Fractionation of Pinus radiata wood by combination of steam explosion and organosolv delignification

The environmental degradation due to greenhouse gas emissions, the increasing energy demand and the progressive depletion of fossils fuels have strengthened the interest in alternative, renewable resources to meet society basic needs of products and services (Alvira et al. 2010). Second generation biorefineries are a promising alternative to the first generation based on sugar plants or cereals. A forest biorefinery is a facility that can generate fuels, energy, and chemical products from lignocellulosic biomass from the primary processing of wood (e.g. sawdust).

However, while lignocellulosic biomass can be used as raw material for producing a wide variety of high valued products, its processing cost must compete with those manufactured from petroleum. The optimal use of lignocellulosic biomass is of paramount importance to make it profitable in the context of a biorefinery (Area and Vallejos 2012). Identifying economically feasible and technically viable strategies for softwood fractionation represents a huge challenge since softwoods are known to be more recalcitrant than hardwoods and other lignocellulosic raw materials.

Pretreatment of lignocellulosic biomass is conducted to improve the lignocellulose accessibility and subsequent enzymatic hydrolysis by removing hemicelluloses and/or lignin, increasing porosity and disrupting the crystalline structure of cellulose (Chiaramonti et al. 2012). There are numerous physical, chemical, physicochemical and biological pretreatments to accomplish the fractionation of forest-industrial wastes into their main components. The type and severity of the selected method are assessed as a function of biomass nature and chemical compositions, involving cost, carbohydrates recovery, and uses of the obtained material (Maity 2015). The pretreatment has been shown to be a critical stage in cellulosic ethanol production, and its impact on the process as a whole is especially highlighted when substrates such as softwoods are used (Galbe and Zacchi 2012). For any pretreatment, the energy/cost involved must not be greater than the energy/benefit obtained from the biomass. Therefore, many strategies have been studied to achieve proper softwood fractionation, including alkaline pulping, steam explosion, organosolv treatment with different solvents, and oxygen delignification, among others (Stoffel et al. 2014, Imlauer et al. 2014, Ewanick et al. 2007).

Steam explosion has been classified as a physicochemical pretreatment, as it introduces morphological and chemical variations of the lignocellulosic biomass like the breakdown of the lignocellulosic structure, the hemicelluloses hydrolysis, and the depolymerization of lignin (Singh et al. 2014). During the steam explosion process, the biomass is heated with high pressure saturated steam (160-180°C) for relatively short periods, followed by a sudden decompression. This pretreatment has a series of advantages: low environmental impact and energy consumption, limited capital requirements and low chemical loads, among others (Wang et al. 2015). Moreover, it has proven to be an effective pretreatment for a wide range of biomass feedstock, though its efficiency is lower for softwoods (Sun and Cheng 2002, Singh et al. 2015). Acid catalyst addition, usually H₂SO₄ or SO₂, is required as they allow shortening the residence time and decreasing the temperature of the treatment (Brodeur et al. 2011).

The amount of lignin present in pretreated softwood has been reported to play a crucial role during enzymatic hydrolysis (Berlin et al. 2006, Chandra et al. 2009). Therefore, a delignification stage is required to enhance the hydrolysis of the cellulose rich fraction obtained after SE. In organosolv pulping, delignification of the biomass takes place using an organic solvent or an organic solvent/water mixture. On the basis of a biorefinery concept, organosolv pulping has been evaluated as a potential pretreatment method for lignocellulosic biomass conversion into valuable products (Bozell et al. 2011). The main drawbacks with the use of alcohols are the high pressures generated during the treatment and the cost associated with the safety measures required for their manipulation (Oliet et al. 2002). In particular, methanol and ethanol are the two most frequently used solvents in organosolv pulping (Zhao et al. 2009). Both low boiling point alcohols improve the cooking liquor penetration into the structure of wood, resulting in a uniform and increased delignification. In addition, they prevent lignin from condensing (Muurinen 2000). Organosolv pretreatment with ethanol has been assessed using softwoods as raw material. Some studies have been carried out to achieve desirable enzymatic hydrolysis yields from softwood material treated with alcohols under several conditions. Moreover, several catalysts have been tested for organosolv pretreatment of softwoods to enhance the hydrolysis yields (Park et al. 2010).

