Introduction

A brief scenario about the solid waste sector and its environmental implications

The waste management sector is amongst the main anthropic sources of greenhouse gases (GHG), principally methane (CH4) and carbon dioxide (CO2), formed during the various decomposition steps of waste mass. Different alternatives aimed at reducing GHG emission have been more and more discussed in agreements between dozens of countries in recent decades. The Kyoto Protocol (1997) and the United Nations Climate Change Conferences of the Parties (COP21 and COP26 in 2015 and 2021, respectively) are examples of important global climate change agreements.

Specifically referring to food waste (FW) management, approximately one-third of all the food produced in the world (edible part) is lost during its passage through the supply chain from its agricultural production to the final consumer. On a global scale, the annual cost of food wastage is estimated as close to US$ 1 trillion, but this value does not consider the enormous social and environmental costs related to food wastage [1, 2]. Just in Latin America, the losses within the food supply chain with the storage (postharvest), processing and distribution of fruits and vegetables can reach nearly 30% [2]. Some environmental impacts associated with food waste management are their disposal in landfills (leachate generation, GHG emission, reduction in the landfill lifespan), odor generation, proliferation of pathogenic organisms, etc. It has been estimated that 3.5 GtCO2eq are annually launched into the atmosphere due to food wastage [1].

Anaerobic digestion (AD): an alternative for treating FVW

Biomass is an important alternative fuel source and has been increasingly used in energy-generating processes, thus increasing the contribution of renewable energies in the worldwide energy matrix, still strongly based on fossil fuels. In this context, anaerobic biodigestion processes become a sustainable alternative (environmentally and economically) to the disposal of food wastes in landfills. Two important subproducts are coupled to the biodigestion processes: the biogas, an important energy source, and the digestate, commonly used as soil fertilizer [3,4,5].

Due to characteristics such as the high moisture content and biodegradability of FVW, various studies have evaluated the biochemical methane potential of these residues [6,7,8,9,10], with the objective of determining the biogas yield under anaerobic conditions and the possibility of its large-scale application.

Chemical pre-treatment applied to the AD of organic solid wastes

FVW are typically lignocellulosic residues, constituted mainly of lignin, cellulose and hemicellulose. To optimize the biodigestion of organic lignocellulosic matter (thus increasing the methane yield), the technical literature has reported the most varied technologies to pretreat these residues (physical, chemical, biological or the association of two or more of these methods). Recalcitrant in nature, lignin constitutes one of the main barriers to the biodigestion of lignocellulosic materials. Thus the pre-treatment becomes an important preliminary step in biodigestion processes, as a way of removing/reducing the lignin content and reducing the crystalline structure of the cellulose, as well as increasing the surface area and porosity of the biomass for the biological activity [11, 12].

Due to their lower costs as compared to other pretreatments, elevated efficiency and degradation kinetics of the organic substrates [12,13,14,15], chemical pretreatments have been applied to a variety of substrates as follows, with a view to optimizing biogas production: fruit and vegetable wastes [16], banana stem [17], sugarcane bagasse [18], corn stover [15, 19]; rice straw [20, 21], wheat straw [22], walnut shells [23], switchgrass [24], etc. In these and other cases, sodium hydroxide (NaOH) was the most used alkali reagent and/or most effective in the pre-treatment of lignocellulosic residues. Apart from the effects related to lignin removal, increase in enzymatic hydrolysis, and reduction in cellulose crystallinity, etc., the residual alkaline reagent (post-treatment) helps buffer the system during the AD steps, avoiding the fall in pH commonly found during the initial steps of the process [12, 18, 20, 25, 26]. It is important to highlight that the reaction time (between substrate and reagent), pretreatment temperature and reagent concentration are some of the parameters that should be evaluated to optimize the biogas generation process. Table S1 (Supplementary section) shows the effect of pre-treatment with NaOH (in terms of methane generation) of some lignocellulosic residues under different process conditions.

