- Email: [email protected]
Optimization of sugar recovery from rapeseed straw pretreated with FeCl3
Accepted Manuscript Optimization of sugar recovery from rapeseed straw pretreated with FeCl3 Inmaculada Romero, Juan C. López-Linares, Manuel Moya, Eulogio Castro PII: DOI: Reference:
S0960-8524(18)31041-1 https://doi.org/10.1016/j.biortech.2018.07.112 BITE 20242
To appear in:
Bioresource Technology
Received Date: Revised Date: Accepted Date:
25 June 2018 22 July 2018 23 July 2018
Please cite this article as: Romero, I., López-Linares, J.C., Moya, M., Castro, E., Optimization of sugar recovery from rapeseed straw pretreated with FeCl3, Bioresource Technology (2018), doi: https://doi.org/10.1016/j.biortech. 2018.07.112
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
OPTIMIZATION OF SUGAR RECOVERY FROM RAPESEED STRAW PRETREATED WITH FeCl3 Inmaculada Romero1,2*, Juan C. López-Linares1, Manuel Moya1,2, Eulogio Castro1,2 1
Dept. of Chemical, Environmental and Materials Engineering, Universidad de Jaén,
2
Center for Advanced Studies in Energy and Environment, Universidad de Jaén, Campus Las Lagunillas, 23071 Jaén, Spain
Highlights
Sugars from FeCl3 pretreatment of rapeseed straw are assessed for the first time
70% of sugars in rapeseed straw were recovered with FeCl3 pretreatment
100 g dry rapeseed straw yielded 37.8 g sugars after FeCl3 pretreatment
FeCl3 enabled pretreatment at mild conditions compared to other methods
ABSTRACT In this work, rapeseed straw was pretreated with FeCl3 to achieve high sugar recoveries. Temperature (120–160 ºC), and FeCl3 concentration (0.1–0.3 M) were selected as factors and modified according to a central composite experimental design. The pretreatment conditions were expressed using the combined severity, which ranged from -0.12 to 2.29. Considering a double criterion that maximizes simultaneously the recovery of hemicellulosic sugars in the liquid fraction from pretreatment and the enzymatic hydrolysis yield, the optimal conditions were found to be 138 ºC and 0.25 M salt concentration. The FeCl3 pretreatment of rapeseed straw under these optimized conditions resulted in 75% hemicellulosic sugar recovery and 53% enzymatic hydrolysis yield. Thereby, 100 g dry rapeseed straw yielded 37.8 g sugars, equivalent to 70% maximum potential sugar in rapeseed straw. Keywords: biorefinery, lignocellulosic biomass, metal salt pretreatment, response
1
surface methodology, severity, sugars.
1. Introduction The path towards a low carbon economy has led to the search for new feedstocks that can replace fossil resources. In this context, the biorefinery concept involves the utilization of biomass feedstocks to obtain different products such as biofuels, chemicals and other added-value products (Uihlein and Schebek, 2009). In this way, the utilization of renewable sources such as agricultural residues in a biorefinery constitutes a very promising option because of their large availability, low cost and environmental benefits, as this can be an alternative disposal method. Moreover, most of them have no practical uses and their elimination is necessary (Dias et al., 2013). Residual biomass has a lignocellulosic structure composed mainly of cellulose, hemicellulose and lignin (Jin et al., 2016; Skiba et al., 2017). Polysaccharides from cellulose and hemicellulose can be hydrolysed to monosaccharides, which can be converted by fermentation or via chemical synthesis into bio-platform molecules to produce green chemicals (Cherubini, 2010). Nowadays, rapeseed straw is considered as an agricultural residue without practical application with high carbohydrate content (Choi et al., 2013). In addition, its production has noticeably increased because of the development of the biodiesel industry. More than 33.7 million hectares of rapeseed were cultivated in the world in 2016, 8.1 million of them in Europe (FAOSTAT, 2018). According to the seed–straw ratio of 1:1.6 estimated by UFOP (2018), more than 110 million tons of rapeseed straw were generated in the world. Concerning straw utilization in the biorefinery context, there are only a few cases in which cereal straw is applied as a raw material for bioethanol production, but only at demonstration scale, i.e., the industrial application of lignocellulosic materials is still under study. Consequently, rapeseed straw is not
2
currently playing any role –but it will do in the near future. Bioconversion of the lignocellulosic biomass requires a pretreatment to disorganize and open its complex structure, making its fractionation possible. The hemicellulose solubilisation, the decrease of the cellulose crystallinity and the increase of the biomass porosity can be considered as the main objectives of the pretreatment step (Elgharbawy et al., 2016). Therefore, a promising pretreatment should be able to enhance the exposure of the cellulose to enzymes with a limited production of toxic compounds that may hinder the subsequent enzymatic hydrolysis and fermentation steps (Ravindran and Jaiswal, 2016). Different pretreatment methods have been assayed, such as steam explosion, liquid hot water, wet oxidation or using chemicals like organosolv, acid or alkali solutions (Ravindran and Jaiswal, 2016). Likewise, new technologies such as ozonolysis (Travaini et al., 2016), microwave, only or combined with alkali (Moodley and Kana, 2017), or fungi treatment (Martínez-Patiño et al., 2018) have been applied in the pretreatment of lignocellulosic biomass. Nevertheless, the utilization of dilute acids is considered as a powerful and effective pretreatment for lignocellulosic biomass (Mosier et al., 2005; Skiba et al., 2017), although they require special equipment materials due to their corrosiveness and they are not readily to recycle. Therefore, the pretreatment with metal salts has been proposed as an environmentally friendly pretreatment able to overcome these issues. Metal salts such as FeCl3 can enhance hemicellulose solubilization with lower sugar degradation associated and consequently with lower formation of inhibitor compounds (Zhao et al., 2011; Romero et al., 2016). In addition, metal salt pretreatment results in a significant improvement of the enzyme activity due to the metal cations form complexes with the lignin, reducing the unproductive enzyme–lignin links (Kamireddy et al., 2013). As a result, pretreatment with ferric chloride has been successfully applied to different lignocellulosic biomass
3
such as wheat straw (Marcotullio et al., 2011), poplar wood (Wang et al., 2013), olive tree biomass (López-Linares et al., 2013a), corn stover (Shen et al., 2014), Eucalyptus wood (Li et al., 2015), rice straw (Chen et al., 2015), corncobs (Shang et al., 2017), spent coffee waste (Ravindran et al., 2017) or sugarcane bagasse (Zhang et al., 2017). In this context, Chen et al. (2015) compared FeCl3 pretreatment with HCl pretreatment for three types of lignocellulosic biomass (bagasse, rice straw and wood fibre) and they reported lower pollution and greater efficiency for metallic salt pretreatment. The objective of this work was to optimize the FeCl3 pretreatment of rapeseed straw as well as to assess the influence of the main operational variables involved in the pretreatment. Taking into account the interest of using the whole material, the optimization was carried out by maximizing the recovery of both cellulosic and hemicellulosic sugars from rapeseed straw, which can be utilized as platform molecules in the biorefinery. In addition, considering the issues caused by the presence of acetic acid and some degradation products such as formic acid, furfural, and 5hydroxymethylfurfural (HMF) in the biological conversion processes, the influence of the pretreatment on the formation of these compounds was also evaluated. To the best of our knowledge, this is the first study on FeCl3 pretreatment of rapeseed straw.
2. Materials and methods 2.1. Raw material Rapeseed straw (8% moisture content) was collected in Seville, Spain, after seed harvest. It was air-dried and milled using a laboratory hammer mill (Retsch, SM-100, Haan, Germany) to a particle size smaller than 0.4 cm. Then, the straw was homogenized and stored in a dry place until use. The feedstock presented the following composition (% w/w): cellulose 31.6 ± 0.3, hemicellulose 17.4 ± 0.1 (xylan 13.2 ± 0.1, 4
galactan 1.9 ± 0.1, arabinan 1.2 ± 0.02, mannan 1.2 ± 0.04), acid insoluble lignin 16.2 ± 0.5, acid soluble lignin 1.6 ± 0.05, acetyl groups 3.4 ± 0.07, ashes 6.7 ± 0.26, and extractives 15.4 ± 1.3 (López-Linares et al., 2015). 2.2. Pretreatment with FeCl3 and experimental design Rapeseed straw was mixed with FeCl3 solution at a solid–liquid ratio of 12% w/v (72 g dry weight material and 600 mL of FeCl3 solution) in a 1 L Parr reactor (Parr Instr. Co., IL, USA). The reactor was heated at a rate of 5 ºC/min and 350 rpm was set as agitation. The reactor was cooled to about 40 ºC after the conditions of the runs were reached according to the experimental design (Table 1). Then, the pretreated material (WIS, water insoluble solids) was separated by filtration, washed with 2 L distilled water and dried at 38 ºC. Next, the WIS were analyzed for hemicellulose, cellulose and lignin content, and used as substrates in the enzymatic hydrolysis tests. Liquid fractions (also referred to as prehydrolysates) were analyzed for sugars, and inhibitors such as formic acid, acetic acid, furfural and hydroxymethylfurfural (HMF). Recoveries of glucose and hemicellulosic sugars were determined as a percentage of the sugar content in the raw material to evaluate the effectiveness of the FeCl3 pretreatment on rapeseed straw.
(1) The pretreatment of rapeseed straw with FeCl3 was studied according to a central composite experimental design (α = 1.414). Temperature (120–160 ºC) and FeCl3 concentration (0.1–0.3 M) were selected as independent variables (factors), keeping the process time at 20 min. A total of 13 experiments including one point and four replicates at the centre of the domain selected for each factor (140 ºC and 0.2 M FeCl3 concentration) were performed in random order. Intervals for the two factors were 5
selected based on previous experience (data not shown). Table 1 shows the coded and actual levels of both the concentration of FeCl3 and the pretreatment temperature. The experimental data were analyzed by the statistical software Design-Expert 8.0.7.1, StatEase Inc., Minneapolis, USA. The harshness of the pretreatment can be determined in terms of severity (S0), which is calculated as the logarithm of the severity factor R0 (Overend and Chornet, 1987). S0 measures the combined effects of time and temperature during the heating, operating and cooling periods (Eq. 2):
(2) where t0 is the time (min) needed to reach TREF, which is a base temperature (usually set to 100 ºC); t1 is the time (min) required to achieve the operating temperature (TMÁX, ºC); t2 is the final time of the operating period (min) at TMAX ; tF is the time (min) needed to achieve TREF during the cooling period; T(t) and T’(t) (°C) are the temperature profiles in the heating and cooling cycles, respectively; w is a fitting parameter, with an assigned value of 14.75 ºC in most studies. To consider also the effect of the FeCl3 concentration, the combined severity (CS) was determined according to the Eq. 2 reported by Chum et al. (1990), which also takes into account the pH of the liquids from the pretreatment: (3)
2.3. Enzymatic hydrolysis The pretreated solids were enzymatically hydrolysed with Cellic CTec2, which is a cellulolytic complex kindly provided by Novozymes A/S (Denmark). A cellulose 6
enzyme loading of 15 filter paper units/g WIS was used. Besides, 15 international units/g WIS of fungal β-glucosidase (Novozym 188, Novozymes A/S) were added to mitigate end-product inhibition caused by cellobiose. The saccharification tests were carried out at 5% (w/v) solids in 100 mL Erlenmeyer flasks with a working volume of 25 mL. The pH was previously adjusted to 4.8 using 0.05 M sodium citrate buffer. Triplicate reaction flasks were incubated at 50 ºC in an orbital shaker (Certomat-R, B-Braun, Germany) at 150 rpm for 72 h. One-millilitre samples were withdrawn from the flasks at 24, 48 and 72 h, and centrifuged at 10,000 g (Sigma 1-14 Centrifuge) for 10 min. Glucose concentrations in the sample supernatant were determined by HPLC. In order to consider the glucose concentration in the commercial enzymes, enzyme blanks were used. The enzymatic hydrolysis yields were calculated as the ratio between the grams of glucose released by enzymatic hydrolysis (EH) and the glucose in the pretreated material (enzymatic digestibility) or in the raw material (enzymatic hydrolysis yield, YEH). (4)
(5)
2.4. Analytical methods The composition of the pretreated solids in cellulose, hemicellulose and lignin was determined using the National Renewable Energy Laboratory (NREL) analytical methodology (Sluiter et al., 2011).
