed-bed bioreactor under mesophilic and thermophilic anaerobic conditions. Process Biochemistry 43, 362–367.




Comparison of different pretreatment methods for lignocellulosic materials. Part I: Conversion of rye straw to valuable products Thomas Ingram a,⇑, Kai Wörmeyer a, Juan Carlos Ixcaraguá Lima a, Vera Bockemühl b, Garabed Antranikian b, Gerd Brunner a, Irina Smirnova a a b

Hamburg University of Technology, Institute of Thermal Separation Processes, Eißendorfer Straße 38, D-21073 Hamburg, Germany Hamburg University of Technology, Institute of Technical Microbiology, Kasernenstraße 12, D-21073 Hamburg, Germany

a r t i c l e

i n f o

Article history: Received 20 May 2010 Received in revised form 8 December 2010 Accepted 1 February 2011 Available online 24 February 2011 Keywords: Lignocellulose Pretreatment Organosolv Biorefinery Enzymatic hydrolysis

a b s t r a c t The conversion of lignocellulose to valuable products requires I: a fractionation of the major components hemicellulose, cellulose, and lignin, II: an efficient method to process these components to higher valued products. The present work compares liquid hot water (LHW) pretreatment to the soda pulping process and to the ethanol organosolv pretreatment using rye straw as a single lignocellulosic material. The organosolv pretreated rye straw was shown to require the lowest enzyme loading in order to achieve a complete saccharification of cellulose to glucose. At biomass loadings of up to 15% (w/w) cellulose conversion of LHW and organosolv pretreated lignocellulose was found to be almost equal. The soda pulping process shows lower carbohydrate and lignin recoveries compared to the other two processes. In combination with a detailed analysis of the different lignins obtained from the three pretreatment methods, this work gives an overview of the potential products from different pretreatment processes. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction According to Kamm et al. (2006) the term biorefinery represents a ‘‘complex (to fully integrated) system of sustainable, environmentally, and resource friendly technologies for the comprehensive material and energetic utilization as well as exploitation of biological raw materials in form of green and residue biomass from a targeted sustainable regional land utilization’’. However, the key issue to replace the present fossil-based petro-chemical industry with biorefineries, is the conversion of carbohydrates from lignocellulosic feedstocks into fermentable sugars (Jørgensen et al., 2007a). An effective conversion of lignocellulosic materials to valuable products first requires a fractionation of the three major components: cellulose, hemicellulose, and lignin, and second, an efficient method to process these components to higher valued products. In general, carbohydrates are hydrolysed to their monomers, whereas the lignin fraction could be utilized as a source for phenolic polymers (Lora and Glasser, 2002). An overview of various fractionation and pretreatment processes is given by Alvira et al. (2009), and includes organosolv-, lime-, ammonia steam explosion-, diluted acid-, and liquid hot

Abbreviations: AL, aquasolve liquefied; AS, aquasolve solid; BCA, bicinchoninic acid; FPU, filter paper unit; LHW, liquid hot water; R, residual biomass. ⇑ Corresponding author. Tel.: +49 (40) 42878 2846; fax: +49 (40) 42878 4072. E-mail addresses: [email protected], [email protected] (T. Ingram). 0960-8524/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2011.02.005

water (LHW)-pretreatment. The effectiveness of these pretreatment methods is generally evaluated by the accessibility of cellulose to cellulases (Wyman et al., 2005; Holtzapple and Humphrey, 1984). The latest developments have clearly shown that future biorefineries can only become economically competitive if the whole range of product streams is processed (Zhang, 2008). Hence, for an effective comparison of different pretreatments, it is essential to take into consideration all potential products. Still, a direct comparison of the pretreatment methods is often difficult, since different biomasses are used for each process. To avoid this, in the present work, three pretreatment methods were compared: liquid hot water (LHW) pretreatment, the soda pulping process and the ethanol organosolv pretreatment while using rye straw as a single lignocellulosic material. The combination of a LHW treatment (Bobleter, 1994; Garotte et al., 1999) with an enzymatic hydrolysis has proven to be an efficient environmentally friendly process to convert lignocellulosic material like straw and corn stover to valuable products. Jedicke et al. (2000) have named the LHW pretreatment AquasolvÒ – Hydrothermolyse. Accordingly, we refer to the whole process as the Aquasolve process, since water remains the reaction media throughout the entire process. 1.1. The Aquasolve process Fig. 1 shows the flow diagram of the Aquasolve process. The lignocellulosic material can be utilized without further conditioning.

5222

T. Ingram et al. / Bioresource Technology 102 (2011) 5221–5228

First, the biomass is fractionated using LHW at T = 473 K. The two resulting product streams are a liquid hydrolysate (1) and a solid fraction (2) consisting almost exclusively of cellulose and lignin. At room temperature, the Aquasolve liquefied (AL) lignin (4) precipitates from the hydrolysate, while the dissolved xylans can be enzymatically converted to xylose (3). The enzymatic saccharification of the cellulose-rich fraction yields two products, the dissolved glucose (5) and a solid lignin fraction (6). This lignin fraction is referred to as Aquasolve solid (AS) lignin. Low sugar concentrations are desirable in this fraction. The latest developments have shown that sulphur-free lignins with properties similar to lignin in the plant material might become an interesting alternative as a source for polymeric materials (Zhang, 2008). The separation efficiency of the fractionation at 473 K is demonstrated in Fig. 2. At temperatures above 473 K, almost all of the xylose and arabinose (representing >96% of the initial hemicellulose) can be recovered in the liquid fraction whereas almost no glucose can be detected (Ingram et al., 2009). At temperatures above 523 K, the cellulose can be depolymerised. However, the glucose recovery in the liquid fraction remains limited due to degradation reactions of glucose to minor products like hydroxymethylfurfural. These results clearly demonstrate that LHW is a selective reaction medium at 473 K that can be used to separate hemicellulose from lignocellulosic materials like rye straw. Franceschin et al. (2010) have demonstrated the advantage of the given technique and the potential applications of the dissolved hemicelluloses. It was shown that the utilization of pentoses can greatly affect the overall process economy of a multi-product biorefinery. The cellulose-rich solid fraction after LHW pretreatment can be efficiently converted to glucose by cellulolytic enzymes. Although this technology has been studied in detail, residence time, sugar yields, and the enzyme consumption all remain major limitations in making it a profitable process (Wyman et al., 2005). In this work, the three major fractions of lignocellulosic biomass: cellulose, hemicellulose, and lignin of the Aquasolve process, the soda pulping process and the ethanol organosolv process were compared. In contrast to previous works (Wyman et al., 2005; Chen et al., 2009), the enzymatic conversion was optimized for each pretreated substrate and for different enzyme mixtures. As

Fig. 2. Hydrolysis profile of rye straw in LHW; amount of dissolved mass related to the initial biomass with increasing temperatures (Ingram et al., 2009).

a result the different fractionation methods could be evaluated based on the overall productivity, consumption of chemicals, and efficiency of the subsequent enzymatic saccharification. The quality and potential applications of the lignins isolated from the different fractionation processes are summarized in the second part of our publication: ‘‘Isolation and characterization of rye straw lignin’’ by Wörmeyer et al. (2011). 2. Methods 2.1. Rye straw Rye straw was purchased from Cordes-Grasberg (Grasberg, Germany). The average length of the straw was 3–4 cm. It was used without further comminution. The straw was dried at 378 K until the weight was constant. The average water content was 5.4% (w/w). The composition of rye straw was obtained from acid hydrolysis (Section 2.5) and is given in Table 1.

Fig. 1. Block flow diagram of the Aquasolve process.

5223

T. Ingram et al. / Bioresource Technology 102 (2011) 5221–5228 Table 1 Amount of lignin and sugar found in solid residue (2) and the residual biomass (R) related to the initial dry mass of rye straw after different pretreatment methods. Pretreatment

Glucose (%)

Xylose (%)

Arabinose (%)

Lignin (%)

R (%)

Untreated LHW 260 LHW Soda Organosolv

42.1 ± 2.1 1.8 ± 0.1 75.4 ± 3.9 62.3 ± 0.8 84.9 ± 4.3

21.2 ± 1.0 ND 3.3 ± 0.9 22.0 ± 1.0 4.7 ± 0.2

2.6 ± 0.2 ND 0.1 ± 0.1.0 2.2 ± 0.2 0.1 ± 0.1

19.5 ± 1.1 95.9 ± 4.8 17.5 ± 0.2 4.63 ± 0.9 2.8 ± 0.2

– 10.5 ± 2.4 48.8 ± 1.0 59.9 ± 1.2 43.3 ± 2.4

ND: not detected.

NaOH on the enzymatic hydrolysis. The washing water was collected and analyzed according to Section 2.6. Lignin was separated from the aqueous fraction according to the method described by Silverstein et al. (2007). First the pH was adjusted with acetic acid to 6.5. The solution was pre-concentrated in a rotary evaporator at T = 313 K and diluted with ethanol (ethanol to water = 5:1). The carbohydrate fraction precipitated and was separated from the solution after 24 h by centrifugation. Subsequently the ethanol was evaporated in a rotary evaporator and the final pH was adjusted with 20% (w/w) hydrochloric acid to 1.6. The precipitated lignin was isolated with a centrifuge and washed with water.

2.2. Liquid hot water pretreatment

2.5. Acid posthydrolysis of pretreated rye straw samples

The flow-through experiments were performed in a fixed-bed reactor (50 mL) according to our previous publication (Ingram et al., 2009). The optimum pretreatment temperature was found to be T = 473 K at p = 5 MPa. The water to biomass ratio was 10:1 (w/w), and the average residence time of water in the reactor was 10 min. At the reactor outlet, the effluent was cooled down to ambient temperature in an ice bath. The sugar concentrations in the hydrolysate were determined according to the method described in Section 2.6. Experiments with temperatures up to 553 K were performed in a pilot plant fixed-bed reactor (1 L). The pilot plant fixed-bed reactor (Hofer, Germany) was equipped with an oil jacket. Approximately 215 g (±1 g) of straw were filled in the reactor. At T = 393 K water (70 L/min) was introduced into the system with a high pressure membrane pump (Lewa, Germany) preheated with two heating sleeves and flowed upwardly through the fixed-bed reactor. At 533 K temperature was kept constant for 15 min before the reactor was cooled down. The average heating and cooling rate of the reactor were 0.8 and 4 K/min. The effluent was cooled down with cooling water in a double-pipe heat exchanger and depressurized in order to collect the hydrolysate. Lignin precipitated from the cool liquid fraction after LHW treatment at 473 and 533 K and was separated from the liquid hydrolysate by centrifugation in 50 mL or 1 L vials at 4000 rpm at room temperature for 30 min.

In order to obtain the total number of sugars (sum of monomers and oligomers) in the pretreated samples, the solid residues as well as the liquid hydrolysates of selected experiments were treated with H2SO4. In the case of liquid samples, the acid hydrolysis was performed for 60 min at 394 K with 4% (w/w) H2SO4. The composition of the solid residues was analyzed by the vTI – Institute of Wood Technology and Wood Biology (Hamburg, Germany). The analytical method was based on a stepwise post hydrolysis with concentrated H2SO4 (Willför et al., 2009). In both cases, the samples were neutralized with CaCO3, centrifuged and analyzed for their sugar content (Section 2.6). The residue after the two-step hydrolysis of the solid fraction is referred to as acid insoluble Klason lignin.

2.3. Organosolv pretreatment Conditions for the organosolv pretreatment were selected following the publication from Sidiras and Koukios (2004). The pretreatment conditions were T = 440 K at p = 3.5 MPa. The method was adapted to the fixed-bed reactor (50 mL) used for LHW treatment. 50% (w/w) ethanol in water with 0.5 N sulphuric acid was used. The liquid to biomass ratio was 10:1 (w/w) and the liquid flow rate was 4 mL/min. After 35 min the residual biomass was flushed and cooled with 140 mL of water to avoid a potential inhibition of the enzymatic hydrolysis by the residual ethanol. This washing water was selected separately and analyzed for carbohydrates (Section 2.6). The liquid fraction was diluted with water to an ethanol to water ratio of 1:4 (v/v). Precipitated lignin was separated in a centrifuge and washed with water.

2.6. Sugar analysis The quantitative determination of the produced sugars and their degradation products including glucose, xylose, arabinose, furfural, and hydroxymethylfurfural were performed by means of HPLC. Pure water served as the eluent and was delivered at a flow rate of 0.6 mL/min to an Aminex HPX 42 A column (Bio-Rad). The components were identified with a refractive index detector. Peak identification and quantification of the components detected with this system were accomplished by injecting standard solutions with known composition at different concentrations. The detection limit was 50 mg/L, while sugar concentrations ranged between 0.1 and 20 g/L. The glucose concentrations after the enzymatic saccharification were additionally controlled using the D-Glucose Kit from Roche. 2.7. Enzymes Celluclast 1.5L and Novozyme 188 were purchased from Sigma– Aldrich and produced by Novozymes Denmark. The activities and protein contents are summarized in Table 2. In cooperation with the Lausitz University of Applied Sciences (Senftenberg, Germany), a mutagenized strain of Penicillium janthinellum was cultivated in a 30 L fermentor (Schulz and Hirte, 1989).

Table 2 Enzyme activities and protein content of Celluclast 1.5L and Novozyme 188.

2.4. Soda pulping process The conditions for the soda pulping process were taken from Silverstein et al. (2007). Forty-five grams of rye straw was pretreated with 2% (w/w) NaOH in water for 90 min at T = 394 K and p = 0.5 MPa in a 500 mL sealed autoclave (Hofer, Germany) equipped with a mechanical shaking system. The biomass to liquid ratio was 1:9 (w/w). The pressure was adjusted with nitrogen from a gas bottle. After the pretreatment the solid residue was washed 10 times with 5 L of water to avoid any influence of the residual

Enzyme

Cellulasea (FPU/ml)

b-Glucosidasea (CBU/ml)

Xylanaseb (U/ml)

Protein contentc (mg/mL)

Celluclast 1.5L Novozyme 188

60.0 ± 3.0

68.9 ± 3.0

0.05 ± 0.01

43.4 ± 4.1

ND

791 ± 42

ND

31.5 ± 0.6

ND: not detected. a Method from Zhang et al. (2009a). b BCA assay: 0.5% beechwood xylan (Sigma), pH 4.0, 20 min, 323 K (Miller, 1959). c Method from Bradford (1976).

5224


Table 3 Enzyme activities and protein content of the enzyme complex from Penicillium janthinellum. Enzyme

Cellulasea (FPU/g)

b-Glucosidasea (CBU/g)

Xylanaseb (U/g)

Protein contentc (mg/g)

P. janthinellum

453 ± 26

166 ± 2.2

0.47 ± 0.01

163.5 ± 8.1

ND: not detected. a Method from Zhang et al. (2009a). b BCA assay: 0.5% beechwood xylan (Sigma), pH 4.0, 20 min, 323 K (Miller, 1959). c Method from Bradford (1976).

fraction remains almost constant compared to the initial concentration of lignin in rye straw. Some lignin was dissolved in the LHW during pretreatment. It is important to note that a certain fraction of the dissolved lignin precipitates from the liquid fraction as soon as it is cooled down to ambient conditions. This fraction is attributed to Aquasolve liquefied (AL) lignin (Fig. 1) and is considered independently from the carbohydrate-rich solid fractions after the pretreatment. The overall recovery of the different lignins is discussed in Section 3.4.

3.2. Composition of the liquid fractions after pretreatment The fermentation broth containing extracellular enzymes was lyophilized. The activities and protein content are summarized in Table 3. 2.8. Enzymatic hydrolysis Enzymatic hydrolysis was performed in triplicates in 5 mL tubes for 48 h at 323 K and 120 rpm in a thermo shaker (Grant, United Kingdom). The wet solid cellulose-rich fractions of pretreated rye straw (water content between 70% and 85% (w/w)) were used without further conditioning. The pH was set to pH 5.0 by the addition of buffer (20 mM acetic acid, 20 mM boric acid, 20 mM phosphoric acid). After 48 h the pH was measured and samples were taken and analyzed according to Section 2.6. The biomass loading is defined as follows:

B¼

Dry biomass ðgÞ 100% Buffer solution ðgÞ

ð1Þ

The enzyme loading of Celluclast 1.5L and of P. janthinellum is represented in relation to the biomass/cellulose content:

E¼

FPU biomass=cellulose ðgÞ

ð2Þ

In addition to the solid residues (2), the hydrolysates (1) of the different processes were analyzed according to their carbohydrate concentrations (Fig. 1). The composition of the liquid fractions at room temperature is shown in Table 4. Since the pH-value of the hydrolysate from the LHW pretreatment (1) ranged between 4.0 and 4.2, no pH adjustment is necessary for the enzymatic conversion of the residual xylans to xylose (Ingram et al., 2009). The precipitated AL lignin can be easily separated from the liquid fraction using a centrifuge. If the enzymatic saccharification of the dissolved carbohydrates is performed before separating the precipitated lignin fraction (compare Fig. 1), the lignin content of the AL lignin can be further increased from 75% (w/w) to 82% (w/w). According to Table 4, the lowest xylose recoveries were found for the soda pulping process. This batch process (Section 2.4) originates from the paper industry and has not been optimized for high pentose recoveries. The organosolv process was developed to achieve high lignin and cellulose recoveries. A certain fraction of the pentoses is degraded to minor products like furfural due to the acetic conditions (Sidiras and Koukios, 2004). In contrast to the Aquasolve process, the soda pulping process and the organosolv process require a pH adjustment and the addition of further chemicals like water or ethanol to separate lignin from the liquid fraction.

For all experiments performed, a volumetric ratio of Celluclast 1.5L to Novozyme 188 of 5:1 was used (Eklund et al., 1990). 3. Results and discussion 3.1. Composition of solid residues In order to compare the efficiency and the selectivity of the different pretreatment processes (Fig. 1), the liquid hydrolysate (1) and the solid residue (2) were analyzed for their carbohydrate content. It should be noted, that the LHW and the organosolv pretreatment were performed in a fixed bed reactor (Sections 2.2 and 2.3), in order to avoid degradation of the pentoses in the liquid hydrolysate (Mosier et al., 2005; Ingram et al., 2009). The soda pulping process is an established batch process (Section 2.4) in the paper industry for the fractionation of lignocellulose from annual plants like rye straw, therefore it was chosen as reference process. Table 1 summarizes the composition of the dried residues and the residual biomass (R) after the different pretreatments. LHW 260 represents the solid residue after a hydrothermal treatment at 533 K. The results demonstrate that almost no carbohydrates can be found in the solid residue using LHW at 533 K. The three other pretreatment methods solubilize roughly 40–60% of the initial dry mass. However, the composition of the pretreated solid residues varies significantly. Organosolv pretreatment provides a solid residue which consists almost exclusively of cellulose. The soda pretreatment yields a residue which is rich in cellulose, and also contains significant amounts of xylans. The LHW pretreated residue is significantly enriched in cellulose and the hemicellulose is almost completely separated from the solid fraction. The lignin

3.3. Determination of the enzymatic digestibility of pretreated rye straw An essential criterion for the comparison of the different pretreatments is the recovery of glucose after the enzymatic hydrolysis of the cellulose rich fraction. In order to analyze the effectiveness of the enzymatic saccharification, the biomass loading must be reduced to a minimum (Wyman et al., 2005). Accordingly, a biomass loading of 1% (w/w) was chosen for the enzymatic conversion. Fig. 3A shows the cellulose conversion as a function of enzyme loading for the enzymes Celluclast 1.5L and Novozyme 188. The cellulose fraction of organosolv pretreated rye straw can be completely converted into glucose in 48 h using an enzyme loading lower than 2 FPU/g cellulose. Whereas LHW and soda pretreated rye straw need 25 and 15 FPU/g cellulose respectively for a complete conversion of the cellulose fraction in 48 h. As a compromise between enzyme consumption and cellulose conversion, the enzyme loading needed to achieve 90% (w/w) cellulose conversion

Table 4 Amount of sugar found in fraction (1) after different pretreatment methods. The values are related to total glucose, xylose, and arabinose content in rye straw; values were obtained by an acid posthydrolysis (see Section 3). Pretreatment

Glucose (%)

Xylose (%)

Arabinose (%)

Organosolv Soda LHW 260 LHW

8.3 ± 0.1 4.7 ± 0.3 29.9 ± 1.5 5.3 ± 0.4

92.5 ± 8.5 41.8 ± 2.1 100.6 ± 5.2 98.7 ± 6.1

77.5 ± 2.4 65.4 ± 3.3 90.7 ± 4.5 80.3 ± 5.5


5225

Fig. 3. Cellulose conversion as a function of enzyme loading for the three pretreatment methods applied in this work; enzymes: Celluclast 1.5L and Novozyme 188 (A) and the enzyme complex from Penicillium janthinellum (B), biomass loading = 1% (w/w), t = 48 h, T = 323 K.

of LHW pretreated rye straw was used for further investigations (13 FPU/g cellulose; Celluclast 1.5L and Novozyme 188). Fig. 3B shows the cellulose conversion as a function of enzyme loading for the three pretreated rye straw samples using the enzyme complex from P. janthinellum. Compared to Celluclast 1.5L and Novozyme 188, a larger amount of enzymes is needed to achieve the complete conversion of cellulose into glucose. Almost no difference can be detected between the soda and LHW pretreated straw. A minimum of 15 FPU/g cellulose is needed to achieve a complete conversion of organosolv pretreated substrate. For both enzyme mixtures, the effectiveness of the enzymatic conversion of pretreated lignocellulose clearly depends on the pretreatment it was subjected to. Organosolv pretreatment provides by far the most suitable substrate for both enzyme mixtures. The residual hemicellulose and lignin content in soda and LHW pretreated rye straw might function as a barrier and reduce the cellulase activity (Jeoh et al., 2007). However, it cannot be concluded from these results that the irreversible adsorption of cellulase on lignin decreases the overall conversion rate. Both the soda (5% (w/w) lignin) and the LHW (18% (w/w) lignin) pretreated straw show similar cellulose conversions with increasing enzyme loadings (Fig. 3). 3.4. Influence of biomass loading on the enzymatic digestibility According to Alvira et al. (2009) key factors for an effective pretreatment of lignocellulosic biomass are highly digestible pre-

Fig. 4. Glucose concentration and cellulose conversion as function of biomass/ cellulose loading for organosolv (A), LHW (B), and soda (C) pretreated rye straw; enzymes: Celluclast 1.5L and Novozyme 188, enzyme loading = 13 FPU/g cellulose, t = 48 h, T = 323 K.

treated solid fractions, high lignin recoveries and high sugar concentrations. Suitable enzyme loadings were found in the previous chapter (Celluclast 1.5L: 13 FPU/g cellulose; P. janthinellum: 28 FPU/g cellulose). The influence of biomass/cellulose loading on the final glucose concentrations and the conversion of cellulose into glucose are depicted in Fig. 4.

