Agave tequilana bagasse for methane production in batch and sequencing batch reactors: Acid catalyst effect, batch optimization and stability of the semi-continuous process

Journal of Environmental Management 224 (2018) 156–163

Contents lists available at ScienceDirect

Journal of Environmental Management journal homepage: www.elsevier.com/locate/jenvman

Research article

Agave tequilana bagasse for methane production in batch and sequencing batch reactors: Acid catalyst effect, batch optimization and stability of the semi-continuous process

T

Luz Breton-Devala, Hugo O. Méndez-Acostaa, Víctor González-Álvareza, Raúl Snell-Castroa, Daniel Gutiérrez-Sáncheza, Jorge Arreola-Vargasa,b,∗ a b

Departamento de Ingeniería Química, CUCEI-Universidad de Guadalajara, Blvd. M. García Barragán 1451, C.P. 44430, Guadalajara, Jalisco, Mexico División de Procesos Industriales, Universidad Tecnológica de Jalisco, Luis J. Jiménez 577-1° de Mayo, C.P. 44979, Guadalajara, Jalisco, Mexico

A R T I C LE I N FO

A B S T R A C T

Keywords: Acid hydrolysis Biogas Biomass pretreatment Central composite design Lignocellulosic biomass Response surface methodology

Agave tequilana bagasse is the main solid waste of the tequila manufacturing and represents an environmental issue as well as a potential feedstock for biofuel production due to its lignocellulosic composition and abundance. In this contribution, this feedstock was subjected to pretreatments with HCl and H2SO4 for sugar recovery and methane was produced from the hydrolysates in batch and sequencing batch reactors (AnSBR). Sugar recovery was optimized by using central composite designs at different levels of temperature, acid concentration and hydrolysis time. Results showed that at optimal conditions, the HCl pretreatment induced higher sugar recoveries than the H2SO4 one, 0.39 vs. 0.26 g total sugars/g bagasse. Furthermore, the H2SO4 hydrolysate contained higher concentrations of potential inhibitory compounds (furans and acetic acid). Subsequent anaerobic batch assays demonstrated that the HCl hydrolysate is a more suitable substrate for methane production; a fourfold increase was found. A second optimization by using HCl as acid catalyst and methane production as the response variable demonstrated that softer hydrolysis conditions are required to optimize methane production as compared to sugar recovery (1.8% HCl, 119 °C and 103min vs. 1.9% HCl, 130 °C and 133min). This softer conditions were used to feed an AnSBR for 110 days and evaluate its stability at three different cycle times (5, 3 and 2 days). Results showed stable reactor performances at cycle times of 5 and 3 days, obtaining the highest methane yield and production at 3 days, 0.28 NL CH4/g-COD and 1.04 NL CH4/d respectively. Operation at shorter cycle times is not advised due to microbial imbalance.

1. Introduction The current global energy supply is mainly based on fossil fuels, which are nonrenewable resources and the main cause for the accumulation of greenhouse gases in the atmosphere (Monlau et al., 2014). Several governments around the world have encouraged the introduction of alternative energy sources that are both environmentally friendly and renewable. In this context, due to its abundance, composition and renewability, lignocellulosic biomass is recognized as a promising energy source for the production of biofuels such as methane (Kumar et al., 2009). Globally, large amounts of lignocellulosic biomass are generated as byproducts of agro-industrial processes. For instance, in México the tequila industry produces ∼380,000 tons per year of Agave tequilana

bagasse (CRT, 2016; Saucedo-Luna et al., 2010). Currently, most of this bagasse is disposed in agave fields causing environmental damage, such as pollution by leachates, odor generation and habitat for pest and diseases (Crespo et al., 2013). The main limitation to use the A. tequilana bagasse as feedstock for methane production is the low conversion rate due to the lignocellulosic recalcitrance (Fig. S1), making necessary to apply pretreatments in order to release the sugars from the hemicellulose and cellulose fractions (Arreola-Vargas et al., 2015). In this context, dilute acid hydrolysis is the most reported method for depolimerization of lignocellulosic biomass because of its effectiveness and low cost (Kumar et al., 2009). This pretreatment can depolymerize the holocellulose to its main sugars: hexoses (glucose, mannose, and galactose) and pentoses (xylose and arabinose); which are suitable substrates for methane

∗ Corresponding author. División de Procesos Industriales, Universidad Tecnológica de Jalisco, Luis J. Jiménez 577-1° de Mayo, C.P. 44979, Guadalajara, Jalisco, Mexico E-mail addresses: [email protected], [email protected] (J. Arreola-Vargas).

https://doi.org/10.1016/j.jenvman.2018.07.053 Received 29 January 2018; Received in revised form 11 June 2018; Accepted 15 July 2018 0301-4797/ © 2018 Elsevier Ltd. All rights reserved.

