Subcritical CO2 shows no effect on liquid hot water pretreatment of poplar wood

Pretreatment of lignocellulosic biomass is required for many biorefinery processes. Previous studies have described hydrolysis of hemicelluloses by using liquid hot water (LHW) pretreatment. We evaluated the effect of carbonic acid originating from pressurized carbon dioxide during LHW pretreatment of poplar. The conditions applied covered temperatures from 120 to 200 °C, pretreatment times from 5 to 240 min and pressures from 1.0 to 2.2 MPa CO2 or N2. The pressure and the type of gas (CO2 or N2) did not have an effect on production of acetic acid, which functioned as a marker of progress of biomass hydrolysis. Results suggested that the presence of carbonic acid in the process does not significantly contribute to acidification. Deacetylation of lignocellulosic biomass can be achieved by LHW pretreatment irrespective of pressure and of gas type used, at the conditions


Introduction
Considerable efforts have been made in the past decades towards investigating options to solve the fossil fuels depletion in combination with the increasing energy consumption and CO 2 emissions by our society (Goldemberg, 2007). This situation has led to a path where biorefineries are promising options, picturing a scenario with industrial activities using bio-based sources and biological operations as main trend (Hassan et al., 2019).
Potential biomass sources are side streams from well-established processes, waste and excess production from agricultural activities, forestry, and crops grown specifically for biorefinery purposes (Kim and Dale, 2004). Lignocellulosic biomass has been investigated intensively, with special interest in the disruption of its structure in order to increase the digestibility of the compounds present in its matrix (Hendriks and Zeeman, 2009;Kim and Holtzapple, 2006;Kumar and Sharma, 2017).
Lignocellulosic biorefinery processes typically consist of four main parts: pretreatment, enzymatic hydrolysis, fermentation and purification. Due to the complex matrix and structure of lignocellulosic biomass, the pretreatment and enzymatic hydrolysis steps usually produce derived substances, apart from the desired and expected sugars, which could limit the activity of the microorganisms and reduces the efficiency during the fermentation (Liu et al., 2019).
In this paper, we focus on pretreatment, which hydrolyses hemicelluloses and improves accessibility to cellulose for subsequent enzymatic hydrolysis (Kumar et al., 2009). Several pretreatment methods have been extensively studied, involving different technologies (Alvira et al., 2010;Mosier et al., 2005b). By using high temperatures and pressures in liquid hot water (LHW) pretreatment, it is possible to hydrolyse and dissolve hemicelluloses, while partially solubilising the lignin with almost no influence on the cellulose fibres (Garrote et al., 1999;Liu and Wyman, 2005;Zhuang et al., 2016). Optionally, acids or other catalysts can be applied to shorten reaction time or decrease temperature and pressure conditions (Cybulska et al., 2013;Esteghlalian et al., 1997;Hu and Ragauskas, 2012). For a green process, one would like to take out and reuse such catalysts. However, acid catalysts usually have to be neutralised to allow subsequent enzymatic hydrolysis, leading to alkali costs and to waste salts that can hardly be recovered. Therefore, using carbon dioxide as volatile (and hence recyclable) acid catalyst has been proposed as a green alternative pretreatment method, due to decrease on pH promoted by the production of carbonic acid (Magalhães da Silva et al., 2014;Morais et al., 2015). The use of supercritical CO 2 has been extensively studied for pretreatment purposes (Alinia et al., 2010;Kim and Hong, 2001;Luterbacher et al., 2010;Narayanaswamy et al., 2011). It has been reported that supercritical CO 2 easily penetrates cellulose such that explosion caused by depressurization increases accessible surface area in cellulose for enzymatic hydrolysis. This is an alternative to steam https://doi.org/10.1016/j.biteb.2020.100442 Received 17 February 2020; Received in revised form 4 May 2020; Accepted 4 May 2020 explosion, but requires no increase in temperature (Zheng et al., 1995). Compared to inert gases (nitrogen, helium), carbon dioxide showed more effective glucose yield when pretreating cellulose (Zheng et al., 1998). However, the use of supercritical CO 2 entails certain limitations at industrial scale due to the level of pressure utilized (Agbor et al., 2011).
At pressures below critical (7.4 MPa), CO 2 use can be considered a modification of LHW pretreatment. In the literature, however, there is no agreement on the influence of subcritical CO 2 . Rogalinski et al. (2008) found no change in the solubilisation of lignocellulosic material by using CO 2 during LHW treatment. In addition, McWilliams and van Walsum (2002) also reported no different outcome on xylose recovery when carbonic acid was present during LHW treatment of aspen wood. On the other hand, other authors noticed enhancement of corn stover hydrolysis from higher values of xylose and xylan oligomers obtained in presence of CO 2 (van Walsum and Shi, 2004), and even a kinetic model has been proposed for the hydrolysis of wheat straw in water at high pressure CO 2 (Relvas et al., 2015). One reason for these discrepancies is that experimental conditions varied, including the type of biomass used, and another is that the performance of pretreatment is judged from the concentration of unstable reaction products or from the performance of subsequent enzymatic hydrolysis.
To understand better the influence of CO 2 on LHW pretreatment, we aim to quantify pretreatment on basis of acetic acid formation. Acetic acid originates from cleavage of acetyl groups in hemicelluloses, and undergoes no subsequent reaction. Most of the previous studies on LHW pretreatment are focused on the monosaccharides release (mainly xylose) as preeminent indicator of the degradation of hemicelluloses (Laser et al., 2002;Mosier et al., 2005a;Negro et al., 2003;Pérez et al., 2008). In this work, acetic acid is considered as predominant indicator of hemicellulose hydrolysis. It is a stable compound, and the main compound contributing to acidity in conventional LHW pretreatment.
For the current study, poplar wood was selected as relevant lignocellulosic biomass type, because it has been widely studied already and has a high acetyl content. Release of acetyl groups from hemicelluloses and depolymerisation of hemicellulose typically occur at 170-190°C, whereas cellulose hydrolysis occurs at temperatures above 200°C (Thomsen et al., 2006;Weil et al., 1997). Focusing on hemicellulose hydrolysis, and therefore acetic acid release, the conditions applied in this work during this hydrothermal pretreatment cover temperatures between 120 and 200°C, pretreatment times from 5 to 240 min and subcritical pressures of CO 2 or N 2 up to 2.2 MPa. Compressed nitrogen is designated as inert gas, in order to provide a comparable reference and assess the possible catalytic effect of carbon dioxide on acetic acid release during LHW pretreatment.