5226


The limit was set to a biomass loading equal to 15% (w/w) LHW treated biomass as recommended by Kristensen et al. (2009). At biomass loadings higher than 4% (w/w), the conversion decreases for all pretreatments. This can be attributed to the so-called solids effect: the conversion decreases mainly due to the incomplete adsorption of cellulases on the biomass at high biomass concentrations (Kristensen et al., 2009). Increasing biomass/cellulose loadings of organosolv and LHW pretreated rye straw lead to increasing glucose concentrations after 48 h (Fig. 4A and B). However, glucose concentrations above 50 g/L are achieved at the expense of a reduction in conversion. The solids effect can be observed for all pretreated substrates using Celluclast 1.5L and Novozyme 188 as well as the enzyme complex from P. janthinellum (see Supplementary data). The conversion of organosolv pretreated rye straw (compare Fig. 4A) decreases at higher cellulose loadings compared to the other two pretreated substrates, consequently at higher glucose concentrations. This can be attributed to the excess of enzymes, which counteracts the solid effect for the conversion of organosolv pretreated rye straw. Comparable digestibility for organosolv pretreated wood using Celluclast 1.5L and Novozyme 188 has been reported previously by Zhang et al. (2009b) and Pan et al. (2005). For LHW pretreated rye straw, the final sugar concentration is around 90 g/L at 17% biomass loading after 48 h. High conversions using LHW pretreated substrate at low biomass loadings between 1% and 2% have also been reported by Liu and Wyman (2005). An almost linear decrease of the conversion of LHW pretreated wheat straw with increasing biomass loading was reported by Jørgensen et al. (2007b). Although biomass loadings up to 40% were used, final sugar concentrations above 84 g/L were not achieved. If LHW pretreated rye straw is enzymatically hydrolysed for 72 h (see Supplementary data), a cellulose conversions can be observed, which is almost identical to the conversion of organosolv pretreated biomass after 48 h (Fig. 4A). Thus sugar concentrations higher than 100 g/L (Table 5) resulting in a conversion around 90% (w/w) are achieved in this work for a biomass loading of 15% (w/w). The maximum glucose concentration of soda pretreated rye straw (compare Fig. 4C) is 50 g/L. Even for an incubation time of 72 h, glucose concentrations do not exceed 60 g/L. In agreement with the result presented in this work, Chen et al. (2009) reported a conversion of alkaline pretreated corn stover of 99%) from Sigma Aldrich, were used as standards to prepare model solutions and to determine the composition of the liquid phase through high performance liquid chromatography (HPLC). Gallic acid monohydrate (>98%), Folin & Ciocalteu’s phenol reagent and sodium carbonate (>99%) from Sigma Aldrich were used for total phenolics determination. MilliQUltrapure water was used. Enzymatic complexes NS50013 (cellulase) and NS50010 (b-glucosidase) were kindly donated by Novozymes (Denmark). Xylanase was not supplemented, as NS50013 complex has been reported to show xylanase activity of 55 U/mL (Zhao et al., 2011; Travaini et al., 2013). 2.2. Pretreatments of wheat straw Four different thermal and thermochemical pretreatments were applied in this study: thermal autoclaving (A), dilute HCl solution autoclaving (B), dilute NaOH solution autoclaving (C), and alkaline peroxide (D), using previously published conditions. In autoclave pretreatments A, B, and C, milled and dry wheat straw was slurried for 5 min with distilled water (Cao et al., 2012b), 1.5% w/w HCl solution (Sun and Cheng, 2005), and 1% w/w NaOH solution (Akhtar et al., 2001), respectively, in a 500 mL screw cap bottle in a solid:liquid ratio of 1:10 w/w, and then autoclaved at 121 °C for 60 min. In alkaline peroxide pretreatment (D), milled and dry wheat straw was slurried for 5 min with 5% w/w H2O2, in a solid: liquid ratio of 1:20, the pH was then adjusted to 11.5 with 2 M NaOH and the mixture was placed in a rotatory shaker at 50 °C and 120 rpm for 60 min (Karagöz et al., 2012). After cooling down to room temperature, the slurry obtained from each pretreatment was recovered and residual solid was separated by filtration and dried in an oven at 45 °C for 48 h. Liquid fraction was stored in a refrigeration chamber for composition analysis. Half of the dry solid fraction from each pretreatment was washed with distilled water in a solid:liquid ratio of 1:10, at 120 rpm, for 60 min at room temperature. It was then filtered and dried in an oven at 45 °C for 48 h. Samples from washing liquid

70

C. Toquero, S. Bolado / Bioresource Technology 157 (2014) 68–76

and solid fractions were collected for composition analysis. Unwashed or washed solid (hereinafter denoted as A, B, C, D or AW, BW, CW, DW, respectively, depending on pretreatment) was used as a substrate in the subsequent enzymatic hydrolysis stage. All experiments were conducted in triplicate and the results were averaged. 2.3. Enzymatic hydrolysis of pretreated wheat straw Enzymatic hydrolysis of untreated raw material and unwashed and washed pretreated solids was performed in 100 mL Erlenmeyer flasks containing 10% w/w dry solid, and a mixture of 10 FPU g 1 (NS50013) and 10 CBU g 1 (NS50010) of cellulose (dry basis) (Travaini et al., 2013). A solution of succinic acid 0.2 M + NaOH 0.2 M was used for pH adjustment (5.0). Hydrolysis was performed in a rotatory shaker at 50 °C and 300 rpm for 72 h. After hydrolysis, samples were collected and stored for analysis of sugars and possible inhibitory compounds. Whole hydrolysates were used as a substrate in the subsequent fermentation stage. All experiments were carried out in triplicate and average data are shown. 2.4. Fermentation with P. stipitis 2.4.1. Microorganism and growth culture The yeast, a culture of P. stipitis DSM 3651 obtained from the German Collection of Microorganisms and Cell Cultures, was kept at 4 °C on YEPX agar plates containing 10 g/L yeast extract, 20 g/L peptone, 20 g/L xylose, and 20 g/L agar. For inoculum preparation, P. stipitis was supplemented from YEPX agar plates to a solution containing 10 g/L yeast extract and 20 g/L peptone (previously sterilized at 121 °C for 20 min in an autoclave) and 20 g/L xylose, which was added after being filtered (0.20 lm Sterile Filters, Ministart Sartorius) to prevent sugar degradation by heat. This inoculum was grown aerobically on a rotatory shaker at 175 rpm and 30 °C for 24 h. 2.4.2. Fermentation with model substrate media and real hydrolysates The effect on fermentation yield of possible inhibitory compounds forming during the various pretreatments was studied. The composition of model fermentation medium was 20 g/L peptone, 10 g/L yeast extract, 12.8 g/L KH2PO4, 0.51 g/L Na2HPO4, 0.47 g/L (NH4)2SO4 and 0.47 g/L MgSO47H2O (Geetha and Gopalakrishnan, 2011). Medium was adjusted to pH 5 with a buffer comprising succinic acid 0.2 M (250 mL/L medium) and NaOH 0.2 M (267 mL/L medium), and was autoclaved at 121 °C for 20 min. As a carbon source, glucose and xylose (in concentrations of 35 g/L and 20 g/L, respectively) were added to the model medium by sterile filtration (0.20 lm Sterile Filters, Ministart Sartorius). Inhibitory compounds (formic acid, acetic acid, HMF, and furfural) were also added to the medium, individually or as mixtures to simulate similar inhibitor concentration to that found in wheat straw hydrolysates. A control without inhibitors was also conducted. Model fermentation experiments were carried out with no oxygen supply in sterile 125 mL serum bottles with cap and needle to remove CO2. Each serum bottle was filled with 25 mL of fermentation medium inoculated with 10% (v/v) growth culture. Real hydrolysates were sterilized at 80 °C for 20 min before fermentation. This temperature is enough to avoid the presence of microorganisms other than P. stipitis while at the same time preventing sugar degradation. Substrates from enzymatic hydrolysis were directly inoculated with 10% (v/v) growth culture. Fermentation of real hydrolysates was performed in the same Erlenmeyer flasks as enzymatic hydrolysis, with cap and needle to remove CO2. Fermentation of both model media and real hydrolysates was performed in a rotatory shaker at 30 °C and 175 rpm for 168 h,

these conditions being selected in previous works (Bellido et al., 2011). After fermentation, samples were collected for composition analysis. All the experiments were carried out in triplicate and the average data are shown. 2.5. Analytical methods Moisture, extractives, ash, lignin, cellulose (as glucose), and hemicellulose (as xylose) in raw and pretreated materials were analysed following NREL laboratory analytical procedures. High performance liquid chromatography (HPLC) was used to measure glucose, xylose, ethanol, formic acid, acetic acid, HMF and furfural in samples, using a Bio-Rad HPX-87P ion-exclusion column, a Waters 2414 refractive index detector and 0.01 N H2SO4 as mobile phase at a flow ratio of 0.6 mL/min and 60 °C. The total content of phenolic compounds in samples was determined by the Folin–Ciocalteu method (Singleton et al., 1999) with gallic acid as calibration standard. All the samples were centrifuged at 5000 rpm for 5 min and filtered before being analysed. All analyses were performed in duplicate. A scanning electron microscope (SEM, Quanta 200FEG, FEI) was used to show the microscopic structure of raw and pretreated wheat straw. 3. Results and discussion 3.1. Effect of pretreatments on sugar solubilisation and pretreated solid composition The percentages of sugars solubilised during the four pretreatments are shown in Table 1. In all cases, the vast majority of cellulose remained in the solid fraction, releasing into pretreatment liquids from 0.67 to 2.26 g glucose/100 g RM. Xylose release was also low for pretreatments A, C, and D, with values below 2 g/ 100 g RM. However, dilute acid pretreatment (B) provided a 16.49 g/100 g RM xylose release, indicating, as expected with this kind of pretreatment, very high hemicellulose solubilisation (around 74.4% of the xylose contained in raw material). These results are within the range of certain values reported in the literature, e.g. glucose release of 2.6% w/w of initial wheat straw (Marcotullio et al., 2011), 0.5% of glucan per dry weight of rice straw before pretreatment (w/w) (Park et al., 2010), and 0.5 g/L of glucose in liquid from 10% (w/v) rapeseed straw pretreatment (Luo et al., 2011) in comparable dilute-acid, alkaline, and alkaline peroxide pretreatments, respectively. Likewise, in terms of xylose, said authors reported 20.1% w/w of initial wheat straw (around 93.4% of the xylose in raw material), 1.3% w/w per dry weight of rice straw, and 3 g/L in the liquid phase after the above-mentioned pretreatments. It is also worth noting that Herrera et al. (2004) reported concentrations of 17.3 g xylose/L and 3.8 g glucose/L in the liquid phase when pretreating sorghum straw with 2% HCl at 100 °C for 300 min at 10% solid loading. Washing liquids also presented low glucose and xylose release (Table 1), with values below 0.90 g/100 g RM in all cases except for xylose in BW, which reached a higher content due to solubilisation of hemicelluloses produced by this type of pretreatment. In terms of solid fraction composition, different variations were observed in cellulose, hemicellulose, and acid insoluble lignin (AIL) compared to the raw material depending on the pretreatment. These results are shown in Table 1. For all the pretreatments tested, cellulose solubilisation and degradation proved much lower than was experienced by hemicellulose and lignin, resulting in a relative increase in cellulose percentage in the solid compared to RM. The higher the solubilisation and reduction of hemicelluloses or lignin, the greater the cellulose composition increase of pretreated solid. The relative percentages of hemicellulose and AIL

71


Table 1 Main components of solid and liquid fractions after different pretreatments of wheat straw and after washing. A: thermal autoclaving; B: dilute acid autoclaving; C: dilute alkali autoclaving; D: alkaline peroxide pretreatment. When followed by W, they refer to the washing liquid or the solid fraction after washing. Sample

A AW B BW C CW D DW

Sugars in pretreatment/washing liquid (g/100 g raw material)

Components of the solid fraction (%)

Cellulose (as glucose)

Hemicellulose (as xylose)

Cellulose (as glucose)

Hemicellulose (as xylose)

Acid insoluble lignin

0.67 ± 0.06 0.13 ± 0.05 1.25 ± 0.11 0.27 ± 0.02 0.96 ± 0.03 0.19 ± 0.03 2.26 ± 0.10 0.67 ± 0.06

0.81 ± 0.01 0.20 ± 0.03 16.49 ± 3.00 2.08 ± 0.62 1.03 ± 0.12 0.14 ± 0.05 1.78 ± 0.07 0.83 ± 0.01

38.66 ± 0.14 38.15 ± 0.20 59.83 ± 0.57 64.47 ± 1.59 50.67 ± 0.81 56.84 ± 0.57 48.86 ± 1.59 52.21 ± 0.70

20.88 ± 0.73 20.11 ± 0.23 5.14 ± 0.31 3.84 ± 0.79 19.70 ± 0.88 20.25 ± 0.66 25.10 ± 0.71 23.59 ± 0.41

19.03 ± 1.12 20.59 ± 0.56 18.98 ± 0.77 19.93 ± 0.17 8.90 ± 1.56 8.93 ± 0.70 15.17 ± 1.10 15.40 ± 0.21

decreased in certain cases and increased in others, in relation to RM, depending on the solubilisation effect of pretreatments. The effect of washing pretreated biomass on solid composition was almost negligible for pretreatment A, and caused slight changes in methods B, C, and D, with drag of sugars in washed water also proving negligible. As regards straw pretreated using method A, the percentage of cellulosic fraction and AIL underwent a slight increase of 9.9% and 4.7%, respectively, compared to RM, while hemicellulose composition dropped by only 5.7%. The effect of pretreatment A was minimal and little AIL was removed. Cellulose and hemicellulose might still therefore be linked to a large amount of lignin, which would hinder subsequent enzymatic hydrolysis and fermentation stages. Cao et al. (2012b) reported that merely autoclaving for 60 min and with a temperature of 121 °C was not sufficient to solubilise the hemicellulose and remove the lignin, as they obtained very similar compositions of raw and autoclave-pretreated sweet sorghum bagasse, with cellulose, hemicellulose, and AIL compositions of 49.78%, 27.72%, and 10.83%, and 54.40%, 22.44%, and 10.71%, respectively. The percentage of cellulose in samples from pretreatment B underwent a sharp increase of 70.0%, a rise which was due to the drastic decrease of 76.8% in hemicellulose composition caused by the high solubilisation of xylose with this pretreatment. However, AIL composition increased slightly by 4.4% compared to the raw material composition. Cao et al. (2012a) reported similar results when applying 0.5% w/w H2SO4 at 170 °C for 30 min to Populus trichocarpa, since pretreated biomass showed an increase of 74.2% in glucan composition, caused by the drastic 92.5% reduction in xylose. Lignin also experienced a slight increase of 1.6%. A milder increase in cellulose composition (36.1%) was obtained by Wei et al. (2012), with a comparable reduction of 91.2% in hemicellulose composition but an increase of 14.8% in lignin percentage when eucalyptus chips were subjected to 0.75% w/w H2SO4 at 160 °C for 10 min. The solid fraction from pretreatment C showed a significant 44.0% increase in cellulose composition due to the sharp 51.0% decline in the percentage of AIL, and the less marked 11.1% decrease in hemicellulose composition. These results are within the range of those obtained by Cao et al. (2012b) in terms of cellulose, since they reported a 57.6% increase in cellulose composition, but higher decreases of 84.5% and 45.6% in lignin and hemicellulose compositions, respectively, when comparing raw and pretreated sweet sorghum bagasse (autoclaved with 2% NaOH at 121 °C for 60 min). Applying the same pretreatment as method C from this study to rice straw, but for 30 rather than 60 min, Oberoi et al. (2012) also reported a 47.6% increase in cellulose percentage, as well as decreases of 43.0% and 47.5% in hemicellulose and lignin percentages, respectively. Dilute alkali pretreatment has been found to partially solubilise hemicelluloses and to cause swelling, resulting in

increased biomass porosity and internal surface, as well as disruption of lignin structure (Kang et al., 2012; Cao et al., 2012b). Pretreatment D provided a 38.8% increase in cellulose composition, a 13.3% increase in the percentage of hemicelluloses, and a substantial 16.5% reduction in AIL composition. The same pretreatment method was applied to rapeseed straw by Karagöz et al. (2012), who reported a slight increase of 14.7% in cellulose composition, and a decrease of 18.0% and 21.1% in hemicellulose and lignin compositions, respectively, delignification proving similar to that obtained in this study. Cao et al. (2012b) performed a variation of method D, which consisted of mixing sweet sorghum bagasse at room temperature and 10% solid loading, first with 2% NaOH for 2 h and later with 5% H2O2 for 24 h in a dark place. Cellulose composition of pretreated biomass increased by 45.5%, whereas hemicellulose and AIL compositions fell by 36.8% and 78.9%, respectively. Method D might be considered moderate compared to methods A, B, and C, due to the mild temperature applied, although it has been found to be quite effective in lignin solubilisation as well as beneficial for retaining cellulose and hemicellulose. Some results have been found for other pretreatments applied to wheat straw. Ballesteros et al. (2006) applied steam explosion under different conditions to wheat straw, reporting that cellulose and lignin compositions experienced sharp increases, whereas hemicellulose was almost completely solubilised. García-Cubero et al. (2012) studied the delignification of wheat straw when applying ozone pretreatment in a fixed bed reactor, obtaining high reductions in AIL content, with negligible cellulose and hemicellulose solubilisation. In short, from the standpoint of sugar recovery in pretreated solids, methods A, C, and D would seem advisable as the pretreatment methods before enzymatic hydrolysis of the solid phase, as sugar losses in the pretreatment liquid were assumable and the amount of cellulose and hemicellulose retained in the solid fraction proved extremely high. However, when using method B, a large amount of the initial xylose was released in the pretreatment liquid and the use of the whole slurry rather than separated solid fraction would be necessary (Kont et al., 2013), as long as there is no negative influence of inhibitory compounds. With regard to lignin removal, the most effective method was C. In any case, the results from enzymatic hydrolysis and fermentation need to be analysed in order to compare the effect and efficiency of the pretreatments applied. 3.2. SEM photo analysis Supplementary Figs. 1 and 2 show scanning electron microscope photos of raw and pretreated wheat straw. In Supplementary Fig. 1, with 100 magnification, it is possible to compare the fibre of raw material (a), which is intact, with different destructive effects on fibres of pretreated samples (b), (c), (d), and (e). Thermal

72


autoclaving (b) caused little change in fibres, thus reaffirming the idea that this pretreatment was not severe enough to remove lignin and modify lignocellulosic structure (Cao et al., 2012b). HCldilute autoclaving (c) caused insignificant delignification (Pu et al., 2013), with slight structural changes being observed in the fibre despite the high hemicellulose solubilisation. The destructive effect of NaOH-dilute autoclaving (d) on fibres was considerable due to high removal of lignin and xylan side chains (Park et al., 2010). Biomass subjected to alkaline peroxide (e) evidenced greater porosity and less structured fibres, although polysaccharide chains remained less deteriorated (Monte et al., 2011). When observing Supplementary Fig. 2, with 1000 magnification, differences in morphology, surface area, and porosity are clearer. Thermal autoclaving (b) left a surface layer of re-deposited polymers on the fibre, wherein no microfibrils are visible and some distinctive globular lignin deposits appear, thus perhaps indicating an incipient disruption of hemicellulose and the onset of the delignification process. HCl-dilute autoclaving (c) acted mechanistically on hemicellulose making the overall structure more porous and disorganized due to hemicellulose disruption. NaOH-dilute autoclaving (d) promoted the separation of some fibers, thus increasing the external surface area and the porosity. The breakage could be clearly observed, resulting in a rough surface wherein most of the structure of cellulose was preserved and there were some hollow areas that could increase the accessibility of enzymes. Although it was a mild pretreatment condition, alkaline peroxide (e) left a surface with re-deposited polymers and full of droplets and globular deposits, similar to those seen on wheat straw pretreated by thermal autoclaving (b), which indicates some delignification and preservation of cellulosic chains.

3.3. Inhibitor concentration in hydrolysates The presence and concentration of inhibitory compounds can significantly affect enzymatic hydrolysis and fermentation processes and is key to ethanol production. Table 2 summarises inhibitor concentration in wheat straw hydrolysates after 72 h of enzymatic hydrolysis. The only inhibitory compounds found in hydrolysates from RM were acetic acid, a typical by-product from enzymatic hydrolysis, and phenolics, which are degradation products of lignin origin. Formic and acetic acids and phenolic compounds appear in the hydrolysates of all the pretreated samples whereas HMF and furfural were only detected in hydrolysates from sample B. The highest concentrations of inhibitory compounds were detected in samples C, with 1.35, 2.36, and 2.13 mg/g RM of formic acid, acetic acid and total phenolics, respectively. The water washing removed completely formic and acetic acids, furfural and HMF, except for CW samples with some acetic acid remaining, but in a lower

concentration than the control; and reduced the phenolic concentration of all the washed samples. Inhibitor concentrations obtained in this study are in the range of some results found in the literature. Díaz et al. (2009) obtained higher concentrations of formic acid, acetic acid, and furfural (2.53, 6.81, and 3.42 g/L, respectively) in initial hydrolysates when applying 1% w/w H2SO4 at 190 °C for 10 min to olive tree biomass at 20% solid loading. Herrera et al. (2004) reported 2.5 g acetic acid/L and 1.0 g furfural/L in hydrolysates from sorghum straw pretreated with 2% HCl at 100 °C for 300 min at 10% solid loading. No furfural and HMF were detected by Díaz et al. (2013) when applying alkaline peroxide to rice hulls. Hendriks and Zeeman (2009) claimed that alkali and oxidative pretreatments do not promote the formation of such inhibitors, which is consistent with the results obtained from methods C and D. About other pretreatments, Bellido et al. (2011) obtained 1.56, 0.05, and 0.16 g/L of acetic acid, HMF, and furfural, respectively, in whole hydrolysates from wheat straw pretreated by steam explosion at 210 °C and 10 min. Alvira et al. (2013) obtained 0.371, 1.731, 0.069, 0.209, and around 2.7 g/L of formic acid, acetic acid, HMF, furfural and total phenolics, respectively, in hydrolysates at 17% solid loading from wheat straw pretreated by steam explosion at 210 °C and 2.5 min.

3.4. Glucose and xylose release yields in enzymatic hydrolysis Fig. 1 shows the effect of different pretreatments on the amount of glucose and xylose released from unwashed and washed biomass after 72 h of enzymatic hydrolysis. Glucose and xylose yields were calculated as the percentage of cellulose or hemicellulose present in the initial hydrolysis material. Glucose and xylose concentrations in the hydrolysates were higher than those of the control (RM) for all the experiments, except for xylose concentration in hydrolysates from dilute-acid pretreatment, due to high xylose solubilisation during this pretreatment. Glucose and xylose release was very low in hydrolysates from pretreatment A and was similar for washed and unwashed samples. These poor results are consistent with the low acid insoluble lignin removal obtained using this pretreatment, which hindered cellulose and hemicellulose hydrolysis, therefore yielding low sugar release. Cao et al. (2012b) obtained similar results with glucose and xylose concentrations of 1.46 g/L and 0.62 g/L, respectively (corresponding to a total sugar yield of 10.12 g sugar/100 g RM) in hydrolysates of sweet sorghum bagasse pretreated by thermal autoclaving. The washing effect on enzymatic hydrolysis is very clear for glucose release after dilute acid pretreatment. Washing the pretreated straw had a significant effect on glucose recovery, where glucose concentration in washed samples was more than double that of

Table 2 Inhibitors concentration in hydrolysates from raw and both unwashed and washed (W) wheat straw subjected to different pretreatments (A, B, C and D) prior to fermentation. A: thermal autoclaving; B: dilute acid autoclaving; C: dilute alkali autoclaving; D: alkaline peroxide pretreatment. Inhibitors in hydrolysates (g/L)

Natural Straw (RM) A AW B BW C CW D DW

Formic acid

Acetic acid

HMF

Furfural

Total phenolics

0.00 ± 0.00 0.07 ± 0.01 0.00 ± 0.00 0.18 ± 0.01 0.00 ± 0.00 2.06 ± 0.16 0.00 ± 0.00 0.62 ± 0.11 0.00 ± 0.00

0.64 ± 0.14 1.02 ± 0.23 0.00 ± 0.00 0.82 ± 0.14 0.00 ± 0.00 3.59 ± 0.09 0.49 ± 0.03 0.69 ± 0.04 0.00 ± 0.00

0.00 ± 0.00 0.00 ± 0.00 0.00 ± 0.00 0.04 ± 0.03 0.00 ± 0.00 0.00 ± 0.00 0.00 ± 0.00 0.00 ± 0.00 0.00 ± 0.00

0.00 ± 0.00 0.00 ± 0.00 0.00 ± 0.00 0.84 ± 0.37 0.00 ± 0.00 0.00 ± 0.00 0.00 ± 0.00 0.00 ± 0.00 0.00 ± 0.00

0.42 ± 0.02 0.50 ± 0.01 0.15 ± 0.02 0.96 ± 0.02 0.37 ± 0.03 3.24 ± 0.07 1.23 ± 0.04 0.29 ± 0.01 0.14 ± 0.02


73

Fig. 1. Sugars concentration in hydrolysates prior to fermentation, and ethanol concentration after 168 h of fermentation with Pichia stipitis of raw and both unwashed and washed (W) wheat straw subjected to different pretreatments (A, B, C and D). (A) Thermal autoclaving; (B) dilute acid autoclaving; (C) dilute alkali autoclaving; (D) alkaline peroxide pretreatment.

non-washed samples. Dilute acid hydrolysates contained furfural, acetic acid, formic acid, and HMF, although no detectable quantity of these compounds was found in the hydrolysates from washed pretreated solids. These compounds may be inhibitory for hydrolytic processes, causing changes in substrate reactivity or decreasing the adsorption of enzymes (Kont et al., 2013). Frederick et al. (2013) recovered 5.3 times more glucose when washing dilute acid pretreated poplar biomass. In any case, sugar concentration in hydrolysates was low with this pretreatment. Glucose release yield was only 19.6% for the washed material. Despite the higher xylose recovery yield (around 40% for both washed and unwashed material), xylose concentration in hydrolysates was, as expected, very low since most of the xylose was solubilised during pretreatment. Panagiotopoulos et al. (2009) obtained higher sugar releases when barley straw pretreated with 1% w/w HCl at 160 °C for 30 min was enzymatically hydrolysed for 24 h, concentrations being 16.9 g/L for glucose and 7.0 g/L for xylose in hydrolysates. Likewise, Wei et al. (2012) reported 11.09 g of glucose and 0.49 g of xylose per 100 g raw material after 60 h enzymatic hydrolysis of eucalyptus chips pretreated at 140 °C, and obtained higher glucose yields when increasing pretreatment temperature, for example, 32.19 g of glucose/100 g raw material at 160 °C. Glucose and xylose concentrations in the hydrolysates of wheat straw pretreated with method C were higher than with methods A and B. Around 26.8% of glucose and 52% of xylose release yields were obtained despite the substantial formic acid, acetic acid and phenolic compounds concentrations in samples from unwashed hydrolysates. This rise in hydrolysis yields is due to the effect of this pretreatment on biomass porosity and internal surface, caused by the high removal of lignin and partial solubilisation of xylan side chains (Park et al., 2010), thus enhancing enzyme attack. Cao et al. (2012b) obtained 12.55 g/L of glucose and 4.98 g/L of xylose in hydrolysates of alkali-pretreated sweet sorghum bagasse, and Akhtar et al. (2001) reported around 30% percentages of saccharification when enzymatically hydrolysing for 20 h wheat straw that had been pretreated with 1% w/w NaOH at 121 °C for 4 h. For dilute alkali pretreatment, the washing stage increased glucose and xylose release yields to 35.2% and 54.6%, respectively. Compared to dilute acid results, furfural, and HMF seemed to exhibit a higher enzymatic hydrolysis inhibitory effect than formic and acetic acids or phenolic compounds. The highest glucose concentrations were found in hydrolysates from wheat straw pretreated with method D, with 48.5% and 63.7% of glucose release yield for unwashed and washed samples, respectively. Alkaline peroxide pretreatment is good for retaining cellulose and opening the lignocellulosic structure, even maintaining

relatively high lignin concentrations. Xylose concentration in hydrolysate D was lower than in hydrolysate C, yet still proved quite significant. Xylose losses were slightly higher during alkaline peroxide pretreatment, and xylose release yield was a little lower (48%) than in hydrolysates from pretreatment C. The best xylose release yield (55.2%) and the highest xylose concentration were obtained in hydrolysates of washed materials from alkaline peroxide pretreatment. The inhibition effect on enzymatic hydrolysis was very similar to that found in alkaline pretreatment, despite the lower concentrations of formic acid, acetic acid, and phenolic compounds found in hydrolysates from pretreatment D. A possible explanation is the end-product inhibition effect of already soluble sugars (as glucose and cellobiose) on cellulolytic enzymes (Alvira et al., 2013). Karagöz et al. (2012) reported 15.08 g/L of glucose and 8.29 g//L of xylose when applying the same pretreatment to rapeseed straw, and Cao et al. (2012b) obtained 10.41 g/L of glucose and 4.98 g/L of xylose in hydrolysates of sweet sorghum bagasse subjected to alkaline peroxide pretreatment. In both cases, xylose concentration was considerably lower than was obtained in this study in hydrolysates from alkaline peroxide treated wheat straw. As regards the results from enzymatic hydrolysis of wheat straw, the concentrations obtained in this study from both unwashed and washed alkaline peroxide pretreated biomass, 35.73 and 45.57 g/L of total sugars, respectively, were considerably higher than those reported by Ballesteros et al. (2006), who obtained 23 g glucose/100 g raw material in hydrolysates from wheat straw pretreated by steam explosion. Bellido et al. (2011) obtained 23.48 and 6.21 g/L of glucose and xylose, respectively, in hydrolysates from wheat straw pretreated by steam explosion, and García-Cubero et al. (2012) reported around 40% and 32% glucose and xylose yields, respectively, in enzymatically hydrolysed wheat straw previously subjected to ozonolysis. 3.5. Fermentation to ethanol 3.5.1. Effect of inhibitors on model media fermentation The performance of P. stipitis was evaluated on model fermentation media containing inhibitory compounds as well as glucose and xylose as carbon sources. The influence of inhibitors, individually or as a mixture, was studied in terms of ethanol production, considering two parameters: theoretical ethanol yield, defined as the percentage of stoichiometric ethanol obtained from all sugars available in the initial fermentation material; and ethanol yield, defined as the ratio between ethanol production and sugar consumption. Table 3 summarises inhibitor concentrations in model solutions, as well as ethanol concentrations and concentrations