Journal of Environmental Management 224 (2018) 156–163

L. Breton-Deval et al.

production via anaerobic digestion. However, the main disadvantage of dilute acid hydrolysis is the formation of by-products (furans, week acids and phenolics) that are considered inhibitory to microorganisms at high concentrations (Monlau et al., 2014). Thus, optimization of hydrolysis conditions is required to maximize sugar recovery and minimize the generation of potential inhibitory compounds. Several studies have evaluated the acid pretreatment of different type of biomasses by mainly using hydrochloric (HCl) or sulfuric (H2SO4) acids as catalysts (Dussán et al., 2014; Kumar et al., 2009; Rodrigues et al., 2010). However, only few studies have aimed to compare the efficiency of these two acid catalysts over the same biomass (Hutomo et al., 2015; Meinita et al., 2015, 2012; Shi et al., 2012). Overall, these works suggest that H2SO4 is more effective than HCl for sugar recovery. For instance, Meinita et al. (2012) found higher sugar concentrations by pretreating marine algal biomass with H2SO4 as compared to HCl (38.45 vs. 22.68 g/L). Nonetheless, regarding A. tequilana bagasse, the results are inconclusive on this respect, since only individual efforts have been made to evaluate the acid hydrolysis of this feedstock under different conditions and type of acid. For instance, Saucedo-Luna et al. (2010) reported sugar concentrations up to 26.9 g/ L by using H2SO4, while we previously reported sugar concentrations up to 27.9 g/L by using HCl (Arreola-Vargas et al., 2015). Another aspect to consider on the selection of the acid catalyst is the downstream process. In the case of methane production, to the best of our knowledge there are no studies that compare the effect of HCl or H2SO4 hydrolysates over the anaerobic digestion process. This is relevant because chloride and sulphate ions present in the hydrolysates may cause negative effects during this process; for instance, by promoting the competence between methanogen archaea and sulphate reducing bacteria. On the other hand, regarding the full scale application of the A. tequilana bagasse hydrolysates for methane production, it is clear that anaerobic digestion studies on continuous or semi-continuous operation modes need to be addressed. In the current literature, there is only one study that has evaluated the use of a semi-continuous system, an anaerobic sequencing batch reactor (AnSBR) (Arreola-Vargas et al., 2015). Nonetheless, such study lasted only 15 days, making clear that the stability of the process needs to be evaluated for longer periods of time. It is worth to highlight that high methane yields were obtained in such study regardless of the short AnSBR operation (average value of 0.26 NL CH4/g COD). The latter is mainly due to the advantages of this reactor configuration, such as the high degree of process flexibility, the better control of the microbial population and the decoupling of the solids retention time (SRT) from the hydraulic retention time (HRT) (Arreola-Vargas et al., 2015; Mace and Mata-Alvarez, 2002). Due to all the above, this work aimed i) to optimize the hydrolysis conditions for maximizing sugar recovery from A. tequilana bagasse by using HCl and H2SO4 as acid catalysts, ii) to compare the efficiency of both acid catalysts for sugar recovery and methane production in batch assays, iii) to optimize the hydrolysis conditions for maximizing the methane production and compare such conditions with those obtained for sugar recovery optimization, and iv) to evaluate the long-term stability of an AnSBR operated at different cycle times for the production of methane from the acid hydrolysates.

Table 1 Natural and coded values in the central composite design for optimization of sugar and methane production. Variable

Coded symbol

Temperature (°C) Acid concentration (% w/w) Reaction time (min)

X1 X2 X3

Coded level -αa

−1

0

1

αa

83 0.3 19

100 1 60

125 2 120

150 3 180

167 3.7 221

α (axial distance) = 4√nf = 1.68, where nf is the number of experiments of the factorial design. a

2.2. Acid pretreatments Acid pretreatments were carried out by dispersing the A. tequilana bagasse at 5% (w/v) in HCl or H2SO4 solutions at different concentrations. The reaction took place in 1 L corning® bottles that were introduced in an oven for certain periods of time at different temperatures. The acid concentration, temperature and hydrolysis time varied according to the experimental design presented in the following section. At the end of both treatments, the hydrolysates were filtered through a 0.45 mm membrane for further analyses. 2.3. Experimental design The central composite design (CCD) is the most popular response surface design for optimization of desired variables. In this study, such design was employed to optimize the hydrolysis conditions (acid concentration, temperature and hydrolysis time) for maximizing the sugar recovery from A. tequilana bagasse by using HCl and H2SO4 as acid catalysts. Furthermore, the CCD was also employed to optimize the hydrolysis conditions for maximizing the methane production by using the best acid catalyst found in the former stage. Table 1 shows the natural and coded values of the independent variables for all the CCDs. It is worth mentioning that the CCD arises through sequential experimentation, i.e. an orthogonal first order design (e.g. 2k factorial design) is used to fit a first order model and once it exhibits lack of fit, axial points are added to complete the CCD and allow the quadratic terms to be incorporated into a second order model (Montgomery, 2013). The 2k factorial design must be augmented with central points in order to allow estimation of the experimental error without altering the orthogonality property of the design (our design was augmented with 5 central points). The first order model is represented in equation (1), where B0 is the constant coefficient, Bi is the linear coefficient; Y is the response variable, and Xi are the independent variables that have influence on the response variable.

Y = β0 +

∑ βi Xi +

ε

(1)

Once the first order model exhibited lack of fit, 6 axial points were added to the 2k factorial design to complete the CCD. The method of steepest ascent was employed when the 2k factorial design was well fitted to the first order model until no further increase on the response variable was observed. Then, a new 2k factorial design augmented with 5 central points was applied and the procedure continued until lack of fit to the first order model was observed. The second order model is represented in equation (2), where B0 is the constant coefficient, Bi is the linear coefficient, Bii is the quadratic coefficient, Bij is the interaction coefficient, Y is the response variable, and X are the independent variables that have influence on the response variable.

2. Materials and methods 2.1. A. tequilana bagasse A. tequilana bagasse was provided by a local distillery and was composed of 10.86% hemicellulose, 56.44% cellulose, 15.24% lignin and 17.46% extractives. The composition was determined by using a semiautomatic ANKOM fiber analyzer (Macedon, NY, USA) (ArreolaVargas et al., 2015). Prior to be subjected to any pretreatment, the fibers were dried at room temperature and reduced to an average length of 0.5 cm by using a blade mill.