The principal premise of this work was to identify the best alkali pre-treatment conditions for fruits and vegetables in a biodigester already operating at the central food distribution market (CEASA, in Portuguese abbreviation) in the city of Recife, in northeastern Brazil. The CEASA biodigester is fed with ‘in natura’ residues, operates at room temperature and can process up to 1 ton of food waste per day. It is important to mention that most of the studies report on the biodigestion of only one or two substrates, or even already-processed FW from restaurants, canteens or households, whose physicochemical characteristics (and hence biogas generation) do not represent the same conditions as the codigestion of raw fruits and vegetables discarded from supply centers, open markets and supermarkets. The present study dealt with the codigestion of fourteen substrates generated by CEASA in large amounts, each with distinct properties (pH, moisture, volatile solids, etc.), with the objective of verifying the interaction of the different components in the mixture during the generation of biogas. When surveying the literature, we did not find studies which had dealt with such a large amount of residues, allied to factors such as low pretreatment time, low pretreatment temperature (ambient), low NaOH concentrations, and low biodigestion time considered simultaneously (as we did in our study). Indeed, on a full-scale application, working at room temperature with low reagent concentrations could be convenient, not only from the economic and environmental points of view but also to convert the substrate into biogas with a minimal production of inhibitory side products.

Materials and methods

Substrate and inoculum

Fourteen raw fruits and vegetables, unfit for consumption, were collected from the central food distribution market (CEASA) in the city of Recife (Brazil). The residues were chosen as a function of the large volume generated in the post-harvest steps, availability for the CEASA biodigester and/or their elevated potential for biogas generation per ton (according to Da Silva and Palha [27]). The digestate from the biodigester itself was used as the inoculum, already climatized to the selected residues. After sampling, the residues were placed in plastic bags and stored in a freezer, and the digestate in a refrigerator (at ≅ 4 ºC) for preservation purposes. The digestate was collected less than 30 h before its use as the inoculum (BMP assays).

Experimental methodology

Pretreatment conditions

The FVW were previously mixed in a proportion of 0.2 g volatile solids per residue (total of 2.8 gvs added), the material comminuted and the mixture homogenized in a blender. Table 1 shows the pretreatment conditions with sodium hydroxide used for each mixture. Considering the high moisture content of the mixtures (Table 2), the sodium hydroxide (solid, 99.2% purity, brand Neon) was effectively diluted by the samples without the addition of extra water and homogenized manually. No buffering agent was added to the mixtures (even the sample of untreated waste–UTW) so as not to mask the action of the alkali reagent in terms of biogas generation.

Table 1 Chemical pretreatment conditions
Full size table
Table 2 Physicochemical characteristics of the FVW and FVW + digestate mixtures
Full size table

The mixtures 3–24, 3–48 and 6–24 were maintained at room temperature (mean of ≅ 25 ºC) throughout the pretreatment period. The pretreatment conditions were simplified with a view to scale-up of the process.

Anaerobic digestion

After the pretreatment period, digestate was added to each mixture (the three pretreated mixtures and the UTW) in a substrate/inoculum (S/I) proportion of 0.5 on a VS basis. The biodigesters used consisted of 250 mL borosilicate bottles sealed with nylon lids equipped with a manometer (to monitor the biogas generated) and valves to discharge the biogas. Approximately 70 mL of the mixture (FVW + inoculum) were added to each bottle and placed in an orbital shaker incubator (70 rpm) at 37 ºC until a minimum generation of biogas was detected (less than 3% of the total produced). Before sealing the bottles, they were flushed with gaseous nitrogen to obtain anaerobic conditions in the headspace of the digesters. Each sample of FVW + inoculum was treated in triplicate. Bottles only with inoculum were also prepared (blanks), and the biogas produced by the blanks was subtracted from the biogas produced by the mixture.

Analytical methods

The physiochemical characterization of the substrates took into account the speed and low cost of the analysis, with a view to its large-scale application or a very high frequency of the analyses. The parameters of moisture content, pH, total solids (TS) and volatile solids (VS) were analyzed under the following conditions before adding the inoculum:

  • - raw FVW (separately) and digestate;

  • - mixtures UTW, 3–24, 3–48 and 6–24. For these mixtures the composition of the lignocellulosic material (lignin, cellulose and hemicellulose) was also determined according to Embrapa [28].

