Detection of early imbalances in semi-continuous anaerobic co-digestion process based on instantaneous biogas production rate
Introduction
Anaerobic digestion (AD) is the microbial degradation of complex organic matter producing a biogas, mainly composed of methane (CH4) and carbon dioxide (CO2) and a digestate. It is a multi-stage bioprocess that requires the interaction of at least four trophic groups of microorganisms associated with steps such as: hydrolysis, acidogenesis, acetogenesis and methanogenesis.
The last step, methanogenesis, is the most critical part of anaerobic degradation with respect to inhibition and variations in the process conditions (pH, temperature, alkalinity, hydraulic retention time (HRT), substrate, etc.) (Rajagopal et al., 2013). Methanogens, strict anaerobes classified as Archaea, allow the production of methane from acetate or hydrogen (H2) and carbon dioxide, according to two main metabolic pathways: hydrogenotrophic and acetoclastic methanogenesis. The efficient conversion of organic matter into methane in AD depends on the mutual and syntrophic interactions of functionally distinct microorganisms (Akuzawa et al., 2011). For a stable operation the process must be well balanced for a successful use of the intermediate by-products by the bacterial consortia of each stage. However, failures do occur in anaerobic digesters and are mainly due to these variations in the process conditions and the accumulation of inhibitors (volatile fatty acids (VFA), ammonia …). The microbiological mechanisms underlying these inhibitions have been the subject of numerous studies.
CH4 can be produced by acetotrophic methanogens (AM) and hydrogenotrophic methanogens (HM). As illustrated in Fig. 1, CH4 formation from H2/CO2, produced by the previous steps, is promoted by the activity of HM, whereas CH4 formation from acetate can be achieved through two different mechanisms: (i) direct cleavage of acetate by AM or (ii) syntrophic acetate oxidation (SAO) to H2/CO2 by syntrophic acetate oxidizing bacteria (SAOB) followed by the subsequent conversion of these products into methane by HM (De Vrieze et al., 2012).
It was generally accepted that the acetoclastic pathway is responsible for 60%–70% of methane production in a digester. However, well established molecular methods indicated that the microbial population structure in a digester depends on environmental conditions and on the characteristics of the substrate, as well as on the temperature in the digester (Karakashev et al., 2006; Nettmann et al., 2010; Franke-Whittle et al., 2014). Accurate dynamic quantification of the prevailing microbial activity showed that, at high VFA and ammonia concentrations, the main acetate degradation mechanism was SAO (Gehring et al., 2015; Goux et al., 2015; Frank et al., 2016; Mosbæk et al., 2016; Westerholm et al., 2016). These results were obtained using different substrates and digestion conditions: a reactor fed with sugar beet pulp with VFA accumulation up to 10,589 mg/L (Goux et al., 2015) or a reactor spiked with acetate at 5900 mg/L (Mosbæk et al., 2016). Frank et al. (2016) also reported a reactor fed with slaughterhouse and municipal waste containing 1300 mg/L of VFA and 360 mg/L of NH3. A review of conditions favoring the SAO mechanism reported several and different ranges of NH3/NH4+ and VFA depending on the experimental conditions used in the study (Westerholm et al., 2016). These conditions caused a shift from the acetoclastic degradation pathway to SAO followed by hydrogenotrophic pathway (SAO-HM).
The shift from AM to SAO-HM in ammonia-induced digestion was even found to maintain stability and compensate for biogas production in digesters (Werner et al., 2014). It was shown that from 60% to 75% of methane was produced via the SAO-HM pathway during mesophilic digestion of samples with high ammonia content (244–1214 mgN-NH3/L) (Jiang et al., 2018). Up to 89% of methane was derived from CO2 and H2 reduction in a thermophilic reactor containing high levels of acetate (3840–5900 mg/L) (Hao et al., 2011). Indeed, high abundance of SAOB in different digesters despite differences in their operating parameters and conditions, demonstrated that they are an enduring and important component of biogas-producing consortia (Sun et al., 2014). A particular case is lipid-rich substrate digestion which leads to degradation of long chain fatty acids (LCFA) (Ma et al., 2015). Although SAO was not found to contribute to acetate degradation in a reactor fed with fats, oils and grease (70% VS) (Kurade et al., 2019), a shift from AM to HM activity was reported in lipid-enriched reactors (1.5–2 g of oleate added in the feed) due to the prevalence of HM along with hydrogen-consuming bacteria in these thermophilic (Baserba et al., 2012; Amha et al., 2017) and mesophilic (Regueiro et al., 2016) reactors.
Different threshold tolerances to ammonia, VFA or LCFA found in the literature during the pathway shift, were highly dependent on the substrate and on digester operation. Several studies highlighted a temporary adaptation of the prevailing microbial community to ammonia (Westerholm et al., 2016), VFA (Li Y. et al., 2018) and LCFA (Amha et al., 2017). AM are reported to be more severely affected by these conditions than HM, which have a similar, or in some cases, higher tolerance to high ammonia levels than SAOB (Wang et al., 2015). In all cases, a decrease of AM microbial activity always affects AD process. All these results show that three pathways (AM, HM and SAO-HM pathways) could occur in a digester to produce methane. However, even if they probably co-exist in digesters, only one is responsible for most of the methane production, depending on the process conditions. When these digestion conditions change, a shift between pathways could occur and could cause instabilities and eventually, failures of the digester.