Y = β° +

∑ βi x i + ∑ β iix i2 + ∑ ∑ βij x i x j

+ε

(2)

Data analysis, response surface plots and analysis of variance (ANOVA) were performed by using the software Statgraphics centurion XV (Statpoint, Technologies, Inc. USA). 157

Journal of Environmental Management 224 (2018) 156–163

L. Breton-Deval et al.

2.4. Anaerobic digestion experiments

3. Results and discussion

2.4.1. Batch assays HCl and H2SO4 hydrolysates obtained at optimal conditions for sugar recovery were characterized according to section 2.5 (analytical methods) and used as substrates in anaerobic batch assays to determine their effect on methane production. Subsequently, the acid catalyst that promoted the best response was subjected to a second hydrolysis optimization by considering methane production as the response variable. As previously mentioned, the initial conditions of acid concentration, temperature and time of hydrolysis were the same for methane optimization as those for sugar optimization (Table 1). All of the batch experiments were carried out in an automatic methane potential test system (AMPTS II-Bioprocess Control AB, Lund Sweden), which has the capacity for incubating 15 reactors in parallel, allowing to automatically measure the methane production at standard conditions. Batch assays were inoculated with 10 g VSS/L of anaerobic granular sludge (collected from a full-scale UASB reactor treating tequila vinasses effluents) and fed with 400 mL of hydrolysate at an adjusted concentration of 8 g COD/L. Prior to incubation, the initial pH was adjusted to 7.5 and the reactors were flushed with nitrogen gas to ensure anaerobic conditions. All the experiments were performed in triplicate at 35 °C with constant agitation (120 rpm).

3.1. Sugar recovery optimization and comparison of HCl and H2SO4 as acid catalysts Total sugar concentration in the hydrolysates from the different experiments of the CCD are presented in Table 2. Overall, the HCl hydrolysates outperformed the H2SO4 ones. It is worth mentioning that the proposed hydrolysis conditions in Table 1 exhibited lack of fit to the first order model (p < 0.05, Table S1) when HCl was used as acid catalyst, indicating curvature in the response and allowing to add axial points to the 2k factorial design to complete the CCD. However, the same proposed conditions in Table 1 were well fitted to the first order model when H2SO4 was used as acid catalyst (p > 0.05 for lack of fit, Table S2), indicating a linear behavior between independent and response variables. Therefore, the method of steepest ascent was employed and then a new 2k factorial design augmented with 5 central points was proposed (Table S3). This new conditions exhibited lack of fit to the first order model (p < 0.05, Table S4), allowing to add the axial points to the 2k factorial design to complete the CCD. Equations (3) and (4) show the simplest second order models that accurately represent the responses for the pretreatment of A. tequilana bagasse by using HCl and H2SO4 as acid catalysts, respectively.

Y = −117.846 + 1.5577X1 + 13.2087X2 + 0.3698X3 − 0,0059X12 − 3.3869X22 − 0.0014X32

2.4.2. AnSBR set up and operation Optimal hydrolysis conditions for methane production were used to produce hydrolysates and feed an AnSBR at a constant concentration of 10 g COD/L. The AnSBR was inoculated with 10 g VSS/L of anaerobic granular sludge and operated during 110 days, in which the stability and performance were evaluated at three different cycle times (5, 3 and 2 days). The AnSBR was made of polyvinyl chloride with a working volume of 1.2 L. The temperature of the reactor was controlled at 35 °C by using an internal water jacket, while the pH was regulated at 7.0 ± 0.05 by adding a 2N NaOH solution. Filling and discharging of the medium was carried out by means of peristaltic pumps. Each cycle consisted of the following stages: filling (10 min), reaction (from 5 to 2 days), settling (30 min) and discharging (8 min). In order to induce homogenous conditions during the reaction time, the reactor was mixed by means of a recirculation loop. The exchange ratio was 80% of the working volume, which means that practically all the liquid content was removed at the end of each cycle. The reactor was fully automatized, allowing the on line measurement of pH, temperature, pressure, biogas flow rate and biogas composition. Daily production of methane was calculated by reading the biogas flow rate (μFlow, Bioprocess Control AB, Lund Sweden) and methane fraction (BCP-CH4, BlueSens, Herten Germany). A National Instruments cRIO9004 device was used for acquisition, treatment, and storage of the data (Fig. 1).

(3)

Y = −41.5927 + 0.4999X1 + 4.7860X2 + 0.1273X3 − 0.0016X12 − 1.1620X22 − 0.0004X32

(4)

Regression coefficients for equations (3) and (4) were 0.85 and 0.86 respectively, indicating appropriate approximations to experimental data with 95% of confidence given the p-values of 0.0002 in both cases. Both equations were used to generate the response surfaces showed in Fig. 2a and b. Such Figures show that harsher hydrolysis conditions are required to optimize sugar recovery with H2SO4 as compared to HCl; 157 °C, 144 min, and 2.1% vs. 130 °C, 133 min and 1.9%. At these hydrolysis conditions, the predicted optimal sugar concentrations for HCl and H2SO4 were 21.06 and 11.82 g/L, while experimental values were 19.53 ± 1.02 and 12.77 ± 0.98 g/L, respectively. These values are equivalent to 0.39 and 0.26 g sugar/g bagasse, representing sugar recovery yields of 52 and 34% by applying the following expression: yield (%) = sugars (g) * 0.9/holocellulose content (g) * 100 (Van Dyk and Pletschke, 2012). Sugar recovery yield of 52% obtained by using HCl as acid catalyst is slightly higher than the previously reported value by our work group, 48% (Arreola-Vargas et al., 2015). However, the sugar recovery yield by using H2SO4 (34%) resulted much lower than the value reported by Saucedo-Luna et al. (2010). Nevertheless, in such study a sequential batch hydrolysis was carried out, obtaining a sugar recovery yield of 35% for the first hydrolysis, which agrees with the value obtained during our study. Significant variables (ρ < 0.05) by using both acid catalysts were the same and are presented in Tables S5 and S6. Neither the acid concentration nor the interactions influence the sugar recovery. Table 3 shows the chemical characteristics of both hydrolysates obtained at optimal conditions. Overall, higher concentrations of total sugars and monosaccharides were found in the HCl hydrolysate as compared to the H2SO4 one. The observed differences between total sugars and the sum of monosaccharides is very likely due to presence of oligosaccharides, while the predominance of xylose in both hydrolysates is due to the main lignocellulosic fraction that is hydrolyzed by the acid pretreatment, i.e. hemicellulose (Kumar et al., 2009). On the other hand, although higher concentrations of acetic acid, furfural and HMF were generated during the H2SO4 pretreatment, the high difference on sugar recovery between both acid catalysts may not only be