7
The sugar content in the prehydrolysates was determined by high-performance liquid chromatography (HPLC) using a Waters 2695 liquid chromatograph (Milford, MA, USA) equipped with a refractive index detector (Waters 2414) and a Transgenomic CHO-782 carbohydrate analysis column. Ultrapure water was used as a mobile phase (0.6 mL/min) at 70 ºC. The inhibitor concentration (acetic acid, formic acid, furfural, and HMF) in the prehydrolysates was analyzed using the HPLC method with the refractive index detector mentioned above, but with a Bio-Rad HPX-87H column at 65 °C and 0.6 mL/min of 5 mM H2SO4 mobile phase. All analytical determinations were performed in triplicate, and the average results and the relative standard deviations are below 3%.
3. Results and discussion 3.1 Effect of FeCl3 pretreatment conditions on the fractionation of rapeseed straw The splitting of lignocellulosic biomass into its main components is required to make a full utilization of the material, which is the aim of the bioconversion processes, and the pretreatment contributes to this separation (Cherubini, 2010). The reaction mechanism of the FeCl3 as catalysts in the biomass degradation is related to the Lewis acid character and the Fe3+ cation helps to rupture the glycosidic linkages of the hemicellulose (Kamireddy et al., 2013). Pretreatment with metal salts has been reported as a low pH system (Romero et al., 2016). Thereby, the pretreatment of rapeseed straw with FeCl3 yielded acid prehydrolysates with pH ranging between 0.6 and 2.7. A clear relationship between the pH of these liquids from the pretreatment and the salt concentration can be observed (Table 1). In this study, the effect of the pretreatment on the biomass was evaluated using CS, which considers the combination of temperature, time and FeCl3 concentration. The experimental values of CS were adjusted with the
8
factors according to a mathematical model and the corresponding 3D surface is shown in Fig. 1. The model was significant and without lack of fit (R2=0.995). A positive influence of both the temperature and the FeCl3 concentration on CS as well as the lack of interaction between both factors can be observed in Fig. 1. During the pretreatment, the rapeseed straw was partially solubilized, mainly nonstructural fractions like extractives or labile fractions such as hemicellulose. The biomass was solubilized by the combined effect of the temperature, the time and the metal salt. Thereby, the solid recovery was related to the pretreatment severity and this relationship was determined according to a non-linear adjustment (Eq. 6): Solid recovery = 73.46 exp (-0.23·CS)
(6)
showing an excellent adjustment of the experimental data (R2 = 0.9711). Solid recoveries ranged from 46% to 75% corresponding to CS of 2.29 and 0.06, respectively (Table 2). At the central point conditions (140 ºC, 0.2 M FeCl3), around 50% of the biomass was recovered (runs 2, 4, 6, 9, 11). As can be appreciated, the pretreatment made it possible to completely solubilize the hemicellulose when the combined severity was higher than 2 (runs 1 and 3). On the contrary, when the pretreatment was carried out at the lowest severities, the resulting solids accounted for about 15% hemicellulose, which means that the pretreatment did not achieve the full solubilization of this fraction (runs 10 and 13). This can be an inconvenience because the presence of small amounts of hemicellulose in the substrate can affect the performance of the enzymatic hydrolysis because it hinders the enzyme access to the cellulose chains (Jørgensen and Pinelo, 2017). As shown in Table 2, the solubilization of the hemicellulose and extractives of rapeseed straw during the pretreatment involved a cellulose enrichment of the pretreated solids. Thereby, the initial cellulose content in raw straw, 31.5%, was increased up to more
9
than 55% for runs 6 and 7 (1.52 (runs 1 and 3), although in these cases it can be attributed to some degradation of this fraction caused by the high severity. Table 3 shows the sugar composition of the prehydrolysates with concentrations ranging from 7.5 g/L (run 1) up to 25.8 g/L (run 2). Most of these sugars were determined in monomeric form, although the presence of oligomers was detected in all experiments except when the catalyst concentration was 0.3 M or higher. This fact is related to the pretreatment severity, and mainly to the concentration of the catalyst, because an acidic pH is required to break the oligomers solubilized during the pretreatment. It can be noted that xylose was the main sugar in the prehydrolysates and its highest levels were determined at pretreatment of intermediate severity, whereas the highest CS (runs 1 and 3) yielded prehydrolysates with lower xylose content due to partial degradation of this sugar. The presence of glucose was also detected in all prehydrolysates with concentrations below 5 g/L, probably due to the solubilization of a fraction of amorphous cellulose. In addition to sugars released during the pretreatment, the prehydrolysates also contain acetic acid and sugar degradation products. The presence of acetic acid from the deacetylation of hemicellulose is common in the acid liquors from lignocellulosic biomass. 3D response surface and contour plots illustrated the interaction effect of the factors on the acetic acid concentration (Fig. 2a). As can be seen, both the temperature and the catalyst concentration showed a positive influence on the concentration of this organic acid. Concentrations above 3.5 g/L were determined when the pretreatment was
10
carried out at CS>1.4. Furfural originates from the thermal degradation of pentoses, primarily from xylose, since arabinose accounted for only 1.2% in the raw rapeseed straw. Therefore, the presence of furfural in the prehydrolysates suggests some xylose degradation during the pretreatment. As can be seen in Fig. 2b, the influence of the pretreatment temperature on the furfural concentration was positive, although more significant at high FeCl3 concentration, whereas the catalyst concentration was only significant at high pretreatment temperature. The highest concentrations of this furan were determined at the highest values for both temperature and metal salt concentration, reaching furan levels above 5 g/L (data not shown), which coincide with low xylose concentrations in the prehydrolysates (runs 1 and 3). Concerning HMF originating from the degradation of hexoses at high temperatures, Fig. 2c shows the effect of the FeCl3 concentration and temperature pretreatment on the formation of this compound. It can be observed that the catalyst concentration did not affect the HMF concentration at low pretreatment temperature, but a significant positive influence can be appreciated at the highest FeCl3 concentration. However, the pretreatment temperature showed a slight positive influence on the formation of HMF, regardless of the catalyst concentration. Nevertheless, it is worth noting that the presence of HMF in the liquids was scarce, lower than 1 g/L, which could indicate low glucose degradation, although when the severity of the pretreatment increases, HMF generates formic acid, which is even more toxic than furans for microbial growth (Chaabane and Marchal, 2013). As can be appreciated in Fig. 2d, the lowest levels of formic acid were determined at the simultaneous minimum level of both factors, with similar behaviour to that determined for acetic acid (Fig. 2a). Table 4 shows the sugar recovery in the pretreated solids and the prehydrolysates obtained from the different combinations of the temperature and FeCl3 concentration in
11
the pretreatment. As can be seen, a minor fraction of glucose was recovered in the liquid fractions even when the pretreatment was carried out at CS<1 (runs 8, 10 and 13). This fact can be attributed to the solubilization of the amorphous cellulose, which is easily hydrolysable and was consequently hydrolysed during the pretreatment. Nevertheless, cellulosic glucose was mainly recovered in the solid fractions, reaching values above 80% except for CS > 2 (runs 1 and 3) with glucose recoveries lower than 70%. Hemicellulosic sugars can represent an important part of the total carbohydrates in the lignocellulosic materials. For this reason, their valorization can be crucial to contribute to the viability of the bioconversion processes. Iron salts have been reported as a successful system to dissolve hemicellulose due to their low pH (lower than 4) versus other metal salts with higher pH (Romero et al., 2016). In this work, the solubilization of hemicellulosic sugars has been evaluated by determining their recovery in the prehydrolysates, although, in some cases, these sugars have been partly degraded during the pretreatment and therefore, although they were solubilized they could not be recovered. Hemicellulosic sugar recoveries in the liquids between 68% and 81% were determined when the rapeseed straw was pretreated at CS in the range 1.4–2. However, at CS>2 (runs 1 and 3), the pretreatment caused a noticeable hemicellulosic sugar degradation, and less than 16% hemicellulosic sugars were recovered in liquids, even if the presence of hemicellulose was not detected in the pretreated solids. It can be noted that the lowest pretreatment severity (CS<1, runs 10 and 13) also resulted in low recoveries of hemicellulosic sugars in the liquids but, in these cases, it can be explained by the mild conditions that did not favour the hydrolysis of the hemicellulose and, consequently, this fraction remained primarily in the WIS without being solubilized (Table 4).
12
3.2 Effect of the pretreatment conditions on the enzymatic hydrolysis Enzymatic saccharification requires a previous treatment that is able to break down the recalcitrant structure of the lignocellulosic biomass to enhance its enzymatic digestibility (Jørgensen and Pinelo, 2017). According to Liu et al. (2009), FeCl3 breaks ether and ester linkages between lignin and carbohydrates and the enzymatic accessibility of the substrates consequently improves. However, metal salt pretreatment hardly removes the lignin. Therefore, purer lignin could be obtained for other applications. The solids resulting from the FeCl3 pretreatment of rapeseed straw were enzymatically hydrolysed to assess the efficiency of the FeCl3 pretreatment on the rapeseed straw. Thus, the enzymatic saccharification of the pretreated straw yielded hydrolysates with glucose concentrations ranging from 6.2 g/L to 20 g/L (Table 4). The lowest glucose production during the enzymatic hydrolysis occurred when the rapeseed straw was pretreated at CS=0.06 (120 ºC and 0.1 M FeCl3, run 13) with an enzymatic digestibility as low as 30.6%, which suggests that the cellulose crystallinity needs higher pretreatment severity to be altered. Cellulose conversion higher than 58% was determined at the central point of the experimental design (runs 2, 4, 6, 9, and 11) corresponding to CS>1.4. It is worth noting that enzymatic digestibility higher than 65% was only achieved when the pretreatment was carried out at 140 ºC or higher. Thereby, the highest enzymatic digestibility determined for rapeseed straw, 93% (run 3), was reached at the highest temperature assayed, 168.28 ºC and 0.2 M FeCl3. The effectiveness of the FeCl3 pretreatment has been tested with other types of lignocellulosic biomass. This result is comparable to those reached at similar conditions of FeCl3 pretreatment using other lignocellulosic materials such as corn stover (Liu et
13
al., 2009), poplar wood (Wang et al., 2013), Eucalyptus waste fibers (Chen and Fu, 2013) or rice straw (Chen et al., 2015). In addition to high enzymatic digestibility, it is crucial to determine the glucose of the raw straw that can be recovered by the enzymes. For this purpose, the enzymatic hydrolysis yields (YEH) have been determined, expressed as grams of glucose obtained by EH per 100 grams of glucose in raw rapeseed straw. It is common for the value of YEH to be lower than the corresponding value of the enzymatic digestibility due to some sugar loss during the pretreatment. Thus, the highest Enzymatic digestibility determined in this work, 93% (run 3) corresponded to an enzymatic hydrolysis yield of only 50.8% (Table 4). This fact can be explained because the high severity of that pretreatment experiment (CS=2.16) enhanced the cellulose conversion during the enzymatic hydrolysis but, at the same time, these severe conditions determined a high degradation of cellulose during the pretreatment. Consequently, this degradation resulted in low glucose recovery in both the solid and liquid fractions (Table 4). Nevertheless, the maximum value for YEH, 63%, was determined at CS = 1.41 (run 12). Similar enzymatic hydrolysis yields were determined for rapeseed straw pretreated at 180 ºC with 3.7% phosphoric acid (López-Linares et al., 2013b), and for steam exploded rapeseed straw at 200 ºC (López-Linares et al., 2015). Table 4 also includes the overall sugar yield determined for each pretreatment experiment. This yield considers the sugars recovered in the prehydrolysate, mainly hemicellulosic sugars, plus the glucose recovered by enzymatic hydrolysis of the pretreated straw on the basis of 100 g of raw rapeseed straw. Overall sugar yields ranged from 16%, determined at low combined severity (CS=0.06, run 13) to 37% determined at intermediate combined severity (CS=1.73, run 2).
