- Email: [email protected]
Optimization of uncatalyzed steam explosion pretreatment of rapeseed straw for biofuel production
Bioresource Technology 190 (2015) 97–105
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Optimization of uncatalyzed steam explosion pretreatment of rapeseed straw for biofuel production Juan C. López-Linares a, Ignacio Ballesteros b, Josefina Tourán a, Cristóbal Cara a, Eulogio Castro a, Mercedes Ballesteros b, Inmaculada Romero a,⇑ a b
Department of Chemical, Environmental and Materials Engineering, University of Jaén, Campus Las Lagunillas, 23071 Jaén, Spain DER-CIEMAT, Avda. Complutense 22, 28040 Madrid, Spain
h i g h l i g h t s Ethanol production from steam exploded rape straw at high solids loading is reported. Operation at high solids loading (20% w/v) allows complete conversion of cellulose. High ethanol concentrations (5% v/v) are obtained, allowing industrial operation.
a r t i c l e
i n f o
Article history: Received 4 March 2015 Received in revised form 17 April 2015 Accepted 18 April 2015 Available online 23 April 2015 Keywords: Rapeseed straw Steam explosion SSF Fuel ethanol High solids
a b s t r a c t Rapeseed straw constitutes an agricultural residue with great potential as feedstock for ethanol production. In this work, uncatalyzed steam explosion was carried out as a pretreatment to increase the enzymatic digestibility of rapeseed straw. Experimental statistical design and response surface methodology were used to evaluate the influence of the temperature (185–215 °C) and the process time (2.5–7.5 min). According to the rotatable central composite design applied, 215 °C and 7.5 min were confirmed to be the optimal conditions, considering the maximization of enzymatic hydrolysis yield as optimization criterion. These conditions led to a maximum yield of 72.3%, equivalent to 81% of potential glucose in pretreated solid. Different configurations for bioethanol production from steam exploded rapeseed straw were investigated using the pretreated solid obtained under optimal conditions as a substrate. As a relevant result, concentrations of ethanol as high as 43.6 g/L (5.5% by volume) were obtained as a consequence of using 20% (w/v) solid loading, equivalent to 12.4 g ethanol/100 g biomass. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Biofuels production can be a sustainable alternative to fossil energy sources as well as a solution to combat climate change (Gupta, 2014). Lignocellulose biomass like forest and agricultural residues can be considered an interesting feedstock for second generation bioethanol production without competition with food or feed uses and more environmentally friendly. They are renewable, abundant, cheap and they can be alternative renewable resources. As a source of raw materials, rapeseed opens up a multitude of possibilities in the food and feed industry, as well as in the energy and material sectors (UFOP, 2014). The land surface planted with rapeseed in the world has increased from approximately 26 million hectares in 2000 to more than 36 million hectares in 2013 (FAOSTAT, 2015). Rapeseed straw can be left on the fields to ⇑ Corresponding author. E-mail address: [email protected] (I. Romero). http://dx.doi.org/10.1016/j.biortech.2015.04.066 0960-8524/Ó 2015 Elsevier Ltd. All rights reserved.
maintain soil health. Currently, the greatest demand for straw is for burning in power stations (Ryden et al., 2014). Besides, this cellulosic biomass may constitute a renewable energy resource for fuel ethanol production by biological conversion. The first step in the conversion of lignocellulosic biomass into fuels or chemicals typically involves a biomass pretreatment step in order to make the cellulose easier to hydrolyze by removing the hemicellulose, reducing the crystallinity of cellulose, and increasing the surface area (Wettstein et al., 2012). Effective pretreatment is fundamental to a successful enzymatic hydrolysis (EH) since thus the saccharification of the pretreated feedstock is enhanced by improving the accessibility of enzymes to the cellulose fibers (Limayem and Ricke, 2012). Steam explosion is recognized as a low-cost option for agricultural residues pretreatment with significantly lower environmental impact since the addition of external chemicals is not necessary (Singh et al., 2015). Moreover, it is considered as an effective pretreatment that causes the breakdown of lignocellulose structure which results in the
98
J.C. López-Linares et al. / Bioresource Technology 190 (2015) 97–105
partial hydrolysis of hemicellulose and disruption of the lignocellulosic matrix by the sudden pressure drop (Romaní et al., 2013). Although an acid presoaking can improve the treatment, the highly acetylated nature of some lignocellulosic materials allows for uncatalyzed (or autocatalyzed) steam pretreatment due to the release of acetyl groups and formation of acetic acid (Alvira et al., 2010). In addition, steam explosion pretreatment without external catalyst addition avoid the additional chemical costs and associated problems derived of equipment corrosion. Thus, the uncatalyzed alternative means a significantly lower environmental impact for the chemical utilization (Romaní et al., 2013). Uncatalyzed steam explosion has previously been shown to be very effective for the pretreatment of others lignocellulosic materials like wheat straw (Ballesteros et al., 2006), Eucalyptus globulus (Romaní et al., 2013) or corn stover (Lu et al., 2010). The performance of the subsequent enzymatic hydrolysis and fermentation at a high solids loading allows to increase the final ethanol concentration. It is essential to improve the economy of the process by reducing the energy consumption in the distillation step (Larsen et al., 2008). However, mixing difficulties in the reactor due to the high viscosity can occur (Sassner et al., 2006) and the yeast can be inhibited reducing ethanol yield in the simultaneous saccharification and fermentation step or even the cessation of ethanol production (Hoyer et al., 2013). This works deals with the steam explosion pretreatment of rapeseed straw, evaluating the effect of the temperature and process time by the response surface methodology. The performance of enzymatic hydrolysis and fermentation processes at high solids loading was tested using the solid pretreated obtained under optimal steam explosion conditions as a substrate. 2. Methods 2.1. Raw material Rapeseed straw (8% moisture content) was collected in Seville, Spain, after seed harvest, air-dried at room temperature, milled using a laboratory hammer mill (Retsch, SM-100, Haan, Germany) to a particle size smaller than 1 cm and stored in a dry place until use. 2.2. Steam explosion pretreatment Rapeseed straw was pretreated by steam explosion, without previous impregnation, in a pilot unit based on Masonite technology and equipped with a 2-L reaction vessel, as described elsewhere (Ballesteros et al., 2006). The reactor was charged with 300 g (dry matter) of feedstock per batch and heated to the desired temperature, directly with saturated steam. The selected temperature was reached in 40–50 s and then, time counting was initiated. After the explosion, the material was recovered in a cyclone, cooling to about 40 °C and then filtered for liquid and solid recovery. The solid fractions (water-insoluble solids, WIS) were washed with 5 L distilled water to remove the remaining prehydrolysate and stored wet in plastic bags. The WIS were analyzed for hemicellulosic sugars, glucose and lignin content and used as substrates in enzymatic digestibility tests. Liquid fractions (prehydrolysates) were analyzed for sugars, acetic acid, phenolic compounds and sugar degradation products as formic acid, furfural and 5-hydroxymethylfurfural (HMF). 2.3. Statistical design of experiments Rapeseed straw was pretreated at different operational conditions according to a rotatable central composite design (a = 1.414) including one point and four replicates at the center
of domain selected for each factor under study (13 runs). Pretreatment assays were performed in random order. Temperature (in the range 185–215 °C) and time (2.5–7.5 min) were selected as variables. Center values and intervals were selected based on previous experience. The coded and uncoded values of factors in the rotatable central composite design are shown in Table 1. The experimental data were analyzed via response surface methodology by the statistical software DesignExpert 8.0.7.1, Stat-Ease Inc., Minneapolis, USA. 2.4. Enzymatic hydrolysis tests The washed WIS fractions of rapeseed straw obtained after pretreatment were used as substrates for enzymatic hydrolysis assays. These WIS fractions were hydrolyzed with a cellulolytic complex (Celluclast 1.5 L) kindly provided by Novozymes A/S (Denmark), which contains an activity of 69 filter paper units (FPU)/mL. The cellulase enzyme loading was 15 FPU/g substrate. To supplement b-glucosidase activity of cellulases, fungal b-glucosidase (Novozym 188, Novozymes A/S), with an activity of 530 international units (IU)/mL was added, at an enzyme loading of 15 IU/g substrate. The pH was adjusted to 4.8 with 0.05 M sodium citrate buffer and enzymes were added to the pretreated substrate (5% w/v dry basis) for a total working volume of 20 mL in 100 mL Erlenmeyer flasks. Triplicate reaction flasks were incubated at 50 °C in an orbital shaker (Certomat-R, B-Braun, Germany) at 150 rpm for 72 h. Two milliliter samples were withdrawn at 24, 48 and 72 h, and they were centrifuged at 10,000g (Sigma 1–14 Centrifuge) for 10 min. Glucose and xylose concentrations in the sample supernatant were determined by HPLC. Once the steam explosion conditions were optimized to maximize the enzymatic hydrolysis yield, different substrate and enzyme loadings were tested. Then, the WIS fraction obtained under optimal conditions was washed and submitted to enzymatic hydrolysis tests as described above. In order to evaluate the performance of enzymatic hydrolysis tests, two parameters were determinate as follows. Saccharification efficiency (SE) was calculated as the ratio of grams of glucose released by enzymatic hydrolysis per 100 g glucose in the WIS, and the enzymatic hydrolysis yield (YEH) was referred to the glucose content in the raw material. All enzymatic hydrolysis experiments were performed in triplicate. Average results and standard deviations are given. 2.5. Microorganism, medium and yeast cultivation Saccharomyces cerevisiae (Fermentis Ethanol Red, Cedex, France) was used for fermentation assays. Yeast was maintained on solid Table 1 Operational conditions assayed expressed as dimensional and dimensionless independent variables. Run
1 2 3 4 5 6 7 8 9 10 11 12 13
Temperature (°C)
Time (min)
Real value
Coded value
Real value
Coded value
221.21 200.00 185.00 200.00 215.00 200.00 185.00 215.00 200.00 200.00 200.00 200.00 178.79
+1.414 0 1 0 1 0 1 1 0 0 0 0 1.414
5.00 8.54 2.50 5.00 7.50 5.00 7.50 2.50 5.00 5.00 5.00 1.46 5.00
0 +1.414 1 0 1 0 1 1 0 0 0 1.414 0
J.C. López-Linares et al. / Bioresource Technology 190 (2015) 97–105
culture media (g/L): glucose, 20; yeast extract, 5; peptone, 5; NH4Cl, 2 g/L; KH2PO4, 1, MgSO47H2O, 0,5; and agar 20, at 4 °C. Inoculum was prepared by transferring the yeast maintained on solid culture medium into 25 mL of the growth media consisting of (g/L): yeast extract, 5; ClNH4, 2; KH2PO4, 1; MgSO47H2O, 0.3; and glucose, 30. Cells were grown in 100 mL Erlenmeyer flasks with a volume of 25 mL and agitated on a rotatory shaker (Certomat-R, B-Braun, Germany) at 150 rpm and 30 °C for 16 h. Inoculum (4% v/v) yeast cultures correspond to a cell addition of 0.25 g/L. 2.6. Simultaneous saccharification and fermentation (SSF) Saccharification and fermentation assays were carried out using WIS fraction obtained under optimal conditions. The SSF experiments were performed under microaerobic conditions in 100 mL Erlenmeyer flasks with a total volume of 20 mL in citrate buffer (0.05 M, pH 4.8) containing the nutrients described above for inoculum cultivation except for glucose. Substrate concentration was 20% and 30% (w/v). SSF was started by adding simultaneously enzymes and S. cerevisiae inoculum (4% v/v) yeast cultures, corresponding to a cell addition of 0.25 g/L. The commercial enzymes used were the same as in the enzymatic hydrolysis tests, with cellulase enzyme loading of 20 and 40 FPU/g substrate. The assays were carried out during 72 h at 37 °C and 200 rpm and compared to controls that contained the same cellulose loading (Sigmacell). The flasks were equipped with a thick rubber stopper, through which one stainless-steel needle had been inserted to permit evolved CO2 to leave and maintain microaerobic conditions. All SSF experiments were performed in triplicate and the average results are given. Samples were withdrawn at 24, 48 and 72 h, centrifuged at 10,000g for 10 min for the determination of produced ethanol and consumed sugars. SSF results are reported as percentage of the theoretical yield considering that all the potential glucose in the pretreated solid is available for fermentation, and a theoretical fermentation yield of 0.51 g ethanol/g glucose. 2.7. Prehydrolysis and simultaneous saccharification and fermentation (PSSF) A previous hydrolysis for 6 and 24 h at 50 °C was carried out in triplicate before inoculation using the pretreated solid obtained under optimal conditions as substrate. The pretreated material concentration used in these assays was 30% (w/v). The enzyme loadings were the same as in the SSF test. Afterwards, the temperature was decreased to 37 °C and the flasks were inoculated with harvested yeast at the same concentration as in the SSF process. The flasks were incubated in a rotatory shaker at 200 rpm during 72 h (including the presaccharification time). Samples were taken and analyzed for carbohydrates and ethanol concentrations at 24, 48 and 72 h of process. All PSSF experiments were performed in triplicate. Average results are given. 2.8. Analytical methods The composition of the raw material and the solid fractions obtained after steam explosion pretreatment were determined according to NREL methodology as described in López-Linares et al. (2014). Sugar concentration was determined by high performance liquid chromatography (HPLC) in a Waters 2695 liquid chromatograph with a refractive index detector. A CARBOSep CHO-682 LEAD column (Transgenomic, Omaha, NE) operating at 80 °C with Milli-Q water (Millipore) as mobile-phase (0.5 mL/ min) was used.