2.5. Analytical methods Hydrolysates were characterized in terms of: i) total sugars by the phenol-sulfuric acid method (DuBois et al., 1956); ii) monosaccharaides and potential inhibitory compounds such as 5-hydroxymethylfurfural (HMF) and furfural by HPLC (Arreola-Vargas et al., 2015); iii) chemical oxygen demand (COD) by using a HACH digester DRB200 and spectrophotometer DR2800; iv) volatile fatty acids (VFAs) such as formate, acetate, propionate, butyrate, isobutyrate, valerate, and isovalerate by HPLC (Arreola-Vargas et al., 2016). Additionally, the digestate from the anaerobic experiments were analysed in terms of total sugars, VFAs and COD as previously mentioned. Gas production is reported at normal conditions (0 °C and 1 atm).

158

Journal of Environmental Management 224 (2018) 156–163

L. Breton-Deval et al.

Fig. 1. AnSBR layout. 1.- feeding pump, 2.sampling port, 3.-effluent pipe, 4.- settler, 5. water bath with thermostat, 6.-water recirculation to reactor jacket, 7.- NaOH feeding pump, 8.- pH/ORP sensor, 9.-pH/ ORP transmitter, 10.- medium recirculation pump, 11.- pressure sensor, 12.- temperature sensor, 13.- biogas collector bag, 14.methane sensor, 15.- gas counter, 16.- automation system with cRIO9004 and data acquisition interface.

Table 2 Responses obtained during the CCD for sugar recovery and methane optimization. Run

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 a b

Coded symbol and level

Responses

X1

X2

X3

Total Sugars (g/L)a

Total Sugars (g/L)b

Methane (L)

0 0 0 −1 1 1 −1 0 0 1 1.68 −1 1 −1.68 0 −1 0 0 0

0 0 1.68 −1 −1 1 1 0 0 1 0 −1 −1 0 0 1 −1.68 0 0

−1.68 0 0 −1 1 −1 −1 0 1.68 1 0 1 −1 0 0 1 0 0 0

2.77 19.26 12.71 2.35 11.64 5.45 3.72 22.35 14.49 9.38 12.68 4.20 11.46 11.70 19.73 8.07 13.66 20.21 20.98

5.24 10.73 8.25 5.82 10.71 9.94 3.82 10.60 7.16 8.96 11.43 5.93 5.67 4.32 11.76 7.61 6.55 10.34 11.14

0.214 0.427 0.210 0.352 0.104 0.115 0.107 0.542 0.083 0.020 0.202 0.115 0.362 0.386 0.459 0.160 0.173 0.489 0.504

pretreatment with HCl. pretreatment with H2SO4.

explained as the consequence of sugar dehydration to furans and further byproducts (Dussán et al., 2014). In this context, Shi et al. (2012) reported that even though the reaction mechanism of lignocellulosic hydrolysis proceeds mainly by protonation, the anions of the acids (conjugate bases) may play an important role on the further degradation of monosaccharides and probably on the hydrolysis of the lignocellulosic material. Furthermore, other studies suggest that the hydrolysis efficiency not only depends on the acid catalyst but also on the nature of the raw materials, i.e. on the lignocellulosic complexity of each material (Herrera et al., 2004; Meinita et al., 2015, 2012). For instance, Herrera et al. (2004) demonstrated that HCl is more efficient for the hydrolysis of sorghum straw, while H2SO4 is more efficient for sugarcane bagasse. Therefore, in the case of the A. tequilana bagasse, it can be concluded that HCl is more efficient for sugar recovery than H2SO4.

Fig. 2. Response surfaces obtained from quadratic equations (3)–(5) for figures a, b and c, respectively. Where a and b are the response surfaces for sugar recovery optimization using HCl or H2SO4 as acid catalysts, and c is the response surface for methane optimization by using HCl as acid catalyst. Optimal conditions are displayed on each single optimization.