14
3.3 Optimization of the FeCl3 pretreatment parameters The bioconversion of lignocellulosic biomass into fuels and chemicals in the context of the biorefineries implies its fractionation to obtain bioplatform molecules such as sugars, mainly glucose and xylose. FeCl3 pretreatment of this feedstock was evaluated according to the production of sugars from the hemicellulose and cellulose fractions. Thereby, the response HSRL considers the hemicellulosic sugar recovered in the liquids as a fraction of the hemicellulosic sugars in raw straw. The other response chosen, YEH, accounts for the glucose obtained by EH of the cellulose fraction on the basis of 100 g of glucose in raw straw. From the experimental results of HSRL and YEH (Table 4), by means of response surface methodology, quadratic models have been obtained in terms of coded factors to predict both responses (Eq. 7 and 8): (7)
(8)
where T (ºC) is the pretreatment temperature and C (M) is the concentration of FeCl3. These models were used to estimate the released hemicellulosic sugars during the FeCl3 pretreatment (Eq. 7) and the subsequent enzymatic hydrolysis of the pretreated solids (Eq. 8) as a function of both factors T and C, maintaining the pretreatment time at 20 min and the solid loading at 12% (w/v). The coefficients of these equations predict the effect of the temperature and the concentration of catalyst on the responses. The temperature showed a clear negative influence on HSRL whereas the effect of the FeCl3 concentration on this response was positive although less significant (Eq. 7). However, a more significant interaction between T and C with a negative influence on HSRL was determined, indicating that the combined effect of both factors, at the highest and the
15
lowest levels simultaneously, results in a drop of HSRL. Probably, the high severity of the pretreatment, at the highest levels of both factors, caused sugar loss by degradation whilst at the lowest levels the severity was not sufficient to hydrolyse the hemicellulose fraction. This behaviour can be observed in the response surface plots representing the influence of temperature and catalyst concentration on HSRL (Fig. 3a), where the highest values for this response are reached near the central point conditions. With regard to the response YEH, the coefficients of the independent variables were positive, although the temperature was more significant, which indicates that higher levels of the two factors mean higher values for YEH (Eq. 8). Besides, a slight interaction between the two factors but with a clear negative influence on YEH was detected. This behaviour can be observed in Fig. 3b and corresponds to the 3D response surface plots between FeCl3 concentration and temperature on the response YEH. As can be seen, when the pretreatment was carried out at low levels of both factors simultaneously, the enzymatic hydrolysis was poor, resulting in low yields. Variance analyses (ANOVA) for HSRL and YEH are shown in Table 5. As shown, the mathematical model for HSRL has a high significance and reliability of the experimental data at a 95% confidence level. In this model, the interaction effect of temperature (T) and FeCl3 concentration (C) as well as the quadratic terms T2 and C2 were more significant than the linear terms (T and C). Concerning YEH, the mathematical model has a higher F-value (120.27) and lower p-value (<0.0001). The value of R2 was found to be 0.9918, which indicates that 99.18% of the total variation in YEH is attributed to the pretreatment variables studied. On the other hand, the observed value of adjusted R2=0.9835 indicates that the model accounts for 98.35% of the variability in the enzymatic hydrolysis yield. In this case, the temperature also showed a more significant
16
effect than the FeCl3 concentration on YEH due to its higher F-value and its lower pvalue (Table 5). The aim of the FeCl3 pretreatment of rapeseed straw was to recover as many sugars as possible from this feedstock. Therefore, the pretreatment was optimized by maximizing simultaneously the hemicellulosic sugar recovery in the prehydrolysate (HSRL) and the glucose recovery by enzymatic hydrolysis of the WIS (YEH). According to the central composite design, the optimal conditions were found at 138 ºC and 0.25 M FeCl3 and the values predicted by the model for HSRL and YEH were 73% and 51.4%, respectively. The optimal conditions were experimentally reproduced and they yielded a pretreated solid with 43% cellulose and only 1.8% hemicellulose, which, after being hydrolysed by enzymes, reached 75% cellulose conversion. Experimental values for the two responses were determined as 75% hemicellulosic sugar recovery and 53% enzymatic hydrolysis yield. Both results are slightly higher than those predicted by the model, although within the limits of variability at the confidence level of 95%. The mass balance of the sugar production process from rapeseed straw, including FeCl3 pretreatment under optimal conditions and the subsequent enzymatic hydrolysis of the pretreated solid, is shown in Fig. 4. An overall sugar yield of 37.8% was attained, which means that FeCl3 pretreatment made it possible to recover 37.8 g sugars per 100 g of raw rapeseed straw at those conditions; this sugar recovery is equivalent to 70% maximum potential sugars in raw rapeseed straw. With the same feedstock, an overall sugar yield of 41% (63% maximum potential sugar) was reported using pretreatment with 0.3% phosphoric acid at 202 ºC (López-Linares et al., 2013b), while steam explosion pretreatment at 215 ºC for 7.5 min resulted in 31% sugar yield or 58% maximum potential sugar (López-Linares et al., 2015).
17
4. Conclusions Rapeseed straw is an interesting source of biomass for biorefineries, with more than 50% of carbohydrates, which could be converted into bioenergy and high-value chemicals. The optimal conditions for the FeCl3 pretreatment were determined at 138 ºC and 0.25 M FeCl3 (CS = 1.65), yielding 70% of sugar recovery. This means that FeCl3 pretreatment may be an effective method for this agricultural residue. Further research should be focused on improving the performance of the enzymatic saccharification to enable a full conversion of the cellulose fraction.
References Travaini, R., Martín-Juárez, J., Lorenzo-Hernando, A., Bolado-Rodríguez, S., 2016. Ozonolysis: An advantageous pretreatment for lignocellulosic biomass revisited. Bioresour. Technol. 199, 2-12. Chaabane, F.B., Marchal, R., 2013. Upgrading the hemicellulosic fraction of biomass into biofuel. Oil Gas Sci. Technol. – Rev. IFP Energies nouvelles, 68(4), 663-680. Chen, L., Chen, R., Fu, S., 2015. FeCl3 pretreatment of three lignocellulosic biomass for ethanol production. ACS Sustainable Chem. Eng. 3, 1794−1800. Chen, L., Fu, S., 2013. Enhanced cellulase hydrolysis of eucalyptus waste fibers from pulp mill by tween80-assisted ferric chloride pretreatment. J. Agric. Food Chem. 61, 3293−3300. Cherubini, F., 2010. The biorefinery concept: Using biomass instead of oil for producing energy and chemicals. Energy Conv. Manag. 51, 1412–1421. Choi, C.H., Um, B.H., Kim, Y.S., Oh, K.K., 2013. Improved enzyme efficiency of rapeseed straw through the two-stage fractionation process using sodium hydroxide and sulfuric acid. Appl. Energy 102, 640–646. 18
Chum, H.L., Johnson, D.K., Black, S.K., Overend, R.P., 1990. Pretreatment-catalyst effects and the combined severity parameter. Appl. Biochem. Biotechnol. 24–25, 1– 14. Dias, M.O.S., Junqueira, T.L., Cavalett, O., Pavanello, L.G., Cunha, M.P., Jesus, C.D.F., 2013. Biorefineries for the production of first and second generation ethanol and electricity from sugarcane. Appl. Energy 109, 72–78. Elgharbawy, A.A., Alam, M.Z., Moniruzzaman, M., Goto, M., 2016. Ionic liquid pretreatment as emerging approaches for enhanced enzymatic hydrolysis of lignocellulosic biomass. Biochem. Eng. J. 109, 252–267. FAOSTAT, 2018. Food and Agriculture Organization of the United Nations. http://faostat3.fao.org/ (Accessed 13 May 2018). Jin, M., da Costa Sousa, L., Schwartz, C., He, Y., Sarks, C., Gunawan, C., Balan, V., Dale, B.E., 2016. Toward lower cost cellulosic biofuel production using ammonia based pretreatment technologies. Green Chem. 18, 957–966. Jørgensen, H., Pinelo, M., 2017. Enzyme recycling in lignocellulosic biorefineries. Biofuels, Bioprod. Bioref. 11, 150–167. Kamireddy, S.R., Li, J., Tucker, M., Degenstein, J., Ji, Y., 2013. Effects and mechanism of metal chloride salts on pretreatment and enzymatic digestibility of corn stover. Ind. Eng. Chem. Res. 52, 1775-1782. Li, J., Zhang, X., Zhang, M., Xiu, H., He, H., 2015. Ultrasonic enhance acid hydrolysis selectivity of cellulose with HCl–FeCl3 as catalyst. Carbohydr. Polym. 117, 917–922. Liu, L., Sun, J., Li, M., Wang, S., Pei, H., Zhang, J., 2009. Enhanced enzymatic hydrolysis and structural features of corn stover by FeCl3 pretreatment. Bioresour. Technol. 100, 5853−5858.
19
López-Linares, J.C., Ballesteros, I., Tourán, J., Cara, C., Castro, E., Ballesteros, M., Romero, I., 2015. Optimization of uncatalyzed steam explosion pretreatment of rapeseed straw for biofuel production. Bioresour. Technol. 190, 97–105. López-Linares, J.C., Romero, I., Moya, M., Cara, C., Ruiz, E., Castro, E., 2013a. Pretreatment of olive tree biomass with FeCl3 prior enzymatic hydrolysis. Bioresour. Technol. 128, 180-187. López-Linares, J.C., Cara, C., Moya, M., Ruiz, E., Castro, E., Romero, I., 2013b. Fermentable sugar production from rapeseed straw by dilute phosphoric acid pretreatment. Ind. Crop. Prod. 50, 525–531. Marcotullio, G., Krisanti, E., Giuntoli, J., Jong, W.d., 2011. Selective production of hemicellulose-derived carbohydrates from wheat straw using dilute HCl or FeCl3 solutions under mild conditions. X-ray and thermo-gravimetric analysis of the solid residues. Bioresour. Technol. 102, 5917–5923. Martínez-Patiño, J.C., Lu-Chau, T.A., Gullón, B., Ruiz, E., Romero, I., Castro, E., Lema, J.M., 2018. Application of a combined fungal and diluted acid pretreatment on olive tree biomass. Ind. Crop. Prod. 121, 10-17. Moodley, P, Kana, E.B.G., 2017. Development of a steam or microwave-assisted sequential salt-alkali pretreatment for lignocellulosic waste: Effect on delignification and enzymatic hydrolysis. Energy Conv. Manag. 148, 801–808. Mosier, N., Wyman, C., Dale, B., Elander, R., Lee, Y.Y., Holtzapple, M., Ladisch, M., 2005. Features of promising technologies for pretreatment of lignocellulosic biomass. Bioresour. Technol. 96, 673−686. Overend, R.P., Chornet, E., 1987. Fractionation of lignocellulosics by steam-aqueous pretreatments. Philos. Trans. R. Soc. Lond. 321, 523–536.