74


of residual sugars obtained in such solutions after 168 h fermentation and the parameter biomass yield (YX/S, g biomass/g substrate) calculated when the exponential growth was finished. In this work, the individual effect of inhibitory compounds was only studied for formic and acetic acids, results with HMF and furfural were already reported by Bellido et al. (2011) under identical fermentation conditions as those used in this study. Media called SimB, SimC, and SimD simulated the composition of hydrolysates from methods B, C, and D, in terms of formic and acetic acids, HMF and furfural. With regard to the individual effect of formic acid, almost complete inhibition of ethanol production was observed in medium containing 2 g/L of formic acid. When formic acid was present in the medium, theoretical ethanol yield decreased from 77.9%, corresponding to the control, to 61.2% and 7.8% in media containing 1 and 1.5 g/L of formic acid, respectively. The same decrease was obtained for ethanol yield, which was 0.40 g ethanol/g sugars for the control, while 0.5 and 1 g/L formic acid medium gave 0.34 g ethanol/g sugars, falling to 0.13 and 0.03 g ethanol/g sugars for 1.5 and 2 g/L formic acid, respectively. Díaz et al. (2009) developed a similar study of inhibition of P. stipitis on 20 g/L glucose and 15 g/ L xylose synthetic medium and reported that ethanol yield after 178 h fermentation (when xylose was exhausted) was null when formic acid was at a concentration of 2 g/L. Experiments performed with 1 g/L of this inhibitor resulted in 0.35 g ethanol/g sugars of ethanol yield after 48 h fermentation. Yeasts may suffer stress as a result of high osmotic pressure or high concentrations of inhibitory compounds and the combination of these factors can act synergistically, affecting ethanol yields (Hoyer et al., 2013). The inhibitory effect of acetic acid on ethanol generation was similar to that of formic acid. Theoretical ethanol yield decreased with acetic acid concentration to 60.1%, 51.9%, and 35.9% for media containing 0.5, 1, and 2 g/L acetic acid, respectively. The medium containing 4 g/L of acetic acid displayed almost complete inhibition of ethanol production with 2.1% theoretical ethanol yield. The trend in ethanol yield was also similar to experiments with formic acid with 0.05 g ethanol/g sugars for 4 g/L acetic acid medium and values around the control (0.40 g ethanol/g sugars) for the remaining samples. These results are in agreement with those obtained by Bellido et al. (2011), who reported complete inhibition of both cellular growth and ethanol production using P. stipitis, when media contained 3.5 g/L of acetic acid. As regards the combined effect of inhibitors: SimB, containing different concentrations of formic and acetic acids, HMF and furfural, presented a theoretical ethanol yield of 50.2% and an ethanol yield of 0.29 g ethanol/g sugars. This result is consistent with the inhibitory behaviour of acetic acid at a concentration of 1 g/L. This slight reduction in ethanol production may be attributable to a cumulative effect of all the inhibitors present in the broth,

although this effect was small since the concentration of formic acid in the medium was low. As shown by Bellido et al. (2011), HMF may even enhance cell growth and ethanol production, and furfural has hardly any inhibitory effect on P. stipitis (theoretical ethanol yields of 82.5% and 74.5% were obtained when 0.1 g/L HMF and 1 g/L furfural were present in the medium, while control yielded 77.9%). Díaz et al. (2009) reported an ethanol concentration of 12.5 g/L and an ethanol yield of 0.40 g ethanol/g sugars when 1.5 g/L acetic acid, 0.5 g/L formic acid, and 1 g/L furfural were present in fermentation media, these parameters being determined after 24 h fermentation when maximum performance was achieved. As expected, when analysing sample SimC, with 4 g/L of acetic acid and 2 g/L of formic acid, ethanol production was almost negligible, the theoretical and ethanol yields being 1% and 0.03 g ethanol/g sugars, respectively, values that are even lower than those of the individual inhibitor samples. Finally, SimD, with 0.6 g/L of acetic and formic acids, reached 55.9% of ethanol theoretical yield and 0.31 g ethanol/g sugars. This might again be due to a cumulative effect of both compounds. Díaz et al. (2009) combined acetic and formic acids at concentrations of 3 g/L and 1 g/L, respectively, and obtained total inhibition of both cell growth and ethanol production. They also affirmed that acetic acid, formic acid or furfural, when appearing individually and not in a concentration high enough to totally inhibit cell growth, exert a positive effect on ethanol yields. When used together, however, these inhibitors negatively affect ethanol yield. Results from samples SimB, SimC, and SimD served as a basis for comparison with real wheat straw hydrolysates in the following section. 3.5.2. Ethanol production from wheat straw hydrolysates Fig. 1 shows ethanol production after 168 h fermentation of wheat straw hydrolysates. A comparison between RM and unwashed and washed biomass from all pretreatments is also shown in Fig. 1. Ethanol production from samples of both unwashed and washed hydrolysates from all pretreatment methods was higher than that of the control (RM), which was almost negligible, representing 8.2% of theoretical ethanol yield and an ethanol yield of 0.04 g ethanol/g sugars. Sugars remaining after raw material fermentation were negligible, with values being below 0.1 g/L. Ethanol concentrations detected in fermentation liquids of thermal-pretreated samples were extremely low, the difference in the average ethanol concentration in samples A and AW not being statistically significant (p-value >0.05 in the Student t-test). Sugars were completely consumed and ethanol theoretical yields were 76.8% and 81.4% for unwashed and washed experiments, respectively. These results indicate that, despite the low concentration of ethanol obtained, it was produced at high sugar uptake, given

Table 3 Fermentation of model solutions (35 g/L glucose, 20 g/L xylose and inhibitory compounds) using Pichia stipitis. Initial inhibitors concentration, final sugars and ethanol concentrations after 168 h of fermentation, and biomass yield (YX/S), calculated when exponential growth was finished. Sample

Control HFor2 HFor1.5 HFor1 HFor0.5 HAc4 HAc2 HAc1 HAc0.5 SimB SimC SimD

Initial inhibitors concentration (g/L)

Final sugars concentration (g/L)

Formic acid

Acetic acid

HMF

Furfural

Glucose

Xylose

0 2 1.5 1 0.5 0 0 0 0 0.2 2 0.6

0 0 0 0 0 4 2 1 0.5 1 4 0.6

0 0 0 0 0 0 0 0 0 0.1 0 0

0 0 0 0 0 0 0 0 0 1 0 0

0.0 20.9 19.6 0.0 0.0 24.5 13.5 9.5 0.0 0.0 26.8 0.0

0.0 15.8 18.0 3.6 0.8 18.6 15.9 9.7 2.1 6.9 17.1 5.1

Final ethanol concentration (g/L)

YX/S (g/g)

21.9 0.6 2.2 17.2 18.3 0.6 10.1 14.6 16.9 14.1 0.3 15.7

0,12 0,01 0,03 0,07 0,08 0,01 0,03 0,07 0,09 0,07 0,02 0,07

(24 h) (48 h) (48 h) (24 h) (24 h) (48 h) (48 h) (24 h) (24 h) (48 h) (48 h) (24 h)


the low amount of glucose and xylose released in hydrolysates A and AW. The theoretical yields obtained were higher than those expected from the results of model media fermentation, which were 51.9% in the sample containing 1 g/L acetic acid and 77.9% for the control. Cao et al. (2012b) obtained 0.89 g/L of ethanol from hydrolysates of thermally pretreated sweet sorghum bagasse, having worked at 2% of substrate loading in enzymatic hydrolysis. Dilute acid pretreatment provided lower ethanol production than the values predicted by the model media study. The average ethanol concentration in sample B was significantly lower than the average value in sample BW (p-value 0.05) differences in percent xylan solubilization for any of the treatment combinations. The solubilization of glucan during NaOH pretreatment was between 12.82% (1%, 30 min, 121 C/15 psi) and 30.14% (2%, 60 min, 90 C) as illustrated in Fig. 2c. Glucan solubilization increased significantly with increasing concentration for 90 C at 90 min and the temperature effect was significant for 2% NaOH for 30 and 60 min. However,

the standard deviations among some replicates were rather large. This could be attributed to the heterogeneous nature of cotton stalks and the fact that amount of free cotton fiber could vary from one sample to the other.

3.4. Effect of hydrogen peroxide pretreatment Hydrogen peroxide pretreatment utilizes oxidative delignification to detach and solubilize the lignin and loosens the lignocellulosic matrix thus improving enzyme digestibility (Martel and Gould, 1990). The lignin reduction and xylan and glucan solubilization due to H2O2 pretreatment in this study are shown in Fig. 3a–c and the solids recovered after pretreatment are presented in Table 2. There was no evidence (p > 0.05) of any effects of either


3007


40 30 20 10 0 30

60

90

30

90

60

90


121/15 psi


40 30 20 10 0 30

60

90

30

90

60

90

Time (m in) Te m p. (o C)

121/15 psi


40 30 20 10 0 30

60

90

90

30

60 121/15 psi

90


Fig. 3. (a) Lignin reduction, (b) xylan solubilization, and (c) glucan solubilization in hydrogen peroxide pretreated samples as a function of residence time, temperature, and concentration.

of the treatment factors on the lignin content of pretreated solids. Hydrogen peroxide pretreatment led to 6.22% (0.5%, 90 min, 90 C) to 32.01% (2%, 60 min, 121 C/15 psi) delignification (Fig. 3a). These lignin degradations are lower than those reported in literature at alkaline conditions where pretreatment of sugar cane bagasse with 2% alkaline H2O2 resulted in 50% decrease in lignin and solubilization of most of the hemicellulose within 8 h at 30 C (Azzam, 1989). Determination of simple treatment effects for delignification showed that increasing the concentration from 0.5% to 2% did not significantly increase delignification for 30 min at 90 C probably because the residence time was too short at the lower temperature. The simple time effect was significant for 121 C/15 psi at 0.5% and 1% H2O2, which indicates that increasing the residence time

from 30 to 90 min showed significant improvements only for the two lower concentrations at the higher temperature. Temperature played a significant role in improving delignification for 0.5% at 60 min and 2% at 30 and 60 min but an increase in temperature significantly reduced the mean delignification for 0.5% at 90 min. The most severe pretreatment in the autoclave at 121 C for 90 min with 2% H2O2 had lower levels of delignification than the treatments at 30 and 60 min at 0.5% and 1%. This could be attributed to the decomposition of H2O2 at high temperature thus diminishing its oxidative delignification potential and to the long residence time which could result in recondensation or repolymerization of solubilized lignin. The solubilization of xylan due to H2O2 pretreatment averaged between 8.18% (0.5%, 60 min, 90 C) and 30.56% (2%, 30, 121 C/15 psi) (Fig. 3b) while the xylan

3008


content ranged from 8.69% (2%, 30 min, 121 C/15 psi) to 10.87% (2%, 90 min, 90 C). Concentration had a significant effect (p 6 0.05) on xylan solubilization for 90 min, 90 C treated samples and 30 min, 121 C treated samples. The simple temperature effect was significant for xylan solubilization for 0.5% and 1% at 60 min and 2% at 30, 60, and 90 min. The percentage of glucan in the pretreated solids remaining as a result of H2O2 pretreatment ranged from 28.4% (1%, 30 min, 90 C) to 34.1% (2%, 90 min, 90 C). Glucan solubilization on average was between 14.91% (0.5%, 60 min, 90 C) to 29.10% (2%, 30 min, 121 C/ 15 psi) as presented in Fig. 3c. Concentration did not have a significant effect on glucan solubilization. Significant differences (p 6 0.05) between percent glucan solubilization due to changes in temperature were noted for 2% at 90 min, while time played a significant role for 0.5 and 1% H2O2 at 90 C. 3.5. Effect of ozone pretreatment Pretreatment of lignocellulosic biomass with ozone gas has been reported to reduce both the lignin and hemicellulose contents of the treated materials (Ben-Ghedalia et al., 1980). The most substantial effect of ozone pretreatment is on degradation of the lignin polymer, followed by hemicellulose and cellulose solubilization (Quesada et al., 1999). In this study, ozone pretreatment reduced lignin in the range of 11.97–16.6% with no significant difference (p > 0.05) noted for treatment times of 30, 60, and 90 min (Table 3). The amount of xylan solubilized during ozone treatment ranged from 1.9% to 16.7%, while the amount of glucan solubilized was between 7.2% and 16.6%. The percent solubilization of xylan and glucan for 90 min treatment was significantly (p < 0.05) lower than the solubilization for 30 and 60 min. The concentrations of ozone measured in pure water after sparging for 30, 60, and 90 min were 16.96, 17.74, and 18.52 ppm, respectively. Ben-Ghedalia et al. (1980) reported a 50% decrease in both lignin and hemicellulose in ozone treated cotton stalks. Possible explanations for the differences between the results from this study and those from past studies include insufficient treatment times, inadequate ozone concentration, and

poor distribution of ozone gas throughout the cotton stalks because of inefficient sparging. 3.6. Enzymatic hydrolysis Acid pretreated samples resulting in maximum glucose availability (2% H2SO4, 60 min, 121 C/15 psi) were chosen for enzyme hydrolysis. This selection criterion was based on the fact that acid pretreatment has little effect on lignin degradation and the main treatment effect is on hemicellulose and cellulose solubilization. Alkali pretreatment caused delignification and glucan solubilization. The selection for NaOH pretreated samples was based on a compromise between having the lowest percentage of lignin in the pretreated solids, while maintaining a high percentage of glucan (2% NaOH, 60 min, 121 C/15 psi). For hydrogen peroxide, there were no significant differences between percentage glucan, xylan, or lignin in the pretreated solids for any of the treatments. Hence, the treatment with the highest percentage of glucan and the lowest percentage of lignin was chosen (2% H2O2, 60 min, 121 C/15 psi). Cellulose conversion of pretreated samples after 72 h of enzymatic hydrolysis is shown in Table 4. Sodium hydroxide pretreated samples had the highest cellulose conversion of 60.8%, followed by hydrogen peroxide (49.8%) and then sulfuric acid (23.8%). Differences in mean cellulose conversions for all the treatments were statistically significant (p 6 0.05). Hydrolysis of sodium hydroxide pretreated samples resulted in the highest xylan to xylose conversion (Table 4) at 62.57%, whereas hydrogen peroxide averaged 7.78% conversion. For the acid pretreated samples, no xylan was detected in the solids during the initial carbohydrate analysis, but an average of 14.3 mg xylose/g dry biomass was detected in the supernatant after enzymatic hydrolysis. This confirms the hypothesis that there was xylan in the stalks after pretreatment, but the amount was below the detection limit during sugar analysis. Detection of xylose in the hydrolysate may be attributed to a higher sugar concentration resulting from the hydrolyzed sample (5 g) being larger than that analyzed for carbohydrate content of pretreated solids (0.3 g). The difference in cellulose conversion during enzymatic hydrolysis is largely dependent on the difference in lignin

Table 3 Effect of ozone pretreatment on cotton stalks Time (min)

30 60 90c RMSE R-square Tukey’s HSD a b c

Reduction (%)a,b

Solids recovery (%)

Lignin

Xylan

Glucan

11.97 (2.91) 16.63 (2.60) 15.15 (3.02)

16.76 (7.32) 10.61 (8.12) 1.92 (2.89)

16.62 (7.80) 13.74 (3.64) 7.19 (0.36)

7.04 0.52 20.19

5.45 0.42 15.63

2.82 0.46 8.08

Percentages calculated from values on a dry-weight basis. Data are averages of three replicates. Numbers in parentheses represent standard deviations. Only two samples were used for 90 min treatment because the third replicate was an outlier.

90.44 (2.47) 88.66 (2.82) 91.66 (0.47) 2.18 0.32 5.46


3009

Table 4 Glucan and xylan conversion after enzymatic hydrolysis Pretreatment agent

Composition of hydrolysis supernatant and pretreated solida,b,c

Sulfuric acid

– 40.68 (1.44)

1.43 (0.16) 0.00 (0.00)

11.03 (0.66) 46.3 (2.89)

23.85 (1.21)

Sodium hydroxide

– 18.40 (0.16)

8.34 (0.15) 12.13 (0.40)

30.57 (0.56) 50.33 (1.84)

60.79 (2.75)

62.57 (2.57)

Hydrogen peroxide

– 25.59 (2.30)

0.90 (0.14) 10.00 (0.26)

17.21 (0.84) 34.53 (0.86)

49.82 (1.40)

7.78 (1.13)

Lignin

a b c d

Xylose

Glucan conversion (%)

Xylan conversion (%)

Glucose 0.00 (0.00)d

Composition percentages calculated from values on a dry-weight basis. Data are averages of three replicates. Numbers in parentheses represent standard deviations. Compositions of xylose and glucose in the hydrolysis supernatant are in upper rows while compositions of pretreated solids are in bottom rows. See text for explanation.

composition. The sulfuric acid and hydrogen peroxide pretreated samples had 2.2 times and 1.4 times the amount of lignin, respectively, compared to sodium hydroxide pretreated samples. Lignin is not attacked by the enzymes and therefore shields the cellulose during hydrolysis (Mansfield et al., 1999). Solubilization of xylan, on the other hand, seems to have a limited impact on cellulose digestibility. 3.7. Modeling Empirical quadratic models using time, temperature, and concentration as continuous variables and linear models relating a modified severity parameter to these variables were developed to predict xylan solubilization in sulfuric acid pretreatment and lignin reduction in sodium hydroxide pretreatment. These two treatment agents were chosen for modeling because they have been widely studied and seem to be the most promising pretreatments for use on cotton stalks. After eliminating the insignificant terms (p > 0.05) from the model based on the p-values from the Type III Sum of Squares ANOVA table (data not shown), the reduced empirical quadratic model used to quantify the percentage of xylan solubilized from the cotton stalks during sulfuric acid pretreatment was

The appropriate values for C, T, and t were plugged into the equations and the plots between fitted vs. observed values, for both percent xylan solubilization (Eq. (5)) and percent lignin reduction (Eq. (6)), had slopes of 0.97. Both models had high R2-values and slopes close to 1 thus indicating good agreement between the experimental data and the models. Linear models relating a modified severity parameter that combines the effects of time, temperature and concentration to the percentage solubilization of xylan by sulfuric acid pretreatment and to the reduction in lignin by sodium hydroxide pretreatment resulted in R2 values of 0.89 and 0.78, respectively. The n-values for sulfuric acid and sodium hydroxide pretreatments that provided the best model fits while keeping log (M0) positive were 0.849 and 3.90, respectively. The resulting equations were T r 100 0:849 exp M 0 ðsulfuric acidÞ ¼ tC ð7Þ 14:75 T r 100 ð8Þ M 0 ðsodium hydroxideÞ ¼ tC 3:90 exp 14:75 The model equation for determination of xylan solubilization during sulfuric acid pretreatment was developed by plotting log (M0) vs. % xylan solubilization (Fig. 4).

120

þ 0:2644t 22:6728C þ 0:6347CT 11:0451C

2

ð5Þ

The square of the correlation coefficient (R2) for the xylan solubilization model was 0.964. The percent lignin reduction model for sodium hydroxide containing significant terms from the Type III Sum of Squares ANOVA table (data not shown) had an R2 of 0.924 and was given by % lignin reduction ¼ 1:3705 þ 0:0002T þ 0:5554t

y = 53.508x - 55.043 R2 = 0.8926

100 80 60 40 20 0 0.8

1.3

1.8

2.3

2.8

-20 log Mo

þ 49:6254C þ 0:0904Ct 15:9216C 2 0:0047t2

Xylan Solubilization (%)

Xylan solubilization ð%Þ ¼ 117:6194 þ 1:0798T

ð6Þ

Fig. 4. Percent xylan solubilization vs. log (modified severity parameter) for sulfuric acid pretreatment.

3010


70

y = 8.6438x + 33.68 R2 = 0.7826

Lignin Reduction (%)

65 60 55 50 45 40 35 30 25 0

0.5

1

1.5

2 log Mo

2.5

3

3.5

4

Fig. 5. Percent lignin reduction vs. log (modified severity parameter) for sodium hydroxide pretreatment.

M0 was calculated using n = 0.849 and the model is represented by Eq. (9) % xylan solubilization ¼ 53:508 logðM 0 Þ 55:043

ð9Þ

The model equation for the reduction of lignin during sodium hydroxide pretreatment, using n = 3.90 to calculate M0, was obtained from Fig. 5 and is represented as % lignin reduction ¼ 8:6438 logðM 0 Þ þ 33:68

to water at high temperatures. Ozone pretreatment also did not perform as effectively as expected. Possible explanations include insufficient time, low ozone concentration, or uneven distribution of ozone throughout the sample. Compared with other pretreatments, sodium hydroxide pretreatment resulted in significantly (p < 0.05) higher cellulose conversion during the subsequent enzymatic hydrolysis. The empirical quadratic models successfully predicted percent xylan solubilization and percent lignin reduction and may be used in the development of better estimation tools. In addition, this study can serve as a step towards the optimization of pretreatment of cotton stalks. Nevertheless, different combinations of treatment factors, perhaps using higher temperatures or concentrations and application of higher pressures could be investigated. In addition, enzymatic hydrolysis using optimized pretreatment factors and ethanol fermentation need to be studied for bioethanol production since they could not be addressed in this study.

ð10Þ

The modified severity parameter model was validated by plotting the experimental values of percent xylan solubilization and percent lignin reduction against the model predicted values (Silverstein, 2004). The R2 from the plot of experimental vs. predicted % xylan solubilization was 0.88 with a slope of 0.95 indicating good predictive ability of the model. Predicted and experimental values for % lignin reduction during sodium hydroxide pretreatment resulted in an R2 of 0.72 and a slope of 0.99. Variation in predicted and experimental values may have likely been due to heterogeneity of cotton stalks and inability of the modified severity parameter to fully capture dependence of response variables on independent variables in the absence of variables such as stalk to cotton fiber ratio and solids loading. 4. Conclusions Sulfuric acid pretreatment substantially solubilized xylan in cotton stalks and temperature had the most significant effect on xylan solubilization. Data analysis indicated that there is a linearly increasing relationship between xylan solubilization and pretreatment severity. The most significant effect of sodium hydroxide pretreatment was on delignification with concentration of sodium hydroxide being the significant factor. Lignin reduction increased linearly with increase in pretreatment severity of sodium hydroxide. Hydrogen peroxide pretreatment resulted in lower lignin and xylan solubilization than expected. This was probably due to decomposition of hydrogen peroxide

References Agblevor, F.A., Evans, R.J., Johnson, K.D., 1994. Molecular-beam massspectrometric analysis of lignocellulosic materials. I. Herbaceous biomass. J. Anal. Appl. Pyrol. 30, 125–144. Agblevor, F.A., Batz, S., Trumbo, J., 2003. Composition and ethanol production potential of cotton gin residues. Appl. Biochem. Biotechnol. 105, 219–230. Azzam, A.M., 1989. Pretreatment of cane bagasse with alkaline hydrogen peroxide for enzymatic hydrolysis of cellulose and ethanol fermentation. J. Environ. Sci. Health B 24, 421–433. Badger, P.C., 2002. Ethanol from cellulose: a general review. In: Janick, J., Whipkey, A. (Eds.), Trends in New Crops and New Uses. ASHS Press, Alexandria, VA, pp. 17–21. Ben-Ghedalia, D., Shefet, G., Miron, J., 1980. Effect of ozone and ammonium hydroxide treatments on the composition and in vitro digestibility of cotton straw. J. Sci. Food Agric. 31, 1337–1342. Chang, V., Holtzapple, M., 2000. Fundamental factors affecting biomass enzymatic reactivity. Appl. Biochem. Biotechnol. 84–86, 5–37. Chum, H.L., Johnson, D.K., Black, S.K., Overend, R.P., 1990. Pretreatment-catalyst effects and the combined severity parameter. Appl. Biochem. Biotechnol. 24–25, 1–14. Ehrman, T., 1994. Method for determination of total solids in biomass. In: Laboratory Analytical Procedures No. 001. Golden, CO, National Renewable Energy Laboratory. Ehrman, T., 1996. Method for determination of acid-soluble lignin in biomass. In: Laboratory Analytical Procedures No. 004. Golden, CO, National Renewable Energy Laboratory. Gould, J.M., 1985. Studies on the mechanism of alkaline peroxide delignification of agricultural residues. Biotechnol. Bioeng. 27, 225–231. Han, J., Rowell, J., 1997. Chemical composition of fibers. In: Rowell, R., Young, R., Rowell, J. (Eds.), Paper Composites from Agro-Based Resources. CRC Lewis Publisher, New York, pp. 83–134. Hsu, T.A., 1996. Pretreatment of biomass. In: Wyman, C.E. (Ed.), Handbook on Bioethanol: Production and Utilization. Taylor & Francis, Washington, DC, pp. 179–195. Ingram, L.O., Doran, J., 1995. Conversion of cellulosic materials to ethanol. FEMS Microbiol. Rev. 16, 235–241. Kim, S.B., Um, B.H., Park, S.C., 2001. Effect of pretreatment of reagent and hydrogen peroxide on enzymatic hydrolysis of oak in percolation process. Appl. Biochem. Biotechnol. 91–93, 81–94.