In the present work, a sequential combination of steam explosion and organosolv pulping is proposed to evaluate steam explosion impact on the enhancement of Pinus radiata delignification and enzymatic hydrolysis to obtain second-generation bioethanol.

The chemical composition of P. radiata used as raw material is summarized in Table 2.

Table 2:
Pinus radiata chemical composition.

Results of SE and EW delignification are shown in Table 3.

Table 3:
Steam Explosion and Ethanol-Water delignification of P. radiata chips.

Steam explosion decreased as the severity factor increased due to the solubilization of the material, varying from 93,3% to 76,8% (Table 3). These values are higher than those obtained by other authors in catalyzed steam explosion of softwoods (Cotana et al. 2014, Stoffel 2016).

The solid fraction obtained as a result of the steam explosion pretreatment consisted mainly of cellulose and lignin. The content of hemicelluloses significantly decreased because of their selective removal during the pretreatment. SE delignification was poor and ranged from 5,2% to 24% for all experiments, but the percentage of lignin in the fibrous solid increased due to the extraction of hemicelluloses. The highest delignification was obtained with the two-cycle treatment. Although steam explosion does not significantly remove lignin, this process has proven to disrupt the cell wall, weakening the lignin-carbohydrate complex (Wang et al. 2015). There is also evidence of lignin redistribution in dense particles within the cell wall, modifying the porosity of the material and the enzyme accessibility (Donaldson et al. 1988). Moreover, studies on chemical changes of steam-exploded lignins have revealed depolymerization and repolymerization reactions as a consequence of the acidity of the medium (Li et al. 2007, Laureano-Perez et al. 2005, Trajano et al. 2013, Shevchenko et al. 1999). There is an increase in the delignification extent as the SE severity increases, up to a point where condensation reactions prevail. Both physical and chemical changes of the lignin present in the steam exploded material would have a direct impact on downstream processes, but the influence of the SE severity on delignification depends on the type of post-treatment (Kumar et al. 2011).

Studies conducted on the chemistry of the organosolv process have suggested that α-ether linkage cleavage is the most important reaction in lignin breakdown, especially in acidic systems (Aziz and Goyal 1993). Control EW experiments (non-exploded materials) showed poor delignification values, in contrast with previous studies of organosolv pulping of softwoods species, including P. radiata (Donaldson et al. 1988, Araque et al. 2008, Pan et al. 2008). Lignin removal in the EW stage decreased with increasing time, from 18,8% (60-minute treatment) to 16,8% (90-minute treatment). Since during acidic organosolv pulping a reactive intermediate is formed as a consequence of α-ether cleavage of lignin (McDonough 1992), this result may suggest that lignin undergoes a series of condensation reactions when the duration of the treatment is extended under the conditions tested. These results agree with another study in which high severity conditions in acidic organosolv treatment (acetone: water 50:50% (v/v), sulfuric acid 0,9% w/w, 195°C, 40 min) produced the lowest delignification (20,6%) (Ewanick et al. 2007).

Delignification in the shortest organosolv treatment decreased with the severity of the SE, from 19,8% to 16%. Conversely, the highest delignification was achieved for the material from the two-cycle SE in the case of the 90-minute organosolv treatment. However, the global delignification always increased when severity increased, but the highest delignification was obtained when combining two SE cycles. SE performed at the lowest tested pressure but in two cycles showed between 30 and 40% higher delignification percentages than the one cycle experiments. The combination of two SE cycles, in comparison with only one cycle of equivalent duration clearly improves delignification and yields a solid with a global delignification of about 55%. Delignification time did not have a significant effect on yield.

Enzymatic hydrolysis was performed to evaluate the effect of the steam explosion on enhancing the accessibility of enzymes to the cellulose-rich material. Samples were taken after 72 h of incubation and the conversion of cellulose to glucose (hydrolysis yield) was determined. Steam exploded materials and the substrates obtained from the combination of SE and EW pulping were subjected to this procedure. Hydrolysis yields are shown in Figure 1.