20
Ravindran, R., Jaiswal, A.K., 2016. A comprehensive review on pre-treatment strategy for lignocellulosic food industry waste: challenges and opportunities. Bioresour. Technol. 199, 92–102. Ravindran, R., Sarangapani, C., Jaiswal, S., Cullen, P.J., Jaiswal, A.K., 2017. Ferric chloride assisted plasma pretreatment of lignocellulose. Bioresour. Technol. 243, 327–334. Romero, I., Ruiz, E., Castro, E., 2016. Pretreatment with metal salts, in: Mussatto, S.I. (Ed.), Biomass Fractionation Technologies for a Lignocellulosic Feedstock Based Biorefinery. Elsevier, pp. 209–227. ISBN: 9780128023235. Shang, Y., Chen, M., Zhao, Q., Su, R., Huang, R., Qi, W., He, Z., 2017. Enhanced enzymatic hydrolysis of lignocellulose by ethanol-assisted FeCl3 pretreatment. Chem. Eng. Trans. 61, 781-786. Shen, Z., Jin, C., Pei, H., Shi, J., Liu, L., Sun, J., 2014. Pretreatment of corn stover with acidic electrolyzed water and FeCl3 leads to enhanced enzymatic hydrolysis. Cellulose 21, 3383–3394. Sluiter, A., Hames, B., Ruiz, R., Scralata, C., Sluiter, J., Templeton, D., 2011. Determination of structural carbohydrates and lignin in biomass. Golden, Colorado: National Renewable Energy Laboratory. Jan, Report No. TP-51042618. Skiba, E.A., Budaeva, V.V., Baibakova, O.V., Zolotukhin, V.N., Sakovich, G.V., 2017. Dilute nitric-acid pretreatment of oat hulls for ethanol production. Bioch. Eng. J. 126, 118–125. Uihlein, A., Schebek, L., 2009. Environmental impacts of a lignocellulose feedstock biorefinery system: an assessment. Biomass Bioenerg. 33, 793–802. UFOP Union Zur Förderung Von Oel - Und Proteinpflanzen E.V. https://www.ufop.de/ english/news (Accessed 13 July 2018).
21
Wang, W., Yuan, T.Q., Cui, B.K., 2013. Fungal treatment followed by FeCl3 treatment to enhance enzymatic hydrolysis of poplar wood for high sugar yields. Biotechnol. Lett. 35, 2061–2067. Zhang, H., Ye, G., Wei, Y., Li, X., Zhang, A., Xie, J., 2017. Enhanced enzymatic hydrolysis of sugarcane bagasse with ferric chloride pretreatment and surfactant. Bioresour. Technol. 229, 96–103. Zhao, J., Zhang, H., Zheng, R., Lin, Z., Huang, H., 2011. The enhancement of pretreatment and enzymatic hydrolysis of corn stover by FeSO4 pretreatment. Bioch. Eng. J. 56, 158–164.
22
Table 1. Experimental design for FeCl3-pretreatment of rapeseed straw, pH of the prehydrolysates and combined severity factor (CS).
Run
Temperature, ºC
FeCl3 concentration, M
pH
CS
0.06
2.73
-0.12
-1
0.1
1.93
0.06
111.72
0
0.2
1.18
0.79
1
160
-1
0.1
2.53
1.32
5
-1
120
1
0.3
0.59
1.40
4
0
140
0
0.2
1.19
1.44
9
0
140
0
0.2
1.13
1.47
6
0
140
0
0.2
1.05
1.55
11
0
140
0
0.2
1.05
1.56
2
0
140
0
0.2
0.98
1.73
7
0
140
+1.41
0.34
0.61
1.98
3
+1.41
168.28
0
0.2
1.38
2.16
1
1
160
1
0.3
0.89
2.29
Coded
Real
Coded
Real
10
0
140
-1.41
13
-1
120
8
-1.41
12
23
Table 2. Recovery of total solids and composition of raw rapeseed straw and FeCl3pretreated solids (%).
WIS composition Run
CS
Untreated straw
Solid Recovery
Cellulose (%)
Hemicellulose (%)
Lignin (%)
--
31.6 ± 0.3
17.4 ± 0.1
17.8 ± 0.5
10
-0.12
74.1
36.9 ± 1.2
14.7 ± 0.2
25.9 ± 0.1
13
0.06
75.1
37.1 ± 0.5
15.6 ± 0.3
23.8 ± 0.2
8
0.79
68.2
37.3 ± 2.8
12.0 ± 0.9
23.0 ± 0.3
12
1.32
54.3
48.4 ± 2.6
5.3 ± 0.3
33.0 ± 0.6
5
1.40
53.2
51.4 ± 0.2
5.8 ± 0.0
30.0 ± 0.4
4
1.44
51.2
51.8 ± 2.6
4.0 ± 0.1
32.3 ± 0.3
9
1.47
50.7
50.1 ± 1.8
3.5 ± 0.2
30.7 ± 0.1
6
1.55
50.4
55.1 ± 0.8
4.6 ± 0.1
30.4 ± 0.2
11
1.56
50.4
50.1 ± 5.0
4.7 ± 0.2
30.7 ± 0.4
2
1.73
49.1
42.7 ± 0.5
3.2 ± 0.1
33.6 ± 0.6
7
1.98
47.1
56.3 ± 0.3
2.1 ± 0.1
33.6 ± 0.2
3
2.16
46.4
37.0 ± 0.7
0.0 ± 0.0
40.1 ± 0.3
1
2.29
45.7
46.4 ± 3.0
0.0 ± 0.0
43.5 ± 0.1
Solid recovery: g WIS/100 g rapeseed straw
24
Table 3. Sugar composition (g/L) and oligomer content (%) in the prehydrolysates
Run
CS
Glucose
Xylose
Galactose
Arabinose
Mannose
Oligomers
10
-0.12
2.65 ± 0.20
2.22 ± 0.03
1.97 ± 0.03
1.94 ± 0.07
0.41 ± 0.08
50.61 ± 0.25
13
0.06
2.51 ± 0.14
1.64 ± 0.13
1.89 ± 0.11
1.99 ± 0.00
0.29 ± 0.02
34.14 ± 0.04
8
0.79
2.76 ± 0.01
4.36 ± 0.07
2.33 ± 0.01
2.27 ± 0.11
0.40 ± 0.07
33.98 ± 0.21
12
1.32
3.02 ± 0.07
8.01 ± 0.10
2.28 ± 0.10
1.94 ± 0.01
1.27 ± 0.01
23.96 ± 0.04
5
1.40
3.39 ± 0.13 10.49 ± 0.07 2.53 ± 0.10
1.83 ± 0.04
1.07 ± 0.07
5.30 ± 0.15
4
1.44
3.39 ± 0.09 10.51 ± 0.15 2.39 ± 0.03
2.17 ± 0.10
1.46 ± 0.04
0.43 ± 0.05
9
1.47
3.87 ± 0.08 11.83 ± 0.05 4.19 ± 0.01
3.94 ± 0.08
1.42 ± 0.06
14.95 ± 0.13
6
1.55
3.78 ± 0.14 11.14 ± 0.11 2.82 ± 0.00
2.32 ± 0.13
1.26 ± 0.00
3.80 ± 0.13
11
1.56
3.78 ± 0.00 10.83 ± 0.07 2.80 ± 0.04
2.07 ± 0.10
1.31 ± 0.07
4.23 ± 0.11
2
1.73
4.57 ± 0.05 13.57 ± 0.07 4.18 ± 0.07
2.14 ± 0.05
1.37 ± 0.01
7.37 ± 0.10
7
1.98
4.11 ± 0.07 10.25 ± 0.08 2.47 ± 0.04
1.87 ± 0.11
1.65 ± 0.08
n.d.
3
2.16
4.16 ± 0.10
1.54 ± 0.11
0.86 ± 0.07
0.48 ± 0.01
0.75 ± 0.03
2.99 ± 0.12
1
2.29
4.93 ± 0.02
1.09 ± 0.06
0.69 ± 0.04
0.24 ± 0.00
0.57 ± 0.00
n.d.
25
Table 4. Recovery (%) of glucose and hemicellulosic sugars in the pretreated solids and the prehydrolysates. Glucose production by enzymatic hydrolysis and overall sugar yield Glucose recovery (%)
Hemicellulosic sugar recovery (%)
Run
CS
WIS
Prehydrolysate
WIS
Prehydrolysate
10
-0.12
86.82
6.37
62.53
26.42
13
0.06
88.31
6.05
67.33
23.10
8
0.79
80.76
6.64
47.31
37.02
12
1.32
83.56
7.27
16.69
56.04
5
1.40
86.95
8.16
17.84
66.48
4
1.44
84.17
8.15
11.75
67.95
9
1.47
80.51
9.31
10.18
73.66
6
1.55
88.19
9.09
13.22
70.00
11
1.56
80.20
9.10
13.72
68.94
2
1.73
84.81
11.00
9.05
80.82
7
1.98
84.07
9.90
5.70
67.56
3
2.16
54.59
10.01
0.00
15.45
1
2.29
67.30
11.86
0.00
10.96
Glucose concentration (g/L)
Enzymatic digestibility (%)
YEH (%)
Overall sugar yield (%)
10
-0.12
7.35 ± 0.14
36.21
31.44
18.28
13
0.06
6.23 ± 0.27
30.55
26.98
15.97
8
0.79
8.17 ± 0.36
39.87
32.20
20.71
12
1.32
20.11 ± 0.76
75.49
63.08
35.36
5
1.40
14.34 ± 0.70
50.68
44.06
31.13
4
1.44
16.68 ± 0.77
58.56
49.29
33.22
9
1.47
17.18 ± 0.15
62.03
50.31
35.10
6
1.55
17.72 ± 0.70
58.49
51.59
34.75
11
1.56
17.88 ± 0.13
64.85
52.01
34.69
2
1.73
17.78 ± 0.30
59.47
50.44
37.13
7
1.98
18.42 ± 0.79
59.53
50.04
34.01
3
2.16
18.96 ± 0.47
93.09
50.82
24.10
1 2.29 18.11 ± 0.20 71.04 47.79 22.81 WIS: water insoluble solids; Enzymatic digestibility: g glucose by enzymatic hydrolysis/100 g glucose in WIS; YEH: g glucose/100 g glucose in rapeseed straw; Overall sugar yield: sum of glucose released by enzymatic hydrolysis and sugars in the prehydrolysate/100 g rapeseed straw.
26
Table 5. Analysis of variance for responses (HSRL and YEH). Source Model Temperature (T) FeCl3 concentration (C) TC T2 C2 Residual Lack of Fit
Sum of p-value F-value Remarks squares (Prob > F) Hemicellulosic sugar recovery (HSRL) 6492.47 38.14 0.0002 Significant 352.25 10.35 0.0182 10.42 0.31 0.6001 1956.29 57.47 0.0003 3849.67 113.08 < 0.0001 155.80 4.58 0.0762 204.26 94.32 Not significant
SDb C.V. %c
5.83 10.97
Model Temperature (T) FeCl3 concentration (C) TC T2 C2 Residual Lack of Fit
Enzymatic hydrolysis yield (YEH) 709.29 120.27 < 0.0001 Significant 183.05 155.20 < 0.0001 19.89 16.86 0.0093 33.17 28.13 0.0032 146.01 123.80 0.0001 19.61 16.62 0.0096 5.90 1.19 1.01 0.3726 Not significant
SDb C.V. %c
1.09 2.36
R-squared AdjR-squareda
R-squared AdjR-squareda
0.9695 0.9441
0.9918 0.9835
a
Adjusted R2 Standard deviation c Coefficient of variation b
27
Figure captions Fig. 1. Response surface for combined severity (CS) as a function of pretreatment temperature and FeCl3 concentration at 20 min. Fig. 2. Response surface curves representing the interactive effect of concentration of ferric chloride and pretreatment temperature on the concentration of acetic acid (a), furfural (b), HMF (c) and formic acid (d) in the prehydrolysates. Fig. 3. Response surface for hemicellulosic sugar recovery, HSR L (a) and enzymatic hydrolysis yield, YEH (b) as a function of pretreatment temperature and FeCl3 concentration at 20 min. Fig. 4. Material balance flow of the FeCl3 pretreatment of rapeseed straw under optimal conditions and its enzymatic hydrolysis
28
Fig. 1.