99
Furfural and HMF were analyzed by HPLC (Agilent, Waldbronn, Germany) employing an Aminex HPX-87H column (Bio-Rad Labs, Hercules, CA) at 65 °C. The mobile phase contained 89% (5 mM H2SO4) and 11% acetonitrile at flow rate of 0.7 mL. A 1050A Photodiode-Array detector (Agilent, Walsbronn, Germany) was employed for detection. Formic and acetic acid were quantified by HPLC (Waters, Milfors, MA) using a 410 Water refractive index detector. An Aminex HPX-87H (Bio-Rad Labs, Hercules, CA) column maintained at 65 °C and a mobile phase of 5 mM H2SO4 at a flow rate of 0.6 mL/min were employed. Phenolic compounds were quantified in the liquid fractions a slightly modified version of Folin–Ciocalteu method (Zhang et al., 2006). Twenty microliter of sample and the serial standard solution were diluted with 88 lL of water on a 96-well microplate. After the addition of 12 lL Folin–Ciocalteu reagent, the plate was incubated for 5 min at room temperature in dark conditions. The reaction was stopped with 80 lL of 7.5% sodium carbonate solution. Before reading, the plate was incubated for 2 h at room temperature in the dark. The absorbance was measured at 765 nm with a spectrophotometric microplate reader (Anthos Zenyth 200rt, Biochrom, UK). Glucose and xylose concentrations from EH samples was measured by HPLC with the above described Waters equipment. Glucose and ethanol concentrations from SSF and PSSF samples were also measured by HPLC, using Bio-Rad HPX-87H column. All analytical determinations were performed in triplicate and the average results are shown. Relative standard deviations were below 3%. 3. Results 3.1. Characterization of rapeseed straw The initial composition of rapeseed straw was determined as 49% total carbohydrates (31.6 ± 0.3% cellulose and 17.4 ± 0.1% hemicellulose), 17.8 ± 0.5 lignin, 3.4 ± 0.1 acetyl groups, 15.4 ± 1.3 extractives, and 6.7 ± 0.3 ashes. These results are mean values of three determinations. The hemicellulose consisted of 76% xylose. By comparing with the composition of rapeseed straw analyzed by other researchers (Karagöz et al., 2012; Garmakhany et al., 2014), the raw material used in this research had less carbohydrates and a higher content of other minor fractions like acetyl groups, extractives and ash. According to Godin et al. (2013), the individual composition for each type of biomass is relatively constant regardless of the crop conditions (year, area, cultivar, nitrogen fertilization level) although the biomass chemical components are mainly influenced by the crop dry matter yield. A significant amount of non-structural compounds, about 15%, was determined in the rapeseed straw as extractives, mainly water soluble materials. 3.2. Effect of steam explosion conditions on WIS and prehydrolysates The solid recovery yield, as based on the initial biomass weight, ranged between 55%, at the highest temperatures (runs 1, 8), and 74%, at the mildest conditions (run 3), Table 2. Steam explosion pretreatment of rapeseed straw meant a cellulose enrichment in WIS due to the removal of extractives and partial solubilisation of hemicellulose fraction since in all conditions tested, a part of this fraction remained in the solid fraction. Similarly, the lignin content in WIS increased, which can be attributed to condensation reactions between lignin, extractives and sugars degradation products according to Ballesteros et al. (2011). Thus, WIS with the highest cellulose content, about 50%, were obtained at 200 °C and 215 °C (runs 2, 5), since at those conditions high solubilisation of
100
J.C. López-Linares et al. / Bioresource Technology 190 (2015) 97–105
Table 2 Solid recovery and WIS composition. Run
Solid recovery (%)
1 2 3 4 5 6 7 8 9 10 11 12 13
54.8 59.8 74.0 62.7 59.4 67.5 73.2 54.5 70.4 67.3 61.7 69.1 72.0
Table 3 Sugar and inhibitor composition (g/100 g raw material) of prehydrolysates.
WIS composition (%) Cellulose
Hemicellulose
Lignin
47.3 ± 1.16 50.3 ± 0.40 40.7 ± 0.55 44.1 ± 0.37 49.9 ± 0.18 42.9 ± 0.64 41.3 ± 0.47 48.3 ± 0.26 42.3 ± 0.34 44.9 ± 0.74 46.5 ± 0.70 44.2 ± 1.00 43.4 ± 0.95
1.8 ± 0.11 5.8 ± 0.07 18.5 ± 0.23 7.2 ± 0.15 1.8 ± 0.15 8.6 ± 0.15 14.0 ± 0.02 4.6 ± 0.02 8.5 ± 0.21 8.7 ± 0.06 9.4 ± 0.05 14.8 ± 0.21 19.6 ± 0.56
44.2 ± 0.30 40.8 ± 0.28 29.5 ± 0.33 35.7 ± 0.18 47.8 ± 1.12 33.9 ± 2.89 32.6 ± 0.47 40.4 ± 0.39 39.5 ± 2.49 37.6 ± 0.78 38.7 ± 0.10 31.9 ± 0.51 29.5 ± 0.71
hemicellulosic sugars was reached, yielding WIS with a hemicellulosic content below 6%. However, at 221.21 °C (star point), a mild decrease in cellulose content was detected, which can be attributed to some degradation of the cellulose fraction by severe conditions of pretreatment. At 185 °C (runs 3 and 7), regardless of the pretreatment time, the cellulose content was about 41%, while the hemicellulose content was over than 14%, which can be a drawback for the subsequent enzymatic hydrolysis since hemicellulose abundance can hinder enzymatic access to the cellulose (Ballesteros et al., 2006). In the prehydrolysates obtained after steam explosion pretreatment, sugars were present mainly in oligomeric form, between 60% and 90%, so that a posthydrolysis step was performed to determine the total amount of sugars. Hemicellulose was partially hydrolyzed resulting in hexoses (glucose, galactose and mannose), pentoses (xylose and arabinose) and a variety of oligosaccharides. As can be observed in Table 3, low recoveries of hemicellulosic sugars in prehydrolysates were attained, yielding sugar solutions with maxima concentrations about 20 g/L. This behavior indicates high degradation of these sugars at the pretreatment conditions assayed. As a consequence, the prehydrolysates contained also non-sugars compounds, as shown in Table 3. In general, inhibitors concentration increased when the severity of pretreatment did. The pretreatment usually degrades the hemicellulose, leading to the formation of products such as aliphatic acids (primarily acetic acid and formic acid), and furan aldehydes (5-hydroxymethylfurfural and furfural). Similarly to hemicellulose, lignin is also partially degraded to the phenolic compounds. Phenolics derived from lignin and other plant components can create significant inhibition on enzymatic conversion of cellulosic biomass materials to useful chemicals (Tejirian and Xu, 2011). Furthermore, the degradation of cellulose can also lead to the formation of byproducts such as HMF (Palmqvist and Hahn-Hägerdal, 2000). The most abundant inhibitory compounds detected in the prehydrolysates were acetic and formic acids. High acetic acid concentrations can be attributed to the high content of acetyl groups in the raw rapeseed straw (3.4%), greater than that in other raw materials such as barley straw (Duque et al., 2013) or olive tree pruning (LópezLinares et al., 2013). The highest acetic acid concentrations, above 3 g/100 g, were achieved at maximum temperature, which can be attributed to a greater hemicellulose solubilisation, corresponding to WIS with the lowest hemicellulose content (runs 1, 5) as can be seen in Table 2. The presence of formic acid in the prehydrolysates can be related to a high degradation of hemicellulosic sugars, since the highest concentrations of this inhibitor compound (more than 2 g/100 g rapeseed straw) were achieved in runs 1 and 5, coinciding with xylose recoveries lower than 0.6 g/100 g rapeseed straw (Table 3). In addition, the high concentrations of formic acid detected can be also related to the low concentrations of furfural
Run
Glucose
Xylose
Galactose
Arabinose
Mannose
1 2 3 4 5 6 7 8 9 10 11 12 13
0.28 ± 0.02 0.42 ± 0.09 0.83 ± 0.15 0.58 ± 0.09 0.19 ± 0.01 0.38 ± 0.03 0.72 ± 0.11 0.57 ± 0.10 0.54 ± 0.01 0.60 ± 0.07 0.57 ± 0.08 0.72 ± 0.04 0.85 ± 0.10
0.62 ± 0.09 2.42 ± 0.21 0.89 ± 0.13 4.10 ± 0.11 0.52 ± 0.08 1.72 ± 0.19 2.53 ± 0.14 3.19 ± 0.26 2.51 ± 0.12 3.41 ± 0.13 2.58 ± 0.25 2.20 ± 0.07 0.79 ± 0.05
0.13 ± 0.01 0.32 ± 0.03 0.92 ± 0.02 0.61 ± 0.05 0.06 ± 0.01 0.36 ± 0.07 0.83 ± 0.11 0.47 ± 0.03 0.46 ± 0.10 0.60 ± 0.13 0.44 ± 0.07 0.73 ± 0.12 0.86 ± 0.06
0.04 ± 0.00 0.12 ± 0.01 0.90 ± 0.10 0.22 ± 0.02 0.03 ± 0.00 0.17 ± 0.01 0.53 ± 0.05 0.18 ± 0.02 0.25 ± 0.04 0.30 ± 0.06 0.17 ± 0.01 0.67 ± 0.03 0.85 ± 0.11
0.14 ± 0.03 0.23 ± 0.04 0.21 ± 0.03 0.24 ± 0.01 0.00 ± 0.00 0.17 ± 0.01 0.19 ± 0.01 0.29 ± 0.07 0.18 ± 0.05 0.25 ± 0.07 0.11 ± 0.01 0.19 ± 0.03 0.18 ± 0.02
Acetic acid
Formic acid
Furfural
HMF
Total phenols
1 2 3 4 5 6 7 8 9 10 11 12 13
3.23 ± 0.17 2.87 ± 0.11 0.63 ± 0.09 1.71 ± 0.12 3.40 ± 0.19 1.81 ± 0.11 1.10 ± 0.13 2.58 ± 0.15 2.06 ± 0.13 2.06 ± 0.07 2.09 ± 0.22 0.96 ± 0.09 0.63 ± 0.08
2.28 ± 0.12 2.07 ± 0.18 0.81 ± 0.07 1.36 ± 0.05 2.18 ± 0.19 1.39 ± 0.11 1.05 ± 0.13 1.96 ± 0.05 1.70 ± 0.06 1.66 ± 0.04 1.59 ± 0.09 1.04 ± 0.11 0.73 ± 0.03
0.22 ± 0.03 0.18 ± 0.02 0.01 ± 0.00 0.10 ± 0.01 0.23 ± 0.02 0.11 ± 0.01 0.04 ± 0.00 0.19 ± 0.03 0.10 ± 0.01 0.10 ± 0.00 0.13 ± 0.01 0.03 ± 0.00 0.01 ± 0.00
0.10 ± 0.01 0.08 ± 0.02 0.05 ± 0.01 0.07 ± 0.01 0.09 ± 0.02 0.06 ± 0.00 0.05 ± 0.00 0.08 ± 0.01 0.06 ± 0.01 0.06 ± 0.02 0.06 ± 0.01 0.05 ± 0.00 0.04 ± 0.01
2.77 ± 0.11 2.38 ± 0.05 1.65 ± 0.09 2.05 ± 0.15 2.06 ± 0.13 1.63 ± 0.12 1.54 ± 0.08 2.38 ± 0.20 2.06 ± 0.17 1.92 ± 0.16 2.06 ± 0.13 1.79 ± 0.11 1.55 ± 0.08
and HMF in these prehydrolysates since the formation of formic acid has been reported due to the degradation of furfural and HMF (Palmqvist and Hahn-Hägerdal, 2000). However, these inhibitor concentrations are slightly lower than those obtained by Wood et al. (2014) at similar conditions of steam explosion pretreatment, obtaining also low concentrations of sugars in the liquid fractions with the same feedstock. It is noteworthy that in all conditions assayed, these sugars were present mainly in oligomeric form, between 63% and 92%. The pH of prehydrolysates was about 4 and no significant differences were detected at the different conditions tested. 3.3. Evaluation of steam explosion at different operational conditions The effects of the steam explosion pretreatment conditions on the glucose recovery in WIS (GRWIS), overall glucose recovery (GRoverall), saccharification efficiency (SE) and enzymatic hydrolysis yield (YEH) were investigated using the response surface methodology. The response GRWIS is expressed as percentage of glucose remained in WIS referred to glucose content in raw material. Overall glucose recovery, GRoverall, is expressed as percentage of glucose recovered in both solid and liquid streams referred to glucose content in raw material. The responses SE and YEH are expressed as the percentage of the glucose content in WIS or in raw material which is released by enzymatic hydrolysis. The regression equations in terms of normalized values are shown in Eqs. (1)–(4).
