3.1 Characterization of initial pig manure (PM) and degassed pig manure (DPM)
Table 1 shows the values for physicochemical parameters and bacterial pathogen counts in pig manure, before and after degassing (PM and DPM).
Table 1
Physicochemical parameters and bacterial pathogen counts in initial pig manure (PM) and after degassing (DPM). Values represent mean ± SD of three different experiments.
Physicochemical parameters
|
PM
|
DPM
|
Total Solids TS (% m/m)
|
5.15 ± 0.05
|
2.25 ± 0.07
|
Volatile Solids, VS (% m/m)
|
33.35 ± 0.69
|
28.06 ± 0.38
|
Chemical Oxygen Demand, COD (g/l)
|
58,81 ± 0.95
|
16.39 ± 0.50
|
Biochemical Oxygen Demand, BOD (g/l)
|
23.90 ± 0.76
|
8.65 ± 0.66
|
pH
|
7.44 ± 0.03
|
7.69 ± 0.09
|
Total Ammoniacal Nitrogen TAN(g/l)
|
2.02 ± 0.06
|
1.17 ± 0.03
|
Free Ammoniacal Nitrogen FAN(g/l)
|
2.08 ± 0.04
|
1.01 ± 0.05
|
Total Kjeldahl Nitrogen, TKN (% m/m)
|
2.24 ± 0.05
|
1.54 ± 0.07
|
Total Carbon, C (% m/m)
|
19.39 ± 0.40
|
16.31 ± 0.38
|
C/N
|
8.69 ± 0.18
|
10.59 ± 0.20
|
Volatile Fatty Acids VFA (g/l)
|
0.97 ± 0.01
|
0.51 ± 0.02
|
Total Alkalinity TA (g/l)
|
2.96 ± 0.03
|
2.87 ± 0.03
|
Sodium Absorption Ratio SAR
|
35.11
|
31.54
|
Percentage exchangeable sodium PES
|
35.82
|
33.21
|
Potassium (mg/kg DM)
|
72 ± 3
|
80 ± 2.2
|
Phosphorus (mg/kg DM)
|
134 ± 0.20
|
175 ± 0.16
|
Sodium (mg/kg DM)
|
2322 ± 3
|
1978 ± 2.1
|
Arsenic (mg/kg DM)
|
0.82 ± 0.01
|
0.90 ± 0.01
|
Zinc (mg/kg DM)
|
5.50 ± 0.10
|
5.20 ± 0.04
|
Chromium (mg/kg DM)
|
1.70 ± 0.02
|
1.54 ± 0.04
|
Magnesium (mg/kg DM)
|
55 ± 3
|
47.30 ± 1.2
|
Mercury (mg/kg DM)
|
0.002 ± 0.001
|
ND
|
Nickel (mg/kg DM)
|
0.88 ± 0.001
|
0.72 ± 0.009
|
Lead (mg/kg DM)
|
0.04 ± 0.001
|
0.02 ± 0.001
|
Bacterial pathogens
|
|
|
Escherichia coli (UFC/100ml)
|
0.80
|
ND
|
Fecal coliforms (NMP/100ml)
|
0.84
|
ND
|
Salmonella spp. (UFC/100ml)
|
ND
|
ND
|
Yield parameters
|
|
|
Organic matter removal OMR (% VS)
|
63.23 ± 1.07
|
-
|
Biogas yield (Nml/g VS)
|
298.77 ± 1.04
|
-
|
Methane yield (Nml/g VS)
|
138.02 ± 1.03
|
-
|
Analysis of samples in triplicate. N: STP. ND: not detected
The parameters that decreased most significantly after degassing included chemical oxygen demand (COD) and biochemical oxygen demand (BOD5). In the first case, the difference between PM (58.81 g/l) and DPM (16.39 g/l) was 42.42 g/l. In the second case, the difference between PM (23.90 g/l) and DPM (8.65 g/l) was 15.25 g/l. Moreover, total solids (TS) decreased by 2.9%, (from 5.15% m/m in PM to 2.25% m/m in DPM). The changes in these three parameters (COD, BOD5, and TS) were likely due to a reduction in organic matter after degassing.
On the other hand, the increase in the C/N ratio in DPM can be attributed to a reduction in total nitrogen, among other factors. As described by Shen et al. [26], a low C/N ratio hinders continuous biogas production. Consequently, high biogas yields would not be possible with PM as the sole substrate, and supplementary substrates would be necessary to improve the C/N balance [26–28].