29
a)
c)
b)
d)
Fig. 2
30
a)
b)
Fig. 3
31
Fig. 4
32
S0960-8524(18)31041-1 https://doi.org/10.1016/j.biortech.2018.07.112 BITE 20242
To appear in:
Bioresource Technology
Received Date: Revised Date: Accepted Date:
25 June 2018 22 July 2018 23 July 2018
Please cite this article as: Romero, I., López-Linares, J.C., Moya, M., Castro, E., Optimization of sugar recovery from rapeseed straw pretreated with FeCl3, Bioresource Technology (2018), doi: https://doi.org/10.1016/j.biortech. 2018.07.112
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
OPTIMIZATION OF SUGAR RECOVERY FROM RAPESEED STRAW PRETREATED WITH FeCl3 Inmaculada Romero1,2*, Juan C. López-Linares1, Manuel Moya1,2, Eulogio Castro1,2 1
Dept. of Chemical, Environmental and Materials Engineering, Universidad de Jaén,
2
Center for Advanced Studies in Energy and Environment, Universidad de Jaén, Campus Las Lagunillas, 23071 Jaén, Spain
Highlights
Sugars from FeCl3 pretreatment of rapeseed straw are assessed for the first time
70% of sugars in rapeseed straw were recovered with FeCl3 pretreatment
100 g dry rapeseed straw yielded 37.8 g sugars after FeCl3 pretreatment
FeCl3 enabled pretreatment at mild conditions compared to other methods
ABSTRACT In this work, rapeseed straw was pretreated with FeCl3 to achieve high sugar recoveries. Temperature (120–160 ºC), and FeCl3 concentration (0.1–0.3 M) were selected as factors and modified according to a central composite experimental design. The pretreatment conditions were expressed using the combined severity, which ranged from -0.12 to 2.29. Considering a double criterion that maximizes simultaneously the recovery of hemicellulosic sugars in the liquid fraction from pretreatment and the enzymatic hydrolysis yield, the optimal conditions were found to be 138 ºC and 0.25 M salt concentration. The FeCl3 pretreatment of rapeseed straw under these optimized conditions resulted in 75% hemicellulosic sugar recovery and 53% enzymatic hydrolysis yield. Thereby, 100 g dry rapeseed straw yielded 37.8 g sugars, equivalent to 70% maximum potential sugar in rapeseed straw. Keywords: biorefinery, lignocellulosic biomass, metal salt pretreatment, response
1
surface methodology, severity, sugars.
1. Introduction The path towards a low carbon economy has led to the search for new feedstocks that can replace fossil resources. In this context, the biorefinery concept involves the utilization of biomass feedstocks to obtain different products such as biofuels, chemicals and other added-value products (Uihlein and Schebek, 2009). In this way, the utilization of renewable sources such as agricultural residues in a biorefinery constitutes a very promising option because of their large availability, low cost and environmental benefits, as this can be an alternative disposal method. Moreover, most of them have no practical uses and their elimination is necessary (Dias et al., 2013). Residual biomass has a lignocellulosic structure composed mainly of cellulose, hemicellulose and lignin (Jin et al., 2016; Skiba et al., 2017). Polysaccharides from cellulose and hemicellulose can be hydrolysed to monosaccharides, which can be converted by fermentation or via chemical synthesis into bio-platform molecules to produce green chemicals (Cherubini, 2010). Nowadays, rapeseed straw is considered as an agricultural residue without practical application with high carbohydrate content (Choi et al., 2013). In addition, its production has noticeably increased because of the development of the biodiesel industry. More than 33.7 million hectares of rapeseed were cultivated in the world in 2016, 8.1 million of them in Europe (FAOSTAT, 2018). According to the seed–straw ratio of 1:1.6 estimated by UFOP (2018), more than 110 million tons of rapeseed straw were generated in the world. Concerning straw utilization in the biorefinery context, there are only a few cases in which cereal straw is applied as a raw material for bioethanol production, but only at demonstration scale, i.e., the industrial application of lignocellulosic materials is still under study. Consequently, rapeseed straw is not
2
currently playing any role –but it will do in the near future. Bioconversion of the lignocellulosic biomass requires a pretreatment to disorganize and open its complex structure, making its fractionation possible. The hemicellulose solubilisation, the decrease of the cellulose crystallinity and the increase of the biomass porosity can be considered as the main objectives of the pretreatment step (Elgharbawy et al., 2016). Therefore, a promising pretreatment should be able to enhance the exposure of the cellulose to enzymes with a limited production of toxic compounds that may hinder the subsequent enzymatic hydrolysis and fermentation steps (Ravindran and Jaiswal, 2016). Different pretreatment methods have been assayed, such as steam explosion, liquid hot water, wet oxidation or using chemicals like organosolv, acid or alkali solutions (Ravindran and Jaiswal, 2016). Likewise, new technologies such as ozonolysis (Travaini et al., 2016), microwave, only or combined with alkali (Moodley and Kana, 2017), or fungi treatment (Martínez-Patiño et al., 2018) have been applied in the pretreatment of lignocellulosic biomass. Nevertheless, the utilization of dilute acids is considered as a powerful and effective pretreatment for lignocellulosic biomass (Mosier et al., 2005; Skiba et al., 2017), although they require special equipment materials due to their corrosiveness and they are not readily to recycle. Therefore, the pretreatment with metal salts has been proposed as an environmentally friendly pretreatment able to overcome these issues. Metal salts such as FeCl3 can enhance hemicellulose solubilization with lower sugar degradation associated and consequently with lower formation of inhibitor compounds (Zhao et al., 2011; Romero et al., 2016). In addition, metal salt pretreatment results in a significant improvement of the enzyme activity due to the metal cations form complexes with the lignin, reducing the unproductive enzyme–lignin links (Kamireddy et al., 2013). As a result, pretreatment with ferric chloride has been successfully applied to different lignocellulosic biomass
3
such as wheat straw (Marcotullio et al., 2011), poplar wood (Wang et al., 2013), olive tree biomass (López-Linares et al., 2013a), corn stover (Shen et al., 2014), Eucalyptus wood (Li et al., 2015), rice straw (Chen et al., 2015), corncobs (Shang et al., 2017), spent coffee waste (Ravindran et al., 2017) or sugarcane bagasse (Zhang et al., 2017). In this context, Chen et al. (2015) compared FeCl3 pretreatment with HCl pretreatment for three types of lignocellulosic biomass (bagasse, rice straw and wood fibre) and they reported lower pollution and greater efficiency for metallic salt pretreatment. The objective of this work was to optimize the FeCl3 pretreatment of rapeseed straw as well as to assess the influence of the main operational variables involved in the pretreatment. Taking into account the interest of using the whole material, the optimization was carried out by maximizing the recovery of both cellulosic and hemicellulosic sugars from rapeseed straw, which can be utilized as platform molecules in the biorefinery. In addition, considering the issues caused by the presence of acetic acid and some degradation products such as formic acid, furfural, and 5hydroxymethylfurfural (HMF) in the biological conversion processes, the influence of the pretreatment on the formation of these compounds was also evaluated. To the best of our knowledge, this is the first study on FeCl3 pretreatment of rapeseed straw.
2. Materials and methods 2.1. Raw material Rapeseed straw (8% moisture content) was collected in Seville, Spain, after seed harvest. It was air-dried and milled using a laboratory hammer mill (Retsch, SM-100, Haan, Germany) to a particle size smaller than 0.4 cm. Then, the straw was homogenized and stored in a dry place until use. The feedstock presented the following composition (% w/w): cellulose 31.6 ± 0.3, hemicellulose 17.4 ± 0.1 (xylan 13.2 ± 0.1, 4
galactan 1.9 ± 0.1, arabinan 1.2 ± 0.02, mannan 1.2 ± 0.04), acid insoluble lignin 16.2 ± 0.5, acid soluble lignin 1.6 ± 0.05, acetyl groups 3.4 ± 0.07, ashes 6.7 ± 0.26, and extractives 15.4 ± 1.3 (López-Linares et al., 2015). 2.2. Pretreatment with FeCl3 and experimental design Rapeseed straw was mixed with FeCl3 solution at a solid–liquid ratio of 12% w/v (72 g dry weight material and 600 mL of FeCl3 solution) in a 1 L Parr reactor (Parr Instr. Co., IL, USA). The reactor was heated at a rate of 5 ºC/min and 350 rpm was set as agitation. The reactor was cooled to about 40 ºC after the conditions of the runs were reached according to the experimental design (Table 1). Then, the pretreated material (WIS, water insoluble solids) was separated by filtration, washed with 2 L distilled water and dried at 38 ºC. Next, the WIS were analyzed for hemicellulose, cellulose and lignin content, and used as substrates in the enzymatic hydrolysis tests. Liquid fractions (also referred to as prehydrolysates) were analyzed for sugars, and inhibitors such as formic acid, acetic acid, furfural and hydroxymethylfurfural (HMF). Recoveries of glucose and hemicellulosic sugars were determined as a percentage of the sugar content in the raw material to evaluate the effectiveness of the FeCl3 pretreatment on rapeseed straw.
(1) The pretreatment of rapeseed straw with FeCl3 was studied according to a central composite experimental design (α = 1.414). Temperature (120–160 ºC) and FeCl3 concentration (0.1–0.3 M) were selected as independent variables (factors), keeping the process time at 20 min. A total of 13 experiments including one point and four replicates at the centre of the domain selected for each factor (140 ºC and 0.2 M FeCl3 concentration) were performed in random order. Intervals for the two factors were 5
selected based on previous experience (data not shown). Table 1 shows the coded and actual levels of both the concentration of FeCl3 and the pretreatment temperature. The experimental data were analyzed by the statistical software Design-Expert 8.0.7.1, StatEase Inc., Minneapolis, USA. The harshness of the pretreatment can be determined in terms of severity (S0), which is calculated as the logarithm of the severity factor R0 (Overend and Chornet, 1987). S0 measures the combined effects of time and temperature during the heating, operating and cooling periods (Eq. 2):
(2) where t0 is the time (min) needed to reach TREF, which is a base temperature (usually set to 100 ºC); t1 is the time (min) required to achieve the operating temperature (TMÁX, ºC); t2 is the final time of the operating period (min) at TMAX ; tF is the time (min) needed to achieve TREF during the cooling period; T(t) and T’(t) (°C) are the temperature profiles in the heating and cooling cycles, respectively; w is a fitting parameter, with an assigned value of 14.75 ºC in most studies. To consider also the effect of the FeCl3 concentration, the combined severity (CS) was determined according to the Eq. 2 reported by Chum et al. (1990), which also takes into account the pH of the liquids from the pretreatment: (3)
2.3. Enzymatic hydrolysis The pretreated solids were enzymatically hydrolysed with Cellic CTec2, which is a cellulolytic complex kindly provided by Novozymes A/S (Denmark). A cellulose 6
enzyme loading of 15 filter paper units/g WIS was used. Besides, 15 international units/g WIS of fungal β-glucosidase (Novozym 188, Novozymes A/S) were added to mitigate end-product inhibition caused by cellobiose. The saccharification tests were carried out at 5% (w/v) solids in 100 mL Erlenmeyer flasks with a working volume of 25 mL. The pH was previously adjusted to 4.8 using 0.05 M sodium citrate buffer. Triplicate reaction flasks were incubated at 50 ºC in an orbital shaker (Certomat-R, B-Braun, Germany) at 150 rpm for 72 h. One-millilitre samples were withdrawn from the flasks at 24, 48 and 72 h, and centrifuged at 10,000 g (Sigma 1-14 Centrifuge) for 10 min. Glucose concentrations in the sample supernatant were determined by HPLC. In order to consider the glucose concentration in the commercial enzymes, enzyme blanks were used. The enzymatic hydrolysis yields were calculated as the ratio between the grams of glucose released by enzymatic hydrolysis (EH) and the glucose in the pretreated material (enzymatic digestibility) or in the raw material (enzymatic hydrolysis yield, YEH). (4)
(5)
2.4. Analytical methods The composition of the pretreated solids in cellulose, hemicellulose and lignin was determined using the National Renewable Energy Laboratory (NREL) analytical methodology (Sluiter et al., 2011).