GRWIS ð%Þ ¼ 93:10 5:64T þ 1:23t 1:63T 2 ðR2 ¼ 0:915; R2 adjust ¼ 0:873Þ GRoverall ð%Þ ¼ 94:79 5:88T þ 0:92t 1:91T 2 ðR2 ¼ 0:930; R2 adjust ¼ 0:895Þ SE ð%Þ ¼ 61:98 þ 22:24T þ 4:90t ðR2 ¼ 0:929; R2 adjust ¼ 0:913Þ
ð1Þ
ð2Þ
ð3Þ
101
J.C. López-Linares et al. / Bioresource Technology 190 (2015) 97–105
YEH ð%Þ ¼ 59:33 þ 17:35T þ 4:50t 2:67T 2 2
ð4Þ
2
ðR ¼ 0:988; R adjust ¼ 0:982Þ
The values of R2 and R2-adjust in each equation show that good agreement between experimental and predicted data was achieved. It can be observed from the coefficients of the equations that the effect of the temperature was more important than the effect of the pretreatment time. Moreover, no response was affected by the interaction between factors. Regarding glucose recovery in solid fraction (GRWIS) and overall glucose recovery (GRoverall) the effect of temperature was negative while the pretreatment time exerted a positive effect although less significant (Eqs. (1) and (2)). The highest GRWIS was obtained in run 13 (178.79 °C, 5 min) and on the contrary, the lowest glucose recovery in solid fraction was found at the highest temperature (run 1, 221.21 °C, 5 min), Table 4. This fact can be attributed to some degradation of cellulose fraction at these severe conditions. However, it is worth noting that the GRWIS was over 82% in all conditions assayed (Table 4), indicating low levels of degradation and therefore remaining solids rich in cellulose after steam explosion pretreatment with negligible loss of glucose. Figs. 1 and 2 illustrate
Table 4 Glucose recovery in pretreated solids and results of enzymatic hydrolysis. Run
Glucose recovery in WIS (%)
Overall glucose recovery (%)
Saccharification efficiency (%)
Enzymatic hydrolysis yield (%)
1 2 3 4 5 6 7 8 9 10 11 12 13
82.2 ± 2.03 95.5 ± 0.76 95.6 ± 1.29 87.9 ± 0.73 94.1 ± 0.33 92.0 ± 1.38 96.1 ± 1.09 83.6 ± 0.45 94.5 ± 0.75 95.9 ± 1.57 91.1 ± 1.37 97.1 ± 2.20 99.3 ± 2.16
83.0 ± 2.11 96.7 ± 0.78 98.0 ± 1.30 89.6 ± 0.76 94.7 ± 0.32 93.1 ± 1.39 98.2 ± 1.15 85.3 ± 0.44 96.0 ± 0.77 97.7 ± 1.62 92.8 ± 1.40 99.2 ± 2.28 100.0 ± 2.15
95.7 ± 2.63 71.1 ± 2.24 34.9 ± 2.03 69.9 ± 0.59 84.6 ± 0.94 61.9 ± 1.21 42.9 ± 1.95 83.8 ± 0.42 69.0 ± 4.31 62.5 ± 1.30 56.6 ± 2.25 45.6 ± 1.18 32.1 ± 3.95
78.6 ± 2.16 67.9 ± 2.14 33.3 ± 1.94 61.5 ± 0.52 79.6 ± 0.88 57.0 ± 1.11 41.2 ± 1.88 70.1 ± 0.35 65.2 ± 4.08 60.0 ± 1.25 51.6 ± 2.05 44.3 ± 1.14 31.9 ± 3.92
the influence of temperature and process time on GRWIS and GRoverall. As can be observed in these figures, the behavior of both GRWIS and GRoverall was similar/the same since almost all glucose was recovered in pretreated solids while the fraction of glucose solubilised was minimal. It can be observed that an increase in temperature limited the glucose recovery in both fractions. Concerning SE and YEH (Eqs. (3) and (4)), both the temperature and the time had a significant positive effect in these responses. However, the process time did not affect hydrolysis process significantly as can also be observed in the response surface for saccharification efficiency (Fig. 3) and enzymatic hydrolysis yield (Fig. 4). Both figures show that an increase in the temperature implies a substantial increase in the enzymatic hydrolysis yields, referred to glucose content in WIS (SE) or in raw material (YEH), respectively. Quadratic effect on YEH (Eq. (4)) indicates that the values of the temperature stabilized, i.e., there is a point from which the enzymatic hydrolysis yield did not increase due to that factor. This behavior can be also observed in Fig. 4 where a light curvature indicates that there is a point from which the enzymatic hydrolysis yield does not increase due to the temperature. At the lowest temperatures tested (178.79 °C and 185 °C), regardless of the process time, the enzymatic hydrolysis yields (YEH) were lower than 42%. These temperatures, although yielded pretreated solids with the highest glucose recoveries, (GRWIS), above 95% (Table 4), and higher hemicellulose content, above 14% (Table 2), did not achieve to alter the lignocellulosic structure enough to make the enzymatic attack easier. It can be observed in Table 4 that the lowest cellulose conversion corresponds to the assay carried out at the lowest temperature (178.79 °C, run 13) while the maximum value of SE (95.7%) was achieved in the experiment carried out at 221.21 °C (run 1). However, the enzymatic hydrolysis yield (YEH) attained at this temperature (78.6%) was slightly lower than that achieved at 215 °C and 7.5 min (run 5). It means that cellulose digestibility increased as temperature did, regardless of the pretreatment time, although at the maximum temperature assayed (star point) some degradation of cellulose fraction occurred, reducing the enzymatic hydrolysis yield referred to glucose content in raw rapeseed straw (YEH).
101
Overall glucose recovery (%)
102
GRWIS (%)
95 89 83 77
1.00
175.00
3.25
Tim
187.50 5.50
e( mi n)
200.00 7.75
212.50 10.00 225.00
Tem
a per
tu r
e(
ºC)
Fig. 1. Response surface and contour plot of temperature and pretreatment time on glucose recovery in WIS.
96 90 84 78 175.00
1.00 187.50
3.25
Tim e
200.00
5.50
(m in)
212.50
7.75 10.00
225.00
Te
m
pe
ra
t
e ur
(º
C
)
Fig. 2. Response surface and contour plot of temperature and pretreatment time on overall glucose recovery.
102
J.C. López-Linares et al. / Bioresource Technology 190 (2015) 97–105 Table 5 Results obtained after 72 h of enzymatic hydrolysis at different solid loadings and enzyme dosages using as substrate WIS under optimal steam explosion conditions.
Saccharification efficie ncy (%)
109
86
63
40
Enzyme loading (FPU/ g WIS)
Glucose concentration (g/L)
SE (%)
YEH (%)
5 20 20 30 30
15 20 40 20 40
22.0 ± 0.45 82.0 ± 1.22 85.9 ± 0.68 108.4 ± 0.09 114.0 ± 0.29
80.9 75.3 78.9 66.3 69.8
72.3 67.3 70.5 59.3 62.4
SE (%): grams of glucose released by enzymatic hydrolysis per 100 g glucose present in the WIS. YEH (%): grams of glucose released by enzymatic hydrolysis per 100 g glucose present in the raw material.
17
10.00
225.00 7.75
212.50
Tim 5.50 e( mi n)
200.00 3.25
187.50 1.00
175.00
re atu per m Te
) (ºC
Fig. 3. Response surface and contour plot of temperature vs. pretreatment time on saccharification efficiency.
90
71.25
YEH (%)
Solids loading (%)
52.5
33.75
15
10.00
225.00 7.75
212.50 5.50
Tim e (m in)
200.00 3.25
187.50 1.00
175.00
Te
t era mp
( ure
ºC)
Fig. 4. Response surface and contour plot of temperature vs. process time on enzymatic hydrolysis yield.
3.4. Model optimization The mathematical model that was developed from the experimental results is able to predict the operational conditions that should be used in steam explosion pretreatment to optimize model responses. In this study, the optimization focused in the maximization of YEH and the maximum yield is predicted to be reached if pretreatment is performed at 215 °C for 7.5 min. These conditions coincide with the experimental run 5, yielding the highest value of YEH (79.6%). To validate the model, a new experiment was carried out (by triplicate), setting the process variables to the optimum ones given above. Enzymatic hydrolysis yield (YEH) at optimum conditions was 72.3%, equivalent to 81% of potential glucose in WIS. This result is slightly lower than that predicted by the model, YEH (78%) although within the limits of variability in confidence level of 95%.