R.A. Silverstein et al. / Bioresource Technology 98 (2007) 3000–3011 Mansfield, S.D., Mooney, C., Saddler, J.N., 1999. Substrate and enzyme characteristics that limit cellulose hydrolysis. Biotechnol. Prog. 15, 804–816. Martel, P., Gould, J.M., 1990. Cellulose stability and delignification after alkaline hydrogen-peroxide treatment of straw. J. Appl. Poly. Sci. 39, 707–714. McKendry, P., 2002. Energy production from biomass (part 1): overview of biomass. Bioresour. Technol. 83, 37–46. McMillan, J.D., 1994. Pretreatment of lignocellulosic biomass. In: Himmel, M.E., Baker, J.O., Overend, R.P. (Eds.), Enzymatic Conversion of Biomass for Fuels Production. American Chemical Society, Washington, DC, pp. 292–324. Milne, T.A., Chum, H.L., Agblevor, F.A., Johnson, D.K., 1992. Standardized analytical methods. Biomass Bioenergy 2, 341–366. Moiser, N., Wyman, C., Dale, B., Elander, R., Lee, Y.Y., Holtzapple, M., Ladisch, M., 2005. Features of promising technologies for pretreatment of lignocellulosic biomass. Bioresour. Technol. 96, 673–686. Neely, W.C., 1984. Factors affecting the pretreatment of biomass with gaseous ozone. Biotechnol. Bioeng. 26, 59–65. Overend, R.P., Chornet, E., 1987. Fractionation of lignocellulosics by steam-aqueous pretreatments. Philos. Trans. R. Soc. Lond. 321, 523–536. Quesada, J., Rubio, M., Gomez, D., 1999. Ozonation of lignin rich solid fractions from corn stalks. J. Wood Chem. Tech. 19, 115–137. Ruiz, R., Ehrman, T., 1996. Determination of carbohydrates in biomass by high performance liquid chromatography. In: Laboratory Analytical Procedures No. 002. Golden, CO, National Renewable Research Laboratory. Schell, D.J., Farmer, J., Newman, M., McMillan, J.D., 2003. Dilutesulfuric acid pretreatment of corn stover in pilot-scale reactor – investigation of yields, kinetics, and enzymatic digestibilities of solids. Appl. Biochem. Biotechnol. 105, 69–85. Sharma, R.R., Demirci, A., Beuchat, L.R., Fett, W.F., 2002. Inactivation of Escherichia coli O157:H7 on inoculated alfalfa seeds with ozonated water and heat treatment. J. Food Prot. 65, 447–451.

3011

Shoemaker, S., 2004. Advanced biocatalytic processing of heterogeneous lignocellulosic feedstocks to a platform chemical intermediate (lactic acid ester). Final report for award number DE-FC02-99CH11007. (accessed September, 2006). Silverstein, R., 2004. A comparison of chemical pretreatment methods for converting cotton stalks to ethanol. MS thesis, North Carolina State University. Available from: . Sun, Y., Cheng, J.J., 2002. Hydrolysis of lignocellulosic materials for ethanol production: a review. Bioresour. Technol. 83, 1–11. Tarkov, H., Feist, W.C., 1969. A mechanism for improving the digestibility of lignocellulosic materials with dilute alkali and liquid ammonia. Adv. Chem. Ser. 95 (1), 197–218. TBWEF. 2004. Texas Boll Weevil Eradication Foundation (TBWEF). (accessed 29.09.04). Templeton, D., Ehrman, T., 1994. Determination of acid-insoluble lignin in biomass. In: Laboratory Analytical Procedures No. 003. Golden, CO, National Renewable Energy Laboratory. Torget, R., Walter, P., Himmel, M., Grohmann, K., 1991. Dilute acid pretreatment of corn residues and short-rotation woody crops. Appl. Biochem. Biotechnol. 28–29, 75–86. USDA 2004. National Agricultural Statistics Service Crop Production. (accessed May, 2005). Varga, E., Scengyel, Z., Recaey, K., 2002. Chemical pretreatments of corn stover for enhancing enzymatic digestibility. Appl. Biochem. Biotechnol. 98–100, 73–87. Yang, B., Wyman, C.E., 2004. Effect of xylan and lignin removal by batch and flowthrough pretreatment on the enzymatic digestibility of corn stover cellulose. Biotechnol. Bioeng. 86, 88–95. Yang, B., Boussaid, A., Mansfield, S.D., Gregg, D.J., Saddler, J.N., 2002. Fast and efficient alkaline peroxide treatment to enhance the enzymatic digestibility of steam-exploded softwood substrates. Biotechnol. Bioeng. 77, 678–684.

Journal of Scientific & Industrial Research 778

J SCI IND RES VOL 70 SEPTEMBER 2011

Vol. 70, September 2011, pp. 778-783

Evaluation of different pretreatments to enhance degradation of pine needles by Aspergillus niger F7 under solid state fermentation Nivedita Sharma, Divya Tandon, Rakesh Gupta and Sanjeev Kumar Department of Basic Science, Dr Y S Parmar University of Horticulture and Forestry, Nauni, Solan (HP) 173 230, India Received 20 January 2011; revised 28 July 2011; accepted 01 August 2011 This study presents degradation enhancement of pine needles by Aspergillus niger F7 , isolated from soil. A modified alkali pretreatment [NaOH+H2 O2 (1M); ratio, 9:1] is found the best among all methods when needles are soaked in this solution for 2 h followed by thorough washing with tap water. Degradation of pine biomass was measured in terms of enzyme (cellulase & xylanase) production, biodegradation index (BI) and hydrolysis (%). There are high enzyme units, BI and hydrolysis (%) in pretreated material as compared to untreated one. Keywords: Biodegradation, Lignocellulose, Pretreatment, Solid State fermentation

Introduction Pine (Pinus roxburghii Sarg syn. P. longifolia Roxb.) is a predominant forest species, widely scattered in Alpine range globally. Accumulation of pine needles (PNs) on forest floor leads to infertility of soil and forest fire1 . PNs being rich in cellulose can be used as substrate for biodegradation. Among different physical and chemical pretreatment methods, solid state fermentation (SSF) of lignocellulosic material holds several advantages as compared to submerged fermentation (SmF). In SSF, enzymes produced are many folds more than SmF and thus it has direct impact on biodegradation of biomass2 . This study presents pretreated PNs as a substrate for degradation under SSF by a potential isolate to enhance hydrolysis of PNs. Experimental Section Extractives of Pine Needles (PNs)

PNs were collected from different forest site of northern India. Components [holocellulose (cellulose + hemicellulose), lignin and other extractives] of PNs were estimated by following standard methods of Technical Association of Pulp and Paper industry (TAPPI). For alcohol benzene extraction, oven dried PNs (2 g) were placed in a porous thimble and extractives were derived by TAPPI method3 . For holocellulose extraction, oven

*Author for correspondence Tel: +91-1792-252560; Fax: +91-1792-252242 E-mail: [email protected]

dried PNs (5 g), pre-extracted with alcohol benzene, were taken in 250 ml conical flask, distilled water (160 ml) was added and holocellulose was estimated following TAPPI method4 . For lignin extraction, oven dried PNs (2 g), pre-extracted with alcohol benzene, were treated with 15 ml of 72% sulphuric acid for 2 h at 18-20°C with constant stirring following TAPPI method5 . Pretreatment of Pine Needles (PNs) Grinding

Chipping of PNs gave small pieces, which were grinded (mesh size, 1.5 mm), soaked in water for 24 h, and then air dried for 24 h, followed by drying at 50°C overnight. Completely dried biomass was stored in air tight containers. Alkali Pretreatments

In NH3 pretreatment6 , PNs (10 g) were soaked in 100 ml of 1% ammonia solution. Under NH3 pretreatment (modified), PNs (10 g) were soaked in 100 ml of 5% ammonia solution for 2 h at room temperature (RT) and autoclaved for 15 min. After thorough washing with tap water (until solution became neutral) and dried at 50°C. Under NaOH+ H2 O2 pretreatment (modified), PNs (10 g) were soaked in NaOH+ H2 O2 solution (9:1) for 2 h at RT followed by washing with tap water and dried at 50°C. Acid Pretreatment

In hydrochloric acid (HCl) pretreatment6 , PNs (10 g) were soaked in 1% HCl solution (100 ml) for 2 h.

SHARMA et al: DEGRADATION ENHANCEMENT OF PINE NEEDLES

Table 1—Estimation of holocellulose and lignin in untreated and pretreated biomass of pine needles using TAPPI standard method Treatments Untreated 1%NH3 5%NH3 1%HCL 1%H2 SO4 1%NaOH+H2 O2

Holocellulose % 57.00 85.87 88.92 70.15 77.50 87.10

Lignin % 23.00 4.63 5.53 19.00 12.50 6.90

Extractives % 20.00 9.50 5.55 10.85 10.00 6.00

Estimation of Reducing Sugars and Soluble Proteins

Reducing sugars produced during degradation of PNs were estimated 11 . Soluble proteins formed during biodegradation were quantified by Lowry’s method12 . Biodegradation Index (BI)

BI13 of PNs is calculated as BI = [reducing sugar (%) released + protein (%) formed] / 2. Hydrolysis (%)

Hydrolysis% is calculated on dry matter basis as14 Hydrolysis (%)=

In sulphuric acid (H2 SO4 ) pretreatment6 , PNs (10 g) were soaked in 1% H2 SO4 solution (100 ml) for 2 h. Biodegradation of Pine Needles (PNs)

Aspergillus niger F7 , capable of producing high amount of hydrolytic enzymes (cellulase and xylanase), was procured from Microbiology laboratory of Basic Sciences, UHF Nauni, Solan (India). Biodegradation of PNs was studied under SSF by using water and modified basal salt medium (BSM) (1: 2). Modified BSM 7 contained Na2 HPO4 (6.0 g), KH2 PO4 (3.0 g), NaCl (0.5 g), NH2 CL (1.0 g) and separately sterilized solutions of 1 M MgSO 4 (2 ml) and 1 M CaCl2 (0.1 ml) were added after medium was autoclaved. It was supplemented with urea (2%), yeast extract (1%), peptone (0.1%), NaNO3 (0.1%), 1M CoCl2 (0.2/l) with pH 6.80 to final volume of 1000 ml. To each 20 g of untreated and pretreated biomass of P. roxburghii, water (35 ml) and of inoculum (5 ml) containing 1x107 spores/ml of A. niger were added in 500 ml of Erlenmeyer flask. Flasks were incubated for 30 days at 28 ± 2°C. Extraction of Enzymes

Hydrolytic enzymes and other fermented products produced during biodegradation of PNs were extracted by Repeated Extraction Method8 . To 5 g of biomass, 50 ml of phosphate buffer (0.1M, pH 6.9) was added in 250 ml Erlenmeyer flask and contents were kept at 120 rpm for 1 h and then filtered through muslin cloth. The process was repeated twice with additional 25 x 2 ml of phosphate buffer making final volume of extracted products to 100 ml. After filtration, contents were centrifuged at 5000 rpm for 5 min at 4°C. Supernatant was collected to estimate enzymes, biodegradation index (BI) and hydrolysis (%). Enzyme assays were performed to quantify CMCaseÿÿ:9 , FPAase9 and ß-glucosidase of cellulase10 and xylanase11 activities.

779

Total reducing sugar (g)× 0.90 × 100 Weight of substrate (g)

Statistical Analysis

Completely randomized design was applied. Different regression models (linear, power, exponential and quadratic) were used to predict hydrolysis (%) and BI activities on the basis of enzyme activities in two mediums (water and modified BSM). Results and Discussion Estimation of Different components in Pine Needles (PNs)

Analysis of untreated biomass of PNs gave: holocellulose, 57.00; lignin, 23.00; and extractives (alcohol, benzene, fibers, resins etc.), 20.00% (Table 1). For efficient biodegradation of holocellulose of PNs, different pretreatments were given to wash lignin and extractives out of PNs. Holocellulose was found highest in alkali treated biomass of PNs (5% NH3 , 88.92; NaOH+H 2 O2 , 87.10; and 1% NH3 , 85.87%). Maximum lignin (19%) was retained by HCl pretreated material while lowest (4.63%) was in 1% NH3 pretreated needles. Extractives were: 1% NH3 , 9.50; 5% NH3 , 5.55; and NaOH+H 2 O 2 , 6.00%. Similar studies showing an increase in cellulose contents and decrease in lignin after pretreating wood biomass has earlier been reported15 . Thus different pretreatments temper lignin shield and take it out of lignocellulosic materials, thereby exposing most of the cellulose in active form for better enzymatic digestion. All pretreated PNs though have shown higher saccharification with enzymes secreted from A. niger during degradation but their BI and hydrolysis% values vary from treatment to treatment (Table 2). SSF of pretreated and untreated materials was carried out under substrate: moisture ratio (1: 2), which has been optimized for other lignocellulosic forest wastes7 . Moistening agents (tap water and modified BSM) were used during SSF with an ultimate aim to enhance biodegradation of PNs.

780


Table 2—Enzyme activity, biodegradation index (BI) and hydrolysis% of pine needles (Pinus roxburghii) after solid state fermentation by Aspergillus niger F7 using water and modified BSM as medium Treatments

Untreated 1% NH3 5% NH3 1% HCl 1% H2 SO4 NaOH+H 2 O2

H2 O as moistening agent Total Enzyme U/g 28.16 49.90 72.28 44.91 51.21 96.14

BSM as moistening agent

BI

Hydrolysis%

7.16 13.00 17.20 8.00 9.00 27.76

0.89 2.85 3.67 1.22 1.40 7.00

Total Enzyme U/g 95.40 144.80 185.06 106.00 137.80 259.60

BI

Hydrolysis%

13.45 26.84 33.40 15.56 16.56 53.30

1.85 5.88 7.08 2.36 2.72 12.87

U/g (on dry matter) of hydrolytic enzymes.

When water was used as moistening agent, degradation was very low in case of untreated biomass as follows: enzyme production, 28.16 U/g; BI, 7.16; and hydrolysis, 0.89%. Among pretreatments, NaOH+H 2 O2 pretreated PNs had led to maximum values as follows: cellulase and xylanase production from A. niger, 96.14 U/g; BI, 27.76; and hydrolysis, 7%. Biodegradation carried out with HCl pretreated biomass of PNs gave minimum values as follows: enzyme production, 44.91 U/g; BI, 8; and hydrolysis, 1.22%. When modified BSM was used as moistening agent, NaOH+H 2 O2 pretreated PNs have led to maximum cellulase and xylanase production from A. niger (259.6 U/g), thus resulting in highest BI (53.30) and hydrolysis 12.87%. Acid pretreatments showed comparatively lower production of enzymes, BI and hydrolysis% as compared to alkali pretreatments. On the other hand, untreated biomass in modified BSM also has shown least production of enzyme as compared to all pretreated biomass of PNs, consequently resulting in marginally low degradation as follows: enzyme production, 95.40 U/g; BI, 13.45; and hydrolysis, 1.85%. Since PNs are exceptionally inert biomass for biodegradation, therefore alkali pretreatment of PNs [NH3 (1%, 5%) and NaOH + H2 O2 ] has been chosen with an idea of removing maximum lignin and other hindering substances like resins etc. Alkali pretreatment is reported to decrease crystallinity of cellulose, remove lignin shield around cellulose and increase pore size of biomass, thus increasing digestibility of lignocelluloses16 . Compared with acid or oxidative reagents, alkali pretreatment appears to be the most effective methods in breaking ester bonds between lignin, hemicellulose and cellulose and avoiding fragmentation of hemicellulose polymers17 . Alkaline pretreatment in combination with

H2 O2 (NaOH + H2 O2 ) additionally promotes to loosen the linkage of hydrogen bonds, resulting in easy enzymatic hydrolysis of biomass18 . Alkaline pretreated biomass is hydrolyzed 40% faster than native cellulose. Under acid treatment of lignocellulosic materials, sulphuric acid is the most applied acid 19 and found most effective in dissolving lignin, and thus increasing cellulose’s susceptibility to enzymatic attack20 . Pretreatment consists of collection, transportation, manipulation, storage, grinding or chipping to reduce particle size and opening fibrous material in order to transform it into a suspension that can be pumped and enable further penetration of chemical hydrolysis agents21 . Overall an appreciable increase has been observed in pretreated PNs as compared to untreated ones. When water was used as moistening agent, in acidic pretreatment (HCl & H2 SO4 ), increase was observed as follows: enzyme activity, 59.48, 81.85%; BI, 11.73, 25.69% and hydrolysis, 37.07, 57.30%. In NaOH + H 2 O2 pretreated PNs, increase was observed as follows: enzyme, 241.40; BI, 287.70; and hydrolysis, 686.51%. When modified BSM was used, in NaOH + H2 O 2 pretreated PNs, increase was observed as follows: enzyme production, 172.11; BI, 296.28; hydrolysis, 595.67% (Fig. 1). A positive correlation has been derived between enzyme activity, BI and hydrolysis% of PNs. Thus when enzyme activity increases, BI and hydrolysis% also increases (Fig. 2). Parameters of various regression models and R2 for estimation of BI, hydrolysis% and enzyme activities show a direct correlation between these parameters (Table 3). Different regression models were tried and higher value of R 2 was found in quadratic model (Y = a + bx + cx 2 ) for prediction of BI activities on the


781

Increase in enzyme activity, %

300 250 200 150 100 50 0 Ammonia (1%)

Ammonia (5%)

Hydrochloric acid

Sulphuric acid

Pretreatments a)

Sodium hydroxide + Hydrogen peroxide

350

Increase in BI, %

300 250 200 150 100 50 0 Ammonia (1%)

Ammonia (5%)

Hydrochloric acid

Sulphuric acid


Pretreatments b) 800

Increase in hydrolysis, %

700 600 500 400 300 200 100 0 Ammonia (1%)

Ammonia (5%)

Hydrochloric acid

Sulphuric acid


Pretreatments c) Water

BSM

Fig.1—Comparison in pretreated pine needles by Aspergillus niger F7 using water and modified BSM as medium over untreated pine needles of increase in: a) Total enzyme activity; b) BI; and c) Hydrolysis%

782 120

30

100

25

80

20

60

15

40

10

20

5

0

% hydrolysis, B. I.

Enzyme activity


0

Ammonia (1%)

Ammonia (5%)

Hydrochloric acid

Sulphuric acid


300

60

250

50

200

40

150

30

100

20

50

10

0

% hydrolysis, B. I.

Enzyme activity

Pretreatments a)

0

Ammonia (1%)

Ammonia (5%)

Hydrochloric acid

Sulphuric acid


Pretreatments b) Fig. 2—Correlation between enzyme activity, B.I. and per cent hydrolysis of pine needles after SSF with Aspergillus niger F7 using: a) water as medium; and b) Modified BSM as medium Table 3—Parameters of various models to predict enzyme activity, BI and hydrolysis% Characters

Prediction model for BI activities and per cent hydrolysis (Y) on the basis of enzyme activities Y=a+bX R2 Y=abX R2 Y=aX b R2 Y=a+bX+cX 2 R2 a b a b a b a b Water -6.48 4.37 0.95 5.11 1.21 0.92 2.264 1.183 0.915 1.204 1.844 0.968 BSM -2.66 1.121 0.90 1.82 0.66 0.85 0.01 0.22 0.86 -0.289 0.339 0.93 X=Enzyme activity, Y=BI/hydrolysis, R2 =Coefficient of determination

basis of enzyme activities of water as well modified BSM. Thus quadratic model can be used for the prediction of enzyme activities, BI and hydrolysis%. Thus it has been established that with increase in extracellular cellulase

Y=aExpbX a b 5.11 0.196 1.82 0.664

R2 0.920 0.85

and xylanase production from hydrolytic microorganisms, biodegradation of PNs is enhanced. Though biodegradation of lignocellulosic wastes like agricultural biomass (corn cob, corn straw, baggase) and


other forest residues has already been reported22 , but successful biodegradation of PNs are rarely reported. This study strongly proves that pretreated PNs under SSF with A. niger F7 can serve as an inexpensive substrate for its saccharification into fermentable sugars, which in turn can be fermented to ethanol to be used as biofuels. Conclusions PNs, which are highly resistant to biodegradation, can be degraded successfully with A. niger after suitable pretreatment. Modified alkali pretreatment [NaOH+H 2 O2 (1M); ratio, 9:1] in 2 h followed by steam explosion for 15 min has been found the best pretreatment for PNs hydrolysis by a hypercellulolytic isolate, A. niger F7 . Modified BSM mediated SSF was found better over tap water. A positive correlation is drawn in three parameters (enzyme activity, BI and hydrolysis%) and has been proved statistically by using regression model. References 1 2

3

4

5 6

7

8

1 Bhasin R, Forest fire ravages Himachal flora and fauna, Biores Technol, 5 (2008) 39-45. Kondo P, Investigation on mechanism of biological dezincification by solid state fermentation, J Sci Ind Res, 55 (1996) 394-399. Alcohol-benzene solubility of wood, Official Standards, T6 M59 [Technical Association of Pulp and Paper Industry (TAPPI), New York] 1950. Holocellulose in wood, Official Standards, T12M-59 [Technical Association of Pulp and Paper Industry (TAPPI), New York] 1954. Lignin in wood, Official Standards, T12M-59 [Technical Association of Pulp and Paper Industry (TAPPI), New York] 1959. Fan L T Y H, Gharpuray M M & Beard M D H, The nature of lignocellulosics and their pretreatments for enzymatic hydrolysis, Adv Biochem Bioeng, 23 (1982)157-187. Sharma N, Bansal K L & Neopaney B, Effect of moisture level on biodegradation of forest waste under solid state fermentation, J Sci Ind Res, 65 (2006)675-679. Bollag D M & Edestein S J, Protein Methods (Wiley - Liss, John Wiley and Sons Inc, New York) 1993, 230.

9

10

11

12

13

14 15

16 17

18

19

20

21

22

783

Reese E T & Mandel M, Enzymatic hydrolysis of cellulose and its derivatives, in Methods Carbohydrate Chemistry, 3rd edn, edited by R L Whistler (Acad Press, London) 1963, 139-143. Bergham L E R & Petterson L G, Mechanism of enzymatic cellulose degradation: Purification of cellulolytic enzyme from Trichoderma viride active on highly ordered cellulose, J Biochem, 37 (1973) 21-30. Miller G L, Use of dinitrosalicylic acid reagent for determination of reducing sugars, Analyt Chem, 31 (1959) 426-428. Lowry O H, Rosebrough N J, Farr A L & Randall R J, Protein measurement with folin phenol reagent, J Biol Chem, 193 (1951) 265-275. Sharma N, Bhalla T C, Aggarwal H O & Bhatt A K, Saccharification of physico-chemically pretreated lignocellulosics by partially purified cellulase of Trichoderma viride, Sci Lett, 19 (1996) 141-144. Szczodrak J, Rogalski J & Liczuk Z, Cellulolytic activity of molds, Acta Microbiol Polonica, 33 (1984) 217-225. Jan K, Lisbeth G, Thygesen C, Henning J & Thomas E, Cell wall structural changes in wheat straw pretreated for bioethanol production, Biotech Biofuels, 24 (2008) 1-5. Fan I T, Lee Y H & Beard D H, Mechanisms of the enzymatic hydrolysis of cellulose, Biotechnol Bioeng, 22 (1980) 177-199. Gaspar M, Kalman G & Reczey K, Corn fiber as a raw material for hemicellulose and ethanol production, Process Biochem, 42 (2007) 1135-1139. Saha B C & Cotta M A, Ethanol production from alkaline peroxide pretreated enzymatically saccharified wheat straw, Biotech Prog, 22 (2006) 449-453. Taherzadeh M J & Karimi K, Acid based hydrolysis process for ethanol from lignocellulosic materials, Rev Biores, 2 (2007) 472-499. Yang B & Wyman C E, Effect of xylan and lignin removal by batch and flow through pretreatment on the enzymatic digestibility of corn stover cellulose, Biotech Bioeng, 86 (2004) 88-95. Muzzy J D, Robertis R S, Fiebe C A, Fieber G S & Mann T M, Pretreatments of hard wood by continuous hydrolysis, in Wood and Agricultural Residues, vol 25 (Acad Press, London) 2009, 351-368. Damisa D & Ameh J B, Effect of chemical pretreatment of some lignocellulosic waste on recovery of cellulase from Aspergillus niger, Appl Microbiol, 22 (2008) 209-213.




Comparison of seven types of thermo-chemical pretreatments on the structural features and anaerobic digestion of sunflower stalks F. Monlau, A. Barakat 1, J.P. Steyer, H. Carrere ⇑ INRA, UR050, Laboratoire de Biotechnologie de l’Environnement, Avenue des Etangs, F-11100 Narbonne, France

h i g h l i g h t s " Comparison of seven types of thermo-chemical pretreatment on the structural features of sunflower stalks. " Cellulose crystallinity, lignin removal and the solubilisation of cellulose, hemicelluloses, uronic acids and proteins were considered. " Highest methane potential increase was obtained after 55 °C, 24 h, 4% NaOH. " Methane potential and rate were linked with structural features changes.

a r t i c l e

i n f o

Article history: Received 20 February 2012 Received in revised form 14 May 2012 Accepted 14 June 2012 Available online 22 June 2012 Keywords: Lignocellulosic biomass Chemical composition Crystallinity Biogas Methane potential

a b s t r a c t Sunflower stalks can be used for the production of methane, but their recalcitrant structure requires the use of thermo-chemical pretreatments. Two thermal (55 and 170 °C) and five thermo-chemical pretreatments (NaOH, H2O2, Ca(OH)2, HCl and FeCl3) were carried out, followed by anaerobic digestion. The highest methane production (259 ± 6 mL CH4 g1 VS) was achieved after pretreatment at 55 °C with 4% NaOH for 24 h. Acidic pretreatments at 170 °C removed more than 90% of hemicelluloses and uronic acids whereas alkaline and oxidative pretreatments were more effective in dissolving lignin. However, no pretreatment was effective in reducing the crystallinity of cellulose. Methane production rate was positively correlated with the amount of solubilized matter whereas methane potential was negatively correlated with the amount of lignin. Considering that the major challenge is obtaining increased methane potential, alkaline pretreatments can be recommended in order to optimize the anaerobic digestion of lignocellulosic substrates. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Sunflower residues and sunflower stalks represent interesting feedstocks for methane production. They are available in large quantities, have few suitable other end uses and are generally burnt in the fields, causing environmental pollution. Methane production from sunflower residues has been investigated by Antonopoulou et al. (2010) who found a methane potential of 240 mL CH4 g1 sunflower residues. However, lignocellulosic substrates like sunflower stalks present a major problem for biomethane production on account of their complex structure that limits their biodegradability (Monlau et al., 2012). Effective pretreatment prior to anaerobic digestion should break down the linkage between polysaccharides and lignin to make cellulose and hemicelluloses more accessible to bacteria ⇑ Corresponding author. Tel.: +33 468425168; fax: +33 468425160. E-mail address: [email protected] (H. Carrere). Present address: INRA, UMR IATE 1208, Ingénierie des Agro polymères et Technologies Emergentes, 2, place Pierre Viala, F-34060 Montpellier, France. 1

0960-8524/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biortech.2012.06.040

(He et al., 2008). To this end, different pretreatment methods (mechanical, chemical, thermo-chemical and biological) have been applied for second generation bioethanol production and, to a lesser extent, to enhance the anaerobic digestion of lignocellulosic residues (Taherzadeh and Karimi, 2008). Ensiling of plant biomass can be considered as a biological pretreatment (Nizami et al., 2009). Biological pretreatment can also be integrated into a twostage system consisting of a dry batch leaching stage (hydrolysis and acidogenesis) followed by an upflow anaerobic sludge blanket where the leachate produced in the first stage undergoes methanogenesis (Nizami et al., 2010). During the leaching stage hemicellulose was loosened and ruptured and lignin and cellulose structures which originally stretched in straight lines were distorted (Nizami et al., 2010). Dilute-acid pretreatment has also been successful in hydrolyzing hemicelluloses, changing the structure of lignin and increasing the cellulosic surface area (Mosier et al., 2005; Taherzadeh and Karimi, 2008). Dilute-acid pretreatments are often used at high temperatures (160–220 °C) for a few minutes (Monlau et al., 2012). Sulfuric acid is most widely used but hydrochloric, phosphoric, nitric and acetic acids and inorganic salts, especially

242

F. Monlau et al. / Bioresource Technology 120 (2012) 241–247

FeCl3, are also efficient in the removal of hemicelluloses (Liu et al., 2009). However, acidic pretreatment can be responsible for the formation of byproducts such as 5-hydroxylmethylfurfural (5-HMF) and furfural (Larsson et al., 1999). Though these byproducts did not inhibit methane production from xylose (Barakat et al., 2012), the methanogenic microorganisms required a period of adaptation that resulted in a lower methane production rate (Benjamin et al., 1984). In contrast to acidic pretreatments that use high temperatures and short time periods, alkaline or oxidative pretreatments are often carried out at low temperature over a longer period (6–24 h) (Mosier et al., 2005; Taherzadeh and Karimi, 2008). Alkaline and oxidative pretreatments (NaOH, Ca(OH)2, H2O2) are effective in increasing the accessible surface by removing some of the lignin and hemicelluloses (Taherzadeh and Karimi, 2008). The objectives of this study were to compare the effects of seven types of thermo-chemical pretreatments (two temperatures and five chemical reagents: NaOH, H2O2, Ca (OH)2, HCl and FeCl3) on the structural features (chemical composition and crystallinity) of sunflower stalks, to compare the impact of the thermo-chemical pretreatments on methane production; and to establish correlations between changes in the structural features of sunflower stalks before and after pretreatments and their methane potential. 2. Methods 2.1. Raw materials ‘‘NK-Kondi’’ sunflower stalks were milled into a particle size of 2–3 mm using a SM-100 cutting mill. Pretreated and untreated sunflower stalks were analysed for TS (Total Solids) and VS (Volatile Solids) in accordance with APHA standard methods (APHA, 1998). The main characteristics and composition of the samples are presented in Table 1. 2.2. Pretreatment The pretreatment conditions are presented in Table 2. NaOH, H2O2, Ca (OH)2 pretreatments were done in 500-mL flasks with a chemical agent concentration of 4% (g/100 g TS) for 24 h at 30, 55, and 80 °C. The assays were performed on an ‘‘Edmund Butler’’ heating shaker series SM-30-control with an agitation of Table 1 Composition of sunflower stalks after drying and milling. Characteristics

Mean ± S.D.

