Potential of anaerobic co-digestion of acidic fruit processing waste and waste-activated sludge for biogas production

2.2.1 Batch experiments

Batch experiments were carried out in 500 mL reactors (effective volume of 400 mL), each covered with a rubber septum. Triplicate bottles were used in all batch experiments, and all data were expressed as the mean of triplicate ± standard deviation. Among the various samples, FW1 and FW7 were 100% FPW and 100% WAS as controls. Then, the effects of the VS ratios of FPW and WAS on biogas production were investigated. Typically, 250 mL of inoculated sludge and feedstock under different feeding ratios of FPW and WAS at the same organic load was added to each group of reactors, which were then added with water to a volume of 400 mL. After stirring evenly, it was purged with 99.99% pure nitrogen for 5 min so that the oxidation-reduction potential reached to −300 mV to ensure an anaerobic environment. The fermentation proceeded at 35 ± 0.5°C for a period of 40 day under 20 min of stirring at 100 rpm every 12 h. Biogas and methane production was detected online using a fully automated methanogenesis potential testing system (AMPTS II, Bioprocess Control). The feed-to-microbe (F/M) ratio was calculated based on the initial VS of the substrate and inoculum. Detailed experimental conditions and mixing ratios are summarized in Table 2. The F/M ratio was calculated based on the initial VS of the substrate and inoculum in Eq. 1:

2.2.2 Continuous experiments

A lab-scale one-phase anaerobic digester (Bioprocess® CSTR-5S, Sweden) with a 4 L working volume was used for the co-digestion of FPW and WAS (Figure 1). A water bath was used to maintain the digester in a mesophilic environment (35 ± 0.5°C). The reactor was mixed using dedicated heaters (3619 Aquarium Heater 300 W, Eheim Jager, Germany) until a homogeneous state was reached. Then, 4 L of the inoculum was added initially and flushed with N₂, ensuring that the oxidation-reduction potential reached to −300 mV to establish anaerobic conditions. FPW was diluted with water to a VS concentration of 8% before feeding from Day 1 to Day 230.

Figure 1

Schematic and material object diagram of the lab-scale FPW and WAS co-digestion processes.

Five operation stages were applied (Table 3). The feedstock was added once a day, and the digestate was discharged every three days. The effluent pH, ammonium nitrogen, volatile fatty acids (VFAs), and alkalinity were analyzed every three days. Biogas and methane were detected online.

3.1.1 Performance of biogas production and methane content in batch experiments

To study the methanogenesis capacity of substrates with different VS ratios, the cumulative biogas production obtained from the reactors under 7 feeding ratios of FPW to WAS is illustrated in Figure 2. Under the same total feeding VS, the biogas production first increased and then decreased following a gradual increase in the proportion of WAS. Because of its low pH and easy degradation, FPW fermented alone can be rapidly converted into small molecular acids at the early stage, resulting in abundant acid accumulation that inhibits the activity of methanogens and leads to fermentation failure (Figure 2a). With the addition of WAS, the acidification during fermentation was significantly improved. When the VS ratio was 2:1, the amount of cumulative biogas produced in the reactor was maximized to 4695.47 mL (Figure 2c), which increased by 5.2% and 10.5% compared with the VS ratios of 3:1 and 1:1, respectively. Then, with a further increase in the WAS proportion, the biogas production gradually decreased, because the hydrolysis stage is the rate-limiting step in WAS methane production. When WAS was fermented alone (Figure 2g), the cumulative biogas production was only 2,234.4 mL, which is consistent with some previous studies [15].

Figure 2

Daily and cumulative biogas production under different VS ratios: (a) FW1, (b)FW2, (c) FW3, (d) FW4, (e) FW5, (f) FW6, (g) FW7. (h) Overall methane content under different VS ratios.

With an increase in the proportion of WAS, the overall methane content gradually rose (Figure 2h). Raising the proportion of WAS improved the lipid and protein contents of the feed, including lipid (1,014 mL CH₄·g VS⁻¹), carbohydrate (415 mL CH₄·g VS⁻¹), and protein (496 mL CH₄·g VS⁻¹) contents, according to theoretical methane production (TMP) [36]. Hence, changes in the components of the feed will affect TMP and thus the overall methane content (Table 5).