Materials
Acetic acid (≥99.7%) was from Sigma-Aldrich. The lignocellulosic material (poplar wood with stems and bark) was provided by Woodoo GmbH & Co. KG, through the Bioprocess Pilot Facility B.V. Samples were pre-milled in a bench mill and sieved to a mesh size of~1 cm.

Pressurized reactor
A pressurized reactor was used, which consists of a stainless steel 100 mL vessel with an electrical heating and an agitation system. The reactor was connected to independent inlet lines for CO 2 and N 2 , and to two outlet lines for gas, one of them with a safety relief valve set at 2.2 MPa. The reactor is depicted in Fig. 1.
The reactor vessel was filled with 5 g of biomass (on wet basis) and 45 g of water. Insulation fabric was placed around. Before starting the experiments, all pipes and valves were shortly flushed with CO 2 or N 2 , assuming negligible evaporation of water. Then, agitation was switched on and the pressure of gas (CO 2 or N 2 ) was progressively increased until the desired "initial pressure" was reached in the reactor vessel. After 30 min (estimated time to reach liquid/gas equilibrium), the heating system was switched on, increasing the temperature inside the vessel up to the set point (see Supplementary data for heating times), which increased the pressure, up to the "pretreatment pressure" value. The reactor was kept at the set temperature by the temperature controller during a time that will be indicated as "pretreatment time". Subsequently, the heating system was switched off and the insulating fabric was removed from the reactor, allowing a fan to speed up cooling of the vessel (see Supplementary data for cooling times). Once the reactor was cooled to room temperature, the outline gas valve was opened to proceed with decompression. When ambient pressure was reached, the reaction suspension was centrifuged and filtered (Whatman 0.2 μm), and the pH of the liquid fraction was measured. Several liquid samples were taken and stored at −20°C until further analysis by HPLC, while the solids were discarded.