7
The sugar content in the prehydrolysates was determined by high-performance liquid chromatography (HPLC) using a Waters 2695 liquid chromatograph (Milford, MA, USA) equipped with a refractive index detector (Waters 2414) and a Transgenomic CHO-782 carbohydrate analysis column. Ultrapure water was used as a mobile phase (0.6 mL/min) at 70 ºC. The inhibitor concentration (acetic acid, formic acid, furfural, and HMF) in the prehydrolysates was analyzed using the HPLC method with the refractive index detector mentioned above, but with a Bio-Rad HPX-87H column at 65 °C and 0.6 mL/min of 5 mM H2SO4 mobile phase. All analytical determinations were performed in triplicate, and the average results and the relative standard deviations are below 3%.
3. Results and discussion 3.1 Effect of FeCl3 pretreatment conditions on the fractionation of rapeseed straw The splitting of lignocellulosic biomass into its main components is required to make a full utilization of the material, which is the aim of the bioconversion processes, and the pretreatment contributes to this separation (Cherubini, 2010). The reaction mechanism of the FeCl3 as catalysts in the biomass degradation is related to the Lewis acid character and the Fe3+ cation helps to rupture the glycosidic linkages of the hemicellulose (Kamireddy et al., 2013). Pretreatment with metal salts has been reported as a low pH system (Romero et al., 2016). Thereby, the pretreatment of rapeseed straw with FeCl3 yielded acid prehydrolysates with pH ranging between 0.6 and 2.7. A clear relationship between the pH of these liquids from the pretreatment and the salt concentration can be observed (Table 1). In this study, the effect of the pretreatment on the biomass was evaluated using CS, which considers the combination of temperature, time and FeCl3 concentration. The experimental values of CS were adjusted with the
8
factors according to a mathematical model and the corresponding 3D surface is shown in Fig. 1. The model was significant and without lack of fit (R2=0.995). A positive influence of both the temperature and the FeCl3 concentration on CS as well as the lack of interaction between both factors can be observed in Fig. 1. During the pretreatment, the rapeseed straw was partially solubilized, mainly nonstructural fractions like extractives or labile fractions such as hemicellulose. The biomass was solubilized by the combined effect of the temperature, the time and the metal salt. Thereby, the solid recovery was related to the pretreatment severity and this relationship was determined according to a non-linear adjustment (Eq. 6): Solid recovery = 73.46 exp (-0.23·CS)
(6)
showing an excellent adjustment of the experimental data (R2 = 0.9711). Solid recoveries ranged from 46% to 75% corresponding to CS of 2.29 and 0.06, respectively (Table 2). At the central point conditions (140 ºC, 0.2 M FeCl3), around 50% of the biomass was recovered (runs 2, 4, 6, 9, 11). As can be appreciated, the pretreatment made it possible to completely solubilize the hemicellulose when the combined severity was higher than 2 (runs 1 and 3). On the contrary, when the pretreatment was carried out at the lowest severities, the resulting solids accounted for about 15% hemicellulose, which means that the pretreatment did not achieve the full solubilization of this fraction (runs 10 and 13). This can be an inconvenience because the presence of small amounts of hemicellulose in the substrate can affect the performance of the enzymatic hydrolysis because it hinders the enzyme access to the cellulose chains (Jørgensen and Pinelo, 2017). As shown in Table 2, the solubilization of the hemicellulose and extractives of rapeseed straw during the pretreatment involved a cellulose enrichment of the pretreated solids. Thereby, the initial cellulose content in raw straw, 31.5%, was increased up to more
9
than 55% for runs 6 and 7 (1.5
10
carried out at CS>1.4. Furfural originates from the thermal degradation of pentoses, primarily from xylose, since arabinose accounted for only 1.2% in the raw rapeseed straw. Therefore, the presence of furfural in the prehydrolysates suggests some xylose degradation during the pretreatment. As can be seen in Fig. 2b, the influence of the pretreatment temperature on the furfural concentration was positive, although more significant at high FeCl3 concentration, whereas the catalyst concentration was only significant at high pretreatment temperature. The highest concentrations of this furan were determined at the highest values for both temperature and metal salt concentration, reaching furan levels above 5 g/L (data not shown), which coincide with low xylose concentrations in the prehydrolysates (runs 1 and 3). Concerning HMF originating from the degradation of hexoses at high temperatures, Fig. 2c shows the effect of the FeCl3 concentration and temperature pretreatment on the formation of this compound. It can be observed that the catalyst concentration did not affect the HMF concentration at low pretreatment temperature, but a significant positive influence can be appreciated at the highest FeCl3 concentration. However, the pretreatment temperature showed a slight positive influence on the formation of HMF, regardless of the catalyst concentration. Nevertheless, it is worth noting that the presence of HMF in the liquids was scarce, lower than 1 g/L, which could indicate low glucose degradation, although when the severity of the pretreatment increases, HMF generates formic acid, which is even more toxic than furans for microbial growth (Chaabane and Marchal, 2013). As can be appreciated in Fig. 2d, the lowest levels of formic acid were determined at the simultaneous minimum level of both factors, with similar behaviour to that determined for acetic acid (Fig. 2a). Table 4 shows the sugar recovery in the pretreated solids and the prehydrolysates obtained from the different combinations of the temperature and FeCl3 concentration in
11
the pretreatment. As can be seen, a minor fraction of glucose was recovered in the liquid fractions even when the pretreatment was carried out at CS<1 (runs 8, 10 and 13). This fact can be attributed to the solubilization of the amorphous cellulose, which is easily hydrolysable and was consequently hydrolysed during the pretreatment. Nevertheless, cellulosic glucose was mainly recovered in the solid fractions, reaching values above 80% except for CS > 2 (runs 1 and 3) with glucose recoveries lower than 70%. Hemicellulosic sugars can represent an important part of the total carbohydrates in the lignocellulosic materials. For this reason, their valorization can be crucial to contribute to the viability of the bioconversion processes. Iron salts have been reported as a successful system to dissolve hemicellulose due to their low pH (lower than 4) versus other metal salts with higher pH (Romero et al., 2016). In this work, the solubilization of hemicellulosic sugars has been evaluated by determining their recovery in the prehydrolysates, although, in some cases, these sugars have been partly degraded during the pretreatment and therefore, although they were solubilized they could not be recovered. Hemicellulosic sugar recoveries in the liquids between 68% and 81% were determined when the rapeseed straw was pretreated at CS in the range 1.4–2. However, at CS>2 (runs 1 and 3), the pretreatment caused a noticeable hemicellulosic sugar degradation, and less than 16% hemicellulosic sugars were recovered in liquids, even if the presence of hemicellulose was not detected in the pretreated solids. It can be noted that the lowest pretreatment severity (CS<1, runs 10 and 13) also resulted in low recoveries of hemicellulosic sugars in the liquids but, in these cases, it can be explained by the mild conditions that did not favour the hydrolysis of the hemicellulose and, consequently, this fraction remained primarily in the WIS without being solubilized (Table 4).
12
3.2 Effect of the pretreatment conditions on the enzymatic hydrolysis Enzymatic saccharification requires a previous treatment that is able to break down the recalcitrant structure of the lignocellulosic biomass to enhance its enzymatic digestibility (Jørgensen and Pinelo, 2017). According to Liu et al. (2009), FeCl3 breaks ether and ester linkages between lignin and carbohydrates and the enzymatic accessibility of the substrates consequently improves. However, metal salt pretreatment hardly removes the lignin. Therefore, purer lignin could be obtained for other applications. The solids resulting from the FeCl3 pretreatment of rapeseed straw were enzymatically hydrolysed to assess the efficiency of the FeCl3 pretreatment on the rapeseed straw. Thus, the enzymatic saccharification of the pretreated straw yielded hydrolysates with glucose concentrations ranging from 6.2 g/L to 20 g/L (Table 4). The lowest glucose production during the enzymatic hydrolysis occurred when the rapeseed straw was pretreated at CS=0.06 (120 ºC and 0.1 M FeCl3, run 13) with an enzymatic digestibility as low as 30.6%, which suggests that the cellulose crystallinity needs higher pretreatment severity to be altered. Cellulose conversion higher than 58% was determined at the central point of the experimental design (runs 2, 4, 6, 9, and 11) corresponding to CS>1.4. It is worth noting that enzymatic digestibility higher than 65% was only achieved when the pretreatment was carried out at 140 ºC or higher. Thereby, the highest enzymatic digestibility determined for rapeseed straw, 93% (run 3), was reached at the highest temperature assayed, 168.28 ºC and 0.2 M FeCl3. The effectiveness of the FeCl3 pretreatment has been tested with other types of lignocellulosic biomass. This result is comparable to those reached at similar conditions of FeCl3 pretreatment using other lignocellulosic materials such as corn stover (Liu et
13
al., 2009), poplar wood (Wang et al., 2013), Eucalyptus waste fibers (Chen and Fu, 2013) or rice straw (Chen et al., 2015). In addition to high enzymatic digestibility, it is crucial to determine the glucose of the raw straw that can be recovered by the enzymes. For this purpose, the enzymatic hydrolysis yields (YEH) have been determined, expressed as grams of glucose obtained by EH per 100 grams of glucose in raw rapeseed straw. It is common for the value of YEH to be lower than the corresponding value of the enzymatic digestibility due to some sugar loss during the pretreatment. Thus, the highest Enzymatic digestibility determined in this work, 93% (run 3) corresponded to an enzymatic hydrolysis yield of only 50.8% (Table 4). This fact can be explained because the high severity of that pretreatment experiment (CS=2.16) enhanced the cellulose conversion during the enzymatic hydrolysis but, at the same time, these severe conditions determined a high degradation of cellulose during the pretreatment. Consequently, this degradation resulted in low glucose recovery in both the solid and liquid fractions (Table 4). Nevertheless, the maximum value for YEH, 63%, was determined at CS = 1.41 (run 12). Similar enzymatic hydrolysis yields were determined for rapeseed straw pretreated at 180 ºC with 3.7% phosphoric acid (López-Linares et al., 2013b), and for steam exploded rapeseed straw at 200 ºC (López-Linares et al., 2015). Table 4 also includes the overall sugar yield determined for each pretreatment experiment. This yield considers the sugars recovered in the prehydrolysate, mainly hemicellulosic sugars, plus the glucose recovered by enzymatic hydrolysis of the pretreated straw on the basis of 100 g of raw rapeseed straw. Overall sugar yields ranged from 16%, determined at low combined severity (CS=0.06, run 13) to 37% determined at intermediate combined severity (CS=1.73, run 2).