3.5. Enzymatic hydrolysis of WIS obtained under optimal pretreatment conditions Pretreated rapeseed straw obtained under optimal steam explosion conditions (215 °C, 7.5 min) was used as substrate for enzymatic hydrolysis tests at high solids loading (20% and 30% w/v). Moreover, in order to assess the influence of enzyme dosage, 20 and 40 FPU/g WIS were tested at both 20% and 30% (w/v) substrate concentration. Table 5 shows conditions and results from these experiments. For comparative purposes, a standard assay at 5% (w/v) solids loading and 15 FPU/g WIS was carried out. All enzymatic hydrolysis experiments were performed in triplicate and average results are given. An increase in the solids loading limited the saccharification efficiency, probably due to enzyme inhibition exerted by high glucose concentration or by mixing difficulties (Koppram et al., 2014) although some authors suggest that the limitations due to mass diffusion are more prevalent than the sugar inhibition (Modenbach and Nokes, 2013). The efficiency saccharification decreased from 75.3% to 66.3% as solids loading increased from 20% to 30% (w/v) at 20 FPU/g WIS. Using the highest dose of enzyme, similar reductions in the conversion of cellulose (by 10%) are obtained with increasing solids loading. However, it can be essential the use of high solids loading during enzymatic hydrolysis to achieve concentrated glucose solutions resulting in high ethanol concentrations in the subsequent distillation step (Larsen et al., 2008). Glucose concentrations of 82 g/L and 85.9 g/ L were achieved after 72 h of enzymatic hydrolysis for 20% (w/v) solids loading at 20 and 40 FPU/g WIS, respectively (Table 5). The system containing 30% solids achieved maxima glucose concentrations of 108.4 g/L and 114 g/L at enzyme dosage of 20 and 40 FPU/g WIS, respectively. It can be observed for both enzyme loadings that glucose concentrations increased although not proportionally with substrate loading. Thus, the highest values of SE and YEH were achieved when the enzymatic hydrolysis was performed at standard conditions, 5% (w/v) solids loading and 15 FPU/g WIS. At these conditions, enzymatic hydrolysis yielded 81 g of glucose per 100 g of glucose content in WIS and 72 g of glucose per 100 g glucose content in raw rapeseed straw. Under these conditions product inhibition or mixing problems are not expected. A maximum saccharification efficiency and enzymatic hydrolysis yield can be obtained. However, glucose concentration was as low as 22 g/L. Nevertheless, to achieve the required ethanol concentrations for efficient distillation (4.4% w/w), substrate concentration in EH must be at least 20% (w/w) which can release glucose solutions of 8% w/w (Larsen et al., 2008). In this research, this assumption holds for solids loading of 20% (w/v) and enzyme dosage of 20 FPU/g WIS, yielding a glucose solution of 8% (w/w). Considering the obtained results, it is noteworthy that the rapeseed straw pretreated by steam explosion is a good substrate than
103
J.C. López-Linares et al. / Bioresource Technology 190 (2015) 97–105
Concentration (g/L)
a
Time (h)
b Concentration (g/L)
can be enzymatically hydrolyzed at high solids loading with relatively low drop in the yield. High enzymatic hydrolysis yields have been achieved by other researchers using steam explosion (Wood et al., 2014) with rapeseed straw although only at 1% (w/w) solids loading. High enzyme loadings can improve enzymatic hydrolysis yield at high solids loading (Xue et al., 2012). However, in this research this improvement was not detected, maybe due to both solid loadings used (20% and 30% w/v) can be considered high substrate loadings. As can be seen in Table 5, the increase in enzyme dosage from 20 to 40 FPU/g WIS increased the saccharification efficiency by only 3.5% regardless of the substrate loading. According to Olofsson et al. (2008), an increase of 50% in enzyme loading should be justified if an increase of yields greater than 6% is achieved. Therefore, it may not be worthy converting an extra 3.5% of glucose since the enzyme cost has been recognized as a considerable contributor to bioethanol cost (Wang et al., 2012). According to Modenbach and Nokes (2013), a sequential addition of enzymes can be advantageous because the addition of fresh enzyme can replace the non-productive enzyme bound to the lignin. This feeding strategy was tested for exploded rapeseed straw at optimal conditions, to improve the enzymatic hydrolysis performance (data not shown), although it did not achieve an improvement in hydrolysis yields compared to the assays carried out adding complete enzyme dosage at the beginning of the process. Therefore, the relatively low enzymatic hydrolysis yields are not due to enzyme adsorption to lignin, and rather must be due to mass transfer limitations.
3.6. SSF and PSSF for pretreated solid obtained under optimal conditions Exploded rapeseed straw at optimal conditions was submitted to an SSF process at the same solids and enzyme loadings tested for EH assays as described in Section 3.5. At 20% solids loading, complete conversion of glucose to ethanol was realized within 48 h for both enzyme loadings assayed (Fig. 5a) without any significant difference in time course of glucose and ethanol concentrations during SSF process. Final ethanol concentrations of 42.8 g/L and 43.6 g/L were attained at 20 and 40 FPU/g WIS, respectively. It should be noted that the main ethanol production occurs at the beginning of the SSF process. Thus, at lower solid loading tested, in the 0–24 h period of process, there was a production about 34 g/L of ethanol, while in the 48–72 h period, the increase was only 1.5 g/L (Fig. 5a). According to this behavior, the process time could be shortened from 72 to 48 h. However, as can be seen in Fig. 5b, when substrate loading increased to 30% (w/v), fermentation resulted in incomplete glucose consumption yielding ethanol concentrations of 54.9 g/L and 50.5 g/L at 20 and 40 FPU/g WIS, respectively. After 72 h, 13.5 g/L and 19.2 g/L of glucose remained in the broth/medium as residual glucose with 20 and 40 FPU/g WIS, respectively, and no additional increases in ethanol concentration were detected after this time (data not shown), indicating the final of the fermentation process. Thus, fermentation performance was the limiting factor in SSF process since the yeast ceased after that time. Similarly as it occurs in EH, the influence of enzyme dosage in the SSF process, in the range tested, was minimal although at 30% solids, glucose conversion decreased with increasing enzyme loading. This fact can be related to an inefficient mixing of the enzyme with the substrate as the solids loading increased. Furthermore, by employing high loadings of solids (30% w/v) and high enzyme dosage (40 FPU/g WIS), some compounds could be released in the fermentation medium affecting the microorganism. Zhao et al. (2013) have reported that high cellulase loadings might
Time (h) Fig. 5. Time course of glucose (dashed lines) and ethanol (continuous lines) concentrations during SSF at 20 FPU/g WIS (d) and 40 FPU/g WIS (j) (a) for 20% (w/ v) and (b) 30% (w/v) substrate concentration.
cause inhibition on S. cerevisiae due to the presence of some additives in the solution of the commercial enzymatic complex. Table 6 shows the results attained in the different SSF tests. The ethanol yield was expressed as a percentage of the maximum ethanol yield (0.51 g ethanol/g glucose). And the process yield is determined as the ratio of the grams of ethanol after 72 h fermentation per 100 g dry straw. SSF at the high 30% (w/v) solids concentration resulted in less efficient process than at 20% (w/v) solids, probably because of
Table 6 Ethanol yields attained by SSF and PSSF at different conditions. Solids loading (%)
SSF 20 30 PSSF (presac. time) 30 (6 h) 30 (24 h)
Enzymatic loading (FPU/g WIS)
Ethanol concentration (g/L)
YE (%)
Process yield (%)
20 40 20 40
42.8 ± 0.1 43.6 ± 0.0 54.9 ± 0.5 50.5 ± 0.7
77.0 78.5 65.8 60.6
12.2 12.4 10.4 9.6
20
56.2 ± 0.3 52.5 ± 0.1
67.5 63.0
10.7 9.9
YE: expressed as percentage of theoretical ethanol yield (0.51 g ethanol/g glucose). Process yield: g ethanol/100 g rapeseed straw.
104
J.C. López-Linares et al. / Bioresource Technology 190 (2015) 97–105
limitations associated with laboratory-scale mixing and poor performance of the yeast. This behavior of the microorganism can be attributed to the combined effect of high stress by reducing the amount of water available and the hyperosmotic conditions in the fermentation broth (Koppram et al., 2014). The system containing 30% solids achieved the highest final ethanol concentration with nearly 55 g/L. However, the conversion decreased with increasing solids loadings, although the conversion of the 30% solids reaction was only 10% below systems at 20% solids (33.6% vs. 39.3% and 30.9% vs.40%, at 20 and 40 FPU/g WIS, respectively), Table 6. Comparing to other pretreatment options for rapeseed straw, steam explosion processing performed in this work followed to an SSF process enabled similar ethanol yields and higher ethanol concentrations than those achieved by liquid hot water followed also by an SSF process at 7.5% w/v (López-Linares et al., 2014). In order to evaluate the effect of a previous hydrolysis on the ethanol yield, prehydrolysis for 6 and 24 h followed by SSF was performed at 30% (w/v) solids and considering the results obtained in the SSF tests, the enzyme loading used in PSSF experiments was only 20 FPU/g WIS (Table 6). Two experiments were conducted to assess the timing of hydrolysis for better yield. Only the experiment carried out with 6 h of prehydrolysis yielded better results of ethanol production than the SSF process (56.2 g/L vs. 54.9 g/L) although the increase in ethanol yield was lower than 1%. Therefore, by comparing SSF and PSSF processes there was no difference between both configurations with regard to ethanol yield. Arvaniti et al. (2012) achieved also an ethanol yield of 67% (referred to the theoretical ethanol yield) from rapeseed straw pretreated by wet oxidation using a presaccharification followed by an SSF process at 12.5% (w/w) solids loading. The highest ethanol yield, i.e., 12.4 g ethanol/100 g biomass, corresponds to 70% of the maximum theoretical yield. This result was obtained from an SSF process at 20% (w/v) solids loading and 40 FPU/g WIS, resulting in an ethanol solution of 43.6 g/L (5.5% v/v). This result indicates a very high efficiency in the use of the cellulosic fraction of the rapeseed straw. Lu et al. (2009) and Luo et al. (2011) with the same feedstock achieved higher process yields (14% and 15% referred to 100 g of biomass) by an sequential process, although they used solid loadings lower than those used in this study, only 10% and 5% (w/w), respectively. 4. Conclusions From the results obtained in this research, it can be concluded that the uncatalyzed steam explosion is an effective pretreatment to increase the enzymatic digestibility of rapeseed straw. WIS obtained under optimal steam explosion conditions, 215 °C and 7.5 min, resulted in a good substrate for bioethanol production at high substrate concentration. A configuration of simultaneous saccharification and fermentation, at 20% (w/v) solid loading, allowed obtaining an ethanol solution of 5.3%, by volume, which corresponds to 77% of the theoretical ethanol yield. Further research will be focused on improving hemicellulosic sugar recoveries to increase ethanol production. Acknowledgement J.C. López-Linares gratefully acknowledges the research Grant received from University of Jaén (Plan de Apoyo). References Alvira, P., Tomás-Pejó, E., Ballesteros, M., Negro, M.J., 2010. Pretreatment technologies for an efficient bio ethanol production process based on enzymatic hydrolysis: a review. Bioresour. Technol. 101, 4851–4861.
Arvaniti, E., Bjerre, A.B., Schmidt, J.E., 2012. Wet oxidation pretreatment of rape straw for ethanol production. Biomass Bioenergy 39, 94–105. Ballesteros, I., Negro, M.J., Oliva, J.M., Cabañas, A., Manzanares, P., Ballesteros, M., 2006. Ethanol production from steam-explosion pretreated wheat straw. Appl. Biochem. Biotechnol. 130, 496–508. Ballesteros, I., Ballesteros, M., Cara, C., Sáez, F., Castro, E., Manzanares, P., Negro, M.J., Oliva, J.M., 2011. Effect of water extraction on sugars recovery from steam exploded olive tree pruning. Bioresour. Technol. 102, 6611–6616. Duque, A., Manzanares, P., Ballesteros, I., Negro, M.J., Oliva, J.M., Saez, F., Ballesteros, M., 2013. Optimization of integrated alkaline–extrusion pretreatment of barley straw for sugar production by enzymatic hydrolysis. Process Biochem. 48, 775– 781. FAOSTAT, 2015.
J.C. López-Linares et al. / Bioresource Technology 190 (2015) 97–105 Wood, I.P., Elliston, A., Collins, S.R.A., Wilson, D., Bancroft, I., Waldron, K.W., 2014. Steam explosion of oilseed rape straw: establishing key determinants of saccharification efficiency. Bioresour. Technol. 162, 175–183. Xue, Y., Jameel, H., Phillips, R., Chang, H., 2012. Split addition of enzymes in enzymatic hydrolysis at high solids concentration to increase sugar concentration for bioethanol production. J. Ind. Eng. Chem. 18, 707–714.
105
Zhang, Q., Zhang, J., Shen, J., Silva, A., Dennis, D.A., Barrow, C.J., 2006. A simple 96well microplate method for estimation of total polyphenol content in seaweeds. J. Appl. Phycol. 18, 445–450. Zhao, X., Dong, L., Chen, L., Liu, D., 2013. Batch and multi-step fed-batch enzymatic saccharification of Formiline-pretreated sugarcane bagasse at high solid loadings for high sugar and ethanol titers. Bioresour. Technol. 135, 350–356.