3.2. Methane production from HCl and H2SO4 hydrolysates The hydrolysates obtained at optimal conditions were evaluated in anaerobic batch assays in order to compare their potential as substrates for methane production. The kinetic profile was evaluated during 5 159

Journal of Environmental Management 224 (2018) 156–163

L. Breton-Deval et al.

Table 3 Chemical composition of the acid hydrolysates obtained at optimal conditions. Compounds (g/L)

Sugar recovery optimization

COD Total sugars Acetic Furfural HMF Glucose Galactose Mannose Xylose Arabinose

Methane optimization

HCl

H2SO4

HCl

20.55 ± 1.44 19.53 ± 1.02 1.48 ± 0.14 0.50 ± 0.05 ND 3.841 ± 0.32 1.13 ± 0.15 0.71 ± 0.22 3.11 ± 0.07 0.62 ± 0.18

14.570 ± 1.37 12.77 ± 0.98 1.75 ± 0.14 0.55 ± 0.08 0.78 ± 0.17 0.80 ± 0.24 1.23 ± 0.11 ND 2.27 ± 0.09 ND

20.74 ± 1.71 19.58 ± 1.12 1.13 ± 0.26 ND ND 3.10 ± 0.28 1.56 ± 0.19 0.58 ± 0.09 2.89 ± 0.10 0.46 ± 0.11

ND=Not detected. Table 4 Comparison of the anaerobic digestion experiments by using both types of acid hydrolysates. Type of hydrolysate

Cumulative production (NmL CH4)

Production rate (NmL CH4*h)

Yield (NL CH4/gCOD)

Yield (NL CH4/g-VS bagasse)

HCl H2SO4

531 ± 14 130 ± 18

4.02 ± 0.12 0.98 ± 0.15

0.17 0.04

0.08 0.014

days until methane production was exhausted. Table 4 shows that the HCl hydrolysate outperformed the H2SO4 one by obtaining a four-fold increase on methane yield and production. The low performance of the H2SO4 hydrolysate may have several explanations, such as the low content of sugar (the assays were performed at a normalized COD of 8 g/L), the presence of potential inhibitory compounds that could affect some microorganisms (Jönsson et al., 2013), or the presumable growth of sulphate reducing bacteria due to the presence of sulphate ions (Nozoe, 1997). Regarding the first hypothesis, the HCl hydrolysate contained 53% more sugars than the H2SO4 one, which does not explain the four-fold increase on methane production. Concerning the second hypothesis, according to Monlau et al. (2014), the concentrations of potential inhibitory compounds presented in Table 3 are not high enough to cause inhibition issues on the anaerobic digestion assays. Finally, as regards to the presence and growth of sulphate reducing bacteria, samples of biomass were taken at the beginning and end of the experiment that was fed with H2SO4 hydrolysate. DNA was extracted from these samples and sequencing was carried out by using next generation sequencing as previously described (Arreola-Vargas et al., 2017). Fig. 3 shows that abundance of sulphate reducing bacteria increased from 15% in the inoculum to 19% at the end of the experiment, confirming the presence and growth of these type of microorganisms. Among sulphate reducing bacteria, Syntrophobacter genus that are a group of facultative syntrophic bacteria were the most abundant at the end of the experiment. These microorganisms have demonstrated their capability to use propionate as electron donor and sulphate as electron acceptor (McInerney et al., 2008; Plugge et al., 2011; Worm et al., 2010). Furthermore, Fig. 3 also shows that Methanosaeta that has been previously reported growing in syntrophy with Syntrophobacter (Arreola-Vargas et al., 2017), decreased its relative abundance from 28% to 23% during the anaerobic batch assay. This suggests that Syntrophobacter may have grown toward sulphate reduction and propionate oxidation, while not sufficient acetate was produced to generate methane via Methanosaeta.

Fig. 3. Relative abundance obtained for a) Archaea and b) Bacteria domains during the anaerobic digestion assay fed with H2SO4 hydrolysate. I: Inoculum; E: End of the batch assay when methane production was exhausted. SRB: sulphate reducing bacteria.

3.3. Methane optimization by using HCl as acid catalyst during A. tequilana bagasse hydrolysis Even though sugar recovery was optimized in section 3.1, the main objective of the present work was methane production. Therefore, a second CCD was carried out to optimize the hydrolysis conditions of A. tequilana bagasse by using HCl as acid catalyst but considering methane production as the response variable. For this optimization, the same hydrolysis conditions proposed in Table 1 for sugar optimization were evaluated, while the anaerobic digestion assays were carried out at normalized conditions (8 g COD/L). To the best of our knowledge, this is the first work that compares the effect of sugar and methane optimization, since most of the reported studies have been devoted to sugar production (Arreola-Vargas et al., 2015; Kumar et al., 2009; SaucedoLuna et al., 2010). Table 2 shows the experimental responses from the different experiments of the CCD. It is worth mentioning that methane responses exhibited lack of fit to the first order model (p < 0.05), allowing to complete the CCD with axial points and fit the data to a second order model (Table S7). The optimal conditions were obtained by using the simplest second order model shown in Equation (5).

Y = −1.7289 + 0.0282X1 + 0.2935X2 + 0.0056X3 − 0.0001X12 − 0,1094X22 − 0.00003X32 + 0.0009X2 X3

(5)

The regression coefficient for equation (5) was 0.9, indicating an appropriate approximation to experimental data with 95% of confidence given the p-value of 0.0001. Significant factors (ρ < 0.05) in the ANOVA test are presented in Table S8, while the characteristics of the hydrolysate obtained at optimal conditions is presented in Table 3, in both cases the results are similar to those obtained for sugar optimization. 160