TS (% wet weight) VS (%wet weight) Cellulose (%VS) Hemicelluloses (%VS) Klason Lignin (%VS) Uronic acids (%VS) Protein (%VS)

94 ± 1.0 88 ± 1.0 34 ± 0.6 20.8 ± 0.8 29.7 ± 0.6 7.6 ± 0.3 5.2 ± 0.1

150 rpm. HCl and FeCl3 pretreatments were performed in a Zipperclave autoclave, series 02-0378-1 (Autoclave France), with a concentration of 4% and 10% (g/100 g TS), respectively, for 1 h at 170 °C. The stainless steel autoclave, with a capacity of 1 L, can reach a temperature of 250 °C and a pressure of 79 bars. The reactor content was agitated by a rod with two propellers at a rate of 150 rpm and was heated by a ceramic furnace. Treatments at 30, 55, 80 and 170 °C without chemical reagents served as controls. Directly after pre-treatment, BMP (Biochemical Methane Potential) tests were carried out on the untreated and pretreated samples. Remaining pretreated samples were filtered through a 0.25-mm pore size sieve to separate the solid from the liquid fraction for further chemical analysis. 2.3. Chemical composition The carbohydrate (glucose, xylose and arabinose) and uronic acid (galacturonic and glucuronic) in solid phases were measured in duplicate using the strong acid hydrolysis protocol adapted from Effland (1977). All pretreated and untreated samples were milled into 1-mm particles using an Ika Werke MF 10 cutting mill. Samples (200 mg) were hydrolyzed with 12 M H2SO4 for 2 h at room temperature, then diluted to reach a final acid concentration of 1.5 M and kept at 100 °C for 3 h. The mixture was filtered through paper fiberglass (GFF, WHATMAN). This insoluble residue was washed with 50 mL of deionised water and then placed in a crucible. The crucible and fiberglass were dried at 100 °C for 24 h and burnt at 550 °C for 2 h to determine the Klason lignin content. The filtrate was filtered through nylon filters (20 lm) for analysis of glucuronic acids, galacturonic acids, glucose, xylose and arabinose by high performance liquid chromatography (HPLC). The analysis was done with a combined Water/Dionex system, using a BioRad HPX-87H column at 50 °C. The solvent was 0.005 M H2SO4 and the flow-rate was 0.3 mL/min. A refractive index (RI) detector (Water R410) was used to quantify carbohydrates. The system was calibrated with glucuronic acid, galacturonic acid, glucose, xylose and arabinose (Sigma–Aldrich). Kjeldahl nitrogen (TKN) was titrated using a Buchi 370-K after mineralisation of the samples. Proteins were determined by multiplying TKN by 6.25. For the determination of furfural and 5-hydroxylmethylfurfural (5-HMF), liquid fraction from each chemical pretreatment were filtered through nylon filters (20 lm) and subjected to HPLC with a combined Water/Dionex system, using a BioRad HPX-87H column at 35 °C. The solvent was 0.005 M H2SO4 and the flowrate 0.4 mL/min. A refractive index (RI) detector (Water R410) was used. The system was calibrated with furfural and 5-HMF (Sigma–Aldrich). 2.4. FTIR-ATR measurements All spectra were recorded at 4 cm1 resolution intervals and 64 scans at room temperature. Spectra were collected in the 4000–600 cm1 range using a Nexus 5700 spectrometer

Table 2 Pretreatment conditions. Chemical

Temperature (°C)

Concentration (g/100 g TS)

Time (h)

pH intial

No chemical

30; 55; 170 30; 55; 30; 55; 30; 55; 170 170

80

–

7.5

80 80 80

4 4 4 4 10

24 1 24 24 24 1 1

NaOH Ca(OH)2 H202 (pH = 11.5, adjusted with NaOH) HCl FeCl3

11.9 12.2 11.6 2.3 1.8


(ThermoElectron Corp.) with a built-in diamond ATR single reflection crystal and a cooled MCT detector. Three spectra were recorded for each sample. All spectra pre-treatments were performed using Omnic v7.3. 2.5. Methane potential Treated and untreated samples were digested in batch anaerobic flasks. The volume of each flask was 550 mL, with a working volume of 400 mL, the remaining 150 mL volume serving as head space. The flask contained a macroelements solution (NH4Cl, 26 g/L; KH2PO4, 10 g/L; MgCl2, 6 g/L; CaCl2, 3 g/L), an oligoelement solution (FeCl2, 2 g/L; CoCl2, 0.5 g/L; MnCl2, 0.1 g/L; NiCl2, 0.1 g/L; ZnCl2, 0.05 g/L; H3BO3, 0.05 g/L; Na2SeO3, 0.05 g/L; CuCl2, 0.04 g/ L; Na2MoO4, 0.01 g/L), bicarbonate buffer (NaHCO3, 50 g/L), an anaerobic sludge at 5 gVS/L and the raw or pretreated substrate at 5 gTS/L. Once the flasks were prepared, degasification with nitrogen was carried out to obtain anaerobic conditions and the bottles were closed with red butyl rubber septum-type stopper which were air tight. Duplicate bottles were incubated at 35 °C. Biogas volume was monitored by the water displacement method. Acidified water (pH = 2) was used to minimize the dissolution of carbon dioxide in water. Biogas composition was determined using a gas chromatograph (Varian GC-CP4900) equipped with two columns: the first (Molsieve 5A PLOT) was used at 110 °C to separate O2, N2, CH4, the second (HayeSep A) was used at 70 °C to separate CO2 from other gases. The injector temperature was 110 °C and the detector 55 °C. The detection of gaseous compounds was done using a thermal conductivity detector. The calibration was carried out with a standard gas composed of 25% CO2, 2% O2, 10% N2 and 63% CH4 (special gas from Linde gas S.A.).

243

Eq. (4) leads to methane potentials of 415 mL CH4 g1 cellulose (C6H10O4)n, 424 mL CH4 g1 xylan (C5H804)n, 288 mL CH4 g1 uronic acids (C6H10O7) and 420 mL CH4 g1 proteins (C14H1207N2)n. Since sunflower stalks are composed of at 34%, 20.8%, 7.6%, and 5.2% cellulose, hemicelluloses, uronic acids and proteins, respectively, per VS and thus a methane potential of 272 mL CH4 g1 VS can be theoretically expected. The biodegradability of substrates before and after pretreatment can be determined as follows:

BD ð%Þ ¼ ½Experimental methane potential ðmL CH4 g1 VSÞ= 272 mL CH4 g1 VS 100

ð5Þ

The analysis of variance (Anova) method was used to analyse the impact of temperature in each kind of chemical pretreatment, the confidence level considered was 95%. Cellulose contains crystalline and amorphous structures (Eq. (6)). The crystalline cellulose content was determined from IR spectra using the LOI (Lateral Order Index). The LOI is the ratio of the 1430 cm1 peak area divided by 898 cm1 peak area (Akerholm et al., 2004). The absorbance at 1430 and 898 cm1 is sensitive to the amount of crystalline cellulose versus amorphous cellulose (Li et al., 2009). So, using the combination of Eqs. (6) and (7), it is possible to estimate the crystalline cellulose content according the Eq. (8).

Cellulose ¼ Crystalline Cellulose þ Amorphous Cellulose

ð6Þ

LOI ¼ Crystalline Cellulose=Amorphous Cellulose

ð7Þ

Crystalline Cellulose ð%TSÞ ¼ Cellulose ð%TSÞ LOI=ð1 þ LOIÞ

ð8Þ

4. Results and discussion 3. Calculations 4.1. Methane potentials In order to evaluate the impact of each pretreatment, the cellulose and hemicelluloses content was determined on the basis of the monomeric sugar content as follows:

Cellulose ð% TSÞ ¼ Glucose ð%TSÞ=1:11

ð1Þ

Hemicelluloses ð% TSÞ ¼ ½Xylose ð%TSÞ þ Arabinose ð%TSÞ=1:13;

ð2Þ

with 1.11 the conversion factor for glucose-based polymers (glucose) to monomers and 1.13 the conversion factor for xylosebased polymers (arabinose and xylose) to monomers (Petersson et al., 2007). All methane potentials were expressed in mL CH4 g1 initial VS so the eventual losses of organic matter during pretreatments were considered in the results. To quantify the kinetic advantage of the pre-treatment on anaerobic methane production, the first order kinetic constants were calculated by using least-squares fit of methane production data during time (t) to the following equation:

B ðtÞ ¼ Bo ð1 ekt Þ;

4.1.1. Impact of temperature The methane potentials (mL CH4 g1 initial VS) for pretreated and untreated sunflower stalks are presented in Fig. 1. The methane potential for sunflower stalks was estimated at 192 (±2) mL CH4 g1 VS which is lower than that reported by Antonopoulou et al. (2010) with 240 mL CH4 g1 VS. Thermal pretreatment alone (30, 55 and 80 °C) did not increase the methane potential of sunflower stalks and a small increase of methane potential (219 (±8) mL CH4 g1 VS) compared to raw sunflower stalks was noticed at 170 °C. Antonopoulou et al. (2010) recorded an increase around 13% when sunflower straw was pretreated at 121 °C for 60 min. For alkaline and oxidative pretreatments, 55 °C seems to be the optimal temperature even though no significant impact of temperature on hydrogen peroxide and lime pretreatments

ð3Þ

where B(t) is the volume of methane produced at time t(d), expressed in mL CH4 g1 VS, Bo is the maximum producible methane volume (mL CH4 g1 VS) and k is the hydrolysis kinetics constant (d1). Bo and k were determined using Microsoft Excel’s Solver function. A theoretical methane potential (mL CH4 g1 VS) can be calculated according to the elemental composition of each degradable compounds of the substrate CaHbOcNdSe (Frigon and Guiot, 2010).

Y Theroretical ðL=gsubstrate Þ ¼ CH4

22:4ð4a þ b 2c 3d 2eÞ 8ð12a þ b þ 16c þ 14d þ 16eÞ

ð4Þ

Fig. 1. Methane potentials of pretreated and untreated sunflower stalks. Values correspond to means of two replicates of independent values ± standard deviations (error bars).

244


Table 3 Methane potential and hydrolysis kinetics constant for selected conditions of raw and pretreated sunflower.

a

Conditions

BMP (mL CH4 g1 initial VS)

Increase BMP (%)

BDa (%)

k (d1)

Increase k (%)

R2

Raw 24 h, 55 °C 24 h, 55 °C, 24 h, 55 °C, 24 h, 55 °C, 1 h, 170 °C 1 h, 170 °C, 1 h, 170 °C,

192 (±2) 198 (±11) 259 (±6) 256 (±2) 241 (±13) 219 (±8) 248 (±6) 233(±2)

– 3 35 33 26 14 29 21

71 73 95 94 89 81 91 86

0.022 0.022 0.028 0.027 0.023 0.036 0.04 0.039

– 0 27 23 5 64 82 77

0.97 0.99 0.96 0.97 0.94 0.94 0.95 0.96

4% NaOH 4% H202 4% Ca(OH)2 10% FeCl3 4% HCl

BD (%): Biodegradability calculated from the Eq. (5).

was detected with Anova p-values of 1.16 and 0.39 respectively. No clear conclusion could be drawn about the impact of temperature on NaOH pretreatments (p = 0.051). No conclusions can be made on the effect of temperature during HCl and FeCl3 pretreatments as the treatments were performed only at 170 °C. However, increases in methane potentials of 21% (233 mL CH4 g1 VS) and 29% (248 mL CH4 g1 VS) compared to raw straw were observed, respectively, for 4% HCl, 170 °C, 1 h and 10% FeCl3, 170 °C, 1 h. The effect of thermo-chemical (H2SO4 and NaOH) pretreatments on sunflower straw was investigated by Antonopoulou et al. (2010) and, contrary to our results; these authors did not observe an increase in methane despite the fact that solubilisation of sugars was enhanced. It is possible that the different results were caused by the differences in temperature as Antonopoulou et al. (2010) conducted pretreatment at 121 °C. 4.1.2. Impact of thermo-chemical pretreatment on methane potential and hydrolysis kinetic constant for the selected conditions For all the conditions studied, except for 55 °C alone, an increase in the methane potential was observed compared to that of the raw sunflower stalks (Table 3). A methane potential of 192 mL CH4 g1 VS was observed for raw sunflower stalks that corresponded to a biodegradability of 71%. Increases in the methane potential of 33 and 30% were observed for 4% H2O2, 55 °C, 24 h and 10% FeCl3, 170 °C, 1 h, respectively. The highest increase in methane potential of 35% was observed for 4% NaOH, 55 °C, 24 h, as a methane potential of 259 (±6) mL CH4 g1 VS was achieved. A biodegradability of 95% was observed for this condition. However, this value should be considered as a rough estimate insofar as the lipid content was not estimated for the assessment of biodegradability. Similar results were observed after alkaline pretreatment of sorghum forage with 10% NaOH (w/w TS) at 40 °C for 24 h as an increase in the methane potential of 29% was achieved (Sambusiti et al., 2011). Lower biodegradabilities were observed for thermal pretreatments at 170 °C and with HCl as reagent (81% and 86%, respectively, for 170 °C alone and 170 °C, 1 h, 4% HCl). For all samples, first-order kinetics was successful in modeling the experimental methane production during the first 18 days (R2 > 0.94). The hydrolysis kinetics constants k (d1) are shown in Table 2. For oxidative and alkaline pretreatments, small increases were observed for the kinetics hydrolysis constant k: as 0.023 d1 (+5%), 0.027 d1 (+23%) and 0.028 d1 (+27%) were observed for Ca(OH)2, H2O2 and NaOH, respectively. Thermo-pretreatment at 170 °C increased the hydrolysis constant by 64% (0.036 d1). Higher hydrolysis kinetics constants of 0.039 and 0.040 were recorded for acidic pretreatment (HCl and FeCl3) that correspond to 77% and 82% increases, respectively. 4.2. Impact of chemical reagents on biochemical changes 4.2.1. Chemical composition Results of the investigation of biochemical changes induced by thermo-chemical pretreatment on protein, uronic acids,

Fig. 2. Biochemical composition of raw sunflower stalks and of the solid residue after basic and oxidative pretreatments. Values correspond to means of two replicates of independent values ± standard deviations (error bars).

Fig. 3. Biochemical composition of raw sunflower stalks and of the solid residue after thermal and acidic pretreatments. Values correspond to means of two replicates of independent values ± standard deviations (error bars).

hemicelluloses, cellulose and Klason lignin content are presented in Figs. 2 and 3. All thermo-chemical pretreatments were effective in protein removal; even at 55 °C for 24 h, more than 60% of the proteins were removed from raw sunflower stalks.Thermo-chemical pretreatments were also effective in uronic acid removal and their solubilisation was complete for all pretreatments at 170 °C. These uronic acids originate from hemicelluloses and pectins. According to Chandel et al. (2011), hemicelluloses and pectin bind to cellulose to form a network of cross-linked fibres. Removal of uronic acids can increase the accessibility of enzymes to hemicelluloses and cellulose (Pakarinen et al., 2012). In contrast to oxidative and basic pretreatments at 55 °C, a high level of hemicelluloses removal was observed for thermal pretreatment at 170 °C, with or without acid reagent (Fig. 3).

245

F. Monlau et al. / Bioresource Technology 120 (2012) 241–247 Table 4 Concentrations of furfural and 5-hydroxymethyl furfural in the liquid fraction of pretreated samples. Conditions


Liquid fraction (g/100 g initial VS)


5-HMF

Furfural

– 0 0 0 0 0 0.3 (±0.02) 0.4 (±0.00)

– 0 0 0 0 0.7 (±0.02) 2.4 (±0.03) 4.1 (±0.00)

The chemical pretreatments were not effective in removing cellulose insofar as the highest removal of 12% was observed at 170 °C with HCl. Such results were also obtained by Liu et al. (2009) with corn stover. These authors found that pretreatment at 140 °C for 20 min in the presence of 0.1 M FeCl3, resulted in removal of 91% of the hemicelluloses but only of 9% of the cellulose. Dilute sulfuric acid treatment has been used successfully to hydrolyze hemicelluloses to sugars with high yields and to change the structure of the lignin (Mosier et al., 2005). Generally, acid pretreatments are known to enhance degradation of xylans which are the main components of hemicelluloses (Nizami et al., 2009). Oxidative and alkaline pretreatments were more efficient than acidic pretreatment for lignin removal, with 30%, 35%, and 36% for Ca (OH)2, H2O2 and NaOH, respectively, compared to 24% and 27% for FeCl3 and HCl, respectively (Figs. 2 and 3). Alkaline and oxidative pretreatments have been shown to be efficient in lignin removal by preserving most of the carbohydrates, in particular cellulose (Zhu et al., 2010; Taherzadeh and Karimi, 2008). Partial delignification of 24% was obtained by further increasing the temperature to 170 °C, but at this temperature, the addition of acid (HCl or FeCl3) did not lead to further delignification. Guo et al. (2011) investigated different pretreatment strategies for corn stalk and concluded that acid treatment was effective in hemicellulose

Table 5 Lateral Order Index (LOI), crystalline cellulose and H lignin/H carbohydrates ratio for raw and pretreated sunflower. Conditions

LOI

Crystalline cellulose (% initial VS)

H lignin/H carbohydrates


1.22 1.2 1.21 1.26 1.15 1.21 1.16 1.10

18.7 18.6 18.8 19.1 19.3 18.5 17.7 16.0

0.11 0.11 0.11 0.11 0.12 0.14 0.15 0.16


removal while alkaline pretreatment led to a significant decrease in lignin content. It was also recorded that during this thermal pretreatment at high temperature with or without acid reagent, in addition to soluble sugars, furans derivatives such as furfural and 5-hydroxyl-methylfurfural (5-HMF) can be generated (Larsson et al., 1999). The furfural and 5-HMF contents were determined for conditions applied in this study (Table 4). Formation of furfural was observed for acidic pretreatments, whereas at 170 °C, only 0.7 g/100 g VS of furfural were recorded, which is in accordance with the results obtained by Diaz et al. (2011). 4.2.2. FT-IR spectra The fingerprint regions of the FT-IR spectra of raw and pretreated sunflower stalks are presented in Fig. 4. The peaks were assigned as follows: 1610 cm1 to holocelluloses (Shafiei et al., 2010); 1512 cm1 is characteristic of aromatic skeletal vibration C = C of lignin; 1430 cm1 assigned to C–H deformation in cellulose; 1375 cm1 assigned to deformation in cellulose and hemicelluloses (Pandey and Pitman, 2004); 1160 cm1 assigned to C–O–C vibration in holocelluloses and 898 cm1 to CH deformation in cellulose (Yang et al., 2009). The peak at 1610 cm1 assigned to cellulose and hemicelluloses decreased significantly with the increase in temperature (170 °C), indicating that part of the holocelluloses

Fig. 4. Fingerprint region (600–3000 cm1) of the FTIR spectra of raw and pretreated sunflower stalks.

246


Fig. 5. Correlation between methane potential and the lignin content of raw sunflower stalks and residual solid fraction of pretreated sunflower stalks (a); correlation between the hydrolysis kinetic constant and the sum of solubilisation of protein, cellulose, hemicelluloses and uronic acids of pretreated and untreated sunflower stalks (b).

was solubilised. Moreover, the intensity of the peak at 1512 cm1 was higher after thermal (170 °C) and acidic thermal pretreatments (170 °C, HCl and 170 °C, FeCl3) than for raw sunflower stalks, indicating that the relative content of lignin in sunflower stalks increased after thermal and acidic pretreatment. The previous observations on spectra can be quantified using the H lignin/H carbohydrates ratio which shows the relative intensity of lignin peaks at 1512 cm1 as opposed to carbohydrates peaks at 1610, 1430, 1375, 1160 and 895 cm1. An increase in this ratio was observed after thermal-pretreatment at 170 °C and for acidic conditions, which means that during such pretreatment more holocelluloses were removed than lignin (Table 5). The ratio of H lignin/H carbohydrates increased from 0.11 (raw sunflower stalk) to 0.16 (170 °C, 4% HCl, 1 h). Such results are in accordance with the observations on chemical composition recorded previously showing that at 170 °C, 24 h, 4% HCl, all the hemicelluloses were removed. The LOI ratio which is sensitive to the amount of crystalline cellulose versus amorphous cellulose was calculated and results are presented in Table 5. A small decrease in the LOI ratio was noticed for acidic pretreatment (HCl and FeCl3), indicating a small decrease of crystallinity of cellulose in sunflower stalks. The crystalline cellulose content was then determined using Eq. (7). No significant changes in crystalline cellulose were observed after alkaline and oxidative pretreatments. With acidic pretreatments, a small decrease in crystalline cellulose was noticed as 5.3% and 14% of crystalline cellulose removal was observed for 170 °C, 1 h, 4% FeCl3 and 170 °C, 1 h, 4% HCl, respectively.

4.3. Correlation between biochemical changes and anaerobic digestion performance A strong negative correlation (R2 = 0.92) was found between the methane and the lignin content (Fig. 5a). These results were also observed by Kobayashi et al. (2004) with steam-exploded bamboo where linear regression showed a strong negative correlation (R2 = 0.95) between the amount of methane produced and the amount of lignin. Triolo et al. (2011) also found a good negative correlation (R2 = 0.88) between lignin content and methane potentials of energy crops and manure. The lignin content plays a major role in methane production by limiting access of microorganisms to holocelluloses. Moreover, due to its non-water soluble nature, lignin represents the most recalcitrant part of the plant structure (Nizami et al., 2009). The holocelluloses (hemicelluloses and cellulose), which are anaerobically-biodegradable in their pure form, appeared to be less biodegradable or even completely refractory when combined with lignin (Tong et al., 1990). According to Zhu et al. (2010), removing lignin enables a high rate of biomass conversion. Moreover, the presence of furfural and 5-HMF during

thermal pretreatment (170 °C), with or without acid reagents, did not seem to affect either the methane potentials or the methane production rate, possibly due to the low concentrations observed. A good correlation (R2 = 0.91) was found between the sum of solubilised proteins, hemicelluloses, cellulose and uronic acids with the hydrolysis kinetic constant (Fig. 5b). Consequently, soluble matter originating from the removal of proteins, hemicelluloses, cellulose and uronic acids plays an important role in the hydrolysis kinetics constant of lignocellulosic residues; the greater the soluble matter, the higher the kinetic constant. Contrary to Zhu et al. (2010) who showed that decreasing crystallinity accelerates hydrolytic reaction in the biomass, no significant correlation was found between crystalline cellulose removal and the hydrolysis kinetic constant, perhaps due to the low range of crystalline cellulose concentration (16–19.3% initial VS) in our study. 5. Conclusion Considering that the major challenge for biomass conversion to methane is to increase the methane potential, alkaline pretreatment at 55 °C with 4% NaOH for 24 h, can be recommended for lignocellulosic substrates. However, these results have to be validated by experiments in continuous reactors to confirm the absence of inhibitors formed during thermo-chemical pretreatments. The process also has to be supported by an economic study to assess full scale application of thermo-chemical pretreatment. Moreover, it will be interesting to extend such pretreatments to other substrates with different chemical composition. Acknowledgements The authors are grateful to ADEME, the French Environment and Energy Management Agency, for financial support in the form of F. Monlau’s Ph.D. Grant. References Akerholm, M., Hinterstoisser, B., Salmen, L., 2004. Characterization of the crystalline structure of cellulose using static and dynamic FT-IR spectroscopy. Carbohydrate Research 339 (3), 569–578. Antonopoulou, G., Alexandropoulou, M., Lyberatos, G. 2010. The effect of thermal, chemical and enzymatic pretreatment on saccharification and methane generation from sunflower straws. Proceedings Venice 2010, Third International Symposium on Energy from Biomass and Waste. APHA – American Public Health Association, 1998. Standard Methods for the Examination of Water and Wastewater, 20th ed. Barakat, A., Monlau, F., Steyer, J.-P., Carrere, H., 2012. Effect of lignin-derived and furan compounds found in lignocellulosic hydrolysates on biomethane production. Bioresource Technology 104, 90–99. Benjamin, M.M., Woods, S.L., Ferguson, J.F., 1984. Anaerobic toxicity and biodegradability of pulp-mill waste constituents. Water Research 18 (5), 601– 607.