Figure 1:
Enzymatic hydrolysis of the cellulose-rich residues vs SE conditions.

The hydrolysis yields of the steam exploded materials from all SE conditions ranged between 8,9% and 11,6% (corresponding to the single 10-minute cycle). The high lignin content of these materials could explain the poor hydrolysis yields, as it has been proven that lignin form non-productive bonds with enzymes (Kumar et al. 2012, Rahikainen et al. 2013). H₂SO₄-catalyzed SE of pine sawdust achieved hydrolysis yields of about 24% under severe conditions (200ºC, 5 min and 3% w/w H₂SO₄) (Stoffel et al. 2017). However, these results are significantly lower than those of other authors who obtained cellulose to glucose conversions of 77% using a high severity SO₂-catalyzed SE (215°C, 5 min, 4,5% w/w SO₂) with beetle-killed lodgepole pine chips (Ewanick et al. 2007). Presumably, the combination of high severity and the addition of SO₂ produced further disruption of the lignocellulosic matrix and improved hemicellulose removal. Similar results were obtained by other researchers working with higher severities than those used in the present work and/or with acid catalyst (Söderström et al. 2003, Cotana et al. 2014), (Asada et al. 2015).

The two-cycle steam-exploded material, which experienced the higher level of delignification, did not produce the most hydrolysable residue, what could be due to the removal of the most accessible cellulose.

Many studies carried out on softwood samples have suggested that a further treatment of the cellulose rich fraction is required to achieve good hydrolysis yields (Kumar et al. 2010, Kumar et al. 2011, Pan et al. 2004, Kumar et al. 2012), whereas others have highlighted that readily hydrolyzed material is obtained at the expense of high sugar loss (Wu et al. 1999, Söderström et al. 2003, Yang et al. 2002).

EW delignification significantly increased the hydrolysis yield of the steam exploded materials, even though the combination of both treatments did not produce an easily hydrolyzable solid. Hydrolysis yields did not exceed 35% and, once again, higher delignification did not necessarily correspond with higher EH yield. These results are considerably inferior to those obtained for Douglas-fir chips subjected to the SO₂-catalyzed steam explosion followed by an alkali-oxygen delignification stage (90% EH yield) (Pan et al. 2004). This sharp difference could be due to a more extensive delignification of the steam pretreated material and the fact that an inverse non-linear correlation between lignin content and hydrolysis susceptibility has been previously observed (Pan et al. 2005, Kruyeniski et al. 2016). The addition of an SE pretreatment has a positive effect on the EH when comparing the results herein with those obtained after a single organosolv treatment of Pinus taeda and Pinus elliottii sawdust mix (below 10%) (Kruyeniski et al. 2015). The disruption of the lignocellulosic structure that takes place during SE may be the reason why higher EH yields are attained with this work strategy.

Results reported in the literature of EH yields after SE treatments applied to different raw materials are shown in Table 4.

Table 4:
Hydrolysis yields of different raw materials subjected to steam explosion pretreatment.