Journal of Environmental Management 224 (2018) 156–163

L. Breton-Deval et al.

The response surface obtained from equation (5) shows that softer hydrolysis conditions are required to optimize methane production as compared to sugar recovery; 1.8% HCl, 119 °C and 103min vs. 1.9% HCl, 130 °C and 133min (Fig. 2), suggesting that other components of the hydrolysates besides sugars may influence methane production (e.g. extractives, potential microbial inhibitors, etc.). At optimal conditions, an experimental methane production of 552 ± 18 NmL was obtained, which is equivalent to a yield of 0.19 NL-CH4/g-COD (0.09 NL-CH4/gVS bagasse). Such methane yield is slightly higher than the yield obtained at optimal conditions for sugar recovery (0.17 NL-CH4/g-COD) even when softer hydrolysis conditions were used during methane optimization. In both cases the methane yield is higher than the value of 0.16 NL-CH4/g-COD previously obtained under no optimal conditions (Arreola-Vargas et al., 2016) but lower than the theoretical one of 0.35 NL CH4/g COD obtained with model substrates such as glucose. The latter is very likely due to the lack of biomass adaptation to the hydrolysate components, mainly heterogeneous mixture of sugars and presence of oligosaccharides and acetic acid. Therefore, the use of reactor configurations that promote biomass adaptation is encouraged. 3.4. AnSBR performance and effect of the cycle time

Fig. 5. Methane production profiles during a single cycle of operation at 5 (a) and 3 (b) days of cycle time in the AnSBR.

One of the advantages of the cyclic operation of the AnSBR is the adaptation of the biomass to complex substrates and toxic compounds (Mace and Mata-Alvarez, 2002). According to a previous study, methane yields as high as 0.26 NL CH4/g COD have been reached in this reactor configuration from hydrolysates of A. tequilana bagasse (Arreola-Vargas et al., 2015), however the stability of the reactor has not been evaluated so far. Therefore, in the present study, HCl hydrolysates were produced at the conditions found in section 3.2.1 (1.8% HCl, 119 °C and 103min) and used as substrate in an AnSBR to evaluate its stability at three different cycle times (5, 3 and 2 days). The AnSBR was operated during 110 days. Fig. 4 shows that the first 60 days, corresponding to 12 cycles of 5 days, presented a stable performance, obtaining COD removals up to 90% and average methane yield and production of 0.28 NL/g-COD (0.13 NL-CH4/g-VS bagasse) and 0.69 NL/d, respectively.

In order to evaluate the stability of the AnSBR at shorter cycle times, the methane production profile was followed online during a single cycle. Fig. 5a shows that the methanogenic activity was predominant during the first three days of the cycle. During this period, more than 80% of the methane was produced. Interestingly, the methanogenic activity was recovered during the last day of the cycle, which can be explained as a result of either, endogenous metabolism because of biomass starvation or degradation of recalcitrant compounds of the hydrolysate such as furans or phenolics (Monlau et al., 2014). Because of the results observed in Fig. 5 and to the fact that shorter cycle times should theoretically increase the methane productivity (Li et al., 2016; Sivagurunathan et al., 2016), the stability and performance of the AnSBR was evaluated at a cycle time of 3 days. According to Fig. 4, at this cycle time the reactor was stable with an average methane yield of 0.28 NL/g-COD, equal to that obtained at 5 days of cycle time. However, the methane production was increased by 50%, obtaining a value of 1.04 NL/d and confirming that productivity can be improved by shortening the cycle time without affecting the stability of the reactor. It is worth to mention that electricity issues caused the lowest values of COD removal, methane yield and methane production observed at day 78. However, the reactor performance was completely recovered in the following cycle, demonstrating the excellent robustness property of this reactor configuration. It is also remarkable that even though the obtained yield is slightly lower than the theoretical one (0.35 NL CH4/g-COD), it is higher than other reports using semi-continuous or continuous systems and fed with hydrolysates from A. tequilana bagasse or wheat straw (Arreola-Vargas et al., 2015; Kaparaju et al., 2009). According to Table 5, only Gomez-Tovar et al. (2012) and Corona and Razo-Flores (2018) have achieved similar values to the theoretical one (0.35 NL CH4/g-COD) but using continuous systems and enzymatic hydrolysates that would negatively impact the cost of a full scale process. Finally, from day 90–110 the reactor was operated at a cycle time of 2 days aiming the further improvement of the productivity. This cycle time was selected because the methane production profile during the last cycle of 3 days demonstrated that the maximum production was obtained at two days of operation (Fig. 5b). Nonetheless, when the cycle time was shortened, the COD removal decreased below 50% and both methane production and yield were considerably affected (Fig. 4). According to Fig. 6, at this cycle time, occurrence of lactate production was detected, indicating the growth of lactic acid bacteria. These bacteria can compete with hydrogenogenic microorganisms for the

Fig. 4. AnSBR performance in terms of COD removal and methane production rate and yield at different cycle times. 161

Journal of Environmental Management 224 (2018) 156–163

L. Breton-Deval et al.

Table 5 Reported methane yields in one and two stages from different lignocellulosic hydrolysates. Reference

Type of hydrolysate

Reactor configuration

System in one or two stages

Yield (NL CH4/g COD)

Arreola-Vargas et al., 2015 Arreola-Vargas et al., 2016 Gomez-Tovar et al., 2012 Kaparaju et al., 2009 Corona and Razo-Flores, 2018 This study

Acid hydrolysate from Agave tequilana bagasse Enzymatic hydrolysate from Agave tequilana bagasse Enzymatic hydrolysate from oat straw Hydrothermal hydrolysate from wheat straw Enzymatic hydrolysate from Agave tequilana bagasse Acid hydrolysate from Agave tequilana bagasse

