Available online at www.sciencedirect.

com

Bioresource Technology 99 (2008) 65056511

Bioethanol production from non-starch carbohydrate residues in process streams from a dry-mill ethanol plant
M. Linde *, M. Galbe, G. Zacchi
Department of Chemical Engineering, Lund University, P.O. Box 124, SE-221 00 Lund, Sweden Received 6 March 2007; received in revised form 13 November 2007; accepted 13 November 2007 Available online 20 February 2008

Abstract Slurries obtained from process streams in a starch-to-ethanol plant, Agroetanol AB in Norrkoping, Sweden, were used to assess the potential increase in bioethanol yield if heat treatment followed by enzymatic hydrolysis were applied to the residual starch-free cellulose and hemicellulose fractions. The eects of dierent pretreatment conditions on our (the raw material), the stream after saccharication of starch, before fermentation, and after fermentation were studied. The conditions resulting in the highest concentration of glucose and xylose in all streams were heat treatment at 130 C for 40 min with 1% H2SO4. Mass-balance calculations over the fermentation showed that approximately 64%, 54%, 75% and 67% of the glucan, xylan, galactan and arabinan, respectively, in the our remained water insoluble in the process stream after fermentation without any additional treatment. Utilizing only the starch in the our would theoretically yield 425 L ethanol per ton our. By applying heat pretreatment to the water-insoluble material prior to enzymatic hydrolysis, the ethanol yield could be increased by 59 L per ton our, i.e. a 14% increase compared with starch-only utilization, assuming fermentation of the additional pentose and hexose sugars liberated. 2007 Elsevier Ltd. All rights reserved.
Keywords: Starch; Cellulose; Pretreatment; Enzymatic hydrolysis; Bioethanol

1. Introduction Interest in bioethanol is growing rapidly worldwide due to the expected shortage of fossil fuels. In 2006 approximately 16.3 billion litres (4.3 billion gallons) of ethanol were produced in the USA, which is twice the quantity produced in 1999, and an additional capacity of 4.9 billion litres (1.3 billion gallons) of ethanol is planned (Wright et al., 2006). The USA is the largest ethanol producer in the world, closely followed by Brazil. Together they contributes approximately 70% to the worlds total production. Europe, contributes only a few percent to the total ethanol produced in the world today, but has the goal of replacing 5.75%, calculated on energy basis, of all petrol

Corresponding author. Tel.: +46 46 222 82 97; fax: +46 46 222 45 26. E-mail address: marie.linde@chemeng.lth.se (M. Linde).

and diesel for transport purposes with renewable fuels by 2010 (Directive 2003/30/EC,). Brazil utilizes cane sugar for bioethanol production while the USA and Europe mainly use starch from corn, and from wheat and barley, respectively. Starch is converted into ethanol in dry-mill starch plants where the raw material, e.g. grain from wheat, barley, or corn, is milled and then treated with a combination of heat and enzymes without prior separation of its constituents (Bothast and Schlicher, 2005). The starch is hydrolysed into monomeric glucose, which is fermented, while the protein-rich residues, the distillers dried grains with solubles (DDGS), are marketed as animal fodder (Bothast and Schlicher, 2005; Gulati et al., 1996). To meet the increasing demand for ethanol it is essential not only to increase the number of ethanol plants but also to utilize the raw material more eciently. The technology used today only converts the starch in the raw material and

0960-8524/$ - see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2007.11.032