In addition, the moisture, pH, TS, VS and chemical oxygen demand (COD) were also verified prior to and at the end of the biodigestion of the FVW + inoculum mixtures.

The pH was measured using a Digimed pHmeter, model DM23; moisture was determined in an oven at a constant temperature of 105 ºC (first 24 h) followed by 65 ºC for the remaining days of drying, to constant mass; VS was determined in a muffle furnace at 550 ºC for 4 h; and COD was measured using the 5220 Method [29].

Biogas monitoring

The headspace pressure of the flasks was monitored daily and converted to the volume of biogas at STP (standard temperature and pressure), according to the ideal gas law. The results were reported in terms of the volume of methane generated daily and the total accumulated volume (mLCH4.gVS−1FVW).

The frequency of determining the methane and carbon dioxide concentrations depended on the volume of biogas generated during the biodigestion process, since the analyses required a minimum amount of gas accumulated in each flask. On average two measurements were made per week (total of 9). The biogas composition was determined in a portable Dräger gas analyzer, model Xam-7000, with infrared sensors to measure CH4 and CO2 (0–100 Vol.-%). Several studies analyzed the biogas composition in the biodigestion of food wastes employing the same method (and equipment) used in this study [30,31,32,33,34,35].

Statistical analysis

A completely randomized experimental design was used, with two repetitions for the parameters of cellulose, hemicellulose and lignin, and three repetitions for t90, total volume and methane concentration generated. The differences between the means of these parameters were determined by an analysis of variance (one-way ANOVA) and by Fisher’s least significant difference (LSD). The Gaussian assumption of the residues and homogeneity of the variances were checked by the Shapiro–Wilk’s and Bartlett tests, respectively. A 95% confidence interval (values of p ≤ 0.05) are significant) was used for all analyses, and the analyses were carried out using the RStudio software.

Kinetic study of biogas production

Bench scale biodigestion studies commonly approach the system kinetic data determination. Such data are highly relevant for the determination of the reactor type suitability for the biochemical process, verification of the biodigestion system best kinetic parameters, and the full scale process physical parameters such as reactor volume, hydraulic retention time, etc. [9]. The pre-definition of kinetic parameters must be considered to estimate the feasibility (and adjustments) of the large-scale process.

The experimental methane production data obtained in the BMP assays were fitted to the First-order (Eq. 1) and Modified Gompertz (Eq. 2) kinetic models, both commonly used in batch systems [6, 36, 37].

$${\text{y}}\left( {\text{t}} \right) = {\text{y}}_{{\text{m}}} \left( {1 - {\text{exp}}\left( { - {\text{k}} \cdot {\text{t}}} \right)} \right)$$
(1)
$${\text{y}}\left( {\text{t}} \right) = {\text{y}}_{{\text{m}}} \cdot {\text{exp}}\left( { - {\text{exp}}\left( {\frac{{{\upmu } \cdot {\text{e}}}}{{{\text{y}}_{{\text{m}}} }}\left( {{\uplambda } - {\text{t}}} \right) + 1} \right)} \right)$$
(2)

Where: y(t) cumulative methane production (NmL.gVS−1), t experimental execution time (d), k constant hydrolysis rate (d−1), \({\mathrm{y}}_{\mathrm{m}}\) maximum methane production (NmL.gVS−1), \(\upmu\) maximum methane production rate (NmL.gVS−1.d.−1), \(\uplambda\) lag phase (d), n shape constant (dimensionless). Sources: [36, 37]

Each experimental condition (mixtures UTW, 3–24, 3–48 and 6–24) was fitted to an exponential curve (First-order) and non-linear regression (Modified Gompertz) using the Origin Pro 8.0 software. The performance of the proposed models was determined by comparison with the experimental values, calculating the residual sum of squares (RSS) (Eq. 3) and the determination coefficient (R2) (Eq. 4), where \({\mathrm{Z}}_{{\mathrm{f}}_{\mathrm{i}}}\) is the predicted value and \({\mathrm{Z}}_{{\mathrm{o}}_{\mathrm{i}}}\) the value observed.