Steam explosion has proved to be a good option as pretreatment for non-woody lignocellulosic materials like wheat straw and sugarcane bagasse (Cui et al. 2012, Neves et al. 2016) (Table 4). Also, EH yield of steam exploded Eucalyptus globulus (log R₀=3,56) was 24,7% (Martín-Sampedro et al. 2012) whereas in this work, Pinus radiata sawdust pretreated with similar severity (log R₀=3,35), did not exceed 12%. Moreover, acid-catalyzed steam explosion of Eucalyptus grandis chips achieved a EH yield of about 57% (Emmel et al. 2003), twice as high as the one recorded for Slash pine sawdust obtained under similar conditions (Stoffel et al. 2017). These differences could be related to the amount and type of lignin present in softwoods (Asada et al. 2015). In a study on isolated lignins from pretreated corn stover, poplar, and lodgepole pine on the EH, Nakagame et al. (2010) found that softwood lignin has a stronger negative influence on EH than that of other sources. Also poplar chips (Table 4) proved to render a more readily hydrolysable material (over 85% cellulose to glucose conversion) (Panagiotopoulos et al. 2013) than Pinus radiata sawdust (up to 35%), both pretreated under a similar SE severity. The addition of an acid catalyst may be suggested as an explanation for the above-mentioned difference. Nevertheless, other studies on softwood did not report hydrolysis yields higher than 36% for Slash pine and 68% for Douglas-fir (Stoffel et al. 2016a, Kumar et al. 2011). Steam explosion is one of the best processes to extract hemicelluloses from softwoods, but it can be negative for the digestibility of the remaining solid fraction, highly dependent on the quantity and quality of lignin. Stoffel et al. (2017) extracted 90% of the hemicelluloses with the steam explosion pretreatment but the hydrolysis yields of the resulting cellulosic fraction were poor (24,3%), (Table 4). Even alkaline delignifying post-treatments (Stoffel et al. 2016a) did not prompt a significant improvement of the EH yield, in agreement with our results. The chemical and structural changes produced in lignin during the steam explosion are presumably responsible for the decrease in the ability of delignifying post-treatments (Pielhop et al. 2015). High steam explosion severities prompt lignin condensation reactions which counteract the depolymerization reactions. This applies to softwood lignins, with a high content of derivatives of guaiacyl-propane structures, which are more susceptible to condensation reactions than the syringyl-propane ones, predominant in hardwood species (McDonough 1992). Pielhop et al. (2015) concluded that repolymerization reactions, especially important in softwoods lignins, modifies lignin nanostructure in a way that significantly increases its specific surface area. Therefore, condensed lignin in substrates from autohydrolysis pretreatments represent an obstacle to EH. However, lignin removal does not necessarily correlate with the ease of hydrolysis. Lignin redistribution during steam explosion treatment, and lignin nature because of its susceptibility to depolymerization/condensation reactions and hydrophobicity, have a significant effect on downstream EH (Kumar et al. 2011), The correlation between lignin hydrophobicity and cellulases adsorption (hydrophobic interaction) has been previously suggested in the literature (Berlin et al. 2006). i.e., a higher hydrophobicity of lignin in the pretreated substrate could mean a higher interaction with cellulases (non-productive binding) and, thus, a decrease in EH yield. Compared with other pretreated lignocellulosic materials, Nakagame et al. (2010) findings suggested that isolated softwood lignins show higher hydrophobic behavior.

The literature in Table 4 shows that, even when higher hydrolysis yields may be obtained from softwood than those recorded in this work, they are only comparable to those obtained from steam exploded hardwood or nonwoody lignocellulosic materials when using high severity factors with the aid of an acid catalyst or/and a post-delignification stage, proving softwood recalcitrant nature. This work demonstrated that the sequential application of SE followed by an EW delignification for Pinus radiata sawdust did not significantly enhance the digestibility of the cellulosic residue obtained during the SE. These results agree with other authors, who noticed that softwoods subjected to autohydrolysis or acid-catalysed pretreatments exhibit a recalcitrant behavior towards downstream processes, such as delignification and enzymatic hydrolysis.

A sequential combination of steam explosion and organosolv post-treatment was performed to evaluate the impact of the steam explosion pretreatment on Pinus radiata delignification and subsequent digestibility.

Low delignification was recorded after the steam explosion pretreatment. Moreover, a further delignification stage (organosolv post-treatment) did not considerably remove the lignin content of the steam-exploded material, presumably due to the chemical changes underwent by softwood lignins. The maximum global delignification corresponded to materials subjected to the explosion stage in two-cycles (54,7%).

EH yields of the materials obtained after the steam explosion and the organosolv treatment did not exceed 35%. Though the tested strategy improved the exploded material digestibility, it did not allow to obtain a readily hydrolyzable material.

The recalcitrant nature of the tested raw material, in addition to the presumably lignin chemical modifications and redistribution within the cellulose-rich fraction after the treatments, prove to have a synergetic negative effect on the ease of hydrolysis of the resulting material.

The authors deeply acknowledge PROVALOR network (Productos de valor agregado a partir de residuos agro y forestoindustriales), CONICET (National Council of Scientific and Technical Research), UNaM (National University of Misiones), MINECO (Project CTQ 2010-17702) and the Madrid Regional Government (Project RETO PROSOST P2013-MAE2907) for the financial support.