Anaerobic co-digestion (AcoD), which entails the simultaneous digestion of two or more feedstocks, is widespread in several countries, allows biogas production enhancement and maximizes waste valorization. In AcoD cases, the characteristics of co-substrates (type, quality and quantity) generally vary over time. Therefore, the variations in the feedstocks induce variations in the process conditions which favor pathway shifts in the digester. Monitoring the microbial populations along with their activity and/or growth could be used to monitor AcoD processes and provide an early warning of imbalances (Poirier et al., 2016). A recent study showed that conventional or innovative early-warning indicators varied among different types of AD systems (with different operating conditions) and mainly focused on biogas and its associated parameters and, digestate compositions inside digesters (Wu D. et al., 2019). These indicators are: pH, VFA (total and individuals VFA), alkalinity (partial, total and bicarbonate), biogas composition, methane yield, the concentration of specific intermediate metabolites (ammonia, glycerol, aromatic compounds, H2, carbon monoxide, etc.), the composition of the microbial community and enzyme activity (Lebuhn et al., 2015; Li et al., 2017; Poirier et al., 2016). There are also coupled indicators such as ratios between intermediate and partial alkalinities, VFA and alkalinity (VFA/Alk) or VFA and total inorganic carbon that performed better than individual indicators (Lebuhn et al., 2015). A recent review exploring the effectiveness of these indicators as well as an extensive study of their threshold values showed that no unique indicator and threshold value were applicable to all biogas plants (Li L. et al., 2018). Moreover, the proposed indicators were only effective for specific substrates and specific operating conditions because of their sensitivity to the process conditions and substrate compositions (Li et al., 2017, Wu et al., 2019b). In addition, the specificities of the determination of some parameters also made it difficult to use them at full-scale plants. Therefore, several studies focused on developing online sensors for VFA, ammonia and hydrogen determination mainly using spectroscopy, chromatography or automatic titration for VFA/Alk (Nguyen et al., 2015). However, experience in using these sensors in full-scale plants is limited. Robustness, simplicity, accuracy and reliability are some of the key parameters that any operator will look for when choosing one of these sensors and/or indicators for process monitoring. As none of these indicators was universally valid for all AD systems, more work is needed to design an indicator and monitoring tool that meet AD plant operators’ requirements.
A shift in metabolic pathway appears to be a relevant indicator of the state of the process. Determination of the pathway shift could be used as a monitoring strategy of the process. However, the existing methods for the determination of metabolic pathways (carbon isotope analysis, molecular methods, etc.) and subsequent pathway shifts are complex and time-consuming. In this context, the aim of this study was to develop a monitoring tool to diagnose imbalances in methanogens activity as early as possible in order to avoid failures of the digester. As biogas and methane production are directly impacted by changes in methanogens, their dynamics could reflect methanogens activity. The tool developed in the present study is therefore based on biogas production rate (BPR) and associated kinetics. A principal component analysis (PCA) model was built based on BPR kinetics obtained during normal operating conditions of the digester. This model was evaluated during the unstable periods of the experiments to determine its ability to rapidly identify the shift between degradation pathways and hence to provide early warnings.
Section snippets
Pilot-scale digester
The pilot-scale digester was a continuously stirred tank reactor (CSTR) with a working volume of 35 L and was maintained under mesophilic conditions (38 °C). The reactor was mounted on a scale for the monitoring of its weight, therefore enabling control and recording of the doses of feed. The feed substrate, prepared every week, was kept in a refrigerator (4 °C) containing a hopper placed above the reactor to allow gravity feeding. Before each feed, the mixed substrate was recirculated from the
Digestate characteristics
The characteristics of digestate samples collected from each experiment are listed in Table 2. TS were similar in each series of experiments (1.5–2.1%), despite variations in the TS of the substrate mixture (2.7–6.5%). This indicates the high biodegradability of the co-substrates added to PS. There was no significant accumulation of matter in the digester. TKN and NH4+ showed similar variations. A significant increase in TKN and NH4+ was observed during experiment B2 due to the addition of PW
Conclusion
Methanogenic pathway shifts occurred in digesters at high concentrations of LCFA, VFA and ammonia. This study was conducted in order to detect these shifts as an early sign of instability in AcoD processes. A PCA-model was built based on the BPR kinetics of a digester operating under normal conditions and the acetoclastic methanogenesis was therefore the dominant pathway. The model PCs scores were used to evaluate its accuracy in predicting instabilities in two inhibition induced experiments.
Funding source
L. Awhangbo is the beneficiary of a PhD scholarship funded by Irstea and the Bretagne Region. This project was also funded by the French Environment & Energy Management Agency (ADEME) (COMET project N°1606C0010).
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References (49)
- et al.
Elucidating microbial community adaptation to anaerobic co-digestion of fats, oils, and grease and food waste
Water Res.
(2017) - et al.
Effect of continuous oleate addition on microbial communities involved in anaerobic digestion process
Bioresour. Technol.
(2012) On reducing false alarms in multivariate statistical process control
Chem. Eng. Res. Des.
(2010)- et al.
Methanosarcina: the rediscovered methanogen for heavy duty biomethanation
Bioresour. Technol.
(2012) - et al.
Investigation into the effect of high concentrations of volatile fatty acids in anaerobic digestion on methanogenic communities
Waste Manag.
(2014) - et al.
Anaerobic digestion of swine manure: inhibition by ammonia
Water Res.
(1998) - et al.
Quantifying the percentage of methane formation via acetoclastic and syntrophic acetate oxidation pathways in anaerobic digesters
Waste Manag.
(2018) - et al.
Acetoclastic methanogenesis led by Methanosarcinain anaerobic co-digestion of fats, oil and grease for enhanced production of methane
Bioresour. Technol.
(2019) - et al.
Transfer of a static PCA-MSPC model from a steady-state anaerobic reactor to an independent anaerobic reactor exposed to organic overload
Chemometr. Intell. Lab. Syst.
(2016) - et al.
Instability mechanisms and early warning indicators for mesophilic anaerobic digestion of vegetable waste
Bioresour. Technol.
(2017)