$$\begin{gathered} \hfill \\ {\text{RSS = }}\sum\nolimits_{{\text{i - 1}}}^{{\text{N}}} {{\text{(Z}}_{{{\text{(f}}_{{\text{i}}} {)}}} {\text{ - Z}}_{{{\text{(o}}_{{\text{i}}} {)}}} {)}^{{2}} } \hfill \\ \end{gathered}$$
(3)
$$R^{2} = \left( {\frac{{\sum \left( {Z_{{f_{i} }} - \overline{{Z_{{f_{i} }} }} } \right)\left( {Z_{{o_{i} }} - \overline{{Z_{{o_{i} }} }} } \right)}}{{\sqrt {\sum \left( {Z_{{f_{i} }} - \overline{{Z_{{f_{i} }} }} } \right)^{2} \sum \left( {Z_{{o_{i} }} - \overline{{Z_{{o_{i} }} }} } \right)^{2} } }}} \right)^{2}$$
(4)

Results and discussion

FVW and inoculum characteristics

Tables S2 (Supplementary section) and 2 present, respectively, the physicochemical characteristics of the substrates and inoculum used and the mixtures (before and after the incubation period).

Table S2 shows that most of the FVW presented elevated moisture contents and volatile solids (VS) and a highly acid nature, corroborating the values reported in similar studies (also dealing with FVW) in the technical literature [6, 7, 10, 38, 39].

Table 2 shows that the moisture content of all the mixtures (with and without inoculum) was above 90%, favoring: dilution of the alkali reagent in the mixture thus increasing the pretreatment efficiency; distribution of the microorganisms and nutrients in the digester, and the dilution of toxic compounds (inhibitors of the biodigestion process) reducing their concentration in the medium.

The reductions in VS may be related to the action of the chemical reagent in degrading the lignin, increasing the biodegradability of the residue [40]. In fact, as can be seen in Table 2, the mixtures pretreated with NaOH showed greater percent reductions in VS (≈60%) as compared to the untreated residue (54.3%). Nevertheless, with respect to alkaline pretreatment, the literature review showed that elevated doses of the reagent did not necessarily imply in greater reductions in VS or even greater biogas generation. Zheng et al. [15] (see details in Table S1) observed that under the best biogas generation conditions (loading rate of 65 g.L−1), the reductions in VS were 61.5, 61.7 and 63.6% (for doses of 6, 4 and 2% NaOH, respectively), compared to 38.6% for the untreated residue. The results of Zhang et al. [17] (see details in Table S1) showed VS reductions of 60.6 and 70.4% for residues pretreated with 2 and 6% NaOH (the latter being identified as the best biogas generation condition) and only 21.3% for the residue pretreated with 10% NaOH, compared with a 59.8% reduction for the untreated residue. In the present study, duplication of the NaOH dose (from 3 to 6%) and of the pretreatment time (from 24 to 48 h) did not increase VS removal. Also, VS reductions in residues pretreated with NaOH can vary considerably as a function of the substrate characteristics, pretreatment conditions and biodigestion process. Alkanok et al. [41] evaluated the codigestion of fruit, vegetable and flower wastes from supermarkets under mesophilic conditions in batch digesters (S/I of 9:1 on w/w basis) and found that, despite the elevated initial volatile solids content (85%) and moisture content of these residues, the VS removal was relatively low (50%), suggesting that food preservatives may have impacted on the efficiency of the conversion of organic substrate to biogas.

A low pH value is characteristic of FVW (strongly acid in the present case: 4.7). All the mixtures started biodigestion with slightly alkaline pH values, more appropriate to methane generation and process stability (7 to 7.8 according to Raposo et al. [42]), and at the end of the process, the pH of all the mixtures was still above neutrality, indicating a well-balanced system during the biodigestion period. Treatment with NaOH has the advantage of maintaining the AD steps above neutrality as a function of the residual alkaline reagent provided by the system in the pretreatment steps. Nevertheless, in the present study, even the UTW mixtures (without NaOH or a buffering solution) maintained the biodigestion process without souring of the AD. This stable condition was guaranteed both by the capacity of the inoculum (already adapted to the substrate) to buffer the system, as to the greater proportion of inoculum (pH = 8.2) in relation to the substrate (0.5 gVSsubstrate.gVS−1inoculum ratio).