6506

M. Linde et al. / Bioresource Technology 99 (2008) 65056511

carbohydrate-containing residues, such as stover, straw, husks and bagasse, are obtained. In a dry-mill starch plant the non-starch constituents of the grain, which contain cellulose and hemicellulose, end up in the DDGS. The building blocks of cellulose are similar to those in starch, i.e. glucose; but instead of a-1,4-linkages, as in starch, cellulose has b-1,4-linkages, generating a linear and more rigid structure than starch. In hemicellulose the backbone consists of pentose sugars, which in wheat and barley consists mainly of arabinoxylan. Thus, a-amylase and amyloglucosidase, which are the hydrolysing enzymes used for starch hydrolysis (Nigam and Singh, 1995), cannot hydrolyse cellulose or hemicellulose. Instead, cellulases and hemicellulases are required, which are currently not widely used in the dry-mill process. The complex structure of the cellulose and hemicellulose, together with the lignin inhibits the enzymes and pretreatment is often necessary to increase the rate of hydrolysis. Heat treatment prior to enzymatic hydrolysis has been shown to considerably increase the rate of hydrolysis (Galbe and Zacchi, 2002; Saddler et al., 1993). Since cellulose is a polymer of glucose the yeast strains presently used in starch-to-ethanol plants ferment the monomeric units. The pentose sugars, on the other hand, require pentose-fermenting yeast (or some other microorganism), not currently used in industrial processes. However, decades of metabolic engineering have resulted in the development of pentose-fermenting yeasts that have been shown to work well on lab scale (Hahn-Hagerdal and Pam ment, 2004; Ohgren et al., 2006; Sonderegger et al., 2004). Thus, it is probably only a matter of time before pentosefermenting yeast will be used in industry. In this study, we investigated whether the overall ethanol yield in a wheat and barley based dry-mill ethanol plant could be increased by pretreatment and enzymatic hydrolysis of the residue, which was withdrawn at various stages of the process. The process slurries investigated were the our, the slurry after saccharication of the starch, before fermentation, and after fermentation. Three microwave pretreatment conditions were evaluated. 2. Methods 2.1. Process samples

Grain
15 ton/h

F
Enzymes H2 O

7 ton/h

Liquefaction Saccharification
AS BF
CO2

Backset

Fermentation

AF
16 ton/h

Ethanol

Distillation

52 ton/h

DDGS

Separation
Fig. 1. Simplied process scheme of the dry-mill ethanol plant Agroetanol AB in Norrkoping, Sweden. Approximate ow rates were given by the plant.

lysed regarding their starch content. The starch-free WIS samples were heat treated using a method developed by Palmarola-Adrados et al. (2004, 2005), which includes microwave pretreatment and enzymatic hydrolysis, to investigate the content of anhydrous glucose and anhydrous xylose in the WIS available for ethanol production. 2.2. Starch removal

Process samples were obtained from Agroetanol AB in Norrkoping, Sweden, and Fig. 1 shows a simplied process scheme. The raw material in the plant consisted of 80% wheat and 20% barley. Samples were obtained from four dierent process streams (Fig. 1): the our (F), the slurry after saccharication of the starch (AS) before fermentation (BF), and after fermentation (AF). The process samples were separated into a solid part and a liquid part by ltration. The solid part was washed with deionised water until all water-soluble solids (WS) had been removed. The remaining solids were denoted waterinsoluble solids (WIS). The WIS and the liquid were ana-

The starch content in the process samples was determined by hydrolysis of the starch in the separated WIS and liquid using Termamyl 120 L and AMG 300 L, kindly donated by Novozymes A/S (Bagsvrd, Denmark). The starch hydrolysis of the solids was performed at a WIS-towater ratio of 1:5. Starch hydrolysis of the liquid fraction was performed at a liquid concentration corresponding to a WIS-to-water ratio of 1:5, although the insoluble solids had been removed. The starch removal procedure was performed in two steps. First, the starch was liqueed by thermo-stable a-amylases (4 mL Termamyl 120 L per kg