ASBR Batch UASB UASB CSTR ASBR

one two one one two one

0.26 0.24 0.36 0.27 0.32 0.28

sulphate reducing bacteria with methanogens. This work also found that softer hydrolysis conditions are required to optimize methane production as compared to sugar recovery, revealing that not necessarily the best conditions for sugar recovery are the best conditions for biofuel production. Finally, results on the AnSBR demonstrated the suitability of this reactor configuration for methane production from this type of substrate. The AnSBR showed stable methane productions and yields close to the theoretical one during more than 90 days of operation at cycle times of 5 and 3 days. Nonetheless, and even though methane productivity may be theoretically increased at shorter cycle times, this is not advised due to microbial imbalance and reactor failure. Acknowledgements This research was supported by Fondo Sectorial CONACyT-SENER, CEMIE-Bio project No. 247006. Luz Breton-Deval thanks SENERCONACyT for postdoctoral fellowship. Appendix A. Supplementary data Supplementary data related to this article can be found at https:// doi.org/10.1016/j.jenvman.2018.07.053. References Arreola-Vargas, J., Flores-Larios, A., González-Álvarez, V., Corona-González, R.I., Méndez-Acosta, H.O., 2016. Single and two-stage anaerobic digestion for hydrogen and methane production from acid and enzymatic hydrolysates of Agave tequilana bagasse. Int. J. Hydrogen Energy 41, 897–904. https://doi.org/10.1016/j.ijhydene. 2015.11.016. Arreola-Vargas, J., Ojeda-Castillo, V., Snell-Castro, R., Corona-González, R.I., AlatristeMondragón, F., Méndez-Acosta, H.O., 2015. Methane production from acid hydrolysates of Agave tequilana bagasse: evaluation of hydrolysis conditions and methane yield. Bioresour. Technol. 181, 191–199. https://doi.org/10.1016/j.biortech.2015. 01.036. Arreola-Vargas, J., Snell-Castro, R., Rojo-Liera, N.M., González-Álvarez, V., MéndezAcosta, H.O., 2017. Effect of the organic loading rate on the performance and microbial populations during the anaerobic treatment of tequila vinasses in a pilot-scale packed bed reactor. J. Chem. Technol. Biotechnol. https://doi.org/10.1002/jctb. 5413. Corona, V.M., Razo-Flores, E., 2018. Continuous hydrogen and methane production from Agave tequilana bagasse hydrolysate by sequential process to maximize energy recovery efficiency. Bioresour. Technol. 249, 334–341. https://doi.org/10.1016/j. biortech.2017.10.032. Crespo, M.R., González, D.R., Rodríguez, R., Rendón, L.A., del Real, J.I., Torres, J.P., 2013. Evaluación de la composta de bagazo de agave como componente de sustratos para producir plántulas de agave azul tequilero. Rev. Mex. Ciencias Agrícolas 4, 1161–1173. CRT, 2016. Consejo Regulador del tequila-Statistical information. Accesed on November 2017. https://www.crt.org.mx/EstadisticasCRTweb/. DuBois, M., Gilles, K.A., Hamilton, J.K., Rebers, P.A., Smith, F., 1956. Colorimetric method for determination of sugars and related substances. Anal. Chem. 28, 350–356. https://doi.org/10.1021/ac60111a017. Dussán, K.J., Silva, D.D.V., Moraes, E.J.C., Arruda, P.V., Felipe, M.G.A., 2014. Dilute-acid hydrolysis of cellulose to glucose from sugarcane bagasse. Chem. Eng. Trans. 38, 433–438. https://doi.org/10.3303/CET1438073. Gomez-Tovar, F., Celis, L.B., Razo-Flores, E., Alatriste-Mondragón, F., 2012. Chemical and enzymatic sequential pretreatment of oat straw for methane production. Bioresour. Technol. 116, 372–378. https://doi.org/10.1016/j.biortech.2012.03.109. Herrera, A., Téllez-Luis, S.J., González-Cabriales, J.J., Ramírez, J.A., Vázquez, M., 2004. Effect of the hydrochloric acid concentration on the hydrolysis of sorghum straw at atmospheric pressure. J. Food Eng. 63, 103–109. https://doi.org/10.1016/S02608774(03)00288-7.

Fig. 6. Metabolic byproducts (a) and COD distribution (b) obtained at the end of the different cycle times during the AnSBR operation.

substrate, affecting the final fate of the electron equivalents that originally should be disposed on methane molecules (Park et al., 2016). Furthermore, lactic acid bacteria can also produce bacteriocins that have a detrimental effect over other microorganisms (Noike et al., 2002). The cycle time effect is clearly evidenced when a COD balance is applied to the last cycles of operation at each stage (Fig. 6b). During the first two cycle times, around 80% of the original COD was converted to methane, while this percentage decreased to less than 40% for the last cycle time of two days. As previously mentioned, it is very likely that constant feeding of substrate at this short cycle time promoted the growth of nocive organisms for the process and cause microbial imbalance and reactor failure. 4. Conclusions Comparison of HCl and H2SO4 as acid catalysts during the hydrolysis of A. tequilana bagasse demonstrated that HCl is more effective to release fermentable sugars at optimal conditions. Furthermore, H2SO4 hydrolysates induced lower methane yields during the anaerobic digestion assays due to its lower content of sugars and the competence of 162