An elevated COD is also characteristic of FVW. In the present study, the mixtures showed initial values above 20,000 mg.L−1, indicating a high biodegradation potential [43], but the final COD values of all the mixtures were still high (the greatest conversion rate observed was ≈25% for mixture 3–24). Santos et al. [9] analyzed biogas production in the codigestion of orange bagasse, passion fruit peel and cashew bagasse (bio-digested individually) with industrial and sewage anaerobic sludge as the inoculum in BMP digesters (250 mL) identical to those used in the present study, under mesophilic conditions (37 ºC) and orbital agitation (60 rpm), but the COD removals were much higher (> 62%) than those found in the present study, independent of the FW and inoculum used.

Thus, the final VS values (all above 30%) and the percent COD removals (all below 25%) indicate an organic matter with degradation potential, even after 34 biodigestion days. According to Decottignies et al. [44], a residue can be considered stabilized when it shows a final VS content between 10 and 17.4%, a range well below the percentages found here.

Since cellulose and lignin are important carbon sources for the biological process and lignin constitutes a physical barrier to microbial degradation, there is a tendency for the removal of the lignin from the residue during the pre-treatment process to directly affect biogas production during AD. As can be seen in Table 2, although the alkaline pretreatment reduced the lignin contents by 17.3% when comparing the UTW mixture with the 3–24 mixture, only 2.3% was removed when comparing mixtures 3–24 and 3–48 and 7.1% when comparing mixtures 3–48 and 6–24, showing that duplication of the pretreatment time and of the chemical reagent concentration did not result in a significant and proportional removal of lignin from the mixtures. The same conclusion can be made from an analysis of the (C + H)/L ratios, where an increase of 8.4% was observed between the UTW and 3–24 mixtures, but only 2.7% between the 3–24 and 3.48 mixtures and 3.2% between the 3–48 and 6–24 mixtures. Hence, although the pretreatment step may have been fundamental to the carbohydrate-lignin complex degradation process (facilitating the enzymatic hydrolysis of C + H) under the conditions of the present study, greater exposure times of the residues to the NaOH and higher NaOH concentrations did not result in the expected delignification process. Zheng et al. [15] pretreated corn stover with different concentrations of NaOH (2%, 4% and 6% based on the dry matter of the corn stover) for 3 days at 20ºC and observed a reduction in the L/C ratio of 7.6% between the untreated corn stover and the corn stover treated with 2% NaOH, but only 2.6% between the corn stover samples treated with 4% NaOH and 6% NaOH. Pellera et al. [45], pretreating olive pomace with different concentrations of NaOH (0 to 4 mmol.gVS−1, from 1 to 24 h at 25ºC), couldn’t identify significant changes in lignin removal for reagent concentrations above 1 mmol.gVS−1 after 16 h of pretreatment. These and other studies reveal that, in fact, the behavior of the pretreatment in terms of the variation in the chemical composition of the substrate can change a lot. The limited variation observed in the lignin contents with increase in concentration of the alkaline reagent could be the result of a recondensation process of the solubilized lignin (which occurred during the pretreatment step), which could also imply in a reduction in digestibility of the substrate [46].

Biogas and methane yield

Figure 1 shows the daily production of biogas and methane during the 34 days of monitoring. The values reported are a mean of the triplicate of each treatment (UTW, 3–24, 3–48 and 6–24), having already subtracted the mean volume of biogas generated in the blank assays.

Fig. 1
figure 1

Methane (continuous line) and biogas (dashed line) daily rates for untreated and pretreated wastes. The methane concentration was determined as from the second day of monitoring of the process

Full size image

In general, the large biogas generation peaks seen at the start of biodigestion must be due to the quick adaptation of the microorganisms to the substrate and to the elevated hydrolysis rate of the easily digested organic compounds, such as carbohydrates (cellulose and hemicellulose), present in considerable concentrations in FVW [47, 48]. Since the mixtures evaluated could be composed of hundreds of molecules (with different degradation kinetics), the “zigzag” behavior with peaks distributed throughout the whole period follows the same pattern seen in the technical literature for the biodigestion of typically lignocellulosic residues.