Contents lists available at ScienceDirect
Bioresource Technology journal homepage: www.elsevier.com/locate/biortech
Optimization of uncatalyzed steam explosion pretreatment of rapeseed straw for biofuel production Juan C. López-Linares a, Ignacio Ballesteros b, Josefina Tourán a, Cristóbal Cara a, Eulogio Castro a, Mercedes Ballesteros b, Inmaculada Romero a,⇑ a b
Department of Chemical, Environmental and Materials Engineering, University of Jaén, Campus Las Lagunillas, 23071 Jaén, Spain DER-CIEMAT, Avda. Complutense 22, 28040 Madrid, Spain
h i g h l i g h t s Ethanol production from steam exploded rape straw at high solids loading is reported. Operation at high solids loading (20% w/v) allows complete conversion of cellulose. High ethanol concentrations (5% v/v) are obtained, allowing industrial operation.
a r t i c l e
i n f o
Article history: Received 4 March 2015 Received in revised form 17 April 2015 Accepted 18 April 2015 Available online 23 April 2015 Keywords: Rapeseed straw Steam explosion SSF Fuel ethanol High solids
a b s t r a c t Rapeseed straw constitutes an agricultural residue with great potential as feedstock for ethanol production. In this work, uncatalyzed steam explosion was carried out as a pretreatment to increase the enzymatic digestibility of rapeseed straw. Experimental statistical design and response surface methodology were used to evaluate the influence of the temperature (185–215 °C) and the process time (2.5–7.5 min). According to the rotatable central composite design applied, 215 °C and 7.5 min were confirmed to be the optimal conditions, considering the maximization of enzymatic hydrolysis yield as optimization criterion. These conditions led to a maximum yield of 72.3%, equivalent to 81% of potential glucose in pretreated solid. Different configurations for bioethanol production from steam exploded rapeseed straw were investigated using the pretreated solid obtained under optimal conditions as a substrate. As a relevant result, concentrations of ethanol as high as 43.6 g/L (5.5% by volume) were obtained as a consequence of using 20% (w/v) solid loading, equivalent to 12.4 g ethanol/100 g biomass. Ó 2015 Elsevier Ltd. All rights reserved.
1. Introduction Biofuels production can be a sustainable alternative to fossil energy sources as well as a solution to combat climate change (Gupta, 2014). Lignocellulose biomass like forest and agricultural residues can be considered an interesting feedstock for second generation bioethanol production without competition with food or feed uses and more environmentally friendly. They are renewable, abundant, cheap and they can be alternative renewable resources. As a source of raw materials, rapeseed opens up a multitude of possibilities in the food and feed industry, as well as in the energy and material sectors (UFOP, 2014). The land surface planted with rapeseed in the world has increased from approximately 26 million hectares in 2000 to more than 36 million hectares in 2013 (FAOSTAT, 2015). Rapeseed straw can be left on the fields to ⇑ Corresponding author. E-mail address: [email protected] (I. Romero). http://dx.doi.org/10.1016/j.biortech.2015.04.066 0960-8524/Ó 2015 Elsevier Ltd. All rights reserved.
maintain soil health. Currently, the greatest demand for straw is for burning in power stations (Ryden et al., 2014). Besides, this cellulosic biomass may constitute a renewable energy resource for fuel ethanol production by biological conversion. The first step in the conversion of lignocellulosic biomass into fuels or chemicals typically involves a biomass pretreatment step in order to make the cellulose easier to hydrolyze by removing the hemicellulose, reducing the crystallinity of cellulose, and increasing the surface area (Wettstein et al., 2012). Effective pretreatment is fundamental to a successful enzymatic hydrolysis (EH) since thus the saccharification of the pretreated feedstock is enhanced by improving the accessibility of enzymes to the cellulose fibers (Limayem and Ricke, 2012). Steam explosion is recognized as a low-cost option for agricultural residues pretreatment with significantly lower environmental impact since the addition of external chemicals is not necessary (Singh et al., 2015). Moreover, it is considered as an effective pretreatment that causes the breakdown of lignocellulose structure which results in the
98
J.C. López-Linares et al. / Bioresource Technology 190 (2015) 97–105
partial hydrolysis of hemicellulose and disruption of the lignocellulosic matrix by the sudden pressure drop (Romaní et al., 2013). Although an acid presoaking can improve the treatment, the highly acetylated nature of some lignocellulosic materials allows for uncatalyzed (or autocatalyzed) steam pretreatment due to the release of acetyl groups and formation of acetic acid (Alvira et al., 2010). In addition, steam explosion pretreatment without external catalyst addition avoid the additional chemical costs and associated problems derived of equipment corrosion. Thus, the uncatalyzed alternative means a significantly lower environmental impact for the chemical utilization (Romaní et al., 2013). Uncatalyzed steam explosion has previously been shown to be very effective for the pretreatment of others lignocellulosic materials like wheat straw (Ballesteros et al., 2006), Eucalyptus globulus (Romaní et al., 2013) or corn stover (Lu et al., 2010). The performance of the subsequent enzymatic hydrolysis and fermentation at a high solids loading allows to increase the final ethanol concentration. It is essential to improve the economy of the process by reducing the energy consumption in the distillation step (Larsen et al., 2008). However, mixing difficulties in the reactor due to the high viscosity can occur (Sassner et al., 2006) and the yeast can be inhibited reducing ethanol yield in the simultaneous saccharification and fermentation step or even the cessation of ethanol production (Hoyer et al., 2013). This works deals with the steam explosion pretreatment of rapeseed straw, evaluating the effect of the temperature and process time by the response surface methodology. The performance of enzymatic hydrolysis and fermentation processes at high solids loading was tested using the solid pretreated obtained under optimal steam explosion conditions as a substrate. 2. Methods 2.1. Raw material Rapeseed straw (8% moisture content) was collected in Seville, Spain, after seed harvest, air-dried at room temperature, milled using a laboratory hammer mill (Retsch, SM-100, Haan, Germany) to a particle size smaller than 1 cm and stored in a dry place until use. 2.2. Steam explosion pretreatment Rapeseed straw was pretreated by steam explosion, without previous impregnation, in a pilot unit based on Masonite technology and equipped with a 2-L reaction vessel, as described elsewhere (Ballesteros et al., 2006). The reactor was charged with 300 g (dry matter) of feedstock per batch and heated to the desired temperature, directly with saturated steam. The selected temperature was reached in 40–50 s and then, time counting was initiated. After the explosion, the material was recovered in a cyclone, cooling to about 40 °C and then filtered for liquid and solid recovery. The solid fractions (water-insoluble solids, WIS) were washed with 5 L distilled water to remove the remaining prehydrolysate and stored wet in plastic bags. The WIS were analyzed for hemicellulosic sugars, glucose and lignin content and used as substrates in enzymatic digestibility tests. Liquid fractions (prehydrolysates) were analyzed for sugars, acetic acid, phenolic compounds and sugar degradation products as formic acid, furfural and 5-hydroxymethylfurfural (HMF). 2.3. Statistical design of experiments Rapeseed straw was pretreated at different operational conditions according to a rotatable central composite design (a = 1.414) including one point and four replicates at the center
of domain selected for each factor under study (13 runs). Pretreatment assays were performed in random order. Temperature (in the range 185–215 °C) and time (2.5–7.5 min) were selected as variables. Center values and intervals were selected based on previous experience. The coded and uncoded values of factors in the rotatable central composite design are shown in Table 1. The experimental data were analyzed via response surface methodology by the statistical software DesignExpert 8.0.7.1, Stat-Ease Inc., Minneapolis, USA. 2.4. Enzymatic hydrolysis tests The washed WIS fractions of rapeseed straw obtained after pretreatment were used as substrates for enzymatic hydrolysis assays. These WIS fractions were hydrolyzed with a cellulolytic complex (Celluclast 1.5 L) kindly provided by Novozymes A/S (Denmark), which contains an activity of 69 filter paper units (FPU)/mL. The cellulase enzyme loading was 15 FPU/g substrate. To supplement b-glucosidase activity of cellulases, fungal b-glucosidase (Novozym 188, Novozymes A/S), with an activity of 530 international units (IU)/mL was added, at an enzyme loading of 15 IU/g substrate. The pH was adjusted to 4.8 with 0.05 M sodium citrate buffer and enzymes were added to the pretreated substrate (5% w/v dry basis) for a total working volume of 20 mL in 100 mL Erlenmeyer flasks. Triplicate reaction flasks were incubated at 50 °C in an orbital shaker (Certomat-R, B-Braun, Germany) at 150 rpm for 72 h. Two milliliter samples were withdrawn at 24, 48 and 72 h, and they were centrifuged at 10,000g (Sigma 1–14 Centrifuge) for 10 min. Glucose and xylose concentrations in the sample supernatant were determined by HPLC. Once the steam explosion conditions were optimized to maximize the enzymatic hydrolysis yield, different substrate and enzyme loadings were tested. Then, the WIS fraction obtained under optimal conditions was washed and submitted to enzymatic hydrolysis tests as described above. In order to evaluate the performance of enzymatic hydrolysis tests, two parameters were determinate as follows. Saccharification efficiency (SE) was calculated as the ratio of grams of glucose released by enzymatic hydrolysis per 100 g glucose in the WIS, and the enzymatic hydrolysis yield (YEH) was referred to the glucose content in the raw material. All enzymatic hydrolysis experiments were performed in triplicate. Average results and standard deviations are given. 2.5. Microorganism, medium and yeast cultivation Saccharomyces cerevisiae (Fermentis Ethanol Red, Cedex, France) was used for fermentation assays. Yeast was maintained on solid Table 1 Operational conditions assayed expressed as dimensional and dimensionless independent variables. Run
1 2 3 4 5 6 7 8 9 10 11 12 13
Temperature (°C)
Time (min)
Real value
Coded value
Real value
Coded value
221.21 200.00 185.00 200.00 215.00 200.00 185.00 215.00 200.00 200.00 200.00 200.00 178.79
+1.414 0 1 0 1 0 1 1 0 0 0 0 1.414
5.00 8.54 2.50 5.00 7.50 5.00 7.50 2.50 5.00 5.00 5.00 1.46 5.00
0 +1.414 1 0 1 0 1 1 0 0 0 1.414 0
J.C. López-Linares et al. / Bioresource Technology 190 (2015) 97–105
culture media (g/L): glucose, 20; yeast extract, 5; peptone, 5; NH4Cl, 2 g/L; KH2PO4, 1, MgSO47H2O, 0,5; and agar 20, at 4 °C. Inoculum was prepared by transferring the yeast maintained on solid culture medium into 25 mL of the growth media consisting of (g/L): yeast extract, 5; ClNH4, 2; KH2PO4, 1; MgSO47H2O, 0.3; and glucose, 30. Cells were grown in 100 mL Erlenmeyer flasks with a volume of 25 mL and agitated on a rotatory shaker (Certomat-R, B-Braun, Germany) at 150 rpm and 30 °C for 16 h. Inoculum (4% v/v) yeast cultures correspond to a cell addition of 0.25 g/L. 2.6. Simultaneous saccharification and fermentation (SSF) Saccharification and fermentation assays were carried out using WIS fraction obtained under optimal conditions. The SSF experiments were performed under microaerobic conditions in 100 mL Erlenmeyer flasks with a total volume of 20 mL in citrate buffer (0.05 M, pH 4.8) containing the nutrients described above for inoculum cultivation except for glucose. Substrate concentration was 20% and 30% (w/v). SSF was started by adding simultaneously enzymes and S. cerevisiae inoculum (4% v/v) yeast cultures, corresponding to a cell addition of 0.25 g/L. The commercial enzymes used were the same as in the enzymatic hydrolysis tests, with cellulase enzyme loading of 20 and 40 FPU/g substrate. The assays were carried out during 72 h at 37 °C and 200 rpm and compared to controls that contained the same cellulose loading (Sigmacell). The flasks were equipped with a thick rubber stopper, through which one stainless-steel needle had been inserted to permit evolved CO2 to leave and maintain microaerobic conditions. All SSF experiments were performed in triplicate and the average results are given. Samples were withdrawn at 24, 48 and 72 h, centrifuged at 10,000g for 10 min for the determination of produced ethanol and consumed sugars. SSF results are reported as percentage of the theoretical yield considering that all the potential glucose in the pretreated solid is available for fermentation, and a theoretical fermentation yield of 0.51 g ethanol/g glucose. 2.7. Prehydrolysis and simultaneous saccharification and fermentation (PSSF) A previous hydrolysis for 6 and 24 h at 50 °C was carried out in triplicate before inoculation using the pretreated solid obtained under optimal conditions as substrate. The pretreated material concentration used in these assays was 30% (w/v). The enzyme loadings were the same as in the SSF test. Afterwards, the temperature was decreased to 37 °C and the flasks were inoculated with harvested yeast at the same concentration as in the SSF process. The flasks were incubated in a rotatory shaker at 200 rpm during 72 h (including the presaccharification time). Samples were taken and analyzed for carbohydrates and ethanol concentrations at 24, 48 and 72 h of process. All PSSF experiments were performed in triplicate. Average results are given. 2.8. Analytical methods The composition of the raw material and the solid fractions obtained after steam explosion pretreatment were determined according to NREL methodology as described in López-Linares et al. (2014). Sugar concentration was determined by high performance liquid chromatography (HPLC) in a Waters 2695 liquid chromatograph with a refractive index detector. A CARBOSep CHO-682 LEAD column (Transgenomic, Omaha, NE) operating at 80 °C with Milli-Q water (Millipore) as mobile-phase (0.5 mL/ min) was used.