Journal of Environmental Management 224 (2018) 156–163

L. Breton-Deval et al.

Montgomery, D.C., 2013. Design and Analysis of Experiments. John Wiley & Sons, Inc. Noike, T., Takabatake, H., Mizuno, O., Ohba, M., 2002. Inhibition of hydrogen fermentation of organic wastes by lactic acid bacteria. Int. J. Hydrogen Energy 27, 1367–1371. https://doi.org/10.1016/S0360-3199(02)00120-9. Nozoe, T., 1997. Effects of methanogenesis and sulfate-reduction on acetogenetic oxidation of propionate and further decomposition of acetate in paddy soil. Soil Sci. Plant Nutr. 43, 1–10. https://doi.org/10.1080/00380768.1997.10414709. Park, J.-H., Lee, S.-H., Ju, H.-J., Kim, S.-H., Yoon, J.-J., Park, H.-D., 2016. Failure of biohydrogen production by low levels of substrate and lactic acid accumulation. Renew. Energy 86, 889–894. https://doi.org/10.1016/j.renene.2015.09.016. Plugge, C.M., Zhang, W., Scholten, J.C.M., Stams, A.J.M., 2011. Metabolic flexibility of sulfate-reducing bacteria. Front. Microbiol. 2, 81. https://doi.org/10.3389/fmicb. 2011.00081. Rodrigues, R. de C.L.B., Rocha, G.J.M., Rodrigues, D., Filho, H.J.I., Felipe, M., das, G.A., Pessoa, A., 2010. Scale-up of diluted sulfuric acid hydrolysis for producing sugarcane bagasse hemicellulosic hydrolysate (SBHH). Bioresour. Technol. 101, 1247–1253. https://doi.org/10.1016/j.biortech.2009.09.034. Saucedo-Luna, J., Castro Montoya, A., Rico, J.L., Campos-Garcìa, J., 2010. Optimization of acid hydrolysis of bagasse from agave tequilana weber. Rev. Mex. Ing. Quim. 9, 91–97. Shi, Y., Yokoyama, T., Akiyama, T., Yashiro, M., Matsumoto, Y., 2012. Degradation kinetics of monosaccharides in hydrochloride and sulfuric acid. BioRessources 7, 4085–4097. Sivagurunathan, P., Anburajan, P., Kumar, G., Kim, S.-H., 2016. Effect of hydraulic retention time (HRT) on biohydrogen production from galactose in an up-flow anaerobic sludge blanket reactor. Int. J. Hydrogen Energy 41, 21670–21677. https://doi. org/10.1016/j.ijhydene.2016.06.047. Van Dyk, J.S., Pletschke, B.I., 2012. A review of lignocellulose bioconversion using enzymatic hydrolysis and synergistic cooperation between enzymes-Factors affecting enzymes, conversion and synergy. Biotechnol. Adv. https://doi.org/10.1016/j. biotechadv.2012.03.002. Worm, P., Müller, N., Plugge, C.M., Stams, A.J., Schink, B., 2010. Syntrophy in methanogenic degradation. In: (Endo) Symbiotic Methanogenic Archaea. Springer Berlin Heidelberg, Münster, Germany, pp. 143–173.

Hutomo, G.S., Rahim, A., Kadir, S., 2015. The effect of sulfuric and hydrochloric acid on cellulose degradation from pod husk cacao. Int. J. Curr. Microbiol. Appl. Sci. 4, 89–95. Jönsson, L.J., Alriksson, B., Nilvebrant, N.-O., 2013. Bioconversion of lignocellulose: inhibitors and detoxification. Biotechnol. Biofuels 6, 16. https://doi.org/10.1186/ 1754-6834-6-16. Kaparaju, P., Serrano, M., Angelidaki, I., 2009. Effect of reactor configuration on biogas production from wheat straw hydrolysate. Bioresour. Technol. 100, 6317–6323. https://doi.org/10.1016/j.biortech.2009.06.101. Kumar, P., Barrett, D.M., Delwiche, M.J., Stroeve, P., 2009. Methods for pretreatment of lignocellulosic biomass for efficient hydrolysis and biofuel production. Ind. Eng. Chem. Res. 48, 3713–3729. https://doi.org/10.1021/ie801542g. Li, Y., Zhang, Y., Liu, Y., Zhao, Z., Zhao, Z., Liu, S., Zhao, H., Quan, X., 2016. Enhancement of anaerobic methanogenesis at a short hydraulic retention time via bioelectrochemical enrichment of hydrogenotrophic methanogens. Bioresour. Technol. 218, 505–511. https://doi.org/10.1016/j.biortech.2016.06.112. Mace, S., Mata-Alvarez, J., 2002. Utilization of SBR technology for wastewater Treatment: an overview. Ind. Eng. Chem. Res. 41, 5539–5553. https://doi.org/10. 1021/ie0201821. McInerney, M.J., Struchtemeyer, C.G., Sieber, J., Mouttaki, H., Stams, A.J.M., Schink, B., Rohlin, L., Gunsalus, R.P., 2008. Physiology, ecology, phylogeny, and genomics of microorganisms capable of syntrophic metabolism. Ann. N. Y. Acad. Sci. 1125, 58–72. https://doi.org/10.1196/annals.1419.005. Meinita, M.D.N., Hong, Y.K., Jeong, G.T., 2012. Comparison of sulfuric and hydrochloric acids as catalysts in hydrolysis of Kappaphycus alvarezii (cottonii). Bioproc. Biosyst. Eng. 35, 123–128. https://doi.org/10.1007/s00449-011-0609-9. Meinita, M.D.N., Marhaeni, B., Winanto, T., Setyaningsih, D., Hong, Y.K., 2015. Catalytic efficiency of sulfuric and hydrochloric acids for the hydrolysis of Gelidium latifolium (Gelidiales, Rhodophyta) in bioethanol production. J. Ind. Eng. Chem. 27, 108–114. https://doi.org/10.1016/j.jiec.2014.12.024. Monlau, F., Sambusiti, C., Barakat, A., Quéméneur, M., Trably, E., Steyer, J.P., Carrère, H., 2014. Do furanic and phenolic compounds of lignocellulosic and algae biomass hydrolyzate inhibit anaerobic mixed cultures? A comprehensive review. Biotechnol. Adv. https://doi.org/10.1016/j.biotechadv.2014.04.007.

163