The hydrolysis phase, the first step in AD, directly influences the biodegradation velocity of lignocellulosic biomass. Although the patterns of the daily biogas rate curves for the first days of biodigestion were very similar for the four treatments (including UTW), the first peaks in Fig. 1 (see enlargement) show a slightly larger biogas production by the mixtures pretreated with NaOH, clearly demonstrating that the chemical reagent aided access of the hydrolytic bacteria to the substrate at the start of the process in the case of pretreated residues. During the pretreatment, the alkali reagent favored rupture of the bonds in the lignin-carbohydrate complex, contributing to microbial access to the cellulose and, consequently, to biogas production [40, 49]. In fact, Table 2 shows that the alkali reagent contributed to an increase in biodigestion of the principal components of the lignocellulosic biomass. The reductions in the LCH contents between the mixtures UTW/3–24, 3–24/3–48 and 3–48/6–24 were 11.5; 4.5; and 4.6%, respectively, and these values are compatible with the biogas production observed in the first days of the process. At the end of the fourth day of biodigestion, the biogas productions by the mixtures 3–24, 3–48 and 6–24 were already 62, 24 and 16% higher than the biogas volume produced by the untreated sample (UTW).

Figure 2 shows the accumulated production of biogas and methane throughout the 34 days of monitoring, the values reported being the means of triplicates for each treatment (UTW, 3–24, 3–48 and 6–24) after subtracting the mean biogas volume generated by the blank assays.

Fig. 2
figure 2

Accumulated volumes of methane (continuous line) and biogas (dashed line) for the untreated and pretreated wastes. The methane concentration was determined as from the second day of monitoring the process

Full size image

The pattern of the biogas accumulation curves ratified the fact that the substrates evaluated were readily biodegradable, thus indicating that the hydrolysis phase was not a rate-limiting step of the process [50]. The values obtained for the total volume of biogas generated by treatments 3–24 and 3–48 were very close (293 and 300 NmL, respectively) and higher than those generated by mixtures 6–24 and UTW (247 and 191 NmL). These results corroborate that stated previously concerning lignin removal that, in terms of biogas production, duplication of the pretreatment time and chemical reagent concentration did not result in a significant and proportional increase in biogas between the mixtures. In a study of the pretreatment of corn stover with three doses of NaOH (2%, 4% and 6%), Zheng et al. [15] found that the lignin contents decreased and the (C + H)/L ratio increased with increase in NaOH concentration (from 2 to 6%), but the methane volumes remained practically constant (211 and 208 NmLCH4.gVS−1 for 2 and 6%, respectively, for a loading rate of 65 g.L−1). Calabró et al. [51], pretreating tomato processing waste with different NaOH concentrations (1% for 24 h, and 5% for 4 and 24 h), observed no significant differences in the volume of methane generated, independent of exposure time and reagent concentration. In this case the authors attributed the accumulation of salts in the reactors with 5% NaOH as one of the probable causes for the low biogas generation. Günerhan et al. [16], evaluating methane production from the simultaneous biodigestion of various fruit and vegetable harvesting wastes pretreated with different NaOH concentrations, reaction times and temperatures, also showed that an increase in NaOH concentration (from 3.3 to 6.5%) did not contribute to greater methane generation. The authors emphasized that methane production depended a lot on the process conditions and on the lignocellulosic biomass characteristics.

In fact, in addition to the reagent concentration, pretreatment efficiency (be it acid or alkaline), can vary a lot as a function of the residue characteristics, process conditions (operational parameters), inoculum used, BMP assay protocol, etc., making it difficult to compare different studies in terms of biogas generation per mass of substrate [3, 52]. Even the digestion of different parts of a fruit or vegetable (stalk, seed, peel, etc.) can influence a greater or smaller generation of biogas [53, 54]. Nevertheless, despite the difficulty in comparing the results of the biodigestion of different lignocellulosic feedstocks, for the reasons mentioned above, many of the values seen in Table S1 call attention to their “elevated” methane production as compared to the accumulated generation in the present study (Table 3).