99
Furfural and HMF were analyzed by HPLC (Agilent, Waldbronn, Germany) employing an Aminex HPX-87H column (Bio-Rad Labs, Hercules, CA) at 65 °C. The mobile phase contained 89% (5 mM H2SO4) and 11% acetonitrile at flow rate of 0.7 mL. A 1050A Photodiode-Array detector (Agilent, Walsbronn, Germany) was employed for detection. Formic and acetic acid were quantified by HPLC (Waters, Milfors, MA) using a 410 Water refractive index detector. An Aminex HPX-87H (Bio-Rad Labs, Hercules, CA) column maintained at 65 °C and a mobile phase of 5 mM H2SO4 at a flow rate of 0.6 mL/min were employed. Phenolic compounds were quantified in the liquid fractions a slightly modified version of Folin–Ciocalteu method (Zhang et al., 2006). Twenty microliter of sample and the serial standard solution were diluted with 88 lL of water on a 96-well microplate. After the addition of 12 lL Folin–Ciocalteu reagent, the plate was incubated for 5 min at room temperature in dark conditions. The reaction was stopped with 80 lL of 7.5% sodium carbonate solution. Before reading, the plate was incubated for 2 h at room temperature in the dark. The absorbance was measured at 765 nm with a spectrophotometric microplate reader (Anthos Zenyth 200rt, Biochrom, UK). Glucose and xylose concentrations from EH samples was measured by HPLC with the above described Waters equipment. Glucose and ethanol concentrations from SSF and PSSF samples were also measured by HPLC, using Bio-Rad HPX-87H column. All analytical determinations were performed in triplicate and the average results are shown. Relative standard deviations were below 3%. 3. Results 3.1. Characterization of rapeseed straw The initial composition of rapeseed straw was determined as 49% total carbohydrates (31.6 ± 0.3% cellulose and 17.4 ± 0.1% hemicellulose), 17.8 ± 0.5 lignin, 3.4 ± 0.1 acetyl groups, 15.4 ± 1.3 extractives, and 6.7 ± 0.3 ashes. These results are mean values of three determinations. The hemicellulose consisted of 76% xylose. By comparing with the composition of rapeseed straw analyzed by other researchers (Karagöz et al., 2012; Garmakhany et al., 2014), the raw material used in this research had less carbohydrates and a higher content of other minor fractions like acetyl groups, extractives and ash. According to Godin et al. (2013), the individual composition for each type of biomass is relatively constant regardless of the crop conditions (year, area, cultivar, nitrogen fertilization level) although the biomass chemical components are mainly influenced by the crop dry matter yield. A significant amount of non-structural compounds, about 15%, was determined in the rapeseed straw as extractives, mainly water soluble materials. 3.2. Effect of steam explosion conditions on WIS and prehydrolysates The solid recovery yield, as based on the initial biomass weight, ranged between 55%, at the highest temperatures (runs 1, 8), and 74%, at the mildest conditions (run 3), Table 2. Steam explosion pretreatment of rapeseed straw meant a cellulose enrichment in WIS due to the removal of extractives and partial solubilisation of hemicellulose fraction since in all conditions tested, a part of this fraction remained in the solid fraction. Similarly, the lignin content in WIS increased, which can be attributed to condensation reactions between lignin, extractives and sugars degradation products according to Ballesteros et al. (2011). Thus, WIS with the highest cellulose content, about 50%, were obtained at 200 °C and 215 °C (runs 2, 5), since at those conditions high solubilisation of
100
J.C. López-Linares et al. / Bioresource Technology 190 (2015) 97–105
Table 2 Solid recovery and WIS composition. Run
Solid recovery (%)
1 2 3 4 5 6 7 8 9 10 11 12 13
54.8 59.8 74.0 62.7 59.4 67.5 73.2 54.5 70.4 67.3 61.7 69.1 72.0
Table 3 Sugar and inhibitor composition (g/100 g raw material) of prehydrolysates.
WIS composition (%) Cellulose
Hemicellulose
Lignin
47.3 ± 1.16 50.3 ± 0.40 40.7 ± 0.55 44.1 ± 0.37 49.9 ± 0.18 42.9 ± 0.64 41.3 ± 0.47 48.3 ± 0.26 42.3 ± 0.34 44.9 ± 0.74 46.5 ± 0.70 44.2 ± 1.00 43.4 ± 0.95
1.8 ± 0.11 5.8 ± 0.07 18.5 ± 0.23 7.2 ± 0.15 1.8 ± 0.15 8.6 ± 0.15 14.0 ± 0.02 4.6 ± 0.02 8.5 ± 0.21 8.7 ± 0.06 9.4 ± 0.05 14.8 ± 0.21 19.6 ± 0.56
44.2 ± 0.30 40.8 ± 0.28 29.5 ± 0.33 35.7 ± 0.18 47.8 ± 1.12 33.9 ± 2.89 32.6 ± 0.47 40.4 ± 0.39 39.5 ± 2.49 37.6 ± 0.78 38.7 ± 0.10 31.9 ± 0.51 29.5 ± 0.71
hemicellulosic sugars was reached, yielding WIS with a hemicellulosic content below 6%. However, at 221.21 °C (star point), a mild decrease in cellulose content was detected, which can be attributed to some degradation of the cellulose fraction by severe conditions of pretreatment. At 185 °C (runs 3 and 7), regardless of the pretreatment time, the cellulose content was about 41%, while the hemicellulose content was over than 14%, which can be a drawback for the subsequent enzymatic hydrolysis since hemicellulose abundance can hinder enzymatic access to the cellulose (Ballesteros et al., 2006). In the prehydrolysates obtained after steam explosion pretreatment, sugars were present mainly in oligomeric form, between 60% and 90%, so that a posthydrolysis step was performed to determine the total amount of sugars. Hemicellulose was partially hydrolyzed resulting in hexoses (glucose, galactose and mannose), pentoses (xylose and arabinose) and a variety of oligosaccharides. As can be observed in Table 3, low recoveries of hemicellulosic sugars in prehydrolysates were attained, yielding sugar solutions with maxima concentrations about 20 g/L. This behavior indicates high degradation of these sugars at the pretreatment conditions assayed. As a consequence, the prehydrolysates contained also non-sugars compounds, as shown in Table 3. In general, inhibitors concentration increased when the severity of pretreatment did. The pretreatment usually degrades the hemicellulose, leading to the formation of products such as aliphatic acids (primarily acetic acid and formic acid), and furan aldehydes (5-hydroxymethylfurfural and furfural). Similarly to hemicellulose, lignin is also partially degraded to the phenolic compounds. Phenolics derived from lignin and other plant components can create significant inhibition on enzymatic conversion of cellulosic biomass materials to useful chemicals (Tejirian and Xu, 2011). Furthermore, the degradation of cellulose can also lead to the formation of byproducts such as HMF (Palmqvist and Hahn-Hägerdal, 2000). The most abundant inhibitory compounds detected in the prehydrolysates were acetic and formic acids. High acetic acid concentrations can be attributed to the high content of acetyl groups in the raw rapeseed straw (3.4%), greater than that in other raw materials such as barley straw (Duque et al., 2013) or olive tree pruning (LópezLinares et al., 2013). The highest acetic acid concentrations, above 3 g/100 g, were achieved at maximum temperature, which can be attributed to a greater hemicellulose solubilisation, corresponding to WIS with the lowest hemicellulose content (runs 1, 5) as can be seen in Table 2. The presence of formic acid in the prehydrolysates can be related to a high degradation of hemicellulosic sugars, since the highest concentrations of this inhibitor compound (more than 2 g/100 g rapeseed straw) were achieved in runs 1 and 5, coinciding with xylose recoveries lower than 0.6 g/100 g rapeseed straw (Table 3). In addition, the high concentrations of formic acid detected can be also related to the low concentrations of furfural
Run
Glucose
Xylose
Galactose
Arabinose
Mannose
1 2 3 4 5 6 7 8 9 10 11 12 13
0.28 ± 0.02 0.42 ± 0.09 0.83 ± 0.15 0.58 ± 0.09 0.19 ± 0.01 0.38 ± 0.03 0.72 ± 0.11 0.57 ± 0.10 0.54 ± 0.01 0.60 ± 0.07 0.57 ± 0.08 0.72 ± 0.04 0.85 ± 0.10
0.62 ± 0.09 2.42 ± 0.21 0.89 ± 0.13 4.10 ± 0.11 0.52 ± 0.08 1.72 ± 0.19 2.53 ± 0.14 3.19 ± 0.26 2.51 ± 0.12 3.41 ± 0.13 2.58 ± 0.25 2.20 ± 0.07 0.79 ± 0.05
0.13 ± 0.01 0.32 ± 0.03 0.92 ± 0.02 0.61 ± 0.05 0.06 ± 0.01 0.36 ± 0.07 0.83 ± 0.11 0.47 ± 0.03 0.46 ± 0.10 0.60 ± 0.13 0.44 ± 0.07 0.73 ± 0.12 0.86 ± 0.06
0.04 ± 0.00 0.12 ± 0.01 0.90 ± 0.10 0.22 ± 0.02 0.03 ± 0.00 0.17 ± 0.01 0.53 ± 0.05 0.18 ± 0.02 0.25 ± 0.04 0.30 ± 0.06 0.17 ± 0.01 0.67 ± 0.03 0.85 ± 0.11
0.14 ± 0.03 0.23 ± 0.04 0.21 ± 0.03 0.24 ± 0.01 0.00 ± 0.00 0.17 ± 0.01 0.19 ± 0.01 0.29 ± 0.07 0.18 ± 0.05 0.25 ± 0.07 0.11 ± 0.01 0.19 ± 0.03 0.18 ± 0.02
Acetic acid
Formic acid
Furfural
HMF
Total phenols
1 2 3 4 5 6 7 8 9 10 11 12 13
3.23 ± 0.17 2.87 ± 0.11 0.63 ± 0.09 1.71 ± 0.12 3.40 ± 0.19 1.81 ± 0.11 1.10 ± 0.13 2.58 ± 0.15 2.06 ± 0.13 2.06 ± 0.07 2.09 ± 0.22 0.96 ± 0.09 0.63 ± 0.08
2.28 ± 0.12 2.07 ± 0.18 0.81 ± 0.07 1.36 ± 0.05 2.18 ± 0.19 1.39 ± 0.11 1.05 ± 0.13 1.96 ± 0.05 1.70 ± 0.06 1.66 ± 0.04 1.59 ± 0.09 1.04 ± 0.11 0.73 ± 0.03
0.22 ± 0.03 0.18 ± 0.02 0.01 ± 0.00 0.10 ± 0.01 0.23 ± 0.02 0.11 ± 0.01 0.04 ± 0.00 0.19 ± 0.03 0.10 ± 0.01 0.10 ± 0.00 0.13 ± 0.01 0.03 ± 0.00 0.01 ± 0.00
0.10 ± 0.01 0.08 ± 0.02 0.05 ± 0.01 0.07 ± 0.01 0.09 ± 0.02 0.06 ± 0.00 0.05 ± 0.00 0.08 ± 0.01 0.06 ± 0.01 0.06 ± 0.02 0.06 ± 0.01 0.05 ± 0.00 0.04 ± 0.01
2.77 ± 0.11 2.38 ± 0.05 1.65 ± 0.09 2.05 ± 0.15 2.06 ± 0.13 1.63 ± 0.12 1.54 ± 0.08 2.38 ± 0.20 2.06 ± 0.17 1.92 ± 0.16 2.06 ± 0.13 1.79 ± 0.11 1.55 ± 0.08
and HMF in these prehydrolysates since the formation of formic acid has been reported due to the degradation of furfural and HMF (Palmqvist and Hahn-Hägerdal, 2000). However, these inhibitor concentrations are slightly lower than those obtained by Wood et al. (2014) at similar conditions of steam explosion pretreatment, obtaining also low concentrations of sugars in the liquid fractions with the same feedstock. It is noteworthy that in all conditions assayed, these sugars were present mainly in oligomeric form, between 63% and 92%. The pH of prehydrolysates was about 4 and no significant differences were detected at the different conditions tested. 3.3. Evaluation of steam explosion at different operational conditions The effects of the steam explosion pretreatment conditions on the glucose recovery in WIS (GRWIS), overall glucose recovery (GRoverall), saccharification efficiency (SE) and enzymatic hydrolysis yield (YEH) were investigated using the response surface methodology. The response GRWIS is expressed as percentage of glucose remained in WIS referred to glucose content in raw material. Overall glucose recovery, GRoverall, is expressed as percentage of glucose recovered in both solid and liquid streams referred to glucose content in raw material. The responses SE and YEH are expressed as the percentage of the glucose content in WIS or in raw material which is released by enzymatic hydrolysis. The regression equations in terms of normalized values are shown in Eqs. (1)–(4).
