Substrates and inoculum

The study was carried out from January to May 2019 in the Soils Laboratory of CIIDIR Oaxaca, in Santa Cruz Xoxocotlan, Oaxaca, Mexico (17º 01 ‘30.3’ ‘N and 96º 43’ 12.5 ‘’ W, at 1530 meters above sea level). The BM was donated by the municipal slaughterhouse of Santa Cruz Xoxocotlan, Oaxaca. The vinasses were obtained at the end of the distillation process at the “Don Agave” mezcal factory, located in the municipality of Tlacolula de Matamoros, Oaxaca. Both substrates were stored at 4 oC until their use. The inoculum was prepared from the effluent of an active biodigester from a pig farm in Santa Ana Zegache, Oaxaca. A mixture of the effluent, water, vinasses, and BM in 4:2:1:1 volume/volume ratio (v/v), respectively, was prepared for its adaptation (^{Angelidaki et al., 2009}) and incubated under mesophilic conditions (36 ± 1 oC) in a water bath, in two airtight plastic containers of 10.1 L each, with 9 L of liquid volume per container, for 25 days until gas production stopped. The pH, total solids (TS), and VS of the substrates and inoculum were determined in triplicate according to APHA (^{APHA, 2012}). The pH was determined with a Hanna brand digital potentiometer (Hanna Instruments^®), TS and VS were determined using an oven at 105 and 550 oC until constant weight with a digital analytical balance (model MSL, Brand Metter Toledo^®), respectively.

Experimental design and data collection

A completely randomized design with six repetitions per treatment was used. Five vinasses-BM (V-BM) mixtures at different percentage (v/v) proportions (25-75, 50-50, 75-25, 100-0, and 0-100) were used as treatments, with no pH adjustment nor TS balance to reproduce as normal conditions as possible. The experimental unit consisted of a 496 mL glass bottle with a metal lid and hermetic seal, with 300 mL of active volume, consisting of 75 mL of V-BM mixture, according to the proportions described above, 75 mL of water to avoid overload and keeping TS content below 10 % which is optimal for AD performance (^{Lorenzo -Acosta and Obaya-Abreu, 2005}; ^{Angelidaki et al., 2009}), and 150 mL of previously prepared inoculum, and immersed in water in mesophillic conditions (36 ± 1 oC) for 10 days to promote methanogenic activity (^{Espinoza-Escalante et al., 2009}; ^{Abdoli et al., 2014}). The experiment was stopped when the volume of biogas produced per day during the last three days was <1 % of the total accumulated biogas (^{Holliger et al., 2016}).

The biogas produced per day and accumulated during the evaluation period, normalized (NmL) to standard temperature and pressure (273.15 °K, 101.33 kPa), was obtained using the ideal gas law equation (^{Holliger et al., 2016}). Biogas methane content was determined by the volumetric displacement method of 5 % NaOH solution (^{Víquez, 2017}). The methane yield was obtained with equation 1:

where ϒ _CH4 is the methane yield (NmL CH₄ g^-1 VS), A_V is the accumulated volume of biogas during the period of the experiment (NmL CH₄), and VS_R is the mass of volatile solids in grams added to the reactor. The daily methane yield was calculated dividing the accumulated yield by 10, which corresponds to the number of days of the experiment.

Statistical analysis

An analysis of variance was used to determine significant differences between treatments in the accumulated volume of methane, a means separation Tukey test was used to determine the best V-BM proportion according to data. To identify the critical stages of AD, such as the start of methane production and the microbial adaptation phase (^{Zwietering et al., 1990}) a nonlinear regression (p<0.05) was performed with the collected data, using the SPSS v. 22 software (Statistical Package for the Social Sciences) to adjust the methane yield to the Gompertz bacterial growth model (Equation 2).

In this model, methane production is proportional to microbial activity. P is the prediction of methane production in a certain period (t), A is the accumulated volume of methane in a given time (NmL CH₄ g^-1 VS), μm is the maximum rate of methane yield per day (NmL CH₄ g^-1 VS d^-1) defined by the tangent of the slope in the exponential growth phase of bacteria, and λ is the adaptation period (d) defined by the intercept of the line of the slope with the x-axis (^{Zwietering et al., 1990}).

Characterization of the inoculum, substrates, and mixtures

Fresh vinasses presented acidic pH compared to bovine manure. The pH of the inoculum adapted to the vinasses was slightly alkaline, higher than vinasses. The highest vinasses content presented a more acidic pH than those with the highest manure content. The pH in 100-0 treatment was higher than the original substrates because of the 75 mL of water added to avoid overload of the system. The content of total solids and volatile solids of vinasses was lower than that of the inoculum, and bovine manure. The vinasses-bovine manure mixtures showed different total solids content. The 100-0 mix had the lowest total solids and volatile solids content, while the 0-100 mix had the highest content (Table 1).