Table 3 Total methane volume accumulated during the monitoring period (mixtures UTW, 3–24, 3–48 and 6–24)
Full size table

Table S1 shows that all the substrates pretreated with doses of up to 6% NaOH produced amounts of methane superior to the best conditions observed in the present experiment (mixture 3–24 with ≅ 127 NmLCH4.gVS−1). In the studies of Calabrò et al. [51], Mancini et al. [55] and Zhang et al. [17], for example, even the biodigestion of substrates with much higher lignin contents than those determined in the present study, obtained final methane volumes close to 200 mLCH4.gVS−1 (frequently above 300 mLCH4.gVS−1). A great variety of inhibitory compounds have been cited in the literature as related to the low accumulation of methane in anaerobic digestion processes (depending on the concentrations in which they are present in the medium): ammonia, sulfide, light metal ions (such as sodium, potassium, magnesium) and heavy metals (iron, copper, lead, zinc, chromium, mercury, etc.), chlorophenols (used as pesticides), halogenated aliphatic compounds, etc. [56,57,58]. Specifically in the case of pretreatment with NaOH, some of these inhibitors may be formed during the process, and even during biodigestion (methanogenesis), the presence of sodium ions (liberated in the aqueous phase due to dissociation of the NaOH) could have an inhibitory effect. Phenolic compounds, furfurals and hydroxymethyl furfurals, aromatic compounds, may also form in the presence of excess NaOH [13, 19, 45, 59]. Some inhibitors may be related to the nature of the biodigested substrate, such as d-limonene from the digestion of orange peel [60] and food preservatives in the codigestion of fruit, vegetable and flower wastes from supermarkets [41]. Thus, in the context of the present study, although the simultaneous codigestion of various substrates may (theoretically) favor a nutrient balance and a positive synergism of the microbial community in the medium, it must be pointed out that some of the food residues present in the mixtures may contain elevated pesticide contents (not evaluated in the present study). Fruits and vegetables such as apple, kale, collard, pepper, nectarine, tomato and oranges are amongst the unprocessed foods that could contain various pesticides in terms of quantity and/or concentration [61, 62] which could have compromised biogas generation in all the mixtures. The process yield is also dependent on a complex consortium of microorganisms present in the system and whose characteristics were not evaluated in the present study.

Table 3 shows the time required for each mixture to produce 90% of the total volume of accumulated methane (t90). Although the higher alkali reagent (6%) concentration shortened the digestion time by 16% in relation to the residues pretreated with 3% NaOH, the net volume of methane generated in the mixture 6–24 was considerably smaller than that of treatments 3–24 and 3–48 (≅ 25% less). Thus, although the addition of a chemical reagent frequently reduces the biodigestion time [15], it must be observed that this reduction in the hydraulic retention time may also compromise the efficiency of the process with respect to a reduction in biogas production. A comparison of treatments 3–24 and 3–48 with the UTW mixture showed a 10% shorter digestion time for the mixtures pretreated with 3% NaOH, with a 65% greater total methane volume, proving the efficiency of the pretreatment process with this dosage of the chemical reagent, at least under the conditions of this study.

Figure 3 shows the methane concentrations in the biogas for each mixture evaluated (UTW, 3–24, 3–48 and 6–24) for the nine measurements distributed throughout the 34 days of monitoring.

Fig. 3
figure 3

Composition of the biogas from the treated and untreated FVW mixtures

Full size image

In addition to producing the greatest biogas volume per substrate mass, the mixtures with 3% NaOH also presented the largest methane contents in all the trials. Considering all the nine measurements of the biogas composition, the mean methane contents in the biogas were 63.0(± 1.4), 65.7(± 1.2), 66.1(± 0.9) and 64.0(± 1.4)% for the mixtures UTW, 3–24, 3–48 and 6,24, respectively (mixtures 3–24 and 3–48 presented statistically equal mean methane concentrations, significantly higher than those of mixtures 6–24 and UTW). These values, both for the pretreated and untreated residues, also show great difference in the methane content produced from FVW when compared to other studies found in the technical literature which used the same alkali reagent. Dasgupta and Chandel [59], in their best pretreatment conditions, obtained biogas with 68.9% of methane from the residue treated with 6% NaOH and 50.3% of methane in the biogas from the untreated residue. Even with the generation of more than 370 NmL.gvs−1 of biogas, Zhu et al. [19] obtained a corresponding methane content of only 52% for the pretreatment with 5% NaOH. Pavi et al. [39], codigesting various FVW (inoculated with anaerobic sludge) collected from a supermarket in Brazil, and found maximum methane contents of 78.7%, despite obtaining a considerable volume of biogas (350 NmL.gvs−1).