The values for AGV, TA, pH, and TAN in DPM were adequate for the production of biomethane, according to reference values by Holliger et al. [22]
and the VS value exceeded TS by 50% (VDI 4630. 2016) [17]. Other factors also point to DPM being an appropriate inoculum for AD. First, its VFA/TA ratio (which indicates the organic load) was 0.17. Values between 0.3 and 0.4 signal optimal biogas production; those above 0.4 indicate substrate overload, while those below 0.3 correspond to low substrate input [29]. In other words, DPM appears to have reached a high degree of starvation. Additionally, due to supplements added to animal feed, DPM contains and preserves macro- and micronutrients and metals that optimize the metabolic reactions in AD. Cai et al. [30] reported improvements in biomethane yield after incorporating micronutrients into animal feed.
Moreover, starvation under anaerobic and mesophilic conditions reduces or eliminates pathogens, which explains why pathogenic bacteria were detected in the initial slurry but not in DPM.
During degassing, PM yielded 298.77 NmL/g VS of biogas and 138.02 NmL/g VS of methane. The latter was approximated to the volume calculated on the basis of chemical composition (157.84 NmL/g VS) [31]. Even higher potential values than these (181 NmL/g VS) were reported by Flotats et al. [32]. The organic matter removal rate (OMR) in our assay was also significant (63.23%).
Figures 1a and 1b respectively show the daily and cumulative production of biogas and methane in each batch (or biodigester), as well as their positive controls. The test lasted 30 ± 2 days, until less than 1% of the total volume was produced. In terms of composition, 46.21% of the biogas before degassing was made up of CH4, which is 9% lower than what was reported by Regueiro et al. [33]. During degassing (Fig. 1c), the percentage of CH4 peaked on day eight (64.08%). That same day, the rest of its composition consisted of 31.87% CO2, 3.86% H2/N2, and 0.19% H2S. In general, CH4 percentages were at their highest when production was also at its highest (days 8–15), while CO2 and N2/H2 remained stable in the ranges of 20–30% and 10–15%, respectively.
The calorific value of biogas is known to climb when its CH4 content increases, and to drop as CO2 rises. A possible cause for CO2 increasing in our trial might be the accumulation of VFA during AD. This shifts the bicarbonate balance towards CO2 to maintain the pH, as long as the alkalinity of the system supports it. Moreover, high percentages of CO2 may create an acidic environment in engines during the transformation of biogas into electricity [34].
The percentages registered for H2/N2 during degassing were also high. Such values make partial pressure higher than what is suitable for acetogenesis, and thus render it thermodynamically impossible. High H2 concentrations may indicate that the digester is overloaded [35]. Finally, the specific methanogenic activity (SMA) of DPM was 0.12 CH4. gCOD/gVSinoc, despite its degree of dilution and its relatively low content of VS. This value is higher than that reported by Angelidaki et al. [20] and Holliger et al. [22] (0.1), and that the one obtained by Astals et al. [23] for inocula from effluent ponds.
3.2 Acclimation of the inoculum
Figures 2a and 2b respectively show the daily and the cumulative production of biogas/CH4 during each acclimation pulse (P1, P2, P3, and P4). Between P1 and P4, the volume of biogas doubled and that of methane tripled. The lag phase for CH4 production shortened as the trial went on (Fig. 2b), which means that the system was able to withstand the stress caused by the organic load of each pulse. This increase in production and the ability to tolerate high organic loading rates may be attributed to the specialized microbial community that emerges in the digester during acclimation [9].
Besides, due to the low solid content applied with each pulse and the characteristics of CSW (high percentage of starch), organic matter was transformed into biogas more quickly. In other words, the pulse feeding pattern appears to have made the biodigesters more tolerant to organic loads and higher levels of total ammonia nitrogen (TAN), and thus more efficient at producing biogas.
The composition of biogas in each batch during each acclimation pulse, in terms of component percentages, is shown in Fig. 3c. After an initial increase, the percentage of CH4 remained at around 60–64% from P3 onwards. This was correlated with a decrease in CO2, whose values then stabilized between 20 and 25%. Although H2/N2 rose dramatically upon the application of P1, they rapidly dropped to 10% and remained stable during the other three pulses. This may have prevented additional stress on the system, and better allowed the microbial biomass to adapt. Syntrophic microorganisms can proliferate during acetate oxidation (SAO) by transforming acetic acid into H2 and CO2 [36] while CH4 production remains stable. The biogas/methane yield was 1.6 and 2.8 times higher than the initial values after P1 and P2, respectively (Fig. 2d). The values for OMR based on VS remained between 55 and 64% throughout the four pulses (Fig. 2e).