F. Monlau et al. / Bioresource Technology 120 (2012) 241–247 Chandel, A.K., Chandrasekhar, G., Radhika, K., Ravinder, R., Ravindra, P., 2011. Bioconversion of pentose sugars into ethanol: a review and future directions. Biotechnology and Molecular Biology Review 6 (1), 008–020. Diaz, M.J., Cara, C., Ruiz, E., Perez-Bonilla, M., Castro, E., 2011. Hydrothermal pretreatment and enzymatic hydrolysis of sunflower stalks. Fuel 90 (11), 3225– 3229. Effland, M.J., 1977. Modified procedure to determine acid-insoluble lignin in wood and pulp. Tappi 60 (10), 143–144. Frigon, J.C., Guiot, S.R., 2010. Biomethane production from starch and lignocellulosic crops: a comparative review. Biofuels Bioproducts and Biorefining-Biofpr 4 (4), 447–458. Guo, P., Mochidzuki, K., Cheng, W., Zhou, M., Gao, H., Zheng, D., Wang, X.F., Cui, Z.J., 2011. Effects of different pretreatment strategies on corn stalk acidogenic fermentation using a microbial consortium. Bioresource Technology 102 (16), 7526–7531. He, Y., Pang, Y., Liu, Y., Li, X., Wang, K., 2008. Physicochemical characterization of rice straw pretreated with sodium hydroxide in the solid state for enhancing biogas production. Energy and Fuels 22 (4), 2775–2781. Kobayashi, F., Take, H., Asada, C., Nakamura, Y., 2004. Methane production from steam-exploded bamboo. Journal of Bioscience and Bioengineering 97 (6), 426– 428. 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 and Microbial Technology 24 (3–4), 151– 159. Li, C.L., Knierim, B., Manisseri, C., Arora, R., Scheller, H.V., Auer, M., Vogel, K.P., Simmons, B.A., Singh, S., 2009. Comparison of dilute acid and ionic liquid pretreatment of switchgrass: biomass recalcitrance, delignification and enzymatic saccharification. Bioresource Technology 101 (13), 4900–4906. Liu, L., Sun, J.S., Cai, C.Y., Wang, S.H., Pei, H.S., Zhang, J.S., 2009. Corn stover pretreatment by inorganic salts and its effects on hemicellulose and cellulose degradation. Bioresource Technology 100 (23), 5865–5871. Monlau, F., Barakat, A., Trably, E., Dumas, C., Steyer, J.-P., Carrere, H., 2012. Lignocellulosic materials into biohydrogen and biomethane: impact of structural features and pretreatment. Critical Reviews in Environmental Science and Technology. http://dx.doi.org/10.1080/10643389.2011.604258. Mosier, N., Wyman, C., Dale, B., Elander, R., Lee, Y.Y., Hotzapple, M., Ladish, M., 2005. Features of promising technologies for pretreatment of lignocellulosic biomass. Bioresource Technology 96, 673–686.

247

Nizami, A.S., Korres, N.E., Murphy, J.D., 2009. A review of the integrated process for the production of grass biomethane. Environmental Science and Technology 43 (22), 8496–8508. Nizami, A.S., Thamsiriroj, T., Singh, A., Murphy, J.D., 2010. The role of leaching and hydrolysis in a two phase grass digestion system. Energy and Fuels 24 (8), 4549–4559. Pakarinen, A., Zhang, J., Brock, T., Maijala, P., Viikari, L., 2012. Enzymatic accessibility of fiber hemp is enhanced by enzymatic or chemical removal of pectin. Bioresource Technology 107, 275–281. Pandey, K.K., Pitman, A.J., 2004. Examination of the lignin content in a softwood and a hardwood decayed by a brown-rot fungus with the acetyl bromide method and Fourier transform infrared spectroscopy. Journal of Polymer Science Part A: Polymer Chemistry 42, 2340–2346. Petersson, A., Thomsen, M.H., Hauggaard-Nielsen, H., Thomsen, A.B., 2007. Potential bioethanol and biogas production using lignocellulosic biomass from winter rye, oilseed rape and faba bean. Biomass and Bioenergy 31 (11–12), 812–819. Sambusiti, C., Ficara, E., Rollini, M., Manzoni, M.F., Malpei, F. 2011. Alkaline pretreatment of sorghum and wheat straw for increasing methane production. In: Proceeding of International Symposium on Anaerobic Digestion of Solid Waste and Energy Crops, Vienna, 28 August–1 September 2011, Austria. Shafiei, M., Karimi, K., Taherzadeh, M.J., 2010. Palm date fibers: analysis and enzymatic hydrolysis. International Journal of Molecular Sciences 11 (11), 4285–4296. Taherzadeh, M.J., Karimi, K., 2008. Pretreatment of lignocellulosic wastes to improve ethanol and biogas production: a review. International Journal of Molecular Sciences 9 (9), 1621–1651. Tong, X., Smith, L.H., McCarthy, P.L., 1990. Methane fermentation of selected lignocellulosic materials. Biomass 21 (4), 239. Triolo, J.M., Sommer, S.G., Moller, H.B., Weisbjerg, M.R., Jiang, X.Y., 2011. A new algorithm to characterize biodegradability of biomass during anaerobic digestion: influence of lignin concentration on methane production potential. Bioresource Technology 102 (20), 9395–9402. Yang, S.G., Li, J.H., Zheng, Z., Meng, Z., 2009. Lignocellulosic structural changes of Spartina alterniflora after anaerobic mono- and co-digestion. International Biodeterioration and Biodegradation 63 (5), 569–575. Zhu, L., O’Dwyer, J.P., Chang, V.S., Granda, C.B., Holtzapple, M.T., 2010. Multiple linear regression model for predicting biomass digestibility from structural features. Bioresource Technology 101 (13), 4971–4979.

Journal of Scientific & Industrial Research 778


Vol. 70, September 2011, pp. 778-783

Evaluation of different pretreatments to enhance degradation of pine needles by Aspergillus niger F7 under solid state fermentation Nivedita Sharma, Divya Tandon, Rakesh Gupta and Sanjeev Kumar Department of Basic Science, Dr Y S Parmar University of Horticulture and Forestry, Nauni, Solan (HP) 173 230, India Received 20 January 2011; revised 28 July 2011; accepted 01 August 2011 This study presents degradation enhancement of pine needles by Aspergillus niger F7 , isolated from soil. A modified alkali pretreatment [NaOH+H2 O2 (1M); ratio, 9:1] is found the best among all methods when needles are soaked in this solution for 2 h followed by thorough washing with tap water. Degradation of pine biomass was measured in terms of enzyme (cellulase & xylanase) production, biodegradation index (BI) and hydrolysis (%). There are high enzyme units, BI and hydrolysis (%) in pretreated material as compared to untreated one. Keywords: Biodegradation, Lignocellulose, Pretreatment, Solid State fermentation

Introduction Pine (Pinus roxburghii Sarg syn. P. longifolia Roxb.) is a predominant forest species, widely scattered in Alpine range globally. Accumulation of pine needles (PNs) on forest floor leads to infertility of soil and forest fire1 . PNs being rich in cellulose can be used as substrate for biodegradation. Among different physical and chemical pretreatment methods, solid state fermentation (SSF) of lignocellulosic material holds several advantages as compared to submerged fermentation (SmF). In SSF, enzymes produced are many folds more than SmF and thus it has direct impact on biodegradation of biomass2 . This study presents pretreated PNs as a substrate for degradation under SSF by a potential isolate to enhance hydrolysis of PNs. Experimental Section Extractives of Pine Needles (PNs)

PNs were collected from different forest site of northern India. Components [holocellulose (cellulose + hemicellulose), lignin and other extractives] of PNs were estimated by following standard methods of Technical Association of Pulp and Paper industry (TAPPI). For alcohol benzene extraction, oven dried PNs (2 g) were placed in a porous thimble and extractives were derived by TAPPI method3 . For holocellulose extraction, oven

*Author for correspondence Tel: +91-1792-252560; Fax: +91-1792-252242 E-mail: [email protected]

dried PNs (5 g), pre-extracted with alcohol benzene, were taken in 250 ml conical flask, distilled water (160 ml) was added and holocellulose was estimated following TAPPI method4 . For lignin extraction, oven dried PNs (2 g), pre-extracted with alcohol benzene, were treated with 15 ml of 72% sulphuric acid for 2 h at 18-20°C with constant stirring following TAPPI method5 . Pretreatment of Pine Needles (PNs) Grinding

Chipping of PNs gave small pieces, which were grinded (mesh size, 1.5 mm), soaked in water for 24 h, and then air dried for 24 h, followed by drying at 50°C overnight. Completely dried biomass was stored in air tight containers. Alkali Pretreatments

In NH3 pretreatment6 , PNs (10 g) were soaked in 100 ml of 1% ammonia solution. Under NH3 pretreatment (modified), PNs (10 g) were soaked in 100 ml of 5% ammonia solution for 2 h at room temperature (RT) and autoclaved for 15 min. After thorough washing with tap water (until solution became neutral) and dried at 50°C. Under NaOH+ H2 O2 pretreatment (modified), PNs (10 g) were soaked in NaOH+ H2 O2 solution (9:1) for 2 h at RT followed by washing with tap water and dried at 50°C. Acid Pretreatment

In hydrochloric acid (HCl) pretreatment6 , PNs (10 g) were soaked in 1% HCl solution (100 ml) for 2 h.


Table 1—Estimation of holocellulose and lignin in untreated and pretreated biomass of pine needles using TAPPI standard method Treatments Untreated 1%NH3 5%NH3 1%HCL 1%H2 SO4 1%NaOH+H2 O2

Holocellulose % 57.00 85.87 88.92 70.15 77.50 87.10

Lignin % 23.00 4.63 5.53 19.00 12.50 6.90

Extractives % 20.00 9.50 5.55 10.85 10.00 6.00

Estimation of Reducing Sugars and Soluble Proteins

Reducing sugars produced during degradation of PNs were estimated 11 . Soluble proteins formed during biodegradation were quantified by Lowry’s method12 . Biodegradation Index (BI)

BI13 of PNs is calculated as BI = [reducing sugar (%) released + protein (%) formed] / 2. Hydrolysis (%)

Hydrolysis% is calculated on dry matter basis as14 Hydrolysis (%)=

In sulphuric acid (H2 SO4 ) pretreatment6 , PNs (10 g) were soaked in 1% H2 SO4 solution (100 ml) for 2 h. Biodegradation of Pine Needles (PNs)

Aspergillus niger F7 , capable of producing high amount of hydrolytic enzymes (cellulase and xylanase), was procured from Microbiology laboratory of Basic Sciences, UHF Nauni, Solan (India). Biodegradation of PNs was studied under SSF by using water and modified basal salt medium (BSM) (1: 2). Modified BSM 7 contained Na2 HPO4 (6.0 g), KH2 PO4 (3.0 g), NaCl (0.5 g), NH2 CL (1.0 g) and separately sterilized solutions of 1 M MgSO 4 (2 ml) and 1 M CaCl2 (0.1 ml) were added after medium was autoclaved. It was supplemented with urea (2%), yeast extract (1%), peptone (0.1%), NaNO3 (0.1%), 1M CoCl2 (0.2/l) with pH 6.80 to final volume of 1000 ml. To each 20 g of untreated and pretreated biomass of P. roxburghii, water (35 ml) and of inoculum (5 ml) containing 1x107 spores/ml of A. niger were added in 500 ml of Erlenmeyer flask. Flasks were incubated for 30 days at 28 ± 2°C. Extraction of Enzymes

Hydrolytic enzymes and other fermented products produced during biodegradation of PNs were extracted by Repeated Extraction Method8 . To 5 g of biomass, 50 ml of phosphate buffer (0.1M, pH 6.9) was added in 250 ml Erlenmeyer flask and contents were kept at 120 rpm for 1 h and then filtered through muslin cloth. The process was repeated twice with additional 25 x 2 ml of phosphate buffer making final volume of extracted products to 100 ml. After filtration, contents were centrifuged at 5000 rpm for 5 min at 4°C. Supernatant was collected to estimate enzymes, biodegradation index (BI) and hydrolysis (%). Enzyme assays were performed to quantify CMCaseÿÿ:9 , FPAase9 and ß-glucosidase of cellulase10 and xylanase11 activities.

779

Total reducing sugar (g)× 0.90 × 100 Weight of substrate (g)

Statistical Analysis

Completely randomized design was applied. Different regression models (linear, power, exponential and quadratic) were used to predict hydrolysis (%) and BI activities on the basis of enzyme activities in two mediums (water and modified BSM). Results and Discussion Estimation of Different components in Pine Needles (PNs)

Analysis of untreated biomass of PNs gave: holocellulose, 57.00; lignin, 23.00; and extractives (alcohol, benzene, fibers, resins etc.), 20.00% (Table 1). For efficient biodegradation of holocellulose of PNs, different pretreatments were given to wash lignin and extractives out of PNs. Holocellulose was found highest in alkali treated biomass of PNs (5% NH3 , 88.92; NaOH+H 2 O2 , 87.10; and 1% NH3 , 85.87%). Maximum lignin (19%) was retained by HCl pretreated material while lowest (4.63%) was in 1% NH3 pretreated needles. Extractives were: 1% NH3 , 9.50; 5% NH3 , 5.55; and NaOH+H 2 O 2 , 6.00%. Similar studies showing an increase in cellulose contents and decrease in lignin after pretreating wood biomass has earlier been reported15 . Thus different pretreatments temper lignin shield and take it out of lignocellulosic materials, thereby exposing most of the cellulose in active form for better enzymatic digestion. All pretreated PNs though have shown higher saccharification with enzymes secreted from A. niger during degradation but their BI and hydrolysis% values vary from treatment to treatment (Table 2). SSF of pretreated and untreated materials was carried out under substrate: moisture ratio (1: 2), which has been optimized for other lignocellulosic forest wastes7 . Moistening agents (tap water and modified BSM) were used during SSF with an ultimate aim to enhance biodegradation of PNs.

780


Table 2—Enzyme activity, biodegradation index (BI) and hydrolysis% of pine needles (Pinus roxburghii) after solid state fermentation by Aspergillus niger F7 using water and modified BSM as medium Treatments

Untreated 1% NH3 5% NH3 1% HCl 1% H2 SO4 NaOH+H 2 O2

H2 O as moistening agent Total Enzyme U/g 28.16 49.90 72.28 44.91 51.21 96.14

BSM as moistening agent

BI

Hydrolysis%

7.16 13.00 17.20 8.00 9.00 27.76

0.89 2.85 3.67 1.22 1.40 7.00

Total Enzyme U/g 95.40 144.80 185.06 106.00 137.80 259.60

BI

Hydrolysis%

13.45 26.84 33.40 15.56 16.56 53.30

1.85 5.88 7.08 2.36 2.72 12.87

U/g (on dry matter) of hydrolytic enzymes.

When water was used as moistening agent, degradation was very low in case of untreated biomass as follows: enzyme production, 28.16 U/g; BI, 7.16; and hydrolysis, 0.89%. Among pretreatments, NaOH+H 2 O2 pretreated PNs had led to maximum values as follows: cellulase and xylanase production from A. niger, 96.14 U/g; BI, 27.76; and hydrolysis, 7%. Biodegradation carried out with HCl pretreated biomass of PNs gave minimum values as follows: enzyme production, 44.91 U/g; BI, 8; and hydrolysis, 1.22%. When modified BSM was used as moistening agent, NaOH+H 2 O2 pretreated PNs have led to maximum cellulase and xylanase production from A. niger (259.6 U/g), thus resulting in highest BI (53.30) and hydrolysis 12.87%. Acid pretreatments showed comparatively lower production of enzymes, BI and hydrolysis% as compared to alkali pretreatments. On the other hand, untreated biomass in modified BSM also has shown least production of enzyme as compared to all pretreated biomass of PNs, consequently resulting in marginally low degradation as follows: enzyme production, 95.40 U/g; BI, 13.45; and hydrolysis, 1.85%. Since PNs are exceptionally inert biomass for biodegradation, therefore alkali pretreatment of PNs [NH3 (1%, 5%) and NaOH + H2 O2 ] has been chosen with an idea of removing maximum lignin and other hindering substances like resins etc. Alkali pretreatment is reported to decrease crystallinity of cellulose, remove lignin shield around cellulose and increase pore size of biomass, thus increasing digestibility of lignocelluloses16 . Compared with acid or oxidative reagents, alkali pretreatment appears to be the most effective methods in breaking ester bonds between lignin, hemicellulose and cellulose and avoiding fragmentation of hemicellulose polymers17 . Alkaline pretreatment in combination with

H2 O2 (NaOH + H2 O2 ) additionally promotes to loosen the linkage of hydrogen bonds, resulting in easy enzymatic hydrolysis of biomass18 . Alkaline pretreated biomass is hydrolyzed 40% faster than native cellulose. Under acid treatment of lignocellulosic materials, sulphuric acid is the most applied acid 19 and found most effective in dissolving lignin, and thus increasing cellulose’s susceptibility to enzymatic attack20 . Pretreatment consists of collection, transportation, manipulation, storage, grinding or chipping to reduce particle size and opening fibrous material in order to transform it into a suspension that can be pumped and enable further penetration of chemical hydrolysis agents21 . Overall an appreciable increase has been observed in pretreated PNs as compared to untreated ones. When water was used as moistening agent, in acidic pretreatment (HCl & H2 SO4 ), increase was observed as follows: enzyme activity, 59.48, 81.85%; BI, 11.73, 25.69% and hydrolysis, 37.07, 57.30%. In NaOH + H 2 O2 pretreated PNs, increase was observed as follows: enzyme, 241.40; BI, 287.70; and hydrolysis, 686.51%. When modified BSM was used, in NaOH + H2 O 2 pretreated PNs, increase was observed as follows: enzyme production, 172.11; BI, 296.28; hydrolysis, 595.67% (Fig. 1). A positive correlation has been derived between enzyme activity, BI and hydrolysis% of PNs. Thus when enzyme activity increases, BI and hydrolysis% also increases (Fig. 2). Parameters of various regression models and R2 for estimation of BI, hydrolysis% and enzyme activities show a direct correlation between these parameters (Table 3). Different regression models were tried and higher value of R 2 was found in quadratic model (Y = a + bx + cx 2 ) for prediction of BI activities on the


781

Increase in enzyme activity, %

300 250 200 150 100 50 0 Ammonia (1%)

Ammonia (5%)

Hydrochloric acid

Sulphuric acid

Pretreatments a)


350

Increase in BI, %

300 250 200 150 100 50 0 Ammonia (1%)

Ammonia (5%)

Hydrochloric acid

Sulphuric acid


Pretreatments b) 800

Increase in hydrolysis, %

700 600 500 400 300 200 100 0 Ammonia (1%)

Ammonia (5%)

Hydrochloric acid

Sulphuric acid


Pretreatments c) Water

BSM

Fig.1—Comparison in pretreated pine needles by Aspergillus niger F7 using water and modified BSM as medium over untreated pine needles of increase in: a) Total enzyme activity; b) BI; and c) Hydrolysis%

782 120

30

100

25

80

20

60

15

40

10

20

5

0

% hydrolysis, B. I.

Enzyme activity


0

Ammonia (1%)

Ammonia (5%)

Hydrochloric acid

Sulphuric acid


300

60

250

50

200

40

150

30

100

20

50

10

0

% hydrolysis, B. I.

Enzyme activity

Pretreatments a)

0

Ammonia (1%)

Ammonia (5%)

Hydrochloric acid

Sulphuric acid


Pretreatments b) Fig. 2—Correlation between enzyme activity, B.I. and per cent hydrolysis of pine needles after SSF with Aspergillus niger F7 using: a) water as medium; and b) Modified BSM as medium Table 3—Parameters of various models to predict enzyme activity, BI and hydrolysis% Characters

Prediction model for BI activities and per cent hydrolysis (Y) on the basis of enzyme activities Y=a+bX R2 Y=abX R2 Y=aX b R2 Y=a+bX+cX 2 R2 a b a b a b a b Water -6.48 4.37 0.95 5.11 1.21 0.92 2.264 1.183 0.915 1.204 1.844 0.968 BSM -2.66 1.121 0.90 1.82 0.66 0.85 0.01 0.22 0.86 -0.289 0.339 0.93 X=Enzyme activity, Y=BI/hydrolysis, R2 =Coefficient of determination

basis of enzyme activities of water as well modified BSM. Thus quadratic model can be used for the prediction of enzyme activities, BI and hydrolysis%. Thus it has been established that with increase in extracellular cellulase

Y=aExpbX a b 5.11 0.196 1.82 0.664

R2 0.920 0.85

and xylanase production from hydrolytic microorganisms, biodegradation of PNs is enhanced. Though biodegradation of lignocellulosic wastes like agricultural biomass (corn cob, corn straw, baggase) and


other forest residues has already been reported22 , but successful biodegradation of PNs are rarely reported. This study strongly proves that pretreated PNs under SSF with A. niger F7 can serve as an inexpensive substrate for its saccharification into fermentable sugars, which in turn can be fermented to ethanol to be used as biofuels. Conclusions PNs, which are highly resistant to biodegradation, can be degraded successfully with A. niger after suitable pretreatment. Modified alkali pretreatment [NaOH+H 2 O2 (1M); ratio, 9:1] in 2 h followed by steam explosion for 15 min has been found the best pretreatment for PNs hydrolysis by a hypercellulolytic isolate, A. niger F7 . Modified BSM mediated SSF was found better over tap water. A positive correlation is drawn in three parameters (enzyme activity, BI and hydrolysis%) and has been proved statistically by using regression model. References 1 2

3

4

5 6

7

8

1 Bhasin R, Forest fire ravages Himachal flora and fauna, Biores Technol, 5 (2008) 39-45. Kondo P, Investigation on mechanism of biological dezincification by solid state fermentation, J Sci Ind Res, 55 (1996) 394-399. Alcohol-benzene solubility of wood, Official Standards, T6 M59 [Technical Association of Pulp and Paper Industry (TAPPI), New York] 1950. Holocellulose in wood, Official Standards, T12M-59 [Technical Association of Pulp and Paper Industry (TAPPI), New York] 1954. Lignin in wood, Official Standards, T12M-59 [Technical Association of Pulp and Paper Industry (TAPPI), New York] 1959. Fan L T Y H, Gharpuray M M & Beard M D H, The nature of lignocellulosics and their pretreatments for enzymatic hydrolysis, Adv Biochem Bioeng, 23 (1982)157-187. Sharma N, Bansal K L & Neopaney B, Effect of moisture level on biodegradation of forest waste under solid state fermentation, J Sci Ind Res, 65 (2006)675-679. Bollag D M & Edestein S J, Protein Methods (Wiley - Liss, John Wiley and Sons Inc, New York) 1993, 230.

9

10

11

12

13

14 15

16 17

18

19

20

21

22

783

Reese E T & Mandel M, Enzymatic hydrolysis of cellulose and its derivatives, in Methods Carbohydrate Chemistry, 3rd edn, edited by R L Whistler (Acad Press, London) 1963, 139-143. Bergham L E R & Petterson L G, Mechanism of enzymatic cellulose degradation: Purification of cellulolytic enzyme from Trichoderma viride active on highly ordered cellulose, J Biochem, 37 (1973) 21-30. Miller G L, Use of dinitrosalicylic acid reagent for determination of reducing sugars, Analyt Chem, 31 (1959) 426-428. Lowry O H, Rosebrough N J, Farr A L & Randall R J, Protein measurement with folin phenol reagent, J Biol Chem, 193 (1951) 265-275. Sharma N, Bhalla T C, Aggarwal H O & Bhatt A K, Saccharification of physico-chemically pretreated lignocellulosics by partially purified cellulase of Trichoderma viride, Sci Lett, 19 (1996) 141-144. Szczodrak J, Rogalski J & Liczuk Z, Cellulolytic activity of molds, Acta Microbiol Polonica, 33 (1984) 217-225. Jan K, Lisbeth G, Thygesen C, Henning J & Thomas E, Cell wall structural changes in wheat straw pretreated for bioethanol production, Biotech Biofuels, 24 (2008) 1-5. Fan I T, Lee Y H & Beard D H, Mechanisms of the enzymatic hydrolysis of cellulose, Biotechnol Bioeng, 22 (1980) 177-199. Gaspar M, Kalman G & Reczey K, Corn fiber as a raw material for hemicellulose and ethanol production, Process Biochem, 42 (2007) 1135-1139. Saha B C & Cotta M A, Ethanol production from alkaline peroxide pretreated enzymatically saccharified wheat straw, Biotech Prog, 22 (2006) 449-453. Taherzadeh M J & Karimi K, Acid based hydrolysis process for ethanol from lignocellulosic materials, Rev Biores, 2 (2007) 472-499. Yang B & Wyman C E, Effect of xylan and lignin removal by batch and flow through pretreatment on the enzymatic digestibility of corn stover cellulose, Biotech Bioeng, 86 (2004) 88-95. Muzzy J D, Robertis R S, Fiebe C A, Fieber G S & Mann T M, Pretreatments of hard wood by continuous hydrolysis, in Wood and Agricultural Residues, vol 25 (Acad Press, London) 2009, 351-368. Damisa D & Ameh J B, Effect of chemical pretreatment of some lignocellulosic waste on recovery of cellulase from Aspergillus niger, Appl Microbiol, 22 (2008) 209-213.