3.1.2 Conversion efficiency of hydrolysis, acidogenesis, acetogenesis, and methanogenesis in anaerobic co-digestion

To analyze the rate-limiting step in the anaerobic co-digestion process, we calculated the conversion ratio of the four stages of hydrolysis, acidogenesis (VFA), acetogenesis (acetic acid, CO₂, and H₂), and methanogenesis [37]. The conversions were calculated from the initial and final COD values of anaerobic co-digestion at different VS ratios. As shown in Figure 3, as the proportion of sludge in the substrate increased, and the methane conversion ratio showed a trend of first increasing and then decreasing. For FW1, FW2, and FW3, apple content was higher, and apple is a substrate that is highly anaerobically digestible. Apples are rich in a variety of soluble sugars (fructose, sucrose, glucose, and sorbitol), which can be converted to pyruvate through the Embden–Meyerhof–Parnas pathway [38]. It is rich in a variety of organic acids, such as malic acid content of 29.4–32.2 mg·g⁻¹ TS and citric acid content of 4.44-5.19 mg·g⁻¹ TS, in addition to lactic acid, succinic acid, fumaric acid, oxalic acid, and other small-molecule organic acids, so that the initial concentration of organic acids in apple raw materials is relatively high. In addition, the macromolecular carbohydrates (pectin, cellulose, and hemicellulose), proteins, and fats contained in apples are also easily converted into small molecules such as monosaccharides, amino acids, and fatty acids, and finally into acetic acid, CO₂, and H₂, which can be used by methanogens. FW1 and FW2 rapidly accumulated a large amount of VFA during the early stage of fermentation, which inhibited the conversion of acetic acid and the production of methane so that VFA could not be converted into methane smoothly. The conversion rate of the four stages is hydrolysis > acidogenesis > acetogenesis > methanogenesis. The methane conversion ratio was the highest in FW3, and there was no inhibition of the methane conversion process.

Figure 3

Hydrolysis, acidogenesis, acetogenesis, and methanogenesis conversion ratios under different VS ratio conditions.

On the other hand, in the fermentation process of FW3, FW4, FW5, FW6, and FW7, the organic matter produced by hydrolysis can quickly enter the stages of acid production and acetic acid production, and finally be converted into biogas. The conversion rates of the four stages are basically equal. In Figure 3, the hydrolysis conversion ratios in the system decreased with the decrease in apple feed components, and the influence of the hydrolysis process as the rate-limiting step became increasingly obvious. This is because WAS is a microbial matrix (floc) composed of microorganisms and exopolymers. These microbial-originated extracellular polymeric substances (EPS) are a complex mixture of biopolymers comprising polysaccharides, proteins, nucleic acids, uronic acids, humic substances, and lipids, amongst others [28]. It is well known that certain compounds in EPS are resistant to anaerobic digestion, resulting in hydrolysis as the rate-limiting step. Overall, when the VS ratio is 2:1, the problems of VFA accumulation and WAS hydrolysis rate-limiting can be alleviated at the same time.

3.1.3 Kinetics analysis

To further study the fermentation efficiency of different VS ratios, we chose FW2, FW3, and FW4 with higher cumulative biogas yields for kinetic simulation analysis. The results of the applied kinetic models are shown in Figure A1 (in the Appendix), while the calculated kinetic parameters are listed in Table 6. For FW2, FW3, and FW4, the correlation coefficients obtained by the Monod model are above 0.9914, 0.9992, and 0.9998, respectively. The correlation coefficients obtained by the first-order kinetic model were higher than 0.9901, 0.9991, and 0.9989, respectively. This indicates that the Monod model is best suited for FPW and WAS co-digestion of cumulative biogas production trends. The correlation coefficients of the two models of FW2 were lower, which was due to the slight fluctuation of the cumulative biogas curve due to acid suppression in the early stage of fermentation.

The kinetic constant K is an important index to determine the operating efficiency of the anaerobic reactor, which reflects the speed of the co-digestion system. In both the Monod model and the first-order model, the K (0.065 and 0.076 day⁻¹) of FW3 is the largest, indicating that the cumulative biogas production rate of FW3 is the fastest, which is consistent with the analysis in 3.1.2. The K (0.043 and 0.059 day⁻¹) of FW2 was the smallest because the generated excess VFA inhibited the subsequent acetogenic and methanogenic stages and reduced the overall velocity of the reaction system. Similarly, the K (0.060 and 0.074 day⁻¹) of FW4 is smaller than that of FW3, which is mainly due to the flocculation of sludge and the difficulty of cell wall degradation, which reduces the overall progress of the reaction system.

3.2.1 Continuous experiments performance

In order to further study the performance of continuous co-digestion for biogas production, the co-digestion of FPW and WAS was operated for 300 days over the five stages, and the biogas and methane production under different VS ratios (FPW:WAS) is illustrated in Figure 4. During the whole co-digestion process, as the OLR increased from 1.80 to 2.88 kg VS·m⁻³·day⁻¹, the MPR rose from 0.6 to 0.89 L·L⁻¹day⁻¹, and biogas production increased from 1.16 to 1.48 L·L⁻¹day⁻¹. The methane yield is always above 300 mL·g⁻¹, indicating that the continuous anaerobic fermentation process can achieve high-rate co-digestion with FPW and WAS. The pH was maintained between 6.76 and 7.29, and VFA/TA was below 0.16, indicating the reactor was always in a steady state.

Figure 4

Changes in the pH, TA, VFA, NH 4 + –N, methane yield, MPR, biogas production, and CH₄ content during acidic FPW and WAS co-digestion.