Analysis of samples
Acetic acid was analysed on a Waters HPLC system with a Bio-Rad HPX-87H column (300 × 7.8 mm) at 59°C. Phosphoric acid (1.5 mmol/L at 0.6 mL/min) was used as eluent. Quantification was by refraction index detection (Waters 2414) using external standards. Volumes of injection of standards and samples were 10 μL at 15°C. All samples were analysed in duplicate for acetic acid content.

Severity factor
Overend and Chornet (1987) have proposed a severity factor (log R 0 ) to combine parameters applied to a pretreatment process: T is the temperature in°C and t is the pretreatment time in minutes. This severity factor is used to represent the results of the present study.

Effect of severity
During LHW treatment, in presence of CO 2 or N 2 , different combinations of temperature and pretreatment time were applied to the lignocellulosic biomass, associated in the severity factor of the reaction conditions (see Table 1). Conditions applied covered severity values from 0.7 to 5.6 (log R 0 ). Fig. 2A shows that with increasing severity the concentration of acetic acid increased up to approximately 5.3 g/L, regardless the gas used. This maximum acetic acid concentration was obtained at log R 0 of 5.17, corresponding to 200°C and 170 min. For beech wood LHW pretreatment, Nitsos et al. (2013) reported a similar trend at a severity range (log R 0 ) between 2.0 and 5.0, obtaining up to 3.4 g/L acetic acid. Analogous results were also obtained by Morais et al. (2014) when processing wheat straw, although they included higher pressures of CO 2 (closer to the critical point) within a narrower range of severity.
Furthermore, our results are in line with the absence of significant acetyl group release for LHW pretreatment of poplar below log R 0 values of 3.5 (Bouchard et al., 1991) and the trend observed for carbonated subcritical water pretreatment of switchgrass at higher temperatures (Dhamdere et al., 2012). Comparing results from this work with literature values of acetyl content of poplar (Kim et al., 2009) suggests complete deacetylation of poplar at values of log R 0 > 5. However, that was not confirmed since the structure and composition of the solid fraction of the pretreated biomass was not analysed in this work. At less severe conditions, some acetyl groups remained in the solid phase linked to hemicelluloses, while part of it was bound to xylooligosaccharides Kabel et al., 2007).
Like generally observed in studies on biomass pretreatment, higher severity promoted not only acetic acid release, but also release of other hydrolysis products such as xylose and glucose, and of their degradation products such as formic acid, furfural and 5-hydroxymethylfurfural (Kim et al., 2009). Data for these other compounds are available in the Supplementary material but will not be discussed, because the observed patterns are in line with those generally described in pretreatment literature and here. Still other components, such as lignin, were not measured, though.