GRWIS ð%Þ ¼ 93:10 5:64T þ 1:23t 1:63T 2 ðR2 ¼ 0:915; R2 adjust ¼ 0:873Þ GRoverall ð%Þ ¼ 94:79 5:88T þ 0:92t 1:91T 2 ðR2 ¼ 0:930; R2 adjust ¼ 0:895Þ SE ð%Þ ¼ 61:98 þ 22:24T þ 4:90t ðR2 ¼ 0:929; R2 adjust ¼ 0:913Þ
ð1Þ
ð2Þ
ð3Þ
101
J.C. López-Linares et al. / Bioresource Technology 190 (2015) 97–105
YEH ð%Þ ¼ 59:33 þ 17:35T þ 4:50t 2:67T 2 2
ð4Þ
2
ðR ¼ 0:988; R adjust ¼ 0:982Þ
The values of R2 and R2-adjust in each equation show that good agreement between experimental and predicted data was achieved. It can be observed from the coefficients of the equations that the effect of the temperature was more important than the effect of the pretreatment time. Moreover, no response was affected by the interaction between factors. Regarding glucose recovery in solid fraction (GRWIS) and overall glucose recovery (GRoverall) the effect of temperature was negative while the pretreatment time exerted a positive effect although less significant (Eqs. (1) and (2)). The highest GRWIS was obtained in run 13 (178.79 °C, 5 min) and on the contrary, the lowest glucose recovery in solid fraction was found at the highest temperature (run 1, 221.21 °C, 5 min), Table 4. This fact can be attributed to some degradation of cellulose fraction at these severe conditions. However, it is worth noting that the GRWIS was over 82% in all conditions assayed (Table 4), indicating low levels of degradation and therefore remaining solids rich in cellulose after steam explosion pretreatment with negligible loss of glucose. Figs. 1 and 2 illustrate
Table 4 Glucose recovery in pretreated solids and results of enzymatic hydrolysis. Run
Glucose recovery in WIS (%)
Overall glucose recovery (%)
Saccharification efficiency (%)
Enzymatic hydrolysis yield (%)
1 2 3 4 5 6 7 8 9 10 11 12 13
82.2 ± 2.03 95.5 ± 0.76 95.6 ± 1.29 87.9 ± 0.73 94.1 ± 0.33 92.0 ± 1.38 96.1 ± 1.09 83.6 ± 0.45 94.5 ± 0.75 95.9 ± 1.57 91.1 ± 1.37 97.1 ± 2.20 99.3 ± 2.16
83.0 ± 2.11 96.7 ± 0.78 98.0 ± 1.30 89.6 ± 0.76 94.7 ± 0.32 93.1 ± 1.39 98.2 ± 1.15 85.3 ± 0.44 96.0 ± 0.77 97.7 ± 1.62 92.8 ± 1.40 99.2 ± 2.28 100.0 ± 2.15
95.7 ± 2.63 71.1 ± 2.24 34.9 ± 2.03 69.9 ± 0.59 84.6 ± 0.94 61.9 ± 1.21 42.9 ± 1.95 83.8 ± 0.42 69.0 ± 4.31 62.5 ± 1.30 56.6 ± 2.25 45.6 ± 1.18 32.1 ± 3.95
78.6 ± 2.16 67.9 ± 2.14 33.3 ± 1.94 61.5 ± 0.52 79.6 ± 0.88 57.0 ± 1.11 41.2 ± 1.88 70.1 ± 0.35 65.2 ± 4.08 60.0 ± 1.25 51.6 ± 2.05 44.3 ± 1.14 31.9 ± 3.92
the influence of temperature and process time on GRWIS and GRoverall. As can be observed in these figures, the behavior of both GRWIS and GRoverall was similar/the same since almost all glucose was recovered in pretreated solids while the fraction of glucose solubilised was minimal. It can be observed that an increase in temperature limited the glucose recovery in both fractions. Concerning SE and YEH (Eqs. (3) and (4)), both the temperature and the time had a significant positive effect in these responses. However, the process time did not affect hydrolysis process significantly as can also be observed in the response surface for saccharification efficiency (Fig. 3) and enzymatic hydrolysis yield (Fig. 4). Both figures show that an increase in the temperature implies a substantial increase in the enzymatic hydrolysis yields, referred to glucose content in WIS (SE) or in raw material (YEH), respectively. Quadratic effect on YEH (Eq. (4)) indicates that the values of the temperature stabilized, i.e., there is a point from which the enzymatic hydrolysis yield did not increase due to that factor. This behavior can be also observed in Fig. 4 where a light curvature indicates that there is a point from which the enzymatic hydrolysis yield does not increase due to the temperature. At the lowest temperatures tested (178.79 °C and 185 °C), regardless of the process time, the enzymatic hydrolysis yields (YEH) were lower than 42%. These temperatures, although yielded pretreated solids with the highest glucose recoveries, (GRWIS), above 95% (Table 4), and higher hemicellulose content, above 14% (Table 2), did not achieve to alter the lignocellulosic structure enough to make the enzymatic attack easier. It can be observed in Table 4 that the lowest cellulose conversion corresponds to the assay carried out at the lowest temperature (178.79 °C, run 13) while the maximum value of SE (95.7%) was achieved in the experiment carried out at 221.21 °C (run 1). However, the enzymatic hydrolysis yield (YEH) attained at this temperature (78.6%) was slightly lower than that achieved at 215 °C and 7.5 min (run 5). It means that cellulose digestibility increased as temperature did, regardless of the pretreatment time, although at the maximum temperature assayed (star point) some degradation of cellulose fraction occurred, reducing the enzymatic hydrolysis yield referred to glucose content in raw rapeseed straw (YEH).
101
Overall glucose recovery (%)
102
GRWIS (%)
95 89 83 77
1.00
175.00
3.25
Tim
187.50 5.50
e( mi n)
200.00 7.75
212.50 10.00 225.00
Tem
a per
tu r
e(
ºC)
Fig. 1. Response surface and contour plot of temperature and pretreatment time on glucose recovery in WIS.
96 90 84 78 175.00
1.00 187.50
3.25
Tim e
200.00
5.50
(m in)
212.50
7.75 10.00
225.00
Te
m
pe
ra
t
e ur
(º
C
)
Fig. 2. Response surface and contour plot of temperature and pretreatment time on overall glucose recovery.
102
J.C. López-Linares et al. / Bioresource Technology 190 (2015) 97–105 Table 5 Results obtained after 72 h of enzymatic hydrolysis at different solid loadings and enzyme dosages using as substrate WIS under optimal steam explosion conditions.
Saccharification efficie ncy (%)
109
86
63
40
Enzyme loading (FPU/ g WIS)
Glucose concentration (g/L)
SE (%)
YEH (%)
5 20 20 30 30
15 20 40 20 40
22.0 ± 0.45 82.0 ± 1.22 85.9 ± 0.68 108.4 ± 0.09 114.0 ± 0.29
80.9 75.3 78.9 66.3 69.8
72.3 67.3 70.5 59.3 62.4
SE (%): grams of glucose released by enzymatic hydrolysis per 100 g glucose present in the WIS. YEH (%): grams of glucose released by enzymatic hydrolysis per 100 g glucose present in the raw material.
17
10.00
225.00 7.75
212.50
Tim 5.50 e( mi n)
200.00 3.25
187.50 1.00
175.00
re atu per m Te
) (ºC
Fig. 3. Response surface and contour plot of temperature vs. pretreatment time on saccharification efficiency.
90
71.25
YEH (%)
Solids loading (%)
52.5
33.75
15
10.00
225.00 7.75
212.50 5.50
Tim e (m in)
200.00 3.25
187.50 1.00
175.00
Te
t era mp
( ure
ºC)
Fig. 4. Response surface and contour plot of temperature vs. process time on enzymatic hydrolysis yield.
3.4. Model optimization The mathematical model that was developed from the experimental results is able to predict the operational conditions that should be used in steam explosion pretreatment to optimize model responses. In this study, the optimization focused in the maximization of YEH and the maximum yield is predicted to be reached if pretreatment is performed at 215 °C for 7.5 min. These conditions coincide with the experimental run 5, yielding the highest value of YEH (79.6%). To validate the model, a new experiment was carried out (by triplicate), setting the process variables to the optimum ones given above. Enzymatic hydrolysis yield (YEH) at optimum conditions was 72.3%, equivalent to 81% of potential glucose in WIS. This result is slightly lower than that predicted by the model, YEH (78%) although within the limits of variability in confidence level of 95%.