In Run 1, the average methane yield was 354.51 mL·g⁻¹, but the pH and TA continued to decrease. This is because FPW is easily hydrolyzed organic matter, and the rate-limiting step of its anaerobic fermentation is methanation, so acid accumulation will occur and the TA consumption in the buffer system will be increased. This result is similar to the batch experiment. In Run 2, the average methane yield was 351.87 mL·g⁻¹, and the downtrend of pH in the first 20 days was alleviated, but the downtrend of pH and TA in the reactor remained unchanged with the accumulation of VFA. Similarly, Fonoll [29] found the co-digestion experiment with apple waste and sewage sludge at a mass ratio of 30:70 (ww/ww, VS ratio of 3:1) was also showed the increase of VFA, a decrease of TA and methane production within the first 20 days, and the average methane yield was 260 mL·g⁻¹. In particular, on the 60th and 120th days, the ammonia nitrogen concentration in the reaction system decreased from the initial 227.54–138.11 and 123.67 mg·L⁻¹, respectively. The C/N ratio in the feed was 22.2 and 19.8, respectively, which met the C/N requirements in the general anaerobic fermentation process [39]. However, because WAS is not easily degraded, the consumption of ammonia nitrogen is less than the production, resulting in insufficient nitrogen source supplementation during the reaction.

In Run 3, the VFA was significantly reduced from 533.45 to 438.12 mg·L⁻¹, the TA content increased from 3,755.67 to 3,958.32 mg·L⁻¹, and no acid accumulation occurred. At this time, the average methane yield was 346.69 mL·g⁻¹ and did not change significantly from the previous two stages, indicating that FPW and WAS maintained high hydrolysis efficiency. The ammonia nitrogen content increased from 123.67 to 138.34 mg·L⁻¹, so the problem of nitrogen source consumption was solved. In Runs 4 and 5, as the addition of WAS further increased, the pH in the system rose to 7.37, and the buffering capacity was significantly improved, but the average methane yield decreased significantly to 310.21 and 299.07 mg·L⁻¹ respectively. This is because although co-digestion of FPW and WAS can improve the hydrolysis efficiency of WAS, excessive WAS can reduce the hydrolysis efficiency.

During the whole fermentation process, the methane content increased slightly with a rise in the WAS proportion. This is because WAS is rich in proteins and fats, and the biogas produced by its decomposition contains a higher percentage of methane [39]. This result is the same as that of the batch experiment.

3.2.2 Microbial community dynamics

The species and abundance of bacteria have a close relationship with the hydrolytic and acidogenic capacity to degrade complex organic matter during CSTR digestion. Relative abundance (RA) and distribution of bacterial communities at 120, 180, and 240 days during continuous fermentation were analyzed at the phylum and genus levels. Four main phyla were selected (Figure 5a), including Firmicutes, Bacteroidetes, Proteobacteria, and Chloroflexi. The RA of Firmicutes increased in the whole digestion, which was 10.2%, 24.3%, and 31.9% at 120, 180, and 240 days, respectively. This phylum is one of the major hydrolytic bacteria that can decompose substrates into SCOD and especially degrade proteins and polysaccharides into soluble organic matter [40]. It helps to promote the hydrolysis of incremental sludge in the feed. Bacteroidetes can degrade complex compounds such as celluloses and hemicelluloses and convert them into glucose and organic acids and further to VFAs [41]. On the 180th day, the RA of Bacteroidetes increased to 21.3%, which helped to improve the hydrolysis efficiency of long-chain compounds. Proteobacteria and Chloroflexi are commonly found in the decomposition reaction of monosaccharides, polysaccharides, and small-molecular compounds to form acetic acids [42]. With the increase in WAS content, the proportion of carbohydrates in the substrate, and the RA of Proteobacteria and Chloroflex all decreased, which are also consistent with the gradual decrease in VFAs.

Figure 5

The RA of bacteria communities at (a) the phylum and (b) genus levels.

The reasons for the changes in hydrolysis efficiency under different VS ratio can also be explained by analyzing the changes in microbial community structure at the genus level. On the 180th day, the RAs of Bacteroides, Paludibacter, Litorilinea, Levilinea, and Smithella reached 5.4%, 5.4%, 5.0%, 7.9%, and 3.8%, respectively (Figure 5b), which were higher than those on Day 120 and Day 240. Among them, Bacteroides and Paludibacter are carbohydrate-fermenting bacteria that promote the utilization of carbon sources such as cellulose and hemicellulose during hydrolysis and acidification [43]. Litolinea can utilize proteins to facilitate WAS hydrolysis. The monosaccharides from the hydrolysis of carbohydrates, peptides, and amino acids from the decomposition of protein can be decomposed into volatile organic acids by Levilinea [44]. At last, Smithella can degrade butyrate and propionate and maintain the stability of the acetogenic methanogenesis reaction [45]. The enrichment of the above-mentioned bacterial groups is conducive to the simultaneous hydrolysis and acidification of carbohydrates, proteins, and other substrates, which can accelerate the feed hydrolysis rate and reduce the impact of organic acids in an anaerobic system.