Effect of pressure and gas type
Several initial subcritical pressure values of CO 2 and N 2 were applied (0.2, 0.4, 0.6, 0.8 and 1 MPa) at otherwise the same reaction conditions: 180°C during approximately 2 h (log R 0 ≈ 4.4). Pressures increased once the reactor was heated up, slightly more in case of CO 2 than in case of N 2 , while the concentrations of acetic acid obtained in all these experiments were similar (see Fig. 3A). This outcome was opposite to those of some other studies in which hydrolysis of biomass increased in the presence of CO 2 , although for different biomass types and at higher but still subcritical pressure values Relvas et al., 2015;van Walsum and Shi, 2004;Zhang and Wu, 2013). Fig. 3A shows that the gas type (CO 2 or N 2 ) and its pressure during pretreatment have no significant effect on the hydrolysis of poplar. Fig. 2B shows a decrease in pH with severity, as expected on basis of increased carboxylic acid amounts at higher severity (Jacobsen and Wyman, 2002). These values of pH were measured at room temperature, after cooling and filtering the liquid samples. Again, no difference is observed between using carbon dioxide and nitrogen during pretreatment, and pressure has no effect (Fig. 3B). During LHW pretreatment of lignocellulosic biomass, the auto-ionization of water produces hydronium ions, which, in combination with the high temperatures, provokes hydrolysis of hemicelluloses, and therefore cleavage of acetyl groups, producing acetic acid. This acetic acid, and other carboxylic acids released such as uronic acids, contributes to the hydrolysis reaction, widely known as autohydrolysis Gírio et al., 2010). In addition, those carboxylic acids contribute to subsequent carbohydrate degradation towards other side-products (Mosier et al., 2005b). Although acetic acid is not yet present in the beginning of pretreatment, in previous studies of hydrothermal treatment, the hydronium ions generated from acetic acid were considered more important than those of water origin (Garrote et al., 1999;Gírio et al., 2010). The use of carbon dioxide during this process has been mentioned to enhance the hydrolysis by further lowering the pH due to the formation of carbonic acid (Luterbacher et al., 2012;Magalhães da Silva et al., 2014;Morais et al., 2015). Some authors suggested that the autocatalysis due to release of acetyl groups has similar influence as the presence of carbonic acid (McWilliams and van Walsum, 2002), while other researchers suggested that effect of dissociation of carbonic acid in water combined with the release of acetic acid from deacetylation of hemicelluloses is crucial for the hydrolysis of lignocellulosic biomass (Gurgel et al., 2014). At our conditions, CO 2 clearly plays no role in acidification.

Effect on pH
At ambient conditions, an aqueous solution of 5 g/L acetic acid is at pH 3.0, according to pK a = 4.82. The minimum pH in Fig. 2B is 3.4, corresponding to 0.5 g/L acetic acid and thus indicating the presence of buffering compounds. Their source might be minerals or proteins originating from the biomass. Other authors van Walsum and Shi, 2004) found somewhat higher final values of pH, probably because their biomass contained less acetyl groups or more buffering compounds. At final (ambient) conditions, carbon dioxide does not contribute to low pH because of its low solubility. At 225°C and 5.4 MPa initial CO 2 pressure, it has been calculated that the pH of an aqueous CO 2 solution is 3.77 (van Walsum and Shi, 2004), suggesting that acetic acid is a stronger contributor to low pH than CO 2 .
As a consequence, improvement of lignocellulosic biomass pretreatment should occur due to addition of acetic acid prior to LHW pretreatment. However, the literature does not report a clear picture on this (Laser et al., 2002;Lyu et al., 2019). Therefore, further research regarding the effect on acetyl group cleavage on hemicelluloses by initial addition of acetic acid to LHW pretreatment of lignocellulosic biomass could provide deeper understanding of this methodology. Besides, systematic studies of different reaction conditions and different biomass types might reveal why different studies find different effects of CO 2 .

Conclusions
Liquid hot water pretreatment at log R 0 > 5 is an efficient methodology for deacetylation of hemicelluloses from poplar. Acetic acid released was proportional with the severity of the process but it has no direct relation with the pretreatment pressure up to 2 MPa, nor with the gas used for achieving this pressure. At the conditions applied, subcritical CO 2 does not affect acetic acid formation. Its role in acidification, even associated to the initial phase of pretreatment, is negligible. At different reaction conditions and with different types of biomass than tested in this work, CO 2 might still improve pretreatment.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.