M. Linde et al. / Bioresource Technology 99 (2008) 65056511

6507

slurry) for 4 h at 85 C and pH 6. Then saccharication of the liqueed starch was performed using amyloglucosidase (12 mL AMG 300 L per kg slurry) at 55 C and pH 5 for 48 h (Palmarola-Adrados, 2005). The liquefaction and saccharication of the starch in the WIS was performed in a 1-l Buchi Rotavapor (RE121) with a Buchi 461 water bath (Buchi Labortoriums-technik AG, Flawil, Switzerland). The starch present in the liquid was removed in Erlenmeyer asks on a temperature-controlled shaker (Lab-Therm, Kuhner, Switzerland). The dierence in glucose content in the liquid before and after starch hydrolysis was considered to originate from starch. 2.3. Microwave pretreatment The starch-free WIS was pretreated with dilute sulphuric acid at high temperatures to increase the enzymatic digestibility of the WIS, thus increasing the yield of sugars. Pretreatment of the starch-free WIS was performed in a Mega Microwave workstation (MLS-1200, Milestone, Sorisole, Italy), as described elsewhere (Palmarola-Adrados, 2005; Palmarola-Adrados et al., 2005). Using a microwave oven decreases the amount of sample required and the vessel also provides a closed system with a constant amount of material throughout the process. The pretreatment vessel, which had a total volume of 350 mL and was made of Teflon to withstand high temperatures and pressures, was loaded with starch-free WIS, deionised water and sulphuric acid to at total weight of 75100 g of slurry. The WIS loading in the pretreatment vessel was 5% (w/w). Three pretreatment conditions, all reported to give high yields of either glucose and/or xylose from wheat bran or the starch-free residues in a waste stream from a combined starch and ethanol plant, were studied. Pretreatment at 170 C with no sulphuric acid has been reported to give high yields of glucose and xylose when performed for 20 and 30 min, respectively (Palmarola-Adrados et al., 2005). However, high yields of both glucose and xylose have also been reported at 170 C and 40 min (Palmarola-Adrados et al., 2004). Pretreatment condition A, 170 C for 30 min, was thus chosen. Pretreatment temperatures below 170 C have been shown to yield low concentrations of HMF and furfural, and increased enzymatic digestibility of glucan (Palmarola-Adrados et al., 2005), thus pretreatment conditions B, 160 C for 40 min with 0.1% H2SO4, was chosen as an alternative to A. Pretreatment condition C, 130 C for 40 min with 1.0% H2SO4, was chosen as it has been reported to result in high pentose yields (Palmarola-Adrados et al., 2005). 2.4. Enzymatic hydrolysis Enzymatic hydrolysis of hemicellulose and cellulose was performed on pretreated, undiluted samples. The enzymes added were Ultrao L (45 FBG/g (Fungal b-glucanase units)) and Celluclast 1.5 L (65 FPU/g and 17b-glucosidase IU/g), both kindly provided by Novozymes A/S. The

enzyme loading was 1 g of each enzyme mixture per 100 g of slurry. Hydrolysis was performed in a laboratory rotary-shaker-incubator (LSR/L-V Adolf Kuhner AG, Schweiz) at 200 rpm and 50 C for 72 h. The pH was adjusted to 5 using sodium hydroxide prior to hydrolysis. Enzymatic hydrolysis was also performed on non-pretreated starch-free WIS with 5% WIS to assess the eect of the pretreatment. The content of oligomeric and monomeric glucose, xylose and arabinose in the starch-free liquid part of the process samples was determined using enzymatic hydrolysis. Enzymatic hydrolysis was performed in the same way as for the WIS and the enzyme loading was based on WIS equivalents for the liquid, i.e. as if the whole stream had been added. The oligomeric content was calculated as the dierence between monomeric sugars before and after enzymatic hydrolysis. 2.5. Analysis Dry matter contents were determined by drying samples in an oven at 105 C until constant weight was obtained. The liquids from the process streams, starch hydrolysis, pretreatment and enzymatic hydrolysis were analysed regarding their content of monomeric sugars using HPLC (Shimadzu, Kyoto, Japan) with a refractive-index detector (Shimadzu). The column used for the separation of the sugars glucose, xylose, galactose and arabinose was an Aminex HPX-87P (Bio-Rad, Hercules, CA, USA) at 85 C, with water as eluent, at a ow rate of 0.5 mL/min. Liquid from the process streams and the samples from enzymatic hydrolysis were also analysed regarding their contents of ethanol, lactic acid, acetic acid, HMF and furfural, using an Aminex HPX-87 H column (Bio-Rad) at 65 C, with 5 mM H2SO4 as eluent, at a ow rate of 0.5 mL/min. All samples were ltered through a 0.2 lm lter before analysis to remove particles. 2.6. Mass balance calculation During the starch-to-ethanol process in the plant some of the cellulose and hemicellulose is already converted to soluble sugars. To be able to evaluate the eects of pretreatment and enzymatic hydrolysis on the remaining WIS it is necessary to estimate the extent to which the cellulose and hemicellulose were already degraded. In dry-mill processes some of the liquid in the stillage, containing WS, is recycled as backset (Bothast and Schlicher, 2005), which complicates the mass balance calculations. In the process investigated in this study, backset is added before starch hydrolysis, mostly between starch hydrolysis and fermentation, see Fig. 1. However, the only change in ow rate of the main stream in the fermentation vessel is the release of carbon dioxide. The release of carbon dioxide and the total ow rate of the process stream out of the fermentor were calculated by assuming that the change in mass of total glucose during fermentation was due to ethanol