Effect of fungal and phosphoric acid pretreatment on ethanol production from oil palm empty fruit bunches (OPEFB) Mofoluwake M. Ishola a,b,⇑, Isroi c, Mohammad J. Taherzadeh a a

Swedish Centre for Resource Recovery, University of Borås, Sweden Department of Chemical and Polymer Engineering, Lagos State University, Nigeria c Indonesian Biotechnology Research Institute of Estate Crops, Bogor, Indonesia b

h i g h l i g h t s

g r a p h i c a l a b s t r a c t

OPEFB was pretreated with

phosphoric acid, white rot fungi and their combination. Phosphoric acid pretreatment improved the digestibility of OPEFB by 24 times. Fungi pretreatment gave the highest recovery of 98.7% of OPEFB. Ethanol yield of 89.4% of the theoretical value was obtained after pretreatment.

a r t i c l e

i n f o

Article history: Received 12 December 2013 Received in revised form 11 February 2014 Accepted 14 February 2014 Available online 22 February 2014 Keywords: Oil palm empty fruit bunches Ethanol White rot fungi Phosphoric acid Lignocellulose pretreatment

a b s t r a c t Oil palm empty fruit bunches (OPEFB), a lignocellulosic residue of palm oil industries was examined for ethanol production. Milled OPEFB exposed to simultaneous saccharification and fermentation (SSF) with enzymes and Saccharomyces cerevisiae resulted just in 14.5% ethanol yield compared to the theoretical yield. Therefore, chemical pretreatment with phosphoric acid, a biological pretreatment with white-rot fungus Pleurotus floridanus, and their combination were carried out on OPEFB prior to the SSF. Pretreatment with phosphoric acid, combination of both methods and just fungal pretreatment improved the digestibility of OPEFB by 24.0, 16.5 and 4.5 times, respectively. During the SSF, phosphoric acid pretreatment, combination of fungal and phosphoric acid pretreatment and just fungal pretreatment resulted in the highest 89.4%, 62.8% and 27.9% of the theoretical ethanol yield, respectively. However, the recovery of the OPEFB after the fungal pretreatment was 98.7%, which was higher than after phosphoric acid pretreatment (36.5%) and combined pretreatment (45.2%). Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction Ethanol is the most important renewable fuel in the transportation sector considering its volume and market value, as its global production reached 85 billion liters in 2012 (Renewable Fuels Association, 2012). Ethanol production from lignocellulosic biomass has been preferred than production from starch or ⇑ Corresponding author at: Swedish Centre for Resource Recovery, University of Borås, Sweden. Tel.: +46 33 435 4585; fax: +46 33 435 4008. E-mail address: [email protected] (M.M. Ishola). http://dx.doi.org/10.1016/j.biortech.2014.02.053 0960-8524/Ó 2014 Elsevier Ltd. All rights reserved.

sugar-based crops, as it does not compete with food and takes care of agricultural and plant residues in an environmentally sustainable process (Farrell et al., 2006). Oil palm empty fruit bunches (OPEFB) is a lignocellulosic residue of palm oil mill. Nigeria produces 8 million metric tons of palm oil fruit in 2011, resulting in about 2 million metric tons of OPEFB yearly. Indonesia and Malaysia also produce about 90 million metric tons/year palm oil fruit resulting in accumulation of about 20.7 million metric tons of OPEFB every year (FAOSTAT, 2012). OPEFB constitutes environmental problem and has low economic value; it is conventionally disposed in landfills, burned openly or used

10

M.M. Ishola et al. / Bioresource Technology 165 (2014) 9–12

in composting for fertilizer. OPEFB primarily contains 82.4% hollocellulose and 17.6% lignin (Law et al., 2007). Considering the high carbohydrate content of OPEFB, it can be a cheap feedstock for ethanol and lignocellulosic derivatives such as glucose, xylose, mannose and animal feed (Piarpuzán et al., 2011) Ethanol production from lignocellulosic biomass involves four basic steps, which are pretreatment, hydrolysis, fermentation and distillation. Lignocellulosic materials are recalcitrant in nature due to the lignin content and the highly crystalline nature of the cellulose (Himmel and Picataggio, 2009). Hence, an efficient pretreatment method is necessary for the process optimization. Lignocellulosic biomass can be pretreated by chemical, mechanical, biological methods, or their combinations (Alvira et al., 2010). Biological pretreatments involves the use of microorganisms usually among white-rot, brown-rot and soft-rot fungi or their enzyme to break down lignin barrier and alter lignocelluloses structure (Adney et al., 2009). Biological pretreatment requires no chemicals, low energy input and seems to be the best environmentally friendly method of pretreatment (Kirk and Chang, 1981). Whiterot fungi are the primary agents of delignification that function through their extracellular enzymes namely Lignin peroxidase (LiP), Laccases (Lac) and Manganese peroxidase (MnP), which degrade lignin macromolecules into CO2 and H2O (Isroi et al., 2011). Chemical pretreatment of lignocellulosic biomass using phosphoric acid has been shown to fractionate the lignocellulosic biomass and enhanced cellulose digestibility (Zhang et al., 2007). To our knowledge, combination of biological pretreatment using white-rot fungi and chemical pretreatment using phosphoric acid for ethanol production has not been reported in the literature. This study aims at evaluating the effect of biological pretreatment using white-rot fungi, phosphoric acid pretreatment and their combination on ethanol production from OPEFB during simultaneous saccharification and fermentation (SSF). Effect of these pretreatments on the digestibility and structural changes of OPEFB were also investigated.

2. Methods 2.1. Oil palm empty fruit bunches (OPEFB) Fresh OPEFB was obtained from an oil palm mill in North Sumatera, Indonesia. It was air-dried until the moisture content was about 35%, and then grounded to pass through a 10 mm screen. It was stored in an airtight plastic container before the pretreatments. It was analyzed for its cellulose, hemicellulose and lignin contents.

2.3. Fungal pretreatment Medium containing 7 g/L KH2PO4, 1.5 g/L MgSO47H2O, 1 g/L CaCl2H2O, 0.3 g/L MnSO4H2O, and 0.3 g/L CuSO4H2O was prepared. A volume of 120 mL of this medium was added to 200 g of air dried OPEFB (moisture content 35%) in order to reach moisture content of 59.4%. It was inoculated a with two-week old fungus inoculum on agar plate for four weeks at 31 °C and neutral pH in a 500 mL Erlenmeyer flask with cotton plugs, followed by freezing the culture to stop the fungal growth. The fungal-pretreated OPEFB was then freeze-dried (Freezone 7670530, Labconco, Kansas City, MO, USA) at 52 °C for 6 h. Material weight before and after the pretreatment process was recorded. 2.4. Phosphoric acid pretreatment Phosphoric acid pretreatment was carried out on both fungalpretreated OPEFB and untreated OPEFB according to Zhang et al. (2007) with slight modifications. Both fungal-pretreated OPEFB and untreated OPEFB were ball-milled separately (RetschÒ MM400) at a frequency of 29.6 s1 for 4 min before the pretreatment with phosphoric acid. One gram of the milled materials was mixed with 8 mL phosphoric acid (85.7%) in a 50 mL-centrifuge tube and stirred using a glass rod until it was homogenised. The mixture was then incubated in a shaking water bath (Grant OLS200, Grant Instruments Ltd., Cambridgeshire, UK) at 90 rpm and 50 °C for 5 h after which it was washed by addition of 40 mL acetone and centrifuged at 1900g, for 15 min. The remaining acid-free pellets were then washed three times using 40 mL distilled water until a clear supernatant with neutral pH was obtained. The pretreated materials were then frozen in airtight container until use in the enzymatic hydrolysis and SSF. 2.5. Enzymatic hydrolysis Enzymatic hydrolysis of untreated and pretreated OPEFB were carried out according to the NREL method with slight modifications (Selig et al., 2008). Pretreated and untreated OPEFB were hydrolyzed in a 50 mL E-flask with a working volume of 10 mL at pH 4.8 using a water bath at temperature of 50 °C and agitation of 150 rpm for 72 h using enzyme loading of 60 FPU/g cellulose. Samples were taken at 12 h interval and analyzed for glucose using the auto-sampler high performance liquid chromatography (HPLC) with a lead-based column (Aminex HPX-87P, Bio-Rad, Hercules, CA, USA). Digestibility (%) was calculated based on the following equation (Selig et al., 2008);

Digestibility ð%Þ ¼

Glucose produced upon hydrolysis ðgÞ Initial cellulose in the substrate ðgÞ 1:11 100%

2.2. Microorganisms and enzymes White-rot fungus Pleurotus floridanus LIPIMC996, obtained from Laboratory of Microbial Systematic and LIPI Microbial Collection (Lembaga Ilmu Pengetahuan, Cibinong, Indonesia) was used in this work. The fungus was maintained on lignocellulosic medium at room temperature before it was used as inoculums. Yeast strain Saccharomyces cerevisiae CBS 8066 obtained from Centraalbureau voor Schimelcultures (Delft, the Netherlands) was used for the fermentation experiments. It was maintained on YPD agar plate containing 20 g/L agar (Scharlau), 10 g/L yeast extract (Scharlau), 20 g/L peptone (Fluka), 20 g/L D-glucose (Scharlau) and stored at 4 °C. Cellulase enzyme CellicÒ Ctec2 (Novozymes, Denmark) was used for the hydrolysis. The enzyme had 168 FPU/mL activity, determined according to the National Renewable Energy Laboratory (NREL) method (Adney and Baker, 2008).

ð1Þ where 1.11 = conversion factor for cellulose hydrolysis to glucose. All experiments were performed in duplicate and the average values are reported, while the error bars show standard deviation. 2.6. Simultaneous saccharification and fermentation (SSF) SSF was performed in a 100 mL Erlenmeyer flask with 20 mL working volume for all the samples (Dowe and McMillan, 2008). Untreated, fungal-pretreated, phosphoric acid-pretreated and fungal followed by phosphoric acid pretreated OPEFB’s were added in order to obtain 5% glucan. Medium containing 100 g/L yeast extract (Scharlau) and 200 g/L peptone (Fluka) was also prepared, as a nutrient supplement for the yeast growth and 2 mL of this was added to each of the flasks containing the pretreated and untreated

11


OPEFB. For pH control, 1 mL of 50 mM citrate buffer (pH 4.8) was added to all the flasks, 60 FPU/g cellulose enzymes loading and 0.2 mL of the inoculums was added to each of the flasks and deionized water was used to make up the volume to 20 mL. SSF was performed at temperature of 35 °C for 96 h in a water bath at agitation of 130 rpm, samples were taken at 24 h interval and analyzed for ethanol concentration with the HPLC. The ethanol yield was calculated based on the cellulose content of the untreated and the pretreated OPEFB’s. 2.7. Compositional analysis

Table 1 Composition of empty fruit bunches (OPEFB) before and after pretreatment. Component (%)

Untreated OPEFB

Fungal pretreated OPEFB

Phosphoric acid pretreated OPEFB

Fungal and phosphoric acid pretreated OPEFB

ASL AIL Total lignin Cellulose Hemicellulose Material loss

7.81 ± 0.03 26.56 ± 0.14 34.37 ± 0.17 39.13 ± 2.26 23.04 ± 2.79 –

8.39 ± 0.40 25.95 ± 0.36 34.34 ± 0.36 34.17 ± 0.40 27.48 ± 9.07 1.31 ± 0.13

4.30 ± 0.09 40.13 ± 0.51 44.66 ± 0.18 43.16 ± 0.43 9.07 ± 0.05 54.84 ± 1.37

4.53 ± 0.05 32.92 ± 0.60 37.22 ± 0.51 53.81 ± 1.14 9.07 ± 0.14 63.55 ± 0.76

ASL = acid soluble lignin; AIL = acid insoluble lignin; total lignin = ASL + AIL.

The OPEFB’s before and after pretreatments were characterized. Cellulose, hemicellulose, and lignin content of the untreated, fungal-pretreated and phosphoric acid-pretreated OPEFB were determined according to the NREL method (Sluiter et al., 2011).

A Fourier transform infrared (FTIR) spectrometer (Impact 410 iS10, Nicolet Instrument Corp., Madison, WI, USA) was used for investigating the changes of OPEFB after the pretreatments. Spectrum of each of the pretreated and untreated OPEFB samples was obtained with an average of 32 scans and resolution of 4 cm1 from 600 to 4000 cm1 (Jeihanipour et al., 2010). The spectrum data was controlled by Nicolet OMNIC 4.1 (Nicolet Instrument Corp., Madison, USA) software and analyzed by eFTIRÒ (EssentialFTIR, USA). The crystallinity of untreated and pretreated OPEFB’s cellulose was determined using crystallinity index as A1429/A894 which is the ratio between absorbances at wavenumber 1418 and 894 cm1 (O’Connor et al., 1958)

pretreated OPEFB shows decrease in the percentage of ASL (4.4%) compared with untreated OPEFB (7.8%) as shown in Table 1. It also shows an increase in percentages of cellulose (43.2%) compared with untreated materials (39.1%), which implies that phosphoric acid pretreatment can successfully fractionates OPEFB and opens up the crystallinity of its cellulose as earlier reported (Zhang et al., 2007). Combination of fungal and phosphoric acid pretreated OPEFB (Table 1) also shows a decrease in the percentage of the ASL (4.5%) compared to untreated OPEFB (7.8%). The percentage of the cellulose after combined pretreatment was increased (53.8%) compared to untreated OPEFB (39.1%). The material loss for fungal, phosphoric and the combination of fungal and phosphoric pretreatments were 1.3%, 54.8% and 63.6%, respectively. This indicates that fungal pretreatment preserves the material, while phosphoric acid pretreatment results in high loss of material.

3. Results and discussion

3.3. Effect of the pretreatments on functional groups in OPEFB

3.1. Effect of different pretreatment on OPEFB digestibility

The changes of OPEFB functional groups were analyzed based on FTIR spectra of the untreated and pretreated materials. The peaks that correspond to cellulose structure were 3338, 2980–2835, 1460, 1375, 750, and 715 cm1, whereas the peaks that correspond to hemicellulose or carbohydrate structures were 1738–1709, 1315, 1230–1221, 1162–1125 and 897 cm1. Wavenumber around 3300 cm1 was assigned to hydrogen bonded (O–H) stretching absorption. The highest cellulose loss was observed in combination of fungal and phosphoric acid pretreatment as indicated with the lowest intensity on O–H stretching absorption. Reduction in the peak of 3338 cm1 indicates a reduction in the hydrogen bond of cellulose of the pretreated OPEFB. Changes in intensity were also found in the band at around wavenumber 1032 cm1 that was assigned to C–O stretch in cellulose and hemicellulose. Intensity at this band increased after fungal pretreatment but reduced after phosphoric acid and the combined pretreatment, also indicating a higher percentage of hemicellulose loss after phosphoric acid and the combined pretreatment. Both phosphoric acid pretreatment and the combined pretreatment showed similar intensities at wavenumbers 1646, 1607, 1593, and 1506 cm1. These spectra explained the results shown in Table 1 that phosphoric acid pretreatment and combined pretreatment resulted to high loss of material and similar percentage of ASL. Crystallinity index of untreated, fungal pretreated, phosphoric acid pretreated, and combined pretreated OPEFB are 2.8, 1.4, 0.7, and 0.6, respectively. This further strengthened the fact that phosphoric acid pretreatment is effective in reducing the crystallinity of cellulose in OPEFB, which suggest the reason for its highest digestibility. However, this crystallinity index cannot be alone an index on the effectiveness of the enzymatic digestibility of the OPEFB,

2.8. Structural and crystallinity analysis

Untreated OPEFB has digestibility of only 3.4%, fungal pretreatment improved the digestibility to 15.4%, which is 4.5 times higher than the untreated. As the fungus primarily deligninfies lignocelluloses, this improvement could most likely related to the reduction of the lignin by fungal pretreatment. Phosphoric acid pretreatment significantly increased the digestibility to 81.4% and combined pretreatment resulted in 56.1% digestibility. It means the acid could improve the digestibility by 24.0 times. Digestibility of combined pretreatment was increased 16.5 times. Phosphoric acid pretreatment shows highest improvement of digestibility and could be related to the reduction of the crystallinity of OPEFB cellulose, this corresponds with earlier report about phosphoric acid pretreatment reducing the crystallinity of cellulose (Jeihanipour et al., 2010). 3.2. Effects of different pretreatment on the composition of OPEFB biomass The different pretreatments show specific effects on the composition of OPEFB. After fungal pretreatment, the hemicellulose content increased from 23.1% in the untreated to 27.5% in the fungal pretreated OPEFB while cellulose content reduced from 39.1% in the untreated to 34.2% (Table 1), this could be due to attack on the linkages between lignin and carbohydrate that exist in hemicellulose by the fungi. There was a reduction in the ASL lignin of fungal pretreated OPEFB (25.9%) compared to untreated one (26.6%), which indicates that white rot fungi may have degraded the lignin. This is in agreement with previous report on lignin degradation by white rot fungi (Isroi et al., 2011). Phosphoric acid

12


62.8% and 27.9%, respectively. Pretreatments with phosphoric acid, combination of both methods and fungal pretreatment improved the digestibilities of OPEFB by of 24, 16.5 and 4.5 times, respectively. Acknowledgements This work was financially supported by European Commission program EM-Euro Asia, and Swedish Energy Agency. The authors are grateful to Novozymes (Denmark) for supplying the enzymes. References

Fig. 1. Concentration (g/L) of ethanol from untreated-, fungal pretreated, phosphoric acid pretreated-, and combination of fungal and phosphoric acid pretreated OPEFB biomass during simultaneous saccharification and fermentation (SSF) with cellulase and baker’s yeast.

as the combined method has lower index than phosphoric acid pretreatment. 3.4. Effect of the pretreatments on ethanol production during Simultaneous saccharification and fermentation (SSF) Without any pretreatment, the SSF of OPEFB resulted in the highest ethanol concentration of only 4.1 g/L (Fig. 1) in 96 h, which was equal to a yield of only 14.5% of the theoretical ethanol yield. The highest ethanol concentrations of 6.8, 20.3 and 21.8 g/L were produced from fungal, phosphoric and combined pretreatments, respectively. Fungi pretreatment slightly improved the ethanol yield (27.9%), achieved in 72 h. However, phosphoric acid pretreatment greatly improved the ethanol yield to 89.4% obtained in 48 h. Ethanol yield of 62.8% was obtained from the combined pretreatment in 48 h of SSF. This indicates that phosphoric acid pretreatment of OPEFB or its combination with fungal pretreatment can greatly increase the rate of ethanol production and approach the theoretical yield. The reduction in the ethanol yield of the combined pretreatment compared to the phosphoric acid pretreatment could be related to 63.6% for its material loss (Table 1). Comparing to previous reports, ethanol yield of 89.4% within 48 h from phosphoric acid pretreated OPEFB shows a similar yield compared to 90% reported from two stage hydrolysis of OPEFB by Millati et al. (2011). One the other hand this study shows an improved yield than previous reports of 65.6% ethanol yield at 168 h when OPEFB was pretreated with aqueous ammonia (Jung et al., 2011) and 70.6% ethanol yield at 95 h when OPEFB was pretreated with alkali (Park et al., 2013). 4. Conclusion Enzymatic hydrolysis and fermentation of palm oil empty fruit bunches (OPEFB) gave ethanol yield of 14.5% without any pretreatment. It was therefore pretreated with white-rot fungi and phosphoric acid as well as the combination of the two methods in order to improve the ethanol yield. Phosphoric acid pretreatment resulted in the highest yield of 89.4% within 48 h, while combined pretreatment and fungi pretreatment resulted in ethanol yield of

Adney, B., Baker, J. 2008. Measurement of Cellulase Activities. Laboratory Analytical Procedure, NREL/TP-510-42628. Adney, W.S., van der Lelie, D., Berry, A.M., Himmel, M.E., 2009. Understanding the Biomass Decay Community. Biomass Recalcitrance. Blackwell Publishing Ltd., pp. 454–479. Alvira, P., Tomás-Pejó, E., Ballesteros, M., Negro, M.J., 2010. Pretreatment technologies for an efficient bioethanol production process based on enzymatic hydrolysis: a review. Bioresour. Technol. 101 (13), 4851–4861. Dowe, N., McMillan, J. 2008. SSF Experimental Protocols — Lignocellulosic Biomass Hydrolysis and Fermentation. Laboratory Analytical Procedure, NREL/TP-51042630. FAOSTAT ‘‘Country Crop Production’’, 2012. Food and Agriculture Organization of the United Nations. http://faostat.fao.org/site/291/default.aspx (accessed 16.07.13). Farrell, A.E., Plevin, R.J., Turner, B.T., Jones, A.D., O’Hare, M., Kammen, D.M., 2006. Ethanol can contribute to energy and environmental goals. Science 311 (5760), 506–508. Himmel, M.E., Picataggio, S.K., 2009. Our Challenge is to Acquire Deeper Understanding of Biomass Recalcitrance and Conversion. Biomass Recalcitrance. Blackwell Publishing Ltd., pp. 1–6. Isroi Millati, R., Syamsiah, S., Niklasson, C., Cahyanto, M.N., Lundquist, K., Taherzadeh, M.J., 2011. Biological pretreatment of lignocelluloses with whiterot fungi and its applications: a review. BioResources 6 (4), 5224–5259. Jeihanipour, A., Karimi, K., Taherzadeh, M.J., 2010. Enhancement of ethanol and biogas production from high-crystalline cellulose by different modes of NMO pretreatment. Biotechnol. Bioeng. 105 (3), 469–476. Jung, Y.H., Kim, I.J., Han, J.-I., Choi, I.-G., Kim, K.H., 2011. Aqueous ammonia pretreatment of oil palm empty fruit bunches for ethanol production. Bioresour. Technol. 102 (20), 9806–9809. Kirk, T.K., Chang, H.-M., 1981. Potential applications of bio-ligninolytic systems. Enzyme Microb. Technol. 3 (3), 189–196. Law, K.-N., Daud, W.R.W., Ghazali, A., 2007. Morphological and chemical nature of fiber strands of oil palm empty-fruit-bunch (OPEFB). BioResources 2 (3), 351– 362. Millati, R., Wikandari, R., Trihandayani, E.T., Cahyanto, M.N., Taherzadeh, M.J., Niklasson, C., 2011. Ethanol from oil palm empty fruit bunch via dilute-acid hydrolysis and fermentation by Mucor indicus and Saccharomyces cerevisiae. Agric. J. 6 (2), 54–59. O’Connor, R.T., DuPré, E.F., Mitcham, D., 1958. Applications of infrared absorption spectroscopy to investigations of cotton and modified cottons. Text. Res. J. 28 (5), 382–392. Park, J.M., Oh, B.-R., Seo, J.-W., Hong, W.-K., Yu, A., Sohn, J.-H., Kim, C.H., 2013. Efficient production of ethanol from empty palm fruit bunch fibers by fed-batch simultaneous saccharification and fermentation using Saccharomyces cerevisiae. Appl. Biochem. Biotechnol. 170 (8), 1807–1814. Piarpuzán, D., Quintero, J.A., Cardona, C.A., 2011. Empty fruit bunches from oil palm as a potential raw material for fuel ethanol production. Biomass Bioenergy 35 (3), 1130–1137. Renewable Fuels Association, 2012. ‘‘Global ethanol production to reach 85.2 billion litres in 2012’’. http://www.ethanolrfa.org/ (accessed 21.07.13). Selig, M., Weiss, N., Ji, Y. 2008. Enzymatic Saccharification of Lignocellulosic Biomass, Laboratory Analytical Procedure, NREL/TP-510-42629. Sluiter, A., Hames, B., Ruiz, R., Scarlata, C., Sluiter, J., Templeton, D., Crocker, D. 2011. Determination of Structural Carbohydrates and Lignin in Biomass Laboratory Analytical Procedure, NREL/TP-510-42618. Zhang, Y.-H.P., Ding, S.-Y., Mielenz, J.R., Cui, J.-B., Elander, R.T., Laser, M., Himmel, M.E., McMillan, J.R., Lynd, L.R., 2007. Fractionating recalcitrant lignocellulose at modest reaction conditions. Biotechnol. Bioeng. 97 (2), 214–223.


Coordinated development of leading biomass pretreatment technologies Charles E. Wyman

a,*

, Bruce E. Dale b, Richard T. Elander c, Mark Holtzapple d, Michael R. Ladisch e, Y.Y. Lee f

a Dartmouth College, Hanover, NH 03755, United States Michigan State University, East Lansing, MI 48824, United States National Renewable Energy, Laboratory, Golden, CO 80401, United States d Texas A&M University, College Station, TX 77843, United States e Purdue University, West Lafayette, IN 47907, United States f Auburn College, Auburn, AL 36849, United States b

c

Available online 26 February 2005

Abstract For the first time, a single source of cellulosic biomass was pretreated by leading technologies using identical analytical methods to provide comparative performance data. In particular, ammonia explosion, aqueous ammonia recycle, controlled pH, dilute acid, flowthrough, and lime approaches were applied to prepare corn stover for subsequent biological conversion to sugars through a Biomass Refining Consortium for Applied Fundamentals and Innovation (CAFI) among Auburn University, Dartmouth College, Michigan State University, the National Renewable Energy Laboratory, Purdue University, and Texas A&M University. An Agricultural and Industrial Advisory Board provided guidance to the project. Pretreatment conditions were selected based on the extensive experience of the team with each of the technologies, and the resulting fluid and solid streams were characterized using standard methods. The data were used to close material balances, and energy balances were estimated for all processes. The digestibilities of the solids by a controlled supply of cellulase enzyme and the fermentability of the liquids were also assessed and used to guide selection of optimum pretreatment conditions. Economic assessments were applied based on the performance data to estimate each pretreatment cost on a consistent basis. Through this approach, comparative data were developed on sugar recovery from hemicellulose and cellulose by the combined pretreatment and enzymatic hydrolysis operations when applied to corn stover. This paper introduces the project and summarizes the shared methods for papers reporting results of this research in this special edition of Bioresource Technology. 2005 Elsevier Ltd. All rights reserved. Keywords: Corn stover; Pretreatment; Hydrolysis; Sugars; Enzymatic digestion; Biomass

1. Introduction Biomass, whether as sugar crops, starch crops, or cellulosic materials, provides a unique resource for sustainable production of organic fuels and chemicals that are now primarily made from petroleum. Furthermore, cel*

Corresponding author. E-mail address: [email protected] (C.E. Wyman).


lulosic materials including agricultural (e.g., corn stover) and forestry (e.g., sawdust) residues and herbaceous (e.g., switchgrass) and woody (e.g., poplar trees) crops can be sufficiently abundant to provide a major resource for making commodity products. They are also inexpensive with cellulosic biomass costing about $42/dry ton competitive with petroleum at about $6/barrel based on mass or about $13/barrel based on energy content (Lynd et al., 1999).