Even though VFA increased at the end of P1, they then decreased considerably and from thereon stayed more or less the same during P2, P3, and P4 (Fig. 3a). Put otherwise, there was an initial spike in VFA associated with the system's rapid adaptation to substrate consumption, but once this had been achieved, the levels fell and each successive pulse brought about slightly decreasing values. This reduction was correlated with an increase in TA, whose values then stayed at around 6 g CaCO3/L throughout the trial (Fig. 3b). Thus, the VFA/TA ratio (which is widely taken to indicate biodigester performance) was balanced enough to allow the system to withstand pH variations. Similarly, Kim et al. [37], who attempted to reduce the AD dormancy phase and attain higher yields by using substrates with a high organic load, concluded that for this to happen the VFA/TA ratio should be below 0.4 and the initial VFA/TS ratio should be less than 10%. Finally, the values registered for FAN, TAN, and NH3 in our trial were within the stable range reported for other inocula (0.5–1.2 g/L) [22].
Although, as seen in Figs. 3a and 3b, the percentages of CH4 had tripled by the end of the trial, they increased mainly during the first two pulses and did not significantly rise after P2. This means the system reached its maximum CH4 production at that point, and it is the reason why the test was stopped after four pulses. In addition to the lag phase being shorter with each successive pulse, the composition of the biogas also became progressively less variable, which is advantageous in terms of final biogas quality.
In general, these findings demonstrate that biodigesters exposed to perturbation in the form of organic feeding achieve better biogas and methane yields (without considering endogenous biogas production), a similar conclusion to that reported by Wang et al. [12]. Other authors propose using disturbance as a strategy to influence methanogenic microbiomes and improve the co-digestion of critical waste, and suggest that acclimated inocula are essential to enhance biogas production by anaerobic co-digestion [38].
All these parameters described so far are important and complement each other. However, it is biogas composition in particular which is essential to implement this kind of feeding regime at a larger scale, since it determines biogas quality. Higher solid concentrations than those used here (< 10%) might produce higher yields, but the system’s stability might not be guaranteed. Nevertheless, each AD system has its own working limits and interactions that arise between variables.
3.3 Microbial characterization
The microbiological characterization was based on 227,590 sequences found in the three samples analyzed: PM (initial pig manure), DPM (degassed manure), and APM (acclimated manure). The average number of sequenced fragments (251 bp) was consistent with the sequencing method used. PM contained 17219 sequences and DPM, 14993. In contrast, APM had a smaller number (9151). Amplicon sequence variants (ASVs) were obtained on DADA2, and taxonomically classified (with 100% identity) using the SILVA database (v 138.1).
Table 2 shows the number of ASVs and the values of Shannon's diversity index and inverse Simpson's index for the three samples. APM had fewer taxa and lower diversity than PM and DPM, which is consistent with its smaller number of sequences. The change in conditions in the biodigesters (acclimation with pulse feeding) was likely responsible for this reduction in microbial diversity.
Table 2
Number of amplicon sequence variants (ASVs), Shannon’s diversity index (richness), and inverse Simpson's index (equity) for PM, DPM, and APM.
Sample
|
Number of ASVs
|
Shannon
|
1/Simpson
|
PM
|
418
|
5.68
|
232
|
DPM
|
417
|
5.68
|
228
|
APM
|
256
|
5.12
|
132
|
Figure 4a shows the relative abundance of phylotypes (% RA), separated into taxonomic levels (domain (K), phylum (P), order (O), and genus (G)) in PM, DPM, and APM. Bacteria were the most prevalent domain in all three. This does not mean there were no Archaea, but rather that the sequencing primers likely originated from bacteria and thus have less affinity with Archaea. The predominant phyla were Bacteroidetes, Proteobacteria, Firmicutes, and Desulfobacterota, particularly in PM and DPM. Bacteroidales and Burkholderiales were the most relevant orders, the latter mainly in DPM. At the genus level, Sulfuritalea was only present in PM (9% RA).
The relative abundance of the 70 most significant ASVs in the samples can be seen in Fig. 5b. Abundance was low for all the taxa and none of them was dominant. Moreover, the most abundant sequences represented only 2% of the total. The samples did not share many sequences, but instead each had its own specific composition.