14
3.3 Optimization of the FeCl3 pretreatment parameters The bioconversion of lignocellulosic biomass into fuels and chemicals in the context of the biorefineries implies its fractionation to obtain bioplatform molecules such as sugars, mainly glucose and xylose. FeCl3 pretreatment of this feedstock was evaluated according to the production of sugars from the hemicellulose and cellulose fractions. Thereby, the response HSRL considers the hemicellulosic sugar recovered in the liquids as a fraction of the hemicellulosic sugars in raw straw. The other response chosen, YEH, accounts for the glucose obtained by EH of the cellulose fraction on the basis of 100 g of glucose in raw straw. From the experimental results of HSRL and YEH (Table 4), by means of response surface methodology, quadratic models have been obtained in terms of coded factors to predict both responses (Eq. 7 and 8): (7)
(8)
where T (ºC) is the pretreatment temperature and C (M) is the concentration of FeCl3. These models were used to estimate the released hemicellulosic sugars during the FeCl3 pretreatment (Eq. 7) and the subsequent enzymatic hydrolysis of the pretreated solids (Eq. 8) as a function of both factors T and C, maintaining the pretreatment time at 20 min and the solid loading at 12% (w/v). The coefficients of these equations predict the effect of the temperature and the concentration of catalyst on the responses. The temperature showed a clear negative influence on HSRL whereas the effect of the FeCl3 concentration on this response was positive although less significant (Eq. 7). However, a more significant interaction between T and C with a negative influence on HSRL was determined, indicating that the combined effect of both factors, at the highest and the
15
lowest levels simultaneously, results in a drop of HSRL. Probably, the high severity of the pretreatment, at the highest levels of both factors, caused sugar loss by degradation whilst at the lowest levels the severity was not sufficient to hydrolyse the hemicellulose fraction. This behaviour can be observed in the response surface plots representing the influence of temperature and catalyst concentration on HSRL (Fig. 3a), where the highest values for this response are reached near the central point conditions. With regard to the response YEH, the coefficients of the independent variables were positive, although the temperature was more significant, which indicates that higher levels of the two factors mean higher values for YEH (Eq. 8). Besides, a slight interaction between the two factors but with a clear negative influence on YEH was detected. This behaviour can be observed in Fig. 3b and corresponds to the 3D response surface plots between FeCl3 concentration and temperature on the response YEH. As can be seen, when the pretreatment was carried out at low levels of both factors simultaneously, the enzymatic hydrolysis was poor, resulting in low yields. Variance analyses (ANOVA) for HSRL and YEH are shown in Table 5. As shown, the mathematical model for HSRL has a high significance and reliability of the experimental data at a 95% confidence level. In this model, the interaction effect of temperature (T) and FeCl3 concentration (C) as well as the quadratic terms T2 and C2 were more significant than the linear terms (T and C). Concerning YEH, the mathematical model has a higher F-value (120.27) and lower p-value (<0.0001). The value of R2 was found to be 0.9918, which indicates that 99.18% of the total variation in YEH is attributed to the pretreatment variables studied. On the other hand, the observed value of adjusted R2=0.9835 indicates that the model accounts for 98.35% of the variability in the enzymatic hydrolysis yield. In this case, the temperature also showed a more significant
16
effect than the FeCl3 concentration on YEH due to its higher F-value and its lower pvalue (Table 5). The aim of the FeCl3 pretreatment of rapeseed straw was to recover as many sugars as possible from this feedstock. Therefore, the pretreatment was optimized by maximizing simultaneously the hemicellulosic sugar recovery in the prehydrolysate (HSRL) and the glucose recovery by enzymatic hydrolysis of the WIS (YEH). According to the central composite design, the optimal conditions were found at 138 ºC and 0.25 M FeCl3 and the values predicted by the model for HSRL and YEH were 73% and 51.4%, respectively. The optimal conditions were experimentally reproduced and they yielded a pretreated solid with 43% cellulose and only 1.8% hemicellulose, which, after being hydrolysed by enzymes, reached 75% cellulose conversion. Experimental values for the two responses were determined as 75% hemicellulosic sugar recovery and 53% enzymatic hydrolysis yield. Both results are slightly higher than those predicted by the model, although within the limits of variability at the confidence level of 95%. The mass balance of the sugar production process from rapeseed straw, including FeCl3 pretreatment under optimal conditions and the subsequent enzymatic hydrolysis of the pretreated solid, is shown in Fig. 4. An overall sugar yield of 37.8% was attained, which means that FeCl3 pretreatment made it possible to recover 37.8 g sugars per 100 g of raw rapeseed straw at those conditions; this sugar recovery is equivalent to 70% maximum potential sugars in raw rapeseed straw. With the same feedstock, an overall sugar yield of 41% (63% maximum potential sugar) was reported using pretreatment with 0.3% phosphoric acid at 202 ºC (López-Linares et al., 2013b), while steam explosion pretreatment at 215 ºC for 7.5 min resulted in 31% sugar yield or 58% maximum potential sugar (López-Linares et al., 2015).
17
4. Conclusions Rapeseed straw is an interesting source of biomass for biorefineries, with more than 50% of carbohydrates, which could be converted into bioenergy and high-value chemicals. The optimal conditions for the FeCl3 pretreatment were determined at 138 ºC and 0.25 M FeCl3 (CS = 1.65), yielding 70% of sugar recovery. This means that FeCl3 pretreatment may be an effective method for this agricultural residue. Further research should be focused on improving the performance of the enzymatic saccharification to enable a full conversion of the cellulose fraction.
References Travaini, R., Martín-Juárez, J., Lorenzo-Hernando, A., Bolado-Rodríguez, S., 2016. Ozonolysis: An advantageous pretreatment for lignocellulosic biomass revisited. Bioresour. Technol. 199, 2-12. Chaabane, F.B., Marchal, R., 2013. Upgrading the hemicellulosic fraction of biomass into biofuel. Oil Gas Sci. Technol. – Rev. IFP Energies nouvelles, 68(4), 663-680. Chen, L., Chen, R., Fu, S., 2015. FeCl3 pretreatment of three lignocellulosic biomass for ethanol production. ACS Sustainable Chem. Eng. 3, 1794−1800. Chen, L., Fu, S., 2013. Enhanced cellulase hydrolysis of eucalyptus waste fibers from pulp mill by tween80-assisted ferric chloride pretreatment. J. Agric. Food Chem. 61, 3293−3300. Cherubini, F., 2010. The biorefinery concept: Using biomass instead of oil for producing energy and chemicals. Energy Conv. Manag. 51, 1412–1421. Choi, C.H., Um, B.H., Kim, Y.S., Oh, K.K., 2013. Improved enzyme efficiency of rapeseed straw through the two-stage fractionation process using sodium hydroxide and sulfuric acid. Appl. Energy 102, 640–646. 18
Chum, H.L., Johnson, D.K., Black, S.K., Overend, R.P., 1990. Pretreatment-catalyst effects and the combined severity parameter. Appl. Biochem. Biotechnol. 24–25, 1– 14. Dias, M.O.S., Junqueira, T.L., Cavalett, O., Pavanello, L.G., Cunha, M.P., Jesus, C.D.F., 2013. Biorefineries for the production of first and second generation ethanol and electricity from sugarcane. Appl. Energy 109, 72–78. Elgharbawy, A.A., Alam, M.Z., Moniruzzaman, M., Goto, M., 2016. Ionic liquid pretreatment as emerging approaches for enhanced enzymatic hydrolysis of lignocellulosic biomass. Biochem. Eng. J. 109, 252–267. FAOSTAT, 2018. Food and Agriculture Organization of the United Nations. http://faostat3.fao.org/ (Accessed 13 May 2018). Jin, M., da Costa Sousa, L., Schwartz, C., He, Y., Sarks, C., Gunawan, C., Balan, V., Dale, B.E., 2016. Toward lower cost cellulosic biofuel production using ammonia based pretreatment technologies. Green Chem. 18, 957–966. Jørgensen, H., Pinelo, M., 2017. Enzyme recycling in lignocellulosic biorefineries. Biofuels, Bioprod. Bioref. 11, 150–167. Kamireddy, S.R., Li, J., Tucker, M., Degenstein, J., Ji, Y., 2013. Effects and mechanism of metal chloride salts on pretreatment and enzymatic digestibility of corn stover. Ind. Eng. Chem. Res. 52, 1775-1782. Li, J., Zhang, X., Zhang, M., Xiu, H., He, H., 2015. Ultrasonic enhance acid hydrolysis selectivity of cellulose with HCl–FeCl3 as catalyst. Carbohydr. Polym. 117, 917–922. Liu, L., Sun, J., Li, M., Wang, S., Pei, H., Zhang, J., 2009. Enhanced enzymatic hydrolysis and structural features of corn stover by FeCl3 pretreatment. Bioresour. Technol. 100, 5853−5858.
19
López-Linares, J.C., Ballesteros, I., Tourán, J., Cara, C., Castro, E., Ballesteros, M., Romero, I., 2015. Optimization of uncatalyzed steam explosion pretreatment of rapeseed straw for biofuel production. Bioresour. Technol. 190, 97–105. López-Linares, J.C., Romero, I., Moya, M., Cara, C., Ruiz, E., Castro, E., 2013a. Pretreatment of olive tree biomass with FeCl3 prior enzymatic hydrolysis. Bioresour. Technol. 128, 180-187. López-Linares, J.C., Cara, C., Moya, M., Ruiz, E., Castro, E., Romero, I., 2013b. Fermentable sugar production from rapeseed straw by dilute phosphoric acid pretreatment. Ind. Crop. Prod. 50, 525–531. Marcotullio, G., Krisanti, E., Giuntoli, J., Jong, W.d., 2011. Selective production of hemicellulose-derived carbohydrates from wheat straw using dilute HCl or FeCl3 solutions under mild conditions. X-ray and thermo-gravimetric analysis of the solid residues. Bioresour. Technol. 102, 5917–5923. Martínez-Patiño, J.C., Lu-Chau, T.A., Gullón, B., Ruiz, E., Romero, I., Castro, E., Lema, J.M., 2018. Application of a combined fungal and diluted acid pretreatment on olive tree biomass. Ind. Crop. Prod. 121, 10-17. Moodley, P, Kana, E.B.G., 2017. Development of a steam or microwave-assisted sequential salt-alkali pretreatment for lignocellulosic waste: Effect on delignification and enzymatic hydrolysis. Energy Conv. Manag. 148, 801–808. Mosier, N., Wyman, C., Dale, B., Elander, R., Lee, Y.Y., Holtzapple, M., Ladisch, M., 2005. Features of promising technologies for pretreatment of lignocellulosic biomass. Bioresour. Technol. 96, 673−686. Overend, R.P., Chornet, E., 1987. Fractionation of lignocellulosics by steam-aqueous pretreatments. Philos. Trans. R. Soc. Lond. 321, 523–536.