6508

M. Linde et al. / Bioresource Technology 99 (2008) 65056511

production. The dierence in the amount of sugars in WIS before and after fermentation was due to degradation of the cellulose and hemicellulose. The degree of hydrolysis in the process, however, is only calculated over the fermentation step, while it is probable that some hydrolysis of the cellulose and hemicellulose has also occurred parallel to starch hydrolysis. However, as the cellulases used in the process are added just before fermentation, hydrolysis is assumed mainly to take place during fermentation. Composition analysis of the process samples was performed with a commonly used, standard procedure from NREL (Sluiter et al., 2004). However, the standard procedure resulted in lower concentrations of sugar constituents in the WIS than those obtained with the microwave pretreatment (data not shown). This has also been observed in other studies on wheat bran (Palmarola-Adrados et al., 2004). Therefore, the concentrations of sugar constituents in the WIS obtained through microwave pretreatment and enzymatic hydrolysis were used in the mass balance calculations. 3. Results and discussion Table 1 gives the concentration of WIS in the process samples and the starch content in the WIS and in the liquid of the various process samples. As it is not possible to determine whether the monomeric glucose in the streams originates from starch or cellulose, all monomeric glucose present in WS is shown as starch. Flow rates obtained from Agroetanol AB are presented in Fig. 1. Figs. 24 show the glucose, xylose and arabinose yields, respectively, after pretreatment A, B and C, and enzymatic hydrolysis of the starch-free WIS without pretreatment. The yields obtained from enzymatic hydrolysis without pretreatment show, as expected, that pretreatment of the starch-free WIS increased the total yield signicantly. The highest yields of glucose from cellulose in the our (F) and the stream after saccharication of the starch (AS) were 3.1 and 1.0 g/100 g process stream, respectively, and were obtained with pretreatment B: 0.1% H2SO4, 160 C for 40 min. The highest yields of glucose in the process stream before fermentation (BF) and after fermentation (AF) were 1.2 and 0.9 g/100 g process stream, respectively, and they were obtained with pretreatment conditions A and C, respectively. However, the dierence in glucose
Table 1 WIS content and concentration of starch in WIS and in liquid expressed as g/100 g of process stream Stream WIS Starcha WIS F AS BF AF
a

Fig. 2. Overall yield of glucose after pretreatment (PT) and enzymatic hydrolysis (EH) of the starch-free WIS. F = our, AS = after saccharication, BF = before fermentation and AF = after fermentation.

Fig. 3. Overall yield of xylose after pretreatment (PT) and enzymatic hydrolysis (EH) of the starch-free WIS. F = our, AS = after saccharication, BF = before fermentation and AF = after fermentation.

Liquid 0.3 19.0 15.7 0.5

Total 58.9 20.6 17.9 1.0

79.2 8.7 11.1 6.3

58.6 1.6 2.2 0.5

Calculated as anhydrous glucose.

yield between the three pretreatment conditions was small, except for the our (F), for which pretreatment A was insucient, (Fig. 2). The glucose yield from our increased signicantly when a small amount of acid was added to the pretreatment. The highest yields of the hemicellulose constituents xylose and arabinose, for all process streams, were obtained with the pretreatment conditions that had previously been reported to give high pentose yields, i.e. pretreatment C: 1% H2SO4 at 130 C for 40 min. Nearly all the xylan was hydrolysed under these pretreatment conditions, thus enzymatic hydrolysis was in fact not necessarily required to release xylose. However, the yield of arabinose increased signicantly with enzymatic hydrolysis in streams

M. Linde et al. / Bioresource Technology 99 (2008) 65056511

6509

Fig. 4. Overall yield of arabinose after pretreatment (PT) and enzymatic hydrolysis (EH) of the starch-free WIS. F = our, AS = after saccharication, BF = before fermentation and AF = after fermentation.

BF and AF. The highest yield of galactose was also obtained with pretreatment C and was low (0.6 g/100 g process stream of F and 0.1 g/100 g of process stream AS, BF and AF). The small dierences in glucose yield between pretreatment A, B and C resulted in the highest yield of the total amount of sugars being obtained with pretreatment C for all streams. For F, AS, BF and AF the total yields were 8.1, 2.5, 3.1 and 2.1 g/100 g process stream, respectively. In all process streams pretreatment B produced the highest concentration of HMF and furfural, which are produced from the degradation of hexose and pentose sugars, respectively, during high-temperature pretreatment. The concentrations of HMF were 0.02 g/100 g process stream, or lower. Pretreatment C, which yielded the highest hydrolysis of hemicellulose, resulted in concentrations of furfural between 0.03 and 0.36 g/100 g process stream, see Fig. 5. This should be taken into consideration in future process design as HMF and furfural inuence the productivity of the yeast. The concentration of acetic acid was also measured to estimate the hydrolysis of the acetyl groups in the hemicellulose, as acetic acid inhibits yeast in the subsequent fermentation step (Palmqvist and Hahn-Hagerdal, 2000). The highest yield of acetic acid in the process streams was obtained with pretreatment C. This conrms that pretreatment with 1% H2SO4 at 130 C for 40 min hydrolysed most hemicellulose. The highest concentration of acetic acid, 0.8 g/100 g process stream, was obtained in stream F. Fig. 6 shows the concentration of xylose and arabinose in the WS fraction of the samples. Oligomeric glucose was only detected in the process stream after fermentation (0.5 g glucose/100 g process stream). It was not possible to discriminate between soluble monomeric glucose originating from starch and from cellulose, thus glucose was not included in Fig. 6.