3.5. Enzymatic hydrolysis of WIS obtained under optimal pretreatment conditions Pretreated rapeseed straw obtained under optimal steam explosion conditions (215 °C, 7.5 min) was used as substrate for enzymatic hydrolysis tests at high solids loading (20% and 30% w/v). Moreover, in order to assess the influence of enzyme dosage, 20 and 40 FPU/g WIS were tested at both 20% and 30% (w/v) substrate concentration. Table 5 shows conditions and results from these experiments. For comparative purposes, a standard assay at 5% (w/v) solids loading and 15 FPU/g WIS was carried out. All enzymatic hydrolysis experiments were performed in triplicate and average results are given. An increase in the solids loading limited the saccharification efficiency, probably due to enzyme inhibition exerted by high glucose concentration or by mixing difficulties (Koppram et al., 2014) although some authors suggest that the limitations due to mass diffusion are more prevalent than the sugar inhibition (Modenbach and Nokes, 2013). The efficiency saccharification decreased from 75.3% to 66.3% as solids loading increased from 20% to 30% (w/v) at 20 FPU/g WIS. Using the highest dose of enzyme, similar reductions in the conversion of cellulose (by 10%) are obtained with increasing solids loading. However, it can be essential the use of high solids loading during enzymatic hydrolysis to achieve concentrated glucose solutions resulting in high ethanol concentrations in the subsequent distillation step (Larsen et al., 2008). Glucose concentrations of 82 g/L and 85.9 g/ L were achieved after 72 h of enzymatic hydrolysis for 20% (w/v) solids loading at 20 and 40 FPU/g WIS, respectively (Table 5). The system containing 30% solids achieved maxima glucose concentrations of 108.4 g/L and 114 g/L at enzyme dosage of 20 and 40 FPU/g WIS, respectively. It can be observed for both enzyme loadings that glucose concentrations increased although not proportionally with substrate loading. Thus, the highest values of SE and YEH were achieved when the enzymatic hydrolysis was performed at standard conditions, 5% (w/v) solids loading and 15 FPU/g WIS. At these conditions, enzymatic hydrolysis yielded 81 g of glucose per 100 g of glucose content in WIS and 72 g of glucose per 100 g glucose content in raw rapeseed straw. Under these conditions product inhibition or mixing problems are not expected. A maximum saccharification efficiency and enzymatic hydrolysis yield can be obtained. However, glucose concentration was as low as 22 g/L. Nevertheless, to achieve the required ethanol concentrations for efficient distillation (4.4% w/w), substrate concentration in EH must be at least 20% (w/w) which can release glucose solutions of 8% w/w (Larsen et al., 2008). In this research, this assumption holds for solids loading of 20% (w/v) and enzyme dosage of 20 FPU/g WIS, yielding a glucose solution of 8% (w/w). Considering the obtained results, it is noteworthy that the rapeseed straw pretreated by steam explosion is a good substrate than
103
J.C. López-Linares et al. / Bioresource Technology 190 (2015) 97–105
Concentration (g/L)
a
Time (h)
b Concentration (g/L)
can be enzymatically hydrolyzed at high solids loading with relatively low drop in the yield. High enzymatic hydrolysis yields have been achieved by other researchers using steam explosion (Wood et al., 2014) with rapeseed straw although only at 1% (w/w) solids loading. High enzyme loadings can improve enzymatic hydrolysis yield at high solids loading (Xue et al., 2012). However, in this research this improvement was not detected, maybe due to both solid loadings used (20% and 30% w/v) can be considered high substrate loadings. As can be seen in Table 5, the increase in enzyme dosage from 20 to 40 FPU/g WIS increased the saccharification efficiency by only 3.5% regardless of the substrate loading. According to Olofsson et al. (2008), an increase of 50% in enzyme loading should be justified if an increase of yields greater than 6% is achieved. Therefore, it may not be worthy converting an extra 3.5% of glucose since the enzyme cost has been recognized as a considerable contributor to bioethanol cost (Wang et al., 2012). According to Modenbach and Nokes (2013), a sequential addition of enzymes can be advantageous because the addition of fresh enzyme can replace the non-productive enzyme bound to the lignin. This feeding strategy was tested for exploded rapeseed straw at optimal conditions, to improve the enzymatic hydrolysis performance (data not shown), although it did not achieve an improvement in hydrolysis yields compared to the assays carried out adding complete enzyme dosage at the beginning of the process. Therefore, the relatively low enzymatic hydrolysis yields are not due to enzyme adsorption to lignin, and rather must be due to mass transfer limitations.
3.6. SSF and PSSF for pretreated solid obtained under optimal conditions Exploded rapeseed straw at optimal conditions was submitted to an SSF process at the same solids and enzyme loadings tested for EH assays as described in Section 3.5. At 20% solids loading, complete conversion of glucose to ethanol was realized within 48 h for both enzyme loadings assayed (Fig. 5a) without any significant difference in time course of glucose and ethanol concentrations during SSF process. Final ethanol concentrations of 42.8 g/L and 43.6 g/L were attained at 20 and 40 FPU/g WIS, respectively. It should be noted that the main ethanol production occurs at the beginning of the SSF process. Thus, at lower solid loading tested, in the 0–24 h period of process, there was a production about 34 g/L of ethanol, while in the 48–72 h period, the increase was only 1.5 g/L (Fig. 5a). According to this behavior, the process time could be shortened from 72 to 48 h. However, as can be seen in Fig. 5b, when substrate loading increased to 30% (w/v), fermentation resulted in incomplete glucose consumption yielding ethanol concentrations of 54.9 g/L and 50.5 g/L at 20 and 40 FPU/g WIS, respectively. After 72 h, 13.5 g/L and 19.2 g/L of glucose remained in the broth/medium as residual glucose with 20 and 40 FPU/g WIS, respectively, and no additional increases in ethanol concentration were detected after this time (data not shown), indicating the final of the fermentation process. Thus, fermentation performance was the limiting factor in SSF process since the yeast ceased after that time. Similarly as it occurs in EH, the influence of enzyme dosage in the SSF process, in the range tested, was minimal although at 30% solids, glucose conversion decreased with increasing enzyme loading. This fact can be related to an inefficient mixing of the enzyme with the substrate as the solids loading increased. Furthermore, by employing high loadings of solids (30% w/v) and high enzyme dosage (40 FPU/g WIS), some compounds could be released in the fermentation medium affecting the microorganism. Zhao et al. (2013) have reported that high cellulase loadings might
Time (h) Fig. 5. Time course of glucose (dashed lines) and ethanol (continuous lines) concentrations during SSF at 20 FPU/g WIS (d) and 40 FPU/g WIS (j) (a) for 20% (w/ v) and (b) 30% (w/v) substrate concentration.
cause inhibition on S. cerevisiae due to the presence of some additives in the solution of the commercial enzymatic complex. Table 6 shows the results attained in the different SSF tests. The ethanol yield was expressed as a percentage of the maximum ethanol yield (0.51 g ethanol/g glucose). And the process yield is determined as the ratio of the grams of ethanol after 72 h fermentation per 100 g dry straw. SSF at the high 30% (w/v) solids concentration resulted in less efficient process than at 20% (w/v) solids, probably because of
Table 6 Ethanol yields attained by SSF and PSSF at different conditions. Solids loading (%)
SSF 20 30 PSSF (presac. time) 30 (6 h) 30 (24 h)
Enzymatic loading (FPU/g WIS)
Ethanol concentration (g/L)
YE (%)
Process yield (%)
20 40 20 40
42.8 ± 0.1 43.6 ± 0.0 54.9 ± 0.5 50.5 ± 0.7
77.0 78.5 65.8 60.6
12.2 12.4 10.4 9.6
20
56.2 ± 0.3 52.5 ± 0.1
67.5 63.0
10.7 9.9
YE: expressed as percentage of theoretical ethanol yield (0.51 g ethanol/g glucose). Process yield: g ethanol/100 g rapeseed straw.
104
J.C. López-Linares et al. / Bioresource Technology 190 (2015) 97–105
limitations associated with laboratory-scale mixing and poor performance of the yeast. This behavior of the microorganism can be attributed to the combined effect of high stress by reducing the amount of water available and the hyperosmotic conditions in the fermentation broth (Koppram et al., 2014). The system containing 30% solids achieved the highest final ethanol concentration with nearly 55 g/L. However, the conversion decreased with increasing solids loadings, although the conversion of the 30% solids reaction was only 10% below systems at 20% solids (33.6% vs. 39.3% and 30.9% vs.40%, at 20 and 40 FPU/g WIS, respectively), Table 6. Comparing to other pretreatment options for rapeseed straw, steam explosion processing performed in this work followed to an SSF process enabled similar ethanol yields and higher ethanol concentrations than those achieved by liquid hot water followed also by an SSF process at 7.5% w/v (López-Linares et al., 2014). In order to evaluate the effect of a previous hydrolysis on the ethanol yield, prehydrolysis for 6 and 24 h followed by SSF was performed at 30% (w/v) solids and considering the results obtained in the SSF tests, the enzyme loading used in PSSF experiments was only 20 FPU/g WIS (Table 6). Two experiments were conducted to assess the timing of hydrolysis for better yield. Only the experiment carried out with 6 h of prehydrolysis yielded better results of ethanol production than the SSF process (56.2 g/L vs. 54.9 g/L) although the increase in ethanol yield was lower than 1%. Therefore, by comparing SSF and PSSF processes there was no difference between both configurations with regard to ethanol yield. Arvaniti et al. (2012) achieved also an ethanol yield of 67% (referred to the theoretical ethanol yield) from rapeseed straw pretreated by wet oxidation using a presaccharification followed by an SSF process at 12.5% (w/w) solids loading. The highest ethanol yield, i.e., 12.4 g ethanol/100 g biomass, corresponds to 70% of the maximum theoretical yield. This result was obtained from an SSF process at 20% (w/v) solids loading and 40 FPU/g WIS, resulting in an ethanol solution of 43.6 g/L (5.5% v/v). This result indicates a very high efficiency in the use of the cellulosic fraction of the rapeseed straw. Lu et al. (2009) and Luo et al. (2011) with the same feedstock achieved higher process yields (14% and 15% referred to 100 g of biomass) by an sequential process, although they used solid loadings lower than those used in this study, only 10% and 5% (w/w), respectively. 4. Conclusions From the results obtained in this research, it can be concluded that the uncatalyzed steam explosion is an effective pretreatment to increase the enzymatic digestibility of rapeseed straw. WIS obtained under optimal steam explosion conditions, 215 °C and 7.5 min, resulted in a good substrate for bioethanol production at high substrate concentration. A configuration of simultaneous saccharification and fermentation, at 20% (w/v) solid loading, allowed obtaining an ethanol solution of 5.3%, by volume, which corresponds to 77% of the theoretical ethanol yield. Further research will be focused on improving hemicellulosic sugar recoveries to increase ethanol production. Acknowledgement J.C. López-Linares gratefully acknowledges the research Grant received from University of Jaén (Plan de Apoyo). References Alvira, P., Tomás-Pejó, E., Ballesteros, M., Negro, M.J., 2010. Pretreatment technologies for an efficient bio ethanol production process based on enzymatic hydrolysis: a review. Bioresour. Technol. 101, 4851–4861.
Arvaniti, E., Bjerre, A.B., Schmidt, J.E., 2012. Wet oxidation pretreatment of rape straw for ethanol production. Biomass Bioenergy 39, 94–105. Ballesteros, I., Negro, M.J., Oliva, J.M., Cabañas, A., Manzanares, P., Ballesteros, M., 2006. Ethanol production from steam-explosion pretreated wheat straw. Appl. Biochem. Biotechnol. 130, 496–508. Ballesteros, I., Ballesteros, M., Cara, C., Sáez, F., Castro, E., Manzanares, P., Negro, M.J., Oliva, J.M., 2011. Effect of water extraction on sugars recovery from steam exploded olive tree pruning. Bioresour. Technol. 102, 6611–6616. Duque, A., Manzanares, P., Ballesteros, I., Negro, M.J., Oliva, J.M., Saez, F., Ballesteros, M., 2013. Optimization of integrated alkaline–extrusion pretreatment of barley straw for sugar production by enzymatic hydrolysis. Process Biochem. 48, 775– 781. FAOSTAT, 2015.
J.C. López-Linares et al. / Bioresource Technology 190 (2015) 97–105 Wood, I.P., Elliston, A., Collins, S.R.A., Wilson, D., Bancroft, I., Waldron, K.W., 2014. Steam explosion of oilseed rape straw: establishing key determinants of saccharification efficiency. Bioresour. Technol. 162, 175–183. Xue, Y., Jameel, H., Phillips, R., Chang, H., 2012. Split addition of enzymes in enzymatic hydrolysis at high solids concentration to increase sugar concentration for bioethanol production. J. Ind. Eng. Chem. 18, 707–714.
105
Zhang, Q., Zhang, J., Shen, J., Silva, A., Dennis, D.A., Barrow, C.J., 2006. A simple 96well microplate method for estimation of total polyphenol content in seaweeds. J. Appl. Phycol. 18, 445–450. Zhao, X., Dong, L., Chen, L., Liu, D., 2013. Batch and multi-step fed-batch enzymatic saccharification of Formiline-pretreated sugarcane bagasse at high solid loadings for high sugar and ethanol titers. Bioresour. Technol. 135, 350–356.