The acidity of vinasses is determined by the content of organic acids, such as acetate, which is synthesized during the fermentation of the agave sugars to obtain alcohol; therefore, as the vinasse in the mixture increases, the pH becomes acidic. On the other hand, bovine manure addition to the mixture increased pH to a methanogenic level. The vinasses pH was comparable to that reported by other researchers for mezcal vinasses with values between 3.5 and 4.77 (^{Gómez-Guerrero et al., 2019}; ^{López-Velarde et al., 2019}). The initial pH of the mixtures studied was at optimal levels for anaerobic digestion (6.8 and 7.4), except for the 100-0 mixture with 5.49, which is closest to the optimal pH level of acidogenic bacteria, between 5.5 and 6.5 (^{Khanal, 2009}). When inhibitions to vinasses anaerobic digestion occur with pH values lower than 5.0 (^{Gómez-Guerrero et al., 2019}), pH adjustments are needed to carry out the anaerobic digestion (^{Espinoza-Escalante et al., 2009}).

The total solids content of vinasses studied was similar to those obtained from industrial-type mezcal factories in Oaxaca (43.4 g L^-1) (^{Villalobos et al., 2009}) and San Luis Potosí (49.17 g L^-1) (^{López-Velarde et al., 2019}), but lower than artisan-type factories (91- 94.7 g L^-1) (^{Gómez-Guerrero et al., 2019}). In contrast, the volatile solids content of vinasses was lower than those from artisan-type mezcal factories (72 g L^-1) (^{Gómez- Guerrero et al., 2019}) and Comiteco liquor vinasses (62.90 g L^-1) obtained from agave juices (^{Cruz-Salomón et al., 2017}). Regarding total solids content, it was less than 10 % in all vinasses-bovine manure mixtures, values below which the anaerobic digestion “usually operates in better conditions” (^{Lorenzo-Acosta and Obaya-Abreu, 2005}). It was found that food waste anaerobic digestion performance was improved by increasing total solids content from 5 to 20 % (^{Yi et al., 2014}). In these studied mixtures, the volatile solids content was recorded in the 20 to 60 g L^-1 range, recommended for anaerobic digestion tests (^{Holliger et al., 2016}), except for treatment 0-100, which had lesser values.

pH changes in mixtures

In the V-BM mixtures the pH increased as a function of the bovine manure content, decreasing significantly (p < 0.0001) in every mixture at the end of the experiment (5.64 ± 0.87). The lowest pH value was obtained in the mixtures with the highest vinasses content (Figure 1).

Figure 1 pH changes of the vinasses-bovine manure (V-BM) mixtures at the beginning and at the end of the experiment. 

Figura 1 Cambios de pH de las mezclas de vinazas-estiércol bovino (V-EB) al inicio y al final del experimento. 

Archaea bacteria synthesize volatile fatty acids during the acidogenesis stage for methane synthesis, thus the pH values of the mixtures during anaerobic digestion decreases (^{Khanal, 2009}; ^{Pibul and Towprayoon, 2015}). The pH variation at the end of the anaerobic digestion process is expected when no pH conditioner is added (^{Gómez-Guerrero et al., 2019}), as in the present study. The highest and most significant pH variation in the mixtures with the highest bovine manure content could be due to a higher volatile fatty acids production, associated with higher total solids content used for its synthesis (^{Parra-Orobio et al., 2015}).

Methane production

The accumulated volume of methane was significantly different (p<0.05) between vinasses-bovine manure mixtu-res. The highest volume was obtained with 0-100 and 75-25 mixtures, with no significant difference between them, and the lowest volume with 25-75 and 50-50 mixtures. The 100-0 mix produced an intermediate methane volume (Figure 2).

Figure 2 Methane accumulated volume from the vinasses-bovine manure (V-BM) mixtures. Different letters indicate significant differences between treatments (p<0.05). 

Figura 2 Volumen acumulado de metano de las mezclas de vinazas-estiér-col bovino (V-EB). Letras diferentes indican diferencias significativas entre tratamientos (p<0.05). 