Fig. 5. Concentration of furfural produced during pretreatment of WIS.

0.6 0.5 Concentration in the liquid (g/100 g process stream) 0.4 0.3 0.2 0.1 0 Xylose (mono) Xylose (oligo) Arabinose (mono) Arabinose (oligo)

AS

BF

AF

Fig. 6. Concentrations of monomeric and oligomeric xylose and arabinose in the liquid part of the process streams.

Table 2 Yield of non-starch polysaccharides in WIS and WS, in g/100 g of process stream Stream WIS Glucose F AS BF AF 3.1 1.0 1.2 1.4 Xylose 3.1 1.1 1.2 1.2 Arabinose 2.2 0.9 1.1 0.9 WS Xylose 0.0 0.1 0.1 0.6 Arabinose 0.1 0.3 0.3 0.4

Table 2 summarizes the highest experimental yields of glucose, xylose and arabinose obtained from starch-free WIS and WS in the process streams. The increase in concentrations of water-soluble sugars during fermentation

6510

M. Linde et al. / Bioresource Technology 99 (2008) 65056511

(Table 2) was not only the result of a decrease in total mass due to carbon dioxide emissions, but also due to the degradation of WIS. Approximately 64% of the water-insoluble glucan from cellulose entering the fermentation vessel was found, through calculation, to remain in the solid fraction after fermentation. Of the hemicellulose constituents xylan and arabinan approximately 54% and 67%, respectively, were found to remain as water insoluble. Thus, partial degradation of cellulose and hemicellulose had occurred in the plant. If all the starch in the our was hydrolysed to glucose and fermented, the theoretical ethanol yield would be 425 L/ton, assuming an ethanol density of 0.785 g/L at 25 C (Aylward and Findlay, 1994), and that 1 g glucose yields 0.51 g ethanol. The calculated degree of hydrolysis of the cellulose and hemicellulose in the fermentor was applied to the our, making it possible to estimate an increase in ethanol yield based on the composition of the our. Adding cellulases and hemicellulases without pretreatment would hydrolyse oligomeric soluble cellulose and hemicellulose present in the streams, which could theoretically increase the ethanol yield by approximately 7 L per ton our compared to when only starch was fermented. Yeast that could ferment all monomeric sugars would theoretical increase the ethanol yield by approximately 22 L per ton our without pretreatment. The increase in ethanol yield from the remaining WIS if pretreatment and enzymatic hydrolysis were applied was 37 L per ton our, of which the cellulose contributes 13 L per ton our. Thus, if pretreatment is performed a total increase, with and without pentose-fermenting yeast, of 20 L and 59 L ethanol per ton our (i.e. 5% and 14%), respectively, is possible. 4. Conclusions The raw material in the dry-mill ethanol plant is grain kernels from both wheat and barley, and large quantities of cellulose and hemicellulose pass through the process to become constituents of the DDGS. It is of interest to utilize the cellulose and hemicellulose to increase the total yield of bioethanol from cereal grains. However, the backset makes it dicult to calculate how much cellulose and hemicellulose were degraded without any additional treatment. By performing a mass balance over the fermentation step, which has no addition of backset or other streams, it was possible to estimate the hydrolysis of the cellulose and hemicellulose. Approximately 64% of the water-insoluble glucan was left after fermentation. Of the hemicellulose constituents xylan and arabinan, approximately 54% and 67%, respectively, remained as water insoluble. Using the calculated degree of hydrolysis, it was found that by adding enzymes that hydrolyse the cellulose and hemicellulose, and utilizing pentose-fermenting yeast, it would be possible to theoretically increase the ethanol yield by 5% (22 L per ton our) without any pretreatment. However, pretreating the water-insoluble solids, which contain cellulose and hemicellulose, prior to enzymatic hydrolysis,