These 70 ASVs were put through BLAST + to find matches (99% identity) with full-length ASVs available on the NCBI (National Center for Biotechnology Information) database. Sixty of them matched 1269 sequences on the database, while the other ten had no matches and might correspond to as yet unreported microorganisms. Most of the sequences belonged to uncultured microorganisms.
Figure 5 shows the sources from which were isolated the sequences on the NCBI database that matched the ASVs in our study. Most of them came from anaerobic bioreactors, followed by a large percentage that is typical of wastewater treatment processes (“WWTP”). Others were found in anaerobic environments unrelated to reactors, or were associated with livestock farming (manure lagoons, slurry, etc.) [39]. Importantly, all these environments are related to the process studied here, and they cover most of the sequences found (~ 80%). There were, nevertheless, other matching sequences that came from the petrochemical industry (“Oil & Gas”); lakes, rivers, and rhizospheric environments associated with agriculture; mining activity, urban landfills, and other laboratory processes (“Other”).
The most important ASVs detected in PM were ASV 7, ASV 12, ASV 4, and ASV 17. The first two are Sulfuritalea spp., and neither of them matched the NCBI data. This makes sense considering that only one species of this genus has been isolated so far (Sulfuritalea hydrogenivorans) [40]. This is an autolithotrophic, sulfur-oxidizing, and nitrate-reducing neutrophil which has been described not only in several aquatic environments with low carbon loads, but also in activated sludge and hydrocarbon-contaminated sites [41].
ASV 4 is a member of Novosphingobium, a genus of facultative aerobic organotrophs that can reduce nitrate and are involved in the degradation of aromatic compounds [42]. Sequences identical to ASV 4 were isolated from landfills and WWTP. ASV 17, a Pelospora sp., showed identity with sequences from an activated sludge, an anaerobic digester of livestock waste, and municipal landfills. Only one species of this genus, a strictly anaerobic glutarate fermenter, is known so far (Pelospora glutarica).
The most predominant ASV in DPM was ASV 1, from the family Anaerolineaceae, which represents only 2% of the community. Furthermore, the same sample featured ASV 9 (genus Mesotoga), ASV 11 (also belonging to Anaerolineaceae), ASV 32 (genus Pseudomonas), and ASV 6 (a taxon of the class Clostridia). Sequences similar to those of ASV 1 and ASV 11 have been widely reported in wastewater treatment systems and anaerobic digesters, including biogas plants [43]. Members of the Anaerolineaceae family include strict anaerobes, mesophiles, thermophiles, and chemoheterotrophs. In addition, some studies have shown that they can interact syntrophically with methanogenic microorganisms and generate hydrogen [44].
ASV 9 was classified as Mesotoga infera, a species within the order Thermotogales which has been found in anaerobic digesters containing high loads of carbonaceous compounds (including hydrocarbons such as toluene, benzene, and xylene) at high temperatures (65–85°C). Mesotoga spp. are mesophilic, and they use sulfur compounds as electron acceptors to produce sulfur, acetate, and CO2, but do not seem able to produce hydrogen [45].
The genus Pseudomonas, to which ASV 32 belongs, comprises a wide variety of species that are capable of obtaining energy from complex carbon compounds. Sequences similar to ASV 32 have been found in bioreactors, rhizospheric environments, sludge, and wastewater. As degraders of complex carbon sources, Pseudomonas spp. may be some of the most important microorganisms in anaerobic digesters [46]. ASV 6, from the class Clostridia, is related to anaerobic reactors as well.
The most relevant ASV in APM was ASV 13, another Pseudomonas sp. (2.1% RA). Other ASVs which were identified in this sample are ASV 2 and ASV 3 (both from the order Bacteroidales), ASV 19 (from the order Synergistales), and ASV 39 (from the class Clostridia). ASV 2 and 3 were similar to each other and had matches on the NCBI database, which corresponded to uncultured microorganisms from anaerobic bioreactors that produce biogas from agricultural waste (lignocellulolytic waste, bovine albumin, and pig farming waste). ASV-19 like sequences (synergists) have been reported in anaerobic bioreactors used for municipal wastewater treatment. Synergists are associated with animal microbiota and frequently found in anaerobic digesters that produce amino acids and degrade proteins, where they establish a syntrophic relationship with methanogens [47–48]. Finally, only five sequences on the NCBI database matched ASV 39 (Clostridia), four of which came from an anaerobic reactor fed with swine waste (the same one where the sequences that matched ASV 2 and 3 were detected).