Substrates such as vinasses, with dissolved organic matter and highly degradable sugars, allow higher methane yields (^{Yi et al., 2014}) due to fructans present in the agave “piñas” (^{Vera-Guzmán et al., 2009}; ^{Chávez-Parga et al., 2016}). During their cooking in the mezcal process, fructans are hydrolyzed into fermentable sugars, mainly fructose, glucose, xylose, and maltose, at different concentrations according to the degree of plant ripeness. These sugars are used in microbial fermentation for alcohol synthesis (^{Michel-Cuello et al., 2008}). However, a portion of the sugars cannot be fermented and go to vinasses; for example, 77.1 mg g^-1 of total sugars were found in the wort after fermentation, i.e., 14 % of the initial total present in the juice of the cooked agave “piñas” (^{Vera-Guzmán et al., 2009}), and 51 mg g^-1 of total sugars in vinasses after distillation (^{López-Velarde et al., 2019}). These sugars present in vinasses, along with other suspended organic solids (^{Espinoza-Escalante et al., 2009}), are used by bacteria during anaerobic digestion to obtain energy and forward substrates for methanogenesis. In the case of the 75-25 mix, the bacterial community present in bovine manure uses these sugars and the optimized pH conditions to produce methane. The 75-25 and 0-100 mixtures had the highest cumulative methane production with 286.40 and 286.23 NmL CH₄, respectively. The 100-0 mix produced 225.98 NmL CH₄, ranking third; the 75-25 mix produced 26.7 % more methane than the 100-0 mix, indicating that vinasses-bovine manure co-digestion can be used for improving methane production from the mezcal vinasses.

Methane yield kinetics

The bacterial kinetic curve coincided with the Gompertz prediction model growth. The mixtures with higher vinasses content obtained a higher accumulated methane yield, while mixtures with higher bovine manure content presented lower yields; in addition, they produced 80 % of the methane produced in the first four days, while the 100-0 mix obtained that production after six days, indicating that co -digestion with bovine manure reduces the digestion time of vinasses by two days (Figure 3).

Figure 3 Methane yield kinetics from vinasses-bovine manure (V-BM) mixtures during the experimental period. 

Figura 3 Cinética de los rendimientos de metano de las mezclas de vinazas-estiércol bovino (V-EB) durante el periodo experimental. 

The methane yield expresses the volume of methane produced per gram of volatile solids invested in the system; this explains that despite having lower accumulated production (Figure 2), in the 100-0 mix methane yield was higher (28.27 NmL CH₄ g^-1 volatile solids) compared to the 0-100 mix (14.43 NmL CH₄ g^-1 volatile solids). Regarding the performance of the yield curve, the conversion rate of complex substrates such as vinasses with acidic pH presents a more prolonged hydrolysis stage at the start of anaerobic digestion, which was evidenced in the S-shaped methane yield curve. In contrast, simpler substrates with more balanced properties, such as neutral pH in mixtures with bovine manure, have shorter incubation periods, as evidenced in the inverted L-shaped curve (^{Ware and Power, 2017}).

The cumulative methane yield estimated by the Gompertz model was adjusted to the results obtained in the experiment with R2 values close to 1 (Table 2). The bacterial kinetics for each of the vinasses-bovine manure mixtures described the expected bacterial growth behavior by the theoretical model. The 100-0 mix had the highest daily and cumulative return rate; nevertheless, it had the highest adaptation phase due to the lower pH value (Figure 2). This implies the need for a longer hydraulic retention time and larger reactors, or in the best case, pH adjustment. On the other hand, in the mixtures with bovine manure, the degradation began before its preparation, indicated by the λ negative values. This phenomenon is possibly caused by the microbiological activity present in the manure before the anaerobic digestion, which could start vinasses digestion in a shorter time, i.e., the 75-25 mix adaptation period was -1.83, which means that methane production started 4.12 days earlier than in the 100-0 mix, and had the best performance regarding the maximum rate of methane yield per day and cumulative methane yield, comparable to the 100-0 mix. This indicates that the vinasses-bovine manure co-digestion in a 75-25 proportion best favors the mezcal vinasses anaerobic digestion treatment process.

The accumulated methane yield was higher with a higher vinasses content because “vinasses are more efficient substrates for anaerobic digestion than other substrates due to the quantity of soluble sugars they contain”, contrary to what happened with the cellulose present in the bovine manure (^{López-Velarde et al., 2020}). During the adaptation period, bacteria grow and prepare to hydrolyze organic compounds; it takes longer for complex substrates under acidic pH (^{Parra-Orobio et al., 2014}), such as the vinasses alone. Nevertheless, this period can be shortened by adding bovine manure, as shown by the 75-25 mix. The co-digestion of substrates demonstrates the shortening of the hydrolysis period, i.e., when maize waste was added to cow dung digestion, the incubation period was 4 days in contrast with 11 days for cow dung digestion alone (^{Abdoli et al., 2014}). Negative λ values have been observed in anaerobic digestion bacterial growth curves of substrates with an initial bacterial load, such as bovine manure (^{Ware and Power, 2017}; ^{López-Velarde et al., 2020}). This implies that vinasses-bovine manure co- digestion favors the initial presence of a bacterial community for vinas-ses anaerobic digestion in a shorter time.