could potentially increase the ethanol yield by 14% in a starch-to-ethanol plant compared to the case when only the starch was hydrolysed and fermented. To hydrolyse the cellulose and hemicellulose in the process streams, pretreatment was necessary. Further investigations are needed to determine where in the process an additional pretreatment step would be optimal. This study clearly showed that with heat pretreatment at 130 C for 40 min with 1% H2SO4 and subsequent enzymatic hydrolysis with the commercial cellulases Celluclast 1.5 L and Ultrao L, it was possible to hydrolyse the WIS in the process streams to monomeric sugars. However, low acid concentrations are preferable and pretreatment at 160 C for 40 min with only 0.1% H2SO4 gave the same high glucose yields but decreased the xylose yield by 25% for F, AS and BF, and by almost 50% for AF. A decreased xylose yield will dramatically inuence the overall ethanol yield when a pentose-fermenting yeast is available but, in the meantime, pretreatment with a low acid concentration, which results in lower cost, would be preferable. Acknowledgement The Swedish Energy Agency is gratefully acknowledged for its nancial support. References
Aylward, G., Findlay, T., 1994. SI Chemical Data, 3rd ed. Peter Storer, Milton. Bothast, R.J., Schlicher, M.A., 2005. Biotechnological processes for conversion of corn into ethanol. Appl. Microbiol. Biotechnol. 67, 19 25. Directive 2003/30/EC, 2003. Directive of the European parliament and of the council of 8 May 2003 on the promotion of the use of biofuels or other renewable fuels for transport. Directive 2003/30/EC. Galbe, M., Zacchi, G., 2002. A review of the production of ethanol from softwood. Appl. Microbiol. Biotechnol. 59, 618628. Gulati, M., Kohlman, K., Ladisch, M.R., Hespell, R., Bothast, R.J., 1996. Assessment of ethanol production options for corn products. Bioresour. Technol. 58, 253264. Hahn-Hagerdal, B., Pamment, N., 2004. Microbial pentose fermentation. Appl. Biochem. Biotechnol., 12071209. Nigam, P., Singh, D., 1995. Enzymes and microbial systems involved in starch processing. Enzyme Microb. Technol. 17, 770778. Ohgren, K., Bengtsson, O., Gorwa-Grauslund, M.-F., Galbe, M., HahnHagerdal, B., Zacchi, G., 2006. Simultaneous saccharication and co fermentation of glucose and xylose in steam-pretreated corn stover at high ber content with Saccharomyces cerevisiae TMB3400. J. Biotechnol. 126, 488498. Palmarola-Adrados, B., 2005. Thesis: Hydrolysis of By-Products from Cereal Kernels for Improved Ethanol Production, Department of Chemical Engineering, Lund University, Sweden. Palmarola-Adrados, B., Choteborska, P., Galbe, M., Zacchi, G., 2005. Ethanol production from non-starch carbohydrates of wheat bran. Bioresour. Technol. 96, 843850. Palmarola-Adrados, B., Juhasz, T., Galbe, M., Zacchi, G., 2004. Hydrolysis of nonstarch carbohydrates of wheat-starch euent for ethanol production. Biotechnol. Progr. 20, 474479. Palmqvist, E., Hahn-Hagerdal, B., 2000. Fermentation of lignocellulosic hydrolysates, II: inhibitors and mechanisms of inhibition. Bioresour. Technol. 74, 2533.

M. Linde et al. / Bioresource Technology 99 (2008) 65056511 Saddler, J.N., Ramos, L.P., Breuil, C., 1993. Steam pretreatment of lignocellulosic residues. Biotech. Agric. Ser. 9, 7391. Sluiter, A., Hames, B., Ruiz, R., Scarlata, C., Sluiter, J., Templeton, D., 2004. Determination of structural carbohydrates and lignin in biomass. Sonderegger, M., Jeppsson, M., Larsson, C., Gorwa-Grauslund, M.-F., Boles, E., Olsson, L., Spencer-Martins, I., Hahn-Hagerdal, B., Sauer,

6511

U., 2004. Fermentation performance of engineered and evolved xylosfermenting Saccharomyces cerevisiae strains. Biotechnol. Bioeng. 87, 9098. Wright, L., Boundy, B., Perlack, B., Davis, S., Saulsbury, B., 2006. Biomass Energy Data Book, rst ed.