20
Ravindran, R., Jaiswal, A.K., 2016. A comprehensive review on pre-treatment strategy for lignocellulosic food industry waste: challenges and opportunities. Bioresour. Technol. 199, 92–102. Ravindran, R., Sarangapani, C., Jaiswal, S., Cullen, P.J., Jaiswal, A.K., 2017. Ferric chloride assisted plasma pretreatment of lignocellulose. Bioresour. Technol. 243, 327–334. Romero, I., Ruiz, E., Castro, E., 2016. Pretreatment with metal salts, in: Mussatto, S.I. (Ed.), Biomass Fractionation Technologies for a Lignocellulosic Feedstock Based Biorefinery. Elsevier, pp. 209–227. ISBN: 9780128023235. Shang, Y., Chen, M., Zhao, Q., Su, R., Huang, R., Qi, W., He, Z., 2017. Enhanced enzymatic hydrolysis of lignocellulose by ethanol-assisted FeCl3 pretreatment. Chem. Eng. Trans. 61, 781-786. Shen, Z., Jin, C., Pei, H., Shi, J., Liu, L., Sun, J., 2014. Pretreatment of corn stover with acidic electrolyzed water and FeCl3 leads to enhanced enzymatic hydrolysis. Cellulose 21, 3383–3394. Sluiter, A., Hames, B., Ruiz, R., Scralata, C., Sluiter, J., Templeton, D., 2011. Determination of structural carbohydrates and lignin in biomass. Golden, Colorado: National Renewable Energy Laboratory. Jan, Report No. TP-51042618. Skiba, E.A., Budaeva, V.V., Baibakova, O.V., Zolotukhin, V.N., Sakovich, G.V., 2017. Dilute nitric-acid pretreatment of oat hulls for ethanol production. Bioch. Eng. J. 126, 118–125. Uihlein, A., Schebek, L., 2009. Environmental impacts of a lignocellulose feedstock biorefinery system: an assessment. Biomass Bioenerg. 33, 793–802. UFOP Union Zur Förderung Von Oel - Und Proteinpflanzen E.V. https://www.ufop.de/ english/news (Accessed 13 July 2018).
21
Wang, W., Yuan, T.Q., Cui, B.K., 2013. Fungal treatment followed by FeCl3 treatment to enhance enzymatic hydrolysis of poplar wood for high sugar yields. Biotechnol. Lett. 35, 2061–2067. Zhang, H., Ye, G., Wei, Y., Li, X., Zhang, A., Xie, J., 2017. Enhanced enzymatic hydrolysis of sugarcane bagasse with ferric chloride pretreatment and surfactant. Bioresour. Technol. 229, 96–103. Zhao, J., Zhang, H., Zheng, R., Lin, Z., Huang, H., 2011. The enhancement of pretreatment and enzymatic hydrolysis of corn stover by FeSO4 pretreatment. Bioch. Eng. J. 56, 158–164.
22
Table 1. Experimental design for FeCl3-pretreatment of rapeseed straw, pH of the prehydrolysates and combined severity factor (CS).
Run
Temperature, ºC
FeCl3 concentration, M
pH
CS
0.06
2.73
-0.12
-1
0.1
1.93
0.06
111.72
0
0.2
1.18
0.79
1
160
-1
0.1
2.53
1.32
5
-1
120
1
0.3
0.59
1.40
4
0
140
0
0.2
1.19
1.44
9
0
140
0
0.2
1.13
1.47
6
0
140
0
0.2
1.05
1.55
11
0
140
0
0.2
1.05
1.56
2
0
140
0
0.2
0.98
1.73
7
0
140
+1.41
0.34
0.61
1.98
3
+1.41
168.28
0
0.2
1.38
2.16
1
1
160
1
0.3
0.89
2.29
Coded
Real
Coded
Real
10
0
140
-1.41
13
-1
120
8
-1.41
12
23
Table 2. Recovery of total solids and composition of raw rapeseed straw and FeCl3pretreated solids (%).
WIS composition Run
CS
Untreated straw
Solid Recovery
Cellulose (%)
Hemicellulose (%)
Lignin (%)
--
31.6 ± 0.3
17.4 ± 0.1
17.8 ± 0.5
10
-0.12
74.1
36.9 ± 1.2
14.7 ± 0.2
25.9 ± 0.1
13
0.06
75.1
37.1 ± 0.5
15.6 ± 0.3
23.8 ± 0.2
8
0.79
68.2
37.3 ± 2.8
12.0 ± 0.9
23.0 ± 0.3
12
1.32
54.3
48.4 ± 2.6
5.3 ± 0.3
33.0 ± 0.6
5
1.40
53.2
51.4 ± 0.2
5.8 ± 0.0
30.0 ± 0.4
4
1.44
51.2
51.8 ± 2.6
4.0 ± 0.1
32.3 ± 0.3
9
1.47
50.7
50.1 ± 1.8
3.5 ± 0.2
30.7 ± 0.1
6
1.55
50.4
55.1 ± 0.8
4.6 ± 0.1
30.4 ± 0.2
11
1.56
50.4
50.1 ± 5.0
4.7 ± 0.2
30.7 ± 0.4
2
1.73
49.1
42.7 ± 0.5
3.2 ± 0.1
33.6 ± 0.6
7
1.98
47.1
56.3 ± 0.3
2.1 ± 0.1
33.6 ± 0.2
3
2.16
46.4
37.0 ± 0.7
0.0 ± 0.0
40.1 ± 0.3
1
2.29
45.7
46.4 ± 3.0
0.0 ± 0.0
43.5 ± 0.1
Solid recovery: g WIS/100 g rapeseed straw
24
Table 3. Sugar composition (g/L) and oligomer content (%) in the prehydrolysates
Run
CS
Glucose
Xylose
Galactose
Arabinose
Mannose
Oligomers
10
-0.12
2.65 ± 0.20
2.22 ± 0.03
1.97 ± 0.03
1.94 ± 0.07
0.41 ± 0.08
50.61 ± 0.25
13
0.06
2.51 ± 0.14
1.64 ± 0.13
1.89 ± 0.11
1.99 ± 0.00
0.29 ± 0.02
34.14 ± 0.04
8
0.79
2.76 ± 0.01
4.36 ± 0.07
2.33 ± 0.01
2.27 ± 0.11
0.40 ± 0.07
33.98 ± 0.21
12
1.32
3.02 ± 0.07
8.01 ± 0.10
2.28 ± 0.10
1.94 ± 0.01
1.27 ± 0.01
23.96 ± 0.04
5
1.40
3.39 ± 0.13 10.49 ± 0.07 2.53 ± 0.10
1.83 ± 0.04
1.07 ± 0.07
5.30 ± 0.15
4
1.44
3.39 ± 0.09 10.51 ± 0.15 2.39 ± 0.03
2.17 ± 0.10
1.46 ± 0.04
0.43 ± 0.05
9
1.47
3.87 ± 0.08 11.83 ± 0.05 4.19 ± 0.01
3.94 ± 0.08
1.42 ± 0.06
14.95 ± 0.13
6
1.55
3.78 ± 0.14 11.14 ± 0.11 2.82 ± 0.00
2.32 ± 0.13
1.26 ± 0.00
3.80 ± 0.13
11
1.56
3.78 ± 0.00 10.83 ± 0.07 2.80 ± 0.04
2.07 ± 0.10
1.31 ± 0.07
4.23 ± 0.11
2
1.73
4.57 ± 0.05 13.57 ± 0.07 4.18 ± 0.07
2.14 ± 0.05
1.37 ± 0.01
7.37 ± 0.10
7
1.98
4.11 ± 0.07 10.25 ± 0.08 2.47 ± 0.04
1.87 ± 0.11
1.65 ± 0.08
n.d.
3
2.16
4.16 ± 0.10
1.54 ± 0.11
0.86 ± 0.07
0.48 ± 0.01
0.75 ± 0.03
2.99 ± 0.12
1
2.29
4.93 ± 0.02
1.09 ± 0.06
0.69 ± 0.04
0.24 ± 0.00
0.57 ± 0.00
n.d.
25
Table 4. Recovery (%) of glucose and hemicellulosic sugars in the pretreated solids and the prehydrolysates. Glucose production by enzymatic hydrolysis and overall sugar yield Glucose recovery (%)
Hemicellulosic sugar recovery (%)
Run
CS
WIS
Prehydrolysate
WIS
Prehydrolysate
10
-0.12
86.82
6.37
62.53
26.42
13
0.06
88.31
6.05
67.33
23.10
8
0.79
80.76
6.64
47.31
37.02
12
1.32
83.56
7.27
16.69
56.04
5
1.40
86.95
8.16
17.84
66.48
4
1.44
84.17
8.15
11.75
67.95
9
1.47
80.51
9.31
10.18
73.66
6
1.55
88.19
9.09
13.22
70.00
11
1.56
80.20
9.10
13.72
68.94
2
1.73
84.81
11.00
9.05
80.82
7
1.98
84.07
9.90
5.70
67.56
3
2.16
54.59
10.01
0.00
15.45
1
2.29
67.30
11.86
0.00
10.96
Glucose concentration (g/L)
Enzymatic digestibility (%)
YEH (%)
Overall sugar yield (%)
10
-0.12
7.35 ± 0.14
36.21
31.44
18.28
13
0.06
6.23 ± 0.27
30.55
26.98
15.97
8
0.79
8.17 ± 0.36
39.87
32.20
20.71
12
1.32
20.11 ± 0.76
75.49
63.08
35.36
5
1.40
14.34 ± 0.70
50.68
44.06
31.13
4
1.44
16.68 ± 0.77
58.56
49.29
33.22
9
1.47
17.18 ± 0.15
62.03
50.31
35.10
6
1.55
17.72 ± 0.70
58.49
51.59
34.75
11
1.56
17.88 ± 0.13
64.85
52.01
34.69
2
1.73
17.78 ± 0.30
59.47
50.44
37.13
7
1.98
18.42 ± 0.79
59.53
50.04
34.01
3
2.16
18.96 ± 0.47
93.09
50.82
24.10
1 2.29 18.11 ± 0.20 71.04 47.79 22.81 WIS: water insoluble solids; Enzymatic digestibility: g glucose by enzymatic hydrolysis/100 g glucose in WIS; YEH: g glucose/100 g glucose in rapeseed straw; Overall sugar yield: sum of glucose released by enzymatic hydrolysis and sugars in the prehydrolysate/100 g rapeseed straw.
26
Table 5. Analysis of variance for responses (HSRL and YEH). Source Model Temperature (T) FeCl3 concentration (C) TC T2 C2 Residual Lack of Fit
Sum of p-value F-value Remarks squares (Prob > F) Hemicellulosic sugar recovery (HSRL) 6492.47 38.14 0.0002 Significant 352.25 10.35 0.0182 10.42 0.31 0.6001 1956.29 57.47 0.0003 3849.67 113.08 < 0.0001 155.80 4.58 0.0762 204.26 94.32 Not significant
SDb C.V. %c
5.83 10.97
Model Temperature (T) FeCl3 concentration (C) TC T2 C2 Residual Lack of Fit
Enzymatic hydrolysis yield (YEH) 709.29 120.27 < 0.0001 Significant 183.05 155.20 < 0.0001 19.89 16.86 0.0093 33.17 28.13 0.0032 146.01 123.80 0.0001 19.61 16.62 0.0096 5.90 1.19 1.01 0.3726 Not significant
SDb C.V. %c
1.09 2.36
R-squared AdjR-squareda
R-squared AdjR-squareda
0.9695 0.9441
0.9918 0.9835
a
Adjusted R2 Standard deviation c Coefficient of variation b
27
Figure captions Fig. 1. Response surface for combined severity (CS) as a function of pretreatment temperature and FeCl3 concentration at 20 min. Fig. 2. Response surface curves representing the interactive effect of concentration of ferric chloride and pretreatment temperature on the concentration of acetic acid (a), furfural (b), HMF (c) and formic acid (d) in the prehydrolysates. Fig. 3. Response surface for hemicellulosic sugar recovery, HSR L (a) and enzymatic hydrolysis yield, YEH (b) as a function of pretreatment temperature and FeCl3 concentration at 20 min. Fig. 4. Material balance flow of the FeCl3 pretreatment of rapeseed straw under optimal conditions and its enzymatic hydrolysis
28
Fig. 1.
29
a)
c)
b)
d)
Fig. 2
30
a)
b)
Fig. 3
31
Fig. 4
32