1960

C.E. Wyman et al. / Bioresource Technology 96 (2005) 1959–1966

Petroleum is the largest single energy source in the United States (40%) and the world (35%). Although production and consumption of most US energy sources (e.g., coal, natural gas) are almost balanced, almost two thirds of petroleum consumed in the US is imported, creating balance of trade deficits and energy security concerns. Of the total US petroleum consumption, about two thirds is used to power a transportation sector that is almost totally dependent on oil (>96%). Further, transportation is the largest emitter of the greenhouse gas carbon dioxide. Although chemical production consumes much less petroleum than transportation, a large fraction of chemicals are made from petroleum, making this sector vulnerable to price hikes or disruptions. Thus, conversion of cellulosic biomass to transportation fuels and chemicals presents a powerful opportunity to improve energy security, reduce the trade deficit, dramatically reduce greenhouse gas emissions, and improve price stability (Wyman, 1999). Because cellulosics are competitive in price with oil, a key challenge to commercializing production of fuels and chemicals from cellulosic biomass is to reduce processing costs enough to achieve attractive investor returns (Lynd et al., 1999; Wyman, 1999). Biological conversion promises such low costs because it has the potential to achieve nearly theoretical yields and the modern tools of biotechnology can improve key process steps. Overall costs of ethanol production from cellulosics have been dropped enough to be competitive with corn processing to ethanol, but the risk of financing first-of-a-kind technology stalls commercial use. Further processing advances are thus needed to improve returns for first projects and eventually compete with fossilderived products without subsidies (Wyman, 1999). Cellulosic biomass must be pretreated to realize high yields vital to commercial success in biological conversion (Mosier et al., 2005). Pretreatment is among the most costly steps and has a major influence on the cost of both prior (e.g., size reduction) and subsequent (e.g., enzymatic hydrolysis and fermentation) operations (Wooley et al., 1999; Lynd et al., 1996). For example, better pretreatment can reduce use of expensive enzymes. Thus, more attention must be given to gaining insight into interactions among these operations and applying that insight to advance biomass conversion technologies that reduce costs. In addition, although several pretreatments are promising, their relative attributes differ, but comparisons have been difficult due to differences in research methodology and substrate use. Improving the understanding of differences among pretreatment technologies and the effect of each pretreatment on other operations can facilitate selection, reduce commercialization risk, and suggest opportunities for step-change improvements. On this basis, a team of researchers from Auburn University, Dartmouth College, Michigan State Univer-

sity, the National Renewable Energy Laboratory (NREL), Purdue University, and Texas A&M University undertook the first coordinated project to develop comparative information on the performance of leading pretreatment options based on the use of a single feedstock, common analytical methods, and a consistent approach to data interpretation. University participation in the project was made possible through funding by the US Department of Agriculture Initiative for Future Agriculture and Food Systems (IFAFS) Program through a competitive solicitation with participation by NREL made possible by the support of the Office of the Biomass Program of the US Department of Energy. This research team is a major part of a somewhat larger Biomass Refining Consortium for Applied Fundamentals and Innovation (CAFI) that was formed in 2000 to understand and advance biomass conversion technologies and train students in this field. Corn stover was chosen as the single biomass source because of its abundance and potential low cost. This paper gives an overview of this project, the pretreatment technologies considered, and the materials and methods used by all participants. Subsequent papers in this volume and other journals provide more in-depth information on the results for each of the pretreatment systems studied through this coordinated research effort.

2. Overview of biological processing of cellulosic biomass Cellulosic biomass contains 40–50% cellulose, a glucose polymer; 25–35% hemicellulose, a sugar heteropolymer; and 15–20% lignin, a non-fermentable phenyl-propene unit; plus lesser amounts of minerals, oils, soluble sugars, and other components (Holtzapple, 1993). Biological routes are built around using enzymes to break down cellulose (cellulase) and perhaps hemicellulose (hemicellulase) to sugars. These sugars are then fermented to ethanol or other products, which are recovered and purified to meet market requirements. A number of references including a recent review paper by this team of authors provide far more detailed information on biological conversion technologies than can be covered here (Mosier et al., 2005). The key economic driver in biomass conversion is high product yields to enhance revenues and reduce waste treatment costs. Next in importance tends to be the concentration of products formed as concentration strongly influences product recovery costs. Rate is also important, although its impact depends on the process conditions employed with rates being less important for reactions in a mild environment than when more demanding environments, such as those requiring high pressures or extensive mixing, are needed. The cost of materials of construction and chemicals can be

C.E. Wyman et al. / Bioresource Technology 96 (2005) 1959–1966

important, and unit costs must be kept minimal for lowvalue products, such as fuels. Enzymatic conversion is attractive because nearly theoretical yields of sugars are possible, a key to economic success. However, a pretreatment step is essential to effectively prepare cellulose for enzymatic hydrolysis with these high yields. An effective pretreatment disrupts cell wall physical barriers as well as cellulose crystallinity and association with lignin so that hydrolytic enzymes can access the biomass macrostructure. In addition, pretreatment affects the cost of most other operations including size reduction prior to pretreatment and enzymatic hydrolysis after pretreatment. Pretreatment can also strongly influence downstream costs by determining fermentation toxicity, enzymatic hydrolysis rates, enzyme loadings, mixing power, product concentrations, product purification, waste treatment demands, power generation, and other process variables. And of course, the pretreatment operation itself must be low in cost and avoid high consumption of expensive chemicals, high energy demands, and feedstock degradation. Despite its importance, limited resources have been focused on developing promising pretreatment technologies or comparing how they affect other process steps.

3. Pretreatment technologies evaluated Although many biological, chemical, and physical methods have been tried over the years, pretreatment advances are still needed for overall costs to become competitive with conventional commodity fuels and chemicals (Wyman, 1999). However, only hemicellulose or lignin removal by adding acids or base, respectively, has been effective at a reasonable cost (Hsu, 1996). In particular, pretreatment by dilute sulfuric acid, with pH control, by ammonia, and with lime appear among the most promising options, with favorable processing conditions summarized for each in Table 1. Dilute-acid (0.5–1.0% sulfuric) at moderate temperatures (140–190 C) effectively removes and recovers most of the hemicellulose as dissolved sugars, and glucose yields from cellulose increase with hemicellulose removal to almost 100% for complete hemicellulose

1961

hydrolysis (Knappert et al., 1981). Although little lignin is dissolved, data suggest that lignin is disrupted, increasing cellulose susceptibility to enzymes (Yang and Wyman, 2004). Pretreatment has also been performed without adding acid in what is often called autohydrolysis (Heitz et al., 1991; Saddler et al., 1993), but hemicellulose sugar yields are lower than when acid is added and hemicellulose sugars are primarily in oligomeric form (Garrote et al., 1999; Parajo et al., 2004). Use of SO2 enhances yields in a way similar to dilute sulfuric acid (Mackie et al., 1985), but many currently prefer dilute sulfuric acid because it is cheap, up to 90% hemicellulose yields are achieved, and enzymatic hydrolysis yields of glucose can be over 90% (Hsu, 1996). Nonetheless, dilute acid pretreatment results in costly materials of construction, high pressures, neutralization and conditioning of hydrolyzate prior to biological steps, slow cellulose digestion by enzymes, and nonproductive binding of enzymes to lignin (Wyman, 1999; Hsu, 1996; Ooshima et al., 1990). Co-current/ batch reactors are typical, but percolation reactors reduce times for sugars to degrade. Forcing liquid through a packed biomass bed enhances hemicellulose and lignin removal rates and gives high yields of hemicellulose and cellulose sugars even without acid addition (Bobleter, 1994; Allen et al., 1996; Liu and Wyman, 2003). However, percolation or flowthrough are challenging to implement commercially, and the high amounts of water used result in high energy requirements for pretreatment and product recovery. An alternative approach is based on maintaining the pH at about 4–7 (Weil et al., 1998). Water under pressure can penetrate the cell structure of biomass, hydrate cellulose, and remove hemicelluose. The pKa of water is affected by temperature such that the pH of pure water at 200 C is nearly 5.0. Water also has an unusually high dielectric constant that enables ionic substances to dissociate and dissolve hemicellulose, and one half to two thirds of the lignin dissolves from most cellulosic biomass when they are treated at 220 C for 2 min. In addition, hot water cleaves hemiacetal linkages and liberates acids that catalyze breakage of ether linkages in biomass (Antal, 1996). The preferred temperature for this approach has been shown to be between 180 and 190 C for corn stover

Table 1 Technologies and representative reaction conditions for biomass preparation by pretreatment considered in this project Pretreatment technology

Chemicals used

Temperature, C

Pressure, atm absolute

Reaction times, min

Concentrations of solids, wt.%

Dilute sulfuric acid—cocurrent Flowthrough pretreatment pH controlled water pretreatment AFEX/FIBEX ARP Lime Lime + air

0.5–3.0% sulfuric acid 0.0–0.1% sulfuric acid water or stillage 100% (1:1) anhydrous ammonia 10–15 wt.% ammonia 0.05–0.15 g Ca(OH)2/g biomass 0.05–0.15 g Ca(OH)2/g biomass

130–200 190–200 160–190 70–90 150–170 70–130 25–60

3–15 20–24 6–14 15–20 9–17 1–6 1

2–30 12–24 10–30 5 MPa are commonly used (Sánchez and Cardona, 2008). It has potential to release high fraction of hemicellulosic sugars mostly in the form of oligomers contributing to reduction of undesired degrading products (Hendriks and Zeeman, 2009; Mosier et al., 2005a,b). Temperature and time showed the most significant effect on the recovery of hemicellulosic sugars and the yield of subsequent enzymatic hydrolysis of pretreated wheat straw (Perez et al., 2007). Perez et al. (2008) reported maximum recovery of hemicellulose-derived sugars (HDS) and the highest yield of enzymatic hydrolysis was achieved under various conditions of time and temperature. Simultaneous optimization of both response variables occurred at 188 °C and 40 min suggesting a two-step pretreatment would be the most adequate process configuration. A pilot scale (up to 100 kg h 1) plant, using two-step hydrothermal pretreatment in continuous operation has been developed for wheat straw (Petersen et al., 2009). The first step is soaking of straw at 80 °C for 5–10 min followed by the second stage treatment at higher temperature. In the second stage, pretreatment at 195 °C and 6–12 min resulted in 70% and 93–94% recovery of hemicellulose and cellulose, respectively. Steam explosion (autohydrolysis) is one of the most cost-effective and widely used pretreatment methods for wheat straw (Alfani et al., 2000; Ballesteros et al., 2006). In this method, size-reduced biomass is rapidly heated by high pressure steam for a period of time and then the pressure is suddenly reduced which makes the materials undergo an explosive decompression. Temperatures in the range of 160–230 °C for a time period of several seconds to a few minutes are generally applied in SE of wheat straw. The efficiency of SE is affected by several factors including temperature, residence time, particle size and moisture content. Addition of chemicals such as H2SO4, or SO2 in SE can improve the rate and extent of hemicellulose removal and lead to enhanced yield of enzymatic hydrolysis at lower temperatures (Jurado et al.,

2009). Beltrame et al. (1992) studied the effect of SE on the fractionation of wheat straw. The maximum delignification occurred with pretreatment at 210 °C and 1–2 min whereas the maximum solubilization of cellulose-rich solid fraction (83.7%) and the highest glucose production (93.5%) during enzymatic hydrolysis attained at 230 °C and 1 min. These values could be compared with maximum glucose production of untreated sample which was only 11.8%, indicating a great improvement of wheat straw digestibility after SE. Ballesteros et al. (2006) investigated the effect of SE on the diluted acid (0.9%) or water-impregnated wheat straw in various temperatures and residence times. The best results obtained from SE of acid-impregnated wheat straw at 180 °C and 10 min where almost complete dissolution of hemicellulose components and the highest yields of glucose and ethanol based on the initial raw material were achieved. Although, SE at higher temperatures increased enzymatic hydrolysis yield of cellulose-rich solid fraction, however the overall yield of glucose was lower due to partial loss of cellulose. In a similar study by Linde et al. (2008) less severe condition in impregnation step was investigated. Interestingly, the highest overall yield of the sum of glucose and xylose was obtained with treatment at 190 °C and 10 min for the wheat straw. It is uncommon for other lignocellulosic materials to reach the maximum yield of hemicellulosic and cellulosic sugars at the same pretreatment conditions. Ammonia fiber explosion (AFEX) is an alkaline thermal pretreatment during which the lignocellulosic materials are exposed to liquid ammonia at high temperature and pressure for a period of time followed by a rapid pressure release. Herbaceous and agricultural residues are well suited for AFEX pretreatment. This method does not produce inhibitors for the downstream processes and small particle size is not required for efficacy (Mosier et al., 2005a,b; Sun and Cheng, 2002). The major parameters influencing AFEX process are ammonia loading, temperature, blowdown pressure, moisture content of biomass, and residence time (Holtzapple

F. Talebnia et al. / Bioresource Technology 101 (2010) 4744–4753

et al., 1991; Teymouri et al., 2004). This pretreatment has drawbacks of being less efficient for biomass containing higher lignin content (e.g. softwood newspaper) as well as solubilization of very small fraction of solid material particularly hemicellulose (Belkacemi et al., 1998; Sun and Cheng, 2002). Mes-Hartree et al. (1988) made a comparison between steam and ammonia pretreatment of wheat straw. The authors reported that enzymatic hydrolysis was improved by several folds and more or less in the same order of magnitude for both pretreatments. However, the highest glucose concentration (0.38 g/g dry mass) was achieved with ammonia treatment. The AFEX pretreatment of wheat straw unlike other lignocellulosic materials has been rarely reported. While near theoretical sugar yields after AFEX treatment under optimum conditions have been reported for various agricultural residues (Alizadeh et al., 2005; Teymouri et al., 2004), more experimental works are necessary to address the feasibility and efficiency of this pretreatment method for wheat straw. 4.3. Chemical pretreatment Chemical pretreatment for wheat straw employ different chemicals such as acids, alkalis, and oxidizing agents e.g. peroxide and ozone (Fig. 1). Among these methods, dilute acid pretreatment using H2SO4 is the most-widely used method. Depending on the type of chemical used, pretreatment could have different effects on lignocellulose structural components. Alkaline pretreatment, ozonolysis, peroxide and wet oxidation pretreatments are more effective in lignin removal whereas dilute acid pretreatment is more efficient in hemicellulose solubilization (Galbe and Zacchi, 2002; Sánchez and Cardona, 2008; Tomas-Pejo et al., 2008). 4.3.1. Acid hydrolysis Inorganic acids such as H2SO4 have been used for pretreatment of wheat straw to improve downstream enzymatic hydrolysis. Based on the dose of acid used in the process, it could be identified as concentrated- and/or dilute-acid hydrolysis (Fig. 1). In the first case, the biomass is treated with high concentration of acids at ambient temperatures, which results in high yield of sugars. Concentrated acid treatment offers advantage of not using any enzymes for saccharification, however, this process has drawbacks including high acid and energy consumption, equipment corrosion and longer reaction time as well as obligation for acid recovery after treatment that largely limit its application (Galbe and Zacchi, 2002; Sun and Cheng, 2002). In the second approach, low-concentration acids e.g. 0.5–1% H2SO4 and high temperatures are exploited. High temperature is favorable to attain acceptable rates of cellulose conversion to glucose. Despite low acid concentration and short reaction time, application of high temperatures in dilute-acid hydrolysis accelerates the rate of hemicellulose sugar decomposition and increases equipment corrosion (Galbe and Zacchi, 2002; Taherzadeh and Karimi, 2007). The main drawbacks of this method are formation of many inhibiting by-products and pH neutralization requirement for downstream processes (Sun and Cheng, 2002; Talebnia, 2008). In order to decrease sugars degradation, a two-stage process has been developed where hemicellulose sugars are released in the first stage under milder conditions followed by the second-stage hydrolysis of cellulose-rich solid residue performed under harsher conditions. Depending on the nature of lignocellulosic feedstock, a range of temperatures between 140 and 190 °C in the first stage and 190–230 °C in the second stage is commonly used (Galbe and Zacchi, 2002; Saha et al., 2005). Delgenes et al. (1990) treated wheat straw with 72% (w/v) H2SO4 for 30 min at 30 °C and obtained 11.1 g monomeric sugars in total from 18.8 g dry raw material accounting for 59% of maximal theoretical value. In a more comprehensive work, the effects of both concentrated- and dilute-sulfuric acid pretreatment on

4747

wheat straw was evaluated by Saha et al. (2005). Concentrate and dilute acid treatments yielded 49% and 63% of the total sugars content. Optimum acid dose for maximum yield of carbohydrates in dilute acid treatment was 0.75% (v/v). The effect of temperature was also studied and formation of furfural was reported only at the highest temperature (180 °C). Two-stage dilute acid was also tried in order to improve the yield of total sugars and to avoid formation of furfural; however, the result was not satisfactory. An investigation on dilute mineral and organic acid pretreatment of wheat straw was made by Kootstra et al. (2009) where the efficiency of sulfuric acid was compared with fumaric and maleic acids. Pretreatments were performed at 130, 150 and 170 °C for 30 min. At the highest temperature, the yield of glucose after enzymatic hydrolysis reached to 98% and 96% for sulfuric acid and maleic acid, respectively. Fumaric acid was less effective than maleic acid. The highest yield of xylose (ca. 80%) was obtained with H2SO4 pretreatment at 150 °C, however, the yield decreased at 170 °C due to partial degradation of xylose to furfural. The maximum xylose yield at 170 °C was obtained with treatment using maleic acid. The authors concluded that maleic acid pretreatment of wheat straw is almost as effective as sulfuric acid in respect with enzymatic digestibility where more xylose and less furfural is produced. 4.3.2. Alkaline Alkaline process is based on utilization of dilute bases in pretreatment of lignocellulosic feedstocks. Sodium, potassium, calcium and ammonium hydroxides are suitable alkaline agents for pretreatment, among which sodium hydroxide has been studied the most (Kumar et al., 2009a,b). Alkaline pretreatment processes utilize lower temperatures and pressures than other pretreatment technologies (Mosier et al., 2005a,b). The effectiveness of this method depends on the lignin content of biomass and therefore, it is well suited for agricultural residues such as wheat straw (Sánchez and Cardona, 2008; Sun and Cheng, 2002). Alkaline pretreatment can largely improve the cellulose digestibility and sugars degradation is less than acid treatment, however, the application is hindered by high cost of alkalis. Utilization of calcium hydroxide (lime) as a cheap alkaline agent and its recovery and regeneration or ammonia which is recyclable due to its volatility could be a solution to this problem (Kim and Holtzapple, 2006; Wyman et al., 2005). Lime pretreatment of wheat straw was explored in a work presented by Chang et al. (1998). They found that for short pretreatment times (1–3 h), high temperatures (85–135 °C) were required to reach high sugar yields, whereas for long treatment times (e.g. 24 h), lower temperatures (50–65 °C) were more effective. The optimal lime loading was 0.1 (g/g dry mass). Under all recommended conditions, the yield of reducing sugars was increased by a factor of 10 compared with untreated wheat straw. 4.3.3. Oxidizing agents 4.3.3.1. Alkaline/oxidative pretreatment. In this pretreatment, an oxidizing compound such as hydrogen peroxide (H2O2) or peracetic acid (C2H4O3) is used in combination with an alkaline (e.g. NaOH) and it is usually carried out under mild temperature. This treatment is more effective in improving of crop residue digestibility compared with NaOH treatment alone. Gould (1984) demonstrated the use of H2O2 for delignification of agricultural residues including wheat straw. Treatment was performed with 1% H2O2 and pH 11.5 at 25 °C for 18–24 h. Under this condition, more than half of the lignin and most of hemicellulose were solubilized. These values were higher than those of NaOH treatment (without H2O2 addition) by several folds. Enzymatic hydrolysis of treated wheat straw in presence of H2O2 also showed a clear enhancement where nearly 100% conversion was attained at pH 11.5. Treatment with NaOH alone under identical condition exhibited a significant increase in cellulose digestibility only at pH > 12, however, the max-

4748

F. Talebnia et al. / Bioresource Technology 101 (2010) 4744–4753

imum efficiency did not exceed 65%. Alkaline peroxide delignification of agricultural residues showed to be strongly pH-dependent, with an optimum pH of 11.5–11.6. The author also reported that at pH lower than 10, delignification was negligible and at a pH 12.5 and higher, alkaline peroxide treatment showed no real effect on the enzymatic digestibility. The variables of alkaline peroxide pretreatment of wheat straw was optimized by Saha and Cotta (2006). Treatment with NaOH (0% peroxide) yielded about 250 (mg/g) total sugars. This yield was almost doubled with addition of peroxide. The optimum dose of peroxide was 2.15% (v/v) and sugar recovery after enzymatic hydrolysis was slightly higher at 35 °C compared to 25 °C. Increase of treatment time from 3 to 24 h showed a minor effect on enhancement of total sugars yield. The results are in agreement with previous reports suggesting a minimum peroxide/biomass weight ratio of 0.25 is necessary for a good delignification (Gould, 1984). 4.3.4. Wet oxidation (WO) In wet oxidation, the lignocellulosic biomass is treated with water and high pressure oxygen or air at elevated temperatures (above 120 °C). Typical oxygen pressure range is 120–480 Psi (Schmidt and Thomsen, 1998). WO is an effective pretreatment method for the fractionation of wheat straw into a solubilized hemicellulose fraction and a cellulose-rich solid fraction with high susceptibility to enzymatic hydrolysis. Combination of alkali and WO not only improves the rate of lignin oxidation (and in turn enzymatic hydrolysis) but also prevents formation of furfural and HMF (Bjerre et al., 1996). Acids formed during initial reaction in WO due to solubilization of hemicellulose components catalyze the subsequent hydrolytic reactions through which hemicelluloses are broken down into lower molecular weight fragments that are soluble in water. Lignin degradation is also significant especially at the higher temperatures because phenol-like compounds and carbon–carbon linkage are very reactive under wet oxidation conditions. Lignin is decomposed to CO2, H2O and carboxylic acids (Klinke et al., 2002). Bjerre et al. (1996) investigated the alkaline wet oxidation of wheat straw at 10 bar of oxygen pressure and various temperatures. A maximum delignification of about 65% and hemicellulose solubilization of about 50% occurred at 170 °C within 10 min. The left over solid residue showed the highest enzymatic convertibility where 85 (%w/w) of cellulose was converted to glucose. The effects of oxygen pressure and alkaline (Na2CO3) addition on WO of wheat straw were studied and the results indicated slight effects of these two parameters on solubilization of hemicellulose. Almost all hemicellulose (96%) was solubilized by pretreatment with oxygen pressure and without carbonate. However, alkaline addition, regardless to presence or absence of oxygen, decreased formation of furfural by more than 10-fold. The lignin removal, on the other hand, was significantly affected by presence of oxygen (60% compared to 11% without oxygen pressure) (Ahring et al., 1996). Schmidt and Thomsen (1998) optimized the WO of wheat straw with respect to oxygen pressure, reaction temperature and time as the main process variables. The optimum condition was 185 °C, 15 min and 12 bar oxygen pressure. In overall, temperature was found to be more important process parameter than time and oxygen pressure. Formation of some inhibitors in WO of wheat straw has been reported including aliphatic carboxylic acids (mainly formic acid and acetic acid), phenols and 2-furoic acid (Klinke et al., 2001). 4.3.5. Ozonolysis In this process, ozone is used to solubilize lignin and a small fraction of hemicellulose from wheat straw. Ozonolysis is carried out at room temperature and can effectively remove the lignin without producing any toxic residues. The main drawback of this

process is a large amount of ozone utilization that makes the process expensive (Sun and Cheng, 2002). Binder et al. (1980) investigated the delignification of wheat straw by ozone treatment and biodegradability of resultant solid residue. It was demonstrated that a 50% reduction of the original lignin content is optimal for enzymatic hydrolysis. After treatment, 75% of the cellulose was digested within 24 h as compared to 20% in untreated straw. The improved digestibility was attributed to lignin removal as well as reduced degree of polymerization of the treated cellulose. The ozonolysis pretreatment of wheat and rye straw was investigated in a fixed bed reactor under room conditions (Garcia-Cubero et al., 2009). Among the studied variables, moisture content found to be the most significant variable and a reaction controlling parameter for values below 30%. Enzymatic hydrolysis yield of up to 88.6% was attained compared to 29% in non-ozonated wheat straw. 4.4. Biological pretreatment Biological pretreatment comprises of using microorganisms such as brown-, white-, and soft-rot fungi for selective degradation of lignin and hemicellulose among which white-rot fungi seems to be the most effective microorganism. Lignin degradation occurs through the action of lignin-degrading enzymes such as peroxidases and laccase (Fig. 1) (Okano et al., 2005). The suitable fungi for biological pretreatment should have higher affinity for lignin and degrade it faster than carbohydrate components. Biological pretreatments are safe, environmentally friendly and less energy intensive compared to other pretreatment methods. However, the rate of hydrolysis reaction is very low and needs a great improvement to be commercially applicable. Hatakka (1983) investigated the pretreatment of wheat straw using 19 white-rot fungi and found that 35% of the wheat straw was converted to reducing sugars after five weeks pretreatment with Pleurotus ostreatus compared to only 12% conversion of the untreated straw. Five different fungi obtained from screening were evaluated for pretreatment of wheat straw in a study performed by Patel et al. (2007). Pretreatment with Aspergillus niger and Aspergillus awamori showed the best results regarding to yields of total sugars and ethanol after fermentation. 4.5. Summary of pretreatment methods Pretreatment plays a significant role in ethanol production from lignocellulosic materials such as wheat straw. The objectives are to increase the surface area and porosity of the substrate, reduce the crystallinity of cellulose and disrupt the heterogeneous structure of cellulosic materials. Thus far, no single pretreatment method was found to meet all these requirements; instead a combination of different methods might be applied. Table 1 describes some the most promising methods and corresponding conditions, recently used for pretreatment of wheat straw. Diverse advantages and drawbacks are associated with each pretreatment method. The shorter reaction times (desired) are generally accompanied with higher temperature (undesired) (Table 1). The choice of appropriate pretreatment method for wheat straw relies on some technological factors including energy balance, higher solid loading, and minimum use of chemicals as well as some environmental factors such as wastewater treatment, catalyst recovery and solvent recycling. Steam explosion is probably the most suitable method for pretreatment of wheat straw in terms of lower reaction time, higher solid loading and minimum use of chemicals (Table 1). However, since the pretreatment has an enormous influence on the efficiency and economy of the subsequent stages, the final decision must me made in the framework of the entire process.

4749

F. Talebnia et al. / Bioresource Technology 101 (2010) 4744–4753 Table 1 Promising methods and corresponding conditions for pretreatment of wheat straw. Pretreatment technology

Procedure/chemicals

Temp. (°C)

Reaction times

Solid loading (wt.%)

References

Dilute acid Steam explosion Alkaline peroxide Wet oxidation (alkaline)

0.5–5.0% H2SO4 Saturated steam >0.25 g H2O2/g biomass, pH 11.5 6–12 bar oxygen pressure (+0.11 g Na2CO3/g biomass) 0.05–0.15 g Ca(OH)2/g biomass

120–180 160–230 25–35 185–195

5–60 min 5–30 min 3–24 h 10–15 min

5–30