Since the energy content of biogas takes into consideration the volume of biogas produced and the methane concentration in this biogas [63], the concentration of 6% NaOH in the pretreatment step failed to collaborate with a representative increase in energy gain by the system, probably due to the excess reagent in the medium and the previously mentioned implications of this.

Kinetic analysis

Figure 4 shows the kinetic graphs of methane production under the experimental conditions of this study.

Fig. 4
figure 4

Experimental curves for methane production and the curves generated by the kinetic models. Legend: (A) UTW mixture; (B) 3–24; (C) 3–48; (D) 6–24

Full size image

Although Fig. 4 shows that all the treatments (mixtures) presented a good fit to the proposed kinetic models, Table 4 shows that the values of the determination coefficients (R2) for the curves of the First-order kinetic model were higher for all the experimental configurations. Also, the values for RSS were smaller for the fitting of the data to the first-order kinetic model, showing that this model was the most adequate to simulate the methane production curves under the conditions of the present study.

Table 4 Kinetic and statistical parameters of the models, for each experimental configuration
Full size table

As with the other process parameters already considered, the hydrolysis constant rates (k) can also vary considerably as a function of the biodigestion conditions and substrate composition. In general, low k values indicate slower biodegradation. In the present study, the values obtained for this parameter were low, if compared to the range of 0.39 to 0.66 d−1 obtained by Edwiges et al. [6], or 0.14 to 0.49 d−1 found by Li et al. [48] (both working with FW biodigestion). On the other hand, Oliveira et al. [58], biodigesting FW from a university restaurant, obtained a hydrolysis constant rate of only 0.0012 d−1, linking this low value to poor degradability of the substrate. The values for ym generated by the kinetic models presented values very close to those obtained for maximum methane productions under experimental conditions. The maximum methane production rates (μ) varied from 2.466 to 2.944 NmL.gVS−1.d−1 and presented the same tendency observed for the early peaks of the methane daily rate during the first days of biodigestion (Fig. 1). We also observed that the pre-treatments adopted favored the optimization of the μ parameter, mainly regarding mixture 3–24 (Table 4). Finally, none of the mixtures evaluated showed lag times (λ), corroborating with that observed in Fig. 1, that is, rapid digestion of the organic matter as from the first hours of the experiment.

Conclusions

The results obtained showed that, even with the synergistic effect of the codigestion of various FVW (plus inoculum), slow-degrading substrates (such as lignocellulosic compounds) require a pretreatment step before anaerobic digestion to improve the biodegradability of the residue and, consequently, biogas production. Under the conditions of the present study, duplication of the pretreatment time (from 24 to 48 h, with 3% NaOH) did not increase methane production. Duplication of the reagent concentration (from 3 to 6% NaOH) produced a significant decrease in methane production from ≈127 to 102 NmLCH4.gVS−1. Specifically in the case of pretreatment with sodium hydroxide, some inhibitors (Na + , phenolic compounds, furfurals and hydroxymethyl furfurals, aromatic compounds) could be associated with higher concentrations of the chemical reagent used in the pretreatment step, which could explain the smaller methane generation in mixtures pretreated with 6% NaOH.

It is important to point out that the residue characteristics, operational parameters, bio-digestion conditions, type of inoculum, inoculum/substrate ratio, and the bioreactor characteristics, etc., directly influence the total volume of methane generated, making it difficult to compare with other studies or propose “optimum pretreatment conditions” for different types of residue and/or process. Hence for large-scale operations, such as the biodigester operating at the central market for food distribution (CEASA) in Recife, the fact of working under biodigestion temperature conditions slightly above room temperature, lower hydraulic retention times (including the pretreatment step) and low alkali reagent concentrations, certainly contribute to the low operational costs of the plant, corroborating with the conditions proposed in this study.